Directoraat-Generaal Rijkswaterstaat

Directie Zeeland

jcour
Nummer: » I

Bibliotheek, Koesir. 30, tel: 0118^86362

postbus 5014, 4330 KA Middelburg

Manual
’••T5
%

’ ■ 'a itj , -
¥'■4

f ,

GJ.C.AA. Hoffmans and HJ. Verheij

■nr.r rTT'sSRmwwcBffr’"'^
SCOUR MANUAL
SCOUR MANUAL
G. J.C.M. HOFFMANS
Road and Hydraulic Engineering Division of the Ministry of Transport,
Public Works and Water Management, Delft

H. J.VERHEIJ
Delft Hydraulics, Delft

A. A. BALKEMA / ROTTERDAM / BROOKFIELD / 1997

Scour Manual, P-DWW-96-120

Editorial Board
G.J.C.M. Hoffmans, Senior Research Engineer, Road and Hydraulic Engineering Division of the
Ministry of Transport, Public Works and Water Management
W. Leeuwestein, Senior Research Engineer, Road and Hydraulic Engineering Division of the Ministry of
Transport, Public Works and Water Management
K. W. Pilarczyk, Manager Research and Development of the Road and Hydraulic Engineering Division of
the Ministry of Transport, Public Works and Water Management
G. J. Schiereck, Senior Scientific Officer, Delft University of Technology
H. J. Verheij, Senior Project Engineer, Delft Hydraulics

Contributors
G.J. Akkerman, Senior Project Engineer, Delft Hydraulics
J. van Duivendijk, Senior Scientific Officer, Delft University of Technology
W H.G. Klomp, Project Engineer, Delft Hydraulics
M. van der Wal, Senior Project Engineer, Delft Hydraulics
R.O.Th. Zijlstra, Project Engineer, Haskoning

The Ministry of Transport, Public Works and Water Management and Delft Hydraulics and those
associated with this report have exercised all possible care in compiling and presenting the information
contained in it. This information reflects the state-of-the-art at the time of publication. Nevertheless, the
possibility that inaccuracies may occur in this publication cannot be mled out. Anyone wishing to use the
information in it will be deemed to do so at his or her own risk. The Ministry of Transport, Public Works
and Water Management and Delft Hydraulics decline to accept any liability whatsoever in respect of loss
or damage that may arise in consequence of such use on behalf of all persons associated with this
publication.

Cover: Ingrid La Haye, 1994, Scouring, 28 x 18 cm. The Hague

Authonzation to photocopy items for internal or personal use, or the internal or personal use of speeific
clients, is granted by A.A.Balkema, Rotterdam, provided that the ba.se fee of US$1.50 per copy, plus
US$0,10 per page is paid directly to Copyright Clearance Center, 222 Rosewood Drive, Danvers,
MA 01923, USA. For those organizations that have been granted a photocopy license by CCC, a separate
system of payment has been arranged. The fee code for users of the Transactional Repotting Service is:
90 5410 673 5/97 US$ 1.50 + US$0.10.

Published by
A.A.Balkema, RO.Box 1675, 3000BR Rotterdam, Netherlands (Fax: -t-31.10.4135947)
A. A. Balkema Publishers, Old Post Road, Brookfield, VT 05036-9704, USA (Fax: 802.276.3837)
ISBN 90 5410 673 5
© 1997 A.A.Balkema, Rotterdam
Printed in the Netherlands
Contents

FOREWORD XI

LIST OF MAIN SYMBOLS XIII

1 INTRODUCTION 1
LI General I
1.2 Scope of this manual 2
1.3 Summary 3

2 BASIC CONCEPTS 7
2.1 Introduction 7
2.2 Design process 8
2.2.1 General 8
2.2.2 Boundary conditions 8
2.2.3 Fault tree analysis 12
2.2.4 Protective measures 13
2.2.5 Characteristics of bed protection 14
2.3 Initiation of motion 15
2.3.1 General 15
2.3.2 Non-cohesive sediments 15
2.3.3 Cohesive sediments 19
2.3.4 Turbulence 22
2.3.5 Examples 24
2.4 Scour processes 25
2.4.1 General 25
2.4.2 General scour 26
2.4.3 Local scour 29
2.4.4 Conditions of transport 31
2.5 Geotechnical aspects 33
2.5.1 General 33
2.5.2 Porosity of sand 34

V
Vt Contents

2.5.3 Empirical criteria 34

2.5.4 Example 39

3 SILLS 40
3.1 Introduction 40
3.2 Flow characteristics 40
3.3 Dutch scour depth modelling 42
3.3.1 General 42
3.3.2 Characteristic time 46
3.3.3 Relative turbulence intensity 47
3.3.4 Scour coefficient 49
3.3.5 Non-steady flow 51
3.3.6 Upstream supply of sediment 53
3.4 Upstream scour slopes 57
3.4.1 General 57
3.4.2 Hydraulic and morphological stability criterion 57
3.4.3 Undermining 59
3.5 Additional measures 59
3.6 Field experiments 60
3.6.1 General 60
3.6.2 Hydraulic and geotechnical conditions 61
3.6.3 Discussion 61
3.7 Examples 65

4 JETS 68
4.1 Introduction 68
4.2 Flow characteristics 68
4.2.1 Introduction 68
4.2.2 Flow velocities 69
4.2.3 Hydraulic jump 70
4.3 Time scale 72
4.4 Plunging jets 73
4.4.1 Introduction 73
4.4.2 Calculation methods 73
4.4.3 Discussion 76
4.5 Two-dimensional culverts 78
4.5.1 Introduction 78
4.5.2 Calculation methods 79
4.5.3 Discussion 81
4.6 Three-dimensional culverts 82
4.6.1 Introduction 82
4.6.2 Calculation methods 83
4.6.3 Discussion 84
 Contents VII

4.7 Scour control 85

4.8 Examples 87

5 ABUTMENTS AND SPUR DIKES 91

5.1 Introduction 91
5.2 Geometry characteristics 92
5.2.1 Introduction 92
5.2.2 Wing-wall abutments 93
5.2.3 Spill-through abutments 94
5.2.4 Vertical-wall abutments 94
5.2.5 Flow pattern 95
5.3 Time scale 96
5.4 Equilibrium scour depth 100
5.4.1 Introduction 100
5.4.2 Calculation methods 100
5.4.3 Discussion 104
5.5 Types of scour 105
5.5.1 Introduction 105
5.5.2 Combined local scour and bend scour 105
5.6 Failure mechanism and measures to prevent local scour 106
5.6.1 Introduction 106
5.6.2 Scour slopes 106
5.6.3 Outflanking 107
5.6.4 Riprap protection 107

6 BRIDGE PIERS 109

6.1 Introduction 109
6.2 Characteristic flow pattern 109
6.3 Time scale 110
6.4 Equilibrium scour depth 113
6.4.1 Introduction 113
6.4.2 Calculation methods 114
6.4.3 Discussion 115
6.5 Effects of specific parameters 117
6.5.1 Introduction 117
6.5.2 Pier shape 118
6.5.3 Alignment of the pier to the flow 118
6.5.4 Gradation of bed material 120
6.5.5 Group of piers 120
6.6 Scour slopes 122
6.6.1 Introduction 122
6.6.2 Single cylindrical pier 122
6.6.3 Other types of piers 123
VIII Contents

6.7 Measures to prevent local scour 125

6.7.1 Introduction 125
6.7.2 Riprap protection 125
6.7.3 Mattress protection 126
6.7.4 Deflectors 126

7 COASTAL AND OFFSHORE STRUCTURES 128

7.1 Introduction 128
7.2 Pipelines 129
7.2.1 Introduction 129
7.2.2 Development regimes 130
7.2.3 Time scale 132
7.2.4 Equilibrium scour depth 132
7.2.5 Methods to prevent scour 135
7.3 Piles 135
7.3.1 Introduction 135
7.3.2 Equilibrium scour depth 136
7.3.3 Large diameter piles 137
7.3.4 Scour protection 139
7.4 Vertical breakwaters 139
7.4.1 Introduction 139
7.4.2 Time scale 139
7.4.3 Equilibrium scour depth 140
7.4.4 Influence of a protective layer 141
7.5 Coastal defences 142
7.5.1 Introduction 142
7.5.2 Type of processes 143
7.5.3 Effects of structures 143
7.5.4 Time scale 143
7.5.5 Equilibrium scour depth 145
7.5.6 Scour protection 148
7.6 Gravity structures 151
7.6.1 Introduction 151
7.6.2 Time scale 151
7.6.3 Equilibrium scour depth 152
7.6.4 Dishpan scour 152
7.6.5 Measures to mitigate scour 152
7.6.6 Uncertainties in scour predictions 155

8 PROTOTYPE SITUATIONS 156

8.1 Introduction 156
8.2 Lengthening of groynes in the Waal River, the Netherlands 157
8.2.1 Introduction 157
 Contents IX

8.2.2 Available data 157

8.2.3 Design philosophy 158
8.2.4 Evaluation 161
8.3 Submerged vanes in the Waal River, the Netherlands 162
8.3.1 Introduction 162
8.3.2 Available data 163
8.3.3 Design philosophy 164
8.3.4 Evaluation 168
8.4 Improvement of the access channel to the port of Barranquilla,
Colombia 169
8.4.1 Introduction 169
8.4.2 Available data 171
8.4.3 Design philosophy 172
8.4.4 Evaluation 176
8.5 Damming of Chao Phraya and Nan Rivers for navigation
improvement, Thailand 178
8.5.1 Introduction 178
8.5.2 Available data 180
8.5.3 Design philosophy 181
8.5.4 Evaluation 182
8.6 Gravity based structure 185
8.6.1 Introduction 185
8.6.2 Boundary conditions 186
8.6.3 Assessment of the scour depth 187
8.6.4 Small-scale model tests 188
8.6.5 Evaluation 193

REFERENCES 195
Foreword

In February 1953 many dikes in the south-western provinces of the Nether¬

lands failed during a severe stormtide. As a consequence of this disaster sev¬
eral hydraulic and soil mechanical issues had to be dealt with in order to be
able to draw up appropriate solution for the breaches in the flood defences.
In part the solutions consisted of repairing the dikes and in part by construct¬
ing closure dams in the existing estuary. To study the effects of enclosures
small-scale experiments were carried out to obtain general information about
the critical velocity for the stability of stones and concrete blocks, the over¬
lapping of mattresses, the water movement, and the scouring effects down¬
stream of revetments.
Based on a systematic investigation of the time scale for two and three
dimensional scour in loose sediments relations were derived for predicting
the maximum scour depth as function of time. In the nineties these scour re¬
lations were slightly modified and used for the design of the storm surge
barrier in the New Waterway near Rotterdam and for the prediction of the
scour process downstream of the barrier in the Eastern Scheldt.
This manual concerns the scour processes and phenomena taking place
near several hydraulic structures. It is intended primarily for hydraulic engi¬
neers in the field; however, it may have some appeal also to researchers in
hydraulic engineering. The scour process has not yet been explained in a
general accepted manner, and therefore it would be only appropriate to dis¬
cuss their mechanism and formulations for the simplest cases. I hope this
manual will offer the practical engineer as much interest as the authors ex¬
perienced in preparing it.

A. Hoekstra
Former head of Hydraulic Engineering Division

XI
List of main symbols

A Cross-section (L^)
b Length of structure (perpendicular to flow direction) (L)
B Width of flow (L)
c Coefficient (-)
C Chezy coefficient (L'''^ T“')
d Particle diameter (L)
^50 Median grain size (L)
D Height of sill (L)
D, Sedimentological diameter, d(Aglv^y^^ (-)
e Embedment (L)
Friction coefficient, g/CP- (-)
Friction coefficient (-)
Roughness function, C/Cq, Cq = 40 Is (-)
Fr Froude number, (_)
g Acceleration of gravity (LT"^)
h Flow depth (L)
/zq Initial flow depth (L)
/z, Tail water depth (L)
H Drop head across structure or wave height (L)
Significant wave height (L)
/ Volume of scour hole per unit width (L^)
k Turbulence energy (L^T"^)
Effective or equivalent bed roughness (rough: k^ = Sz/gg, smooth:
k^ = 2d,,) (L)
K Coefficient (-)
KC Keulegan Carpenter number, Tib (-)
L Length of bed protection (L)
Failure length (L)
p Porosity (-)
q Discharge per unit width (L^T"')
Q Discharge (L^T"')
Reduction transport per unit width (L^T"')

XIII
 XIV Scour manual

s Sediment transport per unit width (L^T-')

Re Reynolds number, U^^h/v (-)
r Discrepancy ratio (-)
Tq Relative turbulence intensity, (sJU (-)
t Time (T)
Characteristic time at which the maximum scour-depth equals (T)
T Wave period (T)
Tp Peak wave period (T)
u Local longitudinal flow velocity (LT“')
u^ Bed shear-velocity (LT'-')
Critical bed shear-velocity (LT“')
I/q Mean velocity, Q/A (LT ')
U Depth-averaged velocity (LT“')
Critical mean velocity (LT“')
Fall velocity (LT *)
X Longitudinal distance (L)
y Vertical distance (L)
Maximum scour depth (L)
a Flow and turbulence coefficient (or angle) (-)
a„ Coefficient, a - UJ Uq (-)
P Upstream scour slope (or angle) (-)
Y Coefficient (or angle) (-)
5 Slope angle downstream of the point of reattachment (-)
A Relative density, pjp - 1 (-)
K Constant of von Karman (-)
X Wave length or angle (L)
jl Shape factor or roughness factor (-)
V Kinematic viscosity (L^T"')
p Fluid density (M^L^*)
Material density (M^L*')
o Relative standard deviation (-)
Standard deviation of the instantaneous longitudinal velocity averaged
over the depth (LT“')
Xq Bed shear-stress (ML“'T“2)
x^ Critical bed shear-stress (ML“‘T^2)
p Angle of repose (-)
())' Angle of internal friction (-)
'P Mobility or Shields parameter (-)
CO Turbulence coefficient (-)

Subscripts
b Bed
c Critical or current
 List of main symbols XV

e Equilibrium
H Horizontal
m Maximum or model
p Prototype
V Vertical
w Waves
0 Initial or reference
CHAPTER 1

Introduction

1.1 GENERAL

A hydraulic structure is generally intended to provide a practical measure to

solve an identified problem. After problem identification subsequent stages
are determined by a series of decisions and actions culminating in the crea¬
tion of a structure or structures to resolve the problem. Aspects that may af¬
fect the eventual outcome of the design process have to be assessed. In addi¬
tion to hydraulic, geotechnical and engineering characteristics, phenomena
such as social conditions, economics, environmental impact and safety re¬
quirements also influence the design process.
Within the scope of the Dutch Delta works, systematic research with re¬
spect to the prediction of the formation of scour holes was conducted by the
Dutch Ministry of Transport, Public Works and Water Management and
Delft Hydraulics. After the catastrophic flood disaster in 1953, the Delta
plan was formulated to protect the Rhine-Meuse-Scheldt delta against future
disasters. Dams with large-scale sluices were planned in some estuaries. The
severe scour that was expected necessitated the acquisition of a better un¬
derstanding of the scour process.
To obtain detailed information about the physical processes playing a role
in scour development, many experiments in which various parameters of the
flow and the scoured material were varied, were carried out by Delft Hy¬
draulics. Prom the results of experiments in flumes, with all the attendant
difficulties of scale effects and limitations in instrumentation, some semi-
empirical relations were obtained that describe the erosion process as a
function of time and position (Breusers 1966, 1967; van der Meulen & Vinje
1975). In addition, design criteria for the length of the bed protection were
deduced. These were based on many hundreds of shear failures and flow
slides that occurred along the coastline in the south-western part of the Neth¬
erlands.
Understanding of the physical processes and mathematical modelling of
the water and sediment movement in rivers, estuaries and coastal waters
have made much progress in recent years. This has led to a number of more

1
2 Scour manual

or less ready-to-use mathematical model systems, but it has also raised many
new research questions. In the early nineties a morphological model for the
generation of scour holes behind hydraulic structures was developed
(Hoffmans 1992). This morphological model was based on the 2-D Navier
Stokes and convection-diffusion equations and used for the calibration and
verification of semi-empirical relations that predict the scour process.
This manual highlights the so-called Breusers-equilibrium method which
describes the maximum scour depth as function of time, including the practi¬
cal equilibrium value near hydraulic structures. Scour due to three-
dimensional flow can easily be predicted when this method is applied in
combination with computational results of depth-averaged models or with
measurements obtained from scale models. The accuracy of the scour com¬
putation depends mainly on the accuracy of the flow velocities and the turbu¬
lence intensities just above the protected bed. According to Breusers (1966),
the development of the scour process depends entirely on the average flow
velocity and relative turbulence intensity at the transition of the fixed to the
erodible bed. By applying this concept, the scour prediction is restricted to
one computation; no information is needed concerning the near bed veloci¬
ties and bed turbulence in the scour hole.

1.2 SCOPE OF THIS MANUAL

The purpose of this scour manual is to provide the civil engineer with practi¬
cal methods to calculate the dimensions of scour holes and to furnish an in¬
troduction to the most relevant literature. The manual contains guidelines
which can be used to solve problems related to scour in engineering practice
and also reflects the main results of all the Dutch research projects directed
to the phenomena of scour which have been realised in the Netherlands dur¬
ing recent decades. A complete review of all the available references on
scour is beyond the scope of this manual. A more general introduction to
scour is given by Breusers & Raudkivi (1991). The scour depth as function
of time can be predicted by the so-called Breusers-equilibrium method.
Basically, this method can be applied for all situations where local scour is
expected. However, the available knowledge about scour is not sufficient, to
apply the method for the prediction of scour at each type of structure. Struc¬
ture specific scour prediction rules are presented then. The treatment of local
scour is classified according to different types of structures. Each type of
structure is necessarily schematised to a simple, basic layout. There is a brief
description of the main parameters of a structure and of the main parts of the
flow pattern near a structure, in so far as they are relevant to the description
of scour phenomena. Detailed and theoretical descriptions of the flow phe¬
nomena are not included because at this stage the consequences of such de-
 Introduction 3

scriptions are minimal in relation to engineering practice. As many scour

problems are still not fully understood attention is paid to the validity ranges
and the limitations of the formulas, and to the information about the accu¬
racy of the calculated parameters, i.e. the maximum scour depth expected
during the lifetime, the upstream scour slope and the expected failure length.
Due to shear failures or flow slides, the scour process can progressively
damage the bed protection; this will lead to the failure of hydraulic struc¬
tures.
Summarising, the presented Breusers-equilibrium method can be applied
directly in the engineering practice for nearly all types of structures. One
may expect that in the near future accurate local flow velocities and turbu¬
lence intensities as a result of three-dimensional flow models will act as in¬
put for the Breusers-equilibrium method, which can be considered as a con-
stination and an expansion of the work of Breusers (1966). In other words,
one may speak of a revitalisation of the Breusers formula, with which a lot
of experience has been gained, mainly in the Netherlands but also abroad.

1.3 SUMMARY

The manual is divided into seven parts, the first of which is a general intro¬
duction to the subject. The next five parts deal with calculation methods for
predicting scour near hydraulic structures and, in the final part, some exam¬
ples of scour at prototype scale are described. A brief summary of each
chapter follows.

Chapter 2 - Basic concepts

The most relevant characteristics that influence the scour process, such as
hydraulic, morphological and geotechnical conditions are discussed. The
functional design of hydraulic structures is also introduced. If safe long term
functioning of hydraulic structures is to be ensured, access to information
relating to failure mechanisms and boundary conditions is indispensable.
The scour which may occur at the site of a structure may be considered as a
combination of bed scour resulting from different processes (general scour
and local scour). In addition, time phases can be distinguished in the scour
process. These phenomena and relations for the critical velocity (based on
the Shields diagram) are presented for both current and waves. The influence
of nonuniform flow is usually expressed by either a turbulence coefficient or
a dominating flow velocity or by both. Scouring is a more or less continuous
process which may suddenly be disturbed by the occurrence of geotechnical
instabilities along the upstream scour slope. Shear failures and flow slides
influence the stability of hydraulic structures. In the extreme case these in¬
stabilities involve large masses of sediment and cause a major change of the
4 Scour manual

shape of the upstream side of the scour hole in a relatively short period of
time. Some design criteria which are based on storage models are presented.
In the subsequent chapters the basic scour concept for a number of hydraulic
structures and applications is discussed.

Chapter 3 - Sills
Calculation methods with respect to sills are summarised. A distinction is
drawn between sills with a broad or a sharp crest and between sills with and
without bed protection. Usually, the flow above a sill is subcritical but, de¬
pending on the downstream water level, the flow may become supercritical.
The time-dependent and equilibrium behaviour of scour holes in sandy beds
related to closure works (broad-crested sills) in tidal channels is discussed.
Special attention is paid to the effects of turbulence and flow pattern on the
scour process. An approximate method (reduction method) for calculation of
the maximum scour depth is described. This takes into account the influence
of upstream sediment supply. In addition, a method to adjust this calculation
method for unsteady flow, especially tidal flow, is given. These methods
were successfully applied during the design of the Eastern Scheldt Storm
Surge barrier. The upstream scour slope determines the stability of the up¬
stream part of the scour hole and the adjacent bed protection. A relation for
the upstream scour slope, based on a probabilistic bed load model for bed
load transport, is presented. Relations derived from the so-called systematic
scour investigation were verified by two field experiments.

Chapter 4 - Jets
Scour due to several jet forms, such as plunging jets, submerged jets, hori¬
zontal and vertical jets, and two and three dimensional jets is discussed. In
addition, the complex flow pattern of jets is treated. Semi-empirical relations
for the scour process behind a short-crested sill are presented. These rela¬
tions are often used in grade-control structures, where the flow above the sill
is supercritical and for the time dependent development of the maximum
scour depth downstream of a hydraulic jump. The structure of these relations
show a good similarity with the Breusers approach. Since there is no univer¬
sal scour relation that gives a good prediction of the practical equilibrium
scour depth in all cases, some semi-empirical relations are presented. These
relations must be clearly understood prior to any attempt to use them for de¬
sign purposes.

Chapter 5 - Abutments and spur dikes

Relations for predicting local scour at the head of abutments, for which sev¬
eral names are used in the literature are presented. The flow characteristics
around blunt and streamlined abutments are briefly discussed. Attention is
also paid to the time scale of the scour process and types of scour (e.g.
 Introduction 5

overall degradation, bend scour, constriction scour). Since the literature

contains many scour relations, a number of generally acceptable predictors
have been selected for this report. Finally, attention is paid to failure
mechanisms and measures to mitigate scour near abutments.

Chapter 6 - Bridge piers

Relations for estimating scour around bridge piers are summarised. To date
it has not been possible to determine the scour depth around bridge piers
from a theoretical analysis of the water movement around the pier, so em¬
pirical relations with correction factors and design graphs for the equilibrium
scour depth are discussed. Attention is paid to both the equilibrium scour
depth and to the time scale of the scour process and some methods used to
predict bridge piers from scour are cited.

Chapter 7 - Coastal and ojfshore structures

This chapter deals with the scour induced by wave and current action near
pipelines, piles, vertical breakwaters, coastal defences and other submerged
structures. For accurate prediction of the scour process it is necessary to
have detailed descriptions of the nearshore hydrodynamics. However, these
processes fall outside the scope of this manual, so the required information
should be obtained from site-specific studies. Here, simple prediction meth¬
ods which relate the scour depth to the incident wave conditions, the flow
depth, the structure geometry and/or reflection coefficient are described. No
effects of angled wave attack or of tidal or wave induced current have yet
been accounted for. It must be stressed that our knowledge of scour due to
waves or due to the combined effect of current and waves is still in a rather
rudimentary stage and that more research and practical experience is still
needed.

Chapter 8 - Prototype situations

Five cases on prototype scale, based on feasibility studies or design studies,
are evaluated in order to determine the practical use of the scour relations al¬
ready discussed. These cases are:
- Lengthening of abutments in the Waal River, the Netherlands,
- Submerged vanes in the Waal River, the Netherlands,
- Improvement of the access channel to the port of Barranquilla, Colom¬
bia,
- Damming of the Chao Phraya River, Thailand, for navigation improve¬
ment,
- Gravity based structure. North Sea.
Acknowledgement
The discussions with colleagues at Delft Hydraulics were much appreciated.
6 Scour manual

especially the discussions with H.N.C. Breusers, GJ. Klaassen, T. van der
Meulen, L.C. van Rijn, N. Stmiksma and A.M. Talmon.
The support and useful comments of A. Franken, R.E. Jorissen, A. Kraak,
H.E. Klatter, J. Lindenberg and F.M. Stroeve of the Ministry of Transport
Public Works and Water Management was highly appreciated.
Einally, R. van der Laag was gratefully acknowledged for the drawing.
CHAPTER 2

Basic concepts

2.1 INTRODUCTION

Hydraulic structures cause disturbances in uniform flow and sediment trans¬

port. Downstream of these structures flow velocities increase due to con¬
striction of the channel. When the flow velocities decrease (i.e. in the decel¬
eration zone), a higher degree of turbulence is generated and therefore a
stronger erosion capacity is present. In most cases this leads to scouring and,
depending on the specific hydraulic conditions, there are sometimes steep
upstream slopes. Bed protection is often constructed in order to decrease the
maximum scour hole and to shift the scour holes that involve a potential risk
to structural stability to a greater distance from the hydraulic structure. The
main dimensions of the scour hole can be characterised roughly from the
maximum scour depth expected during the lifetime of the structure and by
the upstream scour slope. Both hydraulic and geotechnical characteristics
influence these two design parameters and are treated in more detail in sub¬
sequent sub-sections.
To ensure the safety and long term functioning of hydraulic structures, it
is necessary to consider information with respect to failure mechanisms,
boundary conditions and design criteria (Section 2.2). General aspects of the
interaction between the flow and sediment movement are presented in Sec¬
tion 2.3. Attention is paid to the processes relevant to sedimentation and
consolidation, because these affect the erosion behaviour of beds of cohesive
sediments. The scour which may occur near a structure can be considered to
be a result of bed scour caused by different processes. Usually various time
phases occur during the process of bed scour. Initially the development of
scour is very fast but eventually a state of equilibrium is reached (Section
2.4). Owing to shear failures or flow slides, the scour process may progres¬
sively damage the bed protection; this will lead to the failure of hydraulic
structures. Section 2.5 provides empirical criteria for determination of the
damage caused by shear failures or flow slides.

7
8 Scour manual

2.2 DESIGN PROCESS

2.2.1 General
In recent years the need for reliable information on modelling of sediment
exposed to wave and current action has been increasing. This need arises
partly from an increase in the construction of structures which have to be
protected to meet to higher safety standards.
When designing structures, the following aspects must be considered
(Pilarczyk 1995);
- Eunction of the structure; erosion as such is not the problem as long as
the structure can fulfil its function,
- Physical environment; the structure should offer the required degree of
protection against hydraulic loading, with an acceptable risk and, when pos¬
sible, meet the requirements resulting from landscape, recreational and eco¬
logical viewpoints,
- Construction method; the construction costs should be minimised to an
acceptable level and legal restrictions must be adhered to,
- Operation and maintenance; it must be possible to manage and maintain
the hydraulic structure.
Elaboration of these points depends on specific, local circumstances, in¬
cluding the type of terrain (lowland or highland) and its development
(economic value), availability of equipment, and availability of labour and
materials. The cost of construction and maintenance is generally a control¬
ling factor in determining the type of structure to be used. The starting points
for the design should therefore be carefully examined in co-operation with
the future manager of the project. Most research problems concerning water
defences are multidisciplinary in character, specially in the technical sense,
as characterised by all relevant interactions between the soil, water and
structure, and may lead to combined hydraulic, geotechnical and structural
research. These interactions may be brought together in the diagram shown
in Figure 2.1.
An example of the design process for a bed protection in relation to ex¬
pected (predicted) local scour is illustrated in Figure 2.2. In subsequent sec¬
tions some design aspects are discussed in more detail.

2.2.2 Boundary conditions

In order to design hydraulic structures, loading (hydraulic conditions) and
strength (morphological and geotechnical conditions) parameters have to be
specified. Flow characteristics (flow velocities, water levels, discharges) and
turbulence intensities determined by the geometry of the hydraulic structure
and bed roughness characterise the flow pattern in the vicinity of the hy-
 Basic concepts 9

bed pro-
’ tection (geo¬
metry/slopes,
\toplayer, sub¬
layers)

response:
damage armour,
migration filter,
deformation
toplayer/sublayers
hydraulic
geotechnical
conditions
conditions/
(waves & interactions
internal loads
currents) <:^=> Figure 2.1. Soil-water-structure interaction
(SOWAS concept).

boundary geotechnical
conditions hydraulic (geometry)

function bed sand impenetration

protection/design stability bed protection
requirements local scour
flow pattern/
scour process
length bed
protection methods
scour protection/
practical
scour applications
calculation

model
reliability loosely packed sand
investigation
,,yes

failure flow slides

translation (liquefaction)
analysis/
prototype
criteria shear failures

prototype
experiment

additional yes improvements final

measures necessary design

sheet pile execution

ground replacement
compacting
real (proto)
penetrable protection
scour (P,ym)
rock/slag protection
-1-
_ future_ _
measures
Figure 2.2. Design phases of the bed protection.
10 Scour manual

draulic structure and thus provide a measure of the erosion in the scour hole
just downstream of the hydraulic structure. On the other hand, the scour
process is also determined by the composition of the bed material (sub-soil).

Hydraulic conditions
The most simple flow pattern is generated by a steady, uniform flow. How¬
ever, special conditions for the flow pattern can be distinguished, for exam¬
ple, an accelerated flow in a local constriction, a river bend with well devel¬
oped bend flow, an unsteady flow due to flood waves or tidal movement or
when the direction of the flow downstream of hydraulic structures is per¬
pendicular or inclined to the flow direction of a receiving river or estuary.
An ice cover can divert the flow to the part of the bed near the hydraulic
structure, resulting in an additional increase of local scour. The main hy¬
draulic and special conditions are summarised in Table 2.1.

Morphological conditions
For convenience, the sediments forming the boundaries of a flow are subdi¬
vided into cohesive and non-cohesive sediments, although there is a fairly
broad transition range. In non-cohesive sediments such as sands and gravels,
the particle or grain size and material density are the dominant material pa¬
rameters for sediment transport. Bed material which is widely graded will be
more resistant to scour than uniform material of the same median grain size.
During a flood, the finer grains of a nonuniform bed material are eroded in
preference to the coarser grains, and the median grain size of the bed mate¬
rial will increase. This process is known as armouring.
The shape of grains, the surface packing of grains and multiple layers of
different bed materials are additional properties of the scour process which
are of secondary importance and therefore not treated in detail in this man¬
ual. The physio-chemical properties of cohesive sediments play a significant
role in the resistance of cohesive sediments against current and waves. These
properties depend strongly on granulometric, mineralogical and chemical
characteristics of the sediment involved (Table 2.1 and Section 2.3.3).

Geotechnical conditions
A purely hydraulic and morphological approach to a given geometry (struc¬
ture, bed protection) and hydraulic boundary conditions leads to scouring in
which the maximum scour depth gradually increases and the upstream scour
slope steepens (at least the steep part will become longer) until it reaches the
equilibrium phase. This more or less continuous process may suddenly be
disturbed by the occurrence of geotechnical instabilities along the upstream
scour slope. In the extreme case these instabilities involve large masses of
sediment and cause a major change of the shape of the upstream side of the
scour hole in a relatively short period of time. The steeper this slope, the
 Basic concepts 11

Table 2.1. Boundary conditions.

Loading Hydraulic con- Flow velocities (water levels and discharges)
ditions Wave characteristics (flow depth, wave height and wave
period)
Turbulence energy or turbulence intensity (geometry of
structure)
Kinematic viscosity (influence of water temperature)
Fluid density
Special condi- Supercritical/subcritical flow
tions Unsteady/steady flow
Ice cover
Earthquakes

Strength Morphological Grain size distribution (representative diameter)

conditions Material density
Non-cohesive/cohesive sediments
Additional properties of non-cohesive sediments:
- Shape of grains
- Surface packing of grains (homogeneity of bed material)
- Multiple layers of different bed materials
Additional properties of cohesive sediments:
- Sodium Adsorption Ratio (SAR)
- Cation Exchange Capacity (CEC)
- pH-level of pore water
- Temperature
- Organic content
- Porosity
Geotechnical Stratification of sub-soil
conditions Bulk density of sand (loosely or densely packed)
Angle of internal friction
Porosity
Cohesion

greater the probability of slope failure. Although of minor importanee, the

maximum scour depth also plays a role.
Besides these geotechnical aspects, the soil properties are extremely im¬
portant, especially with regard to the type of geotechnical instability that
may occur. Two types of instability are distinguished for cohesionless sedi¬
ment, namely shear failure and flow slide. To predict the occurrence of a
shear failure, the steepness of the upstream slope has to be assessed in rela¬
tion to the angle of internal friction of the bed material. A flow slide is a
more complex geotechnical phenomenon which can only occur in loose to
very loose sand. However, the final geometrical characteristics of the up-
12 Scour manual

stream slope are generally of much greater importance in relation to flow

slide instability than they are for a shear failure. The main geotechnical pa¬
rameters are presented in Table 2.1. In Section 2.5 more information about
geotechnical instabilities is given.

2.2.3 Fault tree analysis

The design process is characterised by solving design problems in an itera¬
tive manner. Since the processes involved are dynamic it is impossible to
reach the optimum solution at the first attempt. Though the optimum solu¬
tion will never be attained, the design philosophy which has been adopted
helps to prevent a haphazard approach to design and research. A general
overview of the failure mechanism of an open bed protection is given in
Figure 2.3 (Vrijling 1990). To produce a safe and reliable design, the total
reliability as a function of all modes of failure should be approximated, at
least at a conceptual level.
The fault tree is a useful tool for integrating the various failure mecha¬
nisms into a single approach. For example. Figure 2.4 shows the fault tree
for a bed protection in which the foundation of the hydraulic structure is the
central point. The bed protection has to prevent or slow down a change in
the geometry of the foundation. A failure of the bed protection does not di¬
rectly imply the loss of the structure, however, when the subsoil becomes
unstable owing to the existence of a well-developed scour hole, the resis¬
tance of the foundation is reduced.
A further advantage of fault tree analysis that this makes it possible to in¬
corporate the failure of mechanical or electrical components as well as hu¬
man errors in the management and maintenance of the structure. For in¬
stance, the safety of a sluice can be dramatically improved by regular echo¬
sounding of the bed protection and by subsequent maintenance if the initia¬
tion of a scour hole is discovered. The probability of instabilities affecting

shear failure erosion of top layer

vzrrrryricxrrnnrn _
flow slide loss of material
through bed protection

Figure 2.3. Failure modes of an open bed protection.

 Basic concepts 13

loss of the
sluice
structure

foundation S = surcharge (loading)

failure
S>R R = resistance

head load foundation

resistance

soil pore geometry of

properties pressures foundation

and-gate

failing
inspection Q or gate
maintenance

shear failure In flow slide in shear failure at flow slide at

the middle of the middle of the end of the end of
bed protection bed protection bed protection bed protection

scour hole in scour hole scour hole

bed protection too steep too deep
5
erosion of loss of material bed protection
toplayer through filter too short
S>R S>R S>R
Figure 2.4. Fault tree for bed protection

the foundation is thus reduced to the coincidence of scour hole formation

and failings in inspection and maintenance.
More information about the design process, including outlines of main
considerations relating to deterministic and probabilistic design processes
can be found, for example, in CUR/DWW (1995).

2.2.4 Protective measures

Several methods may be used to protect hydraulic structures from damage
due to scouring. A conservative measure is to place the foundations of struc-
14 Scour manual

Table 2.2. Protective measures.

Load reduction Strength enlargement
Lengthening bed protection Compacting sub-soil
Roughening bed protection Grouting sub-soil
Streamlining hydraulic structure (guide Protecting bed/upstream scour slope (mat¬
vanes, collars, deflectors) tress protection, sand and stone bags,
Energy dissipators hinged concrete slabs, artificial seaweed,
Wave dissipators (berms, shallow fore¬ flexible mats)
land, vegetation, reed)

tures at such a depth that the deepest scour hole will not threaten the stability
of the structure. Another way is to prevent the generation of erosive vortices.
Hydraulic structures placed in waterways are often streamlined in order to
reduce the drag exerted by flow and to reduce the effects of wake and turbu¬
lence intensity. Streamlining by means of deflectors and guide vanes, how¬
ever, is effective only when the hydraulic structure is aligned with the flow
within narrow limits.
Placing a bed protection downstream or around hydraulic structures is a
common method of local scour protection. In principle, two types can be
distinguished: the permeable, which is sand-tight and the impermeable.
Scour occurs in the area of the bed beyond the flexible bed protection and, as
the scour hole is formed, the bed protection slides down into it. When rock
mattresses or loose riprap are used, consideration has to be given to the pos¬
sibility of erosion of fines from underneath the bed protection. This will be
discussed in Section 2.2.5. Local scour can be reduced or prevented by either
reducing the loading parameters or by increasing the strength parameters
(Table 2.2).

2.2.5 Characteristics of bed protection

From a geotechnical point of view, the stability of the upstream scour slope
is of prime importance both during the scour process and in the final situa¬
tion when the equilibrium geometry has been attained. Besides sediment
transport in the scour hole, soil particles in the fdter structure below the bed
protection can also be transported in both vertical and horizontal directions.
If the groundwater seepage flow becomes concentrated in narrow passages
or pipes, the hydraulic structure may fail due to the transport of soil particles
within the filter structure, so special attention must be paid to the sand perco¬
lation.
Furthermore, the stability of both the upper layer and the end of the bed
protection against current, waves and eddies has to be safeguarded, so that
the bed protection and thus the hydraulic structure will not be undermined.
 Basic concepts 15

2.3 INITIATION OF MOTION

2.3.1 General
The first treatise on initial bed grain instability using the concepts of Prandtl
and von Karman on boundary layer flow mentioned in the bibliography was
produced by Shields (1936), who described the problem using the following
parameters: the fluid density, the sediment density, the kinematic viscosity,
the grain size and the bed shear-stress. When the flow velocity over a bed of
non-cohesive material has increased sufficiently, individual grains begin to
move in an intermittent and random fashion. Bed instability results from the
interaction between two stochastic variables. At first, every grain on the bed
surface can be assumed to be potentially susceptible to an instantaneous
critical bed shear-stress. The grain becomes unstable if the instantaneous bed
shear stress exceeds the critical one. Due to the random shape, weight and
placement of the individual grains, these critical shear stresses will have a
probability distribution. The other random variable is the instantaneous bed
shear stress generated by the flow. The probability that the instantaneous bed
shear stress is greater than a characteristic critical shear stress is a measure
of the transport of sediment. The mechanics of interaction between fluid
flow and sediment is the subject of numerous papers and a number of text¬
books, for example, Bogardi (1974), Graf (1971), Raudkivi (1993), van Rijn
(1993), Yalin (1972, 1992).
The physio-chemical properties of cohesive sediments play a significant
role in the resistance of cohesive sediments to currents and waves. These
properties depend strongly on the granulometric, mineralogical and chemical
characteristics of the sediment involved. Up to now, direct quantitative rela¬
tions between the physio-chemical properties and the erosion rate have not
been established. Nevertheless, design engineers require information to pre¬
dict scour in cohesive sediments, because these soils are widespread natural
sedimentary deposits. The governing parameters for the erosion of non-
cohesive and cohesive sediments are discussed in Sections 2.3.2 and 2.3.3
respectively. In Section 2.3.4 the phenomenon turbulence is treated in an in¬
troductory way. Finally some examples are given in Section 2.3.5.

2.3.2 Non-cohesive sediments

In 1936 Shields published his criterion for the initiation of movement of uni¬
form granular material on a flat bed. The experimental data used by Shields
was mainly obtained by extrapolating curves of sediment transport versus
applied shear stress to the zero transport condition. Originally the data points
were plotted by Shields and the curve (averaged critical value), constituting
the ‘Shields diagram’, as usually quoted (Neill 1968), was drawn by Rouse.
16 Scour manual

Actually, because of the nonuniform distribution of the mixtures and the ef¬
fects of grain and imbrication (i.e. the preferred orientation of natural sands
and gravel particles under certain conditions of transport), Shields drew not a
single curve but a broad belt. In the literature numerous relations can be
found for the critical velocity for particle movement (e.g. Brahms 1753;
Izbash 1970). The critical bed shear-stress (or critical mobility parameter)
can be obtained graphically, directly from the modified Shields diagram
(Fig. 2.5) or by using expressions that fit the Shields diagram.

Current
The critical bed shear-stress (x^) is defined as:
Xc^pul^ (2.1)

According to Shields (1936), the critical mobility parameter is:

u^
(2.2)
^ Agd
in which: d = particle diameter (m) {d = d^^ is median grain size)
g = acceleration of gravity, g = 9.81 m/s^
= critical bed shear-velocity (m/s)
= relative density (-)
= fluid density (kg/m^)
= critical mobility parameter (-)

0.12
measured
occasional particle movement at some locations
frequent particle movement at some locations
Shields curve (Wc)
frequent particle movement at many locations
0.10 frequent particle movement at nearly all locations —
frequent particle movement at all locations
0.08 permanent particle movement at all locations
general transport (initiation of ripples) r, 7
0.06
0.05
W
0.04

0.03

0.02

0.01
8 10 8 10
D.
Figure 2.5. Shields diagram.
 Basic concepts 17

For a uniform flow (logarithmic velocity profile) over a hydraulically rough

bed the critical mean velocity is:
(2-3)

where the Chezy coefficient is given by:

rs 'hr'
C= In (2.4)

in which: k^= equivalent roughness of Nikuradse (m), hydraulically rough

flow: = Sr/go, hydraulically smooth flow: k^ =
R = hydraulic radius (m)
K = 0.4, constant of von Karman (-)

If the width B of the flow is large compared to the flow depth h, can be
rewritten as:
= 2.5 sj^c^gd In (l2h / k^) (2.5)

Vanoni (1967) has noted that the critical mobility parameter for the fully-
rough turbulent zone, i.e. = 0.06, corresponds to a low but measurable
bed load. At values of 0.03 and even less, occasional movement of single
grains may occur. Empirical relations for as proposed by van Rijn
(1984), are presented in Table 2.3. As given by van Rijn (1984), the critical
Shields parameter is related to the sedimentological diameter D*:
40 X 10^
A=J with (2.6)
20+0
in which: v = kinematic viscosity (in m^/s)
0 = temperature (in °C)
The equivalent roughness of a plane bed is usually related to the largest par¬
ticles of the bed (J^j, or J^q). The influences of the gradation, the shape
of the particles and the flow conditions are generally disregarded. In the

Table 2.3. Empirical relations for (van Rijn 1984).

Tc as function of D* (= £?(Ag/v2)'^3)
4DT‘ for D*<4
for 4<D»< 10
^Dp0.io for 10<Z)*<20
for 20<D*< 150
for D*> 150
 18 Scour manual

systematic research (Delft Hydraulics 1972) on scouring in which relatively

uniform sediments were used, the equivalent roughness was assumed to be
equal to the median particle diameter. However, in the literature several ex¬
pressions for can be found (e.g. van Rijn 1982).
According to van Rijn (1982), the equivalent roughness of a plane bed
varies from about 1 to 10 times of the bed material. These values, which
are rather large, indicate that a completely plane bed does not exist. For
conditions with active sediment transport the relatively large scatter of the
equivalent roughness is probably caused by the initial unevenness (initial
bed forms).

Waves
In oscillatory flow there is no generally accepted relation for the initiation of
motion on a plane bed. Many equations have been proposed. Silvester &
Mogridge (1970) present 13 different equations collected from the literature.
Sleath (1978) and van Rijn (1989, 1993) established a modified Shields cri¬
terion using various sets of published data (Fig. 2.6). Sleath introduced cor¬
rection factors to include the different criteria for the initial motion condition
used by the different investigators. Van Rijn used the time-averaged bed
shear-stress and not the maximum bed shear-stress, because the Shields
curve which was originally proposed for unidirectional flow, is based on
time-averaged parameters.

10°
8 Bosman (flume)
6 0 Dingier (flume)
Q Bagnold (flume)
4
0 Silvester and Mogridge (flume)
3 Q Ranee and Warren (flume)
^ Manohar (flume)
2
Y Davies (field)

lo-i
W 8

10-2
8 10° 2 34 68 101 2 34 68 102 2 34 68 103

D. -►
Figure 2.6. Initiation of motion for waves over a plane bed based on critical bed shear-
stress.
 Basic concepts 19

Figure 2.6 represents a critical stage at which only a minor part of the bed
surface is moving. According to van Rijn (1993), it seems justifiable to con¬
clude that the Shields curve can also be applied as a criterion for the initia¬
tion of motion for oscillatory flow over a plane bed. Usually, small scale
ripples are present and in sucb cases the critical velocities for initiation of
motion are considerably smaller owing to the generation of vortex motions
near the bed.
Following van Rijn (1989, 1993), the time-averaged value of the absolute
bed shear-stress acting on the bed is the sum of the current-related and the
wave-related bed shear-stress, and can be given by:
V™ = P/.f'"+'/4P/„t/5 (2.7)
in which: = g/C^, friction coefficient (-), C is Chezy coefficient,
g = 9.81 m^/s, acceleration of gravity
f
Jw = friction coefficient (-)
u = depth-averaged (current-induced) velocity (m/s)
= wave induced velocity (m/s)
X0,cw = bed shear-stress exerted by combined wave and current
In the hydraulically rough regime (U^ ^Jv> 10^ and AJk^ < 100) Swart
(1976) proposed:
/ ^ +0.19 3

/w = exp -6 + 5.2 (2.8)

in which: - U^TI^ln), peak value of orbital excursion near the bed (m)
T = wave period (s)
fw,m ^ Q3 for AJk^< 1.57

2.3.3 Cohesive sediments

The initiation of motion and the transport of non-cohesive sediments are
both determined by the submerged weight of the particles. For cohesive
sediments, relatively large forces are necessary to break the aggregates
within the bed and relatively small forces are necessary to transport the ma¬
terial. Experiments by Mirtskhoulava (1988, 1991) have shown that the
scour of clay soils with a natural structure in a water saturated state occurs in
several stages. In the initial stage loosened particles and aggregates separate
and those with weakened bonds, are washed away. This process leads to the
development of a rougher surface. Higher pulsating drag and lift forces in¬
crease the vibration and dynamic action on the protruding aggregates. As a
result the bonds between aggregates are gradually destroyed until the aggre-
20 Scour manual

gate is instantaneously tom out of the surface and carried away by the flow.
The above mentioned scour process is influenced by the following parame¬
ters: cohesion, Cation Exchange Capacity (CEC), salinity, Sodium Adsorp¬
tion Ratio (SAR), pH-level of pore water, temperature, sand, organic con¬
tent, porosity (e.g. Winterwerp 1989).
The erosion rate of a cohesive bed is determined by the mutual effects of
the sediment and pore water properties. A parameter describing the proper¬
ties is the SAR, which is indicative of the processes in the diffusive double
layer. In general, the critical bed shear-stress will increase with decreasing
SAR and the critical bed shear-stress will increase when salt is added to the
pore water. An increase in pH-level of the pore water and in temperature will
decrease the strength of the bed. In addition to the commonly determined pa¬
rameters such as the granulometry and the mineralogy, the specific surface
of the sediment and the CEC are also important. With a larger specific sur¬
face, the van der Waals forces become larger and the sediment becomes
more cohesive. An increase in CEC with low SAR will also result in an in¬
crease in cohesion: the critical bed shear-stress will increase and the erosion
rate will decrease with increasing CEC.
In general, an increase in organic content will cause an increase in the co¬
hesiveness of the sediments, resulting in a larger critical bed shear-stress and
a smaller erosion rate. However, this effect is known only qualitatively, and
no quantitative information is found in the literature. Another natural aspect
is biological activity, such as bioturbation. The effect of sand on the strength
of a cohesive bed seems to be dependent upon the value of the SAR: at low
SAR the strength of the bed will decrease with increasing sand content, at
high SAR the reverse trend is expected. This effect also is known only
qualitatively, and no quantitative information is found in the literature.
In general, no applicable design equations for the depth of scour holes are
available for cohesive sediments. In the literature most equations are related
to one or two particular parameters influencing the erosion of cohesive
sediments, moreover, they are often related to a specific sediment. In many
scour predictions a critical velocity is applied, for example, the Breusers
method. Eor a first estimate, the following values may be used. For fairly
compacted clay with a voids ratio (i.e. the ratio between the volume of voids
and the volume of the mineral part of soil) of 0.50, the critical depth-
averaged velocity is about ~ 0.80 m/s, while for stiff clay (voids ratio
0.25) is about 1.50 m/s (Table 2.4).
Based on the work of Mirtskhoulava (1988, 1991), a simplified expression
for the critical depth-averaged velocity for cohesive sediments is:

(2.9)
 Basic concepts 21

Table 2.4. Critical depth-averaged velocities for cohesive sediments (rough estimates).
Type of soil h (m) fC (m/s)
Loamy sand, light loamy clay with low compactness 1 0.4
Heavy loamy clay with low density 3 0.5
Low density clay 10 0.6

Light loamy clay with medium compactness 1 0.8

Heavy loamy clay with medium density 3 1.0
Clay of medium density 10 1.3
Light loamy clay (dense) 1 1.2
Heavy loamy clay (dense) 3 1.5
Hard clay 10 1.9

Table 2.5. in kPa and (j)' in degrees (given in brackets); rough estimates (Mirtsk-
houlava 1988).
Type of soil and range of Soil property at voids ratio
liquidity index 0.45 0.45 0.65 0.75 0.85 0.95 1.05
Loamy sand
0-0.25 14.7 10.8 7.85
(30) (29) (27)
0.25-0.75 12.7 8.83 5.88 2.94
(28) (26) (24) (21)
Loamy clay
0-0.25 (low plasticity) 46.1 36.3 30.4 24.5 21.6 18.6
(26) (25) (24) (23) (22) (20)
0.25-0.5 (medium plasticity) 38.2 33.3 27.5 22.6 17.7 14.7
(24) (23) (22) (21) (19) (17)
0.5-0.75 (high plasticity) 24.5 19.6 15.7 13.7 11.8
(19) (18) (16) (14) (12)

Clay
0-0.25 79.4 66.8 53.0 46.1 40.2 35.3
(21) (20) (19) (18) (16) (14)
0.25-0.5 55.9 49.0 42.2 36.3 31.4
(18) (17) (16) (14) (11)
0.5-0.75 44.1 40.2 35.3 32.4 28.4
(15) (14) (12) (10) (7)

in which: Cj = 0.035Q Pa (1 Pa = 1 N/m^), fatigue rupture strength of clay,

Cg is cohesion (Table 2.5)
= size of detaching aggregates, = 0.004 m
h = flow depth (m)
= critical mean velocity for cohesive sediments (m/s)
22 Scour manual

p = fluid density (kg/m^)

= material density (kg/m^)
According to Mirtskhoulava (1988, 1991), the cohesion at saturation water
content and the size of the particle diameter appear to be the most significant
features among the extensive complex of physio-mechanical and chemical
properties of cohesive sediments. However, it does not suffice to rely on
averaged properties, the inhomogeneity of the river bed must also be
considered. The erosion characteristics of cohesive sediments are not yet
fully understood. For specific sediments at a given location, quantitative
information relating to the erosion parameters is available but, for most
situations the designer has to perform an erosion test.

2.3.4 Turbulence
The bed turbulence (standard deviation of the instantaneous bed shear-stress)
together with the mean bed shear-stress, determine the bed load. For uniform
flow the ratio between the bed turbulence and the bed shear-stress is ap¬
proximately constant (Compte-Bellot 1963),whereas for nonuniform two-
dimensional flow the bed turbulence is strongly influenced by turbulence en¬
ergy generated in the mixing layer, while for a three-dimensional flow the
bed turbulence and the bed shear-stress is influenced by a combination of
vortices with both a horizontal and vertical axis. Hoffmans (1992) and
Hoffmans & Booij (1993a, b) have shown that in a two-dimensional scour
hole the bed turbulence can be represented by a combination of the turbu¬
lence energy generated at the bed and the turbulence energy from the mixing
layer.
The influence of turbulence on bed load has been previously investigated
by several researchers. As given by Kalinske (1947) and Einstein (1950), the
instantaneous velocity varies according to a Gaussian distribution. Kalin-
ske’s idea was extended by van Rijn (1985), who postulated an instantane¬
ous transport parameter in which the instantaneous bed shear-stress is nor¬
mally distributed. The weakness of these classical stochastic models is that
they do not incorporate modem understanding of turbulence. As a result,
they are not based on the mechanics of turbulence but use only the continu¬
ity equation for sediment transport and the parametric probability density
functions. For example, the measurements of Lu & Willmarth (1973) show
that for uniform flow the influence of sweeps and ejections is not included in
the Gaussian distribution. Sweeps, which are directed towards the bed, and
ejections, which are moving away from the bed, contribute most to the turbu¬
lent shear stresses (Fig. 2.7). However, there is more to the difference than
the enhanced skewness in the instantaneous bed shear-stress. Under non-
uniform flow conditions, there is no clear relation between the instantaneous
 Basic concepts 23

Figure 2.7. Schematic descriptions of

sand roughness
turbulent flow over smooth and rough
retarded fluids beds (Nezu & Nakagawa 1993).

sediment transport and the instantaneous bed shear-stress. Near-bed meas¬

urements of turbulent correlations (Reynolds stresses) are estimates of mo¬
mentum flux, but are only related to the force acting on the bed when it is
averaged over a long period of time.
The relation between instantaneous products of velocity components and
the instantaneous force on a sediment particle is not fully understood but
even so, some general premises have been used to model the two important
design parameters of the scour process; maximum scour depth and upstream
scour slope. According to Breusers (1966, 1967), the sediment transport in a
scour hole is related to the difference between a maximum and a critical ve¬
locity raised to a power. The maximum velocity is a function of the local (or
mean) velocity and the relative turbulence intensity at the end of the bed
protection. The relative turbulence intensity (- oj U) is defined as the ra¬
tio between the standard deviation of the instantaneous velocity averaged
over the depth, and the time and depth-averaged velocity U. For hydrauli¬
cally rough and smooth flow conditions, is about 0.1 and 0.05 respec¬
tively. Directly downstream of a hydraulic jump or a recirculation zone near
the bed, Tq lies in the range of 0.3 to 0.6. At a distance of 20 to 50 times the
flow depth from the eddies is about equal to the uniform turbulence level.
Downstream of a sill the relative turbulence intensity depends strongly on
the height of the sill (relative to the flow depth), the distance from the sill
and the roughness of the bed.
Nezu & Nakagawa (1993) have surveyed a broad spectrum of technical
and scientific literature on the subject of turbulence in open-channel flow.
The text of their state-of-the-art monograph is up-to-date and contains an in¬
valuable source of information. They propose an interaction mechanism
between the coherent structure of turbulence and the initiation of bed con¬
figuration such as ripples, dunes and antidunes. Further development of
physical models and formulae incorporating these coherent structures are
necessary to further progress in this field.
 24 Scour manual

2.3.5 Examples
Example L A wide, open channel has a flow depth of /z = 5 m. The bed is
covered with sand dunes and the bed material characteristics are = 300
ftm, dgQ = 500 fxm, = 2650 kg/m^. The water temperature is 20°C (v = 10“®
m^/s and p = 1000 kg/m^). What is the critical mean velocity?
Solution:
a) Sedimentological diameter
D* = Jjo X (A X g/v2)‘/3 = 300 X 10”6 x [1.65 x 9.81/(10-6)2]'/3 ^ 7 5
b) Critical mobility
= 0.14 X = 0.14 X 7.6-0-64 ^ o.038 (4 < £)* < 10)
c) Critical mean velocity
f4 = X Ax g X 45o)/k X In [12 X h/(3 x J90)]
= V(0.038 X 1.65 X 9.81 x 300 x 10-6)70.4 x In [12 x 5/(3 x 500 x lO-®)]
= 0.36 m/s

Example 2. The flow depth in a coastal sea with a plane bed is /z = 8 m. The
wave period is 10 The depth-averaged current velocity is t/ = 0.3 m/s.
The bed material characteristics are d^^ = 250 [im, dg^ = 400 [im, p^ = 2650
kg/m3. Other data are v = 10-6 m^/s and p = 1025 kg/m4
a) What is the current-related bed shear-stress?
b) What is the critical bed shear-stress?
c) What is the critical wave-related bed shear-stress?
d) What is the critical wave-related bed shear-stress height (wave
length is = 84 m)?
Solution:
a) Current-related bed shear-stress Xg ^
C = -[glK X In [12 X /z/(3 X dgd)] = V^/0.4 x In [12 x 8/(3 x 400 x 10-6)]
= 88.4
X [/2 = 1025 X 9.81/88.42x 0.32 = 0.116 Pa
b) Critical bed shear-stress X^
D* = dso X (A X g/v2)'/3 = 250 x 10-6 x [1.59 x 9.81/(10-6)2]’/3 ^ 5 2
T,, = 0.14 X £),-0-64 = 0.14 X 6.2-0-64 = o.043 (4 < Z)* < 10)
x^ = p X (m^ J2 p X T^x A X g X 450 = 1025 x 0.043 x 1.59 x 9.81 x 250 x 10-6
= 0.173 Pa
c) Critical wave-related bed shear-stress
The critical wave-related bed shear-stress = '^c-'^0c = 0.057 Pa
d) Critical wave height
The critical wave-related bed shear-stress is given by:
= */4 X p X/^^^ X
= exp (-6 4- 5.2 X = exp (- 6 -r 5.2 x x 77(6 x zt x 49o)]-0->0)
- 0.109 m/s (obtained by trial and error)
- l/zt X xTx sinh (2 x zt x /z/LJ = 1/zt x 0.109 x 10 x sinh (2 x zt x 8/84)
= 0.22 m
 Basic concepts 25

Example 3. A channel has a constant discharge of Q = 12 mVs and a flow

depth of = 1 m. The liquidity index (i.e. the difference between liquid limit
and plastic limit; more information can be found in ASTM 1992) lies in the
range of 0.25 to 0.50 and the voids ratio is about 0.8. The channel bed con¬
sists of heavy homogeneous loamy clay of medium density. Other data are
= 2650 kg/m3, p = 1000 kg/m^, 0 = 20°C, v = lO-^ m2/s.
a) Give a rough estimate of the critical mean velocity (Table 2.4).
b) Give a rough estimate of the cohesion Q (Table 2.5).
c) The average cohesion of the cohesive sediment saturation of soil has
been determined in a geotechnical laboratory and amounts to Q= 15.5 kPa.
What is the critical depth-averaged velocity according to the expression
based on the work of Mirtskhoulava (1988)?
Solution:
a) Critical mean velocity
A first estimate for the critical mean velocity gives ~ 0.8 to 1.0 m/s
b) Cohesion Q,
Applying Table 2.5, the cohesion is about C„ ~20 kPa
c) Critical mean velocity 14
14 = log (8.8 X h/da) X V[(0.4/p) x ((p^ - p) x g x h- 0.6 x 0.035 x C^)]
= log (8.8 X 1/0.004) X V[(0.4/1000) x ((2650-1000) x 9.81 x 0.004
+ 0.6 X 0.035 X 15500)] = 1.3 m/s
(following Mirtskhoulava (1988) 14= 1-21 m/s)

2.4 SCOUR PROCESSES

2.4.1 General
Hydraulic structures that obstruct the flow pattern in the vicinity of the
structure may cause localised erosion or scour. Changes in flow characteris¬
tics (velocities and or turbulence) lead to changes in sediment transport ca¬
pacity and hence to a local disequilibrium between actual sediment transport
and the capacity of the flow to transport sediment. A new equilibrium may
eventually be reached as hydraulic conditions are adjusted through scour.
Scour which may occur at a stmcture can be divided into general scour and
local scour. These possible processes have different length and time scales.
As a first approximation, the scour caused by each process separately may
be added linearly to obtain the resulting scour. In addition, scour in different
conditions of sediment transport can be distinguished. In general, the time
scale of local scour is relatively short. The time-dependent scour process in
prototype situations, however, may be significant. These morphological as¬
pects are considered below.
26 Scour manual

2.4.2 General scour

The time scale for general scour is generally longer than the time scale for
local scour. Commonly occurring examples of general scour are the long
term change in the bed level of a river, scour due to a constriction, scour in a
bend or scour at a confluence.

Overall degradation
Overall degradation results from modification to the stable regime condi¬
tions to which a river has become adjusted. This may be the result of
changes in water or sediment flows in a river. Some examples of features,
resulting either from human interference or from natural changes, which
could respond to the flow regime are:
- Flood embankments, flood detention basins, weirs,
- Channel improvement schemes involving dredging, weed clearance,
mining of sand and gravel,
- Changed in water patterns (confluence or bifurcation of river channels),
- Schemes for transfer of water between river basins,
- Meander cut-offs.
The result of overall degradation at, for example, a bridge site will be the
lowering of bed level, which may place the foundations at greater risk. Al¬
though degradation will increase the risk from scour, in some cases aggrada¬
tion may occur: this will cause higher water levels, and will reduce the risk
from scour. The degradation process that is already in progress has to be
considered as well as the possibility of induced degradation in the future
(e.g. seasonal degradation). Overall degradation can affect a long reach of
river, extending over many kilometers and over periods of many years
(de Vries 1975).
For some conditions, information on possible river bed changes can be
obtained from a one-dimensional morphological model. Besides mathemati¬
cal models of various degree of sophistication, some analytical models
which provide insight into the nature of morphological processes are also
available. More information concerning the prediction of morphological
changes in rivers due to natural causes or human interferences can be found
in de Vries (1993).

Constriction scour
Constriction scour (Fig. 2.8) occurs in confined sections of a river and re¬
sults in a lowering of the bed level across the width of the river. The increase
in depth over a long constriction can be easily computed from the equations
of motion and continuity for sediment and water. For the condition of gen¬
eral movement, the solution of Straub in 1939 reduces to (Laursen & Toch
1956):
 Basic concepts 27

floodplain

Qr Q Q — Clf- + Qf
Figure 2.8. Constric¬
tion scour.

ym,e + K _ 1 (2.10)
/Iq (l-m)P

in which: b - width of the constriction (e.g. width of pier diameter) (m)

B = width of the river (m)
/zq = flow depth upstream of the constriction (m)
m = bIB or 62/-^ o constriction ratio (-)
= scour depth in constriction (m)
P = coefficient; which lies between 0.67 and 0.8
If there is bank overflow with discharge Qj^, Equation (2.10) becomes:

ym,e+^0 _ 1 Q ^2 1 1
bo (l-m)P g-Gy

Bend scour
In general, bend scour depends on local parameters (bend curvature, flow
depth, grain size) and upstream influences (redistribution of flow and sedi¬
ment transport). In the outer part of bends excess scour occurs as the result
of spiral flow. Struiksma et al. (1985) showed that the excess bed scour is
due to this spiral flow and an overshoot phenomenon. The bed adjusts to
changing conditions by a damped response, overshooting the fully developed
solution (Fig. 2.9). The magnitude of this overshoot depends strongly on the
width to depth ratio and the overshoot grows with increasing ratio. Due to
this overshoot effect it is difficult to formulate a simple predictor for the
bend scour although such a predictor is often needed to provide a first esti¬
mate of the scour. In such cases local parameters have to be used, which
implies that the fully developed bend solution (Odgaard 1981) has been
adapted. An example of this method is shown in Figure 2.10.
The time scale for the changes of the cross-sectional profile can be
given by:
52^
T = 0.85 (2.12)
Tt?-S

in which: s = sediment transport per unit width (m^/s)

'F = Shields parameter (-)
28 Scour manual

overdeepening

C
o c
ra
_Q asymptote

OJ 3
"O o

straight curved
reach reach

Figure 2.9. Over

Ym, deepening the bed
bathymetry in a
cross-section A - A river bend.

B/R = 0.8

Km,e + "0

10
AvV
Figure 2.10. Bend scour
A = 10.6 (1 - 2.5 ^/Q (Struiksma & Verheij 1995).

Following Thome (1993), the scoured depth at a bend can be given by the
empirical relation:
= 1-07-log (7?/5-2) for 2<R/B<22 (2.13)

in which R is the radius of curvature of the centre-line.

Equation (2.13) is based on both flume experiments and prototype ex¬
periments in large rivers (flow depths up to 17 m), in which the mean parti¬
cle diameter varied from 0.3 to 63 mm. The error band between the measured
 Basic concepts 29

scour and the scour predicted by using the empirical relation was found to be
25%. For a first estimate the scoured depth in the bend can be assumed to be
equal to the flow depth, which is a somewhat conservative estimate for large
values of R/B.

Confluence scour
When two branches of a river meet, both the angle of confluence and water
level may differ. Though mathematical models are available at present
(1997), these models are of limited value because of the one-dimensional
modelling. For reasons of simplicity, the scour downstream of a confluence
is related to the following variables (e.g. Breusers & Raudkivi 1991):
= % +0-037 0 (2.14)
in which: Cg = coefficient depending on material properties (-),
Co= 1.29-2.24,
/ig = average flow depth of the two branches (m)
= equilibrium scour depth (m)
0 = angle between the two upstream branches
For the Jamuna river in Bangladesh (fine sand), the constant Cg is 1.29
(Klaassen & Vermeer 1988). Based on field data for gravel rivers, Ashmore
& Parker (1983) found that Cg is about 2.24.

2.4.3 Local scour

Local scour results directly from the impact of the structure on the flow.
Physical model testing and prototype experience have permitted the devel¬
opment of methods for predicting and preventing scour at different types of
structures. Information with respect to scour can be obtained by testing
physical models and this approach may be particularly appropriate for un¬
usual stmctures not covered by existing formulae or for field measurements
of scour at existing structures.
This manual highlights the Breusers method because, when this method is
used in combination with computational results of horizontal models or with
measurements obtained from scale models, scour due to three-dimensional
flow can easily be predicted. The accuracy of the scour computation mainly
depends on the accuracy of the measurements of flow velocities and the tur¬
bulence intensities. According to Breusers (1966), the development of the
scour process depends on the flow velocity and turbulence intensity at the
transition between the fixed and the erodible bed. By applying this concept,
the scour prediction can be restricted to one computation; no information is
needed concerning the near bed velocities and bed turbulence in the scour
hole.
 30 Scour manual

When dealing with local scour problems, only the maximum scour depth
in the equilibrium phase is relevant. This is especially true for isolated
structures such as bridge piers, spur dikes, abutments and other permanent
structures (sills, weirs, final closure works). However, there are cases in
which the time factor is important particularly, for example, in the case of
closure of estuary branches.
From model tests on different scales and with different bed materials, re¬
lations were derived between the time scale and the scales for velocity, flow
depth, and material density (Breusers 1966, 1967; Dietz 1969; van der
Meulen & Vinje 1975). The main conclusions were that the shape of the
scour hole is independent of bed material and flow velocity.
The scour process as function of time can be given with reasonable accu¬
racy, provided the prediction of the equilibrium scour depth is satisfactory.

y m.i
(2.15)
m,e

in which: t = time (s)

h = characteristic time at which y^ = 'k (s)
= maximum scour depth at t (m)
= equilibrium scour depth (m)
Y = coefficient (-)
= characteristic length scale (m)
In the development phase (i.e. when t < tj) Equation (2.15) reduces to the
scour relation originally given by Breusers (1966):
f AY
t
(2.16)
\J\ J
Values of y range from 0.2 to 0.4 for two-dimensional flow. More compli¬
cated functions were required to describe the three-dimensional case.
Many definitions of characteristic length scales can be found in the litera¬
ture including, for example, the Kolmogorov length scale which represents
the micro-turbulent eddies. The size of these eddies is in the range of 0.1 to
1 mm. The largest eddies play an important part in any turbulent flow. In
shallow water conditions, the size of the largest eddies is in the order of di¬
mension Aq, which can be used as the characteristic length scale X. For deep
water conditions the dimensions of these macro-turbulent eddies are nearly
equal to those of the hydraulic structure (e.g. bridge pier width, X = b.
Chapter 6).
The time scale can be represented by:
 Basic concepts 31

-4.3
,1.7
(n a,Ui-U^ (2.17)

in which: n = scale ratio (parameter value in prototype divided by para¬

meter value in model)
Uc = critical mean velocity (m/s)
= local depth-averaged velocity (m/s)
«/
= coefficient depending on the local turbulence intensity (-)
A = relative density (-)
Understanding of the physical condition and mathematical modelling of
water and sediment movement in rivers, estuaries and coastal waters has
made much progress in recent years, but this progress has also raised many
new research questions.
Operational dynamic models are available for one-dimensional and two-
dimensional vertical and horizontal depth-averaged simulations. In the early
nineties a morphological model of the development of scour holes behind
hydraulic structures was developed (Hoffmans 1992). This morphological
model, which is based on the Navier Stokes and convection-diffusion equa¬
tions, can simulate two-dimensional scour. At present (1997) this model is
not able to predict three-dimensional scour, which is the type of scour most
frequently encountered in hydraulic engineering. In order to extend this so¬
phisticated model into three-dimensions it will be necessary to make great
efforts with respect to the grid generator and the Reynolds closure problem.
Moreover, currently (1997) the frequent use of three dimensional morpho¬
logical models for long-term morphological calculations is not yet attractive
because of the relatively high computation costs.

2.4.4 Conditions of transport

Clear-water scour occurs when no upstream sediment is present, that is when
the bed material in the natural flow upstream of the scour hole is at rest or
when the bed upstream of the scour hole is fixed.
If the scour is caused by flow that is not transporting sediment (bed load
and suspended load), the depth of scour should approach a limit asymptoti¬
cally. When the approach velocity is greater than the critical mean flow ve¬
locity, the upstream bed is usually covered to prevent the approaching flow
from moving the bed particles. Live-bed scour is scour with sediment trans¬
port over the upstream undisturbed bed. Sediment particles which are con¬
tinuously transported by the flow enter the scour hole. In such cases, the
equilibrium scour depth is smaller than that in clear-water scour conditions.
In general, for the live-bed case, the scour increases rapidly with time
(Fig. 2.11) and then fluctuates about a mean value in response to the bed
features which are being passed. The maximum scour depends on the varia-
32 Scour manual

equilibrium scour depth

Vm
live-bed scour

clear-water scour

time -► Figure 2.11. Scour depth as a function of time.

lions in the depth of flow and is approximately plus half the height of
bed features.
Based on clear-water scour experiments using scale models with small
Froude numbers (Breusers 1966; Dietz 1969), Zanke (1978) distinguished
four phases in the evolution of a scour hole (Fig. 2.12): an initial phase, a
development phase, a stabilisation phase and an equilibrium phase.
In the initial phase, the flow in the scour hole is nearly uniform in the
longitudinal direction. This phase of the scour process can be characterised
as the phase in which the erosion capacity is most severe. Observations with
fine sediments (e.g. Breusers 1966) showed that at the beginning of the scour
hole development some bed material near the upstream scour slope goes into
suspension. Most of the suspended particles follow convectional paths
within the main flow and remain in suspension due to the internal balance
between the upward diffusive flux and the downward flux due to gravity.
Some of the particles will settle and will be resuspended owing to the large
bursts of the turbulent flow near the bed, while some particles with a jump
height smaller than a defined saltation or reference height are transported as
bed load.
During the development phase the scour depth increases considerably, but
the shape of the scour hole does not change. In this phase the ratio between
the maximum scour depth and the distance from the end of the bed protec¬
tion to the point where the scour hole is at its maximum is more or less con¬
stant. Measurements by Hoffmans (1990) showed that the upper part of the
upstream scour slope is in equilibrium, whereas the lower part is still devel¬
oping. The suspended load close to the bed has decreased significantly com¬
pared to the condition in the initial phase. This can mainly be ascribed to the
decrease in the flow velocities near the bed over time, despite the increase of
the turbulence energy. Though bed particles are picked up and carried by the
flow, the time-averaged value of the sediment transport in the upper part of
the upstream scour slope is negligibly, since the contribution of the sediment
transport due to the instantaneous velocities in the downstream direction is
approximately equal to the transport resulting from the instantaneous veloci¬
ties in the upstream direction.
 Basic concepts 33
u
<—>

bed protection

Y = 0.4

In^
phase 1 phase 2 phase 3 phase 4 h

(initiation) (development) (stabilization) (equilibrium)

Figure 2.12. Development of the scour process.

In the stabilisation phase the rate of development of the maximum scour

depth decreases. The erosion capacity in the deepest part of the scour hole is
very small compared to the erosion capacity downstream of the point of reat¬
tachment, so that the dimensions of the scour hole increase more in the
longitudinal direction than in the vertical direction. The more the scour
process continues, the more the flow velocities above the lower part of the
upstream scour slope decrease. In the stabilisation phase the equilibrium
situation for both the upstream scour slope and the maximum scour depth is
almost achieved. The equilibrium phase can be defined as the phase in which
the dimensions of the scour hole do no longer change significantly.

2.5 GEOTECHNICAL ASPECTS

2.5.1 General
In the south-western part of the Netherlands many shear failures and flow
slides have occurred along banks in estuaries. In a number of such cases.
34 Scour manual

more than a million cubic meters of sand slid into the tidal channels. As a
side effect, the sea dike in the vicinity was sometimes damaged or it com¬
pletely disappeared in the area affected. Such slides still occur from time to
time, but as a result of protection works there are now fewer failures affect¬
ing the protected shores. When hydraulic structures are planned in suscepti¬
ble sites, the stability of these structures must be duly safeguarded. With re¬
gard to the cause of shear failures and flow slides, it is generally assumed
that two major factors are involved (Lindenberg & Koning 1981). Firstly,
there is scour due to high flow velocities and turbulence along the slopes of
coastal gullies in the delta region. Secondly, there is the porosity of the sand.
It is assumed that small disturbances can initiate extensive flow slides only
in loosely packed sand. With the exception of the steeper parts, the slopes
that establish themselves after a flow slide are nearly always less steep than
1V:10H and in some cases may be only 1V:20H to 1V:25H.

2.5.2 Porosity of sand

Sand is considered to be loosely packed if the porosity (percentage of voids)
of the sand is higher than a critical value, which in the southwestern part of
the Netherlands is approximately 40%. If shear stresses are exerted on
loosely packed sand, the sediment particles tend adopt a denser packing (Fig.
2.13). As the pores are filled with water, over-pressure occurs in the pore
water, diminishing the effective stresses and thus reducing the frictional re¬
sistance. Depending on circumstances, the increase in the over-pressure may
even be so great that there is no longer any contact between the individual
sediment particles. Thus the frictional resistance is lost and the sand behaves
more or less as a heavy liquid. If such conditions occur in the subsoil adja¬
cent to steep slopes, a quantity of sand slides away, so that a part of the slope
higher up loses its support. This in turn leads to an increase in shear stresses,
followed by a flow slide. In this way very large quantities can liquefy and
flow into the deeper gully or into the scour hole.
The occurrence of a flow slide is dangerous, especially when the affected
soil mass extends up to a hydraulic structure. In such cases the flow slide
initiated by the upstream scour slope can undermine the structure by pro¬
gressive failure under the bed protection, thus leading to a major structural
failure. Therefore it is essential that the porosity or the critical density in situ
is known.

2.5.3 Empirical criteria

The consequences of flow slides, as compared to shear failures, are usually
more serious for hydraulic structures. The length of the expected damage of
the bed protection caused by progressive shear failure can be determined by
 Basic concepts 35

,.WW

!■' " ii
loosely packed water is driven out
shear strength reduction

densily packed
water is sucked
shear strength increment
Figure 2.13. Effect of shear on loose and dense sand.

using standard methods, as developed, for example, by Bishop (1954) and

Fellenius (1947). However, the mechanism of flow slides is still not fully
understood so it is difficult to predict the damage and rate of the backward
erosion accurately. Although little information is available concerning in¬
stabilities, some rough design criteria based on a two-dimensional approach
have been established. Figure 2.14 shows the principle of design with re¬
spect to the minimum length of the bed protection and the upstream scour
slope (p). When the length of the bed protection is too short, or when P ex¬
ceeds the critical upstream scour slope, additional measures have to be car¬
ried out to ensure the safety of the hydraulic structure.

Critical upstream scour slopes

Based on field measurements (Wilderom 1979) regarding dike failures in the
South-west of the Netherlands, Silvis (1988) determined some empirical cri¬
teria for critical slopes (Table 2.6). These criteria are critical values or incli¬
nations for which the probability of geotechnical instabilities approaches 1.
However, these criteria do not include the retarding effect of the bed protec¬
tion, so that the upstream scour slope might be somewhat steeper if the bed
upstream of the scour hole is protected. As the lower value (conservative)
36 Scour manual

hydraulic/
morphological
conditions

maximum slope
scour depth angle p

no protective no
measures

yes no measures yes

(safe)

slope angle
after Instability
angle of
loosely/densely
internal
packed
friction (b

geotechnical
conditions

Figure 2.14. Relation between loading and strength parameters.

Table 2.6. Upper values of upstream scour slopes for soil mechanical instabilities (Silvis
1988).
Bulk density of fine sand Angle of internal Critical slope Phenomenon
friction (j)' angle [3
Very loose 30°(1V:1.7H) 24° (1V:2.25H) Flow slide
Loose 33°(1V:1.5H) 27° (1V:2.00H) Flow slide
Dense 37°(1V:1.3H) 30°(1V:1.75H) Shear failure
Very dense 40°(1V:1.2H) 34°(1V:1.50H) Shear failure

for loose to very loose sand, a critical slope more gentle than 1 to 4 should
be introduced.

Side slopes
The side slopes can be expressed by the angle of repose because the angle of
repose is an upper limit for the side slope in a two dimensional geometry. In
 Basic concepts 2>1

a local scour hole the concave three dimensional surface may result in some
lateral support so that a stable slope steeper than the angle of repose can be
obtained. However, this effect has been observed only in sand with some co¬
hesion caused by clay particles, for example, in the prototype tests in the
Brouwers Sluice, but not in pure non-cohesive sand used in laboratory tests.
The angle of repose depends on the type of soil and also on the compaction
of the soil. For some characteristic values see Table 2.7.

Critical failure length

Wilderom (1979) reported that the failure length due to instabilities of the
sub-soil is strongly dependent on the storage capacity of the channel in front
of the foreland. When the bed of the scour hole downstream of the point of
reattachment is assumed to be horizontal, the failure length can be approxi¬
mated with the following equation presented by Silvis (1988):
f ^
h = yd 1/2 ^-1 (cot Y2 - cot Yj) (2.18)
\ -'m J

+ V2 y^(C0tY2-C0tpJ

in which: = failure length (m)

yd = scour depth at transition of steepest and second slope (m)
ym = maximum scour depth (m)
P. = average slope angle before instability (°)
Y. = sliding erosion slope angle after instability (°)
Y2 = sliding deposit slope angle after instability (°)
Equation (2.18) is based on a two-dimensional storage model in which the
deepening (volume of liquefied sand) equals the deposit 0^ = O2 (Fig. 2.15).

Table 2.7. The angle of repose for different soils.

Soil type Angle of repose (])
Coarse sand Compact 45°
Sand and gravel Firm 38°
Loose 32°
Medium sand Compact 40°
Firm 34°
Loose 30°
Fine sand, silty sand, sandy silt Compact 30-34°
Firm 28-30°
Loose 26-28°
Clay (saturated) Medium 10 20
- °

Soft 0-20°
38 Scour manual

profile before instability

profile after instability

>^2
o.1-^2
=o
Figure 2.15. Two-dimensional schematisation of a flow slide.

Table 2.8. Statistical parameters of shear failures and flow slides.

Storage parameters Flow slides and shear- Shear failures Flow slides
failures (Silvis 1988) (Konteretal. 1992) (Konter et al. 1992)
Mean values Extreme values Extreme values
cot 3.7 41 22 41 22
cot Yi 2.7 2 2 2 2
cotY2 16.3 8 8 20 20
cot 5 40
Vdlym 0.43 0 0 0 0
4/3'm (no storage) 1.7 2.0 3.0 8.0 9.0
(storage) 1.5 2.0 2.9 7.5 8.3
'Protected slope. 2Unprotected slope.

Table 2.8 gives an overview of mean and extreme values of the parame¬
ters in the aforementioned storage models. These results are based on ap¬
proximately 200 instabilities caused by shear failures or flow slides
(Wilderom 1979). Applying default values, the computed failure length must
be considered as a first estimate (L^ ~ 2y^), since specific information re¬
garding the sub-soil (e.g. porosity of sand, clay layers) has not been taken
into account.
A rough but conservative failure length is obtained if - 0. With this as¬
sumption, Equation (2.18) simplifies into:
4 = */2y„(cotY2-cotpj (2-19)
The failure length is reduced when the volume of the scour hole limits the
storage capacity of the flow slide. For triangular scour hole, the failure
length can be given by (Silvis 1988):

Ls = ^(cot Y2 + cot 5) (ayl + byl)-ay^ -by^ (2.20)

in which: a = cot + cot 5 (-)

b = cot 72 - cot Y, (-)
6 = slope angle downstream of the deepest point of the scour
hole (°)
 Basic concepts 39

De Graauw (1981) derived a similar equation which was based on a less ac¬
curate schematisation. The execution of a thorough in-situ soil investigation
may lead to the conclusion that the subsoil consists of dense sand. In such
cases the value of can be decreased significantly, leading to a shorter bed
protection length. It should be noted that the critical failure length depends
strongly on yi (slope angle after instability). At present (1997), no relations
are available which relate 72 accurately to soil parameters. The use of the
following rough values is recommended: flow slides 15 < 72 < 20 and shear
failures 6 < 72 < 8.

2.5.4 Example
A hydraulic structure is built on loosely packed sand. The bed downstream
of the hydraulic structure is protected against current and eddies. The up¬
stream scour slope is about cot = 4.0 and protected with rock. In addition,
to ensure the safety of the structure, it is protected against the occurrence of
flow slides and of shear failures with an extreme probability. The angle of
internal friction amounts to ([)' = 33°.
a) What is the critical slope angle according to Silvis (1988)?
b) What is the minimum length of the bed protection?
c) What would be the critical failure length if the subsoils consisted of
both loosely and very densely packed sand? Up to 5 m below the bed pro¬
tection the soil is loosely packed.
d) What would be the critical failure length if the scour slope downstream
of the deepest point of the scour hole was about cot 6 = 25? Other data are:
cot 7i = 2.0, cot 72 = 25, = 0.2y^.
e) What would be the critical slope angle if the sub-soil consists of
densely packed sand and what would be the corresponding critical failure
length following Silvis (1988)? ((])' = 38°).
Solution'.
a) p, = 27°
b) Lj = (no storage capacity) or = 1.5y^ (storage capacity)
c)
d) 4 = V[(cot 72 + cot 5) X (a X y2 + ^ X y^)] _ a X X
a = cot + cot 5 = 4 + 25 = 29
b - cot 72 - cot 7i = 25-2 = 23
4 = V[(25 + 25) X (29 X y2 + 23 X (0.2 x yj2)]_ 29 x y„ - 23 x 0.2 x y„
= 38.7 X y„ - 29 X y„ - 4.6 X y,„ 5. ly„
e) Pc =31° and Lj = 2.0y,„
CHAPTER 3

Sills

3.1 INTRODUCTION

In an estuary or a river, a sill may be the initial foundation or the lower part
of a structure that has to be constructed on a bed of alluvial material.
Sometimes a sill is used to reduce the mixing of different types of water in
an estuary. In an estuary a sill has to be designed for flow in two directions:
flood flow and ebb flow. In rivers, for example, a sill may be used as part of
a scheme to maintain a minimum water level.
The flow pattern downstream of sills is discussed in Section 3.2. Section
3.3 summarises relations used to predict the maximum scour depth for sills
with horizontal bed protection (Fig. 3.1). The upstream scour slope and the
gradual undermining at the end of the bed protection due to high turbulence
level are treated in Section 3.4. Section 3.5 deals with additional measures to
reduce the risk of failure and Section 3.6 describes some field experiments
with loosely packed material. Finally, some examples are discussed (Section
3.7).

3.2 FLOW CHARACTERISTICS

The bed in the direct neighbourhood of hydraulic structures is generally

protected against current, waves and eddies. The length of the bed protection
depends on the permissible scour (maximum scour depth and upstream scour
slope) and the geotechnical conditions of the soil involved (densely or
loosely packed sand). When the length of the bed protection increases, the
scour process is less intense due to the decay of turbulence energy and
the adaption of the velocity profile downstream of the hydraulic structure.
However, the measures required for protection against scouring are costly,
especially when the bed protection is constructed under water. The two-
dimensional flow pattern downstream of hydraulic structures where the flow
above a sill is subcritical is shown in Figure 3.2. Downstream of a hydraulic
structure and in a scour hole the separated shear layer appears to be similar

40
 Sills 41

acceleration deceleration
zone zone
k--k

Ym

bed protection
scour hole

subcritical flow hydraulic

supercritical flow jump
7
subcritical
flow
short-crested sill

flow separation
recirculation zone

Figure 3.1. Definition

broad-crested sill sketch of sills.

to a simple plane mixing layer. Initially, the axis of the mixing layer is
slightly curved due to the influence of the bed.
The curvature increases in the downstream direction, especially near the
point of reattachment (reattachment length equals about six times the sill
height). A reverse flow develops with a flow direction opposite to the main
flow direction. In both the mixing layer and the recirculation zone the turbu¬
lence energy is relatively high compared to uniform flow conditions. Vor¬
tices with a vertical axis will occur when the flow pattern is influenced by
vertical end or wing walls because the bed material is then picked up by the
rotating ascending current in the vortex and is thrown out sideways. The in¬
tensity of such a vortex street may attain such a magnitude that it will en¬
danger the stability of the structure unless effective protective measures are
taken. The flow in a scour hole is more or less two-dimensional when the
influence of a vortex street on the scour process is marginal, i.e. when the
velocity gradients in the transverse direction are relatively small.
42 Scour manual

relaxation zone
Ou U

mixing layer

recirculation zone
sill
= reattachment point \
^////////7//////7f.
new wall-boundary layer
u

bed protection Ym

/
scour ho e

Figure 3.2. Schematised flow pattern below a dam.

3.3 DUTCH SCOUR DEPTH MODELLING

3.3.1 General
Eor a designer, the most important scour parameter is the maximum scour
depth in the equilibrium phase. However, in deltaic areas, generally charac¬
terised by large flow depths, the time factor is also important, especially in
the closure of estuary branches. The magnitude of the maximum scour depth
depends on the bed shear-stress and the turbulence condition near the bed,
on one hand, and on sediment characteristics (density of the bed material,
sediment-size distribution, porosity, cohesive or non-cohesive bed material)
on the other hand. Several hundreds of tests in which no sediment transport
from upstream was present were carried out at Delft Hydraulics (Delft Hy¬
draulics 1972, 1979; Buchko 1986; Buchko et al. 1987). In general the ex¬
perimental relations derived from these are considered to be reasonably
valid. The scour process as function of time can be given with reasonable
accuracy, provided the equilibrium scour depth iy^ ^) is greater than the ini¬
tial flow depth (/zq) (Eig. 3.3):
 Sills 43

0.01 0.01
0.001 0.01 0.1 1 0.001 0.01 0.1 1 10 . 100
two-dimensional scour ( b/B = 0.0) three-dimensional scour (t>/6 = 0.1)

10
. D//7o =0.6 (M863)
7 = 0.6 Ym.e^^Q =5.0

= =2.0
--y = 0.8-1

Brjusers equil ibrium approach ibrium approach

0.01
0.001 0.01 0.1 1 100 0.001 0.01 0.1 1
three-dimensional scour (6/6= 0.1) three-dimensional scour (6/S = 0.1)

Figure 3.3. Maximum scour depth as function of time.

= \-e 7m,e ) (3.1)

y.
in which: r = time (s)
fj = characteristic time (s) at which
- maximum scour depth at t (m)
y = coefficient (-)
In the development phase when t is smaller than Equation (3.1) reduces to:
( V
Zm t
(3.2)
K V^I J
44 Scour manual

Based on an extensive analysis of the measured bed levels at which the

maximum scour depth was about O.S/iq, Breusers (1966) reported that for 2D
scour the averaged value of the coefficient y amounted to 0.38. According to
Mosonyi & Schoppmann (1968) and Dietz (1969), the coefficient y lies
within the range of 0.25 to 0.40, which can be considered as a confirmation
of Breusers’ results (Table 3.1).
In 3D situations, for example, scour in a horizontal constriction, a vortex
street (vortices with a vertical axis) occurs (Fig. 3.4). In this zone the flow is
very turbulent and large vortices intermittently erode and transport bed ma¬
terial. For 3D flow, y is strongly dependent on the degree of turbulence gen¬
erated by the vortex street. Figure 3.3 shows the results of about 110 tests
from the systematic research (van der Meulen & Vinje 1975), as a function
of yjh^ versus t/t^ for the deepest points of the scour hole in the cross sec¬
tions (D/Tiq = 0.0, 0.3 and 0.6). The scatter in the results is generally within
reasonable limits, so that the scour patterns were similar in shape as the
scour progressed. However, a different coefficient y was required for the
prediction of the maximum scour depth as a function of time.
For small scour depths the exponential function of the form of Equation
(3.2) gave a good fit if y values larger than 0.38 were used. In the initial and
development phases a conservative value for the maximum scour depth is
obtained if y is 0.4 for two as well as three-dimensional situations. In the
equilibrium phase the maximum scour depth approaches a limit. To obtain
equilibrium conditions, a relatively long period of time is needed, especially
for experiments with small Froude numbers and no upstream sediment sup¬
ply (clear-water scour), so with scale models the equilibrium situation was
not always achieved (Delft Hydraulics 1972). Following Dietz (1960) the
equilibrium scour depth is given by:
3^ ^ (oUq-U,
(3.3)
ho
in which: = critical mean velocity (m/s)
Uq - QIA, mean velocity (m/s), Q is discharge (mVs), A is cross
section (m^)
(0 = 1 + 3rQ (Dietz 1969), is relative turbulence intensity (-)

Table 3.1. Coefficient y.

Investigator Flow condition
Breusers (1966) 0.38 Two-dimensional
Mosonyi & Schoppmann (1968) 0.27-0.35 T wo-dimensional
Dietz (1969) 0.34-0.40 T wo-dimensional
van der Meulen & Vinje (1975) 0.4-0.8 Three-dimensional
 Sills 45

1:2.5 ^ 1:2.5
Id

L 2hci 2ho
->)<—>)<-S+ x = ho
b = 2h, .CCCCUCCCCCCCCC
xcccccccc
Wf
Karmin
Vortex
Street
8ft
6 ho
0.4 fto
8=20/7,
1
10.2 fto
\
V

a/
o.Uq = Uf Uf
Un a
0.2

0.4

0.6

0.8

Figure 3.4. Definition sketch of Ui.

Measurements by Dietz (1969) showed that the average value of the turbu¬
lence coefficient oo was about co = 2/3 + 2rQ. Dietz reported that the maxi¬
mum value of to amounted to O)^ = 1 -t- Sr^. Following Popova (1981), the
turbulence coefficient is given by to = 0.87 -t- 3.25 Fr -1- O.SrQ (Fr = Froude
number just before the scour hole). According to Rossinskiy (1956; e.g.
Blazejewski 1991), co lies in the range of 1.05 to 1.7, depending on the geo¬
metry upstream of the scour hole. Upstream sediment supply and/or the
presence of more resistant layers will reduce the maximum scour depth. If
sediment is supplied from upstream, values predicted by Equation (3.3) seem
to be high in the prototype situation.
46 Scour manual

The rate at which the scour hole approaches an equilibrium phase depends
strongly on the turbulence intensity in nonuniform flow and to a lesser extent
on material characteristics. The characteristic time, a non-steady flow, such
as tidal flow, and the influence of upstream supply of sediment will be dis¬
cussed in subsequent sub-sections.

3.3.2 Characteristic time

On the basis of dimensional considerations, the time needed to erode a vol¬
ume V per unit width could be given by ? = Vis, in which the area of the
longitudinal scour section is related to the maximum scour depth (V = c^ y^,
in which c^ is a geometrical factor and 5 is the sediment transport). The de¬
pendence of the characteristic time on the hydraulic conditions and material
characteristics has been investigated by several researchers. From model
tests it appeared that for steady flow the influence of the various parameters
could be described by:

(3.4)

in which: K = 330 hours m^ W in hours (or 1.19 x 10*’ m^ tj in

seconds)
a = coefficient depending on the flow velocity and turbulence
intensity (-)
A = relative density (-)
The characteristic time was originally expressed in hours, and the coefficient
K is not dimensionless. Using the invariables g (acceleration of gravity) and
V (kinematic viscosity). Equation (3.4) can be rewritten as:

/zq K^A^
(3.5)
Uq Fr'^ Re'^
in which: Fr = Froude number (-), g is acceleration of gravity
^1
= A7(g‘ '*3v0-43) {K in m2-3/s3-3) coefficient (-)
Re = U^h^v, Reynolds number (-), v is kinematic viscosity
h = characteristic time at which - /iq (s)
a u = coefficient (-), to be determind as a„ = a - UJUq
Dietz (1969) performed extensive research on two-dimensional scour down¬
stream of horizontal beds and low sills. Several non-cohesive materials
(sand, lignite and polystyrene) were used in the experiments on model scale
in which the initial flow depth varied from 0.125 to 0.25 m. The research of
Dietz (1969) confirmed the considerations of Breusers. Although the relation
 Sills M

Table 3.2. Empirical coefficients and exponents in Equation (3.5).

Investigator /f./lO*’
Breusers (1966) 0.94 1.62 4.0 2.7 0.3
Dietz (1969) 9.96 1.5 4.0 2.5 0.5
van derMeulen & Vinje (1975) 12.9 1.7 4.3 2.87 0.43
Zanke(1978) 1.33 4.0 2.67 0.33
de Graauw & Pilarczyk (1981) 17.1 1.7 4.3 2.87 0.43

is identical, Dietz proposed different values for the empirical coefficient

and the exponents (Table 3.2).
Hoffmans (1992) showed that the scour relation based on the sediment
transport presented by Zanke (1978), which is of dimensional homogeneity,
can be written into the form of Equation (3.4). However, the coefficient K is
not constant, but is a weak function of the sedimentological diameter. Van
der Meulen & Vinje (1975) studied the three-dimensional scour process
downstream of a partial channel constriction and reported that K was equal
to 250 (based on more than 100 tests with a hydraulically smooth, medium
and rough bed protection). They concluded that the shape of the scour hole is
independent of bed material and flow velocity, and that Equation (3.4) is
equally applicable to three-dimensional situations, provided a is assigned
correctly. For three-dimensional scour the characteristic time is not constant
at each cross section, owing to the three-dimensional character of the scour
hole. Therefore it is necessary to determine the development of scour in
many longitudinal sections separately.
De Graauw & Pilarczyk (1981) found that K was equal to 330 for tests
with rough bed protection. In addition, they specified the factor a for both
two and three-dimensional flow where the scour depth is at its maximum.
The influence of the roughness of the bed protection on the characteristic
time can be taken into account by fitting either the coefficient K or the
flow and turbulence factor a. De Graauw (1981) analysed the predictability
of by applying 150 to 200 tests in a scale model and found that the relative
standard deviation a(ti) was about 30%. The accuracy of the computed
(maximum scour depth) is characterised by a relative standard deviation of
about 10%.

3.3.3 Relative turbulence intensity

To analyse the decay of turbulence in the relaxation zone, an analogy with
the decay of turbulence energy and the dissipation of grid turbulence can be
used (Launder & Spalding 1972). If the geometry of the tests consists of a
horizontal bed where the flow is subcritical above a sill, the following rela-
48 Scour manual

tion for the relative turbulence intensity Tq can be deduced (Hoffmans &
Booij 1993a):
V2,^ V108
D L-6D
= J 0.0225 1- - + 1 + 1.45 for L> 6D
h.-0 y 6.61k (3.6)
in which: C = Chezy coefficient related to bed protection (m’^Vs)
D= height of sill (m)
8 = acceleration of gravity, g = 9.81 m^/s
L = length of bed protection (L > 6D) (m)
The length of a safe bed protection will always have to be extended beyond
the point of reattachment. Although some characteristics of three-dimen¬
sional flow and additional phenomena, such as vortices with a vertical axis,
have not been taken into account in Equation (3.6), promising results for
three-dimensional flow fields have been obtained (Fig. 3.5). For example, in

■■ D/ho=0.20

0.3 — D/Hq^O.SO
♦ 0.17 <D//jo< 0.23
0.2 H 0.30 <D//io< 0.35

0.1

10 20 30 40 50

L/ho-► (-)

two-dimensional experiments (rough)

r — D/ho=030 (rough)
Cc Iculations I — D/ho=0.60 (rough)

0.3 L ♦ 0.30 (rough)

0 D//Jo=0.60 (rough)
ft D/ho~0.30 (smooth)
0.2
V D/ho=0.60 (smooth)

0.1

Figure 3.5. Rela¬

0.0
20 30 40 50 tive turbulence in¬
L/ho-► (-) tensity as function
three-dimensional experiments
of zy/iQ.
 Sills 49

the centre of the flow where the scour depth is about at its maximum, the in¬
fluence of the vortex street on the scour process is marginal compared to the
influence of the maximum of the velocities.

3.3.4 Scour coefficient

The magnitude of the coefficient a (flow and turbulence coefficient) de¬
pends largely upon the upstream geometry and can be interpreted as a meas¬
ure of the erosion capacity. For a storm surge barrier a is determined for the
greater part by the height of the sill (relative to the initial flow depth), and
for a dam with a spillway by the presence or absence of energy dissipators.
The length and roughness of the bed protection downstream of the structure
also play an important role in the determination of a.
Several empirical expressions for a have been deduced from the tests in
the systematic series for both two and three-dimensional scour (de Graauw
1981; de Graauw & Pilarczyk 1981).
According to Breusers (1966, 1967), in a scour hole, turbulence, which
plays an important role in the scour process, fully depends on the flow veloc¬
ity and turbulence level at the transition between the fixed and the erodible
bed (a; = (1 + 3rQ)c^ with c^, is a correction factor for the velocity profile). In
three-dimensional situations the factor a is dependent on the turbulence
level which in turn depends on the three-dimensional flow pattern (Fig. 3.4).
Jorissen & Vrijling (1989) reported that for hydraulically rough flow condi¬
tions the turbulence effect a, can be expressed by:
a;=15-l-5rQ (3.7)
A more general relation for the local turbulence coefficient is (Hoffmans &
Booij 1993a):
a, =1.5 + 4.4a-o/c (3.8)
in which: ffi = roughness function related to bed protection,/(- = C/Cq
C = Chezy coefficient related to the bed protection upstream
from the scour hole
Cq = 40 if C < Cq then/^ = 1
For hydraulically rough flow conditions, that is for C - 45 m'^Vs, Equation
(3.8) reduces to Equation (3.7). Equation (3.8) was based on about 250
measurements in a two and three-dimensional flow pattern for > 0.05. For
two-dimensional flow, the local depth averaged flow velocity equals ap¬
proximately the mean flow velocity (Ui ~ Uq thus ~ a), whereas for three-
dimensional flow Ui and a, depend strongly on the geometry upstream of the
scour hole. Consequently, the development of the scour process in a three-
50 Scour manual

dimensional scour hole can be determined at any cross section, provided a,

and Ui are known variables. (Note that a is linked to Uf^.)
Combining Equations (3.6) and (3.8), the local turbulence coefficient can
be given as a function of the geometry upstream of the scour hole. Figure 3.6
shows a, (a for two-dimensional flow) as a function of Lth^ for several
heights of the sill. Hence, a increases strongly with sill height, which dem¬
onstrates that scour increases owing to the decelerated flow downstream of
the sill. The value of a may be reduced by either lengthening the bed pro¬
tection or by making the bed rougher. The flow pattern downstream of a sill
combined with a horizontal constriction is three-dimensional, at least to
some degree, due to the vortex street. In such cases a is significantly greater
than the value of a for two-dimensional flow. When no information is avail¬
able with respect to the three-dimensional flow pattern, the use of design
curves presented by de Graauw & Pilarczyk (1981) is recommended (Fig.
3.6). These curves are valid for D/h^ < 0.8 and Llh^ > 5 and bIB = 0.1, where
b is the width of abutment and B is the width of the flow. A smooth bed
protection results in a faster scour process because the flow above a smooth
bed has greater momentum than the flow above a rough bed. For a hydrauli¬
cally smooth bed, a has to be increased by a term that lies in the range of 0.3
to 0.5 = arough + 0.4).

9.0
I A# M 847 III fig. 41 (moderately rough)
(-) □ M 847 I fig. 63 (moderately rough)
- 8.0 x+ M1321 (syphon in the Grevelingen)
O M 505 11 fig. 20 (Haringviiet)
Q Q935 50% T8 T10
a * Q935 30% T3 T6
7.0 V Q935 10% T1 T2T7

3D (5/B = 0.1) - D/h 0 = 0.0

6.0 -D/ho = 0.30

■ . ■ D/ho = 0.60

5.0
1-
A^
A V

4.0

3.0 )>•
2D

2.0 1= □

1.0
0 10 20 30
L/ho ■* (-)

Figure 3.6. a as function of

 Sills 51

In this section and the foregoing sections it is assumed that the main flow
is subcritical (Froude number is smaller than 1) during all construction
stages of the closure. In an estuary this will be often the case, but in a clo¬
sure of a river branch supercritical flow is possible.

3.3.5 Non-steady flow

In tidal areas the following aspects should be studied:
- The frequency distribution of the discharge, which is influenced by
wind (wind set-down and wind set-up) as well as by neap tide and spring
tide,
- The relation between discharge and the mean velocity during a tide,
- The effects of a series of tides on the development of a scour hole,
- The effects of one tide with a constant flow velocity,
- The schematised effects of one tide with a constant flow velocity.
These aspects have been studied and tested in the field verification in the
Brouwers Dam sluice (Fig. 3.7).
A scour hole will develop under non-steady flow conditions and therefore
the equations for steady flow conditions are extended to unsteady flow
conditions. In situations where the equilibrium phase has still not been
reached, the maximum scour depth can be given by:
(
yjo t
(3.9)
h,(0) \h,u j

in which: h^lfS) - tide-averaged flow depth (m)

= characteristic time at which = hifOi) (s)
Equation (3.4) can be adapted for unsteady flow by taking into account a
succession of infinite short-lasting steady situations. If the flow is cyclic (i.e.
tidal movement) the characteristic time can be represented by:
/:/io(0)A>-7
(3.10)
l|((xI/o(0-I/,)^-3^^
h,(t)

in which: T =t2- ty, half tidal period where aU^ > (s)
ty - time at which aU^ first exceeds during flood tide (s)
^2 = time at which aU^ drops below during ebb tide (s)
In Equation (3.10) the mean velocity and the flow depth are the only vari¬
ables which depend on time. If h^it) and U^it) are given as a function of
time, y can be calculated by numerical integration. In tidal areas with a
dominant tidal period of 12 hours, a time step of 0.5 hour is recommended
52 Scour manual

North Sea Crevelingen Lake

side construction
side wail
M.H.W.«=+1.4
+1,0 y

M.LjW-

I
121

elevation

m
P* 1.4
B=5C =
A = 54 m ^
top-view
prototype situation sizes in m
1 : 2

(m)

10
■S" experiment A
Ym
X experiment B
— calculations

Hq = 10m
t-\ = 1000 hours

y m,e = 35m
y =0.5

10 10
nett scour-time (hours)
Figure 3.7. Brouwers Dam Sluice.

for the numerical integration. In rather deep water the variations in h^^it) are
relatively small. However, the variations in Ugit) may be quite important. As
a result of ebb and flood, both the mean velocity and the sediment transport
vary in time. The scour process in tidal areas can be simulated reasonably by
applying a characteristic constant mean velocity. When the mean velocity in
 Sills 53

a tidal area is approximated by a sine function, the following relation is ob¬

tained (e.g. Hoffmans 1992):
(3.11)
in which: = characteristic mean velocity (m/s)
, = maximum velocity during a tide (m/s)
T) = coefficient (-),ri = 0.75-0.85
The coefficient r) is almost independent of the type of sediment transport
equation. For a first approximation, the characteristic time can be approxi¬
mated by:
/i:/i2(0)Ai7
(3.12)
(at/, -C/,)«

3.3.6 Upstream supply of sediment

The theory in the foregoing sections has been based mainly on the results of
physical scale models in which generally no upstream supply of sediment is
present. However, in prototype situations an upstream supply of sediment is
often present and in such cases the development of a scour hole is reduced,
because part of the erosion capacity of the flow is used to supply sediment
from upstream and is therefore not available for the transport of sediment in
the scour hole. An approximate calculation method for this reduction is dis¬
cussed here. The volume of a scour hole can be expressed by:
^(0 = c„y™(0 (3.13)
in which: - shape factor of scour hole (-), = 20
V{t) = volume of scour hole per unit width (mVm)
yjf) = maximum scour depth (m)
Mosonyi & Schoppmann (1968) reported that the shape factor is approxi¬
mately 6 for dear-water scour experiments. In their experiments the flow
was nearly uniform at the end of the bed protection. When the scour hole is
schematised as a triangle with an upstream slope of 1V:2H and a down¬
stream slope of IV:8H, the shape factor is 5. However, in a situation with
a considerable supply of upstream sediment the upstream scour slope is
about 1V:4H and the downstream scour slope 1V:40H, giving a shape factor
of 22.
In the stabilisation and equilibrium phases the shapes of the scour holes
are not similar. In these phases the distance from the end of the bed protec¬
tion to the cross section, where the scour depth is at its maximum, increases
more than the maximum scour depth. Therefore the shape factor is not con¬
stant. The upstream sediment supply is defined as the volume of the sedi-
54 Scour manual

ment particles, including the porosity of the sediment, so that the volume of
the scour hole is reduced by;
V(t)^V(t)-qJ (3.14)
in which: V^(t) - reduced volume of scour hole per unit width (mVm)
= reduction transport per unit width (including porosity)
(m^/s)
t - time (s)
Consequently the scouring capacity can be given by:
dV; dV
(3,15)

The reduced maximum scour depth on prototype scale can be deduced either
by combining Equations (3.13) and (3.14) or directly from Equation (3.15).
As given by Konter & van der Meulen (1986) the maximum scour depth as a
function of time is (Fig. 3.8);
y^ (tp + Atp ■')— ym,p^(tp'') y^ (tc + Atc ■')— y^m,c^(tc ')
^ (3.16)
At c

in which: t^ - time referring to conditions where = 0 (s)

tp - time referring to live bed conditions (s)
At = time step (s)
For the application of this method of reduction, it is necessary to determine
the magnitude of q^, which can be obtained with the aid of a scale model or
by using morphological models based on sediment transport theories. For a
first approximation the unknown q^ at the end of the bed protection can be
represented by:

(3.17)

(<75 = 0)

(m)
prototype
At

Vm
V'.m.p

Af, At

f, Figure 3.8. Reduction

f method.
 Sills 55

in which: 5^ = bed load (mVs)

= suspended load (m^/s)
= coefficient (-)
r|j = coefficient (-)
Van Rijn (1984) verified several transport predictors, using 486 sets of river
data and showed that bed load and suspended load could be described at best
by:
V-V M.2
-^50
Sf^ = 0.005 [/q/iq (3.18)
Agd,50 y V ^0 y

Y'V
-*50
= 0.012 Uh, (3.19)
50 y V ^0 y
where D* is the sedimentological diameter (Eq. 2.6).
In the Eastern Scheldt in comparison to bed load (5^ ~ 105^), suspended
load is dominant, largely due to the relatively large flow depths (15 to 40 m)
and the fine sediments in the deltaic area. Computational results (Stroeve
1994) have shown that the reduction method yields reasonable results for the
scour process in the Eastern Scheldt when the magnitude of q.. is taken as
being equal to the sum of bed load and part of the suspended load (r)^ = 0.15)
(Fig. 3.9 and Table 3.3).

Table 3.3. Hydraulic data, Eastern Scheldt (Stroeve 1994).

Section Section Section
RO960 RO1680 HO620
Experimental data
Length of bed protection (m) 650 650 650
Flow depth (m) 40 26 25
Sill height (m) 16.5 17.5 15.5
Mean particle diameter (pm) 200 200 200
Density fluid (kg/m^) 1025 1025 1025
Density sediment (kg/m^) 2650 2650 2650
Computational data
Relative turbulence intensity (-) 0.16 0.23 0.19
Turbulence coefficient (-) 2.29 2.63 2.47
Maximum (local) velocity' (m/s) 1.20 1.56 1.49
Characteristic flow velocity (m/s) 0.96 1.25 1.19
Critical velocity (m/s) 0.4 0.4 0.4
Reduction transport (m^ per day) 4.6 17.8 14.3
•Tidal averaged value.
 56 Scour manual

Eastern Scheldt
Hammen East (section 620)
bed level (NAP in m)

■O" measurements
-15 — reduction method

-30

-45

-60

-75
1988 1992 1996 2000
time
Eastern Scheldt
Roompot East (section 960)
bed level (NAP in m)

4. measurements
-15 — reduction method

-30

-46

-60

-75
1988 1992 1996 time ^000
Eastern Scheldt
Roompot East (section 1680)
bed level (NAP in m)

■}■ measurements
— reduction method

30
■s-w

45

60

-75
1988 1992 1596 time 2000
Figure 3.9. Scour process in the Eastern Scheldt.
 Sills 57

3.4 UPSTREAM SCOUR SLOPES

3.4.1 General
The upstream slope of the scour hole (upstream scour slope) determines the
stability of the upstream part of the scour hole and the adjacent bed protec¬
tion. In general, this part of slope reaches an equilibrium value which is less
steep than the initial tangent at the end of the bed protection. When this
slope exceeds a critical value in non-cohesive sediments, a shear failure can
occur or liquefaction of the soil under the bed protection may even be pos¬
sible. A hydraulic and morphological relation for upstream scour slopes has
been derived (Hoffmans 1993). This was calibrated by using a large number
of flume experiments, in which the material properties and the hydraulic and
geometrical conditions are varied.

3.4.2 Hydraulic and morphological stability criterion

The stability of the upstream scour slope is the result of the interaction be¬
tween fluid motion and material properties. The equilibrium situation of up¬
stream scour slopes for non-cohesive material is achieved by equating bed
load due to the instantaneous bed shear-stresses sloping downward and bed
load due to the instantaneous bed shear-stresses sloping upward. Using a
probabilistic bed load model for bed load transport, a semi-empirical relation
for the slope angle p has been found (Hoffmans 1993; Hoffmans & Pilar-
czyk 1995). The equilibrium condition of bed load transport due to up slope
and down slope shear stress resulted in a relation with a shear stress factor
and a turbulence factor representing the skewness of the instantaneous hed
shear-stress (Eig. 3.10):
f \

P = arcsin 2.9x10-4 + (0.11 + 0.75ro)/c (3.20)

V ^8^50

in which: d^Q = median grain size (m)

Tq = relative turbulence intensity at end of bed protection (-)
U = depth-averaged velocity at end of bed protection (m/s)
Erom Equation (3.20) it appears that turbulence is important in the develop¬
ment of the upstream scour slope. With more turbulent upstream conditions
much steeper slopes are found. A smooth bed protection results in a steeper
upstream scour slope because the near bed velocities have greater momen¬
tum and cause a more rapid expansion of the flow in the scour hole. In gen¬
eral, the influence of the shear stress factor can be neglected provided the
flow velocities are smaller than 1 m/s. In deltaic areas with fine sediments
and flow velocities larger than 1 m/s, the shear stress factor determines the
58 Scour manual

6
ro = 0.1
tc = 1.0
ro = 0.2
5 (rough)
(-) I'd = 0.3

\
cot P
*
rc=1.5 (smooth)

3
dgQ = 200 urn
A = 1,65
2
<() '= 30°

4) '= 40°
1

unstable
0
0.0 0.5 1.0 1.5 2.0 2.5
Uq -^ (m/s)

6
Uo7(A g dso) = 100 ■
fc = 1.0
Uo/{Agdso) = 250
5 (rough)
Uq/ (A g rfso) = 500 .
5 (smooth)
4
I
cot P
3

2
4)’=30°

0 '= 40°
1

unstable
0
0.0 0.1 0.2 0.3 0.4 0.5

ro-► (-)

Figure 3.10. Upstream scour slopes.

Upstream scour slope to a large extent. For the design it is obvious that the
value of P has to be much less than the critical value of the natural slope of
sediment in water. Equation (3.20) yields results that compare reasonably
well with measured developments of a scour hole in the case of a subcritical
flow upstream of the scour hole (Fig. 3.11).
 Sills 59

Z ■

Figure 3.11. Calculated and meas¬

measurements ured cot p.

3.4.3 Undermining
The gradual and dangerous undermining of the edge of the bed protection re¬
sults from the turbulence energy and the erosion capacity of the flow in the
recirculation zone. When P exceeds a critical value (angle of internal fric¬
tion), the bed protection could be gradually undermined owing to small scale
shear failures. In addition to the gradual undermining, a sudden undermining
may occur (shear failures and flow slides), leading to a possible failure of the
hydraulic structure upstream of the scour hole. However, these phenomena
are strongly dependent on geotechnical conditions (de Groot et al. 1992).
According to Konter et al. (1992), a failure length equal to eight times the
maximum scour depth can be conceived as a conservative extreme length
when the subsoil consists of fine and unconsolidated sand (Section 2.5). For
densely packed sand the maximum failure length can be decreased consid¬
erably and is estimated to be = 2y,„. If the soil is non-homogeneous, no
general rules are available, although = 8y„, can be taken as conservative
for inhomogeneous soil profiles too.

3.5 ADDITIONAL MEASURES

Hydraulic structures placed in waterways or coastal seas are often stream¬

lined in order to reduce the drag exerted by flow and to reduce the size of the
wake and turbulence intensity. Streamlining by means of deflectors and
guide vanes, however, is effective only when the hydraulic structure is
aligned with flow to within narrow limits.
When the subsoil has the potential to liquefy, measures have to be taken
to ensure the safety of hydraulic structures. The most obvious method is to
focus the design procedure on a bed protection length such that, with a cho¬
sen bed protection, backward erosion does not influence the structural sta-
60 Scour manual

bed protection rock protection

71/.
<5m

71/.
<5m

Figure 3.12. Additional measures.

bility. Local scour downstream of hydraulic structures can be reduced by

lengthening and roughening the bed protection. Consequently, the probabil¬
ity of flow slides and shear failures decreases. When the risk is too great, for
example, due to lack of space or for economic reasons, other measures can
be considered. By protecting the upstream scour slope with rock or slag and
by compacting the subsoil, shear failures and flow slides can be avoided.
The construction of a retaining wall, i.e. sheet piling, at the edge of the bed
protection, can also be considered. It is always advisable to monitor the de¬
velopment of the scour hole frequently during construction and operation, so
that necessary measures can be taken in time to prevent dangerous situations
from arising. During the construction phase, equipment and materials are
available to stabilise the upstream scour slope in order to minimise the scour
that may occur under operational conditions. In the case of a land-based bed
protection, it is recommended that an artificial scour hole should be made
beforehand and that the upper part of the upstream scour slope should be
protected by gravel or slag (Pilarczyk 1984). In general, this will lead to a
shorter and more reliable design of the bed protection. Some additional
measures are illustrated in Figure 3.12.

3.6 FIELD EXPERIMENTS

3.6.1 General
When sluices and dams have to be built on loosely packed sediments in del¬
taic areas, the scour time factor is very important, since the closure dam usu¬
ally has a temporary function and equilibrium depths will not be reached for
every building stage. Therefore the scour process as a function of time has to
be known, since it may play an important role in construction strategy. In
such cases the application of the model relations for the maximum scour
depth as given in Section 3.3 is recommended.
Within the scope of research activities with respect to scour behind the
 Sills 61

storm surge barrier and compartment dams in the Eastern Scheldt, field ex¬
periments were carried out (de Graauw & Pilarczyk 1981). For this purpose
the sluice in the Brouwers Dam was chosen. This sluice was built to refresh
the brackish water in the Grevelingen lake for environmental reasons. The
experiments were executed in order to study the influence of clay layers on
scour and to verify scour relations obtained from scale models.

3.6.2 Hydraulic and geotechnical conditions

The discharges and flow velocities encountered during two field experiments
were almost identical, whereas the soil characteristics were different. The
discharges, the flow levels and the bed configuration were measured fre¬
quently and some flow velocity and concentration measurements in the cen¬
tre of the sluice were also carried out. During the experiments the sea water
was let into the lake during the flood and was released during the ebb. The
outflow had no influence on the development of the scour hole, because of
the relatively small flow velocities above the scour hole during the ebb,
while the suspended load transported from the sea into the lake was also
negligible.
A 5.4 m high sill was constructed on the lake side of the sluice with two
side constrictions measuring 2.5 m on the left side and 1.5 m on the right
side. The flow depth was about 10 m and the length of the bed protection
from the toe of the sill measured about 50 m. The effective roughness of the
bed protection was estimated to be 0.4 m. The other dimensions of the sluice
are presented in Figure 3.7.
The soil characteristics with respect to Experiment A were measured be¬
forehand. The diameter of the bed material varied with the depth from 0.2 to
0.3 mm. Some thin clay lenses were present, especially in the soil layer be¬
tween 2 and 4 m below the original bed. The thickest clay layer of 0.2 m was
situated at about 3.5 m below the bed. The other clay lenses were mostly in
the range of 10 to 20 mm. The scour hole that developed was refilled with
loosely-packed material. The bed material regarding Experiment B consisted
of fine sand with a particle diameter of about d^^ - 0.26 mm. The particle di¬
ameter for which 90% of the mixture is smaller than measured 0.29 mm.
As a result of tidal variations, both the flow velocity and the sediment trans¬
port vary. To simulate the scour process a characteristic mean velocity was
introduced which was defined as the mean velocity; this would give the
same average sediment transport.

3.6.3 Discussion
Figure 3.13 shows some measured bed profiles of the prototype experiments
at different times. The gradual undermining, including a shear failure, is
62 Scour manual

shown in Figure 3.14, and Figure 3.7 shows the maximum scour depth as a
function of time.

Upstream scour slope

Due to the tidal influence, the flow velocities vary in time. In experiments A
and B the maximum velocities averaged over 140 tides were approximately
1.2 m/s (Hoffmans 1992). From this velocity and using Equation (3.20) it
follows that cot P == 1.8. More details of experimental and computational re¬
sults can be found in Tables 3.4 and 3.5, and in Delft Hydraulics (1979).
With respect to the upstream scour slope, it was observed that the varia¬
tion in the slope over the first few meters was rather large. Due to the flow
velocities and the high turbulence generated in the mixing layer and in the
vortex street, the slope became steeper until the critical slope of about
1V:1.5H was reached. Then a shear failure (Experiment B) occurred, result¬
ing in a milder slope for the upstream scour slope. After this instability the
steepening of the upstream scour slope started again. At the end of the proto¬
type experiments the upstream scour slope measured approximately 1V;2H.

Undermining
In general, the angle of internal friction for sand lies in the range of 30 to
40°, depending on the porosity, the particle diameter and the grain size dis¬
tribution of the mixture. Applying Equation (3.20), small scale shear failures

water surface

experiment A experiment B

X after 140h A after 140h

o after 265h a after 486h

V after 844h ■S’ after 850h

IV 2H

Figure 3.13. Bed pro¬

0 10 20 30 40 60 files of scour holes
distance from end of bed protection (m) (Brouwers Dam).
 Sills 63

O experiment A

X experiment B

shear failure

0 100 200 300 400 500 600 700 800 900

nett scour-time-► (hours)
Figure 3.14. Undermining as function of time (Brouwers Dam).

Table 3.4. Experimental results of field experiments at Brouwers Dam.

Experimental parameters Experiment A Experiment B
Initial flow depth (scour hole) (m) 10.6 9.6
Height of sill (m) 5.4 5.4
Length of bed protection (m) 50 52
Effective roughness of bed protection (m) 0.4 0.4
Averaged discharge (m^/s) 271 270
Maximum discharge (m^/s) 380 380
Particle diameter (mm) ^50 “ 0.25 d^Q = 0.26
^90=0.29 dgo=0.29
Upstream scour slope at end of test 1V:2.2H 1V:1.5H
Undermining just before shear failure slide (m) 2.9
Undermining at end of test (m) 2.3 5.0
Condition subsoil Clay/sand Sand
Characteristic time (hours) (extrapolated) 2000 800
 64 Scour manual

Table 3.5. Computed values for the field experiments at Brouwers Dam.
Computational parameters Experiment A Experiment B
Characteristic discharge (m^/s) 9.89 9.89
Characteristic mean velocity (m/s) 0.93 1.03
Critical mean velocity (m/s) 0.41 0.41
Relative turbulence intensity (-) (Eq. 3.6) 0.28 0.29
Turbulence coefficient (-) (Eq. 3.8) 2.91 2.92
Roughness function (-) (Eq. 3.8) 1.13 1.12
Upstream scour slope (Eq. 3.20) 1 V:1.8H 1V:1.8H
Characteristic time (hours) (Eq. 3.12) 2400 1200
Equilibrium scour depth (m) (Eq. 3.3) 33.6 35.5

can be expected when the flow velocities in the prototype situation are larger
than 1.3 m/s for (<])' = 30°) or 1.9 m/s for ((|)' = 40°). During the experiments
flow velocities were measured varying from 1.5 to 2.0 m/s. Since both small
scale shear failures and a shear failure that occurred after approximately 450
hours (Experiment B) were observed, Equation (3.20) seems to be feasible
for use in practical engineering. When the subsoil consists of clay and sand
layers the results obtained from Equation (3.20) must be interpreted care¬
fully, because the influence of the cohesion of the subsoil has not been taken
into account.

Time scale
Both experiments were ended after about 800 hours net scour time. At that
time the maximum scour depth in both experiments was approximately equal
to the initial flow depth. The characteristic time at which the maximum
scour depth equals the initial flow depth was extrapolated from the meas¬
urements when no shear failures would occur. These measured times were
compared with the computed ones obtained with Equation (3.12). The dis¬
crepancy ratio r (ratio between the calculated and measured time scale) for
both experiments lies in the range of 2/3 < r < 1.5 (Tables 3.4 and 3.5). Note
that the degree of turbulence and the flow velocities determine the scour
time scale to a large extent. Hoffmans (1992) showed that r is larger than 2
(or smaller than 0.5) if the error in the discharge is greater than 15%.

Equilibrium scour depth

Besides the upstream scour slope, the maximum scour depth in the equilib¬
rium phase also determines the optimal length of the bed protection down¬
stream of hydraulic structures. The calculated equilibrium scour depth is
about three times the initial flow depth. The time required to reach this depth
in the prototype situation is extremely long. After about 25 years nett scour
 Sills 65

time, the equilibrium scour depth will almost be achieved, as can be seen in
Figure 3.7.

Closing remarks
The objective of these field experiments was to verify the scour relations
obtained from scale models (Breusers 1966, 1967; van der Meulen & Vinje
1975; de Graauw & Pilarczyk 1981). The computed results compare fa¬
vourably with the measured ones, so the hydraulic and morphological rela¬
tions seem to be applicable for practical engineering. However, the scour
process is not only influenced by turbulence parameters but also by the
geotechnical ones. Currently, the instabilities of the subsoil, especially the
phenomena of flow slides are being researched extensively. To produce a
safe and reliable design the total reliability of all modes of failure should be
approximated at least to a conceptual level. A fault tree is a useful tool for
integrating the various mechanisms into a single approach.

3.7 EXAMPLE

A storm surge barrier is constructed in a wide estuary. The height of the

broad-crested sill is 10 m and the distance between the piers is 50 m. During
operation, when the gates between the piers are raised to the upper state, the
flow far downstream the sill is subcritical. If the water level at sea exceeds a
critical value which is, on average, three times a year, the barrier is closed.
The flow depth in the estuary is about 25 m and the mean tidal range is
2.5 m, with maximum velocities of 1.5 m/s. During spring tide, flow veloci¬
ties of 2.0 m/s are observed. The tidal curve is simplified to a sine function.
The length of the bed protection is L = 500 m and the effective roughness of
the bed protection is estimated to be = 0.5 m. The non-cohesive bed mate¬
rial characteristics are = 180 pm, = 200 pm and d^^ = 300 pm. The
angle of internal friction is (j)' = 35°. The average water temperature is 15°C,
V = 1.1 X 10”® m^/s. Other data are = 2650 kg/m^, p = 1025 kg/m^.
a) What is the characteristic mean velocity which is defined as the flow
velocity that will result in the same average transport?
b) What is the relative turbulence intensity at the end of the bed protec¬
tion?
c) What is a according to the methods of Jorissen & Vrijling (1989) and
Hoffmans (1993)?
d) What is a if the length of the bed protection is L = 125 m?
e) What is the characteristic time at which the maximum scour depth
equals the initial flow depth?
f) What is the slope angle of the scour hole in the equilibrium phase? Will