Manual

Hydraulic Structures

Edition 2019
Delft University of Technology
Manual Hydraulic Structures
edition February 2019

dr.ing. M.Z. Voorendt

ir. W.F. Molenaar
PREFACE TO THE 2019 EDITION
The Manual is officially part of the lecture notes for the course “Hydraulic Structures 1” at Delft
University of Technology (course code CTB3355 / CIE3330). It is the result of group work and origins
in Dutch lecture notes that have been used since the execution of the Dutch Delta Works. Amongst
the employees of the Hydraulic Engineering Department that contributed to this work are dr.ir. S. van
Baars, ir.K.G. Bezuijen, ir.G.P. Bourguignon, prof.ir.A. Glerum, dr.ir.P.A. Kolkman, ir. H.K.T. Kuijper,
dr.ir. H.G. Voortman and prof.drs.ir. J.K. Vrijling. The latest years, this manual has been clarified,
revised and expanded by ir. W.F. Molenaar and dr.ing. M.Z. Voorendt. We have received much
feedback from students and got good input from our student-assistants, most recently from Jordi de
Leau, Sjoerd van Til, Lodovico de Vito, Joris den Uijl and Ruben Frijns. All their help is highly
appreciated and has been taken into account in this new edition.

The idea behind the Manual was to have only those tools together that are needed for a first
conceptual design of a hydraulic structure: nothing more than a small collection of formulas, data,
graphs, etc., just from the relevant civil engineering fields. The available tool in the box might not be
the perfect tool for the job, but it will be something to push design of the hydraulic structure a bit
further. Select what is needed, using a good understanding of basic physical laws.

It has become a bit of a challenge to quickly find what is needed in the Manual. An effort has been
made to keep it as accessible as possible by splitting it in 4 parts: General, Loads, Materials and
Structures, dealing with more or less similar subjects and of course by providing an extensive table of
content. Do use the 4-part structure and the table of content to find faster what is needed.

It should be kept in mind that this Manual is intended as educational material as a 'tool box' to learn
how make conceptual designs of hydraulic structures. For professional, detailed designs, more
advanced methods will often be required that strictly comply to official standards, using the right
safety approach.

In the 2019 edition, Chapters 10 and 11 that deal with the changes in water level, flow velocity and
discharge due to the presence of hydraulic structures, have been re-written, for most part by Ruben
Frijns. A new section on types of foundation piles has been added (Section 38.1) and the explanation
of scour near structures has been improved (Section 37.5). The explanation of sheet pile design
(Section 43.2) has been improved as well.

We keep on working on the Manual, because new materials are being introduced, better ways to
calculate loads etc. Suggestions and comments to improve this Manual will be appreciated.

Mark Voorendt
Delft, February 2019

Picture on the front cover: Haringvliet Dewatering Sluices (picture by Mark Voorendt, 2014)
 Manual Hydraulic Structures

Table of contents

Part I: General

1. STANDARDS AND GUIDELINES.....................................................................1

1.1 Standards ..................................................................................................................................... 1
1.2 Guidelines .................................................................................................................................... 3

2. SAFETY OF STRUCTURES .............................................................................4

2.1 Introductory definitions ................................................................................................................. 4
2.2 Failure of structures ..................................................................................................................... 4
2.3 The need for safety factors .......................................................................................................... 5
2.4 Deterministic design (level 0) ....................................................................................................... 6
2.5 Semi-probabilistic design (level I) ................................................................................................ 6
2.6 Probabilistic design (levels II and III) ......................................................................................... 15
2.7 Literature .................................................................................................................................... 18

3. STRUCTURAL MECHANICS .........................................................................19

3.1 Moment and deflection formulas ................................................................................................ 19
3.2 Second moments of area and properties of plane figures ......................................................... 25
3.3 Natural oscillation frequencies ................................................................................................... 29

4. TRIGONOMETRIC, CYCLOMETRIC AND HYPERBOLIC FUNCTIONS .......30

4.1 Trigonometric functions .............................................................................................................. 30
4.2 Cyclometric, or reverse trigonometric functions ......................................................................... 30
4.3 Hyperbolic functions ................................................................................................................... 30

5. UNITS AND CONVENTIONS ..........................................................................31

5.1 Units ........................................................................................................................................... 31
5.2 Conventions for the notation of units and numbers ................................................................... 33

6. DUTCH TRANSLATION OF HYDRAULIC STRUCTURES KEYWORDS ......34

 Part II: Loads

7. WEIGHT ............................................................................................................... 41
7.1 General ............................................................................................................................................. 41
7.2 Design .............................................................................................................................................. 41

8. WIND .................................................................................................................... 42
8.1 Theory .............................................................................................................................................. 42
8.2 Wind loads for the preliminary design of structures ......................................................................... 45
8.3 Wind loads according to Eurocode 1 ............................................................................................... 49
8.4 Wind loads on vessels ..................................................................................................................... 56
8.5 Literature .......................................................................................................................................... 57

9. HYDROSTATIC PRESSURE ............................................................................... 58

9.1 Theory .............................................................................................................................................. 58
9.2 Water pressure on gates .................................................................................................................. 59

10. WATER, FLOW, THEORY ................................................................................... 61

10.1 Conservation of mass ................................................................................................................... 61
10.2 Conservation of energy ................................................................................................................ 62
10.3 Flow in open water ways .............................................................................................................. 65
10.4 Conservation of momentum ......................................................................................................... 68
10.5 Literature ....................................................................................................................................... 69

11. WATER, FLOW, WIDE STRUCTURES ............................................................... 70

11.1 Flow across wide structures ......................................................................................................... 70
11.2 Forces on wide structures ............................................................................................................ 77
11.3 Potential flow and pressure distribution ........................................................................................ 77
11.4 Literature ....................................................................................................................................... 78

12. WATER, FLOW, SLENDER STRUCTURES........................................................ 79

12.1 Drag and lift forces........................................................................................................................ 79
12.2 Drag and lift forces, static part ...................................................................................................... 79
12.3 Drag and lift forces, dynamic part (vibrations) .............................................................................. 82

13. WATER, TIDE AND WIND SET-UP ..................................................................... 84

13.1 Astronomical tide .......................................................................................................................... 84
13.2 Wind set-up................................................................................................................................... 86
13.3 Storm surge .................................................................................................................................. 87
13.4 Other influences............................................................................................................................ 88

14. WATER, WAVES, THEORY................................................................................. 90

14.1 Translation waves ......................................................................................................................... 90
14.2 Wind waves .................................................................................................................................. 91

15. WATER, WAVES, WAVE HEIGHTS .................................................................... 96

15.1 Estimate of wave height and period if no measurements are available ....................................... 96
15.2 The design wave height ................................................................................................................ 97
15.3 Literature ....................................................................................................................................... 99

16. WATER, WAVES, SHALLOWS + BREAKING .................................................. 100

16.1 Shallows: refraction .................................................................................................................... 100
16.2 Shallows: shoaling (non-breaking waves) .................................................................................. 102
16.3 Shallows: breaking waves .......................................................................................................... 103
16.4 Obstacle: reflection ..................................................................................................................... 105
16.5 Obstacle: diffraction .................................................................................................................... 105

17. WATER, WAVES, RUN-UP + OVERTOPPING ................................................. 107

17.1 Wave run-up ............................................................................................................................... 107
17.2 Wave overtopping ....................................................................................................................... 109
17.3 Literature ..................................................................................................................................... 114

18. WATER, WAVES, WALL, NON-BREAKING ..................................................... 115

18.1 Rule of thumb (hydrostatic pressure) ......................................................................................... 115
18.2 Linear wave theory ..................................................................................................................... 115
18.3 Sainflou method .......................................................................................................................... 116
18.4 Rundgren method ....................................................................................................................... 117
18.5 Goda method .............................................................................................................................. 117

19. WATER, WAVES, WALL, BREAKING .............................................................. 119

19.1 Introduction ................................................................................................................................. 119
19.2 Minikin method............................................................................................................................ 119
19.3 CERC 1984 method ................................................................................................................... 120
19.4 Goda-Takahashi method ............................................................................................................ 121
19.5 Comparison ................................................................................................................................ 122

20. WATER, WAVES, SLENDER STRUCTURE, NON-BREAKING ....................... 125

20.1 Theory......................................................................................................................................... 125
20.2 (Preliminary) design .................................................................................................................... 126
20.3 CD coefficient (Shore Protection Manual 1984) ......................................................................... 127
20.4 KD coefficient (Shore Protection Manual 1984) ......................................................................... 128
20.5 KI coefficient (Shore Protection Manual 1984) ........................................................................... 129
20.6 SD coefficient (Shore Protection Manual 1984) .......................................................................... 130
20.7 SI coefficient (Shore Protection Manual 1984) ........................................................................... 131

21. WATER, WAVES, SLENDER STRUCTURE, BREAKING ................................ 132

22. ICE ...................................................................................................................... 133

22.1 Thermal expansion ..................................................................................................................... 133
22.2 Ice accumulation ......................................................................................................................... 133
22.3 Collision ...................................................................................................................................... 134
22.4 Ice attachment ............................................................................................................................ 134
22.5 Design rules ................................................................................................................................ 135
22.6 Literature ..................................................................................................................................... 137

23. TEMPERATURE ................................................................................................ 138

23.1 General ....................................................................................................................................... 138
23.2 Unobstructed thermal deformation ............................................................................................. 139
23.3 Restrained thermal deformation ................................................................................................. 141

24. SOIL - LOADING AND STRESSES ................................................................... 143

24.1 Vertical soil stress ....................................................................................................................... 143
24.2 Horizontal soil stress .................................................................................................................. 148
25. SOIL, SETTLEMENT ......................................................................................... 155

26. SOIL, EARTHQUAKE ........................................................................................ 157

26.1 General ....................................................................................................................................... 157
26.2 Richter scale ............................................................................................................................... 157
26.3 Design ......................................................................................................................................... 159

27. SOIL, GROUNDWATER .................................................................................... 162

27.1 Groundwater pressure ................................................................................................................ 162
27.2 Groundwater flow........................................................................................................................ 163
27.3 Drainage ..................................................................................................................................... 164
27.4 Influences on strength ................................................................................................................ 164
27.5 Influence on stiffness .................................................................................................................. 164

28. SHIPPING, HYDRAULIC ASPECTS .................................................................. 165

29. SHIPPING, BERTHING AND COLLISION ......................................................... 166

29.1 Introduction ................................................................................................................................. 166
29.2 Theory......................................................................................................................................... 166
29.3 Design berthing forces ................................................................................................................ 167
29.4 Design berthing and collision forces ........................................................................................... 171
29.5 Fenders ....................................................................................................................................... 173
29.6 Literature ..................................................................................................................................... 174

30. SHIPPING - MOORING FORCES ...................................................................... 175

30.1 Theory......................................................................................................................................... 175
30.2 Preliminary design ...................................................................................................................... 175
 Part III: Materials

31. SOIL - PROPERTIES ......................................................................................... 179

31.1 Stiffness and strength of soils ..................................................................................................... 179
31.2 Soil investigation ......................................................................................................................... 179
31.3 Determination of soil types ......................................................................................................... 180
31.4 Determination of soil parameters from laboratory tests.............................................................. 180
31.5 Determination of soil parameters from cone penetration tests ................................................... 181
31.6 Determination of soil parameters with the Eurocode table ......................................................... 183
31.7 Soil parameters and models ....................................................................................................... 185
31.8 Literature ..................................................................................................................................... 187

32. SOIL - STRENGTH (RESULTING IN BEARING CAPACITY) ........................... 188

32.1 Theory: strength schematization (Mohr-Coulomb) ..................................................................... 188
32.2 Vertical bearing capacity ............................................................................................................ 189
32.3 Horizontal bearing capacity (resistance against sliding) ............................................................ 194
32.4 Stability of slopes (Fellenius and Bishop) ................................................................................... 195

33. SOIL - STIFFNESS ............................................................................................ 198

33.1 Spring schematisation ................................................................................................................ 198
33.2 Modulus of subgrade reaction .................................................................................................... 199
33.3 Vertical modulus of subgrade reaction (using Flamant) ............................................................. 200
33.4 Horizontal modulus of subgrade reaction for (sheet piling) walls ............................................... 201
33.5 Horizontal modulus of subgrade reactions for piles ................................................................... 202

34. SOIL - SETTLEMENT ........................................................................................ 203

34.1 Consolidation .............................................................................................................................. 203
34.2 Primary settlement and creep..................................................................................................... 205
34.3 Soil relaxation ............................................................................................................................. 207
34.4 Literature ..................................................................................................................................... 208

35. CONCRETE........................................................................................................ 209

35.1 Properties of concrete ................................................................................................................ 209
35.2 Properties of reinforcement steel................................................................................................ 210
35.3 Properties of pre-stressed steel.................................................................................................. 211
35.4 Reinforced and pre-stressed concrete ....................................................................................... 212
35.5 Stiffness of the concrete structure .............................................................................................. 219
35.6 Concrete cover ........................................................................................................................... 222
35.7 Crack width ................................................................................................................................. 224
35.8 Strut & tie modelling ................................................................................................................... 231
35.9 Literature ..................................................................................................................................... 239

36. STEEL ................................................................................................................ 240

36.1 General ....................................................................................................................................... 240
36.2 Strength ...................................................................................................................................... 242
36.3 Stability (buckling)....................................................................................................................... 245
36.4 Welded connections ................................................................................................................... 248
36.5 Bolt connections ......................................................................................................................... 253
36.6 Fatigue (of welds) ....................................................................................................................... 260
36.7 Literature ..................................................................................................................................... 262
 Part IV: temporary and permanent structures

37. STABILITY OF STRUCTURES ON SHALLOW FOUNDATIONS .....................265

37.1 Horizontal stability....................................................................................................................... 265
37.2 Rotational stability....................................................................................................................... 266
37.3 Vertical stability ........................................................................................................................... 267
37.4 Piping (internal backward erosion) ............................................................................................. 268
37.5 Scour protection.......................................................................................................................... 272
37.6 References ................................................................................................................................. 278

38. PILE FOUNDATIONS: COMPRESSION PILES ................................................279

38.1 Pile types .................................................................................................................................... 279
38.2 Bearing capacity of compression piles, general theory .............................................................. 282
38.3 Preliminary design: Pile resistance according to Prandtl and Meyerhof .................................... 282
38.4 Bearing capacity, method Koppejan ........................................................................................... 284
38.5 Stiffness ...................................................................................................................................... 287

39. PILE FOUNDATIONS: TENSION PILES ...........................................................288

39.1 Bearing capacity of tension piles: comparison with compression piles ...................................... 289
39.2 Bearing capacity: cone resistance method ................................................................................. 290
39.3 Bearing capacity: clump criterion................................................................................................ 296
39.4 Edge piles ................................................................................................................................... 298
39.5 Stiffness ...................................................................................................................................... 298
39.6 References ................................................................................................................................. 298

40. PILE FOUNDATIONS: LATERALLY LOADED PILES ......................................299

40.1 Introduction ................................................................................................................................. 299
40.2 Blum: pile theory ......................................................................................................................... 299
40.3 References ................................................................................................................................. 303

41. PILE FOUNDATIONS: PILE GROUPS ..............................................................304

42. STABILITY OF FLOATING STRUCTURES.......................................................305

42.1 Static stability .............................................................................................................................. 305
42.2 Dynamic stability ......................................................................................................................... 312
42.3 References ................................................................................................................................. 314

43. RETAINING STRUCTURES...............................................................................315

43.1 Gravity walls ............................................................................................................................... 315
43.2 Sheet piling ................................................................................................................................. 316
43.3 Combi-walls ................................................................................................................................ 327
43.4 Diaphragm walls ......................................................................................................................... 329
43.5 References ................................................................................................................................. 335

44. CONSTRUCTION PITS: STRUTS AND WALES ...............................................336

44.1 Struts .......................................................................................................................................... 336
44.2 Wales .......................................................................................................................................... 339
44.3 References ................................................................................................................................. 340

45. CONSTRUCTION PITS: ANCHORS ..................................................................341

45.1 General ....................................................................................................................................... 341
45.2 Extreme tensile force .................................................................................................................. 343
45.3 Strength of a tie rod .................................................................................................................... 347
45.4 Total stability ............................................................................................................................... 347
45.5 References ................................................................................................................................. 349

46. CONSTRUCTION PITS: UNDERWATER CONCRETE FLOOR .......................350

46.1 General ....................................................................................................................................... 350
46.2 Limit states .................................................................................................................................. 350
46.3 2-D Arch effect ............................................................................................................................ 351
46.4 3-D Dome effect.......................................................................................................................... 352
46.5 Transfer of forces to piles ........................................................................................................... 353

47. CONSTRUCTION PITS: DEWATERING ...........................................................354

47.1 General ....................................................................................................................................... 354
47.2 Design ......................................................................................................................................... 357

48. FLOOD DEFENCES: THE WATER-RETAINING HEIGHT ................................362

48.1 Introduction ................................................................................................................................. 362
48.2 Deterministic design ................................................................................................................... 362
48.3 Exceedance probability approach .............................................................................................. 363
48.4 Failure probability approach ....................................................................................................... 367
48.5 References ................................................................................................................................. 368

49. FLOOD DEFENCES: GATES ............................................................................369

49.1 Flat gate ...................................................................................................................................... 369
49.2 Mitre gate .................................................................................................................................... 371
49.3 Water pressure on radial gates .................................................................................................. 373
49.4 Water pressure on sector gates ................................................................................................. 376
49.5 Water pressure on arcs .............................................................................................................. 377
Manual Hydraulic Structures

Part I: General
Manual Hydraulic Structures Part I - 'General'

1. Standards and guidelines

Standards, or 'codes', are statutory documents that have to be dealt with by force of law. In the
structural engineering field, they are often material specific, or structure type specific. For instance,
demands regarding the reference period and probability of failure of large hydraulic engineering
projects can deviate from general standards. This means that, for instance, the partial safety factors
(load and material factors) given in these standards cannot be applied. And also, generally,
characteristic values or load factors for typical hydraulic engineering loads, such as waves and
current, are not specified in standards. The reason for this is obvious, as the statistics of waves and
currents differ from one location to another and therefore cannot be generalised in a norm. Moreover,
calculation rules regarding deviating reference periods are lacking in most standards and it is not
specified how to deal with different acceptable probabilities of failure.

This is the reason why more specific guidelines, design handbooks and recommendations have been
published and can be prescribed by the client. For probabilities of failure that deviate from the
standards, one has to resort to probabilistic calculation techniques to determine the design values of
the load and strength. For this, reference is made to course CIE4130 ‘Probabilistic design’.

1.1 Standards
The old Dutch TGB-standards ("Technische Grondslagen voor Bouwconstructies") were officially
withdrawn per 31 March 2010 and were replaced by the Eurocodes, which have a similar structure as
the TGB's. This manual sometimes still refers to the TGB-standards, which could be considered as
outdated, however, for this course this is not a major problem because it deals with main principles in
the first place.

Eurocodes
The basic Eurocode, EN 1990, describes the basic principles and load combinations. The loads for
the design of buildings and other structures are elaborated in the ten parts of EN 1991. Material
properties follow in EN 1992 (concrete), EN 1993 (steel), EN 1994 (steel-concrete), EN 1995 (timber),
EN 1996 (masonry), EN 1997 (soil) and EN 1999 (aluminium). EN 1998 should be used for the design
of structures for earthquake resistance. In addition to the general European standards, obligatory
national supplements have been issued.

Below follows a list of Eurocodes. These codes are available (free for our students!) from the website
of the library of Delft University of Technology (accessible from within Delft campus):
http://www.library.tudelft.nl/, or more directly from: http://connect.nen.nl/.

EN 1990 – Eurocode: Basis of structural design

EN 1991 – Eurocode 1: Actions on structures

EN 1991-1-1 Densities, self-weight and imposed loads
EN 1991-1-2 Actions on structures exposed to fire
EN 1991-1-3 Snow loads
EN 1991-1-4 Wind loads
EN 1991-1-5 Thermal actions
EN 1991-1-6 Actions during execution
EN 1991-1-7 Accidental loads due to impact and explosions
EN 1991-2 Traffic loads on bridges
EN 1991-3 Actions induced by cranes and machinery
EN 1991-4 Actions in silos and tanks

EN 1992 – Eurocode 2: Design of concrete structures

EN 1992-1-1 Common rules for buildings and civil engineering structures
EN 1992-1-2 Structural fire design
EN 1992-2 Bridges
EN 1992-3 Liquid retaining and containment structures

1
Manual Hydraulic Structures Part I - 'General'

EN 1993 – Eurocode 3: Design of steel structures

EN 1993-1-1 General rules and rules for buildings
EN 1993-1-2 Structural fire design
EN 1993-1-3 Cold formed thin gauge members and sheeting
EN 1993-1-4 Structures in stainless steel
EN 1993-1-5 Strength and stability of planar plated structures without transverse loading
EN 1993-1-6 Strength and stability of shell structures
EN 1993-1-7 Strength and stability of plate structures loaded transversally
EN 1993-1-8 Design of joints
EN 1993-1-9 Fatigue strength
EN 1993-1-10 Fracture toughness assessment
EN 1993-1-11 Design of structures with tension components made of steel
EN 1993-1-12 Use of high strength steels
EN 1993-2 Bridges
EN 1993-3-1 Towers, masts and chimneys – towers and masts
EN 1993-3-2 Towers, masts and chimneys – chimneys
EN 1993-4-1 Silos, tanks and pipelines – silos
EN 1993-4-2 Silos, tanks and pipelines – tanks
EN 1993-4-3 Silos, tanks and pipelines – pipelines
EN 1993-5 Piling
EN 1993-6 Crane supporting structures

EN 1994 – Eurocode 4: Design of composite steel and concrete structures

EN 1994-1-1 General – common rules
EN 1994-1-2 Structural fire design
EN 1994-2 Bridges

EN 1995 – Eurocode 5: Design of timber structures

EN 1995-1-1 General rules and rules for buildings
EN 1995-1-2 Structural fire design
EN 1995-2 Bridges

EN 1996 – Eurocode 6: Design of masonry structures

EN 1996-1-1 Rules for reinforced and un-reinforced masonry
EN 1996-1-2 Structural fire design
EN 1996-2 Selection and execution of masonry
EN 1996-3 Simplified calculation methods and simple rules for masonry structures

EN 1997 – Eurocode 7: Geotechnical design

EN 1997-1 General rules
EN 1997-2 Ground investigation and testing

EN 1998 – Eurocode 8: Design of structures for earthquake resistance

EN 1998-1 General rules, seismic actions and rules for buildings
EN 1998-2 Bridges
EN 1998-3 Strengthening and repair of buildings
EN 1998-4 Silos, tanks and pipelines
EN 1998-5 Foundations, retaining structures and geotechnical aspects
EN 1998-6 Towers, masts and chimneys

EN 1999 – Eurocode 9: Design of aluminium structures

EN 1999-1-1 Common rules
EN 1999-1-2 Structural fire design
EN 1999-1-3 Structures subjected to fatigue
EN 1999-1-4 Trapezoidal sheeting
EN 1999-1-5 Shell structures

2
Manual Hydraulic Structures Part I - 'General'

Other standards
Some commonly used standards:
Nederlands Normalisatie instituut: NEN 6702 Belastingen en vervormingen (TGB 1990)
Nederlands Normalisatie instituut: NEN 6740 Geotechniek, Basiseisen en belastingen
Nederlands Normalisatie instituut: NEN 6743 Geotechniek, Drukpalen
Nederlands Normalisatie instituut : NEN 6720 Voorschriften beton (VBC 1995)
Nederlands Normalisatie instituut, NEN 6008: “Steel for the reinforcement of concrete”, july 2008.
Nederlands Normalisatie instituut, NEN_EN 10080: “Steel for the reinforcement of concrete –
Weldable Reinforcing steel - General”, juni 2005.
Nederlands Normalisatie instituut, NEN-EN 10138-1 Draft: “Prestressing steels – Part 1 to 4,
september 2000.

1.2 Guidelines

Empfehlungen des Arbeitsausschusses "Ufereinfassungen" Häfen und Wasserstraßen EAU 2012.

Arbeitsausschusses "Ufereinfassungen" Hamburg; Deutsche Gesellschaft für Geotechnik, Essen,
Germany.
Handbook Quay Walls, second edition. SBRCURnet, Municipality Rotterdam, Port of Rotterdam.
Published by CRC Press/Balkema, the Netherlands, ISBN: 978-1-138-00023-0 (eBook: ISBN
978-1-315-77831-0), 2014.
Handboek damwandconstructies, SBRCURnet-publicatie 166, 6e druk, Stichting CURNET, Gouda,
2012.
Hydraulische randvoorwaarden 2006 voor het toetsen van primaire waterkeringen. Technische
Adviescommissie voor de Waterkeringen.
Ontwerp van Schutsluizen (2 delen). Ministerie van Verkeer en Waterstaat, Rijkswaterstaat,
Bouwdienst.
Prediction of wind and current loads on VLCC's, second edition. Oil Companies International Marine
Forum: Witherby & Co Ltd., London, 1994.
Richtlijn Vaarwegen RVW 2011. Ministerie van Verkeer en Waterstaat, Rijkswaterstaat Adviesdienst
Verkeer en Vervoer. Rotterdam, December 2011.
Wave Overtopping of Sea Defences and Related Structures: Assessment Manual ("European
Overtopping Manual"). EurOtop 2007.
Coastal Engineering Manual. US Army Corps of Engineers; Coastal and Hydraulics Laboratory, 2006,
United States.

Other useful guidelines and technical reports are produced by the 'Technische Adviescommissie voor
de Waterkeringen' which has been succeeded by the 'Expertise Netwerk Waterveiligheid'
(www.enwinfo.nl).

Many guidelines and other documents related to hydraulic engineering can be found at
www.kennisbank-waterbouw.nl.

3
Manual Hydraulic Structures Part I - 'General'

2. Safety of structures
major revision: February 2015

2.1 Introductory definitions

Safety is defined in the ISO-code 8402 as 'a state in which the risk of harm to people or material
damage is limited to an acceptable level. This means that safety is complementary to risk.

Risk in engineering is often quantified as the product of the probability of failure and the
consequences of failure. Probability is the likelihood of an event and its consequences are quantified
as the direct or total economic damage, or the loss of life, given that the hazardous event occurs.

What risk is considered acceptable is usually determined using three criteria:

• Individual risk: the probability that an individual, continuously residing at a certain place during a
year, will perish due to an undesired event;
• Societal risk (group risk): the probability that a (large) number of individuals perishes due to an
undesired event. This is considered a measure for societal disruption;
• Economic optimum: the investments in reducing the failure probability should balance the
prevented loss of economic value.
The risk level that is considered acceptable is subjective and depends, amongst others, on the extent
of voluntariness of exposure to the threat, the recognisability of the threat and social advantages of
opposing the risk.

Reliability is the probability of a structure or system performing its required function adequately for a
specified period of time under stated conditions (Reeve, 2010). In other words: reliability is the
probability of non-failure.

2.2 Failure of structures

To determine the dimensions of structural elements of a civil engineering work, one needs to know
the expected loads and material characteristics. One also needs structural design rules. Nearly all
design rules are derived from failure modes and describe a certain limit state.

A failure mechanism (faalmechanisme) is a description of the way in which a structure is no longer

able to fulfil its function. Not being able to fulfil a function can relate to persistent, transient, accidental
or seismic situations. Failure is permanent if a structure collapses.
Limit states (grenstoestanden) are conditions just before failure. Several limit state types can be
distinguished. Eurocode gives the following overview of limit states:

• Serviceability limit state (SLS, bruikbaarheidsgrenstoestand), indicating disruption of normal use

• Ultimate limit state (ULS, uiterste grenstoestand), indicating collapse of all or part of the structure
- Loss of static equilibrium of the structure or any part of it, considered as a rigid body (EQU).
- Internal failure of the structure or structural elements, including footings, piles, basement walls,
etc., in which the strength of construction materials or excessive deformation of the structure
governs (STR)
- Failure or excessive deformation of the ground in which the strengths of soil or rock are
significant in providing resistance (GEO)
- Fatigue failure of the structure or structural elements (FAT)

Instead of FAT, Eurocode 7 (Geotechnical Design) mentions:

- Loss of equilibrium due to uplift by water pressure (buoyancy) or other vertical actions (UPL)
- Hydraulic heave, internal erosion and piping caused by hydraulic gradients (HYD).

An example of an ultimate limit state of a breakwater (golfbreker) is the overturning of the breakwater
as a result of the collapse of its foundation. Due to its collapse, the breakwater can no longer fulfil its
function. An example of the serviceability limit state is the overtopping of a large number of waves
over the breakwater, in which case there is no guarantee of calm water behind the breakwater.

4
Manual Hydraulic Structures Part I - 'General'

Sometimes also a damage limit state is distinguished, indicating unacceptable damage but no
immediate failure. However, the damage limit state is often included in the ultimate limit state.

During the design process one must take both the ultimate limit state and the serviceability limit state
into account. In our breakwater example, the ultimate limit state refers to the stability and strength of
the structure and subsoil whilst the serviceability limit state is related to the safety of use. (In the
ultimate limit state, stiffness is of importance when deformation induces 'collapse' of the structure or
structural element.)

In general, a structure does not collapse if its loading (solicitation) can be resisted:

S<R

where: S = the load (Solicitation)

R = the resistance to failure (Resistance), or the strength

The type of solicitation and resistance depends on the regarding failure mechanism. It could be a
force if a horizontal or vertical equilibrium is checked or a turning moment if a rotational equilibrium is
considered. If the water retaining height of a flood defence has to be determined, loading and
resistance are expressed as an elevation above reference level (m above NAP in the Netherlands).

For example, in case of the serviceability limit state of a breakwater, the resistance is defined as the
maximum allowed wave height in the harbour and the solicitation is the occurring wave height in the
harbour basin, which is influenced by the geometry of the structure.

In modern standards like the Eurocodes, this is often expressed as a dimensionless unity-check:

S/R<1

The relation between solicitation and resistance can also be expressed as a limit state function
(grenstoestandsfunctie). The general form of a limit state function is:

Z=R-S

If Z < 0 the structure will fail according to the given mode.

2.3 The need for safety factors

In practice, several kinds of uncertainties have to be taken into account while making an engineering
design. There are four main categories of uncertainties:
1. physical or inherent uncertainties;
2. statistical uncertainties;
3. modelling uncertainties;
4. human error.

Physical uncertainties consist of randomness or variations in nature. Variables can differ in time
(water level, for example), or in space (dike height). These uncertainties are mainly caused by a lack
of data of loading or strength. Statistical uncertainties occur if the distribution function of the possible
values for loading or strength is not exactly known, or if the parameters of the distribution function are
determined with a limited number of data. Modelling uncertainties consist of imperfectness of the
models, or failure modes, describing natural phenomena. This can be caused by a lack of knowledge
of these processes, or of over-simplification. Financial uncertainties (like construction costs and
damage costs) are comprised in this category of modelling uncertainties. Finally, human error often
forms a big threat to the reliability of a structure.

5
Manual Hydraulic Structures Part I - 'General'

All these uncertainties can be taken into account by introducing a safety margin between loading and
strength. There are various calculation techniques available to incorporate this margin in a structural
design. These techniques are classified according to the following levels:
• Level 0: deterministic design;
• Level I: semi-probabilistic design;
• Level II: simplified probabilistic design;
• Level III: full probabilistic design.

These methods are briefly explained in the following sections.

2.4 Deterministic design (level 0)

Based on experience, or engineering judgement, overall safety factors (γ) were applied to create a
margin between loading and strength. In general, a structure is considered safe, if:

S·γ < R, where γ > 1,0 [-]

In the Netherlands, the crest height of flood defences used to be based on the highest observed
water level (often the water level that caused most recent flood), plus a freeboard (fb, waakhoogte) of
0,5 to 1,0 metres to account for wave overtopping and uncertainties:

S + fb < R, where 0,5 m ≤ fb ≤ 1,0 m

The estimation of these overall safety factors was not based on a quantification of the uncertainties,
so it was very difficult to determine the extent of over-design (or under-design) relative to some
desired level of safety. This can be overcome by using (semi-) probabilistic techniques, which is
explained in the following sections.

2.5 Semi-probabilistic design (level I)

2.5.1 Theory
In semi-probabilistic design methods, load and strength variables are treated as stochasts, which
means that their possible values are distributed around a mean value µ (Figure 2-1). The
characteristic value of the strength Rk is the value that is exceeded by 95% of the samples. The
characteristic value of the load Sk is the value that is exceeded by only 5% (in other words: the single
tails represent 5% of the possible values).

Figure 2-1 Characteristic values for load and strength

The idea is that, by assuming 95% of the upper limit of the load and by multiplying this with a load
factor, a design value is acquired with a small probability of exceedance. The failure probability then
is very low, especially when these characteristic values are multiplied by partial safety factors.

The characteristic values deviate from the mean values depending on the 'width' of the distribution,
which can be expressed as a function of the standard deviation:

Rk = µR − k ⋅ σ R and Sk = µS + k ⋅ σ S

6
Manual Hydraulic Structures Part I - 'General'

where: µ = mean value of strength µR or load µS

σ = standard deviation of strength σR or load σS
1 N 1 N
(σ = ∑ ( xi − µ )2 or σ = ∑ ( xi − µ )2 for a limited number of samples)
N i =1 N − 1 i =1
k = multiplication constant for the standard deviation to obtain the 5% / 95%value;
k = 1,64 for a normal distribution

These characteristic values of strength and loading are used to obtain the representative values that
are needed to evaluate the limit states (SLS or ULS).

Every load has four representative values:

• the characteristic value (the main representative value), which is found using statistical methods
on a preferably large number of samples, as described above.
• the combination value
If the loads are time-dependent, it is too conservative to add up the representative values of all
loads and to multiply them all by the same partial safety coefficient. After all, the maximum values
of the loads do not necessarily all act on the structure at the same time. This can be overcome by
using the Turkstra rule for the variable loads. According to Turkstra, one load is considered
dominant in every combination of loads. In that case, only averages of the other loads should be
taken into account. The Eurocodes don't work with averages of variable loads, but contain
reduction factors for load combinations (see the following section on load combinations).
• the frequent value
The frequent value is chosen in such a way that it can only be exceeded during a short period. It is
mainly used in the serviceability limit state and in extreme ultimate limit state.
• the quasi-permanent value
It is permissible that quasi-permanent values are exceeded during long periods of time. These
values could be considered as time-averaged values. They are used for long-term effects in SLS,
accidental combinations and seismic design in ULS.

These representative values can be obtained by multiplying the characteristic values by the
combination factor ψ0, frequent factor ψ1, or quasi-permanent factor ψ2 (see the following section for
their values according to the Eurocodes).

The representative values for material properties are mostly the same as the characteristic values,
anyway in the Eurocodes. In some foreign codes the representative material factors could differ from
the characteristic values.

When determining the dimensions of the design in a limit state check, the required strength has to
have a design value larger than the design value of the load. These design values are related to the
representative values through partial safety factors:

Rrep
Rd ≥ Sd ⇔ ≥ γ S Srep
γR

in which: Rrep = representative value for the strength

Srep = representative value for the load
γR = partial safety factor for the strength (material factor) = γM
γS = partial safety factor for the load (load factor) = γg, γq. γR
Rd = design value of the strength
Sd = design value of the load

In hydraulic engineering, for the estimation of water levels needed to estimate hydrostatic loads and
water retaining heights, characteristic values are used with a low probability of exceedance, like
1/1250 or 1/10 000. This is based on statistic calculations of water level measurements, like explained
in Section 13.3.

7
Manual Hydraulic Structures Part I - 'General'

2.5.2 Load combinations

The steps that have to be followed to obtain a design value needed for a design calculation of a load
are:
1. Estimate the types of the load (permanent, variable or accidental);
2. Discern all realistic loads;
3. Estimate the partial load factors (see below) for all relevant combinations of loads;
4. Combine the loads in such a way that the most critical circumstances are obtained.

In case of a load combination with only one variable load, the magnitude of this load is obtained by
multiplying with the concerning partial load factor. If more than one variable loads are combined, the
main variable load should be distinguished from other, possibly simultaneously occurring, loads. A
simultaneously occurring load is always considered as a combination value.

Fundamental load combinations

For fundamental load combinations, the Eurocode distinguishes permanent and variable loads. Loads
from pre-stressing are treated as a separate permanent load and the main variable load is treated
apart from other variable loads.

The design value of the load effect Ed (combined loads) for persistent and transient load combinations
(fundamental combinations) should, according to Eurocode 0, be calculated as:

 n n 
Ed = E  ∑ γ G, j ⋅ Gk , j + γ p ⋅ P + γ Q,1 ⋅ Qk ,1 + ∑ γ Q,i ⋅ψ 0,i ⋅ Qk ,i 
 j ≥ 1 i >1 

For the limit states STR (internal failure or excessive deformation of the structure or structural
member) and GEO (failure or excessive deformation of the soil), the most unfavourable of the
following equations should be used:

 n n 
Ed = E  ∑ γ G, j ⋅ Gk , j + γ p ⋅ P + γ Q,1 ⋅ψ 0,1 ⋅ Qk ,1 + ∑ γ Q,i ⋅ψ 0,i ⋅ Qk ,i 
 j ≥1 i >1 
 n n 
Ed = E  ∑ ξ j ⋅ γ G, j ⋅ Gk , j + γ p ⋅ P + γ Q,1 ⋅ Qk ,1 + ∑ γ Q,i ⋅ψ 0,i ⋅ Qk ,i 
 j ≥ 1 i >1 
where:
E{...} = the combination of the permanent, pre-stressing and variable loads
Gk,j = characteristic value of permanent load j
γG,j = partial factor for permanent load j
ξj = reduction factor for unfavourable permanent load j
P = representative value for the pre-stressing load
γP = partial factor for the pre-stressing load
Qk,1 = characteristic value of the main variable load
γQ,1 = partial factor for the main variable load
ψ0,1 = combination reduction factor for the main variable load
Qk,i = characteristic value of variable load i
γQ,i = partial factor for variable load i
ψ0,i = combination reduction factor for variable load i

It should be judged by the structural engineer what possible loads are useful to combine. The national
annexes to Eurocode 0 give tables with values for the reduction factors γ, ψ and ξ, depending on
failure state, load type and type of building.

8
Manual Hydraulic Structures Part I - 'General'

Load combinations for accidental design situations

In case of combined loads for accidental design situations (fire or impact), the design value of the
load effect Ed should be calculated as:

 
Ed = E  ∑ Gk , j + P + A d + (ψ 1,1 or ψ 2,1) ⋅ Qk ,1 + ∑ψ 2,i Qk ,i 
 j ≥1 i >1 

The choice between ψ1,1Qk,1 or ψ2,1Qk,1 should be related to the relevant accidental design situation
(impact, fire or survival after an accidental event or situation). Ad is the design value of the accidental
action.

Load combinations for seismic design situations

The design value of the load effect Ed during earthquake situations should be calculated as:

 
Ed = E  ∑ G k , j + P + A Ed + ∑ψ 2,i Qk ,i 
 j ≥1 i ≥1 

Reduction factors for the combination of loads

Eurocode 0 recommends values for the load combination factor ψ. This factor is either 1, or ψ1, ψ2 or
ψ3 as indicated in Table 2-1.

Table 2-1 Factor for the combination of loads for buildings (Eurocode 0)
Note: values mentioned in national annexes to Eurocode 0 may differ from this table

9
Manual Hydraulic Structures Part I - 'General'

2.5.3 Partial load factors

Eurocode 0 gives design values of actions in persistent and transient design situations in ultimate limit
state. Static equilibrium (EQU) for building structures should be verified using Table 2-2. The design
of structural members (STR), not involving geotechnical actions should be verified with help of Table
2-3.

For the design of structural members like footings, piles and basement walls (STR), involving
geotechnical actions and the resistance of the soil, verification should be done using one of the
following three approaches:
• Design values from Table 2-4 are applied in separate calculations and Table 2-3 is used for
geotechnical loads as well as for other actions on/from the structure. The dimensioning of
foundations is carried out with the values mentioned in Table 2-4 and the structural resistance
with Table 2-3.
• Table 2-3 is used for both the geotechnical loads as well as for other loads on or from the
structure.
• Table 2-4 is used for geotechnical loads and Table 2-3 is simultaneously used for partial factors
to other actions on/from the structure.

Table 2-2 Partial factors for loads in EQU ultimate limit states (Eurocode 0)
Note: values mentioned in national annexes to Eurocode 0 may differ from this table

10
Manual Hydraulic Structures Part I - 'General'

Table 2-3 Partial factors for loads in STR/GEO ultimate limit states (Eurocode 0)
Note: values mentioned in national annexes to Eurocode 0 may differ from this table.

Table 2-4 Partial factors for non-geotechnical loads on structural members in STR/GEO ultimate limit states (Eurocode 0)
Note: values mentioned in national annexes to Eurocode 0 may differ from this table

Design values of load combinations in accidental and seismic design situations

For an overview of design values of combinations of accidental and seismic loads, see Table 2-5. The
partial factors for loads for the ultimate limit states in the accidental and seismic design situations
should be 1,0. Values for ψ are given in Table 2-1.

11
Manual Hydraulic Structures Part I - 'General'

Table 2-5 Design values of loads for use in accidental and seismic combinations of loads (Eurocode 0)
Note: values mentioned in national annexes to Eurocode 0 may differ from this table

Partial load factors for serviceability limit states

For serviceability limit states the partial factors for loads γS should be equal to 1,0, except if differently
specified in EN 1991 to EN 1999. See Table 2-6.

Table 2-6 Design values of loads for use in load combinations for SLS (Eurocode 0)
Note: values mentioned in national annexes to Eurocode 0 may differ from this table

2.5.4 Partial material factors

Concrete
Partial factors for plain, reinforced, or prestressed concrete in ultimate limit states, γc and γs should be
used as indicated in Table 2-7.

Table 2-7 Material factors for structures in plain, reinforced or prestressed concrete in ULS (Eurocode 2)
Note: values mentioned in national annexes to Eurocode 2 may differ from this table

The partial factor γc should be multiplied by a factor kf (recommended value is 1,1) for the calculation
of the design resistance of cast-in-place piles without permanent casing.

Steel
For steel structures, partial factors as indicated in Table 2-8 should be used.

type of material resistance partial material factor γm

resistance of cross-sections of all steel classes 1,0
resistance of members to instability assessed by member checks 1,0
resistance of cross-sections in tension to fracture 1,25
resistance of various joints see Eurocode 1993-1-8
Table 2-8 Partial material factors for steel structures in ULS (Eurocode 3)
Note: values mentioned in national annexes to Eurocode 3 may differ from this table

12
Manual Hydraulic Structures Part I - 'General'

Soil
The General Eurocode 7 'Geotechnical design' (EN 1997) specifies the use of partial factors for soil
properties, which have to be prescribed per country in a national appendices to the standard. The text
below describes the general use of partial factors, so for practical use, the appropriate national
appendix shall be used. In general, the following partial factors (γM) shall be applied for the verification
of the equilibrium limit state (EQU), when including minor shearing resistances:

soil parameter symbol value

1
angle of internal friction ) γφ' 1,25
effective cohesion γc' 1,25
undrained shear strength γcu 1,4
prism compressive strength γqu 1,4
specific weight γγ 1,0
1
) This factor relates to tanφ'
Table 2-9 Partial factors for soil properties for equilibrium state verification (EQU) (Eurocode 7)
Note: values mentioned in national annexes to Eurocode 7 may differ from this table

For the verification of structural (STR) and geotechnical (GEO) limit states set M1 or set M2 of the
partial factors on soil parameters (γM) shall be applied as given in Table 2-10. The most unfavourable
set M1 or M2 shall be used.

Table 2-10 Partial factors for soil properties For the verification of structural (STR) and
geotechnical (GEO) limit states verification (Eurocode 7)
Note: values mentioned in national annexes to Eurocode 7 may differ from this table

For spread foundations and verifications of structural (STR) and geotechnical (GEO) limit states, set
R1, R2 or R3 of the following partial factors on resistance (γR) shall be applied: γR;v on bearing
resistance and γR;h on sliding resistance, see Table 2-11.

Table 2-11 Partial resistance factors (γR) for spread foundations (Eurocode 7)
Note: values mentioned in national annexes to Eurocode 7 may differ from this table

Partial resistance factors for various types of piles are given in Table 2-12, Table 2-13 and Table 2-14

13
Manual Hydraulic Structures Part I - 'General'

Table 2-12 Partial resistance factors (γR) for driven piles (Eurocode 7)
Note: values mentioned in national annexes to Eurocode 7 may differ from this table

Table 2-13 Partial resistance factors (γR) for bored piles (Eurocode 7)
Note: values mentioned in national annexes to Eurocode 7 may differ from this table

Table 2-14 Partial resistance factors (γR) for continuous flight auger (CFA) piles (Eurocode 7)
Note: values mentioned in national annexes to Eurocode 7 may differ from this table

For retaining structures and verifications of structural (STR) and geotechnical (GEO) limit states, set
R1, R2 or R3 of the partial factors on resistance (γR) shall be applied as mentioned in Table 2-15 : γR;v
on bearing capacity, γR;h on sliding resistance and γR;e on earth resistance.

Table 2-15 Partial resistance factors (γR) for retaining structures (Eurocode 7)
Note: values mentioned in national annexes to Eurocode 7 may differ from this table

For slopes and overall stability and verifications of structural (STR) and geotechnical (GEO) limit
states a partial factor on ground resistance (γR;e) shall be applied. The recommended value for the
three sets R1, R2 and R3 is given in Table 2-16.

14
Manual Hydraulic Structures Part I - 'General'

Table 2-16 Partial resistance factors (γR) for slopes and overall stability (Eurocode 7)
Note: values mentioned in national annexes to Eurocode 7 may differ from this table

Eurocode 7 also gives partial material factors for the verification of the uplift limit state and hydraulic
heave limit state. Furthermore, partial specific load factors are given for the geotechnical limit states
mentioned above.

Important note. This chapter shows a selection of load and material factors from the
General Eurocodes. Therefore, for design calculations in engineering practice, one is advised
to consult the complete text of the Eurocode standards, including the relevant national annex.

2.6 Probabilistic design (levels II and III)

Both level II and level III calculations are probabilistic design methods. Level II methods are
simplifications of full probabilistic design methods, level III. The full probabilistic design, level III, is
explained first in this Manual (Section 2.6.1) and then the simplified methods, level II (Section 2.6.2).

2.6.1 Full probabilistic design (level III)

Level III-methods are full probabilistic approaches in which the probability density functions of all
stochastic variables are described and included in the analysis. A probability density function is a
function that describes the relative likelihood for a random variable to take on a given value. Figure
2-2 shows the probability density functions of the loading fS(S) and strength fR(R) as well as the
resulting probability density function of the limit state fZ(Z). The failure probability pf is represented by
the area where Z < 0 (the small grey area).

Figure 2-2 Probability density distributions of loading and strength, plus corresponding limit state function

A 'wide' distribution around the average limit state value µZ implies a large uncertainty, while a 'tight'
distribution indicates a high certainty. The 'wideness' of the distribution should be judged relative to its
mean value to obtain a good impression of the reliability. A useful expression for judging the reliability
of a structure is the reliability index β, which is related to the mean value and the standard deviation of
the limit state distribution:

µZ
β=
σZ
where: β = reliability index
µZ = mean value of the limit state density function (µZ = µR - µS)

15
Manual Hydraulic Structures Part I - 'General'

σZ = standard deviation of the limit state density function ( σ Z = σ R2 + σ S2 )

It can be seen in Figure 2-2 that µZ = β·σZ.

The influence of the distribution of the load or resistance on the distribution of the limit state function
is usually expressed by the influence coefficient (invloedscoëfficiënt):

σR σ
αR = and αS = − S
σZ σZ

where: α = influence coefficient for the strength (αR) or load (αS)

σ = standard deviation of the strength (σR), load (σS), or limit state function (σZ)

Level III and level II calculations can be used to calculate the partial factors used in level I
calculations, if the reliability index β and influence coefficient α are known:

1 − kR ⋅ VR 1 − αS ⋅ β ⋅ VS
γR = γM = and γ S =
1 − α R ⋅ β ⋅ VR 1 − kS ⋅ VS

where: k = factor indicating the limit of the representative value of strength (kR) or load (kS)
V = coefficient of variation for strength (VR = σR /µR ) or load (VS = σS /µS )

For the calculation of partial factors for building codes for steel and concrete structures it is usual to
adopt kR = 1,64 and kS = 0, but other values may also be chosen. The value of the reliability index β
used for the determination of the partial factors depends on the severity of consequences (in
Eurocode 0 indicated by consequence classes) and the reference period (life time) of the structure.
Values of influence coefficients used for the determination of partial factors are based on calculations
of failure probabilities for a number of reference cases.

The relation between the stochastic variables of loading and strength can be mathematically
described and the probability of exceedance can be calculated for the considered limit state. The
probability of failure pf is the probability that the loading exceeds the resistance:

pf = P (R < S ) = P (Z < 0)

If loading and strength are independent, the failure probability can be calculated using:

pf = ∫∫ fR (r ) fS (s ) dr ds
r <s

where: fR(R) = probability density function of resistance

fS(S) = probability density function of loading

The product of fR(r) and fS(s) is the joint probability function fRS(r,s):

fR (r ) fS (s ) = fRS (r , s )

Figure 2-3 shows such a joint density function fR,S, including the Z = 0 line.

16
Manual Hydraulic Structures Part I - 'General'

Figure 2-3 Joint density function (Melchers, 1999)

The failure probability can therefore be expressed as a function of the joint density function:

r s
pf = ∫ ∫ fR,S (r , s ) dr ds
−∞ −∞

If more than one failure mode or more than one structural element is considered, the failure
probability can be formulated as an n-fold integral. For independent variables xi this looks like:

n
pf = ∫∫ ...∫ Π f x ( xi ) dxi
i =1 i
Z( x ) < 0

Solving this kind of integrals is a tough task, especially if n exceeds 5. However, in some cases with
not too low failure probabilities, the integral can be solved with help of a Monte Carlo simulation.
However, the difficulty of applying probabilistic techniques is that it requires a detailed knowledge of
each variable and the relationship between these variables.

2.6.2 Simplified probabilistic design (level II)

Because of the drawbacks of a full deterministic design, methods have been developed to
approximate the distribution functions of loading and strength. As a simplification, the limit state
function is linearized and for most methods all parameters are considered to be independent and the
probability density functions of loading and strength are replaced by normal distributions (also called
Gaussian distributions):
( x − µx ) 2
−
1 2 σ x2
f (x) = ⋅e , where x can represent load (s) or strength (r).
σ x 2π

Depending on the order of approximation, first-order risk methods (FORMs) or second-order risk
methods (SORMs) can be used.

These methods and more backgrounds on probabilistic design are treated in the lecture notes
CIE4130 'Probability in Civil Engineering'. These probabilistic methods can be used to derive the
partial safety factors needed for a semi-probabilistic design, as prescribed in many building standards
like the Eurocodes (see also Section 2.5).

17
Manual Hydraulic Structures Part I - 'General'

2.7 Literature

CUR/TAW report 141 (1990) Probabilistic design of flood defences. Gouda, CUR.

CUR report 190 (1997) Probability in Civil Engineering, Part 1. Gouda, CUR

Dominic Reeve (2010) Risk and reliability: coastal and hydraulic engineering. Abingdon, Spon Press.

Robert E. Melchers (1999) Structural reliability analysis prediction. Chichester, John Wiley & Sons.

18
Manual Hydraulic Structures Part I - 'General'

3. Structural Mechanics

3.1 Moment and deflection formulas

The 'moment and deflection formulas' (vergeetmenietjes) consist of a set of equations of rotation
angles and deflections for standard load situations on supported and fixed beams.

19
Manual Hydraulic Structures Part I - 'General'

20
Manual Hydraulic Structures Part I - 'General'

Table 3-1 Moment and deflection formulas (from: Hartsuijker / Welleman: Engineering Mechanics, Volume 2)

21
Manual Hydraulic Structures Part I - 'General'

Table 3-2 Singular statically indeterminate beams (Ir.E.O.E van Rotterdam: Sterkteleer 2 toegepaste mechanica)

22
Manual Hydraulic Structures Part I - 'General'

Table 3-3 Twofold statically indeterminate beams (Ir.E.O.E van Rotterdam: Sterkteleer 2 toegepaste mechanica)
(Point C is in the middle of the beam)

23
Manual Hydraulic Structures Part I - 'General'

Table 3-4 Statically determinate beams (Ir.E.O.E van Rotterdam: Sterkteleer 2 toegepaste mechanica)
(Point C is in the middle of the beam)

24
Manual Hydraulic Structures Part I - 'General'

3.2 Second moments of area and properties of plane figures

25
Manual Hydraulic Structures Part I - 'General'

Table 3-5 Second moments of area

Note
The notation for the second moments of area (= moments of inertia) used in this Manual complies to
the notation used in the BSc-courses on structural mechanics at Delft University of Technology. This
notation is different from international literature. If, for example, the x-axis is horizontal and the y-
axis vertical, the notations are as follows: In this Manual: I yy = ∫ y 2dA and in often, in other literature:
I x = ∫ y 2 dA . To put it in other words: Iy y ≠ Iy. This could lead to confusion, so be aware of this.

26
Manual Hydraulic Structures Part I - 'General'

Table 3-6 Properties of plane shapes

27
Manual Hydraulic Structures Part I - 'General'

Table 3-7 Properties of 3D-objects (from: http://cshsyear10maths.global2.vic.edu.au/)

28
Manual Hydraulic Structures Part I - 'General'

3.3 Natural oscillation frequencies

Natural frequencies of structures with a uniform and homogeneous section can be calculated
according to fn = ωn / 2π [Hz] .
EI
The angular velocity is ωn = C [rad/s] ,
µ L4
where:
E [N/m2] = Youngs modulus (elasticiteitsmodulus)
L [m] = length of the beam/girder
I [m4] = moment of inertia (traagheidsmoment)
µ [kg/m] = mass per running metre beam
C [m] = coefficient according to Table 3-8

Table 3-8 Natural frequency (eigen frequency) fn and principal modes of vibration (patterns of motion).
The position of the nodes have been indicated by their distance to the support on the left

29
Manual Hydraulic Structures Part I - 'General'

4. Trigonometric, cyclometric and hyperbolic functions

Chapter added: February 2017

4.1 Trigonometric functions

The trigonometric functions are functions of an angle. They are commonly defined as ratios of two
sides of a right triangle containing the angle, and can equivalently be defined as the lengths of various
line segments from a unit circle (which has a radius of 1).

opposite π 1
sin α = = cos( − α ) =
hypotenuse 2 csc α
adjacent π 1
cos α = = sin( − α ) =
hypotenuse 2 sec α
opposite sin α π 1
tan α = = = cot( − α ) =
adjacent cos α 2 cot α
adjacent cos α π 1
cot α = = = tan( − α ) =
opposite sin α 2 tan α
hypotenuse π 1
sec α = = csc( − α ) =
adjacent 2 cos α
hypotenuse π 1
csc α = = sec( − α ) =
opposite 2 sin α
π
(These equations apply to angles in radians. For angles in degrees: replace by 90°)
2

4.2 Cyclometric, or reverse trigonometric functions

Cyclometric functions are the inverse functions of the trigonometric functions.

So, if x = sin(α), α = arcsin(x) . In an analogous way, arccos(x), arctan(x), arccot(x), arcsec(x) and
arccsc(x) can be defined.

4.3 Hyperbolic functions

Hyperbolic functions are analogues of the ordinary trigonometric functions. They are related to
hyperboles and often occur in the solutions of many linear differential equations. A ray through the
2 2
origin intersects the hyperbole x - y = 1 in point (cosh(A), sinh(A)), where the hyperbole angle A
equals the area enclosed by the ray, the ray mirrored by the x-axis and the hyperbole. The hyperbolic
functions may be defined in terms of the legs of a right triangle covering this sector.
The hyperbolic functions can be expressed as exponential functions:

e x − e− x
sinh( x ) =
2
e x + e− x
cosh( x ) =
2
sinh( x ) e x − e − x
tanh( x ) = =
cosh( x ) e x + e − x
1 e x + e− x
coth( x ) = = x
tanh( x ) e − e − x
1 2
sech( x ) = =
cosh( x ) e x + e − x
1 2
csch( x ) = = x
sinh( x ) e − e − x

30
Manual Hydraulic Structures Part I - 'General'

5. Units and conventions

5.1 Units
The world's most widely used system of physical units, both in everyday commerce and in science, is
the International System of Units (abbreviated SI from the French "Système International d'unités"). A
notable exception is the United States of America, which still uses many old units in addition to SI.
The SI-standard is maintained by the Bureau International des Poids et Mesures in Sèvres, France.
The seven SI base units are presented in Table 5-1.

quantity unit
name symbol name symbol
length ℓ metre m
mass m kilogram kg
time t second s
electric current I ampere A
thermodynamic temperature T kelvin K
amount of substance n mole mol
luminous intensity Iv candela cd
Table 5-1 base units

Base units can be put together to derive units of measurement for other quantities. Some have been
given names. Table 5-2 gives an overview of derived SI units relevant for hydraulic engineering.

quantity unit
name symbol name non-SI units SI base units
(generally used)
-2
acceleration a metre per second squared ms
angle α, β,... radian rad 1
-1
angular velocity ω radian per second rad / s s
2
area A square metre m
4
area moment of inertia I metre to the fourth power m
-1 -2
axial rigidity EA newton per metre Nm kg s
2 -1 -2
bending stiffness EI newton per square metre Nm kg m s
-3
density ρ kilogram per cubic metre kg m
2 -1 -1
dynamic viscosity η pascal second Pa s (N s / m ) kg m s
2 -2
energy, work, heat W joule J (N m) kg m s
3
first moment of area S cubic metre m
-2
force, weight F, G newton N kg m s
-1
frequency f hertz Hz s
-1
impulse, momentum p newton second Ns kg m s
2 -1
kinematic viscosity ν square metre per second m ·s
2 -2
moment of force, torque M newton metre Nm kg m s
3
section modulus W cubic metre m
2 -3
power P watt W (J / s) kg m s
2 -1 -2
pressure, stress σ pascal Pa (N / m ) kg m s
-1
speed, velocity v metre per second ms
3 -2 -2
specific weight γ newton per cubic metre N/m kg m s
3
volume V cubic metre m
-1
wave number k reciprocal metre m
-2 -1 -2
Young's modulus E newton per square metre Nm kg m s
Table 5-2 Selection of derived SI-units

The 'area moment of inertia', or 'second moment of area' in hydraulic engineering is mostly referred to
as 'moment of inertia'. In some literature, the 'section modulus' is indicated as the 'moment of
resistance'. The 'bending stiffness' is also known as 'flexural rigidity'. The 'Young's modulus' is also
known as 'modulus of elasticity', 'elastic modulus' or 'tensile modulus'.

31
Manual Hydraulic Structures Part I - 'General'

Depending on age and place, many non-SI units are in use. The most important non-SI units still in
use are presented in Table 5-3. Extensive lists with conversion factors can be found on internet, see
for example http://www.unc.edu/~rowlett/units/.

quantity unit
name symbol name symbol relation to SI-units
2
area A are a = 100 m
energy W kilocalorie Cal, kcal = 4,1868 kJ
energy W kilowatt-hour kWh = 3,6 MJ
force F ton / tonnes (force) tnf = 1000 kgf ≈ 9,81 kN
force F pound lb (lbf) ≈ 4,448 N
length L mile mi ≈ 1609 m
length L nautical mile nmi, NM = 1852 m
length L yard yd = 0,9144 m
length L foot (international) ft = 0,3048 m
length L inch in (") = 0,0254 m
2
mass m slug slug (lb-s /ft) = 14,59 kg
mass m ton / tonnes (mass) t = 1000 kg
power P horsepower (metric) hp ≈ 735,499 W
pressure σ atmosphere atm = 101 325 Pa
5
pressure σ bar bar = 10 Pa
-1
speed, velocity v knot (international) kn (kt) = 0,514 m s (= 1 nmi/h)
temperature T degree Celsius °C = TK − 273.15
temperature T degree Fahrenheit °F = TK × 1.8 − 459.67
volume V litre l or L = 1 dm³ = 0.001 m³
-
volume V gallon (imperial) gal ≈ 4,546 × 10 ³ m³
-
volume V gallon (USA) gal ≈ 3,785 × 10 ³ m³
Table 5-3 Non-SI units

Prefixes may be added to units to produce a multiple of the original unit. All multiples are integer
powers of ten. See Table 5-4 for an overview of the most used prefixes.

name symbol factor

15
peta- P 10
12
tera- T 10
9
giga- G 10
6
mega- M 10
3
kilo- k 10
2
hecto- h 10
1
deca- da 10
-1
deci- d 10
-2
cent- c 10
-3
milli- m 10
-6
micro- µ 10
-9
nano- n 10
-12
pico- p 10
-15
femto- f 10
Table 5-4 SI-prefixes

For direct conversion of °C to °F and v.v., next formulae apply:

TC = (TF - 32)/1,8
TF = 1,8 TC + 32

Besides the related quantities of density and specific weight, in some anglo-saxon literature 'specific
γ specific material ρspecific material
gravity' (sg) is used. This is defined as: sg = = [-].
γ water ,4°C ρwater ,4°C

32
Manual Hydraulic Structures Part I - 'General'

5.2 Conventions for the notation of units and numbers

2
In design practise, and also in this Manual, stresses and material properties are expressed in N/mm ,
only values for soundings are in MPa. Forces are generally expressed in kN.

The decimal mark and the thousands separator in numbers are written in the style that is used in
most of Europe: a comma is used as decimal delimiter, and a blank space is used as digit grouping
delimiter (for reading comfort). So, for example, one million newton with a precision of two decimals is
written as 1 000 000,00 N.

Many authorities recommend that in scientific notation, when numbers are represented using powers
4
of ten, the exponent of the 10 should be a multiple of 3. So, for example, 1,234·10 should be written
3
as 12,34·10 .

33
Manual Hydraulic Structures Part I - 'General'

6. Dutch translation of Hydraulic Structures keywords

A
abutment landhoofd
angle of internal friction hoek van inwendige wrijving
apron vlonder
aquifer waterhoudende grondlaag
area moment of inertia traagheidsmoment

B
backwater curve stuwkromme
barrier kering
beam ligger
bearing ondersteuning(sconstructie)
bearing capacity draagvermogen
bentonite bentoniet
berth ligplaats
berthing dolphin meerstoel
blinding werkvloer
bollard bolder
bolt bout
breakwater golfbreker
bridge clearance doorvaarthoogte
buckle (to buckle) knikken
building dock / construction dock bouwdok
building pit / construction pit bouwput
bulkhead slingerschot / scheidingswand
buoyancy opdrijving
buttress schoor / steunbeer

C
cable bridge / rope bridge tuibrug
cantilever kraagligger / console
capping beam deksloof
cellular cofferdam cellenwand
chamber lock schutsluis
clearance kielspeling / vrije ruimte
cofferdam bouwkuip / kistdam
compression stress drukspanning
confined aquifer afsgeloten grondlaag met spanningswater
conceptual design voorlopig ontwerp
construction height aanleghoogte
construction pit / building pit bouwput
creep kruip
crest kruin
culvert duiker
cut and cover method wanden-dakmethode
cut-off coupure
cut-off screen kwelscherm (ondoorlatend scherm)
cutting-edge snijrand

D
dam stuwdam
deep foundation paalfundering
density dichtheid (soortelijke massa)
design value rekenwaarde
dewatering gate uitwateringssluis
diaphragm wall diepwand
dike section dijkvak

34
Manual Hydraulic Structures Part I - 'General'

dike segment dijktraject

discharge sluice uitwateringssluis, spuisluis
dissipation chamber woelbak
dolphin dukdalf
dragline sleepgraver (dragline)
draught diepgang
drop gate zakdeur
dry dock droogdok

E
effective stress korreldruk
embedded depth inheidiepte
engineering structure kunstwerk
entry ramp toerit (naar tunnel)
equilibrium state evenwichtssituatie
excavator graafmachine
expansion joint dilatatievoeg

F
falsework steigerwerk / bekisting
fascine mattress zinkstuk (van rijshout)
fender stootblok
fetch strijklengte
fillet weld hoeklas
fixed end inklemming
flange flens
flexural rigidity buigstijfheid
flood overstroming
flood defence waterkering
footing fundament
formwork bekisting
freeboard waakhoogte
frontline waters buitenwateren

G
gantry crane portaalkraan
girder draagbalk / steunbalk
gravity floor gewichtsvloer
ground level maaiveld
ground water table grondwaterstand
guard lock keersluis
guide beam geleidebalk
guide work geleidewerk
guiding system (of a pile ridge) makelaar (van een heistelling)

H
hawser force troskracht
heave hydraulische grondbreuk
high water slack hoogwaterkentering
hinge scharnier
hoist takel
hydraulic structure waterbouwkundige constructie

I
immersed tunnel afgezonken tunnel
inclined plane hellend vlak
inflatable barrier balgkering of schulpkering
influence coefficient invloedscoëfficiënt
inland waters binnenwateren
inundation inundatie (opzettelijk veroorzaakte overstroming)

35
Manual Hydraulic Structures Part I - 'General'

J
jack vijzel
jetty steiger, golfbreker

L
lay-by berth wachtplaats
lean concrete stampbeton
liquefaction zettingsvloeiing
load-bearing structure draagconstructie
lock chamber schutkolk
lock head sluishoofd
low water slack ebkentering

M
masonry metselwerk
mitre gate puntdeur
modulus of subgrade reaction beddingsconstante
mole havenhoofd
moment diagram momentenlijn
mooring facility aanlegvoorziening
mooring force aanlegkracht
mooring post meerpaal
mould mal

N
natural oscillation period eigenperiode
navigation lock schutsluis
navigation width doorvaartbreedte
neap tide doodtij
normal force diagram normaalkrachtenlijn
nut moer

O
overtopping (golf)overslag

P
pad footing poer
pier pijler
pile paal
pile ridge heistelling
pillar pilaar
piping zanduitspoeling, pijpvorming
pivot taats
plunger zuiger
point of fixity inklemmingspunt
prop stempel
pulley katrol
pumping station gemaal

Q
quay kade
quay wall kademuur
quoin wig

R
radial gate segmentdeur
raft foundation strookfundering
raised outer portion of a dike crest tuimelkade
raking pile schoorpaal
recess uitsparing, inkassing
reference crest level dijktafelhoogte, kruinhoogte
reinforcement wapening

36
Manual Hydraulic Structures Part I - 'General'

reinforcement bar yard wapeningsvlechterij

reliability index betrouwbaarheidsindex
roller bearing roloplegging
roller gate roldeur

S
scour ontgronding, wegspoeling
screw schroef
screw thread schroefdraad
screwed pile schroefpaal
section modulus weerstandsmoment
sector gate sectordeur
seepage onderloopsheid, achterloopsheid
seepage screen kwelscherm
self-weight eigen gewicht
settlement zetting
settling klink
shallow foundation fundering op staal
shear force dwarskracht
shear force diagram dwarskrachtenlijn
sheet pile wall damwand
ship-lift scheepslift
shop floor werkvloer
shovel laadschop
shutter schuif / schot
shutter weir klepkering / klepstuw
sill dorpel, drempel
sill beam dorpelbalk, drempelbalk
slab foundation plaatfundering
slackening structure remmingwerk
slack water kentering / doodtij
sliding gate roldeur
sliding surface glijvlak / schuifvlak
sluice caisson doorlaatcaisson
slurry wall diepwand
spalling force spatkracht
span overspanning
specific gravity relatieve dichtheid
specific weight soortelijk gewicht
stop lock spuisluis
stop log schotbalk
storm surge stormvloed
storm-surge barrier stormvloedkering
strip foundation strookfundering
strut stempel / schoor(balk)
subsidence daling (bijv. van bodem, of pneumatisch caisson)
surcharge bovenbelasting
suspension bridge hangbrug / kettingbrug
sway slingering
swell deining
swivel wartel / oogbout

T
tender aanbesteding
tensile stress trekspanning
tidal divide wantij
tie rod / tie bar trekstang, ankerstaaf
timber balk / hout
toe berm plasberm / kreukelberm (bij een dijk)
tow boat sleepboot
trestle aanloopsteiger
truss spant, ligger

37
Manual Hydraulic Structures Part I - 'General'

tug boat sleepboot

tunnel entrance ring dike kanteldijk
turnbuckle spanschroef

U
unconfined aquifer niet-afgesloten grondlaag met freatisch water

V
valve klep
vertical lift gate hefdeur
visor dam vizierstuw
void ratio poriëngetal

W
wale verstevigingsbalk
waling gording
walled building pit bouwkuip
water level waterstand
waterway dike schaardijk
web lijf (bijvoorbeeld van een H-balk)
weir stuw
wedge wig
wind set-up opwaaiing
winch lier

Y
yield stress vloeigrens
Young's modulus elasticiteitsmodulus

An extensive online technical dictionary (English-Dutch / Dutch-English) can be found at:

http://www.tecdic.com/. A dictionary more specifically intended for hydraulic engineering is:
http://www.waterdictionary.info/.

38
Manual Hydraulic Structures

Part II: Loads

Manual Hydraulic Structures Part II - 'Loads'

7. Weight

7.1 General
In many cases, the most important load on a structure is gravity, because in most cases a structure’s self-
weight (a.k.a. 'dead load') constitutes the largest load by far. Sometimes the weight works in favour of the
structure:
• underwater concrete: to reduce the buoyancy (groundwater pressure)
• caisson: more weight means it can take more horizontal shear force
• immersed tunnel: weight is necessary to sink the tunnel to the bottom (of the river)
Sometimes the weight works to the disadvantage:
• bridges and gates: in every span, weight causes extra forces (and moments)
• shallow foundations: the larger the weight, the bigger and more expensive the footing

7.2 Design
The total weight of a structure can be calculated as follows:

G = ∑Vi γ i
i

in which: G [kN] = weight of the structure

3
Vi [m ] = volume of element i
3
γi [kN/m ] = specific weight of element i

The specific weight, or unit weight, γ, is the force on an object due to gravity per unit volume of a material:
γ = ρ·g
3
in which: ρ [kg/m ] = density (mass per unit of volume)
g [m/s²] = acceleration due to gravity
2
g ≈ 9,81 m/s in the Netherlands, at sea level.

The following table gives rough estimates of the specific weights of the most important construction
materials:
material specific weight
3
[kN/m ]
Steel 78
Concrete (reinforced or prestressed) 25
Concrete (not reinforced) 24
Bentonite slurry 11 to13
Wood (soft) 6 to 7
Wood (hard) 8 to 10
Stone (excl. pores) 26
Gravel (dry) 16 to 17
Gravel (wet) 19 to 20
Sand (dry) 17 to 18
Sand (wet) 20 to 21
Clay 15 to 17
Peat 10 to 11
Table 7-1 Volumetric weights

For a more extensive overview of material and fluid densities, see for example http://www.simetric.co.uk/.

Note. Not only the structure’s self- weight causes a load. Objects in the vicinity of the structure can also
cause a (horizontal) load via the ground. This is why in calculations, one assumes a minimum evenly
distributed vertical load (excluding the safety factor γ = 1,5) of qv = 20 kPa next to a sheet pile or a
retaining wall, unless it can be proved that no load (e.g. a concrete mixer truck) could ever reach the
location.

41
Manual Hydraulic Structures Part II - 'Loads'

8. Wind
revision: February 2008; Section 3 updated to Eurocode: February 2015; TGB-method included again (Section 2): 2018

8.1 Theory
Wind is caused by uneven pressures in the atmosphere that exist in so-called high- and low-pressure
areas. The wind climate is a stochastic process, which is described by a number of parameters, such as:
• wind direction
• wind velocity
• turbulence

The wind climate is determined by other factors besides the pressure differences, amongst which are the
topography and the altitude. The parameters mentioned above are stochastic variables, which are
estimated by statistically manipulating measured data concerning the wind climate. The data are gathered
by meteorological institutes. Using a large number of meteorological stations, statistical data is collected
for the description of the wind climate. The collection of data takes place at a set height: 10 m, above
ground level and with a standardised terrain roughness (0,03 m).

Figure 8-1 Wind velocity: one-hour average and fluctuation

The average wind velocity in an hour is described by the so-called long-term distribution (see Figure 8-1).
The fluctuation of the wind velocity within the period of one hour is described by the short-term
description, whereby an average value is distinguished from a value that fluctuates in the course of time:

u x (t ) u x  uɶ x (t )
u (t ) = uy (t ) = uy  + uɶ y (t )
uz (t )  uz  uɶ z (t ) 

in which: u (t ) [m/s] = vector of the wind velocity in all three directions given as a function
of time within an hour
ux [m/s] = average wind velocity in an hour in x-direction
uɶ x (t ) [m/s] = fluctuation of the wind velocity within an hour

The wind direction, wind velocity and turbulence in the lowest layers of the atmosphere play an important
role in the design, realisation and use of hydraulic structures. The friction between the wind and the
earth’s surface is of large importance in this layer. Aspects such as the vegetation and buildings
determine the friction.

In a neutrally stable atmosphere, the time-average wind velocity close to the surface follows a logarithmic
profile (Figure 8-2):

42
Manual Hydraulic Structures Part II - 'Loads'

u*  z 
u (z ) = ln  
k  z0 

where: u ( z ) [m/s] = average wind velocity in an hour at a height z

u* [m/s] = friction velocity, which describes the amount of turbulence
τw
u* = , where τ w = surface shear stress and ρ = fluid (air) density
ρ
k [-] = Von Kármán constant (≈ 0,41)
z [m] = height above the surface (normally this is the earth or marine surface, but in
urban areas the average height of buildings should be taken as surface level)
z0 [m] = roughness height of the surface, dependent on the nature of the terrain
The roughness height varies from 0,0002 m for water surfaces to 1,6 m for
very large cities with tall buildings and skyscrapers.

Figure 8-2 Logarithmic wind velocity profiles (from 'De Ingenieur', 31 July 1970)

The wind fluctuation occurring within an hour can be expressed as a standard deviation from the average:

1 hour

σu = ∫ (u(t ) − u )2 dt
0

The coefficient of variation σ u / u is known as the turbulence intensity I, which at a height z above the
ground surface can be approached by:
k
I (z ) =
 z 
ln  
 z0 

where k [-] = factor dependent on the surface. In rural area: k = 1,0; in built-up area: k = 0,9

From this, the standard deviation of the momentary wind velocity can be calculated with σu = u ⋅ I

The wind velocity is subject to fluctuations with maximum and minimum velocities. The maximums are
known as wind gusts. The maximum momentary wind velocity in a gust can be described with a Rayleigh
 (ξ − u )2 
distribution: P (uˆ > ξ ) = exp  − 
 2 σ n2 
 

43
Manual Hydraulic Structures Part II - 'Loads'

The probability distribution of the maximum wind velocity in a period with n gusts can be determined with:

n
  ( ξ − u )2    ( ξ − u )2 
P (uˆn < ξ ) = P (uˆ < ξ ) =  1 − exp  −
n
  = exp  − n ⋅ exp  − 
  2 σ n2    2 σ n2 
     

This probability distribution is a conditional probability distribution because u must be known. For the
determination of the unconditional probability distribution one is referred to the course CIE4130,
“Probabilistic design”.

The load of the wind on an object is determined by the flow around the object. This flow is an analogue of
the flow around an object in a fluid flow (see Chapter 12).

The force on a structure in the wind direction (drag) and perpendicular to the wind direction (lift) is
described by the following formulas:

1 1
FD = ρ u 2 CD A and FL = ρ u 2 CL A
2 2

1
The term ρ u 2 is known as the dynamic pressure. A is the surface area, CD and CL are coefficients
2
that depend on, amongst other things, the shape of the structure.

The wind velocity is a random variable. Because the dynamic pressure is a function of the wind velocity,
the dynamic pressure is also a random variable. The following is valid for the average dynamic pressure
and the variation coefficient of the dynamic pressure:

1 ∂ pw I ⋅ u
pw = ρ u2 Vpw = = 2⋅ I
2 ∂ u pw

The design dynamic pressure at a height z above ground level can generally be described as:

1
pw ( z ) = (1 + α ⋅ 2 ⋅ I ( z ) ) ⋅ ⋅ ρ ⋅ (u (z ))
2

The old Dutch Technical Basic Rules for Structures (“Technische Grondslagen voor Bouwconstructies,
TGB”) use the value α = 3,5. This value depends on the desired safety level and can be determined by a
risk analysis.

Just like the wind velocity, the force on an object will be part constant and part variable. The variable part
is caused by variations of the wind velocity, but also by possible separation vortices. The latter is
particularly relevant in cases of relatively slender elements (round towers, suspension cables and so
forth). The variable part of the load can cause vibrations of the structure or object. The same way as for
an object in a stationary flow in a fluid, the sensitivity to vibrations can be analysed using the Strouhal
value (see Section 12.3).

44
Manual Hydraulic Structures Part II - 'Loads'

8.2 Wind loads for the preliminary design of structures

Former standard NEN 6702 'Loadings and deformations' (part of the Technische Grondslagen voor
Bouwconstructies, TGB 1990) specified a relatively simple method to determine wind loads. This method,
which could be used for the preliminary design of structures, is summarised below.

If the wind direction is perpendicular to a structure, the wind load can be calculated according to:

prep = Cdim ⋅ Cindex ⋅ Ceq ⋅ φ1 ⋅ pw [kN/m ],

2

in which:
2
prep [kN/m ] = wind load as a result of wind pressure, suction, friction and over- or underpressure
2
pw [kN/m ] = peak velocity pressure (stuwdruk), depending on the height and location of the
structure
Cdim [-] = factor for the dimensions of the structure
Cindex [-] = wind type factor (windvormfactor)
Ceq [-] = pressure dissipation factor (drukvereffeningsfactor)
Φ1 [-] = magnification factor for the dynamic wind component

In the case of most hydraulic structures (h < 50 m and h/b < 5), the wind load equation can be simplified
to:

prep = Cdim ⋅ Cindex ⋅ pw [kN/m ]

2

Peak velocity pressure

The extreme value of the wind velocity pressure (dynamic wind load), working in the direction of the wind,
can be estimated according to:
1
pw = (1 + 7 ⋅ I ( z )) ⋅ ⋅ ρ ⋅ v w2 ( z ) [N/m2 ] ,
2
in which:
I(z) = turbulence-intensity at height z above the surrounding plane:
k
I (z ) =
z−d 
ln  
 z0 
3 3
ρ [kg/m ] = mass density of air (1,25 kg/m )
z0 [m] = roughness length
d [m] = displacement height (reference level)
u* [m/s] = friction velocity
k [-] = factor
z−d 
v w ( z ) = 2,5 u* ln   (wind velocity at height z)
 z0 
The values for z0, d, u* and k are presented in Table 8-1. They depend on the openness of the field and
the location of the structure in the Netherlands, indicated by the velocity pressure zone (I, II or III).

For the calculation of the wind velocity pressure, it should be noticed that for built-on areas, h should be
not less than 9,0 m (even if it is in reality). If h ≤ 2,0 m, the velocity pressure in open areas should be
considered like in built-on areas.

factor open area built-on area

I II III I II III
u* 2,25 2,30 2,25 3,08 2,82 2,60
z0 0,1 0,2 0,3 0,7 0,7 0,7
d 0,0 0,0 0,0 3,5 3,5 3,5
k 1,0 1,0 1,0 0,9 0,9 0,9
Table 8-1 Factors to calculate the peak velocity pressure

45
Manual Hydraulic Structures Part II - 'Loads'

Figure 8-3 wind velocity pressure zones in the Netherlands (from: NEN 6702 / Eurocode 1)

Instead of using the equations above to calculate the peak velocity pressure, the values of Table 8-2 can
be used.

Table 8-2 Peak velocity pressure values pw or qp [kN/m2] in the Netherlands as a function of the height above the surrounding level
(from: NEN 6702 / Eurocode 1)

46
Manual Hydraulic Structures Part II - 'Loads'

Factor for the dimensions of the structure Cdim

The dimensions of the structure should be taken into account, because wind gusts will not always affect
the entire surface of the structure, depending on the proportion of height and width. The correcting factor
for the dimensions Cdim is:

1 + 7 ⋅ I (h) B
Cdim = (≤ 1)
1 + 7 ⋅ I (h)
where:
1 1
B= and I (h ) =
0,4 + 0,021 h 2/3 + 0,029 b 2/3  h 
ln  
 0,2 
h (height) and b (width) are the dimensions [m] in a plane perpendicular to the wind direction.

The value of Cdim can also be found in Table 8-3.

Table 8-3 Factor for the dimensions of the structure Cdim (from: NEN 6702)

Wind type factor Cindex

Combinations of wind pressure, suction, friction and over- or under-pressure should be taken into
account. For the values in case of wind pressure and suction on structural elements, see Figure 8-4.
For over- and under-pressure inside open buildings, a factor with value 0,3 should be applied. In the case
of closed buildings, the factor for overpressure is 0,3 too, but for under-pressure it is:
V 
Cpi = 1,0 − 0,1⋅ log   (V = volume of the building in m ; A = total area of the openings in m ). For walls
3 2

 A
parallel to the wind direction and roofs, different values apply (see NEN 6702). For the total load on a
wall, the various combinations of wind types should be summarized, e.g., for a wall in an closed building
with wind pressure on the outside and wind under-pressure on the inside, Cindex = 0,8 + 0,3 = 1,1.

47
Manual Hydraulic Structures Part II - 'Loads'

Figure 8-4 wind type factors on vertical walls (for structural elements > 10 m2) (from: NEN 6702)

Abroad
If the wind situation (and particularly the wind velocity) is not comparable to the coast or inland areas of
the Netherlands, the dynamic pressure must be determined with the following formula:

q= 1
2 ρ u 2 CD
3 3
In which: ρ [kg/m ] = air density (1,25 kg/m )
CD [-] = drag coefficient, dependent on Reynolds value: CD ≈ 0,7

For a more accurate approximation of the drag coefficient CD see §12 "Water, flow, slender structures".
This CD value depends on a number of different factors, including the shape and size of the structure or
object. The factors are dealt with in larger detail in NEN 6702.

The average velocity is dependent on the altitude, the location and the measurement time according to
the formula:
α
 h 
u = β ∆T   u1h,10m
 10 m 
in which: β∆T [-] = correction factor for the duration of the gust
h [m] = altitude of measurement
α [-] = parameter for the roughness of the area
(α = 0,13 (coast), α = 0,19 (inland))
u1h,10m [m/s] = average wind velocity at an altitude of 10 m, during one hour

The wind velocity is given as an average for one hour. A gust of 3 seconds can reach a higher average
velocity. The difference is indicated by the parameter β∆T (see following table).
The average wind velocity u1h,10m is dependent on both the location and the probability of exceedance
Tyear. According to the TGB, the following values are applicable in the Netherlands.

β ∆T Tyear = 1 × per 5 Tyear = 1 × per 75

years years
Land 1,75 u1h,10m = 20,5 m/s u1h,10m = 25,0 m/s
Coast 1,40 u1h,10m = 26,0 m/s u1h,10m = 32,0 m/s
Table 8-4 Wind velocities according to TGB

For other exceedance frequencies u1h,10m may be linearly interpolated and extrapolated over log(Tyear).
The given wind velocities apply to the Netherlands. These values must be measured again for other
areas.

Notes
• For slender structures it is necessary to consider the load in transverse direction as well ( FL = ½ ρ u 2 CL ).
• For slender structures it is necessary to test the structure for vibrations caused by gusts.
(For this see the TGB 1972, or: Wind load, a commentary on the TGB ("Windbelasting, toelichting op de TGB”)
1972).

48
Manual Hydraulic Structures Part II - 'Loads'

8.3 Wind loads according to Eurocode 1

According to Eurocode 1991, part 1-4 'General actions - Wind actions', wind loads on structures (lower
than 200 m) and structural elements shall be determined taking account of both external and internal wind
pressures. This section of the Manual endeavours to present the main design procedure for wind loads,
but especially the determination of the structural factor and the force coefficient is quite complex. For
details and additional information, one is referred to the original text of Eurocode 1-4 (about 120 pages!).
Furthermore, details can vary per country, so mind the National Annexes!

Important preliminary note

The Eurocode distinguishes between wind pressure, which is relevant for the design of cladding (gevel-
bekleding), fixings and structural parts (Section 8.2.1), and wind force, which should be used for overall
wind effects (Section 8.2.2).

8.3.1 Wind pressure on surfaces

The net pressure on a wall, roof or element is the difference between the pressures on the opposite
surfaces taking due account of their signs. Pressure, directed towards the surface is taken as positive,
and suction, directed away from the surface as negative. See Figure 8-5 for some examples.

Figure 8-5 Pressure on surfaces (Eurocode 1)

The wind pressure acting on external surfaces we should be calculated according to:

w e = q p ( ze ) ⋅ c pe

The wind pressure acting on the internal surfaces of a structure wi should be obtained from:

w i = q p ( zi ) ⋅ c pi

2
where: qp(zx) [kN/m ] = peak velocity pressure (extreme stuwdruk) at reference height zx
zx [m] = reference height for the external pressure ze or internal pressure zi
cpx [-] = pressure coefficient for the external pressure cpe or internal pressure cpi

The value for the peak velocity pressure qp can be calculated according to:

1
q p ( z ) = [1 + 7 ⋅ Iv ( z )] ⋅ ⋅ ρ ⋅ vm
2
(z )
2

where: Iv(z) [-] = wind turbulence intensity at height z

3 3
ρ [kg/m ] = air density (recommended value is 1,25 kg/m )
vm(z) [m/s] = mean wind velocity at height z

49
Manual Hydraulic Structures Part II - 'Loads'

The mean wind velocity vm(z) depends on the terrain roughness, orography (= topographic relief of the
terrain) and the basic wind velocity, but for the Netherlands a table is available where the value for the
peak velocity pressure qp can easily be found (Table 8-2). The three pressure zones in the Netherlands
can be found in Figure 8-3.

The external reference height ze for windward walls of rectangular plan buildings depends on the aspect
ratio h/b and is always the upper height of the different parts of the walls, see Figure 8-6. The
recommended procedure is to take the height of the building as the reference height, but this can vary per
national annex of the Eurocode.

Figure 8-6 Reference height for wind loads (Eurocode 1)

The reference height zi for the internal pressures should be equal to the reference height ze for the
external pressures on the faces which contribute by their openings to the creation of the internal pressure.
If there are several openings, the largest value of ze should be used to determine zi.

External and internal pressures should be considered to act at the same time. For design conditions, the
worst combination of external and internal pressures shall be considered. This has to be done for every
combination of possible openings and other leakage paths.

The external pressure coefficient cpe depends on the wind direction and on the size of the loaded area A
that produces the wind action in the section to be calculated. The external pressure coefficients are given
2 2
for loaded areas A of 1 m (indicated with Cpe,1) and 10 m (Cpe,10) in the tables. Values for Cpe,1 are
2
intended for the design of small elements and fixings with an area of 1 m or less, while values for Cpe,10
may be used for the design of the overall load bearing structure of buildings.

The different configurations are presented in Figure 8-7 and the corresponding external pressure
coefficients in Table 8-5.

Figure 8-7 Key for vertical walls (Eurocode 1)

50
Manual Hydraulic Structures Part II - 'Loads'

Table 8-5 External pressure coefficients for the Netherlands (Eurocode 1 - Dutch national annex)

Reference is made to Eurocode 1 for the external pressure coefficients of all kinds of roofs and domes.
Internal pressure does not only depend on the shape of the structure, but also on the number of openings
and their location relative to the wind direction. For structures with a 'dominant face' the internal pressure
should be taken as a fraction of the external pressure at the openings of the dominant face. The following
two equations can be used for internal pressure coefficient for faces with less than 30% openings:
• When the area of the openings at the dominant face is twice the area of the openings in the
remaining faces,

cpi = 0,75 ⋅ cpe

• When the area of the openings at the dominant face is at least 3 times the area of the openings in
the remaining faces,

cpi = 0,90 ⋅ cpe

The internal pressure coefficient of open silos and chimneys is cpi = -0,60 and of vented tanks with small
openings it is should be based on cpi = -0,40. In these cases the reference height zi is equal to the height
of the structure.

For free-standing walls and parapets, the values of the resulting pressure coefficients cp,net depend on the
solidity ratio ϕ. For solid walls, ϕ = 1, and for walls which are 80% solid (i.e. have 20 % openings) ϕ = 0,8.
The pressure coefficients cp,net mentioned in Table 8-6 should be used for the zones indicated in Figure
8-8. For porous walls and fences with a solidity ratio of less than 0,8 reference is made to Eurocode 1.

Figure 8-8 Key to zones of free-standing walls and parapets (Eurocode 1)

Table 8-6 Recommended pressure coefficients cp,net for free-standing walls and parapets (Eurocode 1)

51
Manual Hydraulic Structures Part II - 'Loads'

For details and specific regulations per country, reference is made to Eurocode 1, part 1-4 and the
national annexes to this standard.

8.3.2 Wind forces on whole structures

For the determination of wind forces on whole structures or structural components, the peak velocity
pressure should be multiplied with the area of the wall and several coefficients:

Fw = cs cd ⋅ cf ⋅ q p ( ze ) ⋅ A ref

where: Fw [kN] = wind force on a structure or structural component

cs [-] = coefficient that takes into account the effect on wind actions from the
non-simultaneous occurrence of peak wind pressures on the surface
cd [-] = coefficient for the effect of vibrations of the structure due to turbulence
cf [-] = force coefficient
2
qp(ze) [kN/m ] = peak velocity pressure at reference height ze (Table 8-2)
2
Aref [m ] = reference area of the surface

An alternative calculation method uses the wind pressures as calculated in the previous section:

• For external forces:

Fw ,e = cs cd ⋅ ∑ (w
surfaces
e ⋅ Aref )

• For internal forces:

Fw ,i = ∑ (w
surfaces
i ⋅ Aref )

• For friction forces (on external surfaces Afr parallell to the wind direction):

Ffr = cfr ⋅ q p ( ze ) ⋅ A fr

where cfr [-] = friction coefficient according to Table 8-7.

Table 8-7 Wind friction coefficients cfr

The force coefficient cf of structural elements of rectangular section with the wind blowing normally to a
face is determined by:

cf = ψ r ⋅ψ λ ⋅ cf ,0

where: ψr [-] = reduction factor for square sections with rounded corners. ψr depends on the
Reynolds number and has a minimum value of 0,5.
For sharp edges ψr ≈ 1,00

ψλ [-] = end-effect factor for elements with free-end flow.

ψλ depends on the effective slenderness λ and the solidity ratio φ of the structure.

52
Manual Hydraulic Structures Part II - 'Loads'

The effective slenderness λ of polygonal, rectangular and sharp edged sections of

structures is:
• for ℓ ≥ 50 m, λ =1,4·ℓ/b or λ = 70, whichever is smaller
• for ℓ <15 m, λ = 2,0·ℓ/b or λ = 70, whichever is smaller
• for 15 m ≤ ℓ < 50 m: linear interpolation should be used
The solidity ratio φ is the ratio between the net area (= the area excluding openings)
and the overall envelope area Ac = ℓ·b. For closed structures, φ = 1,0.
The end-effect factor can be determined when λ and φ are known with help of
Figure 8-13.

Figure 8-9 End-effect factor as a function of the effective slenderness (Eurocode 1)

cf,0 [-] = force coefficient of rectangular sections with sharp corners and without free-end flow
as given by Figure 8-10.

Figure 8-10 Force coefficient for rectangular sections with sharp corners and without free end flow

53
Manual Hydraulic Structures Part II - 'Loads'

The coefficients cs and cd can be combined into one structural factor cscd. For buildings with a height less
than 15 m the value of cscd may be taken as 1. For (rectangular) civil engineering works (other than
bridges) the structural factor should be calculated according to:

1 + 2 ⋅ k p ⋅ Iv ⋅ B 2 + R 2
cs cd =
1 + 7 ⋅ Iv ( zs )

where: kp [-] = peak factor defined as the ratio of the maximum value of the fluctuating part
of the response to its standard deviation:
0,6
k p = 2 ⋅ ln(ν ⋅ T ) + with a minimum of kp= 3,0
2 ⋅ ln(ν ⋅ T )
T [s] = averaging time for the mean wind velocity (T = 600 s)
ν [Hz] = up-crossing frequency as given in:
R2
ν = n1,x ⋅ and ν ≥ 0,8
B2 + R 2
n1,x [Hz] = natural frequency of the structure. The limit of ν = 0,08 Hz
corresponds to a peak factor of 3,0.

Iv [-] = turbulence intensity = the standard deviation of the turbulence σv divided by the mean
wind velocity vm(z):
σv kI
Iv ( z ) = = for zmin ≤ z ≤ zmax
v m ( z ) co ( z ) ⋅ ln( z z0 )
Iv ( z ) = Iv ( zmin ) for z < zmin

kI [-] = turbulence factor. The recommended value for kI is 1,0.

co [-] = orography factor. Where orography (e.g. hills, cliffs etc.) increases
wind velocities by more than 5% the effects should be taken into
account using the orography factor co. Otherwise co = 1,0.
z0 [m] = roughness length as given in Table 8-8
zmin [m] = is the minimum height defined in Table 8-8.
zmax [m] = is to be taken as 200 m

Table 8-8 Roughness length z0 and minimum height zmin (Eurocode 1)

2
B [-] = background factor, allowing for the lack of full correlation of the pressure on the
structure surface:
1
B2 = 0,63
b+h
1 + 0,9 ⋅  
 L( zs ) 
L(zs) [m] = turbulent length scale
2
It is on the safe side to use B = 1,0.
2
R [-] = resonance response factor, allowing for turbulence in resonance with the vibration
mode:

54
Manual Hydraulic Structures Part II - 'Loads'

π2
R 2= ⋅ SL ( zs , n1,x ) ⋅ Rh (ηh ) ⋅ Rb (ηb )
2⋅δ

δ [-] = total logarithmic decrement of damping:

δ = δs + δa + δd
δs = logarithmic decrement of structural damping, for some
structural types indicated in Table 8-9, otherwise see
Eurocode 1, Annex F.5

Table 8-9 logarithmic decrement of damping (Eurocode 1)

δa = logarithmic decrement of aerodynamic damping for the

fundamental mode. If the modal deflections are constant over
each height z, which is mostly the case, δa may be determined
with:
c ⋅ ρ ⋅ b ⋅ v m ( zs )
δa = f
2 ⋅ n1 ⋅ me
n1 [Hz] = fundamental flexural frequency of the structure
me [kg] = equivalent mass per unit length
3 3
ρ [kg/m ] = air density (≈ 1,25 kg/m )

δd = logarithmic decrement of damping due to special devices

(like tuned mass dampers and sloshing tanks etc.). If these
special dissipative devices are added to the structure, δd should
be calculated by suitable theoretical or experimental techniques.

SL [m] = is the non-dimensional power spectral density function:

α
z
L( z ) = Lt ⋅   for z ≥ zmin
 zt 
L( z ) = L( zmin ) for z < zmin
zt [m] = reference height = 200m
zmin [m] = minimum height (Table 8-8)
α = 0,69 + 0,05 ⋅ ln( z0 ) , where z0 = roughness length (Table 8-8)
Lt [m] = reference length scale = 300 m
zs [m] = reference height for determining the structural factor, see Figure 8-11

Figure 8-11 General shapes of structures and the reference height zs

2
For a preliminary design one could consider to use R = 1, which could result in a over-estimation
of the structural factor cscd of up to 15% (the approved minimum value of cscd is 0,85).

55
Manual Hydraulic Structures Part II - 'Loads'

8.4 Wind loads on vessels

The Oil Companies International Marine Forum (OCIMF) issued an information paper with coefficients
and procedures for computing wind and current loads on oil tankers. According to this paper, the resultant
wind force on a moored tanker (a very large crude carrier, i.e. 150 000 < DWT < 500 000) can be
calculated using the following equations.

1 1
FXw = C Xw ρw v w2 ,10 AT and FYw = CYw ρw v w2 ,10 AL
2 2
in which:
FXw [N] = longitudinal wind force (surge wind force)
FYw [N[ = lateral wind force (sway wind force)
CXw [-] = longitudinal wind drag coefficient, see Figure 8-12
CYw [-] = lateral wind force coefficient, see Figure 8-13
3 3
ρw kg/m ] = density of air (≈1,28 kg/m )
vw,10 [m/s] = wind velocity at 10 m elevation
2
AT [m ] = transverse (head-on) wind area
2
AL [m ] = longitudinal (broadside) wind area

(Longitudinal and transverse directions are with respect to the vessel length axis)

The wind force coefficients depend on the wind angle of attack, the extent of loading by cargo (fully
loaded or only ballasted) and the bow configuration.

If the wind velocity is known at a different elevation than 10 metres above water surface, the following
equation can be used for conversion:
1/7
 10 
vw ,10 = uw ,h  
 h 
where uw,h [m] = wind velocity at elevation h above ground/water surface.

Figure 8-12 longitudinal wind drag coefficient for tankers

56
Manual Hydraulic Structures Part II - 'Loads'

Figure 8-13 lateral wind force coefficient

8.5 Literature
• Nederlands Normalisatie-instituut: NEN 6702 "Technische Grondslagen voor Bouwconstructies",
Belastingen en vervormingen, (TGB 1990), Delft 1990.
• Eurocode 1: Actions on structures - Part 1-4: General actions - Wind actions (NEN-EN 1991-1-4)
• Oil Companies International Marine Forum: Estimating the environmental loads on anchoring systems,
first edition. Witherby & Co Ltd., London, 2010.

57
Manual Hydraulic Structures Part II - 'Loads'

9. Hydrostatic pressure

9.1 Theory
Shear stresses only occur in fluids if there are velocity differences between the fluid particles. In a fluid at
rest there is no shear stress and the pressures in one point are equal in all directions. This characteristic
is known as Pascal’s law. For flowing fluids, Pascal’s law gives rather good approximations.

The hydrostatic (water) pressure in any given point under water is a function of the pressure head and the
density of the water:

p = ρw g h

where: p [Pa] = water pressure

h [m] = pressure head
2
g [m/s ] = acceleration due to gravity
3
ρw [kg/m ] = density of water (ρw,salt = 1025; ρw,fresh = 1000)

In water at rest or in a uniform flow (no acceleration or deceleration) in an open waterway, the pressure
head is equal to the water depth at the point considered. Largely varying flow velocities in open
waterways correspond to largely curved flow lines. In such cases the pressure head is not equal to the
water depth (see Chapter 11).

Figure 9-1 Water pressure

The force on a plain under water is calculated by integrating the pressure over the plain:

F = ∫ p dA
A
in which: F [N] = the force perpendicular to the plane
2
dA [m ] = a small part of the area (see Figure 9-1)
2
A [m ] = the total surface area

If the plane on which the force is working is straight, the force is equal to the product of the area of the
plane and the water pressure in the mass centre of the plane. The components of the force in the x, y and
z directions can be determined with:

Fx = ∫∫ p dy dz
A
Fy = ∫∫ p dx dz
A
Fz = ∫∫ p dx dy
A

The co-ordinates of the line of action of the force in the x direction can be found with:

∫∫ p y dy dz ∫∫ p z dy dz
yx = A
and zx = A

Fx Fx

The lines of action of the forces Fy and Fz can be found in a similar manner.

The previous equations correspond exactly to the equations for the volumes and mass centres of the

58
Manual Hydraulic Structures Part II - 'Loads'

pressure diagrams. This makes it possible to determine the forces and their points of application for
known shapes using relatively simple standard formulas.

9.2 Water pressure on gates

Water pressure has a linear relationship with the location beneath the water surface:

σ w ( z ) = ρw g z
where:
2
σw(z) [N/m ] = water pressure at depth z
3
ρw [kg/m ] = density of water (Notice that the densities of sea water and fresh water are
different)
g [m/s2] = gravity acceleration
z [m] = depth below water surface

Water pressure is equal in all directions (Law of Pascal), so, the horizontal water pressure at a certain
point equals the vertical pressure. The direction of the water pressure on a surface is perpendicular to
that surface (a wall or a gate, for example) (Figure 9-2). The total water force on a straight surface can be
calculated by multiplying the average pressure on the surface with the area of the surface:

σ min + σ max
F= ⋅ l gate ⋅ w gate
2
where wgate is the width of the gate.

Figure 9-2 Determination of the force on a straight surface

In case of a circular gate, the principle is the same (Figure 9-3). For practical reasons, however, it could
be more convenient to use a polar system of coordinates.

Figure 9-3 Determination of the force on a curved surface

In most cases, however, it is easier to consider the vertical and horizontal component of the water
pressure on an object separately. The resulting force in x-direction, Fh, is equal to the summarized
horizontal pressure on a projection of the gate to a vertical plane. The total force in z-direction, Fv, is
equal to the weight of the water volume above the gate. This is indicated in Figure 9-4. The total force on

59
Manual Hydraulic Structures Part II - 'Loads'

the gate then is Fgate = Fh2 + Fv 2

Figure 9-4 Simple determination of the water force on a straight gate (left) and a curved gate (right)

In the case of a radial gate with a vertical axis, the gate is curved in the horizontal plane to an arc with a
centre that coincides with the position of the hinge. Thus the resultant of the water pressure load goes
precisely through the hinge in the horizontal plane. The direction of the resultant depends on the
lengthwise profile of the gate. In the vertical plane, the direction and the position of the resultant of the
water pressure depends on the shape of the gate in the vertical cross section. Figure 9-5 shows a
number of possible cross sections and the forces on the gate per metre. The position of the seal with the
bottom is an important detail which determines the size of the upward force.

Horizontal water pressure Uplift water pressure

Figure 9-5 Loads dependent on the shape of the gate

For more information about the design of gates, see Chapter 49 of this manual (part IV).

60
Manual Hydraulic Structures Part II - 'Loads'

10. Water, flow, theory

revised: 2019

In hydraulic engineering the conservation of quantities is often used to deduce equations. Conservation of
mass will be discussed first. After that, the conservation of energy will be dealt with, followed by the flow
in open waterways which is related to that. Conservation of momentum is included in this chapter, which
is useful for determining the flow across and forces on wide structures (walls), as explained in the next
chapter.

10.1 Conservation of mass

In hydraulic engineering, the law of conservation of mass is used to determine the flow characteristics.
Starting point is the choice of a control volume. The quantity mass within this control volume can change
during a time interval by flow through the boundaries. The so-called continuity equation is the applied
conservation law of mass:
d
m = ∑ min −∑ mout
dt

The accumulated mass in the specified control volume can only change by in- and outflow through the
boundaries, since mass cannot disappear or originate from nothing.

For water-related problems it is assumed that water is incompressible. In this case the mass balance and
volume balance are equal. The in- and outflow of water is named discharge in hydraulic engineering. The
continuity equation turns into:
d
V = Qin − Qout
dt

This equation is even more simplified since many situations consist of steady conditions. Steady flow
means that the flow at each location is constant in time. The flow conditions, including the water volume
inside the control volume, don’t change. Therefore, the discharge of the inflow and outflow should be
equal.
d
V = 0 ⇒ Qin = Qout ⇔ A in uin = A out u out
dt
2
in which: A [m ] = area
u [m/s] = average flow velocity

For prismatic channels (having a constant width), the equation can be reduced to a 2D-situation, in which
the area is replaced by the flow depth d.

Since the cross-sectional area at location 1 in Figure 10-10-1 is smaller than at location 2, the average
flow velocity at location 2 is smaller than at location 1. The increase of the cross-section affects the flow
velocity, not the discharge. This can be beneficial, as the narrowing reduces the friction of the flow, which
prevents the erosion of river beds.

Figure 10-10-1 Widening of pipe flow (Engsoft Inc., n.d.)

The continuity equation can be used for small basins, like ports and creeks, and polder systems with
pumping stations as well. In these unsteady situations, the in- and outflow of water results in a change of

61
Manual Hydraulic Structures Part II - 'Loads'

water level. For small systems, inertia can be neglected and the slope of the water level is negligible: the
water level will go up and down and is nearly horizontal all times, see Figure 10-10-2 for an example: a
small basin with a wide gap (the water level will not be horizontal in case of a narrow gap). The
associated equation is:
dh
Qin = Ab b
dt
3
in which: Qin [m /s] = discharge (inflow is positive)
2
Ab [m ] = surface area (of the basin)
hb [m] = water depth (in the basin)

Figure 10-10-2 Small-basin approach (Battjes & Labeur, 2014, modified)

This equation assumes that the surface area is constant for every water depth. However, especially for
creeks, estuaries and rivers this is not true, because the boundaries consist of sloped embankments and
floodplains will flood during periods of higher water levels.

10.2 Conservation of energy

The concept for the conservation of energy is the same as for the conservation of mass. The conserved
quantity is energy, which can be transported by the flow in and out of the control volume. Besides that,
pumps add energy, turbines and friction lead to extraction or loss of energy.

dE
= ∑ Ein − ∑ Eout + E pump − Eturbine − Eloss
dt

Again, a steady situation is analyzed almost continuously, so the change of energy is zero. Furthermore,
in hydraulic engineering, the amount of energy is expressed in a total head. This enables a graphical
presentation of the energy head.

The next step is the definition of the in- and outflow. Fluid particles flowing through the boundaries of the
control volume contain potential and kinetic energy. The well-known formulas for these types of energy
are presented below.
Ekin = 1 mu 2
2
m
E pot = mgz + p
ρ
where: Ekin [J] = kinetic energy
Epot [J] = potential energy
m [kg] = mass
u [m/s] = velocity
z [m] = height in reference system
3
ρ [kg/m ] = density
p [Pa] = pressure

62
Manual Hydraulic Structures Part II - 'Loads'

The first term of the potential energy represents the height of the fluid particle, the second term is related
to the work done by the pressure. The energy head is expressed in meters by summarizing the equations
above and dividing by the term mg. The result is described by:

p u2
H =z+ + ,
ρ g 2g
where H is the total energy head. This energy head consists of the piezometric head ( z + p ) and the
ρg
2
velocity head ( u ).
2g

These definitions are also indicated in

Figure 10-10-3. The piezometric head is constant for a static body of fluid of constant density. At the
bottom of the filter in
Figure 10-10-3, the elevation head is quite small, but the pressure head is large, because there is a
significant amount of water pressure acting on the bottom of the filter. At the level of the water table, the
water pressure and thus the pressure head are nearly zero. However, the elevation head is large at this
point by which the resulting piezometric head is the same. Therefore, no water will flow upwards or
downwards.

Figure 10-10-3 The components of the hydraulic head (Hendrayana, 2012)

The reference level can be chosen by the engineer; often the bottom of the water body is practical. With
this choice, the piezometric head in a lake is equal to the water depth. The same applies to rivers,
provided that the streamlines are straight and parallel.

In comparison to the mostly static conditions in lakes, the velocity head plays a more dominant role in
pipe flow and rivers. In these situations, the energy conservation law for steady incompressible flow, as
defined by Bernouilli, is frequently written as:

2 2
u u
h1 + 1 + hp = h2 + 2 + ht + hL
2g 2g

where: h1 [m] = piezometric head at boundary 1

h2 [m] = piezometric head at boundary 2
hp [m] = head added by a pump
ht [m] = head extracted by a turbine
hL [m] = head loss
u1 [m/s] = average flow velocity through boundary 1
u2 [m/s] = average flow velocity through boundary 2

63
Manual Hydraulic Structures Part II - 'Loads'

The energy equation for pipe flow is shown graphically in Figure 10-10-4. The energy at both boundaries
is equal, because of the absence of pumps and turbines, and the neglect of the head losses. However,
the composition of the total energy head changes: From the continuity equation, it follows that the velocity
increases if the cross-sectional area decreases. The velocity head therefore increases and the
piezometric head decreases, resulting in the same energy head at both boundaries.

Figure 10-10-4 Energy balance for pipe flow (Zitman, 2017)

For the situation of Figure 10-10-4, head loss can be neglected, since turbulence is absent in the control
volume and the control volume has a small length for which friction is negligible. Because there are no
pumps and turbines, also these terms can be left out of the energy equation. One ends up with an
equation similar to Bernoulli’s equation:

2 2
u u
h1 + 1 = h2 + 2
2g 2g

This equation is still derived by using a control volume, therefore the average velocity through the
boundaries has to be used. If the continuity equation is used to calculate the flow velocities, these can be
used in the equation above. The Bernoulli equation is very similar, but it is derived by equating the energy
terms at two points on a streamline: In this case, the flow velocities in these two points are filled in in the
equation.

If turbulence is generated inside the control volume, Bernoulli’s equation is not valid: a term for head loss
has to be added. Head losses are associated with fast variations of flow patterns, for example at hydraulic
jumps and overflow weirs. Widening, bifurcations and outflow of pipes into still water bodies, see also
Figure 10-10-1, are called minor losses. Generally, these energy losses are expressed as a part of the
velocity head:
u2
∆H = K L
2g
in which: ∆H [m] = head loss
KL [-] = loss coefficient
u [m/s] = flow velocity at the structure

The head losses of all components have to be summarized, each calculated with its own loss coefficient
and flow velocity. For several components the loss coefficient is constant, like for the exit (KL=1,0), but for
others it is very dependent on the configuration. This can be seen in Figure 10-10-5for the entrance
losses.

For pipe flow components as contraction, expansion and bends are resulting in extra minor losses. These
are not treated here, the loss coefficients can be found in other literature.

For long sections the major losses, caused by the friction along the riverbed and walls, comes into play.
To determine the resulting head loss the well-known Darcy-Weisbach equation is used, comparable to
the equation for minor losses:

64
Manual Hydraulic Structures Part II - 'Loads'

L u2
∆H f = f
D 2g
in which: ∆Hf [m] = head loss due to friction
f [-] = friction factor
L [m] = length of section
D [m] = diameter

Figure 10-10-5 Entrance losses (Vano Engineering, 2012)

This equation should only be used for steady flows through conduits. The friction factor is depending on
the Reynolds number and/or the relative pipe roughness. The Moody Diagram (see Figure 10-10-6) can
be used to determine the friction factor.

Figure 10-10-6 Moody diagram (STEM, 2017)

10.3 Flow in open water ways

Water flows if there is a difference of piezometric head between two points. The flow velocity depends on
the extent of the difference in piezometric head and the resistance encountered by the water. Friction and
sudden changes of the flow profile cause the resistance encountered by the flowing water. The friction in
open water ways can usually not be neglected because they commonly consist of long sections.

65
Manual Hydraulic Structures Part II - 'Loads'

10.3.1 Friction
Friction between water particles causes a logarithmical flow velocity change over the water depth. The
average velocity can be calculated by using the traditional methods: the Chézy equation and the Manning
equation. These equations relate the amount of bed and bank friction directly to the mean flow velocity.
They are only valid for steady and uniform conditions.

Chézy’s equation is:

12R
um = C R i b with C = 5,75 ⋅ g ⋅ log
k

where: um [m/s] = mean flow velocity

C [ m1/2 /s ] = Chézy’s coefficient
Bd
R [m] = hydraulic radius of the flow profile ≈ ≈d (for: B >> d )
B + 2d
d [m] = depth of flow, i.e. water level
B [m] = flow width
ib [-] = the slope of the river bed
k [m] = roughness of the edges of the flow profile

Note. For beds with sediment transport: k≈ 0,05 - 0,10 m. On other cases k = 2 ⋅ D90 can be used. (D90 is
the diameter of the sieve through which 90 percent of the grains pass).

The disadvantage of the approximation of the flow velocity according to Chézy is that C isn’t a constant.
Therefore, internationally, Manning’s equation is given preference.

Manning’s equation is:

1 2/3
ugem = R i
n

in which: n [-] = Manning’s coefficient, dependent on the roughness of the riverbed [-].

This coefficient can easily be measured by determining the discharge Q, flow section A, slope ib and
water depth d. Henderson (1966) found the values given below:

Glass, smooth metal n = 0,010

concrete made with wooden formwork (untreated) n = 0,014
smooth earth, bare (no vegetation) n = 0,020
river: clear and straight n = 0,025 – 0,030
river: meandering and with shallows/ depths n = 0,033 – 0,040
river: overgrown, meandering n = 0,075 – 0,150
Table 10-1 Manning's coefficient (in m-1/3 s )

Chézy can be converted to Manning and vice versa, because:

1 2/3 1
C Ri = R i ⇒ C = R1/ 6
n n

Notice that with increasing roughness of the bed, n, Chézy’s coefficient decreases. Chézy’s coefficient is
therefore actually a smoothness coefficient. By approximating, among other things, the log-term in the
Chézy equation, the riverbed roughness parameters k and n can be converted into each other:

n −1 ≃ 8,3 ⋅ g ⋅ k −1/ 6

Both equations can only be used for straight, prismatic channels. Bends and structural elements in the
section result in more head loss, so the Chézy and Manning equation cannot be used in these cases,
unless the coefficients include these local losses.

66
Manual Hydraulic Structures Part II - 'Loads'

The resulting flow depth in the uniform, steady conditions is called the normal depth de. It can be found by
using the Chézy equation and the continuity equation (this expression is also known as the Q-h relation,
where h = de):
1/3
 Q2 
de =  2 2 
 C B ib 

10.3.2 Backwater curves

The flow depth isn’t always equal to the normal depth. Spatial variations in friction and width and
regulating structures in the river cause so-called quasi-uniform flow conditions. The effects of the
variations and structures is noticeable over a large distance. From a spatial perspective, the flow adapts
slowly to these conditions. The way of adaptation can be derived with help of the energy balance.

If the control volume includes a long river section, friction cannot be neglected. According to Figure 10-
10-7, the slope of the total energy head H in downstream direction s is equal to iw. For uniform conditions,
the slope of the energy line is the same as the slope of the river bottom ib, for non-uniform conditions it
isn’t.

Figure 10-10-7 Energy balance for long sections (Blom, 2016)

For the derivation, the specific energy E is used:

u2
E =d+
2g
The gradient of the specific energy head is:

dE d Q2
= (d + ) = i b − iw
ds ds 2gA2

If the discharge and width are constant, this reduces to the backwater equation:

dd i −i
= b w2
ds 1 − Fr
u
in which: Fr [-] = Froude number (= )
gd
gFr 2
iw [-] = friction slope (= )
C2

The Froude number correlates the kinetic force to the gravitational force. Due to the small slope of river
beds in the Netherlands, the flow depth (and gravitation force) is large and therefore, the flow (Fr ≤ 1,0) is
most commonly subcritical. The velocity of subcritical flow is smaller than the wave celerity. Therefore,
the flow conditions in a river are determined by the boundary conditions downstream.

Using the formula for the friction slope, the backwater equation can be written in a different form:

67
Manual Hydraulic Structures Part II - 'Loads'

dd d 3 − dn 3
= ib 3
ds d − dg 3
13
 2 
in which: dg [m] = critical depth =  q 
 g

At the critical depth, the Froude number is equal to 1,0. For flow depths very close to the critical depth,
the backwater equation no longer holds.

As mentioned before, subcritical flow conditions are most common in the Netherlands. The normal flow
depth is larger than the critical depth, the flow depths are determined by downstream boundary
conditions. Solving the backwater equation reveals that the flow depth approaches the normal flow depth
exponentially with increasing distance from the boundary condition. Various types of backwater curves for
subcritical flow/ mild slopes are presented in Figure 10-10-8.

Figure 10-10-8 Theoretical types of backwater curves (Blom, 2016)

The water depth gradually approaches the normal depth more upstream of the boundary condition. The
water depths in the influence area of the boundary condition can be calculated with the following formula:

( s − s0 )
∆d (s ) = ∆d (0) ⋅ e L

in which: ∆d(s) [m] = water depth difference with normal depth at point s.
∆d(0) [m] = water depth difference with normal depth at point s=0.
s-s0 [m] = river length between boundary condition and point of interest.
d
L [m] = adaptation length ≅ e
3i b
de [m] = normal depth

Note. If the point of interest is more upstream than the boundary condition, the term s-s0 is negative.

10.4 Conservation of momentum

Similar to the derivation of the mass and energy balance (see Sections 10.1 and 10.2), the derivation of
the momentum balance (impulsbalans) starts with a control volume. The conserved quantity is
momentum, the multiplication of mass and velocity. External forces can cause change of momentum,
which has been introduced by Newton:
d (mu )
∑ Fext = dt
in which: ΣFext [N] = sum of external forces
m [kg] = mass

68
Manual Hydraulic Structures Part II - 'Loads'

u [m/s] = velocity
Fluid particles contain mass and velocity, so the flow transports momentum. Momentum at the upstream
boundary is transported inwards, at the downstream boundary outwards. Using the discharge instead of
the mass, the momentum balance can be transformed into:

∑F ext = ρQ ( uin − uout )

This simplification is only applicable for steady conditions, since the discharge is taken constant in time.

The potential external forces are very divergent. However, the pressure force at both boundaries is
always part of the balance. Consequently, it is important to make a valid choice for the location of the
boundaries. In case of parallel, straight flow lines at the boundary, the pressure force can be calculated
as:
Fh,in = 1 ρ gBhin 2 and Fh,out = 1 ρ gBhout 2
2 2

Where Fh,in stands for the hydrostatic force at the inflowing boundary, Fh,out for the one at the outflowing
boundary.

10.5 Literature

Battjes, J.A., & Labeur, R.J. (2014). Open Channel Flow. Delft: Lecture notes Delft University of
Technology.

Bezuyen, K.G., Stive, M.J.F., Vaes, G.J., Vrijling, J.K., & Zitman, T.J. (2012). Inleiding Waterbouwkunde.
Delft: VSSD, Lecture notes Delft University of Technology.

Blom, A. (2016). River Engineering. Topic 3 - Steady flow II: Backwater curves. Lecture notes Delft
University of Technology.

Elger, D.F., Williams, B.C., Crowe, C.T., & Roberson, J.A. (2014). Engineering Fluid Mechanics. Delft:
VSSD.

Engsoft Inc. (n.d.). ES_CriFlow - Compressible Flow Analysis Engine of Steam, Rev. 1. Retrieved from
http://www.engsoft.co.kr/download_e/steam_flow_e.htm#TOC

Hendrayana, H. (2012). Introduction to groundwater modelling. Yogyakarta: Lecture notes Gadjah Mada
University.

STEM. (2017, August 10). How to Read a Moody Chart (Moody Diagram). Retrieved from Owlcation:
https://owlcation.com/stem/How-To-Read-a-Moody-Chart

Vano Engineering. (2012, December 30). Head loss coefficients. Retrieved from Vano Engineering:
https://vanoengineering.wordpress.com/2012/12/30/head-loss-coefficients/

Zitman, T.J. (2017). Stationaire stroming. Delft: Lecture notes Delft University of Technology.

69
Manual Hydraulic Structures Part II - 'Loads'

11. Water, flow, wide structures

revised: 2019

There are two ways of determining the load caused by a flow against wide structures (walls):
1. Making use of the principle of conservation of momentum (impulsbalans);
2. Calculating the complete flow pattern along the wall, so that the pressure on the wall is known over
the entire wall.

The advantage of the conservation of momentum is that it’s easier much faster to elaborate. The
disadvantage is that only the total force is calculated and not the distribution of pressure on the wall. To
derive the latter, it is necessary to solve the entire flow pattern.

11.1 Flow across wide structures

Civil engineering works such as weirs, piers and sills, can influence the water level in a river. The extent
of the influence is described by the discharge function of the structure. Hydraulic structures are
constructed to change the flow conditions. For example, weirs provide sufficient water depth for ships
during low river discharges. This effect is only noticeable in the river section near the weir due to the
development of backwater curves. In Figure 11-1,the appearance of the theoretical curves around weirs
are shown.

Figure 11-1 Backwater curves due to weirs (Bengston, 2012)

Two possible types of flow are distinguished for weirs, namely:

• overflow: the upstream water level is can be well controlled.
• underflow: in case of submerged underflow, the discharge through the structure is well-controlled.

The flow type depends on the situation and the desired result. Both types are explained in the following
sections.

11.1.1 Overflow weirs

The overflow weirs can be divided into two categories: the broad-crested and the sharp-crested weirs.
Weirs are classified as broad-crested if the streamlines above the weir itself are straight and parallel.
Examples of broad-crested weirs are the sills of storm surge barriers. The streamlines on top of sharp-
crested weirs are curved, which results in a different discharge function. Figure 11-2 can be used to
determine the category of a particular structure.

70
Manual Hydraulic Structures Part II - 'Loads'

Figure 11-2 Broad-crested versus sharp-crested weirs (Ankum, 2002)

Broad-crested weirs
The discharge over a broad-crested weir mainly depends on the water levels upstream and downstream
of the weir. Obviously, if there’s no difference, there is no flow. If the downstream water level is slightly
lower, the conditions occur as indicated in Figure 11-3, where water levels and energy heads are
indicated. The top of the weir is chosen as reference level.

Figure 11-3 Flow conditions for broad-crested weir, submerged overflow (Bezuyen, Stive, Vaes, Vrijling, & Zitman, 2012)

Application of Bernoulli’s equation between points 1 and 2 results in:

u2 = 2g (H1 − h2 )

The energy head H1 is difficult to measure. The use of water level h1 is preferred, which is accompanied
by the introduction of a discharge coefficient m1. Without this coefficient the calculated flow velocity would
be too small. Furthermore, it is desired to relate the discharge over the weir to the water level difference
over the weir, so h3 is used instead of h2. The discharge coefficient is adapted to include the energy
difference between point 1 and 3. The final result is:

Q = msof ⋅ B ⋅ h 3 ⋅ 2g (h 1 − h 3 )

in which: msof [-] = discharge coefficient for submerged overflow ≈ 1,1

B [m] = flow width

The formula above is only applicable when subcritical flow occurs above the weir. Similar to the mild
slope conditions of the backwater curves, subcritical flow means that the upstream flow conditions
determine the downstream boundary condition (in this case the water level h3). By decrease of the
downstream water level, the water depth on top of the sill decreases. Consequently the Froude number
and discharge over the structure increase, until critical flow occurs (Fr = 1). Figure 11-4 shows that, in this
flow situation, the discharge is maximum for a given energy level.

71
Manual Hydraulic Structures Part II - 'Loads'

Figure 11-4 Discharge as a function of the water depth for a constant energy level (Nakayama & Boucher, 2000)

If the downstream water level decreases even more, the discharge over the weir doesn’t increase further,
because it is already maximum. The resulting changed flow condition can be seen in Figure 11-5. The
flow situation has changed from submerged to free flow, with a different discharge function.

Figure 11-5 Flow conditions for broad crested-weir, free overflow (Bezuyen, Stive, Vaes, Vrijling, & Zitman, 2012)

Again, Bernoulli’s equation is used to obtain the discharge formula. assuming that critical conditions occur
above the weir.

u2 2 u2
H1 = h2 + and Fr = = 1,0
2g gh2

Combining these equations results in the discharge formula for free overflow of broad-crested weirs.

2 2 3
Q= ⋅ g ⋅ mfof ⋅ B ⋅ h1 2
3 3

in which: mfof [-] = discharge coefficient for free overflow ≈ 1,0

The discharge coefficient considers the neglect of friction and the use of water level instead of energy
level. The most important difference with the submerged overflow is that for free overflow only the
upstream water level has influence. The downstream water level doesn’t have any effect.

Sharp-crested weirs
Most weirs in the Netherlands are classified as sharp-crested weirs. These weirs are used to control the
water levels in the polders and the rivers. Examples of these weirs are shown in Figure 11-6.

72
Manual Hydraulic Structures Part II - 'Loads'

Figure 11-6 Sharp-crested weirs in the river Maas (left, Planthof, 2018) and a polder (right, Pannekoek GWW B.V., n.d.)

The crest level of a sharp-crested weir is adapted in order that always a free flow situation is created. The
flow situation for sharp-crested weirs is comparable with the free overflow situation for broad-crested
weirs. The discharge formula for free flow can also be used for sharp-crested weirs albeit with a different
discharge coefficient

Figure 11-7 Flow conditions for sharp-crested weirs (Bezuyen, Stive, Vaes, Vrijling, & Zitman, 2012)

2 2 3
Q= ⋅ g ⋅ msc ⋅ B ⋅ h1 2
3 3

in which: msc [-] = discharge coefficient for sharp-crested weirs

Due to the curved streamlines above the weir the flow velocity and discharge is larger than for broad-
crested weirs. Therefore, the discharge coefficient for sharp-crested weirs is larger than 1,0. The radius of
the crest R, the height of the crest p1 and the upstream water level h1 determine the discharge coefficient.
Spillways of large dams make use of so-called ogee crests, which have a profile which maximizes the
discharge coefficient (msc ≈ 1,35).

By experimenting, Kindsvater and Carter developed a more accurate formula than the standard weir
equation. It only holds for rectangular sharp-crested weirs (ISO, 1980):

2
Q= ⋅ (B + K B ) ⋅ 2g ⋅ (h1 + K h )3 2 ⋅ Cd
3

in which: B + KB [m] = effective width

h1 + Kh [m] = effective head
KB [m] = coefficient to account for the viscosity (see Figure 11-8)
Kh [m] = coefficient to account for the surface tension = 0,001 m
Cd [-] = discharge coefficient (see Figure 11-9)

73
Manual Hydraulic Structures Part II - 'Loads'

Figure 11-8 Determination coefficient KB (LMNO Engineering)

Figure 11-9 Discharge coefficient Kindsvater-Carter equation (ISO 1980)

In case of small discharge, often the a triangular V-notch weir is applied. For example, in a fish trap all
water flows through the v-shaped part and provides sufficient water depth for the migrating fish.
Furthermore, discharge measurements are carried out often, because the discharge coefficient of V-notch
weirs is known very accurate.

Figure 11-10 V-notch weir

The width, varying with the water depth, can be described with (see Figure 11-10):

θ 
B = 2(H − z ) ⋅ tan  
2

The general equation for the theoretical discharge over a weir is applied to derive the discharge relation
for a V-notch weir:

H
Qtheory = ∫ (u ⋅ B ) dz
z =0

By substituting u (critical flow conditions) and B, integrating and assuming that far upstream the velocity
head is negligible to the energy head, the discharge over the V-notch can be described as:

8 θ 
Qtheory = ⋅ 2 g ⋅ tan   ⋅ h5 2
15 2

74
Manual Hydraulic Structures Part II - 'Loads'

Also for V-notched weirs, losses occur due to the edges of the weir and contractions in the area of flow.

Q actual = Q theory ⋅ Cd

The value of discharge coefficient Cd for a specifically shaped weir can be found experimentally. From
dimensional analysis and experiments, the average value for V-notch weirs appeared to be 0,6.

11.1.2 Underflow weirs

Underflow weirs control the discharge by adjustment of the flow profile by tilting a gate vertically. In the
Netherlands, several weirs are underflow weirs, like the one at Driel (see Figure 11-11). During opening
of lifting gates of sluices the same flow patterns occurs, although the water level differences are small
during these procedures.

Figure 11-11 Underflow weir at Driel (Rijkswaterstaat, 2018)

The flow conditions of underflow are separated into two categories: free flow and submerged flow.

Free underflow
The flow conditions for free flow are shown in Figure 11-12. Directly downstream of the weir, the flow is
supercritical. More downstream a hydraulic jump (according to an M3 backwater curve) is located, which
indicates the transition to subcritical flow.

Figure 11-12 Flow conditions for free underflow (Bezuyen, Stive, Vaes, Vrijling, & Zitman, 2012)

The discharge relation is obtained by applying Bernoulli’s equation between point 1 and 2. Loss of energy
is only significant more downstream of point 2. This results in the following:

75
Manual Hydraulic Structures Part II - 'Loads'

Q = µ ⋅ B ⋅ a ⋅ 2g (H1 − µ a )

in which: µ [-] = contraction coefficient ≈ 0,6

a [m] = height of flow outlet

However, the contraction coefficient is dependent on the shape of the edge of the weir and cannot be
calculated precisely. Besides that, the water level h1 is rather used than the energy level H1. Again, a
discharge coefficient is set, leading to the discharge equation for free underflow conditions:

Q = mfuf ⋅ B ⋅ a ⋅ 2g (h1 − a )

in which: mfuf [-] = discharge coefficient for free underflow

This discharge coefficient is variable, since it is dependent on the contraction coefficient, the upstream
water level and the height of the flow outlet.

Free underflow only occurs when the downstream water level is low. Flow conditions will change from
free to submerged underflow when the downstream water level rises. For preliminary design, it can be
assumed that free underflow will occur when the following applies:

h3 h
≤ 1,42 ⋅ 1 − 0,3
a a

Submerged underflow
In submerged flow conditions the hydraulic jump is located directly downstream of the weir, like in Figure
11-13. This is caused by a higher downstream water level, which is on its own caused by a downstream
boundary condition.

Figure 11-13 Flow conditions for submerged underflow (Bezuyen, Stive, Vaes, Vrijling, & Zitman, 2012)

In this flow situation, energy is dissipated directly downstream of the weir. Therefore, preferably the
downstream water level h3 is linked to the discharge relation than h2. Consequently, the discharge
coefficient for submerged underflow should take into account the considerably amount of energy loss.
The resulting formula is:

Q = msuf ⋅ b ⋅ a ⋅ 2g (h1 − h 3 )

in which: msuf [-] = discharge coefficient for submerged underflow ≈ 0,8

Like all other discharge coefficients, the presented value is only useful for preliminary design. For more
detailed design, experiments should give more accurate discharge coefficients.

76
Manual Hydraulic Structures Part II - 'Loads'

11.2 Forces on wide structures

The forces on wide structures can be calculated if the flow conditions around the structure are known.
Often these conditions are calculated with help of the mass and/or energy equation. Subsequently, the
force can be calculated, since it is the only unknown in the momentum equation.

In Figure 11-14 the significant forces for a momentum balance of an underflow weir are shown. For a
prismatic channel and steady conditions, the specific discharge q is the same at boundary 1 and 2.

Figure 11-14 Flow around sluice gate in canal (Bezuyen, Stive, Vaes, Vrijling, & Zitman, 2012)

Drawing up the momentum balance for this situation per unit width results in the following force on the
gate Fs:

F1 − F2 − Fs = ρ q (h1 − h2 )

Fs = 1 ρ gh22 + ρ u2 h22 − 1 ρ gh12 − ρ u1h12

2 2

By vertical movement of the gate the flow conditions change and so does the force on the gate. It is good
to realize that the largest static force usually originates at a closed gate. In that case, the force is simply:

(
Fs = 1 ρ g H12 − H 22
2 )
For other weirs the flow conditions are different, but the procedure is quite the same, so these are not
addressed here. In the same way the forces on a culvert can be calculated.

The flow under the gate will cause resonance of the gate (with tidal movements, two directions of flow are
involved!). The determination of the size and frequencies of dynamic forces is beyond the scope of this
book.

11.3 Potential flow and pressure distribution

The pressure distribution on a structure in the flow, and thus also on the point of application, cannot be
determined using the momentum balance. To do this, the potential flow theory may be used. However,
one condition for the use of this theory is that the flow is free of rotations, which means that shear
stresses are to be ignored. The potential flow theory will not be explained here, one is referred to the
lecture notes “Open Channel Flow” (Stroming in open waterlopen), CTB3350 (Battjes & Labeur, 2014).
An important property of a two-dimensional potential flow is that the flow lines and the equipotential lines
are perpendicular to each other.

By drawing lines of equal potential at a distance ∆φ of each other and flow lines at a distance ∆ψ of each
other, ∆φ being equal to ∆ψ, one creates a flow net consisting of squares. Such a net can be drawn
iteratively by using the above-mentioned properties. The diagonals of the flow net in turn form another
square flow net with the same properties. This characteristic can be used as a visual test of the
constructed flow net. The following example serves as a clarification.

77
Manual Hydraulic Structures Part II - 'Loads'

11.4 Literature
Ankum, P. (2002). Design of Open-Channels and Hydraulic Structures. Delft: Lecture notes Delft
University of Technology.

Bengston, H. (2012, December 15). Backwater Curve Calculations Spreadsheet. Retrieved from
Engineering Excel Spreadsheets:
https://www.engineeringexcelspreadsheets.com/2012/12/backwater-curve-calculations-spreadsheet/

Bezuyen, K. G., Stive, M. J., Vaes, G. J., Vrijling, J. K., & Zitman, T. J. (2012). Inleiding Waterbouwkunde.
Delft: VSSD.

ISO (1980). International Organization of Standards. ISO 1438/1-1980(E). Water flow measurement in
open channels using weirs and venturi flumes - Part 1: Thin plate weirs. 1980. Available from Global
Engineering Documents at http://global.ihs.com

Kindsvater, C.E. and R.W. Carter. 1959. Discharge characteristics of rectangular thin-plate weirs.
Transactions, American Society of Civil Engineers. v. 24. Paper No. 3001.

Nakayama, Y., & Boucher, R. (2000). Introduction to Fluid Mechanics. Oxford: Butterworth-Heinemann.

Pannekoek GWW B.V. (n.d.). Foto's: Stuw 2 Burggenhoek. Retrieved from Pannekoek GWW B.V.:
https://www.pannekoekgww.nl/projecten/detail/fotos-stuw-2-bruggenhoek

Planthof, S. (2018). Stuw Borgharen is open tijdens Monumentendag. De Limburger, 4 September 2018 .

Rijkswaterstaat. (2018). 'Gelijk water' bij stuw Driel. De Binnenvaartkrant, 18 Octoober 2018.

78
Manual Hydraulic Structures Part II - 'Loads'

12. Water, flow, slender structures

12.1 Drag and lift forces

An object, entirely in a uniform flow field, is subject to forces both in the direction of flow (drag force) and
perpendicular to the flow (lift force). Drag and lift forces fluctuate in time. These forces cannot be
determined theoretically with a linear flow theory such as Bernoulli’s. The reason why the flow cannot be
schematised as a potential flow is because of the occurrence of rotation: eddies behind the cylinder. A
wake with a lower pressure is created directly behind the cylinder, which causes a resultant force in the
direction of the wake. This turbulence also causes oscillations.

Experiments have shown that the resultant force on an object in the flow F [N] is reasonably proportional
to the velocity head times the density of the fluid (compare this with Bernoulli’s equation):

u2 1
F∼ρg = ρ u2
2g 2
3
in which: ρ [kg/m ] = density of water
u [m/s] = de undisturbed flow velocity

The empirical formulas for drag and lift are:

1 2 1 2
FD = ρ u (CD + CD' )A and FL = ρu (CL + CL' )A
2 2

in which: FD [N] = drag force parallel to the flow direction

FL [N] = lift force perpendicular to the flow direction
CD [-] = drag coefficient (static)
CL [-] = lift coefficient (static)
CD' [-] = dynamic drag coefficient (time dependent)
CL' [-] = dynamic lift coefficient
2
A [m ] = area facing flow, projected perpendicular to the flow direction

The fluctuation in time of the shapes of the wake and the eddies causes a fluctuation of the size and
direction of the force, which, in turn, can cause oscillations. The drag force and lift force therefore consist
of a static and a dynamic part. The corresponding coefficients are treated separately below.

Notes
• The drag force is parallel to the direction of flow (and not to the axis of the object!), the lift force is
perpendicular to this.
• It is very important to notice that the turbulence of the water can wash out the soil behind a pile, thereby
endangering the founding property of the pile. This erosion of the bed is called scour and is not discussed any
further. When dimensioning a post, one must take this scour into account (less fixation and more effective pile
length) or include bed protection measures. The scour behind a pile is 1.5 - 2 times the pile diameter in sandy
soils.

12.2 Drag and lift forces, static part

The coefficients CD and CL are largely dependent on the shape of the structure and the flow around the
structure, which are expressed in the Reynolds value, Re [-]:

ud
Re =
ν

in which: Re [-] = Reynolds value

u [m/s] = flow velocity
-6
ν [m/s] = kinematic viscosity = 10
d [m] = measure of length dependent on geometry

79
Manual Hydraulic Structures Part II - 'Loads'

4
In the field of hydraulic engineering, the flow is usually turbulent (Re > 10 ). The CD coefficients for three
different shapes (cylinder, plate and a streamlined wing) are given in
5 6
Figure 12-1. This figure shows that the CD coefficient of a cylinder (pile) in the area 10 < Re < 10 is
sensitive to fluctuations of Re. However, the CD coefficients of angular objects, such as plates and
beams, are barely dependent on Re in this turbulence area.
5 6
In the region 10 < Re < 10 , the wake behind a cylinder decreases for a larger Reynolds value, the
separation points move further back. A smaller wake means a smaller resultant force. For objects with
sharp edges, like flat pates and square piles, the separation points of the flow and thus the wake are
more or less fixed. The resultant force on such objects is therefore less sensitive to fluctuations of Re.

Figure 12-1 shows that the shape of a pile is very important for the forces that result from flow against the
pile.

Figure 12-1 CD value of a cylinder, plate and streamlined wing

The static lift coefficient (CL) is zero for straight approaching flow.

The CD and CL coefficients of a bridge piers approached by flow at an angle are given in Figure 12-2.
Three different shapes for the pier are indicated. The given coefficients are largest for the cylinder.
However, this does not mean that the force on the cylinder is actually the largest, because the projected
surface perpendicular to the flow is smaller for a cylinder than for the other given piers. The more the
shape of the pier deviates from a cylinder, the larger the force on the pile if the flow approaches at an
angle.

80
Manual Hydraulic Structures Part II - 'Loads'

Figure 12-2 CD and CL coefficients of various pile shapes in the case of flow at an angle (Apelt and Isaacs, 1986)

81
Manual Hydraulic Structures Part II - 'Loads'

12.3 Drag and lift forces, dynamic part (vibrations)

A structure that has been placed in a stationary flow can start to vibrate. These lecture notes only include
an introduction to the phenomenon of vibrations as a result of stationary flow, as the complexity of the
subject doesn’t allow for a full description here.

The most important vibrations in a structure in a stationary flow are caused by instability of the wake and
the vortices behind and alongside the structure or by hydrodynamic instability. Due to fluctuations in time
of the vortices and the wake, the size and direction of the resultant force on an object in a stationary flow
also fluctuates in time. Figure 12-3 indicates the pressure distribution around a cylinder due to changing
vortices. This figure clearly shows that there is a dynamic resultant load.

Figure 12-3 Fluctuation in time of the pressure and the wake with Re=1.12 105 (Drescher 1956)

The frequency fs of vortex shedding behind an object standing in the flow is given by:
uS
fs =
D
in which: u [m/s] = the flow velocity in an undisturbed flow
S [-] = Strouhal number
D [m] = the characteristic diameter perpendicular to the flow

The Strouhal number (S), named after the Czech physicist Vincenc Strouhal, is a dimensionless number
describing oscillating flow mechanisms. For cylinders, this number is given as a function of the Reynolds
number (Re) in Figure 12-4. For shapes with non-circular cross sections, reference is made to Figure 12-
5.

Hydrodynamic instability occurs if a slight movement of an object is enforced by the flow, so that a small
displacement can lead to vibrations with increasing amplitude. Depending on the stiffness of the structure
hydrodynamic instability may occur. If fs > fn , no resonance will occur; here fn is the natural oscillation
frequency of the system (structure + water). This dynamic phenomenon is identical to a chimney swaying
in the wind. Calculating the natural oscillation frequency is a study in its own right and is not covered in
this course.

82
Manual Hydraulic Structures Part II - 'Loads'

Re

Figure 12-4 Strouhal number for cylinders

Re

Figure 12-5 Strouhal number for non-circular cross sections

For the maximum dynamic drag coefficient (time dependent):

CD' = 0,10 to 0,5 ⋅ CD if: fs > fn

C = ? (danger of resonance)
'
D if: fs < fn

The static CL coefficient is of course zero for a straight approaching flow (α = 0). The dynamic CL'
coefficient is not zero due to the time dependent turbulent oscillations. The value of CL' is strongly
fs f
dependent on the ratio and increases largely near s = 1 . Still no reasonably accurate expression is
fn fn
available for the value CL' . For a preliminary design, one may therefore assume:

CL' ≈ CD' .

83
Manual Hydraulic Structures Part II - 'Loads'

13. Water, tide and wind set-up

Extended and updated: February 2015

Water levels along the coast are mainly determined by the tide and by wind set-up. Furthermore,
phenomena such as storm surges, shower oscillations, shower gusts and seiches are involved. These
effects are explained below. Chapter 48 explains how they contribute to the determination of the height of
flood defences.

13.1 Astronomical tide

13.1.1 Theory
One of the most characteristic properties of coastal waters is the tidal movement. Different tides can be
distinguished according to the water motion direction, namely:
• vertical tides: raising and falling of the water level
• horizontal tides: tidal flows in tidal inlets/ outlets and creeks and along the coast.

The tidal movements on earth are caused mainly by the following 5 factors:
1. The gravitational pull of the moon on the seawater.
2. The monthly rotation of earth around the shared earth-moon axis (centrifugal force).
3. The daily rotation of earth around its own axis.
4. The inclination of the earth’s rotation axis.
5. The gravitational pull of the sun on the seawater.

1. The gravitational pull of the moon on seawater

The moon, and in lesser extent the sun, cause an uneven force distribution on the water due to the
difference in distances between the points on earth and the moon, see figure below. The result is a semi-
diurnal tide.

Figure 13-1 Pull of the moon and the sun

2. Rotation around the shared earth-moon axis

The earth has a monthly rotation around the shared earth-moon axis. This axis is approximately 0,6 r
from the centre of earth. Because every point on earth covers the same distance for this rotation, the
centrifugal force is the same everywhere on earth. (In this case one must not consider that the earth is
rotating around its own axis, as this is taken into consideration in point 3.) Together these two forces form
an elliptical force distribution around the earth. If the earth were covered entirely in water, the water mass
would take the shape of an ellipsoid, as shown in Figure 13-2. This fictional situation is known as the
equilibrium tide due to the moon.

3. Rotation around own axis

The earth spins as it were under the water mass, thus causing two high tides and two low tides a day
(semi diurnal tide) in an arbitrary location. There are, however, places on earth where the tidal flow is
influenced by the presence of a continent to such an extent that there is only one high tide per day
(diurnal tide) in these locations.

84
Manual Hydraulic Structures Part II - 'Loads'

4. Obliqueness rotation axis

Due to the obliqueness of the rotational axis of the earth, the tidal movement during a semi diurnal tide is
AB one time and A’B’ another. This causes a daily inequality of the maximum tide (see Figure 13-3). In
winter, the high tide at daytime (AB) is generally smaller than the high tide at night (A’B’). In the summer,
the opposite applies.

Figure 13-2 Two different high tides

5. Gravitational pull of the sun on seawater

The influence of the sun is confined to reinforcing or weakening the moon tide. If the sun and the moon
are on the same side of the earth, they reinforce each other. The resulting tide is known as spring tide, if
they counteract each other it is called neap tide. Spring tide occurs at full and new moons; neap tide
occurs in the first and last quarters.

13.1.2 (Preliminary) design

In reality the astronomical tide bears little resemblance to the theoretical equilibrium tide. This is due to
the presence of continents and the Coriolis effect, which causes the diversion of tidal flows. The true tidal
water levels and the times of high and low tide can be predicted with a “harmonic analysis” or “Fourier
analysis” of the data from measuring stations. These techniques are not discussed here.

Every year, the National Institute for Public Works and Water Management, Rijkswaterstaat, publishes
predictions made using these methods in the booklet Tide tables for the Netherlands ("Getijtafels voor
Nederland"). For a number of measuring stations, the booklet lists when high and low tides will occur at
and what their corresponding water levels are.

Figure 13-3 Water level in one month in Vlissingen

85
Manual Hydraulic Structures Part II - 'Loads'

For hydraulic engineering structures along the coast, the water level during an extreme storm surge is
usually particularly important. See Section 13.3 "Storm surge" for tables with storm surge water levels.

Data on tides can be found at www.rijkswaterstaat.nl/geotool/waterhoogte_tov_nap.aspx.

13.2 Wind set-up

In shallow seas, deltas, closed off creeks, and lakes, wind fields can influence the water level quite
considerably by heading up the water (wind set-up). Figure 13-4 shows a model to approximate the wind
set-up (opwaaiing).

Figure 13-4 Balance of forces in case of wind set-up

The wind set-up in the equilibrium state is approximately:

dS u2
= C2
dx gd

in which: S [m] = total wind set-up

constant ≈ 3,5⋅10 to 4,0⋅10
-6 -6
C2 [-] =
d [m] = water depth
u [m/s] = wind velocity

dS
If the wind set-up is small compared with the water depth, in an area with a horizontal bed, the slope
dx
is constant. The formula shows that the wind set-up increases with increasing wind velocity and fetch and
decreasing water depth. The wind set-up is therefore of importance in river deltas, lakes and shallow
seas. In coastal areas where the sea is deep, wind set-up hardly ever occurs. In the North Sea, the
Wadden Sea and the IJsselmeer, the set-up can be as much as a couple of meters. In 1953 the rise in
Vlissingen was 3,05 m.

If the water level at the edge of a basin is known, the course of the water level in the basin can be
calculated with a simple numerical solution (the 'Heun method'):

dS
d x = d x −∆x + ∆x − ∆zx −∆x ,x
dx

in which: dx [m] = the water depth in point x

dx-∆x [m] = the water depth in point x-∆x
∆zx-∆x,x [m] = difference between the height of the bed in x and x-∆x

In a closed basin or a lake, the total amount of water cannot change. This means that, provided the slope
may be assumed constant, the surface of the water (by approximation) will tilt around the gravity line of
the basin surface, perpendicular to the wind direction. The water in the area between the down-wind edge
and the gravity centre is subjected to wind set-down (see Figure 13-5).

86
Manual Hydraulic Structures Part II - 'Loads'

Figure 13-5 Wind set-up in a closed basin

The combination of tides and wind set-up is important, particularly for the design of most hydraulic
engineering structures. If high tide coincides with strong winds (storm) a storm surge occurs. See the
following section for tables with water levels in storm surges.

13.3 Storm surge

It is more or less possible to predict the timing and the water levels of a spring tide. The wind set-up near
the coast line can be calculated with a given wind velocity. By analysing wind data, it is possible to
calculate the probability of occurrence of a spring tide and a certain wind set-up. Such a combination is
known as a storm surge (stormvloed).

Because the calculation model for wind set-up contains a number of uncertainties, a different approach is
used in practice. The Dutch Delta Committee based the determination of basic water levels upon as many
observed storm surge levels as possible. In this way, it constructed a frequency distribution (an empirical
relation). The normative water level (storm surge level) for a chosen small exceedance frequency was
then determined by extrapolation of measuring points, see the high water level measurements for Hoek
1
van Holland in the following figure (measured at a gauge (peilschaal) near the coast).

Figure 13-6 Water level exceedance frequencies at Hoek van Holland

1
Notice that x- and y-axis are swapped in this figure, as usual with water levels, and the y-axis is plotted on a logarithmic scale. Also the
direction of the y-axis deviates from the usual direction (y-values decrease towards the right of the graph).

87
Manual Hydraulic Structures Part II - 'Loads'

HW (high water level) in this figure includes two phenomena: astronomical tide and wind set-up. In the
Netherlands, the Delta Committee assigned normative water levels to areas surrounded by dikes or other
flood defences (dike ring areas). The heights of the normative water levels were related to an economic
optimum of investments in flood protection and consecutive risk reduction per dike ring area. For the
economically most valuable parts of the Netherlands (the Randstad, dike ring number 14) the design
water level is related to an average exceedance frequency of 1/10 000 per year. Chapter 48 of this
Manual explains how the normative water level is used to determine the height of flood defences.

After a profound analysis of the available data, it was found that the frequency of exceedance of a certain
water level h0 can be described with an exponential function:
h −A
− 0
f (h > h 0 ) = e B

where:
-1
f [year ] = exceedance frequency
h [m] = water level
h0 [m] = reference level, for instance the height of structure
A, B [m] = constants

With help of this equation, the design water level can be found, if the exceedance frequency is given. So,
for Hoek van Holland the design water level is NAP + 5,00 m.

The table below gives water levels corresponding to certain probabilities of exceedance for a number of
places (see Tidal Tables for the Netherlands, "Getijtafels voor Nederland").

HW/year Delfzijl Den Helder Scheveningen Vlissingen Bath

10 −1 4,10 m 2,75 m 3,05 m 3,85 m 4,75 m
10 −2 4,95 m 3,40 m 3,70 m 4,40 m 5,45 m
10 −3 5,60 m 3,95 m 4,40 m 4,95 m 6,10 m
10 −4 6,20 m 4,45 m 5,15 m 5,50 m 6,75 m
1 Feb. 1953 - 3,25 m 3,97 m 4,55 m 5,60 m
Table 13-1 (Extreme) storm surge levels

The probability of exceedance is closely related to the frequency of exceedance. A difference is that a
frequency can theoretically be more than 1,0, whereas a probability per definition cannot exceed this
value. For small numbers, however, the difference is negligible. The relation is expressed as a
2
complementary Gumbel distribution :

P (h > h0 ) = 1 − e − f ( h > h0 ) ⋅ n

For exceedance probabilities frequencies of less than 0,1, the exceedance frequency has about the same
magnitude if the exceedance probability is considered over a one period of a year (n = 1), which we
3
usually do .

For foreign projects, however, there are often too few data to determine the design storm surge level on a
statistical basis. In such cases the models for the tide and the wind set-up have to be used to estimate
the storm surge level.

13.4 Other influences

Besides tides and wind set-up, other factors can influence the water level, such as:
1. Short-term atmospheric depressions and oscillating depressions
2. Seiches
3. Relative sea level rise

These factors are explained below and are also included in Chapter 48 of this Manual.
2
The Gumbel distribution is a probability distribution of extreme values which can be applied to storm surge levels. Various types of extreme
value distributions are treated in courses 'Bed, Bank and Shore Protection' (CIE4310) and 'Probabilistic Design in Hydraulic Engineering
(CIE5310)
3
In practice, exceedance frequency and probability are therefore often confused. It should, however, be noticed that the units are different: a
frequency is expressed as a number of occurrences per year, and a probability is dimensionless.

88
Manual Hydraulic Structures Part II - 'Loads'

13.4.1 Short-term atmospheric depressions and oscillating depressions

Short-term atmospheric depressions (buistoten) cause single elevations of the sea water level. These
depressions are generated when a storm with low air pressure (e.g. 300 Pa = 0,03 m of water pressure
less than the surrounding area) approaches the coast with the same velocity as the wave, thus building
up an increasing large wave. The duration of a short-term atmospheric depression can vary between a
couple of minutes and more than one hour. The resulting water level elevation at the Dutch coast can be
up to 0,50 to 1,00 m during storm conditions.

If atmospheric depressions appear in a more or less regular series, the phenomenon is called oscillating
depressions (bui-oscillaties). The period of an oscillating depression can vary from a couple of minutes to
more than an hour; the amplitude in Dutch waters is usually about 0,20 to 0,30 m. During the passing of
hurricane Hugo in Guadeloupe, a water level rise of 0,30 to 0,60 m was observed in deep water. In
shallow water the rise was as much as 2,50 to 3,00 m.

In Dutch waters, the effect of short-term depressions dominates the effect of oscillating depressions; this
is why it is assumed that the effect of the latter is discounted in the contribution of the short-term
depressions. The Dutch guide on the design of river dikes (TAW Leidraad voor het ontwerpen van
rivierdijken - deel 2: benedenrivierengebied, 1989) recommended to reduce the magnitude of the water
elevation due to short-term depressions, depending on the extent of wave run-up. These values have
been adopted by the Dutch guide on lake and sea dikes (TAW Basisrapport bij de Leidraad voor meer- en
zeedijken, 1999).

13.4.2 Seiches
A seiche is an oscillatory rise of the water level in a basin enclosed on three sides, caused by oscillation
of the water level outside the basin. Particularly the shower oscillation within periods ranging from a
couple of minutes to an hour can cause seiches, for instance a sheltered harbour. In general, when
designing structures in closed areas or harbour basins seiches must be taken into account.

Because of the presence of the Hartelkering and the Maeslantkering, the importance of shower gusts,
shower oscillations and seiches has become much less relevant behind these barriers. The Dutch
committee on flood defence (TAW) therefore recommended to neglect these effects behind the barriers,
until further research has been carried out (TAW Technisch Rapport Ontwerpbelastingen voor het
rivierengebied, 2007).

13.4.3 Relative sea level rise

Relative sea level rise is caused by a number of factors:
• Subsidence of the sea bottom due to geological processes and gas and oil extraction
• Increase of the average temperature, which causes melting of the polar caps and thermal expansion
of the ocean water.

The relative sea level rise is still subject to research. An estimate of the relative level rise of the North Sea
is 0,40 to 0,80 m per century (Note on coastal defence 1990 – TR6 – Sea level rise, in Dutch: “Nota
kustverdediging van 1990 - TR6 – Zeespiegelrijzing”). The Dutch Delta Committee 2008 made a
prognosis of 1,30 m in 100 year as a worst case scenario, based on forecast models. The observed trend
until now, however, is 0,19 m relative sea level rise per 100 year (inclusive land subsidence) and a
deviation from this trend has not yet been observed.

The latest reports suggest that the increased average temperature is causing an increased precipitation
in Antarctica, which compensates the accelerated melting of the southern polar cap. The melting of the
northern polar cap does not affect the average sea level. The expanding volume of water due to
decreasing density caused by higher temperatures is said to form the main contribution to sea water level
rise.

89
Manual Hydraulic Structures Part II - 'Loads'

14. Water, waves, theory

In hydraulic engineering different types of waves can be distinguished, such as:

• Translation waves (sloping front!), caused for instance by:
• quickly emptying or filling a lock chamber on or from a canal section
• opening or closing a hydropower station
• wind waves, caused by the wind skimming large water surfaces
• tides, caused by the position and rotation of celestial bodies
• discharge waves caused by precipitation or thaw
• ships’ waves
• pressure waves, which can be created in closed pipes by sudden discharge changes.

With the exception of pressure waves, all the above are waves which occur on the water surface. These
waves are divided into short and long waves. Short waves, such as wind waves and ships’ waves, cannot
be disregarded in the vertical component of the wave velocity in relation to the horizontal component, as
is the case with long waves, such as tide and translation waves.

This chapter briefly describes translation waves and wind waves. Tides are covered in Section 13.1.
Discharge waves are discussed in Section 10.1. Pressure waves in closed pipes are not be treated in this
Manual. When approaching shallows and obstacles, waves change their properties and can start to
break. This is covered in Chapter 16 "Water, waves, shallows + breaking".

14.1 Translation waves

Translation waves can have different causes. One of the most common causes is emptying or filling a
lock chamber. There are also natural processes that can cause a translation wave. Examples of these are
dam breaches (of ice dams and natural soil dams) and tidal bores. Figure 14.1 gives a sketch of a
translation wave. The propagation velocity of the wave is:

(2 d + η )(d + η )
c =u ± g
2d

in which: u [m/s] = the normal flow velocity

d [m] = the water depth
η [m] = the displacement of the water surface

For a relatively small η, the propagation velocity can be approached with:

c = u ± gd

The displacement of the water surface depends on the change of the discharge, the propagation velocity
and the flow width:
∆Q
η=
Bc

3
in which: ∆Q [m /s] = the change of discharge
B [m] = the flow width

In the translation wave the pressure is hydrostatic. The pressure under a translation wave with elevation η
is:

p = ρ g (d + η )

90
Manual Hydraulic Structures Part II - 'Loads'

Figure 14-1 Translation wave

When the flow profile changes, a wave continues in the original direction and a wave reflects in the
opposite direction. The same happens at the branching of a river.
When determining the displacement of the water surface with the continuing and reflected wave, the
following applies, after the passing of the wave, at the point of profile adjustment:
• the water level to the left and the right of this point are equal
• the sum of the discharges equals zero.

14.2 Wind waves

14.2.1 Linear wave theory (regular waves)
This chapter briefly gives a number of characteristics of wind waves. For a more extensive explanation
and the derivation of the equations one is referred to the course Ocean Waves (CT4325). The terms
involved in wave theory can best be defined using a picture of a single wave.

Figure 14-2 Sinusoidal wave shape

The most important terms are:

L The wavelength is the horizontal distance between two successive wave crests (or troughs).
T The period is the time, which passes at a certain fixed point between the passing of two
consecutive wave crests
f The frequency is defined as the number of waves per unit of time, which equals the inverse
of the period (= 1/T)
c The wave velocity (also known as the propagation velocity or phase velocity) with which the
wave crest travels
η The displacement of the water surface
a The amplitude of the wave
H The wave height is the difference in height between the highest and the lowest point of the
wave profile (H=2a)
H/L The steepness is the ratio wave height to wavelength
u and w The velocity components of the orbital movement of the water particles
p The water pressure
d The water depth

Figure 14-2 shows a schematised, sinusoidal, vertical cross section of the water surface, which is
perpendicular to the direction of the wave crests.
This schematisation is used in linear wave theory. The true shape of short waves is highly dependent on
the steepness (H/L).
If H/L ≤ 1/40, the shape of the wave is comparable to a sinusoid. For larger values of H/L , the shape will
bear more resemblance to a trochoid.

The following applies in case of the sinusoidal approximation of the displacement of the water surface:

91
Manual Hydraulic Structures Part II - 'Loads'

2π t 2π r
η (r , t ) = a sin( − +α) or η (r , t ) = a sin(ω t − k r + α )
T L
2π
in which: ω= [rad/s] = angular frequency
T
2π
k= [rad/m] = wave number
L
t [s] = point in time
r [m] = place measured in the wave direction
α [rad] = phase

Note.
In American literature the following equation is used for the displacement of the surface:
η (r , t ) = a cos(k r − ω t + α ) . This means a phase shift of π/2.

The wave velocity c, the wavelength L and the period T are related to each other according to the
following equation:
L ω
c= =
T k

From theoretical derivations, relations follow between the wave velocity, the wavelength and the water
depth, such as:
g
c= tanh(kd ) or c = co tanh(kd )
k
gT 2
L= tanh(kd ) or L = Lo tanh(kd )
2π
gT
in which: co =
2π
gT 2
Lo =
2π

1
In shallow water ( d ≤ L ) , tanh( x ) ≈ x . Therefore, the wave velocity in shallow water can be written
25
as:
g
c= tanh(kd ) ≈ gd
k

When observing waves it is noticeable that a wave group does not move at the same velocity as
individual waves. A wave group propagates with a smaller velocity than the individual waves, whereby the
waves overtake each other. The wave front moves at the so-called group velocity:

1 kd 1
cg = n c with n = + for deep water: n =
2 sinh(2kd ) 2

All water particles in a wave move in a more or less elliptic course. The movement of the particles
decreases the deeper they are below the water surface. The elliptical movement of the water parts is
known as orbital movement.

92
Manual Hydraulic Structures Part II - 'Loads'

shallow water wave

deep water wave

Figure 14-3 Orbital movement

The maximal horizontal and vertical components of the orbital velocity of a water particle are given by:
cosh(k(d + z )) sinh(k(d + z ))
uˆ = ω a and wˆ = ω a
sinh(kd ) sinh(kd )

in which: û [m/s] = maximal horizontal velocity

ŵ [m/s] = maximal vertical velocity
z [m] = co-ordinate of the considered depth relative to the average water surface

As is evident from the formula, the water velocity is largest at z = 0 and smallest at z = -d.

The pressure in a given point at a depth z below the average water line is:

cosh(k (d + z ))
p = − ρ gz + ρ ga sin(ωt − kr ) for z < 0
cosh(kd )

The pressure in a point above the average water line can be approximated by linear interpolation:

p = ρ g (η − z ) for z > 0

In the preceding, a possible flow velocity of the water was not taken into account. The given equations
only apply for a co-ordinate system fixed to the undisturbed water surface. In a flow the equations are for
a co-ordinate system that moves along with the flow. Relative to a fixed point, the propagation velocity
and phase velocity are to be adjusted as follows:

c ' = c + Uer and ω '=ω + k ⋅ U = gk tanh(kd ) + k ⋅ U

A wave has kinetic and potential energy. The potential energy is related to the displacement of the water
surface and the kinetic energy is related to the orbital movement.
The total mechanical energy of one wave is:
1
E = ρ g H 2L (per m′ wave crest)
8
The average energy per unit of surface area is thus:
1
E = ρ gH 2
8
This energy plays an important part in the description of waves in a wave field.

A summary of linear wave theory is given in Table 8-1.

93
Manual Hydraulic Structures Part II - 'Loads'

Relative depth Shallow Water Transitional water depth Deep Water

Characteristics h 1 1 h 1 h 1
< < < >
L 20 20 L 2 L 2
Wave Celerity L gT L gT
L c= = c = c0 = =
c= = gh tanh kh
T T 2π T 2π
Wave Length gT 2 gT 2
L =T gh L= tanh kh L = L0 =
2π 2π
Group Velocity 1 2k h  1 gT
cg = c = g h cg = n c = 1+ ic cg = c0 =
2  sinh 2 k h  2 4π
Energy Flux T
1 1 F= ρ g 2a 2
(per m width) F = E cg = ρ ga 2 g h F = E cg = ρ ga 2 n c
2 2 8π
Particle velocity

Horizontal g cosh k ( h + z )
u =a sinθ u = ωa sinθ u = ω a e kz sinθ
h sinh kh
Vertical
 z sinh k ( h + z )
w = ω a  1 +  cos θ w =ω a cosθ w = ω a e kz sinθ
 h sinh kh
Particle
displacement
Horizontal
a g cosh k ( h + z ) ξ = − a e kz cosθ
ξ =− cosθ ξ = −a cosθ
ω h sinh kh
Vertical sinh k ( h + z ) ζ = a e kz sinθ
ζ =a sinθ
sinh kh
Subsurface p = − ρ g z + ρ g a sinθ p = −ρ g z p = −ρ g z
pressure
cosh k ( h + z ) + ρ g a e kz sinθ
+ ρga sinθ
cosh kh
H 2π 2π
a= ω= k= θ = ωt − k x
2 T L

Table 14-1 Summary of linear (Airy) wave theory - wave characteristics

14.2.2 Wave fields (irregular waves)

The previous section described the so-called regular wave. In a wave field generated by wind, there are
many waves, with different periods, phases, wave heights and wave directions.
The water surface in a wave field is erratic and single waves cannot be distinguished. With a Fourier
analysis, however, it is possible to describe this erratic surface as the sum of a large number of sinusoidal
waves with different wave heights a, angular frequencies ω and phases α.

The movement of the water surface can be described as:

η (r , t ) = ∑∑ ai sin(ωi t − ki , j r + α i , j )
i j

ki cos(θ j )
in which: ki , j =  k sin(θ ) 
 i j 

θj = angle between the wave direction and the positive x-axis

94
Manual Hydraulic Structures Part II - 'Loads'

If the consideration is limited to one point, the model becomes a one dimensional random-phase model:

η (t ) = ∑ ai sin(ωi t + α i )
i

The properties of a wave field are set in an energy spectrum. For the theoretical background and the
formal definition of the energy spectrum, one is referred to the book used for the course on wind waves
(Waves in oceanic and coastal waters, Holthuijsen, 2005).

Continuing with the one dimensional random-phase model, the energy spectrum is:

f+∆f
1
Eηη (f ) ∆f = ∑ ai2
fi = f 2

A two-dimensional spectrum of frequency and direction is:

f+∆f θ +∆θ
1
Eηη (f,θ ) ∆f∆θ = ∑ ∑ a 2
i, j
fi = f θ θ 2
i =

A number of important characteristics of a wave field can be expressed in the moments of the spectrum.
These moments are defined as:

∞
mn = ∫ f n Eηη (f ) df
0

One of the characteristics is the significant wave height. The significant wave height Hs is the average of
the highest 1/3 of the waves. The relation between this wave height and the spectrum is:

Hs = 3,8 m0

The momentary wave height has a Rayleigh probability distribution. The exceedance probability of a
given wave height within a given wave field is:

  x 2 
P(H > x ) = exp  −2   
  Hs  
 

For the continuation of wave heights and exceedance probabilities, see the next chapter.

95
Manual Hydraulic Structures Part II - 'Loads'

15. Water, waves, wave heights

updated: February 2015

15.1 Estimate of wave height and period if no measurements are available

The most important waves are waves generated by the wind. If a structure is to be dimensioned for a
certain wave height, this wave height has to be known (measurements). If no measurements are
available, the significant wave height and wave period can be estimated by using equations as proposed
by Charles L. Bretschneider. These equations were later on improved by Young and Verhagen (1996)
and Breugem and Holthuijsen (2006):

0,572
  
( )
0,79
 1,14  4,41⋅ 10 − 4 F 
H = H∞ tanh 0,343 d
ɶ ɶ ⋅ tanh 
( )
1,14 
  tanh 0,343 d 
  
0,187
  
( )
1,45
 2,01  2,77 ⋅ 10 − 7 F 
T = T ∞ tanh 0,10 d
ɶ ⋅ tanh 
( )
2,01 
  tanh 0,10 d 
  
g H gTp
in which: Hɶ [-] = m0
Tɶ [-] =
U10 2 U10
gF gd
Fɶ [-] = 2
dɶ [-] = 2
U10 U10
F [m] = fetch (strijklengte)
d [m] = average water depth over the fetch
U10 [m/s] = wind velocity at an altitude of 10 m
Tp [s] = peak wave period
H∞ [-] = dimensionless wave height at deep water = 0,24
T ∞ [-] = dimensionless wave period at deep water = 7,69
Hm0 [m] = significant wave height, estimated from a wave spectrum (Hm0 ≈ Hs)

An alternative for the improved Bretschneider formulas is provided by the nomograms of Groen and
Dorrestein (Figure 15-2 and Figure 15-3). Figure 15-2 should be used for deep and transitional water (d/L
> 0,1) and Figure 15-3 for shallow water (d/L < 0,1). One should use these nomograms cautiously: neither
is dimensionless. Because Groen and Dorrestein used different data sets than Bretschneider, their
findings are not identical.

For the water depth d it is recommended by the TAW to use one value, like the average depth of a lake or
river, if suitable, or the water depth in a river-foreland (uiterwaard). The influence of local deep trenches is
usually neglected.

For the estimation of wave heights just in front of flood defences, it is recommended to use an effective
fetch, which takes the shape of the water body into account. The effective fetch in a random situation is
the weighed average of the projections ℓ(α) on the wind direction of all fetches in all directions α (see
Figure 15-1), according to:
αm
∫ w (α ) ⋅ ℓ(α ) ⋅ dα
−α m
Fe =
αm
∫ w (α ) ⋅ dα
−α m
where: w(α) [-] = weight function. Recommended function is: w(α) = cos(α)
ℓ(α) [m] = fetch in wind direction
α [°] = deviation from wind direction in point of interest
αm [°] = boundary angle for wind directions that influence the effective fetch

96
Manual Hydraulic Structures Part II - 'Loads'

Figure 15-1 Definition sketch for determining the effective fetch (Holthuijsen, 1980)

15.2 The design wave height

The significant wave height Hs is the average of the highest 1/3 of the waves. This wave occurs regularly
and is therefore a much lower than the design wave height Hd. If the effects of shallow water can be
disregarded with a small wave height, a Rayleigh distribution can be assumed. The probability of
exceedance of a given wave height within a given wave field is:
2
 x 
−2  
Pr(H > x ) = e  Hs 

Therefore, the probability that the design wave height Hd is exceeded during a storm with N waves is:

−2( Hd / Hs )2
Pr(H > Hd ) = 1 − e−N ⋅e

For a storm along the coast one can assume Tstorm = 2 h. For rivers and the IJsselmeer, Tstorm = 4 h can
be supposed. Presuming Twave = 3 s , the number of waves N along the coast is:

Tstorm
N= ≈ 2400
Twave

If one allows an exceedance probability Pr(H > Hd ) = 0,10 , the design wave height Hd is:

Hd = 2,25 Hs

To ascertain the design wave length one may assume that the shape of the energy spectrum essentially
does not change for light and heavy storms, so:

Ld ≈ Ls

This wave can change shape because it reaches shallow water (refraction, shoaling and breaking) or
obstacles (diffraction and reflection). This is treated in the next chapter.

97
Manual Hydraulic Structures Part II - 'Loads'

30
100 nautical miles = 185 km; 100 km = 54 nautical miles 3000 4000
20 2000 30
1 knot = 0.51 m/s; 1 m/s = 1.95 knots 150
0
17
600 25
14 400 15
300 14
10 200
150 12
20
8 100
11 17.5
60
6 10
40 15
5
wave height Hs in metres

30 9
4 20 8
15 12.5
3 10
7

6 6 10
2 wind velocity U 4
3
in m/s. 5
1.4 2
7.5
1,5
1 1 4
0.8 0.6 30
0.7
0.6 25 5
3
0.5 20
17.5 wave period Ts
0.4
15 in s wind velocity U
0.3 12.5 in m/sec.
2
10
0.2
7.5

5
0.1
0.1 0.5 1 2 3 6 9 12 18 24 36 48 72 96
duration t in hours

Figure 15-2 Nomogram, valid for deep and transitional water (d / L > 0,1) (Groen and Dorrestein, 1976)

98
Manual Hydraulic Structures Part II - 'Loads'

Ts
2
r d
te
wa
ep
de
1
0.8
Hs
0.6
d

0.4

0.2
r
te
wa
ep
de

0.1
0
00 50
0.08 10
=
d
F/

0.06

0.04

0.02

0.01
1 2 3 4 5 6 8 10 20 40

U
d

Figure 15-3 Nomogram, valid for shallow water (d / L < 0,1) (Groen and Dorrestein, 1976)

15.3 Literature
Bretschneider, C.l. (1958) Revisions in wave forecasting: deep and shallow water. Proceedings of the 6th
Conference on Coastal Engineering, pp. 30-67

Groen, P. and R. Dorrestein (1976) Zeegolven. Opstellen op oceanografisch en maritiem meteorologisch

gebied nr. 11. Derde herziene druk. KNMI. Staatsdrukkerij- en uitgeverijbedrijf 's-Gravenhage

Holthuijsen, L.H. (1980) Methoden voor golfvoorspelling. TAW.

Holthuijsen, L.H. (2008) Waves in oceanic and coastal waters. Cambridge University Press, ISBN 978-0-
521-86028-4

Young, I.R. and L.A. Verhagen (1996) The growth of fetch-limited waves in water of finite depth. Part 1:
Total energy and peak frequency. Coastal Engineering 29, pp. 47-78

99
Manual Hydraulic Structures Part II - 'Loads'

16. Water, waves, shallows + breaking

When waves approach the coast, a number of changes occur, caused by the change of water depth. Due
to the smaller depth, the wave velocity decreases and the wave front turns so it runs increasingly parallel
to the depth contours (refraction). As a result, the wave crests become narrower, the wave becomes
more concentrated and the wave height increases. At the same time, the wave velocity decreases,
thereby reducing the wavelength, causing a further increase of the wave height (shoaling). So, the wave
height increases and the wave length reduces. At a certain point, the waves are so steep that they break.
This section considers these three phenomena:
1. Refraction
2. Shoaling
3. Breaking of waves

Besides that, the effect of an obstacle is also discussed. The two most well-known consequences are:
1. Diffraction
2. Reflection

Unless stated otherwise, regular waves are assumed in this chapter.

16.1 Shallows: refraction

If a wave approaches a sloping coastline at an angle, the propagation velocity will vary along the wave
crest due to the difference in water depth along the wave crest. After all:

g
c= tanh(kd ) = c0 ⋅ tanh(kd )
k
In shallow water the propagation velocity is smaller. Therefore, with decreasing depth, the wavelength
shortens. The wave front decelerates in the first part to reach shallow water. The wave front will thus turn.
This causes a bend of the propagation velocity towards the coast. This phenomenon is known as
refraction.

deep water

θ
depth contours

shallow water

Figure 16-1 Refraction

In the case of a coast with straight parallel depth contours, the angle between the wave crest and a depth
contour line can be derived directly from the local depth and the angle between the wave crests and a
parallel contour line in deep water (Figure 16-1). This relation is called Snell’s law and reads as follows:

sin(θ1 ) c1 c0 tanh(k1 d1 ) L
= = = tanh(k1 d1 ) = 1
sin(θ o ) c0 c0 L0

in which: θo = the angle between the wave crest and the depth contour line in deep sea
θ1 = the angle between the wave crest and the depth contour line in shallow water

100
Manual Hydraulic Structures Part II - 'Loads'

16.1.1 From shallow to deep

If one knows the angle of the wave ray, the wave height and wavelength in shallow water, one can
calculate the wave ray angle in deep water using the above equation or the dotted line in Figure 16-2.

16.1.2 From deep to shallow

If the angle of the wave ray, the wave height and the wavelength in deep water are known, the angle of
the wave ray in shallow water cannot be calculated using the equation above. The equation requires the
wavelength in shallow water, for which the following applies:
2π
L1 = L0 tanh(k1 d1 ) met: k1 =
L1
sin(θ1 )
So as a function of d / L0 is an implicit function, which has to be solved iteratively. The inverse
sin(θ o )
solution is given as a solid line in the graph below, plotted against the relative water depth d / L0 (water
depth d in shallow water, divided by the wavelength L0 in deep water).

10
L0/L1 = 1/tanh(k1d )= sin(θ0)/sin(θ1)

9
8
7
6
5
4
3
2
1
0
0.001 0.01 0.1 1
d/L 0 en d/L 1

d/L0 d/L1

Figure 16-2 Relation between depth and wavelength (d = depth, L1 = wavelength at considered depth, Lo = wavelength in deep water)

Depending on the concentration or spread of wave rays, the wave height will increase or decrease.
Generally, for the wave height in shallow water:

H = K s K r Ho

in which: Kr [-] = the refraction coefficient

Ks [-] = the shoaling coefficient (see next section)

In the case of straight parallel depth contours, the wave height decreases with a factor:

b0 cos(θ o )
Kr = =
b1 cos(θ1 )

with: b = the wave crest width

This is because the wave crest width b continues to increase while the wave crest turns, which causes a
reduction of the energy density and thus also of the wave height. The change of wavelength in shallow
water also leads to a change of the energy density and the wave height, but that phenomenon is called
shoaling and is covered in the next section.

Refraction also occurs when a wave enters an area with a current (along the coast). In this case, the
wave will turn more or less in the direction of the current.

101
Manual Hydraulic Structures Part II - 'Loads'

16.2 Shallows: shoaling (non-breaking waves)

When the water depth decreases, the propagation velocity and the wavelength are reduced with a
constant period. This influences the wave height.

16.2.1 Theory
Chapter 14 showed that the wave energy per unit of surface area equals:
1
E = ρ gH 2
8
The group velocity is:
1 kd g
cg = n c with n = + and c = tanh(kd ) = c0 ⋅ tanh(kd )
2 sinh(2kd ) k
The energy flux is the amount of energy that passes a certain point per unit of width. This energy flux
equals:
F = E ⋅ cg = constant
and is constant for non-breaking waves (no loss of energy) and straight approaching waves (no change of
width).

The wave height is therefore:

constant
H2 =
n⋅c
The wave height in a shallow area, H1, is therefore dependent on the wave height in deep water, H0 ,
according to:
1
H1 cg ;0 c0 n0 2 1
= Ks = = = =
H0 cg ;1 c1n1 tanh(k d )n1  2k d 
tanh(k d )  1 + 
 sinh(2 k d ) 
De shoaling coefficient is therefore a function of the wave number k and the water depth d:

1 d
Ks = with: k d = 2π
 2k d  L
tanh(k d )  1 + 
 sinh(2 k d ) 

This solution is represented by the dotted line in the figure below. If the wave height and wavelength are
known in a certain shallow area, these can be used to calculate the wave height in deep water. The
inverse, using this solution and a known wave height and a known wave length in deep water to calculate
the wave height in shallow water is not possible. The problem is that L and thus also k are dependent on
the depth d and on themselves, for:
2π d
L = Lo tanh(kd ) = Lo tanh( )
L
This is an implicit function. De shoaling coefficient, a function of the water depth d and wavelength in
deep water L0, can therefore only be solved iteratively. This solution is represented by the solid line in
Figure 16-3.

Besides shoaling, refraction also influences the wave height, for this see the previous section.

16.2.2 (Preliminary) design

The wave height H1 of regular (non–breaking) waves in shallow water depends on the wave height in deep
water H0, the refraction coefficient Kr and on the shoaling coefficient Ks, according to:

H1 = K r K s Ho

In Figure 16-4 the shoaling coefficient is given as a function of the relative water depth (water depth d in
shallow water, divided by the wavelength L0 in deep water).

102
Manual Hydraulic Structures Part II - 'Loads'

Figure 16-3 Relation between Ks and d/L (L = wavelength at considered depth, Lo = wavelength at deep sea)

For the refraction parameter, the following equation has been found:

cos(θ o )
Kr =
cos(θ1 )

To calculate the refraction ( θ0 ⇒ θ1 ), see previous section.

The theory given above only applies for regular waves that don’t break (or haven’t yet broken).

16.3 Shallows: breaking waves

Distinction should be made between individual waves (for the calculation of loads on structures) and
significant waves (for the calculation of run-up, overtopping, stability of stones).

Individual waves
Due to the decreasing wavelength and the increasing wave height in shallow areas, the steepness of the
wave increases. Waves will theoretically break if they become too steep, or if the water depth becomes
too small with respect to the wave height:
1
• H ≥ ⋅L
7
or:

• H ≥ 0,78 ⋅ d
There are also more complex formulas for the shallowness criterion, e.g. by Miche. In spite of the
theoretically deduced criterion, individual non-breaking waves with a ratio of H = 1,2·d have been observed
under very specific circumstances.

Significant waves
When calculating breaking for a wave spectrum, the significant wave height Hs should be used. This is the
average height of the 1/3 highest waves. For gentle bed slopes (≤ 1/100), it is often assumed that breaking
occurs if:

Hs
• ≥ 0,4 to 0,5
d
The way in which a wave breaks on a smooth slope with a constant slope angle depends on the
steepness of the wave and the slope of the bed. This is characterised with the

103
Manual Hydraulic Structures Part II - 'Loads'

tan α
breaker parameter (Iribarren): ξ =
Hs Lo

in which: α [º] = angle of the slope

Hs [m] = significant wave height
Lo [m] = wavelength in deep water

Depending on the value of the breaker parameter, different types of breaking occur, as is shown in Figure
16-4.

Figure 16-4 Types of breaking

Notes
• The depth profile need not be constant in time. Shallows can appear or disappear. It is therefore important to
find out if the bed consists of rock or sand.
• The depth can depend on the tide and on the wind set-up (storm surge).
• Changes of depth also mean changes of refraction, shoaling and breaking.

The validity of wave theories is depicted in Figure 16-5.

Figure 16-5 Validity of wave theories (Le Méhauté, 1976 / Introduction to bed, bank and shore line protection, CIE4310)

104
Manual Hydraulic Structures Part II - 'Loads'

16.4 Obstacle: reflection

If waves run into a structure they can break or reflect. The reflection can be partial or complete.
Characteristics of completely reflected waves are:
• the energy of the reflected wave equals the energy of the incoming wave
• the period of the reflected wave equals the period of the incoming wave
• the reflected wave is in phase with the incoming wave.

The consequence of the above is that a standing wave with (in case of complete reflection) a wave height
twice the size of an incoming wave is created in front of the structure. If the reflection is partial, the wave
height of the standing wave will be less. In general the following applies:

H = (1 + χ ) ⋅ Hi

in which: H [m] = the wave height of the standing wave

Hi [m] = the wave height of the incoming wave
χ [-] = the reflection coefficient ≤ 1

The value of χ depends on the permeability, roughness and slope of the structure and on the steepness
of the incoming waves and the water depth in front of the structure.

Figure 16-6 Reflection

The load on the wall can be calculated according to Chapter 18 of this Manual.

16.5 Obstacle: diffraction

If there’s an obstacle in the course of a wave (e.g. a breakwater or an island), wave motion still occurs in
the shadow zone behind the obstacle. The transfer of energy apparently not only takes place in the wave
direction. The wave crests bend round the object shaped like circular arcs. This phenomenon is called
diffraction.

The wave height changes due to diffraction, whereby the wave height on the lee side of the object is
smaller than that of the incoming wave, whilst the wave height next to the object is often larger than that
of the incoming wave.
Generally:

H = K d Ho

where: Kd [-] = the diffraction coefficient

105
Manual Hydraulic Structures Part II - 'Loads'

1H 0.5 H

0.15 H

0.05 H

>1 H half infinite rigid

impermeable breakwater

1H
Figure 16-7 Diffraction

The “Shore Protection Manual” (CERC 1984) gives a large number of diagrams for the estimation of Kd
with different wave directions. The Shore Protection Manual can be found on the internet on
www.google.com with search request 'coastal engineering manual'. Figure 16-8 is taken from the book
'Oceanographical Engineering' by Robert Wiegel (1964). It shows the diagram for waves moving straight
towards a breakwater. The distances in x- and y-direction are divided by the wavelength L.

Figure 16-8 Diffraction coefficients for straight incoming waves (Wiegel, 1964)

For waves approaching from other directions relative to the breakwater, reference is made to literature,
like Diffraction of periodic sea waves along a vertical breakwater by Penny and Price (1952).

106
Manual Hydraulic Structures Part II - 'Loads'

17. Water, waves, run-up + overtopping

SLS overtopping requirements added: February 2016

For the design of slanting structures (a slope or a wall) it is important to know how far a wave can run up
the slope, or how much water can go over the structure. In the first place, this depends on the average
water level (= highest astronomical tide + storm set-up, or the highest river level during a peak discharge)
and in the second place on the height of the waves attacking the structure.

This chapter covers the following two phenomena:

1. Wave run-up (golfoploop)
2. Wave overtopping (golfoverslag)

These phenomena play a role in determining the crest height of flood defences and the top of structure
level of other hydraulic structures, see Chapter 48. The 2%-run-up calculation for the determination of
dike heights was used in the Netherlands and Germany until the end of the twentieth century, but has
4
been replaced by the more universal calculation of critical overtopping discharges .

Figure 17-1 Principle sketch for the 2% run-up height

17.1 Wave run-up

17.1.1 General
The wave run-up R is the vertical distance between the average water level and the highest point on a
slope that is reached by water running up the slope. This applies to both breaking and non-breaking
waves.

The wave run-up depends on:

• The wave field
• The geometry of the flood defence
• The roughness of the surface

As a rule of thumb, the following 'old Delft formula' could be used to calculate the crest freeboard related
to an acceptable amount of run-up on dikes:

R2% = 8·Hs· tan(α)

where:
R2% [m] = the value of run-up exceeded by 2% of the waves
Hs [m] = the significant wave height
α [º] = the angle of the slope with the horizontal

4
The 2% run-up requirement was introduced for the design of the 32 km long Afsluitdijk. It was assumed that the inner slope of a dike would
fail after 40 to 60 overtopping waves, which equals about 2% of the waves in a 'typical' Dutch storm), but since 1999 this has been replaced
with an overtopping criterion (TAW- Leidraad Zee- en Meerdijken, 1999).

107
Manual Hydraulic Structures Part II - 'Loads'

This old Delft formula is valid for the Dutch coast under storm conditions and:
• relatively gentle slopes (1:3)
• “normal” wave steepness (between 4 and 5%)

With help of several extra parameters, the old Delft formula becomes more sophisticated:

R 2% = 8 ⋅ HS ⋅ tan(α ) ⋅ γ f ⋅ γ b ⋅ γ β
where:
γf [-] = factor taking the slope roughness into account
γb [-] = berm influence factor
γß [-] = influence factor for oblique wave attack
See below for an explanation of these factors.

I.A. Hunt found out that the factor of '8' in the old Delft formula actually is related to the wave steepness,
which can be represented by the breaker parameter (Iribarren parameter) ξ (see section 16.3):

Ru = f (ξ ) ⋅ HS ,

or to be more precise:

Ru 2% = 1,5 ⋅ γ f ⋅ γ b ⋅ γ β ⋅ Hs ⋅ ξ p

where the maximum of Ru 2%max = 3 ⋅ Hs

J.W. van der Meer fine-tuned Hunt's equation. For deterministic calculations it reads:
 1,6 
R2% = 1,75 ⋅ γ b ⋅ γ f ⋅ γ β ⋅ ξ m −1,0 ⋅ H m 0 with a maximum of: R2%,max = 1,00 ⋅ γ f ⋅ γ β ⋅  4,3 − ⋅H
 ξ  m0
 m −1,0 

where the breaker parameter (Iribarren parameter) is defined by:

tan α tan α
ξ m −1,0 = =
Hm0 / L 0 Hs / 1,56 ⋅ Tm2−1,0( )
Hm0 = estimate of the significant wave height from spectral analysis = 4√m0 . Hm0 ≈ Hs
Tm-1,0 is a calculated wave period that follows from the wave spectrum. In about 80% of the cases,
Tm-1,0 ≈ 0,9 ·Tp

The Van der Meer equation has been adopted by the CUR-TAW-guidelines and the European Overtopping
Manual 2007.

Roughness (factor γf)

The reduction factor γf that takes the roughness and the permeability of the surface into account is:
• 1,00 for asphalt, concrete with a smooth surface
• 0,95 for concrete blocks, block mats
• 0,70 for gravel, gabions
• 0,60 for quarry stone (rip-rap)
• 0,50 for cubes (random positioning)
• < 0,50 for X-blocs, tetrapods, dolosses (see European Overtopping Manual for more data)

Berm influence (factor γb )

A berm reduces the wave run-up and its influence can be taken into account by using a berm influence
factor γb (also known as the shoulder reduction factor). The magnitude of the berm influence factor
depends on the length of the berm Bb and the water depth above the berm hB.

108
Manual Hydraulic Structures Part II - 'Loads'

Figure 17-1 Definitions berm reduction

At both sides of the berm, the slope is intersected at a vertical distance HS from the horizontal centre
plane of the berm, giving a length LB. hB is the distance between SWL and the berm level (can be
negative or positive). γb finally becomes:

BB   hB  
γ b = 1− 0,5 + 0,5 cos  π 
LB   x 
x = R2% for R2% > − hB > 0 (berm above SWL)
x = 2HS for 2HS > hB ≥ 0 (berm at or below SWL)

With limits: 0,6 ≤ γb ≤ 1. The equation above shows that a berm at storm surge level is most efficient. In
that case, the optimal width of the berm is BB ≈ 0,4 LB, (see Figure 17-1 for an explanation of LB) and the
maximum berm width is 25% of the wave length: BB < 0,25 Lwave. The slope of the berm is 1:15 at
steepest, otherwise it should be considered as a slope rather than a berm. For more information, see
TAW Technisch Rapport Golfoploop en Golfoverslag bij Dijken (2002).

If there is no berm, no berm influence factor should be used in overtopping or run-up calculations.

Angle of incidence (factor γβ)

For oblique waves the angle of wave attack is defined as the angle between the direction of propagation
of waves and the axis perpendicular to the structure (for perpendicular wave attack: β = 0°). For short-
crested waves:

γβ = 1 - 0,0022 β for : 0° ≤ β ≤ 80°

γβ = 0,824 for β > 80°

So, at an angle of 80° (fetch waves) the reduction is only 18%, compared to perpendicular wave attack.

Foreshore reduction
On very shallow foreshores waves will break; the wave spectrum transforms and is flattened. Also the
waves are no longer Rayleigh-distributed. According to TAW (2002) one should use for the calculation of
ξ not the Tm0, but the Tm-1,0, which is m1/m0 (the first negative moment of the spectrum divided by the zero
order moment of the spectrum). In case no detailed spectrum information is available, one may use
Tm-1,0 ≈ 0,9 Tp. For the wave height on the shallow foreshore one should use the Hm0 on the foreshore.

17.2 Wave overtopping

Overtopping waves can jeopardise a civil engineering work if they cause erosion or softening of the
foundations. Wave overtopping can also cause a nuisance for the surroundings. To prevent severe wave
overtopping, the design of the structure should therefore include a sufficient freeboard above the design
water level.

For wave overtopping considerations, the wave run-up Rn simply exceeds the crest height z. The wave
overtopping is usually characterised by an overtopping discharge q per metre of the water defence,
averaged over time. This discharge depends on the wave height, the wave steepness, the slope and the
existing freeboard. This concerns overtopping discharges averaged over time. In reality, a far larger
discharge can occur for a short time, depending on the percentage of overtopping waves. With the results
of several investigations, a global relationship between all of these factors can be derived.

109
Manual Hydraulic Structures Part II - 'Loads'

17.2.1 Overtopping of structures with a slope

In August 2007, the EurOtop team released the "Wave Overtopping of Sea Defences and Related
Structures: Assessment Manual", in short: "European Overtopping Manual" (EurOtop team 2007). That
manual gives "guidance on analysis and/or prediction of wave overtopping for flood defences attacked by
wave action." It replaces the older Dutch "Technical Report Wave Run-up and Wave Overtopping at
Dikes" of TAW, Technical Advisory Committee on Flood Defences, author: J.W. van der Meer, and two
other (foreign) reports.

The Overtopping Manual gives maximum overtopping discharges, to be used for the design of hydraulic
structures where waves could overtop. Distinction should be made between serviceability limit state (SLS)
and ultimate limit state (ULS) requirements. ULS requirements concern the structural integrity (structural
safety) of the hydraulic structure itself. SLS requirements are related to the direct impact on pedestrians,
vehicles and property behind the defence. Maximum overtopping discharges and volumes are given in
Table 17-1 for ULS and Table 17-2 for SLS.

Next to these ULS and SLS requirements, the storage capacity of the water system behind the defence
could restrict the allowable overtopping volume. This could become critical if sea water overtops into a
fresh water basin or lake (regarding salt intrusion).

Hazard type and reason mean discharge

q (l/s/m)
Embankment seawalls / sea dikes
No damage if crest and rear slope are well protected 50-200
No damage to crest and rear face of grass covered embankment of clay 1-10
No damage to crest and rear face of embankment if not protected 0,1
Promenade or revetment seawalls
Damage to paved or armoured promenade behind seawall 200
Damage to grassed or lightly protected promenade or reclamation cover 50
Table 17-1 ULS requirements for overtopping (European Overtopping Manual 2007)

Hazard type and reason mean max. volume

discharge Vmax (ℓ/m)
q (ℓ /s/m)
For pedestrians
Trained staff, well shod and protected, expecting to get wet; 1 - 10 500
overtopping flows at lower levels only, no falling jet, low danger of at low level
fall from walkway
Aware pedestrian, clear view of the sea, not easily upset or 0,1 20 - 50
frightened, able to tolerate getting wet, wider walkway at high level
or velocity
For vehicles
Driving at low speed, overtopping by pulsating flows at low flow 10 - 50 100 - 1000
depths, no falling jets, vehicle not immersed
Driving at moderate or high speed, impulsive overtopping giving 0,01 - 0,05 5 - 50
falling or high velocity jets at high level
or velocity
For property behind the defence
significant damage or sinking of larger yachts 50 5000 - 50 000
sinking small boats set 5 - 10 m from wall; damage to larger 10 1000 - 10 000
yachts
Building structure elements 1 -
Damage to equipment set back 5 - 10 m 0,4 -
Table 17-2 SLS requirements for overtopping (European Overtopping Manual 2007)

110
Manual Hydraulic Structures Part II - 'Loads'

The principal equation used for wave overtopping is:

 - b Rc 
q  
H
= a ⋅ e m 0 
3
g ⋅ Hm 0

where:
q
[-] = dimensionless overtopping discharge
g ⋅ Hm3 0
3
q [m /s/m] = overtopping discharge
Rc
[-] = the relative overtopping height (relative crest freeboard)
Hm0
Rc [m] = overtopping height (crest height)
Hm0 [m] = estimate of significant wave height from spectral analysis = 4√m0 ≈ Hs

(See Figure 17-1for a definition sketch)

Further needed parameters according to the Overtopping Manual 2007:

0,067
a [-] = γ b ⋅ ξ m -1,0
tan α
4.3
b [-] =
ξ m -1,0 ⋅ γ b ⋅ γ f ⋅ γ β ⋅ γ v

ξm-1,0 [-] = breaker parameter (see previous section)

γb [-] = influence factor of a berm (see previous section)
γf [-] = influence factor for the permeability and roughness of the slope, sometimes written
as γR (see previous section)
γβ [-] = factor for oblique wave attack (notice that γβ differs from γβ for wave run-up)
γ β = 1- 0,0033 β for 0° ≤ β ≤ 80°
γ β = 0,736 for β > 80°
β [°] = the angle between the direction of propagation of waves and the axis perpendicular
to the structure
γv [-] = influence factor for a vertical wall on top of the crest (see below)

The breaker parameter also referred to as surf similarity or Iribarren number is defined as:
tan α
ξm-1,0 =
Hm 0 / Lm -1,0
where tan α is the slope of the front face of the structure and Lm-1,0 being the deep water wave length:
g ⋅ Tm2-1,0
Lm -1,0 = [m]
2⋅π

In another shape, the overtopping equation reads:

 -b Rc 
 
H
q = a ⋅ e m 0  3
⋅ g ⋅ Hm 0 ,
which is valid for ξm-1,0 < 5,0. In case of shallow foreshores, other formulas are recommended.
 −2,3 ⋅ Rc 
 
 H ⋅γ ⋅γ 
The European Overtopping Manual 2007 gives a maximum of qmax = 0,2 ⋅ e m 0 f β  ⋅ g ⋅ Hm
3
0 .

The wave steepness s0 (= Hm0 / Lm-1,0) can vary from 0,04 (steep storm waves) and 0,01 (long waves due
to swell or wave breaking).

111
Manual Hydraulic Structures Part II - 'Loads'

Additional wave overtopping experiments have led to improved equations, as specified in the new version
of the Overtopping Manual (2016):

This translates to the following design equation for the overtopping height:

1/1,3
ξ m −1,0 ⋅ Hm 0 ⋅ γ b ⋅ γ f ⋅ γ β ⋅ γ v   q tan α  
Rc = ⋅  − ln  
2,5  g ⋅ H ⋅ 0,026 ⋅ γ ⋅ ξ
3 
  m0 b m −1,0  
and a minimum of:
1/1,3
Hm 0 ⋅ γ f ⋅ γ β ⋅ γ *   q  
Rc > ⋅ − ln  
 
 0,1035 ⋅ g ⋅ H mo
2,5 
3
 

where γ* can be the same as γv, but it can also include several other effects, like the presence of a
promenade, a smooth slope or a parapet ('bull nose').

Influence of vertical walls on top of a slope (factor γv)

The reduction due to relatively small vertical walls on top of the slope can mount up to 35%. This topic
has to be studied further, but based on experience up to now the next equation could be applied:

γv = 1,35 - 0,0078 · αwall

where αwall is the slope of the wall (for vertical walls, αwall = 90º).

The European Overtopping Manual restricts the equation above to the next conditions:
• the average slope of 1,5 Hm0 below the still water line to the foot of the wall (excluding a berm) must lie
between 1 : 2,5 to 1 : 3,5.
• the width of all berms together must be no more than 3 Hm0.
• the foot of the wall must lie between about 1,2 Hm0 under and above the still water line;
• the minimum height of the wall (for a high foot) is about 0,5 Hm0. The maximum height (for a low foot) is
about 3 Hm0.

If there is no vertical wall on top of a slope, of course no correction factor γv should be used in the
overtopping calculation.

112
Manual Hydraulic Structures Part II - 'Loads'

17.2.2 Overtopping of vertical walls

This section deals with wave overtopping over entire vertical walls, so not with (relatively small) vertical
walls on top of a slope. To calculate the amount of wave overtopping over vertical walls or vertical
structures like caisson breakwaters and navigation locks (including the gates), Franco and colleagues
proposed the following equation (Franco et. al., 1994):

Rc
q −b
γ ⋅Hs
= a⋅e
g ⋅ Hs3

where:
Rc [m] = crest freeboard
3
q [m /m/s] = specific discharge
a, b [-] = empirical coefficients
γ [-] = geometrical parameter
Hs [m] = significant wave height
2
g [m/s ] = gravity acceleration
For rectangular shapes: a = 0,192, b = 4,3, γ = 1.

In the European Overtopping Manual two types of wave conditions in front of the vertical wall are
distinguished: If the waves are relatively small compared to the local water depth, or in case of gentle
wave steepnesses, they will not critically influenced by the toe or approach slope of the structure. The
resulting loads on the structure are rather smooth-varying. These conditions are called 'non-impulsive'.
The other type of conditions, the 'impulsive' conditions, occurs if the waves are larger in relation to the
local water depth. Some waves will break violently against the wall with forces up to 40 times more than
under non-impulsive conditions. This also leads to higher overtopping volumes.

Whether or not the wave conditions are impulsive, can be determined with help of the impulsiveness
parameter h* . Conditions are considered to be impulsive if h* < 0,2 and non-impulsive if h* > 0,3. For the
intermediate range overtopping volumes should be calculated for both conditions and the largest value
should be used for the design.

The impulsiveness parameter can be calculated according to:

hs 2 ⋅ π ⋅ hs
h* = 1,35 ⋅
Hm 0 g ⋅ Tm2−1,0
where:
hs [m] = water depth at the front of the structure
Hm0 [m] = estimate of significant wave height from spectral analysis ≈ Hs
2
g [m/s ] = gravity acceleration
Tm-0,0 [s] = average wave period (defined by m-1/m0)

The overtopping volumes for deterministic design can then be calculated according to:

−3.1
q  R  Rc
= 2,8 ⋅ 10 −4  h* c  , valid for 0,03 < h* < 1,0 for impulsive conditions (h* < 0,2)
h 2
* g ⋅ hs
3
 Hm 0  Hm0

and
Rc
q −1,8 Rc
= 0,04 ⋅ e Hm 0
, valid for 0,1 < < 3,5 for non-impulsive conditions (h* > 0,3):
g ⋅ Hm3 0 Hm 0

For equations for probabilistic design, battered walls and vertical walls with parapets, composite walls and
other shapes, the reader is referred to the European Overtopping Manual.

113
Manual Hydraulic Structures Part II - 'Loads'

Note
If non-breaking waves are reflected by a vertical structure, interference of incoming and reflecting
waves will occur. This will result in a standing wave with a height twice as much as the incoming wave.
In the overtopping volume equation, however, the wave height of the incoming wave, not the standing
wave, should be used, because the parameters in the equations already include this effect.

17.3 Literature
Franco, L., Gerloni, M. de, Meer, J.W. van der (1994). Wave overtopping on vertical and composite
breakwaters. Proceedings of the 24th International Conference on Coastal Engineering. Kobe, pp 1030-
1044

EurOtop (2007). Wave overtopping of sea defences and related structures: Assessment manual
"European Overtopping Manual". To be downloaded from: www.overtopping-manual.com (mind the
errata!)

EurOtop (2016). Manual on wave overtopping of sea defences and related structures. An overtopping
manual largely based on European research, but for worldwide application. Van der Meer, J.W.,
Allsop, N.W.H., Bruce, T., De Rouck, J., Kortenhaus, A., Pullen, T., Schüttrumpf, H., Troch, P. and
Zanuttigh, B., www.overtopping-manual.com.

114
Manual Hydraulic Structures Part II - 'Loads'

18. Water, waves, wall, non-breaking

Updated: February 2015

Civil engineering works located at the sea side of a breaker zone, can be subjected to loads by non-
breaking waves. Non-breaking waves also occur in waterways and lakes in which the wave height is
limited. Unlike around slender structures, the wave pattern is influenced by a wall. The wave height in
front of the wall is determined by refraction and diffraction (see Sections 16.1 and 16.5).

There are five methods to calculate the load on a wall due to non-breaking waves. They are given in the
table below with a description of when they are applied.

No Method Design phase Notes

1 Rule of thumb preliminary estimate conservative
2 Linear theory preliminary (and final) design -
3 Sainflou preliminary design simple!
4 Rundgren final design not in this handbook
5 Goda final design also for sills!
Table 18-1 Summary of methods

18.1 Rule of thumb (hydrostatic pressure)

According to linear wave theory for non-breaking waves against a vertical wall, the wave height H in front
of the wall is double the incoming wave height Hi, in the case of total reflection. In short:

H = 2H i and with H = 2a : a = Hi

The wave force is maximum when the wave in front of the wall reaches its highest level (wave top). If this
is considered as a stationary load, the following rule of thumb can be applied to calculate the maximum
wave force Fmax [N/m] against a wall (per metre length):

Fmax = ½ ρ gHi2 + d ρ gH i

3
in which: ρ [kg/m ] = density of water
3
g [m/s ] = gravity acceleration
Hi [m] = the wave height of an incoming wave (= 2 ai)
a [m] = amplitude of the wave (half the wave height)
d [m] = depth of the breakwater

This can be used for a quick estimate of the upper boundary value of the wave load.

18.2 Linear wave theory

For non-breaking waves against a vertical wall, the force on a wall can be determined using the pressure
distribution in a vertical, taken from wave theory. As mentioned before, according to linear wave theory,
the wave height H in front of the wall:

H = 2H i and with H = 2a : a = Hi is valid.

The maximum pressure against a wall in case of reflection is then:

cosh(k (d + z ))
p = ρ gHi for − d < z < 0
cosh(kd )
 z 
p =  1-  ρ gHi for 0 < z < Hi
 H i 

in which: Hi [m] = wave height of an incoming wave

-1
k [m ] = the wave number of the incoming wave (= 2π/L)

115
Manual Hydraulic Structures Part II - 'Loads'

The force per metre follows from integration over the water depth:

H
0 cosh(k (d + z )) i
 z 
F= ∫ ρ gH i dz + ∫  1-  ρ gHi dz
-d cosh(kd ) 0  Hi 
 ( exp(kd ) - exp(-kd )) H i 
= ρ gH i  + 
 2 k cosh(kd ) 2 

In the case of a large wavelength, the wave pressure approaches the hydrostatic pressure (= rule of
thumb). Figure 18-1 gives an example of this. The figure illustrates the wave pressures for different
wavelengths, which are to be added to the hydrostatic pressure corresponding to the still water level.

Figure 18-1 Rule of thumb, linear wave theory, and Sainflou method: wave pressure

18.3 Sainflou method

In practice another simple approximation is often used for the calculation of the total force on a wall. This
method is known as Sainflou’s method. This approach is shown schematically in Figure 18-2 and only
applies to non-breaking waves. The approach is based on Stokes’ second order wave theory. The
incoming waves have the shape of a trochoid and complete reflection (χ = 1) is assumed. Due to
interference of reflected waves with incoming waves, peaks in front of the wall will reach an amplitude
with the magnitude of the incoming wave height.

The still water level in front of the vertical wall will increase with h0 [m]:

1
h0 = ⋅ k ⋅ H in 2 ⋅ coth ( k ⋅ d )
2
where:
h0 [m] = increase of the mean water level in front of the structure
Hin [m] = height of the incoming wave; not influenced by the presence of the wall
d [m] = water depth in front of the sill, 2 or 3 wave lengths away from the wall
-1
k [m ] = wave number of the incoming wave:
2π 2π
k= or k =
L0 L
L [m] = wave length

For 100% reflection: Hrefl = 2·Hin , where Hrefl is the height of the wave that origins from the interference of
the incoming and the reflected wave.

116
Manual Hydraulic Structures Part II - 'Loads'

Sainflou and Stokes’s second order wave theory lead to the same maximum pressures at mean water
level and near the bed as the linear theory; viz.:

p1 = ρ ⋅ g ⋅ ( Hin + h0 )

ρ ⋅ g ⋅ Hin
p0 =
cosh ( k ⋅ d ' )
where:
d' [m] = water depth above foundation level of the structure.

The pressure between p0 and p1 is assumed to be linear. Therefore Sainflou leads to an overestimation of
the load for steep waves.

Figure 18-2 Sainflou: wave pressure

18.4 Rundgren method

Based on adapted higher order wave theory, Rundgren adapted Sainflou’s formulas. The adapted
formulas were used to make the graphs in the Shore Protection Manual (CERC, 1984). In these graphs,
overtopping and oblique approach are taken into account, which reduces the load. Rundgren’s wave
theory is not covered in this manual.

18.5 Goda method

Goda (1985, 1992) made a general expression for the wave pressure on a caisson on a rockfill sill. This
expression can also be used for broken and breaking waves. Worldwide Goda’s equations are used often
for the design of vertical breakwaters, see Figure 18-3. Goda’s equations don’t have an analytical base
but rather an empirically foundation.

For the determination of the design wave height HD and the design wavelength LD, see the methods in
Chapters 15 and 16. Goda proposed his own formula for HD and LD however, these are not dealt with in
this manual.

117
Manual Hydraulic Structures Part II - 'Loads'

top caisson

top caisson

Figure 18-3 Goda (modified by Tanimoto): wave pressure

The sill height is h - d.

The sill width is Bm.

The maximum wave pressures are:

p1 = 0,5 (1 + cos( β )) (λ1α1 + λ2α 2 cos2 ( β )) ρ g HD

p3 = α 3 p1
p4 = α 4 p1
pu = 0,5 (1 + cos( β )) λ3 α1 α 3 ρ g HD

in which: β = the angle of the incoming wave

η *
= 0,75 (1 + cos( β )) λ1 HD
2
 4π h / LD 
α1 0,6 + 0,5 
 sinh(4 π h / L ) 
=
 D 

 (1 − d / hb )(HD / d ) 2d 
2
α2 = min  , 
 3 HD 
 1 
α3 = 1 − (h ' / h)  1 − 
 cosh(2π h / LD ) 
1
≈ (without sill)
cosh(k d )
hc*
α4 = 1−
η*
hc* = min(η * , hc )
λ1, λ2, λ3 = factors dependent on the shape of the structure and on wave conditions;
(vertical wall and non-breaking waves: λ1 = λ2 = λ3 = 1)
hb = water depth at a distance 5HD from the wall
HD = design wave height (see Chapters 15 and 16)
LD = design wavelength (see Chapters 15 and 16)
d = water depth above the top of the sill
h’ = water depth above the wall foundations plane
h = water depth in front of the sill.

118
Manual Hydraulic Structures Part II - 'Loads'

19. Water, waves, wall, breaking

19.1 Introduction
For the description of conditions in which a wave breaks, see Section 16.3 "Shallows: breaking".
For unbroken waves, the pressure distribution in the wave is a measure for the force on the wall. In the
case of breaking waves this is not so. For those waves it is mainly the velocity with which the water
particles hit the wall that is of importance. The shape of the breaking wave and possible air that is caught
between the structure and the breaking wave largely influence the maximum wave shock and the course
of the pressure distribution in time. The load due to breaking waves is still a point of research. The
dynamic character of the load is an essential facet of breaking waves. Due to the collision between the
wave and the structure a transfer of impulse takes place. At the moment of impact a relatively high
pressure occurs, which only lasts a very short time (in the order of 0,01 s). Because of the short time
span, this pressure is not representative for the stability of a structure (due to the inertia of mass). This
pressure can be of importance for the strength of the structure (partial collapse).
If possible, it is better to prevent waves to break before they hit the structure, so that the shock will be
less. In most cases it is therefore more economical not to place too high a sill in front of a straight wall, so
that the waves won’t break and the load of the non-breaking waves is governing (maatgevend).

The sections below describe three models for breaking and broken waves. These models are:
• Minikin
• CERC 1984, broken waves
• Goda-Takahashi
These models are no more than rough estimates.

19.2 Minikin method

Minikin’s model is based on both laboratory tests and on prototype measurements. Figure 19-1 gives a
diagram of the model. It is based on a maximum dynamic pressure at the still water level and on a
parabolic decline to zero over the distance Hb/2 above and below the still water level plus an increase of
the hydrostatic pressure as a result of the displacement of the water surface.

Dynamic component

Hydrostatic component

Figure 19-1 Minikin: broken wave pressure

The maximum pressure is:

1 H d
pm = Cmk π ρ g b s (D + d s )
2 LD D

where: Cmk [-] = coefficient of the impact ≈ 2

Hb [m] = breaker height
ds [m] = depth in front of the wall
D [m] = depth at one wavelength in front of the wall
LD [m] = wavelength at depth D
Minikin found Cmk ≈ 2.

119
Manual Hydraulic Structures Part II - 'Loads'

The resultant force according to Minikin is:

Pm Hb ρ gHb H b
F= + ( + ds )
3 2 4
Note
Minikin’s method is unfortunately described incorrectly in CERC (1984). In the original publication by Minikin
(1963), the pressure on the wall was expressed in tonnes per square foot. This is not correct. It should be ton force
per square foot. This mistake was overseen in conversion to SI units for the CERC 1984 and has led to a formula for
pm which gives values that are far too large. This is why many publications warn against Minikin’s method,
mentioning that the equation gives values that are 10 to 15 times too large, whilst the original method actually gave
far lower values. One is advised not to use equations derived from Minikin (except for the corrected equations given
above).

19.3 CERC 1984 method

The model of CERC (United States Army Coastal Engineering Research Center) 1984, for broken waves,
merely gives an indication of the load. If accurate estimates are needed of the maximum load on a
structure due to breaking waves, more thorough research must be carried out for the specific situation.
Like Minikin’s model, the model assumes a dynamic and a hydrostatic component of the water pressure
on the structure.

Figure 19-2 CERC 1984: broken wave pressure

The dynamic component is derived from the wave propagation velocity c at the moment the waves started
to break. The broken wave is considered a translation wave with the propagation velocity:

c = gd b .

The dynamic pressure is:

c2 ρ gd b
pm = ρ g =
2g 2

where: db [m] = the water depth where the wave broke

As in Minikin’s model, the hydrostatic component of the load is caused by the displacement of the water
surface. The total load as a result of the broken wave is therefore:

d h 
F = ρ ghc  b + c + d s 
 2 2 

in which: hc [m] = the height of the broken wave (translation wave) = 0,78 Hb

120
Manual Hydraulic Structures Part II - 'Loads'

19.4 Goda-Takahashi method

Goda’s model was already given in the previous chapter. According to Takahashi and others (1994), a
couple of factors need to be adjusted for waves that break on the berm of the sill on top of which a
caisson has been placed:

λ1 = λ3 = 1
 αI 
λ2 = max  1, 
 α2 
where: α I = impulse coefficient

The impulse coefficient is determined with the following equations:

α I = α nα m
 HD 
α m = min  ,2 
 d 

cos(δ 2 )
αn = if δ 2 ≤ 0
cosh(δ1 )
1
αn = if δ 2 > 0
cos(δ1 ) cosh(δ 2 )

δ1 = 20 δ11 if δ11 ≤ 0
δ1 = 15 δ11 if δ11 > 0

δ 2 = 4,9 δ 22 if δ 22 ≤ 0
δ 2 = 3,0 δ 22 if δ 22 > 0

 BM  h−d 
δ11 = 0,93  − 0,12  + 0,36  − 0,6 
 LD   h 
 BM  h−d 
δ 22 = −0,36  − 0,12  + 0,93  − 0,6 
L
 D   h 

where: BM [m] = width of the berm in front of the wall (see Figure 18-3)

The dimensions of the berm have an important influence on the extent of the load. Figure 19-3 shows this
influence for an example.

121
Manual Hydraulic Structures Part II - 'Loads'

Figure 19-3 Influence of a berm on the wave load (HD = 6 m, h = 9 m, h' = 7 m, d = 5 m, hc = ∞)

19.5 Comparison
Figure 19-4 shows a comparison of the different models. The wave load was calculated for various
different wave periods, for a given configuration (a caisson on a quarry stone sill).
[kN/m]

Goda
Minikin
Minikin according to CERC
Total force F

Broken wave CERC

Wave period T [s]

Figure 19-4 Comparison of Minikin, Goda and broken waves according to CERC 1984
(HD = 6 m, h = 9 m, h' = 7 m, d = 5 m, hc = ∞)

The comparison of the different calculation models revealed that there are considerable differences
between the results of the models, particularly when waves with periods between 6 and 15 seconds are
considered. When applying the models, one could consider upper and lower boundaries, where Minikin’s
model is clearly a lower boundary for the above-mentioned periods.

One reason for the deviation between the models could be the different researchers’ ways of measuring
the wave load. The wave force of a breaking wave varies in time. The moment the wave breaks against

122
Manual Hydraulic Structures Part II - 'Loads'

the structure, the impact is the largest, during a very short time span the pressure against the wall is
extremely high. Directly after the impact the pressure decreases very quickly and stays constant a while
at a certain level of pressure. This is shown schematically in Figure 19-5. This schematisation is known
as the “church roof” load. The response of the structure under a “church roof” load depends on the
stiffness and on the inertia of the structure. These determine the speed of reaction of the structure. A
structure with little inertia (mass) and a large stiffness will be more sensitive to short impulse loads than a
structure with a larger inertia.

Figure 19-5 Course of the load in time

The wave load and the response of the structure are still subject to research.

Figure 19-6 shows the force F on a smooth wall for different wave steepnesses (s = H / L), according to
the four given methods. Here, the sill height in the Goda formulas is kept equal to zero, even though, the
Goda formulas were derived for a caisson on a sill.

rule ofvuistregel
thumb

Sainflou

linear wave theory

Goda

Figure 19-6 Wave load F of non-breaking waves (H = 4 m) as a function of the steepness s (h = h’ = d = hb = 15m)

The wave load of a four-metre high wave on the same wall is plotted as a function of the wave period in
Figure 19-7.

123
Manual Hydraulic Structures Part II - 'Loads'

rule of thumb

Sainflou

linear wave theory

Goda

Figure 19-7 Wave load F of non-breaking waves (H = 4 m) as a function of the period T ( h = h’ = d = hb = 15m)

The graphs do not begin at “zero”, so the graphs give a distorted view. The differences between the linear
wave theory and Sainflou or Goda are no more than 10% or 20%. Particularly for long waves (hence not
very steep), the rule of thumb has a small error.

124
Manual Hydraulic Structures Part II - 'Loads'

20. Water, waves, slender structure, non-breaking

Analogous to the formulation of the force on an object in a stationary flow, Morison et al (1950) found a
relation between the force on a vertical pile as a function of the velocity and the acceleration of water
particles in a wave. The part of the force that is caused by the flow velocity is the drag force and the part
that is caused by the acceleration of the water particles is the inertia force.

Figure 20-1 Wave load on a slender structure

20.1 Theory
Morison’s formula for a fixed body in an oscillatory flow includes an inertia part and a drag part:

dF π D 2 du 1
= p(t , z ) = pI + pD = CI ρ + CD ρ D u u
dz 4 dt 2

where: p(t,z) [N/m] = pile force per unit of length of the pile at time t and in position z [N/m]
pI [N/m] = inertia force (per unit of length of the pile)
pD [N/m] = drag force (per unit of length of the pile)
CI [-] = inertia coefficient ≈ 2,0
CD [-] = drag coefficient (for low flow velocities CD ≈ 1,2 , see Section 20.3)
3
ρ [kg/m ] = density of the water
D [m] = diameter of the pile
cosh(k(d + z ))
u [m/s] = horizontal velocity of the water particles: u = ω a sin( ωt)
sinh(kd )
du du cosh(k(d + z ))
= ω 2a cos( ω t)
2
[m/s ] = horizontal acceleration of the water:
dt dt sinh(kd )
ω [rad/s] = angular frequency (2π/T)
k [rad/m] = wave number (2π/L)
d [m] = depth

Because the velocity and the acceleration of the water particles are not in phase, the maximum total force
is smaller than the sum of the maximum drag force and the maximum inertia force.

The force on the pile over time is found by integration:

η η η η η
π D2 du 1
∫ ∫ ∫ pD dz = CI ρ dz + CD ρ D ∫ u u dz
4 −∫d dt
F (t ) = p(t , z ) dz = pI dz +
−d −d −d
2 −d

125
Manual Hydraulic Structures Part II - 'Loads'

The moment on the pile relative to the bed is found by integrating the wave pressure multiplied by the
height on the pile:

η η η
M (t ) = ∫
−d
p(t , z )( z + d ) dz = ∫
−d
pI ( z + d ) dz + ∫p
−d
D ( z + d ) dz

20.2 (Preliminary) design

The maximum force and the maximum moment in the above integrals are generally solved as follows:
π D2 1
Fmax = FI + FD = CI KI H ρ g + CD K D H 2 ρ g D
4 2
Mmax = FI d SI + FD d SD

where: CI [-] = inertia coefficient ≈ 2,0

CD [-] = drag coefficient (for small flow velocities CD ≈ 1,2 , see Section 20.3)
KI [-] = correction for extent of inertia force
KD [-] = correction for extent of drag force
SI [-] = correction for position of resultant inertia force
SD [-] = correction for position of resultant drag force
H [m] = wave height
D [m] = diameter pile
d [m] = depth

The “Shore Protection Manual” (CERC 1984) gives graphs with the maximum values of the coefficients
CD, Ki, KD, Si, SD. These graphs are also included in the following sections. The values of the coefficients
depend on the wave period, the phase, the water depth and the applicable wave theory for the
determination of the velocity of the water particles.
The graphs show various different curves. These depend on the ratio:

H
Hb

in which: Hb [m] = wave height when breaking (see Section 16.3)

Just like a body in a constant flow, a pile in a wave is subjected to a lifting force perpendicular to the
direction of the wave and the axis of the pole. This force is caused by fluctuations of the vortices next to
and behind the pile. Due to resonance of the pile, these forces have been known to be 4,5 times larger
than the drag force. In most cases, the lift force is of the same order of magnitude as the drag force.

For a preliminary design, the maximum lift force can be approximated by:

1
FL,max = CL ρ g D H 2 K D,max
2

in which: CL [-] = lift coefficient

KDmax [-] = maximum value of KD

The lift coefficient CL is unknown, but in most cases:

0 < CL < CD

is valid. This can be used to determine a safe upper bound.

126
Manual Hydraulic Structures Part II - 'Loads'

20.3 CD coefficient (Shore Protection Manual 1984)

127
Manual Hydraulic Structures Part II - 'Loads'

20.4 KD coefficient (Shore Protection Manual 1984)

128
Manual Hydraulic Structures Part II - 'Loads'

20.5 KI coefficient (Shore Protection Manual 1984)

129
Manual Hydraulic Structures Part II - 'Loads'

20.6 SD coefficient (Shore Protection Manual 1984)

130
Manual Hydraulic Structures Part II - 'Loads'

20.7 SI coefficient (Shore Protection Manual 1984)

131
Manual Hydraulic Structures Part II - 'Loads'

21. Water, waves, slender structure, breaking

In a breaking wave the water particles have large horizontal velocities. In this case the accelerations are
small compared to the velocity of the water. Consequently, the inertia force on a slender structure is
negligible relative to the drag force. The maximum force on a pile can therefore be approximated using
the reduced formula:
1
Fmax = FD = CD* K D H 2 ρ g D
2

In this case, the maximum moment on the pile is:

Mmax = FD d SD

Observations revealed that the value of CD* in breaking waves is 2,5 times larger than in non-breaking
waves:

CD* ≈ 2,5 ⋅ CD

Thus, the known values of CD for non-breaking waves can be used (see Section 20 "Water, waves,
slender structures").

Breaking waves in shallow water involve super-critical flow ( CD = 0,7 ). In these conditions, one can
assume the following for the preliminary design of cylindrical piles:

CD* ≈ 2,5 ⋅ 0,7 ≈ 1,75

K D ≈ 1,0
SD ≈ 1,11
H = Hb [-] = height of the breaking wave
d = db [-] = depth at which the wave breaks

Section 16.3 "Shallows: breaking" mentions the following about breaking:

Theoretically the wave breaks at a steepness of H/L = 1/7. The depth also limits the wave height. It has
H
been theoretically deduced that an individual wave will break when ≥ 0,78 (there are also more
d
H
complex formulas e.g. by Miche). However, individual waves with a ratio of = 1,2 have been observed.
d
When calculating breaking for a wave spectrum:

Hs
= 0,4 ∼ 0,5
d

132
Manual Hydraulic Structures Part II - 'Loads'

22. Ice
extended with design rules: February 2015; improved: February 2017

One distinguishes four ways in which ice can exert a load on a hydraulic engineering work. These are:
• Thermal expansion.
• Ice accumulation.
• Collision.
• Ice attachment.

22.1 Thermal expansion

During the freezing process of a layer of ice on water, the thickness of the ice increases without an
increase of the ice area. There is hardly any expansion perpendicular to the direction of growth of the ice.
Therefore, the freezing process itself does not generate any loads.
After the freezing process, changes of the ice temperature can cause loads. This thermal expansion of
the ice can sometimes cause static horizontal loads. This only applies if the ice is restricted in its
expansion. Examples of this are the forces on structures on the banks of frozen lakes and the forces on
the piles of berths as a result of the expansion of the ice between the piles.

Figure 22-1 Thermal expansion.

The thermal expansion of ice is comparable to the temperature load as described in Chapter 23
"Temperature". On average, the coefficient of thermal expansion is about α = 5,5 10 °C .
-5 -1

The expansion causes stresses both in the structure loaded by the ice and in the ice itself. These
stresses create equilibrium in the interface between the structure and the ice. The upper limit of these
stresses is the yield stress of the structure or of the ice.

The yield stress of the ice largely depends on the density and the composition of the ice. The salinity (salt
concentration), the extent of water pollution and the temperature are hence important factors involved. It
is therefore not possible to give a general yield stress of ice. It is very much linked to the location where
the ice was formed and also the temperature history of the ice.

Research into the yield stress of ice has been carried out in several different locations. The yield stress
2
varies between 20 and 300 kN/m . The large spread of the values of the strength show that knowledge of
the local ice strength is of large importance.

An important phenomenon associated with thermal expansion is the buckling of the ice surface in places
where the ice shows irregularities. Thus, a completely new situation is created with respect to thermal
expansion.

22.2 Ice accumulation

As a result of a slight current, ice can accumulate against a structure. A current beneath the ice causes a
shear force along the ice, which is in equilibrium with forces on the structure. Due to the accumulation of
ice against the structure in slowly flowing water, a static horizontal load is created.

133
Manual Hydraulic Structures Part II - 'Loads'

Figure 22-2 Accumulation of ice

22.3 Collision
The dynamic horizontal load on a structure is caused by colliding blocks of ice, which were carried along
by the wind and the current. This is particularly of importance for structures in rivers and along coasts with
a considerable current.

Current

Figure 22-3 Collision of ice floes and breaking on an ice breaker

The impact of large ice sheets that collide against a structure can be compared with the impact of ships
that run into a structure. The fact that the mass and particularly the velocity of the colliding ice sheet
determine the load on a structure follows from the equations in Chapter 29 "Shipping, berthing". Here,
too, the stresses are in equilibrium in the interface between the structure and the ice. If the maximum
yield stress of the ice is exceeded, it will collapse. It is known that the stresses are maximal in the outer
fibres, as a result of bending. Therefore, collapse of the ice takes place sooner in a case of bending than
in case of a uniformly distributed normal force. This knowledge can be used to reduce or prevent forces
on structures. An angled slope or an oblique edge in front of a structure can make the ice buckle and thus
prevents the ice sheet from sliding into the structure.

22.4 Ice attachment

The static vertical load on a structure as a result of the attachment of ice onto a structure is of particular
importance when water levels vary. Ice attachment can occur anywhere between the high and the low
water levels. Under water this mass causes an upward force and above water it results in a downward
load on the structure.

Figure 22-4 Ice attachment

134
Manual Hydraulic Structures Part II - 'Loads'

22.5 Design rules

It should be judged per situation whether ice loading is relevant for the design of hydraulic structures. For
example, there can be ice loading from frozen surface water acting on a quay wall, but at the same time
the ground water on the other side of the wall can be frozen too, mainly or partly eliminating the outer ice
load.

The Dutch Guidelines Design Hydraulic Structures (Richtlijnen Ontwerp Kunstwerken, ROK, 2013) of
1
Rijkswaterstaat prescribe an ice load of 400 kN/m on the walls of lock chambers, at the expected water
level during ice conditions. ROK also specifies the directions and locations of ice loads on gates:
• For thermal expansion: in length direction of lock chambers, 0,20 m below the upstream water level;
• For accumulating ice: perpendicular to gates, at upstream water level;
• For attached ice: in vertical direction, equally distributed over horizontal beams (regels) that are
located under water.

22.5.1 General design rules for determining horizontal ice loads

5
The Dutch Manual Sheetpile Walls (Handboek Damwandconstructies, CUR-publication 166 ) states that
in most cases, ice loads will appear not to be critical for the design of sheet pile walls. For retaining walls
adjacent to open water, the combination of ice load, ship impact and wave load should not be considered.
Either the ice load or the combination of ship impact and wave load is governing. For the estimation of
horizontal ice loads on sheetpile walls, it can generally be assumed that the ice layer has a thickness of
0,50 m and a compressive strength of 1,5 MPa for sea water and 2,5 MPa for fresh water. Because this
load will vary over the contact surface from 0 to the compressive strength of ice, a levelling coefficient of
0,33 may be applied. This leads to the following design values for corresponding line loads, acting at the
most unfavourable water level: 250 kN/m for sea water and 400 kN/m for fresh water (corresponding with
the value mentioned in the ROK).

For the strength calculation of structural members also a concentrated load of 1500 kN should be
6
considered according to CUR 166, if ice could affect these members . In intertidal areas, where ice floes
(schotsen) can origin, a horizontal line load of 100 kN/m can be used and possibly also vertical loads (not
quantified in CUR 166, but below there a design rule is given).

The Ice Handbook for Engineers gives the following equation for static ice load:

Fstatic = kc ⋅ σ c ⋅ D ⋅ h

where Fstatic [kN] = static ice load

kc [-] = degree of confinement (opsluiting) of the ice in front of the structure,
varying from 1 to 3
2
σc [kN/m ] = average uniaxial compressive strength of the total ice thickness
D [m] = width of the structure
h [m] = ice thickness

5
Reference is made to Section 3.2.5 of part 2 of the 5th edition of this publication.
6
This seems a very high value, so for the dimensioning of specific structural members one should judge whether situations can occur that can
indeed inflict such a force.

135
Manual Hydraulic Structures Part II - 'Loads'

22.5.2 Horizontal loads caused by thermal expansion of ice

According to Korzhavin (1962), if an ice cover is prevented from expanding (e.g. ice between two bridge
piers) the ice will exert a maximal horizontal linear load of:
1/3
h2 (0,35 ⋅ Ti + 1)5/3  ∆T 
Fmax = 54,13 ⋅ψ ⋅ ⋅ ⋅ i 
h + 9,1⋅ hs Ti 0,22  ∆t 
where:
Fmax [kN/m] = maximum force per running metre of ice cover (linear loading)
h [m] = ice thickness
hs [m] = thickness of snow layer on the ice
Ti [°C] = average air temperature of the previous 24 hours

∆Tl [°C] = maximum rise in air temperature within a given period ∆t within the preceding
24 hours
∆t [h] = period during which the temperature is rising
ψ [-] = coefficient to include the effect of the length (L) of the ice-cover:
ψ = 1,0 for L < 50 m
ψ = 0,9 for 50 < L < 75 m
ψ = 0,8 for 75 < L < 100 m
ψ = 0,7 for 100 < L < 150 m
ψ = 0,6 for L > 150 m

CUR 166 mentions line loads coming from thermal expansion of enclosed ice. The magnitude of these
loads depends on the initial temperature of the ice, the temperature gradient and the thickness of the ice
[m], see Figure 22-5.

Figure 22-5 Loads caused by thermal expansion of ice (CUR 166)

It should additionally be checked whether this force Fmax does not exceed the failure compressive
strength of the ice (see below for the compressive strength of ice).

22.5.3 Compressive strength of ice

The German manual on quay walls, the Empfehlungen des Arbeitsausschusses "Ufereinfassungen"
(EAU, 2012) gives rules to calculate the compressive strength of ice and measured maximum thickness
of ice layers (in Germany, varying from 0,35 to 0,80 m).

136
Manual Hydraulic Structures Part II - 'Loads'

According to the EAU 2012, concentrated ice loads on vertical piles can be calculated with:

Pp = k ⋅ σ 0 ⋅ d ⋅ h1,1

where: Pp [kN] = concentrated ice load (force)

2
σ0 [kN/m ] = uniaxial compressive strength of ice related to a specific expansion
-1
rate ε = 0,001 s
d [m] = width of the pile
h [m] = thickness of the ice
0,4
k [m ] = empirical coefficient; for floating ice: k = 0,564;
for ice frozen onto the pile: k = 0,793

The horizontal uniaxial compressive strength of ice, σ0 , can be calculated according to empirical
equations as found by Austin Kovacs (1996) (adopted by the EAU 2012):

2,7 ⋅ 106 ⋅ ε 0,33

σ0 =
ϕB
and
−0,69
ϕB = 19,37 + 36,18 ⋅ SB0,91 ⋅ ϑm

2
where σ0 [kN/m ] = uniaxial compressive strength of ice
φB [m‰] = porosity of the ice layer
-1 -1
ε [s ] = specific expansion rate = 0,001 s
SB [m‰] = salinity of the ice
(usual values are: < 1 ‰ for fresh water, and 35 ‰ for sea water)
ϑm [°C] = temperature in the middle of the ice layer

22.5.4 Vertical ice loads

Vertical loads on single piles due to attached ice on rising or falling water levels can be calculated
according to the Russian national standard SNiP 2.06.04.82 (1995):

 0,15 ⋅ D 
AV =  0,6 +  ⋅ 0,4 ⋅ σ 0 ⋅ h
2

 h 

where: AV [kN] = vertical ice load on a pile

h [m] = thickness of the ice layer
D [m] = diagonal of the pile
2
σ0 [kN/m ] = compressive strength of the ice layer

For piles being part of a pile group, or piles next to fixed structures, a geometrical reduction factor fg may
be applied to AV (see the EAU or SNiP for details).

22.6 Literature
• CUR 166: Handboek Damwandconstructies, CUR, 2012
• EAU, Empfehlungen des Arbeitsausschusses "Ufereinfassungen", Springer Verlag, 2012
• Ice handbook for engineers, version 1.2. Luleå Tekniska Universitet - Institutionen för
samhällsbyggnad. Luleå, 2009
• Korzhavin, K.N. Action of ice on engineering structures, Novosibirsk, Akad. Nauk. USSR 1962.
(Translation: TL 260 of CRREL Hanover, New Hampshire, 1971)
• ROK, Richtlijnen voor het Ontwerp van Kunstwerken. Rijkswaterstaat 2013
• SNiP 2.06.04.82 Нагрузки и воздействия на гидротехнические сооружения (волновые, ледовые и
от судов). Russian standard 'Loads and actions on hydraulic engineering structures (wave and ice
generated and from ships). Moscow, 1995.

137
Manual Hydraulic Structures Part II - 'Loads'

23. Temperature

23.1 General
Temperature changes of structures and parts of structures usually lead to deformations and stresses. The
cause of the temperature change can be both internal and external. Examples of external causes are:
• accumulation of warmth due to solar rays (daytime)
• loss of heat due to warmth emission (night time)
• adaptation to a changing air temperature
• cooling due to wind
• cooling due to precipitation
• cooling due to the evaporation of water in or on the structure
• warming or cooling due to activities in the structure (heating, air conditioning)

An internal cause of temperature change is the development of heat during the setting and hardening of
concrete (hydration heat).

The temperature spread in a cross section of a structure element depends on the heat flow through the
element. Some examples of possible temperature profiles are shown in Figure 23-1.

uniform temperature non-linear non-linear

no heat transport one-sided heating of the structure one-sided cooling of the structure

linear non-linear non-linear

stationary heat transport from 1 to 2 two-sided heating of the structure two-sided cooling of the structure

Figure 23-1 Temperature profiles

138
Manual Hydraulic Structures Part II - 'Loads'

23.2 Unobstructed thermal deformation

23.2.1 Temperature change

When the temperature of an unobstructed object changes its length increases or decreases linearly
according to:

∆ℓ
ε = = α ∆T
ℓ

in which: ε [-] = the strain (rek)

∆ℓ [m] = the change in length
ℓ [m] = the length at the reference temperature
-1
α [°C ] = the linear expansion coefficient
∆T [°C] = the change in temperature

The linear expansion coefficients of a number of materials are given below:

1,0⋅10 °C
-5 -1
Concrete
1,2⋅10 °C
-5 -1
Steel
5,5⋅10 °C
-5 -1
Ice
Table 23-1 Linear coefficients of expansion

It is noticeable that the linear expansion coefficients of concrete and steel are almost equal. This is useful
because otherwise temperature fluctuations in reinforced concrete would create more problems. Figure
23-2 shows an example of the elongation of a simple beam due to a temperature increase. The shape of
the beam is not relevant for the linear expansion.

Figure 23-2 Increase in the length of a beam due to a total rise in temperature

After a temperature change, one can determine the position of all points on an arbitrary object relative to
a fixed point because the shape of an object is not of importance. An example is given in Figure 23-3.

Figure 23-3 Change of the area of a sheet due to a total rise in temperature

139
Manual Hydraulic Structures Part II - 'Loads'

23.2.2 Temperature gradient

The preceding applies for a total temperature change (of the entire body). However, it is also possible that
a temperature gradient occurs in the body. Figure 23-4 shows an example of this. There is a temperature
rise on the underside and a temperature decrease on top. The consequence is a shortening on the upper
side and a lengthening at the bottom of the beam, causing a curvature of the beam.

Figure 23-4 Temperature gradient in a beam

The curvature of the beam is equal in all places along the beam and amounts to:
ε −ε α ∆T
κ = b o =−
h h
The angle of rotation of the beam is found by integrating the curvature over the length:
x
ϕB = ϕ A + ∫ κ dx = ϕ A + κ x
0
The deflection is found by integrating the angle of rotation:
x
 x
 1
fx = ∫  ϕ A + ∫ κ dx  dx = ϕ A x + κ x 2
0 0  2
Using the boundary conditions at the two supports one finds:
1 α ∆T ℓ
ϕA = − κ ℓ =
2 2h

The deflection in the middle is therefore:

1 1 1 α ∆T ℓ 2
fm = − κ l 2 + κ ℓ 2 = − κ ℓ 2 =
4 8 8 8h

In the case of a plate instead of a beam, the deflection must be considered in two directions. If the plate
can deform unrestrainedly, it will rest on the corners if it bulges out.

For a temperature change of the beam, in which the temperature change is in the same direction both on
top and in the bottom, one can distinguish an average temperature change and a temperature gradient.
The change in length and the deflection can be calculated using the equations derived for the
deformations.

The preceding, assumed a linear course of the temperature over the height of the structure. In reality, the
course of the temperature is often not linear but curved. In such a case, the temperature course is divided
into an average temperature, a linear temperature gradient and a so-called characteristic temperature.
The linear temperature gradient is chosen such that the remaining characteristic temperature does not
cause any deformations. An example is given in Figure 23-5. As regards the characteristic temperature,
the following applies:

h h

∫ (Te − T0 ) dz = 0
0
and ∫ (T
0
e − T0 ) z dz = 0

in which: h [m] = height of the structure

Te [°C] = characteristic temperature
T0 [°C] = reference temperature

140
Manual Hydraulic Structures Part II - 'Loads'

developed temperature profile Linear components Average temperature Temperature gradient residual temperature variation

Figure 23-5 Arbitrary temperature distribution

The characteristic temperature does not cause any resulting deformations, but it does cause stresses.
For a construction material such as concrete, one must take these stresses into account. This is of
particular importance for the two-sided cooling of a thick structure during and after the setting of the
concrete (hydration heat), causing tensile stress in the concrete.

23.3 Restrained thermal deformation

23.3.1 Temperature change

If the deformations of a structure due to temperature changes can take place without restraints, no
stresses occur in the material. Generally, however, the deformation is restrained. This is usually the result
of the boundary conditions of the supports, which in turn are determined by possible adjoining structure
elements and the surroundings. As an example, consider a structure buried in the ground. In the case of
restrained deformation, the total deformation equals the unrestrained temperature deformation minus the
deformation resulting from the stress increase.

The restriction of the linear expansion of a beam can be schematised as a beam with a spring in the roller
bearing (see Figure 23-6).

Figure 23-6 Restrained deformation due to temperature change

The elongation of the beam results in a force in the spring equal to:

F = k ∆ℓ

where: k [kN/m] = the stiffness of the spring

The occurring deformation is equal to the unrestrained temperature deformation minus the linear elastic
deformation:
Fℓ
∆l = α ∆T ℓ −
EA
This leads to the force F:

F Fℓ α ∆T ℓ
= α ∆T ℓ − ⇒ F=
k EA 1 ℓ
+
k EA
If k = EA , 50% of the unrestrained deformation occurs. If k > 100 ⋅ EA , a fully restrained deformation is
involved.

141
Manual Hydraulic Structures Part II - 'Loads'

23.3.2 Temperature gradient

The restraint of the rotation of a beam due to the supports can also be schematised by a rotation spring
(see Figure 23-7). The moment in the bearing then equals M = c ⋅ ϕ

Figure 23-7 Restrained rotation

The rotation in the bearing equals the unrestrained temperature deformation minus the linear elastic
deformation:
α ∆T ℓ M ℓ
ϕ= −
2h 3 EI

-1
where: α [°C ] = linear expansion coefficient
h [m] = height of the beam

The moment is therefore:

M α ∆T ℓ M ℓ α ∆T ℓ
= − ⇒ M=
c 2h 3EI 1 ℓ 
2h + 
 c 3EI 
23.3.3 Longitudinal direction
If the deformations of a plate are restrained, one must take the deformation in all directions into account.
Close to the supports of the plate, where the deformation is restrained, shrinkage can cause a
considerable tensile stress. This tensile stress is parallel to the supports. This is why concrete slabs
should contain sufficient reinforcement in the direction perpendicular to the span (longitudinal
reinforcement). The tensile stress will propagate further into the field as the slab is wider (see Figure 23-
8).

Figure 23-8 Effect of the width of a slab

Note
For compound structures one has to take in account that different temperatures create different expansions. For
example an oil pipe with warm oil on a (cold) jetty needs an expansion loop.

142
Manual Hydraulic Structures Part II - 'Loads'

24. Soil - loading and stresses

The effective soil stress is often the most important load on a structure, next to water pressure. After all,
the effective soil stress is the only load on the entire structure that adjusts itself (up to a certain point) to
the other loads: It resists the forces acting on the soil, creating equilibrium. Unless the ground collapses
(actively or passively), the ground pushes back, both in normal direction and in shear direction, as much
as the structure pushes the ground.

The effective stress load caused by the loads on the structure is calculated by calculating all other loads
and considering the effective stress as the balancing item. One should not forget to check whether this
load will ever exceed the maximum compression stress and the maximum shear stress of the soil during
any construction or operational phases. For the maximum soliciting stresses, see Chapter 37 and for the
bearing capacity of the soil see Chapter 32.

One must also remember to check that this load never causes the maximum displacement (and
particularly the displacement difference) to be exceeded during any construction or operational phases.

If the maximum forces or displacements are likely to be exceeded, the design of the structure should be
adjusted. The calculations of the loads then start anew and the design is therefore an iterative process.

This chapter describes the distribution of loads in soil in vertical and horizontal direction. The resistance,
or strength, of soil against loads is described in part III of this manual (Chapter 32).

24.1 Vertical soil stress

The exception to the rule given above, which states that the effective pressure can only be calculated as
a balancing item of the other loads, is the vertical effective pressure on top of a structure. The vertical
loads occur on horizontal and diagonal planes of structures in the ground, e.g. cellars and tunnels. The
total load corresponds to the weight of the ground. One has to divide total stresses into inter-granular
stresses and water pressures (see Chapter 27 "Soil, groundwater").

The vertical effective pressure (load) for a soil system with n dry layers and m wet layers can be
determined according to:

n m
σv ' = σv − p i.e.: σ v ' = ∑ γ d ,i d i + ∑ γ n, j d j − p
i =1 j =1

2
in which: σ'v [kN/m ] = vertical inter-granular stress (= effective pressure)
2
σv [kN/m ] = total vertical stress
γ d ,i 3
[kN/m ] = dry volumetric weight of soil layer i : γ d ,i = ρd ,i g
γ n, j 3
[kN/m ] = wet volumetric weight of soil layer j : γ n, j = ρn, j g
3
σ [kg/m ] = volumetric mass of a soil layer
di [m] = thickness of soil layer i above the considered plane
n [-] = number of dry layers above the considered plane
m [-] = number of wet layers above the considered plane
2
p [kN/m ] = water pressure in the considered plane

The interaction between the structure and the ground is different from the interaction between the
structure and the groundwater. The groundwater does not have any shear stiffness and the groundwater
pressures do not depend on possible deformations of the structure. This is different for the soil. In soil it is
possible that deformations of a structure or the surrounding ground cause a redistribution of the total
stresses. For instance, in a tunnel of limited dimensions arching can occur in transverse direction, thereby
more or less relieving the roof of the tunnel of stresses. A stiff structure can also be subjected to a larger
load if there are settlements of the surrounding ground, because the surrounding ground is “suspended”
from the structure.

143
Manual Hydraulic Structures Part II - 'Loads'

arching
arching

deflection

soil layer sensitive to

settlements

Figure 24-1 Arching (Left Ka < 1, right Kp > 1)

24.1.1 Pressure under a structure / surcharge

Loads on ground level result in a change of the stresses in the subsoil. The horizontal and vertical earth
pressures on a structure will consequently increase. There are various models to calculate the increase of
the earth pressure as a result of a load on ground level. The best known are models by Boussinesq,
Flamant and Newmark. For these methods the reader is referred to the lecture notes Soil Mechanics (A.
Verruijt / S. van Baars).

The ratio between the stiffness of the structure and the stiffness of the soil has an important influence on
the soil pressure distribution under the structure. There are several models or schematisations in use for
soil stiffness. Depending on the model, the influence on soil pressure distribution, especially the
spreading of pressure, is either less or more significant, see below in this Section under 'stress
distribution'. Generally, the Winkler model is used for soil stiffness, which models the soil as a system of
mutually independent vertical springs with stiffness k.

Considering the stiffness ratio soil-structure, an extreme case is to assume an infinitely stiff structure. This
assumption implies that the foundation surface remains level. In combination with the Winkler soil
stiffness model the calculation of the soil pressure distribution is analogous to the calculation of a cross-
section of a beam with normal and shear forces, and bending moments (For a foundation ΣV and ΣH are
the equivalent of normal and shear force respectively; the resulting overturning moment ΣM is the
equivalent of the bending moment). Therefore the stress in point (x,y) under an infinitely stiff structure, the
whole contact surface remains under pressure, is:

V My M
σ x,y = + x+ x y
A Iy Ix

in which: Iy = the moment of inertia of the foundation around the y-axis

Ix = the moment of inertia of the foundation around the x-axis
A = the area of the foundation = b·h
V = the vertical load
My = moment in y direction relative to the centre of mass of the foundation
My = V e x
Mx = moment in x direction relative to the centre of mass of the foundation
Mx = V ey

144
Manual Hydraulic Structures Part II - 'Loads'

Figure 24-2 Stiff structure on a shallow foundation

Since the entire foundation surface remains under pressure, the working line of the resulting force lies
within the core of the foundation surface. The dimensions of the core of the surface can easily be
determined using:

Wb W W W
eb = ; eo = o ; el = l ; er = r
A A A A

in which: eb, eo, el, er: dimensions of the core according to Figure 24-3
I Iy
Wb = section modulus x Wl = section modulus
yb xl
Ix Iy
Wo = section modulus Wr = section modulus
yo xr

core core

Figure 24-3 The cores of a square and a circular cross-section

If the resultant of the loads lies outside the core of the foundation surface, the previous equations for the
stresses in the foundation plane do not apply.

Large eccentricities, the resulting force has a working line outside the core of the foundation surface,
considerably complicate the calculation of soil pressure distributions under the foundation.

A well-known solution for a rectangular foundation assumes a triangular soil pressure distribution for the
part of the base that remains in contact with the soil. For equilibrium of forces (ΣV=ΣH=ΣM=0) the
working line of the resultant of the soil pressure should be the same line as the one for the resultant of the
loads. Using the position of the centre of gravity for a triangle, one-third of the base, the maximum
(required) soil pressure can be determined easily (see Figure 24-4).

145
Manual Hydraulic Structures Part II - 'Loads'

½B ½B

Fv

Fv = Fresult = 1
2 ⋅ σ 'max,soil ⋅ 3 a ⇒

BASE
2 ⋅ Fv
σ 'max,soil =
3a

Figure 24-4 Pressures in the case of a very eccentric load

If the structure is not infinitely stiff relative to the soil, use of the Winkler model for the soil stiffness only
yields approximate results. For finite stiffness of the structure, relative to the soil stiffness, the way to
determine the soil pressure distribution more accurate is to use the theories on elastic or elasto-plastic
half spaces. Reference is made to the work of Hetenyi – “Beams on elastic foundation” (1946).

Vertical stress distribution under a structure or top load

A stress at ground level will spread in depth due to the shear stresses in the soil. The course of the
stresses was solved by Flamant. The spread of the stress somewhat resembles a Gaussian curve.
Because this is difficult in calculations, for a preliminary design one can spread the stress under an angle
of 1:1 for stiff soil layers (as shown in the figure below), 2:1 (vertical : horizontal) for medium stiff layers
and 1:2 (vertical : horizontal) for stony materials and granulates.

b0

q
x
45 º
d
clay b1
h

sand

z
Figure 24-5 Simplified stress spread in depth

The contributing width b1 and length ℓ1 in case of a spread of 1:1 then are:

b1 = b0 + 2 d and: ℓ1 = ℓ 0 + 2 d

The average increase in stress in the clay layer due to the load is then:

b0 ℓ 0
∆σ ' = q
b1ℓ 1

The strain in the clay layer is subsequently calculated with: ∆σ ' = Eoed ∆ε (Hooke) or, even better, with
the theory developed by Koppejan.

146
Manual Hydraulic Structures Part II - 'Loads'

The total settlement u then follows from:

u = h∆ε

Notes
• Computer programmes are faster at calculations than man and therefore do not use this simplified
stress spreading method, but methods such as Flamant.
7
Flamant's solution is based on one of Boussinesq’s solutions (point force on an infinite half space ); which
is why computer programmes often state the stress spreading method is based on Boussinesq. Flamant
(1892) found a solution for the stresses in an infinite half-space subjected to a line load. On the basis of
superposition, this solution can be transformed to a solution for a strip-shaped load.
a a
q
x

θ1 θ2

r1 r2

z
Figure 24-6 Strip load (Flamant)

The stresses in an arbitrary point in this homogenous half-space are:

q
σ 'zz = v (θ1 − θ 2 ) + sinθ1 cosθ1 − sinθ 2 cosθ2 
π
qv
σ ' xx = (θ − θ ) − sin θ1 cosθ1 + sinθ 2 cosθ 2 
π  1 2
qv
σ ' xz = cos2 θ 2 − cos2 θ1 
π 

7
A half-space is either of the two parts into which a plane divides the three-dimensional (Euclidean) space

147
Manual Hydraulic Structures Part II - 'Loads'

24.2 Horizontal soil stress

When determining the total horizontal soil stress acting on a structure, the (ground)water pressure (p) and
the effective vertical soil pressure (σ') should be considered separately:

σ soil ,h = σ h ' + p

The reason for this is that the magnitude of water pressure at a certain depth is the same in all directions
(law of Pascal), but this does not apply to soil pressure:
• The horizontal water pressure p at a certain depth is equal (in magnitude, not in direction) to the
vertical groundwater pressure at that same depth. It can be calculated according to Section 27.1.
• The relation between the horizontal and the vertical effective pressure is usually assumed to be
constant:

σ h ' = K ⋅ σv '

There are three types of soil behaviour under stress. The first type occurs when soil is at rest, so when it
does not deform due to deflection or sliding aside of a structure. This state of stress is called 'neutral
stress'. The second type of soil behaviour under stress occurs when the soil becomes less compacted
than at rest, due to, e.g., deflection of a sheetpile wall, or the sliding aside of a gravity structure away from
the soil body. This is called 'active soil stress'. The third type of soil stress, the 'passive stress', occurs
when the soil is compressed due to movement of a wall in the direction of the soil.

Bear in mind that if the soil retaining structure cannot deform or displace, or if displacements are not
tolerable, neutral stresses should be used in design calculations. If displacements will not or cannot be
calculated, it cannot be checked whether these displacements would be acceptable. Also in that case
neutral stresses should be used, leading to conservative designs.

The three types of soil behaviour are explained in the following sections.

24.2.1 Neutral soil stress (Jáky's theory)

Hungarian geotechnologist Jószef Jáky studied the relation between vertical and horizontal soil at rest.
He assumed a constant relation between the horizontal and the vertical effective soil pressure. The
coefficient expressing this relation is unknown and difficult to measure. However, for soil at rest (= without
deformation) one often uses Jáky’s formula to simplify reality:

σ h' = K 0 ⋅ σ v' with: K 0 = 1 − sin(φ )

where:
σ'h [Pa] = horizontal soil pressure
σ'v [Pa] = vertical soil pressure
K0 [-] = neutral soil pressure coefficient
φ [°] = angle of internal friction

This formula was first published in 1944 and still seems to be relatively accurate for the limit values. It
works well for water (φ = 0°, so K0 = 1,0) and seems to work fairly well for sand (φ = 30°, so K0 = 0,5).
This formula is very popular, for want of anything better, but not very scientific. Overloaded soil generally
shows a far higher value. However, the accuracy of Jáky's neutral soil pressure coefficient is often
disputed, but at least it gives values between 0 and 1, and leads to conservative results, so it is still being
used. A study in 2005 even lead to the conclusion that it is surprising that the equation is a good
representation of the true stress ratio in soils at rest (Michalowski, 2005 Journal of Geotechnical and Geo-
environmental engineering ASCE).

Large shallow foundations expand due to temperature changes, so in these cases one must take a larger
pressure on the vertical walls into account than predicted by Jáky. In this case:

σ h' ,n = K 0 ⋅ σ v' with: K 0 ≈ 1

Eurocode 7 also takes the over-consolidation ratio (OCR) into account for the horizontal soil pressure.
The over-consolidation ratio is the relation between the original effective vertical soil pressure and the

148
Manual Hydraulic Structures Part II - 'Loads'

actual effective vertical soil pressure at the same depth. The neutral soil pressure coefficient should then
be calculated according to:

K 0 = (1 − sin φ ) ⋅ OCR

This equation should not be used for 'very high values of OCR'.

Notes
• If a structure is sensitive to seasonal temperature changes (resulting in stress increment in the
structure), a soil pressure coefficient of K = 1,0 should be used under the following two conditions:
- The horizontal effective pressure on the wall of a tunnel or open excavation acts unfavourably;
- The bottom of the structure is wider than 15 m.
• For the variable load at ground level next to an open excavation or an excavated building site, an
evenly distributed surface load of at least 20 kN/m2 or VOSB class 45 should be used.

24.2.2 Active and passive soil stress (Rankine's theory)

Soil under active stress develops less horizontal soil pressure than soil under passive stress. In principle,
the horizontal soil pressure is undetermined; only a lower and an upper limit can be given, corresponding
to two types of ground collapse: active or passive deformation. Figure 24-1 shows the various types of
soil deformation due to movement of walls or structures.

Figure 24-1 Soil deformation due to rotation, displacement or deflection of a wall or structure

Using Mohr’s circle, William Rankine, a Scottish civil engineer, found a definition for the maximum and
minimum horizontal effective stresses with a given vertical pressure. When a straight wall without shear
stress collapses, a straight slip plane develops, see Figure 24-2.

Figure 24-2 Slip planes caused by displacement of a retaining wall (left: theoretical; right: empirical)

The maximum and minimum horizontal effective stresses for a straight wall without shear stresses can be
derived with use of trigonometry:

149
Manual Hydraulic Structures Part II - 'Loads'

1 − sin ϕ '
Ka = coefficient of active soil pressure [-]
σ h' ,min = K a σ v' 1 + sin ϕ '
with
σ h' ,max = K p σ v' 1 + sin ϕ '
Kp = coefficient of passive soil pressure [-]
1 − sin ϕ '

If the soil is non-cohesive (like sand), c = 0 which simplifies the equation.

The value of the real horizontal effective soil pressure lies somewhere between the maximum and
minimum pressure and depends mainly on the horizontal displacement of the wall. The horizontal soil
pressure develops gradually between the two limit states, see Figure 24-3.

Figure 24-3 Horizontal soil pressure versus deformation of the soil

The horizontal soil pressure can be influenced by cold flow, temperature changes and the loading history.

Please bear in mind that in reality, Kp is limited to a maximum value of about 7. More accurate, especially
for Kp, are the equations of Fritz Kötter, who assumed curved slip planes. These equations can be found
in the user manual of D-Sheetpiling, or in the original publication of Kötter (1903).

24.2.3 Horizontal soil pressure on a vertical wall including friction

To include the effect of friction at the soil/wall interface, the equations for the active and passive soil
coefficients can be extended according to Heinrich Müller-Breslau:
2
 
2
 
   
cos ϕ  and K =  cos ϕ 
K a =     ,
sin(ϕ + δ a ) ⋅ sin ϕ  1 − sin(ϕ − δ p ) ⋅ sin ϕ
p
 1+  
 cos δ a   cos δ p 
   

where δ represents the angle of wall friction, normally ½φ < δ < φ and δa taken positive and δp negative.
These equations apply to horizontal ground surfaces and a vertical wall.

24.2.4 Horizontal soil pressure on a vertical wall including cohesion

In case of cohesive soil types, the horizontal soil pressures can be calculated according to:

σ h' ,min = K a,h,σ σ v' + K a,h,c c

σ h' ,max = K p,h,σ σ v' + K p,h,c c

in which: Ka,h,σ = active coefficient for the effective stress

Ka,h,c = active coefficient for the cohesion
Kp,h,σ = passive coefficient for the effective stress
Kp,h,c = passive coefficient for the cohesion

150
Manual Hydraulic Structures Part II - 'Loads'

24.2.5 Horizontal pressure on an oblique wall (hellende wand)

If the ground level behind the wall deviates from the horizontal plain with an angle β, the horizontal soil
coefficient for soil at rest may be calculated according to Eurocode 7:

K0,β = K0 ⋅ (1 + sin β )

Müller-Breslau found the following equations for active and passive horizontal soil stresses for inclined
walls and inclined ground levels:

cos2 (φ + α )
K a,h,σ = 2
 sin(φ + δ )sin(φ − β ) 
cos2 (α )  1 + 
 cos(α − δ )cos(α + β ) 
cos2 (φ − α )
K p,h,σ = 2
 sin(φ − δ )sin(φ + β ) 
cos (α )  1 −
2

 cos(α − δ )cos(α + β ) 
2cos(φ )cos( β )(1 − tan(α )tan( β ))cos(α − δ )
K a , h, c =−
1 + sin(φ + δ − α − β )
2cos(φ )cos( β )(1 − tan(α ) tan( β ))cos(α − δ )
K p , h, c =
1 − sin(φ − δ + α + β )

in which: α = the obliqueness of the structure (see Figure 24-4)

β = the angle of the ground level (see Figure 24-4)
φ = angle of internal friction
δ = the angle between the resultant force exerted on the soil retaining wall and the normal
to this wall (see Figure 24-4)) (assumption δ ≈ 0,8 to 0,9 ⋅ ϕ )

Figure 24-4 Definition of angles (be careful: in some literature these angles are defined differently in combination with also
differing equations. The combination leads to the same results, just don't mix up the combination of angle definitions and
equations!)

The aforementioned formulas assume straight slip planes. For angles of internal friction (φ) higher than
30°, curved slip planes are more realistic (the theory of curved slip plane is not discussed in this manual).
If nevertheless the Müller-Breslau equations are used for these higher values of φ, the resulting values for
Kp,h,c are unrealistically high. In reality, Kp,h,c is limited to a maximum value of about 7. This also applies to
the simplified equations.

151
Manual Hydraulic Structures Part II - 'Loads'

24.2.6 Infinite surface load next to a wall

In the case of a surface load q that reaches infinitely far in two directions, the additional horizontal load on
an oblique wall is:
cos(α )cos( β ) cos(α )cos( β )
∆σ h' ,min = K a,h,σ q ∆σ h' ,max = K p,h,σ q
cos(α + β ) cos(α + β )

Figure 24-5 Infinitely extended surface load

For vertical walls, where α = β = 0°, these equations reduce to:

∆σ h' ,min = K a,h,σ q ∆σ h' ,max = K p,h,σ q

24.2.7 Finite surface load next to a wall

If the load is merely exerted on a strip along the side of the soil retaining structure, the horizontal load on
the structure is smaller at the bottom of the wall. An approximation of the horizontal load on the soil-
retaining structure is given in Figure 24-6. In this case one assumes that the surface loads spreads to a
depth b that is determined by the active slip plane that could occur behind the surface load. It is also
assumed that the load reduces to zero over a height of ½b.

Figure 24-6 Load on a strip next to the soil retaining structure

The influence at a depth b is calculated using angle ϑa for active soil and ϑb for passive soil:

(1 + tan2 ϕ ) ⋅ tan ϕ (1 + tan2 ϕ ) ⋅ ( − tan ϕ )

tanϑa = tan ϕ + and tanϑp = − tan φ +
tan ϕ + tan δ − tan ϕ + tan δ

By approximation the additional horizontal load ∆σ 'h is ∆σ 'h = p1 = q K a,h,σ

152
Manual Hydraulic Structures Part II - 'Loads'

24.2.8 Infinite surface load at a distance of a wall

In the case of a very extensive load at some distance from the soil retaining structure, the top of the
structure will have a smaller load. An approximation of the horizontal load on the structure is given in
Figure 24-7.

Figure 24-7 Extensive load at a distance from the structure

In this case, too, the maximum horizontal soil pressure is: p2 = q K a,h,σ . As a rule-of-thumb, the following
angles could be assumed: ψ = 45° and ϑa = 45° + ½φ.

24.2.9 Finite surcharge load at a distance of a wall

One can approximate the horizontal load due to a uniformly distributed vertical load at some distance
from the soil retaining structure according to Figure 24-8.

Figure 24-8 Strip load at some distance from the structure

2q s θ
In this case, the maximum horizontal earth pressure caused by the surface load is p3 =
c −a
sin(ϑa − ψ )cos(δ )
where: θ =
cos(ϑa − ψ − δ )
s = the width of strip on which the load is acting
a, c = see Figure 24-8
δ = the angle between the resultant force exerted on the soil retaining wall and the normal
to this wall (see Figure 24-4))

153
Manual Hydraulic Structures Part II - 'Loads'

24.2.10 Horizontal stress distribution next to a finite surface load

The stress on a retaining wall next to a surface load at a certain distance of that wall may be assumed to
vary over angles of ±45°, like depicted in Figure 24-9. It gradually decreases further aside, until the
retaining wall does not experience any influence of the surface load any more.

Figure 24-9 Top view of a situation with horizontal stress distribution next to a finite surface load (adapted from EAU 2012)

24.2.11 Horizontal soil pressure on piles

For a three-dimensional situation, as for piles, one can increase Kp and the cohesion to include the effect
of shell behaviour. The Danish structural engineer Jørgen Brinch Hansen, who was specialised in soil
mechanics, developed one method for this in his book “The ultimate resistance of rigid piles against
transversal forces”. The 3-D values for the coefficient of passive earth pressure and the cohesion as a
result of shell behaviour is included in the computer program D-sheetpiling. The calculation is too
complicated to discuss in this Manual.

154
Manual Hydraulic Structures Part II - 'Loads'

25. Soil, settlement

Settlements (zettingen) occur as a result of raising the effective stresses in the subsoil. For instance, this
could be by placing a foundation or by lowering the groundwater level. Settlements as such are usually
not a problem for a structure, as long as the settlements are even. Only for flood defences the lowering of
the crest due to settlements should be taken into account (regarding critical overtopping discharges), also
when the settlement is even.

Uneven settlements can create large stresses in a structure. There are several causes for this:
• Uneven foundation pressure
• Different sized foundation elements (even if the pressure is equal!). Wide elements will settle more
than narrow elements with an equal foundation pressure because the stress in the heart of the
structure is spread out less and thus works to a larger depth.
• Local deviations from the soil properties. Locally the soil may be stiffer or less stiff, which causes
uneven settlements for even loads.
• Changes of the horizontal stresses due to a nearby structure excavation or a (bored) tunnel.

According to NEN 6740, article 5.2.2.2, no span (foundation beams!) may undergo a rotation due to
settlement differences larger than:

∆z
= θ = 1/ 300
ℓ

For shallow foundations, the settlement can be calculated according to Chapter 34 (Part III). For pile
foundations, the above requirement is usually met. To check the design of pile foundations, see Chapter
38 (compression piles) and the lecture notes on deep foundations (part of the course material for
"Hydraulic Structures 1").

For excavations (building site) the consequences of the swell of weaker soil layers located at a larger
depth must also be taken into account. For excavations (building site), a pipe located nearby could
undergo a rotation (see figure below). This can also be tested using the rotation condition given above.

Figure 25-1 Displacement and rotation of a pipe near a building site excavation (top view)

If one expects problems due to excessive settlement differences, one can take the following measures:
• Using equal foundation pressures.
• Using equal-sized foundation elements.
• Applying soil improvement measures.
• Demanding stricter requirements for the structure excavation or (bore) tunnel.
• Applying pile foundations.
• Introducing movement joints.
• Reducing the distance between joints.
• Changing the statically indeterminate structure into a statically determinate structure.

155
Manual Hydraulic Structures Part II - 'Loads'

Figure 25-2 indicates the difference between a statically determinate beam (two supports) and a statically
indeterminate beam (three supports). In the first case the forces and moments are not influenced by the
settlement of a support. In the second case the moment above the settled middle support decreases and
the moment in the field thus increases.

Figure 25-2 Forces and moments as a result of the settlement of a support

Settlements can be taken into account in the structural calculations in two ways, namely:
• Taking an imposed deformation into account
• Schematising the ground as a spring

The schematisation of the ground as a spring is not an accurate method because the ground does not
behave like a spring. It is preferable to take the settlements into account as imposed deformations,
whereby the settlements are determined using the distribution of forces in the structure initially neglecting
the settlements. This way the settlements are overestimated and the calculated load is conservative. It is
possible to carry out the calculations a number of times, thus iterating, to estimate the loads more
accurately. For both methods of calculation, see Chapters 32 and 33 (Part III).

156
Manual Hydraulic Structures Part II - 'Loads'

26. Soil, earthquake

26.1 General
For the design of a hydraulic engineering work it is important to know whether earthquakes occur in the
area where construction is to take place. The horizontal accelerations induced by earthquakes largely
affect:
1. The ground
Particularly due to the ratio between the horizontal and the vertical ground stresses. This can lead to
plastic deformation of the ground.
2. Structures (above and below ground level);
These structures are also subjected to the same accelerations in both horizontal and vertical
directions.

The theory of the generation of earthquakes assumes that the earth’s crust consists of plates that move
very slowly relative to each other. At the interface between the plates (the fault line), elastic deformation
takes place until a maximum shear stress is exceeded. At that moment the earth’s crust deforms
plastically and the potential energy of the elastic deformation is released. This occurs as a jerky
deformation (waves and vibrations).
Earthquakes can also be caused by deformations in the earth’s crust. Examples are earthquakes due to
gas and oil extraction.

26.2 Richter scale

The Richter scale is mostly used to register and classify earthquakes. The Richter scale is derived from
the maximum registered amplitude of the quake and the distance from the point of registration to the
epicentre of the earthquake. The epicentre is the projection of the point of origin of the quake on the
Earth’s crust. The Richter scale value is called the magnitude of the earthquake. This magnitude is
independent of the location of the measurement of the amplitude. Though every earthquake has a unique
magnitude that is the same in all locations on earth, the effect of the earthquake is not the same
everywhere. Therefore, the magnitude of an earthquake cannot be used for a good estimation of the
amount of damage to or load on a structure. The table below, however, can be used to give a global
indication of the effects.

Earthquake severity
Richter Magnitudes Earthquake effects
less than 3,5 Generally not felt, but recorded
3,5 - 5,4 Often felt, but rarely causes damage.
under 6,0 At most slight damage to well-designed buildings.
Can cause major damage to poorly constructed buildings over small regions.
6,1 - 6,9 Can be destructive in areas up to about 100 kilometres across where people live.
7,0 - 7,9 Major earthquake. Can cause serious damage over larger areas.
8,0 or larger Large earthquake. Can cause serious damage in areas several hundred kilometres across.

Table 26-1 Effects related to the Richter scale

Another indication is that the energy related to magnitude 1 corresponds to a car (1000 kg) falling to the
ground from a height of 100 metres. For every single increase of magnitude, the number of cars can be
multiplied by thirty. Hence, for magnitude 8, one would have to drop 22 billion cars from the Dom tower in
Utrecht.

To determine the Richter scale a nomogram can be used, which requires the distance to the epicentre
and the maximum amplitude as input. An example of such a nomogram is given in Figure 26-1.

157
Manual Hydraulic Structures Part II - 'Loads'

Figure 26-1 Determination of the magnitude of an earthquake on the Richter scale

Figure 26-2 shows the areas where earthquakes of magnitude IV or larger were registered in 1996.

Figure 26-2 Observed earthquakes of 1996 (M ≥ 4)

158
Manual Hydraulic Structures Part II - 'Loads'

26.3 Design
In the Netherlands, designs rarely take earthquakes into account, unless large consequences (and thus
large risks) are involved, for instance, such as for nuclear power plants. Even the storm surge barriers in
the Delta Works were not dimensioned for earthquakes. The probability that (serious) damage is caused
by an earthquake is very small and the probability that it happens during a storm surge is negligible.

In cases where earthquakes are used for the dimensioning, civil engineers need a measure with which
the load on a structure can be determined. In fact, a constructor is only interested in the maximum
acceleration of the ground. This is why the intensity according to the Mercalli scale is a better measure
than the Magnitude according to Richter. Amongst other factors, the intensity depends on the distance to
the epicentre and the structure of the earth’s crust.

The propagation and damping of waves and vibrations in the earth’s crust depend on a large number of
factors, such as, for instance, the soil types and the layered composition and the density variation of the
earth’s crust. Due to damping, the intensity of the earthquake decreases as the distance relative to the
epicentre increases. Local registrations and experiences are of large importance for the estimation of the
nature and extent of earthquake loads.

Figure 26-3 can be used for the estimation of earthquake intensities in the Netherlands.
For the estimation of the horizontal acceleration a at a certain intensity, one can use the classification
given in Table 26-2.

Intensity Description Acceleration

2
[m/s ]
I Only registered by seismographs 0,01
II Very weak; only felt in favourable circumstances 0,02
III Weak; felt by a few people; vibrations similar to those of passing traffic 0,05
IV Moderate; felt by many; tremors as vibrations cause by heavy traffic; doors and 0,1
windows rattle
V Fairly strong; generally felt by everybody; hanging objects sway; clocks stop running 0,2
VI Strong; shock reactions; objects in houses fall over; trees move; insufficiently solid 0,5
houses are damaged
VII Very strong; many buildings are damaged; chimneys break; waves in ponds; church 1,0
bells ring
VIII Destructive; panic; general damage to buildings; weak buildings are partially 2,0
destroyed
Table 26-2 Intensities according to the Modified Mercalli Scale (1931)

26.3.1 Buildings
For buildings, the accelerations in all directions of the earth’s surface are of particular importance. These
accelerations cause dynamic loads. The dynamic character of the loads does not always stand out
equally well in different structure regulations and norms. In the various regulations the earthquake load is
often calculated as a static load, a function of the mass of the elements of the structure, in all directions.
The stiffness of the elements of the structure is not always taken into account.

The equivalent static method (ESM) approaches the dynamic behaviour of the structure by also relating
the size of the load to the structure’s lowest natural frequency. This and other methods to determine the
earthquake load are described in specialised literature on seismic design of reinforced concrete
structures. The essence of most of these methods is that the structure is schematised as a mass-spring
system with freedom of movement in all directions.

For a description of earthquakes as a statistic process, one is referred to the lecture notes Statistic
vibrations (in Dutch: “Stochastische trillingen”) (Vrouwenvelder, 1997).

159
Manual Hydraulic Structures Part II - 'Loads'

IV

Figure 26-3 Earthquake intensities in and around the Netherlands (Modified Mercalli Scale)

26.3.2 Soil
For soil retaining structures it is unusual to schematise the structure as a mass-spring system. The
difficulty with such a schematisation is estimating the mass of the soil to be retained in earthquake
conditions.

The German manual on quay walls, the EAU (Empfehlungen des Arbeitsausschusses Ufereinfassungen,
Wilhelm Ernst & Sohn, München), gives an approximation method that is simple in application. This
approach considers the maximum horizontal acceleration constant and assumes that the maximum
horizontal and vertical accelerations do not occur simultaneously (the fact that this assumption is not
always correct was shown in Kobé, Japan).

With these assumptions, the resultant acceleration can be found by vectoral addition of gravity and the
horizontal acceleration caused by the earthquake. The calculation is based on a fictitious rotation of the
system, such that the resultant acceleration coincides with the gravitational acceleration. This means that
both the sheet piling and the ground level in front of and behind the retention structure are given a
rotation according to Figure 26-4.
a
The rotation amounts to tan   .
 g

160
Manual Hydraulic Structures Part II - 'Loads'

Figure 26-4 Fictitious rotation of the soil retaining structure and the ground level as a result of horizontal accelerations

Using the general horizontal ground pressure coefficients Ka and Kp, as given in Chapter 24, the
horizontal ground pressure can also be determined for a rotated schematisation. The true horizontal load
is found by resolving the force into its true horizontal and vertical components.

For the soil below the groundwater level, the mass of the water must also be taken into account. This is
done by correcting the rotation, for this one is referred to the German guidelines on quay wall design, the
EAU 2012.

161
Manual Hydraulic Structures Part II - 'Loads'

27. Soil, groundwater

An important aspect of the soil is the presence of groundwater. Groundwater largely influences both the
strength and the stiffness behaviour of the ground. Furthermore, groundwater is also responsible for a
large load on the structure.
Sometimes the load of the groundwater is only temporarily undesirable. If that is the case, one can opt for
drainage. That is why this chapter also considers groundwater flow and drainage. Groundwater flow can
cause erosion, this is known as piping.

These six facets of groundwater are discussed below in the following order:
• Load
• Groundwater flow
• Piping
• Drainage
• Influencing strength
• Influencing stiffness

27.1 Groundwater pressure

In the Netherlands, many structures placed in the ground are also found to be in the groundwater, due to
the high groundwater level. The structure therefore undergoes an upward load. If there is no groundwater
flow (e.g. seepage or consolidation), this load is identical to that of a floating object:

p = ρ ⋅g ⋅h

where: p [Pa] = (ground)water pressure

3 3 3
ρ [kg/m ] = the density of the water (salt = 1025 kg/m ; fresh = 1000 kg/m )
2
g [m/s ] = the gravitational acceleration
h [m] = the piezometric head or piezometric height

For excavations, one must always check if this load on the bottom of a clay layer, building site floor or
cellar floor won’t allow the structure to burst open. There should be a sufficient safety margin between the
upward groundwater load and the downward load of the weight of the clay layer or structure ( γ m,g = 1,1 ) or
the tensile force of the tension piles ( γ m,b 4 = 1,4 ).

By loading a layer of little permeable soil (clay or peat), the piezometric head temporarily increases.
Consolidation gradually reduces the piezometric head back to its initial value (see Part III, section 34.1).
Due to groundwater flow some layers (particularly the deeper Pleistocene sand layers) have a different
piezometric head, which causes seepage. In the vicinity of rivers, the piezometric head in sand layers is
closely related to the river water level (see note).
The water level in a soil layer is easy to measure using a piezometer. That is why the groundwater stress
in deeper layers is always measured using a piezometer. For soil layers close to ground level the water
levels in ditches are conveniently assumed. Due to run-off ground level is often considered a maximum
piezometric head in polders.
Naturally, groundwater flow occurs around a water retaining structure and thus the groundwater pressure
is not constant. This is described in the next section.

Note
Before the fall of the Berlin Wall, the German government built a new 100 million Euro parliament building
(Reichstag) in Bonn, beside a river. Once the car park in the basement had been completed, high water levels were
encountered in the river. The levels above ground had not yet been constructed and thus did not provide
compensatory weight. The company that had to place sheet piling around the excavated building site had
experienced delays, leaving the groundwater and river water free reign to destroy the structure, which
consequently happened. Nobody dared to implement the only alternative (flooding the car park to provide
additional weight) to save the Reichstag for fear of water damage to the basement.

162
Manual Hydraulic Structures Part II - 'Loads'

27.2 Groundwater flow

27.2.1 Theory
In many cases of groundwater flow one can assume that the specific discharge is linearly proportional to
the hydraulic gradient:

q = −k ⋅ i

in which: q [m/s] = specific discharge (= Q / A)

k [m/s] = permeability coefficient
dh
i [-] = hydraulic gradient: i =
ds
h [m] = potential, piezometric head
s [m] = distance along a fictitious flow line

This is called Darcy’s law. This law only applies if the potential flow is laminar. Thereto the following
applies:
ud
Re = <1
ν
in which: Re [-] = the Reynolds value
u [m/s] = the fluid’s filtration velocity
d [m] = diameter of the soil particles
2
ν [m /s] = kinematic viscosity

Re >1 occurs for soil particles with large diameters (stone) and for large flow velocities, which don’t
usually occur in groundwater flow.

A rough indication of the permeability of various types of soil is given below:

Soil type k [m/s]

-2
Gravel 10
-3
Coarse sand 10
-3 -4
Moderately coarse to moderately fine sand 10 to 10
-4 -5
Fine sand 10 to 10
-9 -11
Clay 10 to 10
Table 27-1 Permeability versus soil type

Groundwater flow can thus be described using the theory of potential flow. This means that a two-
dimensional groundwater flow can be analysed using a flow net (see Example).
The following applies between the flow lines:

dh
q da = −k da = constant
ds

in which da [m] = distance between two successive flow lines

Because the successive flow lines and equipotential lines create a square, the following applies:
da
= constant . Because k is a constant, the potential difference between two successive equipotential
ds
lines is constant.

With a given potential in two locations, the course of the potential and the specific discharge along a flow
line are fixed in a flow net.

Erosion below structures due to groundwater flow ('piping') is treated in section 37.4 (part IV of this
manual).

163
Manual Hydraulic Structures Part II - 'Loads'

27.3 Drainage
When an area is being drained the groundwater flow is usually 3-dimensional, so the given 2-dimensional
flow net method cannot be used. For calculations involving drainage, the water levels and the
groundwater pressures, see Chapter 47 (in part IV of this manual).

27.4 Influences on strength

The soil’s shear strength depends on the cohesion, the angle of internal friction and the so-called inter-
granular stress (also known as effective stress). The inter-granular stress, also known as effective soil
stress (σ’) depends on the total stress and the water pressure:

σ'=σ −p

in which: σ [Pa] = total stress

p [Pa] = water pressure (also denoted as σw )

This shows that the presence of groundwater reduces the effective stress and thus the shear stress of the
soil, as:

τ = c '+ σ 'n ⋅ tan(ϕ )

As regards the shear stress τ , it is therefore important to keep the water pressure under the structure as
low as possible. That is why the seepage screens (sheetpile walls that reduce the length of seepage
paths to prevent piping) are usually installed on the upstream side under the bottom slab of the structure.
A separate chapter is dedicated to loads caused by inter-granular stress σ’ (Chapter 24).

27.5 Influence on stiffness

Another soil property which is largely influenced by the presence of groundwater is the course of the
compaction of the soil as a result of an increase of the ground pressure. In the case of impermeable soil
(clay, peat), an increase of the total stress often causes a proportional increase of the water pressure.
This is because water cannot flow out of the pores quickly enough when the soil is compressed. The
water can be considered incompressible and accounts for the entire increase of pressure. The result is a
high total stress and water pressure and a relatively low effective stress. As a consequence of the
increase of the water pressure, groundwater will flow to a place with a lower potential. This way, the water
pressure and the effective stress adjust to the load.

In this case, the compression process depends on the speed with which the groundwater can flow out of
the pores. This process of reducing the water pressure and increasing the effective stress is called the
consolidation process. The presence of groundwater therefore causes a temporary increase of stiffness in
compressible, impermeable soil.

For this see also in part III, Chapters 33 ('Stiffness') and 34 ('Settlement').

164
Manual Hydraulic Structures Part II - 'Loads'

28. Shipping, hydraulic aspects

Water movements created by sailing and manoeuvring ships can be of importance for structures in ports
and shipping lanes. These water movements are:
• Return current
• Water level depression
• Ships wakes
• Propeller wash

Generally, these hydraulic loads on a structure are inferior to the other loads. However, the hydraulic
loads are of large importance to the stability of flexible structures adjoining the hydraulic engineering
work. For the design of such structures one is referred to the book for course CIE4310: "Introduction to
Bed, bank and Shore protection" (Schiereck & Verhagen, 2016).

165
Manual Hydraulic Structures Part II - 'Loads'

29. Shipping, berthing and collision

Extended: February 2018

29.1 Introduction
Berthing a ship (aanleggen van een schip) to a structure and collision of a ship against a structure are,
theoretically, the same process. The only differences are the velocity and the extent of control of the
process.

A number of aspects are of importance in relation to the load of a berthing ship on a structure. These
aspects are:
• dimensions of the ship
• mass of the ship
• velocity of the ship
• stiffness of the ship’s hull
• mass of the water moving along with the ship, known as the hydrodynamic or added mass
• approach angle
• stiffness of the mooring structure
• geometry of the mooring structure
• place where the ship hits the structure
• current
• wind

29.2 Theory
A schematic representation of a berthing vessel including the main parameters relevant to calculating the
berthing impact is depicted in Figure 29-1.

kship = stiffness of the ship’s hull

kwater = hydrodynamic stiffness
kstructure = stiffness of the structure

mship = mass of the ship

mwater = hydrodynamic-, added mass
mstructure = mass of the structure

cship = damping of the ship’s hull

cwater = hydrodynamic damping
cstructure = damping of the structure

Figure 29-1 A berthing vessel and main parameters related to berthing impact

Together with the water that moves along, the ship to be moored has the following amount of kinetic
energy:
1
Ekin = ( ms + mw ) v s2
2

in which: Ekin [Nm] = kinetic energy

ms [kg] = mass of the ship
mw [kg] = mass of the water moving with the ship; additional mass
vs [m/s] = velocity of the ship and water (⊥ structure)

The kinetic energy has to be absorbed by the berthing structure and it is a measure of the load on a
structure. The maximum force that develops between the ship and the berthing structure largely depends
on the stiffness of the whole ship-water-structure system. This system can be schematised as a mass-
spring-dashpot system (massa-veer-dempersysteem).

166
Manual Hydraulic Structures Part II - 'Loads'

ksystem
kship kwater kstructure
mship mwater mstructure = msystem
cship cwater cstructure csystem
Figure 29-2 Schematisation of a berthing vessel as a mass-spring-dashpot system

In a preliminary design damping will be neglected, which results in a conservative calculation of the
berthing load on the structure. Without damping, which dissipates energy, all of the kinetic energy is
transformed into potential energy. The potential energy in a spring is:

∆x ∆x
1
E pot = ∫ k ( x ) x dx E pot = ∫ k x dx = 2 k ∆x
2
for a linear spring system:
0 0
where
∆x [m] = displacement of the spring
k [Pa] = spring stiffness of the total system

F
In general, k = . For the spring stiffness of the berthing system, the combination of the structure and
∆x
the ship should be considered and the water stiffness can be neglected. The stiffness of the total system
is:
1 1 1
= + .
ksystem kstructure kfender

Usually kfender << kstructure.

Compared to the mass of the ship plus added water mass, the mass of the structure (the part that would
move) is negligible. Hence, the maximum potential energy in the spring equals the maximum kinetic
energy of the ship and the water. For a linear elastic structure, equating potential and kinetic energy
1 2 Ekin,max
gives: E pot ,max = Ekin,max ⇔ k ∆x 2 = Ekin,max ⇔ ∆x =
2 k

29.3 Design berthing forces

The current practice in the design calculations and dimensioning of fendering systems and flexible
structures is based on energy considerations. In these, it is assumed that the loss of kinetic energy of the
ship is transformed into an equal amount of energy absorbed by the fender or the structure. The basic
principle was first developed by Gottfried Leibniz and Johann Bernoulli:

1
Ekin = mv 2 [J = Nm]
2

Including the effects of added water mass, eccentricity and structural deformation, the equation,
according to F. Vasco Costa (1973), transforms into:

1
Ekin = ms v s2 CH CE CS
2

CETMEF, a French research institute, recommended to as well include the effect of water confined in
between the ship and the structure:

1 1
Ekin = ms v s2 CH CE CS CC or: Ekin = ( ms + mw ) v s2 CE CS CC
2 2

167
Manual Hydraulic Structures Part II - 'Loads'

ms + mw
where: CH [-] = hydrodynamic coefficient =
ms
CE [-] = eccentricity coefficient
CS [-] = ship coefficient
CC [-] = confinement coefficient
ms [kg] = mass of the ship
vs [m/s] = velocity of the ship (component ⊥ structure)

This approach is globally used and accepted, e.g. prescribed in the BS6349 (British Standard) and
mentioned in PIANC report 151.

The force on the (linear elastic) structure can then be computed with:

F = k ∆x = 2 k Ekin

This formula shows the influence of the stiffness of the structure on the resultant force of the ship on the
structure and therefore also of the structure on the ship. Because the stiffness of the structure is not
known until after the design phase, the load of a ship on a structure is not given as a force, but as an
amount of kinetic energy that needs to be absorbed. Besides the amount of kinetic energy to be absorbed
during mooring, other forces on the ship must also be taken into account, such as wind and current.
These forces are transferred to the structure via the ship.

The parameters to calculate the kinetic energy are discussed below.

Velocity of the ship vs

The velocity with which a ship berths is an important factor, after all, the kinetic energy is proportional to
the square of the velocity. A large number of measurements revealed that the berthing speed depends on
the ship’s dimensions, the type of load (cargo) and the berthing conditions.

In good berthing conditions, one could use Table 29-1 for mooring speeds.

ship Observed velocity Design velocity

[ton] [m/s] [m/s]
<10 000 0,10 ~ 0,30 0,20
10 000 ~ 50 000 0,10 ~ 0,20 0,15
>50 000 0,10 ~ 0,15 0,15
Table 29-1 Berthing speeds in good conditions (velocity component vs ⊥ structure)

As the observed velocities exceed the design velocities one can conclude that the design velocities are
too low. One is therefore advised to assume the maximum observed velocities instead.

The design velocity as given in Figure 29-3 is as a function of both the ship’s dimensions and the berthing
conditions. As the velocities in this figure are sometimes lower than the observed velocities, one is
advised not to assume velocities lower than the maximum observed velocities given in the table above.

168
Manual Hydraulic Structures Part II - 'Loads'

Figure 29-3 Design velocity

Hydrodynamic coefficient CH
The hydrodynamic coefficient is in fact simply the ratio between the mass of the ship plus the water
moving with the ship and the mass of the ship:

ms + mw
CH =
ms
The mass of the ship that should be used to calculate the hydrodynamic coefficient is the total water
displacement of the loaded ship (FLD = full load displacement). The additional mass of the water
depends on, amongst other factors, the movement of the ship in three-dimensional space. Several
models have been developed to determine the hydrodynamic additional mass. They are based on the
potential theory or on the theory regarding the preservation of impulse. However, here a simple
approximation that can be used for a preliminary design suffices. This approximation is based on the
potential theory and is known as Stelson Mavils’ equation:

1
mw = ρ L π D 2
4
3
in which: ρ [kg/m ] = density of (sea)water
L [m] = length of the ship
D [m] = draught of the ship

(To compare this formula with Morison’s formula, see Chapter 20 "Water, waves, slender structure")

The equation for the hydrodynamic coefficient can be simplified further with an assumption for the block
coefficient Cb:
π
Cb ≈
4
ms = ρ LBDCb
ms + mw D
CH = ≈ 1+
ms B

It is of current practice to take CH = 1,2 (PIANC report 151, 2014), but it is clear that this hydrodynamic
aspect is directly influenced by the shape of the striking bow.

2D
Note: Vasco-Costa formula reads as follows CH = 1 + use of this well-known formula is wide
B
spread. Engineering judgement is required!

169
Manual Hydraulic Structures Part II - 'Loads'

Coefficient of eccentricity CE
The coefficient of eccentricity takes into account the energy dissipation caused by the yawing (bijdraaien)
of the ship when it moors eccentrically against the structure. This yawing is shown in Figure 29-4.

L
φ
Yaw B
Vcosφ
Vsinφ
β Vship,perpendicular
V γ
r φ= 5-10 o

lbow
berthing point berthing line
Figure 29-4 Eccentric berthing

If the movement of the ship at time of impact with the structure can be described solely as a translation of
the centre of gravity, the coefficient of eccentricity is:

k 2 + r 2 cos2 (γ )
CE = γ = 90o − β − ϕ r = ( 1 L − ℓ bow )2 + 1 B 2
k2 + r 2 2 4
in which:
k [m] = radius of gyration (traagheidsstraal) of the ship
r [m] = the radius between the gravity centre of the ship and the berthing point
o
β [ ] = the angle between radius r and the ship’s velocity along the axis of the ship
o
γ [ ] = the angle between radius r and the ship’s velocity perpendicular to the berthing line
o
φ [ ] = the berthing angle; the angle between the ship’s axis and the berthing line
ℓbow [m] = distance between the bow tip and the point where the straight part of the hull starts
to curve to the bow

The radius of gyration of the ship can be approximated by:

k = (0,19 Cb + 0,11) L
V
in which: Cb [-] = block coefficient = Cb =
L ⋅B ⋅D
L [m] = length of the ship
B [m] = width of the ship
D [m] = draught of the ship
3
V [m ] = (volumetric) water displacement of the ship

The smaller the value of Cb, the larger the load F. Therefore a small Cb is governing.
The following approximation applies: 0,6 (slender ships) ≤ Cb ≤ 0,95 (container ships)

Ship coefficient CS
The ship coefficient takes the elasticity of the ship’s side into account. This coefficient depends on the
stiffness of the structure and that of the ship’s shell and takes into account the part that is taken on by the
structure. In the case of weak structures, such as, for instance, a wooden pier with fenders, nearly all of
the energy is absorbed by the structure because the deflection of the ship’s shell is negligible. In this
case, CS = 1.

For a relatively stiff structure, such as a quay with wooden support beams and car tyres, the deflection of
the ship’s shell will not be entirely negligible compared to the deflection of the structure. In this case, CS =
0,9 can be used for design purposes.

170
Manual Hydraulic Structures Part II - 'Loads'

Confinement coefficient CC
The confinement coefficient, or cushion factor, takes the dissipation of energy into account, caused by the
fact that water has to be squeezed away from the space between the ship and quay structure. If the
structure is closed, e.g. a straight vertical sheetpile wall, a cushion of water between the structure and
ship will slow down the ship. For a closed quay and parallel mooring (θ = 0°) this can lead to a 20%
reduction of the amount of energy to be absorbed. If there is a small angle between the ship and the quay
(θ = 5°) this reduction can disappear almost entirely because the water can simply flow away.
If the structure is open, e.g. a jetty, no hydrodynamic damping will occur.

Therefore, the limit values of the confinement coefficient are: 0,8 ≤ CC ≤ 1,0 (PIANC report 151, 2014).
For safety reasons, one can assume CC = 1,0 for a preliminary design.

Structure stiffness k
The total stiffness of the system is determined by both the structure (including the foundation) and the
used fenders. In very stiff structures the fenders are governing for the design. Assuming a cell-fender with
the spring characteristics given below and a diameter of 0,80 m, one finds a fender stiffness of:

F 1100 kN
kfender = ≈ ≈ 14 000 kN/m
∆ x 10% ⋅ 0,80 m

2000
reaction force [kN]

1500
cell fender
1000
airbag fender
500

0
0 10 20 30 40 50 60
deformation [%]

Figure 29-5 Force versus deformation of a cell-fender and an airbag fender.

This stiffness of the cell-fender only applies up to a deformation of 15%. After that, the behaviour is no
longer linear and the reaction force remains fairly constant at 1,5 MN. The advantage of the cell-fender is
that it absorbs a lot of energy by means of a relatively large deformation, without increasing force
(rectangular shaped area under the curve). The disadvantage is that the mooring force is very large for
possible small boats; during the first initial deformation the force rapidly increases to a first maximum
(triangular shaped area under the curve). Airbag fenders deform more than cell-fenders, but also cause
smaller mooring forces for small boats.

29.4 Design berthing and collision forces

Design according to Eurocode 1

For preliminary designs, Eurocode 1, part 1-7 (NEN-EN 1991-1-7+C1+A1, 2015), distinguishes between
a frontal impact force Fdx, in the normal sailing direction, and a lateral force Fdy, perpendicular to the
sailing direction. Two tables are given with indicative values for the dynamic collision forces, including the
effect of the hydrodynamic mass. One table is for inland ships (Table 29-2 in this manual) and the other
one is for sea-going vessels (Table 29-3). These tables should not be used for berthing structures that
resist the impact load of ships under normal conditions, like quay walls and breasting dolphins.

In case of non-frontal impacts, an additional friction force FR acts in the same direction as Fdx, and can be
calculated with:

FR = µ ⋅ Fdy

where µ represents a friction coefficient [-], for which usually a value of 0,4 is used.

171
Manual Hydraulic Structures Part II - 'Loads'

Table 29-2 Berthing forces per inland ship class, according to Eurocode 1 (NEN-EN 1991-1-7+C1+A1)

Table 29-3 Berthing forces per sea-going vessel class, according to Eurocode 1 (NEN-EN 1991-1-7+C1+A1)

It is recommended to adjust the indicative values given in Table 29-2 and Table 29-3, depending on the
consequences of the structural failure due to a ship collision. In case of small consequences, the forces
may be reduced, in case of large consequences, the values should be increased. Eurocode 1
recommends a factor of 1,3 for a frontal collision and 1,7 for a lateral collision. In ports and harbours, the
values in the tables may be reduced by a factor 0,5.

Advanced methods for calculating the collision forces are given in appendix C of Eurocode 1-7 (NEN-EN
1991-1-7+C1+A1).

Design according to AASHTO

The USA standard for bridge design, AASHTO LRFD Bridge Design Specifications (2013) gives the
following equation to calculate the kinetic energy of a sailing ship:

Ekin = 500 CH ms v s2 [kJ = kNm]

where:
CH [-] = hydrodynamic coefficient
ms [ton] = weight of the ship displacement
vs [m/s] = velocity of the ship

172
Manual Hydraulic Structures Part II - 'Loads'

The value of the hydrodynamic coefficient depends on the keel clearance and the draught of the ship:
• CH = 1,05 if the keel clearance > 0,5 x draught
• CH = 1,25 if the keel clearance < 0,1 x draught
Values in between should be interpolated.

The hydrodynamic coefficient for berthing ships is higher than for colliding ships, because of the larger
mass of water that is affected if a quay is approached in parallel direction.

AASHTO gives the following equation for the static impact force of a frontal collision:

Fs = 1, 2 ⋅ 105 v s DWT [N]

where:
vs [m/s] = velocity of the ship
DWT [ton] = deadweight tonnage of the ship

The static force in lateral direction is 50% of the frontal force Fs.

29.5 Fenders
Fenders absorb kinetic energy of mooring ships in such a way, that the force on the structure (thereby
also the force on the ship) remains below an acceptable limit. The amount of energy that needs to be
u ,max
absorbed is E = ∫
0
F(u ) δu , in which F(u) is the force and u,max is the entire displacement. This leads

to the conclusion that for a given amount of energy, the force is determined by the total displacement and
therefore the stiffness of the structure. In principle, the more flexible the structure is, the smaller the force
is and the larger the displacement. The relationship between the force and the displacement is
determined by the spring characteristics of the structure. The spring characteristics of a linearly elastic
structure are linear. In the case of rubber fenders, the spring characteristics are generally not linear. An
example for a rubber cell fender is given in Figure 29-6.

Figure 29-6 Spring characteristics of Cell-fenders with different rubber compound (source unknown)

Fender panels are used to spread the concentrated force of the fender over a larger part of the ship’s hull.
2
The distributed load on the ship should not exceed 400 kN/m to prevent damage to tankers. It is
important to take the kind of ship and its cargo into account when selecting a fender type. For instance, at
mooring places for Nafta tankers, it is important to keep the forces on the ship small, resulting in soft
fenders. A disadvantage of such soft fenders is that the movements of the moored ship will be larger.

Guidelines for the design of fender systems are given in PIANC report 33 (2002) .

173
Manual Hydraulic Structures Part II - 'Loads'

29.6 Literature

AASHTO LRFD Bridge Design Specifications - SI Units (8th Edition). American

Association of State Highway and Transportation Officials (AASHTO), Washington,
DC, 4th ed. edition, 2017. ISBN 9781560516545.

Eurocode 1: Actions on structures, Part 1-7: General actions - Accidental actions. NEN-EN 1991-1-
7+C1+A1, 2015.

PIANC (2002). Guidelines for the design of fender systems. Report 33. Brussels, Belgium.

PIANC (2014). Design of lock gates for ship collision. Report 151. Brussels, Belgium.

Vasco Costa, F. (1973). Dynamics of berthing impacts. NATO Advanced Study Institute on Analytical
Treatment of Problems in the Berthing and Mooring of Ships. Wallingford, England.

174
Manual Hydraulic Structures Part II - 'Loads'

30. Shipping - mooring forces

30.1 Theory
Mooring forces are relevant for the design of bollards (bolders) on quays, and quay walls, for the design
of ships and for the choice of suitable mooring ropes, also called 'hawsers' (trossen).

Forces in mooring ropes can be caused by:

• Loads on the ship such as wind, current and water level differences (filling a lock!)
• Movements of the ship, the mass of the ship and the stiffness of the mooring cables are of impor-
tance in this respect.
• Pre-stressing forces in the mooring cables

Pre-stressing forces in the cables are generated during the hauling in and after mooring. The fracturing
force of the cables and the number of cables limit the mooring force. Increasingly, modern quays and
mooring structures are equipped with “quick release hooks”. These release the mooring cables
automatically if a maximum force is exceeded. So-called yielding bolts are used to connect the structure
with conventional bollards and mooring bitts, which are still used often in heavy quay structures and
sluices. The yielding bolts yield when the maximum acceptable force is exceeded.

30.2 Preliminary design

For a preliminary design one can assume the mooring forces for seagoing vessels, self-propelled barges
(for inland waterways) (binnenvaartschepen) and yachts (plezierjachten) as specified in the tables below,
as can be found in several standards, like the German EAU 2012 and British Standard 6394 (part 4), with
more or less similar values.

Seagoing vessels
water displacement of mooring force
ship [ton] 8) [kN]
< 2 000 0 - 100
2 000 ~ 10 000 100 - 300
10 000 ~ 20 000 300 - 600
20 000 ~ 50 000 600 - 800
50 000 ~ 100 000 800 - 1000
100 000 ~ 200 000 1000 - 1500
> 200 000 1500 - 2000
Table 30-1 Mooring forces of sea going ships per bolder

Inland barges
9
class of ship ) mooring force
[kN]
CEMT Class I + II 150
CEMT Class III + IV 200
CEMT Class V 250
CEMT Class VIa 300
CEMT Class VIb 350
Table 30-2 Mooring forces of inland barges per bolder (RVW 2017)

8
Example: Water displacement of a seagoing ship = 70 000 tonnes: Mooring force = 800 + 200*(70-50) / (100-50) = 880 kN.
9
The CEMT barge class definitions can be found in the 'Richtlijnen Vaarwegen 2017' of Rijkswaterstaat (downloadable from internet).

175
Manual Hydraulic Structures Part II - 'Loads'

Yachts
class of ship mooring
force [kN]
Yachts 55
Table 30-3 Mooring forces of yachts per bolder

Guidelines for mooring systems can be found in PIANC reports 168 (Single point yacht mooring design
and 186 (Mooring of large ships at quay walls) and RecCom report 10 (Mooring systems for recreational
crafts, 2005).

176
Manual Hydraulic Structures

Part III: Materials

Manual Hydraulic Structures Part III - 'Materials'

31. Soil - properties

Section 31.5 extended: February 2015

31.1 Stiffness and strength of soils

Before collecting or calculating parameters one must be aware of whether the type of problem that has to
be dealt with is:
• a strength problem (e.g. the stability of a dike),
• a stiffness problem (e.g. the settlement of a road) or
• a mixed problem (e.g. the foundation of a structure).

Specific tests exist to determine the strength parameters (e.g. the direct shear test) and the stiffness
parameters (e.g. the oedometer test). Tests to determine both types of parameters also exist (e.g. tri-axial
compression test).

Most calculations also distinguish between problems of strength and stiffness. Koppejan’s settlement
calculation, for example, is a stiffness calculation. Prandtl and Brinch Hansen’s bearing capacity
calculation is a strength calculation and the spring model used to calculate sheet piling (like D-Sheet
Piling) is a mixed model; the model produces estimates of both strength and displacement(s). The
problem, the parameters and the calculation must correspond.

31.2 Soil investigation

Loads on hydraulic engineering works are transferred to the subsoil. The properties of the soil are
therefore of great importance for the design and the dimensioning of a structure. The nature of the loads
on the structure and the properties of the subsoil determine the choice of the type of foundation.
A number of techniques are available to investigate the subsoil. This includes:
• soundings
• borings
• water pressure measurements
• vane shear tests (determination of cu)
• pump tests
• (seismic research)
• (nuclear research)

These research methods are suitable to investigate the layered composition of the soil and to estimate
some soil parameters. The subsoil is usually composed of different types of soil layers. First, an attempt
is made at establishing the structure of the soil layers. For this, see Section 31.3 "Determination of soil
type". After that, the necessary soil parameters are determined for every soil layer. Fieldwork provides all
the required information, but to determine the soil parameters more accurately laboratory tests are
needed. It is best to obtain soil samples by boring; these samples can then subsequently be tested in a
laboratory.

The most important laboratory tests for alluvial soil types (sand and clay) are:
• consolidation test (= oedometer test) (determination of stiffness and permeability)
• triaxial compression test (determination of stiffness and strength)
• (permeability test)
• (direct shear test)
• (Cassagrande’s test)

For rock and stony materials, other parameters are relevant, such as the compression strength, the
splitting strength, the density and the permeability. Tests mentioned above can be used to calculate the
representative parameters of a given soil layer. For this see Section 31.4 "Determination of soil
parameters from laboratory tests".

If no laboratory tests are available, a few soil parameters one can be estimated empirically, using the qc
values of the soundings. For this refer to Section 31.5 "Determination of soil parameters from cone
penetration tests values". Alternatively, values coming from Dutch practice can be used. Table 31-4 of
Section 31.6 gives the representative low values of several important soil parameters for a number of soil
types. Besides the aforementioned parameters the table mentions the following parameters: volumetric
weight (γ), sounding value (qc), swelling coefficient (Csw , to be used instead of Cc in case of pressure
reduction), Young's modulus (E), and the undrained shear strength (fundr = cu).

179
Manual Hydraulic Structures Part III - 'Materials'

31.3 Determination of soil types

The best way to determine what soil type a soil layer consists of is to carry out borings in different
locations and to analyse the samples in a laboratory. Usually one opts for a cheaper alternative. In this
case only a few borings and a (much) larger number of cone penetration tests (CPT) are carried out (e.g.
every 25 metres). The cheaper CPT’s (sonderingen) are then linked to the more expensive boring
samples, to determine the soil profile and the soil characteristics of every layer that has been
distinguished.

There are two types of CPT’s, which depends on the cone being used. One measures both the cone
resistance and the local friction and one measures the cone resistance and the (pore) water pressure.
The CPT including water pressure measurements clearly shows:
• where the impermeable layers (clay and peat) are.
• how well the layer seals off water.
• what the water pressure is at a greater depth, in the permeable layers (sand and gravel).

The CPT including local friction measurement clearly shows which soil layers are present and is therefore
used most often. There is a strong correlation between the soil type and the friction ratio, see table below.
The friction ratio is the sleeve friction divided by the cone resistance and is therefore dimensionless.
Nowadays the cylindrical electronic penetrometer with friction sleeve is often used for soundings.

Soil type Friction ratio > than: Friction ratio > than:
(cylindrical electronic cone) (mechanical friction sleeve cone)
moderate to coarse sand 0,4 % 1,0 %
moderately fine sand 0,6 % 1,3 %
fine sand 0,8 % 1,6 %
silty sand 1,1 % 1,9 %
clayey sand 1,4 % 2,6 %
sandy clay or loam 1,8 % 2,8 %
silt 2,2 % 3,1 %
silty clay 2,5 % 3,6 %
clay 3,3 % 4,3 %
humus clay 5,0 % 6,3 %
peat 8,1 % 8,0 %
Table 31-1 Friction ratio versus soil type

Notes
• It is advisable to always carry out a couple of borings; firstly because the percentages in the table
above are not always correct, secondly because a boring provides a lot of visual information and
thirdly because a boring is necessary to obtain soil samples for laboratory tests.
• It is important that one first compares the table above with the values found by borings. One should
therefore always carry out a sounding near a boring.
• If weak soil layers are present (especially for an excavation (building site)), it is sensible to have a
couple of CPTs that include water pressure measurements carried out, because these measurements
distinguish well between permeable and non-permeable soil layers.

31.4 Determination of soil parameters from laboratory tests

For the weak soil layers compression test are usually carried out in case of settlement questions, for
strength questions triaxial test are usual. For stronger soil layers (sand and gravel), usually only triaxial
tests are performed. From these tests one can determine the strength and stiffness parameters for a
certain soil layer.

In calculations the 5% upper or lower limit values are to be used. For soil both the real average value ×avg
and the real average spread σ are unknown. That is why an average value ×avg and the standard deviation
σ are first determined with tests. Subsequently, the student-t method can be used to calculate the
representative upper or lower limits of the average ×rep:
1
×rep = ×avg ,5% char = ×avg ± t 5% ⋅ σ ⋅ (31.1)
n
in which n is the number of tests.

180
Manual Hydraulic Structures Part III - 'Materials'

The value of t5% depends on the number of samples n:

n 2 3 4 5 6 7 8 9 10 11 21 ∞
t5% 6,31 2,92 2,35 2,13 2,01 1,94 1,89 1,86 1,83 1,81 1,72 1,64
Table 31-2 Student-t distribution

31.5 Determination of soil parameters from cone penetration tests

If no laboratory tests have been carried out, one can estimate the stiffness and strength parameters of the
soil layers for a preliminary design with rules of thumb using the sounding values (qc), or else by using
Table 31-4 of Section 31.6.

Electric Cone penetration tests (CPTs) provide a fast, repeatable and economical way to determine soil
stratigraphy and soil types. There are various ways to determine soil types from these tests, using the
measured cone resistance qc, the sleeve friction fs and, if available, the pore pressure u. Instead of the
sleeve friction, the friction ratio Rf is often used:
f
Rf = s ⋅ 100 (31.2)
qc
where Rf [%] = friction ratio
fs [MPa] = CPT sleeve friction
qc [MPa] = CPT cone resistance

Peter Robertson developed a classification chart that is used in practice for the determination of soil
(behaviour) types. The simplest version of this chart, which was published by Robertson in 1986 and
updated in 2010, uses the CPT-values of the cone resistance and the friction ratio, see Figure 31-1. The
cone resistance is made dimensionless by dividing it with the atmospheric pressure pa (0,1 MPa) .

Figure 31-1 Soil classification chart based on CPT cone resistance and friction ratio (Robertson 1986, 2010)

The chart is global in nature and can provide reasonable predictions of soil behaviour type for CPT
soundings up to about 20 m in depth. Overlap in some zones should be expected and the zones should
be adjusted somewhat based on local experience.

Robertson improved this chart by using a normalized cone resistance Qt and a normalized friction ratio
F R:
q − σ vo
Qt = t (31.3)
σ 'vo
fs
FR = ⋅ 100 (31.4)
q t −σ vo

181
Manual Hydraulic Structures Part III - 'Materials'

where: Qt [-] = normalized cone resistance

FR [%] = normalized friction ratio
qt [MPa] = measured CPT cone resistance, corrected for pore pressure:
qt = qc + (1 - α)·u2 if the pore pressure filter is directly behind the cone tip
qt = qc + (1 - α) · {β(u1 - u0) + u0} if the filter is situated in the cone tip
qt = qc in sandy soils, or if pore pressures were not measured
u0 = hydrostatic water pressure
u1 = water pressure if the pore pressure filter is situated in the cone tip
u2 = water pressure if the pore pressure filter is situated behind the cone tip
α = net area ratio of the cone because of the slit behind the cone tip
α is usually determined from laboratory calibration, with typical values
between 0,70 and 0,85
β = factor for the differing soil types for the conversion of u1 to u2
mostly β = 0,8 is used
σvo [MPa] = total vertical soil stress determined using the specific weight per soil layer
σ'vo [MPa] = effective vertical soil stress determined using the specific weight per soil layer
fs [MPa] = measured CPT cone resistance (not corrected)

The normalized chart is shown in Figure 31-2.

Figure 31-2 Normalized chart for the classification of soil types on basis of SPT (Robertson 1990, 2010)

This normalized chart corrects for unequal cone end area effects and some other effects related to
specific characteristics of the CPT. The normalized chart in general provides more reliable identification of
soil types than the non-normalized chart, although when the in-situ vertical effective stress is between 50
and 150 kPa, there is often little difference between the two charts. The advantage of the non-normalized
chart is that it can be used in real-time to evaluate the soil type during and immediately after the CPT.
The normalized chart can only be used after the CPT during post-processing, because it requires
information on soil specific weights and groundwater conditions that are not available during the CPT.

In this double-logarithmic chart, the boundaries between the soil types can be approximated by concentric
circles. A soil type index, describing these circles, can be defined by:

2 2
 qc   
Ic =  3,47 − log( )  +  log Rf + 1,22  (31.5)
 pa   

(use the basic CPT-values for the non-normalized chart)

Instead of the charts, the following table can be used now to identify soil types (which is very suitable for
automated identification of soil types).

182
Manual Hydraulic Structures Part III - 'Materials'

Table 31-3 Soil types related to soil type index Ic

Sand (rules of thumb)

The Young's modulus E results from:
(1 − ν )
Eoed = E (31.6)
(1 + ν )(1 − 2ν )
with:
Eoed = constrained modulus (modulus bij zijdelingse opsluiting):
Eoed = 4 qc 0 < qc < 10 MPa
= 2 qc + 20 MPa for: 10 < qc < 50 MPa
= 120 MPa 50 MPa < qc
ν [-] = Poisson's ratio (dwarscontractiecoëfficiënt): ν = 0,3

qc
Other parameters for sand: c'= 0 and φ ' = 29° +
4 MPa/ °

For sand calculations assume the drained values (so the parameters with an ×', instead of ×u).
If there are no known values of qc or if these values vary too much, one can use the following as a rough
estimate for clean sand: E = 250 σv' and ϕ = 32°.

For unloading and reloading, the following applies (for sand, clay and peat): Eunload / reload ≈ ( 3 ∼ 5 ) ⋅ E '

Clay, peat and loam (rules of thumb)

q
cu = fundr = c (clay/loam)
20
q
cu = fundr = c (peat, humous clay)
30
Eu ≈ E = 100 ⋅ fundr (clay/peat)
Eu ≈ E = 40 ⋅ fundr (loam)
ν = 0,4
ν u = 0,5

For clay, peat and loam one must distinguish between a short term load (undrained, so parameters with
an ×u) and a long term load (drained, so parameters with an ×'). For the drained stiffness E ' the
assumption Eu ≈ E ' does not take creep into account. For calculations concerning the stiffness of
drained soil, it is preferable to use the primary and secondary settlement constants Cp and Cs .

The following applies for unloading and reloading (for sand, clay and peat): Eunload / reload ≈ ( 3 ∼ 5 ) ⋅ E

31.6 Determination of soil parameters with the Eurocode table

The table on the next page is table 1 from Eurocode 7. It shows the representative values Xrep for different
soil types. The table is based on values found in practice in the Netherlands. These values can be used in
the Netherlands as an estimate if no laboratory research is available.

183
Manual Hydraulic Structures Part III - 'Materials'

Table 31-4 Indicative soil properties according to Eurocode 7 NEN-EN9997 (to be verified by on-site soil investigation!)

184
Manual Hydraulic Structures Part III - 'Materials'

31.7 Soil parameters and models

In all engineering sciences an effort is made to link analytical, numerical and physical models. Below the
analytical Mohr-Coulomb model and some numerical models are linked, by means of the soil properties,
to soil investigation results (=’real’ physical model).

Mohr-Coulomb (general)
Strength:
In the Mohr-Coulomb model the following expressions are derived for drained and undrained shear
strengths respectively:

Drained: τ max = c '⋅ cos(φ ) + ½ (σ h '+ σ v ' ) ⋅ sin(φ ) (31.7)

Undrained: fundr = cu = τ max = c '⋅ cos(φ ) + ½ (σ h '+ σ v ' ) ⋅ sin(φ ) (31.8)

The stresses σv’ and σh’ can be assumed equal to the stresses before loading because during the quick
loading the normal effective stresses and thus also the strength of the undrained soil does not change.

The difference between the above mentioned Mohr-Coulomb strength and the Coulomb strength is quite
significant:

τ max = c '+ σ n '⋅ tan(φ ) (Coulomb) (31.9)

Stiffness:
(1 − ν )
Eoed = E (31.10)
(1 + ν )(1 − 2ν )
E
K= (31.11)
3 (1 − 2ν )
E
G= (τ = G γ ) (31.12)
2 (1 + ν )
ν
σ h ' = K 0 ⋅ σ v ' with: K 0 = (which differs from Jáky’s formula: K 0 = 1 − sin φ ) (31.13)
1 −ν

According to Mohr-Coulomb, the difference between the stiffness to be taken into account for drained or
undrained loading of clay is merely a factor:

Eu 1 + ν u Eu
= with : ν u = 0,5 and ν ' = 0,3 ⇒ = 1,15 (31.14)
E ' 1+ν ' E'

In reality the difference is bigger because Mohr-Coulomb does not take creep into account. For clay and
peat it is better to assume:

Eu
≈ 2 to 3 (31.15)
E'

It is even better to use Koppejan’s settlement formulas (Section 34.2) and the primary and secondary
settlement constants CP and CS in calculations.

In the case of unloading and reloading one can assume that a little creep occurs, thus one can presume:

Erelief / reapplication ≈ ( 3 ∼ 5 ) ⋅ E ' (31.16)

Winkler
Soil is often modelled as a set of uncoupled linear springs, known as the Winkler model. See Chapter 33
for a further description of spring schematisations.

185
Manual Hydraulic Structures Part III - 'Materials'

D-sheet Piling (spring model)

D-Sheet Piling (named Msheet in earlier days) is a computer programme for the calculation of (mostly)
sheet pile walls and laterally loaded piles. It is based on Winkler’s spring model. When using the
programme, a choice has to be made between the use of either straight or curved slip planes in the soil.
The physical reality is such that the friction between the sheet piling or pile and the soil results in curved
slip planes instead of straight planes. For hand calculations curved slip planes are far too difficult to
calculate. For convenience, one uses straight slip planes instead. However, the straight slip planes model
calculates a strength that is too optimistic, hence less safe. To arrive at a more conservative, safe,
strength the wall friction angle δ should be reduced more in straight than in the curved slip planes model.
Using D-Sheet Piling, or similar computer programmes, it is anyway better to choose for curved slip
planes, because it is more accurate (the computer does the extra/more difficult computation).

For the calculation of slip planes with wall friction angle δ Eurocode 7, NEN 9997-1+C1:2012 applies:

Qualitative description More specific definition Wall friction angle δ

wall surface Roughness wall surface Straight slip surface Curved slip surface

Toothed > 10 d50 2/3 φ’k ≤ φ’k

o
Rough 0.5 d50 - 10 d50 2/3 φ’k ≤ φ’k – 2.5 with a
o
maximum of 27.5
Half rough 0.1 d50 – 0.5 d50 2/3 φ’k 2/3 φ’k
o o
Smooth 0.1 < d50 0 0
* For clay, loam, sand and pebbles
**
For diaphragm walls the wall friction angle has to be reduced more than for steel walls

The stiffness (modulus of sub grade reactions for the springs) is discussed in Section 33.2 "Modulus of
subgrade reaction". The strength of the soil is covered in Chapter 32.

Plaxis (finite elements model)

Plaxis can calculate using several soil models. All models require a reduction of the friction at the soil-
structure interfaces.

For vertical interfaces one can assume: Rinterface ≈ 0,8 so: tan(δ ) ≈ 0,8 ⋅ tan(φ )

For horizontal interfaces one can assume: Rinterface ≈ 0,67 so: tan(δ ) ≈ 2 ⋅ tan(φ )
3

For this see Section 32.3 "Soil - strength, Horizontal bearing capacity (resistance against sliding)".

If the Soft-Soil (Creep) Model is used in Plaxis, the following applies for clay and peat:

1
λ* = (31.17)
C 'p
1 − ν unloading 1 + 2K 0
κ* ≈ ⋅ with: ν unloading ≈ 0,15 and: K 0 ≈ 0,60 (31.18)
1 + ν unloading Cp
1 1
µ* ≈ ⋅ (31.19)
ln(10) C 's

186
Manual Hydraulic Structures Part III - 'Materials'

31.8 Literature
Paul W. Mayne, Matthew R. Coop, Sarah M. Springman, An-Bin Huang and Jorge G. Zornberg.
Geomaterial behavior and testing. Proceedings of the 17th International Conference on Soil Mechanics
and Geotechnical Engineering, 2009.

Peter K. Robertson. Soil classification using the cone penetration test. Canadian Geotechnical Journal,
NRC Research Press, 1990

Peter K. Robertson. Soil behaviour type from the CPT: an update. 2nd International symposium on Cone
Penetration Testing. 2010

Peter K. Robertson and K.L. Cabal (Robertson). Guide to Cone Penetration Testing for Geotechnical
Engineering. Gregg Drilling & Testing, Inc. 5th Edition, November 2012

A. Verruijt. Soil mechanics. Delft University of Technology, 2001, 2012

187
Manual Hydraulic Structures Part III - 'Materials'

32. Soil - strength (resulting in bearing capacity)

(improved: February 2017)

In general, the bearing capacity of a soil should be sufficient to resist the acting loads:

S < R → σ 'k ,max < p 'max

The maximum value of the resulting vertical pressure applied by a structure, σ'k,max, can be calculated
according to Section 37.3 of this Manual. The present chapter covers some of the classic methods
concerning the bearing capacity of soil, p'max:
• Mohr-Coulomb (general strength theory)
• Prandtl and Brinch Hansen (vertical bearing capacity)
• Horizontal bearing capacity
• Fellenius and Bishop (stability of a slope, including reinforced earth)

32.1 Theory: strength schematization (Mohr-Coulomb)

Various models are available to schematise the strength of soil. The most important models are the Mohr-
Coulomb model and the Cam Clay model. This manual only discusses the Mohr-Coulomb model. Mohr’s
circle of stress shows all stress combinations possible in the soil. Coulomb’s line of collapse shows which
combination leads to collapse and when. These two parts together form the Mohr-Coulomb model.

Mohr
Mohr’s circle of stress shows all stress combinations in a two-dimensional space, given that there are two
planes with zero shear stress, in which only a normal stress acts. The directions of the normal stresses in
these two planes are called the principal directions. If the stresses in two planes are known, all states of
stress are defined by Mohr’s circle (Figure 32-1).

Figure 32-1 Mohr’s circle

Coulomb
Coulomb assumes a strength that is expressed by the shear resistance. If this resistance to shear is
exceeded a plane of slip is created. The maximum shear resistance is:

τ max = c '+ σ n ' tan(δ )

in which: τmax [kPa] = max shear stress on a given plane

c’ [kPa] = cohesion
δ [°] = angle of internal friction
σ’n [kPa] = the effective normal stress in a given plane

As long as the shear stress τ is smaller than c '+ σ n ' tan(δ ) , no sliding will occur.

188
Manual Hydraulic Structures Part III - 'Materials'

Mohr-Coulomb
Contrary to Coulomb’s collapse, the Mohr-Coulomb model assumes that not one (σn’ ) but two stresses
(σv’ and σh’) determine the maximum failure value of the shear stress (or shear resistance) τ. After all,
Mohr’s circle considers whether collapse is possible in all directions. Figure 32-2 shows the maximum
absorbable shear stress in the σ,τ plane.
τ’

r
r
τ’max
φ c’
σ’n σ’v
σ’
σ’h
τ’max
r

Figure 32-2 Maximum absorbable shear stress according to Mohr-Coulomb

According to Mohr-Coulomb the following applies: τ max = c '⋅ cos(ϕ ) + ½ (σ h '+ σ v ' ) ⋅ sin(ϕ )

Therefore, for undrained materials (clay and peat): cu = τ max = c '⋅ cos(ϕ ) + ½ (σ h '+ σ v ' ) ⋅ sin(ϕ )

For this the stresses σv’ and σh’ before the application of the load can be assumed because the water
pressure absorbs all normal stresses during the application of the load. Consequently, the shear stress
remains constant.

32.2 Vertical bearing capacity

Rule-of-thumb
Only for very preliminary designs, it could be considered to use some standard values for the vertical
bearing capacity under static loading, as given by British Standard 8004 (not prevailing any more), see
Table 32-1.

Table 32-1 Presumed allowable bearing values under static loading (from former BS 8004)

In the Netherlands, the bearing capacity for vertically loaded structures, used for preliminary designs,
should not exceed 500 kPa (the design value should not exceed 400 kPa), according to CUR Handboek
Funderingen.

189
Manual Hydraulic Structures Part III - 'Materials'

Prandtl & Brinch Hansen

The Brinch Hansen method is often used for determining the maximum bearing capacity of the soil under
a foundation. This method is based on Prandtl’s theoretical slip surfaces (Figure 32-3).

Figure 32-3 Slip surface according to Prandtl and maximum depth and width of the slip body according to Buisman
(De Smedt, 2013)

Buisman found equations for the maximum depth (Dmax) and width (Bmax) of the slip body (Figure 32-3).

π 
B ⋅ cos ϕ αf ⋅tanϕ  ⋅ tanϕ 
Dmax = e and Bmax = B ⋅ tan α f ⋅ e 2  = B⋅ N
q
2 ⋅ cos α f
where
φ = angle of internal friction,
π ϕ
αf = angle as depicted in Figure 32-3: α f = + (α and φ in radians)
4 2
Nq = bearing capacity factor (see next page)

These equations apply to homogeneous soil, but they could give an indication of the maximum
dimensions of the slip body in layered soil. The depth Dmax of the slip body is restricted to 3B and the
horizontal dimension Bmax to 10B. For undrained soil, the angle of internal friction φ is zero, in which case
Dmax = B/√2 en Bmax = B.

NEN 9997-1+C1, article 6.5.2.2, relates Dmax and Bmax to the angle of internal friction, φ', see Figure 32-4.

Figure 32-4 Maximum depth and width of the slip body as a function of the angle of internal friction (NEN9997-1+C1)

190
Manual Hydraulic Structures Part III - 'Materials'

The maximum bearing force that a soil can resist can be calculated with:

Vmax = p 'max ⋅ A eff

where:
p'max [kPa] = bearing capacity of the soil, see the equations below for drained and undrained
conditions
2
Aeff [m ] = effective foundation area = L·B, see explanation at the end of this Section

The bearing capacity p'max depends on the condition of the soil, which can be drained or undrained.

The drained condition indicates a soil state in which any excess pore water pressure has fully dissipated.
This does not mean that all the water has flowed out of the pores, but it means that there is no stress-
induced pressure in the pore water. Granular soils drain very quickly, and hence they are always treated
as drained and analysed in terms of effective stresses using φ' (c'=0). Also the bearing capacity of clay
and peat, exposed to long-term loads (months or years), should be calculated as under drained condition.

The undrained condition occurs in cohesive impermeable soils, in which pore water pressures increase
directly after the load is applied. In undrained soil, an undrained shear strength, the shear strength cu (=
su) is used instead of the effective cohesion c', which is used for drained soil conditions. For undrained
conditions, an internal friction angle of φ' = 0 is used.

Both conditions should be checked in a bearing capacity calculation. In addition, punching through a
possible lower soft layer should be checked for the drained condition (not explained in this Manual).

Drained conditions
The soil bearing capacity in drained materials is usually expressed by:

p 'max, drained = c ' Nc sc i c + σ 'q Nqsq i q + 0,5γ ′B ⋅ Nγ sγ iγ

which consists of contributions from cohesion (index c), surcharge including soil coverage (q) and the
specific weight of the soil below the foundation (γ) (Figure 32-5).

Figure 32-5 Collapse of soil under a structure according to Prandtl and Brinch Hansen

The main bearing capacity factors can be calculated with:

1 + sin ϕ ′ π tan ϕ ′
Nc = ( Nq − 1) ⋅ cot ϕ ′ Nq = ⋅e Nγ = (Nq − 1) ⋅ tan(1,32 ⋅ ϕ ′)
1 − sin ϕ ′

The shape factors ( B ≤ L ≤ ∞ ) are:

sq Nq − 1 B B
sc = sq = 1 + sin ϕ ′ sγ = 1 − 0,3
Nq − 1 L L

The presence of an inclined load, implying a horizontal component, reduces the bearing considerably.
The load inclination factors to deal with an inclined direction of the resulting force ( B ≤ L ≤ ∞ ) are:

i q N q −1 H
If H parallel to L and L /B ≥ 2 : i c = ; i q = iγ = 1 −
Nq − 1 V + Ac ′ cot ϕ ′

191
Manual Hydraulic Structures Part III - 'Materials'

i q N q −1
3 3
 0,70 ⋅ H   H 
If H parallel to B : i c = ; iq = 1 −  ; iγ =  1 − 
Nq − 1  V + A ⋅ c ′ ⋅ cot ϕ ′   V + A ⋅ c ′ ⋅ cot ϕ ′ 

Only the part of the foundation slab which has effective stresses underneath is included in the effective
width B. The factors for the bearing force are also given in Figure 32-6.

Figure 32-6 Bearing force factors as function of the angle of internal friction

Undrained conditions
The bearing capacity for undrained can basically be calculated according to the equation for drained
conditions, but now the angle of internal friction, φ, equals zero, which implies that Nγ = 0 and Nq =1,0.
Instead of the cohesion c', the undrained shear strength cu should be used.

The equation for the bearing capacity thus modifies into:

p 'max, undrained = cu Nc sc i c + σ 'q

where:
 d
Nc = (2 + π )  1 + 0,27 
 B 

B
s c = 1 + 0,2
L
 H 
ic = 0,5 ⋅  1 + 1 − 
 A ⋅ cu 

d [m] = foundation depth
for the rest, see above, under 'drained conditions'

192
Manual Hydraulic Structures Part III - 'Materials'

Clarification of the symbols used in this Section:

p'max [kPa] = bearing capacity of the soil
c' [kPa] = (weighted) cohesion
cu [kPa] = design value of the undrained shear strength (= su)
σ'q [kPa] = effective surcharge pressure surrounding a foundation element at construction
i =n
level: σ q′ = q '+ ∑ ( d i .γ 'soil ,i ) , where:
i =1

q' [kPa] = effective surcharge acting on the ground surface

n [-] = number of horizontal soil layers between the construction
level and the ground surface
di [m] = thickness of soil layer i
γ'soil,i [kN/m ] = effective volumetric weight of soil layer i
3

γ'soil
3
[kN/m ]= (weighted) effective volumetric weight of the soil below construction level
×c [-] = factor for the influence of cohesion
×q [-] = factor for the influence of the soil cover
×γ [-] = factor for the influence of the effective volumetric weight of the soil
φ' [°] = (weighted) angle of internal friction
H [kN] = horizontal component of the resulting force acting on the soil
V [kN] = vertical component of the resulting force acting on the soil
2
A [m ] = effective foundation area: A = B·L
L [m] = length of the effective foundation area, for circular slabs: L = B
B [m] = width of the effective foundation area, for circular slabs: L = B
For an eccentrically loaded foundation, B is approximated by Btot - 2eB
and L by Ltot - 2eB , see the following figure from NEN9997-1+C1:

Notes
• The factors for the bearing force are too conservative. This is because it is assumed that the factors
do not influence each other, which is incorrect.

• The factors for the shape of the foundation and particularly for the horizontal load are not
substantiated scientifically but they are based on empirical relations, experiments, calculations, etc.

• For a horizontal load H one not only has to apply a reduction in the calculation of the vertical
bearing force F, but one also has to check if sliding can occur (e.g. using Coulomb).

• The reduction of the vertical bearing force V as a result of the horizontal load H is considerable.
Assuming H/V = 0,30 results in iq = 0,50!

193
Manual Hydraulic Structures Part III - 'Materials'

Layered soil
If the foundation’s subsoil is layered, one can fill in the weighted soil properties in the previous equations
according to:
i =n i =n i =n

∑h i ϕi; X i ∑h i c i Xi ∑h i γ ′i X i
ϕ' = i =1
i =n
c′ = i =1
i =n
γ′ = i =1
i =n

∑h
i =1
i Xi ∑h
i =1
i Xi ∑h
i =1
i Xi

in which:
φ'i [°] = angle of internal friction for layer i
c'i [kPa] = cohesion of layer i
γ'i
3
[kN/m ] = effective specific weight of layer i
n [-] = the number of horizontal layers between the influence depth Dmax and the
construction level
hi [m] = thickness of layer i (within the influence depth Dmax)
Xi [m] = distance between the centre of layer i to the influence depth Dmax
Dmax [m] = the influence depth, see Figure 32-3

The previous equations can be used to determine the dimensions of a foundation in an iterative way. As a
first estimate, one often uses a rule of thumb which assumes that the soil under and next to the
foundation does not collapse if the work-line of resultant force intersects with the core of the foundation
surface (see Section 37.2).

32.3 Horizontal bearing capacity (resistance against sliding)

Theory
Horizontally loaded foundations transfer their load to the subsoil by means of friction in the foundation
plane and possibly passive earth pressure behind the foundation. To calculate the maximum friction
against the foundation surface one normally uses Coulomb’s formula:

τ max = f ⋅ σ n '

where: τmax [kPa] = maximum shear stress in the foundation surface

f [-] = coefficient of friction = tan(δ)
δ [°] = angle of friction between foundation slab and soil
σn’ [kPa] = the effective normal (vertical) stress under the foundation

One might wonder whether the cohesion of a soil layer should also be included in the formula above, but
on grounds of sensitivity to settlements, foundations are never situated directly on cohesive materials but
always on sand or gravel, which makes this question irrelevant.

The coefficient of friction between concrete and sand is usually in the order of 40% to 50%. For the
coefficient of friction between steel and sand roughly the same values are found.

The angle of friction parameter δ is a parameter that belongs to Coulomb’s model of friction. The angle of
internal friction φ is a parameter that corresponds to Mohr-Coulomb’s model of friction. These are
different models. Coulomb’s model of friction is based on one stress: the normal stress σn’, whilst Mohr-
Coulomb’s model of friction is based on two stresses: σv’ and σh’.

Assuming the normal stress σn’ and the stress perpendicular to this σ ⊥ ' it is possible to work φ into δ :

1 σ⊥ '
tan(δ ) = (1 + K ) ⋅ sin2 (ϕ ) − (1 − K )
2 2
with: K =
2 σn '

If K = 1 then: tan(δ ) = sin(ϕ )

194
Manual Hydraulic Structures Part III - 'Materials'

Design
Usually σ h ' < σ v ' is valid for soil layers, because K 0 < 1.
For horizontal surfaces the normal stress (vertical) is therefore greater than the stress perpendicular to
this (horizontal):

σn ' > σ⊥ '

By approximation one can assume:

τ max = tan(δ )σ n ' with: δ ≈ 2 φ

3

For vertical surfaces the normal stress (horizontal) is therefore smaller than the vertical stress perpen-
dicular to this:

σn ' < σ⊥ '

By approximation one can assume:

τ max = tan(δ ) ⋅ σ n ' with: δ ≈ (0,8 to 0,9) ⋅ ϕ

The values given above can be derived theoretically. Furthermore, these values are also found
experimentally.

32.4 Stability of slopes (Fellenius and Bishop)

Many methods for checking the stability of slopes are based on circular slip surfaces. The earth mass
above the failure arc is divided into slices, of which one, with a width b and (average) height h, has been
drawn. If one considers the moments relative to the centre of the circle, the following applies for the
driving moment:

Ma = Σ γ ⋅ b ⋅ h ⋅ R ⋅ sin (α )

which considers the sum of the contributions of all slices and where γ is the volumetric weight of soil. If
the ground mass is composed of several layers, one may have to deal with different volumetric weights
over the height h. If the water table is found above the slip plane one must take into account the
volumetric weight of wet soil below the water table and the weight of dry soil above it (unless there is also
capillary rise, as tends to be the case particularly for clay and peat). Possible water above ground level
should also be taken into account.

Figure 32-7 The principle of a slip surface calculation

The resisting moment is:

1
Mw = Σ τ b R
cos( α )
in which b/cos(α) approximates the length of the circumference over the slice width b.
Fellenius assumes that the shear stress τ along the whole circumference is a factor F smaller than the
maximum shear stress that can be developed. This factor is also called the stability factor and it can be
considered a type of safety coefficient.

195
Manual Hydraulic Structures Part III - 'Materials'

Therefore, the shear stress can be denoted as:

1
τ= ( c ' + σ n '⋅ tan(ϕ '))
F

in which: c' [kPa] = cohesion

σn' [kPa] = normal effective stress, with σ n ' = σ n - p
φ' [-] = angle of internal friction
p [kPa] = water pressure

Fellenius
Fellenius assumes that the slices do not exert any force on each other. The balance of forces along the
radial leads to:

1
σ n '⋅ b = γ b h cos (α ) ⇔ σ n ' = γ h cos2 ( α )
cos α

Inserting the normal stress produces:

τ=
1
F
( c'+ ( γ h cos2 (α ) - p) tan (ϕ ') )
A balance of moments around the centre of the circle means that the driving moment Ma equals the
resisting moment Mr. Together this results in the following formula by Fellenius for safety:

F=
{
Σ c '+ ( γ h cos2 α - p ) tan ϕ ' / cos α }
Σ γ h sin α

Bishop
Due to the shear force that is created under a slice, the vertical balance should really be corrected, so
that:
sin α
γ h = σ n '+ p + τ ⋅
cos α

Inserting this in the driving moment Md leads to Bishop’s iterative formula:

c '+ ( γ h - p ) tan ϕ '

Σ
cos α (1 + tan α ⋅ tan ϕ '/ F )
F=
Σ γ h sin α

The stability check consists of dividing the soil, for instance into 10 slices of equal width, measuring α and
h for each slice and then determining the safety factor F. The verification is carried out for several slip
planes (different centre points and radii). The smallest value of F (corresponding to the most critical slip
plane) has to be greater than 1,2 for temporary building works (e.g. excavated building sites) or 1,3 for
permanent works. The aforementioned method of calculation entails a lot of work. Particularly if one
considers that several slope angles have to be investigated in order to find the steepest possible slope.
This is why the calculations are always carried out by a computer programme (e.g., D-Geo Stability).

tan ϕ '
For an infinite slope of dry, non-cohesive material, one finds: F =
tan α
Note:
One should not value the calculated safety factor too much, because the entire method is questionable.
Some of the flaws are:
• the circular slip plane is not entirely correct because the soil can deform.
• the calculations only include the balance of moment and not the complete balance of force.
• the transfer of forces between the slices is not taken into account.

196
Manual Hydraulic Structures Part III - 'Materials'

The groundwater pressures are not always well known, but they are very important. A high
groundwater pressure reduces the safety factor. In little permeable soil types, the water pressure
certainly needs (a lot of) time to adjust to the drainage and to the dewatering of an excavated building
site. Using values for the water pressure that are too optimistic in calculations has lead to the collapse of
slopes (sliding) in many cases. Measures against this include:
• designing with more realistic values (and thus shallower slopes).
• dewatering the area earlier.
• draining the excavated building site slowly.
• applying sand drains or geo-drains for faster drainage of little permeable soil layers in places
where slip planes can be expected.

Reinforced earth (Terre Armée)

A reinforced earth wall is a wall consisting of prefab slabs piled on top of each other. Two strips of geo-
textile (or zinc-coated steel), which disappear horizontally in the soil behind the wall, are attached to each
slab. Due to the weight of the soil, the strips can absorb a tensile force and hold the individual wall
elements in place. The collapse mechanism of the entire wall can be solved with Fellenius and Bishop but
that requires calculating the extra tensile force of the strips correctly. The part of the strip outside of the
slip circle (see Figure 32-8) can absorb an additional force in the order of:

Fstrip = 2bl eff σ 'v tan δ

O

Fstrip

Figure 32-8 Extra stability due to reinforced earth

Note
Next to no computer programmes can calculate reinforced earth. The programmes that do calculations
with geo-textiles (nearly) never include the correct effective length in calculations (including Mstab).
Plaxis, however, does calculate this length properly.

197
Manual Hydraulic Structures Part III - 'Materials'

33. Soil - stiffness

Schematizations of soil stiffness are used for determining the flow of forces through the structure and for
estimates of the displacement of the structure. The two most frequently used models:
1. Winkler model, based on the schematization of the soil into a set of uncoupled linear springs.
2. Elastic solid foundation model. This model is based on the relation between stress, strain and the
Young’s modulus.
For conceptual design the use of the Winkler model is often more convenient, therefore the Elastic solid
foundation model will not be elaborated here.

33.1 Spring schematisation

For design calculations involving the stiffness properties of soil, the soil is often modelled as a set of
uncoupled linear springs, known as the Winkler model. This is a popular schematization because it is a
simple model, easy to use.

Foundation strip

Beam on linear Shear stress between

uncoupled springs soil springs neglected
Figure 33-1 Winkler foundation model

Since a linear relationship between the action on the foundation, suppose pressure p, and the deflection
w is assumed:

p = ko · w
2
where ko [N/m /m] = foundation modulus

For beams with width b, we use p = k · w = ko · b · w (unit of k: N/m/m).

If the foundation in infinitely stiff the foundation modulus ko is in fact fully determined by the soil
characteristics. For the infinite stiff foundation, ko is better known as the modulus of subgrade reaction k
(beddingsconstante).

Note: the modulus of subgrade reaction is in fact not a soil parameter; i.e. not only soil parameters
influence its value but the geometry of the structure as well.

As downside of the simple Winkler model the following can be mentioned:

• The springs are not linked or coupled to each other (see Figure 33-1), while in reality the soil
elements are linked by means of shear stresses.
• The spread of stress in the subsoil is not really included in the schematisation because of the
independent, uncoupled behaviour of every spring.
• The soil is not a spring and definitely does not behave as a linear one.
• Springs have a spring constant k (generally k = EA/l). The schematisation as spring therefore
requires a “contributing depth”, to account for l, which is unknown.

Figure 33-2 shows the real and the calculated pressure distributions under a very stiff foundation, using
the linear spring model for soil. The biggest difference is the soil pressure near the edges of the
foundation.

198
Manual Hydraulic Structures Part III - 'Materials'

schematised beam
on springs

True situation

calculated pressure true pressure

distribution distribution

Figure 33-2 Calculated versus true pressures

In the case of an infinitely stiff plate, the foundation pressure is, neither in reality nor according the
continuum theory, constant. The pressure along the edge is, after all, higher than in the middle. According
to the “spring” method, the pressure is the same everywhere.

For foundations of limited dimensions, settlements not only depend on the stiffness of the soil and the
stresses in the foundation plane, but also on the dimensions of the structure. Figure 33-3 shows the
settlements for a foundation slab of ø 75 cm and for a slab of ø 150 cm. In both cases, the settlement of
the largest foundation slab is bigger than the settlement of the smaller plate, for the same soil stiffness
and equal pressures on the foundation surface.

Figure 33-3 measured settlements for two foundation slabs Figure 33-4 Influence depth of a foundation strip

This can be explained by the spread of the load in subsoil. Figure 33-4 shows the influence of the size of
the foundation slab on the spread of the load in the subsoil. By approximation the depth of influence of
the foundation is directly proportional to the dimensions of the foundation area. The area of the foundation
should therefore be taken into account in the calculation of the modulus of sub grade reaction.

33.2 Modulus of subgrade reaction

Estimating the stiffness of a structure is usually relatively easy. The bending stiffness (or flexural rigidity)
of a beam is characterised by its EI value. Estimating the stiffness of soil, however, is more difficult.
The modulus, or coefficient, of subgrade reaction (beddingsconstante) is a spring constant for soil and is
defined as:

199
Manual Hydraulic Structures Part III - 'Materials'

σ
k=
δ

in which: k [Pa/m] = modulus of subgrade reaction

σ [Pa] = the (effective) stress at the bottom of, under the foundation surface
δ [m] = the displacement of the foundation surface

This section discusses three different types of modulus of sub grade reactions:
• Vertical modulus of subgrade reactions
• Horizontal modulus of subgrade reactions for (sheet pile) walls
• Horizontal modulus of subgrade reactions for piles

Often estimates are made from global tables, like Table 33-1. However, it is best to obtain estimates
based on in-situ testing.

Table 33-1 Modulus of subgrade reaction according to Terzaghi (1955)

su is the undrained shear strength

33.3 Vertical modulus of subgrade reaction (using Flamant)

The solution for surface loads was already treated in Section 24.1. The stresses in an arbitrary point in a
homogenous half space are:

qv qv
σ 'zz = (θ − θ ) + sinθ1 cos θ1 − sinθ 2 cosθ 2  σ ' xz = cos2 θ 2 − cos2 θ1 
π  1 2 π 
qv
σ ' xx = (θ − θ ) − sinθ1 cosθ1 + sin θ 2 cosθ 2 
π  1 2

Because all stresses have been solved, all strains and therefore also all displacements can be solved.
The following is valid for the z-axis:

2 qv
ε zz( x =0) = θ (1 − 2ν ) + sinθ1 cosθ1 (1 + ν )  ⇒
πE  1
z =∞
2 a qv
u z( x = 0) = ∫ ε zz ( x = 0)δ z ≈ (1 + ν ) 1,762 ⋅ (1 − 2ν ) + 1,443 
z =0 E

where ν = poisson's ratio of the soil [-].

The modulus of subgrade reaction for the z-axis is then:

qv E 1
k( x = 0) = ≈ ⋅
uz ( x =0) 2a (1 + ν ) ⋅ (1,762 ⋅ (1 − 2ν ) + 1,443 )

A reasonable estimate for the modulus of subgrade reaction of drained soil is:

E
k ≈ 0,31⋅ (ν = 0,3)
2a

200
Manual Hydraulic Structures Part III - 'Materials'

A reasonable estimate for the modulus of subgrade reaction of undrained soil is:

E def def 1+νu

k ≈ 0,53 ⋅ (ν u = 0,5 and Eu = E)
2a 1+ ν

The geometry (width of the load: 2a) is clearly present in the modulus of sub grade reaction, which proves
that the value of k depends on the geometry of the structure.

For line loads one does not use the modulus of subgrade reaction per area unit, but per length unit of the
2 3
line load. This modulus of subgrade reaction k' has a different dimension ([kN/m ] instead of [kN/m ]) and
is independent of the width of the line load, because:

k
k' = ≈ 0,31⋅ E (ν = 0,3)
2a

33.4 Horizontal modulus of subgrade reaction for (sheet piling) walls

This section explains what modulus of subgrade reaction should be used according to CUR 166 (Sheet
pile walls) to calculate a sheet pile wall with a spring model. The figure below is used to calculate the
horizontal stress-displacement relationship.

σ’ h
1.0 σ’ h,max
k3
0.8 σ’ h,max 1

k2
0.5 σ’ h,max 1

k1
1
σ’ h,n
u
σ’ h,min

Figure 33-5 Horizontal stress-displacement relationship

In the figure the following applies:

σ h' ,min = K a σ v' − 2 c K a

σ h' ,max = K p σ v' + 2 c K p
σ h' ,n = K 0 σ v'
where:
K0, Ka and Kp are the neutral, active and passive soil pressure coefficients (see Section 24.2).

The stress-displacement relationship is split into three branches. The moduli of subgrade reaction for
these three branches are given in a table of CUR-report 166 (Table 33-1), which is based on
measurement data obtained in the Netherlands.

k1 applies to σ'0,h < σ'h < 0,5 σ'h,max

k2 applies to 0,5 σ'h,max ≤ σ'h ≤ 0,8 σ'h,max
k3 applies to 0,8 σ'h,max < σ'h < 1,0 σ'h,max

where σ'0,h [kPa] = neutral soil pressure

σ'h [kPa] = max. horizontal soil pressure for the concerning part of the modulus of subgrade
reaction
σ'h,max [kPa] = representative value of the max. horizontal soil pressure (passive) after
increase of the pressure

201
Manual Hydraulic Structures Part III - 'Materials'

3 3 3
k1 [kN/m ] k2 [kN/m ] k3 [kN/m ]
klow khigh klow khigh klow khigh
sand qc [MPa]
loose 5 12 000 27 000 6000 13 500 3000 6750
moderate 15 20 000 45 000 10 000 22 500 5000 11 250
firm 25 40 000 90 000 20 000 45 000 10 000 22 500
clay fundr [kPa]
weak 25 2000 4500 800 1800 500 1125
moderate 50 4000 9000 2000 4500 800 1800
firm 200 6000 13 500 4000 9000 2000 4500
peat fundr [kPa]
weak 10 1000 2250 500 1125 250 560
moderate 30 2000 4500 800 1800 500 1125
Table 33-1 Representative values of the horizontal moduli of sub grade reaction (CUR 166, 5th edition)

Note
• To calculate the displacement normative k-values are used, i.e.: klow.
• To calculate the forces and moments, two calculations are carried out. One with klow./1.3 and a
second with khigh. It is impossible to tell which of the two is normative in advance, although in
most cases the calculation with the low modulus of sub grade reactions is normative.

33.5 Horizontal modulus of subgrade reactions for piles

For horizontally loaded piles horizontal modulus of subgrade reactions are used. The pressure distribution
around a pile is completely different to the distribution around a wall. This also means different modulus of
subgrade reactions are to be used. According to Menard these constants can best be based on the radial
elasticity modulus Emenard . According to Menard, the following relationship between the modulus of
subgrade reaction and the radial elasticity modulus is valid for piles:

α
1 1   D  D
=  4,0 ⋅  2,65 ⋅  +α ⋅  (D ≥ 0,60 m )
K h 3 ⋅ Emenard   0,60m  2
 
4,0 ⋅ ( 2,65 ) + 3α
α
1 D
= ⋅ (D < 0,60 m )
K h Emenard 18

One can measure Emenard by blowing up a balloon in a boring hole and by measuring which radial defor-
mation occurs at a given pressure. In most cases, however, Emenard is not measured. Usually the cone
resistance qc has been found from soundings. In that case the following empirical relationship between
Menard’s elasticity modulus and the cone resistance can be used:

Emenard = qc ⋅ f

Type of soil f α
gravel - 1.4
sand 0,8 to 1,0 1/3
silt 2,0 to 3,0 1/2
clay 3,0 2/3
peat 3,0 to 4,0 1.0
Table 33-2 Empirical relationships

One can calculate using the dynamic stiffness for short-term loads, such as a collision against a bridge
piling. The dynamic stiffness is greater than the static stiffness, for instance:

Edynamic = 3 ⋅ Estatic

This linear stiffness behaviour is limited by the strength behaviour of piles. The derivation of the three-
dimensional Kp values (with shell behaviour) is complex and goes beyond the scope of this manual. See
"The ultimate resistance of rigid piles against transversal forces" by Brinch Hansen. This is a standard
part of the single pile module of the computer programme D-Sheet Piling.

202
Manual Hydraulic Structures Part III - 'Materials'

34. Soil - settlement

slightly improved: February 2011; revised: February 2015; slightly improved: February 2017

Settlement is the process of soil compression. It is an important phenomenon for hydraulic engineering
because it results in a decrease of the ground level including all structures resting on shallow foundations.
Especially for flood defences this is relevant because settlement will reduce their water retaining height,
which immediately affects their primary function if the design does not sufficiently takes this effect into
account. Furthermore, uneven settlements should be avoided or compensated for, in order to prevent
structural failure due to too high stresses in the structure.

Settlement can take place in the subsoil (zetting) or in an embankment (klink). Land subsidence, by the
way, also causes a drop in ground level, but this takes place at another geological scale and should be
taken into account when designing flood defences.

Settlement can mainly be caused by two different natural phenomena. It can be caused by expulsion of
water out of the pores between the soil particles due to an increase in vertical loading, which is called
'consolidation'. In can also be the result of an ongoing, slower, densification process during which the soil
properties gradually change, which is called 'creep' (kruip), 'secondary settlement' or 'secular effect'.

These phenomena and the opposite effect of compression (relaxation of soil) are described in the
following sections, starting with consolidation.

34.1 Consolidation
Theory
Consolidation is the process that involves a decrease in the water content of a saturated soil due to
loading, without replacement of water by air. Especially long-term static loads cause expulsion of water,
resulting in a reduction in volume without changing its shape. The degree of consolidation, U, indicates
how much water (pressure) has already dissipated. The following applies for one-dimensional (= vertical)
consolidation:
 
 
2 π 2 ⋅ cv t 
 −  2 j −1 ⋅

∞ 
8 1    4 d 2

U = 1 − ∑ ⋅e
c
 
,
π2 ( 2 j − 1)
2
j =1

or by approximation:

π 2 ⋅ c v t 
−


2 cv ⋅ t 8  4 d c2 
U≈ ⋅ if (U ≤ 0,5) and U ≈ 1− ⋅e  
if (U > 0,5)
π d c2 π2

2
in which: cv [m /s] = coefficient of consolidation (oedometer test)
dc [m] = drainage path length (drainage height). If a soil layer is drained on the top and
from below, the drainage path length is half the height of the soil layer.
t [s] = duration

The consolidation "constant" cv can be derived from Darcy's law:

kv
cv =
γ w mv

No sensible estimates can be made for general values of cv, as both the vertical permeability kv and the
vertical soil stiffness mv vary too much per soil layer. Furthermore, kv and mv decrease during
consolidation due to their dependence on the stresses! The only reasonable solution is to do oedometer
tests for the correct stress path. To have an idea of the order of magnitude, the value of cv of clay can
2 2
have values of 0,5 to 10 m /year and for organic silt it can be 1,6 to 50 m /year.

203
Manual Hydraulic Structures Part III - 'Materials'

According to Terzaghi (1943), the duration of one-dimensional consolidation until U = 1 (= hydrodynamic

duration) can be computed with:
d c2 ⋅ Tv
th =
cv
where: th [s] = hydrodynamic duration
dc [m] = (vertical) consolidation path length
2
cv [m /s] = coefficient of consolidation
π
Tv [-] = dimensionless time factor: Tv = U2 for U < 0,6
4
Tv = −0,933 ⋅ log(1 − U ) − 0,085 for U > 0,6

At the end of the hydrodynamic period, U =1 and Tv = 2.

Note
Depending on how the excess pore-water is dissipated, two or three-dimensional consolidation is
involved. In literature one can find equations for the adjustment of the water pressure in time for both
two and three-dimensional consolidation (Lambe and Whitman, 1969). Theoretically, in the case of one-
dimensional loading (oedometer test, horizontal strain is zero), 100% of the load goes to the water
pressure and 0% to the effective stress (so: U0 = 0). In the case of plain stress loading (triaxial test,
horizontal stress is constant), 33% goes to the water pressure and 67% to the vertical effective stress (U0
= 0,67). In a combination, plain-strain loading (biaxial test) 50% goes to the water pressure and 50% to
the effective stress (U0 = 0,50). In reality, most loads are in principle something in between a one-
dimensional load and a plain-strain load. That is why it would be better to use U0 ≈ 0,25. The total
consolidation process goes therefore much faster than the equation of the degree of consolidation
suggests. More information on this topic can be found in other literature (A.Verruijt, 1987).

Vertical drainage
The best-known vertical drainage systems are sand drains and synthetic geo-drains. Vertical drainage is
applied to accelerate the consolidation process. The most important reason for accelerating the
consolidation process is the problem that one starts construction too late. An additional matter is that the
drainage costs can sometimes be earned back by interest savings by being able to purchase the land
later.

Drawbacks of vertical drainage are:

• the extra costs,
• the environmental drawbacks (plastic in the soil),
• the degradation of the seal against vertical water transport (“piercing the polder”).

One must realise that the vertical drainage does not accelerate the secondary settlement (creep). If one
doesn’t satisfy a certain residual settlement requirement, vertical drainage is not a solution in all cases.
Preferably one should consider working with a longer settlement period or with temporary embankments.

In the case of vertical drainage, there is a constant horizontal drainage distance for all soil layers. This
makes calculating the degree of consolidation per layer easier:

−8Th
µ
Uh = 1 − e

where:
Uh [-] = horizontal degree of consolidation (for vertical drainage)
Th [-] = time factor for horizontal flow:
c ⋅t
Th = h 2 , where:
D
2
ch [m /s] = horizontal consolidation coefficient (oedometer test)
D [m] = equivalent drain distance:
D = 1,05 ⋅ ℓ (for triangular pattern of drainage)
= 113 ⋅ ℓ (for square pattern of drainage)
ℓ [m] = distance between two vertical drains

204
Manual Hydraulic Structures Part III - 'Materials'

kh
ch =
mv ⋅ γ w
µ [-] = a function mainly related to drain spacing and size and the extent of soil disturbance
due to drain installation (smear effect). The basic form of µ for ideal drain with
no smear effect can be expressed as follows:
n2  3 1 1 
µ= 2  ln(n ) − + 2  1 − 2   , where:
n − 1 4 n  4n  
D
n [ −] =
d
2(b + t )
d [m] = equivalent drain thickness: d =
π

Notes
• In some literature 'consolidation' has the meaning of the primary settlement. In this Manual
'consolidation' is considered as the cause of primary settlement.

• Densification of an embankment ('klink' in Dutch) can consist of three effects: primary settlement,
secondary (creep) settlement and lateral deformation. Lateral, or horizontal, deformation
(expansion) does not influence the total volume, so it results in a decreased embankment height.

• The horizontal permeability of a soil layer can be greater (or smaller) than the vertical permeability.

• The distance ℓ between two drains is generally between 1 and 3 m.

• A triangular pattern is always more economical than a square pattern.

• The cross-section of narrow synthetic drains is b × t = 100 × 4 mm2. The cross-section of wide
synthetic drains is b × t = 300 × 4 mm2.

34.2 Primary settlement and creep

In practice, the compressibility of soil is an important property regarding settlement problems.
Compression is the decrease in volume without the change of shape.

Soil is not a linearly elastic material. Generally the stiffness increases with an increase of the average
compressive stress (σ'0) caused by higher loading. Due to this increase in loading the soil particles will
come closer together (the porosity decreases), increasing the number of contact points and enlarging the
areas of contact. As a result, water residing in the pores will be expulsed. This phenomenon is called
'primary settlement'.

A related phenomenon is secondary settlement, also called 'creep' or 'secular effect'. Creep is the
ongoing compression under a constant load, where the soil properties gradually change. Also here water
is expulsed out of the pores. This compaction goes on practically forever, also if the pore pressures have
been long reduced to zero. Internationally, the Anglo-Saxon method is often used to determine soil
settlement. This method comprises primary and secondary settlement and is expressed by the symbol ε
that represents the relative settlement, so the reduction of the height of a soil layer (∆H) divided by the
initial layer height (H):

ε = εp + εs ,

n
and ∆H = ε·H . The total settlement of n soil layers thus is ∆Htotal = ∑ ∆Hi .
i =1

205
Manual Hydraulic Structures Part III - 'Materials'

A well-known equation for the primary compression comes from Karl von Terzaghi:

Cc  σ ' + ∆σ 'v 
εp = log  v ;i
 
1 + e0  σ 'v ;i 

in which: εp [-] = relative primary settlement = ∆Hp / Hp

Cc [-] = primary compression coefficient
e0 [-] = initial void ratio (poriëngetal) (see Section 34.3)
σ ’v;i [kPa] = initial vertical effective stress
σ ’v [kPa] = (new) vertical effective stress = σ ’v;i + ∆σ v’

The following equation for creep has been developed by prof. Keverling Buisman:

 ∆t 
ε s = Cα log  
t
 ref 

in which: εs [-] = relative secondary settlement (= creep) = ∆Hs / Hs

Cα [-] = secondary compression coefficient
∆t [day] = duration of the secondary settlement
tref [day] = reference duration (one day)

This formula has no advantages, but it does have three disadvantages:

• The initial pore value e0 is unknown and difficult to estimate.
• The formula for secondary compression is not intended for weak soil.
• The formula for secondary compression is nonsense. The value of the pressure difference ∆σ’ is not
even included in the formula!

The equation according to linear elasticity is also used:

ε = mv ∆σ

in which: mv = (vertical) compression constant = 1 / Eoed

As soil is not a linear elastic material, this formula has limited applicability. The constant mv depends on
the state of stress and can only be assumed constant in a limited stress course. The most used method
for determination of the compressibility in the Netherlands is proposed by Koppejan. The following
equation is based on Koppejan’s original formula from 1948, but includes the degree of consolidation:

 U 1  ∆t    σ 'v ;i + ∆σ 'v 
ε = + log    ⋅ ln  
C' σ 'v ;i
 p C 's  t ref   

in which: ε [-] = relative compression = ∆H / H

H [m] = layer thickness
U [-] = degree of consolidation (see Section 34.1)
C’p [-] = primary compression coefficient
C’s [-] = secondary compression coefficient
t [day] = duration after the application of the additional loading
tref [day] = reference duration (one day)
∆σ'v [kPa] = increase of the vertical effective stress in the weak layer
σ’v;i [kPa] = initial vertical effective pressure

This equation comprises both primary settlement and creep, taking the degree of consolidation into
account.

206
Manual Hydraulic Structures Part III - 'Materials'

Notes
• Be aware that the formula is confusing. It contains both an e-log and a 10-log.
• Usually t ∞ = 10 000 days (= 30 years) is considered the maximum settlement time.
• On completion of a structure, the usual requirement is that the total settlement < 0,30 m, provided
that the settlement is even and resulting level differences are dealt with in an appropriate way (for
instance by applying concrete slabs (rijplaten) at the interface of an abutment and an embankment).
• If both sand layers and clay or peat layers are present, the compression of the sand layers is usually
negligible.
• Take spreading of the load in the soil into account. A spread of the stress under an angle of 45
degrees is customary (for slabs both in lengthways and widthways directions: For a square slab
with length = D, ∆σ’ is reduced to 1/9th at a depth D (see Section 24.1).
• The parameters C’p and C’s follow from oedometer tests on samples obtained by (tube sample)
borings. For a preliminary design the values in Table 31-4 (the Eurocode table with soil properties)
would suffice.
• After a situation of preloading, the soil will behave much stiffer (approximately a factor 3 to 5) as
long as the maximum effective vertical stress that has acted on the soil in the past is not exceeded.
This maximum stress is referred to as the pre-consolidation pressure, σ’max (grensspanning).
Compression will be usually be small, if the vertical effective stress remains below σ’max. This implies
that the settlement of preloaded soil will be smaller than of non-preloaded soil. For detailed designs,
it is better to divide the settlement calculation into a stiff part before the vertical effective soil
pressure reaches σ’max and a weak part thereafter. In that case,

U 1  σ '   U 1   σ' 
ε =  + log(t )  ⋅ ln  max  +  + log(t )  ⋅ ln  ' v 
     
 Cp Cs   σ 'v ; i   C ' p C 's   σ max 
where:
Cp [-] = primary compression coefficient below pre-consolidation pressure (σ'v < σ'max)
Cs [-] = secondary compression coefficient below pre-consolidation pressure (σ'v < σ'max)
C’p [-] = primary compression coefficient above pre-consolidation pressure (σ'v > σ'max)
C’s [-] = secondary compression coefficient above pre-consolidation pressure (σ'v > σ'max)
σ ’max [kPa] = pre-consolidation pressure of the soil

Reference is made to specialised readers on soil mechanics for more details on this calculation
method.

34.3 Soil relaxation

Excavations (e.g. for a construction site) relieve the soil situated at a greater depth. This causes
relaxation (= swell) of the soil. In turn this causes the bed to rise in time.

To determine the total relaxation ∆htotal the following formula is used:

∆htotal Csw σ ' + ∆σ

= · log( i ' )
h (1 + e0 ) σi
in which:
h [m] = initial thickness of the relaxing layer
Csw [-] = swell or relaxation coefficient (zwelcoëfficiënt, or zwellingsmodulus)
(see Table 31-4)
e0 [-] = initial void ratio (poriëngetal)
σi’ [N/m2] = initial effective stress
∆σ [N/m2] = reduction of the effective stress due to the excavation

The void ratio is defined with:

Vpores
e=
Vgrains
in which:
3
Vpores [m ] = volume of the pores, which is the volume of water if the soil is completely saturated
3
Vgrains [m ] = volume of the solid material (sediment grains)

207
Manual Hydraulic Structures Part III - 'Materials'

If the relaxation is (partially) restricted, for instance by the presence of a submerged concrete floor with
tensile piles, this leads to an increase of the effective stress below the floor. For relaxation a secondary
effect (creep) is not taken into account in calculations, so the extent of the effective stress against the
floor depends on the consolidation of the swelling layer and on the extent of the change of the effective
stress due to the excavation.
If the stiffness of the soil is linear when it is relieved of its load, one can pose:

σ 'relaxation = (1 − U ) · ∆σ 'excavate

The calculation of the degree of consolidation U, the moment the floor is constructed is carried out
according to Section 34.1 "Consolidation". For this one can assume that:

cv ,relaxation ≈ 4 ⋅ cv ,

because the stiffness ( 1 mv ) is far greater when being relieved of pressure, which has consequences for
cv according to:

kv
cv =
γ w mv

34.4 Literature
Barron, R.A. (1948) Consolidation of fine grained soils by drain wells. Transactions of American Society
for Civil Engineers, Vol. 113, No. 2346, pp. 718-724.

Knappet, J.A. and R.F. Graig (2012). Craig's soil mechanics (8th editon). Spon Press, London.

Terzaghi, K. (1943) Theoretical soil mechanics. John Wiley & Sons, New York.

Verruijt, A. and S. van Baars (2005). Soil mechanics. VSSD, Delft, 2005.

208
Manual Hydraulic Structures Part III - 'Materials'

35. Concrete
major revision: February 2011
crack width and strut & tie modelling added: February 2018

Concrete is a commonly used and very suitable construction material, particularly for non-moving parts of
hydraulic structures. The design of concrete structures is a profession in itself. It should be realised that
the theory of concrete for hydraulic engineering purposes has other emphases than the theory for the
more common non-residential building (utiliteitsbouw) branch, because of the following reasons:
1. In non-residential building, structural elements can often be schematised as bending beams.
Hydraulic structures are often not slender and often have complex 3-D shapes.
2. In hydraulic structures, the concrete parts below the water surface are under pressure. A pre-tensile
stress is present on all sides. This circumstance differs form non-residential building.
3. The reinforcement steel in hydraulic structures in sea water must be well protected from corrosion.
This is why often pre-stressed reinforcement is used to reduce the crack width to zero.

35.1 Properties of concrete

For design calculations European standards should be used: for concrete NEN-EN 1992-1-1:2005
(Eurocode 2: Design of concrete structures). For the theory about this subject, the course of CT2051 and
CT3051 is recommended. The prescribed characteristics for concrete classes currently available in the
Netherlands are presented in Table 35-1 (EN 206-1 Table 7 and EN 1992-1-1 Table 3.1).
concrete strength
class (old) class
fck fc cub k fcm fctm fctk, 0,05 fctk, 0,95 Ecm
(MPa) (MPa) (MPa) (MPa) (MPa) (MPa) (GPa)
B15 C12/15 12 15 20 1,6 1,1 2,0 27
B25 C20/25 20 25 28 2,2 1,5 2,9 30
B35 C30/37 30 35 38 2,9 2,0 3,8 33
B45 C35/45 35 45 43 3,2 2,2 4,2 34
B55 C45/55 45 55 53 3,8 2,7 4,9 36
B65 C53/65 53 67 63 4,2 3,0 5,5 38
B75 C60/75 60 75 68 4,4 3,1 5,7 39
B85 C70/85 70 85 78 4,6 3,2 6,0 41
Table 35-1 Characteristics of concrete classes

fck [MPa] = characteristic compressive cylinder strength of concrete at 28 days

fc cub k [MPa] = characteristic compressive cube strength of concrete at 28 days
fcm [MPa] = mean value of concrete cylinder compressive strength after 28 days (fcm= fck+8)
fctm [MPa] = mean value of axial tensile strength of concrete:
(2/3)
fctm = 0,30·fck for qualities ≤ C50/60;
fctm = 2,12·ln(1+(fck/10)) for qualities > C50/60
fctk, 0,05 [MPa] = characteristic axial tensile strength of concrete (fctk, 0,05 = 0,7fctm 5% fractile)
fctk, 0,95 [MPa] = characteristic axial tensile strength of concrete (fctk, 0,95 = 1,3fctm 95% fractile)
0,3
Ecm [GPa] = secant modulus of elasticity of concrete (Ecm = 22 (fcm /10) ) (fcm in MPa)

The strength class, or concrete quality, is indicated by the letter C (from Concrete), followed by the
characteristic compressive cylinder strength and the characteristic compressive cube strength at 28 days.
The choice for a certain strength class depends on the design, the way of construction, and the costs.
The lowest quality, C12/15 is not very suitable as a construction material. For in situ cast concrete,
qualities C20/25, C25/30, C30/37 are very common, because they provide sufficient strength and
stiffness for non-residential building and can be handled very well during construction. The strength of
prefabricated concrete elements is usually higher than in situ cast concrete , because manufacturing
circumstances can be better controlled and quick hardening is required for mass production: C35/45 is
the minimum, but more common are C45/55 and C50/60.

The design value for concrete compressive strength can be computed as follows:

α c c ⋅ fc k
fc d = (35.1)
γC

209
Manual Hydraulic Structures Part III - 'Materials'

The design value for concrete tensile strength can be computed as follows:

α ct ⋅ fctk ,0.05
fctd = (35.2)
γC
where:
fcd [MPa] = design value of concrete compressive strength
fctd [MPa] = design value of concrete tensile strength
αcc [-] = coefficient taking account of long term effects on the compressive
strength and of unfavourable effects resulting from the way the
load is applied (αcc = 1,0)
αtc [-] = coefficient taking account of long term effects on the compressive
strength and of unfavourable effects resulting from the way
the load is applied (αtc = 1,0)
fck [MPa] = characteristic compressive cylinder strength of concrete at 28 days
fctk,0.05 [MPa] = characteristic axial tensile strength of concrete (5% fractile)
γc [-] = partial safety factor for concrete

An overview of the partial safety factors for materials for ultimate limit states for concrete ( γ C ) and steel
( γ S ) is given in Table 35-2.

Design situations γC for concrete γS for reinforcement γS for prestressing

steel steel
Persistent & Transient loads 1,5 1,15 1,1
Accidental loads 1,2 1,0 1,0
Table 35-2 Partial safety factors for material

35.2 Properties of reinforcement steel

For reinforcement steel TGB 1990 gives material properties for some steel classes:

Steel type fyk fyd εuk

2 2 2
[N/mm ] [N/mm ] [N/mm ]
Bars FeB 220 HWL 220 190 5,00
FeB 400 HWL, HK 400 350 4,00
FeB 500 HWL, HK 500 435 3,25
FeB 500 HKN 500 435 2,75
Wire fabrics FeB 500 HKN, HWN 500 435 2,75
(wapeningsnetten)
Table 35-3 Characteristics of reinforcement steel classes according to the old TGB 1990 standard
where:
2
fyk [N/mm ] = characteristic yield strength of reinforcement
2
fyd [N/mm ] = design yield strength of reinforcement
εuk [-] = characteristic strain (rek) of reinforcement or pre-stressing steel
at maximum load

The TGB 1990 standard has been replaced by a NEN standard that shows some differences regarding
the characteristic values of reinforcement steel classes. The NEN standard (NEN 6008) applies only in a
limited part of Europe. The most frequently used reinforcement steel class is B500B.

Reinforcement ∅ Re Rm / Re Agt
steel classes [mm] [MPa] [-] [%]
B500A 4-16 500 1,05 (1,03 for ∅ ≤ 5.5mm) 3,0 (2,0 for ∅ ≤ 5,5mm)
B500B 6-50 500 1,08 5,0
B500C 6-50 500 1,15 (1,13 for ∅ ≤ 12mm) 7,5 (7,0 for ∅ ≤ 12mm)
Table 35-4 Characteristics of reinforcement steel classes according to the NEN 6008 standard

210
Manual Hydraulic Structures Part III - 'Materials'

where:
∅ [mm] = nominal diameter
Re [MPa] = characteristic yield strength of reinforcement (fyk)
Rm [MPa] = characteristic tensile strength of reinforcement (ftk)
Rm/Re [-] = minimum ratio tensile strength/yield strength (ftk/fyk)
Agt [-] = minimum percentage total elongation at maximum force
A [-] = indicates a smooth, dented or ribbed profile
B [-] = indicates a dented or ribbed profile
C [-] = indicates a ribbed profile

Commonly used reinforcement bar diameters in Hydraulic Engineering are ∅ 12, 16, 20, 25 and 32 mm.
5 2
The Young’s modulus of reinforcement steel (Es) is 2,0·10 N/mm

35.3 Properties of pre-stressed steel

In principle the function of pre-stressing is to prevent the occurrence of cracks in the concrete structure by
creating compressive stresses in a structural member where one normally would expect tensile stresses.
The elimination of tensile stresses does not only result in the prevention of cracks in the concrete, but
also in a more economical use of materials (slender structures). In Figure 35-1 the principle of pre-
stressing is visualized for a simply supported beam (vrij opgelegde balk) on two supports. The load on the
beam results in compressive stresses above the centroid (neutrale lijn) and tensile stresses below the
centroid as indicated in the left stress diagram. As the result of pre-stressing a normal force is exerted,
resulting in an evenly distributed compressive stress over the whole cross-section of the beam (middle
stress diagram). This compressive stress eliminates the tensile stress at the underside of the beam and
reinforces the compressive stress at the topside, resulting in compressive stresses over the whole cross-
section (right stress diagram).

Figure 35-1 the principle of pre-stressing.

The following methods for pre-stressing concrete are used:

1) Pre-tensioning (voorspanning met aanhechting, VMA): This principle is mainly used in the
prefabrication of concrete members. In the factory the tendons (voorspanwapening) are pre-stretched
before the concrete is poured. After the concrete has hardened enough the tendons are released. The
force present in the tendons is absorbed via adhesion (friction) between the steel and the concrete.
2) Post-tensioning (voorspanning zonder aanhechting, VZA): Here the tendons are situated in a
protective tube. After pouring the concrete they are slightly tensioned in order to prevent attachment
with the cement water. After a certain period, when the concrete has reached a strength determined
by the structural engineer, the tendons are pre-stressed to approximately 20% of their capacity. This is
called pre-stressing the self-weight (eigen gewicht aanspannen). When the concrete has reached its
ultimate strength the tendons are pre-stressed to 100% of their capacity.

There are three types of pre-stressing steels, namely: wire (voorspandraad), strands (voorspanstreng)
and bars (voorspanstaven). The properties of these three types are described in the following standards,
NEN-EN 10138-2 (draft) “wire”, NEN-EN 10138-3 (draft) “strand” and NEN-EN 10138-4 (draft) “bars”.
Furthermore the NEN- EN 10138-1 (draft) “general requirements” and NEN-EN 1992-1 are applicable. In
Table 35-5 the characteristic values for certain diameters of all three types of pre-stressing steel are
presented, for information regarding other available diameters the reader is referred to the standards
mentioned above.

211
Manual Hydraulic Structures Part III - 'Materials'

Steel type d Sn fpk fp0.1k εuk

2 2 2
[mm] [mm ] [N/mm ] [N/mm ] [-]
Wire
Y1860C 4,0 12,57 1860 1599 0,035
Y1770C 6,0 28,27 1770 1521 0,035
Y1670C 8,0 38,48 1670 1437 0,035
Y1570C 10,0 78,54 1570 1299 0,035
Strand
Y1860S3 class A 6,5 23,40 1860 1598 0,035
Y1770S7 class A 16,0 150,00 1770 1587 0,035
Y1960S3 class B 6,5 21,10 1960 1687 0,035
Y1960S7 class B 9,0 50,00 1960 1680 0,035
Bar
Y1030H 26,0 531 1030 834 0,035
Y1030H 40,0 1257 1030 835 0,035
Y1230H 26,0 531 1230 1079 0,035
Y1230H 40,0 1257 1230 1080 0,035
Table 35-5 Characteristic values for pre-stressing steels.
where:
d : nominal diameter
Sn : nominal cross-sectional area
fpk : characteristic value for the tensile strength of pre-stressing steel
fp0.1pk
: characteristic 0,1% yield boundary for pre-stressing steel
εuk : characteristic strain (rek) of reinforcement or pre-stressing steel at maximum load
fp 0.1k
The design value for the tensile strength is equal to: fpd =
γs
The design value for the characteristic yield boundary can be computed as follows: ε ud = 0,9 ⋅ ε uk .
5 2 5 2
The Young’s modulus for pre-stressed steel (Ep) is 2,05·10 N/mm for wire and bars and 1,95·10 N/mm
for strands.

35.4 Reinforced and pre-stressed concrete

To design reinforced or pre-stressed concrete structures the following limit states have to be considered:
1) Ultimate limit states, leading to failure of the structure;
2) Serviceability limit states; leading to restriction of use of the structure.

ultimate limit state:

• fracture due to bending and / or normal force
• fracture due to shear force
• fracture due to punching
• fracture due to torsion

serviceability limit state:

• unacceptable deformation
• unacceptable cracking (scheuren)

For a more elaborate consideration of the limit states, reference is made to the Eurocode 0, Section 3.
Bending and shear force are discussed briefly here because they are of importance in a preliminary
design.

Bending and/or normal force

The limit state involving bending and normal force is:

MEd = MRd and NEd = NRd

in which: MEd [Nm] = design value of the maximum occurring bending moment
MRd [Nm] = maximum allowable bending moment
NEd [Nm] = design value of the normal force
NRd [Nm] = maximum allowable normal force

212
Manual Hydraulic Structures Part III - 'Materials'

Figure 35-2 Forces and strains in concrete

The maximum allowable moment and normal force are:

MRd = (N p + NEd )( zb − y ) + ∑ Ns (d s − y ) + ∑ ∆Np (dp − y ) (35.3)

NRd = Nc + Ns 2 − N p − ∆N p − Ns (35.4)

where: Np [N] = design value for the effective normal compression force as a result of the
pre-stressing force
Mp [Nm] = design value for the effective moment force as a result of the pre-stressing
force
NEd [N] = design value of the normal force (excluding pre-stressing force); if the
normal force is a tensile force replace +NEd with -NEd
Nc [N] = design value of the compression resultant = 0,75 ⋅ xu ⋅ fcd
2
fcd [N/m ] = design value of concrete compressive strength
Ns [N] = tensile force in the reinforcement steel
Ns,2 [N] = compressive force in the reinforcement steel
∆Np [N] = increase of the force in the pre-stressing reinforcement relative to the initial
pre-stressing force (∆Np = Ap·∆σpu)
2
Ap [m ] = cross-sectional area of the pre-stressed element
∆σpu
2
[N/m ] = increase of the stress in the pre-stressing reinforcement relative to the
initial pre-stressing stress
y [m] = distance between the compression stress resultant and the edge with
the highest compression = 7/18 xu (for ≤ C 50 / 60 )
xu [m] = height of the concrete compression zone
ds [m] = the distance between the tensile reinforcement and the edge with highest
compression
ds2 [m] = the distance between the reinforcement in the compression zone and the
edge with most compression
dp [m] = the distance between the pre-stressing steel and the edge with most
compression
zb [m] = the distance between the elastic line of gravity and the edge with most
compression
h [m] = total height of the structure
εcu3 [m] = ultimate compressive strain in the concrete

When determining xu one must take into account that: ε cu 3 = 0,0035 .

Furthermore, there are requirements for the maximum value of xu if the normal force is small
( NEd < 0,1⋅ fcd ⋅ Ac ) due to the rotation capacity, for this the reader is referred to TGB 1990 (NEN 6720
article 8.1.3).

To calculate the required reinforcement, the requirement should be satisfied that the reinforcement steel
must yield before the concrete will fail and the minimum of the reinforcement percentage must be large
enough to be sure there will be no brittle failure when cracking of the concrete occurs (brosse breuk). If
the structure is mainly loaded by a moment force, the required reinforcement steel can easily be
calculated with help of Table 35-6. Note that Table 35-6, Table 35-7, Table 35-8 and the flowchart below
only apply to reinforced concrete and not for pre-stressed concrete. The flowchart is used to compute the
213
Manual Hydraulic Structures Part III - 'Materials'

reinforcement percentage needed in a structural member when the bending moment for the ultimate limit
state is known.

Figure 35-3 Flowchart for the preliminary design of reinforcement using the GTB-tables.

214
Manual Hydraulic Structures Part III - 'Materials'

Table 35-6 Reinforcement percentages for rectangular cross-sections, reinforced with B500B, loaded by bending without normal
force, With Mu in kNm; b and d in m1 and fcb in N/mm2

where:
Mu [kNm] = ultimate absorbable bending moment (breukmoment)
 fyd 
k [-] = ratio between the strength of concrete and steel  k = 
 fcd 
 fyk 
[N/mm ] = design yield strength of reinforcement  fyd =
2
fyd 
 γs 
 f 
[N/mm ] = design value of concrete compressive strength  fcd = ck 
2
fcd
 γc 
ρ [%] = reinforcement percentage
d [m] = total thickness of the concrete member (so, not only the effective width)
b [m] = width of the concrete member, perpendicular to the considered cross-
section (so, if forces and moments are per running metre,
b = 1 m).

215
Manual Hydraulic Structures Part III - 'Materials'

ψ [%] = mechanical reinforcement percentage: ψ = k·ρ

x z
kx = u kz = u
d d
ρ ⋅k
xu = height of the of the compressive zone (hoogte drukzone): xu = d ⋅
0,75
zu = arm of internal leverage (inwendige hefboomsarm):
zu = d ⋅ (1 − 0,52 ⋅ ρ ⋅ k ) = d − β ⋅ xu , where β = 0,75 ⋅ 0,52 = 0,39
d[mm] = effective height of the cross-section (nuttige hoogte):
d = h - (c + ½ ∅ )
h [mm] = height of the cross-section
c [mm] = concrete cover
∅ [mm] = bar diameter (kenmiddellijn)
2
As [mm ] = total cross-sectional area of the reinforcement

C20/25 C28/35 C35/45 C45/55

ρmin 0,15 0,18 0,21 0,24
Table 35-7 Minimum reinforcement percentage (ρmin) for B500B.

C20/25 C28/35 C35/45 C45/55

ρmax 1,38 1,94 2,49 3,05
Table 35-8 Maximum reinforcement percentage (ρmax) for B500B.

To check an existing concrete structural member, the maximum allowable bending moment can be
computed using the following equation (see the book Constructieleer Gewapend Beton of course
CTB2220):

Mu = As ⋅ fyd ⋅ d ⋅ (1 − 0,52 ⋅ ρ ⋅ k ) and Med ≤ Mu (35.5)

where:
Med [Nm] = design value for the bending moment in the ultimate limit state
Mu [Nm] = ultimate absorbable bending moment
2
As [m ] = total cross-sectional area of reinforcement
 fyd 
k [-] = ratio between the strength of concrete and steel  k = 
 fcd 
2
fyd [N/m ] = design yield strength of reinforcement
2
fcd [N/m ] = design value of concrete compressive strength
A
ρ [-] = reinforcement ratio ( = s )
b ⋅d
b [m] = width of the concrete structure

Automatically the equation can also be used to calculate the necessary reinforcement, when the load is
known.

216
Manual Hydraulic Structures Part III - 'Materials'

Shear force of concrete elements without shear reinforcement

The design value for the shear resistance VRd,c without shear reinforcement is given by

VRd ,c = CRd ,c ⋅ k ⋅ (100 ⋅ ρ1 ⋅ fck ) + k1 ⋅ σ cp  ⋅ bw ⋅ d [N]

1/ 3
(35.6)
 

with a minimum of

VRd ,c = (v min + k1 ⋅ σ cp ) ⋅ bw ⋅ d [N], (35.7)

2
where fck [N/mm ] = characteristic compressive cylinder strength of concrete at 28 days in MPa
200
k [-] = 1+ ≤ 2,0 with d in mm
d
Asl
ρ1 [-] = reinforcement ratio for longitudinal reinforcement = ≤ 0,02 ;
bw ⋅ d
2
Asl [mm ] = the area of the tensile reinforcement, which extends more than (lbd + d)
beyond the considered section
bw [mm] = the smallest width of the cross-section in the tensile area
σ cp 2
[N/mm ] = compressive stress in the concrete from axial load or prestressing:
σ cp = NEd / AC < 0,2 ⋅ fcd
NEd [N] = the axial force in the cross-section due to loading or prestressing
(NEd > 0 for compression)
2 2
AC [mm ] = the area of the concrete cross-section [mm ]
k1 [-] = a coefficient, in the Netherlands: 0,15
CRD,c [-] = a coefficient, in the Netherlands: 0,18 / γ c = 0,18 / 1,5 = 0,12
vmin [-] = 0,035 ⋅ k 3 2 ⋅ fck 1 2

Figure 35-4 Reinforced concrete structural member without shear reinforcement.

Shear force of concrete elements with shear reinforcement

The design of members with shear reinforcement is based on a truss model (vakwerkmodel), see Figure
35-5.

Figure 35-5 Inclined shear reinforcement

217
Manual Hydraulic Structures Part III - 'Materials'

The following notations are shown in Figure 35-5:

α : angle between shear reinforcement and the beam axis perpendicular to the shear force
(measured positive as shown in the figure)
θ : angle between the concrete compression strut and the beam axis perpendicular to the shear force
Ftd : design value of the tensile force in the longitudinal reinforcement
Fcd : design value of the concrete compression force in the direction of the longitudinal member axis.
bw : the smallest width of the cross-section in the tensile area
z : arm of internal leverage, for a member with constant depth, corresponding to the bending moment in
the element under consideration. In the shear analysis of reinforced concrete without axial force, the
approximate value z = 0,9·d may normally be used.

The angle θ should be limited. The recommended upper and lower limits are: 21,8° ≤ θ ≤ 45° .

Vertical shear reinforcement

For members with vertical shear reinforcement, the shear resistance, VRd is the smaller value of :

Asw
VRd ,s = ⋅ z ⋅ fywd ⋅ cot θ (stirrups governing) (35.8)
s
where:
Asw = the cross-sectional area of the shear reinforcement (be aware that the reinforcement
crosses the cross-sectional area of the concrete element twice).
s = the spacing of the stirrups
fywd = the design yield strength of the shear reinforcement

and:
α cw bw z ⋅ v1 ⋅ fcd
VRd ,max = (concrete compressive struts governing) (35.9)
cot θ + tanθ

where:
v1 = strength reduction factor for concrete cracked in shear. Recommended is that v1 = v
 f 
and v = 0,6  1 − ck 
 250 
α cw = coefficient taking account of the state of the stress in the compression chord. The
recommended value of α cw is as follows:
-1 for non pre-stressed structures
- (1 + σ cp / fcd ) for 0 < σ cp ≤ 0,25 ⋅ fcd
- 1,25 for 0,25 ⋅ fcd < σ cp ≤ 0,5 ⋅ fcd
- 2,5 ⋅ (1 − σ cp / fcd ) for 0,5 ⋅ fcd < σ cp ≤ 1,0 ⋅ fcd
σ cp = the mean compressive stress, measured positive, in the concrete due to the design axial
force. This should be obtained by averaging it over the concrete section taking account of
the reinforcement. The value of σ cp need not be calculated at a distance less than
0.5 ⋅ d ⋅ cot θ from the edge of the support.

The maximum effective cross-sectional area of the shear reinforcement, Asw,max for cot θ = 1 is given by:

Asw ,max ⋅ fywd 1

≤ ⋅ α cw ⋅ v1 ⋅ fcd (35.10)
bw ⋅ s 2

218
Manual Hydraulic Structures Part III - 'Materials'

Inclined shear reinforcement

For members with inclined shear reinforcement (schuine dwarskrachtwapening), the shear resistance is
the smaller value of:
A
VRd ,s = sw ⋅ z ⋅ fywd ⋅ ( cot θ + cot α ) ⋅ sin α (stirrups governing) (35.11)
s
and
α ⋅ b ⋅ z ⋅ v1 ⋅ fcd ⋅ ( cot θ + cot α )
VRd ,max = cw w (concrete compressive struts governing) (35.12)
1 + cot 2 θ

The maximum effective shear reinforcement, Asw,max for cot θ = 1 follows from:

Asw ,max ⋅ fywd 1

⋅ α cw ⋅ v1 ⋅ fcd
≤ 2
(35.13)
bw ⋅ s sin α

Note
In the walls of many hydraulic structures, there are large areas in which the shear force has reached its maximum
while the bending moment is zero. In this case pure tension is found in the concrete wall, which needs special
attention.

35.5 Stiffness of the concrete structure

For statically indeterminate structures, the stiffness (EI) of the elements used in the calculations has
considerable influence, not only on the resulting deformation and displacements, but on the flow of forces
through the structure (global force effect) and the resulting internal forces in each individual member
(local force effect) as well. Figure 35-6 illustrates this effect for a U-shaped cross section on a pile
foundation. The correct stiffness has to be used in hand or computer calculations to find the governing
(internal) load distributions M, N and V.

Figure 35-6 Influence of foundation stiffness on force distribution in the structure

Unfortunately, the stiffness of a reinforced concrete section or element changes depending on crack
development. There is a significant difference in bending stiffness between the non-cracked and the
cracked concrete cross-section. After occurrence of the first cracks, further loading will go hand in hand
with a decreasing stiffness of the concrete. This is easily demonstrated by a M-κ diagram, here κ is
curvature (kromming), see Figure 35-7.

219
Manual Hydraulic Structures Part III - 'Materials'

Figure 35-7 M-κ-diagram

The bending stiffness of a concrete section, having to resist a certain M, is equal to the tangent of the line
in the M-κ-diagram:

Mx Mx
EI x = tan α x = and κx = (35.14)
κx EI x
So, for a certain bending moment the intersection with the line in the M-κ-diagram has to be determined
first to find the value of the curvature on the x-axis. Finally the tangent of the line connecting the origin
with the intersection point can be found, using the equation above and hence the bending stiffness
corresponding to that moment.

To construct a M-κ-diagram, all the M-κ combinations have to be computed. The curvature κ can be
determined as follows:

ε b' + ε s
κ= (35.15)
d

To find the correct stiffness of the whole structure the M-κ-diagram has to be constructed for every
different concrete section, for each type of concrete and reinforcement percentage (ρ). This is a lot of
work, often too much work for the level of precision required. In the following subsections first an
approximation of concrete stiffness will be presented, then development of the M-κ-diagram will be further
explained for detailed calculations.

First design calculations with concrete EI estimate

For uncracked cross-sections the bending stiffness of concrete EI0 can be estimated/computed as
follows:

EI0 = Eb' ⋅ I (35.16)

where:
Eb' = 22 250 + 250 ⋅ fck for 15 ≤ fck ≤ 65 (NEN6720)
E = 35 900 + 40 ⋅ fck
'
b for 65 ≤ fck ≤ 105 (CUR 97)
1
I = ⋅ b ⋅ h3 for rectangular cross-sections
12
220
Manual Hydraulic Structures Part III - 'Materials'

For cracked cross-sections the bending stiffness EIg can be computed as follows:
EIg = 0,5 ⋅ Es ⋅ As ⋅ h 2 (35.17)
where:
2
Es [N/m ] = Youngs' modulus of steel
2
As [m ] = area of the reinforcement steel present in the cross-sectional area of the beam
2
fck [N/m ] = characteristic compressive strength
b [m] = width of the cross-section
h [m] = height of the cross-section
(source: 'construeren in gewapend beton' - part 2, Kamerling 1978)

More detailed calculation of concrete stiffness with M-Κ diagram

In this subsection the critical points of pure bending, i.e. bending moment M without normal force N, will
be explained using stress-strain diagrams (spanning-rek diagrammen) in the end leading to the M-κ-
diagram, see Figure 35-7.

Non-cracked beam (ongescheurde balk)

At the instant that the concrete tensile strength fctd is reached, the deformation diagram and stress
diagram look like depicted in Figure 35-8. The bending moment equals the moment of rupture Mr
(scheurmoment) and the concrete is just not cracked. In this stage the concrete’s compressive strength is
still very small because the mean value of the axial tensile strength of concrete fctm is much smaller than
the design value of the concrete compressive strength fcb, so that ε b' << 1,75 0 00 .

Cracked beam
When the load only increases a little the tensile zone in the concrete will crack and the tensile forces will
be concentrated in the existing reinforcement. The centroid (neutrale lijn) displaces in upward direction.
The load can be increased further until the reinforcement reaches its yield stress fyd. The corresponding
deformation and stress diagrams are depicted in Figure 35-9. The concrete is cracked so it does not have
a tensile strength any longer. The deformation of the concrete at the compression side of the beam ε b' is ( )
still smaller than 1,75‰. The corresponding bending moment is the yield moment (vloeimoment). At this
point the deformation of the steel changes from elastic to plastic.

Figure 35-8 Deformation and stress diagram for a non-cracked beam

Figure 35-9 Deformation and stress diagram for a cracked beam

221
Manual Hydraulic Structures Part III - 'Materials'

Compression strain in concrete (betonstuik)

When the load on the beam is increased further, at a certain moment the deformation of the concrete at
the compression side of the beam will reach the value of 1,75‰ in the extreme pressure fibre (uiterste
drukvezel). At the moment the compression strain of 1,75‰ is reached and the corresponding bending
moment is equal to the plastic moment Mb,pl. The corresponding deformation and stress diagrams are
depicted in Figure 35-10.

Figure 35-10 Deformation and stress diagram when the compression strain in the concrete has reached a value of 1,75‰.

When the load on the beam increases even further, the compression strain in the concrete will reach
eventually a value of 3,50‰ in the extreme pressure fibre. If the beam reaches its point of collapse, the
corresponding bending moment is the moment of fracture Mu (breukmoment). The corresponding
deformation and stress diagrams are shown in Figure 35-11.

Figure 35-11 Deformation and stress diagram when the compression strain in the concrete has reached a value of 3,50‰.

35.6 Concrete cover

The concrete cover on the outer reinforcement bar of a structure protects the reinforcement against
external influences, such as rain water, soil, corrosive liquids or fumes or the like, which can lead to
corrosion of the reinforcement. Reinforcement bars could oxidise if the concrete cover is too thin or
insufficiently dense. This will lead to a reduction of the bar diameter and hence to a decrease of the force
that the reinforcement can resist. Since rust has a larger volume than the original steel, there is a
probability that the concrete cover is pushed off the reinforcement. This will lead to further corrosion and a
further decrease of the strength. It is obvious that the reinforcement in an aggressive environment
requires a thicker concrete cover than in a dry environment. The thickness of the concrete cover therefore
depends on the environment in which the concrete structure is located. The environment characteristics
are expressed via an exposure classification, see Table 35-9.

Besides the type of exposure, the required concrete cover depends on the design life time of the
structure, concrete quality, the geometry of the structural member and the quality control of concrete
manufacturing, which is indicated by the 'structural class'. The standard 'structural class is S4, but may be
adjusted as indicated in Table 35-10.

222
Manual Hydraulic Structures Part III - 'Materials'

Exposure classification
Class Corrosion induced by Class measure of humidity
X0 no risk very dry
XC carbonation XC1 dry or persistently wet
XC2 wet, seldom dry
XC3 moderate humidity
XC4 alternating wet and dry
XD chlorides (excl. seawater) XD1 moderate humidity
XD2 wet, seldom dry
XD3 alternating wet and dry
XS seawater XS1 exposed to salt in the air, no direct contact with seawater
XS2 persistently submerged
XS3 tidal-, splash- and spray-zone
XF freeze/thaw attack XF1 not fully saturated with water, without de-icing salt
XF2 not fully saturated with water, with de-icing salt
XF3 fully saturated with water, without de-icing salt
XF4 fully saturated with water, with de-icing salt
XA chemical attack XA1 weakly aggressive chemical environment
XA2 moderately aggressive chemical environment
XA3 highly aggressive chemical environment
Notation: XC1 X stands for exposure; the letter indicates process that causes the corrosion and the number indicates the humidity

Table 35-9 Exposure classification of the environment in which the structure is situated (Eurocode 2)

Table 35-10 Modification of the standard structural class S4 (design life time of 50 years), depending on the design life time,
strength, geometry and quality control of concrete manufacturing (Eurocode 2, Table 4.4N) Note: values in national codes
can deviate form this table.

According to Eurocode 2, the nominal concrete cover cnom is determined by the minimum cover required
for durability or for bond (aanhechting), and a construction tolerance:

cnom = cmin + ∆cdev (35.18)

The minimum cover cmin is the highest value of:

• cmin,dur , the minimum cover because of durability, see Table 35-11.
• cmin,b , the minimum cover because of bond:
cmin,b > Øbar , or:
cmin,b > Øbar + 5 mm if the largest grain size is 32 mm
• 10 mm

The margin for construction tolerances ∆cdev in the Netherlands usually is 5 mm.
223
Manual Hydraulic Structures Part III - 'Materials'

Table 35-11 Minimum concrete cover on reinforcement steel, related to durability (Eurocode 2, Table 4.4N)
Note: values in national codes can deviate form this table.
For pre-stressed steel, higher values of the minimum required cover apply, see Table 4.5N in Eurocode 2.

For finished concrete surfaces (nabewerkte betonoppervlakken), the minimum cover cmin should be
increased with 5 mm. For concrete floors on uneven surfaces (like work floors), cnom ≥ cmin,dur + 10 mm,
but for concrete floors directly cast on the ground surface, cnom ≥ cmin,dur + 50 mm.

For hydraulic structures, a nominal concrete cover of 50 mm is not unusual.

35.7 Crack width

Theory
A characteristic of concrete elements that are exposed to tensile and bending stresses is the presence of
cracks (scheuren). Cracks will develop if the tensile strength of concrete is exceeded.

Figure 35-12 Cracks in a reinforced concrete beam subjected to tensile forces

Cracks are necessary for the well-functioning of reinforced concrete, but they should not become too
large, because that would affect the concrete cover. The concrete cover protects the reinforcement steel
against aggressive substances, such as groundwater and salts, but if this protection is not sufficient,
corrosion of the reinforcement steel will be the result. The volume of corroded steel is larger than of the
original steel and can push off the concrete cover, after which the steel is directly exposed to the
aggressive substances. Furthermore, corroded steel is less capable of resisting tensile forces, because of
a decreased effective cross-sectional area. Therefore, a minimum concrete cover and maximum crack
width are required.
For the calculation of the crack width of reinforced or pre-stressed concrete elements due to short-term
loading, two stages can be distinguished:
1. The crack formation stage: The number of cracks increases while the strain increases. The width of
the newly formed cracks increases, until a maximum is reached just before a new crack is formed. If
the strain increases further, the tensile force N does not exceed the cracking force Ncr. The maximum
crack width can be calculated with:

1 fctm d
w max = σ sr (35.19)
4 τ bm ρ Es
where:
wmax [mm] = maximum crack width
2
fctm [N/mm ] = mean tensile strength of concrete, according to Table 35-1
d [mm] = diameter of the reinforcement bars
τbm [N/mm2] = bond strength (hechtsterkte) ≈ 2,0 fctm
ρ [-] = reinforcement ratio = Asteel / Aconcrete
2
Es [N/mm ] = Youngs' modulus of steel
2
σsr [N/mm ] = maximum tensile stress in the steel bar (in the crack)
224
Manual Hydraulic Structures Part III - 'Materials'

2. The stabilised cracking stage, where a crack pattern has fully developed. New cracks will not develop
when the strain increases, but the width of existing cracks will increase. The maximum crack width
depends on the steel stress in the crack and can be calculated with:

1 fctm d
w max = (σ s − 0,5 σ sr ) (35.20)
2 τ bm ρ Es

where:
2
σs [N/mm ] = tensile stress in the steel bar
other symbols: see above.

Causes of long-term crack development are:

1. Shrinkage of the concrete, which is usual during the hardening process of concrete. During the crack
development stage, concrete shrinkage does not lead to larger crack widths, because of the
development of additional cracks (the external force cannot exceed the cracking load). However, in he
stabilised crack stage, no new cracks are formed and the width of existing cracks will increase. The
crack width caused by load and shrinkage can be calculated with:

1 fctm d
w max = (σ s − 0,5 σ sr + ε cs Es ) (35.21)
2 τ bm ρ Es

where:
2
σs [N/mm ] = steel stress in a crack under external tensile load
εsc [-] = shrinkage of the concrete (no steel-concrete bond assumed)
other symbols: see above.

2. Long-term constant or varying loads, that reduce the bond stress. During the crack development
stage, the reduced bond strength results in an increase of the transfer length of 25% and thereby 25%
increase of the crack width. Equation (35.19) can be used to calculate this crack width, but now τbm =
1,6 fctm.
For the stabilised crack stage, the bond stress decreases and therefore also the influence of the
concrete in between the cracks. It can be assumed that this reduction is about 40%. This can be taken
into account by replacing the coefficient 0,5 in Equation (35.21) by 0,3.

In the theory above, it was assumed that the tensile stresses are uniformly distributed over the entire
concrete cross-section. In many cases, however, only a part of the concrete cross-section effectively
transfers tensile forces to the reinforcement bars (Figure 35-13).

Figure 35-13 Wide cracks in a reinforced concrete beam under extreme loading

The effective tensile areas are indicated in Figure 35-14 for three types of cross-sections. The effective
height is 2,5(h-d) for beams, (h-x)/3 for slabs and h/2 for elements subjected to tensile forces.

225
Manual Hydraulic Structures Part III - 'Materials'

Figure 35-14 Effective tensile area

If one dimension of the cross-section of the concrete element is much larger than the other, the
distribution of the tensile stresses in the concrete is not uniform, see Figure 35-15a. As a result, the crack
does not proceed over the full width of the element. Such a cross-section can only crack, if the force
introduced by the reinforcement is uniformly spread over the width of the element (Figure 35-15b). Only a
few cracks reach the outer surface of the concrete element. If the reinforcement would be concentrated
near the surface of an element, many cracks with small widths will occur at the surface, but wide cracks
occur in the middle (Figure 35-15c).

A similar phenomenon occurs in elements in bending (Figure 35-15d). The main reinforcement limits the
crack widths over an area close to the reinforcement and the relatively small cracks at the bottom join
higher in the element and develop into wide cracks, if there is no additional reinforcement over the height
of the element.

226
Manual Hydraulic Structures Part III - 'Materials'

Figure 35-15 Crack behaviour of elements with a high width / transfer length ratio or a concentrated reinforcement

The strain in beams decreases from the bottom, where the highest tensile stresses occur, in upward
direction. Therefore, also the crack width decreases in that direction and it is not needed to extend the
reinforcement to the neutral axis (Figure 35-16).

Figure 35-16 Stressless area in a concrete beam above the reinforced area

The crack width follows from the crack spacing that occurs even without reinforcing steel present. It can be
assumed that the crack spacing is about equal to the height of the unreinforced area (t in Figure 35-16).
The maximum crack width occurs about halfway this area:
227
Manual Hydraulic Structures Part III - 'Materials'

w max = t ε c2 (35.22)
εs 1 t
where the fictitious concrete strain halfway down the unreinforced area is: ε c2 = 2
(d - x )
t2 1
The expression for wmax then becomes: w max = εs2
(35.23)
(d - x )
Figure 35-17 shows graph that can be used to determine the area where reinforcement should be applied.
Additional crack distributing reinforcement is only required for relatively high beams. For smaller beams it is
sufficient to verify only the crack width in the main tensile tie.

Figure 35-17 Part of the web where crack width limiting reinforcement should be provided to avoid the occurrence of large cracks
(Jungwirth, 1985)

It appears that in high beams, crack width limiting reinforcement is required over a certain height. This does
not automatically imply that in the remaining parts no reinforcement is needed. In general, a practical
minimum reinforcement should always be applied. Summarising one may conclude that the following
reinforcement is to be provided:
• Horizontal reinforcement Ø10 – 100 mm as crack width limiting reinforcement
• Other parts: horizontal reinforcement Ø10 – 300 mm.
• Vertical reinforcement, for instance also Ø10 – 300 mm.

Crack width design (Eurocode 2)

For the design of new structures, it is not necessary to calculate the crack widths. The procedure is to
design the reinforcement of the structural member in such a way, that the crack width will not exceed the
maximum allowable value. So, for designs, the maximum crack width wmax is given.

Crack widths are usually restricted to values between 0,1 and 0,4 mm to avoid unacceptable damage to
the concrete cover. The maximum value depends on the type of steel (reinforced or pre-stressed), the
durability of the structure and aesthetics. Table 35-12 specifies the maximum values for the maximum
crack width, as recommended by the general Eurocode 2. Maximum crack widths can deviate from these
recommended values per country!

228
Manual Hydraulic Structures Part III - 'Materials'

Table 35-12 Recommended values for the maximum crack width (Eurocode 2, 2011)
Notice that the values vary per country, see the national annexes to Eurocode 2!

In general, the following steps can be followed:

1. calculate the maximum moment Mmax (with a moment diagram (momentenlijn))
2. choose concrete class, steel quality, exposure class (milieuklasse)
3. calculate the required reinforcement area per cross-section, As,req
4. choose reinforcement bars / lay-out, and therewith the provide reinforcement area As,prov
5. calculate the stress in the reinforcement bars σs according to:

Mfreq A s ,req
σs = ⋅ ⋅ fyd
MEd A s, prov

where: Mfreq [kNm] = bending moment in SLS

MEd [kNm] = bending moment in ULS
2
As,req [mm ] = required reinforcement area
2
As,prov [mm ] = provided reinforcement area
2
fyd [N/mm ] = design value of the yield strength of the reinforcement steel

6. find the maximum allowed crack width in Table 35-12

7. find the maximum bar diameter in Table 35-13
8. find the maximum distance between the bars in Table 35-13

steel stress max. bar diameter (mm) max. distance between bars
2 wmax = 0,4 mm wmax = 0,3 mm wmax = 0,2 mm wmax = 0,4 mm wmax = 0,3 mm wmax = 0,2 mm
(N/mm )
160 40 32 25 300 300 200
200 32 25 16 300 250 150
240 20 16 12 250 200 100
280 16 12 8 200 150 50
320 12 10 6 150 100 -
360 10 8 5 100 50 -
Table 35-13 Maximum bar diameter and maximum distance betwee reinforcement bars,

The value for fyd in the equation for the steel stress σs is the design value of the yield strength of the
reinforcement steel, which is the characteristic yield strength fyk (Table 35-3), divided by a material factor
(1,35 for example). The bending moment in ULS conditions can be about 1,5 (or maybe even 2) times
more than during SLS conditions. The reinforcement diameters obtained with help of Table 35-12 and
Table 35-13 can still be smaller than the diameters commonly used in hydraulic engineering (32 or 40
mm). These larger diameters are usually obtained by a more detailed crack width control calculation
(rather than using the table), which is not required for this course.

229
Manual Hydraulic Structures Part III - 'Materials'

Crack width calculation (Eurocode 2)

For the structural verification of existing structures, it should be checked whether the actual calculated
crack widths exceed the maximum allowed values (Table 35-12).

A general equation for calculating the crack width, taking the effects described above into account, is:

1 fctm d
w max = (σ s − α σ sr + β ε cs Es ) (35.24)
2 τ bm ρ Es
where:
wmax [mm] = maximum crack width
2
fctm [N/mm ] = mean tensile strength of concrete
d [mm] = diameter of the reinforcement bars
τbm [N/mm2] = bond strength (hechtsterkte), see Table 35-14
ρ [-] = reinforcement ratio = Asteel / Aconcrete
2
Es [N/mm ] = Youngs' modulus of steel
2
σs [N/mm ] = tensile stress in the steel bar
2
σsr [N/mm ] = maximum tensile stress in the steel bar (in the crack):
f
σ sr = ctm (1 + α e ρ ) , where αe = Es /Ec (35.25)
ρ
εsc [-] = shrinkage of the concrete (no steel-concrete bond assumed)
α, β [-] = coefficients, see Table 35-14

Table 35-14 values for the coefficients α, β and τbm. The values in brackets are recalibrated values, according to Eurocode 2.

In the Eurocode2, Equation (35.24) appears in a different form:

fct ,eff
σ s − kt (1 + α e ρ p,eff + ε cs Es )
ρ p,eff
w max = sr ,max (35.26)
Es
where:
sr,max [mm] = maximum crack spacing:
Ø
sr,max = k3 c + k1 k2 k4
ρ p,eff
c [mm] = concrete cover to the longitudinal reinforcement
k1 [-] = bond stress coefficient:
k1 = 0,8 for high bond (ribbed) bars
k1 = 1,6 for bars with an effectively plain surface
k2 [-] = coefficient for the distribution of the strain over the height of the concrete area:
k2 = 0,5 for bending; k2 = 1,0 for pure tension
k3 [-] = coefficient; usually recommended: k3 = 3,4
k4 [-] = coefficient; usually recommended: k4 = 0,425
kt [-] = load duration coefficient: kt = 0,6 (short-term loading) or 0,4 (long-term loading)
2
fct,eff [N/mm ] = mean value of the concrete tensile strength of the concrete at the time
when the crack may first be expected to occur:
fct,eff = fctm or lower
(fctm(t), if cracking is expected earlier than at a concrete age of 28 days)
αe [-] = ratio between Youngs' moduli of steel and concrete: Es /Ec
ρpf [-] = effective reinforcement ratio of the tensile element:

230
Manual Hydraulic Structures Part III - 'Materials'

As
ρ pf =
Ac,eff
2
As [mm ] = total area of steel in a concrete element cross-section
2
Ac,eff [mm ] = effective tensile area of a concrete element:
A c,eff = h eff b
b [mm] = width of the concrete element
heff [mm] = effective height of the virtual tensile element:
the smallest value of 2,5(h-d), (h-x)/3 and h/2, according to
Figure 35-14.
εsc [-] = shrinkage of the concrete (no steel-concrete bond assumed)
2
Es [N/mm ] = Youngs' modulus of steel

The total shrinkage of the concrete εsc consists of two components: dehydration shrinkage and
autogenous shrinkage. The dehydration shrinkage is a function of the migration of water through the
hardened concrete (related to the humidity of the air) and therefore develops gradually. Autogenous
shrinkage takes place during the hardening of the concrete, mainly during the first days after casting and
is a linear function of the concrete quality. After a long duration, the effect of autogenous shrinkage is
much smaller than the effect of dehydration shrinkage. For not too high concrete qualities and outdoor
-3
concrete elements (80% humidity), the total shrinkage εsc is in the order of magnitude of 0,20·10 , which
could be used for preliminary designs. For more detailed designs, the equations and tables of Eurocode
2, article 3.1.4 can be used.

35.8 Strut & tie modelling

Theory
Reinforced concrete members are usually designed to resist shear and bending stresses, based on the
assumption that strains vary linearly over a section. The mechanical behaviour of a beam is commonly
determined by the assumption that plane sections remain plane (Bernoulli’s hypothesis) and that a linear
strain distribution can be assumed for all loading stages. The internal state of stress can be derived from
the equilibrium of forces at a cross-section as described in Section 35.4, hence the name 'sectional
methods'. For deep beams (hogewandliggers), however, these assumptions are not valid, because the
shear strains are non-linearly distributed. The non-linearity is caused by abrupt changes in geometry or
by concentrations of loading. Examples of deep beams are beams on two supports with a small height-to-
length ratio, and a bottom plate of a shallow pier foundation. Figure 35-18 shows the difference between
a 'normal' and a deep beam.

Figure 35-18 A typical 'normal' beam (left) and a typical deep beam (right)

The effects of the specific application of loading on a structure will dissipate or smooth out within regions
that are sufficiently far away from the point of application of the load (Saint-Venant’s principle). Usually,
the distance where the way of load application influences the internal stress situation is one member
depth d, see Figure 35-19.

Figure 35-19 Saint-Venant's priciciple illustrated on a beam loaded with compression forces

231
Manual Hydraulic Structures Part III - 'Materials'

In Figure 35-20, the shear span (which is the distance between the applied load and the support) is
indicated with a and the depth d (actually, the distance between from extreme compression fibre to the
centre of the longitudinal tension reinforcement) with d.

Figure 35-20 Strain distribution in the deep and slender portion of a beam (Tuchscherer et.al. 2011)

The figure shows the two types of regions that can be distinguished depending on the shear span to
depth ratio a/d:
• 'B regions' are well understood and can be determined according to sectional methods (B stands for
Beam or Bernoulli).
• 'D regions' indicate the regions of discontinuity (D stands for Discontinuity or Disturbed). Deep beams
are a typical example of D regions, where the shear span is less than twice the depth. For D regions,
the load transfer is assumed to follow an arch action mechanism and the strain distribution across the
section is nonlinear.

Where sectional methods cannot be used, the strut & tie model (staafwerkmodel) is a useful hand
calculations method. Detailed design of complex structures, such as caissons, locks or bridge piers, often
requires a finite element approach. The strut & tie model is a rational approach to represent the structures
by a simplified truss model. However, there is no single or unique strut & tie model that can be generally
applied, so there is a possibility of less efficient or even wrong strut & tie schematisations. Fortunately,
there are techniques and rules that help the designer develop a model to correctly schematise the
structure. If the principles for a lower-bound solution are satisfied, it can be assumed that the resulting
design is safe.

The strut & tie model provides a conceptual framework, where the stress paths in a structure are
schematised as a system of struts, ties and nodes, like a virtual truss. The ties are tension bars in the
truss and consist of reinforcement bars. The struts are compression bars and are formed in the parts of
the concrete in the deep beam. The nodes, connecting the ties and the struts, are the locations where the
reinforcement bars are anchored in the concrete (Figure 35-21).

Figure 35-21 Typical strut & tie model for a deep beam

Struts
There are multiple type of struts, of which the three basic compositions are: the prism, the fan and the
bottle composition (Figure 35-22). The prism is a strut with a constant width, the fan is a composition with
a number of struts with varying inclination that converge to, or diverge from, a single node. The bottle is
somewhat alike the fan type, but there is only one strut that converges or diverges.
232
Manual Hydraulic Structures Part III - 'Materials'

Figure 35-22 Types of struts

Ties
Tension ties include all kinds of reinforcement bars. A critical part in setting up the strut & tie modelling is
providing sufficient anchorage for the reinforcement. If this is not provided, anchorage failure is likely to
occur and the structure will fail at a load smaller than the design load.

Nodes
At the locations where ties and struts intersect, they are connected via nodes. The loads are redirected at
the nodes into other components of the strut & tie model. There are four types of nodes in a 2D-plane:
CCC, CCT, CTT and TTT, in which 'C' stands for compression and 'T' stands for tension. So, for instance,
in a CCC node, three compression components converge (Figure 35-23).

Figure 35-23 Node types (Schlaich et al, 1987)

Design
For the design of a deep structural component with strut & tie modelling, the next steps can be followed
(modified from Williams et al, 2012):

Step 1. Determine the geometry of the strut & tie model

Based on a first notion of the work lines of the external forces and the location of the supports, the
geometry of the strut & tie model can provisionally be determined. This is an intuitive step for experienced
structural designer, which implicitly includes the steps 2 through 9. The difference is that steps 2 through
9 are executed explicitly. The initially chosen geometry (step 1) can be adapted if later design steps give
rise to (so, it is an iterative approach).

Several strut & tie models can be created for the structural component under the considered load case.
233
Manual Hydraulic Structures Part III - 'Materials'

The design of a strut & tie model can be considered conservative, if there is an equilibrium with external
forces and sufficient deformation capacity to allow the distribution of forces along the paths enforced by
the model. A better alignment of the strut & tie model with the flow of forces within the structural element
will result in a lower likelihood of cracking. The struts will generally transfer compression forces with less
deformation than the reinforcement in tension. The ties represent the steel reinforcement within the
structure and should thus be placed where tensile stresses occur.

The model that features the fewest and shortest ties is typically the most efficient and realistic. Loads are
mainly resisted by the strongest and stiffest parts of a structural component. The loads will therefore be
distributed along the strongest and stiffest paths within that component, which, as a result, minimizes the
deformations. In the correct model of Figure 35-24, the forces will naturally flow along the paths, because
it has fewer ties and closely matches the flow of forces.

Figure 35-24 A correct and an incorrect strut & tie model (modified from Williams et al, 2012)

The angle between a strut and a tie entering the same node should not be less than about 25°, because
otherwise tensile and compression stresses would occur within the same vicinity of the node, which is
undesired and unrealistic. The model should therefore include the least number of truss panels while
satisfying the 25° rule. The model in the right part of Figure 35-25 is therefore less efficient than the
model in the left part.

Figure 35-25 An efficient and an inefficient strut & tie model (modified from Williams et al, 2012)

Step 2. Distinguish B- and D-regions

The first step in the design strut & tie process consists of dividing the structural component into B- and D-
regions. In general, D-regions extend to a distance equal to the member depth d from the applied load or
support. A region of a structural member is therefore assumed to be dominated by nonlinear behaviour, if
the shear span, a, is less than about 2 or 2,5 times the member depth, d. These regions are D-regions.
The other regions can be assumed to be B-regions, which can be designed using the sectional design
approach (Section 35.4). If only a small part of the structural component is a B-region, the entire
component can be designed using strut & tie modelling.

Step 3. Define critical load cases and determine the magnitude of the acting loads
The loads acting on the structural component should be determined for each critical load case. The strut
& tie model should be analysed for each critical load case, and, if needed, the geometry of the model
should be modified. These modifications comprise:
• Replacing a moment acting on the structural component by a couple of forces, since moments cannot
be applied to a truss model;
• Concentrated loads at a very short distance form each other can be resolved together as a
simplification;
• A distributed load must be divided into a set of point loads that act at the nodes of the strut & tie
model.
This step is often carried out simultaneously with the development of the strut & tie model (step 5).
234
Manual Hydraulic Structures Part III - 'Materials'

The loads acting on the boundaries of the D-regions are determined. These loads are relevant for
defining the geometry of the component and for determining the forces to be carried by the struts and
ties.

If the structural component consists of both a B-region and a D-region and only part of the component will
be designed using strut-and-tie modelling, the internal forces and moment within the B-region should be
applied at the boundary of the D-region. A linear elastic stress distribution can be assumed at the
interface between the B- and D-regions as shown in Figure 35-26 (C = compression, T = tension). The
calculated loads and boundary forces are then applied to the D-region under consideration.

Figure 35-26 linear stress distribution at the interface of a B- and D-region (modified from Williams et al, 2012)

Step 4. Determine the initial component geometry

To initially determine the dimensions of the structural component, the geometry of the member can be
chosen in a manner that reduces the risk of diagonal crack formation. This can be accomplished by
predicting the likelihood that diagonal cracks will form within the D-regions of an in-service beam. For
beams, the estimated load at which diagonal cracks begin to form, Vcr, appears to be related to the shear
span - effective depth ratio a/d and the effective shear area d·bw. The properties a, bw and d should be
chosen in such a way, that the actual shear force, to be determined with a shear force diagram
(dwarskrachtenlijn), should not exceed the critical shear stress Vcr:

 a
Vactual < Vcr =  6, 5 − 3  f ' bw d
 d 

where:
Vactual [N] = actual shear force
Vcr [N] = critical shear force
a [mm] = shear span
d [mm] = effective depth
2
f' [N/mm ] = compressive strength of the concrete
bw [mm] = width of the structural web

If the actual shear force is larger than Vcr, several measures may be taken:
• increase of the effective shear area, d· bw
• decrease of the ratio a/d
• increase of the compressive strength of the concrete
• apply additional distributed crack control reinforcement

The first three measures are usually the most effective.

Step 5. Determine the forces in the struts and the ties

It is preferable that the truss model is statically determinate, because it will only require equilibrium to
calculate the forces in each member. Figure 35-27 shows the steps to determine the forces in the struts
and ties.

235
Manual Hydraulic Structures Part III - 'Materials'

Figure 35-27 Procedure to determine the forces in a struss & tie model (modified from Williams et al, 2012)

Step 6. Proportion the ties

The area of reinforcement of each tie in the model should be sufficient to resist the acting forces. Ignoring
the use of partial safety factors, this area can be calculated according to:

Ftie
A tie =
fy

This applies, if the centroid of the bars coincides with the position of the tie in the model.

Step 7. Perform strut strength check near the nodes

The actual compression stresses follow the paths of the struts but spread further away from the nodes,
resulting in bottle-shaped struts (Figure 35-28).

Figure 35-28 Actual shape of a strut (left) vs the idealised shape (right)

Nodes are therefore the highest stressed regions and each node should be checked to ensure that it can
resist the imposed forces without crushing the concrete. The width of the path, or strut, is mainly
determined by the dimensions of the bearing plate (the contact surface of the load application) and the
angle of the strut. Furthermore, it matters whether the node is proportioned in a hydrostatic, of non-
hydrostatic way (Figure 35-29). Hydrostatic nodes are proportioned in a manner that causes the stresses
applied to each face to be equal. Non-hydrostatic nodes are proportioned based on the origin of the
applied stress. This proportioning allows the geometry of the nodes to closely correspond to the actual
stress concentrations at the nodes. As the use of hydrostatic nodes sometimes result in unrealistic nodal
geometries and impractical reinforcement layouts, non-hydrostatic nodes are preferred in design.

236
Manual Hydraulic Structures Part III - 'Materials'

Figure 35-29 Two types of nodes: hydrostatic (left) and non-hydrostatic (right)

The actual compression stress in a strut should not exceed the compressive strength of concrete. The
compression stress can be computed by dividing the strut force, as determined in step 5, by the cross-
sectional area of the strut. The cross-sectional area of a strut equals the strut width ws, multiplied with the
strut height hs. In a 2D-schematisation, the strut width is the same as the width of the bearing face. The
height of a strut differs per node type (Figure 35-30):

For CCT nodes: hs = l b sinθ + w t cosθ

For CCC nodes: hs = l b sin θ + a cosθ
where:
hs [m] = height of the strut
lb [m] = length of the bearing plate, or the part of the bearing plate that can be assigned to
the strut under consideration
θ [°] = angle of the strut with the plane perpendicular to the main load direction
wt [m] = length of the back face of a CCT node (usually twice the distance from the bottom
of the beam to the centroid of the longitudinal reinforcement bars)
a [m] = length of the back face of a CCC node

Figure 35-30 Geometry of a CCT node (left) and a CCC node (right) (modified from Williams et al, 2012)

The length of the back face of a CCC node, a, can be taken as the depth of the rectangular compression
stress block, determined form a flexural analysis:

As fs
a=
0,85 fc' bw

2
where: As [m ] = area of tension reinforcement
2
fs [N/m ] = stress in tension reinforcement
2
f'c [N/m ] = compressive strength of concrete
bw [m] = width of the strut = width of the web of the structural member

CTT nodes do not have a clearly defined geometry, because the diagonal strut entering the node is able
to disperse, or smear, over a large volume of concrete (Figure 35-31). Smeared nodes can therefore be
considered as not critical and a check of concrete stresses in these nodes is not necessary, at least not
for preliminary designs.

237
Manual Hydraulic Structures Part III - 'Materials'

Figure 35-31 'Smeared' CTT nodes (modified from Williams et al, 2012)

Step 8. Proportion the crack control reinforcement

Crack reinforcement should be provided to restrain cracks in the concrete caused by the transverse
tension that crosses diagonal bottle-shaped struts (see Figure 35-22c), except for slabs and footings.
0,3% reinforcement must be provided in each orthogonal direction to restrict crack widths. This
percentage is based upon tests performed by Birrcher et al (2009). In equation form, the requirement is:

Av Ah
> 0, 003 and > 0, 003
bw s v bw s h
where:
2
Av [m ] = total area of horizontal crack control reinforcement within spacing sv
2
Ah [m ] = total area of vertical crack control reinforcement within spacing sh
bw [m] = width of the web of the structural member
sv [m] = spacing of vertical crack control reinforcement
sh [m] = spacing of horizontal crack control reinforcement

See Figure 35-32 for an explanation of the variables.

Figure 35-32 Illustration of the variables in the crack control reinforcement requirements (modified from Williams et al, 2012)

Step 9. Provide the necessary anchorage for the ties

The ties must be properly anchored to ensure that the structure can achieve the stress distribution
assumed by the strut & tie model. The yield strength of the reinforcement should therefore be developed
at the point where the centroid of the bars exits the extended nodal zone, as indicated in Figure 35-33.
This implies that the critical section for the development of the tie is taken at the location where the
centroid of the bars intersects the edge of the diagonal strut. The available length should be larger than
the required development length.

238
Manual Hydraulic Structures Part III - 'Materials'

Figure 35-33 Available development length of ties (Williams et al, 2012)

35.9 Literature
Braam, C.R. and Lagendijk, P. (2011). Constructieleer. Gewapend beton deel 2. Aeneas / Cement &
Betoncentrum, Boxtel / 's -Hertogenbosch.

Jungwirth, D. (1985). Begrenzung der Rißbreite im Stahlbeton- und Spannbetonbau aus der Sicht der
Praxis. Beton- und Stahlbau 7/1985, pp 137-178 & 8/1985, pp. 204-208.

Eurocode 2: “Design of concrete structures", various parts and national annex, Nederlands Normalisatie
instituut.

NEN 6008: “Steel for the reinforcement of concrete”, July 2008, Nederlands Normalisatie instituut.

NEN_EN 10080: “Steel for the reinforcement of concrete – Weldable Reinforcing steel - General”, June
2005, Nederlands Normalisatie instituut

NEN-EN 10138 Draft: “Prestressing steels, various parts, Nederlands Normalisatie instituut.

Sagel, R. and A.J. van Dongen, Constructieleer: Gewapend beton, December 2004, ENCI media,
‘s Hertogenbosch.

Williams, C., Deschenes, D. and Bayrak, O. (2012). Strut-and-tie design examples for bridges: final
report. Center for Transportation Research at the University of Texas at Austin.

239
Manual Hydraulic Structures Part III - 'Materials'

36. Steel

major revision: February 2011

Steel is a frequently used material for particularly moveable parts of hydraulic structures and for
temporary structures. A well-known example of a temporary structure is the building pit, where the sheet
pile wall, the struts and the wales consist of steel elements. Steel structural elements for building pits are
dealt with in Chapters 40 and 41 in part IV of this Manual. The present chapter deals with the
characteristics of steel and the most important calculation checks for beams, according to European
standard EN 1993-1-1. More detailed information can be found in the books “(over)spannend staal”, and
the lecture notes of CTB2220 ("Beton- en staalconstructies"). Reinforcement steel to be used in concrete
structures is dealt with in Section 35.2 of this Manual. For an example calculation of struts and wales in a
cofferdam, see part IV, Chapter 44.

36.1 General
Characteristics of steel
The most common types of construction steel are S235 and S355. For the most important properties, see
Table 36-1.

material property ↓ / steel quality → S235 S275 S355

2
yield stress fy at 20 °C [N/mm ] 235 275 355
2
tensile strength fu [N/mm ] 360 430 510
2
Young's modulus E = 210 000 N/mm
Shear modulus (glijdingsmodulus) E
G= ≈ 81 000 N/mm2
2 (1 + ν )
Poisson’s ratio in elastic stage ν = 0,3
Coefficient of linear thermal expansion α = 12 ⋅ 10−6 per °C (for T ≤ 100 °C)
Table 36-1 Steel characteristics for S235, S275 and S355

In this chapter, the partial material factors γ M should be applied to the various characteristic values of
resistance as follows:
• Resistance of cross-sections for all classes: γ M 0 = 1,0
• Resistance of members to instability assessed by member checks: γ M 1 = 1,0
• Resistance of cross-sections in tension to fracture: γ M 2 = 1,25
• Resistance of various joints, see EN 1993-1-8

Ultimate limit state (ULS)

The following aspects should be checked in the ultimate limit state (ULS):
• Exceedance of the yield stress caused by tension, compression, bending, shear and torsion
• Global buckling (of bars, beams or columns) (knik)
• Lateral buckling (of beams, columns) (kip)
• Local (plate) buckling (plooi)
• Fatigue

Serviceability limit state (SLS)

For the serviceability limit state the following should explicitly be checked:
• Deformation / deflection (also plays a significant role for second order effects in ULS)

First or second order analysis?

The internal forces and moments may generally be determined using either:
• First-order analysis, using the initial geometry of the structure or
• Second-order analysis, which also takes into account the influences of the deformation of the
structure.

First order analysis may be used for the structure if the increase of the relevant internal forces or
moments or any other change of the structural behaviour caused by deformations can be neglected. This
condition may be assumed to be fulfilled, if the following criterion is satisfied:
240
Manual Hydraulic Structures Part III - 'Materials'

Fcr
α cr = ≥ 10 for elastic analysis
FEd
Fcr
α cr = ≥ 15 for plastic analysis
FEd
Where:
αcr [-] = amplification factor to further increase the design load that will result in elastic
instability in a global mode
FEd [kN] = design force acting on the structure
Fcr [kN] = elastic buckling force for global instability mode based on initial elastic
π 2EI
stiffnesses: Fcr = 2
ℓ

For single storey frames on the basis of elastic global analysis, second order sway effects due to vertical
loads may be calculated by increasing the horizontal loads HEd (e.g. wind) and equivalent loads VEd due
to imperfections and other possible sway effects according to first order theory by a factor:
1
1
1−
α cr
Classification of cross-sections
Calculation methods for structural elements need to be appropriate with respect to the deformation
characteristics of the profile (mostly dependent on the width/thickness-ratio). The plastic theory, for
instance, should only be used if the profile is able to sufficiently deform in a plastic way. For the sake of
convenience, standards (like the old NEN 6770 - TGB staal, and Eurocode 3) make use of classifications
to indicate what calculation methods have to be used. This classification is useful to determine the extent
to which the resistance and rotation capacity of cross-sections is limited by its local buckling resistance.

Four classes of cross-sections are defined, as follows (see also Table 36-2):
• Class 1: plastic cross-sections which can form a plastic hinge with the rotation capacity of plastic
analysis without reduction of the resistance. Plastic theory may be used for the calculation of
cross-sections and determination of the load distribution.
• Class 2: compact cross-sections (gedrongen doorsneden) which can develop their plastic moment
resistance, but have limited rotation capacity because of local buckling. Plastic theory may
only be used for the calculation of cross-sections.
• Class 3: semi-compact cross-sections in which the stress in the extreme compression fibre of the steel
member assuming an elastic distribution of stresses can reach the yield strength, but local
buckling is liable to prevent development of the plastic moment resistance. Elastic theory
should be used for both cross-section calculations and load distribution determination. No
buckling calculation is required.
• Class 4: slender cross-sections where local buckling will occur before yield stress occurs in one or
more parts of the cross-section. Elastic theory should be used for both cross-section
calculations and load distribution determination. Buckling calculation is required.
•

Table 36-2 Cross-section classes for steel profiles (S235) (from: (over)spannend staal Construeren A).

241
Manual Hydraulic Structures Part III - 'Materials'

36.2 Strength
Possible internal forces in a cross-section are:
• tension • torsion
• compression • combination of bending moment and shear
• bending moment • combination of bending moment and axial force
• shear

Tension
The design value of the tension force NEd at each cross-section shall be smaller than the design tension
NEd
resistance (maximaal toelaatbare trekkracht) Nt,Rd: ≤ 1,0
Nt ,Rd
For sections with holes the design tension resistance Nt,Rd shall be the smallest value of:
A ⋅ fy
a) the design plastic resistance of the gross cross-section N pl ,Rd =
γ M0
0,9 ⋅ Anet ⋅ fu
b) the design ultimate resistance of the net cross-section at holes for fasteners Nu,Rd =
γM2

Compression
NEd
The design value of the compression force NEd at each cross-section shall satisfy: ≤ 1,0
Nc,Rd
The design compression resistance (maximaal toelaatbare drukkracht) of the cross-section for uniform
compression Nc,Rd shall be determined as follows:

A ⋅ fy
Nc ,Rd = for class 1, 2 or 3 cross-sections
γM0
Aeff ⋅ fy
Nc ,Rd = for class 4 cross-sections
γ M0

Bending moment
MEd
The design value of the bending moment MEd at each cross-section should satisfy: ≤ 1,0
Mc ,Rd
The design resistance for bending around one principal axis of a cross-section is determined as follows:
Wpl ⋅ fy
Mc,Rd = M pl ,Rd = for class 1 or 2 cross sections
γ M0
Wel ,min ⋅ fy
Mc,Rd = Mel ,Rd = for class 3 cross sections
γM0
Weff ,min ⋅ fy
Mc,Rd = for class 4 cross sections
γ M0
Where Wel,min and Weff,min correspond to the fibre with the maximum elastic stress.

Shear
VEd
The design value of the shear force VEd at each cross-section should satisfy ≤ 1,0
Vc ,Rd
Where Vc,Rd is the design shear resistance (maximaal opneembare dwarskracht).

For plastic design Vc,Rd is the design plastic shear resistance Vpl,Rd. In absence of torsion the design
plastic shear resistance is given by:

Vpl ,Rd =
(
Av fy 3 ) Where Av is the shear area [mm ].
2
γ M0

For verification of the design elastic shear resistance Vc,Rd the following criterion for a critical point of the
cross-section may be used unless the buckling verification in section 5 of EN 1993-1-5 is applicable:
242
Manual Hydraulic Structures Part III - 'Materials'

τ Ed VEd ⋅ S
≤ 1,0 Where τ Ed may be obtained from: τ Ed =
fy ( 3γ M 0 ) I ⋅t
Where:
VEd [N] = design value of the shear force
3
S [mm ] = first moment of area (statisch moment) with respect to the centroidal axis of
the portion of the cross-section between the point where shear has to be
considered and the boundary of the cross-section.
4
I [mm ] = second moment of area (traagheidsmoment) of the entire cross-section.
t [mm] = thickness at the considered point

VEd VEd VEd

τ Ed = τ Ed = τ Ed =
b⋅h 1
2 ⋅ π ⋅ d ⋅ t − 41 ⋅ π ⋅ t 2 tw ⋅ hw

Torsion
For members subject to torsion for which distortional deformations may be disregarded, the design value
T
of the torsional moment TEd at each cross-section should satisfy: Ed ≤ 1,0
TRd
Where TRd is the design torsional resistance (torsiemomentweerstand) of the cross-section.

The total torsional moment TEd at any cross-section should be considered as the sum of two internal
effects: TEd = Tt ,Ed + Tw ,Ed
Where Tt,Ed is the internal St. Venant torsion (inwendige St. Venantse wringing)
Tw,Ed is the internal warping torsion (inwendige wringing met verhinderde welving)

The following stresses due to torsion should be taken into account:

a) the shear stresses τ t ,Ed due to St. Venant torsion Tt,Ed
b) the direct stresses σ w ,Ed due to the bi-moment BEd and shear stresses τ w ,Ed due to warping
torsion Tw,Ed.

The above method is prescribed in Eurocode 3 (NEN-EN 1993), which is meant for use in mechanical
engineering. For this course, it will go too far for detailing this. For checks on torsion, the calculation
method mentioned in TGB1990 (Technische Grondslagen voor Bouwconstructies 1990) may be used in
the Netherlands (the way of calculation differs per country).

Combination of bending and shear

Where shear force is present, its effect on the moment resistance (maximaal opneembare moment)
should be taken into account, except if the shear force is less than half the plastic shear resistance (in
which case it may be neglected). Otherwise the reduced moment resistance calculated using a reduced
yield strength should be taken as the design resistance of the cross-section: (1 − ρ ) ⋅ fy for the shear area,
where:
2
 2 ⋅ VEd  Av ⋅ (fy / 3 )
ρ = − 1 and Vpl ,Rd =
V  γ M0
 pl ,Rd 

243
Manual Hydraulic Structures Part III - 'Materials'

in which Av = shear area (slide plane) and γM0 = partial factor for cross-sections for the resistance to
instability (in most cases γM0 = 1,0).

2
 2 ⋅ VEd 
If also torsion is present, ρ should be obtained from ρ =  − 1 , but should be taken 0 if
V 
 pl ,T ,Rd 
VEd ≤ 0,5 ⋅ Vpl ,T ,Rd , where Vpl,T,Rd depends on the profile shape (see NEN-EN 1993-1-1, article 6.2.7).

Combination of bending and axial force

For class 1 and 2 cross-sections, the following criterion should be satisfied: MEd ≤ MN ,Rd
Where MN,Rd is the design plastic moment resistance reduced due to the axial force NEd.
For a rectangular solid section without fastener holes MN,Rd should be taken as:

MN ,Rd = M pl ,Rd 1 − ( NEd N pl ,Rd ) 

2

 

For double symmetrical I- and H-sections or other flanges sections, the effect of the axial force on the
plastic moment resistance around the y-y axis does not have to be taken into account if both the following
criteria are satisfied:
0,5 ⋅ hw ⋅ tw ⋅ fy
NEd ≤ 0,25 ⋅ N pl ,Rd and NEd ≤
γ M0

The effect of the axial force on the plastic moment resistance around the z-z axis may be neglected if:

hw ⋅ tw ⋅ fy
NEd ≤
γ M0
For bi-axial bending the following criterion may be used:

α β
 M y ,Ed   M z,Ed 
  +   ≤ 1,0
 MN ,y ,Rd   MN ,z,Rd 

in which α and β are constants, which may conservatively be assumed to be 1, otherwise as follows:
• I- and H-sections: α = 2 and β = 5 ⋅ n but β ≥ 1
• Circular hollow sections: α = 2 and β = 2
1,66
• Rectangular hollow sections: α = β = but α and β ≤ 6
1 − 1,13 ⋅ n 2
Where n = NEd N pl ,Rd

In absence of shear force, for Class 3 cross-sections the maximum longitudinal stress σ x ,Ed should
fy
satisfy the criterion: σ x,Ed ≤ , where σ x ,Ed is the design value of the local longitudinal stress due to
γM0
moment and axial force taking account of relevant fastener holes (boutgaten).

In absence of shear force, for Class 4 cross-sections the maximum longitudinal stress σ x ,Ed calculated
fy
using the effective cross-sections should satisfy the criterion: σ x,Ed ≤ , where σ x ,Ed is the design value
γM0
of the local longitudinal stress due to moment and axial force taking account of fastener holes, which are
relevant here. The following criterion should be met:

NEd M y ,Ed + NEd ⋅ eNy M z,Ed + NEd ⋅ eNz

+ + ≤ 1,0
Aeff ⋅ fy γ M 0 Weff ,y ,min ⋅ fy γ M 0 Weff ,z,min ⋅ fy γ M 0

244
Manual Hydraulic Structures Part III - 'Materials'

where:
2
Aeff [mm ] = effective area of the cross-section when subjected to uniform compression
3
Weff,min [mm ] = effective section modulus (corresponding to the fibre with the maximum
elastic stress) of the cross-section when subjected only to moment about
the relevant axis
eN [mm] = shift of the relevant centroidal axis when the cross-section is subjected to
compression only

Combination of bending, shear and axial force

Provided that the design value of the shear force VEd does not exceed 50% of the design plastic shear
resistance Vpl,Rd, no reduction of the resistance defined for bending and axial force need be made, except
where shear buckling reduces the section resistance, see EN 1993-1-5.

Where VEd exceeds 50% of Vpl,Rd the design resistance of the cross-section to combinations of moment
and axial force should be calculated using a reduced yield strength:

2
 2 ⋅V 
(1 − ρ ) ⋅ fy for the shear area where ρ =  Ed − 1 .
 Vpl ,Rd 

36.3 Stability (buckling)

Buckling (knik) is a failure mechanism that occurs due to a normal force acting on a beam (or other
structural element). Lateral buckling (kip) occurs if there is a shear force acting in the element. Both are
displacements in the direction perpendicular to the load direction. Dependent on the forces, this failure
mechanism must be checked.

Buckling resistance
NEd
A structural part under compression should be verified against buckling as follows: ≤ 1,0
Nb,Rd
Where:
NEd [kN] = the design value of the compression force
Nb,Rd [kN] = the design buckling resistance of the compression member

For structural members with non-symmetric Class 4 sections the additional moment ∆MEd due to the
eccentricity of the centroidal axis of the effective section should be taken into account.

The design buckling resistance of a compression member should be taken as:

χ ⋅ A⋅ fy
Nb,Rd = for Class 1, 2 and 3 cross-sections
γ M1

χ ⋅ Aeff ⋅ fy
Nb,Rd = for Class 4 cross-sections
γ M1
Where χ is the reduction factor for the relevant buckling mode, to be taken from Figure 36-1.

Buckling curves
For axial compression in members the value of χ for the appropriate non-dimensional slenderness λ
should be determined using the relevant buckling curve.
A ⋅ fy
λ= for Class 1, 2 and 3 cross-sections
Ncr
Aeff ⋅ fy
λ= for Class 4 cross-sections
Ncr

Ncr is the elastic critical force for the relevant buckling mode based on the gross cross-sectional
2 2
properties. Ncr = Fcr = π EI/ℓ for classes 1, 2, and 3.

The appropriate buckling curve should be obtained from Table 36-3. Values of the reduction factor χ for
the appropriate non-dimensional slenderness λ may be obtained from Figure 36-1.
245
Manual Hydraulic Structures Part III - 'Materials'

NEd
For slenderness λ ≤ 0,2 or for ≤ 0,04 the buckling effects may be ignored and only cross-sectional
Ncr
checks apply.

Table 36-3 Buckling curves for various cross-sections.

Figure 36-1 Relation between the reduction factor χ and the non-dimensional slenderness (from: NEN-EN 1993-1-1)
Note: in the older standard NEN 6770, χ is indicated with ωbuc (buckling factor) and λ with λrel (relative slenderness).

246
Manual Hydraulic Structures Part III - 'Materials'

Slenderness for flexural buckling (buigingsknik)

The non-dimensional slenderness λ is given by
A ⋅ fy Lcr 1
λ= = for Class 1, 2 and 3 cross-sections
Ncr i λ1
Aeff
Aeff ⋅ fy L A
λ= = cr for Class 4 cross-sections,
Ncr i λ1

Where: Lcr [m] = buckling length in the buckling plane considered (= lef in Figure 36-2)
i [m] = radius of gyration around the relevant axis, determined using the moment of inertia
I
and the area of the cross-section =
A
E
λ1 [-] = π⋅ = 93,9 ⋅ ε
fy
235
ε [-] =
fy

Figure 36-2 Effective buckling length.

Uniform members in bending

Buckling resistance (kipweerstand)

A laterally unrestrained member subject to major axis bending should be verified against lateral torsional
buckling as follows:

MEd
≤ 1,0
M b,Rd

Where MEd [Nm] = design value of the moment

Mb,Rd [Nm] = design buckling resistance moment

247
Manual Hydraulic Structures Part III - 'Materials'

The design buckling resistance moment of a laterally unrestrained beam should be taken as:
fy
Mb,Rd = χ LT ⋅ Wy ⋅
γ M1
Where:
3
Wy [m ] = the appropriate section modulus as follows:
Wy = Wpl,y for Class 1 or 2 cross-sections
Wy = Wel,y for Class 3 cross-sections
Wy = Weff,y for Class 4 cross-sections
χ LT [-] = the reduction factor for lateral-torsional buckling

Lateral torsional buckling curves (kipkrommen)

For determining the value of the reduction factor χ LT the curve of Figure 36-1 can be used, and

W y ⋅ fy
λ LT = ,
Mcr

in which Mcr is the elastic critical moment for lateral-torsional buckling and is based on gross cross-
sectional properties and takes into account the loading conditions, the real moment distribution and the
lateral restraints. Recommended values for lateral buckling curves for cross-sections, see Table 36-4.

cross-section Limits Buckling curve

h/b ≤ 2 a
rolled I-sections
h/b > 2 b
h/b ≤ 2 c
welt I-sections
h/b > 2 d
other cross-sections - d
Table 36-4 buckling curves per cross-section

36.4 Welded connections

The most often used type of weld in steel structures and especially temporary steel structures, is the fillet
weld (hoeklas). This is the only type of weld discussed here, based on the general weld calculations.

Three types of connections in the building pit example have to be designed:

1. Pipe strut – long waling (Figure 36-3)
2. Short waling – long waling (4 corners of the building pit) (Figure 36-4)
3. H-beam strut – waling (struts shoring the wales in the corners with an angle of 45º).

These connections can be chosen to be designed as rigid connections or hinges. Especially for a
temporary structure such as a building pit, rigid connections are an appropriate choice for all 3 types of
connections, but most of the time hinges are preferred. The welds will have to transfer normal forces,
shear and bending moments, but the span moments will mostly be smaller. The connection can be
strengthened with extra steel plates or stiffeners (schotjes, kopplaat).

248
Manual Hydraulic Structures Part III - 'Materials'

Figure 36-3 Welded connection of pipe strut and waling. The picture on the left (cofferdam near Sebastiaansbrug in Delft, 2007)
shows a wale consisting of single H-profile, whereas the drawing on the right depicts the more common case of a
double profile.

The theory that is stated in this section applies only to weldable structural steel with a material thickness
larger or equal to 4 mm. For welds on thinner material the reader is referred to EN 1993-1-3. Also the
mechanical properties, such as the yield strength (fy); ultimate tensile strength (fu); etc., of the weld
material should be equivalent or better than the parent material, this ensures that the weld is at least as
strong as the materials welded together.

Fillet welds may be used for connecting parts where the fusion faces form an angle between 60° and
120°. Fillet welds finishing at the ends or sides of parts that are connected should continue, full-size,
around the corner for a distance of at least twice the leg length of the weld; unless access or the
configuration of the joint renders this impracticable. Since a fillet weld is roughly rectangular in cross-
section the leg length is defined as the length along the base material, or in other words the length of the
sides of the triangle that make an angle of 90° (see Figure 36-4). Note that the length of the leg can also
be unequal, which, however, is not common practice.

Figure 36-4 Connection between two walings, with front plate (σ1 = σ┴ and σ2 = σ//)
Normal forces in the structural members will either act parallel or rectangular to the weld. They cause a
stress in the direction parallel or rectangular to the weld, σ// or σ⊥ respectively. The bending moments on
a rigid welded connection also cause stresses in these directions.

249
Manual Hydraulic Structures Part III - 'Materials'

Figure 36-5: Standard cases 1, 2 and 5 of a normal force and shear force in a connected member.
(Overspannend staal– Construeren A)

Figure 36-5 shows three basic standard cases of fillet welds connecting rectangular members with one
weld per side; ℓef is the effective length of the member, a is the throat thickness.

1. Normal force in connected member

Fd ⋅ 2
σ⊥ = τ⊥ = ; τ =0
4 ⋅ a ⋅ ℓ eff

2. Shear force or load in connected member, perpendicular to the weld length

2 ⋅ Fd
σ⊥ = τ⊥ = ; τ =0
4 ⋅ a ⋅ ℓ eff

5. A bending moment on the connected member

A concentrated load on the member is modelled as a concentrated load plus a bending moment on the
connection. The stresses resulting from the bending moment alone are:
3 ⋅ Ms;d
In points A and B: σ⊥ = τ⊥ = ; τ =0
2 ⋅ a ⋅ ℓ eff 2
In point C: σ⊥ = τ⊥ = 0 ; τ = 0

According to European standard EN 1993-1-8 the design resistance of a fillet weld can be determined
using either the directional method or the simplified method, which will be discussed in the remainder of
this section. Important notions are the effective length (leff) and the effective throat thickness (a) of the
weld.

The effective length of a fillet weld should be taken as the length over which the fillet is full-size. This may
be taken as the overall length of the weld reduced by twice the effective throat thickness. Provided that
the weld is full-size throughout its length including starts and terminations, no reduction in effective length
need be made for either the start or the termination of the weld. Note: a fillet weld with an effective length
less than 30 mm or less than 6 times its throat thickness, whichever is larger, should not be designed to
carry load.

250
Manual Hydraulic Structures Part III - 'Materials'

Figure 36-6 Throat thickness of a (deep penetration) fillet weld.

The effective throat thickness of a weld should be taken as the height of the largest triangle (with equal or
unequal legs) that can be inscribed within the fusion faces and the weld surface, measured perpendicular
to the outer side of this triangle, see Figure 36-6. The effective throat thickness of a fillet weld should not
be less than 3 mm for practical reasons. In determining the design resistance of a deep penetration fillet
weld, one may take the its additional throat thickness into account, provided that preliminary tests show
that the required penetration can be achieved consistently.

Directional method
In this method the forces transmitted by a unit length of weld are resolved into components parallel and
transverse to the longitudinal axis of the weld and normal and transverse to the plane of its throat (a). The
design throat area (Aw) should be taken as Aw = ∑a·leff and should be assumed concentrated in the root of
the weld.

An uniform distribution of stress is assumed on the throat section of the weld, leading to normal and shear
stresses as shown in Figure 36-7.

Figure 36-7 stresses on the throat section of a fillet weld.

where:
2
σ┴ [N/mm ] = normal stress perpendicular to the throat
2
σ// [N/mm ] = normal stress parallel to the axis of the weld
2
τ┴ [N/mm ] = shear stress (in the plane of the throat) perpendicular to the axis of the weld
2
τ// [N/mm ] = shear stress (in the plane of the throat) parallel to the axis of the weld

The design resistance of the filled weld is sufficient if the following equations are both satisfied:

0,9 ⋅ fu
σ d = σ ⊥ 2 + 3 (τ ⊥ 2 + τ )
fu
2
≤ and σ⊥ ≤
βw ⋅ γ M 2 γM2
where:
2
fu [N/mm ] = nominal tension strength of the weaker part joined
βw [-] = appropriate correlation factor taken from Table 36-5
γM2 [-] = partial factor for the tensile strength (γM2 = 1,25)

Using these criterions the throat thickness of the weld can be determined, as stated before the practical
minimum amounts to 3 mm. Note that the stress parallel to the axis, σ//, is not considered when verifying
the design resistance of the weld.

When welds connect parts with two different material strengths, the weld should be designed using the
properties of the material with the lower strength grade.

251
Manual Hydraulic Structures Part III - 'Materials'

Steel grade Tensile stress Correlation factor

2
S235 360 N/mm 0,80
2
S275 275 N/mm 0,85
2
S355 360 N/mm 0,90
2
S420 360 N/mm 1,00
2
S460 360 N/mm 1,00

Table 36-5 correlation factor for fillet welds

Simplified method
Alternatively the design resistance of a fillet weld may be assumed to be adequate if, at every point along
its length, the resultant of all the forces per unit length transmitted by the weld satisfy the following
criterion:

Fw .Ed ≤ Fw .Rd

Where:
Fw,Ed [N/mm] = design value of the weld force per unit length
Fw,Rd [N/mm] = design weld resistance per unit length

Independent of the orientation of the weld throat plane to the applied force, the design resistance per unit
length should be determined from:

fu
Fw .Rd = fvw ,d ⋅ a where fvw ,d =
3 ⋅ βw ⋅ γ M 2
Where:
2
fw,Rd [N/mm ] = design shear strength of the weld

NB. The deformations of temporary structures, such as, struts are not often checked. They could be
estimated using the deformation diagrams (vergeet-me-nietjes) from structural mechanics and then
checked against the requirements.

252
Manual Hydraulic Structures Part III - 'Materials'

36.5 Bolt connections

The partial safety factor γ M for the tensile strength of bolts (bouten) is γ M 2 = 1,25 . The partial safety factor
for the preload of high strength bolts is γ M 7 = 1,1 . Detailed information about connections made by bolts
can be found in EN 1993-1-8. The rules in this standard are valid for the bolt classes given in Table 36-6.

Figure 36-6 Two angular sections (hoekprofiel) are bolted to a T-profile

The yield strength fyb and the ultimate tensile strength fub for bolt classes 4.6, 4.8, 5.6, 5.8, 6.8, 8.8 and
10.9 are given in Table 36-6. These values should be adopted as characteristic values in design
calculations.

Bolt class 4.6 4.8 5.6 5.8 6.8 8.8 10.9

2
fyb (N/mm ) 240 320 300 400 480 640 900
2
fub (N/mm ) 400 400 500 500 600 800 1000
Table 36-6 Bolt classes (Nominal values of the yield strength and the ultimate tensile strength for bolts.)

Figure 36-7 Cofferdam for the construction of the metro line in Thessaloniki, with bolt connections struts-wales (August 2009)

253
Manual Hydraulic Structures Part III - 'Materials'

Categories of bolted connections

Shear connections
Bolted connections loaded in shear should be designed as one of the following categories:

Category A : Bearing type

In this category bolts from class 4.6 up to and including class 10.9 should be used.
No preloading and special provisions for contact surfaces are required. The design
ultimate shear load should not exceed the design shear resistance, nor the design
bearing resistance.

Category B : Slip-resistant at serviceability limit state

In this category preloaded bolts from classes 8.8 and 10.9 should be used. Slip
should not occur at the serviceability limit state. The design serviceability shear load
should not exceed the design slip resistance. The ultimate shear load should not
exceed the design shear resistance, nor the design bearing resistance.

Category C : Slip-resistant at ultimate limit state

In this category preloaded bolts from classes 8.8 and 10.9 should be used. Slip
should not occur at the ultimate limit state. The design ultimate shear load should
not exceed the design slip resistance, nor the design bearing resistance. In addition
for a connection in tension, the design plastic resistance of the net cross-section at
bolt holes should be checked at the ultimate limit state.

Tension connections
Bolted connections loaded in tension should be designed as one of the following categories:

Category D : Non-preloaded
In this category bolts from class 4.6 up to and including class 10.9 should be used.
No preloading is required. This category should not be used where the connections
are frequently subjected to variations of tensile loading. However, they may be used
in connections designed to resist normal wind loads.

Category E : Preloaded
In this category preloaded bolts from classes 8.8 and 10.9 with controlled tightening
in conformity with 1.2.7 Reference Standards: Group 7 should be used.

254
Manual Hydraulic Structures Part III - 'Materials'

Positioning of holes for bolts (and rivets)

Minimum and maximum spacing and end and edge distances for bolts (bouten) and rivets (klinknagels)
are given in Table 36-7.
1) 2) 3)
Distances and Minimum Maximum
spacings,
see figure 26-10 Structures made from steels conforming to Structures made from
EN 10025 except steels conforming to steels conforming to
EN 10025-5 EN 10025-5
Steel exposed to the Steel not exposed to Steel used unprotected
weather or other the weather or other
corrosive influences corrosive influences
End distance e1 1,2 d0 4t + 40 mm The larger of
8t or 125 mm
Edge distance e2 1,2 d0 4t + 40 mm The larger of
8t or 125 mm
4)
Distance e3 in 1,5 d0
slotted holes
4)
Distance e4 in 1,5 d0
slotted holes
Spacing p1 2,2 d0 The smaller of The smaller of The smaller of
14t or 200 mm 14t or 200 mm 14tmin or 175 mm
Spacing p1,0 The smaller of
14t or 200 mm
Spacing p1,i The smaller of
28t or 400 mm
5)
Spacing p2 2,4 d0 The smaller of The smaller of The smaller of
14t or 200 mm 14t or 200 mm 14tmin or 175 mm
1)
Maximum values for spacing, edge and end distances are unlimited, except for the following cases:
- for compression members in order to avoid local buckling and to prevent corrosion in exposed
members and;
- for exposed tension members to prevent erosion.
2)
The local buckling resistance of the plate in compression between the fasteners should be calculated
according to EN 1993-1-1 using 0,6 p1 as buckling length. Local buckling between the fasteners need
not to be checked if p1 / t is smaller than 9 ε. The edge distance should not exceed the local buckling
requirements for an outstand element in the compression members, see EN 1993-1-1. The end
distance is not affected by this requirement.
3)
t is the thickness of the thinner outer connected part.
4)
The dimensional limits for slotted holes are given in 1.2.7 Reference Standards: Group 7.
5)
For staggered rows of fasteners a minimum line spacing of p2 = 1,2 d0 may be used, provided that the
minimum distance, L, between any two fasteners is greater or equal than 2.4 d0, see figure 26-10 b).
Table 36-7 Minimum and maximum spacing, end and edge distances of bolts

Figure 36-8 Explanatory sketch of 'ends' and 'edges'

255
Manual Hydraulic Structures Part III - 'Materials'

Where:
e1 [mm] = the end distance from the centre of a fastener hole to the adjacent end of any part,
measured in the direction of load transfer
e2 [mm] = the edge distance from the centre of a fastener hole to the adjacent edge of any part,
measured at right angles to the direction of load transfer
e3 [mm] = the distance from the axis of a slotted hole to the adjacent end or edge of any part
e4 [mm] = the distance from the centre of the end radius of a slotted hole to the adjacent end or
edge of any part
p1 [mm] = the spacing between centres of fasteners in a line in the direction of load transfer
p1,0 [mm] = the spacing between centres of fasteners in an outer line in the direction of load transfer
p1,i [mm] = the spacing between centres of fasteners in an inner line in the direction of load transfer
p2 [mm] = the spacing measured perpendicular to the load transfer direction between adjacent
lines of fasteners
d0 [mm] = the hole diameter for a bolt, a rivet or a pin

Figure 36-9 Symbols for end and edge distances and spacing of fasteners

256
Manual Hydraulic Structures Part III - 'Materials'

Design resistance of individual fasteners

The design resistance for an individual fastener subjected to shear and/or tension is given in Table 36-8.
For preloaded bolts the design preload, Fp,Cd, to be used in design calculations is:

Fp,Cd = 0,7 ⋅ fub ⋅ As γ M 7

where:
Fp,Cd [N] = design preload
2
fub [N/mm ] = ultimate tensile strength
2
As [mm ] = tensile stress area of the bolt
γm7 [-] = partial factor for preload of high strength bolts (γm7 = 1,1)

In single lap joints with only one bolt row, the bolts should be provided with washers (ringen) under both
the head and the nut (moer). The design bearing resistance (stuikweerstand) for each bolt should be
limited to:

Fb,Rd ≤ 0,7 ⋅ fu ⋅ d ⋅ t γ M 2
where:
Fb,Rd [N] = design bearing resistance
2
fu [N/mm ] = tensile stress
d [mm] = nominal bolt diameter
t [mm] = thickness of the thinner outer connected part
γm2 [-] = partial factor for the tensile strength (γm2 = 1,25)

Figure 36-10 Single lap joint with one row of bolts

257
Manual Hydraulic Structures Part III - 'Materials'

Failure mode Bolts Rivets

Shear resistance per shear plane α ⋅f ⋅ A 0,6 ⋅ fur ⋅ A0
Fv ,Rd = v ub Fv ,Rd =
γ M2 γM2
- where the shear plane passes through the
threaded portion of the bolt (A is the tensile
stress area of the bolt As)
- for classes 4.6, 5.6, 8.8: αv = 0.6
- for classes 4.8, 5.8, 6.8 and 10.9: αv = 0,5
- where the shear plane passes through the
unthreaded portion of the bolt (A is the gross
cross- section of the bolt): αv = 0,6

k1 ⋅ α b ⋅ fu ⋅ d ⋅ t
1) 2) 3)
Bearing resistance
Fb,Rd =
γ M2
fub
- where αb is the smallest of αd ; or 1,0
fu
- in the direction of load transfer:
e p1 1
- for end bolts: α d = 1 ; for inner bolts: α d = −
3 ⋅ d0 3 ⋅ d0 4
- perpendicular to the direction of load transfer:
e
- for edge bolts: k1 is the smallest of 2,8 ⋅ 2 − 1,7 or 2,5
d0
p2
- for inner bolts: k1 is the smallest of 1,4 ⋅ − 1,7 or 2,5
d0

2)
Tension resistance k2 ⋅ fub ⋅ As 0,6 ⋅ fur ⋅ A 0
Ft ,Rd = Ft ,Rd =
γ M2 γ M2
where k2 = 0,63 for countersunk bolt,
otherwise k2 = 0,9

Punching shear resistance 0,6 ⋅ π ⋅ d m ⋅ t p ⋅ fu No check needed

Bp,Rd =
γ M2
Combined shear and tension Fv ,Ed Ft ,Ed
+ ≤ 1,0
Fv ,Rd 1.4 ⋅ Ft ,Rd
1)
The bearing resistance Fb,Rd for bolts
- in oversized holes is 0,8 times the bearing resistance fro bolts in normal holes.
- in slotted holes, where the longitudinal axis of the slotted hole is perpendicular to the direction of the
force transfer, is 0,6 times the bearing resistance for bolts in round, normal holes.
2)
For countersunk bolt:
- the bearing resistance Fb,Rd should be based on the plate thickness t equal to the thickness of the
connected plate minus half the depth of the countersinking.
- for the determination of the tension resistance Ft,Rd the angler and depth of countersinking should
conform with 1.2.4 Reference Standards: Group 4, otherwise the tension resistance Ft,Rd should be
adjusted accordingly.
3)
When the load on a bolt is not parallel to the edge, the bearing resistance may be verified separately
for the bolt load components parallel and normal to the end.
Table 36-8 Design resistance for individual fasteners subjected to shear and/or tension

258
Manual Hydraulic Structures Part III - 'Materials'

where:
Fv,Rd [N] = the design shear resistance per bolt
Fb,Rd [N] = the design bearing resistance per bolt
Ft,Rd [N] = the design tension resistance per bolt
Fv,Ed [N] = the design shear force per bolt for the ultimate limit state
Ft,Ed [N] = the design tensile force per bolt for the ultimate limit state
Bp,Rd [N] = the design punching shear resistance of the bolt head and the nut
2
fu [N/mm ] = tensile stress
2
fub [N/mm ] = ultimate tensile strength
2
fur [N/mm ] = the specified ultimate tensile strength of the rivet
αv [-] = factor defined in the table
αd [-] = factor defined in the table
αb [-] = factor defined in the table
2
A [mm ] = the gross cross-section area of bolt
2
As [mm ] = the tensile stress area of the bolt
2
A0 [mm ] = the area of the rivet hole
γM2 [-] = partial factor for the tensile strength (γM2 = 1,25)
tp [mm] = the thickness of a plate
k1 [-] = factor defined in the table
k2 [-] = factor defined in the table
e1 [mm] = the end distance from the centre of a fastener hole to the adjacent end of any
part, measured in the direction of load transfer
e2 [mm] = the edge distance from the centre of a fastener hole to the adjacent edge of any
part, measured at right angles to the direction of load transfer
p1 [mm] = the spacing between centres of fasteners in a line in the direction of load
transfer
p2 [mm] = the spacing measured perpendicular to the load transfer direction between
adjacent lines of fasteners
dm [mm] = the mean of the across points and across flats dimensions of the bolt head or
the nut, whichever is smaller
d0 [mm] = the hole diameter for a bolt, a rivet or a pin

Group of fasteners

The design resistance of a group of fasteners may be taken as the sum of the design bearing resistances
Fb,Rd of the individual fasteners, provided that the design shear resistance Fv,Rd of each individual fastener
is greater than or equal to the design bearing resistance Fb,Rd . Otherwise the design resistance of a group
of fasteners should be taken as the number of fasteners multiplied by the smallest design resistance of
any of the individual fasteners.

259
Manual Hydraulic Structures Part III - 'Materials'

36.6 Fatigue (of welds)

This section presents an introduction to the phenomenon of fatigue of welds. Of all structural members,
welds are, in general for hydraulic structures, the most vulnerable to fatigue. Reference is made to the
lecture notes of CIE5126 on fatigue and NEN-EN 1993-1-9 “Fatigue”.

Fatigue is the progressive and localized structural damage that occurs when a material is subjected to
cyclic loading. Final failure occurs in regions of tensile stress, if the reduced governing cross-section of
the structural member becomes insufficient to bear the repetitive peak load without rupture. Fatigue is a
major threat for structures under dynamic loads, such as bridges, cranes and offshore structures, where
the live load presents a high portion of the total load. This is mostly the case in steel structures, where the
dead weight is relatively low compared to the total load. Fatigue mainly leads to static failure by rupture
(cracking) or buckling.

For a simple check, follow these design steps:

1. Locate the stress areas in the structure
2. Locate geometrical stress concentrations in the structure
3. Determine the fatigue strength
4. Determine the dynamic loads resulting in fatigue
5. Define the number of cycles n, using a cycle counting method
6. Check if damage will occur (Dd = Σ (ni / Ni) ≤ 1) (Miner’s summation)

Ad 1/2. Design simple structural details. Start checking fatigue in critical sections, such as welds,
connections, changes in structural member and high live load/dead load ratio etc.

Figure 36-12: Discontinuities, stress peaks and cracks at the toe of a fillet weld

Ad3. Fatigue strength

The fatigue strength of a weld component is defined as the stress range (∆σR) between the minimum and
maximum stress in the cycle which would cause failure of the structural member when fluctuating with a
constant amplitude for a specified number of cycles (NR). The number of cycles to failure NR is known as
the endurance or fatigue life.

Figure 36-13: Typical crack growth curve

d
NR = or log NR = log d - m ⋅ log ∆σ R
∆σ Rm
where:
m = slope of the fatigue strength curve (3 for most welded details)
d = design weld strength parameter, or the fatigue detail category coefficient, dependent on
the stress range, see Table 36-9.
260
Manual Hydraulic Structures Part III - 'Materials'

Detail Category
∆σ c d m
2 [-]
[N/mm ]
12
160 7,962 10 3
12
140 5,636 10 3
12
125 3,990 10 3
12
112 2,825 10 3
12
100 2,000 10 3
12
90 1,416 10 3
12
80 1,002 10 3
12
71 0,710 10 3
12
63 0,502 10 3
12
56 0,356 10 3
12
50 0,252 10 3
12
45 0,178 10 3
12
40 0,126 10 3
12
36 0,089 10 3
Table 36-9 Classification table

∆σ c is the characteristic value for the weld-class applied, and represents the value for the stress range
7
that is exceeded 10 times during the lifetime of the structure (see Figure 36-11).

Figure 36-11: Constant amplitude stress history Figure 36-12: Typical S-N curve for constant amplitude tests

Beyond 5-10 million cycles, the stress range is generally too small to allow crack propagation under
constant amplitude loading. Beneath this limit stress range cracks will not grow. The force fluctuation shall
be calculated using the elasticity theory. No plastic redistribution is permitted because that would
decrease the safety margin of the design.

Ad. 6 Calculation of damage.

The load spectrum is often simplified into a limited number of bands, see Figure 36-13. The damage per
band is defined as n / N, where n is the number of cycles in the band and N is the endurance belonging to
that stress range.

Figure 36-13 Simplification of stress spectrum and determination of endurance for each band (from lecture notes CT5126)
261
Manual Hydraulic Structures Part III - 'Materials'

The damage during the design life (Dd) as a result of all stress bands should not exceed unity:

ni
Dd = ∑ ≤1
Ni

In the design, the following aspects should be taken into account:

• Fatigue depends on the whole service loading sequence, not just one extreme load event.
• Fatigue is very sensitive to the geometry of details. Details should be simple and smoothness of the
stress path should be ensured.
• In welt steel structures, fatigue cracks will almost certainly start to grow from welds, because
- most welding processes leave minute metallurgical discontinuities from which cracks may grow
(peak stresses at discontinuity).
- At the toes of butt welds, sharp changes of direction often occur. Discontinuities at these points
cause local stress concentrations, therefore cracks will grow faster.
• Fatigue requires more accurate prediction of elastic stress.
• A structure designed on fatigue makes more demands on workmanship and inspection.
• The static design safety margins are not sufficient for fatigue, therefore during the conceptual design
should be checked whether fatigue is likely to be critical.
• Check areas of high live load/dead load ratios first on fatigue.

36.7 Literature
Nederlands Normalisatie instituut, Eurocode 3: “Design of steel structures – Part 1-1: General rules for
buildings” (NEN-EN 1993-1-1), januari 2006, Nederlands Normalisatie instituut.

Nederlands Normalisatie instituut, Eurocode 3: “Design of steel structures – Part 1-8: Design of joints”
(NEN-EN 1993-1-8), januari 2006, Nederlands Normalisatie instituut.

Nederlands Normalisatie instituut, Eurocode 3: “Design of steel structures – Part 1-9: Fatique” (NEN-EN
1993-1-9), januari 2006, Nederlands Normalisatie instituut.

262
Equation Chapter 37 Section 1

Manual Hydraulic Structures

Part IV: temporary and permanent structures

Manual Hydraulic Structures Part IV 'Structures'

37. Stability of structures on shallow foundations

New chapter: February 2011; updated February 2015; 'Piping' section slightly improved in 2016 and 2019
'Scour protection' improved in 2018, 2019

A foundation is denominated 'shallow', if the structure is supported by a soil stratum near the ground sur-
face. Shallow foundations are usually applied on non-compressible soils (sand) and can consist of pad
footings (poeren), rafts/strips (stroken), or slabs (platen). For large structures, like most hydraulic struc-
tures, slab foundations are usual, where a part or the whole slab area is used to transfer forces to the
subsoil. Frost edges (vorstranden) should reach as deep as the frost free level.

Shallow foundations are usually less expensive than deep foundations, and also easier to construct.
Therefore it is advised to first check whether a shallow foundation would suffice. The bearing capacity of
the shallow soil stratum should suffice and the deformations (settlement) should be acceptable, possibly
after soil improvement (grondverbetering) by replacing soft layers with sand, or by dewatering (bemaling).
Potential uplift of the entire structure should be checked as well for this purpose. If a shallow foundation
would not suffice, a deep foundation is required.

In addition, hydraulic structures have to be stable and and not get into motion due to loads. This implies
that there has to be a horizontal, vertical and rotational stability:
ΣHtotal = 0
ΣVtotal = 0
ΣMtotal = 0,
where
ΣHtotal [kN] = total of the horizontal components of the soliciting and resisting forces
ΣVtotal [kN] = total of the vertical components of the soliciting and resisting forces
ΣMtotal [kNm] = total of the moment caused by the soliciting and resisting forces

Another threat to the stability is groundwater flow under the structure, which can lead to internal erosion
or undermining of the structure if sediment particles are moved away with the flow. External erosion, i.e.,
erosion in front of a structure, can also lead to collapse. These five criteria are elaborated in this chapter.

37.1 Horizontal stability

The total of the horizontal forces acting on a hydraulic structure based on a shallow foundation will be
transferred to the subsoil (Figure 37-1). The friction force of the subsoil should resist the resulting total
acting horizontal force, otherwise it will slide aside. This friction force is determined by the total of the
forces soliciting (acting) on the structure in vertical direction (or the total of the vertical components of the
forces, ΣV), multiplied by a dimensionless friction coefficient f. The friction force should not be less than
the total of acting horizontal forces (ΣH), to prevent the structure from sliding aside. In equation form:

ΣH < f ⋅ ΣV (37.1)

The self-weight of the structure and buoyant forces should be included in ΣV.

Figure 37-1 Slip-off principle sketch

265
Manual Hydraulic Structures Part IV 'Structures'

Depending on the failure mechanism, more specifically the position and alignment of the slip surface, a
certain type of friction develops. Potential critical slip planes are:
1. friction between structure and subsoil: f = tan(δ), with δ = friction angle between structure and
subsoil. If δ is unknown, it can be approximated: δ ≈ ⅔ φ (φ is angle of internal friction of the sub-
soil). The friction coefficient for concrete-rubble is about 0,5; in case of a ribble bed: 0,6 (Dutch
Deltaworks). Values for cast concrete on soil are given in Table 37-1.
2. Internal friction of the subsoil: f = tan(φ), with φ is the angle of internal friction of the subsoil.
3. A deeper soil layer with a low sliding resistance.

Table 37-1 Friction coefficient for cast concrete on soil [USACE Technical Letters]

37.2 Rotational stability

Between the soil and the structure only compression stresses can develop. For the stability of shallow
founded structures, tensile stresses should not be taken into account for force equilibrium. Especially the
adhesive and cohesive properties of sand are very poor, so in general, tensile stress cannot be provided
by the subsoil. Therefore, it is usually stipulated that the soil stresses necessary for rotational stability
may only be compressive (so, no tensile stresses allowed). This is the case if the resulting action force
intersects the core of the structure. The core is defined as the area extending to 1/6 of the structure width
on both sides of the middle of the structure (point K). In equation form:

ΣM 1
eR = ≤ b (37.2)
ΣV 6

where: eR [m]
= distance from the middle of the structure (K) to the intersection point of the
resulting force and the bottom line of the structure
ΣV [N] = total of the acting vertical forces (or vertical components)
ΣM [Nm] = total of the acting moments, preferably around point K, halfway the width
b [m] = width of the structure

See Figure 37-2 for an illustration.

Figure 37-2 The action line of the resulting force should intersect the core of the structure
The rotation depicted on the left side is very unusual. In reality, if there is a maximum load acting in just one point (the corner),
the soil under the corner will immediately collapse. The adjoining soil will then start bearing the load.

266
Manual Hydraulic Structures Part IV 'Structures'

If, however, the action line of the resulting force is located outside the core of the structure, only part of
the soil below the structure will contribute to the bearing, see Figure 37-3. In that case, the no-tensile cri-
terion is not met. It actually can occur that even in this case the maximum occurring soil stress is not ex-
ceeded and the situation is not really problematic.

Figure 37-3 Action line of resultant force does not intersect te core of the structure

37.3 Vertical stability

The vertical effective soil stress, required to resist the acting loads (σk.max), should not exceed the maxi-
mum bearing capacity of the soil (p'max), otherwise the soil will collapse:

σ k ,max < p 'max (37.3)

The maximum acting load on the soil can be calculated with:

σ k ,max =
F M
+ =
∑V + ∑ M (37.4)
A W b⋅ℓ 1 ℓb 2
6

where: ΣV [N] = total of the acting vertical forces (or vertical components)
2
A [m ] = area of the bottom plate
3
W [m ] = section modulus of the contact area of the bottom plate
b [m] = width of the structural element
ℓ [m] = length of the structural element
ΣM [kNm] = total of the acting moments, preferably around point K, halfway the width

The bearing capacity p'max can be calculated according the Brinch Hansen method for determining the
maximum bearing capacity of a foundation. This method takes into account the influence of cohesion,
surcharge including soil coverage and capacity of the soil below the foundation, see Section 32.2 of this
Manual (Part III).

As a rule of thumb (instead of the Brinch Hansen calculation), the bearing capacity of densely packed
2 2
sand is often assumed to be 400 kN/m (= 0,4 N/mm ), in accordance with the Dutch Handboek Funder-
ingen (CUR, 2010).

The minimum acting load on the soil can be calculated with:

σ k ,min =
F M
− =
∑V − ∑ M (37.5)
A W b⋅ℓ 1 ℓb 2
6

Because soil cannot or barely cope with tensile forces,

σ k ,min > 0 (37.6)

267
Manual Hydraulic Structures Part IV 'Structures'

37.4 Piping (internal backward erosion)

Theory
A difference in water level (the 'hydraulic gradient') across a hydraulic structure or dike can cause a flow
of groundwater in permeable soil under or besides that structure. The phenomenon of groundwater pour-
ing out of the soil is called 'seepage'. The consequences of seepage are usual not severe, because the
quantity of seepage water mostly does not lead to a flood, but only to an amount of excess water that can
usually be dealt with by the lower water system. The flow under the structure can best be described by a
flow net as in Figure 37-4 (left), where a piping screen (= 'seepage screen'), usually a sheetpile wall, has
been installed under the structure.

If the flow velocity is sufficiently high, sediment particles will start extruding from permeable soil at the
land side of the structure. This internal erosion leads to the formation of cavities or channels in the sub-
soil: the so-called pipes. Little sand boils (zandmeevoerende wellen) appear where the water flows out at
ground level. The pipes can grow until they reach the water side of the dike or structure, undermining its
foundation, which can lead to collapse or sliding of the dike body. Pipes need a so-called roof (a cohesive
layer or a hard structure) that prevents immediate collapse of the pipes. This effect causes the piping
paths to follow the geometry of the structure, so they are usually orientated horizontally or vertically and
do not follow the lines of the original flow net, see Figure 37-4 (right).

Figure 37-4 Piping: Initial flow pattern under the structure (left) and final piping path (right)

There are two types of internal backward erosion:

• Piping (onderloopsheid) is the formation of erosion pipes under a structure.
• Outflanking (achterloopsheid) is the formation of erosion pipes around (besides) a structure.

It is important to realise that two basic conditions have to be met before piping can occur:
1. The duration of the water level difference has to be sufficiently long to complete the process of pipe
formation. This is the reason why piping under normal Dutch sea dikes has not yet been observed:
water levels along the coast highly depend on tidal fluctuations and the duration of a storm is usually
not longer than approximately 24 hours with a peak of about 2 hours. However, for hard structures
along the coast, piping measures are usually included in the design.
2. Sand particles must have a possibility to extrude. So, if the ground level at the side of the lower water
level is relatively high, piping becomes unlikely, just like in case of a closed and sufficiently strong sur-
face, like an underwater concrete floor in a building pit. Before erosion can start, rupture of the top
layer should occur. In case of low-permeable cohesive soil or a filter layer this phenomenon is called
uplift; for sand (aquifer material) the term heave is used.

Piping under pile-founded structures is very critical, because the subsoil can easily settle a bit while the
pile-founded structure remains at its place (because of its foundation in a deep, stable ground layer). This
results in an open space under the structure, where the water can flow without significant hindrance and
very likely cause internal erosion.

268
Manual Hydraulic Structures Part IV 'Structures'

Design tools
Empirical formulas based on research describe the critical situations in which piping can occur. The most
famous are the Bligh and Lane formulas. According to these formulas there is a limit state with a critical
ratio between the differential head and the seepage distance. More recent research has confirmed this.
For the results of this research and the design rules that have been derived from it, one is referred to
Sellmeijer, 1988 (only for standard dikes, not for structures with seepage screens). It should be realised
that the Bligh and Lane formulas are inaccurate and that they should only be used for a first impression
about possible piping problems.

W.G. Bligh was among the first to find a formula for a safe seepage distance. He studied several struc-
tures like weirs, dams and barrages, mainly in the India. Based on his observations, he developed a 'criti-
cal average gradient model' which determines the critical ratio between head difference and seepage
length. Bligh mainly assumed horizontal seepage paths.

E.W. Lane extended Bligh’s theory based on an investigation of more than 200 masonry dams. Lane dis-
tinguished between horizontal and vertical seepage paths. He found that vertical paths offer a three times
higher resistance against piping than horizontal paths, mainly due to the fact that vertical pipes tend to
collapse due to gravity, thereby blocking the pipe. In Lane’s formula pipes with a slope angle of 45° or
larger are treated as vertical, slopes at less than 45° are considered horizontal.

A sketch with horizontal and vertical seepage distances under a structure is shown in Figure 37-5.

Figure 37-5 Horizontal and vertical seepage paths

Bligh assumes L =∑ L vert + ∑ Lhor and Lane assumes L =∑ L vert + ∑ 31 Lhor . On ground of Lane’s assumption,
the soil constants also undergo changes compared to Bligh's constants (see Table 37-2). It should be
noticed that for pile-founded structures, Lhor should not be included in the total seepage length L, unless it
is certain that the soil volume below the structure cannot settle (as it would lead to interspace between
the soil and the structure).

Piping method: Bligh Lane

criterion L ≥ γ ⋅ CB ⋅ ∆H L ≥ γ ⋅ CL ⋅ ∆H
used seepage length L =∑ L vert + ∑ Lhor L =∑ L vert + ∑ 31 Lhor
Soil type CB imax CL imax
very fine sand / silt /sludge 18 5,6 % 8,5 11,8 %
fine sand 15 6,7 % 7,0 14,3 %
middle fine sand - - 6,0 16,7 %
coarse sand 12 8,3 % 5,0 20,0 %
gravel 4-9 11,1 – 20,0 % 4,0 25,0 %
clay 3 2,0
stiff clay - 1,8
Table 37-2 Safe seepage distance for piping (partly derived from Calle & Weijers, 1994)

269
Manual Hydraulic Structures Part IV 'Structures'

where: L [m] = total seepage distance, which is the distance through the soil where the water
1
flow is impeded by the soil structure
CB [-] = Bligh’s constant, depends on soil type
CL [-] = Lane’s constant, depends on soil type
∆H [m] = differential head across structure
γ [-] = safety factor (1,5)
imax [-] = maximum (allowed) hydraulic gradient = ∆H / L

In the Dutch design practice, both methods are being applied. For a first check on piping under dikes,
Bligh's method is most suitable, whereas both Bligh's and Lanes' method should be used to estimate
whether piping will occur under water retaining structures.

Hans Sellmeijer more recently (1988) developed a mathematical model to describe piping. The design
rules resulting from this model lead to more favourable dimensions for the required horizontal piping line,
compared to Bligh's method. Sellmeijers original model, however, is not applicable for vertical piping lines
and is only suitable for the dimensioning of dikes. That is why he formulated additional design rules for
the dimensioning of heave (hydraulische grondbreuk) behind seepage screens. These new design rules
can be used if the piping criterion of Lane doesn't suffice.

According to Sellmeijer a piping channel becomes critical if its length exceeds half of the seepage length.
If this occurs, the piping channel will progressively increase until failure. The observation of sand boils
therefore does not immediately imply failure, but it is not very well possible to determine how safe the
structure then still is.

The critical water level difference can be computed with an empirical equation (Sellmeijer, 2011):

∆H = F1 F2 F3 L > 0
γ 
F1 = η  s − 1 tan θ
 γw 
0,4
d70m  d70 
F2 =   (37.7)
3
ν kL  d70m 
g
0,28
+ 0,04
( D / L ) 2,8 −1
F3 = 0,91⋅ ( D / L )

where:
L [m] = (horizontal) seepage length
3 3
γs [kN/m ] = specific weight of sand particles (for sand 26,5 kN/m )
3
γw [kN/m ] = specific weight of water
θR [°] = bedding angle
D [m] = thickness of the aquifer (sand layer)
η [-] = drag factor coefficient
d70 [m] = 70% fractile of the grain size distribution
d70m [m] = reference value of d70
2
ν [m /s] = kinematic viscosity
k [m/s] = specific conductivity
2
g [m /s] = gravitational acceleration

This method can only be used if sufficient specific soil parameters are known.

Effect of piping screens on water pressure under a structure

The water pressure in front and behind a hydraulic structure depends on the water depth on both sides. It
linearly changes along the length of the seepage path directly under the structure, see Figure 37-6a. The
same applies, if there is a piping screen under the structure. The presence of the screen causes a discon-
tinuity in the upward pressure under the structure at the location of the screen, see Figure 37-6b.

1
The horizontal distance under pile-founded structures should not be included in the horizontal seepage distance Lhor.

270
Manual Hydraulic Structures Part IV 'Structures'

The piping screen is preferably placed at the side of the structure with the highest water level, as it then
maximally reduces the vertical pressures, thus reducing the uplift force.

Figure 37-6 Water pressure under a structure a) without a piping screen and b) including a piping screen

Measures to prevent piping

If a structure does not meet requirements for piping, the following solutions could be considered:
1. apply a piping berm (consisting of clay) at the inner side of a dike
2. use (longer) sheet piling upstream as a screen against seepage
3. apply grout columns upstream to make the soil impermeable and cohesive
4. insert a diagonal protective textile in the soil in front of a structure
5. insert a filter structure downstream
6. install gravel piles (grindpalen) to lower hydraulic head at the inner side of a dike
7. (temporarily) increase the downstream water level (for instance in the ditch behind the dike)
8. apply a clay core (kleikist) in the aquifer as an impermeable blocking system
9. as an emergency measure: sandbagging around a sand boil (opkisten)

Notes
1. The full and one third part of respectively the vertical and horizontal piping paths may only be taken into ac-
count if the following conditions are satisfied:
• The building material must be in direct contact with undisturbed soil
• The walls must be closed and must have a water-tight connection with the rest of the structure
2. The first condition means that if a structure has a pile foundation, which may allow settlements that create a
split between the structure and the soil, the seepage path along the bottom slab may not be counted, unless ex-
tra measures are taken to seal the split.
3. One should not only consider under-seepage, but must also take backward seepage into account. Cut-off walls
to prevent this are obviously placed in the same plane as the normal cut-off walls to prevent under-seepage.
Furthermore, these walls are extended sideways beyond the loose ground of the building site.
4. It is very important to compact the soil around the structure so it erodes less easily.

271
Manual Hydraulic Structures Part IV 'Structures'

37.5 Scour protection

The presence of a hydraulic structure along or in the water usually affects the flow current and the waves.
This leads to a change of the motion of water particles near that structure, which can lead to scour of the
bed in front of it. A gradual dislocation of a sill or geotechnical instability of the structure can be the result.
Scour can be prevented by applying granular filters, geotextile, or and more or less impermeable layers
2
like asphalt . This section first explains when scour can be expected, followed by the basic design princi-
ples of (geometrically closed) granular filters. More information on bed protection can be found in the lec-
ture notes on 'Bed, bank and shore protection' (Schiereck & Verhagen, 2016).

Scour of sand beds, caused by waves

The interaction of waves with a structure can lead to a scour hole that could threaten the stability of the
structure (Figure 37-7).

Figure 37-7 Bottom scour induced by waves in front of a breakwater

If no information is available, only for preliminary designs, the maximum wave-induced scouring depth
hmax in front of structures with vertical walls may be assumed to be equal to the height of the maximum
unbroken wave, which cannot become more than 0,5 to 0,7 times the original water depth h0, which could
occur under the worst conditions (Shore Protection Manual, 1984):

hmax ≈ Hmax < 0,7·h0 (37.8)

Xie (1981) proposed a simple equation for scour near vertical or steeply sloping impermeable structures,
related to the wave height, water depth and wave length:

−1,35
  2 ⋅ π ⋅ h0 
h max = 0,4 ⋅ H  sinh    (37.9)

  L wave 

The equation is based on experiments on both fine and course sediments with regular waves.

Scour of sand beds, caused by current

Bed protection prevents structural damage by water currents near hydraulic structures. Turbulence can
cause scour holes at the end of the scour protection, if the amounts of sediment transported over the pro-
tection are small (Figure 37-8).

Figure 37-8 Scour due to current

2
To prevent scour at both sides of a closure dam, much research on the required length of a bed protection has been carried out during the
execution of the Deltaworks in the Netherlands (Huis in ‘t Veld, 1987, § 2.4.9).

272
Manual Hydraulic Structures Part IV 'Structures'

The angle of the upper slope of the scour hole, β, is usually considerably less steep than the natural slope
of sediment under water. Usual values for β are between 18° and 26°.

The maximum depth of the scour hole at a given moment can be calculated with the formula of Herman
Breusers of Deltares (1991), but the method requires much information, which is difficult to obtain. A sim-
plification can be achieved by calculating the equilibrium depth of the scour hole, while assuming that
there is no sand coming from upstream (clear water scour). In that case the maximum (= equilibrium)
scour depth hmax is given by:

0,5 ⋅ α ⋅ u − ucr
hmax = ⋅h0 for 0,5 ⋅ α ⋅ u − ucr > 0 (37.10)
ucr

where:
hmax [m] = maximum depth of the scour hole (= equilibrium depth)
h0 [m] = initial water depth
u [m/s] = depth-averaged flow velocity at the end of the scour protection
ucr [m/s] = critical velocity regarding begin of motion of sand particles
α [-] = turbulence coefficient, depending on the upstream disturbance.
The value of α is in the order of 3.

The critical velocity ucr for the initiation of sand transport under uniform flow can be calculated with the
Shields equation:

ucr = C Θcr ⋅ ∆ ⋅ D50 (37.11)

where:
D50 [m] = median nominal diameter of sand particles at the end of the sour protection
3
C [√m/s] = Chézy coefficient
ϴcr [-] = Shields (stability) parameter (= τ*), the dimensionless critical shear stress:
τ cr
Θcr = (37.12)
g ⋅ ∆ ⋅ D50
ρ s − ρw
∆ [-] = relative density: ∆ = (usually about 1,65)
ρw

The Chézy coefficient can be calculated with:

 12 ⋅ R 
C = 18 ⋅ log   (37.13)
 kr 
where
kr [m] = equivalent bed roughness:
kr ≈ 2 Dn50 for narrowly graded gravel and rock
kr ≈ 3 Dn90 ≈ 6 Dn50 for widely graded gravel and rock
kr ≈ 1 to 5 dm for sand, which is the height of the bed form
(ripples or flow channels)
R [m] = hydraulic radius of the flow channel at the end of the scour protection:
R = wet area / wet circumference of the flow channel
R ≈ h0 for wide channels

∆⋅g
The Shields parameter depends on the dimensionless grain diameter: d∗ = D50 ⋅ 3 , where ν is the
ν 2

2
kinematic viscosity [m /s]. Figure 37-9 shows the relation between the Shields parameter ϴcr and the di-
mensionless grain diameter d* (lower horizontal axis). For normal circumstances (temperature, density),
the value of ϴcr can be directly related to D50 (upper horizontal axis).

3
If this equation is used to calculate the critical velocity for the initiation of rock, for example to determine the stability of a bed protection, Dn50
should be used in stead of D50.

273
Manual Hydraulic Structures Part IV 'Structures'

Figure 37-9 Relation between the Shields parameter and d* or D50 for determining the scour depth of sand beds
under usual conditions (after Schiereck & Verhagen, 2016)

Scour caused by a plunging jet

Scour caused by plunging, free falling jets can be a relevant failure mechanism for the gates of locks and
cut-offs (coupures) (Figure 37-10). The jet that hits the water surface with an impulse that is proportional
to the product of the discharge and the velocity of the jet. The jet propagates into the lower water body
and will reach the bed with a reduced velocity (and impulse). There it can leade to erosion of the bed par-
ticles or rock. Eddies originate in front of and behind the jet and circulate the eroded particles, part of
which is consequently transported by the current in downstream direction.

Figure 37-10 Scour caused by a plunging jet (Achtergrondrapport toetsspoor hoogte, 2004)

The maximum scour depth due to plunging jets according to D'Agostino and Ferro (2014) can be calcu-
lated with:

− 0,3
hmax = 0,4 ⋅ q 0,6 ⋅ H 0,4 ⋅ D50 − 0,5 ⋅ h 0 (37.14)
where:
hmax [m] = maximum scour depth
3
q [m /s/m] = specific discharge per m width
H [m] = water level difference
h [m] = water depth downstream
D50 [m] = median nominal particle diameter

Scour behind piles and piers, caused by waves and currents

The scouring process around vertical piles and piers is mainly caused by:
• local disturbance of the flow field (local scour)
• local reduction of cross-section (contraction scour)

274
Manual Hydraulic Structures Part IV 'Structures'

Other general relevant scour effects are:

• general degradation effects (downstream of weirs, reservoir dams, etc)
• bend scour; deeper part of cross-section in outer bend area (variability in river planform)
• confluence scour; deeper parts of cross-section downstream of confluence
• thalweg variations (deepest point of coss-section may shift in lateral direction)
• bed-form variations

The flow pattern around a cylindrical pipe is shown in Figure 37-11.

Figure 37-11 Scour pattern behind a pile or pier (Melville, 1988)

Several researchers have studied the scour develeopment behind cilindrical piers and piles. Based on
their data, Van Rijn (2013) proposes the following equation for the scour hole depth:

0,2
 Dpile 
h max = 0,35 ⋅ Dpile   ⋅ Vpar (37.15)
 h0 
where:
hmax [m] = scour hole depth
Dpile [m] = pile diameter
h0 [m] = water depth
u c2 + (0,7 ⋅ uw ) 2
Vpar [-] = dimensionless velocity parameter: Vpar = (37.16)
u cr
uc [m/s] = near-bed peak orbital velocity due to storm
uw [m/s] = maximum flow velocity upstream due to tide and wind
ucr [m/s] = critical velocity for initiation of motion:
ucr = 0,35 m/s for 100-300 µm; 0,4 m/s for 400 µm and 0,5 m/s for 600 µm

Scour of cohesive beds, caused by current

Cohesive forces cause a distinct increase of the soil against erosion compared to homogeneous sand
beds. The amount of erosion mainly depends on the type of fine minerals, biological activity at the bed
and the rate of consolidation of the soil. A sand bed with small amounts of silts and clays already shows a
considerably increased resistance against erosion and old compact clay soils are even highly resistant
against erosion. There are two main ways of clay erosion: erosion of (many) single particles and chunk
erosion (kluiterosie).

Based on tests performed at Delft Hydraulics (1989) on bed samples from the North Sea with particle siz-
es of 100 µm < D50 < 200 µm, and mud-silt percentages of 2 to 20%, the critical bed shear stress τcr,c was
related to the critical bed shear stress for sand according to Shields, τcr,sand, and the percentage of fines
(mud or silt) smaller than 50 µm, ps (Van Rijn, 1993, §4.1.2):

τcr ,c = ps0,5 ⋅ τcr ,sand (37.17)

Similarly, it can be stated that

Θcr ,c = ps0,5 ⋅ Θcr (37.18)

275
Manual Hydraulic Structures Part IV 'Structures'

where ϴcr is the critical Shields parameter of sandy beds as indicated in Figure 37-9. Be aware that this
equation only applies to homogeneous soil layers and not to chunk-type erosion.

Thus, the procedure to find out whether a mixed sand-clay bed will scour is to determine the percentage
of fines (ps), then calculate the Shields parameter for this clayy bed (ϴcr,c), draw a new line in the graph of
Figure 37-9, and determine the ϴcr,c for the location specific sediment particle size D50. The critical veloci-
ty can then be calculated with Equation (37.11), with which the maximum scour depth (Equation (37.10))
and the required length of the scour protection can be found (Equation (37.25)).

For chunk-type erosion of cohesive beds, still no design formulas are available.

Bed protection
The sediment particles of the bed protection should be sufficiently large to avoid erosion. The D50 of these
particles can be estimated with help of the adapted Shields Equation (37.19) for uniform flow:

u2 K 2v
D 50 > ⋅ (37.19)
Θcr ⋅ ∆ ⋅ C 2 K s
where:
u [m/s] = depth-averaged flow velocity on the scour protection
ϴcr [-] = Shields (stability) parameter for no motion: ϴcr = 0,03 (so, don't use Figure 37-9!)
ρ − ρw
∆ [-] = relative density: ∆ = s (usually ∆ ≈ 1,65)
ρw
C [√m/s] = Chézy coefficient of the top filter layer
Kv [-] = factor for non-uniform (deceleration) flow = 1 to 2 (≈ 1,5)
sin2 α
Ks [-] = reduction factor for slope: K s = 1− (37.20)
sin2 ϕ
where α = the slope angle and φ = the angle of internal friction.

Extrusion of particles of underlying layers through the pores of the bed protection should be prevented.
This can be accomplished by granular filters, or with help of geotextile (Figure 37-12).

Figure 37-12 Definition filter structures

Granular filters can be divided in geometrically open and closed filters. Geometrically open filters allow
loss of filter material (sediment particles) if the load becomes higher than a critical value. To establish a
stable open filter, a critical filter gradient has to be determined, which requires detailed information on the
loading gradients. In geometrically closed filters, the grain sizes are chosen in such a way, that they can-
not move in the filter. Classical closed-filter rules have already been developed by Karl von Terzaghi in
the 1930s.

In general, there are three types of geometrically closed filters, viz.:

• granular filter
• asphalt concrete
• geotextile

The granular filter type is the most commonly used and in the following paragraphs the design of geomet-
rically closed filters is explained.

276
Manual Hydraulic Structures Part IV 'Structures'

A geometrically closed granular filter should be designed in such a way, that grains in the basic layer
cannot pass the holes of the filter. The holes in the granular filter consist of the pores in the grain packet
that are interconnected by small pore channels. If the diameter of the pore channels Dc is smaller than the
diameter of the governing grains of the basic layer Db, no transport can take place, irrespective of the val-
4
ue of the water level slope, the direction and type (stationary or not) of the flow .

The governing grain size of the base layer (the lower layer) is about D85. If grains with the governing di-
ameter cannot be transported, smaller grains can be transported neither. A layer is therefore considered
geometrically closed, if

Df 15
<5 (37.21)
Db 85

This only applies if the sieve curve of the basic material does not differ too much from the curve of the
filter material. If a layer is widely graded, internal instability can occur. In that case, the small grains can
be transported through the channels of the big ones. Internal stability can be expected, if

D60
< 10 (37.22)
D10

which should also be the case for a filter layer.

To avoid overpressure perpendicular to the separating layer, the filter layer should be more permeable for
water than the basic layer. In general, smaller grain diameters imply smaller permeability. In graded lay-
ers principally small grains of about D15 determine the permeability. During the lifetime of the structure,
the diameters of the pore channels can decrease because of siltation with material from the basic layer,
or because of deterioration (verwering). It is difficult to check up on this effect and eventual repairs are
practically nearly impossible. Therefore, an extra requirement for permeability should be met:

Df 15
>5 (37.23)
Db15

For geotextile filters, similar rules apply. A geotextile is considered to be tight, if the larger fraction of the
grains of the basic layer cannot pass the characteristic pores of the geotextile:

O90 < Db90 , which is very strict, and usually: O90 < 2Db90

where O90 = the diameter of the fraction of which 90 weight-percentage remains on the geotextile after 10
minutes of sieving (so it is not an indication for the pore size distribution!)

If the basic material has a wide gradation (Db60/Db10 > 10), Db90 in the equation can be replaced with Db50.

To avoid overpressure in vertical direction, the filter should be more permeable than the basic material. In
this case the permeability coefficient of the geotextile should be more than the permeability coefficient of
the basic layer:

kng > γ · kb (37.24)

where:
kng [m/s] = permeability coefficient of the geotextile
γ [-] = safety coefficient, dependent on the composition of the basic material:
γ = 2,5 for a uniform grain distribution,
γ = 10 for a highly graded material (D60/D10 > 10)
kb [m/s] = permeability coefficient of the basic layer

The permeability can be determined with help of a straightforward test, using a glass tube filled with soil,
connected with two reservoirs of water - see the book on 'Soil Mechanics' by Verruijt/Van Baars, Section
8.1.

4
Dxx is the size of the holes in a sieve, through which passes xx percent by weight of the material, see Dutch standard NEN 2560.

277
Manual Hydraulic Structures Part IV 'Structures'

The required length of the bed protection is determined by the depth and the slope of the scour hole hmax
near the structure (Figure 37-13).

Figure 37-13 The required length of a bed protection near a structure

If the scour hole would become too deep, a critical slope 1:ns will be exceeded, in which case sliding is
probable to occur. For normal packed sand this will result in normal sliding, but for loosely packed sand
this may lead to liquefaction (zettingsvloeiing). The values of ns can vary very much, but for tentative de-
signs, ns ≈ 6 can be assumed for densely packed, or cohesive material, and ns ≈ 15 for loosely packed
material.

The length of the scour protection should be such, that the sliding plane does not reach the structure, to
prevent its undermining. For a first estimate, the required length can be calculated with:

L ≥ γ · ns · hmax (37.25)
where:
γ [-] = safety factor (> 1,0)
1:ns [-] = average slope of the slide plane
hmax [m] = maximum scour depth

37.6 References
Bligh, W.G. (1915). Submerged weirs founded on sand. In: Dams and weirs.
Breusers, H.N.C. and Raudkivi, A.J. (1991). Scouring. Balkema, Rotterdam.
Calle, E.O.F. and Weijers, J.B. (1994) Technisch rapport voor controle op het mechanisme van piping bij
rivierdijken. Grondmechanica Delft.
Hoffmans, G.J.C.M. and Verheij, H.J. (1997). Scour manual. Balkema, Rotterdam.
Huis in ‘t Veld, J.C. (1987). Closure of tidal basins, Delft University press, ISBN 90-6275-287-X.
Kanning, W. (2013) The weakest link. Spatial variability in the piping failure mechanism of dikes. PhD-
thesis Delft University of Technology.
Lane, E.W. (1935). Security from under-seepage. Masonry dams on earth foundations. Transactions
American Society of Civil Engineers, no. 100.
Rijn, L.C. van (1993). Principles of sediment transport in rivers, estuaries and coastal seas. Part 1. Aqua
publications, Amsterdam.
Schiereck, G.J. & Verhagen, H.J. (2016). Introduction to Bed, Bank and Shore Protection, VSSD, ISBN
ISBN 978-90-407-1683-6.
Schweckendiek, T. (2014). On reducing piping uncertainties. A Bayesian decision approach. PhD-thesis
Delft University of Technology.
Sellmeijer, J.B. (1988) On the mechanism of piping under impervious structures. PhD-thesis Delft Univer-
sity of Technology.
Weijers, J.B.A. & Sellmeijer, J.B. (1993). A new model to deal with the piping mechanism. In: Filters in
Geotechnical and Hydraulic Engineering.

278
Manual Hydraulic Structures Part IV 'Structures'

38. Pile foundations: Compression piles

Section on piles types added in 2019

38.1 Pile types
Before considering pile foundations, it should be checked whether a shallow foundation would be feasible
regarding:
• bearing capacity of the soil to resist the self-weight of the structure and its loads (Chapter 32)
• settlements of the subsoil under the structure (Chapter 34)
• the stability of the shallow-founded structure (Chapter 37)

A shallow foundation is generally (but not always!) easier to construct and less expensive than a pile
foundation. If a shallow foundation does not have sufficient bearing capacity, soil improvement can be
considered in the form of a replacement of compressive layers by sand. However, if the soil improvement
reaches to a level under the groundwater table (1 m or more), it generally becomes too expensive. In
cases where a shallow foundation is not feasible, piles should be used (or in a few special cases, pneu-
matic caissons).

An overview of pile types is presented in Figure 38-1. All the displayed pile types are soil displacing
(grondverdringend), except for the screwed soil replacing piles (grondverwijderend).

Figure 38-1 Overview of foundation pile types. Only the most common types are included.

There are different installation methods that all have their advantages and disadvantages: driving/drilling
(heien), pulsing (drukken), boring (boren), screwing (schroeven), vibrating (trillen). The choice for installa-
tion method and pile type is usually made by the contractor. There are too many pile types to make a
simple decision tree. If possible, the least expensive and simplest to install prefab & pre-stressed con-
crete piles shall be used (these piles are also suitable for tension). However, if no noise or no vibrations
are allowed, other pile types should be considered. This is also the case if the cone resistance of the soil
is too high for proper driving (especially in gravel layers). The choice for a suitable pile type mainly de-
pends on the following considerations.

Consideration 1: Disturbance of the surrounding soil

A first consideration for selecting pile types is disturbance for the surrounding soil. The installation of piles
can lead to undesired or not permitted shock waves and vibrations in the soil. Within a radius of 8D (D =
pile diameter), the installation of soil displacing piles can lead to excess water pressures (wateroverspan-
ning) that can temporarily decrease the bearing capacity of the existing piles. All types of screwed piles
result in a release of water pressures (ontspanning) within a radius of 4D to 6D. To not influence the bear-
ing capacity of existing piles, the influence sphere of screwed piles should not intersect the soil cylinders
around the existing piles (radius 2D) that contribute to their bearing capacity.

Consideration 2: Disturbance of surrounding buildings and structures

The second consideration usually is whether driving (heien) of the piles is permitted, because driven piles
have several advantages. Driven piles can be selected if the vibrations or shock waves (under the ground
surface) and noise (through the air, above the ground surface) caused by the pile driving are acceptable
to the surroundings.

279
Manual Hydraulic Structures Part IV 'Structures'

Several of the main advantages of driven piles are:

• the blows of driven piles can be counted ('kalenderen') to verify whether the soil layer composition is
according to the design conditions
• the soil around the piles is displaced (verdrongen), leading to an increase of the bearing capacity
• the pile shaft is relatively insensitive to lateral soil pressures and insensitive for groundwater flow.

Driving piles at a distance of more than 2 m from adjacent buildings or structures usually does not lead to
structural damage, but vibrations can cause nuisance and (minor) cracks. If vibrations cannot cause big
problems, prefab piles or in situ created piles can be used. If vibrations are expected to cause problems,
other pile types should be chosen, like screwed piles, Fundex and Tubex piles, auger piles (avegaarpa-
len), screw-injection piles, etc.
A noise sleeve (geluidsmantel) can be used to reduce noise levels with max. 5 dB(A).

Consideration 3: Construction site and transport

The available space, work height, accessibility and obstacles can influence the suitability of certain pile
types. For instance, if there is a limited work height, only small equipment can be used, which often leads
to a choice for steel tube piles or segment piles. (Horizontal) transport of prefabricated piles restricts their
slenderness (= length-diameter ratio) to 80 (like driveability requirements). Because of accessibility rea-
sons for transport over road, the length of prefabricated piles in inner cities can usually not be more than
25 m.

Figure 38-2 Indication of available pile lengths (consult the information of manufacturers for exact data)

Consideration 4: Pile installation equipment

The length of all driven piles is restricted by the piling rig (heistelling), more specifically by the length of
the guiding system (makelaar). The maximum length of a guiding system is about 40 m, of which 5 m is
needed for the driving block (heiblok), which restricts the length of driven piles to 35 m. Furthermore, not
all equipment is suitable for installing piles under inclinations and the force of driving hammers should not
damage the pile heads. A universal piling rig for prefab piles is shown in Figure 38-3 to give an impres-
sion of its dimensions. The driveability (heibaarheid) of prefabricated piles poses further restrictions to the
length of the piles: usually, the slenderness (= length-diameter ratio) should not be larger than 80. An
overview of possible lengths per pile type is shown in Figure 38-2.

Consideration 5: Type and magnitude of loading

The suitability for resisting compression forces, tensile forces or both varies per pile type. This also ap-
plies to the resistance against bending moments. An overview of all load conditions during the life time of
a structure can be helpful to determine which of these load types can occur. If there are only tensile forc-
es, anchor piles can be considered. Several other piles that are suitable for tension are: prefabricated
concrete piles, Tubex piles, screwed grout piles and bored piles.

The magnitude of the pile load not only depends on the self-weight and external loads of the superstruc-
ture, but also on the distance between the piles (determined by the pile grid). The pile load should be re-
sisted by the surrounding soil (bearing capacity) and by the material properties of the pile itself (strength).
A balance should be found between creating sufficient bearing capacity (by increasing the length of the
piles to achieve sufficient tip resistance or increasing the diametre of the piles to achieve sufficient shaft
resistance) and reducing the load per pile (by reducing the centre-to-centre distance between the piles).

280
Manual Hydraulic Structures Part IV 'Structures'

Figure 38-3 A universal piling rig for prefab concrete piles (Handbuch der Ramm- Vibrations- und Einpresstechnik, 2012)

The required strength of the piles is determined by the magnitude of its loading, see Figure 38-4 for the
strengths of a few common pile types.

Figure 38-4 Indicative design values for the pile strength (consult the information of manufacturers for exact data)
Notice that the soil bearing capacity is governing in most cases

Consideration 6: Soil composition

The soil composition determines the magnitude of the tip resistance and the shaft friction. In addition, the
driveability (heibaarheid) is highly dependent on the soil type. For a pile foundation on a gravel layer, the
driveabilty is a problem, next to the porosity which complicates grout injection. However, screwed injec-
tion piles (mortelschroefpalen) can be used in gravel. For gravel-sand layers, bored piles (with bentonite
injection) can be used with some precautions during construction.

281
Manual Hydraulic Structures Part IV 'Structures'

38.2 Bearing capacity of compression piles, general theory

The ultimate resistance of a compression pile is determined by the strength of the pile and the bearing
capacity of the pile-soil system. The strength of compression piles is determined by the characteristics of
the material and shape of the piles (mainly the compressive strength). The strength per pile type can be
obtained from the manufacturers of these piles. The strength of prefabricated pre-stressed concrete piles
depends on the concrete quality, steel reinforcement and length of the piles and varies from about 500 kN
to about 4000 kN. Critical loading can occur during transport and installation (during hoisting (inhijsen) or
driving (heien)), prior to loading by the completed structure.

The bearing capacity of compression piles is determined by both the tip bearing capacity and the shaft
bearing capacity:

Fr ;max = Fr ;max;tip + Fr ;max;shaft − Fr ;max;nk

where:
Fr ;max;tip = Atip ⋅ pr ;max;tip
and:
∆L
Fr;max;shaft = Op;avg ∫ pr;max;shaft dz
0
in which:
Fr;max [kN] = bearing capacity of the pile
Fr;max;tip [kN] = maximum tip resistance force
Fr;max;shaft [kN] = maximum shaft friction force
Fr;max;nk [kN] = maximum negative shaft friction force (only to be included in SLS checks!)
2
Atip [m ] = surface area of the tip of the pile
2
pr;max;tip [kN/m ] = maximum tip resistance according to a cone penetration test
2
Op;avg [m ] = average circumference of the pile shaft
∆L [m] = length of the pile
2
pr;max;shaft [kN/m ] = maximum pile shaft friction according to a cone penetration test

Notes
• In case prefab piles with enlarged feet are used, ∆L may not be larger than the height of the en-
larged foot H (see Figure 38-4).
• The maximum tip resistance should be limited to 15 MPa

38.3 Preliminary design: Pile resistance according to Prandtl and Meyerhof

If not cone penetration test results are known for the project location, for preliminary designs, the analyti-
cal methods of Prandtl and Meyerhof could be used to determine the pile resistance. Both Prandtl and
Meyerhof assume a slip plane with a certain shape next to the foot of the pile. To determine the shaft
bearing capacity one can simply total the maximum shear forces along the pile’s surface.

Prandtl
The maximum tip resistance can be determined analogously to the bearing capacity of a shallow founda-
tion, which is based on a slip plane according to Prandtl. This entails simply using the 2-dimensional solu-
tion for 3-dimensional collapse (conservative) and simply disregarding the shear strength, not the weight,
of the soil above the foundation plane (conservative).

Figure 38-1 Slip planes under a pile according to Prandtl

282
Manual Hydraulic Structures Part IV 'Structures'

In this case pr;max;tip can be approached using σ’max,d according to Part III, Section 32.2 "Soil, strength,
Prandtl and Brinch Hansen". For pile calculations, the term 0,5 γ e;d ′ Bef Nγ sγ is negligible relative to
′ Nqsq . Therefore, the tip resistance can most simply be calculated using:
σ v;z;o;d

′
pr;max;tip = σ v;z;o;d Nq sq

Meyerhof
Contrary to Prandtl, Meyerhof assumes that the slip plane developed below and beside the foundations,
continues to above the foundation level. According to Meyerhof, one can differentiate between three basic
cases, see Figure 38.3.

Figure 38-2 Basic cases according to Meyerhof

In his collapse analysis, besides normal stresses, Meyerhof also takes into account the shear stresses
that could develop along the top of the slip plane and along the vertical part of the foundation element.
This means Meyerhof's factors for the limit bearing capacity, especially in the case of high φ values and
relatively deep foundations are larger than Prandtl’s. This leads to a more favourable, more economical
design. The figure below shows the difference between Meyerhof and Prandtl for a foundation pile.

Figure 38-3 Slip planes according to Meyerhof (a) and Prandtl (b)

Figure 38-3 shows that according to Meyerhof, the ground properties above the foundation plane are also
of importance when slip planes occur.

Shaft bearing capacity

A simple method to estimate the shaft bearing capacity is to find the cumulative shear stress of all ground
layers along the pile:

Fshaft = Os ∫ (σ 'v ⋅ K ⋅ tan δ + c ' ) dz

In the case of soil displacing piles (prefab driving piles) the ground is tensed horizontally, which means:
K ≥ K 0 . In the case of drilled piles, one should assume K ≤ K 0 .

283
Manual Hydraulic Structures Part IV 'Structures'

38.4 Bearing capacity, method Koppejan

In the Netherlands, the bearing capacity of a pile is usually deter-
mined with soundings (Cone Penetration Tests). A sounding can be
perceived as a model test, loading a compression pile beyond the
state of collapse. Theoretically sounding entails determining the max-
imum tip resistance at every level in the ground. Because the size of
the slip plane below and beside the pile or sounding cone depends
on the dimensions of the foundation area, the cone resistance cannot
be directly interpreted as the maximum tip resistance of a pile. Gen-
erally one assumes the slip plane continues to some 0,7 times the
pile diameter below the pile and up to approximately 8 times the di-
ameter above the foundation (Figure 38-5).

The Koppejan method was adopted in the Dutch standard NEN 6743
and more recently in NEN 9997-1 that complies to the Eurocode on
Geotechnology.
Figure 38-5 Definition of slip planes
Maximum tip resistance
In this case the maximum tip resistance can be derived from the sounding using:

 qc;l;avg + qc;ll;avg 
pr;max;tip = 1 α p β s  + qc;lll;avg 
2 2
 
in which:
qc;l;avg [MPa] = the average value of the cone resistances qc;z;corr’ along section I, from the level of
the tip of the pile to a level that is 0,7 Deq to 4 Deq deeper. Deq is the equivalent pile
tip diameter (see below). The bottom of section I must be selected between these
boundaries, so that pr;max;tip is minimal;
qc;ll;avg [MPa] = the average value of the cone resistances qc;z;corr’ along section II, from the bottom
of section I to the level of the tip of the pile; the value of the cone resistance to be
used in this calculation may never exceed that of a lower level;
qc;lll;avg [MPa] = the average value of the cone resistances qc;z;corr’ along section III, which runs up
from the pile tip level to a level of 8 Deq above the pile tip; like for section II, the
value used for the cone resistance may never exceed that of a lower level, starting
with the value of cone resistance at the top of section II.
In the case of auger piles (avegaarpalen), this section must start with a cone
resistance equal to or less than 2 MPa, unless the sounding took place within 1
metre of the side of the pile after the pile was placed in the ground;
Deq [m] = equivalent pile tip diameter:
Round pile: Deq = D
4
Square/rectangular pile: Deq = a b/a
π
a = width of pile, shortest side
b = width of pile, longest side (with a ≤ b ≤ 1,5a )
αp [-] = pile class factor according to Table 38-1, according to NEN 9997-1 since 2017.
β [-] = factor that takes into account the influence of the shape of the foot of the pile,
determined using Figure 38-4 and Figure 38-5;
s [-] = factor that takes the shape of the cross-section of the foot of the pile into
accounted, see Figure 38-6.

284
Manual Hydraulic Structures Part IV 'Structures'

pile class / type αp

• soil displacing placement methods
- driven piles 0,70
- driven piles, formed in the soil 0,70
- screwed piles, formed in the ground 0,63
- prefabricated screwed piles 0,56
• piles with little soil displacement, such as steel profiles and 0,70
open steel tubes
• piles made with soil replacement
- auger piles 0,56
- drilled piles 0,35
- pulsated piles 0,35
Table 38-1 Values of the pile class factor αp

pile tip

Figure 38-4 Shape of the pile footing Figure 38-5 Factor for the shape of the pile footing β

Limit value is 15 MN/m2 for

all non-cohesive soil types

1. sand and gravel OCR ≤ 2

2. sand and gravel 2 < OCR ≤ 4
3. sand and gravel OCR > 4

OCR = degree of overconsolidation

Figure 38-6 Values of the pile footing shape factor s Figure 38-7 Value of the maximum tip resistance in sand and gravel

285
Manual Hydraulic Structures Part IV 'Structures'

Maximum pile shaft friction

The maximum pile shaft friction derived from the sounding is:

pr;max;shaft;z = α s qc;z;a

in which:
pr;max;shaft;z [MPa] = maximum pile shaft friction at depth z;
αs [-] = factor that takes into account the influence of the method of pile installation.
Table 38-2 applies for fine to coarse sand. An additional reduction factor of
0,75 applies for extremely coarse sand. For gravel this reduction factor
is 0,50. For clay, silt or peat, Table 38-3 should be used.
qc;z;a [MPa] = cone resistance, whereby the peaks in the sounding diagram with values
2
larger than 15 MN/m are removed if these values occur over sections of at
2
least 1 metre and otherwise at values of 12 MN/m .

Pile class / type αs

• ground displacing placement methods:
- driven smooth prefab concrete pile and steel tube pile with closed tip 0,010
- pile made in the soil, whereby the concrete column presses directly onto the 0,0014
ground and the tube was driven back out of the ground
- in the case of a tube removed by vibration 0,0012
- tapered wooden pile 0,0012
- screwed piles
with grout injection or grout mix 0,009
without grout 0,006
• piles with little ground displacement
- steel profiles 0,0075
• piles made with ground replacement
- auger piles 0,006
- drilled piles 0,006
- pulsated piles 0,005
Table 38-2 Maximum values of αs in sand and in sand containing gravel

ground type relative depth z/D αs

clay/silt : qc ≤ 1Mpa 5 < z/D < 20 0,025
clay /silt : qc ≥ 1Mpa z/D ≥ 20 0,055
clay /silt : qc > 1Mpa not applicable 0,035
peat not applicable 0
Table 38-3 Values of αs for clay, silt or peat
Notes
• The shaft friction factor is in fact the same as the friction value of a CPT-sounding test as mentioned in Part III,
Chapter 31 'Soil properties'. The only difference is that for αs one does not use an average, but for safety’s sake,
a lower limit value.
• In the case of peat one assumes no frictional strength because it is not certain that the shear stress doesn’t relax
to zero due to the creep of the peat.

Maximum negative shaft friction

The effective bearing capacity of the pile is not only determined by the shaft bearing capacity and the tip
bearing capacity, but also by the negative shaft friction (negatieve kleef), which influences the bearing
capacity unfavourably. Settling soil layers can hang onto the pile, causing a downward shaft friction force.
The negative shaft friction depends on the contact area of the shaft, the horizontal soil pressure and the
friction coefficient:

Fs,nk = Os ∑ h ⋅ σ 'v ,avg ⋅ K 0 ⋅ tan δ

where:
Fs,nk [N] = negative shaft friction
Os [m] = circumference of the pile
h [m] = height of the soil layer that 'hangs' on the pile
σ'v,avg [kPa] = average vertical effective soil stress, per soil layer
K0 [-] = neutral soil pressure coefficient
tanδ [-] = friction between soil and concrete

286
Manual Hydraulic Structures Part IV 'Structures'

Notes
• For groups of piles the following applies: The negative shaft friction per pile can never be larger than
the total weight of the settling ground divided by the number of piles. The weight of the weak ground
per pile is therefore an upper limit for the calculation of the negative shaft friction (γm;g = 1,0 !).
• In ULS, the pile will fail, meaning that it displaces downward. This implies that negative shaft friction
will be overruled and should therefore not be taken into account.

38.5 Stiffness
As a rule of thumb, calculation programmes often use a pile length of 2ℓ instead of ℓ. This provides a fairly
good estimate of the stiffness of the pile and the soil together.

It is more accurate to divide the pile settlement into two parts:

1. The shortening of the pile
2. The settlement of the tip of the pile

The calculation method for the pile settlement under a load has been normalised (see NEN 6743) and
needs to be calculated according to this norm for the final design. The principle of this norm is described
briefly below.

Pile shortening
The shortening of the pile ∆ℓ can be estimated using:

σ = Eε
with:
F ∆ℓ
σ= and ε =
A ℓ

in which:
F [kN] = load on the pile
ℓ [m] = representative pile length

The representative pile length is shorter than the true pile length because the load on the pile is already
partly absorbed by the shaft friction, which reduces the compression stress and therefore also the strain
in the pile. See the following table for the Youngs' moduli of several pile materials.

Material E –value
Concrete 20 ⋅ 106 kN/m2
Steel 200 ⋅ 106 kN/m2
Wood 15 ⋅ 106 kN/m2
Table 38-1 Youngs' moduli

Pile tip settlement

For a preliminary design one can assume a rough estimate of the pile tip settlement.
The settlement of the pile tip largely depends on the proportion between the true pile load and the maxi-
mum acceptable load on the pile. Due to the coefficients of safety, there is a ratio of about 1,5 between
these two. In this case the tip settlement of a ground displacing pile (prefab concrete) is:
w tip ≈ 2% to 3% ⋅ Deq
The tip settlement of drilled and auger piles is roughly twice as large because the ground has not been
compressed as for driven prefab concrete piles, but has even been made looser by the drilling. A tempo-
rary excavation (deep building site excavation or drill tunnel) near the pile tip will cause relaxation of the
soil, a drop of the value of qc and a serious increase of the pile tip settlement.

For a more accurate description, refer to: NEN 6743 "Berekeningsmethoden voor funderingen op palen,
drukpalen" (Calculation methods for foundations on piles, compression piles).

287
Manual Hydraulic Structures Part IV 'Structures'

39. Pile foundations: Tension piles

Major revision: February 2011

Equation Chapter 43 Section 43
In some cases, resulting loads will cause structures to lift, if the present piles are not able to bear the ten-
sion forces. For quay walls founded on piles and similar structures, a number of piles will normally be
subjected to tension forces where the soil pressure exerts a horizontal force on the structure. Other ex-
amples of structures that may need tension piles are the access ramps of tunnels, parts of infrastructure
like a railroad that are situated partially deeper than the surrounding ground water level, pump cellars etc.
In such cases one needs to compute the maximum admissible tension force, which is the subject of this
chapter.

A foundation pile that is subjected to a tension force generates shear stresses along the shaft. Since the
tip is not loaded, tip resistance should not be taken into account when calculating the strength. Note that
during the construction phase a tension pile can be exposed to compression forces (due to the lowering
of the ground water table by well-pointing (bronbemaling)). In other cases, e.g. a lock chamber (sluiskolk),
well-pointing results in the opposite situation. When the chamber is in use, a compression load is exerted
on the piles. The same chamber, when it is in maintenance and therefore empty, can be subjected to a
buoyancy force under its floor, causing a tensile load on the foundation piles. In order to determine the
shaft friction it seems logical to use the friction chart (wrijvingsgrafiek) of a sounding. This, however, is not
possible without restrictions since the friction as a result of tension is less than the friction due to com-
pression, see Figure 39-1. This reduction of the shear stress is a result of the fact that in case of com-
pression the vertical soil pressure increases around the mantle of the sounding pipe which leads to an
increased shaft friction, while in case of tension the vertical soil pressure decreases.

Figure 39-1 Friction chart of a sounding illustrating the difference between pressing and pulling of the sounding pipe.

As can be seen in the figure above there is not much difference in the clay layer, but there is a significant
difference visible in the sand layer.

288
Manual Hydraulic Structures Part IV 'Structures'

39.1 Bearing capacity of tension piles: comparison with compression piles

The pile resistance calculation of tension piles resembles that of compression piles. For the bearing ca-
pacity of the pile-soil system, however, there are some differences that should be kept in mind:

1. Tension piles only generate bearing capacity as a result of shaft friction, not tip resistance;
2. Tension piles have smaller coefficients for shaft friction than compression piles;
3. Tension piles are most commonly used in excavated construction sites to withstand the buoyancy
force underneath the structure after construction. As a result of the excavation of the construction
site, the soil pressure is reduced due to a weight reduction. Hence the effective soil pressure is
also reduced and thus the value of qc and the friction force on the pile are reduced;
4. Due to the tensile force exerted by the tension piles, the vertical effective soil pressure is reduced
and as a result also the value of qc, see Figure 39-1;
5. Not only should a single pile be tested for possible slip, but also a group of piles should be tested
as the piles may influence each other in a unfavourable way. If the piles are placed in a group,
the zones from which each pile receives it strength may overlap thus reducing the strength of
each pile. This is checked with the clump weight (kluitgewicht) criterion, which states that the ten-
sile force of a pile can never be larger than the weight of the soil that is, on average, present be-
tween the piles, see also Figure 39-2;
6. The safety factors used in the computation are larger than in case of compression piles. This is a
result of the fact that inaccuracies in determining the shaft friction are larger than when determin-
ing the tip resistance. These inaccuracies are enhanced further as a result of the effect described
in statement 4;
7. The dead-weight of the pile, usually the underwater weight, has a positive effect on the strength
of the tension pile.

As a result of statements 1-6 the tensile capacity of

a pile is much less than the compression capacity
of a pile. No standards exist for the computation of
tension piles. Usually the rules according to the
CUR-report 2001-4 “Design rules for tension piles”
are used, but they are disputable.

Particularly for very large tension forces grout an-

chors are used instead of tension piles. For this the
reader is referred to Section 45.2 “Extreme tensile
force, grout anchor” of this manual.

In the remainder of this chapter the method for

computing the strength of a single tensile pile is
described, after which the method for checking the
clump criterion is explained in order to establish
whether a group of piles is stable or not.

Figure 39-2 Maximum occurring friction in pile groups

289
Manual Hydraulic Structures Part IV 'Structures'

39.2 Bearing capacity: cone resistance method

For the computation of the strength of a tension pile many influences have to be taken into account. For
instance the reduction of the shaft friction due to reduction of the vertical effective soil pressure as a result
of the exerted tensile force and the excavation works. This leads to a complex computation method.
Hence, in general, the strength of tension piles is computed using a computer programme that deter-
mines the maximum shaft friction of the pile. For a first estimate of the strength of tension piles, the cone
resistance method, can be used, or other methods like the slip method and the method according to Be-
gemann. This Section explains the cone resistance method as applied in the Netherlands, also known as
the qc-method, as described in CUR report 2001-4 and Eurocode 7 (NEN 9997-1+C1). The most im-
portant reason for choosing the qc-method is that it does not include the soil-pressure-coefficient (K) and
the angle of internal friction (φ) of the soil, since these parameters are difficult to determine.

The maximum tensile force that a pile can resist without slipping is computed with the following equation:
L
Fr ,tension,d = ∫q c ,z ,d ⋅ f1 ⋅ f2 ⋅ Op,mean ⋅ α t dz (43.1)
z =0
Where:
Fr,tension,d [kN] = design value for the tensile strength of the soil
2
qc,z,d [kN/m ] = design value of the cone resistance at depth z
5
f1 [-] = pile installation factor (f1 ≥ 1)
f2 [-] = cone resistance reduction factor (f2 ≤ 1)
Op,mean [m] = average circumference of the pile shaft
αt [-] = pile class factor (paalklasse factor)
L [m] = length over which shaft friction is computed
z [m] = depth

The parameters are explained below.

Cone resistance
The design value for the cone resistance is determined per soil layer, using the following equation:
qc,z,rep q
qc ,z,d = where qc,z,rep = c,z,max (43.2)
γ m,b,4 ⋅ γ m,var,qc ξ
where:
2
qc,z,d [MN/m ] = design value for the cone resistance at depth z
2
qc,z,rep [MN/m ] = representative value for the cone resistance at depth z
2
qc,z,max [MN/m ] = value of the cone resistance at depth z in the sounding
ξ [-] = correlation factor, depending on the number of performed soundings,
the number of piles and the stiffness of the structure (see Table 39-1)
γm,b4 [-] = resistance factor for tension piles (= 1,35)
γm,var,qc [-] = factor for variable loads (see Figure 39-4)

If the cone resistance in the sounding exceeds 12 MPa, the value of 12 MPa should be used for qc,z,max,
unless the layer for which the cone resistance exceeds 12 MPa is at least one meter thick. In that case
the cone resistance of the sounding can be used with a maximum of 15 MPa.

If an excavation takes place after the soundings were made or there is over-consolidation of the soil, the
design value for the cone resistance has to be corrected for the reduction in cone resistance that is
caused by these processes. In case of an excavation this is done as follows:

• Piles driven after excavation or with a high vibration level:

σ v' ,z,exca
qc ,z,exca,max = qc ,z,max ⋅ ≤ 12 or 15 MPa (43.3)
σ v' ,z,0
where:
2
qc,z,exca,max [MN/m ] = corrected cone resistance at depth z after excavation
2
qc,z,max [MN/m ] = value of the cone resistance at depth z in the sounding
’ 2
σ v,z,exca [kN/m ] = vertical effective soil pressure at depth z after excavation
’ 2
σ v,z,0 [kN/m ] = vertical effective soil pressure at depth z before excavation

5
For conceptual designs, it may be assumed that f1·f2 = 1.

290
Manual Hydraulic Structures Part IV 'Structures'

• Piles driven before excavation or with very low vibration level:

σ v' ,z,exca
qc ,z,exca,max = qc ,z,max ⋅ ≤ 12 or 15 MPa (43.4)
σ v' ,z,0

If the corrected cone resistance is computed with equations (43.3) and (43.4), equation (43.2)
should be used to determine the design value of the cone resistance at depth z by replacing
qc,z,max with qc,z,exca,max.

• In case of over consolidated soil and the piles are driven using vibrations the cone resistance is
corrected using the following equations:

1
qc ,z,nc ,max = qc ,z,max ⋅ ≤ 12 or 15 MPa (43.5)
OCR
where:
2
qc,z,nc,max [MN/m ] = cone resistance at depth z corrected for over consolidation
2
qc,z,max [MN/m ] = value of the cone resistance at depth z in the sounding
OCR [-] = factor expressing the degree of over-consolidation

Again the design value of the cone resistance at depth z is computed using equation (43.2), now
replacing qc,z,max with qc,z,nc,max.

Material factor for variable loads (γm)

Alternating loading, meaning alternately a compression force and a tensile force or a large tensile force
and a small tensile force, on a structure leads often to the normative loading situation. In Figure 39-3 the
test results of a tensile test (trekproef) with alternating loading are showed. As one can see, the effect of
the alternating loading is that the pile head will rise when a large tensile force is exerted and will lower
again when the tensile force is reduced. Alternating loading, however, will cause a gradually rise of the
pile head, which is called creep. Tension piles are very sensitive for creep, which is therefore the most
common shortcoming of this type of piles.

Figure 39-3 Results of a test load on a prefabricated concrete pile (tension)

291
Manual Hydraulic Structures Part IV 'Structures'

The effects of alternating loading are taken into account with the material factor for variable loads, which
is a function of the maximum and minimum force exerted on the tension pile, see the figure below. In
case of tension the force has a positive value; in case of compression the force is assumed to be nega-
tive. The maximum value of the material factor for variable loads is 1,5.
Fs;min;d
Fs;max;d
γ m;var;qc

Figure 39-4 Factor for variable loads

The corresponding equation reads:

Fs,max,d − Fs,min,d
γ m,var,qc = 1,0 + 0,25 ⋅ where γ m,var,qc ≤ 1,5 (43.6)
Fs,max,d
where:
γm,var,qc [-] = material factor for variable loads
Fs,max,d [kN] = maximum design value of the exerted force
Fs,min,d [kN] = minimum design value of the exerted force

Influence of the number of soundings and installed piles

The influence of the number of performed soundings, the number of installed piles and the stiffness of the
construction is expressed by the factor ξ. Table 39-1 is used to determine this factor.

m n
1 2 3 4 5 10 >10
1 or 2 1,39 1,32 1,30 1,28 1,28 1,27 1,25
>3 1,29 1,21 1,18 1,16 1,15 1,15 1,13
Table 39-1 Values for the correlation factor ξ. This table applies to both compression and tension piles!

Where:
m [-] = number of piles installed underneath a structure
n [-] = number of performed Standard Penetration Tests (SPT)

Pile installation factor (f1)

When piles are driven into the ground the soil is compressed under the tip and along the shaft. As a result
the soil becomes stiffer, hence the cone resistance increases. The stiffening of the soil is taken into ac-
count via the pile installation factor (f1).

qc ,z,1 = f1 ⋅ qc ,z,d (43.7)

where:
2
qc,z,1 [MN/m ] = corrected design value for the cone resistance after pile installation
2
qc,z,d [MN/m ] = design value for the cone resistance before pile installation
f1 [-] = pile installation factor

Notice that equation (43.7) already is integrated in equation (43.1).

292
Manual Hydraulic Structures Part IV 'Structures'

The pile installation factor can be computed as follows:

n

∑ ∆e
f1 = e 3⋅∆Re where ∆Re = 1
(43.8)
( emax − emin )
 
n
r − 6 1 + e0 qc ,z
and ∑ ∆e = − ⋅ and Re = 0,34 ⋅ ln   (43.9), (43.10)
(
 61⋅ σ '
)
0,71 
1 5,5 50 
 v , z,0 
where:
f1 [-] = pile installation factor
Re [-] = initial value of the relative density of the soil
∆Re [-] = increase of the relative density of the soil as a result of pile installation
6
∆e [-] = decrease of the void ratio within a radius of 6 Deq of the pile
emax [-] = maximum void ratio of the soil (0,80 in NL)
emin [-] = minimum void ratio of the soil (0,40 in NL)
e0 [-] = initial void ratio of the soil
n [-] = number of piles within a radius of 6 Deq
7
r [-] = distance, expressed in Deq, from an arbitrary pile to the pile that is regarded
2
qc,z [kN/m ] = initial value of for the cone resistance
2
σ’v,z,0 [kN/m ] = initial value of the effective stress at depth z

In Figure 39-5 the relative density of soil is given as a function of the cone resistance for different effective
stresses. The influence area of a pile with respect to the increase of the relative density is a cylindrical
area with a maximum radius of 6 Deq.

Figure 39-5 Relative density as a function of the cone resistance for different effective stresses.

If one uses a pile installation factor that is larger than one, one should perform standard penetration tests
after installation to check whether the effective soil pressure indeed has increased or not. In the latter
case it is not allowed to use a value larger than 1,0 for the pile installation factor. In practice, it is incon-
venient to change the design during construction, therefore f1 is assumed to be 1,0 in most cases.

6
Deq = 1,13 ⋅ D□ , so for a pile 400x400 mm Deq = 1,13 x 400 = 452 mm.
7
When r > 6, the pile will not be effected by the other pile(s), therefore the maximum value for r = 6. This parameter,
in combination with equation (43.9), enables to distinguish between centre piles, edge piles and corner piles.

293
Manual Hydraulic Structures Part IV 'Structures'

In the following cases, the pile installation factor is always 1,0:

• For single piles;
• In clay layers;
• If a soil investigations are performed after the installation of the pile(s) with ground-displacing in-
stalling methods.

Cone resistance reduction factor (f2)

The tensile load on a pile group reduces the effective soil pressure in the soil layers from which a single
pile receives its tensile strength. Since for piles placed in a group these zones may overlap, the strength
of each pile is thus reduced even more. This reduction is taken into account by applying a reduction factor
f2 to the cone resistance. In case of a single pile the reduction factor f2 has a value of 1,0. The value of
the cone resistance reduction factor is always 1,0 for clay, in sand the value may become less than 1,0.

The reduction of the effective soil pressure is taken into account by dividing the soil along the length of
the pile into layers of for example 0,5 m thick. The factor f2 is subsequently computed per layer using the
following equation:
i −1
 
( )
−M i + M i2 + 2 ⋅ σ v' ,z,d ,0,i + γ d' ,i ⋅ d i ⋅  2 ⋅ σ v' ,z,d ,0,i + γ d' ,i ⋅ d i − 2 ⋅ ∑ Td ,n 
 n =0 
f2,i = (43.11)
2 ⋅ σ v ,z,d ,0,i + γ d ,i ⋅ d i
' '

f1,i ⋅ Op,i ⋅ α t ⋅ qc ,z,d ,i ⋅ 1000 ⋅ d i

With: Mi = and Td ,i = M i ⋅ f2,i (43.12), (43.13)
A

Where:
f2,i [-] = cone resistance reduction factor for layer i
2
Mi [kN/m ] = reduction of the vertical effective soil pressure in layer i
2
σ’v,z,d,0,i [kN/m ] = vertical effective soil pressure on top of layer i
3
γ’d,i [kN/m ] = design value of the effective volumetric weight of the soil in layer i
di [m] = thickness of layer i
2
Td,i [kN/m ] = maximum possible friction between the pile and the soil
f1,i [-] = pile installation factor for layer i
Op,i [m] = average circumference of the pile in layer i
αt [-] = pile class factor as indicated in Table 39-2 or Table 39-3
2
qc,z,d,i [MN/m ] = average cone resistance in layer i
2
A [m ] = influence surface area of the pile

The vertical effective soil pressure can be computed using the design value of the volumetric mass of the
soil:

i −1
γ
σ v' ,z,d ,0,i = ∑ γ d' ,n ⋅ dn Above the water table: γ 'd = (43.14)
n =0 γ m ;g
γ
Beneath the water table: γ 'd = − γ water
γ m ;g
Where:
2
σ’v,z,d,0,i [kN/m ] = vertical effective soil pressure on top of layer i
3
γ’d,i [kN/m ] = design value of the effective volumetric weight of the soil in layer i
di [m] = thickness of layer i
γm.g [-] = material factor for the weight of the soil
3
γw [kN/m ] = volumetric weight of water

The material factor for the weight of the soil (γm,g) is to be taken 1,0 if the weight of the soil works favour-
able on the pile’s maximum admissible tension force and 1,1 if the weight of the soil works unfavourable
on the pile’s maximum admissible tension force. As a result the material factor for the weight of the soil
will be 1,0 from ground level to the excavation level and 1,1 beneath the excavation level.

294
Manual Hydraulic Structures Part IV 'Structures'

Effective friction area of the pile shaft (Op,mean)

If a tensile force is exerted on a foundation pile, deformation of the soil occurs (in the direction of the pile
axis), causing a friction force along the pile shaft. The magnitude of this friction force depends, among
other things, on the amount of deformation. The deformation process depends on the soil type, e.g. small
deformations in sand and large deformations in clay. Since the soil structure consists of several layers
with a different deformation process not all layers contribute to the useful friction area of the pile shaft. In
Figure 39-3 the useful friction area is shaded.

When determining the useful friction area of the pile shaft one should keep the following aspects in mind:
1. As long as the subsoil consists of relative homogeneous sand, the circumference of the pile shaft
over the whole length of the pile is considered to be part of the useful friction area;
2. If the upper layers consist of clay or peat with a relative homogeneous sand layer beneath these
layers, only the contact surface area pile/sand is considered to be the useful friction area;
3. If the subsoil consists of two (or more) qualitatively different types of sand (density and composi-
tion) the whole contact surface area pile/sand is regarded as useful friction area. However one
should also account for the fact that:
• when the deformation is too small in the qualitatively poorer sand layer(s) only part of the
shear stresses will develop;
• when the deformation is to large in the qualitatively good sand layer(s) only the residual
shear stresses will occur (12 or 15 MPa);
4. When in a certain section of the subsoil (thin) clay layers occur one should account for the con-
sequences of possible consolidation (as a result of the driving or the occurrence of a tensile
force), such as relaxation or the deformation of the sand layers.

Pile class factor (αt)

The influence of the installation process on the shaft friction resistance is taken into account with the pile
class factor (αt). The magnitude of the pile class factor depends on the roughness of the shaft and the
installing method. In Table 39-2 the pile class factors are given for piles driven in sand and sand contain-
ing gravel. For other pile types, reference is made to Table 7d of Eurocode 7 (NEN9997-1+C1).

Pile class/type αt
• Ground displacing installing methods:
1)
o driven smooth prefab concrete pile and steel tube pile with closed tip 0,007
o pile made in the soil, whereby the concrete column directly presses
onto the ground and the tube is driven back (teruggeheid) out of the 0,012
2)
soil
o ditto, in case the tube is removed by vibration 0,010
o tapered wooden pile 0,012
• Screwed piles:
o with grout injection or mixing 0,009
• Piles with little ground displacement:
o driven steel profiles 0,004
• Soil removing piles:
o drilled piles (and auger piles) 0,0045
1)
The base of a tube pile with a closed tip shall not exceed 10 mm beyond the tube protruding.
2)
For this type of pile the diameter of the base may in principle be 30-50 mm larger than the outside
diameter of the casing.
Table 39-2 Maximum values for the pile class factor in sand and sand containing gravel (CUR 2001-4)

The contribution of cohesive layers to the shaft friction resistance is much smaller. Table 39-3 shows the
pile class factors for piles in cohesive layers. As a conservative value for clay, αt = 0 can be assumed. It
should be noticed that the water pressure in cohesive layers needs some time to adjust after soil dis-
placement. This implies that the shaft resistance in clay needs some time to develop after installation.

For piles that predominantly derive their tensile bearing capacity from sand layers, the contribution of
clay/silt layers may be taken into account as well. In these cases, the values for αt in Table 39-3 should
be reduced with 50% to account for the difference in the stress-strain behaviour of sand and cohesive
soils.

295
Manual Hydraulic Structures Part IV 'Structures'

3)
Soil type Relative depth z/Deq αt
Clay/silt: qc ≤ 1 Mpa 0< z/Deq < 20 0,02
Clay/silt: qc ≤ 1 Mpa z/Deq > 20 0,025
Clay/silt: qc > 1 Mpa not applicable 0,025
3)
In the case that piles derive their tensile resistance for most part from
sand layers, only half of the values for αt may be taken into account! This is
because of the difference in stress-strain behaviour of sand and clay/silt.
Table 39-3 Maximum values for the pile class factor in clay/silt

39.3 Bearing capacity: clump criterion

Not only should one check whether the tension pile will fail due to sliding and whether it can be pulled
“clean” out of the ground. One should also verify if a pile, including the surrounding soil, cannot be pulled
out of the ground. This is achieved by verifying that the clump criterion is met. The clump criterion states
that the tensile force on a pile can never be larger than the weight of the pile plus the clump of ground
around a pile. For a group of piles, the average weight of the ground situated around a pile between the
other piles should be used.

The clump criterion is computed using the following equation:

Fclump = Fr ,tension,max,d = (Vcone + Vcylinder − Vpile ) ⋅ γ d' ,soil + Vpile ⋅ γ d' ,concrete (43.15)

where:
Fclump [kN] = maximum tensile force the soil can absorb
Fr,tension,max,d [kN] = maximum tensile force the soil can absorb
3
Vcone [m ] = volume of the conical soil volume at the tip of the pile
3
Vcylinder [m ] = volume of the schematically cylinder-shaped soil volume near the
pile tip, below the conical part
3
Vpile [m ] = volume of the pile (the part under the ground surface)
’ 3
γ d,soil [kN/m ] = design value for the effective specific weight of the soil
’ 3
γ d,concrete [kN/m ] = design value for the underwater weight of the pile

Since the soil can absorb a tensile load only once, it is not allowed for the zones to overlap. The maxi-
mum influence area of a pile is a cylindrical area with a radius of 3 Deq.

In order to be able to determine the volumes of the cone and cylinder in equation (43.15), the top angle
should be known. This angle depends on the pile type and the position of a pile within the pile group. As a
result of the influence of other piles the top angle of piles within a group will be larger than for piles along
the edge of the group, see also Table 39-4. This means that the shape of the clump around the pile is
asymmetric for edge piles, at the outer edge the top angle is smaller and therefore the volume of the
clump is larger than at the inner side of the edge pile due to the influence of the other piles (larger top
angle).

Pile type Half top angle of piles with- Half top angle of the outer
in a group or the inner side side of edge piles
of edge piles
[°] [°]
Ground displacing installing methods 45 30
Piles with little ground displacement 2/3 ϕ ' 1/2 ϕ '
Soil removing piles 2/3 ϕ' 1/2 ϕ '
( ϕ ' = angle of internal friction of the soil at the pile tip)
Table 39-4 Top angle used for determining the clump volume

296
Manual Hydraulic Structures Part IV 'Structures'

Figure 39-6 Clump volume of tension piles

As said, clumps around piles cannot overlap, so if the centre-to-centre distance of the piles is less than
the maximum clump diameter 6 Deq (Figure 39-7), the clumps will be smaller than a cylinder around the
pile with diameter 6 Deq. Therefore, the shape of these clumps will actually be less round (partly straight
planes). The clump will even be completely box-shaped if the centre-to-centre distance is small enough
(< 6/√2 · Deq ). In this last case, the part of the clump at the tip of the pile has a truncated pyramid shape.

Figure 39-7 Different clump shape for dense pile groups

Self-weight of the pile

Since the self-weight of a tension pile is working favourably for its strength it may be taken into account
as follows:
γ pile
Wpile,d = Vpile ⋅ γ pile
'
where γ pile
'
,d = − γ water (43.16)
,d
γ m ,g
Where:
Wpile,d [kN] = design value of the pile’s self-weight
3
Vpile [m ] = pile volume (see Section 3.2 for the volumes of cones and pyramids)
3
γ’pile,d [kN/m ] = design value for the effective volumetric weight of the pile
3
γpile [kN/m ] = volumetric weight of the pile
γm,g [-] = material factor for the pile weight (1,1)
3
γwater [kN/m ] = volumetric weight of water

The design value for the dead weight of the pile can be added to the maximum design value for the ten-
sile force of the soil computed according to equation (43.1).

297
Manual Hydraulic Structures Part IV 'Structures'

39.4 Edge piles

Particularly around an open access road to a tunnel, edge piles sometimes have larger tensile loads than
the other piles. In this case, to determine the tensile load capacity, one must take into account:
• Limited reduction of the sounding resistance along the edge of the excavated building site due to the
excavations.
• Group action based on the average pile tensile forces in a cross-section and thus not based on the
tension in the edge piles themselves.
• Possible horizontal forces in sheet piling next to the pile.

39.5 Stiffness
This manual does not go into the description of the stiffness/flexural behaviour of tension piles. If grout
anchors are used they must be stressed after they have been placed, so that the subsoil does not float on
the groundwater but builds up effective soil pressure and thereby responds more stiffly.

39.6 References
Civieltechnisch Centrum Uitvoering Research en regelgeving, Cur-rapport 2001-4: “Ontwerpregels voor
trekpalen”, Gouda, juni 2001, Curnet.

Keverling Buisman, A.S., Grondmechanica, Delft, 1944, Waltman.

Verruijt, A., revised by S. van Baars, Soil Mechanics, Delft, 2005, VSSD.

298
Manual Hydraulic Structures Part IV 'Structures'

40. Pile foundations: Laterally loaded piles

improved: February 2016

40.1 Introduction
The forces on the foundation of buildings predominantly work in vertical direction. Therefore, in general,
the use of only vertical foundation piles will suffice for buildings. Hydraulic structures, on the contrary, fre-
quently have to resist considerable horizontal (= lateral) forces, besides the vertical forces. Think of hori-
zontal soil and water pressures behind a soil retaining structure, mooring and berthing forces on a jetty or
quay wall, etc. A deep foundation with only vertical piles is generally not suited to for resisting the large
horizontal forces and the overturning moment accompanying these forces. Therefore, the deep founda-
tion of a hydraulic structure is often a combination of vertical piles and piles driven under an angle to the
vertical. These piles are called 'raking piles' or 'batter piles' (schoorpalen).

Irrespective of the type of structure and the type of pile, vertical or batter, the horizontal load has to be
transferred into deeper soil layers where enough horizontal bearing capacity can be provided (see the
following section). In addition, the strength of the piles should be able to resist the shear forces. The
shear strength can be calculated with help of Chapter 35 for reinforced concrete piles and Chapter 36 for
steel pipes. As a rule of thumb, for the purpose of conceptual designs, a shear capacity of 100 kN could
be used for prefabricated concrete foundation piles, which in most cases is very conservative.

40.2 Blum: pile theory

A commonly used method to compute the required embedded depth (inheidiepte) needed to provide
enough passive soil resistance to lateral loads, is Blum’s method (Die Bautechnik, Heft 5,1932). Blum
used the following assumptions for his theory:
1) the embedded part of the foundation pile is regarded as an elastically supported beam;
2) the soil response is perfectly plastic;
3) the soil reaction on the deeper part of the pile is substituted by a concentrated force known as the Er-
satzkraft (R3), see Figure 40-2;
4) for stiffness considerations the pile is thought to have a fixed support at the depth were the concen-
trated force R3 is acting on the pile.

Ad 2: It is assumed that the elastic range of the soil is extremely small. As soon as there is a displace-
ment (u) the soil response is either immediately at is maximum (passive, σh,max) or minimum (active,
σh,min). Hence the stiffness of the soil is assumed to be very large. It is to be expected that the dis-
placements as a result of the load will be somewhat underestimated (see Figure 44 1).

Figure 40-1 On the left: elasto-plastic soil response. On the right: perfectly plastic soil response.

Ad 3: The basic idea is that the pile is displaced towards the right by the applied force, except at the lower
end were a displacement towards the left occurs because the pile rotates around a point somewhere
above its deepest point. The soil reaction towards the left is replaced by the concentrated force (R3) gen-
erally known as Ersatzkraft (German for substitute force), see Figure 44 2.

299
Manual Hydraulic Structures Part IV 'Structures'

Figure 40-2 Schematic representation of the shearing soil wedge behind a laterally loaded pile.
where:
F [kN] = load
R1 [kN] = resultant force of the soil wedges on both sides next to the soil wedge directly
behind the pile (two half pyramids)
R2 [kN] = resultant force of the soil wedge directly behind the pile (a triangle with width b)
R3 [kN] = substitute force ('Ersatzkraft')
a [m] = width of the pile in the direction of the load
b [m] = width of the pile perpendicular to the load
h [m] = length of the unsupported part of the pile
to [m] = theoretical embedded deptht
t [m] = practical embedded depth t = 1,2·to
φ [°] = angle of internal friction

Strength
The maximum load the soil wedge behind the pile is able to resist can be computed by taking the sum of
the moments at depth t0 below bed level, in this way the concentrated force R3 does not enter into the
equations (see Figure 40-3).

Figure 40-3 Blum’s schematization

300
Manual Hydraulic Structures Part IV 'Structures'

∑M D =0
−Fmax ⋅ ( t0 + h ) + R1 ⋅ 41 ⋅ t 0 + R2 ⋅ 31 ⋅ t0 = 0
Fmax ⋅ ( t 0 + h ) = ( 1
6 )
⋅ γ ' ⋅ K p ⋅ t03 ⋅ 41 ⋅ t0 + ( 1
2 )
⋅ γ ' ⋅ K p ⋅ b ⋅ t02 ⋅ 31 ⋅ t0
Fmax ⋅ ( t 0 + h ) = 1
24
⋅ γ ' ⋅ K p ⋅ t04 + 61 ⋅ γ ' ⋅ K p ⋅ b ⋅ t03 (40-1)

t03 to + 4 ⋅ b
⇒ Fmax = γ ' ⋅ K p ⋅ ⋅
24 to + h
where:
Fmax [kN] = maximum load that can be resisted
R1 [kN] = resulting force of soil at either side of the pile
R2 [kN] = resulting force of soil directly behind the pile
b [m] = width of the pile perpendicular to the load
h [m] = length of the unsupported part of the pile
to [m] = theoretical embedded depth
Kp [-] = passive soil pressure coefficient
3
γ' [kN/m ] = effective volumetric weight (γs - γw)
δ [°] = (pile) wall friction; generally δ = -⅔ φ

cos2 (ϕ − α )
K p,h,σ = 2
with α = β = 0
 sin(ϕ − δ )sin(ϕ + β ) 
cos (α )  1 −
2

 cos(α − δ )cos(α + β ) 

Because all (shear) forces are known, the moment diagram can be determined using the following equa-
tion:
x3
M x = F ⋅ ( h + x ) − γ ' ⋅K p ⋅ (x + 4 ⋅ b ) ⋅ (40-2)
24
where:
Mx [kNm] = moment at depth x
F [kN] = load
x [m] = distance below bed level
δM
The maximum moment Mmax occurs at depth Xm, where = 0 . This depth can be found by solving the
δx
following equation iteratively:

t 30 t 0 + 4 ⋅ b
xM2 ⋅ ( xM + 3 ⋅ b ) = ⋅ (40-3)
4 t 0+ h

Notes
1. The soil resistance only depends on the angle of internal friction and the volumetric weight of the soil.
Almost always the effective (submerged) weight of the soil has to be used in combination with δ=- ⅔φ, to
calculate the Kp, and arrive at the lateral bearing capacity.

2. In case of an impact load (stootbelasting) the load duration is short; in cohesive soils there is quite a
possibility that the groundwater does not have time to flow out of the soil wedge behind the foundation
pile or flow into the soil in front of the pile. For that reason Blum proposes to use the volumetric weight
of saturated soil (γs), however in combination with δ=0 to calculate the Kp.

3. For the strength considerations in the above, the position of the fixed support of the pile is at a dis-
tance XM below ground level.

301
Manual Hydraulic Structures Part IV 'Structures'

Stiffness
A vertical pile, fixed in the soil, will bend if it is loaded perpendicular to its axis. The displacement of the
part of the pile embedded in soil causes a passive resisting force. The bending moment diagram of, ac-
cording to Blum’s strength considerations, shows a curved line, see Figure 40-4. The pile could also be
modelled as a cantilever beam with concentrated horizontal load. For this last schematization, the bend-
ing moment diagram and the formula for the maximum displacement are well-known:

F ⋅ L3i
δ = (40-4)
3 EI

where: δ [m] = maximum displacement

F [N] = concentrated horizontal load
Li [m] = not-supported part of the pile (= h + 0,65·t)
-1
EI [Nm ] = bending stiffness of the pile

Figure 40-4 Moment diagrams for laterally loaded piles

Equating the static moments of both moment diagrams, Blum found the following expression for the dis-
placement of the pile head as a result of a lateral load F:

F ⋅ ( h + 0,65 ⋅ t )
3

δ = (40-5)
3EI
where:
δ [m] = displacement of the pile head
F [N] = load
h [m] = length of the unsupported part
t [m] = practical embedded depth
2
E [N/m ] = Young’s modulus of the pile material
4
I [m ] = moment of inertia of the pile
Li [m] = not-supported length of the pile h + 0,65·t

Hence, the overall pile stiffness, the spring stiffness k of the pile, which includes a contribution of both soil
and pile material, when the maximum load is acting, can be estimated as follows:

Fmax 3EI
k pile = = [N/m] (40-6)
δ ( h + 0,65 ⋅ t )
3

302
Manual Hydraulic Structures Part IV 'Structures'

As long as the soil can be schematized as one layer, Blum’s schematization results in a reasonable ap-
proximation of the maximum absorbable lateral load on a foundation pile. However when the soil profile
consists of multiple layers, then the displacements better be analysed using a finite element model or an
elasto-plastic spring model. In the latter case the foundation pile will be schematized as a beam on an
elastic bed. For subgrade moduli (k-values) of the soil (beddingsconstanten van grond) the reader is re-
ferred to Section 33.5 of this Manual.

40.3 References
Blum, H., Wirtschafliche Dalbenformen und deren Berechnung. Die Bautechnik, Heft 5,1932.

N.V. Houthandel v/h G. Alberts Lzn. & Co., Het ontwerpen, berekenen en vergelijken van ducdalven in
tropisch hout en staal, Middelburg, september 1957.

Veen, van der, C,. Horvat, E., Kooperen, van, S.H., Grondmechanica met beginselen van de funderings-
techniek, Waltman, Delft, oktober 1981.

Verruijt, A., Laterally loaded pile, March 29, 1999.

303
Manual Hydraulic Structures Part IV 'Structures'

41. Pile foundations: Pile groups

The previous chapters gave calculation methods for the determination of the pile load. The choice of
method depends on the schematisation of the structure. Figure 41-1 is a flow chart to determine which
calculation method to use. It is always possible to use a method shown to the right of the calculation
method selected with the flow chart.

Configuration
of piles and
loads

no no
Statically determined Rigid structure
structure

yes yes

only vertical no
piles

yes

all pile tips in a no

horizontal
plane

yes

Flat cross-section
Beam on two loaded by shear Displacement method Framework
supports forces, normal for rigid bodies (EEM)
forces and bending

Figure 41-1 Flow chart for the determination of a calculation method

A common problem in structural hydraulic engineering is a group of horizontally loaded piles.

304
Manual Hydraulic Structures Part IV 'Structures'

42. Stability of floating structures

Equation Chapter (Next) Section 1
Minor revision: February 2011, additional clarifications: February 2015

To ensure that floating elements do not undesirably move or rotate, they should be statically and dyna-
mically stable. The stability of a floating element depends on forces and moments, and the shape of the
element.

42.1 Static stability

Equilibrium of vertical forces

Vertical forces establish an equilibrium if the buoyant force (opwaartse kracht) equals the weight of the
floating body (including all ballast). For a floating body, this buoyant force has the same magnitude as the
weight of the displaced volume of fluid (Archimedes' principle: a floating body displaces its own weight of
fluid).

A vertical equilibrium is usually reached if the element is floating, or if it is resting on the bottom of the
water body. If there is no equilibrium, a completely immersed element will move upward or downward un-
til an equilibrium state is reached. An element will move upward if the buoyancy is more than the total
weight of the element. Then, at a certain moment, the part of the element rising out of the water will result
in a decrease of the buoyant force in such an amount that this buoyant force equals the weight of the el-
ement. It then stops moving upward. An element, conversely, will sink if the weight of the element ex-
ceeds the buoyant force, until it reaches the bottom. The bottom will resist the downward directed force
and will stop the element moving down.

Equilibrium of moments
For the design of large-scale prefabricated elements, vessels and dredging equipment, not just the weight
is of importance. There must also be an assurance that the elements do not tilt in an unacceptable de-
gree during the floating transport or the immersing procedure. Tilting can be initiated by the mooring forc-
es, wave motions, and inlet of water during immersion, etc. Elements must therefore be designed or
equipped in such a way that a rotation, caused by external factors, is corrected by a righting moment that
will return the element to its original position.

If the sum of moments around the point of rotation (= point of gravity) equals zero, the element will not
incline to tilt. This principle is illustrated in Figure 42-1, where equilibrium is not reached yet. It will tilt in
such a way that an equilibrium will be reached.

Figure 42-1 Forces acting on a floating element (Nortier, 1991)

305
Manual Hydraulic Structures Part IV 'Structures'

Static stability - metacentric height

A check of the equilibrium of moments (previous paragraph) is sufficient if an element is floating in still
water. In reality, however, this is rarely the case. This is why also the 'sensitivity to tilting' has to be taken
into account. A measure for the resistance to tilting is given by the 'metacentric height'. The principle is
illustrated in Figure 42-2. The left side depicts the cross-section of a floating element (like a caisson). On
the right side the same element is showed in tilted position. The rotation angle is φ.

Figure 42-2 Floating element

Indicated are three points, which are of importance in the evaluation of the stability:
• B is the centre of buoyancy (drukkingspunt), the point of application of the buoyant force Fb in state of
equilibrium (the state in which the axis of symmetry of the element is vertical). B is therefore the cen-
tre of gravity of the displaced water. In a rectangular container (caisson), B is found halfway between
the water surface and the bottom of the element. In tilted position the centre of buoyancy shifts to a
new position because the geometry of the displaced volume has changed. The shifted centre of
buoyancy is indicated with Bφ and the horizontal shift is a [m].

• G is the centre of gravity (zwaartepunt) of the element. If the element is filled with a layer of gravel or
water for the benefit of the immersing procedure (not shown in Figure 42-2), this weight should also
be taken into account when calculating G. Not only will this ballast lower the centre of gravity, it will
also increase the draught and will therefore raise B relative to the bottom. If the element heels over,
the centre of gravity generally remains fixed with respect to the element because it just depends upon
the position of the element's weight and ballast. The centre of gravity at the same time is the rotation
point.

• M is the metacentre; the point of intersection of the axis of symmetry, the z-axis, and the action line of
the buoyant force in tilted position. For small rotations (φ < 10°) the metacentre is a fixed point (see
lecture notes OE4652, 'Floating Structures' for a proof). The determination of point M is explained be-
low.

For static stability, rotation of the element should be compensated by a righting moment caused by the
buoyant force and the weight of the element. This is the case if M is located over G: the line segment
GM , also known as the metacentric height hm, must be positive.

306
Manual Hydraulic Structures Part IV 'Structures'

Figure 42-3 Tilted element

Figure 42-3 shows an element with a rotation φ. The part dx, which has been forced under water by the
rotation, experiences a buoyant force:

dFb = ϕ x dx ℓ ρw g (44.1)

in which ρw is the volumetric mass of water. This equation is only valid for rotations smaller than 10° in
which case tanϕ ≈ ϕ [rad]!

Relative to G this gives a moment dM = x dFb = ϕ x 2 dx ℓ ρ g .

+½b
Over the entire width this means a righting (corrective) moment of: M = ∫ ϕ x 2 dx ℓ ρ g , which could
x = -½b
+½b
be rewritten as M = ϕ ρ g ∫
2
x ℓ dx . (44.2)
x = -½b

+½b
Because ∫
x = -½b
x 2 ℓ dx represents the area moment of inertia (a.k.a. 'second moment of area') relative to

the y-axis, Iyy, the expression for the moment can be formulated as:

1
M = ϕ ρ g I yy , in which I yy = ℓ b3 . (44.3)
12

The point of application of the buoyant force (Fb) in a state of rest, is the centre of buoyancy B. A rotation
φ leads to a translation of the line of action of Fb over a distance a (see Figure 42-2). M = Fb · a, so:

M ϕ ρ g I yy ϕ I yy
a= = = (44.4)
Fb ρ gV Vdw

In this equation, Vdw is the volume of the immersed part of the element (= the volume of the displaced
water).

a a a
Because sinϕ = , BM = ≈ (44.5)
BM sin ϕ ϕ

I yy
So, BM ≈ (44.6)
Vdw
For small rotations (φ < 10°) the metacentre is a fixed point, but BM increases because the position of B
will go down. In case of considerable rotations, the metacentre displaces upward and sideways in the op-
posite direction in which the ship has rolled and is no longer situated directly above the centre of gravity.
If M is positioned above G, a righting moment F b h m ϕ = ρ g Vdw h m ϕ is created, which tries to return the
element to its stable position.

307
Manual Hydraulic Structures Part IV 'Structures'

For small seagoing vessels a metacentric height hm of at least 0,46 m is required. A ship with a small
metacentric height will be "tender" - have a long roll period. A low metacentric height increases the risk of
a ship capsizing in rough weather and more likely to develop "synchronized rolling". It also puts the vessel
at risk of potential for large angles of heel if the cargo or ballast shifts. If a ship with low hm is damaged
and partially flooded, the metacentric height will be reduced further and will make it even less stable.

For large ships hm should be at least 1,1 m, but not too large because in that case the vessel will be too
'stiff': it will snap back upright too quickly after a wave or wind gust has passed, which will cause heavy
stresses in the structural parts of the vessel, maybe shifting of the cargo and not unlikely sea sickness of
the persons on board.

The requirements for caissons and tunnel elements are less tight: 0,5 m suffices for hm. If M is positioned
below G, the element is unstable and will tilt.

Check design static stability

The check of the static stability (in this case also known as the outset stability, because only small rota-
tions of the element are investigated) is made up of the following steps:

• Calculate the weight Fw and the position of the gravity centre G of the floating element with reference to
K ( KG ). K is the intersection of the z-axis with the bottom line of the element. In general,

KG =
∑V ⋅ e ⋅ γ
i i i
, (44.7)
∑V ⋅ γ i i

3
where Vi [m ] = volume of element i
γi 3
[kN/m ] = specific weight of element i
ei [m] = distance between gravity centre of element i and reference level
(which is a horizontal plane through point K)

• Calculate the draught d of the element.

• Locate the centre of buoyancy B and calculate its position above the bottom of the element. This dis-
1
tance is KB . In case of rectangular elements, KB = d .
2
• Determine the shape of the area at the fluid surface and compute the smallest area moment of inertia I
for that shape (this is the most unstable). For rectangular elements: I yy = 121 ⋅ ℓ ⋅ b3 .

• Compute the volume of the displaced water Vdw.

I yy
• Compute BM =
Vdw

• Calculate hm = GM = KB + BM - KG

• Theoretically, if hm > 0, the body is stable. In practice, hm > 0,50 m is recommended. If hm < 0,50, addi-
tional measures are required.

Besides the static stability, the dynamic stability (the oscillation) must also be checked. This will be ex-
plained later in this chapter.

308
Manual Hydraulic Structures Part IV 'Structures'

Measures for unstable elements

If the element is unstable, the design should be altered or additional measures have to be taken.
Examples of design alterations are:
• widening of the element, thereby increasing the area moment of inertia I (usually the floor thickness will
increase too, because of strength requirements).
• making the floor of the element heavier. This lowers G and increases the draught (if the transport route
allows for this), which raises B relative to the bottom of the element. Unfortunately V also increases,
which decreases BM , but the other effects dominate, so the stability is increased.

Examples of additional measures are:

• adding ballast to the element (below the point of gravity) during transport.
• the use of stabilising pontoons or vessels (see
• Figure 42-4a), which increases Iyy.
• linking two elements during the floating transport (see
• Figure 42-4b), which increases I. Before the elements are disconnected at their destination, extra bal-
last must be applied to ensure the stability of the individual elements.

Figure 42-4 Stabilizing measures

If the required stability is achieved, one can opt to alter the design or use additional measures. Of course
combinations of both are also possible.

Stability during immersion

Floating tunnel elements generally owe their stability to their large area moment of inertia. Once the ele-
ments have been given extra ballast, they immerse under water and no longer have a plane intersected
by the waterline. The area moment of inertia is then zero. Stability is then only achieved if B is positioned
above G. However, the elements are lowered on four cables using winches placed on four pontoons (see
Figure 42-4c). This way, the elements are lowered accurately and in a controlled fashion. The element
and the four pontoons together act as one system, which, around the pontoons, does have a plane that is
intersected by the waterline and thus has an area moment of inertia. By positioning the pontoons as
closely to the corners of the element as possible, large moments of inertia arise, both in the transverse
direction and alongside the element.

Water ballast in the immersing process (free surface effect)

The use of water as ballast to immerse elements is attractive because it is a fast method and filling up
with water can be simply accomplished by opening the valves (mostly placed in or just above the floor).

If a closed element (i.e., with a roof) is completely filled with water, it acts in effect as a solid mass. This
means that its weight can be regarded as being concentrated at its centre of gravity. If the element is only
partly filled, or completely but not having a roof, the water surface is free to move and therefore possess-
es inertia. This causes a destabilising effect for the element, which can also be observed from Figure
42-5a. Due to the rotation, the depth of the ballast water increases on the left and reduces on the right.
This results in a moment that amplifies the rotation. If the inner space of the element is partitioned, as is
shown in Figure 42-5b, the moment driven by the ballast water decreases.

309
Manual Hydraulic Structures Part IV 'Structures'

Figure 42-5 Water ballast

The unfavourable influence of the ballast water on the stability can be discounted by defining I as:

I yy = Icaisson - ∑ I comp,i (44.8)

in which:
4
Icaisson [m ] = area moment of inertia of the plane intersected by the waterline if there
would be no water in the caisson
4
Icomp [m ] = area moment of inertia of the ballast water area relative to the gravity centre
line of the compartment concerned

It should be noticed that Steiners theorem (verschuivingsregel van Steiner) is not applicable here. This
can be explained with help of a somewhat different approach to the free surface effect (just another ex-
planation of the same phenomenon):

The centre of gravity of a compartment filled with a fluid will shift if the caisson (including the compart-
ment) is tilted. It shifts towards the lowest side of the compartment (moving away from the bottom of the
caisson). The shift of the gravity centre of the compartment in horizontal direction, similar to equation
(44.4) that expresses the shift of the centre of buoyancy, is:

Icomp
gg' = ⋅ϕ (44.9)
Vcomp
where:
4
Icomp [m ] = area moment of inertia of a compartment relative to its own axis through
the water surface in the compartment
3
Vcomp [m ] = volume of water in the compartment
φ [radians] = rotation angle

If the density of the liquid in the compartment differs from the water outside the caisson, this should be
taken into account according to:

ρcomp Icomp
gg' = ⋅ ⋅ϕ (44.10)
ρoutside water Vdw

The shift of the gravity centre of the compartment results in a shift in the gravity centre of the entire cais-
son. This shift of the gravity centre of the entire caisson in horizontal direction is:

 Icomp 
 ⋅ ϕ  ⋅ Vcomp
gg' ⋅ Vcomp Vcomp  Icomp
GG' = =  = ⋅ϕ (44.11)
Vdw Vdw Vdw
where:
gg' [m] = horizontal shift of the gravity centre of the compartment
3
Vcomp [m ] = volume of the water in the compartment
3
Vdw [m ] = total volume of 'outside water' displaced by the caisson
φ [radians] = rotation angle

310
Manual Hydraulic Structures Part IV 'Structures'

This corresponds to a rise of the gravity centre of the caisson of GG'' (in z-direction):

GG' Icomp ϕ Icomp

GG'' = = ⋅ = , (44.12)
ϕ Vdw ϕ Vdw

which is also called the 'free surface correction'.

For multiple compartments, the free surface corrections for the separate compartment have to be added
up to find the total free surface correction.

The rise of the gravity centre could be expressed as a decrease of the value of BM :

Icaisson Icomp,i Icaisson − ∑ Icomp,i

BM' = BM − GG'' = −∑ = , (44.13)
Vdw Vdw Vdw

so actually the shift of the caisson gravity centre can be expressed as a decrease of the area moment of
inertia:

I yy = Icaisson - ∑ I comp,i , indeed without applying Steiners theorem. Q.E.D.

Not only in transverse direction, as treated above, but particularly lengthways is the creation of compart-
ments beneficial for the stability!

The partitions, which incidentally won’t always be used and not for every type of prefab element, also
have the following advantages:
• Smaller spans (and so smaller moments and shear forces) of outer walls, floor and if present, roof.
• Realization of a better flow of forces. Concentrated loads (puntlasten) like shipping and crane wheel
loads can be resisted more easily.
• Correcting the tilt (trimming) during the immersing process, by letting more (or less) water into one
compartment than into the other.

Notes
• Water is let into ballast tanks during the immersion of tunnel elements. These tanks are of limited sizes, in order
to keep Icomp as small as possible.
• Working without tanks: letting water straight into the elements, immediately leads to tilting, especially length-
ways.
• Rainwater tanks used by farmers are often used as ballast tanks. These tanks are easy to fix and remove and are
also relatively inexpensive.
• The mentioned stability problems do not occur if sand or gravel is used as ballast material, provided that the ma-
terial is spread evenly (so there is no unnecessary tilt) and that a coincidental unwanted tilt does not lead to the
sliding of the ballast material.

311
Manual Hydraulic Structures Part IV 'Structures'

42.2 Dynamic stability

Not only the static stability but also the dynamic stability should be checked. If an element is transported
over water, it will be affected by waves or swell (deining). This can cause the element to swing
(slingeren), which can cause problems with respect to navigability and clearance (kielspeling).

Swing
If the dimensions (length or width) of a floating element are too small compared to the length of the waves
or swell, the element will start swinging on the waves (rolling or pitching). If there is only a limited clear-
ance beneath the the floating element, it is likely to hit the river or estuary bed and get stuck or damaged.
This is especially relevant for caissons for closure dams that have to be towed through rivers or estuaries
(with limited depth) to their final destination.

The following rule of thumb is being used in engineering practice:

Lw < 0,7 ⋅ ℓ e or Lw < 0,7 ⋅ be

(dependent on the direction of the waves relative to the caisson)

where:
Lw [m] = wave length
ℓe [m] = length of the floating element
be [m] = width of the element

If this condition does not apply, problems due to swinging of the element can be expected.

Natural oscillation
Worse than just swaying on the waves or on the swell is the movement of an element if the period of the
water movements comes close to the natural oscillation period (eigenperiode) of the element. In order to
prevent this, one must ensure that the natural oscillation period of the element differs significantly larger
form that of the waves or swell. For example, long swell was a problem for the transport of caissons in the
bay of South Africa. If the natural oscillation period could be a problem and adjustments of the design or
additional measures do not offer a solution, or are too expensive, the transportation and positioning at the
final location should be postponed until favourable wave conditions come back. This can, however, lead
to serious delays and thus larger costs. Therefore one should conduct a cost optimisation: on the one
hand the costs of additional measures and/or design alterations, on the other hand the costs of possible
delays.

Ignoring the hydrodynamic mass (the additional water mass) and damping, the natural oscillation period
T0 can be calculated by considering the angular acceleration and the righting moment of the floating
body. For small rotational angles, the body can be assumed to oscillate around its metacentre. The angu-
lar acceleration by definition is:

dω d 2ϕ
α (t ) = − =− 2 [rad/s2 ] (44.14)
dt dt
where:
ω [m/s] = angular velocity
φ [rad] = rotation after time t
t [s] = time

The direction of the angular acceleration α(t) is towards the equilibrium position, hence the minus sign.
The mass moment of inertia (so not the second moment of area, also called the area moment of inertia!)
of the floating body about the centre of gravity is:

G 2
IG = j [kg m2 ] (44.15)
g
where:
G [N] = weight of the floating body
2
g [m/s ] = gravity acceleration
j [m] = radius of gyration about the horizontal axis through the gravity centre
(sometimes indicated with i )

312
Manual Hydraulic Structures Part IV 'Structures'

The radius of gyration (polaire traagheidsstraal) j can be found according to:

I polar
j= , (44.16)
A

where A is the area of concrete in a vertical cross-section.

The polar moment of inertia Ipolar is a measure for accelerated rotation and should in this case be
considered around the y-axis:

I polar = ∫ r 2dA = I xx + Izz , (44.17)

A
where Ixx = rectangular moment of inertia respective to the z-axis, and Izz = rectangular moment of
inertia respective to the x-axis, both in relation to the centre of gravity G.

The mass moment of inertia of the body about the metacentre is:
G
IM = ( j 2 + h m2 ) (44.18)
g
where:
hm [m] = metacentric height

If the metacentric height is small compared to the radius of gyration, IG ≈ IM. For small angles of rotation,
the righting moment equals the mass moment of inertia multiplied with the angular acceleration:

− d 2ϕ
M = I ⋅ ϕ = G hm ϕ = IG (44.19)
dt 2

After substituting the equation for IG, it is obtained that:

G 2 d 2ϕ
G hm ϕ = − j (44.20)
g dt 2

It can then be deduced that:

d 2ϕ g ⋅ hm ⋅ ϕ
+ =0 (44.21)
dt 2 j2

d 2y
The solution of an equation in the shape + a 2 y = 0 is given by y = A ⋅ sin(ax ) + B ⋅ cos(ax ) , so in our
dx 2
case the solution is:

g ⋅ hm g ⋅ hm
ϕ = A ⋅ sin 2
⋅ t + B ⋅ cos ⋅t (44.22)
j j2

If t = 0, also φ = 0 and therefore B = 0, it follows from equation (44.22) that :

g ⋅h 
A ⋅ sin  2 m ⋅t = 0 (44.23)
 j 

If still φ = 0, and t = ½T0 (where T0 is the period of a complete oscillation), this can be rewritten as:

 g ⋅h  T
A ⋅ sin  2 m  ⋅ 0 = 0 (44.24)
 j  2

313
Manual Hydraulic Structures Part IV 'Structures'

Since A ≠ 0, this is only true if:

g ⋅h  T
sin  2 m  ⋅ 0 = 0 (44.25)
 j  2

The simplest solution of this equation is:

g ⋅ hm T0
⋅ =π (44.26)
j2 2

from which the expression for the natural oscillation period can be found:

2π j
T0 = (44.27)
g ⋅ hm

If the natural oscillation frequency is much larger than the wave or swell frequency, the element is dynam-
ically stable for oscillations (rotation). Short-term forecasts (a few days ahead) of wave characteristics can
for many locations be found on the internet, for instance www.windfinder.com, or https://waterinfo.rws.nl
(the last one only for the Dutch North Sea).

A large natural oscillation period is gained by a large polar inertia radius. A large metacentric height,
however advantageous for the static stability, decreases the natural oscillation period.

The same type of calculation has to be carried out for dynamic stability for pitching (vertical translation
instead of rotation), but this calculation method is not discussed here. A good example of pitching is the
rolled up mat that was used for the Oosterschelde storm surge barrier. This floating roll weighed 9000
tonnes and moved 0,60 m up and down in calm seas (H = 0,20 m en T = 9 s). Pitching can also be a
problem for dredging vessels, because the cutter can hit the seabed.

42.3 References
• Barltop, N..D.P. et al: Floating Structures: a guide for design and analysis. Centre for Marine and Petro-
leum Technology (CMPT), 1998.

• Eugene A. Avallone and Theodore Baumeister, eds.: Mark's Standard Handbook for Mechanical Engi-
neers. McGraw-Hill, New York, 1996.

• J.J.M. Journée and W.W. Massie: Offshore Hydromechanics. Lecture notes OE4620. Delft University of
Technology. August 2001. Chapter 2.

• Robert L. Mott: Applied Fluid Mechanics. Pearson, Prentice Hall, 2006. Chapter 5

• I.W. Nortier and P. de Koning: Toegepaste Vloeistofmechanica. Hydraulica voor Waterbouwkundigen.

7th edition. Stam Techniek, Houten,1991. Section 2.5

• H.J. Pursey. Merchant Ship Stability. 4th edition. Brown, Son & Ferguson, Glasgow, 1969.

• http://www.golfklimaat.nl, Rijkswaterstaat, accessed: July 2007.

• http://www.codecogs.com/library/engineering/fluid_mechanics/floating_bodies, accessed: August 213.

314
Manual Hydraulic Structures Part IV 'Structures'

43. Retaining structures

(major revision: 2011, improved and extended: 2014; sheet pile design improved in 2019)

A soil retaining wall is a structure intended to resist lateral (= horizontal) soil pressure. Retaining walls are
needed if the required slope of the surface exceeds the angle of repose (hoek van natuurlijk talud) of the
soil. If this maximum angle to the horizontal is exceeded, the soil will start sliding if not hindered by a wall.
Retaining walls are also applied if water has to be retained.

There are mainly three types of soil retaining structures (Figure 43-1):
1. Gravity walls (gewichtsconstructies)
- including L- walls, standard caissons, reinforced earth (terre armée), often masonry (metselwerk) or
(reinforced) concrete;
2. Sheet pile walls (damwanden)
- cantilever (onverankerd) or anchored (verankerd), including combi-walls (combiwanden), mostly
steel profiles, sometimes pre-stressed concrete sheetpiles (spanwanden);
3. Diaphragm walls (also called slurry walls; diepwanden), in reinforced concrete.

Figure 43-1 Five types of soil retaining walls (from: Wikimedia Commons, modified)

The following sections elaborate on the most interesting of these types. The designs of all these types
have in common that:
• The structure and the soil body as a whole should be stable
• The retaining wall is sufficiently embedded in the soil
• The wall elements are able to resist the load (especially the largest occurring bending moments)

43.1 Gravity walls

Gravity walls derive their stability from a combination of their shape and mass. They typically have a shal-
low foundation and their stability can be calculated according to Chapter 37 'Stability of structures on
shallow foundations'. For strength calculations, normal force, shear force and moment diagrams (nor-
maalkrachten-, dwarskrachten- en momentenlijnen) can be used to find the stresses occurring in the wall
material, which have to be resisted by the material properties.

If the displacements are not known, it cannot be judged whether the magnitude of the displacements are
acceptable. In that case, it is better to not allow displacements and to use neutral soil stresses, which
leads to conservative designs. The neutral soil pressure coefficient can be estimated using the formula
proposed by Jáky: K 0 ≈ 1 − sin ϕ ⋅ OCR , where φ is the angle of internal friction of the soil, and OCR
the Over Consolidation Ratio. If the friction angle is very small, e.g. for soft clays, the neutral soil pressure
coefficient is close to 1. The neutral soil pressure coefficient of sand varies from about 0,5 to about 0,7
(see also Section 24.2).

315
Manual Hydraulic Structures Part IV 'Structures'

43.2 Sheet piling

The least expensive soil retaining wall is usually made of steel sheet piling, particularly if it is only needed
temporarily (excavated building site) and can be reclaimed. Sheet pile walls are generally used in soft
soils and tight spaces. They are usually made of steel and can be installed one by one, or more usual in
pairs. The sheet piles are connected by special joints or interlocks, see Figure 43-3. If installed in pairs,
two sheets are connected before installation by tack welds (hechtlassen) or by continuous filled welds
(hoeklassen) along the joints (or interlocks, sloten). Another method to join two sheets to make pairs, is
by pinching (vastknijpen) the joints.

Sheet piles can also be made of pre-stressed concrete (spanwanden), wood or synthetic materials. Instal-
lation into the soil can be done by driving (heien), vibrating (trillen) or pressing-in (drukken) (Figure 43-2).

Figure 43-2 Driving a sheet pile Figure 43-3 Various joint types
(http://www.hjfoundation.com/Services/Sheet-Piling/) (ThyssenKrupp sheet piling handbook)

There are two main profile types of sheet piles: U-profiles (like the 'Larssen' profiles) and Z-profiles (like
the 'Hoesch' profiles). Two U-profiles together are not symmetrical which leads to oblique deflection when
subjected to load. This reduces the strength and stiffness of these profiles by tenths. Z-profiles don’t have
this problem. If the heaviest sheet pile profile is insufficient regarding strength or stiffness, one should
choose a combi-wall. It is least expensive to use steel quality S235. If this meets the demands for stiff-
ness but not for strength, it is advisable to opt for steel with a higher yield stress rather than for heavier
profiles. The AZ-type profiles from ArcelorMittal are relatively stiff.

Table 43-1, Table 43-2 and Table 43-3 give details for the most common sheet pile profiles in the Nether-
lands.

Notes
• See http://www.tk-steelcom.com.au/documents/sheetpile-handbook-full.pdf for details of many
Hoesch and Larssen profiles, connections and combinations. ArcelorMittal profiles can be found on:
http://sheetpiling.arcelormittal.com/
• Sheet pile walls do not only need a strength and stiffness design, but also a vibrating or driving de-
sign.

316
Manual Hydraulic Structures Part IV 'Structures'

Table 43-1 Sheetpile profiles, Hoesch type (from: Sheet piling handbook)

317
Manual Hydraulic Structures Part IV 'Structures'

Table 43-2 Sheetpile profiles, Larssen type (from Sheet piling handbook)

318
Manual Hydraulic Structures Part IV 'Structures'

Table 43-3 Sheetpile profiles, ArcelorMittal AZ type (from: ArcelorMittal brochure Z-profiles 2013-2)

319
Manual Hydraulic Structures Part IV 'Structures'

Calculation of sheet pile walls: general principles

Differences in the horizontal soil pressures in front of and behind a sheetpile wall, cause it to deform and
displace. Deformations occur due to the relatively low stiffness of the sheetpiles. Displacements can con-
sist of a rotational movement of the entire wall about a deep point, and, to a lesser extent, of a transla-
tional movement. These movements cause deformations of the soil, resulting in an active soil state where
the wall moves away from the soil, and a passive state where the wall moves towards the soil (see Sec-
tion 24.2.2). The embedded depth (t), the part of the wall under excavation level where it has soil at both
its sides, should suffice to provide a balance in horizontal pressures and forces, as well as a moment
equilibrium to prevent overturning of the wall.

If vertical stresses or forces act on the wall, e.g., the vertical component of an inclined anchor rod, or a
vertical line load coming from a top structure, the friction along the wall should be high enough to prevent
subsidence of the wall. The toe resistance of the wall helps to resist vertical loads as well, but this effect
is relatively very small.

This Manual explains the basics of the calculation of the required length and profile of the sheetpile ele-
ments with help of classical methods. These methods are suitable for hand calculations, hence very suit-
able for educational purposes, though the number of soil layers should not be more than two or three (a
simplification of the real situation might be needed). The results of the classical methods are not very pre-
cise, but could be used for preliminary designs. The classical methods assume soil coming into motion
because of the deflection or translation of a sheetpile wall in case of unequal loading. These methods
assume that soil wedges (grondwiggen) will start moving along straight slip planes (glijvlakken), if the sys-
tem reaches its limit state, see Figure 43-4.

Figure 43-4 Soil wedges and deformation of a sheet pile wall due to loading (very much exaggerated)

Hermann Blum proposed to simplify sheetpile wall calculations by assuming that the local displacement of
sheet pile walls would result in immediate yielding of the soil at both the active and passive side, instead
8
of the real gradual development of shear stresses in the soil . This means that soil deformations are as-
sumed to be maximum. At the active side, minimum soil stresses will occur and at the passive side max-
imum soil stresses (Figure 43-5). As a result, it is assumed that no elastic soil deformation occurs. This is
of course a simplification of the real situation, just as the assumption that the active and passive wedge
have straight slip planes. The consequence is that these methods are not suitable to calculate the defor-
mations of sheet pile walls.

8
Active and passive soil pressures are explained in Section 24.2.2 of this Manual.

320
Manual Hydraulic Structures Part IV 'Structures'

Figure 43-5 Blum's assumption regarding horizontal soil pressure

Next to the stability check of the active and passive slip planes, the stability of deep slip planes should be
checked as well, because larger soil bodies with slip planes below the active and passive wedges could
be mobilised as well. The stability of these soil bodies can be checked with help of the methods of Krantz,
who assumed a straight slip plane, and Bishop, who assumed a circular slip plane (see Figure 43-6 and
Section 45.4).

Figure 43-6 Deep slip planes: a straight plane (left) and a circular plane (right)

Other points of attention regarding the stability of sheet pile walls are piping and scour, which could po-
tentially lead to collapse of the wall (see sections 37.4 and 37.5 for the calculation methods).

A general point of attention for sheetpile wall calculations is the schematisation of water pressures. Usual-
ly, the water levels on both sides of a sheetpile wall are not equal. The water level of the surface water in
a harbour or river often differs from the ground water table at the other side of the wall. When assuming
that the water pressure increases linearly with the depth, according to σw = γw · z (where γw is the specific
weight of water and z is the depth below ground surface), this would lead to different water pressures at
the left and right side of the toe of the wall, as in case of an infinitely deep wall, see the thick lines in Fig-
ure 43-7, where σw,p = h1· γw and σw,a = h2· γw.

Figure 43-7 Hydrostatic pressures. Thick line: in case of no free flow of water; dashed line: after redistribution of unequal stresses

321
Manual Hydraulic Structures Part IV 'Structures'

In reality, this situation cannot exist. In sand, immediately after excavation and lowering the water level in
the construction pit, groundwater will start flowing, which leads to levelling of the water pressures on both
sides of the toe. The Handbook Sheetpile Walls (CUR 166, 5th edition) proposes an approximation meth-
od for the water pressures acting on the wall, which could be used for preliminary designs. It proposes the
following water pressures at the pile toe:

0,7d 3 0,7d3
σ w, ,p = h1 γ w + and σ w, ,a = h 2 γ w − .
d1 + d1 d 2 d 2 + d1 d 2

A more accurate method to determine the real water pressures uses the theory of potential flow (flow
nets), for which reference is made to Section 4.6.3 of CUR 166.

The situation in cohesive soils (clay) is different from sandy soils. Initially, the water pressure on the ex-
cavated side at the toe will be the same as in the original situation, thus the same as the water pressure
at the active side: σw,p = σw,a = h2· γw, and it linearly decreases to zero towards the bottom level of the
construction pit.

The following sections explain 'classical' methods to calculate the length and profile of sheet pile walls.
These simple methods can be used for analytical (hand)calculations as a first rough estimate. They are
all schematised in such a way that the internal moments and shear forces at the top and the toe of the
wall are equal to zero. Elastoplastic and finite element methods can be used for more detailed dimension-
ing of sheet pile walls.

Calculation of cantilever sheet pile walls

A sheetpile wall that is not supported by anchors or struts, just only embedded in the soil, mechanically
acts as a cantilever (uitkragende ligger). The pressure of the soil at the active side leads to a tendency of
the wall to move towards the passive side. This movement consists of a translation effect (because of an
imbalance of horizontal forces) and a rotational effect (because of an imbalance of moments). Because of
the flexibility of the wall, the sheet piles will deform as well. Figure 43-4 shows the deformations of a can-
tilever sheet pile wall in a rather extreme extent (failure state).

An increase of the height to be retained, and accordingly an increase in acting loads (until a certain
depth), results in rapidly increasing deformations and an increasing claim on the bending capacity of the
structure. Cantilever sheetpile walls are therefore usually not used as permanent structures. For tempo-
rary structures, like most building pits, they can only be used in the case of relatively low retaining
heights. Otherwise, the profiles would become too big and too long, hence too expensive compared to
supported sheetpile walls.

Schematisation of cantilever sheetpile walls according to Blum

As a result of the acting forces exerted on the sheet pile wall and the assumption that the sheet piles are
infinitely rigid, the sheet pile wall inclines to turn around point D, see Figure 43-8A. Accordingly, a clamp-
ing load (inklembelasting) will develop under point D, which resists the turning of the wall (Figure 43-8B).
If the length of a sheet pile wall is somewhat longer than strictly necessary to ensure equilibrium (so,
deeper than point D), the passive soil pressure does not develop over the lowest part of the wall. This
extra length, however, allows the constitution of a fixed end (inklemming).

Figure 43-8 A) sheetpile with rotation point; B) 'real' resulting soil pressure diagram; C) simplified resulting pressure diagram

322
Manual Hydraulic Structures Part IV 'Structures'

To calculate the required embedded depth of the wall, the real pressure diagram is simplified to a sche-
matised diagram as shown in Figure 43-8C. Blum proposed to replace the clamping load under point D by
a concentrated load (Ersatzkraft, or substitute force Q) acting through point D. In addition, the fixed end is
replaced by a free support at point D (hence, M = 0 at point D). The theoretical embedded depth can then
be calculated considering an equilibrium of moments about point D. Since the passive pressure at the
front side of the wall is underestimated in this schematisation, the deviation is corrected by increasing the
computed theoretical embedded depth (t0) by a factor α (1,20). This is done in such a way that the need-
ed shear force, Q, in point D can be delivered (Q = Fp,h – Fa,h). Notice that the magnitude of the field mo-
ment is not influenced by this schematisation.

The increment factor α that is used to increase the theoretical embedded depth t0, varies between 1,20
and 1,60, depending on the water level difference. Table 43-4 gives an overview of the various values for
this factor α.

Table 43-4 Increment factor α for the embedded depth of cantilever sheetpile walls (from: Hoesch Spundwand-Handbuch 1980)

Calculation method of the length (depth) and the profile of a cantilever sheet pile wall:

Step 1. Determine the governing load situation regarding water levels, surcharge and other external
loads;
Step 2. Calculate and draw the vertical soil stresses and the (ground) water pressures;
Step 3. Calculate the horizontal loads stresses at the active and passive side as a function of the theoret-
ical embedded depth (inheidiepte) t0.
Step 4. Calculate the embedded depth t0 considering an equilibrium of moments around the theoretical
pile tip, point D.
Step 5. Calculate the substitute force Q considering the equilibrium of horizontal forces.
Step 6. Calculate and draw the shear force and the moment diagrams.
Step 7. Determine the maximum moment in the sheetpile wall and calculate the minimum required sec-
tion modulus (Weff,y).
Step 8. Multiply the theoretical embedded depth t0 with the increment factor α, as given in Table 43-4 to
determine the real required length (t = α·t0).
Step 9. Choose a suitable sheetpile profile from the product specifications of the manufacturer.

Notice that the presence of water at the passive side of the sheet pile wall works favourable for the
strength of the sheet pile wall. Therefore, one shall either assume a certain safe (= guaranteed!) water
level at the passive side of the sheet pile, or assume that no water is present at all above the excavation
level.

Calculation of horizontally supported sheet pile walls

If the design of a cantilever sheetpile wall does not suffice, horizontal supports can be applied as alterna-
tive to large or long profiles. The most common types of supports are struts (Chapter 44) and anchors
(Chapter 45).

The mechanical schematisation of the supported sheet pile structure as a beam on two supports, or on a
support and a (partly) fixed end, is a simplification of reality, but acceptable for a first conceptual design
as it leads to conservative designs if the right safety factors are used. Figure 43-9 shows different, sche-
matisations of the toe of the piled wall. If the wall has insufficient embedded depth, it will move towards
the passive side (a), but this translation of the lower support point will be prevented if the embedded
depth increases (b). The sheet pile wall could be more or less considered as a beam on two supports. If
the wall becomes longer, a partly fixed end originates (c), which transforms into a fully fixed end at a cer-
tain depth, where the deflection angle is zero (d). A further increase of the embedded depth does not lead
to a change of the position of the fixed support and is needless.

323
Manual Hydraulic Structures Part IV 'Structures'

Figure 43-9 Effect of the embedded depth on the pressure distribution and the deflection

The first schematisation (a) is commonly called the free earth support schematisation. The passive
stresses in front of the wall are insufficient to prevent lateral movement and rotations in point C. The next
schematisations show the effect of increasing the embedded depth. In cases (b) and (c) the passive
stresses pressure has increased enough to prevent lateral deflection at point C; however, rotation still
occurs. In case (d) passive pressures have sufficiently developed on both sides of the wall to prevent
both lateral deflection and rotation at C. This schematisation is commonly called the fixed earth support
schematisation, because point C is essentially fixed.

For the free earth support schematisation (a), the soil at the lower part of the piles is incapable of inducing
effective resistance, so that it would not result in negative bending moments. The passive pressures in
front of the sheet piles are insufficient to prevent lateral movement and rotations at the lower end of the
piling. No passive resistance is developed on the backside of the wall below the line of excavation.

For the fixed earth support schematisation, the piling is driven deep enough so that the soil under the line
of excavation provides the required resistance against deformations and rotations. In other words: the
lower end of piling is essentially fixed.

Free and fixed earth schematisations are mutually independent; their results should not be mixed. Sheet
pile walls designed with free earth schematisations tend to be shorter and have a higher moment of iner-
tia against bending stresses than fixed earth schematisations, while the latter tend to be longer with a
lower moment of inertia.

The following two subsections explain the free and fixed schematisations in more detail.

1. Free earth support schematisation of supported sheet pile walls (American method)
The wall is schematised as a beam on two free supports and it is assumed that the anchor provides the
necessary resistance of the sheet pile (Figure 43-9b). Free earth methods result in a statically determi-
nate structure, but that structure is only stable under certain load conditions, namely 1) the sum of the
distributed loads equals to the anchored load and 2) the passive earth pressure resistance below the ex-
cavation line is sufficiently large to prevent overturning. Generally, the first criterion governs, with the re-
sultant horizontal support force being a result of the calculations.

324
Manual Hydraulic Structures Part IV 'Structures'

Calculation steps of the embedded depth and anchor force per running metre:
Step 1. Determine the governing load situation regarding water levels, surcharge and other external
loads;
Step 2. Calculate and draw the vertical soil stresses and the (ground) water pressures;
Step 3. Calculate the horizontal loads stresses at the active and passive side as a function of the theoret-
ical embedded depth (inheidiepte) t0.
Step 4: Compute the equilibrium of moments around the anchor point ∑ME =0 in order to determine the
embedded depth t0.
Step 5: Compute the support force T from the horizontal equilibrium ∑H =0.
Step 6: Determine the maximum moment in the sheetpile wall and calculate the minimum required sec-
tion modulus (Weff,y).
Step 7: Choose an appropriate sheet pile profile and length from the manufacturers information (t = t0).

To increase the safety of the structure, often either a larger maximum bending moment is assumed than
computed, or the passive ground pressure is reduced by using a conservative value for Kp.

2. Fixed earth support schematisation of supported sheet pile walls (European method)
With the fixed earth schematisation, the wall has two supports: a simple support at the anchor and a fixed
support at the toe. The system is therefore statically indeterminate, forcing the designer to consider sys-
tem deflections. This would become very elaborative for hand calculations, which is the reason why Blum
proposed to replace the fixed end of the toe of the wall by a free support plus a substitute force. The
loads on the sheet pile wall are schematized as shown on the right hand side of Figure 43-10.

Figure 43-10 left: 'real' situation; right: Blum’s schematisation.

The substitute force R (= Ersatzkraft) in the above schematisation is equivalent to the horizontal soil
pressure at the lower part of the right hand side of the sheet pile wall. Furthermore it is assumed that the
bending moment at the toe of the sheet pile is zero, but a shear force (of the magnitude Q) is allowed. To
ensure that this shear force can develop, the actual length of the sheet piles should be longer than the
sum of the retaining height and the embedded depth. Usually the computed embedded depth is increased
by 20%, similar to the cantilever method (see Table 43-5 for the exact values of this increment factor).

Table 43-5 Increment factor α for the embedded depth of supported sheetpile walls (from: Hoesch Spundwand-Handbuch 1980)

The computation of the main sheetpile characteristics of supported sheet pile walls differs from cantilever

325
Manual Hydraulic Structures Part IV 'Structures'

sheetpile walls, because now there are not two, but three unknown parameters: the embedded depth (d ),
the anchor force (T) and the substitute force (R). They can be solved by considering the displacement at
the anchor point, in addition to the horizontal equilibrium of forces, and the moment equilibrium. The dis-
placement of the sheetpile wall at the anchor point is prevented by the anchors, so it equals zero.

Compared to the free earth schematisation for supported sheetpile walls, Blums' schematisation will
mostly lead to somewhat longer sheetpiles, but the required profiles will be lighter and the displacements
will be less. Hence the advantages of Blum’s method, a lighter profile and smaller displacements, comes
at the expense of a longer sheet pile wall.

The calculation of the embedded depth, anchor force and substitute force per running metre thus consists
of the following steps:
Step 1. Determine the governing load situation regarding water levels, surcharge and other external
loads;
Step 2. Calculate and draw the vertical soil stresses and the (ground) water pressures;
Step 3. Calculate the horizontal loads stresses at the active and passive side as a function of the
theoretical embedded depth (inheidiepte)) t0.
Step 4: Compute the theoretical embedded depth (d) using the condition that the horizontal displacement
at the level of the anchor (point E) must be zero.
Step 5: Compute the equilibrium of moments around the toe of the sheet pile (point D) in order to deter-
mine the anchor force (T).
Step 6: Compute the substitute force (Q), from the horizontal equilibrium ∑H =0.
Step 7: Determine the maximum moment in the sheetpile wall and calculate the minimum required sec-
tion modulus (Weff,y).
Step 8. Multiply the theoretical embedded depth with the increment factor α given in Table 43-5 to deter-
mine the real required length.
Step 9. Choose a suitable sheetpile profile from the product specifications of the manufacturer.

Note
• The mechanical schematisation considered in the foregoing description of the 'classical methods' is a
statically determined system. However, in the case of multiple horizontal supports (struts or anchors),
a hand calculation is not possible (or very elaborative), because it cannot be solved (too many variables
compared to equations). This is also the case for a building pit with a cast and hardened under-water
concrete floor. The underwater concrete floor acts as an extra support and, if there is no water in the
pit, the horizontal forces in the upper supports will become higher than in the schematisation used in
the hand calculations. This also applies to the maximum moments in the wall. De Bruin (2018) did a
series a computer simulations, from which it appeared that the support force could become up to 40%
higher and the maximum moment 30% higher than in the 'classical load case' as explained above. The
required embedded depth remains the same.

Vertically loaded sheetpile walls

Sheetpile walls can be exposed to vertical forces, originating from:
• self-weight of the sheetpile elements
• vertical component of resulting soil force (active + passive)
• vertical components of inclined anchors
• surcharge or vertical forces from superstructures, usually transferred via a coping beam (kesp)
• possible other vertical forces acting on the wall
These forces have to be transferred to subsoil, which should have sufficient bearing capacity.

The bearing capacity of the subsoil consists of two components:

• bearing capacity of the tip of the sheetpile elements
• bearing capacity of the wall friction of the sheetpile elements

According to CUR-publication 166 'Quay Walls', the calculation of the vertical bearing capacity can be
done in analogy to the method for compression piles (see also Chapter 38 of this Manual):

Fr , max = Fr , max, tip + Fr , max, wall

and:

Fr , max, tip = A tip ⋅ pr , max, tip

326
Manual Hydraulic Structures Part IV 'Structures'

Fr , max, wall = 2 ⋅ Awall ⋅ ∫ pr , max, wall dz

where:
Fr, max, tip [kN/m'] = bearing capacity of the tip
Fr, max, wall [kN/m'] = shaft friction capacity
2
Atip [m /m] = cross-sectional area of the sheet pile wall
2
Awall [m /m] = wall area (one side of the wall)
2
pr, max, tip [kN/m ] = maximum tip resistance
2
pr, max, wall [kN/m ] = maximum wall friction

For the estimation of the maximum tip resistance pr, max, tip and the maximum wall friction pr, max, wall, refer-
ence is made to section 38.4.

The max. vertical force that can be resisted by wall friction is:

Fv,max = Fh · tan(δ),

where δ represents the angle of wall friction (wandwrijvingshoek). This angle can be determined with help
of tests on soil samples (direct shear stress test), but these tests are often difficult to carry out and the
result is anyhow not very reliable. Therefore δ is often related to the angle of internal friction (φ') and also
the coarseness of the soil. Eurocode 7 restricts δ to 2/3 φ' (for steel or concrete sheet piles in sand or
gravel) and to φ' for concrete cast in soil.

Based on long experience, Rijkswaterstaat found the following values for the angle of wall friction:
- for gravel with D50 > 8 mm: δ = 0°
- for course sand with D50 > 2 mm: δ = 1/3 φ'
- for sand with D50 < 2 mm: δ = 2/3 φ'
- for loam δ = 1/2 φ'
- for clay δ = 1/3 φ'

The direction of the wall friction respective to the wall is downward at the active side and upward at the
passive side. In case of an external load on the sheetpile wall, these directions can reverse.

For the calculation of the tip resistance of sheet pile walls, the following parameters should be used:
4
αp = 1,0; β = 1,0; s = 0,62. The tip cross-sectional area can be computed with Deq = A tip .
π
2
If the average vertical load does not exceed 12,5 kN/m , the vertical bearing capacity can be calculated
separate from the horizontal equilibrium check. Otherwise, an interaction calculation should be carried out
of vertical and horizontal forces. Usually finite element or spring models are used for these cases.

Note
• An eccentric vertical load will introduce a second order moment the wall, which should be taken into
account.
• Vertically loaded sheetpiles can be susceptible to buckling (knik), so this should be verified during the
design (see Section 36.3 of this Manual).

43.3 Combi-walls

Combi-walls (combiwanden) are walls that consist of a combination of different elements, for instance
sheet piles and steel pipe piles. Usually a combi-wall section, see Figure 43-11, consists of two pipe piles
with two to three Z-shaped sheet piles or one U-shaped sheet pile in between (The most commonly used
configuration has two Z-sheet piles in between). The pipe piles are fitted with interlocks, thus providing a
soil and water tight connection between the sheet piles and the pipe piles. The pipe piles have a far larger
flexural rigidity (EI) than the sheet piles and hence give the structure strength and stiffness. The sheet
piles provide a watertight seal in between two pipe piles.

327
Manual Hydraulic Structures Part IV 'Structures'

Figure 43-11 Cross-section, side-view and welded interlock of a combi-wall

An overview of combi-wall types is given in Figure 43-12.

Figure 43-12 Types of combi-walls (from: Grundwerke in Beispielen)

The disadvantages of combi-walls are that the pipe piles make the wall much more expensive and that
removing the pipe piles leaves a large hole, because the ground usually stays stuck within the tube (es-
pecially in cohesive ground layers).

Combi-walls are generally used for deep excavations. However as a result of the use of the pipe piles
they can withstand larger loads than conventional sheet pile structures. Examples of the application of
combi-walls are: underground car parks, quay walls and tunnel entrances. In quay walls the toe of sheet
piles are often placed less deep than pipe piles. The customary external diameter of the steel pipe piles
lies between about 800 and 2000 mm, the customary wall thickness lies between 10 and 20 mm.

As stated before, combi-walls have a far larger flexural rigidity. A combi-wall Ø 1620 mm × 19 mm + 2
6 2
PU20 has an EI equal to 2,28·10 kNm /m. That is almost 10 times as stiff as an AZ48! In this case the
flexural rigidity of the PU20 sheet piling is almost negligible compared with that of the tubular pile.

It is easy to calculate the flexural rigidity of tubular piles. After all, for tubular profiles:

I y = Iz =
(
π D4 − d 4)
64
where:
4
I [m /m] = area moment of inertia
D [m] = outer diameter pipe pile
t [m] = thickness tube wall = (D - d)
d [m] = inner diameter pipe pile = D - 2t
5 2 8 2
For the Youngs' modulus one can use: Esteel = 2,1·10 N/mm (= 2,1·10 kN/m ).

The bending stiffness (flexural rigidity) EI per metre combi-wall therefore mainly depends on the type of
pipe pile and on the collective width of sheet piling between these tubular piles. The area moment of iner-
tia (I) of a sheet pile profiles can be found in Table 43-1, Table 43-2, Table 43-3 or in the manufacturers’
fact sheet.

Note
• Combi-walls have to be checked for local buckling of the pile shaft.

328
Manual Hydraulic Structures Part IV 'Structures'

43.4 Diaphragm walls

The scarcity of space, especially in urban areas, has in recent years led to an increasing demand for un-
derground structures. The required depth of these structures has increased over the years, and also
stringent requirements regarding noise and vibration nuisance are imposed during construction. Further-
more often nearby adjacent properties are situated close to the building site, as a result of which the re-
taining walls should be very stiff in order to keep the deformation of the wall small. All in all diaphragm
walls, also called slurry walls, (diepwanden) have become an indispensable alternative.

Especially for diaphragm walls the relationship between design and construction is important. Critical
parts of the design and construction process are:
- the design of the reinforcement cage (good concrete flow ability and hence a large enough mesh width);
- control of the properties of both the concrete and bentonite;
- the process of de-sanding or exchanging the bentonite;
- the cleaning of the joints, if necessary;
- the process of casting the concrete.

Design
In this section only the basic outline of the design rules regarding diaphragm walls are treated. For more
detailed information the reader is referred to the Handbook Sheetpile Structures (CUR 166, in Dutch), the
Handbook Quay Walls (CUR 211, in English). CUR-report 76 gives design rules for diaphragm walls (in
Dutch).

Strength
In essence the strength that a diaphragm wall needs is computed in a similar way as for sheet pile walls
computations using Blum’s method, see Section 43.2 of this Manual. However, there are some differ-
ences. First of all the bending stiffness (EI) is rather different. As shown in the following computation the
bending stiffness of a diaphragm wall is much larger than that of a very stiff sheet pile profile.

As a result of the large bending stiffness of the diaphragm wall, the deformation at the top of the wall is
very small, resulting in an equally very small deformation of the soil behind the wall. The horizontal pas-
sive soil pressure should therefore be calculated with an passive soil coefficient (Kp) with an angle of in-
ternal friction of maximum ⅔φ ≤ 20°.

Another major difference with the computation of a sheet pile wall according to Blum’s method is that after
the initial computation, using Blum’s method, the diaphragm wall has to be computed again using a sim-
ple spring model, e.g. D-Sheetpiling. This is necessary because Blum’s method gives a good initial esti-
mation of the strength needed, but for a final design the computation with a simple spring model is neces-
sary in order to establish the real occurring bending moments and shear forces, as these are slightly un-
derestimated by Blum’s method (in case of a diaphragm wall). Usually the diaphragm wall will turn out to
be somewhat longer and heavier. The computed horizontal force and moment are used to determine the
amount of reinforcement steel needed.

An additional benefit of using a computer model is that curved sliding planes can be used, which gives a
more accurate representation of reality. For a hand calculation this is too complicated, so straight sliding
planes are used then. This, however, leads to an overestimated strength, which is compensated by a re-
duction of the wall friction angle (δ). According to NEN 6740, table 4, the following values for the wall fric-
tion should be used in diaphragm computations:

Curved sliding planes: - in sand and clay: δ/2

- in peat: δ = 0

Straight sliding planes: - in sand and clay: 2/3 δ

- in peat: δ = 0

Reinforcement and concrete cover

With respect to the computation regarding the required reinforcement, reference is made to Chapter 35 of
this Manual.

As a result of the way in which a diaphragm wall is constructed, several essential additional requirements
have to be met with respect to the reinforcement cage:
1) due to the long panel length (6-8 m), usually two reinforcement cages are placed alongside (in top

329
Manual Hydraulic Structures Part IV 'Structures'

view). Since the forces are mostly directed perpendicular to the wall the cages are usually not coupled
width-wise;
2) due to the large depth the reinforcement cages consist of parts of transportable lengths which are
coupled on site before or during the lifting;
3) stiffness of the reinforcement cages is of minor importance because they are suspended vertically.
Nevertheless, they should be stiff enough in order to be able to be transported and to prevent lasting
deformations during the placement. Usually welded cages are used, instead of twined cages
(gevlochten korven);
4) the flow ability of the reinforcement cage has to meet high standards since the support fluid has to be
displaced during the placement and because the concrete has to be poured continuously over a large
height;
5) the concrete cover has to be at least 75 mm under normal conditions and at least 100 mm in case of
weak soil layers, such as peat and clay;
6) the free space between the reinforcement cage and the grout float of the previous section must be at
least 100 mm, but preferred is 200 mm. The distance between the bottom of the wall and the rein-
forcement cage must be at least 200 mm;
7) the minimum centre to centre distance between horizontal reinforcement bars amounts 200 mm, be-
tween vertical bars this distance is 100 mm;
8) the minimum distance between reinforcement cages within one panel must be at least 200 mm, but
preferred is 400 mm.

The concrete used for diaphragm walls is usually C20/25 or C30/37 and the exposure classification is
XC4, alternating wet and dry (see also Chapter 35 of this Manual). Additional requirements are:
- a high resistance against segregation (ontmenging);
- a good consistency (plastic behaviour) and maintaining sufficient consistency;
- a good coherence;
- certain self-compacting properties.

Deformation
Both failure of the soil- and/or water retaining function of a diaphragm wall and the deformation of the wall
are likely to lead to settlements of the soil behind the diaphragm wall. In order to prevent damage, the
allowable deformations of the diaphragm wall and/or the adjacent land are limited. These limits are de-
fined as an allowable combination of settlement, relative rotation (β) and horizontal displacement (εh) that
adjacent structures may undergo. In appendix H of NEN-EN 1997-1 recommendations are stated for the
limit values for the deformation of normal, conventional structures. For many structures a maximum rela-
tive rotation of 1:500 is acceptable in order to prevent the occurrence of the serviceability limit state. The
ultimate limit state occurs around a relative rotation of 1:150, see Table 43-6. The ratios mentioned apply
to the sagging mode, see Figure 43-13. For a hogging mode (edge settling is larger than the part in be-
tween) the ratios mentioned should be halved.

Figure 43-13 Structural deformation

Furthermore it is stated in the NEN-EN 1997-1 that for separate foundations under normal conditions a
total settlement of maximum 50 mm is allowable. A larger settlement may be allowable provided that the
relative rotation is limited and no problems with house connections and public utilities (nutsvoorzieningen)
come into play.

Table 43-6 is suitable for newly constructed structures. For an older structure, that is likely to have al-
ready undergone a differential settlement, stricter rules must be applied e.g. an allowable rotation that is
1.5 times smaller than otherwise (1:500 instead of 1:300).

330
Manual Hydraulic Structures Part IV 'Structures'

Qualification Limit state Limit relative rotation (β)

Framing (skeletbouw) Masonry
1:300 (general) 1:600 (downward direction)
Architectonic damage
SLS 1:600 (skyscrapers) 1:1200 (upward direction)
(cracks up to 5 mm)
1:1000 (warehouse)
Structural damage 1:150 1:300 (downward direction)
ULS
(cracks 15-25 mm) 1:600 (upward direction)
1:75 1:150 (downward direction)
Collapse ULS
1:300 (upward direction)
Table 43-6 Allowable limits for the relative rotation

In Figure 43-14 damage thresholds related to a combination of the relative rotation and the horizontal dis-
placement are presented. In this figure, the relative rotation is the result of the differential settlement due
to the self-weight of the structure and the horizontal displacements represent the influence of the nearby
excavation.

Figure 43-14 Damage thresholds of diaphragm walls

Limits for the horizontal displacement in the subsoil at the location of pile foundations are not found in
literature. However during the design of the Willems railroad tunnel in Rotterdam a maximum horizontal
displacement equal to 1/6 of the shaft diameter was allowed as a safe limit for prefabricated concrete
piles. This limit value is related to the fact that in the case of a centric load on the pile no tensile stress
occurs in the pile shaft as a result of eccentricity. For wooden foundation piles a horizontal displacement
of 0,10 m was considered allowable.

Water-tightness
In many cases diaphragm walls are used to enable dry excavation in a deep building pit. Often the lowest
point of the wall is situated in an impermeable soil layer, but also underwater concrete can be used to
close off the bottom of the building pit.

The under normal circumstances occurring amount of leakage water depends on the following environ-
mental factors: - permeability of the soil layers behind the diaphragm wall;
- water pressure present in these layers;
- quality of the diaphragm wall and the joints.

To be able to relate the usage of an underground space to the allowable amount of leakage water, a
classification is desired. A possible classification based on Austrian guidelines is presented in Table 43-7.

331
Manual Hydraulic Structures Part IV 'Structures'

Classification Description Quantification water Functionality

locking ability
1 completely dry - specific goods storage
damp stains acceptable up to 1‰ of the spaces for public use,
2 relative dry visible surface. Water trails up to 0,20 m storage
damp stains acceptable up to 1‰ of the garages, infrastructural
3 slightly damp visible surface. A few water trails are projects
permitted
1
Maximum leakage per spot or per m garages, infrastructural
4 damp joint 0,2 l/h; and the average amount per projects with additional
2
m wall 0,01 l/h measures
1
Maximum leakage per spot or per m
5 wet joint 2 l/h; and the average amount per -
2
m wall 1 l/h
Table 43-7 Classification regarding the amount of allowable leakage water.

Taking into account a high water table, which is customary in the western part of The Netherlands, and
excavation depths for which the usage of a diaphragm wall is realistic, generally speaking the water pres-
sure against the wall is 10 metre water column or more. For a situation where aquifers are present behind
the diaphragm wall it turns out that in practise only structures belonging to classes 4 and 5 are feasible. In
case of water pressures exceeding 15 metre water column only class 5 structures are feasible. The con-
sequence of these findings is that usually a front wall (voorzetwand) must be constructed in front of the
diaphragm wall.

Construction
The construction stages of diaphragm walls are illustrated in Figure 43-15. Frist, the trench is excavated
in panels with help of guiding beams and special rectangular diggers (stages 1 and 2). A standard panel
has a width of 0,6 to 1,5 m and a length of 2,8 to 8,0 m. The depth can amount up to 30 m. As a rule of
thumb the minimum dimensions of the guide walls have a height of 1,00 m and a width of 0,20 m. These
guide wall dimensions are governed by the horizontal load (soil and equipment) and the changing fluid
level in the trench during the excavation. The fluid level is not allowed to drop regularly beneath the guide
walls as the fluid motion will erode the soil underneath the walls.

Figure 43-15 Construction stages of a diaphragm wall (from: http://www.wtcb.be)

The trench is filled with a support fluid (mostly bentonite) already during excavation to avoid collapse of
the soil walls. So, the excavation takes place in the support fluid. The support fluid becomes mixed with
soil falling down from the wall, so it is replaced by clean bentonite (stage 3) before the joints and the rein-
forcement cages are positioned (stage 4). Synthetic joint ribbons (voegbanden) are integrated in these
joints so that they can be cast-in in the concrete.

Subsequently, concrete is cast with help of cast tubes (stortbuizen) (stage 5). Since the density of con-
crete is larger than that of the bentonite mixture and because the concrete mixture is pumped in the
trench near the bottom, the bentonite mixture “floats” on top of the concrete and the surplus can be col-
lected for re-use. There is some mixing, hence the upper 1 to 2 m concrete is of poor quality and must be

332
Manual Hydraulic Structures Part IV 'Structures'

removed (and eventually replaced) afterwards. Finally, the joints are removed, but the cast-in joint ribbons
remain at their place so that they form a water-tight connection between the panels (stage 6).

Leakage of the joints

If the diaphragm wall appears to contain soil or bentonite lumps adjacent to the joints, this will most prob-
ably lead to leakage. This can be repaired by attaching steel plates to the inside of the diaphragm walls
and injecting PUR-foam behind the plates in the arisen gap. If the flow of the soil-water mixture through
the gap is too strong to use this repair method, the gap can otherwise be blocked with a 'dam' consisting
of clay or sand with a clay core, after which the soil behind the diaphragm wall is frozen by freeze lances.
After the flow is stopped the temporary dam is removed and again steel plates are positioned over the
weak spots. When this is done, the freeze lances can be removed.

Trench stability
If no special measures are taken during excavation of a deep trench, it is very likely that the walls of the
trench will collapse. In the case of diaphragm walls the trench is filled with a support fluid, usually a ben-
tonite mixture, to enhance the stability of the trench walls.

One distinguishes between micro instability and macro instability of the trench walls, see Figure 43-16.

Figure 43-16 Micro and macro instability of trench walls

Micro-instability is a failure mechanism that only occurs in non-cohesive soils. It occurs when individual
grains situated at the edge of the trench lose their stability and fall into the trench. Also the collapse of
small sections of the trench walls is regarded as micro instability. The presence of support fluid in the
trench can prevent micro instability. This will work if the support fluid does not move any more, in which
case it will have a certain shear strength. The fluid that has infiltrated the grain skeleton of the soil will
then stabilize the grains at the edge of the trench wall.

This yield stress (bezwijkspanning) can be computed as follows:

d10 ⋅ γ '
τf ≥ ,
tan(ϕ )
where:
2
τf [kN/m ] = yield stress
d10 [m] = grain diameter of the 10% fraction from the sieve curve of the soil
3
γ’ [kN/m ] = effective volumetric weight of the support fluid
φ [°] = angle of internal friction of the soil

Note that the grain size is a very important parameter in determining the micro instability of a trench wall.
In case the grain diameter is smaller than 0,2 mm or the layer thickness of the coarse material is smaller
than 0.5 m micro instability can be disregarded.

Macro instability is failure due to the sliding of a large soil section. In order to prevent this failure mecha-
nism the fluid pressure within the trench must be larger than the water pressure in the surrounding soil.
This overpressure, however, does not have to be very large: According to DIN 4126 the hydrostatic ben-
tonite mixture pressure must be 1.05 times larger than the hydrostatic water pressure at all depths. NEN-
EN 1538 prescribes that the level of support fluid must be at least 1 m above the rise (stijghoogte) of the
ground water in the layer under consideration. In general one strives to choose the support fluid level 2 m
above the highest rise. Macro instability is often checked with a finite elements model (FEM), important

333
Manual Hydraulic Structures Part IV 'Structures'

for the computations is the fluctuating support fluid level as a result of the digging process (which may
lead to erosion of the trench walls).

One should also be aware of the dynamic effects in the neighbouring area that can influence the trench
stability, e.g. strong vibrations caused by the hammering down of piles or sheet piles.

Sequence of work and panel width

From an economical point of view it is recommended to make the panel width as large as possible. An
additional advantage resulting from this is that the number of joints is as small as possible. It should nev-
ertheless be kept in mind that in most cases trench stability and impact on the surroundings are the gov-
erning factors.

Important for the sequence in which the panels are dug are the shape of the panel, the type of joint
(round or flat) between successive panels and the type of panel under construction.

One can distinguish the following panel types:

- start panels: do not connect to already constructed panels;
- follow up panels: connect to an existing panel at one side;
- end panels: connect to existing panels at both sides.

In the following table the different sequences of work are presented.

Table 43-8 applicability panel types and sequence of work

Usually one-movement panels (eengangspanelen) are constructed in special situations, e.g. to limit the
settlements of adjoining structures (higher trench stability!), or for logistic reasons (limitation of the
amount of concrete needed to construct a panel, so less change on disruption of the pouring process)
and as heavy foundation piles.

Types of joints between successive panels:

- joint pipe (voegbuizen);
- flat grout floats (voegplanken);
- permanent grout floats;
- cascade joint.

334
Manual Hydraulic Structures Part IV 'Structures'

43.5 References

Blum, H. Einspannungsverhältnisse bei Bohlwerken, 1931, Technische Hochschule Braunschweig.

Blum, H. Beitrag zur Berechnung von Bohlwerken. Die Bautechnik, 27 Jahrgang Heft 2 Februar 1950.

Blum, H. Beitrag zur Berechnung von Bohlwerken unter Berücksichtigung der Wandverformung, ins-
besondere bei mit der Tiefe linear zunehmender Widerstandsziffer. Berlin, 1951. Verlag von Wilhelm
Ernst & Sohn.

Bruin, A. de. Schematisering van een onderwaterbetonvloer in een bouwkuip. BSc-thesis Delft University
of Technology, 2018.

Civieltechnisch Centrum Uitvoering Research en regelgeving (CUR), CUR-publicatie 166: “Damwand-

e
constructies”, 5 druk, Gouda, 2008, CUR.

Civieltechnisch Centrum Uitvoering Research en regelgeving (CUR), CUR/COB-rapport 231: “Handboek

e
diepwanden – Ontwerp en uitvoering”, 1 druk, Gouda, 2010, CURNET.

Civieltechnisch Centrum Uitvoering Research en regelgeving (CUR), CUR-publicatie 211: “Handboek

kademuren”, Gouda, september 2005, Stichting CUR.

Nederlands Normalisatie instituut, Eurocode 3: “Design of steel structures – Part 1-1: General rules for
buildings” (NEN-EN 1993-1-1), januari 2006, Nederlands Normalisatie instituut.

Nederlands Normalisatie instituut, Eurocode 7: “Geotechnical design – Part 1: General rules” (NEN-EN
1997-1), maart 2005, Nederlands Normalisatie instituut.

Rowe, P.W., 1952. Anchored Sheet Pile Walls, PICE, Vol. 1, part 1, pp. 27-70.

Simpson, B, and Powrie, W. Embedded retaining walls: theory, practice and understanding. Perspective
lecture 15th International Conference on Soil Mechanics and Geotechnical Engineering. İstanbul, 2001.

Vanhoutte, B. and Verstrynge, F. Ontwerpmethodes voor damwandconstructies. MSc thesis Katholieke

Hogeschool Vives Noord. Oostende, 2007.

Verruijt, A., revised by S. van Baars, Soil Mechanics, Delft, 2005, VSSD.

Websites:
http://www.spoorzonedelft.nl/Het_project/Heden/Bouwtechniek_2

http://www.thyssenkrupp-bautechnik.com/products/sheet-piles/

http://sheetpiling.arcelormittal.com/

335
Manual Hydraulic Structures Part IV 'Structures'

44. Construction pits: Struts and wales

major revision : February 2011

Sheet pile walls are often horizontally supported by struts (stempels) or anchors. The advantage of a
support is that the embedded length of sheet piling doesn’t need to be as long (less passive soil pressure)
and heavy (smaller moment) compared to an unsupported wall, which saves money. Another advantage
is that there is less settlement of the ground level behind the sheet piling. The disadvantages of supports
are the higher costs and the possibility of obstruction during construction works.

As a rule, excavations of a couple of metres are carried out without supports, unless settlements behind
the sheet piling are unacceptable, e.g. for buildings. Deeper excavations are normally spoken carried out
with supports. Building site excavations can be supported in front of the sheet piling with struts. Support
behind the sheet piling is possible by using anchors. The forces of the strut or anchor are usually not
transferred to the sheet piling directly, but via wales (horizontal beams).

For the dimensioning of struts, see the following section; for wales see Section 44.2 "Wales" and for an-
chors see Chapter 45 "Anchors".

44.1 Struts
Struts consist of steel profiles that can absorb axial compression forces (strut forces). These forces cause
stress in the steel, which should not exceed the compression strength of the steel to avoid failure. Anoth-
er possible way of failure is buckling. Buckling is possible both in the vertical and horizontal plane, there-
fore it should be resisted in both planes. That is the reason why circular pipes are often used for large
spans and loads. For smaller spans and loads usually H-profiles are applied because they are less ex-
pensive and still sufficient against buckling.

So, struts should be checked on both strength and stability, which concerns their ultimate limit state
(ULS). In theory, the deflection in the serviceability limit state (SLS) should also be checked, but experi-
ence has learned that this normally spoken is not necessary for the design of struts.

Strength check
The dominant load in struts is caused by the normal (axial) compression force in horizontal direction. The
stress σc due to the combination of axial force and bending moments in two directions should not exceed
the yield stress (vloeispanning) of the steel: σn + σb < fy.

The required unity check thus is as follows:

NEd M y ,Ed M z,Ed

+ + ≤ 1,0 where: NEd = 1,5 ⋅ NRk ; M i ,Ed = 1,5 ⋅ Mi ,Rk
Aeff ⋅ fy γ M 0 Weff ,y ,min ⋅ fy γ M 0 Weff ,z,min ⋅ fy γ M 0

If the design shear force (VEd) exceeds 50% of the plastic shear force (Vpl,Rd) , the cross-sectional design
resistance to combinations of moment and axial force should be calculated using a reduced yield
strength:
fred = (1 − ρ ) ⋅ fy

 2 ⋅ VEd
for the shear area where ρ = 

− 1
2

and where: Vpl ,Rd =

(
Av fy 3 )
V  γ M0
 pl ,Rd 

τ Ed
The required unity check for the shear force is ≤ 1,0 and the local shear stress ( τ Ed ) may be
fy ( 3γ M 0 )
VEd ⋅ S 0,9 ⋅ VEd
obtained from: τ Ed = = .
t ⋅I tw ⋅ h

336
Manual Hydraulic Structures Part IV 'Structures'

where:
My;Ed [Nmm] = design value for the bending moment in the strut about the y-y axis
Mz;Ed [Nmm] = design value for the bending moment in the strut about the z-z axis
FEd [N] = design force of the load
2
Aeff [mm ] = effective cross-sectional area of the member
π π
⋅ D2 − ⋅ (D − 2 ⋅ t )
2
For a cylindrical profile: Aeff =
4 4
D [mm] = diameter of the cylindrical profile
t [mm] = steel thickness
2
fy [N/mm ] = yield strength
γM0 [-] = partial safety factor for resistance of cross-sections (γM0 = 1,00)
3
Weff;y;min [mm ] = minimum section modulus (moment of resistance) in y-y direction
3
Weff;z;min [mm ] = minimum section modulus (moment of resistance) in z-z direction
2
Av [mm ] = projected shear area ( for a cylindrical profile Av = Aeff)
4
S [mm ] = first moment of area

For calculating the moment of inertia I of cylindrical profiles see Section 43.3 “Combi-walls”.

The moment in a strut is caused by:

• a uniformly distributed load q, consisting of the self-weight of the strut and dropped sand laying on
the strut (design value of 1 kN/m', according to CUR166);
• a concentrated load FEd,grab (design value of 10 kN, according to CUR166) from a grab (grijper) that
accidentally hits the strut in the middle;
•
9
a second order moment introduced by the initial bending of the strut and the normal force .

The maximum moment in the strut in y-direction can therefore be computed with:

1 1
M y ,Ed = q ⋅ ℓ 2 + FEd ,grab ⋅ ℓ + NEd ⋅ eNy (The struts are considered to be hinge-supported.)
8 4
where:
NEd [N] = design value of the compression force in the wale
eNy [mm] = eccentricity of the axial force in y-y direction:
5 q ⋅ ℓ4 1 FEd ⋅ ℓ3
eN ; y = ⋅ + ⋅
384 EI 48 EI
q [N/mm] = uniformly distributed load
2 5
E [N/mm ] = Youngs' modulus (E = 2,1·10 )
4
I [mm ] = moment of area of the profile
ℓ [mm] = length of the strut

Stability check
A strut is considered to be stable if the stress caused by normal force (σn) and bending moment (σb), in-
cluding a second order effect introduced by initial bending of the strut, does not exceed the yield stress of
the steel:
 
 1  F π 2EI
(σ n + σ b ) ⋅  1  ≤ fy ;d where α cr = cr and Fcr = 2
 1−  FEd ℓ
 α 
 cr 

Fcr is the elastic critical buckling load (Euler force) for a global instability mode based on initial elastic
stiffnesses and the design buckling load FEd.

9
Additional forces caused by temperature changes should be included in the value of the normal force.

337
Manual Hydraulic Structures Part IV 'Structures'

NEd
It should also be checked that ≤ 1,0 (unity check)
Nb,Rd
where:
NEd [N] = design value of the compression force in the strut
Nb,Rd [N] = design value for the buckling resistance of the compression member:
χ ⋅ A ⋅ fy
Nb,Rd =
γ M1
χ [-] = buckling factor (formerly indicated with ωbuc)
2]
A [mm = cross-sectional area of the member
2
fy [N/mm ] = yield stress
γM1 [-] = partial safety factor for resistance to instability (γM1 = 1,00)

The buckling factor χ is a measure for the sensibility for buckling instability, and is dependent on the
relative slenderness λ of the profile. The relationship between the relative slenderness of a profile and its
buckling factor is given by the instability curve (see the graph or the equation in part III, Section 36.3). For
closed, hot formed tubular profiles, instability curve a is valid.
A ⋅ fy L 1
The relative slenderness is defined by: λ = = cr ,
Ncr i λ1
Where:
λ [-] = relative slenderness
Ncr [N] = critical Euler load
Lcr [mm] = the buckling length in the buckling plane considered
i [mm] = radius of inertia
E
λE [-] = Euler slenderness = λ1 = π ⋅ = 93,9 ⋅ ε
fy
ε [-] = strain

A third unity check deals with the ultimate limit state for the combination of flexural and lateral buckling
(knik en kip):

NEd M y ,Ed
+ ≤ 1,0
χ ⋅ NRk / γ M 1 χ LT ⋅ M y ,Rk / γ M 1
Where:
NEd [N] = design value for the compression force
NRk [N] = characteristic value for resistance to compression for the critical cross-section
NRk = A·fy
My;Ed [Nmm] = design value for the bending moment in the strut about the y-y axis
My;Rk [Nmm] = characteristic value of resistance to bending moments about the y-y axis
χ [-] = reduction factor (= buckling factor) (see Figure 36-1)
χLT [-] = lateral-torsional reduction factor (see Section 36.3)

It is not necessary to check a cylindrical pipe strut on lateral buckling (kippen), because all the axes are
equally strong. Therefore, the strut will sooner fail on vertical deformation than on lateral buckling. H-
profiles should be checked separately on lateral buckling.

Notes
• The strut (and also the sheet piling!) should be checked for fluctuating temperature loads during day
and night.
• The construction site should be checked for strut failure, especially the zip effect, resulting from a
strut being knocked out of position by a digger or crane.

338
Manual Hydraulic Structures Part IV 'Structures'

44.2 Wales
If the wales (gordingen) of a cofferdam are supported by struts, they usually exist of two H-beams placed
against the wall, with the webs (lijven) horizontal, and the flanges (flenzen) one under the other. If the
wales are supported by anchors, sometimes two U-profiles are used that are positioned opposite each
other, with the end of the anchor in between.

Figure 44-1: wales supported by anchors (left) and wales supported by struts (right).

The collar section of the H-beam will be the governing section of the entire cross-section. The combina-
tion of the normal, bending and shear stresses will be at its maximum here.

F
The stress due to the normal force F is σ n = , and is equally spread over the entire cross-section.
A
M
Stress caused by the bending moment can be calculated according to σ b = . The occurring bending
W
moment is resisted mainly by the flanges. For a beam on two supports, the maximum moment in a bend-
1 1
ing beam is: M = q ⋅ ℓ 2 and for a beam on multiple supports: M = q ⋅ ℓ 2 . Furthermore the maximum
8 12
1
moment at the end of a beam on multiple supports is set in between these two values: M ≈ q ⋅ ℓ 2 . This
10
is a good estimate for the moment working on the wales in the corners of a building pit. The design of
wales is therefore based on this moment.

Since it is assumed that the anchor force Fanchor is evenly spread across the sheet pile wall via the wales,
F
the following applies: q = anchor . In which ℓ is the distance between the anchors. The maximum bending
ℓ
1
moment in the wale is therefore: M ≈ ⋅ Fanchor ⋅ ℓ
10

The maximum shear stress is located at the centre of the cross-section of the H-profile and thus mainly
born by the web of the H-beam. Again the unity check is:
τ Ed V ⋅ S 0,9 ⋅ VEd
≤ 1,0 where τ Ed may be obtained from: τ Ed = Ed = ,
fy ( 3 ⋅ γM0 ) t ⋅I tw ⋅ h

where:
2
τEd [Nmm ] = local shear stress
tw [mm] = nominal web thickness
h [mm] = height of the web

The combined stress at the collar section as a result of both the bending moment and shear force is de-
termined from the stress distribution over the cross-section. Where the design shear force (VEd) exceeds
50% of the plastic shear force (Vpl,Rd), the design resistance of the cross-section to combinations of mo-
ment and axial force should be calculated using a reduced yield strength fred = (1 − ρ ) ⋅ fy for the shear

339
Manual Hydraulic Structures Part IV 'Structures'

2
 2 ⋅ VEd 
area, where ρ =  − 1 .
V 
 pl ,Rd 

The required unity check for the ultimate limit state is as follows:

NEd M y ,Ed M z,Ed

+ + ≤ 1,0
Aeff ⋅ fy γ M 0 Weff ,y ,min ⋅ fy γ M 0 Weff ,z,min ⋅ fy γ M 0
Where:

My;Ed [Nmm] = design value for the bending moment in the wale about the y-y axis
Mz;Ed [Nmm] = design value for the bending moment in the wale about the z-z axis
2
Aeff [mm ] = effective cross-sectional area of the member
2
fy [N/mm ] = yield strength
γM0 [-] = partial safety factor for resistance of cross-sections (γM0 = 1,00)
3
Weff;y;min [mm ] = minimum section modulus (moment of resistance) in y-y direction
3
Weff;z;min [mm ] = minimum section modulus (moment of resistance) in z-z direction
2
Av [mm ] = projected shear area ( for a cylindrical profile Av = Aeff)

The wales, being beams on many supports (the struts), and being evenly supported by the sheet piling,
do not have be checked on stability (buckling).

44.3 References
Nederlands Normalisatie instituut, Eurocode 3: “Design of steel structures – Part 1-1: General rules for
buildings” (NEN-EN 1993-1-1), januari 2006, Nederlands Normalisatie instituut.

Nederlands Normalisatie instituut, Eurocode 3: “Design of steel structures – Part 1-5: Piling” (NEN-EN
1993-1-5), februari 2008, Nederlands Normalisatie instituut.

340
Manual Hydraulic Structures Part IV 'Structures'

45. Construction pits: Anchors

45.1 General
A soil retaining wall can be anchored in several ways. The most important anchor types are:
1. Grout anchors
2. Screw grout anchors (schroefinjectiepaal) and self-drilling anchors
3. Anchor piles
4. Continuous anchor walls (doorgaande ankerwanden) and anchor screens (ankerschotten)
5. Screw anchors

Grout anchors
A grout anchor consists of a cylindrical body of cement grout, created in the soil. It usually has a length
between 4 and 20 metres and a diameter between 110 and 200 mm, depending on the diameter of the
drill pipe used. The anchor is connected to the sheet piling by means of a steel rod or steel wire that is
installed by driving or drilling a hollow drill pipe into the ground (see Figure 45-1, picture 1). Subsequently
the steel rod or wire is inserted into the drill pipe (2) and the cement grout is injected under high pressure
into the surrounding ground through the drill pipe while pulling back the pipe (3). When a grout body of
sufficient length is created, the drill pipe is pulled out completely. When the cement of the grout body has
hardened enough, the steel rod or wire is tightened up to 70-90% of the needed anchor force (4), after
which the external load is exerted on the sheet piling. As a result of this pre-stressing of the steel rod or
wire, most of the elastic strain is removed from the anchor. This thus limits the horizontal deformations of
the soil retaining structure.

Figure 45-1 Installation of a grout anchor

The way in which the anchors are installed influences their yield strength. Usually one achieves higher
tensile strengths using driven (geheide) anchors instead of drilled (geboorde) anchors. The anchors are
installed under an angle of 25-45° to a depth were a sufficiently strong sand layer is present in order to
form the grout body. Since the rods and wires are made of high quality steel with a relatively high allowa-
ble stress, their cross-section can be kept relatively small. Minor damage to a rod can therefore consider-
ably decrease its cross-section area, which has large consequences for the reliability of an anchor. The
maximum ultimate bearing capacity is around 3000 kN.

Screwed grout anchors

An anchor type derived from the grout anchor is the screw grout anchor, e.g. the Leeuwanker. This an-
chor type consists of a thick-walled tube with a spiral drill head welded onto the end. The grout body is
formed by injecting grout through the drill head while installing the anchor. Compared to the grout anchors
mentioned above the grout body is formed under relatively low pressure. The thick-walled pipe is made of
normal quality steel, which means that the maximum allowable stresses in the tube are a lot less than for
the high quality steel rods or wires in the grout anchors. The screw grout anchor, however, is a lot stiffer
than the grout anchor.

Another subtype of grout anchor that is often used nowadays is the self-drilling grout anchor, e.g. titan
and jetmix anchors. The main advantage compared to the other grout anchors is that the system is
cheaper to install (less laborious), but still has a large bearing capacity (1500-2000 kN). This anchor con-
sists of a tube with thread (schroefdraad) and a drill head. These threads are formed much like the ribs on

341
Manual Hydraulic Structures Part IV 'Structures'

a reinforcement bar and result in a higher bond friction with the grout body compared to standard drill
steel. During installation the drill head is moved back and forth while at the meantime grout is being in-
jected. In this way the soil is removed from the drill hole and replaced by the grout mixture. This results in
a drill hole with a diameter between 150 and 200 mm. After reaching the planned position of the grout
anchor, the grout pressure is increased and the drilling is continued until the thus created grout anchor
has reached its required length. Note that the drill hole widens after increasing the pressure (Figure 45-2).
After installation the tube serves as a draw bar (trekelement).

Drilling the drill hole Creating the anchor body

Cross-section of anchor body Drill tube

Figure 45-2 Installation of self-drilling grout anchor. (http://www.ischebeck.de)

Anchor piles
Anchor piles are used when very large anchor forces are involved. Also this type of anchor transfers the
load to the subsoil by means of friction. The presence of sand along the effective length of the pile is re-
quired. The effective length of the anchor pile is the part of the pile that is situated outside the active slid-
ing zone behind the sheet piling, see Figure 45-3. Nearly all pile types that are suitable as vertical tension
piles are also suitable as anchor piles. The strength of the anchoring pile is determined in the same way
as for tension piles and the stability is checked by means of the clump criterion, see Section 39.3 of this
manual.

An example of an anchor pile that is often used for heavy quay structures is the MV-pile (Müller Verpress
Pfähle, which is German for 'Müller grouting piles'). This type of pile consists of a steel H-beam which is
coated with cement grout during installation. The cement grout coating can be applied in two ways, first
by welding an enlarged foot at the end of the H-beam and second by welding steel strips on the outside of
the flanges of the H-beam. In both cases the cement grout fills the cavities that are formed in the ground,
thus forming the coating.

342
Manual Hydraulic Structures Part IV 'Structures'

Figure 45-3 Effective length of an anchor pile.

Continuous anchor walls and anchor screens

A continuous anchor wall and anchor screens can be used if there is:
• a reasonably strong and stiff top layer and fill.
• sufficient space to install the anchor wall and to excavate in order to install the tie rods.

As a rule, the advantage of anchor walls or anchor boards is that they are cheaper than other anchor sys-
tems. The disadvantage is that one must have sufficient space to dig trenches for the tie rods and to be
able to drive the walls or boards into the ground.

Screw anchors
Screw anchors consist of a steel tie rod with a diameter of approximately 30 mm width and at the end a
steel screw blade of approximately 250 to 500 mm that is welded to the tie rod. To install the anchors only
a small excavation behind the sheet piling is necessary. The anchors are installed using a vibrating drill.
This type of anchors can be installed horizontally or at an angle. The maximum capacity is attained when
the anchors are installed at a depth of approximately 12 times the diameter of the screw blade. When the
depth over blade diameter ratio is increased further, the capacity of the anchor hardly increases. In case
the depth over blade diameter ratio is smaller than 5, the capacity of the anchor will reduce rapidly. In
general the bearing capacity of a screw anchor is rather small (< 75 kN), therefore they are mainly used
for campsheeting (beschoeiingen).

For all types of anchor structures the bearing capacity is extremely dependent on constructional aspects.
To eliminate anchors with a small bearing capacity in an early stage, each anchor has to be tested before
it is connected to the retaining structure. The magnitude of the test load is around the design load of the
anchor force (Fs;A;d).

45.2 Extreme tensile force

Anchor wall and continuous anchor screen

When regarding limit states in determining the ultimate resistance of anchor walls, one normally assumes
straight slip planes according to Coulomb. These slip planes are shown schematically in Figure 45-4,
along with the horizontal soil pressures working directly onto the both sides of the anchor screen. Gener-
ally one assumes slip planes that extend unto ground level. This situation corresponds with an anchor
screen that extends till ground level. However this assumption only applies to anchor screens with a ratio
h
of 2 ≤ 1,5 , since in such situations the ground above the anchor screen will slide in the same way as if
h1
the screen would extend all the way up to ground level. In other words the anchor screen is fully embed-
ded.

343
Manual Hydraulic Structures Part IV 'Structures'

Figure 45-4 Unfavourable location of surcharge next to an anchor wall

Using Coulomb’s theory and assuming that the anchor screen consists of steel rods connected to steel
plates the maximum anchor force can be computed as follows:

FA;max;d = K p ⋅ 21 ⋅ γ ' ⋅ h 2 ⋅ a − K a ⋅ 21 ⋅ γ ' ⋅ h 2 ⋅ a = (K p − K a ) ⋅ 21 ⋅ γ ' ⋅ h 2 ⋅ a

Where:
FA;max.d [kN] = maximum anchor force the anchor screen can resist
Kp [-] = passive soil pressure coefficient
Ka [-] = active soil pressure coefficient
3
γ' [kN/m ] = effective volumetric weight of the soil (= γd or γs – γw)
h [m] = depth of the bottom tip of the anchor wall below ground surface
a [m] = centre-to-centre distance between the tie rods, with a ≤ amax
amax [m] = maximum centre to centre distance allowed between tie rods

In case a variable external load (q) is present at ground level, see Figure 45-4, the most unfavourable
situation is the presence of the load above the active slip plane of the anchor wall and the absence of the
external load above the passive slip plane. The maximum anchor force is now computed as:

FA;max;d = K p ⋅ 21 ⋅ γ ' ⋅ h 2⋅ a − K a ⋅ 21 ⋅ γ ' ⋅ h 2⋅ a − K a ⋅ qd ⋅ h ⋅ a = (K p − K a ) ⋅ 21 ⋅ γ ' ⋅ h 2⋅ a − K a ⋅ qd ⋅ h ⋅ a

where:
2
qd [kN/m ] = design value of the external load

The design value for the maximum anchor force (Fs;d) is determined as follows:

Fs; A;d = 1,1⋅ FA;max;d

In engineering practice a sheet pile wall is often anchored by a series of individual anchor boards spaced
at certain distances. When regarding these anchor screens, the maximum anchor force can be computed
in the same way as for a continuous anchor wall as long as the distance between the tie rods is not too
large. For rectangular anchor boards the maximum distance between tie rods can be expressed as:
amax = h1 ( β + α − 1) ,
where:
amax [m] = maximum centre to centre distance allowed between tie rods
h1 [m] = height of the anchor board
α [-] = ratio between the height and width of the anchor screen: b/h1
β [-] = Buchholz’s factor, according to Figure 45-5.

Figure 45-5 Buchholz’s factor

344
Manual Hydraulic Structures Part IV 'Structures'

If the centre to centre distance turns out to be larger than amax then amax is considered to be the total width
of the ground slice that slips.

Screw anchor
The way in which the ground around a screw anchor will fail depends both on the installation depth of the
anchor and on the angle of the tie rod relative to horizontal.

If the depth over diameter ratio of a screw anchor is smaller than 5 (H / D ≤ 5), the anchor is considered
to be shallow. This implies that the sliding planes reach unto ground level, see the upper anchor in Figure
45-6. In case of a deep anchor (H / D > 5), the collapse pattern of the soil largely resembles that of the
pattern around the tip of a compression pile, see the lower anchor in Figure 45-6.

Figure 45-6 Collapse pattern around an anchor blade

Hergarden (1983) researched the grip force of the anchor. For deep screw anchors in sandy ground he
found a relationship between the anchor force and the cone resistance in the area of influence of the an-
chor. This relationship reads as follows for sand:

Fr ; A; scr ; min = 0,4 ⋅ A ⋅ qc ; avg ≥ Fs; A; d

Where:
Fr;A;scr;min [kN] = minimum grip force of the screw anchor
Fs;A;d [kN] = design value for the maximum anchor force the screw anchor can resist
2
A [m ] = area of the anchor blade
qc;avg [kPa] = average cone resistance in the area of influence (3·D above and beneath
the anchor’s axis)

In cohesive soils the relationship reads:

Fr ; A; scr ; d = 10 ⋅ Cu; d ⋅ A ≥ Fs; A; d

Where:
Fr;A;scr;d [kN] = design grip force of the screw anchor
Fs;A;d [kN] = design value for the maximum anchor force the screw anchor can resist
Cu;d [kPa] = design value for the undrained cohesion of the soil
2
A [m ] = area of the anchor blade

345
Manual Hydraulic Structures Part IV 'Structures'

Grout anchor
When determining the maximum strength of anchors, one usually assumes the values given in Figure
45-7.

Figure 45-7 Maximum grip force for grout anchors with Danchor is 100 - 150 mm and H ≥ 4 m.

The reader is referred to CUR report 166 for information regarding a more detailed design of grout an-
chors. All anchors have to be subjected to a test load equal to the design load (Fs;A;d). The test results
may lead to a lower safety factor, which in turn leads to the use of fewer anchors. One should be aware
that over time the soil pressure, which is temporarily increased as a result of the installation of the grout
anchors, will creep back to lower values (this holds in sand too). Since the test loads only have a short
duration one tends to find too large values for the maximum anchor force (FA;max). Hence one should not
simply use the test values found for determining the maximum grip force with Figure 45-7.

Screw grout anchor

Screw grout anchors closely resemble auger piles (avegaarpalen), since no soil displacement occurs dur-
ing construction of the pile. The minimum grip force and hence maximum anchoring force of the anchor is
determined using:

Fr ; A;gr ;min = α t ⋅ qc ;gem ⋅ O ⋅ LA ≥ Fs; A;d

where:
Fr;A;gr;min [kN] = minimum grip force of the screw grout anchor
Fs;A;d [kN] = design value for the maximum anchor force the screw grout anchor can resist
αt [-] = pile class factor (αt = 0,011)
qc;gem [kPa] = average cone resistance along the embedded part of the anchor
O [m] = average circumference of the embedded part of the anchor
LA [m] = length of the embedded part of the anchor

Again all anchors should be tested with a load equal to the design value for the maximum anchor force.
This may lead to a lower safety factor, which in return leads to fewer anchors.

346
Manual Hydraulic Structures Part IV 'Structures'

45.3 Strength of a tie rod

The design calculation of a sheet pile wall results in an anchor force per running meter, which has to be
converted into an anchor force, FA, per anchor. The design value used for the tensile force in a tie rod
(Fs;A;st;d) is found by increasing the design anchor force with an additional load factor:

Fs; A;st ;d = FA;max;d = γ f ,b ⋅ FA;max with: γ f ,b = 1,25

Where:
Fs; A; st; d [kN] = design value for the tensile force in the tie rod
FA; max; d [kN] = design value of the maximum anchor force
FA; max [kN] = maximum anchor force the anchor can resist
γf; b [-] = load factor

The minimum cross-section of the tie rod is calculated using the maximum calculated force in the tie rod,
FA;max;d, according to:
A ⋅f
FA;max; d = s u with: γ m,a = 1,4
γ m,a
Where:
FA; max; d [kN] = design value of the maximum anchor force
2
As [m ] = cross-section of the tie rod
2
fu [kPa] = tensile strength of steel (see Table 36-1, where fu is in N/mm )
γm; a [-] = partial material factor

45.4 Total stability

One can distinguish three failure types for an anchored sheet pile wall:
- along a deep sliding plane;
- along a circular sliding plane;
- as a result of bursting of the bottom of the excavation pit, during construction of the anchor wall
See Figure 45-8.

Figure 45-8 sliding planes

In the remainder of this section only the first failure mechanism will be discussed in detail, since checking
for a circular slip plane is only required in specific circumstances such as short anchors and a small sheet
pile length. Bursting of the bottom of an excavation pit occurs if the water pressure under an impermeable
layer becomes too large compared to the layer thickness of the impermeable layer. This phenomenon is
also known as 'heave' (hydraulische grondbreuk).

In case one keeps the areas of influence of the anchor plates, or other tension elements, and the sheet
piling strictly separated, it is very unlikely that the whole structure consisting of sheet piling, anchor plates
and soil will slide. The areas of influence are shown in Figure 45-9. As a rule of thumb there should al-
ways be a distance of at least 1 meter between the straight sliding planes at ground level (This is added
safety!). If the distance between the sheet pile wall and the anchor becomes too small, the areas of influ-
ence intersect, see dashed lines in the figure below. This results in higher pressures at the active side of
the retaining structure (here sheet pile wall), while the resistance of the anchor is reduced. That is the
pressures at the passive side are lowered.

347
Manual Hydraulic Structures Part IV 'Structures'

Figure 45-9 Areas of influence using straight slip planes.

Using the method of Kranz, one can compute the required minimum distance between the anchor and the
sheet pile wall. If this distance is too short, a straight sliding plane will occur approximately between the
point where the shear force in the sheet piling is zero and the bottom of the anchor plate. More specific,
for free supported sheet piles and anchor screens the sliding plane will occur between the bottom of the
sheet pile wall and the bottom of the anchor structure. For clamped sheet pile walls and anchor screens,
the sliding plane will occur between the lowest zero point of the shear force in the sheet pile wall and the
bottom of the anchor structure.

In order to prove that the total stability is sufficient one should demonstrate that the design value of the
maximum anchor force can be supported by the soil retained between the sheet pile walls and the anchor
structure. This is done by regarding the balance of the soil body enclosed by ground level, the vertical
planes just behind the sheet pile wall and just in front of the anchor structure and the assumed deep slid-
ing plane, see Figure 45-10.

All the forces working directly on this soil body have to be taken into account when using this method.
Forces working directly onto the retaining structure are already accounted for in the computation of the
design value for the maximum anchor force. For example; if a bollard is situated in the ground behind the
sheet pile wall, this force is working directly on the soil body regarded with Kranz’s method. Thus, it has to
be accounted for. However if the bollard is connected to the sheet pile wall it has already been taken into
account when computing the design value for the maximum anchor force. Therefore, it is not considered
in the Kranz method.

Figure 45-10 Computational scheme according to Kranz.

348
Manual Hydraulic Structures Part IV 'Structures'

On the soil body ABCD (light shaded area in Figure 45-10) the following forces are working, which should
balance each other:

G [kN] = weight of the soil body ABCD

Ea [kN] = force exerted on AB by the sliding active soil mass in case of failure. This force
can be deduced from the active pressure on the sheet piling and the soil mass
Ga of the soil body AA’ B working on the active sliding plane A’B.
E1 [kN] = soil pressure exerted on the vertical through the anchor structure. In this force
also the presence of a possible external load (q) should be included.
Q1 [kN] = force exerted on the plane BC, that in case of failure makes an angle (φ’) with the
normal on this plane.
FA [kN] = anchor force exerted on the soil body by the anchor structure, as computed in the
sheet pile computation.
Fbollard [kN] = bollard force, if present.
c’ [kN] = friction force as a result of cohesion, if present. This force works in the direction
of plane AB.
φ’ [°] = angle of internal friction

In case the above forces do not balance each other, the soil body ABCD will slide along plane BC. As a
result the direction of force Q1 is determined and the force polygon (krachtenveelhoek) can be drawn for
the chosen distance between the sheet piling and the anchor structure, see Figure 45-10. Or one can
determine the maximum allowable anchor force analytically.

Note that the presence of a horizontal groundwater table behind the sheet piling only affects the weight of
the soil masses; there is no resulting horizontal water pressure. Water pressure as a result of a head dif-
ference is already included via the computed maximum anchor force as computed in the sheet pile com-
putation. In case of a curved groundwater table behind the sheet piling one should include the influence
of this curved groundwater table by including a resulting horizontal force caused by the difference in water
pressure over the length of the anchor, see Figure 45-11.

Figure 45-11 Influence of a curved groundwater table.

When after computation the maximum allowable anchor force is too large one can either increase the
length of the anchor bar or increase the length of the sheet piles.

45.5 References
Civieltechnisch Centrum Uitvoering Research en regelgeving (CUR), CUR-publicatie 166: “Damwand-
e
constructies deel 1”, 5 druk, Gouda, 2008, CUR.

Civieltechnisch Centrum Uitvoering Research en regelgeving (CUR), CUR-publicatie 166: “Damwand-

e
constructies deel 2”, 5 druk, Gouda, 2008, CUR.

Webpages:
http://www.gebr-vanleeuwen.nl
http://www.ischebeck.de
http://www.jetmix.nl

349
Manual Hydraulic Structures Part IV 'Structures'

46. Construction pits: Underwater concrete floor

46.1 General
Underwater concrete is mostly used to create a watertight floor, usually of a building pit. It takes away the
necessity of dewatering, which can lead to considerable cost savings. In addition, the risk of settlements
due to groundwater extraction is reduced by using an underwater concrete floor. Underwater concrete
can be applied in floors with a maximum slope angle of 6 to 7°.

The composition of underwater concrete is quite similar to normal structural concrete. During application,
special attention is paid to the prevention of washing out of the concrete when in contact with water,
which could lead to a badly mixed concrete that does not harden sufficiently. In exceptional cases, un-
derwater concrete can be reinforced with steel bars, but this is problematic regarding construction. How-
ever, if it is reinforced, it can be considered structural concrete, for which one needs to take larger toler-
ances concerning the shape, the dimensions and the quality of the concrete into account. Loads on rein-
forced underwater concrete are usually transmitted by bending and shear forces. Non-reinforced concrete
transmits its loads different from reinforced concrete, as non-reinforced concrete cannot take tensile forc-
es. Hence, as a rule, more concrete is required in case of non-reinforced concrete. On the other hand,
one does not need to place reinforcement cages under water.

The forces in a non-reinforced underwater concrete floor can be compared to those in masonry arches.
Between the floor supports (piles and sheet piling), pressure arches are created, which transmit the loads
to the supports.

sheet piling tensile piles sheet piling

underwater
concrete
pressure arch

Figure 46-1 Pressure arches in underwater concrete

An arched structure owes its strength to so-called arch thrust. This arch thrust is a horizontal reaction
force in the support (see Figure 46-1 and Figure 46-2). The arch thrust makes an equilibrium with the
strutting force in the floor, which originates from the sheet piling.

The non-reinforced concrete floor can also be compared with a pre-stressed floor. In this case the strut-
ting forces can be considered pre-stressing forces in the floor. Because the floor is subjected to a com-
pression load, the non-reinforced concrete is able to absorb moments without the presence of tensile
stresses.

The non-reinforced concrete floor can therefore be schematised as a number of arches between the sup-
ports and as a plate that is pre-stressed in two directions with supports in a number of points and lines.

At the supports, however, the reaction force has to be transmitted by a shear force between the concrete
and the pile. If steel piles are used, the concrete is usually well attached to the pile, however, if the shape
of the pile encourages the creation of gravel pockets or soil enclosures, one cannot simply depend on the
attachment. Clean contact areas are of large importance for a good attachment. If prefab concrete piles
are used, the top metres of the sides of those piles are often given ribs to improve the join.

46.2 Limit states

There are three ultimate limit states that apply for an underwater concrete floor:
1. Floatation of the floor
2. Fracture of the floor
3. Fracture of the join between the floor and the tension piles

350
Manual Hydraulic Structures Part IV 'Structures'

Floatation of the floor

In the case of this ultimate limit state; one must consider the equilibrium of the floor. The downward forces
are:
• the weight of the floor
• the weight of the piles and anchors, plus the attached clump weight (weight of the soil under water that
is pulled up by the piles) or the maximum friction along the pile shaft
The upwards forces are:
• the upwards groundwater pressure under the floor
• the effective soil pressure under the floor (loads caused by swell of lower clay and peat layers)

Fracture of the floor

Fracture of the floor is possible when the ultimate bearing capacity of the arch is exceeded. The determi-
nation of the ultimate bearing capacity of the arch created in the underwater concrete floor is not simple
because the geometry of the arch is not set beforehand. To determine the maximum bearing capacity one
often assumes an arch with a height of 75% of the floor thickness and with a thickness of 10% of the floor
thickness. Generally, this limit state is not normative for the thickness of the underwater concrete.

Fracture of the join between the floor and the tension piles
A fracture of the join between the floor and the pile is possible if the maximum acceptable shear stress
along the piles is exceeded.
To describe this limit state one cannot simply use the limit state “Punch” of TGB 1990 for columns punch-
ing through floors, because there is no firm attachment to the reinforcement steel. A safe approach is to
only consider the transfer of forces from the floor to the pile in the contact areas between the pile and the
underwater concrete.
In the case of a steel pile the force from the floor is transferred in the entire contact area, assuming a
good attachment of the pile and the concrete. The slide plane in the concrete is assumed to be the small-
est envelope around the pile.
In the case of prefab concrete piles, one usually assumes that the shear force in the floor is transferred to
the pile along the ribbed sides of the pile.
The limit state involved assumes that the calculation value of the occurring shear force in the mentioned
slide planes equals the maximum acceptable shear force in the underwater concrete.

46.3 2-D Arch effect

The shape and height of the pressure arches depend on the type of load and the arch thrust that occur.

Figure 46-2 Schematisation of a pressure arch

Before we look more closely at the (3-D) arch that is created between the supports, the general distribu-
tion of forces in a two-dimensional arch is considered first.
The value of the arch thrust can be found by considering the balance of moments in the centre. The verti-
cal reaction force and the uniform distributed load to the left of the centre cause an anticlockwise moment
equal to 1/8 q l . The arch thrust in the left support causes a clockwise moment equal to H⋅ f.
2

The balance of moments leads to:

ql 2 ql 2
H= or f =
8f 8H

The optimum shape of an arch with a uniform distributed load is derived from the balance of moments for
every point on the arch:
1
H y = q x (l − x )
2

351
Manual Hydraulic Structures Part IV 'Structures'

The position of the median of a pure pressure arch, loaded by an equally distributed load is:

q x (l − x )
y=
2H

For a horizontal non-reinforced concrete beam, in which a uniform distributed load leads to a pressure
arch, the height of the arch is determined by the point of application of the horizontal arch thrust and the
centroid of the stress diagram in the middle of the beam. The shape of the stress diagram depends on the
deformations in the cross-section of the beam and on the δ-ε diagram of concrete. Figure 46-3 shows a
number of possible stress diagrams.

Figure 46-3 Centroid of the pressure arch in concrete, dependent on the deformation

The position of the line of action of the horizontal arch thrust depends, amongst others factors, on the way
the pressure is applied to the side of the beam. In this case the position of the line of action depends on
the angle of rotation of the plane on which the force acts. For an underwater concrete floor in an excavat-
ed building site, this angle of rotation occurs when the wall and floor deflect after the water has been re-
moved from the site.

46.4 3-D Dome effect

The previous section only discussed the arch. In an underwater concrete floor with tension piles, pressure
arches will occur between the piles in different directions. Between these pressure arches, the load is al-
so transferred by means of arches. Between four piles the arches will form an inverted dome. Figure 46-4
schematically shows the main load bearing structure, drawn in thick lines.

Figure 46-4 Dome effect between four piles

The horizontal arch thrusts on the dome can be calculated by considering the balance of moments
around a horizontal axis through the centre of the dome.

352
Manual Hydraulic Structures Part IV 'Structures'

Figure 46-5 Balance of moments around an axis through the centre

q l 2 l12 q l1 l 2 2
The equilibrium leads to: H1 = and H2 =
8f 8f
H1 l1
Consequently: =
H2 l2
V q l1 l 2
The vertical reaction force is: =
4 4

46.5 Transfer of forces to piles

The reaction force that the piles have to provide is transferred from the floor to the pile through the con-
tact area between the floor and pile.
For a clean steel pile, the steel and concrete attach like reinforcement steel attaches to concrete. Next to
the pile, the load is transferred to the underwater concrete floor by means of shear stress (see Figure 46-
6).

For a smooth prefab concrete pile, the attachment between the underwater concrete and the pile is inferi-
or. For this reason prefab piles usually have ribs along the top few metres, to ensure a certain amount of
catch resistance.
Steel vibro combination pile prefab concrete pile

ribs

shear stress shear stress

water pressure water pressure

pile, shaped in the ground

Figure 46-6 Shear stress along a pile shaft

The underwater concrete floor is loaded by the upward water pressure less its own weight. The floor
transfers the load to the piles. A test calculation will have to decide whether the thickness determined
practically is sufficient.

353
Manual Hydraulic Structures Part IV 'Structures'

47. Construction pits: Dewatering

Major revision: February 2008

This chapter first covers dewatering in general. Subsequently, it shows how (theoretical) designs are
made.
For comprehensibility of the dewatering mechanism, one should distinguish between two types of aqui-
fers (waterhoudende grondlagen):
• An "unconfined aquifer" is an underground layer of porous material with a free groundwater level
(niet-afgesloten grondlaag met freatisch water).
• In a "confined aquifer" the groundwater is trapped between impermeable ground layers (afgesloten
grondlaag met spanningswater). The water level usually rises above the top of the aquifer, in which
case the well is called an artesian well (artesian water, overspannen water). Sometimes the water
level is below the top of the aquifer (onderspannen water).

47.1 General
Dewatering methods
Theoretically two methods of dewatering are possible:
• open dewatering, where drainage trenches along the foot of a slope carry the water to wells. From
there the water is removed from the construction site using dirty-water pumps (Figure 47-1). Dewater-
ing in the open is only possible for small water level reductions in impermeable soil types (at most 4 to
5 metres in clay).

Figure 47-1 Dirty-water pump (from www.pedrollo.co.uk/)

• wellpoint drainage (bronbemaling), for which the following systems can be distinguished (see Figure
47-2):
vacuum drainage with closed wells, where flexible pipes connect the riser (stijgbuis) of every well
to a header (= suction pipe, zuigbuis). The under-pressure in the header is created using centrif-
ugal or plunger pumps.
vacuum drainage with open wells. The system is the same as the vacuum drainage mentioned
above, though here each well has a draw pipe (haalbuis) suspended in it. The advantage over
the system mentioned above is that the water level in the wells can be reduced maximally, pro-
vided the draw pipes are long enough.

Because the under-pressure can never be more than atmospheric pressure (1 atm = 101 kPa ≈
10 m column of water), the water level reduction that can be achieved by means of vacuum
drainage is rarely more than 5 to 6 metres. Vacuum drainage in deep excavations must therefore
be carried out in stages. The method is applicable to a range of soil types, from silty to fine sand
3
sills. The yield per well is approximately 1,5 to 2,0 m /hour.

354
Manual Hydraulic Structures Part IV 'Structures'

header header header

ground level ground level ground level

clay clay
clay

filter encasing
filter pipe 2"
delivery pipe
draw pipe
casing
filter pipe 3" - 5"
reflux valve
filter encasing
underwater pump

1 - 2 m perforation perforation
motor

perforation

vacuum drainage with closed wells vacuum drainage system with open wells underwater pump

Figure 47-2 Types of wellpoint drainage

a pumping system with underwater pumps, in which a pump is placed in every well (perspomp).

Underwater pumps are suspended in the groundwater and force up the water mainly through com-
pression instead of suction. Thus the limit of the suction head (zuighoogte) is not relevant and large
groundwater level reductions can be realised. This makes this type of pump attractive for the con-
struction of hydraulic engineering works. The installation is more reliable than a vacuum drainage
system, in which air leaks can lead to problems (air suction in one well can stop the water yield of
all other wells connected to the same pump). One must ensure that the pump is suspended at suf-
ficient depth to prevent it from falling dry. Underwater pumps are used in soil types ranging from fi-
3
ne sand to coarse gravel. The capacity of the pumps varies between 5 and 250 m /hour. In hydrau-
3
lic engineering pumps with capacities between 40 and 120 m /hour are usually employed.

Dewatering problems
A pump system (e.g. consisting of wells with underwater pumps) can start to function less efficiently or
even fail entirely due to a number of reasons. Causes could include:
• electricity failure (power station, transformer station, cables etc.)
• mechanical or electrical defects of the pump(s)
• breakage, tearing and leakage of the pipes
• blocked pipes (for instance due to iron oxide deposits) and/or filters encasings (filteromstortingen) and
blocked perforations in the casings (see Figure 47-3).
• air in the drainage pipes or suction of air via the pumps. The air can accumulate in high points of a
non-level discharge pipe, increasing the resistance.

Figure 47-3 Blocked pipe used during construction of the Haringvliet discharge sluice in the building pit

355
Manual Hydraulic Structures Part IV 'Structures'

Reduced efficiency or even failure can have the following consequences:

• in an unconfined aquifer:
the excavation can be partly or entirely flooded, which will lead to a work delay and possibly to
damaged equipment, certainly as regards electrical or electronic items.
• in a confined aquifer:
as for an unconfined aquifer, but also bursting open of the bottom of the excavation, in case of
an aquitard (slechtdoorlatende laag) or impermeable aquifer. This can cause disturbance of
the foundation soil of the structure under construction. For an explanation of the bursting open
of a clay layer, see CTB2410 (Waterbouwkunde), section 3.7.3 "Opbarsten".

Depending on the extent of the possible damage, measures must be taken such as placing emergency
generators, which partially or fully take over the power supply in case of power failure, and double cables
on the building site. The examples above are based on the assumption that the building site uses power
from the national grid. However, if the building site is in an isolated location, it can be worthwhile to use
generators for the daily power supply. In this case one could also consider backup generators.
Depending on the extent of the possible damage and on the duration of the drainage, the following mat-
ters will also be necessary:
• regular inspection
• preventive maintenance
• monitoring operations
• emergency procedures for calamities, etc.

Particularly the last two points depend largely on the situation. In small projects stagnation of the drainage
needs not be very grave (though disruption of the foundation base is also serious in these projects).
However, in large projects it can be beneficial to register certain parameters continuously and to sound
alarms and trigger certain procedures automatically when certain limits are exceeded. Limits that could be
exceeded are the groundwater level measured in a well (upper limit), and the minimum drainage dis-
charge (lower limit). The alarm can be sounded in a permanently staffed area, in somebody’s home, close
to the site and with an occupant who is on guard duty outside normal working hours or by phone. Proce-
dures that can be initiated automatically include starting emergency generators if the grid fails.

Trouble for the surroundings

Often a building pit (with slopes and a dewatering system) leads to lower building costs than a coffer dam,
even if one includes the possible damage to the surroundings. Disadvantages of the building pit construc-
tion method can be the consequences of dewatering on the surroundings and the large use of space (the
slopes require a lot more space than is necessary for the completed structure, which can cause problems
in built-up areas and if the new structure is to be built next to the an existing structure). Very often de-
watering causes no damage at all or only limited damage.

In agricultural areas, dewatering dehydrates soil, thereby ruining crops. Nature reserves and recreational
areas can also suffer badly from a lowered groundwater level. Sometimes the surface layer will “retain”
sufficient amounts of precipitation to counter the effect of the lowered groundwater level. Furthermore,
precipitation wells for agriculture and horticulture can dry up.

Another important point is the possible settlement as a result of the temporary lowering of the groundwa-
ter level. This causes an increase of the effective soil pressure (see part II Loads, Soil, Groundwater). In
soil types that are susceptible to this, it will lead to compression of the soil and settlement of the ground
level, as well as to an increase of negative shaft resistance on piles. If houses or buildings settle evenly
across the whole of their foundation, this does not have to result in damages. It can be worse if settle-
ments are uneven for one building. The breaking of unfounded sewer pipes and cables at the point where
they connect to a house or building on a pile foundation is not uncommon. Wooden piles can start to de-
cay above groundwater level.

Even settlements of the ground level in agricultural areas are not a problem as long as the water level in
the ditches can be adjusted accordingly. In old agricultural areas with considerable differences in ditch
levels this adjustment is not always possible, because lowering of the ground level could lead a decrease
of the distance between groundwater and ground level, which can result in a permanent reduction of crop
productivity (however, if the distance between ground water and ground level was too high, productivity
could be increased). Uneven settlements can be a nuisance for sowing and harvesting. Drains can get an
unfavourable slope.

356
Manual Hydraulic Structures Part IV 'Structures'

In water extraction areas dewatering is usually forbidden, because, as a rule, drainage water is pumped
out to open water and is therefore lost. Return pumping can offer a solution: the water pumped out of the
site is returned to the ground through injection wells, at some distance from the building site. During the
construction of the tunnel under the Dordtse Kil, for example, a large amount of drainage water was sup-
plied directly to the water works. The water was taken from a soil layer that was normally not used for
drinking water extraction, but with sufficient quality. The extraction from deeper layers could be reduced
during the construction of the tunnel.

Sometimes the water acquired by means of drainage is of bad quality and may not be discharged into
open water. Drained water in coastal areas and deep polders can be salty, which could cause open water
to become brackish (this was a problem for the construction of the Schiphol tunnel near Amsterdam). In
that case separate discharge is required to protect agriculture, horticulture or livestock industry. Moreo-
ver, people in the past few years have become aware that groundwater flow (however slow) can lead to
unwanted diffusion of harmful substances from toxic waste dumps.

Once again: these problems can occur although they often don’t. If damages occur they will have to be
compensated by the construction company. Good observation (groundwater levels, height measure-
ments, descriptions of buildings and crops) before, during and after dewatering is necessary to identify
the damages. Photos can be used to describe houses and buildings. In fact, in the Netherlands this is
prescribed: an independent committee makes an estimate of the expected damage. Constructors gener-
ally buy off the risk.

47.2 Design
Groundwater can be divided into two types: groundwater in an unconfined aquifer and groundwater in a
confined aquifer. In an unconfined aquifer the flow profile equals the piezometric height. The groundwater
level is therefore free. In a confined aquifer the flow profile is determined by an impermeable layer. A bad-
ly sealed layer that leaks, is called a semi-confined aquifer. The water level can be above or below the
top of the aquifer (overspannen of onderspannen water).

Unconfined aquifer
A drainage well will affect the phreatic surface in a certain volume of permeable soil around it. The phreat-
ic surface drops gradually until a stationary situation is achieved. In an ideal situation, the affected soil
body has a cylindrical shape with radius R, see Figure 47-1. The subsoil is impermeable, in contrary to
the soil that surrounds the soil body.

Figure 47-1 Dewatering of groundwater in an unconfined aquifer in a cylindrical soil body

The discharge of the well depends on the slope of the groundwater level and the permeability of the soil,
which is expressed in Darcy’s law:

q = k ⋅i

in which:
3 2
q [m /s/m = m/s] = specific discharge in the soil body
k [m/s] = coefficient of permeability (see Table 47-2)
dh
i [-] = slope =
dr

357
Manual Hydraulic Structures Part IV 'Structures'

For the total discharge Q, we can write Q = q ⋅ F [m /s], where F is the circumference area of the soil
3

body: F = π ⋅ d ⋅ h = π ⋅ 2r ⋅ h [m ].
2

For the well this means:

dh
Q = k ⋅i ⋅F = k ⋅ ⋅ 2π r h
dr
or
Q dr
2 h dh = , so:
π k r
Q 1
2∫ h dh
πk ∫ r
= dr ,

Q
where the solution is: h 2 = ln r + C
πk

Q
If r = R (and h = H) then: H 2 = ln R + C
πk

If we subtract the equations from one another:

(ln R − ln r ) = − ⋅ ln  
Q Q r
H 2 − h2 =
πk πk R

In practice it is customary to determine the maximum allowable water level first, to avoid that water will
enter the building pit. The phreatic water level should be at least 0,50 m below the pit floor. (This applies
to all places on the construction site to avoid problems with saturated soil, freezing and the accumulation
of rainwater; see Section 4.1 ('Construction Pit') of the general lecture notes of the course "Hydraulic
Structures 1"). Knowing this, the value of the acceptable groundwater height h can be affixed. With h, the
required total discharge can be calculated if we rewrite the previous equation:

π k (H 2 − h 2 )
Q=−
r 
ln  
R 

Confined aquifer
Similarly, we can derive the following relation for the groundwater in a confined artesian aquifer, in the
permeable layer shown in Figure 47-2:

Figure 47-2 Dewatering of confined aquifer in a cylindrical soil body (artesian well)

358
Manual Hydraulic Structures Part IV 'Structures'

dh Q Q 1 Q
Q = k ⋅ 2π r ⋅ D ⋅ → dh = dr → ∫ dh = 2π kD ∫ r dr → h = ln r + C ,
dr 2π krD 2π kD

Q
and also: H = ln R + C
2π kD

Q r 
Subtraction yields: H − h = − ln   ,
2π kD  R 

2π kD (H − h )
from which the discharge can be derived: Q = −
r 
ln  
R

This equation can be used when the maximum allowed piezometric height h has been determined with
respect to the bursting open criterion (an impermeable layer should not burst open because of too high
water pressure from the layer underneath).

Semi-confined aquifer
Usually the sealing layer is not entirely impermeable and the water is in a semi-confined aquifer. Figure 6-
14 shows a common profile: two very permeable layers separated by an almost impermeable layer of clay
with thickness d. We assume that any withdrawal of water from the bottom layer does not cause a reduc-
tion of the groundwater level in the top layer (for instance, this is realised by polder ditches in the top lay-
er). This means that the bottom layer is fed by the top layer through the layer of clay.

piezometric height
upper layer

piezometric height
lower layer

almost impermeable

Figure 47-3 Dewatering of semi-confined aquifer in a cylindrical soil body

According to Darcy, this supply q equals

H-h
q = k '⋅ i = k ' ⋅
d

per unit of area, in which k' is the vertical coefficient of permeability of the layer of clay. In this situation
the following relation can be derived (see also lecture notes Soil Mechanics by Verruijt / Van Baars):

Q
H −h = − K 0 ( r λ)
2π k D
in which:
K0 ( r λ ) = a modified Bessel-function of the first type, order zero.
(see following table, or see lecture notes Soil Mechanics; appendix A , or see
Abramowitz & Stegun: 'Handbook of mathematical functions').
λ = the leakage factor = k D C [-]
d
C = the resistance of the clay layer = [s] or [d]
k′

359
Manual Hydraulic Structures Part IV 'Structures'

r λ 0,1 0,5 1,0 1,5 2,0 2,5 3,0 4,0

K0 ( r λ ) 2,43 0,92 0,42 0,21 0,11 0,062 0,035 0,011

Table 47-1 Values for the Bessel function

Schematization effects
The formulas above have all been derived for so-called perfect wells, i.e. their filters run along the full
height of the layer from which water is extracted. This will not always be the case. In an unconfined aqui-
fer the following applies: If the bottom of the well does not coincide with an impermeable layer the formula
for unconfined aquifer is still used, as it has been shown that the water beneath the bottom of the well
hardly participates in the flow (i.e. assume that the impermeable layer does not extend beyond the bottom
of the well). If the impermeable layer is very deep, Sichardt advises us to increase the calculated dis-
charge by 20%.

The formulas have been derived for a cylindrical volume of soil, not an everyday situation. In reality the
excavated site that needs to be dewatered is usually on land. R is understood to be the distance from the
well to a point where the groundwater level reduction approaches zero. For small dewatering systems, R
is sometimes approximated with Sichardt’s empirical formula:

R = 3000 (H - h) k (units in metres and seconds).

For a first rough estimation, the following table gives an overview of order of magnitude of the permeabil-
ity k, it is certainly is not absolute:

Soil type k [m/s]

-2
gravel 10
-3
coarse sand 10
-3 -4
moderately coarse to moderately fine sand 10 to 10
-4 -5
fine sand 10 to 10
-9 -11
clay 10 to 10
Table 47-2 Permeability

For large scale dewatering one uses experience or pump tests. The influence of a “wrong” estimate of R
on the discharge calculation is limited by the use of the natural logarithm of r / R.

The formulas are derived for one well in the centre. In practice a number of wells will be placed around
the building site. In this case the principle of superposition applies, i.e. the water level reduction in any
one point is the sum of the reductions caused by each individual well. In an unconfined aquifer this is not
the case, as this requires the superposition of the differences of the squares of H and h.

The formulas are derived for a stationary situation. Other formulas apply for the preceding non-stationary
phases. However, dewatering systems are usually designed for the stationary situation, though it is im-
portant to start pumping well in time because it takes some time to reach the stationary situation (and the
desired situation). Furthermore, the pump system needs to have sufficient overcapacity to deal with pre-
cipitation (a greater reduction of the groundwater level than technically necessary can create a buffer for
extreme rainfall) and the possible failure of one of the pumps.

Certain parameters, such as the coefficient of permeability k, can be found by carrying out laboratory
tests on undisturbed samples. However, they are usually not representative for the extensive area in
which the groundwater flow occurs. It is therefore desirable, if not indispensable, to carry out a pump test;
a dewatering well surrounded by a number of measuring wells. The pump test can determine parameters
such as the coefficient of transmissibility kD (usually in m /day) and the leakage factor λ ([m].
2

Influence of supply borders

Building sites as a rule are close to waterways (rivers, canals), which are to take on the hydraulic engi-
neering work some time later. This waterway can act as a supply for the construction pit. The water level
in the waterway (H) will not be affected by the dewatering, so the piezometric height at this supply border
remains the same. Mathematically this can be approximated by introducing a fictitious “mirror well” (i.e., a
supply) on the same distance from the waterway as the construction pit, but on the opposite side of the
waterway (see Figure 47-4).

360
Manual Hydraulic Structures Part IV 'Structures'

Figure 47-4 Schematisation of the mirror well principle

So, draw a line from the dewatering well perpendicular to the supply border and project an injection well
on the opposite side of the border, which injects as much water as the first well discharges, on this line.
By making the distances to the border (l) the same, the piezometric height remains the same along the
border line (middle of the waterway), provided it is straight, so (h2 + h1)/2 = H. The reduction of the water
level caused by the drainage well is compensated by the injection of water in the other well. With both of
these wells one can calculate the water level at any point on the side of the drainage well.

In case of a confined aquifer the following applies to a given point on the line between the well and the
fictitious injection well:

n
Q  rj  Q r  Q r 
h = H + ∑ 2π kD ln  R  = H + ln  1  −
2π kD  R 
ln  2 
2π kD  R 
j =1  

(derived from the formula for n wells).

Given that the distance from the drainage well -and hence also from the injection well- to the border is l,
and expressing r1 and r2 in r (r1 = r and r2 = 2l - r) this equation reads:

Q r  Q 2l −r  Q  r 
h = H+ ln − ln = H+ ln  
2π kD  R  2π kD  R  2π kD  2 l − R 

from which can be derived:

Q  r 2l − r 
H - h= −  ln − ln 
2π k D  R R 

For an unconfined aquifer the formula is:

Q  r 2l − r 
H 2 - h2 = − ln − ln
π k  R 
R 

Note that the supply border normally is not perfect. In wide, deep rivers (e.g. the Nieuwe Maas at Pernis
near Rotterdam, measuring approximately 600 m by 13 m) the influence of the dewatering system on one
side was still detectable on the other side. A layer of silt on the riverbed creates a large entry resistance,
which means the river does not qualify as a perfect supplying border: no reduction of the piezometric
height at the border as a result of dewatering. In that case the border should really be moved, from the
waterway to further inland, on the fictitious injection side. If the river has just been dredged, the layer of
silt will have been removed and the river can act as a perfect supplying border.

361
Manual Hydraulic Structures Part IV 'Structures'

48. Flood defences: The water-retaining height

New chapter: February 2016, improved: February 2018

48.1 Introduction
Flood defences are hydraulic structures that protect the hinterland from being covered by water coming
from oceans, seas, rivers, lakes and other waterways. There are four main types of flood defences:
• natural soil structures (dunes);
• artificial soil structures (dikes, dams);
• specific water-retaining structures (cofferdams, gravity walls, sheetpile walls, etc.);
• engineering structures (kunstwerken) (sluices, navigation locks, cut-offs, storm surge barriers, pump-
ing stations).
Water-retaining engineering structures are constructed for functions that are combined with flood defenc-
es, mainly as part of intersecting infrastructures. They are often equipped with movable closure means
(gates) (ENW, Grondslagen voor Hoogwaterbescherming, 2016, TAW, Grondslagen voor Waterkeren,
1998).

The main property of a flood defence is its height, which should be sufficient to retain design water levels
and reduce the discharge of overtopping waves to acceptable quantities. This implies that the flood de-
fence is sufficiently strong and stable to provide structural safety.

The height should be determined with respect to two types of limit states:
• serviceability limit state (SLS): the discharge of overtopping waves or overflow causes unacceptable
difficulties, or damage, without collapse of the flood defence structure. Consequences of exceedance
of the serviceability limit state are:
o the amount of water flowing into the hinterland is more than the capacity of the hinter-lying water
system, leading to flooding of the polder;
o pedestrians, vehicles and properties on and directly along the flood defence suffer too much from
water passing the flood defence.
• ultimate limit state (ULS): collapse of the flood defence causes flooding of the hinterland, leading to
unacceptable consequences. Flood consequences can consist of loss of lives, cultural and natural
values and such alike, and economic damage. Collapse can be caused by wave-overtopping or over-
flow leading to erosion or softening of the inner dike slope, but also by other failure mechanisms, such
as sliding-off of soil bodies (macro-instability), erosion of revetment, piping and horizontal shear of the
entire embankment.

There are several design approaches for flood defences:

• deterministic design (in the Netherlands used until 1960)
• exceedance probability approach (in the Netherlands used from 1960 to 2016)
• failure probability approach (in the Netherlands used since 2017)
These approaches are explained in the following sections.

48.2 Deterministic design

A deterministic design is the classic approach, where the construction level of (the top of) a dike is based
upon the highest observed water level, which was usually the water level that caused the last flood. An
extra height, the freeboard (waakhoogte), was added to account for wave overtopping and uncertain-
10
ties . The freeboard, was usually chosen between 0,5 and 1,0 m (excluding the surplus height to com-
pensate for settlement).

hconstr > h water + fb

where:
hconstr [m] = construction level of the flood defence
hwater [m] = highest observed water level
fb [m] = freeboard

10
Freeboard = waakhoogte in Dutch, which is sometimes erroneously translated into English as wake height and into Dutch
as vrijboord, which has a completely different meaning! In Dutch, waakhoogte is also known as kruinhoogtemarge, as propo-
sed in Voorschrift Toetsen op Veiligheid voor de tweede toetsronde 2001 - 2006.

362
Manual Hydraulic Structures Part IV 'Structures'

The other dike properties, like the materials of the core and cover, the angles of the inner and outer slope
and the presence of berms, were chosen based on experience (mainly 'trial and error'). The deterministic
method was applied for a long time, when it was not yet possible to make strength calculations, due to
lacking numerical insight into failure mechanisms and difficulties in estimating the governing loads. The
freeboard, a safety factor, was not based on the quantification of the uncertainties; therefore, with this
method, it was not possible to determine the extent of over-design (or under-design).

48.3 Exceedance probability approach

Exceedance probability approach in general

The performance of a flood defence is usually expressed by its reliability, which is the probability that the
flood defence fulfils its function. The required reliability of a flood defence should be such, that it complies
11
with an acceptable risk of flooding of the hinterland . The settling of the acceptable flood risk level (safety
level) is based on answering the question 'how safe is safe enough', because 100% safety cannot be
achieved.

Three criteria can be distinguished in settling the safety level:

• Economic risk: the investments in flood protection should balance the therewith obtained risk reduc-
tion;
• Individual death criterion (individueel risico): the probability that a human being, residing at one loca-
tion during one year, will die because of a flood;
• Multiple deaths criterion (groepsrisico): the probability that a number of people, residing at one loca-
tion during one year, will die because of a flood.

The most severe of these criteria should determine the safety level (if affordable). Once the allowable risk
level has been determined, the corresponding maximum allowable probability of flooding of the hinterland
can be found, if the damage caused by a potential flood can be estimated. The allowable flood probability
can then be related to the maximum failure probability of the flood defence, which in its turn is related to
an extreme water level with a certain exceedance probability (see Section 13.3). In this design approach,
flood defences are primarily designed to resist this design water level. Therefore, the height of a dike is
the basic design parameter. Other properties, like the geometry (shape) and materials, are then designed
to provide the flood defence with sufficient stability.

In general, the required height of a flood defence is mainly determined by the following two components:
• Design water level (maatgevende hoogwaterstand = MHW in Dutch )
12

• Maximum allowable discharge of wave overtopping or overflow

Figure 48-1 depicts the components that determine the height of a flood defence.

Figure 48-1 Determination of the construction level of a dike

The design water level, which is the water level under governing conditions, is based on the still water
level, which occurs if there would be no wind, no waves and no tides. To find the design water level,
several effects require extra height of the flood defence:

11
Flood risk is defined as the probability of flooding multiplied with the consequences of a flood.
12
In English, MHW = mean high water, but that is not the same as the Dutch MHW. This can be rather confusing!

363
Manual Hydraulic Structures Part IV 'Structures'

• Highest astronomical tide, spring tide (springtij). In most seas, the semi-diurnal range (the
difference in height between high and low waters over about half a day) varies in a two-week cycle.
Approximately twice a month, the Sun, Moon, and Earth form a line and the tidal force due to the Sun
reinforces that due to the Moon. The tide's range is then at its maximum.

• Wind set-up (opwaaiing). Water level changes along the coast by astronomical tides are explained in
Section 13.1 of this Manual. Wind fields can head up the water in shallow seas, deltas, closed-off
creeks and lakes, because of friction between wind and the water surface (Section 13.2). Wind set-up
is usually included in the estimate of the normative high water level by using the peak-over-threshold
method.

• Funnel-effect (trechtereffect). If the wind blows in a direction where the water surface becomes nar-
rower, the water level will be pushed up. This set-up is called the 'funnel effect'. Wind set-up and fun-
nel-effect are comprised in the term 'storm set-up.'

• Short-term atmospheric depressions (buistoten). When a storm with low air pressure approaches
the coast, it will cause an elevation of the sea water surface. It thus builds up an increasing large
wave. The duration of such a depression can vary between a couple of minutes and more than one
hour. If these depressions appear in a more or less regular series, the phenomenon is called 'oscillat-
ing depressions' (bui-oscillaties).

• Seiches (staande golven in waterbasins) are oscillatory elevations of the water level in a water basin
enclosed on three sides, for instance sheltered harbours and lakes. They are caused by oscillations of
the (sea) water level outside the basin, particularly by shower oscillations with periods ranging from a
couple of minutes to an hour. Typical periods of seiches are between 10 and 120 minutes.

• Tidal resonance (getijresonantie). Tidal waves in estuaries, bays or river mouths can possibly be re-
flected. Interference of the reflected wave with the incoming wave will then lead to an increase of the
water level, which has to be taken into account when determining the height of a flood defence.

• Climate change effects (sea level rise and increase of river discharges). An increase of the aver-
age temperature would cause thermal expansion of the ocean water, resulting in a water level eleva-
tion. The melting of ice masses on land (flowing into the oceans) has only a relative small effect on the
sea water level. It is also believed that the rivers will be influenced by climate change, because the
highest and lowest river discharges will become more extreme, due to more fluctuations of rainfall
over land. The combined effect of (absolute) sea level rise and land subsidence is indicated as ’rela-
tive sea level rise’. In addition, climate change can influence wind set-up, atmospheric depressions,
seiches and tidal resonance. Climate change effects should be considered over the design life time of
structures, in case of designs, or over the return period of the assessments, in case of assessments.

• Land subsidence (bodemdaling). Geological processes and gas and oil extraction result in a general
subsidence of deeper ground layers. The effect of land subsidence is often included in sea level rise,
and the combined effect is called 'relative sea level rise'.

A flood defence should be higher than the design water level, to account for waves and uncertainties. The
thus obtained height is the 'crest level' of the flood defence. To compensate for settlement, a 'surplus
height' should be applied on top of the crest level. This leads to the 'construction height'. The required
flood defence height higher than the design water level is called the 'freeboard'.

The effects that have to be included when determining the crest level of a flood defence are:
• Wave-overtopping height (overslaghoogte). This height is required to limit wave overtopping to ac-
ceptable amounts (Section 17.2). This mainly applies to coastal defences and flood defences along
the lower rivers. For the upper rivers, where there is no tidal influence, overflow is usually more critical
than overtopping. The height needed to reduce wave overtopping is known as the 'overtopping height'.
The waves can be caused by wind, or induced by vessels.

• Mathematical surplus heights:

Length effect. The failure probability of a long dike is higher than the failure probability of a short
dike because of partial correlation or independence between different cross-sections and/or struc-
tural elements. An additional height compensates this effect.

364
Manual Hydraulic Structures Part IV 'Structures'

Robustness surcharge (robuustheidstoeslag). For uncertainties regarding changes in the estima-

tion of water levels, changes in prescribed calculation principles and changes in strength criteria, a
robustness surcharge is used as compensation (TAW, TR Ontwerpbelastingen voor het Rivi-
erengebied, 2007). The aim of adding this surcharge is to cope with eventual setbacks during the
lifetime of a structure.
Two main types of robustness uncertainties can be distinguished:
• Technical uncertainties:
– magnitude of discharges
– modelling uncertainties (water levels and wave parameters)
– climate change scenario uncertainties
– statistical uncertainties in load parameters
• Policy uncertainties

The following surplus heights for geotechnical effects have to be taken into account to find the initial con-
struction level (if applicable):
• Settlement. Soil gradually settles because of extra loading, namely by the weight of the dike. This has
to be compensated by a surplus height. Compaction can occur in the subsoil (zetting) and the em-
bankment itself (klink). The rate of compaction depends on the development of the consolidation in
time, which indicates how much water has already been expelled. In the Netherlands, mainly two
methods are used in engineering practice: Koppejan (modified) or Terzaghi/Bjerrum (Section 34.2).
For a very rough indication, when no data is available, as a rule of thumb, a value of 10% settlement
could be used for clay, or 5% if the clay is carefully applied and densified. For sand, a value of 5%
could be used, but if it is well compacted, settlement of sand can be neglected.

• Earthquake surplus height. Earthquakes induced by seismic movements and by extraction of natural
resources can cause deformation of dikes, next to subsoil subsidence. This could cause additional
lowering of the crest by compaction of the dike core, squeezing of softened soil (zijdelings wegpersen
van verweekte grond), or macro-instability induced by an earthquake (rotational slip of the slope).
Usually, advanced calculation methods are used to compute the magnitude of these effects and re-
search is still in progress. These calculation methods are not included in this Manual.

The term 'freeboard' is not uniformly defined, so the exact definition varies per guideline. The freeboard
is more or less defined as the difference between construction level and (local) assessment level (for as-
sessment purposes) or between construction level and design water level (for design purposes). Often, a
minimum value of 0,50 m is prescribed for the freeboard. This minimum freeboard usually applies to the
crest level, not the construction level!

Exceedance probability approach in the Netherlands

In the Netherlands, the target safety level, based on an exceedance probability approach per dike ring,
was used in internal instructions of Rijkswaterstaat since 1960, and was stipulated in the Act on the Flood
Defence since 1996 (Wet op de Waterkering) and since 2009 in the Water Act (Waterwet). The safety
level used in the Dutch exceedance probability approach was mainly based on economical considera-
tions, but imponderable values were taken into account in the safety philosophy as well.

In the exceedance probability approach, the design water level is based on a series of water level meas-
urements. The water level appeared to be related to the probability of occurrence: the occurrence of low
water levels is very likely, while storm surge levels are very rare (see Section 13.3 of this Manual). To
protect against extreme circumstances, the measurement points have to be extrapolated to find the water
level that corresponds to a probability of occurrence that is still considered acceptable. The Delta Com-
mittee reasoned that for the central part of Holland, the probability of exceedance of an extreme water
level during a random year, leading to failure of a flood defence, should not be more than 1/10 000. This
exceedance probability appeared to correspond to a water level of NAP + 5,00 m near Hoek van Hol-
13
land . Since 1 January 2017, the policy changed to a failure probability approach per dike segment
(change of the Water Act).

The normative high water level (NHW) along the coast is determined at locations where measurements
have been carried out for many years, which have been extrapolated to extreme circumstances. The
normative high water level is found by relating the water levels to acceptable exceedance probabilities
(see also Section 13.3). Along rivers, the normative high water level is determined with help of computer
models that calculate the distribution of river discharges over the river branches. The thus determined

13
The Delta Committee followed three ways of reasoning to find the basic water level of NAP+ 5,00 m at Hoek van Holland, which are ex-
plained in Voorendt, 2017.

365
Manual Hydraulic Structures Part IV 'Structures'

critical water level along coasts and rivers is the assessment level (toetspeil), as listed in the book Hy-
draulische Randvoorwaarden (Hydraulic Boundary Conditions) of Rijkswaterstaat. This book is used to
evaluate the reliability of existing flood defences.

NHW along the coast

The measurements along the coast include:
• astronomical tides,
• wind set-up, including a possible funnel effect (trechtereffect).

NHW along rivers

The governing water level along river dikes can be based on the design river discharge for the location
where it enters the Netherlands (like Lobith for the Rhine) and then making a discharge computation for
locations further downstream. The levels in the lower rivers (benedenrivieren) are influenced by both river
discharge and tidal influences from sea. Control objects like weirs and discharge sluices can influence the
water level in lower and upper rivers (bovenrivieren), so these effects have to be taken into account as
well, as they are not included in the normative high water levels. Usually, software like Hydra-NL is used
to compute the NHW at specific locations.

Below, several values are given for the Netherlands of the effects that determine the crest level of a flood
defence, to give a first idea of their magnitude. In addition, some information is given how these effects
are dealt with in the guidelines that are used for the design and assessment of flood defences.

• Short-term atmospheric depressions. The water level elevation caused by short-term atmospheric
depressions along the Dutch coast can be up to 0,50 m, or even more, but it is not very likely that this
coincides with extreme water levels, so usually only a reduced effect of atmospheric depressions is
taken into account. This reduced value depends on the height needed to reduce wave run-up to ac-
ceptable amounts. The maximum value is 0,30 m along the Dutch coast, and only 0,05 m 20 km land-
ward (TAW, 1989). The Dutch guide on lake and sea dikes (TAW, 1999) has adopted these values.
In Dutch waters, the effect of single depressions dominates the effect of oscillations. Therefore, it is
assumed that the effect of the latter is discounted in the contribution of the single short-term depres-
sions (TAW, Grondslagen voor Waterkeren, 1989).
Because of the presence of the Hartelkering and the Maeslantkering, short-term atmospheric depres-
sions have become much less relevant behind these barriers. The Dutch committee on flood defence
(TAW) therefore recommended to neglect these effects behind the barriers, until further research has
been carried out (TAW, TR Ontwerpbelastingen voor het Rivierengebied, 2007).

• Seiches. Because of the presence of the Hartelkering and the Maeslantkering, the importance of
seiches has become much less relevant behind these barriers, like short-term atmospheric depres-
sions. Seiches in the IJsselmeer, with wave heights up to 0,113 m, have been reported in Bottema,
2004

• Relative sea level rise. The average trend of relative sea level rise observed in the Netherlands since
1900 is 0,19 m/century (Dillingh, 2008) and a deviation from this trend has not yet been observed.
Forecast models, however, come with a wide range of results, up to 2 metre per century (and even
more) in the very near future. It is the task of engineers to find a way to deal to cope with these alter-
natives, ranging from adaptive to robust design strategies.
The relative sea level rise should be considered during the entire life time of the structure, or during
the reference period of assessments. The design life time of flood defences usually is usually 50 years
(for dikes if they can be easily adapted) or 100 years (for engineering works). The reference period for
the assessment of existing flood defences is much shorter, usually a few years.

• Land subsidence. The rate of subsidence varies per location in the Netherlands. Outside the Nether-
lands, the deeper ground layers can also gradually rise, dependent on the geological history of that
area. De Lange & Gunnink studied land subsidence in the Netherlands and produced the map as pre-
sented in Figure 48-2. Their results should be carefully dealt with, because of the uncertainties of the
assumptions.

366
Manual Hydraulic Structures Part IV 'Structures'

Figure 48-2 Land subsidence between 2011 and 2061 if the climate and ground water table would not change
(De Lange & Gunnink, 2011)

• Local wind set-up. The normative high water level includes the effect of wind set-up, but this has only
been estimated for certain measurement points along the coast. So, if there is local wind set-up that is
not included in the normative water level, this effect should be added.

• Mathematical surplus heights. The Dutch TAW-guidelines do not prescribe statistical and policy un-
certainties, because a useful estimation could not be made. For the Design Toolkit 2014 (Ontwerpin-
strumentarium 2014) a robustness surcharge of 0,30 m (for rivers) up to 0,70 m (for the Ketelmeer) is
recommended to account for modelling uncertainties (Rijkswaterstaat, Handreiking ontwerpen met
overstromingskansen, 2014). For the new Scheveningen boulevard, the following surcharges were
taken into account for a design life time of 100 years: 0,30 m for the length effect, extra wind set-up
due to more severe storms: 0,40 m, higher shower oscillations, also caused by more severe storms:
0,10 m. Water Board Rivierenland uses a robustness surcharge of 0,30 m for their primary river dikes
(unless a probabilistic model is used for calculation) (TAW, 1999a).

48.4 Failure probability approach

In a failure probability approach, the height of a flood defence is calculated as part of a probabilistic com-
putation, where the failure probability per dike section is determined by the failure probabilities per failure
mechanisms. Usually, only the main failure probabilities are considered, like erosion of the inner dike
slope due to overtopping waves, macro instability (sliding of soil bodies), piping or uplift, failure of revet-
ment or erosion of the outer slope, and non-closure of closing means. A dike section will fail if one, or
more, of these failure mechanisms occur. The allowable failure probabilities per failure mechanism are
derived from the required failure probability requirement per dike section by means of a standard distribu-
tion of failure probabilities over the main failure mechanisms (Figure 48-3).

367
Manual Hydraulic Structures Part IV 'Structures'

Figure 48-3 Fault tree of a dike section

In this approach, the height of a flood defence depends on the allowed probability that a critical overtop-
ping discharge is exceeded. However, it is allowed to deviate from the standard failure probability. This
implies that the height of a dike can also be lower than standard required, if the failure probability due to
overtopping is compensated by, for example, an lower failure probability due to macro-instability (slip of a
soil body), which can, for example, be obtained by a stability wall.

The determination of the height of a flood defence according to the failure probability approach is ex-
plained as part of the course of Flood Defences (CIE5314) and is not comprised in this Manual.

48.5 References

Bottema, M. (2004), Verrassend snelle golfgroei op het IJsselmeer, Meteorologica, maart

2004, p. 1519.

Lange, G. de and J.L. Gunnink (2011). Bodemdalingskaarten. Deltares, TNO.

Dilingh, D. (2008). Zeespiegelstijging langs de Nederlandse kust. Personal correspondence. Deltares.

Jonkman, S.N., R.E. Jorissen, T. Schweckendiek and J.P. van den Bos (2017). Flood defences. Lecture
notes Delft University of Technology.

Ministerie van Verkeer en Waterstaat (2007). Hydraulische Randvoorwaarden primaire waterkeringen

voor de derde toetsronde 2006-2011 (HR 2006). Ministerie van Verkeer en Waterstaat.

Rijkswaterstaat (2014a). Achtergrondrapport ontwerpinstrumentarium.

Rijkswaterstaat (2014b). Handreiking ontwerpen met overstromingskansen.

TAW (1998). Grondslagen voor waterkeren. Balkema uitgevers BV.

TAW (1985). Leidraad voor het ontwerpen van rivierdijken. Deel 1 - bovenrivierengebied. Staatsuitgeverij
’s-Gravenhage.

TAW (1985). Leidraad voor het ontwerpen van rivierdijken. Deel 2 - benedenrivierengebied. Staatsuitge-
verij ’s-Gravenhage.

TAW (1999). Leidraad zee- en meerdijken. Staatsuitgeverij ’s-Gravenhage.

TAW (2003). Leidraad kunstwerken. Rijkswaterstaat, Dienst Weg- en Waterbouwkunde.

TAW (2007). Technisch Rapport Ontwerpbelastingen voor het Rivierengebied.

Visschedijk, M.A.T., P. Meijers, R van der Meij, E.H. Chbab, G.A.M. Kruse, Zuada Coelho, D.S. Nugroho,
C.M. Wesselius, W.J. de Lange and G.A. van den Ham (Deltares) (2014). Groningse kades en dijken bij
geïnduceerde aardbevingen. Globale analyse van sterkte en benodigde maatregelen.

Voorendt, M.Z. (2017). Design principles of multifunctional flood defences. PhD report, Delft University of
Technology.

368
Manual Hydraulic Structures Part IV 'Structures'

49. Flood defences: Gates

There are many different types of closure mechanisms, namely:

• flat gate
• vertical lift gate
• drop gate
• sliding gate
• mitre gate
• single gate
• shutter weir
• visor dam
• radial gate
• sector gate
• inflatable dam

This chapter, however, only covers the action of forces in the following types:
• flat gate
• mitre gate
• radial gate
• sector gate
• arc

49.1 Flat gate

A flat gate (vlakke schuif of deur) transfers the load onto its rabbeted stops. These rabbeted stops can be
positioned both on the sides and on the top and bottom. If rabbeted stops are only positioned on the
sides, the total load must be transferred horizontally. This horizontal transfer of load can be accomplished
using various systems. The flat sluice can, for instance, be equipped with vertical girders that transfer the
load a number of horizontal rails, which, in turn, transfer the load to the supports. It is also possible to
mainly use rails that transfer the load to the supports in one go. The use of rails and/or vertical girders
depends highly on the shape of the gate or sluice and the position of the rabbeted stops.

crossbeam door girder door crossbeams and girders

Figure 49-1 Principle support system with rails and girders

The rails and girders can be constructed as a solid web girder, a Vierendeel girder, a box girder or a truss
(see also Figure 49-2). For small span girders (= c-c distance of horizontal rails), one can also use a so-
called trough girder.

369
Manual Hydraulic Structures Part IV 'Structures'

Structural shapes crossbeams

web plate beam half span

truss possible designs girders

Vierendeel girder

parabolic girder
standard truss solid web
profiles Vierendeel
Figure 49-2 Some possible structural shapes of a gate or sluice on double supports

The vertical girders outside of a gate that rest on the rabbeted stops are known as “posts”.

If the gate also transfers load to a sill and/or an upper rabbeted stop, the use of rails and girders is obvi-
ous. The transfer of loads and the forces and moments in the girders and rails of such a gate can be cal-
culated by schematising the structure as a beam grid. In such a calculation the torsional rigidity of the pro-
file must be estimated correctly, because this can have an important effect on the spread of the load. Par-
ticularly when applying box girder profiles or plates on two sides, this is of large importance.
An alternative to a gate with rails and girders is a gate with plates on both sides, where the rails and gird-
ers are plates that keep the main plates together and create a rigid profile. This structure is used in par-
ticular for sliding gates. The forces and moments in such a gate can be calculated using the three dimen-
sional finite elements method, where the gate is modelled with plates and discs.

A gate with plates on two sides has reasonable to large torsional rigidity. The result of this is that, in the
case of hinged supports on the sides and top and/or bottom of the gate, the gate will want to curl inwards
and will therefore start to leak. For this reason such a method of load transfer is avoided or is combined
with measures in the bottom rabbeted stops to counter the curling. Such a measure could be to insert a
jack in the gate, which fixes the corner of the gate. However, the edges of the structure have to be very
stiff to ensure that it is fully waterproof.

C ur li ng o f t he p la t e a s a r e s u lt o f
t he p la t e b e n d i ng i n t wo d i r e c t i o ns

Figure 49-3 Curling of a gate or sluice supported on three sides

The supports of the sluice can be single points or linear. For single point supports it seems obvious that
the supports should be positioned at the ends of the rails and girders. For linear supports, the posts need
to be sufficiently stiff to be able to spread the load from the girders over the entire rabbeted stop. The
same applies for a structure, supported on three or four sides, for the top and bottom rails.

370
Manual Hydraulic Structures Part IV 'Structures'

49.2 Mitre gate

In mitre gates the load is transferred to the side rabbeted stops of the gate. The bottom rail is generally
not considered as a support, because torsional moments will occur in the gate. Both outer vertical girders
of the mitre gate are called posts (“har”s, after the old Dutch word for hinge). The girder nearest the rota-
tional axis is called the rear post, whilst the girder at the loose end of the gate is called the front post. If
the gate is closed, the support of both vertical girders can be uniformly distributed or concentrated in two
or more points.

For an even spread transfer of the load at the rear post, it is necessary that there is some give in both
points of rotation. The lower hinge usually has a little give, because the pivot would otherwise have to be
set infinitely accurately, relative to the vertical rabbeted stop strip. The upper hinge needs more give, be-
cause, during closure, the gate can be pushed against the rabbeted stop in the sluice wall by the water
pressure. If there is no give available in the upper hinge, the reaction forces will have to be supplied by
both hinges. In this case, too, a small amount of give is allowed in the bottom hinge in order to be able to
lower the pivot socket over the pivot when the gate is put in place.

First, the gate is considered in its closed position, subjected to a load caused by the water pressure, due
to the head over the gate. This situation is shown in Figure 49-4. By regarding the mitre gate as a triple
hinge truss, symmetry considerations dictate that the forces in the contact point between the front posts
have to be in equilibrium i.e. equally large but aimed in opposite directions, perpendicular to the sluice
axis.

Figure 49-4 Forces and reactions on a closed gate

The reaction force S can be resolved in a force parallel to the gate (N) and a force perpendicular to the
gate (F). These reaction forces can be denoted as a function of the resulting force from the water W and
the angle α between the gate doors and the sluice axis (see Figure 49-4).

W
S=H =
2 ⋅ sin α

F=½W
W
N=
2 ⋅ tan α

were: S [kN] = reaction force of the sluice head

H [kN] = force from the other door (perpendicular to the length axis of the sluice or lock)
W [kN] = resulting hydrostatic force on the door
α [°] = angle between the gate door and the length axis of the sluice or lock

From the formulas one can deduct that if the angle α is small, the reaction force is very large. This advo-
cates a large angle α. However, if angle α is large, the gate doors become longer and hence the load
(W) becomes larger. Furthermore, longer gate doors would mean heavier gate doors as well as longer
gate chambers and lock heads. For economic reasons, one often usually chooses a ratio of tan α = 1:3.

The way the reaction forces S and H are provided depends strongly on the choice of supports along the
vertical sides. One could opt for a uniformly distributed load transmission or a load transmission with sup-
port points.

371
Manual Hydraulic Structures Part IV 'Structures'

Uniformly distributed load transfer

As mentioned earlier, a uniformly distributed load transfer, requires a certain amount of give in both hing-
es. The transfer of load at the lock head then takes place according to the linear loads N and F. Due to
structural considerations, one also opts for a uniformly distributed load transfer at the front post, there in
the direction of H. In this case, an adjustable gate is selected, in order to approximate a uniformly distrib-
uted load as well as possible.

Load transfer via point supports

To transfer the loads via point supports there may not be any give in the top hinge. At the rear post, the
hinges provide the reaction forces, which point in the direction of S. Two supports are also made in the
front post. It seems obvious to make two heavy horizontal rails between the four supports and to transfer
the load on the plates to these rails via girders (girder gate). For the height of these rails, an optimum be-
tween the moments in the girders and in the rear post is sought (see Figure 49-5).

Figure 49-5 Internal transfer of forces in the girder gate

Nowadays, the latter solution is employed often. Setting vertical rabbeted stop plains (for N and F) at the
rear post, which has to carry the load as evenly as possible, is very difficult. However, according to the
solution given in Figure 49-5 only two points have to be positioned accurately. Moreover, techniques have
been developed to equip the top hinge with ball bearings that tolerate barely any give and that can absorb
large forces.

Another load situation occurs when the door is open, in which case the hinges only need to be able to
take on the self-weight of the gate. The vertical load, the self-weight of the gate, is nearly always taken on
by the lower hinge. The moment that is thus created results in a tensile force in the upper hinge and a
compression load in the lower hinge (see Figure 49-6).

Figure 49-6 Balance of forces for an open gate

As the length of the gate increases, the gate is more and more inclined to start to sag (zakken) at the
loose end, due to its self-weight, so-called tilting. This problem especially affects wooden adjustable
gates. This is because they are not very rigid. Tilting can be avoided by creating rigid triangles. Starting
from the hinges one can create a triangle by placing a compression bar.

372
Manual Hydraulic Structures Part IV 'Structures'

compression bar

tension bar

Figure 49-7 Diagonal compression and tension bar

Further structural measures include placing a tension bar and applying planks in the direction of the com-
pression bar.
Steel mitre gates can be made sufficiently structurally stiff, hence the tilting problem is no longer relevant.

49.3 Water pressure on radial gates

See also Chapter 9 of part II of this manual for hydrostatic loads.

One distinguishes between radial lock gates with horizontal and vertical axes.

Radial gate with a horizontal axis

Figure 49-8 gives an example of a radial lock gate with a horizontal axis. The moving parts are not in-
cluded. Basically, the radial lock gate consists of a gate, which can be built up out of beams and girders
with plates on the high water side, and an arm that transfers the forces on the gate to the hinge. It is im-
portant that the curves of the gate correspond to a circle and the hinge is placed in the middle of this cir-
cle.
The transfer of the loads on the gate is completely dependent on the type of load. Hence, such the (par-
tially immersed) self-weight of the structure is carried by both the sill and the hinges, but the load caused
by the water pressure is taken on entirely by the hinges.
g ird e rs

tru s s
b e a rin g a rm

P la n F ro n t v ie w

m a in g ird e rs

C ro ss -s e c tio n
Figure 49-8 Radial gate with a horizontal axis

A special transfer of load to the hinge occurs if the gate is equipped with watertight plates on two sides,
or if the rear side of the gate is not circular. Figure 49-9 gives an example of such a gate. The horizontal
component of the load remains the same but the vertical component is anomalous and is wholly deter-
mined by the lift force of the hatched areas shown in Figure 49-9. In this case, the resultant of the water
pressure does not pass through the hinge. However, the resultant can be resolved into a force through
the hinge (water pressure, high water side) and a force in another direction (water pressure, low water
side). In the calculations it is sometimes also useful to combine the water pressure on the side of the low
water level with a part of the self-weight, so this combined load also passes through the gate’s hinge.

373
Manual Hydraulic Structures Part IV 'Structures'

Cross-section Uplift force

Figure 49-9 Radial gate with water proof plating on both sides

To calculate the load on the gate with the line of action through the hinge, the structure can be schema-
tised as a two-dimensional framework.

Depending on the shape of the arms and the way the arms are attached to the gate, a force parallel to the
hinge axis is applied to the hinge, as well as a force perpendicular to its axis. Figure 49-10 shows this for
a structure in which the gate is hinged to the truss arms. The line of action of the resultant load on the
axis of the hinge passes through two structural hinges. In the case of a small angle between the arms and
the perpendicular of the axis and a load perpendicular to the gate, this results in a small force parallel to
the axis is. The downside of the given structure is that it is not stable for a load parallel to the gate. Such
a load can occur if the gate is open and there is a side wind. That is why such a system requires guid-
ance equipment for the gate that offers sufficient lateral support to guarantee stability.

line of action hinged connection

reaction force

Figure 49-10 Force parallel to the axis of the hinge in the case of hinged arms

If the truss arms are fixed to the gate (not hinged), one creates a stable system in the shape of a portal
frame (see Figure 49-11). In this case, for a load q perpendicular to the axis, the force F// parallel to the
axis of the hinge will be larger than with hinged arms. This parallel force is caused by the deflection of the
radial gate. This generates an outwards rotation of the arms, which is prevented in the hinges.
The relationship between the forces perpendicular and parallel to the axis of the hinge depends on the
stiffness of the truss arm and the gate. The stiffer the arms, the larger the load parallel to the axis. The
theoretical maximum parallel load caused by a load perpendicular to the wall is:

L
F// = F⊥
4R

However, this can only occur with infinitely stiff truss arms.

374
Manual Hydraulic Structures Part IV 'Structures'

Figure 49-11 Force parallel to the axis of the hinge in the case of fixed arms

Radial gate with a vertical axis

In the case of a radial gate with a vertical axis, the gate is curved in the horizontal plane to an arc with a
centre that coincides with the position of the hinge. Thus the resultant of the water pressure load goes
precisely through the hinge in the horizontal plane. The direction of the resultant depends on the length-
wise profile of the gate. This is apparent in Figure 49-12, which shows two different profiles.
symmetry axis

symmetry axis

hinge foundation hinge foundation

resultant
resultant (uniformly
distributed)
resultant
(trapezium)

outside water level

inside water level

rectangular profile profile with submerged slopes

Figure 49-12 Direction of the resultant water pressure on a radial gate with a vertical axis

In the vertical plane, the direction and the position of the resultant of the water pressure depends on the
shape of the gate in the vertical cross-section. Figure 49-13 shows a number of possible cross-sections
and the forces on the gate per metre. The position of the seal with the bottom is an important detail which
determines the size of the upward force.

375
Manual Hydraulic Structures Part IV 'Structures'

Horizontal water pressure Uplift water pressure

Figure 49-13 Loads dependent on the shape of the gate

A force with a line of action that does not go through the hinge can be replaced with a force through the
hinge plus a moment. An example of this is given in Figure 49-14. The moment is taken on by the system
as for a beam on two supports.

Figure 49-14 Schematisation of the load from the gate on the arms

49.4 Water pressure on sector gates

The sector gate transfers its loads directly to the sill. The line of action of the resultant of the water pres-
sure on the plates on the side of the high water level goes through the hinge. The water pressure on the
plates on the side of the low water level causes a moment relative to the hinge. This moment is in equilib-
rium with the moment on the hinge as a result of the self-weight of the gate and the possible reaction
force created by a rabbeted stop.

with or without rabbeted stop

without stop

with stop

Figure 49-15 Action of forces in a sector gate

If, on the low water side, the gate is not only subject to hydrostatic loads, but also to dynamic loads such
as waves, one will always opt to include a rabbeted stop for the gate. The water level in the gate is then
kept high enough to ensure that there is always a force present in the rabbeted stop.
Waves on the high water side do not influence the position of the gate, because the resultant water pres-
sure goes through the hinge.

376
Manual Hydraulic Structures Part IV 'Structures'

49.5 Water pressure on arcs

Arcs that are subjected to a load on all sides (fluid or gas pressure) do not develop any shear stresses or
moments, only normal forces. One can make use of this property to optimise the use of materials. Con-
crete, for instance, will be used in a compression arc (dam) and steel will be used in a tension arc (visor
dam).

The transfer of load in an arc is relatively simple because the direction of the forces in the supports is
known; these naturally follow the tangent of the circle. This is shown in Figure 49-16.

valley

dam

reservoir

Figure 49-16 Load transfer in a circular arc

377