World Tunnel Congress 2008 - Underground Facilities for Better Environment and Safety - India

The effects of soil conditioning on the operation of earth pressure balance

machines
F.X. Borghi
Kaelin Toscano Ltd, Zurich, Switzerland

R.J. Mair
Cambridge University Engineering Department, UK

SYNOPSIS: This paper is concerned with the effects of soil conditioning on the control and on the
performance of earth pressure balance (EPB) machines. Field data collected during the construction of the
Channel Tunnel Rail Link (CTRL) in London are presented. The difficulties arising with the selection of soil
conditioning treatments are highlighted, and the quantities of foam and polymer used in different ground
conditions are discussed. In addition, the effects of soil conditioning on the control of the head chamber
pressure are analysed. It is intended that the data summarised herein will provide guidance for projects in
similar ground conditions.

INTRODUCTION

1.1 State-of-the-art and research needs

Earth pressure balance (EPB) machines are
increasingly being selected for the excavation of
tunnels in soft ground. Advances in technology over
the last decade now enables EPB machines to
excavate large tunnels with diameters exceeding
12 m in a wide range of challenging ground and
hydrological conditions. The viability of EPB
machines to maintain appropriate control of ground
movements during excavation has often been
demonstrated. However, serious difficulties with the
operations of these machines still arise, sometimes
leading to excessive ground settlements [1], large
sinkholes [2], or even loss of face stability [3]. These
difficulties almost invariably result from poor
control of the supporting pressure applied to the
tunnel face, which, in EPB machines, is
substantially provided by the confinement of the
excavated ground into the head chamber of the
machine (Fig. 1). The pressure is then controlled by
balancing the mass flow rate of spoil entering the
head chamber and that being extruded from that
chamber by the screw conveyor.

Because natural ground (unless soft in-situ)

seldom possesses properties allowing it to be easily
remoulded and confined in the head chamber to the
desired pressure level (typically in the range of 50300 KPa), the viability of EPB machines critically
depends on soil conditioning to modify the
properties of the excavated ground. This usually
occurs by injecting foams, polymer solutions, or a
combination of these products into the head
chamber in order to remould the excavated soil into
a homogeneous mixture of relatively low shear
strength and low permeability. Should the spoil
properties be inadequate, the EPB machines may
fail to operate properly or only do so under unsafe
or much impaired conditions. In spite of the crucial
role of soil conditioning, guidelines remain scarce
and the common sense and experience approach
has tended to be preferred in the field when
selecting soil conditioning treatments. This attitude
is perfectly reasonable for less demanding projects
or in cases where only a moderate standard of
performance is required. However, poor control of
an EPB machine in the vicinity of sensitive
structures or when substantial ground water
pressures have to be withstood, may initiate a
sequence of events leading to the serious
consequences discussed above.

1469

Figure 1. Diagrammatic section through an EPB machine (courtesy Herrenknecht Ltd)

1.2 Scope of the paper

2.2 Geology and hydrogeology

This paper discusses the effects of varying soil

conditioning parameters on the behaviour of EPB
machines. The field data collated during
construction of the CTRL in London are discussed
to highlight, in particular, the effect of the injected
conditioner quantities on the control of the head
chamber pressure, which is one of the factors
governing the ground movements induced by tunnel
excavation. It is intended that these data offer
guidance for a first estimate of foam and polymer
requirements in ground conditions similar to that of
the CTRL.

The typical geological sequence encountered along

the tunnel route is summarised in Table 1 and
shown in the longitudinal section of Fig. 3. Upon
launching from Stratford Box, excavation began in
mixed face conditions of the Upnor and Thanet
Formation (see Table 1) before continuing for about
560 m in full face Thanet Sand. For the next
1550 m, chalk was encountered in the tunnel invert,
sometimes forming up to 30% of the face. Mixed
face conditions then continued, with Thanet Sand in
the bottom half of the face overlain by the basal
layer of the Lambeth Group. This layer consists of
the Upnor Formation, which is 3 to 12 m thick and
contains continuous horizons of gravel or pebble
beds. About 870 m were then tunnelled in the
Lambeth Group strata (formerly known as the
Woolwich and Reading Beds). These strata are
often believed to be problematic for tunnelling due
their extensive lateral and vertical variability. This
formation comprises very stiff (undrained shear
strength, su = 300 KPa) intermediate to high
plasticity clay as well as frequent lenses of water
bearing silts and dense sands. As the tunnel
ascended towards the west portal, the thin Harwich
Formation overlying the Lambeth Group
intermittently appeared at the tunnel horizon, giving
rise to rapidly changing and difficult face
conditions. The remainder of the excavation took
place in full face London Clay, which was high to
very high plasticity and firm to very stiff with
undrained shear strength increasing linearly from
about 40 KPa at the top of the stratum to 250 KPa
or more at its base.

2.

CHANNEL TUNNEL RAIL LINK:

CONTRACT 220

2.1 Project overview

The data presented hereafter were obtained from
CTRLs Contract 220 [4], which comprised 7.5 km
twin-bore tunnels excavated between late 2002 and
early 2004 from Stratford Box to St Pancras station
in Central London (Fig. 2). The excavation diameter
was 8.15 m and the tunnel depth to axis increased
from about 8 m at Stratford Box to approximately
40 m, before reducing to 5 m as the tunnel ascended
towards the west portal. The structures overlying
and surrounding the tunnels comprised 5 km of
retaining walls, 28 bridges, as well as the North
London Flood Relief Sewer and the four railway
tunnels of the Victoria Line and the Great Northern
& City Line. The closest of these existing tunnels is
about 10 m, and the flood relief sewer only 2 m,
above the crown of the new CTRL tunnels.

1470

Figure 2. Channel Tunnel Rail Link and Contract 220, London

The ground water table in the lower aquifer

(Thanet Sand) was lowered by deep well pumping
prior to tunnelling in order to reduce the water
pressures and ease the construction of the cross
passages.
2.3 EPB machines
The sensitivity of the structures overlying the tunnel
demanded rigorous control of the ground
movements and hence dictated the use of closedface tunnel boring machines. The wide range of
ground conditions called for the selection of EPB

machines, and two Kawasaki machines were

commissioned for Contract 220. Both machines had
an outer diameter of 8.15 m and were equipped with
a 15 m long central shaft screw conveyor of 1.2 m
internal diameter. The cutting wheel of the machine,
illustrated in Fig. 4, had an opening ratio of 57% of
the excavated area and no shutters were provided to
close the openings between the cutting arms, so that
support to the face had to be continuously provided
and rigorously maintained by the pressurised spoil
in the head chamber.

Table 1. Simplified stratigraphy along the tunnel route [5]

Formation

Depth to tunnel Length of the tunnel drive

axis

Description

London Clay
(LC)

12-25 m

840 m

Stiff and very stiff to fissured clay of high to

very high plasticity.

Harwich
Formation (HF)

25-35 m

700 m in mixed face

condition with LC and LG

Fine sand, fissured clay / shelly clay with

some limestone / fine and medium flint
gravel with clay, sandy clay or sand matrix.

870 m + 1600 m in mixed

face condition with LC and
HF or with UP

Interlaminated fine sand, silt and clay / stiff

and very stiff clay / clayey sand (Woolwich
and Reading Beds).

Lambeth Group
(LG)

32-35 m

Thanet Sand (TS)

25-35 m

560 m + 3450 m in mixed

face condition with UP or
UC

Very dense, structureless fine and medium

silty sand with permeability between 5x10-8
and 1x10-5 m/s.

Upper Chalk
(UC)

35-42 m

1550 m in mixed face

condition with TS

Weak to moderately weak chalk.

1471

Medium sand / fine and medium flint gravel

in a clay, sandy clay or sand matrix (Upnor
Formation).

West Portal
Groundwater table in upper aquifer

40

Stratford Box

E levatio n (m )

30

Made Ground &

Terrace Gravel

20
10

London Clay

Woolwich & Reading Beds

-10

Upnor Formation
Thanet Sand

-20
-30
Groundwater table in lower aquifer

-40
2000

3000

4000

5000

Chalk

7000

6000

8000

9000

Chainage (m)
Figure 3. Longitudinal geological section of contract 220 (after Woods, 2003)

Figure 4. Cutter head of CTRLs 220 EPB machines (by courtesy of Rail Link Engineering)

Seven injection nozzles for soil conditioning

were distributed along the radius of the cutting
wheel and two additional ports were located in the
screw conveyor. Each machine was instrumented
with about 130 instruments including flow and
pressure measuring devices monitoring the foam
and the polymer injection ratio (FIR and PIR as
defined in section 3.1). The total pressure in the
excavation chamber was measured at five locations
on the pressure bulkhead (2040 mm behind the
face) at vertical distances of between -600 mm
below and 2000 mm above the tunnel axis. The data
was logged at a rate of 0.2 Hz for the duration of the
tunnel construction.

3.

SOIL CONDITIONING TREATMENTS

3.1 Generalities and definitions

Although the type and quantities of soil
conditioning are critical to the machine operation as
well as to the transport and disposal of the spoil,
guidelines for their selection remain scarce. The
quantities of soil conditioning agents used in the
different ground conditions of Contract 220 are
summarised here to provide additional guidance for
future projects in similar ground conditions. The
foaming agents and polymer type, as well as their
concentration were varied as the excavation
progressed (details on chemical composition of the

1472

conditioning agents are not discussed herein). In

this paper, the acronyms FIR and PIR refer to the
foam and the polymer injection ratio, respectively,
i.e. the injected volume of foam and polymer
solution expressed as a percentage of the excavated
soil volume. The liquid injection ratio, LIR, refers
to the total amount of soil conditioning liquid
(foaming agent and polymer solution). Values of
FIR, PIR and LIR were calculated for each ringexcavation (1.50 m) of both the tunnel drives;
average values and their standard deviation
measured in each of the main types of ground
conditions are summarised in Fig. 5.
3.2 Soil conditioning in the Thanet Sand
In the Thanet Sand, the average FIR and PIR were
51% and 7.5% (LIR = 10%, Fig. 5), respectively.
This FIR falls within the range recommended by the
EFNARC [6] recommendations for sandy soils.
While no evidence exists to treat the measured
values as optimum quantities, the safe and
efficient excavation of over 1,000 m in full face
Thanet Sand indicates that they may at least be
taken as robust guidance figures, especially in view
of the lack of previously published guidance for
suitable rates of polymer injection in sand.
3.3 Soil conditioning in the Lambeth Group
In the Lambeth Group, an average FIR of 47% and
an average PIR of 11% (LIR = 16%, Fig. 5) were
recorded. The FIR showed a relatively large
standard deviation, perhaps illustrating the difficulty

in determining appropriate conditioning treatments

in this heterogeneous soil stratum. In many
instances, FIRs in excess of 200% were used in the
Lambeth Group, but no direct benefits of such large
quantities of foams could be observed. On the
contrary, detailed analysis of the data consistently
showed that the increased volume of foam injected
did not result in a smoother control of the head
chamber pressure, as could perhaps have been
expected. Instead, the amplitude of chamber
pressure fluctuations increased with increasing FIR
(see section 4.2). Surprisingly, the cutter head
torque did not reduce with the thrust despite
increasing FIR and LIR values. The screw torque
showed no consistent trend with the FIR, and did
not appear to be reduced as the LIR was doubled.
While excessive quantities of soil conditioning
agents (in terms of FIR and LIR) did not appear to
be associated with significant benefits in terms of
performance (advance rate) or control (fluctuation
of the face pressure) of the machine, the absence of
conditioning impaired satisfactory operation of the
machine. This was demonstrated for a short drive of
about 300 m in the Lambeth Group, where no
conditioning could be used, presumably as a result
of mechanical damage of the rotary joints of the
cutter head. For this portion of the drive, the
chamber pressure was never higher than 10 KPa,
indicating that the EPB machine was in fact driven
in open-mode, and that, in the absence of soil
conditioning, the chamber pressure could simply
not be maintained at the required levels.

Figure 5. Injection rate of conditioning agent in different ground conditions [5]

1473

Repeated measurements of the moisture

content of spoil samples recovered at the rear of the
screw conveyor indicated that the clay-conditioner
mixture did not mix into a homogeneous paste, and
that some significant proportion of free liquid
remained in the head chamber. Free liquid gushing
out of the screw conveyor was also frequently
observed, suggesting that the injected quantity of
liquid was higher than optimal. Heterogeneous
spoil may compromise smooth control of head
chamber pressure, as the flow rate of spoil through
the screw conveyor is very sensitive to the
properties of the extruded material.

of fluid and compressed air blows. This

heterogeneity is believed to be the result of foam
breakdown following sorption of the foaming liquid
into the clay, a process also observed in the
laboratory [10]. This may eradicate the benefits of
foam and the large quantities of compressed air
resulting from the foam breakdown in the head
chamber may be counterproductive and even give a
false sense of security by raising gas, rather than
earth pressure, in the head chamber.

3.4 Soil conditioning in the London Clay

4.1 Overview

London Clay is a more homogenous material and its

engineering properties, although variable, may be
more easily described than that of the highly
heterogeneous and variable Lambeth Group
Formation. The average in-situ moisture content of
the London Clay was measured at about 24%, the
plasticity index at 47% and the plastic limit at 23%.
Therefore the consistency index, Ic, of the natural
London Clay can be assumed to have been about
unity, i.e. much higher than the ideal value for EPB
machine operation (Ic = 0.4 [7]). In addition, the
undrained shear strength of the London Clay at
tunnel axis depth ranged from 150 to 250 kPa or
more, hence one order of magnitude higher than the
strength deemed appropriate for EPB machine
operation in closed-mode. Therefore, unless treated
with suitable conditioning, London Clay is
inappropriate for closed-mode EPB operations.
The volumes of foam used in the London Clay
were on average half that injected in the Thanet
Sand and in the Lambeth Group. The average PIR
was 13% for the first tunnel, but was reduced to 6%
in the second. The FIR used in the London Clay,
26% on average, fell below the range of 30-80%
recommended by EFNARC, and contrasted with the
need for large quantities of 100% and beyond
sometimes reported in the literature [8, 9]. Analyses
of machine data suggested that the principal effect
of the foam in clay was mostly that of its liquid
phase, and that the mechanisms which make foam a
suitable additive in sand, cannot be expected to be
effective in clayey material. Observation of the
conditioned London Clay at the outlet of the screw
conveyor revealed poor mixing when large
quantities of foam were used: intermittent discharge
of large and stiff clay lumps alternated with gushing

The behaviour of the machine in relation to soil

conditioning treatment is discussed next. The
simultaneous alteration of several machine
parameters (e.g. shield thrust, conveyor rotation
speed, location and quantity of conditioner
injection) renders the identification of clear
correlations between the soil conditioning
treatments and the machine performance
notoriously difficult. Insightful patterns of the
chamber pressure evolution during excavation and
ring-build cycles may nevertheless be recognised
and two examples are discussed hereafter to
illustrate what is considered to be poor and ideal
control of chamber pressure.

4.

CONTROL OF THE HEAD CHAMBER

PRESSURE

4.2 Machine behaviour in the London Clay

Similar patterns of chamber pressure were observed
in the London Clay when large volumes of foam
were injected into the head chamber. However, the
FIR used in the London Clay was usually much
lower than in the Lambeth Group and resulted in
better control of the chamber pressure. Over 500 m
were excavated in full face London Clay with
chamber pressure often above 200 kPa. Fig. 7
shows an example of excellent control of the
chamber pressure when tunnelling in London Clay,
contrasting with that of Fig. 6 in the case of the
Lambeth Group: in Fig. 7 only minor pressure
fluctuations may be discerned and, importantly,
little or no reduction of chamber pressure occurred
during the ring-build. In such conditions, the shield
thrust and the screw rotation speed, not shown in
this paper, could also be held constant, indicating
that steady state and controllable operation of the
machine had been achieved. Interestingly, and

1474

perhaps unexpectedly, the FIRs used during

construction of the rings in Fig. 7 were extremely
low, with values of the order of 5% (at foam
expansion ratios close to unity). Therefore,

negligible quantities of compressed air were

injected, suggesting that polymer alone is, under
these circumstances, sufficient to promote very
good control of the chamber pressure.

Figure 6. Difficult head chamber pressure control in the Lambeth Group formation [5]

Figure 7. Ideal head chamber pressure control in London Clay [5]

1475

The sensitivity of the screw conveyor

operation to the material properties often
demonstrated that the excavated soil should best be
conditioned as early as possible in order to
maximise the mixing time and hence improve the
homogeneity of the material in the head chamber.
Conditioners should be injected directly from the
cutting head. They should only be injected into the
screw conveyor as a last resort measure to reduce
the driving screw conveyor torque.
5.

ACKNOWLEDGEMENTS
The authors would like to acknowledge the Joint
Venture Nishimatsu/Skanska/Cementation, the
Engineering and Physical Science Research
Council, the British Pipe Jacking Association, the
British Tunnelling Society for their financial
support and their collaboration during the research.
REFERENCES
1.

Boone, S.J.; McGaghran, S; Bouwer, G & Leinala, T

(2002). Monitoring the performance of earth
pressure balance tunnelling in Toronto. Proc. North
American Tunnelling: 293-302.

2.

Shirlaw, J.N.; Ong, JCW; Rosser, HB; Tan, CG;

Osborne, NH & Heslop, PE (2003). Local
Settlements and Sinkholes due to EPB Tunnelling.
Proc. ICE, Geotechnical Engineering, 156(4): 193211.

3.

Gaj, F.; Guglielmetti, V; Grasso, P & Giacomin, G

(2003). Experience on Porto: EPB follow-up.
Tunnels & Tunnelling International, 35(12): 15-18.

4.

Woods, E. (2003). C220Stratford to London

Portal. Tunnels & Tunnelling International, special
supplement CTRL Section 2 Tunnelling, 35(9): 810.

5.

Borghi, F.X. (2006). Soil conditioning for pipe

jacking and tunnelling. PhD Dissertation,
Cambridge University, UK.

6.

EFNARC (2005). Specification and guidelines for

the use of specialist products for soft ground
tunnelling. European Federation for Specialist
Construction Chemicals and Concrete Systems,
Surry,
UK.
Available
at
website:
http://www.efnarc.org/publications.html

7.

Maidl, U. (1995). Erweiterung der Einsatzbereiche

der Erddruckschilde durch Bodenkonditionierung
mit Schaum. PhD Thesis, Ruhr Universitt, Bochum,
Germany. In German.

8.

Janscecz, S; Krause, R. & Langmaack, L (1999).

Advantages of soil conditioning in shield
tunnelling: experiences of LRTS Izmir. Challenges
for the 21st Century, Alten et al. (eds), Balkema,
Rotterdam: 865-875.

9.

Leinala, T; Grabinsky, M; Delmar, R & Collins, J.R.

(2000). Effects of foam soil conditioning on EPBM
performance. Proc. North American Tunneling:
543-552.

CONCLUSIONS

While the feasibility of closed-mode operation in

the stiff plastic clays of the London area was
sometimes questioned prior to the CTRL
construction, the data presented in this paper
demonstrates that
carefully
selected
soil
conditioning allowed chamber pressures of 200 kPa
or more to be accurately controlled in the stiff
London Clay and, to a certain extent, also in the
very stiff clays of the Lambeth Group. Average
FIRs of 50% and PIRs of 7 and 10% were used in
the Thanet Sand and in the Lambeth Group Clays,
respectively. In contrast, much lower quantities of
foam were used in the London Clay. PIRs of about
15% with little or no foam were found adequate to
remould the clay mixtures and allow accurate
control of the machine operation with little or no
pressure decay during ring build. While the
complete absence of conditioning prevented the
machine from being operated in closed-mode, the
injection of large quantities of foam or polymer
solutions may act counterproductively by
exacerbating the drop of pressure in the head
chamber during ring build or by preventing the soil
from being remoulded into a homogeneous mass.
For most of the drives, almost ideal operation
in steady state conditions could be achieved, i.e.
without the need for continuous adjustment of the
thrust, the screw speed or the conditioning
parameters. This resulted in a good control of the
chamber pressure and, in turn, the volume losses
induced by the tunnel construction could generally
be reduced to values of between 0.2 to 0.8% [5]. For
tunnel axis at a depth of 25 m, these volume losses
correspond to maximum settlements of between 1
and 5 mm and demonstrate the quality of tunnel
driving that can be achieved even in the wide and
complex range of ground conditions in the London
area.

10. Merritt, A.S.; Borghi, F.X. & Mair, R.J. (2003).

Conditioning of clay soils for earth pressure
balance tunnelling machines. Proc. Underground
Construction, London Dockland, UK: 455-466.

1476

BIOGRAPHICAL DETAILS OF THE AUTHORS

Dr. Xavier Borghi graduated in Civil
Engineering from the Swiss Federal
Institute of Technology in 2000. He
obtained a Ph.D. in Civil Engineering
at the University of Cambridge (UK).
From 2006 to 2008 he worked in
industry, specializing in the design and
construction of underground structures.
He worked on several large infrastructure projects,
including the CTRL in London and the Gottard Base
Tunnel in Switzerland. In 2008, he joined the newly
founded Kaelin Toscano Ltd as a Partner. Kaelin Toscano
is an consulting firm specialising in the construction
management of large infrastructure projects, as for
example AlpTransits Gotthard and Ceneri Base Tunnels
in Switzerland.

Professor Robert Mair graduated in

Civil Engineering from the University
of Cambridge in 1971. He worked
continuously in industry until 1998,
except for a three year period in the
late 1970s when he returned to
Cambridge to work for his PhD on
tunnelling in soft ground. He was
appointed Professor of Geotechnical
Engineering at Cambridge University in 1998. Throughout
his career he has specialised principally in underground
construction, providing advice on numerous projects
world-wide involving soft ground tunnelling, retaining
structures, deep excavations and foundations. Recent
projects have included railway tunnels in the cities of
London, Barcelona, Bologna, Florence, Rome and
Warsaw, and motorway tunnels in Turkey. He is a Fellow
of the Royal Academy of Engineering and a Fellow of the
Royal Society.

1477