THE

ART OF TUNNELLING
by
KAROLY SZE.CHY
Member of the Hungarian Academy of Sciences
Professor of Foundation Engineering and Tunnelling
at the

Tcchni:al University f or Civil and Commu nication Engineering

Bu:lapest

.,,.--- ----

111Iiilil11
K150928

AKADEMIAI KIAD6 · B UDAPEST 19 70

1988 808338
The original
ALAGUTEPiTESTAN

published by Tanki:inyvkiad6, Budapest 1961

First English edition: Akademiai Kiad6, Budapest 1966

Reprinted 1967, 1970

Translated from the Hungarian

by
Denis SzkHY, I van L1ESZKOVSZKY, Zoltan Sz1LVASSY, Miklos EsZT6

© Akademiai Kiad6, Budapest 1970

Printed in Hungary
 ,,
.'•
··.•

Dedicated to MY WIFE
helpful and affectionate consort of my life and work
PREFACE TO THE ENGLISH EDITION

It has been a very flattering satisfaction to the Author that since this book was
published in Hungarian, it has met with a favourable reception, not only in
Hungary, but also in foreign countries. This has proved, on the one hand, that
the book was meeting the actual demands of related Civil Engineering Practice
and, on the other hand, it was encouraging for its publication in a foreign language.
I am grateful to the Publishing House of the Hungarian Academy of Sciences
for having kindly undertaken this English publication because it makes the book
accessible to interested engineers in a very great part of the world. Minor revisions
and the addition of some supplementary material seemed to be necessary in this
English version, of course, but it represents the translation of the Hungarian
edition, as a whole.
I should like to express my thanks here especially to Mr Denis Szechy and
to Mr Ivan Lieskowsky (Toronto) for their careful and excellent translation and
to Mr Franklyn North for the careful revision of the English manuscript.
Dr Miklos Eszt6 and Mr Zoltan Szilvassy have also participated in the work of
translation. English inscriptions on the figures and additional drawing work was
done by Mr Tibor Manyai whereas typing and revision was again done by Mrs
Julia Tompa.
Budapest, 1966
K. Szechy
 PREFAC E TO THE HUNGA RIAN EDITION

Tunnels represent the most expensive engineering structures where in addition

the greatest difficulties may be encountered during achievement. This may account
for the fact that since the completion of the major railway networks in Europe
and North America i.e. since the second half of the last and the first decades of
the present century, the technique of tunnelling did not keep pace with the devel-
opment in other technical fields and the literature of tunnelling has also shown
the signs of a certain decline.
• After the second World War however the large-scale construction of hydraulic
power-stations just as the rapidly increasing tasks of urban communicat ion and
, utility conduits as well as the defence of big industrial plants - against war
perils - again lent a new impetus to' subterranean construction in general. It must
be considered, on the other hand, that during the elapsed period construction
technique (that of tunnel construction included) had undergone a very significant
development. The adyent of new construction materials, technique and equipment
had to transform tunnelling methods, too. This experience, however, was accessible
only in short articles or reports scattered in various technical journals and publi-
cations.
In Hungary recent progress was launched by the start of the construction of the
new Budapest Subway in 1951. To make modern methods and structures accessible
to the engineers concerned, a compilation of studies and articles was published
by Kozlekedesi Kiad6 (Communica tion Publishing House) in two volumes in
• 1952-54 under the title: "Selected chapters from tunnel and mining engineering"
which included relevant experience from Mining, too. Additional material was
published in some special publications , and in some leaflets of the Institute for
Advanced Engineering Education (Mernoki Tovabbkepz o Intezet). About this
time the valuable literature dealing with the construction of the Soviet Subways
became available, and this was rendered more applicable and was enlarged by
their direct interpretatio n by the visiting Soviet Metro experts. About this time
more attention was given in academic education, too, to the science of tunnel
construction at the Technical University of Budapest. Nevertheless, no compre-
hensive book covering the whole field of the art of tunnelling has been published,
either in Hungary or abroad. This book, of course, cannot claim to be completely
comprehensive in this vast field of civil engineering science; it may be regarded
rather as an attempt to give a fair information about its actual position as regards
10 PREFACE

theory, design and execution based on a critical compilation of recent literature,

on personal experience and information gained at the realization of the new
Budapest Subway· and on the lectures delivered by me for about a decade at the
Technical University of Budapest. An attempt has been made to complete the
book with up-to-date references to enable the reader to find the details lying in
his further special interests.
A warm expression of .gratitude is due to all ~ho helped me in the preparation
of the manuscript either by active participation o"r by-putting some data, photos
and designs to my disposal. This is due in the highest degree to Chief Engineer
Gyorgy Fazakas and to Dr Vilmos Vincze, who through the numerical presen-
tation of the Soviet dimensioning methods in Chapter 4, respectively through the
presentation of the manuscript of Chapter 5 (Surveying of tunnels) took a valu-
able active share in the contents of this book. My thanks must be extended to
the readers of the book i.e. to Director Istvan Gyorgy and to Superintendent
Gusztav Almassy, not only for their revisional work, but also for their kind advice
and suggestions. Further, thanks are due to my direct helpers: to Mrs Julia Tompa
for the typing and to Mr Laszlo Zsolnay for the drawings and· also to Dr Laszlo
Rozsa, to Assistant Professors: Geza Petrasovits, Laszlo Szijarto, and Dr Laszlo
Varga for their assistance and, finally, to the staff of the Publishers and Printers:
Tankonyvkiado Vallalat and Allami Nyomda for their conscientious work and
for the pleasing layout of the book.
Budapest, August 1961
K. Szechy
 CONTENTS

CHAPTER I

INTRODUCTION 19

1.1. Purposes and classification of tunnel5 21

1.1 L Traffic tunnels . . . . . . . 22
11.11. Classification by position or alignment . 22
11.12. C lassification according to purpose 26
1.12. Transportation tunnels 30
11.21. H ydroelectric plant tunnels 30
11.22. Water supply tunnels 31
l 1.23. Public utility tunnels . 32
11.24. Sewer tunnels '.i2
11.25. Other tunnels 33
l.13. Storage tunnels (garages, parking areas, shelters and storehouses) 33
1.2. A brief history of tunnel construction . . . . . . . . . . . . . 36

C HAPT E R 2

PRELIMINARY STUDIES AND GENERAL DESIGN

CONSIDERATIONS 47

2.1. Preliminary studies . . • . . 47

2.11 . Economic analysis . . . 47
2.1 2. Geological survey and exploration 52
21.21. General geological survey 55
21.22. Detailed geological site investigations prior to planning 60
21.23. Site exploration during design . . . . . . . 62
21.24. In situ exploration during construction . . . . . . 63
2.13. Typical geological factors and their effect on tunnelling . . . 64
21.31. The situation and orientation of layers to be penetrated 64
21.32. Condition of strata to be perforated . . . . . . . 68
21.33. Stress, strength and deformation properties of rocks 70
21.34. H ydrological survey . . . . . 79
·, 21 .35. Gases and rock temperatures 84
2.14. Geological profile along the tunnel axis 94
12 CONTENTS

2.2. Factors influencing the location of the tunnel . 106

2.21. Selection of the line . . . . . . . . 106
2.22. Selection of the longitudinal elevation and gradients 110
2.23. Determination of the cross-section . . . . . . . 117
22.31: Tunnel clearance . . . . . . . . . . . 123
22.32. Influence of geological environn;ient on shape of cross-section 124
22.33. Influence of construction method on shape of cross-section . 128
22.34. Influence of tunnel lining on shape of cross-section . 128
22.35. Size of the tunnel cross-section . 129
References . . . . . . . . . . . . . . . . . . 132

CHAP'!'ER 3

ANALYSIS OF LOADS ON TUNNELS AND UNDERGROUND

STRUCTURES . . . . 133

3.1. _Causes and types of rock pressure 133

3.11. Rock pressure due to loosening 136
3.1-2. Genuine mountain pressure 148
3.13. Swelling pressure . . . . . 15 6
3.14. F actors affecting the magnitude and local types of rock pressure 15 7
'
3.2. Determination of vertical rock pressures. Rock pressure theories 15 9
3.21. Estimates and approximate methods based on the extent _o f upbreak . 15 9
3.22. Theories based on theoretical stress conditions in the rock mass . 166
32.21. Theory of elasticity and estimation of stresses . . . . 16 7
32.22. Results of investigations by Fenner . . . . . . . . . 171
32.23. Stress conditions around circular_and elliptical cavities . . 18 3
3.23. Theories based on various displacement and equilibrium assumptions 191
32.31. Theories taking into account the effect of depth 191
.
32.32. · Theories neglecting the effect of depth .
3.3. Determination of lateral pressu!es on tunnels- . ._ .
.
. 208
. 219
3.31. Approximate determination of lateral pressures 219
3.32. Exact determination of lateral pressures ·. . . 22 1
3.33. Experimental determination and in situ measurement of lateral pressur~s . 221
3.4. Bottom pressures . . . . . . . . . . . . . . . . . . . . . . 223
3.41_. Determination of bottom pressure according to Tsimbaryevitch 225
3.42. Determination of bottom pressure according to Terzaghi . . . 227
3.5. Development and superposition of rock pressures . 230
3.51. Development of rock pressure during excavation . 230
'3.52. Superposition and interaction of roof loads above adjacent tunnels 234
3.53. Loads on rock-pillars . . . . . 238
35.31. Protodyakonov's, theory . 239
35.32. Tsimbaryevitch's theory . 240
 CONTENTS 13

3.6. Critical treatment of rock pressure theo_ries and in situ rock pressure measurements 241
3.61. Measurements on· the rock face . . . . . 243
3.62. Measurements in the interior of the rock . 245
3.63. Measurements on tunnel supports· . 246
3.64. Measurement of pressure changes in completed linings . 248
3.65. Determination of pressures by model tests . 255
3.66. Main types of pressure measuring instruments 256
3.7. Water pressure 259
3.8. Live loads 262
3.81. Internal loads . 262
3.82. Surface loads . 263
References . 263

CHAPTER 4

DESIGN OF TUNNEL SECTIONS . 264

4.1. Design loads . 267
4.11. Estimated design loads for deep tunnels in solid ground . 267
4.12. Estimated design loads for shallow tunnels in loose, saturated soil . 269
" 4.13. Excerpts from the specifications for the design loads of the Lisbon subway . 270
4.14. Excerpts from the specifications for the design loads of the Budapest subway 272
4.15. Soviet standard specifications for the design of underground and motorway
tunnels . . . . . . . . .- 277
4.2. Design of horseshoe-shaped tunnels . 281
4.21. Design by members 281
42.11. Graphic investigation (Kommerell) . 282
42.12. Analytical design by members . 291
42.13. Common deformations and composite action of the surrounding
ground (Davidov's method) . . . . . . . . . . . . . . . . 316
4.22. Analytical design method treating the section as a whole and considering com-
posite action of the ground . . . . . . . . . . . . . . 321
42.21. Method of Zurabov and Bougayeva . . . . . . . . . 321
42.22. Numerjcal example of the Zurabov-Bougayeva method . 326
4.3. Design of circular tunnel sections . 341
4.31. Graphical investigation . 341
4.32. Approximate method for the calculation of ring sections . 342
43.21. Design by dividing the section into segments . 342
43.22. Design of a monolithic ring section . . . . . . 343
43.23. The Hewett-Johannesson method . . . . . . 356
4.33 . The design of circular tunnels as sections elastically embedded in ·the subsoil 365
43.31. The method of Bodrov-Gorelik . 365
43.32. The polygonal method . 378
43.33. Bougayeva's method 396
43.34. Davidov's method . 405
 14 CONTENT S

43.35. Varga's develop ment

43.36. Meissne r's and Orlov's methods 408
410
4.34. Design of tunnels with double lining .
412
43.41. Soviet specifications (based on Galerkin 's theory)
. 412
43.42. Design with steel plate lining (design after Miihlho fer)
43.43. Design of tunnels with laminate d linings . . . . . . 415
43.44. Design of laminate d linings for non-rad ial (external) loads 418
420
4.4. Design of culverts and condllits
. 423
4.41. D esign of circular culverts
. 424
44.11. Determi nation of vertical earth pressure (accordi ng to Yaroshe
nko) . 424
44.12. Determi nation of traffic loads . . . .
44.13. Effect of the bedding . . . . . . . . 433
. 434
44.14. Cross-se ctional design of culvert sections
. 435
44.15. Longitu dinal design of culverts . . . .
. 438
4.42. Design of circular conduits . . . . . . . .
4.43. The semi-graphical design of egg-shap ed culverts . 442
4.44. Flexible metal culverts of corruga ted steel sheets 444
. 454
4.5. Design of rectangu lar tunnel sections • . . . . . .
. 459
4.51. One bay rectangu lar section on rigid foundat ions
4.52. Two bay box section on elastic foundat ions . 459
. 462
4.6. Design directives for ancillary works, services and installat ions_
480
4.61. Compos ition and structtlr e of tunnel walls
4.62. Waterpr oofing of tunnels . . . . . . . 480
481
46.21. Multi-layer plasterin g and shotcret e
. 483
46.22. Gunite plaster . . . . . · . . . .
. 485
46.23. Bonded waterpro ofing . . . . . .
. ,486
46.24. Reinforc ed bitumin ous waterpro ofing
. 487
46.25. PVC waterpro ofing . . .
. 487
46.26. Use of thermop lastic sheets
489
4.63. Drainag e of tunnels . . . . . . .
4.64. Protecti on against corrosio n 490
. 493
46.41. Problem s of corrosio n by soil
46.42. Ground water corrosio n . 493
. 498
4.65. Tunnel ventilati on
. 498
46.51. Ventilat ion during construc tion
. 499
46.52. Natural ventilati on . . . . .
. 503
46.53. Mechan ical ventilati on . . . .
. 505
4.66. Lighting , and noise control in tunnels
519
46.61. Lighting
46.62. Noise control 519
520
4.67. Ancillar y works . .
520
46.71. Safety recesses
46. 72. Portals . . 521
. 521
Referenc es
. 523
 CONTENTS J5

CHAPTER 5

TUNNEL SURVEYING 524

5.1. Surveying from portals 524

5.1 J. Horizontal layout 524
51.11. Layout of short tunnels . 524
51.12. Layout of long tunnels . 526
5.12. Marking the stations . . . . . 532
5.13. Layout of details . . . . . . 533
5.14. Methods to improve accuracy . . 535
5.15. Vertical layout . . . . . . 537
5.2. Layout of .tunnels starting underground 538
5.21. Horizontal layout . . . . . . 538
52.11. Surface reference net 539
52.12. Transfer of directions down shafts . 544
52.13. Underground directional layout . 556
52.14. Possibilities for increasing the accuracy of the layout 559
5.22. Vertical layout . . . . . . . . . . . . . 559
5.23. Layout of ring sections (segments) and shields 561
52.31. Layout of segments . . . . . . . 561
"52.32. Determining the position of the shield 564
References . 566

C HAPTER 6

CONSTRUCTION AND DESIGN.OF TUNNELS 568

6.1. Tunnelling in solid rocks . . . . . . . 569

6.11. Means of excavation in solid rocks . 569
61.11. Mechanical drilling and cutting · 569
61 -12. The use of explosives and blasting technique 571
6.12. Full-face tunnelling without supports . . . . . . 581
6.13. Full-face tunnelling with supports . . . . . . . . 584
61.31. Temporary support independent of the permanent lining . . 586
6.2. Tunnelling in moderately firm rocks and ground . . . . . . . . . . 609
6.21. Tunnelling by successive excavation and lining of smaller independent head-
ings (classical or mining methods) 609
62.11. Scope and construction of headings . . . . 61 5
62.12. Vertical break-ups and pits .- . . . . . . 629
6.22. Single-stage mining methods of tunnel construction 630
62.21. The crown bar (English) method . 630
62.22. The cross-bar (Austrian) method . . . . . 635
 16 CONTENTS

62.23. The alternate ring method 638

62.24. The centre-cut method . 638
6.23. Multiple-stage classical methods . 639
62.31. The Belgian or underpinning (flying arch) method . 639
62.32. The core-leaving or German method . 648
62.33. The Italian or invert method . . . . . . . . . 652
62.34. Combined tunnelling methods . . . . . . . . . 654
6.24. Recent tunnelling methods employing both timber and steel-supporting elements 65 6
62.41. Tunnelling with liner plates (the nee9le beam method) . 657
62.42. The full-face erector method . . . . . . . 658
62.43. Tunnelling with liner plates and stiffening rings . 658
62.44. The Kunz method . . . . . . . . . . 664
.62.45. The Col,o gne method . . . . . . . . 665
6.3. Tunnelling in loose ground and under watercourses . 668
6.31. Tubnelling by sinking caissons . . ._. . . . 673
63.11. Sinking caissons in the form of working chambers . 673
63.12. The floating caisson·method . 675
6.32. Shield tunnelling . . . . . . . . 678
63.21. Structure and dimensions of tunnel shields . 681
63.22. The main working procedures of shield tunnelling . 699
63.23. Lining segments for shield-driven tunnels . 738
63.24. Application of compressed air (the plenum process) and soil stabilization
in tunneUing . . . . • . . . . . . . , . . 771
6.33. Special underground structures and railway stations . 783
63.31. Access or ventilation shafts . . . . . . . . 784
63.32. Connecting drifts (ventilation ducts) . 795
63.33. Underground halls (shield chambers, power and substation rooms, etc.) 797
63.34. Underground stations and adjoining service and communication
localities . . . . . . . . . . . . . . . . . . 799
6.4. Urban and public utility tunnels constructed at a shallow depth . . . . . . . 811
6.41. Si,i~:wall construction in wall drifts . . . . . . . . . . . . . . . . 812
_ 6.4.J.,8pec1at.~ethods for the construction of pedestrian subways, highway and public
. ~ --- service line ,underpasses . . . . . · 812
64.21. Subways in ~ienna . . . . 814
64.22. Margaret brid~e in Budapest . . . 817
64.23. Dimitrov Squ:fre subway in Budapest 818
6.43. Construction of culver,.ts and sewers by pipe jacking 819
6.44. Cut and cover constru'ction method with precast panels . . 824
\
6.5. Safety measures and health p}~~ection in tunnelling . 828

6.51. :safety measures . . S28

6.52. Health protection . 329
65.21. Silicosis . 330
65.22. Caisson disease . . i\31
References .
.. \•
·,,, . 'i . 333
 ....

CONTENTS 17
CHAPTER 7
8
8
SERVICE, OPERATION AND MAINTENANCE OF TUNNELS . 835
9
~ 7.1. Organization and responsibilit_ies of maintenance and service departments 835
~
7.11. Inspection of the shape of the tunnel . . . . 835
z 7.12. Inspection of the track drainage and insulation . 838
I
71.21. Inspection of the tracks . . . . . . . 838
71.22. Inspection in the Budapest Subway . . . 839
7.13. Inspection and maintenance of lining and tracks . 840
7.14. Organization of maintenance services . 842
7.15. Storage of maintenance equipment . 843
7.2. Deterioration and repair of tunnels . . 843
7.21. Causes of deterioration of tunnels 844
72.11. Deterioration due to defective materials and workmanship . 844
72.12. Deterioration caused by water . . . . 844
72.13. Damage caused by smoke .. . , . . . 846
72.14. Damage due to atmospheric conditions . 846
72.15. Impact damage . . . . . . . . 846
72.16. Damage caused by overburden . . . 847
72.17. Deterioration in pressure tunnels . . 847
72.18. Damage to railway tracks 848
72.19. Maintenance of cables . . . 848
7.22. Repair and reconstruction of tunnels 848
72.21. Repairs to the drainage system 849
72.22. Repair of linings • . . . . . 857
7.23. Remodelling and reconstruction of tunnels . 861
72.31. Reconstruction for operational demands . 861
72.32. Reconstruction because of decay and external effects . 862
72.33. Examples of tunnel remodelling . . . . . 864
7.3. Surface subsidences resulting from tunnel construction . 869
7.31. Estimation of surface subsidences . . . . . . : ·.. .
• jl
. ~~- -· . 8~
..
73.11. Estimating surface subsidence on a theoretical basis . 870
73.12. Estimating subsidence from actual observations· . . 877
7.32. Measuring the -subsidence of buildings due to su~way construction 879
References 882
General literature ,. 883
Index . ..
. ·' . 885

2 Szec:hy: The Art of Tunnelling

 ..,

CHAPTER 1

INTRODU CTION

Even in ancient times underground structures presented a challenge to Man.

The arduous labour and considerable danger associated with underground con-
struction was at all times justified only by exceptional reasons of defence, assault,
production, (transportati on) or traffic.
Apart from natural caves which offered shelter to primitive man, tunnels driven
to undermine, or underpass fortifications constituted an important, and in many
instances effective, means in ancient warfare. Various secret underground passages·
into and from fortresses in the Middle Ages, as well as the assault tunnel systems
built for undermining and blowing up fortress-walls in more recent military
history have played similarly important roles.
Undergroun d structures have retained their military significance to this very
day, in spite of the change in their character from an o ffensive to a defensive one.
Thus, the role of shelters affording protection against air raids and of "bomb-
proof" underground plants, power stations, factories and warehouses is still
considerable in modern warfare.
Second to military purposes was the extraction of valuable mineral resources
from the hidden depths; in other words, major systems of considerable extension
became necessary for mining purposes. It will be recalled that the origins of ore
mining in particular can be traced back to the beginnings of human history.
Besides military and mining purposes tunnels were often incorporated in water
supply and drainage systems. The tunnels introducing domestic water into some
of the ancient towns (Jerusalem , Athens, Rome) concealed to a certain extent
from the enemy, served the first conveyance function.
Neither could man dispense with the benefits afforded to traffic by underground
structures. Starting with the secret underground passages mentioned above,
underground traffic routes, i.e. tunnels constructed to avoid surface obstacles
and for a wide variety of traffic and transportatio n purposes, became an indispens-
able feature in modern transportatio n systems.
Tunnels, therefore form an important section of subterranean structures, as they
can be defined as underground passages constructed for the purpose of direct
traffic, o r a transportatio n connection between two points. Simply defined,
tunnels are " underground passages made without removing the overlying rock
or soil".
Undergroun d structures can be classified broadly into four categories:

2•
 20 INTRODIJC T.ON

(a) The most ancient are those bl.tilt for the mining industry. These comprise
in part the permane nt transport ation network e?(cavated to final dimensions and
provided with a permanent lining, and in part of the interconnected exploitation
drifts and· galleries in which only temporary supports are employed. This latter
system is not permane nt but changes continuously in keeping with the mining
faces. There is, therefore , a fundamental difference between these changing mining
engineering structures of a temporary character and tunnels proper, not only on
account of their motility and temporary support system, but also because of the
basic aims.
The purpose of mining is the exploitation of minerals. The cavities are pro-
duced as unwanted by-products creating an additional problem by the neces-
sity for their subsequent maintenance in order to prevent the collapse of the
entire system and the occurrence of unwanted surface subsidence. Tunnels, on the
other hand, are invariably built as permane nt structures. The basic objective of
their construction is the creation of cavities suitable for the admission of traffic
and transportation. The unwanted by-product in this case is the material exca-
vated from the cavity, the removal and disposal of which must be provided for
separately.
Another distinction between underground structures associated with mining
and tunnels can be drawn as regards depth, since the former are constructed
mainly at great (several hundred, or even thousand metres) depths in rock
formations of earlier · geological ages · according to the occurrence of the mi-
nerals to be exploited, while tunnels may equally be driven at great depths, e.g.
base tunnels, or close to the surface, e.g. subway tunnels, in younger geological
formations and thus different methods of construction can be adopted.
(b) The second category of underground engineering structures is constituted
by tunnels, the typical features of which have just been described.
(c) The third category comprises shelters offering protection against air attacks
to persons, materials, warehouses, even to entire factory plants. The objective
in this case is to ensure protection and the conditions necessary to life for a limited
period or sometimes even for normal operating conditions. Safety is the only
consideration in selecting the depth for these structures ·and consequently they are
usually built in young ·geological formations. Their dimensions may vary within
wide limits and are governed primarily by the effort to arrive at an economical
structural design.
Chambers of abandoned quarries and natural underground cavities have
been frequently used for this purpose.
(d) Finally, increasing use has been made in recent times of the subsurface
space in highly developed city areas of metropolitan centres with dense traffic,
not only for solving traffic problems and to provide for eventual wartime protec-
tion, but also for . various purposes of peacetime storage. Typical examples of
this category are underground garages and parking areas.
 CLASSJJICATION 21
I.I. PURPOS ES AND CLASSI FICATIO N OF TUNNE LS

The demands on passenger and goods transport ation have increased immensely
with social development. Not even the moc;t ai;lverse natural obstacles could
curtail it. Rivers, mountains and oceans could delay transport ation for certain.
periods only, but were eventually conquered by human endeavour. People and
territories formerly isolated from each other are now linked by bridges spanning
rivers and by tunnels underpassing mountains in increasing numbers and with
increasing dimensions, each representing a tie and contribut ion towards the cul-
tural unity of mankind .
The purpose of tunnels is to ensure the direct transport ation of passengers or
goods through certain obstacles. Depending on the obstacle to be overcome and
on the traffic, or transport ation objective to be achieved, tunnels can be classified
into various groups.
The obstacle may be a mountain, a body of water, dense urban, or industria l
areas (traffic, etc.). Tunnelc; may pass accordingly under mountains, rivers, sea
channels, dense urban, or industrial areas, buildings and traffic routes. Their
purpose may be to carry railway, road, pedestrian, or water traffic, to convey
water, electric power, gas, sewage, etc. or to provide indoor transport ation for
industrial plants. Tunnels may thus be classified according to their purpose,
location and geological situation. Depending on their purpose the following two
main groups of tunnels may be distinguished:

(A) Traffic tunnels:

1. railway tunnels
2. highway tunnels
3. pedestrian tunnels
4. navigation tunnels
5. subway tunnels.

(B) Conveyance tunnels:

1. hydroelectric power station tunnels
2. water supply tunnels
3. tunnels for the intake and conduit of public utilities
4. sewer tunnels
5. transport ation tunnels in industrial plants.

In addition to purpose, importan t cla!.sification criteria are location, position

reiative to the terrain and alignment as well, these having a decisive influence on
the tunnel section, the method of construction, the design and the acting forces
(cf. Section 21.1).
Tunnels will hereafter be understood as being underground structures, which
apart from serving the above purposes, are constructed by special undergro und
tunnelling methods generally without disturbing the surface. Tunnels are also
22 INTRODUCTI ON

constructed in open pits excavated from the terrain surface and are backfilled
subsequently (cut and cover). The means employed in these "cut and cover"
.methods do not differ essentially from those associated with the constructi on
of foundation pit enclosure, dewatering, excavation and backfilling methods used
in foundation engineering.

1.11 . TRAFFIC TUNNELS

Traffic tunnels may be classified according to their position, or alignment and

purpose.

l l.11. Classification by Position or Alignment

11.111. Saddle and base tunnels. When leading from one side of a watershed
to the other and when connecting parallel valleys, railways and highways are
bound to cross mountain ridges. Tunnels being the most expensive of all
struc-
tures, the general tendency is to follow an open route as far up as possible in
order to minimize the length of the tunnel. The open track is, in such cases, :ed
along the valleys. As long as the natural slope of the valley does not become steeper
than the ruling gradient of the route, the length necessary for the climb can be
ensured without difficulty. As soon as the valley becomes steeper (Fig. 1/ 1) the
t-~
e
Ot' < t
~~==~~=~~=~~-::=.~~=-=:.====-~~=-===::.-~-:=;~-;;:r---~:=:::

Fm. 1/ 1. Location of the traverse line in comparison

/ with the natural rise of contour lines of a valley

line must be developed in length. At first one, then both sides of the valley and
in many cases even the side valleys must be utilized for this purpose. Instead of
a straight track the necessary length must be created by sinuous curves. Both the
waste in distance, and the number of bridges and structures increase in this process
 CLASSIFICATION 23

(cf. Fig. 1/3). The specific construction cost related to the distance between the
points to be connected (termini) increases similarly. The valleys become gradually
narrower and steeper until any further lenghtening of the line, when adhering to
the specified smallest radii of curvature and steepest ruling gradients, proves to
be technically impossible (cf. Fig. 1/2). The highest ~levation is then attained
beyond which the line must inevitably be led into the interior of the mountain,
in order to arrive at the adjacent valley in compliance with the specified alignment
and traction requirements. Tunnels built thus at the highest possible elevation
are referred to as saddle, or watershed tunnels. They are usually adopted on railway
lines of secondary importance and light t~affic and on highways, where it is justi-
fiable to save construction costs at the price of higher operating costs. On main
lines, however, where the traffic density is great and traction loads are heavy,
the increased travelling distance and the lost height would increase oper-
ating coi1ts sufficiently to make the construction of a tunnel at a greater

WaterJl,ed
'

____ !/ass tvnnel

' c'.
:tj
~~
~

~ j " Lmedevelopment
'fil _ --,;,.;==l==f: At. max.height
-!:i -~:r7M3.19i11"rJ.h)e.ei1Q,th"ii"tt~off1--=Z!IT:======-t- ---- __ ofembank-
~ jJ ~ embankmenU, t16 max.depth of, -- ment
~ e cut t12
e1 > e1
Yz >!I,
FIG. 1/2. Comparison of base and pass-tunnels

depth and of a greater length at the bottom of the valley more economical
in the long run than to bear the constant losses arising from increased
operating costs. It should also be borne in mind that the carrying capac-
ity of a shorter line with gentle gradients is appreciably greater than of a longer
one with steep gradients and sharp bends, and so the construction of a tunnel
at greater depth, a base tunnel (Fig. 1/2) by which sometimes even the construction
of a _second line may be avoided, will represent the best solution economically.
Another fact revealed by the figure is that secondary railway lines and particularly
highways with a steeper ruling gradient e2 attain the highest permissible embank-
ment heights and cut depths (.1max) only at steeper terrain slopes and over short
distances. On main lines with a more gentle ruling gradient, ei, this circumstance
also will soon limit any line development. The increase of this .1 value is at the
same time representative of the topographical difficulties encountered in the line
development. In conclusion it can, thus, be established that on railway trunk lines
base tunnels of increased length are usually preferred whereby both the length
of the track and lost height can be minimized, whereas on secondary railway lines,
24 INTROD UCTION

Until recently
and especially for highways, saddle tunnels are given preference.
allowable steep
even saddle tunnels were rarely adopted for highways since the
pment of the line
grades and small radius of curvature made the necessary develo
conducted in an
on one side of the valley possible and thus the highway could be
of being made
open cut across the pass. In fact, the open route runs the risk
ered by ava-
impassable by snow and ice, while other sections may be endang
of highw ay traffic such
lanches. Owing to the sparcity and secondary significance
ly. tolerat ed. Nowadays,
seasonal interruptions of highway routes were former larly
must be ensure d particu
however, uninterrupted and safe traffic conditions
of saddle tunnel s instead of
on busy motorways which justify the construction
tunnel s have been adopte d on
open routes for highways as well. Even base
6 km St. Bernard
highways of international importance, as exemplified by the
and the 12 km Mont Blanc highway tunnel s.
ed lengthening
11.112. Spiral tunnels. In the course of developing the requir
becom es inevita ble. In valleys too narrow
of the line the use of spiral tunnels often
one side with the minim um permis sible radius
to return the track arriving from
steep slope of the valley can no longer be follow ed
to the other and where the
d by penetr at-
(cf. Fig. 1/1. t > ..1;), the necessary additional length must be ensure
of the mount ain and the length require d for overco ming the
ing into the interior
s spiral tunnels
rise must be obtained by forming a loop (Fig. 1/3). The famou
mentio ned here.
on the Gotth ard and Albula railways in Switzerland may be

FIG. J/3. Loopin g tunnels

 CLASSIFI CATION 25

Geologically, the widest variety of rock conditions is likely to be encountered

in driving spiral tunnels, since in forming the {oop the outer weathered and inner
sound rocks must be traversed both parallel and normal to the dip.
ll.113. Offspur tunnels are used to shortcut minor local extruding obstacles
(Fig. 1/4), when it is desired to maintain a straight line for any traffic route (railway,
highway, canal), or when a sharp and steep projection cannot be followed with
the minimum permitted curvatures (cf. Fig. 1/4). Another indication for this type

FIG. 1/4. Offspur tunnels

of tunnel is if the cut leading through the obstacle would be too deep, or if the
open cut would induc~ sliding in the sloping layers. In general beyond 20 to 25 m
cut depths the construction of a tunnel is more economical.
11.114. Slope tunnels. These are also constructed with a local character. Their
objective is to ensure safe operation and protection to railway and highway
routes in steep mountain hangs. In such cases the dip of the layers, whether
inward, or outward is of paramount importance. Outward dip of the layers is
especially dangerous for open routes, and even slope tunnels must be located
deeper in the mountain side under such circumstances (Fig. 1/ 5). Even where the
stratification of the rock formations is favourable enough, i.e. inclined inward,
it is not advisable to apply mixed cut and fill sections even though there is no
danger of sliding, for the danger of rock and earthslides, or avalanches still
remains. Th~ primary aim of these tunnels is the provision of safe traffic conditions,
but sometimes they are adopted for economic reasons, as the construction of the
26 INTRODUCTION

Fm. 1) 5. Hang t.unnels

Retammg w a l l ~ "\ _._

I \
\
\.'\ /
\. / /
V

alternative cut-and-fill section may call for the construction of very long slopes
and of very"high retaining walls the cost of which may well exceed that of a tunnel.
·Tunnelling, besides being cheaper, may also be appreciably safer.

11.12. Classification According to Purpose

11.121. Railway tunnels. Among traffic tunnels the importance of railway
tunnels is unquestionably the greatest. The majority of railway tunnels are to be
found in ·mountainous country, but there are some, too, which have been con-
structed under watercourses, or in urban areas (e.g. New York). As far as location,
site and functions of mountain railway tunnels are concerned, the same types can
be distinguished as discussed in section 11.11. Otherwise these distinctions apply
equally to highway tunnels in mountainous regions. The characteristic features
of railway tunnels will be discussed in detail in connection with their design,
location, cross-section and profile (cf. Section 2.2).
 CLASSIFICATION 27

11.122. Urban underground railways. Although essentially railway tunnels, the

tunnels of urban underground railways constitute an individual group among
them, differing from other railroad tunnels as regards location, construction
methods, materials and purpose. The latter are mostly of horseshoe section, while
the tunnels of underground railways are usually circular, rectangular, or polygonal
in cross-section, depending on whether highly located "cut and cover", or deeply
located tunnels are concerned (cf. Fig. 2/51). In design they most resemble sub-
aqueous highway tunnels. The obstacle to be underpassed in the interest of traffic is,
in this case, the densely overbuilt urban area, the network of streets and buildings.
Underwater railway tunnels belong essentially to the same group. Such tunnels
were relatively few in number, owing primarily to the difficulties encountered with
steam and diesel traction.
The stations of underground railways are of special structural design and of
a great variety of types, especially in the case of deep line!! (see Figs 6/ 160 to
6/171).
The principal special requirements of underground railway tunnels, distinguish-
ing them from other railway tunnels, can be grouped _as follows:

1. Increased safety requirements due to the great density and high speed of
traffic (freedom of displacement and deformation of track and tunnel lining);
2. Careful watersealing;
3. A high standard of cleanliness and ventilation.

ad 1. Derailing due to breakage or even to the slightest loosening of rails may

involve catastrophic consequences in view of the high speed of trains and to the
confined space.
ad 2. The tunnels, being located under groundwater level in the geologically
youngest, waterbearing top formations, practically stand in water.
ad 3. Since the stations themselves are accommodated in the same tunnel as
the track, increased attention must be paid to the sanitation a nd ventilation
requirements of the large number of passengers. Also, the removal of heat and
dust caused by the sudden braking and acceleration of high speed trains must
be provided for, and the tunnels must be kept clean.

11.123. Highway tunnels. Highway tunnels can be classified broadly into three
groups (see Figs 2/44 and 2/45).
(a) Tunnels constructed on motorways, or modern main traffic routes may
belong to the same types as railroad tunnels; pass tunnels (Giovi), base tunnels
(Mont Blanc, St. Bernard), offspur, or slope tunnels are equally' possible. Owing
to the steeper ruling gradient on highways_ spiral tunnels have not been built
so far. The only difference in cross-section in comparison with railway tunnels
is the reduced height, since the ratio of width to height "in the case of two-lane
highway traffic differs from railway clearance requirements. Highway tunnels of
this nature have been constructed in recent times only, since the significance of
28 INTRODUCTION

highway traffic increased tremendously with the development of the motor vehicle.
In earlier times highways always followed the slopes of the terrain, since the sparse
horse-drawn traffic tolerated very narrow curves and steep gradients and the lost
height in the alignment was not considered. The claims of modern motorways
built for a great traffic volume are similar to those of railroads.
(b) Interconnection tunnels, underpassing minor hills in too interior of towns
differ from the former group mainly in their dimensions. These tuooels are usually
direct continuations of wide urban road arteries and are distinguished by a wide
roadway and a great sectional height as well, in the interests of good natural
ventilation, e.g. the urban tunnels in San Francisco, Budapest, Prague, Naples,
Rome, Rio de Janeiro, etc., or they are artificially ventilated when of a flat
section (Brussels, Paris, etc.).
(c) The third group of vehicular tunnels, those under watercourses are also
encountered in urban areas. In carrying urban traffic, these tunnels compete
with bridges and present an especially attractive alternative to the latter where
high clearances are required to ensure the passage of seagoing vessels at all times.
In highly frequented sea ports attempts were previously made to solve the problem
of simultaneous navigation and road traffic by the compromise of lifting, or revolv-
ing bridges (Leningrad, London, Copenhagen, etc.). With traffic of increasing
density on both water and road, the situation created by such solutions became
soon intolerable. The accommodation of the approaches of fixed high level bridges
may encounter unsurmountable town planning difficulties in congested urban
areas, interfere with the very important t.rnffic on quays and docks, and in addition
may produce such an unaesthetic apr.earance to ~ city that this alone might
justify the abandonment of this solution. In many instances also the construction
costs of a permanent bridge may exceed those of a tunnel. Thus, e.g. in Hamburg
the required navigation clearance for seagoing vessels would have been 55 m
and while the tunnel actually cost 15 million marks, the construction costs of
a bridge were estimated at 25 million. A similar clearance of 55 m would have
been necessary in Antwerp as well, and a suitable bridge would have cost from
260 to 360 million francs in 1932, whereas the actual construction costs of the
tunnels amounted to 245 million franc<; only. For the same reasons tunnels re-
m;\ined the only choice in the most valuable city areas of New York. The reduc-
tion in construction costs of tunnels resulting from the use 9f prefabricated large
R.C. caissons has further added to their advantages over 'bridges (Rotterdam,
Amsterdam, etc., see Section 63.12).
In Budapest, on the other hand, where the required navigation clearance is not
more than 15 m, the costs of a tunnel would be three times as high as those of
a permanent bridge.
The advantages offered by such tunnels to town planning may even outweigh
economic considerations. In fact open cuts for the access and egress roads of the
tunnel are inevitable beyond a certain depth. This depth is, however, usually no
more than 5-6 m and with gradients of 3-5 % permitted for vehicular traffic the
disturbance of the town area extends to a length of 100--150 m only. On the other
 29
CLASS lFICA TJON

ioned must be at least 1150--1500 m

hand , the appr oach ramps of the bridges ment
long.
er immunity to military attac k,
Furt her advantages of tunnels are their great
t to deviate from the straight axis,
as well as the greater freedom in align men
approaches be shor ter; but they can
as a consequence of which not only may the
street patte rn of the city, and the
be fitted much more readily into the existing
traffic value and often of historical
disturbance of emb ankm ent areas of grea t
significance can be avoided.
c engineering study prep ared for
As demonstrated by FEUCHTINGER in a traffi
s are superior to bridges as far as
a parti cular project, • tunnels unde r waterway
1 1

of view of traffic engineering four-lane

lost height is concerned, and from the poin t
. The summary and com paris on of
constructions are preferable to two-lane ones
:
relative costs yielded the following sequence
1. four-lane tunnel = 82 %,
2. four-lane bridge = 89 %,
3. two-Jane tunnel = I 00 %,
4. two-lane bridge = I 13 %-
crossing time and construction
As regards the steepest permissible gradient,
d vehicular traffic, the most econom-
costs are conflicting factors. Assuming mixe
consumption are 3· 5 % and 3·0 %,
ical gradients for crossing time and fuel
r the more impo rtant , be specifies
respectively. In ratin g the role of time facto
the most favourable gradient at 3·5 %-
a grea t similarity to unde rgrou nd
Vehicular tunnels unde r waterways show
extent in cons truct ion _methods.
railway tunnels in cross-section and to some
ngs and unde r identical unde rgro und
They are built in identical urba n surroundi
oaches may be steeper. For tunnels
conditions. Natu rally, the grad ient on the appr
value usually ranges from 3 to 5 %-
designed mainly for moto r traffic the gradient
the Elbe River in Ham burg should be
In particular, the vehicular tunnels unde r
ned approaches and valuable urba n
mentioned, where in orde r to save the incli
them, connection to the street network
territory which would have been affected by
cal shafts at the tunnel ends. The
was made by elevators accommodated in verti
ted caissons previously mentioned
employment of the sinking meth od of prefabrica
reduction in the slope and length of
also offers a possibility for a considerable
access ramp s.
belong to the grou p of road tunnels,
11.124. Pedestrian tunnels. Essentially, these
t radii of curvature and steep per-
but owing to their smaller cross-section, shor
the possibility of providing for their
missible gradient (up to 10%) as well as to
construction and design is sufficiently
access in vertical shafts instead of ramps, their
rate group. They are usually employed
simplified to justify their discussion as a sepa
ior of towns. Beginning with the
unde r surface traffic centres i.e. in the inter
1•1 FEUC HTING ER, M. E.: Hochbri.icke oder Tunn el, Strasse
und Autobahn 1956 226
30 INTRODUCT ION

underpasses at congested intersections and under bridge approaches (London,

Paris, Frankfur t, Munich, Vienna, Genoa, Rome, etc.) up to the pedestrian
tunnels under wide watercourses (Hambur g, Antwerp) similar structure s are built
under the greatest variety of conditions. Among pedestrian tunnels those under
the Elbe in Hamburg and under the Scheide at Antwerp have gained the widest
reputation. These are built without inclined approach es b1:1t terminate, in vertical
shafts, where pedestrians are carried by passenger lifts to and from the surface.
The tunnel in Antwerp was built especially for pedestrian traffic, while the one
in Hamburg carries mixed traffic, with l ·25 m wide sidewalks at both sides of the
central single-lan e roadway in a way c;imilar to the deck arrangem ent of highway
bridges (cf. Fig. 2/44).
11.125. Navigation tunnels. Among tunnels for traffic, navigation tunnels were
among those whose importance and the economic interests concerned justified the
expense and labour involved in their construction. In the development of civili-
zation the transport ation potential of waterways was first utilized, a-nd in order
to establish connections for navigation from one river system to another, the
construc tion of inland navigation canals was started as early as the 17th century.
In crossing dividing mountain ridges the necessity for constructing navigable
tunnels soon arose. From the Malpas tunnel opened in 1679 there was a long
development to the construc tion of the Rove canal tunnel in the vicinity of Mar-
seilles in 1922. This is still the world's largest tunnel. Typical features of navigation
tunnels are the impermeable lining and careful sealing, as well as the great width,
which is, relative to the vessel towed, the largest among all traffic tunnels in the
interest of keeping towing resistances as small as possible (cf. Table 1/I and Fi;.
6/59). An additiona l factor calling for great width is the poor manoeuverability
of floating craft. For these reasons the wide cross-section is usually covered with
a flat arch. Watertight internal seals, or linings were frequently found necessary
(cf. Fig. 2/41 ). In regions where there is a possibility of tectonic movement, or
where the rock formations are faulted or cracked and show a tendency to slip,
navigable canal tunnels must not be constructed.
Here again, single and two-lane tunnels can be distinguished, and in the case
of long tunnels intermediate passing points are usually provided.

1.12. TRANSPO RTATION TUNNEL S

11.21. Hydroelectric Plant Tunnels

The water diverted for utilization in hydroelectric stations is conveyed from
the raised reservoir to the adjacent valley, or the powerhouse, located at a lower
elevation. Two main groups of tunnels for this purpose can be distinguished .
Those which discharge with full section directly to the turbines are referred to
as pressure tunnels. Owing to the considerable difference in elevation they are
subjected to very large internal pressures . In contrast, tunnels constructed for the
 CLASSIFIC ATION 31

sole function of conveying water by normal gravity from one side of a mounta
in
to the other, or from the point of diversion to the head of the steep incline,
are
tl:rmed discharge tunnels (Fig. 1/6).
Between the two groups significant differences exist as far as loads, cross-section
and shape are concerned. Pressure tunnels are subjected to an internal water
pressure which is frequently many times in excess of the external rock and ground
water pressure. The resultan t tensile stresses can be resisted most econom
ically
by a circular cross-section. Pressure tunnels are therefore built with a circular
or

Waler carrier tunnel

Storage
reservoir
Pressure conduit

F 1G. 1/6. Location of water-con veying and of pressure tunnels

horseshoe-shaped cross-section, the latter being more readily adaptab le to drilling

methods. In order to resist tensile forces these cross-sections are provided
with
a lining of at least reinforced concrete , but steel linings are also quite commo
n.
These linings have, of course, a sealing function as well, since water under
high
pressure is more likely to escape through the fissures and cracks in the
rock
(cf. Fig. 2/40).
Simple discharge tunnels, on the other hand, are usually designed with a horse-
shoe cross-section similar to that of railway tunnels, the internal pressure exerted
by the water conveyed being negligibly small in compar ison with the rock pressure
acting on the tunnel. The horseshoe section with an inverted base has been found
most economic for tunnels of this type (cf. Fig. 2/40). It will be realized that
the
applicat ion of a reliable interior waterseal is essential.
These tunnels, even more than the navigation tunnels mentioned previous
ly,
cannot be built, unless the rock is completely immobile and solid, i.e. pressure
tunnels, or pressure. shafts must not be constructed in rocks interlaced with
faults
and cracks and tending to slip, or in those liable to tectonic movements.

11.22. Water Supply Tunnels

These are essentially identical with discharge tunnels dealt with in the precedin
g
paragra ph, the only difference being in their respective functions. The task in
this
case is the conveyance of domestic water from springs, reservoirs, or river diver-
 INTRO DUCTI ON
32
(cf. the Delaw are tunne l in the
sions to the storag e tanks of the city water work s
is that, being built for the con-
suppl y system of New York ). The main difference
ments , suppl y tunne ls are in many
veyance of dome stic water to majo r settle
, while press ure tunnels trave rse moun -
instances locat ed in recen t alluvial layers
rocks. Again, the distin ction
taino us regions, i.e. are usually locate d in more solid
ls and press ure or high-head tunnels can
betwe en free-surface, or low-head tunne
be made here (cf. Figs 1/6 and 2/40).

11.23. Publi c Utility Tunnels

powe r, teleph one and other
These are usually built in city areas for carry ing
under watercourses, roadw ays,
cables, gas, water and other impo rtant utility pipes
de for conti nuous inspection,
railway track s, blocks of houses, etc. and to provi
main tenan ce and repai r of occasional damage.
previously menti oned is that
A significant difference from severa l tunne l types
, are used for conne cting them with
vertic al shafts, instea d of inclined access paths
publi c utility tunne ls is the possi-
the surface. A great advan tage of these urban the necessity of
metho ds, elimi nating
bility of const ructin g them with tunnelling
. Their maint enanc e and inspec-
open trenches and the distur bance of surface traffic
ls of small dimensions have
tion can be perfo rmed similarly. Short utility tunne
units throu gh the groun d start-
recently been built by jackin g prefa brica ted lining
and 6/179). Several utility lines
ing from head shafts (cf. Section 64.3 and Figs 6/178
a comm on tunne l especially in
(gas, water, cables, etc.) are accom moda ted in
areas.
indus trial plant s and in newly established residential
be rectan gular or circul ar in cross- sectio n, depen ding on
Utilit y tunnels may
the cut-an d-cov er metho d. In
whether they are const ructe d by tunnelling, or by
urban pedes trian tunnels (cf.
locati on and dimensions they usually resemble
Fig. 2/42).

11.24. Sewer Tunnels

sewage and are, in certa in
These are const ructe d for the removal of domestic
tunne ls, being more similar to water suppl y
respects, different from other utility
under the actio n of gravity,
tunnels (cf. Figs 2/40.and 2/42b). Sewage flows in them
to prote ct the walls again st the
and the prima ry role of intern al watersealing is
ined in the liquid, the preve ntion
aggressive actio n of harm ful chemical agents conta
on. For this reaso n an intern al
of seepage losst.5 being the secon dary considerati
acid-resistant and bitum inous
lining of sintered bricks, ceramic tiles, laid in
the main sewers of very large
morta r is necessary, as a rule. Of course, it is only
iques, becau se these are at
diame ter which are const ructe d by tunnelling techn
it the use of such metho ds.
a sufficient depth below the groun d surface to perm
e, cross-sections of special
Owing to the considerable varia tions in sewage volum
smaller troug h (cunette) is pro-
shape are used for these tunne ls, inasm uch as a
flows . Occasionally, only this
vided at the botto m of the inver t to collect small
troug h is lined with brickwork.
 CLASSIFICA TION 33
The construct ion of sewer tunnels may become necessary under watercou rses
as well as under urban traffic lanes not only for conveyin g but also for introduci ng
sewage into the recipient (cf. Fig. 6/77).

11.25. Other Tunnels

A wide variety of tunnels is required for the operation of particula r industria l
plants. Example s of these are the sand conveyan ce tunnels in foundries , scale and
scrap tunnels in rolling mills and various other utility tunnels.

1.13. STORAG E TUNNEL S

(GARAGE S; PARKING AREAS, SHELTER S AND STOREHO USES)

As a result of the rapid developm ent of urban traffic and the growing congestio n
in their centres, no garage facilities are nowaday s available to thousand s of auto-
mo biles and badly needed public space is occupied at night by vehicles parked at
kerbs and at every conceiva ble place. During daytime motor traffic in the city
streets is practical ly impossib le. Since the roadways , which are usually not suf-
ficiently wide, are blocked by parked automob iles, city authoritie s levy heavy
parking charges to discourag e motorists from unnecess ary parking. (Accordi ng
to estimates traffic congestio n, e.g. in Paris, is responsib le for an annual loss of
34 thousand million francs.) Followin g electric tramway s, buses and trolley buses
are gradually being excluded from city centres and the undergro und railway is
becoming the only means of transport ation. These are the reasons which make
the systemati c developm ent of the space available under valuable city areas a
compelli ng necessity. Also, the unobstru cted rapid access to the city centre by
public transport must be available . This demand has led to the idea of the under-
ground developm ent of metropol itan areas, conceive d 25 years ago. The objective
pursued is the advance planning of the undergro und transpor t network, under-
ground railways, express motorwa ys, subways, garages and parking lots, and the
systemati c co-ordina tion with the requirem ents and network of public utility
services. As indicated by the evidence already available the random location of
utility pipelines and other miscellan eous structure s (pedestri an subways, etc.)
results in wasting space which is of immense potential significance for future
town developm ent, and the relocatio n of scattered structure s at a later date involves
excessively high costs. Since the construct ion of undergro und garages and parking
areas has become an actuality , the space requirem ent of undergro und road and
railway traffic must be determin ed in advance and the necessary areas must be
reserved for them.
Thus, at present, the main obstacles to the construc tion of undergro und garages
and parking areas are:
1. the irregulari ty of existing undergro und conduits and structure s;
2. the difficulties encounte red in construct ion work (mainten ance of surface
traffic, or difficult circumsta nces of undergro und construct ion, turnouts ;

3 Szechy: The AIi of TunneUing

 INTROD UCTION
34
;
3. the high investment costs of this type of constr uction
4. ventil ation and sanita tion problems.
areas utilized first were
Proceeding along the line of least resistance the urban
parks , quays , river beds, hillsides,
those offering certain natura l advantages, e.g.
bridge appro ach viaducts. ,
construction work does
Utility conduits are less frequent under )?ark areas and
es are simila rly infreq uent under quays and
not interfere with traffic here. Utiliti
quay, or appro ach can
bridge appro aches , and in addition, the side-wall of the
while in the latter case the
be utilized as a latera l wall either at one side or both,
as well.
bridge deck may even serve as the roof for the garage
hillsid es the length of access ramps can be reduced to
When utilizing steep
The space proble m of
a minimum and a saving in lining can also be expected.
in many instances by pro-
entran ce and exit ramps bas recen~ly been eliminated
to and from the garage.
viding special vertical lifts for transp orting the vehicles
of the River Limmat at
A highly original idea is the utilization of the bed
two-st orey garage of 126 m
Zurich, where it is contemplated to accom moda te a
ns to be sunk below the bed level.
by 41 m surface area in huge pneumatic caisso g areas
groun d garage s and parkin
A furthe r aim in the construction of under the city
offerin g adequ ate protec tion to
is to provide air raid shelters capable of
population.

FIG. 1/7. Design of a

subter ranean
parkin g place
'::
Oe/ail f'
Travel/mg crane

r51 110 1151 - flfm • / __ Jl O - ~ ~;~::::r ---· .. ..- ...., .

~~
---- ~ '0' . ,

0 ~
-·~ <
-'.¾ ~
D -.E ~ -~. - ~~~~ ~ ~
• ¼ • •• -. 6 r
' L : = : t...t:;;;;:;;.:;::;:; ("}
A • Inflow tonne/
~'rane (Jenerator-s !
-.,,
~
r
~
:i
B • Storage _dham~er Gale-valve chamber ~
"' ;("}
C • Eqllllibr1um shaft lilrbines ~
O Vertical shaft
O
~
E • Underground power JhJlion
F • Head breaking chamber
G - 011_tflow off tube
H • Oblique access tunnel
J • Stores and offices
I( • Transformer station ' Pres.sure tunnel
v • Sloice gate

PtG. 1/8. Location and t11rbine house of a subterranean hydra11lic power station
 INTROD UCTION
36
of the parkin g area
As an il1ustrativc example the schematic arrang ement
in Fig. 1/7. Abou t 1900
planned under the Monceau park in Paris is shown
three-bay, five-storey
vehicles can be accommodated in the 30 m wide, 181 m long,
80 000 persons. Access
structure, which is capable of affording shelter to about
se protec tion against
ramps are designed with helical bends in order to increa
d caused by fuel vapours
bomb damage. In order to eliminate the explosion hazar
ing the necessity of a
a high stand ard of ventilation must be ensured, involv
complete air exchange every 4-6 minutes.
ications relating to the
A detailed treatm ent of ~arious regulations and specif
be out of place here, and the
design and comtr uction of air-raid shelters would
large under groun d machine halls and
space requirements and dimensions of
for each partic ular installation and
other plants for hydroelectric stations vary
suit actual condi tions. Geolo gical considera-
their arrang ement is also adapt ed to
tion of turbin es and gener ators of high-head
tions also favou r the accom moda
1/8). Such h_alls are pre-
hydroelectric plants in large under groun d halls (Fig.
l caves, or in t>ld quarries
ferably located in mountain-sides, or in existing natura
adequate enlargement.
by providing for their prope r perma nent suppo rt and

TION
1.2. A BRIE F HIST ORY OF TUNN EL CONS TRUC
prehistoric times when
The histor y of the art of tunnel driving reaches back to
weath er and his enemies,
primitive man, seeking shelter against the inclement
drove headin gs to enlarge the
excavated cavities resembling natura l ones, or
latter.
underground cavities
There is ample archeological evidence to indicate that
in the search for miner al deposits. These
were excavated as ·early as the stone age
of circul ar cross- sectio n and reveal a certain
primitive headings were already
regular patter n.
the express purpo se of
The oldest tunnel, i.e. under groun d structure built for
t knowledge, 4000 years
communication was constructed, according to presen
ancient Babylon to under-
ago under the reign of the famous Queen Semiramis in
underground connection
pass the bed of the River Euphrates and to establish an
length of this tunnel was
between the royal palace and the Temple of Jove. The
n dimensions of 3·6 m
1 km and it was built with the considerable cross-sectio
origin al bed for the construc-
by 4·5 m. The River Euphrates was d1verted from its
erable project even according
tion period and the tunnel, which would be a consid
cut. The wall of the tunnel consists
to mode rn standa rds, was built in an open
the sectio n is covered from above
of brickwork laid into bituminous morta r and
exten t of the under taking point to the fact
by a vaulted arch. The vast scope and
to be built by the Babyl onians and that
that this tunnel was not the first of its kind
practi ce with severa l tunnel s built earlier.
they must have acquired skill and
d be remembered that the
To appreciate tile grand eur of the under taking it shoul
 HISTORY OF TUNNEL CONSTRUCTION 37

neitt subaqueous tunnel was opened about 4000 years later, in 1843. This was the
tunnel under the River Thames in London.
Numerous tunnels were driven later for military purposes to undermine hostile
fortress walls, and to open access thr~ugh them into the interior of besieged
towns. Tunnels built for undermining or underpassing fortifications were an impor-
tant feature of R~man and Greek warfare as well.
The 200 m long tunnel built with 0·7 m by 0·7 m internal dimensions 2700 years
ago by Hezekiah, King of Judea, in Jerusalem for introducing the water of a
nearby spring into the town also served military purposes, as it was prompted
by the fear of the impending siege by the Assyrian army. The most famous Greek
tu nnel on the Island of Samos was also built 2600 years ago for water supply
purposes with a length of about 1·5 km and with 1·8 m by 1·8 I)). internal dimen-
sions!
The aqueduct tunnels built in Roman times also acquired great fame and repu-
tation. Of these the supply tunnel built in the reign of the Emperor Hadrian,
i.e. 1800 years ago in Athens, was reconstructed in 1925 and is still operated by
the municipal water works.
Tunnels several thousand years old and driven in rock have been found in
great numbers during the exploration of Egyptian tombs-and temples. The number
of tunnels built in ancient India and Mexico, as well as in Peru, is similarly large.
The 900 m long and 7·5 m wide road-tunnel under Posilipo hill on the Naples-
Pozzuoli route was built 2000 years ago and can be attributed to Roman civiliza-
tion. A most extensive network of tunnels was developed at the time of the per-
secution of the Christians in Rome, where countless c:rnmbers were excavated
along the passages and halls connected by the catacomb system, which in turn
were started from abandoned quarries. In these chambers over 6 million <leads are
deposited.
In the cultural decline of the Middle Ages tunnels were built for military pur-
poses only. The construction of the Col dt Tenda road tunnel was started under
the Maritime Alps between Nice and Genoa in 1450, but it was soon abandoned
and left unfinished to this very day. Mention must be made of the 5·6 km
long drainage tunnel of the Biber mine at Selmecbanya, Hungary, which
was started around 1400 and was the most remarkable underground project of
its time.
Attention was again focussed on tunnelling by the construction of navigation
tunnels in the 17th century. The differences in elevation which can be overcome
by navigation canals .are, indeed, very slight and so comparatively low terrain
obstacles, such as bills, made tunnelling inevitable. Gunpowder was introduced
for rock blasting in the construction of a tunnel on the Languedoc canal in France
in 1679. Prior to this, rock was broken with hammers and wedges by manual
labour with heavy toil and immense effort. Occasionally, the rock was heated
and then quenched with cold water so thafthe rock fissured by sudden contraction,
and so could be broken and excavated more readily. Gunpowder had been used
half a century earlier in mining. Its first application was also in Selmecbanya,
38 INTRODUCTION

Hungary, in 1627. Difficulties similar to those of excavation were encountered

also by underground water inundating the mines, and trus led to the development
of water lifting devices. The importance of these in tunnelling was less pronounced,
since drainage could always be achieved by the relatively simple means of con-
structing the tunnel at a gradient, or by a separate drainage gallery.
From this date, navigation canal tunnels were built both in France and England
in rapid succession as navigation canals became the economical means of transport
not only for agricultural products, but also for those of industry and mining,
e.g. coal. On the canals traversing hilly mining regions in the two countries a great
number of navigation tunnels have been built, the total length of which attained
several hundred kilometres, during one and a half century.
The impetus imparted by inland navigation was further increased by the advent
of railways, wruch, with their gradient limitations, were soon confronted with the
first terrain obstacles. The first railway tunnel for horse-drawn operation was
constructed in France in 1826 on the Roanne-Andressieux line. The first railway
tunnel for steam traction was constructed between 1826 and 1829 on the Liver-
pool-Manchester line. The rapid development of the railway network and the
recognition of its tremendous importance in establishing close contact over an
ever widening area led shortly to the necessity to construct tunnels under huge
mountain ranges. The realization of these gigantic tunnel projects, involving the
piercing of several kilometres of hard rock, would have been impossible without
the development and application of new rock excavation and construction
methods.
It was during the construction of the Mont Cenis tunnel between France and
Italy, started in 1857, that the hydraulic rock drill was introduced, which was
superseded by the pneumatic drill improved by G. SoMMELLIER. Dynamite was
invented at about the same time, in 1864, by A. NOBEL. As a result of the new
techniques the 12·8 km long tunnel was already opened to traffic in 1871. The
improvement of tunnelling techniques remained associated with the famous Alpine
tunnels (in Switzerland the Gotthard, Simplon and Lotschberg tunnels; in Austria
the Semmering, Tauern, Karawanken and Arlberg tunnels), and later with the
Italian tunnels (Ronco, Col di Tenda, Monte Adone, etc.), during the construction
of which not only practical methods and equipment were improved and perfected,
but also the theory of rock pressure and the structural analysis and dimensioning
of tunnel linings was evolved. Parallel with these outstanding engineering a-
chievements countless tunnels of minor or major significance were built all over
the world, in Russia, Germany, France-, the United States, Norway, Japan, New
Zealand and more recently in the Soviet Union, Iran and Turkey as well, using
construction methods and equipment some of which may already be regarded
as standard, while others are still developing (e.g. the frazing machine " mole"
pp. 711).
A relatively recent outstanding accomplishment in the field of railway tunnel
construction is the Great Apennine tunnel built on the new Bologna- Florence
line. During the 11 years of construction 1 970 000 m3 (2 570 000 cuyd) of rock
 HISTORY OF TUNNEL CONSTRUCTION 39

were excavated. The development will be appreciated if it is remembered that

during the 14 years of construction 610 000 m 3 (800 000 cuyd) were excavated for
the Mont Cenis tunnel.
The 8 km long double tr;ack Tanna tunnel built on the Tokyo-Kobe route
fro m 1918 to 1934 has won a special place in the history bf tunnelling because
cif the difficulties encountered during construction. These included the caving in
, •fa 60 m (200 ft) high roof and the repeated flooding of the headings by an influx
of water under 200 m (600 ft) head and at 31 °C. The drainage of this water
necessitated the driving of five parallel drainage drifts in sequence. Owing to the
enormous rock pressure the thickness of the concrete vault roof had to be increased
in places to 2 m (6 ft 8 in).
The lessons learned with railway tunnels were adequately utilized during the
construction of water tu~nels built in increasing number since the turn of the
century with similar cross-sections for conveying do mestic and industrial water,
as well as for the conveyance of water to hydroelectric stations. New problems
arose in these tunnels in connection with the proper design of the waterproof
lining.
The demand for constructing subaqueous tunnels arose in the wake of the rapid
development of traffic in cities situated along watercourses and the tremendous
increase in the 'speed' of urban life. Subaqueous vehicular tunnels were the first
representatives of this group, but were soon followed by the tunnels of under-
ground railways and of v1rious public utilities. A wide variety of construction
methods and materials has been developed for these, partly in response to the
new problems encountered. The construction of the first tunnel under the River
Thames in London was started for the first time in 1807, but it was soon flooded
and abandoned. Construction work was resumed only 15 years later by the
engineer BRUNEL using the shield invented by him. After many difficulties involving
several floodings, the double track tunnel only 150 m (500 ft) long under the bed
and with 6·7 (22 ft 4 in) by 11 ·3 m (37 ft 8 in) cross-section was completed in
1842. Brunel's first shield was rectangular and the tunnel was lined with brick-
work.
The construction method involving the use of a cylindrical shield was greatly
improved by GREATHEAD, who employed it successfull y in London in 1869 for
the construction of the Tower Tunnel, for which cast-iron segments were already
used. A pedestrian tunnel was built almost simultaneously in New-York City
by BEACH, also using the shield method. After the first successful applications the
shield method was combined with compres~ed air dewatering and it soo n became
the almost exclusive tool for building subaqueous and underground railway
tunnels. It would be a difficult task to attempt to list all the famous underwater
tunnels built with the shield method. The most recent examples are the vehicular
tunnels between Blackwall and Rotherhithe in London, the Battery, Lincoln and
Midtown tunnels in New York, the tunnel under the River Scheide in Antwerp
and the largest, the highway tunnel under the River M ersey of 14·3 m (47 ft 8 in)
diameter from Liverpool to Birkenhead.
40 INTRODUCTION

The shield method of tunnelling has recently been further developed in the
construction of the Moscow, London, and Leningrad underground railways.
The shields have been equipped with efficient exca-vating and mucking machinery
as well as lining erectors, which enabled the daily peak advance of 25 cnt attained
by BRUNEL to be increased to 10-15 m.
In soils which are highly pervious both to air and water, and which are at
a moderate depth under the river bottom, the shield method has been superseded
by one in which prefabricated tunnel elements are sunk into place. For this pur-
pose either precast caissons incorporating the tunnel sections are sunk side by
side (Rotterdam and Amsterdam tunnels and the Deas Island tunnel, the sewage
Main Data of Some Remarkable Tunnels

Construction Line. Depth

Name Purpose Rock quality
period country of cover
(m)

RAILWAY

Simplon I. 1895- 1906 Brig- Chiasso Single Gneiss, 2150

II. 1912- 1921 Switzerland- track limestone,
-Italy clay shale

Mont Cenis 1857-1871 Modane- Double Limestone, 1600

-Bardonecchia track gneiss,
France-Italy sandstone

Gotthard 1872- 1881 Goschenen- Double Gneiss, 1800

- Airolo track breccia
Switzerland boulders

New Cascade 1924-1927 USA Single

track

Suramsk 1890- Caucasus Double Hard rock

Sowiet Union track

Great 1920-1931 Prato Double Sandy, 300-600

Apennine Italy track clayey shale
and marl

Tanna 1918-1934 Atami Double Lava, volcanic 378

Japan track tuff, clay

Hokoriku 1957- 1960 Japan Double Sandstone, 60

track granite

Monorfah,a 1941- 1942 Monorfalva Single Cl~y 100

Transylvania track
 HISTORY OF TUNNEL CONSTRUCTION 41

octfall tunnel in Budapest), or a continuous steel pipe section is sunk into a dredged
trench (Houston). The prospects for this method are still very promising and ic
ma) be regarded as the latest stage attained in tunnelling technique, concluding
at the same time the history of development.
A comparison of the dimensions, cross-section and construction times of several
transportation tunnels is given in Table 1/1, illustrating to a certain extent the
de,elopment outlined above. As will be perceived from the table, the tunnel
\\ith the largest cross-section is the Yerba-Buena Tunnel in San Francisco, but the
Ro\'e Tunnel near Marseilles may still be regarded as the largest tunnel in the
orld, where the total excavated volume was in excess of 2 million m3 (2·66
TABLE 1/I
Exca-
,tu.
Lining
Total Shape and d imensions Section vated Excavation
temp. length of cross-se~tion area volume per year Remark
material (m')
C CJ (w) (m) (m')
(IO00m')

56 Ashlar I 19 730 Horseshoe 2 X 27 1070 53 500

stone, 5·0 X 5·35
concrete

Stone, 12 700 Horseshoe 48 610 43 000

brick 8·0 X 7·50

30·8 Stone 14 980 Horseshoe 47 704 37 100

8·0 X 7·35
I
I

Concrete 12 400 Horseshoe 46 450 112 000 4·3 km/year

4·87 X 7·32

3 998 8·95 X 6·98

·63) Brick and J 8 510 9·40 X 8·30 46 1970 178 000 Fire, gas,
concrete (1 ! ·40 X 10·20) water
460000 m 3 I inrush

35 I Stone 7 800 8-40 X 5·26 40 310 20000 Fire, gas,

20 water inrush
I
Concrete 13 871 Horseshoe 75 1000 250000 Water
block 8·54 X 5·30 inrush 15 °C

Concrete, 906 Horseshoe 45 40 25 000

precast 5·50 X 5·70
I block
42 INTRODUCTION

Construction Depth
LineJ
Name period country Purpose Rock quality of cover
( m)

HIGHWAY

Forest-Hill 185!'.-1853 Budapest Urban. Marl 40-50

Hungary vehicular
--- -
Yerba-Buena 1934 San-Francisco Urban Sandstone 200- 300
USA vehicular

Giovi 1932-1 934 Milano- Genova Motorway Clay shale, 200-300

Italy limestone

Mont Blanc 1959- 1965 France- Italy Highway Gneiss, 2000

protogin (24 500)

St. Bernard 1959-1964 Switzerland- Highway Granite, 1500-

-Italy gneiss -1800

Under the 1907-1911 Hamburg Vehicular+ Sand, 6'0

Elba River Germany pedestrian clay shale

Under the 1931 - 1933 Antwerp Vehicular+ Silt, 7·0

Scheide Belgium pedestrian fine sand, 9·5
clay

Lincoln 1934-1939 New York High way Silt 10- 12

! USA and rock

-- -
Rotherhite 1904- 1908 London Highway Silt, 14
Great Britain sand and gravel,
clay

Brooklyn- New York Highway Silt 10-12

- Battery USA
-

Queens- Mid town 1936- 1941 New York Vehicular+ Silt, (4·5)
USA pedestrian rock 14

Mersey 1925-1934 Liverpool Vehicular+ Sandstone 1-10

Great Britain pedestrian 6
 HISTORY OF TINNEL CONSTRUCTION 43

TABLE 1/ I (cont.)

~
Exca-
Total Shape and dimensions Section vated Excavation
Lining
material I length
(m)
of cross-section
(m)
area
(m')
volume
(IOOO m')
per year
(ms)
Remark

- Brisk 349·6 Horseshoe 80 28 9 300 Water

9·48 X 10·60
- 432 65 65 000
Reinforced 150 Horseshoe
concrete 24 X 18

Stone 2868 Horseshoe 74 210 70000

(892) 9·20 X 10·30
- I
Concrete 12 650 9 X 12·45, 80 101 300000
9 X 9·65

R . C. 6 600 Horseshoe 95 660 120 000

10·64 X 9·56
Utilized
cross-section
4·8· X 9·74
-~ - -- -- --
Structural 448·5 Circle 2 X 6·0, util- 28·2 25 6200 Vertical
steel ring ized cross-section shaft access
4·70 X 4·50
--
Cast iron 575 Circle 9·40, util- 69·5 160 54000 Grade
ring 2111 ized cross-section 18·0 3-1 %
6·75 X 4·50

Cast iron and 2 729 Circle 9·62, util- 72·5 198 40000
steel lining ized cross-section (12 000)
segments 4·14 X 6·54

Cast iron Z095 Circle 9· 14, util- 65-61 137'5 35 000

ized cross-section
2x1·4+4·87
-- 71 210 000
Cast iron 2 970 Circle 9·5 0
(6-30 + 4·05)

Cast iron 2 X 1900 Circle 9·50, utilized 71 270 54000

ring cross-section
6-30 X 4·10

Cast iron 3425 + ICircle

14·10
155·96 720 80000 Grade
3·5 %
segment + 1195
' I
44 INTRODUCTION

Construction Depth
Line.
Name period country Purpose Rock quality of cover
(m)

------- -~---- -- ---------~- - - - -~ -- - -- -

H IGHWAY

Maas 1937-1 941 Rotterdam

Netherlands pedestrian
I
Vehicular + Silt 3- 4

I
I
UNDERGROUND
- - - - - -- -- - - - -- -
I

Underground 1939- 1946 London Under- Upfiil, J0-20

llford line Great Britain ground clay, max. 60
railway ballast

Metro 1900- Paris Under- Silt, 3- 5

France ground upfill sand
I ----
railway
- -
U -Balm Berlin Sand 3-5
N ord- Si.id Germany " I
- -- -
Subway Chicago Blue clay JO
USA I "

-- - - - - -- --
Tunnclbanan 1914- 1953 Stockholm Granite rock 5- 30
Sweden "

Metro 1934- 1964 Moscow ,. Sarmatian

Soviet Union limestone,
carbonitic clay

Underground 1950- Budapest I Oligocene clay

"
Hungary and

- I silty clay
- --
NAVIGATION

1--
1
Languedoc 1679- Malpas, France Navigation
-- --
Rove 1911 - 1922 Marsei lle Navigation D olomite,
France :imcstone

Genissiat ]946- 1950 I Bcllegarde Press ure Limestone

France tunnel
I
 l!ISTORY OF TUNNEL CONSTRUCTION 45

TABLE 1/i (cont.)

11 Exca-
,Ox
Lining Total Sha~ e a nd dimensions Section vatcd Excavation
~
material length of cross-section area volume per year Remark
("C) (m) (m) ( m ') ( l000 m ' ) (m')

n; ,ELS

I R. C. 1 070 24·8 X 8·40 208 223 55 800 2 lane, l pe

destri:.in, I
cycle+ pub
lie utilities
I

Partly RC
lining
15 200
(414 000)
Circle
3·73
I 11 167 24000

segments
- - - --
I concrete
Brick+ 164 600 Inner
6·7 X 4·5

-- - -- ---
R.C. 65 000 Inner 6·24 x 3·30
6·90 X 3·60
-- ------
R. C. and 22 500 Horseshoe 32 72
cast iron (4·5 X 4·5)
6·0 X 6·40
-
R.C. 10500 8·80 X 6·0 52·8 560 46000

-- - --
Cast iron 59 500 Circle 2 x 6·0 57 5500 255 000 Station tun-
and R. C. 36 500 Circle 2 x 5·5 48 4000 400 000 nels 3 x 9·5
0 3x8·5m 0

Concrete+ 8 200 2 X 5·5 48 800 Station

cast iron tunnels
3 X 8·5 m 0
I
TUNNELS

Brick 157 8-4 X 6·9 47 25

Concrete 7118 Flat arch over rect- 305 21 70 168 000

angular section
22·0 X 15·4
-- - - -
Concrete 1 150 H orseshoe 84 97 24000 Pressure
11·4 X 8·65 tunnel
I
 1:--:TRODUCT IO:-
46
excav ated annua lly is conce rned,
millio n cuyd) , and, as far as the vol ume of rock
ls of the World , in spite of the
it occup ies the seco nd place amon g all the tunne
basis of total excav ated vo lume
interr uptio n cause d by World Wa_r IL On the
tun nel betwe en· Bolog na and
the seco nd larges t tunne l is the Great Apen nine
n m3 (2·6 millio n cuyd) , altho ugh
Flore nce, the volum e being little less than 2 millio
ed a recor d figure here. Tts lengt h
the volum e of rock excav ated annua lly has attain
tunne l, the Simp lon tunne l,
closel y appro ximat es to that of the longest traffic
s (S handa ken tunne l is 27 km
altho ugh there are some water suppl y tunne l system
.
[ 16· 8 miles] long) of an even great er total length
most rapid const ructio n rate was attain ed at the New Casca de tunne l in
The
advan ce was 4 km (2·5 miles). Com-
the Unite d States , where the annua l rate of
l advan ce of the Mont Ce nis tunne l,
parin g this ,,ith the 0·9 km (0·56 miles) annua
ry is manif ested. lt will be perceived
the striki ng devel opme nt in half a centu
can be accele rated to an appre ciable exten t not
furthe r that the rate of tunne lling
but also by aband oning conve n-
only by increa sing the numb er of work ing faces,
ucing and perfe cting mode rn
tional minin g metho ds and by progr essive ly introd
large scale use of high capac ity
free face and shieid metho ds, wh ich perm it the
mech a nical equip ment.
recen tly built St. Berna rd and
Outst andin g amo ng vehic ular tunne ls a re the
ueous tunne ls those under the
Mont Blanc tunne ls and of the nume rous subaq
. These techn ical achie veme nts
River Mersey and under the Ij River at Amst erdam
railwa y tunne ls.
appro ach very closely to those of the larges t
tunne ls a rc far in excess of
The volum es excav ated for undergroun d ra ilway
tunne l proje cts. Thus ,exca va tion
those ever attain ed in any of the larges t vehic u lar
totall ed4·5 millio n m 3 (6 millio n cuyd) ,
for the Mosc ow under groun d railwa y system
of the same order of magn itude .
with the vo lume for the Lond on 'Tube ' being
n will be the realiz ation of the
The most darin g ventu re in tunne l const ructio
e which proje ct has been lately
Chan nel tunne l betwe en Engla nd a nd Franc1 y
-~ It will most likely be a railwa
a pprov ed in princ iple by the two gover nmen ts. es on
tracks conve ying road vehicl
tunne l comp rising three track s, the two side
rail platfo rm-ca rs.
icance of tunne lling in trans por-
As demo nstra ted by the abo,e figures the signif
show s a tende ncy to increa se.
tation and eco nom y in gener al is very great and
navig ation . but the emph asis was
Tunn els were built first in the intere st of inlan d
ls, ,, herea s today , tunne l con-
soo n place d o n the const ructio n of railwa y tunne
under groun d railwa ys, as we ll
struct ion is ma inly co ncern ed with highw ay and
ure or divers ion t unne ls and with
as with water -supp ly and hydro electr ic press
struct ures of ever growi ng impo r-
public utility tunne ls. The la test und ergro und
ng facilit ies and the vario us types of
tance a re subte rra nean stores , garag es, parki
subw ays, under passes and expre ssway s.

Gr,;.up
•·• BRUCK SHAW, GoGU EL, HARDI NG
and MALCQ R: The work of the Chann el Study
1961 Febr.
1958 - 1960, Proc. Inst. Ch-. Eng. Londo n
 CHA PT E R 2

PRELIMINARY STUDIES
.\ND GENERAL DESIGN CONSIDERATIONS

2.1. PRELIMINARY STUDIES

! .I I. ECONOMIC ANALYSIS

The planning of every traffic route must be preceded by an economic analysis

and no actual design can be prepared unless the construction of the route is justi-
fied economically (analysis of the efficiency of investment). This is of particular
. ignificance in the case of tunnels, which are the most expensive civil engineering
~tructures and thus their specific costs per unit length of the traffic, or transpor-
tation route are many times higher than those of an open line. Consequently
the saving in distance by the tunnel, and the volume of traffic handled by it must
be considerable enough to result in sufficiently large savings in length and operating
costs to offset the high specific construction costs.
In Table 2/1 an attempt has been made to compile from available published
data the specific costs of various tunnels belonging to different categories. Although
dollar prices are quoted uniformly, it should be remembered during the compar-
ison that the value of the dollar before World War I was three to four times as
high as it is to-day.
As will be perceived from the table, the cost per lineal metre of a tunnel in
mountainous country is from about $ 1000---1500, $ 300-450 per lineal foot,
whe reas subaqueous tunnels built as a rule in loose soil by compressed air methods
are 6-7 times as expensive. A saving of up to 30 % over this latter method can
be attained by sinking the tunnel into place (Baltimore).
The economic analysis prior to construction must be very thorough and may
proceed on the following lines.
The anticipated traffic volume must be determined first, fo r which purpose the
effective area of twice 12 km (7 1/2 miles) width, its population, consumption and
industrial production.at present and for at least the next 10 years may be taken
into consideration. This provides a basis for an estimate of the number of vehicles
(trains, automobiles, vessels, etc.) using the route. Once these data and the operat-
ing costs per unit weight and length are known together with the specific investment
costs per unit length, the economy of construction can be determined by comparing
various methods. .
Let B1 denote the construction cost of the line involving the tunnel, expressed
as B1 = La Ka + Lv1 Kv, and B 2 the construction costs of the alternative without
a tunnel, expressed as B 2 = Lv2 K"' where La, L. 1 and Lv2 are the length of the
tunnel, the open line leading to, and from the tunnel, and the open line of the
alternative without a tunnel respectively, and Ka and Kv are the total capital
 .i:,..
TABLE 2/ 1 00
Construction Costs of Some R emarkable Tu1111els

L ocation Length
km Shape
or
Diameter
width/height
Lining Rock material I Cost d ollar
per lineal m R ema rk

Railway tunnels

Mont-Cenis Brick and Volcanic rock 910

1857- 72 12-7 Horseshoe 8·0/7-30 ash lar masonry
Simplon I. Ashlar masonry Mixed rock 800
1895-1906 19·8 4·9/5·40 formations
" -ci
:,,
Simpton II. 1914- 15 19·8 4·9/5'40 400 "'
r-
" " " i
Lotschberg 1910-13 ]4·6 850 z>
:,,
Moffat 1924-27 9·9 ,, 7'4/4·80 Concrete Limestone -<
1 550
"'
..;
Great Apennine 18·6
1923-34 "
8-70 Ashlar
masonry
Marl,
limestone
l 200 ~
"'
"'
Mont d'Or 1912- 15 6·1 Limestone 855
"

Highway-tunnels

Pennsylvania turnpike 10·6 Semi-circle 6·90/4·30 R . C. Marl slate, 1 165

1939- 40 vault sandstone
Memorial turnpike 0·54 7'20/4·30 R. C. Sandstone 6200 Including
1954 a nd slate ventilation
 ...
Cl) lloll aml N, Y.
1920 27
5,ns circle (100/l tJS I (~:,,t iron ~tit m"c,1 with 'l \()() 'fotuf .t\'ClaRC
ii· rock debris (5200) Shiel.I ,cctiu11
:,-
'<
Mersey 3· I8 circle+ 19·0/5·70 Cast iron rissurcd 11 100
--1 Including ve11t1l.1-
:T 1925 34 segmen t sandstone
"' 11011 and access
~ . ramps
£, Lincoln N. Y. 4·68 circle 6·45/4·00 Cast iron Silt, mixed with 10000 Including ven tila-
--1 1934-45 rock debris tion and access
c;

""g,
ramps
;·
., Queens-M idtown N. Y . 3-84 (9·3) 6·3/4·0 Rock mixed 10000
1936-41 with debris 17 000 I Shield section only
Antwerpen 1930 33 (·74 (9· I) 6·65/4 · I Clay and loam GOOO
Baltimore m
2 X 3·77 D o uble 2 X 6-60/4·20 Steel sheet with Silt, sand 6650 n
1954-57 0
circle R . C. lining and clay 7
0
3
n
Water Supply Tu nnels )>
--- z)>
,...
Shandaken Catski ll 29·1 Ho rseshoe 3·08/3'45 Gunite Gneiss,
;:;
930 ;;;
191 7-27 limestone
Colorado aqueduct Horseshoe + 4·80 Concrete Hard rock 480

-
1935- 41 circle l ·80 Concrete Hard rock 145
--
Underground railways

Mosen"

1
circl 6·00 Cast iron Limestone 15000Rbl
circl 9·50 Cast iron Limestone 45 000 R bl
Budape~t circl 6·0 Cast iron Varved clay 45 000 Ft
circl 8·50 Cast iron Varved clay 150 000 Ft
.j>.

""
50 PRELIMIN ARY STUDIES

U denotin g
costs per lineal metre of tunnel and open line, respective1y. With 1
and U 2 the
the total annual operating costs of all vehicles using the tunnel route
the excess
correspo nding costs of vehicles using the alternati ve without a tunnel,
capital invested in the tunnel will be recoupe d after a period
B 1 -B.,
t = -
Uz- U1
than a cer-
The construc tion of a tunnel appears to be justified where tis smaller
tain common ly accepted amortiz ation period, usually taken as about JO years.
ed on the basis of consider ing relative virtual
Operati ng costs may be compar
point of the tunnel alternati ve is lying over one
distance s as well. 1f the highest
end point of the line, higher by the vertical distance m;_ and by the vertical distance
of the
m; over the other end point, and provide d that the correspo nding values
distance s are:
open-lin e alternat ive are m~ and m;, respectiv ely, then the virtual
I 000 (mi + m ~)
Hva = H + ---·-
µ
and
l 000 (mi + m;) (2.1)
H <·v =H+ - - - - - - - '
µ

where H = the actual distance

µ = the specific traction resistanc e per mill.

bends, this
Where the open-lin e alternati ve involves a conside rable number of
distance s, by introduc ing the curve
must be allowed for in calculat ing the vertical
resistanc e Kand introduc ing it into the formu la reduced in the ratio h/H, where
h denotes the total length of all curves.
in tons
Operati ng costs will obvious ly be proporti o nal to the weight of vehicles
{n), the virtual distance and the specific traction cost (a)
using the route annually
of a vehicle for a distance of I km . ~o that we may write U = Hw a n.
while
The value of n is to be determi ned from the prelimin ary traffic survey,
ance and
that of a includes primaril y fuel consum ption, track and vehicle mainten
the period
marshal ling. On introduc ing these relation ships into the formula of
of refund , we obtain:

cents/ mile)
The values of a for 1 ton and 1 km may be taken at 30·8 fillers {I ·6
and
for railways , 182 + 23 = 205 fillers (1 1 2 cents/mi le) for motor vehicles
9·3 fillers (0·5 cents/ mile) for navigati on.
 ECONOMIC ANALYSIS 51

C mparison may be effected not only on the period of refund but also on the
· of minimum annual costs of construction (capacity demand) and of operation,
basic relationship for which may be written in the following form:

Umin = LJ ·B + U, (2.3)

\\here Ll = the standard efficiency or capacity demand coefficient

B the first investment cost
U = the annual operation cost.

According to Soviet data, the value of LJ may vary from 0·05 to 0· 10.
The economic analysis must be extended to aspects of regional development
a nd of defence as well, which may in many instances overrule those of a purely
economic nature.
The construction cost of short tunnels must frequently be compared with that
of a deep-cut alternative (cf. Section 22.1). The problem arises especially in con-
nection with offspur tunnels and with the entrance section (appropriate choice
of portal location) of tunnels. The limit depth can be calculated from the for-
mula:
(2.4)

where Kb = the cost of expropriation

Kr = the cost of excavation
K,z = the cost of earth transportation
M = the cost of supplementary structures in the cut (lined ditches, drain-
age, lines and ribs, etc.)
fb the annual maintenance cost of the cut
Ka = the eventual expropriation or indemnification cost of the tunnel
site
A the construction cost of the tunnel per lineal metre
fa the annual maintenance cost of the tunnel
t = the standard time of refund.

With an assumed value for the latter the cross-sectional area Fb can be obtained
from the formula and the relationship

eh2
Fb = kh + ---
2 ,

where e = the slope

k = the crest width

yields the limit depth of the cut.

4•
 52 PRELIMINARY STUDIES

2.12. GEOLOGICAL SURVEY AND EXPLORATION

The most important phase of preliminary work in tunnelling io the careful

exploration of geological conditions. Reference has already been made to this
in the classification of tunnels according to their location (Section 1.11) where
it was pointed out that the geological environment decisively affects both the
loads acting on the tunnel and the choice of the preferable tunnelling method to
be employed.
For this reason the engineer, who is fully aware of his responsibility, consults
a geologist when considering the first draft plans for the tunnel, or other extensive
underground structure. The expert, as a result of previous local experience, or
existing geological maps, may indicate at this early stage of design certain geolog-
ical formations, or stratifications which may render construction work difficult
or even impossible. Therefore, his help may be invaluable in the selection of the
first choices. The information gained from large-scale geological maps is of a
general character only and no detailed picture of geological conditions' can be
obtained unless detailed soil and rock explorations are made.
The general location of the tunnel is governed by existing traffic or transporta-
tion interests, while the exact location is controlled by the geological conditions
prevailing in the area. An important consideration in selecting the location are
the points where the tunnel penetrates into, and emerges from the .mountain,
i.e. the location of the tunnel portals. These acting as retaining walls, are especially
sensitive to adverse stratification which may result in a tendency to sliding. On the
other hand they are just to be built in the most weathered, weakest surfacial
crust (see Fig. 2/ la).
The more carefully and accurately the geological conditions of the proposed
location and its environment are explored, the more confidently the plans of the
tunnel can be prepared and tunnelling methods seiected, i.e. essentially, the more
rapidly and economically can the tunnel be constructed.

Depending on geological conditions tunnels may be classified, according to Stini, into the
following categories:
Hony7-/.r.r.r.rnel.r (/u.rl'nel.r l'/1' .rno,m/oh .r.ri:/e.r/ . Their location is usually delicate, since t/Je rock
in mountain sides is as a rule more decomposed (weathered), of lower strength, and may ia
the case of adverse stratification involve danger of sliding as well as that of unbalanced creep
pressures (see later). A cover of adequate depth is of primary importance and should be
<leeper (50-60 m) in older rock, more exposed to weathering, than in fresh, solid young rock
formations (10-30 m) from which the decomposed surface layers were removed in relatively
recent times by glacier action.
Tunnels in mountain bases. Their location is less affected by geological conditions, since
th:: equilibrium of the mountain structure is less disturbed by driving a tunnel through tne
wide, flat formation of mountain bases. Covers of smaller thickness may also prove satis--
factory.
Offspring tunnels. These are usually adopted for shortcutting sharp bends of watercourses
on the convex side. The general principle according to which the younger the rock the less
weathered it is and the lower the rock pressures likely to arise, applies to these tunnel5 as well.
 GEOLOG ICAL SURVEY
53
Tunnels under mountain spurs, or under ridges. Rock pressur
es will here again be control led
pnman ly by the geological structur e, but the influence
of surface formati ons will also be
manifes ted. Differe nt situatio ns will arise depend ing on
whethe r the spur is penetra ted at
the base of the younge r geologi cal rock formati
ons, or higher up on the mounta in sid:c-.
In the first case the length of the tunnel section to be driven
in weathe red and fracture d rock
\\ill usually be less than in the second.
Tunnel s under offspur s constitu te a special categor y also
as far as geologi cal conditi ons
are concern ed, when an extrudi ng rock mass which
is general ly of higher strength and with
greater abrasiv e resistan ce to externa l effects than
the neighbo uring rocks has to be pierced.
Tunnels under watersheds and mouma in ranges. Tunnel s under
watersh eds are constru cted,
accordi ng to Stini, under minor ridges separat ing two adjacen
t smaller river basins, while
those under mounta in ranges pass under formati ons of
major geograp hical significance.
Conseq uently tunnels under watersh eds usually connec t valley
throats an d in the debris and
morain e deposit s accumu lated there are subject ed to pressur
es of conside rable magnitu de.
Tunnel s under mounta in ranges penetra te the rock formati
ons necessarily in a directio n
perpend icular to the strike (see later) and thus more favoura
ble pressur e conditi ons may be
anticipa ted. The usually great depth of cover may, on
the other hand, in volve difficulties
of a differen t charact er (temper ature, water, etc.).
Urban and underwater tunnels. These are built under conditi
ons and in a geologi cal environ -
ment basically differen t from the previou s ones. Solid
rocks are only rarely encoun tered,
the formati ons to be dealt with most frequen tly consisti ng
of loose rocks and soils. The difficul-
ties arising from the low mechan ical strength of the layers
are aggrava ted by their saturat ion
with water and by the necessity to minimi ze surface subside
nces which would be detrime ntal
in densely overbui lt urban areas. In tunnels under waterco
urses there is an increase d danger
of over-flooding. These conside rations have led to the develop
ment of special tunnel structu res
and tunnelli ng method s. Urban vehicul ar, underg round
railway , public utility and storage
tunnels , both subterr anean and subaqu eous, belong to this
categor y.

The purposes of geological explor ation are as follows:

1. The determ ination of the origin and actual condit ion of
rocks;
2. The collection of hydrological data and inform ation on
underg round gases
and soil tempe ratures ;
3. The determ ination of physical, mechanical and streng th
proper ties of rocks
along the proposed line of the tunnel ;
4. Determ ination of geological features which may affect
the magni tude of
rock pres~ures to be anticip ated along the propos ed locatio
ns.
Explor ations should be extend ed:
I. to the investi gation of the top cover
2. to the determ ination of the positio n and quality of subsur
face rock
3. to surface draina ge condit ions
4. to the positio n, type and volum e of water and gases
contai ned by the
subsur face rocks
5. to the determ ination of the physical proper ties and resista
nce to driving
of the rocks encoun tered.

The sequence of geological explor ations referri ng to tunnel

constr uction s may
be divided into three groups :
54 PRELIMINARY STUDIES

(a) Investigations of a general character prior to planning, which should include

the bibliographical and statistical survey of morphology, petrography, stratigraphy
and hydrology of the environment. 2 •1 This should be completed by a thorough
field reconnaissance and by surface explorations. The field reconnaissance on foot
where possible will amplify and crystallize previous data obtained from preceding
bibliographical study. From aerial photographs not only much of the above data
may be spotted, but the trained observer by'identifying the vegetative plant types
can often draw conclusions concerning the gross chemical characteristics and thus
the origin (igneous or sedimentary) of the underlying bedrock, not to mention
the clearer tracing of fault outcrops, folds, etc.
Hydrological studies involve stream drainage and spring locations, pH deter-
minations. The thermal, chemical and mineralogical characteristics of local
spring and/or surface w&ter can contribute very valuable information relative to
the nature of the bedrock and to its condition.
Geochemical investigations are designed to establish a correlation between
bedrock and topsoil. The total composition, unusual concentrations .of elements
and their relative percentage in residual soils may be a helpful asset to assess the
na;ure, extent and, within limits also the degree of decomposition of the underlying
bedrock.
Geophysical explorations (either electric-res;stivity or seismic methods) are help-
ful in the exploration of the rock-soil boundary in delineating fault and shear
zones, igneous or ore bodies, geological structures and similar phenomena.
However, comparative check core-borings are advisable.
Careful study and mapping of the joint and fracture system may greatly facil-
itate conclusions on and predictions concerning the gross geological structures
at depth.
(b) Detailed geotechnical (subsurface) investigations parallel to planning but
prior to construction, by which an improved information should be obtained
on the physical strength and chemical properties of rocks to be penetrated, as
well as on their condition (weathering, fissuration, relative density, consistency).
Information on the location and dip of layers, folds, faults, bedding planes, and
joints, as well as on the location, quantity and chemical composition of under-
ground waters associated therewith is of paramount significance. The determina-
tion of gas occurrence and rise in rock temperature in both location and extent
is similarly important.
(c) Geological investigations should be continued during construction, not only
in the interests of checking design data but also for ascertaining whether the driv-
ing method adopted is correct or needs to be modified. For this reason, a pilot
heading should be driven in advance of the working face to explore actual rock
conditions and to take rock samples on which strength tests and chemical analyses
can be performed, and occasionally for the in-situ measurement of rock stresses.

2 •1 Le !Jonde Souterrain Apr. and Oct. 1959

 - ------ -- - - - - -- - - -

GEOLOGICAL SURVEY 55

_1.21. General Geological Survey

Geological investigations of a general nature should be extended to the geolog-

acal history of the region, and to the structure and age of the various component
roe· . The layers deposited and formed in the course of the development of the
can.h's crust have only rarely remained in their original position. Cooling of the
earth's solid crust and the accompanying contraction, as well as continuous vol-
a nic action, both internal and reaching the surface, keep the deposited rock
la)ers in continual motion. Layers subject to compressive forces acting towards
~ch other, or towards an abutment which may be considered as fixed , will undergo
a deformation and become creased, ruffled and distorted, as a result of which
the} become folded (Fig. 2/ Ia). According to more recent theories the formation

iveat/Jered zone

w'eathered
_ zone
1

Solid core

s M H

H • Heavy pressure conditions

M = Medium pressure conditions
8 = Small pressure cond,t/ons
F1G. 2/ Ja. Extension of the weathered zone

of folds is not necessarily the result of lateral thrust due to tectonic effects, but
deep seated layers may become compressed by'the increasing pressures of the upper
layers whose thickness is increased by continuous surface depositions. Under this
compression the deep'"'Iayers yield laterally and in so doing give rise to folding.
Movement of the same kind is caused by isostasy, which term can be explained
by the fact that the earth's crust is composed of layers of different density which
may be regarded as floating on the fluid magma. Owing to the fluid-like behaviour
of the latter the layers of greater densities 5ink deeper than the lighter ones or,
in keeping with the law of hydrostatic equilibrium, the latter rise to a greater
height. Equilibrium conditions are continuously disturbed by erosion which
causes denudation and must be compensated by continuous rises and depressions
to satisfy isostasy.
S6 PRELIMINARY STUDIES

The folds may assume a wide variety of shapes, such as synclines, anticlines,
overfolds, etc. the main types of which are illustrated in Fig. 2/ 1b. The develop-
ment of an overflow and of a subsequent fault is demonstrated in Fig. 2/2a, whereas
the typical effects of thrust and tension on folding resp. fault-building are shown
in Fig. 2/2b.

fold-wing

Cpre

Eroded saddk

Lying folds
flexure

FIG. 2/ 1.b
Main fold formation

Movement, however, results not only in the creasing of the layers (folds), but
also in their failure and relative displacement {cf. faults).
A fault is developed where the continuity of layers is interrupted under the
bending or compressive forces set up as a result of the causes mentioned above.
The layers are broken and their ends will be displaced relatively to each other so
that the more or less plastic deformation develops into a failure either owing to
the brittleness of the material or to the magnitude of movement.
Beside the magnitude of acting forces, the shape and magnitude of folds and
faults is greatly affected by the quality of the rock. In this respect two kinds may
 -- =---=--------:---= --
-. - --

GEOLOGICAL SURVEY 57

he distinguished. Some rocks are highly resistant in transmitting the acting pres-
::.ures but owing to th"eir rigidity fail at relatively small deformations (solid rocks).
On the other hand, some will suffer deformations at low pressures but are capable
of large deformations without failure (pseudosolid or plastic rocks). The first

a;

CJ

f;

F1G. 2/2a. Development stages of an overthrust

58 PRELIMlNARY STUDIES

Compressim group is made up of all igneous rocks,

quartzites, sandstones and, up to cer-
¾~¼,$ - - ~ - tain pressure limits, limestones. All
Folding / kinds of slates, shales, mylonites and
' marls as well as gneiss and limestone
''
belong to the second group. Conse-
quently the rocks belonging to the first
: Fracfured fol.d
I I group are in a highly fractured, fissur-
' . · ed condition around the folds, whereas
~
'
:' those of the second group may show a
: \\\\\' ' : largely sound texture in the same area.
~
I I
'' The same applies to faults as well.
'
Upcast f~ult Where the fault occurred in solid rocks,
'

~~
it is likely to be surrounded by a frac-
tured zone of irregular extension, but of
considerable width and depth. In plast-
ic, or pseudosolid rocks, however, the
fault plane may be hardly noticeable.
Tension Various types of faults are shown in
Fig. 2/3. The plane along which move-
ment has taken place is termed the
Downcast fault
fault plane, or surface. The movement
itself may be downward, upward, or
even lateral, as revealed by the sketch-
es. For identifying a fault the most
F10. 2/2b. Effect of tectonic forces in the important problem is to determine the
building up of various geological direction of movement. Inferences as to
formations
the latter may be derived from the fact
that the surface of the immobile part
remains sound and smooth as if ground off, while the moving layers are fractured
and crushed because of friction developing between them. In fact, movement occurs
not only along a single surface, but along a series of sliding surfaces with de-
creasing intensity, so that there is a gradual transition along a series of sub-faults.
These latter, together with the main faults constitute the fault zone, or disloca-
tion zone. The primary significance of these crushed zones, disturbed by slips,
is that they may offer conduits for the infiltration of large volumes of water or
sometimes gases, giving rise to excessive leakage when penetrated by a tunnel.
Driving of headings and the construction of tunnels across the loose, crushed zone
is only possible by using the greatest care ·and, in order to eliminate hazards,
under the protection of heavy supports. The installation of supports is difficult
owing to excessive mountain pressures. The efficiency of blasting is reduced since
the gases escape without major resistance through the voids in the loose mass.
All these difficulties may also occur in the vicinity of a single main fault, and
in thin-layered, varved rocks similar conditions may also be expected. In the
 ~ - . . ------
GEOLOGICAL SURVEY 59

a)
Verlical (au/, Steep fault flat (au/!
V
v ~Heave

~
roW
Net slip nrow ------:1__
Heave Tftrow v '
V

Diitorlion
(fold) 1one ~ - - -
v v-fau/t
b}
Normal fault Reverse fault

FIG. 2/3. Various fault

formations

case of joints and bedding planes water seepage and consequent slipping
may appear.
Inrush of water shows itself in the dislocation zones by scouring the loose debris
into the tunnel. It continuously widens its passage until it may assume catastrophic
proportions.
Inclined faults are associated with non-uniform rock pressures on the tunnel
and necessitate the construction of box sections with invert arches . They involve
additional expense and greater hazards in construction (cf. Fig. 4/7).
Least harmful to tunnelling are those geological formations in which the earth
crust is broken under the action of tectonic forces into large separate blocks.
These blocks slide along the bedding planes over each other without greatly frac-
turing the adjacent masses. A formation like this affects a short section of the
tunnel only and the difficulties can be overcome with relative ease. The joints may
be filled with clay deposits, or may be open and convey water. Faults of this kind
were not regarded as especially dangerous by STINI even when considerable move-
ment occurred along the surfaces. Such formations are rare, however, and the
main fault is usua'lly accompanied by a number of minor sub-faults, similar to
the sets of sliding surfaces. Conditions may be regarded as favourable where the
main fault face is not extensively fractured by these sub-faults. As a rule, however,
the dislocation surface itself becomes more or less shaly by the secondary move-
60 PRELIMIN ARY STUDIES

ment of the adjacent rock masses. The dislocation zone may, thus, vary in width
from a few centimetres to several hundred metres. Fault zones are frequently
accompanied by sliding layers which cause additional difficulties in tunnel con-
struction.
The degree of fissurization of a certain rock and the extension and magnitude
of the fault zone may differ widely, depending on the rock material and on the
magnitude of the acting forces. Brittle rocks,e.g. dolomite, quartz-shale and quartz-
ite, fracture readily and over a great wid th, their crushing zone frequently being
filled with sand and debris. The degree of crushing of the contacting rock surfaces
along the sliding plane may go as fa r as the pulverization of the rock material,
when the dislocation zone will become filled with a clay-like dust. The internal
moisture content of the mountains then turns these fine-grained masses into
a paste-like substance which increases the tendency to sliding.
The site inspection of the tunnel region in the company of a geoloffe-cal expert
may be considered part of the general geological survey. This comprises the
observation of surface formations, tracing of past landslides, of vegetation and
the outcrop of springs, the shape of rock blocks and other circumstances which
may help in reconstructing the geological history of the region and deciding on
its geological structure. Particular attention should be devoted to traces of earth
crust movements. Such movements are usually indicated by surface unevenness
in the pattern of ridges, hills and valleys. The latter are practically in continual
movement owing to isostasy.
Younger orogen effects are almost invariably evidenced by the landscape in the
form of depressions, ridges, or sharp edges. These represent the weak points of
the surface, most susceptible to infiltration, denudation, weathering and other
effects. Under such points the rock is frequently crushed and fractured by move-
ments of the earth crust, even if it is found to be solid at the surface.

21.22. Detailed Geological Site Investigations prior to Planning

The exact location and method of detailed soil investigations at the site prelim-
inary to design will all be determined on the basis of the preliminary site-
inspection surveys and studies previously described.
Additional valuable information may be gained by geophysical soil investigation
which may be performed simultaneously with the preliminary site-inspection work.
This includes geoelectrical, seismic and dynamic soil investigations in which the
electric resistance of various soil layers is measured, or the changes in the velocity
or propagation of explosion or oscillatory shock waves are observed. From these
measurements conclusions may be drawn concerning the position and variation
of soil and rock layers, as well as of the contained water layers. With these
methods large areas can be explored in a relatively short time and the reconstruc-
tion of the general geological pattern is greatly facilitated. Actual soil and rock
investigations·, comprising exploratory borings, shafts and drifts can then be com-
 G::OLvGICAL SITE JNVESTIGA TION 61

menced. Of these, the rewlts obtained by exploratory borings are likely to be the
!east accurate, primarily because dry boring methods are hardly applicable (owing
to their restricted depth range) and thus in the majority of cases wash boring must
be resorted to which is not well suited for sampling.
The methods to be taken into consideration are:
1. percussion boring
2. rotary-percussion boring
3. rotary boring (Craelius method).

The first two methods are used for moderate depths and in loose sediments,
while the third is used at greater depths and in solid rock.
The depth of dry borings is limited to 100-150 m by the necessity of sampling.
This depth is most frequently insufficient fo r tunnelling purposes. In these boring
methods the rock is smashed and mixed, so that from greater depths no represen-
tative samples showing the strength or bedding and dip orientation of the rock
can be taken and no information on changes, bedding and water content of the
layers can be gained. The most useful results may be obtained by core drilling,
but reliable cores can be extracted only from solid rocks. (The orientation of the
sample, i.e. indication of the side facing North is very important.) Special
television cameras have been developed recently (Grundig- Fernauge) for inspecting
the orientation and original condition of the rock layers in the boreholes.2;~
In France instruments have been devised for measuring rock stresses and thus,
indirectly, rock pressures within the borehole (cf. Section 21.33).
A borehole spacing of 300--500 m is usually sufficient for preparing preliminary
designs, but for working plans boreholes should be spaced no wider than 50--100 m.
A dense network of boreholes is required in geologically disturbed regions and
where a great deal of variation in tunnelling conditions may be anticipated. This
is especially true for the design of underground railway tunnels. In such cases,
especially where compressed-air methods are likely to become necessary, the
boreholes should be located at a lateral distance of 10--15 m off the contemplated
tunnel axis. These should be carefully backfilled, or even concreted after explora-
,ion in order to prevent water seepage from the upper water-bearing layers into
the tunnel and later the escape through these boreholes of the compressed air
necessary to keep water away from the working-face of the tunnel.
T he boreholes are seldom deeper tha n 100--150 m. In exceptional cases, e.g.
at the Great Apennine tunnel betwee n Bologna and Florence, there were seven
exploratory boreholes of 390 m depth. In the interest of satisfactory exploration
the boreholes should, in principle, be sunk at least 20- 50 m deeper than the con-
templated tunnel bottom. Information on the stratification of soil under the tunnel
is desirable for drainage design and for the control of the proper location of the

2 2
• MOLLER, L.: Optische Sondierung im Festgestein, Vortriige der Baugrundtagung in
Hamburg 1958
'52 PRELIMINARY STUDIES

tunnel and not so much for safety against excessive settlements or overloadings.
In fact a favourable permeable layer underneath the tunnel may offer advanta-
geous possibilities for drainage.

21.23. Site Exploration during Design

Boreholes spaced several hundred metres apart yield information for preliminary
design only, but are insufficient for preparing working.plans, so that intermediate
ones must be inserted as planning proceeds. Designs must be based on accurate
information about the physical, strength, and chemical properties of the rock
layers to be penetrated and on hydraulic, gas and temperature conditions prevail-
ing in them. The data obtainable by boreholes are, however, usually insufficiently
accurate for this purpose (cf. Section 21.3).
A far higher degree of reliability can be attained by exploration shafts. These
are usually vertical, but inclined shafts may also be used exceptionally. The shafts
should be located in a manner to permit their subsequent use for constructional
and later for operational purposes. During construction they may be used for
material transportation and for starting intermediate faces, and during operation
for ventilation and drainage purposes. It is obvious that exploratory shafts permit
the direct inspection of bedding and dip conditions as well as the thickness of
layers. Furthermore, the physical and strength properties of the materials can be
determined with great accuracy by tests on samples taken. They yield reliable
information on the position of permeable layers, on the presence and character
of water passages, as well as on physical and chemical properties.
Laboratory experiments and tests from which physiological effects of water
and gases can be estimated should also be carried out during the design stage.
The drawbacks of exploration by shafts are of an economic character only,
as they are considerably more expensive than boreholes and cannot be sunk to
greater depths unless built with a large diameter and a solid lining corresponding
to the ultimate purpose to be served. The depth of shafts braced with relatively
cheap timbering is very limited and in loose soils can extend only to the ground-
water table. Even in solid rocks their attainable depths are insignificant in com-
parison with the location of the tunnel. Temporary exploration shafts are usually
built with cross-sectional dimensions ranging from about 3 m by 1·5 m to 3 m
by 2 m, while permanent shafts are, as a rule, circular in cross-section with dia-
meters ranging from 3-6 m, depending on their ultimate purpose.
Shafts afford a possibility for the more accurate measurement of rock stresses
and thus indirectly of rock pressures as well (cf. Section 21.33). The main draw-
backs of exploration with shafts is that because of the expense involved they are
spaced at greater distances and consequently changes over the intermediate section
may be overlooked by this exploration where soil conditions, stratification and
hydraulic conditions are not uniform which, in turn, may lead to unpleasant
surprises during construction.
 GEOLOGICAL SITE INVESTIGATION 63

21 .24. In situ Exploration during Construction

Horizontal pilot headings are used to explore the section between individual
shafts. As a rule they a re not driven jn the design stage but during construction,
a integral parts of the particular tunnelling method adopted. Exploration drifts
in the real sense of the word, i.e. for exploration purposes only ary.d driven before
tunnelling proper commences, are not resorted to unless the particula r section
appears to be especially da ngerous-, or where a great deal of uncertainty exists.
The usual practice is to drive exploratory headings immediately prior to construc-
tion. or ratl)er elements of the Construction method and advanced always a few
JOO metres ahead of the face. At any rate jt is desirable for thes~ drifts to be
incorporated into the permanent tunnel system. While drifts serving purely
exploratory purposes and lying outside the tunnel section may be utilized as
lateral drainage, or ventilation ducts, t he pilot headings which later become an
integral part of the tunnel section n ot only direct tunnel driving but may be
used for removing the soil . excavated fro m the section, for the transportation
"f lining and supporting material, for drainage and for the accommodation
.,f service conduits, etc. (cf. Section 62.11).
Very little can be done, of course, in adopting another tracing jf the pilot
heading reveals, e.g. adverse stratification, a crushed fault zone, or major under-
grou nd water flo w. The information .gained is much more u seful for taking pre-
cautions against construction haza rds to be encountered, and for making minor
deviations from the contemplated line in order to avoid them. On the other hand,
pilot headings afford the mos·t reliable data on the strength, stratification, water-
bearing joints, fissures of the rocks to be penetrated and give timely warning where
a change in the contemplated method or structure is necessary.
Exploratory headings are the most accurate means for determining geologi::al
conditions in tunnelling and suppl y the most reliable data on the fundamental
design problems such as rock properties, hydraulic and gas conditions, as well
as o n expectable rock pressures. R ock pressure measurements can be performed
most read ily in the headings and thus even the inaccuracy caused by the difference
between the strength observed on laborato ry samples· and within the mountain
itself - which is often considerable - can be eliminated.
According to STINI the following rock properties should be of particular interest during
the geological investigation :

1. The orientation of rock stratification (whether horizontal, sheet-like, mo derately inclin-

ed, steeply sloping, reserved, overfold, etc.);
2. The thickness of individual layers, the regularity of sequence of rock layers, o r changes
in mountain types;
3. Mineralogical co mpostition (detrimental components);
4. The crystal structure of rocks (uniformly grained, porphyric, etc.);
5. The bonds between the individual grains (strong, weak, direct, indirect);
6. The hardness, workability (drilling, blasting, etc.) of rocks;
7. The structural form of rocks (massive, stratified, shaly, etc.);
8. Internal structure (whether solid," or porous, with closed, or open voids).
'64 GENERAL DESiGN

9. Deformations suffered during the orogen process (cleavages, crushed zones, faults),
or other effects (weathering, mylonitization, kaolinization, etc.);
10. The probable bearing and tensile strength of the mountain (not rock!) at various tunnel-
sections;
11. The stability of the mountain: the character and magnitude of probable rock pressure.
12. The bulk densities and dead weights of component rocks;
13. The anticipated durability of various rock types to be penetrated, the length of entrance
sections to be lined with, regard to the danger of frost effects;
14. The depth and composition of cover above each point of the tunnel and separately
for each rock constituent;
15. Temperature conditions within the mountain;
16. Hydraulic conditions at the construction site and its wider environment. Springs and
watercourses in the vicinity should be observed continuously for years starting fmm
the preliminary inspection in order to obtain evidence concerning both draw-off and
inflow. The water should be tested also for harmful constituents;
17. The possibility of the occurrence of harmful gases;
18. The susceptibility of structures to earthquakes and artificial vibrations;
19. Surface formations;
20. Safety against air escape in anticipation of compressed air operation;
21. Hazards to structures and especially of entrance portals by forces of Nature (e.g. slides,
rock falls, avalanches, mountain-slides, slumping, etc.).

The fundamental objective of geological surveys, reconnaissance, borings,

exploratory shafts and investigations is to obtain as clear a picture as possible of
the gMlogical structure of the mountain, or river bed, as is required for selecting
the most favourable location and construction method for the tunnel.

2.13. TYPICAL GEOLOGICAL FACTORS AND THEIR EFFECT

ON TUNNELLING

21.31. The Situation and Orientation of Layers to be Penetrated

The effect of mountain structure on tunnelling is quite obvious. Tunnel con-

struction is simplified, accelerated and made cheaper by the uniformity of rock
and the greater the variation of layers, the more involved, expensive and time
consuming the tunnelling methods will be. Mountain formations, devoid of strati-
fication are much more favourable for tunnelling than mountains composed of
several layers, or shales, or granular masses of varyi_ng degrees of solidification.
The adverse effects of stratification and shaling are the more pronounced, the
better defined and the thinner the individual layers are. The direction (strike) and
dip of the layers are of paramount importance.
The location of the layers in space can be described in terms of strike and dip.
Strike may be defined as the direction of the horizontal extension of the layer,
i.e. the direction of the horizontal straight line which can be drawn on the Ja~,er.
The dip is the inclination of the layers and is perpendicular to the strike. The
straight line of intersection of the layers with the horizontal plane is the strike.
 GEOLOGICAL FACTORS 65

For complete information on the spatial position of the layers the angle of dip, i.e.
t'1e angle of inclination of the layer with the horizontal must also be determined,
since a koO\\ ledge of the strike alone does not yield positive information on the
direction of the dip, but the strike is completely defined by the dip (Fig. 2/4).
In strata that are simply tilted both
dip and trike are relatively constant
O\er \\ide dista nces, but in folded beds
nri.ations fro m both the regional dip
and regional strike are numerous.
In the selection of the location and
derth of the tunnel axis its position
rt. ative to the stratification should be
thoroughly studied.
Where the tunnel axis is perpendic-
ular to the strike of a steeply dipping
rock stratum (similar to a cross drift,
Fm. 2/4. Schematic representation
Fig. 2/ 5a), the excavation of the tunnel of strike and dip
1S likely to succeed under favourable
rock pressure conditions. However,
where the tunnel axis is parallel to a)
the strike (Fig. 2/5b) higher rock
pressures may be expected to occur.
In general, steeply dipping strata
facilitate the penetration of the weath-
ering action of atmospheric effects
into the interior of the mountain, pro-
ducing a loose crust of increased liJnnels running parallel with the strike
thickness. Otherwise steeply dipping,
or even vertical layers may be ad- b)
vantageous as far as strength condi-
tions are concerned. When driving the
tunnel perpendicular to the stratifica-
tion (i.e. to the strikes) each individ-
ual stratum must act as a girder
wi th a span equal to the width of the
cross-section, and with a considerable
depth (Fig. 2/6a). The only disadvan- Tunnels running at right angles to strike
tage of such stratification is the gen- F1G. 2/5. Location of tunnels relafive
erally poor efficiency of blasting to the strike
operations.
When, on the other hand , the tunnel axis is parallel to the strikes and
bedding planes of the vertical strata (Fig. 2/6b), bridge action is limited to the
extent until the shear strength (due to friction and cohesion between adjacent
layers) is fully mobilized, while · the inherent bending strength of the layer is
5 Szechy: The Art of Tunnelling
66 GENERAL DESIGN

not utilized unless an appro-

priate span is developed in
the longitudinal axis of the
tunnel.
Horizontal stratification,
with relatively thick layers, is
advantageous for the driving
of small drifts or passages,
since the thick layer can
safely overbridge the small
span by acting as a beam
(Fig. 2/7a).
Wide underground cavities,
on the other hand, cannot be
overbridged by thin layers (cf.
Figs 2/7b and 2/7c:) e~ecially
if the layers are fissured (Figs
2/7d, e). Under such condi-
tions a pointed arch-roof may
be advantageous (Fig. 2/7r.i).
This type of stratification is
even more dangerous where
the fissured strata dip at about
5 to 10 degrees, as in this case
the roof may spall off in the
form of blocks as the heading
is driven forward (Fig. 2/7e).
The more steeply the strata are
inclined, the greater mutual
support will develop between
them.
· Should the tunnel axis fol-
low the strike of horizontal
or only slightly dipping strata
it must be located at a depth
where solid and impervious
rocks are encountered over
the entire length and a suffi-
ciently thick cover of this stra-
tum should be left to support
the overlying waterlogged and
loose burden and to prevent
any breaking off into the
FIG. 2/6. Tunnel location in relation to various
stratifications tunnel.
 ---- - -~----

GEOLOGICAL FACTORS 67

The folding of strata gives rise to pressure on the core and tension in the crown.
icline a nd syncline folds are of special significance in tunnel driving.
th terms denote a wave-like fold, but whereas a syncline is the trough of
\\-a,e, the crest is called the anticline. If circumstances necessitate that tunnels
• the strike, they should always be located in the anticline, since on passing
ough the crest of the fold they will then be subject to lower pressures. In the

1..10-12 m_/

F1G. 2/7. Usual cavity

sections in horizontal
stratification

FIG. 2/8. Location of tunn~ls_in

syncline and anticline
formations
68 GENERAL DESIGN

in
syncline, however, they would be exposed to overpressure from both sides and
danger of inrushes ,
addition the accumulation of water there would increase the
in the anticline the water would tend rather to seep away from the tunnel (Fig. 2/8).
For tunnels running perpendicular to the strike uniform pressure conditions
-
will also be slightly disturbed - although over a rather considerable length
both in synclines, and in anticlines. In anticline s the entrance sections of the tunnel
will be subjected to higher pressures and the central portions to lower ones,
whereas in tunnels in longitudinal synclines the pressure conditions will be reversed
(Fig. 2/9a, b).

Fm. 2/9. Variation of pressures in syncline and anticline formatio n

when alignment is normal to strike

Not only the dip and strike but also the sequence of layers plays an importa nt
role in tunnelling. Uniform stratification will usually afford easy conditions both
for driving and for constructing the final tunnel section, whereas serious diffi-
of
culties are likely to be encountered where strata are highly variable. Instead
rings should be
a continuous type of lining a system composed of adjoining
adopted in this case.
Tunnels being extremely susceptible to earthquake damage, particular care
should be devoted during geological investigation to ascertain that the tunnel
should not be located in an earthquake zone. According to observations made
in mining the effect of seismic waves are much less perceptible in solid rocks
than in loose, unconsolidated layers.
In conclusion the purpose of geological inve5tigation is essentially to provide
to
advance information on pressures likely to act on the tunnel, on conditions
be expected during driving, i.e. on rock pressure s, rock strength , excavati on prop-
erties, water pressures, and volumes and temperature conditions in the interior
of the n:o:mtain.

21.32. Condition of Strata to be Perforated

The determination of the condition of rocks along the tunnel axis is one of
d
the primary tasks of geological investigation. Investigations should be extende
to tb.-e possible physi:;al, chemical or biological action to which the rocks may
ed
have been exposed during their geological history and which may have influenc
their strength to a certain extent and in a certain location .
 GEOi.OGICAL FACTORS 69

In order to explain the phenomena involved let us consider first the chemical
d!ccts ,, bich may cause changes in the condition of the rock. The principal agent
bringing about these changes is always water whose activity may be increased by
lved chemical agents, gases and by temperature conditions. In fact, water
Sttping into the rock is never chemically pure although many of the rock consti-
tatnt,. e.g. feldspar, react even with chemically pure water. The main passages
of infiltration are naturally the fissures and faults through which water can gain
acttS~ to the mineral constituents of rock. Incidentally, water may be surface pre-
cipitalion water inducing chemical weathering, or thermal water emerging
from greater depths, which niay produce changes referred to as hydrothermal
dfect;;.
Precipitation water infiltrating from the surface has a double effect, as chemically
unstable rock components 'such as feldspar, are decomposed and the water-soluble
components, together with the decomposition products are leached away. As a
R ult of these processes the surface of otherwise sound rocks is completely envel-
oped by a more or Jess weathered layer, resembling the rust on steel. The top soil
covering the external layers may be regarded as the end product of the weathering
process. The thickness of the weathered zone depends on climate (it is greater
m warm climates) and on the geological factors shaping the earth surface. The
measure of progress in weathering is the quantity of components (CaC0 3 , Si0 2 ,
1~03 , Fe 2 0 3) soluble in low concentration acids.
In water-soluble rocks, e.g. limestone and dolomite, initial fissures may develop
into regular passages and eventually into a network of coherent watercourses
and cave systems (see hydrological exploration). Deposits consisting of reddish
cla~ and limestone debris are often encountered in these cavitie5 . Dolomite is
mu..:h less soluble in water than calcite, which frequently may be dissolved and
kached away, leaving behind only the dolomite crystals from a mixed rock. These
Cl)stals disintegrate on blasting into a mass of sharp-grained sand particles com-
pletely devoid of cohesion.
The decay of plastic and metamorphic rocks depends entirely on their original
,nstituents and minerals. Sometimes it is only the cementing agent which is
,eathered, but eventually the mineral constituents may also decay. Rock compo-
nents containing feldspar may alter into clay; quartzite is insoluble, while marble
i! subject to solution only. Metamorphic rocks in the stage of weathering constitute
the group of expansive rocks (cf. Section 31.3).
In igneous rocks it is again the constituent feldspar, the chemical weathering
c,f which may reduce these rocks to clay. The prerequisite for this alteration is
again an abundant access of water. Consequently, veins of clay are likely to occur
rnainly in the vicinity of fissures. In granite this is due primarily to hydrothermal
~'fects. A special weathering process of diabase is the transformation of its olivine
C"lntent into serpentine whose expansion exerts a disruptive effect on the rock.
The fi rst sign of chemical weathering of rock is a dull (dead) and locally tar-
nished surface, accompanied by a heavy response to striking. Once subjected
'·' pressure such rocks disintegrate rapidly before any appreciable deformation
·' 70 GENERAL DESIGN

can be observed. The places of occurrence are usually coincident with those of
mechanical deficiencies, i.e. fissures and faults, aµd are usually close to the sur-
face. In other words, the portal areas are likely to be most affected. The. thickness
of the weathered outer layer may range from a few centiP1etres to 20-30 metres,
and is difficult to predict owing to the large number qf incalculable factors.
The place and extension of decomposition due to hydrothermal reactions are
most difficult to predict. Sometimes the sound, solid rock, e.g. granite, encountered
in the vicinity of the portals may show signs of hydrothermal decomposition witb
increased inward penetration.
Weathering, naturally, has a great influence on the kind and magnitude of rock
pressures and, thus, on the selection of the tunnelling method as well.

21.33. Stress, Strength and Deformation Properties of Rocks

The stress conditions in rocks are basically different from those prevailing in
soils or in younger sediments. Whereas in soils vertical stresses predominate
and can be related directly to the weight of the overlying layers, and lateral pres-
sures are lower in accordance with Poisson's ratio, in rocks the original pressure
conditions are of a hydrostatic character brought about by both hydraulic and
tectonic agents. Stratification, on the other hand, results in the development of
inclined stresses. As can be concluded further from the highly variable fissure
pattern in rocks, the direction of pressures are also highly variable during the
development of folds. However, the rock movements still taking place up to the
present day indicate a trend towards the gradual correction of existing irregulari-
ties and the creation of an ideal state of equilibrium in which the shear is of zero
magnitude. HEIM's theory (cf. Section 32.3), according to which the pressure
around any point in the interior of the rock is the same in all directions and of
a magnitude corresponding to the weight of the overburden, may be regarded
as giving a close approximation to actual conditions in rocks at great depths.
The development of such a similar distribution of stresses is hardly conceivable
unless purely mechanical pressures are combined with thermal effects, or when
other agents contribute to bringing the rock to a plastic state (e.g. the pseudo-
solid character of such rocks as gneiss, marl, slates, etc.). The usual depths in
tunnelling are too small for the occurrence of a similar stress distribution and
inevitably there will be a difference between the magnitude of vertical and hori-
zontal stresses resulting in shear stresses. As indicated by stress observations at
different places, the largest stresses are not vertical and cannot always be related
to the depth of cover. Evidence points rather to the influence of mountain building
(orogen) pressures, i.e. stresses depend on local tectonic conditions.
Deformation properties of rocks cannot be investigated, in general, according
to the simplifying ~ssumptions commonly used for metals. Rock d~formations
cannot be described by a single parameter. The main reason for this is thch-
 GEOLOGICAL FACTORS 71

anisotropy; in addition, their deformations depend on the magnitude and duration

of the acting force (rheology). Their heterogeneity is responsible for their strength
properties and stresses being non-uniformly distributed. A typical stress-strain-
time diagram of rocks is shown in Fig. 2/10. As will be seen plasticity plays an

Rupture
I
I
I "'
I~
11
,""
'-"+- - " /Principal deformation
Compaction . ~-:---,- -7
--) lo
, I n
,," I /
/ I I
I /

FJG. 2/ 10. Typical rock

deformation
diagram as a
function of time

{z}

important role in deformation and the elastic phase is preceded by the compaction
phase characterized by the progressive failure of minute internal ·particles and
bonds to resist the reduction of porosity and by the start of loss of water from
the pores (cf. Fig. 2/ 12). Full reversibility does not exist even in the elastic phase,
i.e. the original deformation condition is not restored after complete release of
stress. (Perfect elasticity is even impossible theoretically unless the load increase
1s effected at zero velocity, as otherwise part of the kinetic energy absorbed by
the molecules during deformation would be transformed into heat.)
Non-elastic deformation incurs the transformation of rock by internal modi-
fication of its structure or shape by kneading, or crushing. This also may be rever-
sible or irreversible. Irreversible, i.e. permanent deformation may again be instan-
taneous (e.g1 within 1 hour), or slow. In the case of rocks the slowness or"defor-
rnation plays an important role and may extend over months· or even years.
Finally, as already mentioned, deformations may be accompanied by volume
changes, or less frequently by external changes in shape only. Elastic deformations
entail changes in volume, whereas plastic ones (the relative sliding of the planes
of the crystal grid, the rupture planes) usually occur at constant volume. As seen
from Fig. 2/ 10 the significance of the elastic phase of the stress-strain curve of
rocks does not predominate, all phases being of equal importance. The typical
72 GENERAL DESIGN

stress-strain curves under repeated loading cycles are indicated in Fig. 2/11 ,
which shows that the three main phases persist and even the slope of the elastic
phase remains unchanged. It can also be seen, however, that different deformation
coefficients (YouNG's moduli) pertain to each of the phases. The value of this
coefficient also depends within the compaction and plastic phases on the magnitude
of the acting load. Also, the
6 value of Poisson's ra tio (µ)
characteristic of lateral defor-
mation, varies according to
the type of deformation. Its
value is usually small, e.g.
for hard and solid rocks it is
within the elastic phase 0· 15.
It continues to have a moder-
ate value in the compaction
phase also but increases in
e the neighbourhood of failure
0 to about 0·30. In the case of
FIG. 2/ 11. Typical rock deformation curves under deformations at constant vol-
repeated loading ume a value of µ = O· 5 will
be obtained.
The phase of elastic deformation is followed by the plastic, or residual defor-
mation, phase. Depending on the quality of rock this phase may result in a failure
or plastic deformation. The deformation process may thus extend over periods
of different lengths in rocks with different properties.
Slow deformation is a feature of rocks similar to the creep of concrete and,
as will be seen, plays an important role, e.g. in the development of genuine moun-
tain pressures (cf. Section 31.2).
Deformation can be measured accurately either on test specimens, or in cavities
excavated at the site, but the values observed cannot be used for the calculation
of stresses unless the modulus of elasticity is also known. Therefore, the most
accurate knowledge possible of the modulus of elasticity of the rock is essential
for estimating the forces acting on the tunnel and for dimensioning the lining.
Recently, extensive research has been carried out in this subject especially in
France where the value of E has been determined from strains observed under
loading tests performed in situ using hydraulic jacks, on the basis of the funda-
mental relationships of elasticity theory (Habib 2·1).
On the basis of these sets of experimental observations rock types have been
classified by DELAR.UE and MAR!OTTI2· ' into the following main groups.

2 •3 HABIB, M . P. : Determination du module d'elasticite des roches en place, Annales de

I. T. B. T. P. 1950 Sept.
2
-J DELARUE and MARIOTTI: Quelques problemes de mecanique des sols au Maroc, Anna/es
/ . T. fl. T. P. 1950 Sept.
 GEOLOGICAL FACTORS 73

(a) Hard and solid rocks which may be regarded as behaving approximately
like elastic materials.
(b) Fissured, massy rocks, suffering considerable deformations (compactioh)
at first loading, but which display a more or less elastic beha_viour when loaded
repeatedly. In this latter condition their £-value is considerably higher than under
the first load. 1

(c) Soft rocks, the internal structure of which is destroyed beyond a certain
limit load, and which afterwards suffer a residual deformation of increasing mag-
nitude under each successive loading cycle.
Typical stress-strain curves are shown for each type in Fig. 2/ 12. These curves
indicate that the characteristic feature of hard and solid rocks (quartzite, quartz,
sandstone and solid shales) is that their shape remains fairly constant under repeated
loading (it represents, essentially, a linear function and thus E remains constant;
Fig. 2/12a). A further observation is that the value of Eis higher when the load
acts in a direction perpendicular to the stratification, than under loads parallel to it.
In the case of fissured rock an appreciable compaction phase can always be
observed first, which is reflected by the upward concave trend of the successive
loading loops. This may also be interpreted as a sign of the closure of the fissures,
rather than of the compaction of the rock mass. Beyond this phase the chords
which can be drawn in the hysteresis loops of successive loading cycles do not
differ appreciably from one another, and the curves may be regarded essentially
as straight lines (Fig. 2/ 12b). A further observation made during repeated loading

bJ CJ
8) load mtons Load 111 /,011s
Load in Lons
75 0 10 20 25 40 50
0 52 104 156 0 50 100

f
O·Of
£,
I:;
0·02 z
!:;

~ .3
.,::i.._ O·OJ

t ~ 4
~
!}04
·!:::
-~ 5
0·05 :;/
t
I:,
6
~

Fm. 2/ 12. Typical load

deformation diagrams of various rock types under repeated loading
 74 Ol!NERAL DESIGN

tests is that the rock does not resume its original loose condition upon removal
of the load, but retains, to a certain extent, the increased modulus of elasticity
produced by the closure of fissures.
The deformations of soft rocks, irrespective of whether they are caused by
compaction or sudden failure, are irreversible. Therefore, their deformation curves
display a downward concave trend (Fig. 2/ 12c).
Accordingly, DELARUE and MARIOTTI suggesHhe use in computations of a single
value for the modulus of elasticity for hard rocks, or for fissured rocks where
lateral yield can be prevented, e.g. by suitable grouting. In soft rocks, on the
other hand, where deformat'ions are small up to a certain load limit but increase
rapidly thereafter, different moduli of elasticity, corresponding to the average
slope of the tangents over these two phases should be assumed in accordance
with the magnitude of the acting load.
It should be remembered that where the modulus of elasticity is determined
from in situ loading tests the value obtained under the first test load is not charac-
teristic for the sound rock, as the rock at the surface is necessarily fissured and
loosened owing to the redistribution of stresses inevitably taking place during
the excavation of the cavity, and so it is bound to suffer larger deformations.
In such cases the mean value of the strains observed during the first and last
series of load cycles should preferably be used.
It was also found that the modulus of elasticity can be considerably increased
by grouting the fissures and voids (BERNHARD). Furthermore, the modulus of elastic-
ity measured perpendicular to the stratification (bedding planes) and to the
fissuration was always higher than that obtained in a direction parallel to it.
The value of E can be determined experimentally, e.g. by a diametrical load
test in which a circular plate loading is pressed by a hydraulic jack against the
tunnel wall and the compression [> of the rock is observed. For this case, according
to BOUSSINESQ:
<> = 4P 1- µ
2

nD E
from which

(2.5)

where P =the resultant of the uniformly distributed load acting on the circular
plate with diameter D
µ = Poisson's ratio.

If, on the other hand, uniformly distributed water or air pressure of intensity
p is applied in a circular tunnel and the elongation [> of the diameter is measured,

pD
E = -<>- (1 + µ). (2.6)
 GEOLOGICAL FACTORS 75

(With a knowledge of E this latter method can also be used for the determination
of the rock pressure p.)
The value obtained with a loading with a hydraulic jack is always higher than
that obtained under a load which is uniformly distributed along the perimeter.
The closest agreement between the two values will always be obtained in solid
rocks.
The failure of rocks may be broadly classified into three main groups which
are, in the order of frequency of occurrence:
(a) Transformation and displacement due to mechanical sliding occurs where
the stresses developing on the exist-
ing fissure planes exceed the passive a;
frictional forces.
(b) Brittle failures comprise those Tension
resulting from the exhaustion of
internal intermolecular bonds. In
the case of such failures the re-
sidual deformation is practically
zero.
(c) Plastic failures include all
types of failures taking place with-
out a volume change and the ex-
haustion of intermolecular cohesive
bonds (flow, creep, plastic slidings,
Tension Compression
etc.), which are all accompanied by
a theoretically infinitely large defor-
mation.
In practice the three failure types
set in simultaneously, but their rela-
tive importance in any particular
case is a function of the heteroge-
neity of the rock.
Brittle failures set in without any
appreciable residual deformation, Compression
provided that the load is applied
suddenly. The strength under con-
tinuous loading may be 40 % lower
than this, since in this case the sep-
aration of internal components
owing to plastic deformation tends
to become the predominant factor. Fm. 2/13. CAQUoT's envelopes for various cases
Eventually failure may be governed of rupture
by the following three factors:

l. The highest (dominant) stress (uniaxial normal stress);

76 OENERAL DESTGN

l\1ain Physical and Structural Characteristics of Some Rocks

Water Crus hing Tensile Bending Shearing

Dry density absorption strength in dry
R ock. strength strength strength
(% of condition
(t/ m') weight) (kg/cm') {kg/cm' ) (kg/cm ') (kg/c m')
----
Batholites

Granite, granodiorit 2-50-2·75 O·l- 2 1200- 2800 40-70 100-200 50-80

Gabbro 2·92-3·05 2- 5 1500-2000 50-80 100-220 40-85

Extrusive:

Riolite 2·45-2·60 0"4- 4 800-1600 50-90 100-220 40-110

~

Fonolite 2·54- 2·66 0·5-t-2 1500-3400 60-110 110-230 40-120

Dacit 2-50-2·75 0·5-5 800- 1600 30- 80 90-200 30-100

Andesite 2-30-2-75 0·2-8 400-3200 50- 110 130-250 50-120

Basalt 2-75- 3·00 0·2- 1·5 800-4200 60-120 140-260 50-130

Diabas 2-90-3·10 0·3-0·7 1200-2500 60-130 120- 260 60- 100

-
Volcanic tuff 1·30-2·20 8- 35 50- 600 5- 45 30- 80 10-40
---
Sedimentary
.
Sandstone 2·10-2·50 J- 8 100-1200 15- 60 40-160 20-60

Coarse grained
limestone 2·60-2·85 O·l-0·8 500-2000 40- 70 50-150 30- 70

Coarse grained
limestone 1·55- 2-30 2- 16 40- 600 10- 35 25-70 15-35

Fresh-water
limestone l ·55- 2-50 I ·5-6 400- 2000 15- 50 30-90 20- 50

Dolomite 2·20-2·70 0·2- 4 150-1200 25-60 40-160 25-70

Clay-shale 2·45- 2·72 0·2-0-4 - - 200-300 -

Metamorphic

Marble 2·65- 2"75 O·l- 0·5 500-1800 50- 80 80-120 35-80

Gneiss 2·60- 2·78 1- 5 800-2500 40-70 80-200 30-70

Remark. see MoSONYI, E. and PAPP, F . ( 1959): Muszakt foldta'n, (Engineering geology) pp. 122 and LEGGET R. F . :
Geo/opy and Energineering, McGraw Hill, New York
 GEOLOGICAL FACTORS 77

T ABLE 2/II

Impact Abrasi ve Young's Po isso n's H eat con- SpeciRc

resistance resistance modulus number duc ti vity electrical Freezing
E 1000 resistance resistance
(kgcm/cm') (cm' /cm' ) (kg/cm' ) (m = I /µ) (kcal/ h 0
C) (ohm xcm)

110-160 5-7 300-700 5- 8 2·7- 3·5 l0L l0" Highly resistive

110-240 7-8 600-1000 5- 8 2·7- 4·1 1QS-J06 Resistive

120-150 5-10 100-200 5-10 l ·8- 3·0 104--105 Resistive

140-300 4---7
I 100-250 5- 9
I 1·2- 2-5 10'- 105 Resistive
I I
120-160 5- 10 80-180 5- 11 1·2- 2-4 l0L lQS Resistive

130-190 4-10 120-350 5-9 l·l - 2-5 1os-106 Generally resistive

150-290 3-8 200-1000 5- 7 I ·l-2·5 1QLJ06 Highly resistive

140-290 4- 8 300-900 5-8 I·0-2'8 105-106 Resistive

/

- - - 5- 10 0·5- 1 ·0 103-106 Variable

30-120 8- 14 150- 170 8-15 1·1- 1·6 107-108 Generally resistive

- -

70-120 15-40 500-800 5-10 ] ·5- 2·8 107-108 Resistive

- - - 8- 0·5- I ·0 105-106 Variable

- - - 8- 15 0·7-1·9 103-]05 Resistive

60-110 20-35 200-300 5- 12 1·3- 2·7 106-107 Generally resistive

- - - - 0·6-1-1 - Resistive

80-1 30 10-37 600-900 1·8- 3·0 10s Variable

- ---
- 5-10 250-600 ~ I I 1·7- 3·5 10~- 106 Resistive
7g GENERAL DESIGN

2. The difference between the maximum and minimum principal stresses

(complex stress);
3. A certain amount of potential energy of the deformation.
The first case is an exception in the case of rocks, occasionally occurring with
test specimens.
The second case is conceivable in isotropic rocks only, and can be described
by the CAQUOT envelope curve of the MOHR circles representing the smallest and.
largest principal stresses (Fig. 2/13). The shape of these curves depends on the-
rigidity of the particular material. For instance, the pure tensile strength u~ of
concrete is only about one-tenth of the pure compression strength u 3• As a con-
sequence of this the elemenatry surface AB comes to lie very close to the apex
of the curve and the angle p included by the two sliding planes developing in
tension approaches 180°. In metals, on the other band, o-3 and u; are almost
equal and the largest difference between the principal stresses does not exceed the
constant values corresponding to the pure compression strength. Brittle and plas-
tic failures are represented in the diagrams (a) and (b), respectively, while (c) repre-
sents the well-known mechanical sliding process characteristic of granular materials.
The third case can be described by MISES' characteristic failure surface. (If the
three principal stresses are plotted at the origin of the coordinate system and their
resultant is traced, the characteristic surface will be their envelope. A convenient
method of representation consists of projecting them on a plane passing through
equal intercepts on the three axes.)
These considerations naturally apply to uniform rocks only, which occur very
rarely in Nature and even these are necessarily anisotropic. The mechanism of
failure can thus be best approached on the basis of the third case, i.e. of the
rupture surface developing in the mass which is assumed to be fissured. Since
the rocks may be regarded as a heap of finely textured crystalline grains held
together by some kind of surface tension, the grains are either crushed by the
uniaxial pressure (combined with lateral expansion), when the surface stress is
superior to the inner strength between th~ grains or they may be separated in
groups if the surface force is inferior to their inner strength. If lateral expansion
is prevented, the surface stress cannot become effective and the entire load must
be carried by the fine internal texture of the grains. Should this no longer afford
the required resistance to the increasing load, the crystal surfaces will be displac-
ed on each other and the rock will be brought into a plastic state. In this--
condition a further stress increase will bring about no volume change but only
deformations of shape will take place. Energy is transmitted in the form of
shearing stresses and the deformation work is consumed by the changes in:.
shape (cf. Section 31.2; genuine mountain pressure!).
Data of major importance on some of the physical strength and deformation,:-
characteristics for the more frequent rock types are given below2-5 (Table 2/Il).
2·5 STOCKE, HERMANN and UDLUFT : Gebirgsdruck und Plattenstatik, Zeitschr. fiir Berg-
hiitten u. Salinenwesen 1934, 1936
 -~ -

HYDROGEOLOGICAL FACTORS 79

These should naturally be regarded merely as rough approximate information

values. In general, the modulus of elasticity decreases with increasing porosity,
particle size and water content. (The above figures apply only to the elastic por-
tion of the deformation!) The effect of moisture varies in extent from rock to
rock and is much more pronounced, e.g. in ·t he case of clays and clayey shales
than m sandstone. According to Stini the modulus of elasticity is reduced by mica
formation and kaolinitization, whereas it is increased by carbonic acid minerals
and lime content. It shou}d be evident that deflectjons also increase appreciably
v.ith the water content. According to STOCKE this increase may be from 3· 5 to
16-fold in clayey slates, and from l · 5 to 5-fold in sandy shales.

21.34. Hydrological Suney

The hydrological survey is carried out simultaneously with the geological

nploration to which its importance is not to be subordinated since watyr is a
governing factor in tunnel loads as well as in construction possibilities and con-
ditions.
The appearance of water in drifts and
tunnels depends primarily o n the char- : . : ....... - .· · ..
. . . ·... : :· ·.· _·..
acter and distribution of water-con-
.. · ....
veyi ng passages . The length and depth
below the terrain surface of the cav- . ·.-.
ities, precipitation and local geologi- .,. . . .

cal conditions are also important.

The passages may extend along .....
surfaces, as e.g. exfiltrations appear- .. : '. ·.: .
ing in fissures and joints, where one
·... : . t " ·II ....
dimension of the conveying cross-sec- . _· • . ~?,~

tion is negligibly small in comparison .. ~~ ~ •. . . . .· .. . . :... .. .

. I:;~\\'
.
• ..
with the other. They may again be tube- ~t ·. ·-.--:-:-_-. :·._ ·:·_._:_ :-: :·-.::: _: .-·:_.:.-·<: :-·;._:
like, ranging in size from cavities of . . . . ·. ...
several metres in diameter down to
FIG. 2/ 14. Curtain-like water infiltration
tiny seepage ways called "thread-like" from a joint
water passages.
Furthermore, subterranean local wa-
ter masses may be discriminated filling
up a certain volume (water lenses, pockets and dislocation zones). Finally,
groundwater and especially that of intercalated aquifers should be distinguished.
Solid rocks made watertight by the density of their texture are always inter-
woven with fissures, seams and cleavages in which water is moving in curtain-like
films. In igneous rocks the water film moves in fine cracks (Fig. 2/ 14) but highly
extended, distributed to countless films whereas in sedimentary rocks the dispersed
water has a tendency to unite in definite ducts by scouring the rock. Once the
,
 80 G £r,ERAL DESIG N

fissures are wiu.-: enough and commun icate with each other through a great
numu.!r
of passages, the moveme nt will be directed by the pressure bead. With the
decrease
of the quantity of water (end of thaw or dry season) the moveme nt will be
a gravity
flow. The lo nger the period of migratio n through the rock, and the smaller
th~
resistanc e of the latter against the dissolvi ng action, the greater the progress
from
water-films towards the formati on of tubular ducts. This difference
can be
observe d on the ground- surface as well, since water issues from igneous
rocks
at numcro m points, but the discharg e at individu al points is small. On
the other
hand, in mounta ins built up from more soluble sedimen tary rocks there
are fewer
springs, but with an ample discharg e, althoug h subject to wide seasona
l fluctu-
ations includin g complet ely dry periods.
Ground water and the water of intercal ated aquifers , where the voids
of the
rock are saturate d with a coheren t mass of water extendin g over the entire
thick-
ness of the layer, or at least over a conside rable part of it is the most dangero
us
in tunnelli ng. If possible , the tunnel should not be located under the
pbreatic
groundw ater table. Howeve r, where construc tion in such a layer is unavoid
able
(e.g. tunnels of undergr ound railways), special tunnelli ng methods and techniqu
es
must be resorted to (shield driving, dewater ing by compres sed air). If the
tunnel
can be located above the groundw ater table then only drainag e of periodic
ally
percolat ing meteoric water need be provide d for (Fig. 2/ 15-1). In tunnels
under
the groundw ater table rain-like dripping from the rocf and entrance
of water
through fissures of side walls can be expected (Fig. 2/ 15-2). The volume
of water
entering the tunnel in such cases depends exclusively on its height, relative
to
the groundw ater table, and decreases with this hydraul ic head. Locatio
n 3 in
Fig. 2/ 15 is the least favoura ble .
Special referenc e must be made to karstic (cavern) waters, resulting from
per-
colating precipit ation, and stored in the continu ous duct-lik e passage cavity
system
eroded in the rock by the water itself. These represen t a serious source of
danger
in tunnellin g. In the vicinity of the ground- surface the water percolat es in numero
us
fine fissures into the interior of the rock and widens these fissures in the
process.

Fm. 2/15. Po~sibletunnel

elevations
relative to
water-levels
 HYDROGEOLOGICAL FACTORS 81

The hair cracks develop into larger channels, decreasing, however, in number
accordingly (Fig. 2/ 16a). Such a continuous passage system will not develop except
in water-soluble rocks. For this reason caverns are common in dolomite or lime-
stone mountains only. Two types of karstic (cavern) systems, namely perched
(mounds) and basin-like ones can be distinguished (Fig. 2/16b, c). Water from the
interior or perched cavern formations seeps out to the surface and the water is
stored only for the period of percolation. On the other hand, in deep karstic basins
meteoric water percolating from the surface is stored in the formation as in an
underground reservoir (Fig. 2/ 16c).
When the drift pierces such an underground reservoir containing large volumes
of water in the communicating passage system of the karstic basin, this water
\\ill burst into the tunnel under a high hydrostatic head and at high velocity.

cJ

b) Ant,clmal karstic formation

(with no permanent gro1Jndwaterstorage)
Precipitation

c)
Synr:lmal carsitc for- Hydrostatic level
mal/on
Carst1c s1Jrface
'-

FIG. 2/ 16. Karstic water-flow net

6 S z~chy: The Art of Tunnelling

82 GENERAL DESIGN

Water inrushes exceeding several hundred m3/min have been experienced (Mont
d'Or tunnel, Great Apennine tunnel, Tanna tunnel, etc.). According to experience
gained in mining practice, karstic water is not stored, as a rule, in the interior
of the mountain at depths greater than 100 to 200 m, so that the likelihood of
danger from it decreases with depth. Blasting operations below the karstic ground-
water table should be undertaken with great care in order to anticipate the inrush
of karstic water through blasting-fissures.
Water ·moving in larger cavities may carry
sand, debris and even pieces of wood as well
as suspended matter, as do surface streams
(Fig. 2/ 17). The volume of inflow also depends
on the dip of the layers, as well as on the
position of the heading relative to the eleva-
tion of water m:isses stored in the passage
....... -- ..
system. High initial inflow volumts have been
--~,, found to decrease rapidly in relation with
-------
the emptying of the underground reservoir.
Geological basins and depressions (synclines)
are more likely to be filled with such water,
which is a further reason why tunnels should
Carst1c c;;vems not be located in synclines. As each drift pen-
tapped al Mont d'or etrates the interior of the mountain it gen-
tunnel
erally draws off the water from there and
also drains the surroundings to a certain ex-
tent. It has also been experienced that water
constantly entering through the face may be
drawn ahead of it by driving a pilot heading
which results in a certain dewatering of the
space in front of the face.
The inflow into the heading may also in-
crease with time if mountain pressure also
increases as a consequence of the excavation
Fm. 2/ 17. Natural underground
reservoir crossed by of the cavity involving a displacement of the
Mont d'Or tunnel surrounding rock masses towards the lining.
As a result, the rock masses become loosened
and internal fissures are opened, so incurring
an increase of the inflow. Similarly the inflow may increase with the progress of
time in fine-grained loose soils, or limestone and dolomite masses as a con-
sequence of the development of cavities due to continued solution, leaching and
weathering.
The discharge of internal springs tapped in the interior of limestone mountains
generally appears to follow a decreasing trend with considerable variabiliiy. They
may become entirely dry in periods of frost or extended .draughts and may reach
peaks at times of thaw or in wet seasons, i.e. they behave like surface streams.
 HYDROGEOLOGICAL FACTORS 83

Tur...nels retain their draining and dewatering function after their completion.
For the interception and removal of attracted water an extensive and costly system
of drains and ducts is built in the completed tunnel through which water is con-
tinuously removed, and the tunnel lining is carefully insulated. Water is permitted
to enter the tunnel in order to reduce the hydraulic pressures acting on the tunnel,
to relieve the tunnel lining and to increase its durability. Recently, objections
have been raised against this generally accepted practice. These are based on the
assumption that as long as there is a continuous movement towards the drai.nage
system, fine particles will be diss~lved, leached and scoured from the rock, which
15 thus constantly loosened around the lining and so is prevented from attaini.ng
a final state of equilibrium. In addition, where the water contains active chemicals
as well, a destructive action upon the lining itself is intensified by the continuous
renewal of detrimental agents. No progressive damage occurs, however, once the
harmful substances in still water are neutralized. The maintenance of a steady
water-table is advocated for this reason and the employment of collar-like imper-
,ious buttresses extending far into the rock is suggested as a means of attaining
:.uch conditions. These r~tain water at a higher level and flow is diverted from the
,icinity of the tunnel lining. Such buttresses may be realized in some cases by
grouting.
Others recommend the construction of intake shafts through which
water is reintroduced at some points into the fissured interior of the mountain
and flow towards the lining is diverted. The method of water reintroduction
cannot be regarded as entirely co'rrect owing to the action of the dammed water
in softening rocks due to the great increase in pressure. In fact, the problem of
drainage must be solved individually in every particular case, after due consider-
ation of all local circumstances (cf. Section 72.21).
The determination of the detrimental substances contained in the water is of
paramount importance. The damage done is usually the greater the higher the
concentration of the solution and the higher the temperature of the water (hot
springs!), and the larger the volume gaining access to the tunnel lining. It is for
this reason that groundwater is much more dangerous when moving than when
still.
The aggressive action of acids is in general the more pronounced, the higher
the proportion of cementing agents - free or weakly bound lime - in the natural
rock. Based on these experiences the use of cements with a high calcium content
is to be avoided in dangerous zones and those containing little lime (blast-furnace
slag cement, aluminous cement, trass, etc.) should be used instead. The disadvan-
tages of honeycombed ,porous concrete are well known, since an increased surface
is exposed to· the attack of aggressive water and the leaching of the free lime con-
tent of concrete by soft waters is facilitated. Of the natural rocks, dolomite and
rocks containing lime as a binding agent are most susceptible to the aggressive
action of acids.
Sulphuric acid, which is present in water originating from peaty areas and from
rocks and soils containing gypsum and sulphide ores, is generally known to be
6·
84 GENERAL DESIGN

highly aggressive, but similar damage may occur in clays containing pyrites upon
contact with atmospheric moisture.
Carbonic acid occurring in various mineral waters, peaty and moorland water,
may leach the cement of concrete and dissolve limestones and dolomites. Calcium
and magnesium are removed in the form of double carbonic acid salts. Carbonic
acid is usually a weak acid and its agg;ressive action becomes conspicuous only
after a considerable period. As revealed by recent research, detrimental action
can be attributed to free carbon dioxide alone, but its actual quantity can be meas-
ured only on the site. Of the salts, sodium sulphate (Na 2SOJ occurs in sea-water,
in certain mineral waters and in some rare salt deposits. Calcium sulphate occurs
frequently both as gypsum and anhydrite (CaS04). It is highly aggressive and
necessitates special care and precaution. Magnesium sulphate, or Epsom salt
(MgSOJ is even more detrimental than gypsum, since magnesium is a weaker
base than calcium.
Sodium chloride (NaCl) can be found in sea-water and in a great number of
rocks representing sea-water sediments. Solutions of sodium chloride are not too
dangerous, in contrast to the contaminations of MgC1 2 and MgS04 accompanying
it, which are always present in marine sediments.
Soft waters, i.e. those containing little dissolved salt are also highly variable;
they occur in igneous rocks, peat areas and glacial creeks (snow-melt water).
In the same mountain region the softness of the water increases with the coldness
of the climate. Tile smaller the salt content of such waters, the more readily v.ill
they dissolve salts, e.g. the lime salts of mortar. This type of aggressivity is not
especially dangerous except in flowing groundwater where saturated masses are
continuously replaced by unsaturated ones. Aggressivity is increased by the carbon
dioxide content and is characterized by the pH value which should be taken into
consideration wherever it is lower than 6· 5.
In many instances tunnelling affects surface water courses. Springs may become
dry, lakes may disappear and creeks may vanish as a consequence of tunnel driv-
ing, the reasons for which should be self-evident. The determination of the number
and yield of existing springs and water courses therefore constitutes an important
task of the preliminary site survey.

21 .35. Gases and Rock Temperatures

Another important part of the preliminary exploration work is the exploration,

estimation and study of gas outbursts, gas exfiltrations and rock temperatures.
Gas and temperature are significant for the safety and health of workmen, but
their influence on the technical feasibility of tunnelling is less pronounced although
the necessary precautions, the selection of working and transportation equipment
are factors certainly affecting the technology adopted.
 GASES, TEMPERA TURES 85

21.351. Gases occurring in tunnelling. The occurrence of gases is governed fun-

oamentally by the geological structure of the mountain. Detrimental gases may
be expected primarily in igneous rocks and especially in fissured, weathered for-
mations of such an origin.
Carbon dioxide (CO2 ) is the most important of the detrimental gases which is
likely to be encountered in igneous rocks in the proximity of coal layers and other
organic deposits and in fault zones. It is usually the product of the slow oxidation
of coal, carbonatization processes, or of the decay of organic compounds. Owing
to this its probable penetration into underground cavities can often be predicted.
mctimes it rises from the molten magma core of the Earth along fissures and
faults. Carbon dioxide also occurs in the form of post-volcanic exhalation. It is
precipitated in the form of secondary acids but is never decomposed. The gas is
colourless, odourless, slightly sour, I ·53 times heavier than air and therefore accu-
mulates at the bottom of the drifts. The detrimental effect to the human organism
cesults from the lack of oxygen associated with its occurrence. CO 2 is harmless
as long as oxygen is present in a proportion higher than 20 %, but a CO 2 content
of even 4-6 %causes difficulties in breathing, I (,L 20 %results in a loss of conscious-
ness, while 20-25 % may cause lethal poisoning (see table on pp. 87). Its biolog-
ical effect is reflected by agitated lung action as the CO 2 level in the blood increases
and breathing becomes deeper. It is aggressive to the concrete of the lining, but
its corrosive action is particularly dangerous to steel structures.
Carbon monoxide (CO) is lighter than air (0.97) and is much more toxic than
carbon dioxide. The gas is tasteless, colourless and odourless and its presence is
impossible to detect in time to prevent casualties without the use of deiicate
instruments. It occurs in the vicinity of coal fields, as a consequence of methane
or coal dust explosions and also in the exhausts of internal combustion engines
and in explosion gases. The gas is somewhat unstable and turns to carbon dioxide.
It readily combines with the red blood cells to form carboxyhaemoglobin which is
sufficiently stable to prevent those cells from performing their normal function
of forming oxyhaemoglobin which is essential to life. Poisoning can be detected
at 25 % blood saturation. Above this limit agitated heart pounding, headache,
dizziness, general weakness occur, while loc;s of consciousness occurs at 50%.
Above 75 % saturation, rescue attempts are unsuccessful. Since saturation depends
on the retention period as well, even a small CO concentration may become dan-
gerous with prolonged exposure (cf. Section 46.52 and Fig. 4/ 134).
Marsh gas, or methane (CH4) is a tasteless, odourless, combustible gas, also
lighter than air (weighing only 0·558 times as much). It occurs principally in the
vicinity of coal and oil fields, but it may also result from the decay of organic
substances. It may also be found in the vicinity of salt deposits and bituminous
layers. In some exploration boreholes of the Budapest underground railway the
methane gas encountered resulted from the decay of large quantities of fish
remains enclosed in deep layers of clay in geological times, while the methane
inflows responsible for large fires in the Great Apennine t•mnel issued from
shale strata. Methane may be tapped from similar inclusions in dolomite and lime-
r

86 GENERAL DESIGN

stone rocks. Where escape to the atmosphere is prevented by an impervious top

cover (clay), the gas may migrate laterally to great distances through fissures in
adjacent rocks.
Methane, especially in 5· 5- 13·) % proportions forms a highly explosive mixture
with air (pit gas), but is not toxic otherwise. The use of open flames or burning
lamps in an atmosphere contaminated by methane is dangerous. Concentrations
even as low as 2 %are inflammable and may initiate an explosion in the presence
of coal dust.
After the fire catastrophe caused by a gas explosion in the Great Apennine
tunnel, safety precautions were taken, the main items of which may be summarized
as follows 2· 6 :

1. The use of battery-powered electric lamps;

2. The installation of gas-indicator lamps in all drifts;
3. The use of remote controlled electric detonators ;
4. Permanent supervision of all working activities by gas experts;
5. The use of compressed air locomotives for the transportation of debris and
construction material;
6. The installation of high-pressure water mains for fire extinguishing;
7. Air-extracting ventilation at all points of accumulation (especially at the
roof) and the provision of ample artificial ventilation in general;
8. Strict prohibition of smoking and of the use of open flame lamps in the entire
tunnel.

Hydrogen sulphide (H 2S) is l ·7 times as heavy as air, smells like rotten eggs,
has a sweetish taste and is highly toxic. It supports combustion and in a quantity
of 6 %in air it is explosive. The gas enters drifts and cavities through fissures and
underground streams which have come into contact with decaying organic sub-
stances or volcanic fumes. It is essentially the disintegration product of orga nic
substances containing sulphur, or sulphates and sulphides (decaying wood),
which is readily absorbed by water. For this reason water inrushes are frequently
accompanied by the inflow of hydrogen sulphide. It occurs in blasting stythes
and may fill the cavities and fissures of salt-bed formations. Hydrogen sulp,hide
is dangerous on account of its toxic effects, rather than on account of fire hazard.
As a strong blood poison in small quantities it causes dizziness and sickness;
it attacks the eyes and causes the slightest wounds to suppurate. In a concentration
of 0·05 % it causes sickness, at 0· 1 % unconsciousness and is lethal at higher
levels (cf. table given below). It is highly corrosive to concrete linings.
Sulphur dioxide (S0 2) is a colourless, pungent, asphyxiating gas which dissolves
readily in water and combines with it to form sulphuric acid. This gas also occurs
in volcanic fumes and is very detrimental to concrete linings.

2 6
• La Direttisima Bologna-Firenze, Annali dei Lavori Publici 1934 LXXII 37
 -- - -- --- - -- - ~-
- _-

GASES, TEMPERA TURES 87

HJ-drogen (HJ is 15 times lighter than air and is not detrimental to health,
t occause of its inflammability its presence always involves the hazard of explo-
n and fire. It develops readily on salt deposits, or in their vicinity, but it may
be an adulteration in methane, carbon dioxide and nitrogenous gases as well.
Nirrogen (N) is a non-toxic gas, slightly lighter than air and is not inflammable.
It may become dangerous by displacing oxygen from the upper part of headings,
thus causing suffocation. Nitrogen occurs in young eruptive rocks, or as the
decay product of organic substances of floral origin, but it has been encountered
m coal layers and coal bearing rocks, in fissures of rock-salt deposits and sometimes
mi.'{ed with methane and hydrogen.
Vitrous gases (NO, N0 2) are the by-products of explosion fumes and may be
~') dangerous in blow-out blasts, since they are even more detrimental to health
than carbon monoxide. They have a sharp odour and are pulmonary irritants
C"t"Cn in minute quantities.
Water vapour {H 20) is in itself entirely harmless, but may adversely affect the
progress of work. In an atmosphere saturated with vapour the working ability
of the human organism is greatly reduced, the corrosion of metal machines and
tools, as well as the decay of timber supports is considerably increased. Since the
,-apour content of the atmosphere in underground cavities is naturally raised
by the moisture in the interior of the mountain, one of the basic aims of ventilation
should be the reduction of this vapour content (cf. Section 46.5).
The"more frequent gases encountered cause lethal poisoning in the percentage
concentrations given below:
co, co [ H ,S so, NO or NO,

A spiration for 20 to 30 min 10-20 0·05 0·05 0·0l 0·0l

short exposure 20-25 0·l 0·2 0·05 0·025

21.352. Rock temperatures to be expected in the interior of a mountain. Temper-

atures on. the surface of the Earth's crust are subject to wide variations and are
governed primarily by external conditions, such as season, geographical location,
climate, etc. While sand and rock soils alike are heated to high temperatures by
the sun in summer, their surface in winter will be colder than the.air. Temperature
fluctuations may exceed 50 °C . These surface fluctuations, however, become less
a nd less perceptible in the temperature of rock with increasing depth below the
surface and are no longer effective below a depth of 20--25 metres. Below this
crust affected by external influences there is a consistent increase in rock temper-
ature with depth. The rate of increase is not uniform and is governed by several
factors. It is measured by the geothermal step defined as the vertical distance over
which there is a temperature increase of 1 °C. The inverse of this is the geothermal
gradient, expressing the temperature increase for every I m depth. The geothermal
step depends on several factors, the principal one being the material of the moun-
tain itself, i.e. the thermal conductivity of the rock. The higher the conductivity,
 88 GENERAL DESIGN

the higher is the value of the geothermal step. Data on the thermal conductivity
of the more frequent rocks are compiled in Table2/I ll. (For references : the thermal
conductivity of water at 18 °C is 5·0, that of rock sa!t is 41-45 and that of afr
0·205 kcal m- 1 h- 1 for every I °C difference in temperature.)
Accordingly, thermal conducti on in air is 25 to 30 times lower than in water
and 10 to 20 times lower than in most rocks. Actual values may deviate consid-
erably from these laborator y results owing to the effects of fissures, cavities and
the water content of the rocks.
The value of the geothermal step is lower in loose, frozen and dry rocks and
may be reduced by chemical processes that may take place in the rock. The step
is reduced and consequently rock temperat ure is increased by gases trapped in
the rock. 2·7
Temperatures are further frequently increased by mineral oil, coal and especi-
ally by ore deposits, i.e. they reduce the value of the geothermal step. Temperatures
increase similarly as a result of fissuration caused by rock pressures , or of the
increase in porosity. The influence of porosity can, naturally, be traced back to
the presence and movemen t of air in the voids.
An influence still greater than that of air on thermal conductivity is the infiltra-
tion of meteoric water, which, apart from the approximately 25 times higher ther-
mal conductivity of water, results in the expulsion of air from the voids and the
wetting of rock surfaces. These factors explain why the value of the geothermal
step in cavernous limestone mountains is appreciably higher than in fissured
eruptive rocks. Water moving slowly in the fine system of joints in eruptive rocks
is a much less effective cooling agent than when flowing at higher velocity in wide
gouges in sedimentary rocks.
The value of the geothermal step is considerably a.ffected by the topograp hy of
the terrain. Under otherwise identical conditions the geothermal step is higher
under hills than under valleys (Fig. 2/ 18). Accordingly, the lines connecting

b;
a;

Ground surface

~ Geo,solherms
--== =--~ ----
fx
!J·fbJT !J• blr
FIG. 2/ 18. (a) Variation of isotherm distances in valley and ridge formation s;
(b) compared with theoretical sinusoidal curves

M During the constructi on of the Great Apennine

tunnel under a cover depth of some
hundred metres, temperatu re suddenly i.:1creased in clay-shale from 27 °C to 45 °C
and excep-
tion ally to 63 °C asa consequen ce of gas inrush (4000 g/lit CH content).
4
 GASES, TEMPERATURES 89

points of the same temperature (geoisotherms) will be more widely spaced under
hills tha!l u!lder valleys.
Finally, the value of the geothermal step is affected to a considerable extent
by the stratification and dip of the rock layers as well. Heat in rocks is conducted
better in a direction parallel to their stratification bedding, or shelving than per-
pendicular to it. For this reac;on the geothermal step is higher in steeply indined,
or vertically stratified rock layers than in almost horizontally bedded ones. This
difference is especially great in rocks containing high proportions of mica where
it may be even 2-5-fold. Dense stratification, i.e. a close succession of thin layers,
tends to minimize the value of the geothermal step owing to the insulating effect
of layer interfaces.
Maximum temperatures in the tunnel depend, ·finally, on its length, as will be
demonstrated by the following theoretical considerations and by the tabulated
values.
Maximum temperatures and the value of the geothermal step as observed in
the major Alpine tunnels have been compiled in Table 2/III.

TABLE 2/i.I

Name
Length
of tunnel
Cover
depth
I Maximum
tempera-
Geo-
thermal
Inflow
of water
ture step Remark
of tunnel
(m) (m) I •c ("F) (m/"C) (1/s)

Ricken 8 604 530 23·5 (74) 27 2 (N) Dry sandstone under

(450) 21 (S) 45° dip, Methane
Simplon 19729 2135 55·4(132) 37 1300 Geothermic step
under ridge 43·5 m,
under valley 29 m
Tauern 8 551 1567 23·9 (75) 49 Granite-gneiss,
locally mica schist
Arlberg 10250 715 18·5 (65) 38·6 so Gneiss with mica
Lotschberg 14605 1673 34 (95) (25-49) Granite archaic slate
45
Gotthard 14 998 1752 40·4 (87) 47 30Q-400 Geothermic step
45 resp. 29 m
Albula 5 886 750 15 (59) 49 100 Albula granite
(912) (I 1-25) (58-55', (250)
(52-77)
Mont Cenis 12 236 1610 29·5 (85) 58-4 7 Sandstone with 50-80°
dip and limestone
with 20-30° dip
K arawanken 7 976 370 17-9(64) 28 60 Under a cover depth
916 15·0 (59) 144 of 370 m 17·9 °C a
very steep stratifi-
51.!fam cation
I
(Cauca~usJ 3 990 510 20·5 (69)
Apennine
I 18 500 2000 63-8 (145)

i I
90 GENERAL DESIGN

It can be clearly seen that under the influence of a variety of combinations

of factors the temperature within the tunnel and the value of the geothermal
step (G) may vary between very wide limits (the latter fro~ 27 to 144 m!). In
general the geothermal step G may be taken at around 3()-40 m f° C . In Hungary,
in view of the geologically younger covering layers a value from 12 to 25 m/°C
may be assumed. Stini has given the following values for the long European tun-
nels:
Depth Geothermal step
Tunnel
(m) (ft) (m/OC)
I (ft/° F )

Simplon 2100 7003 65·3 121

St. Gotthard 1725 5748 85·3 154

Mont Cenis 1565 5282 104 192

The temperature likely to be encountered in the interior of the mountain is

governed, according to Andreae, by the following factors:
1. The position of the geoisotherms under the mountain ranges (geothermal
step);
2. The soil temperature on the surface over the tunnel;
3. The thermal conductivity of the rock and hydrological conditions;
4. The elevation of the tunnel.
The annual mean temperature of the ground surface (t0) can be derived from
the annual mean temperature of the air (/1) as:

(2.7)

the annual mean air temperature at a known location

= the height difference between the point under consideration and the
one with the known mean temperature /10
X = the height difference causing a 1 °C drop in air temperature
(150--220 m).
For the temperature within the tunnel to be built at depth h we may write

h-C
T = /1 + k + -G-, (2.8)

where G = the geothermal step

/1 = the annual mean air temperature
 GASES, TEMPERATURES 91

C = the thickness of the cover affected by the external temperature

It = the total overburden over the tunnel
k = a correction factor expressing the difference between the air tem-
perature and terrain temperature, given by Bendel in the following
tabulated form:

~-ation above sea level

(m) 0 500 1000 1500 2000 2500

Factor k . c·q 0·8 l·O I-3 1·7 2'3 3·0

The temperature of the terrain surface is also affected by its slope, topsoil
co,er and vegetation.

In a mathematical analysis of the problem ANDREAEu concluded that when considering a

,.-cnical section along the tunnel axis in order to examine it according to the law of cooling,
•1th the primary aim of determining the influence of surface formation, the two-dimensional
beat distribution in this section in a ~·-:ady state can be described according to Fourier by
the following differential equation:

(2.9)

As demonstrated by THOMA, one of the solutions of this equation can be written in

the for m
,(} = C1 In (A + .JA 2
- 1) +C 2
, (2.10)
where
A = exp T2x + [1 - (2 cos T
2y - 2x) exp T
exp T 2x] 11
'
, (2.11)

if the surface is bounded by a regular wave form. The curves represent wave forms. The
period of the function is/= nb (Fig. 2/ 18b), while C 1 and C2 are constants. Wave troughs
are spaced at y = nbn, while the wave crests are at y = (n + 1/ 2) bn.
From these equations we obtain for points under a wave crest:

(2.12)

the simple relationship, from which the temperature at any depth x under the peak of the
wave crest can be calculated, once the values of the constants C1 and C2 , as well as the origin·

2 ·8 ANDREAE, C.: Die Temperaturprognose im Tunnelbau, Schw. Bztg 19S3 41-42

 92 GENERAL DESIGN

of the coordinate system and the position of the y axis are determined. At the same
point
the value of the geotherma l gradient is

d{} 2C1 exp ; (exp ; + )exp 2bx + 1)

dx = - b - · - -x- (~ -x-----;= ,==2===-)-- (2.13)
exp b exp b + \J exp T + I + 1

wh,.:h approache s with increasing depth to the maximum· 2Cif.b. Consequently, the
inverse
thereof, the value of the geothermal slep decreases under a wave crest with increasing
depth
until the minimum b/2C1 is attained.
Under wave troughs the value correspon ding to y = bn is

{} = 2C1 1n (exp;+ J exp -

2: - 1) + C2 , (2.14)

ar:cl the geotherma l gradient is

d{} 2C 1
dx = b (2.15)

i.e. it decreases with increasing depth x and tends to the minimum 2Ci/b. The geotherma
l step,
on the other hand, continues to increase until the maximum value of b/2C is attained.
1 Thus,

(2 .16)

will represent the normal geothermal gradient at great depths. Once this value is known
for
a specific type of rock, the magnitude of the cnnstant C can be computed as
1

(2.17)

For the determina tion of the position of they axis, the depth x under wave crests should
be expressed from Eq. 2.14 and the depth x 2 under wave troughs from Eq. 2.12. The amplitude
of the {} isotherm lines is thus obtained as the difference D = x - x •
1 2
Consequently
exp _f>_-_ C_2 + exp (- _{}_-_C_z)
2c 1 2c 1
D = b In ------- -'-- - -- --_;;_-'- (2.18)
exp _iJ_-_c_z - exp I - _{}_-_c_z) •
2c1 ~ 2c1
wher,ce
D
exp-,; + 1
Ci = {} - C 1 In - / J- . (2.1)
e,.p , ; - 1
 GASES, TEMPERATURES 93

It follows that once the amplitur.;i of a {} isotherm has been found, the C2 constant can
also be determined. No more than the geothermal gradients and geothermal steps at the tunnel
adits and exits and under the greatest overburden need be computed. Intermediate isotherm
points can be plotted by interpolation. If the vertical section of such a wave does not deviate
too much from a regular sinusoidal wave line, the theoretical shape of the isotherm can be
caJ..,ilated from the expression:

)l
x =-
bl
2
l
In
( exp 4_f>_~ 1_C2·+l+ 2exp2-f}-~_1C_2
f} -
8exp2---2
,C
C1
1\
- 2)- lo
(exp _f>_
2
- _ C_2 + 1
C2
exp---
~~
C1
- cos 2: l·
(2.20)
In spite of the simplifying assumptions made, the information supplied by this analytical
method concerning the influence of terrain formations on the value of the thermal gradient
and the geothermal step is very valuable.
An important role is played in the above equations by the constant Ci, the magnitude
of which depends on ex, i.e. the normal (standard) gradient specific for the rock, or essentially
on the thermal conductivity of the rock material.
Concerning the magnitude of ex it was found by KoNIGSBERGER and TuoMA that the influence
of the type of rock is much less pronounced than its stratification and especially the inclination
of the layers, ' as is clearly revealed by Table 2/IV.

TABLE 2/IV
Effect of Dip 011 Geothermie Gradient

Kind of rock Vertical 45° Horizontal

c·c) c c>
0
c c)
0

Gneiss, protoginc 0·027-0·028 0·033 0·034-0·036

Granite, limestones 0·027-0·028 0·033 0·034-0·036
Crystalline schist 0·027-0·028 0·034-0·036 0·037-0·041
Fillites 0·027-0·028 0·031 0·032-0·033
When in wet conditior., values may
be subject to a further reduction of 8% 5% 8%

The above method is also illustrated by ANDREAE by a numerical example. 2 · 9

When plotting the thermal conductivity values for different directions referred
to a given point in the rock,. a rotational ellipsoid is obtained, the minor axis
of which will be perpendicular to the stratification.
The effect of moisture is very considerable and temperature values appreciably
lower than the computed ones are an indication of the proximity of water.

29
• ANDREAE, C. : Die Temperaturprognose im Tunnelbau, Schw. Bztg 1953 41-42
94 GENERAL DESIGN

The estimation of probable temperatures in the tunnel are important because

of the deterioration of working conditions at temperatures above 24-28 °C.
Conditions can be improved first of all by increased ventilation, secondly by water
spraying, and sometimes even by the introduction of refrigerated air. The output
and power demand of cooling and ventilating equipment to be provided will,
thus, depend on tunnel temperatures and, therefore, its determination is required
in advance. Occasionally even a different alignment must be selected for the tunnel
where preliminary estimations reveal excessively high temperatures over thre orig-
inally selected route.

2.14. GEOLOGICAL PROFILE ALONG THE TUNNEL AXIS

All results of preliminary geological survey and exploration should be united

in the geological profile.
The main items to be indicated here are the location and depth of boreholes,
exploration shafts, drifts, etc. together with all information on the rock obtained
otherwise.
Beside the bore log in the tunnel axis and the location of the tunnel, the geolog-
ical profile should display all rock types, their condition (fissured, weathered,
etc.), detailed information on stratification, folding and fault zones and, where
possible, even strength properties. Hydrological conditions (groundwater table,
intercalated £quifers, artesian water level, etc.) must also be shown, together
with water gouges, springs and water-bearing layers. A very important supple-
mentary feature of the geological profile is the curve of the estimated internal
temperatures (cf. Figs 2/20 and 2/21).
Geological profiles of subaqueous and urban tunnels would be incomplete
without the indication of the bottom of ground-surfaces, of the extent of level
fluctuations, of the river bed material, its physical properties, especially its imper-
meability. In addition to these the weight, foundation conditions of major buildings
on the surface, elevations of possible access roadways, the location of public
utilities, elevations of various groundwater stages together with the pertinent
beads should also be entered (cf. Figs 2/ 19 and 2/26).
The object of the survey preceding actual tunnel construction is, essentially,
to furnish preliminary information on all circumstanc~s affecting the site, loca-
tion, construction and dimensions of the tunnel, in particular the quality and
position of the layer to be penetrated, on rock and water pressures and on water,
gas and temperature conditions within the mountain.
With a detailed knowledge of the geological pattern the correct method of
construction can be selected and the tunnel lining, drainage, insulation, etc. can
be designed economically and with adequate safety.
Tunnelling usually presents no particular difficulties in eruptive and paleotypal
hard rocks. Although the co~t of excavation is usually higher, no temporary
 GEOLOGICAL PROFILE 95

Butldmg block N! 1 Butldtng block NP II 8lJlldmg bluck N! Ill

FIG. 2/ 19. Typical geological profile of a subway

supports are necessary and the possibility of a full face method ensures rapid
advance and economical work. Thin linings will provide ample safety against
the small rock pressures.
In soft rocks (limestone, dolomite, Tertiary rocks) excavation is cheaper, but
the installation of temporary supports is slower, more expensive and less safe.
In fissured rocks large-scale timbering, solid linings and costly drainage are
usually required.
The greatest difficulties are likely to arise in alluvial rocks in the vicinity of
the surface and at moderate depths. These rocks are usually completely saturated
with water. Especially adverse conditions may be encountered in young soft
clay, sand and fine sand soils, especially if they are waterlogged. These types of
soil are likely to be penetrated during the construction of underground railway
and subaqueous tunnels, and special construction methods (shield driving) have
to be developed to overcome the difficulties.
Critical sections in tunnel construction may be those around portals, i.e. adits
and exits which in the case of a stratification running parallel to the slope may
be subjected _to loads (due to creep-pressure) exceeding the geological pressure
~fi~~~- .
In the light of the foregoing considerations let us study the geological
profile of some of the better known tunnels at home and abroad.
 '°
""
Gn • &neiss
J • Jvra limestone
T , Triassic limestone
P • Schrst

\~t·
\
\

\ I

~.,,.
I

I
I C')

,,"'"'z
I
I
I
/Jse!le >
t"'
0
!;l
g
km

.:-m. 2/2C. Geological profile of the Simpion tunnel

 GEOLOGICAL PROFILE 97

A 6r tinA p
Gr Fp Sp 6r8fld &'rl
0 I 2 J 4 5 6 7 7 6 5 4 J 2 I tl km
G, - Granite P - Mica schist
Gn - Gneiss Fp - 8/ack slate
C - Cir,o/ine Sp - Serpendne
0 Dolomite A - Amphibolite
FIG. 2/21. Geological profile of the St. Gotthard tunnel

1be geological profiles of the Simpton and St. Gotthard tunnels are shown in Figs 2/ 20
and 2/ 21, and both have been completed from actually observed temperature data. As can
be seen, the rocks to be penetrated consisted mainly of igneous, paleotypal rocks in the case
of the St. Gotthard tunnel, while gneiss, limestone and shale layers predominated over the
length of the Simpton tunnel. In the case of the St. Gotthard tunnel the layers are not folded,
are mostly vertical and the tunnel axis runs perpendicular to the strike. The rock layers
over the Simplon tunnel are extensively folded, the dip is highly variable, sometimes approach-
ing the horizontal, yet the tunnel runs generally in a direction perpendicular to the strike.
There are differences in the greatest depth of overburden also, this being 1725 m in the case
of the St. Gotthard tunnel and 2135 m over the Simpton tunnel. It will be perceived that
considerable differences exist between the geological profiles of these two deep Alpine tunnels
in spite 6f their similar character. These differences were reflected by the conditions
encounter-
ed in actual tunnelling work. First of all the observed differences in temperature were conspic-
uous. While this was 30·8 °C in the St. Gotthard tunnel, in the Simplon tunnel it reached
55·2 °C. This difference cannot be explained by the difference of 380 m between overburden
depths, particularly because the occurrence of the peak temperature in the St. Gotthard tunnel
did not coincide with the section under the greatest cover. An essential difference is to be
detected, however, in the dominant direction of stratification. Whereas the layers over the St.
Gotthard tunnel are vertical and . consist of granite having an average heat conductivity
coefficient ;. = l ·81, the layers over the Simpion tunnel are inclined mostly parallel to the
slope of the terrain and consist of Jurassic limestone, gneiss and shale with heat conductivity
coefficients of 1 · 12, 2 ·20 and 0 ·9, respectively. The other significant difTerence was in the rock
pressure encountered, insofar that no genuine mountain pressure was to be coped with at
all in the case of the Gotthard tunnel because of the high strength of rocks perforated, true
rock pressures (i.e. genuine mountain pressures) of high intensity had to be overcome, at
a dista,,ce of 4 km from the adit and over the central section in the Jurassic limestone
of the Simplon tunnel, howeve(, in agreement with the theory that under deep overburden
layer-a limestone suffers plastic deformation and behaves like a plastic rock (cf. Section 3.12)_

7 Szechy: The Ah of Tunnelli111

98 GENERAL DESIGN

- · - Asnmed svrface line Jbove tunnel

~ G/3ciers
- - CJ!c11/Jted rock lemperJt11re(al}er Prof. Bibolinj
- - /1eJsvred temperJtvres.tJ distJnce 2140
of om from tbe f;ce
ITALY FRANCE'.

Glwerde
To11/e

51
50 50
45 --- __ _

'S
JO JO
25 25
20 20
15
10
5 '
O.....__-.u.._....L....._.__-"',.,_-'-.......,__J.......,.___._--'-~...t_'--'-.L.L.L.LIO
Sedimentary &, ~ &, &,~ ~ ~ &, ~ ~~ ~ ~
strau ~ "' ~ ~"' ~ ~ ~ ~ <>;~ ~ ~
FIG. 2/22. Geological profile of the Mont Blanc vehicular tunne1

Fig. 2/22 shows the geological profile of the recently built Mont Blanc highway tunnel.
The rock penetrated consisted chiefly of protogine (a transitory species between granite and
gneiss) and of crystalline shales. Serious difficulties were encountered in the sections where
the protogine was modified, i.e. where it was in a crushed state owing to the action of previous •
excessive geological pressure (milonitization). Genuine mountain pressure had to be overcome
here (see Section 3.12) and densely spaced roof-bolting and a considerable inner supporting
system had to be applied as a protective measure.
Another interesting feature of the figure is the representation of isotherms (Travaux, 1963
Jan.). The temperature values actually measured are far below those obtained by prior calcula-
tion (after Prof. BIBOLINI). This nearly 100 % difference may be well ascribed to the tremendous
cooling effect of the covering glaciers.
The geological profiles of the Great Apennine and Tanna2 • 10 tunnels are reproduced in
Figs 2/23a and 2/23b. The comparison of the two tunnels with each other and with the forego-
ing examples is highly illustrative. First of all, overburden depths of these two tunnels range
from 400 to 600 m (1300-2000 ft) only, so that temperatures calculated on the basis of the
geothermal step should have been appreciably lower than those actually observed in the
Alpine tunnels. Yet temperatures as high as 60 °C were encountered in the Great Apennine
itunnel, whereas the highest temperature observed in the Tanna tunnel was not more than

2-1o Cf. Report of the Construction of the Tanna Tunnel, Proc. of the World's Engineering
·--:ongr. Tokyo 1929
 -------=-- ~--

GEOLOGICAL PROHLE 99

S: 41 42 4J 44 45 50 50 Statlon ( km)

Fm. 2/23a. Geological profile of the Great Apennine tunnel

Fm. 2/23b. Geological profile of the Tanna tunnel (Japan)

28 °C, although temperatures up to 59 °C were estimated. This latter estimation took into
account the porous volcanic tuff overburden material, the expectable gas and hot spring
inbreaks and postvolcanic action (solfatara, fumarola). In the Apennine tunnel the discrep-
ancies are due to intensive methane infiltration through the Eocene clay shale (2·3- 3·0 km
fro m the northern adit).
This raised the temperature considerably, whereas the great quantity of water rushing
into the Tanna tunnel exerted a considerable and unexpected cooling effect. True rock pres-
sures were encountered in both tunnels. In the Apennine tunnel this was experienced in the
clay shale mentioned before, where the Belgian method and the trapezoidal timber-frame
supported drifts had to be abandoned in favour of the Austrian method and of
circular drifts lined with hardwood blocks (cf. Fig. 2/23a). In the Tanna tunnel
genuine mountain ·pressures approaching 20 kg/cm2 were measured in the solfatara
clay which is a product of andesite and andesite tuff weathered by hydrothermal
action. This clay was encountered i fault zones and ur.d;r the action of water flowed like
a fluid mass into the tunnel. In the re~ion which displayed signs of violent volcanic action
the rock was densely interwoven with faults, whose width exceeded 15 m in many places.

7•
JOO GENERAL DESIGN

Driving through two similarly wide faults (indicated in the figure) took 8-9 months to accom-
plish. Attempts were made to drain the inflowing large quantities of water with the help of
an auxiliary drift system (cf. Section 72.333). In the loose crushed rock section extremely
great difficulties had to be overcome before the tunnel could be completed at last by the use
of shield driving, compressed-air dewatering and cement grouting. The Tanna Basin itself,
as shown on the geological profile, is the result of a depression caused by tectonic action
which was accompanied by the development of a close succession or'a series of faults. Along
many of the closely spaced faults water burst into the tunnel and the inrush of vast quanti-
ties of water (Q max= 3-4 m 3/sec) caused several fatal mass accidents. Considerable inflow was
encountered during the construction of the Apennine tunnel too, (Qmax = 0·35 m 3/sec).
As revealed by the comparison of the geological profiles of the two tunnels, the penetratio
of young sedimentary and igneous rocks or rock fragments may involve serious difficultie
even in the case of smaller overburden depths. The preliminary exploration of geologica
conditions may prove impossible for long tunnels, since exploratory boreholes and shafts
cannot be spaced closely enough to obtain a complete picture of the possible variations.
Seven boreholes were drilled for the Apennine tunnel, and a close network of boreholes
was established in the Tanna Basin, but these failed to furnish all the information required
for the reliable prediction of geological conditions. Boreholes are generally unsuitable for
this purpose, and exploratory shafts are too expensive and time consuming so that they are
rarely resorted to. (For example, the driving of a 100 m [330 ft] deep shaft takes a year!)

755
lri,mc ,meslont
f;u//1d t;ym
-J¾. J768m - J956m

Fm. 2/24. Geological profile of the Karawanken tunnel

The geological profile of another tunnel known also for the difficulties encountered
during their construction is shown in Fig. 2/24. The 'carbon' section of the almost 8 km long
Karawanken tunnel on the Villach-Trieste line is famous for the extraordinarily high 'genuine'
mountain pressures experienced there. Although the overburden was not more than 650 m,
intense pressures pushed the heavy ashlar stone masonry linings almost 1 m towards the
axis and the cross-section of timber headings was reduced by lateral squeezing, bottom up-
heaval and roof subsidence to such an extent that it necessitated constant reconstruction and
reinforcement. Conditions could not be improved by the installation of double 300 mm steel
section frames applied at the bottom heading, these being deformed as badly as timber frames.
Conditions became worse when the top heading was driven and the 4 m thick natural deck
left between the two headings suffered a heave at the bottom of the crown heading, while
it suffered a downward movement at the roof of the bottom heading as a consequence of the
enormous side pressure resulting in a complete distortion of the cross-sections. These high
pressures occurred principally in sections driven through coal- shale layers interwoven by
clayey shale and slate layers sometimes in combination with quartz sandstone and quartz
conglomerates. The shales showed a tendency to fracture, while the quartz rocks were as
hard as glass. The difficulties will be fully appreciated if it is realized that timber supports
amounting to 25 % of the cross-section volume were required for constructing one linear
metre of the tunnel. The difficulties were obviously due to the plastic deformation of the
low-strength coal. and clay shale layers which were capable of considerable plastic de.f ormation
 GEOLOGICAL PROFILE 101

and in which the development of the stress-relief took place very slowly and was accompanied
by large deformation (cf. genuine mountain pressure).
The geological profile in Fig. 2/25 is that of the Semmering tunnel.~- 11 This tunnel is one
of the typical examples of pass tunnels, where the overburden is minimal and its depth varies
rrom 40 to 100 m only. This naturally results from the attempts to minimize the length of
the tunnel. This advantage is, however, offset by the weathering effect of meteoric waters,
frost, and other atmospheric and biological actions which is more pronounced in the thin

New
Highway
Dark i1nd grey q113rfz,tes w!lh
m ~ f)Jltles and c/ay-schiJ/s
7000
ll1mlJ Coloured fiUites H111h
C~!l'!I Jehists

950
,\orlhern portal

i{onronla/ Downgrade
Station ?i}?..

!Jwc level
'? ~ '<,
~~
~~
I:;~
'-c~

FIG. 2/25. Geological profile o( the Semmering tunnel

overburden layers. Consequently, a greater loss in strength was produced and water found
more ready access to the tunnel. Moreover, in the location selected, the rocks were in a highly
folded, faulted and fractured condition interwoven with numerous clayey layers and at the
interfaces, layer boundaries and faults the situation was aggravated by the infiltration of
water. At a distance of a few hundred metres towards the mountainside sound strong
li mestone would probably have been encountered, at the price of the tunnel being longer.
The mountain range consists of Triassic and Jurassic sediments which were subject to violent
tectonic action in the vicinity of the tunnel. The rocks can be classified broadly into two
groups, the first of which (I) comprises grey quartzites and grey phyllitic clay shales. The
more or less coarsely grained quartzites are of a block-like structure, and were subject both
at the surface and at greater depths to considerable tectonic forces which is reflected by the
strong folding and fracturing. The residual pressures in these rocks were extremely high as
was indicated by the spalling off and disintegration of the individual layers into incoherent
shingle. This process was reinforced by seepage water. The greatest overburden depth is
100 m composed of dark phyllitic clay shales of not more than 10 m thickness and of 40 m
of pure quartz layers.
The second group (II) comprises brown, closely jointed dolomites of low strength, and
reddish brown phyJlitic clay shales, the latter alternating closely wi_th the dolomite. Almost

2 11
· HANKER, R.: Der neue Semmeringtunnel, Eisenbahntechnische Rundschau 1952 H. !2.
102 GENERAL DESIGN

white mylonitic clay shales occurred in the vicinity of the southern portal. This other kind
of overburden has likewise a maximum thickness of 100 m, within which the total thickness
of the brown dolomite amounts at most to 20 m, the phyllitic clay shale to 15 m, the light
coloured dolomite to 25 m, and the quartzite to 20 m. Usually, however, these rock types are
thoroughly mixed.

An/w,rp [sc;ut
11m

El pi.,sl,t c/;y Ea fine S3nt/ f2ZI P'-'I

ll!ill /lNJ! , ;ml, grn t / mJ to,rs, und , gr;r,/ ~ J'ro--y ~rt1,3n,1
all 113rk coarse ~ nd 111D ~pf1/I

Ntw ~rk ((/uun's - M1dtown )

Rollfrdam

l!verpooi 61rktnhead

llam/Jur;
f,nes.nd

FIG. 2/ 26. Geological profiles

of subaqueous
tunnels
C/;iy

Plastic flow of the rocks could be observed along almost the entire length of the tunnel
varying between a higher and a lower rate, depending on whether the clay shales, or the
dolomites and quartzites were predominant. From these flow phenomena it could be
concluded that the rock was either originally in a limit state of plastic equilibrium, or was
brought into such a condition as a consequence of stress concentration produced by the excava-
tion of the artificial hole. Whichever the cause the pressures experienced were genuine mountain
 GEOLOGICAL PllOFILE 103

i-n,ssures with intensities ranging from 10 to 25 km/cm2 corresponding to overburden depths

beiv.een 40 and 100 m.
Owing to the limited construction time . specified for the completion of the tunnel (27
months) it was impossible to wait for the progressive development of genuine mountain
pressure and for the building up of the natural protective zone.
A circumstance of special interest was the considerable heave of the bottom resulting from
tile genuine mountain pressure, the magnitude of which attained as much as 2 ·m in some
places. The maintenance of the bottom drift caused serious difficulties during the entire con-
i.truction period and the tunnel was lined for this reason all over the high pressure zones with
rings made of l ·5 m long beech beams (cf. Great Apennine tunnel and Section 62.114).
The extent of bottom heave increased with the size of the section excavated (cf. Section 3'4
and Fig. 3·43).
The above observations were made during the reconstruction of the tunnel in the years
1950 to 1952, and during the construction of the new single track tunnel ·92 m nearer the
mountainside. Geological conditions here hardly differed from those of the. original double
track tunnel constructed between 1838 and 1852.
Even more serious difficulties may be encountered when tunnelling through
ecent geological deposits, particularly under river beds.
Geological profiles of four famous subaqueous tunnels are shown in Fig. 2/26. As mentiont
ed previo\lily in connection with the location survey, in this case the considerations of los-
height and of watertightness assume decisive importance since the tunnel section must usually
be designed to withstand the full geostatic and hydrostatic pressures. The extent of permeabil-
ity of the layers overlying the extrados of the tunnel is, thus, one of the governing factors
especially where shield driving combined with compressed-air dewatering must be resorted
to. Favourable conditions were ensured in this respect by the plastic clay surrounding the
Antwerp tunnel, and the red Triassic sandstone through which the Mersey tunnel was driven .
The Oligocene Kiscell clay in which .the Kaposztasmegyer tunnel was built for the Budapest
waterworks was also sufficiently impervious. The fine loose sand over the Elbe tunnel in
Hamburg, the silty peat soil in Rotterdam and the gravelly moraine debris penetrating into
the rock above the Queen's Midtown tunnel proved to be highly inefficient. For this reason,
the method of sinking into place from above ·was preferred to shield driving in Rotterdam,
while a clay blanket seal was spread over the river bed at the Queen's Midtown tunnel. This,
however, was insufficient to prevent the inrush of water from above which occurred once
during construction. This also happened, but with a much greater intensity, in the Hamburg
tunnel where the internal overpressure of the compressed air used for dewatering the excava-
tion, exceeded the hydrostatic head and burst through the air-pervious and water-pervious
overburden. Afte, the escape of air, water and ~and from the•river bed filled the shield cham-
ber. A single wa·ter-bearing passage in the otherwise impermeable cover is sufficient to cause
a similar disaster. Where underground conditions are difficult or dubious the procedure
followed at the Haarlem River Syphon on the Croton aqueduct may serve as a model (Fig.
2/27). As will be perceived, an experimental drift was driven at a depth of 36 metres and as
pressure and seepage conditions were found to deteriorate exploratory boreholes were driven
from it. All these signs indicated the proximity of a fault filled with weathered and crushed
rock. The exploratory drift was then abandoned and the syphon tunnel was constructed at
90 m depth, where sound rock layers were _e ncountered. Complete information on the geology
could thus be obtained here too, by means of an exploratory drift only. For this reason the
advisability of pilot and exploratory headings driven as far as possible in advance of the en-
largement to full section cannot be overemphasized (cf. Section 62.11).
Geological profiles for several other tunnels both in Hungary and a broad are shown in
Figs 2/28 to 2/30 (Monorfalva, Pereces, and Varhegy).
The geological profile of the bigger Monorfalva tunnel on the Szeretfalva- Deda railway
line is reproduced in Fig. 2/28. The tunnel, as will be seen, is located in younger Tertiary
104 GENER AL DESIGN

,4bandoned on account
of soft rock

1 JO·O - :---------+----------,---~--:;----;;------------~
LeY1!/ marks Assumed surfoce ofsound rock
0

·- 30·0

·-720-0 - ~ - - - - - - - - - - - - - - - - - - - - - - - - - - -

FIG. 2/27. Haarlem River Syphon (New Croton aqueduct)

m
,;/Jove Sea level
2
500 cp tF,1 • 5·0 kglcm
cp • 24 .
6"d
550
NW SE Fte. 2/30.
500
450
-····~·Bkg/tm' .. _
400 --- rt. • Jr -ffioierie i:!ay layers inter1r11xed with_11_nd ···-·
(IV -···· - ··-· · - · - ··- •••••••••

J50
!lOOm

FIG. 2/28. Geological profile of the railway tunnel at Monorfalva (Transylva nia)

2/0m
li'econslrvtted ,n 1g4, s,,t 1n!r1JJ1011 Jm1141red sec/fen R,rons/rvcle,t
(dr41ned 1111951-54) (rep,,rtd 1111955) ,n 19n
Q M,rl w!III t!,y B loose s,ntl ,nrf SJndslant
~ Marl w,lh lo,m E::J Co,Jt.,y,rs

I• cca Z450m
FIG. 2/29. Geological section of a small-gau11:e service tunnel at Pereces (Hungary}
.I
 GEOLOGICAL PROFILE 105

(Miocene) clay layers. As a consequence of the volcanic action

which was responsible for the surrounding ranges, andesite
rubble was deposited over these layers. The clay layers in the
vicinity of the portals were saturated to a greater extent and,
consequently, higher pressures developed there. Proceeding to-
wards the interior of the mountain the clay displayed an increas- ·
ing shear strength while towards the SE a transition to shales
could be observed. Highest pressures developed around the
northwestern portal. Under atmospheric influence the clay
was found to expand to a high degree which was revealed by
the continuous inward squeeze and difficult maintenance of
the cross-section of a drift driven a little too far ahead.
The Pereces tunnel shown in Fig. 2/29 was also built in
Tertiary and even younger layers in the direction of the strike
in almost horizontally bedded layers. Considerable difficulties

Danube L.WL.
l<risztinaviros Vizivaros
Fm. 2/30. Geological section of the vehicular tunnel under the Castle Hill in Budapest

were encountered not only during its construction but also later during its maintenance (cf.
Section 72.334).
The geological profile of the tunnel under the Castle Hill (Varhegy), Budapest, is shown
in Fig. 2/30.

In the absence of data provided by an accurate geological profile not even plan-
ning work can be started on proper lines, because of the uncertainty in selecting
the location, the method and dimensions of drainage, and lining of the tunnel ,
and there is a great deal of uncertainty as to the most economic construction
method. The geological profile is also indispensable for the economical and co::rect
performance of maintenance and of occasional repair operations.
Naturally, in the design and planning stage this geological profile cannot be
limited to a single vertical section taken along a given line, but in order to find
optimum conditions it should be extended to the surroundings of the probable
alignment. Thus, for instance, in the case of underground railway tunnels it is
usually specified to locate the exploratory boreholes alternately to the right and
to the left of the contemplated tunnel axis, at distances from 50 to 100 m, rather
than in the axis itself. Where the· exploratory data display disturbed geological
106 GENERAL DESIGN

conditions, it is advisable to complete it by geological cross-sections. Unfortu-

nately, owing to the gaps and deficiencies of geological exploration, the accurate
geological pattern will be disclose_d, especially in geologically disturbed regions,
in the course of the actual tunnelling work only (cf. the Tanna tunnel).

2.2. FACTORS INFLUENCING THE LOCATION OF THE TUNNEL

2.21. SELECTION OF THE LINE

The selection of the general alignment is governed primarily by the traffic,

or transportation interests necessitating the construction of the tunnel and is,
thus, a function of the purpose which the road or railway line is intended to serve.
The exact location, however, is controlled by the particular geological and hydro-
logical conditions prevailing in the surroundings. The tracing of tns: route proper
should be subject to the same principles as on the surface, the train or other
vehicle to be carr.ied through the tunnel being the same. The only difference is,
perhaps, that in tunnels the adherence to a straight line is even more desirable.
The reasons for this are not only those of shortening, economy in traction, or
better visibility, but also the simplification of construction and surveying, as well
as the setting out, and finally better ventilation. The tunnel axis itself is frequently
straight but the adit sections are at a tangent owing to the necessity to connect
to the open line. For instance at the southern adit to the Lotschberg tunnel the
line climbs for a considerable distance on the steep rock hang, bordering the
narrow valley of the River Rhone and penetrates abruptly with a 90° curve into
the mountain. A straight tunnel axis was originally plan ned but, as to be seen
in Fig. 2/31, this had to be abandoned during construction when the Kander
Creek which flowed about 100 m (330 ft) above the tunnel found access into the
heading through a gravel seam and filled the whole section with sandy silt for
a length of about 1500 m. It was therefore founq necessary to by-pass this hazar-
dous section in the manner indicated.

FIG. 2/ 31. Alignment alteration scheme of Lotschberg tunnel

 LOCATION 107

Similar solutions are adopted more frequently for subaqueous vehicular tunnels
built in the city area of large ports. Owing to the great depths the access ramps
would have their exits at the surface far from the waterside where they would
disturb the most valuable urban areas of the city and where they would fail to
connect the related embankment sections. For this reason the tunnel is constructed
in the shortest possible straight line under the river, usually with a rectangular
crossing, while the ramps are curved and have their exits to the surface paraUel
to the river banks.
Curves may sometimes be inevitable in long tunnels, too, in order to avoid
adverse geological formations, water-bearing strata, fault zones, etc. Minor offspur
and hang tunnels may also be curved, in an attempt to follow terrain c:mtours.
Spiral tunnels built to permit the necessary gradient are, naturally, always
curved.
According to Soviet specifications the minimum radius of curvature for curved
tunnels is 250 m (830 ft) on main line railway tunnels, and 200 m (660 ft) on
branch lines.
In the choice of the tunnel axis the careful selection of the adit and exit locations
is important. It has occurred more than once that an incorrectly select~d portal
has broken down shortly after the commencement of excavation and has had to
be abandoned for another site. Conditions are especially unfavourable where
weathered, loose, fractured layers slope towards the portal. If these are pierced
by an•approach cutting before a sufficiently resistant tunnel portal structure is
built, the entire slope may be mobilized and sliding can no longer be arrested.
The correct selection of the tunnel adit proper, i.e. the length of the approach
cutting is also of great significance. For economic reasons its depth should not
exceed 20-25 m (65-83 ft), yet at the same time the adit section should not be
sited in a sliding layer. According to STINI highly resistant structures must be
designed in uncertain layers and the coping of the portal should reach a few
decimetres above the fractured layers covering the slope (Fig. 2/32).
Tunnels should, in general, not be constructed in undermined regions (e.g.
utility and underground railway tunnels in mining districts!), because of the
unpredictable magnitude of the settlements and forces to which the tunnel may
be subjected.
The alignment of urban underground railway and subaqueous vehicular tunnels
depends on traffic and town-planning considerations. The necessity for co-ordi-
nating the location of the stations and their entrances with the street network and
with the town picture, and thus ~djusting the line to town-planning will make the
use of curved sections inevitable. The problems arising from the selection of an
appropriate tunnel line are especially difficult in the case of sub-pavement under-
ground railways, which are bound to follow the street network, so that ai;ca-
sionally very sharp bends must ~ resorted to. This limitation is more tolerable
where the railway is located at a greater depth, where the curve resistances can
be reduced and higher speeds can be assured. This is well exemplified by Fig. 2/33,
showing details of the old Budapest underground and a connecting section of the
 \

108 GENERAL DESIGN

new line. The sharp bends necessitated by adherence to existing streets a,·e clearly
shown.
Curvature and speed data for some of the more famous underground railways
are compiled in Table 2/V.

- - - -- --~...~/ Portal
F1G. 2/3-2. Forces acting upon a tunnel portal (creep-pressure!)

<;;
t7-,.,,-.....

FIG. 2/33. Comparison of admitted curvatures of the old and new underground railway line~
in Budapest
 TABLE 2/Y
Comparatiue Data of Underground Railways
-1- -- - - -
r- Max. number
I

I
~ !-
C: -l::0 of trains/ hour ·g
,;;- ~ ~
·;; :< .,.6 °'E .,><~
_
-g "C. ,;;; '0 "'
'0 .;
I E ... ] ·3~ EE "'
u
.g
.... C: ....0 ":;! C:
....0 "' u "'"'
C. C
....
:;I :;I
;:
"'c,.,
Ciu,
C:
0
-~ _g
-a~
:;!
:;I ...
0
,,_.!?
~
GJ O::::
.0 ,., 13
...u :;I
0
-s .c II C~
[ O
..
= "'
:.;; u u
.. .c .
=
C.

-;;
.l1..
=
0
.c
:;;
~ :2 .. "
"' 0
0

~ ~
·.::: ~
:::,.':!
C. 0
~~
~ 8
!
~ .""
~
.E
!g
.0 u
= 6 ...
u =
E
= °'E
ii-"
"' .... 0 ~
8 0 -~ t u ... E .£: .!:: '--" E 3
o

~
C.
ov
p.. z c.:.:-
:;I
z
:;I
:.:, ...:i~
"'u
&; C. < ~"' < I >
<
0
u0 u
> :;;:. 0 <>;
~ "
I
New York
8000 1574 4300 380·0 6'2 40 32 3-0 1 I 29·6 n 1 I 38·8 100
0·1s I
1- -- ! - - -1- - -1- - -1- - - -
Paris 4776 1032 3000 169·0 16·4 36 2·21 45·5 0·51 41 ·6 39(75) 40
- -- - - -- -- -
60·8 241 (100)
London
- - - -- - -1
9785
3357
702
267
2000
732
357·0
75·9
4·9
- - - - - - - --
I 7'7
40
40
38
18
1·91
J·08
34·8
24·6
1·3
1- - - -1- - 1-
0·77 47·2 76
s
(l
),
Berlin new 201 --1

Philadelphia
Boston
2600
1500
J 82 621 40·2
58·5
1~2._ ~ ~ ~ 26'9
38 25 9·5 30·5
0·7 I
0·93 57·0
40
0
z

· ------ - - -- --1---1
Chicago 5000 J 53 504 28·6 5·6 40 25 3·2 33·4 0·83 81 ·4 103
- - - - -- - - · --
Hamburg 1700 J 56 426 41 ·5 12·7 40 24 1·48 29·6 1·22 47·0 68(180)
-------- --
Stockholm 753 80 219 15·5 21 ·4 40 24 3·7.5 30·4 0·5 70·5 201
------
G lasgow 1090 37 100 6·6 19·6 20 4·18 22'4 0·69 38·4 201
------- - -
Rome l 680 6·0 0·99 201
---- -- - - - 1- - - 1 - - -1- - -1- -- 1- -- - -
I I
Oslo
Budapest
429
1800
5500
20
,--1
I 1000
60
80
1 - - - 1-
8·2
59·5
21·6
-----1
40
42 - - 1 - - 1 - - -
30 35·0 J · J7
1- -1 - -1 - -
40 30 40·0 I ·45
360
300
30
33
Moscow* (35)
(6000) ( !000) (3500) ( 121 ·O)

ll,111,~k; • Data given for 1955. Data in bracke t< refer to 1964. -
0
\0
110 GENERAL DESIGN

The line of utility tunnels in urban areas is adjusted to the needs of public
utilities and mainly to the existing road network and housing arrangements.
Water supply tunnels connect the site of the source with the centre of consumpti on
and their alignment is chosen accordingly, with due considerat ion for construc-
tion and safety requirements. These are largely the same as those g<;>Verning the
selection of the line for communic ation tunnels. For increased watertightness
regions showing a tendency to movement and fault zones should be avoided with
even greater care. Water conveying or navigation tun'nels should not be built
in regions exposed to earthquake s.

2.22. SELECTIO N OF THE LONGITU DINAL ELEVATIO N AND GRADIEN TS

The choice of the vertical location and profile of railway tunnels is in gen-
eral governed by the destination of the traffic route, i.e. by the aspects of estab-
lishing and maintaining traffic operations . The more specific choice between
potential alternatives, however, will be governed by traction a nd constructi on
requirements.
The elevation, i.e. altitude or depth at which the tunnel is located is equally
significant for both constructi on and traction, the layer in which the tunnel will
be driven and its position relative to the ground water table, and also grade and
lost height conditions being dependent on this decision. While for tunnels in
mountains the aspects of traction and operation, and sometimes the position of
ground water (cf. Fig. 2/ 14) may be decisive factors in the selection of elevation
- constructi on methods being hardly affected by slight chan[es in altitude in the
usually uniform, solid rock - the depth of underwate r tunnels is controlled pri-
··marily by geological conditions. It is, naturally, preferable to locate ·mountain
tunnels in a rock formation of adequate strength and thickness, if possible above
the ground water table (cf. Fig. 2/ 14). This requireme nt can, however, usually
be met in an economical manner even if traction and operation remain the primary
considerations.. No unsurmoun table difficulties are likely to arise even where these
latter considerations call for a more expensive and more involved method of
constructi on entailing the necessity of dewatering and of providing stronger sup-
ports for the loose soil masses.
For subaqueou s tunnels, on the other hand, the vertical position is determined
primarily by constructio nal aspects, i.e. by geology; grade, traction and operation
must be subordina ted and adjusted accordingly. Accurate informatio n must
be obtained first of all on the bed contours on the elevation and succession of soil
layers and particularly on the depth (t) and thickness (A t) of the impervious
layeJ (cover).
The tunnel must be located in the impervious layer under an adequately thick
cover, whose thickness should be sufficient to prevent the inrush of water and
eventually the escape of compressed air so ensuring a wide margin of safety in
the constructio n of the tunnel (Fig. 2/34).
 VERTICAL LOCATION 111

Fro. 21 34. Vertical alignment

of subaqueous
tunnels

Considerable restrictions imposed by the prevailing geology at the site can be

avoided by constructing the tunnel from above, i.e. by sinking prefabricated tunnel
sections into place. This· is one of the reasons for. the growing popularit y of this
method.
The typical profile of such tunnels resembles that of a trough, as the central
horizonta l section is flanked by inclined ones with the steepest permissible grade.
The vertical position of urban undergro und railway tunnels is also governed
by geology and aspects of construction, but aside from traction and operation ,
considerations of accessibility, the avoidance of interference with municipal
utilities, as well as the safety of surface structures assume equal importan ce.
Elevated stations may be responsible for special profile arrangem ents (see Fig.2/37).
The vertical position of canal and navigation tunnels is determined mainly by
the elevation of the waterways to be connected, or by the elevation of the n_a tural
water supply to them, although geological conditions are also of significance both
for construction and watertightness. The slope in the tunnel itself must be limited
to a very small value.
Utility tunnels can be located most readily, in general, to comply with geological
and constructional requirements. Operatin g restrictions will assume significance
in the case of water supply tunnels, although the required watertightness calls
for a careful considera tion of geology as well.
As far as traction is concerned, the ruling grade permissible in tunnels is appre-
ciably smaller than in the open air, owing to the reduced adhesion and increased
air resistance in tunnels.
The first reduction factor is the decre~se in the tractive force due to the reduced
adhesion-coefficient in tunnels.
Under favourable conditions, in summer, on a hillside exposed to the sun the
adhesion-coefficient for steam traction varies from f = 0· 17-0· 18, and may drop
under adverse conditions, e.g. in a deep cut, or in a moist atmosphere at altitudes
around 1000 m above sea level, to f = 0· 14-0· 15. The atmosphere in tunnels is
usually saturated with moisture, owing to the vapour which precipitates from
warm air in summer, and to ground water infiltration. Moisture thus precipitated
on the rails or on the pavemen t causes a significant reduction of the adhesion-
coefficient between wheels and rails. The degree t>f adhesion is affected by several
112 GENERAL DESIGN

other factors as well, such as the elevation of the tunnel above sea level, the situ-
ation of the tunnel portals, wind and sunshine conditions, the cross-sectional
dimensions of the tunnel, ventilation and draught, etc.
This reduced / 1 value was found experimentally to be 0· 11-0· 12 for steam
traction.
Consequently the traction force of the locomotive will drop from the open-air
value
V = lOOOfQ0
to

where Q 0
= the adhesive weight of the locomotive.
The ruling grade is naturally also affected thereby, since in the case of constant
motion the tractive force must equal the sum of resistances, or, on an open line

(2.21)

where Q = the weight of the train

µ = its specific resistance
em = the ruling grade per mile.

The weight Q of the train to be handled remains the same in the tunnel, neither
does the value ofµ change, so that the ruling grade e,,, must decrease to e1 in order
that the reduced tractive force V 1 shonld not lead to deceleration, since

(2.22)

by division of one equation by the other we obtain

µ+ e,,,
µ + e1
whence
(2.23)

Assuming the ·conventional value of 3-3·5 kg/ton for µ (slow, heavy freight
trains with v = 30 to 55 km/h are decisive in this respect), the value of e 1 can be
computed in terms of e,,,. Various e1 values pertaining to corresponding e,,, values
are presented in the following table which also indicates the reduction involved.
As will be perceived the reduction is appreciable and depends, naturally, on the
absolute value of the ruling grade as well.
 GRADES 113

Conditions are different to a certain R uling grade

extent in the case of electric traction Open line I Tunnel
Reduction
(°I,,)
where the value of adhesion, i.e. the coef- em ('/..) I_ _e,_<•_J ,_,)_~--- -
ficient of friction is in itself considera-
bly higher than in the case of steam 10 6-5 3-5
traction, and where the rate of reduc- 15 10·2 4·8
20 14·0 6·0
tion is also different. In the case of 25 17"5 7·5
electric traction e 1 may be taken at 30 21 ·2 8·8
20-25°/ 00, whereas it cannot exceed
15°/ 00 in the case of steam traction.
Besides the reduction in the traction
force the increase in air resistance is also a factor necessitating a reduction in
the ruling grade. Denoting this share of the reduction by Lie, we obtain the
previous equation in the following completed form:

(2.24)

The magnitude of air resistance is known to depend on the relative velocities

of wind and train (i.e. it is greater upwind than downwind) as well as on the rela-
tive cross-section areas of tunnel and train. The magnitude of the resistance and
the factors involved have been obtained experimentally by measurements in
wi nd-tunnels and existing tunnels. Observations in the Simplon tunnel have
already pointed to the fact that resistance is especially large in single track tunnels
which are comparatively narrow. The train proceeds in a relatively restricted
space and is preceded by a compression wave and followed by a depression wave.
The first experimental measurement on the increase of air resistance was carried
out in the Simplon tunnel which consists of two single track tunnels spaced at
12 m and provided with artificial ventilation . Air resistance was measured first
on the open line, then in the tunnel in the direction of 'the air stream and sub-
sequently against the forced draught. The · train proceeded at different speeds.
The results of these measurements are compiled in the following table:

Train velocity, v (km/h) 50 60 70

Air resistance, open line, .de,

(kg/ton) 3-3 4·1 5·0

A ir resistance in tunnel in the

direction of draught, Lie, (kg/ton) 5·0 6·4 8·0
..
A ir resistance in tunnel against the
draught, Lie, (kg/ ton) 7·5 9·5 J2·0
I
Sz~;hy : T;i: Art of T unn:llins
114 GENERAL DESIGN

As is shown the air resistance Lie may assume values considerably in excess
of the specific train-resistance µ, and may amount to a considerable portion of
the total traction resistance.
The most recent resistance measurements have been performed in the tunnels
of the London Underground Railway. 2 ·12 The parameters included the velocity
of the train, the ratio between the tunnel cross-section, the front area of the train,
and the internal surface roughness. The results are represented graphically in
Fig. 2/35. A5 can be seen from Fig. 2/35a, the measured resistance increased
in a practically linear ratio with the speed but the gradient of the straight line
was a function of the tunnel diameter. Beyond a certain limit this influence ap-

BJ

0 ~ ,o..._~2 4~-J.~'2-- 40-'------ 4... _'8_ _ _55

..__ ___,
54
Speed of train (lm/nour)

nm---- ~----- --------~

18001-------1-~ ---\-- f - - - + -- - +- - + - - - - t ----i

Smooth mnersur(3ce}·- Resistance
~
----- Ribbed surface in tunnel
~ 1850

~
~
~ g(KJ l--------,-l-¥------,L-'>k'-~-+-
.~ Tunnel '-.:. ""+---,--
~ diameter ~-
:i"-:::.b::::=:::'.+'.=--==tcc=~===:!~
450 using existin!J ro/2ffi1/l:;;ng Fm. 2/35. Results of air
.stock ~ 'I'). resistance tests
~
~
~ ~ ~
~ ~ ~
made
O L____ _IJ.,'8~:,_;z.;i,.7,.::i.....:LJ6
~ - -4+.5=----:!-5-4~- -::6J:;--- 7·2 in the London
free area between vel11cle ,1nd tunnel cross-section (LJFmz) Undergro und

2 · 12 T U RNER, J.: Newest Trends in the Design of Underground Ra ilways, J ourn. Inst. Civ.
Eng. 1959 I
 G RADES 115

peared to d~crease appreciably. In Fi_g. 2/ 35b more exact information is given on

the limit and extent of the influence exerted by the ratio of cross-section areas
1free area between train and tunnel) and by the roughness of the internal surface
on air resistance. The differences between the cross-section area of the tunnel
and the front area of the train (i.e. the free area A F) are indicated here on the
abscissa and the magnitude of the resistance on the ordinate axis. Values pertaining
to different speeds are connected by corresponding curves. As indicated by the
trend of these curves, the effect of the free cross-section area becomes negligible
as soon as the value of AF reaches 4-5 m2, i.e. about 40% of the total cross-
section area. The influence of the ribbed tube lining becomes entirely negligible
at the same limit, but it was found to be altogether appreciably smaller than that
of the previous factors.
Curve resistances in tunnels are the same as those on the open line.
The reduction of grade due to the increased resistance should be started at
a distance corresponding to one half of the train length before the adit, rather_than
at the portal itself.
Grade conditions within the tunnel are influenced to a certain extent also by
considerations of drainage which must be ensured in every tunnel. The minimum
slope required for this reason is 20/o0 , but 3%0 is preferable. Tunnels are usually
made to slope in two directions outward from a central apex in the mountain.
If the tunnel is not at the highest elevation of the line and is on a gradient, or is
not too long, a unidirectional slope may be used. Outward slopes towards the
portals have the advantagl? of ready drainage away from the driving faces at both
ends. In the case of a unidirectional slope, on the other hand, the water must
be removed artificially from the heading started from the upper end. The water
from subaqueous tunnels should be drained to collecting sumps constructed under
the adit or ventilation shafts at the banks (cf. Fig. 2/26), whence it must be removed
by continuous pumping. Drainage by gravity can be effected only in tunnels
elevated above the ground-surface.
In designing the longitudinal profile of urban underground railways special
problems have to be considered. First of all the absolute value of the ruling grade
may be greater than in surface tunnels, since underground tunnels are carefully
sealed, drained and ventilated. Constant movement of air at relatively high veloc-
ity is induced further by the rapid sequence of trains in the comparatively narrow
cross-sections, so that no reduction in adhesion due to vapour precipitation is
likely to occur. This is accompanied by the initially higher adhesion value of .
electric traction. Narrow cross-sections, on the other hand, involve the danger
of greatly increased air resistance (see above) with co:r;i.sequent adverse effects
on the ruling grade.
The stations of deep underground railways are usually higher than the stand-
ard line. The main purpose of the elevation is to bring the platform level
closer to the surface and to reduce the time of access to .and from it, i.e. to
minimize the loss in height. This, however, is insignificant where escalators or
lifts are used.
116 GENERAL DESIGN

An additional and more significant advantage of this arrangement is the possi-

bility of the insertion of a steep braking slope before and an even steeper starting
slope after the station level (Fig. 2/36). The twofold benefits so obtained are
a reduction of time loss caused at the stop and at the start by more rapid decel-
eration, and acceleration, and the considerable saving in braking and starting

SI.at/on -tJ.
SzytioP

1~ ,
L j .,__~2=L_ __,~ -j___w
.,
; • 1,5V
2L Braki11g slope
Li, , ITllr,
<::::
~
-<:: Acee/era
t:::, tmg slopt
J2 -L1, • 117" CQI/St. ~ _§:i
L • .SL8tion length
v • Trai11 length
FIG. 2/36. Elevated location scheme of underground railway stations

energy consumption (permitting the use of smaller motors). The elevation of sta-
tions may be accomplished by a hump, or a saw-tooth type profile arrangement
(Fig. 2/37). Th~ saw-tooth type may be used advantageously where the stations
are closely spaced, or where the line descends from the surface. Elsewhere the
hump type is commonly used.

Station Station

C>!---1~-~_________). .-;
Accelerating
.slope Hump profile

St,at,on St.ation

Fm. 2/37. Accelerating and braking slope arrangements before and after stations
 CROSS SJ!CTl'ON '117

The benefits of a raised station arrangement can, however, be obtained only

where they are all elevated approximately by the same measure, and where the
slope of the ground surface and, thus, that of the line keeping spac~ with it is
insignificant'. With usual train speeds (55 to 65 km/h) the stations are usually
elevated by m = 3·0-4·0 metres, affording 3·0-3·3 % starting and l ·6-2·0 %
braking gradients in starting slopes 100-130 m long, and braking slopes 150-200 m
long. According to calculations carried out in connection with the design of the
Budapest Underground the saving in energy over a horizontal line is 20 % and
25 % when using 2 % and 3 % starting gradient over 145 m and 125 m lengths.
A braking gradient of 2·0 % results in a 20 % saving in braking energy. Owing
to differences in station elevations and upgrades over the intermediate line sections
usually occurring in practice, the saving in energy which can be obtained does not
exceed 10 % to 15 %- It should be noted that the elevated station arrangement is
likely to involve difficulties in drainage. The current optimum gradients for braking
and starting slopes and the lengths of the latter relative to the length of the sta-
tions and intermediate line sections are shown in Figs 2/36 and 2/37.
Slopes are disadvantageous if it is necessary to stop a train between stations,
where it must start on an upgrade after the line has been cleared. This case is
more likely to occur in the case of the saw-tooth type of arrangement.
In the case of steam traction counter-slopes should be avoided owing to the
ianger of stagnant smoke accumulation.

2.23. DETERMINATION OF THE CROSS-SECTION

The selection of the tunnel cross-section is influenced by:

1. The clearances and gabarit specified in view of the vehicles moving and
goods transported in the tunnel;
2. The type, strength, water content and pressures of soil;
3. The method of driving;
4. The material and strength of the tunnel lining as well as the internal loads
acting on it;
5. The necessity of accommodating a single, or a double track in the tunnel.

For certain traffic and transportation purposes, together with soil conditions
and construction methods, the most advantageous and commonly used tunnel
cross-sections already developed are shown in Figs 2/38 to 2/47.
(a) For railway tunnels the horseshoe cross-section is commonly used which
is more pointed for single track and flatter for double track tunnels. Sometimes
the internal face of the side walls is vertical. In the case of subaqueous railway
tunnels - driven by the shield method - ci.rcular cross-sections have also been
used on rare occasions (Fig. 2/38).

•
IJS GENERAL DESIGN

[lectr1c
traction

--,
t-rr-=~ \ . /
\ I/
}1
j)Jo ,--
C ---------+-

IIi
ij !
-- ...../) ·l,.__________
' _/.

Track/eve

FIG. 2/ 38. Typical cross-sections of railway tunnels

(b) Underground railway tunnels of the sub-pavement type usually have a

flat-arched cross-section. For deep lines two separate tubes of circular cross-
section are usually constructed (Fig. 2/ 39).
(c) Horseshoe cross-sections are adopted similarly for water-conveying tunnels
under gravitational flow conditions. However, the higher the internal pressure
in these tunnels the nearer the circular cross-section is approached (Fig. 2/40).
Sewage tunnels are usually circular or oval in cross-section with an inverted
flume.

•
 --- =- - -~-~-

CROSS SECTION 119

Sui>pavemenl a/1gnme111 Deep alignment

Ground surlacl:

-----w
bi'ound surf/Jee

F10. 2/ 39. Typical sections of subway tunnels

water conveymg (g1-av1-

Pressure tunnel !,aftonal) tunnel
/?emforced-
concrete
wall will!
eventual steel· I
sheet /mmg

~ i 1~- - = · ~- -i,;tll _ Steel posls

embedded
m concrete

Concrete blocks Sewer-tunnels

Inner rein~::e:~a -
coocre/e !tmng fl 2"
1 JJ _ Remforced-concrete

Vttrtfi'erf bnck
F 10. 2/40. Typical sections of water-conveying tunnels

(d) The typical cross-section for navigation tunnels consists of a trough-

shaped invert covered with a flat barrel vault (Fig. 2/41).
(e) Utility a nd pedestrian tunnels may be constructed with rectangular or
circular cross-sections, depending on whether they are located close to the surface
(subpavement), or at greater depth o r under a watercourse (Figs 2/42 and 2/43).
]20 GENERAL DESIGN

FIG. 2/41. Typical section of navigation tunnels

a; Pvblic vtt!tly tvnne!s

Precast-con-
crete block
!ming
Water-
proojrng
l?einlbrced- J60
concrete

b} Sewer tvnnel
Pret,Jsf reinforced-concrete
,...:.,tilif,ifof,Si~,f,,A,. Concrete r,11
Vttrifted brick !ming
Reinlbrced-
concrete
lining segments

Hardwood wedges
Concrete /illin{I

i
k
j 66
: _
I-
_ _
1-ffS I I
______::3-:..::.,55:___ _ _- - i) r - ' ~ - · ·
V fm
FIG. 2/42. Ty'pical sections of public utility tunnels
 CROSS SECTION
121
Hamburg

Burf;;pest
N:1rgaret Brio>

Jl~l'.i!!i~,,~P:,~.P}_'.'~~-'-'L).t_;;~~-% :~...~~~~~~- ~JF.t~tfl

Outside msu...f - · - =rr Precasf rem/brr,'d-
lat,on with ! Gumte tnis 1 concrete beams
I
,'l /

precast ,,
protective i 2-ffl._
Plates I
l c-00 Cast m slfu
1 mner remforced-
F1G. : -13. Typical sections of 1 concrete framewor,
passenger tunnels I1/i'~-,::----..,= ~~"s"i~ ~' ,~~

.Slcndarrl pro/ileof the 6errnan

61ov1 express way Re1rhsautobahnen
(Cat1111JtfJ/eJ

Ccsfle Hill Tunnel

(Buda)

Goat Island

FIG. 2/44. Typical sections of vehicular tunnels

122 GENERAL DESIGN

lirlf'/)l!(JI

. Rotterdam
Slope of
ck ii!
Steel sheet limng
70mm
J1"tiateif"Jir

•. .-
:.·.·
\".-:·
):."IJredged gullet
.. . . ,:-c.-/f'''.t
. ,· . .. .:--- ...· .·. ·'ioad iun;ei/""· :::-·

FIG. 2/45. Typical sections of subaqueous tunnels

(f) The widest variety of cross-sections has been adopted for road tunnels.
Flat horseshoe sections are common for highway tunnels, while more pointed
ones are used in city areas (Fig. 2/44; Budapest, Castle Hill). Subaqueous road
tunnels, (Fig. ·2/45) on the other band, are usually circular in cross-section -
when driven by the shield method (Antwerp, New York, London, etc.), but flat
rectangular cross-sections are usually adopted when they are composed of sunk
prefabricated reinforced-concrete caissons (Rotterdam, Amsterdam, Deas
Island, etc.).
 CROSS SECTIO1' 123

22.31. Tunnel Clearance

Th~ first step in the design of a tunnel is the determina tion of the cross-secti on
required by the communic ation, i.e . the necessary clearance.
The clearance of railway tunnels should be designed in accordanc e with the
specificatio ns of the Central European Railway Associatio n so that it should be
at least 30 to 40 cm larger than the clearance required for the open line. The pur-
pose of this precaution is to provide additional safety against constructi onal
inaccuracies and deformatio ns of the section due to rock pressure, rock displace-

Fm. 2/46. Required

clearance for l/ppl'r level
railway tunnels ofsleeper

111ent or water inrush. Considerat ions justifying a larger cross-section are improved
ventilation and smaller air resistance, the former being made imperative by the
necessity of removing the stack and exhaust gases of locomotive s and internal
combustio n engines. For this reason the tunnel cross-secti on is made about 50-60
cm larger than the clearance on the open line (Fig. 2/46).
Accommo dation must be provided not only for the vehicles using the tunnel but
also for the pipelines and cables of va rious services mainly for tunnel and track
maintenan ce (water pipelines, ven tilat ion ducts, compresse d air pipes, supply
cables for lighting, power, safety equipment , etc.), which must be arranged in the
most practicable manner. This aspect assumes particular significanc e for urban
undergrou nd railway tunnels.
Cross-sectional dimension s should be selected with potential electrificat ion in
mind a nd the overhead clearance should be increased accordingl y.
!24 GENERAL DESIGN

In road tunnels this dimension is usually even larger because of the increased
toxicity of exhaust fumes to the passengers of motor vehicles. This calls for im-
proved ventilation. The circumstance that the lateral movement of road vehicles
is greater than that of track-bound ones, should also be taken 'into consideration
(additional safety measure beyond clearance of about 1 m). It should be remem-
bered finally that road vehicles are still in a stage of development and their dimen-
sions, and consequently clearance requirements, are likely to increase in time.
Cross-sectional duµensions specified for tunnels on the German Autobahns are
shown in Fig. 2/47.
An important factor in designing the
cross-section of navigation tunnels is the
ratio of the wetted area (F) to the
submerged cross-sectional area (() of
the vessel since there is a rapid in-
crease in towing resistance below a ratio
m = F/f = 2·5 to 3·0. This demands
adequate depth in the first place. Owing
to the inevitable lateral movement of
towed vessels, a lateral play of at
FIG. 2/47. Clearance
least 1·5 m should be provided for on
prescription for motorway both sides.
tunnels The best possible utilization of space
can be achieved in rectangular cross-
sections, while circular ones are least favourable from this respect. It is interesting
to note that conditions are reversed in this respect in water-conveying tunnels.
In the case of water-conveying tunnels the size of the cross-section is governed
by their discharge capacity. Thus the magnitude of the cross-section is affected
by the permissible flow velocity as well. In free-surface conduits there is a decrease
in the discharge capacity as soon as the section becomes filled. For this reason
a clearance or about 0·5-1·0 mis usually provided between the crown and the
water level in the section, which is usually of the horseshoe type.
Water in pressure tunnels, on the other hand, is conveyed in the entire cross-
section, which in this case is .not horseshoe-shaped but circular to resist the high
internal pressures most effectively.

22.32. Influence of Geological Environment on Shape of Cross-section

A factor of considerable influence on the shape of the cross-section is the type

of geological environment in which the.tunnel is to be constructed. The magnitude
of external loads, i.e. of foe rock pressure, depends on the inherent strength .of
the rock material, the quality of which is responsible also for the ratio between
vertical and lateral pressur~s acting on the tunnel. In loose, soft and weak rock
 CROSS SECTION 125

materials large rock pressures and a relatively considerable lateral thrust may
be expected. The greater the relative magnitude of the lateral thrust the more
advantageous a circular cross-section will be found to be. In solid rocks showing
no tendency to weathering, tunnel sections excavated with an arched roof may
serve without resort to any lining. A thin layer of gunite may offer adequate
protection against weathering. In the choice of the ideal shape of the tunnel
section, in addition to structural considerations, earth pressure theory may serve
as a guide. Adopting Rankine's ratio for the constant ratio of vertical and hori-
zontal pressures an elliptical shape may be advised for which the ratio of the major
and minor axes can be expressed as (Fig. 2/48.)

2a : 2 b = l : tan2(45° - ¢/2).

For instance in the case of sandy soils (with ¢ = 30°),

2a :2b = 1 : 0·577 = 1·6,

:.e. the height of the section should

be 1·6 times its width which is ap-
proximately in conformity with the
requirements of rail',1/ay clearances.
In the case of saturated clays, on
the other hand (where ¢ = 0),
2 a : 2 b = unity and the advisable
section becomes circular. The pres
sure distribution is hydrostatic in
character with average pressures
Fm. 2/48. Theoretical shape of tunnel section
related to the centre being uniform according to Rankine's ratio
in every direction. In rocks this ra-
tio may be taken on the basis of
confined compression as approximately equal to

p ( 1 )
l - µ = ~-=-i-· ,
where µ = Poisson's ratio for the particular rock
m = 1/p.
As demonstrated by FENNER, 2· 13 an elliptical cross-section proportioned accord-
ing to this ratio has an advantage over circular ones, in that the considerably
large tensile stresses in the crown are reduced, but at the same time large compres-
sive stresses may be involved there. He investigates, therefore, the transitional

2 •13 FENNER, R . : Untersuchungen zur Erkenntnis des Gebirgsdruckes, Giiickauf 1938 681
 126 GENERAL DESJON

shape (cf. Sectio n 32.22), where the stresses develo ping in

the crown are of zero
magni tude. The shape of this section was found to be elliptic
al with the major
b
axis standi ng vettica lly and with an axis ratio - = - - 2
- . (It should be noted
a m-2
that this relatio nship is valid only for m ,I: 2.)
Recently the effect of the shape of the section on stresses
was investigated by
TERZAGHI and RICHARr·14 who studie d the norma l stresr.es in
the crown and spring -
P,
P,' HJ uHI -f(J -2.(J - ,XJ -4(1
6,
P,
a)
l 4-5
7 60
' 4.0
'I
t 10 J.5
/() t-.--+ ---+-- 1-+--h -+.,L-.j 6() ~
\ 1"

~ M
4 8 r----t-- -+--+-- -1-+ 5(1
\cl' \'"
4(J
I \ '\
J(J 11 ~\ 'O'
' ...<,:ii

r-- - 4; l "" I~ .:_,

--
I{ \ d
/'
?-- :.-::--
, ...._..... -
f(J -HJ
~ ,:----1 -:NJ -JO -40
X
P,
"i

1-tL-
FIG. 2/49. (a) Influen ce of section shape and of side pressur
e coeffici ent on roof-
(b) on side pressur es

ings of circula r and elliptical section s under -identical loads.

As is shown in
Fig. 2/49b, the stress in the crown changes from 0·25 Pz (comp
ression ) to - 0·50 p,
(tension) while the axis ratio of the ellipse change s from 1/2
to 2. This change is
accom panied by an increa se in the compressive stress from
l ·75 Pz to 4·15.p,
at the spring ings. The coefficient of lateral pressu re was
assum ed in this case as
)._ = ~ = 0·25. The effect of the coefficient of lateral pressures ()._) on the
Pz norma l
(tange ntial) stresses uh and uv arising at the crown and spring
ings, respectively,
is clearly illustra ted by Fig. 2/49a. It should be noted that the
compressive stresses
in the crown increa se consid erably with the lateral pressu
re (with increas ing l),
wherea s the normal stresses at the springings decrease at
a much slower rate.
2
-u TERZAOHI, K. and RICHART, L.: Stresses in Rock
about Cavitie s, Geotechnique 1952
57-75
 CROSS SECTION 127

Stresses due to bending are, of course, neglected in this comparison although they
are of a sign opposite to that of the normal stresses and may - es~ecially in the
case of large dimensions - dominate the stress pattern.
The conclusion to be derived from the foregoing considerations is that the
shape of the cross-section should always be selected in accordance with the relative
magnitude of vertical and horizontal loads. Where the vertical loads are relatively
larger, an ellipse with a vertical major axis is preferable, while for relatively larger
horizontal pressures one with a horizontal major axis will prove more suitable.
The resulting stress distribution is also affected by the degree of restraint, inasmuch
as normal stresses will dominate where deformations are restricted , while the
influence of bending stresses will predominate where deformations can develop
freely. For this re .. son under certain circumstances an elliptical section may be
preferable to a circular one even where pressures act in the direction of the minor
axis, since in this case the superimposed
:ompressive stresses will be decisive for
he section. On the basis of such consid-
erations, F. MOHR2 · 15 suggested that in
the case of restrained deformation it is
more reasonable to use an ellipse with
the major axis perpendicular to the di-
rectio n of pressure rather than parallel
lo it, provided that only a uniaxial (vertical)
mountain pressure is likely to develop
in the rock.
In practice a shape composed of circle
segments is used instead of an ellipse
(cf. Fig. 2/46). When plotting this line
the upper corner points of the clearance
a re fixed first , the size of the section
being determined by them. The specified
safety of- 30 to 40 cm is added and then
a circular segment is drawn through the
points thus obtained. The sides are form-
FIG. 2/50. Asymmetrical tunnel
ec! by circle segments of larger radius 9tetion
which join the crown arch tangentially.
The invert is again a circle segment with
its centre lying on the vertical axis. There is no need to join this segment tangen-
tially into the sides; they may intersect each other. The internal surface of the
li ning is defined by these curves and its thickness should be taken to keep the
resultant of pressures acting on the external surface and dead weight within its
inner third; otherwise a lining capable of resisting tensile stresses must be used.

: .u MOHR, F. : Gebirgsdrucl.: und Ausbau, G/iickouf 1952 27-28

128 GENERAL DESIGN

Vertical loads (roof loads) are carried by an arch. In order to resist lateral thrust,
horseshoe sections must be strutted by an invert arch at the bottom and it is
advisable to increase the dimensions of the side walls as well.
Inclined loads and unbalanced thrusts may call for asymmetrical sections. Such
conditions may be encountered in tunnels located at a small depth under a sloping
terrain, i.e. primarily in adit sections or in hang tunz.els. Similar ~ituations may
arise in tunnels driven in mountain sides parallel to the strike of steeply dipping
layers where the possibility of slides is not excluded, or in the vicinity of faults
and in slide zones. Under such conditions the structural axis of the arch should
be parallel. to the direction of load, provided this direction and load intensity
can be reliably identified. The section in such instances is fully enclosed (with
an invert arch). Asymmetrical sections are always more expensive than symmetrical
ones, owing to the unutilized area in excess of the necessary clearance, and to the
increased thickness of lining (Fig. 2/50). The shape of the section is influenced
further by the method of construction the choice of which is, in turn, governed
by the prevailing rock or soil conditions.

22.33. Influence of Construction Method on Shape of Cross-section

The influence of the construction method on the shape of the cross--section is
reflected by the following facts:
1. Conventional mining methods are suitable for driving horseshoe and flat
arched sections, and, less economically, for circular sections;
2. The shield method is restricted almost exclusively to circular sections;
3. The caisson or other sinking method is economical for rectangular and
circular sections alike;
4. The free-face method can be used for cross-sections of any desired shape.
The method of construction must, on the other hand, be chosen in accordance
with the prevailing soil conditions but may be influenced, to some extent, by the
availability of equipment, machinery and materials.
The considerable influence of the construction method on the shape of the
cross-section is clearly illustrated by the wide variety of cross-section designs for
the stations of underground railway tunnels (cf. Section 6.34, and Figs 6/161
to 6/ 172).

22.34. Influence of Tunnel Lining on Shape of Cross-section

The material of the tunnel lining must also be taken into consideration when
choosing the cross-section, since materials capable of resisting compressive stresses
only (stone and brick masonry, and concrete) are limited to structures composed
of arches and of robust buttressed walls, i.e. to those called upon to carry purely
compressive stresses, or tensile ~resses not exceeding the tensile strength of the
mortar used. Examples for tllJ.ese are horseshoe, circular and elliptical sections,
 CROSS SECTION 129

(the major axis of the last may be vertical, or horizontal). Materials capable of
resis~ing tensile and.bending stresses alike (reinforced concrete, steel and, tempo-
rarily, timber) can be used for lining sectjons of any desired shape, designed in
the most economical manner and permitting the fullest possible utilization of
space. (Rectangular sections with a flat, or arched roof, thin shell-like linings, etc.)

22.35. Size of the Tunr.el Cross-section

This is affected by the decision to build a single double-track tunnel, or two
separate single-track tunnels on Jines where two tracks must be accommodated
owi ag to traffic density. In the case of road tunnels the counterpart of this problem
is the choice between a multi-lane and two single, or double-lane tunnels.
Of fundamental significance in this respect is whether traffic density is already
large enough to justify the construction of a double-track tunnel, or whether the
enlargement to accommodate the second track will become necessary in the near
or more distant future only. Where the need for a double-track tunnel is imminent,
the problem is to decide whether a horseshoe, rectangular, or circular section
should be used · under the prevailing soil conditions considering the possible
methods of construction. A double-track tunnel will always offer advantages
over two single-track tunnels where conditions permit the use of horseshoe, flat
arch or rectangular sections, and will usually be adopted for surface tunnels,
in solid rocks, and for underground railways located at moderate depth under
the terrain and, finally, for subaqueous tunnels sunk immediately below the
river bed.
Where, on the other hand, large lateral pressures make the use of circular
cross-sections imperative, the construction of two separate single-track tunnels
will be more economical. Large rock pressures may necessitate the construction
of two separate tunnels with smaller cross-sectional dimensions even where a
horseshoe section is used. Reference is made to the construction of the Simplon
and Tanna tunnels, where the great depth of cover and unfavourable soil condi-
tions resulted in pressures which prohibited the economical realization of
a single larger cross-section.
The relative magnitudes of cross-sectional areas- of horseshoe and circular
railway tunnels accommodating a single and a double track. respectively, are clearly
illustrated by the following comparison:

Number
·Shape of tracks Area (m1 ) % Notes

Horseshoe 1 35- 50 55} No invert

Horseshoe 2 75- 85 100 arch
Horseshoe 1 45- 60 58} With invert,
Horseshoe 2 80- 100 100 arch
Circular 1 29·6 43'5
Circular 2 68·0 100
Elliptical 2 52-8 77'5

9 Szechy: J:he Art of Tunncllini

130 GENERAL DESIGN

As will be perceived, the cross-sectional area of two single-track tunnels of

horseshoe section is larger than that of one double-track tunnel of the same shape,
while the situation is reversed in the case of circular sections. Consequently, the
construction of two single-track tunnels with.horseshoe cross-section may be
more expensive, while for circular cross-sections it may be more economical to
construct two separate single-track tunnels.
Tunnelling costs depend,
however,not only on the cross-
sectional area, but also on the
length of the perimeter, which
may be regarded as a measure
of the volume of lining. The
perimeter lengths of single-
track tunnels are about two-
thirds those of double-track
tunnels irrespective of whether
they are of horseshoe or cir-
cular cross-section. This would
Tvnne/ ivitll!t 8500 point to the advisability of
constructing a double-track
FIG. 2/51. Clearance for single-track and double-track
tunnels (Soviet specification) tunnel in both cases. It should
be remembered, however,
that the lining thickness of
horseshoe sections, usually adopted in solid rock, hardly changes with the
span, a thickness usually being adopted for safety on the basis of past experience,
so that the perimeter length may well be regarded with minor corrections as
representative of the volume of lining. The lining thickness of circular sections
commonly adopted in loose soils where large lateral pressures are likely to occur
depends, on the other hand, to a great extent on the diameter used. When using
cast iron tubing elements, the lining volume of the single-track tunnel is only
35 % to 40 % of that required for the double-track tunnel although the perimeter
ratio of the two tunnels is 0·66.
It would thus appear that the economical solution in solid rock is a double-
track tunnel with horseshoe cross-section, while in loose, soft soils two single-track
tunnels of circular cross-section will be preferable. Standard cross-sections of
single and double track railway tunnels specified on Soviet raiiways are shown
in Fig. 2/51.
A considerable advantage of double-track tunnels is the improved ventilation
and the possibility of extensive repair and mai.ntenance work without closing the
tunnel to traffic.
Two single-track tunnels must be located at a suitable distance from each other
to eliminate superposition of the pressure zones developing around them. The
spacing depends, thus, on the depth of cover, the width of the tunnel, and on the
strength of the surrounding rock or soil (cf. superposition between rock pressures).
 CROSS SECTION 131

The tunnei axes should be spaced not closer than 25-30 m. (For the Budapest
Underground Railway this spacing is 23·5 m.) Particular care should be taken
in cases where the second tunnel is constructed at a later date, since the first tunnel
may then become subjected to additional loads; the shift of loads may result in
a rearrangement of the stress pattern brought already to an equilibrium and this
may lead to the development of cracks, water inrush and even to the break of the
existing lining. This is exemplified by the construction of the second Simplon

Case /. n,se/1.

l'1fO

F ro . 2/52. Possible methods for later widening of a single-track tunnel

tunnel,. where the tunnel axes are spaced at no more than 17 m at a depth· of
2200 m. The rearrangement of the pressure pattern due to the construction of the
second tunnel resulted here in deformations of the first which attained over some
sections a magnitude of. almost half a metre. During the enlargement of the
London Underground the tubings of several existing tunnels at Piccadilly Station
broke when the new crossing tunnels were constructed.
Where, in anticipation of future traffic requirements the construction of a
double-track tunnel is decided upon, consideration should be given to the possi-
bility of a two-stage development, i.e. to that of constructing a single-track tunnel
first and enlarging it to accommodate a second track only when this is actually
j ustified by the increased traffic density. In stable soils or rocks a two-stage devel-
opment of this kind is well feasible and may offer economical advantages, as
demonstrated by the example of the section along Lake Baykal of the Trans-
Siberian R~ilway (Fig. 2/52).
In designing the tunnel cross-section, allowance should be made for super-
eleva.tion in curves, as well as for the accommodation of drainage and ventilation
facilities. These will be dealt with in greater detail later in the appropriate chapters.
It should only be noted here that in the interest of improved ventilation the
entrance sections of road tunnels are constructed with a higher cross-section than
rhe central part. The profile of the tunnel will thus resemble a funnel ensuring
continuous draft in the tunnel even without artificial ventilation (cf. Section 46.52).
Enlarged portal sections also result in improved lighting conditions (cf. Section
46.61).
132 GENERAL DESIGN

Common internal dimensions of railway tunnels are shown in Table 2/VI,

but similar information can be obtained from Tables I/I and 2/I as well.

Clear Dimensions of Railway Tunnels TABLE 2/ VI

Double-track . tunnels
Country
Width (m) Height of arch above rail
level (m)

Italy 8·00 6-00

France 8·0-8·20 5·80-6·00
Germany 8·20-8·50-8·70 6·20
Austria 8·20 6·40

Single-track tunnels

Italy 4·60- 5·50 5·00

Austria 5·50 5·00
Switzerland (R_icken-tunnel) 5·20 5-80

REFERENCES
BENDEL, L. (1948): Tunnelgeo/ogie, lngenieurgeologie. Springer, Vienna II 507
BERGER, R . (1951): Untergrundbahnen und ihre Einsatzgrenzen. Ernst, Berlin
BLACK, W . T.: Great Engineering Achievements
CLEAVES, A . B. (1958): Geological Factors in Tunnel Construction, Proc. ASCE 84 1648
KRYNINE, D. P. and JUDD, W. R. (1957): Principles of Engineering Geology. McGraw Hill,
New York
LEGGET, R. F. (1962): Engineering Geology. McGraw Hill, New York
LEROY SCHARON, H. and CLEAVES, A. B. (1958): Geophysical Investigations for the Lehigh
Tunnel. Proc. ASCE 84 1650
MAYER, A. (1963): Recent Work in Rock Mechanics, Geotechnique, 1963 June
MOLLER, L. (1963): Der Felsbau. Ferd. Enke Verlag, Stuttgart
MosONYI, E. and PAPP; F. (1959): Miiszaki foldtan (Engineering Geology). Miiszaki Kiad6,
Budapest
NEUMANN,_ R. (196.4 ): ·Geologie fur Bauingenieure. Ernst, Berlin
Proceedings of the First Congress on Rock Mechanics, Lisbon 1966, Laboratorio de
Engenharia Civil, Lisbon
REEVES, A. B. (1958)': Geological Factors in Tunnel Construction, Proc. ASCE 84 1649
SCHREIER, F . (1960): Verlegung offentlicher Verkebrswege in Unterpflaster oder Tieflage?
Strassen u. Tiefbau 1
STINI, J . (1950)_: Tunr.elbaugeologie. Springer, Vienna
TALOBRE, J . (1957}: La mecanique des roches. Dunod, Paris

-
 CHAPTER 3

ANALYSIS OF LOADS ON TUNNELS AND

UNDERGROUND STRUCTURES

The most important potential loads acting on underground structures are earth
pressures, i.e. rock pressure and water pressure. Moving (live) loads due to vehicle
traffic on the surface can be safely neglected unless the tunnel is of the cut and
cover type, i.e. the depth of overburden is very small. Neither the loads resulting
from vehicles moving in the tunnel, nor from the material conveyed therein,
e.g. water, are usually taken into consideration unless the tunnel is cl'ose to the
surface, or in loose soil of very low bearing capacity. Exceptions to this are
water pressure tunnels-.

3.1. CAUSES AND TYPES OF ROCK PRESSURE

Rocks in nature, especially deep lying ones, are affected by the weight of the
overlying strata and by their own weight. Stresses develop in the rock mass
because of these factors. In general every stress produces a strain and displaces
individual rock particles. But to be displaced a rock particle needs to have space
available for movement. While the rock is confined, thus preventing its motion,
the stresses will be accumulated or stored in the rock and may reach very high
values, far in excess of their yield point. (It has been proved by KARMAN that
apparently rigid materials such as solid rock and concrete may be stressed beyond
their plastic-limit in this confined condition.) As soon as a rock particle, acted
upon by such a stored, residual or latent stress, is permitted to move, a displace-
ment occurs which may take the form either of ' plastic flow' or 'rock bursts'
(popping) depending upon the deformation characteristics of the rock-material.
The displacement may be of an elastic nature, should the elastic limit of the rock
material not be exceeded by the residual stresses.
Thus, whenever artificial cavitfos are excavated in the rocks forming the external
crust of the. earth, the weight of the overlying rock layers will act as a uniformly
distributed load on the deeper strata and consequently on the roof of the cavity
excavated. The resisting forces - as passive forces (shear strength) - are scarcely
mobilized prior to the excavation of the cavity, since the deformation of the loaded
rock mass is largely prevented by the adjacent rocks. By excavating the cavity
opportunity is given for deformation towards its interior. In order to mair.tain
 134 ANALYSIS OF LOADS

the cavity, the intrusion of the rock masses must be prevented by supporting struc-
tures and the load acting thereon is referred to as rock pressure. This terminology
also indicates that the magnitude of pressure is greatly affected by the magnitude
of the deformation produced.
The determinatio n of the magnitude of rock pressure is one of the most intricate
problems in engineering science. This complexity is due not only to the inherent
difficulty of predicting the stress conditions {primary stresses) prevailing in the
interior of the the non-uniform rock mass, but also to the fact that in addition
to the strength properties of the rock the magnitude of secondary pressures
developing after excavation around the cavity is governed by a variety
of factors, such as the size of the cavity, the method of its excavation, rigidity
of its support and the length of the period during which the cavity is left
unsupported .
Within any particular rock the pressures to which it was exposed during its
history are best indicated by the pattern of folds, joints and fissures, but it would
be extremely difficult to determine how far these pressures are still latent. There is,
in addition, a significant difference between the internal stress condition of rocks
and soils. Soils - particularly gravel and sand - were deposited uninterruptedly
and uniformly and this regular sedimentation and stratification was but rarely
disturbed by external forces . (The probability of external disturbance increases
with the age of the formation.) In soil mechanics only vertical stresses due to the
weight of the overlying layers, and horizontal stresses resulting from lateral defor-
mation - determined with the help of Poisson's number - are taken into consider-
ation. According to TALOBllE this approach cannot be accepted in rock mechanics.
At first a similar differentiation of internal stresses according to the orientation
of layers would be unimaginable there. Yet even in this case groundwater condi-
tions,· settlement and plastic properties ought to be considered as these all tend
to bring about an equalization between vertical and horizontal stresses. In rocks,
similar effects need not be considered. In sedimentary limestones or in rocks crys-
tallized from the molten state, for example, there are frequent indications from
which the existence of an initial hydrostatic, rather than an oriented stress distri-
bution may be concluded. The enormous changes produced by tectonic effects
act in the same stress field. Tectonic effects act in a sense opposite to gravitational
ones. The rock joints and :fic;sures are seldom vertical and bear signs of the numer-
ous changes which they have undergone in the course of time as a result of
various external effects even within the same block of rock.
In soils, especially in clays, it may frequently be observed that lateral pressures
do not develop in accordance with lateral deformations but undergo a gradual
transition from this initial stage to hydrostatic stress conditions. This process is
stimulated and aided by repeated seismic movements which contribute over a
lengthy period to the equalization of stratification, fissuration and the effects of
other irregularities. ·
In the hydrostatic state, however, the principal stresses are of equal magnitude
and perpendicula r to each other, i.e. shear stresses are zero. This assumption
 J35
CAUSES Of ROCK PRESSU RE

prevailing at great depth s.

was made by HEIM as early as 1878 for stress conditions
at suppo rt for this theory
Tunnelling practice was unable to furnish even gener
because of the wide variety of factors involved.
ally granu lar soils,
There is a furthe r difference between rocks and soils, especi
r being depen dent on three vadab les
in their stress -strain curves; those qf the forme
(cf. Fig. 2/ 10), i.e. load and deform ation and also time.
l, shoul d be under-
According to TERZAGHl secondary rock pressure in genera
a certai n heigh t above the tunnel, which
stood as the weight of a rock mass of
drop out of the roof, and the only con-
when left unsup ported would gradually
would be that this rock mass would fall into
seque nce of installing no propp ing pment of
ts would result in the gradu al develo
the cavity. Successive displacemen the
necessarily involv ing
an irregular natura l arch above the cavity witho ut
tunne l itself. Earth pressu re, on the other hand, would
complete collapse of the
c masses on the tunnel
denot e the pressure exerted by cohesionless, · or plasti
the absence of suppo rts -
supports, witho ut any pressure relief which would - in
complete disappearance.
soone r or later completely fill the cavity leading to its
enden t of the streng th
The magnitude of earth pressure is, in general, indep
it is only its distribution
and installation time of the suppo rting struct ure and
magn itude of rock pressures,
which is affected by the deformation of the latter. The
streng th and time of installation
on the other hand, is influenced decisively by the
ing the excav ation of the cavity
of props . This is because- the deform ation follow
of a plastic nature and extends over
in n;>ck masses surrou nding the tunnel is
the final deform ations and, thus, for
a period of time. This period required for
ses with the plastic ity of the rock and
the pressures to develop, generally increa
its cross- sectio n. The magn itude of deformations
with the depth and dimensions of
d by sufficiently strong
and consequently that of stresses can, therefore, be limite
bered, however, that
propp ing installed at the prope r time. It should be remem
decrease with increa sing
the intensity of plastic pressures shows a tendency to
by the tunnel lining and
deformations. Furth ermor e the loads are carried both
to utilize this co-op;}r-
the surrou nding rock and every attem pt should be made '
ation. true
sor RztHA , 'The
According to one of the authorities on tunnelling, Profes res,
of large rock pressu
art in tunnelling lies in the anticipation of the development which
resisti ng rock pressu res
which is far more effective than to find the means of
existing between menta l
have already developed. This contra st is similar to that
howev er mainly for rock, whe-
work and physical labou r.' This remar k holds
ies, the acting pressure will be rat-
reas in soils with plasti c deform ation qualit
her reduced by the yield of suppo rts.
res can be classified
The reason s for the development of secondary rock pressu
ing three main catego ries:
acco rding to RABCEWICZ in the follow

l. Loosening of the rock mass

2a. The weight of the overlying rock masses
2b. Tectonic forces
 136 ANALYSIS OF LOADS

3. Volume expansion of the rock mass, swelling due to physical or chemica

l
action.
These rea_sons lead in general to the development of the following three types
of rock pressure:
1. J,ooseni ng pressure
2. Genuine mounta in pressure
3. Swelling pressure.
The conditions under which tllese develop, the probability of their occurren
ce
and their magnitude differ greatly from one another and require the adoptio
n
of different constru ction methods. The possibility of their simultaneous action
is,
nevertheless, by no means excluded. The combined action of genuine mounta
in
pressure and swelling pressure particularly must be taken into account and
it is
difficult to discriminate between them.
The type of rock pressure encountered will depend mainly on the quality
of
rock and on the depth of the tunnel. As regards probabl e rock pressure conditio
ns
three basic groups are distinguished by RABCEWICZ:
(a) solid rocks
(b) pseudo-solid, soft and weathered rocks
(c} loose rock (soils).

3.11. ROCK PRESSU RE DUE TO LOOSEN ING

Of the pressure types listed above loosening pressure and its develop ment may
be dealt with first because it is characteristic of the process of the development
of
secondary pressures in general. Because of the deflection of the loaded roof
the
behaviour of the overlying rock mass above the excavated cavity is, accordin
g
to RABCEWICZ, similar to that of a soil mass piled up in a silo provide d with
a slot
at its bottom, which in turn is closed by a downward moving cover plate. Looseni
ng
pressures will more closely resemble silo pressures the smaller the cohesion of
the
soil mass .
When the slot is opened the pressure drops suddenly to a minimum value
but
increases again gradually as the slot is opened wider, without attaining, however
,
,, the intensity of the original geostatic pressure acting on the cover (Fig.
3/ 1,
experiment of KONCZL ). If the slot is left open, i.e. the roof is unsupported, a wedge-
shaped mass will gradually drop out from the overlying rock mass into the cavity
until equilibrium corresponding to the changes in stress conditions is re-established
.
First the rock mass 1 will drop out from the centre (Fig. 3/2) and the moveme
nt
will extend sidewards and upwards in the shape of a pointed arch until the
half
arches become capable of supporting each other at a certain inclination angle
oc.
 LOOSENING PROCESS 137
~

,.,t·I
2-fj

-t
2-0
~ ~
~
~ ~
II
1·5 '<
c:}

0 Z -1 6 8 10 12
Magnitude oj' cover plate openiflg i117(mm) Cover plate
FIG. 3/1. Pressure upon a cover plate at a silo bottom

The an,gle a. is in direct ratio to the cohesion of the particular soil. The height
of the arch is approximately

b b
h -- - - - -
malt - 2 tan a./2 = 2 sin <J, '

(cf. the theories-of TERZAGHI and

PRoTODIAKONov; Sections 32.314.
and 32.324.). It may be con-
cluded fro.m this that the magni-
tude of pressure is proportional
to the area of the wedge, i.e. it
increases as the square of the
span of the cavity. A phenome-
non frequently observed in cohe-
sionless soils is the development
b
of such chimney-like natural up-
breaks if the installation of sup-
ports is delayed. The roof of nat-
ural caves also displays a similar
shape. The weight of the loosen-
ed, propping wedge is thus re-
sponsible for the loosening pres- FIG. 3/2. Upbreaking process of wedge above
sure, which acts on the tunnel roof cavity
138 ANALYSIS OF LOADS

supports. 1:his pressure can, thus, be traced back primarily to the deficiencies
of excavation and secondly to those of temporary underpinning,- its magnit~de
depending consequently on the care exercised during construction. Excessive
blasting and careless overbreak may, for example, result in considerable pressures
even where the solid rock would otherwise require no support at all. Blasting
operations should therefore not be carried out in the vicinity of the boundaries
of the excavation unless unavoidable, and less active explosives should be employed

Arching ---~-- Ila!/ dome 3cfion

Arcl11ng action 118!/ dome action
Overbrea/t.

Supported to.,,
here
I I I
f!jasted. Supported.. j j ~lasted. j
to here
Overbreak if unsupported section Overbreak if unsupported section
is ver!f short is long
FIG. 3/3. Increase ·or overbreak as a function of the length of unsupported section
(after TERZAG HI)

for such work. The perimeter should be excavated to its final dimensions pref-
erably by hand tools, or pneumatic hammers. The installation of the temporary
supports offers another opportunity for the development of loosening pressure.
The height of the lo.osened zone increa5es with time and the longer the section is
left unsupported the greater span will develop to be overbridged. The load-bearing
capacity of the rock around a cavity decreases, or the amount of overbreak
increases as the square of this span (Fig. 3/3). There is always a gap left between
the lagging and the excavated face (overbreak). 3 ·1 Cavities remain behind the
forepoling spiles driven (especially when breaking out the crown segment owing
to the repeated change in the direction of the spiles, cf. ' Fig. 6/24). The upper row
of lagging settles together with the cap beam as the packing-wedges are replaced

3 1
· MOLLER, L.: Der Mehrausbruch in Tunneln und Stollen, Geo/ogie u. Bauwesen 1959
24 3-4
 LOOSENING P.ROCESS 139

by blocking (cf. Fig. 6/24), The bearing surfaces of the side posts are pinched
and even the sill beams may be pressed into the soil. These inaccuracies may
involve a crown settlement of 10 - 20 cm, but in excavating the top heading or
the full face, the settlement of the crown may attain as much as 50- 100 cm.
The necessity for erecting the supports rapidly and accurately to minimize the
extension of loosening, affects the selection of the construction method. This type
of pressure may be reduced effectively by the shield method, and by the cross-bar
mining methods if interstitial dimensions are smaller and the rigidity of supports
is higher than in the longitudinal beam methods and which are, consequently,
superior to the latter. Continuous construction methods are similarly superior
to ring-section methods (cf. Section 62. I).
When excavation is carried out without proper care, at shallow depth and in
loose rock, the loosened range may extend to the surface and appear there in
the form of open rupture lines. The loosening pressure in this case equals the full
geostatic load.
The height of the loosening zone will be determined otherwise by the rise of
the natural load-transmitting arch developing in the rock, or by the position of
an overlying resistant rock layer capable of carrying the developing rock pressure.
Stress conditions produced by loosening will the more closely resemble pressure
conditions in silos the less cohesive is the soil. Similar effects will act unquestion-
ably in cohesive soils as well, except that both processes, i.e. the increase of pressure
and the new state of equilibrium, will be established after a longer period than in
granular soils, and the lower the minimum pressure the higher the cohesion
will be.
The largest extent of loosening, expressed as a percentage of volume is, according
to STINI (cf. Section 32.321, KOMMERELL's theory):

Transitor y Permanent

Sand, fine gravel 10-20 1-2

Slightly cohesive loam 2
Loam, coarse gravel 25-50 3-5
shale, shale marl 25- 30 4- 6
Marl, clayey gravel, weathered sandstone 25-30 6-8
Hard clay 30- 50 8-10
Weathered rock 30-50 8- 10
Solid rock with short fissures 35-50 8-15
Diabase gravel 15/ 30 rnm 45
Solid rock, greatly fissured 40-55 10-25

It should be noted that tunnel linings must be dimensioned in most instances

to resist either geostatic or loosening pressure.
140 ANALYSIS OF LOADS

In a broader sense any pres:.ure due to the weight of the rock mass within the
boundaries of the protective zone (Trompeter zone, stress-free body) .may also
be referred to as loosening pressure. This rock mass is largely fissured and broken
by the redistributio n of stresses in the interior of the roe~ mas:; - following the
excavation of the cavity - and has lost its load-bearing capacity.
Solid rocks transfer the load acting on them by beam action to the sound
supports, while in loose rocks and soils load transfer to the undisturbed lateral
parts is bound to rely on friction developing during mass displacement. Deforma-
tion occurs in both cases, but whereas in rocks this is necessarily limited, in prin-

~b
Pro~blt JNximvm
_J_
DttiKaOWtr If UtrJUp·
porl,d .,.,,..-·-.....
_L
'I/ ·"· t
{ \
i. 0
- I
i
L __ ___ .Ji

Fm. 3/4. Extension of ultimate overbreak if no support is installed with

(a) horizontal stratification; (b) vertical stratification (after TERZAGHI)

ciple, to the deflection of the beam, in soils it must be of a magnitude sufficient

to mobilize a frictional force along the sliding surface capable of transmitting the
additional loading. The load-transferring formations developing in vertically and
horizontally stratified rocks over unsupported cavities are shown in Fig. 3/4,
while the same is shown for soils in Fig. 3/5. At first the load exceeds the bending
resistance of beams formed by the individual layers and determined by the width
of the open cavity (the span) and by the thickness of the load-transmitting layers.
Consequently those layers next to the cavity fail and break down. Deflections and
bending stresses decrease, however, with the height above the cavity, as the spans
become smaller due to the cantilever action of the failed layers (corbel arch).
Beyond a certain limit (which is mostly at 0·5 b in .solid rocks and B + m in
!oose sands) the stresses developing over the diminishing spans no longer exceed
the bending rec;istance of the rock or the compressive and shearing strength of the
soil, respectively. The thinner the layers, the lower the resistance of the rock and
the more pointed and the higher the arch will be. The caving zone is overlaid by
one of deformations within which the strains gradually become elastic. A certain
 -- - ---- =-- - - - . -
' - - - 4. .

LOOSENING PROCESS 141

£03ding tne loJding upon

FIG. 3/5. Arching wedge ace. wedge bdf. m
above cavity Oirection of disp/.3ce-
(after TERZAGHI) r11Ct1t during eKcaraLion
l03ding dim/fl!
uno!'rM"

time is, however, required to elapse before this phenomenon develops fully and
the length of this period increases with the plasticity of the rock. It has already
been mentioned that the rupture loosening process can be limited by installing
artificial supports and the development of the protective zone can be speeded up.
Consequently, the shorter the period for which the roof is left unsupported the
smaller the zone affected by loosening and so the smaller the load on the lining
will be.
B:::fore proceeding to a more detailed discussion of genuine mountain pressures
let us investigate more closely the process of development of rock pressures in
gen:::ral as it has been investigated and described by BRANDAU, WIESMANN and
WILLMANN, and demonstrated experimentally by LEON and WILLHEIM. This process
defines how primary pressures are converted in the vicinity of the cavity - as a
consequence of its excavation - to secondary pressures.
An undisturbed horizontal surface element in the interior of the' undisturbed
soil or rock mass is acted upon by the weight of the overlying soil, or rock column,
constituting the vertical force referred to as geostatic pressure:

II

Pv = L Yi hi • (3.1)
i=l
 142 ANALYSIS OF LOADS

The horizontal force, on the other hand, may vary between very wide limits
and is dependent upon the magnitude of lateral deformations. In undisturbed
solid rocks the horizontal pressure is very likely to be much smaller than the ver-
tical one and their relative magnitude can be expressed with the help of Poisson's
number derived by the ratio of vertical to horizontal elastic deformations. If it
is assumed that lateral deformation is entirely prevented, then the specific hori-
zontal deformation

eh
1 (
=E Ph -
Pv +Ph.) = ·0,
rn
and because
ev = pvfE,
and
Ph 1 µ
A0 = - = -- - or - -
Pv rn - I 1 - fL '

I
where µ =- = Poisson's ratio and consequently
m

µ
Ph= Pv - - - ·
1 -µ

It must be noted strictly, however, that this applies only to elastic deformations
and must not be used where plastic deformations are encountered.
In rocks, however, which have undergone fo lding in the course of their geological
history, the horizontal pressure may vary in magnitude from a value corresponding
to undisturbed conditions to one approximating the crushing strength of the rock,
depending on whether the lateral thrust causing the fold has ceased to exist or is
still latent.
There being no method available for predicting the horizontal pressures actu-
ally prevailing in the interior of the rock, their magnitude can only be estimated
from external manifestations (such as propping of the sides at moderate depths).
Yet, as will be seen, the role of horizontal pressures in the development of genuine
mountain pressures is considerable.
A solution for ·t he analytical determination of stresses developing around a cir-
cular cavity in elastic media was given by MINDLIN.'·2 Owing to the uncertain
magnitude of horizontal pressures, three "fundamental cases are discussed:

1. The hydro-static stress condition, where

Ph = Pv = 'I h · (3 .2)

32
• MINDLIN, R. D.: Stress Distribution around a Tunnel, Proc. Am. Soc. Civ. Eng. 1939
Apr. 6/9- 642

(
 LOOSENING PROCESS 143

II. Horizontal pressures sufficiently large to confine lateral deformation and

increasing linearly with the depth, where

µ
Ph=Pv - - - . (3.3)
1 -µ

III. The case of no lateral pressure, i.e. where Ph = 0.

The results derived are theoretically correct, but the solutions are highly involv-
ed: the bipolar co-ordinate systqn of Jeffery is used and the results are expressed
in terms of transcendent functions.
For the first case an approximate solution is given by KERISEL, according to
whom the tangential stresses can be computed from the relationship

a 1 = yh ( 1 + :: ) ,

the radial stresses being given as

2

- ,.a
~
) ,

\\ here h the depth of cover

a =
the radius of the circle
r = the distance of the point under consideration.

The distribution of secondary stresses along the horizontal diameter based on

these expressions is shown in Fig. 3/6a. At the perimeter the tangential stress will
increase to double its original value, while the radial stress will be zero.
Proceeding away from the perimeter the tangential stress will gradually decrease,
'-\hcreas the radial stress will increase towards unity, i.e. to the original stress
c,,ndition. As will be seen, beyond the valuer/a> 4 the stress increment induced
1 less than 6 %, so that the extent of the disturbed stress. zone may safely be taken
.1 twice the width of the cavity. In the exact theory derived by M1NDLIN the
influence of depth is also considered, i.e. the deeper the location of the cavity,
tile smaller will be the stress peak, but the larger will be the extension of the
di.::.turbed stress zone.
The disturbance of the original state of eq1Jilibrium produced by the excavation
of the cavity will, thus, extend not only to the zone above the roof but also to
the rock masses under the cavity and at both sides of it. According to WILLMA~
the process can be visualized as in Fig. 3/66. Following the excavation of the
ca, ity of width b the load bp is transferred to the adjacent intact rock walls in
-...bich the stress is increased thereby to Pi· The stress increment p' decreases
gradually to zero in the undisturbed range (2). The peak ordinate Pi, i.e. the
magnitude of the stress increase p', as well as the extent of the range disturbed
144 ANALYSIS OF LOADS

BJ Theoretical red1str1bvtion ofstresses

after the exc4v,9tio11 oft1 circu/3r. hole
IJ

-
~
~
~~ • I
~

2 J 4 5
r/4

Or1y111a/ stress
1 0
distribution
1l 1111 1111 11111111 11 111 11111 11111111 11111t11
e e Brttfle rigid
t' p' rock (p1 ;:p2
_\ Plastic rock
Stress distr1bvL1on
after eutbreak of
the /Joie
Extension ofprotective

Stress distribution
after development of
tl?e protective zone

FIG. 3/6, Redistribution of stresses around the hole

(a) on the basis of elasticity theory (MINDLIN-KERISEL) ;
(b) according to WILLMANN's practical observations

depend on the quality of rock and on the depth al which the tunnel is located.
From practical experience gained in plastic rock the stress appears to increase by
50 % and the width of the disturbed range may be up to three times the width
of the tunnel. In rigid rock, on the other hand, the 5tress may increase to twice
the original value (Mindlin.'s theory of elasticity) but the extent of the disturbance
is limited to about another width of the cavity itself.
The rock, in turn, in its partly unconfined state is incapable of resisting the
increased stresses and will gradually fail in the course of the process outlined in
 LOOSENJN G PROCESS 145

Fig. 3/7, as its load-bearing ability is gradually exceeded in the immediate vicin-
ity of the cavity leading to a consequent redistribution of stresses with a transition
from state 2 into state 3 as shown in Fig. 3/6b. In this redistribution process the
rock is either crushed or suffers plastic deformation, and the process is, thus,
the development of the previously mentioned loosening process i.e. that of an
inert stress-free body.

Development of rock-pressure
__ _ Rock
Rock ~ ::--=- looseningII.
---
I.
f3stic Roof beam is
Stress distribution Drift rt i
at the moment of deflecting
outbreak

Ill.a. f.ateral spal!tng Ill. b. Elastic ext ension

/ Pressure body

Plastic rock , Pr· 1-5p

Rigid rock, Pr • 5p Stress relieved
IV. body
Plastic
expansion

Slidtng plane..---
format1on

F1G. 3/7. D evelopment of stress-relieved zone owing to the deformation processes

occurring around the cavity

This process is illustrated in Fig. 3/7. The weight of rock above the cavity acts
as a load on its roof a) deflecting the roof beam. The rock masses follow this
movement and become cracked and loosened by bending (II). Owing to bridging

10 Szecb y: The Art o f Tunnelling

146 ANALYSIS OF LOADS

action the side walls receive additional reaction forces as supports of the over-
bridging beam, or arch. If the resultant load exceeds the unconfined compressive
strength of the rock wall, its edge will crush and pop out, and the thrust line shifts
further inward until the magnitude of-the increased stress decreases to the strength
of the rock, the stress-condition of which changes towards the inkrior from an
unconfined to a confined one (Illa and b). As a further result of this redistribution,

FIG. 3/8. Development stages

of stress-relieved zones around a tunnel
section broken out in successive steps

FIG. 3/9. Order of succession of fractures observed

in LEON's compression tests around a hole
in a perforated marble slab

the vertical overload at the bottom level of the cavity induces lateral pressures in
the rock under the cavity as well. This lateral pressure, in turn, gives rise to vertical
pressures acting upward towards the unsupported floor of the cavity. Thus, the
rock forming the invert will be subjected to an upward-directed bending, producing
phenomena similar to those taking place at the roof, but on a reduced scale. This
bending crushes and loosens the rock of the invert (IV). In this manner a con-
tinuous fissured and crushed zone is gradually developed around the excavated
cavity. The rock in this zone is incapable of resisting stresses, and no longer
participates in the transmission of loads. It forms the so-called stress-free body,
but its weight acts as an inert mass on the structure supporting the excavated
 LOOSENING PROCESS 147

cavity. Under this load the supporting structure suffers further deformation
towards the cavity which, in turn, leads to further loosening, and so to the fully
continuous formation of the stress-free body. The new state of equilibrium is
gradually reached in this way (V).
The development of stress-free zones around a tunnel section excavated in
several stages is shown in Fig. 3/8. The shape of these zones depends, naturally,
not only on the shape and dimensions of the cavity but also on geological con-
ditions. The time factor which plays a decisive part in the development of the
protective zones must by no means be neglected. The deformation of rocks has
been shown to be a process extending over a certain period of time (cf. Fig. 2/ 10).
A certain lapse of time is therefore required for the development of the protective
zone which is a function of the nature of the rock (elastic or plastic), of the magni-
tude and depth of the cavity, and the rigidity and installation time of supports.
Cases are known where the lining was completed before this development ter-
minated. A very frequent cause of failure of tunnel linings is, in fact, the delayed
development of the protective zone, or the too early construction of the rigid
lining. The development of the protective zone is of fundamental importance in
tunnelling. Without such zones the construction of tunnels. at several thousand
metres depth would have been impossible.
The above phenomena were demonstrated experimentally by Leon, who used
perforated marble blocks for his experiments. Cracks due to flexural tension were
obseryed to develop first at the roof and at the floor along the boundaries of the
cavity when placed under compression (Fig. 3/9), which led at first to a stress
relief at the side-walls. As the load was increased external roof cracks opened (3),
starting at about 45 degree inclination and converging at the centre, while at the
same time the first tension cracks (1) and (2) closed slightly. The appearance of
such cracks (3) could not be observed at all in the blocks perforated with circular
holes. These cracks are therefore obvious indications of the fact that the shape
,,f the cavity attempts to adjust itself to the stress flow. Popping of the side walls
,-.ccurred as the load was still further increased (4). The side walls themselves
,uffered ~lastic deformation and where their yield in any other direction was
preYented, they moved towards the cavity. Partial failure of the cross-section was
f, •llowed by the redistribution of stresses everywhere. As, after popping of the
side walls, the compressive stresses shift towards the still sound part of the cross-
section. only compressive stresses will prevail above and below the cavity once
the roo f and invert are cracked. The cracked zones become compressed and bulge
imilarly towards the cavity. Owing to this bulging effect temporary roof pressu res
ma) considerably exceed the value corresponding to the weight of the stress-free
body developed. The pressures actually developing are, however, not only vertical
but horizontal as well, the latter in the form of a passive resistance preventing
the deformations due to lateral expansion and corresponding to Poisson's ratio,
rather than in the form of a direct force. The lateral pressures, in turn, prevent
the development of tensile cracks in the roof and the floor and this is why tension
cracks are seldom observable- in practice in these places.
IG•
148 ANALYSIS OF LOADS

3.12. GENUINE MOUNTAIN PRESSURE

Let us now proceed to a discussion of genuine mountain pressure. This phenom-

enon will appear when the secondary stresses developing around the excavated
cavity, exceeding the primary stresses prevailing in the undisturbed condition,
exceed the strength of the rock material not only in the roof but jn the side walls
as well, i.e. they attain an order of magnitude lying at the limit of plasticity. This
condition may be produced, however, not only by the fact of the absolute proxim-
ity of the excavation proper (i.e. by the stress superposition involved) but may
be encountered anywhere, where the rock has been originally in a so-called
latent-plastic state. This denotes the condition where the plastic-flow of the rock
bas been prevented by its confined state, i.e. its plastic deformation was prevented
by the surrounding masses. It is known from KARMA.N's experiments, that under
compound (triaxial) compression (i.e. under perfect prevention of lateral expan-
sion) even the most rigid and solid rock materials (including concrete) may be
brought into a plastic state. Consequently, whether any rock material at any
depth may be brought into such a stress condition is in practice merely a function
of the magnitude of the applied pressure (depth of the overburden), of the compres-
sion strength of the material and of the degree of the prevention of lateral expan-
sion (side-pressure coefficient, lateral rigidity, etc.).
This condition cannot be manifested, however, until plastic deformation
takes place, i.e. until the lateral support (confinement) either as a consequence of
its insufficient supporting capacity or of its entire removal (excavation of a cavity)
is eliminated. When dealing with genuine mountain pressures it is not forces
but movements which have to be considered; movements brought about by break-
ing out a cavity and extending over definite surrounding. Forces will become
apparent only when attempts are made to prevent these movements. The formation
of a protective zone or stress-free zone will come to realization here through the
gradual extension of the movements of the adjacent zones and through the gradual
decrease in its intensity.
Genuine mountain pressure is, thus, essentially some revelation of geostatic
(overburden) pressure, the magnitude of which depends largely on its geological
structure and on its tectonic disturbance.
Genuine mountain pressure is, thus, a sort of primary pressure. The transfor-
mation process itself is initiated, according to KASTNER,3· 3 essentially at that
moment when the secondary tangential pressures acting on the side walls attain
about double the primary value as a result of the stress redistribution described,
and exceed the unconfined compressive strength of the rock. Thus, the side wall
fails and breaks in or, at least, the rock is brought within a certain zone into
a plastic state in which its resistance is exhausted.

3· 3 KASTNER, H. : Ober den echten Gebirgsdruck beim Bau tiefliegender Tunnel, Ost.
Bauzeitschrift 1949 10-11
 GENUINE MOUNTAIN PRESSURE 149

Genuine mountain pressure with its inherent magnitude is thus of a temporary

character only and will manifest itself in a form depending on the rock group
classification given by RABCEWICZ quoted in the introduction.
Evidence is supplied in solid rock by popping at the side walls, or even spat-
ling on the roof and floor, and exceptioDally by some scaling observable at the
ground surface. Thus, it is actually responsible, in general, for the development
of a n oval self0 supporting cavity corresponding to the redistributed pressure
pattern, i.e. for the crushing of the rock lying beyond the excavated cavity,
and consequently it initiates the development of the protective zone men-
uoned above.
In pseudosolid rocks the development of the protective zone is a much slower
process and is characterized by their plastic intrusion into the excavated cavity
instead of successive fracturing around it. This plastic encroachment, in turn,
tS made possible by the plastic properties of the rock. The rock in this disturbed
zone of stress-relief is at a plastic limit rather than in a crushed state. In the
development of the protective zone the pseudosolid rock becomes loosened as
a result of the stress reduction ensuing at the boundary of the cavity, and of the
subsequent plastic deformation. The decrease in stress intensity computed after
3
F ENNER as a function of the deformation is shown by MOHR. .4 In similar cases
when there is no time to wait for the full development of plastic deformations the
use of a compressible bedding behind the rigid lining - in this particular case
boiler ash - is suggested by MOHR. This layer would absorb most of the deforma-
tions by its own consolidation and would, thus, relieve the lining of the correspond-
ing part of the load. It should be noted that the higher the stresses and the softer
the rock, i.e. the more pronounced its plastic properties are, the farther away
from the cavity the peak stress will shift and the wider will be the extension of the
stress disturbance (cf. Fig. 3/6). The plastic deformation initiated will exert
a direct pressure, genuine mountain pressure, on the supporting structure. The
magnitude of this pressure will increase in proportion to the attempts to exclude
the rock material from the cavity, i.e. the more tightly spaced rigid linings will be
a pplied. Since deformations last until the protective zone is fully developed, and
in such rocks this may be delayed for several months, early supports may also
entail adverse consequences. Genuine mountain pressure accompanying deforma-
tion may attain - depending on the depth of overburden - an order of magni-
tude of several thousand tons per square metre and there is no temporary support
system capable of resisting such a pressure. In such cases the immediate installa-
tio n of the supports may prove harmful and if the permanent lining must be
co nstructed before the full deformation can develop, an unfilled gap should be
left between the extrados of the lining and the rock, or this gap should be filled
wi th a compressible . material such as slag.

3 · 1 MOHR, F . : Kraft und Verformung in der Gebirgsmechanik untertage, Vortriige de r

IJaugrundragu.-:g 1956, W. Ern st, Koln 1957

 150 ANALYSIS OF LOADS

The poorer the strength of rock the smaller the cover depth required to produce
genuine mountain pressures. For example, in marl and clay-shale genuine moun-
tain pressure may be encountered at a geostatic pressure of 200 tons/m2 ; in lignite,
under an overburden depth of 30---40 m it is at a geostatic pressure of 60-80
tons/m2 . This is frequently accompanied by swelling pressures (cf. Section 31/3).
The only remedy against the occurrence of pressures of tnis type is the continued
removal of the intruding material during construction. The material tends to
intrude into the cavity from both sides, and from the roof and bottom alike,
so that it is manifested both in roof settlement and also in bottom upheaval.
Initiated movements are usually slow and not necessarily instantaneous, but
generally cannot be forestalled.
Whereas in counteracting loosening pressure the most effective construction
method is for the cavity to be excavated as fast and with as little settlement as
possible followed by support at the earliest possible moment with a rigid and
permanent structure, in the case of genuine mountain pressure the t,ype of rock
must be first ascertained. The methods required for supporting the excavated
cavity against genuine mountain pressure will be entirely different in solid rocks
than in pseudosolid ones. The occurrence of genuine mountain pressure in solid
rock is altogether exceptional and reveals itself in popping which, however
unpleasant during construction, does not affect the final stability of the tunnel.
Linings of moderate thickness cast with relative rapidity tightly against the rock
wall have been found to ensure the required solidity. A rigid lining should be
required everywhere, but especially at the roof. In pseudosolid rocks, on the other
hand, (such as clay, clayey shale, phillite shale, crushed and modified gnt_iss, etc.)
the early construction of the lining has been found to lead to failure whenever
the plastic state of stress has developed. Temporary supports especially must not
be designed to resist genuine mountain pressures. No practical means are avail-
able to resist these tremendous pressures and therefore the development of
the protective zone must be awaited. This is a question of time and space. It
is essential to provide for easily replaceable temporary supports. The pressure
of the material intruding forcibly along the unlagged sides, can be com-
pensated for most effectively by continuous yielding. This can be effected
in three different ways:

1. By prolonged maintenance, repeated replacement and renewal of the tem-

porary timbering until the deformation process is completed;
2. By the use of yielding supports;
3. By leaving a space between the extrados of the lining and the rock (cf.
Fig. 3/ 10).

In the last case the possibility of grouting should be provided for. The structural
advantages of grouting are illustrated in Fig. 3/75.
 151
G ENUIN E MOUN TAIN PRESS URE

loping arou nd a circular section

Concernin g the extent of the plastic zone deve
theory has been expounded by KAST ·
and the impo rtant role of lateral pressures, a
NER · which, using the nota tions
35 of Fig. 3/ 11, will be reproduced below.

80

70 BJ Compre~s1on of fuel ash

~ I
~ 60
~
~
l(f 50 I

40 ~- -
)47-
1 - - - ----1-"- '----
I 1 .
b; Hock pressure on ltmng as a 1/Jnc-,....__ _ ....J

JO t10n ofits deformalton

I
20

10

00 '5(J !al /5(} 200 250

I
44 l?adte/ s/Jortemnf1 ( cm)
F10. 3/10. Decrease of n, representing the pressure
pressures as a function oflin ing defor matio
reducing.effect of comp ressib le beddi ng (after FENN ER)

i R, - fh
-- CJ ~fi =kj h, J. · I

t · - Elastic zone
6rr , Dtr

- Plastic zone
6,.p I t1p
~rp "~r
Dtp ·qr
F G. 3/ 11. A 1nt1 ti'.>l i of KAST NER's theor y

Ost.
gsdru ck beim Bau tiefliegender Tunn el,
:s-. KASTNER, H.: Ober den echte n Gebir
Bauzeitschrift 1949 JO- 11
 152 ANALYSIS OF LOADS

Considering the state of stress developing in the vicinity of a circular cavity, the radial
stresses u, at the boundary will be seen to diminish to zero for lack of any internal support.
At the same time the tangential stresses u, assume peak values, which are in excess
of the
strength of the rock, leading to a plastic stress-condition. The deeper the tangential stresses
propagate inward into the rock their magnitude will decrease in proportion to the increase
of cross-sectional area affected until at a certain distance it will fall below the elastic
limit,
beyond which the rock will remain in the elastic stress condition. On the other hand, the
magni-
tude of radial stresses will increase within the plastic zone towards the interior of the
rock,
since the particles of the plastic material, being acted upon by the tangential stress as a normal
force and under the influence of friction, will be more and more capable of resisting
radial
shearing stresses. This variation of stresses is well. illustrated by a numerical example
by
KASTNER (Fig. 3/12). In the case when the external stress is increased
by opening a cavity to
twice its original value p , this value according to the elasticity theory remains below the
uncon-
fined compression strength of the rock u.,; the distribution of radial u, and tangential
stresses is shown on the figure by the dashed line. This case is compared with that in
u,
which
2p > u., (full line).

60 70 80
r(m)

FIG. 3/12. Development of radial and tangential stresses

around a circular cavity in elastic 6
and in plastic rock (KASTNER )

FIG. 3/ 13. Stress-str ain diagram of an ideal plastic material 0 e

 GENUINE MOUNTAIN PRESSURE 153

/
./
/
C
/
/

T-l~cot.9n~:, ~P.1 I IM I
I
, .. 617!( ·I I I
f
1 -,
- _. _ _ Oto ;6ro ----l I
I
1 ..- 6,ta -I
FIG. 3/ 14. Relation between normal and shearing stresses at the plastic limit condition

The boundary of the plastic zone developing around the circular cavity was determined
by KASTNER on the basis of the followi~g consideration:
The rock material was assumed to be an ideally plastic material having a stress-strain
curve similar to that in Fig. 3/ I 3. By adopting MOHR's theory of failure and with reference
to Fig. 3/ 14 the limit condition for plastic deformation can be expressed as

1 +sin</> cos</>
a,, - a,. .
1 - sm <I>
- 2c .
1- sm <I>
=0. (3.4)

On introducing Airy's stress function it is derived further that

a,. = i.a: [r: r--·- J] '

1

a,, = a [ (r)'·,-• -
}., : I A• \-;- 1
·1
J, (3.5)

r, = O,
where
1 + sin </>
).• = - - - - ,
1 - sin <I>
a = the radius of the circular cavity.
In the above expressions the strength properties of the rock material only are included
and these are unaffected by rock pressure for which a hydrostatic distribution (p, = Ph = p)
u assumed .
154 ANALYSIS OF LOADS

In contrast, the elastic stresses in a disc of infinite length perforated by a circular opening,
are3 ·8 :

-r .) + ----.-
0 Ta
u,, = p ( 1 - u,,
r· r·
(3.6)
Uv=P (l +r!)
,
: --.a,,r • r·
Ta

T, = 0.
At the boundary of the plastic and elastic ranges, where r = ra, the above stresses assume
the following simple form:
<1,, = a,,

u 1, = 2p - a,,

T, = 0. (3 .7)
On the boundary of the plastic and elastic zones the annular stresses calculated by the
plastic and elastic theories must be equal. From this compatibility condition the radius r0
of the boundary circle is obtained as

Ta = a [-2-
"• + 1
Uny + p ()., -
a.Y
1)] ,.~1_1. (3.8)

The extent of the plastic zone calculated in this manner is, however, too small to account
for the great masses of material found to intrude in practice into tunnels whenever genuine
mountain pressures are encountered. It should be remembered, however, that the extension
of the plastic range depends to a considerable degree on the ratio of vertical to lateral pressures
(1.). By starting from the condition of full plasticity, in which

_!_ (u, - u' ) + a.y) 2

T = 2 T2 = ( = k 2 _(y h)! (3.9)
mu 4 2 4'

a theory was derived by KASTNER for this case also, but the stresses a,. u, and T were calculated
from the theory of elasticity. In the above expression yh denotes the geostatic pressure and
k is the ratio of its unconfined compressive strength, i.e. k = u.y/yh. The formula derived
on this basis for any direction t/1 is

cos 2 2 t/J + 2 cos 2 t/1

1 + ). 1 - 2-x + 3a
-r=-.r - 4- (2- _- 3a--=,-
2 4
-
l-r=-.r
+ ). )' -
1 a•
- (.,...- _- -a-')-
4 2 3

(1 + 2a:
- -- -3a - + --~-------
2 4 2
- )k 2
- 4a (2 - 3!X
2 (1 - 1.) 4a: (2 - 3!X
2)
= 2 2 2)
0' (3.10)

where a = a/r = the relative distance in a radial direction .

The boundaries of the plastic range obtained when using this relationship are shown for
different k values and for the lateral pressure ratios ). = 0, ). = 0·5 and ). = I ·O in Fig.
3/ 15a. In these drawings the shaded areas indicate the zones where k < 1·O, i.e. where the

3 ·8 TlMOSHENKO: Theory of Elasticity. Mc Graw Hill, New York 1951 79

GrRKMANN: Fliichentragwerke. Springer, Vienna 1956 142
 GENUINE MOUNTAIN PRESSURE 155

geostatic pressure exceeds the unconfined compressive strength of the rock. Beyond the limit
). = 0·5 the plastic ranges will be seen to be confined to a defined area, but at relatively small
lateral pressures they may extend to infinity, starting from the springings at 45 °. Adopting
MoHR's failure theory instead of the above fundamental equation for Tmax, the boundaries ;hown
in Fig. 3/ 15b were obtained with 4> = 30°, c = 25 kg/cm\ h = 300 m, y = 2·5 tons/m3, for
various lateral pressure ratios J. in a given case. As may be seen a confined plastic zone restricted

a) b) c)

f' 4
a-

FIG. 3/ 15. Variation of stress isobars around a circular hole as a function of the lateral earth
pressure coefficient

to a relatively thin annular area is characteristic when external pressure distribution is uniform
ti. = l ·O), while for .l. = 0· 5 the plastic zone is still relatively narrow and closed and is limited
by a curve returning around the springings. With a further decrease of the ). values the exten-
sion of the plastic zones increases rapidly, chiefly in the direction of the 45 ° diagonals. At a
rat io ). = 0· 141, the plastic zone extends to infinity.
When a forced intrusion of external soil masses is experienced in spite of uniform external
pressure distribution it may be accounted for only by the presence of dilatation pressures
~hich is quite common in clays and clayey marls due to an increase in their water content
upon load release. 3 •7

As is clearly demonstrated by the theoretical evidence presented above, the

plastic zone in pseudosolid plastic rock will develop around the tunnel in a less
.. nd less favourab le manner as the relative magnitude of the lateral pressure
decreases and a corresponding increase of genuine mountain pressure is incurred,
\\ hich reveals itself by the intruding earth masses mainly at the springings. The

·; T ERZAGHT, PROCTOR and WHITE: Rock Tunnelling with Steel Supports. The Commercial
.!iearing and Stamping Co. Youngstown, Ohio, 1946
156 ANALYSIS OF LOADS

development of the protective zone around the tunnel is also disturbed by tht
large plastic zone, since the ring of necessarily increased tangential stresses is
either interrupted or shifted towards the interior of the mountain to such an extent
as to render it ineffective as far as arching action over the tunnel and load release
are concerned.

3.13. SWELLING PRESSURE

Under certain conditions swelling pressure is encountered in clays, in decomposed

clayey rocks, or in rocks intermingled with clay strata. The explanation offered
for this phenomenon by TERZAGHI is that with non-uniform stress relief, these rocks
start to swell in a manner similar to clay. Water is drawn away from adjacent parts
which have a higher load intensity (pore water migration), whereby swelling is
intensified, strength is reduced and compressibility is increased. As a result of
excavation, swelling pressure may be experienced primarily at all inner surfaces,
especially at the working face, which are left unsupported and are, to a certain
extent, relieved of the· load as well as the floor and to a lesser extent the sides.
After opening the cavity pore water will migrate from the more loaded inner
parts towards the released zones around the cavity. The increase of watet content
at these inner faces is accounted for by this pore-water migration rather than by
the extraction of water vapour from the intruding fresh air.
Swelling pressures are of an unpredictable magnitude and may be extremely
large. Their p~riqd of development may vary from a few weeks to several months.
Initially, i.e. immediately after excavation this pressure is insignificant, then it
increases at a higher rate and in the final stages the increase is again slowed down.
In shallow tunnels this pressure may be considerably higher than the geological
(overburden) pressure and in preloaded clays it may attain intensities as high as
10-20 kg/cm2• Deformations produced result in a reduced load intensity on the
supports. The method of protection developed from experience consists of impos-
ing no restriction on swelling until this has attained a ~rtain limit, and of construct-
ing the permanent solid lining at a later date. It could be stated in addition that
the development of the loosening core discussed in the preceding paragraph is
not hindered by swelling pressure.
The external manifestations of swelling pressure differ so little from those of
genuine mountain pressure that it is very difficult to separate them according to
the inconvenience, deformation and damage caused. This is rendered all the more
difficult because swelling rocks usually have a low modulus of elasticity and are,
thus, capable of exerting genuine mountain pressure even at moderate overburden
depths. Considerable lateral and roof pressures encountered under shallow cover
in the entrance sections of the tunnel in relatively stable rock indicate that the
pressure is due to swelling. Exact identification is possible on the basis of physical
soil investigations only. The danger of swelling exists, in general, in all types of
clays, in clayey rocks, clayey shales, slates and marls. Several marls display an
 --- - - - - --.-:;

LOCAL ROCK PRESSURES ]57

increased tendency to swelling upon repeated wetting and drying. Pore water
squeezed out by blasting, subsidence, sliding or by genuine mountain pressure
may also lead to saturation and to swelling pressures.
The most familiar among the swelling pressures caused by chemical action is
that following the transformation of anhydrite (CaSO4) into gypsum (CaSo4 +
+ 2H 2 O = CaSO4 • 2H 2O). Besides unwelcome pressure this transformation is
accompanied by the development of aggressive agents which have a deleterious
effect on the lining. The decomposition of iron sulphide (FeS 2) also results in
an increase in volume, although in this case the aggressive action is usually the
more dangerous factor. Volume and pressure increases are further involved in
the transformation of olivine-bearing rocks into serpentine under the action of
water and atmospheric effects, as well as in the transformation of rocks containing
iron oxide into iron hydroxide. The expansion (upheaval, scaling) of shales and
phyllite rocks accompanied by corrosion, especially if they are rich in mica, iron
sulphide and other scale-shaped minerals, also gives rise to a certain swelling
pressure.
It should be noted that the gradual development of the protective zone can be
observed in the case of sw<,!lling pressure as well.

3.(4. FACTORS AFFECTING THE MAGNITUDE AND LOCAL TYPES

OF ROCK PRESSURE

Rpck strength and the residual (orogen) stresses introduced into the rock mass
in the course of its geological history are the most important factors in the reduc-
tion of the magnitude of rock pressures. Stratification, water content and per-
colation effects are similarly important. The effect of stratification has already
been dealt with in connection with preliminary geological studies (cf. Section 21.3).
It should be added here that in the case of tunnels following the strike of steeply
inclined' layers, the vertical component of the weight of the inclined strata may
be larger than the geological pressure related to the tunnel axis, H/sin a. > h
(Fig. 3/ 16). In steep hillsides tunnels may be subjected to creep pressures as well
(cf. Section 31.4), regardless of their direction. Creep is understood here as the
slow downward motion of rock and soil masses on steep slopes.
Besides creating direct additional pressure and reducing the internal shear
strength of particles, water may prevent the development of favourable internal
pre5tressing and wedging. The effect of topography has been illustrated in Fig. 2/9
a nd it was shown that no uniform pressure distribution can be expected above
tunnels under ridges even when driven perpendicular to the strike_:>.s Special

3·8 EXNER, F. M.: Ober den Druck von Sandkugeln, Sitzungsber. Ak. Wiss. Wien , Abt. 2a
133 7:....8
 158 ANALYSIS OF LOADS

J. • 0, 0-!4 !, O·Z, Off, 0-5, ind 10

m ·J+1-oo, 8-f, 6<J, 4-J, J<J ,ind 10

Fm. 3/ 16. Stress isobars around a

circular hole plotted on the basis of Mohr's rupture condition

attention should be devoted in this respect to the tunnel portals which may carry
increased loads where the layers follow an arch-like (anticline) pattern (cf. Figs
2/9 and 2/32). The previously mentioned creep pressures (slow sliding of the hangs)
may also act on the portals. Owing, furthermore, to the smaller overburden depth,
the development of the protective zone is less complete and extends over a longer
period of time than at greater depths. Consequently in pseudosolid rocks the full
geostatic pressure may act on the portals, while further in this may be replaced
by loosening pressure.
As already mentioned during the discussion of rock pressures, the eventual
pressure on the cavity will be affected by the shape, width and height of the
cavity, as well as by the method of excavation and installation time of supports.
This influence is due not only to differences in the position of the structural axis
which govern the load pattern, but also to the interdependence between the
intensity and distribution pattern of pressures and the shape of the tunnel.
 VERTICAL PRESSURliS 159

3.2. DETERMIN ATION OF VERTICAL ROCK PRESSURE S.

ROCK PRESSURE THEORIES

After the general discussion and introduction of rock pressure phenomena the
preliminary determinatio n of the magnitude of the secondary rock pressures
which the tunnel lining must be designed to resist must be dealt with. The first
problem for any particular case is to ascertain the type of rock pressure likely to
occur during tunnelling operations on the basis of the_geological profile and with
a knowledge of the physical properties and strength of lhe rocks encountered
(primary pressures). It may be concluded from the foregoing considerations that
only a very approximate estimate of loads can be expecte·d. Prediction of the
magnitude of genuine mountain pressure is the least reliable, and its occurrence
1s basically a geological problem. The same applies to swelling pressure, although
m this latter case the role of soil mechanics also becomes important. The deter-
mination of creep pressures is also largely a geological problem, in spite of the
landslide and slope ·stability aspects involved.
The magnitude of secondary rock pressure which will actually act upon the
tunnel lining after the development of the ·protective zone, obviously varies between
t"'o limit values; one of these being the full geostatic pressure, the other the case
cf zero load encountered in very solid and strong rocks which are capable of
carrying the loads without plastic deformation and popping by mere internal
redistributio n of stresses without showing any external evidence.
When dealing with the secondary pressures responsible for the dimensioning
of tunnel sections, vertical, lateral and invert bottom pressures should be discuss-
ed separately.

3.21. ESTIMATES AND APPROXIMA TE METHODS

BASED ON THE EXTENT OF UPBREAK

First of all some practical values are given in the literature which indicate,
i thin rather wide limits, the expectable vertical, bottom and lateral pressures
a a function of the rock material.
The values given in Table 3/1 are based on observations made at the failure
of timber supports.:i.9
For the estimation of vertical rock loads on the basis of past experience the
following hints are given by TERZAGHI:

M Cf. BENDEL: lngenieurgeo/ogie. II Table 202

Observed Rock-Pressure Values TABLE 3/ (

Lateral pressure
Roof pressure Bottom
I I Temporary timber support
Pv (t/ m') PII = 2p• -3p• pres-
R ock material R e mark

At out-
break
After com- 1 Initial
pletion of
drift
( / ')
t m
I
Outbreak
completed
(t/m ' )
sure

(t/m') Mode
of execution
Degree
o f stressing

Rock, more or less blocky 0 8-12 - - - Skeleton 0 to in- Loosening pressure small
Jagging, significant
light

Very seamy rock, cemented 10 30-35 - 3 4-6 Skeleton Small Loosening pressure increasing z>
conglomerate, soft rock, lagging, at the moment of outbreak
>
with small overburden height solid not perceivable ~
~
Heavily fractured rock 15- 25 30- 40 5-10 5-15 10 Tight, Mean Bigger pressures perceivable
(roof breakdown), rolling
gravel and conglomerate
strong
lagging
simultaneously with out-
break. Ensuing of equi- ~
"'
librium, condition, very
prolongated

Loose rock under heavy pres- 25- 35 40-60 10 10 15 Very tight, Con- Stabilization of pressure
sure (eventually in saturated solid siderable conditions very difficult
condition). Bigger over- lagging
burden height

Loose and soft (pseudosolid) 40-60 100-150 20 15 30 Very tigl;it, Going up Stabilization possible only
rock under heavy pressure. lagging and to after the completion of very
Very big overburden height. strong rupture protracted deformations
hard-wood (months even years;
sill-beams Karawanken tunnel)
 VERTICAL PRESSURES 161

(a) In solid and horizontally stratified rocks the greatest load results from the
largest upbreak likely to develop when the unsupported rock is blasted. In the
case of horizontal stratification this height (cf. Fig. 3/4a) will be equal to half
the tunnel width: p = 0·5 by. In the case of vertical stratification, p = 0·25 by
(cf. Fig. 3/4b). In the case of inclined stratification a lateral pressure corresponding
to the direction in which the layers are inclined will also occur and its horizontal
compone nt will induce bending of the side wall (Fig. 3/ 17). The height of the
vertical dropping wedge is again 0·25 b.

I,

FIG. 3/17. Rock pressure

increasing effect of outward inclined stratificati on (RABCEWI CZ)

(b) In moderately fissured, uniform solid rock the height of the vertical dropping
wedge whose weight constitutes the load on the supports may again be assumed
as 0·25 b.
(c) In fractured rocks such as are common, e.g. in the vicinity of faults, where
the extent of fracturing may vary from intense jointing to pulverization in rocks
which are otherwise sound and unweathered, the load on .the roof becomes
- beyond a certain height - independent of overburden depth and depends solely
on the width and height of the excavation. Friction between individual particles
leads to the development of "arching action", which extends generally to a height
of 1·5 B, where B = b + m (cf. Fig. 3/5). A very slight downward movement is
sufficient to reduce the roof load appreciably below the weight of the soil arch
(hp min)- Upon further movement the load again increases slightly (hp max) but
remains still considerably less than the weight of the ground arch. In fact, the load
can be expressed in the form hp = (1. B. According to TERZAGHI the value of hp
depends on whether the layers are above, or below the water table, and on whether

11 Szechy: The Art of Tunnelling

 162 ANALYSIS OF LOADS

the sand is in a dense, or loose condition. The corresp onding values

of or: are
summarized in the following table:

Condition
I
hp min
I hp max
I
B
hpmln
I hpma,.

Dense sand Initial 0·27 0·60 0·54 1'20

(lab. expt)
Final 0·31 0·69 0·62 J-38

Loose sand Initial 0·47 0·60 0·94 1·20

(lab. expt)
h u·d 0·54 0·69 J·08
i J-38
Moderately fractured Increases from 0 to 0·35
Based on observations
Highly crushed in railway tunnels
and mixed Increases from 0·60 to 1·10

(The assumed vertical movem ent in the case of dense sand was initially
0·09 B,
in the final conditi on 0· 15 B; in the case of loose sand initially 0·02
B and in the
final conditi on 0· 15 B.)
In practice, actual roccf loads in sand soils and fractur ed rocks were
found to
be much nearer to the minimu m initial values than to the highest
final ones,
indicat ing that the slight displacements assumed are sufficient to produc
e arching.
An intensified kind of a1.:ching can be t~bserved at the working face of
the tunnel.
Here the rock above the roof is suppor ted at the face as well as the
sides and the
loads are carried by the resulting half dome, which has a higher load-be
aring capac-
ity (cf. Section 61.2, full-face excavation method).
The roof load in fractured rocks may vary between very wide limits
and may
range from the weight corresponding to the afore-mentione d wedge
of height
0·25 Bas a lower limit to several times this value. The actual load
will be affected
not only by the degree of fracturing but also by the moisture conten
t and surface
percola tion conditions. The time of suppor t and the density of back-p
acking may
also play an import ant part (cf. Sections 6.1 and 6.2).
For soils devoid of interna l cohesion, but displaying considerable
internal
friction (dry sand, gritstone, crushed rock, rubble, etc.), an approx
im,.te formula
was propos ed by BIERBAUMER. The roof load derived by taking into
accoun t the
droppin g wedge i.5 accordingly (cf. Fig. 3/2):

p' =: b2 cotan </>, {3.11)

 VERTICAL PRESSURES 163

where b = the width of the cavity

</J = the angle of internal friction
y = the bulk density of the material.

A conspicuous shortcoming of this formula is that for </J = 0, p = oo, which

1s obviously impossible, inasmuch as the highest theoretically possible load is
the geostatic pressure p = yH. The obvious validity limit of the formula is
defi ned by H = ~ cotan </J, i.e. where the wedge emerges to the surface. Actually

surface subsidence occurs x metres earlier so that the limit is H = x + i cotan </J,
\\ here x varies from 5 to 10 m.
The height of the expectable upbreak and hence the shape and weight of the
rock mass loading upon the tunnel lining were determined somewhat more exactly
by SZECHY3·oa on the basis of a statical analogy. His basic assumption is that it
1s the tensile strength of the rock itself which will terminate the upbreak above
a r~tangular cavity. The developing girder-form in the rock will be a three-
hinged arch (Fig. 3/18a) subject mainly to compression stresses. The two lower
hio~!s may be assumed sidewards from the cavity in the centre of gravity of the
additional stress area at roof level resulting from the redistribution of stresses
(Fig. 3/ 18b, after WILLMANN; cf. Fig. 3/6b), whereas the upper plastic hinge will
develop in the centre line at a height where the stresses set up by the overbridging
action will no longer exceed the strength of the rock material. The mutual support
between the two half arches established in this central hinge will transmit compres-
sion stresses ·and because of the relatively high compression strength of all rock
materials its location will not be defined by these, but by the tensile stresses.
For the determination of these, SzECHY assumes that the rock mass overbridging
the cavity roof acts as a quadrangular disc-like beam. It is known from the theory
Jc\-eloped by D1scHINGER3 ·9b that beyond a certain ratio of girder depth (d)
,ersus span (when 1;;?: 1) any further increase of the depth does not affect the
dbtribution and magnitude of inner stresses. Therefore it is justifiable to assume
2n inner stress distribution as shown in Fig. 3J18c indicati~g the flow of compres-
sion stres["!S above and that of tensile stresses below the neutral axis, considering
hat in a homogeneous solid rock mass no change is to be expected either in the
trength properties, or in loading. The only factor which leads to a gradual
decrease of inner stresses is the decrease of effective span due to progressive arching.
Let us denote the measure of this span diminution by the factor 'Cl (where ex < 1).

.,.. Sztcttv, K.: Angenaherte Bestimmung des Gebirgsdruckes auf Grund einer statischen
'1lalogie, Proc. Int. Conf Soil Mech 1963, Budapest 521
SztcHY. C: Approximate . Determination of Rock Pressure on the Basis of a Statical
am.logy. lnt. Symposinm on Rock M-~chanics Lisbo,n 1966
" ' DISCHJ«GER, F. : Beitrag zur Theorie der Halbscheibe und des wandartigen Tragers,
A.Hand/. der j_ V. B. H. Ztirich, 1932

11 .
164 hNAI.YSIS OF LOADS

FIG. 3/Iea Upbreak of the roof in

analogy with a three-hinged
arch (SZECHY)

A 8

FIG. 3/18b Determination

of fhe assumable location of
lateral plastic hinges

FIG. 3/18c Assumable rock pressure

diagram based on the statical
anakgy (SztcHY)
 - - _ - - = --

V ERTICAL PRESSURES 165

The assumed stress diagram in the unknown limit height h will determine the
location of the line of action a. lying in the centre of gravity C of inner compression
stresses N and defined actually by the condition that the maximum tensile stress
at the bottom of this cross-section should be exactly equal to the tensile sterngth
of the rock mat~rial.
lhe location of the two bottom hinges may be assumed, as mentioned, at the
line of action of superimposed vertical stresses. The extension (b') of this stress
zone will depend upon the e1astic and/or plastic properties of the rock material
and may be expressed, after WILLMANN, as a function of the width (b0 ) of the cavity
as b = pb0 and its intensity will decrease according to a parabolic law (see
Fig. 3/ 18b). The value of the factor fJ may be taken between 1/2 and 1/3 for plastic
materials and bet~een 2 and 3 for elastic ones. (For lack of more exact investi-
gations it may be assumed that fJ = I in solid rock and fJ = 2·5 in plastic rock
and soils.)
Based on the investigations of DISCHINGER as to the inner stress distribution in
a quadrangular disc-like beam and-as to its correlation with a uniformly distri-
buted external loading (p), SzECHY concludes finally that if the decrease of effective
span is assumed to follow a linear rnle in correspondence with the decrease in the
distance between the reaction forces of the assumed three-hinged arch (i.e.
:z = 0·5) then we obtain for the height of the upbreaking zone:

h .= [1·13(1 + /J/2)2 - 0·5] b0 , (3.12a)

'" hich indicates that, ·besides the width (b0) of the rectangular cavity', it will be
a function of the fJ value characterizing the rock-plasticity (in conclusion the
increase· of active span). Should the disc-like beam be regarded as not fixed but
imply supported at both ends (e.g. in loose soil), then the above equation will
have the form
h = f1 ·31(1 + /3/2)2 - 0·39] b0 • (3.12b)

In conclusion the rock pressure acting upon the cavity may be assumed from
the determined upbreak limit as the triangular zone defined by the location of the
binges of the three-hinged arch. Fig. 3/18d shows this pressure area which repre-

I7
G
vq
FIG. 3/ 18d Loading estimate for an inclined
stratification (TERZAGHI)
166 ANALYSIS OF LOADS

sents the ioading which is decisive for dimensioning the lining of the cavity. This
loading, naturally, constitutes the maximum, developing when no inner temporary
support is applied and the upreaking process is entirely defined by the strength of
the rock material itself.
The application of any temporary inner strutting would, naturally, reduce the
overbridging period and so also the extent of the upbreaking process.

3.22. THEORIES BASED ON THEORETICAL STRESS CONDITIONS IN

THE ROCK MASS

In one group of these methods the theory of elasticity is applied to identify

first the stress conditions prevailing in the rock in its original undisturbed condition
(Fig. 3/ 19a), and then the distribution of stresses and strains disturbed by the
excavation of the tunnel and developing around the excavated cavity are derived
which, at the same ti.me, define the pressures (stresses) acting on the tunnel section .
The calculation of stresses within the material and the concept of Airy's stress
function are assumed to be known from the general theory on the strength of
materials.
In the cylindrical coordinates commonly used in the calculation of rock pressures
on underground spaces the stresses
., arvz A
lgz-,.-:,;-::-UL 0'f.
are obtained from Airy's stress func-
tion (<P) in the following form:
>OZ t. + _E dz
I 'J/l OZ
1 a 2 <1> 1 a<1>
(1
,
=r -
2 -
a0-2 + ,.--or- '

and (3.13)

a { 1 o<P )
'= - or [7ao ·

Fm. 3/ 19a Relation between stresses,

deformations and distortion
on the unit cube cut out
from the r.ock: mass
 - - -

THEORIES BASED ON TIIEORETICAL STRESS CONDITIONS 167

32.21. Theory of Elasticity and Estimation of Stresses

The first theory based on the above fundamental relationships of the theory
of elasticity and permitting the estimation of stresses within the rock both in the
original undisturbed condition and after the disturbance caused by excavation
was developed by SCHMIEo. 3• 10•
In this theory SCHMIED also considers the effect of mass forces (rock weight).
Thus, the general stress relationships for the two-dimensional problem, where no
component acting in the direction of tunnel axis is involved, are

(3.14)
,h;lr 1 ou, 2 .,, 0
--+---+
or r o0
-+T=
r
'

where R = y sin 0 and T = y cos 0 are the mass forces.

Introducing Airy's stress function ( <I>) the stresses will be ·obtained in the form

1 o<I> l o2q> •
u = - - - +-2 --
2 + 2yr sm 0
r r or r 00 '

02 <P
u, = - ~ + 4yrsin 0, (3.15)

•,: = •rr= - :r (+ ~! )-yrcos0.

In addition, it must be checked whether these stresses correspond with the actual
deformation conditions. To do so they must satisfy the following compatibility
equatio~:

(3.16)

The stress function derived by SCHMIED for the undisturbed rock (primary stress condition) is
3
<P =- - r + -·-
s1H 2
r cos 20 + -r
SoH 2 (
s1 -
So
_:::_ -
) •
4y sm () + -·
So
r 3 , ()
sm , (3.17)
2 2 6 2 12
where
H = the overburden depth
s = a factor expressing the effect of sloping terrain surface in the cas:! of hang tunnels
(for tunnels at.considerable depth s = 0)
,.ioa SCHMIED, J.: Statische Probleme des Tunnel- und Druckstollenbaues. J. Springer, 1926
168 ANALYSIS OF LOADS

Si=-
y (1
2
+---
µ- 1
1-s)
and

are abbreviated notations.

The stresses in the undisturbed rock are obtained by differentation a3

a~= s H + (s + i) r sin 8 + s
1 1 2 H cos 20 - i r sin 3~,

a0 =s1 H+ (s 1 --1)rsin8-s2Hcos28+ i rsin38, (3.18)

0 S2 . Sz
1:,1 =
2 r cos 8 - s2H sm 28 - 2 r cos 38 .

The stress function developed for the rock masses disturbed by the excavation of the tunnel
(secondary stress condition) is of a more involved form:

1 1
<1>' =a0 1n r + b0 r2 - -~ r8 cos8 + (d1 r 3 + : + d 1r In r) sin 8 +

+ ( a2 r • + a.,
,; + fl 2 ) cos 28 + ( Ca r3 + 7s
Y3 + -;-
d 3 ) sm
•
38, (3.19)

where the constants a0 , b0 , Ci, di, y 1 , d1 , a 2, a 2, fl:, Ca, d3, and y 3 have to be determined from
the boundary conditions by lengthy and tedious calculations.
The stress equations, accordingly, will be much more involved:

a,= -;;-
r- 1
{Cl
ao - s H + -. (1 + m) + s -
1
r·
-
S2
2
-
~11 .
2 -r. r sm 8 +

+ s.H-6 (
-a: -4 __::_ cos28- {Sz Ya d3}rsm3
-+12-+10- . 8 fl.1
• t
r• ,2 , 2 r6 r' '

+ - { S2H +7 6a 2}
cos 28 + (Si Ya
T + 12 -;s + 2~ , d3} . 38 .
SID

-rr1 = - {m -C1 Sz 2y1}

,z - - 2 - -r•- r cos 8 -
a-2 + 2 -flz } sm28-
. ' {-
s~- 1 2y-3
- 6d-
3}
(3.20)
-
{s 2 H+ 6rt r2 2 r6 r•
rcos38.
1
 THEORIES RASEU ON THEORETICAL STRESS CONDITIONS 169

The problem may thus be regarded as solved in principle, but the computation required
for the evaluation of the great number of constants from the boundary conditions is enormous
and disproportiona te to the results obtained. In fact, there is a fundamental discrepancy
between the basic assu mptions introduced (elastic, isotropic, uniform, homogeneous medium)
and actual conditions and even the assumption of boundary conditions is rather arbitrary.
For this reaso n the method may be regarded as suitable for demonstrating the influence of
individual factors, e.g. that the magnitude of tensile stresses above the crown is decisively
controlled by Poisson's number. Tensile stresses develop above the roof and below the bottom,
but compressive stresses affect the sidewalls. A conclusion of great significance is the theoretical
demonstration of the fact that the elastic range affected by the distu rbin2 effect of the tunnel
is rather na rrow. This distance is, according to SCHMIED,

m~ (m - 1) £ t j (3.21)
R = r:r ( -y(m+l}(m -2)H '

where £ = the modulus of elasticity

m = Poisson's number of the rock.
SCHMIED has also developed a solution for including the strength parameters of the artificial
support of a circular tunnel lining in the stress functi on, which renders calculations, naturall).
even more complicated.
SzllNAGYI also gives a solution for the determination of the tangent ial and radial stresses
arou nd a circular cavity resu lting from the disturbance of the stress field owing to the estab-
lishment of the cavity in an elastic medium.3·10b
The solution is based on AIRY's stress function as given by TIMOSHENKO. He clearly demon-
strates the effects of tunnel linings of varying flexibility upon the distribution of stresses and
pressures around the cavity.
He concludes, in general , that stress distribution will be rendered much more favourable
even with a quite fle xible lining, and will be made quite uniform by a lining of mediu m
rigidity.
Tension stresses at the crown and bottom will disappear and the disturbance of the original
stress field will be reduced to zero at a lateral distance equal to the diameter of the cavity.
The co nstructi on o f the lining will counteract deformations, but prior to its installation
deformations will be already initiated which may go on as a consequence of defective embed-
di ng. To allow for this action a factor '.X is introduced expressing the rigidity of encasement.
:x I expresses an absolutely tight embedding, whereas '.X ,.., 0 expresses free deformation.
The construction of the lining will increase radial stresses so exerting a beneficial :nfluencc
upon rock strength. whereas the dominant tangential stresses wi ll decrease even with a small
,a Jue of :i: . ln addition, the distribution of stresses will be rendered more uniform as '.X increases
producing a nearly uniform distributio n when ,:i = 0·5. With the increase of '.X , stresses on
the top will increase as against those at the further hor izontal diameter.
The distribution of pressure will become more uniform corresponding to the distance
from the cavity (Fig. 3/ 19b).
Omitting the detailed derivation and with reference to the literature the formulae derived
by CAQUOT and K t RISEL3 •11 fo r rock pressures acting on a circular cavity are also deemed
\\Orthy of mention. Radial rock pressure in a plastic state of equilibrium is, accordingly,

P, = "/I' ( I - cos 0)
••H
+ T: _
[. r (Hr)·,..-,_I,
H - (3.22a)
2

,,,,,b Sz1LVAGYI, l. : Die Bestimmung der Spannungen um einen kreisformigen

Tunnel,
Proc. Jnr . Conf Soil M ech. 1963, Budapest 533
11
CAQUOT and K ERISEL : Traite de mecanique des sols. Dunod, Paris I 956 475
 170 ANALYSIS OF LOADS

a: • 0·7 a-o';J· a .o,g

r -4
a•Z

r•S
a •2·5

r•6
a-J

FIG. 3/19b Variation of pressure distribution around a circular cavity as a function of the
rigidity of embedment a and of the distance r

where r = the radius of the circular cross-section

H = the overburden depth
y = the bulk density of the rock
J., = the coeffident of passive earth pressure
(J = the angle included between the vertical and the radius drawn to the circumferen-
tial point under consideration.

The above formula for crown pressures was presented by K:ERISEL in a lightly modified form
at the Fourth International Congress on Soil Mechanics and Foundation Engineering in
London, 1957:

P, = yH{- 1
}., - 2
[.!_ - (.!._);.']-
H H
_c
yH
. _l [1 - (.!._),."]}.
tan ~ .H
(3.22b)

The theory is claimed to be applicable for the prediction of pressures on flexible pipes
buried at moderate depths, but there is, as yet, no conclusive evidence available to confirm
this. An interesting feature, however, is that radial pressures are assumed.
 ONS 171
THEORICS BASED ON THEORET ICAL STRESS CONDITI

32.22. Results of Jm:estiqations by Fenner

More practic able resu lts have been develop ed by F ENNER3 -1 2 , also on the basi)
of inner stress-c onditio ns.
to KIRSCH, in terms of
In perfectly elastic media the stresses are obtained , accordin g
provided that nowhere do they exceed the proporti onality limit, from
the vertical pressure p,
the following expressi ons:

a,~ __ P__ __ [ m+ (m - 2) cos 2:x],

2(m - 1)

_
_ _P - [ 111 - (111 - 2) COS 2:x], (3.23)
a,= I)
2(111 -

p(m - 2)
-c = - - - - s i n2:x,
2(m - 1)

of the stress a,, while

\\ here :x = the angle included between the vertical and the direction
and minor axes of the stress ellipsoid (a,, ax, ay) are related to the vertical pressure a,
the major
p
a. = p; ax = a,. = - -- . (3. 24)
111- I

with the help of Mohr's

In plastic media the normal and shear stresses can be expressed
l1rcle in terms of the principa l st resses as

a 1 +a .• a,-u ..
"• - - -- : .
2
+- 2
- - cos 2x,

•
T = -<l1 -
2 -<l:: sm2x ,

of the principal stresses:

-u.hile the state of limit equilibr ium is expressed by the following ratio
+
1 sin <I> (3.25)
I - sin q, ·

nd.
In the elastic state the ratio of principal stresses is, on the other ha
a, (cf. 3.24) (3.26)
- = m-1.
<l2

In ~ohesionless material normal stresses and shears are related as

f 1938 68 I
:a,,: R.: Untersu chungen zur Erkennt nis des Gebirgs druckes , G/iickau
FE'l'iER,
Pressure s (in English) TTS 15 NRC Div. of Bidding
FENNER R. : Study of Ground
Rtstarch Ottawa
172 ANALYSIS OF LOADS

where µ 1 is the coefficient of internal friction. Lateral displace

ment is characte rized on the
basis of the relations hips (3.26) and (3.25) by the inequali ty:

~ = 1 + sin tf, c!:: m _

a: 1 - sin tf, <: 1 (3.27)
·

The upper sign applies to the case where the shear T is greater than
the frictiona l resistance:
µ 1an and the particles will display relative displace ment along the
shear planes inclined at
90° - tf,, as long as the limit conditio n Eq. 3.25 is satisfied. The
equality in Eq. 3.27 should
. 2
then be equated to K - 1, i.e. K = m, where K = ·
. , and then on substitution into
J - SID a
Eq. (3.23) the following relationships will be obtained :

a, = -2-( K- ~- 1-) [ K +(K - 2) cos h] ' l

a, = i(K ~ I) [ K - (K - 2) cos 2~)

pK - 2 .
r = - - - sm2~.
I (3.28)

2K - 1

The ratio of principal stresses can further be expressed by the ratio

of friction coefficients as

The rocks as far as practica l construc tion is concern ed display

considerable cohesion ,
consequently both the conditio n of displacement and the ratio of
principal stresses must be
modified accordin g to Fig. 3/20. Thus,

T = c + Un tan tf, ,
= f-a ,+ a..
-2-:._ + c cotan tf, ) si n tf, = a 1 + a: sin tf, + c cotan ef, •
2

r;,

o;,

Fie. 3/20. Stress display

at rupture
conditio n
(after FENNER)
 THEORIES BASED ON THEORETICAL STRESS CONDITIONS 173

With the notation

a -+
= -01 -
2
- + c cotan '+'"- '
02

we have
o, + o 2
___ _;;_ = a - .c cotan <P •
2

The stresses in undisturbed rock can now be written logically in the form:

a, = o(I + sin <P cos 2-x) - c cotan <P ,

l

and in disturbed rock as:

o, = a(I - sin <P cos 2-x) - c cotan <P,

.- = a sin <P sin 2-x , l (3.29)

01 = o(l + sin <P cos 2-x) - c cotan <P ,

i
ax = o(l - sin <P cos 2-x) -

.- = a sin <P sin 2-:r •

c cotan <P,

Considering the actual shear strength and other physical properties of rocks encountered in
I {3.30)

practice the respective depths to which Eqs 3.23, 3.28 and 3.29 are valid can be determined:
I. In igneous and high-strength rocks (granite, basalt, gneiss)

.- = 200 kg/cm', y = 2·7 t/ m", m = 0·5 and µ 1 = 0·75 .

From the third relationship of Eq. 3.23 the critical depth at which the ultimate shear strength
1 attained (-:r = 45 °) can be determined:

2·7 h 5 - 2 . ,,
7.000 t/ m 2 = - - - - - - - S111 90
2 5- 1 '

h = 2000 ~ ~ 1975 m.

llus ,alue indicates that practical tunnel or mining depths are usually less than this limit.
f tt>is reason the stresses should be calculated from Eq. 3.23 according to elasticity theory.
" In sand<tones of medium strength

.- = 400 t/m', y = 2 t/ m3 , m = 5, and µ 1 = 0·7.

Formulae based on elasticity theory yield 533 m, but Eq. 3.29 indicates 1200 mas the critical
depth. It appears, thus, that in these soft rocks the critical depth should be somewhere be-
t,a.-ccn 500 and 1200 m, depending on the magnitude of the smaller principal stress (cf. Section
31.~J. on Poisson's number and on the coefficient of friction.
3. ln clays, marls, etc. T = 10 t/m2, µ 1 = 0·36, m = 5, y = 2 t/ m3 . The yield point is
a:re.ad) attained at a few metres depth and thus the stresses have to be calculated, in general,
rding to Eq. 3.28.
 174 ANALYSIS OF LOADS

After the above general introductory consideratio ns the state of stress developing
in an elastic rock around the cavity is derived first by FENNER. The derivation is
based on the stress (AIRY) function of FoPPL

m P m - 2 (r 2 _ a2)2
er,= -P - - - ( r 2 - 2a2 lnr)- - -- -· 2 cos2cc. 3.31)
4m-1 4m-1 r

From the foregoing exposition the stresst!s can be obtained directly from the
stress function IP:

1 o2 ~ I 011' p r2 _ a2 -,
(l
r
= - -2
r
- - + -r
• 002
-- - - . -m m
. or - 2 --
- l
. r2 +

+ -p2 · -
m- 2 (
- l - -4a2 + -3a4-
m - l , r r
2 4
I
cos20
'

(l = a21P
or 2
= !!.... • _
2 m- l
m_ . r2 + a2 - !!.... . m - 2
2
r 2
(1 + 3a4.) cos 20, (3.32)
I m- I r4
I p p m - 2 2a2
o·z = - (a,+ a,)= - - - - - · - - - · - 2- cos20 ,
m m-1 m m- I r
t = __!__ (_!__.
or r
o!/1)= !!.... • m -
80 2 m - I
2 (- 1 - 2a2
r2
+ 3a4) sin20.
r4
j
By plotting the principal stresses around a cylindrical cavity on the basis of
the known relationship between principal stresses

<T':,'/:. = a,+2 a, ± 2I J 4t 2
+ (a,- a,)2 ,

the diagram shown in Fig. 3/21 is obtained. It is to be seen that the rock fails
to resist at the boundary the tensile stress 0·25 p even at moderate depths. Con-
sequently, the above relationships have significance only in rocks of high tensile
strength. Owing, however, to the'various tectonic movements to which mountain-
forming rocks have been subjected, their tensile strength has been greatly reduced,
for which reason the above formulae are unsuitable for practical use.
The optimum tunnel cross-section for minimum tensile stress is then developed
by FENNER, as an ellipse the major axis of which is vertical and the axes are related
I
to each other as bfa = - -- . In this case the tangential normal stress will be
m- 1
mp
(l
1
=- --
m- l '

-
 THEORIES BASED ON THEORETICAL STRESS CONDITIONS
175

\
ff) 10

FIG. 3/2.1. Stress isobars

around a circular hole in elastic rock as a function of the rock pressure p

which would completely prevent sliding even of the fractured layers. This relation-
ship would lead again to a shear resistance high enough to anticipate the formation
of a protective zone. Since this zone can actually be observed in practice, this is
again a condition which can be regarded as a theoretical limit only. It is reasoned
hereafter that if large tensile stresses occur in the crown of circular cross-section s
and large compressive stresses in that of a standing ellipse with an axis ratio
1
- - - , thet ,' is bound to be a shape for which the stress in the crown will be zero.
m- 1
. 2
This shape is developed as a vertical ellipse having an axis ratio bfa = - - -
2
m-
(Cf. Section 22.31). The isobars of maximum and m.inimum stresses around a tunnel
shaped accordingly are shown in Fig. 3/22.
In this theoretical approach the effect of dead weight was neglected by FENNER.
An even more serious deviation from reality is inherent in his foregoing expositions
 176 AI-IALYSIS OF LOADS

8:b • (tn-2),2

10 HJ
I

0-2

FIG. 3/ 22. Isobars of horizont al and ,,f vertical

stresses around an elliptical hcle

by the assump tion of a uniform rock which follows HOOKE's law. The
relationships
derived on the basis of the elasticity theory are, nevertheless, suitable
for the
theoret ical verification of the develop ment of the stress-free body
around the
cavity and that the pressur es should be carried entirely by the rock
around the
cavity. This overloa ded range is shown theoretically to be limited
in extent from
four to five times the major half axis a, in spite of the high stress
peaks in the
crown and at the springi ngs.
Greate r practic al import ance may be claimed for FENNER's theoret
ical investi-
gations concern ing rock pressures developing in the plastic state.
The magnit ude of rock pressures acting on a vertical shaft is investig
ated first
and is found to be uniformly ~istribu ted at a given depth around
the circular
perimeter. It is assumed that rock can freely intrude into the interio
r of the shaft
throug h openings. Since the displacements are directed radially
inward and as
meanwhile no relative displacement between individual pa_rticles can
take place,
the shear stress is zero and the radial directio n as well as the tangent
ial direction
perpen dicular thereto will constit ute the directions of princip al
stresses. The
direction of sliding surface s ,;an thus be determ ined from the conditi
on that these

- - - - -- -- - -
 - - - -- - = -- ~ - - - --
- - -· - - - - ~

THEORIES BASED ON THEOREllCAL STRESS CONDITIONS

must intersect the principal directions at a constant angle. If it is taken foto con-
sideration that the radial stresses u, decrease as a result of radial displacement
while the shaft is being sunk, and the circumferential stresses u, increase as a
1esult of being pressed together, it will be readily appreciated that no equilibrium
can take place between the principal stresses uR and Ur unless Ur assumes a maxi-
mum value which occurs when

(3.33)

Conversely, the value of uR cannot be smaller than that determined above, other-
wise the material would flow into the cavity. The problem, once the above rela-
tionship is established, is to determine the value of uR. In developing the differen-
tial equation it may be noted that owing to axial symmetry the stresses become
independent of 0 so that this may be neglected in the calculations. Thus,

1 a<P a2 <T>
uR = - - - and ur = - -
ar ar ar ,
also
r = 0.
On substituting these values into Eq. 3.33 we obtain the differential equation

a2 <P 1 o<P
- -(K-1)- - = 0
a, 2 r ar '
the solution of which is
C
q, = - rK,
K
whence
Ui- = C(K - 1) ,K-2.

For evaluating the constant C, let us consider the boundary condition at the shaft
wall, i.~. for r = a; uR = uRa and thus C = uRa aK- 2. Substituting this value we
obtain
K- ~
UR= URa
(
:
)
- ; (3.34a)

If K = I + (µ 1 + .j 1 + µi) 2 and the specific friction µ 1 = tan </>, then for

q, = 10°, at µ 1 = 0·36, we have K = 3 and
r
<Tr= 2uRa - ,
a

1 e. th<: stresses increase proportionally to the distance from the shaft of radius a.

I :! Sz~by; The Art o f Tunnelling

 =----= --

178 ANALYSIS OF LOADS

Since the rocks encountered in practice may safely be assumed to rossess an

internal friction of at least ¢ = 20°, the rock is bound to reach a state of equilib-
rium around cavities of any arbitrary shape. The stresses arising at the boundary
of the plastic region are obtained from the condition that beyond this boundary
the fundamental relationships of elasticity theory regain their validity. The sum
of the two principal stresses must be constant, i.e. O', + 0'1 = 2 p, which when
combined with the relationship of plasticity O'r = O'R(K - I) yields O'R +
+ O'R(K - I) = 2 p; p = O'R = 2 p/ K. Denoting the distance of the boundary of
the plastic zone by b and taking Eq. 3.34a into consideration, the condition at
this boundary may be written as

}!!_ _
- O'Ra
(!!_) K-2
,
K a

whence b can be calculated (cf. Eq. 3.21).

It can be seen that b varies inversely with O'Ra, the latter being the radial stress
at the perimeter of the shaft. The magnitude of this stress is, in turn, governed
essentially by the radial support of the cavity, i.e. by the strength of timbering.
Consequently, the stronger the supports installed, the smaller the extension of
the plastic region will be.
Beyond the plastic region the rock material is again in an elastic stress condition
and the stresses can be calculated there from the following relationships :

To illustrate the method the rock pressures acting on the wall of a circular
shaft at h = 1000 m depth in a soil having a bulk density of y = 2·4 ton/m3
have been calculated by FENNER for O'Ra = 1. The geostatic pressure y h = 2400
ton/m2 = 240 kg/cm2• The numerical results obtained are shown in the table
below and in Fig. 3/23.

Coefficient of friction Ratio Exponent Lateral

Boundary of the pressure p
µ, K-1 K-2 plastic region b
(kg/cm' )

0·75 (36° 52') 4 3 2·8845 a 60

0·5773 (30°) 3 2 6·3246a 80
0·35354 (19° 29') 2 J 80·0a 120

As can be seen, the distance of the boundary of the plastic region increased
as a function of the coefficient (angle) of friction from 2-8845, a to 80,0 .;,
 THEORTES BASED ON THEORETICAL STRESS CONDITIONS 179

~ 120
~
~
100

~ ~~--P.'1=t:tf":"ii:;o:!3-:=~~~~=~~==
60~~,;t--,-- t--::z.~==~==~=~-......=--
40

20 40 60 '··_g1 ' 100 •1io ' 140 · tio ·1io ·200 •i20 '24/Ja I'
t----- b-----l
F• ,. 3.'23. Propagation of tangential and radial stresses around a hole
as a function of soil quality µ 1

bile the angle decreased from 37° to 19-5°. This considerable range is evi-
dence that a very long time is required for the development of calculated
tresses and deformations in clay, clayey silt and clayey sand soils.
The value of u Ra can never be predicted in practice since at low friction values
s depends on the quantity of material removed from the cavity to which it is
\ ersely proportional, (cf. genuine mountain pressure). In the case of high coef-
Scients of friction, on the other hand, the new stress equilibrium does not take long
· ~ develop; the pressure on the shaft will be very low and independent of the depth
c r overburden.
A plastic zone of great extension also involves deformations of considerable
magnitude (the specific volume change can be expressed as e = (ux + Uy +
2
- o:) : ~ and V1 = V0 e, where V0 denotes the total volume within the plastic
region which is manifested by the quantity of material (V1) entering the cavity
 180 ANALYSIS OF LOADS

and to be removed theref rom. Pressu re intensities have

been found to decrease
rapidl y with the excavated volume, and, as can be seen
from the diagra m for the
above numerical example (Fig. 3/24) the, remov al of 5-10
% of the material results
in a reduc tion of pressures to 1/ 15-1/20 of their origin
al value.

b;
kg/cm'
80
rg
70
68 {iJ

50
~ 1/4 40
] J.'Q \ ,___
C: JO -
.,....
0 JIJ \.
20
::s '\. ...
~
.,...."'"' 10
_i
00
! -- 10
'-..,
/0 20 XI 40 !ii 6/J 'JO !O .W/J OO 10
20 JO 40 50 fj(J "/0 60
Extension radius of plastic zone (m) Volumetric increase (m')
FIG . 3/24. Variati on of pressu res acting upon the lining as
a functio n
(a) of the extens ion of the plastic zone;
(b) of the quanti ty of intrude d soil mass

A simila r appro ach is adopt ed hereaf ter in determ ining

the pressures acting
on a horizo ntal {unnel, but now the difference in loadin
g due to self-weight is
also taken into consid eratio n. The stress functi on in
the elastic zone will now be
of the form <fJ = rK 'E An cos n 0, where the An factor
s repres ent param eters
depen ding on bound ary condi tions.
Assum ing appro priate symm etry condi tions for the plastic
zone the following
simple differential equat ion can be writte n:

(3.35)

Expressing now the stress values by a stress functi

-obtain
on cf:', and with 0 =0 we

1 a<P'
UR= - -- - -
yr a2<P'
r or Ur=~ -r=O . (3.36)
2
 . .

THEORI ES BASED ON THEORETICAL STRESS CONDITIONS 181

Substituting these values into Eq. 3.35 we obtain

o2 <I>'
_ _?_ _ _ __
o<P'- + -K· --yr=O
K - I - - 1
or r or 2

a nd from this after double integration

C K.- 1
<I>' = _ ,-K + - ~ - - yrs.
K 6(K - 3)
Determining the integration constant C, again at r = a, from the condition
<1R= <1Ra and expressing the stresses according to Eq. 3.36, we have

0 2<J>' (r)K- 2-(K-l)K_ (r)K-

<1, =~=(K-I)e1Ra~
ya
3
a 2+

In view of the three-dimensional stress condition Eq. 3.34a is modified to the

for m
<1 = (<1 - _ Y_<l ) (!_)2K-2+ _ yr _
a 2K - 5 a 2K - 5 '

whence with µ 1 = 0·35354 as the lowest value we have

.,here the sign of the second term in brackets depends on the value of 0; for
,J = 0 the upper, for 0 = n the lower sign being applicable.
At the boundary of the plastic zone the sum of the two principal stresses must
again be equal to the sum of the two original (initial) principal stresses, where-
fro m as an extreme value for r = h/2

O"Ra = ay [++In ;a),

in other words, the condition for any possible equilibrium is that o-Ra should
auain this value at least.
 182 ANALYSIS OF LOADS

In the general case

ya 2- 5K + 7
K [ 2a ]K - 2 (3. 38)
O"Ramin = K-3 - {K- l)·(K- 3) yh (K-2)h

and now the plastic region extends to the height r = K; 2

h, and for K = 4,

O" Ra min = ya ( 1 - ~- ) ·

This immediately suggests the conside ration of the plastic region as a vertical
ellipse with the cavity at its lower focus.
Again for a depth h = 1000 m, a cavity with a rad ius a = 3 m, a bulk density
y = 2·5 t/m3, and friction µ 1 = 0·35354, and for internal support
with strength s
O"Ra = 4·215 kg/cm2 , 4·5 kg/cm 2 , 6 kg/cm 2, 10 kg/cm2
and 30 kg/cm2 respectively,
the ellipses representing the plastic regions and the sums of correspo
nding
principa l stresses are shown in Fig.
3/25. The considerable restricti ng
effect of support strength upon the
extension of the plastic zone is
Extension of clearly illustrated.
plastic zones The case of stratified rocks was
41.KJ also investigated by FENNER , and he
found that a relatively thin clay
layer may conside rably modify
stress conditio ns and will lead to a
significant extensio n of the loosening
core both upward s and down-
wards.
In conclusion he has stated that
an elliptical stress-free zone is de-
veloped in elastic rocks which does
not depend on the overbur den depth
but only on the width b of the cavity
and Poisson's number . lt can be
described by the equatio n
1000m
Ofst,ance
4x2 Y2
(m - 2)2 b2 + b2 = I.

The stresses will be equal to zero at

FIG. 3/25. Extension of plastic zones as a
function of the rigidity of supports the crown and springings of the
around a circular tunnel lying at a ellipse and will attain peaks at their
depth of 1000 m (FENNER) diagonals.

)
 STRESS CONDITIONS 183

On the other hand the magnitude of the stress-relieved zone developing in

plastic rocks depends not only on the width of the cavity but also on the depth,
on the coefficient of friction, on the bulk density and on the pressure supporting
the tunnel lining.
For the radius of the plastic range developing around a cylindrical tunnel the
following formula was derived by RENDULic3 · 13
r
(3.39)
R = (1 - p3)!'
where

P=(1+2~r
In addition, the tangential and radial stresses at the boundary of the plastic range
are given as

.
<Tr =C cotan ¢
[
1-
( c cotan
r +eo
¢) ,..- 1
]
·. (3.40)

32.23. Stress Conditions around Circular and Elliptical Cavities

This problem was discussed by pure elasticity theory by TERZAGHI and RICHART3 • 14 who dealt
not with the actual determination of the magnitude of rock pressures but, taking the magnitude
of vertical (p0) and horizontal pressures as given, investigated the effect of their ratio, and the
~hape of the cavity upon the distortion resulting from pressures around it. The approach
followed in calculating the stresses is to superimpose the values obtained for the uniaxial
~tate of stress produced by the vertical load on those obtained for the biaxial state of stress
produced by the horizontal pressure. Horizontal pressures around a cavity will not develop
unless the horizontal deformation due to the vertical load is prevented. O therwise the only
horizontal pressures to be taken into account would be those of a residual (orogenic) nature.
For the basic determination of stresses around a circular cavity the already mentioned
Rlationship of KrRSCH is applied:

a,= -p2 ( a2 ) + -p
1 - --;;-
(
1
Ja4
+ -- - a2)
4 --;;- cos 20,
r· 2 r 4 r-
4
a 8 = -p ( 1 + --;;-
aZ ) - -p2 (I + -3ar - 4
)
cos 20 , (3.41a)
2 r·
4 2
T, 0 = - -2
p (
\1 - -3ar'.- + -.-
2a
r·
)
sin 20,

$U RENDULIC, L.: Spannungszustand in der Umgebung eines Hohlraumes, Wasserwirr-

schoft 1934 168

a.u TERZAGHI, K . and RICHART, L.: Stresses in Rocks about Cavities, Geotechnique 1952
Ji-75
 184 ANALYS~ OF LOADS

where , 1 and e are polar coordin ates, and for the vertical axis z, e= 0
p = p,. the uniformly distribu ted vertical rock pressure
a = the radius of the cavity
Px= ).p is the uniform ly distribu ted horizon tal rock pressur
e
a, and u 0 = the radial and tangent ial normal stresses
r, 0 = the shear acting in the r - O plane (Fig. 3/26a).

With a lateral pressur e of intensity p ., = ).p, the stresses are to be calcula ted in a co-ordi nate
system transfo rmed by 9 = :r/2

4 2
a, = -pi. ( a!) p). ( 3a 4a )
,- - - 2 1 + --
2 I - --. ,, - -.-
,- cos 28,

a ,0
2
2
= -pi. ( I + :a ) + -pJ. r1 + -3a,- 4
)
cos 2/J, (3.41 b)
r 2 , r

4 2
T;o = -pi. (
I - -3a.- 2a
+ -.- )
sin 2IJ •
2 r- r-
These equatio ns apply to the
stresses develop ing around a circu-
lar hole in a plate. In the case of
a three-d imensio nal tunnel the
stress acting in the directio n of

- the longitu dinal axis will be, in

accorda nce with HOOKE'S law,
<1: = µ (a,+0 0 ).

-- a
--
p•• ). .P,
-- F1G. 3/26a Annota tions used for the
comput ation of stresses around

tttttttfptttrttttt a circular hole

(TERZA GHI- RICHAR T)

b;

X
\
\
'- FIG. 3/26b Coordi nate system used in the
comput ation of stresses
around an elliptical hole
z (TERZA GHI- RICHAR T)
 STRESS CON DITIONS 185

Horizontal and vertical stresses can be obteined by the known stress transformation cor-
relations
a.+2- -
a,. =-~ c;,
-
a 0 - a, .
cos 20 - r, 0 sm 20 ,
2

a6 + a, U8 -
- -
a,
-- cos 20 + r, 6 sin 20 , (3.42)
a,, =
2 2

T vh =- sin 20 + r, 0 cos 20.

The equations for stresses around an elliptical cavity are much more involved. Stresses
around an elliptical hole in a plate were determined by NEUBER by introducing the elliptical
co-ordinates ct and {3 in the following form:
(a) For an ellipse with the major axis horizontal (parallel to the x direction)
x = cosh ct sin {3,
z = sinh ct cos {3 (Fig. 3/26b).
(b) For an ellipse with the major axis vertical (parallel to the z directfon)
Case (i). The equation related to the edges of the cavity:

2 2
1
( sin~ a0 ) + (, cos: ct0 ) = '

and the major, and minor axes are

x,=o = a = cosh a0 ; Zx=o = b = sinh ct0 •

The stresses are:
2

a.
= !!.._ { [ sin 2{3
8 h4
+ 4 cos
h" 2{3] · [ cosh 27 + 1 - 2 Be- 2" - c (I - e-2•) ] +
2
+ si~;. 0: [ 2 sinh 27 (1 - cos 2{3) + 2Ce- 2" (1 + cos 2{3) + 2A - 4Be- 2 • • cos 2{3]}

a, = ~ {5in;,2x [- sinh 2a (1 - cos 2{3) - ce- 2• (1 + cos 2{3) - A+ 2Be- 2"(cos2{3)] +

+ sin~/{] [<- cosh 2a) - 1 + 2Be- 2" + C (l - e- 2") ]+ (3.43)

+ ;2 [ cos 2{3 (- cosh 2a - Ce- 2• + 2Be- 2") + cosh 2a - ce-

2
• ]}

r. = ~ {2 s~~ 2{3 [- 2 sinh 2a - 4Be- 2" + 2Ce- 2e] + sinh~{J -[ sinh 2!X [ 1 + cosh 2a -

- 2Be- 2" - C (1 - e- 2•)] - sinh 2:i - Ce- 2" - A+ co~ 2&(sinh 2-, - Ce- 2• +

+ 2Be- 2
") j},
186 AN ALYSIS OF LOAl>~

where A = - l -- cosh 2'X0

+ -43 - e••,
1
B = - e2•,
2
-4
C = 1 +e 2 0
"

the factor of distortion: h = 2 sinh2 a + 2 cosh 2 {J.

At any point of the elliptical cavity a= a0 and hence [a,.]•=•, = ,; 2 [sinh 2a0 - 1-
- e 2" 0 - cos 2P J at the crown,. wbere P = 0, a_. =- p while at the springings (x axis),
where P= :i/2, a_. = p ( 1 + 2 : ) .

Case (ii). The equation related to the edges of the cavity:

( cos~ a0 -r+ ( . z
smh a0
'J ! = 1 ,

and the major and minor axes are

Zx= o =b= cosh a 0 and x,=o = a = sinh ,x0 •

The stresses are given as

2
+ sinh 4 a
2h
[2A + 2 sinh 2-x - 2Ce- 2" + cos 2P (- 2 sinh 2a - 2Be- 2• - 2Ce- 2") ll
JJ
a_. = p { sinh
h' 2-x [ - A - . h 2,z ( 1 - cos R)
sm 2,-, + Cr •-• (1 + cos 2fJ ) +
8

+ 2Be- 2
• cos 2/J] + : [ cosh 2-x (1 - cos 2P) + Ce- 2 • (1 + cos 2/J) + (3.44)

+ 2Be- 2" cos 2P] -

2
sin 2P [
h• .
1 - cosh 27 + 2Be- 2" + C ( 1 - e- 2•) ]1J
-r = ~ { si:,2/J [sinh 2:x {-1 + cosh 2a - C (1 + e- 2·) - 2Be- 2 '} +

+ A + sinh 2:x (1 - cos 2P) - c e- 2• (1 + cos 2P) - 2Be- 2• cos 2P] -

- 2 sin l.
h: 2P 2 smh 27 + 2 (C + 2B) e- 2• ]} ,
 STRESS CONDillO NS 187

e!.So e - 211'0
where A= l - - - - - - -
2 2
3 e211'e e41Zt
B= - - - --
4 4 4

the distortio n factor is h = sinh~cx + sin fJ.

2 2

At the crown where fJ = 0, a/J = -p, while at the springings where

fJ = :r./ 2, a/J =p (1 +
.. .
,.J,
+ b.
)
to the above
The stress values around various cavities, compu ted accordi ng
in Fig. 3/27.
equatio ns with an assume d value of A = 1 µ- µ = 0·25 a re plotted

50,--, ~~~ ~
z t--t---l½- -1-fr+--+ H lH
a ~.-4
40t-+--tt-+il -+-~4i J.-r-r- ~.--,-- .--.-~
(5 r-t-t- -t--+-- +--+-+-
.i,_;
-1

F10. 3/27. Variation

<Jf vertical and
horizont al
--
tangential stresses
as a function of
shape

ely in the
lhe vertical stresses uv and the horizon tal stresses uh are shown separat
, and plotted against the
fig ure, reduced by the origina l vertica l rock pressur e p
related to the half axes a and b.
relative distanc e from the cavity, this latter being
diagram s that the stresses above
it is interest ing to note from a compa rison of the
rock pressur e p
the crown do not attain the value of the uniform ly distribu ted
the value of uv approa ches that of p
acting on the cavity. Procee ding upward
than 0·5 p and decreas es with incre~sing
asymptotically, while uh remain s smaller
whethe r the cross-s ection is a circle or an
height. The shape of the ~avity, i.e.
vertical ly or horizon tally, does not appea r to
eiiJpse with the major axis lying
signific antly. The stresses arou nd the springi ngs,
affect the stresses at the crown
 188 ANALYSIS OF LOADS

o n the other hand, depend much more on the shape of the section. fhe vertical
stress ac in the case of a circular cross-section amounts to 3 p ; it is slightly Jess
for a vertical ellipse, and considerably bigger for a horizontal one, and decreases
3symptotically towards the initial value p. The rate of decrease is moderate for
circles and vertical ellipses, and is rapid for horizontal elliptical cross-sections.
The horizontal stress ah is also much smaller here and approximates 0·5 p in
circular cross-sections and 0·25 p in elliptical ones. A peak value of 0·8 p appears
only in the immediate vicinity of the horizontal ellipse, but the region affected
thereby is very narrow.
Thus it is demonstrated that cross-sections of a vertical elliptical shape- are the
most adva ntageous also as far as the stress distribution in the vicinity of the
cavity is concerned (cf. Section 22.31).

Stresses around spherical cavities a nd those having the shape of an ellipsoid of revolution
were compared subsequently by TERZAGHI and RICHART with those around two-dimension al
cavities with circular and elliptical cross-sections.
The stresses produced around a spherical cavity by the un iformly distributed vertical load
p are, according to NEUSER:

15B 3(1-a)C] ··o 12B A 2(,z - l)C

a• = [~ + R• sin· - ~ + R.3 + Ra '

248
T,. = [- p +~ + -(,z-- R.4)-
C] .
Sill Ocos 6, T •• = 0, r:,,. = 0, (3.45)

where R, T and cf, denoting the polar co-ordinates

a = 2(1 - µ),

2 + Sa
A =- 2(4 + 5,z) P
p
B=
4 + 5a
Sp
C=
4 + Sa
The stresses are, thus, no longer unaffected by Poisson's ratio and, for instance, the tangen-
tial stresses a 9 at the springings A and at the crown C vary as functions ofµ in the following
manner :
µ= 0 UeA = 1·929p U8 c = -0·21 4p
= 0·2 = 2·0 = -0·500
= 0·3 = 2·045 = -0·682
= 0·5 = 2·167 = -1·161
T he variation is especially pronounced at the crown.
 STRESS CONDITIONS 189

The theoretical determination of stresses around an eliipsoid of revolution was solved by

EDWARDs3 • 1sa. The stresses are again obtained by superimposing the uniaxial vertical stress
fie ld on the biaxial horizontal stress field in the following general form:
i =3
R ; = R 01 + L a,/R u,
i= l

where R0 = the stress in the undisturbed condition prevailing prior to the excavation of
the cavity
a ;; · R ;; = the additional stress caused by the excavation.

Here R u denotes the stress itself and a;; the constants depending on diameter relations and
sa tisfying the boundary conditions.
Any detailed description of the theory involved would exceed the scope of this work and
reference is, therefore, made to the literature3 .i;;b. TERZAGHI and RICHART compiled the stress
coefficients a/p in a tabulated form for various diameter ratios and µ values.

J{)
I
Jf ,,I
- - -Circular cavity
2-5 q ')5 I
z
8
:I ----Spherical cavity
'I
I
I

F10. 3/28. Vertical (a ,) aJJct 15 Z.0 2-5 X/a

horizontal (ah) Px
tangential stressr.s
around circular
and spherical
cavities

Assuming again the value 0·25 for the lateral pressure ratio ),, it is to be seen from Fig.
3 ~8 that the stresses around a spherical cavity are, in general, smaller than those around
. tunnel with a circular cross-section. This should be especially conspicuous at the springings,
• here the peak vertical stress is l ·8 p, which compares favourably with the value of 2·8 p
dc,eloping in the case of a circular cross-section . The vertical stresses above the crown, on
· he other hand, will be slightly higher in the case of the spherical cavity. The difference in
orizontal stress is slight, both at the springings and the crown.
The differences found when comparing stresses around elliptical tunnels and spheroidal
<.u ,ties are largely similar. The difference between the tangential stresses arising around

1.a EDWARDS: Stress Concentration around Spheroidal Inclusions and Cavities, Journ. App.

(uh. 1951 March

~•.- TERZAGHI, K. and RICHART, L.: Stresses in Rocks about Cavities, Geotechnique 1952
ji-75
190 ANALYSIS OF LOADS
8)
b;

f'
J
i:~
fOtfljj
M OMJ /-4 1-8 2-2 Jo
X
r
Stress sc;/e
Z/ 1' "'
b · ,~ 'z'o fr:
FIG. 3/29. Stress distribution around a horseshoe section
(after ZANGAR and PHILIPS) •

three-dimensional cavities and linear tunnels increases in general as the ratio of the principal
axes decreases, i.e. as the spherical shape is approached and as the lateral pressure coefficient
). becomes smaller.

The conclusion to be reached by the above considerations is that not only are
rock pressures smaller when acting on confined cavities (cf. Section 32.313), but the
stress distribution around them is also more favourable.
The distribution of circumferential stresse saround a horseshoe-shaped cavityis illustrated in
Fig. 3/29 after the results obtained by ZANGAR and PHILIPS by the use of photoelastic methods.
Drawing (a) represents the distribution around the perimeter of the cavity, while drawing
(b) shows the variation of stress with distance from the cavity along the line F-F, again for
a lateral pressure coefficient). = 0·25. The peak circumferential stress again remains, apparent-
ly, below the value 3p and extends not further than three times the width. It is of interest to
note the stress peak at the lower corner of the section which can be traced back to the abrupt
change in direction at this point (stress concentration).

The considerable influence of the lateral pressure coefficient l on the distribution

and state of stresses around the cavity was also demonstrated by these authors
as it was by KASTNER (cf. Section 3.12). Horizontal stresses at the crown of a cir-
cular cavity were found by TERZAGHr, e.g. to vary according to the relationship
uh = p(3}. - 1), and the vertical stresses at the springings according to a 0 =
= p(3 - A.), the uniformly distributed vertical rock pressure being denoted by p
and the horizontal rock pressure by pA (cf. Fig. 2/39a and Section 3.3).
The foregoing considerations apply to perfectly elastic, homogeneous and iso-
tropic materials and, thus, are suitable primarily for the demonstration of the
effects and role of various factors governing stresses rather than for the numerical
evaluation of stress intensities around particular cavities.
 --- -

BASED ON DISPLACEMENT AND EQUILIBRIUM ASSUMPTIONS 191

3.23. THEORIES BASED ON VARIOUS DIS.PLACEMENT

AND EQUILIBRIUM ASSUMPTIONS

The foregoing expositions and especially those of the theory of FENNE"R have
revealed that there are two possibilities when deciding how to determine the loads
for practical tunnel-lining design. The one is when the extension of the protective
zone only is considered and the depth of cover is disregarded. The other also takes
this latter factor into consideration. The choice depends on whether the rock in
question is solid, or pseudosolid, or whether it is loose. In solid rocks and in the
case of stresses not exceeding the elastic limit, the development of the protective
zone is unaffected by the depth of the overburden; consequently, the application
of theories disregarding the effect of depth is fully justified. In pseudosolid and
plastic rocks with stresses exceeding the plastic limit, the effect of depth is a relevant
factor not only in the magnitude of rock pressure but also in the time required
for the development of a protective zone; consequently, the use of theories taking
this factor into account is warranted. The design theories used in practice may thus
be classified into these two groups, namely, into those which take into account
the effect of the overburden depth, and those which disregard it.

32.31. Theories Taking into Account the Effect of Depth

32.311. Bierbiiumer's theory. The theory of BIERBAUMER was developed during
ihe construction of the great Alpine tunnels. According to this theory the tunnel
is acted upon by the load of a rock mass bounded by a parabola of height h = a.H
l Fig. 3/30).

1-- _40_1_., .... - (/round surface

----
Q

-
.,. I
I
.,.,.

,1
1 I
.
"l::
0
I i
"' I I LoadmgIbody ~

I
IIJ
B. I ~
Ill I I
·,
~ i~
FIG. 3 30. Rock pressure bulb after
BIERBAUMER .-
\~l b

I
I
-,--
fffdl.
\45"+ r
192 ANALYSIS OF LOADS

THO methods, yielding almost identical resuJts, were developed for the deter•
mination of the value of the reduction coefficient GI'..
One approach was to assume that upon excavation of the tunnel the rock mate-
rial tends to slide down along rupture planes inclined at 45° + </>/2 (Fig. 3/31).
F

FIG. 3/31 . Assumption model

of BrERBAUMER'S
theory

The weight of the sliding rock masses is counteracted by the friction force
2
S = 2/E = 2 tan</> tan.2 (45° - </>/2) · ~ y developing along the vertical sliding

planes and therefore a rock mass of height a.H only instead of H must be taken
into account during the calculations. Consequently, the pressure on width
b+2m tan (45° - </>/2) at the crown will be

p = a.1HY .

Taking into consideration the load diagram shown in Fig. 3/30 the value of a 1
is derived as follows:

P = Hy[b + 2 m · tan (45° - <f>/2)) - H 2 y tan 2 (45° - </>/2) tan <I>,

since
P [ tan</> · tan2 (45° - </>/2)H
p = - - - -- - - - = H y l - ----'----- - - - 1 , (3 .46)
b + 2m · tan (45° - <J>/2) b + 2m · tan (45° - <f>j2)
thus
2
._ a.1 = 1 - _tan</>·
__:._tan ( 45°
_.....;___ </>/2)H
- __:__.:__
b + 2m · tan (45° - </>/2) '
 BASED ON DISPLACEMENT AND EQUILIBRIUM ASSUMPTIONS .193

implying that geostatic pressure is diminished by the friction produced by the

horizontal earth pressure of the wedges EC and DF acting on the vertical shear
planes.
Values of the reduction coefficient for single- and double-track tunnels, as well
as for various angles of internal friction cp and depths Hare compiled in Table 3/11.

The Values of a1 TABLE 3/ Il

For single-track tunnels

b = 1 m, m = 8 m

H~ \, ~ 50
I 20°
I
I 25°
I 30°
I 35°
I 40° 45°

20 0·80 0·79 0·78 0·77 0 ·76 0·74 0·72

30 0·70 0·69 0·67 0·65 0·63 0·61 0·58
40 0·60 0·58 0·56 0·54 0·51 0·48 0·44
50 0·50 0·48 0·44 0·42 0·38 0·34 0·30
75 0 ·42 0·38 0·32 0·26 0·21 0·17 0·12
100 0·36 0·32 0·26 0·20 0·15 0·12 0·09
125 0·35 0·28 0·22 0·17 0·12 0·09 0·07
150 0·35 0·24 0·19 0·14 0·10 0·08 0·06
175 0·35 0·24 0·17 0· 12 0·08 0·06 0·04
:oo 0 ·35 0·24 0·17 0·11 0·07 0·05 0·0~-
min. 0·35 0·24 0·\7 0· l I 0·07 0·05 0·03

For double-track tunnels

b = !Om, m = 8 m

H~~\ 15° 20' 25°

I 30' 35° 40° . 45"

::!O 0·86 0·84 0·84 0·83 0·83 0·83 0·83

3-0 0·79 0·76 0·76 0·73 0·73 0· 73 0·73

40 0·72 0·68 0·66 0·64 0·64 0·63 O·G}

50 0·65 0·60 0·58 0·55 0·54 0·53 0·53
iS o·,rn 0 ·42 0·37 0·33 0·31 0·29 0·29
100 0·39 0·36 0 ·29 0·24 0·18 0· 15 O·t I
l ::!5 0·35 0·30 0·24 0·19 0·14 O· l l 0·08

150 0·35 0·28 0·20 0· 16 0· 11 0·09 (,•fv<

1,5 ()·:'5 0·24 O·!o 0·13 0·09 ()·() 7 0 05

200 0 35 G·24 0·\7 0·12 0·08 0·06 0·04

min. 0·35 0·24 0· Ii 0· J l 0·07 0·05 0·03

U Sdchy: The Art of Tunnelling

194 ANALYSIS OF LOADS

The reduction coefficient a: 1 has two limit values, namely for very small over-
burden depths ex = 1 and at several hundred metres depth whenever H > 5B
the magnitude of a: 1 is no longer affected by depth ~.nci i:ie,;omes

a: 1 = tan4 (45° - ¢ /2).

The second approach of BIERBAUMER for developing a reduction factor a: 2 was
also based on friction arising in the sliding earth mass but a detailed description
of this approach may be omitted as it is not of particular importance. The correct-
ness of BIERBAUMER's formulae could not be completely verified in practice. The
best results were obtained for cavities excavated at great depths in materials
displaying high internal friction (shear strength).

32.312. Mail/art's theory. The basic feature of anothzr theory, that of MAILLART,
is his refusal to accept the belief that the originally uniform distribution of stresses in the
mountain mass is fundamentally changed by excavation. He maintains that as the size of the
cavity is negligibly small in comparison with the volume of the mountain masses, its effect
on stress distribution is presumably insignificant. The determination of pressures is based
on rock strength. A cavity of rectangular cross-section is considered in the axis of which the
rock pressure p 1 is assumed to be smaller than the cube strength of the rock (Fig. 3/32) .

Rock strength .- x,,. ~ Pin.

-----------------------

II. Highly resistant

p<k

!I
Inner spa//ill!l
&tera! spJll!i7g

FIG. 3/32. Basic assumptions of

MAILLART's theory
 BASED ON DISPLACEMENT AND EQUILIBRIUM ASSUMPTIONS 195

This is the condition existing in solid rocks where pressures are of no concern as the rock is
fully capable o f resisting the loads. If, on the other hand, the pressure intensity increases to
Pu which is greater than the compressive strength of the rock, the latter will fail towards the
interior of the cavity so resulting in a structurally more advantageous circular form. On the
.:oncave s urface of the circle the compressed rock is in a more favourable condition than on
that of the rectangle owing to the favourable inward arching action , as a result of which its
strength wiU increase from k to k 1 • For k 1 > Pu equilibrium is again established regardless c f
lhc fracturing which has taken place on the sides. Where the pressure Pu is higher even than
the increased strength of the rock (p111 > k 1 ), the latter will gradually fail again. Considering
now the equ ilibrium of an annular rock zone of radius z and thickness dz, situated concentri-
'4llly around a cavity of radius r, and subjected to the circumferential compression p, and
a ,o from both outside and inside to the radial stresses of magnitudes a, and a, + da, acting
an opposite di rections, we may write:

pdz = (a, + do,) (z + dz) - a,z = a, dz + zda,,

'lcnce
dz do,
z p- a
hich o n integration yields
In z = C - In (p - ( a,).

At the perimeter of the circular cavity, for z = r we have a, = 0 and thus In r = C- In p,

from which C = In r + In p. After s ubstitution

z p
whence z (p - a ,) = rp.
,. p - a,

c, is, consequently, described b y a symmetrical hyperbola with the horizontal line represent-
ing the pressure p and the vertical in the tunnel axis as its asymptotes. Owing to the resulting
confined state of compression the rock strength is increased by these radial stresses.
Relying o n experiments of CoNSIDERE and KARMA N, MAILLART assumed the following rela-
wr ship to exist between the confined compression strength k 0, the unconfined compressive
wength k and the lateral pressure a,:

a(l + k)'
here the value of a is 3 t/cm 2 for concrete and marble and 6 t/cm 2 for sandstone.
Putting 3 arbitrarily for a and using k 1 instead of k the magnitude of the compound stress
obtained as

This relationship together with the radial stresses can be seen plotted in Fig. 3/ 32. The
roe" pressure p 11 at the side walls is greater than k0 , so that the rock will necessarily fail. The
process may be visualized as a progressive fracturing of subsequent thin layers extending over
,1U1ous periods of time. However, by shifting the line representing the a, values which start
at zero at the side wall, higher by a small amounta, 0 , i.e. by providing a certain radial strength
by an internal radial support for the cavity, the pressure Pu will not exceed the rock strength
at the side-faces, so that failure will not even be started. Actually it is this pressure a,o
,-tuch rr:ust be carried by lhe supports. Where the installation of the support system does
oot follow excavation immediately and fracturing of the rock is permitted to proceed, the
required supporti ng force will also be greater, as given by the expression a,oz/r. The importance
of1I.Stalling the supports at the earliest possible time is, thus, again demonstrated theoretically.
 196 ;.NAl,YSIS OF LOADS

The rock p.-1:~sure itself around a circular cavity, according to MAILLt..RT, will oe uniformly
distributed and :;f the magrtitude <1,0. It can be obtained in terms of thi: seo3tatir pressure
by
replacing k 0 by hy in the basic equation. Thus

(3.47)

(All values should be substituted into the formula in t/cm2 units.)

Since the rock strength k 1 increases also with the depth, the pressure on the t11M1el
is not
directly proportion al to the geostatic pressure.
The required thickness of a circular lining can then be calculated from

<l,o
v = r- ,
u.,
where, = the radius of the circle enveloping the cavity
u., = the permissible compressive stress for the wall.
For example, the wall thickness of a tunnel at 2000 m depth, in a rock of strength ,'(
kg/cm2, of a density y = 2·8 t/ m3 with a radius r = 5 m is obtained by taking the 1
= 500
hcto r a
at 3, and the permissible stress in the wall: <1,. at 100 kg/cm2 as

(2·8xJ00 - 3 x2000xl0 0) 2 - (0·5 t/m 2) 2

a ,o = 3(1 + 0·5)2 = 0·009 t/cm:

0·009 t/cm2
v = 500cm - - -- = 45 cm.
0·l t/cm2

32.313. Eszto's theory. The effect of tunnel width is also taken into considera-
tion in the rock pressure theory developed by Eszr6 on the basic observation
made in mining that excavation is followed by the development of rupture sur-
faces outcropp ing to the ground surface. These rupture surfaces become gradually
steeper as fissures appearin g at the ground surface have been observed to start
almost vertically, their inclination decreasing with depth. Rupture failure, thus,
takes places along a curved surface rather than along a plane, and the profile
of this surface is, according to Eszr6, a curve of second order, preferably a para-
bola (Fig. 3/33). Any curve of second order could be selected for the profile of
the rupture surface since the magnitude of the error introduced thereby is µradi-
cally the same for all curves, particularly because the overburden rocks are not
uniform and consequently the curvature of the surface will not change regularly.
The assumpti on of a parabola of second order is otherwise justified by the fact
that in beams exposed to bending and consisting of a material of uniform strength
the magnitude of tensile stresses varies according to a parabola of second order
in proportio n with the elasticity of the beam.
As a most suitable paramete r for the characterization of the entity of over-
burden materials the angle of rupture f} has been chosen by Eszr6. That is, th~ one
included by the horizontal and the straight line connecting the eorner of the ro0f
and the point of intersection of the rupture surface with the terrain. The field
observation of this angle renders no undue difficulty, while the strength properties
 BASED ON DISP LACEMENT ANO EQUILIBRIUM ASSUMPTIONS 197

H. cotan ?J
.Y

II

•f}
X
I
l. /J.tano•h.cotJnc:/,. j. b _ 1

I
I
FIG. 3/33. Assumption model of EszT6's rock pressure theory

are likely to change from layer to layer and although their measurement is, in
principle, possible on specimens obtained from each layer, their behaviour and
interaction in situ cannot be evaluated. In mining areas the angle of rupture is
in ma ny cases known from past experience. Naturally, the actual value of this
angle depends not only on the strength and stratification of the overburden mate-
rial but also on the dip of these layers, inasmuch as it will be steeper proceeding
upwards than downwards along the dip. Nevertheless, the rupture angle is always
futter than the end-tangent of the profile as it reaches the terrain. EszT6 assumed
1he tangent to be normal to the terrain, i.e. vertical, where the terrain is horizontal.
The rupture surface must show an increasing slope towards the terrain, and accord-
fogly a parabola is adopted whose axis is formed by the horizontal terrain and
hose apex coincides with the point of intersection of the chord plotted under
t e a ngle of rupture and the terrain.
The axis of the rupture parabola is thus the line ABB' , the apex lying at B',
'hile a point on the parabola is the upper corner of the rectangular cavity D .
ince BB' = H cotan {}, the equation of the parabola is H 2 = 2 pH co tan fr ,
H .
, ence the parameter p = tan 19. The equation of the parabola is thus (cf.
Fig 3, 33): 2
 ~ - - - - -

198 ANALYSIS OF LOADS

and using the notations of the figure the tangent of the parabola at the elevation
EF is, from the shaded triangle,

dy cotan {} 1 Htan {}
tano: = - = 2x-- - or - -=tanb= - - - - .
dx H tan a. 2(H - h)

In this theory EsZT6 assumed that the cavity created at depth Hand of width
CD = b would not be called upon to carry the weight of the entire rock prism
CDAB extending to the terrain, but one part of the rock load would be trans-
mitted by friction and cohesion, i.e. by its internal strength, to the intact environ-
ment. The weight of the elementary layer EF of thickness dh, and of weight
dp = dhy considered at height h above the cavity would be distributed not only
on surface CD but on the larger surface GH. The distribution is assumed to be
uniform and the straight line bordering the stressed zone is assumed to be normal
to the rupture parabola passing through point F, and to include the angle t,
corresponding to the tangent of the parabola with the vertical. 3 · 16
The elementary load dp, produced by the weight of the layer of thickness dh
at height h, is distributed in the plane of the roof over a width of b + 2h tan t,,
so that from the equilibrium of weights it follows that

yb · dh = y(b + 2h tan <>) dm .

where dm = the average thickness of loading layer in the lower plane GH .

The unit load over width b of the crown is thus expressed as:

• dp = bdhy .
b +2h tanl,

The full roof load is obtained by integrating the above expression between the
crown and the terrain:

H
bydh
H
. bydh
P= I _b_+_ 2_h-tan t, = j
0 0
HhtanD
b+ H-h

3 16
' NOTE. The correctness of this assumption is highly questionable, as it would follow
from the principle of the rupture plane that instead of the earth mass between the boundary
verticals of the cavity, it is the weight of it between the rupture surfaces which is to be distri-
buted. This weight would be partly resisted by friction along the rupture surfaces.
 BASED ON DISPLACEMENT AND EQUILIBRIUM ASSUMP'ITON5 199

After integration we obtain

b cotan {)
p = y - - -- - -~2
[i Htan {) b
n -- - - 1 - - -- - -
]-
(3.48)
1_ ( b ) b Htan 1)
Htan{}

=y (1 -
b'
~ r[
In b'
H
- 1-
b']
H '
where b' = b cotan fJ.
For lack of practical observation values of {}, it may be taken either as {) =
= 45° + cp/2 (where cp is the angle of internal friction), or as 63· 5° (tan {} = 2)
as is common practice in mining.
As pointed out by EszT6 himself3· 17, the pressure calculated according to his
theory should not be used as a design criterion in practice. It is, however, con-
sidered suitable for clarifying potential pressure varia_tions under varying condi-
tions by establishing proportions and relationships. In fact, the greatest merit of
EsZT6's theory is that it provides a better insight into the influence of the 'factors
governing the magnitude of rock pressure. His considerations relating to the
vertical and horizontal distribution of pressures around cavities and comparative
studies on confined underground spaces are of special interest.

Considering the role of various factors included in the equation the rock thickness capable
of transmitting 50 % of the total pressure is determined first. It follows that

p H dh
2 = yb fb+ 2h tan 6 '
the solution of which yields:

1n[1+~(Htan{}-1)]-~(1-
H b H
b
H tan ,f}
)-__!_[lnHtan{}-t+
2 b
b
H tan {} -
]-o '

whence the value of x can be calculated by substituting the appropriate H, b and tan ,{} values.
Depending on the width and depth of the cavity, x was found to range from 8 to 15 m.
Table 3/III is compiled for various combinations of width and depth by taking,{} = 63·5°.

The Values of x TADLF. 3/111

Depth (H) b = 2m b = 4m b=Sm

100 m 5·5 7-5 10·0 =x

200 m 8·0 11·0 14·9 ·-: X
400 m 11·5 16·0 22·0 = x

a. 17 EsZT6, P.: A kozetmozgasok mechanikai elemei (Mechanical elements of rock move-

ments), Banyaszati cis Kohaszati Lapok 1939 24
200 ANALYSIS OF LOADS

It is interesting to note that even under several hundred metres of overburden the magnitude
of pressure is controlled by_the layers situated immediately above the crown, consequently
the quality of these layers is of paramount importance.
Concerning the influence of depth it is shown by substituting numerical values that pressures
vary according to its one fourth (1/4) power, but the rat,i: of increase is more rapid in layers
of poorer strength and in wider cavities. The influence of both the angle of rupture ,f} and
the width of the cavity b on the magnitude of pressure is rriuch more pronounced, inasmuch
as pressures vary as the 3/4-th power of the width and of the tangent of rupture angle as
illustrated by the values of Table 3/IV calculated for y = 2t/m 3•

TABLE 3/IV

b (m)
H (m)
2·0 4·0 8·0

100 p= 12·2 19-6 30·3

200 14·8 24·4 39·2
400 17·4 29·5 48·8

fJ = 63·S 0

Confined cavity
b (m)
H (m) d (m)

4·5 9·0 2 ·0 4·0 8·0

100 p-3·9 7·0 7·4 12·2 19·6

200 4·1 7-8 8·7 14·8 24-4
400 4·3 8·3 10·0 17-4 29·5

fJ = 76°
b (m)
H (m)
2·0 4·0 8·0

100 p =4·3 7·4 12·2

200 5·0 8·7 14·8
400 5·7 10·8 17"4

A closer study of the table shows that by doubling the width the pressure is increased by
63 % for f} = 45°, by 67 % for f} = 63·5°, and by 71 % for ,f} = 76°, furthermore that the in-
crease is 63 % at 100 m depth, 67 % at 200 m and 71 % at 400 m. Pressures are thus seen to
increase with width at a higher rate in rocks of better quality and at greater depths (although
producing reduced absolute values!).

Pressures developing under identical conditions above confined underground

spaces (halls) are subsequently investigated {cf. Section 32.23). Denoting the dia-
 BASED ON DISPLACEMENT AND EQUILIBRIUM ASSUMPTIONS 201

meter of a cavity of circular plan by d, the differential equation describing equi-

librium conditions assumes the form

d2 n . n n
y- -dh = ydm (d + 2h tano) 2 - = dp(d + 2h tnno) 2 - '
4 4 4
whence
H H yd2dh
p =Y .f dm = f
0 0
(d ~ 2h tan o)2

Substituting the value

Htan f}
tano= 2(H-h)'
we have 11
(H- h)2 dh
p = yo 5 H + h dh tan
[ (
f} - 1) ] '' ,

taking the form after integration

d cotan f} [ d d H -,
(3.49)
p=
l _ d )3 1 - Htanf} - Htanf} ln7 tan{} ·
( Htan f}

For comparing these results with the pressure obtained in the foregoing for
a lineal tunnel it is necessary to reduce them to the same area. This may be accom-
plished by using the ratio d = 1· 128 b obtained from the relationship d = b ~ -
By computing the pressures pertaining to various widths and depths in this way,
again for an angle of failure f} = 63·5°, the values compiled in Table 3/V are

TABLE 3/V

b = 4·0 m b = 8·0 m
H (m) a = 9·0 m
a = 4·5 m

100 p = 3·9 (12"2) t/m p = 7·0 (19'6)

200 p = 4·1 (14·8) p = 7'8 (24-4)
400 p = 4·3 (17'4) p= 8·3 (29·5)

obtained . As revealed by a comparison with the previous table the pressures

above a linear tunnel are 3 to 4 times higher than above a cavity of circular plan .
The advantages offered thereby are utilized in practice when using an excavation
front which is circular in plan (cf. Section 61.2).
202 '\NALYSIS OF LOADS

32.314. Terzaghi's rock pressure theory. This theory was developed origin~lly
for cobesionless , dry, granular soils, but it can be extended to cohesive soils as
well. In corresponde nce with actual conditions he assumes a moisture content in
the sandy soil sufficient to secure the cohesion value necessary to maintain the
vertical position of the face in minor headings. The sand masses around the
cavity are already disturbed by excavation, and movement continues when the
temporary supports are installed after the tunnel has been excavated to full
section. These displacemen ts are sufficient to lead to the development of a set
of sliding planes characterizin g the state of imminent rupture in the sand. It is
therefore justifiable when determining the width of the earth mass suffering
displacemen t to use the inclination 45° + </>/2 of the plane of rupture associated
with active earth pressure. This width is thus obtained as

B = 2 [: + m tan (45° - <J>/2)] ·

The displacemen t of the earth mass is counteracted by friction developing on the

vertical shear planes.
The vertical boundary planes of displacemen t may be represented by the ver-
ticals drawn at the ends of the element of width B. The shear strength acting along
these verticals in resisting displ acement may be assumed as

-r = c + <J tan</> = c + <Jh tan</>.

If the unit weight of the soil is denoted by "I and the uniformly distributed
load q is assumed to act on the ground surface, and further, if the ratio of hori-
zontal and vertical pressures is expressed by the empirical constant K, then the
horizontal normal stress uh acting normally to the shear plane and due to the
vertical stress u 0 developing at any depth z under the ground surface can be ex-
pressed in the form uh = Ku0 •
Forces acting on the prism of width B and height dz at depth z are illustrated
in fig. 3/34. The equilibrium of these forces may be expressed as

By dz = B(uv + du,.) - B<1,, + 2-r dz.

Substituting

we have
By dz= B(uv + duv) - Bu0 + 2c dz + 2kuv dz tan cf,,
whence
i<Tv = y _ 2c _ K u tan</>
2
dz B " B ·
 -- - -

BASED ON DISPLACEMENT AND EQUILIBRIUM ASSUMPl!ONS 203

\
\
H \
\
\
\
\
\
\
-l pf-~ \
rn I \
I \
I I
I

I 01str1buitoo of vertical pre.ssures m tl7e /10r1zon

t(#,~ ID!illillU~~--T-~mlliwm+: / plane df the crowr

F,, ,. 3/34. Basic assumptions of TERZAGm's rock pressure theory

\,;suming the boundary condition <lv =q at depth z = 0, the solution of the

J1fferential equation is

s(y- ~ J ( 1- e
<l = - - - - ~
-Ktan4> ~)
+ qe
- Ktan 4> -
2z
B •
" 2Ktan¢

for c = 0 and q = 0
2z
By ( - K tan 4> Fl )
<l = - - - - 1-e ,
v 2Ktan¢ .

and by substituting in any particular case the depth of overburden H into the
equation, the rock pressure is obtained as

2H
_ By ( _ - Ktan 4> 8 ) , (3.50)
Pv - 2K tan <P 1 e ·

At greater depths arching action no longer extends to the ground surface

In his experiments TERZAGHI found the value of the coefficient K to increase
gradually from 1 to 1·5 over a height corresponding to B, and beyond heights
greater than 2·5 B the displacement of the lower layers did not affect stress con-
204 ANALYSIS OF LOADS

ditions in the upper sand layers, so that no arching developed there. Rock pressure
may consequently be regarded as being composed of two terms. Denoting the
heights up to which arching action extends and thence up to the ground surface
by H 2 and Hi, respectively, the soil mass in this latter may be considered an
external load of magnitude q = H 1 y acting on the soil arch and, thus, the pressure
acting on the roof is, in accordance with the theory derived (Fig. 3/35),

2H. 2H,
yB ( - Ktan 4>"Tj - Ktan 4>-B
-
p,. 2K tan ¢ 1 - e J+ Y Ht e . (3. 51)

Height H 2 increases as the roof is placed lower and lower, and as soon as H 2
increases to at least 1/5 of the entire overburden depth H , the effect of the second

Ii.I

FIG. 3/ 35. Rock pressures

at greater depths
(after TERZAGHI)

term becomes negligibly small, while the expression in brackets of the first term
approaches unity. The earth pressure at very great depths thus becomes

yB
Pmax = 2K tan¢ '

and is unaffected by overburden depth. Relying on experiments and experience

TERZAGHI found the value of K to be approximately unity. Taking this into con-
sideration the modified formula yields a linear relationship between rock pressure
and increased width of excavation, and shows a strong resemblance to the for-
mula developed by PROTODYAKONOV (cf. Section 32.324) from which it differs
essentially only in yielding a uniform, rather than a paraboiical pressure distri-
bution in the tunnel cross-section.
 BASED ON DISPLACEMENT AND EQUILIBRIUM ASSUMPTIONS 205

8n the basis of the fundamental relationship TERZAGHI's theory can in principle

be extended to cohesive soils as well, in which case the rock pressure would be
obtained as

B ( -2c)
- 2H
= Y B ( 1 - e- K tan ef, B \. (3.52)
Pv 2Ktan cp J

No rock pressure would develop at all for y = ~. i.e. for B< ~ . In practical
B y
2c
terms this implies, however, no more than the fact that in the case where B >-
y
,
the cavity must be supported temporarily because of the tensile stresses developing
above the roof, and the roof itself is preferably excavated with an arch-like form
in order to protect it against the outbreaking wedge. Owing to the basic assumptions
themselves the theory yields useful results in dry granular soils and in the case of
moderate overburden depths (H < 3B) only.
As demonstrated by GOLDSTEIN3 · 18 the delayed shear strength of clayey soils is
c,nly 50-70% of the initial value, i.e. the value of -r decreases with time. Since in

cohesive soils the roof pressure p 0 = H (Y - ~) , an explanation would be offered

thereby for the gradual increase of roof loads acting on tunnels built in clay soil:
(cf. Fig. 3/67).

32.315. Jaky's concept of theoretical slope. This concept was introduced by lAKY fr,
1925 o n the basis of CuLLMANN's cohesion-parabola. The slope on the margin oflability become:,
~,ccper with decreasing height and assumes an overhanging form at very low heights (Fir.
~ 36a). According to lAKY the cavity is acted upon by the soil masses remaining beyond thL

F10. 3/ 36a. Concept of IA.KY'S theoretical slope

•·• Go:.osnrn, M. and VJRTUMSKY, V.: Vtiyanie polzutsesti na gornoye davleniye (The
clfcct of soil creep on roof pressures), Transp. Strait. 1954 4
206 ANALYSIS OF LOADS

theoretical slope lines fitting to the cavity. The theoretical slope line can be determined by
constructing the tangents on the basis of CuLLMANN's parabola. The depth where the tangent
to the slope line becomes vertical is determined theoretically as

q cos (30° - </>/ 3)

mo =2 sin 2 (60° - 2/3 </>)

(a) In the case of headings located above the limit depth m0 the intersecting curves yield
a roof load only, which is distributed in the shape of a triangle.
(b) In cases where the heading is located below the limit depth m 0 a lateral pressure will
also be present. By fitting the the<'retical slope lines to the lower corner points of the excavated
section, the area enclosed by ttie i_ntersecting curves yields the magnitude of the earth mass
exerting the pressure (Fig. 3/36b). Owing to its conceptual deficiencies this theory has found
no use in practice.
BALLA assumes in his theory 3 • 193 that the medium lying above the cavity will_suffer loosen-

b;

FIG. 3/ 36b. Plot of the

pressure
bulb based on the
theoretical slope

ing and a downward movement as a consequence of the breaking out of the cavity.
This displacement will suffice for the mobilization of the shear strength of the material and
consequently downward movement must take place along some kind of sliding surface.
He arbitrarily assumes the presence of circular sliding surfaces and that they will start
from the upper corners of the rectangular cavity. The radius of the circle, in addition, will
be defined by the assumption that in the intersection of the two cylindrical sliding surfaces,
in the axis of symmetry they must join to a tangent inclined to the horizontal at (45° -
- </>/2), corresponding to that of the sliding planes of passive earth pressure (Fig. 3/37a).
With the distribution of stresses along the sliding surface being determined after Kotter's
differential equation, the rock pressure is obtained from the equilibrium o f all forces acting
vertically upon the sliding ground mass:

G + Q, + K, + P = 0,
where G denotes the weight of the sliding earth mass enclosed within the circular sliding sur-
faces, Q, and K , represent the resultants of shear stresses, viz. of cohesion acting along the slid-
ing surfaces and with P, the resultant of rock pressure, acting upon the roof of the cavity.

· • BALLA, A.: Rock Pressure Determined from Shearing Resistance, Proc. Jnr. Conf S oil
3 19

Mech. Budapest 1963, 461

 BASED ON DISPLACEMENT AND EQUILIBRIUM >SSUMPTIONS 207

After ,lie: aetermination of these

\'alues as functions of the dimensions '?.: :~:'.:'.::~'~:::.:\;.'.·: ~.'/:$j!L~:~;'.-;:<::. ~_:;;::.:>.;: ;_-;
and location, depth of the cavity and
of the strength characteristics of the
overlying ground mass and assuming
I
a uniform distribution of the roof
pressure (p0 ) he obt3.ins
lx+t/>
,2

In this expression Fn, F 8 and Fe

represent factors depending on the
angle of inner friction (<J,) of the
ground material, the values for which
may be taken either from Fig. 3/37b
, ,r from the following table:

b
•
10° +0·6814 +0·1502 +1·8066 Fm. 3/37a. Principle of BALLA' s theory
20° +0·4145 +0·2577 +1·6084
30° +0·2109 +0·3277 + 1·3667
.
400 +0·0757 +0·3671 +1·1016
45 +0·0333 1 +0·3774 1 +0·9667 f
2·0 -r----,---, ---,----,-- -,---,-----,
An inherent defect of this theory
I es in the fact that both the loosening
and pressure of the lateral earth mas-
,cs are neglected as the sliding sur-
faces are started from the corner
points. In addition, numerical compa-
r~ tive results make it clear that the
~.,.cct of cover depth (F H factor) has
a predominant and very slowly di-
cunishing effect.

Fro. 3 ·37b. Resistance factor diagrams

o f BALLA 10 ?O 2,5 ,J!j J5
 208 ANALYSIS OF LOADS

32.316. Suquet's theory of rock pressure. This theory of SuQUET3- 19b was develope
d
from observati ons made during the construc tion of the Paris Under~o
und system. The
assumpti on underlying the theory is that the rock massACB B'A' above the
cavity - bordered
by failure planes extending to the ground surface and having a slope which
depends on the
strength of the rock - will move downwar d and in doing so will bear as an
arch on the spring-
ings AA' and BB' (Fig. 3/ 38).
The soil arch is, .however, capable of transmitt ing to the springings only a
part of the total
load proportio nal to its strength, leaving the rest acting on the cavity.
This excess load is
determin ed by calculating from the unconfined compressive strength
of the soil the unit
weight y' which does not result in a load exceeding the bearing capacity
of the soil. The
maximum stress developing in the soil mass is

a=1y' [f (H+i)
~v,-lllce subsidence

H tan a +
(H+ m):].
2

The total load is

-b(H+ -m i ]
0 = ~[ 2 '.l ) + (H + "!)~ ,
- H tan a 2
8 \
\ I (3.53)
v··
,a 0/ I
\ I which assumes the simplified form
\ I

Fm. 3/ 38. Principle of SuQuET's theory

P= 1 (--b-
y
2 tan
+ H
2
)' a
if the height m of the cavity becomes negligibly small in comparis on with
the overburd en
depth H. By equating the pressure obtained with the unconfined compress
ive strength a~
of the soil, the limit unit weight y ' can be expressed as

, Ge Uc
y = (. b
2 - -+-
HJ = b
--+H
2 tan a 2 tan a
and the partial pressure to be carried by the cavity is

Po = H(y - y').

32.32. Theories Neglecting the Effect of Depth

The second group of rock pressure theories deal; es..sentially, with the deter-
mination of loosening pressure since the existence of any relationship between
the overburden depth and mounta in pressure is denied. A common feature
of
these theories is the assumption that the temporary supports, or permanent lining
3 • 19
b Anna/es des Ponts et Chaussties 1920 4, 1928 5-6
 BASED ON DISPLACEMENT AND EQUILIBRIUM ASSUMPTIONS

of the cavity will be acted upon only by the weight of the stress-free body devel-
oped as a consequence of relaxation following excavation. This assumption is,
of course, incorrect. It is invalid in the vicinity of the ground surface, and at small
depths, but it is equally invalid at several thousand metres depth where the rock
material is in an almost plastic condition, just as in all cases where the quality
of rock (pseudosolid rocks, plastic soils) is such as to permit plastic deformation
to take place on the side walls as a consequence of original geostatic pressure
which is relieved by excavation. Neither is the theory valid in perfectly plastic
materials having an angle of friction equal to zero.
Tbese theories assume the development of arching action in the rock material
a nd the size of tJie load mass depends, according to one approach, on the strength
characteristics such as angle of internal friction of the rock material and on the
size of the cavity, while the other approach attributes a decisive role to the manner
in which the stress distribution is changed in the vicinity of the cavity.
32.321. Kommerell's theory. The oldest and most widely known of these theories
is that developed by KOMMERELL, who determined the height of the loading body
from the deformations of the structure_ supporting the cavity excavated. The
theory is justified by the consideretion that the displacement, or deflection,
of the supporting structure is representative of the
displacement suffered by the disturbed soil mass.
As a consequence of this displacement the mountain
ma terial is relaxed to a height h, which is equal to
the height of the soil column capable of filling this
space of height e by loosening. If the ensuing spe-
cific strain is denoted by b ( %),

e = -hb- and thus h-- IOOe (Fig. 3/39)

100 b

"!I
.,here b is the loosening coefficient of the moun- FIG. 3/39. Annotations
:ain material as a percentage, having for various and shape
---oils and rocks the values given below : of KOMMERELL' s
pressure diagram

0 ( %)
loose granular soil (sand) . . . . . . . . . . . . . . . . 1- 3
moderately cohesive soil (dry clay) . . ... . .. . 3- 5
cohesive soil (marl, gravelly clay) .. .. . ..... 5- 8
soft r"ocks (sandstone, limestone) . . ......... 8-12
solid rocks ... . .. . ...... . .......... . ... 10-15

The supporting structure is required to carry the weight of a rock mass bounded
b~ a parabola, or half ellipse of height h. The equation of the ellipse would be

14 Suchy : The Art of Tunnelling

210 ANALYSIS OF LOADS

and since
4 x2 y2 c52
-- +--~ =1,
b2 (100e) 2
while the total load

(3.54)

Where the mountain material is loose, or soft enough for lateral pressures also
to be anticipated, the curve enveloping the loading body should be started from
the points of intersection of the extended roof line and the straight lines drawn
to the slope 45° + <f>/2 from the lower corner points of the cross-section. Other-
wise the curve should be started from the corners of the roof of the cavity.
The theory of KOMMERELL should be regarded - in spite of its popularity -
merely as a rough approximation for the following reasons:

1. Loosening is possible in granular soils only as a result of the redistribution

of individual particles. Loosening as such is impossible in solid rocks, where
only elastic expansion upon load release can occur, but this is of the order
of a few millimetres only.
2. The linear relationship assumed to exist between the deflection of the roof
beam e and the height h of the loading column could not be ver\fied by
measurement and cannot be explained theoretically.
3. Neither the dimensions or material of the roof beam, nor the width of the
cavity are accounted for in the formula. Yet in evaluating the deflection,
both the dimensions and span, as well as the material of the deflected beam
are important factors, since the load producing a given deflection will vary
within very wide limits depending on those factors.

An attempt has been made to partially eliminate these shortcoxnings by

T. HARosy3 •20 who expanded the formula to take into account the deformations
of the beam itself (1'/p) and of the backspace (Wo 3), and the settlement of the
invert.
The intensity of the uniformly distributed rock pressure is thus:

(3 .55)
p=
1-
(1 ) '
ay C + '1F

Mo HAROSY, T.: Betonalagutak tervezese (Design of concrete tunnels), MTA Oszt. Kozie-
menyek 1958 xxm 1- 2
 BASED ON DISPLACEMENT AND EQUILIBRIUM ASSUMPTIONS· 211

v.bere a - 100/f>
C - the coefficient of elastic subgrade of the soil under the invert.

32.322. Forchheimer's rock pressure theory. On the basis of careful experiments, the
rock pressure theory of FoRCHHEIMER3 • 21 was developed for cohesionless, dry, granular soils
on the analogy of pressure acting on a hole at the bottom of a bin. From these experimental
results it was concluded that the pressure acting on the movable boltom plate is not related
to the height of the sand column, and is proportional to the weight of a body with a fixed
~ape which, in turn, is governed by the size of the bottom opening. Consequently the pressure
as essentially proportional to the cube of the diameter.
The pressure on the movi.ng bottom plate is

yF 2 1 2 tan cf,
P= - - - - - -- -
+
2U tan cf,

Vob1ch in the case of a circle is

1+ 2 tan 2 cf, nr3

pk = y - -tan
- -,1.- - - -
'I'
4- '

an that of a square with sides a is

1 + 2 tan 2 cf, a 3
Pn = y ~ - c f ,- - 8 '

11oh· e fo r a rectangle with sides a and b

l + 2 tan~ cf, a 2 b2
P, = y --ta~ 4(a + b)

For a linear" tunnel ~'tructure of width b the pressure intensity is, finally,

1
p = - yb - --
1 + 2 tan'- -
cf,
. (3.56)
4 tan cf,

It should be noted here that STINI also found this theory of little practical value and suitable
ll'.llmly in stratified rocks over adit sections adjacent to tunnel portals.

32.323. Ritter's theory. In a second group of theories an attempt is made to estimate

the height of the developing natural arch rather than friction along failure planes, and to
ck :ermine therefrom the weight of the enclosed loosening core, considered as the load acting
~ the tunnel.
The first to adopt this approach to the theoretical determination of vertical rock pressure
..s W. RITTER, Professor at Riga, according to whom the magnitude of the vertical pressure
ma} be regarded as the difference of the weight of material within the principally unknown
pn:ssure line and the vertical components of tensile (frictional) forces developing along it:

P = G- z.

" FoRCHHEIMER, P.: Ober Sanddruck und Bewegungserscheinungen im Inneren trockenen

S.ndes, Zeitschr. 0. I. A. V. I 882
212 ANALYSIS OF LOADS

If the tensile forces acting normal to the arch surface are denoted by ZJ', the vertical projec-
tion thereof over the arch length ds will be zy ds (Fig. 3/40). The weight of material within
cos (I(
the pressure line will be

G = i' fy dx,
D 0

a nd thus, the roof load will be

b b ds
p = y fy
O
dx - 1/Z f-
O
-.
cos (I(

{'

X dx
This differential equation is solved
for the criterion of maximum roof
load, i·.e. the surface enclosed by the
curve ABC should include the.greatest
possible weight of material at the
shortest arch length. The equation
derived from this criterion is

dx

FIG. 3/40. Assumptions in RITTER'S theory where the unknown constants are
to be determined from the bound-
ary conditions, i.e. y = 0 for x = 0,
and x = b , and further the criterion of a maximum is that the second derivative of
1
the curve should be yn = - - . Substituting accordingly, the equation of the meridian sec-
2z
tion of the arch is obtained as
X
y = 4z (b - x),

which describes a vertical parabola with its apex at B . The greatest height of the arch is at
the midpoint x = b/ 2 and has the value

The intensity of the roof load is

b'
p = yb [. 48-z - z) · (3.57)

In spite of the considerable mathematical apparatus used in its development, the theory has
not been corroborated in practice. The theory was, however, completed by very valuable re-
search into the most advantageous shape of the cross-section, for which an ellipse was suggest-
 BASED ON DISPLACEMENT AND EQUILIBRIUM ASSUMPTIONS 213

ed 3 · 22 where the square ratio of the minor and major axis equals Ranki_n e's lateral pressure
coefficient:
b'
--.
a-
= tan2 (45 ° - 4>/2)

(cf. Section 22.32. and Fig. 2/48).

32.324. Protodyakonov's theory. This is similarly founded on the determination

of natural arching in the rock. The theory, which has gained wider popularity in
practice following favourable experiences in Soviet tunnel and underground con-
struction, is very useful within certain limits. In developing his theory for
granular materials, PROTODYAKONOV assumed the development of an arch above
the cavity of which the equilibrium is not ensured unless the stresses along the
line ACB (Fig. 3/41) are purely compressive and are not associated with bending.
The arch produced under this assumption will follow a parabolic line with good
approximation. The forces acting on any section DO of
the arch are: p
(a) The horizontal resultant T of reac- T
tions acting from the right at the
crown 0.
(b) The resultant px of vertical pres-
sures.
(c) The tangential reaction R' at point
D of fqrces acting from the left on
the lower half of the arch.
TaJcing the moments about D we have 8
px2 px2
,\ID= - Ty+ - - =0 - 2-= Ty.
2

At the springing A, pressure is exerted F1G. 3/41. Annotations

b} the resultant R which can be resolved and assumptions of
into a vertical component V and a ho- PROTODYAKONOV's rheory
rizontal one H. The vertical component
depresses the arch to the supports, while the horizontal one tends to displace it.
The strength of the natural load-bearing arch can be derived from the condition
that any displacement by the horizontal force N is prevented by the frictional
~sistance developing in the plane A-B under vertical pressure.
With symbols:· N = VJ, where V = p !!_ , and f, the coefficient of internal fric-
tion = tan ¢. 2

~-'~ HORVATH, J.: Kozetnyomasi elrneletek (Rock pressure theories), Banyaszati Lapok
J9S4 Febr.
 214 ANALYSIS OF LOADS

An additional safety reserve against displacement is provided by the shear

stresses -r mobilized in the lateral rock masses. Thus

Substituting the corresponding boundary values x = b/2, y = h and T = N,

into the general equation of the parabola, r.elative to point D we have

whence
b 4/h - b
-r = p2 4h2

Height h of the arch is derived from the condition that it is associated with the
maximum value of shearing resistance -r. The above expression is therefore differ-
entiated with respec' to h

whence,
b b
h= - = - - - (3.58)
2f 2 tan¢
Substituting this value into the expression for -r we obtain

and inserting this into the boundary condition

after substitution the equation of the parabola becomes

px2 pfb 2x2

2 = Ty = 4 Y, y = bf . (3.59)

The load acting on the tunnel is, thus, the weight of the rock masses confined
within the parabola; those lying outside this line being transmitted by arching
action, The area of the parabola is

2
F= - bh,
3
 BASED ON DISPLACEMENT AND EQUILIBRIUM ASSUMPTIONS
215

and the load .per unit length is

2
P =yS= - bhy,
3
and substituting the value
b
h= - - -
2 tan¢ '
we have
I b2
P= - y - -
3 tan¢ '

which when transformed to specific pressure (p = + y ta! <p) shows a certain

resemblance to the simple formula of TERZAGHI for great depths (see Eq. 3.51).
His theory was tested by model experiments, which revealed that, with the
exception of small overburden depths, pressures are not affected by the depth at
which the tunnel is located. These experiments involved granular materials, but
the theory was applied to cohesive soils as well by using an appropriate value for
the coefficient f (to which he referred as strength coefficient). In cohesive soils
f = tan <p + _c_, whereas in rocks f = 1<JOk0 , where uc is the unconfined comp-
Uc
ressive strength and uk is the cube strength of the rock. Empirical values were
collected and compiled for various soils in Table 3/VI.
When selecting the value for the strength coefficient, the condition as well as
the type of rock must also be taken into consideration.
The theory of PROTODYAKONOV has been found to yield satisfactory results at
depths from b to ~ in Soviet practice (Volkov). Satisfactory results
2 tan¢ tan'!-'
were obtained, although rather on the conservative side, in the construction of
the Budapest Underground, as well. The two main objections which can be
raised are:
1. The height of the load-carrying arch, as obtained by the formula, varies as
a linear function of width, although actually this relationship should be
much more involved.
2. The tabulated determination of strength coefficients is inaccurate and may,
in many instances, be selected arbitrarily within very wide limits.
In Soviet practice the principles observed when detennining rock pressures
may be summarized as:
(a) The basic assumption underlying any computation of rock load is the
development of a load-carrying arch which will always develop above the
216 ANALYSIS OF LOADS

Strength Factors after Protodyakonov

TABLE 3/VI

Cate- Strength Crushing

Unit weight Strength
gory &radc Denotation of rock (soil) strength
factor
a,
(k11/m') (kg/cm') I

I Highest Solid, dense quartzite, basalt and 2800 2000 20

other solid rocks of exceptionally 3000
high strength
II Very high Solid, gra9ite, quartzporp hyr, silica 2600-2700 1500 15
shale. Highly resistive sandstone s
and limestones
III High Granite and alike. Very resistive 2500-2600 1000 10
sand- and limestones. Quartz.
Solid conglomerates.
Illa High Limestone, weathered granite. 2500 800 8
Solid sandstone , marble. Pyrites.
IV Moderatel y Normal sandstone 2400 600 6
strong
IVa Moderatel y Sandstone shales 2300 500 5
strong
V Medium Clay-shales. Sand- and limestones 2400-2800 400 4
of smaller resistance. Loose
conglomerates.
Va Medium Various shales and slates. 2400-2600 300 3
Dense marl.
VI Moderately Loose shale and very loose lime- 2200-2600 200-150 2
loose stone, gypsum, frozen ground. Com-
mon marl. Blocky sandstone, cem-
ented gravel and boulders, stoney
ground
Vla Moderately Gravelly ground. Blocky and fis-
loose
2200-2400 - l ·5
sured shale, compressed boulders
and gravel, hard clay.
Vil Loose Dense clay. Cohesive ballast. 2000-2200 - l ·O
Cl11,yey ground.
---
VIia Loose Loose loam, loess, gravel. 1800-2000 - 0·8
VIII Soils Soil with vegetation, peat, 1600-1800 - 0·6
soft loam, wet sand.
IX Granular Sand, fine gravel, upfill 1400-1600 - 0·5
soils
X Plastic Silty ground, modified loess and - - 0·3
soils other soiti in liquid condition
 - -------- .---= cc - - - - --=-------=-=

BASED ON DISPLACEMENT AND EQUILIBRIUM ASSUMPTIONS 217

tunnel. Within this arch a loosening (fragmentation, plastic deformation)

will take place and the tunnel supports have to be dimensioned to resist
the weight of this loosened rock mass (Section 4.15.)
(b) Where the underground structure is built in loose sedimentary rocks where
tan¢ < 0·8(¢ < 40°) and the overburden depth is less than 2·5B, i.e.
H < 2·5[b + 2m tan (45° - ¢/2)], the structure must be designed to carry
the full geostatic pressure calculated at the full depth below the terrain.
(c) In soils composed of various layers of adequate thickness the physical prop-
erties of the layer immediately above the tunnel only must be taken into
consideration, the development of the stress-free body being no longer
affected by subsequent layers. Their load-increasing effect may be taken
into account in the distribution of assumed stresses.
(d) In the case of tunnels built in clay under the groundwater table swelling of
the clay must be anticipated. The effect of swelling may be taken as equiva-
lent to the weight of a soil column extending up to the groundwater table.

32.325. Engesser's theory. Arch action on the analogy of masonry arches is again the
principle underlying the theory developed for rock loads by ENGESSER3· 23, who considered
cohesionless rock masses of infinite extension below a horizontal plane in his investigations.
For determining the minimum specific pressure ENGESSER arbitrarily assumed this to occur
when the angle 1" included between the end tangent of the load-carrying arch and the horizontal
equals the angle of internal friction </> of the rock. With this assumption

hb tan 2 (45 ° - <f>/2) tan <f>

(3 .60)
Pmin = 2h tan </> +
b tan 2 (45° - </>/2) +- 6- ·

The full load on the plane AB, in~luding the weight of the circle segment, is

., ( h tan 2 (45° - <f>/2) tan </> )

p = 4b- y 2h tan <f> + b tan 2 (45 ° - </>/2) +- 6- '

er at a great depth and if the angle is not excessively small

cotan
P = 4b2 y tan 2 (45° -
[
</>/ 2) - - --
2
</>
+ -tan6 -<f> ] ,

1.c. the pressure becomes independent of depth.

3 .: 3 HORVATH, J. : Koietnyomasi elmeletek (Rock pressure theories), Bdnydszati Lapok

1954 Febr.
218 ANALYSIS OF LOADS

The theoretical results were checked against experimental ones. ENGESSER used sandy soil
for these experiments and observed fair agreement provided h ~ I ·Sb. For small depths there
appeared to be a discrepancy between actual conditions and those obtained by theoretical
calculation.
Further, in cases where excavation is carried out with the help of blasting he pointed to
the necessity for taking the loosening effect of the latter into account. For this purpose the
width B of the range effected by the blast should be substituted into the formula instead of
the width b of the tunnel (Fig. 3/42). The height x of the fractured rock material is, however,

45+!l
2 Fm. 3/42. Arching in the cracked
rock layer above
the cavity (ENGESSER)

in general unknown and an approximate method was developed to obtain it. Assuming
that the tangent of the parabolic loosened core will include an angle equal to the angle <f>
of internal friction between the extreme point A and the horizontal secant of the parabola,
the half-length AD of the chord can be expressed from the triangle ADC in terms of the
load parabola as
2h b m+x
CD= tan <f> = 2 + tan (45° + </>/2)'
whence
2h b
x = --,,.- - - tan (45° + </>/2) - m,
tan 'I' 2
and
B = 2 (m + x) tan (45 ° - </>/2) + b.
Roof load is thus due to the combined effect of the rock shattered by blasting below line
AB, and above by that bounded by the parabola:

cotan
p = xy + By [ tan2 (45° - </>/2)
2
</>
+ -tan6- </> ]
. (3.61)
 - --- ~ -= - - - =- - -

LATERAL PRESSURES 219

3.3. DETERMINATION OF LATERAL PRESSURES ON TUNNELS

Although the analysis of loads acting on tunnels is concerned mainly with the
vertical or roof pressures, ·the magnitude of lateral pressur~s is not less significant
for the structural dimensioning of sections. In fact, as demonstrated during the
detailed discussion on genuine mountain pressures, the side walls of the cavity
excavated are th~ first to fail owing to the less favourable structural conditions
developing in the rock there. In some instances lateral pressure may therefore
play a more important role than the roof load. Also, its theoretical estimation
is more involved than that of the latter, since its magnitude is even more effected
by the extent of deformations of the section, so that its value depends increasingly
on the strength of lateral support- besides the properties of the rock and dimen-
sions of the cavity. Lateral pressures are also much more affected by residual
geological stresses introduced into the rock mass during its geological history
which are released upon excavation and whose magnitude depends on the defor-
mation suffered by, and the elasticity of, the rock, but is unpredictable. Genuine
mountain pressure and swelling pressure which cannot be evaluated numerically
may act in full on the side walls.

3.31. APPROXIMATE DETERMINATION OF LATERAL PRESSURES

Lateral pressures in soils are determined approximately from earth pressure

theory, as a product of geostatic pressure, or roof load and the earth pressure
coefficient, respectively in terms of the lateral strain. (It should be noted that on
some occasions the lateral thrust actually observed was greater than anticipated
o n the basis of geostatic pressure.)
Lateral pressures are, according to SnNI, in contrast to roof loads, in a linear
~elationship with overburden depth. This has been proved also by recent model
ests and in situ stress measurements in the Donnerbiihl tunnel in Bern (cf. Sec-
t ion 3.33)
At greater depths pressures on the springings are usualJy higher than on the
-0of, but at the same time frictional resistance is also higher. Again, the effect
~ f geostatic pressure on lateral thrust is less disturbed by stratification than in
the case of roof loads.3 •24
Lateral pressures compiled according to practical experience are given in Table 3/I.
A.s may be noted from the table, lateral pressures range from one-fourth to
c ne-third of the roof pressure. Latent geological stresses in the rock may radicaIJy
:ilter this ratio. Thus, e.g. lateral pressures three times as high as the roof load were
.:ctually observed in one of the tunnels at Hoover dam.

a.::a :SOTE. As an example, TSIMBARYEVITCH suggests the determination of the lateral pressure
fox each layer separately by reducing the vertical pressure computed from the thickness and
t "-~1ght of the layers above the cavity by RAN KINE'S coefficient.
 220 ANALYS IS OF LOADS

According to TERZAGHI a rough estimate of lateral pressure is given by the fol-

lowing formula:
p,, = 0·3y (0·5m + hp) ,
(cf. Fig. 3/ 5), where hp is the height of the loosening core representing the roof load;
in granular soils and rock debris, on the basis of Rankine's ratio

Ph= yHtan2 (45° - ¢/2),

and finally, in solid rocks, relying on Poisson's ratio

µ
Ph= - - - Pv· (3 .62)
1 -µ

According to Soviet practice, and the design code developed for the Budapest
Undergroun d Railway lateral pressures should be assumed in a linear distribution
and should be based on the vertical pressure estimated by one of the rock pressure
theories instead of on geostatic pressure. The parabolic distribution shown in
Fig. 4/6 should be assumed for the vertical pressure having a peak ordinate cor-
responding to the estimated roof load. If the pressure ordinate of the parabola
on the vertical erected at the side of the cavity is h 2 y = p 2 , then the lateral pressure
intensity at roof level will be

e1 = p 2 tan2 (45° - ¢ /2) - 2c tan (45° - ¢/2) ,

and at invert level

e2 = (p 2 + my) tan2 (45° - ¢/2) - 2c tan (45° - ¢/2). (3.63)

Owing to the favourable effect of lateral pressure on bending moments arising

in the section, cohesion which tends to reduce the magnitude of this pressure must
not be neglected in the interests of safety.
The magnitude of lateral pressure is determined also by ENGESSER using earth
pressure theory, simply by multiplying the roof load estimated for fissured rock
(cf. Section 32.325) by Rankine's coefficient.
Adopting an approach similar to that of ENGESSER, PROTODYAKONOV assumed a
loading parabola of increased width over a roof plane extended by the rupture
planes of passive earth pressure and adhered to Rankine's ratio. Accordingly, the
full lateral thrust acting on the side wall of height m of the cavity is

E = my tan2 (45° - ¢/2) { t:n ¢ [b + m tan (45° - ef,/2)] + ; } .

3
 LATERAL PRESSURES 221

As will be perceived, the lateral pressure coefficients involved m the above

formulae are either ). = _ µ_ , or ), = tan (45° - </J/2). The value of 2 has been
2

I- µ
observed to depend not only on Poisson's number but also on the geological
formation. Thus, e.g. it has been definitely established that lateral pressures in
a tunnel under the bottom of a valley differ from those in tunnels under mountain
ridges, inasmuch as the value of 2 is appreciably higher under a valley bottom
than under a mountain ridge. Especially high values are likely to occur where
the valley sides are steep. Laboratory experiments and tests are not very suitable
for obtaining information on J., since HABIB, for example, reported E values deter-
mined in the laboratory by the conventiona l compression test which were five
times as high as those obtained by the in situ compression test. In view of all
these uncertainties, TERZAGHI recommende d, for instance, the adoption for a partic-
ular rock formation of a range of 2 values rather than a single figure. The sec-
tions must then be dimensioned for the critical loads determined by the limit
values of the range.

3.32. EXACT DETERMINA TION OF LATERAL PRESSURES

For a more exact determinatio n of lateral pressure, latent stresses and the passive
resistance mobilized by deformation and lateral outward displacemen t of the
section are taken into account, as well as the pressure which acts as an active
force. This problem is of great significance for improving the economy of tunnel
design, and has recently been investigated theoretically , mainly by Soviet authors
(ZURABOV-BOUGAYEVA, DAVIDOV; see Sections 43.34. to 43.37). The width of the
tunnel section tends, in general, to increase under the vertical loads, and this
movement is resisted by the lateral rock masses. In the majority of cases the active
pressure at the extrados is insufficient and the lateral masses are brought from
an active into a passive stress condition. The actual magnitude of this resistance
'--annot be determined unless the displacemen ts are known. The theories developed
1n connection with this problem will, therefore, be treated in more detail in the
~ hapter on design calculations (cf. Sections 4.33 and 44. 13'.

3.33. EXPERIMEN TAL DETERMINA TION AND IN SITU MEASUREM ENT

OF LATERAL PRESSURES

During the recent construction of a large diameter shield-driven tunnel in Bern

S , itzerland), laboratory and in situ measuremen ts have proved that if local condi-
taon:) do not admit sufficient displacemen t and loosening (e.g. in case of shield
lunnelling) which should be required for the mobilization of the soil's shear
 222 ANALYSIS OF LOADS

Fm. 3/43. Lateral pressure determination after HousKA

strength, horizontal pressures must be derived rather from the overburden weight
(geostatic pressure), than from the rock pressure acting at crown level. Thus
it
may be written after BoNNARD for Ph

/ Ph = H · Y · A0 - 2c JJ:,,
with H denoting the total overburden depth. In commenting BONNARD's experi-
ments, HousKA suggested that owing to the restricted displacement possibil
ities
a "local failure" may be assumed in the surrounding soil-mass and the shear-
strength components should be considered - in conformity with TERZAG
HI's
bearing capacity theory - only with 2/3 of their value i.e. the soil-physical values
of</>' = 2/3 ¢ and c' = 2/3 c should be used.
In this way the vertical rock pressure may be calculated after TERZAGHI accor-
ding to eq. (3.52) with the reduced cf/ and c' values and a difference of

u , = ( H,y-p .') yb/2

may be assumed as acting at crown level on the top of the sliding lateral earth-
wedges (see Fig. 3.43).
Thus the horizontal pressure at any depth z below crown level may be gained
as

Ux =Ph= [y(H + z) + u'] A 0 -2c'J}.~

with d' = m/2. J;.~ + H. J;,; strip width of the superimposed vertical pres-
sure a' upon the l~teral sliding wedges. When calculating pressures -
both
vertical and horizontal - after this method, they were in much better agreeme
nt
with the measured values exceeding slightly TERZAGHI's vertical pressures
(Eq.
3.52), but very considerably the lateral pressures, as calculated after
Eq.
3.63.
 BOTTOM PRESSURE$ 223

3.4. BOTTOM PRESSURES

Bottom pressures should be essentially the counterparts of roof loads, i.e.
reactions acting on the tunnel section from below if the tunnel section is a closed
one having an invert arch. A certain part of this load is, however, carried by the
surrounding rock masses, so that this situation does not occur even in the ca~e
of closed sections, and bottom p.-essures have usually been found to be smaller
th.an roof loads. TERZAGHI quoted empirical evidence imlicating that bottom pres-
sures are approximately one-half and lateral thrusts one-tbird of the roof load
inter.sity. This is due not only to the fact that pressures acting on the roof become
more uniformly distributed with increasing depth in the interior of the rock and
their intensity is thus proportionately reduced, but also to the fact that the weight
of overlying mountain ma.sses tends to increase roof loads but has an opposite
effect on bottom pressures. In the case of sections open at the bottom, i.e. having
no invert arch, pressures of different intensity develop under the side walls and
under the unsupported bottom surface. The pressures arising under the solid
side walls must be compared with the load-bearing capacity, or ultimate strength
of the soil, but do not otherwise affect the design of the tunnel. The magnitude
of rock pressure acting upward towards the interior of the open tunnel section is,
however, unquestionably affected by these pressures (cf. the theory of TSIMBARYE-
\HCH). The development, distribution and magnitude of bottom pressures are
greatly influenced by the method of construction adopted, i.e. by the sequence in
which various structures and components of the tunnel are completed (cf. Figs
3 8 and 3/62).
Genuine mountain pressure and swelling pressures act in a manner similar
·o lateral pressures on the bottom of the tunnel as well. In practice the problem
f bottom pressures is usually encountered in loose soils and especially in plastic,
...iturated clays. The rock at the bottom of the open tunnel section is exposed
· 1 no pressure from above, while at a slight distance therefrom the load trans-
·erred in the plane of the bottom is the same as that acting on the side walls.
; :, long as the side walls consist of solid rock this load is transferred with its full
naznitude to the rock at the bottom which acts as a lower support. In cases,
t ,wever, where the sides of the cavity are not strong enough to prevent the devel-
c !lment of lateral pressures along them, only a part of the original pressure will
be transmitted to the rock at the level of the bottom, the remainder being dissi-
p..ted by the deformat10n work of lateral pressure.
The pressure transmitted to the rock at the bottom is naturally propagated
fLrt her and since the rock displays a certain strength (friction and perhaps cohe-
i n as well) the area over which the pressure is distributed becomes bigger _with
1r.:reasing depth. The extent of propagation can be described here again by a
c ~r. e. Consequently, the rock under the bottom of an open tunnel section is
p -e..:.ure-free to a certain depth, i.e. it will be subjected to the load due to the
:eight of overlying rock only. A pressure-free core having the shape of an inverted
arch will thus develop under the bottom as well, similar to the relieved core
 224 ANALYSIS Of LOADS

above the roof (cf. Figs 3f7 and 3/8). The compressive stress at the top of the
pressure-free core, i.e. at the bottom of the tunnel is zero and increases but slightly
as a consequence of dead weight. At the sides of the core, on the other hand,
high compressive stresses prevail owing to the transferred load. If this pressure
is higher than the strength of the underlying loaded rock, the latter will necessarily
yield and by losing its strength will expand towards the free surface, i.e. towards
the bottom of the cavity, making it heave. This phenomenon should not be mistaken
for the bottom heave occurring when water (in the form of either groundwater
or atmospheric moisture) finds access during tonstruction to highly plastic clays
and these expand as a result of physical transformation. Nevertheless, boUom
pressures may still be affected by such water, inasmuch as water reduces the
strength of clay soils, rendering them plastic, so that the critical condition occurs
under a smaller stress and the clay heaves into the tunoel at a smaller load. The
intimate relationship between bottom and lateral pressures is clearly demonstrated
by the observations carried out during the construction of the new Semmering
tunnel (Fig. 3/44). As may be seen, the-bottom did not start to heave appreciably

JO
-~
;:;,
-!Ii
.....
t20 -~
~
-,ls

-~ j
!?
~
-<:::
~
~
t
"'
~
"'
els 10 ""'
29·4
v,., "v •l-4cm/week
F10. 3/44. Measured bottom
swelling values as a
function of the
0 ~~~""F.::::::;::;::::::;:;;:i..~-:-;--r-::--~---r--::---k.1., progress of
Apr May Jun_ Jul Aug Sep_ Oct. Nov Dec. Jan Feb. Mar.
excavatioc

before the excavation of the crown section was commenced, i.e. until the loads
were transferred increasingly to the sides. This trend increased when the excavation
for the side walls was started, i.e. when the load of an even larger cavity roof
span had to be transferred by the remaining sides to the bottom plane.
Bottom heave may thus occur not only in clay, but, e.g. in sandy soils as well.
The greatest part of the material causing the heave encroaches directly from the
edge of the side walls. This is the point where largest settlements occur, showing
a decrease towards the interior of the sound rock. An inward inclined veritable
arch abutment is thus formed under the side walls in the bottom plane, under
 BOTTOM PRESSURES 225

which the soil will continue its consolidation owing to pressure. Consolidation
will also increase with the degree of inclination. Therefore bottom heave will
stop automatically once a certain inclination is attained, since the load on the
s upporting sublayer is relieved by the inclination itself.

3.4 1. DETERMINATION OF BOTTOM PRESSURE ACCORDING

TO TSIMBARYEVITCH

The problem of bottom pressures was investigated theoreticall y by TsIMBA-

RYEYITCH. As in the stability analysis of the bottom plane of a braced construction
pit. he assumed that a soil wedge is displaced to wards the cavity under the action
of active earth pressure originating from the vertical pressure on the lateral parts.
This displacement is resisted by the passive earth pressure on the soil mass lying
under the bottom of the cavity (Fig. 3/45a).

I
f xcavated hole

FIG. 3/45a. Computation method

of bottom pressures
(after TSIMBARYEVITCH)

The active earth-pressure diagram at the perpendicular of the corner point

of the excavated cavity is a trapeze. The earth pressure at depth x will be

ea = (p + xy) tan2 (45° - ef>/2) - 2c tan (45° - ef>/2) .

At the same time the specific passive earth pressure at depth x is

eP = xy tan2 (45° + ef>/2) + 2c tan (45° + ef>/2).

D~pth x: where ea= eP can be computed by equating the above two expressions.
The layers above this depth will be involved in bottom pressure:

p tan 2 (45 ° - ef>/2) - 2c [tan (45° + ef>/2) + tan (45° - ef>/2)]

X=
y[tan2 (45° + ¢ /2) - tan2 (·;5° - ¢ /2)]

pJ. 0 - 2c(JT; + JT;)

= y(A.p - ).a)

IS Szec:hy : The Art of Tunnelling

226 ANALYSIS OF LOADS

The magnitude of the horizontal force acting towards the cavity above depth x
is given as the difference between the areas of the diagrams for ea and eP. This
force induces a set of sliding surfaces inclined at (45° - ¢/2) to develop in the
soil mass under the cavity.
The force of magnitude E = Ea - EP may be resolved into components T
and S, parallel to the sliding surfaces and perpendicular to them, respectively:

T = E cos (45° - ¢/2)

S = E sin (45° - ¢ /2).

Force T tends to displace the soil and is resisted by the frictional component
of the normal force
T' =Stan¢.

After trigonometric transformations and remembering that the soil is displaced

by forces acting from both corners, the magnitude of forces acting on the bottom
plane is obtained as:
sin2 (45° - ¢/2)
T 0 = 2£ A, • (3.64)
cos 'I'

The resultant T0 acts at the centre line and is vertical. This upward pressure
can be counteracted either by loading
- - - - - - - - - . - - - -------- the bottom with the counterweight of
intensity %, or by a suitably dimen-
sioned invert arch.
The counter load q0 must be applied
Ii over a lengthy, which can be obtained
from the expression
X

/J y = tan (45° - ¢/2)

The pressure acting on the bottom of
the cavity in the practical case illustrat-
ed in Fig. 3/45b and assuming a gran-
ular soil, it can be determined in the
following way:
If the bottom reaction under the side
walls is p = ~ , the height of the soil
s
column at the side of the cavity can be
F1G. 3/45b. Bottom pressure computation obtained from the relation H = .!.
in an actual case (DAvmov) y
 BOTTOM PRESSURES 227

Since
tan2 (45° - ¢/2)
X = H -----;:------,----,-'-----,,-------,--.
tan (45° + ¢/2) - tan (45° - ¢/2) '
2 2

E = Ea - EP = + yx(x + 2H) tan2 (45° -¢/2) - + 2

yx tan (45° + ¢/2),
2

a nd
T = E _si_n _(4_5°_ _- <P/_2)_
cos¢

The pressure acting from below on the cavity is

T0 = 2T sin ( 45° - ¢/2) .

Closure of the section with a bottom slab and the application of an internal
ballast are the only possible counter-measures to this pressure. In the. interest
of safety a coefficient n = 1· 3 to 1·5 has been specified .
With ~ - s denoting the actual loading width, P 5 and Pb the weight of
the bottom slab and internal ballast, respectively, the coefficient of safety can be
determined by comparing the resulting downward stress and the upward pressure
N/y. In other words, it is required that

11 = _ _Y_ ~ 1·3 - 1·5.

b
--s
2

3.42. DETERMINATION OF BOTTOM PRESSURE ACCORDING TO TERZAGH

The relationships developed by TERZAGHI for the bottom pressure in open construction
pits can be applied to bottom pressures in un_d erground cavities as well, by considering the
equilibrium conditions at the bottom of a cavity excavated in the interior of an earth mass
mstead of the open construction pit. Here again two extreme conditions can be distinguished,
one of them being the case of the ideally cohesive soil, where c #- 0 and cf, = 0, the other
that of the cohesionless granular soil, where c = 0 and cf, #- 0. This analysis yields information
on the stability of the bottom, rather than on the magnitude of bottom pressure.
Consider first the case where the soil surrounding an underground cavity covered by a
"ault has a specific density y, cohesion c and an angle of internal friction cf, (Fig. 3/46). The
soil masses above the planes ae and bd adjacent to the cavity and similarly of width b may,
in this case, be regarded as external loads and the problem now is to determine the critical
equilibrium condition of the soil under these planes. The only possible displacement of the
loaded soil wedge is again towards the cavity and the sliding mass consists, according to the
theories of failure relating to the case when cf, = 0, of the active earth pressure wedge bounded
b>· straight lines inclined. at 45° under the loaded section, the adjoining radial shear zone

1~·
 228 ANALYSIS OF LOADS

e'

ti___ J. ~ b -- It
1 I
tI H
it
I I
C ti b/2 b
,-:.........4- ---=---,
ltc
tI
I
F1G. 3/46. Bottom pressure computati on
_ _§_
t: analogue to open excavation
pits (TERZAGHI)

b
formed by section of a circle with a central angle :-z/2 and radius .J2. and finally by the passive
earth pressure wedge bounded similarly by straight lines inclined at 45°.3• 25 The end
points
of the sliding planes are, accordingly, the points denoted by e and d. Verticals are
erected
at these points from the extension of the sliding planes to the ground surface. Consider
now
the magnitude of forces acting on the plane ae, which must not exceed the ultimate
strength
of the soil which can be mobilized in this plane.

Q0 = (b
,
+ !:_)
2
Hi y - mb y - ~ b/y - Hi c=
2 3
(3.65)
= ~ (3Hi - m - !f)- Hi c.
On the other hand , the load-beari ng capacity of the soil strip of width ae = b is

Q = Pmax b = (2 + ;r)cb.
The margin of safety can thus be expressed by the following quotient :

2c (2 + :-z)

in order to obtain adequate safety against sliding into the interior of the tunnel,
n should
have a value of at least I ·3-1 ·5.

3 5
-~ cf. SzECHY, K. : Grundbau l . Springer, Vienna 1963, Fig. 126c on p . 206
 BOTTOM PRESSURES 229

A similar approach is then adopted to investigate the case where c = 0 and cf, =I 0. The
width of the strip ae bounded on one side by the wall and on the other by the intersection
of the bottom plane and the slip surface, and which depends now on the magnitude of the
angle of internal -friction cf,, wiil be denoted by A. The shearing resistance mobilized along the
shear plane, which is again assumed to be vertical, is obtained as the product of the normal
force (active earth pressure) and the specific friction (tan cf,). The active forces are, thus,

b) H, y - -b my - -
Q0 = ( A+- bfy - --yHjtan-(45°
2 - I • • - cf,/2)tancf,
. 2 2. 32 2 '

or, in another form,

Q0 = y H,A - Q,,
where
Q1 = I yH ., tan cf, tan 2 (45 0 -
2 1 cf,2)-
/ by (H - m - 2
2 1
3
I) .
The ultimate load representing the bearing capacity of the soil can be expressed for c = 0
in the following form (cf. the failure theory of OUBRov3· 26):

Q = A 2 yn~

where n0 is a resistance coefficient characterizing the resistance of the soil to sliding, the value
of which may be taken from the following table in terms of cf,:

cf, = 0° 10° 20° 25° 30° 35° 40° 44° 48°

n0 = 0 5 10 20 40 140 260 780

Sa(ety from sliding is again expressed oy the ratio n = Q/Q0 • The minimum width A can be
determined from the extreme condition dn /dA. In other words, the expression
2
d 1m 0 A
dA y H 1 A - Q
must be zero. Hence
(yH,A - Q,) 2yn0 A - yn0 A 2 yH 1 = O
(yH 1 A - Q,)2

which is impossible unless the numerator is zero, i.e.

From this equation of the second degree A = 2 ~ and substituting the value of Q,
y H,
determined above, the minimum vaule of A is obtained. It is necessary that

A = H 1 tan cf, tan 2 (45° - cf,/2) - _!!__ (H 1 -

H1
m - 2.1)
3
> 0.

..:, D UBROV, A.: Gidrotechnitcheskoye Stroyitelstvo 1951 52 and cf. SztcHY: Grundbau.
Springer, Vienna 1963 I 177
230 ANALYSIS OF LOADS

which can be further increased to attain the specified degree of safety. The value of n should
again be higher than 1·3 to I ·5. For A = 0 the safety against sliding is insufficient at the bottom
of the cavity, so that special support, or ballast must be provided.

3.5. DEVELOPMENT AND SUPERPOSITION OF ROCK PRESSURES

3.51. DEVELOPMENT OF ROCK PRESSURE DURING EXCAVATION

It was mentioned during the general description of rock-pressure phenomena

that the magnitude of rock pressures is greatly affected by the time before the
internal support is installed and by the strength of the support.
Driving and (:Onstruction of any "tunnel consist of repeated cycles of several
operations such as excavation, installation of temporary supports (propping)
and construction of the permanent lining. Driving is always accomplished ,at the
excavation front which, in rocks, is usually advanced by blasting. Support con-
ditions of the rock surrounding the cavity are more favourable here than over
normal sections, since the rock here is supported on three sides, but only on two
sides in the latter case. The rock over the face forms a half dome which is capab"Ie
Progressive f'a1/vre of half dome ifno
...------..._ / fvrther advance in made and no
I
/'.,,,---- - - -''<' , svpport is installed
I *"
/ / '-:,--
. . . ., ..... _ ____,---------------
I I
// /'-118/f dome
I
--------
-t-- 6rovnd arch 8)
I/ I..._,
'4 I ~ I
\:o I 9-_'l::, I
~ f1 I~~ I
"S~

: ~~
01Adv,1nce pen z 4 x/1 •
cgcle OisMnce from
working j'ace

1 Unsvpported
_:xcavatmo r o)
.::: Cf!c/e .
E::
~ r--- - - -
te
~

Fm. 3/47. Relation

between time,
overbreak and
rock-load
in blocky and
2 J 4x~ seamy rocks
rtme in excavation C!!cle vnits (after
TERZAGHI)
 DEVELOPMENT OF PRESSURES 231

of carrying a larger load than the arch or vault developing over normal sections.
After blasting, the half dome may remain stable for a certain period even without
supports, but gradualJy increasing quantities -of material will drop into the unsup-
ported cavity thereby leading to the formation of new arches extending higher
and to increasing distances (Fig. 3/47). Theoretically , the process may go on until
the entire cavity is filled with loose rubble.
This process of gradual breakdown depends not only on the quality, bedding,
jointing and fissuration of rock, as well as the width of the excavation, but also
on the distance between the last support and the rock face (/J. The process is
extended in time because of the viscous, tough behavio ur of the cementing material
filling the joints and fissures of the rock. For a certain period prior to breakdown
the loosened rock itself is capable of overbridging.tbe unsupported cavity. This is
referred to as the bridge-action period (th). The sequence and method of excavation
should be selected to enable the installation of the necessary new support before
the bridge action period has expired. The position of the bridge action period in
relation to the sequence of operations during the construction of a tunnel exca-
vated by blasting in solid rock is shown in Fig. 3/48. Considerable overbreak is
likely to occur if tb is shorter than the ventilation period, t.,, or slightly longer,
so that it expires early during the mucking cycle. On the other band, very little or
no overbreak will occur where the bridge action period is sufficiently long to permit
the erection of the new supports (t,). In some instances it may therefore be advisable
to reduce the length / 1 (see Fig. 3.47) consistent with the character of the excavated
material and the width of the cavity. It should be noted that in wide cavities
the bridge action period is appreciably shorter than in short spans, i.e. narrow

Br1r,ye aclton per1od(t6)shorter than (ts,Jmvolm

ri t",., /I
overbreak
Ttme required to sc.?le down and mampulate crown
bars or install temporary posts
/ I
i i, I
tiz "'
<:,
.;:, ,,,"'
c:,._

~ ~
~..., 't::,
~ !:': ~
..:::; ~ ....
t.. g-.
~ -i;; <:,
.;::. ::::::
<..i
~ t:
~ ~ ~ c::,

0 Time

FIG. 3/48. Diagram representing operating cycle for one round

 232 ANALYS IS OF LOADS

sections. The rock load to be carried by the suppo rts shows

, however, an increasing
tendency even if the supports are installed in good.time
and are sufficiently strong,
if wedging is carried out carefully and the overbreak
is backpacked before the
bridge action period expires. This increa se can be traced
back to two causes,
namely, to the transition from half-dome action into pure
arch action as the face
is driven forward, and to the fact that rock movement
is not entirely prevented
by backpacking and wedging until the loosened rock above
the roof adjusts itself
to the changed equilibrium conditions. The joints in the
rock immediately above
the suppo rt become closed by the suppo rt, while those.
higher upward will widen,
owing to a loosening of the rock. Consequently the load
on the roof increases
-as long as movement occurs. The extent of overbreak
and the magnitude of the
space backpacked either poorly or not at all, play, theref
ore, equally impor tant
roles in the development and magnitude of rock load. The
final magnitude of roof
load and the period of movement, or load-increase depen
d greatly on the thor-
oughness of backpacking and wedging. Where these. opera
tions are carried out
carefully, equilibrium conditions and the final load HP
may even develop after
the first week (curve c 1 in Fig. 3/47). On the other hand,
insufficient care, inad-
equat e backpacking and loose wedging will extend the
load increase period to
several weeks as illustrated by curve c~, and the final load
H'max will also be higher,
in spite of the fact that the initial load may have been
smaller. The consistent
displacement of the rock towards the suppo rts will result
in a progressive failure
of the rock material in the load-carrying arch.

The increase of roof loads is illustrated after B ENDEL

in Fig. 3/49 for a tunnel constru cted
accord ing to the Belgian metho d (cf. Section 63.21). In
the figure the excava tion of the front
is indicat ed by " I", the excava tion of the roof and the
constru ction of the roof arch by "2",
and the constru ction of the side walls by "3".Ac cordin
g to BENDE L the magnit ude of the minor
axis of the load ellipse is governed not only by the strengt
h of the rock but also by the speed
of constru ction. Therefore, the angle charac teristic of
its magnit ude is defined in solid rock

flock pressi,·re w/Jensvpport oft!rr(t.• ,~

_ _ _ _ _ _ _ eJ/!£!!.d with wual lm6m,jg _-

flock pressure .,~ fvn,/Jon of loJd!nfl

el1psc w/Jr.1 tunnel ronslruc/Jon
h pro;ress 1s low

+~~ rt-'f-.!!..f '~~~ =~~= ~~=~ ..:.., -~;.'~ ~~~i7

,,7,,~ ~~~ Slr:u;.11/rJ:'. ~-.1
,.,,,.,.,f",--"-"--...........-.....~_.,=:.,_.,___;""""':v4 I. fl'i/Vf',SS
,:sr1tfCOIIJ{rl',!19fl
: , . ; . . { , J L . , - - - - - - - - - - 11. prorra:s
/}
70 JO 5(} 50 70
Nvmb~r of da;,1s

F10. 3/49. Rock pressure develc pment with the Belgian

m~thod for both slow and rapid
progress in constru ction ( BENDE L)
 DEVELOPMENT OF PRESSURES 233

on the basis of Rankine's theory as cx 1 = 45° + </!/2, in highly squeezing rock or in the case
of slow construction progress, this may be obtained on the basis of the limit condition of flow
as -x 2 = 45° - </!/2.
In low-strength, cohesionless soils the length of the bridge-action period natu r-
ally approaches zero incurring an appreciable reduction of the excavation span
(increase in the number of drifts), the unsupported length from the face must be
minimized (employment of forepoling) and tunnelling methods based on these
principles (cf. mining methods, shield driving) must be resorted to. It is interesting
to note that the bridge-action period is considerably longer in weathered rocks
than in crushed ones, but the load increase in the former may be many times that
of the latter, and may take a very long time to develop. Pressure conditions in
rocks of different quality plotted by TERZAGHI in terms of the bridge-action period
are shown in Fig. 3/50, where time is plotted in units required to complete one

Successive overbreJK Imes 1j'excavat1011 1s diScOfllml'i!d a11d 110

wpport is msta!led (ordir..~Les of dasl, l!f7es m fig b)

a)
l_____~
H ,-
,' , -- ~
-t t,- ---
1 ·
Overbreak !me immediately after rou11d 1s (Ired
.,,, Backpack
'
I,) I
I
I
I

,H ¼

Supported
•

{J11suppofed 1 1 ,
Dashed lines indicate increase ofoverbreak
1n workmg cycle umts

I
R1g1d support
(!iiled ,--- --...+;..,c:._

2 J I- 7 8 11 12
Bridge acl!o11
period

F1G. 3/50. Relation between time, overbreak and rock load (after TERZAGHI)

phase of operation. The extent of the breakdown of roof is indicated by a dashed

line, and the magnitude of roof loads involved by full ones. In sound and mode-
rately fissured, seamy rocks the length of the bridge-action period is seen to be
practically unlimited, and supports are installed for the sole purpose of protection
against spalling. The bridge-action period in crushed and shale rocks may extend
fro m a few hours to several days. For plotting Fig. 3/50b the length of the bridge
action period was assumed to be 80 % of the time required for completing one
phase of operation.
 234 ANALYSIS OF LOADS

3.52. SUPERPOSITION AND INTERACTION OF ROOF LOADS

ABOVE ADJACENT TUNNELS

The distance at which a new tunnel may be driven parallel to an existing one,
or at which various elements of a system consisting of parallel galleries and halls
may be arranged is of interest to the tunnel builder in many practical cases.
Owing to the nature of rock pressure not only may the construction of a new
tunnel too close to an existing one transfer. a load considerably in excess of the
original to the latter, but the load on the new tunnel may also be appreciably
higher than that estimated for a single tunnel. In this connection two problems
arise. The first is to determine the minimum thickness of wall-masses or pillars
between parallel tunnels or adjacent cavities that is necessary to prevent them
from being crushed by the load superimposed from both sides; the other, to
determine at what distance parallel tunnels should be located to prevent the devel-
oping neighbouring loading parabolas, stress relieved cores, or protective zones,
from fusing together. Bridge action in this case would be the same as over a single
continuous cavity.
Load-interference, naturally, is affected by all the factors affecting the magni-
tude of rock load over a single tunnel, such as rock strength, depth of overburden,
width of tunnel and the method of construction. According to the elasticity theory
concerning stress distribution around circular cavities, tangential stresses in undis-
turbed ·rock extend to distances equal to three or five times the diameter (cf.
Section32.23). Thus if it is intended to avoid any stress accumulation by the exca-
vation of a s.econd tunnel, the latter must be located at least two diameters distant
from the first (Fig. 3/51).

[',MtyN: 1
a)

Fm. 3/ 51. Tangential pressures

acting on the side of
b) two adjacent cavities:
(a) when undisturbed
(b) when disturbed by the
superimposed pressures

In cases, however, where adjacent cavities serve the same purpose, or plant,
a rigorous adherence to this principle would cause considerable difficulties and
would involve extremely long connecting galleries besides rendering fast indoor
traffic of a plant as one unit impossible. Stresses transferred to sound rock decrease
according to a hyperbolic function so that the increment load caused by them
 SUPERPOSITION OF PRESSURES 235

beyond a certain distance - smaller than that mentioned - becomes insignificant.

According to various elasticity theories (cf. KlRsctt's formula in Section 32.22) the
peak circumferential stress around a circular cavity occurs at the side of the cavity,
at the springings, and its variation with the distance from the cavity is illustrated
in F.ig. 3/52.

FIG. 3/ 52. Decrease of tangential stresses around a circular cavity

Relying on his rock pressure theory (see Section 32.313) EszT6 computed the reduction
of transferred load over the distance x from the cavity for different tunnel' widths and depths.
Using the notations adopted in Fig. 3/33, pressure increase is calculated by the same differen-
tial equations, with the modification, that the distance x from the cavity is adopted as the
mdependent variable, while the dependent variable is, again, the thickness h of the layer,
·which is proportional to the transferred load. Thus

X y b dh
Px = ! b + 2x '
where
h Htan {}
x = h tan o= 2(H- h) '

and
2xH
h = - - - - .
Htanff+ 2x

On integration this becomes

b tan {} [ H tan {} + 2x H tan {} - b ] (3.66)

Px=Y(l- b ) 2 ln b+2x -Htan{}+2x·
Htan {}

Assuming an angle of failure (rupture) {} = 63·5°, transferred loads, in t/m 2, are obtained
for different depths and tunnel widths as given in the following table:
 236 ANAL'rSlS OF LOADS

"(m) - 0 I I 2
I 5
I 10
I 20
I 50
I x,
--
H(m)
I b (m) I I

100 2 7·4 6·0 5·2 3·9 2-9 1·9 0·9 6

4
8
12'2
19·6
10·6
17·8
9·5
16·:5 I N
13·4
5·5
10·3 I 7·0
3-7 1·7
3·3
8
1!

200 2
4
I 8·7
14·8
7·3
13· 1 I 6·5
12·0
5·2
9·8
I 4·0 3·9
I 115·7·0
1·6 8

8 24·4 22·6
..
21·2 18· I I 7-8
14·1
I
3-2
6·2 I 12
16

400 2 I 10·0 8·7 7·8 I 6·5 ! 5·3 4·1 2·6 12

4 17·4 15·8 ! 4·6 i2·4 10·3 II 8·1 5·2
8
I 29·5 27-7 26·3 23· I I 19·7 15·7 10·4
17
23

The distance x, at which the transferred load decreases to half the roof load above the
cavity is indicated in the last column.
As clearly illustrated by both the graph and the table, the load transferred to this s1,rround-
ing rock is, according to EszT6, of considerable magnitude and its effect can be felt even at
fairly large distances. Consequent ly drifts and tunnels driven close to each other (e.g. sub-
ways) are subject to considerable superimpose d loads. The increase is especially large if a wider
cavity is excavated beside one of smaller width at a great depth. The exam ple of a si ngle-track a nd
double-trac k drift, located at 200 m depth a nd spaced at 20 m distance is quoted by EszT6.
In this case, if the original load on the 4 m wide double-trac k drift is 14·8 t/ m\ the transferred
load will be 3·9 t/m 2, correspondi ng to an increase of 26 %. In the 2 m wide single-track
drift the original load is 8·7 t/ m 2 , whereas the transferred load is 5·7 t/ m 2• The increase is
65 %, and this considerable superimpose d load could not be neglected in designing the interna
l
supports. EszT6's hi gh values have, however, not been encountered in tunnelling practice and
the load increases actually observed have been appreciably lower. On the o ther hand, the
effect of depth must not be neglected when estimating load increase.
According to EszT6's theory, transferred loads increase with depth at a higher rate than
do original rock pressures. Whereas roof loads were shown in Section 32.3 J 3 to increase as
the fourth root of depth, transferred loads appear to be proportiona l to its square root.
Considering the distance at which transferred loads decrease to half their original val ue,
this will be seen to occur within two or th ree times the tu nnel width, and the wider the tunnel
the sooner this will take place, but the greater the depth the later it will do so.

Experiments of considerab le interest have been performed by LEON and WJLL-

HEJM, who demonstra ted in a very convi ncing manner the effect of tunnel spacing
on the developme nt and superposition of stresses (Figs 3/53 a nd 3/54). Variations
of pressure around a double section are illustrated, revealing the consequences
of the neglect of a certai n minimum distance between the units.

I n the first case (Fig. 3/ 53), where the distance beiween the tunnels is 0·4 b the developmen t
of stresses takes the followi ng course:
Tensile stresses of relatively high magnitude develop even under low pressures at the roof
and at the bottom of the tunnel, as well as in the axis of symmetry of the double section.
These stresses result in cracking in the roof and bottom, as a consequenc e of which they are
relieved, while stresses in the axis of symmet ry increase at about the same rate (a).
 ----

SUPERPOSITION OF PRESSURES 237

At the same time compressive stresses develop in the external walls as well as in the internal
ones facing each other. In the central buttress wall these stresses are superimposed upon
each other and, sooner ot later, exceed its compressive strength (b). Consequent crushing of
this buttress will lead to a reduction of compressive stresses there, but will be accompanied

~11

FIG. 3/53. Development

of stress conditions around closely spaced cavities

by a largely proportional increase of those in the exterior side walls. As a consequence of

the redistribution of pressures, the tensile stresses in the central axis will increase and result
Ill cracking of the roof and bottom. The tension zone will shift up or down, resulting in a
further increase of pressure on the side walls. No load will be carried at all by the crushed
central buttress and the stress pattern around the double tunnel is identical to that which
would have developed around a single tunnel of a width 2b + 0'4b = 2'4b.
lo the second case (Fig. 3/54) the wall between the two tunnels has a thickness of 3·3b,
and the redistribution of stresses takes a course essentially different from that in the former
case. The first phase consists, here again, of the rupture of the roof and bottom in both

FIG 3. 54. Development of slress conditions around cavities spaced at greater distances
238 ANALYSIS OF LOADS

tunnels, brought about by tensile stresses. As a result the tension zone above the tunnels will
be moved upward towards the sound rock and shows a considerable increase in the a.r.is
of symmetry as well. Its centre of gravity is shifted further upward or downward.
Hereafter the development of shear cracks (sliding planes) can be observed in the side
walls, the increased tensile stresses in the axis of symmetry cause the rock to crack, bringing
about the splitting and upward or downward shifting of the tension zone. The tension zones
above and below the tunnels are, however, relieved by this process and further opening of
initially developed cracks is stopped.
Popping eventually takes place on the side walls and compressive stresses move towards
the interior of the rock, yet the intermediate buttress wall is not crushed and excluded from
load bearing, although from the figure, the exterior walls would still appear to carry a greater
share of the load.

In the above experiments LEON and WILLHEIM took vertical loads only into
consideration. Actually these are always accompanied by a higher or lower lateral
pressure which is naturally not without effect on the magnitude and distribution
of stresses. The behaviour of the rock mass remains, however, essentially the
same.
Relying on site observations a tunnel width to buttress wall-thickness ratio
of 1 : 3 is suggested by SnNr. In so und and favourably stratified rock this ratio
may be reduced to 1 : 2.
The effect of stresses transmitted from adjacent cavities can be estimated on
the basis of the theory of elasticity as well. For this purpose the values u, should
be plotted in the springing line (0 = 0) at the points r = a; r = 2a; r = 2·5a and
r = 3a, as obtained from the relationship

2
u1 p m r
= - - - - - --2
+ a2 -
p m - 2 [
-
3a 4
- -- 1 + - 4
)
cos20
2 m- 1 r 2 m- 1 r

(cf. Fig. ~/52). As may be seen, the stress increment at distances 2a and 4a from
the side of the cavity is 22 %and 4 %, respectively, so that a distance corresponding
to two tunnel diameters should, in agreement with SnN1's suggestion, afford
adequate safety in practice. Interference intensity increases slightly with rock
strength and with the value of m.
Therefore, stress interference should always be taken into consideration when
excavating tunnels side by side, or when driving a new tunnel in the vicinity of
an existing one. There are several instances on record where the lining of old
tunnels has been badly damaged by the increased load it was called upon to carry
when a new tunnel was driven in its vicinity (London, Piccadilly Station).

3.53. LOADS ON ROCK-PILLARS

Rock zones left between underground cavities are referred to as pillars. Their
correct dimensioning is of paramount importance for the stability of adjacent
tunnels.
 SUPERPOSITI ON OF PRESSURE~ 239

35.31. Protodyak onov's Theory

According to trus theory, the additional load to wruch the pillar is subject
consists of the weight of the rock mass extending to the surface ab_ove the pressure
arch {shaded area in Fig. 3/55), since the load due to the masses under this arch
are carried by the tunnel lining. The additional load is assumed to be linearly

---,;;;-- ·

Loadmg area for

pt/far

h,

b d b

Fie. 3/55. Compressio n stresses produced in undergroun d rock pillars (PROTODYAK ONOV)

distributed in the pillar cross-section, showing peak values at the pillar side and
decreasing to zero over the distance s. The condition of equilibrium may be
expressed as:

-Hb y - [h' -b2 Y + -23 -b2 (h 1 - I1') y ] - - 21 ns = 0 .

2

'-leglecting the term in brackets in the interest of safety we have

Hb ns
2 y- 2 = o·
For the ratio of peak stress p to distances PROTODYAKONov's experimen ts yiel-
ded the value
p
-s = 0-0745/.
 240 ANALYSIS Of LOADS

After substitut ion we have

Hb 0 . 0745/s~
- y- 2 = o.
2
whence

Conseque ntly the buttress must be at least 2s wide, but a safety factor of 2 may
be introduce d and the minimum distance between adjacent tunnels is thus obtained
at about
d(cm) = 20 l 2/ )¼,
ybH

where/ is PR0TODYAKONov's strength factor (cf. Section 33.324). All values in the
formula should be substitut ed by kg and cm units. Using metres and ions the
formula becomes

d<m> = 0 · 65 ( ybH )t
2.f .

35.32 Tsimbary evitch' s Theory

Another distributi on of compress ive stresses in the pillar wall was assumed b y
TsIMBARYEVITCH, as shown in Fig. 3/56. The slopes of the broken line are defined
by the sections q and p for which the following relationsh ips are given:

q S: 0·5 (K - 1) and p S: 0·6b

where K is the stress concentr ation factor measured at half the height of the tunnel
side wall, for which we obtain the following tabulated values calcuiate d according
to the theory of elasticity :

II
__Jb - - -----

m
p FIG. 3/56. Distribution
of compression stresses
actina upon rock pillar~
b s
(TS!MBARYEVITCH)
 - - - --- - = -= c;;c- =- - =- - -
---- - --- .

CRITICAL TREATMENT OF ROCK PRESSURE THEORIES 241

'funnel shape Circular Elliptical Rectangular

--le---- - - ---/--- - -- - -- - + - - -- - - - -- - -
Ratio m/b 2:33:23:1 1 : 1 1 : 3 1 : 5 1 : 18
K 3-75 2·2 l·0 1·3 2·0 2·23 5·0

Distance a over which additional stresses are spread, according to the above
elasticity theory, is equal to 3b. A corresponding value for s=2·5b may be assumed.
A further condition to be satisfied is to keep the stress in the buttress below the
permissible maximum level.
For the case d > s the
maximum stress becomes
u = (K - 1) yH .
Ford < s, however, the com- d /J
ponents of the superimposed
stresses may be obtained from
the relationships (Fig. 3/57):

<I~= <Ib + q-u-(1 _ !!_),

---
J J s,

U;= <Ij+ qb<Ib(l - ~),

which values must remain

below the permissible stress
for the rock in unconfined
compression Uc. Of course, Fm. 3/57. Pressure distributions between rock-pillars
superimposed stresses may
also be obtained by an ap-
propriate summation of circumferential stresses calculated according to the
theory of elastic.ity.

3.6. CRITICAL TREATMENT OF .ROCK PRESSURE THEORIES

AND IN SITU ROCK PRESSURE MEASUREMENTS

From the wide variety of rock pressure theories presented a series of sound and
intuitive assumptions, rather than satisfying uniform results may be concluded.
This is but natural because of the essential deviations in the presumptions and
of the impossibility of the theoretical consideration of all the relevant factors.
Besides, it must be added that the preliminary theoretical consideration of factors
like the method of excavation, rigidity and time of propping, consolidation process

16 Sz;Echi: The Art of Tunnelling

 242 ANALYSIS OF LOADS

of deformations, stratification and relative location of overlying layers, magnitude

of inherent residual stresses, etc. seems to lie beyond the capacity of both elastic
and plasticity theories.
Bearing in mind their inherent defects, practical design may be based on the
following rock pressure theories, with due regard to the circumstances and local
conditions:

1. Up to a depth of cover of H < 2·5B full weight of the overburden (full

geostatical pressure) must be taken into account.
2. For cover depths exceeding this value in all kinds of grounds the use of
TERZAGHI's theory or that of PROTODYAKONOV is permissible; for approximate
values, and chiefly in cohesive and solid ground, the author's theory based
on a statical analogy may be advised (cf. Eq. 3.12).

Because of the incertitude and inexactitude of rock-pressure theories, the dimen-

sioning of tunnel linings must be regarded as uncertain, particularly as recent
research shows that tunnel lining can never be considered a free-standing supported
structure but an embedded one exerting a composite action with the surrounding
ground-material against external loadings. Thus, the surrounding ground represents
not only external loading but simultaneously affords support and composite
action against it.
Difficulties and uncertainties inherent in theoretical rock pressure computations
cannot be restricted, nor can new better-founded computation methods be devel-
oped unless reek pressure estimates are checked against site observations.
The newer trend of investigations has, therefore, renounced the method of
developing further rock-pressure theories, but takes the strength and deformation
characteristics (crushing and tensile strength, Young's moduli, Poisson's ratio,
shearing strength and plastic properties, etc.) of the surrounding ground for a
basis. Thus, the aim is to obtain information about prevailing stress conditions
and pressures from deformations and strains actually measured in exploration
drifts and borings, and on the actual loads to be considered 'for the dimensioning
of tunnel lining.
Laboratory or model tests can afford satisfactopY information on stress condi-
tions prevailing in the rock in exceptional cases only, since actual stresses in the
rock are governed to a great extent by its natural bedding and geological history.
None of these lends itself to laboratory modelling.
Natural conditions may perhaps be most truly reproduced in the model if the
r,rototype consists of a relatively dry granular soil, or rock debris, but even then
no means are available for simulating moisture conditions, construction method
and dimensions.
An auxiliary, supplementary role may be attributed to model tests in the labo-
ratory only, where the latest trend is to simulate actual stratification conditions
by the application of "equivalent materials" in the models produced. This is to
reproduce the same deformation and strength properties (including thickness,
 -=-= - - =~ -= - - ----=--

ROCK PRESSSURE MEASUREMENTS 243

inclination, etc.) on a proportional scale to natural conditions. In this way quali-

tative information may be obtained on the influence and interaction of various
factors such as shape and method of excavation of the cavity, the relative thick-
ness, succession and strength of overlying and underlying strata, etc.
Model tests of this type are rather advanced in mining engineering with special
regard to the most economical and safe exploitation methods.
For the above reasons increasing attention is being devoted in rock pressure
determination to the in situ measurement of rock pressures performed either:
(a) in exploratory drifts driven ahead of the tunnel face, sometimes over expe-
rimental sections, or
(b) in the lining of the completed tunnel, or in the surrounding rock. These
observations are of a long-term character.
Both methods are of significance for the design and construction of tunnels,
although information available before construction is undeniably more valuable.
Measurements may be aimed at the determination of the so-called static stress,
lf undisturbed mountain pressure prevailing before the tunnel is excavated, or of
the stress in the disturbed surrounding of the cavity.
The methods by which these measurements can be performed have been sum-
marized by TALOBRE as follows.

3.61. MEASUREMENTS ON THE ROCK FACE

This information relates to stress conditions prevailing at the rock surface

"'hich must not be mistaken for the natural stress condition existing in the interior
!" the rock mass prior to excava-
tion. Little is known about the re-
lationship between the two states
of stresses and the former is not
indicative of the latter.
The basic principle underlying
these so-called mechanical meth-
ods as used at present is the
..relaxation of stresses". Essen- Extensometer
tially, the compression prevailing Strain gauge-
in the rock is relieved by a cut and
the ensuing deformation is ob-
ned with the help of a strain
gauge.
The oldest of the methods aim-
mg at full stress relaxation is the
one outlined in Fig. 3/58, where
an cxtensometer is mounted be- FIG. 3/58. Arrangement of pressure measurement
l ttn two fixed measuring points, on a bare rock-face
244 ANALYSIS OF LOADS

or a resistance gauge is applied on the smooth and dry rock surface. Stress relieving
cuts are then made beyond the limits of the measuring device by sawing, chiselling,
or boring. Strain gauges arranged in a rosette yield results from which the prin-
cipal stresses can be determin~d.
Partial stress relaxation has been used more recently, which is similar to the
method used with metals, and is suitable
for measuring stresses in hard rock.
fxtensometer Ex'tensometers or strain gauges are here
arranged in a delta configuration {Fig.
3/ 59). A stress relieving borehole is then
drilled at the centre. The borehole must
be small enough in comparison with the
measuring length of the device for the
stress reduction produced not to exceed
one third of the original value. In this case
the behaviour of the investigated zone re-
/{KJflllTl mains completely elastic and the actual
modulus of elasticity will closely approx-
FtG. 3/59. Stress relieving borehole
imate the value obtained from a sample .
The accuracy of the method can be im-
proved, by gradually deepening the borehole meanwhile continuously observing
deformations. Deformations thus obtained are then plotted against depth, and
inferences as regards local irregularities or disturbances can be drawn from the

I
I
I

I I/
I
I
I
I
I
Of----0 I
Stress gauge arrangement I
in elevation I
L ______________ J II
L ________________ J
FIG. 3/60. Pressure measurement on the principle of pressure reestablishme;it
 ROCK PRESSURE MEASUREMENTS 245

discontinuities of the plot. At a side length of 200 mm satisfactory results have

been obtained with 56 mm diameter -:boreholes drilled with very fine diamond-
tipped bits.
The third surface method relies for its operation on the principle of stress resto-
ration. Tensioned wires are first mounted on the intact rock surface (Fig. 3/60),
and their vibration frequency is measured . In the second step the rock is stress-
relieved by a cut at one side which results in partial elastic stress relaxation. A flat
Freyssinet jack is then placed in the slot and fixed with mortar. After the mortar
has hardened the jack is operated until the original vibrating frequency of the
wires is restored. The necessary jack force and, thus, the stress is then recorded.
This method was found successful in softer rocks. A modified version was devel-
oped by BERTHIER, who used an annular slot around a 250 mm central borehole.
Stresses were restored by radial jacks placed into the slot and recorded with
a special microscope. A drawback of this method is the full relaxation of stress
which leads to questionable results in view of the doubtful reversibility of stresses
in rock.
In order to combine the advantages of partial stress relaxation obtained with
a central borehole with those of the method using stress restoration,a cylindrical
jack was recently introduced which fits into the central borehole of 56 mm dia-
meter and is capable of exerting pressure in two diametrically opposite directions.

3.62. MEASUREMENTS IN THE IN TERIOR OF THE ROCK

The aim o{ measurements in situ depth is to determine, within the attainable

limits of boring, the original pressures prevailing in the interior of the rock. The
necessary equipment and method of measurement is, however, expensive and the
method itself is not without limitations.
(a) For the method based on full stress relaxation, a cylindrical device, the so-called mov-
ing star, has been developed by BERTHlER. Three extensometers taking measurements in three
different directions are equipped with six elastic feelers. The instrument can be lowered into
a 76 mm diameter borehole: When in the required position, the feelers are fixed and readings
are taken. The rock around the instrument is slotted with a 200-250 mm diameter crown
drill. Readings taken with the feelers now in the same positioii correspond to the fully stress-
n:lieved condition. Feeler positions at the bottom may be recorded either photographically
or by printing, depending on the type of equipment used. Besides the uncertainty with which
u11tiaJ and final positions are indicated, the impossibility of determining correctly the mag-
nitude of the deformation modulus in the case of. .full stress relaxation (see above) is a
senous drawback to this method.
<b) The second practical method is based on the principle of stress restoration. A 185 mm
diameter borehole is driJled first to the depth where stresses are to be detected. The instrument
lowered and fixed at the bottom of the borehole. The extensometers in this position are
ad;usted to zero. In the second phase of partial stress relaxation a. 56 mm diameter borehole
drilled in the axis of the first. The third step consists of lowering a cylindrical jack into the
crntral 56 mm diameter hole, operating the jack and taking readings before and after final
pressure is applied. Essentially the instrument perll)its the application of surface methods
the interior of the rock, but requires a borehole of considerable diameter.
 246 ANALYSIS OF LOADS

Instead of jacks, pressure cells or membrane gauges are also used (Section 36.6; Carlsson
or Philips cells), in which the functions of pressure exertion and deformatio n measurem
ent
are combined. The applicatio n of these devices appears, however, to be limited to softer
rocks
only.
(c) Attempts have been made to correlate the velocity of propagation of sound
waves
introduced into the rock with the magnitude of stress prevailing in it. Velocity
was ob-
served to increase with stress. It will be noted from the diagrams published after HABIB
and
DoWANCE 3· 07 (Fig. 3/61) that a small percentage change in
the velocity of sound waves
correspon ds to a change in the compressive stress of several hundreds of kg/cm',
so that
the method is of limited accuracy and suitable for rough approxima tion only.

Dolom,te
I
10"/4 f - - -- - + - - - --L-'--- --+-- -----. ---..._ .- = - - ~
-·

100 200 300 400 500

ComJ)l'ession stress (kg/cm2)
FIG. 3/61. Stress determina tion on the basis of velocity of sound propagatio
n
(HABIB and DOWANCE )

3.63. MEASUR EMENTS ON TUNNEL SUPPORT S

Informat ion directly utilizable for tunnel design is supplied by methods indi-
cating the forces acting upon the proposed lining·an d so leading to its reliable
and economic al dimensioning. These rely mainly on measurin g the deformat ion
of timber and steel structures used as temporar y supports . A wide variety of
methods has been develope d - especially in mining - for installing different types
of dynamom eters on ordinary supports.
Measurem ents are carried out either prior to construc tion in pilot headings
driven ahead of the main face, or after construc tion on the complete d structure
for control purposes , or for statistica l records. In the majority of cases only defor-
mations have been measured hitherto; direct pressure observati ons became pos-
sible only on the more recent introduc tion of new measurin g devices (pressure
cells).
3 · 27 HABm, P.: Determina tion du modul~ d'elasticite des roches en place, Ann. I.
B. T. P.
1950 Sept.
 --- -- --- - -- - - .

ROCK PRESSURE MEASUREMEN TS 247

Tunnelling is usually started by driving a pilot heading in which measurements

can be performed conveniently. Target marks are established at about 20 m
intervals on the timbering, expediently by driving nails into it (Fig. 3/62). The
horizontal and vertical position of these points is observed in the first phase of
construction weekly, later fortnightly, or
at monthly intervals. Measurement can
be speeded up- by a suitable measuring
frame composed of simple elements. The
frame is set up in the plane of measure-
ment; it is wedged tight between the roof
_-2.L.._
and floor beams and the central strut is
positioned on the centre line. Its vertical 2
position is then measured with an in- 7
strument. The feeler arms are adjustable,
with coarse and fine adjustment, over a
rack and pinion mounted on the central Fm. 3/62. Simple deformation measuring
pole. The feelers are then brought into installatiol'I in an advance
heading
contact with the target point, and the ver-
tical and horizontal position of the arm
is recorded. Once the position of the
central pole is known, both the vertical and the horizontal positions of the points
can be determined in an absolute manner. With the frame set up, readings can
be taken by several observers simultaneously.
For evaluating observation data the movements of individual points are plotted
in a diagram (Fig. 3/63). The period after which equilibrium in the rock is restored
- following individual stages of construction - is distinctly recognizable from the
graphical representation. Increasing displacements during construction give timely
warning of growing pressure, permitting counter measures to be taken.

F10. 3/ 63. Measured

deformation
values as
a function of
time and of
individual
operation stages
 248 ANALYSIS OF LOADS

Recorded deformations are compared with information obtained by geological

exploration (which is completed as the heading is driven), samples are taken from
the rock or soil layers penetrated, and laboratory tests are run on them. In obser-
vation of actual stress conditions samples should be tested by increasing loads
slowly and maintaining each load increment for a considerable period rather than
by rapid load increase. In the elastic stage of deformations the elastic limit should
be determined, then the compressive, tensile and shear strengths, as well as the
ratio of vertical and horizontal deformations should be noted (cf. Fig. 2/10)
F.inal decisions on construction method and lining dimensions can be taken on
the basis of measurements in the pilot heading, observations of mountain structure
and investigations of samples taken.

3.64. MEASUREMENT OF PRESSURE CHANGES IN COMPLETED LININGS

For the measurement of pressures acting on completed tunnel linings, use can
be made of rod extensometers arranged radially, or in a star-shaped configuration,

I or pressure cells can be installed between the rock and the lining, or electrical
strain gauges can be mounted directly on the tunnel lining. Use can also be made
of a variety of other mechanical and acoustic devices.
Rod extensometers in radial arrangement are shown in Fig. 3/64. Rod ends

fixation
Jnvar rod

Pressure cell

Acoustic cell

[lectric cable
F10. 3/64. Radial measuring
feeler arrangement

are telescopic and displacements (o) are measured by acoustic measuring cells,
transmitting the results to a central electric recording instrument. If pressure (p)
were distributed uniformly over a circle of diameter D and of unit length, and with
E and µ denoting Young's modulus of elasticity and Poisson's ratio respectively,
the pressure could be determined from the expression

p = D (1 + µ) .
 ROCK PRESSURE MEASUREMENTS
249

s
Taking into account the presence of a lining of thickness v and having a modulu
eters, the
of elasticity Eb, between the surroun ding soil and the set of extensom
pressure is approximately

LI
Instead of the radial 'pattern , extensometers can also be arrange d in a double
or standing i.e. in star-sha ped configur ations. In complet ed
with its apex either lying
tunnels this arrange ment causes ihe minimum interference with traffic.
Methods described in Section 36.1. and suitable for the measurement of stresses
in natural rock faces can also be used. The only difference will be that devi~es
pres-
attached to the internal surface of the lining record stresses instead of rock
bedding conditio ns on the
sures which, however,reflect the effects of strength and
lining (voids, and grouting, drainage, etc.) as well.
Pressure boxes or cells of various types are available for the direct measurement
of rock pressures. These are usually mounted on the external face of the lining.
The method is recently gaining in popularity, especially for underground railway
ad
tunnels. Measurements of this type were carried out on the Moscow, Leningr
currentl y taking place on that in Budapes t.
and Chicago undergrounds and are
encount ered in the evaluati on of observa tion data,
Serious difficulties are still
since besides the elapse of time, the development of pressures is seriously affected
the
by the proper and firm installation of gauges, by all changes produce d in
on
s•moundings and by a variety of other factors, the identification and separati
of which is practically impossible.
The installation of pressure cells under the invert of a tunnel is shown in Fig.
3/65, while their arrangement in a circular tunnel is illustrated in Fig. 3/66. Obser-
vations conducted in Chicago soft clay led to the following conclusions:

---
H. Tube ,.
,, I. Tvbe

N---
OJx undertzv
/Jelow rib

Arrzn;emenl ofmusurtn; ,mtrvmenu

)
Fm. 3. 65. Arrangement of pressure cells below bottom plate of tunnel (Chicago
250 ANALYSIS OF LOADS

I. The ten-year observation record revealed that final equilibrium in the soil
did not occur until five years after completion.
2. Pressure changes showed two basically different trends, inasmuch a~ they
decreased during the first three months, indicating that surroundin g soil

Arrangement scheme ofpressvre cell,

St,at,on 8 f 8U Sf.:Jtion 10+6,2

Sf.:Jt1on !0,9J
Summan?ed /ocat,on scheme
FIG. 3/66. Location of pressure cells in the circular tunnel section

layers were only gradually affected by subsequent soil movements. Thereafter,

vertical pressure increased for several years until static equilibrium was
obtained and soil stresses increased to the limit soil resistance, determined
also by independent shear tests.
3. Lowest values were observed at the end of the first phase when stresses
became practically uniform around the entire circumference of the circle.
4. In the final equilibrium condition vertical loads attained the magnitude of
the geostatic pressure, while lateral pressures, which showed small variation
during the observation period, remained appreciably below the hydrostatic
value.
 ROCK PRESSURE Ml!ASUREMENTS 251

:J,/1
~ p;
. :::: ---
~--- --- -- --·---- --
I I
,.St3 1c head at 110 _ _ fl
4.5 p'
1
St;Jt1c heJd at H
4,0
,r
-
---
~ StJt1c llea'd at H.
-~, - Pv - .-~
J.5
... -

J.O
-:--.
~
-- ~
Jt H~~-ri,..r ..
J
./~
~
,--- . . . · 1- -- -- -- ,1,,_ -···. -
, >--- -~
i
- j~ -1?! - -
0 ~I ~~ ~
~ ~
!:'"'i
~
"'
~
~ I~ - -
~ ,,...._ it-~ - ;~ ~- "'
<..:
""
~
....
~-~ <:--;-

~
"" - ~ ~~ ~ '-; ~ ~
- ~
'---~ - It-- 't:: ~-
~
~

-~ ~ ~',~-
~ ,~~-~
S::il~-..;
~ ~
"'
~I
~- - - --
1,.J :c..,ijs ~ ~
~ic!::~ ~ i I
- ~ - ~

0
(; ·i I I - -
O 10 20 JO 4() 50 60 70 730
Time (Months)
I.. 1eJO .. I.. 1931 .. I.. tr112 .. 1., 1911 .. I.. 19J4 ,., I.. 191s .. w16 - 1940., IJ94'
Time (years )
FIG, 3/67, Variation in pressures with time (plastic clay, Chicago)

The development of pressures is plotted against time in Fig, 3/67, demonstrating

that vertical pressures in the plastic clay soil even at relatively small .overburden
depth attain the value of the geostatic pressure after a certain period, while owing
to cohesion lateral ones remain appreciably lower.
Observations included changes in the horizontal diameter of the tunnel as well
as the deformations in tunnel shape (Fig. 3/68). The horizontal diameter was
found to decrease as long as the driving of the second tunnel did not reach the
cross-section investigated. Later, it increased showing a decreasing rate in time
as long as the air pressure in the tunnel was maintained. From the moment when
compressed air was shut off, the rate of increase again became steeper to flatten
out again after a further three months onJy (Fig. 3/68a). Eventually the deformed
shape of the 7·5 m diameter tunnel lined with cast iron elements changed from
that under (b) to the one under (c). Deformation and stress observations thus
revealed the considerable influence of various factors (adjacent tunnel, pneumatic
support) besides that of time.
The final conclusion derived from these observations by TERZAGm3·28 was that
lining walls in plastic soils should preferably be constructed as flexible and as

s.:a TERZAGHI, K . : Shield Tunnels of the Chicago Subway, Journ. of the Boston Soc. of
Ci11. Eng. 1942 3
 252 ANALYSIS OF LOADS

bJ
I Compressed air
5 7 I I
,, Erection of/irst tube / Concretmg of
Slite!d of,seco.r10· tube amve.s ...-
I
'- first lube (Ir.st tube
',<I 7 T
,,,..----, './
0 nl..x
·1 Diameter change ofsecond tube
,/'1 I. I I :

) I
5v
Coocret:t of.secollr tube

5
100 ?00 JOO
I
400 600 700 800
I Erection ofsecond tube Time "days)

I
I
Deformatioo of cast ,roo Jmmg 47 Deformation afterJ42 days
days after erecL!on
FIG. 3/68. Progress of deformations of the circular tunnels of the Chicago Subway

late as possible. This is in full agreement with the remarks made in Section 3.12
in connection with genuine mountain pressure.
As revealed by careful deformation and pressure observation conducted in
recent years in the circular tubing-lined tunnels of the Moscow and Leningrad
underground3 ·w and elsewhere the stresses developing in the tunnel lining are
frequently affected more by other circumstances than by rock pressure itself.
Concentrated initial eccentric stresses leading to high bending moments and grow-
ing deformation of the wall may thus be induced especially by inaccurate elliptical
erection of segment rings and by improper grouting and backfilling. Cracks and
stress peaks were ·noted to occur especially at the crown and in the elements
immediately adjacent thereto.

3 9
·~ GELMAN, G. and BODROV, B. P. : Naprashennoe sostoyanie sbornikh tonnelnikh obdyelok
metropolitena (State of stress in prefabricated linings of underground railway tunnels), Sborn.
Statei 31
 --- -- -- - = ;:c - ---= =- - =-=--- - -

ROCK PRESSURE MEASUREMENTS 253

The horizontal diameter of the circular tunnel in clay of the Carboniferous Era
was found to expand at a rate decreasing considerably with time (Fig. 3/69),
while normal forces were observed to increase more rapidly than bending moments,
indicating that the rate at which lateral pressures increased was higher than that
of vertical ones. No appreciable load was transferred from the parallel tunnel

If?. .,_. i.---- --

L-
.
i-::-_:-r..::. .-- ·---
--
.. -· ---- .. -
If
10 ·"-~..... -- --
- --~

·-
_-;:: t~. I=
l.--c::
.... !l
~--; _,::. ;:::...
,;,'
~
~ 8 ---
-~~lnnertdge ofsegment _ - - Oerfom3tton curve of r111g N! 55 .__
~ 7 -· b<" /,_. -
~ '·-':. ~>,: ~ ri/J 1 ·- -- Oerform,tion rvrveofri,,_q N'57
..
~6
~ 5
....... ,
i~ 1/ Inner fece ofsegment 1
p/;te 2 --- Deform.lion curve ofring N'56 .__
Rem;rk, S1mJ/;r observ.t,ons were m;de .__

I <i>\ ~-
on the other rings
~
Jl
4
J
f?.
,' f)
- I
I

1
J
0
I
60 tPO f8Q f40 360 420 480
lime {d,1ys)

FIG. 3/69. Progressive increase of the horizontal diameter of circular tunnel

(in Carbonaceo us clay)

as long as the connecting bolts of lining segments were not tightened and per-
mitted relative rotation between them. Otherw1se the parallel tunnel caused an
overload of almost 50 % on the stress diagram. In the case of careful construction
no significant differences between calculated and actually measured stresses could
be observed.
Grouting between the rock and lining is of considerable significance for the
distribution and development of pressures. As can be seen from Fig. 3/70, the
stresses in the rock behind the lining can be restricted by grouting to a much
narrower range and the peak stress occurring at the interface can be materially
reduced which is essentially a consequence qf reduced pore volume behind the
lining.
The effect of imperfect grouting on stress distributio n around the tunnel is
shown in Fig. 3/71. Averages of test results obtained at an undergrou nd gas- 3 10
container with a 6 mm thick welded steel-plate lining are illustrated here. ·
An internal air pressure of 15 atmospheres should have caused a uniform circum-
ferential stress of 1134 kg/cm2 in the lining. At the crown and invert, where owing

- Kov.-lcstt.-lzv, F.: Obudai barlanggaztart6 (Cavities as gas containers), Epites- es Kozleke-

destudomdny i Kozlemenyek V 2
 254 ANALYSIS OF WADS

to imperfect grouting no firm

.·.: Lmmg contact with rock was estab-
lished, peak stresses far in
excess of the average were ob-
Grouted cavtly
served instead. Apart from ini-
tial stress differences due to
W/tl, groulmg imperfect grouting, further
Without droutmg definite discrepancies were
observed between the rates of
growth of roof and invert
stresses. This, however, may
be interpreted as an indication
of differences between vertical
and horizontal rock pressures
in this particula r rock. The
'without grouting ratio of the slopes of lines
Void volume
Wtlh qroutmg represent ing increase appears
~¾ to be suitable for determining
]
................u..Ju...L.........u...L.L.J..1...1..LL.L.L.L.L..L...
the lateral pressure coefficient,
since pressures prevailing after
FIG. 3/70.. Change of pressure distributio n produced
by grouting in surroundin g rock the inaccuracies of grouting
are eliminated, may be re-
garded as 'balancin g' the ac-
tive rock pressures. The slopes
of the two straight lines were
found to yield the coefficient
A.= 460/ 1200 = 0·382, which
Average experience has proved to be
11341<,_q/crrl acceptable for the dolomite
520 5Jf rock found on the site.
A general conclusion from
in situ rock pressure measure-
ment appears to be that pres-
sures around the tunnel are
1524 distributed continuously and
F1a. 3/71. Tangentia l stresses in steel-sheet lining of decrease from the peak value
embedded circular gas-container distorted at the crown towards the
owing to imperfect grouting
springings (cf. Figs 3/68 and
3/71). Stress calculations
should, therefore, preferably be based on a continuous radial pressure pattern rather
than on the combination of uniformly distributed vertical and horizontal
pressure diagrams. The resulting difference in the design bending moment
is considerable, inasmuch as the moment in a circular section due to uni-
formly distributed vertical and horizonta l loads is M = ±0·25 q (l - 1) r2,
 ROCK PRESSURE MEASUREMENTS 255

with q being the vertical and )..q the horizontal load intensity (cf. section 43.22).
On the other hand, in the case of continuous radial distribution we have M =
= ±0·15 q (1 - ).) , 2• The difference amounts to 40%.

3.65. DETERMINATION OF PRESSURES BY MODEL TESTS

The numerical determination or check of real rock pressures on the basis of

model tests is very difficult because the strength and deformation characteristics
of the natural layers built with a true stratification but on an appropriately reduced
scale in the model will undergo the stresses and deformations brought about by
the excavation of the model cavity on a different scale.
In order to obtain correct and characteristic deformations and stresses in the
model, the deformation and strength-properties of the applied model materials
should also be changed in accordance with the model scale. Recent investigation
is therefore more and more governed by the trend (chiefly in mining engineering)
to apply so-called 'equivalent materials' in the models which should be able to
afford a correct reflection of the pressure and deformation conditions brought
about by the excavation of the cavity owing to their altered streng.t h and defor-
mation properties changing in accordance with the requirements of the scnk:
model.3.3la
The ideal equivalent materials should possess similarity of modulus of propor-
tionality and of stresses and strains in their resistance to compression, tension and
shear i.e. they should display a similarity of the Mohr-curves. To imitate sand-
stone, or sandy- and clayey-shales, mixtures of gypsum and sand, whereas for
clays and slates mixtures of sand and paraffin are used. It is not only the similarity
in these qualities of the material but the similarity in the stratification and structure
of the rock which is of primary importance. The main point for which it is gener-
aJly striven for is the construction of models, in which the factors influencing
pseudoplastic deformations and their extension may be reliably observed in order
to conclude uoon the processes, which are about to display in Nature. Eqviva-
lency in this sense was not an object of recent tests carried out prior to the
construction of the Donnerbtihl tunnel in Bern, Switzerland which may also be
mentioned here.3· 31b
Here the factors generally involved in rock pressure expressions were found to be

p =yHA + cB,

where H the geometrical height of the loading rock mass

y = its density
uia Cf. KuSNETsov, G. N. : Investigation of Rock Pressure Phenomena on Models,
Ugletehisdat, Moscow 1958 (in Russian)
:J.3ib Cf. the series of articles in Schweizerische Bauzeitung 1960 III 10-31
 256 ANALYSIS OF LOADS

A and B are dimensionless coefficients whose magnitude depends, in general.

on <p and c (the shearing strength of the rock), as well as on the ratio of vertical
and horizontal dimensions of the loading rock mass.
By adopting y, <p and c values in the model identical to those in the prototype,
the first term of the expression is obtained correctly in the model by multiplication
by the scale factor. The second term has the same value in the model and in the
prototype.
In the case considered, the results observed in the model at small overburden
depth were equal to the geostatic pressure and approached the value calculated
according to TERZAGHI's theory as soon as the overburden depth exceeded the
tunnel diameter.
Lateral pressures on the other hand were found to be higher than those obtained
by RANKINE'S widely used reduction which is explained by the fact that owing
to the small gaps and limited deformation opportunity involved in shield driving
no shear resistance could develop on the sliding plane. 3 · 32
Model tests performed earlier by Y AMAGUT1 3· 33 also demonstrated that the

(
I
difference between roof and lateral pressures is usually smaller than RANKINE'S
heoretical value .

3.66. MAIN TYPES OF PRESSURE MEASURING INSTRUMENTS

A section through the Philips-type pressure cell (plate) which is also used in
Hungary is shown in Fig. 3/72. This meas uring device consists essentially of an
800 mm diameter oil-filled box enclosed by two membranes. Pressure acts on the
large-surfaced (225 mm diameter) external meinbr&ne (A), the defor mation of
225¢

"""""-"'~11
,.____ _ _ _ __ _ D 8IJIJ¢ I
Fm. 3/72. Section of " Philips" pressure cell A , Ovler membrane
8 , Oil space
C , Inner membrane with ex/ensomefer band
D · Compensating band

3· 32 cf. HOUS KA, J.: Beitrag zur Theorie der Erddriicke auf das Tunnelmauerwerk, Schweiz
rische Bauzeitung 1960 38 and see Section 3.33
3 · 33 YAM AGUTI: On the investigation of stress distrfbution in a tunnel. Abh . 3. Int. Kongr.

Techn . Mech. Stockholm, 1930 2 150

 ROCK PRESSURE MEASUREMENTS 25 7

which is transmitted by the 1. Cover piate 6.il. Supports of

oil in space B to the smaller 2. Load transmitting registration bean•
spherical pin 7. Electromagnet
internal membrane C. Defor-
3. Registraton beam 8. Cable
mations of the latter are re- 4 . Encasament 9. Cable inlet
gistered with high sensitivity 5. Bottom plate 10. Steel string
by an electric stra in gauge
mounted thereon . The strain
gauge consists of extremely
thin copper wire of which sev-
eral waves are glued on pa-
per. Any change in the length
of the gauge induces a change
in its total length, the ratio of
tbe two being equal to the
number of full waves. Slight
changes in the gauge length
thus give rise to perceptible
changes in wire resistance
which can be measured by
changes in the strength of the
small direct current passing.
Changes in the current in-
duced by electromagnets sus-
pended on a stressed wire are
registered by the MAIHAK-type
pressure cell. The stress in the
wi re changes with external
load (Fig. 3/73).
Devices relying on vi-
brating wire- and acoustic prin- FiG. 3/ 73. MAIHAK's pressure cell
ciples are also widely used.
Changes in the frequency of
a stressed wire due to its extension are registered by the vibrating-wire
device.
These acoustic strain gauges provide a very sensitive means of measuring strain.
They depend essentially on the variation of frequency of a tensioned wire with
strain. The transverse frequency of vibration of a tensioned wire is governed by a
well-known physical law which states that the transverse vibration of a stretched
\\ ire is proportional to the square root of the tension in the wire. Thu~ if the
tensio n is altered in the wire, subsequently being strained by it, then the frequency
of its fundamental vibration will be altered accordingly. In practice a very small
(0·0 l in) diameter silver-plated steel piano wire is clamped to that region of the
structure in which strain is to be investigated and its frequency is recorded, sub-
sequent strain in that region of the structure is transmitted to the wire via its
Szechy: The Art of Tunnelling
 _' ----=-=- - ~

258 ANALYSIS OF LOADS

rigidly clamped ends, resulting in a change in frequency of the wire, which is

again recorded. The difference between these recorded frequencies is an obvious
measure ofthe alteration in strain of the wire and hence of the structure to which
it is attached.
Many types of this device have been developed in various countries among
which those of COYNE, MAIHAK and Road Research Laboratory {The Engineer
Dec. 1958. p. 964) may be mentioned. It is especially applicable for long term
tests and this makes it very suitable to rock-pressure measurements.

The RANK1-type oil-cushion membrane pressure ·cell, the operating principle of

which is illustrated in Fig. 3/74, was developed at the Technical University,

finin; Glgcmne
segment Ovter membr,1ne Membr,1ne

Deflection dli/ wit//

1/IOOOmm e<.1cltl11de

FIG. 3/74. Directly registrating oil-cushion pressure cell of RANKI

Budapest, in connection with the construction of the Budapest Underground

Railway. Pressures acting directly on a larger external membrane are again trans-
mitted to a smaller one, the deflections of which are indicated mechanically by
a dial gauge reading to 1/100 mm accuracy. Highly sensitive electrical equipment
is thus dispensed with, although the possibility of tele-repbrting is sacrificed
(direct readings are only possible).
Pneumatic pressure cells, which are the most popular in American
practice (GOLDBECK, CARLSSON, etc.), operate in the following manner (Fig. 3/75).
External load acts on a 0·5 mm thick external membrane bearing on a concentri-
cally located contact surface of 6 mm diameter. As the vessel is filled with com-
pressed air, contact is interrupted at the moment when the internal and external
pressures are balanced. The interruption of the electric circuit can be recorded
with high accuracy. Dry compressed air and current are introduced on one side
of the cell, the other lead being on the opposite side. The sensitivity of the
 WATER PRESSURE 259

pressure cell can be adjusted by a

screw at its base.
Aside from pressure cells,.measuring
struts and jacks operating on hydraulic
principles and built into the lining may
be used fo r directly registering the pres-
sures or deformations therein. Their
more detailed treatment would, how-
ever exceed the scope of this book.3· 34 ' ...........
Current outlet ~ Svpporting ribs of
I steel membrane •
· Copper packin;
l'res~n
3.7. WATER PRESSURE

Tunnels, where they are below the

groundwater table, are exposed to the
external pressure of water in addition 'Prespan
to external rock or earth pressure. This
load must be taken into account mostly F10. 3/ 75. CARLSSON's pressure cell
in the case of tunnels built in the upper,
geologically young formations, i.e. especially for underground railway and sub-
aqueous vehicular tunnels where the major part of the total load must be
attributed to it.
In contrast to rock pressure, external water pressure usually acts according to
the full hydraulic head without any loss, especially where in the interior of the
mountain a natural system of communicating caves exists which is bounded below
by an impervious layer.
In the case of such an intercommunicating system the magnitude of water
pressure is at every point equal to the hydrostatic head in a direction perpendicular
to the loaded surface. Such conditions are encountered primarily in sedimentary
rocks, while freely percolating water in volcanic rocks occurs only in the rarest
instances. Yet the development of full hydrostatic head in the interconnected
fissures of these rocks is by no means impossible. For the lining of pressure tun-
nels this effect may even be advantageous. On the other hand , in headings with
no permanent lining such 'weeping water' under pressure may loosen and dislodge
pieces of rock from the roof and from the ·side walls. Water flowing with a free
surface will thus be less dangerous in similar cases. A value lower than the full
hydrostatic head can be taken into consideration as a rule for temporary under-
ground structures only (shield chamber faces, temporary head-walls), provided
that the soil penetrated is of low permeability and the possibility of the percolation
..:i , HAMILTON J. J . : Earth Pressure Cells (NRG. Canada), Rep. No 6009 1960.
J ACOBI D . (1958): Instrumentation for Rock Pressure Research, Colliery Engg. Vol. 35, p . 81.
'-1oHR F. ( l 956): Measurement of Rock Pressure, Min. Quarry Engg Vol. 22, p. 178

11•
260 ANALYSIS OF LOADS

of groundwater from upper water-bearing layers along fissures to the structure is

nil. In some instances water may percolate only very slowly towards the tunnel
lining through silty sand veins a few millimetres thick, embedded in fissures
between clay layers. On the one hand, a very long time is necessary for the ex-
tremely small flow to build up to full hydrostatic head and, on the other, a very
slight leakage or seepage through the lining is already sufficient to drain off the
water and reduce water pressure. In similar soil, interwoven with thin permeable
veins, water pressures considerably lower than the hydrostatic head were observed
even after several years during the construction of the Budapest Underground
Railway (cf. section 63.24). In similarly stratified soils the distribution of water
pressure around the Section was found to be non-uniform. No water pressure
acted at points where the lining was bedded tightly against the impervious layer,
but even the full hydrostatic head prevailed on the faces exposed to water-bearing
veins. Thus it is conceivable

- --¾=--
that water pressure will not
act on the entire surface be-
cause water has no access at
VGWL. all to some parts.
a) Perviovs lining b) lmperv10L lininq In view of the foregoing,
t tunnels need not necessarily
be designed to resist the full
hydrostatic head where the
surrounding impervious soil is
fissured, or woven through by
m permeable veins small enough
to ensure a certain degree of
sealing, admitting no more
water to the lining than it is
capable of draining continu-
ously through its pores and
Fm. 3/76. Water pressures acting on other discontinuities, or which
(a) pervious and can be evaporated by venti-
(b) on impervious tunnel linings
lation. The great advantage of
careful drainage is once again
demonstrated, inasmuch as not only will the excavation be drier, but the load is
also reduced and the lining is subject to smaller forces (Fig. 3/ 76). It will be
readily noted from the figure that the critical water pressure on the tunnel lining
will develop under static conditions, while seepage pressure maintained by con-
tinuous drainage will be least harmful. This condition must, however, be en-
.sured by providing suitable drainage structures, or sometimes artificial means
such as pumping must be employed. It should be remembered at the same
time that upward seepage may reduce the load-bearing capacity of soil under
the tunnel ; it may lead to boiling resulting in increased settlement of the tun-
nel. In this respect the situation is analogous to that in the experiment
 WATER PRESSURE 261

.Sand surface

b)

Overflow

. ;,
Is -

[ H
1cril. ;;; 1

Inflow

FIG. 3/77. Effect of seepage and danger of boiling on tunnel floor in sand (TERZ.AGm)

shown at the right-hand side of Fig. 3/77, where weight W settles suddenly into
the soil as a result of upward seepage pressure as soon as the hydraulic
gradient h/ H = i reaches unity. Since this ratio

At/(t + m)

(cf. Fig. 3/77b) is, in the case of tunnels, not lower than unity unless the tunnel
is relatively very high, or unless drawdown is very slight, the inflow of water into
tunnels in sandy soils is permissible only if the entrance of sand particles can
positively be prevented by careful filtering. The advantage of admitting water
at the bottom of the structure will be noted from Fig. 3/77. The cross-section
of drainage should be capable of handling the flow at which the entire grou~d-
water table can be lowered in the vicinity of the tunnel, which then assumes the
role of a large filter drain. Static pressure is thereby minimized, while seepage
p ressure (h 0) is in most cases negligibly small in view of the low velocities and
small water volumes involved. Otherwise, however low the seepage velocity
and small the fl.ow may be, water will gradually accumulate at the lining which
after a certain period of time will be subject to a pressure corresponding to
the full static head.
 262 ANALYSIS OF LOADS

If a tunnel is designed to resist water pressure, the calculations must naturally

include uplift as well, and the density of the saturated soil must accordingly be
determined from the known relationship

y' = (s - I) (1 - n).
Sometimes the tunnel itself lies entirely in an impervious layer through which
no water has access to the tunnel and this layer is overlain by strata saturated
with water. In such cases the load to be taken into consideration is the weight
of the water-logged layers ·which have a density:

y" = (1 - n) s + n.
No water pressure will act on the section.
Besides the direct external load produced, water has secondary effects as well,
such as the reduction of inherent rock strength upon saturation. Consequently,
the stress-relieved core will be larger and the rock load on the tunnel will be the
saturated weight of an increased rock mass.
In the case of pressure tunnels water pressure may represent not only external
loads, but the internal design loads as well, the magnitude of which depend on
operating conditions while its direction is perpendicular to the internal surface
(cf. Section 4.34.).

3.8. LIVE LOADS

3.81. INTERNAL LOADS

Loads due to the weight and velocity of vehicles moving, or material transporte d
in the tunnel are usually negligible in compariso n with the external earth and water
pressures. For this reason internal loads are usually neglected in the design with
the exception of tunnels subject to high internal pressures (e.g. power tunnels,
undergrou nd gas containers, etc.). This is all the more justifiable, as internal
pressures tend to counterbalance external loads and the empty tunnel will usually
represent the design loading case. Exceptions to this may be listed as follows:
(a) Closed rectangular cross-sections of monolithic reinforced -concrete con-
struction encountered with sub-pavem ent undergrou nd railways where the weight
and pressure of earth are relatively small in compariso n with the internal railway
loads, and where direct loads on the bottom slab are no longer negligible because
of the great width of the rectangular section.
(b) Power or sewage pressure tunnels referred to already, where internal pres-
sures may attain orders of magnitude considerably in excess of the external rock
and water pressures. The internal load furthermore induces tensile and bending
s tresses which are critical to the tunnel lining.
 LIVE LOADS 263

(c) Tunnels in extremely soft, loose soils where the latter surround the tunnel
in a liquid state and may give rise to critical local moments as an asymmetric
external load.

3.82. SURFACE LOADS

Loads due to buildings and vehicles on the surface, or to utility lines running
above the tunnel play a role in the case of sub-pavement tunnels only, which run
at relatively shallow depths under the ground surface. In Hungary these loads
must be taken into consideration according to the specifications of the Highway
Bridge Code. In the design of the Moscow Underground Railway these loads were
allowed for by assuming a 45° spread and no dynamic factor (cf. Section 4.15).
lieyond 8 metres overburden depth the effect of these loads is usually insignificant.

REFE RENCES

A,o.,__ (1 951): Int. Fachtagung fiir Gebirgsdruckfragen , Leaben 1950. Urba n , Vienna
A,o'I. (1958): Vonrdge lnternatio11afer Gebirgsdrucktag1111gen, Leipzig 1958-1964. Akad.
Verlag, Berlin
Bericlue der Ldndertreff en 1- 5 des int. B iiros fiir Gebirgsmechanik, Leipzig 1960- 1965.
Ak. Verlag Berlin
HORVATH, J. (1954): K ozetnyomasi elmelctek (R ock pressure theories), Bdnydszati Lapok Feb.
hECKLIN, P. (1965): Felsmechanik im Tunnelbau, Schw. Bauzeitung, 27
Loos, W. and BRETH, H. (1949): Kritische Betrachtun g des Tunnel- und Stollenbaues und
der Berechnung des Gebirgsdruckes. D er Baui11ge11ieur 5
'-EuMANN, R. (1964): Geologic fiir Bauingenieure, Ernst, Berl in
Proceedi11gs of the first Congress 011 Rock M echanics Lish.on 1966, Laboratorio de Engcnharia
Civil, Lisbon
RABCEW1cz, L. ( 1944): Gebirgsdruck und T1111nelba11. Springer, Vienna
ROTTER, D. (1960): Uber die Diskussion ZIii" Entwickl1111g der Gebirgsdrucktheorie11 in der
Soi•iet-U11ion in den Jah ren 1958/59. Akad. Verlag, Berlin
SPACKELER, H. (1931): D er heutige Stand der Gebirgsdruckfragen, Der Bergbau 34- 35
S21LVAOYI, I. (1952): Kozet nyo masi elmcletek (Rock pressure theories), Viilogatott f ejezetek
a: alag1itepites, bdnydszati ri1elyepites korebol (Selected chap ters on tunnelling and e ngi-
neeri ng constructions in mining), MGszaki Kiad6, Budapest
TERZA0HI, K. (I 946): In troduction to Tunnel Geology. Commercial Shearing and Stamping Co. ,
You ngstown, Ohio
fSl\!BARYEVITCH, P. M. ( 1952): Voprosi dacfeniya gornykl, porod (Questions of rock pressu res).
Moscow
 CHAPTER 4

DESIGN OF TUNNEL SECTIONS

It is a generally recognized fact that stresses and deformations measured in

existing tunnels usually do not agree with those predicted on the basis of certain
assumed loads and/or accepted design methods. There are many reasons, of course,
to account for the discrepancies including:

l. Design loads and assumed pressure distributions may not correspond to

actual conditions;
2. The regarding materials and structure of the tunnel lining may not display
the elasticity, rigidity and the other characteristics assumed in the design;
3. Th.e design assumptions made for the calculation of stresses are not quite
accurate;
4. Certain initial stresses may develop during construction that have not been
allowed for in the design.

It follows, then, that it is advisable to use somewhat higher safety factors in

the design of tunnels and that the loads should be considered in every possible
combination.
In the design of underground structures the surrounding rock at first was
considered only an external load and the effect of any elastic subgrade reaction
was disregarded. Monolithic stone and concrete linings were designed as arches
according to the theory of elasticity, subject to active pressures and ignoring any
passive resistance to the deformations as may have been offered by the surrounding
rock. Flexible linings used in shield driven tunnels were treated in much the same
manner. The bearing pressures wery calculated from equilibrium considerations
and assumed to be distributed as under an infinitely rigid body (Fig. 4/ 1).
More advanced design methods were first introduced by 0. KOMMERELL, 0. N.
R0ZANOV and J. HEWETT, in which the resistance of passive earth pressure to struc-
tural deformations was recommended for consideration. In these methods, how-
ever, the monolithic lining was still treated as a composition of arches and walls,
i.e. of separate members even though rigidly connected. Neither the deformation
of the section as a whole, nor the interaction of the deformations of the composing
members was considered (Fig. 4/.2).
KOMMERELL's method (described in Section 42.1) was modified later by RoZAN0V
to include the friction force S2 under the wall base. This, of course, is a prerequisite
 DESIG N O F TUNN EL SECTIONS 265

Vertical pressur~

FIG. 4/ 1. Monolithic
rigid tunnel
section with no
elastic s_u bgrade
reaction

{Jmformlp distributed bemng pressure

fixed

FtG. 4/2. Built-up section h

supported by
passive earth
resistance

!ryr a base whose horizontal displacement is assumed to be zero. In his method

for monolithic circular ring sections (see Section 43.33) HEWETT also included the
rassive earth resistance. However, the distribution of the latter was taken as trap-
ezoidal, a rather arbitrary assumption, and was calculated assuming no change
m the length of the horizontal diameter of the ring. In these design methods the
effect of passive ground resistance was obviously overestimated which, in turn,
rcaulted in lower safety factors.
[n order to improve design accuracy, the Soviet experts DAvrnov, ZURABOV-
Bol:GAYEVA,BODROV' MATERI and GORELIK have attempted to include the interaction
bet xeen the deformations of the monolithic and flexible structure and the elastic
ubgrade reactions. All the assumptions used in these methods were based on the
formula by W1NKLER p = Cy, i.e. on the linear relationship between pressures
nd displacements (Fig. 4/3; Sections 42.3 and 43.4).
It 1s generally recognized today, that over-rigid tunnel linings are not economical
ro build. Lateral support can, and should, be utilized by designing flexible sections
266 DESIGN OF TUNNEL SECTIONS

Vert1c,1/ pressure
mmmmmnmumumnn1
"1tera/ .lclrve
eart/J pressure l1ler,1/ .S!J.OpC!rt
in proportion to
Elastic deformation
s11bgrade Ci
f'Paction

Dav1dov Z(lrabov - Bu§c!JilVa

F1G. 4/3. Loads on a tunnel section with lateral ground support considering elastic deformations
and subgrade reactions

·6

FIG. 4/4. Load reductions with

increasing strains
in plastic ground
E (by MOHR)

which will expand horizontally under vertical loads. Flexibility will also allow
self-adjustments to follow the thrustline and may prevent damage arising from
loads which increase with time.

In his study, Dr. F. MoHR4 · 1 draws attention to the changes in loads as functions of time .
In rocks with plastic behaviour loads on the tunnel section diminish rapidly in the wake of
struct·Jral deformations. This is represented by a descending line (1) in the o - e diagram
(Fig. 4/4) as an indication of the contin·uous decline of the o stresses with increasing strains.
If certain strains (e 2) have developed prior to the commencing of lining operations, the initial
stresses arc reduced to o 2 accordingly. As strains continue to increase the loads will also
gradually increase on the lining (line I). The load governing the design will be that correspond-
ing to stress o 3 as determined by the point of intersection (D) of the two lines. Continued
deformations in the rock, of course, are only made possible through the flexibility of the lining.

4 ' MoH R, F.: Gebirgsdruck und Ausbau, Glilckauf 1952 27-28

 DESIGN LOADS 267

In the case of a rigid lining the loads would be determined by the stresses prevailing at the
time of construction (uz).
Compressibility of Jhe lining and the modulus of elasticity are also contributing factors
to flexibility. The smaller E is, the flatter the line of load transfer becomes (II); the point
of intersection is shifted to D' and the resulting stresses will be smaller (u 3). It must be remem-
bered, however, that the above observations apply to genuine mountain pressures only. The
development of the stress-free body, giving rlse to relaxation pressures and the resulting loads
cannot be reduced by trying to promote such deformations. On the contrary, solid supports
constructed hurriedly will place a limit on the fragmentation and on the propagation of the
zone of relaxation (cf. Fig. 3/47).
Jn this case the advantage of section flexibility lies in the increased lateral ground support.

It is well to remember that maximum loads are not neces5arily encountered in

the finished stage, but may occur during construction as well as some time after
completion. This will also depend on the construction methods employed and
on the type and condition of the rock surrounding the tunnel.
As discussed in Chapter 3, loads on the tunnels include the weight of the over-
burden, self-weight, weight of installations, live loads both within the tunnel and
on the surface and temporary construction loads (such as compressed air, jacking
thrust, etc.). In order to avoid repetition the following section will deal primarily
with specifications for design loads.

4.1. DESIGN LOADS

As mentioned in Chapter 3, this is about t_he most difficult problem because all
possible factors (such as type of soil, construction methods, structural shape and
flexibility, etc.) cannot be taken into account within the practical limits of any
specification. This is why there are no comprehensive tunnel design-load speci-
fications anywhere in the world and why they are compiled for each particular
undertaking on the basis of prevalent conditions.
A number of general design rules are given below and supplemented by speci-
fications for two particular projects.

4.11. ESTIMATED DESIGN LOADS FOR DEEP TUNNELS

IN SOLID GROUND

1. The sliding surfaces in the entrance area extend to the surface (Fig. 4/5).
There is no lateral pressure in materials with high cohesion values and the over-
burden is confined between two vertical planes, i.e. one on each side of the tunnel.
Because of the static earth pressure along these planes the actual vertical load on
the tunnei will be the total self- weight of the overburden less the frictional forces
due to the static earth 'J)ressure on both sides (Fig. 4/Sa).
 268 DESIGN OF TUNNEL SECTIONS

In loose soil and fissured, fragment ed rock material there will be vertical as
well as horizonta l pressures. In this case the overburd en zone will be limited by
sliding surfaces inclined at 45° + </>/2 rather than vertical. The vertical loads will
be equal to p = hy and the diagram of lateral pressures will be trapezoid al
(Fig. 4/5b).

Entrance 1011e

BJ w,lhoul lateral pressure bJ Wttl, liJteral pressvre

e,
e, =17.t-J. -Zc~
3

.-12 -t;Jl f/5°- f/2)

~ '/J-1-m)y.Ji.0 -2c1/X3
FIG. 4/5. Assumed pn:~~ures in the entrance zone

A more accurate estimate of the overburd en near tunnel entrances may be based
on the theoretic al slope, outside of which the ground masses will stand up by
themselves and only those within will have to be supporte d. While the pressure
diagrams will be similar to the previous ones, the forces will have to be reduced.
2. On the way into the tunnel there will be a point where sliding surfaces no
longer reach the surface and the overburd en is limited to a stress-free body con-
tained within an arching of soil. For covers of 30-50 m (100-170 ft) and more,
tunnel linings should be designed for this condition .
If there is no lateral pressure, a parabola or an ellipse of a calculate d height h
can be fitted over the tunnel to represent the burden zone.
If there are lateral pressures to be considered, the sliding surfaces will . start
at 45° + <J>/2 at the bottom of the walls and be extended to the top of the section.
Above this level the burden zone is once again contained within a parabola , the
height of which, h, can be calculate d by a number of methods. The vertical loads
are to be calculate d by dividing the area into narrow strips. In most cases the
parabola is divided into three parts and then evened out; the two dividing lines
are vertical and are construct ed on either side of the tunnel. The tunnel is then to
be designed for the vertical pressures of the centre strip (pJ and for the lateral
pressures as derived from pressure p 2 (Fig. 4/6b).
3. Except for the adit and exit areas it does not make any difference whether
the surface is sloping or not. In this case the diagram of vertical pressures will
be trapezoid al and the la teral active pressures will be unequal (Fig. 4/7a).
 D!!SIGN LOA:CS 269

In solid ground - with no

lateral pressures to speak of - b,
passive pressures must not be Mt/I 13terJIpres;ure
relied on unless there is a dis-
tance of at least 9-12 m (30-
40 ft) between the exterior face
of the walls and the surface.
4. Special consideration
p-nr
must be given to tunnel sec-
tions passing along dividing
lines between two layers and/
or faults (Fig. 4/ 7b). Thes~ F10. 4/6. Rock-pressure assumptions with a great
loads will constitute the critical depth of overburden
sections where tunnel pressure
will not be symmetrical, and
b)
there will be lateral pressures
also in solid rock.
Once again reference is made Limestone
to Section 32.1. according
to which pressures in solid /J111htera/
pressure
rock depend largely on the
condition of the material. For
example, there will be no pres-
sures at all in sound rock that
is free from fissuration. On the
other band, adjectives such as Fm. 4/7. Asymmetrical loads due to sloping
"laminated", "shale structu- surface and to changes in the type of
re", "block formations", "in- ground material
terwoven", "cracked", "weath-
ered", etc. are all indica-
tions of a tendency to detrition and warrant increases in the assumed lateral
pressures. Moisture content bas a similar effect as far as design loads are
concerned, because it will "lubricate" the cracks and will represent an increase
due to its own weight as well as to its effect on the pressure coefficients.

4.12. ESTIMATED DESIGN LOADS FOR SHALLOW TUNNELS IN LOOSE,

SATURATED SOIL

Vertical loads on tunnels close to the surface (H ~ 2·5B) will consist of the
weight of all materials above such tunnels, including hydrostatic pressure if the
water table is above the tunnel. For a ground cover of depth h and a water table t
270 DESIGN OF TUNNEL SECTIONS

above the tunnel the vertical load is

Pc= (h - t) y,(I - n) + t[y,(l - n) + I ·n], (4.1)

where Ys and n are the specific weight and the void percentage of the soil, respec-
tively.
The assumed design load may be bigger or smaller than this geostatic load
depending on construction methods, tunnel location and depth of cover.
Bigger loads may be encountered on tunnels constructed in wide trenches close
to the surface and bearing a certain amount of backfill that is still subject to con-
solidation. The same rule applies to tunnels constructed on the surface and then
buried under an embankment which is again subject to consolidation. The reason
for this lies in the increased settlement due to consolidation of the adjacent earth
masses which, thus, impose additional loads on the more rigid tunnel'section in
the form of negative skin friction (see also Section 43.4).
Loads may be less than the geostatic pressures, however, if arching develops
above the tunnel, either because of the strength of the cover or because of its
depth. In both cases part of the overburden will be carried by the adjacent undis-
turbed ground. Construction methods causing extensive deformations in the
overlying strata will also result in arching and in a reduction of loads on the tunnel.
If, however, such deformations are too small to develop frictional resistance, the
loads on the tunnel will hardly differ from the initial stage. It must be noted here
that, while the phenomenon of arching should be considered in the de~gn of
temporary structures whenever possible, it is regarded much less frequently in
the design of permanent structures. The reason for this discrimination lies in the
temporary nature of arching itself and in the empirical fact that pressures in general
tend to increase gradually, as time goes on, towards geostatic values. This is partic-
ularly true in clay materials (see also Section 36.4 and Figs 3/67 and 3/68).

4.13. EXCERPTS FROM THE SPECIFICATIONS FOR THE DESIGN

LOADS OF THE LISBON SUBWAY

In 1950 design load specifications were compiled for the construction of the new Lisbon
Subway, based primarily on experiences gained in the capitals of the Latin countries (Paris,
Rome, Madrid, Buenos Aires). Although the vertical alignment in general follows the surface,
because of the hilly nature of the terrain, there is a wide range in the depth of cover as well
as in the geological conditions. Tunnel sections passing through basalt were constructed
with classical tunnel driving methods; those sections passing through clay were built in open
trenches.
The specifications for design loads cover both temporary (scaffolding in headings) and
permanent installations (tunnel lining) and differentiate between sections that were driven
and those that were constructed in open trenches. Temporary structures were designed for
3200 kg/ m 2 (655 lb/ft 2 ).
Tunnel sections constructed in open trenches were designed to carry an overburden equal
to the geostatic pressure multiplied by a factor. This was a function of the ratio of the depth
of cover (H) to the tunnel width (B) and also depen(!ed on the type of soil:
 DESIGN LOADS 271

H
0·5 1·0 1-5 2·0 3·0 4·0 5·0 1 7·0 10·0 12·0 1 15·0 > 15·0
B

Granular material
with no cohesion
y = I ·6 t/ m3 (99·9
lb/ft3) ( %) 91·0 83·0 76·0 69·8 59·3 51 ·0 44·4 34·6 25·4 21·4 17·2 16· I

Silty wet sand

y = 1·9 t/m3
(I 18·6 Jb/ft 3) ( %) 92-2 85·2 78·9 72-2 63·5 55-2 49·0 38·9 29·2 24·8 20·1 19·0

Saturated clay
y = 2·1 t/m•
(131'1 Jb/ft3)( %) 94·8 89·8 85·2 80·9 73'2 66·5 60·6 51 •O 40·4 35·2 29·1 28-4

Surcharge loads due to buildings and vehicles were distributed at 45° and replaced by equi-
valent uniformly distributed loads
Surface t/ m!
6·0
H= l m deep 5·0
H = 2 m deep 4·0
H = 3 m deep 3·0
H = 4 m deep 2 ·0
H = 5 m deep I·O
H = 6 m deep 0

Bearing pressures due to subway cars (weighing 40 t) and due to the self-weight of the
bottom slab ranged from 0·36 kg/cm 2 (5·12 psi) to 0 ·13 kg/cm2 (1 ·85 psi). The depth of the
overburden zone carried by driven tunnel sections, h., varied with the surrounding soil type
and was different for the initial stage and for the final condition (after loosening in the rock
had taken place). Maximum and minimum pressures, as specified, were to be combined with
lateral pressures to produce governing design conditions.
In sand and silt materials the depth of the overburden zone was specified in the form of a
maximum and a minimum (h.)

Above water table Below water table

Type of soil Pressure
hp min h
P max hp min
I hP max

Compact initial 0·27 (b + m) 0·60 (b + m) 0·54 (b + m) 1·20 (b + m)

sand final 0·31 (b + m) 0·69 (b + m) 0·62 (b + m) I 1·38 (b + m)
Loose initial 0·41 (b + m) 0·60 (b + m) 0·94 (b + m) 1·20 (b + m)
sand final 0·54 (b + m) 0·69 (b + m) 1·08 (b + m) 1·38 (b + m)
I
where b and m were to be substituted in metres.
Lateral pressures were specified as Ph = 0·3y (0·5 m + h.) i.e. conforming to a trapezoidal
diagram.
 272 DESIGN OF TUNNEL SECTIONS

Provided that there was not too much fissuration in the material, the depth of overburden
1n the prevailing basalt was specified at hp= 0 in the initial stage and at hp= 0·5b or 0·35
(b + 111), whichever was greater. In basalt material that had disintegrated into blocks the
initial overburden depth varied similarly from 0 to 0·6 (b + m) (initial stage) and from 0·35
(h + 111) to l · l (b + m) (final stage) depending on the extent of fissuration.
In clay the overburden pressure was

25
P, =yH- -,_;- · (4.2)

where S the lateral friction of the soil mass surrounding the overburden body and is calculat-
ed from

s = (C + Aa'Y : tan </> l H

In clay materials, where lateral pressures also had to be considered, the factor for such
pressures was taken as 0·7 in plastic clays and 0·5 in clayey soils.

4.14. EXCERPTS FROM THE SPECIFICATIONS FOR THE DESIGN

LOADS OF THE BUDAPEST SUBWAY

The external loads to be considered in the design were specified to include :

LOADS

A. Permanent loads and effects

(a) overburden (earth pressure, hydrostatic pressl:lfe, surcharge loads
due to buildings, etc.)
(b) seef-weight of structure
(c) loads due to installations and equipment
(d) displacement of supports (where working against safety)
(e) shrinkage of concrete.
B. Transient loads and effects
(f) live loads (\ ehicles, pedestrians, etc.)
(g) temperature changes (within narrow limits)
(h) loads due to construction equipment (such as jacking thrust,
pressure grouting, etc.).

Surcharge loads (buildings, vehicles) are to be distributed at an angle of :x 1 = J0° for a

depth not exceeding the half width of the building, below which the angle of distribution
is to be increased to 45 ° (Fig. 4/8). Buildings wider than three times the depth of their founda-
tion shall be considered in the form of a uniformly distributed surcharge pressure equal to
an additional depth of fill. Allowance for impact shall be considered for depths of cover less
than J ·5 m ( 5 ft) and shall then be decreased gradually to zero as that depth reaches 3 m (10 ft).
The live load of subway trains need not be considered in tunnels with a depth of cover in
excess of 7·5 m (25 ft).
Because temperatures are at a reasonably constant level in underground structures, the
effect of temperature change need be considered only in exceptional cases and then within
narrow limits. In such exceptional cases uniform temperature changes shall be limited to
:!: 10 C (18 °F). Uneven temperature changes apply only to exceptional cases such as ventila-
 DESIGN LOADS 273

o//4'/NdSY/&1-frlN~/4\o//4"dSWS-f'o//40¼
I I
I \
I B ' BI t
t "" I· .
I
I· .
I 17

---,.,..,,r--'-- -
""'-
----->
\
\ m

C)

~$ff~~
:•4. I. 4_: l.4 I.B~ !
II I II II I I
I

~~) \ . I l!H.J\ . I
d

FIG. 4/8. (a) Surcharge distribution;

(b) and (d) Computation of governing rock pressures;
(c) Minimum distance between tunnels (Specifications Budapest Subway)

lion shafts and ducts, where a sudden drop in temperature would affect the inside face but
not the outside face in contact with earth. In such cases temperatures of - 5 °C (23 °F) and
+ JO °C (50 °F) shall be assumed for the inside and outside faces, respectively.
Shrinkage of concrete shall not be considered unless causing substantial stresses, in which
case it shall be considered as equal to a temperature drop of 10 °C (18 °F) and 7 °C (13 °F)
for concrete and reinforced concrete, respectively.
There is no need to comment on loads under (b), (c), (d), and (h) except that they should
be taken into account at their actual values and effects.
ROCK PRESSURES

Vertical pressures, to be calculated according to TERZAGHI's theory, are specified in two

groups depending on the presence or absence of groundwater.
A. No groundwater.
(a) In rock material the pressures may be taken as zero, provided that there are no
cracks and faults to speak of, and that there is no stability problem.

18 Szecby: T he Art o f Tunnelling

214 OESIGN OF TUNNEL SECTIONS

(b) In soft clays (consistency index K 1 < 0·6) and in granular materials subject to vibra-
tions (if caused by traffic and for depths less than 3 m=20 ft) and in backfill ma-
terials (provided that the excavated trench is not wider than l ·5 times the width
of the tunnel) earth pressures shall be taken as equal to the full geostatic loads.
(c) Vertical loads are generally less than the geostatic values in soils that are sufficiently
homogeneous and consist of layers displaying little variation in physical properties.
Vertical pressures will also depend on whether the type of soil underlying the tunnel is
solid, in which case the vertical loads on the structure will be reduced, or whether it is of a
loose texture in which case the vertical pressures will increase because of the compression
of the base.
The following dimensions and notations are listed here for reference purposes:

8 = ~ + m tan r45° - ~ l,
where b = the tunnel width (cf. Fig. 4/Sb)
m = the depth of the interior of the tunnel section
,P = the angle of friction of the surrounding ground,
Also, c = the smallest of the cohesion values to be found in the layers above the tunnel
c0 = the cohesion of the layer immediately overlying the tunnel
g = the geostatic pressure at the level being investigated.

In homogeneous material g = ty, in stratified soils g = Et 1 • y,. (If there is groundwater

present, g is the geostatic pressure excluding buoyancy but including the weight of the water
in the ground. Surface loads are not included in g.) Un denotes the geostatic pressure including
buoyancy, h is the depth measured from the top of the structure to the water table (Fig.
4/Sb) and t is the depth from the top of the structure to the surface.
I. In cases where granular materials - such as gravel, sand, sandy and silty fill (max.
IO% silt) - extend on both sides of the structure for a distance not less than b, the vertical
pressure is
- e-V
p, = I·4g - - - - (4.3)
V
where
t
r; = K__:_tan"'
B ,;, •

and K is the ratio of lateral to vertical pressures, varying from .I.a = tan" r45 ° - ~ ) to I ·O.
In lieu of more accurate information K may be taken as I
The factor 1 ·4 is to be used in the final load design.
2. In case; of overburden materials having a certain amount of cohesion (silt, clay) the
vertical pressures are calcuiated as

P, = I '4 g I -
V
e-v . (1 - _c_)
Ya B
• (4.4)

where Ya = the average unit weight of the layers above the tunnel.
3. Surcharge loads (q) distributed over a distance of not less than 3 times the deptii of
cover (3t) shall be added to the above pressures at the rate of
p,. = I ·4q · e-v (4.5'
regardless of groundwater conditions.
 DESIGN LOADS 275

Surcharge loads less than 31 wide shall be distributed as illustrated in Fig. 4/Sa.
4. For an overburden consisting of granular as well as cohesive soils, two calculations have
to be completed.
First calculation: Assume arching in the lower layer and consider the upper layer as a
surcharge:

where
(4.6)

provided that the lower layer consists of granular material ;

or
11
A = 1·4t 2 y 2 -I --e- • (
- 1 --- ,
c )
Vz Ya B

provided that the lower layer consists of cohesive soil ,

and

(the cohesion is to be taken as that of the lower layer).

Second calculation: Assume arching in the upper layer and assume the full weight of the
lower layer to be carried by the tunnel

p., = A+ t~y,, (4.7)

where, provided that the upper layer consists of cohesive material,

and

(the cohesion is to be taken as that of the upper layer.)

The smaller of py1 and p, 2 shall be considered from here on.
5. Two tunnels close to one another can be regarded as a single structure of width b 3 ,
as far as vertical pressures are concerned, provided that

d < 1·3 (B 1 + B 2)
and

(Fig. 4/Sc). In other cases the loads shall be considered separately over each structure.
6. With p" determined, the active lateral pressures at a given depth y can be calculated
as foll ows:

Pxu = (py + yy) tan 2 ( 45° - ; ) - 2 cotan (45° - !). (4.8)

,s•
276 DESIGN OF TUNNEL SECTIONS

It is possible that the walls may move outward under the thrust from the arch. In this
case the maximum passive earth pressure that may be considered is

Px, = ~ [(g +yy) tan (45° + ;)+ 2 cotan (45° + :)] ,

2
(4.9)

where n is the factor of safety, which should not be less than n = 2 according to the Specifica-
tions.
B. If there is groundwater present, all the above must be modified accordingly.
7. Vertical pressures in pervious, homogeneous, loose granular materials (k ~ I0 - 7cm/s)
shall include the weight of the overburden, hydrostatic pressure and buoyancy (see Fig.
4/Sb for notations)
} - e-V
PY= I·4g. - -- - + h y, (4.10)
V

8. In soils stratified as illustrated in Fig. 4/Sd there are three py values to be calculated:

All loads - referring to a plane at a depth t 1 - are to be included in the product of t 1 y 1

including the weight of the water.

where g 1• denotes the pressure due to the upper layer, including buoyancy.
Denoting the smaller of Pn andPy 3 by p ' the critical load is the greater of p' and p, 1 • The
value of p ' shall be not less than 5 t/ m 2 (I ·025 lb/ft 2) and shall not exceed the geostatic pressure
without buoyancy.
If the arrangement of the layers is reversed, i.e. the impervious layer is above and the pervi-
ous layer is below,the pressures shall be computed as under 4,taking the buoyancy into consid-
eration as required.
9. The lateral pressure in pervious soils is to be calculated as the sum of the earth pressure
reduced by buoyancy, and the hydrostatic pressure, i.e.

g-hy ]
p; = [ I "4 V ' (] - e- ' ) + y (y - y ,) ). + (h + y)), , {4.1 1)

In the above expression the value of the hydrostatic pressure is generally predominant. For
this reason hydrostatic pressures shall be eliminated whenever possible.
Theoretically speaking, hydrostatic pressures need not be considered in solid, impervious
and non-porous materials provided the voids between the tunnel and the ground are filled
by grouting. ln impervious materials it is generally either the hydrostatic pressure or the total
earth pressure plus the weight of water that has to be considered. The reason for taking hydro-
static pressures into consideration lies in the possibility of water eventually reaching the struc-
ture through cracks or faults. However, because the changes of this happening are limited,
the safety factors may be reduced in such cases.
 DESIGN LOADS 277

In the case of temporary structures the water may be allowed to penetrate the structure as
this will lower the water table and eliminate hydrostatic pressure. Special holes shall be provid-
ed above such inlets, however, to allow inspection of the actual lowering of the water level.
C. The directives also cover the distribution of hydrostatic and earth pressures on double
linings. Here are the:, basic principles :
In the case of a single waterproof lining the latter shall resist hydrostatic as well as earth
pressures.
In the case of a combination of permeable exterior walls and interior waterproof linings,
such as brick of concrete wall outside, reinforced concrete lining and waterproofing inside,
the first will have to resist the earth pressure only and the latter shall be designed for the
hydrostatic pressure.
The above design loads, of course, apply not only to circular tunnel sections but also to
shields and any other underground structures used in the construction of subways.

4.15. SOVIET STANDARD SPECIFICATIONS FOR THE DESIGN

OF UNDERGROUND AND MOTORWAY TUNNELS

(Sztroitelnie Normi i Pravila, GI. 3.,

8. SNIP II-D - 3-62 and II-D-62, Moscow 1963)

The principal prescriptions as to loading assumptions of these extensive design

specifications are as follows:
1. Loadings acting upon tunnel structures must be assumed as a function of the
cover-depth, of geological, hydrogeological and seismic circumstances, of the
dimensions of the excavated cavity and of the construction method employed
and of the order of sequence of single items in the excavation-propping
process.
2. Dimensioning of tunnel structures must be effected on the basis of the most
unfavourable combination of all loads and effects acting simultaneously and
in cooperation either on the structure as a whole or on its separate parts
and either during construction or during operation.
For this, the following loads and effeets must be considered :

(a) Basic loadings composed of permanent loads and effects such as

Self-weight of structure
Vertical and horizontal rock (mountain) pressure
External hydrostatic water pressure
Pressures transmitted from buildings and structures on the external surface
lying within the area of subsidence
In addition non-permanent loads such as the weight of vehicles moving in
the tunnel or on the surface above it.
(b) Accessory loading-combinations to be superimposed on basic loadings from
the following non-permanent loads and effects:
 .. - -- - - - - - -

278 DESIGN OF TUNNEL SECTIONS

Temperature effects
Transient loads during construction Uacking pressure, weight of construc-
tion equipment, grouting pressure on the external face, compressed-air
pressure, local earth-pressure concentrations during the various stages of
excavation and filling operations).

(c) Special loading combinations to be eventually added to items (a) and (b),
e.g. seismic effects.
3. Tunnel portals and adits must be dimensioned for rock-pressure (vertical
a nd horizontal), for full dead-weight and for eventual seismic effects.

Permanent loads and effects

4. The magnitude of mountain (rock) pressure may be computed either on
the basis of the experimental values gained at similar constructions executed
under similar circumstances or in compliance with the present prescriptions.
(4.1.) The uniformly distributed vertical rock pressure prevailing in the most
common rock materials· acting upon separate tubular tunnels of 5·5-6·0 m
dia. and driven by the shield method, or by some similar method providing
a continuous and constant support, may be assumed either after the follow-
ing table or after point 5. (It may be noted that a tunnel may be regarded
as a separate unity when the distance to the adjacent tunnel exceeds:
(a) In limestone, marl and Upper-Carboniferous and Cambrian clays half
the diameter.
(b) In Jurassic and Tertiary clays and in sa11ds the whole diameter.)
(4.2.) Should this tabulated load exceed the geological pressure, then the effect
of arching must be considered (see point 5).
- - -- . - --- ------ --
Increased value
Basic value (to be used in Majoration
Kind of rock ma te rial aro und the excavation computatio n) factor

(t/ m ') I

Fissured limestone and marl with a cube- I

strength, (a, = 250- 400 kg/cm 2 ) 6 10 ] ·7

Strongly fissured limestone or clayey marl

(a, = 80- 250 kg/c m 2) 9 14 1·6
- -- !

C lays
I
Upper Carboniferous 13 20 I 1·5
Cambrian 16 24 1-5
Chalk 22 33 1·5
l ·5
Jurassic
Dense sand, with slight moisture content ..I
'
26
15
40
20
I
I ·3
 DESIGN LOADS 279

(4.3.) Should the covering thickness of the layer be less than the diameter of
the tunnel and if it is overlain by less resistant layers, the rock pressure
must be computed according to the following formula :
z (p, - Po)
p=p,- - ~ ·

where p 0 the rock pressure relative to the layer surrounding the tunnel
P, = the rock pressure relative to the weaker layer overlying the
surrounding layer
z = the distance between the tunnel crown and the bottom of
the weaker layer
D = the diameter of the tunnel.
(4.4.) For tunnels lying in clay and deeper than 45 m below the surface
the tabulated values must be majorated by a factor of K = H/45, with
H denoting the depth of cover.
(4.5.) If the tunnel is lying in a clay layer exposed to the seepage of water, the
tabulated values are to be increased by 30 %.
(4.6.) Lateral pressures may be derived from the vertical values when multiplied
by Rankine's ratio, i.e.
q = p tan2 (45° - ¢ /2) = p},0 •

5. For tunnels built under circumstances other than those under point 4.1.
the determination of rock pressures is to be effected as follows:
(5. I.) With a cover-depth sufficiently large for the development of arching, rock
pressure will be defined by the
area enclosed by the arch
started from the intersection
points of the rupture-planes
with the horizontal tangent
b
drawn to the crown of the
tunnel section (Fig. 4/9A). A
precondition for this is, how-
ever, that the distance be-
tween the vertex of the pres- m
sure arch from the bottom
of the weak layer or from the
ground surface must not be
less than the height of the
pressure arch itself. Should
this precondition not be ful-
8 = b t 2hun(45'-r/!2)
filled, the total value of geo-
static pressure is to be taken FIG. 4/9A. Assumed load on a circular cavity
into account. (after Soviet Regulations)
 280 DESIGN OF TUN NEL SECTIONS

(5.2.) The dimensions of the pressure arch may be gained frdm formulae:

B
B = b + 2m tan (45° - </>/2) and h= -
2/
where f = the resistance factor of PROTODYAKONOV.
(5.3.) According to the regulations the rock pressures acting upon the tunnd
lining both vertical and horizontal are to be assumed as uniformly distri-
buted in the case of arching also with the following values

p = hy (t/m2) and q = y(h + 0·5m) J. 0

,

where y = the dry density of the ground.

(5.4.) In loose and saturated grounds (quicksand and silt) the acting pressure is
to be assumed according to the laws of fluid-pressure.
(5.5.) Majoration factors for permanent loads and effects are to be c~nsidered
for all loading-combinations according to the following table:

Nature of loading Majoration factor

Vertical rock lwhen arching action is considered 1·5

pressure when total geostatic load is considered J·l and0·9
Horizontal active pressure 1·2 and0·8
Water pressure 1·1 and 0·9

NOTE: Majoration factors greater or Jess than unity must be applied always in a sense to inc rease safety.

Non-permanent loads and effects

Loads of vehicles must be assumed ~!ways according to the respective regula-

tions. Pressures exerted by hydraulic jacks must be multiplied by a majoration factor
of l · 3. The effect of temperature difference must be computed also in compliance
with the respective regulations, but with regard to the isotherms of the tunnel
site and dimensions and the extent of its exposure to the effect of atmospheric
and temperature conditions. Seismic effects must be taken into account only with
tunnels built in regions liable to earthquakes of 7, 8 and 9 Bells intensity.

Design stipulations

The dimensioning of masonry walls, portals and any kind of permanent tunnel
lining must cover the following three items:
I. Bearing resistance
2. Deformations and displacements
3. Check of fissuration-resistance.
 HORSESHOE-SHAPED TUNNELS 281

The check of deformations may be omitted when the rigidity and stability of
the structures are sufficiently proved either by previous tests or by practical expe-
rience.
The control of fissuration-resistance must be effected for normal operation con-
ditions and must not exceed for ordinary R.C. structures for maximum basic
load combinations the 0·2 mm and for additional accessory-loading combinations
the 0·3 mm value.
Statical calculation of tunnels must be effected after the methods of elasticity
with regard to the strength of materials, i.e. to those qualiti~ of the tunnel
structure proper as well as to the surrounding ground, and also to the specialities
of the construction methods employed.
In the case of lining segments built-in under the shieW. method, stresses may be
computed on the basis that plastic hinges are brought about possibly in the most
stressed cross-sections. Shield-driven tunnel linings must be dimensioned with
regard to the elastic support of the surrounding ground. This elastic support
should be assumed on those sections of the perimeter where deformation will
take place towards the ground.
Tunnels constructed in open cuts must be dimensioned according to the theory
of elastic subgrades with regard to the effect of lateral ground resistance.
Young's modulus (E); Poisson's ratio (µ) and the elastic subgrade coefficient (C)
values must be assumed on the basis of investigations of Engineering Geology,
of laboratory and in situ tests. and of observations and measurements taken on
similar structures.

4.2. DESIGN OF HORSESHOE-SHAPED TUNNELS

With the external loads determined as described in Chapter 3 and Section 4.1 ,
the actual design work can be started basically through reiterated investigations
(successive approximation) as is usual for statically indeterminate structures.
Empirical dimensions based on existing tunnels are assumed first and then checked
either by the earlier graphic or by one of the more recent numerical methods.
The procedure can be greatly expedited by using various design tables developed
for arches and circular sections. Admittedly, these tables do not consider the
exact distribution of loads and use either point loads or knife-:!dge loads or distri-
bute the loads in an easily manipulatable manner. The errors resulting from the
simplified distributions, however, are much smaller than those involved in the ·
computation of rock pressure.

4.21. DESIGN BY MEMBERS

Consistent with the classic step-by-step tunnel construction methods (Sections

62.2 and 62.3) the earlier design methods considered the top, bottom and wall
sections as separate members subject to inter-reactions ensuring their rigid con-
282 DESIGN OF TUNNEL SECTIONS

nection. It was not correct, of course, to ignore the deformations and displace-
ments of the above members because the loads from the supported units will
cause certain displacements which, in turn, will again affect the reactions them-
selves.

42.11. Graphic Investigation (Kommerell)

With a set of empirical dimensions established and the rock pressure computed
either by theory or by specifications the loads and their locations can be deter-
mined as follows (Fig. 4/9B):

\ v- ¥ l;, d
\ Hz X,+X2 l
\ 2 V
V
II
H V

FIG. 4/9B. Loads acting on

6 elementary strips

Sliding surfaces are started from the lower corners of the walls at an angle of
45° + ip/2 extending to the crown and then continuing in a parabola or ellipse
of height h thus defining the " burden zone", i.e. the zone to be supported. A dia-
gram can then be constructed for the lateral pressures in the usual manner and the
arch as well as the wall are separated to form individual members. The loads for
each element of the arch can be obtained by projecting the limits to the diagrams
horizontally and vertically. The vertical load is V = Yi; Y2 thy and the hori-

2
zontal load is H = Xi+ x I, . In addition to V and H each element also has to
2
carry its own weight G; the resultant of the three forces can be constructed as
illustrated in Fig. 4/9b.
 HORSESHOE-SHAPED TUNNELS 283

The above loads can also be determined in a simpler manner, particularly when
used in analytical calculations (cf. Fig. 4/ 6).
Once again the burden zone is constructed first and then the pressure diagrams
are evened out between the vertical lines representing the tunnel limits. Loads
should be computed separately for the centre portion and for the two end sections.
The vertical pressure diagram will thus consist of p 1 ordinates in tlie centre and
the smaller p 2 ordinates at the ends.
Lateral pressures are to be calculated from p 2 . There is no significant difference
between the two methods; the uncertainties in establishing the depth of the burden
zone are considerably greater.
In fact either method may b~ adopted.
ln rock where there is no lateral pressure, the sliding surfaces at 45° + ¢ /2 are
omitted and the burden zone is as wide as the tunnel itself (Fig. 4/6a). The vertical
loads have to be balanced by bearing pressures under the walls or under the entire
inverted arch as well. The horizontal loads on the opposite walls are in equilibrium
in themselves.
42.111. Absence of lateral pressure. We shall now consider the first case
where there is no lateral pressure and therefore no need for an inverted arch.
The problem is one of constructing a thrust polygon under the external loads
plus the self-weight through the crown and the side walls in such a manner
that the stresses due to eccentric loads do not exceed the allowable values in
any cross-section.
There are four conditions to be satisfied:
1. The thrust polygon shall be contained within the inner third of any cross-
section and there shall be no tensile stresses. Both conventional (brick, stone)
and new (concrete blocks) building materials can withstand considerable compres-
sion and, in fact, a certain amount of tension. However, any such tensile strength
cannot be relied upon because of the very nature of the interlying mortar.
2. The maximum compression stress in any cross-section shall not exceed the
allowable value. While the allowable stress depends on the particular building
material (stone or conci:ete) it is not equal to the strength of the latter. The overall
strength of the wall is once again governed by that of the mortar.
An allowable compression stress of a = 20 kg/cm2 (285 psi) was widely accepted
in the design of earlier tunn<.:ls, but would be considered unreasonably conservative
today. Stresses as high as 100 kg/cm 2 (1420 psi) and 134 kg/cm 2 (1900 psi) have
been measured in the crown and walls of the Gotthard tunnel, respectively,
without any visible sign of damage to the brick material laid in cement mortar.
Today, it is customary to allow 40 kg/cm2 (570 psi) compression in walls of medium-
quality and 60 kg/cm2 (850 psi) where better than average quality can be expected.
2
(The allowable compression in the Lisbon Subway [brick walls ] was 70 kg/cm
[1000 psi].)
3. The third condition calls for fairly uniform stresses in the various sections.
This is, of course, generally an ambition of the engineer, serving safety and econo-
mics at the same time.
284 DESIGN OF TUNNEL SECTIONS

4. As a fourth condition the thrust line must cross the neutral axis at least
twice for stability considerations (Fig. 4/ 10 and below). 2
Allowable soil bearing pressures may generally be increased by 4 kg/cm (60 psi)
if an inverted arch is adapted in the design.
In the first case we have assumed that an inverted arch was not required because
the tunnel was subject to vertical loads only.

r.Fa-0
F • [ f1i·. ,/Js1
ff• lf e,

FIG. 4/ 10. Conditions

a governing the
a construction
of a thrust line

In his method, KOMMERELL starts out by dividing the crown into elements and
compiling the loads into a vector diagram (Fig. 4/11). The pole is established
arbitrarily along the horizontal line preceding the first force (O'). The resultant
of the external loaqs can be determined independently of the arch. The thrust
line can now be constructed so that it passes within the inner third points and
crosses the neutral axis at least twice. The area between the thrust line and the
neutral axis represents the moment diagram (N x e) (Fig. 4/10). For best results
the algebraic sum of the areas of the moment diagram should be zero, as should
be the moment of those areas about the springing to eliminate displacement at
the latter. This can only be achieved with moment areas of alternate signs, a con-
dition already mentioned.
The thrust line can be constructed by trial and error. For a symmetrical section
subject to symmetrical loads the thrust line is horizontal at the crown. The first
line of the polygon can thus be drawn tentatively through the outer third point
at the crown and then extended to intersect the resultant. The closing line is drawn
through this intersection and through an arbitrary point at the springing. To make
full use of the designed sections the arbitrary point has to be located again within
the inner third but on the opposite side of the neutral axis. A line drawn in the·
vector diagram from the end of the last force and parallel to the closing line wiil
 HORSESHOE-SHAPED T UNNELS 285

115t 6-8t 10-2 t ,2.gt f4-t 1

i t t i i
P. P, P, P,

6JO (t 6012 \ . 4.,•-•t,lm'

I 6, I f-2
1·2
14·!
IN 14-1
!5·J
v,.
,.
2·0 + 20-; <'l , ...
2
fJJ{} ft-
o.b • N 60·12 '-20-1 tm'
i J I.Z 10·2 11-4 \' 20 J 1/'

4 f.2 6-8 8·0

X •
f4f.!-2C.t
f
J
2 360.M'
· (/·Mm
5 2•7 11·5 14-2
7~
6p - iJ.fJ -28·5 tjm
i[V·~ 2
6 28
7 2-8
8 3>1

!J_JIJD1~ _,., w
F ro. 4/11. Graphic investigation by KoMMEREL L

establish the proper pole (0) as the point of intersection with the first horizontal
line. With the thrust thus determined both at the crown (H) and at the springing
(R) the thrust line can be completed. Now is the time to check whether the assumed
polygon lines have been positioned correctly at the crown and at the springing,
whether the thrust line crosses the neutral axis in at least two locations, whether
it passes within the inner third of all the cross-sections, whether the stresses are
reaso nably uniform, etc.
286 DESIGN OF TUNNEL SECTIONS

Discrepancies, if any, can be corrected by shifting the thrust line at the crown
and at the springing, or - if this is impossible - by changing the shape and/or the
thickness of the arch.
The walls will be investigated next. The reactions from the arch will force
the walls against the natural ground resulting in passive earth pressures behind
them.
First the amount of passive earth pressure as well as its distribution and location
should be computed. The horizontal component can be readily obtained: because
of the equilibrium of all horizontal loads the horizontal component of the passive
earth pressure (Eph) must be equal to the thrust at the crown (H). As the wall will
tilt under the abutment pressure the shape of the passive earth pressure diagram
will be either a triangle or a trapezoid. To determine the line of the resultant one
has to consider all the forces that are acting on the wall, i.e. the thrust from t he
arch (abutment pressure), passive earth pressure, self-weight and the bearing
pressures and friction under the wall. In graphical investigations the .latter car1 be
ignored, even though the base will obviously have to move horiior.ta}ly t(J be
consistent with the partial distribution that has been assumed for the pa ~,ive
earth pressure, and even though a friction force S2 will necessarily have to dew lo p.
According to KOMMERELL, the passive earth pressure will not develop to the f}1ll
depth of the wall but only in the undisturbed area near the springing, as wo uld
follow from the method of construction. The balance of abutment pressuw ,
earth pressure, self-weight and bearing pressures is reflected in the closed polygo1\
of forces. The location of the thrust line and of the self-weight has been determim·d
earlier. When using elementary strips the resultant of the thrust line, self-weight
and passive earth pressure shall be carefully controlled so that it stays within
the inner third at the base as well as half way up the wall where the wall section
is usually the thinnest (Fig. 4/ 11). An attempt should be made to keep the resultant
near the centre of the base to preserve uniform bearing pressures. It is advisable
to assume an arbitrary location for the resultant of the bearing pressures and proj~ct
it vertically upwards to intersect the thrust from the arch. Because of equilibrium
the resultant of the passive earth pressures has to pass through this point as well.
The passive pressures are generally inclined at ¢ = 20°-30° to the horizontal
(reference works recommend the use of tan ¢ = 0·3-0·5). An angle of¢ can
be assumed consistent with the type of ground material which will also determine
the direction of EP at the same time. The diagram for the distribution may be
taken as a triangle. The dimensions of the diagram (dist:ince x) and the exact
location of EP can be calculated by equating the tilting of the rigid wall to that
of the base, as follows:
The bearing pressures are calculated as if their resultant was in the assumed
location. Because of the trapezoidal pressure diagram the base will undergo ·
a rotation, the angle of which will be in proportion to the ratio of the difference
of ak and a~ over the width of the base. Assuming definite proportions between
the passive earth pressure and the horizontal displacement of tbe wall, a similar
ratio can be developed regarding the passive earth pressure diagram. The angle
 HORSESHOE -SHAPED TUNNELS 287

of rotation should be the same for the base and for the wall because of the rigid
connectio n. Thus
(Jk - (Jb (JP u/J · 3x
--- = but - 2- =Eph
b 3x

2Ep1r
and hence (JP= 3x

2Eph (Jk - (Jb I 2Eph b

su bstituting = -- b- and X= - - ~
9x2 3 <Yk - (Jo '

2Eµh uk - ub 1 J?Ephb (4.12)

or in centimetr es and
900x2 = - b- - X = JO (Jk - <Jb

(In the above formulae all quantitie s are in centimetres~ division by 100 refers
to a strip of wall l m wide.)
Figure 4/ 11 shows that the passive earth pressure diagram can now be complete d
and with the location of EP known, the resultant of the bearing pressures and of
the horizonta l loads can be located. Finally the polygon of forces can be complete d
and the funicular polygon can be construct ed.
In the next step the sections should be checked for stresses.
Tunnel sections construct ed of brick, concrete or stone blocks are very often
designed by the above graphic procedur e. It must be borne in mind, however,
that EP and x have been determin ed arbitraril y, because 3x < h implies that
point A is the centre of rotation and point O has to move inward (Fig. 4/ 12).
This movement, on the other hand, will certainly develop a friction force S 2
which should then be taken into account in the balance of the horizonta l forces.
In practical cases, however, there is sufficient friction to keep point O in a fixed
I
position (RozANOv) and thus 3x = h, but Si=--;,- u"1, - H which should be less
than the available maximum of W tan cp. u, can be compute d directly from moment
considera tions with refetence to point O,

(J h2
_P_ = Rr + Wt,..
3

3 (Rr + Wr~) ( 4.13 )

and hence (J p = h2

Actually the value of St could now be determin ed. There is this inconsistency,
however, that the base would have to move over to mobilize friction, in which
 288 DESIGN OF T UNNE L SECTIONS

1541/m 91/m !Mf/m 11-21/m IHl/111

:si ~
~
~ ~ 0 10 20 JOt/m
~~ ~ ~
~ ~

I ft IH trJ/1
..,, 2 f2 m IN ~n
..,~ J 1-2 /# 14-8 ,., 4f /J-6 (){)6(l fll9
4 u 9/J 02 b-b 56 06 ()(J67 156
5 2-7 /5,f I.JI c-c 76 0c ()(J67 211
6 N N d-d 'I() H5 IMtJ # 5
~ 7 N f.8 e-e ~ 1-JJ /J0/6 ~

4 "4 J,I
IV • g().fft

F1G. 4/12. G raphic investigati on considering the horizontal displacement of the base

case the passive earth resistance wo uld not develop behind the full height of the
wall (h) but to a height of 3x only. For a base movemen t of L1 we have

L1 = !!_I!.. . h - 3x
C 3x

where C = the coefficient of subgrade reaction.

Hence x may be expressed as x = - -<Jph- - . The
movemen t L1 I. S 1·1m1te
. d to
3(up + CLI)
1-2 cm (3/8"-3/4") (this being the upper limit required to develop the maximum
friction of tan </J, the value of which has to be determin ed by a shear test) and so
a passive earth pressure distributi on over less than the full depth h is justified
only for large values of C. The construct ion in Fig. 4/ 12 illustrates a case in which
both Sz and EP have been considered. Assuming L1 = 0·3 mm (1 /8") and C = 20
 HORSESHOE-SHAPED TUNNELS 289

k g/cm2 (284 psi), fo r the given h = 4·97 m (16' 3 5/8") the results would be u =
= 23·4 t/m2 (4·79 lb/ft2) and x = 1·32 m (4'4" ). •
42.112. Lateral pressures present. In the second case the tunnel is subject to
lateral pressures, and an invei;ted arch is required. To investigate this problem
the simplest method is, once again, a graphic one.
Tunnels passing through loose and soft materials are subject to vertical as well
as to lateral loads. Because the computation of forces and their locations have
been discussed earlier, we proceed directly to the investigation of the wall.
Fig. 4/ 13 shows a tunnel section with empirically established dimensions.
The investigation should be started with the arch at the top. The resultants
of the forces on the segments are no longer vertical. The resultant of the forces
on the arch is determined fi~st (see the polygon above the arch line). The thrust
line is constructed next; in Fig. 4/13 it passes through the outer third point of
the crown and through the inner third point at the springing. A horizontal line
through the upper third point of the crown is extended to intersect the resultant

tiMl/m §t/m fJ-51/m t7-tt/m

· JW f/m
::'i.
>---.,~ -----,.lr------1----- # i
i :;~ Scale Ill'fbrces

it
:L/J fQ f!J l/m

Weights
.;

I
fl l/m
~ 6, f.21/m ~etio, N (tjm) b(,,,) er"J 4/i,,.,?
:.:., 6j f.21/m a-a 51-50 /}ljO (}(/fJ 1115-5
:§ ~ f.21/m /J-b 51-50 /}6(/ /1(145 f4!M
'<i
~ 2-741/m c-c 72-40 060 /I05J f//4·4
~,
"> Ii;- 2-611/m d-d 84·5/I /J-95 /J-105 f# /1
"'., li1 2-Mt/m ,_,, 8/J80 f.20 (}l)QJ 08-J/J
~ 6i 3-431/m f-f 25·5/J /J,40 /J-(JIJ 55-5

~ {}6 t/m g-g 25-/JO /},//} IJ-/JfJ 74·7

1 'l';
f;i e. (}litjm
~
~ G, (}lit/m
l 5w /J5 l/m

F IG. 4/ 13. Graphic investigation considering active lateral pressures

19 Szechy: The. Art of T unnelling

~ - - - - - - - - -

290 DESIGN OF TUNNEL SECTIONS

(the line is horizontal under symmetrical loads only) and this point is connected
with the inner third point at the springing. A line parallel to this and starting from
the end of the resultant in the force polygon will· yield the values for the thrust
at the crown (H) and at the springing (V). With the pole thus determined the
thrust line can be completed for the entire arch, checked' for the conditions dis-
cussed earlier, and adjusted if necessary. This procedure should be repeated until
all the conditions have been satisfied .
If the arch is found to be satisfactory, the thrust line is extended through the
walls down to the base. The three forces to be balanced are the reaction under
the base (W), the reaction under the inverted arch (T) and the thrust in the bottom
arch (H The bearing pressures under the wall base and under the inverted arch
0
) .

are assumed to be distributed uniformly. To obtain the location and the direction
of the thrust, the last line of the thrust polygon is extended to intersect the vertical
resultant of the bearing pressures under the wall base which is assumed to act
near the centre. This point (a) connected with the intersection of th~ resultant
of the bearing pressures under the inverted arch and of the horizontal thrust
passing within the inner third of the lowest section (b) defines the direction of the
thrust between the wall base and the inverted arch; the force itself can be scaled
on the parallel line in the force diagram (V'). This line also establishes the vertical
reactions; the lower being the reaction under the wall (W) and the upper being
that under the inverted arch (T). The thrust line can then be refined considering
the distribution of the bearing pressures under the inverted arch as well.
Because of the favourable effect of lateral pressures on the thrust line it should
be carefully considered whether in an actual case an active lateral pressu1e will
materialize and its value should be kept to a minimum in doubtful cases. (In other
words the angle of repose shall be taken at its maximum and the cohesion shall
also be included in the calculations).
The passive earth pressure may be considered only on the following conditions:
Owing to the trapezoidal distribution of bearing pressures under the wall the
latter will tilt and thus mobilize the passive earth pressure with an assumed trian-
gular distribution extending over its entire height. Corner O can be assumed as
fixed. If Ea is insufficient to balance Hr and Ha, the passive resistance can be cal-
culated as

i.e. as required for the equilibrium. This, on the other hand, must be accompanied
by a displacement y at the springing. The arch at the top will have to be checked
for this displacement y = ~ (C being the coefficient of subgrade reaction), the
effect of which shall be treated as discussed in Section 42.122. The method of
keeping the thrust line within the inner third using the passive earth pressure can
be seen readily in Fig. 4/13.
 HORSESHOE-SHAPED TUNNELS 291

There is always an interaction between the arch at the top and the wall.supports
a nd this should be taken into consideration even when treating the arch and the
walls as separate members.

42.1 2. Analytical Design by Members

An investigation by calculation rather than graphical construction is not only

more accurate but it also enables the designer to consider other effects within
certain limits. Such a method will be illustrated here for horseshoe sections that
are closed all around, i.e. those that are subject to active lateral pressures and have
an inverted arch.
42.121. Analysis for horseshoe· sections. In this an;tlxs.is both arches (top and
bottom) are considered fixed at both ends; the walls are regarded as completely
rigid and are assumed to be supported elastically by the ·ground in a horizontal
and vertical direction. The external loads consist of the self-weight, vertical and
horizontal earth pressure and the reactions such as the bearing pressures under
the base, the passive earth pressure behind the walls and horizontal and vertical
forces of friction.
Both methods described below are based on the equilibrium of forces interacting
between members designed as separate units.
In the simpler method all end supports are assumed either to stay in place or
not to affect the reactions by any displacements. This assumption is rather arbitrary
and can be accepted only if the conditions warrant it (e.g. heavy sections sur-
rounded by rock). If the relative displacements at the springings cannot be neglected,
then the effects of the support movements have to be included in the arch design.
This method, of course, is much simpler than to design the entire section as
a ~ontinuum and will be sufficiently accurate in ground that is capable of develop-
ing considerable support.
A member-by-member design is justified also beca:use all three members are
constructed separately and are divided by construction joints anyway.
This method should be used only in the design of structures with rigid and
heavy walls.
Details of the design of each men:iber are given below:
l. The arch at the top is assumed to have fixed ends and the combined l9ad of
its self-weight plus earth pressure is assumed to be diEtributed in a uniform manner

Q+ Q0 + LIG
p=
!

where Q = the self-weight of the arch

Q0 = the uniform rock-load
LlG= the variable rock-load (Fig. 4/1 4).

19•
 292 DESIGN OF TUNNEL SECTIONS

FIG. 4/ 14. Loads on separated

members of a
tunnel section

The arch represents a statically three times indeterminate structure, the r~dun-
dants of which can be determined as known from the theory of structures.4 · 2
By replacing the rigid support at one of the fixed ends by a moment (X ), a hori-
1
zontal force (X2 ) and a vertical force (X3) the structure is transformed into a can-
tilever i.e. into a statically determinate structure (Fig. 4/ 15). The three redundants

Fio.4/ 15. Statically determinate

structure equivalent
to the hingeless arch

have to be determined in such a manner that the resulting relative displacements

and rotation shall be zero at the released end. Here are the three equations of
redundanc y:
auX1 + a 12X2 + a 13 X 3 + a 10 = 0,
a 21X 1+ a 22X2 + a 23X3 + a 20 = 0, (4.14)
a 31X1 + a 32 X 2 + a 33 X 3 + a 30 = 0,

4.2 KoRANYr, I.: Tartok statikaja (Theory of structures).

Tank6nyvkiad6, Budapest 1957
II 205
GARTNER: Statically Indeterminate Structures. Concrete Publication s Ltd.,
London 1958
BEYER, K.: Die Statik im Eisenbetonbau I. Springer, Berlin 1930
 HORSESHOE-SHAPED TUNNELS 293

where an, a 1 2 and a 13 represent the rotational, horizontal and vertical displace-
ments, respectively, at the released end, due to a unit moment. Similarly a ii,
a 22 and a 23 denote displacements due to the unit horizontal load and a 31,
a 32 and a 33 denote the respective displacements due to the unit vertical load.
Finally a 10, and a 20 a 30 denote the rotational, horizontal and vertical displace-
ments, respectively, at the released end due to the external loads. The three
equations are, in fact, a mathematical expression of the fact that the released end
will not move or rotate in any way under the combined action of the redundant
forces and the external loads. In other words, the redundants X1, X2 and X3
will cause the cantilever to act exactly as the indeterminate structure.
Transferring the redundants Xi, X 2 and X 3 to the elastic centre of the arch
(point o) will make the calculations considerably simpler as each redundant will
cause only one movement and that in its own direction, i.e. the moment will
produce rotation only, the horizontal redundant force will produce a horizontal
displacement and, finally, the vertical redundant force will produce a vertical
displacement only.
(For arches with a constant moment of inertia the elastic centre is identical
,,ith the centre of gravity of the neutral axis; for arches with a variable moment
of inertia the elastic centre becomes the centre of gravity of the ~s quantities).

Or, if t; = l; , the elastic centre can be determined as the centre of gravity of the
lo
Lis; quantities, where 10 denotes an arbitrarily chosen (constant) moment of
'C;
inertia.

Yo= (4.15)

With the redundants acting in the elastic centre the equations of redundancy
will reduce to the simple form of

X1a11 + ao1 = 0,
X2a22 + ao2 = 0, (4.16
X3a33 + ao3 = 0.
Here we have assumed that the elastic centre is rigidly connected to the released
end and each redundant, acting at the elastic centre, will cause it to move in its
own direction only (Fig. 4/ 15). This thesis can be proved mathematically.
While the factors a0 k denote the relative displacements of the elastic centre
under the external loads, i.e. the rotation (a0 J the horizontal displacement (a02)
 294 DESIGN OF T UNN EL :.r.....'TIONS

and the vertical deflection (003) , the other factors with the double
suffix: 0 11 , 0 22
and 0 33 denote the corresp onding movements due to X = 1
1 tm, x~ = l ton
and X 3 = l ton, respectively.
Factors oOk and okk can be calculated from the equatio ns of virtual
worlr as
follows:

(4.17)

(4.18)
where M0 and N0 = the momen t and the axial compo nent of the externa l load ,
respectively
mk and nk = the moments and axial loads due to the unit loads X1c = l ,
all of which are considered as acting on the equival ent
determ inate structu re.
The second member in each formul a considers the effect of the axial
loads and
can usually be neglected as insignificant compa red with the first membe
r.
The moment diagrams M 0 , mi, m 2 and m a re to be established first
3 and the o
factors then calculated as their produc t.
Fig. 4/16 illustrates the momen t diagram s in question. The arch is
to be divided
into n segments a nd the momen ts calcula ted at the dividing points.
From the momen t diagram s the o factors can be calculated as

n 1 2
1 n Lls
a1 1 = L1 --
EJ;
Lls-
'
= - -L-
EJo 1 T; ' ,
YT I Lfa;

rt::;r
II n 2

o,, - : EJ; ~ ~ y: , As; ( / '

2
1
(4.19)

0 33 = L ---
EJ;
- Lls;= -EJ L - T ; --- x;).
1 2 0 1

For unifor mly distribu ted vertical loads (p)

l
l
p(l
= - L - -- -X;)2 P ~
n
0 01 As; = - - - L -Lis; II
(I - x,)-,
1 2EJ;
2£10 1 t;
p(l - X;)2 Y; (4.20)
n p n As,
0 02 = - L --- -- As = - -;.,- - L -
?

2EJ, 1 (l-x;)- Y;,

i ,_£Jo t t;

where 0 03 = zero if the arch and the loads are symmetrical.

 HORSESHOE-SHAPED TUNNE LS 295

Moment di,1gr,1m due to uniforml!J distributed loads

-p(!-xJ2
Mo - 2

Moment di,1gr,1m due to' x 1 • 7 tm (m1)

Moment diagram due to x2 = 1 t ( m1 )

Lls· ~
0

F 1G. 4/ 16. Moment

diagrams
of the arch
released at
one end

From the above X1 and X 2 can be calculated. If the structure and the loads
are symmetrical about the vertical centre line, X3 will be zero and

(4.21 )
 ---- --
~~·

296 DESIGN OF T UNNEL SECTIONS

The moment and the axial load at any point of the statically indeterm inate
structure is calculate d from

M = M0 + X1m 1 + X2m 2 = M0 + X1 •1 + X2Y;, (4.22)

N = N0 + X 1n 1 + X 2n 2 = N 0 + X 1·0 + X 2 cos</>,

where M0 and N0 = the moment and the axial load on the cantileve r, r~spec-
tively.

For a circular arch with a constant

moment of inertia the bending mo-
ments and the axial loads can be ex-
pressed in the form of closed :ormu-
lae. A derivation for uniformly distri-
buted loads is given below. In this
X particular case where the structure
and the loads are symmetrical about
the vertical centre line the structure is
split through the centre (rig. ' 4/17)
and thus becomes a pair of identical
and statically determina te car•iJevers .
This will greatly simplify th• alcula-
tions.
In computing the relative displace-
ments the only effect considered will
FJG. 4/ 17. Statically determina te structure for a be that due to the bending moments;
circular arch both axial load and shear will be
neglected.
The location of the elastic centre is

••
2 f ds.r (I - cos </>) f••r 2 (I - cos </>) dtf>
0 0
y~ =
••J ds = ••Jrdq,
2
0 0
(4.23)
r2 ( tf>o - sin </>o) r(t/>0 - sin </>0)
Yo=
rtf>o <f>o
In the next step the moments on the two cantilevers are computed, due to the external
loads and then due to X 1 = 1 tm (ftk) and X 2 = 1 ton (kip). According to Fig. 4/ 18 the
moment
due to the external loads is

px2 pr 2 sin' 4>

Mo= - - 2- = - - -2- -
The moment ciue to X 1 = 1 tm (ftk) is constant throughou t
m1 = 1 tm (ftk)
and the moment due to X 2 = 1ton (kip) is
m: = -y0 + r (I - cos</>).
 HORSESHOE-SHAPED TUNNELS 297

With the moments determined the reiative displacements of the elastic centre can be calculat-
ed. Since the moment of inertia is constant, all displacements may be multiplied (magnified)
by El.

J
a; 1 = L mi ds = 2 r dq, = 2r <Po . J••
0
(4.24}

~ ~

a; 2 = L f mids = 2 J[-Yo+ r (1 - cos 4>)] 2 r dq, = 2r J[Yl - 2ry0 (1 - cos 4>) +

0 , Q

+ r 2 (1 - cos 4>) 2 ] dcp = 2r [Yl 4>o - 2ry0 4>o + 2ry0 sin <Po+ r' <Po - 2r2 sin 4>o +

+r 2 (: sin 24>0 + q,0 /2)] = 2r 3 [ (4>o - ;~n4>o)2 - 2 (4>0 - sin 4> 0) +

( q,0 - sin 4>0) sin 4> 0 + ,1, . ,1, + 1 . 2 ,1, + <Po ] =

+2 4>o
'l'O - 2 sm 'l'O -
4
Sill 'I' -
2
2
2sin 4>0 1 . ) (4.25}
= r 3 ( 4>0 - 4>o + Sill 24>0 ,
2
.,,,
a, "f
01 =1..., M 0 m1 ds = -2. r 2
pr Sin
2
2
q> rd4> = -pr3 (_24>o - 41 sm24>o
. ) (4.26)
9

, = LJ
"1oz Mom2ds = -2 .r-
h
2
pr sin q>_
2
2
[-y0 + r(l - cos q,)rdcJ,] = -pr 3 J
·
~

((r -
o n
0
-yo)sin 2 q,-rsin'q,coscf>]d4>=-pr3 [(r-y0) (-: sin24>o+ ~ )-

3
-
sin3 4> = - pr 4 [ -
0
r ---
3
sin -
4>0
4>0 ( - -sm
4
1 . ,1,
2 .,, 0 + -4>o)
2.
- -sin-3 4>o]
- =
2
1 sin 24> 1 sin 4>o )
= -pr4 smcf,0
• (
- ----+----
4 4> 2 3
0
(4.27)
0

(4.28}

.1 sin 24> 1 sin2 4>

- - - -0 + - - - -0
Doz
X2 = - -, =
a22
+ pr sm. 4>0 - -4-2-cf>o 2
--------
sin 4>o 1 . 2
3 (4.29)
- - - - + - sm24>0
cf>o 2

The moment and axial load for any section of the statically indeterminate arch is

M = M 0 + m1 X1 + m2 X 2 = M 0 + l ·X1 + X 2 y, }
(4.30}
N =N0 + n 1 X1 ~ n 2 X 2 = N0 +o+X2 cosq,. ;
_ ,. "'Q, ... i,. .
298 DESIGN OF TUNNEL SECTIONS

For a semicircular arch with ef, = 90° the moments are given in Fig. 4/ 18
X1 = +0·25p r 2 , }
(4.31)
X2 = +0·560 pr .

f§

..(

p
I j j I jj jj j j j I II j Ij j I j j I I j I j I j j j j j I j ti
~i
~-

0. 1/

I
FIG. 4/18. (a) to (c) Moment diagrams on the statically determinate equivalent structure;
(d) Moment diagram on the circular hingeless arch
 HORSESHOE-SHAPED TUNNELS 299

The ma,d mum moment at the springing is

Mmax= - pr• sin' 4'o +0·25pr-+

" ( - y +r-rcosq, ) 0·5,pr
6 = 0 ·107pr-
6 " . (4.32)
0 0
2

These formulae have been developed into convenient tables for the design of
the Budapest Subway. The factors in the tables can be applied directly in the cal-
culation of moments and axial loads at the critical cross-sections under uniformly
distributed vertical loads

M = rx.Mp/2,
H = ct.Hp!,
(4.33)
I
V=p - , e.
2

and under uniformly distri-

buted horizontal loads lf
r1-?;rc /.;n-
,. i

M -- {3M e/2' } (4.34)

FrG. 4/ 19. Notations for the circular arch
H={)Hel.

The notations are illustrated in Figs 4/ 19 and 4/20. Numerical values for ct.M and
r:1.H are listed in Table 4/1 for cross-sections at the springing (1), quarter point (2)
and crown (3).
With the moments and axial loads determined, the stresses due to the eccentric
load can be calculated from
N M
a1. 2 =A± K '

FIG. 4/20. Tabulated data

of the circular arch
 w
0
Coef/icients for Circular Arch Design TABLE 4/ 1 0

Moments at
----- Thrust at springing
--- - -
Ri se to X pl' Central
R adius (r)
span ratio angl e
ff/ Springing (l ) Qua rter I Crown (3)
X pl
I
I point (2)
I "'· ""'
-
"M
I {JM "M
I "M
I {JM "H I {JH

0·10 +0·00120 - 0·00120 - 0·00042 +0·00040 1 -0·00040 1·2750 0·0750 1·3011 22° 36'
O· l i +0·00167 - 0·00167 - 0·00047 +0·00058 - 0·00058 1·0390 0·0574 l · 1013 27° 00'
0·14 +0·00224 - 0·00224 - 0·00066 + 0·00080 - 0·00080 0·9045 0·0816 0·96 15 31° 20' 0
0·16 + 0·00292 - 0·00292 - 0·00087 +0·00110 ' - 0·001!0 0·7950 0·0938 0·8610 35° 30' ~
ci
0· 18 +0·00366 - 0·00366 - 0·00111 +0·00140 - 0·00140 0·7075 0·1031 0·7844 39° 36' z
0·20 + 0·00452 -0·00452 - 0·00138 +0·00 176 I - 0-00116 0·6335 0·1085 0·7256 43° 36' .,,0
0·22 +0·00550 -0·00550 - 0·00168 +0·00214 - 0·00214 0·5800 0·.1218 0·6781 47° 30' -,
0·24 +0·00653 - 0·00653 - 0,00200 +0·00256 - 0·00256 0·5 355 0·.134 1 0·6410 51° 16' C
z
0·26 + 0·00765 - 0·00765 - 0·00239 + 0·00300 - 0·00300 0·4995 0· 1487 0·6109 54° 56' z
0·28 +0·00885 -0·00885 - 0·00280 +0·00346 - 0·00346 0·4685 0·1621 0·5864 58° 30' "'r-'
0· 30 + 0·01010 - 0·01010 -0·00324 + 0·00400 - 0·00400 0·441 5 0 -1749 0· 5666 61 ° 56' "'
"'(")-,
0· 32 +0·01 145 -0·0 1145 - 0·00372 + 0·00458 - 0·00458 0·413 5 0·1829 0·5507 65 ° 14'
0·34 +0·01285 - 0·01285 - 0·00424 +0·00520 - 0·00520 0·3920 0·1944 0·5376 68 ° 26' 0
I ti
0·36 + 0·01434 - 0·0 1434 - 0·00479 +0·00585 - 0·00585 0·3725 0·2053 0·5274 71° 27'
0·38 + 0·01589 - 0·01589 - 0·00540 +0·00648 - 0·00648 0·3540 0·2151 0·5 190 74° 28 '
0·40 + 0·01748 - 0·01748 - 0·00603 +0·00721 - 0·00721 0·3385 0·2260 0·5125 77° 19'
0·42 +0·0 1920 - 0·01920 - -0·00672 +0·00797 - 0·00797 0·3247 0·2371 0·5076 80 04'
0·44 + 0·02105 - 0·02105 - 0·00744 +0·0088 1 - 0·00881 0·3120 0·2479 0· 5041 82 42'
0·46 +0·02296 -0·02296 - 0·0082 1 +0·0097 1 -0·00971 0·3000 0·2583 0·5017 85° 14'
0·48 +0·02478 - 0·02478 - 0·00902 +0·01064 - 0·01064 0·2895 0·2690 0·5004 87° 40'
0·50 + 0·02668 - 0·02668 - 0·00987 +0·0116 1 - 0·0 1161 0·280 1 0·2801 0·5000 90° 00'

fJM =- aM fJM =- aM '

/h, = l
2
cot <f, - a 11 I
I
I

I
 - - - - -- -- - ----

HORSESHOE-SHAPED TUNNELS 301

and the cross-sections can be designed accordingly. Should the initially assumed
cross-sections prove to be inadequate then the calculations have to be reptated
because of the effect of the change in J(K) and A on the deflections and on the
reactions.
2. A similar procedure has to be followed in
the design of the inverted arch. The external
loads have to be estimated with due considera- 4 u....u..u......i....u.u..L.U..1.........u.u..u.u.........~
tion to construction procedures. In most cases
the arch at the top and the walls are constructed
first and so their self-weight is transferred to
the ground at the wall base.
In the design of the inverted arch the bearing
pressures have to be calculated from the self-
weight and from the earth pressure, whether to-
tal or partial. These pressures have then to be
reduced by the friction behind the walls (Fig.
4/21)

The vertical pressure on the inverted arch is

,_ (
Pv- g +
AG+ 2P + 2Qa - 2S
L
lL '
,
LK
m

where rx denotes that fraction of the pressure that

will develop only after the inverted arch has been Fro. 4/21. Loads
on the inverted arch
completed (40-50% in plastic soils).
Moments and axial loads can be calculated
similarly to those of the arch at the top.
3. Design of the walls. Once again the external loads have to be determined
first. In addition to the reactions from the upper arch (Vr, H1, M1 ) and from the
inverted arch (Va , Ha, Ma), all of which have been obtained previously, the walls
are also subject to direct pressure at their top (P), to their self-weight (G) and
to the active earth pressure (Ea)-
The resulting reactions will develop below the wall base (W) and behind the
vertical face of the wall. Owing to the friction between the wall and the ground
the bearing pressures along the plane of contact will have perpendicular as well
as tangential components. The friction forces can be calculated as the product
of the perpendicular pressures and the coefficient of friction (tan ¢ ).
Thus, the friction under the wall is
S2 = Wtan ¢,
where W = the ground reaction under the wall.
 302 DESIGN OF TUNNE L SECTIONS

The friction behind the wall (vertical face) is

S = tan </>(Eu + Ep),

where EP = the horizontal comp onent of the passive
earth pressure.
Finally, the friction along the top of the wall is

S1 =Pta n <J>.

As mentioned earlier, the above methods of calculating

the friction forces S 2
and S, are not accurate because they result in. maxim
um values which canno t
develop witho ut a certai n displacement (a) taking place.
In fact the amou nt of
friction will be limited to that actually required for
the equilibrium and that
which is propo rtiona l to the displacement of the surface
in question. Such displace-
ments can be calculated with the coefficient of subgrade
reaction. The displace-
ments required to mobilize friction forces Si, S and S
2 , are

<> = ep1
i C'

(In our case, illustrated in Fig. 4/22, tan </> can be taken
at its maximum value
as even the full S2 force has failed to balance the horizo
ntal forces and an addi-
tional, passive earth pressure (ePJ was required to develo
p in the plane of point 0.)
In the calculation of the friction forces the respective coeffi
cients of friction are

c51 ep1
/ 1 =tan</> - = tan</> - ,
a aC

f = tan</> -
c5 w
= tan</> - - ,
a bCa

where a can be determined from a surface friction test and

with the appro ximat ion
illustrated in Fig. 4/23. The value of a also depends on
the type of soil, on its
condition and moisture content. In most practical cases
it can be taken as less
 HORSESHOE-SHAPED TU NN ELS 303

F1G. 4122. Loads on

the rigid side wall

,..
Displacement L1

FIG. 4/23. Relationship between the mobilized friction and displacement

than I cm (1 /2 in). Whenever the displacements exceed this limit, the friction
can be taken at its maximum value (tan <J>).
The unknown quantities to be determined are the passive earth pressure (Ep),
the ground reaction under the wall (W) and the respective lines of action, i.e.
distances tP and t.,. Assuming that both passive earth pressure and ground reaction
are distributed linearly and in a compatible manner, the four unknowns can be
determined from the three equations of equilibrium plus a fourth one referring
to the rotation of a rigid body on an elastic subgrade.
.304 DESIGN OF TUNNEL SECTIONS

Because all forces on the wall will have to be in equilibriu m, the sum of the
horizonta l and vertical compone nts has to be zero as bas the sum of the moments
about any point, for example point O (Fig. 4/22).

I. (4.35a)
From the above equation W can be calculate d.

2. (4.35b)

-can be solved for Er

3. The moments about point O are also equal to zero

a
P + V1V1-H 1h1-M1 +Ma-H aha-Va va +Gfg -Wt .. +
2
+ Ea ta+ Ep tp + f1 Ph= 0 . (4.35c)

This equation contains the two further unknown s, tP and t,..

4. The fourth equation is based on the elastic subgrade and on the proportio ns
between bearing pressures and penetrati ons. Because the wall is assumed to be
infinitely rigid, the back face will remain a straight plane and will generate linearly
distribute d bearing pressures according ly. From the above assumpti ons it also
follows that vertical and horizonta l faces will remain perpendi cular to each other
and so the lines of pressure distributi ons will be at right angles as well. These
proportio ns can be expressed as

and since

and w = -
w
a b -

therefore

(4.35d)

The relations between wa, wp, ep 1, eP 2 and fp , t., are given by

(4.36)
 HORSESHOE• SHAPED TUNNELS 305

~nsider two more relations:

(4.37)

Both wb and eP1 can be expressed from Eq. 4.37 and then written into Eq. 4.36.
This will result in expressions for tP and lw which can be substituted into Eq. 4.35c.
The latter combined with Eq. 4.35d can be solved for wa and eP 2 • Finally :vb
and eP 1 can now be calculated.
This concludes the determination of the external loads and brings us to the
construction of the thrust line and to the investigation of stresses in the character-
istic cross-sections. These include the spring lines at top and bottom and the
section half way in between. The thrust line has to stay within the inner third of
each cross-section and the maximum stresses shall not exceed the allowable ones.
From the point of view of mechanics the use of walls with straight interior and
exterior faces is not recommended because they do not follow the thrust line and
result in awkward construction joints at the base.
42.122. The effects of relative displacements. In the above analysis the inter-
action between the members due to relative displacements has not been considered.
Of all the members, the arch at the top is the most sensitive to horizontal displace-
ments of the walls which, in turn, are the results of the reactions of the arch
itself. The horizontal displacements, added tt'l the a0 factors, will increase the
redundants X 1 , X 2 , X3 and
the loads on the arch in gen-
eral and may result in con-
siderable overstresses. It is
customary to assume that an
arch constructed of brick,
concrete or stone will crack
under the increased moments
and will work as a three-
hinged arch from then on. For
t'iis reason arches supported .
on walls subject to displace-
ments should be designed also
for a condition. with three ec-
centric hinges, two of which
Fm. 4/24. Assumed behaviour of
are located inside (at the tunnel section (as three-hinged
springings) and one outside arch) after displacement of the
(at the crown) (Fig. 4/24). The walls

20 Szechy : The Art of Tunnellin11

 306 DESIGN OF TUNNEL SECTIONS

developmen t of these " emergency hinges" will necessarily have to be accompanied

by local crushing of the wall material, resulting in damage to the waterproofin g
and in potential trouble spots.
Once the reaction at the springing is known, the displacement of the wall {LI)
can be calculated according to WINKLER as p=Cy. With the passive earth pre.;sure
and its distribution length determined (graphically or analytically), the displace-
1
ment is calculated as (A) = e~ where eP1 is the ordinate of the passive earth

pressure diagram at the springing.

Similarly the slope at the springing is (K) = epi ~heP2 (Fig. 4/22). Before
substituting the displacemen ts into the equations of redundancy they should be
El
multiplied by n = - - -2 where
l - µ
E and µ denote the modulus of
elasticity and the Poisso-'?- ratio
of the particular constructipn
material, respectively (for con-
crete µ = 0· 18) and I is the aver-
age moment of inertia of the
arch.
For a circular arch of constant
FIG. 4/25. The effect on the arch of horizontal moment of inertia and subject
displacements of lateral support to uniformly distributed loads
the equations of redundancy,
a lso including the effects of dis-
placements and rotations, can be written as (see also Eqs 4.26, 4.27 and Fig. 4/25)

X{ = -M +Ma=M' =- - - - = - 001 +K - pr 3 ( ~ - ¼ sin 2</>0 ) + 11 (K)

Ou 2r c/> 0

and

002 + K (I - Yo) - A
X/ = J-1 + H,, = H' = -- -- ------ = (4.38)
0 22

3
- pr4 sin c/>0 (- I - - 1 -
sin 2c/>0 sin c/>
-- - - - 0
)
+ n(K) (f - • ,.
Yo) - n ,~)
2 4 c/>0 3
=-- ------- ---'-=- -,,....--
2 ---'---- ------
(,1,. 2sin c/>0
+ -l sm
. 2 ,1,. \
3
r '+'o - - - - '+'o
c/>o 2 }
 HORSESHOE· l:HAPED TUNNELS 307

1 - - 1 sin 2¢ 0 - sin2 </>o )

• ,i..
- pr 4 sm'f'o -
(

2 4 <Po 3
i.e. ~=F=---,a-(_,1._0__~2-s-~~2~¢-o-~l-.--,1..~ )--+
'f' </>o +2 sm2'f'o

1
~Jµ 2 ep1 ~/p2 ~_ r(¢0 ~:in¢0) j- _ _~_J_µ_
1 2
e~
1

+ - - - - - - - - - ~- 2 -- - -- - - - -- (4.39)
2 sin
r 3 { </>0 - _ ¢0 1 .
__.:.. + - sm2¢0
)
<Po 2

The increases due to the displacements become higher as the coefficient of

subgrade reaction ( C) becomes smaller. Arches are as a rule very sensitive to
horizontal displacement of the springings.
It should be noted that very thick arches and very low coefficients of sub-
grade reaction (C = 1·0-3·0 kg/cm3) may yield excessive values for M,, and H,,,
which cannot be taken by the arch. It is in these instances that the arch may be
assumed to crack and transform into a three-hinged arch which is indifferent to
any further displacements. This, of course, can only be assumed if the formation
of cracks can be tolerated, i.e. if watertightness is not a requirement (see earlier).
In general, arches are also sensitive to vertical displacements but to a lesser
extent. Differential settlement of the walls, resulting in additional slope at the
springings, is not particularly common. If necessary, it can be taken into account
~imilarly to the above.
Assuming that there is sufficient friction to prevent any displacement (i.e.
eP2 = 0), only the rotation at the springings need be considered in the investi-
gation of the inverted arch. Even if this were not the case, eP 2 would be con-
-,iderably less than eP 1 and thus have a lesser effect on the inverted arch.
Any rotation K will result in an additional moment X{a = n(K) and in an
Gh
.·
a dd 1tiona I honzonta
· 1 t h rust X'?a = - a --Yo)
n(K)(f
-- - , bot h o f wh"1ch have t o b e
a;2
deduced from the original moment and thrust, respectively.

42.123. Example. The use of the analytical method will be illustrated in an actual example
of segment design, shown in Figs 4/26a-d. •
(a) Investigation of the arch at the top. The vertical loads on the arch consist of a uniformly
distributed pressure (geostatic load), a triangular load G1 in the corners and the self- weight
c,f the arch (Q). For a cover of 27·6 m (90 ft) these loads are as follows:

Uo = h0y 1 = 27·6 m X 2·0 t/ m 3 = 55·2 t/m 2 (11 ·3 k/ft2 )

fk fk 2·925 X 10·30
Go =- - y1 = 2·0 = 20·4 t/ m (13'7 k/ft)
3 3

~·
308 DESIGN OF TUNNEL SECTIONS

m'f.·IJ-66(j(JJ

--- -
I
---t----- -- s
I
1 - - - - - ' - - = ' - ' - - - - ---, 1
w f, • /J6/Kl0/l

1 . t-,os 'P. f-0-1lJ711

I lm ~ - 2/F1()711 "/l·2tJ7!

FIG. 4/26. Numerical example for the analytical design of the roof arch and of the invert a rch
of a shield chamber

24'o 120
Q = Yb · n 2m = 2·3 5·35 X 2n = 31 ·0 t/ m (20·8 k/ft).
2 360

Assuming that the entire vertical load is uniformly distributed over the span of the neutral
axis

_1_
0 ·_3_x _5_5_·2_+_ 3_1·_0_+_20_·_
4 = 67 _0 t/ m (
45 _0 k/ft).
9·27

The horizontal load can be calculated from the vertical load with the Rankine ratio. To
be on the safe side, the cohesion will be taken at half value: If cf, = 17° and c = 28 t/m'
(5·73 k/ft2) then

e0 = g 0 Aa - c -Jf,, = 55·2 X 0·5475 - 28·0 X 0·74 = 9·5 t/ m 2 (l ·.95 k/ ft")

el= 91 J.• - c Jr.= 61·1 x 0·5475 - 28·0 X 0·74 = 12·7 t/ m 2 (2·60 k/fi 1 )
where
).• = tan 2 (4'.l 0 - cf,/2).
 HORSESHOE-SHAPED TUNNELS 309

lo the next approximation the trapezoidal stress diagram must be evened out and thus

eo + e1 9·5 + 12'7
e=---
2 2
= 11 ·1 t/m 2 (2'2.7 k/sft)

or, when projected to the extrados

2·975
p
H
= e --
· -
2·675
= 12'3 t/m (8·27 k/ft).

The thrust at the springing is calculated from Table 4/I. For

75
ff I = 2-6 = 0·287
9·27

'XM = 0·00939, aH = 0·4562, {JM = -0·00888 and {JH = -0·1588

;,.nd the thrust a nd moment at the springing (due to vertical loads)

Mv = aMPv / 2 = 0·00939 X 67·0 X 9·27 2 = 54·0 tm (391 ftk),

H. = (T.H Pv I = 0·4562 x 67·0 x 9·27 = 238·0 t (524·7 kips).

The vertical reaction is

V
V
= Pv l
2
= 67 '0 X2 9-27 = 310·0 t (683-4 kips) •

N. = ..jH; + V; = 420 t (925·9 kips).

The thrust and moment at the springing (due to horizontal loads)

MH = fJMPHl 2 = -0·00888 x 12·3 x 9·27 2 = -9·4 tm (-68·0 ftk),

HH = flHPHI = -0·01588 x,12;3 x 9·27 = -18·2 t (-40·1 k),

The reaction due to combined horizontal and vertical loads is then

M 1= M v + MH = 54·0 - 9·4 = 44·6 tm (322·6 ftk),

H, = H v + HH = 283·0 - 18·2 = 264·8 t (583·8 k),

V, = Vv + VH = 310·0 - 0 = 310·0 t (683·4' k).

The stresses are investigated only at the springing and at the crown, and the relief offer~d
by the horizontal load is_neglected, which is another assumption on the safe side.
310 DESIGN OF TUNNEL SECTIONS

The cross-sectional area of the arch (considering a strip of unit width) is

A = IO0X 120 = 12 000 cm' (1860 in 2)

and the section modulus is

100 X 1202
K = - - - - - = 240 000 cm 3 (14.645 in3),
6

the stress at the spring line

N. M, 420 54·0 tm • /
a =- ± -.-= --- ± = 350 ± 22·5 = 372'5 t/m- = 37·25 kg cm (5320 psi)
A K 1·2 m• 2·4 m 3
or
327·5 t/ m 2 = 32·75 kg/cm 2 (4670 psi).

(b) In the investigation of the inverted arch the construction sequence is also to be consid-
ered. Thus the weight of the arch at the top and the weight of the walls - both members having
been completed prior to the inverted arch - is transferred to the ground at the base of the wall~.
In another assumption on the safe side the entire pressure on the arch at the top and on the
walls is again assumed to be transferred to the ground through the walls and through the
inverted arch and is reduced only by the friction behind the walls. Some of the pressure ob-
viously must have been transferred to the ground through the walls completed earlier.

Because
P = afk Yt = I ·60 X 2·925 X 2·0 = 9·36 t (20·64 kips)

and
S = 120 t (264·6 kips)

(see later under wall design}

p; = Uo +
G0 + 2P
L
- 2S
= 55'2
20·4 + 2 X 9·36 - 240
+ ---- -- - - = 55'2 - 14·9 = 40·3 t/ m.
13·50
(27· 1 k/ft)
The components of the horizontal load are

et = [uo + (11·87 x 2·0)] ).0 - c .Jf. = (55·2 + 23·74) 0·5475 - 28 x 0·74 = 22'5 t/ m,
(15· I k/ft)

e3 = [no+ (13-46 x 2·0)] ;.. - c -Jf;, = (55·2 + 26-92) 0·5475 - 28 x 0·74 = 24·3 t/m,
(16·3 k/ft)

22·5 + 24·3 1·59

_ = 24·5 t/m (16·5 k/ft).
2 1 52

Selecting the appropriate ex and fl values from Table 4/ l the reactions at the springing of the
inverted arch can be, calculated.
 HORSESHOE-SHAPED TUNNELS 311

Owing to vertical loads:

Mv -= ~M Pv 1'2 = 0·00485 X 40·3 X 7,35z = 10·6 tm/m (23·4 ftk/ft},

H 'y = aHPvf' =--- 0·6145 X 40·3 X 7'35 = 182·0 t/m (401·2 ftk/ft),
p ' I' 40·3 X 7'35
V'v = _v_ = - - - - - = 148 t/m (99·5 k/ft;,
2 2

Owing to horizontal loads:

M11 = /JMPnl'2 = -0·0049 X 24·5 X 7'352 = -6·50 tm/m (-14·3 ftk/ft),

H11 ~ /JHPnl' = -0·1146 X 24·5 X 7'35 = -20·6 tm/m (-45'4 ftk/ft},

The sum of the forces at the springing:

Ma= Mv + M" = 10·6 - 6·5 = 4·1 tm/m (9·04 ftk/ft),

Ha = Hv + H 11 = 182·0 - 20·6 = 161 ·4 t/m (108·5 k/ft),

v; = v,, = 148 t/m (99·5 k/ft),

N; = ..}161·42 + 1482 = 219 t/m (147'2 k/ft).
With a cross-sectional area of

A = l·0x0·9 = 0·9 m 2 (1395 in2)

and with a section modulus of

1·oxo·9 2
K = --- = 0·135 m 3 (8238 in3 )
6
the stresses are
219 4·1
a = -0·9- ± - -=
0·135
243 ± 30·3 t/m 2

i.e.
Omax = 27·33 kg/cm2 (389 psi)
and
Umin = 21 ·27 kg/cm2 (303 psi).

(c) This brings us to the investigation of the walls. Because the base is held firmly in place
'>)'the inverted arch, the wall is assumed to rotate about its lower outer corner. This will
result in passive earth pressures behind the wall being distributed in a triangular form and
with a resultant acting at 2/3 of the height. As t • is thus known right from the beginning the
calculations become considerably- simpler. (Fig. 4/27).
Other assumptions are:
1. The ground may be regarded as a homogeneous, isotropic material, in which the strain~
can be calculated ~ith a constant coefficient of subgrade reaction in every direction.
312 DESIGN OF TUNNEL SECTIONS

~ • :Jt(HJ t/m

Ht· 264"8tjm
Mr· 44·6 tm/m
Ya • 196-0 t/m
Ha· 21.9·8 t/m
Ma· N tm/m
P - g,. a -6t08. t6·98t
S • 120 t (est,m,;Led r,1/ve}
S, - Pf4n f • .98x O·JP6 - JO-O t Fm. 4/27. Numericai
example of
6 •50t wall design
e, - 12·7 t/m 2
e1 • 22-5 t/m 2

2. The wall is infinitely rigid and the back face remains a plane throughout the rotation.
From the equilibrium of vertical forces

w = v,+ r + G - s - v;
where
P' = 1'60 (27-60 + 2·90) = 98 t (216 kips),

G = (2'15 + 3·07) ~ 8·94 X 2'3 = 50 t (110 kips),

thus
W = 310 + 98 + 50 - 120 - 148 = 190 t (419 kips).

The active earth pressure

Ea = ei + e2 h = 12 ' 7 + 22 ' 5 8 ·94 = 157 ·0 t (346 k'1ps) ,

2 2

2e 1 + e. h 2 X 12·7 + 22·5 8·94 ( f

ta= • - = - - - - -- - - = 4·05 m 13 ·29 t).
e 1 + e2 3 12·7 + 22·5 3

and the moments about corner 0

P'p + S 1h + V,v, - Mr - H,h, + Gg - Haha - Vava +Mn+ E 0 t0 + E,t, - Wt..,= 0,

where S 1 = P tan ¢. This can be rewritten as M 0 + E, t • - Wt.., = 0, where M ~ denotes the

sum of all known quantities. Considering the proportion of pressures under the wall to those
behind the wall and the proportions to the common angle of rotation we get

e, : (w1 - w2) = m : b.
 1 '.0miESHOE·SHAPED TUNNELS 313

From these relations, combinea with the equation of moments, Ep can be expressed as

Wb- 2Mo
Ep = 4 bl •
-3 m+ -3m-2
or, whith numerical values

M 0 = 98 x0·80 + 98xtan17° x8·94 + 310 x2-115 - 44·6- 264·8 x8·58 + 50 x 1·10 -

- 161·4 x0·39 - 148·0 x3·08 + 4·1 + 157·0 x4·05 = 78·5 + 274·0 + 660- 44·6 -
- 2270 + 55 - 63-0 - 456 + 4·1 + 635·0 = 1706·6 - 2833·6 = - 1127·0 tm
(-8152 ftk),

190 X 2·76 + 2 X 1127·0 2778·0

E, = 4 2.763 = +~= 232 t (511 kips),
3 8"94 + 3 X 8·942
S = (Ep +E 0) tan cf, = (232 + 157·0) 0·305 = 118·5 t (261 kips).

(This is in good agreement with the 120 tons (265 kips) assumed at the beginning.)

e. = T2£ =
2 x232
""""s-94 = 51 ·7 t/m 2 = 5·17 kg/cm2 (73·5 psi),

The average bearing pressure

W 190 •
w. = b = _ = 69 t/m1 = 6·9 kg/cm· (98·1 psi),
2 76

w1 = w0 + -Llw
2
- = 6·9 + 21·6 =
.
7·7 kg/cm2 (109·5 psi),

Llw 1·6
w2 = w0 - - - = 6·9 - = 6-1 kg/cm2 (86·8 psi),
1 2
t
2w2
= - - -- - = -
+ w1 b 2x6·1 + 7·7 2·76
- -- - - - = 1·28 m (4-2 ft),
"' W1 + W2 3 6·1 + 7·7 3
and the friction

S: = H , + H 0 - S1 - Ep - E 0 = 264·8 + 161 ·4 - 30·0 - 232 - 157 = 7·2 t (15·87 kips).

Thi~ would require a friction coefficient of

f =-
· = -
s. 7·2
=0·04
W 190

\:Orresponding to cf, ~ 9°, which is considerably less than the available maximum i.e. the actual
angle of internal friction.
314 DESIGN OF TUNNEL SECTIONS

Having determined all forces on the wall the thrust line can be constructed. Because the
maximum eccentricities occur at the crown and at the springing of the inverted arch the stresses
also should be investigated at the same sections.
We shall now investigate the changes in the loads on the arches if the effects of wall displace-
ments at the springings are also to be considered.

With C = 15 kg/cm3 the horizontal displacement at the spring line is y = e~ ;

5 7
(Ll) = ec, = · I kg/cm' = 0·345 cm (O· 136 in)
15 kg/cm3

and the rotation of the wall is

(Ll) 0·345
(x) = - - = - - = 0·385 X J0- 3
h 894 .

The moment of inertia of the arch is

1·oo x 1·2O2
l ,= - - - - - = 0·12 m• (288 240 in 1); Econc = 210 000 kg/cm2 (2987 k/ in~)
12

and the Poisson ratio of the concrete is

µl = 0·0324.

The magnifying factor for the displacements is

EJ 21x105 x O·12
n = --- 2
= - - - - - = 2·6x 10·' tm2 (6 168 000 ft'k).
1- µ 1 - 0·0324

and thus the magnified displacement and rotation

Ll= O·345x1O - 2 x 2·6x1O5 = 895 (tm 3) (69660 ft 3k),

x = O·385 x JO - 3 x 2·6 x 105 = 100 (tm 2) (2372 ft 2k).

The equations of redundancy thus become

aux; + 0 0 1 - " = 0'

0 22 X~+ Onz - " (f - Yo) - Ll = 0.

In the first equation the sign of " is opposite to that of a01 since the displacement will
diminish the rotation of the end section. In the second equation the sign of a 0 2 is opposite
to that of Ll but is the same as that of the displacement due to "· From Eqs 4.23 to 4.27

60° :n:
a 11 = 2r,1..
'l'O
= 2 X 5·35 - - = l 1·2O5 m (36·76 ft) ,
1800
 HORSESHOE-SHA PED TUNNELS 315

2 sin 2 60° \
a! 2 == r 3 ( cf,,, -
2 sin: cf,,.
<Po + 2I . ) _ .
sm 2cf,0 - 5 35 3
( 60;-r
180 - 600 n 180° + 21 sin 120 i ~

= 154 ( ; - 1·43 + 0·433) = 7-9 .

Since a 01 = - pr3 ( ~o - ! sin 2cf,) = - 67·0 x 5·353 (

2
~
0
;
1 00
- ! sin 120° ) = - 322.

3
cf, )
Go: = -
, .
pr sm <Po
(1 1 sin 2cf,0
2 - 4~ - -sin-3 - 0 =

=- 67·0 X 5·35' Jf
2
(~2 - 4~ 32n 3
$ - -4 X-3) = -2190 '

n ,ff)
(.3 - - 2-
r(cf, - sin cf,0)
Yo = ----- = 5·35 - -----'- 3 = 0·93 ,
<Po n

after substitution the redundants can be obtained as

322 - 100
11·2O5X( - 322 + 100 = O; x; = _
11 205
= 19·8 tm (143'2 ftk),

7-9·x;- 2190 - 100 X 2'925 - 0·93 + 895 = 0,

,
Xi =
2190 + 199·5
.
- 895 .
=189 t (417 kips).
79

It must be noted here that the thrust at the springing is greatly influenced by the displace-
ment (in the example 30%), The horizontal thrust at the springing, for example, was reduced
from 283 tons (624 kips) to 189 tons (417 kips), resulting in a substantial increase in the positive
moments at the crown. Any displacement at the top of the wall will cause only rotations at
the springing as the wall was assumed to rotate about its lower corner. Thus the equations
of redundancy become
x; a11 + ~ I + K0 = 0,

The magnifying factor is

l·O X 0·93
21 X 105
12 2
no = - - 1- -- 0--1-8_%_ _ = 1·32x 105 tm 2 (3 131 600 ft k),

and so
X0 = 0·385 X 10- 3 X l ·32 X 105 = 50·8 tm: (1205 ft 2k),

( ; - sin 45°)
<Po) = (1 ·71 f)
a
Yo =r sin
(cf,~ - <Po 5·20 - -- n 4 = 0 ·52 m t ,
 ---------------

316 DESIGN OF TUNNEL SECTIONS

a , :t
a, 1 = 2,¢0 = 2
4 =
X 5·20 X 8·20 m (26·90 ft),

a
= r3 (,I.' sinZ 'Po
2~ . 2¢ ;J
+ 21 sm 1·4 1J
4 --;- + 2 =
a,: ( :i
'l'O- 0 = 5·23 1·85,

furthermore
, = - pr 3 ( -'Po - -I
a01
2
. 2¢0')
sm
4
= 40·3 x 5·23 ( -:i - - I) = - 810
8 4 '

, 2
- -sin-3-¢ 0)
• . , (1 sin 2¢ 0
ao, = - pr sm 'Po 2 - 41 ~ =
= - 40·3 x 5·2• x 0·101 (o·5 - ..!..:i - ..!..)
6
= 444.
Both x and a01 will be of the same sign.

860·8
X;a X 8·20 - 810 - 50·8 = 0; x;a = + 8·20 = + 106-20 tm (768 ftk);

x;a X , 393·2
1'85 - 444 + 50·8 (l ·52 - 0·52) = 0; X2 =- - = 212 t (167 kips).
a 1·85

Here the effect of the rotation is favourable as far as the crown is concerned because the
increase in the horizontal thrust (about 15 %) diminishes the positive moment.

42.13. Common Deformations and

Composite Action of the Surrounding Ground (Davidov's Method)

A more accurate design method has been developed by Prof. DAvmov 4·3 in
conjunctio n with the experiments and research especially carried out for the
design of the Moscow Subway. This method is particularl y applicable to major
undergrou nd structures built by classical constructio n methods. The members
of structures of such proportion s, constructe d in separate successive phases, lend
themselves to design as separate units. According to DAvmov the arch and the
inverted arch are to be designed separately and the calculated reactions a re to •
be applied to the walls as external loads. However, the displacements at the spring-
ings, common to arch and wall because of the contact between them, will be
included in the calculations. In the design of the walls, on the other hand, the
surroundin g rock will be assumed to participate in carrying the loads and support-
ing the walls.

3
•· DAvmov, S. S. : Raschet i proiektiroi:anie podzemnikh konstrukcii (Computati on and
design of subterranea n structures). Gosud, ,I,;q~!- Stroitelnoi Literaturi, 1950

--..\
-~-.
",1. .........,
f \ • · •' '
f;,

~ ,:_
·~·-..-.. ~
,:~
.,..,
{'lo , ~,,ia
-.
,"......,, ,,\,'-._,"-f
~ ••

, ... .:,
• {e
 - ------- --- - - --

HORSESHOE-SHAPED TUNNELS 317

According to DAVIDOV the effective depth of the cooperating ground will

depend on the magnitude of transferred stresses and will be defined by the range
within which these aoditional stresses do not exceed the initial geostatic pressure
by more than 20 %. Thus, the wall can be assumed to be elastically supported
and surrounded by the "composite ground" the thickness of which can be cal-
culated as above and will be
denoted by H 9 and H b for be- p.
hind and under the wall, re-
spectively. In his treatment of
the elastic foundation DAVI-
QOV also considers the inter-
action and accumulation of
penetrations in adjacent areas.
Instead of a continuous bed- e
ding the foundation is assum-
ed to consist of a finite
number of fictitious elastic
rods with hinges at top and
bottom (Fig. 4/28).
The tunnel wall, leaning
against the surrounding
ground, may then be regarded
as a cantilever with a displace-
ment Yo and a rotation </>0 at F10. 4/28. Structural mechanism of the DAvmov
design method
the fixed end and subject to
the external loads, and to
the ground reactions in the I
form of forces Xi, X 2 , ••• X n 0
exerted by the elastic rods
{Fig. 4/29a).
The rigidity of the wall is
to be taken into considera- @
tion as

rtEo 1 - µ2 a
oc= - - - - - c FIG. 4/29. The wall acting as a cantilever
6EJ 1- µ~ •
called the rigidity factor .
For values of oc < 0·05 the wall is practically infinitely rigid and the deforma-
tion of the cantilever may be neglected. However, for oc > 0· 10 the wall has to
be i:egarded as flexible and the deformations considered accordingly.
In the formula, E0 is the modulus of elasticity and µ 0 is the Poisson's ratio of
the ground; El denotes the rigidity andµ the Poisson's ratio of the wall and c is the
spacing of the fictitious bars which can be. 41ken as 1/5 of the height of the wall hr
._\,,. .1•r C> .
.,.'<.•~,,)'I ..'?
;~~t
.;: - Jf~ ~-~
'
<: -f,-

~ ; ' t\t"'~
"1:J] \'1''
 318 OESIG1' OF TUNNE:.. SECTION~

In order to simplify the calculations and to reduce the volume of computation s,

a number of tables have been developed by the author for rigid as well as for
flexible walls; these can be found in his work referred to earlier.
42.131. Davidov's theory for clear spans. An approximate but more practical
theory by DAVIDOV can be used for clear spans up to

I0 -< 8Jfkr
p .
I
where/k, = a factor of rock strength (by PROTODYAKONOV, see Table 3/VI), p = •o
r
Jo
where / 0 = the rise of the arch. Under these conditions the horizontal
resultant of the external volumetric forces, E0 , just about balances the thrust at
the springings. If E 0 and the external loads on the wall are in equilibrium, the
rotation of the base is zero (</>0 ~ 0) which also eliminates any rotation at the
spring line </>~ = 0 and Xn = 0. Based on the above formulae this method can
be used for the following conjugate values:

fkr = 0·6 0·8 1·0 2·0 4·0 6·0 10 15 20,

/0 = 3·5 4·0 5·0 6·0 8·0 10 12 14 16 m

I. The design of a rigid and fixed wall (c/>0 = 0) is simpler because

(a) The elasticity of the surrounding ground need not be considered;
(b) The forces acting in the elastic centre of the arch can be calculated as

where M ~. H ~ and Qg denote the moment, horizontal

and vertical force, respectively, due to the external
loads and acting in the elastic centre of the fixed arch;
(c) The X 1 forces, representing the elastic ground
resistance, will not be mobilized and so X 1 = 0;
(d) The elastic reactions (Y1) under the wall base can
be assumed to be uniformly distributed;
(e) The forces acting in the elastic centre of the
inverted arch can be calculated similarly from
M~ = M~; H° = H~ and Q; = 0
where M~ and H~ denote once again the moment and
the horizontal thrust, respectively, due to the external
loads and acting in the elastic centre of the inverted
h, arch. The walls are assumed to remain updisplaced
under the thrusts from the arches. While the diagram of
lateral pressures behind the wall is assumed to be trian-
gular, the bearing pressures under the base can be taken
Fm. 4/30. Simplified method for as uniformly distributed. Finally the wall is designed
the design of rigid walls as a statically determinate structure.
 HORSESHOE·SHA PED TUNNELS 319

The unknowns u, u" and X can be determined from the equations of equilibrium (Fig.
4/30).
1
~X=0; H 1 +X+ Ha- uh,=0,
2

o o I uh, - =
~Y = 0; Q, + P + G - Qa - µ 1
2
17x hx 0,

•
~M=0; M1- Ma + H I h1 + Ha ha - o o
Q, q1 + Qa qa - Pp - Gg -
1
3
•
uh; + 2I a:,ch; = 0.
(4.40)
Letting
H~ = H 1 + Ha,

Q~ = Qo + P + G - Q~ '

the formulae reduce to

1 l 0
X -
2
uh, + H., = 0,
0
hence X =
2
ah, - H .,,

M,,,0 -
I
3
•
uh; + 21 ux h-;,• = 0 ,

these two equations yield

Q~ 1 h.
and a =-- - - µ1 - u.
hx 2 hx

With these two relations the loads can be computed for any cross-section. For example at
the centre of the wall base

I uh·., - -1 h µ - 1 uh = 0
+ P (-hx2 - p) + G (-hx2 - -dtal
2-
)
- -3 , 2" 1 2 y'
(4.41)

I 1
N = Q~ - Q~ +P+G- µ1
2 u h, = l7x hx and T =
2
uh,-Ha = X. (4.421

2. An elastic and flexible wall can be designed for the cJ,0 = <Pb = 0 conditions as follows :
(a) The range of the ground in compression is disregarded;
(b) The X 1 and Y, forces, representing the elastic ground resistance, are also neglected ;
(c) The bearing pressure diagrams behind and beneath the wall are assumed to be linear;
(d) The wall is assumed to remain undisplaced (Fig. 4/ 31).
 J20 DESIGN Of TUNNEL SECT.''.<1NS

p
6 Because of the flexibility df the wall the pre.~·
sures due to the external loads are assumed to be
transmitted to the ground only over a certain por-
tion of the height. Furthermore, the horizontal re-
action, X, acting at the bottom of the base, is as-
I !/4 sumed to be insignificant and can be neglected.
I The forces acting in the ,elastic centre of the arch
hy are again calculated as they were in the case of the
rigid wall, i.e. assuming the arch to be fully fixed.
Once again the wall in itself is a statically deter-
minate structure for which the unknown quantities
o, a.. and Yo can be calculated from the equations
of equilibrium:

IX = O; 1 . . 2H
- ay0 - H 1 = 0 g1vmg a = - -1
FIG. 4/ 31. Simplified method for the 2 Yo
design of elastic walls
IY= O;

IM = O;

2H (4.43)
Substituting o = - -1 in the second and third equation
Yo

0 X =
hence

hence

r
3 lH, (h,. - h,) + Qr q,+ Pp+ Gu+ hx a
M , - 2<Q, +P + G + µ 1 H,)_]
Y, , =
H,

,.
nnd thus the loads can be determined for any section of the wall; fo t ..:xample, at the centre
of the wall base

(4.44)

ay., ayo
N = Q1 + G +P- µ1 - -
2
= axhx and T = --;;;- - H, = 0. (4.45)
J.
 MONOLITHIC HORSESHOE-SHA PED TUNNEL5 321
DAvmov's r.iethod can also be used to determine the coefficient of subgrade reaction in
cases where f'n varies with the depth of the compressed layer (H).
Regarding the wall

and regarding the base

5n£0
C: = (I - µfJhzYo .

A numerical example can be found in DAvmov's book, referred to above,on pp.251-271.

Because of its complexity and limited accuracy the method has not gained general acceptance
and is used only in the design of exceptionally large underground structures built by classical
construction methods.

4.22. ANALYTICA L DESIGN METHOD TREATING THE SECTION

AS A WHOLE AND CONSIDERI NG COMPOSITE ACTION OF
THE GROUND

42.21. Method of Zurabov and Bougayeva

ZURABOV and BouGAYEVA consider the structure as a continuum and take into
account the support offered by the ground in the form of an elastic foundation.
This method is considerably more accurate than the previous one, not only because
of its more realistic approach of treating the horseshoe section as a continuum
rather than a composition of arches and walls, but also because of its treatment
of the support offered by the surrounding ground. The latter is assumed to develop
through elastic subgrade reactions consistent with the actual deformations rather
tha n through passive earth pressures, which would require displacements that
may not materialize at all under the given set of conditions.
More often than not the vertical loads will cause the walls to spread apart.
This trend is resisted by ground reactions developing in proportion to the defor-
mation of the structure and to its penetration into the ground. On the other hand,
the top of the tunnel will deflect without giving rise to any grouad reactions.
In this method, then, the ground reaction is assumed to be proportiona l to the
structural deformation but only at the points of maximum deformations. In all
other points the bearing pressures are assumed to be distributed in an arbitrary,
approximate manner. Also the locations of the initial zero points of the pressure ...
diagrams are assumed arbitrarily.
The structure is assumed to have elastically fixed ends. The rotations of the
lower corners can be calculated in proportion to the elastic compression of the
gro und and the horizontal displacements are assumed to be zero at these points.
The ground reactions are assumed to be distributed as illustrated in Fig. 4/32.
An equation for the diagram is given as a function of the locations of the maximum
ordinate (q) and the zero points. The upper zero point has been observed to vary
•·,ithin narrow limits and may be taken at cf,0 = 45°.

21 Sz~chy : The Art of Tunnellina

322 DESIGN OF TUNNEL SECTIONS

The maximum lateral pressure

Gnmiimmii nnmn1111111uP is assumed to occur where the
1 section is the widest, but shali
not be taken as lower than 1/3 of
the total height h. The direction
of the reactions is radial.
The maximum ordinate of the
pressure diagram is calculated
h as q = Cbh where bh denotes the
maximum horizontal displace-
ment due to the combination of
external loads and ground reac-
tions and C denotes the coeffici-
ent of subgrade reaction. For
compound circular sections it is
FIG. 4/32. Load diagram for a monolithic customary to assume the follow-
horseshoe section (according to ing approximate pressure distri-
ZURABOV and BoUGAYEVA)
bution:

~
2
at the top q' = q (i _ cos ) • (4.47)
cos2 cf>o ·
2
at the bottom " =( q lsin
q -~ ~ 1 ). (4.48)
sm 'f'n
If ~ = c/>0 then q' = 0 and if ~ = 90° then q' = q and in the second equation,
if ( 1 = 0 then q" = q and if ~1 = </>n then q" = 0.

If the lower section consists of

straight vertical walls

q" = q rI - ~u ,( 4.49)

(For the notations see Fig. 4/32.)

From the above equations it
will be noted that once the zero ,.
points have been assumed , the
only remaining unknown to de-
termine the reactions is q. This
is to be obtained from the dis-
placement which can be calculated
with the coefficient of subgrade
dftp reaction:
FIG. 4/ 33. Deformation qf the monolithic section (4:50)
 - --- -- - - - ----

MONOLITHIC HORSESHOE-SHAPED TUNNELS

. 323

where fh can be written as (Fig.,4~33):

where [)hp denotes the structural deformation due to the external loads and bh 1
is the structural deformation due to the
unit reaction : q = 1.
Eq. 4.50 can be rewritten as
/.
/
I
I
I
--r-+-------,----- ---
hence 1
\
\

(4.51)

If N and M denote the axial load and

the moment due to the external loads,
respectively, the rotation of the elastic- F10. 4/34. The effect of the rotation of the
ally fixed ends is (Fig. 4/34) : fixed end

N 6M N 4M
<Jk =- + - .-. <Jb=-- -2-,
mo mo mo mo
12M M
(4.52)
fJ = Cmg = CJ '

where J = the moment of inertia of the end section.

With the above conditions and assumptions established , the calculation can
be carried out in the following steps:
l. Disregarding the passive earth pressure the statically indeterminate and
elastically fixed structure is to be solved under the external loads (Fig. 4/35).
2. The same structure is to be solved for the unit lateral load q = 1 t/m , distri-
buted as above (Fig. 4/35).
3. The horizontal displacement at the point of maximum lateral pressure shall
be determined under the external loads as well as under the unit lateral load (J,. 1)
( Fig. 4/35).
324 DESIGN OF TUNNEL SECTIONS

4. The maximum ordinate

of the lateral pressures is cal-
0 culated by equating the struc-
tural horizontal deformation
to the elastic compression of
the ground (Eq. 4.51). Thus
all external loads have been
dete_rmined and the moments,
axial loads and shear forces
can be calculated at any cross-
section.
In the above calculation the
same structure is to be solved
for various external loads.
Because it is three times re-
dundant it will have to be cut
through at an arbitrary sec-
tion (e.g. in the middle) and
then subjected to the r~dun-
dants. The latter will consist
of a moment (XJ, a horizon-
tal force (X2) and. a vertical
force (X3) and these will have
Flo. 4/35. Separ~ting the loads into to be determined in such a
(a) upper and way as to eliminate a_ny dis-
(b) unit lateral load continuity at the disconnected
ends. If X 1 , X 2 and X 3 are
the forces acting on one half
of the structure then the other half will have to be subject to - Xi, - X 2 and
-X3, i.e. to forces of the same size but of opposite directions. Because both the
structure and the loads are symmetrical about the vertical axis in most practical
cases, X 3 is usually zero.
In order to simplify the calculations the structure is split through the centre
and the redundants X i, X 2 (and -X1 and - X2 on the other half) are shifted to
the elastic centre, which is once again the centre · of gravity of the elastic quan-
Lls
tities
EI

The eq11ations for X 1 and X 2 are

(4.53)
 MONOLITHI C HORSESHOE-S HAPED TUNNELS 325

where a0 1 and a0 ~ = the relative rotation and horizontal displacement due to

th e external loads, respectively
a 11 and a 22 = the respective movements due to the redundants X 1 = 1
and X 2 = I
a21 the horizontal displacement due to X 1 = 1
a 12 the relative rotation due to X 2 = l.
All these factors are to be calculated in relation to the elastic centre.

FIG. 4/ 36. Equivalent

determinate structure of the horseshoe section

ll·lt

FIG. 4/37. Horizontal

movement at the spring line of the cantilever

In calculating the relative movements the elastic rotations of the supports also
have to be considered

where a~ 1 is the rotation of the structure with fixed ends and /Jo is the rotation
of the structure with elastically constrained ends. Similarly (Fig. 4/36)
"

Substituting in Eq. 4.53

(4.54)
326 DESIGN OF TUNNEL SECTIONS

From the above two equations X 1 and X 2 can be calculated. Note that

a01
, = j' Mom1
EJ ds =
f MoEJ ds ,
au= _I
I r ds
EJ , a 2I 9
-
s y2
= -EJ ds .

Horizontal displacements <>hp and <>hl also have to be computed for the calcula-
tions.

where <>"° is the horizontal displacement of the cantilevers due to the external
loads and <> 1 and [) 2 are those due to X 1 = 1 and X 2 = 1, respectively.
It is best to calculate the displacements from the equations of virtual work .
For example, a fictitious force H = I can be assumed to act at the point in. question
and then the exterior and interior work can be equated (Fig. 4/37):

(4.55)

The value of <>hi can be computed in a similar manner.

The use of the above niethod is illustrated in the following section by a numer-
ical example.

42.22. Numerical Example of the Zurabov-Bougayeva Method

Consider the horseshoe section composed of two .circular sections as illustrated in Fig . .,
4/ 38. The moment of inertia is constant througho·uc.
The tunnel is subject·to a uniformly distributed vertical load of p = 1 t/ m. The ground is
assumed to be elastic and capable of developing lateral support. In the first part of the calcula-
tion the supports are assumed to be r·igidly fixed.
The data of the section are:

7t
r = 3·0 m, t/> 0 = 45°, tJ>. = = 22° 30'. J = cons!. = 0·0104 m•,
8
C = 50 kg/cm3 = 50 000 t/m3, E = 3 000 000 t/m\ a = 2.
- - - - -- - - - --- - - --

MONOLITHIC HORSESHOE-SHAPED TUNNELS 327

!11 1!11111 11111!/llllllli!l lll ll!ill!lll!li!IP =I t/m

FIG. 4/ 38. N umerical example for the q

design of a monolithic
horseshoe section

The ground reactions along the upper part of the section are:
(substituting <Ph into Eq. 4.47)
q' = q(l - 2 cos2 ~) ,
a nd a long the lower part of the section:
(substituting </>0 into Eq. 4.48)
q" = q (1 - 6·8286 sin' ~1).

I. Find the elastic centre.

f yds
c, =Jds'
::r 3:i:
J
;r
ds = r --
2
+ 2r -8 = -4 r = 2·356 r '

y' = r ( I - cos ~) (upper part)

y" = r + 2r sin ~ 1 (lower part)

•n
J J
n /':!

+ 2 sin e1) d~1 = r 1 \ ~ -

12

j y ds = r2 (1 - cos ~) de + 2r 2 ( I sin ~) : +
O 0

+ 2r2 (e 1 - 2cos e 1 ):" = r2 (; .- I) + 2r 2

(</>.- 2cos <Pn+ 2) =

= r2 (} + 3 + 2</>n - 4cos <Pn ) '

substituting the numerical values

J y ds = r2 ( 1·571 + 3 + 0·7854 - 4 x 0·9239) = 1·666r2,

I ·666r2
c1 = _ ,. = 0·708r .
2 356
 328 DESIGN OF TUNNEL SECTIONS

p, f t/m
1101r01110,1111111111~

FIG. 4/ 39. Loads and moment diagrams on the equivalent determinate structure

2. In the next step the structure is to be investigated under the external loads without
considering the lateral ground support. Because of sy.mmetry the structure can be regarded
as one with two redundants. Assuming the two redundants X 1 and X in the elastic centre
2
and assuming rigidly fixed supports (Fig. 4/39) with the notations already given

f
.mr
a 11 = . -EJ ds; au =
f mi
EJ ds;
•

In these expressions M denotes the moment on the equivalent determinate structure due
to the external loads and m 1 and m2 are those due to X 1 = 1 and X: = l, respectively. The
moment diagrams on the equivalent determinate structure are illustrated in Figs 4/39b-
4/39d.
 Jd'.ONOLITHIC HORSESHOE-SHAPED TUNNELS 329

As the moment of inertia is constant, the relative rotations may be 'magnified' by EJ in

the cak:ulations.

a; 1
f. f
= m; ds = ds =
4
3.n r = 2·356r ,

rr./Z .,.

a;z J
= m; ds = I [- c,.+ r (I - c;os e)]Zr cte + J(r - c,
0
+ Zr sin e1)Z 2, cte1 •

~12
••
a 1z = r I'! c1- 2rc 1( 1 - cos e) + r 2 (1 - 2 cos e+ cos2 e) ] de + 2, J[c; ·- 2c, +r
0 0

t- r 2 - 4 (C1 - r) r Sin e1+ 4r2 Sin! e1] del = r[c:e- 2rc,(e - Sine)+ r! (e - 2 Sin{+

+ ¾sin 2e + e12) rz + 2, fc; e, - 2c, rel+ ,z + ,1 4(c, - r) r cos el +

+ 4r
2
(- : sin 2e1 + e 1 /2) ]:" = r 3 [ ( C; - 1 r 3
: + 2 : +

+ 8 (~ - 1) (cos <Pn - 1) + 7- 2 - 2 sin 2,f,.] = r 3 [ 0·2922 x 2·356 + 1·416 -

- is X 0·292 {0·92388 - 1) + 2·356 - 2 - 2 X 0·7071] = + 0·737r3 •

a'0 1 = J m 1 M 0 ds = J 1 X M 0 ds,

, _,zSinZe (upper part)

Mo=-- 2 - -

,2
M~ =- + 2r2 (1 - cos e 1) (!ewer part)
2 •
roz , 2 sin2 e 1·~ [- ,a ]
Q~I =
0
f- 2
r de +.
0
2 + 2rZ(1 - COS ,1) 2r d{1 =

= -r2 [1-
2 4
sin 2e - - e]"'
28
2
+ 2r 3 [ _ e
_!. + 2e1 -
2
2 sin ' 1 ],;s = r ·(.n4
0
3 .:_ - 4 sin .n/8) =
= r 3 (0·7854 - 1·5307) = - 0·745r3 (p).
330 DESIGN Of TUNNEL SECTIONS

,r/ !- • ~in~ /' J~~ [ .:

a~ 2 =
J
m 2 M 0 ds=
o
J Lc,-r(l-cos~)Jr2 ~ r di; +
0
[c1 -r-2rsin~iJ ~-

=
- 2r2 (l - cose,)] 2rdi;, = :.r(c, sin 2 e - rsin' ·i; + rsin 2 {coseld<+ 2r 3
0 0
J (; -
h

- 2c, + 2c,cos <, - ; + 2r - 2r cos ~l - r sin <1 + 4r sin ~l - 4r sin <1 cos e,) de,
= r"[ (c, - r) { 1 sin 2i; + e) +
l- r -sin
3
-- i;]"'
2
+ 2r [(c
3
; - ) + 2c sin i; +
2c1 i; 1 1 1
2 4 2 3 0

+ (2,- ~- ) i;, - 2r sin i; , + r cos ~.. - 4r cos i; , - 2r sin' e,_i:· = r 3

[<c,- r) ; -

3
- 3 (c, - r)cJ>. + 4 (c, - r) sin cJ>. - 6r cos cJ>. - 4r sin 2 cJ>. + : r] = ,., [ - ( ~t - J) ; +

+ 4 (;' - 1)0·3827- 6 X 0·92388- 4 X 0·14644+ 6· 16] = r 1 [+ 0·229 - 0·446-

- 5·540 - 0·586 + 6·16] = -0·J8r 1

(p).

0·745r3
X, = +- - = + 0·3l7r'(p),
2·356r
O· J 80r'
X.,
·
= + -0·737r
- - = + 0·245r (p)
3

This completes the first part of the calculations. The structure has been solved for the con-
dition of fixed ends and no lateral ground support. The corresponding moment diagram is
illustrated in Fig. 4/40. The bending moments are
M = M,, + m 1X1 + m 2 X,,
for the upper part
r• sin 2 i;
M' = - + X, - c1 X 2 + r (1 - c_~ OX,,
2

FtG. 4/40. Moment diagram

for rigidly fixed ends
and no lateral ground support
 MONOLITHIC HORSESHOE·"SHAPED TUNNELS
331
for the lower part

M" = - ~: + 2r 2 (1 - cos <1) + X1 + (r - c1) X i + 2r sin e X: .

1

3. Turning now to the effects of the assumed lateral load of q =

I t/ m the expression for
the bending moment s on the equivale nt determi nate structur e is
(Figs. 4/41 a- b):
(a) For the upper part

«-~

M; = - J q' da rr sin (< - a) ..,. _ ,:

« = 1'14
(

J (1 - 2 cos~ a) sin (e - a) da -,~J

(

[sin ecos y
-x ~ n/4
'J. ~ 11/ 4

e e
- cos sin a - 2 cos 2 cc (sin e cos cc - cos sin cc)] d a = - , 1 !·sin
e sin a + cos ecos J

sin cc cos 2 a
-2sin { ( - - - - + -sin
3
2
3
cos3 a
a ) - 2 cos e --
3
]"=(
«• 111'
I
= - r 2 sin 2 {+ cosz e-
'
cosz ~
.
- 2 Slllz < -- -
3 J4 . • cos' {
Stn· { - 2 - - - 0·7071 Sin { - 0·707 J COS { + 2 Sin ~
I0·7071
---,
x 0·5
+
-
3 1

+
2
3 0·101 1J + 2 cos e
0·70713
3
l .

q'=I - 2 cos 2 5

lt/m

F 10. 4/41. Loads due to the lateral support

acting on the equivale nt
determi nate structur e
332 DESIGN OF TUNNEL SECTIONS

and in a rearranged form

M 0, = - •[
r- - 32 cos
• _.
<,, -
2 . • _.
"3 • •
sin-<,, cos- <; -
4 . •
3 sm· e + O·4714 sin< - 0·4714 cos e + 1] .

The moment at < = 90° i~

M~ = - r
2
[- ; + 0·4714 + 1] = -0·138 r 2

The horizontal component is

90° ;r,/ 2

H = J
ot =4S 0
q' ds sin -x = J
a.-=n/C
(1 - 2 cos 2 a) r sin -x dtz = r [- cos -x +
3
3 90
2 cos a ] "
,s•
=

= r [ 0·7071 - 2 X 0·707!3 ~ = 0·4714 r

3
The vertical component is
eo•

V = J
1X = 4S
0
.q' ds cos -x = f (I - 2 cos2 -x) r cos -x d-x = r r (cos a - 2 cos3 ; ) da =
9
2 4 ] o• 4 2
= r 2
[ sin-x--sin-xcos -x- - sina =r [ 1 - - - 0·7071+ - 0·7071x0·5+
3 3 Gtd5° 3 3

+ J4 0·7071 ·1 = O· 138 r (q).

.

(b) On the .lower part of the structure the bending moments are made up of two parts
q"= (I - 6·83 sin~ a) · (q) M " = M 0; + M 0;. While the first member represe·n ts the moment
due to load q", the second member denotes the moments due to H , V, and M 0 (Fig. 4/ 41 b) .

M ~, = .r••q" 2r dtz 2r sin (e 1 - -x),

"
J .f
(X :;;-:1

M~, =- 4r 2 (1 - 6·828 sin2 Q'.) sin (< 1 - -x) d x = -4r 2 sin<, cos rx - cose, [sin a -
«=0

- 6·83 sin 2 a (sin <1 cos a - cos <1 sin a)] drx = -4r2 [ sin <1 sin -x + cos <1 cos 'X -
·s in3 a
- 6·83 sin 1 - - e + 6·83 COS { 1( - cos tz sin 2 a
----- - -
2
COSl.l
)]"={,= - 4r-·1·sm-
.. <1 +
3 3 3 <t = O _

2
sin•e 1 ( cose1sin <1
+ cos e, -2
6·83 - - 3 - + 6·83 cos <1 - 3 - 32 cos e,) - •
cos<;, +

+ 6·83 cos e1 : = - l 4r2 ( - 2·276 sin• <1 - 2·276 sin2 <1 cos 2 <1 - 4·552 cost <1 +
+ 3·552 cos <1 + 1).
 MONOLITHIC HORSESHOE-SHAPED TUNNELS 333

r 11e moments due to H , V and M 0 are

= M~ - H2r sw. ft.;,.

M ;2 V2r (1 - cos ~1),

M (;2 = r 2 (0·1381 + 0-942t sin~, - 0·2762 + 0·2762 cos ~1)

ind the total moment is

M; = - 0
4r2 (q). (--2·276 sin4 ~, -2·276 sin 2 ~, cos2 ~, -

- 4• 552 cos 2
~ , + 3·622 cos ~ + 0·2357 sin~, + 0·9655)
1

The moment c\iagram on the equivalent determinate structu re due to assumed lateral
= 1 t/m is illustrated in Fi~ 4/4?.
"'rl. of a

lt/lTJ,
Fm. 4/42. Moment diagram on the
equivalent determinate
structure due to the
lateral load

The relative rotation of the elastic centre due to the lateral load is
900 900

!101 = {.Mods = - J[2

r 3 cos',; - 2
-
2
i" sin 2 ecos• e- 3
4
sin' ,; + 0·4714 sin ,; -
u• u•

- 0·4714 cos e+ l) r de - J••

0
4r2 ( - 2·276 sin' el - 2·276 sin 2 e1 cos• el - 4·552 cos• e1 +

+ 3·622 cos el + 0·2357 sin el + 0·9655) 2r de1 = - r3 - 32 e

( sin cos3
4
e+
[

+ -43 J 2
COS ,;
)
d,; - -2 -
3
sin-,; --
4 34
3
cos ,; - -2 -l J.
sm2 ,; de - -4 ( - -I sin 2e+ -,; )
3 4 2
-
cos e1 sin3 el
90
- 0·4714 cos ,; - 0·4714 sin e+ e -
]
•
8r3
[
+ 2·276 -
45• 4
 334 DESIGN OF TUNNEL SECTIONS

1· sin , cos3 ~ 1 . { sin 3 { cos ~ 1 . , 1 .

=- r
3
- - - - - - sm 2~ - - - - - - - + - sm 2{ - - + - sm 2{ -
6 8 4 6 24 12 3
90
- -2 ! - 0-4714 cos ( -0·4714 sin i; + ~] "
- 8r3 [
-
2·276
- cos ~1 sin 3 {
1
+
3 w 4

2·276 X 3 3 X 2·276 2·276 2·276 2·276

+ !
6
sin 2i; 1 -
8
{, - - -
4
sin 3 , , cos { 1 +- - - si n 2{1
16
- - - <, -
8

4 ·552
- - - sin 2{ 1 - -
4 ·554
- ~, + 3·621 sin { 1 - 0·2357 cos,,+ O·O965H 1 ,
]"'8
4 2 0
turning to numerical values

a 01 = - 0·02859 r 3 - 0·3183 r 3 = -0·3469 r 3

The relative horizontal displacement of the elastic centre is

90°

.= r3 [(- _:_cos•.;
3
- 2-sin
3
3
.; cos•.; - ~sin'.;+ 0-4714 sin.; - 0·4714 cos
3
e+ I) (c, -
,50 '
...
- r + r cos.;) de+ 8r 3
J<-
0
2·276 sin' el - 2·276 sin' e. cos' e, - 4·552 cos• el+

+ 3·621 cos e, + 0 ·2357 sine,+ 0 ·9655) (c, - r - 2r sin e1) d.;, = 0·02859 (c, - r) r 3 +

.+ ,•j"(- 2_ 3
cos3 .; - 2_ si~ 2 .; cos3 ~ -
3 .
~sin•.; cos
3
e+ 0·4714 sin .; cos ~ - O·47l4cos•.; +

+ cos e) d .;+ 0·3183 (c, - r) r 3 - 16r•J(- 2·276 sin 3 t1 - 2·276 sin 3 .;, cos•~. --

- 4·552 COS
2
e, sine,+ 3·6214 cos e, sin + 0 ·2357 sin' e, + O·965.'i sin.;,) d.;,=
t1

2 sin .; cos• .; 8 ( sin .; cos• e 2 \

=- 0·02859 x O·292r 4 + r1[- - ---- - - • + - sin .; I -
3 5 15 . 3 3 ,
..
2 .
15
, 4 sin 3 .;
15 3
4 sin3 .;
- - sm 3 .;cos'.;- - - - - - - - + 0·2357 sin 2 .;
3 3
-
( 1
0·4714 - sin
4
U + e12J +

- 2. cos e,)· - 2·276 (- sin

2
e, COS
3
e, _ .2_ cos
3
e, ') + 4·552 cos el
3
3·6214 cos• E1
3 5 5 3 3 2

+ 0·2357 ( - - I sin 2
4
e, + -I el ) - 0·9655 cos e1 ]"" ,
2 <, ~ 0
 - ------

M O1'OLITHI C HORSESHOE-SHAPED TU1'NELS 33 5

turning again to numerical values

0 0: =- 0·8347r 1 + 0·00437r 1 - 0·09295r 1 - O· J 606r' =- 0·2588r 4(q);

0 01 0·3469r 2
x, = - -· = - - - =
0 11 2·356
+ 0· 1473,- 2 (q),

0·2588r'
0·7 366,-3 = + 0·352r (q).

The bending moments due to the assumed lateral load are illustrated in Fig. 4/43 . ALI
values are to be multiplied by ,•.

FIG. 4/43. Moments due

to the Ia teral
load

4. We now have to investigate the horizontal displacements at the points where the section
is widest, due to an assumed vertical load of p = 1 t/m and due to the assumed lateral foad
of q = 1 t/m.
(a) The horizontal displacement due to load p is

where 00 , o1 and 02 are the displacements of the equivalent determinate structure due to the
vertical load p and the redundants X1 and X 2 , respectively.
Once again it is best to compute the displacements with the equations of virtual work.
Assuming a fictitious horizontal force of 1 ton the moment diagram is determined and used
in the above equation (Fig. 4/37).
M H = + 2r sin ~1 ,

01 =
r
•
M m ds
~
EJ
= -4r-
EJ
2
J ¢n

sin ~1 M 1 = - -4r- [ cos ~1 1•n

2

EJ
= -4r-
,0 EJ
2
(1 - cos if,.),
0 -

+ -,;21 )]•n =
3
- -c1 +
r
•
2 sm ,1) d,;1 = -4r
El
[
- cos ,;I + -cr1 cos ,;l + 2 ( - -1 SID
4
. 2,;,
0
 "'" ,, - - - - ---
- -_,_ ....

336 DESIGN OF TUNNEL SECTIONS

3
4r c1 c1 ]
= EJ
[
- COS <Pn + -;: COS <Pn - T1 Sill
.
2</Jn + <Pn + 1 - -;: =
4r [ /
= EJ
3
c,)
· I - -;: (1 - cos <fJ.) + <J>. -
1
2 sin 24>.] ,

2
4
4r [ -1 cos {
= -EJ 1 - 2 cos { 1 - s·in 2 { 1] "n = -4r
o EJ
4
t-23 - -32 cos '+'n
"' - sin2 "' )
'+'n '

substituting the above results into

b, = -~· [1'5 - 1·5 X 0·9238 - 0·14644]= - 0·1290 - r (p),

' EJ EJ
4r1 4r1 r4
61 = - (1-cos<J,.)X, = - (l-0·92388)0·317 =0·0965 - (p),
EJ EJ EJ

n,- - -4r I [-(

l - -Cf J (l - cos <J>.) + <Pn - -1 Stn
, 2</>n ] 0·145r =
EJ r 2

~ ~
= - [0·292 (1 - 0·9238) + 0·3927 - 0·3535] 0·245 = 0·060 - (p),
EJ EJ

~ ~
op= ---
EJ
(- 0· 1290 + 0·0965 + 0·060). (p) = + 0·0275 -
EJ
(p).

(b) The horizontal displacement due to the lateral load q is

where 00 , 0 1 and 52 are the displacements of the equivalent determinate structure due to the
lateral load q and to the redundants X 1 and X 2 , respectively.

00 =
JM H M0
~
4r' r .
ds = EJ. [- 2·276 sm' e1 - 2· 76 sm 2 e 1 cos- e1 - 4·552 cos 2 { 1 +
t . .

+ 3·621 cos ~1 + 0·2357 sin e 1 + .0·9655] 2r sin e1 2r de 1 = 16r

EJ • [- 2·276sin5
4
r ~1 - !

- 2·276 sin 3 <1cos2 <1 - 4·552 sin e. COS 2 el+ 3·6214 sin el cos e. + 0·2357 sin2 e1 +
. ~ • 16r
+ 0·9655 sm ~.J d, 1 = EJ
4
[
+ 2·2762
cos {.sin• e,
-
4x2·276 (
-
cos e.sin 2 e-, -
5 5 3

-
2
-COS < 1
)
-
( sin e . cos e, 2 cos3 e.)
2
cos 3 <1
3
cos2
2·2762 - - - -- - - - - - - + 4·552 - - - 3·621 - -
e, +
3 5 5 3 3 2

+ 0·2357 ( -
4
l - sinH1 +
2 ~. 1 -
I ) 0·9655 cos ~1 I"" = -
,o
0· J 620 _r ,
EJ
4
 -- - - - - -- - - - - - - ---
~

MONOLIT HI C HO RSESHOE T U N N ELS 337

,., ,-4
o, = EJ 0·3044 X 0· [473 = 0·0448 El (q) .
. ,.i y1
~..
·
- o ·352 X 0·2455 -
El
= 0·08642 -El- (q).
,. 1 ,-l
o,, = El ( - 0· J 620 + 0·0448 + 0·0864) (q) = - 0·0308 EJ (q).

5. In the next step we shall now equate the structura l deformation to the elastic compression
of the ground, where the sectio n is the widest.
q
C =Op + qOq ,
hence
,.,
0·0275 -
0·0275
q= ~ -= - - - EJ
J
C
- o q
I
C + 0·0308 E/
,-1 EJ
-Cr' + 0·0308
EJ 3 000 000 X 0·0 104
Cr' = - 50000 ~ = 0·0077 ,

0·02751
q = _
0 0077
+ 0 _0308 = 0·7 14 t/ m' (0·146 k/rn.
Thus, the magnitude o f the latera l pressure has been determined.
X1 = (+ 0·3 17 p + q 0· 1473 x 0·714) r', i.e. with q = p = I t/ m, X 2 = 0-422r~
X2 = (+ 0 ·245 p + q 0·352 X 0 ·714) r, i.e. with q = p = l t/ m X, = 0·496r.
T he bend ing moments due to the assumed lateral load a re illustrated in F ig. 4/ 43. All
values are to be mu ltiplied by ,-2 •

The moment diagram on the laterally elastically supported structure with fixed ends is
illustrated in Fig. 4/44. All ordinates are to be multiplied by r 2•

All ordi11,1tes .Jre to

be multiplied by
1'2

6'fl4t:m

Fm. 4/44. Combined motnent diagram due to vertical and lateral loads, for a structure with
fixed ends and elastic lateral ground suoport

22 Suchy: The Art of Tunnellin&

""
'. ~ .
-------

338 DESIGN OF TUNNEL SECTIONS

The calculation is next to be extended to consider the elastic constraint at the ends.
The rotation of the end section due to a moment of 1 t/ m (7·23 ftk) is
1 1
fl, = CJ
1
= 50 000 X 0·0835 = 0·00024
The moment of inertia at the fixed end is

J1 = 0·0835 m 1 (200 600 in').

1. Consider the relative movements on the structure with elastically constrained ends
and multiply them by EJ

a; 1 =au+ /J1 = 2·356r + 0·00024 X EJ

EJ = 3 000 000 X 0·0104 = 31 260 tm 2 (741 600 ft 2k);

a;, = 2 ·356r + 0·00024 X 31 260 = 7·068 + 7·50 = +14·57,

a; 1 = {1 1 Ye= 0·00024 (0·292 + 0 ·7654) 31 260r = +23·73,
a;2 = a 22 + {1 1 y; = 0·737r3 + 0·0024 X I ·05742 X 31 260r 2 = 19·90 + 75·2 = 95-10.

2. The relative movements a 01 and a~ 2, due to the uniformly distributed vertical load p are
a01 = - 0·745r3 - 0·00024 EJ X 0·3478r2 = -20·1 - 23·5 = -43·6;
a02 = - 0·180r 4
- 0.00024EJ X 0·3478r2 X l ·0574r = -14·6 - 74·5 = -89·1.

The redundants X 1 and X 2 -a re calculated from

-43·6 + 14·57X1 + 23·73X2 = 0,

-89·1 + 23·73X1 + 95·10X2 = O·
Hence
- 79· 1 + 23·73X1 + 175·0 - 58·5X1 = 0,
and hence X1 = +2·47 tm (17'87 ftk),
X2 = 1·84- 0·615x2'47 = 0·32 t (0·705 kips).

Consider now the structure with elastically constrained ends under the lateral load q = t.
t/m (0·672 k/ft)
ao 1 = -0·3469r3 - 0·00024 X 31 260r2 X 0·733 = - 9·36 - 49·5 = -58·86,
a 02 = -0·25:?8r' - 0·00024 X 31260 X l·057r x 0·733r2 = -21·0- 157·0 = -178·8,
58·86
X~ = . - 0·615X, = 2 ·49 - 0·615Xi,
23 73

-178·0 + 23'73X1 + 237-0 - 58·5X, = 0.

X 1 = + 1·70 tm (I 2'3 ftk) and x~ = + l ·44 t (3· l 7 kips).

 MONOLITHIC HORSESHOE-SHAPED TUNNELS 339

The moment diagram due to the uniformly distributed vertical load and for elastically con-
strained end conditions, without considering lateral support, is illustrated in Fig. 4/45. All
ordinates are to be multiplied by , 2 •
We proceed now !o compute the horizontal structural deformation at the points where
the section is widest.

FIG. <-!45. Moments due to vertical load

for elastically constrained
end conditions

Owing to the vertical load p = l t/ m (0·672 k/ft)

~ ~ ~
o, = -0·1290 Ei + 0·3045 EJ 2·47 + 0·245 EJ 0·32 + 0·362 x 0·00024 x 0·7653 ,

9
o, =<- 1·160+0·152 + 0·235)
31260
+ 0·0000666 = - 0·0000498 + 0·0000666=

= + 0·0000168.
Owing to the horizontal load of q = 1 t/m (0·672 k/ft)

fJ
,.
= - (- O·l6206r 2
= 0·3045 X 1·70 + 0·245 X 1·44r) - 0·34 X 0·00024 X 0·7653 =
• El

= 0·0000331 - 0·0000626 = -0·0000295, •

o. 0·0000168
q=-
,-- o-00002 + o-0000295 = o-34 t (0·75 kir~).
__:_ _L ()
C I • ,

With the value of q determined the redundants X 1 and X 2 are

X1 = (2·47 + 1·7 X 0·34) = 3·05 = 0·339r3

X2 = (0·32 + 1·44 X 0· 34) = 0·81 = 0·27r.
340 DESIGN OF TUNNEL SECTIONS

Fm. 4/46. Combined moments

for elastically constrained end
conditions and elastic lateral
support

M
- -- -- -- ·- ·- ·:-T·-
1

I I
I I
I .. I
I .;;I
~
I <!:.J
.. I
~,
' I
~I
I

0 O{}f 0-f
Fm. 4/47. Moments at the
crown as a function of structural rigidity and ground elasticity

The moment diagram considering elastic end constraints as well as elastic lateral support
is illustrated in Fig. 4/46. All ordinates are to be multiplied by r 2•
For variable structural rigidity and ground elasticity the bending moments are a function
of the ratio El4 . For a structure with fixed ends the moments at the crown are illustrated"
Cr
in Fig. 4/47 as a function of the ratio EJ4 . From the graph it would appear that the moment
Cr
for a laterally fully supported structure is 0·053 r 2• In the case of no lateral ground support
the same moment is 0·144 r 2 i.e. 2·7 times of what it was before. The coefficient of subgrade
reaction and the structural rigidity have only a lesser effect on the moments.
Generally speaking it can be concluded that if EJ4 is less than 0· 1, the elastic lateral
Cr
ground support might as well be neglected and the monolithic arch can be treated as if not
supported at all (laterally). For a summary of the above results see Table 4/11.
 CIRCULAR SECTIONS 341

TABLE 4/11

Max. moments
Condition
Crown Quarter point Sprinain1

Fixed ends, no lateral support +0·144r 2 -0·13tr•

I -F()·228r•

Elastically constrained ends, no

lateral support +0·200r 2 -0·194r'
I +0·040r2

Fixed ends (1806 lb/in3) +o·07lr 2 -0·06lr2 +0·078r2

V'ith lateral support (C = 50 kg/cm3)
I
Fixed ends (181 Jb/in3)
with lateral support (C = S kg/cm3)
+0·l18r 2 -0·107r2
I +0·174r 2

Elastically constrained ends +o·I48r 2 -0·129r2 +0·027r2

with lateral support

4.3. DESIGN OF CIRCULAR TUNNEL SECTIONS

4.31. GRAPHICAL INVESTIGATION

A graphical method for the investigation of circular tunnel sections of arch-

type construction is given below (Fig. 4/48):
The dimensions of the circular section are to be.assumed on an empirical basis,
followed by the calculation of the self-weight and the horizontal and vertical
external loads. The section is divided into elementary strips and the components
acting on each strip are converted into a resultant. The problem is one of construct-
ing a funicular polygon with a pole at point 0. The location of O is not known
as yet but the total polar distance is known to be equal to H i.e. to the sum of

"

t,
11· Se
6;;-f/ {tt f)
FIG. 4/48. Graphical j , II,
investigation of a Moments Jbouf point 0,
circular tunnel H. • f.(.e,--[6,·.g,·;;nd II. • [E. -N.
U.- e - U; ..9; -11,.11 · O ' II t II I
section
 -- ~ -- - - - -

342 DESIGN OF TUNNEL SECTIONS

the lateral loads. The position of O shall now be varied in such a way as to arrive
at a thrust line contained within the inner third and at the same time to result in
a set of H 1 and H 2 that will satisfy the equations of equilibrium below. If this
proves to be impossible then the wall thicknesses have to be modified accord-
ingly.
Thus. the problem is how to locate pole O since H 1 and H 2 are unknown. They
can be calculated, however, from the equations of equilibrium. Taking the moments
of the external loads about point 0 1 for example, (i.e. about a point in line with
one of the unknowns, H 2 ) the other unknown can readily be obtained .
Moments about point 0 1

In Fig. 4/48 the formula for ·maximum and minimum stress in the wall due to
an eccentric external load is repeated for convenience.
Because of the lengthy, awkward and inaccurate procedure involved, the above
graphical investigation is rarely used today.

4.32. APPROXIMATE METHOD FOR THE CALCULATION

OF RING SECTIONS

43.21. Design by Dividing the Section into Segments

Should the tunnel be constructed in the classical manner, i.e. the roof, wall
and bottom sections are built separately, then their analysis can be performed in
a similar way to that outlined under Section 4.21.
The roof section (i), the walls (ii), and the base (iii), are designed separately
assuming fixity at the connections and considering the reactions as external forces
on the supporting member. As a result of the construction, the shape of such
sections is far from being an ideal ring. As the walls have to transfer the roof
loads to the subgrade until the bottom section is formed , their bases have to be.
built wider in order not to exceed the bearing capacity of the soil. The connecting
base section itself is seldom a perfect ring segment. Thus, in fact, only the inner
surface has a circular shape (see Fig. 4/49). The finished circular section is then
analysed as a monolithic ring. As most of the external forces will develop and
act on the unfinished or rather the not yet continuous section, it has to be estab-
lished what portion of these forces will also be acting on the finished structure .
This depends mainly on the rate of the consolidation of the surrounding soil
and the rate at which the rock pressures develop. In loose granular soils or soft
saturated silty clays, in practice there will be few, if any, forces acting on the final
 CIRCULAR SECTIONS 343

section. However, in hard clays or solid rocks, experience indicates .that both the
vertical and horizontal forces are steadily increasing even several months after
construction. Therefore, depending on the speed of construction, 20%-40% of
the final pressure can be considered to be transferred to the completed ring section.
The analysis of the monolithic ring will be discussed later (Section 43.22).
If the soil surrounding the sec-
tion is soft and saturated, in loads on th2 outside wall
which case the pressures are near-
ly hydrostatic, and there is no
significant difference between the
pressure on the top and the bot-
tom, then the pressure distribu-
. tion is assumed to be uniform
and acting normal to the section.
This, of course, is the most fa-
vourable loading condition as
only normal forces are produced
in the ring section. The greater
the difference between the verti-
cal and lateral forces, the less
favourable will be the stress dis-
tribution in the ring. This is
usually the case in hard or dense Fro. 4/49. Loads on a circular tunnel section
constructed in stages
soils. However, even under these
conditions, if the walls of the
ring are built against the soil without leaving voids between, then the lateral
earth pressure will be replaced by the lateral earth reaction. On the other
hand, if voids are left between the structure and the soil, the moments at the
springing and the crown will be considerably increased (Section 63.225).
It is often necessary to provide an independent water-sealing ring on the inside
for the completed tunnel section. This is designed to take the full hydrostatic
pressure.
The pressure distribution on this internal ring is illustrated in Fig. 4/49, and is
shown to gradually increase in intensity with depth. The loading -diagram acting
on the outer section as a result of the vertical and lateral earth pressures, is also
given on the left-hand side of Fig. 4/49.

43.22. Design of a Monolithic Ring Section

The stresses in a tunnel section acting as a closed ring, i.e. an indeterminate

structure with three redundant reactions, can be determined exactly. In the fol-
lowing section, however, certain approximations will be made, namely, the magni-
tude of the supporting earth reactions will be arbitrarily selected and assumed
 344 DESIGN OF TUNNEL SECTIONS

to be independent of the deflection of the ring. In additio n, the

ring section will be
assum ed to remain consta nt.
The first step is to bring the structu re to a statically de~rm
inate condit ion by
cutting it throug h at the crown . The ring can be consid
ered now a curved
cantilever beam, - fixed at the left and free at the right side
of the cut - which
uµon the effect of the released interna l stresses would tend
to move at its free
end. To preven t the movement, three unkno wn forces Xi, X
and X3 are applied
at the elastic center u which is considered to be connected to 2
the free end by rigid
brackets. Upon the applic ation of these forces, neithe r rotatio
n nor horizo ntal
or vertical movem ent at the free end can occur (Fig. 4/50a).

8)

CJ

F10. 4/50. Determinate system for a

symmetrically loaded
ring beam

To determ ine the unkno wn forces X , X and X , first a unit

1 2 3 mome nt system
X1 = ± 1 tm, then a unit horizo ntal force system X = ± 1 t
2 and finally a unit
vertical force system X3 = ± 1 t is applied at the elastic centre
.
As a furthe r step, the rotatio n as well as the horizo ntal and
vertica l displace-
ments a 11, a 22 and a 33 can be found successively upon the
effect of these unit
force systems. Furthe r the rotatio nal a i, horizo ntal a , and
0 02 vertical a0 a displace-
ments caused by the external forces at the free end are
determined. Finally ,
to express the fact that the forces Xi, X and X bring the
2 3 elastic centre back to
its original position, the following expressions can be written
:

01 - 0
•
au X1 +Got= 0 from where Xi= ---,
au

- 0
022 X2 + 002 = 0 from where 02
Xz= --- ,
0 22

-a03
a33 X3 + 003 = 0 from where X3= - -. (4.56) .
033
 CIRC ULAR SECTIONS 345

And since the loading is symmetrical

a0 3 = 0 and hence X3 = 0
The displacement factors can be calculated by use of the virtual work equation:

a02 = J M O m 2 ds .
0

a ll = J
0
m~, ds . and a22 = J.m~ ds.
0
(4.57)

Where M0 designates the moments created in the statically determinate cantilever

beam by the external forces and similarly m 1 , m 2 and m3 are the moments caused
there by the imaginary unit moment and the unit horizontal and vertical forces,
respectively. The moment diagrams caused by these unit forces are illustrated
in Figs 4/50b-d.
The moment and axial thrust at any sec- /I
ltllltllilliltlllflfjfjjitliiillii
tion then follow from statics :
Moment : M = M 0 + X 1 + X 2y .
Thrust: N = N 0 + X 2 cos a:.
When considering the external forces act- )(

ing on the section not only the earth and

water pressures should -be taken into account,
but consideration must be given to the mag-
nitude and distribution of the subgrade re- ihi!ilililil!iillfiliill/WID
actions. The simplest case is when we as- p
sume that the reaction is the mirror image FIG. 4/51. The analysis of the
of the external loading, which in many cases, determinate system of
of course, is not true. In hard soils the re- a symmetrically loaded
ring girder with
action is more concentrated and in loose or uniform wall thickness
soft soils is more evenly distributed. But in
any case, this method enables us to determine
the moments on the statically determinate
beam for any reaction distribution and thus to obtain the displacement factors
a0 1 , a 02 and a 03 for the given case.
For a closed ring with a uniform cross-section the unit displacement factors
are given by the following equations:
The displacements caused by the unit forces acting at point <1 if y = r cos a: and
x = r sin a:, (Fig. 4/51)
...-= --- -- - ~ -

346 DESIGN OF TUNNEL SECTIONS

2n

a 11 = f
0
I ds = 2rrr ;

2.1t 21t

a 22 = J0
2
y ds = j', cos ar da =
0
2 2

f
= r 3 2n cos2 a da = r 3 [ ~ sin 2a + ; j~" = r 3 rr. (4.58)
0 a=O

1. The stresses in a closed ring, with a uniform thickness caused by the sym-
metrical and uniformly distributed load as illustrated in Fig. 4/51 can be deter-
mined as follows :
First the moment created by the uniform pressure p on the determinate struc-
ture will be calculated

px2 pr 2 sin2 a
Mo = - - - = -----
2 2

JI
ltifif li ltllit llllii fi lilfl l

FIG. 4/52. Moment diagram for the determinate

system form a uniformly distributed
n,, ,, 1,111•11111,1,1, 1•111, 1,n vertical load
p

..
The moment diagram is shown in Fig. 4/52. The displacement factors:

a01 =2 J
0
M 0 ds = - f
O
3
pr sin a da =
2

= - pr 3 . -I
t
•
4
sm 2a + -
r
.
2 _0
a l" 7t

2
pr3. (4. 59)
 CIRCULAR SECTIONS 347

a 02 = 0 and a 03 = 0 as the positive and negative areas of the m 2 and m 3 moment

diagrams are equal. Consequentl y x 2 = 0

1tpr3 p,2
= + 2 · 2m = +4 ·
The moment at any section
p, 2
M = MO + X1 = (1 - 2 sin2 oc) ,
4
•)
pr (4.60a)
o r expressing it with 2oc M = cos 2<X (Forchheime r equation)
4

The axial thrust of any section is

2 (4.60b)
N = N0 = --px sin oc = -pr sin a..

The final moment and thrust diagram for the continuous statically indeterminat e
ring is illustrated in Fig. 4/53. The same figure gives typical values for moments
a nd thrusts which could be used in approximate calculations.

Thrust Moment dtagram

F1G. 4/53. Moment and thrust

diagram of a continuous ring beam

2. To complete the picture, the stresses caused by the dead weight of the ring
should also be investigated. If the unit load on the developed section of the ring
is g and if we assume that the soil reactions are uniformly distributed along the
full length of the projection of the horizontal diameter, then according to
4
MARQUARDT,4- in the top section, where O < oc < n/2

u MARQUARDT, E.: Handbuch fur Eisenbetonbau. IX 429 and 447

 -- ----------

348 DESIGN OF TUNNEL SECTIONS

Men= _qr
2
[ ! n - ct sinct - ~ cos ct] ,

N 0 ,,_ . - -cosci]
= gr [ ct smct - , (4.61a)
6

and in the bottom half section where n/2 < ct < n

M 0, ,,_=gr 2 [cn - ct)smlX

. - 7t
2
• 2
sm IX - s
6
cosa - 7t
8
l
_,

, l.
N 0 ,,_ =gr n sm2 IX - (n - ct) sm1X
• - -COSIX
-
6
] . (4.61b)

It should be noted here that experience in subway construction in Budapest and

other cities indicates that the stresses are the most severe in the top section,
as the sides and bottom of the ring appear to get sufficient elastic support from
the surrounding soil, to counterbalance ring deformation.
3. The analysis of the ring section for the hydrostatic triangular loading case
illustrated in Fig. 4/54 is as follows:

!I
(e

Fm. 4/54. The moment diagram

of the determinate
structure from the
•f triangular loading

The moments on the determinate structure are:

- px(r -y)2
Mo=
6

Pe(r -y)
Px
2r
 CIRCUI.AR SECTIONS 349

Fig. 4/ 54 shows the moment diagram of the determinate structure for this loading
condition.
The relative movements 'of the elastic centre <1 (displacement factors) are as
follows:

f" M ds = - 2f" Pe r (1 12- cos a) 3 3 3

Pe r [ .
a01 = 2 0 doc= - - - a - 3sma +
6
0 0 ·

+ 3 -1 sm
(4
. 2ex + - - -sinacos a)
- - - ex - -2 sm
. ex
2
]"'=" =
2 3 3 «=O

5
= - u1t1'e r3;

a02 =2 f,. M yds=+2J,. Pe. ,2 (1 12- c;os a)s -rcosexrdex= -Pe6-J

o
0
O
r4 ,.
(cosex-
0

- 3 cos2 a + 3 cos3 a - cos4 ex) da = + Per• [sin a - 3 / .2_ sin 2ex + ~) +

6 \ 4· 2

+ 3 (-sinexcos a 2 . )
2
- - - + --s1na -
sina-cos -a
.
3.
3
- -- sm2a- -
3a]"'=" =
3 3 . 4 16 g· «=O

5
-= - - 1tp
16 e
r• '

The Jll.OID.t;llt at any section M = M 0 + X 1 - X 2 r cos a is:

M = - Pe r2(l -cosa)3 + ~p ,2_ ~ n ,2cosa =

12 24 e 16 re

p ,2
= ~ [ - 4(1 - cosex) 3 + 10 - 15 cos ex] . (4.62a)
350 DESIGN OF TUN1'EL SECTIONS

The thrust N = N 0 - X 2 cos a is:

l p, r2{l - cosa)2
N =px(r -y) · - cos a. - X 2 cosa = - - - - - cos:x -
2 4r

- -5- Per cos a.= --cos

Per a [ 4(1 -· cos a)-., - 5] . (4.62b)
16 16

The triangular loading as well as the moments and thrusts induced by this loading
are illustrated in Fig. 4/ 55.

T/Jrust MomMt

FIG. 4/55. Moment anJ

thrust diagram of the continuous ring beam

4. The stresses due to a trapezoidal earth pressure distribution can be obtained

by the correct summation of cases I and 3 (-1 and + 3). For the sake of sim-
plicity the end results of this summation are given below:
The moment at any point between O and n is

M=-y2r 2 [ t ( -l - sin
a a 4
--a+
2
) sin-
rcosa. ( -
6
a --
2
1
24
2
)j' (4.63a)

and the thrust is

2

Na = - YA.a r [ t cos-? a + r cos a (sin

- --a - 3 )] , (4.63b)
2 8
where t the distance measured from the ground surface to the centre of
the ring
20
= the coefficient of active earth pre_ssure
y = the unit weight of the soil.

Accordingly, the earth pressure at the crown level will be :

 c;IJ!.CULAR S.l;CTIONS 351

and at the invert level:

5. The stresses in the ring due to water pressure can be calculated by considering
first the stresses caused by the uniform water pressure acting at the top level of
the section, and then considering the crescent water pressure diagram below the
crown level.
The water pressure acting at the crown level will be uniformly distributed and
radical to the section and from i.t only normal axial stresses will result. If the height
of water·column above the crown level is h, and y,. is the unit weight of the water,
the axial thrust in the ring will be N = yjzr.
As a result of the crescent pressure distribution of the water pressure below
the top of the ring, as indicated on Fig. 4/56, bending mom~ts will be produced

Fm. 4/56. Triangular load

balancing the ·crescent water pressure

in the section in addition to the normal forces. Considering this loading case alone
the section appears to be .tabile, as it tends to rise as a result of the uplift forces .
This, in fact, is resisted by the dead W!;!ight of the ring and the soil reaction acting
at the top plane, and is indicated by the triangular pressure diagram on Fig. 4/56.
However, as the ratio and distribution of these balancing forces can be quite
variable the analysis is carried out first considering the water pressure acting
alone with an imaginary downward pointing reaction at the invert, and then con-
sidering the triangular shaped pressure di~gram of the balancing earth :i:eaction
replacing the uplift forces with a single ·point load at the invert level. The super-
position of these two, or rather three, if the dead load condition is included,
loading cases will give the correct answer. ·In the summation, the vertical reactions
assumed to be acting ·at the invert, will cancel each other since they are equal
but opposite in direction.
First the crescent loading diagram is considered, assuming a vertical concentrated reaction
acting in the vertical axis.
 352 DESIGN OF TUNNEL SECTIONS

The section is reduced to a determinate structure by cutting it through at the crown.

The moments and moment diagrz.m of the determinate structure can be calculated as shown
in Fig. 4/57a.
The elementary load on a ds element is

p• ds = y,r (1 - cost/>) rdt/> = y,r! (1 - cost/>) dt/>.

The elementary moment of this unit load about point A is

dM0 = - y,r2 (1 - cos t/>) d t/> r sin (a - t/>) = - y,r3 (sin a cos t/> - cos 7 sin ti> -
- sin a cos2 t/> + cos,, sin t/> cos t/>) dt/>."

M, • - t,r' (f - coJcr-'f sincr)

FIG. 4/57. Crescent loading with the concentrate d reaction acting at the invert and the moment
diagram of the determinate structure

The moment of all forces acting from the left of A is

M0 =
.
-J-" y,r3 (sin a cost/>- cos a sin ti>- sinacos 2 t/> + cos a sin t/> cost/>) def>=
• =0

= -y,r3 [sin a sin t/> + cos a cost/> - sin a(_!_ sin 24> + j_) + cos a sin! cp]•-"=
4 2 2 • = 0 ..

=- y, r 3 [sin2 a+ cos2 a - _!._ sin2 a cos IX- ~ sin a+_!._ sin 2 a cos IX - cos IX] =
2 2 2

= - y, r 3 [ 1 - cos IX - ; sin IX].

The moment diagram of the determinate structure is illustrated in Fig. 4/ 57a. Having
established the moment diagram, the relative displacemen t of the centre of the annular beam
can be determined. For a ring of constant cross-section the elastic centre <1 coincides with the
geometrical centre of the ring
 CIRCULAR SJ:CTIONS 353

.. ff

a 0t = 2.f M 0 ds = -2.ry, r 3 ( 1 - cos IX - ; sin IX) rd1X = - 2 y, r 3 [ IX - sin IX -

0 0

4
- ~ (-occost.t+si n1X)]: =-nr y,,

a 02 = 2.f M 0 m 2 ds
O
= JM
2
O
0 rcosoc rd tx = -2y,r5 J(1 - cos IX -
0
; sin IXJcos IX doc =

= -2y,r5 [ sin cc - -} sin 21X - -IX - -1 (-21X cos 21X + sin 21X)
4 2 16 . <X = O
= - -J :.y, r 5
4
]"="
As a next step the imaginary forces X 1 and X 2 are determined.

Dot nr' Y, ,a
X1 = - - = - - -
a 11 2r:n:
+ -y,.
2

Do2 3:n:r5 Y, 3 •
X: = - -
a 22
= -4r-3-:n: = + -4 r· Y,.

And finally the moment at any section of the statically indeterminate d annular beam is

, r3 .
= - y, 4 (2 - cos IX - 2rx SID 1X) .

The variation .of the moment on the annular beam is illustrated in Fig. 4/58.
Returning to Fig. 4/56, we now assume that the distribution of the soil reaction opposing
the uplift pressure acting on the section, is triangular. To satisfy the requirements of statics,

..

FIG. 4/S8. Moment

diagram of the ring girder

23 Szccby : Tbe Art of Tunnellina

 - - - - -- - - - - -

354 DESIGN OF TUNNEL SECTIONS

the equilibrium of the section is temporarily restored by an imaginary reaction-force at the

bottom (Fig. 4/59).
First, the moment diagram for the determinate girder is computed.

!I

a;

Fm. 4/59. Triangular loading with the balancing

concentrated force (R) and the moment
diagram of the determinate structure

As the loading is not continuous, the moments will be expressed separately, for the top
section-for values of 0° < a< 90° - for the bottom part where 90° < a < 180°.

Px xt (p - Px)x% xt (2 )
Mo1 = - - 2- - 3 = - 6 p + Px ,

(r - x)
x = r sin a; Px = p - - - = p (1 - sin a) .
r
Substituting these values:

r 2 sin: a prt sin! a

M0 1 = -
6
[ 2p +p (1 - sin a)] - -- - (3 - sin a).
..
6

In the bottom section if

.r./2 < iz <-n
2
=- pr ( r sm
. a- r ) pr 1)
3 =- 2 3 .
Mo: ( •
(sm a-
2

The variation of the moments on the determinate annular beam are given in Fig. 4/59b.
Next, the relative displacements arc determined.
 CIRCULAR SECTIONS 355
« /! ffl 2.

0 01 = L.J M 0 ds = 2J -
2
pr! s!n r.c (3 - sin r.c)rd1X + 2.f - p; 2
(sin r.c - !) rdr.c =
~ 0

=- pr3 [ 3 .
-3 - -4 Sill 2a
cos IX sin r.c
+ -31X2 + - -3- - + -32
2
cos IX
]•=• - 1
•,
pr 2
(
- cos IX - -
IX )•="=
• '=0 3 «,=,c/ %

L JM
.
w/2
pr2 sin2 IX
0 02 = 0 Y ds =- 2
J0
6
(3 - sin ex) r cos IX rd IX -

"
·-2f 2 pr% (sin a - _!_) r COS a rd a= }
3
pr' [sjn3 a -
3
sin' IX r = w/ !
4 J -n
_

w/ 2
2
_ pr' [ sin IX _ sin IX ·] ·=" = _ .!!!!__ + pr' = _ !!!!_.
2 3 • = w/2 4 6 12

The X 1 and X 2 forces will be

Xi = - ~ = pr3 (~ + ~) = + prZ (_!_ + .2_·) ;

0 11 6m 4 3 6 4 3:n

At any point the moment:

M1 =-
pr2 sin2 1X
- - - - ( 3 - sin a)+- -
p,.1.(1 ') +
+ -3:n pr•
-cos or=
6 6 4 12:n

2 2 . } 14 COS IX]
= - -pr [ 2 sm
. 2
a - - sm3 a - - - - - -- •
4 · 3 6 9:rr: 3:n _ ·

2 1 2
•
M 2 = - pr (sina- 1 ) pr ( 1 7) pr
2 3 + 6 ·4 + Jn - lln C081X=

= - .!!!:._
4
[2 sin a - ~ - _.!i - COS-IX] .
6 9:n 3n

·The final moment diagram is sliown in Fig; 4/60.

The stresses from the cr~scent-shaped water pressure are determined by the
8 uperposition of these two lpading conditions if p is replaced by p = 1vnr.

23•
356 DESIGN OF TUNNEL SECTIONS

Fm. 4/60. Moment dia~ram

of the annular beam for a crescent load ing

For the combined loading for X 1 and X 2 we obtain :

,a ,.a }'v n (1 7 ) yv ,a (7t 81

X1=2 Yv +-6- 4+~ = +-3- 8+3'

The moment at any section if O < a < ~ ,

. 2
3 3
yv r . yv n r ( . 2 . 1 14
M 1 = ---(2-cosa-2crnna)---- 2sm- a --~m 3 rx - - - - - 0

4 4 3 6 9n

- -
3n
-J =-~211:sm
co set y v ,a (
4
. 2 2 11:sm
et- .-
3
. a et-2rxsma--cosa-
. 4
'3
-7t + -4
6 9.'
l
and if n/2 < et < n

M2 = - -'lv r-
4
3
( •
2 - cos rx - 2a sm et) - -Yv -r
4
3
( .
2n sm et - 5n 14 COSet
6 - 9 - - 3- =
l
3
4
yv r Sn
+ 4 .)
9 .
( • •
= - - - 2n smet - 2et sm et - coset - (4.64)
4 3 6
The moment diagram is shown in Fig. 4/60.
..
43.23. The Hewett- Johannesson Method
In the previous paragraphs we ·have discussed the design of an annular beam
which was considered to be independent of the surrounding soil mass and acted
upon by arbitrarily assumed loads. The following method which was developed
for tunnels built by the shield method takes the lateral support and interaction
of.the soil into consideration but is quite arbitrary when assuming its distribution.
 CIRCULAR SECTIONS 357

It assumes that its value and distribution will be such that it will adjust the centre
line of the section to coincide with the -thrust line, and thus no moments will be
created in the section. It is obvious that this method, too, can be regarded only
as an.approximation and its use is limited to hinged sections composed of segments
such as the liner plates.
The method regards the tunnel section
as a continuous rigid structure, analyzing
it for both the long-term and the short-
term (i.e. during construction) cases.
Fig. 4/61 illustrates the loading condi-
tions assumed for the long-term case.
The loadings shown in the diagram
are the full earth and water pressures of
which the values of the lateral earth pres-
sure alone are assumed to be unknown.
FIG. 4/61. The final
Upon the effect of the external loadings load acMng on the ring
the ring will undergo a deformation; it (HEWETT-JOHANNESSON)
becomes shorter along its vertical dia-
meter and expands along its horizontal
diameter. This elongation, however, is counteracted by the passive resistance of
the earth. The horizontal earth pressure, therefore, will have a value somewhere
between the active and passive cases. The coefficient of horizontal earth pressure
k will, thus, be

if Aa = tan2 (45° - </>/2).

The value of k is chosen to satisfy the condition that there should be no moments
in the section. Thus, for the long-term case, only normal stresses will be consid-
ered. Should the above condition result in the value of k becoming less than Aa
1
then it is taken to be equal to }.,a and similarly k = - - is used on the calculation
la
1
if its value is indicated·to be greater than--. Naturally in t~ese cases the condi-
}.,a
tions that M = 0 cannot be satisfied.
The validity of such iateral pressures 1s, of course, not substantiated theoreti-
cally and is only justified by the flexibility of the ring and as a result of experience.
However, in case of liner plates or cast-iron segments there is sufficient reason
to make these assumptions 1 as such sections tend to act as a truly hinged struc-
ture.
In soft and saturated loose soils, the horizontal pressure is considered to be
equal to the vertical pressure k = 1.
Some typical loading conditions and the resulting stresses will be discussed below using
the following nomenclature: (Fig. 4/ 62)
 358 DESIGN OF TUNNEL SECTIONS

Water/eve/

h,
Ground .surface

Fm. 4/62. The position

of the liner plate tunnel

a = the radius of the extrados (m)

a1 = the radius to the center line of the ring (m)
b = the inside radius (m)
y, =the unit weight of water (t/m3)
?' =the unit weight of soil (t/m 3). Below the ground water table the buoyant weight
should be considered
W = the weight of the structure per lineal foot
P = the uniformly distributed part of the vertical earth and water pressure calculated at
the top level of the tunnel (t/ m1 )
Q = the air pressure 'in the tunnel (above atmospheric) (t/mt)
h 1 = the depth of water above the ground surface (m)
h 2 = the thickness of earth cover above the centre line of the tunnel (m)
k = the coefficient of horizontal earth pressure
.i.a = the coefficient of active earth pr~s~ure.

1. Dead load condition:

The calculations consider the stresses creat-
ed by the weight of the top half only, as the
bottom section is supported directly by the
soil and therefore the moments due to the
weight of this part can be neglected.
The final moments in the indeterminate
structure will be (Fig. 4/63) •
M = n4 Wai-
The value of n4 in the upper section is
, 1
n, = -
:n2
+ -323 - -24:n
11
- cos a: -
er .
- _ sm a: ,
2 ,.
and in the bottom section
,, 1 1 cost er' 1 . ,
Fm. 4/63. Moments due to dead load n4 = - - - - - - - - -- sm cr.
:n! 32 8 2411
 CIRCULAR SECTIONS 359

The thrust at any section can be

computed from the following two
expressions (Fig. 4/64):
For the upper half

(}251'1
N' = - w (- 1 -COSCI + - ct smci
• )
'
2411: 2n

and for the lower portion

W ( sin ci'
w _
N --- - +cos:
-2411: -4-ci' )
.·
Fm. 4/64. Thrust due to dead load

2. The vertical earth pressure is aJ

taken to be equal to the full geostatic
pressure. The pressure is calculated Moments
as the sum of the uniformly distribut-
ed pressure at the top level of the
tunnel (4) and the weight of the two
earth-segments below the crown level
(marked 2 in the diagram). The reac- -0043
tion (5) at the bottom is assumed to
be uniformly distributed . The moment
due to the uniformly distributed earth
pressure is given by :

r ~ .,o, . r (t-.Jr/4).1
-1 - I - sm:ci
paa
M = - . )
11 111filtlH 01 jjj jjjjji)jjj\ jjij jjjjjj Ol!IIIIBD '
2 2 '

or, more simply,

M = n5 p

where
1 sin! ci
n5 =4 - - 2- .

Values for the thrust are also given

in Fig. 4/53.
The stresses due to the weight of
the earth between the crown and the
springing are as follows (Fig. 4/65a): FJG. 4/65. Moment and thrust diagrams
In the top section
3
I 8 ·3 cos ci sin: ci 1 . . ci . cos ci]
M' = ya 3 -
[4
- - - - n + - - - - - + - SID Cl SID 2ci + - SID Cl+ - -
9n 32 48 2 4 2 3 '
 360 DESIGN OF TUNNEL SECTION:,

in the bottom section

M" = ya 3 r-4I - --
8 + - :,: - ----
9:,: 32
sin
48 ,
I- a.' (
4 2
• -i
- :-:-)' -cos-7.'
:
'

M = n3 ya 1 a 2 •

where n 3 is the expression between the brackets. Fig. 4/65b indicates the diagram of the axial
thrusts. In the top half

N
, = yaj. ,cosa.
48 -
. • a.
sm- +
4
1 sm
. a. sm
. 2a + z]
2
sin a ;

and in the bottom half

N" = ,,a~ [- (1 - : ) cos 2

7.' -+s- z'].
sin

3. The lateral earth and water pressure diagram is assumed to have a trapezoidal shape.
It is further resolved into a uniformly distributed and triangular-shaped pressure diagram .
The moment diagram for the triangular loading case is given in Fig. 4/ 54. The magnitude
of the moment is:

M = ya, a' [ - 5 - - 5 cos a. - -1 (1 - cos 7.)3 ]

12 8 6 '

5 5 l
where n1 = -12 - -8 cos "'~ - - 6 (1 - cos a) 3 .

The axial thrust

3
N = cos
ya,• (- -- a. - cos' a. - - cos
--
'.1. )
.
2 8

The variation of the normal force s is also the same as given in Fig. 4/54.
The moments and thrusts due to the uniformly distributed part of the lateral pressure
are calculated in the same way as for the vertical pressure except that the values of the dia-
'?,rams will have to be rotated by 90°.
Considering all the loading conditions previously discussed, the final moments are given
by the equation:

-
(4.65)

In this equation the first expression on the righ-hand side represents the moment due to
the triangular lateral earth pressure, the second expression the uniformly distributed lateral
pressure. The third and fourth expressions are the triangularly and uniformly distributed
parts of the lateral water pressure. The fifth expression gives the moments set up by the hori-
zontal components of the internal air pressure; the sixth and seventh expressions yield the
moments produced by the weight of earth and water below the top level of the tunnel. The
eighth expression represents the effect of the dead weight and the ninth gives the moments
due to the vertical earth, water and air pressure components. Values of ni, n 2 , n 3 , n4 and 115
can be obtained from Table 4/111. In the calculations the value of k is assumed to lie between
 TABLE 4/ lll

-1 -
Moment at the crown
--
Moment at the springing
-
-
M oment at the bottom

!'_ I M, M, M,
" I

I
-~--
n, "• n, n, n, "• "• n, n,
I "• "• r-n, I n,
I n,
I "•
0·70 -0·07 -0·1 I 0·02 0·05 0·23 l 0·09 0·13 -0·03 -0·06 -0·27 -0·09 -O·ll 0·03 0·05 0·23

0·75 -0·08 -0· 13 0-02 0·05 0·23 0·I l 0·15 -0·G3 -0·06 -0·26 -0· 11 -0· 13 0·03 0·05 0·23
i
0·80 -0· 10 -0· 15 0·02 0·05 0·24 0·13 O·J7 -0·03 -0·06 -0·26 -0·14 -0·15 0·03 0·05 0·24
(")
0·85 -0·12 -0·17 0·02 0·05 0·24 0·15 0·19 -0·04 -0·06 -0·26 -0·17 -0·17 0·04 0·06 0·24 :ii
(")
C
I"'
0·90 -0·15 -0· 19 0·02 0·05 0·25 0·18 0·21 -0·04 -0·06 -0· 26 -0·21 -0· 19 0·04 0·06 0·25 >
:,0
en
0·95 -0·18 -0·22 0·02 0·05 0·25 0·21 0·23 -0·04 - 0·06 1 - 0·25 -0·25 -0·22 0·04 0·06 0·25 "'
~
l·00 -0·21 -0·25 0·03 0·05 0·25 0·25 0·25 -0·04 -0·06 -0·25 -0·29 -0·25 0·04 0·06 0·25 0
zen
1·05 -0·24 -0·28 0·03 0·05 0·25 0·30 0·27 -0·04 -0·05 -0·25 -0·34 -0·28 0·05 0·06 0·25

l·IO -0·27 -0·3 1 0·03 0·05 0·26 0·35 0·29 -0·04 -0·05 -0·24 -0-4 1 -0·3 1 0·05 0·06 0·26

1· I 5 -0·31 -0·34 0·03 0·05 0·26 0-40 0·31 -0·04 -0·05 -0·24 -0·45 -0·34 0·05 0·06 0·26

1·20 -0·35 -0·37 0·03 0·05 0·26 0-46 0·34 -0·04 -0·05 -0·24 -0·51 -0·37 0·06 0·06 0·26

1·25 -0·39 -0-41 0·04 0·05 0·26 0·52 0·37 -0·05 -0·05 -0· 24 -0·57 -0-41 0·06 0·06 0·26

1·30 -0·44 -0·45 0·04 0·05 0·27 0·58 0·39 -0·05 -0·05 -0·23 -0·64 -0·45 0·06 0·06 0·27

J-35 -0·49 -0·49 0·04 0·05 0·27 0·64 0·42 -0·05 -0·05 -0·23 -0·7 1 -0-49 0·07 0·06 0·27

]·40 -0·54 -0·53 0·04 0·05 0·27 0·71 0·45 -0·05 0·05 -0·23 1 -0·79 -0·531 0·07 0·07 0·27 w
°'

•
362 DESIGN OF TUNNEL SECTIONS

the active and passive cases, so that, as discussed before, there will be no moment in the
structure.
From this point of view soils are divided into three categories:
1. dry soils, where <f, ~ 30° where k lies between 1/3 and 3·0
2. dense permeable soils : 30° > <f, > 19°, 1/: < k < 2
3. soft, saturated soils: <f, ~ 0 and k = 1·0.
For these main soil type& the following. expressions can be developed.
1. For dry soils no water pressure need be considered. The expression for the moment
will be:

and if
k = _ -n: Qa + n3 ya 2 + n4 W + n5 (P - Q) a
n 1 ya 2 +n 2 Pa '
then
M = O.

If the value of k lies between 1/ 3 and 3 then for the long-term case no moments will hav'- to
be considered. If k is less than 1/ 3 or more than 3·0 it is taken to be equal to 1/ 3 or 3, re-
spectively.
The following expressions will give the thrust:
at the crown:

N1 = -by, (0·67b + h 1 + h:) - kby (0·67b + h 2) + Qb,

at the springing:

Ne = - Pa - 0·22 (y, + y) ab - 0·25W + Qa,

and at the invert:

N 3 = - by, (l ·33b + h 1 + h 2) - kby (1 ·33b + h 2) + Qb.

The exact values for the thrust for circular sections differ from those given by H EWETT
and JOHANNESSON; they are:

N1 = -ay, (0·604a + h 1 + h 2) - ka y (0·625a + h 2) + 0·02la 2 y + Q 0 + 0·0l33W,

N2 = -Pa - 0·215 (y, + y) a2 - 0·25W + Qa,

N3 = -ay, (1 ·396a + h 1 + h 2) - kay (1·375a + h 2) - 0·02la 2 y + Qa - 0·0l 33 W.

2. In dense permeable soils the moment is calculated by Eq. 4.65. If

n 1 y,a2 + n: [y, (h 1 + h 2) - Q] a +n 3 (y, + y) a2 + n4 W + n5 (P - Q) a

k=
n 1 ya 2 +n 2 y h2 a
then
M = O.

In this case the moments can be neglected for the long-term case if 1/2 < k < 2·0.
 CIRCULAR SECTIONS 363

3. In soft or loose saturated soils with k = t ·0 the moment is. given by Eq. 4.65. Using
this value for k

introducing the symbols

Y, + Y = Yo,
and assuming that

and
(y, + y) hz = P

then the moment can be written as:

M = a 1 [m 1 Yo a 2 +m 2 W +m 3 (P - Q) a).

The values for mi, m 2 and m 3 are given in Table 4/IV:

TABLE 4/IV

Crown moment Springing moment Bottom moment

b/a M, M, M,

m, I m, I m, m, I m, I m, m, I m, I m,

0·70 -0·05 0·05 0·12 0·06 -0·06 -0·14 -0·06 0·05 0·12
0·75 -0·06 0·05 0·ll 0·08 -0·06 -0·11 -0·08 0·05 0·l I
0·80 -0·08 0·05 0·09 0·10 -0·06 -0·09 -0·11 0·05 0·09
0·85 -0·10 0·05 0·07 0·12 -0·06 -0·07 -0·14 0·06 0·07
0·90 -0·12 0·05 0·05 0·14 -0·06 -0·05 -0·17 0·06 0·05
0·95 -0·]5 0·05 0·03 0·17 -0·06 -0·02 -0·21 0·06 0·03
l ·OO -0·18 0·05 ()·00 0·21 -0·06 0·00 -0·25 0·06 0·OO
I

For a circular section the above equation can be further simplified, since

•
thus

Hewen and JoHANN.ESSON express the moment in terms of the diameter

M = d(m;y 0 D 2 + m; W),
where dis the diameter measured to the neutral axis of the section and D is the outside dia-
meter.
, m:
mz =2
-~ - -- - - --- -

364 DESIGN OF TUNNEL SECTIONS

The thrust for this case: at the crown:

N1 = -ay0 (0·625a + h 2) + 0·0021 a y + Qa + 0·0133 W

2

and writing similar expressions for Nz and N 3 , N can be writt,m as:

N = n,' YoDZ + n'w - 1 ( P - Q) D .

2
2
The values of n; ana n; can be determined from J!ig. 4/66.

~~1 s::.
~- ~ ~
S:,
~- s::.
~-
-;:;,20 <:S ~ <:::, <:::,
... JO' I\ I I I

oo \
5{!

,,.
\
60' ,J ·-
70'
!JO' ~~ '
,w I \ Springmg
100-' I I

110° \ /
120' \ ~ll,.
!JO'
140'
'v'
/,
150' / I"'-.
160° /
~170' 'fl_ ~
~/80' r;:,
..
~
~ '?
' ~
'{,\
~
~ .
' Invert
Neffe!1ve sign ,nd1c1tes
compression
FIG. 4/66. Values of n; and n;

~
1.l5 ..
<;s
FIG. 4/67. Dead load <:>I
moments for two knife-edge supports

In the design the construction stage or short-term case should also be consid-
ered. The stresses created by the dead weight of the structure are then likely
to be the most serious since the ring is not yet receiving lateral support. The heavy
weight of the erecting equipment could also cause considerable stresses.
HEWETT and JOHANNESSON analyse the stresses due to the dead weight during
the construction by assuming that an isolated ring is supported only on two
 SECTIONS ~65
ELASTICALLY EMBEDDED CIRCU LAR

the vertical axis. This load ing stage

knife-edges each of whic h enclose 20° with
is illus trate d in Fig. 4/67.

M=-: a 1 (asi na+ cosc x-1) =0·0 102W a1 +0·0 595W

acos a.

(4.66 )

s can be obta ined from this equa tion.

For most pract ical cases the maxi mum value

NELS
4.33. THE DESI GN OF CIRC ULA R TUN
AS SECT IONS ELAS TICA LLY EMB EDD ED IN THE SUBS OIL

of Mos cow deve loped the theo ry of

The designers of the Subw ay in the City
r more accu rate meth ods take the
the desig n of circu lar secti ons furth er. Thei
oil into full acco unt.
elastic embe ddin g of the tunn el into the subs
rial is elastic and in it the react ions
lt is assum ed that the embe ddin g soil mate
the equa tion of WINKLER- SCHWEDLER
are prop ortio nal to pene tratio ns, that is
is valid, i.e.
p= Cy,

s on the perim eter face

wher e p = the press ure actin g at right angle
circu lar secti on
y = the radia l defo rmat ion of the
rt;ac tion.
C = the coefficient of the subg rade
of the subs oil the defo rmat ion is the
A furth er cond ition is that at any poin t
prov iped by the soil is take n as a
resul t of embe ddin g press ure. The react ion
orts.
series of inde pend ently actin g elastic supp
two basic assu mpti ons of elastically emb edde d tunn el design.
These are the
assu mpti on are, in general, smal ler
The inacc uraci es resul ting from the first
than those resul ting from the seco nd one.
ed in the next para grap hs. The first •
Thre e of the design meth ods will be treat
the poly gona l meth od, and the third is
is that of BODROV- GORELIK, the seco nd is
that deve loped by BouGAYEVA.

43.31. The Meth od of Bodrov-Gorelik

pote ntial energ y in the solut ion of
This meth od appli es the theo rem of least
elastically emb edde d circu lar sections.
n in the illust ratio n (Fig. 4/68) which
Let us analyse the circu lar secti ons show
neve r the secti on defo rms outward,
is acted upon by know n exter nal forces. Whe
~
--- -~~
-

366 DESIGN OF TUNNEL SECTIONS

FIG. 4/68. External

forces acting on an elastically embedded ring

!I

FIG. 4/69. Relation between

radial and tangential displacements
.x

elastic reactions are mobilized, which will work as infinitely small independent
supports whose magnitude is proportiona l to the deformation (compression).
The deformation of an arbitrary point of the tunnel will be resolved into two
components : the radial one will be denoted by u, the tangential one by t.
Neglecting the deformation caused by shearing and normal forces, the following
relationship exists between the components u and t:

t =- f ud<f, .

The correctness of this equation can be shown as follows (Fig. 4/69): The element
abed of the tunnel section takes a new position a'b'c'd' as a consequence of defor-
mation. Hence the tangential specific deformation is

t + ~ d¢ + (r + u) d¢ - r]- rd¢
e,,.= - - -- = I v 'I'
= -
dt + ud<f,
- -
ab rd¢ rd¢

l dt u
e,,.= - - +- .
r d<f, r
 ELASTICALLY EM!lEDDED CIRCULAR SECTIONS 367

Assuming that the specific elongation is e~ = 0 we obtain the following rela-

tionship between u and t:

t= - f ud</>. (4.67)

BODROV and GORELIK apply the theorem of the least potential energy. They
express the radial deformation of the section with the aid of a Fourier Series:
m
u = "f,(amcosm</> + bmsinm<J>). (4.68)
m-0

In the above equation </> is the angle enclosed by the radius drawn to the ana-
lyzed point and by the direction X, whereas ai, a 2 • •• am and b 1, b2 • •• bm are
unknown constants.
The tunnel sections and external loadings are symmetrical about the vertical
axis in most cases. Therefore in the equation for u, the second expression is zero,
hence the radial deformation is
m
u= Lm-1am · cosm</> , . (4.69)

and the tangential deformation:

m m 1
L amcosm</>d</> = - L
t
S
= -
m=t m = t
- amsinm</>.
m
(4.70)

The full potential energy (ll) of the tunnel section consists of three parts:
1. the work of external forces, T
2. the deformational work of internal forces, V 1
3. the work of elastic reactions, V 2
(4.71)
.
The work of external forces
a1 a1

T j~, u ds +fp, t ds ,
=
Di 01

where p, and p, = the radial and tangential components of the external pressures,
respectively.
 368 DESIGN OF TUNNEL SECTIONS

By substituting u and t we obtain

j'
•2 02

T
S
= Pr r
DI
L am· cos m</J
m• t
m
d</> + {'
a1
1 rL
m
-
m=l m
I
am sin m</J d</J. (4.72)

The sign of the work can be positive or negative depending on whether the signs
of the deformation and force are the same or not.
The deformation work done by the internal forces

(4.73)

where M is the moment of the external forces acting upon the tunnel and which
can be expressed by the deformation of the section

(4.74)

Substituting into Eq. 4.73 and as ds = rd¢

Replacing u by (4.69) and integrating

Ehc "'
V1= -2 3
r
L
m=l
(4.75)

The work of the elastic reactions

.;2
..
V2 =- I
2 J'Cu ds . 2
(4.76)
<Pl

On substituting u from Eq. 4.69

<P2
V2 CJ' L
= - _!_
2 m=l
m (aml COS m¢)2 d¢. (4.'/7)
<Pl
 ELASTICALLY EMBEDDED CIRCULAR SECTIONS 369

The limits of integration </> 1 and </> 2 are still unknown in the above expression,
because it is not known from wb.ere the tunnel section will move outward.
In the design the position of </> 1 and <j, 2 must be estimated in advance. Then,
when the results are known, whether the deformations are equal to O at points
<j, 1 and <j, 2 must be checked. If the discrepancy is significant, the calculations must
be repeated by substituting new correct<:d integration limits.
On this basis the potential energy of the 6ection can be expressed in such a form
that only the constants ai,.a2 ••• am are unknown. These can be determined by
applying the theorem of least potential energy because if the partial differential
quotient of the function of the potential energy is calculated and equated to 0
then m equations are obtained from which the constants ai, a 2 • •• a,,, can be
expressed
(4.78)

The calculations can be made with any desired accuracy. If more members are
taken into consideration, the results will be more accurate. Generally, the minimum
number of members to be considered is from 5 to 6. When a 1, a2 ••• am have
been determined, the deformations of the section can be computed at any point
by Eqs 4.69 and 4.70.
The numerical value of the reactions acting over a unit length of the tunnel is

m
q = - Cu = - C L am cos m<J, •
m=l
(4.79)

The bending moment at any cross-section

d2 m
substituting d 2u - ,i..~ =L m 2 am cosm<p and rearranging the equation:
d.,, m= l

El
.
M = - 2-
r
L (1 - 2
m )am cosm<p . (4.80)

The shearing force can be obtained from •

dM El d
Q =-
ds
=- - -
r 2 rdrj,
[ L l1 - m2) am · cos mq>] • (4.81)

The normal stress can be determined as follows:

24 Szcchy: The Art of Tunnelling

o--,=-- .
----
- - _ -_- _- __ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ -····· --

370 DESIGN OF TUNNEL SECTIONS

Let the external active and reactive forces applied upon an elementary section
;;Is of the arch be Pds. Its radial component is Rds.. The equilibrium equation of
all forces acting upon ds is (Fig. 4/70) :

Rds-N sm
I. 2d<j) )
2-dQcos d</> =0, sm
. dq,
2 2
~ d</>
2
, d</>
COST~ 1,

ds=rd<j), Rrd<j)-Nd<f>-dQ=O,

dQ
N= Rr - d</> .

FIG. 4/70. Equilibrium

diagram of the external force5

Furthermore, substituting the value of Q

EJ m
N = Rr - - -
r3
L
m=l
m2 (m2 - 1) am· cos m</> . (.4.82)

In the formula, R denotes the radial component of all e;:xternal active and
reactive forces applied against a unit length of arch.
The application of this method is demonstrated in the next example.

~
18/J-(P.P---.J dy

FIG. 4/71. The symbols

used in the numerical exampl.:

Given a circular tunnel section (Fig. 4/ 71) loaded by a uniformly distributed vertical
pressure. The tunnel has a cast-iron lining with a modulus of elasticity E = 10 000 000 t/ mt
and its moment of inertia J = 0·000144 m•. The coefficient of subgrade reaction C = 10 000
t/m 3•
 ELASTICALLY EMBEDDED CIRCULAR SECTIONS 371

First the sum of the potential energy of the ring is determined.

1. The work done by the external forces

In the calculation the number of members in the equations is taken equal to: m = S

pdx = pds cos (180° - 4>),

dsp, = pdx cos (180° - 4>) = pds cost 4>,

dsp, = pdx sin (180° - 4>) = -pds sin 4> cos 4>.

In the work equations the sign of the work will be either positive or negative depending
on whether the direction of the movement and the force is the same or not.

f
3n/ !

T= - p cost 4>r(a 1 coslf, + az cos 24> + a 3 cos 34> + a4 cos 4tj, + a5 cos 54>) d<f, -
•=ff/2

f
3,r/ !

T1 = - pr (a 1 cos 3 4> +a 2 cosz 4> cos 2tj, + a 3 cosz 3,f, + a, cos2 <f, cos 4,f, +
,.Jt
3,./z

+ a 5 cos 2 cos 5,f,) d<f, - pr f (a 1 cos tj, sinz tj, + ;z cos 4> sin <f, sin 2,f, +
,.Jz

+ ~3 cos <f, sin 4> sin 34> + ~, cos <f, sin tj, sin 4,f, + ~5 cos <f, sin 4> sin S<f,) d<f, .

Solving the integrals separately

J
.
3rc/%

cos3 tj,d<f, =
-sin <f, cosz <f,
3
+ 32 I cos tj,dtj, =
[ 3
2
sin <f, cos <f,
+ 32
3 12
] "
sin <f, ,.,z =- 3 ;
4
.
,c/%

J
3rc/Z

J cost <f, cos 2,f, d<f, = cos• tj,dtj, - f cos2 <f, sin2 <f, d<f, =
sin <f, cos3 4>
4
+ 3
4
cos2 1f,d<f,-
..it

sin3 4> cos 4> - -1

4 4
J. sm2 <f, d 4> r 3
= ._ sin "' 4cos "' 3
+ -16 . 2'1'
sm ,I,
+ -83 cf, -
sin
3
"'
4
cos "' +
1 .
+ -16sm 2tj, - -
"']3,./2 = + - 7t
·
8 ,.,z 4 '

24•
 372 DESIGN Of' TUNNEL SECTIONS

J (
3"1!

cos: cf> cos 34> def>= J 4cos5cf>dcf>-3 J 4 sin tf, cos 4 cf>
cos3 cf> def> = - -- - - + 16
cos3 cf> dtf, -
,r/ %
5 5

- 3
J cos 3 cf> def>
[
= - ----
4 sin cf> cos 4
tf,
+
5
1
- S sin cf> cos2 cf> + - sin tf,
1 IS
2 ]a,riz
n1s
= - -
4
15
;
3"/!

J
:n; /
2
cos cf> cos 4cf, def> = Jcos 8 cf> def> - 6 J sin 2 cf> cos• cf> def> + f os2 cf> sin' cf> def> =

sin cf> cos 5

cf> S sin cf> cos3 cf> 15 sin 2 cf> l S -'- . 3
=- --
6
-+ . + - --+ - } . • •
'l'-sm cf>cos3 tf,-3 sm- cf> cos- 4>dcf>+
24 96 48

+
sinScf>coscf>
6 +
1
6
J.,
SlO
-'- - [sincf>cos5 cf>
cf> d'I' - 6 +
5sincf>cos3 cf>
24 +
1Ssin2cf>
96
IS
+ 48 cf> -

3 3 3 sin Scf> cos cf> l .

- sin3 cf> cos3 tf, - - sin3 cf> cos tf, + - sin 2cf> - - cf> + - - - - - - - cos cf> sm 3 q,-
4 16 8 6 24

3 3 ]3"/! 15 3 n
- -
96
sin 2cf> + -
48
cf> =- n- -n + - = 0.
n l: 48 8 16
31'/!

J 11/ 2
cos2 cf> cos Stf, def> = J J
cos7 cf>dcf> - 10 cos5 cf> sin: cf>dtf, + 5 f sin' tf, cos 3 tf, d<f, _

6
sin tf, cos cf, 6 sin tf, cos' <f, 24 sin ef, coz2 tf,
=- -- -- + - - - - - - + - - - - - - - + -48 . -'-
--sm 'I' - -
l 0 . 3 -'- • -'-
sm 'I' cos 'I' -
7 7 5 35 X 3 105 7

- 740 J 5 2
sin 2 tf, cos3 <f,dtf, + 5 siri tf, cos <f, +
7 7
10 J sin' <f,cos <f,d<f, = [ sin <f, cos 6 tf,
+
7
6 24
+ - sin <f, cos• cf, + - - sin cf, cos 2 cf, + -48- sm
. 10 .
<f, - - sm3 <f>
40 sin3 tf, cos: tf,
cos• cf, - - - - - - -
35 3 X 35 105 7 7 5
40x2 sin3 <f, 5 . • • l0ins5 <f, ] aniz 96 160 20 4
- 7x5 3 + 7 sm <f,cos tf, + ·1 5 ,. : -
1
-105 + 105- 3S - + 105 ;

.
3:n;/!

- -.-"']a111
· 2=
J
11/!
sin 2 tf, cos cp dcp = [sina
3 n/Z
2
- - ;
3

-'3,r/!

J
,r/!
cos j
cp sin cp sin 2,f, dtf, = 2 sin: tf, cos2 tf, d<f, =

= sin3 tf, cos tf,

[ 2 --4- - - -
2 l .
- sm24>+ -
tf, ]w: ~ -
3:-r ,t :-t
- - = + -- ;
4 4 4 ,r/ 2 8 8 4
 ELASTICALLY EMBEDDED CIRCULAR SECTIONS 373

J
lnll

J cos tf, sin tf, sin 3tf,dtf, = 3 sin! tf, cos tf,dtf, -
11/2

]a,,,i =
-
J 4 sin' tf, cos tf,dtf, = sin3 [
.
4 sin5 tf,
tf, - - --
5 •i
4 X 2
5
2
- 2+ -- = - - .
5

J J
311/Z
3
cos tf, sin tf, sin 4tf, dtf, = 4 sinz tf, cos• tf, dtf, - 4Jsin' tf, cosz cf, def,= : sin3 tf, cos cf,+
::t/ l

6
+ 12J'sin2 tf, cos: tf,dcf, - 4 sin tf,costf, - sm' cf, dtf, = [2 sm
4J. . 3 tf, cos3 tf, +
6 6 6 3
3
3
sin tf, cos tf, 2 . cf, 2 . s cf, 2 cos tf, sin tf,
+ 2 ----- 4
- -
16
Sill 2 + -cf,4 - -
3
Sill COS
cf,
+-
3
-- 4--- -

J J
311/ Z

fos cf, sin tf, sin 5tf, def, = 5 sin' tf, coss tf,dcf, - 10 sin' tf, cos3 tf,dtf, +
,r/ Z

5 sin3 tf, cos• tf, 5 x 4 sin3 tf, cosz q, 20 2 sin3 cf,

+ sin 6 cf,costf,dtf,= [ - - - - - + - - - - - - - + - - - - -
7 7 5 7 5 3
J

_
10
sins cf, cos2 tf, _ 10 x 2 _ sin5 tf, + sin cf, ] " = _ ~ _ ~
2 3 12
+~+
7 7 5 7 n /Z 105 105 35

Summarizing the results, the work done by the external forces:

T 1 = -pr - - a 1
3 [ 4 + -n4 a2 - - a3
15
4 +0 X a,+ - - a5
105
4] -pr
[2-3 a
- 1 + -8n a. -
•

.
2. The work done by the internal forces

V1 -_ - -EJn
--
2 X r3
[fm~,
.t.,· (1 - m 2) z Gm
•] ,

substituting the values

V1 = - -EJn- (9azz + 64a 3z + 225a,z + 57 6a 512\ •

7.r3
 3'74 DESIGN OF TUNNE"t SECTIONS

3. The work done by the feaclions:

J•• ~ .
~ (am cos m<J,)' d<J,,
V: = - Cr
2 ., 1

~ ~ ~

V2 = - ~ [J
4'1
J f
af cos• <J,d<J, + 2a1 a 2 cos <J, cos 2<J,d<J, + 2a 1a 3 cos <J, cos 3<J,d<J, +
4i1 4'1
•• th ••

+ 2a 1 atfcos <J, cos 4<J, d<J, + 2a 1 asfcos <J, cos 5cf, d<J, + a: f cos 2cf, def,+
••
+ 2a2 aJcos 2<J, cos 3cf, def, + 2a2 a,Jcos 2<J, cos 4cf, def, + 2a2 a 5 cos 2cf, cos 5cf, d<J, + J
J
+ a;Jcos2 3<J, def, + 2a 3 a4 cos 3cf, cos 4cf, def, + 2a 3 a 5 cos 3cf, cos 5<J, d<J, + J
+ .a1Jcos 2 4cf,dcf, + 2a4 a 5 J cos 4cf, cos 5cf, d<J, + a:Jcos2 5<J, def,],

solving the integ_rals

2
V, = - 2Cr [ a12 ( 2<J, + -sin
2 "'·)
- + 2a1 a2 (-sin2- cf, sin 3cf,) + 2a, aa ( --4-
+ --6- sin84cf,
sin 2cf, + - -J' +

sin S<J,) sin 6cp ) ( cf, sin 4cf, ·i

+ 2a1 of (l-
sin 3<J,
6
- +---
10 .
(sin 4cf,
+ 2a1 a 5 - - + - · - - + a.2 - + - -- +
8 12 - 2 8

·sin cf, 5in 5<J,) ( sin 2cf, sin 6cf,) (sin 3cf, sin 7cf, ·).
+ 2a.a 3 (- - + --- + 2a2 a4 - - - + - -- + 2o. a 5 - - + -- - +
- 2 10 4 12 - 6 J4

+ 032 (cf,
2 + _sin6cf,)
1_2_ + 2oa 04 (sin<J,
- 2- + _sin7cf,)
1_4 _ + 2aa as (sin2cf,
- -4- + _sin8c/>)
1_6_ +

+a!(-!.+ sin8cf,)+ 2ojas(sincf, sin9<f,)+a~(±-+ sinl0cp)1.,,.

2 10 2 18 2 20 4>,

Assuming that the deformation at:

3:n: 3:n:
..
</>2 = + 4' </>1 = - 4 lpcations is zero;

and substituting these limits:

V2 =- ~ (1 ·8562of + 1·8856a 1 a 2 - a 1 o3 + 0· 188560 1 o4 + ·0·333301 ds +

+ 2·356190~ + 1·1337a2 a 3 -0·6667a 2 o4 + 0·26937a 2 o 5 + 2·5228605 +
+ l ·21218a 3 o 4 - a 3 a 5 + 2·35619ai + 1·57135o4 o5 + 2·2561a~,
 375
CIRC ULA R SECTIONS
ELASTICALLY EMBEDDED

and the pote ntia l energy

z
EJn (9az + 64a8 + 225 a, +
z z !\
576a61
2 a, ) - 2r3 -
II = pr ( 2a 1 -
3n az
8
+ 52 a3 -
35

228 6a: + l ·21218a3 a• - a 3 a,+

+ 1·13137az a 3 - 0·6667a: a,+ 0·26937a: a, +2·5

2-2562ai).
+ 2·35619a: + 1·57135a, a, +
and putt ing it equ al to zero
quo tien t of the pote ntia l energy
Form ing the part ial differential

an Cr
(2 1·85619a1 + 1·88562a.• - a 3 + O· l 8856a4 + 0·3333a5) = 0,
-=
-aa1 2pr - -
2
X

Cr
an
-- = -
Eln
pr X 1-17809 - - r 3- 9a: - - 2
(l ·88562a1 + 4·71239a.• + I-13
137a 3 -
aa:
- 0·66667a, + 0·26937ag) = 0,
= 0,
an
- - = 0·4p r -
EJn
- 3- 64a 3 - -
Cr
(- a1 + I-13137a1 + 5·0457a3 + 1·21218a, - a 5)
r 2
aaa

an EJn Cr + 0·6667a 2 + 1"21218a3 + 4·71239a

, +
--;-- = - - r3- 225a4 - - (0·18856a1
ua, 2

+ 1·571348a5) = 0,

an EJn Cr
-0. 333 3 a 1 + (0·26937a 1 - a3 + 1·57135a, +
-- = - pr X 0·05714 - -r - 576 a5 3 -
2
aa,
+ 4·51239a5) = 0 .
the folJowing equ atio ns
subs titut ion and rear rang eme nt
Let nEJ be equ al to D. Afte r
are gained:
4
1·8856a 2 + a 3 -0·1 885 6a, - 0·3333a5 + C = o.,
-3·7 124a1 -
•
2·3562
+ 0·6667a4 - 0·26937a5 - -- C- = 0,
-1·8 856 2a1 - (4·71289 + 18D)a 2 - I-13137a 3

0·8
1·21218a4 + a6 + C = 0,
l-13 137 a: - (5·0457 + 128D)a3
-
+a 1 -

(4·71289 + 450 D)a , - 1·57135a 5

= 0,
-0·18856a 1 + 0·66667a 2 - 1·21216a 3 -

0·11428
(4·51239 + 1152D)a5 - C = 0,
0·26987a 2 + a 3 - I-57135a4 -
-0·3 333 a 1 -
~
- --- - - - -
------------
~I .......

376 DESIGN OF TUNNEL SECTIONS

from where ai, a 2 • a can be determined :

• • 5

::EJ 3·1416
10 000 000 X 0·000144
X
D = - - = - --------,--- -,--- -_- - - - = 0·005585 ,
Cr' 10 000 X 9 2

cl = 0·0001,
and the numerical values are:
-3·71 24a 1 - 1·8856a2 + a 3 - 0·18856a, - 0·3333a5 + 0·00040 = 0,
-1·885601 - 4·8129202 - 1·1313703 + 0·6666704 - 0·2693705 - 0·0002356 = o,
+a 1 - 1·13137a 2 - 5·7606la 3 - 1·21218a4 + a 5 + 0·000080 = 0,
-0·18856a 1 + 0·66667a2 - 1·21218a 3 - 7·22366a, - 1·57135a + 0 = 0,
5
-0·3333a1 - 0·26937a 2 + a3 - 1·57135a, - 10·946380 5 - 0·00001142 = 0.
The above equations are solved by the Gauss elimination method. The end ~ suks are:

01 = + 0·00021561 a 2 = -0·0001611 4

03 = +0·0000928 2 04 = -0·0000383 19
05 = +0·0000103 6.
The moments and reactions can now be readily determined.

The eqllattons for the reactions are:

m
Q= -CL amcosmq, = -2·156cos q, + 1·6llcos2 q, -0·928cos 3cp +
m= l

+ 0·3832 cos 4q, - 0· 1036 cos 5¢.

The values of the reactions calculated in the above manner are illustrated in
Fig. 4/72. The magnitude of the radial displaceme nts is in proportion with the
reactions. Values of the displaceme nts are also indicated at places where OJJ
reaction forces will develop and therefore these are indicated with dashed lines.
..
The moment can be calculated from the following equation:

EJ m
M= - 2 L ( I -m 2)amCC!:m tp,
r m.::!

and substitutin g the numerical values we obtain.:

M = 0·07735 cos 2¢ - 0· 11882 cos 3¢ + 0·09196 cos 4¢ - 0·03977 cos Sep.

 TABLE 4/ V

a, a, a, Additive member Multiplying factors

a, a,
- . -- ---
I -J-71240 -1·885 62 +1·00000 -0·188 56 I -0·333 33 +0·00040
-- I

-1·885 62 -4·812 92 -1·131 37 +0·66667 -0·26937 -0·00023 5 6

2
+0·16931 -0·00020 3 2 a1 = -0·507 92 "'
a,. I +1·885 62 +0·957 75 -0·50792 +0·095 776
-0·10006 6 -0·00043 8 2
~
II -3-85516 -1·639 29 +0·76244 ::i
-- ~
3 +1·00000 - 1·13137 -5·760 61 -1·21218 4 +1·00000 +0·00008 0 ~
-1·00000 -0·50792 +0·269 37 -0·05079 -0·08978 9 +o·ooo 108 az = +0·269 368
• «z . I
P1. II + 1·639 29 +0·69706 -0·32421 +0·04255 0 +o·ooo 186 Pi= 0·425 22 i9
-4·79417 8 -1·58718 4 +0·952 761 +0·00037 43 C
Ill
9
4 -0·188 562 +0·66667 -I-21218 -7-22566 7 -I-57135 0 ~
+0·188 562 +0·095 776 -0·05079 3 +0·009 578 +0·01693 1 -0·00002 032 a8= -0·050 792 (l

~
a8 • I
P: . II -0·76244 3 -0·32420 7 +0·13079 0 -0·01979 0 -0·000 086 78 Pi= 0·197 77 ;,o

Yi. III +1·587 184 +0·525 461 -0·31542 6 -0·00012 3 9 y1 = -0·331 06 "'
"'
IV
I -6·539 837 -1 ·889663 - 0·OOO 231 02 B
0
z
5 -0·333 33 -0·26937 +1·00000 -I-517 348 -10·946 38 -0·000 0 l 1 43 "'
a,. I +0·333 33 +0·16930 8 -0·08978 9 +0·016 931 +0·02993 0 -0·000 035 91 a, = - 0·089 79
+0·10006 6 +0·04255 0 -0·019 790 +0·00259 7 +0·00001 139 Pa= -0·025 956
Pa. II
Yz · III -0·952 761 -0·315 426 +0·189 345 +o·ooo 014 39 Yz = ()·198 733

15 1 . IV .;,yl + 1·889633 +0·545 994 +0·00006 675 (51 =- 0·288 942

V -10·178 51 +o·ooo 105 2
I

substitutina the values already calculated.

Values of a., a., ... , a, are o!:!ained from equations I, II, ... , V by successively
...,,
V. - I0•i7851a, + 0·00010S2 - 0.
IV. - 6·5398470, - t ·889 663a, - 0 ·00023102 - 0.
...
-.J

,.
- ---- --

378 DESIGN OF TUNNEL SECTIONS

FIG. 4/72. The distribution of the calculated

reactions

FIG. 4/ 73. Moment diagram of the

elastically embedded ring girder

The resulting moment diagram is given in Fig. 4/73.

The next step is to check whether the displacement at the first assumed location
3
of </> = - n = 135° is in fact zero or not.
4

u = a1 cos </> + a 2 cos 2</> + a 3 cos 3</> + a4 cos 4</> + a5 cos S<J>.
Having substituted the appropriate values we obtain

u = - 0·0041 cm.

As this could not be regarded as zero the calculated values for the reactions and
moments can only be considered approximate . The accuracy of this method
could be improved by increasing the number of the expressions in the work equa-
tion, and by correcting the assumed position of the locus of zero displacement.

43.32. The Polygonal Method

An alternative method for the analysis of an elastically embedded continuous

ring is to replace it by a polygon with a fixed pole and by replacing the elastic
bedding by elastic rods connected to the corner points of the polygon. The polygon
is then considered a statically indeterminate structure.
 ELASTICALLY EMBEDDED CIRC ULAR SECTIONS 379

The external forces are considered to act a~ '?

concentrated loads at the corner points of the
polygon. The elastic earth support is replaced by
radial rods with hinge connections at the corner
points. The elastic properties of these rods are
assumed to be equivalent to the elastic behaviour
of the soil which they' replace.
It is not difficult to see that th~ more sides the
polygon has, the closer it will approach the shape
of a circle, and hence the calculations will also be
more accurate. Generally a polygon with 16 sides
is considered to give sufficiently accurate results. FIG. 4/74. The polygon
(a) Fig. 4/74 illustrates the polygo n replacing the replacing the ring
circular ring. The polygon is considered to be sup- beam
ported along its entire perimeter. However, the ana-
lysis of this structure will indicate tension in some of the supporting rods but
since the soil cannot take any tension, the analysis must be modified to the
extent that in the revised calculatio ns suppo rt is assumed o nly a lo ng that part
o f the polygon where the first analysis indicated compression . Thus, the design
must be carried o ut in two steps and the second sho uld yield the co rrect answer.
This somewhat lengthy method can be simplified where the length of perimeter
a lo ng which the polygonal section is not supported by the ground can be estimat-
ed with reasonable accuracy in adva nce. However, in this case the end results
have to be checked and if the boundaries of the support were no t assumed
correctly, the analysis must be repeated. To illustrate the calculations both
methods will be demonstrate d.
The first step of the first method is to determine the stresses in all the members
of the polygonal ring which is assumed to be elastically supported along its full
perimeter. For an arbitrary loading the stability of the regular polygon illustrated
in Fig. 4/74 is secured only if at one of its points a tangentia l force (Y) is also
a pplied. With symmetrical loading, however, there will be no stresses on this
rod Y, and therefore in the a nalysis of most of the tunnel sectio ns it can be ignored.
The n-sided polygo na l ring is a statically indeterminat e structure with (n + l)
redundant forces. We start the :analysis by cutting the structure back to a stati-
cally determinate state. Each corner point is substituted by a hinge a nd also one
of the suppo rting rods is cut through (Fig. 4/75).
•
The magnitude of the moments M 1 - M{, M 2 - M;, . . . M 11 - M;, applied
at the hinges, a nd the axial force X 1 - X{ acting on the su pporting rod
which was cut through, have to be such that no relative movement sho uld
take place at the points 1, 2, ... n as a result of these and a ll external forces.
11 + 1 equatio ns can thus be established from which the fo rce X 1 a nd
the
mo ments Af1 , M~ , ... M,, can be computed. The uniformly distributed external
fo rces, of co urse, will ha ve to be replaced by concentrated fo rces ac ting at the
corner points.
 -- - - -- - - - - -

380 Dli,SIGN OF TUNNEL SECTIONS

FmA/75. Determinate system of the polygonal ring

FIG. 4/76. Symbols used in the calculations

The problem is usually solved in two steps. First the stresses caused by the
force X 1 in the statically indeterminate hinged polygon are determined, and only
then is the analysis extended to the solution of the statically indeterminate struc-
ture with n + 1 redundant forces.
The calculations usually ignore the elastic deformations in the bars of the poly-
gonal ring but the deformations of tb supporting bars, since they represent the
elastic earth support, must, ·of course, be taken into consideration.
The symbols and notations· us_ed in the calculations are shown in Fig. 4/76.
The axial forces on the bars of the polygon and in the supporting members
of the indeterminate structure are designated by the letters Si, . . . , S; and
~i, ... , R1 respectively_. The ind.ex O is used to designate the stresses in the cut-
back or statically determinate structure due to external forces e.g. S01, S0 ;, R0;,
whereas the stresses created by the arbitrary force X 1 are noted with the index
e.g. S 11, Su, Rli. The second index refers to the location and the first .one indi-
1:
cates the force responsible for the stresses.
On the other hand, the stresses in the indeterminate structure with one redun-
dant force only; are designated by a bar drawn above the letters: S10, ••. R. 11 • ••
 ELASTICALLY EMBEDDED CIRCULAR SECTIONS 381

A is the cross-sectional area of the individual bars, and E is the modulus of

elasticity. The elastic properties of the supporting members are designated by
the letter D, and since they represent a strip of soil of length S

D=CS, (4.83)

where C = the coefficient of subgrade reaction.

The usual symbols a01 , a 11 , a ;k are used to indicate the relative displacements.
The redundant force X1 is given by the following equation:

(4.84)
where

(4.85)

~ Si; ~ RL
a 11 = L- - - s + '----. (4.86)
1 EA ; 1 D

The stresses created by the unit force X 1 = I ton are as follows (Fig. 4/77):
At point 1 the force X1 = I ton is balanced by the internal forces S 11 and S1 n

1
811
= 2 sin<f>/2 ·

Extending the investigation to points 2 and i, generally, it becomes evident that

because of the constant value of <f> the thrust S 1i will be equal in each rod, and

ti~
R,s
•

FIG. 4/77. Stresses in the

polygon with one redundant
force
-
. .
- - .

382 DESIGN OF TUNNEL SECTIONS

that because of the equilibrium of each joint the axial force in each supporting
link must be equal to 1ton.
- 1
Su= 2sinef,/2 ' (4.87)

R11 = lton (4.88)

and neglecting the deformations due to the axial forces in the polygon:

" R~; " 1 n

On= E-E-=-.
D
1. D D 1
(4.89)

The forces created by the external loading P are determined, starting the analysis
similarly at joint 1. If the external loading is vertical the thrust is

Soi= 2 sin</>/2 '

and at joint i the axial forces are as follows: The tangential and radial components
of all the forces acting at joint i are determined and because of equilibrium they
are equal to _zero.
The radial components (Fig. 4/78b) are

-R0 ; - P; cos (i - 1) </> + sin </>/2(So;-i + S 0;) = 0, (4.90)

and the tangential components are

P1 sin (i - l) </> + cos </> /2 (So;- 1 - S 0 ,) = 0, (4.91)

from where
sin (i - 1) </>
S 0; = P ; ~ ~ + So;-1 · (4.92s)

This equation is then used to compute all the other forces. The expression for
the axial forces in the supporting members is: •

R0 ; = -P; cos (i - 1) </> + sin </> /2 (S0;_ 1 + S0;). (4.92b)

The displacement factor can now be calculated.

(4.93)

and the value of X 1 is obtained from Eq. 4.84.

 ELASTICALLY EMIJLO DED C IRCU L A R SECTIONS 383

The fo ri::es in the structure with the one redundant force will then be:

(4.94)

(4.95)

Next the indeterminate structure with the n + l redundant forces will be con-
sidered.
The M 1, M 2 , . . • , Mn moments are calculated from the following equations :

(4.96}

a0 k = ·
I " - -
D ;~ 1 El i= •
s- -
I, R 0 ;R k;+ 1 n M 0 ;Mk; ds ,
-- I (4.97)

(4.98)

'fo determine the unit displacement factors akm one has to know the internal
fo rces created by the unit moment which is applied at the joint of the indeter-
mjnate structure to the first degree. But :first the stresses in the determinate struc-
ture must be determined .
A unit moment appJjed at joint i will produce the following bar stresses in the
statically determinate structure (Fig. 4/78):
The stresses in the bars S 1, •. . , S,, of the polygon will be equal to zero since
the ro tation of the joint will be balanced by two reactions at the adjacent hinges.

BJ b)

!J
..

Fie. 4/ 78. Reacti ons caused by a

unit moment applied
at a joint .x
~ - -

384 DESIGN OF TUNNEL SECTIONS

Considering the individual bars as simply supported beams the reactions due to
the I tm moment applied at the joint i

1 I
R,, -1 =R11+1 =- A./ = - ~
s cos'f' 2
,
r sin¥'
(4. 99)

2cos<J,
Ru= - 2Ru+1 cos</>= , sin¢. (4.IGU)

There will be no stresses created in the other members by this moment.

From the stresses created by the M 1 = 1 tm moment in the structure which
is indeterminat e to the first degree, first the force X 1 has to be determined :

(4.101)

n
011=[).

Substituting the expressions given by Eqs 4.99 and 4.100

X1 = -1- (2
-cos</>
- - -2- ) = -2- (1-cosq,) (4.102)
n r sin <J, r sin q, nr sin q, ·
Having determined x;, the stresses in the supports of the indeterminat e structure are.

- 2 cos q, 2
R -1=R0 -+X1 ·1 = - .- · + . (I -cos</>)=
' ' r sm q, nr sm </>

2
=---;;:-[I +(n- l)cosq,] , (4.103)
nrsm'f'

1 2 1
Ru_ 1 =R11 + 1 =- . <p+ nr sm . </>(1-coscf>)= . c/>[2(1 -coscf>)-n ],
r 0 sm 0 nr sm

in the other supports only X 1 will induce stresses

- 2
R;m = . <p (1 - cos cf>). (4.104)
nrsm
 ELASTICALLY EMBEDDED CIRCULAR SECTIONS 385

Irrespective of the particular joint at which the unit moment is ap.Rlied, the stresses
created by it will always be the same (Fig. 4/78a)
Of the factors akm for those for which m = k + 2 or m = k - 2 the second
part of Eq. 4.98 will be equal to zero, since the angle change taking place is
caused only by the rotation due to the settlement of the support as a result of
the reactions at joints k + I and k - I (Fig. 4/79)

Elastic selflemement of the

U011111j@11111uu1111 IJ111111 /1111/ t/m
supports 1-2

k-

FIG. 4/ 79. Reactions due Ro1c ,2

to the external forces, and to the elastic displacement of the supports

akk = D1 [ 2
Rkk +
- 9
2R-i.:k-I +
-
(n - 3) Rk;
2 ] l
+ EJ 3 ,
2s
(4.105)

1 - R - - - 2 s
~kk+I = akk~_i = D [2Rkk kk+t + 2Rkk-1 Rk; + (n - 4) R k;] + 6EJ , (4.106)

~
akk+z=akk-2= [Rfk -1+ 2RkkRk;+2Rkk-1Rk;+(n-5)Rt]. (4.107)
..
The displacement factors a0 k as a result of the loading P are as follows (Fig. 4/79):

(4.108)

With the aid of the above expressions the factors a 0k and a;k can be calculated
and by solving Eqs 4.~6, also the moments in the structure.

2S Szechy : The Art of Tunnelling

386 DESIGN OF TUNNEL SECTIONS

FIG. 4/80. Virtual forces to

eliminate reactions along the
roof section

All the above calculations rep,:esent only the first step. So far it has been assumed
that the ring. is .elastically supported around its entire perimeter, but since the
support cannot resist tensile forces, in the second step the calculations will have
to be revised so that there should be no tensile forces in any of the supporting bars.
To achieve this, concentrated loads ex, /3 - /3', y - -y' are applied successively
at the points 1, 2, 3, 2', 3', etc. where tensile forces were indicated in the first
analysis. The magnitude of these forces is so chosen that at points 1, 2; 3, .. . ,
etc. the reactions should be equal to zero (Fig_. 4/80). This condition will yield
the same number of equations as there were supports in tension. From these
equations the forces ex, /3, -y etc. can be computed. Returning with these values
to the solution of the fully supported ring the correct answer is obtained. The ,.
equations expressing the condition of zero reaction are

for rod No. 1:

R1 + (Rd - 1) ex + Rp 1 /3 + R,1 y = 0, (4.109)
for rod No. 2:
(4.110)

for rod No. 3:

R 3 + Ra2ex + Rp3/3 + (R1 3 - 1) y = 0.
 ELASTICALLY EMBEDDED CIRCU LAR SECTJONS 387

In the above equations R 1 , R 2, R 3 represent the reactions in bars I , 2 and 3, respec-

tively, for the case of the fully supported ring. Ra1, Ra2 , Ra3 ••• are the reactions
due to the IX = 1 t load, and similarly Rpi, Rp2 , Rp 3 and R 1 i, R 12 , R 13 are reactions
caused by the unit lo.ads f3 = /3' = 1 t and y = y' = 1 t, respectively.
The values of IX, f3 and y are determined from Eqs 4.109 and 4.110.
Another remark remains to be made. For the last supporting bar, which is
still in compression, the value of D has to be modified since in the expression for
D = C-s the value of s is smaller owing to the fact that the ring is not fully sup-
ported along this section . Usually this value of Dis assumed to be 1/2D.

Fie. 4/81. Partially

embedded polygonal ring beam
and its determinate structure

(b) It is common practice to assume from the beginning the length of perimeter
along which the ring will se parate from its bedding and a nalyze only this partially
supported ring. In this case the calculations can be carried out in one step. How-
ever, if the assumptions prove to be wrong, the computations wi ll have to be
repeated.
A polygon with 16 sides, and which replaces a circular ring loaded by a uni-
formly distributed vertical load, is shown in Fig. 4/81. This structure is statically
.
indeterminate to the 14th degree. However, as both the loading and the structure
are symmetrical only 8 of the unknowns have to be determined. It can be reduced
to a statically determinate structure by replacing 14 joints with h inges. Then the
moments M 1 - M{, M 2 - M ;, .. . , M 9 - M; a re applied at the joints so that
as a result of these moments there will be no movement at the fictive hinges.
(See left-hand side of Fig. 4/81.) Writing the condition of no movement for each
joint we obtain 8 equations from which the 8 unknowns can be calculated. In the
calculations fo r the exterior panel the value of D is taken as D/2.
388 DESIGN OF TUNNEL SECTIONS

The solution of the problem is demonstrated by a numerical example illustrated in Fig.

4/81.
It is assumed that as a result of the deformation of the structure, the ring is not supported
along the top quarter ring segment. The ring is then transformed to a determinate structure
by insta1ling hinges at joints 1, 3, 4, 5, 6, 7, 8, 9 and at the corresponding symmetrical joints
1
(1 , 3', etc). To determine the relative movements at these points one has to calculate the
internal stresses of the structure as a result of both the external forces and the phantom
unit moments applied at the hinges.

r = 3·0 m; J = 0·000144 m4 ; E = 10 000 000 t/m';

C = 10 kg/cm3 = 10 000 t/m 3

the width of t_h e investigated ring = 1 m;

<Jr= 22° 30'; s = 2 sin <J,/2 = 1·17 rn; l = r sin 45° = 2·12;
f = r (1 - cos 45°) = 0·88 m; y = r (cos <J, - cos 2<J,) = 0·65 m;
e= r (sin 2<J, - sin <J,) = 0·97 m.

By reducing the uniformly distributed loading to single concentrated forces acting at the
comer points we obtain :
P1 = ls cos <J,/2 = 1·15 t;

p % = - s (cos -<J,
2 2
+ cos -3<J,)
-
2
= 1·06 t,

2 + cos 2
P3 = s ( cos 3<J, 5<J,) ~0,81 t,
2

2 + cos 2 = 0·44
P, = s ( cos 5<J, 1<J,) t,
2
1
P = ; (cos : ) = 0·11 t.

The calculation of the stresses is started with the analysis of the 3-hinged arch, forming
the top part of the ring (Fig. 4/82).

f.f5t
..
f-QfJI f{J6t

Fm. 4/82. The three-hinged

top section of the polygonal ring
 ELASTICALLY EMBEDDED CIRCULAR SECTIONS 389

Writing the moments about point 1:

S 01 = -H0 cos <J,/2 - (Ve - P 2) sin <J,/2 = -2·62 t,

M 02 = -2·56' x 0·65 + 1·65 'x 0·973 = -0·015 tm.

Next the stresses in the hinged polygon are determined. The tangential components of all
forces acting at joint 3 (Fig. 4/82) are:

- (P3 + Vo) sin 2</, - H0 cos 2</, = s03 cos <J,/2,

1
S 03 = - cos </,/ 2 [(P 3 + V0) sin 2</, + H 0 cos 2<J,] = - 3·62 t,

and the radial components of the same forces are:

R 03 = (P 3 + V0) cos 2</, - H 0 sin 2</, - s 03 sin <J,/2 = -0·787 t.

The stresses in the other members are calculated similarly from the condition of equilibrium
of all forces acting at the joints.
In the remaining members the stresses are all the same and equal to

S 06 = s01 = s08 = -4·15 t;

R 118 = Re, = R 08 = R 09 = 2s08 sin <J,/2 = -1 ·62 t.

Summarizing the stresses and moments from the external loading (Fig. 4/83):

p =I t/m
0011111111 Illlll ll Irnf1l11 I!II Iii Ill Iii I Iiii
1- -0·075

...

FIG. 4/83. Internal forces

 390 DESIGN OF TUNNEL SECTIONS

Axial stresses are:

S 01 = -2·62 t R03 = -0·787 t
S 0: = -3·04 t RM = -1·32 t
S03 = -3·62 t Ros= -1·59 t
So, = -4·03 t R 00 = R07 =R03 = R·09 = -1·62 t
S 05 = -4·15 t
s.G = So7 = s<,8 = -4·5
For the calculation of the internal stresses due to the unit moments first the 3-hinged
arch section is analyzed:
1
V = O·, H =-
f = 1·14t,

S 11 = Hcos <f,/2 = +1·12 t,

Su = Hcos 1·5<f, = 0·95 t,

M, = H, = +1·14xO·65 = +o·74 tm.

To the analysis of the hinged polygon: the tangential components of the forces acting
at joint 3 are:

H cos 2</, - S 13 cos <f,/2 = -0,

Hcos 2</,
S1 3 = - - -- = +0·821 t,
cos <f,/ 2

and the radial components are :

H sin 2<f, + S 13 sin <f,/2 - R, = 0,

Rt= Hsin 2<f, + S 13 sin <f,/2 = +0·965 t,

Su = S16 = S, 6 = S 1, = S18 = +0·821 t.

R 3 = 2S 13 sin <f,/2 = +0·32 t,
R, = R. + R 6 = R1 = R8 = +0·32 t.

The internal stresses produced by the unit moments applied at joints 3, 4, 5, 6, 7, 8 and
9 can be found in a similar manner. The rcsQlts are indicated in Figs 4/84a-d. The stresses
due to M 6 , M 1 and MR are identical with that created by M • Next, the factors a;k and a oic
6
are calculated:

D = Cs = 10 OOOX 1·17 = 11 700 t/m2,

EJ = 10000000 t/m'xO-000 144 m• = 1440 tm~.

 ELASTICALLY EMBEDDED CIRCULAR SECJ1ONS 391

1 ·ftm
0 I

FIG. 4/84. Stresses and

reactions caused by unit moments applied at different joints

The strains due to the axial forces are neglected and for simplicity the EJa11c terms are
calculated.
Thus:

..
EJ J 440
- = - - = 0·123.
D 11100

The value of D' is assumed to be D/2 at the first 45• support and hence

EJ
- = 0·24/:
D'
392 DESIGN OF TUNNEL SECilONS

The calculation of a 11 (see the moments and axial forces indicated in Fig. 4/83;
the moment diagram for mf is illustrated in Fig. 4/85):

a 11 = 32s
2
m 11 +
2s [
1+4
( m 11 + m 12 ) 2
2 ]
+mu
EJ 2
+y2R 13
EJ 2
+Dl1R 12
6 2
2s 2s
= 3 0·548 + .6 (1 + 4 x 0·756 + 0·548) + 0·246 x 2x 0·965 2 +

+ 0·123 X 11 X 0·322 = 2·81.

s - 1-llm
I.
CD ®

FIG. 4/85. Work diagram due to the

rotational displacement of the three-hinged arch
(for the calculation of a 11)

FIG. 4/80. Moment diagrams

of a mixed displacement factor (a 13)

The m 1 and m 3 momC'nt diagrams used for the calculation of a 13 are given in Fig. 4/86.
Determining the 11.reas of the moment diagrams and adding to them the products

2s EJ EJ EJ 2s
a13 = 6 F 0 - D' 2R 13 R 33 + D 2R 14 R 34 - D9 R 15 R35 = 6 (0 + 4 X 0·113 +

+ 0·192 + 0·192 + 4 X 0·233) - 0·246X2 X 0·965 X 1·836 + 0·32 X 0·419 -

- 0·123 X 2 - 0·32 X 4·53 X 0·123 X 9 = -0·309;

au= 2 x 0·965 x 0·871 x 0·246x 2 x I ·61 x 0·32 x 0·123 + 2 x 0·871 x 0·32 x 0·123 =
= +0·355 ;
 ELASTICALLY EMfiEDD'ED CIRCULAR SECTIONS 393

a15 = 0·123 X 2 X 0·32 (0·871 X 2 - 1·61) = +0·0104;

a 19 = +0·005.

The other a1" and a°" factors are calculated in a similar way, the results of which are as
i"ollows:

a 11 = + 2·81 033 = + 3·797 au = + 2·758 a~ = + 2·571

013 = - 0·309 a 3, = - 0·658 045 = - 0·298 a 68 = - 0·298

a 14 = + 0·355 035 = + 0·1725 04R = + 0•187 057 = + 0·187

015 = + 0·0104 a 38 = - 0·0145 a,,= 0 C53 = 0

016 = + 0·0104 037 = - 0·0145 043 = 0 059 = 0

017 = + 0·0104 a 38 = - 0·0145 049 = 0

a 1s = + 0·0104 03q = - 0·007

a 19 = + 0·0052

a 68 = + 2·571 077 = + 2·571 ass = +- 2·758 a,. = + 1·285

a 67 = - 0·298 a18 = - 0·298 039 = - 0·298
a68 = + 0·187 079 = + 0·187
Ogg= 0

001 = - 1·165 a 03 = + 1·325 004 = - 0·154 aOl> = 0

a 08 = - 0·0465 a 07 = - 0·0525 a08 = - 0·0525 009 = - 0·0262

The equations for the moments are as follows:

On M1 + a 13 M 3 + a,1 M 4 + a 15 M 5 + a 16 M 6 + a,, M, + a, 8 M8 + a 1• Mg + a 01 =0
~~+~~+~~+~~+~~+~~+~~+~~+~ = 0

~~+~~+~~+~~+~~+~~+~~+~~+~ = 0

~~+~~+~~+~~+~~+~~+~~+~~+~ = 0
..
~~+~~+~~+~~+~~+~~+~~+~~+~ = 0

¾~+~~+¾~+~~+~~+¾~+~~+¾~+~=0
~~+~~+~~+~~+~~+~~+~~+~~+~=0

~~+~~+¾~+~~+~~+~~+~~+~~+~=0

The equation system is solved by the Gauss elimination process (Tabi.: 4/Vl).
 TABLE 4/VI w
IQ
.p..

M, M, M, M, M, M, M, M, p
I I I I I I
E

I +2-810 -0·309 +0·355 . +0·010 +0·010 +0·010 +0·010 +0·005 -1 ·165 +1 ·736
1
2 - 0·309 +3-797 -0·658 +0·172 - 0·014 -0·014 -O·OJ4 -0·007 1 +1 ·325 1 +4·278
al +0·309 - 0·037 +0·039 +o·OOl +0·001 +0·001 +0·001 +0·001 -0·128 +0·191 a = +0·1100
II +3·763 1 -0·619 +0·173 -0·013 -0·013 - 0·013 -0·006 +1·197 +4·469
3 +0·355 -0·658 +n58 - 0·298 +0·187 0 0 0 - 0·154 +2·190
al - 0·355 +0·03 9 - 0·045 - 0·001 -O·OOJ -0·001 - 0·001 - 0·001 +0·147 -0·219 a = - 0·1262 C,
all +0·620 -0·102 +0·028 -0·002 -0·002 -0·002 -0·001 +0·198 +0·739 a = +0·165 !il0
III +2-611 -0·271 +0·184 I -0·003 -0·003 - 0·002 +0·191 +n10 z
4 +0·010 +0·172 - 0·298 +2·571 -0·298 +0·187 0 0 0 ~
+2·344 ...
aI - 0·010 +0·001 -0·001 0 0 0 0 0 +0·004 - 0·006 a = -0·00356
aII - 0·173 +0·028 - 0·008 +0·001 +0·001 +0·001 0 -0·055 -0·206 a= -0·046
~
ixUC +0·271 -0·028 +0·019 0 0
~
0 +0·020 +0·282 ix = 0·104 t""

"'"'

~
IV +2-535 -0·278 +0·188 +0·001 0 -0·031 +2-414
5 +O·OlO -0·014 +0·167 - 0·298 +2-571 - 0·298 +0·187 0 - 0·046 +2-298
ix l -0·010 +0·001 - 0·001 0 0 0 0 0 +0·004 -0·006 IX = -0•00356 "'
all +o·Ol3 -0·002 +0·001 0 0 0 0 +0·004 +0·016 a = +0·00346
1.1III -0·184 +0·019 - 0·013 0 0 -0·014 1 -0·192 IX= -0·0705
alV + 0·278 -0·031 +~·021 I 0
I 0 -0·002 +0·264 IX= 0·1096
V I +2-527 -0·277 +0·0187 0 I - 0·055 +n81 l
6 + 0·010 -0·014 0 +0·187 -0·298 +2·571 -o-298 +0·187 - 0·052 +2-293
ixI - 0·010 +0·001 - 0·001 0 0 0 0 0 +0·004 - 0·006 IX= -0·00356
all +0·013 - 0·002 + 0·001 0 0 0 0 +0·004 +0·016 a = +0·00346
am +0·003 0 0 0 0 0 0 +0·003 ix = +O·OOll5
aIV -0·189 +0·021 -0·014 0 0 +0·002 -0·179 IX = -0·0745
ixV I +0·277 - 0·030 +0·020 0 - 0·006 I +0·261 a = +0·1096

•
 TAB Lt 4/ V I (cont.)
--
M, M, p E
M, M, M, M, M, M,

+2·5-27 -0·278 +0·187 -0·048 +2-388

VI
-
7
- +0·010 -0·014 0 0 +0·187 --0·298 +2'758 -0·296 -0·052 +2'293
0 0 0 0 0 +0·004 -0·006 X = -0·003 56
al -C0IO +0·001 -0·001
+0·001 0 0 0 0 +0·084 +0·016 a= +0·003 46
~!:j
all +0·013 -0·002
+0·003 0 0 0 0 0 0 +0·003 a= +0·001 15
aIII
aIV -0·001 0 0 0 0 0 -0·001 n
aV -0·187 +0·031 -0·014 0 +0·004 -0·176 ...>
aVI +0·277 -0·031 +0·02 1 -0·005 +0·263 S:
"':s:
VII +2'713 -0·277 -0·045 +2-392 "'tl
0 -0·298 +1·285 -0·0262 +1·146 0
8 +0·005 -0·007 0 0 +0·187
0 0 0 0 0 -0·003 ~ = -0·001 78 "'0
al -0·005 0 0 0
all +0·007 -0·001 0 0 0 0 0 0 +0·008 a = +0·OOI 73 Q
:,0

aIII +0·002 0 0 0 0 0 0 +0·002 ~= G·00076 n

alV 0 0 0 0 0 0 0 a= 0 ...>
C:

av 0 0 0 0 0 0 a= 0 :,0

aVI - 0·1 87 +0·021 -0·014 +0·004 -0·177 a= -0·074 "'

"'
aVII +0·277 -0·028 -0·005 +0·245 a = +0· l023 !:l
5
z
"'
VIII +1 ·243 -0·023 1 + 1·22 1
I results
The M., M ,, . . ., M, values arc obtained from equations l , II, ..., VII! by substituting successively the known
VIII . 1·243 M, - 0·023 - O; VII. 2·713 M , - 0·277 M , - 0·045 = O;
YI. 2·527 M , - 0·278 M, + 0·181 M, - 0·048 - O; V. 2·527 M, - 0·277 M , + 0 ·0187 M , - 0 ·055 = O:
IV. 2·535 M , - 0·278.M , + 0· 188 M , + 0·OO I M, - 0·031 - 0;
Jll. 2·61 1 M, - 0·271 M , + 0·184 M , - 0·003 M , - 0·003 M , - 0·002 M , + 0·191 - O;
II. 3·763 M , - 0·619 M, + 0·173 M, - 0·013 M , - 0·013 M , - 0 ·013 M , - 0·06 M, + 1·197 - O; ""
\0
I. 2·810 M, - 0·309 M , + 0 ·355 M, + 0·010 M, + 0·0I0 M, + 0·010 M , + 0·010 M, + 0 ·005 M, - I · 165 - o. V,

!II
 396 DESION OF TUNNJ!L SECTIONS
Aud from the above set of equation s we obtain the followin g values
:
M, = + 0·0185 M 8 = + 0·0225 M 3 = - 0·331
M8 = + 0·0185 M 5 = + 0·0133 M: = + 0·1265
M 7 = + 0·0197 M, = - 0·0735 M = + 0·388
1

The final momen t diagram is shown superimposed on the ring

in Fig. 4/87.
These results are notably differ-
ent from those obtaine d by the
BooRO V- GOREL IK method . Only
the shape of the two diagrams
is similar and also the maximum
values of the positive momen t at
the crown are in good agreem ent
(see Fig. 4/73).

43.33. Bougayeva's Method 4 •5

For the design of the elastically
embedd ed ring, BOUGA YEVA de-
veloped a simple method which
provides an approx imate but
quick solution. 4 · 6
FIG. 4/87. Momen t diagram The method takes the elastic
embedm ent into accoun t by de-
termini ng the approx imate val-
ues of the elastic reactions so that they satisfy both the conditi on
of equilibrium
and the conditi on which states that the displacement of the structu
re and of the
embedm ent at the springi ng line are equal. Thus, only at these
two diamet ri-
cally opposit e points is the magnit ude of the elastic reactio n correct
. At any other
point the magnit ude and distribu tion of the reactio n are arbitrar ily
assumed val-
ues, which, however, are close to their real values.
The distribu tion of the subgra de reactio n as a result of a uniformly
distributed
vertical loading is illustra ted in Fig. 4/88. Typical values of this
diagram are
given by the following expressions:

if e < 45° the reaction is zero;

..
if 45° < e < 90° ; Cb = Cbv cos 2e;
if 90° < e< 180° ; Cb = C{Jv sin2 e+ Cb1 cos 2 e.
u BoUGAYEVA, 0.: Raschot i tonnelni kh obdyelo k krugovo vo
oscherta niya (Compu tation
of circular tunnel linings) lzvestiya Gidrotek hniki, 1951
u ZURABOV, T. and BoUGAYEVA 0.: Gidrotek hnichesk ie Tunneli
(Tunnel s for hydraul ic
power plants), Gosener gizdat, Moscow 1962 p 352
 ELASTICALLY i!MBEOOEO CIRCULAR SECilONS
397

Fm. 4/88. The distributi on

of soil reactions according
to BOUGAYE VA

p
I Ifliifltl tilliIf I016\ iii if Ii Iii Iii II

FIG. 4/89. Determin ate

structure of the ring girder

The structure is analyzed as a statically indeterminate structure to the third •

the
degree and, therefore, further two equations are required to determine
unknown values of Cb 0 and Cb1. These are: the equatio n expressi ng the equi-
the
librium of the forces, and the equatio n in which the horizontal deflection of
ring is equated to the compression of the soil at this point.
The analysis for a symmetrical and uniformly distributed vertical load (Fig. 4/89)
nate
is carried out in the usual manner. First the ring is reduced to a determi
applicat ion of the
structure by cutting it through at the crown. Then by the
reduced structur e is
moment X1 and forces X 2 and X3 at the elastic centre, the
made to act as the continuous, indeterminate structur e.
398 DESIGN OF TUNNEL SECTIONS

Because of symmetry the force X 3 will be equal to zero. X and X are deter-
1 2
mined from the condition of zero displacement at the crown.

(4.111)

(4.112
Because of the symmetry the displacement factors are calculated for the half
section only :
"
au =Jmi ds = !!!:._
EJ EJ'
0

- a22 ~ ds = 2 f•
=J··•mEJ r3 cosz <P d</J = n,a
EJ 2EJ.
0 0

The coefficients a0 i, a02 are expressed as functions of C, Jv and 1J .

1

p. dt i
fllilUII ijlliIm I~
; I
I

(r.rir,fJ-r;,.sinJ)

b;
ll@lllill/llllllilil~ p
•

Fm. 4/90. The calculation of

stresses in the ring
beam at typical sections
 ELASTICALLY EMBEDDED CIRCULAR SECTIONS

The moment diagram is determined first for the statically determinate girder
from the external loading. The half ring is divided into three sections and the
moments for each are determined separately (Fig. 4/90).
There are no reactions within the zone where: 0 < <Ji < n/4.

"'
M ' = - Jprk cos e (r sin</> -rk sine) de= - ~ prk r (2 - '; Jsin 2
<P ,
0

and adopting the notation

rk M 0' 1 . 2 <i>
(l.=2 --- - : =- prkr(I.Stn (4.113)
r 2
the moment where n/4 < <f> < n/2 (Fig. 4/90b)

M; = -½prkocsin2 </> -J"'

,r/4
Corsin(</>- e)rkde = - ~ prkrocsin2 ¢ +

"'
+ Cov rk rJ cos 2e sin (cp - e) ae = - ~ prk roe sin2 cp -
11/4,

-- Cov rk r [-½-cos 2¢ + 0·4714 (sin¢ -cos cp)] . (4.114)

and in the third section where n/2 < <P < n the moment can be computed as
follows:
The forces transmitted from the upper sections are represented by the resultant
of their vertical and horizontal components and a moment

"'
M~" = M + Pvr(l - sincp) + PHl'COf,e - J Corsin(<P - e)rkde
11/2
In this expression

,r/2

P.,=prk+ J Cocoserkde=rk(p+0·l381Cov);
11/4
n /2

P11 = J Cosine rk de= 0·4714 Co. rk

1t/•i
 400 DESIGN OF TUNNEL SECTIONS

Substituting these values into the expression for M; and integrating

M~" =- rk r [p(sin </> + 0·5a - 1) + Cciv (-0·4714 cos</> - 0·1953 sin</>+

+¼cos 2</> + 0· 5) + Cci1 (0·5 - 0·1667 cos 2</> - 0·6667 sin</>)].

(4.115)
Then writing the expressions for 0 01 and a02

M SMEJ ds
a01 =
f EJ
0 m1
ds =
0

1t/4 1t/2 1t

o01 = _EJ
r [JM' +JM" +fM'"d"'] ·
0
· o d"'
"'
1t/4
o d"'
'I'
1</2
o "' ,

and making the appropriate substitutions

0 01 =- ;J,2 [p(l·l781ix - 0·5708) + 1·0899 Cci + 0·11875 C8

r
0 1 ];

002 =- ,2 J
EJ M 0 cos</> d<J> =
1</4 1</2 1t

= - ;~ [ f0
M~ cos</> d</> + f
1t/4.
M; cos<J> d</> + f
,c/2
M~" cos</> d</> J;.

Writing these values of 0 0 1, 0 02 , a 11 and a 22 into Eqs 4.111 and 4.112 we obtain:

X1 = rkr [p(0·375a - 0·18169) + 0·34694Cci. + 0·03778Ccif]; (4.116)

X2 = rk [p(0·2122la - 0·31831) + 0·52427Cci, + 0·07073CcirJ- (4.117)

To determine ci, and ci1 two additional equations must be established:

(4.118)

and the sum of the vertical components of all forces is equal to zero

EY=O. (4.119)
 ELASTICALLY EMBEDDED CIRCULAR SECTIONS 401

In Eq. 4.1 18 •:iP. v'?J ue 00• is the displacement of a point of the reduced s!ructure
located at the horizontal diameter and caused by the external loadings on the
structure. Simil::i , ly 01v and 0 2, represent the displa~ement of the same point owing
to the action of the unit moment X 1 = 1 tm and of the unit force X 2 = 1 t, re-
spectively.
I "
<> 10 = --jM,, m1 ds , where m1 = ltm, M ,, = - rcos<jJ ,
EJ.

<52, = + El
1 J "
3 2
r cos <P d,P = 4 EJ ,
nr3
11/2

~
uo, =
l
EJ
J M o M v d S = - El ,2 J "
M o cos¢ d¢ =
Ill

1'/2

=r;lr 3
[0·Sp(l - a) - 0·82807Co, - 0·lllllCb1 ]

On substituting these values into Eq. 4.118 we obtain

1
Cb,( / + 0·06937} = p(0·06831 + 0·04167a)- 0·01778Cb1 . (4.120)
r~C ,

and Eq. 4.119 becomes ·

,,
prk + 0· 1381Cb, rk -t- J Crk (b, sin2 ( + b1 cos2 () cos ( d~ = 0 ..
1'/2

Writing the above equation into Eq. 4.120 and rearranging it we have .

Cb
V
= 0·04167 (1 + a)
m + 0•06416 p•
Cb _
'f - P
I.
15
-
0·0122 (l -t- a)
m + 0·0641 .5
j
26 Sz~chy: The Art of Tunnelling
#"
•
-

DiSIGN OI' l1,1NNEL SECTIONS

From where the expressions for X 1 and X 2 as ·giv_e n in Eqs 4.116 ~nd 4.117
will become· · ··

X1 = prk r [o·38~ix - 0· 125 + 4

O·Ol (l +.ix)'] ·
m + 0·06416 ·'

0·02l(l+ix )]·
X2 = prk [ 0·21221 (ix - 1) + m + _ (4.122)
0 06416

Knowin_g X1 and X 2 , the stresses at any point of the ring can be calculated

(4.123)
I

N = N0 + X2 cos cp. (4 124

Writing into these equations the values of X1 and X 2 , BouGAYEVA derived for the
moments and normal forces at an arbitrary point of the ring the following expres·
sions:

N = prk [Dix + F + Gn(l + ix)].

She also gave tabulated values for A, B, C1 , D, F and G:

1 EJ
n= - - - -- m-
m + 0·06416 - rk r 3 C

A B c, D F G
\
</> = 0 0·1628 0·0872 +0·0070 0·2122 -0·2122 0·02100
;r;/4 -0·0250 0·0250 - 0·00084 0·1500 0·3500 0·01485
1i/2 -0·1250 - 0·1250 0·00825 0·0000 1·0000 0-00575
31i/2 0·0250 - 0·0250 0·00022 -0·1500 0·9000 0·01380
:n; 0·0872 0·1628 -0·00837 - 0·2122 0·7122 0·02240

If we now calculate the moments for the previously solved numerical example
we obtain the moment diagram indicated in Fig. 4/91 :

r= 3·0m =rk, 10000000 x0·000144 =~·144 = _ ,

m= 0 001 78
J = 5·000 144 m', 34 xl0000xl·0 81

E = 10 000 000 t/m2, 1

n= - - - - -- -- = 15·2,
C = 10000t/m3, 0·00178 + 0·06416
 ELASTICALLY EMBEDDED CIRCULAR SECTIONS 403

o: =I M = 9[A + B + 30·4Ci]

b = the width of the ring = 1 m.

These results indicate that the ap-

proximate values of this method are
in good agreement with those ob-
tained earlier. The character of the
moment diagram is identical with-
those determined by either the poly-
gon or by the BODROV-GORELIK
method (see Figs 4/87 and 4/73).
The values for the moment at the
crown and the upper quadrants are
within 10% of the moments calcu-
lated by the BODROV-GORELIK Fm. 4/91. Moment diagram
method, and become negligible in
the lower half section of the ring.
In the following we shall demonstrate by a few examples the extent to which
the moment at the crown is affected by the coefficient of subgrade reaction.

C =
50 kg/cm3 Mmax = 0·033pr 2
C =
JO kg/cm3 Mmax = 0 ·034pr2
C =
1 kg/cm3 Mmax = 0·079pr 2
C= 0 Mmax = 0·25pr2
As shown, with an increase of the coefficient of subgrade reaction there is a
decrease in the the moment but this tendency rapidly diminishes if C > 10 kg/cm3 •
The above values, and the diagrammatic representation in Fig. 4/92A. both
illustrate that even a small value of subgrade reaction considerably decreases the
moments and stress~s in the structure as compared with the case of no lateral
support.

M
prt
0-250

0-150

0-079
F1G. 4/92A. Variation of the crown 0-050
moment as a function of
subgrade reaction 10 C

26•
 404 DESIGN OF TUNNEL SECTIONS

The maximum value of the moment is

(4.125)
when neglecting the difference between inside and outside diameters.
By substituting the constants we get

,lfma~ = pr 2 [o·25 - 0·007 X 2

m
1
+ 0·06416
f
] = pr 2 0·25 -
m
O·O
14
+ 0·06416
],

The variation of the term inside the brackets as a function of m is illustrated

in Fig. 4/92B. (For comparison the values obtained by DAvmov's method, as will

Mma,
p.? O-Z5---- ----==:::~~;;.;=;-------
----------r- 7---...~
I ~
I ~
: ~~
0-95
0-10
,
/ :I
0-237 (Bvgajeva) ~ l FIG. 4/92B. Variation of the crown
moment as a function
1 {0·234) Oavidov i ~ of the rigidity of the
005 ' i I ~
0.()32 (o ' 5,) c::, ring and the
subgrade reactions
0 0-1 m-t m.EJ
C.r~

be discussed in the next section, are also shown in the same figure.) An inspection
of Fig. 4/92B also indicates that the maximum moment increases from 0·032 pr 2
to 0·25 pr 2, i.e. about 7 or 8 times while the degree of lateral support ranges from
•
f u11 restramt to no support. When the value o f m = C, EJ . . . f
1s mcreasmg
4
rom Oto I ,
the moment is also considerably increasing. This increase of the moment is quite
rapid for small values of m, then tapers off gradually and becomes unimportant
for m values greater than l ·O, and while the value of m changes from l ·O to oo
the moment increases only from 0·237 to 0·25. That is, even a small lateral support
can considerably reduce bending stresses in the structure. However, the bene-
ficial effect of the subgrade reaction is felt only up to a certain point, beyond which
a further increase in the coefficient of subgrade reaction will have no practical
effect on the magnitude of the moment (see also Fig. 4/92A).
For all practical purposes one can say that if the value of m is greater than l ·0
then the effect of lateral support can be regarded as insignificant. However, for
cast-iron segments and liner plates even if C = 20 kg/cma the value of m is less
than 0·05, i.e. it is in the range where the moment varies rapidly with a small
change of the subgrade reaction (see Figs 4/92A and 4/92B). According to the
foregoing calculations the moment will range between 0·03pr 2 and 0·05pr 2 •
It may be mentioned here, that according to the experience gained with the
dimensioning of the various circular sections of the Budapest subway it is always
 ELASTICALLY EMBEDDED CIRCULAR SECTIONS 405

p p

I
~ go• \
I 'r \\
I
I
F1G. 4/93. Scheme of equivalent I
I
I
arches for rapid \
determination of '\ /'
max. B.M. ', /
--- _..--- /

Lhe top section without subgrades where the maximum B.M. values are acting.
These B.M. values will act on the crown section and its value may be computed
with sufficient approximati on in the following way:
Let us substitute the circular ring either by a hingeless arch with a central angle
of 120° or by a two-hinged arch with a central angle of 90° (Fig. 4/93). When these
equivalent girders are of the same radius and load~d by the same uniformly dis-
tributed external load p, then the computed B.M. values will be about 0-03
pr 2 and will correspond well with the B.M. values acling in the case of an
e11triely rigid embedment. Should the section be elastically embedded and cha-
racterized by an elastic subgrade coefficient C, the upper limit of which cannot
be more than 20 kg/cm3 , then

Mmax=P'
,..., 2 ( 0-014 ) ,..., 2
0-25 - m+0·0 (o·25 - 0-014 )- 2
642 =P' 0 1172 =0·121pr
_

where the value of mis tahn according to BouGAYEVA's method (cf. Eqs4.121
and 4.122).

43.34. Davidov's Method

For the analysis of circular tunnel sections DAvmov also developed an approxi-
mate method in which his assumption as to the distribution of the ground reac-
tion is similar to those made by ZuRABOV and BouGAYEVA for horseshoe sections
(Section 42.2 l)
The variation of the ground reaction is
p
expressed as a second degree trigono- uu1mu11mmmeiim1u110
metrical function, so that the condition
of elastic embedment is satisfied only •
at the points of the horizontal diameter.
For the case of a uniformly distri-
buted vertical load the external forces
acting on a circular ring section are shown
in Fig. 4/94. For simplicity the lateral
active earth pressure diagram, e 1 is assu-
med to have a similar shape to the dis- F10. 4/.94. Forces acting on a ring beam
tribution of the lateral earth resistance, e.,.. (DAVIDOV)
-- - - - - -- - --
----------~ -

406 DESiGN'OF TUNNEL SECilONS

The calculation is carried out in the following steps :

1. The stresses in the _ring are determined, neglecting the effect of the lateral earth reaction
and also the deflection at the horizontal,diameter

where Mp= the moment due to external load

M H = the moment caused by the load H = 1 t.
Thus
pr' (4.126)
u. = 12EJ .

2. The stresses, due to the.horfaontal load e2 are calculated, and the horizontal deflection
Uz of the structure is determined for this load condition:

101 r•
U: = - 1440 EJ .

3, Next the compression of the soil u,, caused by the initial horizontal pressure e, is calculat-
ed at the line of the horizontal diameter:

eH
u, = - -1 (4.127)
Eo

The value of e is that value of the horizontal loading by which the compression of the
soil just begins

Up
Up+ 11:e = 0, e=--.
U:

Substituting the values of Up and u2

e = 1·19p.

4. Then e 2 is determined, utilizing the expression that the deflection of the structure must
be equal to the compression of the soil:

pr•
"• - = - - - --
e: = - - 12 -- (4.128)
u, - u 2 EJH1 101 4
-e;;-+ 1440 r

Having determined e 2 , a ll external loads are known and thus both the moments and-
the axial forces in the structure can be calculated.
Another noted difference between DAvmov's and other methods considering the elastic
embedment of the tunnel section, is that he does not use the coefficient of subgrade reaction
to determine the compression of the soil but calculates it in the way as settlement analyses
are carried out. ·
 ELASTICALLY EMBEDDED CIRCULAR SECTIONS

The compression of the soil

eH
u, = - -1
Eo
where e is the loading, II1 is the thickness of the earth column considered to be compressed .
and £ 0 is the modulus of compressibility of the soil.
When determining H 1 = rJr, he assumes an active zone limited by the condition that there
the maximum value of soil stresses due to lateral pressure just attains 120% of the geostatic
p ressure
Umax = l ·2aaeol. = Ae (4.129)

The corresponding values of T/ and A are tabulated by him. Thus DAvmov eliminates
the use of the coefficient of subgrade reaction.
Substituting the numerical values of the previous example (see Fig. 4/71).
r = 3 m, p = 1 tm
EJ H 1 0·0835 0·0835
if - , -E =1 . 1 + 0·07 = ~ = 0·078
r o

0·0835
if e, = -O·J-+ - - = 0·49
- 0·07

The resulting moments:

pr'
from the loadingp: Mp = (1- 2 sin• t/1),
4
e r2
a nd from the loading e,: M, = ; (21 - 48 cos2 ti>+ 8 cos• t/,).
6

At the crown for the combined effect:

• •
(~-~)
2
pr 19
M = - - - -96 e.- r- = r- 4 96
4

and substituting the numerical values

M = 0·234r2 M' = 0·153r'
Comparing the results with those calculated by BouaAYEVA's method we obtain·

Moment

Bougayeva Davidov
EJ
r4 C =
0·237,• 0·234r2 ..
EJ O·I53r2
-=O·J 0·165r'
r'C
EJ 0·044r1
- =0·0) 0·062r2
r•c
EJ 0·015r'
- 4 =O 0·032,·
r C
 408 DESIGN OF TUNNEL SECTIONS

The values calculated by DAvmov's method are consistently lower, because he considers
the
lateral support to be acting along the full height of the section, while BoUGAYEVA neglects
the
support on the upper quadrant. The deviation between the results becomes increasing
ly
larger with the rigidity of the support.
The moment diagram is symmetric al about both axes, and thus it does not reflect
what
experience and observation indicates, namely, that the moments in the top part are
larger
than in the bottom part. Therefore BouGAYEVA's method is considerecl to be more accurate.

43.35. Varga's Development

By developing and simplifying DAvmov 's assumptions, L. VARGA4 • 7 arrived

at a very simple approxim ate method for the solution of an elastically supporte d
ring section.
He assumed the (jistribution of the subgrade reaction to be triangula r as indi-
l:ated in Fig. 4/95, which, he calculates, would result in about 3 %-4 % inaccuracy
only.

.. l.r,n/.i Fm. 4/95. The lateral

support of a
ring beam,
according to
L. VARGA

When calculating lateral displacements, instead of considering the coefficient

of subgrade reaction, he compute s the compression of the soil. He does this by
considering the two triangula r loadings acting on the vertical tangent of the ring,
as uniformly distribute d vertical pressures , from where he calculates the horizon-
tal stresses and the compres sionofan m0 = 3·5r thick layer, using the modulus
of compressibility of the soil.
The derived values for the double triangula r loading are:
For the upper half

The reduced moments are

if O ~ex~ 45° ..
Ma
-
2
= -4 (0·248613 - 0·378155 cos ex).
qr 3

•· 7 VARGA, L.: Szimmetri kusan terhelt alland6 vastagsagu

korboltoz atok es korszelven yli
alagutfalak meretezese (Dimensioning of symmetrically loaded circular tunnel linings
of
constant wall thickness). EKMOE Tudomtinyo~ Kazi. 1961
 ELASTICALLY EMBEDDED CIRCULAR SECTIONS 4()9

The reduced thrust is

N 4
_L = - 0·378155 cos ex;
qr 3
if 45° ~ ex ~ 90°
the reduced moment is
2
M 0'= -4 r0·165280 - 0·024601 cos ex - -cos- ex 0·235702cos3 ex) ;
qr 2 2
-
2

the reduced thrust is

N'a = -3 cos ex ( 0·024601 + cos ex - .j-O· 5 cos2 IX ) •

_
qr 4

In the lower half

the reduced moment is

Mp . fl - -sin-
qr 2 = 4 (0·165280 + 0·024601 sm
2
fl + 0·166667 . fl) ;
sm3
3 2

the reduced thrust is

N q: + 34 sin fl (sin fl 2 - 2sin fl - 0·024601 J,

the horizontal displacement is
qr2
A= - 0·067408 - - ,
EbJ
the displacement factor is
A Ebv3
J = - ~ = - 0·808896
qr4

(where v = the wall thickness).

The approximate value for the compression of the soil is
..

from where
(4.130)

where E, =;= the modulus of compressibility of the soil.

410 DESIGN OF TUNNEL SECTIONS

The method enables one to take into consideration the initial void h, left between
the structure and the surrounding ground due to improper grouting. On the basis
of the identity of displacements we get

r4 r
J=-E 3 L,P;O;= h+1·75q - ,
· bv E,

where P; is the external loading on the ring, and o is the corresponding displace-
1
ment factor at the springing, which are either known or can be calculated. Thus,
from the equation for a given value of p, q can be calculated and the design of
the ring completed. To facilitate the calculations, in his paper the author provides
tables for the values of M and N.

43.36. Meissner's and Orlov's Methods

MEISSNER bases his computa tions on a further simplifying assumpti

on. In his
paper (Bauingenieur, 1964, No. 4) he deals with the dimensioning of flexible
shield-driven circular tunnels composed of articulated jointed lining segments.
He assumes in his approximate method that both vertical and horizontal pres-
sures are uniformly distributed (Fig. 4/51) and that their original magnitude is
equal to the overburden pressure and to the earth pressure at rest, respectively.
But, owing to the deformation of the circular ring, the intensity of these pres-
sures will undergo a change and as a consequence of the horizontal elongation
the lateral pressure will be raised by a mobilized Pmax value, whereas the vertical
pressures will be decreased by the same value. (This latter assumption is highly
debatable.) Thus, the following formulae are derived:
For the horizontal displacement:

J = yzA r4 (1 - A - 2Pmax) COS 2<X .

12Eb J n YZ.11. '

or' the, llending moment at any section:

M yzA
=- - --,2
4
l1 - An - -2Pmax
- - cos 2a,
YZ.11.
J'
•

and for the additional pressure:

YZA (1 -- An)
Pmax = --- - -£ - - - - - -
4 1 1
2 + __!_ ~
,a E,
where y = the dry soil-density
 ELASTICALLY EMBEDDED CIRCULAR SECTIONS 411

Z,4. = depth of tunnel axis below ground-surface

).n = the coefficient of earth pressure at rest (~ 0·5)
Eb = the modulus of elasticity of the tunnel lining
E, modulus of elasticity of the soil
r = the radius of circular section
µ = Poisson's ratio of the soil
(X the angle between the radius of the section and the vertical.

A stricter solution has been given on the basis of elasticity theory by ORLov.4·7a
Instead of the displacement obtained by WINKLER's theory of elastic subgrades,
he determined the radial and tangential displacements of an annular ring lying
in an infinite elastic body. An intermittent pressq,re distribution was given
by him in the form of a Fourier series. While maintai.n ing his concept a solution
in rigorous mathematical form was derived by L. R6zsA and G. Kov,ks with
the additional consideration of the favourable effect of tangential frictional forces
brought about by the perimeter displacement. The somewhat elaborate method is
more of theoretical than o_f practical importance.uh

Summary

In the foregoing several methods have been put forward concerning the use
of which the following can be said:
First, one has to distinguish between flexible and rigid linings as well as between
soft and stiff soils. Where the value of m =
E~ is greater than unity the effect
Cr
oflateral support need not be taken into account and the ring can be analysed as
a freely standing structure acted upon only by the different loads as discussed
under Section 43.22. (In very soft, saturated soils the distribution of the loads
around the ring will be close to hydrostatic.)
For m values Jess than unity the lateral support should be taken into consider-
ation. If great accuracy is not required, the structure can be analysed either as
a fixed circular arch with an internal angle of 120° or alternatively it could be
designed by one of the methods suggested by DAVID0V or L. VARGA.
Liner plate or cast-iron segment tunnels, however, can be regarded as very •
flexible structures, made up of Jinks and connected to each other with hinges
and therefore should be analysed either by the HEWETT-J0HANNESS0N method or

4•7• Owv, N.: Rastshet konstrukciy lezhashtshih na konture krugovovo vireza f ploskosti.

(Computation of plane structures lying along the perimeter of a circular cut) Iss/ed.
po teorii soorusheniy Vip VI, Moscow /954.
Ub Jt6zsA, L. and KovAcs, G.: Stresses an<l Deformatins due to Tangential Forces
along a Circular Cut, Acta Techn. Hung. 41, 3- 4
.....--=----- -- -~ -- -- -- -~ ~ -

412 DESIGN OF TUl"~EL SECTIONS

by the more accurate polygon method. (This latter was used

in the design of the
Moscow Subway.)
Monolithical sections can be designed with sufficient accuracy
by BouGAYEVA's
method, or where still greate r accuracy is required one should
refer to the B0DR0V-
G0RELIK or to ORLov·s method.
As both the polygo nal and the BonR0V-GORELIK and ORL0V
methods are very
lengthy and cumbersome, their use is justified only in the design
of a long tunnel
with a unifor m cross-section and when the soil can be regard
ed as homogeneou s
along the full length of the tunnel. Under those circumstance
s even a small reduc-
tion in the wall thicknesses or segment dimensions could
result in substa ntial
savings.

4.34. DESIG N OF TUNN ELS WITH DOUB LE LININ

G
Tunne l linings are often divided into separa te rings by the
interlying water-
proofing. If a resistant waterproofing is used, it might be benefic
ial to make use
of the streng th of the waterproofing in the design whenever
possible (see also
Sections 4.622 and 46.26) . In the case of pressure tunnels and
underg round gas-
storage tanks subject to inside pressures a certain percentage
of the load will be
carried by a natura l second layer developed in the surrou nding
ground. Because
most tunnels with double linings are circular, this discussion
will be limited to
circula r ring sections.
The following design metho ds are based on the theory of
elasticity and on
a numbe r of assumptions, such as elastic ground , elastic
pipe and uniformly
distrib uted radial pressures (this last is true in most ~ractical
cases). 4· 8

43.41. Soviet Specifications ( Based on Galerkin's TheoryJu

Aqueducts constructed in solid groun d are usually lined with

concrete or only
plastered with shotcrete. The radial strain in an elastic, isotrop
ic circular ring
embed ded in a homogeneous medium and subject to an inside
pressure p is

..
which can be rewritten as
p
U= -
k

u CABELK A, J. and MENCL , J.: Hydroteclznicke Sto/ne (Hydro

technic al tunnels ) Slovenske
Vydatelstvo Techn. Lit., Bratislava, 1960 272
49
· GALERK IN, B. G.: Naprya zhenno ie sostoia nie cilindritseskoi trubi uprugo i
(Stress conditi ons in cylindrical tubes embedd ed in an elastic srede
mediurr.), Sbornik LIIPSZ 100
 DOUBLE-LINED CIRCULAR SECTIONS
413
where
Ek 1
k= - -- -
r 1 + µk
is called the coefficient of elastic ground resistance in the Soviet Specifications.
For a ring with a radius of 1 rn the value of k becomes k 0 , i.e.

ko = Ek 1
1 1 + µk '

Note that k0 depends only on the elastic characteristics of the ground.

Monolithic concrete linings embedded in solid ground and subject to a uniform
radial pressure p shall have a minimum thickness of

(4.131a)

where rb is the radius to the inside face of the concrete lining, ub,. is the allowable
tension in concrete and
O·O lEb - (1 - µb) k 0
A=
O·OlEb + (I + µb) k 0
where Eb = the modulus of elasticity of the concrete
µb = the Poisson ratio of the concrete.
(Negative v values indicate that the full strength of the concrete section cannot
be utilized and the thickness is· governed by other structural considerations
(waterproofing); the thickness should be not less than 6-8 in.)
The tangential stresses in the concrete are:
on the inside face
r )2
(-/; +A
2
(4.131b)
rk )
( -;:; -A
on the outside face .
(4.131c)

If ther~ is a reinforced shotcrete lining (v, thickJ inside the concrete lining
(Fig. 4/96a) there are 3 cases to be distinguished according to the Soviet Speci-
fications:
k O·Ol E,
1. If -0 > - - -
p (lt
~- ---- -- ~

414 DESIGN OF TUNNEL SECTIONS

\
\
. I
ti,f',__1
~
==:l
P, •tt I
I
I
I

FIG. 4/96. Circular double ring section (shotcrete + concrete) (a) subject to interior pressures;
(b) laminated tunnel lining

where E, = the modulus of elasticity of the shctcrete

u, = the allowable tensile hoop stress in the shotcrete
v, v, (concrete thicknesses) and f v (area of reinforcing steel) can be kept
at a structurally sound minimum because the allowable stresses
cannot be utilized completely.

O·Ol E, k O·Ol E v
2. If - -- > - 0 >---
u, P <lv

then v andfv are governed once again by practical considerations and the thickness
of the shotcrete is
V =0·92r1 (-p - - -k 0 - )
1 u O·OI E, .
and should be between 5 and IO cm.

k0 O·Ol E,
3. If -> - - --
p u.
then the thickness of the concrete lining can be taken at the practical minimum
and the thickness of the shotcrete ring is
p ko \
v, = 0·92 r, (-;; - 0·02 E, } (4.132)

and the area of reinforcement required is

100p k0 )
fv=r, (~ - 0·OOOIEv ·. (4.133)
 DOUBLE· LlN ED CIRCULAR SECTI ONS 415

43.42. Design with Steel Plate Lining ( Design after Muh/hofer)

Both steel lining and concrete ring, with thicknesses of {) and v, respectively,
may be designed on the basis of equal radial deformatio ns with reference to the
interface. A similar relati'onship ,can be established between the concrete ring and
the surroundin g ground. The problem of a steel plate lining and a concrete ring
4 10
embedded in an elastic medium has be(;n solved by MDHLHOFE R • for a variety
of strengths of the componen t layers (concrete, reinforced concrete, shotcrete and
steel plate). The effect of cracks in the concrete, which prevent it from working
in tension, has been also investigated.
Assuming an uncracked concrete section, the hoop stresses in the steel lining
are (Fig. 4/97a)
(4. 134)

where
V- Ev {) [ Q-2
- Eb • -;:; (k~ - I) ; +1

Q = _E_b . _l_ . _m_b_-_l_

C r3 mb

and the radial compressive stresses are:

at the interior face u,0 = - p,
at the exterior face

rr;v= -
1
[ V+I
lT,. P·

In addition to the notations used in Fig. 4/97,

{) the thickness of the steel plate
E 0 and Eb = the moduli of elasticity for steel and concrete, respectively
mb = the Poisson ratio of the concrete
C = the coefficient of elastic subgrade reaction.

The hoop stresses in the concrete at the face in contact with steel are
.
(4.135)

•· 10 MOHLHOFER , L.: Dber die Inanspruchn ahme von Druckstollenauskleidungen, Bau-

ingenieur 1923 8
 416 DESIGN OF TUNNEL SECTION S

and at the face in contac t with earth

, 1 Q- 1
Ub= - - - - -
t V + 1 k1 (k~ - 1) <; +1 ,
where

and the radial stresse s are

1 1 , I I 1
<1 b= - - - - - p and <1 b= - - - - - - - - - - - - - p.
' V + 1 kl ' V + 1 k1 (k~ _ l) _g_ + l
2
Finally the compressive stresses in the surrou nding ground
are

1 1 1
<1k= - - - - - - p; <1,k = 0.
V V +1 k1 (k~ - I) ; + l

Assum ing a cracke d concre te section

the hoop stresses will have to be carried
entirel y by the steel lining:
W v1
u,v= -W+ l 7, P (4.136 )

and the radial stress is <1,0 = - p where

1 1
W = E0 • !._ [~- · - + - - In k 2 ].
'2 C ra Eb
These hoop stresses will be many times
higher than those shared with the con-
crete throug h compo site action .
If a certain amoun t of tension (hoop
stresses) is permissible in the surrou nding
rock, this will also make quite a differ-
ence. Prestressing the concre te lining by
pressu re grouti ng will also have remark -
able effects. In the latter case the hoop
stress in the steel lining is
(I - ). -e)
FIG. 4/97. Stresses in the steel plate lining <1,v =p r, (4.137 )
(1
 DOUBLE-LINED CIRCULAR SECTIONS 417

where i:( %) the portion of the stresses carried by the concrete and the rock
l( %) = the external pressur~ due to prestressing by pressure grouting
expressed as a percentage of the inside pres~ure p.
4 11
i: can be calculated after KNAPP : • .

where R = the radius measured to the borderline between sound and cracked
rock
mk = the
Poisson number of the rock (usually about 6).
R may be calculated from the following expression by trial and error

where <r, = the tensile strength of rock.

At first it seems to be hard to believe that the surrounding rock could take any
tension worth considering. Because of the substantial prestressing, however,
it becomes obvious that, as a result of geostatic pressures, tensile stresses not
exceeding such prestressing could be carried even by materials having no tensile
strength at all. It has been demonstrated by measuring stresses on actual projects
that most of the hoop stresses were carried by the rock and that tension in the
steel lining remained low even if thin plates were used. Larger diameters and
higher pressures will, of course, increase the stresses in the steel lining as well.
For the steel lining an empty duct subject to external hydrostatic pressure
represents a more critical condition. In a buckled section, like that illustrated
in Fig. 4/97b, even the yield stresses can be reached at point A; for a given spacing
of anchor ties the critical external hydrostatic pressure can be calculated from the
formula given by DUBAS :4 ·12
<52
(4.138)
Pko = <r1 2a2 r2

from which expression (cf. Fig. 4/97b)

3

et. =J6pr
<5E
cp

Proc. Inst. Ciu. Eng. July 1955•549

~- 11
,,a JAEGER, C. : Present Trend in the Design of Pressure Tunnels and Shafts, Proc. ln 5 I •
Ciu. Eng. March 1955

2 7 Szechy: The Art of Tunnelling

418 DESIGN OF TUNNF.L SECTIONS

where u1 = the yield stress of the steel

r the radius of lining
8 = the plate thickness.4· 13

43.43 . Design of Tunnels with Laminated Linings

Consider a circular tunnel section composed of rings of different materials and
different thicknesses. Assuming that each ring is uncracked and is supported by
the adjacent one, the stress components for a ring denoted by i and with a radius
r are given as
(a) radial stress
Pi- 1rf-1-P;r 2 P; rl -1 r1
u, = _____ 2 _ _;_ Pi-1 -
(4.139)
rf - r; _ 1 r2 r'f- rl _ 1

(b) hoop stress

Pi-1rf_1-P;r'f + Pi- 1-P; rl - 1r1
u, = (4.140)
rl - rl - 1 r2 rl - rl -1
(c) longitudinal stress
2 2
_ P1- 1 r, -1 - P; r;
u, - 2µi 2 2 (4.141)
r; - r ; -1

At the same time the unknown pressures between the rings, Pi, can be deter-
mined from deformation considerations of the contact surfaces. Narrow, uniform
gaps between the rings (a1_ 1, a;, a 1+1) - due to shrinkage or waterproofing -
can also be taken into account. For a lining consisting of three layers, such as
those illustrated in Fig. 4/96b, the equations to be solved are (i - I = I, i = 2,
i + I = 3)

(4.142) ..

13
' · AMSTUTZ. A. : Das Einbeulen von Schacht- und Stollengrenzungen, Schw. Bauzeitung
1960 9
 DO~LE-UNED CIRCULAR SECTIONS 419

(4.144)

In the last equation Pk may be taken as the initial stress (geostatic pressure) or
as the tensile strength of the rock and rk may be. taken as three times the radius
of the excavated section.: Because 14 and Ek are known (physical properties· cif
the rock material)p 1 andp 2 can be determined from Eqs 4.142 to 4.144 and then
the stresses can be computed for each ring from Eqs 4.139 to 4.141. Experience
bas shown that the gaps (mar.k.ed a) ,average around 0·0015-0·0040 times the
diameter.
Hoop stresses in a single concrete ring completely supported by rock can be
calculated approximately as
Po r1 + ro 1
O', =2 r -+~_E_k__r__+_ r__
- r -
1 0 1 0
1 (4..145)
Eb 2(r 1 ~r0)
Stress measurements carried out in connection with a pressure test on an under-
ground gas-storage tank as shown in Fi?. 4/98 are used to demonstrate the effect

Hoop stresses

1200
I
'
I

I
2000
!'
I
1800

;;--,- 1600
ls
~
~ 1400
.i
@ 1200
~
-.,
!'>
;S 100/J
5
:i:: "'
800

600

400

200

0
F1G. 4/98. Hoopstresses in an 6 8 10 12 14
underground gas tank Interior pressvre (kg/cm1
~ - - - - --

420 DESIGN OF TUNNEL SECTIONS

of gaps on the stresses, particularly on those of the innermost ring. 4·13 As long
as the deformations were smaller than would be required to fill the gaps of inac-
curate construction, the hoop stresses in the steel lining corresponded exactly
with those of an unsupported cylinder (section a). Once the deformations reached
this stage, however, the increases in the stresses gradually slowed down in pro-
portion to the support afforded by the surrounding rock. It is interesting to note
that the maximum stresses occur at the crown, i.e. at the place most difficult to
be grouted, and that the minimum stresses occur at the springing, obviously
because the modulus of elasticity of the rock is lower horizontally than vertically.
It can be concluded, then, that the distribution of stresses is not uniform through-
out the ring section even under uniformly distributed inside radial pressures
(cf. Fig. 3/71).
Cracking in the ring lining will also change the stress pattern. The uncracked
exterior rings will be subject to increases in loads as well as to greater inside pres-
sures (see TALOBRE p. 261). The effect of the composite action of surrounding
fissured rock has been investigated by KovACSHAZY. 4·u

43.44. Design of Laminated Linings for Non-radial (External) Loads

Non-uniform radial and other loads will result in much more complex stress
patterns which are difficult to analyze because of the combination of memlkane
action and bending moments.us
Generally speaking, external loads are not uniformly distributed and thus are
bound to develop bending moments. It is customary to design the ring that is
constructed first (~ually the exterior one) to carry a certain part of the load by
itself and then to consider the support afforded by the first ring in the design of
the subsequent one·, as discussed in Section 43.31, also assuming a certain gap.
With a double lining the loads could also be distributed directly between the two
rings, at least approximately. The hoop forces may be distributed in proportion
to the cross-section areas and the bending moments could be distributed in pro-
portion to the rigidities; the resulting loads will cause compatible deformations.
As an example, the author considered a circular tunnel section with double
lining, subject to uniformly distributed loads, p, at top and bottom. Each ring is
assumed to carry a certain part of the total load and these will also be assumed
to be distributed uniformly and denoted by p 1 and p 2 , respectively. The radii
(measured to the centre of each ring), thicknesses, moduli of elasticity and moments

, i.t 3 K ov.i.csHAzY, F.: K ozetbe telepitett gaztart6k (Gas reservoirs embedded in rock),
Ep. es Kozl. T ud. Kozl. 1960 IV 4
4 • 11 K ov.i.CSHAZY, F.: Berechnung von in Gestei n gebcttcten Druckrohren und Behiiltern,

Acta Tech nica 1957 XVlll 3 - 4

4 •13 CHU-K IA-WANG: Theoretical Analysis of Perforated Shear Webs, J ourn. of Appl. M ech.

1946 A 77-84
 Load on Lhe outside rmg ('? o)

/00% , -
85%
t,
0
cf0% ..-- - - g
r-
':'
60% I I ,If·' +_; ex
flt •1 --+-- - - - - - i
C
z
"'0
'
45:;;-r a- r,4.7, (")

I
40 %.,-- - - - -+-f----+--- r;4J,: ----+-------1
:ii
(")

f,,

""ti p;
v, - 0-2 m
v, - 0·2 m - ----< "'~
v, -0-Z.m
I I I g
z
0 I 15 2 J !!.. "'
~,

Fro. 4/99. Approximate load distribution within a double lining

-
N

•
 422 DESIGN OF TUNNE L SECTIO NS

of inerti a will be denot ed accordingly by , , , ,

1 2 Vi, v2 , E 1 , E_2 and J 1 , 1 2 , respec-
tively (Fig. 4/99). From the above assum ption s it follows that
P 2 + P1 = p. (4.146)
The total load has to be divided in such a mann er as
to cause equal defor matio ns
at the ends of the horiz ontal diame ter as well as
at the ends of the vertical dia-
meter. Now

and if both rings are of the same mater ial

l1 ,: .f ,: 11 }
Pi= P2-. - =exp2
( ICX= - 4 - ,
'1 12 '1 l2
subst itutin g into the equat ion of equili brium

hence
p pcx
P2= - - and Pi=-- (4.147)
l + ex 1 + ex
With the loads deter mined the mome nts are

P ,2 pex '12
M - ___!__2_ - - - - - and
1 - 4 - 4 (1 + ex)
and the stresses due to the eccen tric loads are:
for the inner ring

0"1= -
Ni M1 A '1 M1 6 JJfJ.'1
F1
± -- = -
K1
- -
v1
+ -- = ---
- Vi (l ex) v1 +
±
± __pcx_rf_6__ = _ pcx
__ , 1_( ± 3,1 ) ;
4 (I + ex Vi (1 + ex) v1 1 2v1
outer ring

112 = N2
F2
+ M2
- K2
=
(1
P'2
+ ex) v2
+
-
pr~6
4 (1
= P'2
+ ex) v~ (1 + ex) v2
(l + 2v3r2).
-
(4. 148) •
2
For an initial gap a between the two rings (due to
waterproofing, unfilled voids ,
inacc uracy of const ructio n) the equat ion of defor
matio ns becomes
A2 = A1 +a
and similarly to the above deriv ation

p+q
P2= - - (4.149)
l + ex
 CIJLVJ!RTS AND CONDUlTS 423

where
12EJ
'· q = - -4-a.
Ti

The bending moment!> are

M - (p + q) ari_- qTf
1
- 4(1 + a) 4

and
M.
2
= (p + q)T{ (4.150)
4(l+e<)

and the stresses can be calculated in a similar manner (N1 = p 1 T1 and N 2 = P2 T2)-
Fig. 4/99 indicates the load percentages on the outer ring as a function of the
radius and the ratio of wall thicknesses. It is noted that the radius has little effect
where the wall thickness is the basic 0·20 m. On the other hand, a relative increase
in r: 2 increases P2 rather sharply at first and moderately thereafter. There is hardly
any load ~t all on the inner ring for values of~ exceeding .3. With v 2 < I,
V1 V1
however, the loads carried by the outer ring will gradually disappear. Because
of the lapse of time that is required for earth pressures to develop, the dia-
gram should also give an indication of the relief to the outec ring that will be
offered by the inner ring after its completion.

4.4. DESIGN OF CULVERTS AND CONDUITS

Somewhat different methods have been developed in the design of culverts

and conduits although theoretically the methods already described for under-
ground structures would be applicable. The main differences appear as a result
of the differences in shape and construction of these structures, from the different
longitudinal and transverse loads and also from the greater influence of bedding
conditions. As they are closer to the ground surface, live loads seem to have an
increased effect. These differences are particularly important in the design of
culverts, where the large number of failures has resulted in an extensive research
programme over the last two decades.
The influence of lateral earth support and the flexibility of the structure on
stress distribution in tunnels has already been dealt with. In the case of culverts
and conduits, in addition to these, bedding conditions and the manner in which
the fill or backfill is placed also play an important role. It has been demonstrated
by MEYERHOF that a thoroughly compacted and appropriately chosen backfill
may perform a composite action with a flexible steel culvert. The latter acts,
in fact, merely as a reinforcement for the former.
 424 DESIGN OF TUNNEL SECTIONS

4.41. DESIGN OF CIRCULAR CULVERTS

44.11. Determination of Vertical Earth Pressure (according to Yaroshenko)

According to YAROSHENKo,4•16 underground conduits should be regarded as

structures working together with the mass of soil surrounding it, and each element
of these conduits should be designed in such a way that the forces acting on the
structure be kept to a minimum. To achieve this, the sections should be flexible,
the headwalls should be free from the pressures of the embankment, the bedding
should be flexible rather than rigid and, further, the deformations and separation
at the joints should remain within tolerable limits.
Y AROSHENKO describes the interaction of the culvert and the embankment in the
following way: As a result of the earth pressure on the culvert its crown setties
by an amount of - ,1h due partly to the flexural deflection of the section and
partly to the compression of the supporting soil. Consequently the earth column

3; o,

c P/.lne ofequiv,;,
II 'lent settlement

111

FIG. 4/ 100. Forces acting on

Origma/ ,oosttior. of a culvert
top of cu/rP.rl

above the culvert attempts to settle to an equal degree. However, the movement
of the earth column is opposed by the frictional forces created between the sta-
tionary and moving earth masses. Thus, part of the weight of the earth column
above the culvert will be transmitted to the adjacent soil, and the pressure on
the culvert will be less than the value of the geostatic pressure of yH. On the
other hand, it is also possible that the movement of the culvert is less than the
settlement of the adjacent soil, resulting in a relative upward movement+ ,1h •
of the culvert. In this case exactly the opposite will occur (Fig. 4/ lOOb). Owing
to the downward movement of the adjacent soil mass additional loads will be
transmitted by friction to the earth column above the structure, and the resulting
pressure will be higher than the geostatic one.

4 •16 YAROSHENKO, V. A., ANDREEV, 0. V. and PROKOPOVA, A.G.: Vodopropuskie trubi pod
zheleznodorozhnimi nasupiami (Culverts for discharge of water under railway embankments).
Tranzheldorizdat, Moscow 1952.
 CULVERTS AND COND UITS 425

The frictional forces would only be distribute d along the full height H above
the culvert, if the eartp column above the culvert were incompressible. H owever,
as it settles itself, frictional forces will develop only along that part of the height
along which differential movemen ts take place. This height will be equal to the
height of that earth column which will undergo a t otal settlemen t of Ah. If we
call this height He, then there will be a zo ne above r~, the depth of which is equal
to He within which no relative movemen t between the soil particles and hence
no redistribu tion of the stresses due to friction are occurring. This plane, drawn
at a depth of He, is called the pla ne of equivalen t settlement.
F rom what has been said above, it is .obvious that if the cul verts are rigidly
constructed (stone, concrete, cast-iron, etc.) or arc supporte d on rigid foundatio ns,
the settlemen t of the adjacent soil masses is Jikely to be larger and consequently
the earth pressure on the culvert will also increase. On the other hand, with
flexible structure s and bases the earth pressure will be less than the geostatic
pressure. Experime nts conducte d in the Soviet Union also indicated that the

Embankment crown /eye/

FIG. 4/ 101. The constructi on

of embankme nts around
culverts (after P ECK)

pressures will reach their final val ue in about 4- 5 months. As 4 17

the deflection
at the crown is dependen t on the degree of latera l support, PECK · suggests that
-:onduits through embankm ents should be construc ted as shown in Fig. 4/101.

I. First the layer A 1 is placed , consisting of a uniforml y compacted sandy

gravel, which also serves as a drain.
2. Next, the conduit is construct ed.
3. Then a thorough l y compacted layer (B) is placed on both sides of the con-
duit, for which purpose a clayey sandy soil is the most suitable.
•
4. The material placed on both sides and designate d by A 2 should be of the
same material as the embankment proper and a lso compacted to the same
degree.
5. A loosely compacte d layer (C) is placed directly above the crown.
6. Only after these procedur es should the embankm ent be construct ed.

~-17 TSCHEBOTA RIOFF, G P.: Soil Mechanics and Foundation Structures. McGraw Hill ,
1951 531
426 DESION OF TUNNEL SECITONS

YAROSHENKO determines the magnitude of the pressure increment or decrement

on the conduit by the following reasoning:

First, he assumes that the frictional forces arc directly proportional to the active earth
pressure. Accordingly the frictional force at a depth z is equal to

where ).0 = the coefficient of a.ctivc earth pressure

,p = the angle of internal friction.
ii 0

FIG. 4/ 102. Forces acting on

a rigid culvert
structure

With the above assumption, the earth pressure on a rigid conduit can be obtained in the
following manner (Fig. 4/ 102):
(a) if H e= 0 (see Fig. 4/ 102a)

G =yHB+ 2s,

H
S = yH). tan ,p ,
0
2
where
G = y HB + yH 2 ).0 tan </>, (4.151)

1
G = yB 2 [ -H
B
+ -.-
H ).
B· 0 tan </> ] ,

on introducing the potation

C, = ![+ ; 1 ·;.. tan </> ] , (4.152)

•
G =yB 2 C,;
(b) if He> 0 (sec Fig. 4/ 102b)

G =yHB+ 2S,

H He
S = yH). tan
0 <I> - - - y H, ).a tan </> --;;- ,
2 ~.

G = y HB + yH 2
).0 tan </> - yH~ ).0 tan </> ,
 CULVERTS AND CONDUITS 427

2
H H H' ]
G = y B2 [ B + JP Aa tan ,P - B; Aa tan cf, ,
and if

G = c; y B 2• (4.153)

In the latter case the value of H e has to be determined separately. Its value obtained by
finding the depth of the soil stratum H,, the settlement of which is a result of the additional
loads, will equal the differential settlement between the conduit and the embankment measured
at the crown level.
Let us denote this differential settlement by LJh.
The compression of a soil particle at a depth z is as follows (Fig. 4/103A):

d · O;

FIG. 4/ 103A. The calculation

of the
plane of equivalent
settlement

The load on an element dz from the frictional forces i$:

z'
P = y z)•• tan <f, 2;
2
the unit compression
p
dh = - dz,
EB
where E is the modulus of elasticity of the soil.
Now

dh = yz).atan <f,z' dz2 • ~ = H, . H, •

z' = z -
1EB ' z H ' H '·
y.l. tan <f,z2 H,
dh - - - -- - dz2 ,
2EBH

r~]H =
1f

LJh = Jdh = y).. tan <f,H, fz• dz = y).a tan <f,H, y ).a tan ,f,H,H•
E BH • EBH 3 Z= O . 3£B
,o
(4.154)
~ -~ - - --- - -- - ·-

428 DESIGN OF TUNNEL SECTIONS

introducing the term

Ah3EB
YAa tan r/> '

a= H H 2
' . •
' H ,= H-Hc ,

a = (H - He) H 2 = H3- He H 2 \
from where
Hc = I - ~ .
H H3

To determine the value of H e one has to know the differential settlement Ah between the
embankment and the culvert at the elevation of the crown.
A h can be expressed (see Fig. 4/103A.b) as

LJh = At' + At - AD' - L1D

where Lit'= the settlement of the embankment
Lit = the settlement of the soil below the embankment
L1D' = the deflection of the culvert
AD = the settlement of the soil below the culvert.

lntroducin g the value

.1h
S = --
At' '

Llh = sL1t',

E aD ,
At'=..!!___

where aD = the projection of the culvert above the natural ground.

u=yH ,

A,
t = Ey H aD ,

yHD
Ah= as - - , (4.155)
E

substituting this into the equation for a •

LJh 3ED yHB 3ED 3BDH
a = = as - - - = as (4.156)
J.,.y tan ,j, EJ.a y tan ,j, A0 tan ,j,

In case of circular sections

3D2 H
a =-as - - - .
J.,. tan ,j,

In this expression only s is unknown, for which the followiniz empirical values are used
 CULVERTS AND CONDUITS 429

Rock foundation and rigid structure. . . . . . . . . . . . . . . . . . . . s = 1·O

Rigid structures.founded on dense soil.. ......... ....... s = 0·7
Rigid structures founded on elastic soil . . . . . . . . . . . . . . . . . s = 0·3
Flexible structure on any type of soil .................. : s= O
Knowing a: the values of H,, H e and c; can now be determined. A graphical solution for the
values of as as a functiou of H ,fD was worked out by YAROSHENKO and is given in Fig. 4/103B
In this solution </> was assumed to be equal to 30° (tan <f>),0 = 0·192).

He
7F

"'~
J-0,

~v
--- ~
2-5
I\
\ /
v"
/
2-0
\ V
-
' !·5
.,V
\ V
\ / -~
\ }
\ 0-5
\
0·8 0·4 0 0-4 (}8 1-2 1-6 2-0 2-8
-s fl + S.8
F1G. 4/ 103B. Graphical solution of the plane of equivalent settlement (after YAROSHENKO) ;
</> = 30° and tan</> Aa= 0·192

The final expression for

(4.157)

To determine .the additional loads on culverts, a simplified and approximate •

solution haii been worked out by KLEIN4 · 18

Aq= yt2
[
-
H, t2+ 2 l H
2 )
- - - - - - -- - -
!2
+t
f2
H, + 1
2t2
H + 12 +-2
t2
I ]
' (4.158a)

where t2 represents the depth, measured from the crown elevation of the culvert,
to which the rigidity and thus the compressibility of the soil strata below the

u s Cf. Yarpshenko's work, quoted above.

 - - - - - - - - ----------- --

430 DESIGN OF TUNNEL SECT~S

embankme nt and the culvert is equal and therefore rio , differential movement
between the soil masses will take place. The value of He can be obtained by a
reduction of the upfill height H from PoKROVSZKI¥'s equation:

(4.158b)

where E, and Ea is the modulus of compressibility and y1 and Ya the unit weight
of the soil under the embankme nt and the culvert, respectively.

Fm. 4/ 104. Assumption s of

PRUSKA
Incompressible sub agtr

A somewhat similar assumptio n was made by PRUSKA,4· 18 a when considering

the settlement difference in the embankme nt resp. in the terrain level (..1fl) after
backfill as a measure of the additional load acting upon the extrados of the cul-
vert (Fig/ 4.104). Assuming a constant compression modulus and using the known
equations of elasticity for stress propagatio n in the semi infinite continuum he
concludes that
l y-n·b(Jff - H2)
p = q + Pi = y. H +2 - - -- - H--- -b~2- +-H ~2 (4.159)
H · arc cotg b + b · In b2

is the total maximum vertical pressure acting upon the extrados of the culvert.
In this expression the second member Pt is expressing the additional load - •
the other annotation s being indicated in the figure. This expression holds for all
cases, when the thickness of the upfill is at least four times the height of the cul-
vert (H ~ 4D).
It may be noted, that GUERIN has suggested on the basis of practical observa-
tions that the maximum additional load may not exceed 50 % of the geostatic
value.

'-18.a PRUSKA, M. L.: Pression exercee par un remblai epais sur

une conduite rigidc. Ann .
de l'lnstitut Technique du Bat. et des Tr . Pub/. 1963 mars-avril
 CULVERTS AND CONDUITS 431

The additional weight decreases, however, when the culvert is underlain not by
a rigid but by an also compressible sublayer. Should the additional weight be
found excessive it may be well reduced, when some - very densely compacted- layer
is back.filled at both sides, which would. practically not perform bigge·r deforma-
tions, than the culvert itself.
The loading conditions will be more favourable if ~he culvert is not founded
on the surface of the natural soil but is laid in a trench cut into the natural ground.
There will be no additional loads transferred to the culvert from the weight of
the embankment, as the compressibility of the natural soil will undoubtedly be
smaller than that of the embankment. Therefore not only the weight of the
embankment, but also part of its own weight will be carried by·the trench walls.
In fact, the case will be similar to the loading conditiop.s for sewers and conduits.
If there would be no friction and/or settlement, the weight of the backfill would
be fully carried by the conduit. However, owing to the settlement of the backfill,
frictional forces will develop between the trench walls and the backfill, and part
of the weight of the backfill will be transferred to the soils adjacent to the
trench. The loading conditions on conduits can be expressed as follows:
The equilibrium of a dh thick layer assumes that

G
G + dG+ 2J 0 - tan<5 dh'= G + l Bdh (4.160)
.B

or that the . upward and the downward acting

dn
forces are equal. to each other (Fig. 4/ 105).
By solving this differential equation we obtain

With 2l tan <5 ·

0
K=-.::...._ __
Zr
B
FIG. 4/ 105. Forces acting on a
conduit laid in G
a trench If tant:5 = tan¢ and p = -
B

then
p=. yB (1 _ e-2.!atanef> ! ),
2),,0 tan¢ •
whic~' is simil~r to TERZAGHI's expression for rock pressure. If now

1- e-2J.atanq,!!_
CF= ______B_
2l tan¢ 0

then p = yBCp and

(4.161a)
--=--
-

432 DESIGN OF TUNNEL SECTIONS

The value of CF does not increase beyond the point where the height is equal to,
o r greater than, nine times the width. In three-dimensional analysis the value of
CF remains practically the same if the ratio of height to depth exceeds 3·0. B is
always the width of the trench measured at crown elevation and is increasing the
acting pressure at a progressive rate.

TABLE 4/ Vll .

Values of CF
H/8 Granular cohe-
Item Rati o of earth s ionless fi II (be-
Dry moist
No cover and trench fore compaction gr3nular loam
Wet loam Moist clay Wet clay
width and consolida·
lion)
J..f = 0 ·1924 ).•f •• 0·165
--
J.0 [ - 0· 150 J..f = 0· 130 ;..r- 0·100
I 0·5 0·455 -I 0·461 0·464 0·469 0-474
2 ]·0 0·830 0·852 0·864 0·861 0·898'
3 I ·5 ] ·140 J ·]83 J·208 ]·242 I ·278
4 2·0 ] ·395 ] ·464 ] ·504 1·560 1·618
5 2-5 1·606 l ·702 1 ·764 1 ·838 1 ·923
6 3·0 1·780 ] ·904 1 ·978 2·083 2· 196
7 3·5 1·923 2·075 2· 167 2·298 2·441
8 4 ·0 2·04 1 2-221 2·329 2-487 2·660
9 4 ·5 2·136 2·344 2-469 2·650 2·856
10 5·0 2·219 2·448 2·590 2-798 3·032
11 5·5 2·286 2-537 2·693 2-926 3·190
12 6·0 2·340 2·612 2·782 3·038 3·331
13 6-5 2·386 2-675 2·859 3·] 37 3-458
14 7·0 2·423 2·729 2·925 3·223 3·571
15 7· 5 2-454 2·775 2·982 3-299 3-673
16 8·0 2·479 2·814 3·03 1 3·366 3·764
17 8·5 2·500 2·847 3·073 3-424 3-845
18 9·0 2-518 2·875 3·J09 3-476 3·918
19 9·5 2-532 2·898 3·14 1 3·521 3·983
20 JO·O 2-543 2-918 3·167 3·560 4·042 ~

21 11·0 2·561 2·950 3·210 3-626 4·141

22 12·0 2-573 2·972 3·242 3·676 4-221
23 13-0 2·58 1 2-989 3·266 3-715 4·285
24 14·0 2-587 3·000 3·283 3-745 4·336
25 15·0 2-591 3·009 3·296 3·768 4·378
26 above 2·599 3·030 3·333 3·846 4·545

J.0 = tan 2 (45° - 4,/2), where 4, is the angle of internal friction, f • tan 4, is the coefficient of frict ion.
 ---- -

CULVIRTS AND .CONDUITS 433

A simiral expression was derived by VoELLMY which sounds

H
-2.l.atan~B
2 2
B - B5 B • y 1- e
p = y. 8r cot <p + - 4 , - -10-·tan
- --
<p
- (4.161b)
with
D-n
Bo= - -
4
and
B B-B
H' =H - _ o - - - - 0 cot <p.4:1Bb
2 2

44.12. Determirzation of Traffic Loads

The loads on the conduits from the traffic on the surface or from uniformly
distributed surcharge loads can be determined in the foliowing manner:
For a uniformly distributed load according to TERZAGHI (see Section 32.314):
-2.1. 0 tan~~
G =e B qB = cp qB (4.162)
a nd for concentrated loads :
where k is a dynamic factor, T ABLE 4/ VIII
the value of which is depen-
dent on H. If H = 0 then Values of c,
H
k = 1·5,_a nd if H/B > 1·5 k = B Sand, gravel and
= 1·0. c; is obtained from wet loam
Clay Wet clay

BoussrNE.SQ's equation and can

also be obtained from Table O·0 l ·00 1·00 l·00
4/VIII. 0·5 0·77 0·79 0 ·81
The following expression is 1·0 0·59 0·63 0·66
given by the Soviet building 1·5 0·46 0·61 0 ·54
code to determine the pres-
2 ·0 0·35 0·40 0·44
sure from a concentrated sur-
face load on a conduit con- 2·5 0·27 0·32 0·35
structed in an embankment: 3·0 0·25 0·25 0 ·29
K 4·0 0·12 0·16 0·19
qm = 0·5 H + 1·25 5·0 0·07 0 ·10 0·13
where K = the concentrated 6·0 0·04 0·06 0·08
load (t) 8·0 0·02 0·03 0·04
H = the depth of 10·0 0·0l 0·0l 0-02
cover (m).
4 8
· ' b VOELLMY: Eingebettete Rohre. Mitt. lnsf. Baustatik ETH Zurich No 9. 1937

28 Szcchy: The A rt of Tunnelling

~------ - - - - ~ - - - - - - - - - - - -- - - ~ -

434 DESIGN OP TUNNEL SECTIONS

44.13. Effect of the Bedding

The Scviet experiments mentioned above indicated that the radial and tangen-
tial forces acting on the conduit depend on the flexibility of the conduit, the
physical properties of the soil and, to a large extent, on the method and degree
of compaction of the fill. If, for instance, the pipe is laid directiy on the ground
without shaping the ground to the form of the underside of the pipe, the pressure
distribution on the pipe will be non-uniform and also unfavourable. With circular
sections, for example, at the zone of contact between the conduit and the founda-
tion material the radial forces will be zero and the forces in the vertical axis will
be significantly increased. Whereas the stress distribution for conduits with cradles
will be similar to the case ofs~rip footings, i.e . the maximum values will be observed
under the edges.
The carrying capacity of a pipe increases with the width over which the
base pressures are distributed and also the more uniform these pressures are.
The MSZ 15300 (Hungarian Building Code) shows in its Appendix how the carry-
ing capacity of a circular or egg-shaped pipe section increases under variable
bedding conditions compared with its knife-edge bearing test strength (e.g. for
180° embedment it could increase to 2·5 times of its lowest value).
e
The ratio of the horizontal and vertical pressures ,l = - ranges between 0·2
q
and 0·5 depending on the method of construction, the properties of the soil and
the degree of compaction of the backfill, and ,l could be as high as l ·0 for flexible
pipes. However, this will only: be reached in the final stage, after all vertical and
horizontal deformations have already taken place. Therefore for flexible pipes,
it is not the final loading condition which is the most critical but the one imme-
diately after construction, which eventually leads to the vertical deflection of the
pipe, an increase in its horizontal diameter and to an increase in horizontal
pressure.
According to Y AROSHENKO, the horizontal earth resistance increases linearly with
the increasing size of the horizontal pipe diameter (Fig. 4/106a) :
(4.163)
where C, the coefficient of subgrade reaction, is a function of the soil type. •

FtG. 4/106. Pressures

acting on a conduit
(a) according to
Y AROSHE NKO;
(b) ~ccordi ng to
HA.ROSY
 CULvERTS ANO CON DUITS 435

The change in the diameter size is

r 4 q(I ))
Ad =0·18----
" EJ '
by substitution

and since }, = !.___

q
C
J + - 0·l8r 4
a EJ
),= - - - - - -- (4.164)
c
1 + - - 0·I8r 4
EJ

The resulting pressure distribution is shown in Fig. 4/ 106a, and values for C
and Aa are given on page 438. As for the pressure distribution around tunnels
4 19
HAROSY • suggests the loading scheme shown in Fig. 4/106b which, according
to him, results in moments similar to those caused by the partial distribution of
the vertical pressure. Taking into consideration the void ,1 0 left between the lining
and the soil after construction he arrives at the following equation:
p Q
)q=---C.£10 - - - -
1 +Q l+.Q

EJ
where
Q = 0·068 r4 c
where C is the coefficient of subgrade reaction.
For rigid pipes it is suggested that either RANKINE'S horizontal earth pressure
coefficient ,1,a or the relationship defined by Poisson's ratio _ µ_ = ,1, be used,
1-µ
neglecting the deformation of the pipe.

44.14. Cross-sectional Design of Culvert Sectiof)s

4</>2
Ifweacceptthatthe value of qaround the tube varies as follows: q0 = q(l -,1,)-2-
n
then for both the flexible and rigid sections the following equations can be written.
For the moment and axial compression at the crown: ·
M,., = -
2 0·144r 2q(l - J),

9
u HAROSY, T.: Betonalagutak tervezese (Design of concrete tunnels), MTA Oszt. kozle-
menyek 1958 XXIII

28*
436 DESIGN OF TUNNEL SECTIONS

and

The elongation of the horizontal diameter

4
L1 d =- ,1 d · = O· 18 _Y q (1 - }.)
V 'J EJ

from which the approximate value of the bending stresse~ on the basis of the
equation Mn/2 = 0· 15r 7 q(l - )), wo uld be

O· l5r 2 q(I - Jc)

(4.1 65)
K

(where K denoting the section modulus) to which tht compressive stresses from the
axial load must be added.

- N - l - · (1 - Jc)
a --- ~-----
n F A
rq[ 1 J
(4.166)

From the paper by HA.ROSY already mentioned and according to the approximately
equivalent pressu re distribution diagram shown in Fig. 4/ 1066 the stresses in any
section to the left of the point of discontinuity 17, we obtain

M1 = [(
2
! 0
- I) + cos¢ 0 cos¢ .lq (I - Jc) r 2 ,

N1 = (1 - cos ¢ 0 cos¢) q( l - }.) r + },qr,

T1 = cos ¢ 0 sin ¢ q(l - ).) r,

and to the right of the point of discontinuity

M 11 =[ ! 2 0
- sin ¢ 0 sin ¢lq (1-A.)r 2 ,

Nu = sin ¢ 0 sin ¢ q (I - }.) r + J.qr,

Tn = sin ¢ 0 cos ¢q (I - }.) r.

 CULVERTS AND CONDUITS 437

The above statements and practical -experiments led to the construction of

hinged culverts in the Soviet Union. Hinges were provided on ·the sections at
four points, i.e. at the ~rown, the invert and the springings, by reducing the· cross-
sectional area at these locations. Experiments have shown that the deformation
of the hinged sections took place . more rapidly than that of continuous rigid
sections. This was true only to a certain·point beyond which a further increase
of the load led to a cessation of the deformation of the hinged sections; while
rigid sections showed a continuous deformation, so that the ultimate load on
the flexible sections was twice as large as that on the rigid sections of the same
dimensions and cross-section. The reason for this was that from the very beginning
the lateral support for the hinged sections was larger than that for the rigid ones
and the value of .A. increased from the initial value of O· 18 to almost unity. On the
other hand, the value of .A. was nearly ~qual to O for rigid sections, and increased
only later when as a result of the overstressing, plastic ·hinges developed at the
crown, the invert and the springings.
From this it follows that whenever the deflection of rigid sections is small
(or before plastic hinges can develop) the passive eart~ reshtance is small, _and
the tube does not obtain sufficient lateral support to counteract deformation due
to vertical loads, resulting in considerable bending moments, whereas iri the
hinged tubes plastic hinges are present and acting from the very beginning, and
secure adequate lateral support and reduce the moments in the section. When de-
signing hinged sections, not only the stresses but also the deflections of the struc~
ture have to be checked. The limiting load against buckling was given by LEVI as :

3EJ
qer = ----;::r- ·. (4.167)

At the hinges the stresses in the reinforcement should be at the yield point
and the concrete section should be so proportioned that it can safely carry the
axial loads. The design could be carried out by the following steps:
(1) At the location of the . plastic hinges, the cross-sectional area should be
large enough to transfer the normal forces and, in addition, the section modulus
should be able to provide a safety factor of l ·5 against bending stresses, i.e.
1·5 Muan~ < Kd, where fs is the anticipated yield stress of steel.
(2) · One should investigate whether the reinforcement is sufficient to withstand
the moments \Vhich could be ~reated in the hinges :

(3) The amount of the deformation should be checked by the equation Jd_=
q(.A. - ).a)
for which the values of ..l.0 and C are tabulated below and .A. is
C
taken equal to its maximum value; 0·9.
 438 DESIGN OF TUNNEL SECTIONS

Soil type 1. C (kg/cm')

Sand, sandy clay compact 0·35 2-5

dense 0·25 5- 8

Clayey sand plastic 0·70 2-5

stiff 0·50 5-8
I
Clay plastic 0·75 l 2-5
hard 0·70 5-8

The deformation, as a result of the plastic hinge, will be:

2M a
A dr = lim + I ·2 M hm
- ,2
Ehlh E,.J,e ,

where a =
the length of the hinge in cm
=
the reduced modulus of elasticity of the cross-section
=
the moment of inertia of the cross-section
=
the longitudinal rigidity of the pipe section at the plane of the
hinges.
This deformation should not exceed 1/ 100 of the tube diameter.

44.15. Longitudinal Design of Culverts

Culverts are structures which are likely to undergo differential settlements as

a result of the variable loads acting on the structure and of the non-uniform ity
from stress superpositio n brought about in the underlying strata.
To avoid the developmen t of fissures and fractures, it is customary to construct
,, long culverts from separate and short units, either precast or cast in place. The
functional purpose of the culvert is then secured by providing ·strong and water-
tight joints. As a result of the differential settlements the culvert tries to take the
shape of a trough, thereby causing compressive stresses at the top and tension
at the bottom.
Following TsCHEBOTAR IOFF's suggestions4· 20 and assuming (Fig. 4/ 107) that the
radius of curvature of the longitudinal axis is R the resulting moment according
to NAVIER's hypothesis is
El
M=- - -
R
~-0 °Cf. T SCHEBOTARIOFF, G. P.: Soil Mechanics and Earth Structures, McGraw Hill, 1951531
 CUL VERTS AND CONDUITS 439

.and I
Mm Em
J= - - = - - .
fl
.I
2J 2R I\
: \
i. \\
L2
and if R = SA where A is the maximum I '/?
\
I
EmA \
deflection at the centre, then <T = -~ .
I \
Assuming that the effective inner lever arm i \
is equal to 2/3 of the diameter (m) I ',
M=- -
EJ 2
= - mH
R 3 F:'. ---~
4/ 107. Simplified method for
from where F10 .
the longitudinal design
EJ 12EJ A of culverts
H=- - = ---
Rm L2m ( TSCHEBOTAR!O ff)

is the tensile force on which the longitudinal

design of the section should be based.
RENDULIC expresses the longitudinal tensile force as the sliding resistance evoked
by the weight of the overburden acting as a normal force

1
Hmax = 2 yh2 B tan2 (45° - · </>/2) ,

where h = the height of the embankmen t above the tube

B = the width of the culvert
y = the unit weight of the backfill
</> = the angle of inner friction.

In the Soviet Union the following empirical values are used to determine the
longitudinal force acting on small diameter culverts:

Height of embankment (m)

Diameter of tube 10 15 20
5
(m)
H tensile force (t)

1·00 0·3 1·0 4·0 8·0 13-0

1·25 0·5 1·5 6·0 12·0 20·0
] ·50 0·7 2·0 8·0 J7·0 30·0

If H is known, the longitudinal reinforcement or the connections between the

precast uhits can be designed.
 440 DESIGN OF TUNNEL SECTION S

A simple but accura te method has been worked out by A

. S1K64 • 21 for the
lo ngitudi nal design of cul verts, wlio considers them as beams on
elastic supp o rts.
The same method could a lso be appl ied to condui ts.
From the relation ship establis hed by WIN KLER for elastic founda
tions y = ~ ,
M 2
a nd from the strengt h of materia ls the degree of curvatu re y" = - - - or -d- y =
EJ dx
2
M d2 y
=- - - fro m where M =- EJ -
- . Instead of solving the more compli cat-
EJ dx~
ed differen tial equatio n, one could use the approx imate method
of GOLD and
LEVINE (see Fig. 4/ 108b).

__! _ _

Fm. 4/ 108. Longitu dinal

design of culverts
'°51K6)
DJ

The first change ratios

Lly
LIX
I = _Y;_-LI_X__ an d _LILly_X_ I
x;
Yi-1
x;+ I =
Yi+I - Y;
Llx
fro m where the second change ratios

Lly I - Lly I
LI~ YI =~ x,+ 1 Ll x lx1 = Y;+1 - 2~;_""l_- '_Y,_
- _l
Llx xi L1x 1..r.2
1 1
·~S1K6, A.: Qyakorl ati m6dszer rugalma s a latamasztasi csovek stb.
mereteze sere (Prac-
tical method for the dimensi oning of elastically support ed tubes,
etc.), Viziigyi Kozl. 1958
3-4 203
 CULVERTS AN D CON DUITS 441

Substituting these values into the equation for M

EJ A2 p EJ
M = C Ax2 = C Ax2 CPi+1.- 2p; + A - 1) .

Dividing the total length of the culvert L, into n segments each Ax long, then
Ax= L . At each dividing perpendicular as well as at the ends of the tube con-
n
tact pressures of unknown magnitude will be generated in the soil. However for
each section one can write:
EJ k
M '-=e--Ax2 (P·+ 1 -2p-+p- 1) = M I- ' (4.1{;8)
I l ,-

where Mf is the sum of the moments of the external forces and the unknown
contact pressures Pi, about the section i. For equilibrium the moment calculated
from the difference equation and the moments created by all external forces
(G, weight of embankment + p, soil pressures) must be equal at each section
(Fig. 4/108a)
EJ , Ax .
M ;= CAx2 (P;+1 - 2P; + Pi -1) = LG g - Po
2 (1 Ax - 2/3 Ax) -

- Pt 2Ax (i LJx - -Ax

- ) - Pi Ax [(i- 1) Ax - 2/3 Ax] -
3 2
p;Ax2 Llx
- •••
-p-I - 1 Ax2/3Ax- -2- - 3 .
LJ
(4.169)

A similar equation could be written for any intermediate dividing perpendicular,

thus yielding n - l equations for the solution of n + 1 unknowns. In case of
symmetry the number of unknowns will be reduced and in case of asymmetry
the required two equations can be obtained from statics, namely the sum of ver-
tical forces l:pAx - IG = 0, and the moment of these forces about any con-
venient point is (l:pAx) a - l:Gg = O.
Having thus obtained the pressure ordinatesp;, a catenary polygon with the
forces G and p can be constructed, and the design moments determined graphi-
cally.
Alternatively, the values of P; +1> P;- i can be substituted into the equation
EJ
M '-= - - CAx2
- (P·+1-2P·
' ' +P·, - 11'

and the ordinates obtained from there.

The accuracy of this method can be improved by increasing the number of
divisions.
- - --- ----- - - -- - - --- --

442 DESIGN OF TUNNEL SECTIONS

Table Comparing the Stress for a Circular Conduit Fully, Supported Along its Lcwer Half

I Moment
M e1 - 0 (crown}
Item Loadin11 "-ndition I
'Normal I I II
force I

N Be,:tding

l Dead load IM /gr2

N /gr
+0·3448
-0·1667
+0·2725
0
---
2 Internal water pressure M fyvr3 ·+0·1724 -0·1363
N /y..,.r2 -0·5833 -0·5000
- - - --·
3 External water pressure M fyvr 3 +0·2203 +0·1363
N fyvr 2 +0·5833 +1'5000
4 Excess water pressure in conduit M
-N

5 Uniformly distributed vertical earth M /y' tr2 0·2500 0·2273

pressure N /y'rt 0
I 0·0530
6 Horizontal earth pressure, M /y'Aar" (-0·2500t + 0·0041 7r)
trapezoidal distribution N / y ;,., (t - 0·375r)

- --
7 Uniformly distributed horizontal M /y 0 rt 0·25 0·25

Non:
I earth pressure N/yA0 rt
I
1·00

The above formulae for both bedding conditions were also derived by MARQUARDT, for the
I·OO

case or
panial cmbcdment.

4.42. DESIGN OF CIRCULAR CONDUIT S

The difference in the design of undergrou nd conduits and the culverts discussed
above are mainly due to the different loading, bedding and lateral earth pressure
conditions.
Regardless of whether the conduit is laid under a river-bed, street or structure,
the vertical pressure is always assumed to be equal to the geostatic pressure.
Surface loads do not play a significant role. However, both the external and internal
hydrostati c pressure must be taken into consideration with their full value.
The design procedures of conduits are discussed in detail by E. MARQUARD T4 • 22
who constructed a table, based on the theories presented in Section 43.31 , in which
different loading and bedding conditions are compared. The loading and bedding
conditions on which Table 4/IX is based are shown in Fig. 4/ 109. The table con-
....
M! MARQUARDT ,E.: Rohrleitung en und geschlossen e Kaniile, Handbuch fiir Eisenbetonbau
W. Ernst, 1933, XII 425 -699
 CULVERT S AND CONDUITS 443

= 90°) [MARQU ARDT]

TAs:..:; 4/ IX
Section (!1.'

" - 90° (Springing ) " - 135° (Quarter point) " - 180° (B ottom)
er = 45 (Quarter point)

I
I II I
I II I
I II I
I JI

Bedding Beddina Bed-:H~e

Bedding

+0·0335 +0-0100 -0·3921 I -0·2983 -0·0355 +0-01 00 +0·4406 1 + 0·2725

+0·5554 + 1·5708 + 1·5708 +1-1334 +1·9696 +1·1667 + 2·0000
+0·4375

-0·1964 -0·1492 -0·0168 +0·1363 +0·0050 +0·1363

+0·0168 +0·0050
-0·2146 -0·2146 -0·7868 -0·3687 -1·4147 -0·5000
-0·4277 - 0·3687

+0·0050 -0·1964 -0·1492 +0·0168 +0·0050 +0·1724 +0·1363

-0·0168
-1 ·7854 + 1·7854 + 1·5723 +1·6313 +1·4167 + 1-5000
+ 1·2131 +1·6313

0 1-0·00 72 -0·2500 1 -0·2197 0 0·0141 I

00·2500 0·J967 I I
o
_ _ _ _ 0_·5_
_ 0 ~ __1_·00 o
_~j __ ·
o_ 66
1_ _ _
2 __ _ _ ~_0_·5_8 _36
~ ~ ~ 7_5___ 1_·00 1

(0 - 0·295r) (0·2500t + 0) (0 + 0·0295r) (-0·250 0, -

-0·0417 r)
0 (0·500r + 0·884r) (t + 0·3750,)
(0·5t - 0·0884r)

0·25 0·25 0 0 -0·25 -0·25

0 0
0 0 0·50 0·50 1·00 1·0
0·50 0·50

2) J;
Embedding I

P· -r'
FKJ. 4/ 109. Loadin& and bedding conditio ns according to
MARQUA RDT
 444 DESIGN OF TUNNEL SECilONS

tains only those special cases of bedding conditions I and II when

the pressures
are projected over the entire horizontal diameter (ex' = ; ) . MARQU
ARDT has also
worked out the cases for partial distribution of the bottom reactio
n as well as
the cases of two and three edge suppor ts; however, these problems
are beyond
the scope of this book.

4.43. THE. SEMI-G RAPHI CAL DESIG N OF EGG-SH APED

CULVE RTS

In the previous paragraphs we have discussed in detail the most recent

theories
in the design of rigid and flexible circular sections. However, should
the section
be egg-shaped rather than circular, the analytical solution of the problem
would
be too cumbersome, and therefore it is more practical to design the
statically
indeterminate structure by purely graphical or semi-graphical method
s.
The load q should be determined in the
manner discussed in Section 43.414 and ac-
curately plotted over the section. Next the
section is divided into ele.ments with lengths
Lis and the elastic weights Lis are computed.
EJ
The centre of gravity of the elastic weights
u is determined by drawing a horizontal and
vertical catenary polygon considering the
elastic weights as forces. Having established
the elastic centre, the section is reduced to
a statically determinate structure by cutting
Fm. 4/ 110. Semi-graphical solution it throug h at the crown. Applying a unit
of egg-shaped sections
moment X 1 = ± 1 tm and a unit horizontal
force X2 = ± 1 t the unit displacement factors
a 11 and a2 2 are determined. To expedite the computation, the Lis lengths
and they
distances are scaled from the drawing.
Then a 11 = r Lis 1tm , and a01 represents the displacement of the elastic centre
EJ .
as a result of the external moments; i.e. taking the sum of the produc
ts ~; px
(Fig. 4/ 110).
•
The quotien t- --, or t he r --
Go1 Lis
px ..dx : "'.., - Lis- = X1 expressions
. ·11 . h .
W1 give t e m-
au EJ EJ
ternal moment M 1 as a result of the distributed load p. The displacement
factor
 CULVERTS ANO CONDUITS 445

a 22 is determined similarly, taken as the sum of the moments of the elastic weights
As about the horizontal axis through the elastic centre u . The X 2 = H force is
EJ
002
obtained from the quotient . Owing to the symmetry the vertical internal
0 22
force X3 = V equals zero. Thus, when calculating the internal forces one has to
consider only the moments M = M 0 - X 1 - X 2 y and the axial forces N = N 0 -
-X 2 ct>S a

NUMERICAL EXAMPLE. Given the reinforced concrete sewer pressure main with the
dimensions and sections shown in Fig. 4/ 112, it is required to determine the reinforcement
for a combination of the following loading conditions.

I. dead weight + vertical earth pressure (Fig. 4112a)

Ila. lateral earth pressure (Fig. 4/ 112b)
IIb. the lateral water pressure from the position of the groundwater table at elevation +
+ 6·0 (Fig. 4/ 112c)
III. the internal water pressure at a magnitude of 7·0 t/m 2 (Fig. 4/ 112d).

Owing to the symmetry the calculations are made for one-half of the section only. The
centre line of the half ring is drawn midway between the contours of the sewer section and is
divided into 18 Lis lengths each measuring 50 cm. The distances of the elastic centres from
both the axis of symmetry 9nd' the horizontal line drawn to the extrados at the crown are
scaled from the drawing and are tabulated in Fig.4/ 111. Similarly the depths (v) of the sections
are also scaled.
The calculations, based on the measured dimensions are carried out in tabular form, deter-
mining first the position of a and then the unit displacement factors a 11 and a 2 2• Next, the
displacement due to the external loads for each loading condition is calculated (a01 and a 0 2)-
With the aid of these factors the unknown internal reactions X1 (M) and X 2 (H)can be comput-
ed (see Tables 4/ Xa-f). In the last table (f) all forces acting on the reduced section have been
summarized using the equations

and
N = P sin cc + X 2 cos cc •

The next step is to combine the M and N values for the different loading conditions and
the required cross-sectional area of the reinforcing is then determined for the worst combi-
nation of these cases. When combining the different cases one has to consider the possibility
of such a combination as well. As shown in the example, the maximum value of the moment
is obtained from Case I, and Case III usually gives the maximum axial thrust. The effect of
lateral earth and water pressures is relatively small. In the preceding calculations the lateral
passive earth resistance was not taken into account, but the active case can be considered
by combining Cases I and Ila.
 •

--------------------.. I w------!~
Ground surface ~ + ;,10
.x t
1
ii
f 0·25 0·04 Mc f:50

2 0·74 0·19
J f./6 0·42
4 !·57 0•72
5 1-.95 1·05
6 2·26 1-44
7 2·55 !·84 II.

8 2·845 2·25
g J·OJ 2·71
10 8·05 ,J-20
ff J·Of J.70
111.-·-
12 2·87 4-17
IJ 2·48 4·45
14 NlG' 4·53
f5 149 4·60 +1·18
16 0·§95 4-65 i::
17 0·50 4-69 F 10, 4/ 1I I. The pressure sewer pipe analyzed in the numerical example
 TABLE 4/Xa
Unit Displacement Factors (per unit length: Jm)

Item
Depth of
cross-section
(m)
J ,. -= .
J
J,
ds
cm
-ds T
I
ds
--1
•
Y = Yo - I y' ~y'
'
E~yl
T

I 0·40 0·00534 0·010 0 ·534 50 93·7 0·04 3-75 0·683 0·467 43·75 43 ·75
2 0·41 0·00576 0·010 0·576 50 86·7 0·19 .16·49 0·533 0·285 24·72 68'47
3 0·44 0·00715 0 ·0I0 0·715 50 69·9 0·42 29·38 0·303 0·092 6-43 74·90
4 0·49 0·00984 0·010 0·984 50 50·8 0·72 36·60 0·003 0·00 0·00 74·90
5 0·56 0 ·01470 0·0IO 1·47 50 34·0 ]·05 35·70 - 0·327 0·J07 3-64 78·54 (")
C
6 0·66 0·02410 0·010 2·41 50 20·75 1-44 29·90 -0·717 0·515 10·68 89-22 r
<
7 0·84 0·04975 0·010 4·975 50 10·05 1·84 18·50 -1·117 1·247 12'52 101·74 "';,,
8 J-10 0· 1114 0·0I0 11 ·14 50 4·49 2-25 10·12 -I-527 2'335 10-48 112·22 vi
9 1·45 0·2545 0·0I0 25'45 50 l ·963 2·71 5· 32 -]·987 3·95 7 ·76 119·98 z>
0
10 1·41 0·2342 0·010 23'42 5P 2·135 3·20 6-83 -2·477 6·12 J3·07 133·05 (")
0
11 1-63 0·3620 0·010 36·20 50 J ·38 3'70 5·10 -2'977 8·86 12'22 145·27 z
0
12 2·10 0·7750 0 ·010 77·50 50 0·645 4·17 2·69 -3-447 l 1·93 no 152·97 C

13 1 ·80 0·4875 0·010 48·75 50 1 ·025 4·45 4 ·56 -3·727 14·00 14'37 167·34 a
14 I-50 0·2820 0·010 28·20 50 1·772 4·53 8·02 - 3'807 14·55 25·80 193· ]4
15 J-26 0·1675 0·010 16·75 50 2'983 4·60 J_3 ·72 -3'877 15·02 44·78 237·92
16 1·12 0·l 175 0·010 11·75 50 4·25 4·65 19·77 -3'927 15·44 65·60 303·52
17 1 ·04 0·0941 0·010 9'41 50 ~j-31 4·69 24·90 -3'967 15·76 83'65 387·, 7
18/2 1·00 0·0833 0·0 I0 8·33 25
I:= 394·853
3·00 4·70
I: =
I 284·45
14·10 -3·977 15·82 47·46 4J~·.--.; J

28S·45 ds ds
The po~:tion of point o; Y, = - --- = 0·723 I - = a11 = 2 " 394·853 = 789·706 I - = a,, 2 X 434·63 = .'169·26
394·8S3 t
t
t
-..J
Displacement Factors due lo External Loading TABLE 4/ Xb t
00

I Dead load + Earth load p1 =

(9·50 - 6·58) 1·80Um' + 1·40 X 2·4vm• = 8·60 t/m
P z = (9·50 - 3·88) l ·8011m• + 2·70 X 2·411m' = 16·60 t/m
Considerin g the bottom reaction contact pressure to be uniformly distributed .
-- -
Item <I x p p <I s <Is
X M, -< M ,-
T
EM, -
,1,

T
y M, -
,1,
y
,1 ,
EM, - y
T T

0·0 8·60
1 4·58 0·25 0·29 93·7 27·2
0·495 9·90 27-2 0·683 18·57 18·57
2 4·76 0·74 2·25 86·7 19·5
0·455 11 ·09 46·7 0·523 10·40 28·97 tl
3 4·89 l ·16 6·16 69·9 431 ·0 477'0 ~
0·420 12·20 0·303 130·7 159·67
4 4·96 1·57 12·00 50·8 609·3 1087-3 0
5
0·39 13·21
4·8
0·003 1·84 161'51 z
0·35 1·95 19·31 34·0 656·0 1743·3
6
14·13
4·35 2·26 26·88 20·75 558·0
-0·327 -214·5 - 52·99 .,,
0
0·30 14·92 2301·3 - 0·717 - 400·0 - 452·99
~
7 4·44 2-55 35·22 10·05 354·0
0·29 15·69 2655·3 -1'117 -395·5 - 848-49
8 4· 16 2·845 45·04 4-49 202'2
0·26 16·36 2857·5 -1·527 -309·0 -1157'49 ~
9 1·48 3·03 51 ·98 1·963 102'1 2959·6 r
0·09 16·60 -i·987 -202·6 -1360·09
10

I
12·60 -0·13 3·05 52·74 2·135 112·5
- 0·01 3072-1 - 2-477 -278·2 -1638·29
11 12·60
-0·08 - l ·0l 3·01 5.1 ·50 1'38 71·05
12 12·60 3143·15 -2·977 - 211 ·3 -1849·59
-0·26 -3-28 2-87 46·21 0·645 29·82 3172·97
13 12-60 -3-447 - 102'9 -1952-49
-0·38 -6·05 2·46 32·03 1·02~ 32-84 3205'81
14 12·60 - 3-727 -122'3 -2074·79
-0·49 -6·17 l ·99 18·69 1·772 33·10 3238·91
15 12·60 -3·807 -126·0 -2200·79
-0·49 -6·17 l ·49 7'54 2·983 22'50 3261 ·41
16 12·00 -3-877 -87-2 -2287·99
-0·495 -6·24 0·995 -0·43 4·25 - 1·83 3259·58
17 12·60 -3-927 -t- 7·20 -2280·79
-0·495 -6·24 0·50 -5·21 5·31 -27-65 3231 ·93
18/2 12·60 -3·967 +109·90 -2170·89
- 3·15 0·125 -6·69 3·00 ·-20·07 3211 ·86 -3-977 + 79·90 -2090·99

a~t =2 X +3211 ·86 = 64.23·72; Oo: = 2 X -2090·99 = -41 81·~8

6423 72
M1 = ' = 8·14tm·, - 4181 ·98
H 1 = - --- = -4·81 t
789·706 869·26
 TABl ll 4/Xc
~ Jl a Lateral Earth Pressure
V,
go 1000
c:r The surcharge load is assumed to be 1000 kg/m 2 which is equivalent to the action of an earth column of height h ' = ""Ts= 0 ·60 m.
'< The coefficient of lateral earth pressure is ).3 = 0 ·3.
,,."1 / 1 = (0·60 + 3·12) 1·8 X 0·3 = 2·01 /2 = (0·60 + 9·50 - 1 ·68) 1 ·8 X 0·3 = 4 ·55
" +
~ 2·01 4-55 4·70 2 X 2·01 + 4·55
F= 4·70 - - - - = 15·4t z = - 3- _ -· . . •• = 2·04 m
...
0 2
"1
C
:,
:,
~
~y I F
I
I M, -
,h
I ~.
M, - , - IM, -
~. y IM,~ Y IM, - 6,
y

.
;;· I I I < <
I < ' <

I
()

I 0 2-01 0·204 0·04 0·0 93 ·7 0 0 ·683 0 ?

2 0·JO 2·06 0·423 0·19 0·031 86·7 2'68 2·68 0-523 1 ·40 1·40 ~
~
3 0·30 2·17 0··5 82 0 ·42 0·J75 69·9 12-22 14 90 0·303 3'70 5·10 ;;l
4 0·56 2-31 0 ·792 0·72 0·558 50·8 28·35 43 ·25 0·003 0·09 5·19
5 0·89 2·49 0·930 1 ·05 l ·198 34·0 40·65 83·90 -0-327 - 13-31 - 8·12
zC,>
6 1'25 2·685 l ·088 1·44 2-341 20·75 48·60 132'50 -0:717 - 34'85 - 42·9"7 n
0
7 1-64 2·895 1-ll 1·84 3-95 ]0·05 39·70 172·20 -1'117 - 44·30 - 87'27 zC,
8 2·05 3·12 1 ·357 2·25 6·052 4'49 27-20 199-40 -1 ·527 - 41·50 - 128·77
9
JO
2·47
2·95
3·34
3·6)
1·702
1 ·873
2'71
3-20
9·034
13·0:'il
1 ·963
2·135
17-73
27·83
217 ·13
244·96
-1·987
-2:477
-
-
35'20
68·90
- 163·97
- 232·87
a
11 3-46 3-88 2·008 3'70 18·083 1'38 24·95 269·91 -2·977 - 74·15 - 307·02
12 3·96 4·15 2·13 4·15 23'751 0·645 15-32 285·23 -3·447 - 52-85 - 359·87
13 4·36 4·37 0·617 4·45 27·73 1·025 28-40 313'63 -3-727 -106·0 - 465'87
14 4·50 4·44 0·312 4·53 28·923 1·772 51'22 364·85 -~·807 -195·1 - 660·97
?5 4·57 4 ·48 0·270 4·60 29·967 2-983 89·40 454·25 -3-877 -346·2 -1007'17
I~ 4·63 4·51 0·126 4·65 30·745 4·25 130·70 584·95 -3·927 - 513·0 -1520·17
1'l 4·68 4 ·54 0·090 4·69 31 ·356 5-31 166-30 751 ·25 -3-967 -660·0 -,2180·17
18/2 4·70 4·55 0 4·70 31·471 3·00 94·41 845'66 -3·977 -375·3 -2555-47

a 0 u .1. = 2 X 845·66 = 1691·32; 0011.2. =2X - 2555·48 = - 5110·94

169·132 -5110·94
Mm.=
789.71)6 =2·14tm; H111a = . - = - 5·89 tm.
869 26
£
 ,.... TABLE 4/ Xd "'
V,
C
lib Additional Loading from External Water Pressure
As for loading case Ila the uplift pressure has been neglected; only the excess pressure will be considered po;·osity, n = 27 %;
coefficient of lateral earth pressure, ).0 = 0·3
1 - n = 1- 0·27 = 0·73; V = (6·0 - 1•68) X 1000 (1 - 0·73) 0•3 = 3'38;
Thus , tpe pressure increment for 1 running metre is 0·781 t
- -·
Js Js Js Js Js
Jy V V I M, - t
M -
t
EM, -
t
y M, -
t
y EM, -
t
y

0 Cl
I
0
0 "'f:!l
2 0 Cl
0·38 0 z
3 0·013 0·42 0 69·9 0 0
4
0·56 0·14
0·089 0·72 0·004 50·8 0·203 0·203 0·003 0·001 .,,0
0·89 0·40
5 I ·25 0·68
0·195 1·05 0·037 34·0 1'26 1·463 -0·327 - 0'412 - 0·411 ..;
C
6 0·326 1'44 0·153 20·75 3·18 4·643 -0·717 - 2'28 - 2'691 z
1·64 0·99 z
7 2·05 1 ·30
0·470 1'84 0·402 10·05 4·04 8·683 -1·117 - 4·51 - 7·201 "'r
8 2·47 1 ·65
0·620 2'25 0·851 4·49 3-82 12·503 - 1·527 - 5·83 - 13·031
"'
9
10
2·96 2·02 0·900
1·105
2·71
3'20
1·639
2'918
1·963
2'135
3'22
6·22
15·723
21 ·943
- 1·987
- 2·477
- 6·40
- 15·40
- 19'431
- 34·831 8"'
3-46 2·40
- 54·461 ~
11 1'30 3-70 4·774 1 ·38 6·60 28·543 - 2·977 - 19·63 "'
3-96 2·80 - 70·331
i2 1·18 4·17 7·] 33 0·645 4·60 33·143 -3·447 - 15-87
4·36 3·11 - 104·271
l3 0·443 4'45 8·871 1 ·025 9·10 42·243 -3·727 - 33·94
4·50 3'22
l4 0·227 4·53 9·402 ]·772 16·68 58·923 -3'807 - 63·50 -167'771
4·57 3·27
15 0·198 4·60 9·884 2-983 29·50 88·423 -3'877 -114·20 -281·971
4·63 3·32 - ·452·971
16 0·167 4·65 10·237 4·25 43·50 131·923 -3-927 -171·0
4·68 3·36
l7 0·067 4·69 10·531 5·31 56'0 187·923 -3·967 -222'2 -675·171
4·70 3'38 -126·3 - 801·471
18/2 0 4·70 10·600 3·00 31 ·80 219·723 -3·977

a~ 11•1, = 2 X 219·723 = 439·446 a; u.z. = 2 X - 801 ·471 =- 1602·942;

439'446 -1601·942
Mnb = .
789 706
= 0·556 tm Hm = + 869.26 = - 1·85 t,
 TABLE 4/Xc
~ Jlf. Internal Water Pressure

The normal forces are resolved into horizontal and vertical components (see Fig. 41112d). The values of a a re measured
Ph= P sin oc P, = 1' cos a

,fa' p p r,; sin« cos er P, y'

Ph x'
I t/m'
I I I I
l 0·44 7·0 3·08 12° 0·20791 '0·97815 0·64 3·02 0·235 0·20
2 0·455 7·0 3·18 23°30' 0·38268 0·92388 J-20 2·89 0·39 0 ·645
I
3 0 ·46 7·0 J-22 31 °40' 0·52498 0·85112 1·69 2·74 ()·60 n
1·05
l3
4 0·465 7·0 3·26 40°50' 0·65386 0·75661 2·13 2-47 0 ·89 1·t:-0' ~
5 0·455 7·0 3·18 50°30' 0·77162 0·63608 2-46 2 ·02 1·23 1·73 ~
),
6 0·445 7·0 3·12 60°00' 0·86603 0·50000 2·70 J-56 1 ·61 1·97 z
0
7 0·425 7·0 2-98 68°30' 0·93042 0·3665 2·77 1·09 2 ·00 2·15
8 0·405 7·0 2·84 76°30' 0·97237 0·23345 2-76 0·66 2 ·37 2-27 i
9 0 ·37 7·0 2'59 84°50' 0·9~594 0·09005 2'58 0·23 2·77 2'335 a
10 0·295 7·0 2·06 96°50' 0·99290 -0·11898 2·04 0·245 3·11 2-35
1J 0·245 7 ·0 1·72 114°20' 0·91116 -0·41204 1·57 -0·71 3·34 2-27
12 0·235 7·0 1·64 130°40' 0 ·75851 -0·65166 )·24 -1·07 3-54 2-15
13 0·23 7·0 1·61 147°00' 0·54464 -0·83867 0 ·875 -1·35 3-69 f·98
14 0·29 7·0 2·03 155°25' 0·4160 -0·90935 0·845 ~1·845 3·84 1·67
15 0·38 7·0 2-66 161 °30' 0·31730 -0·94832 0·844 -2·52 4·00 J-28
16 0·42 7·0 2·94 167°35' 0·21502 -0·97661 0·632 -2·875 4·10 0·88
17 0·44 7·0 3-08 173°45' 0·10987 -0·99406 0·338 -3·06
18/2 0·225 7·0 J-58 180°00' 0 -1 ·00000 0 -1-58
II 4·20
4·17 0·46

I ! I0 ~
....
V,
 I
I rM; I ""
Ph y' M; P,
I
x' M: IM·+= M; ~ I - Js
Js
M ,7 y I M --;-y
Js I
IM 7Js Y
V,
N

I I
I m
I tm Im M; t
~s

J 0·64 0·235 0 3·02 0·20 - 0·151 - 0·151 I 93·7 - 14·15 - 14·15 o·683 I- 9·67 - 9·67
2 1·20 0·39 - 0·099 2·89 0·645 - J-343 - 1·442 86·7 -125·00 - 139·15 0·533 - 66·6 - 76·27
3 ] ·69 0·60 - 0·486 2-74 l ·05 - 3·735 - 4·221 69·9 -295·60 - 434·75 0·303 - 89·5 -165·77
4 2·13 0·89 - 1'509 2·47 1·40 - 6·763 - 8·272 50·8 -420·0 - 854·75 0·003 - 1·26 -167·03
5 2·46 1·23 - 3·438 2·02 1·73 -10·437 -13-875 34·0 -471·0 -1325-75 -0·327 + 154·2 - 12·83
6 no 1·61 - 5·75] 1·56 1-97 -13·595 -19·346 20·75 -401·0 -1726·75 -0·717 +288·0 +275 ·17
7 2·77 2·00 -10·729 1·09 2·15 - ]6·231 -26·960 10·05 -271·0 -1997·75 -1·117 +302-4 577'57
8
9
2·76
2'58
2·37
2-77
-15·757
- 22-305
0·66
0·23
2-27
2·335
-18·]15
- 19·186
-33-872
-41·491
4·49
1·963
-152·0
-
-2149·75
81 ·50 -2231·25
-1 ·527
-1·987
+232·0
+ 161 ·9
809·57
917·47
i
z
.,,0
10 2·04 3· I 1 -28·73 6 0·245 2-35 - 19-433 -48·169 2-135 -102'80 -2334·05 -2·477 +254·4 1225-87 -l

II
12
1·57
1·24
3·34
3·54
-33·567
- 38·063
-0·71
-1·07
2-27
2·15
-18·101
-16· 146
-51·668
-54·209
I-38
0·645
-
-
71·40 -2405·45
35·0 -2440·45
- 2·977
-3'447
+212-2
+120·7
1438·07
1558-77
!
r

13 0·875 3·69 -41 ·628 - 1·35 1·98 -13-585 -55·213 1 ·025 - 56·6 -2497·05 -3·727 +2ll·O 1769·9
"'"'
14
15
0·845
0·844
3'84
4·00
-45·328
- 49·415
- 1·824
-2'52
1 ·67
1·28
-
-
9·299
4·610
-54·627
-54·025
1·772
2'983
- 96-9
-161·3
-2593·95
-2755·25
-3'807
-3'877
+369·0
+624·0
2138·77
2762·77
i
"'
16 0·632 4·10 -52·040 -2·875 0·88 - 0·877 -52·917 4'25 -224·6 -2979·85 -3·927 +883 ·0 3645·77
17 0·338 4·17 -53·933 - 3·06 0·46 + I ·870 -52·063 5·31 -276-0 -3255-85 -3·967 +1097-0 4742·77
18/2 0 4·20 -54·774 - 1·58 0·I 15 + 3-086 - 51 ·688 3·00 -155·06 -3410·91 -3'977 +616·0 5358·77
I I I ( I
ao 11r. 1 =2 X - 3410·9 I = - 682 1·82; 00111. J. = 2 X 5358·77 = 10717·54:
- 6821·82 H _ 10717-54 _ .
M111 = ---- = - 8·65 tm · 111 - 869·26 - I 2 33 t.
789·706 '
Tabulation of the Critical Moment and Shear Values In the Critical Sections T ,.BLF. 4/Xf
N= Psinix + H·cosix; H= 4·81 1 ; M= -8·14 ,m,

Cross-section
y
M, -Mi -Hy M p H a s in a
cos« -, ·-· N-

Load I (own weight+ earth)

0 0·683 0 -8·14 + 3-28 - 4·86 0 4·81 oo 0 1·00000 4·81
4 I 0·003 +12·0 -8· 14 + 0·014 + 3'874 19·19 4·81 38°20'
60°00'
0·62024 0·78442 ]5·69
7 -1·117 +35·22 -8·14 - 5·36 +21·12 33 ·27 4·81 0·86603 0·50000 31'06
10 -2·477 +52·74 -8·14 -11·90 +il2'70 38·78 4-81 93°00' 0·99863 0·05234 38·95
14 -3·807 +18·69 -8·14 -18-30 - 7·75 22·27 4·81 171 °00' 0·15643 0·98769 8·24
18 -3-997 - 6·69 -8 -'14 -19·10 -33-93 0 4·81 180°00' 0 1·00000 4·81
Load II/a (lateral earth pressure) (")
C
0 0·683 0 -2·14 4·02 1'75 0 5-89 oo 1·00000 5·89 ~
4 0·003 0·558 -2·14 I- 0·18 -]·43 3'88 - 38°20' 0·78442 3·04
;""
7 -1·117 3-95 -2·14 6·58 ~ 4'90 0·75 - 60°00' 0·500 0·375
10 -2·477 13·051 -2·14 -14·60 - 3'82 -4·18 - 93°00' 0·05234 - 2·00 z>i:,
14 -3'807 28·923 -2·14 -22-44 +4·21 -9·23 - 171 °00' 0·98769 -9·11
18 -3·977 31 ·471 -2·14 -23·40 +s-so -9·71 - 180°00' 1·00000 -9·71 8z
Load II/b (external water) i:,
C
0
4
0·683
0·003
0
0·004
-0·556
-0·556
+1·262
+0·006
0·706
-0·546
0
-
1·85
l ·75
oo
38°20'
1·00000
0·78442
1·85
1·372
a
7 -1·117 0·402 -0·556 - 2·1 62 -2-3 16 - 0·76 60°00' 0·500 0·38
10 -2·477 2-918 -0·556 -4·575 -2·213 - - 1·87 93°0Q' 0·05234 -0·10
14 -3·807 9·402 -0•556 -7·05 +1'796 - -5·02 171°00' 0·98769 -4·96
18 -3-977 10·600 -0·.556 -7-35 +2-694 - I -5·45 180°00' 1·00000 -5·45
Load JIL (internal water)
0 0·683 0 8·65 -8·43 +0·22 -12·33 oo 0 1·00000 -12·33
4
7
0·03
-1·117
- 8·272
-26·96
8·65
8·65
I - 0·04
13-78
+0·338
-4·53
11·12
]5·79
- 6·67
+ 1·26
38°20'
60°00'
0·62024
0·86603
0·78442
0·500
1·66
14·3C
10 - 2'477 -48·169 8·65 30·54 -8·98 16·925 + 8·64 93°00' 0·99863 0·05234 17'37
14 -3·807 -54·627 8·65 47'00 +1·03 12·07 +13·17 171 °00' i 0·15643 0·98769 14·89 .i,.
V,
18 I -3-977 -5 1·688 8·65 49·00 +5·96 2·03 + 15·98 t 80°00' I 0 1·00000 15-98 w
 454

3)
b)
+950
P, -~·50-5·5. l-81/t/m'+ MO. Nllt/m'- 4-60 1/m '
P, - ~50-H~ l-80,2-70.2-4•15@t/m'
p . !!L.!!i.12-50t/m'
z ----~~ ~'.:'"1
M,,,

I
M
8 11
,
1,..,J , 168
I,
/,·11/..:r -----... ,__

Momeni dta!Jram ofthe

static structure
V• h/i-{f-n)J.j -{60-.f.68)(1-0·7J.OJ}· di
·J·J8t/m ' /
if n • 0·27 ,md AJ • OJ Internal water prt>ssure

Fm. 4/112. Loading and moment diagrams

4.44. FLEXIB LE METAL CULVE RTS OF CORRU GATED STEEL

SHEET S
It has already been pointed out, when dealing with the dimensioning of circula
r
tunnels, that an effective lateral earth suppor t may essentially reduce
or even
completely eliminate bending stresses in this type of tunnel. The degree
of lateral
suppor t depends on the stiffness of the passive resistance of the surrqun
ding soil ,
and as the mobilization of this is a function of displacement, the flexibili
ty of the
section is of fundamental importance. The greater the lateral elongat
ion of the
tunnel diameter, the greater the passive earth resistance, i.e. the late·ral
pressure
(see Eq. 4.163 and Fig. 4/101).
This concep t is realized in the design and construction of the "Armc
o''-type
corruga ted steel-plate culverts. (Fig.4/ 112A) Where a uniform soil suppor
t around
the culvert is obtaine d by proper back-fill and compa ction, bendin
g stresses are
small compa red with axial stresses. The maximum radial deflection
of a circular
culvert is
0·167 p:- 4
.dr = - - - - - -
EJ + 0·061 Cr 4 '
CULVERTS AND CONDUTTS 455
456 DESIGN OF TUNNEi. SECTIONS

where p = average vertical pressure at the top cf culvert

r= r adius of culvert
E= modulus of elasticity
J moment of inertia of culvert
C= coefficient of subgrade reaction.

Since the flexural rigidity, EJ, of flexible culverts is generally smali compared
with the second term in the .denomina tor, the former can safely be neglected and,
t hus, the maximum deflection obtained approxima tely is

Field experimen ts on corrugated steel culverts under fills of sa nd, silt and clay
with heights exceeding the c ulvert diameter have sh own that the vertical soil
pressures vary between 50% and 90% of t he overburde n pressure at the
top of the cu lvert. The horizontal soil pressures on such culverts were generally
greater than the vertica l ones and sometimes exceeded the overburd en pressure.
It was-also found tha t the method of backfilling and compactio n as well as foun-
d ation conditions affected the load o n cul verts. In this respect MEYERHOF
then WmTE1 • 23 stated that soft foundation s under the c ulvert, wit h hard
5uppontat ,each side, serve to decrease the load on the structure. Hard foundation s
underneath , with soft support on the sides, tend to increase it and a uniform fo un-
dation viz. support undern_eath a nd at each side subjected to compactio n to such
a n extent that deflection amounts to a minimum tends to make the load on the
structure equal to the height of cover, or the column of material plus superimpo sed
live loads. For heights of fill , exceeding about 3 m (IO ft) under highway loading
and in excess of 9 m (30 ft) under railway loading, the line load is negligible when
taking the dead load equal to the full overburde n pressure. Flexural rigidity of
corrugated steel-sheet culverts is required mainly during constructi on stage, when
it has to resist the forces of handling a nd compactio n of the earth against it,
whereas after completion , co mpressio n resistance is the governi ng factor . Compres-
sion within the ring m ay be d etermined by muitiplyin g its radius by the normal
pressure p 0 , and this product may be kept constant aro und the whole perimeter.
According to this assumptio n, the correspond ing soil p ressure o n the structwre
a t any point may be taken as inve rsely · proportio nal to the radius and may be
given as
p = Por.fr.
Correspon ding soil pressure distributio n figures for flexible culverts are repre-
sented in Fig. 4/ 112B. a nd axial compressio n forces may be obtained, in general.
from the sim ple formula
N = p0 D/2
, .:• 3 WHITE, H. L. : Largest Metal Culvert Designed by
Ring Compressio n Theory, Am
Cii:il. Eng. Jan. 1961 53
 CUJ, VERTS _. ND CONDUITS 457

FIG. 4/ l 12B. Pressure distribution

assumptions around variously shaped corrugated steel plate culverts

F1G . 4/ ! 12C. Calculation of normal force N

N•p•. J}i

A
' '
'
N, ! /1
1: !
,N
FIG. 4/112D. Load assumptions for a flat-bottom culvert N \N
r--, ~
n
..... r:
7H++++H
~
~<
r, P/

as indicated for circular and elliptical sections in Fig. 4/ l 12C and for pipe-arch
se::tions in Fig. 4/ 112D; withp0 denoting the uniformly distributed roof pressure
oc th<> culvert resulting from overburde n pressure and live load on the surface,
A limitation of axial compression stresses is that they cannot exceed the critical
value owing to buckling. With some allowance made for accidental eccentricities
 458 DESIGN OF TUNNEL SECTION S

and imperf ections of culverts in practic e, the critical stress

(o-kr) can be conve-
niently expressed by

where o-r ·- yield stress of culver t materi al or joints

o-k = buckli ng stress of curved -plates· bearin g agains t compa
ct soil to·
be determ ined from elastic theory as

o-k = 2_( CEJ )t

F 1- µ 2

with F denoti ng the unit cross-s ectiona l area of culver t wall,

µ Poisso n's ratio,
beyon d the alread y introd uced annota tions.
As to be seen in this correla tion o-k is indepe ndent of the culver
t radius which is
holdin g, howev er, only for the practic ally impor tant case
when the ratio r/L is
bigger than 2, with L = ( EJ?) )¼ After substit ution we get
(1 - µ- C
<Ti

provid ed that the embed ment of the culver t is of unifor m

stiffness and the height
of fill exceeds its diamet er.
Recen t investigations and experi ments (see MEYERHOF)
have shown that all
plain sheet culverts failed by buckli ng a t about one sixth
of the yield stress of
the materi al, while corrug ated sheets failed generally by
crushi ng at the yield
stress of the steel. Excep t near the critica l section , the bendin
g stresses in the
sheets were small and amoun ted to about one quarte r to
one half of the axial
stresses at failure in the centra l portio n of the sheets. While
the distrib ution of the
soil pressu re on the sheets at failure was simila r to that of
the radial deflections,
the observ ed coefficients of soil reactio n (ratios of soil pressu
re to radial deflection)
va ried consid erably around the sheets (decreasing with
the increa se of radial
deflection).
Accor ding to the experi ments of R. K. WATKI NS 4 · 24 the modul
us of soil reactio n
(M) may be determ ined from the compr ession index or
directly from the triaxia l
compr ession and from this the coefficient of subgra de reactio
n may be obtain ed
from the following formu la:
M M
C= ., ~ --
2(1 - µ,)r l ·5r
where µ, = 0·5, Poisso n's ratio of soil.
•.2, WATKJNs, R. K.: Influence of Soil Charac teristics
on Deform ation of Embed ded Flexibh,
Pipe Culverts, Highwa y Research Board Bull. 223, Nationa l Academ
y of Sciences Washin gtoc
1959
 RECTANGULAR SECTIONS 459

Experiments have also shown that the deformation characteristics (C, M) of

a loose fill may be assumed for one half of those of a compacted fill.
In practical calculations only 50 % of the theoretical culvert stiffness is assumed
with regard to the imperfections of joints.
Based on the beneficial influence of soil support on the statical behaviour of
embedded culverts, MEYERHOF and FISHER4,. 25 describe some examples how the
composite action of underground steel-soil structures can be advantageously
secured and utilized.

4.5. DESIGN OF RECTANGULAR TUNNEL SECTIONS

In the review of the typical tunnel sections (see Section 22.3) it was mentioned
that subway, pedestrian or other utility tunnels, which are generally constructed
under paved roadways, at shallow depths and in open cuts, are mostly rectangular
and are built of reinforced concrete.
This not only makes an economical design possible, but also means that the
available tunnel space can be utilized to the utmost. To speed up the construction
the recent trend is towards the increased use of precast units, resulting in further
reductions in costs which, in turn, explains the growing popularity of rectangular
sections.
The reinforced-concrete sections are designed as closed frames , and their struc-
tural analysis should, in fact, be beyond the scope of the present discussion.
However, because of the special foundation and bedding conditions involved
and for the sake of completeness, in the following paragraphs we shall discuss
the structural design of the two most common cases: rigidly supported narrow
sections, and wide, two bay sections on elastic foundations.

4.51. ONE BAY RECTANGULAR SECTION ON RIGID FOUNDATIONS

For simplicity it is assumed that the stiffness (J) of the roof and base slabs
of the rectangular frame depicted in Fig. 4/ 113 is the same and the stiffness of
the two walls (J0) differs from that value. The elastic weights which play an
important role in the deformation of the sections can be reduced by the factor

-r = J/Jo.

Because of the symmetrical arrangement of the relative stiffnesses, the elastic

centre of the frame a is located at the point of intersection of the axes of symmetry

,.u MEYERHOF, G. G. and FISHER, C. L. : Composite Design of Underground Steel Structures,

Eng. Journ. Canada Sept. 1963
 460 DESIG N OF TUNNEL SECTIONS

BJ A D
J

~
(5
b
1. ~
~
J c_J_
a

FIG. 4/113. Loading and moment diagrams of a single bay rigidly supported closed frame and
its static structure

which, especially with symmetrical loading, greatly simplifies the calculations.

U<;ing the well-known method of analyzing statically indeterminate structures,
the frame is cut at the centre line of the roof slab, and the movement of the elastic·
centre u is investigated while assuming it rigidly connected to the end of the frame.
The movements caused by the external and internal forces should be zero.
In the determination of the moment diagrams for the loading conditions shown
in Fig. 4/ 113, both the unit displacement factors a;; and the movements due to
the loading a0 ; are obtained from the work equations.

I Jm;
a;;= EJ ( ds and a ; = EJ
0
I M m; ds.
0
2 J
Accordingly, the unit displacement coefficients are,
see Fig. 4/ 113c:
2
a 11 = } m 12 -ds= -a 1 2+ab2=2 ( a- +b,
) .
! ! ! I

see Fig. 4/ I 13d:

 RECTANGULAR SECTIONS 461

see Fig. 4/113e :

For simplicity, only one loading condition will be discussed; the case of a
uniformly distributed loa.d p, on the roof slab and an equally uniformly distributed
reaction of the same magnitude on the base slab (Fig. 4/113b). The displacement
factors from the above loading conditions are:
2 2
0 01 = Mm
.,r
0 1-
ds
r
=- (p -a2 I - a 4
S 6r
+ -pa
8
lb 2 ) = - -pa- (-a +
12 r
3b ),

As the areas of the m 2 moment diagram are symmetrical and opposite in sign
their sum will be zero.
Similarly, for the above reasons

0 oa = f Moms~ = 0.

And from the condition of no displacement

a11X1 + oo1 =0

a pa
2
(~+3b)
whence Xi=- -
all
01
=+
24 r~
' r
+ b)
/
.

The moments at the points A, B, C and D can then be determined

MA= Ms= Mc= MD= M 0 + X1 ;

as a clockwise rotating moment, by convention, M 0 has a negative sign.
Thns

pa2
{3-r + 3b]
pa2 paa
.MA - - 8 + a --·- -
I - 12 (a+ br) ·
24 l~ + b)

,,a
The axial forces in the vertical walls will be N0 = ½-· and in the roof and

base slabs N = 0.
• - - - -----~--- - --

462 DESIG N OF TUNNEL SECTIONS

The moment at the centre line of the roof and base slabs will be + M
pa2 paa
= - 8- - 12(a+ br ) and will have a uniform value of MA in the walls.
The second, most common loading condition is the case of lateral earth pres-
sure which, strictly speaking, would give a trapezoidal loading diagram, in which
case neither a 22 nor X 2 would be equal to zero.
However, no serious error is made if one assumes a uniformly distributed pres-
sure diagram based on the average pressure intensity, in which case the analysis
would be similar to the one just discussed.

4.52. TWO BAY BOX SECTION ON ELASTIC FOUNDATIONS

The section shown in Fig.4/114a is the most commonly used section in practice.
The support at the centre line can either be a row of columns or a continuous wall.
In view of the -relatively thin sections
, 6rovnd /ere/ as compared with the overllll dimen-
sions of the structure, the individual
members should be regarded as flexible
and the structure designed as one on
elastic foundations. In the calculations
it is assumed that the structure is sup-
ported on independent elastic sub-
grades.
The elasticity of the soil is repre-
sented by WINKLER's coefficient of
I subgrade reaction according to the
' equation
p=Cy
h where C = the coefficient of subgrade
reaction
p = the loading -
y = the deformation of the soil.
The ideal frame used in the design
FIG . 4/ 114. Elastically embedded two-bay
tunnel structure
is shown in Figure 4/ 114b.
The calculations a·re complicated by
the fact that the base slab is .supported
on elastic foundations. Before discussing the design of the whole frame the case
of a simple beam on elastic supports will be investigated as derived by PASTER-
NAK . 4.26

2
•- • PASTERNAK, P .: Die baustatische Theorie biegefester Balken u. Platten auf elastischer
Bcttung, Belon u. Eisen 1926 9
 RECTANGULAR SECTIONS 463

The principal equation for elasticafly embedded beams is

L 4 d'M .
- - - +M = O (4,170)
4 dx' '
where
L = ' / 4EJ (4.171)
\J bC
and where E = t he modulus of elasticity
J = the moment of inertia
C = the coefficient of subgrade reaction
M = the moment at point x
b = the width of the section.
The ~olution of the differential equation sounds:

M = C1 cos c; cosh c; + C2 sin c; sinh c; + C 3 cos c; sin,h c; + C4 sin c; cos c;,

where
c; = _:_ (4,172)
L

c, are constants, depending on the boundary conditions.

and C 1 , C:, C 3 anq
The shear can be expressed as the derivative of the moment

Q = ..!_ [ - C1 (sin c; cosh c; - cos c; sinh c;) + C2 (sin ( cosh c; + cos c; sinh c;) +
L
+ Ca (cos c; cosh c; - sin c; sinh c;) + c, (cos c; cosh c; + sin c; sinh c;)], (4.173)

and the soil reaction as the second derivative

q = :2 [- C1 sin c; sinh c; + C: cos c; cosh c; - C 3 sin c; cosh c; + c, cos c; sinh c;]. (4,174)

FIG. 4/115. Elastically supported beam loaded

at its end

Forming the third derivative the external rotations are obtained as being proporJional with
the loading q.

" = :3 [ - C1 (sin c; cosh c; + cos c; cosh c;) - C: (sin c; cosh c; - cos c; sinh <!) -

- C 3 (sin c; sinh c; + cos c; cosh c;) + C, (cos c; cosh c; - sin c; sinh c;)]. (4.175)

The different loading conditions can be deduced from the simple case when a single point
load P 1 and a moment Mi ar~ acting at one end of the beam. Therefore, this case will be
investigated first (Fig. 4/ 115).
 464 DESIGN OF TUNNEL SECTIONS

Substitut ing the boundary condition s and the loads of the beam into Eqs
4.172 to 4.175
we obtain: At point O from the condition that M = 0 it follows that
C 1 = 0; and from
the condition that Q = 0 we obtain C3 + C 4 = 0. By introduci ng the
symbols C2 = A 1
and C3 = - C4 = A: Eqs from 4.172 to 4.175 become

M = A 1 sin<sin h<-A 2 (sin < cosh < -cos~si nh <)=A y -A y ~,

1 1 2 (4.176a)

K= :
3
[A 1 (sin ~cosh~- cos~sin h <) -2A 2 cosecosh ~] = : [-A 1 y2 -2A 2 y,].
3

(4.176d)
The value of Ai and A 2 can be determin ed from the boundary condition
s. y 1 , y 2 , y 3 and y4
can be calculated as functions of ~- These values can also be obtained
from Table 4/XI.
The moment at { = 1 should be equal to Mi and si milarly Q = Pi. If 1 is replaced
by ). we can write the following equation :

A 1 sin). sinh). - A 2 (sin). cosh). - cos). sinh i.) = M ;

1

A 1 (sin). cosh). +cos). sinh ).) - 2A: sin). sinh). = LPi.

From the two equation s :

sin ). sinh ). sin ). cosh ). - cos ). sinh ).

Ai = 4 - - - -- - - - M1 - 2 - - - ~ - - - ~ - - LP1 = 4e. M1 - 2e5 LP1 ;
cosh 2). + cos ). - 2 cosh 2.l + cos 2.l - 2
(4. 177)
sin ). cosh ). + cos ). sinh ). sin ). si nh ).
A 2 - 2 -- - - , - - - - ~ - - M - 2 - - - - - -- - - - LP1 = 2 (e6 M 1 -
- cosh 2). + cos 2). - 2 1
cosh 2). + cos 2). - 2 Q4 LP1 •
(4.178)
The rotation and the soil reaction at the two edges of the beam are

at x=O

(4.179)
and at
X =I
4 sinh 2.l + sin 2). . 2 cosh 2.l - cos 2.l.
"1 - - - - -- - - - - Mi - - - - - - - - - - Pi=
- L 3 cosh 2.l. + cos 2i. - 2 L' cosh 2.l. + cos 2.l. - 2

(4.180)
 RECTANGULAR SECTIONS 465

TABLE 4/XI
----- I I

±
Y, 7, !
{ Yi Diff. sin ~ cosh ~ - Diff. .sin ~ cosh { + Dill". Y, Diff.
sin ( sinh ( cos { cosil {
- cos ~ sinh ~ + cos ~ sinh ~

O·OO 0·0000 0·0000 0·0000 1·0000

25 0 1000 0
0·05 0·0025 0·0000 0·1000 1·0000
75 6 1000 0
0·10 0·0100 0·0006 0·2000 1·0000
125 16 1000
0·15 0·0225 0·0022 0·3000 0·9999
175 32 1000 2
0·20 0·0400 0·0054 0·4000 0 ·9997
225 50 1000 4
0·25 0·0625 0·0104 0·5000 0·9993
275 76 998 6
0·30 0·0900 0·0180 0·5998 0·9987
325 106 998 12
0·35 0·1225 0·0286 0·6996 0·9975
375 141 997 18
0·40 0 ·1600 0·0427 0·7993 0 ·9957
-- 424 181 995 25
0·45 0·2024 0·0608 0·8988 0·9932
474 225 991 37
0·50 0·2498 0·0833 0·9979 0·9895
- - - 524 276 988 48
~1 0·60
0·3022
0 ·3595
573
622
0·1109
0·1439
330
390
l ·0967
1·1949
982
974
0·9847
0·9784
63
81
0·65 0·4217 0·1829 l ·2923 0·9703
- - 670 445 965 103
0·70 0 ·4887 0·2284 1 ·3888 0·9600
-- 718 524 954 127
0·75 0·5606 0·2808 l ·4842 0·9473
-- 766 598 940 15 j
0·80 0·6371 0·3406 1'5782 0·9318
812 678 922 18 7
0·85 0·7183 0·4084 1·6704 0·9131
-- 858 761 903 223
0·90 0·8041 0·4845 1 ·7607 0·8908
-- 902 849 877 263
0 ·95 0·8943 0·5694 1 ·8484 0·8645
-- 946 941 851 308
I·OO 0·9889 0 ·6635 1 ·9335 0·8337
987 1038 816 35 7
1·05 1·0876 0·7673 2·0151 0·7980
~
1027 1138 778 I 41 2
1-10 1· l 903 0·8811 2·0929 0·7568
1065 1243 733 471
l · 15 l ·2968 1·0054 2·1662 0·7097
- - 1101 1352 684 53 6
1-20 1·4096 1 ·1406 2-2346 0·6561
1133 1464 626 606
1·25 1·5202 1 ·2870 2·2972 0·5955
1163 1578 562 68 3
1·30 1'6365 1·4448 2'3534 0·5272
1189 1696 490 764
1·35 1·7554 1·6144 2·4024 n.A~I\O
V"'tJVO
-- 1212 1815 409 85 2
]·40 J ·8766 1·7959 2·4423 0·3646
---!---- 1?.30 1938 318 946
( ·4'.i ]·9996 ! ·9897 2·4751 0 ·2710
- - ----~--- - 1243 - - - -- 2062 - - - - - -- 220 1064
1·5() 2· 1239 2·1959 2·4971 0·1664
-- - 1252 2186 110 115 2
I

30 Szcchy : The Art of Tunnelling

~ - - ' - -

466 DESIGN OF TUNNEL SECTIONS

'!'.4.!!!..E 4/ X..! (cont.)

----
{
I Yi Diff.
Y,
sin { cosh { - Diff.
Y,
sin { cosh { +
I
Y,
sin { sinh { Diff. Diff.
- cos { sinh { + cos ( sinh ~ cos { cosh (
I
1·55 2·2491 2·4145 2-508 0·0512
1254 2313 11 1265
1·60 2·3745 2·6458 2·5070 -0·0753
1251 2437 143 1383
1·65 2·4996 2' 8895 2·4927 -0·2136
1240 2561 289 1508
I 70 2-6236 3· 1456 2·4638 -0·3644
1221 2685 445 1640
1-75 2·7457 3·4141 2·4198 -0·5284
1195 2806 606 1776
1·80 2·8652 3·6947 2·3577 -0·7060
-- 1160 2924 800 1920
1·85 2·9812 3·9871 2·2777
-1·90- 3·0927
1115 3037 1001
-0·8980
2069
4·2908 2-1776 - l ·1049
1059 3146 1214 2224
1 ·95 3· l 986 4·6054 2·0562 - .I ·3273
-- - 993 3247 1457 2383
2·00 3-2979 4·9301 1·9115 - I ·5656
914 - - - 3346 1690 - - - - 2549
2·05 3·3893 5·2647 1 ·7425 - I ·8205
824 3431 1955 2718
2·10 3·4717 5·6078 1·5470 -2·0923
719 3509 2225 2891
2·15 3·5436 5·9587 I ·3245 -2·3814
600 3585 2543 3068
2-20 3-6036 6·3162 l ·0702 -2·6882
465 3638 2850 3249
2·25 3·6501 6·6790 0·7852 -3·0131
314 3667 3183 - - -- 3431
2-30 3-6815 7·0457 0·4669 -3·3562
147 3689 3535 3615
2-35 3-6962 7·4146 0·I 134 -3-71 77
- -- - -- - 40 - - -- 3696 3906 3799
2·40 3·6922 7-7842 -0·2772 -4·0976
244 3682 4296 3985
2·45 3·6678 8·4524 -0·7068 -4·4961
469 3646 4702 4167
2'50 3-6209 8·5170 -1·1770 -4·9128
715 3587 5130 4349
2·55 3·5494 8·8757 -1·6900 -5·3477
983 3503 5572 4526
2·60 3·4511 9·2260 -2•2472 -5·8003
1272 - 3390 6034 4698
2·65 3·3239 9·5650 -2·8506 -6·2701
1586 3248 6512 4864
2-50 3· 1653 9·8898 -3·5018 -6·7565
-- 1924 3072 7006 5023
2·75 2·9729 10·1970 -4·2024 -7·2588
2287 2862 7516 5271
2-80 2·7442 10·4832 -4·9540 -7-7759
2676 2614 8040 - - 5308
2-85 2·4766 10·7446 -5·7580 -8·3067
-- 3091 2326 8578 5404
2·90 2·1675 10·9772
3524 1994 - -6·6158
- - - 9126
-8·8471
5568
2·95 1·8141 11·1766
4004 1617 -7-5284 9685
-9·4039
563 0
3·00 l ·4137 l 1 ·3383 -8·4969 -9·9669
 RECTANGULA R. SECTIONS 467
2 cosh 2). - cos 2). 2 sinh 2). - sin 2).
q, =- - - - - - - - , - - - - , - - - M1 +- =
L 2 cosh 2). + cos 2). - 2
- ----,--------,-- - -----,,--- P 1
L cosh 2). + cos 2). - 2

(4.181)

PASTERNAK bas tabulated values for ei , e:, ea, e,, es and ea as functions of J.. With the aid of
these tables (Table 4/ Xll) Eqs 4.177--4.181 can be solved. Ttie tables contain values only
for
the range of ). between 0·5 and 2·75. Namely, if ). < :rt t1'e deformation of the girder
4 is
:rt
negligible and it can be considered a rigid beam; if < ). < :n the beam belongs to the
4
category of siv,rt beams for which Tables 4/XI ·and 4/XII can be used; if ). > :n the beam
can be considered infinitely long.
Returning to the investigation of the so-called short beams, let us consider the case where
the load is applied somewhere around the middle of the beam (Fig. 4/ 116). By cutting through

Fm. 4/ 116. Elastically embedded beam

loaded at its midpoint and the +°
~---.... .;hM
eqµivalent · loading condition
WM, b;

~M

the beam at the line where the load is applied, one obtains two beams, which could be analyzed
in the way discussed before. The value of the imaginary forces Q and M applied at the cut
could be determined from the condition that the displacements and the rotations at
the ends of both beams should be equal.
If the angle of rotation caused by M 1 = 1 tm at point 1 is denoted as au, and that caused
by Pi, as au, and similarly adopting the notations au and au for the diplacements due to
\f1 and P:, respectively, the following equations can be written:

(au+ af1) M - (a:: - af:) Q - ar:P: + af 1 M1 = 0 l (4,182)

(a: 1 - a~ 1 ) M - (ai: + a~:) Q - a~: P: + af: M 1 = 0, J
on solving these equations for M and Q the problem becomes analogous with the case of
beams acted upon by forces applied at one end only.
In the analysis of the frame one has to know the magnitude of the angle change and displace-
ment at the two ends o·· the beam caused by the unit moment (M = 1 tm) and force (P =
= 1 t) (Fig. 4/ 115). Denoting the angle changes by au, au and the1 displacements by au, 1O:z
then from Eqs 4.180 and 4.181

30•
 .:,.
TABLE 4/XJI
°'
,::;:,

,t q, Diff. ·-~ ~ e, Diff. q, Diff. e, Diff. e, DHT.

0·50 24·18661 12·02662 4·00252 2·99592 0·99904 ll ·96809

5·95160 2·07805 0·36315 0·52131 0·09092 58
0·55 18·23501 9·94857 3·63937 2-47461 0·90812 8·98051
4· 12393 1·57786 30190 39682 7743 37
0·60 14·11108 8·37071 3-33747 2·07779 0·83169 6·90614
'.!·94570 1 ·22584 25530 30999 6453 65
0·65 11'16538 7·14487 3·08217 1'76880 0·76716 5·42049
2·15940 0·97125 21855 24575 5635 22
0·70 9·00598 6·17362 2'86362 1'52305 0·71181 4·32827
l ·61671 78152 18894 19846 4822 77
0·75 7'38927 5·39210 2'67468 1·32459 0·66359 3·50750
I ·23318 63750 16492 16255 4235 94 t:)
0·80 6 ·15609 4-75460 2·50976 1 ·16204 0·62124 2'87856 !;J
0·95610
0·85 5·19999 4·22923
52537
2'36457
14519
1·02690
13514
0·58386
3748
2'38806
50 az
75133 44100 12853 11343 3347 4
0·90 4·44866 3-78823 2·23604 0 ·91347 0·55039 1·99792 .,,0
59787 37012 11454 0·09634 3012 02
0·95 3·85079 3·41811 2· 12150 0·81713 0·52027 1·68890
[·00 3'36998
48081
3·10415
31396
2·01891
10259
0·73467
8246
0·49292
2735
I ·43642
48 'z"
c::
39030 26836 0 ·09227 7128 2490 0 zr,,
1·05 2·97968 2·83579 1·92664 0·66339 0·46802 1'22912 t"'
31949 23077 8336 6205 2289 79
l · 10 2·66019 2'60502 1·84328 0·60134 0·44513 l ·05733
V>
26339 19944 7552 5438 2108 "'
l ·I 5 2'39680
21856
2·40558
17320
1·76776
6864
0·54696
4798
0·42405
1952
0·91364
69
1
§z
1'20 2· 17824 2-23238 1·69912 0·49898 0·40453 0·79253 V>
18233 15099 6253 4262 1818 92
1·25 l ·99591 2·08139 1 ·63659 0·45636 0·38635 0·68961
15286 13209 5708 3805 1704 05
1-30 1'84305 1·94930 l ·57951 0 ·41831 0·36931 0·60156
12868 11590 5220 3415 1600 80
1'35 1'71437 l ·83340 l ·52731 0-38416 0·35331 0·52576
10371 10194 4781 3079 1513 67
1·40 i ·60566 I ·73146 1·47950 0·35337 0·33818 0·46009
0·0920~ 0·08983 4381 2790 1433 22
1 ·45 1·51357 1'114163 1·43569 0·32547 0·32385 0 ·40287
8821 7930 4021 2539 1359 006
l ·50 1 ·43536 1 ·56233 1 ·39548 0·30008 0·31026 0·35281
6654 7008 3690 2318 1300 02
I ·55 1·36882 1·49225 l ·35858 0·27690 0·29726 0·20879
5669 6197 3389 2126 1241 8
I-60 !·31213 l ·43028 1-32469 0·25564 0·28485 0·26991
4834 5484 3110 1955 1193 7
 TABLE 4/Xll (cont.)
I

A e,
I Diff. e• Diff. e, Diff. e,
I
Ditf. e,
I
Dilf. II• Diff.

---
1·65 1 ·26379 1·37544 1·29359 0·i3609 0·27292 0·23544
4123 4852 2855 1804 1148 3067
1·70 1 ·22256 1·32692 1-26504 0·2180~ 0·26144 0·20477
3516 4291 2619 1670 007 2743
1"75 1 ·18740 1-28401 1·23885 0·20135 0·25037 0·17734
2997 3794 2401 1549 1071 2440
1·80 I-15743 1·24607 1·21484 0·18586 0·23966 0·15294
2553 3355 2199 1441 1036 2195
1 ·85 1 ·13190 1-21257 1· 19285 0·17145 0·22930 1006 0·13099
2170 2955 2012 1342 1972
1·90 l ·11020 l ·18302 1·17273 0·15803 0·21924 0·11127
1891 2654 1890 1261 0·00986 1779
1·95 1·09129 1·15648 l · 15383 0·14542 0·20938
941
0·09348 ~
1510 2234 1624 1166 1595
2·00 1 ·07619 1·13414 1·13759 0·13376 0·19997 0·07753 ~
1316 2004 1527 1088 925 1440 >
2·05 1 ·06303 0·0l 107 1·11410 1·12232 0·12278 0·01028 0·19072 0·06313 z
901 0·01300 Cl
0·01752 0·01387 C:
2-10 1·05196 1·09658 1·10845 0·l 1250 0·18171 0·05013 t"'
0·00927 1527 1258 966 887 i169 >
0·17294 :io
2·15 1·04269 1·08131 1·09587 0·10284 0·03844
77~ 1326 1138 906 856 0·01059 "'
"'
2-20
2-25
1·03496
1·02855
641
528
1·06805
1·05658
1148
0·00988
l ·08449
1·07423
1026
0·00924
0·09378
0·08528
850
800
0·16438
0·15604
834
813
0·02785
0·01834
951
854
g
z
2-~o 1·02327 1·04669 1·06499 0·07728 0·14791 0·00980 "'
433 846 829 750 792 766
2-35 1·01894 1·03823 1·05670 9·06978 0·13999 771 0·00214
351 722 741 704 685
NO 1·01543
283 1·03101
611
1·04929
661
0·06274 660
0·13228
750
-0·00471
611
2-45 l ·01260 J·02490 1·04268 0·05614 0·12478 -0·01082
225 514 587 619 729 542
2-50 l ·01035 1·01976 l ·03681 0·04995 0·11749 -0·01624
177 429 519 580 708 478
2-55 1·00858 1·01547 ~ ·03162 0·04415 0·l 1041 -0·02102
137 354 459 542 687 420
2·60 1·00721 1·01193 l ·02703 0·03873 0·10354 666 -0·02522
105 290 410 506 365
2·65 1·00616 1·00903 l ·02302 0·03367 0·09688 -0·02887
79 235 351 472 643 366
2·70 1·00537 1·00668 l ·01951 0·02895 0·09045 -0·03203
51 187 305 439 622 268
2·75 1 1·00480 1·00481 1·01646 0·02456 0·08423 -0·03471
...
"'
IQ
 ---- - - - ----- ---- - - --

470 DESIGN OF TUNNEL SECTIONS

'l[
If i. < i.e. the beam is infinitely rigid, the values become
4
12
011 =ya,
1f .i. > n the beam can be regarded as infinitely long. In this cas~ .i. = oo is used in the
calculations, and Eqs 4.172- 4.175 can be rewritten as

C2 = -C4 and cosh ~ - sinh { = e-·~,

(4.183)

(for values of 'I/ see Table 4/XIII).

If the load is applied at the edge of the beam (Fig. 4/ 115) and substituting the boundary
conditions, then

as { =0 and then
f/1 = eOCOS 0° = 1; 17: = e0 sin 0° = 0,

Therefore the value of the two constants is

the rotation at the edge

2 4
K= - P.+ - M1
£2 • £ 3 '
and the displacement

On that basis the rotation and displacement from the unit moment

M1 = 1 tm and unit force P2 = 1

4 2
011 = £3, 012 = - £2,

The calculations will be demonstrated by a numerical example: Determine the

moments in the frame shown in Fig. 4/ 117. The frame is loaded on the top by
a uniformly distributed load and on the sides by trapezoidal earth pressure
diagrams,
 TABLE 4/XHl

{ I ~. I Diff. I ~. r~---= Diff. ~. Diff. I {

I
I -~. ~.

I
- ~- I ,,.

0·0 1·0000 0·0000 1·0000 1·0000 3·6 -0·0245 -0·01209 -0·03659 -0·01241
0·0996 0·0903 0·0093 0· 1900
O·l 0·9004 0·0903 0·9907 0·8100 3'7 -0·0210 -0·01310 -0·03407 -0·00787
980 724 256 1702
0·2 0·8024 0·1627 0·9651 0·6398 3-8 -0·0177 -0·0 1369 -0·03138 -0·00401
946 562 384 1510
0·3 0·7078 0·2189 0·9267 0·4888 3·9 -0·0147 -0·01392 -0·02862 -0·00077
904 412 483 1324
0·4 0·6174 0·2610 0·8784 0·3564 4·0 -0·01197 -0·01386 -0·02583 0·00189
851 298 553 1149 "'
0·5 0·5823 0·2908 0·8231 0·2415 4·1 -0·00955 -0·01356 -0·02309 0·00403 ~
793 191 603 984
~
-i
0·6 0·4530 0·3099 0·7628 0· 1431 4·2 -0·00735 -0·01307 -0·02042 0·00572
732 100 631 832 o
0·7 0·3798 0·3199 0·6997 0·0599 4·3 -0·00545 -0·01243 -0·01787 0·00699 :::
668 23 643 692 >
0·8 0·3130 0·3223 0·6354 -0·0093 4·4 -0·00380 -0·01168 -0·01546 0·00791 "'
602 38 642 564 "'
0·9 0·2528 0·3185 0·5712 -0·0657 4·5 -0·00235 -0·01086 -0·01320 0·00886 /!{
540 89 629 451 ::I
l ·0 0·1988 0·3096 0·5083 -0·1108 4·6 -0·0010 -0·00999 -0·01112 0·00886 ~
478 129 607 349 "'
l'I 0·1510 0·2967 0·4476 -0·1457 4·7 -0·0002 -0·00909 -0·00921 0·00898
418 160 577 259
1'2 0·1092 0·2807 0·3899 -0·1716 4-8 0·0007 -0·00820 -0·00748 0·00892
363 181 544 181
l'3 0·0729 0·2626 0·3355 -0·1897 4·9 0·0009 -0·00732 -0·00593 0·00870
310 196 506 - 114
1·4 0·0419 0·2430 0·2849 -0·2011 5·0 0·0020 -0·00646 -0·00455 -0·00837
261 204 465 57
l ·5 0·0158 0·2226 0·2384 -0·2068 5·1 0·00235 -0·00564 -0·00334 0·00795
217 208 425 9
1-6 -0·0059 0·2018 0·1959 -0·2077 5·2 0·00260 -0·00487 -0·00229 0·00746
177 206 383 30
1·7 -0·0236 0·1812 0·1576 -0·2047 5·3 0·00275 -0·00415 -0·00139 0·00692
140 202 342 62
1·8 -0·0376 0·1610 0·1234 -0·1985 5-4 0·00290 -0·00349 -0·00063 0·00636
108 195 302 86
1-9 -0·0484 0·1415 0·0932 -0·1899 5·5 0·0029 -0·00288 0·00001 0 00578

-•
......

---
 TABLE 4/XIII (cont.) ""N
-..J

~ ~. I Diff.
I ~. Diff. [ ~. Dffi.
7
I ~. Diff. ( ~. ~. I ~. I
I

~.
I I
2·0 -0·0564 80 0· I 231 184 0·0667 265 -0·1794 105 5·6 0·0029 -0·00233 0·00053 0·00520
2-1 -0·0618 54 0·1057 174 0·0439 228 -0·1675 119 5·7 0·0028 -0·00184 0·00095 0·00464
2·2 - 0·0652 34 0·0896 161 I 0·0244 195 -0· 1548 127 5-8 0·0027 -0·00141 0·00127 0·00409
2-3 -0·0668 16 0·0748 148 0·0080 164 -0·1416 132 5·9 0·00255 -0·00102 0·00152 0·00356
2·4 - 0·0669 1 0·0613 135 -0·0056 136 -0·1282 134 6·0 0·0024 - 0·00069 0·00169 0·00307 i::,
2-5 -0·0658 11 0·0491 122 -0·0166 110 -0·1149 133 6·1 0·0022 -0·00041 0·00180 0·00261 ""
;!l
2-6 -0·0636 22 0·0383 108 -0·0254
C)
88 -0·1019 130 6·2 0·0020 -0·00017 0·00185 0·00219 z
2·7 -0·0608 28 0·0287 96 - 0·00320 66 -0·0895 124 6·3 0·00185 0·00003
.,,0
0·00187 0·00181
2-8 - 0·0573 -I
35 0·0204 83 -0·0369 49 -0·0777 118 6·4 c::
0·00165 0·00019 0·00184 0·00146 z
2·9 -0·0535 38 0·01330 71 -0·0403 ~
34 -0·0666 111 6·5 0·00150 0·00032 0·00179 0·00115 I"'
3·0 - 0·0493 42 0·00703 63 -0·04226 20 -0·0563 103 6·6 0·0013 0·00042 0·00172 "'
§z"'
0·00087
3·1 - 0·0450 0·00187 -0·04314 -0·04688 6·7 0·0012 0·00050 0·00162 0·00063
3-2 -0·0407 - 0·00238 -0·04307 - 0·03831 6·8 0·00095 0·00055 0·00152 0·00042
Cl)

3-3 - 0·0364 - 0·00582 -0·04224 -0·03060 6·9 0·0008 0·00058 0·00141 0·00024
3-4 - 0·0322 -0·0853 -0·04079 -0·02374 7·0 0·0007 0·00060 0·00129 0·00009
3·5 - 0·283 - 0·0 1059 - 0·03887 - 0·01769

I
I
I I l I I I I
 RECTANGULAR SECTIONS 473

Fm. 4/ 117. The two-bay, elastically

embedded frame, analyzed
in the numerical example:
(a) its loading diagram;
J -10-42'.J,,
(b) its static structure

5·JJJ,,

b;
/ I ,•
x, x,

J0 = 10-~ m', E = 2·4 x 106 t/m 2, C = 500 t/m 2, b = I m,

6
L= '/4EJ = '/4x2·4x l0 xl8xI0- 3 = 4. 31 m,
V Cb V 500
and thus:
I 4·60
.l = - = - - = I ·0672.
L 4·3I

The bottom slab of the frame belongs to the category of short beams. In view
of the symmetry, the frame can be analyzed as a statically indeterminate structure
with three redundant forces (Fig. 4/117b).
The section can be modified to become a statically determinate structure by
cutting it through at the top centre.

l
The condition of no movement can be expressed by the following equations:

a 11 X 1 + a 12 X 2 +a13 X 3 +a01 =0,

ll21X1 + ll22X2 + ll23X3 + Ooz = 0, (4.184)
ll31 X1 + ll32 X2 + ll33 Xa + lloa = 0,

from which X 1 , X 2 and X 3 can be calculated. When calculating the unit displace-
ment factors, the displacements resulting from the elastic bedding shouid also
be taken into consideration:
 l
474 DESIGN OF TUNNEL SECTIONS

a,k= ~m1mk ds-q 1 ykds,

mkMo
aok = - ~ ds-y0 qkds ,, (4.185)

where the second part of the right hand side of the equation is the product of the
pressure (q) on the bottom slab and the displacement (y) caused by a unit force
acting in the same direction.
First the bottom slab will be considered, regarding it as a beam on an elastic
foundation acted upon by the loading diagram illustrated in Fig. 4/ 118. Taking

r
ttm ® ltm ®
r
I §-20
T f no

I.
@

~ 60 r
: • 4-60 • I
ltm
,,..... (r/) ftm
~
Fm. 4/ 118. Loading cases on the elastically
embedded beam

the centre of the slab as origin, then, because of the symmetrical arrangement of
oadings, the constants C 3 and C4 will be equal to 0. Taking this into account

from thr boundary conditions

These valut's become, in our example:

C1 = 2 X 0·4073 = 0·81466; C2 = 2 X 0·16060 = 0·32120;

M = 0·81466y~ + 0·32120 y 1 ;
 JlECTANOULAll SECTIONS 47$

(the values of y:, y 1, y 3 , y 4 are ~aken from Table ftm

4/XI). Substituting the y values at the eighth
points (Fig. 4/ 118d) we obtain the moment r- ~------------....-----........-.. .
diagram illustrated in Fig. 4/ 119a. The moment a) ~ ®
diagram for the two concentrated unit loads
acting on the two edges can be obtained simi·
iarly. In this case

C1 = -2 X 4·31 X 0·22381 =- l ·92924,

C: =2 X 4·31 X 0·15660 = 1·34989,

M = - l ·92924 y 4 + 1·34989 Yi,
and by substituting the appropriate tabulated FIG. 4/ 119. Moment diagrams due to
y values the moment diagram shown in Fig. unit concentrated loads
4/ 119b can be drawn up. and moments applied at
The influence diagram for the mid section the two edges of the beam
can be plotted as the sum of the moment dia- (PASTERNAK)
grams of a 0·5 t load acting at the end points
of the two half sections and from that resulting
from the end moment (Fig. 4/ 120).
I 4·60 p
1 = T = 4-31 = 1·0612 ; Q = - = 0·5 t;
2

M = _!_ .!'.._ /!z p = ~ 2-1

564
= 1·035 tm.
2 2 1!1 4 2·8698
In general:
from Eq. 4.176a

from Eq. 4.180

from Eq. 4.161

If Q = 0·5 t
thus
A 1 = 2g,LP 1 = -2 X 0·4602 X 4·31 X 0·5 = -1 ·9835
and
A: = -2g4 LP 1 = 2 x 0·6421 X 4·31 X 0·5 = -2·7675
(the values of y 1 and y 2 can be obtained from Table 4/ XI, and Table 4/XII could be used
for e),
M0 = 0
and

M1 = -0·14221 + 0·0360 = -0·1062 tm,

M2 = -0·5653 + 0·2817 = -0·2836 tm,

M 3 = -1-2649 + 0·9440 = -0·320'J.

 476 DESIGN OF TUNNEL SECTIONS

-
I 4·50 P•ft t • ~ro
V
I
Fm. 4/ 120. Diagrams of the moment effect at the
section below the centre column

0-(}5t
) t,(JJ5tm
0-5t
0 2 J 4

0-5t

FIG. 4/1 21. Unit moment

diagrams of the static structure
 RECTANGULAR SECTIONS 477

ln the case of M = 1·035 tm loading P 1 = 0, Therefore

A 1 = 4e,M = 4 X 0·6421 X 1·035 = 2·658;

A 2 = 2e 6 M = 2 X 1-1707 X 1·035 = 2·4231.

Using again Eq. 4.176, M 0 = 0 and M 4 = 1·035

M 1 = +0·19058 - 0·0315 = 0·15908; M 2 = 0·7575 - 0·2467 = 0·5108;

M 3 = I ·6950 - 0·8265 = 0·8685.

Combining the two cases we shall get the influence diagram shown on Fig. 4/120.
Let us calculate now the displacement of the frame, taking the elastic nature of the bottorr.
support into consideration. First the unit displacement factors are calculated for the loads:
X 1 = 1 t, X 2 = 1 tm and X 3 = 1 t, then the displacement due to the uniformly distributed
external load p . The moment diagram for the load X 1 = 1 tis shown on Fig. 4/ 121. The
moment diagram in the bottom slab was calculated by multiplying the influence diagram values
of Fig. 4/119a by the 5·05 tm corner moment. As E is constant and only the value of J is
J
variable, T can be determined from the expression - , and the area of the moment diagrarr.

.
IS
.
calculated from the expression a 11 = _ r~fu-T- :
J.

5·.05 2 X 5·05 5·05 2 + 4·8672 4·867 2 + 4·522

On= - - - - - 2 + - - - - - 1·15 X 2+ - - - - - - 1·15 X 2 +
3x5·33 2xl8 2x18

4·52 2 + 4·2262 4·2262 + 4·1142

+ xl t-15x2+ xl 1·15x2=26·739.
2 8 2 8

Fig. 4/ 12lb illustrates the moment diagram for the X 2 = 1 tm loading case, utilizing the in-
fluence diagram of Fig. 4/ 119a to calculate the moments 1n the bottom slab. From the figure the
unit di~placement factor a = Jm; ds can be directly determined, and by multiplying with
22

the appropriate ordinates of the m 1 diagram, the value of

may be obtained. Wi.t.'lout giving the details:

022 = + 3·233; a 12 = +6·889.

The moment diagram from the load X 3 = It is also given in Fig. 4/ 121. The moment
on the bottom slab is the resultant of the three forces illustrated (see Figs. 4/ 119a-b anc
4/ 120a-b).
Knowing m3 we can calculate the unit displacement factors a i 3 a 23 and C33:

033
J
r
= mids=+ 1.2·462.
......--------- -- -

478 DESIGN OF TUNNEL SECTIONS

FIG. 4/ 122. Moment diagram

of the static structure caused
by the external loading

The moment diagram of the

reduced, statically determinate
frame, from the loading condition
given in Fig. 4/ 117 is illustrated in
Fig. 4/ 122. The moments in the
bottom slab are calculated again
from the combined cases of the
(i 7
y ·fJ-8, 4·85- 68'JJ t 68·JJt
vertical wall reaction and mo-
ments (see Fig. 4/ 122a- b).
0 The loading factors can tnen
be computed from the moment
diagram whose ordinates are the
products of the appropriate mo-
ment diagram ordinates M 0 and
m, using the general relationship

J
M0 m 1
a.,;= - T- ds .

The diagrams, indicating these

products of M 0m1, are illustrated
in Fig. 4/ 123, and from the sum-
mation of the areas the required
values are obtained :

a01 = -1 148·43 ; a02 = -418·72 and a 03 = 911 ·41.

The equations expressing the condition of no displacement :

auX, + a1 2X2 + a1aXa + a o1 = 0

a 21 X 1 + a 22 X 2 + a 23X 3 + a 02 = 0
a a1X1 + a a2X2 + aaaX3 + aoa = 0
26·74X1 + 6·89Xi - 13·63X3 - 1148·43 =0
6'89X1 + 3·23X2 - 5·90X3 - 418·72 = 0
- 13-63X1 - 5·90X2 + 12·46X3 + 911 ·41 =0
From these equations the values of X 1, X 2 and X 3 can be determined

X 1 = +14·38 t; and X 3 = -77·41 t.

With the aid of the known Xi, X 2 and X 3 values the moment and axial force at any point
of the frame can be computed using the expressions
 RECTANGULAR SECTIONS 479

or in the horizontal members

and in the vertical columns.

N= N 0 + X3 •
The final BM values are given in Fig. 4/ 124.

FIG. 4/123. Combined moment

diagram of the static
structure

Oiagr3m lbr M, .m1

-/.'46 - -

-/61{J ~

::::: \% ~

-2f4J5 ~ ~ ~ =
-214'.35 _ -2t4·J5
~ Oi3gr;m lbrM,,.m, i
335-8

FIG. 4/ 124. Moment

diagram of the closed frame
 480 DESIGN OF TUNNEL SECTIONS

4.6. DESIGN DIRECTIV ES FOR ANCILLA RY WORKS,

SERVICES AND INSTALL ATIONS

In this section I wi5h to outline briefly certain aspects and. design details that
are not concerned with the proportion ing of the shape and thickness of the section
and yet are most important from the point of view of constructio n, stability,
durability and maintenan ce.
To cover all the deiails of the above ancillary works would exceed the scope of
this section and for that reason the following discussion will be limited to their
applicatio n in the constructi on of tunnels and to certain special provisions and
specifications.

4.61. COMPOSIT ION AND STRUCTU RE OF TUNNEL WALLS

Tunnels walls can be constructed of iron, steel, concrete, reinforced concrete,

brick or stone. 4 · 27 The loads carried by the walls are typically permanen t in nature,
unlike the live )QadsGarried by-other structures such as bridges which also display
dynamic characteristics. For this reason permanent deformatio ns have to be
considered in both steel and concrete. These can be all the more significallt because
more often than not the walls are constructe d in sections a nd, thus, are not mono-
lithic. The joints between each section undergo certain deformatio ns themselves,
changing in time as a function of the ambient earth pressure. From the point
of view of loads these deformatio ns are beneficial because they help the structure
to adj ust itself to follow the thrust line. However, the same deformatio ns are
undesirable when it comes to clearances and operationa l safety as they may
increase from inches to feet in time (see Section 72.331).
The walls, then, should be designed to have joints that will not become a source
of detrimenta l deformatio ns. The joints in vaults, for example, shall be tight and
radial in direction, with a minimum of mortar (about 1/ 16 in) and/or waterproofing
between the adjacent blocks or units. Concrete, reinforced concrete and stone
blocks shall be chamfered when subject to heavy thrust; bolt connection s shall
be such as to allow re-tightening.
In a design using stone blocks the layout and the arrangeme nt of constructi on
shall be such as to allow the keystone to be placed from the front, unless a special
unit with increasing width toward the centre has been provided to facilitate placing
from the inside. (This is structurall y undesirable, however, as it is liable to drop
out if anything goes wrong with the connection .) Since in the case of radial joints
the keystone cannot be placed from the inside, the resulting gap would have to
be :filled with concrete cast in place. This is a great drawback, particularl y in shield-
driven tunnels, not only because of the delay involved in the concrete setting but

M? DAWSON, J.:TunneJL inings, Trans. Inst. Civ. Eng. Am. 1956-1957 8317-33, 113-122
 WATERPROOFING 48 1

also because of potential deformations. It is similarly undesirable to make no

allowance for lateral pressures in the design of walls.
When designing concrete. or stone blocks, the method of construction would
be borne in mind so that the weight and size of the blocks <!an be handled by the
transporting and placing equipment. Transportation and placing are some of
the most important and difficult problems in organizing the whole job because
of the limited space and illumination. This is why pumping and pneumatic piping
of concrete is gaining increasing- acceptance; they take a minimum amount of
space and deliver the material ·right to the spot of placing (see Fig. 7/J 6). With
the above methods concrete can be conveyed for hundreds of yards and can be
lifted as high as 80 ft. The.re are the,disadvantages, however, that unlike blocks the
poured concrete requires formwork it and cannot carry loads immediately. These
drawbacks can be eliminated easily through the use of sliding forms (Fig. 7/22)
and hardening accelerators. It is worth noting that the Kemano tunnel in Canada,
one of the largest aqueducts in the world, was made with prepact concrete.us
Tunnels, in general, having a linear type layout, particularly lend themselves
to construction with re-usable formwork and scaffolding. To control the placing
and the alignment of steel segments and reinforced-concrete panels, all that is
required is a portal fabricated to the true dimensions and acting as a 'template'.
These portals can consist of dismountable rigid rings, made up of steel rails or
rolled sections and should be spaced to match the width of the panels and/or
the blocks. About 4-5 panels should be braced together to form one unit.

4.62. WATERPROOFING OF TUNNELS

Waterproofing and drainage are the most important supplementary works in

t unnel construction. Good waterproofing is essential not only because of (?pera-
tional considerations, but also because of structural protection and to weathering
of the surrounding rock. While water seepage from the outside will interfere with
the safe operation and usage of the tunnel, water seepage _from the inside -
as in the case of aqueducts - will result in losses and impaired efficiency. At the
same time, water seeping through the wall and the joints is the most common
initial cause of future troubles (see Fig. 7/6).
It follows, then, that waterproofing is not only to stop any groundwater seeping
into a tunnel and conveyed water seeping out of an aqueduct, but also to limit
all surrounding seepage in order to avoid any detrimental washouts and soakings.
This is why any waterproofing has to be supplemented by an elaborate drainage
system.4 · 29

4·!8Cf. Bauingenieur 1959 10

29
•· BoTAR, I.: Vizszigeteles a melyepitesben es osztohezagok vizzdrdsa. (Waterproofing of
engineering structures and of dilatation joints). Miiszaki Kiad6, Budapest 1955. · ·
LUFSKY : K. Bituminiise Bauwerksabdichtung. Teubner Verlag, Leipzig 1951

31 Szecby: The Art of Tunnelling

482 DESIGN OF TUNNEL SECTIONS

Of the two types cf waterproofing, namely interior and exterior, the latter is
definitely better from a structural point of view (Fig. 4/125). The exterior type
bas the "advantage of keeping the water out of all layers of the wall and allows
the wall to carry the hydrostatic pressure with its entire thickness. On the other
hand, this type may be difficult to apply, depending on the method of tunnelling,
and it is certainly difficult to repair (see Section 72.21). In the case of classic tun-
nelling methods the exterior type waterproofing is applied from the overexcavated

Co11ti11uous
dmn.,ge

lfh.ter table

Outside wall (protection

a!Jilinst groond pressure)
Hydrost.lt1c Reinforced-concrete ring
pres.sure (lo carry hydrosutic pressur;
Interior raterproo/1'ng
(1i11permea/J/r)

Fm. 4/125. Interior

and exterior
waterproofing

side, parallel to the walls and with numerable construction joints and under very
unfavourable working conditions; a bonded type exterior waterproofing· could
not even be considered for shield-driven tunnels, because such repairs as might
be necessary would be even more difficult to carry out and would require special
adits (see Section 72.32.). The only ideal exterior type waterproofing for shield-
driven tunnels is. pressure grouting, which can be done from the inside and can
be extended and repaired at any later date; it is, unfortunately not absolutely
reliable.
The interior type waterproofing, on the other hand, can be applied from the
inside and fastened to the inside face of the exterior wall, which, while not capable
of carrying hydrostatic pressures, should certainly be able to carry the earth pres-
sure because it had been completed earlier. The waterproofing on the inside face
 WATERPROOFING 483

of the exterior wall cannot transfer the hydrostatic pressure to the latter, unless
a number of measures have been taken to provide sufficient adhesion or some other
form of anchorage of adequate strength. (For example, :fibreglass-epoxy resin
waterproofing or sheet metal with ribs and anchor rods.) In lieu of such measures
the waterproofing will be unable to carry the hydrostatic pressures and will have
to be supported by a second wall or lining, built inside the first one (Figs 4/49
and 4/125). The outer ring will carry the earth pressures and the inner one will
take the hydrostatic pressures. In the case of impervious units (cast iron, rein-
forced concrete, concrete blocks, etc.) the problem is one of waterproofing the
joints and connections; this will be discussed later in Section 62.23.
The materials for waterproofing are:
1. impervious plaster
2. sheet lead (soldered)
3. bituminous tarpaper or foil
4. thermoplastic materials (PVC)
5. welded steel plate
6. miscellaneous groutings.
(Waterproofing types 2 to 5 can be regarded as flexible and _types 1 and 6 can
be considered brittle.)
In the oldest tunnels waterproofing consisted .of either plain plastering on the
outside ·face, or - in the case of special impermeability requirements - of sheet
lead, protecting the outside face. While the :Single layer of plaster, applied in a
limited working space, soon cracked, weathered and lost its impermeability in
most cases, the application of sheet lead had to be limited to small areas because
of the high cost. Practically speaking, neither of the two types is ever considered
today and so there is no need for a discussion of details.

46.21. Multi-layer Plastering and Shotcrete

Multi-layer plastering and shotcrete are both used as waterproofing today. The
first is used mostly as an interior type waterproofing, requiring careful preparation
of the surface, mixing and application. Plasterings of this type and capable of
resisting hydrostatic pressures up to 2 atm have to be made up and applied as
follows:
(a) Base. The prepared concrete surface is to receive about 1/4 in mortar, applied with
a rounded trowel and then left un~creeded and rough (1 : 4 mix, wa~r- cement ratio. 0·55,
Dmax = 5 mm).
The limits of each layer have to be cut off along a sloping straight line, after setting has
begun. Outside the cut lines the excess material has ·10 ·be scraped off and the surface has
to be cleaned.
(b) A rough layer, about 1/4 in tliick, succeeding the base in about 24 hours, i.e. at a time
when the latter can still be scratched with the handnail. Two hours after placing - initial sei -
it shall be finished smooth with a wooden trowel (mix 1 : 3; water-cement ratio 0·45, Dmax =
= 2·5 mm).
J t•
484 DESIGN OF TUNNEL SECTIONS

(c) Bleeding layer. Also after 20-22 h, appr. 3/ 16 in thick, to be finished smooth
with a
wooden trowel. Mix 1 : 2, w/c = 0·35, Dmox = 0·8-1 ·5 mm.
(d) Cement paste finish, about 1/ 16 in thick, 1 part cement to 1 part of water, to be applied
after 48 h, finished smooth with a wooden trowel and re-finished with a small wooden
trowel
in 8 hours and again in another 3 hours.
Finally another cement paste layer is to be added, about 1/64 to 1/ 32 in thick, and
is to
be finished with a steel trowel after I hour.
The finish may consist of a single layer if the Ce{Tlent paste is applied with a round
brush
directly over layer (c) - just about ready to set still soft - and is finished after 1 hour
with
a steel trowel. If exposed to fresh air frequently, the cement paste layer may be replaced
by
a cement slurry finish of 1 : 1 mix, 5/32 in thick and finished with a rough wooden
trowel
and then with a steel trowel. (The setting times between subsequen t layers can be
reduced
by up to 50% under favourable conditions .) Curing is most important ; regular spraying
shall
be kept up for 14 days. The completed plaster is considered acceptable if the amount
of water
seepage does not exceed 0·5 fl. oz.per ft 2 within 24 h.
In addition to the above preparatio n methods the following rules shall be observed:
1. The cement shall have high tensile strength and low shrinkage characteris tics.
2. The temperatu re of both cement and water shall be around 75-80° F at the time
of
placing.
3. The grading diagram of the sand shall be continuou s
(a) Coarse aggregate (Dmax = 5 mm)
%
0·0 to 0·2 mm grain size 10-15
0·2 to I ·0 mm grain size 25-30
1·0 to 2·5 mm grain size 25-30
2·5 to 5-0 mm grain size 40 - 25
(b) Fine aggregate (Dma. = 2·5 mm) %
0·0 to 0·2 mm grain size 10-15
0·2 to I ·0 mm grain size 25-30
J ·0 to 2·5 mm grain size 65-55
4. Impermea ble plastering can be applied only to consolidat ed structures, already
load-
bearing.
5. Finishing shall be carried out in a space protected from the sun, practically free
from
drafts, with steady temperatu re and humidity conditions (to keep expansion and shrink-
age to a minimum).
6. In order to avoid cracking the difference in temperatu re between concrete wall
and
plaster shall not exceed 35 °F.
7. The ideal temperatu re for plastering is around 50- 70 °F.
8. Plasterirrg on newly constructe d concrete walls shall not be commence d within
8,
12 and 16 days after stripping (70 °F, 50-70 °F and 40-50 °F) and shall not be carried
on under 40 °F.
9. The concrete surface shall be thoroughly cleaned of all dirt (dust , oil, soot etc.)
- in-
cluding any calcium efflorescence - and shall be roughened . The dust shall be cleaned
with water under pressure prior to plastering, so that the wall will not absorb
any
water from the plaster.
10. Honey-com bed areas shall be chiselled out and made good with concrete having
a
higher cement content (200-250 lb/yd 3 more than in the wall). Small areas can be made
good with mortar of 1 : 3 mix.
 WATERPROOFING 485

Here are some of the most common mistakes and errors in plastering work.

The surface is not cleaned properly.

A cement slurry base is used, reducing the adhesion of plaster to concrete.
Dry cement is applied to the wall surface and the properly mixed plaster has water-
surplus to transfer to the cement (haircracks will develop).
The surface is over-rubbed with a steel trowel (the fine cement grains will work themselves
up to the surface and haircracks will develop as a result of shrinkage).
Curing is inadequate.
The work is done by unskilled labourers.

More than three layers of plastering are hardly ever required. In over-thick
plasterwork the different thermal and shrinkage coefficients of the individual
layers made of different mixes have undesirable interactions.

46.22. Gunite Plaster4·:w

This can also be used for interior type waterproofing. Basically, it consists of
plaster shot at the wall by air pressure. There are two methods of mixing and
applying:
1. Sand and very fine gravel (smaller than 5 mm) are mixed with cement in
a dry condition in a pneumatic mixer (see also Fig. 6/114) and then pushed by
air pressure through a rigid pipe system and through a connecting flexible rubber
hose into a spray gun. The spray gun is also connected to a water pressure system
through a regulating valve, controlling the rate of mixing of water with the dry
mix. The amount of added water shall be such as to keep rebound and plaster
losses-to -a minimum.
2. In the other method, also called the "wet method", dry cement and sand
are mixed in the upper drum of a double barrel mixer (Fig. 4/ 126), then mixed
with water in the lower drum and finally enter the spray gun under air pressure
and already in a wet condition. Gunite plastering may consist of one or several
layers. None of the layers shall be too thick, nor shall these be applied at less
_than 8 hour intervals. Prior to applying the succeeding layer, the cement crust of
the previous one shall be blasted with dry sandy material. If the overall thickness
of the plaster is more than 20 mm (3/4") then a wire mesh shall be fastened to
the wall in order to promote adhesion and reduce shrinkage cracks. Horizontal
surfaces do not lend themselves to guniting because most of the plaster will splash
off. Generally speaking, gunite plastering results in high impermeability by virtue
of the compaction effected by the guniting thrust. This impermeability can be
increased further by finishing the top layer, making use of its own moisture.
Because of the approximately 30 % loss on rebound, the remaining plaster will

4 •30 Gunite Handbook. Concrete Proofing Co. Ltd., London 1934

KLEIN, R. : Grundsatzliches tiber Torkretarbeiten im Tunnelbau, Eisenbahntechnische
Rundschau 1952 4'49, Spritzbeton im Tunnelbau, Schw. Bztg. 1956 212
'186 DESIGN OF TUNNEL SECTIONS

have a considerably higher cement

content than that of the original
mix. Since this will result in more
shrinkage, the placing of a wire
Section tl,rouglJ mesh is justified on this account,
t!J concrete gun as well. The elasticity of the gunite
plaster is also increased by the wire
mesh. This is even more desir-
able because of the sensitivity to
subsequent settlements and defor-
mations.
Epoxy fibreglass waterproofing
has recently been introduced for
Mixing 1,ns interior surfaces (see also Section
72.211).
In addition to its original pur-
pose the waterproof gunite ring
also serves as a structural load-
bearing member, particularly in
pressure tunnels (see Section 43.4).

46.23. Bonded Waterproofing

Bonded type waterproofings plac-

ed while hot are more suitable
for interior than for exterior use.
Not only are the working conditions
Fm. 4/126. Section through a concrete gun,
with details
much more favourable in this case,
but also the surface preparation
(cleaning, finishing) and drying.
all of which are essential for sound hot placed waterproofings, can be completed
much more thoroughly and safely. These items are undoubtedly the most important
factors in waterproofing. Tunnel walls with no waterproofing will be moist not
only because of water seeping through from the outside, but also because of the
condensation of water precipitated from the humid atmosphere that is inevitably
present in all tunnels. Bituminous hot-placed waterproofing will adhere only to
dry surfaces. Collector drain grooves shall be provided throughout the entire
area of the walls to collect the seepage water into a subdrain (see also Fig. 7/8).
Before applying the first layer of waterproofing the surface has to be dried with
infrared lamps, gasoline lamps or flame-throwers. Cold bitumen emulsions can
also be applied successfully instead of using the drying process; the hot bitumen
will readily bond to water repellent materials of this type.
 WATERPROOFING 487

It is common practice to use 4 layers of gauge, 120 sheets in the design of water-
proofing. For interior type waterproofings a supporting wall is also required
inside. Once the latter has set, the drain grooves can be grouted either with hot
placed bitumen or with cement plaster and bentorute. Adequate grouting ducts
have to be provided in the supporting wall in any (1lSe. The design, spacing and
placing of the ducts require special care because all joints and connections can
become potential sources of trouble.
For exterior type bonded waterproofing, the drainage, drying, bonding and
finishing of the outside face are a much greater problem. An even finish is partic-
ularly important because the waterproofing is likely to break and fail under the
bending moments caused by uneven support.

46.24. Reinforced Bituminous Waterproofing

Bituminous waterproofings reinforced with foil (aluminum, zinc) have been

used on a number of occasions because of their improved bending strength. The
application of materials of this type did not prove successful in the field of tunnel
construction, however, because of the long term settlements and other movements
that the materials were unable to follow. Also, the l:!ituminous coating has left
a lot to be desired because it has failed to protect the foil material from corrosion.

46.25. PVC Waterproofing

The use of PVC sheets for waterproofing tunnels appears to be much more
promising. The ability of thermoplastic materi:tls (Opanol, Dynogen, Isofol) to
stretch to many times their original length is most helpful in overcoming the
overstresses and stress concentrations caused by settlements and lack of local
support. "lsofol", for example, has a tensile strength of 190-250 kg/cm2 (2700
to 3500 psi) and an ultimate elongation of 280- 340 %.
PVC type waterproofing has been used with good results in Czechoslovakia.4 ·31
A number of admixtures (softeners, stabilizers, lubricants, fillers) have been
added to PVC powder in order to improve impermeability, resistance to heat
and corrosion and also for economic reaso~. The sheets were 1-2 mm (about
1/ 16") thick and have stood up well against corrosion and chemicals. They are
sensitive, however, to gasoline, kerosene, oil, acetone and ether. It is recommended
that prolonged exposure to the sun should be avoided. ·
Because of the lack of bond the structure practically has to be wrapped into
the above waterproofing sheets (Fig. 4/127). Provided that this is possible, the
finishing, drairiing and drying of the surfaces can be dispensed with, which is

'· 31 HAPL, L.: Isolovani stavebnich konstrukci foliami z mekeneho polyvinylchloridu

(Waterproofing of buildings with .softened polyvinylchloride sheets), Inz. Stavbi 1959 4
 488 DESIG :--. or T U1'NEL SECTIO' -1S

Fie. 4/ 127. Plastic waterpr oofing placed on the extrado s

a tremen dous advant age when it comes to exterio r type waterp

roofings. Owing
to their ductility, the sheets can span over surfa~e wrinkles withou
t being damag ed.
Anoth er advan tage over bitumi nous sheets is the fact that
there is no danger of
rotting or loss of ductility (aging). Notwi thstan ding the above,
where good work-
mansh ip is a factor, the sheets are usually glued (hot or cold)
to the wall. This can
be done with hot asphalt, variou s epoxy resins and other specia
l bondin g agents.
Prior to the applic ation of such waterproofing the surface
shall be cleaned thor-
oughly lest the soft material be damag ed by dust or other rough
particles.
The sheets can be spliced both hot and cold . Cold splicing
is similar to rubber
patchi ng; the bonding agent, as a matter of fact, is rubber
cement. Hot splicin g
can be done either by ironing or with high frequency dielect
ric heatin g, melting
togeth er the abutti ng or overla pping sheets at 180-200 °C
(350- 400 °F), withou t
using any foreign materia l. Both factory a nd field splices
have been made in
Czechoslovak ia with very good results particu larly with overla
ppine sheets .u2a

•· 32 • STRAKA , J. : Foldala tti epitmen

yek szigetelesenek helyzete Csehszl ovakiab an (Water -
proofing of underg round structu res in Czechoslovakia), Melyep
tud. Szemle Sept. 1960
 WATERPROOFING 489

46.26. Use of Thermoplastic Sheets

Although the use of thermoplastic waterproofing sheets is promising, there is

no absolute proof of durability because of the limited period of time since their
introduction. The only absolute waterproofing to date that is completely satis-
factory in all respects is the steel plate, welded from inside. However, for economic
reasons, its use can be warranted only in special cases, such as aqueducts, under-
ground gas-reservoirs, engine rooms and electrical control centres. Used exclu-
sively as an interior type waterproofing, the steel lining is unable to carry hydro-
static pressures by itself because of the thin cross-section. It has to be supported
by a reinforced concrete ring, requiring an expensive and time-consuming process.
For this reason it is advisable to use stiffeners on the outside face of the steel
lining in order to take the hydrostatic pressures, or, better still, anchor the stif-
fenersinto the outside concrete wall and take advantage of the resulting composite
action with the steel lining acting as reinforcement.
Stiffeners are always required at the splices of the steel lining. These may con-
sist of rolled steel sections (T, L or [ ) welded to the lining·and embedded in the
concrete (as in Fig. 6/137). Considerable saving can be 2chieveJ with sufficiently
closely spaced stiffeners and anchors by developing coJtposite action. Because
of corrosion considerations, the thickness of the steel plate cannot be less than
6 mm (1/4") in any case; this means an area of steel of As= 50 cm2/m (3·00 in2/ft)
of usable reinforcement. The inevitable gaps between steel lining and concrete
can be pressure-grouted afterwards (cement grout or bitumen); this will provide
structural strength and protection against corrosion at the same time. It should
be realized, of course, that grouting will be rendered more difficult because of the
stiffeners.
Steel lining is particularly important in the waterproofing of pressure tunnels
leading to the turbines of hydroelectric generating plants. The reliability of the
waterproofing and protection against damage are particularly important in this
case to prevent the loss of water. It must be noted that the waterproofing may
be exposed to considerable pressures. It is true that in bearing the internal pres-
sures the waterproofing is supported by the surrounding concrete and rock, but
it is also well to remember.that the distribution and transfer of such pressures will
also set up additional stresses in the lining.
A continuous contact between steel lining and concrete is so important that
many authorities have suggested the omission of all stiffeners and anchors as
these may present serious problems in the making of solid concrete and in pressure
grouting. {The use of expanding cement may be recommended.)
Grouting appears to have a part to play with all types of waterproofing because
of the importance of continuous contact between the structural layers. Grouting,
in itself, can also be used as a type of waterproofing. The structural significance
of grouting in shield-driven and other tunnels will be discussed in Section 63.225,
where mechanical equipment and methods of the first and second application
are also dealt with. As w.ill be seen there, the first, low pressure grouting is aimed
490 DESIGN OF TUNNEL SECTIONS

at filling the voids and the second, high pressure g;-outing is applied for water-
proofing. The most suitable materials for grouting are cement-bentonite suspen-
sions and hot bitumen.
Grouting bas been mentioned here only to complete the list of waterproofing
methods. Reference is made to Section 63.225 for details.

4.63. DRAINAGE OF TUNNELS

The two major components of tunnel drainage systems are the transverse
collector system and the longitudinal gallery. Apart from this, the design will
be basically different for tunnels in mountains where drainage through gravity
is possible and for subaqueous tunnels or those passing under built VP areas
where either intermittent or continuous pumping is required for the removal of
infiltrating water.
As already mentioned (Section 3. 7), tunnels, in general, can also be regarded
as oversize galleries collecting water from the surrounding ground and rock
(Fig. 3/77). As it is almost impossible to divert the seeping water, the problem is
how to keep it out of the tunnel. This can be achieved either by completely imper-
meable waterproofings or by providing an interceptor blanket behind the walls·
which is not susceptible to wash-outs, softening and hoeing, and from where the
water can be drained directly into the longitudinal gallery. This was attell\Pted
earlier by deliberately leaving an interstice between the wall and the surrounding
ground and applying waterproof plaster to the outside face, about 2·5-5 cm
(l "-2") thick and thoroughly finished. The space behind the wall, about 20-30 cm
(8 "-12" ), was then backfilled snugly with dry, hard stone not susceptible to frost.
Thus, the water could not reach the wall at all, for the way was -either blocked
by the finished plaster or the water was intercepted and drained in between the
stones. It was then collected every 8-20 m (25-60 ft) and directed through laterals
(leads) into the longitudinal gallery. However, maintenance problems have proved
this method to be unsatisfactory (cf. Section 72.221). The outside plastering was
of course by no means a perfect waterproofing and while it did retard and localize
the development of flaws in the wall, it did not stop them. The steady stream
of water towards the tunnel, on the other hand, is continually leaching, washing
and weakening the surrounding ground. In mountainous terrain the amount of
seepage water is in proportion to the infiltrating portion of precipitation and is
thus a function of the seasons as well. 4 · 32 -b
The collected water is usually drained through longitudinal .subdrains (also
called drain galleries), which can be located:
I . under the centreline of the road or of the track,
2. next to the wall base.
4 32
· ·b BENDEL: lngcnieurgeologie Vol. II. 521
 DRAINAGE 491

In the first alternative the concrete or stone drain gallery is constructed over tee
inverted arch. This arrangement has one drawback: any repair or cleaning ~f
the gallery necessitates the removal of the pavement and therefore interferes
with continuous operation. In the second alternative the gallery is made part
of the wall. Although easier to clean, it does weaken a critical section. Any crack
in the cantil~ver-type gallery will result in water getting under the wall base,
softening the supporting ground and eventually leading to settlement, tilting of
the walls and cracks in the arch. This design, then, should be adapted only in the
case of water resistant soil and rod: bases where the above problem would never
arise (Fig. 4/128).

Sect1011 b-b
Ject1011 a-a

- --
- -- --
-- - - - - --
- - ~---=--=---

b _j Lateral dram
25x25cm
at wall base at centre
FIG. 4/ 128. Tunnel drainage scheme

In upgrade headings the water in the gallery may be removed by gravity drainage
to the portal or to gathering sumps, in which case it will have to be pumped out
(continually or interrnittc..-ntly).
In case of a heavy inflow of water the drainage problem becomes more compli-
cated and special drainage adits have to be provided parallel to the tunnel (see
also Section 72.333). Frequently, under such drastic measures an underground
pocket which has been tapped will be drained off and exhausted or diminished
to a trickle; in this case the draining ad its become vents drying the tunnel. Such
adits, however, are seldom designed in advance without a thorough geological
survey and are used rather as emergency measures to stem a sudden inrush of
.- --~~~ --- -

492 DESIGN OF TUNNEL SECTIONS

water (Tanna, Mont d'Or, Great Apennine tunnels, etc.) or are constructed
later to facilitate maintenance. Chapter 7 contains a detailed discussion of this
(see also Figs 7/13 and 7/25).
In order to avoid pumping of prohibitive proportions, severe interference with
traffic and a rapid deterioration of the lining, the ground water should be kept out
of subways and subaqueous tunnels completely, using the highest grade of water-
proofing. Although the collector drainage system may be dispensed with in this
case, provision must still be made to drain a certain amount of infiltrating and
cleaning water. The following information is given here as a guide for estimating
purposes:
During the temporary suspension of the construction of the Budapest Subway
the water seepage amounted to 20-25 l/m 2/day {l ·7- 2·0 quarts ft2/day) for concrete
lining (no waterproofing) and 1·3-1 ·5 1/m2/day (0· 10---0· 12 qt/ft 2/day) for cast-iron
lining (first grouting completed); both sections were located in Torton clay and
were 20-30 m (65- 100 ft) below the water table. In another type of clay the
water seepage amounted to 0·9-3·5 l/m2/day {0·1---0·3 qt/ft2/day) and in fissured
chalky clay to 48·5 1/m2/day (4·0 qt/ft 2/day) has been observed. After grouting
these values were reduced to 2- 5 {0· 2-0·4), O· 2---0·6 (0·02---0·05), 0·02---0·5 (0·002---0·05)
and 14·8 1/m2/day (1·2 qt/ft2/day), respectively. Another 50%- 80% reduction
was observed upon completion of the waterproofing of the concrete lining. Inci-
dental water seepage at joints as a result of traffic vibration cannot be avoided
even with the waterproofing completed. The same applies to condensation water,
however good the ventilation may be. StiJI, none of these seepages can be
predicted or estimated. The drainage system of the Budapest Subway has been
designed for the amount of water required to clean the inside face of the lining,
usually done with hoses. This varies between 1·0-1·5 1/m2/week (0·08---0·12
qt/ft2/week), according to data compiled with reference to the Moscow Subway.
In the Budapest Subway the longitudinal gallery is located under the ties which
are embedded in concrete. Subaqueous highway tunnels can have the gallery
either under the centre line or under the curbs. Adequate gradients must be pro-
vided in both cases to drain the water into sumps located at the subway stations
or in the ventilation shafts on the shore, as the C!lse may be (see Figs 6/ 168 and
4/138). Draining very long sections by gravity alone would be impractical because
of the excessive fall in height that would be required. In such cases a number of
sumps must be located at regular intervals with a series of pumps to lift the water
at each sump.
Because of the relatively high silt and dust content of the seeping and rinsing
water there is a distinct danger of the laterals and the gallery becoming clogged
through sedimentation, owing to changes in velocity and discharge. All dimensions
must be established keeping accessibility, cleaning and maintenance in mind.
Washrooms and sanitary facilities must be provided in all subway stations for
the employees and for the general public (see also Section 63.342). The sewage
to be disposed of amounts to approximately 50 I/head (10 gal/head) i.e. about
1000 I/day (250 gal/day).
 CORROSION PROTECTION 493

4.64. PROTECTION AGAI NST CORROSION

In addition to the dangers of water, frost and atmospheric pollution, mentioned

earlier, the greatest enemy of concrete and steel underground structures is corro-
sion due to water and to soil. The protection against such corrosion has to be
preventive in nature because the affected areas, i.e. the outside face of the lining
in contact with soil and water are practically impossible to inspect after construc-
tion. This is why such protection should obviate the need for maintenance and
repairs and should serve as waterproofing at the same time.

46.41. Problems of Corrosion by Soi/4•3i

Because this is an ionic corrosion it does not depend on the presence of water,
although the latter will certainly stimulate the exchange of ions caused by electric
potential difference. Specific electric resistance and electric potential vary with
the type of soil to start with. Variations in oxygen and water content add to this
natural difference in the electric potential of soils. Thus, soil corrosion is of natural
origin and is accelerated or retarded by other factors. It is particularly hard on
metals, because any metal conductor connecting two materials of different electric
potentials will initiate an electrolytic process. In the case of iron, for example,
the loss of an electron and its negative charge will ionize the iron molecules and
cause them to move over from the anode to the cathode. The chemical reactions
are :
at the anode
Fe+ 20H- = Fe(OH)z + 2e-,
at the cathode
2H 20 + 2e- = H 2 + 20H - ,
and the total reaction is
Fe + 2H 2 0 = Fe(OH)z + H 2 •
The characteristics of corrosive soils are: low carbonic acid content, high
degree of acidity, good conductivity and high salt and moisture content. Soils
that are rich in carbonic acid are normally non-corrosive, but the difference in
their oxygen or air content will stimulate the potential difference as well as cor-
rosivity. (The difference in oxygen content is also referred to as differential venti-
lation.) The oxygen content and porosity of the ground above the pipe or tunnel
are of particular importance to corrosion. A steel lining, for example, passing
through soil strata with variable oxygen content, will develop a potential difference
and thus generate corrosive electric currents just as in a galvanic cell. Disintegra-

4 33
· BACSKAY, GY.: A talaj altal okozott korr6zi6 (Corrosion caused by soil), Mern . To1J
Kepzo 1955 3436
LINDSAY, M ., APPLEGATE, M .: Cathodic Protection. Mc Graw Hill, 1960.,
R OMANOFF, M .: Underground Corrosion NBS Circ. No 579 1957.
-- - ----- -- - - - - - - -- - - - - - - - -

494 DESIGN OF TUNNEL SECTIONS

tion of the metal begins at the origin of the current i.e. at the anodic areas (Fig.
4/ 129). Vertical shafts are particularly exposed to the danger of differential venti-
lation because the soil near the surface will always have a higher air content than
deep below, and so corrosion will inevitably develop even if soil conditions are
uniform in every other respect. Another factor in corrosion is grain size. While
oxygen supply is low where the tube is in contact with large lumps or boulders,
the opposite is true where the material is finer and so corrosion pits will develop
under the lumps.

: ._ c :_ .·-_ -:· .. __: .

·:. -~:::o.··.. .·. _··.. _:__
· -·. " _ . o· · - -. .
.o .·.. . • . · . . · . o :
- Sandy 9;~~;,:: -. : .
:(; \:i -~\·<· ~._=~--
.; ·.: ·.. :··.o ,· :_
~ .
_-. . . . . · -.-.·. o . Efectrohte
(Pit) . o ·:;:,e~f_·:
Spot corrosion ·\ ,/:A(lf!d_k Jru
~ e _ _ Electrons
FIG. 4/ 129. Scheme of
soil corrosion
- Iron pipe

Chlorides, sulphates and sulphides increase corrosion as well as the corrosive

activity of soil bacteria. As is the case with variable oxygen content, any difference
in the moistur~ content will also generate potential differences. Oxygen content
in water will stimulate corrosion. While 'black rust' is characteristic of soils with
high moisture content, 'red rust' can be observed where such moisture content
is low.
Corrosion is also stimulated by bacteria in the soil (aerobic and anaerobic)
in more than one way:
(a) The velocity of the anodic or cathodic reaction is affected.
(b) Initial resistance of the metal may be impaired by metabolites.
(c) Corrosive pockets are created in the soil.
(d) Electrolytic concentrations are formed through growth and prolificacy.
Sulphate reducing bacteria in particular may accelerate soil corrosion. The
chemical reaction is

Corrosion also depends on the quality of steel or iron. Cast iron, for example,
is much more resistant to corrosion than is alloy steel. In fact, alloy and iron
particles can generate potential differences themselves in so far as the first will
act as cathodes and the latter as anodes. However, the rate and speed of corrosion
 CORROSION PROTECTION 495

will depend much more on the type of soil than on the type of alloy. Scale on the
steel surface could afford some protection provided it forms a continuous and
uniform cover. Because this is not the case, ~owever, potential differences will
result between iron and scale at every crack in the latter. Because of the difference
in electric resistance of the two materials the electric current will be concentrated
at the relatively small anodic area and will cause severe pitting on the metal.
(Local potential differences can be observed also in welding When the heat treat-
ment of successive zones is. not uniform.)
The corrosivity of soils is measured by their resistance (ohm cm) and is divided
into five groups:

Resistance
Group No. Corrosivity (ohm cm)

I Low ~10000
2 Normal 2000-10 000
3 Aggressive I 000- 2 000
4 High 500- 1 000
5 Very high 0- 500

The following information should give an idea of the pace at which corrosion
progresses through an iron plate 10 mm thick :

Period of time
Group No. Corrosivity of corrosion
(years)

I Low More than 25

2 Normal 10-25
3 Aggressive 5-10
4 High 3-5
5 Very high 1- 2

The process in soil corrosion is similar to that artificially utilized for electric
soil stabilization. It is brought about by an exchange of ions, but what is primarily
aimed at is the source of damage in the latter.
Protection against soil corrosion can be achieved by the application of miscel-
laneous coatings. Tar is not suitable for this purpose, in fact, it may be detrimental
{because of its phenol content). Bituminous talcum mix can be used successfully
if applied over a preliminary ferro-fixol coating.
Soil corrosion can be boosted greatly by stray currents in the ground.Originating
from uninsulated conductors such as t!Je rails of street cars or subways (transmitted
to the ground through the steel tucn.c! lining) they can be a source of considerable
potential differences resulting in rapidly progressing soil corro~i0n.
 496 DESIGN OF TUNNEL SECTIONS

Anodic areas will be created where the current leaves the tunnel lining, develop--
ing heavy corrosion, and cathodic areas will develop at location.s Qf re-entry
(Fig. 4/ 130). Stray currents can also affect the reinforcing bars in the concrete,
as reported by a number of authorities. Because of the expansion arising from
corrosion the reinforcing bars can even cause the concrete cover to spall. In some
cases, concrete stresses due to such expansions as high as 2500 psi have been
reported.
+
Overhead wire

I
I I
/ / I I
/ ,./ / /
<< < :: . ;&
Clthodic lO/le Anodic 7011e
heav!I corros1011
FIG. 4/130. Electric corrosion due to stray currents

The corrosive action of stray currents does not depend on oxygen supply,
moisture content or soil type, and is a function of current intensity only. It varies
with different metals.
In one year an intensity of 1 A can disintegrate

9-128 kg (20 124 lb) Fe++

20-770 kg (45 790 lb) cu++
10-390 kg (22 906 lb) cu+
33-866 kg (74 662 lb) Pb++
10-665 kg (23 512 lb) zn++
40-000 kg (88 185 lb) Mg+-¥

The answer to corrosion by stray currents is cathodic protection. 4 · 34 Artificial

electrodes are driven into the ground in the vicinity of current leaks. These are
more anodic than the lining material of the structure to be protected and thus
transform the latter into a cathode. The electrodes - aluminum, zinc, carbon or
magnesium - disintegrate in time as they protect the structure. It is surprising
indeed that these remarkably low potential differences (even unable to ring a
door bell) are sufficient to protect iron pipes and linings for miles. The electrodes
can be located far away from the pipeline if connected through a cable. This pro-
4 34
- LINDSAY, M., and APPLEGATE, M. : Cathodic Protection. Mc Graw Hill, 1960
 CORROSION PROTECTION 497

cecture can be regarded also as i::ounteracting the local effects of galvanic action.
Fig. 4/131 indicates the reaction taking place in the vicinity of the Mg anode.
Free H+ ions are streaming towards the cathodic iron pipe and cover it with a
protective coating while the anode is being oxidized. Any local effect due to
impurities in the pipe will also be checked by this cathodic protection.

r;:=e-e
t I Two electrons ,1re tuns/erred
l'rom the anode in place ol'eacll
~
M'ion
#iiMIJSi'lM'l&J&/ff/.ffffeJ, l/1&/&IN/lill)
+ Pipe is coated
With JI atoms

>fg-2e·._ Mg~'
Positive charges
transferred from H+
e;_ [;Jc/I Mg++ 1011
the Mg anode to '- emits two e·
Lhe cathode % electrons
H' OH- ·;;; ~
~~
Mg(OH}2 ~~
:.::
FIG. 4 / 131. Scheme Remains in the anodtc
of cathodic protection zone

Special bedding should be provided for long life anodes, consisting of granu-
lated coke, coal and graphite with some alkaline admixture such as lime. The
alkaline admixtures retard the electrochemical process and are effective in:

I. reducing the electric resistance between soil and anode;

2. prolonging the life of the anode by providing relief throughout the electro-
chemical reaction;
3. increasing the electric capacity without r.educing the life of the anode.

There are other methods of preventive protection against stray currents as well,
which concentrate on reducing the opportunities for the current to escape, such as :

(a) butt joints between full sections of adjacent rails i.e. welding (inadequate
connections are the main causes of stray currents);
(b) connections between separate sections of steel lining or steel waterproofing
to prevent current escapes (e.g. connection between two sections with steel
lining separated by one with concrete lining);
(c) careful insulation of the rails and complete current return through the
rails;

32 Szkby : The Art of Tunnelling

~ - -- ·----- - -- - - --

498 DESIGN OF TUNNEL SECTIONS

(d) making the structure cathodic and using corrosion resistant materials for
anodes;
(ej alternating the locations of poles daily in order to reverse the direction of
the current.
Soil corrosion may affect concrete linings as well but only if accompanied by
changes in the groundwater table or by unstable groundwater conditions. Unstable
groundwater is bound to dissolve the calcium or magnesium sulphates (both are
detrimental to concrete) that may be present in the ground and deposit the salts
along the concrete lining. Developing sulphate action and cement bacilli will
gradually corrode the concrete. This phenomenon is very common in channel
construction, particularly with fluctuating water levels. Salts will be filtered out
again and again by the concrete lining at a certain water level and accumulated
over the years. Rising water will then dissolve the salts and form a highly con-
centrated solution that is detrimental to the concrete.

46.42. Groundwater Corrosion

Corrosion by groundwater i.e. the rusting of underground steel structures is

a phenomenon of ferrohydroxide formation and can affect concrete structures
as well. Sulphate action, mentioned earlier, can also originate from remote areas
via seepage flow. Because of its low lime content soft water can be just as danger-
ous for it may dissolve the lime from the concrete, and thereby gradually remove
its ~inder. Even heavier damage can be expected if the groundwater is allowed
to seep through the concrete, either along construction joints or because of exces-
sive porosity. Damage of this nature is not confined to the surface but affects
the concrete lining throughout. Waters with a pH value of about 6·5-6·7 (i.e. less
than the neutral pH = 7) already can be detrimental. The best protection possible
is to make the concrete really solid, using plasticizers and other suitable fillers
and admixtures. Failing this a complete waterproofing system can still provide
adequate protection. No water must be allowed to escape under any conditions.
Decomposition of iron sulphide in clayey soils can cause just as much trouble
as can soft water. Once exposed to the air, the iron sulphide will form a sulphate
solution with the seeping groundwater and, thus, will attack the concrete. Excava-
tions in clay materials with iron sulphide content must not be left open for a long
time but must be lined as soon as possible. Another reason for keeping the exca-
vation open only as long as is absolutely necessary is the fact that exposure to air
helps to develop sulphate-reducing bacteria and other aggressive micro-organisms.

4.65. TUNNEL VENTILATION

There are two types of tunnel ventilation: temporary (during construction) and
permanent. The latter may be either natural or artificial.
 VENTILATION 499

46.51. Ventilation During Construction4 •35

Because of the low permeability of soils natural ventilation is seldom encount-

ered in underground working areas. Air circulation due to differences in temper-
ature, the passage of haulage trains or pneumatic tools cannot supply fresh air
through natural ventilation. In addition to the amount of fresh air required for
the working crew, pollution by explosives, smoke and fumes also has to be taken
care of by the ventilation as also has any uncomfortable excess heat that may be
present.
Air consumption increases with the pace of working and thus becomes
a matter of economics as well. Good quality fresh air must be supplied in sufficient
quantities to all working areas and particularly to the working face without inter-
ruptions or breakdowns in order to improve health and working conditions,
and productivity, and also to increase efficiency.
As soon as the oxygen content of the air drops below 20 % (from the natural
21 %) headache and faintness will develop. Less than 17 % oxygen can cause
unconsciousness after a period of time. (The oxygen content of exhaled air is 17 %. )
In order to estimate the volume of air required the causes of air pollution and
consumption must be considered first. These are:
1. number of men working in each shift
2. haulage system
3. type and quantity of explosives
4. type of operations
5. geological conditions, gas content arid gas production of the ground.

Ad 1. For a given minimum of 20 % oxygen content a maximum of 0·9 % CO 2

4·0
can be tolerated. Because of the 4 % CO 2 content of the exhaled air . = 4· 5
09
times as much fresh air should be supplied as is exhaled. This factor is usually
increased to 10 to take care of losses and also for better dilution. About 30 I (qts)
of air is required for each person per ·minute under normal conditions. Design
values, however, average around 2·0 m 3 (2·5 cuyd)/min/person to allow for fumes
and other pollutions arising from carbide lamps. In long tunnels there are extra
air requirements for cooling, which can be pro-rated also for each person. A prac-
tical minimum total of 60 m3 (80 cuyd)/min represents standard air requirements
for a long tunnel.
Ad 2. All haulage systems, even cars drawn by horses or humans, will have
an effect on the air requirements. (Horses require 4 times as much air as do humans.)
Practically no air is required for electric or pneumatic haulage systems. Diesel
systems require 2·3 m3 (1 ·6 cuyd)/min air for each horsepower according to RABCE-

4·35 RICHARDSON and MAYO : Practical Tunnel Drivin,q. Mc Graw Hill, 1941
Z ANOSKAR, W.: Stollen u. Tunnelbau. Springer, Vienna I 964

32•
- =- - -- -. -
~ - - - - - -

500 DESIGN OF TUNNEL SECTIONS

w1cz. The CO content of exhaust fumes of combustion engines may reach 3 %

when working and 7 % when idling, particularly with improper mixtures and
carburettor settings.
Ad 3. Explosives and their fumes affect air consumption to a large extent.
Although varying for each type, generally speaking all explosives produce vast
volumes of fumes, consisting mainly of CO, CO 2 and N0 2 • The required venti-
lation is determined in terms of the allowable concentration of the poisonous
fumes. In the case of CO for example the concentration that is still harmless is
about 0·02 % and the required volume of air, Q, is

Ab
Q = 5- (m3),
t

where A = the weight of explosives in kg

b the amount of CO produced per kg of explosive in
t = the time of ventilation in min.

The value of b is fairly uncertain as it also depends on the excavation and on

the method of blasting (priming, density of charge); an average value would be
around 100 I. For example: the explosion of I kg dynamite produces 0·6 m 3 of
fumes, of which 22 %-42 % is CO 2, 46 %-28 % is CO, 17 %-25 % is N and 15 %-5 %
is H .
The following practical design method can be recommended. It has be~n
observed that the area contaminated by the fumes of explosives is less than
20 m (66 ft) deep within the specified minimum safe interval of 10 minutes
between successive explosions. (According to ROMANOV the range of pollution is
I= 2·4A + lOm .) It should be more than enough to exchange this volume of
air five times within the 10 minutes waiting period specified in mining regula•
tions, i.e. :

20F-S lOOF . F
Q = --- = -- = l0Fm3/mm = - m 3 /sec
t 10 6

where F = the cross-sectional area of the tunnel in m2•

An empirical formula has been derived by ANDREAE from data of the construc-
tion of the Simplon tunnel:

Q = O·SN + 0·104A m 3/min,

where N = the number of miners

A = the weight of explosives in kg detonated per hour.
Ad 4. and 5. The volume of air required to compensate for the type of opera-
tion and for geological conditions shall be evaluated always on an individual
 VENTILATION 501

basis, considering the intensity of haulage movements, frequency of blasting, type

and quantity of natural gas conte,nt of the ground, etc. It js well to remember
that wood fire g~nerates as much _as 570 cuyd of CO for each cuyd of wood and
considerable volumes of CO and H 2S will develop in the proc~ss of decomposition
of organic m~tters and rotting timber. (Heavy Co and H;iS coJ}.ce_n tr_ations are
bound to be encouptered when re-opening old drainage a,dits and abandoned
headings.) Lubricants in mechanical equipment will also contribute to the e,q:i.aust
fumes under high temperatures of the combustion chambers a_nd i~re,i~,se the CO
and CH 2 content of the working area.

Fm. 4/132. Losses in

head in mechanical ventilatjon

Fresh air has to be blown in under pressure if the ventilation is really to be

effective. The required pre.ssu_re can be computed from t~e losses encountered
in the drift (Fig. 4/132).
There are three factors contributing to the above losses: friction, increased
resistance jn bends and valves, and the requirements of a constant velocity (v)
Jhat has to be maint~ined. For circular drifts:

(4.186)

where ). = the coefficient of friction

'l' the specific weight of air (l ·1255 kg/m3)
L,.... the length of the straight drift section
D ;=; ~he diameter of the drift
e ; ;: an empirical factor representing the proportional extra length due
to all local resistances (bends etc.).
For a given volume of air Q

Q Q 2 - 16Q2
V= - = - - and
p D21t
V ---
7t2 n•
-4-

and the pressure drop due to friction

2
h = ,t !:_.:!_ = 16,lyLQ
• 'l' D 2g 1t2 2g D 5 '
~ - - - ~ -
r"""':'"i -------- ~

502 DESIGN OF TUNNEL SECTIONS

i.e. it increases as the square of the volume of air to be carried and decreases as
me fifth power of the drift diameter.
EXAMPLE (by Kaposztas). Find the depression hn at the intake of a pipeline
L = 300 in long, D = 0·5 m in diameter, for a volume of air Q = 6 m3/sec and
a specified water head of 300 mm at the end of the pipe (,l = 0·014 and y =
= 1·22 kg/m3)

v= _q_ 6 · O x 4 = 30·6 m/sec and h = ,ly _£ ~ = 0·014 x 1·22 300 . 30"62 =

F 0-52 • 1t D 2g 0·5 19·62
= 471mm
the head at the intake is 471 + 300 = 771 mm.
There are three methods of ventilation during construction:

(a) blowing
(b) exhaust
(c) a combination of the two.

(a) In the blowing method fresh air is forced directly into the working area.
The foul air drifts back through the_full tunnel section, suddenly increasing the
rate of contamination and only slowly giving way to gradual clearing. On the other
hand the working face enjoys a positive supply of fresh cooling air resulting in
improved performance all around. The disadvantage of the method is that the
foul air and smoke filter back through the full cross-section of the tunnel and there-
fore at lower velocities. Smoke and dust inundate the entire heading and clear
up rather slowly. As a result, scaffolding and lining operations are impeded by
foul warm air.
(b) Just the opposite occurs in the exhaust method. Foul air is drawn out
through a pipeline and fresh air traverses the entire length of the heading. While
dust and smoke due to blasting are kept out of the tunnel in general, the exchange
of air is not as fast _and is not as uniformly distributed at the working face as it
was by the first method. Temporary pipe extensions are required, resulting in
more loss of air. Fresh air entering at the portal or shaft may absorb heat and
moisture in its path down the heading and arrives at the working face in a some-
what deteriorated and warmed up condition. For better results a second, auxiliary
pipe can be installed to supply fresh air directly to the working face.
(c) One way to combine the blowing and exhaust methods is to use the first
at the working face and the latter in other working areas. This arrangement is
particularly feasible in long tunnel sections.
More advanced modern methods combine the advantages of the first two
methods and apply the proper system at the proper place and at the proper time.
!n this case the common pipeline will :
 VENT"~ATION 503

1. fbrce in fresh air after blasting operations to displace . smoke aqd furoes;
2. exhaµst smoke and foul air at slower velocities to the surface ;
3. blow in fresh air again for a direct and positive supply of fresh air to the
working face.
It is very simple to switch over from one action to the other, even the fans can
be non-reversing; a double pipe connection and appropriate valves are all that
are required. The diagram of a simple layout is illustrated in Fig. 4/133.

f
Exnwst dct,011 8/owinp /clion '

To the
from the
,svrf;ce <l:::+-,:-~c.i-__.__.... svr"1ce

FIG. 4/133. Di~gram for the combination of blowing and exhaust ventilafion methods

46.52. ·Natural Ventilation

Taking advantage ·or natural ventilation is a possibility so important that it

should be kept il_l: mind as early as the time of deciding on the location of the tunnel.
For example it is ·advantageous to lay out the tunnel parallel to prevailing winds,
to expose the approaches and portals to rather than shelter them from such winds
and to avoid an align.pient with sharp curves and reversing grades. Meteorological
records should be consulted and preference should be given to locations exposed
to high winds.
Natural ventilation obviously makes maintenance easier because of the drying
effect on condensation water and minor water seepages. These, as we know,
are both prerequisites for corrosion (particularly if combined with fumes and
smoke) and will affect road, track, rails, lighting and lining at the same time.
Natural ventilation is performed partly by winds on the surface and partly by
differences in outside and inside atmospheric pressures due to differences in tem-
perature and specific gravity. ANDREAE4·36 divides the total pressure differentiai
into 3 components :

4 •36 ANDR£AE, CH.: Probleme,s du projet et de l'etablissement des grands souterrains ro11tier

a/pins. Leemann, Zurich 1949

 504 DESIGN OF TUNNEL SECTIONS

1 a(t
2 - t )
where h1 = y H - - - - (hydrostatic head in mm) is due to the difference in
1 + at 1
temperature
H = the difference in elevation between the two ends of the tunnel
y = the specific weight of air (kg/m3)
t 1 and / 2 = the ambient inside and outs.ide temperature s in °C, respectively
a= ~ hb = 13·6b (hydrostatic head in mm) denotes the effect of the
2 3
difference in barometric pressures
b = the difference of barometric pressures between the two ends of the
tunnel
2
hw = sin2 </> ;; (hydrostatic head in mm) denotes the wind effect
w = the wind velocity (m/sec)
</> = the angle of inclination of the wind to the centreline of the tunnel.
With due consideration of friction losses the velocity of the air moving in the
tunnel becomes

V=
n
-y (i +e RJ.. L)' (4.187)

where ),_ =
the friction coefficient of the lining (0·007 for stone and brick and
0·006 for smooth linings)
L = the length of the tunnel in m
y = the specific weight of air (kg/m3)
R = ~ = the hydraulic radius (F and U being the area and the perimeter,
of the cross-section, respectively).

Although natural ventilation may assume considerable proportions it should

not be relied upon in the design, according to ANDREAE, because of its intermittent
nature. However, it may obviate the need for mechanical ventilation in short
tunnels or tangent sections with non-reversing grades. It should be noted that
longer one-way than two-way tunnels may be served by natural ventilation; this
will also depend on the density of traffic and on the type of traction. Tunnels
carrying electric trains can be efficiently ventilated naturally for miles. Changes
in the cross-section of the tunnel may also stimulate natural ventilation because
of the resulting velocity change.
 . VENTILATION 505

According to ·KREss'· 37, the maximum length of a highway tunnel without mechanical
v:ntilation should not be .determined on the basis of natural ventilation. Such maximum
length should depend on the cross-.sectional area (F), the entire volume of the tunnel, the aver-
age running speed (v) (km/hr), the traffic density and the capacity of the tunnel, and the actual
and maximum allowable CO content of the air in the tunnel. For an average spacing a be-
tween the cars (including braking distance), an allowable CO pollution rate off and a number
of vehicles z = 2000 ~ (at any one time within the tunnel), for an exhaust fume volume of
CO of q (I/min. per ca:) and for each car travelling within the tunnel for a time of / = 3600 !:_ (sec)
the maximum length of the tunnel is v

L = /Fav
120q
where L is expressed in km, and q in I. As an example With f = 0·22 %, v = 15 km/hr.
F=61 ·77 m 2, q = 60 I/min and a = 9·9 m the above formula of KRESS yields L = 0·28 km.

Natural ventilation has been considered lately but only to a limited extent.
It has been assumed that mechanical ventilation could be discontinued as long
as natural wind velocities exceeded a precalculated minimum. This has resulted
in a certain amount of saving in operational costs. 4·38

46.53. Mechanical Ventilation

Generally speaking, mechanical ventilation is warranted for tunnels carrying

highways, subways and also for railway tunnels if the latter are several miles long
and are travelled by steam locomotives. The volume of air required depends once
again on the allowable concentration of air pollution and on the amount of heat
to be removed. It is well to remember that the allowable air pollution varies for
each poisonous gas and is in inverse proportion to the period of time spent in
such environment, i.e. the allowable concentration increases as the length of time
decreases (Fig. 4/134). For this reason consideration should be given to service-
men or maintenance crews who would have to stay longer in the tunnel while
performing their duties than would the travelling public.
Density of traffic and the type of travelling vehicles will greatly affect air require-
ments. It has happened on more than one occasion that tunnels with natural
ventilation performing well for years have had to be provided with mechanical
ventilation later because of an increase in traffic.
It must be noted that the causes of natural ventilation, such as differences in
barometric pressure, temperature and wind, may be working against the mechan-
icai ventilation as well. This is why the latter has to be designed not only to supply
the amount of air required, but also to counteract and overcome the above natural
factors.

u 7 KRESS, H.: Liiftungsentwurffiir den Wagenburgtunnel in Stuttgart, Schw. Bztg. 1953 52)
..:ia BARTH, KLEIN and RAAB: Eine neuartige Liiftungsanlage in Liimmerbuckeltunnel,
Bauing. Dec. 1958
506 DESIGN OF TUNNEL SECTIONS

Fm. 4/ 134. Danger ratings or

prolonged stay
in CO polluted
~oo~-~_ _.___ _.___....,___....,___-'-_ _.J__ _J
atmosphere
I J 4
( flours)

Air suction behind moving vehicles should also be considered in the design.
Because of the vast amounts of air in question air conditioning, heating, cooling
or conditioning of any other type would be absolutely impractical.
46.531. Mechanical ventilation of highway tunnels. Large scale experiruents
have been conducted in the mining experimental station in Pittsburgh regarding
the chemical composition of exhaust fumes in connection with the construction
of the Holland tunnel in New York. Similar experiments have been carried out
at Yale University dealing with physiological aspects !ind with the rate of dilution
employed as a countermeasure. Power requirements and practical ventilation
systems have been studied at Chicago University at the same time. While sub-
stantial amounts of CO have been observed in the exhaust fumes of gasoline
engines, the CO content in diesel exhaust fumes _has been insignificant. (Recent
studies indicate, however, that the latter do contain acrolein, a gas also highly
detrimental to health.) The CO content in gasoline exhaust fumes is a function
 VENTILATWN 507

of fuel consumption and also depends on whether the vehic]e travels downhill,
uphill or on the 1evel and whether the carburettor had been set properly, a factor
reflected in the rate of combustion.
Table 4/XIV indicates 1he volume of CO discharged in I/min a). The same is
related to 1 m of tunnel in line c). Both uphill and downhill gradients are taken as
3·5 %-Similar data are given under b) for 5·0 % gradients, developed by KREss for

TABLE 4/XIV
Volume of CO Content in Exhaust Fumes (I/milt)
--
Speed
I Avcraac size
passenaer cars and
lill,htina
Trucks 2t averase
and lill,htina
Trucks 2t to St and
Jiahtina
Trucks over St and
liahtina

(km/hr)

I Up- \ Down- Hori-

hill hill zontal
Up-
hill
Downi Hori-
hill zontal
Up-
hill
I
Down-/ Hori-
hill zontal
Up- /Down Hori-
hill hill zont.a l

a)
16 40 18·3 29 40·6 31 ·6 80·0 18·0 62"3 85-6 71 ·6
24 56 27 41 52·0 26·3 40·0 83·0 44·7 56·4 42·5 54·0
Starting 52 46·5 26·4 21 ·0 97·0 48·0
Idling 21 19·4 23-8
Driving 39·5 32·0 49·6

5 km b) 35 II ls-20 29
40
15
12
II 53·S 2S
74·0 34
. 6S
90
131
45
15 km
10 km
43 10·8
51·5 14·35 I 40·0 :- - - -
s_t_a_rt_in_g_ _ _ _5_5_·6_1_ _._ 32·70 60·5

c) 1

16 km 0·150 0·067 0·10810·150 0·071' 0·120 0·300 0·142 0·225 0·322, 0·15010·268
24 km 0-150 0·061 0·10010-125 0·06510·100 0-200 0-115 0-140 - 0·0451o·i35
1

the ventilation design of the Wagenburg tunnel in Stuttgart. It can be noted that
an increase in upgrade from 3 %-5 % increases CO production by 13 %- Table
4/XV (by SCHLAEPPER, Zurich) indicates that carburettor setting has an even
greater effect on CO production. A design CO volume of 60 1/min per car is used
in American practice rather than the 32·81/min assumed by KREss.
Health authorities report that CO contents less than ,0·04 % do not affect
human beings. An allowable limit, however, should be set at 0·020 %-0·025 %
to avoid problems arising from impaired visibility. As mentioned earlier, CO
contents in diesel exhaust fumes are much lower because of surplus· air facilitating
complete combustion in the cylinders. 'Because of the acrolein content, however,
mechanical ventilation will still be required to avoid impairl!d visibility.
 508 DESIGN OF TUNNEL SECTIONS

TABLE 4/ XV
Volume of CO per Vehicle (for 1 fin. m of road)

Trucks
Passenger cars with gas
Carburettor setting consumption of
15 1/ 100 km
30 1/100 km
I 501/ lOOkm

Gas consumption

Good 0·023 0·046 0·077

Fair 0·060 0·120 0·200
Poor 0·096 0·192 0·312

Ventilation systems have to be designed for maximum CO pollution conditions,

i.6. for cars travelling at the closest spacing and at the lowest speeds. 4• 39 Although
air cooling is also importan t in highway tunnels, it should be only a secondary
consideration because it requires less air than the amount required for freshening
the air through dilution. For this reason it is not usually a governing factor in
ventilation designs (see also Section 46.532).
According to KRESS each traffic lane can carry

c = 1000v = ___36_o_o_v _--..,,.

a a1 + 0·5v + 0·l66v2 '

where C = the number of cars per hour (disregarding delays caused by possible
intersections at tunnel entrances)
V and v = the average running speed in km/hr and in m/sec, respectively
a = the spacing of cars (m)
a1 = the average car length (m).

The total number of cars travelling within the tunnel at any one time is

L L
z = 2c v = 2000 a .

The amount of exhaust CO is CO=¼ 0·0001 qz (m3/sec) and the volume of fresh

. require
. d.1s Q CO 1000 0·01667 qz ( 31 ) A l
arr F f = f m 1 sec. compete exchange o f air
·
. . d 3·6Q . h
1s require n = -V- times per our.

4 •39
NOTE: Due to a traffic congestion in the Liberty tunnel, Pittsburgh , cars were lined
up
throughou t the entire tunnel. Natural ventilation proved inadequat e to keep the air pollution
down; some people fainted, others panicked. Mechanical ventilation was installed
shortly
after the incident.
 VENTILATION 509

The volume of acrolein gas (mainly in diesel exhaust fumes) is

A
G = 15B IOO (kg/h)

where B = the fuel consumption per hour (kg/hr)

A = the percentage acrolein content of the exhaust fumes (0· 13--0· l 5). uo
3 3
The allowable rate of air pollution is 2 mg/m (20 mg/m for CO). According
3
to GUERIN, 10 minutes spent in a polluted atmosphere (to the extent of 11 mg/m )
may be fatal. 4•41
Permanent ventilation methods may be classified as
(a) longitudinal
(b) lateral
(c) combined.
(a) Longitudinal ventilation systems make use of the entire tunnel section itself
as an oversize vent pipe between entrances or ventilation shafts: the air traverses
the tunnel in a longitudinal direction. Because there are no special vent p'ipes
required this method has the advantage of low installation costs.

F10. 4/135. Scheme for tunnels with several ventilation sections

There are many ways in which the air can be kept moving. In the most common
arrangemen t a number of ventilation shafts are provided (the number depending
on the length of the tunnel) through which the air can be either forced in or drawn
out by fans. In the case of more than one shaft the exhaust and blowing shafts
are spaced alternately (Fig. 4/135). A uniform distribution of the air can be effected
by a proper layout of the shafts.
No matter how carefully the shafts have been laid out, this method has the
disadvantage of interruption s in ventilation due to adverse atmospheric conditions
(wind from the wrong direction, etc.).

uo Utasitasok garazsok fiitesenek es szellozesenek tervezesere (Design directives for the

ventilation and heating of garages), Ep. Gep. Dok. 1954
•·ct GUERIN, H.: Traite de manipulation et d'analyse des gaz. Paris 1952
- - ~ - ~ .

510 DESIGN OF TUNNEL SECTION

As an alternative, ventilation plant may be located at one or both ends of the

tunnel (Simplon). As long as traffic is light, the intermediate ventilating shafts
- sometimes very deep - can be dispensed with.
A gate has to be provided at the portal where the ventilation plant is located,
so that the air, just forced in, cannot escape through the portal in the direction
of lesser resistance, i.e. into the open. With ventilation plants at both ends, gates
should be provided at both portals. Air can then leave the tunnel only through
a shaft in tbe centre. The gates have to be opened to approaching traffic either
manually or automatically.
One of the widely used longitudinal ventilation systems is that of SACCARD0,
ba~ed on a vacuum effect. Part of the fresh air (about 25 %-40 %) is forced into
the tunnel at high speeds (15 m-25 m/sec) and at an angle of approximately 15°,
by fans located near the portal (Fig. 4/136). The vacuum created by the high speed
draws in the fresh air through the entire tunnel cross-section.

forcing in fresh Removing foul

air air

air

FIG. 4/ 136. Diagram of the

SACCARDO ventilation system

The SANGER ventilation system works on a_similar principle and is mostly used
in long tunnels where friction losses are too high to overcome with a ventilation
-p lant located at one end only. A number of intermediate fans are installed at
suitable intervals, designed to draw the air from the preceding tunnel sections
and to force it back into the tunnel in a fOiward direction at high speeds and at
a slight angle thus giving a boost to the longitudinal draft.
This system can be particularly recommended in cases where the cost of con:·
structing intermediate ventilation shafts would be prohibitive.
The main disadvantage of longitudinal ventilation systems is the fact that
conflicts may develop between mechanically forced air on the one hand and
natural ventilation and air pressures built up by moving traffic on the .other. This
will reduce ·ventilation efficiency and call for additional power requirements.
Meanwhile the traffic itself will face highec air resistances and suffer losses in
 VJiNTILATION SU

power. This drawback applies particularly to long tunnels where air has to be
forced in at high velocities at the entrance in order to get through the entire
tunnel. This is another reason for dividing long tunnels into sections. In this
arrangement fresh air is forced in· and foul air drawn out through double barrel
vertical shafts, situated in the centre of each section. In another alternative,
taking advantage of natural ventilation rather than opposing it, the fans are
installed at both terminals of the tunnel and only one group is operated at a time
- that working with the prevailing wind. 4· 42

FIG. 4/ 137. Lateral ventilation

in subaqueous highway tunnels

Longitudinal ventilation systems are used in almost all railway and subway
tunnels, but are hardly ever used by themselves in highway tunnels. The disadvan-
tages cannot be eliminated wherever there is two-way traffic.
(b) In lateral ventilation systems the air is forced to move transversely. Fans
blow in the fresh air at the bottom and exhaust foul air at the top (Fig. 4/ 137).
The tunnel is divided into sections, each of which is served by a number of blower
and exhaust openings. The fresh air is forced into each section from a separate
vent tube (Fig. 4/138) and the foul air is drawn_ou_t from the same sections through
conduits leading to the surface. Because foul air is lighter than fresh air, the
exhaust openings and conduits are.to be located at the top and the blower openings
and vent tubes are to be located '. at the bottom. This arrangement is also good
from the point of view of visibility: a top-to-bottom ventilation system would
impair the visibility in the lower half of the section, i.e. exactly where it is most
important. In highway tunnels it is best to locate the blower openings at curb
level (with no pedestrian traffic) or just above head level if there are pedestrians.
Exhaust openings should be located near the ceiling at a uniform spacing I 5 x 90 cm
or 15 x 180 cm (6" by 3 ft or 6" by 6 ft) openings at 3·0-6·0 (10 to 20 ft) intervals.

J ,I! B ARTH, KLEIN and RAAB : Eine neuartige LUftungsanlage im Lammerbuckeltum1el,

Bauing. Dec. 1~58
 ....._,,
t.;>

Rwer side vent!la!/017 Inland vert!1!alion

2464-Mm bu!ldillg building
Z4_4·5{} --
~ t:,
m
az
River lludsoll vent duct .,,
0
- __-:-.---- - - ---'<--- -tfl ,-j
C:
z
z
m
r
I Rwd qr.1de
I . "'"'
S'I
I C,
'Z-
VMttl3Ltoll secL!oll s-1
(5J6U A} r4 - +- s-2 Pump shaft _:
-
fi'es/Ja1r duct
7 - -- - - - - _!__::_3 J-4 _I _ J-5
I
V•

Fie . ·if; 38 Ventilatio n section~ in a subaqueous highway tunnel with ventilatio

n towers
 VENTILATION 513
While lateral ventilation is certainly more expensive than the longitudinal one
it is the best alternative for long tunnels with high ventilation requirements. It does
not interfere with traffic, which is most important, and removes foul air in the
shortest possible way. This was manifested in the case of a truck carrying chem-
icals (CSJ and which caught fire in the middle of a tunnel. The ventilation
system successfully kept exhausting the fumes until the arrival of the fire engines,
and thus averted disaster.
Both ventilation systems take in and remove the air through ventilation shafts
and towers. These shall be spaced as closely as practical in order to make the
ventilation sections short and keep energy losses down. For short tunnels it is
sufficient to locate them at the entrance and at the exit portals (see Figs 4/ 135
and 4/ 138); for long tunnels, however, intermediate shafts will also be required.
It has been found that in some tunnels - served by longitudinal ventilation - foul
air escaped the intermediate shafts and kept moving towards the portals. In
subaqueous tunnels - where intermediate ventilation shafts may be too expensive -
the ventilation towers are located right on the banks and the approaches are served
by additional towers (Fig. 4/ 138).
Urban underpasses located in congested areas and carrying heavy traffic may
also be ventilated by small fan units if these are properly spaced. In this case
intermediate shafts can be dispensed with. The Avenue Louise underpass in
Brussels, 330 m (1100 ft) long, has 17 fans on one side (2 HP and 21 000 mil
[27 000 cuyd]/hr each) drawing air from the street level. Small exhaust fans along
the other side remove the foul air - at springing level - out into the street (Fig.
4/139).
(c) In semi-lateral layouts the fresh air is supplied as in fully lateral systems,
but the foul air is removed through ventilation shafts rather than special vent
tubes. While not as perfect as fully lateral arrangements, semi-lateral systems may

Intake t~n

Foul air

frhwst fan

Ii
FIG. 4/ 139. Lateral ventilation with independent fan units

33 Szechy: The Art of Tunnellina

- -
------ -

- - --- --- - --~ - ~ - --~- - -

514 DESIGN OF TUNNEL SECTIONS

be considered in oval and horseshoe sections that are fairly deep. Fumes may then
be allowed to gather in the top regions and filter towards the ends or ventilation
shafts without too much trouble.
The following losses should be taken into account when calculating power
requirements for lateral ventilation systems:
(a) blowing
I. losses due to intake filters;
2. losses in vent tubes, curved sections and bends;
3. losses due to air escaping back to the street through the openings.

(b) exhaust
I. losses due to foul air passing through openings at the top;
2. losses encountered in exhaust ducts, curved sections and bends;
3. losses suffered while passing through fans and exhaust stacks.
The drop in the pressure head, resulting in zero velocity at the end of a duct, is given by
ANDREAE as (cf. Fig. 4/132):
2
h _ yvi AL yw ( l AL0 ) (4,188)
c -6R g + 2g +R 0
'

where v 1 and w = the velocities at the intake and at the exhaust to the street, respectively
y = the specific gravity of air \
L = the length of the vent tube
L0 = the length of the intake openings
R = .!... and R 0
= ~ = the hydraulic radii for went tube and intake opening5,
U U0
respectively
). = the coefficient of friction between the air and the vent tube: ). = 0·006.

The exhaust velocity can be determined from

w = .!!_
/kc
where q = the volume of discharged air
f = the area of the exhaust openings
k = the factor of contraction (0·63)
c = the velocity factor (0·97).
In a somewhat more complicated formula by 0. SINGSTAD the pressure head required to
blow the air over a distance x measured from the intake is:
2
h = y -
Vi [ aLz-3 - -1 (I - k)z2]
- +-bLz
-} + -y2g
w ( 0
1 + -J..L ) (4.189)
ex { 2g 3R 2 _ 2gR3 R

L -- -X- t h ereIat1ve
where z = · . d'1stance
L
a= 0·0035
b = 0·0001236
k = 0·615 = the contraction factor.
 VENTILATION 515

The pressure hea d required for the exhaust (by ANDREAE) is

2
yv~ ( J.L) yw ( J.L~) (4.190)
h, = 2g 1'36 + 3R + 2g 1 + R~ .

Note that the factor l ·36 in the first member has been increased from the previous 1·00
because exhaust systems are Jess efficient than blowing fans.
Similarly (according to 0. SINGSTAD):

hsx =
v'-
y { 2;
r(3
aLz
2

+ c) R + 2
3z
+
2

c
]
+
bLz
2gR3 (1 + c)
} yw
2

+ ~ O·S 7 '
where the c velocity factor represents an air volume of 0·3 m 3/sec/m.

Losses have to be calculated separately for each section and then summarized for the total
pressure head required. Additional pressure heads keep the air moving in the vent tubes:
2
yw ( ).' H)
ha = ~ 1+~,
where H = the length of the shaft.
The next pressure head is to include the effect of the temperature difference between the
top and bottom of the shaft and can be positive or negative, depending on the season:
ex (t1- lz)
h,= ±yH - - - - . (4.191)
1 + ext1
For two way ventilation the velocity w shall be doubled. The total pressure head thus
becomes:
h = he + h, + 2 (ha + h,) + h4,
where h4 = the dynamic pressure head required for the end velocity at the ducts:
2 + 2
h4 = ~
2g
- Y•

The required output of the fans is:

Qh
N= - - (HP),
0·7577
where 77 may be taken as 0·6.
Pressure heads are simpler to calculate for longitudinal ventilation because the only losses
to be considered are those due to friction, curved sections and dynamk losses in velocity.
As in the case of the ventilation of headings during construction:

h = h, + ha+ h4 = yv2
2g
(1 + E,_
R
+ ~) + yw•
2g
(1+ HJ. )
R
+ yH (t1 - t,) ex
l + ext 1
(4.192)

where / 1 and t 2 = the ambient air temperatures at shaft top and bottom, respectively.
46.532. Special considerations on ventilation of subway tunnels. While ventilation
requirements for highway tunnels are governed by the allowable CO content
and rnilway tunnels have to be ventilated because of fumes and temperature
considerafions, the ventilation of subway tunnels is required because of dust
and humidity control as well as temperature considerations. Also, while highway
 516 DESIGN OF TUNNEL SECTIONS

tunnels are ventilated laterally, railway and subway tunnels u:se longitudinal
ventilation in most cases. There is also a difference in the extent to which passen-
gers are exposed to the atmosphere of the tunnel. -In highway tunnels there is
direct contact throughout the entire length but there is hardly any such contact
in railway tunnels. In subway tunnels the exposure is limited to that at the stations.
The ventilation of subways is provided by two systems :4 •43

Air supply to a triple section station

~ .....
From Ian

Bottom vent duct

1 lknti/Jtion ch;;mher
2. Muffler chamber
,1. Shaft
4 Valves and gates
5. Oef/eclors
6. }{e,;d of' venli/.1/ion
shaft
Z intake grid

Typical section FIG. 4/ 140. Ventilation scheme for subways

Primary ventilation looks after the tunnel line, stations, cars, removal of such
excess heat as may be present because of motors, passengers and lighting, removal
of excess humidity due to ground water and passenger traffic, fresh air supply
to the passengers and dust control.
Secondary ventilation, on the other hand, serves all the underground power
plants and utility rooms.
Longitudinal primary ventilation works through ventilation shafts (Fig. 4/ 140).
As a general rule there should be one shaft for each station and another one

u 3 EMHO, L.: Varosi alagutak szellozese (Ventilation of urban tunnels), Mern. To v. Kepzo
1953 2150
GARMS: Handbuch der Heizung- 11. LiiftungsJechnik
ARBuzov, G. V.: Ventilaciya tonnelei Metropoliteno v (Ventilation of subway tunnels),
Moscow 1950
Improved Ventilation of London Tube Railways, Engineering l~53 Dec. 713
 VENTILATION 511

between consecutive stations. Air is forced in, and drawn out, by axially set up
fans through the ventilation shafts and through the adjoining vent tubes.
The direction of ventilation can be reversed with the seasons of the year.
Passengers are exposed to ventilation in the entrance halls, escalators and plat-
forms. In summer, people entering the subway have an avidity for fresh air
and this should be supplied directly through the ventilation shafts right at the
station. From there the air will flow towards and co0l the shafts located half way
between the stations. The direction of flow is reversed in winter: cold air is
forced into the tunnel first and by the time it reaches the stations and the passen-
gers, it has been warmed up. From here the air is exhausted through ventilation
shafts Iocateo at the stations. In order to have more flexibility in meeting the chang-
ing ventilation requirements, it is customary to provide axial fans in groups of
two. The fans are installed in the adjoining horizontal vent tubes rather than
in the vertical shafts themselves (Fig. 4/140).
The amount of air required for the ventilation of rapid transit systems is pri-
marily governed by heat considerations. Heat is generated and accumulated in
direct proportion to the density of traffic. Mechanical ventilation thus becomes
a necessity in all cases where heat is generated at a higher rate than can be absorbed
by the walls through conductivity. The amount of heat transmitted to the ground
by the walls is limited and may be expressed approximately as fo11ows:

Q1 = F(ta - t,) k, (4.193)

where F = the surface ~rea of the tunnel lining in contact with the ground
ta and t, = the temperatu_res of tunnel and soil, respectively (in °C)
k = the coefficient of heat conductivity (about 1·2-1·5 ca1fm2/hrj°C)

Because of the low heat conductivity of the ground the absorption of heat by
the tunnel lining cannot be relied upon. The ground cannot maiutain dispersion
of the heat that is constantly being generated and heat storage areas will develop
around the tunnel with alternating flows of heat: "charging" in warm weather
and "discharging" during cold periods (e.g. cycles of day and night, summer and
winter). ·
For this reason most of the heat has to be removed by mechanical ventilation,
which is required anyway for health reasons. The type of longitudinal ventilation
brought about by the moving trains pushing the air like pistons in cylinders is
used only as an auxiliary today, because the pressure afforded by the trains is
inadequate to get the foul air out to the surface.
For a modern subway tunnel with heat being generated· at the rate of 688 000 cal
by trains (800 kW) plus 120 000 cal by passengers (48 000 persons/hr), the total
amount of heat to be removed is about 1 mi11ion cal/km/hr. Assuming 10 °C
or 20 °C difference in temperature (inside vs. outside) and 1 m3 air remov.ing 0·3
cal, the total volume of air required to ventilate the above tunnel a,mounts to
,, \> \1:1 I 1
..~~(ly .I;.,
.. ~ -·' 1 ~ --:.~
-.:.... ~">. 9\ ' -•.
,._._~ ... ,~..:-~ ·7)
..,,.r:,JJ i.L."'\v
'"',;.
-,r -- ~

518 DESIGN OF TUNNEL SECTIONS

333 000 or 167 000 m3/hr, respectively.4· 44 Amounts of heat to be disposed of are
given in Table 4/XVI (EMH6).
Ventilation systems designed to remove excess heat are generally adequate to
provide for sanitation and dust control as well. 4 · 45

TABLE 4/XVI
The Amou,rt of Heat Generated in Urban Tunnels (per k m)

Heat due to auxiliary pJant

B ody Output Total heat
Description Traffic
heat of VC• Gen. per hr
density Traction Output
(cal) hicles
(cal) ( kW)
I heat
(cal)
(cal)

High·way traffic, 20 vehicles - 200HP 126 400 5 4 300 131 000

low volumes simultane- (light-
ously ing)

Highway traffic, 200 vehicles - 2000HP I 264000 30 25 800 I 290 000

high volumes, simultane- (light-
4 lanes ously ing)

Old subway, 9600 30000 200kW 172 000 14 12 000 214 000
shallow passengers
alignment per hour

Modern subway, 38 000 120000 800kW 688 000 210 180 000 988 000
deep passengers
alignment per hour

Ventilation shafts and adits used during construction can be made part of the
final ventilation system, as shown diagrammatically in Fig. 4/ 140. The air can
flow through lateral vent tubes under the tunnel and through longitudinal ones
formed in the voids under the platforms to-reach the station through pipes rising
inside the columns. In a reversed action the foul air can be exhausted in the same
manner. The distribution of air from shafts between stations requires less elaborate
facilities and can be effected through simple lateral galleries.

4
,u ARBUZOV, G . V.: Ventilaciya tonnefei Metropolitenov (Ventilation of subwa y tunnels),
Moscow 1950
us NOTE : For a good example of ventilation design see : DANDUROV, M. I. : Tonneli
(Tunnels) Gos. Transp. Zheleznodorozhnoe lzd., Moscow J 952
 5 19
LIGHTING

4.66. LIGHTIN G, AND NOISE CONTRO L IN TUNNEL S

46.61. Lighting

Tunnel lighting.requirements vary greatly and depend on whether the lighting

is provided for maintena nce purposes only or for operation al safety as well.
Listed in increasing order of importan ce: lighting in railway tunnels may be
limited to directional lights; navigatio nal tunnels and those with utility lines have
to be illuminat ed to facilitate maintena nce and access; lighting in highway tunnels
shall be consistent with traffic re-
quirements and, finally, undergro und c::
stations, workshop s and storage i
~~,.,-:::5000
-~~ t:
areas shall be illuminat ed as required
1~~ a)
for regular working conditions. ~~-~
Maximum illuminat ion in the last ~= ~ !nter,or ,!lumm3Lt0!7
~ ~
case varies between 1000 and 1500 lux, Out.Jide - ->-In tunnel 500 770
and in highway tunnels from 500 to
100 150
150 Jux. Because of the absence of 0 50
D,st.;nce(m)
glare, indirect lighting is preferabl e
in both cases, but particula rly so in
highway tunnels. 900001/IX
b)
Great contrasts in illuminat ion at 6000
tunnel entrances and exits (in day- 5000 Adjvst.;b!e sltades
~ i--.>,, . . . - ' - - - - - --
light) must be balanced gradually. ~ 4000
Daylight illuminat ion is equivalen t c::
~ 3000
to 90 000- 100 000 lux, which is .ti!
~ 7000
almost impossible and certainly
impractical to simulate because of
~ 1000
4 46
Outside -- - Tvnnel
prohibitive costs. • A number of ·so fOO 150
0
experiments have been conducte d o,sl3nce (m)
to establish transition s that are still
comforta ble for the driver's eyes Fm. 4/ 141. How to reduce
(Fig. 4/141). There are two methods illuminatio n intensity at hjghway
to achieve such transition s. In one tunnel entrances
method the funnel-sb,aped entrance
allows daylight to get into the tunnel as far as possible. (For example, an outward
expandin g parabolic -hyperbo loid shell roof, opening up toward the outside, has
been added to the new highway tunnel in Le Havre. 4 468
• In the second . method,
exactly contrary to the first, the daylight is kept out by means of adjustabl e shades.
This method was adapted in the case of a wide, 350 m (1150 ft) long subway

u s Travaux Febr. 1960 8-82

u1a Le Monde Souterrain April 1 960 61
... - - -- ~ --

520 DESIGN OF TUNNEL SECTIONS

along the road l,ading to the new Orly Airport__ H 7 The 90 000 lux of daylight il-
Juminatio n has 'oeen reduced in steps to 500 lux at the entrance and then further
reduced by 20 % over a distance of 40 m (130 ft) and b.y an additional 90 % over
another 10 m (33 ft).
Because lighting is indispensable in all undergro und areas, emergency lighting
must be provided in the subways and in the worksho ps. (Corrosio n of lighting
cables has posed serious mainten:m ce problems in old railway tunnels.)
A discussion of lighting details is nut given here as this would exceed the scope
of this work.

46.62. Noise Control

The need for noise control is a recent development and is limited at present
to highway tunnels with heavy vehicular as well as bicycle and pedestria n traffic.
Observations made in Franceu8 have indicated considerable discomfo rt for
pedestrians and maintenance crew whenever the noise level, due to continuo us
traffic and reverberation, has exceeded 100- 105 decibels. In order to diminish
the noise, a number of special resonator s have been installed. The effect of the
latter cannot be expressed numerically, however, because it was not the intensity
but the nature of the noise that was changed. This was achieved through sound
fractioni ng and through the elimination of reverberation, resulting in a type of
noise similar to that present on a street with tall buildings on either side.
The resonato rs employed consisted of 20 x 30 cm (8" x 12") plastic boxes,
10 cm (4") high, with 2 cm (3/4") dia. spouts at the bottom and were mounted
on aluminum ribs installed in the ceiling. 60 000 boxes have been installed in
a subway 350 m (1150 ft) long. In addition to eliminating reverberation and chang-
ing the nature of noises the resonato rs also have cut down noise intensity by
4 %-14%.

4.67. ANCILLA RY WORKS

Installations required for the operation, ventilatio n, drainage and lighting of

tunnels have been dealt with earlier under the appropri ate headings and will be
discussed further in connecti on with subway stations (see Section 63.34). In this
article the discussion will be limited to portals (to be provided at tunnel entrances
and exits) and to niches.

,.., La traversee routiere de l'aeroport d'Orly, Travaux Fehr. 1960 91

Travaux Fehr. 1960 80
4 -~ 8
 ANCILLARY WOIU.~ 521

46. 71. Safety Recesses

Refuge niches or recesses in the tunnel lining have to be provided for the safety
;:>f maintenance crews. Staggered ,along the two walls, they should be spaced not
more than 50 m (150 ft) apart in railway tunnels. They can be constructed also
in combination with drainage laterals, in which case additional space bas to be
provided for collector sumps. Covered by underpitch vaults the recesses are about
1·90-2·25 m (6'-3" to 7'-6") high, 2·0-2·5 m (6'-6" to 8'-2") wide and 1·0 m
(3'-3") deep (Fig. 4/142) and may be left unlined if carved into solid rock.
Utility rooms, spaced about every 1·0-2·0 km (about every mile), are also to be
provided in long tunnels for the storage of materials and equipment as well as
for the convenience of the maip.tenance crew. These rooms are about 3·04·0 m
(10-13 ft) wide, 2·8-3·0 m (9-10 ft) high and 3·0-8·0 m {1 - 026ft) deep and are
equipped with a telephone. In most cases the entrances are whitewashed so they
can be seen more easily in the dark.

Chamber
Recess

FIG. 4/ 142. Recesses and utility rooms

46. 72. Portals

More important and with a structural function to perform are the portals.
Their functions are:
1. To protect and support the adits, exits and approaches under masses of
earth;
2. To keep out and drain surface water running down the front slope;
522 DESIG N OF TUNNEL SECTIONS

3. To emphasize the structu ral significance of the tunqel throug h

archite ctural
features.
From a structu ral point of view portals may be divided into the
following
categor ies:
(a) In solid rock there may be no need for a portal facade at all and
the tunnel
lining may be extend ed and left exposed at the approa ches.
(b) If there is any danger of rolling rock or surface water seepage
then a facade
wall has to be provide d around the entire openin g (Fig. 4/ 143).

Stope drain
Approach re-
l.ai111ng W3/I
(counte rfort
to f,;c,;de)

l'/'eeping
holes

FIG. 4/ 143. T unnel portal

 ANCILLARY WORKS 523

(c) If there is earth pressure to be expected from the front slope then the
portals have to be designed as retaining walls. Any retaining walls lining
the approach cuts can be considered as counterforts supporting the portal
(Fig. 2/32).
Surface water running down the front slope should be intercepted behind the
portal and drained so as to eliminate any seepage into the tunnel.
Because of differential loads, potential settlement (soaking of the base) and
forward tilting (shifting pressures) it is advisable to separate the portal from the
tunnel lining by an expansion joint.

REFERENCES

ANDREA£, C. (1949): Problemes d11 projel el <le l'etablissemenl de grands sou terrains rouners
tra11salpi11s. L eemann, Zurich
ANDREAE, C. (l 956): Gebirgsdruck und Tunnclbau. Schweiz. Bauzeitung 107
BEYER, K .: Die Statik i111 Eisenbetonbau. I. Spr in ger, Berlin 1930
CLARKE, N. W. B. (1967): The loads imposed on conduits laid under embankments or ~alley
fills, inst. of Civ. Eng., London, Jan.
(DAvwov) J(aauooa C. C. ( 1950): Pac'lem II Jlpoe" 11111poaam1e "011cmpy"111111 (Computation
and design of underground structures). roe. M3AaT. C Tpo1n. JlHT. MocKsa
GAYLORD and GAYLORD (1955) : Structural Design. Mc G ra w Hi ll. Civ. Eng. Series
GARTNER (1958): S1a1ically indeterminate Strucrures. Concrete P ublications Ltd., London
KLEINLOGEL, A. (1925): Roh111e11formeln. Ern°st, Berlin
LANE, K. S. (1957): E ffect o f Lining Stiffnes on Tunnel Lining. Proc. TV. int. Conj. Soil
Mech. London, fl , 223
Metrogiprotans ( l 959): H ay'IHo ,11emoo. pa3pao. pac<1ifm 11101111, 06iJe.w1-.· (Scienti fic method for
the calculation of flex ible tunnel linings) Moscow
MEYER HOF, G. G. ( I966): Composite D esign o f Sha llow Buried Steel Structures. 47. Annual
Convention of the Canadian Good Roads Ass. Halifax
RoHNE, E. (1964): Ober die Liingsliiftung von· Autotunneln mit Strahlventilatoren, Sc/nv.
Bauzeitung, 48
(SANSIEY) Ca11cuea ( 1955): Cmamu'lecf;ue pac,1iim o6iJe.lO" mo1111e,1e1i (Statical computation
o f tunnel linings). rwAp03HepronpoeKT MocKBa
TEODORESCU, P . (1961): Calc11l11l tu11elelor. (Computation of tunnels). Editura T ransporturilor
si Telecommunicatiilor, Bucharest
(VOLKOY) BO./f;Ofi, B , JI. (1945): ToHHe,111 (Tunnels), roe. Tpa11cn. )l(ene3HOAOPO:lKHOe vh-
AaT. MOCKBa
WARD, N. H . and CHAPLIN, T . K . (1957): Existing Stressen in Severa l Old London Under-
grou nd Tunnels, Proc. IV. int . Conj. Soil Mech. London , JI. 256
(ZURABOV and BOUGAYEVA) Jypaooa, T, r. ll EyzaelJa, 0' E. ( 1940): Be3H0/10f)HUe 211iJpo-
mex11u<ie31-.·11e mo1111e. 1u ( Hydra ulic tunnels without water pressures). roe. lh,naT. CTpOHT.
Jl1-1T. M OCKBa
(ZURABOV and BoUGAYEVA) 3ypa6os T. r. l! E yraesa 0. ( 1962): r uopome1111ecwe To11-
11e.111 (Tunnels for hydraulic power p lan ts) roe. E tt. l13AaT. MoCKBa
- - - ~ ------~- -- --

CHAPTER 5

TUNNEL SURVEYING

Tunnel surveying is one of the most difficult chapters in engineering geodesy

and differs from surveying on the surface in many respects. The problem 1s one
of achieving a high degree of accuracy while measuring in underground areas,
impeded by constructio~ work, impaired visibility and limited access conditions.
A high degree of accuracy is necessary because the layout of tunnels is always
based on the least favourable geometric configuration of the open traverse orien-
tated at one end.
Measuring becomes even more difficult if the azimuths have to be carried
through -air locks into the area under pressure. Another problem with low lying
tunnels is the fact that the azimuth has to be dropped down shafts that are 3o-40 m
(l00-130 ft) deep, with optical plumbs, or with mechanical plumb lines and thus
even the 'fixed' points will be moving, at least for a period of time.
Marking of th½' stations is also different from that on the surface, as this is
done ov.e,rhead. The transit thus has to be centred below rather than above the
stations.
All these factors have an adverse effect on the closing error and requfa~ the
surface triangulation, serving as a basis for the tunnel layout, to be all the more
accurate.

5.1. SURVEYING FROM PORTALS

5.11. HORIZONTAL LAYOUT

'funnels starting from the surface can be laid out directly from monuments of
the surface triangulation system.

51.11. Layout of Short Tunnels

In most simple cases short tunnels can be laid out similarly to straight lines
between two points that are not intervisible, as illustrated in Fig. 5/1.
Consider the tunnel starting at Kand ending at V, which are also the two points
to be established ·in line with A and B. In the first step two poles, both visible
from A as well as B, are tentatively -located. Point 3 is then lined up from B with
 HORIZONTAL LAYOUT 525

[ntr3nce termmJ/
A
--
- - -'- , K --
.. , (I

FIG. 5/ 1. Layout of short tunnels by successive approximations

_.,._
f

l_k:-.;;--=f_~-=2--J'
o.. r~~~~~-...:.-===-=-~=~~.. ,
A K V 8
Fro. 5/2. Layout of short tunnels by successive approximations and with an auxiliary range pole

,.j =--- - -L---+----L - _______ fl

-- -- -::--.. !J' ' --- --- - -

8 --,.--- . ~e - -- p
_ tt--~
FIG. 5/3. Layout of short tunnels by angular measurement

B 2 and then point 4 is lined up from A with A3 • Once the two poles on the hill
are in line both from A and B, the line has been established.
If the two poles cannot be located so as to be visible from both A and B, a third
pole has to be placed between them so as to be visible from A and B. In this case
(Fig. 5/2) pole No. 1 is located to be visible from A and B and poles 2 and 3 are
lined up with A 1 and Bi, respectively. Pole I is now lined up with 2 and 3 and the
procedure repeated as before until all poles are in line. Finally K and V are lined
up with A 1 and Bi, respectively. Permanent markers shall be placed at A, B, K
and V so as to be readily accessible throughout the construction.
The above layout can also be obtained with chaining and angular rotation.
Once again points K and V (Fig. 5/3) have to be lined up with the base line as
 526 TUNNEL SURVEYJN (;

defined by A and B. Having chained distances a and b (Fig. 5/3) and having
meas-
ured the angle e, distance s can be calculated from

(5.1)

which, being an approxi mate formula, is valid only for (a + b) ~ AB

and
sine~ e. By repeating this procedu re point P can be located exactly. Because
of the chaining involved this method is more cumbersome than the two previous
ones. Chainin g can be avoided by measuring the-distance optically.

Sl.12. Layout of Long Tunnels

Tunnels several miles long can be laid out by the above methods except that
transits have to be used in all line-ups.
The layout of long tunnels can be tied into existing triangul ation network
s
of national surveys, or can b_e based on special systems developed for the particul
ar
tunnel or on an indepen dent traverse.
51.121. Layout of long tunnels from existing triangulation network s of national
surveys. Let K and V denote two known points of the centre line of the
tunnel
on the ground (Fig. 5/4). The directions KV and VK have to be established
at K
and V, respectively.

F10. 5/4. Layout of the tunnel centre line on the ground between two points

First, the coordinates of Vand Khave to be established in the national coordin

ate
system, based on points A, B, C, D and £ , F, G, H using the method of
inter-
section. The angles are to be measured in two positions of the horizon tal
circle.
Either (a) azimuths, or (b) internal angles may be used in the procedure.
 HORIZONTAL LAYOUT 527

(a) In the method using azimuths the bearings to K and V have to be deter-
mined with reference to A, B, C, D and E, F , G, H , respectively. This can be done
as follows : Setting up the instrument at A the telescope is pointed toward at least
two known points {S, Q) and to point K. Let c5As and c5AQ denote the bearings,
which can be calculated also from the coordinates of A. S and Q :

and let /A s, IAQ and /AK denote the readings referring to points S, Q and K.
respectively. The angles of orientation . zAs and zAQ can be calculated from

and their mean value is

The bearing of the line AK is then

The other bearings can he cierived in a similar manner for Kand V from the other
fixed points. Using more than two points or measuring the angles at K as well
as at V will result in redundant measurements, that can be used to adjust the
coordinates, and so will improve their reliabi\jty.
{b) Internal angles may be used where a lesser degree of accuracy is satis-
factory. In this method only two points are used, forming a triangle with the point
to be determined, and the angles of this triangle are measured only at the two
reference points. (Angle /3 at Band y at C, see Fig. 5/4.) Distance BC can be cal-
culated from the coordinates and the other sides of triangle BCK are then

sin/3 siny
b =a - - and c =a - . - ,
sina Slll<X

and the bearings of BK and CK are

where
Ye- Xs
c5 8c = tan- 1 -X v and br,8 = c5 8c ± 180°.
C - As
 528 TUNNEL SURVEYING

The coordina tes of point K are

or in another form (for checking purposes)

Yx - Ye+ BC sin Dex and Xx= Xe+ CK cos Dex·

Points Kand V may be determined by resection rather than by the above methods
if only limited accuracy is required.
The bearing of KV: (KV) = DKv is calculated from coordina tes:

Yv-Yx
(KV)= b
Kv
= tan- 1 - X --
-X ' (5.2)
v- K
and
(VK) = Dvx = Dxv ± 180°.

With the instrume nt set up over K the direction Dxv can be laid out using the
known points (bvK can be laid out similarly at point V):
Setting up the instrume nt at K the angles to A, B, C and D can be measured
and the mean angle of orientati on can be calculat.ed. Denoting the respective
readings with I,., /8 , le and ID the correspo nding angles of orientati on are

()KA - /KA = ZA ,
Dxs - IKB = Zs,
Dxe - lxe = ze,
{JKD - IKD = zD,

and the mean angle of orientati on at point K is

=zx.

The reading pertaining to the direction to be laid out is

Turning to point V the above procedur e is repeated and the reading per-
taining to the direction to be laid out is

lvx = DvK - Zy.

The national triangula tion network is often not sufficiently accurate for tunnel
layout work and thus the points should be checked before use. The problem of
accuracy is also discussed in Section 5.14.
 HORIZONTAL LAOUU'f 529

51.122. Layout of long tunnels from special networks. In cases where the nationa!
triangulation network is not sufficiently accurate, the layout should be based
either on a special triangulation net, or, in simpler cases, on a special traverse.
Regarding the use of special traverses there are two alternatives. In the first
alternative both terminals are known (Fig. 5/5) and the unknowns are the angles
¢ and tjl. In the second alternative only one terminal is known (K) and one angle
(¢); the other terminal (V) and the other angle (t/1) are to be determined (Fig. 5/6).
In the first case the distances t 0 , t 1 , t 2 , ••• , tn and the interior angles ro 1, w 2 , •••
. . . , wn are measured in the field . From Fig. 5/5

KVsincp _ ,1,. _ VT _ [t;sinp;)l';:;~

- - - - - tan 'I' - -- - . , (5.3)
KVcos<f> KT [t;cosPiJ):~
and with the notation used in Fig. 5/5

Po= 0, P1 = 180° - W1, P2 = W2 - Pi ± 180°, P,, = w,, - Pn-1 ± 180°,

FJG. 5/5. Tunnel layout

by triangulation
between two known
terminals

I
Fm. 5/6. Layout of the
other terminal hy
special oriented
traverse

34 Szec hy: The Art of Tunnelling

530 TUNNEL SURVEYING

and with <p known

(5.4)

where n = the number of the traverse stations.

In the second case the angles as well as the distances have to be measured,
as before. Then, first; the location of V along the traverse side is determined.
Considering Fig. 5/6 and Eq. 5.3 it follows that (with the same notation)

and the unknown x can be expressed as

[t; sin P&:;~- 1 - tan <p [t cos p;)l::;g- 1

X= - (5.5)
tan ¢ cos Pn - sin Pn

As in Eq. 5.4 the expression for the angle 1/1 is

(5.6)

Where, in the layout of tunnels 8-10 km (5-6 mil~s) long, the accuracy of
the national triangulation network is insufficient a special triangulation net is
required.
Once again there are two possibilities regarding the independence of this special
network.

FIG. 5/7. Tunnel layout by

adopting two points of
an existing system and
developing a local
network in between

In the first case at least two points of the network to be .developed are also
points of the national network and the line between these two has to be accepted
and adopted. Thus, in this case there is no need to measure a base line, but al!
the angles of the triangle have to be measured with a theodolite with readings
to the nearest l" and in at least four positions of the horizontal circle. It is best
to establish a chain of triangles along the centre line of the tunnel.
Consider the chain of triangles illustrated in Fig. 5/7 with points I and II
belonging t~ the national triangulation network and with points 1, 2, ... , 7 to
be laid out. The coordinates of these points !1re calculated in two steps. Assuming
first an arbitrary length for line II, for exan:tple, the lengths of all the other sides
can be calculated on the basis of the measured and corrected angles and so can
 HORI ZONTAL LAYOUT 531

all the coordinates of the stations. Taking the coordinates of point I as y i = 0

and xf = 0, the hypothetical coordinates of point II, yf1 and xf1 as well as all
the others ean be computed and marked with a prime so that they can be <listing~
uished from the actual coordinates. Now the ratios of actual to hypothetical
distances are calculated from

(5.7)

In the second step the actual coordinates can be obtained from

Yi= Yi + Cy y{
and
Yz =Yi+ Cyy~ and
X1

x2
= X1 + Cxxi
= x 1 + Cxx;
l (5.8)

y,, = y1 + Cy y~ and x 11 = x1 + Cxx;

Knowing all the coordinates in the local network the tunnel can be laid out
exactly as if working in the national network (see Section 51.121).
In the other case it will be necessary to measure a base line, and it is advisable
to determine the coordinates of at least one station and the bearing of at least
one side in the national coordinate sys-
tem. One side of the local network is s
usually expanded from the base line
through a base net. The length of each
side can then be calculated with the '< Qs::----,-a~----,.---,.Y _ _ _ _E--r--::-::0 i'
measured and corrected angles, and the
bearing for each side can be derived from
the one with the known azimuth. The
coordinates of each station can now be
calculated with reference to the station
whose coordinates are known in the na-
tional coordinate system. With the local FIG. 5/8. Tunnel layout
network established, the tunnel can be by self-contained triangulation
laid out exactly as if working in the
national network (see Section 51.121).
With the notations of Fig. 5/8 the actual closing error is

(5.9)

where i:: 7 and i::1 = the error in the angles and in the lengths of the sides, respec-
tively
11 = the number of sides.
532 TUNNEL SURVEYING

The sign of the first member is positive if </>; denotes the angle as illustrated at
station 1, and is negative if it is the angle supplementary to 360°.
Calculations for the chain of triangles illustrated in Fig. 5/8 are the same as
for a local traverse, because stations K, l, 2, 3, 4, 5 and V may be regarded as
the stations of a traverse.

Fm. 5/9. Tunnel layout by

self-contained triangulation
(direct method)

According to Fig. 5/9 the angle (Ct) required for the layout of centre line k - e
and the length of the tunnel, ke, can be derived directly from the chain of triangles :

~f sin{J - l:'.f sin (/3 + y)

tanix = - - '
fe cos (/3 + y) - hfcosp + kh

ke = L = J [hfsin/3 -.fe sin (/J + y)]2 + [fe cos (/3 + y) - hf cos /3 +kh]2 '

hf= ab s~n (bac) si_n (bca) si~ (def) s~n (gdf) s'.n (kgf)
(5.10)
. SIU (acb) Sill (bdc) sin (cfd) SIU (dgf) Sill (ghf)

5.12. MARKING THE STATIONS

Horizontal reference points on the ground are marked in the same manner as
they generally are in surface surveys. Exceptions to this are the points along the
centre line of the tunnel as these are also to be used as instrument stations. It is
also advisable to mark them with permanent monuments. Permanent monuments
are also recommended for auxiliary target points located along the extension of
the centre line. The monuments should be such as to be able to receive a target
in the form of an iron bar, while surveying is in progress. An example of a per-
manent type monument is illustrated in Fig. 5/ 10; the marker on the left hand
side was used ·in the construction of the Simplon tunnel. In the case of paved
surfaces the permanent markers may consist of a metal plug set in concrete and
encased in a steel box as illustrated in Fig. 5/ l0b.
Inside the tunnel the markers are usually mounted overhead. There are several
types in use; that of BORCHER is generally preferred because it is the best for direc-
tional layouts. The BORCHER staple has to be driven into the roof in an approximate
 LAYOUT OF DETAILS 533

iJ; 20

{i(J

Concrete lfQ
//){J

100 100 16?

a€6m~il,i,f,,~~+-. _r_

Reference ITl3rker
F10. 5/ 10. Horizontal reference point markers

Roof

Brass scale Rider with vem1er Drilled hole

F10. 5/ 11. Overhead type station markers

position so that the horizontal portion will intersect the line to be determined.
Having established the line a hole has to be drilled into the staple in the exact
location to receive the wire suspending the plumb bob (Fig. 5/ 1 I). The rider with
the vernier, illustrated on the left, works on the same principle.

5.13. LAYOUT OF DETAILS

In tunnels with a straight alignment the centre line is laid out every 50- 100 ft.
This can be done with a theodolite set up over the station last established and
backsighting at a point at least 3-4 times as far away as the next station to be
determined.
The telescope is then reversed about the horizontal axis, a BORCHER staple is
driven into the roof at the required distance and in the approximate location,
and the new station is marked on it. The telescope, still in the reversed position,
is now backsighted to the first point, then reversed and the point is marked once
again on the staple. Now a hole has to be drilled in the staple half way between
the two marks: this will be accepted as the new station. Keeping up with the
progress of construction the stations have to be spaced about every 15-30 m
(50-100 ft) along the tunnel. In dark he·a dings the cross hairs require artificial
illumination.
5J4 TUNNEL SURVEYING

jP111mb line
Slot.1

L
'l
~
I

Wnttesf;eef of
Fm. 5/ 12. Metal targets
!Jlackboard
pt1per

Fm. 5/ 13. Tunnel

layout dn a curve
by deflection angle:;

In the case oflong sights the permanent markers consist of metal targets mounted
on the roof. Cross-shaped slots have to be cut into the plates and illuminated
from the back to give a well defined sharp light on the face that is easy to sight on.
A black cross on illuminated opaline, or other easily visible figures with distinct
vertical centre lines also can be employed (Fig. 5/12).
Tunnels on horizontal curves can be laid out by deflection angles using the
following formulae (Fig. 5/ 13) :

• rx S
SIU - - ,= - , (5.11)
2 2R

s2
y= --- (5.12)
2R

The value of ~ should be chosen to suit the length of the chord. s (the length of
the tape).
 METHODS TO IMPROVE ACCURACY 535

5.14. METHODS TO IMPROVE ACCURACY

The degree of accuracy required in tunnel surveying is governed by the allow-

abie closing error. The accuracy of the survey will depend primarily on:
I. a reliable network of surface reference- points
2. careful and repeated measurements.
Better control over the tunnel alignment can be achieved by:
(a) driving pilot drifts ahead of the -construction work
(b} vertical test holes along the centre line (Fig. 5/14)
(c) in the case of twin tunnels, lateral connecting adits to facilitate the use of
closed traverses (Fig. 5/ 15).

Test llo!e

Steel pipe en-

cased in con-
-------- I crete (on tile
centre !me)
mnne/ i

FIG. 5/ 14. Check holes along the centre line of the tunnel

Tunnel

------r,_------·-·r;_-- ·-· ·-A

/ / / LaLeral adit
_ }/ / - or nead1nQ

__ 2 ___ ____ =J_~- ------ ---cB

P<1!'3//el tunnel or l!e,ding
FIG. 5/15. Directional adit connections

1. The reliability of the reference point network - as mentioned earlier - can

be improved by developing a local network of reference points. In this network ..
a relafr;e error less than 1/50 000 is expected in the distances. The reliability of
the network can be checked by measuring angles at each station to be used. The
536 TUNNEL SURVEYING

dependability of each direction may be concluded from the respective angle of

orientation and from its mean value, as determined by measurements taken at
the respectivt station.
[z;]);::j'.
z
i
and if
el= Z1 - z,
e2=z2-z,

en= Zn - z,
then the mean error of the angles of orientation is

The value ofµ; ~hould be less for long tunnels than for short ones. Ifµ;= ± I" ,
for instance, because of the sighting uncertainty of I", a lateral error of ± 5 mm
(0·2 in) may develop and under adverse conditions it may be as much as three
times as great, i.e. ± 15 mm (0·6 in).
2. In tunnel surveying the measurements should be taken very carefully and
it is most important that they be repeated. The latter is important not only because
there is less chance of a gross error occurring in this way, but also because the
greater the number of targets and measurements, the more errors will become
negligible, and the better the overall accuracy becomes. These measurements
should be taken with a theodolite on which readings to the nearest l" can be

Closing Errors of Famous Tunnels TABLE 5/J

Tunnel Horizontal Vertical

Length
Name Closing error (mm)
(km)

Gotthard 14·9 250 100

Simplon 19·7 202 87
Lotschberg 14·5 257 102
London {Underground) 2-5 63-5
Mont Blanc 11 ·6 13-5 200
Tauern 5·55 550 56
Hauenstein 8·13 50 10
Suramsk 3·99 130 40
Wasserfluth 3·55 50 1U
 VERTICAL LAYOUT sr
estimated, and centering control devices should be employed, particularly for
short distances. This will be discussed later in detail. Subways constructed in
open excavations can be laid out in the same .manner as other structures on the
surface. The closing errors of some of the better known long tunnels are listed
in Table 5/I.

5.15. VERTICAL LAYOUT

In tunnels starting from portals the vertical layout is done by levelling as on

the surface, except that the benchmarks are established in a different manner.
It is essential, first of all, that the elevations of both terminals (Kand V) should
be referenced to the same datum, or that there should be a benchmark near the
tunnel, tied in to the national survey, from which levels can be run to both ter-
minals. It is advisable to check the two points by differential levelling even if the
elevations of both points are known. These benchmarks shall be fixed by perma-
nent markers, types of which are illustrated in Fig. 5/ 16. Drilled-in piles with

Iron knob

''
''
so-roocm ''
l:
I '

170-200cm
40
20 -

FIG. 5/ 16. Types of benchmarks Plug

Wa/J
¼

precast caps can serve this purpose very well under favourable soil conditions.
The advantage of piles over precast stones lies in the deeper embedment.
In the headings the gradients are marked temporarily by nails driven in the
timber posts about 5 feet above the bottom. Gradients should be checked over
once again before the commencement of the lining operations. After the lining
has been completed permanent benchmarks can be installed; i.t is best to use
plugs for this purpose (Fig. 5/ 16). These should be located in such a manner as
to allow the levelling rod to rest on top. The levelling rods used in the differential
levelling shall be graduated to one-hundredth of a foot and shall be adequately
538 TUNNEL SURVEYING

illuminated (portable floodlights). The use of levelling

glass plate, such as that of CsETI illuminated from the
back may be considered also inside the tunnel (Fig.
5/17).
Some of the new levelling rods have scotchlite type
graduations and numbering to reflect the incident light.
These rods are very easy to read even in darkness, with
a flashlight mounted on the telescope.
The mean closing error of the vertical layout is

(5.13)

where µkm the mean levelling error per km

L = the distance between the benchmark and
the point where the bore is holed through,
in km.

If the mean error of the reference benchmark is

known (µa) then the total closing error is

FIG. 5/ 17. Levelling (5.14)

glass plate
(by CSETI)

5.2. LAYOUT OF TUNNELS STARTING UNDERGROUND

5.21. HORIZONTAL LAYOUT

Subways in general can be divided into two groups, depending on their location :
(a) subways directly beneath the road surface
(b) subways deep below the surface.
Construction of the latter begins with the sinking of vertical shafts. Both the
shafts and the tunnel itself are laid out from numerical data, i.e. geodetic coordi~
nates which should be listed on the construction drawings. It follows, then, that
a geodetic network should already be available at the time of:the design. Because
of groundwater conditions the shafts very often have to be sunk under pressure
(pneumatic caissons). In such cases a steel air-lock, about 0·90 m (3 feet) in dia-
meter, is the only connection to the surface. Muck is raised and building materials
lowered through the shafts, which also have to provide for access and for dropping
down the lines. These shafts - eventually to serve maintenance and ventilation
purposes - are spaced about 600-1200 m (2000--4000 ft) apart and are connected
to the tunnel and to the shield chamber through ventilation adits about 30-70 m
(100-250 ft) long (see Section 5.21 and Fig. 6/ 150). A full-time survey super-
vision and direction is required during these construction phases, particularly
with shield driven tunnels.
 LA.YOUT STARTING UNDERG ROUND 539

52.11. Surface Reference Net

52.111. Types of reference 11ets. Horizontal surface reference nets for deep
tunnels may consist of:

(a) local precision traverse (self-contained)

(b) local chain of triangles
(c) precision triangulation net covering the entire area.

(a) LAYOUT USING SELF-CONTAINED PRECISION TRAVERSES. If a precision trav-

erse on the surface is to be used in the underground directional layout then
the angles of the traverse should
be measured with extra care, be- Portal
100' a a'
cause any angular errors will
have their' full effect on the trans- A -- ~ - ~ B,. \
. ex \
fer of directions. On the other ex' ~ - :_
hand, any errors in the coordi-
nates or in the lengths of the \ ~
sides of the traverse will result \ \. \
o nly in certain shifts in the un- \ \
derground ar~a (Fig. 5/ 18). Preci- ~ i
Cz CC,
sion traverses may be accurate
enough to be used for tunnel FIG. 5/ 18. Offsets in the underground
triangulation net due to errors of
sections less than l km (1/2 mile) angular measurements and centering
long. Theodolites with readings
to the nearest l" and distance
measuring bars with an invar base
must be used in precision trav-
ersing. Distances have to be 1
measured with the bars held ho-
rizontally (for further details on
self-contained traverses see Sec- 6
tion 51.122). F,a. 5/ 19. Layout by self-contained chain .of
(b) LAYOUT USING LOCAL triangles and intermediate headings,
CHAIN OF TRIANGLES. A local starting for intermediate ~hafts
cliain of triangles should be
used in the layout of long subway tunnels. Although this has been dealt with
earlier (Section 51.122); it is supplemented here with a particular case in which
the construction of the tunnel progresses not only from the terminals but also
from intermediate shafts (Fig. 5/ 19). Angles ¢ and i/1 should be calculated
o nly from the respective polygons and sections of the network, e.g. <P and i/1
should be determined using the angles IXi, IX 2 and Pi, /3 2, respectively.
(c) LAYOUT FROM A TRIANGULATION NET COVERING THE ENTIRE (URBAN)
AREA. Soviet specifications allow triangulation stations of the first and second
 540 TUNNEL SURVEYI NG

order to be used in subway layouts; those of the third order may also
be used if
the error in the length does not exceed 1/50 000.
In his study on the survey of the Warsaw Subway z. KowAL CZYK 5 1
· )las noted
that station s of the City triangulation system can be used only
if working in
open excavations; subways deep below the surfaee require either local
triangula-
tion or precise traversing. There is no doubt that not all City triangu
lation systems
measure up to the higher standar ds of subway surveying and therefo
re should
be checked carefully before use. According to Soviet specifications
the errors in
the measurements of distances should not exceed 1/50 000 throug
hout the net.
In the layo ut of the Budapest Subway the existing City triangu
lation net was
used because it was found that the above requirements had been met.
Direct precise distance measurements - carried out by K. 0LTAY 5 2
· to deter-
mine the reliability of the stations - have shown the errors in
the distances
between triangulation stations of the lowest order to have been about
1/45 000.
This would indicate that in general these errors would be between
1/45 000 and
1/100 000 as far as the stations of higher orders are concerned.
52.112. Refining the reference point net. For the purpose of transfe
rring direc-
tions underg round two or three high signals are required in the vicinity
of each
shaft and at least one marker on the surface in the immediate vicinity
of the
shaft.
The signals (extensions to chimneys, towers, etc.) should be always
of solid
construction and should also serve as instrument stations. Markers on
the ground
are usually plugs embedded in concrete or encased in steel boxes
(Fig. 5/11).
Markers located on the ground around the shafts are bound to
be di!;located
by construction equipment; in such cases they should be installed
on a concrete
block about lm x lm x lm (3 ft x 3 ft x 3 ft). Even if the monum ent
cannot be
preserved throughout the construction, a tempor ary monum ent bas
to be erected
over the steel box to facilitate the droppi ng of the line down the shaft.
All the above reference markers around the shafts should be located
keeping
in mind that:
l. As many of the high signals as possible should be intervisible
{Fig. 5/20);
2. The marker on the ground should be within the triangle formed
by the
signals ;
3. While signals in general should be located near the centre line, there
should
be at least one outside the area that is subject to settlements;
4. One of the signals should be located in the direction of the progres
sion of
constru ction (Fig. 5/21).

5 '1 KOWALC ZYK, Z.: Problems of Closing in the Construction of the Warsaw Subway.
1952 Warsaw ,
5-~ OLTAY, K. :
A budapes ti varosme res haromsz ogelesen ek hosszme ghataroz asaban
pontossag (Accura cy of linear measure ments in the triangul ation of elert
the Budapes t city survey)
Geodeziai Kozlony 1941 2
 LAYOUT STARTING UNDERGROUND 541

F10. 5/20. Recommended

layout for horizontal
reference points ,

FIG. 5/21. Recommended

layout for signals
to facilitate the
transfer of directions
underground

The high signals should be referenced by at least 5 angular tie measurements.

There should be at least four uniformly diverging directions selected at each refer-
ence point and the angles should be measured in four positions of the horizontal
circle, using a theodolite with readings to the nearest O·l". Directions are to be
measured similarly from the station to be referenced, sighting at the reference
points as well as at the signals of the next shaft.
The calculated bearings, as derived from the readings taken with the various
positions of the horizontal circle, should agree within 3", and if they do not,
then the respective series must be repeated.
In general, measurements should be taken early in the morning and late in
the afternoon. Sightings should be kept clear of heat-radiant buildings and chim-
neys so that errors due to refraction may be avoided. Whenever possible, the
instruments should be set up on monuments. To facilitate precise centering and
sighting, centering control plates should be used with three grooves and with
a precision signal. The plates are to be fastened to the monuments with plaster
of Paris. Tripod instrument bases are also often used on monuments (Fig. 5/22).
(For a further discussion of centering control plates see Section 52.13.)
In order to interrelate the indirect measurements (adjustment of coordinates)
the unknowns should be independent and should be grouped in such a manner
as to include the signals of only two shafts at a time (Fig. 5/23).
Each signal will thus have two sets of coordinates. In order to avoid confusion
it shall be recorded clearly for each coordinate as to which adjustment group
and to which tunnel section it refers. Because of the absolute uniformity of layout
 542 TUNNEL SURVEYING

Base plate

FIG. 5/22. Instrument

base for precise centc•ing

- - - - --
----- -,r : :. .:_- r:entre lme -:C:\ _
' . ---- . .
,/A-- .J_-!!..~.l!J.U!!_ll.!.L----r
/ .
.
/ /

\ ,....
/ ' \ .,
_,,,,...--;
--------
,,
\ /
_ _ _ .,,x... ._
/ __ _ _ _ _ ----- ,,,/'

FIG. 5/23. Selection

0roup /. Cr oup II. of groups of adjustment

of each section between two adjacent shafts, the above method is bound to give
the most reliable results. Also, with a . proper network, correct measurements
and calculations, the discrepancies are bound to be very small.
The calculated mean errors will refer to the stations within that particular group
and will give an indication of the closing errors to be expected between the respec-
tive tunnel sections.
The angles of deflection between the adjacent adjustment systems can be checked
directly by measuring the angles at the signals as follows:
Starting at signal Bi, located at shaft B, the angle between the sightings to
points A 1 and C 1 , located at the adjacent shafts A and C, respectively, is to be
measured first. This angle should agree with that calculated from the commonly
adjusted coordinates of A 1 , B 1 and C1 . The same measurement and computation
can be carried out for the other directions. The arithmetical mean of the differ-
ences between measured and calculated angles will represent the deflection
between the two tunnel sections. This, from a geodetical point of view, is inevit-
able. The deflection - only a few seconds _:_ may not be noticeable at a shaft or
shield chamber, but may be noticeable in the case of two headings advancing
from opposite directions. For this reason each coordinate should be used only
for the tunnel section to which it refers.
Provided that the. difference between two corresponding coordinates - adjusted
in two independent groups - does not exceed a few hundredths of a foot, it is
sufficient to consider the arithmetical mean of the two values in subsequent cal-
culations and measurements. A maximum difference of 7 mm (about 9/32") has
been found between any two sets of coordinates of signals of the Budapest Subway,
and the mean error of the coordinates has not exceeded 5 mm (about 3/ 16") for
any station. The difference between measured and calculated deflection angles
was about 5 seconds.
 LAYO'UT STARTiNG UND1'RGROUND 543

Reference points at grade at each shaft are derived individually from the signals,
using a single point for the adjustment. Measurements are to be taken in two .posi-
tions of the· horizontal circle, with the insttumen~ set up over the point at grade
and :under the· signals.
The sign_als of each shaft should be checked annually as part of a comprehensive
plan and programme. Such a systematjc rep~tition of measurements is neces~ary
to facilitate a check on possible movements and also tQ· improve the accuracy in
general.
As mentioned earlier, ground movements due to subway construction constitute
a significant source. of error for_the surface reference point network (Section 73.2).
In the case of the Warsaw Subway there is one traverse over or near the ~ntre
line and two others running outside the area subject to settlement. Space·d about
500 m (1500 ft) apart they are interconnected by lafetal traverses every 500 m
(1500 ft). For settlement zones wider than this a chafo of triangles is required.
In the case of ~e Moscow Subway the travei:ses crossing over settlement zones
are run between triangulation stations. According to the specifications the width
of the ~ettlement zone may reach four times the depth of. the tu.nnei and must be
considered as such.
From the point of view of sur-
veying there is a conflict regarding
the requirements for the location of ,..,..,~.;::;,-,...J..,.....,...,,-,.-,...,....,,...,,....,..., Building
reference points.. On the one hand it
is advisable to locate the :QOints close
to the centre line of the subway so
that -any errors in the coordinates
have the least e.ffecti but on the
other hand it is also important to P/111 l{/pe,~f/e.'liPJ1Nf : -:z-;V.ikl -· -J- ·
keep them away from the centre line
because of inevitable settlement. In Fm. 5/24. _Locii~ion of-benchmarks in a group
view of the above the following ofbuiidirigs,
considerations should be remem-
"t?ered when laying out signals :
I. One of the signals of each shaft has to be locate"d outside the settlement zone
(four times the depth' of the tunnel).
2. The movements of all buildings carrying signals should be checked consta_ntly
by precise levelling. At least three benchmarks should be installed in such
buildings .(Fig. 5/24) and in such a manner .as to facilitate a check on the
movements in two perpendicular planes.
3. The elevations of the points that have been found to be unstable by the precise
levelling should be re-established from other original l?enchmarks of the
network. In the ·overall adjustment the points shall be divided into groups
including the points outside the settle~ent zone at every. shaft. Thus, the
only coordinates that will change to any appreciable extent and as a result
--==-=--=-=---- ~~- - - - - - ~
--
~ -

544 TUNNEL SURVEYING

of such repeated measurements will be those of the point that has been
dislocated.

52.12. Transfer of Directions Down Shafts

Undergrou nd surveys can be connected to the surface coordinate systems by
means of

(a) plumb lines

(b) optical plumbs
(c) gyroscopes.

Compass surveys cannot be used in the constructi on of subway~ partly because

of the limited accuracy (l'-2' error) and partly because of interference from the
steel structures and electric cables that are always present.
52.i21. Plumb lines. Plumb lines should be used whereyer a high degree of
accuracy is required. Plumb bobs, weighing 10-25 kg (25 to 60 lb), are suspended
usually 01;1 high strength steel (piano) wire, about 0· ~-5 mm in diameter, and
fastened to a reel. The reel is clamped to collar timbers about 2-3 m (7-10 ft)
above the shaft.
A vertical plane, with its direction
+X
determined on the surface, can be estab-
lished by means of two plumb lines
4 _4 4 to which the undergrou nd survey will
--9--o----
6
-.O be orientated (Fig. 5/25).
Underground traverse If the plumb lines can be centc;.red
Transfer fer directions on, the deflection angles - both on the
surface and undergrou nd - should be
F IG. 5/25. Layout of traverses
measured directly at the plumb lines
with orientation by plumb lines
(F1 and F 2 ). If this is not possib~e,
then the angles have to be calculated .
The plumb lines should not sway while the operation is in progress. This can be
achieved positively only at the top where the lines are fixed. There are, however,
two ways in which to effect stability undergrou nd as well. One is to damp the
vibrations by immersing the bobs in liquids of high viscosity. In the other meth-
od the mean position of the vibrating line i_s determined in two perpendicular
planes and the wire clamped in this position. The cycle of swaying may be short-
ened by increasing the weight of the bob (as long as the operation does not
become unwieldy).
Stabilizing fins, attached to the bob, are most effective in damping vibrations
in viscous fluids (Fig. 5/26).
An instrumen t for determining the amplitude of swings is illtistrated in Fig. 5/27.
This instrumen t has two perpendicu lar scales, etched on.glass (marked I and II),
 LAYOUT STARTING UNDEROROUND 54S

Fm. 5/26, Reel for housing / T/Jumb screw

and handling
plumb-bob wire
with plumb bob

--- ---·---·
light box and
damping pail at
bottom

RJi/ t,//ed With 0/t

Level(~

'Mirror 181-,,, ,..,.,1,.,,0

Mirror
0
f
2-41
,,

FIG. 5/27. Instrument

for dropping survey lines
(developed in Sopron)

JS Szechy: The Art of Tunnelling

546 TUNNEL SURVEYING

an adjustable mirror, a hooked end (o) that

can be adjusted in two perpendicular direc-
tions and a base plate with three foot screws
and a level (e).
With the wire swinging in front of scales
I and II, the amplitude can be read directly
on scale I and with a mirror at 45° on the
lil.strument other. The two readings cannot possibly be
ba.sewJt/7 confused because a reading of 30 is the maxi-
clamp for
plumb mum on scale I and the minimum on scale II.
c::lgl:::=il~::i:;;::m w,re Swinging the wire first in front of scale I
and then before II, the extreme positions
are to be recorded for 5 cycles on each scale.
In order to improve the accuracy at least 3
sets - or in exceptional cases 6 to 10 sets - of
measurements shall be taken.
As mentioned earlier the angles at the
plumb lines should be measured directly
both on the surface and underground, if it
is possible to centre over the plumb lines.
An instrument for measuring the angles at
the surface, devised by W1LSK1- HoRNOCH, is

Pin marker

Box /eve/
Fm. 5/28. Instrument for
dropping survey
Jines down shafts
(by WILSKI T .-HORNOCH)

Fm. 5/29. Instrument for

centering the
pin marker
 LAYOUT STARTING UNDERGROUND 547

illustrated in Fig. 5/28. It consists of an instrument base with an adaptor

for clamping the wire, and two pairs of cones. The ·mean position of the plumb
-line can "be determined underground as described above and the plumb bobs
<:an then be replaced by a pin marker. This can be done with the instrument
shown in ·Fig. 5/29, which is also supposed to receive the theodolite. The pin
marker can be adjusted by sliding plates e and g along the interJying disc c. It is
suffi.cient"to centre to the ·nearest 1- 2 mm (about 1/ 16") because the marker can
be turned through 180° and the necessary ~djustment can be calculated as the
difference between the arithmetical mean of the two opposite readings and the
reading referring to the proper position. Next, the pin marker is to be replaced
.b y the theodolite, which can be centered with the help.of the three grooves.
In the construction of the Budapest Subway the plumb lines have been laid out
in an eccentric pattern. This method does not require special" equipment and is
also the most accurate. The two plumb Jines are to be lowered as before, with
a vertical plane approximately in line with the respective sides of the adjoining
traverse (both on the surface and underground). An auxiliary point is now to be
chosen on the surface and another one underground. These points will form two
triangles with the plumb lines: one at the top and ~me at the bottom. The auxiliary
points should be located so as to result in small angles between sightings to the
two plumb lines (Fig. 5/30). Also, the signal near the shaft should be v~ible and
should be sighted from the auxiliary point at the surface. While the angles are
being measured at the auxiliary point on the surface, the bearing of the line
connecting this point with th~ signal should also be determined by sighting from

,x.

·SMft

t -~
\ surface point

F10. 5/30. Layout for the

eccentric set-up of plumb
Jines +.'I

35•
548 TUNNEL SURVEYING

the latter. This procedure has to be repeated every time the plumb lines are used,
regardless of whether the coordinate s of the auxiliary point are known or not.
Consider the triangle AF1F2 (Fig. 5/30) on the surface and the other one BF F
1 2
undergrou nd. It is assumed that the angles </>i, </> 2, cxi, Pi, Yi, cx2, P , Y2, I/Ii, 1/12
2
have been measured as have the distances ai, b1 , c1 on the surface and a , b , c
2 2 2
undergrou nd, and it is further assumed that the coordinate s of points A and Care
known. The coordinate s of B and the bearing of (Bl) can be calculated then as
follows:
(Bl)= (AC)+ </>1 + 180° + P1 + 180° - cx2 + 180° + 1/12 - n 180°, (5.15)
Ye= ¥:1 + a1 sin (AF1) c sin (F1 F 2) + b2 sin (F2 B) , l

where
Xe= XA + a 1 cos (AF1) + c cos (F1 FJ + b2 cos (F2 B) J (5.16)

C1 + C2
C= - - -
2
As a check the bearing (Bl) and the coordinate s of B can also be calculated
another way:
(Bl)= (AC)+ </>2 + 180° - cx1 + 180° + P2 + 180° + 1/11 -
l
n 180°, (5.17)

Ye = YA + b1 sin (AFJ + c sin (F2 F1) + a2 sin (F1 B) ,

(5.18)
XB = X,c bi cos (AF2) + c cos (F2 F1 ) + a2 cos (F1 B).
Working with two plumb lines the directional error works out to be about
8-10", but can be reduced if 3 or 4 plumb lines are used. In the latter case there
will be two connecting triangles.
Two survey teams are required for dropping survey lines down shafts. While
the first group (I technician plus 4 labourers) is measuring the sides of the con-
necting triangle, the other group (1 engineer - in charge of both groups - and
3 labourers) is measuring the angles of same.
The steps to be followed in the calculations are:

1. The connecting angles are:

cx 1 = arc sm c
. ( a1 sm
. Y1) ,

. { bi .
.Bt = arc sm le Sill Y1) ,

(5.19)
exz = arc sm c
. / a2 .
\ sm y2) ,

f3 2 = arc sm c
. ( b2 sm
. y2 ) •
 LAYOUT STARTING UNDERGROUND 549

2. The bearing of line Bl is computed as in Eq. 5.15

3. The coordinates of point B are calculated as in Eq. 5.16
4. The calculatiOJlS are checked as in Eqs 5.17 and 5.18
EXAMPLE. Consider the layout illustrated in Fig. 5/3 1. The coordinates of point A are
known YA= +184·092 m; XA = -2038·258 m and the bearings of the two sides of the

P-1 - - - - - - - - - - - ---+~!f_,-,_

X
Flo. 5/31. Transit set-up
outside the plane of plumb
lines

connecting triangle on the surface have been measured as (AF1) = 316° 41 ' 2·5" and (AFz) =
= 316° 39' 46·9 as have the angles at B: ifti = 191° 26' 53·3" and l{t2 = 181° 19' 32·4".
Other measurements are:

a 1 = 12·3056 m; b1 = 7·9217 m; c1 = 3·3843 m; y1 = l ' 15·6";

a 2 = 7"4760 m; b 2 = 11·8602 m; c 2 = 4·3845 m ; y 2 = 7' 21·0";

C =
C1 + C.• = 4·3844 m .
2
The bearings (AF1 ) anc;I (AF2 ) have been calculated using the mean angle of orientation
at A.
1. The connecting angles as in Eq. 5.19 are

(x1) . ( 12-3056
= arc sm _ . 0° 01' 15·6" ) = arc sm
sm . 0·0010290 = 0° 3' 32·2",
4 3844

(Xl = 180° - 0° 03' 32·2" = 179° 56' 27·8" ;

fJ 1 = . (-
arc sm 7-9217
-- sin 0° 01 ' 15·6" ) = arc sin 0·0006622 = 0° 02' 16·6"·
4·3844 '

,.
550 TUNNEL SURVEYING

!l'. 1+ /31 + y 1 = 179° 56' 27·8" + 0° 02' 16·6" + 0° 01 ' 15·6" = 180° 00' 00" ;

a2 = arc sin ( ;:;;:~ sin 0° 07' 21 ·O") = arc sin 01)()36456 = 0° 12' 32·0";
~\~~~
2
{/3 2) = arc sin ( sin 0° 07' 21 ·0" ) = arc sin 0·0057836 = 0° J 9' 53·0",

/3~ = 180° - 0° 19' 53" = 179° 40' 07" ;

et 2 + /32 + Y 2 + 0° 12' 32" + 179° 40' 07" + 0° 07' 21 ·0" = 180° 00' 00"'.
2. According to Fig. 5/ 31 the bearing (Bl) ofline Bl is:

- 180° = -180°
(F1 A) = 136° 41 ' 02·5"
+/31 = 0° 02' 16·6"

+ 180° = + 180°
<F:F1 ) = 316° 43' 19·1 "
-':t2 = 12' 32·0"

(F:B) = 316° 30' 47· J II

-180° = -180°
(BF. ) = 136° 30' 47• I "
+t/12 = +191 ° 19' 32·4"

(Bl) = 327° 50' 19·5"

3. The coordinates of point Bare calculated from Eq. 5.16:

Y:!i = + 184·092 + 12-3056 sin (316° 41 ' 02-5" ) + 4·3844 sin (136° 43 ' 19·1 " ) +
+ 11·8602 sin (316° 30' 47·2" ) = 184·092 - J2·3056 x 0·6860212 + 4·3844 x 0·6855392 -
- 11-8602X 0·6881886 = + 170·494 ;

X 8 = -2038·258 + 12·3056 cos (316° 41 ' 02·5" ) + 4·3844 cos (136° 43' 19· J " ) +
+ 11·8602 cos (316° 30' 47·2") = 2038·258 + 12·3056x0·7275815 - 4-3844 x0·728035 7+
+ 11 ·8602x0·7255318 = -2023·892.
 ~OUT·. STARTING UNDERGROUND

4 The results are now to be checked by Eqs 5. 17 and 5. 18 :

(AF,) = 316° 39' 46·9"

+ 180° ~ + 180°
(F2 A) = 136° 39' 46·9"
-,xl = 179° 56' 27·8"

(F1F 2) = 31·6° 43' 19·1 "

-180° = -180°
(F1Fz) = 136° 43' 19· l
+ {3 2 = + 179° 40' 07·0"

-180° = -180°
(BF1) = 136° 23' 26· 1"
+t/t1 = +191 ° 26' 53·4"

(Bl) = 327° 50' 19· 5"

Y8 = +184·092 + 7·9217 sin (316° 39' 46·8") + 4·3844 sin (316° 43' 19·1 " ) +
+ 7 ·4760 sin (316° 23' 26·0" ) = + 184·092 - 7'9217 x 0·6862882 - 4-3844 x 0·6855392 -
- 7·4760x0·6897389 = +170·494 m;
X 8 = -2038·258 + 7·9217 cos (316° 39' 46·8" ) + 4·3844 cos (316° 43' 19·1" ) +
+ 7·4760 cos (316° 23' 26" ) = - 2038·258 - 7·9217 X 0·7273297 - 4·3844X 0·7280357 -
- 7·4760X 0 ·7240582 = -2023·892 m .

52.122. Optical plumb. Surface points can also be projected down shafts with
optical plumbs. Theoretically speaking mechanfcaf plumb lines could be replaced
by telescopes set in a truly vertical position; however this is not quite so simple
in actual practice. The accuracy to which the telescope can be set vertically is
rather limited to begin with, and the humidity and lack of illumination in the
shaft make sighting even more difficult. The application of optical plumbs in
subway construction is limited to temporary and insignificant applications.
The first optical plu mb was designed by NAGEL (Dresden). It resembled a theodo-
lite with the telescope set vertically and had a base allowing such sighting. After
centering the instrument over each point, the points are projected down the shaft,
marked, and then extended into an underground traverse. This procedure is
rather awkward and inaccurate.
 ssz TUNNEL SURVEYING

The instrument by ZEISS is an improved version of the optical plumb. The pro-
jector is once again combined with a vertical telescope, with a level and with three
levelling screws. It projects a vertical plane rather than single vertical lines (Fig.
5/32). A prism with a small angle of refraction, mounted in front of the objective
lens of the telescope, divides the field of view jnto two halves and the image
viewed through the prism is slightly displace~_

Z,
Ocular

Ver_
t1c:,I telescope

----p
=--- Fm. 5/ 32. Principle of
the optical plumbing

Owing to the fact that single points are not being projected in this method,
the effect of any error in the vertical setting of the telescope will be limited to
a parallel shift or a slight rotation of the plane. A double collimator with cross-
hairs is mounted on the telescope in a position perpendicular to its line of sight.
Also part of the equipment is a rod with three parallel lines marked in white
against a black base. The two white dots, also marked on the base, represent
the two plumb lines.
The rod has to be held horizontally underground in such a position that the
two white dots beco'me two points of the connecting triangle for the direction to
be extended.
The instrument should be set up on a solid pedestal above the shaft with th<'
spindle approximately centered over the centre of the rod. Once the prism is
placed in front of the telescope, another image of the rod will appear in addition
 LAYOUT STARTING UNDERGROUND 553

to the original one, offset corresponding to the refraction of the prism. As long
as the plane of refraction of the prism coincides with the vertical plane passing
through the centre of the rod, the image viewed through the prism will be subject
to a parallel shift only and the single line viewed through the objective lens will
lie between the two lines viewed through the prism. If the two planes intersect
at an angle, then the image viewed through the prism will appear in a laterally
displaced position (Fig. 5/33).
The telescope should now be turned around its spindle with the adjustment
screw until the lines are located as shown in Fig. 5/33. Another theodolite is set
up over point D in line with the horizontal collimator which has followed the
vertical telescope in its turn. While this is being done the focus should be set at
infinity on the second telescope so that the image of the cross-hairs of the colli-
Jnitial position After
wi!Jh no prism settiog
Image of base
line
Image of base

·~~9
line

--
FIG. 5/33. Images
as seen in the optical
plumbing
a . -·
pri_sm.

mator appear in the field of view of the telescope and coincide with the cross-hairs
of the latter.. Once the station of the theodolite has been located in the surface
network and the angle P (Fig. 5/32) has been measured, the orientation and con-
nection of the rod in the shaft can be completed.
It will be noted that the transfer of directions down shafts with optical plumbs
has to be carried out in two steps. First, the two points replacing the plumb bobs
are projected and then the connecting triangle formed and measured.
No matter how carefully the lines of sight of the prism and of the collimator
are adjusted they are never perfectly parallel. In order to eliminate the resulting
error the measurements should be repeated with the telescope turned through 180°.
The error due to imperfect adjustment can be eliminated by taking the arithmetical
mean of the two measurements.
While the above method is undoubtedly more accurate than that of NAGEL and
even though the instrument has been further improved in the Soviet Union, its
application in subw:ay construction is still limited to measurements of lesser impor-
tance. According to OGLOBBIN the refraction in the shafts - due to differences in
air temperature and humidity between surface and shaft bottom - is a significant
source of error and can hardly be eliminated.
52.123. Other methods of orientation. In the early stages of construction (shaft
sinking, layout of ventilation adits, etc.) and in measurements of secondary
importance certain simpler methods with limited accuracy may also be acceptable.
554 TUNNEL SURVEYING

1. In the geometrical method for determining the centre of break-out and the starting direc-
tion for the adit, after the sinking of the shaft has been completed, two vertical plugs, with
holes (I and ID, are to be installed at the top of the shaft and at opposite ends of a diameter
which is at an aogle of approximately 70°-80° to the centre line of the adit (Fig. 5/34). Points
K and L, i.e. the two opposite ends of the diameter coincident with the centre line of the adit
are to be established on the lining.

I
\ ' I
D1rect1011 of\ FIG. 5/34. Establishing
break-out \ the direction of the horizontal heading
.S/Jaft wall

B
,,.,,-,..•,, /.

i
ri!, O
I
I

FIG. 5/35. Checking the A

- Ill-
direction of shaft sinking

I
I
C!bC

Knowing the bearing of the centre line of the adit (<'l,n) and measuring the same for line
1- ll on the surface (<'l) the angle between the two lines is

(5.20)
13ecause
OA = OK = r,
according to the cosine rule

b = ..)2, 2 - 2r 2 cos 11 1 = fo ..j t - cos 11 (5.2 I)

 LAYOUT STARTING UNDERGROUND 555

which can be rewritten as • 1X1

b = 2rsm . (5.22)
2
Similarly
2 .
IX,,
c = 2rsin (5.23)

With the distances b and c determined, points K and L can be laid out.
The rotation of the shaft lining during construction can be investigated similarly with
the geometrical method. It proved particularly useful in the construction of the Moscow
Subway when working in areas under pneumatic pressure. Before sinking the shaft the centre
line of the adit is to be established on the surface with two points A and B (Fig: 5/35) and so
is a direction perpendicular thereto (points C and D). Permanent markers, such as cut stone
or metal plugs, should be used to preserve these points. In the next step points I, II and III,
IV are established on the outside face of the shaft lining, and points 1, 2 and 3, 4 marked on
the top of the working chamber, all lined up with points A, Band C, D, respectively. At the
end of the plumb lines 15x 15 cm
(6"x 6" ) steel plates are to be embedded
in the concrete lining above the cutting
edge. With the shaft in the correct posi-
tion the positions of the plumb lines are
marked on the steel plates and these
marks can then be used to check whether
the shaft is in a truly vertical position
during construction. Any rotation of the
shaft can be determined with an instru-
mentset up over point A (Fig. 5/ 36). The
angle IX and distance c can be measured
and the other data calculated

a=csin ,x (5.24) FIG. 5/36. Measuring the rotation of the shaft

and
a Revolving prism rd. e
f3 = arctg b . (5.25)

When laying out the direction in

which to break out from the shaft, the
direction 1-2 has to be adjusted by {3.
As the markers I-IV are about to dis-
L-'--c--±,--7"'-'i~I-
appear in the ground as construction
progresses, the ro tatioo of the shaft
Projected direction
should be measured with reference to the
old markers and new markers should be FIG. 5/37. Optical directional
mounted in the lining at the same time. transfer with prism adapter
Any subsequent rotation has to be deter-
mined to include the previous one.
2. Optical methods may also be adopted when transferring directions with limited accuracy.
Because of the steep sightings involved, the theodolite should either have an eccentric telescope
or an adapter consisting of a prism or a mirror that can be rotated in front of the objective
Jens about the horizontal axis (Fig. 5/37).
In either case the instrument has to be set up approximately over the centre of the shaft.
Since the plane to be projected will be only as vertical as the spindle, the latter should be made
vertical very carefully, using the level tubes. It should be noted that any error in the plumbin~
operation will result mainly in a parallel shift of the directional plane.
556 TUNNEL SURVEYil-iG

Separate, firm pedestals should be provided over the centre of the shaft hoth for the in-
strument and for the observer. The instrument is set up in the usual manner and a backsight
is taken to a reference point. Now the coordinates of the instrument station are determined
in the surface network and the bearing for the backsight is calculated. Turning the instrument
through an angle that will also be suitable for the connecting triangle, two points are projected
downward and spaced as far apart as possible. This is done either with the eccentric telescope
or with the revolving prism. In the first case the horizontal axis of the telescope has to be made
horizontal very carefully-with a mountable level if required-and in the latter the same ad-
justment is required for the axis of the· prism. For this reason the housing of the prism has
to be adjustable. If required, the adjustments can be made similarly to the correction of colli-
mation non-adjustment.
The above· methods can also be used in the pipes of air locks and with an accuracy lhat
is sufficient in the early stages of construction.
3. The gyroscope was developed by L. FOUCAULT (1819-1868), and the idea of orientation
with gyroscopes is more than a hundred years old, and is based on the fact that once the
axis of a gyroscope is set in the horizontal plane at the observing station it will turn in a north-
south direction.
Gyroscopes with a free axis-suspended through a universal joint-will retain their plane
of gyration regardless of the position of the housing. The use of gyroscopes in ~avigation,
aviation and underground orientation is based on this feature.
· Modern gyroscopes tun\ at 20 000 rpm and are practically free from friction because the
revolving parts are submerged in a fluid. Because their accuracy is limited to ± 60", their
use in subway surveying is limited to less important measurements.

52.13. Underground Directional Layout

Once the shaft siQking is completed, the shaft is to be connected to the adit
through a horizontal lock (Fig. 5/38). The connecting point (B) in the lock should
be located on a temporary monument, erected for this operation. Measurement
of the angles required for the connecting triangle is done with gate II closed.
With the theodolite left in place gate T is also to be closed and the pressure can
be admitted. This should be done gradually, even if there are no men in the cham-
ber, because sudden changes in pressure are liable to cause detrimental stresses and
deformations in the closed telescope. Gate II can now be opened and angle t/F
can be measured by sighting to point A in the adit.
Straight lines can be carried through air locks directly if only a limited deg1ee
of accuracy is required. The straight line passing through A and Bon the pressure-
free side is to be prolonged (Fig. 5/39) and the theodolite set up in the air lock,
joggled into line with A and B so that the images of both points should coincide
exactly with the cross-hair. Having equalized the pressures the inner door can
be opened, the telescope is plunged and points C and D can be set out . In order
to reduce the directional error the measurement should be repeated. The telescope
is kept in the inverted position and turned through 1S0° to sight to points A and B,
then plunged again to sight at C and D. The correct point is half-way between
the points set out by the two sightings.
As mentioned earlier, the station markers are mounted usually on the roof.
A number of types of markers are illustrated in Fig. 5/ 11.
 LAYOUT STARTlNG UNDERGROUND 551

One of these markers can be I

mounted on the roof and consists
A
of a graduated adjustable metal
rod with a hinge at one end. A
rider with a hole can be set to
receive the plumb line and the
operation is repeated with the Pressure
telescope in another position. area
The readin~ obtained are then
averaged and the rider set on
this avebge.
More often than not, under-
ground traverses are of the open Material
type and have to follow rather lock
undesirable alignments, full of
_sharp deflections and 1with long
sides alternating with short ones.
For this reason all angular meas-
urements should be carried out
with particular care and precision.
Even if the marking of the
points and the centering on the
.stations is done with the utmost
care, a certain amount of error
is inevitable, i.e. the. tatgets and
'i:he instruments will not be cen-
tered in exactly the same way
on the same station. For short
legs of the traverse even small
errors can result in considerable
errors in the angles.
Fm. 5/38. Method for carrying the sur vey
line through entrance locks ·:o the
horizontal adit

Fm. 5/39. Method for carrying the survey line through air locks
 0

TUNNEL SURVEYING

To avoid errors of this type a number of centering control devices have been
developed, the simplest form of which is illustrated in Fig. 5/40. It consists mainly
of a base plate with three grooves to receive the three foot screws of the theodolite
or those of the target in such a manner that the spindle of the theodolite and the
centre line of the target coincide exactly when interchanged (Fig. 5/40). The most
precise centering control in-
strument is the one with the
revolving socket, designed by
TARCZY. This version basically
represents an improvement
over an earlier arrangement in
which the spindle of the theo-
dolite had to be lifted out of
the bushing, thereby allowing
dirt to get in between the pre-
cision-finished surfaces. In the
case of the T ARCZY instrument
the revolving socket is put into
FIG. 5/40. Centering control plate the bushing of the spindle and
the spindle of the theodolite is
put into the revolving socket,
larget Target
as is· the spindle of the target
(Fig. 5/41).
Using this instrument cen-
tering can be carried out to
an accuracy of 0·01 mm
(appr. 1/2500" ) because the
target is sighted in two dia-
metrically opposite positions
and the mean of the two di-
rections will be free from any
centering error of the target.
Centering control devices
Enlargement of base are generally required where
traverses with short legs have
Fm. 5/41. Centering to be run. Each instrument
control instrument by TARCZY-HORNOCH
comes with two targets (Fig.
5/41). Because the station
markers are located overhead, the theodolite shall be set up beneath a plumb line
suspended from the markers.
In all other respects underground traverses should be run in the same manner
as those on the surface. Theodolites with readings to the nearest I" should be
used in all angular measurements and the targets should be illuminated from the
back. Linear measurements should be made with calibrated chains O!." ·in var tapes
 VERTICAL LAYOUT UNDERGROUND 559

and on horizontal pedestals prepared for this purpose. Layout procedures are
the same as those followed in a tunnel starting from the portal (Section 51.3).

52.14. PossibiJities for Increasing the Accuracy of the Layout

This problem has been treated earlier in Section 5.14 and the observations made
in connection with tunnel surveying can also be applied to layout work in subway
construction.
Undoubtedly the accuracy of the surface network can be increased by adjusting
the triangulation stations in a comprehensive manner and by checking by levelliog
the settlement of those buildings on which such stations are located. Similar im-
provements can be achieved by using centering control devices for angular meas-
urements on the surface as well as underground. Directions transferred from
the surface will be more reliable if the connecting triangles have been orientated
from the signal next to the shaft in question.

5.22. VERTICAL LAYOUT

The vertical layout of low lying tunnels begins with measuring the depth of
the shaft. The instruments required in this operation include two levels, two rods
and a steel tape 50-200 ft long, depending on the depth of the shaft.
First the tape is to be lowered into the shaft as shown in Fig. 5/42 with a weight
of 5-10 kg (10-20 lb) attached to the end.-A level is then set up on the surface
and another one underground, so as to be able to sight to the rod set up over the
nearest benchmark and to the tape, respectively. Once the swinging of the tape
has subsided, readings are taken on it with_ the levels, at the top and at the bottom.
These readings should be repeated two or three times and their mean values cal-
culated so that the difference in elevation between the heights of the two instru-
ments can be determined .
Denoting the upper and lower readings by 9'; and /01, respectively, the average
of the readings is

(5.26)

where i = the number of the readings.

The difference in elevation between the heights of the instruments is

(5.27)

The tape should be standardized for each measurement because of the changes
in temperatures and because of the pull. While being standardized the tape should
 ~GO TUNNEL SURVEYING

:._,,.Le,,e!lmg rori
FIG, 5/42. Direct
measuremen t of shaft
depth

l J
t ~
·~ T

_1 , f Levelnng rod
~w,~

be subjected to the same amount of pull as under working conditions. It is best

to establish the actual exact length (l) of the tape.
AU undergrou nd benchmarks have to be tied in to the national network and
the entire layout of the subway should be designed and laid out with reference
to the same datum.
Before the vertical layout work can be started a benchmark is to be established
in the vicinity of the shaft on which the measurement of the depth of the shaft
will be based. This benchmark should be checked periodically, however, because
it is located in the settlement zone.
Let F and A denote the upper and lower benchmark, respectively, and let MF
denote the elevation of F above sea level. If the readings on the upper and lower
rods are IF and /A> respectively, then the elevation of A above sea level is

(5.28)
-where I = the difference in elevation as measured on the standardiz ed tape.
 LAYOUT OF SEGMENTS AND SHIELDS 561

The levelling in headings should be run both ways. Because the benchmarks
are located on the walls, suspended levelling boards, like the one shown in Fig. 5/17
(illuminated from the back) Cl!n be used. Nails can also be used as temporary
markers. All elevations have to be checked over again before constructing the
permanent lining in which the permanent benchmarks are to be installed. Plugs
should be located in such a manner as to allow the levelling rods to be set up
vertically over them . The instruments used in the levelling should be able to be
read to the nearest mm (0·01 ft).
For further details on vertical layouts reference is made to Section 51.2.
The following procedure can be followed to carry out vertical layout in areas
under pneumatic pressure. The level is set up in the main lock in such a position
as to be able to sight both ways i.e. the levelling rods set up on the benchmarks
or turning points should be visible in both directions. A backsight reading is
taken first and then the gate is closed for the locking operation. The other gate is
opened to take the foresight. It is advisable to repeat this procedure in reverse
order to eliminate gross errors and to improve the accuracy of levelling. Vertical
closing errors are usually smaller than horizontal ones (see Table 5/T.).

5".23. LAYOUT OF RING SECTIONS (SEGMENTS) AND SHIELDS

52.31. Layout of Segments

In a geodetic coordinate system the centre line of a subway is defined by a set
of equations of tangents and curves. In addition to this the coordinates of stations
at 15-60 m (50-200 ft) intervals are calculated and indicated on the plans.
During construction the alignment
and the positions of the segments
should be checked regularly. This
means primarily the re-establish-
ment of the centre line of the tunnel,
based on underground benchmarks
and reference points.
Consider points 1 and 2 along
FIG. 5/43. Checking the centre line of the tunnel section
the position of lining segments under construction and horizontal
reference points A and B located
nearby. To make it simple the centre line of the tunnel between I and 2 is
assumed to be on a straight line (Fig. 5/43).
There are two possibilities in a case like the one illustrated in Fig. 5/43: either
point I can be set out directly from stations A and B, or - and this is generally
the case - an intermediate point H can be laid out on line 12 from, say, station B.
The general case will be considered first; the calculations and measurements
can be carried out in the following order: (A theodolite with readings to the nearest
1" is required for the measurements).

36 Szechy: The Art of Tunncllins

562 TUNNEL SURVEYING

1. The coordinates of the arbitrary point Hon line 12 are calculated. Point H
should not be farther from B (from which it will be set out) than the length of the
tape 20 m (50 ft) and should be about half way between 1 and D. Also, point A
should be visible from H.
2. Bearings <5 12, {JAB and <58 H are calculated from the coordinates and so is
distance t = BH.
3. With the theodolit~ set up over B, a backsight is taken to A and the reading
on the horizontal circle is recorded ([A). The theodolite is now turned to read
([A + <Jan - <J8 A) on the horizontal circle.
4. A target on a centering control base is lined up with the telescope at a distance
t from point B.
5. This procedure shall be repeated in positions I and II of the telescope. The
centering control base should be set half way between the two points obtained
in the two operations.
6. Once point H has been established the theodolite and the target are to be
interchanged. A backsight is taken to point B and the reading on the horizontal
circle is recorded (!B).
7. With TB known the theodolite is turned to read (!A) = l8 + <JnA - <Jna·
If the difference between this value and the correct one does not exceed 4 to 5
seconds, the calculations and measurements establishing point H were correct;
in the opposite case both must be repeated.
8. Having checked point Hthe theodolite is turned to read !1 = TB + <Jn 1 - <JnB
and a well illuminated, easy to sight target is set along the· line at an arbitrary
point F. The theodolite is now turned through 180° to !ead !2 = /1 + 180° and
a diopter is directed on line at point D. For best results the distance FD should
be about 50 m (150 feet). The layout of points F and D has, thus, been completed
with the telescope in position I . With the telescope in position II a backsight is
taken to point B and the reading on the horizontal circle is recorded (/'8 ). The
telescope is now turned to read I{ = /~ + {JHl - <5n»· Since this will not result
in the same point as before, the target is to be set half-way between the two.
Keeping the telescope in position II, point D can be laid out in the same manner.
Thus the section of line 12 within the construction area has bee.o established on
the ground.
If point 1 can be sighted directly from A as well as from B, point 1 shall be laid
out to begin with. If, for example, station B is the point from which to start,
distance Bl is calculated first and the bearings of lines AB and Bl are derived
from the coordinates. Point A should be laid out with the telescope in both posi-
tions and should be set half-way between the two resulting points.
Now the theodolite is set up over the point just obtained and the target is set
up on B. As before, the telescope is lined up again with 1-2 and diopter D directed
on line. With this line 12 has been established on the ground.
 LAYOUT OF SEGMENTS AND SHIELDS 563

To lay out the last lining segment the procedure to be followed is this:
I. With the theodolite set up over point H (Fig. 5/43) and sighted to point D,
an auxiliary point S - on a centering control base - .is directed on line near the
centre of the last segment. The distance between H and S is measured next and
the theodolite and the target are interchanged. The telescope is sighted to Hand
the reading on the horizontal circle is recorded (/H)-
2. The theodolite is turned to read /H + 90°, i.e. at right angles to line 1- 2.
3. Two points are directed in line with the theodolite on the rib of the segment
and at the opposite ends of the horizontal diameter and marked with a punch.
These points represent the ends of the horizontal line in the perpendicular vertical
plane.
4. The distances from the front face of the segment to the punched marks,
called the right hand and left hand advance of the segment (si and sb, respectively)
are to be measured next. The difference between the two distances, si - sb, indi-
cates the gain of one side over the other and should be marked in red on the
proper side of the segment. If, for example, s1 = 435 mm (17") and sb = 446 mm
(I 7 1/2") then the difference of 11 mm (l/2") should be marked on the left hand
side.
5. The centre of the segment and its horizontal offset from the centre line
should be determined to the nearest 1/2". One end of the tape (with the zero mark)
is to be held against the left end of the horizontal diameter and, under sufficient
pull, the· distance has to be measured to the right end. An intermediate reading
should also be taken at the plumb line centered over- the instrument. If, for example,
the distance between the two .marks. was 5·60 m (18'-4") and the plumb line was
found to be 2·75 m (9'-0") from the mark on the left, then the eccentricity, i.e.
the offset from the centre line of the tunnel is obviously 5 cm (2") to the right.
6. The coordinates of the centre, established as above, can be calculated. The
centre line of the tunnel and the vertical plane passing through the two points
intersect at a distance of 18·45 .m (60'-5" ) from the auxiliary point S. Point K
will be 5 cm (2 in) to the left of the point of intersection, measure~ in a perpendic-
ular and horizontal direction.
7. The direction in which the segments are headed may be determined from the
amount of gains and from the distance between the punched marks (d). The plane
of the segments is inclined to the pllne perpendicular to line 12 at an angle of
Lib which can be calculated from

· Sb- Si .
Lib = arctg d , (5.29)

or, in our example

. 11
Lib = arctg .
5600
564 TUNNEL SURVEYING

8. The angle of skew should be determined in the plane passing through the
vertical diameter and the centre line and is to be measured with a plumb line
suspended from the edge of the rib at the top of the segment. Once the distance
between the vertical projection of the above edge and the lower end of the vertical
diameter has been measured (b), the tilt (slope) of the segment can be calculated
from
(5.30)

If the plane of the segments is leaning backwards (i.e. the gradient is rising)
then distance b should be measured at the top, with the plumb line held over the
bottom of the vertical diameter.
9. The elevation of the centre of the segments above sea-level should be deter-
mined by levelling. The elevations of three points along the bottom of the ring
are determined by levelling and then, because the radius of the ring is known,
the elevation of the centre of the ring can be calculated.

52.32. Determining the Position of the Shield

M9re often than not, the centre line of the shield and that of the tunnel do not
coincide during construction. Accordingly, the centre of the shield is offset hori-
zontally as well as vertically, from the centre of the tunnel. The shield, of course,
has to be steered during construction both horizontally and vertically and any
rotation abo,ut its own axis should be checked regularly from time to time.
The position of centres C1 and C3 of ribs 1 and 3 of the shield have to be deter-
mined in relation to stations 1 and 3 on the centre line of the tunnel.
To determine the gradient (slope) of the shield it is sufficient to level two com-
parable points located in the upper part of the shield on the two end ribs (stiffen-
ers). The slope (%) will be equal to the difference in elevation, multiplied by
100 and divided by the distance between the two points.
The vertical position of the shield may also be expressed in terms of elevations
if the centres are tied in to the nearest benchmarks by levelling.
The following simple arrangement is a practical way of checking the changes
in slope and the rotation about the axis, while the shield is being driven:
A plumb line is suspended from an arbitrary point at the top of the shield and
a checkered plate is placed under it. The grid lines should run parallel and per-
pendicular to the centre line of the shield. With t_he grid lines spaced at 1/1000 of
the length of the plumb line the position of the plumb bob will indicate the slope
to the nearest 0· 1 % when read in the longitudinal direction and will give the
rotation in the same units when read transversely. The plate should be set properly
and made horizontal by measuring with a level in two perpendicular directions.
The position of the centre of the shield has to be determined horizontally as well
as vertically.
Both the eccentricity of the centre of the ring and the rotation of the segments
can be measured with circular scales as well. Graduated in mm and having a
 LAYOUT OF SEGMENTS AND SHIELDS 565

radius of about 2 m (6 to 7 ft), the scales are mounted concentrically with the ribs
near the top of the shield, with the zero mark set vertically above the centre
(Fig. 5/44). The rotation of the shield can be
measured with a plumb line, passing through Zero mark
the centre of the shield and indicating a read- Ctrcvlar scale ·
ing (c) on the scale (Fig. 5/45).
The eccentricity of the centre of the
shield is
h = d - c, (5.31) e
where d = the reading taken at the point
where the vertical plane (plane of the light)
passing through the diopter set up on the
centre line of the tunnel, intersects the scale.
It should be noted that Eq. 5.31 is only Fm. 5/44. Measuring the
an approximate formula. However, since the rotation of segments on
eccentricity of point C is only a matter of a circular scale
inches, ·the horizontal distance (h) may well
be replaced by the length of the arc (c - d).
Distance h can be determined as follows:
1. The centre of the shield is projected
up onto the scale with a plumb line and
the reading (c) is recorded (positive to the
left, negative to the right of the zero mark).
2. Similarly, the light passing through the
diopter is projected on the scale and the
reading (d) is recorded. FIG. 5/45. Measuring
the offset of the centre
Repeating these operations at ribs I and 3,
the values of c 1, c3 , di, d3 , h 1 and h3 can be
determined and the angle of shield deflection can be calculated from (Fig. 5/46):
h3 - h1
tan w = ---=---=- (5.32)
a
where a = the spacing of the ribs.
RibJ

r cfa ID ..____ • /"

Centreline
~---· I - - -- - o( svb~!f_
F10. 5/46. Measuring the
deflection angle
of the shield
566 TUNNEL SURVEYING

It is noted that reading c (in Eq. 5.31) can also be measured on a stiffenc::r plate
as reference. From reading h, as indicated by the plumb line on the plate, and
from the length of the plumb line (f) the value of c can be calculated (Fig. 5/47):

Oi'tvlarscale r
c :h = r :f thus c= h- (5.33)
f

r = the radius of the circular scale.

The relative position of the shield and
that of the last segment may be determin-
ed by separate measurements as above.
However, there is also a direct method.
Two points are selected and marked on
the horizontal diameter and two other
points on the vertical diameter of the
shield, and the horizontal distances are
FIG. 5/47. Measuring measured from the segment (91, 9 2 and
the rotation of the shield g3, 94).
with a circular scale If the centre line of the shield and that
of the tunnel are parallel, then

91 = 92 = 93 = 94· (5.34)
If they are not parallel, then the vertical deflection angle is

</> = tan-1 91 - 92 (5.35)

V

and the horizontal deflection angle is

-1 93-94
K= t an h , (5.36)

where 9i, 9 2 and 93 , 94 refer to the points on the vertical and horizontal diameter,
respectively
v and h = the distance between the points on the vertical and on the .
horizontal diameter, respectively:

REFERENCES
(BARANOV) EapaHoB. (1952): Geodesy in Tunnel Construction. Moo:«a 1
HEWITT, B. H. M. and JOHANNESSON, s.
(1922): Shield and Compressed Air Tunnelling. McGraw
Hill, New York
KONRAD, 0. (1952): Folda/atti meresek tdjekozdsa a bdnydszatban es az a/agutepitesnel (Orien-
tation of underground surveys in mining and tunnel construction). Kozlekedesi Kiad6,
Budapest
 UJBLIOGRAPHY 567

KoWALCZYK, Z. (1952): Hole-through Problems Encountered in the Construction of the

Warsaw Subway. Warsaw
LANE, K. S. (1957): Effect of Lining Stiffness on Tunnel Lining, Proc. IV. Int. Conf Soil.
Mech. London Il 223
OLTAY, K. (1941): A budapesti varosmeres haromszogelesenek hosszmeghatarozasaban elert
pontossag (Accuracy of linear measurements in the triangulation of the Budapest city
survey), Geodeziai Kozlony 2 ·
OLTAY, K. (1948): A Szabadsag-hid epitese alkalmab61 vegzett geodeziai munkalatok (Geo-
detic survey carried out in connection with the construction of Szabadsag Bridge), Geodeziai
Kozlony 8-10
OLTAY, K. (1951): A budapesti invardrotmeres (Nleasurements with invar wire in Budapest).
Akademiai Kiad6, Budapest
S.4.Rosr, SoHA and KELEMEN (1960): Bentonit a melyepitesben (Bentonite in civil engineering
construction). Miiszaki Kiad6, Budapest
TARCZY-HORNOCH, A. (1938): Sokszogelesi alagutkitiizesek pontossagi vizsgalata (An in-
vestigation of the precision of tunnel layouts based on traverses), Magy . Mern. es Ep.
Egy/et Kozlonye 1938 I
TARCZY-HORNOCH, A. (1943): Zur Fehlertheorie der offenen PolygonzUge, A Bdnya- es
Kohomernoki Oszt. Kozlemenyei
T ARCZY-HORNOCH, A. (I 944): A kenyszerkozpontositasok geodeziai jelentosege es szerkezeti
megoldasai (The -geodetic significance of centering control and its types of construction),
Mern. Tov. KepziJ M. 53
TARCZY-HORNOCH, A. (1949a): A legkedvezobb sulyelosztas es jelentosege a geodeziai
mereseknel (Optimum assignment of weights and its significance in geodetic_ surveys),
Mern. Tov, KepziJ M. 75
TARC2Y-HORNOCH, A. (1949b): Haromszogelesi alagutkitiizesek attoresi hibaja (Closing
errors of tunnel layouts based on triangulation), Magyar Technika
TARCZY-HORNOCH, A. (1952): Alagutkituzes pontossdgi kovetelmenyei kiilonos tekintettel a
kiilszfni meresekre esaz irdnyatujtelre (Precision requirements in tunnel surveyingwith special
regard to measurements on -the surface and directional transfer). Kozlekedesi Kiad6,
Budapest
VINCZE, V. (1953): A budapesti foldalatti vasut felszini szabatos geodeziai munkai (Precision
surface surveys for the Budapest subway), Foldmerest~ni Kozlemenyek 2
VINCZE, V. (1955): A budapesti foldalatti vasut felszini szabatos geodeziai munkai (Precision
surface surveys for the Budapest Subway), Mern. Tov. KepziJ
WARD, N. H. and CHAPLIN, T. K. (1957): Existing Stresses in Several Old London Under-
ground Tunnels, Proc. IV. Int. Conf Soil. Mech. London Il 256
. -- - - - - -~

------ -- - ------- - ===--=----------------

CHAPTER 6

CONSTRUCTION A.ND DESIGN OF TUNNELS

Owing to the number of factors influencing the design, loading, location and
building of tunnels, various tunnelling systems have been developed. The major
influencing factors being as follows: geological and hydrological conditions,
shape and cross-sectional dimensions of the tunnel and its intended purpo~e.
The construction of tunnels requires the carrying out of the following operations:
1. excavation
2. support
3. transportation
4. lining or coating, sealing, draining and ventilation.
The combination and effect of these tunnelling operations vary according to
the above mentioned conditions. While the excavation and the transport of the
excavated material is always an indispensable necessity; the type of working tools
and the means of transportation used may differ widely, and the importance and
extent of both the support of the excavated cavities and the processes mentioned
under 4 can vary likewise, within.a wide range. The required operations can thµs
be carried out by various methods which can be grouped into the following five
categories:

1. full-face tunnelling without temporary support

2. mining or classical methods
3. combined underground and open surface (cut and cover) methods
4. precast element and caisson sinking methods
5. shield driving methods.

These methods will be discussed separately according to the type of rock or

ground to be penetrated, i.e. whether applied in solid rocks or in soft ground)
The first method can only be applied in solid rocks where also .the methods belong-
ing to group 2, which afford supports of variable extent and strength can be applied.
In loose and friable rocks .and in cohesive or granular soils all the·methods given
above can ·be used, with the sole exception of that under 1. In exceptionally soft
and loose ground the methods given under 4 and 5 will affoi:d good .results.
In the following, tunnelling methods will be discussed in order of their simplic-
ity, beginning with those applicable in solid rocks only.
 MECHANICAL DRJLLING 569

6.1. TUNNELLING IN SOLID ROCKS

6.11. MEANS OF EXCAYATION IN ,OLID ROCKS

The means of cutting solid rocks and the excavation methods used in them can
be discussed here only briefly, as their detailed description does not constitute
a particular subject of tunnel engineering. Accordingly, only the relations asso-
ciated with rock strength will be treated here in the following order:
1. mechanical drilling and cutting
2. technique of blasting
3. methods for the joint application of drilling and blasting.

61.11. Mechanical Drilling and Cutting

By mechanical drilling and cutting of rocks a local separation of the rock-

forming minerals is aimed at, through local crushing processes, brought about
by a series of fractures, i.e. by brittle fracturization due to tensile stresses. From
this point of view the cutting of rocks differs fundamentally from the machining
of metals, the latter being uniformly based upon plastic fracture or shear.
Fracturization is brought about by percussion or by percussion combined with
cutting, the acting forces being of static or dynamic nature, the former action
being brought abou·t mainly by rotary tools, the latter by percussive ones.
In order to bring about brittle fractures, stresses equalling the crushing strength
of rock (see Table 6/f) have to be created.
The crushing stress produced on the surface of the cutting tool, according to
HERTZ, is:

(6.1)

where p = unit compressive force exerted on the cutting surface

r = radius of the cutting tool bevelling
E 1 = modulus of elasticity of the tool material
E 2 = modulus of elasticity of the rock_.

The specific deformation caused by the dynamic impact - provided that it

may be considered fully elastic - can be expressed by the formula:

,d / V
e=-=K-,
I Vo
...,.... - -
-
...___Ii_ • • - -

570 CONSTRUCTION AND DESIGN OF TUNNELS

TABLE 6/1 (aftei; TALOBRE)

Dry Shear- Resist_a ncc factor .Angle of

Crusrung Tensile

=rc:r
ing internal
Type of ground density strength strength stre ngth
(t/ m') (kg/cm') (kg /cm') 5 frictior.
(kg/cm') (! (! = c; <•)
Hard limestone 2·65 700-1700 50 110 I ·55-2·4 2-5 45- 50

Medium limestone 2·5 300-550 30 30 2·7-3·6 41

Soft limestone 2 100 10 10 6·3

Gneiss 2·1

Dense granite 2·76 1800-2500 80 100 1-3- 1·5 2'1,

Fissured granite 2·8 750-860 80 100 2-2- 2-3 2·5

Quarzite-sandstone 680-1000 2·0-2·5 50-70

--
Marl 35- 180 4·8- 10 34 (sandy)

Porphyry 1360-2400 1-3-1"7

Quartzite 2·65 870-3600 30-50 100 1·1- 2'1 1·9

Slate (shale) 250-800 2-3-4 45

Soft slate (shale) 200 4-5 5 30

Medium sandstone 2-3 · 20 30

Basalt 3 80 200

where v = the velocity of impact

v0 = the velocity of compressed air in the cqtting tool
K a constant, whose value lies between I and 2, depending upon the
magnitude of the impacting masses.

As <Tdy n =E2e, its approximate value can be obtained from the formula :
 BLASTING 571

It is desirable under all circumstances that the percussive force should lead to
a permanent deformation in the rock, i.e. the latter has to absorb the greater
part of this energy. A strong rebound is always evidence of low efficiency.
Considering the existing uncertainties ·as regards the impact, it is worth men-
tioning that the contact stress is directly proportional to bath the velocity of
impact and the square root of the impacting mass and the rock's modulus of
elasticity. The efficacy of cutting depends upon many factors, such as the hardness,
cuttability, fatigue strength, tensile strength and breaking strength of the rock,
its dry or wet condition, the hardnt:ss, strength characteristics and toughness of
the tool material, the shape of the tool and the pressure exerted by it, the angle
of cutting, the frequency and the velocity of the percussions, etc.
The tools can be classified as percussive tools (light, medium and heavy pick-
hammers with a weigh{of G = 18 'kg; 18-27 kg, and 27-60 kg, resp.) and rotary
tools (crown bits, star bits, etc.). These fools are operated by compressed air at
an overpressure of 6·5 atm, the number of percussions or rotations ranging from
200 to 250 and from 100 to 650 per min. The rotations per minute needed for loose-
ning dry rock are higher in number than those needed for wet tocks. Of the cutting
tools - especially those designed for the cutting of blocky rocks, such as are often
encountered in mines, - the cutting and !Qading machines play a prominent part
and can occasionally be used also in tunnelling operatio.ns. Of the tunnelling
machines of Hungarian .make the type F4 (designed by AJTAY and SzIL.Gm) fitted
with a cutting head is worth mentioning as it can also be successfully employed
in tunnelling op_erations in··<iry clays. As a successful device of this kind is iiitro-
duced.the B.RETBY Tunnelling Machine. 6· 1 a

61 .12. The Use of Explosives and Blasting Technique

Explosives, confi"1,ed within a relatively small space (the shot hole) and initiated
there induce chemical reactions in the course of which gases and heat will develop
in such 4\.!antities that the stresses produced in the surrounding rock by the devel-
opir;tg gas pressure will exceed the strength of the ro.ck in a certain zone. Blasting
is, t}lerefore, effective only within certain zones of a given radius, discriminated
accordlng to the succe~sive decrease in blasting efficiency as follows (see Fig. 6/la):
(a) crushing zone
(b) throwing zone
(c) fracturing zone
(d) shattering zone.
In addition to the type and quantity of the explosive used and the range of
blasting, the radius of this zone depends upon the type of the blasted rock; the
81
• • Proc. Inst. Civ. Eng. 1965 April
For the detailed description of these tools and machines and of their construction cf.
Blinyliszati Kezikonyv (Handbook of mining), Vol.. II, Ch. 6, pp 717-822 ·
 572 CONSTRU CTION AND DESIGN OF TUNNELS

\
,
I

I ~
~ \
\
,~\s.
1 !11
,
I

I
\ I
I
\ / F.1ce
'' /

Fm. 6/1. Formati on of a blast cone

distance of the extreme spherical shell, within which the blasting produc
es fissu-
ration or crushing of rocks is determined by the stresses brough t about
by the
shock waves exceeding their tensile strength. This stress, of course,
can only
produce displacement when a free face (e.g. a free rock surface or a joint
or some
rock material fractured by a previously fired charge) is.lying in the
direction of
shock-wave propagation. Under normal conditions, the blasting in
a shothole
will break out a crater in the rock, with sides sloping outwards
at an angle of
about 45°. The choice of the depth of the shotholes, the depth of cut,
and their
spacing, i.e. the drilling pattern can be determined according to the
following
theoretical considerations:
An explosive charge placed at the bottom of a shothole of depth m will
theoreti-
cally break out the rock material within a cone with a conical angle
of IX. The
volume of this cone (see Fig. 6/la) is

m
V= 3n,2 ,
or, as r = m tan IX
7t
V = m 3 tan2 IX •
3
According to LEBRUN, the following relation holds between the weight
of the
charge (L) and the depth of cut (m):
 BLASTING 573

ma
L=C--
cos3 ex
whence
L
m3 = C cos3 ex ,

where C = a coefficient, whose value depends on the strength of rock and the
quality of explosive used. Thus

1t L
V= C cos3 a: tan2 ex ,
3
and its limit value according to oc

dV
-d = -1t -L . • ·a )
(2 sm rx. cos2 a: - sm oc ,
ex 3 C
equalling zero if 2 cos2 oc = sin2 ex, i.e. with tan ex = Ji and ex = 54°44'. Thus
the cut-depth corresponding to the optimum blasting cone will be
3

mo=; J~ . (6.2)

From this value only 70--80 %is to be taken, considering that the blast is also acting
downwards; according to experience, and, thus, the crater-cone will start from
below the bottom of the shothole.
Te.st results have shown, however, that only about 20 % of the energy of the
explosive charge will be actually utilized for breaking the rock.
Explosives can be classified into the following main groups:
l. Slow (or phlegmatic) explosives, which, when confined, can be exploded by
simple exposure to an open flame, and which have a relatively low velocity of
burning, resulting at first in the bursting_of the surrounding rocks then their
displacement by the slowly developed gas pressure, not crushing them, however,
to fine fragments. Such an explosive is blackpowder.
2. High (brisant or quick) explosives which can only be initiated by the acti-
vation of .a detonator, producing a total decomposition of their whole mass at
a very high vel9city (about 6000 m/sec) shattering at the same time the material
in their imm!icliate _surroundings, breaking, rending off and displacing those lying
farther away. Such explosives are, e.g. nitrotoluene, nitroglycerin, nitrocellulose,
dynamite, ecrasite, paxite, etc.
3. Initiating , explosives· (e.g. fulminate of mercury) used for detonators to
initiate the activation of certain less sensitive brisant explosives must also be
mentioned here.
 574 CONSTRU CTION AND DESIGN OF TUNNELS

When blasting in tunnels the task generally aimed at is to use the minimu m
possible weight of explosive for the loosenin g of the maximu m possible
volume
of rock. A suitable fragmen tation of the blasted rock is at the same time
also
desirable to facilitate mucking and disposal of the debris.
The spacing of the shotbole s and the sequence of their firing play a promine
nt
part in the proper accomp lishmen t of this ,task. If the distance between two
shot-
holes is not too great, the blast will break out the rock materia l between
them
(see Fig. 6/ lb). An inclined shothole is generally more efficient than one
drilled
perpend icular to the rock surface, as the thrust of the blast in a sbothole
acts
normall y to its centre line and, thus, there is far less danger of blowing
out the
stemmin g in an inclined hole. A shothole inclinat ion of 45° is generall
y used
(Fig. 6/lc). If the explosive charge is placed equidist ant from two free faces
per-
pendicu lar to each other, the effect of the blasting will be similar to that
shown
in Fig. 6/ ld, breakin g out 2·5 times more material than in the pattern in Fig.
6/ la.
With three free faces the result will be even 3·5 times as much.
Sbothol es are generally drilled with a diamete r of 26--40 mm (1-1 1/2 in)
and
with a depth of 1·2-3·5 m (4-12 ft). It is not worth while to drill boles
deeper
than 2·5 m (8 ft) as the drilJing time of a 3·6 m (12 ft) deep hole is equal
to that
of two holes of a depth of 2·4 m (8 ft). Moreov er the drilling of a deeper
bole
must be started with a greater initial bit diamete r. The hole depth used conside
rably
depends on the type of rock to be blasted. In a rock material of low strength
and
readily loosened, shallow er drillholes and smaller explosive charges are generall
y
used. Anothe r aspect to be conside red is that the depth of boles should
never
exceed the width of the tunnel, and that the round pulled by the boles is generall
y
shorter by about 30--40 cm than the hole depth (averaging one tenth). To
increase
the blasting efficiency of the shot it is advisable to use a bottom charge made
of
a more brisant explosive. The hole depth (m) to be used may be,
81
accordin g to
RANDZI O : · b

m = u·- 1) a. + 0·5, (6.3)

where f = the cross-section area of the tunnel face
a. = a coefficient the value of which depends upon the diamete r (15)
of the shothole varying as follows:

15 = 60 - 80 mm, a.= 0·20,

15 = 45 - 60 mm, a.= 0·18,
15 = 20 - 30 mm, a.=0·10 .
Shothol es are, in general, drilled accordin g to a predeter mined drilling pattern
and fired in a predeter mine_d sequence of roundi. The drilling pattern to be·used
de-

• -ib RANDZJO , E.: Der Stollenbau. Ernst, Berlin 1927. See

further: Furtis-robbanttis a btinyti-
ban (Drilling and blasting in mines). Miiszaki Kiad6, Budapest 1961
 f -::~>Support1d!f brt33t /)(gm--,- \
'ltfr;.:;:- ~~-~ ====
Borin_1/ c1x1s of drtl/

i:-::...- .....
-...;~¾

Supportin!f cross be.1m f{J-f·20m ·

I• •I
I I
-.--- - - ; ·
\' 'i\ ~\ j!j :: :; ij
\l Ill •I' .JIL ..l,,i_
1:jT" \\ ~ :.I 1T ff T~
I I I I

5 ,t.2 !,.:>I "•• 7~"

ti f
. n\
,, ,, ,1
-;f'.ZJC-;'*:
ft2. !;;

~
Ll
,~
Ll I

-;,- : ~
I I I
...KJ
I
...~ a,

i
.J l:l 4
w -~ ,1
+1.r ~ 15 ..
14
J-
,a .._,5
i2 ~3
I ~ '!t,
,,-;J9 ¥ I +. /8 21) +- ,, I Y • i -
ff±

I •
d I,
4 4 ,r
2 2 5
'"<> 'I. 9 er'
J \ / 3
jo-- fo--''--o! - - -oJ
/\
I \

2cl \2 I
\
%- f
I.I\
FIG. 6/2. Various drilling patterns -.I
I.I\
-- - - - - - - - - - - --

576 CONSTRUCTION ANO DESIGN OF TUNNELS

"' pends upon the depth of the holer,,

~~
]~ the stratification of the rock and
<--s ~ Thermodynamic its behaviour. (Shotholes parallel to
pressure zone the stratification are disadvanta-
1j1111t1ating geous!) The practice, in general, is to
I (detonJtors) use V-shaped wedge -'cuts' in the
middle of the face with 'ring' holes
around them, inclining outwards
towards the periphery of the tunnel.
Ill. llig/J explosi/les The holes can be spaced in lines,
rings, or recently, in a spiral se-
quence (Fig. 6/2). The sequence of the
firing of rounds varies widely ac-
Time
cording to the nature of the rock,
to the applied operation methods
and to the position of headings. It
---!:--
.,,..,..-- 4 ...................... is practical to start firing with the
r/
f
oJ
of of
oJ 1
\
cut (or break-in) holes placed in

/
f o2 I oO o O
o2 \
\
the centre of the tunnel face, and
to finish with the outermost rim
02 02 holes, as in this case all the subse-
f I of of I
quent shots will have at least two
L__o~ -~4__- ~J_j free faces. Special attention must be
given both to the bottom holes and
On//m_q 0iLtt:m to the outermost peripheral rim
F10. 6/3. Principle and pattern of holes which are to be drilled with
short delay blasting an inclination downwards, or out-
wards, respectively. The efficiency
of blasting is naturally here the smallest - owing to the high resistance of
the undisturbed 'infinite' rock-mass - and the supplementary removal of the re-
maining thin cover constitutes the most cumbersome and expensive work.
The overall effect of blasting and especially the rock fragmentation is also bene-
ficially affected by the application of delay firing, controlled by the length of the'
delay periods between the consecutive rounds of shots and by the sequence of firing.
The effects of the blasts will be beneficially superimposed if the explosive charges
in the holes adjacent to a previously fired shothole detonate at a time when the
physical effects of the preceding shot have not yet come to an end. This advantage
has recently been exploited by the use of short-delay exploders or millisecond
blasting6· 2 with which the detonations of the consecutive neighbouring charges
follow each other after a delay period of a few hundredths of seconds.

6 ·2 BIERMANN, G.: Neuzeitliche Sprengtechnik . Bauverlag, Wiesbaden 1966

 :BLASTING 577

figure 6/3.sbows the pressure-time curves for explosives of various types. Initiating,
phlegmatic and brisant explosives are here represented by the curves I, II and III,
tre~ectively. •Of these, -the brisant explosives are most often used in tunnelling
o_perations. According to :eurve Ill a shock pressure of short duration but very
hjgb "intensjty js followed by a thermodynamic prtssure of longer duration and
!.W}th_;g radu.aUy -decreasing intensity. With millisecond blasting or delay exploders
'lthe ~thermodynamic pressure of a preceding shot is still in existence when the
sbock pi:essure of the ,detonating explosive charge.in the adjacent hole is already
ib~.,ginning fo ..act, with the result that tlie two effects in the immediate surround-
6.ngs of the hole become sumn;iated advantageously while, at the same time, the
tock 'Vibrations farther away are damping each other. The optimum millisecond
ilelay period is proportional to the half of the oscillation period of the shock wave,
cbaracteristic of the medium, i.e. of the time in which the seismic vector changes
'from '.its ·positive to its negative maximum. The main advantage of millisecond
blasting as against the simultaneous firing being t_hus that the effect of blasting
an :the dose vicinity of the working face is increased, being decreased in areas
farther away.
'The suggested millisecond drilling pattern~·3 is a linear pattern shown in Fig.
6/3b. The two shotholes denoted by Owill be :µred first to be followed in sequence
by ;the ,d_~tonation of the four holes denoted by 2, then the four denoted by 3
Gl;ll_d, <finally, the last two - bottom and top holes - · deri.oted by 4; the distance of
the boles from those preceding in sequence being 0·5 .m and 0·6 m respectively.
7Ihis drilling pattern is generally used for the driving of pilot· headings and is
ca'lled the parallel cut meth~d as its holes must be drilled parallel to one another.
Theoretically, the criss-cross firing pattern, where the delay firing of the hples is
proceeded with alorig a saw-tooth line, is preferable and has recently been sug-
gested as particularly suitable for linear excavations.
Millisecond blasting is not only more efficient than simultaneous firing but even
Tequires less time than firing in separate rounds.
IB!asting in tunnelling ~ used for the foliowing operations:·
{a) cutting or breaking-in
(b) full-face blasting or reaming
(c) tunnel ·Pr9file trimming or contour blasting.
(a) Cutting is made in order to create .cut-holes from which the breaking out
of the ·remaining tunnel cross-section ca_n be started an:d facilitated. The successful
break-out of a cut requires the use of the most .efficient explosives and the largest
<:barges as well as the drilling of the deepest and mo.s t closely spaced drill-holes.
The cut can be blasted in 3 or, possibly, 4 rounds, dependi:rig upon its desired
.dimensions. It must, however, be always: in a central position. With greater cut
dimensions -additional 'blind' (uncharged) boles are to be drilled into the face

.,~ er: HETZEL; K.: Tunnel- und Stollenbau - Wandluogen und Erfolge, Bauingenieur, 1957 9

97 Szcchy: The Art of Tunnelling

-..;;... . . - - - --

578 CONSTRUCTION AND DESIGN OF TUNNELS

with diameters increasing with the decrease of the strength of the rock. The spacing
of the drill-holes in hard rocks such as granite should preferably not exceed 20 cm.
For the blasting out of such cuts one of the most suitable drilling patterns is the
so-called Canadian or 'burn' cut shown in Fig. 6/4, in the centre of which there
is an uncharged 'blind' hole with a greater diameter, around which shot-holes
with smaller diameters, numbered in
Fig. 6/4 in their firing sequence, are
placed spirally at increasing distances.
(b) Drill holes for section reaming are
placed around the cut either in concentric
rings, or along a spiral, spaced at distanc-
es of 0·7 to 0·8 times the blasting radius
(see Fig. 6/2). The spacing of the holes
in the same layer should not exceed 1--2
to l ·4 times the blasting radius. The
-alint1'dr1!/ shot-holes in the same ring are fired in
/Joie
one round. The proper choice of the
length of the explosive rods in the reamer
holes and the control of the fragmen-
tation is of great importance for the
efficiency of the reaming. Shorter ex-
plosive rods are used for shorter delay
periods and longer ones for longer delay
Fro. 6/4. Canadian cut arrangement
periods. The degree of fragmentation
depends upon the specific quantity of
loading in the shot-holes, increasing in addition with the decrease in the dura-
tion of the delay periods.
(c) Contour blasting or trimming of the. tunnel profile can be carried out more
or less accurately.
The shot-holes have an inclination of 10° outwards (see Fig. 6/2) and
their length should not extend beyond the theoretical contour of the tunnel
cross-section by more than 10 cm (4 in). According to experience, the cleavage
effect of the blast should not extend beyond 20 cm (8 in). In order to obtain a
high standard of accuracy in tunnel profile trimming it is advisable to spaca: the
peripheral holes rather closely and to use shorter holes than for reaming. The
explosive charges should be discontinuous (cushion blasting) and of a very high
brisancy.
The formulae used for the calculation of the required quantity of explosives
can be arranged in the following three groups:
(a) The explosive charges of conical 'mines', - mostly used by the military
engineering corps - may be obtained from the following formula:

(6.4)
 BLASTING 579

where L = the quantity of explosives in kg

e 1 = coeffi::ient whose value depends on the kind of explosive
f 1 = the strength factor of rock
d = the smallest cleavage length, with a value equal to 0·5 to 0,75
times the depth of the shot-hole.

GRIMAUD's formula is similar to the above:

where v = a factor depending on the number of free faces, equalling 1 with

only .one free face and 0·25 with 5 free faces
m = the hole depth
b = a coefficient depending on the shape of the shot-hole and varying _
from 0·5 to 1·0, with the· change of ratio m/.J from 50 to 33 (rn/o
denotes the ratio of the hole depth to the diameter).

(b) For linear charges usually used in tunnelling operations the empirical rule
of FRAENKEL can be used, according to which·

(6.5)

where a the blasting radius of the charge (in m)

m = the depth of the shot-hole (in m)
h = the length of the charge (in m)
o = the dia~eter of the charge (in mm)
(! = the coefficient of resistance to be obtained from the formula

(! = J- 4 0
~ ; R = the rock's resistance to. failure (see Table 6/I) .
.
Taking a 35 % dynamite witli a rending capacity of 10 000 kg/cm2, FRAENKEL
gives the (! values as varying from 1·3 to 5·4 for bard rocks and for soft ground,
respectively.
(c) The third group of calcula!ion formulae refers to parallel series blastings,
when the tensile strength of the rock in the planes to be separated and the shearing
stren·g th of the ,;am~ in the shear planes, as well as the resistance to friction due
to the weight of the rock mass to be displaced, must all be overcome by the rending
force acting in the cross-section to be blasted. A detailed discussion of this would,
however, exceed the scope of this book.
The following simple formula published by STINI and used generally in Austria
must also be mentioned
L = 1h2 evfs, (6.6)
37•
 -- - --

580 CONSTRUCTION AND DESIGN OF TUNNELS

where m = the least distance between the shot-hole and the free face
e = a coefficient representing the bri~ancy of the explosive and being
of a value of 0·8 for blasting gelatin and 1· l for gelatin-dona rite
v the rending factor, with a value to be tak~n as 1·6 for the cut holes
of headings with but one free face, 1·4 for other shot-boles in bead-
ings and l ·2 for the blasting of the calotte, where there are two
free faces
/ the strength factor of the rock, its value varying from 0·6 (for
limestone) to l ·5 (for quartzite), while the values for other rocks
may be obtained by interpolation
s a stratification coefficient with values ranging from 0·6 to 2·0,
depending upon the stratification , fissuration and cleavability of
the rock and comprising also the effect of the angle between hole
axis and stratification.

Even more simple is the formula given by LUCAS:

(6.7)

where k = a factor of the specific explosive demand, the value of which can be
taken from the following table:

Paxite
.
Dynamite

for rocks easy to blast 0·25 0·08

0·40 0·10
for rocks moderately difficult to blast 0·40-0·60 o· 10-0·15
for rocks difficult to blast 0·65--0·80 0·15--0·25

The explosive demand for pilot headings is always higher.

Fur full-face tunnelling through hard and homogeneou s rocks the specific
explosive consumption is given by WALLING Table (6/11).

TABLE 6/ll

Cross-sectional area of tunnel (m') IO 20 40 I

I
60 80 100

Specific consumption in explosives (kg/m 3) 2-5 I-8 1· 1 I 0·8 0·7

- - - - - - - -- - -- - --
Specific borehole length required (m/m') 4 2-8 l ·8 1·4 I-2 I

Number of boreholes according to

- - - - ---- - - - - --
drilling scheme 32 42 60 75 84 100
 FULL-FACE TUNNELLING S8t

6.12. FULL-FACE TUNNELLING WITHOUT SUPPORTS

It is obvious from the above that this most simple method of tunnelling can
only be .used in slightly fissured rocks of very high strength. A distinction
. must
be made whether temporary ~upports may. be omitted in the course of tunnel
excavation only ot whether permanent support or lining o( fhe section may be
dispensed with entirely. That is, whether only the operation 2 (support) mentioned
under Section 6.13 is being omitted or whether the lining of the tunnel mentioned
under 4 is to be con'siderably simplified. The latter also includes those cases in
which a coating or rendering is only required for the reduction of traction resistance
and for aesthetic purposes (high.way tunnels), or· for waterproofing seepage-
prevention and drainage purposes (water supply, pressure, and navigation canal
tunnels). This latter case per~ains, J;,owever, to the methods of full-face tunnelling
with support, to be discussed in Section 6.13.
Tunnelling without support actually consists in the repetition of the following
cycles of operations:

(a) drilling shot-holes into the tunnel face

(b) loading and firing of shot-holes
(c) clearing and disposal of the debris (' mucking out' the muck pile)
(d) supplementary operations (contour_trimming, breaking down of -loosened
rock layers and lining or coating for operational purposes, if necessary).

It is possible to blast either the whole cross-sectional area or parts of it in each

cycle-. varying in accordance with the chosen sequence of operations. The working
method depends mainly upon the shape and dimensions of the tunnel and the
nature of the rock to be penetrated; it is also influenced by other factors such as
the available equipment, the cost of labour and the required construction period,
among others. The usual practice is, first to drill all the shot-holes required for
reaming out the whole cross-sectional area of the tunnel or the pilot heading,
in such a number and following such a drilling pattern'as to achieve an adequc,1te
advancement with the minimum consumption of explosives and with a rock frag-
mentation suitable for efficient mucking. In addition, both overbreaking and
underbreaking must be reduced to a misimum in order to avoid subsequent
filling or trimming. Quick and efficient loading and disposal of the debris and
adequate ventilation and drainage are also of outstanding importance.
The great improvement in rock tunnelling in the last two or three decades
comes from the improvement .of drills and drill steel, the increase of the hole
lengths, the introduction of the portable drill carriage (or jumbo), the use of delay
exploders in blasting, and the development of economical and reliable mechanical
muckers. A truck-mounted drilling platform used at one of the stations of the
Stockholm subway driven in granite is shown in Fig. 6/5a as an example, while
Fig. 6/5b shows rock-drills mounted on a simple column leg and operated from
a platform.
- ~ ~ -- - - --- .

582 CONSTR UCTION AND DESIGN OE T UNNEL S

FIG. 6/ 5a Jumbo mounted on a motor car

(Stokholm)

Fro. 6/ 5b Simple
driling-leg on a working platform

Some time ago tunnelling methods employing a heading were customarily used.
In these methods a heading generally with the minimum size for economical
driving, i.e. with a cross-sectional area of 3·6-4· 5 m2 (40--50 ft2) is excavated first.
This heading was subsequently enlarged to the required size by methods which
were relatively less expensive than those necessary for the heading excavation.
The length and location whether at the top, centre or bottom of the full size
cross-section depended on the conditions of the particular undertaking. The
practical location of the heading was at the top, as in this case the remainder of
the tunnel section, the bench, could be broken out stepwise downwards, benefiting
therefrom in the economy of drilling, firing and mucking operations alike tFig.
6/6a). For the advancement of the top, vertical shot-holes can be drilled more
efficiently in its floor, i.e. in the bench, and m~cking could be effected into mine
trucks running on the floor of the tunnel. An additional improvement in this
heading and bench method would be the introduction of horizontal cutting
(rock-milling) devices which could effect the remov~l of the bench in 'slices'.
In a section heading no drilling jumbos, but only simple jack posts have to
be used on which the drills are mounted (Fig. 6/66). For greater speed the bottom
hea ding has been used sometimes and the enlargement started at frequent intervals
along it, permitting the full-scale work to be completed almost as soon as the
heading. The disadvantage of this method is that the bottom heading has to be
completely timbered to protect traffic through the itttermediate working faces.
 (v

@!
@, 8) ..,
(J) ?
t;'
f 'IG. 6/ 6a Heading and bench stages ~
()
t,I

--i
c::
!i:
-------- ~ ) ; ) o . ' ) ,~,.i. >")!a
~ 'Z..""=-"'=.."':11:.
~-~~'"""-............ - ;:!'-- -.:-=
~ - - •. -::-e.::..,:.o,... - <l!,a=:::..~=----
~
- -:~:::::::,
. -- ...~-~---~
-~--~-~
DJ

FIG. 6/6b Heading and bench method

V,
..,
OC,
584 CONSTRUCTION AND DESIGN OF TUNNELS

In large section tunnels and in very hard rocks where no timbering is necessary
the centre heading method can b.:: used advantageously (Fig. 6/7). From this
centre heading radial holes at right angles to the axis of the tunnel are drilled to
the contours of the full-size section and then blasted, thus enlarging the remainder
of the tunnel section in one single round.

A-A

l.
--~
-wwa;
---1
.,,,,,.,.,.,,..,_,,,.,,,,"lr>,_.---,.,,,,....,..,--r~~

IA \'
··;.. Fm. 6/7. Central drift method

R'.ecently a multi-face tunnelling system with a separate parallel heading has

been developed in connection with the centre heading method which, by creating
intermediate tunnel faces, offers the advantages of having not only one or two
faces of attack but an unlimited number. By this system a heading of 2·40 m x 2·40
m (8 ft x 8 ft) cross-section is driven parallel to the axis of the tunnel at a distance
of 15- 25 m (50- 83 ft) aside. From this beading. cross-drifts are driven in every
500-600 m (about 1700-2000 ft) to the axis of the tunnel, from which intermediate
tunnel faces may be attacked - always choosing the most suitable tunnelling
method. This combination offers the advantage of simultaneous driving in both
directions with the use of the same cross-drift as, while blasting is being done in
one tunnel face, mucking can be carried out undisturbed and without interruption
from the opposite tunnel section.

6.13. FULL-FACE TUNNELLING WITH SUPPORTS

First of all let us look into the role and possible degree of the supports. As to
their role, they afford :
1. Protection against the fall of loosened rock fragments, while, at the same
time, no resistance is offered against the displacement of the surrounding
virgin rock disturbed by the excavation;
2. Support against all loosened rock masses without preventing further loosening
of the virgin rock;
3. Support partially stabilizing the excavated cavity by preventing rock dis-
placement beyond a certain extent;
 FULL·FACE TUNNELLING 585

4. A perfectly rigid and powerful support preventing any movement of the

surrounding rock;
5. An opportunity for the consolidation of the surrounding rocks with a view
to relieving the p~rmanent tunnel lining.

The full-face tunnelling method obviously represents the most economic tun-
nelling system. Therefore, it is desirable to extend its field of application by all
practical means to rocks which, though they cannot be left unsupported perma-
nently, may have the installation of their support either postponed for a limited
period, or reduced in extent by establishing an artificial co-operation with the
somewhat further removed undisturbed rock masses (roof-bolting).
The main point is always to make the excavation and to maintain the clearance
of the whole tunnel section without disturbing any inside support, until the load-
bearing permanent tunnel-lining can be constructed. The development of such
methods is rendered possible by the fact that rock pressures do not immediately
reach their maximum value when the excavation is completed but undergo a suc-
ces'live development. It must be understood, however, that their maximum value
is always significantly affected by the erection of the supports at the proper time
(cf. Section 3.51). It is a great advantage in various constructional operations when
the full tunnel section can be excavated even for a certain limited distance by the
full-face tunnelling method, and uniform temporary supports can be installed
subsequently, at intervals extending over the whole cross-section of this newly
driven tunnel stretch. It must be noted that the postponement of supporting a
cavity until the construction of the permanent tunnel-lining, would incur in most
cases a considerable increase of rock pressures with an inevitable consequent
increase of the lining thickness and thus render the whole process uneconomic.
There may, however, be some few exceptional cases when the rate of increase in
the rock pressure is so small, e.g. in ground of a highly plastic character, and the
construction of the load-bearing permanent supports - owing to their design
and construction - is progressing so rapidly (e.g. precast reinforced-concrete
segments or rigid steel arches) that the time elapsing from the excavation of the
whole tunnel section to the installation of the permanent support will not incur
any considerable increase of rock pressure. In such cases the omission of the
-temporary supports will not involve any surplus expenses.
Three methods will be discussed below, each of which offers considerable pros-
pects for further development.

(a) Temporary support by prefabricated steel construction with a subsequent

independent permanent tunnel lining;
(b) Temporary support by a reinforcement-like rigid steel structure followed by
its successive concreting;
(c) Anchoring of the immediate roof layers by rock bolts to upper lying layers
of higher load-bearing capacity (roof bolting) with a subsequent decorative
coating.
 -- - -
---- -
- ~

586 CONSTRUCTION AND DESIGN OF TUNNELS

61.31. Temporary Support Independent of the Permanent Lining

Such temporary supports can be composed of wood, steel or concrete elements

or of a combination of them.
The lining can be composed of arches, ribs or frames fabricated in conformity
with the contour line of the tunnel, with intermediate spacers, braces or struts
between them and with an outside lagging bearing upon them in order to transfer
loads from the rock, and to avoid the risk of falling rock.
61.311. Supporting with steel structures. In American practice the following
types of rigid steel support systems have been developed (Fig. 6/8):
(a) continuous rib type (leg and rib in one piece)
(b) rib and post type (arches on posts)
(c) rib and post wall type (arches on wall plates)
(d) rib wall plate and post type (arches on wall plates and posts)
(e) full-circle :ib type.

All these types can be used for full-face tunnelling methods, with the continuous
rib type offering most advantages. A precondition for the application of this
method, however, is the use of portable drill carriages or jumbos and a rock
quality to secure self-support not only for the time of the drilling, firing and
clearing operations but also for the time required for the erection of the temporary
or: primary supports (cf. 'bridging period' in Section 3.51). The most advantageous
and quickest method is to· use steel arch sets consisting of two halves with butt
. joints at the crown. The set can al.so be composed of more segments if required
by the size of the tunnel cross-section, or by the blasting and excavation method
used. The arches or ribs are to ·be made of I- or H-section steel or of bell section
steels usually used in mines (cf. T.H. frames and Fig. 6/14). U-section steels are
unsuitable, partly because of the small bearing surface offered to the lagging,
partly because of the asymmetry of their sections owing to which they are apt to
get twisted.
If the side walls of the tunnel adjoin its vault at an angle, and whenever large-
section double-track railway or road ~unnels are concerned, the rib and post
type is used if only to ~educe the dimensions of the single supporting elements
and to make their transportation and handling easier.
. In large-section tunnels driven in relatively self-supporting rocks the rib and
wall plate can also be used if the load from the arches can be transferred directly
by wall plates or blocks to the rock at the springing, so eliminating both the use of,
and the surplus excavation needed for setting the posts on the bottom. This
supporting method can also be used when the support is only required as a pro-
tection against rock spalling from the roof, which is most likely to occur there.
Rock spalling at ~unnel walls is, in general, smaller in extent and far less dan-
gerous, as stresses ·induced here in the rock are considerably smaller due to the
relatively large bearing surface left for the seating of wall plates. This latter support
 FULL· FACE TUNNELLING 587

b)

I or H section

,·I·

Post

II
I ~ We,// beam (pl3te )

A/temite solution with

I W3/I p/3/e
I
:-- --·+·- ·--
!
I

Fm. 6/8. Steel propping types (after PROCTOR and WHiTE)

~-
--------

588 GONSTRU CTION AND DESIGN OF TUNNELS

system, however, can by no means be applied in decomposed and

broken rocks,
adversely jointed , or in case of excessive overbreaks. It may" be extrem
ely difficult
to establish adeqiiate suppor t for the wall plate at any point above
the floor line
due to irregularity of the overbreak. Loads must be transferred in this
case directly
to the firm tunnel floor. This suppor t system can be advantageously
combin ed
with posts installed to suppor t the wall plates in tunnels with high
side-walls if
the· 'bridge action period' permits it, with a possibly more distant spacing
of the
posts than that for the roof supporting arches. It may be advantageous
even with
the denser post-spacing required by higher rock pressures. Moreov
er, it permits
the use of a full-face tunnelling method when combined with crown
beams slid
forward (see below), if the ' bridge action period' of the roof is short
and the roof
has to be suppor ted before clearing the muck. Crown beams in this
case can be
suppor ted from the wall plates. The wall plates are suppor ted, in turn,
temporarily
by cantilever brackets fastened into hitches drilled in the rock until
the posts can
be installed - following the mucking out of the debris (Fig. 6/9).
In crushed and swelling rocks the full-circle rib type should be used
as it is
able to resist pressures acting from any direction (Fig. 6/8e). A detailed
description
of this system will be given in the chapte r dealing with soft-ground
tunnelling
_methods. It should be noted here, however, that some horseshoe-shap
ed tunnels
cannot be converted easily to the full circle. Hence, because o( their
high sides
other provisions must be made for resisting side pressures and preven
ting a heave
of the bottom . The most commo n is the installation of an invert strut.
These are
placed across the bottom of each set to resist side pressure by providi
ng a hori-
zontal reaction for the post.
The heading and bench method though , strictly speaking, not represe
ntative
of a full-face method, may be usefully discussed here, too. This
method was
standar d prior to the introdu ction of moving drilling jumbos , as the
bench offered
a firm drilling 'platfo rm' for drilling the holes in the arch section
of the tunnel
thus serving its advancement as a top-heading. Nowadays this method
is only
applied when the bridge action period is insufficient for full face excava
tion includ-
ing subsequent ventilation and mucking with a consequent necessi
ty to provide
for the installation of a tempor ary roof suppor t. In such cases the
method offers
a double advant age:

(a) An occasionally loose roof can be suppor ted from this rock bench
either
directly or by an advanced crown beam immediately following the
blast, before
the erectio n of steel ribs could be commenced.
(b) Mucking can be done simultaneously with the installation of the
steel ribs
(or arches), reducing the tiine required for carrying out the cycle
of operati ons
even with a relatively short bridge action period.

With this method , suppor ting is effected in the following two stages.
First,
the steel ribs bear on the bench with, or without, the insertion of
wall plates to
be suppor ted in the second stage by posts bearing on the floor.
As a result, a
 FU LL-FAC E TUNN ELLI NG 589

Sect1011 A-A

-c; ~-~~ 1,/ - -

* Blo,ki11g belwee11 rock a11d rib

I Blocki11g betweer, crown bar ,111d rock ,111d nb

llalf s~dio11 8- 8

FIG. 6/9. Excavation process with protruded crown bars (after PROCTOR and WHITE)

bigger roof deflection and higher rock pressures are incurred than those encount-
ered in the one-stage support methods described previously.
Both bench and heading are shot out at each round (Fig. 6/7), the bench charges
being fired first. The heading shots throw most of the heading muck out into
the bench muck-pile which is under the protection of the steel support erected
after the preceding round. The first thing after blasting is to place the wall plates
and to erect the ribs bearing on them. Muck which hinders the work can be swept
aside on the top of the spoil from the bench, the removal of which can be postponed
or accomplished at once by the use of loading machines. The depth of the bench,
i.e. the length of the top heading is about 2-5 m (7-17 ft).
 590 CONSTR UCTION AND DESIGN Of TUNNE LS

In this worki ng process the type of steel suppo rt to be

used is limited to one
which includes a wall plate, either the 'Rib and Wall plate'
type or the 'Rib WaJI
plate and Post' type. This latter type may be .supplement
ed by 'truss panels ' or
'crow n beams ' to sustai n sudde nly increased roof loads
due to fauks or to any
unexpected change in stratification conditions. When excav
ating tunnels of larger
cross-section, safety can be increased by driving separa
te side drifts ahead from
the face for the instal lation_of wall plates and posts prior
to the excav ation prope r
of the heading. Anoth er modification of the metho d
consists in excavating only
one part of the upper headin g in the form of a drift for
a preceding instal lation
of the roof suppo rt (cf. mining methods).
The crown bars (see Fig. 6/9) may be built up of doubl
e steel-channels or may
be plain H-bea ms or square timbers. They either rest
upon the outer flanges of
the ribs or are attach ed to them by hangers. Their role
is actually similar to that
of forepoles (cf. Section 62.111), their doubl e purpo se
being to afford :
(a) A direct suppo rt to the roof immediately after ventil
ation •following the
blast, and thereby gainin g time for the instal lation of
ribs.
(b) Suppo rt to the roof or roof-ribs over the bench shot
thereb y relieving or
supplementing the wall plates.

(In case .o f a sudde n increase in the rock pressure emerg

ency crown bars for
imme diate suppo rt may be required. These may also
be successfully combi ned
with other tunnelling metho ds when attack ing the face.)
Crow n bars have to be
suppo rted in front by ribs restin g upon the wall plates
, and in the rear ·by ribs
suppo rt~d by finally installed posts. The manip ulatio n of
crown bars, thoug h useful
in some cases, undou btedly requires considerable extra
work at the most awkward
place in the tunnel , i.e. in the crown. If immediate suppo
rt is requir ed their use
canno t be avoided. Shoul d, however, even the wall plates
be unabl e to carry the
transf erred load, then anoth er type of auxiliary struct
ure-ha s to be used which
eliminates the crown bars and even the wall plates thems
elves. This struct ure is
the 'truss panel ' (Fig. 6/10), Its purpo se is to form, in
comb inatio n with the ribs,
a truss to span the gap produ ced by the bench shot. It
is generally long enoug h
to overs pan at least two or more rib-sets so enabli ng
the work to proce ed more
rapidly. Truss panels are attach ed to the inside face of
the ribs by two bolts at
each rib. They may be designed to carry the roof over two
bench shots, if required.
The initial cost of a set of truss panels is less than that
of contin uous wall plates,.
the labou r cost of movin g them is much less than that
of handl ing crown bars.
(The worki ng principle is similar to that of the needle
beam ; see Section 62.41.)
Memb ers of a tempo rary tunnel-supp ort, which span
over the spaces between
the main suppo rting ribs are denot ed by the term 'laggi
ng'. Lagging can be made
either of wood or steel. Wood lagging consists of hard-w
ood plank s, 8-10 m
(3-4 in) thick and 15, 20, 25 cm (6, 8, 10 in) wide, cut
to lengths slightly less than
the rib spacing. These are generaJly placed on the outsid
e flange, but occasionally
may be placed against the inside flange of the ribs and
may be spaced or tight.
 FULL-FACE TUNNELLING 591

( W3//be,;m
( fJmissible)

~,•..
:;_;;;
/ ,<;

~~~:fll;j:1fr11il~~lr:m:JJ~mil!-~-~-~
- I ;:T:j:;;~:;t,::;,-n~~t~;~:~~>li
FIG. 6/ 10. Turnout of ribs upon truss panels (after PROCTOR and WHITE)

Steel lagging may be composed of rolled channels, pressed steel channels of

various forms, rolled H-sections placed on the ribs, or of pressed liner plates
bolted together (Fig. 6/ 11 ), or of corrugated purlin-plates attached either to the
rib-flange or placed on purlins (Fig. 6/12). Frequently, trough-shaped shed-water
plates are used with or without purlins. These also maintain the distance between
sets (Fig. 6/13).
One undeniable disadvantage inherent to steel supports is that they are less
adaptable to changing rock-pressure conditions than are wooden supports.
Should, namely, the same tun.nel cross-section be maintained, the spacing of
ribs and the length of the lagging must be altered, while, when maintaining the
spacing, the structural strength of the ribs must be altered. Even so, this method
is the one usually employed.
There are, however, some self-adjustable support systems in use which are
freely adaptable for carrying increased loads. These are called yielding supports
as opposed to the rigid ones already dealt with, and are divided into the following
three basic types:
592 CONSTRUCTION AND DESIGN Of TUNNELS

Channel
Wood hgging lafh•'.'..rg,===n~,g==============-
~c-===-

a)

F1G. 6/ 11. Various steel lagging elements: (a) liner plate;

(b) spile (sheet pile, after PROCTOR and WHITE)

(a) Sliding friction supports which retain their shape while getting shorter by
sliding, so affording a high resistance against bending moments and a relatively
low one against axial forces.
(b) Articulated or .hinged supports which do not preserve their shape under
an increased load, offering, thus, a relatively low resistance against bending
moments (torsional effects) and a high one to axial forces.
(c) Combined supports incorporate the characteristic features of both previous
types. They slide when the axial load exceeds a certain limit, and deflect when
bending moments grow beyond a given value. This type of support involves con-
siderable participation by the neighbouring rock.
 FULL·FACE TUNNELLING 593

· In European· practice - especially in the mining industry, - Toussaint-Heinz-

mann rings (briefly T.H. rings) and Moll ribs are in general use.
T.H. ribs are supports of the sliding friction type with bell-shaped sections,
affording, thus, a fairly good seating besides high horizontal rigidity. A T.H. rib
consists of three parts: a crown segment and two legs attached to the crown
segment by bolted friction clamps. Their contact surfaces can slide on each other
under an increased load in proportion to the change in rock loads, and this dis-
placement can be regulated by tightenin? or loosening the clamp-bolts. T.H. ribs

~~:::Jc:_._:::a. .c:::.:_ __ ;.::i

~ {})l'rug,ted ,l,eet,i,g element,

q ==flj)
==~=Pu.='f'/in=C7
===~
= ==~
FIG. 6/ 12. Purlin-plate and steel sheet-pile laggiug (PROCTOR and WmTE)

Connection detail Melwui bottom pressure

Section wed when bottom pressure
is also acting

FrG. 6/13. Types of Toussaint- Heinzmann supporting bents used in Hungary

38 Szechy : The Art of Tunneiling

-- - - -----
~~ - -- ---= -

594 CONSTRUCTION ANO OESJON OF TUNNELS

are, therefore, typical yielding supports. The spacing of ribs varies according to
rock pressure and can be reduced to 0·35 m, with a consequent omission of any
lagging. T.H. ribs, when yielding to excessive rock pressures, imply a high safety
against failure, being especially suitable for temporary supports for a considerable
service period. In Hungary the T.H. ribs shown in Fig. 6/ 13 are used by the mining
industry for the support of transport drifts designed for long service and exposed
to rock pressures acting from all directions. As an open profile is not capable
of resisting bottom upheaval, full rings composed of four segments are used in
such cases. The exact excavation of the tunnel section, the installation of the sup-
porting rings or ribs, and the placing of the lagging require skilled and careful
work, and the recuperation of the rings must be effected before excessive rock
pressures are developed, in order to anticipate their irreparable plastic defor-
mation.
The articulated support type is represented by the Moll ribs used chiefly in
solid rocks. In plastic rock deformation resulting in twisting or buckling of the
support elements is almost unavoidable. Moll ribs used in the Hungarian mining
industry are shown in Fig. 6/14. These ribs are composed of old rails or tubes
with concave bearing-shoes welded to their ends. Articulation of the structure is

llmber log

Fxo. 6/ 14. Moll's bents used in Hungary

obtained by timber logs placed at the crown and at both spring lines between
the shoes of the respective segments and running parallel to the centre line of
the tunnel.
This support system is used again in transport drifts of a more permanent
character, where roof pressures are prevalent, and side and bottom pressures
are of secondary importance. Moll ribs must be thoroughly braced and provided
with a tight lagging. The tunnel section must be excavated accurately also in this
case and eventual overbreaks tightly back-packed. Three piece Moll ribs, A-shaped
and closed at the bottom (Fig. 6/ 14 on the left side) are used in swelling ground
(cf. adits).
 Fl;LL-FACE TUN"-ELl '- G 595

Fm. 6/ 15. Structural steel ribs to be used as rigid reinforcement

Combined types of yielding supports afford both possibilities to deformation.

There is however some difficulty in providing for a hinge-like action, as it is hard
to make it sufficiently resistant to increased loads. A combined type of support
is exemplified in the Lorenz frame.
A detailed specific discussion of the temporary support types to be converted
to rigid steel reinforcing elements of the permanent tunnel lining, may well be
omitted here as any of the previously introduced yielding steel-support systems
would be suitable to act against smaller initial loads and to become later - with
this economical prestress - incorporated as reinforcement into the permanent
concrete lining and to resist the increased loads developed in the meantime
(Fig. 6/15). It is but natural that, in order to promote a better co-operation with
the concrete, the use of sharp edged profiles should be avoided where possible or,
at least, their splitting effect should be reduced by using an ample quantity of
wrapping lateral stirrups which, simultaneously, form part of the structural distri-
butive reinforcement. ·
61.312. Wooden supports ( rimberingj . The supporting methods described above
can also be effected by the use of timber supports. The easy workability of timber,
its adapt:ibility to the variation of rock pressures, its relatively small weight,

~8 *
- =-
.,._________ - ~ - - • = -

596 CONSTRUCTION AND DESIGN OF TUNNELS

its warning squeaks when overloaded, and its apparent deformation when punched
are all in favour of the use of timber. Its main disadvantages are the ever increasing
difficulty of satisfying timber demands, its lower structural strength, the bigger
dimensions required, its liability to decay and the very limited possibilities for
its re-usage. In conclusion, lately there has been an increasing trend from
wood towards steel. This is due, however, mainly to economic reasons.

I
I

I
I· Design line
.. -- - -- -·
'f'bre,1k line

F1G. 6/ 16. Comparison of steel

and timber supports
(after PROCTOR
and WHITE)

Steel support Timber support

The use of steel reduces very considerably the number of skilled workmen
required to erect the support and it increases the speed of tunnel driving as it cuts
hours off the excavation cycle. (Steel supports can. be erected in minutes instead
of hours; they require a smaller excavation area so reducing drilling, mucking,
backpacking time and labour considerably. (The difference is clearly shown in
Fig. 6/16, e.g. 200 mm steel girders can be used instead of timber logs of 30-40 cm
dia.) _
In addition, steel encased in concrete may constitute a permanent lining and
need not be removed after the setting of the concrete.
This has again an influence upon the dimensions as is made clear in the same
:figure. Steel supports can be placed within the concrete section, while timber
must be left outside the concrete lining. Finally, there is usually some water per-
colating into the tunnel, which coupled with the moisture content of the venti-
lating air and with evaporation, makes timber always moist, if not wet. This pro-
motes decay and affects the permissible fibre stress and deformation of the
timber.

-
 FULL-FACE TU NN ELLING 59i

Otherwise, the systems for timber support for the whole tunnel section are
similar to those for steel supports, with the sole exception that its alignment at
the roof cannot be arched but must be polygonal.
Lagging consists, of course, also of timber poling boards. Whenever the exca-
vation and the support are not done at once for the' full tunnel section as a whole,
timber support finds a wider use, as will be seen in the discussion on classic
(mining) tunnelling methods (see Section 62.11).
61.313. Support with reinforced-concrete structures. The ever-spreading use of
precast and more especially of prestressed reinforced-concrete structures has
brought reinforced-concrete elements into use recently as temporary supports in
tunnels driven by the full-face method. Though this is chiefly due to the need
to economize in steel and timber, reinforced-concrete elements may claim the
great advantage that they can also be incorporated into and constitute an integral
part of the permanent lining. Their great disadvantage, on the other hand, is their
heavy weight and, in conjunction therewith, the difficulty of their handling and,
installment. In small section tunnels their load-bearing capacity cannot be fully
utilized, and in large tunnels the weight of reinforced-concrete ribs increases to
such an extent that their erection necessitates the use of mechanical hoisting
cranes. An attempt has been made in some cases to apply precast reinforced-
concrete roof pieces 6.4 for the internal support of larger underground halls. These
structures, however, have not found wide application in practice, chiefly because
their weight and their laborious and difficult installation did not allow them to
compete with simple.brick masonry or with in-situ cast concrete vaults.
61 .314. Rock bolting. The field of application of the full-face tunnelling method
has recently been extended by the development of the roof bolting or rock bolting
method. This is applied mainly not as a temporary, but as a permanent support.
The main points of this method are, according to RABCEw1cz 6· 0 (Fig. 6/ 17),
as follows:
When advancing the tunnel face by a length / 1 (the pull attained in a single
round), the half dome like double arching effect (described in Section 35.1) will
be produced as a consequence of the loosening and expansion of the overlying
strata (cf. Figs 3/47 to 3/49). The time for which this pressure arch will give
a reliable support, i.e. the bridge action period, depends upon the geological con-
ditions, the span /1 and the diameter of the tunnel. It will last for some hours,
however, even in weathered or strongly Jointed rocks for spans used in practice.
If, however, no support should be installed within this period the rock in the roof
will begin to spall off, involving a gradual rise of the pressure arch in a tendency
to obtain stability by decreasing the spans in conformity with the bearing capacity

6
-4 Cf. JAN0s1, J.: Eloregyartott vb. elemek hasznalata foldalatti csarnokok fodemszerkeze-
tenel {The adoption of prefabricated reinforced-concrete elements for roof structure of under-
ground halls), Ep. es Kozl. Tud. Kozlemenyek II J 55
s.5 R ABCEWICZ, L. : Bolted Supports for Tunnels, Water Power 1954 150 and 1955 452
 598 CONSTRUCTION AND DESIGN OF TU'INE'--S

B a)
l,
Hoof bolt~d
arch

clvmao

b)

FIG. 6/ 17. (a) Principle of roof bolting after RABCEwtcz;

(b) spacing of -roof bolts to form a flat arch

of the rock. The purpose of roof bolting is now to set anchoring bolts of lengths
exceeding the thickness of this fractured zone radially into the still undisturbed
rock layers within the bridge action period, to establish by this a resistant rock
vault of suitable thickness able to counteract rock pressures, and to anticipate
thereby the successive ravelling of the overlying rock masse·s.
Experience shows that this natural roof arch, liable to progressive ravelling,
has a parabolic shape and its thickness does not exceed /1 /2. Therefore, the length
of the roof bolts cannot be less than /1 or below the double thickness of the assumed
deteriorated zone. For safety measures, however, the length of rock bolts / must
be bigger than Ii, I> b/3, or I> b/4, as it is reasonable to bring the bolt length
in line with the width of the cavity. This protective rock vault so established will
·b e effective only for the time before it disintegrates. With due regard to this,
the bolts must be installed and fixed within the bridge action period and the bare
interfaces between rock bolts should be lined with a wire mesh and sprayed with
gunite.
 FULL-FACE TUNNELLING 599

Thus, with the aid of rock bolts a load-carrying ring is formed from the rock,
in which not only tangential but also radial stresses are induced, these latter:
having a favourable effect on the stability of the rock vault. With regard to roof
bolting TAL0BRE6 · 6 makes the following observations:

F1G. 6/ 18. Natural

rock arch produced
by roof bolts

1. The extent and dimensions of roof bolting depend upon the nature of the
rock. Roof bolts in less firm rocks are to be spaced more closely and must also
be longer. Plastic ground is completely unsuitable for roof bolting.
2. The thickness of the load-bearing rock vault to be formed must be able to
resist and carry the expected loads.
3. The length of the rock bolts should be at least equal to the thickness of the
rock vault plus the mean distance between adjacent bolts (Fig. 6/ 18).
4. The spacing of rock bolts should be as uniform as possible.
The dimensioning of roof bolts and their spacing must conform t~ lhat of the
rock vault, i.e. the thickness of the vault required to c2rry the expected loads
must be calculated first.
Let us, for example, assume that a cavity of 2·5 m (8 ft 4 in) inner radius is
loaded by 3 ton/m2 • A rock vault of 1 m thickness is to be formed around the
excavation by using roof bolts 2 ·5 m in length, thus establishing a ring-shaped
rock vault of an external radius of:

I 2·50
R = r +- = 2·50 + - - = 3·75m.
2 2

6·6 TALOBRE, J .. : La mecanique des roches. Ounod, Paris 1957 254

600 CONSTRUCTION AND DESIGN OF TUNNELS

The normal force acting on this ring will be:

N = pR = 3·0 t/m2 x 3·75 m = 11 ·25 t.

Assuming now (after Section 32.23) that the minimum radial stress relative to
the normal force of 11 ·25 tons will be 2 t/m 2 , a force of 1·5 2 x 2 = 4·5 tons will
fall on each rock bolt if spaced at 1·5 m for structural reasons. Consequently,
the bolts are to be prestressed, for safety reasons, by a force of 2 x 4·5 = 9·0 tons.
Though a roof-bolted rock vault, clearly, can be employed only in arched tun-
nels, yet roof bolting can also be satisfactory in flat-roofed tunnels. In the latter
case, the composite action of the respective rock layers is secured only by contact
friction. Nevertheless, its magnitude can be considerably increased by roof bolts,
if applied as shown in Fig. 6/17b, when they will effectively resist principal tensile
stresses due to shear, similarly to the
bent-up reinforcing bars in a' concrete
beam.
The spacing of the individual roof bolts
must be determined according to their
Wedge diameter and length. If, e.g., a load of
12 tons (equal to the weight of a rock
volume of 4·5 m 3, having a density of
2·65 ton/m3) is to be transferred safely to
a 25 mm dia bolt 2·50 min length, each
bolt will be capable of anchoring 1 roof
12
area of = l ·8 m 2 • In inclined
2·5x2·65
or vertical stratifications, or under other
particular geological conditions, the
thickness and dip of the layers and the
nature of the rock will be the govern-
ing factors.
According to their system of anchor-
ing, rock bolts can either be simple steel
bolts with slotted ends used in conjunc-
tion with wedges, or special steel bolts
fitted with 'expansion shells' (Fig. 6/19).
Haring bolts Slotted-end bolts with their inserted
-ir-o; t8ncr Conical wedges will be pushed into the bolt holes
head and then pressed · a~ainst the bottom of

~i ~l_lng
~ 1:> . .s/Jell
FIG. t!,/19. End fittings of roof bolts
 FULL·FACE TUNNELLING- 60}

the bole and hammered. As a result of this, their slots willexpand. With the other
type, a shell or a double wedge encircling the head of the bolt will open up on
tightening the bolt. The success of rock anchoring will depend upon whether the
deformation of both the bolt and the rock can be avoided or not. If it were possible
to prevent the slightest crumbling apd subsidence. of the rock, a load-bearing rock
vault could be formed by rock bolting theoretically even in pure sand. This method
of roof suspension, i.e. the formation of a natural rock vault, is in principle similar
to the pre-stressing of load-carrying steel stuctures. Another method is the suspen-
sion of loosened lower roof layers or of a tunnel lining of inadequate load-bearing
capacity on firm and load-bearing upper r-0ck layers. This latter procedure depends
entirely on local geological conditions, and can be considered as the supporting
of the tunnel lining or lower layers of inadequate load-carrying capacity by rock
bolting, considerably reducing th~reby ~he.bending stresses induced in them.
The preceding considerations attribute the beneficial effect of rock bolting either
to the anchorage of loosened lower roof layers into the upper lying solid one, or
to the establishment of a load-carrying ring brought about by the tensioned bolts
holding the fractured mass together and thus increasing considerably its compres-
sive and bending strength. 6 ·6 a
R. RICHTER6· 6 b claims however that the real merit of roof bolting is lying in
an essential modification of inner stress conditions within the affected area, which
is secUFed by the fact that the applied bolts are impeding and restricting consider-
ably the inner strains of the mobilized zone and thus lead to an essential reduc-
tion of inner stresses, which are responsible at last for inner stability and bearing
capacity.
This assumption seems to be somewhat proved also by the experiments of
TrNCELJN 6· 6c who found that inclined and pretensioned bolts are more efficient
than normal ones.
As to the dimensioning and spacing of rock bolts STRAKA6 · 6d has given also an
approximate method, which is giving somewhat more theoretical support to actual
computation work, although the most reliable way for the determination of the
bearing capacity of a rock bolt still remains the previous pulling test executed
at the site.
Ks:.;~nt tests with roof bolts have shown that their position and, thus, the measure
of their prestress cannot be regarded as constaftt iu time, as the bolts tend to slide
gradually outwards. This process is rather fast at first, and slows down gradually.
In the first 'sudden' phase of this ~lip, a 'play' due to the actual incorrect setting
•-•• See No. 5166. R eport of Investigations U. S. Bureau of Mines, 1'956 March 25
•-ob RICHTER, R.: Grundlegende Betrachtungen zum Ankerausbau. Bergbo.uwisssenschaf-
ren, 1964 11
idem. About Rock Bolting. Pub/. of the Techn. University for Heavy Industry, Vol. XXIII
Miskolc, Hungary, 1963
6•6t TINCEUN, E. et. al.: Le soutenement suspendu ou boulonnage. Reveue de l'lnduftrie

Minerale No. B. 2.15.4. 1961

s.sd STRAKA, J.: Kotveni stropu podzemnich vyrubu. (Support of subterrainean cavities by
roof bolting). Inzenyrske Stavhy, 1963
,...-----

....._____
--- - --- . --
- - -- - - ~ - -- - - - -- - -- - -

602 CONSTRUC TION AND DESIGN OF TUNNELS

of the bolt may have a considerable part as well as the elongation of the boll
under stress, while the following flattening phase characterizes the typical resi-
dual deformation of the rock itself. This phenome non clearly demonstrates that
the permissible stresses used for calculation in the determin ation of the area and
spacing of bolts, must be taken at a much lower value than the tensile strength of
the bolt material. Considering that the tests have also indicated that the displace-
ment, i.e. the extent of ·stress relief of the bolts,. also depends upon the quality
of rock, being the greater the softer the rock, consequently, the permissible stress
will finally depend on the quality of the rock. Fig. 6/20 shows the recommendations

25 ~~ ,~ -I~,
- ~fl-
-
Tensile strength ofsteel

·c.,
~ 15 1--¼-J>--+ -----l

10

FIG. 6/20. Admissible loads on

bolts as a function of
0 L - - L --~...Lt}_ __.___j20L--'--J0-'-----'--40--'----' -__J rock quality
50
(after RABCEWIC Z)
Oispl.:;cement (mm)
of RABCEWJ CZ - based upon his test results- relating to 22 mm dia high-tensile
steel bolts anchored by expansion shells. As can be seen, stresses approximating
the tensile strength of the steel can be attained only in the firmest rocks. In addi-
tion , it has also been demonstrated that roof bolting in rocks inferior in strength
to shales without sand will be ineffective. In accordance with this, recent regula-
tions and codes as to the dimensioning of roof bolting are already based on pulling
tests and on the elasticity and strength characteristics of the basic rocks. 6 · 7
The lower the bearing capacity of the rock the greater importance should be
attribute d to rock bolting and the more difficult is its application. In such rocks,
the use of small.er diameter and more closely spaced roof bolts is advisable to
increase the specific adhesion force and· to reduce the possibility of the spalling
out of the intermitt ent rock surfaces. Another method is to resort to the use of
reinforced-concrete anchors. Reinforced-concrete anchors were first used to elim-
6 7
• H uooN, A.: Le boulonnag e des roches en sou
terrain, Le Monde Souterrain , 1960 April 41
H uooN, A . and CoSTES, A., Edition Eyrolles, Paris 1959
S CHURMANN : Richtlinie11 fiir d en A11kerausb a11. Gliickallf 1960
3
 FULL-FACE TUNNELLING 603

inate the upheaval of concrete road pavements placed upon squeezing clayey
soil, then to stop rock squeezing in tunnels excavated in the same kind of soil.
The left side of Fig. 6/21 shows the inward squeeze of the rock into a circular
tunnel without reinforced-concrete anchors, while, on the right side the same is
shovm with reinforced-concrete anchors, for comparison. As can be seen, there
is a considerable difference which might be explained by the fact that when
proceeding inwards to the ground the specific squeezing value is diminishing as
a result of moisture or the increase of rock pressure, with a limited extension

Boumhr!lof Reinforced-concrete ancl!ors

/ squeezing
,,4--/ motion
I
I
I
I
f
r ---,
bl

I
C

I
h I
i\ I
I\ I
I '' I
I
I ', I I
I 'i , ofouthreak

llJl
i ' ' .,,,
I I -~
1 I '-..

I :,.<>
~
.......
Witnoul support Anc/,ored /Jg reinforced-concrete plvgs
~ - lbtl!f squeeze
1
al point ':1
6 - Toll/ sqveez~
Fm. 6/ 21. Effect of reinforced-concrete plug anchorage upon the inward squeeze of ground

zone of squeezing to a distance R. Consequently, if the anchorage extends beyond

this distance, the inward movement of the rock can be considerably slowed down,
or even totally prevented. This experience is of great significance when rock pres-
sures due to squeezing have to be counteracted (cf. Section 31.3). As such pressures
reveal themselves mainly at the sides and at the bottom of the tunnel, reinforced-
concrete bolt anchorages will be most advantageously applied at these places.
According to measurements made in the coal-mines in the German Ruhr district
the floor upheaval of a mine drift 2·40 m (8 ft) wide when provided with rein-
forced-concrete bolts amounted to 10 cm (4 in) only, while that observed on
a section of the same drift without anchorage was 140 cm (4 ft 8 in).
-- - ------------
- -
- ----

604 CONSTRU CTION AND DESIGN OF TUNNELS

Besides the setting of roof bolts, the wire mesh cover of the interme diate surface
is also of great importa nce, not only to provide for immedi ate protecti on
against
spalling rock pieces, but also to afford a suitable base for subsequ ent guniting
or concrete coating.
Summin g up, the following advanta ges may be attribut ed to roof bolting:
1. The applicab ility of full-face tunnelli ng "without support is extende
d to rocks
of lower strength , making the undistur bed and econom ic use of up-to-da te machin-
ery possible here also.
2. It eliminates the use of timber in tunnelling which is significant not only
in
econom y in timber consum ption but also with regard to the reductio n so obtained
in the dimensions of the excavated section. As timber is not allowed to be
incor-
porated into the perman ent lining, an addition al overbre ak is necessary
beyond
the tunnel cross-section, incurrin g tedious and delicate back-pa cking operatio
ns.
3. It is even more advanta geous than the use of steel support s, as the erection
of steel structur es also requires the excavat ion of greater cross-sectional
areas.
In addition , the workers are always disposed to blast for overbre aks in order
to
avoid the cumbers ome and unecono mic work of trimmin g. Moreov er, no support
s
will be exposed to the damage caused by the rock pieces ejected by the
blast.
4. Rqof deflection and subsidence, althoug h not completely avoidab le
on
account of bolt-slips, will be much smaller than when using internal support
s,
especially when the whole tunnel section is excavat ed by steps requirin
g the
removal and replacem ent of support s (cf. Section 62.1 and Fig. 6/24).
RABCEWICZ made an econom ic compar ison between two sections
of an aqueduc t
tunnel construc ted in New York (see Table 6/III), one of which was excavate
d

TABLE 6/llJ
--- --- --
Steel supports Roc k-bolting Savin
Item Unit price -
Un it
(Swedish cro wn) Quan- Quan- Qua n-
tity C ost Cost C o st
tity tity

Steel kg I 520 520 75 75 445 445

Timber m• 250 -
O·J 25 - O·J 25
Working hours hours JO 137 1370
~
- -- - -
108 1080 29 290
-- --
Excavation m• 58 15·5 588 14 812 1·5 87
Concrete m• 87 4·1 356 2·6 226 1·5 130
'
Shutterin g m2 7 12 84 12 84 - -
Sum total Swedish 3254 2277 ~77
crown
' I i
 FULL-FACE TUNNELLING 605

under steel supports, and the other under roof bolts (Fig. 6/22). As can be seen,
the saving in steel amounted to 80 %, that in working time to 20 %, and in excava-
tion work to 20%, giving a total of 30% saving in costs calculated on the basis
of Swedish price conditions.

FIG. 6/ 22. Alternative methods with

steel propping and roof
bolting for a water-supply
tunnel

It must finally be mentioned that roof bolting can only exceptionally and in very
firm and impermeable dry rock be considered to act simultaneously as a perma-
nent support, too. It cannot permanently replace the tunnel lining chiefly because
of the danger of corrosion of its bolts . If, however, the corrosion of the bolts is
prevented (e.g. by a water-tight and air-tight coating or by the use of stainless
steel, etc.), and the deformation process of the bolted rock can be considered to
be completed, the construction of a separate tunnel lining can be dispensed with.
A more recent and very important field of application of roof bolting is to drill
the bolt holes far beyond the circumference of the advanced heading into the sur-
rounding rock sufficiently long to penetrate even beyond the perimeter of the final
tunnel cross-section, thus providing a satisfactory anchorage-support for the
same6 ·8 (Fig. 6/23a). As the stability of an unsupported roof depends on the length
of span and on the elapse of time a~ well as on the strength characteristics of the
rock, the drive of small section headings affords great possibilities for this extended
anchorage in Iese; solid ground. The roof of this heading will be self-supporting
for a bridge-action period sufficiently long to permit the installation of roof bolts
from it extending well beyond the final perimeter of the tunnel cross-section. As to

u Cf. VoLUMARD, P . and BASTIDE, A.: Ancrage du toit en galerie avant abattages, Travaux
J957 October
- -- - -
-- - -----------~----
- --

606 CONSTR UCTIO N AND DESIGN OF TUNNELS

Steel arc/Jes

F1G. 6/23. Roof bolting

accomplished from
bottom drift

the bridge-action periods for given spans in various rocks, some general infoi ma~
tion is given by H . LAUFFER6· 9 in Table 6/IV.
61.315. Support by gunite-rendering. A column in Table 6/IV shows the possi-
bilities of guniting as a means of support (cf. Section 4.62 and Fig. 7/ 10). In relation
to this it must be mentioned 6· 10 that recently gunite rendering has been successfully
utilized as a final support (lining) in addition to its original object of affording
a simple insulating or protective coating against rock spalling and popping incurred
by the redistribution of rock pressures. The considerable development of gunite
installations has allowed a spray of aggregates with grain sizes up to 25 mm (l")
to be used. In addition, the admixture of suitable chemicals affords better adher-
ence, more rapid hardening and a greater thickness of the rendering. According'
to LAUFFER a relatively thin gunite coating offers the following advantages:
1. Owing to its relatively high tensile strength the gunite rendering practically
constitutes a quasi-reinforcement of the involved 'rock bea m' which has compres-
sive strength only, so enabling it to resist both bending and tensile stresses. This
effect is considerably increased by inlaying a wire-mesh web.
6·9 LAUFFER, H .: Neuere Entwicklung der Stollenbautechnik, Osterr. Bauzeitschrift 1960
January
1-1o Cf. ROTTER, E.: Anwendung von Spritzbeton, Schriftenreihe des Ost err. Wasserwirtschaft-
verbandes 35, Springer, Vienna
Field of Applicatio11 of Modem Support Systems TABLE 6/IV

Class of ground and the I Bridging time and I Gunitmg Rock-bolting Steel supports burriccl in
I supports usually applied span of the ground
. --·-
the permanent lin ing

A Firm 20 years not reqiured not required not required

4·0 m
·-· ·-
B Loosening in time 6 months in a layFr thickne,s of2 to 3 cm, spaced at I· 5 to 2 m and using uneconom ic
(above head 4·0 m but only in the arch wire mesh but only in the arch
protection)
--
C Slightly friabl e I week in a layer thickness of 3 to Scm, spaced at l·0 to 1·5 m, only uneconomic
(roof supports) 3·0 m but only in the arch in the arch, applying either
wlre mesh or subsequent gunit-
1
ing in a thickness of 2 cm "'C:
r-
r;..,
D Friable 5 hours in a layer thickness of 5 to 7 cm. spaced at 0·7 to l ·0 m, mairlly occasionally, in the same way )>
()
(sets of light supports) J·5 m sprayed on wire mesh main ly in the arch, applying both wire as under £; m
in the arch mesh and subsequent guniting ,-;
in a thickness of 3 cm C:
z
z
m
E Considerably friable 20 mir,utes in a iayer thickness of 7 to 15 to be applied only after the steel or concrete slabs placed r-
{;;ets of heavy supports) 0·8 m cm, sprayed on wire m,;;sh setting of the roof (temporary)
!:
behind ncel arches z
supports in cases where rock a
bolts spaced at 0·5 to 1·2 m
can offer a supporting effect at
all and immediately followed
by guniting in a thickness of
3 to 5 cm
I
F Immediately exerting ground 2 minutes in a layer thickness of IS t o 20 not to be adopted steel slabs placed behind struitcd
pressure (forepoling without 0·4 m cm, sprayed on wire mesh steel arches, with the appHca-
the use of face supports) supported by steel arches (in tion of subsequent guniting
some cases also the face is
gunited)

G Immediat~y exerting heavy 10 seconds n ot to be adopted not to be adopted steel slabs placed behind strutted
ground pressure (forepoling 0·15 m steel arches and immediatel y
and face supports) gunited 8l
.....
. ~~-- -
---
c:c -

608 CONSTRUCTION AND DESIGN OF T UNNELS

2. As a result of guniting, the rock - as with rock bolting - becomes prestressed

to a certain extent and so promotes the formation of a load-bearing vault.
An additional important effect, though not a structural one, is that rocks sus-
ceptible to disintegratio n or swelling when exposed to weathering or to the action
of moist air become immediately sealed off by the rendering from these adverse
effects, thus preventing their development.
3. Another great advantage of this method is that gunite coating can easily
be repaired when fissured and when the cavity is excavated in parts, each part can
be temporarily supported separately.
According to the experiences described in the publications mentioned above,
gunite rendering can be applied as a support also in detritus, crushed rock, or even
in clayey soils with a certain cohesion (cf. Table 6/IV). Experiences as to the general
validity of this supporting capacity are, however, incomplete as yet.
61.316. Water exclusion by grouting. In waterlogged, fissu red rocks, especially
in karstic limestones, tunnelling can be undertaken also under the protection of
a grouted watertight rock-mantle. In such cases the main task is to exclude water.
inrushing under high pressure, the rock otherwise being quite firm enough to
allow full-face tunnelling with the minimum use of supports. The main points
of this method 6 ·11 are clearly shown in Fig. 6/23b, according to which the working

Jal'eL.!f space Prepad ed concrete

lunne/ profile fr .51,arting grout
--- enclosure

first Concrete bu&l!e;;d·

Second enclosure curtain enclosure curtain

FIG. 6/23b. Tunnel construction scheme under the protection of ~ grouted artificial shield

place, where waterlogged rock is encountered , has to be solidly packed with rock
debris to a certain length and dammed off by a solid watertight dam made of
reinforced concrete. From behind this dam, injection holes, inclining somewhat
outwards, are drilled in a ring around the planned periphery of the tunnel, the
length of the boreholes varying between 5 and 20 m according 'to local conditions.
A cement-bent onite, or, in looser ground, a grout of a suspension of sodium

• · 1L BARBEDETTE, R.: Percement de galeries en terrain difficile, Le M onde Souterrain 1959

Aprii.
 MJNfNG METHODS 609

silicate is injected under very high pressure (up to 150 atm). When the grout bas
hardened, the tunnel face will be advanced up to some metres behind the front
of the injected zone, where, for safety's sake, it must be stopped again for the
drilling and injection of the next ring of injection boles.

6.2. TUNNELLING
IN MODERATELY FIRM ROCKS AND GROUND

6.21. TUNNELLING BY SUCCESSIVE EXCAVATION AND LINING

OF SMALLER INDEPENDENT HEADINGS
(CLASSICAL OR MINING METHODS)

It has been clearly demonstrated that the smaller the cross-sectional dimensions
of an underground cavity, the less dangerous is its excavation, and the longer is
the natural bridge-action period of the rock. Thus, it is an obvious and a long-
established procedure to excavate the tunnel's cross-section not in full-face at
once, but in smaller parts by the driving of smaller- specially arranged individual
headings. The arrangement and sequence of these headings should always be
adapted to the necessary operations to be carried out in them (excavation, installa-
tion and construction of temporary and permanent lining, etc.) and to the nature
of the rock, and also to the rock pressure conditions encountered.
It must be noted here that tunnel construction in parts implies the subsequent
and repeated erection, dismantling and removal of various temporary supports.
This procedure requires not only much surplus work and a considerable additional
consumption of support material but also gives rise to additional roof subsidence
and repeated loosening of the overlying rock, which results in an increase in rock
pressure; all are unavoidably inherent in each dismantling and change of tem-
porary supports (cf. Fig. 6/24).
With regard to these disadvantages, the part-excavation tunnelling methods
(usually called classical tunnelling or mining methods) are being more and more
displaced from up-to-date tunnelling practice.
6 12
RABCEWICZ, in one of his papers, • makes a comparison between mining
(Austrian) methods, and up-to-date steel support, and roof-bolting tunnelling
methods. First of all, he points out -the advantages of full-face tunnelling and com-
pares the respec6ve roof deflections, settlements and rock pressures taking place
during the excavation and lining of the entire tunnel section. He concludes that
while roof settlements of the order of 20-80 cm (8 in-32 in) may be expected with
the segmental arch-timbering (Austrian) method, and, accordingly, roof loads at
least equal to the width of the tunnel excavation are to be considered, but roof
settlement in the full-face tunnelling method with up-to-date steel supports will
6 ·12 RABCEWICZ, L.: Effect of Modern Constructional Methods on Tunnel Design, Water

Power December, 1955

39 Szecby: The Art of Tunnelling

~ ---- - ------ ------------ - -

610 CONSTRUCTION AND DESIGN OF TUNNELS

\~----
''
''I 4
\
''

Jectton 8-8
L11 Oefletlion or crown b.Jrs !. BotLom huding
L12 Sliding by unfjvour.tble dipping 2. Top heading Note , !J,.1nd Ll.1 incrused by softe·
orJointing ning of rock 1173lerial by w.,ter
J. fnl.1rging upper heading
.1nd by loosening from bl.istin;
4-. Enla@illJJ lower he.1din!I
FIG. 6/24. Superposition of crown subsidence brought about by repeated installation and
dismantling of temporary supports in the Austrian method

be but a few centimetres, resulting in a rock load equal to the weight o~the rock
layer of half the excavation width or of the unsupported span / 1 only. This might
also be accounted for by the advanced mechanization in this latter case through
which a cycle of operations can be completed in about as many days as was pre-
viously required in weeks. Finally, the roof settlement in the case of roof bolting
must be even smaller than in the previous case, and no void can be left between
support and excavation-line, and consequently both the measure of loosening
and that of the loosening pressure must be less than with steel supports. This
cannot exceed, in general, the distance / 1/2, with the greatest possible value being
not more than the length of the roof bolts, /.
Another significant procedure used in up-to-date tunnelling practice is the solid
pressure grouting of the voids left behind the overbreak line and.lining as against
the loose back-packing applied with mining methods. Pressure grouting provides
not only a complete back-filling and, thus, a better support for the surrounding
rock on all sides of the tunnel, but, by filling up the inner seams and joints of the
rock, it substantially increases and stabilizes the strength and load-bearing and
arching capacity, and the modulus of elasticity of the rock itself. At the same time
it provides a certain watertightness as well as a useful prestress in water supply
or pressure t unnels. The present endeavour is to establish tunnel constructions'
suffering no deformations separately but only in complete unity with the sur-
rounding rock (see Sections 42.3 and 43.3), and resisting stresses also in that way.
 MINING METHODS 61l

RABCEWICZ compared for a specified case the classical Austrian and the combined
Austro-Belgian part excavation methods (see Section 62.34) with full-face methods
usfog steel supports, and rock bolting, respectively. The tunnel profile is shown
in Fig. 6/25a while the graph of costs can be seen in Fig. 6/25b. The inner width
of the tunnel is 9 m (30 ft), with·a permissible stress for the lining of 50 kg/cm2

SLee!arc!,es Austrian method

Improved A11sLrtan obsolete design
method
,, .t
""~ ~
~
]j
cs ~
f{}/
88
6·5
100 " I /5!f4

1128
1000 ...............
! M,1terta/
-- - -----~---- --- 0 L__..J.__ ___1._ _ _..1.__ ____._

Metl/ods
Fie. 6/25. Comparative data for a given example

(700 lb/in2) without being grouted, rising to 125 kg/cm2 (1680 lb/in2) when pressure
grouted. In Tables 6/V and 6/VI the arch thrusts, wall thicknesses, material, work
and cost demands are given in Swedish Kronor for 1 linear metre of tunnel cal-
culated for a rock density of y = 2·5 tons/m3• Though the data given by RABcEWICZ
may not be considered to be generally valid, nevertheless it can be stated that
full-face methods affording immediate support are essentially less costly than
either the Austrian method or its improved type combined with the Belgian
method.
Temporary excavations of smaller cross-sec1ioilal area than that of the tunnel
are termed ' headings' or drifts when closed within the rock along their whole
periphery, and 'cuts' or 'stopes', if they have no roof. Headings are connected
by rises or chimneys (if driven upwards) and dips (if driven downwards). The
tunnel's cross-section is divided into the following three main parts: the top
(arch) section (calotte), the core (strozze) and the bottom section (Fig. 6/26).
The top heading and the roof arch are set in the top section, while the core is
bounded by the side walls, and the bottom heading and the invert are placed in
the bottom section.
The excavation of a horseshoe section tunnel can be carried out according to
the patterns showu in Fig. 6/27. The methods can be well cla~sified into the two
----
-- - - - - - - -

612 CONSTRUCTION AND DESIGN OF TUNNELS

TABLE 6/ V

Segmental arch
FulJ-section excavating methods
(Austrian method)
Thrust Wall thickness Steel posts and ribs Roo f-bolting
--

-s.-~I r: -~ I
Quan- E_
. I <J-; Quan-
.;
~ - ti -; Quan-
-5 ·~ f .~
0" 0 "
tity r,. • tity r,.- tity
--
---
(t) (cm) (t) (cm) I (t) (cm)

Circular arch Vertical rock

pressure 137 27 27 75 15 25 42 8 15
------ - - - -- -
Pressure
grouting 134 11 134 11 134 11
Circular arch Inclined rock
-------- -- ----
pressure 144 29 29 81 16 25 15
Pressure
- -- ------ - - -45- 9

grouting 134 11 134 11 134 11

---- - - - - --
Composite Vertical rock
curve arch pressure 175
-- -- -- --
35
I
35 93 19
I 25
- -
52 I 10 16
Pressure
grouting 188 I 16 198 16 198 1 16
I I
TABLE 6/ VI

·Method
Austrian
Improved
---1 Supporting
with steel ribs
Roof-bolting
Classical (Belgian com-
bined)

Item I
Unit
Costs in
Swed. Kr.
Quan-
tity
Unit
price
Quan-
tity
Unit
price
Quan-
tity
Unit
price
Quan-_
tity
I Unit
price
-----~
Working time hours I 6 379 2274 I 254 1524 149 894 106 I 636
- - - - - - - - - - - - - - -- --- - - -
Shuttering m2 2·6* 20 104 20 104 20 52 20 52
Concrete ma 5·2** 20·7 1035 10· l 505 8·8 440 6·6 330
50·0
Pressure
grouting ma 110 ••• 4·2 462 2·2 242 1 ·o 110
ma
Over break
_____,___ - - -
28 13-4 375
---
3-1 87 2
- - - - - - - - - - --
56
---
Labour costs value 2274 1524 894 636
Material costs value 1514 1158 790 492

Sum total Swed.

I K r. 3788 I 2682 1684 1128

Notes: • Steel shuttering; •• Wood shuttering; • • • No pressure grouting computed.

 MINING METHODS 613

following groups, depending Tophl'admg(drif t} _i=--1-:::.._.,,/'~ ..qoof :re,?

X7 I I , ·,
on whether the lining is con-
structed in one single course
// / / I
// . . ..>t:"r-~--
j
-- '-....._.,,,-frtrados
I '-.....__ "
or in subsequent courses. In I / L__ .--..I '\ \
the first case, the lining can I I flea mg \ \
be constructed by one of the
following methods:
(a) The method most fre-
(
/ /
...1.. I
---ti-- - -
--+I
(Top seC'tion)
Intrados
Sprmg lm-e--,r
,) \
\ ~
,,
I drif't ~ 1
quently used is to break up with I y s,de wall I II
a rise from the bottom head- I
I
'I
'
Abutment w.111 "
,.,ore
L1-_
I
_ +j
I

K
ing to the top section and to 1
\ (Strom) 1
excavate a ring in a predeter- I \ · · J Bottom /
1

lr __ _yS1de drtft lleac!i ng I I

T\
1
mined length. The excava-
1 (drif t) J I
tion is carried out from the
II
l 1
~ IJ .. \\ I I
top to the bottom according LL _ L.J............ Bottom section __.......-7\.>- _J
J

to the pattern shown in Fig. -----..;;;

' ...... -~-+ ~-
-= ..... .............. -
6/27a, in which case hauling - - - i - - - - -'-Jnvert arch
(in heading No. 1) and exca-
vation (starting from heading F IG. 6/26. Terminology of tunnel ? arts
No . 2) can be carried out in-
dependently and simultaneously. When the excavation of the ring is completed
in this given sequence, the lining will be erected from the bottom to the top un-
interruptedly within the whole width of the excavated ring.
This method has the advantage that not only does it afford good transport
conditions but also by its use a considerable number of working faces are possible
and, consequently, there is a high rate of advance. Its disadvantage lies in the
insufficient ventilation of the working drifts.
{b) Another frequently used method is the benching (or strozze) method with
successive underhand stoping in the sequence given in Fig. 6/27b. In this case also
the lining will be constructed from the bottom upwards. Here, haulage and exca-
vation cannot be entirely separated and a frequent translocation of the rail tracks
is necessary but, on the other hand, good ventilation and clear and undisturbed
working conditions are assured.
(c) Finally, the third method of the first group comprises successive overhand
sloping upwards from below in the sequence shown in Fig. 6/27c. It. may be used
only in firm rocks which do not require any roof support. The lining masonry is
again constructed upwards from below in one single course as in the previous
cases.
The methods of the second group include the effecting of the excavation
and the lining construction in alternate cycles.
(a) One alternative is the top section excavation (Belgian) method in which
the top section of the tunnel is excavated first starting from a top heading in
a predetermined ring-length, to be followed by the erection of the top arch sup-
ported by the rock itself or by wall plates. This is followed by the excavation of
612 CONSTRUCTION AND DESIGN OF TUNNELS

TABLE 6/V

Seamental arch Full-section excavatina methods

(Austrian method)
Thrust , Wall thickness Steel posts and ribs 1· Roof-bolting

Qt~::· 11] I K] I Q~~yn- j ] I K] t Q~~- l~ I!]

(t) (cm) (t) (cm) I (t) (cm)
--- -~----- ---~c- -- I
Circular arch Vertical rock I
s u_r_e_____ _ 1_3_7
pr_e_s_
___ _E__,_E__ ~ ~ ~~ __8_ ~
I
1 11

Pressure
grouting 134 11 134 11 I 34 11
1

1n
-C-ir_c_u.,.la- r- -a-rc_h____ ,-_-_ _--i;-t_-_
-P_c-r~ re-_r-_
ud-_ ck-_-
o-_ 14:4= ==2=9-
_-_ : : 29 81 I 16 25 _ 4_5_1=~-9-- __1_5_
1 1
Pressure
_I_11
I 34
I_
_ u_t_in_g _ _ _1_34 _1_ 1_1_ _ _ ~~ 1_1___
g_ro _ _ __
Composite Vertical rock
curve arch ___p_re_s_s_u_ r_e______1_7_5___3_5 ~ ~ _ 19 I _E__ 52 10 16
1

Pressure
grouting 188 16 ' 198
I 16
I 198 16

TABLE 6/ VI

- Austrian I d - I -Supporting ~ of-bolting

·Method
Classical
I
mprove
(Belgian com-
bined)
I
. h t l 'b
Wit s ee n s

------~i----
Item
C-os_t_s
Unit _ Swed . I uan-
ity
I Unit
price
I Qua n-
t~
I Unit
pri~
I
Qu_an-
ty I
Unit
price
Qu_an- , Un_it
uty . pncc

hours 379 2274 I 254 1524 149 _1_8~ 6 G 3~

Working time 6

Shuttering
Concrete
~
I m3
2·6*
5·2..
20
20·7
-1 104
1035
20
10·1
104
505
20
8·8 440
52 20
6·6
52
330
50·0
Pressure
I
grouting m3 110 ••• 4·2 462 2·2 242 l·0 110

Over break m3 28 13-4 375 3·1 87 2 56

Labour costs value I 2274 1524 894

790
636
492
1514 1158
ts_ ,_v_a_1u_e_. ___ __ __ _
_M_a_t_e_r_ia_1_c_o_s_
Sum total Swed.
I Kr. 3788 I 2682 1684 1128

Notes: • Steel shutterina; •• Wood sbutterina; ••• No pressure grouting computed.

 - -

-- - - - - - - - - - - -- - -~ ~-~-~~

HEADINGS AND DRIFTS -615

a central cut from which narrow cross-cuts are made to both sides at intervals,
from which the masonry arch is underpinned, first in a column-like manner,
and then the underpinning is extended over the whole length (Fig. 6/27d).
(b) Should the ground not be firm enough to support the top arch the side walls
have to be installed first. This is done from headings driven on both sides of the
bottom section, followed by the driving of a separate top heading from which
the calotte is excavated and the top arch constructed, bearing upon the side walls
built previously (cf. German method). The core (strozze) will only then be removed
from the ready-made tunnel section (Fig. 6/27e).
(c) In a very loose ground, the invert aQd the lower part of the side walls have to
be made first to provide a safe bottom-support (cf. Italian method). A central
drift is started and driven at the bottom for a very short distance. Cross-drifts
in which the invert arch and the upper part of the side walls are constructed step
by step are excavated from it on both sides. Advancement at the bottom half is
effected by the repetition of this cycle. Step by step again, the driving of the top
heading then follows. From this the calotte is excavated subsequently on both
sides, and the top arch is erected, supported on the top of the previously built
side walls (Fig. 6/27f). This method constructs the tunnel in very small consecutive
steps.
Before giving a brief survey of classical tunnelling systems (mining methods)
developed from the main principles given above, the construction and driving
of the separate excavation and construction units employed, i.e. the headings,
the rise, dips and cuts will be first dealt with.

62.11. Scope and Construction of Headings

Headings are mainly used for working and haulage purposes (working or
haulage headings, respectively). This distinction is of significance also for the
period of use and, thus, also for the structural strength of the respective heading.
In addition, headings are also important in geological exploration, direction
control, and for housing various service lines and for drainage.
Tunnel construction ~s usually started by driving pilot h_eadings (entries) at
both tunnel adits. These pilot headings serve not only for setting-out the tunnel
and the exploration of soil and hydraulic conditions, but they constitute an integral
part of the tunnel construction from the very beginning. Muck is excavated through
them, and they are the starting points - with certain tunnelling methods - for the
gradual and successive enlargement of the excavation and lining of the final
tunnel section. The pilot beading reveals the geological profile through the
whole length of the tunnel, and also provides points of attack for the excavation
and subsequent lining at any point of the tunnel - even at several spots simul-
taneously. The haulage track, ventilation duct, electric, compressed-air and other
service lines are all installed in the pilot heading. Its driving also exerts a relieving
effect upon the surrounding rock; at the same time it provides for the collection
 616 CONSTR UCTION AND DESIGN OF TUNNEL S

and draina ge of ground water from the workin g places (adits)

. The pilot headin g
is, thus, an impor tant part of tunnel constr uction , and with
due regard to this,
the choice of its locatio n and dimensions is of decisive impor
tance, necessitating
the consid eration of a great many factors which depen d, largely
, on the tunnel ling
system chosen.
Pilot headin gs, in princip le, can be either bottom or top headin
gs. For opera-
tional purpos es it is more advant ageous to place the pilot
headin g in the lowest
part of the tunnel , i.e. to drive a bottom headin g. Thoug
h this choice is of no
conseq uence in the setting -out of the tunnel , the rail tracks
for the transp ort of
materials are laid to the best advan tage on the bottom
of the tunnel section ,
so that there. will be no need for their translo cation throug hout
the whole tunnel ling
operat ion to effect mucki ng, most of which can be loaded by
gravita tion. Simila rly,
the catchw ater drain (with a section of 30 x 30 cm, lined
with planki ng in loose
ground ) is best placed in practic e at the floor of the bottom
heading (Fig. 6/28).

Compressed air tube

f electric ' • - "
, conduit

FIG. 6/28. General arrange ment of bottom drift

A headin g also makes the install ation of new workin g sites

possible. Throu gh
side or top drifts work can also be started in the centra l tunnel
sections. In tunnel s
locate d on slopes or on mount ain hangs,. tunnel ling operat
ions can be started
in interm ediate tunnel section s by means of cross-cuts or drifts
started from out-
side. Shallo w city tunnel s repres ent a special case, as their
constr uction is usually
started from shafts.
In choosi ng the headin g profile, many points have to be taken
into accoun t.
It must be consid ered, which kind of tempo rary suppo rt can be
used best, since
the shape of the headin g profile depend s upon the structu
re emplo yed being tra-

.. . '·· ·.
 READINGS AND DRIFTS 617

pezoidal with timber sets, circular with steel liner-plates, horseshoe-shaped with
steel rails, flat bents and pointed arches with rein(arced-concrete frames. On the
other hand, economy and the uninterrupted fl.ow of material transport must be
taken into account, as well as good ventilation conditions which require cross-
sectional areas as large as possible, while the rate of advancement and the anti-
cipation of excessive rock pressures call for smaller cross-sectional areas. The most
suitable cross-sectional dimensions are always determined by due consideration
of the local conditions, of soil conditions, of tunnel length, and of the prescribed
construction period of the tunnel and of the tunnelling system employed.
62.111. Timbered headings. Timbered headings are usually trapezoidal as this
shape is suitable for resisting both vertical and lateral pressures. The clear height
of the headings when completed varies from2to2·5m(6ft8 in-8ft4 in),the width
at the roof from 2·0 to 2·2 m (6 ft 8 in- 7 ft 4 in), and the width at floor level from
2·2 to 2·5 m (7 ft 4 in-8 ft 4 in). The timbering consists of frames (bents) spaced at
1·0--1·5 m (3 ft-4 ft 5 in) distances with timber lagging supported on them.

Po/mg board

Working de!,a1!s Agamst Agamst

vertical /Jorizonta/
pressure pressure

~
'

ff::
FIG.
t LU
6/29.. Structural elements of timbered drifts with details
 618 CONSTRUCTION AND DESIGN OF TUNNELS

Timbering is usually made of pine wood, of even-grown and snag-free trunks.

The required seats and joints between the single elements of the bent are of simple
craftsmanship, such as can easily be hewn on the spot (Fig. 6/29). The timbers
are usually butt- or lap-jointed; tenons, etc. are not used. Rafter nails are used to
fix and keep timbers in place, while ordinary nail-work must be avoided; moreover,
even rafter nails do not substitute for an appropriate trimming of the joining
timber ends, nor must firm wedging of the lagging be omitted.
The most common timber bents are shown in Fig. 6/30 for cases where no rock
pressure is to be carried by them, their only task being to offer protection

Fm. 6/ 30. Part timbering

against rock-spalling. Lags are placed with interstices (skeleton lagging), the back
space between them and the surrounding rock being tightly packed with rock spall.
In fissured rock, with smaller potential rock pressures, the timber bent shown in
Fig. 6/31 consisting of a cap, two posts and a sill-beam is employed. A strut-
bracing is applied between neighbouring sets. With an adequate bridge-action
period, when practically no immediate rock pressure is to be resisted after the full-
face excavation, the bents are placed and the laggings simply set behind the caps

ij' Skeletoo
, lagging
~, fagging
• (placed svb-
t seqventlj;)
~

Fra. 6/ 31. Timber bents with skeleton lagging

 HEADINGS AND DRIFTS 619

and posts. In more intricate ground, for~?p0/i'7i

' ,
but still with moderate rock pres- .;ii. • • '

sure, forepoling has to be applied.

If the ground is stiff, the face of the
heading-drift is cut down straight
almost at the forward end of the Auxiliary bent
polings and new posts and a cap
are set. If the ground is soft a small
excavation, say 0·60 m (2 ft) high,
under one or two polings is made
and a short board is quickly set ver-
tical under the end of the polings
and acts as a combined breast-
board and prop for the poling. The
rear end of the poling is supported
on the cap or, temporarily, on an
auxiliary cap, and the face is braced
by breast boards (Fig. 6/32).
In loose and running ground,
however, not even a one man bur-
row can be advanced without the
danger of subsidence and collapse,
not even at depths of 60 cm (2 ft),
therefore advance forepoling must
always be employed (cf. Fig. 6/29).
Forepoles are driven into the ground
(advanced support) and the excava-
tion (the advancement of the face)
is made under their protection with
a cut-off length of the forepoles
sufficiently deep to provide for the
stabilization of the face. With a tight
.+
spacing of the bents, however, fore- FIG. 6/ 32. Drift
poles directed by the last cap will advancement with poling boards
have too steep an inclination up-
wards, as a result of which their front ends will lie far above the cap to
be newly set. Consequently the forepoles must either be ti!ted i:! wl:ich case
the ground material lying between the front ends of the forepoles and the cap
will spall off, so resulting in a loosening of the ground, or they must be
supported step by step from the next for.!pole, as shown in Fig. 6/33 (and cf.
Fig. 6/44).
Forepoles are supported on a bridge (sill) board, and are tightened from the
top of the cap by packing wedges (keys) (cf. Fig. 6/29). The next row of forepoles
can then be driven in the interstices left between the keys. These, in turn, will be
620 CONSTRUCTION AND DESIGN OF TUNNELS

"lilted forepoling w1tbout

end support
• ~. ~ . . . " .. . ~ t , Direction of tilt
'b : ,,;:,~I> :.... ~ ... . --:::.;;---"" 'b
~ ·' • :;::.:--,;.::-...-;.T.. .:y=--=-'
~ -- m
Fm. 6/33. Forepoling methods for densely spaced bents

tightened by smailer wedges from the bridge board. Then the keys are knocked
out and further forepoles driven at their place.
If progressively higher pressures are encountered the caps of the timber bents
are supported by purlins applied subsequently (Fig. 6/34a). The purlins are p ropped
up at distances varying between 2 m (6 ft 8 in) and 0·5- 1·0 m (I ft 8 in- 3 ft 4 in)
depending upon rock pressure conditions.

SLr,1ming wedges Packmg ht/,

Packmg wedge \

I 8,1d packing
~ rubble

Bottom sill
8) b)
Fm . 6/34. Reinforced and double ti mbering

Higher loads can be dealt with by the use of double or strutted timber bents
(Fig. 6/34b). In this case each timber bent is strengthened , at the time of its
setting instead of later. An important part of the double bent is a roof beam
placed longitudinally, by the use of which the span of the cap is reduced.
This longitudinal beam is supported by inclined struts from shorter side props.
 HEADINGS AND DRIFTS 621

ln case of lateral rock pressures bot-

tom purlins ·are also installed, braced to
each other by bottom struts (Fig. 6/34b).
All strengthenings of this type consid-
erably reduce the clearance of the m
heading, leaving, e.g, from the originally
trapezoidal section of 3 m,x 2 m (IO ft x 6
ft 3 in) a section of only about 2·2 m+.
1·4 m (7 ft 4 in x 4 ft 8 in) so impairing 28
haulage and working conditions in the FIG. 6/35. Numerical example
heading.
The dimensioning of timber bents may
be illustrated by the following example (Fig. 6/35):

The vertical rock pressure acting on headings can be taken, after DAVIDOV, to have a parabol-
ic form with a peak ordinate:

where a 1 = a + h tan (45° -tp/2)

fk, = tan ,J> (after PROTODYAKONOY) with granular soils or up to 1 % of the cub
strength in solid rocks (cf. Section 32.324).
If the width of the heading is 2a = 1 ·8 m, its height h = 2·2 m and the thickness of the
overburden H = 12 m with y = l ·6 t/ m3 , ,J> = 35° and tan ,J> = 0·7, then

J ·8
a1 = + 2·2 tan (45° - ,J>/2) = 2·09 m,
2
2·09
'1 1 = _ = 3·0 m .
0 70

The height of the ground column loading (h') at the edge of the heading is

( a•) = 3·0 (1 - -2·0

h' = '1 1 1 - --.
a-
I ,
o-9•j
- ' = 2·445 m,
-
2

and the vertical pressure acting at a distance x from it is

p = 1·6 X 3·0 (1 - -x•- ) = 4 ·8 (1 - 0·228x2)

2·09 2 , '

assuming the average rock pressure value as acting at a distance of x = 0·3 m

p = 4·8 (1 - 0·228 X 0·3 2 ) = 4·7 t/ m 2 •

If the spacing of the timber bents is taken as 0·8 m, the moment acting on the cap will be:

1·02
0·8 x 4·7 - - = I ·52 tm = 152 000 kgcm.
3
 622 CONSTRUCTION AND DESIGN OF TUNNELS

Ca'icuiating with a permissible stress of o = 100 kg/cm 2 for tim~r, th~!}

M
W= - ,
0
i.e.
152 000
32 100
whence
d = 23·4 cm~ 25 cm.
If, e.g., no round (log) timber of such a dia is available, the problem can be solved by
installing longitudinal roof girders under the caps (double set), in which c~se its ~pan wjll
be reduced to half, and the acting moment to a quarter of its former value. Hence, e.&,, a
cap dia of

would be sufficient.
The compressive force acting upon a prop will be

1·8 X 0·8
N = 4·7 - - - = 3·4 t
2

The bending stress from lateral loading is calculated for a lateral pressure of

h2 = h tan (45° - cj,/ 2) = 2·20 x 0·52 = 1·12 m,

P2 = h 2 y tan2 (45° - cj,/2)

= 1·12 X 1·6 X 0·27 = 0-484 t/ m~,

from which the bending moment acting upon the props spaced at 0·8 m •.•:;li be:

0·484 X 0·8 (2·2 - 0·3)2

M = 0·128 = 0·09 tm = 9000 k gcm .
2
Thus

where ok = the permissible bending stress reduced according to buckling len gth.
If, e.g., ok = 50 kg/cm', the solution of this mixed equation of third degree will be

d ~ 15 cm,

which dimen~ion must perhaps be rounded up or off according to available supply.

62.112. He,pdings with steel supports. The most simple steel supports used in
headings are tp.e T.H. and Moll ribs mentioned above (cf. Section 61.311), and
wide!y employed jn Hungarian mines in haulage roads and ventilation drifts.
Wooden planks or, most recently, reinforced-concrete boards and sometimes also
sheet-steei piles are used as lags, applied in the same way as in timber headings.
 HEADINGS AND DRIFTS 623

However, stronger lags have to be used here because of the wider spacing of the
steel ribs. The same structural elements can be applied also, of course, in tunnel
driving.
For more permanent headings, ribs of horseshoe-shaped rails with timber skel-
eton lagging are frequently used (Fig. 6/36). It must be noted, however, that the
legs of these ribs may not be placed directly on the floor because of their compar-
atively small contact surface, but concrete blocks, reinforced-concrete plates or
iron-covered hard-wood blocks must be inserted.

· Bent scr,,p r,,,f

Concrete h7vert

Ke,nforced-co,i;ete
sill bOJrd
FIG . 6/ 36. Drift bents made of scrap rail

Naturally, not only bent rails but any other structural steel section can be used,
but, in general, heavily worn steel rails have proved to be the most economic.
Likewise, corrugated or ribbed steel plates or prestressed reinforced-concrete
planks can be used instead of timber lagging.
However, some disadvantages are also inherent in such steel supporting struc-
tures inasmuch as they are suitable for headings of constant cross-sectional dimen-
sions only, and they do not give prior warning of the increase of rock pressures
as timber supports do by squeaking and cracking. On the other hand, they have
the great advantages of simple installation and transfer, and suitability for repeated
re-use.
The foregoing steel structures are, strictly speaking, mixed structures as their
lagging consists mostly of timber or concrete planks.
Liner-plate headings may be regarded as purely steel structures when built
with mining methods. Such headings were used, e.g. on some stretches of the
subway construction in Budapest. They make a continuous steel skin consisting
of pressed steel liner-plates bolted together through their rims. These headings
were driven with the full face, and the rim steel s~gments were pressed from 6-8 mm
(1/4-3/8 in) steel plates and had a surface area of 0·52 m x 0·80 m (2 ft x 3 ft)
and were bolted into rings, the consecutive rings forming a coherent flexible skin.
 - - - - ------- -- -~
. -

624 CONSTRUCTION AND DESIGN OF TUNNELS

As shown in Fig. 6/37, these lining segments have radial joints, and therefore the
placing of the key segment (Z) from inside the ring was made possible only by
providing it with joints of opposite inclination and by providing for two segments
{C) with asymmetric joints to be coupled to the normal segments (N). Because of
the flexibility of this structure the prompt and tight dry packing of the back spaces
behind the lining is of supreme importance before excessive rock pressures can
develop and so also ensuring the supporting effect, i.e. the composite action of
the surrounding ground.

C'ross-sect/011 Longitudinal section

6mm.

Bo/fed c011nect1011

gJomm -j ISpeaa/ muck-car

r~7
8001

Fro. 6/ 37. Liner-plate drift

The headings mentioned were driven in loose sandy, silty soils where the cir-
cular cross-section is also statically the most advantageous. The rimmed segments
were erected according to the top-heading method described in Section 62.41
(cf. Fig. 6/63) or the heading and bench method described in Section 6.12, i.e. not
according to the full-face method. In this case the instaHation of liner plates is
started from the roof proceeding downwards and the key plate is inserted at the
bottom finally. Headings of this type have been used also auxiliary to shield tun-
nelling, partly to stabilize the face (cf. face anchorage in Figs 6/97 and 6/98),
and partly to provide a proper guidance for the big (8· 5 m, 28 ft dia) station shields
(cf. Section 63.2).
 HEADINGS AND DRIFI'S 625

62.113. Precast reinforced-concrete supports. The use of reinforced-concrete

sets combined with laggings either of reinforced-concrete planks or wooden
boards comes repeatedly into the foreground to economize in timber and steel
materials. The considerable advantages of these sets are their long service-life
and the relatively high strength obtained with comparatively little reinforcement.
Their disadvantages include their great weight, rigidity, difficulty in handling and
transport, and lack of workability. Owing to the rigidity of concrete accurate
fitting, wedging and bracing of these elements also present certain difficulties.
In Fig. 6/38 the pointed arch support system of reinforced-concrete elements
suggested for use in the construction o( the underground railways at Budapest
is shown as an example, indicating that pointed arch structures being very suitable
in shape to resist rock pressure can be made of reinforced concrete.
In Hungarian mining industry precast reinforced-concrete supports com-
posed of elements similar to those of timber sets are now being introduced to

/
/
/

59

Fm. 6/38. Pointed three-hinged reinforced-concrete bent

40 ~zecby: The Art of Tunnelling

 15 Section A- A \
2 x ¢ 6 ~ 1 5 7,8 8 8 21
I
• 2J
f80
23 21 8 8 on a,.
t,.>

4t1:1 l;®i 1
~)!1
a,.

0
~
5
2x¢l4

Jection C-C
i
;;J
C
g
0
z
~ z>
tl
--¥-/,IV7'I-- ~
m
C)
z
.,,0
>-i
~
z
A/tem3t/ve tubul:;r post E

-,!~
-~ , Section 8-8 f1G. 6/39. Trapezoidal reinforced-concrete ben t
-iP
 HEADINGS AND DRIITS 627

reduce timber consumption (Fig. 6/39), chiefly in semi-permanent drifts. The

support consists of a cap and two inclined posts placed on a reinforced-concrete
sill. The support seats of the respective elements are curved. Experiments have
also been made with tubular hollow spun-concrete posts in order to reduce weight.
Lagging consists either of wood boards (cf. Fig. 6/38) or of reinforced-concrete
planks 5- 6 cm (2-2 1/4 in) in thickness, l l-12 cm (4 l/4-4 3/4 in) in breadth, and
90-120 cm (3- 4 ft) in length. The drawback of reinforced-concrete planks
is, again, their weight and the difficulty of handling; their main advantage being
their long service-life and strength and the economy obtained in the use of timber.
A more widespread use of reinforced-concrete heading frames may be expected
probably only when the mechanization of their erection is adequately solved.
The weight of these supports has been considerably reduced recently by making
the elements of prestressed concrete posts, the use of which in the uniform haulage
drifts of mines is now being introduced.
62.114. Headings of masonry lining. In many cases it can be both economic
and practical to construct circular beadings of brick or concrete block masonry.
This will be reasonable either because of a demand for a long service-life when
maintenance costs will be at a minimum if permanent supports are used, or in
weak soils where heavy rock pressures acting from all directions have to be resisted
even when a heading of relatively small cross-section is driven so that a closed
ring is essential.
Up to now the following two kinds of masonry lining have been employed in
practice:

Mucking cars

o~z-o- 4-80

FIG. 6/40. Drift, lined with precast concrete blocks

40*
 -~~
~

- - - -

CONSTRUCTION AND DESIGN OF TUNNELS

(a) Unyielding masonry linings are built as closed rings of bricks or concrete
blocks with a circular, frog-mouth, or elliptical horseshoe shape. In Hungarian
mining practice concrete-block linings are used at an ever-increasing rate for both
permanent haulage passages and air ducts and for drifts exposed to high rock
pressures (Fig. 6/40). Though common bricks can be used instead of concrete
blocks, they are seldom employed because of their lower strength (the crushing
strength of brick masonry) which is only 30-35 % that of the brick owing to the
low strength of the mortar, and because of the long construction time required.
While concrete blocks can be used against external pressures up to 5-15 kg/cm2,
brick masonry is employed only for pressures not exceeding 3-4 kg/cm2 • Owing
to the deformation characteristics of concrete such masonry rings can be regarded
as unyielding supports capable of minor deformations. In exceptional cases
pine-wood planks or blocks may be inserted between certain block layers, most
particularly at the abutment, to provide a certain deformation capacity.
(b) Flexible lining-supports are used in headings of short service-life yet exposed
to very high rock pressures. Such a support is constituted first of all by the afore-
mentioned circular rim steel liner-plates with bolted joints (cf. Fig. 6/37). Here
the flexibility or deformation capacity of the steel ring required, e.g. when exposed
to extensive rock pressures, is obtained not only by its small thickness but also
by the articulation of the bolted joints.

F10. 6/ 41. Flexible bottom heading lined with wooden blocks (Great Apcnnine tunnel)
 VERTICAL BREAK•UPS 629

Another kind of flexible circular support is the lining ring made of wooden
blocks (Fig. 6/41). It was used in the construction of the Great Apennine tunnel
in a section of its bottom heading exposed to extraordinarily high genuine moun-
tain pressure when all attempts to maintain the cross-sectional area required for
the work by the usual timber supports had failed. The inner diameter of the head-
ing lined with this &lock ring was 3·25 m (10 ft l 0 in) with a ring thickness of
0·5 m (l ft 8 in). This lining was capable of resisting pressures even in sections
where crown bars a nd props of a dia as large as 0·6. m (2 ft) were crushed under
the rock load. The deformation capacity of wooden rings (due to the high compres-
sibility and to the low modulus of elasticity of wood, especially when in a wet
condition and perpendicular to its fibres) made the deformations produced by the
genuine mount_ain pressures possible without failure of the support.

62.12. Vertical Break-ups and Pits

Interconnections between bottom and top headings or bottom and side headings
are made by vertical break-ups (chimneys, rises), serving for the lowering of muck
to, or the lifting of the varied construction or supporting materials from the bottom
beading. For lifting up materials, simple winches are often used; a communication
ladder must be provided· in every break-up. In certain cases inclined break-ups
may be used.
Break-ups are quadrangular or circuiar in cross-section. In firm rocks, they
stand unsupported or with a few timber frames at most as a protection against
rock spalling. In break-ups with a quadrangular cross-section, timber frames
with timber laggings are ·installed, while in those with a circular cross-section,
bolted steel-plate rings are used. In loose ground, the installation: of complete
rings is likely to fail, and even timber frames have to be placed very tightly or,
if possible, forepoling has to be employed. This requires very delicate and difficult
work in vertical break-ups (Fig. 6/42).
The smallest cross-section used in break-ups is 0·7 m x 0·7 m (2 ft 4 in x 2 ft 4 in),
whj}e those generally used for communication purposes are 1· 3 m x 1·2 m
(4 ft 4 inx4 ft) and 1·3 mx0·5 m (4 ft 4 inx 1 ft 8 in) for box-holes or chutes.
The cross-sectional dimensions of a brea!<.-up must. in no case· exceed the roof
width of the respective bottom heading.
In classical tunnelling ·methods, pits or cuts are also frequently in use. Their
design and supports are similar to those applied in working pits, the strength of
the rock being always the predominant factor in the decision whether tight or
skeleton laggings or forepoles should be used. Cuts are usually deepened in sub-
sequent steps and walled up before the deepening of the next step of the cut is
started. This method is generally needed in the construction of side walls (Figs 6/ 158
and 6/ 159).
 630 CONSTRUCTION AND DESIGN OF TUNNELS

6.22. SINGLE-STAGE MINING METHODS OF TUNNEL CONSTRUCTION

Headings, break-ups and cuts are employed in every mining method of tunnel-
ling, though in various combinations. The full section is usually divided into the
following parts (cf. Fig. 6/26): bottom heading; top heading; arch segment ca-
vity or calotte; core or strozze; side wall space and invert arch. In the mining
methods to be introduced here, the whole tunnel section is excavated by sections
before the construction of wall masonry is started. Following_this, lining is cons-
tructed in a single, continuou_s course under the protection of temporary supports.
Consequently, adequate supports or support systems are required suitable for
carrying the whole excavated tunnel section either as a whole or divided into two
parts during the construction of the lining masonry, offering adequate protection
for all work to be carried out in the section.
An important requirement for these temporary supports is that they should be
of rigid, unyielding and stationary construction, even when exposed to asymme-
trical lateral loading. They must not suffer extensive compression or deflect.i on as
this would increase rock pressure, and the required clearance dimensions could
not be maintained. The respective supporting elements have to be tightened to
each other by wedges allowing repeated re-tightening of any loosened elements.
' The possibilities of punching and settlements involving deformations of this kind
were given in Fig. 6/24 applied to the Austrian method.
Single-course walls can be constructed by any of the following methods:
(a) the crown bar (English) method (c) alternate ring method
(b) the cross bar (Austrian) method (d) the central cut method.

62.21. The Crown Bar (English) Method

In the crown bar (English) method sections 3-6 m (10-20 ft) long are excavated
in full face from top to bottom, and then the entire wall masonry is constructed
in one course from bottom to top. Work commences with the driving of a bottom
(haulage) drift from which at a distance of every 50-60 m (170-220 ft) break-ups
(chimneys) are made to allow the driving of the top (working) heading. The bottom
drift furnishes, in addition, good temporary drainage facilities and a good control
of the alignment and it also allows the tunnel to be attacked at several points.
The top heading has relatively small dimensions [l ·5 m (5 ft) in width and 1·8 m
(6 ft) in height] and is placed at an elevation such that its roof is level with the
crown of the masonry arch extrados. The main attack on the face always begins
here in rings of the previously mentioned width of 3-6 m (10-20 ft) depending
upon the nature of the ground: the looser the ground the shorter the rings will
have to be. The forward ends of the roof (crown) bars rest on posts whereas their
rear ends may be supported by the extrados of the completed work.
Enlargement is started by underpinning the caps of the top heading by two
crown bars of 30-40 cm (12-16 in) dia (Fig. 6/43a). This, in turn, renders the sue-
 S OLE STAGE MINING METHODS 631

'",'-, - - T,...,,
'-
._ I
- -i
 632 CONSTRUCTION AND DESIGN OF TUNNELS

cessive removal of intermediate props possible. Following this, the top heading
will be enlarged across to the full width of the top section. This operation is usually
effected by transversely driven forepoles supported by roof bars and temporary
posts installed subsequently. As these radially placed posts are, in general, too
close to one another, shorter forepoles have _to be used to minimize overbreak
or they have to be tilted (Fig. 6/44a). If loosening of the ground caused by this
procedure is to be avoided, in all cases secondary strutting supported from crown
bars must be installed (Fig. 6/44b). The
heading is widened out for the length of
the roof (crown) bars. At the same time
the face is securely breasted and back-
strutted to the completed lining. The
crown bars, posts and struts are all
stiffened by spacers and bracers.
During widening, and simultaneously
with the install.a tion of additional crown
/
TIited lbrepo/111g bars and posts, the bottom of tlie top
beading must also be deepened and wid-
ened, and posts and sills must, conse-
quently, be continually replaced by long-
er ones. When the bottom of the en-
largement is deepened to the level of the
b) springings, a new transverse sill is laid
spanning the whole section width from
which all the crown bars will be sup-
forepo/ing w1tll St'COll- ported by newly installed inclined posts
dary support (Fig. 6/43c). Then, sills are supported by
FIG. 6/44. Forepoling posts extending down to the bottom of
methods with densely spaced the section and the operation is contin-
props ued down the sides as far as required by
the nature and pressure of the ground,
frequently down to the bottom, the face
being securely breasted and backstrutted
(Fig. 6/45) and the section enlarged to full dimensions (Fig. 6/43d). During
the whole widening operation special care must be taken, especially in loose
ground, that no unpacked back space should be left behind the lagging (cf.
Fig. 6/44) as such cavities may cause later ground movements and an increase
in rock pressure. The back spaces must therefore be tightly dry-packed with rock
spall or - at least temporarily - with timber blocks.
When the entire· tunnel section is excavated, miners give way to masons who
construct the length of lining within the completed timbering. Construction is
started with the invert arch, if any, or where this may be omitted are built up first
starting with the installation of formwork and centering. Concrete is poured or
courses of masonry blocks are placed uninterruptedly until both side walls meet
 SINGLE STAGE MINING MFTHODS 633

-- '

Fm. 6/45. One fully-excavated section of the English method with raking support of the face

in the crown of the arch (Fig. 6/43e). During the construction of walls laggings
are successively removed and crown bars transferred to the next tunnel ring, their
place being tightly filled with concrete together with all cavities packed o nly tem-
porarily at the time of the excavation.
As mentioned, the construction of the adjoining masonry rings can be so
arranged that the rear ends of the crown bars will bea r upon the extrados of the
completed masonry ring (cf. Fig. 6/45), their forward ends resting on the full
support system at the newly opened face (Fig. 6/46). The main supporting sets
must be efficiently stiffened and bound to each other in all directions (by St.
Andrew's cross bracing between the ve rtical supporting elements).
 634 CONSTRU CTION AND DESIGN OF TUNNELS

F1G . 6/46. Isometric view of the English method

The English method of timberin g, though it ensures very good working

con-
ditions and a high rate of advance , is more suitable for rather firm ground,
as in
weak ground the crown bars would not be able to resist higher rock pressure
s,
so that either the spacing of the crown bars and posts ought to be reduced
or the
length of the sections decrease d. Both these procedu res would increase
timber
consum ption. Though , in principle, the crown bars, laggings, posts and
other
timberin g element s could be repeated ly used after being pulled down, the
removal
of crown bars and lagging is impossible when exposed to unduly high pressure
,
and they must be left in place and incorpo rated with the masonry . The dismant
-
ling of timberin g element s and transfer of the load to complet ed rings
must be
done with the greatest care to prevent further rock movements.
Advanta ges of the English tunnelli ng method are that excavat ion, timberin
g,
mucking and erection of masonry can be underta ken with ease and without
any
 SINGLE STAGE MINING METHODS 635

mutual interference in time or space, and both the ventilation of the working
faces and their drainage are properly provided for. Its disadvantages are that the
cross-section to be broken out is far bigger than that required for the masonry
lining because of the room needed for the crown bars and the overbreak behind
the masonry lining is usually only loosely packed with rock spalls. The crown
bars and wood laggings occasionally buried and walled-in may decay, resulting
later in rock settlements and pressure increases. Front faces are ~iways exposed
to high loads and their temporary support by raker posts is rather elaborate and
complicated, necessitating altogether a very high timber demand. Finally, the
eventual subsequent strengthening of the timbering is very complicated. Because
of all these drawbacks, the English tunnelling method is nowadays employed
only rarely.

62.22. The Cross-bar ( Austrian) Method

Unlike the crown bar arrangement of the English method, this method uses
segmental arch sets. Their structural arrangement is shown in Fig. 6/47a and the
sequence of the following steps of its construction are shown in Fig. 6/47b. Work
is started here also by driving a bottom drift from which break-ups are made
every 50--100 m from which the top heading is driven. From this top heading
the top section of the tunnel (the calotte) will be broken out much in the same
way as in the English or the Belgian method, with the exception of the placing
of the poling boards. The poling boards over the segmental arch timbers are all
driven parallel and not perpendicular to the axis of the tunnel and the widening
is done in short lengths, working practically step by step to the full face (cf. Fig.
6/47a). Owing to this system this method is better suited to forepoling than is the
English one, where considerable overbreaks are likely to occur due to transverse
poling (cf. Fig. 6/44). Segmental arch timber sets are spaced at 1·20--2·0 m (4-7 ft)
according to the strength characteristics of the ground. Since segmental arches
can be spaced even more tightly if necessary in looser grounds, this method can
be better adjusted to actual ground conditions than the previous ones. Since,
under higher rock pressures, it is safer to widen out the top heading in a transverse
direction, this may be ·effected by placing densely spaced, short horizontal struts
between the segmental timbering which , in turn, will support the cross-driven
poling boards when the calotte is being excavated. (In firmer ground, poling
boards can be omitted, as segmental timbering and these struts will afford suf-
ficient support.)
After excavation of the top segment, arch timbering is supported on main sill
beams installed in the horizontal diameter at the springings which are supported
by posts placed in cuts extending down to the invert as excavation is not inter-
rupted , but continued downward in full, just as in the English method. With an
appropriately arranged bottom heading, vertical posts do not need to be sup-
ported on the floor but on the caps of the bottom heading. In addition to the
 636 CONSTRU CTION AND DESIGN OF TUNNELS

BJ

Axial /;;gging

Arch construcLton
Topdnft
/ b) Long,tudmal section
l.ateral enlargement

_ Construction o/'abutmenf 11.fJlls

- 8reak down to Lile bot/om drift
"-- ConstructiOfl ofrnvert arch ( XI)
Jnst;;ll;;L!on of drain-off canal (XII)
' Excav3t1on of mvert arc!,
FIG. 6/47. Excavati on for invert arch; structura l arrangem ent and sequence
of operation s for
the Austrian method

sill bea m placed at the springin g, a second sill beam is also often placed
at the
bottom of the section (Fig. 6/48).
After full outbrea k the invert arch is construc ted, the formwo rk erected,
shut-
tering placed and the masonry lining built in a single course, again working
up-
wards from the bottom. Since segment al arch timberin g sets are spaced
too
close to each other, either a suitably short length for the masonry rings
has to
be chosen, or else the timberin g must be removed which will be very cumbers
ome.
With the Austrian method, all the operatio ns menfam ed above are carried
out
-:;:-

::--
>

-;
~
I
<:
t::::
~
~ ·, C

G:5

,...._ ______, _.)

~'\ l'
\ \: I
'
',...... , . . ________ ' J
'-, -}- I
, . . . . . . _______ t-

.,,,, ...-
-------4-, . I

/
_,,/ / ---------+--,
I I
~ / // i \l _
..,- I I I ' -. .

ttl--
'-
0

"'
"-'
"'
"'
-g_
-;;
C:
.2
u::,
~ \ ;;;

\ ' ',.. _ _ ------+~,

\ \ ·1 /

'- I C:
0
u
..)
',-
------ _--l-
0
C
 SINGLE STAGE MIN ING METHODS 637

in steps and staggered in space but in a continuous sequence, thus allowing a con-
tinuous advance and the possibility of a n easy survey. (This continuity of work,
naturally, will involve difficulties as to speed of driving by the establishment of
several simultaneous working faces. as the bottom heading will not be able to
provide adequately for simultaneous mucking fro}Il several working places.)
An inherent disadvantage of the Austrian metfiod lies exactly in the fact that
the most delicate part of the tunnel section, i.e. the'l'oof, is supported by temporary
-supports only, for a considerable time. In addition, this support is subject to
repeated dismantling, each occasion initiating progressive roof subsidences (cf.
Figs 6/24 a nd 6/48) which, in turn, result in increasing rock pressures. In large
tunnels, however, the timber dimensions and spacing of sets required to resist
increased ground pressures must be so big and, respectively, so tight, that in tun-
nels with great spans (e.g. motorway tunnels) the method can no longer be applied.
On account of these facts, the Austrian method has recently been combined
in some cases with the English one, in that the breaking out of the full section is
uninterrupted and staggered in space but in rings excavated from break-ups spaced
more closely, maintaining at the same time the crown bar support system and
constructing the lining masonry upwards from the bottom in one course. The
invert arch is installed also only after the construction of the side walls a nd top
arch. The excavation of the adjacent ring is, as a rule, not to be commenced
before the permanent lining of the previous ring is completed. The length of the
rings can·vary between 1·5 m and 12 m (5 ft and 40ft), depending upon the nature
of the ground. Several rings may be served by one common top heading. The rings
can also be excavated sta_ggered in space, i.e. leaving temporary ground columns
between the respective r~ngs. In this case the subsequent breaking out and the
lining of the intermediate rings is safer and quicker under the protection of adja-
cent completed linings. In any case, when walling the arch, this site must always
be kept in a distance of at least 12 m - 15 m-(40-50 ft) from the excavation face
to protect the masonry from damage due to blasting.
The advantage of the Austrian method is that it affords a safe method in loose
gro und also, although it is very uneconomic in timber and labour demands and,
thus, is very expensive, ~s shown in Table 6/VH.6 ·1 ~

TABLE 6/ YII

Assistant Shift T o tal

Method Pitmen H aulage men ~tfasons Carpenters
pitmen worker s working hr

Belgian 294 185 361 152 293 20 1304

---
Austrian 297 353 665 126 61 54 15:34
I
6·13 MIKLOS, P.: Alagutepitesi rendszerek es m6dszerek. Vtilogatolf fejezetek a foldalalli
vasutepites, btinydszat, me/yepites korebol (Tunnelli ng systems and methods. Selected chapters
from tunnelling constructio ns and mining engineering). K ozl. Kiad6, Budapest 1953 2 376
638 CONSTRUCTION AND DESIGN OF TUNNELS

62.23. The Alternate Ring Method

The new (modified) Austrian method also may be listed among the alternate
ring methods together with full-face tunnelling methods using temporary steel
supports and discussed in Section 6.1. No headings are actually driven here, the
full tunnel section being excavated in l ·2 m wide stretches with the masonry liriing
built in immediately. Then, the second next ring is excavated and walled up,
leaving a ring of ground of the same length in between. Thus, by driving no
advance heading, the preliminary loosening of the ground and the mobilization
of ground pressures are prevented. Tunnel sections exceeding 3·5 m in dia are
not usually excavated in a single operation but by steps in benches, according to
the principles of the Belgian (underpinning) or the German (core leaving) methods
to be discussed later.
In the alternate ring method steel elements are most usually used for tempo-
rary supports (cf. Sections 62.44 and 62.45) so economizing in timber consump-
tion. Other advantages to be ascribed to this method are that the full tunnel section
is opened up only in short stretches and is supported by temporary supports only
for short periods, its permanent masonry lining being erected after a short interval,
resulting in a relatively small disturbance of the original state of equilibrium of
the ground. On account of this, ground loosening, rock spalling and overbreaks
will be reduced to a minimum. The stability of the tunnel section against axial
displacements will be considerable, and the system will greatly favour the repeated
use of movable steel supports. A disadvantage is that the· rather inaccurate junc-
tion of the masonry of intermediate rings with those previously built sometimes
necessitates the application of some more expensive construction especially when
erecting the arches. Should, however, the propping of masonry rings be done in
steps, roof subsidences will also have to be taken into account owing to the
repeated dismantling of temporary supports. On the other band, the ground left
between the completed rings will obtain support from these as a result of arching,
and this reduces the roof-load.
The alternate ring methods cannot be classed strictly among classic tunnelling
methods and, therefore, their detailed description will be given later in Sec-
tion 6.24.

62.24. The Centre-cut Method

In rocks of adequate strength, certain advantages are offered by driving a centre

cut, proceeding upwards to the full height of the tunnel from the bottom beading,
followed by the successive underhand stoping of the sides (cutting method).
After widening out to full width, the masonry walls are built from the bottom
to the top in a single course (cf. Fig. 6/27c). The section can be supported either
by segmental arches or by the crown bar method, depending on roof pressures.
A great advantage is that the roof is immediately supported from the bottom.
 MULTIPLE STAGE MINING METHODS 639

rendering the repeated removal and replacement of various timbering elements

unnecessary, thus avoiding the main source of roof subsidence. A further advantage
of the method is the %tablishment of a clear and continuous site which is easily
ventilated. In addition, timber can easily be recovered and re-used.
A higher rate o'f progress can also be attained since multiple face tunnelling
and the placing of the tunnel lining in alternate rings from separate break-ups
is also made possible. Under heavier pressures it is reasonable to reduce the length
of these stretches in order to allow more prompt installation of permanent sup-
ports. The method is similar to the alternate ring method discussed in the previous
section.

6.23. MULTIPLE-STAGE CLASSICAL METHODS

'
The methods grouped among the classical tunnel construction methods are as
follows:
(a) the underpinning (flying arch) or Belgian method
(b) the core leaving or German method
(c) the invert arch or Italian method
(d) the combined construction methods developed as combinations of the
above methods by selecting and combining some of their characteristic
elements.
These methods are, in general, applied in soft ground, the overriding factor in
their choke being the stability of the gro und, as follows:
(a) The Belgian method is used when the ground in the height of the arch
abutment is stable enough to carry temporarily the full weight of the arch loaded
by the roof pressure without undesirable subsidence.
(b) In grounds of lower load-bearing capacity, but still stable enough to carry
tem porarily at floor level the weight of both the side walls and the arch, weighed
by the roof pressure transferred by a wider footing of the same without harmful
settlements, the German method is employed.
(c) In softer ground where the full tunnel width is required for tbe safe transfer
of the loads, the Italiao method is to be employed.

These methods will be discussed below in the order given above.

62.3 I. The Belgian or Underpinning ( Flying Arch) Method

This method is similar to the English one but more economical and can be
applied to relatively firm ground only. The main difference between the two
methods is that in the English method the entire cross-section of the tunnel is
fully excavated before commencing the construction of the walls, whereas in the
 ~

!""' 1-,.., Ih-i !:....

~,·, ·mnn mnrn11·t
11ti11fuuu, oudi1 1i ·•1n111nuntfll rrir1,1l1io,cm,Ll
~llfflfmtff~M1ntffir1111Jn1 ·'F.fi: ___
~ ,,./__,,.·
I<,

!4,
8z
,.~
i<:ri ,.,_, c::
$ection6-6 Section F-F SecttonE-E Section 0-D SectlonC-C Section 8-8 Section A-A
~
~
1,/~ .

•
. >
/ ' z
II I '\ 0
)~I- If ,.;lf.~"'\ I
I I
I \ \
\ \ 0
I ~
_,, ' .' . I II ·
I·
I I
I I
II I\ c\
4-60 , , :~/ I I . I I
I· I I. / I II l
I I
I· /I II z
I
I I
I I
-t..l
\ I
I
I

I
- -r -
i/
-b,
I I
I I
I I
\ I
1 1
-t:,- - + - !d-
I I
I

I -- d \ I
t:t __
I
1 I
I -r~J_
_,_. ,J I ..
0

--1
C:
z
z
m
r-
FIG. 6/49. Scheme of the Belgian (flying arch) method as started from a top drift "'
 MULTIPLE STAGE MINING METHODS 641

Belgian method excavation and erection of walls proceed alternately in stages.

The method is widely used because of its comparatively great economy. Two
main types may be differentiated: i.e. single heading and double heading methods.
To place the pilot ~eading at tbe top section is reasonable in dry rock and for
short tunnels only. Generally, a bottom heading is driven (although it weakens
the supporting core).
The Belgian method was based originally on starting with a single heading,
and comprised the following features (Fig. 6/49):
A centre crest heading is driven _along the tunnel track from which the lateral
excavation of the upper half-section (the arch or the calotte) is done much the
same as in the English method (cf. Fig. 6/43), i.e. caps of the timber sets are under-
pinned by longitudinal roof bars an,d t,he heading successively widened out fan-
like to the full width of the top section with transverse poling. Each poling set
may be supported by separate roof bars underpinned by props from sills placed
at varying levels keeping pace with the gradual deepening of the calotte, until
reaching down to springing level, i.e. to the abutment provided by the masonry
arch. The transverse polings are frequently not driven over the roof bars in firm
ground but are placed instead against the earth roof after the excavation has been
made for the next side-roof bar. Then excavation is stopped, temporary props
and·the construction of the arch masonry .started and the walls erected in a sec-
tion of 4-9 m length. The roof bars are underpinned at both ends by radial
posts at the same time arid formwork and shuttering are placed for the masonry
arch which is built up from both sides. In ground oflower,or insufficiently uniform
strength a reinforced-concrete wall beam or timber wall-plate is laid first on which
to build the masonry arch of concrete blocks or ashlar stones (cf. Fig. 6/53).
Tlre construction of the masonry arch, the pulling forward of crown bars and
the backpacking are done just as in the English method. When, during walling,
a temporary prop or post is in the way, it will be removed and the roof will be
supported meanwhile by temporary laggings and shores. The place of crown bars
is left out in the same way until they are slid ahead into the. next stretch. Should
the ground loosen in the
course of construction to such
an extent that a breakdown a} b}
of the roof seems probable,
all the timbers, crown bars,
posts, props and polings are
left and incorporated into the
masonry.
If work is started by driving 6 4
a bottom heading, the multi-
stage excavation of the full
tunnel section and its masonry FIG. 6/50. Sequence of part excavations:
lining can be done simultane- (a) single-drift method;
ously from several break-ups (b) double-drift method

41 Szecby : The Art of Tunnelling

642 CONSTRUCTION AND DESIGN OF TU NNELS

in rings of definite length. This method has the advantage that mucking is
separated from construction and excavation work (which, at the same time,
is disadvantageous as regards ventilation) and the rate of advance is also increased.
The difference in the sequence of excavation work compared with the top heading
method is shown in Fig. 6/50.
After the completion and deshuttering of the masonry arch, there follows the
underhand stopping and the construction of the side walls under the protection
of the erected load-bearing masonry arch. This operation, again, can be done in
several ways:
l. Pits of 2-3 m (7-10 ft) width are sunk or deepened at intervals from the earth
core surface, from which the abutments of the masonry arch will be underpinned
by masonry columns on the most important arch sections. The sequence of opera-
tions of this method is shown in Fig. 6/51a, where the construction of the support-
ing masonry columns follows a chessboard pattern. The greatest care must be
taken not to excavate the opposite sides of the same section simultaneously.
In very loose ground, the sections between the masonry columns are also under-
pinned from such pits, the removal of the core of ground following afterwards.
In sufficiently solid ground (such as marl or solid clay) the work can be consider-
ably speed~d up by the simultaneous removal of the ground at the sides and in
the core between the adjacent masonry columns to be followed by the building
of the remaining side wall sections in a single operation.
2. For material transport and working conditions it is more advantageous to
excavate a central trench to the invert leaving a berm of l ·5-2 m on either side
of the tunnel to support the arch of the tunnel lining. From this central trench
narrow cuts to the sides are made at intervals following a chessboard pattern
from which the masonry arch is underpinned by columns. The haulage tracks
are laid at the bottom of the trench, which makes loading and removal of the
muck easier (Fig. 6/51 b). This method can be applied only in firm to medium firm
ground.
Underpinning is a very delicate job and must be done in alternate sections
with at least 2-3 m (7-10 ft) wide temporary ground pillars left between them.
Special care must be taken in the construction of the sections adjoining the abut-
ment of the masonry arch because of the difficulties of accurate compaction work .
Because of this, these sections are usually solidified by subsequent cement grouting.
Side wall construction is followed by the removal of the core and the installation
of the invert arch, if necessary.
3. Finally, the side walls can be constructed in an alternating continuous oper-
ation, following .at a distance of 3-4 m (10-13 ft) the uninterrupted advancement
of the face (Fig. 6/5 lc). In this case, adequate temporary support of the arch
abutments is most important for which wall beams and shores are used. In addi-
tion, strut beams are placed in the spring line of the top arch to prevent the dis-
placement of the side walls during its construction and that of the invert arch
(Fig. 6/52). It is unreasonable not to attack the opposite sides simultaneously,
but to excavate and build up the respective half-sections somewhat staggered in
 MULTIPLE STAGE MINING METHODS 643

space. This procedure enables the invert to be placed as promptly as possible

after the construction of the side walls. This is highly desirable with rapidly devel-

C-0

4-80 4-80 4-80 480

.Section at boundary C-D

D Concrete level I .. 4-8(] , I , 4·80 • I • 4·80 , I • 4·80

b;

C'-0
E-F
F Temporary underpinning !
O
'

~ ,, ,,..,.'\
- Temporary underpinntil!f of'
Ci A -8 arch ,5pringing

Temporary
Jide wall eoncreting
c_, underpinning

FIG. 6/ 51. Vrtriations for the consttuction of side walls

 644 CONSTRUCTION AND DESIGN OF TUNNELS

Wall beam (wood or

steel or co!7crete)

lemporaq; arc/J
I pinl7ifl{l
Boundary ofcentral
cut I
I
I
-+-

FIG. 6/52. Arch underpinning for the construction of side walls

"Ti?mporary bearmg surf8ce

1
1, '
It I
~ --+-&de waII
I I Retnfbrced- '
I \ ·~ concrete wall beam
I I
I A Inc/med post
L __ --<'/ ' ,,
', .'>
' .
'
FIG. 6/5 3. Details for springing support ( wall beam) possibilities

oping lateral pressures. ([t must be noted, however, that where invert arches are
required the Belgian method is not usually applicable.)
ff the height of the side walls exceeds 3 m (10 ft), underpinning operations should
be undertaken in steps.
 MULTIPLE STAGE MINING METHODS 645

In more stable ground the abutment of the masonry arch can also be ·established
by a niche cut in the rock which affords a temporary support along face a while
the side walls are being constructed (denoted in Fig. 6/53 by broken lines). In this
case the lateral drifts driven for the erection of side walls need only sufficient
support to provide protection against spalling.
A modified later alternative to the Belgian method is its double-heading alter-
native which differs frqm the sing\e-heading method in that a second (bottom)
heading is driven in which work is actually started (Fig. 6/54).
The top heading follows the advanced face of the bottom heading at a distance
of about 25- 30 m (80-100 ft) only. The headings are interconnected by break-ups.
The employment of the bottom heading obviates the difficulties in the transport
of material, muck removal and drainage inherent in the single-heading system,
while excavating and walling operations can be done simultaneously without
mutual interference. On the other hand, the ground core - on which the full roof
pressure is temporarily transferred for the time of exC'avating the calotte and
constructing the masonry arch - is weakened by the drift of the bottom heading.
Consequently, whereas the most critical tunnelling phase, i.e. the breaking out
of the ca:Iotte and the construction of the arch was done in the single-heading
method in entirely intact rock, it will be accomplished in the double-heading method
in ground which has been slightly mobilized because of the previous driving of
the bottom heading. Therefore, the double-heading method is employed mainly
in rather firm ground and when driving larger tunnels, where the width and height
of the bottom heading are relatively small ·compared with the dimensions of the
tunnel, or where the quantity of ground to be excavated requires, at any rate,
the driving of two headings to provide for a reasonable rate of progress and for
uninterrupted working conditions. The employment of the double-heading method
can also be justified when driving relatively long tunnels (exceeding 1 km in
length).
Later on, when the section is deepened the bottom heading can be advan-
tageously used also as part of the centre cut (cf. Section E- E in Fig. 6/54).
Not only two, but three or even more headings are employed when tunnels
with large sections are _driven by the Belgian method.
The most important of the merits of the Belgian method is the relatively short
time interval required from the excavation till the erection of-the permanent roof
lining because the rapid construction of the masonry arch makes all subsequent
operations easier and safer under its protection, with the additional advantage
that the development of higher rock pressures is prevented. T_he reduction in
timber consumption and in the demand for long timber logs are added advantages.
It is also advantageous that tunnelling op.:u-ations can be carried out simultane-
ously and undisturbed in a number of separate working sites (especially when
using double-heading methods). Finally, the method can be easily and advan-
tageously combined with other tunnelling methods, e.g. a combination with the
KUNcz method (see Section 62.44) is shown in Fig. 6/55 when a small crown-short
adit is driven ahead with a roof supporte? b,Y forepoleg crown bars instead of
 i
°'

Sectio,r A-A Section 8-8 SectionC-C Sectio11D-O Sect1011f-E SectionF-F

I),'

~
'\ ·~ ~ / ~·-, ,·...
T.
I"' ., r
.·. :: :·:m
. .., : 0_. .,
,\ -,
\ I
8z
l~
I I
I I
11
J J

..,
'I
I
-~.r1m
-11%'
,,,, ~ ~~~
JI
I I
\ t.
--i.--:l
t
' '
.

. '.
I I
I/
!....l
'
I I
I I
I\ <
I\
\..J
'
,
-, .
.
: I
I I
I I
/ /
LJ
'\
I I
\ \
LJ ,
r I /
., l_j
,,
I/
/
~
"'C:
(')
-i
i&o/''2-f(J I 2-10-!"~' -,, 0
z
'"" i i;;::,
,.....,••t
~ - '"" I
>
z
C,

,,,C,

i:,,_
1<::::i
.ILD h ~
"'ci
z
.,,0
-i
C:
z
z
'"f;;
PIG, 6/54, Details of the double-drift Belgian method
 Bl Aj

.. :• 'Fxcavatloh of -~ ,. ~- Top arc/; co11slmdt0n ' I [ZQ I,1-2e1

f~ I, ~side wall and ~ ,~, Advance drtf'I.
-----~..- vtde w<11/ ~ constmcLtoll ~ ' I J ~

I 2-00_,_
read!/ ~ -: '

J::1
C:
. ~_u . , ~
,~ ----~--f---_,_ --_L - --- --- - - --- - -:_-:_ :;;
r'
"'
l?emov;;/';o~;---B j o:0-70m ----------~- ~--------~t Cl>
-I
- - ·--· >
cl ---· Cl
"'
~
SecL1on A -A Section 8-B Section l'-l'
Lo,ngitudmal laggmg z
scrap -rail rtbs steel zCl
formwork ~

Steel bfc1ck '.:l

:i:
0
0
Wall beam "'

I
I
'I
I
\ \ I / ,
\__ J-----r-----i
.... ____ ----- . )
'

0,

FIG. 6/ 55. Flying arch method combined with steel supports (K UNZ) ~
 648 CONSTRUCTION ANO DESIGN OF TUNNELS

a regular top heading. The calotte can be excavated in its full section under the
protection of rail ribs placed on the crown bars and arranged along the annular
space established. The side walls are constructed by the employment of sinking
pits at intervals, serving for the construction of masonry columns performing
the counterforts.
Beside its merits, some drawbacks are also attributable to the Belgian method
principally because of the intricacy and dangerous character of both arch under-
pinning and underhand stoping. The masonry arch when ready often becomes
fissured due partly to the poor compaction possibilities of side wall concrete and
to the blasts applied in breaking out the place for the side walls. A further dis-
advantage is presented by the spring-line interruption joints established not only
at the crown but also at the counterforts of the arch, i.e. at the most exposed
displacement points. Moreover, haulage tracks and drainage ditches must be
repeatedly re-sited and, finally, the casting of the invert arch, if any, for protection
against lateral pressures is performed last of all tunnelling operations, i.e. only
after a long delay. ln conclusion - according to SnNI - the Belgian method can
be advantageously applied in the construction of medium length tunnels driven
in dry ground or in fissured rocks of medium strength. If running or rolling
ground results in heavy rock pressures the application of this method becomes the
less advisable the greater the span of the tunnel and the bigger the rate of water
inflow.

62.32. The Core-leaving or German Method

This method is applied when the load-bearing capacity of the ground is insuf-
ficient to afford safe temporary support for the masonry arch during the construc-
tion of the underpinning side
walls, so that these must be built
up first. The top tunnel section
(the calotte) is only excavated
I
I subsequently and the masonry
I
arch is built upon the already
I I '
6
I . "
r__
I
I
_l ___ _ .i_ _ _
I
l_ ___ _r
completed side walls.
ln its most characteristic form
I I the method starts with driving
I I
I I I _
11 Sprmg two bottom headings, one at the
I I
: I
!me drrft foot of each side wall (see I in
-j f- BotLf'm Fig. 6/56). These are driven ahead
l I
I I dn/'f to the chosen length of the work-
ing section, each of them being
used during advancement for
excavation and mucking, then
Fm. 6/56. Sequence of part excavations tn the during their return for the deliv-
core-leaving or German method ery of masonry material and to
 MULTIPLE STAGE MINING METHODS 649

Bottom drif't drivifl!J afld concreting lns/,3/£9t10/l of lbrmwork afld ,;rc/1 roflcrettfl!J
{'ap left outside ti;
8) .......~ - --+-- ~ ~ - from extrados

Le,1n conc-
rete

Post le bel!ind
masonrg
Drivifl!J oftop drtf't Concrete /afllll!l it tl!e 1ilvert
,.

ElliJrgement of the /Je,1difl!l Construction ofeventual ill/lei'

reinforced-concrete !if7ing
/'J

Remforced-
concrete post
left beyond
ma,501,ry

Fm. 6/57 . German method without spring-line drifts

 650 CONSTRU CTION AN D DESIG N OF TUNNELS

give space for the construction of the side walls. Side walls are compac
ted
tightly to the earth wall outwards, whereas tpe inner space left between
the
inner wall face and the outlines of the heading sets must be fir;mly
back-
packed with boulders or lean concrete , which is to. be cleared away
later
during the removal of the core (Fig. 6/57). If the height of the side
walls
admits, two upper stage headings are also driven above the bottom heading
s
(cf. Section 63.33), rendering the operatio n of muck removal , or the transpo
rt of
construction material , and concreting operatio ns possible . Should upper
stage
(spring-line) headings not be employed, bottom headings are interconnected
at
definite intervals by cross-cuts, making allowance for the simultaneous construc
-
tion of invert arch ribs there, which is essential in loose ground. For very
wide
tunnels also the driving of two double-bottom beadings can be considered . By
the
separation of excavation and haulage work and by providing for undisturbed
and
continuous transpo rt facilities through double headings, the construction
time
of the walls may be reduced to about one half of that required for single-heading
operatio n (Fig. 6/58).

FIG. 6/58. German method

with spring-lin e drift

In higher sections the use of multi-stage headings is preferable, in which case

the upper spring-line heading (/) is to be driven ahead and pits 3-4 m (10-13
ft)
wide are to be cut from this to the bottom heading at intervals, in which
the
erection of side walls may be started.
The construction of the side walls is followed by driving a centre top heading
(3)
which is then widened out in the same way as in the Belgian method (4) and
the
masonry arch is bearing upon the already constructed side walls. In the course
of all thesa operatio ns both the posts and formwork are still support ed from
the
 MULTIPLE STAGE MINING METHODS 651

inner ground core, which is removed only in the subsequent phase, followed ,
finally, by the placing of the actual invert (cf. Figs 6/ 56 and 6/ 57).
Another advantage of this method is its economy in timber consumption as
the ground core left in the section secures the support for arch timbering, provided,
of course, that it is wide and firm enough to ensure an unyielding support. Any
undue compression of the core under loading would result not only in further
loosening of the ground above the roof, but would also lead to cracking of the
concrete arch, especially when movem~nt~ take place during setting time. A further
advantage consists in the excavation of small sections only at one time where the
respective section of the permanent lining is immediately built in. A disadvantage
of the method lies in the difficulties encountered in haulage, as a consequence of
I
,/2-40
2

( \
/,,.------ .....,

L------ -~
tJ I
a ,
\
, ,.,,.,,----- ..... , ,

L- ------~ I

--,,,
\.._
D, .
________
/,
%

5 c
,<}.¼
_
,,,,,. ' ..... ,

~ ~ D;"
~
L ________ _,1'

Sequence ofpart f?XC.illl8tioll

Pv..--,t:::~s.-n''lJ.
l,.., Propp111g for the wide
111ng of' tilearc/I sectior
I
I
I
~- 860 I
~~~~~1=· ---- __ _ _0@1
4tr duct
//-00 ·

F1G. 6/59. Construction of the Rove tunnel with multiple-drift method

 652 CONSTRUCTION AN D DESIGN OF TU NNE LS

which bottom headings must be driven to their full length before side-wall con-
creting can be started. Further disadvantage s are that the ground is always loosened
so mewhat in the calotte before its break-out can be started, and that construction
of the invert has to be postponed to a certain extent.
To eliminate the difficulties in haulage and ventilation a centre bottom heading
may be driven ahead and used in wide-sectio n t•·nnels, from which cross-cuts are
made at certain distances in both directions to permit the driving of the respective
stretches of the side-wall headings from them. · Another possible combination is
to start with the driving of a top heading, as in the Belgian method, and
after widening o ut the top section in the usual way trenches a re cut down for the
construction of the side walls, leavi ng the core between them intact. This also
shows ho w the German method is often combined with the elements of other
tunnelling methods.
Du~ing the whole process care must be taken nor to weaken the core excessively
by the driving of lateral headings, as the core must constitute the support of the
arch timbering. For this reason this method can be employed only fo r building
tunnels of considerable width and is especially suitable for the construction of
underground stores, factory halls or shield chambers. There is no doubt whatever,
that in ground under heavy rock press ure this method affords the greatest safety
in construction . Where the a rch sectio n - owing to the running character of the
gro und - can not be excavated safely by the usual method of widening, additional
headings may be inserted between the standa rd headings driven at the springing
and crown when the tunnel sections are large eno ugh This allows the arch masonry
to be constructed by steps. Naturally, the space left between the outlines of the
arch a nd of the heading must a lways be thorough ly backpacked here as well
(cf. the construction of the R ove tunnel, Fig. 6/59). In addition, of all classical
tunnelling methods this provides the greatest safety against surface subsidence.
As tunnelling costs will increase directly with the nu mber of the beadings and
break-ups, the German method is far more expensive than all the other methods
hitherto discussed , but this is partly compensated for, nevertheless, by a higher
degree of safety.

62.33. The Italian or Invert M ethod

This tunnelling system (Fig. 6/60) is applied in very soft and treacherous ground
where considerable lateral pressures a re a lso likely to develop. On account of
tbe developmen t of these pressures it must be ensured that only small areas are
opened and are supported immediately by temporary timbering or, preferably,
by the subsequent parts of the permanent lining. With regard to th is, the con-
struction of the lining is carried out in steps in two or three vertical stretches
separated by horizontal · ,anes. Construction starts with laying the base (/- 5)
prior to the erection of the side walls, as otherwise the loads upon the walls could
not be tra nsferred safely on to the soft ground without involving undue settlement.
 8.

1 1<, . r,, (() iwlian 111.:tl111d

 MULTIPLE STAGE MINING METHODS 653

Furthermore, this order of sequence is required not only for load transfer, but is
necessary to provide an early support to the side walls against the rapidly increasing
lateral pressures while construction of the walls is proceeding.
A bottom centre-drift is first started and driven at about the top level of the
invert for a very short distance (2-3 m: 6-10 ft) after which it is enlarged to the
full width of the tunnel by driving narrow strips of cross-drifts with very heavy
and tight timbering. The invert and as much as possible of the side walls are
constructed in these drifts and the open area is back-packed by rock spalls or by
lean-concrete (7). Since the timbering of the bottom drift must be very strong,
steel is most often used. The invert is cast from pits which are, sunk below the
floor of the bottom drift and adjoining it and the side-walls are also built up to
the roof level of the drift (6). After the backpacking (7) already mentioned, only
little space remains open for the advancement of the bottom drift, i.e. for transport
purposes. A centre top heading is then driven (7-8) and enlarged to full width (JO)
but limited in length to the previous dimensions where the corresponding section
of the masonry arch will be constructed. This upper section of the tunnel is very
heavily timbered. Tightly spaced transverse polings are supported by radial posts

Crown drift

FIG. 6/61. Enlargement

of the heading with
radial props

seated on a sill placed either on the bench left between the drifts or on the previously
installed backpacking (Fig. 6/61). In higher tunnels an intermediate centre heading
is installed between the top and the bottom drifts, from which the intermediate
part of the side walls is built up, at least, to spring line level according to the
pattern given above. The loads from the roof on the supports can, thus, be trans-
ferred either to the walls or through the dry packing to the permanent invert.
In loose or running ground, however, it will often happen that even the masonry
arch can only be built in portions from several short axial and cross-drifts pro-
ceeding from the springing towards the crown.
Under especially difficult circumstances the cap of the bottom or the inter-
mediate drift is also constituted by an auxiliary masonry arch (cf. 1 I in Fig. 6/60),
offering a steady support for the excavation of the annular top space and for the
construction of the masonry arch.
654 CONSTRUCTION AND DESIGN OF TUNNELS

After completion of the masonry arch the core and the temporary dry-packing
are removed. The removal of this packing, when made of lean concrete, may
require some blasting. This lean concrete must, at any rate. be separated from the
permanent masonry by insulating paper.
The main advantage of this method is that the workings are sufficiently small
to be readily braced and maintained in treacherous ground.
It is adopted only under extraordinar ily heavy ground conditions because of
its high costs and its very slow rate of progress. Modern shield methods would
now be used for all tunnels for which this method was designed, except possibly
where only a short length of very treacherous ground is encountered . For example,
it was employed when driving the Simplon tunnel on a stretch of its southern
section under very high rock pressures which could be perforated only by the
application of this method (cf. Fig. 6/60).

62.34. Combined Tunnelling Methods

The tunnelling methods already discussed are no longer used in their original
design even for tunnels, much less in the construction of large underground halls,
shield chambers, etc. but their working phases are generally mixed and combined.
The most usual comb nations are developed from items of the English and
Austrian' methods, or from those of the Belgian and German methods.
The possible variations of these combination s are numerous; some of them
have already been mentioned in the discussion of the methods themselves. In one
of these variatio ns, e.g. the advantages of the English and Austrian methods are
combined and, at the same time, the radial-post suppo rting method is used for
widening the top section (Fig. 6/61).
In the most frequently used combination s the lining masonry is built in rings,
in the same way as in the English method, carrying out in each ring only one
working phase at a time. Construction proceeds, however, not in a single tunnel
sect ion only but work is going on in several tunnel rings simultaneously, con-
nected by break-ups to an advance bottom heading in order to speed up the rate
of progress. The continuity of work is, thus, secured by the separation of the
working sites without leaving areas open for a long time as opposed to the Austrian
method where the various working phases are carried out simultaneous ly, but are
staggered in space.
Segmental arch timbering which is well-adaptab le to varying pressure con-
ditions, is adopted , but is combined with transverse poling in soft ground. With
the application of the radial-post system the excavation in the a rch segment must
not extend beyond the extrados as the crown bars which would require these
ovcrbreaks in the English method are omitted here. On account of the radial
arrangement of posts the elements of both segmental arch timbering a nd poling
can be removed immediately prior to the construction of the respective st retch
 MULTIPLE STAGE MINING METHODS 655

of masonry in such a small area and for such a short time that the bridge-action
period of the ground will afford protection for a safe casting of the wall section.
The advantages of such combinations were clearly demonstrated in the construc-
tion of the shield chambers for the new subway in Budapest as discussed in
Section 63.33.
II-II 111.-ltl.
Axial lbrepo/iag

Top dnft

Loag1tudlf781 SE?C'tioa -Aux!li8f'!/ C'fOW/1 b8r

I
1'1.-/V.
!

.,
lw

F1G. 6/62. Pentagonal

timbering for the heading :
(a) with vertical posts;
(b) with inclined posts

Among combined tunnelling methods may be mentioned the American method

which is characterized by the use of polygonal arch timbering and by the greater
use of side posts without interior struts.
The face is opened by a top heading, usually rectangular in shape, which is
poled longitudinally and timbered in the usual way. Tn the space between two
6% CONSTRUCTION AND DESIGN OF TUNNELS

bents of the heading the crown segments (caps) of the timber arches are set in
position and held in place temporarily by secondary posts or by strips or scantlings
spiked to the main posts as shown in Fig. 6/62. A short length of 0·60-1 ·80 m
(2-6 ft) is then widened out without polings or other roof support and the seg-
mental arch elements adjoining the crown segments are put in position and held
by iron dowels and by a short prop. If the segmental timbers are not set close
together, lagging is inserted above the timbering and the cavities between the
lagging and the ground are backpacked. The widening out for the next segmental
timber is done in the same way down to the sill. After the two sills have been
placed the roof timber is completed and the bench remaining below the top heading
sill is removed.
During each operation the sill timbers are underpinned by any one of the fol-
lowingmethods. If the material is very firm a longitudinal cut is made in the bench,
leaving a beam on which the sill rests and the room for the counterforts is exca-
vated at intervals under the sill, and posts are placed. If the material is too soft
for this, pits are sunk at interva ls and at first short posts are placed to support
the sills, and, later, if excavation has proceeded these are replaced by long posts.
In sufficiently firm material the sills or wall plates are set in niches at about the
springing line and no posts are used. The number of units in the segments varies
between 3 and 7 depending on the width of the tunnel section . The method has
been successfully and economically used in fairl y firm material which would stand
for a short time without ~upport. Its chief advantages are the large open area
within which the masonry lining can be built continuously from invert to crown,
and the economy in timber. It may be used also in rock, when the top heading
is timbered with caps and posts and, while the arch section is being widened
successively by short lengths, the elements of the segmental arches are joined
directly to the caps. After reaching down to the springing, sills are placed and the
posts may be removed (Fig. 6/62).

6.24. RECENT T UNNE LLI NG METHODS

EMPLOYING BOTH TIMBER AN D STEEL SUPPORTING-ELEMENTS

The tunnelling methods hitherto discussed all have a high timber demand.
It is true that generally only the poles are definitel y lost from the timber used for
s upporting purposes, while elements such as crown bars, sills, posts, etc. are most
often recuperated when the permanent masonry lining is built. Nevertheless, both
their recuperation and re-use always involve some inevitable losses as some will
break under excessive loading while others ha ve to be cut for re-use, in addition
to which considerable losses will also ha ve to be ascribed to decay accelerated
by the damp atmosphere of the tunnel, all resulting in the grad ual consumption
of the whole of the timbering material.
Owing to the fact that in the drivi ng of any particular tunnel or heading of
constant cross-sectional dimensions, the temporary supports required are always
 C"
'.·
'·.l ,
,, "--f- '
. - l '
,,,
I

"i} )
f
" :.-1
.,,
f

I -,.7
1
J,,
l .;
;-~
c-:
'J
C

..c

"'
~
~
>
:fu ~
~
&,
-<:::,
E; Q)
.§ !ll
"" -<:::, ~
~
~ ~
~
\ ~
~
~
~
c:'l
~
~ ~

. -~
1'l
~ ~
~

---~-- .................
',
' '\
\
\
\
I
- __ \
I
i
I
I
I
I
I
/
/
_,,,,,,,/
 MULTIPLE STAGE MINING METHODS 657

of the same type and dimensions, it is a matter of course to favour the employ-
ment of supporting elements suitable for multiple re-use. This has led to the devel-
opment of steel and reinforced-concrete supporting structures and of tunnelling
methods making use of such temporary supports. As supporting elements of this
kind have already been discussed in detail in connection with headings, only the
more usual tunnelling methods which make use of tl~m will be discussed here.
Steel elements can be employed in the following th~e combinati ons:
(a) steel liner-plates instead of timber lagging (poling boards), the eventual
bents being of timber
(b) steel supporting-hents rombined wit.h timber lagging
(c) steel bents with stee!- liner-plates.

62.41. Tunnelling with Liner Plates ( the Needle Beam Method)

The liner-piate J11ethod is generally employed for driving steel-lined small-sec-

tion drifts or headings (2·5-4·0 m : 8- 13 ft dia) in medium soft ground. It can also
be adopted for small cross-section drifts even in running ground when combined
with compressed-air dewatering . Its simplest application is shown in Fig. 6/63,
where steel liner-plates 0·40--0·50 m (16-20 in) in width, 0·90-1·80 m (3-6 ft)
in length and ·3-6 mm (1/8- 1/4 in) in thickness, with 40-50 mm {l 1/2- 2 in) deep
rolled or welded flanges (rims) are used as roof supports (cf. Fig. 6/11). The first
liner-plate of 40-50 mm (l l/'4 -2 in) width is placed as the crown segment in a
previously mined cavity at·the top and two adjacent liner-plates are bolted to it,
one on each side, after the hole has been sufficiently widened. These plates are
temporaril y supported from the ground core by trench jacks or by carefully
tightened props (Fig. 6/63a. The arch section is then gradually widened down to
the springing line and the liner-plate ring so obtained is wedged outwards from
wall plates or wall beams placed at the grooves b. In small-sectiorr tunnels and
in softer ground this plate ring can be left unsupport ed during the· constructio n
of the masonry arch, as in tl}e Belgian method, the formwork being temporarily
supported from the bench. While the a rch concrete is being cast, the liner-
plates are gradually withdrawn and regained. After an interval of 3-4 days the
excavation and constructi on of the lower tunnel section follows. This can also
be accomplished in rings as shown in Fig. 6/63c.
Another variation of the method is the 'needle-beam method' (Fig. 6/64), where
the full cross-section of the tunnel is broken out and the plates, which in the course
of the excavation are set up one by one, are supported by radially set trench
jacks or props from a centrally placed longitudinal girder, the 'needle beam',
which consists of two heavy steel joists bolted to each other, the space between
being filled with hardwood . The length of the needle-beam is chosen to exceed
the daily advance by 1·0-1 ·2 m. It is placed at the bottom of the top heading,

'12 Sz~hy · The Ari c,.r Tu11Drllinr

658 CONSTRUCTION AND DESIGN OF T UNNELS

its rear end being suppo rted during the driving by a post from the concrete invert
of the tunnel which has already been placed in position. When the needle beam is
placed , the trench jacks which formerly stood upon the core will be taken out as
the new ones are set fro m the beam. While in the upper half section also timber
props can be used instead of trench jacks, in the lower half, steel trench jacks
are best to counteract the excessive bending deflection of the beam by their restretch-
ing, es pecially in really bad grou nd with r...<cessive rock ground-pressures. The
needle-beam method is, however, not d..ifoid of some drawbacks. First of all,
th e heavy beam must be moved forward by hand, and the large number of trench
jacks interferes wit h both communication and work. The use of liner-plates
without stiffening ribs is, in additio n, always somewhat risky and m ay be toler-
ated only in firm ground and small diameter tunnels. A further important point
is not to excavate sectio ns ahead to distances exceeding one day's concreting
in order to avoid the development of high rock pressures. Moreover, each row
of trench jacks must be connected by axially placed etchers to prevent its slipping
o r kicking out on the steel washers.

62.42. The Full~face Erector Method

This variation of the previous tunnelling method is often used. In large circular
tunnels, such as subaqueous highway tunnels or subway tunnels, where, on account
of the greater diameter a nd of the higher rock pressures, the tunnel lining must .
consist of heavy cast-iron segments or of precast or prestressed reinforced-
concrete blocks, which are very difficult to manipulate if placed by hand. In such
cases, special equipment, i.e. erectors with mova ble arms, usually employed in
shield tunnelling (see Sectio n 63.22), are resorted to for placing the segments or
blocks (cf. Figs 1/ 10 and 7/20), whereas mining is effected either by the full , face,
or the heading and bench, or the crown-bar method.
Such heavy lining, elements can be placed only when proceeding from the botto m
upwards as a precisely fitting ring of precast concrete blocks o r cast-iron segments
can only be built if their subsequent settlement or displacement is completely
avoided and their geometrical trueness to shape is secured right up to the comple-
tio n of the ring. ln this respect the heading and bench method would o nly increase
the period between excavation a nd lining - involving continued loosening - so
it is generally mo re practicable to apply the full-face method. This system can be
applied in rather firm dry ground possibly only in combination with rock bolting.
In wet or loose ground the add itional application of the compressed-air process
may be of great help.

62.43. Tunnelling with Liner Plates and Stiffening Rings

The liner-plate method is used in very large tunnels in combination with stiffen-
ing ribs.
 M ULTIPLE STAGE MINING METHODS 659

62.431. System employed in the Chicago Subway. A highly practical system was
employed in the construction of the Chicago Subway as shown in Fig. 6/65. The
ribs consisted of 150 mm I-beams and were spaced at 0·60-0·75 01 (2- 2 1/2 ft)
intervals, the liner-plates being bolted inside the outer flange of the ribs instead
of outside (cf. Fig. 6L12). The segmental ribs are composed of several units con-
nected by bolted fish-plate joints allowing the face to be excavated in successive
phases.
150x150 radial prop
I150 steel ribs

Lower benc/1

1' v0)··

Bottom sect10n

FIG. 6/65. Liner-plates

stiffened by
steel ribs

A top heading is driven first and the crown plates placed with the corresponding
rib segments in its roof and supported by trench jacks from the ground core.
In continuation, the whole top section is successively excavated down to the spring-
line on both sides while further liner plates are attached and supported by further
trench jacks from the core. Then pits are sunk at both sides and further liner
plates and rib segments attached, supported now from the inner pit-walls by other
trench jacks or props. The last segment-ribs acting as posts always rest on washer-
plates laid on the bottom of the pits and sufficiently large in surface area to transfer
their load to it safely. Instead of separate washer-plates longitudinal wall-beams
may be used to support the ribs. Finally, the ground core is attacked from all
sides and cleared away.

42•
 Blocking
Shuttering Arch concrete J8Cks
Longitudinal
lrenc/J jack Ril,sing .)8ck stifl'enmg
' \ Arttculated /!Jrmwork rib 0 formwork
~ \ Trc1vellmg cc1r -

t J Adjusuble movmg
arm Temporary c1rr:h rtb
lrencll jacks
Lmer plates
Invert concrete
emporary steel
arc/J rib

~--+---~
~
'/J j3cks Muck OJI'
i,.g;; plates
- - _'.ii-~ .
Aifjust,ablt> 8rm

Wedges
Wedges I
W!J/1 p/8tes
I iroot pi!Jtes

F1G. 6/66. Liner-plate method with strained-steel ribs (SZADECZKY-V:.STAOH)

 MULTIPLE STAGE MINING METHODS 661

The masonry lining is constructed from the bottom upwards in rings as wide
as the width of a liner plate, the liner plate and ribs having been dismantled pre-
viously, thus leaving the ground faces unsupported for this time. For this reason
the masonry work must be accomplished as quickly as possible and therefore
preference must be given to the use of precast-concrete blocks, ashlar stone or
bricks as lining material. For the sake of better bonding between the single courses
of blocks or stones, the masonry ring walled up extends to the width of two liner-
plate rows.
Naturally, the whole working cycle is staggered ·in si,ace, the top heading
being driven ahead, directly followed by its widening, whereupon the various
stages of sinking pits or side cuts are effected, completed by breaking down
and removal of the core, casting the invert and the construction of the
masonry walls and top arch in successive shifts (cf. Fig. 7/33).
The formwork supporting the shuttering of the inner wall faces is usually also
a steel structure, strutted from the core or from the opposite sides of the exca-
vation by trench jacks, the corresponding liner plates and rib segments being
strutted at the same time by another set of trench jacks or props from the form-
work proper.
62.432. The liner-plate method. This method, described below, was employed
with certain modifications in the construction of the concrete-lined sections of
the subway tunnels in Budapest (Fig. 6/66). The flanged liner plates used were
50 cm (1 ft 8 in) in width and had a length of 90 cm (3 ft) and a thickness of
8-10 mm (3/8-5/12 in), with flanges 8 cm (3 in) deep.
The face is attacked at the crown, a drift being broken out equal in width to
a liner plate, at a height of 1·5-2·0 m (5-7 ft), where the first liner plate is placed
supported by a trench jack from the drift-bottom. This plate will also be tempo-
rarily cantilevered by attaching it by bolts to the previously erected plate-ring
section. The drift is then widened out towards both sides and further liner plates
are bolted and propped temporarily by additional trench jacks down to the spring
line. The last flanged plate at the spring line is placed on a longitudinal wall beam
and the plates forming a thin arch are firmly prestressed by wedges driven between
the wall beam and the overlying plate to render it self-supporting while the trench
jacks are replaced by adjustable propping arms mounted on a car moving on
a central track in the heading. This, again, only provides a temporary propping
to support the liner plates until a rib of steel joists composed of two quarter circles
is installed. These steel ribs are butt-jointed at both ends and provided with carrier
plates bearing against each other at the crown and against the wall-plates or laid
upon the wall beams.
After installing the steel ribs the moving car is withdrawn. The steel formwork
is erected a nd its shuttering placed for the construction of the masonry arch.
Prior to this, however, the steel ribs must be propped by short trench jacks from
the formwork ribs. When starting with the construction of the masonry arch
the lowest liner plate and the lower quarter of the stiffening rib are removed, and
662 CONSTRUCTION AND DESIGN OF TUNNELS

the remaining set of plates is simultaneously propped by additional trench jacks

from the steel formwork ribs .

. :--- .

I
+
I
I
i
__ ___ -- - -1- =~~=-~.:_-=-
~--
Fm . 6/67. Side drift IT'"'t!c.od

In this way the masonry ring can be safely built and both the liner plates and the
stiffening ribs can be regained.
The construction is continued and completed according to the Belgian method
by underpinning the masonry arch by pillars constructed in pits or by staggered
bottom-rings built up in transverse cuts.
In firmer ground the method has been modified by abandoning the stiffening
ribs, and directl y supporting the liner plates by trench jacks from the steel form-
work. This, however, has proved to be admissible only in cohesive ground. More-
over, care must be taken that the excavation is closely followed by the construction
of the masonry and that, above all, the liner plate rings shall not be left unsup-
ported except wit hin the bridge-action period.
62.433. Side drift method with steel supports. In fissured or treacherous rocks
and for wide tunnel-sections side drifts are driven ahead of the main excavation
for some convenient dista nce. The steel s upport of these drifts includes the main
tunnel s upport posts and wall plates. Just prior to excavation of the main face,
the drift support proper is removed, leaving the main posts and wall pla te in
position. These extend above the muck-pile, thus permitting erection of the main
arch-ribs before mucking out . Where rock conditions are hard, break-ups are
made towards the crown leaving the core in place, thus offering a convenient
surface from which the roof can be quickly supported by jacks or props until
 MULTlPLE STAGE MINING METHODS 663

Crown dnft

Tran,sverse cvt

'Stde dr,rt
I
·----.. ! /
--- --=----L.-;---- -
1

Fm. 6/68. MultiplQ--drift method

the arc~ ribs can be placed. The ribs are usually multi-piece for manual handling
in the restricted space. The main tunnel is excavated to the full face where rock
conditions permit. The ribs are then usually installed in two pieces for speedy
erection and early support of the main roof. Crown bars may be used to hold up
the roof till the ribs are placed. When all temporary supports of the full section
have been placed, the casting of the concrete lining is started. In practice, the
formwork is strutted from the sides of the ground core. Finally, the ground core
is excavated and cleared away and the invert is constructed last.
A variant of the side-drift method is to make the side drifts large enough to
permit the construction of the respective part of the permanent lining just up to
the springing. The main arch ribs are then installed as an arch on these concrete
abutments. The main support and the drift support are of the continuous rib
type (cf. Section 61.311) in this case. As can be seen, this procedure is even more
similar in principle to the German method.
A further variant is obtained when an additional top drift is driven, in addition
to the side drifts, and the supporting rib segments of the arch are placed in trans-
verse cross-cuts upon the top of the side walls previously cast in the side drifts
(cf. Fig. 6/68).
 664 CONSTRUCTION ANO DESIGN OF TUNNELS

62.44. The Kunz Method

Another widely used tunnelling method is the KUNZ method using mixed mate-
rials for temporary support, namely steel formworks and arch ribs combined with
longitudinal timber forepoling and timber props (Fig. 6/69). In this method the
top heading is driven ahead in step-like stages. First, a short drift is advanced
in the crown section of the top heading, "the roof of which is supported by crown

Aux1!1ar!J /brmwork - ....

A
----
-f-48
• 0
Crown
•
:- _ t.
~- .
,,,,-- 4-cm timber '3gging
Scrap rail

-~
'"
I

I \ 1h111.r:
Deep cd i'or side wall
--~+----

footing mat of'

6cmpltlllks

~/ Ar(!ustable steel bloc/r7ng

')1"

FIG. 6/ 69. Kunz method of mixed supporting material

 MULTIPLE STAG E MJNlNG METHODS 665

bars pushed forward as cantilevers. The drift is then widened out towards both
sides and steel supporting ribs (usually of small profile rails) are placed piece by
piece at the extrados of the heading bearing the axially forepoled lagging. The
front face of the heading is supported by breast boards strutted from the ribs,
if necessary. These multiple-piece ribs are· composed of 3 or 4 segments con-
nected to each other by bolted fish-plates as widening proceeds. In the same drift
is the multiple-piece steel formwork also erected by a step lagging behind. The
lower pieces of both stru·ctures are seated upon a longitudinal wall
beam, the external ribs being braced from the formwork by special
.... ., steel collar braces, known as riders. These riders are short steel-collar
props consisting of two:channel or four-angle sections which can be
·D• •

.; -
seated into the 100-120 mm wide interstice of the double-channel
steel segments of the formwork, where they are fixed and adjusted
in length by hardwood wedges (cf. "A" in Fig. 6/69). When the
riders have been placed, the top heading is broken out and the steel
(otmwork supported by props from transverse sills. Side cuts are
sunk next, while the transverse sill is underpinned by posts from
the bottom of·these cuts.
The construction of the masonry lining is started in the side cuts
. :o .. •. from the bottom upwards, the cuts themselves being supported by
.o: ·.
struts from the ground core. When the masonry is raised to the
spring-line level, the lowest rider is replaced by a timber block and
the masonry arch raised_to the level of the next rider, proceeding in
~ .. this way until the masonry arch is completed at the crown. As the
construction of the arch proceeds the external supporting ribs are
\ flau/age track
withdrawn piece by piece. In loose or running ground, however, it
•1/(
is better not to attempt the recovery and disturbance of both fore-
poling and steel ribs; it is better to leave them in place and incor-
porate them into the concrete. (This is also one reason why these
ribs are usually made of used small-section rails.) Otherwise, it is
rather common practice, in this and in other tunnelling methods for
laggings, ribs and cmwn bars not to be recuperated but to be incor-
porated in the masonry concrete. In such cases, timber will decay
with ti me resulting in surface subsidences. With the KUNZ method
this decay is restricted to the wooden forepoling only and so is of little
consequence.

62.45. The Cologne Method

In order to prevent undue subsidence owing to the decay of lost

timber elements a special method, the ' Cologne method' has been
developed where all lost elements encased in the concrete are of steel.
The main features of this method are shown in Fig. 6/70. As shown,
 666 CONSTRUCTION AND DESIGN OF TUNNELS

corrugated steel boards (spiles) are used here for forepoling instead of timber
poling, bearing on the steel stiffening ribs, and only the props and the transverse
sill are of timber. The forepoling of the l ·8 m (6 ft) long spiles is effected by
pneumatic hammers while the supporting ribs are spaced at l ·20 m (4 ft) only.
Beside this considerable protrusion a substantial overlapping of the single spiles
is a perfect safeguard against any intrusion even with the most squeezing and
very loose grounds.

J,01/e lbrm "B •

-o
~ -::-:::.:_;:_:n
Jpile lbrm r·
~

FIG. 6]70. Section of the 'Cologne· method

Work is started here, also, with the driving of the top heading, which is con-
stantly advanced during the following operations by a distance equalling the
spacing of the ribs. Then the segments of the stiffening steel rib are placed in an
annular cut made transversely from the heading, supported by props from the
core. Then, forepoling of the spiles is effected around the outer flange. The core
in the top heading is then removed, and the props replaced by longer ones provid-
ing a new support, on a transverse sill. The excavation face is supported by breast
boards in the meantime, as are also the faces of the annular cut. Jn the following
stage, the steel ribs are lengthened in pits sunk at both sides down to the bottom
of the tunnel, the transverse sill being underpinned by posts in the sa me way
from the bottom. After this, the concrete lining is poured, starting from the
 / J/Jaft
Road level

&'m= cr______ I . .;i=- -

©:=:::---r.
--------,r.--------
.

z
;
3::
;,:,
0
(")

Steel s!ieet pile

FIO. 6/71. Isometric view of the 'Cologne' method

°'
°'....,

,~
;\.
 668 CONSTR UCTION ANO DESIGN OF TUNNEL S

bottom , simult aneous ly with the erectio n of the formw ork

and the remov al of
the inner props and posts. Thus, the inner formw ork is
erected here only after
the excava tion of the full tunnel section , contra ry to the
KUNZ metho d, where
it is erected simult aneous ly with the suppo rting ribs.
The great ad va ntage of the Colog ne metho d lies in its safe applic
ation in shallow
depths (under a cover of 4-5 m) and in quite loose ground
withou t incurri ng sur-
face subsid ence. Its great drawb ack, on the other hand, is
that the steel of both
the suppo rting ribs and the poling boards is lo~t as they
will be embed ded in the
concre te. ln Fig. 6/7 l one shaft excava ted in a sheet-piled
enclos ure is shown ,
from which tunnel ling may be started by simply cutting out
the sheets in the form
of the tunnel cross-section.

6.3. TUNN ELUN G fN LOOS E GROU ND

AND UNDE R WATE RCOU RSES

The most si mple metho d is the combi ned 'cut and cover·
tunnel ling.
Public utility tunnel s, sewers and shallow underg round-
railwa y tunnel s are
often built with the cut a nd cover metho d. This is usually
cheape r and more
practicable than tunnel ling, up to depths of IO m (35 ft) in
open trench es. the sides

FIG. 6/ 72. Constru ction of the old subway in Budape st with

'cut and cover'
 CIJT AND COVER METHOD 669

of which are supported by sheet piling or by simple timber bracing (e.g. 'Berliner
Bauweise'), dewatering being effected either by draining to sumps or by ground-
water lowering depending upon the nature and stratification of the ·ground. This
method was employed, for example, in the construction of the first Underground
Railway in Budapest (Fig. 6/72) in 1896 and was developed and perfectioned in
many respects in the construction of a series of modern subways (Paris, New York,
Toronto, Montreal, Stockholm , Frankfurt, etc.). Although this is the most econo-
mic method , it interferes with street traffic, restricts the alignment to the street
network and requires the transfer and reconstruction of the intervening conduits
of public utilities. It cannot properly be included in tunnelling methods, but consti-
tutes rather a foundation method.
When traffic does not permit an open cut in the entire width, the pavement is
taken up, necessarily at night and replaced by a temporary deck under the cover
of which the subway may be built. Where wide cutting would be objectionable side-
walls and interior supports are built first in trenches, bore-holes or pits and roofed,
after which the core is removed, and the bottom, insulation and interior coating is
placed (Fig. 6/73). After the deck has hardened , the pavement may be replaced
and surface traffic re-established. (Additional savings in time and in manual
labour are obtainable when prefabricated deck elements are used .) The last
items, i.e. the removal of the core and the construction of the inner coating, etc.
constitute proper tunnelling elements (Fig. 6/74). Interference with street traffic
is, thus, restricted to a minimum, i.e. for the time required to place the deck,
and a ;easonable transfer of the enclosed public utility conduits is also made
possible (though it is more practical if these are all transferred beforehand).
A prerequisite for the application of this method is the existence of a resistant
soil layer underneath the temporary footing of the side walls, capable of bearing
both the roof loads and the moving load of surface traffic above without undue
subsidence. The economy of this method as opposed to the full width cut and
cover is afforded by the omission of formwork and staging for the roof and of
the extensive strutting of the sides across the wide trench.
A very economic and clever improvement of the same idea was first applied in
the construction of the Milan Subway (Italy), the Icos-VEDER method /- 14 compris-
ing the following elements.
In the first stage of the construction, as shown in Fig. 6/75, a 1 m deep and
1·50 m wide trench is excavated by a special trench excavator, bordered on both
sides by longitudinal reinforced-concrete walls 25 cm thick at the outer side of
the trench and 15 cm thick at the inner side. Then the trench is filled with a dense
bentonite suspension, and a 60 cm wide cut is deepened by the same excavator
from below the level of the bentonite suspension which is kept at a constant level
to afford an effective support to the earth faces within the trench. Trenches of

6
·u Cf. !(RU PINSKI, H. J.: Die U-Bahn in Mailand unter Anwendung eines neuartigen
Bauverfahrens, Bautechnik 1959 JO
670 CONSTRUCTION .\, D DESIGN O F TUNNELS

rlouses
iootpaih

&mn1 stratum

Bearing stratum

Bearm!J stratum

!?emov.1/ ofmuck in f?e!llforced- concrete

Jxial direction
frame construction

fleJrmg stratum

FIG. 6/73. Subway construction combined

with 'cut and cover'
tStockholm)
 CUT AND COVER METHOD 671

FIG. 6/74. Earth excavation under the protection

of roof built from
the surface

12 m depths can be safely excavated (detail I in Fig. 6/75) in this way. In the course
of dredging, care must be taken not to lift too much of the suspension with the
soil as the loss has to be made good. When a stretch of 5 m has been finished,
the prefabricated reinforcement skeleton is sunk into the trench still filled with
bentonite suspension, and the concrete is placed according to the rules of sub-
aqueous concreting. The bentonite suspension.., gradually squeezed out from the
trench by the concrete, will overfl.o\\' into the next trench section and secure exca-
vation (detail 2 in Fig. 6/75). ; (According to experience the suspension cannot
be re-used more than 3 or 4 times, because of the contaminating effect of concrete
and earth particles and because of the loss in its bentonite content due to sedi-
mentation and adhesion.) When the level of the side wall concrete reaches the
bottom of the roof, the soil above this level is excavated between the walls to make
room for the reinforced-concrete deck construction. Prior to its casting, however,
the side walls are strutted by a beam placed between the wall tops protecting
the uppermost part of the trench (detail 3 in Fig. 6/75). After the concrete has
hardened, the inner earth core is axially removed and the inner outfit of the tu~nel
completed (detail 4 in Fig. 6/75).

'-
 Lonq1t udma/ sect1017
-7 °'
......
N
I
I
I
~ ~ I
~~ I ~
'~ !..sztOOO
~

8z
AL__ ~
l?einfbrced-concrete :,Jl

{Jl'Otect1on wall 1-5 m

:Jeep 0
B
z
z>
0
0
/;l
ci
z
Section A-A ..,0
J 4
>-l
C:
z
z
"'!;;

h'ootconstruction
Frc . 6/75. An up-to-date
alternative to 'cut
and cover'.
The Milanese method
(lcos-VrnER)
 CAISSON SINKING. METHODS 673

6.31. TUNNELLING BY SINKING CAISSONS

Another practical tunnelling method at a shallow depth below the ground-

surface or under river beds is either to sink adjoii:i,ing tunnel units as caissons from
the surface or - in the case of subaqueous tunnels ~ by launching and sinking
prefabricated caissons in a previously dredged trench in the river bed.

63.11. Sinking Caissons in the Form of Working Chambers

Caissons used for this purpose may be constructed so that the working chamber
is arranged separately underneath the prefabricated caisson or subsequently
within a watertight enclosure of the established tunnel-section (see Fig. 6/76)

r --------, I r 5-50
------ -

I
FIG. 6/76. Scheme of the Wor king
chamber

.,
steel caisson used
for Paris
Metro !160

or the inner space of the working chamber itself is converted later to the tunnel
prop:r (see Figs 6/77 and 6/78.) Because of the higher elevation of such tunnels
resulting from a much smaller cover depth, the length of the · approaches can
be kept to ·a minimum.
The method js mainly used for the construction of subaqueous tunnels when
the river bed is composed of loose, permeable, silty layers to a consid,::rable depth
in which drainage by either direct pumping or compressed air would be impos-
sible because of the hazard of blowing out with the latter and of boiling up with
the former (cf. S::ction 63.24 and Fig. 6/ 147).
There is, actually, no difference between this construction method and the
compressed-air foundation methods, as the. caissons are assembled either on the

43 Szechy : The Art of Tunnelling

674 CONSTRUCTION AND DESIGN OF TUNNELS

Long1tudm,1/ sect10n

;,/ater level

-440 Stream /Jed level

-8-20
Outfa!,

Pian

Fm. 6/77. Longitudinal section and plan of southern sewer outfall tunnel in Budapest

FIG. 6/78. Scheme of caisson sinking

 CAISSON SINKING METHODS 675

shore, then floated and sunk between guiding stages under the river bed by pneu-
matic dewatering, or they are erected on temporary sinking trestles (Fig. 6/78),
and lowered from these in the same way. The only problem involved is the junction
of adjacent units when sunk into position. This may be effected either by ·sealing
the joints with subaqueous concrete or clay-fill poured into a temporary enclosure
made with sheet piles, or under the protection of a floating working chamber
(diving bell). When this external sealing is effected the bulkhead walls may be
broken through frotn the inside and internal insulation and outfitting completed
(cf. Fig. 6/81).

63.12. The Floating Caisson Method

Recently, vehicular and public utility tunnels have been constructed in increasing
numbers by prefabricating tunnel sections in dry docks or on launching ways,
floating them to the site (Fig. 6/ 79a-b) and sinking them into a dredged trench

a)

b)

FrG. 6/79. Launch ing of (a) prefabricated stell tunnel secti on for the Washburn tunnel ;
(b) of the r cinforced-conctete tunnel section fo r the Rotterdam tunnel

4 3*
676 CONSTRUCTION AND DESIGN OF TUNNELS

i
'·
~

i~

'\:::,~

~
~ -!:>,

""'~
<:::: ";j

i 1?'
~~
1§-~

,( \

~1 \'
I

1··,
-hI t
.

E
I
"'...
-0
-I ~ ~
\' I
0
0::

-=
"
E
-5"
"'
.2
 CAISSON SINK.ING METHODS 677

with a suitably prepared and levelled sand bed, then joining the sections together
into a _continuous tunnel and backfilling the trench (Fig. 6/80). The length of
the units varies from 30to 100 m (100-330 ft). ·At present 90 m (300 ft) is considered
an optimum length. The adjacent sections are joined as in the case of pneumatic
caissons. Sinking operations .are controlled by floatmg derricks or other means.
An example of this is shown in Fig. 6/81, as appliefl for the junction of the 65 m
(218 ft) long and 21 m ·(70 ft) wide reinforced-concrete units of the Rotterdam
Maas tunnel In this _pase a 1 m wide interstice was left between the units. First
the siqes of this inters\jce were closed bY. pushing two lock plates into· the vertical
grooves on the outer edges of the respective units, one of the plates being convex
outwards and the other being plane. Then the space between these plates was
filled with subaqueous concrete, and a diving bell; floated exactly over the site,

a;
Air or water 1311ks _ _

Driven steel sheeting

CJ Subaqueous concrete Temporarfl' bulkhead wolf _ /3/eral water-sealing p!vg

I-
I
.5eal
I dr_:
I J[:_
__-=_~ -- --=-: :i"!
I I
I I
I I
I
I I
I I
I I
I I
Driven dovble steel f(ori7ont3! lock Jnterspace dewatered by
sl;eet pile wall l'ompresserf air

FIG: 6/81 . Subaqueous connection of sunk units (Rotterdam)

678 CONSTRUCTION AND DESIGN OF TUNNELS

was seated on tne eages of the adjacent units. The working chamber was then
dewatere.d by compressed air and the deck construction (i.e. the gap) carefully
completed above the interstice and bound thoroughly to the depth of the immersed
units. Afterwards, an air lock was mounted upon the deck of one of the: units,
from which the bulkhead wall on the contact side was then broken through (d)
in Fig. 6/81) and the water excluded from the whole of the adjacent · unit. The
invert was placed monolithically connected to the side wall enclosures. Joints
were finally sealed by special grouting from inside the tunnel.
Another type of joint especially used with steel-tube tunnels is to apply circum-
ferencial watertight sleeves protruding over the end of the adjacent units, over-
lapping on all sides. The required connection can also be effected in this case
by compressed air, but directly from the inside of the adjacent tubes.
The internal pavements and coatings, ducts, etc. of the tunnels can all be pre-
pared in the prefabricated units when in the dry docks or when floating, with the
exception ofjoint sections. The uniformity and evenness of the subgrade of immers-
ed units can be secured by subsequent hydraulic filling underneath or by more rigid
end supports such as subaqueous bearing plates on pile groups, supporting pile-
bents (e.g, Ij tunnel, Amsterdam).
This tunnelling method is now in full development and will apparently soon
replace tunnel shields in the construction of many subaqueous tunnels. As further
examples, the Washburn6 · 15 (a steel tube tunnel), the Rotterdam, 6·16 the-Amsterdam.
(Velsen and lj), 6· 17 the Rendsburg6 ·1Sa and the Deas ·Island, (Vancouvei:)6·18b
tunnels (reinforced-concrete tubes) may be mentioned, and that a suggestion has
been made that the Channel tunnel should be constructed also by using this
method. 6 ·18c

6.32. SHIELD TUNNELLING

Whatever mining methods are applied in tunnel driving, certain drawbacks will
be :inherent to them.
I. The first, and as regards rock pressures, the most important is that the
construction of tup.nel lining can follow the excavation of the cavity only after
a certain lapse of time and, in the meantime, only temporary supports are
installed, admitting a certain amount of displacement" even with the best
workmanship, involving an extension of the loosened zone.
6 15
· Le Tunnel Washburn, O.ssature Metallique 1949 JO
0 6
·l FRITZLIN, M . <!:. : Le construction du tunnel sous la Meuse a Rotterdam. La Technique
des Travaux 1949 March
61
· l EGGINK, I. R. A . : De tunnels de Velsen, De lngenieur 1953 Sept.
De lj-tunnel. Werk in Uitvoering Dienst der Publieke Werken, Amsterdam 1961 Sept.
6 18
· • The Street Tunnel under the Kiel Canal at Rendsburg. Bull. of Wasser u. S chiffahrts-
direktion, Kiel 1961
8 15
· b PER HALL : The Deas Island Tunnel, Proc. AMSCE 1951 Nov. 1436
MB< LASSEN, M, and NJELSEN, C. N.: Projekt ti kanaltunnelen lngenioren Febr. 1962
 CAISSON SINKING METHODS 679

2. Much time passes· between the excavation of the face and the perfection of
the permanent lining which, in addition, is not always effected in a conti-
nuous course -along its whole perimeter, but with interrupt~ons leading to
intermediate settlements.
3. The excavation and walling in several parts slow down construction progress,
essentially disturbing the unity o.f the working area and rendering progress
altogether difficult.
4. A special drawback is the necessity for repeated installation and demolition
of temporary proppings.
5. The larger the tunnel and the smaller the strength of the rock, the stronger
and more densely spaced propping will be required leading to increased
slow-down of progr_ess, to difficulties in transport and to increased timber
co·n sumption.
These drawbacks are responsible for the trend to apply free-face tunnelling
methods in solid rocks and to extend its use to rocks of moderate strength by-
means of rock-bolting and by the reasonable combination of various steel-sup-
porting systems.
Difficulties show themselves to a greater degree in recent alluvial deposits,
i.e. in soils where a coherent united working area can be secured by the shield
method only.
The tunnel shield is a moving metal casing, which is driven in advance of the
permanent tunnel lining, to support the ground surrounding the tunnel-bore and
to.afford pq:,tection for construction of the permanent lining without any tempo-
rary support or timbering.
In fact, the shield is a rigid steel cylinder open at both '!nds, providing facilities
at its front for the excavation of the ground material and at its rear for.the erection
of the prefabricated lining. Thus, the shield is a1ways forced ahead by steps keeping
pace with the progress of excavation and erection work to the extent that the
excavated hole should be well supported until1the permanent lining is constructed.
A full cycle of shield-tunnelling comprises the following items (Fig. 6/82):
(a) excavation and temporary support of the front face at an appropriate depth
(b) advancing the shield, taking support on the pre_viously erected lining
(c) ·placing another course or ring of the permanent lining.

As the only face left unsupported during the working cycle is at the front, the
amount of advance and of unsupported face area must always be carefully adjusted
to actual soil and ground-water conditions.

Shield-tunnelling offers four essential advantages:

1. The tunnel section can be advanced with its full dimensions.
2. It offers a moving but constant support to the advanced tunnel.
68C CONSTRUCTION ANO DESIGN vi-' TUNNELS

t Cvtting edge 'Trvnk T81/

BJ

I· t+x
~~
Placed new r111g af'ter shove
{,)

==UL__ .J

Tout height of void left 3f'ter.s/;ove

Fm. 6/ 82. Principle of scheme of shield tunnelling

3. The omission of any temporary support is compensated for by the imme-

diate installation of permanent tunnel lining.
4. By speeding up construction work, it prevents the development of higher
rock loads.

In hard rocks where excavation must be effected by blasting, the shield tunnelling
mt:thod has to be -excluded partly because of the sensitivity of its inherent machin-
ery, partly because of the loss of its steerability.
 SHIELD TUNNELUNG 681

63.21. Structure and Dimensions of Tunnel Shields

. The principal element of the shield structure is the skin which is constructed
of steel plates, bent to the shape·of the tunnel section and slightly larger than it.
The skin may be divided into three main parts, differing in their' inner rigidity
and arrangement in accordance with their purpose.
1. The front end of the skin, where excavation is effected is heavily reinforced,
generally with steel castings to form. the cutting edge, its inner rigidity being
increased by stiffening rings. Its principal purpose is to facilitate the smoothest
possible advance and steerability of the shield skin by cutting the face, and to
provide for as uniform a distribution as possible of the mighty pressures induced
by its being forced ahead.
Its secondary task is to give an adequate shelter to the workmen engaged in
the excavation, through aff~rding a certain support for the front face.
2. The intermediate or trunk part is destined for the housing of pushing machin-
ery (hydraulic jacks, high-pressure pump installations, etc.).
3. The tail part of the shield is designed for the erection of lining segments,
i.e. for the building up of the tunnel lining.
4. In addition, some important supplementary elements are incorporated in the
interior of the shield mostly in combination with its stiffening elements, such as
working platforms mounted upon the horizontal and vertical stiffening posts
and girders, or front-support (face.) jacks mounted upon the division walls of
working-boxes, etc.
5. The shield construction proper must be supplemented by special machinery
for excavation, mucking, material conveyance, erection and grouting, all indispens-
able with shield-tunnelling but still independent of the shield-construction proper
(Fig. 6/83).
In the design of the main parts of a shield, the following principles must be
taken into account.
ad l. The cutting-edge, of which the primary purpose is to facilitatate the
penetration of the shield skin ahead into the ground-mass, while it has to over-
come earth resistance. In order to reduce this resistance, the previous excavation
of the front-face is desirable to the greatest possible extent. This cannot be
effected, however, in full as th~ steering and safe direction of the shield requires
that a ce_rtain area along the perimeter should be 'peeled off', too.
Namely, in a totally excavated · face the shield would loose its support and
would become liable to wobbling and torsion. It can keep its position and
a
direction only when it has certain ' hold' on the surrounding ground along the
perimeter of the cutting-edge. This hold is necessary not only because of its
·steerability (effected by the·exertion of unequal pressures by the jacks on advanc-
ing the shield) but a certain cut-off is a great asset against the intrusion of the
682 CO STRUCTlON A D DESIGN OF T UNNELS
 SHIELD TUNNELIN G 683

FrG. 6/ 84. Model

of a roof shield

surrounding loose grourfd into the interior of the shield. In particular, the layers
at the crown are liable to intrude through the front, and they are also the most
harmful as regards surface subsidence. This accounts for tile use of special roof
shields in loose ground with the top third of their perimeter projecting, by some
2-4 dm (8-16 in) ahead in proportion to the shield diameter (cf. face-stabilization
methods,· Section 63.221).
This hood or roof is actually a projecting part bf thC! cutting edge and struc-
it turally is built up entirely in the same way (cf. Figs. 6/84 and 6/85), with the sole
exception that its stiffening brackets need to be of heavier design correspond-
n_g to the increased projection. It is, however, more difficult to steer and to
keep in line a roof-shield and, therefore, it is usually of such a construction that
when the loose and weak layers have been passed over the hood can be dismantled.
An additional task of the cutting edge is to resist and to transmit the pushing
force of the jacks over the shield skin. This must be taken into account also in
the structural design, and stiffening brackets must be distributed opposite to
the seats of the hydraulic jacks (rams).
The section and shape of the cutting edge must comply partly with the require-
ments of cutting and partly with the transmission of jacking pressures.
This is best fulfilled by the unequal angle shape, with the projecting horizontal
leg providing for cutting, the rectangular vertical leg for the support of jack pres-
sures.
The cutting edge of smaller diameter shields may be constituted by the skin-plate
itself, when extended, pointed and stiffened by brackets accordingly. The pointed
surface may be artificially hardened (carbidization) or provided with a special
wear resistant coating (Fig. 6/86a).
684 CONSTRUCTION AND IJESIGN OF TUNNELS

f--' .... ~
i~~ .i?' t -~
~ §:: ~~ ~
i
rPr'~~ t~
·~
& ~~
<..::
c., I'.I
~ -;:::;
!i i?' ~~ ~~ n
[1= ~ Qi~ ~~ ~ di
.::j ::s~ <:::,
·--.. ~
::t=:

-0
Q)
:a
....0
VJ

...0
....
«S

0
C:
co
·;;;;
.,
Cl
 Counterw.r,k bolt 8J . __!fP_/J_ _ _
JOO -- -- - Cutting edge constituted by tile
760
+- · !ti«)

pointed skin-plate
I
Abras/Ofl-l'l?SIS·
tMt C'Oa/lllfl
Msct1111ed
/ /letractlflgJ8Ck
'~k ' ' ' 'l~~ ,r I .
+++I +

. '
5/

' ' .

I "&to b) --. - - '""·--····

/l(J!J

Cutting edges wit/7 cl7angea/Jle

_ St1flen111g IYng N'/ St1ffer,/r,g rir,g N'Jl steel-casting points
~
·~ "':5
"'r
0

CJ -- - d;
Steel cast111g
-'I
C:
z
Cutting edge o/'steel casting I IA z
"'rr
~
~
--] ~
--~

Section 1-7 iA
Section A-A

~85
~ " Lo11g1tudi11al
stiffening rib

F10. 6/86. Cutting edges (a- d)

°'
00
VI

I
686 CONSTRUCTION ANO DESIGN OF TUNNELS

In earlier shields the cutting edge was constructed as a riveted or welded steel
structure reinforced by a cast-steel inlay along the perimeter of its extreme edge
to offer a greater resistance to wear due to the quartz content of the ground.
(The effective design of a sufficiently resistant structural bond of this inlay to the
steel-structure proper is of major importance.) At present there is a tendency to
form the cutting edge entirely of cast-steel blocks. These blocks have to be joined
to the skin very carefully. The usual method is to give an immediate support to
the vertical leg of the casting by a stiffening ring, which in turn must be supported
by rather densely spaced longitudinal ribs acting also for the transmission of
stresses to the skin.
The support offered by, and the connection to, the inner stiffening ring must
be resistant enough to prevent both the outward and inward turnouts of the cutting
edge. As the outward turnout is greatly encouraged by the inclined face of the
cutting edge, sections with rather thin bare projecting legs have been employed,
recently, with stiffening brackets instead of the former nearly full edge shapes
(Fig. 6/86d). Inner turnout may be encountered when the cutting edge is con-
fronted with some hard rock boulders or the like. Moments of turnout are the
bigger th6 longer the projecting horizontal leg, therefore this hazard may be
lessened by reducing its length.
The influence exerted upon the advance and the stressing of a shield by the inclination of
the cutting-edge point may be best illustrated by the experience gained with the first shield
of -the new Budapest Underground . In this case the lower cutting-edge inlays were mounted
with a slight inward inclination (1 : 20) at the bottom quarter in order to anticipate progressive
sinking (wobbling) of the shield into the previously assumed soft bottom layer. The layer
was in fact not as compressible as assumed and the· shield, after being forced only a few metres
ahead suffered a vertical diametrical deformation of 300 mm (12 in) when it had to be stopped.
After the removal of the cutting-edge point insets mentioned and replacing them with parallel
faced ones, similar to those applied along the whole circumference, the work was able to
proceed without any further deformation. 6 · 19

It sometimes happens that insets have their points rising outwards in order to
reduce friction resistance through progressive loosening. This arrangement, how-
ever, may tear off the connections and lead to a turnout.
The same care and foresight which is necessary in choosing the wedge-shape and
inclination of the cutting edge is necessary in the choice of the measure of its projec-
tion relative to the shield skin, which is in correlation with the magnitude of actbg
earth pressures and through that with the pushing force required for advancement.
Whereas a cutting-edge shape outwardly inclined and slightly (1- 2 cm) in excess
of the shield diameter is advisable in hard grounds (cf. Fig. 6/86a) producing
a slight annular space around the skin incurring a certain loosening of the sur-
rounding material and, thus, involving a certain decrease in the acting earth
pressure, such a shape is unsatisfactory in weak ground because the annular void
initiates a progressive loosening-action leading to an increase of rock pressures.
6 19
· Similar distortions were reported by PrnRIE from the construction of the Toronto sub-
way cf. Proc. Inst. Civ. Eng. 1966 May, 71.
~IIIELD TUNNELLING 687

~
"--

.,
"O

:2
"'
,.._
....
Q)
~
~
....
2
.;
E
"'
~
'-
0
,::
.2
t,
Q)

.;"'
,::
:.;:;
3
"ii
,::
,:i
....l
,-.:
<D
\O'
ci
ii;
 688 CONSTRUCTION AND DESIGN OF TUNNELS

ad 2. The following shield section, the trunk, is attached to the cutting edge.
It is primarily a housing for 'the hydraulic equipment necessary for the advance-
ment of the spield. As the shield is sliding on t4e outside surface of the steel skin
ahead, all possible projections and sockets should be avoided there and it should
be kept as smooth and even as. possible. The thickness of the skin varies between
15 mm and 70 mm (5/8 in and 2·5 in). Thicker skins are multi-layered, connected
with alternately overlapping splices either riveted or welded. Riveting is rarely
used now, but if it is outside, cover plates are to be avoided and counter-sunk
rivets are to be used. Any projecting heads are quickly worn off by the soil in
addition to involving a considerable and superfluous increase in frictional resis-
tance and to increasing the extent of the back-space. For the ·same reason an
outside gradiation of the skin-plate;5 is also to be avoided.
The trunk part of the skin can be stiffened in several ways. For example, in the
Soviet Union where shields are destined for constantly repeated and long-term
use the skin is composed of cast-steel elements provided with stiffening ribs
(cf. Fig. 6/85). Another solution, the simply-ribbed mild-steel skin plate as custo-
mary for smaller diameter shields is shown in Fig. 6/87. The skin of larger-diameter
shields is usually stiffened by ring-shaped riveted or welded inner diaphragm
girders of I-sections. The stiffened girders are spaced at 0·60-1 ·80 m (2-6 ft)
unequal distances varying in accordance with the inner arrangement and with
the acting forces . Although their number and dimensions depend primarily on
shield dimensions their inner arrangement is also a function of the type, dimen-
sions and operating machinery of the hydraulic jacks employed. For example,
with double-acting jacks, they are usually placed between two stiffening rings,
whereas with single-actingjacks, where additional retracting jacks are also operated,
three rings are applied as a rule. The first ring does not have to resist the stresses
aroused from pushing the shield ahead, .but is merely to provide a safe support
for the cutting edge. The main advancement jacks (rams) are seated upon the second
ring, whereas retracting jacks have their anchorage in the third. The interstices
between the stiffening girders in the bottom _third · of the shield are covered and
used as an operation platform. Large-diame_ter shields cannot be satisfactorily
stiffened merely by rings, but must be additionally braced by vertical posts and
horizontal beams so that the whole area is divided into separate working boxes
(see Figs 6/85 and 6/88), where se_veral gangs can do excavatiop. work simultane-
ously (see Fig. 6/84). The boxes should have a minimum width of I ·20 m
(4 ft) and a minimum height of 1-80 m (6 ft). The working pockets of the large
shields are generally equipped with sliding platforms _which are operated by auxili-
a ry hydraulic jacks and which can be extended to the face of the excavation and
used riot only for the workmen to, stand upon, but also to form a support for
the face. When the shield is advanced, the platform- and face- -supporting auxiliary
jacks are slackened, permitting the jacks to recede as the shield moves forward
and, at the same time, maintaining a constant pressure on the face supports.
In the extreme case of liquid ground the stiffening ring next to the cutting edge
 SHIELD TUNNELLTNG 689

FIG. 6/88. 8·5 m dia shield, stiffened with division walls, under erection
in the shield chamber

may be developed to a perfectly closed dia,phragm wall (bulkhead) perforated with

sluice-gates for material transport or communication only (see Section 63.221).

Stiffening members must be dimensioned for tl}e most unfavourable loading conditions,
represented by the action of full overburden weight from above counteracted by an equal
bottom reaction in the case when no lateral support is afforded owing to an excessive o ver-
break on the sides. For this loading assumption we obtain, after FoRCHHEIMER (cf. Sectio n
43.22), a bending moment acting upon the ring-wall:
.,
pr-
M = - 4 cos2"-
'f',

and an axial force :

N = pr sin 2 </>.

Maximum bending moment at the crown and springing:

pr•
M,nax = ± 4·

Maximum axial force at the springing N"'"' = pr and minimum axial force at the crown
N"''" = 0 .
44 Szechy: The Art of Tunnelling
 690 CONSTRUCTION AND DESIGN OF TUNNC:LS

The radial deformation of the ring:

l pr 4
o = -12 -EJ- cos 2tp '

and the shortening of the vertical diameter :

1 pr 4
Lld = 20max = --
6 EJ

wherep = the unit vertical rock pressure taken as equal to the overburden pressure in general,
because only quite exceptional constructional circumstances and ground
characteristics could justify the consideration of arching.

Powerful hydraulic jacks ( rams) are placed in the central part of the shield be- .
tween the stiffening girders to force it ahead. They are located just inside the skin
in order to bring their axis closest to the centre of gravity of the lining segments
and to that of the cutting edge to reduce eccentricity and bending inoments to
a minimum. They are spaced uniformly around the perimeter and are housed in
openings called jack ports in the transverse stiffening girders (rings). They con-
stitute the most important elements of shield operation, and bear on jack seats
on the web of the forward stiffenings girders (behind the cutting-edge structure)
and are supported on the placed lining s·egments (see Fig. 6/82). To effect ad-
vancement they have to overcome the following resistances:
external friction between sqield
skin and surrounding ground
(WJ;
inner friction between tail-skin
F//1!1
and placed lining segments
H!!dravltc ram (W2);
(Jack)
passive earth resistance against
the intruding surface of the
FIG. 6/ 89. Resistance cutting edge (W3);
forces acting against propulsion partial resistance of the free
face depending on whether it
is supported by face-jacks or
some other kind of yielding bracing, or bulkheaded during advancement (W4).

The determination of the magnitude of these resistances in advance is very

difficult, chiefly because of uncertainties about the resistances on the face, which,
however, are usually preponderant compared with the frictional resistances on the
skin.
Approximate formulae for them may be found in the Soviet literature as follows
(Fig. 6/89):
 SHIELD iUNNELLiNd- 691

where Py = the vertical rock pressure

Px = the horizontal rock pressure
L the length of shield
D = the diameter of shiefd
GP = the weigh( of shield
/ 1 = friction coefficient between shield-skin and ground.

the weight of lining segments erected in the. shield-tail

the friction coefficient between skin and segments.
(6.10)

where Dk = the shield diameter measured· in the centre line of the cutting edge
[J the thickness of the cutting edge
).P = the coefficient of passive earth resista nce.

For pY).P SAMOYLOv6·19" gives a correlation based upon the rupture theories defin-
ing the ultimate bearing capacity of foundation strips (qu)

\Vhere k = a shape factor

¢ = the angle of inner friction.

Finally
W 1 = pylF,

where F = the face area supported by the bulkhead or by the face bracing
). = the ratio of vertical and horizontal. pressures, depending upon the
process of loosening or compression involved and ranging between
the active and passive earth pressure coefficients.
19
(An average value of pY)'P may be taken as 60 t/ m = 850 lb/in .)6- b
2 2

The number and capacity of jacks are to be .assumed with a safety factor of
1· 5-2·0 as obtained from the above equations. This is required not only because
of the incertitude of the above computation al estimates and the necessity for

s.,9a SAMOYLOV V. P. : Usilia vosnykayushtshie .. . truboprovodo v (Earth resistance to be

overcome at the propulsion of shields and with pipe-jacking). Sanilarnaya Tehnikal957 10.
· b Cf. ILLESY, J.: A pajzsos alagutepites ktili:ini:is tekintettel a pajzsok
6 19 tervezesere. (Shield
tunnelling with special reference to shield design), M ernoki Tov. Kepzo 1954 2101
692 CO!'ISTR UCTJON AND DESIGN OF TUNNELS

operatio nal safety, but also to provide for safe steerability which is alwayc, affected
by overstressing during the advance ment of one group of jacks on one side
and
releasing the opposite group simultaneously on the other side.
Shield jacks have to perform two movements, namely by the propuls ion
of
their pistons when they have to advance the shield and afterwar ds when they
have to
provide for the necessary space for the erection of the following ring of
lining
segments which is effected by subsequently retractin g the jack-cylinders
to the
pistons. In practice, ·this double operatio n may be .effected either by double-a
ctin~
reversible jacks or by the combina tion of single-acting and separate retractio
n
jacks.
The advanta ge of double-a cting jacks is that the fluid pressure in the cyli
nder
acts on the full piston (plunger) surface when in propuls ion, and after entering

Jack pot

!50 J55 1650

,---- - - -
- - - - - - _: _=::-:_~-_:-_=--.,I

~ --
~
--

2005
FIG. 6/90. Double-a cting jacks
 SHIELD TUNNfcLLlNG 69l

the cylinder, will exert only a reducect effect upon the smalf annula:r surfiee fefe
between the cylinder wall and piston (plunger) rod, but one which. is still sufficient,
to push back the piston proper (Fig. 6/90). In any case, the inner surface of the
cylinder must be smoothly polished with the greatest care which involves con-
siderable costs.

Guide bush (pivot ring)

r Ram

FIG. 6/91. Single-acting jack

As opposed to this, single-acting jacks reqmre much more simple machining,

because the careful polishing can be restricted to two small inlaid rings which,
however, can also be inserted subsequently (Fig. 6/91). On the other hand, pistom
(plungers) after propulsion can only be withdrawn by special devices. This is
most usually done by employing special retracting jacks or auxiliary pistons
(plungers) for this purpose. Retracting jacks are small-capacity jacks with a
stroke equal to that of the main advancement jacks but with an opposite move-
ment. They are usually placed in line with the main jacks in order to avoid undue
lengthening of the shield and drag the jack-shoes forward by means of traction
rods through the intervention of a pressure distribt.ition ring. The disadvantage
of this arrangement is shown not only by the necessity of applying special devices
for this purpose, but it incurs a definite increase in the shield length . Various
auxiliary-piston devices incorporate secondary piston machinery built into the
main piston body. This, of course, will not require any surplus lengthening of
the shield-skin but involves more delicate and more expensive machinery.
The jacks must be supported very carefully. Cast-steel seats are placed for the
bearing of the front and on the cutting edge (cf. Fig. 6/90), whereas support at
the rear on the lining segments is effected either by direct bearing on them by
means of specially formed cylinder shoes or by special pressure distribution rings
inserted between these shoes and the segments. Tt is a general requirement for
the transmission both of propulsion forces to lining segments and tension stresses
therefrom to shield-skin that this should be effected with the smallest possible
eccentricity. Therefore, both the jack shoes and the pressure distribution rings
are asymmetrical in section so that pressure should be transmitted nearest to the
centre of gravity of the lining segments. Such a pressure ring of asymmetrical
U-section is shown in Fig. 6/92.
The second group of shield jacks includes face jacks, which are intended to
afford a steady support for the excavation face independent of the advance of
694 CONSTRUCTION AND DESIGN OF TUNNELS

~ Ram
"" pressure
~

FIG. 6/92. Section of pressure-distribution ring

the shield. These are attached to the inner transverse platforms and vertical
division walls in such a position that the plungers when in the protruded position
extend t o the outside plane of the cutting edge there to afford an intermediate
support to the breast-board planks. Face jacks are single acting and their stroke
should be equal, at least, to the width of one lining segment ring. When mounted
upon sliding platforms operated between the division waits, this operation length
must also be taken into account when choosing their stroke and length. They are
usually 1<1cated in the foremost part of the shield body in the immediate vicinity
of the cutting edge (d Figs 6/ 84 and 6/85). When shield advancement is finished
and at the Very start of a new cycle the face jacks have to support the face with
r~pelled plungers while the fluid is extruded from the cylinders to be recuperated
in turn into the system again. During the commenced new 'push' the plungers
again protrude.
The jacks are operated hydraulically and connected to central distributor blocks
and controlled through a battery of valves under the control of the operator,
so that he may advance or withdraw any one or any combination as required.
The block is a steel casting provided with bore holes having a valve-controlled
outlet for each jack and a central inlet bore with a conduit to the high-pressure
hydraulic pump.
As the advance jacks have to exert pushing forces up to several hundred tons
each, very high fluid (water or oil) pressures (250-500 atm) ·have to be applied
and thick wall (10-15 mm) conduit-tubes must be med from pumps to jacks.
The pumps themselves may be eventually located apart from the shield in the
completed tunnel section whereas distributor blocks and feeding conduits are
always located in the trunk of the shield.
The operation of face jacks does not require the exertion of forces of similar
magnitude, therefore it is usual to feed them separately by a pump distributor
block and pipe network system at a lower pressure (30- 50 atm) which is suitable,
 SHIELD TUNNELLING o::,5

.. ~ the same time, for manoeuvring the sliding platforms, installed eventually
between the stiffening members (beams and posts) of the larger diameter shields.
ad 3. The most delicate section of the shield skin, apart from its joint to the
cutting edge, is its third section, the tail. This is mainly exposed to deformation
as the installation of any inner stiffening is impossible because of the lining seg-
ments placed there.
Its free length is determinated by the width of the lining rings applied and it is
considered that at least an overlap of l 1/2 times this width should be provided for.
In addition the width of an eventual pressure-distribution ring must also be added
and provision has to be made for the removal and replacement of shield jacks
when necessary.

Inside new o! sl!ieltf Pressu~ dlstri6ulio11 anti cor,lro/ block w1U1 v;lve, anil

Fie;. 6/ 93. Arrangement of pressure distribution blocks inside the shield

Certain authors6 ·20 hold the view that the free tail length should be enough
to cover the width of two entire lining rings and to afford, in addition, an overlap
of a few centimetres for the third ring. This provision should make it possible
for the segments of the first or second ring to be replaced by new elements while
still within the shield, should they be damaged under the heavy jack pressure in
the course of advancement. The trend to increase the width of the rings in order
to lessen the number of leakage spots, i.e. the joints, is also in favour of longer
tail sections.
On the other hand, it must not be disregarded that not only can the unduly
long and unsupported tail skin be easily distorted, but through the increase of
the total skin length involved higher jack pressures will be required for the advance-
ment and a long shield loses very much of its steerability. In case of absolute
necessity an increase of the skin thickness must also be considered, but this can

6·20 SINGSTAD, 0.: Tunnels, American Civil Engineering Practice 1 10-24

696 CONSTRUCTION AND DESIGN OF TUNNEi.S

be effected only inwards to prevent any unduly high rise ~n sliding resistance.
The increased thickness will also result in an increase of the hollow annular space
left behind after the advancement (cf. the measure 'A' in Fig. 6/82) leading to
greater loosening of the overlying strata. In this respect, the greater rigidity of the
middle (trunk) section may involve a certain stress relief for the more flexible
tail section, by taking over a bigger portion of the loads.
It is customary to make .a bottom cut in the s~<in plate of the tail section cor-
responding in length to the width of the segment rings and with a central angle
of 90° to 120° in order to prevent a sudden fall of the lining from the invert face
of the shield upon the ground surface left behind at the advancement.
ad 4. As special structures built in the shields the division walls, working plat-
forms, face supports and bulkheads must be mentioned.
Division walls and working platforms usually divide the larger shields into
working boxes and consist of stiffened steel plates connected to the skin plate
by angles or by welding along their full length. The web plates of the vertical
division walls are led without interruption or splicing from the inside top to the
bottom of the s~in plate and it is the horizontal platform plates which are inter-
rupted and connected to the former by angles.
Working platforms usually slide on the horizontal division members and are
operated by the hydraulic system mentioned above. These platforms can be
advanced slightly beyond the front plane of the cutting edge, thus affording not
only excavation facilities, but also partial edge supports to the face. The extension
measu re of the sliding p latforms may have a terrace-like
arrangement mainly for l::irge diameter roof shields working ' Cuttiag ed_qe_
in very soft ground (Fig. 6/94).
Face bulkheads are composed of a steel-joist grid cover-
Bulkhead
ed with double steel plate sheeting and provided with inlet wall
openings. These openings are controlled by sluice gates, /

PrQJecting 1
which are usually opened when the shield is advanced to workmg } ............._~..
admit the entrance of the highly plastic material. The front platform i
bulkhead is stiffened a nd supported by the division walls
and beams which in this case extend to the front plane of
Jlood f
the cutting edge though under normal conditions they ~'
1/,
117/et gates ~
only reach its rear plane. (Effective support of the front
face can be obtained by various mecha nical excavation im- (sluices) :;,:;
plements installed, recently, in the shields ; see below un-
der Section 63.22.)
When the above structural elements and their dimen- OirectiOfl
- ofs/Jove
sions have been decided upon, we may proceed with the
determination of the main shield dimensions, i.e. its dia-
meter and length.
Shields have, in general, a circular shape, as this shape
affords the best resistance to outside pressures, prevailing
in recent alluvial deposits where shield tunnelling methods FIG . 6/94. Operation
 SBIE.LD TUNNELLING 697

are nio'stiy" used. (The pre-ssure conditions in these loose and soft, heavily water-
logged layers may be characterized by the approximate equality of vertical and
horizontal pressures and by their uniform distribution aro und the perimeter.)
In addition, this shape is most suitable for forming easy and exact bolted joints
between consecutive rings. It is a common experience for shields to show a rotating
tendency around their longitudinal axis as they are advanced. This is largely due
to oblique stratification, to asymmetrical external pressures and to non-uniform
excavation at the face. If the shields were oval or rectangular in cross-section
the rotating movement a:·ound a rigorously followed longitudinal axis would lead
to gradual mis-shaping cf the section and an increase in stresses and might even
niake the longitudinal bolted connection of successive rings impossible.
At first, shields of oval and rectangular cross-sections were tested, e.g. the first
shield of BRUNEL was of this type, but it was soon abandoned by GREATHEAD,
because in addition to the drawbacks mentioned above "it was observed that
a bigger pushing force is required for their advancement than for circular shields.
A bigger arching action will take place above circular shields leading to a decrease
in rock pressure and, consequently, in frictional resistance.
Occasionally, arch-like circular segment shields are also used, mainly in the
construction of large underground structures where side walls and springing
Supports have been previously built up either in suitable mining drifts, e.g. German
method, or in the form of smaller diameter tubes. Sometimes a bedrock boundary
lying at an appropriate level . may also favourably constitute such a support.
Cast-irofl lmmg segme1Jts
½ ,· ;/4 ,

.. !
698 CONSTRUCTI0:-1 AND DESIGN OF TUNNELS

These half shields or open shields may undergo considerable mis-shaping owing
to their horizontal deformation. Therefore, they have to be supplied with hori-
zontal ties at their bottom ends and can resist o_nly a relatively small vertical load
and so are suitable only for small depths.
The diameter of the shield is rather definitely determined by the clearance
requirements of the tunnel proper. As all operations (excavation, mucking, trans-
port, erection) must be done and all mechanical equipment (jacks, pressure pumps,
platforms and conduits, erectors, loading machines, etc.) must be installed within
the limited inner space of the shield, a reasonable economical utilization of this
space - chie:1-y with smaller diameter shields - is of primary importance. In the
first instance, the reduction of the number of jacks, the removal of pressure pumps
from the shield and the omission of working platforms, division walls, etc. must
be considered.
The choice of a shield of suitable length is a basic problem. It can be seen from
the above considerations that this is mainly determined by the dimensions of the .
jacks and of the lining segments. The operating conditions of the shield are mainly
affected by the relati ve length, i.e. by the shield diameter compared with shield
length. This ratio governs the steerability, mo bility and the steadiness of its
directio n. The shorter the shield, the more difficult it is to keep it in the correct
line and the easier a change of its direction, e.g. on curves, will be, whereas
the longer the relative length of the shield , the easier it is to keep it in its original
directio n, but the mo re difficult to bring it back from an accidental fincorrect
direction and, in addition, the more difficult is its advancement. Therefore, definite
advantages may be attributed to longer shields in loose and soft ground and in
relatively long straight stretches, but in more resistant ground and in curved
(sinuous) stretches shorter shields will afford greater advantages.
This relative length will become disadvantageous mainly with smaller diameter
shields as the installation of indispensable shield machinery requires a certain
space which cannot be reduced. Ai, the steerability of smaller diameter shields is
rather poor, an attempt has been made to insert an articulated transverse joint
in the middle of the skin. The longitudinal axis of the shield may then be altered
by adjustable bolts, applied in the joints and, thus, it can be brought back to the
correct position.
The ratio of relative length varies between 0·4 < L/D < l ·4 but, according to
present considerations, should not, in any case, exceed 0·70-0·75. The relative
length and some other important data of shields employed in some noteworthy
shield tunnels are indicated in Table 6/VIII.
RICHARDSON and MAYO give the following formula for the approximate steel
weight of a tunnel shield:
W = IS(D - 10), (6.11)

where W = the weight of shield in tons

D = its external diameter in feet.
 SHIELD TUNNELLING 699

TABLE 6/ VII[
Comparative Data of Shields

Dia· Le ngth Number Thickness of

Total Weight I
L pres• of
Name of tunnel meter L of skin plate R emark
D sure shield
D (m) (m) rams (mm) (t) (t)

Antwerpen 9·50 5·50 0·576 32 70 6400 275 Cast steel

London City 3-86 2·13 0·55 6 186 -
Rotherhite 9·35 5·49 0·586 40 6700 -
Moscow Metro 9·50 4·73 0·50 36 3500 340 Cast steel
Moscow Metro 6·00 4·97 0·83 24 1800 120 Cast steel
Budapest FAV 5·50 6·00 J ·09 20 20 (50) 2500 80
Budapest FAV 8·50 6·50 0·764 24 30 (60) 6000 180
Queens- Midtown 9·65 5·70 0·59 28 5600
Mersey (segment) 3·12 3·52 1·13 10 19 770 -
Blackwall 8·61 5·78 0·67 34 63 5785 224
Holland Vehicular 9·17 5·73 0·63 30 70 6000 400
Detroit Dearborn 6·54 4·26 I o·65 20 55 3120 -
Lincoln 9·63 4·7 1 0·49 28 63 + ]2·7 6440 304
in the tail
section
Chicago
underground 7·68 5·93 0·77 24 4800 225
Brooklyn Battery 9·63 4·71 0·49 28 63 + 12'7 6440 315
in the tail
section
I
Actual weight figures are also given in the above table and, as shown, shields
constructed of steel castings have about twice the weight of those made of normal
structural steel.

63.22. The Main Working Procedures of Shield Tunnelling

Under the protection of a shield the following working procedures have to
.take place:
(a) excavation (63 .221)
(b) mucking (haulage) (63.222)
(c) shield advancement (63.223)
'(d) erection of tunnel-lining (63.224)
(e) grouting, caulking and drainage (63.225).
 700 CONSTR'U CTroN AND DESICN OF TUNNELS

63.221. Excavation. This takes place at the front-face ana: as regards

security,
it is the most difficult part of shield tunnelling. Shield tunnell ing
necessitates the
excavation of the full-face, which in turn is not an easy task conside
ring that
shields are used in less resistan t ground and the stabilit y of face
areas exceedi ng
50 m 2 is frequently require d.
The face may be stabiliz ed and suppor ted by the .following
means:
(a) The applica tion of a roof shield (cf. Section 63.21) transfers
the plane of
action of horizon tal pressures (ph) responsible for the intrusio n of
ground masses
toward s the inside of the face by the protfud ing length (x) of the
hood. Throug h

9406
Previously inst3//ed
/,emporary posts Columnar type st3tion
( Moscow}

Wall oe,;m lbrs/rdtnfl

FiG. 6/95. Sketch and view of segment -shaped shidld for columna
r type station (Moscow )
 SHIELD TUNNELLING 701

FIG. 6/96. Face stabilization by means of the Joosten silicatization process

this, the acting external pressures are partly resisted by this earth wall, viz. the
pressures acting on the original face will be reduced in proportion to the weight
of the earth mass loading upon the face. Horizontal pressures (Ph) are produced
by the vertical pressures (p,,) and the corresponding weight is supported in this
case directly by the hood (cf. Fig. 6/97a).
Roof shields also offer advantages in loose ground when the face is breasted
prior to each push, in that the breast support may be installed at such a distance
ahead that the push length can correspond to ·the full width of a lining segment
ring.
(b) The face can be supported by breast-boards either in full or in skeleton
arrangement. Breast-boards are to be supported from division walls or from stiffen-
ing rings either by direct struts or by face jacks. In the former system tedious and
sometimes_hazardous dismantling is necessitated.
(c) A very advantageous supporting effect is obtained by the application of
compressed-air dewatering. This has a double effect because, besides affording
a uniformly distributed supporting pressure on the face, it also stabilizes loose
and plastic soils by expelling water from their voids so raising their shear strength.
(d) Similar and still more definite stabilization may be obtained by the arti-
ficial solidification of the face (cf. Section 63.24), which may be accomplished
directly or indirectly. Direct stabilization as accomplished by injection pipes
driven ahead in a fan-like arrangement and with an outward inclination around
the perimeter of the face is a rather cumbersome procedure requiring continuous
alternation of excavation and solidification cycles, involving the time delay of
tube driving, during grouting and solidification periods. Preference is given,
therefore, to indirect stabilization, when artificial solidification is effected from
the surface or from a special upper drift in order to raise the strength of strata
lying just above the crown, in this way obtaining a corresponding release of hori-
zontal pressures acting upon the face (cf. Section 63.24).
(e) An original method of face stabilization is constituted by its perforation
by holes and anchorage adits driven ahead into the undisturbed ground. This
 ,

_,
C
N

a) b)
Unsupported /:9ce-area witlt
/Iv an anchorage drift (l
0
z
"'
>-i
:,,
Pn = tYPvJ c::
I
g
I 0
z
)>
D 0 z
C
X t,
rn
ci
d z
..,0
>-i

z~
"---- Reduction of pressure Area supported Anchorage-drift
Reduction of pressure directly by friction
diagram acting upon lace diagram "'[;;
to the circumference
of drift

FIG. 6/97. Reduction of earth pressure on shield face (a) with a roof shield;
(b) with an anchorage drift
 SHIELD TUNNELLING 703

procedure reduces the dimensions of the unsupported face area by affording inter-
.mediate supports along the perimeter of these adits. These drifts will be anchored
by frictional. forces into the ground and will be capable of resisting considerable
tensile reactions in proportion to their length and surface area.
In addition, they reduce earth pressure by dividing the earth column at the face
into smaller parts, by taking over and transmitting directly a part of its weight
to the bottom. By shortening the length of the sliding surface, they also reduce
the dimensions of the sliding wedge (Fig. 6/97).
This concept was extended by E. KINTLI and J. FABIAN during the construction
of the Budapest Subway to multiple adit anchorages on the temporary front-walls
of a spacious shield erection chamber. Three anchorage adits were applied here: one
at the bottom and two somewhat above the spring-line; they all consisted of bolted

Section 8-8 Section A-A

Ancliorage liner-plate drift

FIG. 6/98. Application of anchorage drifts for the reduction of temporary bulkhead wall
thickness of shield chambers (Budapest Subway)

liner-plates with a diameter of 2· 50-3·0 m (8- 10 ft) and were driven according
to the system described in Section 62.41. Long-term anchorage adits were also
grouted to prevent excessive loosening and . rock pressure in the surroundings.
In the construction practice at the Moscow Subway such advanced bottom
drifts are also utilized for material haulage when they are lengthened to the next
vertical operation (ventilation) shaft.
A further advantage is afforded by advanced bottom drifts in · soft ground,
where they constit~te a firm bottom support and tracking for the advancing
shield and successfully counteract its wobbling tendency. The concrete invert of
this bottom drift is shaped in the form of the shield perimeter and the two rails
embedded reduce friction and fix the required straight direction. (In loose and
soft ground it is general experience that the front of the shield is impressed more
deeply into the ground than the tail section, because breakdowns and loosenings
lead to increased rock pressures at the face, whereas equilibrium conditions and
arching action have been re-established to a certain extcn-~ ~t the tail.)
704 CONSTRUCTION AND DESIGN OF TUNNELS

(f) Finally, in entirely plastic liquid soils none of the above methods could
provide suppcrt for the face. Under such conditions the whole face has to be
closed by a steel bulkhead through which the liquid soil enters the shield interior
through slots controlled by sluice gates (see Fig. 6/94). The bulkhead wall is usually
arranged in . the inner plane of the cutting edge and each working chamber is
provided with a separ_a te slot. Depending on the nature and compressibility of

Ortlling°rOtf
/ /Cutting edge ProtJuls1on ram J7nmg s_egment

Advance dnft for Working platform Travelling erector Muck car and track
/Jau/age and face arm and groutin.tl
stab1/isati011 p/atl'orm
FIG . 6/ 99. D ri ft ad vanced beyond face in shield t unel (Moscow)

the ground the shield may be driven partially or totally blind, i.e. a certain amount
of material is permitted to flow into the shield through the openings and the rest
is pushed aside as the shield moves forward, or in perfectly liquid material it may
happen that no material is allowed to enter the shield which is then pushed
' blind'. For example, in driving the south tube of the Lincoln tunnel, New York,
in Hudson river silt about 20 % of the total shield displacement entered the
inside through two small"inlet-slots making up about 0:5 % of the total ' blinded'
area. In driving large diameter subaqueous tunnels by this method, enough
material has to be allowed .to enter the tunnel to serve as ballast to counteract
 SHIELD TUNNELLING 705

the rising tendency ~f the lining owing to the uplift of the liquid. In other cases
even more material than necessary for ballast may be permitted to enter in order
to minimize the lateral pressures which might be set up against adjacent structures,
as was the case when driving the second tube of the Lincoln tunnel while the
adjacent tube was already in operation. Again, in constructing some parts of the
Chicago subways by the shield method in highly plastic clay under city streets
adjacent to heavy important buildings, where the control of heaving, or subsidence
was of the utmost importance these were kept at a minimum by adjusting the size
of the bulkhead slots to the varying soil conditions. Almost 100% o'fthe material
was admitted into the shield through openings 5-20 % of the total face area.

Fm. 6/100. Soil displacement brought about by the propulsion of a bulkheaded shield
(lateral squeeze and entrance into shield)

The admission slots are controlled by hydraulically operated sluice-gates

(Fig. 6/100). It may be mentioned, that the rate of advance of 'blind' shields in
such liquid and loose gro,und is rather favourable. An average daily progress of
9 m (30 ft) has been obtained, when in excavated ground it did not exceed 2·2 m
(7 ft 4 in) and in rock was only 1 m (3 ft 4 in).
(g) The next step in the development of ' blind push' was to close the face by
mechanical excavators, i.e. to do excavation work with some device providing
support for the face and preventing its breakdown at the same time.
(i) PRICE's rotary excavator may first be mentioned among the early devices.
This was used for smaller diameter tunnels in ground of medium strength. 6·20a The
excavator head consisting of three cross members of girder construction is inside the
6 20
· • Cf. The Engineer Aug. 1926. 178

45 Szechy: The Art of Tunnelling

 706 CON STRUCTION AND DESIGN OF TUNNELS

Front view Rear view

FIG . 6/ 101. Rotating-bucket excavator of FRlCE

shield's skin. Steel cutters are attached to the cross-members in varying positions
and in the centre there is a larger cutter of special shape. The head is mounted
on a central shaft which is supported by the girder seen in the centre of the rear
view (Fig. 6/101.). The internal rack wheel bolted to the excavator head is driven
through gearing by an electric motor. The buckets which scoop up the material
ex9avated by the cutters, which is then dropped into a chute are mounted on the
revolving cross-members. From the chute the material falls into a container,
whence it is conveyed by an endless belt and dropped into skips which take it
away.
The energy consumption of de'fices· like this is relatively large and, in addition,
the support afforded to the face is not complete.
(ii) Another earlier development is represented by the pitching milling-head
hemisphere of HALLINGER. As shown in Figs 6/ l 02a and 6/ l 02b excavation is effected
here by a hemispherical cutting head built in the front and extending somewhat
beyond the cutting-edge plane of the shield. This cutting head is a sheet-steel
hemisphere mounted on a central transverse shaft. The steel skin of the hemisphere
is provided with cutting slots which peel off the ground as they perform a ' nodding'
movement at a central angle of about 30° - 45°. Peeling is effected in correspon-
dence with the push in 1-2 cm thick slices by the star-shaped cutting inlays and
the slices drop through the slots into the interior of the shield on a transporter
band conveying the material into ships. The cutting head is actuated by an eccen-
trically mounted and hydraulically operated piston-rod. The excavation of that
part of the face lying beyond the surface of the head, i.e. the area corresponding
....
V,

l1iling segments

- ,--·-
I
.l!fL. _ _ _ __ __ __ _ _ _ l z
r-
0
~ 0ti)
I I ~
I ~ "'0
~" :,,
0
1l C
z
1,l'§ 0
>
'--5 z
0
C
z
0
"'
:,,

High pressure pump ~

Segment pl8c1ilg crane ..,>
"':,,('"l
0
~
ti)
FIG. 6/102a Spherical shell segment excavator (HALLINGER type) m

--.J
~
708 CONSTRUCTIO N AND D ESlNG OF TU NNE LS

Fm. 6/102b Front sight of cutting spherical shell segment

to the vertical projection of the cutting edge and that of the strip left between the
two structures has to be sliced off by the cuttini edge during the push. In conclu-
sion, this shield operates in an exactly opposite manner to the roof shield when
stoping the centre of the face involving a potential breakdown of the roof. Thus,
it may produce unwanted roof collapse owing to excessive overbreak in drier
granular soils, whereas in plastic cohesionless soil the slots of the cutting opening
easily become clogged by the squeezed material. This, in turn, may prevent
excavation and stop the push because of the full-face resistance. This scraping
head in motion requires considerable energy which increases rapidly with its
diameter. Therefore, its use will not be advantageous over a diameter of 5·0 m
(17 ft) even in relatively hard cohesionless or mixed material with little moisture
content. This shield secures good steerability through the clutching of the cutting
edges into the virgin soil at the push.
(iii) MAKOVSKIY attempted to make the operation of bulkheaded shields more
economical. He employed water jets in the excavation for the liquefaction of loose
soil in the face and let the slurry flow through bottom slots into an interior sump,
from where it could be pumped out. Push was effected in the form of a slow
constant motion under the pressure of the hydraulic rams. This process, however,
was not accepted in general practice because the extent of blind excavation could
not be controlled. It often led to unexpected and unwanted roof subsidence and
 SHIELD TUNNELLING 709

eventually brought about a considerable increase and non-uniform distribution

of rock pressures and surface subsidence.
(iv) A decisive step forward was made in mechanized excavation in the Soviet
Union by the intt'oduction of rotating cutting discs. The first of this type was
used in the construction of the Leningrad Subway in Cambrian clay. As shown
a
in Fig. 6/103a four-arm cutting head is rotated in front with six smaller cutting
discs mounted upon them and revolving in opposite directions. Tbis planetarian
movement covers practically the whole area of the face and the cutting ribs scrape
off the material which is collected into a chute and carried back into the inside
of the shield whence it is hauled off by an appropriate conveying system.

Cutting rtbs

Propulsion jack·

FIG. 6/ 103a Leningrad shield with rptating cutting discs

Rotation is effected through a central shaft which operates the revolving discs,
in turn, through a somewhat complicated gearing. A very high rate of construc-
tion progress amounting to 10-15 m (33-50 ft) daily has been obtained for 6 m
(20 ft) diameter shields with this device.
Another rotating cutting-disc shield has been used successfully in a limestone
of medium strength during the construction of the Moscow Subway. In this rock
the free·face and the heading and bench methods were used formerly, the cast-iron
lining segments being placed by moving erector arms.
However, blasting involved considerable overbreaks leading to an unwanted
increase of rock loosening and, correspondingly, of rock pressures and considerably
increased the amount of material and labour for backfill and grouting. Therefore,
the application of the rotating cutting-discs shown in Figs 6/ 103b and 6/103c has
710 CONSTRUCTION AND DESIGN OF T UNNELS

Fm. 6/103b Moscow

shield with cutting rliscs

Fm. 6/103c Newer

Moscow-type shidd
 SHIELD TUNNELLING 711

proved more economical although the wearing-down of the cutting blades is rather
considerable (1 to 2 blades for each lineal metre of progress) and the time loss for
the change is also far from negligible. As is illustrated two diametrically placed
rotating discs (each nearly equal t6 half the diameter) are mounted.upon a centrally
driven annular ring, their shafts being driven separately with suitable gearing.
The discs rotate together with the annular ring, their individual rotations being
in opposite directions. Buckets are mounted outside the main ring for the collection
of falling rock which is again transported through chutes to the rear of the discs
and hauled off mechanically by transport bands and lorries. Daily progress has
6 20
attained 10 m (33 ft) for a 6 m (20 ft) diameter tunnel. · b
Tunnel construction with one single revolving cutting disc, 'continuous tunnel
boring' dates back otherwise to· the introduction of the so-called 'moles' patented
by GOODMAN and perfected by ROBBINS. This was used without shields and is
mostly applicable in medium bard rocks where the bridge-action period lasts
until the final lining can be installed and no temporary support is required. It is
used mainly in limestone, sandstone and shale and is unsuitable in plastic clays
or in cobesionless soils as well as in hard igneous rocks.
Basically, the 'mole' consists of a large rotary cutting face which carries rolling
discs and, in some types, fixed drag-type cutters, all arranged to cause spalling
of rock from the working face. 6 ·2 0c
Behind the cutting head are hydraulically activated gripping shoes and propul-
sion jacks, which not only provide torque reaction but maintain the required
cutting pressure against the rock face.
Rock, crumbling from the face as a result of pressure and torque applied to the
cutting head, is scooped up by revolving buckets fixed to the rotating head peri-
meter just behind the cutters. The falling rock is then carried by a conveyor system
to the rear of the machine (see Fig. 6/ 104). Some of the more important advan-
tages claimed for the 'continuous boring' method are:
Circular smooth and unshattered boles devoid of blasting fracturation and
damage are safer than blasted ones and usually require no, or very little,
temporary support.
Overbreak is reduced to a minimum anti therefore a saving in the amount
of permanent lining material can be obtained.
The uniform size of the broken rock renders muck haulage and disposal
much easier.
I11 Fig. 6/ 104 a complete set of machinery for the continuous boring method
is shflwn and it can be seen that mechanical devices can also be attached to provide
for the erection of ring beams if necessary as a temporary support for the heading.
In conclusion, it may be stated that revolving cutting discs are suitable in rocks
of medium strength and in hard clays or shales of moderate water content and
6 20
· b Dr. H. WAGNER: Vortriebsmaschinen im Tunnelbau, Die Bautechnik, June 1963
6 2
• 0c Engineering and Mining Journal, March. 1960
'112 CONSTRUC TION AND DESic.N OF TUNNELS

ca,
Cr)
· &;
of

diJ
M
Jin

Buckets for pr
broken rock
fo
m~
,aa
ar~
an
the
dr

th

FIG. 6/104. The GOODMAN -ROBBINS continuous tunnel boring set

,Sec/,Jon 8-8 Section C-C Section A-A

FIG. 6/105. The drum digger-type shield (KINNEAR MOODIE and Mc ALPINE)
 SHIELD TUNNELLil'IG 713

camtot possibly be used in cohesionless soils or in plastic clays. The costs of mac;hin-
ery and the energy demand increase progressively with the diameter. Available
·soviet data indicate that for the excavation of 1m 3 (1 ·4cuyd) ground about 5-8 kWh
of electric energy consumption must be allowed for.
(v) The most recent development in mechanized shield construction is the drum
digger type of shield, which was developed by Kinnear Moodie and Co. and by
Mc Alpine Ltd. and was successfully used in the construction of the new Victoria
line of the London Underground, and the Toronto Subway.
The drum digger consists of two drums (Fig. 6/ 105) of which the shield skin
proper constitutes the external main drum, its leading edge being bevelled to
form the cutting edge. Within the main drum is a rotating drum of smaller dia-
meter carried on two roller races and provided with a thrust ring to take the axial
load from the rotating cutters arranged in its front. The, Gutting teeth themselves
are mounted on six arms forming a cutting head at the front of the shield. The
arms carry eight changeable teeth each, and are mounted at the outer edge of
the inner rotating drum so that they cut the area between the inner and outer
drums. The area in front of the inner drum is cut by teeth mounted on a removable
arm across the diameter of the inner drum. The distribution of the teeth along
the arms varies from arm to arm and they can be changed.
The inner drum and the cutting teeth with it can be rotated at speeds up to
4-6 rpm in both directions with much lower energy consumption than the full-
face cutting discs driven by a central shaft.
A further reduction in energy consumption has been obtained by omitting the
tail section entirely and reducing the skin length of the outer drum accordingly.
In solid and uniform clays or marls and shales for· which this shield is princi-
pally suited, cutting can be done with sufficient exactitude to allow the linfng seg-
ments to be placed directly upon the excavated earth face and
propulsion can be effected by rams supported on the segment
ring which is already outside the shield proper. Natur~lly, a tail
section ca.n be added, if required, although the direct placement
of the segments on the excavated earth face will prevent any
unwanted difference in level (steps) between the subsequent
rings as they slide off from the inner surface of the tail (cf. Fig.
6/82). The speed record in tunnel construction was obtained by the
Kinnear Moodie and Mc Alpine drum-type shields giving a pro-
gress of 1 m (3·3 ft) per working hour for a 4·20 m (14 ft) exter-
nal' diameter tunnel. 6 ·20 d
The main criticism of the value of shields provided with mecha-
nized excavation devices is (besides the danger of mechanical de-
fects) that their smooth and efficient operation is possible only in a
homogeneous soil. Whenever soil conditions change, e.g. boulders,
quicksand, lenses, etc. are encountered, they fail to function and

s.?Od Engineering, March 1961, and Bautechnik, June 1963

714 CONSTRUCTION AND DESIGN OF TUNNELS

any repeated dismantling or change in the mechanical devices involyes consid-

erable loss in time and labour.
(vi) Based on wide experience, RICHARDSON and MAYO consider the use of the
following excavation methods as most suitable in particular soils;
In liquid silt the use of a bulkheaded shield is advised. The bulkheaded front is
pressed against the soil and an amount of material up to about 20--25 % of the
cross-sectional area is allowed to enter the' shield - chiefly through the top sluice
gates. This material is levelled at the bottom and used as overload to counteract
any uplift tendency due to compressed air. The rest of the material has to be
squeezed around the circumference when propelling the shield.
In soft clays the use of bulkheaded shields is also advisable, but all the material
corresponding to the cross-sectional area has to be allowed to enter the shield.
It is, however, possible that driving shields in such material under a relatively
thin surface cover in urban areas will produce a few centimetres of surface heave
even under such circumstances, but this -is still more tolerable than surface sub-
sidence. It may happen that the liquid clay as a result of being pressed and squeezed
at the front, will pass to the tail section and force an entry into the shield from
the rear across the gap between the skin and lining segments.
In dry clay, conglomerated gravel and in all self-supporting ground open-face
shields. equipped with the previously described mechanical cutting devices can be
advantageously used. The front face can
be excavated freely o~ under the protec-
tion of light strutting nearly in the whole
cross-sectional area, ' and only the ·peri-
meter has to be cut by the cutting edge
4dVJsed Of/ experi-
during the push.
e11ce An awkward situation arises when the
rp,>C/, >C/, ;,q,, shield has to be driven through a mixed
stratification. It is somewhat better when
the top layers are loose or soft, because
the use of a hood or other face stabiliz-
ing methods may be successful. But if the
FIG. 6/ 106. Shield face stabilization
bottom layer is of very hard rock so
with horizontal division
platforms (laminated face) that excavation necessitates blasting, the
greatest precaution and care have to be
taken as this operation may be very dan-
gerous. If, on the contrary, the top layer is of solid rock, then this may be used
as a protective roof for the excavation of the bottom layer and then it can be
removed carefully with the aid of temporary posts and blasting with small charges.
The prior installation of the lower half of the permanent lining segments may also
be helpful.
In sand and gravel and in all cohesionless soils very appreciable help may be
afforded to the stabilization of the working face by the hood, face jacks and hori-
 SHIELD TUNNELLING 715

zontal division (working) platforms. These all count_e ract the inrush of the rolling
material, the latter restricting it to a distance of m cot ¢, where ¢ denotes the
angle of inner friction, and m the height difference between the successive working
platforms (Fig. 6/ 106). The strutting effect of these platforms is also maintained
during the push and, in fact, the sloping material which has entered the shield
provides an elastic moving support. Practical evidence gained in the So-.iet Union
has revealed that the resistance against propulsion can be well regulated by partial
blinding of the face and jacking. Forces were generally lower, when the sides
were bulkheaded by temporary sluice-boards, the quantity of which had to be
adjusted to the actual nature of the soil.-6 ·2 0e In the same way division platforms
may be removed entirely when the improvement of soil conditions admits and,
therefore, must be structurally shaped to allow for eventual dismantling. To
counteract an accidental escape of compressed air through the cohesionless slopes
a clay blanket may reduce losses considerably. The lower third is usually left
more open so that the soil there can enter the shield.
63.222. Mucking (haulage). Effective muck haulage is one of the major problems
of efficient tunnel driving. In the case of shield tunnelling it is performed in two
steps. The first is the immediate removal of soil from the shield body and the
second is its conveyance to the ventilation or working shaft. The importance of
rapid and well-mechanized removal from the shield body cannot be sufficiently
emphasized and the use of efficient belt conveyors is indispensable. Haulage is
usually effected towards the rear, but may be effected sometimes in the case
of special advanced drifts (see Fig. 6/99) towards the front when this involves
least interference with excavation, lining and grouting operations. Conveyors also
provide for the loading of muck trucks (cf. Figs 6/ 102 and 6/83). If track or trailing
arrangements are inadequate or for long distances special loading machines have
to be used for this purpose.
The second step is the conveyance of muck through the completed tunnel
section to the access shafts or to the tunnel portals. This process is completed in
the former case by elevating the cars in cages to the surface and emptying them
into ordinary transport vehicles.
All loading and haulage equipment (belt conveyor, loader cars, traction engines,
track, etc.) must make the smallest possible demand on space. Usually special
cars and locomotives are used with this in view having, in general, a width twice
that of the track gauge and since it is necessary for two cars to pass in the tunnel.
the gauge should be about a quarter of the tunnel width, ranging between 60 cm
and 90 cm (24 in and 36 in). With a single track a quarter of the tunnel width is
the limiting gauge factor and this is also affected by the curvatures. The sharper
the curves the smalJer the gauge should be. Generally, a single track arrangement
may be adequate. provided that appropriate passing tracks are arranged at the
most important spots, i.e. just at the rear of the shield and in front of the lift

6 •2 0c Osnovania i Fundamenti 1964 No. I

 716 CONSTRUCTION AND DESIGN OF TUNNELS

cages where a reserve set of cars must be always at hand to ensure continuous
operation. For longer haulage distances intermediate sidings may be indispensable.
In order to save space the use of portable 'Californian sidings' (Fig. 6/ 107) may
also be considered. Some practical \irrangements as advised by RICHARDSON and
MAYO are shown in Figs 6/108 and 6/109.
A very important factor in efficient and economic haulage is the solidity
and reliability of the permanent way. Poorly laid track will increase haulage
traction, reduce travelling speed and results eventually in numerous and costly
derailments. Track defects are due not only to the rails (often taken from scrap)
and ties, but also to careless maintenance service. This can be prevented by the
employment of separate track crews independent of the mucking crews who are
always in a hurry.

Permanent track

Fm. 6/ 107. Portable Californian siding

CJ
Moving crane

Locomotive

F IG . 6/ 108. (a) Cross-over arrangement with front loading; (b) and (c) storage track arrangement
I·; ,,.t - •
'! •

'\.' . ";,"/'J ; · ,:·'

.,_ I ~
• . . ,,. "1"
.. '. "'.~ :,:
' . 1 '<.;.
~:,,. v~-•
,?.......
•~~o "'H • ·,/
, A_·; !.~ . .,\*"
 SHIELD TUNNELLING 717

~E~r~~r,
A
~t=?Lincage
Sprm.tJ sw,tcll "'- Loaded cars

Storage track J'prmg swttcll Empt!I Ca/I$ Spring sw,tclles

Lo,;ded cars
Storage track Empty CHS

FrG. 6/ 109. Track arrangements in front of lift shafts

Mucking cars are either self-emptying or dumping cars, or when fixed they are
emptied on a special dumping platform adjoining the shaft by a revolving drum
into the central muck-hopper bunker).
The utmost possible mechanization in muck haulage is a very essential item in
tunnelling as the obtainable progress determined by the two main factors, excava-
tion and lining, should not be hindered by haulage and transportation difficulties.
Electric-battery locomotives or trolley locomotives are best for traction; pref-
erence being given to the former type because of safety requirements. The trolley-
battery locomotive is a desirable combination of the two. Outside the tunnel or
in completed sections this locomotive operates off the trolley wire. In the uncom-
pleted sections, past the end of the wire, the trolley pole is lowered and the loco-
motive continues on the battery. These locomotives are self-chargers, part of the
current when working off the wire being diverted to recharge the batteries, thus
cutting short the idle time used for recharging and complying with safety require-
ments.
Internal combustion locomotives cannot be used under poor ventilation con-
itions. The combustion gases of diesel locomotives are not particularly dangerous
 718 CONSTRU CTION AND DESIGN OF TUNNELS

to health because of the absence of carbon monoxide but the

presence of
acrolein must be considered (cf. Section 46.531). lf gasoline locomotives
are used,
efficient ventilation must be provided and the air frequently tested
for the pres-
ence of CO.
Concrete pressure-pumps are very efficient for the conveyance of concret
e for
the lining. These installations convey the plastic concrete from a central
mixing
plant to the face. to distances of several hundre d metres within relative
ly small
(200-300 mm) diamet er pipe lines.
63.223. Shield advancement. Accurate propulsion and direction of the
shield is
a very delicate part of shield tunnelling, because it determines not
only how far
the designed alignment and gradients will be conformed to, but it
also seriously
effects the construction. Should the deviations exceed the prescrib
ed tolerances,
the whole operation will fail, because the joints and connections of
the cast-iron
or reinforced-concrete lining segments cannot be attache d to each
other withou t
forcing and overstresses, or with the required watertightness. Even
the additio n
of subsequent lining rings will be rendered imposs ible beyond a certain
degree of
horizontal and vertical distorti on.
As propulsion is done by the propelling rams (hydraulic jacks) the
first task is
to ensure both their uniform spacing along the shield perime ter
and also their
perfect co-operation. But at the same time their indepe ndent action
must also be
secured as steering can be effected only by exerting unequal jack pressur
es on the
various parts of the shield periphery.
Occasionally an extra jack may be placed at the bottom to preven
t the shield
from ' nosing down' (wobbling).
The relatively small interstice between the extrado s of the segmen
t lining and
the intrados of the shield skin limits the measure of corrections,
however, and
a superposition of deviations rapidly attains a measure which cannot
be correct ed
withou t the regressive dismantling of a couple of erected rings. Someti
mes the
insertion of unwanted counte r curves (bottle necks) in the alignm
ent becomes
indispensable.
The required clearance dimensions may also be impaired.
The efficiency of unequal jack action can be improved by some other
mechanical
means, too, as, e.g. the advance of the face excavation·on one side and
its retentio n
on the opposite one. Forced excavation always has to be applied
on the side
toward s which we intend to direct the shield. A more forceful method
is to instal
props or rakers in front of the cutting edge on one side to divert the shield
toward s
the opposite side. Such rigid and resistan t struttin g may produc e
considerable
torsional stresses in the shield body which may easily involve its perman
ent defor-
mation , too. The installation of ' lead-bo ards' on one side of the shield
at the begin-
ning of the push is also a help in correcting its alignment. An effectiv
e means of
keeping the shield in the correct vertical position is to drive an advanc
ed bottom
drift beyond the face. The invert of this drift is often provide d with
a concrete
bottom which effectively prevents 'nosing down' in weak soils, and, simulta
neously,
assists in stabilizing the face (cf. Section 63.21 [}.
 SHlELD TUNNELUNG 719

The correct direction of the shield must be checked by geodetic instruments,

of course. For this purpose a diopter is hung from the top of the tail section with
adjustable spider lines. During each push the shield operator must check that the
spider lines coincide strictly with those of the diopters hanging from the completed
sections of the tunRel. When driving in curves the exact direction can be given
for each push by the corresponding horizontal or vertical correction of the spider
lines (see Section 5.33).
63.224. Erection of tunnel-lining. Lining segments will be placed by band in
small diameter shields and usually with light pressed structural-steel or thin
reinforced-concrete elements only. The most simple mechanical device is a simple
hoisting winch the rope of which is led on a moving boom. both being mounted
on a travelling platform. This is also used in the erection of the concrete block
or reinforced-concrete segment lining of relatively small diameter public utility
tunnels and the placing into the exact position required is aided and directed by
hand.
In tunnels of larger diameter segmental lining elements ac~ erected with a
hydraulically-operated erector arm which can be mounted either directly on the
axis of the shield tail or on a travelling working platform following closely behind
the shield (Fig. 6/ 110). Although the previous arrangeme nt is mechanically simple,
it is much less adjustable to the given circumstances, owing to its being in a fixed
position.
The erector arm can be rotated around the horizontal axis to any required
position and extended or retracted as required. The end of the erector arm must
be provided with a suitable grip also operated hydraulically (Fig. 6/ 111). This
grip must be accommod ated to the structures and rib spacing of the segment.
63.225. Grouting, caulking and drainage. Lining segments must be followed as
soon as possible after placing, in almost all kinds of ground, by the grouting of
the back space left behind the shield tail after the push (cf. Fig. 6/82), in order
to prevent considerable surface subsidences and increases in rock pressure. The
formation and extent of the back space depend on both the nature of th<> ground
and the type of shield. The annular void left between the respective surfaces of
the circular shield-skin and the smaller dia. lining ring is sickle-shaped, its maximum
height at the crown being equal to the sum of the thickness of the tail skin
(3-6 cm, 1- 2 in), and the width of the safety gap (3-5 cm, 1-2 in) left between the
extrados of the lining a·nd the intrados of the tail.
In temporarily self-supporting ground even the tunnel arch can sag during the
bridge-action period under vertical loads for lack of a lateral support should the
void left in the spring line by the advancing tail not be filled in quickly enough.
Filling of voids at the crown can be carried out later, but this must also be done
before the ground caves in. The sag of tunnel lining under such conditions can
also be prevented by employing horizontal bow-string bars (turnbuckle rods) at
the spring line (cf. Fig. 6/ 167). Their application, however, is rather cumbersome
and connections to concrete blocks are very difficult to make.
 _,
t-.J
0

(1
0
t;
..,
:,,
c::
g
z
z>
C,
C,
"'"'
ci
z
.,,0
>-l
c::
z
.,,z
i;;

Fro. 6/110. General arrangement of travelling platform with erector arm and grouting equipment
 SHIELD TUNNELLING 721

Fm. 6/ 1 ll . H ydraulically
operated erector
a rm when
placing e lining
segment

In plastic, clayey soils voids left behind the shield.tail can be filled by the proper
J1anipulation of the shield. If, e.g., a bulkhead shield is driven, it can be arranged
by a corresponding control of the ~luice gates that the clay pressed aside will
tightly encircle the shield and fill out all the voids left behind, with the exception,
perhaps, of that at the crown. . .
The case is similar in liquid silts, but the possibjlity that horizontal pressures
may temporarily exceed vertical ones must be taken into account.
The most difficult soils through which to drive a shield are running sands and
gravels, where surface subsidences can hardly be prevented. Here, the possibility
of pressing the ground aside for filling the voids is completely out of the question;
the only remedy is to push the shield very slowly so that grouting ca·n keep pace
with the sufficiently retarded advance of the shield.

46 Szccby: The Art• of Tunnelling

722 CONSTRUCTION AND DESIG N OF TUNNELS

Grouting has several functions including:

the establishment of a tight bai:;kfill;

water sealing;
stabilization of the surrounding ground, so contributing to a reduction
of the ground pressure acting on the tunnel lining.

These aims, however, . cannot be simultaneously attained with every 'type of

grouting, nor to the same efficiency, being alway~ dependent upon the nature
of the ground and the type of grouting material.
A general distinction is made between (a) the primary grouting intended to
fill the back space, and (b) the secondary grouting for sealing and stabilizing.
Primary grouting is carried out under low pressure (5 atm-6 atm = 70-85
lb/in2 ) , the injected material being ce-
Lilllng segment ment grout, mortar of gravel or coarse
sand. Secondary grouting takes place
under high pressure (10 atm - 25
atm = 160-400 Ib/in2) using more ce-
ment fluid suspension, bentonite, hot bi-
tumen or bituminous emulsion. The
grout is forced in through grout plugs
A;r vent
(4 cm -- 5 cm; 1 1/2 in - 2 in dia) provided
in the lining segments and closed for the
time of erection (Fig. 6/ 112). In each
lining ring at least four grouting holes
are provided with a staggered distribu-
tion. In less permeable grounds closer
F1G. 6/ 112. Grout plug and connections spacing will be required.
(a) For primary (back-filling) grouting,
fairly wet cement mortar with a cement/
sand ratio from 1 : 1 to 1 : 3 was exclusively used at first; this was forced in at
the bottom grout plugs as the shield tail cleared them, while the plug next was
removed to serve as an air vent. When mortar appeared at the upper hole the
grout hose was then transferred to that plug, where the injection of grout continued
till that sector was filled. As shown in Fig. 6/ 112 grouting is effected by screwing
a stopcock directly into the grout plug. The grout hose with an air vent is connect-
ed to the stopcock. When grouting is completed, the inner stopcock is turned to
retain the pressure in that hole until the grout sets. Meanwhile, the grout hose
is transferred to the next plug.
Besides the large quantity of grout and consumption of cement required in loose
granular soils and with large annular spaces left behind, the great objection is that
owing to its liquidity it is difficult to prevent it from flowing around the segments
and entering the shield. This may be prevented by using 'circle boards', segments
of wood cut to a radius equal to the inside of the tail. These are inserted into thP-
 SHIELD TUNNELLING 723

gap between shield skin and extrados of the segment ring and by fitting tightly
against the tail may prevent leakage with very little caulking. Owing to its liquidity,
much of the grout will be wasted when entering the voids of looser soils and
p'°assing, e.g. through fissures, to adjacent sewers. All losses are disadvantageous
because of the high cost of cement. Moreover, it might also occur that air bubbles
will remain in the injected mortar leading to subsequent settlements.
Considerable savings in cement can be obtained if the voids are promptly filled
with bird's eye gravel or with granulated slag, screenings, etc. of uniform size
(D = 4-6 mm; 1/6-1/4 in) at first. Grouting itself is done in the same way as
above, using a grout hose of similar diameter, in the same order of sequence and
according to the same principles. As the shot has a pore volume of about 33 %,
and with regard to the fact that the gravel will not flow like a grout and will not
pack so tightly, occasional spots may be produced into which gravel fails to flow.
Therefore, gravel 'shooting' is usually followed by a cement grout of 1 :- 1 mix
and quite wet. This consolidates the gravel backpacking by filling the voids.
A secondary benefit of this cement grouting is that it stops most of the leaks
into the tunnel. {This effect may be increased by the addition of 5-8 %bentonite.)
Grouting is a highly skilled operation requiring experience and judgement to
choose both the proper consistency and grain size of the grout and the highest
permissible but harmless grouting pressure.
It must be noted that if the
feeding batcher
injection pressure is raised without
forcing in additional quantities of
grouting material the increased
pressure will act merely as external MixiM
loading upon the erected rings pos- chamber
sibly producing an excessive local Mixing
overload sufficient to bring about rock
the failure of connections or even
of the respective lining segments.
Two types of grout machines (pneu-
matic or mechanical) are distin-
guished as they are used for prim-
ary or for secondary grouting.
For primary or low-pressure grout- Pressure tube
ing, pneumatic grout placers are
Fm. 6/113. Single chamber grout machine
generally used operating at an air
pressure of 5-7 atm (80 lb/in2) sup-
plied by travelling compressors. Some of these appliances are of quite simple
design which can be produced in a local workshop. Their operating air pressure
is, however, limited and not adjustable to the actual resistance encountered in
the ground. Popular machines both mix and place the grout. Some types mix by
agitating the water with compressed air, and some by the action of paddles rotated
around horizontal or inclined axes.

46•
 724 CONSTRUCTION AND DESIGN OF TUNNELS

The main types are as follows:

The single-pan placer with a capacity of 150-200 1/h (Fig. 6/113). The batch
of cement and sand fed into the feeding cone is dropped into the mixer through
a cup valve there to be mixed and agitated with water entrained by compressed
air, the quantity of which can be controlled by a mixing valve. When a homogene-
ous grout has been produced in the mixing chamber, the batch can be forced out
by compressed air into ,the grout pipe b)i opening a stopcock. After discharging
the batch, the feeding and mixing cycle can be repeated.
The double-pan placer with a capacity of 200-3001/h can replace two single-pan
placers and affords a possibility for continuous grouting, because while the mixed
batch is sluiced from the upper pan into the lower one and forced into the grout
pipe, the next batch is being mixed in the upper pan (Fig. 6/ 114).
The inclined drum placer (Fig. 6/ 115) with a capacity of 180-200 1/h is suitable
for placing grouts of lower cement and water content. With this equipment the
mixing and forcing out of the grout are done separately. Its main advantage is
that instead of using compressed air revolving paddles or a spiral agitator fitted
on the inclined axis within an inclined drum provide continuous mixing. The
batch is forced out by compressed air introduced into the drum. The compressed
air was first supplied by an air pump operated by hand (Wolfsholz design).

~ ~ ~ ~~~==1rwmpres
sedair
plant FIG. 6/ J 14. Double-
Airsluice 11. chamber grout machine

Safety
Y3lve
Mixing pipe

·!lopper
Check valve
J)!?ve "

Fm. 6/115. Grout

machine of the To pressure tube
inclined-drum type , 1/, "/,'-
 SHIELD TUNNELLING 725

Mixing paddle

F10. 6/116. Grout machine of the horizontal drum type

4-72 _,

FIG. 6/ 11 7. Soviet type

of gravel
-. 1i'!gger pea shooter

500mm

Manom/?tef'

Modern devices have their drum connected to a separate air compressor. This
step has eliminated the previous objection that the air pressure in the drum
gradually decreases with the progress of grouting.
The mechanical grout machine with mixing paddles fitted with wi1e brushes
is suitable for the injection of the most dense and dry mixtures (Fig. 6/ 116).
 726 CONSTRUCTION ANO DESIGN Of TUNNELS

Special equipment has been developed (pea shooter) for the injection of granular
dry material (sand or bird's eye gravel). This operates under an air pressure of
5-7 atm = 80 lb/in2 (Fig. 6/117). The material fed by air pressure from a container
into the grouting hose is directly entrained by a separate flow of compressed air.
As this material will not flow like a cement grout it is generally unnecessary to
use circle boards to prevent the gravel from running into the tail of the shield.
Should the grout be required - besides filling the back-space - to penetrate
the voids of the surrounding soil with a view to compacting and artificially bind-
ing its grains, then the composition of the grout has to be designed to make this
penetration possible within the permissible and available pressure values.
According to recent American investigations 6· 21, the grain diameter pertaining
to the 85 % ordinate of the grading curve of the grout (D85) must not exceed a
value equal to the one twenty-fifth of the grain diameter pertaining to the 15 %
ordinate of the grading curve of the ground to be penetrated (D 15). According
to MAAG the extension radius of grouting in homogeneous ground can be computed
from the formula

(6.12)

where h = the pressure head

t = the duration of grouting
r 1 = the radius of the grout pipe
n = the void content
k = the coefficient of permeability of soil.

Experience indicates that cement injection is not effective in a ground with

an effective grain-size smaller than 0·8 mm.
(b) The secondary of sealing grouting can be carried out with one of the follow-
ing grout materials:
1. cement grout, cement suspension
2. lean mixture of various types of cement and plasticizing agents
3. bitumen emulsion
4. hot bitumen
5. various chemical agents (particularly a suspension of bentonite or cement
bentonite).
The type of grout must be selected according to the nature of both ground and
groundwater as well as to the type of the respective underground structure.
Secondary grouting is carried out several months after primary grouting, usually
with the sole purpose of sealing.

21
• Pressure Grouting. Proc. AMSCE 1958 August.
 SHIELD TUNNELLING 727

While the immediate object of primary grouting is to fill up voids and the back-
space, with secondary grouting it is essential to wait until ground movements
set up by the construction of the tunnel have come to an end, and the ground
settles into a condition of equilibrium. As long as ground movements occur,
fissures may also arise in the grouted zone causing the seal to fail. Accordingly,
in cases where long-term ground movements are to be expected, the secondary
grouting will possibly have to be repeated several times.
Three types of grout material are used in Hungarian underground-railway
construction practice: pure Portland cement, a mixture of Portland and bauxite
ecments (a fast-setting mixture), and pure bentonite or a mixture of bentonite
and cement (cf. Table 6/l.X).
In general, bentonite grouting is used in fine-grained soils while cement suspen-
sion grouting is employed in coarse-grained soils. Where flowing groundwater
is encountered a fast-setting cement grouting must be used.
The general rules for secondary grouting are as follows:
Grout machines with automatic air-pressure control are to be used only.
The specific water-absorbing capacity of the ground must, in each case, be
determined before grouting to select the suitable grout material accordingly.
For this purpose a mixed sample has to be taken from the ground through the
grout from the back-space holes beyond the lining.
During the grouting process continuous attention must be given to grouting
pressure and grout consumption. It is a basic rule that grouting must be done
under increasing pressure and with a decreasing grout consumption. A decrease
in pressure is an indication of an escape of the grout either into horizontal fissures
or joints or its accumulation under an impermeable dome leading to surface
upheaval (Fig. 6/118).
When high grouting pressures are applied the tunnel lining must also be
checked for deformations and stresses in the .vicinity of the grouted section.

/ Heave
_________ _L _______ -
Ground surface

t t
I t t

FIG. 6/118. Geological scheme

of surface upheaval ·_ pressure .
brought about by
grout pressure
 728 CONSTRUCTI ON AND DESIGN OP TUNNELS

Pressure Groutings Used in the Construction of the Underground Railway in Budapest

Type of ground Gravel Sandy gravel Sand and Mo

D = 200-S mm D - 200-0·2 mm D - 2-0·02

Enainccring: structure - -
I I II I I II I I Ii
Type of grout

Shaft cement cement cement cement cement bento-

mortar grout mortar grout, grout nite
bentonit-
ic
cement
grout
Concrete tunnel and cement cement cement cement cement bento-
large undergroun d hall mortar grout grout grout grout, nite
construction . or ben- bento-
tonite nitic
cement
grout
Cast-iron lined tunnel cement cement cement cement cement bentonite
mortar grout mortar grout grout or fast-
setting
mixture

Non: In cases where two types of grout material have been suggested, the proper type to
be chosen must be decided
practically used.

As is known, great importance is to be attributed to the value of the permissible

grouting pressure. The greater the pressure applied the more perfect the seal
between lining and ground, and the more thorough the compactio n and stabiliza-
tion of the surrounding ground, while on the other hand, an unduly high grouting
pressure may cause the upheaval, even the up-break of ground surface, should
the grout be forced between two impermeable layers or into fissures or bedding
planes of such a layer. An excessive grout pressure may be still more harmful
if the tunnel lining suffers non-uniform and excessive local loading conditions
because of non-uniform loosening of surrounding layers and distributio n of
back-spaces. The increase of the grout pressure beyond a certain limit can -even
lead to a complete failure of the lining. With a view to this eventuality, it is of
great importance to take deformation and stress measurements in the tunnel
lining in the vicinity of the grouting site.

The highest permissible grouting pressure which remains within the safety margin to avoid
failure or upheaval of the layers can be calculated from TERZAGHJ's rock pressure theory
(cf. Section 32.314, Fig. 6/ l 19a) with the following formulae:
 SHIELD TUNNELLING 729

TABLE 6/IX

Silt Clay Rock

Type of ground Ip < 15 % (Marl)
Ip < IS %

Type of engineering structure

I
I II I I II I I II
Type of grout
- - ~

Shaft cement cement cement cement cement cement

mortar grout mortar grout mortar grout

Concrete tunnel and cement cement cement cement cement cement

large underground hall mortar grout grout or grout or mortar grout or
construction fast- pea or pea fast-
setting gravel gravel setting
mixture mixture

Cast iron lined tunnel cement cement cement cement cement cement
mortar grout mortar grout mortar grout
or fast- or pea or fast- or pea or fast-
setting gravel setting gravel setting
mixture mixture mixture

,on the rate of water flow. In places with high rates of flow a fast -setting grout (e. g. PC + bauxite or Tricosal) is

(a) For homogeneous layers:

4c + yd [exp ( 4). t;n <I> h ) _ 1] , (6.13)

Pk = 4.l.tan <f>

where c =
the cohesion in ton/m2
y =
the dry density in ton/ m2
<J, =
the angle of internal friction of the ground
h =
the overall thickness of the overlying layer
l = the ratio between horizontal and vertical ground pressures
d = 36A (where_ A is the volume of injected grout in m3) .

The value of Pk as calculated _with this formula, however, must not exceed the maximum
value determined by the available passive earth pressure

where
Ap = tan 2 (45° + <f>/2).
 730 CONSTRU CTION AND DESIGN OF TUNNELS

(b) Jn a multilay er system the calculat ion must be carried out

separate ly for each layer
in a descend ing order of seq uence. First the maximu m permissi
ble pressure for the upperm ost
layer is calculat ed in its bottom plane (Pk,); then the permissi
ble pressure for the next layer
is compute d (Pk:), taking into consider ation that in its top plane
the already known pj, 1 pressure
will offer a certain resistanc e against upheava l. Accordi ngly, the
maximu m pressure permissi ble
for the 11th layer can be calculat ed supposi ng that the permissi
ble pressure for the (11 - l)th
next overlyin g layer is already known (Fig. 6/ Jl9b):

41 tan ,f, ] 4l tan ,f,

pk• = 4c + y d + Pk.1 _, - -d-
4}. · tg cf, I e--d- - I . h•.
(6.14)

In calculat ing the maximu m grouting pressure, the cohesion

of the layers immedia tely
overlyin g the respectiv e undergr ound structur e up to a height
of 10 m above the crown of
the tunnel must not be taken into account with a higher value
than c0 = 10 ton/ m! with
regard to the possible occurren ce of cavings
BJ in settleme nts a nd fissures in the course of
the tunnel construc tion.
lt is a limit in practice that the' maximu m
pressure should n ot exceed 2·5 times the
relevant geostati c pressure at any point.
T The injection pressure must be controll ed
T
Ii with the utmost care, as e.g. during the con-
struction of the undergr ound railways in
Budapes t, surface upheava ls of several milli-
metres occurred at some points as a result
of excessiv e grouting pressure s.

ad 1. Injections of cement suspen-

b; sions are unsuita ble for filling 1arger
voids and back-spaces as they can only
be effective if the dispersed cement
grains are practically sieved out of the
a fluid and retaine d by the grain skeleto n
Ii of the ground , where they form a gel
on its surface . Cemen t particles will
settle out from the dispersing water at
a distance where the velocity of flow
falls below a certain limit value. The
leaner the suspension the greater this
distance. Moreov er, a high water-
FIG. 6/119. Forces conten t is by no means disadva nta-
acting against grout pressure
geous for the stabilization of the ground ,
(a) in a homoge neous layer;
(b) in a multi-la yer system as water will flow ahead from the
retained cement particles, the fractio n
left behind becomi ng denser and den-
ser, the pores and voids being gradua lly
filled regressively from the settlem ent bounda ry toward s the groutin
g pipe. Thus,
the use of a leaner suspens ion is definitely advantageous considering
both the radius
of action and the use of lower grouting pressure which also reduces
the inheren t
 SHIELD TUNNELLING 731

hazard of ground failure. On the other hand, the strength of the gel formed on
the surface of retained particles will be the higher the thinner the water film
surroundin g them. A high water content will, additionall y, result in a more porous
cement stone. •
ad 2. Both kinematic viscosity and action radius of grouting can be increased
considerab ly also by the addition of various plasticizin g agents (such as plastiment ,
intrusion-a id, bentonite, ~tc.).
ad 3. One of the bitumen-in jection methods used is the cold method (the Shell-
perm method), with which a perfectly fluid, water-like bitumen emulsion is injected
into the ground . This emulsion readily penetrates into the pores of even the most
fine-graine d sand soils as the size of the emulsified bitumen particles is only about
I or 2 microns and the viscosity, of the injected fluid with a bitumen content of
50 % is only l ·5-2 times as high as that of water. The injection pressure required
for such an emulsion is, consequen tly, relatively low. Moreover, the bitumen
particles will settle out only if catalyzed by a subsequen tly injected coagulatin g
agent.
ad 4. In the U.S.S.R. hot bitumen injections are used mainly in cases where
cement or cement-cla y grouting cannot be successfully employed because of
the aggressive properties or of the high seepage velocity of the gro und water en-
countered, i.e. in cases where there is a risk of the cement grains being washed
away or dissolved. A further advantage of the hot bitumen injection is that the
risk of an escape of the injected material into fissures is greatly reduced owing
to the physical properties of the bitumen. Thus, not only can economy be obtained
in the grout material but, in addition, the hazard of fissuration of the surroundin g
ground is greatly reduced. On the other hand, a disadvanta ge of the method is
that the bitumen becomes plastic under pressure, that its radius of action by
injection is rather small and that it cannot be applied in soils such as liquid clayey
silts.
For the sake of compariso n it can be mentioned that the kinematic viscosity
of the bitumen when heated to a temperatu re of 140-160 °C, is about 8 cm /s
2

as against 0·01 cm2/s for water which means that the critical velocity in a 10 cm
dia fissure or void is 0·023 m/s for water, while for hot bitumen it is as much as
18·4 m/s.
Owing to its gradual cooling the injected hot bitumen will deposit on the sides
of the voids forming a layer of increasing thickness soon leading to their being
filled. As the loss of temperatu re of the injected bitumen amounts roughly to
60 °C per lineal metre, the action radius practically never exceeds l ·5-2 m for
bitumen injected at a temperatu re of 180 °C.
Any further increase of injection pressure will produce the gradual deformatio n
of plastic bitumen, pre5sing it more into the surroundin g fissures a nd voids.
In flowing ground water, when hot bitumen is injected, bitumen strips or strings
are immediate ly formed and entrained. These strings when brought.in to coJJision
with obstacles, will be deposited and - wound up into small balls - reduce the
cross-secti onal area of voids. During the back-space grouting of the undergrou nd
732 CONSTRUCTION AND DESIGN OF TUNNELS

railway tunnels in Moscow, the voids were first flushed out with· water, then
filled under pressure with coarse grained aggregates and finally grouted with
hot bitumen. This method proved to be very succe~sful and applicable also in
fast-flowing groundwa ter, e.g. in maintenan ce work, as it offered both a good
water seal and anti-corrosive protection .
ad 5. The use of cement-clay grouting is rapidly becoming more widespread
owing to its following advantage s:
(a) It is economical in cement consumptj on and gives at the same time fairly
good water seal.
(b) The repeated removal of clogging material and the cleaning of injection
pipes- are easier than with cement grouting.
(c) Injections with grouts of high clay content can be carried out in stages
(staggered in time) and with a predeterm ined and controllab le setting time render-
ing the advance of the shaft sinking or tunnel driving easier in the less stabilized
ground.
(d) It can be employed in aggressive and waterlogged soils.
(e) The grout can be more readily pumped, causes far less corrosion to the
pump and owing to its plasticizing properties, readily penetrates into the voids
and fissures filling them uniformly.
During the constructi on of the undergrou nd railway tunnels in Budapest partic-
ularly the application of bentonite-cement grouts proved very successful compared
with the use of pure cement grouts, which when penetrating into finer voids or
fissures form a rather porous cement stone after setting, with poor water sealing
capacity, owing to the large amount of dissipated water. Concrete corrosion
is also promoted in cement grouts owing to this porosity which allow greater
percolation and as soft waters leak out, lime from the concrete percolatio n may
result in considerab le damage (lime sinter).
If the water-cement ratio of the cement suspension exceeds O· 5-0·6, sedimenta-
tion of cement particles will take place and the resulting cement stone will be
porous. This phenomenon is still more marked because a cement grout cannot
penetrate into voids with diameters less than O· l mm, as a result of which they
will be sieved out, so considerably decreasing the radius of action of cement
grouting. In order to reduce the water emitting capacity of cement grouts, suitable
chemical agents will have to be added to it, or finer grained cement must be used,
i.e. the suspension must be stabilized to prevent the early sedimenta tion of cement
particles. The more fine-graine d the solids of a suspension the smaller the risk
of their sedimenta tion, which is also hindered by the surface activity, 8· 22 i.e. by

6-:: NOTE. SKEMPTON

=
expresses the colloid activity by the following ratio : a1 P,/A,
where P 1 denotes the plasti<;it.y index a nd A the clay content. According to this classification
the value of a , is in inactive clays less than 0·75, in normal clays 0·75 and in active clays
more than 0·75.
 SHIELD TUNNELLING 733

the electric charge and by the adsorption film on the particles. It is a prerequisite
for the stability of a suspension that it should have a continuous grain-size distri-
j:>Ution curve, i.e. that no fractions should be missing and no jumps should be
revealed, because otherwise sedimentation will take place at each size limit.
Accordingly, the grade of stability of a suspension can be raised by the admixture
of additives (electrolytes, col-
loids, bentonites, clays) the Add1ltves:
particle sizes of which will 1. S54 cement-bentonite
secure the transition from the 2. S54 ceme11t-cl<1!J
J. S54 cement -o/igocene cl3y
cement particle size to those 180
of the dispersing medium.
Of the most important clay
minerals the particle size of
i150 ~
~
,,
kaolinite grains are nearly
100 %above 0·2 micron, while
in montmorillonite the per-
~ !40 I
-~
~ !20
•
.
J'~
,x
~

centage of this same grain-size ~

'3
!00 \
fraction amounts only to 20 %-
The bigger the specific surface 80 v \
area and the water adsorption
\
capacity of the mineral parti-
cles the greater is the stability
of its suspension. The most
stabe suspensions can, thus,
be produced by the addition
of Na-bentonites with a mont-
50

40

20
\
'-
'"'---
\
~
' ',"'
~
- ...........
M
"""
0
morillonite content. The ad- go 8/} 70 6{J 50 40 JO l,J
dition of bentonite provides Cement-
both a plasticizing effect and 0 10 20 40 llJ 70 80
Add;tfVl'
better water sealing and in-
creases the setting time of the Fro. 6/120. Effect of bentonite or clay addi tives upo n
crushing strength
grout but, at the same time,
it lowers its final crushing
strength (Fig. 6/ 120). Tests carried out in connection with the construction of
the underground railway of Budapest 6·23 clearly showed that the.addition of
bentonite resulted in the following effects:
sedimentation of' the cement particles decreased;
the crushing strength of the set grout was reduced;
the grouts became more readily injectible and, when set, more impervious;

s-: 3 BALAZS, KILIAN, K ELEMEN and SAROSI: Betonok vizzar6sagana k novelese a kotoanyag
javitasa utjan (The increase of the water sealing capacity of concretes by the im provement of
binding additives), Epites- es Kozl. Tud. Kozlemenyek 1957 3-4.
734 CONSTRUCTION AND DESIGN OF TUNNELS

setting time was increased;

an economy in cement consumption was obtained.
All these effects are of considerable practical importance. Great importance
is to be attributed to the fact that cement-bentonite grouts can be more easily
injected, thus requiring lower injection pressures and so affording higher safety
against surface upheavals. According to test results (Fig. 6/121), the addition
of Ca-bentonites will result in
1. · C!l bentonite } higher crushing strength val-
2. Activated bentonite of Mac/ +- P. C. ues but the action radius of
J. Activated Jstenmezeje bentonife·
Duration ofpreswellin!J JO m!fl. grouting is much larger when
Na-bentonites are added. Fig.
120 6/ 122 shows clearly that the
!/0
\ volume of the grout injected
under the same pressure is
I much greater with cement-
I \
I 1, bentonite grouts than with a
I
I \ pure cement suspension.
The advantages of bento-
I ' nite-cement grouting over pure
I \
\ cement may be summarized
\ as follows:
-~ '-
\ I
70

60 \ I (i) better water sealing;

(ii) easier injectibility with a

50
\\ \ I consequent increase of
injected grout volume;

40 \) \ \ (iii) a more complete filling

up of the voids due to
JO
\ \ the decreased possibility
of early water escape
\ I\ 2 from the grout (increased

""
20 r--....
resistance against corro-
10
~~ ~ ~ sion);
(iv) economy in cement con-
0
~ sumption and grouting
2 4 6 8 10 20 J'J costs, greater safety a-
Bentoflltl? admixture(¾) gainst surface upheaval
F1G. 6/ 121. Effect of the quality of bentonite upon hazards.
crushing strength
Grout machines employed
for secondary grouting are, in
general, reciprocating pumps suitable for supplying rather high injection pressures.
The grout machine which has proved most suitable in practice is the single-
piston-type reciprocating pump of HXNY (Fig. 6/123). With this equipment the
 SHIELD TUNNELLING 735

mixing and injection of the groutarecarried out separately. The equipment consists
fundamentally of a reciprocating pump combined with a membrane, with the
following operating principle. Compressed air enters cylinder (a) and by displacing
the fluid (oil) acts on a membrane (M). During its return stroke the piston sucks
back the fluid allowing the membrane to retreat, so creating a pressure reduction
in the pump box (b), as a result of which the cement grout will be sucked up
into this box out of the container (t). This volume of grout is then forced by the
next stroke of the pump into the delivery pipe (d). This process is repeated until
the predetermined maximum pressure is reached when the pump will stop automat-

/4

12
/
.......... -- --- -- --- -
~-- _ _ ,..1.-.

1>.;,·'
_..,_
y"'PermisS,_1b/e maxirnvrn grovL pre~ svre
·-r·---- -
{1~011\t:---
·i-----
-·-
f' '-).,. 'r;f-

>-- l
f
~~~
~7
~lt~

F1G. 6/ 122. Comparison 2

/;V
of injectible
grout volumes 0
V JOO 200 JOO 400
Volvrne of irzjected grovt (I)

T(I prov! hose

Operat111g va/VE
,,,\ CornpressM
air
~ = =~it===i{o)::a,::~$~3
/'1ix/ng
pa//
// Storing (ft
r:ontainer "'-
Perforated gnd and
compressed air inlet CompressM a,r tvbe

r,G. 6/ 123. HANY' s reciprocating grout machine

736 CONSTRUCTION AND DESIGN OF TUNNELS

ically. In case of a fall in pressure, due to the discharge of the grout from the
delivery pipe, the pump is automatically restarted. The. fixed injection pressure
a
is, thus, automatically maintained . (The equipment has capacity of 25·8 I/min,
i.e. 1·5 m 3/h at an injection pressure of 18 atm. Its main advantage compared
with earlier reciprocating pumps is that cement grout does not come into contact
with the oil-lubricated and carefully machined surfaces of the pump which are,
therefore, not exposed to the abrasive action of the grout.
The 'MAVAG' type grout machine, manufactured in Hungary (Fig. 6/ 124) is
designed on similar pripciples to HXNY's machine, as shown in Fig. 6/124a. As
the·piston moves ahead in the fluid chamber, a membrane is pressed into the suction

Air c11amt1e1 ·

Membra11e

A!r pressure equa-

Tlizlllf/ chamber
Brovt hose coo-
11ection
Grovt cootaioer

Rvbbermem-
braoe cylioder
Cock-tap

'00

FIG. 6/124. 'MAV AG' grout pump

 SHIELD TUNNELLING 737

l'rankshaft

Crank c/Jamber Motor

Fm. 6/125. Double piston injection pump of the Leningrad type

chamber where, on its return stroke, it creates a reduction in pressure by which

the grout is sucked thr~ugh the ball check-valve into the same chamber.
During the next stroke this volume of grout is forced through a second ball check-
valve (2) into a pressure tank to be discharged from there into a delivery (injection)
pipe. This process is repeated and maintains the pressure tank (air chamber)
which acts as a pr_essure equalizer to maintain the injection pressure at an approxi-
mately constant value when the injection valve is opened. This single-piston grout
machine has a capacity of 6 m 3/h at a maximum working pressure of Pmax = 15
atm. In this machine the grout does not come into contact with the piston and
the cylinder of the pu:r:np.
The double-piston type grout machine, developed in conjunction with the con-
struction of the underground railways in Leningrad, is also used in Hungary.
Here the grout enters the pump-cylinders and thus comes into direct contact
with the pistons. A schematic plan of the machine is shown in Fig. 6/ 125.
The pistons operate alternately in opposite phases. In the figure, piston I is in
the suction position (inner extreme) at which the grout i'l sucked through the
open inlet-valve into the pump cylinder. At the same time, piston II (shown
in the figure in its outer extreme position) forces the grout through the open
ouilet-valve and discharge pipe into a central pressure chamber and from there
through the pressure pipes into the injection tubes. In this way continuous grour-
ing is possible.

47 Szechy: The Art of Tunnelling

738 CONSTRUCTION AND DESIGN OF TUNNELS

Cleaning, rinsing and checking of the machine, its hoses and valves, especiall)'
after use, as well as frequent greasing of the respective parts is of the greatest
importance. Grain sizes of above 3 mm must be screened out. Operation without
a suitable manometer is forbidden. With pneumatic grout machines no safety
valves are required, while with reciprocating grout pumps they are indispensable:
In major injection work the use of
registering manometers is advisable.
As grouting is a work not without
certain dangers, owing to the high
pressures applied, great care must
be taken in the use of safety equip-
ment and clothing (protective cloth-
ing, goggles, cap and gloves, stable
scaffolds, etc.). The eyes and bodies
Removable plug of employed personnel in particular
must be protected against recoiling
grain particles.
Grouting, as can be seen, is an
important supplementary to shield
tunnelling, where the erector or
crane used for the placing of the
lining elements is directly followed
by the grouting equipment, which
is best mounted on a travelling
FIG. 6/ 126. Details of grouting tubes scaffold (ef. Figs 6/83 and 6/ 110).
(In case of concrete lined tunnels
the injection pipes are previously
placed in the concrete and plugged. Details of the coupling of the injection hose
are shown in Fig. 6/126.)

Grouting is important, not only in the waterproofing of underground tunnels, but also in
the subsequent water sealing of tunnels and other underground structures (cf. Section 72.211 ).
Grouting procedures and materials used in the construction of the underground railway in
Budapest are tabulated in Table 6/IX. The principle followed here was to undertake grouting
in a sequence from the bottom upwards and from the rear towards the front. The operation
was always continued until the grout flowed out from the next adjacent grout hole.

63.23. Lining Segments for Shield Driven Tunnels

In the construction of tunnels by the shield method, it is essential for lining

elements to have instantaneous bearing capacity, to be watertight and to allow
for rapid and simple positioning. The main requirements for such lining segments
are:
 SHIELD TUNNELLING 739

I. Immediate bearing capacity against external earth and water pressure without
detrimental deformation or leakage.
2. Resistance to impact stresses due to rough handling, transport or erection
operations.
3. Resistance to high axial stre:,ses produced during the advancement (push)
of the shield by propulsion jacks.
4. Resistance to moisture and ground water effects of the segment itself and
of its joints and watertightness as well as reiistance to corrosive action .
5. Economy in construction and maintenance. (His to be borne in mind that more
expensive stronger materiais may compete successfully with cheaper materials
of less strength when the difference in dimensions, erection and durability
is also considered.)
Linings may be single or double. Double layers are employed when the double
tasks of resistance to external pressures, and watertightness or aesthetic appear-
ance cannot be obtained with a single layer.
In the earliest periods of shield tunnelling brick was used for lining, but soon
cast-iron segments had to be introduced in order to resist. the large horizontal
pressures during pushing and, at the same time, to provide sufficient watertightness.
More than half a century has passed since concrete· blocks and structural steel
segments and, more recently, reinforced-concrete and prestressed-concrete seg-
ments have been in use. Attempts to introduce brick, timber or ashlar stone blocks
have generally failed, but the application of fresh concrete compressed by the
advance between a rigid inner shuttering and the rock face is possible.
Present practice is to apply cast-iron or reinforced-concrete segments in the
construction of larger diameter (railway and vehicular) tunnels, whereas concrete
blocks are given preference in the construction of smaller diameter (public utility)
tunnels. Fresh concrete is just beginning to compete with these materials.
63.231. Cast iron linings. The permanent lining for larger diameter shield-driven
tunnels is usually (for subaqueous tunnels almost exclusively) constructed of
cast iron. Its main advantages are that it can be quickly erected and is at full
strength to resist shield jacking pressures and external loads immediately upon
erection. It is more watertight than other types of lining and being relatively
thin for its compressive strength is not too heavy and is economical concerning
the required area of excavation. Cast-iron lining for tunnels of circular section
i'l made up of successive segmented rings bolted to one another by means of
circumferential flanges which at the same time act as stiffening elements. The
dimensions of the segments are determined by the weight limit required for easy
handling and erection operations. The length of the segments is usually between
1·5 m and 2·0 m (5 ft and 7 ft) whereas their width, i.e. the length of the
single rings is subject to several considerations including:
I. The distance that a miner can reach in advancing the excavation from the
old face to the new one, which is also a function of the shearing strength of
the ground.

47•
740 CONSTRUCTION AND DESIGN OF TUNNEL<;

N A A N
A A

segment unit
 SHIELD TUNNELLING 741

2. Foundry practice and capacity and dimensions of machinery tools.

3. Weight which can be conveniently and economically handled and erected
and which is also a functiot':t of the shield-diameter as skin thickness is a func-
tion of external stresses.
4. The permissible length of the shield tail, where about two rings have to be
placed considering that excessive tail length increases resistance to advance
and renders steering more difficult.
5. By increasing the ring width,-the advance of the shield is also increased per
push and reduces the number of joints and bolted connections as possible
leakage spots.
Present practice reveals a tendency towards an increase of the dimensions,
and ring widths of 1·0 m (3 ft) may be considered as standard for tunnel diameters
up to 6·0 m (20 ft) and 0·75 m (2 1/2 ft) up to diameters of 9·0 m (30 ft).
Figure 6/ 127 illustrates the section and plan of a lining segment ring of 6·0 m
(20 ft) external diameter tube with a detailed section of a single segment. As can
be seen, the joints of the segments are staggered between adjacent rings in order
to improve both watertightness and longitudinal flexural rigidity.
A full ring is composed largely of equal segments (N), with the exception of
the crown segments (K), where the contact faces of the flanges cannot be radially
directed as this segment has to close the ring by being pushed in from inside.
As a consequence, the contacting flanges of adjacent segments (A) must have
also a corresponding special inclination on one side.
The contact faces of the flanges have to be machined to close tolerances
(±0·5 mm; 0·02 in) for fit and watertightness. This is necessary at first to estab-
lish a smooth and perfectly uniform bearing surface for the adjacent rings when
transmitting the great pressure of the hydraulic rams at the push and then to
keep the possible distortion of the ring due to unavoidable inexactitudes in erection
within the small tolerances of 2-3 cm (1 in) between shield and ring diameters.
Should a perfect bearing contact surface not be available, the relatively rigid
cast-iron material of low flexural strength would suffer hair cracks and fail under
the large non-uniform thrust.
Axial flanges and stiffening ribs of lining segments should be arranged in line
with the axis of hydraulic rams whenever possible (Fig. 6/ 128 upper section).
Practical dimensions for cast-iron segments are recommended by RICHARDSON
and MA YO, who suggest that skin thickness may vary between 25 mm and 40 mm
(1 in and 1 1/2 in) and the depth of flanges should be taken as 42 mm for each
metre of the tunnel diameter, i.e. 1/2 inch per foot of tunnel diameter. Soviet
practice is given by VoLKov 6· 2 4 in the following table (see Fig. 6/129a).

VoLK0V V. P. : Ton elli Vol. I. Gostranszheldorizdat, Moscow 1945

c _et
SINGSTAD,0.: The Queens - Midtown T111111el. New York City Tunnel Authority 1944
Trans. AMSCE, 1943 Paper No 2219
742 CONSTRUCTION AND DESIGN OF TU NNELS

FIG. 6/128. View of a large tunnel section lined with cast-iro n segments

t--
F1G. 6/ J 29 . Determinat ion of the principal dimensio ns of a cast-iron segment (Yoli..:ov)
 SHIELD TUNNELLING 743

Ring width s = 45-100 cm

(l ·5-3· 5 ft)
Depth of flange (after HEWITT)
in clayey soils m = 0·033 .D
in permeable soils m = 0·042 D
Thickness of flange a s;; 0·2 f
Inner inclination of flange i = 1/ 30--1/ 16
Diameter of bolt hole 0 = d + 6 mm
(1 /4 inch)
where d = diameter of bolt.
The skin thickness of the segment may be determined from the forceplay of
a girder subject to eccentric compression brought about by the combined effect
of external rock and water pressures and of axial thrust ( W), from pushing pressure
with fixed supports in the flanges.
The maximum stress for a unit width will be correspondingly

w p/2
(J = - 1- + 2t2 • (6. 15)

In the computation of shearing stresses the jack-thrust may be neglected and the
value obtained for skin thickness (t) is to be rounded up by a measure which will
not sensibly change the moment of inertia and extreme fibre distance of the seg-
ment.
In the determination of the whole segment section the bending and normal
loads (M, N) - as computed in the annular ring beam in correspondence with
Chapter 4 - may be considered as decisive.
Extreme fibre stresses are obtained again from the formulae of eccentric com-
pression
N My1 N _ My2
<J1 =y±-J- and <J2 =p+ - J -.

The next step is the dimensioning of bolted connections (Fig. 6/ 129b). Let us
assume that the joint will open owing to the bending moment and consequently
the full thrust is to be transmitted upon the area of the reduced width b. The produc-
ed tension force T 2 will act upon the bolt rows, its maximum action being upon
those most distant from the extreme fibre in compression.
Taken the moment upon the centre line of compression the following relation
may be derived:
(6.16)

From the moment of opposite sign, compression will be concentrated upon

the other side, and bolt rows on the opposite side will be subject to tension, whence
 744 CONSTRUCTION AND DESIGN OF TUNNELS

similarly
(6.17)

where b is taken according to Soviet practice as 6 cm (2½ in).

After the determination of the tensile forces T 1 and T 2 , for the stresses in the
bolts we get
4T
a=-d? , (6.18)
nn •
where n = the number of bolts
d = the diameter of bolts.

The merit of this computation lies in its simplicity and clarity, but its defects
are the neglect of the resistance of the inner bolt rows and the arbitrary assumption
of the working width b as 6 cm. This width b is, in fact, a function of many para-
meters, such as the specific deformation capacity, casting, the deflection of
flanges, the eccentricity of thrust, etc. which cannot be calculated exactly in ad-
vance.
Taper rings in which the circumferential flanges are not parallel are used in
metal-lined tunnels in order to follow designed vertical and horizontal curvature
and for correcting any possible deviations from the theoretical line and gradient.
These segments are not usually ribbed but have the full depth of the flanges.
For sake of watertightness the flanges are provided with dovetail-shaped machin-
ed recesses. The groves formed by these at the adjoining flanges are carefully
caulked by inserted lead wire or by asbestos thread soaked with bitumen. In the
Soviet Union expanding cement mortar has been successfully used for this purpose.
The watertight caulking of bolted connections must be provided for, and this
is effected by bituminous washers as shown in Fig. 6/130a. These lens-shaped wash-
ers consist of a thin-shouldered steel upper cover filled from underneath with hard
bitumen reinforced by an asbestos thread. When the bolt is tightened the shoulder
of the steel plate becomes flattened and the extruded bituminous material enters
the space between the hole and the bolt thread, and renders the connection water-
tight. Bolts are tightened by hand-wrenches or by hydraulic or pneumatic bolt
tightening -machines. (Tightening has to be repeated after each grouting.)
Cast-iron complies well with the requirements inherent to tunnel-lining. They
require special protection, however, against corrosion by soil (tar or ferrofixol
coatings may be used; cf. Section 46.41).
In addition to its high strength properties it has high resistance to damage
in transport and handling, and its watertightness and even its corrosion resistance
qualities are not bad.
A disadvantage of cast-iron segments is the heavy iron demand, i.e. 3·5 t/m
~ (1·27 t/ft) for a 3·6 m (12 ft) dia tunnel, 6·0 t/ m (2·2 t/ft) for a 5·5 m (18·5 ft)
 SHIELD TUNNELLING 745

_,6ro11t no/e

--
50

Jomt w!Lh haunched

biluminous washer @i
fgf'
Asbestos -bttumen l't/1

/{;1111cheo' sfee/-sh~II was/,u

Section A- A
Fm. 6/130a Bolt connections for cast-iron lining segments

dia tunnel, and 20 t/m (7·35 t/ft) for an 8-5 m (28 ft) dia tunnel. Furthermore, the
considerable weight of the single pieces (500-1500 kg) and the unaesthetic ap-
pearance of the bare surface may also be deprecated. The air is also contaminated
by the dust which settles in its corners and recesses so that a smooth surface cover-
ing is frequently indispensable.
With regard to the advantages offered by a flexible lining in the reduction of
its dimensions, and to avoid the difficulties inherent to bolted connections,
unbolted flexible jointed cast-iron segments were used in one section of the new
Victoria line (Fig. 6/ 130b). The lining of the 4·0 m (13 ft 1 inch) diameter tunnels
was composed of six segments each 2·5 cm (1 inch) in thickness for a ring width
of 0·60 m (2 ft). One end of each segment is concave and the·other convex, so that
the ends of segments form knuckle joints fitting into each other. The erection of
the segments is started from the bottom, the first two forming the invert and
offering provisional supports for auxiliary posts needed during the course of
erection. The segments have shallow interior ribs which are enlarged to form
perforated hinges regularly spaced round the periphery of the ring for handling
purposes and to provide in due course for fixing various railway appliances.
With regard to the reduced area of the flanges, wood packing pieces were inserted
between the cast-iron segments to distribute the thrust of the shield rams. In
conclusion, the application of segments without bolts offered a considerable
saving in material and a reduction of erection time of about 50 %-
It was, again, the heavy demand in iron and the desirability of reducing the
dimensions which gave rise to the idea of the utilization of more resistant steel
746 CONSTRUCTION AND DESIGN OF TUNNELS

Cross -stction
of stgmMt

FIG. 6/ 130b Knuckle jointed boltless connections for cast iron lining segments (Victoria line)

materials. The use of cast-steel was attempted at first for the production and man-
ufacture of the same type of lining segments with flanges and bolted connections
but with a reduced weight. Their application is restricted to lining station tunnels
of very large diameter or of adjacent tubes lying close beside or above each other,
or where the construction of such closely spaced tubes and an inherent stress
superposition initiated therefrom may be expected in the future.
63.232. Structural steel segments. The same endeavour was the source of the
attempt to fabricate steel segments from structural-steel. These were fabricated
from the usual rolled sections (plates) and welded to an appropriate form, or
from plates pressed to the required form - for smaller diameter tunnels. Owing
to its distortive effect, the use of welding has to be kept to a minimum and pressing
must be used wherever possible.
The fabrication of these segments is started by pressing the plates to the required
radius and bending up their transverse edges to form the rims. Rectangular steel
strips can be welded on the bottom of these rings for improved rigidity. On the
longitudinal edges flange-plates of corresponding shape and curvature are welded,
which have previously been perforated according to the spacing of the connecting
bolts. For larger segments, stiffening elements may be of special T-shaped or other
rolled sections. Fig. 6/ 131 shows the bigger section lining segment used in the
construction of the Queens-Midtown tunnel (New York) provided with welded
flanges on all four sides. 6 · 2 :;
The weight of a structural steel lining ring may be computed after RICHARDSON
from the correlation:
6·25 S1NGSTAD, Q.: Queens-Midtown T111111el. New York City Authority 1944
Trans. AMSCE, 1943 Paper No 2219
 SHIF.LD TUNNELLING 747

where W = weight in kg/ m

D = the external diameter of the ring in metres
(W 1b,rt = 7·5D 2 with Din ft).

Experience shows that in dry clay a saving of 50 % and in submerged silt one
of 65 % may be obtained in comparison with the use of cast-iron segments.
In addition, a great advantage of structural-steel lining segments is that
when all subsequent movements after placing have ceased, all joints may be

8
Section Y-Y

f5Q I
~-~~
·~1;y~ .
" ,, . &JI/; angle
vec,,1011 x-x
FIG. 6/ 131. Welded structural steel lining segment (Queens- Midtown)

closed with perfect watertightness by welded seams. Naturally, the contacting

flanges of the single elements can be provided with machined grooves - like
cast-iron segments - and caulked by lead-wire or .expanding cement in the same way.
In conclusion, structural steel segments may be somewhat cheaper than cast-
iron ones, but have two important drawbacks. The first is their very high sensi-
tivity to corrosion (except when corrosion-resistant steel is used, which involves
considerable excess expenditure). The second drawback is the insufficient exacti-
tude in their fabrication, rendering them under normal conditions not 100% suitable
for resisting and transmitting the very big pressures exerted b·y the hydraulic
rams during the push. The required tolerances within a few tenths of a millimetre
cannot be secured even by the most careful fabrication, assembly and welding,
but only by subsequent milling which increhses costs considerably.
63.233. Concrete segments. To avoid the high cost of cast-iron lining early
attempts were made to develop some type of precast concrete blocks. The greatest
difficulties encountered up to quite recent times have, again, been caused by the
required perfect and uniform contact of adjacent block rings. Even when the
considerable difficultie3 of exact casting could be overcome, there remained the
inequalities of placing from the shield tail and the difference in level (skin plate
thickness: A) necessarily resulting in a distortion of the block ring (cf. Fig. 6/82).
748 CONSTRUCTION AND DESIGN OF TUNNELS

The big ram press ures produced by the propulsion of the shield have brought
about bending pressures even with the slighte3t unevenness of contact which
led to immediate cracking - chiefly during the first few pushes.
The meas ure a nd structural importance of cracking can be considerably reduced
by the use of reinforcemen t (which is an additional item of expense) but spalling
off at the edges and hair-cracking cannot be prevented by this measure.
Another early difficulty was revealed by the erection process when it was neces-
sa ry t o find some metho d of holding the upper blocks in position until the ring
was keyed. Cast-iron seg ments were held in position by bolting to the previous
clean cuts and rings, but concrete blocks should be held by so me temporary
props until the ring (arch) is completed. A further difficulty was presented
when placing the key block. The first sta nd-
Key block ard key block could not be placed until
------------- t he shield had advanced another ring, i.e.
the pressure distributing ring was cor-
respondingly retracted. This required a
Concrete grave/ longer tail on the shield in any case. A
hole dummy key block of wood was placed
j 6ro/ltmg /Joie
temporarily in the uncompleted ring to
~Recess assure stability during the push, which
in turn necessitated that the top ram could
~ Sections
not be used, so rendering steering more
Prq;ectton
difficult. A more convenient method
is to apply small cast in place concrete
keys of rapid h a rdening cement which
become a perma nent part of the lining.
The first successful development was
-- -- -
-- the O ' Rourke block. A s sh own in Fig.
Bottom view 6/ 132 these concrete segments, 0·75 m
6rip/J~
(2 1/2 ft) wide and about 2 m (6 1/2 ft) long
have two projections and/or recesses on
FIG. 6/ 132. O ' Rourke concrete blocks
both faces. Those on the forward face
are depressed 32 mm ( 1 1/4 in) and those
on the rear edge project 38 mm l (1 /2 in). The heads of the rams bear against a n
oak cushion in the recess. These blocks are erected t o break joint a separation of
6 mm (l/4 ip) corresponding t o the above difference between the depth of
recesses and the height of projections be tween the faces. Thus, the concentrated
and more u niform transmission of all jacking stresses through these restricted
contact surfaces will b e secured.
Another interesting feature of the system is the key-block which is in t wo
pieces, the two sections h aving an overall lengt h equal to a t ypical block. After
the last piece has been placed in position a concrete pin is pushed into a groove
to lock th e two key pieces which m ay be erected separately fro m inside the ring.
The blocks are placed by erector arms which are mounted on a trailer closely
 SHIELD TUNNELLING
749

Fm. 6/ 133a Cross-section

of the water supply tunnel under the
Danube (1935) with the concrete
biocks of ERDELYI and VAJDA

FIG. 6/ 133b Its view and

the ready-
made tunnel
750 CONSTRUCTION AND DESIGN 0~ TU:-!NE LS

~-7-S'mject,on no!,

Anchor. e p,ns

Grouted bock
spice

lnject,on tube ,: - 7-5'

FIG. 6/ 133c Cross-section of the sewer tunnel in Budapest (1963)

following the shield. At the front of this trailer is a frame carrying a series of tele-
scopic steel beams (as crown bars: cf. Fig. 6/9) which are designed to afford the
required temporary support for all blo<:ks in a new ring above the springline
until the key is set and the ring is made self-supporting.
Gravel boles 38 mm (l l/2 in) dia are provided in some of the blocks for shoot-
ing gravel into the annular void left behind the tail of the shield and other holes
are arranged for grouting the 6 mm (1/4 in) interstice left between the faces
of adjacent rings after sufficient time has been allo\-ved for them to adjust them-
selves to rock pressure.
Considering that the bigger the dimensions of the lining segments the more
difficult it is to secure a uniform bearing for them, the use of smaller concrete
blocks may be advised, particularly in the construction of relatively small diameter
public utility tunnels. The system advised by ERDELYI and VAJDA and successfully
employed in the construction of the subaqueous water supply tunnel at Kaposztas-
megyer under the river Danube is shown in Figs 6/ 133a and b. These conrete
blocks 25-30 cm (I0-12 in) long, 25 cm (10 in) thick and 33 cm (13 in) wide had
 SHIELD TUNNELLING 751

curved contact faces-concave at one end and convex at the other in the direction of
the ring - which made a hinge-like action possible and rendered the lining flexible
enough to adjust its deformation to the thrust line and reduce bending moments
to a minimum. All edges of the blocks were chamfered by 2· 5 m 1(1 in) to prevent
spalling off under the heavy pressure of propulsion rams.
The blocks were cast with the utmost exactitude in accurately (±0·25 mm)
machined cast-steel moulds. The accepted tolerance was ±0·5 mm. The uniform
distribution of pushing pressures by the rams upon the relatively small blocks
was secured primarily by a special rigid st~el pressure-distribution ring and also
by the careful and accurate placing of the blocks themselves.
A similar lining was designed recently for a main sewer in Budapest (Fig.
6/133c). This work was successfully completed although difficulties had to be
overcome when the lining ring slid out from the shield and suffered a drop equal
to the thickness of the skin, which incurred at first some cracks. This was counter-
acted by cutting out a lower segment of the bottom skin so that the blocks could
be placed on the ground directly.
The great development in recent decades in the improvement of concrete quality
(careful grading, water control, steaming, vibration, etc.) has been largely respon-
sible for the use of precast-concrete segments in general practice. This is well
illustrated by the fact that a water-supply tunnel in London6 · 26 was built recently
with 15 m (6 in) thick trapezoidal blocks of the Don-Seg type (Fig. 6/134a)
similar to the T. M. blocks6· 27 (Fig. 6/ 134b). These, however, have no flat sides,
but have tongue and groove joints on all four sides.
The primary function of the shield in the London water-supply tunnel was
t_o cut a circular hole in the good London clay of an exact diameter. The protective
function inherent to the shield was of secondary importance here. The most
important requirement was to cut the hole for the precast lining of fixed diameter
very accurately._When the clay was exposed behind the shield, the pressure was
released and the clay expanded. When the lining was erected and tightened the
clay became recompressed to some extent. It was important, therefore, to establish
a correct relation between shield diameter and lining diameter making proper
allowance for the expansion of the clay and producing the required degree of
compression in the ring and corresponding pressure against the clay.
Each Don-Seg ring was 54 cm (21 in) wide and consisted of ten equal wedge-
shaped segments weighing 163 kg (360 lb) each. The longitudinal edges were
coated with a bitumen paint. Alternaie segments were placed in a ring with the
wide end against the last assembled ring and the new ring was then tightened
by using the shield jacks to wedge home the intermediate segments with their
narrow ends towards the last assembled ring. (Each segment was pushed by a

2 •26 TATTERSAL, WAKELING and WARD: Investigation into the Design of Pressure Tunnels

in London Clay. Proc. Inst. Civ. Eng. 1955 July.

CuTHBERT-Wooo: The Thames to Lee Tunnel Water Main. Proc. Inst. Civ. Eng. 1962 Febr.
6 · 27 :RICHARDSON and MAYO: Practical Tunnel Driving. ~raw Hill, New York 1941
752 CONSTRUCTION AND DESIGN OF TUNNELS

6
Axia/joint~ K e ! I
L·5Jorri t - -
Tr,msversal }oint Don-Seg type
b)~-- J9cks · v--
-· n Jl:" ":Q,

FIG. 6/ 134. Trapezoidal

concrete lining
segments:
T.M. blocks (a) The
Don- Seg type;
(b) the T. M. type
Longit udma/ S?ctio11

separate jack in the shield.) The segments were bearing tightly against the excavated
clay face and became pre-tensioned through the passive resistance.
T. M. blocks are also trapezoidal in shape but are tongued and grooved on all
four sides (Fig. 6/134b). They are laid up within the shield skin with their wide
ends alternately to the front and to the rear; in effect every other block constitutes
a key-block. The blocks P are laid first joining tightly to the last ring whereas
blocks Fare pushed by hand into the interstices. When pushing either the applied
pressure distribution ring will push all the F elements into their position at once
or the separate jacks will do so, simultaneously bringing about a tight connection
and solid bearing against the surrounding ground. In order to avoid bending in
the block, only one jack bears against a block. All the tongues are coated with
thick asphalt which acts as a cushion in the transmission of jack pressures and
is squeezed into the joints where it acts as a water-seal.
It must be noted, however, that bituminous coats as described above do not
comply with higher waterproofing requirements. In such cases a concrete block
lining must generally be subsequently completed by an additional inner water-
proofing whicli must be capable of resisting water-pressure. This inner lining
may consist of a carefully prepared bituminous paper or P.V.C. sheets supported
by an inner, cast in situ, monolithic reinforced-concrete lining to resist water
pressure; or a welded steel sheeting may be anchored to the outer block ring to
ensure its resistance against external water pressure (Fig. 6/ 135a). With adequate
 SHIELD TUNNELLING 753

Mono/it/Jic concrete dnp Precast concrete block lintn{l

wit/I Inner steel s//eetin11

Anchorage
cl3mp

Grouting

FIG. 6/135a Cast in place T rib

concrete lining with anchored 1~~- -- - + - - -- --;r{loall<:t-
steel-sheet waterproofing Cap

anchorage, the external concrete ring and the internal steel sheeting may be dimen-
sioned as a composite structure to resist the united action of external rock and
water pressures.
However, in watertight uniform unfissured clays the inner auxiliary ring may
occasionally be omitted and a thorough caulking of the joints between the lining
blocks may be satisfactory, chiefly when coupled with subsequent grouting which
has to fill partly the back-space of the excavated rock-face with the extra appropri-
ate additives (bentonite) in the grout.
Some favourable results have been obtained recently by the application of cast
in situ fresh concrete linings.6· 28 In this case a rigid and resistant steel-formwork
and shuttering has to be erected in the interior of the shield tail. The pressure
distribution rings bear directly upon the annular space filled with green concrete
(advantageously placed by concrete pumps through suitably spaced fairly large
diameter inlet holes) and the push produces perfect compaction and tight bearing
against the excavated earth-face. The erection employed, the removal and section-
length of the steel formwork, as well as the number of units and the overlapping
length with the tail skin must be in accordance with the setting time of the cement
applied . The length of the shield tail may be shortened to reduce frictional resis-
tances during propulsion (Fig. 6/135b).
A rigid formwork 'and easily dismantled shuttering is another essential part
of this construction, the use of which was given remarkable technical and economi-
cal results in Germany (G. HALLINGER) and, recently, also in the Soviet Union (the
economy obtained is more than 50 %). It is very probable that this method is
suitable only in fairly cohesive ground.
6 28
· P APROTH,E.: Herttellung eines Hauptsammlers im Tunnelbauverfahren. Schriftenri:ih
No 9, 1963

48 Szechy: The Art of Tunnelling

 .....
V,
~

S/11eld skm Pressure dlslnbutmg ring Jnst3/13t1on of' inner l'ormwork

I
,Y/,;A
. ····• ·., . . ... • ,; ........
~ - -----J- .-J-1!-. ~ 2

' - _,,.,y/AV°/((..~

\Prop11/s1on r3rn
w
=~

1/'~ "'/.<W. - -,,... .,. ~~ · ., 'I 1/..<'v"/ A'Y/ A '

8
~
;l
C:
(")
--t
0
z
>
z
0
< 0
,.
9
rn
~

~
J 4. 0
z
s~ ., 0
.,,
0

i F---"" S),"/;,{' --t

~
z
Co3rse Section re3d!I formwork "'t:;
3g_greg3tFs for grout/fig removed

F ro. 6/ l 35b Successive construc tio n stage s o f cast in place co t1crete lini ng wi th the sh ield method
 SHIELD TUNNELLING 755

63.234. Reinforced-concrete lining segments. In order to increase strength and

to reduce weight, concrete lining blocks are replaced - chiefly in larger diameter
tunnels - by reinforced-concrete lining segments. It can be stated authoritatively
that in the course of development over the last 20 years, these principal require-
ments have been fairly well satisfied; the major problem still to be solved is to
secure satisfactory watertightness both for the joints and also for the concrete
of the thinner reinforced-concrete element itself. The solution of this ·problem
has been sought in two directions.
The first trial was made in London in the construction of the Ilford line6· 29a
where ribbed-type reinforced-concrete segments were employed (cf. Fig. 6/ 138a).
This was the prototype for the experiments started at the very beginning of
the construction of the Budapest underground railways in 1951 under the direction
of K. SzECHY, L. PALOTAS and J. lLLESSY6· 29b, Fig. 6/ 136a shows the design
of one reinforced-concrete element designed for a 6·0 m (20 ft) diameter tunnel
with 1·0 x 2· 12 m (3· 3 x 7 ft) surface dimensions and 28 cm (11 in) depth. The
strength of the segments was secured by choosing a carefully graded aggregate,
by the addition of 350 kg P.C. per m3 (590 lb/cuyd) and by thorough vibration
with a consistency of 0·37--0·40 water-cement ratio. These measures provided
only for adequate strength, the watertightness of the 8 cm (3 3/ 16 in) thick seg-
ments of concrete shell against a water pressure of 12 atm was obtained only
by special steam-curing in cylindrical kilns under a pressure of 7 1/2-10 atm at
a temperature of 167 °C (270 °F).
The connections were designed similarly to those of cast-iron segments, swell-
cement caulkings and asbestos-bitumen-filled shouldered steel-shell washers
providing for the watertightness of bolted joints (Fig. 6/136a).
Test loadings carried out in a test pit with a series of vertically placed segment
rings have proved that the bolted hinge-like connections coupled with the residual
deformation of the rings resulted in a considerable decrease in the stresses, compar-
ed with the values calculated for a monolithic circular ring. The deformed shape
of the lining ring adjusted itself much more to the thrust line, as shown in Fig.
6/ 136c, thus eliminating bending stresses to a considerable extent.
This favourable behavio1,1r against transverse loading was not maintained
against' axial (jack-pressure) loads. At an average axial stress of 25 kg/cm2 (356
ib/in2) and with the secured fabrication tolerance of ±0·5 mm (0·02 in) spalling
occurred at the edges of contact surfaces and at the rear face of the lining seg-
ment. The best arrangement and amount of applied reinforcement were found by
direct bending tests in laboratory test machines (Fig. 6/ 136b).
Ribbed-type segments were also used at first in the construction of the newer
lines of the Moscow Underground (Fig. 6/ 137a) and in the extension of the Lenin-

6 · 293 GROOVES, G . L.: The Ilford Tube, Proc. Inst. Civ. Eng. 1946 5
6 ·' 9 b SztcHY. K . : Tests on Concrete Tunnel Linings, Concrete and Constructional Engineering
1959 March

48•
 -.J
V,

°'
Section A-A Section B-B
!000

8
~
-i
:,,
C
Sl
0
z
Plan Detail ofjoint
)>
z

--t4 -A-
m
18
I
+ ~ -4-
Tr,~
Asbestos-bitummous
t;:I
t;:I
Cl
~~ ™ffeMfw r-:Af;-f /, '""1"
haunclied steel- w,;s/Jer
i5
z
I
.,,0
140 -i
C
8!Lumlflous. msu/at1on z
of double slieetlflg ~
/;;
~I~ i;f:Y/MJV//41V/4?- ,;;W,IF,1//4101//41/d/#?.½--JW/27/'. -?,'%]12?d@•
f'.

[.

i
I~~ ~

le
FrG. 6/ 136a Design of an expertmental reinforced-concrete lining segme nt for a 6·0 m dia tunnel ring
 SHIELD TUNNELLING 757

FIG. 6/ l 36b Bending test of experimental reinforced-concrete lining specimen

grad Underground (Fig. 6/ 137b), whereas more recently, attempts have been
made to employ slab-type segments, too (Fig. 6/ 137c). Although the relatively
short period of their use doe~ not allow final conclusions to be drawn, a comparison
of the two types permits tl:\e following observations to be made.
The ribbed segment can be placed more readily and exactly, and, through the
rigidity of its joints, it assure., greater longitudinal rigidity in varying layers against
differential settlements. Troubles have arisen, however, in watertightness at the ribs
and at the connections. Both the roots and the bolted rims of the ribs are inclined
to fissuration induced during fabrication or erection (bolt-tightening, caulking)
processes. The smaller structural thickness of their segment skin offers less resistance
to water infiltration. In pr~sent practice, therefore, preference is given to the slab
type in somewhat greater cover depths where a uniform pressure distribution
may be expected, whereas in shallow depths the more rigid type of segments is
used.
In order to reduce the dangers of rib fissuration during shield propulsion as
well as at bolt tightening, and to reduce the time necessary for placing the bolts,
steel pins are used in Soviet practice in all axial joints and the circumferential
bolted connections are somewhat modified. Only one ~ow of connecting bolts
is applied near the inside edge of the ribs, not piercing through the dovetail-
shaped recesses provided for swelling cement cauiking "(Fig. 6/ 137d). As to the
watertigM sealing of joints, experience has sho·ivii that swelling-cement caulking
758 CONST P.UCTION AN D DESIGN OF TUNNELS

'Loadings l)eformations
2
E• 2f0000kg/Ml
C'ilse :
-ff7
· ·y/' Theoretical value
Original axis ~--...:. i Res1dwl vahe after
· .stress release
0 253
1·

®
.?-1550 kg(05atm 1 oerrin_q

Case Ii

Original axis f?es1dwl va/ue after

dress release

Measured
vahe

®
JJ•1550kg(0-5atm) perrir,g

Case Ill
C

lfeasured
T/7eorelica/ value
0040
value

Residwl va/tJl' al't,er

Ortginal value .stress release

P~J2400 kg(4atm) per dng

FIG. 6/ 136c Test-loading results of experimental reinforced-concrete lining rings

 SHIELD T UNN ELLING 759

frequently crumbles out when exposed to higher variations in temperature and

to greater dynamic effects. Therefore, experiments are being carried out with
various artificial resins and plastics (Polymers, PVC., etc.) with a view to their
use not only in the recesses, but also for their insertion in the form of sheets
between the contact surfaces. This latter application may also have a beneficial
influence on the mc,re uniform distribution of jack pressures upon the contact
faces when the shield is advanced. The contact faces of slab-type segments are
usually not planes but slightly convex or concave to ensure a hinge-like action.
Watertightness is ep.sured by the insertion of plastic sheets and a seal of plastics

Moscow type

a)

~: 7000

FIG. 6/ l 37a Reinforced-concrete lining of the Moscow Underground (rib type)

Leningrad type

FIG. 6/ 137b Reinforced-eoncrete

lining of the Leningrad
Underground (rib type)

Cross-section of

rn
segment

~[
I 1000
760 CONSTRUCTION AND DESIGN OF TUNNELS

Flat type
5500mm

c)

Cross section of segment

1000
- ---------j
F1G . 6/ l37c Reinforced-concrete lining of the Moscow Underground (flat slab type)

Nein/breed-concrete
ribbed segment Bitumen-,1sbestos
se,,/
d)

ll3undled slle!I wasller

r2 > r1
R~~i:.,__ Swelling cement or artific/4,l
resin caulking

Fm. 6/ 137d- e Details of

waterproof
connections of
reinforced-concrete
lining segments
 SIIlELD TUNNELLING 761

or of swelling cement in the recesses arranged on the inside face. To reduce toler-
ances and to obtain greater exactitude, experiments are being carried out to
fabricate the reinforced-concrete segments in suitable squeezer moulding machines
instead of casting ·them on vibrating platforms.
Precast reinforced-concrete lining segments were first used, as already mentioned,
in the construction of the Ilford line of the London Underground. 6· 2 0c The form ,
reinforcement and dimensions of these are shown in Fig. 6/ 138a. AGcan be seen,
the whole design is perfectly simil~r to the cast-iron segments, the difference being
only in dimensions. The thickness of the slab was 5 cm (2 in) and the height of
the ribs 13 cm (4 1/2 in) for a 3·90 m (12 ft 9 3/4 in) outside di.~.meter shield.
Bolt holes were lined with a. steel sleeve, and wooden plate packing with inlaid
bituminous sheeting was used between the contact surfaces for waterproofing.
It may be mentioned here that neoprene rubber gaskets and washers (under the
bolts) have been applied successfully for waterproofing the joints of a sewer tunnel
in New York. 6 29 d
Instead of this early type, precast slab-type concrete lining was used in the con-
struction of the new Victoria line. These were designed for an internal diameter
of 3·80 m (12 ft 6 in). Tests were made with segments of thicknesses varying be-
tween 11-5 cm and 22·9cm(4 l/2 in and 9 in). It is hoped that finally 15·2 cm (6
in) thick segments may be adopted. Apart from the key, the segments are left
completely unreinforced, a fact which is due both to the favourable strength
characteristics of the London clay and to the excellent quality of the concrete.
(In addition, no watersealing was necessary!)
Each tunnel ring is made up of 14 identical segments (Fig. 6/ 138b) having one
end convex and the other concave, so that they fit together with knuckle joints
similar to those for cast-iron segments (cf. Fig. 6/ 130b). When the 14 segments
are erected a gap of about 18 cm (7 in) remains at the top. This gap is filled by
a pair of reinforced-concrete folding wedges having plane contact faces and
convex and concave faces respectively in contact with the segments on each side.
The wedge with the wide end nearer to the shield is driven home by a pair of small
hydraulic jacks, while the other wedge is held in position. These two wedges
hold the whole ring firmly in place.
The segments are cast with fo,ur holes equally spaced round the interior circum-
ference. These are used for handling at the erection stage and, later, with expand-
ing bolts for fixing tunnel equipment, they serve the same purpose as the ribs
of the cast-iron segments. Handling is effected by hoists at the working face,
an expanding bolt placed in one of the holes being used for lifting. This lifting
is done by a manipulator arm mounted at the rear of the shield and held in place
by pull-out pins until the fixing wedges are driven in.
6· 290 GROVES, G. L. Tunnel Linings with Special Reference to R. C. Lining Proc. Inst. Civ.
Eng. 1943 Paper No. 5304.)
PATTENDON B.: ibid. No 5453.
Cf. Watertight Primary Linings Cut Cost of Tunelling Sewer. Constr'.lction Methods
6•29 d

I 965. Sept.
 l

--.I

°'N

28-t'mm (l~e·

J·/mm('le'}
~
"'
..,t

~
Recess for /Jolting arm ~
8
z
~
t:,
t:,
"'
Steel pipe 5z
.,,0
-i

Tnre.derf zc::
SOB mm z
¢ J1·7 mm (1¼ ') inject,on hole (2()')
rn
!;;
i
soammoo·) B1
Section A-A

FIG. 6/ 138a Reinforced-concrete lining segments employed on the Tl ford line (J 943)
 SHIELD TUNNELLING 763

No grouting is used with the tunnel lining and there is nothing between the
rings, except that wood packing pieces are inserted between the segments to
distribute the thrust of the shield rams.
The watertightne~s of reinforced-concrete lining rings may be improved not
only by an increase in wall-thickness, but also by reducing the number of joints,
i.e. by increasing the length of the single elements. This necessitates, of course,
the appiication of more powerful erection machinery.6 · 29e

Extern.;,,., ·
4? ' u1.1meter
~'S'Otnm(Wo•)
73x IJ ( 1z ·x 1/z")
1

chamfers
230
(9")
~f'
Inserts to
Jecure timber
fi4Ckings
H.1d1us ofconcave ends (,11/ segments) lOJ·2inm (8-0")
/ladius ofconvex ends (all segments) 199·3 mm (7-85")

FIG. 6/ 138b Reinforced-concrete lining segments employeq on the Victoria line (1963)

One of the latest developments in this respect is the construction of the Donner-
bGhl tunnel in Bern (Switzerland)6 ·30 where the reinforced-concrete lining ring
with an external diameter of 10· 10 m (32 ft 6 in) was composed of four main
precast segmental slab-elements 7·5 m (25 ft) long, 0·50 m (l ft 8 in) wide and
0·55 m (1 ft 10 in) thick, the weight of which was about 5 tons each. The four
main segmental elements were completed by quite a small key-stone, which was
not placed at the crown but sideways, just adjacent to the bottom piece (Fig.
6/139). This location proved to allow a more ready and exact insertion of this
important piece by pushing it backward from the face. The elements were
designed with projecting collar flanges affording temporary support from the
previous ring. ,The joints between the four main elements are arranged diame-
trically in the axis inclined at 45°, thus practically coinciding with the zero bend-
ing moment points. Careful grouting, following the erection of the rings, pro-
vided a tight contact with the adjacent soil. This was the more necessary because
the tunnel was driven through moraine-debris formations with sand layers inter-
No Bolt Linings Halve Tunnelling Time. Engineering 1961, 414
6·29C

°Cf. DESPONDS, R .:
6 •3 Der Donnerbiihl Tunnel in Bern. Schweizerische Bauzeitung 1960
10-12.
 '-I
~
.....

S2ct1{;n A- A Section 8-8 Section C--C'

Skm L• 5-50
Lir.:,7!1_ styme.7!. left
r-A Cl
' ;,iy ,,Jh;>, (l
0
z
,,"'..;
feeding C:
I,' ., 1, ') ,, tubes ~
0
~ z
~ )>
~ z
0
c::, .~ · .,: ii'orktn!J 0
· platform "'"'
ci
z
0
Cvttmg edge ""..;
C:
z
j ~ ---- z
Bi Bottom segment "'
i;;
t-A

ha , 6/ 139, Reinforced-concrete lining segments of the Donncrbiihl tunnel in Bc:·n

 SHIELD TUNNELLING 765

mingled with gravel at a very shallow depth (H ~ D ) under an urban area,

where surface subsidence had to be kept to a minimum, all of which necessitated
immediate and thorough grouting. (It is interesting to note that the contractor had
to undertake the financial risk of any surface damage and this was effected by a
contract with an insurance company.) To ensure this, not only the backspace of
the rings was grouted but thixotropic suspensions were forced behind the shield
skin not so much for water sealing, as for soil solidification, impregnation and
compaction. The application of thixotropic material is to anticipate the danger
that grouting would increase friction resistance and render propulsion more
difficult. In this way the tunnel could be driven in the heavily water-logged soil
with direct pumping only, without any excessive surface subsidence.
In very much the same way a section of the Hamburg Underground was driven
with an external diameter of 6·60 m (22 ft), where four segments 90 cm (3 ft)
wide, 35 cm (1 ft 2 in) thick and about 4 m (13 ft) long of 15 tons weight made
up a ring. 6 ·31 The segments were connected to each other in a transverse direction
by steel pins, and the rings by steel bars. An inner insulation was applied supported
by an inner monolithic reinforced-concrete ring against water pressure.
As to the dimensioning of precast reinforced-co ncrete lining segments a special
draft specification has been laid down in the Soviet Union.6·3 La According to
this draft, dimensioning should comprise the following three items:
1. Dimensioning of the full annular ring not only against external loading
when completed but also for the erection stages is required . As to the method,
one of those taking the lateral elastic soil support into account is prescribed,
preferably that of 0RLOV, or the polygonal chain method (see Section 4.33).
More exactly defined, statical control is to be effected for the two stages:
(a) erection stage before the accomplishment of the first (back-space filling)
grouting;
(b) completed stage after the accomplishment of the first grouting.
For stage (a) the specification prescribes that the supporting pattern of the ring
is to be assumed in accordance with the distribution of external packings or
blocks as provided in the erection programme. (Packings provided in the interior
of the shield skin should not be taken into account.)
Should there be no definite temporary support for the erection, then a single
knife-edge support is to be assumed below the bottom of the ring. Bending moment
diagrams for this case are given in the specification (Fig. 6/ 140a). When computing
the weight of lining the specific gravity of reinforced-concrete segments is to be
taken as 2·5 t/m3 multiplied by a safety factor of l · 1. The effective weight of erection
machinery is a lso to be increased by a factor of µ = 1· 3.
Both vertical and horizontal rock loads should be taken as uniformly distributed
when projected upon the horizontal or vertical diameter of the ring. Rings with
6 31
· Cf. MANDEL, G.: U-Bahn Bau in Hamburg, Schweizerische Bauzeitung 1959 43 .
3
Vremennye metod ukazhanya po rast. zhet. obdel. tonn. (Temporary prescriptions for
G- ia
dimensioning R.C. lining segments), Mintransstroy. Moscow 1959
766 CONSTRUCTION AND DESIGN OF TUNNELS

no axial connection between one another may be dimensioned independently,

but for those which are effectively connected to each other in an axial direction
the increase from adjacent ri ng loads transmitted by load .superposition owing
to composite deformations must also be taken into account.
2. The single segments must be checked for all st resses induced by initial and
unequal gaps resulting from deficiencies in fabrication and in erection. In addi-

- 0-12 gr 1

4)

5) 6)
f 1G. 6/ 140a B. M. diagrams owing to self-v.cigh1 "1th a single knife-edge support at the bottom
(So, ict regulations)
 SHIELD TUNNELLING 767

tion, initial gaps may be due to improper structural links and connections
between adjacent rings. According to tests, reinforced-concrete lining segments
are very sensitive to these. Concrete faces will crumble at relatively low stresses
and edges will break down readily.
The prescriptions limit the maximum width of allowable fissures to 0·2 mm in
the inner face irrespective of whether the lining is coated with a waterproo f
insulation or not.
The statical behaviour of the- ring will be altered fundamentally by initial gaps,
as stresses will not be transmitted over the whole, but only on one part of the
area of adjoining faces. In case of such a type of load-transmission, the magnitude
of bending moments (i.e. the location of the thrust line will depend not so much
on the non-uniformity of external loading as on the relative situation of the gap)
(Fig. 6/ 1406).
The eccentricity of thrust (N = normal force)
will depend upon the distribution and magnitude
of contact pressures. Assuming certain approxi-
t) ~
mations, the initial eccentricity to be considered
according to the specifications may be comput-
ed as
2)~
where
d = the distance
between· the centre line of
gravity and extreme fibre outside or the J)~
bottom of caulking recess inside (Fig.
6/ 140b) FIG. 6/140b Assumptions
uH = unit compression strength of concrete for defective
kb = homogeneity factor of concrete (kb < l) contact between
b = the width of the segments. The actual reinforced-concrete
segments
gaps due to erection and fabrication
deficiencies (e) have to be added to
this value.
The total additional bending moment will finally be
M = ± (e0 + e)N.
3. FinaJly, the deformation of t he articu lated annular ring is to be computed
with regard to the danger that if this should attain an excessive value not only
will the thrust line be dangerously modified but also the stability of the articulated
structure (which is primarily secured by the embedment itself) may be endan-
gered under yielding embedment conditions.
63.235. Pre-tensioned and post-tensioned segments. Further possibilities of devel-
opment are offered for the widespread use of reinforce ct-concrete lining segments
through the application of pre-tensioning a nd post-tensioning.
 768 CO STR UCTION AND DESIG N OF TUNNELS

i'restressin g, in genera l, eliminates ma ny of the defects of precast concrete-

block linings besides offering considerab le eco nomy in material. In addition, it
increases watertight ness a nd promotes monolithic co-operati on between the indi-
vidual rings, and brings abo ut a favourable reduction of bending effects owing
to the big initial thrust ind uced into the ri ng by prestressin g, which also prevents
an unwanted opening of the joints between the elements in contact. f n addition
to the transverse prestressing, applied first, longitudin al prestressing has, recently,
been coming into practice.
Transverse prestressin g was first realized by a n external hoop bandage after
a project prepared by FREYSSINET for a subaqueous sewer tunnel constructe d
under the Seine. 6· 3~ These tunnels of 4·15 m (14 ft) and 4· 80 m (16 ft) diametZ:r

After tens1011i11g

Concrete
f.l:Jck ''8 ''
Steei mesh ttl.oa'ge

--·-·--,- I

- - ·-
-At:<1 "O •

6rovted flyer

F 1G. 6/ J 41. FREYSS1NET

0

s prestressing method with external flat-steel hoop bandage

were driven under comp ressed air by hand. The concrete blocks J ·45 m (4 ft
10 in) long, 55 cm ( l ft 10 in) wide .and 41 cm (1 ft 4 in) thick were placed with
a rotating erector arm in the interior of the shield skin upon the previously placed
Aat steel tensio n tapes of 270 mm by 5 mm (11 in by 1/5 in) (see Fig. 6/ 141). These
were always arranged in two groups, one half group beneath the blocks of the
previo us r ing and the other beneath the rin g under constructio n. At about the
spring line two 50 ton jacks were inserted in suitable interstices left between the
two half rings. Then the jacks were tensioned unti l the stresses in the tapes attai ned

"·"" LALA ' DF, M. : Diversite des applications du bcton precontrain t. Travaux 1949 February
 SHIELD TUNNELL!NG 769

the strain required for the pre-tensioning. Then, two blocks were pushed on
both sides of the jacks and by driving suitable steel wedges at their top, the
jacks could be relepsed and dismantled and the gap D filled with concrete. In
this case, tensioning was increased to 30 kg/cm 2 (430 lb/in2), a value approximately
corresponding to the external soil pressure. The joints were then clipped and the
back-space around the tensioning groups grouted. Finally, a 3-10 cm (l-3 in)
thick gunite rendering was applied inside. This method, however, could not spread
in practice because of the difficulties and uncertainties in the control of the external
tensioning. The idea of axial (longitudinal) prestressing was initiated in the cons-
truction of subaqueous tunnels floated and then sunk in prefrabricated units. 6· 33
The most recent realizr.tion of
this idea is the construction of
the Liljeholmen tunnel in Stock-
holm, 6 · 34a where 40-85m long sec-
tions were produced from 15-25
m long precast units and sunk,
after prestressing and floating,
to the bottom of Lake Malar
to house the double track of the
Stockholm underground railway.
Another recent application of
prestressing is to be found in cer-
tain sections of the sewage tunnel
under the Rhine at Diisseldorf.6 ·34b
The precast-concrete blocks were
only used here on the embank- F1G. 6/142a Concrete tunnel lining
post-tensioned by the
ment section of the tunnel. The FlNSTERWALDER system (Dlisseldorf)
circular section of 3·80 m (13 ft)
inner diameter was composed of
four 69 cm (2 ft 3 in) wide and 25 cm (10 in) thick, high-strength (600 kg/cm 2
cube strength) plain concrete segments weighing 1·5 tons each. The ring is com-
pleted by a special key placed somewhat above the spring line. The single
elements were cast with shoulders projecting somewhat beyond the face of the
contacting joints. In these shoulders the ·tension holes were housed into which
the high strength (90 kg/cm2 = 130 000 lb/in2) prestressing bars (FINSTERWALDER
method) were anchored. One end of the pre-tensioning bar was supported on
these shoulders bearing upon a steel packing with its head, whereas prestressing
was effected by screwing the nut tight upon the thread of the free bar end. Thus,

6 33
· FOSTER, H. A.: Le beton precontraint applique a la construction des tunnels sous-
fluviaux, La Technique des Travaux 1954 January.
53
- 4a VRETBLAD' E.: Tunnelbanebro under Liljeholmswiken. Teknisk Tidskrift 1959 13
VRETBLAD, E. et al.: Yorn weiteren Ausbau der Stockholmer Untergrundbahn.
Die Bautechnik I 966. Febr.
6 34
• b Rheintunnel Dlisseldorf, Hochtief Nachrichten 1960 Aug/Sept.

49 Szechy: The Art of Tunnelline

770 CONSTRUCTION AND DESIG N OF TUNNELS

Fie. 6/ 142b Interior view of the sewage tunnel at Diisseldo rf

and tensioning forces measured accurately from inside the tunnel. Asbestos insu-
lating :.:1eets were inserted between the contacting surfaces of the segments and a
thorough groutin g applied to ensure tight contact with, and a solid bearin g upon
the surroundin g earth face (Fig. 6/ 142b).
Another economical method of tra nsverse prestressing is preloading by external
grout pressure. An additional advantage which may be claimed for this method
is that in this way a preliminary compaction of the surrounding ground and better
co-operation between the composite structural elements - tunnel lining and sur-
rounding ground - is also obtained. As an example of this, K1ESER's method may
be mentioned (Fig. 6/ 143).6· 35 As a first step, a circumferential concrete bedding
is cast in situ, and tightly compacted against the excavated rock face in order to
provide a tight connection between them and for the production of a regular and
as smooth an inner face as possible. The former purpose is served also by a prelim-
inary grouting by which all fissures, cracks and voids are filled in the rock so
increasing its strength . The lining ring of concrete segments is then erected, leaving
only adjacent elements were always stressed to one another (Fig. 6/ 142a). A great
advantage of the method was that prestressing could be calculated conveniently,

, .as Cf. Ro usSELIER, M. : Le revetement des galeries Anna/es de l'lnstitut Technique d~

B. T. P. 1952 59.
KIESER Druckstollenbau. Springer, Wien 1961.
 snu:LD TUNNELLING 771

----VJ'/1/---, £. Grout /loft? _for tl?f7- j £-"2~~,u,;4,lllt-

sion groutm_q
1- ·--·- ·--·
1· Angular s,1Ece to be
entered by tension
grouting

F ;-:;. 6/ 143. K1EsER's me1hod

of pre-tensioning

an annular gap of 3 cm (! 1/4 in) between the previously installed bedding con-
crete and the extrados of the ring. This gap is filled first through grout holes left
in the segments with a cement-mortar mix under low pressure. This operation is
always continued until the mortar appears in the next hole above. Then, a cement
suspension is forced in at a pressure of 150 % of the working load upon the tunnel
lining. The pressure is maintained until the hardening process is completed m
order to counteract slackening effects.

63.24. Application of Compressed Air ( the Plenum Process)

and Soil Stabilization in Tunnelling

Shield tunnelling is usually employed in recent waterlogged alluvial deposits

with a small bearing capacity for the construction of metropolitan subways or
subaqueous tunnels. The most difficult problem in shield tunnelling is presented
- besides the support of the front face - by dewatering the working site. Two
methods are commonly used for this purpose. The infiltrating water is either
collected in drains and canals and led to sumps, whence it is pumped off, or it is
pressed back or rather retained in the pores of the surrounding soil by increased
air oressure. (It may be noted that, exceptionally, horizontal well points may also

49•
 772 CONSTR UCTION AND DESIGN OF TUN NELS

FIG. 6/ 144. Horizon tal locks embedd ed into the bulkhea d wall

be used. These are driven and kept ahead of the front foce, in
shorcc::r tu nnel
section s.)
The precon dition for admitti ng a ny seepage of water into the tunnel
is that it
should not contain and wash out any fine particles from the surroun
ding soil.
This may be the case when water is infiltrating from the fissures
of hard clay or
from coarse grained soils. The most delicate dewate ring problem s
are encoun tered
in fine sand, and in very plastic clayey soils, and for this purpos e
the compre ssed
air method (plenum -process) is best suited. The compre ssed air workin
g chamb er
must be established just at the front face and restrict ed to a minimu
m tunnel
length backwa rds. This is greatly facilitated by the fact that in shield
tunnelling
a perman ent bearing and sufficiently watertight lining is installe
d a t once and
this may be expose d to the full design load soon after the first groutin
g. Therefo re,
the airtight bulkhe ad separat ing the comple ted tunnel section from
the workin g
area, where excava tion and lining are accomplished, may be placed
at quite a short
distanc e (some 10-20 rings) from the front. Comm unicati on and
transpo rt from
the area under compressed air are effected throug h boiler-shaped
horizo ntit air
locks (Fig. 6/ 144).
Someti mes this workin g area is divided into two compa rtments because
excava -
tion work and dewate ring at the face dema nd much higher stabiliz
ation tha n do
placing and caulkin g of the lining segmen ts. This idea was realized
at the construc-
tion of the Spree tunnel, Beilin (Fig. 6j l45) where in the front
co'llpa rtment
 SHIELD TUNNELLING 773

a o

--~ Compressed air · · ·• · · · ·•

Tube tor the conveyance
.;;·.
·· · ' or ltq11id silt
,;;:_~~ . ::-;--- ·. .. : .-..
·. . ·-"="<- -:·; ·.
·}:~:>;~ -:: <,: :_: :;,'_.-; .
".:_~=-·c::: .· ~-; ~; -.
:;· ,·.

FrG. 6/ 145. Double-stage compressed-air working chamber

a pressure PH equal to the full hydrostatic head H was maintained, whereas in

the following propulsion and tail section of the shield as well as in the adjoining
not fully completed (ungrouted or uncaulked) tunnel section a reduced pressure
of 1/2 PH to 2/3 PH only had to be maintained. This was done with due consider-
ation of the fact that in the section already supported either by the shield skin
or by the erected lining rings the essentially reduced infiltration of water cannot
possibly wash soil particles through the not fully caulked interlocks of the lining
segments and involves cave-ins. On the other hand, work under the reduced air
pressure is more efficient, less injurious to health and, in addition, assures a gradual
transition to the open air for the gangs.
Naturally, this section of reduced air-pressures has to be separated with a bulk-
head wall from the completed section, and communication and transport opera-
tions through it must be effected through horizontal air locks. This divided pressure
space has not met with a favourable reception in practice and is very rarely resorted
to, chiefly because of the double time loss due to waiting before the air locks.
The greatest problem of pneumatic dewatering in tunnel construction, in addi-
ti o n to the costs already mentioned and reduced rate of progress, is the danger
of a ' blow-out'. It must be considered that in contrast to pneumatic caisson sink-
ing, where the escape of the compressed air from the working chamber can take
place only beneath the horizontal cutting edge where it has to overcome the
maximum water head and the friction resistance of overlying earth masses every-
where equally, the pneumatic working chamber of a shield tunnel is a horizontal
cylinder, where the uniform air pressure is resisted by a smaller water head and
earth friction resistance at the top than at the bottom, the difference (m) being
in proportion with the external shield diameter. For the complete exclusion of
water the air pressure Pn in the cylindrical working chamber must be adjusted
co the full extern.al water he1d related to the bottom level (H). Consequently, an
 CONSTRUCTION AND DESIGN OF TUNNELS

,,. w.tter pressur,,

rlt1grJm

Fm. 6/146. Pressure conditions in a tunnel driven under compressed air

uplift pressure m ( = D) will be acting which will be counteracted by the resistance

against air ,seepage of the soil layers lying between the shield top and external
water level (Fig. 6/146). Should this measure not be sufficient either because of
its insufficient thickness (t) or of its insufficient air-insulating capacity then com-
pressed air will' catastrophically burst out owing to the excessive pressure from
the front working chamber (cf. Tunnel under the Elbe at Hamburg, Fig. 2/26).
The escape of the air will be followed by a sudden drop in the pressure within
the working chamber and when below the value of PH it will no longer be able to
keep the outside water pressure in balance, with the result that water will rush in
catastrophically across the face, washing all the soil into the working chamber.
As shown above (Fig. 6/ 146) the bigger the shield diameter, the greater is the blow-
out hazard. Therefore, it is always essential to check the thickness and specific
air-sealing capacity carefully, i.e. its resistance to the exfiltration of compressed
air through the covering layers above the top of the tunnel. The rate of expectable
escape must, in any case, be comparable with the supply capacity of compressed
air and if an ample safety margin cannot be assured, either an artificial increase
of the air insulating capacity of the cover (e.g. by covering the river bed with
a clay blanket or raising the air-tightness of the original cover by various grouting
processes) must be provided, or the idea of shield tunnelling must be abandoned
and other tunnelling methods (e.g. caisson sinking, cf. Section 6.31) be resorted to.
An example of this may be found in the construction of the Queens-Midtown Tunnel under
the East River, New York, where a natural cover of 3·1 m (10 ft) thickness only was available
above a 9·6 m (32 ft) dia tunnel. Considering that the submerged specific density of the air-tight
soil was 0·7 t/m 3 (48 lb/ft3 ) only, the unbalanced pressure amounted to 9·6-3·1 x 0·77 = 7 ·2
t/m 3 (450 lb/ft3), thus necessitating an additional cover from a soil material of 0 ·9 t/m 3 with
a submerged specific density of y' = 0·9 t/m3 (68 lb/ft 3 ), t = 7·2/0·9 = 8·0 m (27 ft).
 SHlELD TUNNELLING 77S

A successful example of the artificial increase of water-tightness and air-tightness

of the shallow and permeable cover was demonstra ied in the constructi on of the
reccntlyop enedDartf ord-Purfle et tunnel · a under the river Thames (Fig. 6/ 147a).
6 36

As shown in the figure, constructi on of the 8·50 m (28 ft 2 in) dia 6hield-driven
tunnel was started with a circular pilot tunnel of 3·60 m (I 2 ft) dia designed in
addition to soil exploratio n, for preliminary artificial solidification of the surround-
Oetai/ of' Soletancl!e
if!Jection tube

J) Bore CilStn§ tube e) frpandmg stopper

b) tirout !'ti/ I) Rubber sleeve
c) Perforated tube JB mm 1/J g) Perforations 5mm¢
d) ln;ect,on tube 25 mm 1/J

FIG. 6/ 147a Artificial waterproofi ng of covering layers

by combined chemical and bentonite consolidatio n (Dartford- Purfleet tunnel)

ing river ballast and fissured limeston.:- layers. This was effected by cement gro uting
pressed in through radia lly arranged injection tubes filling the larger voids of the
Thames gravel directly overlying the weathered upper 2011e of the chalk to a
thick ness of 5-15 ft. This was fo llowed by a special clay-chemica l grouting system
6, 363 DROSCHA, H .: Bau des neunen Themse-Kra ftfahrtunnels, Ba11i11ge11ie11r 1960 2 and

River Gravel Solidified for British Tunnel. Eng. News. Rea. 1957 Mov. 14; K ELL, J .: The Dart-
ford Tunnel. <zoum, Inst. Civ. Eng. 1963 March
776 CONSTRUCTION AND DESIGN OF TUNNELS

'Soletanche' carried out. from ground level (or from staging)

through a series of vertical holes spaced at 3 m (IO ft) intervals
over ihe area to be treated. Points at which the tunnel had to
penetrate open gravel with its upper face were the main objective
of this grouting for here the risk of excessive loss of compressed
air during tunnelling was the greatest, with the consequent dan-
ger of 'blows' and difficulties in maintaining sufficient working
pressure. Where gravel occurred at these levels it was grouted to
a distance of 6 m (20 ft) on each side of the tunnel, and up to a
level where this layer started over a minimum cover of 1·50 m (5
ft) of chalk. This width of cover was considered adequate ta
afford protection against failure.
First, a cased bore hole 76 mm (3 in) or mor~ in diameter was
sunk to the bottom of the gravel. Inside this a special sleeve injec-
tion tube (see detail of Fig. 6/ 147a) was placed consi5ting of a 38
mm (1 1/2 in) dia tube (c) with groups of radial perforations
arranged at intervals of about 300 mm (12 in); each perforated 'J 5 10 20 40 50 feet
section was covered by ~ flexible rubber sleeve that fitted clo~ely
around the tube.
When the sleeved tube- was inserted the outer casing tube was
withdrawn and the sleeved tube simultaneously filled with the spe-
cial grout. This did not penetrate the ground but formed a water-
tight plastic seal around it. Then a smaller 25 mm (1 in) dia tube
connected _by a flexible hose to the grout pump was inserted into
the sleeved tube, with its bottom end closed but provided in its
lower portion with a short perforated section. Above and below
th.is are expanding ·stoppers (e) which form a tight seal against
the inner surface of the sleeved tube. By raising or lowering the F10. 6/ 14b Application O' ..al consolidation
inner grouting tube its perforated end could be brought into
coincidence with the sleeved perforations of the outer tube. When grouting fluid
was pumped into the inner tube it emerged into the sp..ace between the isolating
,toppers, lifted the rubber sleeves and, rupturing the sheath of the special filling valley so
mrrounding the outer tube, flowed into the ground, filling the voids and fissures. of cover
[njections started normally at the bottom of the gravel and proceeded upwards. pressure
[n the given case the original permeability coefficient k varying between O· l m/5ec escape o
md 0·3 m/sec was reduced to 0·l-0·5x 10- 0 m/sec. high loss
Once the sleeved tubes had been installed in the ground, grouts of different became s
;haracteristics · could be repeatedly applied at any interval of time without the ficient sat
!Xpense of reboring. A further advantage is that separate injections can be made Chemi.:
1t short vertical intervals. The grout travels radially outwards from the injection cessfull}
:ube at the chosen level which makes selective applications possible suited to ground,'-
1ariations in the bedding.
A similar successful method of making airtight was obtained by the Soletanche
;oil-solidification method in the construction of the new ~imress Line of the .:.a&b Im
 776 CONSTRUCTION AND DESIGN OF TUNNELS

'Soletanche' carried out. from ground level (or from staging)

through a series of vertical holes spaced at 3 m (10 ft) intervals
over the area to be treated . Points at which the tunnel bad to
penetrate open gravel with its upper face were the main objective
of this grouting for here the risk of excessive loss of compressed
air during tunnelling was the greatest, with the consequent dan-
ger of ' blows' and difficulties in maintaining sufficient working
pressure. Where gravel occurred at these levels it was grouted to
a distance of 6 m {20 ft) on each side of the tunnel, and up to a
level where this layer started over a minimum cover of 1·50 m (5
ft) of chalk. This width of cover was considered adequate ta
afford protection against failure.
First, a cased bore hole 76 mm (3 in) or more in diameter was
sunk to the bottom of the gravel. Inside this a special sleeve injec-
tion tube (see detail of Fig. 6/ 147a) was placed consisting of a 38
mm (1 1/2 in) dia tube (c) with groups of radial perforations
arranged at intervals of about 300 mm (12 in); each perforated 'J 5 10 20
section was covered by a flexible rubber sleeve that fitted clo~ely
around the tube.
When the sleeved tube- was inserted the outer casing tube was
withdrawn and the sleeved tube simultaneously filled with the spe-
cial grout. This did not penetrate the ground but formed a water-
tight plastic seal around it. Then a smaller 25 mm (1 in) dia tube
connected. by a flexible hose to the grout pump was inserted into
the sleeved tube, with its bottom end closed but provided in its
lower portion with a short perforated section. Above and below
th.is are expanding stoppers (e) which form a tight seal against
the inner surface of th·e sleeved tube. By raising or lowering the FIG. 6/ 14b Application <-
inner grouting tube its perforated end could be brought into
coincidence with the sleeved perforations of the outer tube. When gro uting fluid
was pumped into the inner tube it emerged into the sp..ace between the isolating
stoppers, lifted the rubber sleeves and, rupturing the sheath of the special filling
surrounding the outer tube, flowed into the ground, filling the voids and fissures.
Injections started normally at the bottom of the gravel and proceeded upwards.
In the given case the original permeability coefficient k varying between 0· 1 m/sec
and 0·3 m/sec was reduced to 0· 1- 0·5 x 10- 5 m/sec.
Once the sleeved tubes had been installed in the ground, grouts of different
characteristics · could be repeatedly applied at any interval of time without the
expense of reboring. A further advantage is that separate injections can be made
at short vertical intervals. The grout travels radially outwards from the injection
tube at the chosen level which makes selective applications possible suited to
variations in the bedding.
A similar successful method of making airtight was obtained by the Soletancbe
iOil-solidification method in the construction of the new ~imress Line of the
 SHIELD TUNNELLIN G 777

Section A-A
Water/eve/
-- -- ------ - - - . -- - -· -7
I
Mud

tJne/

Log of borehole
NEJ
..al consolidation in the extension works of the London Underground

Paris Metro. This double-track tunnel was driven with a 10 m (33 ft) diameter
shield through various alluvial layers. Unfortunately, when nearing the Seine
valley some waterlogged loose permeable layers had to be pierced and the depth
of cover gradually diminished. Owing to the big shield diameter a very considerable
pressure (m) developed at the top (see Fig. 6/ 146), considerably augmenting the
escape of air upwards. To eliminate the blow-out hazard and uneconomically
high loss in compressed air, bentonite grouting was applied in the cover, which
became soaked and clogged to such a degree that work could proceed with suf-
ficient safety and with a much reduced loss of compressed air.
Chemical soil-stabilization (the Joosten process) has been repeatedly and suc-
cessfully applied in the construction of the more recent lines of the London Under-
ground,6·36b where driving tunnel shields under compressed air was rendered

.;.asb Improving London's Transport, Railway Gazette 1946 May 21-23

 118 CONSTRUCTI ON AND DESIGN OF TUNNELS

!l.'O· inlmul du. rv1111i11; tv1111tl

FIG. 6/147c Roof stabilization with silicatization

po~sible under a few metres of cover consisting of very permeable layers of sand
and gravel. In other instances, when the tunnel bad to be driven at a shallow
depth under existing buildings, their undue subsidence was prevented by prior
thorough silicatization (Figs 6/ 147b and c).
The silicatization process was applied not only to prevent the escape of com-
pressed air, but prtmarily for the stabilization of the face, particularly that of the
arch section (cf. Section 63.221). Continuous stabilization of the face is somewhat
cumbersome and appreciably retards the rate of progress, therefore the stabili-
zation of the roof is more advantageous. This can be accomplished either from
an advanced exploration drift, or where there is inadequate depth of cover directly
from the ground surface. Some examples of the latter procedure undertake n
during the extension of the London Undergrou nd are shown in Fig. 6/147b,
whereas Fig. 6/ 147c presents an example when the driving of a 3·60 m (12 ft) dia
tunnel could not be attempted directly even with the shield method because of
the small bearing capacity and looseness of the overlying cover. Therefore, a pilot
tunnel of 2· IO m (7 ft) internal diameter was driven with grout holes arranged
radially in its liner plates. Chemical grout was pressed through the holes to such
a distance that a solidified cover of 60-90 cm (2- 3 ft) was obtained above the
crown of the tunnel to be subsequently driven. In addition, chemical consolidation
was also effected from the ground surface under the foundations of heavier build-
ings to act as a curtain wall and establish a connection with the solidified layer
underneath. The extent of the solidification varies according to the level of the
underlying clay layer and the elevation of the tunnel crown relative to it· (see
cross-sections in Fig. 6/ 147c).
Should either the establishment of an additional upper airtight cover or the
artificial sealing of the pervious layers not be feasible, the inner air pressure must
be reduced so that it balances the hydraulic head relative to the top, the axis or,
 SHIELD TUNNELLING 779

occasionally, the lower third of the shield only. This latter solution is frequently
employed and, in order to reduce the blow-out hazard, water infiltrating into the
bottom section of the tunnel is given due attention. The action of an unbalanced
water head is sometimes utilized, e.g. in shields with a bulkheaded face, for increas-
ing the rate of flow of the outside material through the inlet openings into the
shield.
The resistance offered by the soil against the escape of air is also shown by the
fact that for the dewatering of the working area (chamber) it is not necessary to
provide an air pressure corresponding to the full hydrostatic head. This can be
accounted for by the fact that part of the static head is transformed to seepage
pressure subject to friction losses similar to the friction losses encountered in
the escape of the air. In this way, effective water pressure - as manifested in the
working chamber - will be less than the theoretical hydrostatic head. According
to an American practical rule about 0·227/0·3 = 0·76 i.e. 76 % of the theoretical
value is generally needed. The actual value depends on the nature of the ground
and the depth of cover.
Under exceptional circumstances it may also happen that hydrostatic pressure
values greater than the theoretical head have to be balaµced . For example, this
is so when an impervious dome-like and airtight clay blanket overlies the water-
bearing strata. Air escaping from the working area through the face is trapped
under the blanket bringing about an increase in the pressure on the water surface
(cf. the superposition of grout pressures under similar geological formations;
Fig. 6/ 118).
Naturally, the length of the sections under compressed air are to be shortened
whenever possible to reduce the compressed air demand as well as to extend
undisturbed working conditions over the greatest possible length. Its minimum
length, however, is determined by the back-space grouting and caulking operations
and by their setting time. During this latter period all preliminary work for the
rapid installation of the next bulkhead wall with the portable air locks is to be
done, so that after the elapse of the hardening period the section length of the
working area under compressed air is bordered by this new diaphragm and the
previous wall can be removed.
Compressed air is not only the most efficient means of dewatering in treacherous
soils, but it may offer considerable support for the ground and may also improve
its strength properties. This latter is mal}ifested by the action of compressed air
in expelling water from the voids of various soils, in this way reducing their water
content. As is well known, this is accompanied by an increase in their shear
strength. The degree and performance of the ground-supporting action of com-
pressed air are manifested in various ways according to the nature of the ground.
(In general, the more airtight the ground the more effective this support.) Accord-
ing to RICHARDSON and MAYO this effect is the greatest in clay which is both air-
and watertight and , therefore, pressure supports the ground to such a degree
that very little and sometimes no support of the face is required; unbalanced
pressures on the face whether high or low rarely make any difference. Even in
 780 CONSTR UCTION AND DESIGN OF TUNNE LS

tunnels driven by minin g metho ds very little or no timbe

ring is necessary. This
suppo rt offered by compr essed air will usually extend
well over the bridge action
period (Section 3.51) up to the instal lation of the perma
nent lining.
Assum ing a tunne l throug h solid clay (with a specific densit
y of l ·77 t/m 3 = 110
lb/ft3) under a cover of 9 m (30 ft) at the crown , then
the maxim um geosta tic
load will be 16· l t/m 2 (3300 lb/ft 2). Thus, an air pressu
re of l ·61 atm (23 lb/in 2)
will, theoretically, suppo rt the clay. Owing to the natura
l archin g of the groun d
(bridg e action period), in practi ce about 1/3 to 1/2 of
this pressure is sufficient
with minim um surfac e subsidence if the instal lation
of the perma nent lining is
accomplished within 24 hours .
Silt or liquid mud when first encou ntered and exposed
to comp ressed air looks
and acts like clay. It will stand on a vertical face and the
roof will be self-su pporti ng
while being trimm ed and liner plates or other tempo
rary suppo rts a~e placed.
Howe ver, silt is a pervio us soil. Any unbal anced air pressu
re at the crown drives
water back to the soil, drying the silt until it cracks and
finally spalls off. At the
same time a deficiency of pressure at the bottom alJows
water to percola te throug h
the groun d. In large diame ter tunnels, where consid erable
aifferences are encou nt-
ered between the extern al top and bottom pressures,
the lower part of the face
may resum e its satura ted mudli ke consistency and start
to flow. Owing to the
variability of pressu re and soil condi tions, consta nt
carefu l contro l is desira ble
and the actual pressu re must be adjust ed to the place
and nature of the work in
hand. A low pressu re will be used while the arch is being
lined, but when the liner
plates are set here and the upper part of the face bulkh
eaded and well niudd ed
to preve nt the escape of air, the pressu re may be increa
sed to dry up the silt in
Surface ot:·support the bench and invert . This highe r pressu re will
be maint ained until the invert is excavated and
lined.
Water -beari ng sand or quicksand is entirely
pervious and theref ore the air will not act on
the surface as in clay, but will penet rate a
certai n distan ce into the sand until a point of
equili brium is reache d where it forms the surfac e
of suppo rt somew hat in the shape of Fig. 6/ 148.
(Provided that the sand layer is of ample
thickness and of relatively small perme ability !)
The bulk of the groun d benea th or inside the
F1G. 6/1 48. Drying effect air-su pporte d mater ial is then, theoretically, all
of compre ssed air that has to be held by the tempo rary liner
in front of the plates. At the invert , the deficiency of air pressure
sh ield face will allow a flow of quick sand into the tunne l.
There fore, the lower part of the face must be
tightly breast ed and caulke d to preven t sand runnin
g into the tunnel. The
instal lation of well points in the invert sectio n may
be a very helpful asset
to count eract this tendency, and could rende r any
undesirable increase in air
 SHIELD TUNNELLING 781
pressure superfluous. (Generally speaking the employment of horizontal well
points is also important in the stabilization of the whole face.) The apparent
cohesion which is characteristi c of most fine sands also assists the face to stand
vertically for a certain period ~ hich can be prolonged by drivieg horizontal steel
rods or tubes as ' face-bolts' ahead into the undisturbed zone. Through the com-
bination of these agents it may be possible with sma ller diameter tunnels for the
face a nd top section ·to become self-supporti ng until crown plates are placed.
Coarse-grained granular sands without any cohesion are just as bad as quick-
sands and are similar in behavio'ur to gravel or ballast, constituting the wo rst
ground for tunnelling. Because such ground offers practically no resistance to the
escape of the air, it is very difficult to maintain any unbalanced pressu re at the
crown and to establish suita°ble working conditions at the face. All that can be
done is to apply ample clay-mud sealing in addition to tight breasting which
makes the work very cumbersome and the progress very slow, besides involving
the blow-out hazard with· any undue increase of air pressure above that balanced
by the water head. At a ny rate, it is advisable always to attack 011ly a small area
at any one time.
The best way to conserve compressed air in porous ground is to keep the concrete
lining close to the heading (i.e. to the shield) and to 'mud' the joints of the liner
plates. The actual volu me of air required in a tunnel to maintain the desired
pressure depends entirely on the porosity of the soil and on the number of passages
through the lock. For informative use the following rule of thumb has been devel-
oped in tunnelling practice. About 5·7 m3 (20 ft3) of free air per minute is to be
calculated for every m 2 of the face area, when estimating the probable number
and size of co mpressors.
The question of compressed air demand and cover depth has recently been
dealt with theoretically by S CHENCK and WAGNER. 6· 37
In their study they give a correlation for the determinatio n of the time delay (T),
necessary fo r drying out a soil stratum of thickness H under the action of com-
pressed air pressure p 1
l H2
T= - - - , (6.19a)
2 kPi

a nd when the initial resistance of capillary wacer is also considered

1 H2
T= - ---- (6.19b)
2 k' (Pt - hk) '

where hk = the capillary raise

k' = the actual seepage velocity.

83
• ' S C HENCK, W. and WAGN ER, H .: Luftverbrauch u. Oberdeckung
beim Tunnelvortrieb
mi t D ruckl uft. Bauteclmik 1963 2
782 CONSTRUCTION AND DESIGN OF TUNNELS

Fro. 6/149 Air-seepage conditions

Minimum tl11clrness with the application of
Qr; 1m (a) a safety blanket
of porous material
- Assumed bound,1rg (b) an airtight blanket
of seep;,ge zone against blow out

a)

Then they discuss the question of a protective cover on the river bed necessary
to anticipate blow-out hazards. They discriminate between an impervious and
a pervious blanket which might also secure temporary safety when the rate of
driving progress is greater than that of the formation of the air-percolation cone
(Fig. 6/149a). The re.quired dimensions of such a blanket, as indicated in the figure
are 6D in width and I m thickness, at least. The application of an impervious
blanket is more effective (Fig. 6/ 149b), because the escape of air will be possible
1n this case only beyond the blanket and in this way the percolation distance is
lengthened and the velocity reduced. The minimum dimensions may be given for
a top soil of 1·5 t/m3 (100 lb/ft3) dry density, as a thickness of D/2 (or after HEWETT
and JOHANNESSON at least 3·0 m ~ IO ft) over a width of 6D.
The compressed-air demand is given in first approximation by the equation

Q = 70 kiF, (6.19c)
where i = the hydraulic gradient
F = the area of face.
With somewhat more exact assumptions it is given as

Q __ w -5:._ Pi - P2 Pi + P2 -F D - hk (
D
/
m3 sec
)
y., L 2 Pi
where
Pi~ Pz =- the hydraulic gradient
 SHIELD TUNNELLIN G 783

Pi+ P2 = the mean capillary pressure

2
kh (~ ?Ok.,) = the coefficient of permeability for the air
w = a correctio n factor.
Finally, the minimum cover depth against upheaval is given as H ~ D. No
mention is made, however, about the determin ation of the depth of cover necessary
to prevent the escape of compress ed air. Some approxim ate conclusio ns about
this can be drawn from Eq. 6.19. Namely, should the ai r-percola tion time T
exceed the time period necessary to drive the tunnel over the safe length of 6D
then the danger of a blow-out is most unli kely to occur.

S
6.33. SPEClAL UN D ERG RO UN D ST R UCTURE S AND RAILWA Y STATION

The undergro und structures discussed in this chapter a re mainly built in con-
j unction with undergro und railways though they can also serve fo r other purposes
(access shafts, undergro und stores, wo rkshops, machine halls, etc.).

Section
8111>l'~ atea
------
6r. ,5urlace /?{J,!Jd
r •• • • •• , •. r. • • ·. • f1 •

·:-.-." Permeable water

· ~-:' bear.mg la,1Jcr
. ~. :·.-: . ·. ·,_ : , : , : ; . ·.

·r--- -1

',---------1
Connecting adtt
fsh;e/d:
[ NP I I
! ;
I•-sn,eld I!
I
N P.ff
L_ _ _ J I
S
I

L--':.-- -.J

P/811
Sheld chamber N!I. Sll!e/d chamber N-vI/.
FIG. 6/ J 50. Layout of undergrou nd structures necessary to start shield tunnelling
 784 CONS TRUC TION AND DESI GN
OF TUNN ELS

The con stru ctio n of any und ergr oun

d stru ctur e mus t be star ted with the
of a, gene rally , vert ical access shaf sink ing
t and from this , afte r outb reak at
level, a hori zon tal tran sver se adit of the requ ired
inde finit e leng th will be driv en. The
tion of the vari ous und ergr oun d con stru c-
stru ctur es can be star ted from this
(cf. Figs 6/ 150 and 1/ 8). The ir con drif t onl}
stru ctio n will be deal t with belo w
of sequ ence ado pted in the con stru in the orde r
ctio n of und ergr oun d railways, whe
requ ired to prov ide access to the re they are
shie ld cham bers from whe re the
star ted afte r erec tion . shields are

63.31. Access or Ventilation Shafts

A shaf t requ ired for the con stru

ctio n of a tunn el can be eith er
inclined. The sink ing of the mos t freq vert ical or
uent ly emp loye d vert ical shaf t can
out in vari ous ways dep end ing upo be carr ied
n the natu re of the grou nd as well
requ ired dep th of the shaf t (Fig. 6/ as on the
15la -f).
(a) A shaf t can be sun k in solid rock
s to any dep th, in med ium grou nds
dep ths of abo ut 10-1 2 m by the up to
min ing met hod with which timb er
stee l ring s and plan ks eith er fore pole fram es or
d or subs eque ntly plac ed are used
pora ry sup port , the perm ane nt mas for a tem -
onry walls then bein g buil t from
upw ards (Fig . 6/ 15Ia ). the bott om
(b) In cohe sive grou nd an<.i in wet,
but not wate rlog ged, gran ular soils
drai nab le by• pum ps and self- supp read ily
ortin g at a dep th of 80-1 00 cm for
at leas t 5- 6 hou rs, circ ular shaf ts a time of
can be deep ened in successive sect
leng th and imm edia tely sup port ed ions of that
by perm ane nt rings of prec ast con
or by cast -iro n linin g segm ents by cret e bloc ks
und erpi nnin g the ring nex t abov e.
is not sufficiently self-supp orti ng, If the grou nd
shaf t sink ing can be carr ied out
exca vati on, imm edia tely plac ing by part ial
the respective sect ion of the ring
whi ch is ram med into plac e (Fig . of con cret e
6/ 151 b-c) .
(c) In wate rlog ged gran ular soils
or in soft silts, shaf ts are sun k in
way as is don e with wells or ope n a similar
caissons, i.e. in a pit exca vate d to
wat er tabl e, a cutt ing edg e is asse mbl the grou nd-
ed upo n whi ch the shaf t walls are
successively incr ease d in heig ht with buil t a nd
the prog ress of sink ing while the eart
from its inte rior is con stan tly rem h mat eria l
ove d (Fig . 6/ 151 d).
{d) If the shaf t is to be sun k to a
grea ter dep th in wat erlo gge d grou
sink ing mus t be com bine d_with -ple nd, ope n
num (com pres sed air) met hod s in
allow a mor e precise rem ova l of orde r to
the grou nd and of the poss ible obst
ben eath the cutt ing edge. This , agai acle s from
n, can be acco mpl ishe d in two way
a wor king cham ber is buil t ben eath s. Eith er
the perm ane nt bott om of the shaf
exclusively for the purp oses of shaf t to serve
t sink ing, while abo ve it the wat erpr
man ent shaft-walls can be grad uall oof per-
y buil t up (e'), or the low er part of
sepa rate d from its upp er part hy a load the shaf t is
-bea ring and airti ght deck to form a
cha mbe r, whe n the shaf t is sun k wor king
to its final dep th and a wat erpr oof
inve rt arch
 C:
~
0
-0
OJ
~
~~
.8
V)
~
u
.~
u
::,

(
,._0
-0
i

t J ~
~
4
L
LJ
j
~

I
I
17 II
I
 -
....

l
"'
.:::
oj
.r::.
"'
"'
8
oj
'-
0
"'
"O
0
.r::.
~
E
C
._g
u
2
"'
C

8
...:
V)

. -
_,
.
.....'
 ACCESS SHAFTS 785

installed. This deck will be removed (e). The latter method has the advantage that
the working chamber allows a safe break out and construction of the eventual
connecting adits (see also Fig. 6/157). On the other hand, a disadvantage of this
method is that when shaft sinking comes to an end before reaching the required
depth, construction must be continued and completed by successive underpinning
(see Fig. 6/15lb), which means that the level of the airtight deck remains constant
and thus the volume of the working chamber and the compressed-air supply
required will steadily increase as excavation work progresses.
(e) Finally, if the shaft is to be sunk through waterlogged soils to great depths
preference is to be given to the artificial freezing method, where watertight shaft
walls can be built in a continuous course within this enclosure (Fig. 6/ ISlf).
In plan the shaft is to be designed generally with a circular shape in spite of
the fact that this offers less advantageous utilization of space than the rectangular
one. However, this is outweighed by far by the considerable advantages offered
in the reduction of external pressures due to arching effects in a horizontal plane. 6·38
Around a circular shaft much more efficient arching will develop in the ground
masses sliding towards the shaft when it is being sunk than around a rectangular
shaft. This arching effect besides promoting the self-supporting capacity of the
soil, so enabling it to stand with vertical walls during underpinning operations,
also reduces friction resistance, thus facilitating a more ready sinking. As access
shafts generally must attain such depths that their sinking presents the greatest
construction difficulty both advantages are in favour of a circular shape. In those
cases, however, when the shaft deepening is done not by sinking but by the method
of successive deepening under the protection of strutting frames as shown in
Fig. 6/15la, the circular shaft is no longer a requirement. Moreover, timber frames
can be installed more readily with a rectangular shape.
The material and construction of the shaft wall depend not only on its intended
purpose but mainly on the method of sinking. In the cases mentioned under
(a), (b) and (f) concrete or brick masonry made of high-quality bricks coated
with a suitable waterproofing can be used, in which case the water sealing, struc-
tural and frost-resisting capacity of the walls and the invert is adequately secured.
In case (c) prccast-concrete, reinforced-concrete or, exceptionally, cast-iron lining
segments or rammed concrete cast in situ is used, while in cases (d), (e') and (e)
the shaft lining should be made of rP-inforced concrete. Details of the shaft walls
and inverts can be seen also in Fig. 6/ 15la-f.
When a shaft has to be sunk in loose and waterlogged soils, where direct exca-
vation and accurate undercutting with direct pumping is not feasible, it often
occurs that in spite of the artificial reduction of the frictional resistances on the
sides (cf. Fig. 6/ 155) the shaft cannot be sunk to the prescribed depth. In such
a case an airtight roof has to be installed at the bottom of the shaft to allow for
the formation of a working chamber above the cutting edge. Additional sinking

•· 30 SZECHY, K .: Foundation failures. Concrete Publications Ltd. London 1961 101

50 Szechy ; The Art of Tunnelling

 786 CONSTRUC TION AND DESIGN OF TUNNELS

may then be secured under compressed air, by an exact undercut ting of the cutting
edge in the dry and by placing a surcharge on the airtight deck. As shaft sinking
usually involves an undesirable surface subsidence due to the development of
a depression cone, the application of the compressed-air method also offers
advantages in this respect by preventing the entrance of water and soil under the
cutting edges into the working chamber and thus reducing the possibility of
loosening in the surroundings. As it cannot be foreseen with safety at the beginning
of sinking operations whether the application of the compressed-air method will
be required or not, in practice, shafts are designed, when providing for any later
attachmel)t, e.g. precast roofs, in the form of an inverted cone6· 30 (cf. the dotted
line in Fig. 6/15le).
The forces acting against the shaft walls are, in general, those due to external
earth and ground-water pressures, beside which, if compressed-air sinking is
applied, internal forces may act due to a possible external surcharge of the working
chamber deck by water or piled up soil, or by both (Fig. 6/152). While the water

fl I
J'/Jal't
skin-w,;//
E I
w

Working c/J,;mber Airtight deck

FIG. 6/ 152. Loads acting upon shaft skin anc! wnrkin

chamber deck

'·39 Cf. SzECHY, K.: A pneumatik us alapozasr61 (Pneumati

c foundation s). Vfziigyi Kozl
1936 July-Sept . 38-42

--
 ACCESS SHAFTS 787

pressure can be determined definitely if the hydrostatic bead is known, the deter-
mination of both earth pressure and the vertical tensile stresses induced by friction
is, even nowadays, more or less uncertain.
The use of water for internal surcharge, though very convenient as regards
filling and removal, results in considerable internal Toadings with greater filling
heights unless offset by external ground-water pressure and, in addition, is very
sensitive to any inequality in ve::tical sinking.
On the other band, if the interaal loading is attained by dumped soil, the normal
pressure acting on the inside of both the shaft walls and the roof of the working
chamber will be relatively low, as the greater part of its weight will be transferred
only by friction like the pressures developing in silos. The horizontal press ure
can be calculated from the formula

Fy ( K~z ).
= - - - 1-e (6.20a)
Pbx Utan b

while the vertical pressure, also acting at the depth z, from

(6.20b)

where F = the cross-sectional area of the shaft

U = the circumference of the shaft
y = the dry density of the soil
K = the coefficient of earth pressure ~hich is equal to tan 2 (45° - ¢ /2) x
x tan b and b ~ ¢ /2.

The internal earth pressure acts on a horizontal ring as uniformly distributed ,

thus, inducing thrust only in the walls of a circular shaft, being not decisive upon
the dimensions of the wall thickness. On the other hand, the vertical pressure
acting on the roof of the working chamber must be considered as a standard load.
The external pressure - though its determination is more complicated - is of
about the same importance as the internal pressure for the structural dimensioning
of the shaft walls. This pressure can be considered also as uniformly d istributed
along the perimeter of the shaft, thus inducing no bending stresses but exclusively
normal stresses in its masonry walls. These normal stresses can, of course, be
safely dealt with by the masonry walls which are usually made of materials of
high crushing strength, even when built with the minimum wall thickness to reduce
the weight. The problem of their determination has an influence upon both the
structural calculation of the de!p mining shaft and of the preliminary estimation
of sinking resistance, as these depend upon the magnitude of the normal pressures
acting on the shaft walls from outside ai:id their determination is a precondition
for setting up a reliable sinking programme.
788 CONSTRUCTION AND DESIGN OF TUNNELS

According .to the classic earth-pressure theory of COULOMB, earth pressure

increases linearly with depth. At the beginning of this century it was already
realized, however, that this has no general validity. The relative displacement of
earth masses has a great influence upon earth pressure and where displacement
produces arching action in the soil qi.ass this will lead to changes in pressure
distribution, too. This is responsible for the fact that in the mining industry very
deep shafts, exceeding several hundred metres, can be safely constructed with
masonry walls of relatively small thickness. The fundamental tests carried out
by TERZAGHI have shed the first light upon this problem, indicating that earth
pressure depends also on the displacement and deformation of the supporting
wall. TERZAGHI, when calculating the active earth pressure acting on the shaft
walls, considered that the shearing resistance developed on the vertical mantle
surface of a co-moving concentric earth cylinder reduces the displacements required
for the mobilization of active earth pressure. The radius of this earth cylinder
depends upon the ratio between this shearing stress and the radial pressure.
According to TERZAGHI, the minimum external radial earth pressure6 · 40 is

(6.21
where

and n 1r0 = r, denotes the critical radius of the earth cylinder, on the mantle of
which the pressure-reducing shearing stresses are acting (n1 = P A.~ ) ; if z = 0,
2 2
then n1 =1

and

when z = co n
' 1
=J-~
A.p -2 (6.22)

(¢ 2 can be taken as equal to ¢ - sc).

KARAFIATH6· 41 made the assumption that when an elementary earth segment in

the course of shaft sinking has a tendency to slide on to the shaft wall and exerts
a pressure there, this tendency will be counteracted by the lateral earth pressure P
exerted by the earth masses adjacent on both sides (Fig. 6/ 153a). The resultant
of these lateral pressures, the force H will be directed outwards opposite to the

uo TERZAGHI, K.: Theoretical Soil Mechanics. Chapman and Hall - John Wiley 1948
208-211
6 41
· KARAFIATH, L. : On Some Problems of Earth-pressure, Acta Technica 1953 7 3-4
328-357
 ACCESS SHAFTS
789
earth pressure. Thus in the vector diagram (c) the resultant forces due
to the self-
weight of the sliding earth mass (G) and the friction resistance (Q)
acting on the
plane sliding surface are counterbalanced by the earth pressure (E)
minus the
vector H (Fig. 6/153b) .
The remaining frictional forces exerted by the adjacent lateral wedges
as assumed
by KARAFIATH will be propor tional to the earth pressure at rest origina
lly prevailing
in the earth mass. (His force diagram , however, does not satisfy the
equilibrium
condition of moments.) With these approximations he eventually
2
derives that
E = K0 m y , giving tabulat ed values for K for various ' silo
2 0 factors' m/r and
coefficients of natural earth pressur e(~); the respective values for K
0 are calcula ted
from the minimum condition of the sliding surface angle. (Experience,
however,
does not suppor t KARAFIATH's values!)

_1 __ 1~--
I
m. cot,m a:
-- - - -~-1
I
0
m '
i ---, \
- ~ - - - l " - -C(---'- -

2r • ,
1--

(§)

£ H
£ • 5. tan(cx- fJ)-H

0
H -2P31il 1-
8

5!--0
&"

m J .rm2{g- P} - sin2a
7'- 2 cos 1a,Jco.r 1 (« - P)-col3na ..rtn! ?a-Cl1
Fm. 6/153. Assumptioh of external loads acting on shaft wall (after
KARAFIA TH)
790 CONSTRUCTION AND DESIGN OF TUNNELS

SrEINFELDu2 assumed arching effect in a similar way as KARAFIATH, glVlng in

addition an equation in closed form after double integration for the ang/e a.'-of
the developing critical sliding surface or rather that of the truncated cone surface,
when arching (or ring) effect is also taken into account (Fig. 6/154a), as follows':

bJ e(l/m')
I 2 J 4 5 6 7 8 9 0 M ~ 0 #
2~ I I
-, · :2r.45-;-tl/i~m'-i--t-----r-c I i I
rr'--tr.: L-

l
4
2-BP't/m' SM~ !'5.~mmeter _
6 ,. ~ J65f/m' 0 .,, "
8 Soil character- -
'"" 1 ~ '455 l/m' . isticH/·~~; -
]'· 1-8 t/rrr ~
10
' !'..
,2 \.U I\ N
14 C-!5fii, 1/m" "'
16 t--LL
--+--ttll-lt---lrt-t+-+-t,t--t--i,...",_i~---+.--+---+---+--

~I & ·-~- • " ,,

,--- .. J \ '-,
221--ll---l......-...l,.+4+-+- -s.1 ---1-.---+-+--'!c~-
\ ,r+-
. -
~ 24
t 26 ""'z"'r1""'1m...............+aaa=F=-lt9i==l==+=l'=+~ol/m'-
~ 24 I
1 X/t--il---l....-+--+---+---+--tt--t---t--t-.--+-r-t--
~ 32 1--ll---l-½+-+---+---+---+--t---t--t- I I
~ J4 (j) lem;hi t//eor!I
~ I
~ 35 (j) Cl3s.sica/ t/Jeorf/{3fter
2 tSteinf'eld}
36 (j) [Kperimen/,,1/ 1r1'1es@!te1
4'0 tSchoklit.sch- I o//meuer)
e,... · M5t/m'
fc.,,,.; • 2-76 t/m'
FIG. 6/154. External loads acting upon shaft wall (after STEINFELD)

+ tan ix (3 -m
r
A, - 1 m) m
tan cp + 3 + - - - + 2 - = 0.
tan¢ r r
(6.23)

The greatest pressure will act at the bottom of the shaft (at depth m), its appro-
ximate value being e = y m tan (ix - ¢) , in which formula the value for a. must
tan a.
be calculated from the above equation of the third degree. In this equation ,l,
denotes the earth pressure coefficient of the lateral earth masses which, owing
to their friction, reduce the active earth pressure (thus having the same role as
the coefficient of KARAFIATH) and having - according to STEINFELD - a considerable

u 2 STEINFELD, K,: Ober den Erddruck auf Schacht und Brunnenwandungen, Vortriige der
Baugrundtagung, Hamburg 1958, 3 126 Ernst u. Sohn, 1959
 ACCESS SHAFTS 791

influence on the pressure acting on the wall (the greater its assumed value the lower
the pressure acting against the shaft). STEINFELD suggests plotting the pressure
diagram as an envelope of these maximum pressures calculated for the shaft
bottom. The bottom line of the sh!1ft can be considered as a boundary beneath
which the pressures acting against the shaft should be calculated by the following
formula:

e= - 1- [(!!!....z+ziancx- __!_z2 ) tan(cx-</>) -).,(mr z - -r z2

~ncx r r ~ncx
1
)l· (6.24)

But, in practice, it is sufficient to consider the earth pressure beneath this

boundary line as constant.
The lateral pressure coefficient depends - according to STEINFELD - also upon
both the rigid yielding of the shaft wall and the compactness of the soil, inasmuch
as with unyielding walls and loose soils its value will be low (0·5 < ,l,< 0·7),
while with yielding walls and compact soils it will have a higher value (,l, ~ 1).
According to both KARAFIATH and STEINFELD, the angle of the sliding surface (ex)
increases with depth. (This has been confirmed also by the theoretical and research
results of M. HoRN 6•43 obtained for pressures acting upon closed tunnel
faces .)
STEINFELD compared the values obtained by his formula with various lateral
press ure coefficients (from ),, = 1 to ,l, = 0·6) by a numerical example with those
calculated with the formulae ofCouLOMB and TERZAGHI and also with the actual
values observed by LOHMEYER and ScHOKLITSCH which correspond with those ob-
tained by Cou LOMB's formula up to a depth of 10-14 m below which they are con-
stant (Fig. 6/ I 54b).
The specific pressure acting against a masonry shaft ring with a unit height of
I metre can be taken as p = e + w, where the value for e is to be calculated,
of course, in consideration of both the uplift and the submerged soil density,
w being the hydrostatic head. It can be written that

pr
pr= m: , V= - (6.25)
<J

where r = the radius relative to the centre line of the shaft wall
v = the thickne5s of the shaft wall.

Transverse dimensioning of rectangular shafts must be carried out for the same
effects, but the sections must be dealt with as closed frames.
The sinking of a shaft to a greater depth is impeded by the frictional resistance
acting on the sides of the shaft due to the ~xternal earth pressure. Both this resis-
tance and the external press ure can be considerably reduced if a mantle of thixo-

5 43
- HOR N, M .: Alagutak tengelyirtinydba lzato vizszintes lzomloknyomds (On the horizontal
axial pressure acting on tunnel faces). Scientific thesis for the Master's Degree
 CONSTRUCTION AND DESIGN OF TUNNELS

feeding tvbe
&Jr/ace level

Oi.,g,;1111 of' /J!Pr,1/ l}egram of'l leral pressurt'

-pres.rvres w,th without tile action of'
tllixofropic,fnkin thixotropic fluid

liqvid
U,ixotropk
1/vid
I

Dense t111xotropic
livid, mixed with flb
fOl/Smateriel
Pressure of'
&Jndand thixotropic
grarel fluid
\
I
i
~
i
M!ter pressure ---.1
Earth pressure llfllter pressvre

FIG. 6/ 155. Loads acting when sinking with thixotropic fluids

tropic fluid (bentonite with barite suspension) is poured between the shaft wall
and the surrounding soil. This thixotropic fluid exerts a counterpressure opposed
to both earth and water pressure, keeps the soil away from the wall and at the
same time suppresses tht: formation of a subsidence cone. Moreover, as the coef-
ficient of friction between the wall and the soil is considerably reduced, the sinking
resistance is decreased correspondingly. Likewise, the pressure diagram will be
more favour~ble (see Fig. 6/155, showing one of the shafts built for the
underground railway in Budapest. The horizontal steps shown in the diagram
are due to changes in the physical properties of the soil layers encountered).
The thixotropic flµid is continually poured into the annular void formed around
the circurnfer~nce· d_uring shaft sinking, produced by a 10 cm wide outside pro-
jection at the ·bottom section of the shaft wall. 8-44
6 · " Cf. B _
ALOGH, J.: Aknak siillyesztese tixotropos anyagok segitsegevel (Shaft sinking
with the use of thixotropic fluids). Me/yep. Tud. Szemle 1954 May
LoRENZ H. : Senkkastengriindung mit Reibungsverminderung durch thixotrope Fliissig-
keiten. Bautechnik 1957 250
 ACCESS SHAFTS 793
The shaft wall must be dimensioned against axial stresses for tension and bend-
ing, encountered during the course of uneven or sudden sinking. According to
the author, the most unfavourable tensile stresses will be induced when the fric-
fional forces acting on the upper x metre high section of the shaft's mantle is
exactly equal to its dead weight, that is, is sufficient to suspend the shaft. The
most unfavourable situation may, thus, arise at the last stage of sinking (Fig.
6/ 156a), the frictional force being in this case as follows :

x2
S = U y tan2 ( 45° - <J>/2) tan<> .
2
While the weight of the total wall section will be

a;

Area ol'
a/lnular rmg

n • t;na,
H· ' n-6

I
I I
V
I
I I
FIG. 6/ 156. Approximate
computation of
I
\
I
I
I
vertical loads from
I\
friction resistance and
of bending moments
from torsion (SZECHY)
794 CONSTR,UCTJON AND DESIGN OF TUNNEµ .

from which
x2 ..
Uy tan2 (45° - cf,/2) tan<> = F0 hyb,
2
i.e.
x2
Uy tan 2 (45° - <J,/2) tan<>
2 - F 0 hyb = 0. {6.26)

Based upon observations and measurements obtained with shafts of the under-
ground railway in Budapest1 HORN and KovAcs 6·45 arrived at the conclusion that
the highest tensile force occurs during a sudden sinking and not in the static state,
but as ·a condition of the sinking movement, its value amounting to
z g-a
H= L Sc-[l'Gz+Gi]--
0 g
(6.27)

where a = the acceleration of sinking, being equal to ~ ,

AG = EG - l'Sc (the resultant of the friction forces acting against the shaft
lining)
f Sc= [.Ea .tan cf>+ he] u
0
and

h2y
E0 = tan2 (45° - c/>/2)- 2ch tan (45° - cf>/2) .
2
From this relationship the greatest value of the tensile force can be obtained
from the scheduled sinking programme.-That the value thus obtained is smaller
than those obtained previously may be partly due to the consideration_of cohesion
here.
Also, the bending stresses due to shaft-tilting during the sinking process can
be calculated again if certain approximative·assumptions are made. If tan a = 1/n
is assumed to be a measure of shaft-tilting; the maximum value of the horizontal
component deriving from the overall weight G of the shaft wall will occur at its
state of suspension (no reaction force acting on the cutting ·edge) and will have
the value H = G/n. Considering this force to be linearly distributed on the shaft
mantle proportionally to its displacement (Fig. 6/ 156b), the moment M = 2/3Hh =
= 2/3Gh tan a has to be considered and the wall thickness and the axial r ~in-
forcement have to be calculated accordingly. ·
The wall thickness (v) obtained from the forces mentioned above is generally
increased in practical design considerat_ions, the chosen wall thickness·in practice
being equal to about 1/ 12 of the shaft diameter ( v = ~ }·
u 5 HORN, M. and KOVACS, G.: SUllyesztendo miitargyak kopenyfalan fellepo htiz6erak
uj szamitasi m6dja (A novel calculation method of tensile forces produced in shaft wall in the
course of sinking), Me/yep. Tud-,· Szemle 1954 May
 ACCESS St!AFTS 795

Inclined shafts, though commonly employed in the mining industry are seldom
built in conjunction with tunnelling work. The special sinking or underpinning
methods are unsuitable for the construction of inclined shafts which are driven
exclusively by the same methods as tunnels; i.e. either by the full-face method,
by classical mining methods or by shield driving, depending upon the nature of
the ground. With small inclinations the lining of an inclined shaft can be built
with vertical masonry rings, stepped at the joints. In inclined accern shafts special
hoisting and safety equipment (car brakes, catches, cable sheaves, hoists, etc.)
must be provided, owing to the increased hazard of the running away of muck
cars. If shield driving is used, the inclined shaft is driven, in practice, from its
highest level downwards, as in this case its face will be self-supporting when
excavated at the natural angle of inner friction of the ground. If the mechanical
properties of the soil permit, inclined shield driving must be replaced, if possible,
by the full-face tunnelling method, the lining elements being placed by using
a special oblique erector arm. If the lining elements can be directly supported
from the precisely trimmed earth face, they are often placed in an opposite
sequence to the usual order (from bottom to top).
Shield driving is most frequently used in the construction of the inclined shafts
for escalators. In the upper, waterlogged loose or soft soils the ground water is
kept at bay either by compressed air or by applying the freezing method with an
oblique frost mantle as first used in the construction of the underground railway
in Moscow. 6 .46 In such cases the freezing wells, which are of a considerable length
and have an inclination of about 30°, must be drilled with the utmost care and
accuracy. A section of a completed escalator shaft is shown in Fig. 6/168.
63.32. Connecting Drifts (Ventilation Ducts)
There is no difference between the design of connecting drifts and other circular
cross-section tunnels, but as connections are relatively short, the use of shield
driving would be uneconomical, so they are driven either by classical mining
methods or by employing full-face attacks, movable steel supports and roof bolting.
The driving scheme of a ventilation tu~nei driven by the core-leaving (German)
method is shown in Fig. 6/58.
The most delicate structural parts of these connections are the out-breaks from
the shaft and the crossing and junction to the tunnel (Fig. 6/ 157). In the shaft
walls the area of out-break is lined previously with bricks or lean concrete, or
blocked by a temporary steel door to facilitate its opening and the load-trans-
mitting arches or beams above the future opening are also built-in previously.
The dimensioning of a spatial vault or of a break-through is a highly compli-
cated problem. 6.4 7 Further, such work requires special care and attention with
6 46
· Cf.TRUPAK: Zamorazhivanie gruntov v stroitelnoi industrii (Soil freezing in the
construction industry). Stroizdat 1949
• -47 ESSLINGER, M .: Berechnung von Rohrstutzen, Stahlbau 1951 Nov.-Oct; cf. also
NEMETH, F . : Foldalatti csokereszt erojdteka (Force diagram of underground pipe crossings) .
Scientific thesis.
796 CONSTRUCTION AND DESIGN OF TUNNELS

!§(}

Fio. 6/157. Outbreak of the

sides with
connecting adit

.foll solidil}e1ting .
and water tea/Ing 5i - <
grovting eflec- .:
ted /Je/'ore out-'\=' ··
break =j; ;

PtHTlping sl!c1ft

respect to the possibility of the formation of a funnel of loosened soil around and
along the shaft wall, in which groundwater and loosened soil could be drawn
down even through impervious layers. Without .drainage and preliminary support!
of this zone of subsidence a hazard of an inrush of groundwater -and run_niri.g
ground into the working place is to be expected and the outbreak may even lead
to fatal accidents. 6· 48 In practice drainage is effected by compressed air, while
the loosened soil can be consolidated with preliminary grouting through gr.out
holes installed in the shaft walls. If drainage is not done by the compressed-air
(plenum) method, grouting must be carried out with such grouting materials and
to such an extent as to provide an efficient water seal, too (Figs 6/ 151 and 6/1_57).
When a short adjoining section of the connection adit is completed it is advisable

6 48
• Cf. SzkHY, C.: Foundation failures. Concrete Publications Ltd. 1961 101
 UNDERGROUND HALLS 797

to build an airtight bulkhead (diaphragm-wall) for the installation of air locks

into this section instead of using the airtight roof in the shaft for the same purpose
(cf. Section 63.24).
When arriving at the junction point to the underground structure, e.g. shield
chamber, with the connecting adit, a load transferring spatial vault must be built
before constructing the side walls and roof arches of the underground structure.
63.33. Underground Halls (Shield Chambers, Powar and Substation Rooms, etc.)
The difference between tunnels and underground halls lies only in their purpose,
size, the degree of safety required and the magnitude of the loads to be carried
(e .g. air raid shelters). Factory and stor~ge halls, turbine chambers, etc. as well as
shield chambers are of considerable width and consequently of considerable
height so the formation of a statically favourable horseshoe section is usually
desirable. These halls and chambers being relatively short, are driven by one of
the classical mining methods or by the more up-to-date rock bolting and full-face
tunnelling methods, as the use of shield driving would be uneconomic because
of the costs and the trouble associated with their assembly being charged for a
relatively short section only. As regards rock pressures, the closed (spheroid)
shape of these structures offers definite advantages (cf. Section 32.23).
The construction of an underground hall 13·0 m wide and nearly 20 m high
driven in limestone has been described by P. M1KL6s 6.4 9 (Fig. 6/ 158). First, the
side walls were built by the aid of bottom and wall drifts (1). From the wall drifts
cuts were .made down to the bottom, where the concrete for the side walls had
been placed (2). The concrete was poured from the surface through feeding pits
specially sunk for this purpose down to the wall drifts, while mucking was done
separately through a central bottom drift. The construction' of the walls in this
way did not interfere with haulage and communication. The concreting of the
side walls up to the spring line was followed by the construction of the arch (3).
For this purpose, a central top heading was driven to a length of 6-7 m, then
widened out and concreted from the spring line upwards. The excavated soil was
dumped into cars in the central bottom drift through pits or chutes, the concrete
having been poured in again from the surface through chutes leading into the top
heading. When the concrete of the arch was set, the inner earth core was broken
out and cleared away through the central bottom heading and a central haulage
tunnel, which had previously been driven to a length of 500 m before tunnelling
was started. An underground power plant had also been built previously to supply
the electric energy required for lighting, the operation of the compressors and the
working machines.
The shield erection chambers of the underground railways in Budapest had relati-
vely smaller dimensions, but had to be driven in unfavourable ground. These cham-
bers were three-bayed, the two lateral bays being actually the shield chambers and
the smaller central one housed the transformer substation (Fig. 6/159a-c). These
6 · 49 V<ilogatott fejezetek a foldalatti vasutepites es bdny<iszati melyepites korebo/ (Selected

chapters from underground railway and mining construction). Miiszaki Kiad6, 1953 2 381
 ~
~
-!:,
~
- ---.... /
~
/ rr -·---..'-LUIIN ::::::
~
--; -"- --
, - ...L -
--/
/ /'
.,
~ ~

I
I
I
/ C::
-~
~

~-~~
I I §

I
I
LL
I
I
I _____ _
__ -
I
~~
SG
-- - ~ .-
.S - -
~
~
___ _

c::
s:,
- ·,
I
'I ~
......
~
~ ~
G I \
\ \
~ "'l
\
\
\
',
~ <ii
.c::
cS \ ~' -0
C
::,

_ \. __ - - - -...__,,...._-J~:::::!!,=Ldb::::!!:::==--- 2Oil
:;;
-0
C
::,
......
0
C
0
-~::,
----- ---------7

I
I
I
/
/
/
/
,, /
,
----
--r-- , I .,,
I~
!::

8
"'
C

I
I .I I i5
- ·,.- <,
--~-- - ·- ·-t-
I~~-
\
I
I
\ "' I ~
\ ' \
,,
-----J
\ \
\
'
"-, '
' ----..._ ______ ---- ___ __ J
 ([) Bottom Bild wall drift system
0 Secolldar!J dri/'i With upbreaks to top sediorJ
f
F

5
Combilled wall- dril't s11stem

n•~-- -- - - - -- -- - - - - ~ - ~ - - - -
Comb111ed wa/1-
dn f! Sf!stem
 I'

(!_)-

(j) Bottom ,Md w,g// drift system

(!) Second,;r!I drtft with upbreaks to top section
F
1

5
['ombmed w,1/I- drift ,~1/,~tem
 UNDERGROUND HALLS 799

shielderection chambers had overall widths of 10 m, 6 m, and 10 m, their height being

8-1 l ·5 m and their length 10 m. Their construction , like that of the above example,
was commenced by driving a peripheral bottom and wall drift system (1), from
which the side walls were built (Fig. 6/ 159a). When the side and face walls had
been built up to the level of the spring line, a second central heading was driven
out on a somewhat higher level (2), from which to break up (F) and drive separate
top headings (G) in the arch of each bay. These top beadings were then widened
out (3) and the masonry arch finally built upon the side walls from the springline
upwards (Fig. 6/ 159b).
The use of the double bottom-drift system, as well as the building of the under-
ground structure in two separate stages proved to be uneconomic, requiring too
much time and resulting in an undesirable weakening of the supporting ground
core. For this reason, the method shown in Fig. 6/ 159c was resorted to later,
the general working procedure being as follows.
Somewhat above the b9ttom level of the hall a straight liner-plate drift (2) was
driven which crossed all three bays in their complete width. From this drift
break-ups (F, F 1) were eccentrically set out and raised in the plane of the separating
walls up to the spring line. From these break-ups a peripheral wall drift system
was driven (2). From the wall drifts pits were sunk in which to place the concrete
of the side walls and separating walls (3) in bulk from the bottom upwards.
These works were followed - staggered in time - by the driving of other break-ups
(Fz) laid out on the other side of the bottom drift and driven up to the full height
of the top heading. Then from these break-ups top drifts (G) were driven out and
enlarged, and here the masonry arch was placed (4). In both cases the inverts
were placed after the core had been excavated and cleared away. The introduction
of this method reduced 'the time required for the construction of these under-
ground halls by nearly 50 %. ~.!Ml

63.34. Underground Stations and Adioining Service

and Communication Localities

Undergroun d stations, in general, represent subterranean localities of consid-

erable width and height whose structural arrangemen t and shape depend pri-
marily on their depth below the surface, on soil conditions and on construction
methods, whereas their dimensions must be determined to enable them to meet
all traffic demands. From the point of view of their structural arrangemen t it is
of ·decisive importance whether subsurface or deep-level stations are to be built.

, .so Cf. TAVASZY, F.: Pajzs nelkiili alaguthajtasi m6dszerek es rendszerek hazai tapasztalatai
(Experiences obtained in Hungary with tunnelling systems and methods without using tun-
nelling shield5), Mern. Tov. Kepzo 2752 1954
With a view to reduce the dimensions of shield chambers, the headroom left free for
erection maneuvering may be materially reduced by the introduction of special shield skin
plate splicing and erection procedures. Cf. Proc. Inst. Civ. Eng. May 1966 7 I
 a)

1. So/Jsor/'ace entrance MIi With ticket 6.Macmnec/Jamberofesca~~r

offices · 7 f/eftric siritch, r:l!a(!l/Jers ·
2. Refreshments 8. ,;,tqres /ll.r mainte11,1fl(;(1 terric'e,
J. Ttciet - qontrol fl. Diaifi9i, .svmf)$ 6117d p11mpi11; stafl~'d
4. Escalator (Stairway) ltkli'ai11·
5. Plat/'orm

b)

Subsurface si,ation
Fm. 6/160. A subsurface station: (a) isometric section; (b) ·cross-sectional arrao~ement

As to their traffic capacity the central or lateral arrangement, width ani:i acces$·
·racilities to the irplatforms is are vital importance. The decisions qnboth questions.
have to be taken at a very early stage of design and are· intimately ~rui~(¢d
with general town planning and public traffic planning. A full discussion of thes~.
factors exceeds the ~cope of this book and the following treatment is restrict~d
to a dist:ussjon of some constructional problems.
 UNDERGROUND HALLS 801

FIG. 6/161. Subsurface station hall excavated by the free-face method (Stockholm)

As a rule, subsurface stations are always built from the surface by 'cut and cover'
methods and deep-level stations by some tunnelling method (usually by the shield
method but also by some free-face or mining method). The differenc~ in construc-
tion methods has a considerable influence upon their shape, construction materials
and structural arrangement.
63.341. Subsurface underground stations. The isometric arrangeme11t and cross-
section of a subsurface underground station is shown in Fig. 6/ 160a-b. This
structure was built in an open excavation pit. The conditions and i.nethods used
for enclosure, excavation and dewatering are the same as used iu 'foundation
engineering'. The figure is intended to indie-at~ the flow of traffic, its facilities and
the reasonable arrangement and connections of the single units an<l installations
assuming that a subterranean :ticket hall is provided and direct es.;alators give
access to the centrally arranged ·platforms. Under exceptionally favouable ground
conditions combined with hea;vy urban surface traffic even subsurface stations
may be built with real tunnelling methods. For example, in Stockho~m the appli-
cation of the open-face method was assisted in some sections of the subway by the
solid granite (Fig. 6/ 161), whereas in Paris, Lisbon and Rome the he~y urban traf-
fic excluded any application of the 'cut and cover' method. Owin to the rather
unfavourable subsoil conditions and to the care which h d to be ta ~n to prevent
surface subsidence, the 'German' method (cf. Section 62.32) was ipplied here,
in principle, with some variations, however, in the construction of the arch.

51 Szechy : Tne Art of T11nn-elling

802 CONSTRUCTION ANO DESIGN ' OF TUNNELS

In Rome and Lisbon, iii firmer: ground and with a somewhat greater depth of
cover this· was effected wiih the successive transverse widening of the top-heading,
whereas in Paris a·special segment-shape shield was installed for some sections
on the top of the previously constructed· side walls and propelled forward at the
spring line to protect the erection of the precast lining segments, forming the arch
(Fig. 6/162a). Stations built by these methods usually have a lateral platform
arrangement because in this case a bigger rise can be obtained for the arch, owing
to the smaller clearance requirements at th.e sides (Fig. 6/162b).
Prefabricated elements were used on a large scale in the construction of the
stations along the· recently built extensions ·or the Moscow Subways. In these
areas the Underground lines and subterranean public utility conduits (tunnels)
were installed first, preceding the construction of residential buil9ings. It was
possible here to combine the cut and cover method with the utilization of pre-
fabricated reinforced-concrete elements very advantageously (Fig. 6/163).
63.342. Deep stations. These are split actually into two levels. The upper level
under shallow cover accommodates the ticket-hail.
From there, access to the lower platform level is obtained by escalators or lifts
each serving a single or, occasionally, both traffic directions. This lower part of
the station houses the railway ·tr11cks, platforms and access passages to and from
the · trains and escalators (Fig. 6/164a). The construction of this lower level
requires real tunnelling methods and structures differing from those common to
foundation ·engineering and therefore will be discussed in greater detail · below
(see Fig. 6/164b). . ..
As the shield method ·is predominantly used for their construction, their sec-
tional shape represents various combinations of circular sections.
The most simple· system was developed in . London. It consisted of two larger
section tubes {Fig. 6/165a). These accommodate one track and one platform each.
The running tunnels enter the larger station tunnels eccentrically at their sides,
headwalls being built at the entrance to retain the ground. The twin tunnels are
connected by· suitably spaced passages connecting the. platforms. The first of.
these transits has ample dimensions to receive the escalator. and distribute passen-
gers in both directions. The same principle was .realized in a somewhat more up-
to ·date form in Toronto where a shorter. and smaller central tube is pr.ovided
'for passenger interchange _(Fig. 6/165b).6·51a ·
A special l!,rrangement was employed in the construction of Ga:nts Hill station6·51b
during .the extension of the underground system in 1950 (cf. Fig. 6/164a), where
several ·tubes were combined. The successive stages of construction are·shown in
Fig. 6/164b a.nd · it may be seen ·that altogether 3 smaller diameter secondary
shields, 4 full diameter station shields and .o ne segmental°shield were used rendering

'"61" BARTLETT, NosKIEw1c·z and RAMSAY: Soft Ground Tunnelling for the Toronto Subway
Proc. .Inst. Civ Eng.. 1965 Sept. · · ·
f.5lb CARTY, C.; Les travatix d'extension du reseau metropolitain de Londres La Technique
de Trdvaux 1949 Sept. and: Improving London's Transport. .Railwav Gazette 1.950 March.
 j / ///,

/
-- ----- '-"--~
.,...- -- -- -
I "
-......... . ..........
Stage II.

/
/
/
/
/

/
/ ..........
'-..
"
'-..

"' '
I / '-
\.
'
I I \ \
I I \ \
I / Steel Joist gn/1,;ge platfo rm \ \

I Leaa coacrete
I
\ I $/de wall
\ .'i1dedr1/'f
\ I
\..
-- -- -- ~-!-- -- -- --
~ ~ -- -- -- +------
Propulsion ram
Stage Ill
' j

.,1-r
- -
II= m111.
W'/4%%W.d'a'1')'/4'/,@'/4)
I
I
.Segmeat-s/Japed
.r sllteld

- Steel _qr1ll2ge platform

I
I
--- --- --- t-- - - - - - - -
!..-==~ ~ - - - - - - - -r - -- -- - _L= === .~~
,, j '%'' ''
, ·:wvwr4'/47/%l/%WA%,%1/~
I /= mif7.
Stage It(
Stage V.
 S!3ge lll

Segme11t·sllaoed
/" shield

t ~~~~~----'-"-
i ::J!:= ·-=~~~~ ~ ~ ~~

/ > " . 1/7, "

«J?>V/¾',,l f-~?«~~WA9&Yd"/4?
_..%? Wd %' ~
I ,n· m,tl.
~tage W
Stage V.

,. ,, • I J"J •, {
 % _ L
'""S
-<>
°"'I ~
~ K,.'

I-I Qs
I
I
V'
I ;1
~ -·- I '- I
l _J_
5·JJ 4"/2l __ 2-00 _ s, (~I
-,--
18·/4
,- - - - - -,

Lisbo/7 - - r

Rome ..
, ~ \
__ ,·, 1-

.
~ l
'
c:::;:::::;::::====:t - - - -1-'-'.L
_________ L________ _

fc'd/7 C'OllC/'E'lE'
L~ -r- :a-12 ?·JO
~ - - - - - - ----=-22;!2~
\11
'y'
FIG. 6/ 162b. Se.;tion of subsurface stations in various cities
 Shallow subsurface arrangement of prefabricated elements

F1G. 6/ 163 Subsurface station from prefabricated in Moscow

_ /lenl!laL/0,1 sl!art
Ttcket oflic:>s C'!c

Oeep- /eve/ st,1t10,1

Fm. 6/ 164a Isometric section of a deep-level station (G~ nts Hi ll)

 ~
Y'-
l8·60
- --1

0 I

I /.g5_ _
f5l_ 40,_'0~ _j_
7J..,____ J&J_ _ ~ J.80

.l2-!;(
!J,

, ,:-i ,'. rd _,, n , /'

Ill, . (,, 16--lh Succc,,i,c cn11,1ruc1ic11 ,1agc, l,r (ianh H ill 'itation (Lnndnn)
0 -tJ I

O!HI

I
r

{D I
0 I

tltilfl/OJ /iJiJJS'
//!?M )jJ!Jf/ p31jlJJ_III
an;- - - - - - -l

-d)
_ J
09·6'/

-($}- -·-·-·-·-·--·--¢-1
 Shallow subsurface arrangement of prefabricated elements

F1G. 6/ 163 Subsurface station from prcfabrica1ed in Moscow

- Ve,1/1/aLion sld 't

TtckeL of/rct's etc

Deep -level stati on

F1G. 6/ 164a Isometric section of a deep-level station (Gnnts Hill)

 Paris 1--··· _ __ 14-/4

1 1
'/
~ o-;; ·
20
--
;
I '
~
.;,

~. I
~
I
~

J
""
-· ~l
._ 2-.00+, 407 tJ.]__ -tOZ _ _ _ Z.00 _, '"
~
18-14
I

Lisbon ---- - r

1-- 20£10_

:i,:,12 - - - ---
22-12_ - - --

FIG. 6/ 162b. Se.:tion of subsurface sta tions in va rious cities

/
 ------------ -
H-m:n.
fl.age I
Stage //

/
/
..,.....- -........
........ .........
''
/
/
/
/
/
/ ' '' '\
'\
// I/ '\\ \
I I \ \
/ / Sled JOtst grt//,;ge platfo rm \ \
I
I /!1311 CO/ICl'etC
I
\ (i"!df' dn/'t I Side wall
...
\
\
\.. _______ _(_______ _ I

'-==:~ =- L- --- --+ ---- --J === =;: 2=: :~

Stage Ill

Propu/s1011 ,ilm

Segmcot·skped
.,- S,'ltf'ld

,, ,, ?; '0> ~~i¾ l!i@ Jm ,;,2 m"J /2

Stage It~
Jf3ge V.

,r ri 171~ ~
'" ,,.,.. .

J
--'--=
s\\ ------ -~~~
-~1 --- - - ·--- --- --'·-~---.......:.....r.;.:;......--'
r 1,, 1, 162n ("o,1ru ction ~lagcs of some ,u h,urraL·~ 'L'
L"l1011, ,,r Pa ris Metro
 r--- -- ____;o::u___ __
(!)-- ---·---·---® -
- ~ _ !NO -~- -----i

-Q
r-
_ !8&'i ~
,_
.~- 6)
Vttr!/it'l/ bnrk wall ~
$/ePI saddle Steel colum/J ~
~.,
I I ~
~

CD~ ,4 0-
18·:x/
7?1 7

(!)-
7r/_) r
!8-60
-- -1
I

0
__ -r__ ''<_

- --
\9j l§O
- 400
~-·, _• .,..,

<
 18-60

SI•

,,,

?a ,/ heap' /{!;, e/
1· ,t ,. (, I 6.Jh Sun:c"i' c construct ion stages of Gan ts l-1 ill S t ation ( London)
~----===-

.
"'

Escalator of' lifts

----11r--
,,,1 f ~
llccess passB!Jl' C:

Jt
_JL_
z
0

~
~
C:
~
~
E

FIG. 6/ l65a Section of standard deep-level station in London

00
0
w
 804 CONSTI.UCTION AND DESIGN OF TUNNELS

Ccnfiru:Jus NN /is}h/ing fixfurrs · - Cat"'1]/lt

F10. 6/165b Section of deep-level station in Toronto

the construction work very lengthy arid expensive. The total width of this station
is 22·5 m (75 ft). Because of its long construction period and high cost no other
station has since been built in London by this method, but the twin tube ar-
rangement mentioned previously is µsually used in spite of its rather restricted
_space for interchange and communication.
Various types of deep-level stations have been developed during .the course of
the construction of the Moscow and Leningrad Underground systems. Here,
the intention from the very beginning was to construct spacious coherent under-
ground halls. The first type is the standard three-bay type constructed in Moscow.
The_cross-sectional arrangement and the successive construction stages are shown
in Fig:.6/166, ·a nd the complete arrangement in model scale is shown in Fig. 6/167.
As shown, construction was started here by driving the two outer shields of 9· 5 m
(311/2 ft), later 8·5 m (28 1/2 ft) diameter. These gave accommodation for the running
tracks and platfoims. The third central tube is shorter and lies in the prolongation
of the inclined escalator tunnel used not only for access and exit but also for inter-
change of train traffic. This central tube extends about half the length of the out-
side tubes and is connected to them by closely spaced lofty openings -which give
· the impression of a coherent hall unit.
'IJ).e subsequent driving of the central tube induces stress superposition in the
previously driven outside tubes (cf. Section 35.2) and to counteract undue defor-
_mation_and overstressing, horizontal ties were applied there (see Fig. 6/16-6).
(In Budapest, temporary vertical columns built of cast-iron lining segments were .
erected to strut the crown directly for this purpose.) ·
The _central tube was constructed first, by mining methods (bottom and top
heading) but quite soon a third shield of the same diameter was employed. Open-
ings to be formed between the tubes were foreseen in the design by fonning frames
or .'eyes' with special castings in the lining, from which the filling segments .could ·
be removed when a junction had to be established (Fig. 6/168).
~ -~---- ~ - . - ···

Stage I.

Stage//.
S/,art of drm'rlg of ceaLral tube

N~::--+----c;:r'\ '<
I

~fJriVl'rl.9' ofdn~s ';bando~~;-in/Jrour of

Stage (II. t/lird snield
hlater/Jroo/'i11g 1

F10. 6/ 166. Successive construction stages of a three-tube deep level station tunnel

F10 . 6/ 162. Model section of a standard

three-tube underground station
(Moscow-Leningrad)
 Stage IV
Ccnstrudion ofaccess opeomgs

folaterprool'tr,g oraccess opeamgs

,, •,,,, /1'" 'I

•bandoned in /Jvovr of .. j..,

t .sl!ield
S/3ge VI
hla(er_/Jroofi11g Lflllt:r OU/j'!/(i17g 8/Jd p/afform COl7S/~ucfiOl7
Back space /'1//;:?_!J
a/ld grouti11g

u
'
,
j

'
'

. p---
.-.,

Jeep level station tunnel

 UNDERGROUND HALLS 805

F1G. 6/ 168. Special

lining segments and frame segments around a future access opening between
adjacent tubes

The overall width of such a station was 31 ·2 m (104 ft), a very great width,
for which the bigger clearance and dimension requirements of the running lines
used in the Soviet Union are partly responsible. With due consideration to the
progressive construction time, material demand and cost which vary at about the
square of the section width a steady trend is noticeable in the development of
various structural and constructional arrangements of stations in order to conti·
nuously decrease these width requirements.
The simplest method of obtaining this decrease was the reduction of the tube
diameters from 9·50 m to 8·50 m, which resulted in an overall width of 28·20 m
(94 ft). (This solution was also adopted in the construction of the new under·
 ' '
806 CONSTRUCTION AND DESIGN OF TUNNELS

ground in Budapest.) Various columnar types baye been developed subsequently.

One type, the Ma.yakovski type, is shown in Fig. 6/169 with an indication of the
successive construction stages. .Here again,. the two large diameter outside tubes
are driven first in which· cast-iron columns and strong over-bridging purlins are
erected near the margin of the·· somewhat smaller-platforms. In succession, a third
segmental top shield is driven in the centre line of the whole section bearing against
the inside purlins ap.d columns. The installation of this segmental arch makes
the removal of .the earth core left between the outside tubes possible, as well as
the.successive demolition of the inner sides of the lining rings of.the outside tubes
themselves: Afterwards a tmiforn{platform could be constructed for all passenger
traffic and interconnection purposes. and providing a uniform coherent hall
space. .
Further development ·_was !!~cessary because the operation with · segmental
arch shield was very difficult afid the omission of the space ·below the platform
deprived the track and tunnel maintenance services of very important stores and

Stage V Stage JI/.

Fm. 6/169. Successive construction stages of a three-bay columnar-type station (Moscow)

 UNDERGROUN D HALLS
807

F10. 6/ 170. Modified columnar-tvoe station (Moscow, Arbat)

operation rooms, which ought to be provided in all cases and which had been avai-
lable in the previously introduced three-tube station arrangements (see Fig. 6/ 166).
A more recent and somewhat modified type of columnar station is shown in Fig.
6/ 170 (Moscow, Ar bat) where purlins are replaced by cast-iron arches over-
bridging the openings. Another new feature in the construction of this type was
that the segmental arch was not driven by the shield method but by free-face
mining methods, the lining segments being placed by erector arms. The arch had
a bigger rise to reduce horizontal thrust. The continuous platform was dismantled
in the same way as before.
The development of the Leningrad Metro stations indicates some deviations
from the above ideas. In the columnar type station 'Kirov', the columns and over-
bridging purlins were erected within t he central full diameter circular tube, the
shield of which was driven first. Then two outside shields (5·50 m dia) followed
with truncated plane faces towards the central tube, bearing at the truncated top
and bottom lines on the support offered by the inner row of columns. The- fairly
watertight solid clay allowed the contacting parts of the lining of the central tube
to be dismantled within this shield which was practically open towards this
lining, without serious difficulties in retaining the ground and in dewatering. 6· 52
A more recent development of this idea is shown in Fig. 6/ 171 which represents
the ' new type' station. This allows for a further reduction in width to 18·33 m
(61 ft) although assuring the same clearances for running tunnels as well as for
the entering escalator. The column row, however, leaves no free platform space

C-S% LIMANOV, Ju. A.: Metropoliteni (Subways). Transheldoriz dat, Moscow 1960

- ----- ----- ----- ---

 808 CONSTRUCTION AND DESIGN OF TUNNELS

l?ein/lm:etl-roncrete
TT-- - - - - -~=::;;ail!E!I·~~ linin;se!1Tlent

1832·6

FIG. 6/ 171. Section of new type station (Leningrad)

for the clearance of the trains and therefore a moving gate system closes the open-
ings in the column row which can be automatically operated from the train only.
When the train stops these moving gates are opened simultaneously with the car
-doors. This arrangement also provides for the store and administration rooms
necessary.for the maintenance services below the platform (see Fig. 6/ 171).
· At the construction of the deep-level stations for the new under.ground line in .
:Budapest the· three-tube system: (cf. Fig. 6/166) ·was employed first with 8·5 rri
(28 ft4 in).dia tubes and with a total widthof28·2Qm (94ft). Recently,a newlicens.:·
ed five-bay station type has ·been developed (Fig. 6/172) based on existing or pre~
:viously built outside running tunnels . (5·5 m; 18 ft 4 in dia). The successive
co11;struction ph_a~es are numb~red on the figure as follows: ·
1. Outbreak through the previously built composite concrete and inner rein-
.forced-concrete or cast-iron segment lining at the place of the main underpinning
beam at the bottom in 4·0 m (13 ft) long staggered sections followed by the resto-
·ration of the support for the lining ring by establishing a lean concrete junction
:to ·t he cast iron ring at the bottom beam.
2. Placing of structural steel columns upon the underpinning bottom beam
sections spaced at 4 m with up-breaks through the lining.
3. In situ construction of the main longitudinal top beam under the protection
of a brick-lined longitudinal drift produced by manual mining methods.
4. Driving of bottom and shoulder drifts for the construction of side walls
cast in 4·0 m (13 ft) long staggered sections.
· 5. Excavation and construction of the arch of the adjacent bay housing the
passenger platform using mining methods (flying arch o·r Kunz method, see
Sections 62.31 and) 62.44).
 r~~~- 31 lfJCh Im
distance

-- -- ---·
~
"'Cl
"'
-4===~~~~~-1 ------~~~ "'g
zt,
:i:
>
t"'
f;;

Existing tunnel ltmng with

inner monolithic reinforced-concrete ring

~l 2'70 l W l I
' . ' ' 2:J!l75 -+

Fm. 6/ 172. New five-bay type station of the Budapest Subway (licensed by R6zsA and ENYEDI)

00
0
ID
 00
.....
0

~rel
lt.l''
\1Ped
1rc

()
0
z
1/l
..;
;<J
C
()
..;
0
z
>
z
Cl
Cl
tT1
1/l
Cl
z
0
'11
..;
C
z
z
tT1
r-
1/l

\ Dramage pump

FIG. 6/ 173. Section of escalator Tunnel

.... -- -

SHALLl:JW URBAN AND PUBLIC UTILITY TUNNELS 811

6. Excavation and construction of the arch of the central bay for traffic inter-
change also by the above mining methods.
7. Removal of the earth core from the central bay.
8. Removal of tll.e earth core from the platform bay.
9.-10. Excavation of the invert section of platform and of central bays for
placing invert concrete and constructing platform.
All work was executed under compressed air and in varved clay. The method
has some similarity to the construction procedure of the Gants Hill station
(Fig. 6/164b) but all auxiliary drifts were driven bv mining methods instead of
shield-driving.
In addition to the stores and administration rooms located below platform
level all stations must be equipped with other accommodation rooms such as cabins
and lavatories for the attendants, sewage sumps, punwing stations, exhaust
ducts, transformer rooms. electric transformer substations, ventilation ducts
(cf. Fig. 4/ 140) and accommodation for emergency generating units. compressor
houses, etc.
Modern deep-level stations are provided with escalators. Wherever possible,
escalators should be arranged to convey passengers from street level to platform
ievel. Escalators are accommodated in inclined tunnels adjoining at both ends
to special machine chambers provided for the operation and control of the escala-
tors and accommodating all driving or stretching machinery. Heavy concrete foun-
dations support the escalator trusses which, in turn, give support to the moving
stairways with all their equipment (Fig. 6/ 173 and cf. Fig. 6/ 167).
Since modern escalators have a slope of 30° from the horizontal, special emphasis
must be placed upon the accuracy of tunnelling work

6.4. URBAN AND PUBLIC

UTILITY TUNNELS CONSTRUCTED AT A
SHALLOW DEPTH

During the development of big cities the need for tunnels is continually increas-
ing. These tunnels are, as a matter of course, constructed at shallow depths, lying
directly beneath the pavement, l)lostly in loose, soft, and frequently waterlogged
soil layers. Their main purpose is to reduce surface traffic by providing a separate
route for vehicles and pedestrians, or to avoid interference with traffic when
maintenance work has to be done on the public service lines and other conduits
accommodated in them. The special conditions and demands encountered have
led to the development of special construction methods employing, more or less.
the elements of the cut-and-cover principle and making progressive use of precast
reinforced-concrete structures in order to speed up the work.
 ----- -

812 CONSTRUCTION AND DESIGN OF TUNNELS

6.41. SIDE-WALL CONSTRUCTION IN WALL DRIFTS

When building tunnels at shallow depths under .roads with heavy traffic, this
modified version of the tunnelling method described in Section 6.31 is generally
applied. The side walls of the tunnel are built in horizontal drifts instead of vertical
pits, which may be driven from some laterally situated pit with small dia shields
or with common mining methods. ·Then the road pavement is lifted and the prefab-
ricated roof~elements ·are placed. Should the removal of the road pavement have
to be completely avoided even for short night periods, the method shown in Fig.
6/162a (Paris Metro), i.e. a segmental shield travelling on the completed side
walls has to be· used.

6.42. SPECIAL METHODS FOR THE CONSTRUCTION

OF PEDESTRIAN SUBWAYS, HIGHWAY AND PUBLIC SERVICE LINE
UNDERPASSES

The first step in reducing the heavy traffic at cross-over points in big cities is
the elimination of pedestrian level crossings by the construction of subwayn
A further measure is to transfer one of the cross roads or at least one sectoo.
(tramway tracks) of them under the ground surface. Similar measures can ailss
be taken at other multiple junctions of heavy traffic (bridge approaches, etc.).
In the latter cases, however, level crossings can also be eliminated by means of
elevated viaducts (as e.g. in Stockholm).
In busines~ areas with dense traffic it is generally attractive. to accommodate
small shops and refreshment rooms in pedestrian subways (cf. Fig. 6/174).
Pedestriaµ subways though mostly needed under metropolitan cross-over
points are often built, e.g. under watercourses, railway stations or other big
traffic obstacles in preference to bridges to establish safe and uninterrupted -traffic.
The platforms of up-to-date railway stations or the bigger stations of underground
railways can usually only be reached by way of pedestrian subways which, there-
fore, are built not only to secure safety for pedestrian traffic or for the safe avoi-
dance of railway traffic but also to provide for direct and uninterrupted access
to traffic centres from several places. As these structures have their peculiarities
in construction and design, it seems justifiable to deal with them separately.
Design and construction methods of subways are similar to those used for
subsurface public utility tunnels, differing only in cross-sectional dimensions
and in water-sealing methods. The following remarks may therefore be relevant
for the construction of such tunnels.
Pedestrian cross-over subways can be constructed by one of the following
four methods:
.1. Cut-and-cover method, when the subway is built in sections in open trenches,
and surface traffic is simultaneously diverted to parallel streets or restricted alter-
.... - - -- -

SHALLOW URBAN AND PUBLIC UTILITY- TUNNELS 813

nately to smaller strips of the route. (For example, the construction of the sewer
syphon in Deak Square, or the temporary loop line of the subsurface underground
railway in Bajcsy Zsilinszky Street, both in Budapest.)
2. Cut and cover method, with partial diversion of -surface traffic and partial
underpinning of certain traffic lanes (tramway) by provisional bridges combined
with the brief closure to total traffic in the hours of least traffic density (cf. Fig.
6/175).
3. Intermittent cut and cover method with the application of prefabricated
roof elements placed on the side walls built previously either in underground
drifts driven from the sides by mining methods, or in narrow trenches cut from
the surface, or constituted by sheet piling driven from the surface. Traffic restric-
tions are limited to narrow strips and streets are closed only for short periods
during construction (cf. Figs 6/176 and 6/177).
4. Underground tunnelling by mining methods or, most often, by shield
driving without interfering with surface traffic (pedestrian tunnels in Antwerp,
Hamburg and some main sewer tunnels in various cities, e.g. Essen, Hanover,
Budapest).
Of the methods, mentioned above, that under 1. is the most satisfactory, where
it is possible. Should there be no place for the required diversion lane for traffic,
either because of the narrowness of the street, or because of the importance of
unimpaired maintenance of traffic (tramway tracks), then the traffic lanes to be
maintained undisturbed must be underpinned by previously installed bridge struc-
tures, and the subway has to be built by the cut-and-cover method.
The use of load-bearing prefabricated elements are extremely important when
building in metropolitan areas with dense traffic. In the construction of subways
built under access ramps at the approach of bridges or, generally, at places where
the reduction both of construction time and the restriction period for traffic
are of outstanding importance, the method under 3. must be applied. In such cases
not only must the provision of prefabricated side wall and arch elements be attend-
ed to, but arrangements must also be made to place these elements in the shortest
possible time in the brief periods of traffic blocks tolerable in the hours of low
traffic density. Side walls can be constructed in sections either in open trenches
deepened from the surface and filled with concrete cast in situ or with temporary
sheet piles which may eventually form permanent elements of bo-th lateral and
vertical supports.
If the lifting of the road surface in strips is restricted only to the time for placing
the roof elements, the side walls may be built up in drifts started from pits, sunk
laterally from arbitrary areas not affected by traffic.
The prefabricated roof elements are always placed on the previously erected
side walls in the shallow trench strips excavated from the surface during the partial
or complete traffic stoppage in the night hours.
Finally, if a traffic stoppage is quite out of question even for the shortest
time, the whole subway must be constructed either by mining methods from
814 CONSTRUCTION AND DESIG N OF TUNNELl>

drifts started from the sides (e.g. by means of the Cologne method, see Fig.
6/71) or, when the subway or public service tunnel is to be constructed at a shallo"
depth or in a subsidence-sensitive environment, by the shield method.
The above-mentioned constructions are usually drained by direct pumping
from sumps, although in finer grained soils or silts ground-water lowering can
also be resorted to. The soil stabilizing effect of this latter method offers consider-
able advantages for the support of open trenches or tunnels driven with mining
methods. The compressed-air dewatering method is seldom employed in the con-
struction of subways, as they are located at a shallo·w depth below the surface,
and consequently a ' blow out' could be prevented only in the exceptional case
when the waterlogged layer would be overlaid by an air- and water-tight cover of
suitable thickness.
The construction methods referred to under points 2. and 3. will be introduced
by the presentation of some practical examples.

64.21. Subways in Vienna

In the years 1955-1956 a pedestrian subway system was built in Vienna under
the cross-over point of the Kartnerstrasse and the Ring in front of the Opera House,
with accommodation for underground shops and refreshment rooms. 6 · 51 According
to the previously made traffic census these extremely heavy traffic routes were
crossed in peak hours by about IO 000 pedestrians, interfering with the passage
of about 3200 vehicles. The pedestrian level crossing was not only the cause of
considerable time losses to traffic but a lso a constant source of accidents (e.g.
in 1954 it headed the relevant statistics with some 80 accidents).
The reinforced-concrete underground structure built to relieve these troubles
is elliptical with axes of 51 m and 56 m. Jts flat reinforced-concrete roof is supported
by soild walls both on its periphery and at its centre. where two concentric rows
of columns suplpy the intermediate supports (Fig. 6/ l 74a). The height of the hall
is 2·90 m (9 ft 8 in). while that of the acce% galleries is 2·57 m (8 ft 7 in), and the
whole underground area is artificially ventilated . Seven short flight escalatorl-.
lead to the tramway stations and to the footpaths of the intersecting streets.
Refreshement rooms and shops installed in the interior of the structure are attrac-
tively fitted out and decorated .
The subway was built in an open trench . The tramway tracks crossed the trench
at a level elevated 60 cm over the street surface and underpinned by temporary
steel beams and trestles (Fig. 6/ 174b). The reinforced-concrete framework and
cover of the structure together with all roads and structural installations, etc.
were completed in about seven months by uninterrupted work day and night
to reduce the period of serious hindrance to public traffic to an absolute minimum .
Internal coating, fitting out and decoration proceeded afterwards assisted by arti-
ficial drying under the completed roof structure.
6 53
· KOLLER, R.: Massnahmen der Stadt Wien zur Behebung der Verkehrsnot, dst. Ing .
Zeitschrifr 1958 0k t. und : Die Erweiterung des Wiener Stadtbahnnetzes. Stadt Wien
1965 o kt.
 SHALLO W UR BA N AN D PUBLIC UTl LlTY TUNNELS 815

FIG. 6/ l74a M odel of the Opera Subway in Vienna

.r

Fm . 6/ 174b Opera Su bway under construction in Vienna

816 CONSTRUCTION AND DESIGN OF TUNNELS

Since then a series of underpasses have been built in Vienna, making use of the
lcos-VEDER method (cf. Fig. 6/75) for lateral support of open excavation pits.
The most remarkable of these might have been that at the Mariahilfer Strasse
where a working pit of several thousand sqm surface area had to be covered by
a temporary deck construction in order to maintain street and tramway traffic
undisturbed during the whole construction period. s.st
The Icos-VEDER method has been extensively used also at the realization of the
Park-Lane improvemen t scheme London. • a A typical cross-section of these
6 54

vehicular underpasses is shown in Fig. 6/ 175.

... .

..
~

....
·.·'
23' o· carriageway
1f dia. 9"dia. C.I. pipe

4• dia. fire main

28'3"
.·.:

!COS wall

FIG. 6/ 175. Cross section of a vehicular underpass of the Park-Lane improvement

A remarkable subway system in Brussels in conjunction with the Internationa l

Exhibition in 1958 was constructed with interesting construction al features .
6 51
The reader, however, is referred to reference · b owing to lack of space in this
book.

u 4
cf. Bauingenieur 1966 1; Der Aufbau 1966 1-2
6 ·5 , . E.: Park-Lane Improveinent..Scheme: Design and Construction. Proc. Inst.
GRANTER,
Civ Eng. 1964. Oct. No. 6754 · • "" •, • ,,:.
B,SCb Cf. DUBOURG , L.: Le viaduc et:'fes pas,Sijges.;~~terrain s de Bruxelles.
La technique des
Travaux January 1958 · · • ,••,
---~ • '..;, .--'\(.•..r
'>,•'7" ' L",:,$:n"t
-_,~•._.. 1,.• '

~. - ,
.... ::. :f ~-, _....I'
" '
 Longdudioa/ section

I
3-50
- - .fctJt!J

() 0
er -·er~·d1 -e,-·ot,.,ring
S: ""' sheet p1/io§'

~/t.. ..... ,.."-..,::·\_-.:·. '-"_:'c"·~-s SectionA -8 Section C- 1)

-<I • .,...... • : 7, .;' "':: ,... ..,~. I' '7!::t~

".":""' ,. . , ."[Qr "f.r ~"'"'·-TIJtt=-"":;ccc==:;x:=c,c-==-"'r::::::m=::t:r.cc.c=-,-,=rn

i =-t
2
-/rr~• -,. ~ .~
1
} UT, w,;:cr:;.rc(il'!,' ' ) 'T ' Reir:for::ed-co11cre/e wme
12cm pnt!:dWf!
/Jf'!U . ;_,;,r ,,
_,
I I
- 7J!e cover
Is:,
1
,·.< ~~::c-$~~
...c;.b., ~-a;~;.;-~-;::-;;;-~.,-,....,;4,J,
. ·' .~ ~ jG
7cm waler{'roc-r."g
10 cm protecttte
. '"
CO/JCl'ele
~ Steel s/Jeet o I: 1;-

Structural detail
4 cm protective B.I' Q:l ! 2·/8¢
/plates I
~

I I I
rt I 111" fl"- "I I.I ,~;, I
~i T ,;,,1
.. I
"L) " I I

0 : l) / JI '

A . /
(§Olflill!J pills
/1

"(. I ~-18¢ '

2-<1:- g. -
nn I

'- ' '- 400

· oo o~r -J
,J-265 Section A-B ~
------L~

i' L. ¢18t ~
~~ •~ /
~~
- 1l

~- . . rt
""
~ re--~
~-
0-20 0·20 O?O o-20 0-20

Fw. 6/176. G~ncral dc:,ign and dctaih or th~ pcdcssrian ,ub\\ay under the left acccs ramp of the Margaret Brid.

---
 Longitudinal section
_ __ _ _ _ _ _ _ __.___J$
--ged road Width
- -- - --.----4 -§0
North
--
I
2 cm sheet 11sphalt
I I

.I TTfi Preca~t rein/breed-con- 0

i ITT I crete roof beams

I I I
)
"' ,,:..J'.~ _, .,;,I.

Section C-0

Layout
.strip constructed in one

'--~- 0-f,J I I I 17l§ht

1cm waterproofinq
10 cm protectwe
concrete
Steel s,?eet 01/mg
Prenous roadwa§
Structural detat!
'8 8.;,
~! 2·/8¢ haterprC'o1og
I

I 4· 18¢
"( i 2- 4 cm gumte rr?dering
\
- ---400- - - - - -- - - 020 Precast mnlbrced-co(lcrete cap
Section A- B i ~
. ..
--l-/ "1-

~
rt Inner monolttll1c
0-10 0-20 020M'oiJo ':!: remforced-coflcrete ;ia//

e left acces ramp of the Margaret Bridge, Budapest

 SHALLOW URBAN AND PUBLIC UTILITY TUNNELS 817

64.22. Margaret Bridge in Budapest

In the year 1936 when the Margaret Bridge in Budapest was being widened,
pedestrian subways had to be built under its access ramps first on the Pest (left)
side and then on the Buda (right) side of the Danube, in both cases without inter-
fering with public traffic. 6·55 The construction of these subways was designed to
eliminate the jamming effect of heavy pedestrian-crossing traffic and to provide
for safe and undisturbed access by pedestrians to the platforms of the centrally
located tramway station island. Fig. 6/176 shows the layout of the subway on the
Pest bank and the details of the roof structure.
The two outside sections of the subway lying under the newly built approach
sections, could be built in an open trench as closed monolithic reinforced-concrete
frames without any interference with bridge traffic. However, the section lying
under the central tramway platforms, i.e. under the used approach could be built
without interference to day traffic only if construction work was restricted to the
night hours from 11 p.m. to 5 a.m., in which period tramway tracks and road
pavement had to be torn off in two narrow strips in the axis of the proposed side
walls in widths of 2-3 m. In the shallow pits excavated, steel sheet piles were
driven in two parallel rows for the temporary side supports and also as structural
elements of the permanent side walls. On the top of the driven sheet piling, rein-
forced-concrete inverted trough-shaped cap elements were placed. In the morning
hours constructional work was stopped, the trenches refilled and tramway tracks
and road pavement replaced and traffic flow restored during the day. The next
night the road pavement and the tramway tracks were again removed, but this
time not only in the continuation of .:he narrow strips of the two side walls but
in the length where piling had been done during the previous night and also in
the section in between, to allow the precast reinforced-concrete roof beams to
be placed side by side. When the corresponding group of 20 cm wide elements
had been placed, their top was coated with bituminous waterproofing and covered
with precast-concrete plates for protections. After this had been completed the
soil was backfilled again and road pavement and tramway tracks replaced. The
step by step construction of a load-bearing deck made it possible for earth to
be excavated and removed from both interior ends of the subway, without requir-
ing any additional support and without interference with traffic at the approaches.
The permanent reinforced-concrete side walls and the invert slab of the subway
as well as its internal sealing and glazed tile cover could also be placed without
any disturbance. The waterproofing of pedestrian subways must be undertaken
with great care, in general, not only to maintain their stability and to afford the
necessary protection against corrosive action but also to satisfy aesthetic demands
for a stainless condition which necessitates the absolute exclusion of water infiltra-

6·55 SZECHY, K.: Bau der FussgangerunterfUhrungen der Margarethen Donaubrlicke in

Budapest. Bautechnik 1937 and The Reconstruction and Widening of the Margaret Biidge,
The Structural Engineer 1937 March.

52 Sz~chy : The Art of Tunnelling

 817
SHALLOW URBAN AND PUBLIC UTILITY TUNNELS

64.22. Margaret Bridge in Budapest

d,
In the year 1936 when the Marga ret Bridge in Budape st was being widene
the Pest (left)
pedestr ian subway s had to be built under its access ramps first on
withou t inter-
side and then on the Buda (right) side of the Danub e, in both cases
was designed to
fering with public traffic. · The constru ction of these subways
6 55
and to provid e
elimina te the jammin g effect of heavy pedestr ian-cro ssing traffic
ms of the central ly
for safe and undistu rbed access by pedestr ians to the platfor
of the subway on the
located tramwa y station island. Fig. 6/176 shows the layout
Pest bauk and the details of the roof structu re.
approa ch
The two outside sections of the subway lying under the newly built
thic reinfor ced-con crete
sections, could be built in an open trench as closed monoli
Howev er, the section lying
frames withou t any interference with bridge traffic.
could be built
under the central tramwa y platforms, i.e. under the used approa ch
only if constru ction work was restrict ed to the
withou t interference to day traffic
a.m., in which period tramwa y tracks and road
night hours from 11 p.m. to 5
two narrow strips in the axis of the propos ed side
paveme nt had to be torn off in
In the shallow pits excava ted, steel sheet piles were
walls in widths of 2-3 m.
as structu ral
driven in two paralle l rows for the tempor ary side suppor ts and also
piling, rein-
elements of the perman ent side walls. On the top of the driven sheet
In the mornin g
forced-concrete inverte d trough- shaped cap elements were placed.
tramwa y tracks
hours constru ctional work was stopped , the trenche s refilled and
the day. The next
and road paveme nt replace d and traffic flow restore d during
remove d, but this
night the road paveme nt and the tramwa y tracks were again
of the two side walls but
time not only in the continu ation of ~he narrow strips
the previou s night and also in
in the length where piling had been done during
reinfor ced-con crete roof beams to
~he section in between, to allow the precast
wide elements
be placed side by side. When the corresp onding group of 20 cm
bitumin ous waterp roofing and covered
nad been placed, their top was coated with
ions. After this had been comple ted the
with precast -concre te plates for protect
paveme nt and tramwa y tracks replace d. The
soil was backfilled again and road
of a load-be aring deck made it possibl e for earth to
step by step constru ction
d from both interio r ends of the subway , withou t requir-
be excavated and remove
the approa ches.
ing any additio nal suppor t and withou t interfer ence with traffic at
ed-conc rete side walls and the invert slab of the subway
The perman ent reinforc
placed withou t
as well as its interna l sealing and glazed tile cover could also be
be underta ken
any disturb ance. The waterp roofing of pedestr ian subway s must
y and to afford the
with great care, in general, not only to mainta in their stabilit
aesthet ic deman ds
necessary protect ion against corrosive action but also to satisfy
on of water infiltra-
for a stainless conditi on which necessitates the absolut e exclusi

Donaub rUcke in
SzECHY, K.: Bau der Fussgiingerunterfiihrungen der Margare then
• ·55
ruction and Widenin g of the Margare t Bridge,
Budapest. Bautechnik 1937 and The Reconst
The Structural Engineer 1937 March.

52 Szecby : The Art of Tunnellin&

818 CONSTRUCTIO .\ND DESIG N OF T UNN ELS

tion. Waterproofing was achieved with a continuous gunite rendering of 2 cm

thickness in the given case. in addition to the continuous bituminous coating.

64.23. Dimitrov Square Subway in Budapest

Another example worth mentioning is the subway built under Dimitrov

Square6 · 56 in Budapest crossing the access ramp of the Szabadsag Bridge over
the Danube just behind its left abutment. (This subway accommodates a subsurface
tramway station completed with an adjoining network of pedestrian subways.)
The reinforced concrete side walls were built, in this case, in a strutted open
trench and placed on a foundation of sand piles and on a well-compacted soil
cushion (Fig. 6/177), because of the relatively low load-bearing capacity of the origin-

s,o'ewall constroctlon 11noer Sflb constr11ct1on after pli

temporarp overbr!dgmg . ang ofprecast rootelemenL< . ongitod/lJal sectJor
fO·tiJ

ICIISI//OI/ 'Bundle
" orpre-tens1011ed
I vr
V roof beams

Plcn

Attendants (Toll},
l!o11Se .

0
,_,____,__,_ __...,_ ,ramwa11 track

Fm. 6/ 177. Design of subway at Dimitrov Square, Budapest

6 36
• Kov,kSHAZY, F.: Dimitrov teri aluljar6 epitese (The construction of the subway at
Dimitrov Square), Me/yep. Tud. Szemle April. 1952
 PIP~ JA CKING 819

al soil. Following this. trenches crossing the most frequented traffic lanes were
spanned by short temporar y bridge~ to a width about equal to that of the roadway
of the bridge to allow the pits to be deepened and the reinforced-concrete side
walls were cast in them without interference with surface traffic. The permane nt
deck girders placed on the side walls completed in t,bis way were of prestressed-
concrete (Fig. 6/ 177). As these could, of course, not he placed without tearing off
the surface, bridge traffic had to be stopped for two consecutive holidays. First,
the temporary overbridging was removed from above the side walls. Then the
permanent prestressed-concrete roof girders which were previously arranged in
groups of 9 to 11 units, were placed on the side walls and covered with bitumino us
waterproofing, o n the top of which protective concrete plates were placed. The
girders in each group were tightly forced to one another by transverse post-tension-
ing. This not only helped to speed up construc tion work, but, above all, also
secured a better co-op~ra tion between the single beam elements and a better and
more uniform distributi on of external surface loads among them. In conclusion,
this post-tensioning resulted in an increase of 18 % in the load-bea ring capacity
of the single elements.
subways
The problem of load distribu tion is of great importanc e in the constructi on of
with a roof composed of prefabrica ted elements, as it could not be considered econom-
designed
the load
ical to design the single elements with a load-beari ng capacity sufficient to carry
bearing grids
of a heavy vehicle's wheels individually. Also, the installatio n of cross ribs or
Trans-
would be expensive and would, at the same time, involve a loss in structural height.
solution of
verse tensioning of the beam elements is, without doubt, the most advantage ous
(see above)
the problem. The co-operati on of the girders of the M argaret B ridge Subway
upper load-distri buting concrete layer witb steel mesh reinforce-
was obtained by the use of an
contacting
ment. There are, neverthele ~, also some other solutions, as, e.g. when the lateral
the adjacent
surfaces of the beams are tong'ued and grooved. If these beams are deflected,
heavily
grooves and tongues will interlock as a result of which part of the load on the mostly
however,
loaded girder will be transferred to the adjacent ones. This constructi on requires,
a seldom obtainable exactitude of one millimetre in prefabrica tion.

The great advantages in the use of prefabricated elements are that not only
is there a reduction in construc tion time and disturbance to traffi~, but also, and
particularly, a lower structura l height is necessary, e.g. as opposed to the usual
tunnelling methods, where, owing to the use of timbering, a greater height is
required and this applies to the use of temporar y bridges over the trenches excavat-
ed from the surfa,ce for the maintenance traffic. (This was why the tramway
tracks in front of-the Opera House in Vienna had to be elevated above the level
of the road; cf. Fig. 6/ 175.)
Some examples for the use of prefabric ated elements are presenlt>d in Chapter
6.44.

6.43. CONSTR UCTION OF CULVER TS AND SEWERS BY PIPE JACKING

Pipe jacking can be considered fundamentally as a kind of shield tunnelling,

where the propulsion jacks remain in place and the permane nt lining (a pipe)

s2•
820 CONSTRUCTION AND DESIGN OF TUNNELS

is driven ahead instead of the shield. The principle of the method is that vertica
shafts are sunk to the design level of the pipe, one at each of its end points. A

I
r----------- i
I I
! I
I I
I I
·~- ---------j _l-iJrtfwroo dtstr1/Jutlng frame
4~~~ §'l'J'l/ ..,ack

FIG. 6/ 178a Principal

arrangement of
pipe jacking
Surde timbers

the bottom of these shafts bearing walls of suitable strength and jack frames
are built with pipe-guiding rails or beams in front of the jack (Fig. 6/ 178a). The
shaft must have a width suitable not only for ho using the jack and the load
distributing frame but for a lining unit (a pipe) as well.
A reasonable arrangement fo r the shaft is given in Fig. 6/ 178b. As shown, the
best constructional arrangement may be obtained, when bottom slab, side walls
and the rear abutment wall are cast in situ in form of a monolithic R.C. unit
and the excavation for the shaft i!i done within a sheet-piled enclosure. In this
way not only the fol) cooperation and resistance of the bottom-slab and side-walls
are secured against the propulsion reactions, but the recesses and cut-off depth of
the sheet pile wall are additionally increasing this value. The constructional arrange-
ment of the opening on the front wall of the shaft has to offer first of all a definite
and firm guidance for the pipes, but a n appropriate sealing around its circum-
ference against the infiltration of ground water o r against eventually applied
thixotropic fluids must also be provided for. Finally, the embedment of guiding
rail!> and the location of water collecting cana ls in tbe bottom slab are also an
important requisite.
For jacking the pipes the application of about 4 propulsion rams is advised.
The propulsion itself may be effected in rwo ways:
(a) The ram1, are supported directly on the abutment wall, when appropriate
extension pads must be inserted directly between the ram and wall after each
stroke.
(b) Th~ rams are built into a sliding frame, when this frame is advanced
according to the stroke and the altering position of the frame is to be fixed from
the wall by the insertion of pads and wedges (Fig. 6/ 178c). In this latter case the
repeated dismounting and mounting of the rams is omitted.
 SHALLOW URBAN AND PUBLIC UTILITY TUNNELS
821

.-1

I
I
I
I
I
:~
1
Concrete
/'.II blinding
111
111
-_------=---=--=--
=:::~_:-_-_--:_-::.-::::-__ -::-::-__
,· · Sandor
~--------- -----Drain
- ~------
off canal Sump
' slag cushion

Fro. 6/ 178b Shaft arrangement for pipe jacking

The stroke of the jack is practically equal to the length of one pipe, as otherwise
an extension pad inserted between the jack and the pipe has to be used after the
stroke length of the jack has been exhausted.
The pipes to be jacked can be of flat steel, of precast concrete or of corrugated
steel units and should have an inside diameter of at least I m to allow a man
to work inside them. (In practice, even pipes of smalJer diameter are jacked,
in which cases, however, special tools, such as scraper rods must be used for
excavation and guiding.) The soil is excavated to a distance of 30 cm from the

j
822 CONS1'R UCTION AND DESIGN OF TUNNELS

~ l To prmu~ ~ump--

------t+---------+-1--I-· ----t-l>e t 11

- o,le

Commanding .olve
4 hydc ja cks \
If .
,;:stribution Sliding f,·c.me :,

'/

Rubbe; collar Clamps/'oe

FIG. 6/178c Arra ngement of pro1:11l5io n apparall!S w ith a sliding frame

front of the pipe, at the same time taking great care to prevent the soil from
caving in. Following this, the pipe is pushed forward by the jacks until an overall
distance equal to the length of the pipe is reached, when, after the jacks are with-
drawn, there will be sufficient room to place and join up the next JJipe to be jacked.
The pipes are connected by lap joints fo r concrete pipes, or by lap joints or Wflding
or threaded internal fish-plate joints for steel pipes. Constructionally sound joint
arrangements for R.C. pipes are shown in Fig. 6/178d. Care must be taken that no
protrudin g bosses, ribs or rivet heads appear on the outer surface of the pipes. Pipes
made of corrugate d steel-sheets may also be provided with flat guiding strips
welded to the outer surface of the pipe to reduce frictional resistance. The magni-
tude of the specific frictional resistance depends upon the nature of the soil, the
extent of overbreak around the pipe and the diameter of the pipe. The extent of

Solid rubber gasket

Plywood ,n/ay
Pf',/WOOd inlO',/

FIG. 6/ I 78d Watertigh t joints for RC. pipes

 SHALLOW URBAN AND PUBLIC UTILITY TUNNELS
823
overbrea k around the pipe is usually about 2- 3 cm. With grt>acer dia pipes,
however, such a gap behind the pipe would produce a considerably bigger settle-
ment of the roof layers than with small dia pipes (arching of the ground). Owing
to this fact, the specific frictiopal resistance also increases with the diameter of
the pipe, its value varying from 0·5 to 2·5 ton/m 2 on an average.
The specific frictional resistance and thus the required propulsio n force may
be effectively reduced by lubricatin g the pipe skin with some thixotrop ic fluid
(bentonite). This is to be effected mainly around the top section of the pipe and
the fluid must be directed to the cutting-edge which in this case must be formed
to leave a sickle-shaped space free· around the circumference. A practical arrange-
ment is shown in Fig. 6/178e.

Bentornte

Section A-A

Fie. 6/ 178e Layout of bentonite luhrication

This method ii; genera1ly employed for the construc tion of culverts in newly
backfilled embankm ents or for sewers or public service line passages under railways
or highways. Accordin g to RICHARD SON and MA YO the longest distance attainabl e
by pipe jacking from one direction is, a t m..ost, 30 m ; for longer distances the
pipe jacking must be carried out from both sides. More recently, even longer
distances have been obtained by the injection of thixotrop ic fluids mentione d
above. 6 · 57
The maintena nce of both the proper direction of the pipes and the concentr ic
actio n of the pushing forces is of the highest importan ce, as otherwise pipe
failures could arise, jeopardiz ing the success of pipe jacking. Direction control
can be carried out by sighting through sighting squares fixed in the pipes (Fig.
6/ 179a). Smaller deflections can be corrected by overbreaks made in the required

s-s 7 LORENZ, H.: Erfahrung en mit thixotrope n Fliissigkeiten im Giundbau , Bautechni k

1953 232 and 1957 200
 824 CONSTRUCTION AND Dl!SJGN OF TUNNELS

8.)
direction, and at the same time setting
an adjusting sc;ew-jack between the op-
posite side and the edge of the pipe
(Fig. 6/179b) and with the uneven oper-
ation of propulsion jacks. Excessive
excavation ahead the face is danger-
2-5, 5cm
ous and it must be always cared for that
\ the cutting e~ge should not be set free,
""- but it should always "cut". This is the
best warrant for a definite direction.
lrenc/J .J8CK Many difficulties have to be over-
come when using the method be low
groundwater level, as drainage by com-
pressed air can be employed only at
a very high cost and limited to bigger
diameter pipes, and open drainage into
. /.leading enlarged Ill sumps could be dangerous in loose
deslf'ed dtrect1on ground. Lowering the ground-water
Fm, 6/ 179. (a) Sighting square for checking level with vacuum well-points would
alignment of pipe in jacking; be the most suitable solution of the
(b) method for alignment problem, but this, however, would be
correction
effective only up to a certain distance
from the outfall heads.

6.44. CUT AND COVER CONSTRUCTION METHOD WITH PRECAST PANELS

As indicated in the title, the method to be discussed here can only be considered
as a tunnelling method in so far as its aim is the construction of underground
tunnels. This aim, however, is not achieved by underground excavation but by
excavating an open trench either with vertical sides supported by sheet piles or
,,.-1th sloped sides without support; in the trench, prefabricated lining elements
(panels) are placed from the surface. The use of large lining elements is, thus, made
possible as with this construction working possibilities are not limited by th~
small cross-sectional area of the tunnel either for hoisting capacity or the available
working area. All these circumstances greatly improve the economy and efficiency
of the tunnel construction, and, at the same time, reduce the number of unavoid-
able joints in the tunnel-lining, so alleviating the problems and difficulties of drain-
age and water staling.
A fairly good example of the application of the method is represented in the
construction of the latest extensions of the Metro in Moscow {the Kaluga line).
Owing to far-sighted and carefuJ town planning, the construction of the under-
ground railway preceded the building of houses, roadways and public service
lines in this district, so that the adoption of the safe cut and cover method with

--
 SHALLOW URBAN AND PUBLIC UTILITY TUNNELS 825

the use of sloped trenches was made possible without the slightest danger to the
surroundings.
The tunnels built for the Metro are, in general, double-track tunnels of rectang-
ular cross-section, separated into two single-track tunnels only just ahead of
6 59
the .stations to allow the necessary central platforms to be construct ed • {cf.
Fig. 6/163).
These single-track tunnels which have a cross-section of 4·56 m x 5·30 m are
schematically shown in Fig. 6/ 180a, while the double-track tunnels can be seen
in Fig. 6/180b.

FIG. 6/ 180a Subsurfac e

(single) tunnel section
composed of
3z-z prefabrica ted elements
(Moscow)

l 1---- ----'.;z..o .t.~------.~

,{b.7:rete
socket
.,Jomts f't!led m situ w1tl!
concrete
Double -track tunnel of prefabricated elements

F1G. 6/JS0b Isometric view of the twin tunnel section of prefabricated elements (Moscow)

6·59 KRrVOSHEI N and RI ZNICHENKO : Sooruzhenie tunnelei Metropoli tena otkritim sposo-
. Stroit.
bom (Construc tion of the tunnels of the Metro with open-surface methods), Transportn
J959 Nov. 20- 23
 826 OONSTRUCTION AND DESIGN Of TUNNELS

After the trench had been excavated with a bucket excavator and the botton,
had been levelled, a simple concrete socket was placed to give an exactly plane
surface coated in succession with waterproof sheeting. Following this the invert
(marked on the figure by L-3) and two side-wall panels (Sz) were placed, each
having a length of 3 m. When assembling these lining elements, a 50 cm wide
gap w_as left at the joints to allow for the welding of the exposed ends of the
reinforcement bars; this gap was then filled in situ with concrete. After that
followed the placing of the •·5 m long, trough-shape d and ribbed roof panel
(P-/ ), and the concreting of the abutment joints. The double-track tunnel section
is similar in arrangement , with the exception that, prior to placing the roof panel
and concreting the gaps in the joints, a central bottom panel and a separation
panel are placed. A single panel weighs more than IO tons, and this weight
is expected to be further increased with the development of mechanizatio n, which
again will result in the further speeding up of tunnel construction .
Several public service tunnels - though much smaller in size - have been built
in Hungary also according to the same principles, mainly in new settlements
for the placing and accessibility of public service lines.6 ·80 Channelled concrete
frames weighing I· 3 tons and measuring 50 cm in length, 12 cm in thickness and
wit_h_cross-sectional measurements of 2·0 m x I·90 m were placed in dry trenches

FIG. 6/ 18 1 Section of Hamburg subway composed of prefabricated elements

8 • 60 R6zsA
L.: Zartkeretu elo regya rtott vasbeton kozmualagut (Pu blic service tunnels of
precast reinforced-concrete panels), M e/yep. Tud. Szem/e Oct. 1959
 SHALLOW URBAN AND PUBLIC UTILITY TUNNELS 827

with supported vertical sides, by a travelling crane. The channels between the adja-
cent frames were subsequently filled with cement grout and then made watertight
from the 'i:op.
Other examples can be found in Europe of the application of this method6· 61
which is exceedingly suitable for mechanization. Thus, e.g., some precast prestressed
reinforced-concrete panels, 5 m high, 8 m wide and 2 m long with central inter-
mediate posts are being adopted in open trenches excavated under the protection
of ground-water lowering in the recent extension of the double-track underground
railway tunnel in Hamburg (Fig. 6/ 181). All these panels (each weighing 37 tons)
have iron bands protruding from their sides which, after they have been placed
on carefully levelled concrete sockets and after an adjustment to an accuracy
measured in millimetres, are welded to the iron bands of the adjacent panels
along their entire periphery. This results in an absolutely watertight lining. Though
the operations were carried out in the night hours during the absence of traffic
only four sets of panels making a double-track tunnel section 8 min length could
be built per night.
Whereas in the previous examples the whole cross section was composed of
prefabricated elements, a partial application of them may be demonstrated in

1 to 20 l' . 10/f'Precast roof sla~. . {asphalt

;on~rete 8 d1a. 4· n d1a. 8 ,d1a. ~ I A insulation
,
• 'l 0
o ' O • ' . :\. ·· . ,.a :.. ,_;_,..,,,~ - ,,_::.;....,...=-·
0 . ~.
.:._·. :l ·<·; . •-·•\?k:\.-Jns1tu concrete . :.- ·_. X' • •:
~ , f- • : ~- : : :,

:• 0: ·1n11
~--- ----+i. :,._c_';_us-t-ic_c_e_il
1
..,."i'_'' ,-·ng--~II
• 0

.'. ;·, ·• •.. o . •It

.--- Precast units

·• , I
:: .: 10' 0 'clear . ... . ~ '
. .. ,
:·~ ·.;·-- i ) 1' .
• •• · )
•• ·1
2 d1a. {Mosaic tiles..---- ,
0
..
~ :

• ·.1.
...
0 . -~ .

•
,
3'max. floor finish 7i/erazzo floor tiles
·,
' ..
•• • 1' \
• •· .,- sf!a. 1"arout
... ·W~~~~~;;;:;:;;::;;;;::;::;::;:;~~=:=fil_J!
/ \ .
;:~· ·j ·-~.; ;:;·.~-:->:~~>._::.\.:i..~:·>~·.'·.P/ 0...•0 o; .,\ : 0 • ,.• ~- o'·;:~-·-~--~;·" •.. t>";;.) • • •
-~ Q

1'Granolithic pad 1 to tJ concrete 1 asphalt f ~ 1' Granolithic

A pad
CROSS SECTION
Fm. 6/182. Cross-section of a pedestrian subway with U-shaped precast bottom units

6 · 61 Cf. Neue Bauweise im U-Bahnbau, Der Stadtverkehr 1958 8. Cf. also : Der U-Bahnbau

in Hamburg, Schweizerische Bauzeitung 1958 43

 1128 CONST RUCTI ON AND DESIG N OF TUNNE
LS

connection with the construction of the pede

strian subways of the Park-Lane
improvement scheme London. 6· 614 Here "prec ast
concrete U-shape units 4 ft long
and 9 in thick form the floor and walls, with in
situ joints between them . The
roofs are of composite construction consisting
of 9 ft 6 in long precast units
with in situ reinforced concrete on the top" (Fig.
6/182).

6.5. SAFETY MEA SUR ES AND HEA LTH

PRO TEC TION
IN TUNNELLING

6.51 SAFE TY MEA SURE S

Tunnelling is inherently a hazardous operation.

The small working areas
available in beadings or inside the shields, the wet
or slippery working platforms,
floors and footings, the artificial and most often
inadequate undergro"und lighting,
unseen weaknesses in rocks , the treacherous loose
ned rock roofs , the loading,
hauling and unloading of muck, coupling of cars
and trains in limited space and
bad visibility as well as the operation of loading,
hoisting, drilling and other machin-
ery are all sources of accidents, apart from the
danger in handling and using
explosives .
Still, it can be stated that accident prevention
is essentially a matter of good
organization, as the percentage of 'unavoidab
le· accidents arising from natural
catastrophes (e.g. those caused by sudden break
down of ground, or water and gas
inrushes, etc.) is comparatively small compared with
the percentage of those arising
from negligence, lack of order , tidiness or discip
line.
Strict safety regulations and rules must, therefore,
be laid down for each work-
ing place before the start of tunnelling operations
, and these regulations must
be strictly obeyed. Permitted violation of any
safety rule - no matte r how slight -
nullifies the whole safety programme. Some of the
most important safety measures
are detailed below :

I. All machines and tools are to be kept in prope

r and efficient condition .
2. Reserve tools, construction materials or mach
ines must not be stored in
the tunnel sections.
3. Gangways and working platforms must be
maintained in solid and clean
condition.
4. Working sites must be provided with efficient
lighting, and electric wirin g
should be properly insulated.
S. Telephone communication must be established
between working places,
hoisting and lift stations, machine house and
the works headquarters.
uu Cf. GRAN TER, E.: Park- Lane lmpro
meve nl Scheme: Design and Construction, Proc
Inst. Civ. Eng. 1964. Oct. No. 6754
 HEALTH AND SAFETY MEASURES
829
6. There are many circumstances from blasting6·62 to muck haulage which
conduce to accidents in excavation work. Mine cars must not be filled beyond
the provided clearance and above their safe bearing-capacity as projecting tools,
pipes or machine parts can easily come into collision with supporting elements,
especially when hauled on uneven, rough tracks scattered over with muck falling
from the overloaded cars. The coupling and uncoupling of cars and trains, as
well as runaway cars Gack-catches!) are all importan t sources of many accidents.
7. All elevators, lifting and hoisting machines and equipment must be designed
with a safety factor higher than usual.to allow considerable overloadings. They
should not be operated without strict service instructions. The use of automati c
safety brakes and reliable sign:111ing apparatu s is, similarly, of the highest impor-
tance. Many serious accidents have occurred when hoisting was started before
the mine car or the load had been properly encased or fixed in the cage.
8. The breakdow n of rock or ground is a source of accident characteristic
of tunnelling. Therefore, the careful and frequent check of the roof, and the sides
as well as that of the tunnel supports is of primary importance. It is especially
importan t that the rock surfaces freshly exposed to blasting should be carefully
cleared from the loosened spalls or fragments and that the supporting elements
should be inspected frequently in order to detect possible deformations and breaks.
9. Proper protective clothing (steel helmet, rubber boots, working suits, goggles)
are essential requisites for successful accident prevention. The wearing of a proper
steel helmet is most especially importan t as it will protect the head not only against
falling objects but also "\\'hen colliding with props, caps, etc. Feet can be best
protected against falling 9bjects, soaking and cold by waterpro of and solid boots.
10. In tunnelling work a well-organized fire-fighting service, and its supply
with foam extinguishers, etc. cannot be dispensed with. Fires in sites under com-
pressed air are particularly liable to involve catastrophic consequences.

6.52. HEALTH PROTECT ION

Hygienic regulations relate, first of all, to the proper ventilation of the under-
ground working places already dealt with under Section 46.51. Suitable working
clothes (proper footwear and light, yet warm clothing suitable for protectio n against
rheumatic diseases which are easily caught in moist places and atmospheres)
are of similar importance. Still more importan t, perhaps, is the insia1lation and
maintenance of warm shower-bath rooms, comfortable and clean dressing rooms,
rest rooms and lavatories at tunnelling sites.
The special occupati onal diseases of tunnelling, i.e. silicosis and caisson disease
the latter of which results from working under compressed air, will be briefly
discussed below.

a.et Cf. Robbantas (Blasting), Bdnyabiztonsagi Szabalyza t V 2 1259

830 CONSTRUCTION AND DESIGN OF TUNNELS

65.21. Si.licosis
Silicosis is a serious pulm,;mary disease caused by prolonged ex~osure to high
concentrations of rock dust containing minute particles of silica \Si0 2) . Most
rock formations contain silica; quartz is almost pure silica, and sandstones are
largely built up of silica. The silica dust particles cause fibrosis of the lung tissue
which so far is incurable and the progress of the disease often continues after
the victim is no longer exposed to the dust. According to physicians, silicosis
is caused by chemical processes which take place when the dust particles, less
than 10 microns in diameter, react with the pulmonary lymph. The minimum
exposure time sufficient for the development of silicosis has not yet been determin-
ed. lt is, however, supposed to be somewhere about 3 years, depending, of course,
also upon the state of health and the liability to pulmonary diseases of the individu-
al. The symptoms of the disease will only become manifest after 5-6 years. The
blood passages in the lung will narrow due to cicatrization, as a result of which,
the work of the heart is intolerably increased and finally leads to failure. The
causes of the disease cannot be eliminated, nor the disease cured, although its
symptoms can be alleviated, and the illness always terminates fatally.
In moist air - thus, also, in tunnelling conditions - the silicosis hazard is less
dangerous than in dry atmospheres, such as in quarries, or stone dresser shops,
i.e. in plants where the air is usually dry and polluted with silicous dust. in Switzer-
land, for example, it is legally binding to report to the relevant authorities all
those activities where the quartz content of the air-borne dust exceeds IO% by
weight, whether in the course of excavation, dressing, transport or utilization
of minerals. in such cases it is for the Accident insurance institute to decide
whether the plant is to be considered as imperilled by the silicosis hazard or not.
ff so, the labourers employed there must undergo certain medical examinations.
while suitable labour-safety measures and equipment must be installed at the
plant (dust control, respirators, forced v.entilation, spraying, moistening, wet
drilling or dressing, etc.). 6·63
Of the methods used to control the dust from drilling, the vacuum hood method
has proved satisfactory especially with the use of dry drills. The system consists
of a hood that fits around the steel drill at the rock face, from which a suction
hose leads all the dust created by the drill to separators and filters. With wet
drilling it is no longer considered sufficient for the prevention of silicosis. for the
microscopic dust particles responsible for the disease to escape with the water
stream.
The most effective dust control methods are water spraying the air-borne
dust and the use of wet and ventilation drilling. For exampie, in the costruc-
tion of the Mont Blanc tunnel the dust in the working place appeared only as
a fine sludge.

u 3 Cf. Btin_vtiszati Lapok 1959 4- 5

 HEALTH AND SAFETY f>tEASURES 831

65.22. Caisson Disease

In conjunction with tunnelling jobs a separate discussion of caisson disease
(decompression-sickness) is justified, the work under compressed air in tunnelling
being of longer duration than work in bridge pier foundation jobs. While the latter
are usually completed in one or two months and, even when built in succession,
the overall time of their construction does not last longer than eighteen months,
in the course of which there are generally interim pauses, or, at least, the air
pressures applied will successively vary between lower and higher values, tunnel-
ling jobs are mostly of a duration of many years with no pauses, and are usually
carried out under considerably higher air pressures. The common misgiving
that if the human organism is exposed to high air pressures for long, it will, sooner
or later, become exhausted and more susceptible to caisson disease, seems to
be fairly justifiable. This apprehensio n, however, was not supported by the experi-
ence obtained in the construction of the underground railway in Budapest. It
could, nevertheless, be stated that the specific number of cases of caisson disease:
B (i.e. the number of cases related to the number of decompressi ons: k) did sudden-
ly increase when the air pressure reached a value of l ·7-1 ·8 atm (24-26 lb/in2)
overpressure. This limit value could be clearly shown, e.g., in the ye·ar 1953 for
the 11th working site, where this specific number was found to be B/k % = 1·8 %
at an air pressure of l ·8 atm overpressure, immediately falling back to a value of
0·2-0·4 % as soon as the air pressure was reduced to 1·5 atm. Or, e.g., in the
case of the 9th working site, where the value obtained for B/k at an air pressure
of 0·7 atm overpressure was only 0· 15 %- At the same time the average value of
B/k obtained in the year 1952 for an air pressure of 1·7 atm overpressure was
found to be 0·84 %- The same figures at the construction of the Clyde tunnel in
Glasgow were 0· 18 %and 0·33-0·92 %rnspectively. Accordingly, in tunelling work
an attempt should be made to ensure that the air pressure applied should, wherever
possible, not exceed 1·7 atm (24 lb/in2) overpressure. Thus, it is far better to carry
out the work under an air pressure lower than that required fo r full dewatering at
the bottom level of the tunnel - so also reducing the risk of dangerous blow-outs
due to unbalanced pressure - and to apply air pressures equal to the water head at
the axis or springline level of the tunnel depending upon the critical height differ-
ence which would make soil particl~s also flow with the entering groundwater .
Thus, in larger diameter tunnels air pressures some tenths of an atm lower than
required at bottom· level may be applied (cf. Fig. 6/ 147).
The incidence of sickness is, generally, not influenced by the seasons or by
climatic factors and weather conditions. According to American reports, however,
the specific number of cases of caisson disease (bends) varied according to the
day of the week. For example, during the construction of the Queens-Mid town
tunnel6 • 64 it was bighe.;t on Sundays (0·4 %) and lowest on Mondays (0· 12 %)
u, SINGSTAD, 0.: Cf. The Queens-Midto wn tunnel, Transact Am. Soc. Civ. Eng. 1944
109 679-762. Cf. also: HA XTON and WHITE: Constructiona l Problems of the Clyde Tunnel.
Proc. Inst. Civ. Eng. 1965 Febr. 337

.....I....
 -
;; 32 CONSTRUCTION AND DESIGN OF TUNNELS

and this latter percentage increased successively from Monday to :Saturday up

to 0·2 %- The first value (for Sunday) can be ascribed·to fatigue caused by Saturday
entertainments, while the last value (for Saturday) must be attributed to' fatigue
caused by the whole week's labour. According to American regulations, the longest
permissible uninterrupted working time under compressed air is 4 hours. With
increasing air pressure the working time must be reduced and the time of rest
increased according to the table below

----- -
Air pressure I I
(atm
overpressure) 0·OO-l ·26 I ·26- 1·84 1·84- 2·3) 2·3) - 2·66 2· 66-3·0) 3·0)- 3·36 3·36-3·5
0
-- - -
Working time
(hours)
7 4·0
3·0 2·0 I ·5 )·0 0·75 0 ·5

~
T ime of rest · -- -
(hours) 3·0 4·0 4·5 6·0 6·25
I I
7·0
-- -
Working hours
I 3·0 i 2·0 )·5 l·0
I 0·75 o ·5

lt is beneficial to vary the rate of decompression as follows : for air pressures

of 0-1 ·05 atm overpressure it must be, at most, 0·21 atm/min; for l ·03-1•4 atm
overpressure 0· 14 atm/min; for l ·4-2· l atm overpressure 0· I05 atm/min and for
air pressures exceeding 2· I atm overpressure 0·07 atm/min. The so-called ·stage
decompression' has also been applied where the overpressure is reduced to
half its value at a higher rate of decompression (O·35 atm/min) using the
remainder of the prescribed decompression time for the reduction of the rest of
the overpressure to atmospheric pressure at a constant, much lower rate of
decompression.
The use of two air chambers, i.e. of double bulkheads has proved beneficial
in practice. Medical investigations have also shown that the use of oxygen while
under decompression accelerates the desaturation of excess nitrogen from th~
body fluids and tissues. As a by-product of the oxygen breathing tests conducted
on the Queens-Midtown tunnel work the use of helium-ox ygen mixtures for the
alleviation of ear blockage was also developed successfully. A comparison of
the percentages of types of symptoms on several tunnels in New York as indicated
by the table below shows fair agreement with experience gained in the construc-
tion of the Budapest underground.
As the joints, fibrous tissues and fats of the body (which have poorer circulation
than other parts of the body) desaturate more slowly during decompression,
this accounts for the high percentage of symptoms which are classified as various
localized pains in the body.
......- -

-- -- -

HEALTH AND SAFETY MEASUP.fS 833

Symptoms Budapest N ew York Glasgow

<Y.> (%) ( %)

Localized pains in the_joints and limbs 93 89-4 85

Localized pains in the trunk 1·7 0·2
Pains in the central nervous system and 2-2 2·7
the head
Localized pains in the vePlebra 5·2 7-3 8·0
Choking 0·9 0· l
Unconsci ousness (collapse ) 0·l 0·2 4·0

It can be stated in gen~ral that the number of cases of decompression diseases

decreases considerably when a smaller number of more skilled and disciplin
ed
workmen are employed.

REFE RENC ES
ANDREAE, C. (1926): Der Bau /anger, 1iefliegender Gebirgstw mel. Springer
Verlag 1:Se11111
ANON. (1953): Leichtme tallschalu ngen fiir Seolien und Tunnelba uten.
S chweiz. Bauzeitung I
ANON. ( 1953): Untcrirdi sche Strassen und Parkierpl atze. Z. der 0.
I. A. V. 13- 14
ANON. Tunnel u. Stollenba u H ochtief Nachrich ten 1963
A NON. (l 958 and 1964): The Stockhol m Underground Railway ( A technical
description) . Public
Works Commi,s ion
A NON. ( 1961) : F1iras es robba11tas a banyasza tban. (D rilling and blasting
in mines.) Muszaki
Kiad6, Budapest
ANON. ( 1962): flpoxodKa nl//Hlle.1e1i (Tunnel driving). JleHHHrpa.QCKHll:
HHCTHTYT HRlKeHepOB
JKene3HOAOpO)1(11OfO TpaHcnop Ta. C6opHHK 195
ANON. (I 963): T yHHe.111 u. Mempono .1111neHbl (Tunnels and undergro
unds). JleHHHrpa.QcKHil
HHCTHTYT HH)l(eHepoB lKeneJHO,IOpOll<HOro TpaHcnop Ta . C6opHHK
302
BARDOUT, G . and BERNY, P .: La construction des tunnels, galeries
et souterrai ns (Eyroles
Editeur, Paris J 954
CAPLAN, I . B. (1967): ~afontain e Tunnel Montreal , int. Construct
ion, March- April
D sm A, L. et al, (1952): A lagutepite si rendszere k es m6dszere k (Method
s and systems of
tunnel construct ion) Nehezipa ri Kiad6, Budapes t
ELLINGER, M . a nd WEINHOLD, H . (1966): Die Slidbahn unterfiih rung
K etzergass e in Wien
Der Bauingen ieur 1.
HAVERS, H. C. P. ( 1966): Undergro und R ailways at the World, T emple
Press Book.;, London
H ERRMANN, H. U. ( 1967): Planung und D urchflihr ung von Tunnelar
beiten in grossen Profilen,
Sc/llV. Bauzeitung 50
HEWETT, B. H. M. and JOHANNESSON, S. {1922): Shield and Compress ed
AirTwme lling. McGraw
Hill, New York
KASTNER , H . ( 1959): Einige beim Bau des Drucksto llenabsch nittes
Lend gewonne ne Er-
fahrungc; n. Ost. Z. fiir Elektrizit iitswirtsc haft 2
KIESER, A. ( 1960): Drucksto l/enbau. Springer Verlag, Vienna
KMENT, W. and PARRER, F . (1966) : Wiener Fussgang erpassage n.
Der Aufbau. 1-2
LIMANOY JJ11,11a110B, JO. A. (1960): Mempo110.111me11&1 (Undergr ound
railways) :>Ken,1opHJ-
l:\aT, MOCKBa
MENCL, J . and MENCL, V. (1962): Sto/ne a T1111ely (Tunels) Statn.
Vyd . Techn. Lit. Bra-
tislava and Prague
MEREGALLI, G. (1957): La situazion e attuale d elle metropol itane
nel mondo. Transporti
pubblicis

53 Sze<:hy: The Art of Tunnelling

834 CONSTRUCTION AND DESIGN OF TUNNELS

MORGAN, H. D. , HASWELL, C. K. and Prn1E, E. S. (1965): C lyde Tunnel Design , Construc-

tion , Services, Proc. In st. Civ. Eng. Febr.
NIEMANN, H. J. (1965): U-Bahnbau in Berlin , VD! Zeitschrift, 3
PATTERSON, W. H. (1963): Subway Construction in Toronto, Toronto Transit Commission
PEQUIG NOT, C. A. ( 1963): Tunnels and Tunnelling. Hutchinson, London
PROCTOR, R. V. and WHITE, T. L. (1946): Rock Tunnelling with Steel Supports. The Com-
mercial Shearing and Stamping Co., Youngstown, Ohio
RABCEWICZ, L. (1952): Da s Boh r problem im rnodernen Tunnelbau. S chweiz. Bauzeitung 297
RANDZ10, E. (I 927): Der Stollenhau. Ernst, Berlin
RICHARDSON, H . W. and MAYO, R. S.: Pracrical Tunnel Driving. Mc.Graw Hill , New
York 1941
R67SA, L. and K~L~M~N, J. ( 1962): Sch1>.ellzement zur Abdichtung von Tunnels aus guss-
c isc}ncn Ttibings Bautechnik, 5
SAR0SI, SOHA and KELEMEN ( 1960): Bentonit az epitoiparban (Bentonite in Civil Engineering).
Muszaki Kiad6, Budapest
SINGSTAD, 0. and BAILLIE, D. (1956): Tunnels. (Part. 10. in Abbett's Amer. Ch•. Eng. Pract.
Vol. I. ) Wiley, New York
SKEMPTON, 8JrnRUM et al. ( 1960): Frum Theory to Practice in Soil Mechanics. J. Wiley 338
Studia i proje!.ty M erra w Warszawie 1928-1958. W ydanictwo ,,Arkady'', Warszawa 1962
SWIDA, JI. (1959): Yorgespannte Druckstol len, B aurechnik 4
SzECHY, K. ( 1958): Der Bau der neuen Buda pester Untergrundbahn, Ost. Jngenieur z.
SzECHY, K. (1967): Besonderhcitcn beim Bau der neuen Untergrundbahn in Budapest Ost.
lngenieur Z. 4
Ttto~is0N, J. C. (1967): Hori,ontal Earth Boring, Proc. Inst. of Civ. Eng., London, April
VAJDA, H. ( 1959): Tunnelausbau in bedingt standfestem Gebirge, Baupl. · Baut. 1959 3.
V<ilogatott Jejezetek az alag1itepites es b<iny<iszati melyepires korebul (Selected chapters from
tunnelling and mining construction). Mtiszaki Kiad6, Budapest 1952
VRETIJLAD, E. (1959): Tunnelbanebro under Liljeholmesviken (Tunnel construction under
Liljeholmcn Bay), Teknisk Tidsknft 13
VRETllLAD, E. et al. (1962): Yorn Weiteren Ausbau der Stockholmer Untergrundbahn.
Die Bautechnik 2
WIEDEMANN, R. (1947): Srol/enballfen in der neuzeitlichen Tunnelbautechnik . Ernst, Berlin
ZAN0SKAR, W. (1947 u. I 964): Stollen- und T1111nelba11. Springer Verlag, Vienna
... -- - - -

CHAPTER 7

SERVICE, OPERATION AND MAINTENANCE

OF TUNNELS

Generally speaking, all tunnels are rather special and delicate types of structures
from the point of view of design as well as construction. Because of restricted
clearances, poor visibility and the continuous destructive ·action of natural forces
there is special need for careful maintenance, regular inspection and rather compre~
bensive renovation and reconstruction.

7.1. ORGANIZATION AND RESPONSIBILITIES

OF MAINTENANCE AND SERVICE DEPARTMENTS

The problems of tunnel maintenance may be divided into two groups: such
as those concerned with the track or pavement and the others relative to the tunnel
lining. While this differentiation entirely suits real railway tunnels 7· 1 (particularly
subways), 1t is less fitting for highway tunnels and is practically inapplicable
to navigational and service tunnels.
Here are the main aspects relating to the maintenance of railway tunnels:
1. Regular inspection of the tunnel shape, dimensions and clearances;
2. Inspection and maintenance of the drainage system, waterproofing and track;
3. Regular inspection of the condition of the tunnel lining (freezing, corrosion
by fumes and water, weathl!ring, leaching, cracks, deformations etc).

7. 11. INSPECTION OF THE SHAPE OF THE TUNNEL

The shape of the tunnel is usually checked by mechanical or, more recently,
by optical measurements, using the centre line of the track as refer.ence line.
A special carriage with an adjustable platform and telescopic clearance arms
can be used to measure the coordinates x and y, as illustrated in Fig. 7/ 1. As an
alternative the measurements can also be taken from a moving scaffold, provided
that the latter is sufficiently rigid and the track is in a solid condition, i.e. that the
position with reference to the centre line of the tunnel is known and well established.

1-1 BIRKEL: Die neueren Verfahren der Tunnelunterhaltung, Eisenbahntechnische Rundschau

19S9 223
 8~6 SERVICE, OPERATION AND MAINTENANCE

--- The completed sections of the Buda-

pest Subway were checked every 13 ft
from a special scaffold. The dimensions
under observation consisted of the distan-
ces from the centre of the section to the
ends of the vertical, horizontal and 45°
dfameter. The results were plotted to
scale and superimposed on the clearance
diagram so that any inaccuracies of con-
struction could be observed at a glance.
All this was done with a drawing board,
about 24" x 32", mounted on the frame-
work illustrated in Fig. 7/2 and set up in
the plane of the section to be checked.
FIG. 7/ 1. Carriage with
The plane itself was determined by a
cross-sectioning frame
chalk line between the two ends of the
horizontal diameter, using a level to set
A
the board in a vertical position.
The board itself and the sheet of paper
2
on it should be positioned near the axis
of the tunnel; the point of intersection of
the latter on the sheet can be determined
as follows . The theodolite is set up over
B the nearest station and backsighted to
J the one before. With the upper plate
clamped, the telescope is turned about the
horizontal axis and a point is directed on
line on the sheet. Now the upper plate is
turned to backsight again to the station
before, the telescope is turned agaiu 2nd
F another point is marked on the sheet. If
Fm. 7/2. Cross-sectioning frame used in the two marks do not c0incide, then the
the Budapest subway half way between them is to be accepted
as the one required. A vertical line is
drawn next with the aid of a level through
this point and the axis (0) is determined and marked using a level and a levelling rod.
this is followed by marking the horizontal, vertical and 45° diameters on the sheet.
With one end of the tape held at point 0, the distances can be measured along
the marked diameters to points B, J, A, F, I, 2, 3, and 4, re;pectively.
In the Moscow Subway as well as in some western European subways the
fixed frame is substituted by feeler swinging arms and telescopic rods.7 · 2

1-2 KoPPENWALLNER, F.: Lichtschnitt Profilmessung in Stollen, Geologie und Bauwesen

1959 25 I
-
-
-

ORGANIZA TTON, INSPECTION 837

FIG . 7/ 3. Cross-sectioning with

projected light

A better degree of accuracy can be achieved with the optical clearance diagram
projector, developed recently 7•3 • This projector is cylindrical and emits a band
of light (d wide) all around, clearly outlining the tunnel lining in that particular
section. This image can be photographed with a camera, mounted concentrically
with, and at a certain distance from , the projector. A reference distance of known
length should also be photographed. en the same picture so that the outline of
the tunnel lining can be evaluated precisely with reference to the axis of the projec-
tor and the camera. (Fig. 7/3 indicates an eniarged detail of the section enclosed
in the square inset.) For best results and sharpest pictures the width of the light
band, d, should not exceed 1/ 100 of the radius of the section. This procedure
can be developed readily into a quick method for checking and comparing a num-
ber of sections, provided that the relative position of the track and the carriage
is fixed with reference to the tunnel axis. In practice, the accuracy of this method
has been shown to be about 1/ 16"- l/8".
A special type of instrument has been designed in the Soviet Union for measur-
ing the projected line on a trigonometrical basis7.4. This instrument consists
of a revolving horizontal axis mounted on an upper plate with foot screws. At
one end of the horizontal axis and perpendicular to it is a fixed telescope with a
built-in light; at the other end there is another telescope which can be turned in
a plane passing through the fixed telescope and parallel to the horizontal axis.
The distance between the two telescopes (A) is the base length. The point projected
on the lining by the fixed telescope is to be sighted on by the second one; by
reading the vertical angle the ordinate e can be calculated from (! = A tan y
(Fig. 7/4). To cover the entire section the horizontal axis has to be lined up with

7 3
- Transportnoie Stroit. 1959 and Bautechnik 1960 6 222
7·• MI NJN,P. I.: Pribor dlya proverki gabaritnosti tunnelei (Device for measuring the
clearance of tunnels) Transportnoie Stroit. 1959 1
-----
- - - - ------=-- - - - - --- --- - --

838 SERVICE, OPERATION AND MAINTENANCE

I/·' I the tunnel and the first (project-

I
ing) telescope has to be turned
about the horizontal axis; each
point so projected has to be sight-
ed on by the second telescope
and the polar coordinates of each
point can then be calculated from
the recorded angle.
Revolvmg Deformations of the lining,
telescope and particularly the development
of such deformations as a
fixed projector function of time will provide val-
telescope uable information on the forces
acting on the lining and on the
FIG. 7/4. Optical cross-
sectioning on a trigonometrical basis resistance of the· latter, even
though the picture is often dis-
torted by inaccuracies in construc-
tion. This, of course, is another reason for starting the measurements immediately
after construction and for keeping up this practice over a long period of time.

7.12. INSPECTION OF THE TRACK DRAINAGE AND INSULATION

The condition of the track as well as the waterproofing and drainage system
should be checked regularly in order to detect any discoloration of the walls, con-
densed water, water seeping through the joints, efflorescence and corrosion. It is
particularly important that the components of the drainage system, such as the
subdrains, drop drains and laterals be checked regularly for silting, clogging
and cracking.

71.21. Inspection of the Tracks

The inspection of the tracks should include a check on the rails and ballast;
it should also detect a ny dislocations that may have occurred in the track layout.
This is particularly important not only because tracks and rails have been observed
to wear faster in tunnels than on the surface (see Section 72. 18) but also because
of the heavier losses suffered· if trains are derailed in tunnels. Since broken rails
can be the sources of disastrous accidents, the inspection of connections, fasten-
ings and the general condition of the rails should be carried out with exceptional
care. For this reason and because of poor visibility it is not sufficient to have daily
routine inspection runs but the condition of the rails should be checked by tapping
as well, and special mirro;:s should be used to facilitate inspection from underneath.
In addition to this, the gauge of the track an9 _the soundness .of the rails in sub-
ways have to be checked with special cars (defectoscopes) about twice a month.
 ORGANIZ ATION, INSPECTI ON
839

for disloca tions.

The horizon tal and vertical alignm ent also has to be checke d
to the third rail
bumps and sags. In the case of subway s this also applies
d with referen ce to a
(power rail). The positio ns of the tracks should be checke
at closer interva ls along
fixed networ k tied into the walls and this should be done
curved section s than on straigh t ones.

71.22. Inspection in the Budapest Subway

are carried out
In the Budape st Subwa y the laying and inspect ion of the tracks
on the following basis:
ate system
As mentio ned earlier (Sectio n 5.21), a geodeti c networ k and coordin
structu res are being designe d
must be already availab le when the tunnels and
with great precisio n, particu larly
and laid out. The laying of tracks has to be done
ent of the trains at high speeds
on curved sections, because the smooth movem
depend s partly on the precise laying of the tracks.
design even
Because the actual layout of the tunnel will always differ from the
ng and in constru ction, the final layout of the tracks
with the utmost care in surveyi
precise cross-s ectionin g of the actual tunnel has
can only be determ ined after
n 7.11). In the design the curved and tangen t section s of
been comple ted (Sectio
ce to the geodeti c
the track alignm ent are given in the form of equatio ns with referen
coordin ate system.
k has to be
While the tunnel is being cross-s ectione d a reference point networ
checke d regular ly later
establis hed from which the tracks can be laid out and
with metal plugs
while the tunnel is in service. This networ k can be marked
opposit e positio ns and at equal
(screw spike5) embedd ed io the concre te lining in
markin g the referen ce point
heights, about every 50- 70 t't, with a 1/ 16" hole
an underg round
itself. Such points must be referen ced by orthog onal coordin ates to
precise ly into the coordin ate system of
travers e, which in turn should be tied in
For each point the offset from the under-
the surface triangu lation net (Fig. 7 /5).
ined with a transit set up over one of the adjacen t
ground travers e should be determ
the travers e leg in questio n; the distanc; e can be read
station s and lined up with
on the tape. The latter should be under sufficie nt pull
directly by the instrum ent
the minimu m
and moved slowly back and forth to enable the observe r to take
reading pertain ing to the distanc e at right angles.

FIG. 7/ 5. Referencing the completed tunnel for track layouts

--
--=--- --- - - - - - - --
-

840 SERVICJ:, OPERATION AND ~NANC E

The accuracy of this operation can be improved by measuring ' the ,dist~lices-
between opposite markers and, with the ordinates left unchanged,' adjµstihg the
abscissae as follows:
Let o1 , o 2, o 3 and a4 denote the abscissae of points 1, 2, 3 and 4, r~pectively
(Fig. 7/5), with the respective ordinates b1 , b 2, b8 and b4 and with s arid · s~
12
denoting -the distances between the respective reference points. Thus ·

{7.:1)
In most cases
(1.2)

where (oJ and (~J = the measured distances.

Introducing L1o as the difference between tlie two diff~rehces Wt; get tpe a4J~s-
ted abscissae as -

(7 .-3)

Jn Eq. 7.3 the sign of o 1 is positive and that of tz2. is ne~ative-(Fjg. "?/5) if

(a1) - (a2) < 01 - 02,

and is reversed jf

If the co.ordinates are calculated with the adjuste<;l abscissae then the ca_l,cl)lated
distances will e.lways agree with those actually measured.
Once the coordinates of the reference points have been determined, the. poi11ts
of intersections of the tunnel axis with the lines between opposite markers can
be calculated and actually laid out.
While the reference points are being measured in, the other characteristic
points (columns, platform corners, etc.) are also to be included in the measure-
ments, although not necessarily to the same degree of acc;uracy.

7.13. INSPECT ION AND MAINTE NANCE OF LINING AND TRACKS

The inspection and maintenance of the tunnel lining and the ~racks ~re usually
done together. In the case of the Moscow Subway, how_ever,_tliey were separated
because of the particular importance of the matter. The main object of the lining.-
 I

ORGANIZATION, INSPECTION 841

inspection is to detect any spalling, deterioration of joints, wet spots, cracks,

leaching and to determine their spread and also to correct any distortions and
deformations that may have been found. It is most important that the formation
of cracks be checked and followed through regular inspection. Such cracks are I
most likely to develop with changes either in the wall thickness or in the external
loads. The most sensitive points are the portals and the recesses in the linings.
Changes in external loads may be caused by the physical properties of the soil
as well as by abrupt changes in the formation of the surface ground (steps).
Cracks caused by changes in the loads or in the cross-section are perpendicular
to the tunnel axis. Similar transverse cracks will appear as a result of uneven
longitudinal support if the tunnel is passing through different layers that are
perpendicular to its axis. On the other hand longitudinal cracks, i.e. those running
parallel to the axis, can be traced to inadequate load-bearing capacity of the tunnel
section. Cracks of this type usually appear along the crown or the spring line.
The immediate cause, of course, is either an excessive load condition (not allowed
for in the design) or unexpected bearing conditions. The walls, for example
may have moved outward either because of the low coefficient of subgrade reaction
o( the surrounding ground or because of voids left between the walls and the
excavated rock face _in the course of construction. This is certainly the cause of
cracks showing a widening trend toward the intrados. In the opposite case the
walls must have yielded under excessive lateral pressures, which can also be mani-
fested by cracks along the spring line widening toward the inside. (The cause can
also be the lack of an inverted arch.) Cracks may also develop along the bottom
of the walls, being an indication of foundation settlements. Vertical movements
in general do not cause cracks unless they are uneven, in which case the cracks
are diagonal in direction and are relatively minor.
The above causes may result in quite a number of cracks running parallel to
the centre line of the track. Short cracks are an indication of local overstress and/or
flaws in the lining material or deficiencies in workmanship: In fact these erratic
and local cracks may be caused by a wide variety of causes (water table, frost,
smoke, draft, etc.).
It is very important to know whether the size of the cracks is increasing or
not. For this purpose the ends of the cracks should be marked and any increase
in width should be measured between markers (metal insets, wooden plugs)
embedded in the wall on either side of the.crack. Minor cracks can also be pasted
over with a piece of glass or plaster. By observing the distance between the markers
regularly the widening, 1breathing', of the cracks can be plotted as a function
of time.
If the crack has ceased to show any signs of increase it should be closed
in order to avoid damage caused by air or water. It should be chiselled out and
filled with good quality mortar {shotcrete) in such a manner as to ensure
that it will stick (dovetail shape, wire mesh anchorage etc.). Should the crack
continue to increase, the cause should be cleared and remedial measures taken
acco!dingly.
842 SERVICE', 01'.ERATIO.N AND ¥AINTENANCE

ln addition to checking deformatioos and wall movements· inspection ~h~mld·

also be extended to detect cavities behind the liaing. In the· case of-thin secti9ns
this can be done by tapping on·each storie or ~n eyery square yard of monolithic
lining.
A muffled, hollow s.ound js always an indication of a hole ·behind the . wall.
In the cas~ of heavy se_ctioil:S tapping is notJ~asible and such ·s~tions him: to be
checked by means of dynamic, electric.instruments .Qr radioactive isotopes similar
to thds~ used in the· investigation ofsoils. · The cavities so discovered. ha~e to be
marked on th~ wall for further observations and/or repair. It.is _commonly·known.
that holes behind the. walls are detrimental,_not only because ·of the loss ·of compo.-
site action with the surrounding ground, but also because of the resulting c.rosive
action of groundwater and the spread of we·~thering, all of which .m·ay_cause.
entire blocks of rock to .separate and add.to the loads on the lining as well as,- -tp
further enlarge the cavity itself..· · ·
With linings consisting o.f cast-iron. _segments it is usually ·sufficie~t to tap. on
them and to ·check on the grouting h·oles. The amount .of water disc~arging_.after
removal of the plugs from these hole.s s\lould give an idea of the size of the ·oack~
space (cavity), -the direction and maximu~ d1~ensions of whi¢h can be explored
with a piece of test wire.
The maintenance and cleaning of th.e bench~s; decorative tiles and of.the entire
station area also have to be provided for. In the Moscow Subway regular c;leaning·
is effected by special mechanical brushes within the stations and by hosing be-
tween the stations.

7.14. ORGANIZATION OF MAINTENANCE SERVICES

The organization of maintenance services is bound to be more. elaborate in

subway systems which form an interconnected tunnel network often -~l!nch-ed~-
of miles long, than in single railway or highway tunnels where_suC:h; _.service, is.
necessarily only a branch of the general track or highw~y maintenance and wiie~e
less special attention can be given to ali tµe aspects of maintenance of' the tut?,~~l
sections. Iii subways, the maintenance of the _tracks as well as ·that of the ti.nin~i
itself is divided into sections, centered around the stations .and/or storage arid
maintenance yards. Materials and small working. tools can pe i.i~red unde~···the
platforms of the stations (Fig. 6/167); small offices can als9 ~e housed within
the ~tations. Storage yards usually have their own workshops ~µd also· accommo-
date the heavier machinery, equipment_and :inst1'."umeQts.
Ano.t hei:item to be mentioned here is the importance of fire st_ations an:(j .h~l!.VY:-·
duty. fire e}(tinguishers which should always be kept ready for use. rti'e da_ng~i;
of sparks genera,ted by electric subway trains combin~d with..dl!5t, 'oii, vari~Js
electric cables and ties soaked ii) grease and oil _represents a <,qp.stants· fire haza.rd
and requires alertness and preventive safety measures.
 -
- - -- -- - - - - - -
~

DETERIORAT ION AND REPAIR 843

7.15. STORAGE OF MAINTEN ANCE EQUIPMEN T

Maintenan ce materials and equipment are stored in various utility and service
rooms which may also accommod,ate pumping stations, ventilation plant, lighting
and power plants, etc. Other premises, serving passengers and traffic in general,
and housing mechanical equipment have been dealt with earlier in Section 63.34.

7.2 DETERIO RATION AND REPAIR OF TUNNEL S

Because of the importanc e of railways in transporta tion and because of the

practical limitations on maximum gradients most of the tunnels existing today
are railway tunnels, built in the latter half of the past century, i.e. during the boom
period in railway constructio n. This also means that most of our tunnels are
just about 100 years old, a span of time long enough to warrant the reconstruc tion
of almost any type of engineering structure. In the case of bridges, factory buildings
and other structures the need for reconstruc tion come,; with increased loads
and traffic for the most part, although it may also be warranted by other reaso~s
such as fatigue of materials, weathering and other forms of decay. As far as
tunnels are concerned the increases are rarely in the volume of traffic, but rather
in additional clearance requiremen ts (electrification, second track). Increases in
natural loads are more common (ground pressure) but are still less significant
than the destructive agents of weathering and the inadequate strength and resis-
tance of old building materials.
As the developme nt of railway networks may well be regarded as being complete
in Europe as well as in North America, and because of the world-wide reduction
in the rate of constructi on of new railways, it is safe to say that the main problem
with railway tunnels is not concerned with the constructi on of new ones but rathei
with the remodelling or reconstruc tion of old ones. As far as navigational tunnels
and aqueducts are concerned , however, the situation is different as most of these
are being built at the present time. The same goes for highway and subway tunnels,
both of which have been started fairly recently and both of which are still far
from being fully developed . In these cases, then, the problems of new constructi on
outweigh those concerned with reconstruc tion. We enjoy today, of course, the
benefits of better and more durable constructi on materials and the benefits of
the experience derived from the constructio n of the older railway tunnels. All
these factors have much to contribute when it comes to arresting and retarding
natural deteriorati on. 7 5
In his review of the existing railway tunnels in Germany · BIRKEL arrives at
the conclusion that all these tunnels - unless they have been repaired at least
once already - are in poor condition to a larger or lesser extent. Actually the same

,.. BIRKEL, F. : Gegenwarts aufgaben im Tunnelbau, Eisenbahnte chnische R11ndscha11 1952

Dec.

.
 844 SERVICE, OPERATION AND MAINTENAN CE

is true for any other country. This is why most tunnels to be dealt with in the
following section will be railway tunnels.

7.21 . CAUSES OF DETERIO RATION OF TUNNEL S

The causes of the deterioration of tunnels can partly be traced to passive ones
such as poor building materials and construction methods and partly to active
ones such as the external loads. Practically speaking, the two groups cannot be
separated as clearly as that, because the active ·caus·es also affect the weathering
of building materials and in the final analysis the resulting deterioration is a prod-
uct of the effects of both groups.

72.11. Deterioration due to Defective Materials and

Workmanship

One of the most importan t causes is improper excavation, loosening up extensive

areas in the surrounding ground, prolonging the consolidation period and also
opening the way to the erosive action of ground water (excessive blasting, improper
temporary supports, etc.) Next in order are the old building materials of inferior
strength and durability, affording less resistance to the forces of nature. Materials
of this type include the brick, soft sandstone, limestone with Marl, slate, certain
types of cement, plaster made with lime, impervious plaster and waterproofing
sheets containing organic fibre materials subject to decay. These materials have
much lower resistance to water, frost, smoke and atmospheric pollution than
have the modern high strength and well compacted concrete, mortars made with
Portland cement, waterproof coatings and plastic and foil waterproofing. This
is why old tunnels deteriorate to a much larger extent than do their modern
counterparts. Poor construction methods also have a part in this; in the old
days tunnel linings were not built right up against the excavated ground face
but were separated from it by loose backpacking that was poorly grouted or
left ungrouted. Another factor is poor workmanship, particularly common in
the superficial application of mortar between lining blocks, resulting in poor
embedment.

72.12. Deterioration Caused by Water

The active causes and the effect of water in particular are more importan t as
they damage the tunnel in more than one way. Groundwater will stream towards
the tunnel like to a subdrain and meanwhile will have a number of mechanical
effects such as wash-outs and silting up the drainage system, enlarging the holes
behind the lining and loosening the surrounding ground. It will also dissolve
certain chemicals; its carbonate and chloride content will attack concrete and
 DETERIORATION AND REPAIR 845

FIG. 7/6. Icicles in a tunnel

metal members, its sulphate content will be detrim~ntal to concrete and to the
cement mortar in the joints. Even clean, unpolluted water is bound to dissolve
the lime of mortar and concrete. The most severe damages, however, are those
due to physical effects. The first of these is frost and the bursting action that goes
with it (Fig. 7/6). The water resulting in frost or ice action may com·e not only
from leaching through the joints, and through the lining itself, but also from
condensation water (natural humidity as well as that caused by steam locomotives).
Short tunnels, and particularly the vicinity of portals are most liable to suffer
from frost. It has be;en observed that tunnels running parallel to an .easterly or
northerly wind direction will experience the heaviest frost damage. Frost will
affect the railway tracks as well, partly through the formation of icicles (water
dripping on the rails and freezing up) and partly through freezing the ballast.
All this may lead to broken rails and derailed trains. The frozen ballast is less
elastic and creates unpredictable uneven support conditions which may cause
the rails to break.
Another physical hazard caused by water lies in the saturation of the surround-
ing cohesive materials accompanied by a reduction in strength and an increase
in compressibility. For example the softening of the ground under the walls may
lead to excessive settlements, tilting towards the inside and cracking in the arch.
If, for example, this softening occurs in a tunnel with no inverted arch, the plastic
1'146 SERVICE, OPERATION AND MAI NTENANCE

clay may be pushed from under the walls into the ballast and. may dangerously
impair its strength and elasticity.

72.13. Damage Caused by Smoke

Smoke7 • 6 also can be a source of damag,e but is less dangerous than water and,
in fact, would have to be accompanied by water to be dangerous. The sulphur
dioxide in the smoke will form sulphuric acid with water, and this is aggressive
to cement and lime. This effect, however, has been observed to be of little danger
and to be confined to the surface. Old mortar made with lime, for example, was
affected only to a depth of a few centimetres and the same effect was only a matter
of millimetres in the case of cement mortar. Slightly more extensive scaling was
observed on relatively soft limestqne and sandstone, whereas the change in the
structure of vitrified brick and shotcrete was limited to 1 or 2 mm only. Stones
with a dense structure are not affected by the sulphates at all. The reason for this
lies partly in the relatively short length (a few hundred yards) of the majority
of tunnels, where natural ventilation is very effective. In frost-susceptible materials
the effect of smoke combined with frost is more destructive and can penetrate
to ·greater depths. In any case smoke is more dangerous to iron components
because it greatly accelerates the corrosion of iron in the damp atmosphere. In
the case of cracks in the concrete cover this will also apply to reinforcing steel.
The corrosion of reinforcing steel, in turn, will result in more extensive concrete
spalling through expansion and this is bound to intensify.

72.14. Damage Due to Atmospheric Conditions

Atmospheric conditions and temperature changes also affect deterioration.

Among these are the mechanical effects of drafts, and cracking arising from
temperature changes; combined with drifving moisture in the air these can be
summarized as weathering caused by atmospheric action. The weathering of
stone materials used in tunnel linings varies with their crystalline structure and
with their sensitivity to the dissolving and aggressive chemical actions of water
and smoke.

12.15. Impact Damage

Impact effects under rolling traffic may also contribute to the decay of tunnels,
not as a primary cause, but as a factor accelerating and aggravating general deterio-
ration. It is particularly detrimental in widening the cracks and contributing to
the saturation of the ground as well as to its remoulding.

7-6 RAAB: Die Schadgaskonzentration in Stollen und Tunneln, Eisenbahntechnische Rund-

schau 1959 410

 847
DETERI ORATIO N AND REPAIR

12.16. Damage Caused by Overburden

more than one form.
Damage caused by the overburden pressure may take
overb urden pressu re which will occur primarily
The most dangerous is the actual
• (see Section 31.2).
7 7
in tunnels at great depths and in pseudosolid materials
depths in which case it
It is not uncommon in clayey groun d even at shallower
and cracks due to actual
is associated with squeezing pressures. Deformations
in limestones and clay
overburden pressures are most likely to be encountered
marls of the Tertiary Era.
by wash-outs and by
Increases in overburden pressures may be broug ht about
is why nowadays the objec-
the loosening effect of the groundwater as well. This
of the tunnel and push it
tive of tunnel waterproofing is to keep the water out
interc ept it and drain it inside .
back into the ground rather than to attem pt to
ses in the overb urden pressure is
Anoth er possible cause of subsequent increa tunnel
ped equili brium aroun d the
the disturbance and disruption of the develo or
cavitie s or by the install ation of exteri
by any subse-quent excavation of adjacent and in
result in pressu re superp ositio n
type waterproofings etc. All of this may be expec ted
zone. Proble ms of this nature are to
further extension of the loosened
due to excessive charges
particularly with excessive blasting operations either
or when too long drill-holes are used.
It is common experience that tunnel' portals are subjec
ted to considerable addi-
of the suppo rted slopes.
tional pressure due to progressive weathering and sliding
separated from the actual
Because of these differential pressures the portal s may be
develop over the entire
tunnel itself. Additional pressures of this type may also
frost action and in-
adit section of the tunnel. It goes without saying that both
the adit sectio n to the largest
creases in the pressures are most likely to affect
extent.

12.17. Deterioration in Pressure Tunnels

e spots are the cracks
Turning now to pressure tunnels the most impor tant troubl
d by the hoop stresses
and the corrosion of reinforcing steel in the lining, cause
may be a numb er of
in the lining and by the escaping water, respectively. There
of the high allowable
reasons for the development of cracks: one only has to think
able gaps and imperfect
stress (and strain) in the reinforcing steel, of the inevit
of the excessive blasting
grouting between tunnel lining and surrounding rock,
the vast difference between
operations and the resultant loosening all aroun d and.o f
nding groun d. The absence
the modulus of elasticity of concrete and the surrou
water throu gh the cracks.
of groundwater can also expedite the escape of

zerische Bauzeilung
7-7 R OTHPLE TZ, F.: Woran leiden unsere Eisenb ahntun nel, Schwei
1918 Febr.
 848 SEP.V!CE, OPERATI ON AND MAINTEN ANCE

72.18. Damage to Railway Tracks

High humidity, corrosion and unfavourable maintenance conditi

ons have
rather adverse effects on the tracks in railway tunnels. Because of
limited space,
poor visibility, inadequate ventilation and limitations on the time
available for
maintenance the cost of track-maintenance in tunnels is twice as much
as on the
surface . Even the rails will wear faster because of the corrosive atmosp
here and
because of the skidding and grinding of wheels on the wet and often
sand-strewn
surface. Steel ties are not practicable for use in tunnels because of corrosi
on and
because of the creatio n of undesirable noise. As a result of these conside
rations,
the object today is to reduce the labour required for maintenance,
leading to the
use of heavy rail sections over extended lengths and to a solid connec
tion between
the tracks and the bottom of the tunnel. The above considerations
have led to
special track systems, particularly in subways. 7 • 8

72.19. Maintenance of Cables

Because of the danger of corrosion the various cables, gallerie5,

lights, etc.
inside the tunnel also require special attention. In railway tunnels used
by steam
locomotives, for instance, the electric cables have to be protected by a
lead coating.
On the other hand, in subway tunnels the danger of stray-currents
- due to the
electric trains - and soil corrosion require extra care in the insulati on
of all metal
parts.

7.22. REPAIR AND RECON STRUC TION OF TUNNE LS

Repair of tunnels generally comprises the following items :·1•2

1. Repairs to the drainage system to avoid frost and icing damage
to the lining;
2. Reinforcement and/or replacement of sections of the existing lining
because
of inadequate strength;
3. Reinforcement of face lifting of.certa in parts and members of objecti
onable
appearance;
4. Additional clearance requirements, such as those required for the constru
ctiop
of a se.cond track;
5. The additio n of utility rooms and plants (refuge niches, ventilat
ing plant,
storage rooms, weepers and subdrains, shafts, pumping stations, etc.).
Items 1, 2 and 4 will be considered in detail below; the others are
of specific
nature a::id can be carried out accordingly.
7·8 DANDUR OV M.
I.: Tonneli (Tunnels). Gos. Transheldor. Zdat., ~/.oscow 1952
Cf. also: PATERSO N, W. N.: Subway Constru ction in Toronto Address 513.
to International Cong-
ress in Berlin on Underground Traffic Facilites 1963 and Cf. also:
LUCAS, G .: Der Tunnel.
Ernst. Berlin 1 920
7· 9 DANDUR
OV, M. I.: Tonne/i (Tunnels). 389
 DETERIORATION AND REPAIR 849

72.21. ReJlairs to the Drainage System

As mentioned earlier, the most severe damage is caused by water and, thus
reconstructio n and repair work, in general, should be aimed primarily at protection
against water. There are two alternatives for such protection: the water can be
kept out of the tunnel altogether in which case the infiltration of groundwater
should be prevented and it should be sealed out completely, or the water may
be allowed to penetrate the tunnel but in a controlled manner through special
subdrains so that it can be intercepted and removed regularly. A disadvantage
of the first method is that the tunnel lining has to carry the hydrostatic pressure as
well as the overburden. It has to be constructed in a box-like (closed) form (with
an invert); in addition, waterproofin g can be also rather complicated and expensive
to construct, particularly if it is done on the extrados. On the other hand, there
is no danger of disturbing the equilibrium of the surrounding water-bearing
ground, no continuous underground streams and no danger of wash-outs accom-
panied by a loosening and saturation of the surrounding ground. The second method
is not without drawbacks either, because the construction on an interceptor
drainage system behind the tunnel lining is a complicated and expensive proposi-
tion, not to mention the problems of maintenance and potential damage by wash-
outs, loosening of the surrounding ground and increases in the overburden
pressures brought about by the facilitated streaming of water directed towatds
the tunnel. The construction of special drainage galleries and adits is also expensive,
and it relieves the immediate vicinity of the tunnel from wash-outs and loosening
only to a limited extent. This is why the first method, i.e. exclusion of water
s the one being favoured today.
72.211. Repairs aimed at the exclusion of water. It is a common observation
in old tunnels that the lining is often not in tight and firm contact with the excavat-
ed face and the overbreak has been filled by stone backpacking, acting as a filler,
laid in lean mortar or with no mortar at all. While this layer was supposed to
intercept and drain the groundwater this could not be done without the cement
being dissolved and without wash-outs developing in the surrounding loosened
ground. This has led to the collapse of the " stone packing", leaving innumerable
holes and cavities behind, which in turn have induced additional overburden
pressures. The best way to repair a tunnel suffering from such defects - and to
solve the problem of drainage at the same time - is to remove the old lining, ring
by ring, and replace it with a good quality, solid and impermeable concrete lining,
built tight up against the natural ground. If the inner load-bearing wall is in an
otherwise acceptable condition, it may be worth while to try to grout the stone
fill behind the lining thoroughly. By the way, grouting _is not only an effective
and economical way of waterproofing, but it is also very good for reinforcing exist-
ing linings and for reducing overburden pressures. This is particularly effective
for near-circular sections, because the pressure due to grouting will create ring-
stresses, which, while adding to the compressive stresses, may also pull in the thrust

54 Szcchy : The Art or Tunnellin&

850 SERVICE, OPERATION AND MAINTENANCE

line within the inner third. At the same time, it will fill the holes and cracks and
will eliminate loosening in the surrounding ground to a large extent. Furthermore,
it will increase the modulus of elasticity of the ground, reduce the external pressures
on the lining and promote composite action between lining and supporting ground
(Section 63.22). It should be remembered, of course, that hydrostatic pressures
are also bound to increase as a result of grouting. The grouting material should
be selected with the pore content and void ratio of the surrounding ground kept
in mind. While in the case of sizeable voids the grout should consist of a cement-
mortar made with coarse sand, soils with a lot of fine particles should be grouted
with a cement slurry, with sodium silicate and bentonite added as plasticizing
agents. Grouting holes should be spaced evenly about every 2-3 m (7-10 ft)
around the perimeter with variable penetration depths up to 6 m (20 ft) deep.
Reference is made to Section 63.22 regarding the details of grouting operations.
(It should be noted that if, in the course of repairs, grouting is applied only
to certain sections that appear to be saturated or show other signs of damage,
the displaced water has to find some other outlet and may leach through other
sections of the lining that were dry before, thus creating new problems and potenti-
al frost-susceptible trouble spots.)
Another way to exclude water from the lining is to apply some waterproof
coating. Of the two alternatives the exterior type waterproofing has the advantage
that in carrying the hydrostatic pressure it is supported by the lining, which also
means that there is no need for a special internal load-bearing wall. Also, it will
protect the lining against aggressive groundwater. On the other hand, it does
have the drawback that its complementary application is very difficult and expen-
sive and almost impossible to achieve in a perfect manner. As it requires the exca-
vation of an additional annular space outside the extrados of the lining (Fig. 7/13)
it will inevitably lead to loosening the surrounding ground and, thus, increasing
the overburden pressure. The subsequent lining and backfilling of this space
require much material and labour, by far exceeding the cost of the interior type
waterproofing including the inner load-bearing wall. And, still worse, any imper-
fection or damage inflicted on the exterior type waterproofing is extremely difficult
to locate and the extrados has to be re-exposed for repairs.
The subsequent application of exterior type waterproofing has to begin with
cutting a hole in the lining and excavating a parallel drift either at the crown or
at the spring line (Fig. 7/7). The application of a crown-drift is to be avoided if
possible at all, because it will require special props in the vicinity of the hole
rendering access and material transportation extremely difficult. Before placing
the exterior type waterproofing, the extrados has to receive a mortar finish which
must be artificially dried. Regarding the quality and preparatiqn of the water-
proofing the reader is referred to Section 46.2. The simplest form of the exterior
type of waterproofing is grouting, provided that this is feasible. The materials
for grouting include various cement mortars and suspensions with plasticizers
and bentonite added as admixtures. Hot-poured bitumen may be used for this
purpose also and for the repair of waterproofing, too.
_,_,

DETERIORATION AND REPAIR 1!51

a) Bre.1k-out ~t springing
Protective
waterproofing

' Longitudin,1/
1 3d!l

b) 8re.1k-out at crown

.~-.
_,

Existing ·
/ming

Arch support mad~~

of r3t/ section

Fm. 7/ 7. Break-outs facilitating lining repairs

and the subsequent construction of exterior type waterproofing:
(a) at the springing
(b) at the crown

Once the tunnel lining has been completed it is much simpler and less dangerous
to apply the interior type of waterproofing as a supplement. Admittedly, interior
type waterproofings do not protect the lining from aggressive chemicals and they
 852 SERVICE, OPERATION AND MAINTENANCE

do require a separate support to resist hydrostatic loads. A practical method of

application is to rub down the intrados and provide drain grooves in the lining
(Oberhasli-method, Fig. 7/8), apply a coating of bituminous or plastic sfieets and
construct the required load-bearing reinforced-concrete rin:g inside. In the simpler
cases the drain groove system is covered by impervious shotcrete only, in which
case there is no need for a load-bearing wall because any small quantity of leaching
water will be removed by the drains thus preventing the building up of a head .

M1in p l/tr!/

Livers of imperviovs p/1ster

Stction A - A

d ?0-40 mm met1/ or pkstic pipe

FIG. 7/8. Surface drains provided after censtruction

It is most important with interior type waterproofings, however, that the lining
should be in a consolidated condition and that it should be dry. In other words,
it can best be used in solid and stable rock where further deformations of any
consequence are not to be expected. Whenever there are considerable overburden
pressures to cope with, the waterproofing should not be applied until the settle-
ment period is over. Another problem may arise from the settlement and shrinkage
of the inner load-bearing reinforced-concrete ring. In such a case the disconti-
nuities and cracks may be repaired by grouting with hot asphalt. Unfortunately
the joints of grouting tubes to the waterproofing are usually vulnerable spots.
Recently, plastic sheets have also been used successfully in interior type water-
proofings (Section 46.2). It takes only minor preparations to the old lining (rubbing,
drying) and does not require the lining to be dried out completely. Thermoplastic
splicing of the plastic sheets eliminates the usual problems of joining waterproofing
units and affords perfect impermeability and unimpaired strength. From inside,
the plastic sheets are supported and protected by precast-concrete blocks. Since
 DETERIORATION AND REPAIR 853
there is no hydrostatic pressure
to carry, this lining need not be
more than 6" - 8" thick, which
can be accommodated" in most
cases from the point of view of
clearances (Fig. 7/9). If there is a
continuous infiltration of water
as well, it is best to collect Jhis
by the drain-net illustrated in Fig.
7/8 before the waterproofing is
applied and to drain it through
a bottom conduit.
It is more economical and still
simpler to cover the prepared,
cleaned and roughened lining FIG. 7/ 9. Interior waterproofing with plastic sheets
with shotcrete. This can be made provided after construction
self-supporting by using a wire
mesh, anchored in the wall and Existing concrete
applying the shotcrete in layers _..--- !tnm;
to an overall thickness of several ~ 1t8mm oper 10 cm a)
inches (Section 61.315). Through
the anchors the resulting solid
self-supporting ring will also
bring about a certain amount of
composite action between the
new and the old lining (Fig. Ex,stm; Imm; lo be
roughened, with dmr.
7/ 10). With this method it will be grooves ifreq/J/red
Old cut stone blocks ( Oberhas/1)
necessary, however, to dry out the
existing lining completely and/or
to provide the drain grooves
(Oberhasli) described above. Anchor
A new method, called "Aero-
.cem"7· 10 can also be used with
good results for waterproofing.
An air entraining agent (Terpol) New aerocem !ming
is added to the water before the
latter is added to the mixture of Fm. 7/10. Anchorage and structure of
dry sand and cement, resulting in reinforced-sho tcrete lining (" Aerocem")
a shotcrete plaster with entrained
air. The entrained air renders the plaster more elastic and more suitable for un-
interrupted application and enables it to penetrate deeper into the joints (Fig. 7/ lOb).

1-1o PROTOPOKAD!S, P.: A New Method for the Mechanical Pointing of Brick Work,
Civil Eng. and Public Works Review 1954 May
854 SERVICE, OPERATION AND MAINTENANCE

The application of impervious cement mortar is an even simpler method of

waterproofing. Details of its composition and preparation can be found in
Section 46.2.
In old tunnels it is most important that the old surface be stripped completely
off the weathered layers before plastering and that the joints be scraped out
thoroughly. It may also be necessary to apply sand blasting and to use wire mesh
reinforcement properly .anchored into the lining.
Interior type waterproofings have been constructed most recently with plastic
materials shot at the existing lining. The 'Torkret GmbH.' construction company
has used epoxy resin and fibreglass materials for this purpose after drying the
lining with infrared lamps and has produced a material that adheres to the wall
perfectly, has a tensile strength of 1500 kg/m2 (21 000 psi) and is highly resistant
to acids, alkalines and tear and wear. Because of the perfect adhesion a supporting
wall is not required. However, experience with materials of this type is rather
limited at present (cf. also Aquarex, made in Austria).
If the amount of infiltrating water is relatively small and is entering through
the joints only then the repairs may be limited to scraping and cleaning the joints
thoroughly and filling them with waterproof mortar. Under dry conditions this
may consist of applying a number of layers of various mixes; in a wet environ-
ment half-round pipes should be placed behind the pointing (Fig. 7/11) to form
a system suitable for draining the intercepted water.

a) Wt1terproof pointing

Half round
pipe

New mortar mt1de with Jika, Tricos,1/

or other seatin t1dmixture

b) Dry pointing

0
1st 18!/er 1.-1 cement .- s,;nd
2nrl layer 1, 2 cement , s,;nd
3rd lauer 1,J cement , Mnd
4th layer f, t l'ement .- snd

t7Jd mort,;r
Fm. 7/11. Re-pointing the joints of old, weathered stone or concrete block linings
_....- ---- -

C..:TERL)R.',TiON .!.ND REPA IR

Both interior waterproofing and pointing will require a mobile stage such as
the one used for the control survey and inspection of tunnel cross-sections. Various
types of stages are illustrated in Fig. 7/12a-c. The most difficult problems in
tunnel repair are the need for maintaining traffic during constructional operations,
the limited working space that is available and the short periods of time during

Bmt rail

; 175
I
~C/earanje
( dlilgram
\
' .....__ --- --- _ , , ,,
',. mtn 700_,.._m,n ?0O_j S!.1ge for storing buildtng
,;;3feria/s
Fm. 7/12. Scaffolding types for lining inspection and repair

which work can be carried on undisturbed by traffic. For this reason, special
consideration must be given to the structure, mobility and clearance requirements
of the stage.
72.212. Repair by controlled drainage. In this method the object is to control
the groundwater that is seeping towards the tunnel. This can be done in either
one of two ways: in the first the danger of wash-outs is avoided by intercepting
the water by collector drains and discharging it into a conduit ; in the second a
special drain gallery is constructed parallel to the tunnel. The latter method has
been known for centuries in the mining industry; generally speaking, it is more
effective but also more expensive.
Tunnel drainage by interceptor drains outside the tunnel used to be a standard
procedure in the last century; the water was piped and discharged into a drain
conduit constructed inside the tunnel.
Although stone backfill placed outside the extrados also has a number of prob-
lems (such as subsequent increases in pressures, etc.) it does intercept the ground-
water seeping towards the tunnel and drains it into longitudinal drain pipes running
at a slope along the spring Jine or along the wall base. These pipes are connected
in turn to laterals passing through the lining at a gradient of 3 %- 5 % and spaced
at about every 25 - 50 ft (Fig. 4/128).
From the point of view of statics such drainage of the groundw.:iter is most
desirable because most of the hydrostatic pressure is eliminated. More detaih
about this type of drainage can be found in Section 46.3.
- ~ -- - - - -- - -- - - - . -

SERVICE, O PER ATION AN D MAI N TE N A NCE

Tr.1nsverse adtt
above the arch Side vent1/at1ng r-
ad1t 1
Side vent11/acmg
'-,j,,!1.m.~~==....,=iimim.a.iiimi~::....=amaa.sam"""''"~-
r =~¥1!.....=~ = ~~ ~="'"'
ad,
...1t..
s ---- · ·.....,....,
- __ ~--M~
J.!J Alagu!
Str ~

Drain Collector
gallery gallery

Bent scrap m l
frame work
Dram
fPllery

7/ 13. Layo u t o f ventilation and drainage arrangements in a tunnel in B udapest

'm. 7/ 13. Layout of ventilation and dr ainage arrangements in a tunnel in Budapest

Repairs of t
more drastic drainage measures are required, the alternative to be considered creting. Th15
at in which a special drain drift is constructed outside the tunnel. An example rather than
certain struct
11pplementary drainage arrangements is illustrated in Fig. 7/ 13, showing the
ut and cross-section of the drain gallery constructed in a tunnel in Budapest.
The first pr
1ould be noted that the size of the gallery was established to allow inspection lining elemen
maintenance. ments.) In m
·om the aspect of vertical layout it is better to keep the bottom of the drain by new ones
,ry somewhat below that of the tunnel. In stratified ground the drain gallery a number of
1ld be located on the side from which the strata are falling in order to inter- cast in situ re
the water before it reaches the tunnel. Collector drains, like 'feelers', can be for cast-iron
id out from the drain gallery if required, primarily in the direction of the have to be sur
el so as to draw the water away from the tunnel area. This can be done from in concrete.
cal risers which are spaced along the gallery and from which the •feeler' Entire sect:
1s can be started in a radial pattern (Fig. 7/ 13). forcing ring~
:ain galleries may be already needed during the construction of the tunnel if be dried out "
,r inrushes of water are being encountered. In this case it is obviously advan- protect the m
>us to design such temporary galleries in such a way that they can be made tha t this can
of the final drainage system once the tunnel construction has been completed.
~ - - - ~ - - -~~~~ - -- - -- - . -

856 SERVTCE, OPERATTON AND MAJNTENANCE

Tr.1r,sverse adit
Side ver,ti!Bimg

---- 3dlt'
5 -- --- -- -

---=

DrBi11 Collector
!Jii!lery g3!!ery
- ~
Bnck

~
!ming

8er,t scrap r,1/(

framework~
Dr3m
gallery

FJG. 7/ 13. Layout of ventilation and drainage arrangements in a tunnel in Budapest

Fm. 7/13. Layout of ventilation and drainage arrangements in a tunnel in Budapest

If more drastic drainage measures are required, the alternative to be considered

is that in which a special drain drift is constructed outside the tunnel. An example
of supplementary drainage arrangements is illustrated in Fig. 7/13, showing the
layout and cross-section of the drain gallery constructed in a tunnel in Budapest.
It should be noted that the size of the gallery was established to allow inspection
and maintenance.
From the aspect of vertical layout it is better to keep the bottom of the drain
gallery somewhat below that of the tunnel. In stratified ground the drain gallery
should be located on the side from which the strata are falling in order to inter-
cept the water before it reaches the tunnel. Collector drains, like 'feelers', can be
spread out from the drain gallery if required, primarily in the direction of the
tunnel so as to draw the water away from the tunnel area. This can be done from
vertical risers which are spaced along the gallery and from which the 'feeler'
drains can be started in a radial pattern (Fig. 7/13).
Drain galleries may be already needed during the construction of the tunnel if
major inrushes of water are being encountered. In this case it is obviously advan-
tageous to design such temporary galleries in such a way that they can be made
part of the final drainage system once the tunnel construction has been completed.
 DETERIORATION AND REPAIR 857

The profile of drain galleries should be such as to allow

drainage by gravity, if possible. If this is not possible, then a
system of collector sumps and pumping stati011s must be in-
corporated in the layout (Kandevan tunnel, see Fig. 7/24).
72.213. Repair by mechanical ventilation. In simple cases
where the amount of percolation and condensation water is
small, the use of mechanical ventilation may be effective. The
drying effec~ of agitated air may be sufficient to avoid frost
damage in the lining. As an additional benefit, the smoke and
fume content of the air will be diluted and removed. This will
be hygienic and will also prevent the formation of aggressive
solutions from gases and water. Natural ventilation can also
be helpful, particularly if augmented by the circulation of the
air by traffic. In subways, for example, there is no need for
the removal of percolating water or wet spots on the lining at
all - unless there is a danger to the condition of the lining itself.

72.22. Repair of Linings

Some of the simpler problems such as weathering, scaling,

cracking, etc. have already been dealt with in connection
with maintenance and inspection. The scaling, weathering
and chipping of natural rock due to fumes and smoke in
' tunnels left with a bare rock face fall into the same category.
Repairs of these fypes are most easily performed by sandblasting and shot-
creting. This section deals with repairs designed to reach the causes of the trouble
rather than superficial repairs. Such repairs will necessarily have to involve
certain structural changes and reconstructions.
The first problem to be mentioned here is the replacement of broken or cracked
lining elements: (Stone or concrete blocks, cast-iron or reinforced-concrete seg-
ments.) In most instances it is extremely difficult to replace the defective units
by new ones - one only has to think of the radial joints - and for this reason
a number of substitutes will have to be used, such as concrete for stone blocks,
cast in situ reinforced-concrete or bent steel liner-plate with welded rib stiffeners
for cast-iron segments. While the replacement is in progress, the adjacent units
· have to be supported by steel props which may be left there eventually and encased
in concrete.
Entire sections of lining can be repaired under the protection of inside rein-
forcing rings or a second lining. This method also allows the existing lining to
be dried out and covered with continuous waterproofing which in turn will also
protect the inside wall constructed subsequently. It is not very often, however,
that this can be done without encroaching into the clearance area. To keep such
 858 SERVICE, OPERAT ION AND MAINTE NANCE

encroa chmen ts to a minim um, it is advisable to

use fairly thin second ary linings acting as slabs
between steel or reinforced-concrete ribs or
beams counte rsunk and embed ded into the
existing lining (Fig. 7/14).
It may happe n in exceptional cases that the
existing tunnel carryin g two tracks may be re-
constr ucted to carry a single track only (recon-
structi on of the Semmering tunnel , Sectio n
72.331) ·in which case there is plenty of space for
the constr uction of a heavy load-b earing lining
inside (Fig. 7/20). Provisions for waterp roofin g
betwee n the old and new lining do not presen t
any proble ms in this case.
Fm. 7/14. Layout of If the deform ations are caused· by excessive
tempor ary steel rib extern al loads rather than by defects in the lining
suppor ts fqrming itself the cracks have to be stoppe d by elimin-
part of final
reinforcement ating the movem ents that cause them. The first
thing to do is to suppo rt the lining by curved
scaffolds, made up of bent rails, T-, H- or /-sec-
tions well braced in the longitu dinal direction. Most cracks
can be traced back to
movem ents of the walls. An outwa rd t!lting of the walls about
the bottom corner
is usually an indica tion of additio nal loads on the arch
or of erosion due to
leakag e behind the wall itself. Deform ations of this type
are accom panied by
settlem ent of the crown and haircra cks nearby , enlarg ement
of the span at
spring ing height and wide longitu dinal cracks along the
spring lines. Repairs
should begin with the grouti ng of cavities and soil stabili zation
behind the spring -
ings and with stabili zation of the strata above the arch.
Should this prove in-
sufficient or if the arch shows signs of early collap se it has
to be replac ed, ring
by ring, with reinfor ced-co ncrete or cast-ir on· segments
of greate r streng th.
A paralle l adit should be driven behind the spring line at
the same time, from
which the spring section can be thorou ghly reinfor ced and
bricked up tightly
agains t the natura l ground . Care should be taken in driving
the adit that the
b)

Oe!ormeil
sh;pe

Fm. 7/15. Cracks due to wall

movements Invert constr11Clad /3ttr
 DETERIORATION AND REPAIR 859

excavation is closely followed by the backfilling operation. Openings should

be provided every 5-8 m (10--20 ft) from the tunnel to facilitate the transportation
and handling of materials (Fig. 7/ l5a).

Strut

Bucket

4dlt for haulage

Fm. 7/16. Tunnel

reconstruction
'Nhile
maintaining ll- 65m
traffic
- --- - - - ~~~---~ - ~

860 SERVICE, OPERATION AND MAINTENANCE

There may also be another possif:>ility: the walls may move and/or tilt inward
under excessive lateral pressures. Repairs should begin with the construction of
an invert and with soil stabilization behind the walls in order to reduce lateral
pressures (Fig. 7/l5b). The invert should be constructed ring by ring while main-
taining traffic - the use of structural steel encased as rigid reinforcement in con-
crete has many advantages to offer in this respect. If the movements have reached
such proportions as to endanger the stability of the walls and particularly that of
the arch, the answer lies once again in the ring by ring construction of a thin but
firm lining inside (steel or reinforced concrete, Section 72.332).
If the load-bearing capacity of the existing lining is rather doubtful - either
because of movements or wash-outs and frost action - and there is no room for
a second lining inside because of clearance considerations, then the existing lining
has to be removed and replaced, ring by ring. The new lining must be of adequate
strength and must be waterproofed.
Fig. 7/16 illustrates a case in which the entire existing lining had to be recon-
structed while maintaining traffic.7· 11 An adit for the haulage of materials was
constructed first from which a number of riser shafts were driven up above the
crown line. From here another horizontal adit was constructed to the centre
line of the tunnel to provide access to the longitudinal adit above the centre line
from which the reconstruction of the lining was to be carried out. A steel frame-
work, supporting the existing lining, was
erected prior to starting demolition and
then an opening was cut into the ground
under the adit right down to the crown-
line. The section was then fanned out
fully with the aid of lateral forepoles and
I crown bars while longitudinal beams
I ' resting on the steel framework were sup-
Securing steel porting the posts. After removing the ex-
frame
isting lining and the ground immediately
behind it, the new lining can bestbe con-
structed with concrete pumped into place.
In some instances there is no neeq
for reconstructing the entire lining and
the repairs can be confined to the arch.
It may be possible . to do this from the
inside (Fig. 7/17) by erecting a steel
F1G. 7/17. Arch reconstruction
while maintaining traffic
framework and then working toward the
outside; the overburden loads in the
meantime have to be transmitted by tem-
porary posts to the steel framework. All this can be done best from a mobile stage ;
there will, however, be lengthy interruptions caused be the movement of traffic.
7 11
· DANDUROV, M. I.: Tonne/i(Tunnels). Gos. Trans. Zheleznodor. lzdat., Moscow 1952
 DETERIORATION AN D REPAIR 86 1

7.23. REMODELLING AND RECONSTRUCTION OF TUNNELS

Because of additionltl clearance requirements of the electrification of railway
lines and the construction of additional tracks, a tunnel carrying such lines
has to be reconstructed. Although Jess frequently, a tunnel may also have to be
reconstructed because of damage caused by subsidences, breakdown and excessive
blasting.

72.31. Reconstruction for Operational Demands

In order to accommodate a second track, the tunnel section may be widened
on both sides, in which case the existing lining is not to be demolished until the
new one has been completed (Fig. 7/ 18a). This method is most practicable when
working under poor soil conditions and considerable pressures. Work can be
started from a top heading above the crown line, with excavation proceeding
from the top and then constructing the lining in an uninterrupted sequence from
the bottom (Fig. 7/ 18a) or from drifts at the bottom of the section. In the latter
case the new lining must be constructed in lifts following the excavated height of
the drifts (Fig. 7/18b). The first method is limited to firm soil conditions, the
second can be used in poor soil as well.

Fm. 7/ 18. Symmetrical

widening of a
single-track tunnel .empor,;ru support
to two tracks (steel f r,;me)
I

-~ ~ . / 4 \ \ - -+--
 862 SERVICE , OPERAT ION ANO MAINTE NANCE

It is also possible to widen the tunnel section to one side

(Fig. 7/19). This is
feasible only in ground s having consid erable streng th; the
existing arch has to
be suppo rted tempo rarily at the joint with the new lining.
In case of electrificatjon the tunnel section has to be deepen
ed, requir ing the
recons tructio n of the arch and the heightening of walls. Basica
lly the proced ure
is the same as the one illustra ted in Fig. 7/18a. The load-c arrying
capaci ty of the

Old tunnel "--

Lining to be
demo/ts/Jed

FIG. 7/ 19. Tunnel widenin g to one side (single track to two

tracks)

section i~ usually not impair ed by makin g it higher, unless

there are heavy lateral
pressures to conten d with, in which case a system of counte
rsunk rib stiffeners
may be effective withou t extensive remodelling of the walls.

72.32. Reconstruction because of Decay and External Effects

Recon structi on due to breakd owns and blastin g operat ions

is an extremely
difficult proble m as it amoun ts to tunnel constr uction throug
h a mixtur e of mate-
rials that have been loosen ed-up over extensive areas and
throug h rubble from
the collap sed lining. If the cover above the tunnel is not too
deep then it may be
well worth aband oning the tunnel altoge ther and going into
an open cut.
If the cover is consid erable , howev er, the chance s qf archin g
effects re-developing
during constr uction are much better. Using the Germa n or
Italian metho d com-
bined with careful driving and effective underp inning (steel
sections encased in
 RECONSTRUCTION 863

concrete cf. Fig. 7/ 14) it may be possible to reconstruct the tunnel even though it
has to be designed to take heavier loads. It is advisable to use pilot beadings in
order to explore the sections of the walls and invert which may have been left
undamaged so they can be reutilized in the new structure. In this respect the pro-
cedure is similar to widening a tunnel from one track to two tracks.
The reconstruction of subaqueous tunnels is even more difficult. As a matter
of fact it may be simpler as well as more economical to construct a new tunnel in
a new location rather than to attempt to reconstruct the old one. In reconstructing
a shield-drive)l tunnel the first thing to do is to build up a heavy impervious and
airtight blanket of clay from the top. Next, an airtight floor can be constructed
in the undamaged shaft. Thus, the tunnel can be cleaned and the defective seg-
ments replaced under compressed air and the entire interior can be reconstructed
in successive stages. Keeping pace with the progress of the work the area under
pressure is moved forward all the time with shifted bulkheads and airlocks
(cf. Fig. 6/174). The problems..become enormous at places where the old lining has
totally disintegrated so that the remnants have to be removed as well. This can
best be done with a shield assembled within an undamaged tunnel section (i.e.
smaller than the original shield) and using temporary lining segments. The tem-
porary lining has then to be changed into the final one, piece by piece, with an

Concrete supports
under r,11/way ties New concrete
support
FIG. 7/20. Expanding a tunnel section with segments
f

864 SERVICE, OPERATION AND MAINTENANCE

erector ~mn (i.e. without a shield, Fig. 7/20).7· 12 Because the difference between
the sizes of the temporary and final linings is rather limited, in large tunnels it
may be worth considering driving a pilot heading 2·50-3·0 m (8-10 ft) diameter
first of all, from which the disintegrated zone can be . artificially stabilized by
injections (cf. Fig. 6/149). Such stabilization may be helpfulin the construction of
the temporary and final lining as well.
Relatively speaking, it is more fortunate if the approaches of the tunnel have
been destroyed rather than. the section below the river bed as the risks of a blow-
out are less. Even in this case, however, reconstruction· from underneath is very
complicated and expensive as manifested by the example of the tunnel under the
Escaut in Antwerp. This tunnel was blown up during World War II with 24 tons
(metric) of explosives detonated in one of the ventilation shafts. Some 540 m
(1800 ft) of tunnel lining was damaged by the terrific blast, a considerable por-
tion of which consisted of cast-iron lining segments located in the shield driven
section. In spite of the more than 25 m (80 ft) cover ab0ve the tunnel the recon-
struction was carried out from the top in a huge open cut, protected by multi-
stage well points, rather than from the bottom, as mentioned earlier. 7· 13 This,
in itself, is an indication of the magnitude of the problems to be encountered in
the method of tunnel reconstruction described above.

72.33. Examples of Tunnel Remodelling

72.331. The Semmering tunnel. This tunnel, about 1470 m (4900 ft) long, was
built between 1848 and 1852. Because of unfavourable geological conditions (Chapter 2)
several problems arose during the actual construction (cf. Fig. 2/25). These problems were
solved although the lining was built to construction standards toot would be considered
unacceptable today. In spite of serious infiltration of water the lining has not been waterproof-
ed, resulting in all sorts of repairs immediately after opening the tunnel and these repairs
went on for 100 years, i.e. until the tunnel was completely reconstructed. Owing to the
defective lining, considerable pressures and continuous damage, caused by fumes and water,
the lining has deteriorated and deformed to the point where maintenance costs reached prohib-
itive proportions and the tunnel could no longer perform its function safely. About 30 %
of the tunnel lining was loaded to the limit of its bearing capacity and another 50 % was
rapidly approaching this condition. Between 1947 and 1949 about 260 m (850 ft) of scaffold-
ing had to be provided to protect trains from spalling rock and icicles. Because of plans
for electrifying the railway line the tunnel finally proved to be inadequate from the point
of view of clearances as well.
All these considerations have led to the reconstruction of the old double-track tunnel into
one with a single track . A separate tunnel was constructed for the other track_;.tt The cro~s-
section of the reconstructed tunnel is illustrated in Fig. 7/21 . Preliminary studies have indicated
that the most economical solution was to construct the new lining right up against the old

;.u JONES and CuRR'Y: Enlargement of the City and South London Railway Tunnels,
Proc. Inst. Civ. Eng. 1927 224 176
; .ia DE BROE, H.: Les tunnels sous l'Escaut a Anvers. La Technique des Travaux 1947
Mars
,.u RAINER, H.: Der Umbau des alten zweigleisigen Semmeringtunnels, Ost. Bauzeitschrift
1953 11-12
 RECONSTRU CTION 865

one even though this was to result in a surplus

of clearances that otherwise would not have
been necessary. Because of the considerab le de-
formation s of the old lining and because the
new lining was constructe d to a constant thick-
ness, the reconstruc ted section is not uniform
and has a variable width of 6·60-7·0 m (21 '-8"
to 23'-0"). The concrete was placed with the
ajd of mobile scaffoldin g (Fig. 7/22} and con-
crete pumps (5·0 -5·5 atm and 9 m (30 ft) sec
velocity). A total of 3500 m 3 (45 0OOcuyd) of con-
crete was poured while traffic was being main-
tained. To provide an imperviou s concrete the
concrete mix had to be designed very carefully
on the basis of trial mixes (half and half 0- 7
mm (0--1/4" ) and 7- 50mm (1 /4"-2" ) aggregate FIG. 7/21. Cross-section
with about 25 % hard rock quartz-sili cate con- of the reconstruc ted
tent, 370 kg/m 3 (440 lb/cuyd). Portland cement, Semmerin g tunnel
50 k'g/ m3 (80 lb/cuyd) trass and occasionally 13·5
kg/m 3 (22 lb/cuyd) plasticizer for better worka-
fing
bility). The resulting concrete was, indeed, impervious and for this reason no waterproo
old and the new lining - at least not for the time being. It
has be6n provided between the
by frost,
would have been rather expensive to adjust the old lining, deformed and damaged
fing
to afford a smooth surface for the waterproo fing, so it was decided that should waterproo
lining.
still be required at some future date it should be applied to the inside face of the new
suitable
This would be supported by another (third) lining, made of ashlar stone blocks,
along both edges of the invert slab (Fig. 7/21).
bearing surfaces for which were provided
cracks
This arrangeme nt has not proved to be altogether successful, however, as shrinkage
Drain
have developed in the new lining, allowing water to percolate in a number of places.
system.
grooves were then provided to collect and discharge the water into the drainage
artificial-
Similar provisions were made at all joints between abutting lining sections. "Opanol"

Fm. 7/22. Mobile scaffolding in the

Semmerin g tunnel

SS Szechy: The Art of Tunnelling

 -
866 SERVICE , OPERAT ION AND MAINTE NANCE

resin waterproofing sheets were provide d at these constru ction

joints right from the beginning.
The old lining has been dried out very carefully, mainly through
groutin g the holes behind
the lining and through drain grooves on the inside faces (Oberh
asli method , cf. Fig. 7/8).
Except for fluate treatme nt to protect against corrosi on by
fumes, the inside concret e face
was left unfinished. Traffic was maintai ned during constru
ction and the second, new single
track tunnel was constru cted simultaneously (1949-5 2). Excava
tion for the inverts and con-
structio n of the arch had to be done with extra care through
out these operati ons because of
the previously experienced bottom heaving.
72.332. The Bo-P'eep Tunnel. This tunnel Is 1180 m (3850
ft) Jong, near Hastings
(England), and was built in "Wadh urst clay" i.e. in alterna ting
l:iyers of sand, clay, sandsto ne
and marl. • ~ Passing mostly under built up areas the tunnel
7 1
has a shallow cover not exceeding
35 m (115 ft). The cross-section, accomm odating two tracks,
is elliptical and consists of a
brick lining about 80 cm (31") thick; in some sections there
is also an invert 60 cm (24")
thick. According to the constru ction records a great many problem
s were encoun tered during
constru ction. First, a considerable amoun t of water had to
be coped with and saturate d clay
kept squeezing in at the roof. This was attribut ed to water
finding its way from the upper
layers of sand down along the initial shafts_into the clay at
the bottom . The amowr t of water
varied in proport ion to the amoun t of precipitation on the surface
. The tunnel itself is located
in hard marl which was very difficult to excavate but, exposed
to air, weathered very rapidly.
The toughness of this material is also borne out by the fact
that the waUs do not extend
to the same constan t depth: in very hard material they were
stoped higher up than in soft
and weathe red material where they were carried all the way
down, and even supplem ented
with an invert.
Problem s started to arise in 1927 when clay intrudin g from
under the walls started to
heave the subdrai n, caused extensive silting and created soft
pockets in the ballast which had
to~ removed ever more frequently. In 1932 the first wall movem
ent of 1/ 10 mm (1/ 32" ) was
observe d; this movement came to a halt fairly soon. A number
of cracks appeare d in 1947
and reopene d soon after having been repaire d. The cracks
were particu larly pronou nced
in the recesses. Throug h systematic inspection and recordi
ng it soon became obvious that
while both tracks were rising, the walls were sinking and started
leaning towards the inside.
The deform ations have continu ed to focrease at an ever acceler
ating rate. Geological studies
have traced the centre of the movem ents to a fault. The saturati
on and remoulding of the
material under the walls was aggrava ted by continu ous vibratio
ns caused by traffic. At the
time reconst ruction was started the leaning toward s the
inside has reached 15-18 cm
(6-7 in) with 13 cm (5 in) measured level with the rails.
An investigation of the soil
conditi ons has revealed a shear strength of 0·5~-0-86 kg/cm•
(7·5- 12 lb/in 2) below the walls
and 0·12-0·22 kg/cm• (1·7- 3·0 lb/in 2) under the tracks. [A
check on the geostatic pressur e
has indicate d that the latter was just enough to overcome the
shear strengt h of the soil behind
the wall. Althou gh the angle of repose has varied widely
(</> = 10°-50°) this had but little
effect on the vertical and horizon tal rock-pr essures ; p = 3·9-
4·2 kg/cm• (55-60 lb/in')]. After
deducing the shear strength the resultan t vertical pressure
was about 3·5 kg/cm 2 (50 lb/in 2 ).
In the design of the new lining the lateral pressure was assume
d at 50 % of this value.
The point was reached in 1949 where reconst ruction could not
be postpon ed any longer and
indeed this was begun with a section 140 m (460 ft) long, adjoinin
g the west portal where the
movements and cracks had been observed (Fig. 7/ 24). An invert,
consisting of structu ral steel
encased in concrete, was installed first. Because of rapidly increasi
ng deform ations there was
no time left for doing this work and maintai ning traffic at
the same time and so for this
period the section had to be closed and passengers transferred.
The old lining had then to be
removed and replace d by cast-iro n segments and the space
behind it was thoroug hly grouted .
In additio n to adequa te strengt h and instant load-be aring
capacity this arrange ment also
7•15 CAMPIO N, F
. E.: Part Recons truction of Bo-Peep Tunnel at St. Leonar ds
on Sea, Journ.
Inst. of Civ. Eng. 1951
 RECONSTRUCTION 867

New cast-iron
.sepments

L,m,t of fort-
poles
J5~m
V!tt'lfted brick
Steel pl,;te svbdra,ii
New invert

FIG. 7/23. Tunnel reinforcement by cast-iron segments and by an invert consisting of a steel
frame encased in concrete (Bo-Peep tunnel)

cm
0 18

2-5

S·O 1.7
.,,.
.....,
..,<::: 7-5
..,t,:..
10·5
:l
13 5·0

15'5 2·5 t----f--7.;,f==ff--\+----+--'-'----t"<-t----t----t--'-'<>-

18(cm) oc,=......____.__........._c.=..c...c.......cL.c:..::c,__.__,__'--__,__,__- " = = -

~ ~ ~ ~ ~ ~ ~~~" t)-;, :st- ~
""" :>,; """ ~ ~ ~ ~ ~~ ~ ~~ ~
FIG. 7/24. Diminishing wall movements as a function of the progress of invert construction

afforded additional clearance. With the invert completed, the wall movements have tapered
off rapidly and so has the heaving of the tracks (Fig. 7/24). No temporary supports were provid-
ed while construction of the invert was in progress, partly because regardless of the type
of bracing there was hardly any chance of preventing the settlement and leaning over of the
walls at all, and also because such bracing would have required extensive and sizeable mud-
sills that would have seriously interfered w:'(h the construction of the invert. Speed of construc-
tion was the most i~portant factor as could be concluded from the study of the available
measurements and records.

55*
868 SERVICE, OPERATION AND MAINTENANCE

73.333. The Kandevari highway tunnel. This tunnel which is situated between Teheran
and the Caspian Sea, 1884 m (6200 ft) long, 5·0 m (16'-6" ) wide and about 2300 m (7500
ft) above sea level, was built in the 1930s. The pass itself is about 2700 m (8800 ft) above
sea level and the bare mountain, with no vegetation to speak of, is made up mainly of warped
layers of marl interwoven with limestone .and silty gravel conglomerates. Due to heavy
autumn and winter precipitation great amounts of water penetrated the tunnel through
the extensive cracks in the marl. Since the tunnel had only an inadequate drainage system
and no waterproofing at all, the joints in the lining were leached out, leaving voids behind
and resulting in additional overburden pressures and heavy frost damage to the limestone
lining. Frost has also caused spalling and, thus, has represented a constant traffic hazard.
Even though it slopes towards both, ends from the centre the profile of the tunnel is not very
good (Fig. 7/25); the grade towards the north portal is only 0·2 %, providing inadequate
draw off through the central drain and causing steady silting, at the same time.

Schematic profile of tile

+2700m l(andevan Tunnel

French drains

-south
To Teheran
Jnl'low ol' water
( ' 'i
so·;..
v
. Inflow of water
, ",,
'lateral drains
I ; 2%,
North
lo, lasi,ian Jea

1884m

Cross section

New f)3vement
lateral g;llery New chm
gallery
.
I
-/--- ---- -
--- -----::-
Old pavement centre drain --- - -
12·50

FIG. 7/25. Repairs to the Kandevan Tunnel

The repairs were concentrated on improving the drainage system and filling the voids
behind the lining. This was achieved through a drainage drift 750 m (2450 ft) long, 2·0 m
(5'-6'") deep and 2·20 m (7'-3" ) wide, with a grade of0·8 %, 12·5 m (41 ft) off the tunnel centre
line and on the side intercepting the -streaming water. Starting 4 ft below the bottom of
the tunnel the gallery ended 5·50 m (18 ft) below the surface. French drains above the extrados
filled with stone and gravel were provided in locations beset by large amounts of water and
were extended to intercept, as far as possible, the water before it reached the tunnel. The
central drain was reconstructed and relocated at the sides, the grade was increased and
interrupted by steps every 150 m (500 ft). (At the entrance the grade was reversed because
of the frost danger.) At each step the water was discharged into the side drift through lateral
drains. The road itself was repaved (concrete pavement) and the cross-section changed from
a sag to a crown. Finally, the holes behind the lining were filled with concrete (pumped and
 SURFACE SUBSIDENCE 869

piped in place) for the major ones and trass cement grout for the minor ones. "Sika" and
"Darex" plasticizers were added to the second grouting; in the first grouting sawdust was
added to the trass cement. Even though the frost-damaged segments have also been replaced,
the elaborate reconstruction cannot be a final solution to the problem without perfect main-
tenance and subsequent extension of the drainage system, as required from time to time.
To construct the tunnel at such a high altitude was a fun\lamental error in the first place ;
it would have been much better to build a longer tunnel at a much lower altitude in more
solid ground and causing less height losses for communica\ion.
72.334. The Pereces-Baross shaft tunnel. This is known as the Griinzenstein adit, and
was built in the 1870-s (cf. Fig. 2/29). It was later expanded to accommodate haulage trains run-
ning on a narrow-gauge track. During this operation the supports were replaced several times
until the derailment of a train in 1947, at which time a number of steel supports were hit
involving the breakdown of the section. After this the tunnel was lined with concrete blocks.
In 1954 cracking was observed in the top of the concrete block lining. Because the cracks
were considerably wider on the intrados than on the extrados, they must have been caused
by movements at the spring line or by excessive bending moments due to heavy vertical loads.
It was found later that the vertical loads were not abnormal and that the reason must have
been the voids behind the spring line and the consequent yield of the springing. In the cracked
sections the entire lining was removed and replaced and the holes were grouted properly.
Because of uneven settlement of the drain gallery, caused by the numerous repairs, the drainage
system was also in need of reconstruction. In certain sections the water was pooled rather
than drained and so the low spots had to be transformed into sumps from where the water
was pumped into a gravity drainage system.

7.3. SURFACE SUBSIDENCES RESULTING FROM TUNNEL

CONSTRUCTION

7.31. ESTIMATION OF SURFACE SUBSIDENCES

Surface settlement and subsidence have presonted serious problems throughout

the centuries of the history of mining. The prevention, reduction, as well as the
prediction of the range of anticipated settlement are all the more -important as in
most cases the mines are located under populated and built up areas. As regards
tunnel construction, however, the problem is relatively new and is tied up with
the construction of subways, pedestrian and highway underpasses and utility
tunnels. In earlier days it made hardly any difference as to what happened to the
surface where railway, highway or navigational tunnels were passing under
unpopulated areas.
Tunnel construction causes discontinuity in the ground masses resulting in
settlements, loosening, and in a stress increase within a certain area. Both factors
are bound to cause deformations, which, depending on the strength of the soil,
on the depth of cover, and on the rate of disturbance caused by the operation,
may extend to the surface and affect it to a larger or -lesser extent. There is no *
doubt that, for a given set of soil conditions and depth, surface settlements and
the extent of loosening will depend largely on the quality of workmanship, on the
care exercised in construction, on the stability, strength and orompt installation
- - ~- -
- -- -

870 SERVICE, OPERATION AND MAINTENANCE

of temporary and final supports and on keeping the number of turnouts to a

minimum. It follows, then, that future settlements cannot be predicted but, at best,
estimated only, and that such computations can only serve to underline the role
and importance of the contributing factors more than anything else. An estimate
of the subsidences should be based on actual observations, and on: this basis the
computation may follow either theoretical or empirical lines.

73.11. Estimating Surface Subsidence on a Theoretical Basis

All theories dealing with surface subsidence start out from the size of the surface
depression i.e. of the discontinuity, left in the wake of excavation.
73.111. Empirical formulae. A number of empirical mining formulae have
been developed over the last 100 years, relating the amount of surface subsidence
to the size of the excavation and to the depth of cover.
One of these was devised by BRIGGS:

2-2
S=m- - - - -
2-2+ J_!!__
100
where m = the· depth of the drift
H = the depth of cover.

There is another empirical formula by GOLDREICH:

s = m - KH, (7.5)

where IC = the coefficient of expansion of the volume of the ground and is cal-
culated from
v.- V,
K= - - -
V

where V. - the volume of earth excavated underground

V, - the volume of the ground within the cone·-of subsidence
V = the volume of the total displacement zone.

As a minimum the value of K may be taken as IC= 0·01; from this minimum it
will increase as the depth of the heading increases; for m = 4·0 m (13 ft) it is
about K = 0·03.
GOLDREICH has also investigated the dynamics of the movements.
Owing to the excavation of the drift shown in Fig. 7/26 blocks A, B1 and B 2
will be set in motion, the effect of which will extend to the surface (abed). The
underground excavation will give rise to movements on the surface, both hori-
zontal and vertical as illustrated. While, according to GoLDREICH, the maximum
 SURFACE SUBSIDENCE

a b C (J

\ I /

' .\ 1
\
A

t
i }//
I I
\ I
I I
' I I ,

~ ~
- Tens1011 - Verti cal move
~ ~ents
~ ~
+ Compressi on
FrG. 7/26. Movements due to
underground excavation Hor1zonwl movemenLs
(GOLDREICH)
~ ~

vurface

£xcavat1on

.Surface
FIG. 7/27. Various phases of
building subsidence in /I
relation to the pr,)gress Excavation
of excavation

will occur at b and c, the minimum will be found at a and d. Once block A has
moved downwards, wedges B 1 and B 2 will follow; this is why there will be hori-
zontal movements as well as vertical ones in areas ab and cd* and this is how
th~ depressed area will develop on the surface. There will be a lapse in time between
the underground excavation and the subsidence on the surface; the faster the exca-
vation underneath, the faster the buildings on the surface will have to go through
the various stages of settlement and accompanying stressing (Fig. 7/27). In areas
with extensive mines underneath there is more than one slip surface along the
boundary of the subsidence zone and these are marked by gaping cracks, some-
times several dm (1 ft) wide (Fig. 7/28).
Based on investigations and observations made by the Mining Research Insti-
tute of the Soviet Union, S. G. AVERSJN 7· 16 has studied the shape and extent of the

7·16 AVERSJN, S. G.: Gornie raboti pod sooruzhe11ia11i i vodoemami (Mining works under sur-
face and groundwater), Ugletehizdat, Moscow 1954
 872 SERVICE, OPERATION AND MA!NTENANCE

FIG. 7/ 28 Wall of a
building
in Karvin,
cracked through
and ttlted due
lo undermining

iH
Fie. 7/29. Theoretical
treatment of
i
-... I
surface
settlements by
AVERSIN

depression area on the surface. (Fig. 7/ 29): His equation for the depression line is

11 (x) = 110 ( 1 - ~) ( ~ )2 e ~ x • (7.6)

It has been observed that L = 2·13 I and with this the equation is
X
4·53 2'13 -
1
11(x)=110 ( 1- _~ e · (7.7)
2 31 )
where / = the-distance between the point of counterflexure and the maximum
depression.
 SURFACE SUBSIDEN CE
8'73
The approxi mate value of the maximu m depression in the centre, 11 , is calculat
0 ed
from
m - 110
K=---
H ,
i.e.
11 0
=m - KH,

The maximu m depression, 170 can also be obtained approxi mately from
the
area ( V) between the depression line and the original surface as follows:

In his quoted work AVERSIN also presents some more accurate and complex for-
mulae, taking into conside ration the effect of time and the progress of consolid
a-
tion. Because all the above formulae conside r homogeneoµs soil, which is
hardly
ever encount ered, the computa tions are of rather academic value.
In the Hungari an Mining Researc h Institute , F. MARTOS has also inves-
tigated the problem of surface subsidence. 7 · 16 a From a statistic al evaluati
on of
actual observa tions he arrived at

where 11(x) = the vertical compon ent of surface subsidence

H the depth of cover
= a constan t for a given set of local conditio ns
the efficiency of backfilling
= the distance of the point of counterflexure

I = ± H cotan fJ J _l _
In ~
Vs:
,

where {} = the angle measured from the horizon ta l to the line connect ing the
edge of the undergr ound excavation with the boundar y of the surface
subsidence
v.. = the mean error of levelling .

;.J6a Cf. MARTOS, F.: Banydszati Kut. Int. Kozl. 1960 1-2
. - - ~ - -

874 SERVICE, OPERATION AND MAINTENANCE

1'he maximum settlement is

Km
r,0 = - -{1 - 81) ,
K + H

and the horizontal component of the movement is

~ (x) =' xr, (x) .

H
With reference to the studies and settlement records made in connection with
the construction of the Leningrad Subway, LIMANov 7· 17 has developed a method
to calculate surface settlements in the case of shield-driven tunnels in Cambrian
clay. Going back to the Maxwell theorem L tMANOV considers surface settlements
as deformations in a semi-infinite elastic space caused by pressures exerted by the
tunnel lining on the surrounding ground - except for the sign which has to be
reversed. For, if the pressure from the surrounding ground causes stresses and
deformations in the tunnel lining then surely the problem can be reversed and
considered from the other end as well. Assuming a set of reversed stresses to be
caused in the lining by a hypothetical pressure inside the tunnel, the lining would
also be bound to exert opposite pressures on the surrounding ground which in
turn would result in deformations on the surface that would be equal to the
settlements but would be opposite in direction.
Two approximations are made in this theory: the surrounding ground is assumed
to be perfectly elastic and the fictitious inner pressure, i.e. the cause of the ring
deformations, is assumed to be uniformly distributed and radial in direction.

flugrtiary top layers

'1r

Cla!I h,

FIG. 7/30. Surface movements due to pressure inside the tunnel

7 ·17 LIMANOV, I. A. : O_
sadki zadnoi poverhnosti pri sooruzhenii tonnelei v kembriskih glinah
(Surface settlements in Cambrian Clay due to tunnel construction). Inst. Inzh. Zhelezn.
Transport. Leningrad 1957
 SUIUACE SOBSIOENC I
875
Depending on the degree ro which these as-
sumptions are correct, the results will have to
be treated as only an estimate of the range of
settlements.
Consider the circular underground ,?pening at
a depth h0 and subject to a uniformly distribut-
ed pressure pas illustrate d]trFig. 7/30, with the
surface deformations plotted upside down. While
the settlements can be descril:,ecl by their vertical
components, they will haw horizonta l compo-
nents as well.
The movement of the po1nt inside the semi-infinite
body caused by undergrou nd excavation , is divided by
JEFFERY1o1 8 into two parts, the one due to the inside
pressure p ( W2) and that due to the external loads on
the surface ( W3). The problem is investigate d in a bipo- lx
lar coordinate system,7 • 19 and in this case the radius Fto. 7/ 31. Determina tion of
of the circle round one of the poles is zero (Fig. 7/ 3 J ). movement s due to
The resulting deformati on vector W is the vectorial pressure inside the
sum of the above two movement s: W = W + W • tunnel in a bipolar
2 3
Let U and V denote the componen ts of W parallel to coordinate system
axes x and y, respectivel y, then:

U = - (1 + µ) -Ep · {[ 2(1 -
r0 µ) -h,, - (I - 2µ) ] -
cosfr
- -
1
[ 2(1 - µ) -h" +
a r1 a

+ j
(I - 2µ) - - -
-
COS{}.
+ -X r( h- a
) COS 2
• f}I
- (h 0 + x) COS 2{}. ]}
• - , (7.8'
'2 a 'i '2

V = - (I+µ)~ ri {[2(1 - µ) - (I ·-· 211) ~ ] si:

1
°
1
+ (2(1 - µ) +

+ (J - h0 ]
sin {}.
2µ)- -
a
- -
r~
+ -x
a
[
(h,,- a) -sin-2&
'i
1 sin 2iJ 2
. - - (h.+a) - -.-
r;
l} ,
where, in addition to the notations used in Fig. 7/ 31

the radius vectors ar~

a =Jh~ - rJ,

r 1 = J<x - a)'+ y2; r2 = J<x+ a)2 + y2,

the polar angles are 1J 1 , 1J 2 , calculated from tan {) 1 y y

= - - - and from tan fJ- 2 = ---
x - a x+a

7 18
· COKER, F. and FAILON, L.: Fesziiltsegmeresek sztatikai a/apoh {Calculati on of stresses
on the basis of statics), ONTI 1936
7 19
• JEFFERY, G. B.: Plane Stress and Plain Strain in Bipolar Coordinat es, Transactio
ns of
the Royal Society London. Series A 1920 221 265-293
~
-- ~ -- - _ --

876 SERVICE, OPERATION AND MAINTENANCE

LIMANoV, using the above equations for homogeneous material, arrives at a ma)limum
vertical movement in the line of the vertical diameter of
_
2
p 4r5/t0 (7.9)
Umax - (1 - µ) E -h, 2.
o - 'o
This, however, can also be calculated from the loosened area (F) that has not been back-
.filled around the tunnel
F
Um.1x=-,
an
where
p
F = (I - µ•) - 4r02 h 0
E
+-
J
dy
a2 + y•
• = (l - µz)
p 4ri h0 :t
E a
(7.10)

However, the opening of the undergi:ound hole diminishes the initial strength of the soii
and will increase its compressibility at ~he same time. As a result a multiplying factor has to
be introduced in the the calculation of F; this factor is

K = 2(1 - µ) h 0 (7.11)
a + 2 (l - µ) (b - a) ·

Knowing the amount of maximum

L
surface settlement the shape of the
7 subsidence area can be determined on
F the basis of the probability curve of
L• Za +II,. t:mf/5"-t/z) AvERSIN. Thus, if half of the diameter
of the cone is 2a, then the settlement
at a distance x from the centre is
4x
4
X -
'1
= U
max (
1- -
2a )
e2a
· (7.12)

From observations made in

the construction of the Leningrad
multitube subway stations LIMA-
FIG. 7/32. Computation of NOV developed a number of em-
settlements in stratified soils (two layers)
pirical formulae relating to the
settlements resulting from the
driving of adjacent tunnels and also considering the lapse of time between the
two operations. The anticipated settlement in the two-layered soil in Leningrad
was also investigated , as shown in Fig. 7/32.
The maximum ordinate of the depression curve of 2L diameter can be calcu-
lated by equating the areas above the depression curves of the two layers

F
11o= -
L
where F = Umax arr ,
 SURFACE SUBSIDENCE 877
an,:!.

L = 2a + h1 tan (45° - ±-J,

2;
(7.13}
also

Um&x -_ (l - µ 2).!!_
E
4r5h0
2 2 = UA + UB ,
ho - 'o

73.112. Estimation according to the theory of probability. An interesting method

by LITVINISZYN 7 · 20 is based on the theory of mathemati cal probability . Studies of
minirrg-
areas have led him to consider the ground above
the hole to be of a fractured and granular rather
than of a continuou s nature. Assuming that the
fractured elements are all of the same size, the
empty space created by the removal of block l (Fig.
7/33) could be filled by either of the two blocks di-
rectly above it. The probability of the left ·block
or the right block sliding down is I /2. These blocks,
however, would have to be replaced once again by
one of the blocks above them, the probabilit y of the
three blocks sliding down being 1/4, 2/4, and 1/4
respectively. Proceedin g in this manner each block in
each layer can be assigned a 'chance fact1,1r', result-
Fm. 7/33. Probabilit y of settlement
ing in the probability curve (K) according to Gauss
for elementary blocks
(Fig. 7/33). From the above considerat ions LITVI-
above an opening
NISZYN derives a mathemat ical operator by which
any depression surface can be computed from the
one above it. The expression is a function of time (.;), horizontal distance (x) and maximum
depression in the centre (w )
0

- (e, x) =
,w J~
ivo J s
e- ' ds'
- ½xo"
which is in good agreement with actual measurem ents. For further details the reader
is referred
to the references.

73.12. Estimating Subsidence from Actual Observations

The developm ent of surface subsiden ce as a function of the various construc tion
stages has been recorded carefully in connectio n with the construc tton of the
Chicago Subway. 7•21 Passing through plastic clay the tunnel has a horsesho e

7 0
-~ LITVINISZYN, J.: Displacem ents in Loess Bodies as Stochastic
Processes, Bull. de I' Acade-
mie Polonaise des Sciences 1955 III 4
LITVINISZYN, J. and NEUBERG, K.: Gebirgsbewegungen Uber einem Abbau als stochastisc
her
Prozess aufgefasst, Freiberger Forschungshefte 1956 H. C. 22
7 21
• TERZAGHI, K.: Liner Plate Tunnels on the Chicago Subway, Proc. Am. Inst. Civ. Eng
.
1942 Jan.
-- - - -- - - - -
---
- - - - - - - - --

878 SERVICE, OPERATI ON AND MAlN'TEN ANCE

average cover
cross-section 6·0 m (20 ft) wide, 7·50 m (25 ft) deep and has an
with steel liner-pl ates (cf. Section 62.43
of 8-9 m (27-30 ft). The tunnel was built
tion, illustra ted in Figs 7/34a-b , shows
and Fig. 6/65). The sequence of excava
measur ed subside nces plotted at a numbe r
a road above the tunnel in plan with the
tunnel is shown above the plan (c), also indicat ing
of station s. The profile of the
l'he settlem ent and constru ction progres s lines,
the progress of constru ctioh.
indicate
relating to the same date and numbe red for easier identification, clearly
of construction
that the development of surface subsidence followed the progress

c) SFttlemff/t st.;tio11s
a)
G 764+,W G 164,20
V
~~
~~ "'
"-
~~
..,..., --- '
<._<-..
~ Prortle
"'~ "' "'
"'"" "<,~

• 01rect11J11
I
1Jf'~ f11Jr
Exav-1ted between Apr 77, 9PM ~, _L
a/Id /,pr ?8, 4 A"I
JOO rt ch.1m;ges

b)

,J
_l_, _

F1G. 7/34. Some su,rface settleme nts at the Chicago Subwa)'

a maximum of
in a regular manner. Along the centre line subsidences reached
the tunnel width (b) to cover (H) ratio
15 cm (6 in), i.e. about 20 % of
(s = 0·2 !),which is considerable. The locatio ns with the fastest developing sub-
the sections
sidences were the same as those with the highest rate of excavation;
nce rates. (Note the differ-
behind and ahead indicate rapidly diminishing subside
es in the settlem ents have been
ences between lines I and 6.) S_u dden increas
ting the starter crown plates on the
observed whenever the underpinnings suppor
allow the excava tion to be deepen ed.
steel ribs were removed and replaced to
ed: wall plate drifts (2) and (3) were
To avoid this, the method was later improv
m (6-16 ft)
driven on each side (Fig. 7/ 34a) at the spring line, kept 2·0-5·0
in which longitu dinal steel beams were placed
ahead of the general excavation
ting the crown section . After excava ting area No. 4
to hold the steel ribs suppor
the two beam, to enable the rib stiffene rs to stand up
struts were placed between
 SURFACE SUBSIDENCE 879

under the lateral pressures while working on the side drifts. With these improve-
ments the surface settlements could be reduced by a half. On the same project
the surface settlements doubled, however, in about 12-15 days after the pneumatic
pressures were cut off.
Once again the depression diagrams closely follow the Gauss probability curve,
which would indicate that the shape of the depression area is the same for shallow
I as well as for deep covers. No subsidences were observed on either side beyond
a distance equal to 1· 5 times the depth of the invert, i.e. the width of the depression
• zone amounted to
3(H + J?'!) + b
l where H = the cover
m = depth of tunnel section
I b = width of tunnel section.

7.32. MEASURING THE SUBSIDENC E OF BUILDINGS

DUE TO SUBWAY CONSTRUCT ION

In metropolitan areas subsidences due to subway construction may adversely

affect the stability and safety of buildings. The primary object of regula r observa-
tions is not so much to discover the absolute magnitude of subsidences as the
effect of various construction phases on the subsidence of buildings, the develop-
ment of such subsidences as a function of time, the type of consolidatio n, the
reasonably accurate ·predicti.on of final subsidences by extrapolatio n, and the
discovery of danger signats sp that the necessary measures may be taken in time
to prevent disaster. All this requires regular checking by levelling on all buildings
within a distance equal to twice the depth of the tunnel on either side. The zone
under observation in Budapest is 160-200 m (530-660 ft) wide.
l A number of benchmarks have to be established near the subsidence zone to
facilitate the levelling of well defined points on the buildings to be observed. Such
benchmarks should be checked regularly by precise levelling tied in to other
benchmarks well outside the subsidence zone. The instrument used in the precise
levelling should read to the ne1rest O· I mm and the mean levelling error (per km)
shall not exceed

• where L = the distance covered by levelling, in km.

Reference points on the buildings can be checked from the nearest benchmark
with a level capable of reading to the nearest mm. Such reference points should
be spread out evenly on the buildings subject to subsiden~e and marked with an
iron plug or spike and the observations should commence before construction
starts, i.e. in the undisturbed state. In this way subsidences can be related to static
880 SERVICE, OPERATION AND MAINTENANC E

conditions . The movements should be recorded and processed in two ways.

A plan must be prepared for each building showing a sketch of the layout of the
lmilding and the reference points together with their identification numb~rs.
On the reverse side of this plan the movements of the reference points are plotted
diagramma tically; each point should have its number clearly indicated together
with the range of levelling and with the amount of movement. Elevations should
always refer to sea level (geodetic).
A general picture of the subsidences can best be obtained by preparing a so
i;alled isoanabase map, similar to a map with contour lines. The map, covering
the entire area under observatio n
and to a scale of 1 : 500 (l" = 40
ft), has to indicate the buildiT'lgs
with the reference points and
with the centre line of the
subway.
The subsidence of each refer-
ence point, computed from the
previous reading, is marked next
l?elerence to the appropriat e point. On the
.OOill/.f
basis of the maximum and mini-
mum subsidence a suitable inter-
Fm. 7/35. Isoanabase Jines showing settlement val can be selected for the 'con-
conditions tour' lines, after which these
lines can be plotted by connecting
all points indicating the same
amount of subsidence. The lines will be valid only around the buildings, how-
ever, as the areas not built on may settle in a different manner (Fig. 7/ 35).
This map, indicating the movement s relating to a certain period of time, is use-
ful for two reasons: it will give an indication of the movements of the buildings
not included in the observatio ns and it will indicate the location of maximum
subsidence. If there is no reference point in this location, one should be estab-
lished immediate ly and should be observed from then on. A number of con-
clusions can be drawn by comparing two maps of successive periods and plotting
special graphs for the points indicating the maximum movement s: it can be
established which are the points that have slowed down or even stopped and
which are those with increasing subsidences, calling for a closer spacing of refer-
ence points for future observations or, if necessary, calling for appropria te con-
struction and safety measures.
Regarding the frequency of measurements there is a distinction between areas
under constructi on and those that have been completed. From past experience
those in the first category should be checked every 1 or 2 days, the other need to
be checked only every 2 to 3 months.
Benchmarks located within the subsidence zone must he checked at least every
6 months.
 ----
\
'

- '·-::-
-- -- ~ - -, - ~- -: --
I Shield chamber II.
____ __,I ~ :_ ___ __:; L -~ 5 ,,-: - -- -~
-- -- -- -~ - -- -- -
- - - - ------ - -

-3
~7 -

dap est Sub wa y

stru ctio n site s of the Bu
e sur fac e sett lem en t lines at one of the con
Fie . 7 36. Iso ana bas
1

•
 ----- -

_ _.:i.. _ _ _ _ _ _

~hield chamber II. ------------

~ --------
----------

_// /
/
/"

I Settlement (tsoanabase) lines at 1mm intervals

apcst SJbway

• CT
 SURFACE SUBSJOENCE 881

Fig. 7/ 36 illustrates an isoanabase map of one of the construction sites of the

Budapest Subway. Note that the subsidences centre around the excavated under-
ground areas. It is also worth noting that the subsidences were considerably
larger over the shield chambers constructed by classical methods and requiring
larger areas to be excavated than those over shield-driven line sections.
This is even more striking in the case of one of the deep,-lying stations of the
Leningrad Metro (Subway), of which the isoanabafse lines are illustrated in Fig.
7/37. The soils consisted of Cambrian clay underlying recent alluvial layers.
Figs 7/37a-b illustrate the conditions after the completion of the shield chambers

Arter complet1011 of the 0

shield cMmbers

b)
0 0

After l'Ompletion of the entire

sl.,;tion

Fm. 7/ 37. Development of surface settlements

------=----=--~ - - - - --

882 SERVICE. OPERATION AND MAINTENANCE

0
20 3 7
40 Jettltmtnt of' 1vrl.,u
~ 60 /Jtncllm1rks
la
.. BO
100
120

Orirt o'rlVttTf1 frc,,v.,tion of Construction of wills anti invtrt Canstrvct,on of

L-------=----'l,uding (top) conntct,-,,; chm-
.section btr (tnnsformtrs)
HJ
(}8 Mun r,1tt of' stttltment o'tvtlopmtnl
of svrlict 61:ncllm1rks
(}4

0-Z
o 1--,,,,e=:;;;;=::+=:..,,,,.~.,,,.....!.,.,.J..,,:;L~L_;:=c::::~__;~~.,.fil--=:::r:~~~

FIG. 7/38. Development of surface settlements during construction of twin shield chambers

and of the entire station, respectively. It can be noted that a trough shaped depres-
sion surface, about 13 cm (5 in) deep and uniform in the longitudinal direction,
has developed as a result of driving the tube type stations 3 x 8·5 m (10 x 28 ft)
in diameter and about 200 m (650 ft) long, and this subsidence was in addition
to the cone-shaped subsidences up to 15 cm (6 in) deep, which were the result
of the mining operations required for the construction of the shield chambers.
Surface subsidences and their rate of development as a function of time and
construction phases are illustrated in Fig. 7/ 38 for the construction of twin-shield
chambers, connecting corridor and ventilation adit. It can be noted readily that
the maximum settlements have developed following the excavation and lining of
the top section of the wide (/ ~ 12 m [40 ft]) shield chambers.

REFERENCES

ANON. (1952): Eisenbahntechnische Rundschau . Fachheft Tunnelbau, 12 Carl Rohrig Verlag.

Cologne and Darmstadt
(A VERSIN ) A eepC}IH, Cr . . (195 4)·. Tonnue
y
pa6omb1 noo coopy:J1Celll/fl.1tU u B oooe.11a,1111 yrne-

TeXH3AaT, MocKBa . . . . .
· Movement of Articulated Bu1ldings on Subsidence Sites, J o u111.
H EATHCOTE, F . W . L . ( 1965) · ,
inst. Civ. Eng. Febr. _ _
LuETKENS O ( 1957): Bauen im Bergbaugebiet. Springer, Berlin
STREIT, J.'
an.d SEFCIK, K _ (1959): Tune/y (Tunnelling). Dopnivni Nak_lad. Prague
WILKE, (I 959): Tunneluntersuchungswagen der DBB, _fisenbahntech111sche Rundschau 20 5
ZANOSKAR, W . (1959): Rekonstruktion von Tunneln. Ost. Ing. Z. Dec.
---.-=- - - --

GENERAL LITERATURE 883

GENERAL UTERATURE
AussENDORF, K. (1955): Tunnelbau. VEB Verlag Technik, Berlin
BARDOUT, G. and BERNY, P. (I 954): La construction des tunnels, galeries et sou terrains.
Eyrolles Editeur, Paris
CABELKA, J. and MENCL, J. (1960): Hydrotechnicke Stolne (Hydrotechnical tunnels). S lovenske
Vydatelstvo Literatury, Bratislava
Compres rendus du /Veme Congres International de I' Urbanisme Souterrain et des Techniques
des TravatlX Souterrains. Varsovie 2-5 Juin 1965
(DANDURov) J(anoypoB, M.H. (1952): To1111e1111 (Tunnels). roe. Tpattcrr. :lKene3HO.!IOPOl!CHoe
l13,11., MocKBa
DESIO, A. (1949) Geo/ogica Applicata a/1'/ngegneria. V. Heepti, Milan
DUHM, K. (1947): Tunnel und Stollenbau. Georg Fromme und Co., Wien
HAMMOND, R. (1959): Tunnel Engineering. Heywood, London
KIESER, A. (1961): Drucksto/lenbau. Springer, Vienna
KOGLER und HARMANN (1932): Bergbau unter Tage, Tunnelbau, Handbuch fiir Eisenbetonbau.
Ernst, Berlin XII
KRYNINE, J. and JUDD, P. (1957): Engineering Geology. McGraw H ill, New York
LUCAS, G . (1926): Der Tunnel. Ernst, Berlin
(MAKOVSKY) Ma,;06cK1111, 8.1. (1947): To1111e.111 (Tunnels). J,1311. Ai<a.11. Apx1neKTypuii CCCP
MocKea
MENCL, V. and MENCL, J. (1962): Staine a Tunely (Tunnels). Statni Naklad. Techn. Lit. Prague
MEREGALLJ, G. (1957): La situazione a1tuale de/le metropolitane nel momlo. Transporti pubblici
PEQUIGNOT, C. A. (1963.): Tunnels and Tunneling. Hutchinson, London
RANDZIO, A. (1927): Der Stolle11ba11. Ernst, Berlin
RICHARDSON, H. W. and MAYO, R. S. (1941): Practical T111111el Driving. McGraw Hill, New
York
RZJHA (1874): Lehrbuch der gesamten Tunnelbaukunst. Berlin
SEDLACEK, J. (1954): Staine a Tunely. (Tunnels) Statni Naklad Techn. Lit. Prague
SINGSTAD, 0 . and BAILLIEIN, D. (1956): Tunnels. (Part 10. in Abbett's Amer. Civ. Eng. Prac.
Vol. I) Wiley, New York
STINT, J. (1960): Tunnelbaugeologie. Springer, Vienna
STREIT, J. and SEFCIK, K. (1959): Tunely (Tunnelling). Dopravni Naklad. Techn. Lit. Prague
SztcHY, K. (1961): Alagutepitesta11 (Tunnel construction), Tank6nyvkiad6 Budapest
TAL'.)BRE, J. (1958): La mecaniques des roches. Dunod, Paris
TEODORESCU, P. (I 958): Construc1ia Tune/or (Tunnel construction). Ed. Cailor Ferate
Bucharest
TEODORESCU, P. (I 961): Calculul Tune/or (Tunnel dimensioning). Editura Transporturilor si
Telecomunicatiilor, Bucharest
Valogatott fejezetek az alag1i11ipites es banyaszati melyepites korebol (Selected chapters
from tunnelling and mining construction). Miiszaki Kiad6, Budapest IS52.
(VOLKOY) Bo111,06, B II. (I 9'15): To1me.111 (Tunnels). roe. Tpattcn. :lKene3H0.!I0PO)l(HOe l13t\.
MocKsa
 INDEX

A bend (caisson disease) 831

bent 618
accelerating slope 116 - (double) 620, 623, 625
access opening 805 bentonite (grouting) 732
acce5s shaft 784 - (lubrication)· 823·
acoustic strain gauge 258 BIERBAUMER (rock pressure theory) 191
acrolein 509 blasting (short delay) 576
adhesion coefficient 111 - technique 571
adit (location of) 107 blow-out 773
adjustment groups 542 BODROV-GORELIK (design method) 365
admission slot 697, 704 boiling danger 261
advancement of shield 718 BONNARD (lateral pressure) 222
"Aerocem" method 853 boring set (mole} 712
agressive constituents 83 bottom friction (consideration of) 303
air resistance 113 pressure 223
- seepage 782 - stability 227
airtight cover (artificial) 778 - swelling 224
alignment of tunnels 106 BOUGAYEV A (ring design method) 396
alternate ring method 638 box section design 459, 462
American (polygonal) timbering 655 braking slope 116
anchorage drift 702 BRANDAU 141
ANDREAE, C. 91 , 503, 514 break-up 629
anticline 67 breakdown of roof 233
ARBUZOV, G. V. 518 breast board 619
arch (circular) 296 bridge action period 231, 597
- computation 292 bulkhead (shield) 689, 696, 704
arching 138 - wall 704, 772
"Armco" culverts 454
Austrian (cross-bar) method 635, 610
AVERSIN, G. 871 C

B caisson-disease (bend) 831

- floating methods 675
- sinking methods 673
back space 688 Californian siding 716
bacteria (sulphate reducing) 494 Canadian cut 578
BALLA (rock pressure determination) 206 CAQUOT-KERISEL (rock pressure) 169
base tunnel 22, 53 carbon monoxide 85, 509
bearing (surveying) 527
CARLSSON (pressure cell) 258
Belgian (flying arch) method 639 cast iron lining segments 739
benchmark 537 cathodic protection 496
 ~--
- - ----- - - - - - - - - - - - - -- -
-
- --

886 I NDEX

caulking of segments 744, 760 D

causes of tunnel detcriorl!ti on 843
cavern (Karstic) water 80 damages in tunnels ;;46
centering 542, 546 DAvmov (bott0m pressure) 226
- control instrumen t of T ARCZY 557 (design method. for circular sections) 405
central drift method 584 - (desi gn method for horseshoe sections)
centre-out method 638 316
check hole 535 deep station (five-bay type) 802, 809
chemical consolidat ion 776 - - (for subways) 802
circular arch (tabulated data) 300 - - (three-bay type) 801, 808
circular hole (stresses around) I 52, 171, deformatio n process (around a cavity) 145
183 delay firing 576
circular tunnel (computat ion of) 341 design (analytical ) 291, 343
- - (elastically embedded ) 365 by members (graphic) 282, 341
classical (m ining) tunneling methods loads 267
609 - for deep tunnels 267
clearance 123, 130, 132 - for shallow tunnels 269
closing error 536, 538 of horseshoe shaped tunnels 281
Cologne method 665 of circular tunnels 341
combined tunneling methods 654 of rectangula r tunnels 459
composite action (tunnel + ground) 316 of tunnel sections 264, 341
composite design (of culverts) 459 specificati ons (Budapest Subway) 272
compresse d air (applicatio n of) 771 - (Lisbon Subway) 270
- - demand 781 - (Soviet standard) 277
- - (supportin g effect of) 780 deteriorati on of tunnel s 843
concrete lining (cast in place) 753 determina te system 292, 325, 346
- - segments 747 developme nt of design procedure s 265
conduits (circular) 442 dip 64
confined cavity (shape effect) 126, 187, dislocation zone 58
200 division platforms (laminated face) 714
connecting adit 796 Don-Seg lining segment 751
contour blasting 578 double lining (design of) 412, 42:?
core leaving (German) method 648 drainage of tunnels 490
corrosion (by groundwa ter) 498 - system repair 849
- protection 493 drift (see heading) 583, 613, 616
- (by soil) 493 drilling (mechanic al) 569
corrosivity of soils 495 drilling platform 582
cost figures 48 dropping survey lines down shafts 548
crescent loading 351 dnrln-digg er type shield 712
cross-bar (Austrian) methDd 635, 610 dry boring (exploratio n) 61
cross sectional shape 117
crown bar (beam) 589, 630
- (English) method 630 E
- moment variation 340, 403
- subsidence 610 economic studies 47
culvert 423, 8 I 9 elastic centre (a) 293, 324, 334, 344, 349
constructi on 425 elastic rod supports (fictitious) 317, 379
- (longitudi nal design) 438 elastic subgrade 302, 321 , 365, 381, 396, 404,
- of corrugated steel ("Armco" ) 454 415, 440, 458, 463
cut and cover 668 , 824 elliptical hole (stresses around) 185
cutting edge (of shield) 68 L ENGESSER (rock pressure theory) 21 7
(mechanic al) 569, 577 English (crown bar) method 630
- segment (''Hallinge r") 708 entrance lock 557
 INDEX 887

equivalent model material 242, 255 (for water exclusion) 608

erection-chamber (shield) 783 materials 726, 730
erector arm 719 pressures 728
escalator tunnel 810 gunite (plastering) 485
EszT6 (rock pressure theory) 196 - rendering support 606
excavation (in shield) 700
- (mechank?.1) 705
excess air pressure 774 H
exploration boring 61
- headings (drifts) 63 "Hallinger" -shield 706
- shaft 62 hang (slope) tunnels 25, 52
explosive charges 579 HAROSY T. 210, 434, 436
- consumption 580 heading and bench method 583
explosives 573 headings (liner-plate) 624
extensometer arrangement 244 (masonry) 627
(pilot) 616
(scope and construction of) 615
F (timber) 617
health protection 829
face stabilization 701 heat (generated in subways) 518
failure surface (Mieses) 78 HEIM J. 70
- types in rocks 75 HEWETT-J0HANNESSON (design method) 356
fan 512 highway tunnels (ventillation of) 506
faults 56, 59 - (vehicular) tunnels 27, 42, 121
FENNER (rock pressure) 171 hinged-culvert 437
five-bay type stations 801 , 809 hingeless arch 292
flying arch (Belgian) method 639 hole depth (blasting) 574
folds 55, 56 hoop stress 418
FoRCHHEIMER (moment equation) 347 horizontal air lock 557, 772
- (rock pressure) 211 horizontal layout 525
forepoling 619, 632 horseshoe shaped section (analytic design)
full-face tunnelling (with supports) 584 291
- - - (without supports) 581 - - - (graphic design) 282
HousKA J. (lateral pressure) 222
hump profile (underground stations)
G 116
hydrological survey 79
GALERKIN, B. G. (design theory) 412
Gants- Hill type station 802
garages 33 I
gases 84
genuine mountain-pressure 148 icicles 845
geoisotherm (computation of) 91 Icos-VEDER (Milanese) method 669,
geological investigation 63 816
- profile 94 illumination intensity 519
- survey and exploration 52 inclined stratification 161 , 165
geothermal gradient 87, 93 inspection of axis 840
- step 88 (of lining) 840
German (core-leaving) method 648 - (of shape) 835
grade (ruling) 113 - (of tracks) 838
grout machines 723, 735 isoanabase contours 880
grouting 719 isostasy 55
- (effect on pressure distribution) 254 Italian (invert) method 652
 888 JNDEX

J - (prestressed) 768
- (reinforced concrete) 755
jack (face) 694, 718 LIMANOV I. A. 874
- (ram) for advance 690 Lisbon Subway (design loads) 270
HKY (theoretical slope) 305 live loads 262
joints (for jacked pipes) 882 load-carrying ring 599
jumbo 582 loads on shafts 786
- on tunnels 133
local effect (pit corrosion) 447
K local failure (HOUSKA) 222
location of tunnels 106
KARAFIATH (loads on shaft wall) 788 longitudinal elevation (gradients) 110
karstic (cavern) water 80 loosening coefficient (KOMMERELL) 209
KASTNER H. 148, 151 - pressure 136
KAPOSZTAS P. 502
KERISEL J. 143, 169
knife-edge support 364 M
KOMMERELL (graphic design) 282
- (rock pressure theory) 209
MAIHAK (pressure-cell) 257
KOVACSHAZY F. 420
MAILLART (rock pressure theory) 194
KUNZ-method 647, 664
maintenance service 836, 842
marking 532
MARQUARDT (conduit design) 442
L
MARTOS F. 873
mechanical excavators (in shield) 705
lagging 590 mechanical ventilation 505
- (skeleton) 618, 623 MEISSNER (design method) 410
laminated face (division platforms) 714 methane 85
laminated lining 418 Milanese (Icos-VwER) method 672
latent-plastic state 148 MINDLIN R. D. 142 .
lateral ground support 317, 321, 326, 331 mining methods 609, 630
lateral pressure coefficient (influence on model tests (for pressure determination) 255
stress distribution) 190 MOHR F. 127,149,266
- - determination 219 mole (boring set) 712
layout (horizontal) 525, 538 Moll-ribs 594
(of details) 533 moments of monolithic section (variation of)
(of long tunnels) 526 340
(of segments) 561 monolithic section (design of) 321
(of short tunnels) 524 monument 532
(of underground tunnels) 538 Moscow type of shield 710
(underground directional) 556 - - station (three-bay) 806
- · (vertical) 537, 559 mucking 715
Leningrad type of shield 709 multiple drift method 651, 663
- type station 808 multiple stage mining methods 639
LEON and WILHELM 141, 146, 236 MOHLHOFER (design theory) 4i5
lighting 519
liner-plate 623
- - method 657, 660, 661
- - with stiffening 658 N
lining segments 738
(concrete) 747 navigation tunnels 30, 44, 120
(of cast iron) 739 needle beam method 657
- - (of structural steel) 746 noise control 519

.......
 INDEX 889

numerical example (BODROV-GORELIK) 370 tunnels 31

(egg-shaped culvert) 445 prestressed lining segments 768
(horseshoe design) 308 primary (orogen) stresses 134, 157
(polygonal method) 388 propulsion (of pipe) 820
(timbering) 621 - (of shield) 690, 718
(two-bay box section) 473 protective (stress-free) zone 140
(ZURABOV-BOUGAYEVA) 326 PR0T0DYAKON0V (pillar load) 239
PROT0DYAroNOV (rock pressure) 213
PRUSKA J. (culvert design) 430
0 pseudosolid rock 136, 149
public utility tunnels 32, 120, 811
Oberhasli drain-system 852 PVC caulking 759
offspur tunnel 25, 52 - waterproofing 487
optical plumb 551
orientation angles 528
- methods 553 R
ORLOV (computation method) 411, 76S
orogen (primary) stress 134, 157 RABCEWICZ L. 135, 161, 602, 609
overbreak (effect on pressure) 138, 140, 233 railway tunnels 26, 40, 118
overthrust 57 raking support 633
ram (jack) for propulsion 690, 718, 820
RANK! (pressure cell) 258
p reconstruction of tunnels 858, 861
rectangular section (design of) 459
parking areas 33 reference net 539
Park-Lane underpass system 816, 828 - point 535
PASTERNAK (box section on elastic foun- reinforced concrete lining segments 755
dation) 462 relative displacement of members (con-
pedestrian tunnels (subways, upderpasses) sideration of) 305
29, 119, 121, 812 remodelling (examples) 864
pentagonal (polygonal) timbering 655 RENDULIC L. 183
"Phillips" (pressure cell) 256 repair of linings 857
pillars (loads on) 239 - of tunnels 848
pin marker 546 RICHARDSON and MAYO 698, 714, 716,
pipe jacking 819 823
pit 629 RICHART L. 126, 183
- corrosion (local effect) 497 RICHTER R. 601
plastering 483 rider (KUNZ) 665
plastic zones (extension of) J.82 rigidity (influence of) 340, 404
plumb lines 544 ring (design of monolithic) 343
- - (eccentric-in shafts) 547 RITTER W. (rock pressure) 211
- (optical) 554 rock (crushing) strength 570
poling board 617 deformation diagrams 71, 73
polygonal method (ring design) 578 pressure (development process) 230
- (with roof separation) 387 measurement 241
portals 520 superposition 234
potential energy (BODROV-GORELIK) 367 theories 159, 191
precast panels 824 - (considering overburden) 191
prefabricated elements (the use of) 669, 817, - (neglecting overburden) 208
825 (types and causes) 133
pressure cells 249, 256, 258 properties 70, 76
distribution ring (shield) 694 (roof) bolting 597
- reestablishment 244 temperature 87
890 INDEX

roof bolts 600 ~tray-currents 496

- shield 683 strength factor (PROTODYAKO NOV) 216
- stabilization (by silicatization) 778 stress conditions (around cavities) 189, 23:.
ruling grade 113 isobars 175
R :l!HA 135 measurements in the rock interior 245
- on the rock face 243
- on the tunnel supports 246
s - relaxation (BERTHIER) 245
- releasing bore hole 244
- stress releasing cut 243
SACCARDO (ventilation system) 510
stress-relieved zone (development of) 145
saddle (watershed) tunnel 22
stress restoration 245
safety measures 828 strike 64
- recesses 521
subgrade reaction (influence of) 340, 404
SANGER (ventilation system) 5 I 0
subsurface stations 801
sawtooth profile (station) 116
subway (Metro) tunnel 27, 44
SCHMIED J . (elasticity theory) 167
sulphate-reduc ing bacteria 494
secondary stresses (in rock) 134
superposition of roof loads 234
section design by member 28 L
support (reinforced concrete) 597
- reaming 578
(temporary) 587
segmental timbering 635
- (timber) 596
semi-graphica l design (culverts) 444
- (yielding) 591
settlement zone 543
SuQuET (rock pressure theory) 208
shaft arrangement (for pipe-jacking) 821
surface loads 263
- - (for shields) 783
subsidence 869
shape (check of) 836 (actual observations) 877
- effect (confined cavity) 126, 200
- (due to subway construction) 879
shield advancement 718 - (theoretical estimate) 870
chambers 797
surveying 524
- (dimensions and structure) 681
survey line (through air Jocks) 557
- position (determination of) 564
swelling pressure 156
- tunnelling 678
r,y ncline 67
short delay blasting 576
SZECHY (computation of sinking resistance)
shotcrete (inner ring) 414, 483
793
shothole pattern 575
- (computation of upbreak height) J 62
side drift method 662
Sz1LVAGY1 I. 169
side wall construction 643
sigma point (elastic centre,) 293, 324, 334, 344
S1K6 A. (culvert design) 440
silicosis 830 T
site-explor.a tion 62
skeleton lagging 618 tail section (of shield) 69S
slope (hang) tunnel 25, 52 target 534
soil corrosion 493 TARCZY·HORN OCH A. 546, 558
sound propagation 246 television sets (exploration by) 61
specifications for the design 270, 272, 277 temporary steel support 587
spherical cavity (stresses around) 187 TERZAGH( K. 126, J 35
spile 592 TERZAGH! (bottom pressure) 227
spiral tunnel 24 (inclined stratification) J 65
steel plate (lining) 415 (loads on shaft wall) 788
STEINFELD (loads on shaft wall) 790 (rock pressure theory) 202
STINI, J. 63 (st ress conditions) 183
stratification (effect on location) 65 (tabu lated pressure values) 160
- (effect on rock pressure) 165 (upbreak height) J40, 162
 INDEX 891
theoretical slope (HKY) 205 V
thermoplastic sheets 489
- (waterproofing) 487
VARGA (design method) 408
three-bay type station (Moscow) 806
vehicular (higway) tunnels 27, 42, 121
T. H. ribs 593
ventilation 499
thrust line 282
ducts 795
thyxotropic fluids (appl. in sinking) 792
timbering 595 (during construction) 499
(lateral) 510, 513
track arrangements 716
(longitudinal) 511
transfer of directions 544
(mechanical) 500, 505
traverse 529, 544
(natural) 503
triangular loading 348
(of highway tunnels) 506
triangulation network (local) 530
(of subways) 51f
- (national) 526
shaft 518, 784
- - (special) 529
tower 512
- (self-contained) 532
vertical layout 517
trompeter (stress-free) zone 140
Vienna (underpass systems) 814
trunk (of shield) 688
VOELLMY A. 433
truss-panel 591
TSIMBARYEVITCH (bottom pressure) 225
- (pillar load) 240
tunnel (construction history of) 36 w
deterioration and repair 843
drainage 490 wall beam 587, 644
storage 33 - design (computation) 301
subaqueous (underwater) 42, 53, 102, 122 washer (haunched) 760
surveying 524 water-exclusion (by grouting) 608
walls (linings) 480 - pressure 259
tunnelling in firm ground 569 waterproofing 481, 851
- in loose grounds 668 (bonded) 486
tunnels (classification of) 21 - (of covering layers) 775
(construction and design of) 568 - (PVC) 487
(public utility) 32, 120 watershed (pass-, saddle-) tunnel 22, 53
(subway) 119 weathered zone 55, 69
(traffic) 22, 119 WILLMANN, R. 143
(transportation) 30 wobbling (of shield) 686
working chamber 673, 773
- platform 697
u
unbalanced air pressure 774
y
underground railways (comparative data) 109
station-profile 116
- stations 799 YAROSHENKO V. 424
- structures (classification of) i 9 yielding springing support 306
underpass system (Budapest) 8 I 7
- - (Park-Lane) 816
- - (Vienna) 814 z
upbreak after BIERBAUMER 162
after SZECHY 163 ZURABOV-BOUGAYEVA (design method of)
- after TERZAGHI 137, 140, 162 321

- --------------------
 - - -. --- ~~~
----------- -
~-

Responsible for publication: Gy. Bernal

Director of the Publishing H ouse of the Hungarian Academy of Sciences and of the Academ) Press

Jackel and binding oesigned by T. Piros

Responsible editor: E. K. Kallay-T6tb . Technical editor: L. Vidosa

AK 380 g 6776

70.69274 Akademiai Nyomda, Budapest - F elelus ,·ezelo: Bernat G yorgy