Hydraulic gates and valves

In Free Surface Flow and Submerged Outlets

Jack Lewin (Hon) DEng, CEng, FICE, FIMechE,

FCIWEM

Second Edition
Published by Thomas Telford Publishing, Thomas Telford Ltd, 1 Heron Quay, London
E14 4JD. URL: http://www.thomastelford.com

Distributors for Thomas Telford books are

USA: ASCE Press, 1801 Alexander Bell Drive, Reston, VA 20191-4400, USA
Japan: Maruzen Co. Ltd, Book Department, 3^10 Nihonbashi 2-chome, Chuo-ku, Tokyo 103
Australia: DA Books and Journals, 648 Whitehorse Road, Mitcham 3132, Victoria

First published 2001

Also available from Thomas Telford Books

Bulk water pipelines, Tim Burstall. ISBN 07277 2609 9
Pipeflowsoftwareversion2 CD ROM, W Chojnacki, S Kulkarni and J Tuach. ISBN 07277 2684 6
Tables for the calculation of friction in internal flows, D I H Barr. ISBN 07277 2046 5
Tables for the hydraulic design of pipes, sewers and channels, 7th edition, Vol 1, D I H Barr.
ISBN 07277 2637 4
Tables for the hydraulic design of pipes, sewers and channels, 7th edition, Vol 2, D I H Barr.
ISBN 07277 2638 2

A catalogue record for this book is available from the British Library

ISBN: 0 7277 2990 X

Jack Lewin and Thomas Telford Limited, 2001

All rights, including translation, reserved. Except as permitted by the Copyright, Designs
and Patents Act 1988, no part of this publication may be reproduced, stored in a retrieval
system or transmitted in any form or by any means, electronic, mechanical, photocopying
or otherwise, without the prior written permission of the Publishing Director, Thomas
Telford Publishing, Thomas Telford Ltd, 1 Heron Quay, London E14 4JD.

This book is published on the understanding that the author is solely responsible for the
statements made and opinions expressed in it and that its publication does not necessarily
imply that such statements and/or opinions are or reflect the views or opinions of the
publishers. While every effort has been made to ensure that the statements made and the
opinions expressed in this publication provide a safe and accurate guide, no liability or
responsibility can be accepted in this respect by the author or publishers.

Typeset by MHL Typesetting Ltd, Coventry

Printed and bound in Great Britain by MPG Books, Bodmin, Cornwall
Contents

1. Introduction 1

2. Types of gates 3
Gates in free surface flow 4
Radial gates 4
Radial automatic gates 9
Other water operated gates 15
Vertical-lift gates 17
Rolling-weir gates 22
Flap gates 23
Fuse gates 33
Barrier and barrage gates 35
Drum and sector gates 48
Bear-trap gates 50
Gates in submerged outlets 51
Intake gates 51
Control and guard gates 53
Emergency closure gates, maintenance gates and stoplogs 59
Summary of types of gate 63

3. Valves 71
Sluice valves 71
Butterfly valves 71
Cavitation in valves 75
Hollow-cone valves and hoods 78
Hollow-jet valves 84
Needle valves 85
Pressure-reducing valves 85
Sphere valves 87
Matching terminal discharge valves and guard valves 89
Summary of types of valve 90
 Contents 4. Trashracks, screens and debris 93
Trashracks and screens in submerged intakes 93
Trashracks and screens in culverts and river courses 96
Screen instrumentation 96
Screen raking 97
Debris 97

5. Structural considerations 101

Design criteria 101
Structural design of radial gates 103
Structural design of vertical-lift gates 107
Stiffening members of skin plates 107
Composite construction 109

6. Operating machinery 115

Electromechanical drives 116
Oil hydraulic operation of gates in free surface flow 121
Hoist speed 126

7. Detail design aspects 127

Seals 127
Guide and load rollers 137
Trunnion assembly 140
Trunnion bearing failure and lubrication 142
Trunnion mounting for radial gates 144
Limit switches 145
Ropes 147
Chains 148

8. Embedded parts 151

9. Hydraulic considerations pertaining to gates 157

Flow under and over gates 158
Hydraulic downpull forces 166
Limited ponded-up water 171
Three-dimensional flow entry into sluiceways 171
Reflux downstream at a pier and flow oscillation 171
Hysteresis effect of gate discharge during hoisting/lowering 172
Hydraulic considerations pertaining to gates in conduit 172

10. Gate vibration 185

Types of gate vibration 185
The vibrating system 186
Excitation frequencies 188
Added mass 191

iv
 Preliminary check on gate vibration 192 Contents
Vibration due to seal leakage 194
Flow attachment, shifting of the point of attachment and
turbulent flow 196
Hydraulic downpull forces and flow reattachment at
the gate lip 199
Unstable flow through small openings 200
Flow over and under the leaves of a gate 201
Vibration of overflow gates due to inadequate venting 201
Vibration due to a free shear layer 201
Two-phase flow 202
Slack in gate components 202

11. Control systems and operation 207

Control objectives 207
Operating rules and systems 209
Telemetry 216
Factors in the choice of automatic gate control systems 217
Fall-back system and standby facilities 218
Instrumentation 223

12. Hazard and reliability of hydraulic gates and valves 227

Events at spillway gate installations 227
Incidents and failures of bottom outlets 232
Fault frequency by gate type 234
Risk assessment of gated hydraulic structures 235

13. Ice Formation 247

Ice 247

14. Earthquake effects on gates 251

Spillway gate installations 252
Methods of analysis 254
Allowable stresses 255
Vertical-lift spillway gate installations 255
Bottom outlet tunnel gates 256
Operating machinery under earthquake conditions 257
Systems analysis 257
Control buildings 261
Sample event tree for a seismic event on a dam 261

15. Materials and protection 265

Materials 265
Steel corrosion and painting 266
Cathodic protection 268

v
 Contents 16. Model studies 271
Froude scale models 271
Two-phase flow problems 272
Two- and three-dimensional models to Froude scale 272
Models for investigating vibration problems 273

17. Environmental considerations 277

Gated river control structures 277
Barriers 281
Barrages 281

18. Maintenance and operation of gate installations 285

Appendix: Calculation of hydrostatic load on radial gates 289

Index 295

vi
Preface

The success of the first edition of this book has encouraged me to persist in
writing a second edition. Some chapters of the book are in their original form,
with minimal additions. Other chapters have been substantially enlarged. New
chapters deal with hazard and reliability, earthquake effects and environmental
impact of gate installations. These subjects have become more prominent since I
wrote the first edition, and I have attempted to provide an introduction to the
particular problems at gated structures.
As in all fields of technology, the position is never static and a constant
stream of technical papers adds to our knowledge. It is impossible to deal
comprehensively with all the material which has become available; I therefore
had to be selective about what to include. Even without absorbing new
material, it is difficult for a practising engineer to cover the field of hydraulic
gates and valves exhaustively as it incorporates aspects of many different
engineering disciplines. As a compensation, I like to think that my extensive
hands-on experience puts me in a better position to bridge the gap between
theory and practice, which is essential for the success of any engineering
endeavour.
I am indebted to all those who made valuable contributions to the book
based on their own specialist knowledge, particularly Mr Derek Wilden for
the first edition, and Drs Geoff Ballard and Mike Gardner for the second
edition. I owe much to Dr Paul Kolkman, formerly of Delft Hydraulics, for
his generous permission to use material from his extensive papers, fundamental
to an understanding of gate vibration. Thanks are due to Sue Lamb for tracing
the diagrams and to Mr Mark Noble for supplying material.
I would also like to thank clients for whom I have carried out projects for
permission to use information acquired in the course of this work.

Jack Lewin, 2001

vii
Acknowledgements

My thanks are due to my wife Barbara and daughter Jaqueline, who struggled to
translate the manuscript into a readable form.
Acknowledgement and thanks for permission to reproduce material from the
author's papers are made to CIWEM for Figs 2.8, 2.16, 2.56, 2.57, 2.58, 2.60,
2.61, 2.64, 3.16, 6.2, 7.13, 10.10 and 10.14.

Acknowledgement and thanks for permission to reproduce material are

similarly due to:

Bridon Ropes Ltd for Table 7.2.

British Hydromechanics Research Group, Cranfield for Figs 3.4^3.7.
Computational Mechanics Centre, Southampton for Figs 2.11^2.15.
The Consorzio Venezia Nuova for Figs 2.41^2.44.
Delft Hydraulics for Fig. 16.2.
The Electricity Corporation of New Zealand, Waikato Hydro Group and Water
Power and Dam Construction for Figs 11.8^11.12.
Goulburn^Murray Rural Water Authority and the Murray^Darling Basin
Commission, Australia for Figs 5.5 and 7.17.
The Environment Agency for the photograph in Fig. 2.37 and Figs 2.45^2.48,
5.4 and 7.4.
Hydraulic Research, Wallingford for Fig. 9.18.
Hydroplus for Fig. 2.36.
The Institution of Civil Engineers for Figs 2.38^2.40.
The Institution of Mechanical Engineers for Fig. 2.37.
Ishikawajima Harima Heavy Industries Co. Ltd for Figs 2.49, 2.59 and 2.63.
Hans Knz GmbH for Fig. 4.3.
Mannesmann Rexroth for Figs 6.7 and 6.8.
Renold plc for Fig. 7.19(c).
Scottish and Southern Energy plc for Fig. 2.23.
SKF (U.K.) Ltd for Figs 7.14 and 7.16.
Voest Alpine GmbH for the photograph in Fig. 3.18.
J.M. Voith GmbH for Figs 3.1^3.3, 3.8, 3.16 and 3.17.
 1
Introduction

The control of rivers, canals and reservoirs requires weirs or gated structures.
From considerations of reliability and maintenance, the fixed weir is the
preferred control structure. Similarly, the fixed crest free overflow spillway is
the most advantageous arrangement for reservoirs. Wherever weirs or fixed
crest overflow spillways cannot be accommodated, or where the backwater
stages of a flood or variable river levels are unacceptable, a device which
provides a movable crest or a submerged variable discharge opening has to be
provided. Gates and valves are therefore an essential and critical part of many
flood control schemes, of reservoir management and the control of water in
river courses.
Many types of gate are in successful operation. However, only a few of these
may be suitable or cost-effective for a specific situation. The challenge is to select
and design a gate of the most appropriate type and size which will meet the
hydraulic, operational, site specific and economic requirements.
Gates and valves control the flow of water, so the hydraulic conditions are
basic to the success of the installation. This comprises not only the flow under or
over a gate but also the upstream and downstream hydraulics.
Since gates are designed for extreme events, personal experience of their
performance under these conditions is by definition limited. Some gate
installations have met with serious difficulties in service. Subsequent research
and published papers, sometimes presented at a specialist congress, are not
always disseminated widely enough to prevent repeated use of flawed design
features. This book attempts to combine available knowledge with practical
experience of gates and valves in civil engineering structures.
The first task is to provide a guide to selecting the right gate or valve.
Information on details of gate design is provided to help in an assessment of
the suitability of a gate for its task, and the chapter on valves is intended to assist
in identifying the correct type of valve for its duty. Ancillary equipment
required in control structures, such as screens, stoplogs and handling
equipment, is covered in separate chapters.
Later on, space is given to the consideration of hazard and reliability of gates
and valves. This is part of dam safety, but it is also of paramount importance at
tidal defence barriers. Gated weirs and barrages control flood flow and their
failure can result in inundation and serious damage.

1
Hydraulic gates and Note on units
valves
The symbols used in the equations of the text, with few exceptions, have no
designated units. Wherever this is the case consistent SI units can be used.

Note on terminology
The term `barrage' is used in the book for structures which impound water.

The term `barrier' defines a structure which is brought into operation to protect
a river against an extreme event, such as storm surge.

2
 2
Types of gates

In this chapter, gates are divided into two groups: in free surface flow and in
conduits. While similar types of gate are found in both groups, the design of
gates in submerged outlets, especially at high heads, is more demanding and
requires special consideration. The main applications, advantages and
disadvantages of the various types of gate are summarised in the table at the
end of this chapter.
While there are many types of gate, a limited number predominate because of
their advantages. In open channels, at spillways and barrages, radial gates are the
first choice. Except for very large span gates in navigable rivers, new vertical-lift
gate installations are infrequent apart from situations involving rehabilitation
of old barrages, such as the Sukkur and Kotri Barrages on the River Indus in
Pakistan where old gates have been replaced by similar vertical-lift gates of
modern construction.
The bottom-hinged flap gate or tilting gate is sometimes preferred in river
courses because it is considered less visually obtrusive than a radial gate. Also its
ability to discharge debris over the gate may be important (a radial gate requires a
flap section to carry out the same function). In tidal barrages a tilting gate can
completely prevent ingress of saline water to the upstream pond or reach, and is
often chosen on this account. Undershoot gates can, under drowned flow con-
ditions, permit a lens of saline water to penetrate upstream against the flow of water.
In conduits, vertical-lift gates are used more frequently than radial gates,
mainly due to greater flexibility in installation (radial gates require a large
chamber to retract). The gate slots of vertical-lift gates create hydraulic
problems at high velocity flow, whereas radial gates do not require slots. In spite
of this advantage, vertical-lift gates are often selected because of the installation
problems associated with radial gates.
Detail aspects of gate design are dealt with in separate chapters because they
apply to a number of different gates. The exception is radial automatic gates
where many design considerations are specific to this type of gate.
A number of gates, such as tidal barrages or storm surge barriers, have been
developed for special conditions. So far they have not found more general
application.
Gate types which were once common, such as the rolling-weir gate, drum
and sector gates, Stoney-roller gates and others not dealt with in this chapter
have been largely superseded because of their complexity, cost and associated
civil engineering construction costs.

3
Hydraulic gates and Gates in free surface flow
valves
Radial gates
Radial gates are the most frequently used movable water control structures.
They consist of a skin plate formed into a segment with a radius about the pivot.
The skin plate is stiffened by vertical and horizontal members which act
compositely with the skin plate. The skin plate assembly is supported by two
or more radial arms at each side which converge downstream at the trunnions,
where they are anchored to the piers and transfer the entire thrust of the water
load to the civil engineering structure. The resultant of the water load passes
through the trunnion pins and there is no unbalanced moment to be overcome
when hoisting the gate. The hoisting load consists of the weight of the gate, the
friction force between the side seals and the seal contact plates embedded in the
piers, and the moment of the frictional resistance at the trunnions.
Radial gates are under most conditions the simplest, most reliable and least
expensive type of gate for the passage of large floods. Their advantages are:
(a) Absence of gate slots, because side seals are used. This benefits pier
structural design and hydraulic flow. Pier slots can produce cavitation and
at low flows collect silt. It also avoids the build-up of ice.
(b) Gate thrust is transmitted to two bearings only.
(c) Less hoisting capacity is required than for a vertical-lift gate.
(d) It is mechanically simpler and mechanical equipment usually costs less.
(e) The geometry of the skin plate provides favourable hydraulic discharge
characteristics.
(f) Location of the bearings above the level of discharge or flood flow protects
them from damage by debris, simplifies corrosion protection and permits
some degree of inspection with the gate in service.
(g) The superstructure required for the gate lifting gear is generally much
lower, and in cases where a road spans the sluiceway usually no additional
structure is required.
(h) They are stiffer structurally.
(i) They have a better appearance.
(j) There is no possibility of trash jamming in the wheels.
The disadvantages of radial gates are:
(a) The flume walls must extend downstream at a sufficient height to provide
attachment for the gate trunnions.
(b) The gate water load is taken by the piers as concentrated loads at the gate
anchorages. Because of this, integrity of the anchorages and distribution of
the load into the piers require special consideration. In gates of 150^200 m2
aspect area and larger, this has led to the use of prestressed steel and concrete
anchorages.
(c) Increased fabrication complexity.
Radial gates will not pass floating material until they are at least 75% open.
This can be overcome by adding a flap or overflow section to the top of the gate
(Fig. 2.1(b)).

4
 Types of gates

Figure 2.1. Types of radial

gate

Overflow flaps are usually operated by an independent motor and hoist gear
or a separate hydraulic cylinder. The overflow discharge section is curtailed on
this type of gate for the nappe to clear the gate arms. Tilting crests may be
required when ice floes have to be discharged in spring or where the river carries
an abnormal amount of debris during part of the year.
The shape of the tilting crest has to be checked so that flow separation cannot
take place in any one position when the crest section is lowered. If the crest
section is curtailed the nappe will be vented and problems due to nappe collapse
will not occur.
Radial gates have been designed to be submergible to provide both overflow
and discharge under the gate (Fig. 2.1(c)).

Hydraulic forces acting on a radial gate

In a closed radial gate the resultant hydrostatic forces act through the trunnions,
see Fig. 2.2. This applies to upstream and downstream forces.
Provided the skin plate has been formed in a true radius with the origin at the
pivot point there are no unbalanced hydrostatic forces about the pivot. When
the gate is lifted the forces are the mass of the gate with the centre of gravity
fairly close to the weir plate assembly, the frictional resistance of the side seals
and the frictional resistance of pivot bearings. During discharge under the gate a
frictional resistance at the skin plate is assumed to act tangentially (see Fig. 2.3).
The seal friction will be significantly higher on starting compared with the

Figure 2.2. (left) Hydro-

static forces acting on a radial
gate
Figure 2.3. Lifting and
lowering forces acting on a radial
gate

5
 Hydraulic gates and
valves

Figure 2.4. Distribution of

pressure head for radial gate
under free discharge conditions

running friction, and this also applies to bushed bearing although not to roller
bearing pivots.
Figure 2.4 shows the distribution of pressure head for a radial gate under free
discharge conditions. If the pressure curves are available from flow-net analysis
or actual measurement on a model or prototype, the required components of the
forces may be obtained by graphical integration. An approximation assumes
hydrostatic distribution of pressure over the gate as indicated by the broken
lines in Fig. 2.5. The corresponding force components are equal to the enclosed
areas multiplied by the specific weight of water and the gate width.
This will result in an overestimate of the hydraulic forces acting on the gate.
Rouse1 has given a method of computing the hydraulic forces by taking into
account the discharge characteristics of the gate. This leads to a closer
approximation.
An Appendix at the end of this book sets out the calculations of the hydraulic
forces acting on a radial gate for water pressure on a closed gate, for a gate in the

Figure 2.5. Approximation

of pressure head for a radial gate
under free discharge conditions,
by assuming hydrostatic
distribution of pressure over the
gate

6
 Types of gates

Figure 2.6. Conventional

arrangement of arms of a radial
gate

open position with drowned discharge and for a gate under overflow
conditions.

Constructional features of gate arms

Gate arms are usually offset (Fig. 2.6). This reduces the bending moment on the
horizontal girders connecting the arms and, where necessary, permits the
trunnions to be recessed into the pier. In gates operating under drowned
discharge conditions in rivers it also prevents the trapping of debris between
the gate arms and the piers, and ice bridging between the gate arms and the sluice
walls under severe winter conditions.
The lowest gate arm is usually positioned as low as possible for structural
reasons (Fig. 2.7).
Under drowned discharge conditions severe turbulence is set up in the
stilling basin and an unsteady roller occurs.2 If the roller acts on the submerged
gate arms or on other structural members vibration is likely to occur.3 In this
case hydraulic considerations should override structural priorities so that
members are disposed in the most efficient manner (Fig. 2.8).
The hydrodynamic effects on the gate, such as wave action, are taken into
account in calculating the load on a gate. Hydrodynamic downpull forces (see
Chapter 9) have little structural effect on radial gates in open channels but must
be taken into consideration when calculating hoisting forces. They become

Figure 2.7. Positioning of

arms of a radial gate

7
 Hydraulic gates and
valves

Figure 2.8. Effect of reverse

roller on the lowest arm of a
radial gate under drowned
discharge conditions

more significant when radial gates are used as culvert valves under high head
conditions (see Fig. 2.9).
Two-dimensional analyses of the strength of the weir plate assembly ignore
the additional strength and rigidity due to its curvature. To take this into
account, a three-dimensional analysis, such as finite elements, is required. This
is used mainly when gate vibration or the dynamic response to an earthquake is
investigated.
There are two main methods of constructing the skin plate assembly. The
skin plate is either stiffened vertically, as shown in Fig. 2.10(a), or horizontally
as shown in Fig. 2.10(b). The advantages and disadvantages of each method are
discussed in Chapter 5. Both methods of reinforcing the skin plate use main
structural members to tie the gate arms. Together with the gate arms they form
a portal which transmits the load to the trunnions. All members which are
welded to the skin plate are assumed to act compositely with the plate.

Figure 2.9. Radial gate used

as a culvert valve

8
 Types of gates

Figure 2.10. Main methods of

reinforcing and stiffening the
skin plate of a radial gate

Radial gates can be operated by electric motor driven hoists or by hydraulic

rams. Some arrangements of hoists and of hydraulic rams are described and
illustrated in Chapter 6.

Radial automatic gates

Radial automatic gates have been built and operated during the last 60 years.4 They
require no outside source of power, are simple and reliable. Provided their design
is hydraulically correct, they will operate consistently with minimum attention
and require very little maintenance. In general, radial automatic gates can be
installed where upstream level control is required under variable inflow con-
ditions and where the downstream stage does not rise disproportionately quickly.5
They can also be used to control the downstream level in irrigation canals.

Water level control

The gates are actuated by changes in water level and either upstream or
downstream water level can be controlled. The former is usual in rivers and
the latter in irrigation channels. Figures 2.11 and 2.12 illustrate the most
frequent arrangement of upstream water level control. The radial arms which
support the skin plate assembly extend downstream of the trunnions to carry
counterweights which balance the gate. Operation is by displacers in each side
pier. The displacers are attached to the radial arms by pivots, so that the forces
exerted on the gate, due to changes in displacer submergence, cause the gate to
open and close. The displacers have to overcome side-seal and pivot friction
which act in both directions as shown in Fig. 2.3.
The gate control system comprises an intake upstream of the gate protected
by a screen. The intake is controlled by a sluice valve and discharges into the
weir chamber. The discharge over the weir flows into the displacer chambers
which are interconnected by a pipe passing under the sluiceway. An outlet from
the displacer chamber, also controlled by a sluice valve, returns the flow
downstream of the gate. This is shown in Fig. 2.12.
An increase in the upstream level causes flow through the intake pipe into the
weir chamber and discharge over the weir. In turn, this causes a rise in level in

9
 Hydraulic gates and
valves

Figure 2.11. Radial

automatic gate for upstream
water level control

the displacer chambers and flow to the river downstream of the gate. Due to
increased buoyancy of the displacers, the gate rises and discharge occurs under
the gate. Rise of the gate may cause a slight drop in upstream water level, which
will result in reduced flow over the weir and hence a lowering of the level in the
displacer chambers. The gate may close slightly as a consequence until balanced
conditions are achieved. The stability of the system is provided by head loss in
the inlet system and side-seal friction.

The inlet system

The inlet is positioned some distance upstream of the gate. It should be
placed facing downstream and at two-thirds of the retention level to

Figure 2.12. Radial

automatic gate control system

10
 Types of gates

Figure 2.13. Inlet system for

the control of a radial automatic
gate

minimise obstruction of the inlet due to flotsam. The pipe is usually made
large to enable rodding out. The head loss, which should be 50 to 65 mm, is
provided by the setting of the inlet valve. Figure 2.13 shows a typical
arrangement.

The control weir

The control weir acts as an amplifier and takes the form of an adjustable sharp
edged weir set 50^65 mm below the retention level. A convenient way to
achieve a long weir is shown in Fig. 2.13. A stub wall at the end of the weir
permits aeration of the nappe.

Displacers and displacer chambers

The flow from the weir is baffled to provide quiescent conditions in the
displacer chambers. Movement of the gate lip relative to that of the displacers
is amplified by the ratio of the gate radius to the radius of suspension of the
displacers. Similarly the frictional resistance of the side seals referred to the
displacer buoyancy is amplified by the same ratio. Since the direction of the
friction force changes on opening and on closing of the gate, the change in water
level within the displacer chamber required to reverse the movement of the gate
will be double.
Upstream level retention can be achieved within close limits. It can be as low
as 12 mm, although under high flow conditions 50^80 mm rise will be required
for the gate to discharge the increased flow.
Under flood conditions when the upstream level cannot be retained (a
condition which is also associated with a substantial rise in downstream level),

11
 Hydraulic gates and the inlet weir is drowned and the gate operates according to the ratio R1/R2 as
valves shown in Fig. 2.12. The gate will therefore lift more rapidly than the rise in flood
water and will cut out of the water to permit unobstructed flow under these
conditions.
If the initial difference between upstream and downstream level is 2 m or less,
the stage^discharge relationship must be known when the design is carried out,
so that the gate control system is arranged to avoid the downstream level taking
over control of the gate during some stage of the rising flood. This can also be
important during a falling flood when the downstream level falls more slowly
than the upstream level, resulting in a condition where the downstream water
level keeps the gate open and causes loss of retention level.
The displacers are designed so that their specific gravity is 1.05. Construction
takes the form of a stiffened box with watertight access covers for loading weights.
The total assembly forward of the pivot is out of balance to the extent that the
displacers are half immersed when the gate is in a steady condition. Equal forces
are then available for opening and closing the gate. With this arrangement the
mass of one displacer must be at least equal to the friction force of one side seal
multiplied by R1/R2 plus the pivot friction. A gate with displacers sized on this
basis would have no margin and would provide very coarse level control. With
displacers designed three times this size they will give good service.
The wide piers required to accommodate displacer chambers can be a
disadvantage in a sluice installation.
If self-aligning roller bearings are used for the pivots the effort at the
displacers to overcome friction is negligible. For bronze bushed bearings it will
amount to 1^2.5 kN at the displacer.
It is normal practice to use displacers, but there are some gates where floats
have been used. Figure 2.14 shows a design where floats are held stable by the
link `a' so that pivots `b' and `c' and the link form a parallel motion. In a 6 m wide
gate, using floats would result in an overall saving of 2.8 tonne.
To ensure consistent operation throughout the range of travel of the gate, the
centroid of the counterweight, the centre of the pivot bearing, the centre line of

Figure 2.14. Radial gate using

floats (11.5 m radius, 15 m
width)

12
the displacer suspension pivot and the centroid of the skin plate assembly should Types of gates
be in one line.4 Departure from this is possible, provided the gate operation is
calculated separately for different openings.

Trunnions
Most gate manufacturers use phosphor bronze bushed bearings. Self-
lubricating bronze bearings have a starting coefficient of friction of 0.1^0.12,
which reduces to 0.07 during running. The gates in Figs 2.11 and 2.15
incorporate self-aligning roller bearings. Their coefficient of friction is 0.0018.
Using self-aligning roller bearings eliminates uneven bearing pressure due to
deflection of the pivot pin in a bushed bearing. Self-lubricating bearings,
mentioned in Chapter 7 and illustrated in Fig. 7.16, significantly reduce
maintenance. They are available as plain bushes or self-aligning bearings.

Counterbalance
This can take the form of a reinforced concrete beam with pockets cast into the
upper section so that final balancing of the gate can be carried out. Where

Figure 2.15. Downstream

level control gate

13
Hydraulic gates and environmental conditions require a less obtrusive appearance, as at Pulteney
valves Weir in the City of Bath, the kentledge can be of cast iron sections bolted to a
structural beam.

Downstream water level control gates

Gates for downstream water level control use a float chamber downstream of
the trunnions (Fig. 2.15). These are only suitable for small gates, because the
discharge under the gate causes turbulent conditions. Larger gates have to be
severely damped, or displacer chambers in the piers must be constructed using
an intake positioned downstream of any turbulence. It is usual to baffle the
intake. Gates of this type are not subject to size limitations.

Causes of instability and malfunction of radial automatic gates

Gate instability can be due to:
insufficient head loss in the inlet system
insufficient side-seal friction
limited ponded up water in the upstream reach
reflux of flow downstream of the gates around piers dividing two sluiceways.

Malfunction of gates can be caused by:

blocked inlet system or screen
lodgement of an obstruction at the sill beam (this also applies to motorised
gates)
flooded displacer
failure to flush silt from the weir and the displacer chambers.

Dividing piers between sluiceways

If the length of the dividing piers between the sluiceways is short, the discharge
from one sluiceway may, under drowned discharge conditions, reflux around the
pier and set up oscillating waves acting on the transverse stiffeners of the gate in
the adjoining sluiceway. This can set up harmonic motion of the second gate
which, in turn, can affect the first gate which will respond in a similar manner,
but out of phase. Such motion will not amplify more than 150^300 mm. It can
be a nuisance to river navigation. In an installation in the River Lee (a tributary
of the Thames) in Essex, the dividing pier had to be extended in order to eliminate
this effect, so its length downstream of the gates was increased from 15 m to 23 m.

Computer program for radial gates

Because unstable operating conditions can arise in radial automatic gates, and
because design and testing for instability can be carried out only on a trial and
error basis varying parameters in turn, a computer program forms a useful
design tool so that changes in variables can be rapidly examined.
The program is used to determine the relationship between upstream and
downstream water levels, gate opening and the discharge under the gate.
Relative values for the rising and falling river or reservoir stages and for the
condition when the gate is clear of the water have to be computed.

14
 Such a program was run for the gate shown in Fig. 2.11 during the Types of gates
preliminary design phase. The results demonstrated that the gate cut out of
the water too quickly because the rapid rise in downstream level took over the
control of the gate. This caused an unstable condition because the upstream
level then dropped below the retention level. The program was rerun,
increasing the water level in the displacer chamber and the outlet from the
displacer, until stable conditions were obtained.

Other water operated gates

In Portugal a number of spillway radial gates have been installed which depend
for their operation on a system of floats and counterweights (see Fig. 2.16).6
These gates were developed so that their operation did not require a power
supply and changes in reservoir level directly caused raising of a gate,
characteristics which are similar to radial automatic gates. The gates operate
unattended and respond rapidly to changes in water level, a requirement at
spillway gates at Portuguese dams due to rapid rises to the peak of flood
hydrographs.

Figure 2.16. Float and

counterweight system for
automatic operation of a radial
gate (after Quintela et al.6)

15
 Hydraulic gates and Control of the gate response to inflow into the float chamber is by a penstock
valves or a weir at the inlet to the pipe system and a valve at the outlet from the float
chamber. The tripod above the float chamber is for manual raising of the gate.
Gates wider than 10 m have a float and counterweight at both sides. Quintela
et al.6 list 37 gates of this type in Portugal. The gates appear to have been reliable
in service except for two gates where the hoist ropes failed due to corrosion. The
risk of malfunction due to obstruction of the inlet to the control pipe and
jamming of the counterweight or of the float is mentioned in the paper. The
possibility of the hoist or the control ropes leaving the diverter pulleys must
be a further risk, also silt deposition in the chambers when the flood flow carries
silt in suspension.
In South Africa, other automatic water operated gates have been developed
by Flowgate Projects of Randburg, South Africa. Their operation depends on
buoyancy tanks which are integral with the skin plate.
Operation of the automatic crest gate is shown in Fig. 2.17. It requires a
spillway weir with a vertical face on the upstream side to permit the gate to recess
under flood conditions. Level control achieved by this type of gate is coarse
compared with that of radial automatic gates. The inlet weir has to be protected
from wave action and debris. The face seal at the crest of the weir requires

Figure 2.17. Automatic crest

gate (after Townshend7)

16
 Types of gates

Figure 2.18. Automatic scour

gate (after Townshend7)

accurate fabrication of the skin plate. Under overflow conditions, the nappe has
to be aerated (see also section on flap gates later in this chapter).
The automatic scour gate shown in Fig. 2.18 can provide scouring and
maintain upstream water level.

Vertical-lift gates
The advantages of vertical-lift gates are:
(a) short length of flume walls
(b) distribution of the gate water load.
The disadvantages are:
(a) the requirement for gate slots
(b) possibility of trash jamming in the wheels
(c) overhead structure
(d) wheels have to rotate underwater
(e) greater cost because of the need for an overhead structure.
The majority of vertical-lift gates are counterbalanced to reduce the hoisting
load. To prevent the counterbalance from entering the water when the gate is

17
Hydraulic gates and lifted, the counterbalance is reeved 2:1 so that it travels for only half the distance.
valves This results in an additional load on the superstructure of the order of 2.7 times
the mass of the gate, and requires a substantial support structure, adding to the
cost of gate installation. Most gates of this type that are used in open channels are
of the fixed roller type (Fig. 2.19(a)).
In gates in open channels, rollers are usually spaced out to take an equal load
of the hydrostatic forces acting on the gate. Roller alignment is critical, as
uneven contact of a roller can overload adjoining rollers. One method of
adjusting rollers to ensure equal contact on the roller path is shown in Fig. 7.13.
Downstream sealing of a gate is preferred because the water load compresses
the seals. Upstream sealing is required where a gate is located upstream in a shaft,
and access for inspection of the tunnel and the gate is via the shaft.
Fig. 2.19(b) is a diagram of a Stoney-roller gate. The roller trains, one on either
side within the gate slots, move at half the speed of the gate and are reeved 2:1 on
their suspension. In practice, the load distribution from the gate via the rollers to
the track is not even, and the permissible contact (Hertz) pressures for Stoney
rollers is half of the pressure for load roller wheels (see also Chapter 7). The rolling
load is transmitted from the rolling face on the gate directly to the track face and
Stoney-roller axles should theoretically be subject to only nominal rotational
friction. In fact this is not the case, since the rollers deform elastically imposing a
load on the axles or bushes, if they are fitted. In addition, inaccuracies in alignment
of the rollers can impose considerable additional loads. Breakdown of individual
rollers has occurred as a consequence. The slack in the tracking of a Stoney-roller
gate can cause gate vibration, especially under conditions of high velocity flow.
Stoney-roller gates have fallen into disfavour. There are, however, many gate
installations of this type which are being refurbished or will require
replacement.
Vertical-lift gates can also be fitted with overflow sections (see Fig. 2.20),
where limited overflow is required.
Where substantial overflow must be effected a hook-type gate is used (Fig.
2.21). This results in considerable complexity of construction, and high
accuracy of manufacture is required to maintain the seal between the two
sections.
Hook-type gates use either a single hoist so that the upper section is hoisted
first and when it reaches the full extent of travel it moves together with the lower
section, or two hoists. The latter arrangement is required for combined over-
and underflow or when the gate is to be used for underflow without moving the
upper leaf (sometimes required for aesthetic reasons). The hydraulic conditions
caused by combined over- and underflow can induce gate vibration and the
range of operation of such a gate may have to be restricted.
For long span vertical-lift gates where the skin plate structure is backed by
girders and where there is drowned discharge, similar considerations to those
mentioned earlier in this chapter concerning arms for radial gates apply. The
turbulent discharge conditions downstream of the gate, and an unsteady roller
^ if it develops ^ can act on the structural members and cause local or general
gate vibrations. This is discussed further in Chapter 10.
Larger vertical-lift gates are often manufactured in sections with an
articulated joint between them. Figure 2.22 shows how this can be effected.

18
Figure 2.19. Types of vertical-lift gate

19
 Hydraulic gates and
valves

Figure 2.20. Vertical-lift gate

with overflow section

The connecting bolts permit limited back and forth movement between
sections to allow the guide rollers of any one section to centre. The connection
bolts are prestressed to limit deflection under load. Apart from the saving in site
welding that such designs permit, it is often possible to mount four wheels on
any one section and omit adjustment of the guide rollers.
Figure 2.22 also shows the seal between the sections. This has to be jointed to
the side seals. This is more easily effected if the seal is on the downstream side of
the gate. An intermediate short section of a seal may be required if the seals are
not in the same plane.
An articulated vertical-lift gate has to make provision for transfer of shearing
forces from one section to the other so that racking forces do not shear the seal
between sections.
Transverse guidance of vertical-lift gates is provided by separate guide
rollers or slides.

20
 Types of gates

Figure 2.21. Hook-type gate

Free rolling gates

Free rolling gates are vertical-lift gates where the water load on the gate is
resisted by a series of horizontal tubes which are positionally fixed, but free to
rotate. The steel tubes span the opening and rotate on axles held within frames at
their ends. They act as rollers travelling on the roller paths at each side of the
opening.
The rollers are not in contact with each other and the gaps between them are
sealed on the upstream side by brass tubes which are loosely held in position by
stop plates on the end plates. Under water pressure they are moved against the
steel rollers to seal the gaps.
The top and bottom of the gate is sealed by the top and bottom steel rollers.
The uppermost roller makes contact with the lintel and the bottom roller with
the sill beam lining (Fig. 2.23).

21
 Hydraulic gates and
valves

Figure 2.22. Vertical-lift gate

manufactured in sections, detail of
connectors

The gates are used to shut the intake to the turbine and can be operated under
balanced head or, in emergency conditions, at an unbalanced head. Gates
capable of operating at differential heads of up to 30 m are still in use and have
proved to be very reliable. They were manufactured up to the 1970s.

Rolling-weir gates
Figure 2.24 illustrates a rolling-weir gate where `a' is the drum and `b' a lip to
effect flow separation. The gate has a spur gear segment `c' which engages with a
rack `d'. To raise or lower the gate it is lifted by chains `e' which make it roll
upwards on the rack `d'.
Rolling-weir gates were designed for wide sluiceways where their structural
rigidity and high torsional resistance were advantageous. Some gates have been
designed so as to be submergible to clear ice floes in spring. This can impose
difficulties in flow control at low discharge.
This type of gate is complex to manufacture, imposes difficulties in designing
effective side seals and is vulnerable to jamming if debris becomes lodged on the
rack. Few, if any, drum gates have been manufactured during the last 30^40
years for this reason.

22
Figure 2.23. Free rolling gate

Flap gates
Bottom-hinged flap gates
Bottom-hinged flap gates are used in tidal rivers to prevent the ingress of saline
waters or where special environmental considerations apply. They are
sometimes selected for reservoir spillways because they can be made to operate

23
 Hydraulic gates and
valves

Figure 2.24. Principle of

rolling-weir gate

in an emergency on a `fail-safe' basis. For gates operated by hydraulic cylinders,

fail-safe gate lowering can be accomplished by operation of a bypass valve which
diverts the oil from the annulus side of the cylinder to the piston side at a
controlled rate. Since the volume of oil in the annulus side is less than that in
the cylinder side for a given length of stroke, the cylinder may have to be vented
to atmosphere or a degree of cavitation in the cylinder may have to be accepted.
For fish-belly flap gates in river courses or barrages where the downstream
water level can rise above the level of the gate hinges, complete fail-safe opening
is possible only if water is admitted to the fish-belly section, otherwise the gate
will open only to the stage when the fish belly becomes buoyant, that is when the
mass of water above the gate and its submerged mass are equal to the buoyancy
and pressure under the gate. It is not usual to admit water into the fish-belly
section because it is difficult to repaint, and silt and sediment can accumulate
there.
Because tilting gates discharge by overflow, floating debris can be cleared
without having to introduce a separate flap or tilting section as in bottom
discharge gates.

24
 A disadvantage of bottom-hinged flap gates in river courses or tidal barrages Types of gates
is permanent immersion of the hinge bearings and inability to inspect the
bearings and the hinge seal where the downstream water level is always above
the level of the hinges, except by placing stoplogs.
Figure 2.25 illustrates different versions of bottom-hinged flap gates. Figure
2.25(a) shows a flap gate operated by a hydraulic cylinder positioned underneath
the gate. This can present difficulties in servicing the cylinder when the
downstream water level does not fall below the downstream bed level. Figure
2.25(b) shows an arrangement where the operating cylinder is housed in the piers.
A torsion tube transmits the operating force laterally. If the operating shaft passes
into the pier chamber the seal adds to the complexity of this design. Figure 2.25(c)
shows a single operating cylinder positioned at the side of the gate. The structure
of such gates is made torsionally rigid and often takes the form of a `fish-belly'.
Bottom-hinged gates open by tilting in a downstream direction until they lie
flat or form the required crest profile to prevent flow separation. They are
usually sealed along the hinged edge by a flexible flat rubber strip clamped to
the embedded sill member and extended to rest against the upstream face of the
gate skin plate. Because the gap to be sealed varies throughout the movement of
the gate it is possible to extrude the seal under maximum hydrostatic head, and
designs clamping the seal both to the sill and the gate face plate have been
evolved. Side-sealing over the full travel of the gate is essential, since debris in
the water tends to be drawn into any gap. If the gates are made to close against an
abutment in the sluiceway in order to seal by making direct contact with the
upstream face of the recess, any material trapped is compressed or wedged and
there is a danger that the gate will become jammed. Recessed abutments are also
not recommended for hydraulic reasons. It is therefore good practice for side
seals to sweep the sluice walls throughout the total movement of the gates. This
requires machined side plates, of either cast or fabricated construction,
embedded flush with the civil structure.
Bottom-hinged flap gates which seal against the concrete face of an abutment or
a pier have been constructed. This requires that the tolerance of the concrete face is

Figure 2.25. Different

versions of bottom-hinged flap
gates operated by hydraulic
cylinders

25
 Hydraulic gates and 3 mm along both axes over the distance of gate movement. While this saves the
valves cost of embedded side plates, it exacerbates wear on the seals and increases
construction costs due to the requirements for accuracy and finish of the concrete.
Seal designs for bottom-hinged flap gates are illustrated in Chapter 7.
Where gates have to be fully retracted, the shape of the skin plate must be
formed so as to avoid flow separation and subatmospheric pressure.
The arrangement shown in Fig. 2.26 can be a trap for debris. Where
operational reliability is paramount, it may be necessary to provide means of
clearing debris without dewatering the sluiceway or using a diver. This can be
done using jets of water, compressed air or by flushing with river water. No
information regarding practical experience in the use of compressed air
distribution systems for this application has been published. If river water is
used to flush out the recess, the design must avoid the creation of eddies and
dead pockets which will cause the transfer of debris from one part of the recess
to another without dislodging it. Debris will be trapped behind the overflow jet
and a bypass system, shown in Fig. 2.27, is necessary to clear flotsam. The gate is
elevated to stop overflow when flushing out.
Tilting gates can be manually or motor operated through screws, or actuated
by means of two hydraulic cylinders on either side of the gate or by a single
cylinder centrally positioned or on one side only. Small gates are more often
operated by ropes. The arrangement of the hoist machinery is then similar to
that shown in Fig. 6.1.
Rising screw-type gearing with twin lifting screws operated from a central
headstock gives the required large mechanical advantage and also provides a
self-locking feature which resists the water load tending to reverse drive the
gears. The use of exposed lifting screws either in the sluiceway or in a recess in
sluiceway walls is a potential source of malfunction due to contamination of the
threads by dust, insects and silt, especially when coated by an adhesive lubricant.

Figure 2.26. Recess in

sluiceway to permit complete
retraction of flap gate under
flood conditions

26
 Types of gates

Figure 2.27. Bypass system

for flushing debris accumulated
downstream of a flap gate

The resulting increase in friction of the screw threads has caused seizure of
screws, particularly when multistart threads have been used. Totally enclosed
screws operating in an oil bath are preferred.
Oil hydraulic operation by placing the ram under the gate makes it possible to
dispense with an overhead structure. Maintenance of the ram then requires
dewatering of the sluiceway, often considered a major drawback of this
configuration.
At the Bala sluices of the Dinorwic Pumped Storage Scheme in Wales,
planning requirements stipulated that no overhead structure should be
provided. The gates were therefore designed as a torque tube structure with
the pivot shaft extended into the hollow piers, where they were operated by a
hydraulic cylinder acting through a lever (see Fig. 2.28).
A special case of a bottom-hinged flap gate is the velocity-control structure in
the River Orwell at Ipswich.8 In its elevated position, the gate acts as a
submerged weir reducing the cross-sectional area of the river, reducing the flow
and scouring action on the river bed.
Automatic tilting gates have been constructed with counterbalance above or
below the gate (see Fig. 2.29). The counterweight is arranged to balance the
overturning moment of the upstream water load at normal retention level. With
a rise in level the gate becomes overtopped and the overturning moment is
increased. When this overcomes the resistance of the counterweight, the gate
opens. By careful proportioning and positioning of the counterweight system
and the pivot point these gates can be arranged to open and close in a series of
movements on a rising and falling upstream level. The degree of control is not

27
 Hydraulic gates and
valves

Figure 2.28. Bala Sluices of

the Dinorwic Pumped Storage
Scheme

28
 Types of gates

Figure 2.29. Automatic

tilting gate

accurate and a significant variation in level is required to effect full gate travel. If
on discharge the downstream level starts to rise, thus creating an overturning
moment tending to close the gate, it may be impossible to prevent the upstream
water level from rising.
The gate arrangement shown is inherently hydraulically unstable.
Disturbance may be set up due to wave motion in a reservoir, or a pulsating
surge or wave may be set up in the upstream reach of a long approach channel
of uniform section. Surging can be initiated by level drawdown immediately
upstream of the gate following a downward movement. The loss of water load
on the gate then causes a closing movement, and if the frequency of gate
oscillation coincides with that of the surge wave the gate movement is
accentuated and can become dangerous.
The only damping force present is the friction of the side seals, hence
hydraulic dashpots have to be added to the counterweight system. A
control system which is more stable utilises the same principle as radial
automatic gates by arranging the counterbalance weights so that they act
as displacers.

Venting
The nappe has to be vented to prevent gate vibration and nappe collapse. Flow
dividers are used to vent the nappe under moderate overflow conditions. The

29
Hydraulic gates and design and spacing of flow dividers is critical.3 The dividers must project beyond
valves the gate lip and they must be wide enough to form an adequate opening in the
nappe for the admission of air. The flow of water over the gate lip expands but
as soon as it is no longer in contact with the divider it tends to close it up again.
Experimental work and prototype trials have been carried out to study nappe
oscillation and resulting vibrations.914 Model studies to Froude scale of
bottom-hinged flap gates incorporating flow dividers are ineffective in
preventing self-excited nappe oscillations and resulting vibrations.15 The
spacing of flow dividers is important. Pulpitel16 gives some information on
flow dividers which can be applied in practice. Initially, flow dividers were
spaced at 2100 mm centres. Additional flow dividers were added between the
original ones. These projected 280 mm beyond the skin plate and had a crest
width of 300 mm and a width over the tapering side sections of 450 mm. The
maximum overflow depth was 720 mm.
When the head of water above the gate lip is appreciable, flow dividers
become ineffective17 and additional venting through the sluiceway walls or
the piers has to be provided. A method for calculating the air demand of an
overflow jet is given in Chapter 9.

Top-hinged flap gates

Top-hinged flap gates are used in tidal structures to prevent flooding of an
inland region by sea waters during rising tides or flood surges and to permit
inland waters to drain off into the sea during ebb tides. They are also used in
culverts and pumped drainage outfalls to rivers.
They do not require an outside source of power and operate automatically.
The construction of the gates is simple and little maintenance is required. The
gates will not entirely exclude ingress of saline water if the downstream water
level rises above the sill during discharge under the gate, when a lens of saline
water can penetrate upstream against the flow.
They control water in one direction only and perform like a non-return
valve. They cannot control the upstream level. In stormwater discharge this
facility is not required. Top-hinged flap gates can be operated under clear or
drowned discharge conditions. When designing gates of this type a gravity bias
is required in the closed position so that the gates close immediately before
reversal of flow occurs. This can be effected by slanting the closed gate position,
an arrangement which was investigated as part of the Severn Tidal Project Study
(Bondi scheme), or the flap may be in the vertical position when closed, with the
bias to closure due to eccentric hinges (Fig. 2.30).
Flap gates using an elastomeric gate leaf (Fig. 2.31) deflect to provide the
fluid passage for the flow of storm water. Under flow conditions top-hinged
flap gates deflect the discharge downwards, and scour can occur where the gate
sill is close to a river bed or the bed of an estuary. A slanted gate provides part of
the antiscour apron and can offer some economies in civil engineering
construction, but the apparent gain can be cancelled out where there is a
requirement for stoplog grooves downstream of the gate.
In a vertical leaf arrangement the closure bias can be provided by an eccentric
hinge, or by a weight mounted ahead (downstream) of the flap. If the discharge

30
 Types of gates

Figure 2.30. Top-hinged flap

gate

conduit runs full, the eccentric hinge opens a gap which will cause flow over the
gate leaf in the open position. However, the eccentric hinge arrangement
reduces the total mass of the flap compared with the weighted flap and therefore
provides an increased discharge for a given opening. This is due to the
relationship between the discharge and the mass of the flap. At a given gate
opening these are exponentially related.
Sealing is usually effected by face-to-face contact between the flap and body
of the gate. This requires exact positioning of the hinges. To ensure even contact

Figure 2.31. Top-hinged flap

gate with elastomeric leaf

31
 Hydraulic gates and
valves

Figure 2.32. Adjustable pivot

lugs for a top-hinged flap gate

between the sealing faces the pivot lugs are made adjustable. One example of this
is shown in Fig. 2.32. The cushioning of the gate leaf in Fig. 2.33 is effected by
the movement of the projecting section of the flap in the seat. This acts like a
piston moving in a cylinder. The extent of cushioning is determined by the
clearance between the faces. The section in the seat in contact with the flap must
be tapered because the flap moves in an arc. If elastomeric seals are fitted they
should not be located on the face of the flap because this causes the flow to
separate. Seals should therefore be mounted on the frame. Under flow
conditions the flap rides the discharge jet. Under free discharge as well as some
conditions of drowned discharge a reverse roller forms at the lip of the flap (Fig.
2.34). This is similar to the discharge conditions at underflow gates of the radial
and vertical-lift type. If the conduit flow is near full or under supercharged
conditions, transverse flap stiffeners at the lip of the flap can cause flow
reattachment which is likely to cause vibration. Transverse flap stiffeners should
be set back and the lip of the flap should be stiffened by webs leading from the
transverse stiffener to the leading edge.
When a series of flap gates are close to one another, such as in estuarial tidal
outfalls, the sideways discharge under the gate causes flow interference between
adjoining gates and results in hydraulic losses (Fig. 2.35). If the discharge through
the gates has to be maximised, the losses must be reduced by training walls. These
should extend as far as the length of the sector swept by the opening of the flap.
The theoretical treatment of the stage^discharge relationship of a flap gate by
Pethick and Harrison presupposes that there is no sideways discharge. It is
reproduced in Chapter 9.

32
 Types of gates

Figure 2.33. Top-hinged flap

gate hydraulically cushioned

Wave action or flow reversal will cause rapid movement of the flap and can
result in severe slamming. This can be damped by a hydraulic cushion (Fig.
2.33), or by an oil hydraulic damper. For a large gate it can be of the oleo-
pneumatic kind, or a torsional damper can be introduced in the hinge assembly.
Another method of suspending top-hinged flap gates was used on the
Ishmalia gates. This incorporated a cycloidal rocker-type bearing instead of
pivots or hinges and proved reliable in service.

Fuse gates
Fuse gates consist of an alignment of elements standing on the crest of a dam on a
concrete sill with a toe abutment (Fig. 2.36). Under low flood conditions they act
as a labyrinth weir (Fig. 2.36(b)). At higher floods, water flows into the well and
into the hollow base set over the sill (Fig. 2.36(c)). The uplift pressure on the base
and the change in the centre of gravity of the fuse gate causes the gate to overturn

Figure 2.34. Reverse roller at

the lip of a flap gate

33
 Hydraulic gates and
valves

Figure 2.35. Flow

interference due to sideways
discharge between adjoining flap
gates

Figure 2.36. Fuse gate

34
when the design flood level is reached (Fig. 2.36(d)). Fuse gates are made of steel, Types of gates
concrete, or a combination of both. A seal is provided between adjoining fuse
gates. Their main application is to increase the reservoir capacity of existing
uncontrolled spillways, although they have also been fitted to new dams.
The discharge characteristics of fuse gates are similar to those of an ungated
spillway until the elements start to tilt. For large floods, the elements tilt
independently of each other, progressively following the rise in head water
level. Discontinuities arise as each element is activated. The reservoir retention
level, after the flood has subsided, is then the level of the concrete sill until the
fuse gates have been recovered and repositioned. When the elements overturn
they can be recovered just downstream of the dam and can sometimes be
reutilised, which may require reworking. If the velocity of flood discharge is
higher than 3 m/s, the elements are carried away by the current.
Fuse gates can be set on a multiple level sill to refine flood routing.
Floating debris is said to have no effect on block stability and little impact on
the precise timing of tilting.
Existing dams, if fitted with fuse gates, have to be checked for the higher load
due to an effective increase in height and the transfer of the shearing forces
exerted at the sill.
For small and moderately large dams elements ranging from 0.5^2.5 m high
have been used and for larger scale projects fuse gates up to 6.5 m high have been
designed.
The main criteria for selection of fuse gates on an existing dam are:
the additional loads imposed by fuse gates.
The main criteria for selection of fuse gates on all dams are:
The required limits of control of the reservoir level. This may have to include
the reservoir level after tilting of fuse gates during a flood.
The time and effort required to recover fuse gate sections after a flood.
Possible replacement of damaged gate sections or repair of reusable units.
The frequency of flood events which cause fuse gates to tilt.

Barrier and barrage gates

Rising sector gates
The novel concept of a rising sector gate was developed for the Thames Barrier
at Woolwich. This was a response to the requirement for overhead clearance or
no obstruction with the gate open to provide unrestricted navigation within the
main openings.
The barrier is formed by four 61 m clear width main navigational openings
and two 31.5 m wide subsidiary navigational openings. Four falling radial gates
31.5 m wide flank the navigation gates, one on the south bank of the river and
three on the northern side.
The gates are supported by a series of nine piers and two abutments, with
concrete sill units spanning between piers at river bed level. In the open position
the rising gates are housed in shaped recesses in the sill units so that the flat upper
surface of the gate is flush with the sill surface and does not intrude on the full

35
Hydraulic gates and
valves

Figure 2.37. Thames Barrier and principle of operation of the Thames Barrier gates (after
Ayres18)

36
navigational depth of the opening (Fig. 2.37). To close the gates they are rotated Types of gates
through approximately 90 until the flat skin plate is near to the vertical and the
curved skin plate is facing seawards. For maintenance, the gate is rotated
through a further 90 until it is fully inverted. A gate segment is supported at
its ends by two disc-shaped arms (Fig. 2.38).
The arms rotate on a shaft and bearing assembly (Fig. 2.39). The bearings are
self-aligning and are of the self-lubricating type described in Chapter 7.
Provision for supplementary lubrication is provided.
The main gate operating machinery is shown in Fig 2.40. It consists of two
hydraulic cylinders (1). They rotate a rocking beam (2) through crossheads and
links (3). The cylinders are disposed on either side of pivot shaft (4) of the
rocking beam. The cylinders develop a thrust of 15.21 MN at a pressure of
17.24 N/mm2. The rocking beam operates the disc shaped arm through the
connecting link (8).
To raise the gate, the upper cylinder pulls and the lower cylinder thrusts the
rocking beam causing it to rotate about its pivot, raising the gate link and so
rotating the gate. To rotate the gate through the link dead centre position a shift
and latch mechanism is provided. It is actuated by a hydraulic motor driving a lead
screw. It is also used to lock the gate in the open, closed and maintenance positions.

Figure 2.38. Thames Barrier gate (after Clark and Tappin19)

37
 Hydraulic gates and
valves

Figure 2.39. Thames Barrier

gate shaft and bearing assembly
of rising sector gates (after
Clark and Tappin19)

Identical units are provided at each end of the gate. They can be used in an
emergency to raise a gate should the main machinery be out of action.
A high degree of redundancy is built into the system, specifically to the
power packs and gate raising machinery. The gates can be elevated into the

Figure 2.40. Thames Barrier

gate, methods of gate operation
(after Fairweather and
Kirton20)

38
 Types of gates

Figure 2.41. Venice Barrier

buoyant gates

defence position by the machinery on either side or by the two sets working
together, the fall-back being provided by the latch mechanisms. This also
applies to the operating controls.

Bottom-hinged buoyant gates

The design of the Venice Barrier gates, which have so far not been built, was
evolved to avoid any piers within the navigation ways from the Adriatic Sea into
the Venice Lagoon. Cruise ships up to 30 000 tonne displacement pass through
the Lido passage to the port of Venice and very large supertankers enter the
Lagoon through the Malamocco opening.
The gates recess into caissons in the navigation ways (Fig. 2.41). To raise
them, compressed air is admitted into the gates and water is expelled causing
the gates to rise into their operating position at an angle of 50 to the horizontal.
The gates can withstand a differential head between the waters of the Adriatic
and the Lagoon of up to 1.8 m. This is due to their buoyancy and mass. The four
barriers, each 400 m wide, are comprised of gates 20 m wide, moving
independently of one another (Fig. 2.42). There is a leakage flow between the
gates but the effect on the water level of the Lagoon is negligible because of its
large area (about 500 km2).

Figure 2.42. Barrier across

navigation opening into Venice
Lagoon

39
 Hydraulic gates and
valves

Figure 2.43. Venice Barrier, method of exchanging gates for maintenance

Figure 2.44. Venice Barrier,

detachable gate hinge

40
 The gates can be transported by crane to a platform on the shore for servicing Types of gates
and a replacement gate can be substituted (Fig. 2.43). Detachment and
reattachment of the hinges is by a series of hydraulic cylinders radially disposed
around the mandrel of the hinge (Fig. 2.44).
Air is admitted and exhausted via pipework in the service gallery of the caisson.
There are two independent, physically separate systems. The air is routed through
the detachable hinges. Detachment of the hinges automatically blocks the air
admission ports. In case silt is deposited on a mandrel of the hinge when a gate
has been removed, a flushing system is incorporated in the permanently attached
part of the hinge and is actuated prior to engagement of a replacement gate.
Sand and sedimentary matter will be deposited in the caissons during slack
tides and wave action. The caisson recess for the gate is in the form of a series of
troughs. A hydraulic transport and ejection system, consisting of evenly spaced
jets positioned above the crest of the troughs, clears about 80% of the material
which accumulates in the troughs. This is effected by supplying water under
pressure to the jets which then move the deposited material along the troughs
to the end, where jet pumps elevate and expel the silt and sand into the
navigation way.21
The barriers and gates have been described by Lewin and Scotti.22

Large span vertical-lift gates providing navigation clearance

A large span vertical-lift gate forms the River Hull Tidal Surge Barrier (Fig.
2.45). When the gate is hoisted to its uppermost position it rotates and comes
to rest in the horizontal plane, reducing the overall height of the support
structure. This method of storing a gate in the fully open position is sometimes
adopted for vertical-lift lock gates. The upstream lock gate of the Kotri Barrage
on the River Indus in Pakistan is an example of this type of construction.
The flood surge protection gate at Barking Creek on the Thames,
downstream of the Thames Barrier, is of similar design (Fig. 2.46). The
navigation gate is 38.6 m wide and 10.8 m high. The gate can be raised 39.4 m
for ships to pass underneath. In the open position the gate is stored vertically.
The navigation gate is flanked on one side by two vertical-lift gates and on the
other side by one gate. The gate is counterbalanced and each end is suspended by
two Renold chains of 165 mm pitch shown in Fig. 7.19. Each roller has a grease
nipple to lubricate the roller bush.
Two sets of hoist machinery are provided, located at high level in each tower.
A shaft interconnecting the hoists runs along the bridge tying the towers. Either
one of the hoists can be used to lift the gate but they cannot be used together.
In an emergency, the navigation and all the side gates can close under gravity.
Closure is then controlled either by hydraulic retarders or by hydraulically
operated disc brakes, mounted on the shaft which drives the chain hoist
sprocket. Figure 2.47 shows the arrangement of the overspeed disc brake at
one of the hoists of the navigation gate.

Flap gates for storm surge protection

An unusual type of gate was developed for the storm surge protection of King
George V lock on the Thames downstream of the Thames Barrier. The

41
42

Hydraulic gates and

valves
Figure 2.45. River Hull Tidal Surge Barrier
 Types of gates

Figure 2.46. Barking

Creek Barrier

navigation opening leading to the dock is 30.498 m wide. The bed level at the
apron on the Thames side is at 11.85 m and the flood defence level is +7.2 m.
The structure consists of a bridge moving on a roller track (Fig. 2.48). The
overall length of the moving bridge is 72.5 m. A flap gate is hinged from the
bridge. To close the navigation opening, the bridge moves laterally so as to span
the lock entrance. The flap gate is then lowered to abut on a step in the
navigation way. The hydrostatic thrust due to a surge flood is resisted at the
bottom of the step and at the top it is transmitted via the gate hinges to the
moving bridge. The bridge transfers the force to two thrust blocks on either
side of the lock entrance.
The bridge is surmounted by two hoist winches for lifting and lowering of
the gate. The winches are driven by low speed oil hydraulic motors. The bridge

Figure 2.47. Barking Creek

Barrier hoist, disc brake which
can be used for emergency gravity
lowering of the navigation gate

43
 Hydraulic gates and
valves

Figure 2.48. A top-

hinged flap gate mounted on
a moving bridge (King
George V lock on the
Thames, storm surge
protection gate)

is moved by rope sheaves which are also driven by oil hydraulic motors. The
sheaves utilise fixed haulage ropes. A stationary emergency winch can launch
the bridge but cannot retrieve it.
The gap between the gate and the underside of the bridge structure, where
the gate hinges are located, is sealed by hydraulic cylinder operated flaps.

Lock gates acting as storm surge barriers

Two main types of lock gates are used to protect river mouths from high tides or
storm surges. These are mitre gates and vertically hinged sector lock gates. In a
few instances small caisson gates and pointing gates are also used for this
purpose.
At the storm surge barrier of the River Hunte in northern Germany,
conventional mitre gates are arranged in two pairs in parallel. The 26 m wide
navigation passage is protected by the 12.7 m high gates. Two radial gates of
20 m span are disposed in parallel on each side of the navigation opening. In
all cases the second gate or pair of gates acts as a back-up to the seaward gate
or pair of gates.
The disadvantage of mitre gates when used for flood protection is the
requirement for near balanced conditions for opening and shutting. This makes
it impossible to operate the gates in anticipation of a flood. It is presumed to be
one of the reasons why two pairs of mitre gates form the main part of the Hunte
Barrier. In the event of the failure of one pair of gates to close, there is little time
to activate emergency procedures. Another disadvantage is the heavy mitre
thrust exerted by large gates, which is considerably in excess of the hydrostatic
load on the gate. This increases the cost of the civil engineering works.
These disadvantages are not present in the vertically hinged sector lock gate.
Figure 2.49 shows the gates for a 24 m wide 12 m deep navigation passage for
storm surge protection of Tokyo City.
In sector lock gates the skin plate is formed in a true radius. The resultant
forces acting on the gate pass through the hinges and do not produce
unbalanced moments which would tend to open or close the gate. The river
level can be higher than the estuary level or the reverse can occur without

44
 Types of gates

Figure 2.49. Vertically-

hinged sector lock gates for
storm surge protection of Tokyo
City

affecting the sealing of the gate. The gate can be opened and shut against flow in
either direction. A disadvantage of this type of lock gate is the requirement for a
large recess in the river wall to accommodate the gate leaves when they are
opened.
Large sector lock gates safeguard a river on Rhode Island on the Atlantic
coast of the USA against surges due to hurricanes.
The largest example of a sector gate is the New Waterway storm surge barrier
in The Netherlands.23,24 It protects the Rotterdam area against floods from the
North Sea by closing off a 360 m wide waterway. The barrier consists of two
movable box shaped sector gates, each having an arc length of 203 m and a
height of 22 m. The radius at the sea side of the box is 240 m. Two trussed steel
arms support each of the buoyant boxes and transfer the water load to a ball and
socket joint (Fig. 2.50). The joints, of 10 m diameter, allow the gates to be
floated into position. When water is admitted the gates settle on the sill 17 m
below datum.
In the open position the gates are parked in dry docks shaped in an arc. The
operation of closing and opening the gate leaves is by a `locomobile'. It is fixed
and transfers traction to rails on top of the gate wall, thus the locomobile rides
on top of the gate sector, imparting horizontal movement. Figure 2.51 shows
the arrangement. The connection between the locomobile and the support on
the dock allows for the vertical motion which occurs when the gate is ballasted
and when it is buoyant. While the locomobile remains stationary, the gates
move horizontally.
The gate is required to lift to equalise a differential seaward head, or a higher
level on the river side. The initial design (Fig. 2.52(a)), was unstable under these

45
 Hydraulic gates and
valves

Figure 2.50. New Waterway

storm surge barrier and ball and
socket rotational joint (after
Ieperen23)

conditions because the gates were strongly influenced by their own bottom side
geometry. Model studies arrived at a design which ensured stable operation
under both positive and negative heads (Fig. 2.52(b)).
Other model tests investigated bed protection. Riprap bed protection was
required on both sides of the barrier in order to withstand hydraulic loads
during the closure operation (flood) and the opening operation (ebb).
Conventional vertical-lift gates were used at the Eastern Scheldt Storm
Surge Barrier in The Netherlands. The movable barrier comprises 63 gates
of 42.5 m width with height varying from 5 to 11.9 m and a depth of 6 m.
They are operated by a hydraulic cylinder at each end. The largest rams
have a stroke of 12 m and an internal diameter of nearly 2 m. If severe wave
action is encountered the cylinders are required to hold the gates rigid
under all hydraulic conditions, including waves impacting upwards and
downwards.

46
 Types of gates

Figure 2.51. `Locomobile'

driving a sector of the New
Waterway storm surge barrier

Bottom-hinged flap gates

At impounding barrages the requirement is to prevent saline water intrusion
into the basin formed by the barrage. At the same time, the river flow has to
be discharged to sea. Any form of undershoot gate will permit a lense of saline
water to progress upstream irrespective of the discharge under the gate. The
bottom-hinged flap gate (Fig. 2.25(c)), is usually selected for this purpose,
because discharge is by overflow. At the Tees Barrage and the Langham Weir
in Belfast, the gates are operated by a single cylinder located at one side of the
gate.

Hook-type double leaf gate

Hook-type gates were used at the Cardiff Bay Barrage. Figure 2.21 illustrates the
type of gate but is not a reproduction of the gates used at that barrage. The upper

Figure 2.52. New Waterway

storm surge barrier in the
Netherlands, determination of
a hydraulically stable cross-
section of the floating sectors
(after Ieperen23)

47
Hydraulic gates and leaf can be lowered to provide overflow from the upstream reservoir and the
valves lower leaf can be raised for discharge under the gate when the ebb tide is below
the sill level and it is necessary to discharge silt accumulated at the gates.

Caisson or sliding gates

Caisson gates move horizontally on rollers and are retracted, when not in
operation, into rectangular chambers at right angles to the navigation way.
Actuation is by an electric motor driven winch and haulage chains passing over
sprockets.

Pointing gates
Pointing gates are vertically hinged double leaf flap gates. They operate
automatically without external power on movement of the tide. They are used
on small rivers to prevent the ingress of tidal water and provide protection
against storm surges. Older gates were constructed of hardwood. New gates
of this type have been constructed in steel. The gates open at low tide due to
river flow and close on the rising tide. Sealing is by face contact with the
masonry structure. The gates can slam on closure. Damage to masonry works
sometimes occurs as a result of violent closure. No attempt appears to have been
made to damp or to provide buffers on gates on the Somerset Levels in England.
In existing installations, pointing gates are backed by vertical-lift gates which
provide a standby in case of gate breakdown or malfunction due to debris being
caught between the gate leaves, preventing complete closure.

Possible gates for tidal power barrages

The Severn Estuary is one of the world's best sites for tidal power. Part of the
reason for the high tidal range in the Severn is that the estuary is close to
resonance. In 1978^81, a committee under Sir Hermann Bondi considered a
tidal barrage from Lavernock Point, south of Cardiff, to Brean Down near
Weston-Super-Mare.25 The installed capacity would have been 7.2 GW if the
barrage had been built. The firm power contribution worked out at 1.1 GW.
Two different gate designs were considered to facilitate admission of sea
water to the upriver estuary for ebb generation, and were subject to model
studies. Figure 2.53 shows top-hinged flap gates arranged at three levels. The
attraction of this scheme was automatic operation of the gates without an
external power source. Figure 2.54 illustrates an application of vertical-lift gates
where the water passage was designed as a Venturi to minimise hydraulic losses.

Drum and sector gates

Drum and sector gates are acute circular sectors in cross-section. Gates hinged
on the upstream side are referred to as drum gates (Fig. 2.55(a)), and those
hinged on the downstream side as sector gates (Fig. 2.55(b)).
The gates are designed so that they can be fully retracted, making the upper
surface coincident with the crest line. Control of the gates is automatic and is by
admission of the upstream water level into the float chamber.
Drum gates float on the lower face of the drum, whereas sector gates are
usually enclosed only on the upstream and downstream surfaces.

48
 Types of gates

Figure 2.53. Severn Estuary

tidal power study, flap gates for
admitting water to the upper
estuary (Severn Barrage
Committee25)

These gates are not suitable for low dams because of the deep excavation
required and the possibility of flooding of the float chamber due to downstream
water level. Some very large drum gates have been built, up to 40 m long and
9 m high.
Drum and sector gates have been superseded by radial gates at spillways
because the former are more complex to manufacture and therefore more costly.
The cost of civil engineering works associated with drum and sector gates is
significantly higher than with radial gates.

Figure 2.54. Severn Estuary

tidal power study, vertical-lift
gates for admitting water to the
upper estuary (Severn Barrage
Committee25)

49
 Hydraulic gates and
valves

Figure 2.55. Drum and sector

gates

Bear-trap gates
A bear-trap gate consists of two leaves, one hinged upstream, the other
downstream (see Fig. 2.56). Both leaves are sealed at their side and pivots and
are free to slide or roll relative to one another with a sliding seal at their juncture.
When the gate is lowered, the leaves come to rest in the horizontal position
with the upstream leaf on top of the downstream one. When the upstream water
level is admitted to chamber `a' the gate can be raised.
The water pressure under the gate is controlled either by an adjustable weir or
by setting the inlet and outlet sluice valves in a control chamber in the sluiceway
abutment.
Bear-trap weirs have been used in the USA for log-sluicing operations, when
the skin plates are usually protected by hardwood skid timbers.
The accumulation of silt under a bear-trap weir set on the river bed has been a
source of trouble and various methods have been developed for the removal of
silt by sluicing.
The control system and the seals of a bear-trap gate are critical. There is a
recorded instance of breakdown of a bear-trap weir due to vibration caused by
bad design of the hinge seal. This problem applies equally to bottom-hinged
flap gates.

Figure 2.56. Bear-trap gate

50
 The calculations for each equilibrium condition of the gate have to be carried Types of gates
out separately, taking into account the external water load, the differential head
required to raise the gate and the water level in the chamber to maintain the gate
in position. The gate can be arranged to operate automatically to maintain
upstream water level, although close control with variable tailwater levels is
difficult to achieve. Raising the gate by admission of water to chamber `a'
requires effective side and sill seals. Bear-trap gates are now seldom constructed
and when used they are often raised by mechanical lifting gear travelling across
the weir.

Gates in submerged outlets

Intake gates
Vertical-lift intake gates can be of the upstream or downstream sealing type as
shown in Fig. 2.19(a). Upstream sealing gates are located in a shaft, a short distance
from the intake. Gates which control flow or have to be shut against flow in an
emergency must be operated by an oil hydraulic servo-motor. The reason why a
gate on an elastic suspension such as a wire rope can be subject to vibration due to
high velocity flow under the gate is discussed in Chapter 10. If the servo-motor is
located above the reservoir level, it has to be connected to the gate by a series of
stem sections interconnected by knuckle joints, as shown in Fig. 2.57.
The gate is raised to its maintenance position by hoisting it through the full
stroke of the servo-motor, dogging the next lowest knuckle joint, then
removing the uppermost stem and lowering the piston so that the piston rod
can be reconnected to the next lowest stem. This operation is repeated until the
gate reaches its service position.
Downstream sealing gates, and in some cases upstream sealing gates, require
air supply pipes. Upstream sealing gates are usually opened under conditions of
balanced head, and this invariably applies to bulkhead gates. To effect balanced
head conditions, either tunnel filling valves are incorporated in the gate or a
valve controlled bypass system is provided. In rope suspended gates, tunnel
filling valves integral with the bulkhead or intake gates, as shown in Fig. 2.58,
are opened by initial tensioning of the hoist ropes. A short movement of the
hoist lifts the valves. Tension in the ropes is sustained until the pressure on both
sides of the gate is equalised. Further hoisting of the ropes then lifts the gate.
The type of valve suitable for this application is also used in stoplogs where it
performs the same function and is illustrated in Fig. 2.67.
Valves controlling bypass systems of high head gates should be checked for
cavitation conditions. In a piped bypass controlled by an operating valve and a
guard valve, the system may require a tapered outlet to cause a back pressure, so
that cavitation occurs external to the discharge section. Protection of the sluice
wall from an impinging jet may also be required.
Intake gates which are not required to shut against flow, and bulkhead gates,
can be rope suspended and lowered or hoisted by a conventional winch (Fig.
2.58).
Intake and bulkhead gates can be of the fixed roller type or for very heavy
duty a caterpillar gate, known in the USA as a coaster gate (Fig. 2.59), is used.

51
Hydraulic gates and
valves

Figure 2.57. Intake gate, servo-

motor operated

Slide gates are also used as intake gates. If the gate is only operated under
balanced conditions, the slide material has to be able to withstand the
hydrostatic forces with no head on the downstream side of the gate. The
frictional properties of the slide material in this application are of secondary
importance. Impregnated woven asbestos is sometimes used. If the slide gate
controls the intake flow, the slide-bearing material is similar to that of the
control gates.
Intakes can be controlled by radial gates, as shown in Fig. 2.60. The
advantage of using radial gates is the absence of gate slots (which can cause
hydraulic problems in high velocity flow), guide rollers which have to operate
totally immersed, or slides. The gate is rigid with no slack in its movement and
operating forces are less than those required for vertical-lift gates.
The disadvantages are the requirement for a chamber to retract the gate, the
fact that the gate cannot be withdrawn to the surface for maintenance and, in
some cases, the immersion of the operating cylinder or cylinders while the gate is
in the open position.
Cylinder gates (see Fig. 2.61), are used where the controlling gate must
operate in a shaft or intake tower. They are used as shut-off gates and for
regulating the intake. The gates are guided by rollers operating on tracks fixed
to the tower walls, and therefore have little mechanical friction to overcome

52
 Types of gates

Figure 2.58. Intake gate, rope

operated

hydrodynamic excitation. Long operating stems or suspension chains of

cylinder gates can result in low resonance frequencies.
Researchers have reported vibration problems with cylinder gates. Vibration
experienced at some cylinder gates appears to have been due to lack of a sharp
cut-off point at the lower lip. Vibration which has occurred at low gate openings
is consistent with the variation of hydraulic downpull forces due to unstable
flow. The general principles of gate vibration are dealt with in Chapter 10.
Ball26 conducted a model study of a high head cylinder gate which
demonstrated that cavitation could occur due to the preliminary gate seat design
and also vibration of the gate. Bixio et al.27 found asymmetric pressure
distribution on the shell of a cylinder gate due to unsteady flow through the
eight openings of the intake. Negative pressures were recorded under
emergency closure conditions, especially at the lower edges of the gate.

Control and guard gates

Control and guard gates may be vertical-lift, roller or slide gates which are
servo-motor operated and retract into a bonnet (Figs 2.62 and 2.63). Frequently
two identical gates are used.

53
 Hydraulic gates and
valves

Figure 2.59. Caterpillar or

coaster gate

The discharge from a slide gate is smooth and the only limitation is discharge
at very small openings when the flow does not spring clear of the gate lip and is
liable to produce cavitation damage at the bottom of the gate.
The gates are designed with the skin plate downstream and with open
stiffener girders on the upstream side. This causes flow circulation between
the girders which is not detrimental. An alternative design is box construction
with the space between the girder flanges filled in. Some gates, such as the
bottom outlet gate at the Victoria Dam in Sri Lanka, have been manufactured
from solid forged steel plate.
The bonnet is designed to withstand the full hydrostatic head without any
structural contribution from the embedding concrete.
The transverse deflection of the gate must be very low so that the slope at the
bearing faces does not cause uneven contact pressure at the slide faces. Keeping
the design contact pressure below the permissible bearing pressure of the slide
material will permit some variation in the imposed pressure.
Unless the deflection of the gate is very low, it may be desirable to cross-
radius the bearing to allow for gate deflection, in which case the contact pressure
(Hertz) calculations have to be carried out accordingly.
The slides can be of conventional bearing materials such as leaded bronze,
aluminium bronze or manganese bronze. Used on their own they require high
pressure grease lubrication.This has to be applied to the slide contact face within
the gate slot by pipes leading from grease nipples at the top of the bonnet to a

54
 Types of gates

Figure 2.60. Intake gate of the radial

type

number of selected points on the slide face. Grease distribution grooves must be
incorporated in the slide face to distribute the grease. When applying grease,
effective distribution will occur only at grease outlets masked by the gate. There
is a danger that grease or lubricant on exposed slide faces can be washed away by
the recirculatory flow within the gate slots when the gate is in a partially or
wholly open position. Possible environmental contamination may have to be
considered.
The use of bearings with lubricant inserts eliminates the sliding seat greasing
pipes and ensures an even lubrication coverage. In this type of slide, the
lubricant is compressed into trepanned recesses in the bearing. The lubricant is
of a permanent, solid, thick film nature and is a compounded mixture of metals,
metallic oxides, minerals and other lubricating materials combined with a
lubricating binder. Graphites containing lubricants should not be used in
conjunction with stainless steel as they cause electrolytic action, which is
accelerated underwater.

55
 Hydraulic gates and
valves

Figure 2.61. Intake gate of the

cylinder type

When a gate is opened and discharges into an empty tunnel, an air demand is
created due to air entrainment in the air/water transition region. The calculation
of air demand is dealt with in Chapter 9.
Difficulties can be experienced at gate slots at high velocity flow. These are
discussed in greater detail in Chapter 9. Hydraulic problems in gate slots have
led to the development of jet-flow gates. The gates incorporate contraction
slopes on the conduit upstream from the gate slots to cause the flow to jump
the slots in order to avoid intermittent flow attachment (Fig. 2.64).
The jet-flow gate shown in Fig. 2.64 is of the United States Bureau of
Reclamation (USBR) circular orifice type. Other types of jet-flow gate with
rectangular outlets have been developed. A rectangular gate seals flush at the
sill and the contraction section is omitted at the sill. The rectangular jet-flow
gate is appreciably cheaper to manufacture. The cost advantage is to some
extent offset by the requirement for a transition section of conduit from
circular to rectangular and on the downstream side from rectangular to
circular.
Radial gates can be used as control gates in conduits, arranged as shown in
Fig. 2.65, or the gate may be located in a shaft (Fig. 10.16).

56
 Types of gates

Figure 2.62. Control and

emergency closure gates of the
slide type

Figure 2.63. Slide gate

57
 Hydraulic gates and
valves

Figure 2.64. Control gate of

the jet-flow (circular orifice)
type

As discussed earlier in this chapter, the main advantage of using a radial gate
is the absence of gate slots. The disadvantages are a substantially larger gate
chamber or shaft, and in many cases difficulty of access for initial assembly and
subsequent maintenance.
A ring-follower gate (Fig. 2.66), is selected as a guard gate where the terminal
discharge is controlled by a valve. It removes the need for the upstream transition
section joining a circular conduit to a rectangular one, and a similar downstream
transition section to a circular cross-section. In the fully open position it provides
an unobstructed fluidway. The gate can therefore reduce hydraulic losses to outlet
works and result in economies in the transition sections.
The gate leaf retracts into the uppermost body, the bonnet section, when the
circular opening aligns with the fluidway to present an unobstructed flow
passage. To close the gate the circular opening is lowered into the bottom
section of the bonnet and the bulkhead portion of the leaf blocks the fluidway.
The lower bonnet has to be drained and designed for flushing of accumulated
sediment.

Figure 2.65. Control gate of

the radial type

58
 Types of gates

Figure 2.66. Ring-follower gate (half

section)

The disadvantage of a ring-follower gate is its size, 312 times the diameter of
the fluidway. This results in increased fabrication cost compared with a valve
serving the same pipe diameter. However, the size of a ring-follower gate is not
limited, unlike that of a valve.

Emergency closure gates, maintenance gates and

stoplogs
An emergency closure gate can close against flow. It will also perform the
function of a maintenance gate. It requires load rollers. A maintenance gate
will not normally close against flow. It can incorporate guide rollers or rely
entirely on slides. It has to be placed under balanced conditions. Both
emergency closure gates and maintenance gates can be in one section or in
several sections assembled into one gate prior to lowering. Other terms in
use, such as bulkhead gate, guard gate or stop gate, designate maintenance
gates in most cases.

59
Hydraulic gates and The term stoplog derives from the time when it was general practice to isolate
valves a sluice installation with wooden beams. Stoplogs cannot be placed in flowing
water because they are liable to vibration during lowering and raising when
combined over- and underflow conditions occur. Stoplogs can be designed to
be guided by rollers, or more frequently by slides. Stoplogs incorporating guide
rollers are easier to place.
Maintenance gates and stoplogs can be positioned by a rail-mounted gantry
crane or by a mobile crane. Emergency closure gates should only be placed by a
rail-mounted gantry crane. Hydraulic downpull forces (see Chapter 9) could
topple a mobile crane.
Figure 2.67 shows the essential features of a stoplog comprising lower seal
`a', upper seal contact plate `b' and side seals `c'. If the side seals are designed to
seal both upstream and downstream it enables the gates to be tested
hydrostatically on commissioning before the reservoir is fully impounded. If
this is considered desirable the stoplogs must also be designed for reversal of
the hydrostatic thrust. Bypass valves `d' are linked with the grappling beam
anchor points so that initial lift movement of the grappling beam opens the
valves to equalise the water level upstream and downstream of the stoplog.
Landing sensing device `e' on the stoplog is positioned on the sill beam or on
top of another section. The rod is displaced upwards and permits disengage-
ment of the grappling beam. If the stoplog jams or meets an obstruction, release
of the grappling beam cannot be actuated or accidentally effected.

Figure 2.67. Stoplog section

60
 Figure 2.68. Grappling beam and stoplog

Types of gates
61
Hydraulic gates and Stoplog guide channels
valves
The same structural design criteria apply to the design of stoplog guide channels
as to those of vertical-lift gates. They are usually simpler with embedded parts
only for the slide face of the stoplog, the contact face of the stoplog side seal and
any other sliding or rolling face. If it is important to enhance the hydraulics of
flow through the sluiceway, stoplog masking plates are used. They are placed
and withdrawn by the crane handling the stoplogs.

Grappling beams
Figure 2.68 shows a typical grappling beam. It is designed to engage
automatically with a stoplog, and will automatically disengage once the
beam is in position and the landing sensing device is activated. By moving
lever `a', the operator determines whether the grappling beam is in the
engage position (for recovering a stoplog) or the disengage position (for
placing a stoplog). Guiding in transverse and longitudinal directions is
effected by rollers.

Cranes
Emergency closure gates must be handled by a rail-mounted gantry crane,
whereas maintenance gate and stoplog handling is either carried out by a rail-
mounted gantry crane or a mobile crane. Gantry cranes can be used to transport
maintenance gates and stoplogs from their storage area to the sluiceway whereas
a mobile crane cannot, as a rule, transport heavy gates or stoplogs. If this is the
case two small rail bogies can be provided.
Different functions are often carried out by the same crane. The crane in Fig.
2.57 can service the operating gate and place the maintenance gate. In other
installations it is used for servicing the operating gate and as a means of placing
stoplogs or an emergency closure gate. When the crane has a cantilever runway
with an auxiliary hoist, it can also raise a removable screen of the type shown in
Fig. 4.2. At freestanding intakes in a reservoir accessed by a bridge, the gantry
crane servicing the operating gate can also place the emergency closure gate, and
in some instances may incorporate screen raking machinery.
Creep speeds on all motions of a gantry crane should be provided to enable
accurate positioning. Creep speeds should be about 1/10 to 1/15 of normal
motion speeds.

62
Summary of types of gate Types of gates

Type Main application Advantages Disadvantages

Gates in open channels
1. Radial gates Sluice No unbalanced forces Extended flume walls
motorised installations Absence of gate slots High concentrated
River control Low hoisting force loads
Spillways Mechanically simple Increased fabrication
Barrages Bearings out of the complexity
water
Can be fitted with
overflow section
Some inspection
possible with gate in
service
2. Radial Sluice No outside source of Wide piers to
automatic installations power required accommodate
gates River control Absence of machinery displacers
Low maintenance Counterbalance
visually intrusive
Can malfunction due
to incorrect design
Can malfunction due
to blockage of inlet or
control system
3. Radial gates Spillways No outside source of Wide piers to
float power required accommodate
operated Absence of machinery counterbalance and
floats
Can malfunction due
to rope system
Can malfunction due
to blockage of inlet or
control system
4. Automatic Limited No outside source of Requires special type
crest gates application at power required of spillway
spillways Absence of machinery Coarse level control
Low maintenance compared with 2.
Limited height of gate
compared with 1. and
2.
5. Automatic Limited No outside source of Limited height of
scour Gates application at low power required opening
level outlets of Absence of machinery Limited upstream
low height dams Low maintenance head
Free discharge only
Coarse control

63
Hydraulic gates and Type Main application Advantages Disadvantages
valves
6. Vertical-lift Sluice Can be fitted with Gate slots required
gates installations overflow sections Load rollers under
River control Short piers water can jam due to
Old installations: Wide span gates can be debris
barrages engineered to provide High hoisting load
spillways good navigation unless
openings counterbalanced
Up and over gates can Overhead support
reduce height of structure visually
supporting structure intrusive
7. Vertical-lift River flow Greater independent Gate slots required
gates control control of overflow and Load rollers under
hook type Barrages where discharge under gate water can jam due to
ebb tide exposes than 6. with a flap debris
gates section for overflow High hoisting load
unless
counterbalanced
Overhead support
structure visually
intrusive unless
operated by hydraulic
cylinders
Seal between the two
moving sections of the
gate can present
problems
8. Free rolling Shut-off of Very reliable No longer
gates turbine intakes Capable of operating at manufactured
differential heads up to Guide slots required
30 m High hoisting load
9. Rolling-weir River flow Can be manufactured Complex to fabricate
gates control very wide Vulnerable side seals
Debris can become
lodged in rack
10. Flap gates Tidal barrages Complete separation of Requires extensive
bottom- Sluice saline and fresh water side staunchings for
hinged installations Overflow to clear debris side sealing or very
River control No visually intrusive accurately constructed
overhead structure pier walls
Can in some cases be Hinge bearings not
engineered to open easily accessible and
under gravity in an permanently
emergency immersed
11. Flap gates Tidal outlets No outside source of Cannot control water
top-hinged power required level
Automatic operation Will not entirely
Simple construction exclude tidal water if
Low maintenance downstream water
level rises above sill
Gate slam can occur

64
 Type Main application Advantages Disadvantages Types of gates
12. Fuse gates Limited Automatic operation Operates as fixed weir
application to Simple construction until it tilts
spillways Easy addition to height Gates have to be
of a dam recovered after tilting
Particularly applicable and may require repair
in an unsophisticated or replacement
environment Use may depend on
frequency of flood
events
13. Rising sector Storm surge Unobstructed Requires piers
gates barriers navigation passage Permits some flow
Not visually intrusive upstream
Can be raised for Complex to fabricate
maintenance and Complex machinery
inspection without
placing stoplogs
14. Buoyant Storm surge Unobstructed Capable of
gates barriers navigation passage withstanding only
bottom- Tidal barriers without piers limited differential head
hinged No structure above Gates move
navigation passage bed independently under
level wave action resulting in
Excellent from visual leakage between gates
considerations Requires detachable
hinges
Gates have to be
interchanged for
maintenance
15. Vertical-lift Storm surge High clearance and High overhead
gates for barriers large span can be structure
navigation achieved Gates are normally
channels No underground counterbalanced
passages required Large gates require
Gravity closure in an lifting chains of the
emergency type where each pin
Conventional hoist can be lubricated
machinery
Can be maintained in
high level position
subject to safety
requirements
16. Vertical-lift Storm surge No underground Lock required
gates for non- barriers passages required Separate overhead
navigation Gravity closure in an structure
channels emergency Gates are normally
Conventional hoist counterbalanced
machinery Large gates require
Can be maintained lifting chains of the
when elevated above type where each pin
waterway can be lubricated

65
Hydraulic gates and Type Main application Advantages Disadvantages
valves
17. Mitre gates Storm surge Unobstructed Width of navigation
barriers navigation passage more limited than
Excellent from visual some other gates
considerations Cannot accept reverse
thrust
Opening has to be
effected when water
levels are nearly equal
Heavy mitre thrust
Cannot close against
high flow
18. Vertically Storm surge Unobstructed Requires wide recess
hinged sector barriers navigation passage in banks to
gates Can be opened or closed accommodate gates on
against flow opening
Can be opened at
differential head
Can be constructed with
wider opening than
mitre gates
Excellent from visual
considerations
19. Drum and Spillways No outside source of Complex gates
sector gates power required Requires extensive
civil engineering
works
Requires zero
downstream level
Control system critical
Can silt up
Not favoured
20. Bear-trap Water control in No outside source of Seals critical
gates logging rivers power required Control system critical
Clears debris Can silt up
Provides unobstructed Rarely used
flow
Gates in submerged outlets
1. Vertical-lift Control and Reliable control gate Gate slots required
intake gates emergency Good load distribution Load rollers or slides
servo-motor closure in the slide version operate under water
operated Damped Requires stem
connections between
servo-motor and gate
Possible cavitation
problems
Slow operation to
raise to maintenance
position
Requires air admission

66
 Type Main application Advantages Disadvantages Types of gates
2. Vertical-lift Bulkhead gate Can be load roller or Cannot be used as a
intake gates slide gate control gate
rope Does not require air Cannot be used as an
operated admission emergency closure
gate
Requires balanced
head for operation
Guide slots required
Possible cavitation
problems
Requires bypass
system
3. Caterpillar or Control and Control gate for very Wide gate slots
coaster gates emergency high heads required
closure Caterpillar train
operates under water
Requires stem
connections between
servo-motor and gate
Cavitation problems
Slow operation to
raise to maintenance
position
Requires air admission
Very costly
4. Radial intake Control and Absence of gate slots Requires chamber to
gates emergency Requires no load rollers retract
closure or slides High concentrated
Intake gate load
Lintel seal critical
Requires dewatering
of tunnel to carry out
maintenance
Requires air admission
5. Cylinder Intake gate Capable of controlling Low natural
gates intake flow and large frequency of vibration
openings due to rope
suspension and low
friction
Possible vibration
problems
Large gates require
counterbalance to
reduce hoisting forces
6. Slide gates Control gates in Reliable control gate or Gate slots required
conduit emergency closure gate Possible cavitation
Back-up gate for Inherently damped due problems
a control gate to sliding friction Requires bonnet for
withdrawal
Requires air admission

67
Hydraulic gates and Type Main application Advantages Disadvantages
valves
7. Jet-flow Control gates in Reliable control gate at Gate slots required
gates conduit for high high heads Requires bonnet for
head application Inherently damped due withdrawal
to sliding friction Requires air admission
Can be circular or
rectangular
8. Radial gates Control gates in Absence of gate slots, Requires chamber to
conduit rollers or slides retract
Requires lower hoist High concentrated
force compared with 6. load
and 7. Lintel seal critical
Requires dewatering
of tunnel to carry out
maintenance
Requires air admission
9. Ring- Back-up gate for Can control flow Large overall height
follower terminal Provides unobstructed approximately three
gates discharge gate or flow times that of fluid way
valve Does not require Requires regular
transition section from flushing
circular to rectangular Drain connection
duct must be provided

References
1. Rouse, H (1964): Engineering hydraulics, Proc. 4th Hydr. Conference, Iowa Institute
of Hydraulic Research, Jun 1949, John Wiley and Sons, Inc.
2. Murphy, T E (1963): Model and prototype observation of gate oscillations, 10th
I.A.H.R. Congress, London, paper 3.1.
3. Lewin, J (1983): Vibration of Hydraulic Gates, Journ. I.W.E.S., 37, 165^179.
4. Thorne, R B (1957): The design, fabrication and erection of radial automatic sluice
gates, Proc. I.C.E., 6th Feb, 126^133.
5. Lewin, J (1984): Radial automatic gates, Proc. 1st Int. Conference Channels and Channel
Control Structures, Southampton, paper 1^195, editor Smith, K V H, Springer
Verlag.
6. Quintela, A C; Pinheiro, A N; Afonso, J R; Cordeiro, M S (2000): Gated spillways
and free flow spillways with long crests, Portuguese dams experience, 20th ICOLD
Congress, Beijing, Q.79^R.12, Vol. IV, 171^189.
7. Townshend, P D (2000): Towards total acceptance of fully automated gates, Dams
2000, Proc. of the Biennial Conference of the BDS, Bath, Jun., editor Tedd, P, Thomas
Telford, 81^94.
8. Randerson, R J (1979): A velocity control structure in the River Orwell, Ipswich,
Journ. I.W.E.S., 38, 135.
9. Petrikat, K (1958): Vibration tests on weirs, bottom outlet gates, lock gates, Water
Power, Feb., Mar., Apr. and May.
10. Naudascher, E (1965): Discussion on Nappe oscillation, Proc. A.S.C.E., Journ.
Hydr. Div., May.
11. Schwartz, H I (1964): Nappe oscillation, Proc. A.S.C.E., Journ. Hydr. Div., HY6,
Nov., paper 4138.

68
12. Partenscky, H W; Swain, A (1971): Theoretical study of flap gate oscillation, 14th Types of gates
I.A.H.R. Congress, Paris, paper B26.
13. Krummet, R (1965): Swingungsverhalten von Verschlussorganen im
Stahlwasserbau, Forschung im Ingenieurwesen, Bd. 31, No. 5.
14. Falvey, H T (1979): Bureau of Reclamation experience with flow induced
vibrations, 19th I.A.H.R. Congress, Karlsruhe, paper C2.
15. Ogihara, K; Ueda, S (1979): Flap gate oscillation, 19th I.A.H.R. Congress,
Karlsruhe, paper C11.
16. Pulpitel, L (1979): Some experiences with curing flap gate vibration, 19th I.A.H.R.
Congress, Karlsruhe, paper C12.
17. Nielson, F M; Pickett, E B (1979): Corps of Engineers experience with flow
induced vibrations, 19th I.A.H.R. Congress, Karlsruhe, paper C3.
18. Ayres, D (1983): The Thames Barrier: the background and basic engineering
requirements, in I. Mech. E. Seminar Proc. The Thames Barrier, 8th Jun.
19. Clark, P J; Tappin, R G (1977): Final design of Thames Barrier gate structures, in
Proc. I.C.E. Conference, Thames Barrier Design, 5th Oct, paper 7.
20. Fairweather, D M S; Kirton, R R H (1977): Operating machinery, in Proc. I.C.E.
Conference, Thames Barrier Design, 5th Oct, paper 8.
21. Hamilton, A J; Prosser, M J (1988): Venice Lagoon flood protection, hydraulic model of
scouring system, B.H.R.A. report RR 2918.
22. Lewin, J; Scotti, A (1990): The flood prevention scheme of Venice: experimental
module, Journ. Inst. Water Environmental Management, 4, 1, Feb.
23. Ieperen, A van (1994): Design of the New Waterway Storm Surge Barrier in The
Netherlands, Hydropower and Dams, May, pp.66^72.
24. Janssen, J P F M; Jorissen, R E; Ieperen, A van; Kouvenhoven, B J; Nederend, J
M; Pruijssers, A F; Ridder, H A J (1994): The Design and Construction of the New
Waterway Storm Surge Barrier in The Netherlands, 18th ICOLD Congress,
Durham, C.15 pp.877^900.
25. Severn Barrage Committee (1981): Tidal power from the Severn Estuary, Vol 1, Energy,
paper No. 46, HMSO.
26. Ball, J W (1959): Cavitation and vibration studies for a cylinder gate designed for
high heads, 8th I.A.H.R. Congress, Montreal, paper 9A.
27. Bixio, V; Cola, R; Garbin, C; Mariani, M (1985): On the hydraulic behaviour of a
cylinder gate in a vertical intake with radial symmetry openings, 2nd Int. Conference on
the Hydraulics of Flood and Flood Control, Cambridge, paper D3.

69
 3
Valves

The flow in pipelines is controlled by valves. This chapter is mainly concerned with
large valves carrying out the function of terminal discharge and providing back-up
in circular conduits. The exceptions are pressure-reducing valves in pipelines (an
example of these is shown in Fig. 3.16), and the needle valve which can be used for
regulating flow in pipelines or as a terminal discharge valve. It is largely superseded
as a discharge valve.The main applications, advantages and disadvantages of the
various types of valve are summarised in a table at the end of this chapter.
Amongst terminal discharge valves, the hollow-cone valve predominates
because of its good energy dissipation characteristics, simplicity of
construction, lower cost and favourable coefficient of discharge. It is also least
prone to blockage.
Compared with the external control sleeve of the hollow-cone valve, hollow-
jet valves and needle valves with their internal moving parts are mechanically
more complex and therefore more difficult to service. The fluid passages are
more restricted and liable to blockage.
Trashracks and screens are essential at intakes which serve conduits
containing valves, and may only be omitted when large hollow-cone valves
are used as these valves are less prone to blockage.

Sluice valves
Sluice valves are the most frequently used control devices in pipelines. In the
open position they provide an unobstructed fluid passage. In spite of their non-
linear flow control characteristics they are often used for control of low velocity
flow. Because the valve blade is unsupported during raising and lowering, and
due to eddy shedding from the blade tip, they are only suitable for closure and
opening against flow at low velocities.

Butterfly valves
The butterfly valve is the most frequently used closure device in pressure
conduits because of its relatively compact arrangement and simple construction.
In the open position the blade lies in the plane of flow of the fluid. Valves are
manufactured in sizes up to 4 m diameter and are able to withstand operating
heads up to 200 m.

71
 Hydraulic gates and There are two types of butterfly valve: the solid-disc valve, sometimes
valves referred to as lenticular (Fig. 3.1(a)), and the lattice-blade valve, also described
as a through-flow valve (Fig. 3.1(b)). The latter offers the advantages of a stiffer
disc assembly and lower loss coefficient.1
Valve blades are normally manufactured in cast iron or carbon steel. Other
more corrosion resistant materials such as high nickel cast iron, stainless steel or
aluminium bronze are used as the material for the blade where the water carries
bed material or is aggressive to cast iron or steel. Valve bodies and blades have
been coated for special applications with epoxy resin or ebonite.
Figure 3.2 shows different arrangements of seals for butterfly valves. Seal (a),
an all metal seal, is suitable only for low operating heads. At higher pressures an
elastomeric seal (b) of Neoprene or Nitrile is used, and the seal is pressurised by
the upstream water head. A diaphragm seal is shown in (c). This seal is also
pressurised by the upstream water.
The seal seat for (a) and (b) is of stainless steel and is welded into the
housing. It is arranged flush, but for clarity is shown in the figure as
projecting.

Figure 3.1. Solid disc and

through-flow butterfly valves

72
 Valves

Figure 3.2. Different

arrangements of seals for
butterfly valves

In the closed position the disc of the versions shown in (a) and (b) is inclined
at approximately 80 to the conduit axis. This applies also to lattice-blade valves.
The latter can be provided with a second seal on the upstream side which can be
normally or hydraulically operated, allowing replacement of the downstream
seal while the conduit remains under pressure.
Butterfly valves are normally opened under balanced conditions and closed
against flow. They are not generally suitable for flow control, only as on/off
devices, because of flutter of the blades and eddy shedding from the blade tips.
Prior to opening of the valve, the pressure is balanced by means of a bypass pipe
which incorporates shut-off valves. At high heads, manual, electrical or oil
pressure actuated filling nozzles are used. A guard valve is arranged upstream
of the filling nozzle. Means of venting the conduit downstream of the valve
during the filling operation must be provided together with facilities for
draining.
Closure of valves controlling conduits or penstocks is usually by gravity (see
Fig. 3.3(a)). During normal operation the lever arm for the falling weight is locked
in the valve open position. The locking mechanism is designed to release the arm

73
 Hydraulic gates and
valves

Figure 3.3. Operation of

butterfly valve controlling
conduits or penstocks

which is then able to rotate to the shut position. The release mechanism may be
triggered manually, electrically or indirectly by excessive velocity of flow in the
conduit or by loss of pressure. The valve is opened by the oil hydraulic servo-
motor which also controls the rate of closure. Shortly before the end of the closure
movement the discharge of oil from the annulus side of the cylinder is throttled to
ensure slow final closure of the valve. The closing time must be controlled so as to
minimise the water hammer in the penstock. This can also be achieved by
cushioning the last part of the movement of the piston. Ellis and Mualla2 have
analysed the closure characteristics of butterfly valves.
The servo-motor in Fig. 3.3(b) is double acting. Opening of the valve is
effected by oil or an emulsion supplied to the piston side of the cylinder from

74
a hydraulic power pack, while the fluid on the annulus side is ported to Valves
discharge. For closure of the valve, mains pressure is admitted from upstream
of the butterfly valve to the annulus side of the cylinder, while the oil is ported to
the tank of the power pack. The water used as a closing medium is filtered before
it reaches the control elements and the servo-motor.
In the event of emergency closure, cavitation will occur where the conditions
of initial operation with positive back pressure change into a situation of
complete separation of flow. The time is usually short enough not to cause
any damage.
The loss coefficient for fully open butterfly valves is shown in Fig. 3.4. The
loss coefficient of a valve Kv is defined as:

H
Kv
V 2 =2g

where
H total head loss

V velocity of flow in the valve
g gravitational constant

The loss coefficients for partially open valves are shown in Fig. 3.5.
Variations of more than 10% occur, particularly when the valve is nearly closed
and when the valve seating arrangement becomes very important.3

Cavitation in valves
Cavitation is caused by the local pressure on the downstream side of a valve, by
the accelerated flow of the water as it contracts to pass through the valve
opening, and by the generation of turbulence. Eddies are formed in the intense
shear layer which surrounds the accelerated flow of water through the valve

Figure 3.4. Loss coefficients

for fully open butterfly valves
(after Miller3)

75
 Hydraulic gates and
valves

Figure 3.5. Loss coefficients

for partially open butterfly
valves (after Miller3)

opening. When the pressure inside the eddies, which is considerably less than
the fluid pressure in the penstock, approaches the vapour pressure, cavitation
bubbles grow from nuclei suspended in the water. As the ambient pressure
increases and the eddies degenerate due to viscous forces, the bubbles become
unstable and collapse. If this occurs next to a solid boundary, it creates noise,
vibration and in its intense form erosion damage. The cavitation index can be
used to express the level of cavitation as the ratio of forces suppressing or
preventing cavitation to the forces causing cavitation:

hd hv
cavitation index 
hu hd

hu hv
or
hu hd

76
 hu hv Valves
or
v2 =2g

where hd head downstream of the valve

hv vapour head at the inlet temperature
hu head upstream of the valve
v average pipe velocity

It should be noted that consistent pressures, either absolute or gauge, must

be used. If gauge pressures are used the vapour head has a negative value.
The higher the cavitation index, the less likely cavitation damage will occur.
Chapter 9 discusses the different intensities of cavitation ^ incipient, critical and
choking. The initial stage, incipient, which consists of light intermittent bursts
of noise, can increase to choking cavitation when intense noise and severe
damage occur. Rahmeyer4 carried out measurements of cavitation intensity in
valves, and the Instrument Society of America5 lays down a test procedure.
Miller3 gives a graph of incipient, critical and choking cavitation for v and Kv
which is reproduced in Fig. 3.6.
The base conditions for the graph are a valve diameter of 0.31 m and an
upstream head minus vapour head of 50 m. To correct these velocities to other
valve sizes and heads the following equation should be used:
 0:39
hu hv
Vir or Vcr C1 Vr
50

where Vir incipient cavitation

Vcr critical cavitation
C1 a correction factor
Vr velocity Vir from Fig. 3.6 for incipient cavitation or the
reference velocity Vcr from Fig. 3.6 for critical cavi-
tation

C1 is taken from Fig. 3.7.

Table 3.1. Vapour pressure of pure water

Vapour pressure
2
Temperature C N/m Head of water: mm
0 610 62
5 870 89
10 1230 125
15 1700 174
20 2330 238
25 3160 323
30 4230 433

77
 Hydraulic gates and
valves

Figure 3.6. Cavitation

velocities for butterfly valves
(after Miller3)

Hollow-cone valves and hoods

This type of valve is more commonly known by the surnames of the inventors,
Howell and Bunger. It is widely used as a regulating valve for free discharge
because of its simplicity. Figure 3.8 shows a typical cross-section through a
hollow-cone valve.
As illustrated, the upper section is the closed position and the lower one the
fully open position. The valve body is cylindrical, flanged at the upstream end
for attachment to the pipeline and connected to a downstream dispersing cone
by streamlined radial ribs forming an annular outlet port. Flow control is
effected by a reinforced stainless steel cylindrical gate which slides over
gunmetal bearing strips, secured to the body, to close the annular port and to
seal against a seat ring attached to the dispersing cone. The dispersing cone
forms the discharging jet of water into a hollow divergent cone in which the
energy of the jet is dissipated by air friction and entrainment.
The discharge of the valve is given by:

78
 Valves

Figure 3.7. Correction factors

for valve size

p
Q CdA 2gH 3:1

where Q discharge
Cd discharge coefficient (approximately 0.85)
A area of the valve based on the inside diameter of the valve
body
g gravitational constant
H net head at the valve entry

Westinghouse quote a Cd value of 0.85 for their valves. A study of a 2.5 m

diameter valve carried out by Boving & Co. gave a value of 0.83.
Hollow-cone valves are manufactured in sizes up to 3.5 m diameter with
operating heads up to 250 m.
Two major types of hydrodynamic problem have been experienced with
hollow-cone valves: vane failure, and shifting of the point of flow attachment.

Vane failure
This has been attributed to a number of causes but the most likely one is
hydroelastic instability causing vibration normal to the chord of the valve and
twisting about the longitudinal axis. Destructive resonance occurs at a critical
velocity at which the flow couples the two forms of vibration in such a way as to

79
 Hydraulic gates and
valves

Figure 3.8. Hollow-cone valve

feed energy into the elastic system. Possible modes of vibration for a hollow-
cone valve are shown in Fig. 3.9.
Mercer6 has suggested a parametric value incorporating a coefficient
depending on the ratio of shell-to-vane thickness and number of vanes. Valves
with a value less than 0.115 have operated successfully and valves with a value
greater than 0.130 have failed. Mercer's parametric value is:

Figure 3.9. Possible vibration

modes of hollow-cone valves
(after Mercer6)

80
Table 3.2 Values of C in equation (3.2) Valves

N 4 5 6 6 6 6 6
Ts/Tv 1.00 1.00 0.50 0.90 1.00 1.20 2.00
C 2.22 2.35 1.98 2.40 2.48 2.53 2.75

Q=CDTv
p 3:2
Eg=e

where Q discharge
C a dimensionless coefficient depending on K and N
D valve diameter
Tv vane thickness
E Young's modulus
g gravitational constant
e mass per unit volume (of the material of the valve)
K ratio of shell thickness to vane thickness Ts/Tv
N number of vanes

Mercer's investigation and tabulation of values was carried out in Imperial

units, therefore consistent Imperial units must be used for equation (3.2). Values
of C in equation (3.2) are given in Table 3.2.
The frequency of vibration f (Fig. 3.9) can be expressed by the equation:
r   
Eg . Tv
f C 3:3
e D2

Inserting the value of C in equation (3.3) shows that six-vane valves have 10%
higher frequencies than comparable four-vane valves, and that the thickness
of the shell relative to the vane does not have too great a bearing on the
frequency.
Nielson and Pickett7 reported a major vane failure of a 2740 mm diameter
hollow-cone valve. The failure was of the fatigue type. Mercer's parametric
value of the valve as originally constructed was 0.176.
Falvey8 cited severe vibration of two 2135 mm hollow-cone valves. The
observed 85 Hz frequency correlated well with estimates of its natural vibration
frequency based on the paper by Mercer. The valve opening in the prototype
had to be restricted to a maximum of 80%.

Shifting of the point of flow attachment

As the hollow-cone valve is opened the flow control may shift from the sleeve to
the valve body (see Fig. 3.10) and intermittent attachment and reattachment
may occur, resulting in severe vibration.7 Under these conditions the opening
of the valve, that is the sleeve travel, has to be limited.
There have been a few instances of vibration of hollow-cone valves due to
excessive length of the sliding sleeve.

81
 Hydraulic gates and
valves

Figure 3.10. Vibration of

hollow-cone valve due to the
shifting of the point of flow
attachment, oscillating between
A and B (after Nielson &
Pickett7)

Deterioration of the seal between the valve body and the sliding sleeve can
result in leakage. In general this does not lead to vibration of the valve but if it
continues for a long time it can result in erosion.
The expanding cone shaped discharge pattern of the hollow cone is very
effective in aerating the water and dispersing the energy. Because of these
features, stilling basins are not normally used for the discharge of hollow-
cone valves. The usual angle of the cone is 45. Experimental investigations
involving valves with cones of different angles9 have been carried out.
Where hollow-cone valves are located in a tunnel or where the spray
from a widely dispersed jet is not acceptable, a hood is used to confine
and redirect the discharge (Fig. 3.11). In tunnels or conduits the hood
prevents erosion by the impinging jets and ensures that air is admitted from
upstream. Guidelines as to the optimum geometry of hoods are given by
Brighouse and Chang.10,11 The hood has to be arranged to minimise
splashback through the upstream opening and ribs are introduced on the
inside of the discharge section of the hood so that air is admitted to the
inside of the discharge jet.
Hollow-cone valves can be installed to discharge into a stilling basin or
submerged as shown in Fig. 3.12. In the former case the main purpose of
the valve, to act as an energy dissipator, is limited because the jet is

Figure 3.11. Installation of a

hollow-cone valve with hood

82
 Valves

Figure 3.12. Installation of a submerged

hollow-cone valve and valve discharging
vertically into a stilling basin

shortened reducing its ability to entrain air. When submerged within a

stilling basin they are not energy dissipating devices and this function is
carried out by sheared flow and air which is introduced into the stilling
basin. A model study is required for the successful installation of
submerged hollow-cone valves. Figure 3.12 illustrates such an application.
Flotsam can become wedged across the vanes of a hollow-cone valve but is
usually only a problem in smaller valves.

83
 Hydraulic gates and Hollow-jet valves
valves
Figure 3.13 illustrates the construction of a hollow-jet valve. Movement of
the cone controls the area of the discharge orifice. It is used as a terminal
discharge and control valve. The jet is compact and therefore entrains less
air than the hollow-cone valve. It can be installed directly after a bend in the
pipework.
The hollow-jet valve is frequently installed so that it discharges at an angle of
30 into a stilling basin. The flow in the conduit past the movable cone and the
body results in a hollow jet, which initially maintains its shape and flares out
shortly before the point of impingement. A jet angle relative to the horizontal
loosens up the jet structure and reduces the intensity of impingement.
Inspection and servicing of the mechanical or the oil hydraulic actuator
requires removal of the valve in its entirety. The oil hydraulic cylinder of the
valve, Fig. 3.13(b), is operated by an external hydraulic power pack with
pipework which has to be routed through the jet.

Figure 3.13. Hollow-jet valve

84
 Valves

Figure 3.14. Installation of

hollow-jet valves

The coefficient of discharge of a hollow-jet valve is about 0.7 at full valve

opening, reducing to 0.4 at half opening and 0.23 at quarter opening.
Hollow-jet valves are manufactured in a similar range of sizes and heads to
hollow-cone valves. Figure 3.14 shows the installation of hollow-jet valves.

Needle valves
Needle valves are used for regulating flow, either as terminal discharge valves or
for controlling high head flow in pipes. Their use in outlet works has been
supplanted in many applications by more economical and hydraulically efficient
valves,suchasthe hollow-cone valve.Figure3.15showsaninstallation andcross-
section through an interior-differential needle valve. The valve is closed by
admitting water pressure to chamber B and connecting chamber A to drain
through the spool valve located at the bottom of the needle valve. To open the
valve, water pressure is admitted to chamber A and chamber B is opened to drain.
To prevent cavitation the discharge opening is arranged so that the
downstream cone angle of the needle is slightly less than the downstream cone
angle of the body. A sharp flow separation point at the body seat is another
requirement if cavitation is to be avoided. The coefficient of discharge for use
in equation (3.1) is about 0.6 at full valve opening. This reduces to 0.4 at half
opening and 0.26 at quarter opening. Because of the low coefficient of discharge
at partial openings, the valve can dissipate energy when controlling high head
flow in pipes. Needle valves are manufactured in sizes up to 2 m and for working
heads up to 200 m.

Pressure-reducing valves
Figure 3.16 illustrates a pressure-reducing valve. The energy dissipation is
effected by discharging some or all of the flow through the orifices of the

85
 Hydraulic gates and
valves

Figure 3.15. Interior-

differential needle valve

perforated cylinder. The cylinder is arranged on a plunger which advances or

retracts the throttling cylinder. The drive can be manual or can be powered by
an electric actuator driving the plunger via a bevel gear. The perforations break
up the flow into numerous concentric individual jets directed against one
another.

86
 Valves

Figure 3.16. Pressure-

reducing valve

This type of valve is suitable only as a regulating valve in closed pipe systems.
It is also used as a bypass valve. They are manufactured in sizes up to 1.5 m
diameter.

Sphere valves
Figure 3.17 shows a sphere valve, sometimes referred to as a rotary valve. The
section shows the lower half of the valve fully open and the upper half fully
closed.
Sphere valves have a clear bore and when fully open have a very low loss
coefficient. Resilient rubber seals are used for valves working at pressures up
to 400 m head. Metal seals are used for higher heads. Valves are normally
supplied with an operating seal at the downstream end and with an additional
maintenance seal on the upstream end which is either operated manually or
hydraulically. Closure is droptight. The application of sphere valves is for
shut-off control on the pressure side of high head turbines and pumps. The

87
 Hydraulic gates and
valves

Figure 3.17. Sphere valve

usual arrangement is gravity closure of the valve and oil hydraulic piston
operation for opening. Alternative operation by servo-motor has been used
with opening effected by hydraulic oil acting on the piston side of the cylinder
and uncontrolled water operating on the annulus side. On opening, the
hydraulic oil on the piston side overcomes the force on the water side displacing
the uncontrolled water from the cylinder. When closure is initiated, or there is a
failure of the oil supply, the valve is closed by the uncontrolled water pressure.
Sphere valves are manufactured in sizes up to 3.5 m diameter for use at
working pressures up to 500 m. Smaller size valves up to 2 m diameter are
available up to working pressures of 1000 m head.

88
Matching terminal discharge valves and guard Valves
valves
Figure 3.18 shows a typical arrangement of bottom outlet valves where a
hollow-cone valve is backed by a butterfly valve. The diameters of the two
valves have to be sized so that the hollow-cone valve is smaller than the butterfly
valve. The change in the diameter of the conduit is effected by a taper section
between the valves. This produces a back pressure and prevents cavitation in the
pipe and the butterfly valve when the hollow-cone valve is in the fully open
position. In order to calculate the difference in the size of the valves, the loss
coefficient of the hollow-cone valve between 70% to fully open must be known.
As a very approximate guide, the area of the terminal section should be about
60% of that of the section where the butterfly valve is located.
Operationally the butterfly valve is used only for emergency closure,
otherwise the valve is actuated under balanced conditions. To facilitate this,
water under reservoir head is admitted to both sections of pipe upstream and

Figure 3.18. Arrangement of bottom outlet valves. Butterfly valve and hollow-cone valve

89
Hydraulic gates and downstream of the butterfly valve. If the section of pipe upstream of the valve
valves remains under pressure, admission of water under reservoir head is required
only to the downstream side. Apart from means of draining the sections of pipe,
provision for releasing air must be made. The closure characteristic of the
butterfly valve must be designed to minimise water hammer.
The matching of terminal discharge and guard valves may also be appropriate
for other valve combinations apart from butterfly and hollow-cone valves. It is
likely to be more critical when a hollow-cone valve is used for discharge because
of the low loss characteristic of a fully open hollow-cone valve.

Summary of types of valve

Type Main application Advantages Disadvantages
1. Sluice valves Controlling flow Low cost Unsupported valve
at low velocities Simple blade during raising
Closure and Reliable and lowering
opening of flow Eddy shedding from
blade tip

2. Butterfly valves Closure device in Relatively low Normally opened

pressure conduit loss coefficient under balanced
Available in large conditions
sizes Possibility of blade
Capable of flutter
working at high Possibility of eddy
heads shedding from blade
Closure by tips
gravity can be
arranged

3. Hollow-cone Terminal Very efficient Seal of sliding sleeve

valves discharge energy may leak
(Howell^Bunger dissipation Can trap debris but
valves) Simple much less than 4., 5.
construction and 6.
Relatively low
cost
Can be operated
electro-
mechanically or
by oil hydraulics
Good discharge
coefficient
Available in large
sizes
Least flow
obstruction of
any terminal
discharge valve

90
 Type Main application Advantages Disadvantages Valves
4. Hollow-jet Terminal Dissipates energy Less efficient energy
valves discharge Can be arranged dissipator than 4.
to discharge into Lower coefficient of
a stilling basin at discharge than 4.
an angle Greater cost than 4.
Fluid passages can
become blocked
Internal moving parts
Inspection and
servicing requires
removal of valve

5. Needle valves Terminal Dissipates energy Less efficient energy

discharge Can be used as an dissipator than 4.
in-line pressure Low coefficient of
reducing valve discharge
Greater cost than 4.
Fluid passages can
become blocked
Internal moving parts
Inspection and
servicing requires
removal of valve

6. Pressure- Pressure control Pressure control Orifices in perforated

reducing valves in closed pipes cylinder can be
(perforated blocked by debris
cylinder type) Internal moving parts
Inspection and
servicing requires
removal of valve

7. Sphere valves Shut-off control Low loss Greater cost than 2.

(rotary valves) in high pressure coefficient
conduits Shuts droptight
Manufactured in
large sizes
Capable of
working at high
heads
Can be supplied
with a
maintenance seal

91
Hydraulic gates and References
valves
1. Bramham H T (1979): Developments in through flow butterfly valves. Water Power
and Dam Construction, Mar.
2. Ellis J; Mualla, W (1984): Dynamic behaviour of safety butterfly valves. Water
Power and Dam Construction, Apr., pp. 26^81
3. Miller D S (1978): Internal flow systems, B.H.R.A., Fluid Engineering.
4. Rahmeyer W J (1981): Cavitation limits for valves, Journ. A.W.W.A., Nov, pp.
582^584.
5. Instrument Society of America: Control valve capacity test procedure, ISA^S39.2,
Pittsburgh, Pa.
6. Mercer A G (1970): Vane failures of hollow-cone valves, I.A.H.R. Symposium,
Stockholm, paper G4.
7. Nielson F M; Pickett E B (1979): Corps of Engineers experiences with flow-
induced vibrations, 19th I.A.H.R. Congress, Karlsruhe, paper C3.
8. Falvey H T (1979): Bureau of Reclamation experience with flow-induced
vibrations, 19th I.A.H.R. Congress, Karlsruhe, paper C2.
9. Rao P V; Patel G G (1985): Hydraulic characteristics of cone valves with different
angles, Irrigation and Power, Jul., pp. 233^243.
10. Brighouse B A; Chang E (1982): Design data on deflector hoods for hollow-cone outlet valves,
B.H.R.A., report 1939, Dec.
11. Brighouse B A; Chang E (1982): Monasavu hydroelectric scheme, Fiji, Part 2, model study
of Howell^Bunger valve for the controlled filling outlet, B.H.R.A., report RR1818, Mar.

92
 4
Trashracks, screens and
debris

Trashracks and screens in submerged intakes

The terms trashrack and screen are synonymous, the former being favoured in
the USA and the latter in the UK. They are provided at intakes to turbine or
power penstocks, at pumping stations, to prevent debris in irrigation canals and
where valves liable to blockage are installed. They are either of the mobile type,
withdrawable to the surface for cleaning, or form a fixed installation cleaned by
raking from an overhead gantry.
Figure 4.1 shows a screen mounted on a rail wheel bogey protecting an
abutment intake on a rock fill dam. This installation is at the Kotmale Dam in
Sri Lanka. A similar arrangement was used at the Mangla Dam in Pakistan. The
screen carriage moves on rails and is hoisted clear of the water for cleaning. The
semicircular shape of the screen increases the free area.

Figure 4.1. Screen protecting

an abutment intake, hoisted to
surface for cleaning

93
 Hydraulic gates and Fixed screen installations are used principally to protect intakes at concrete
valves dams. If it is necessary to provide a hood over the intake to suppress vorticity a
rectangular, withdrawable screen is used, Fig. 4.2. However, problems can arise
if a long section of debris becomes stuck in the screen and the screen cannot be
withdrawn.
At least four criteria must be considered when designing screens:
Differential design head, that is, head loss across the screen plus head loss due
to the accumulation of debris. In submerged screens a differential head of 3 m
is often used.
Spacing of screen bars, which depends on the capacity of turbines or pumps
to pass solid objects. A frequent spacing is 100 mm, although 75 mm or less is
sometimes used.

Figure 4.2. Removable screen

installation

94
 Head loss across the screen. Trashracks, screens
Vibration. and debris

Failure of screens due to vibration of the screen bars has been recorded.1^3
Vibration depends on the natural frequency of the screen bars, the forcing
frequency and the potential development of resonance.4 Vibration occurs when
the two frequencies approach resonance.
The natural frequency of oscillation of screen bars in water is given by:
r 
 EIg
fn
2 m mw L3

where fn natural frequency

 a coefficient depending on the end fixity of the bar (bars
are normally welded to the supporting grid frame;  is
between 16 and 20 for bars 60^70 mm deep having a
thickness-to-depth ratio of 5:1)
E Young's modulus
I moment of inertia of screen bar
g gravitational constant
m mass of screen bar
mw added mass of water; that is the mass of water vibrating
with the bar
L length of bar between supports

From Levin1,2 mw can be approximated:

mw equivalent mass of water of the same volume as the screen bar  b=d
where b effective spacing between adjacent bars, and d thickness of the bars

m b
or mw 
8 d

Work done by Levin2 suggests that the value of the b term be limited to 0.55
times bar depth for a bar with a depth to thickness ratio 10, and to 1.0 times bar
depth for a bar with depth to thickness ratio 5. In practice, the effective spacing
between bars will be greater than the bar depth but the computed value of b should
be based on the suggested relationship, which will yield conservative results.
The forcing frequency due to vortex shedding at the downstream edge of the
screen bar is given by:

ff SV=d

where ff forcing frequency

S Strouhal number
V approach velocity
d thickness of screen bar

95
Hydraulic gates and
valves
The Strouhal number depends on the spacing between bars and the shape of
the bars. Levin1,2 gives detailed information. For most design purposes the limit
value of the Strouhal number applies when the bar spacing to bar thickness
number is 5 or greater. For a bar fully rounded upstream and downstream the
limit value of the Strouhal number is 0.265, and for a screen bar with sharp
rectangular profile the number is 0.155.
Selecting the velocity is complicated by the difference between the flow
across the bars when the screen has been cleaned and when it is partially blinded
by debris. It is suggested that a range of values be used: at the lower end the net
velocity through the bars and at the upper end a value three times greater.
Although screens are not normally operated so that debris is permitted to
accumulate to this extent, the local velocity of a partially blinded screen may
significantly exceed the average velocity. It is recommended that the natural
frequency of oscillation of the screen bars should differ by a factor of 2^2.5 times
from the forcing frequency.
In the diagram of the velocity profile at La Plate Taille,3 the variation of flow
velocities about the mean value appears to lie between 0.5 and 1.8.

Trashracks and screens in culverts and river

courses
Most screen installations are at the entry to a culvert to ensure that any blockage
occurs outside the culvert and to ease removal of debris. A secondary purpose is to
prevent unauthorised entry into the culvert, and for this reason a safety screen is
usually also provided downstream of a culvert. Magenis5 reports an
extraordinarily high proportion of local councils (91%) and water authorities
(95%) in the UK who have experienced serious problems with screens. Structural
failures were significant: 14% for local councils and 18% for water authorities.
The survey highlighted the failure of many screens to satisfy basic criteria:
To pass the maximum flow when partially blocked to match the capacity of
the culvert it protects.
To allow safe clearance of debris under normal and adverse conditions.
To prevent all debris which would cause a blockage from entering the
culvert.
To remain structurally sound under all conditions.
To satisfy the first criterion and take into account the possibility of a screen
becoming completely blocked, a bypass should be provided. Accommodating a
bypass may, however, present physical problems in urban river courses.

Screen instrumentation
Instrumentation for submerged inlets
It is a usual requirement that the head loss across the screen be measured to
indicate when the screen has to be cleaned. At inlet depths of 30 m or more the
most suitable instrument is a bubbler device, because it can measure differential
head by means of a sensitive bridge.

96
 Pressure transducers located at depth at the approach and downstream of a Trashracks, screens
screen should be restricted to measuring a limited range of head. The accuracy of and debris
the instrument is increased if the range of operational head is limited,
irrespective of depth of installation. For example, if pressure transducers with
an accuracy of 0.75% are installed at 30 m depth and measure total head, the
reading can be 0.45 m in error.
Bubbler devices are considerably more expensive to install than pressure
transducers and require maintenance. To overcome this, at least one installation
uses tubes which are brought to the surface and the water level is measured by
tank gauges. The output from the tank gauges is compared and the difference is
displayed.

Instrumentation for screens in free surface water

Pressure transducers or ultrasonic level measuring instruments are used, one
positioned upstream and another downstream of the screen. The instruments
should be located in a stilling well. The output from the instruments can be
displayed separately or integrated to show differential head. It is common
practice for staged warnings to be sounded as the head loss across the screen
builds up.
A difficulty sometimes experienced in small river courses is the rapid build-
up of head loss which can be caused by a single large object such as a tree.

Screen raking
A variety of screen raking machinery for fixed screens is available. Figure 4.3
illustrates two types.
Screens which are hoisted to the surface such as that shown shown in Fig. 4.1
have to be raked by hand using special purpose combs. An auxiliary crane is
provided at the screen's cleaning platform to handle logs or trees which have
been trapped by the screen bars.

Debris
In underflow gates, debris will not normally be discharged until a gate is
70^80% open. Under conditions of drowned discharge, debris becomes
trapped in the hydraulic jump which forms in the stilling basin, and may
recirculate for an appreciable time. Floating oil cans or other metal containers
which repeatedly impact the submerged sections of gate arms or structural
stiffeners of the skinplate can be a noise nuisance at gate installations close
to dwellings. Floating debris can also cause damage to gate equipment and
to paintwork on gates.
At overflow gates or overflow sections of gates, debris and flotsam will be
discharged from upstream but can build up at flow breakers where these have
been provided to vent the nappe, or become trapped by the discharge roller
which forms downstream of the gate (see Fig. 2.27).
In bottom-hinged flap gates which recess into the river bed, floating timber
trapped upstream of the gate discharge roller can cause operational problems.

97
Hydraulic gates and
valves

Figure 4.3. Screen raking machinery

98
 At flood diversion channels which are controlled by gates, and which Trashracks, screens
incorporate a fixed weir alongside, a boom may be fixed across the flood channel and debris
to divert floating debris to the weir.
A floating boom is often placed at the exit of a stilling basin to prevent debris
discharged over the weir from refluxing and entering the stilling basin when the
flood relief gates are shut. It can then remain trapped in the stilling basin when
discharge under the gates commences.
Floating booms for diverting debris are positioned across a waterway at an
angle of 30^45 to assist in driving flotsam towards the bank for clearance.
Slack must be provided in the stringer cable to allow for a rise in water level
during a flood. Most designs of floating booms are only partially successful.
In rivers which carry a large amount of flotsam during the flood season, an
appreciable load can accumulate on the structural stiffeners of gates. To prevent
this the rear of the gate skin plate is sometimes protected by wire mesh, but some
debris will still penetrate the mesh and becomes more difficult to remove.
Another solution is to design horizontal stiffening members of the skin plate
assembly and horizontal tie members of gate arms as closed sections, adding a
fairing section where, due to the angle of the upper face of the stiffening
member, a ledge is still formed for debris to collect.

References
1. Levin, L (1967): Proble mes de Perte de Charge et de Stabilite des Grilles de Prise
d'Eau, La Huille Blanche, 22, No. 3, pp. 271^278.
2. Levin, L (1967): Etude Hydraulique des Grilles de Prise d'Eau, Proc. 7th Gen.
Meeting I.A.H.R., Lisbon, 1, p. C11.
3. Vanbellingen, R; Lejeune, A; Marchal, J; Poels, M; Salhoul, M (1982): Vibration of
Screen at La Plate Taille Hydro Storage Power Station in Belgium, Int. Conference on
Flow Induced Vibrations in Fluid Engineering, Reading, England, Sept., p. B2.
4. Sell, L E (1971): Hydroelectric Power Plant Trashrack Design, Proc. A.S.C.E., J
Power Div., Vol. 97, Jan, No. PO1.
5. Magenis, S E (1988): Trash Screens in Urban Areas, I.W.E.M., River Engineering
Conference, Jan.

99
 5
Structural considerations

Design criteria
The design of gates can be analysed by conventional two-dimensional (2D)
structural analysis. This can also be applied to radial gates if the curvature of
the skin plate is ignored. If it is considered necessary to take curvature into
account, a finite element three-dimensional (3D) analysis is required. Existing
gate design standards require that designers consider the limit states at which
gates would become unfit for their intended use, by applying appropriate
factors for the ultimate limit state and the serviceability limit state.
The extent and magnitude of load factors applied is usually varied for
different design elements, operating conditions and the possible occurrence
of extreme events which cannot be reliably quantified. Examples are ship or
vessel impact, increases in weight due to entrained water or floating debris,
exceptional tide levels, jammed foreign bodies, increased loads due to
impeded movement caused by solid freezing, ice impact and possibly
irregular settlement and deformation of the foundation works. Some
designers or specification writers stipulate that a corrosion allowance be
added after calculations determining the sizes of members and plates have
been carried out. A reduction in working stresses or an increase in the load
factors is a more rational approach.
Detailed structural design standards for hydraulic gates are current in
Germany and the USA. The German standards are DIN 19704, Hydraulic Steel
Structures, Part: 1, Criteria for design and calculation, 19981 and DIN
Handbook 179, Water Control Structures 1, 1998.2 (At the time of preparation
of this edition, the German standards were not available in English. A previous
version of DIN 19704: Sept 1976 is available in an English translation.) The
USA standard is the US Army Corps of Engineers manual, Design of spillway
tainter gates, January 2000.3 The designation `tainter gate' is used in all US
references to radial gates. The design basis in this specification, while similar
to limit state design, is load and resistance factor design (LRFD) which can be
expressed mathematically as:

 i Qni  Rn

101
 Hydraulic gates and where i load factors that account for variability in the
valves corresponding loads
Qni nominal load effects
 reliability factors
 resistance factor that reflects the uncertainty in the
resistance for the particular limit state and, in a relative
sense, the consequence of attaining the limit state
Rn nominal resistance
The load factors used in the Corps of Engineers Manual correspond to the
load factors used in BS 5950: Part 1: 1990 Structural use of Steelwork in
Buildings, when these are interpreted to apply to the load conditions
encountered in hydraulic gates. Table 5.1 lists the load factors.
Calculations of hydrostatic loads on a radial gate for the water pressure on a
closed gate are set out in the Appendix to this book. Approximate calculations
are also shown for a gate in the open position when the downstream water level
is above the gate lip and the discharge is drowned. The third set of calculations is
for the gate load due to water pressure under overflow conditions.

Table 5.1. Load factors for the design of gates (after US Corps of Engineers' Manual, Design of Gates,
January 2000)

Loading Factor i
or f
1. Gravity loads including dead load or weight of gate 1.2
2. Mud weight or debris 1.6
3. Ice weight 1.6
4. Maximum net hydrostatic load that will ever occur 1.4
5. Design hydrostatic load ^ maximum net hydrostatic load of any flood 1.4
up to a 10 year event
6. Normal hydrostatic load ^ temporal average net load due to upstream 1.2
and downstream water levels (water levels which are exceeded up to
50% of the time during the year)
7. Machinery load where the machinery exerts applied forces on an 1.2
otherwise supported gate
8. Maximum compressive downward load that a hydraulic hoist system 1.2
can exert if the gate is jammed while closing, or when it comes to rest
on the sill
9. Hydraulic cylinder at rest ^ downward load while the gate is on the sill 1.2
(cylinder pressure plus weight of piston and rod)
10. Maximum upward load of hydraulic cylinder or wire rope when a gate 1.2
is jammed or fully open
11. Ice impact load (also debris) or lateral loading due to thermal 1.6
expansion (73.0 kN/m) applied uniformly distributed along the width
of the gate at the upper pool elevation)
Impact of debris 1.0
12. Side-seal friction load ( 0.5) 1.4
13. Trunnion pin friction load ( 0.1^0.15) 1.0
14. Earthquake design loading 1.0
15. Wave loading 1.2

102
 When considering the combination of loads under service conditions, the Structural
following are deemed not to act simultaneously (numbering corresponds with considerations
loading factors given in Table 5.1):
The maximum net hydrostatic load (4.) coinciding with either wave load
(15.), ice impact load (11.), or earthquake load (14.) is considered negligible.
The operating condition of maximum downward load of a hydraulic hoist
system (8.), wave load (15.) and ice impact load (11.) will not occur at the
same time.
The likelihood of opening or closing a gate at the same time as an earthquake
occurs is considered negligible.
In the event of failure of one hoist, the likelihood of it coinciding with either
the maximum net hydrostatic load (4.), wave load (15.), ice impact load (11.)
or an earthquake load (14.) is considered negligible.
If the gate jams, the simultaneous occurrence of the following loads is
considered unlikely: the maximum net hydrostatic load (4.), wave load
(15.), ice impact load (11.) or an earthquake load (14.).

Structural design of radial gates

The main element of a radial gate is the skin plate assembly which is stiffened and
supported by either horizontal beams or curved vertical ribs (Fig. 5.1).
The skin plate and the stiffener beams or vertical ribs act compositely. Either
form of construction transfers the load on the gate to the gate arms. The gate
arms form a splayed portal with the beams tying the gate arms in the horizontal
plane. The gate arms converge at the trunnions. The trunnion bearings are
anchored to the piers and the abutments by trunnion beams (Fig. 5.2).
The usual arrangement is to use two gate arms per side and three for larger
gates. A few gates of relatively low height have been constructed with a single
arm per side in the form of a tapered box girder. The gates are inclined to
optimise the design of the horizontal beams. There are also operational reasons
for inclining the gate arms. Debris can lodge between the gate arms which are
parallel to the sluice walls, and during severe winter conditions ice can bridge

Figure 5.1. Radial gates with skin

plate assembly stiffened by horizontal
beams or curved vertical ribs

103
 Hydraulic gates and
valves

Figure 5.2. Trunnion beam

and trunnion bearings

the gap between the gate arms and the piers. Different arrangements of bracing
the gate arms are shown in Fig. 5.3.
When the skin plate assembly is stiffened by horizontal beams, as shown
in Fig. 5.1, the load on the skin plate is transmitted by the horizontal
members to the two main vertical members which tie the gate arms in the
vertical plane.
When the skin plate is stiffened by curved vertical ribs, the load on the skin
plate is transmitted to the horizontal beams tying the gate arms and forming a
portal with the gate arms. The curved vertical ribs are usually of constant section
throughout. Spacing of the vertical ribs is based on the plate stress of the lowest,

104
 Structural
considerations

Figure 5.3. Different

arrangements of bracing gate
arms

most heavily loaded section. As a consequence, the upper section of the skin
plate assembly becomes more lightly stressed.
In horizontal reinforcement, the spacing of the stiffener beams can be varied
to equalise the panel stresses, and the beam sections can be selected to suit the
loading imposed on the adjoining plates.
Structurally, the horizontally reinforced skin plate assembly is more
economical. In radial gates, the reinforcing members of the skin plate constitute
about one-third of the total weight of the gate. The material saving will amount
to 10^15% of the weight of the skin plate assembly.

105
 Hydraulic gates and Where the span of a radial gate is wide in relation to the height, the
valves horizontally reinforced skin plate is a more efficient choice. In such gates, an
arrangement of four arms per side can save weight and result in a torsionally stiff
structure. The four arms form a tapering box. Figure 5.4 shows a counter-
balanced radial automatic gate designed on this basis.
One disadvantage of horizontally reinforcing the gate skin plate assembly of
river control gates is the potential for debris to accumulate on the beams. This

Figure 5.4. Wide span radial

gate with four arms per side

106
can also occur on the lower horizontal tie member of the gate arms of vertically Structural
stiffened skin plates. considerations
When a gate operates with the downstream side partially submerged the sheared
flow inthestilling basinresults instrong recirculation which impinges onthedown-
stream side of the gate. As debris, particularly floating timber, is carried under flood
conditions and discharged under the gate, the recirculatory motion in the stilling
basinwilldeposititonthehorizontalbeams.Thiscanbecomeadifficultmaintenance
problem. AttheTorrumbarry WeirontheRiverMurrayinAustralia,aconsiderable
quantity of timber accumulates during the flood season and is discharged under the
gates. To prevent timber build-up, the horizontal skin plate stiffener beams were of
closed sections, designed so that lodgement of timber was prevented. Figure 5.5
shows a section of one of the six radial gates fabricated in this way.
At very large spillway gates, the skin plate is usually stiffened both
horizontally and vertically.
The potential for a gate to jam on one side when hoisted, or for failure of one
rope or multiple ropes on one side of a gate, has resulted in designs which brace the
area betweenthe gate armsto forma truss. This is implemented at the downstream
side of the horizontal beams tying the gate arms. It is regarded as standard practice
in the US Corps of Engineers' Manual 1110-2-2702. Figure 5.6 shows the load
diagram when a gate jams on one side (load factors have been omitted).

Structural design of vertical-lift gates

Vertical-lift gates are reinforced horizontally to transfer the hydrostatic load to
end frames which mount the guide rollers or slides (Fig. 5.7(a)).
A design which was frequently used on barrage gates stiffened the skin plate
vertically and then transferred the load to the end frames by two open girders
(Fig. 5.7(b)). To obtain the same load on each girder, they were often spaced
equally about the centre of pressure of the hydrostatic force. This resulted in a
girder of appreciable depth near the bottom of the gate. During drowned
discharge, a suppressed hydraulic jump occurs, causing erosion of the structural
members of the lower girder especially in rivers which carry a high silt or bed
load under flood conditions. Some of the barrages on the lower Indus in
Pakistan were subject to such erosion, but it also occurred in similar river
control structures in the UK. Rehabilitation of some of these gates was carried
out by following the original design but substituting all-welded construction
and replacing bow string girders by deep fabricated beams. There have been
instances of the lower beams causing flow reattachment of the discharge under
the gate, resulting in gate vibration (Fig. 10.14).
Relatively high loaded vertical-lift gates are sometimes designed so that the
open girders form a space frame.

Stiffening members of skin plates

Vertical stiffening members of radial and vertical-lift gates are designed on the
basis that they act compositely with the skin plate. This also applies to
horizontal stiffeners and beams. When vertical stiffeners are used, the spacing

107
 Hydraulic gates and
valves

Figure 5.5. Skin plate

assembly, horizontally
reinforced, using closed
trapezoidal sections

of the members depends on the panel stresses of the skin plate at the lowest, most
heavily loaded, part of the skin plate. This is continued to the crest of the gate,
resulting in relatively close spacing of stiffening members.
In all-welded construction T-sections are selected as stiffeners, and in riveted
construction rolled beams have to be used to provide a flange for riveting. To

108
 Structural
considerations

Figure 5.6. Load diagram and

bracing to withstand jamming of
a radial gate

weld the leg of a T-section, a minimum depth of 200 mm of the stiffener is

required. When repainting becomes necessary, the underside of the flanges is
difficult to access, particularly for paint preparation, and the quality of the finish
and its durability tends to suffer.
Horizontal stiffening beams of radial and vertical-lift gates which are welded
to the skin plate are also designed on the basis that they act compositely with the
skin plate.

Composite construction
A form of stress analysis which is specific to gates and is not normally
encountered in onshore structures is the combination of panel and bending
stresses in stiffened plates subject to hydraulic pressure. The stiffening sections

109
 Hydraulic gates and
valves

Figure 5.7. Methods of

reinforcing vertical-lift gates

welded to the skin plate of a gate (Fig. 5.8), act compositely with the skin plate to
form the top flange of a beam.
The composite beam, in bending, transfers the hydrostatic load to other
beams or support members. The skin plate is also subject to a panel stress at right
angles to the bending stress of the beam. These stresses have to be combined.

Figure 5.8. Stiffening

members welded to a plate
subject to hydrostatic load

Stresses inplates have been dealtwith byTimoshenko andWoinowsky-Krieger4

and have been tabulated in convenient form by Roark and Young.5 The two cases
most frequently met in gate analyses are reproduced in Tables 5.2 and 5.3.
For the case of a rectangular plate where all edges are fixed (Fig. 5.9 and Table
5.2) the stress at the centre of the long edge, the maximum stress is given by:

 1 qb2 =t 2

where q intensity of load (hydrostatic pressure), and

t plate thickness.

At the centre of the plate

 2 qb2 =t 2

110
 Structural
considerations

Figure 5.9. Rectangular plate

with all edges fixed, uniform
load over entire plate

and the maximum deflection at the centre of the plate is given by:

y qb4 =Et3

where E Young's modulus.

For the case of a rectangular plate with three edges fixed and one edge free,
Table 5.2. Variables in stress and deflection equations of plates (all edges fixed)

a/b 1.0 1.2 1.4 1.6 1.8 2.0 1

1 0.3078 0.3834 0.4356 0.4680 0.4872 0.4974 0.5000
2 0.1386 0.1794 0.2094 0.2286 0.2406 0.2472 0.2500
 0.0138 0.0188 0.0226 0.0251 0.0267 0.0277 0.0284

uniform load over entire plate (Fig. 5.10 and Table 5.3).
At x 0, z 0, the maximum stress is:

b 1 qb2 =t2 and R 1 qb

where R the reaction force normal to the plate surface exerted by the
boundary support on the edge of the plate in N/mm. The units of q
are N/mm2 and b is in mm.
At x 0, z b:

a 2 qb2 =t2 :

At x  a=2, z b

a 3 qb2 =t2 and R 2 qb

Figure 5.10. Rectangular

plate with three edges fixed, one
edge (a) free, and uniform load
over entire plate

111
 Hydraulic gates and Table 5.3. Variables in stress equations of plates (three edges fixed)
valves
a/b 0.25 0.50 0.75 1.0 1.5 2.0 3.0
1 0.020 0.081 0.173 0.321 0.727 1.226 2.105
2 0.016 0.066 0.148 0.259 0.484 0.605 0.519
3 0.031 0.126 0.286 0.511 1.073 1.568 1.982
1 0.114 0.230 0.341 0.457 0.673 0.845 1.012
2 0.125 0.248 0.371 0.510 0.859 1.212 1.627

Stiffener beams are considered to be continuous. Figure 5.11 shows the

loading condition, bending moment diagram and associated skin plate which
forms the upper flange of the beam. The reduction factors V1 and V2 depend
on the ratio of the panel support dimensions L and B and are listed in Table 5.4.
The methodology used here for computing the width of the panel acting
compositely with the stiffener section to form a beam is the one given in DIN
19704 (1976). This standard also gives guidance on many other aspects of
design. It has been revised (1998)1 so that it is based on load factors and not
on safe working stresses.

Figure 5.11. Composite

action of stiffeners and beams
with skin plate

112
Table 5.4. Reduction factors V1 and V2 for skin plate acting compositely in bending with stiffeners Structural
(Poisson's ratio 0.3) considerations

L/B V1 V2
1 1 1
20 0.984 0.861
10 0.938 0.753
6.67 0.867 0.660
5 0.783 0.580
4 0.697 0.512
3.33 0.616 0.453
2.86 0.545 0.404
2.5 0.484 0.363
2.22 0.433 0.324
2 0.391 0.295
1.67 0.325 0.250
1.43 0.276 0.215
1.25 0.241 0.190
1 0.195 0.155

A method and data for analysing curved skin plates and stiffener beams for
radial gates has been given by Wickert and Schmausser.6 The preferred 3D
analysis of a radial gate is by a finite element program.
The biaxial stresses represented by the panel stress acting at right angles to the
bending stress of the beam are combined so as to calculate the equivalent stress:
p
e x2 y2 x y 3 2

x and y are the normal stresses in orthogonal directions, that is the panel stress
and the beam bending stress. They are substituted with their signs.  is the shear
stress, which is calculated as:

 TS=Id or for a member with I or box section as  T=A

where T shear force

S static moment about the centroid of the section of part of the
cross-section between the point concerned and the extreme
fibres
I moment of inertia
d web plate thickness
A cross-sectional area of web plate.

References
1. DIN 19704 (1998): Hydraulic Steel Structures ^ Criteria for Design and Calculation.
2. DIN Handbook 179 (1998): Water Control Structures 1.

113
 Hydraulic gates and 3. US Army Corps of Engineers (2000): Design of Spillway Tainter Gates, EM1110-2-
valves 27023, Jan. 1.
4. Timoshenko, S P; Woinowsky-Krieger, S (1970): Theory of Plates and Shells, 2nd
edition, McGraw-Hill.
5. Roark, R J; Young, W C (1975): Formulas for Stress and Strain, 5th Edition,
McGraw-Hill.
6. Wickert, G; Schmausser, G V (1971): Stahlwasserbau, Springer Verlag.

114
 6
Operating machinery

Gates may be operated by either electromechanical drives raising the gates by

ropes or chains, or by oil hydraulic cylinders. Screw jacks have been used, in
some installations, these are the preferred means of operating penstocks.
Electromechanical drives consist of electric motors driving hoisting drums
or chain sprockets through multistage reduction gear boxes. The gates close
under their own weight with the motors controlling speed of descent. Large
speed reductions are required from motors to the rope drums or the chain
sprockets. They can be in the range 1500:1 to 2200:1.
In most gate installations, that is at spillways and river control weirs, the
critical emergency operation is opening. This also applies to tunnel gates
controlling bottom outlets. For inlet gates for turbine passages it is the
opposite. They have to close in the event of turbine rejection or runaway
conditions.
A few radial gates have been constructed to open without power. To effect
opening under gravity the gate has to be counterbalanced so that closure motion
requires the drive effort, that is, the mass of the skin plate is overbalanced. The
gates are shown in Figs 2.11, 2.14 and 2.15. The ropes for closure from the hoist
drive are anchored at the counterweight or along the arms sustaining the
kentledge. The gear reduction of the hoist motor is kept low, while high
efficiency spur or helical spur gears are used to ensure that the gear train does
not become self-sustaining, which would prevent gravity lowering.
The same principle has been applied where gates are closed by oil hydraulic
cylinders. The fail-safe operation is gate opening under gravity and power
closure by hydraulic cylinders. The spillway gates of the Victoria Dam in Sri
Lanka1 were arranged in this manner, described in detail in Chapter 11.
Oil hydraulics applied to a hoisting cylinder, usually referred to as a servo-
motor, actuate gate closure as well as opening. Oil hydraulics permit the direct
application of large forces moving slowly, eliminating electric motors, brakes,
large multistage reduction gear boxes and hoisting drums.
A servo-motor can be used at each side of a gate and the cylinders can be
linked by a pipe ensuring that the same forces are exerted by both cylinders.
The elimination of transmission shafting and overhead hoisting machinery is
an advantage at locations where visual considerations are important, and where
the appearance of electromechanical machinery above the abutments or piers is
not acceptable.

115
 Hydraulic gates and In large radial gates, the arrangement of shafts connecting mechanical drives
valves can present difficulties in the layout of the transmission from one side to the
other. The use of servo-motors overcomes this problem.
A factor in the selection of oil hydraulics as the operating medium for gate
installations is possible contamination by mineral oil of the watercourse or the
reservoir due to failure of a hydraulic pipe or a flexible hose. This is overcome by
using an environmentally compatible hydraulic fluid. A selection of such fluids
is available and their use in water control structures is frequent. They require
certain installation precautions.
Gate hoisting by ropes introduces an elastic suspension. To a lesser extent
this is also the case for chains. Certain types of gate, such as bottom-hinged flap
gates, have to be close-coupled to their servo-motor to prevent reversal of
motion. In tidal barrages the tide level can exceed the pond or reach level and
reverse the thrust on the gate. A similar consideration applies to fish-belly gates
where the fish-belly section is sealed and becomes buoyant when submerged. A
rise in the downstream water level may tend to raise the gate, stopping overflow.
The gate then exerts an upthrust which the cylinder must resist.
Tunnel gates subject to high velocity flow must be rigidly coupled to their
servo-motors. They are subject to hydrodynamic disturbing forces which will
cause vibration if the gate is suspended by ropes. Table 7.2 which gives the
moduli of elasticity of wire ropes shows that this is about one-third of that of a
steel rod. Oil hydraulic operation is discussed later in this chapter.

Electromechanical drives
Single motor drives with line shafting driving multireduction gearboxes at each
end are a common method of layout. Apart from simplicity of control, it is
possible to mount two motors to duplicate the drive in the event of one motor
failing, and to arrange for manual winding of the motor extension shaft on
mains failure. Figure 6.1 shows two common layouts of single motor drives.
Squirrel-cage induction motors are invariably used in gate installations
because of their simplicity and minimal maintenance requirement. The hoist
capacity of the motor should be as close to the required load as possible, with
only a small margin for an unexpected load combination. In the majority of cases
of hoist failure, due for instance to a racked gate, the force that caused the failure
of the hoist and sometimes the gate was most likely an oversized motor. The
motor is the easiest part of a power train to replace. It is also protected by
thermal overload devices which can be reset after stalling.
Motors should be able to start against full load, the so-called `hard start',
because of the considerable inertia of a gate system which has to be accelerated
to full speed. A standard squirrel-cage motor will develop about 150% of its full-
load torque on starting, whereas 200% is characteristic of a high torque motor.
Star-delta starting is not suitable for hoist motors, even with Wauchop `no
break' winding. With star-delta winding, the starting torque is only 54% of
the motor full-load torque, and in order to accelerate a hoisting load, the motor
has to be oversized. In the event of an obstruction such a motor can wreck the
power train and the gate.

116
 Operating machinery

Figure 6.1. Arrangements of

single motor drives for radial
gates

During on-line starting and running up to full speed, a squirrel-cage motor

will develop about 280% of full-load torque at 65% of its rated speed.
Autotransformer starting gives greater flexibility in starting characteristics but
also presents difficulties in matching motor performance to the requirement to
accelerate a substantial mass. The most common motor designs produce a
maximum torque at locked rotor of 250%, and for another type 400%, of the
rated torque. The locked rotor torque will be transmitted through shafting,
keyways, coupling and gearing which have to be designed accordingly.
In a wide gate or where a clear, unobstructed water course is required,
independent drive at each side may be necessary. Up to 30 kW per motor size,
electrical cross-synchronisation can be provided by means of power Selsyns.
These are three-phase AC motors coupled to the drive motors. The primary
and secondary windings of the power Selsyns are interconnected and will
transmit full synchronising torque at all speeds (Fig. 6.2). The development of
Selsyn stabilisers to synchronise the receiver Selsyn with the transmitter has
overcome one of the limitations of the system, specifically the requirement to
synchronise the machines at standstill and the violent standstill synchronisation
which could occur.
It is possible to design a system so that one motor can lift a gate in the event of
a breakdown of the other motor by transferring the necessary power through
the Selsyn machines. This requires oversizing of the motors and the Selsyns. If
this is adopted, the start-up characteristics have to be calculated with precision
and the mechanical components designed accordingly. If this is not done and the
pull-out torque is developed by the normal operation of the hoist motors,
shafts, keyways, couplings and gears could be sheared or damaged. It is not
possible to transfer manual winding at one motor to the other in the event of a

117
 Hydraulic gates and
valves

Figure 6.2. Principles and application of a power Selsyn motor drive

118
mains failure, since the primary windings of the power Selsyns have to be Operating machinery
energised from the mains for synchronisation to be effective.
Alternative linked drives can be achieved by synchronous motors fed from a
variable-frequency supply or thyristor-controlled DC motors with servo-
controls. The former system also offers the benefit of speed variation, if
required. The starting characteristics are not as good as those obtained with
DC motors or high torque squirrel-cage AC motors.
Thyristor-controlled DC motors with servo-controls can be up to 55 kW per
motor, or even larger.
With any thyristor control, great care should be taken with signal cables,
particularly those transmitting signals from solid-state devices. Signal cables
should be screened and should not run alongside motor feeder or control cables.
In some circumstances a clean supply should be considered.
Two hoisting ropes per side are frequently used. They are connected at the
gate anchorage point by a compensating beam to allow for differential rope
stretch. Articulation of the compensating beam is restricted so that, in the event
of the failure of one rope, hoisting can be continued with the other (Fig. 6.3).
Similar layouts are used when chains are the means of gate suspension. In this
case, the hoist drums are replaced by chain sprockets.

Figure 6.3. Hoist rope attachment to

radial gates

119
 Hydraulic gates and
valves

Figure 6.4. Loads due to wire rope

anchorage upstream of the skin
plate

The hoist ropes can be anchored to the gate either upstream of the skin plate,
Fig. 6.3(a), or downstream, Fig. 6.3(b). Upstream anchorage with the ropes in
contact with a wear plate welded to the skin plate has the disadvantage that
debris can become lodged between the ropes and the skin plate and that the
ropes are immersed during the majority of their working life. In spite of these
drawbacks this arrangement is sometimes used because the layout of the hoist
machinery shown in Fig. 6.1 is difficult to achieve or is incompatible with the
location of the gear box drive shaft upstream of the gate. However, it should be
avoided where possible.
Figure 6.4 shows the load imposed on the skin plate due to anchorage
upstream of the skin plate. A typical arrangement of a hoist for an intake gate
is shown in Fig. 6.5.

Figure 6.5. Hoist machinery

layout for an intake gate
(Kotmale Dam, Sri Lanka)

120
Oil hydraulic operation of gates in free surface Operating machinery
flow
Most hydraulic cylinder hoist systems for gates in free surface flow consist of
two cylinders disposed on either side of the gate, pressure equalised to provide
even raising and lowering. The exception is bottom-hinged flap gates, Fig.
2.25(c), where a single cylinder one side is often employed. Because of their
fish-belly shape, these gates can be designed so that they are torsionally very
rigid. Although there is some torsional deflection under full load conditions it
is usually acceptable.
Oil hydraulic circuits for gate control have to be designed so that leakage is
detected and the plant is shut down automatically to prevent loss of fluid.
Hydraulic cylinder hoist systems have a number of advantages:
large forces can be applied at low speed without the need for gearboxes
having several reduction stages
high overall efficiency
location downstream of the skin plate of a radial gate
absence of overhead structure resulting in more flexibility for locating a
bridge spanning a spillway or weir
easy arrangement of equipment redundancy and, if necessary, automatic
changeover of pumps
close overload setting
the power pack containing the motors, pumps and valves can be located a
short distance from the gate, and one power pack can operate several gates
in turn
a portable standby power pack can be easily connected
cost effective.
Disadvantages or detrimental aspects are:
Sensitivity to contamination of hydraulic fluid. Contamination of the fluid
when changing oil can result in a common cause failure, an event which
affects the whole system. In practice, in an installation which comprises
several power packs, oil changes are staggered.
Possible contamination through an oil spill or the fracture of a high pressure
hose or pipe. The practice in river or reservoir installations is to use
environmentally compatible hydraulic fluids. A number of different fluids
are available which are not detrimental to fish and which will dissolve
organically.
Operation of gates can be slightly affected by solar heating of long pipe runs.
When gates are partially or wholly open for a long time, a slight leakage
occurs from the piston rod side (the annulus) to the cylinder side, affecting
the gate attitude. This is usually dealt with by monitoring the gate position; if
it changes by more than a preset value, the power pack is automatically
started up and the original gate position is restored.
The cylinders pivot on a gimbal (Fig. 6.6), on pedestals mounted on the
adjacent sluice wall. The piston rods are attached by a clevis to the gate. The
cylinder magnitude of force and its orientation change throughout the lifting

121
 Hydraulic gates and
valves

Figure 6.6. Gimbal mounting

of a hydraulic cylinder

motion. For preliminary design of radial gates it is often assumed that the
cylinder will be at a 45 angle to the horizontal when the gate is closed, although
optimum angles may vary from that. If environmental or aesthetic considera-
tions dictate that cylinders do not project into the skyline, different cylinder
arrangements can be engineered, but they will usually be less efficient. Pistons
are either of high tensile steel, chrome plated or, for very long life, ceramic
coated.
Figure 6.7 shows the hydraulic circuit for a power pack for operating two
radial gates, each having two cylinders. Features of the circuit are:
Two motor and pump units where one acts as standby to the other.
Automatic start-up and changeover from one pump motor unit to the other
in the event of one unit failing to run up to pressure. This is effected by valves
W1 and W2.
For each gate, two directional control valves which select the raise, lower and
hold positions. One valve acts as a standby to the other. The changeover is
carried out manually.
Alternative manual operation of the directional control valves in the event of
failure of the mains supply, when the gate is operated by a hand pump or by a
mobile plug-in power pack.
Figure 6.8 shows a possible circuit for the hydraulic cylinders. Together

122
 Operating machinery
Figure 6.7. Hydraulic circuit for a power pack operating two radial gates
123
124

Hydraulic gates and

valves
Figure 6.8. Hydraulic circuit for operating the servo-motors for a radial gate
with Fig. 6.7, it forms a complete system diagram. Features of the circuit Operating machinery
are:
Pressure equalisation pipe which balances the forces in the two cylinders.
Two pressure switches, one a standby to the other, to sense the loss of
pressure in the pressure equalisation pipe.
The circuit is arranged to arrest gate motion if there is an obstruction to
closure on the sill, even if the obstruction is off-centre.
The speed of operation is sensibly constant.
The cylinder-mounted manifolds ensure that the load is always positive in the
same port of the cylinder, the annulus. Thus the load holding facility within the
manifold is only on the positive load pressure port. Load holding is achieved by
the two pilot-operated check valves and a check valve.
A direct-operated relief valve is connected to the load holding port. It is set
above the normal operating pressure. In the event of an excessive external load,
this valve will relieve the resulting excess pressure and prevent damage to the
gate and the cylinder.
A shuttle valve is required in the manifold to provide a pilot signal which
operates the load holding pilot operated check valve when gate raising or
lowering is carried out. This allows free flow from each cylinder to equalise
the load pressure. It prevents torsional or racking loads in the gate by out-of-
balance cylinder forces.
Slide gates and roller gates, when controlling flow or opening and closing
under unbalanced hydraulic conditions, must be operated by a hydraulic servo-
motor. Fluctuating and variable hydrodynamic forces act on the gate. To
prevent these from causing gate vibration, the piston and the annulus side of
the servo-motor are pressurised during opening and closing of the gate. In gates
subject to high heads of water and therefore high flow velocities, the closure
speed must be controlled to prevent hydraulic downpull forces from
accelerating the movement.
Operating gates are required to hold partial openings for a long time, and the
extent of the opening is often critical. Servo-motors are required to maintain
pressure in the piston and annulus side of the piston during partial opening,
although the hydraulic pumps are not in operation. This is effected by a
hydraulic accumulator. When the pressure in the accumulator falls below a
preset value, the pump or pumps are automatically started.
Leakage of oil occurs in hydraulic cylinders. Oil leakage past the piston seals
becomes progressively worse as seals become worn. Because of this, gates
operated by hydraulic cylinders in the open or partially open position will
gradually close. Gates are fitted with position indicators and these are utilised
to signal a predetermined gate movement, and to initiate a signal to restore gate
position.
Features which should be part of a reliable oil hydraulic system comprise:
(a) Inlet strainers for the tank.
(b) Suction and delivery filters with 5 m apertures.
(c) Two manually operated pumps.
(d) Directional control valves which can be manually operated in addition to

125
 Hydraulic gates and electrical actuation by solenoids. If several valves have to be actuated
valves manually in an emergency, they should be grouped so that they can be
operated by one man and should all move in the same direction for the
emergency condition.
(e) An offloader valve which can be combined with a relief valve. The
offloader valve will prevent heat build-up when the pump is running but
is not delivering oil to a piston.
(f) Pipework of stainless steel fastened by saddles at close spacing, especially
where there is a risk of earthquakes. Movement of pipes due to seismic
action can initiate vibration of high amplitude which can result in pipe
fracture.
(g) Flexible hoses which are rated 50% above the system working pressure.
(h) Means of locking the oil at the piston-raising port in the event of a pipe or
flexible hose fracture. This can be achieved by mounting a pilot valve on the
cylinder which closes in the event of loss of system pressure.

Hoist speed
Hoist speeds of gates are conventionally 300 mm/min. In gates which control
water level, hoisting is either carried out in steps followed by a dwell period,
both of which are controlled by timers, or by set point control.
The final closure speed of servo-motors is usually decelerated. This can be
effected by using hydraulically cushioned cylinders, whose pistons have a
stepped crown which mates with a recess in the cylinder cap. If variable delivery
hydraulic pumps are used, the pump output is reduced when a limit switch is
actuated shortly before the end of the stroke of the piston rod is reached. In dual
pump operation, one pump is shut down by the limit switch.
For constant delivery by a hydraulic pump, the gate opening and closing
speed will vary with the cross-sectional area of the annulus and the piston side
of the servo-motor. This can be compensated by using two pumps, both
delivering only when the piston side is under full pressure. Two pumps also
ensure greater reliability. In the event of failure of one pump, it is accepted that
gate opening and closing speeds will vary. Figure 6.8 shows an alternative
arrangement of an oil hydraulic circuit for spillway gates, which ensures sensible
speeds for raising and lowering the gate when only one pump supplies the
power and the second pump is a standby.

Reference
1. Back P A A; Wilden D L (1988): Automatic flood routing at Victoria Dam,
Sri Lanka, Commission Internationale des Grandes Barrages, 16 th Congress,
San Francisco, Q63, R52.

126
 7
Detail design aspects

Seals
Seals are required to prevent loss of water. Jets emitted at inadequately sealed
sills, sides or soffits are a major source of gate vibration and gate noise. Under
severe winter conditions water leakage results in ice formation and can freeze a
gate to a pier. In addition, seal leakage in gates subjected to high head can cause
long-term damage to downstream concrete works. The selection of seals and the
design of their mountings is therefore important.
Since fluctuating pressure due to flow through gaps is a major cause of gate
vibration, aspects of seal arrangements are also discussed in Chapter 10.
Seals at old gates were either of leather or staunching bars. They were used to
prevent water leakage at piers or abutments.
Block seals of timber or lignum vitae which were fitted to older installations are
not suitable, as explained in Chapter 10. Modern seals are extruded or moulded
from an elastomer. The usual materials are natural rubber or polychloroprene,
known commercially as Neoprene. The elastomers are compounded to produce
the necessary properties such as tensile strength, tear resistance, low water
absorption, compression set and ultraviolet resistance, and they contain
antioxidants. Seals are normally specified to have a Shore A hardness of 65 with
a tolerance of 5. High head gates often have seals of greater hardness. At
hardness significantly lower than 65, the coefficient of friction between a seal
and a stainless steel sliding surface increases. The friction coefficient is also
affected by the surface finish of the seal contact face.
Approximate values are:
Shore A 55 hardness: coefficient of friction 0.8
Shore A 65 hardness: coefficient of friction 0.7
Shore A 80 hardness: coefficient of friction 0.6
For PTFE covered seals the coefficient of friction reduces to 0.1.
Side and top seals rely on water pressure to aid sealing. The gap between a
gate skin plate and the side rubbing strip must be sufficiently wide to permit
inaccuracies of tracking and deflection under load, including dimensional
variations to the seal contact faces. The seals must be able to accommodate
these variations. If the gap is made too wide, the seal may not develop

127
 Hydraulic gates and adequate contact pressure or may extrude through the gap under the
valves hydrostatic head.
Seals should be preset, that is the stem should be placed under deflection, but
not under compression, because the force required to compress a seal bulb is
very much greater than that required to deflect the stem of a seal. If friction is
to be minimised, a preset of 3^5 mm is advisable. Less will be adequate for
effective sealing but is unlikely to be sufficient to allow for dimensional
variations of the seal contact plates and the gate structure.
Bottom seals rely on the weight of the gate to provide the contact pressure for
sealing. Figure 7.1 shows typical seal shapes.

Bulb seals in the shape of a musical note

These are more frequently used in their solid, rather than hollow, bulb
form. The solid bulb seal has a reduced contact area because it deforms less
than the hollow bulb type under water pressure and is less liable to
compression set.

Figure 7.1. Typical seal shapes

128
 Detail design aspects

Figure 7.2. Mounting of bulb

and L-shaped seals

Figure 7.2(a) shows the commonest arrangement of a bulb side seal. Distance
c must allow an adequate area for the upstream water pressure to force the seal
into contact with the stainless steel contact plate embedded in the pier face. The
seal clamping plate must not be too close to the bulb to permit the seal to flex
under pressure, distance b. Mounting the side seal on an angle section permits

129
 Hydraulic gates and site adjustment. Clamping of the seal mounting angle and the skin plate should
valves be on the downstream side of the skin plate to permit the sill seal to be extended
to a junction with the side seal.

Double bulb seals

The main application is in a situation which requires sealing against a reversal of
head, as occurs in gates in a tidal river.

L-shaped seals
This type of seal (Fig. 7.2(b)) can be more effective than the bulb seal, but can
only be used for movement in one plane, that is in radial or vertical-lift gates and
not in bottom-hinged flap gates. The L-shaped seal is inherently more flexible
and can blow through or `fold under' if clearances become excessive.

Double stem (centre bulb) seals

These are used for face contact and particularly for sealing the top edge of
submerged vertical-lift gates and radial gates in a culvert. The seal is suitable
for forward and backward motion which occurs at the overflow flap of a radial
gate or a bottom-hinged flap gate. Seals can be arranged so that they will move
towards the seal contact plate under the influence of the upstream head. At
high head gates when the seal is pressurised there is a risk that it could be
extruded from its clamping. To prevent this, a key is moulded at the end of
the stems.

Sill seals
Sill seals (Fig. 7.3) are formed by a rectangular elastomeric section. They should
be mounted as far as possible downstream of the gate. Bulb seals, elastomeric or
timber blocks should not be used as sill seals.
To ensure even clamping by seal retaining strips, the bolts passing through
bulb and L-shaped seals should incorporate ferrules, as shown in Fig. 7.2.
Bolt holes through the skin plate for seal clamping are a frequent source of
leakage. This is avoided by fitting nylon washers or washers faced with an
elastomer under the nuts.
Splices in seals should be hot vulcanised, whether in the factory or on site,
and corners should be moulded. Special moulds are available for the junction of
different seals.
When designing a gate sealing system it is desirable to arrange sill and side
seals to be in the same plane. This applies also to the uppermost or lintel seal for
gates in conduit. It simplifies the sealing of the junction between the seals. If
this cannot be achieved a block seal must be introduced to bridge the gap, but
these are difficult to engineer successfully (95% of seal leakages occur at corner
seals).
Rubber has a tendency to contact bond when kept under high compression
for a long time. Where gates are infrequently operated and are under high
hydrostatic head, it is advisable to uprate the coefficient of friction by 20%.

130
 Detail design aspects

Figure 7.3. Sill seal for a high head,

vertical-lift gate
The types of gate shown in Fig. 7.4 are:
(a) vertical-lift gate
(b) radial gate
(c) submergible radial gate
(d) bottom-hinged flap gate
(e) mitre gate
(f) drum gate
(g) double leaf vertical-lift gate
(h) vertical-lift gate with flap
(i) bear-trap gate
(k) cylinder gate
(l) vertical-lift tunnel gate
(m) culvert valve (reverse tainter gate)

Figure 7.4. Gate boundaries

requiring sealing

131
 Hydraulic gates and Sealing radial gates
valves
Side sealing The elastomeric seal bridges the gap between the skin plate and the
sluice wall. On the pier or sluice wall a stainless steel contact plate provides the
rubbing surface.
Figure 7.5(a) is a common arrangement. The gap between the skin plate and
the seal contact face is of the order of 10 mm to permit minor inaccuracies in the
width of the skin plate, the perpendicularity of the seal contact plate and the
tracking of the gate. A greater gap could cause the seal to be extruded through
the gap under the hydrostatic pressure of the upstream water.
Variations in tightening the bolts of the seal clamping plate can cause
undulations of the seal. To ensure even pressure a ferrule is inserted in the holes
in the seal, one millimetre less in thickness than the seal itself.
The arrangement of side seals shown in Fig. 7.5(b) is used when the pivot
centre and the origin of the weir plate radius do not coincide. This is sometimes
adopted on large radial gates to reduce the hoisting force. In this case the seal
mounting plate is on a radius whose origin is the pivot centre. This ensures that
the seal is only subjected to radial sliding forces and not to lateral movement.
In Fig. 7.5(b) debris can lodge in the space between the weir plate and seal
mounting plate. This can be reduced by keeping the gap between the skin plate
and the flume wall to a minimum.

Sill sealing The sill seal for a radial gate (Fig. 7.6), is formed by a rectangular
elastomeric section. The face of the seal should be angled to provide a contact

Figure 7.5. Arrangement of

side seals for radial gates

132
 Detail design aspects

Figure 7.6. Arrangement of sill seals for

radial gates

area with the sill beam. Other seal profiles have been used but are incorrect for
hydraulic reasons. This is more extensively discussed in Chapter 10 where it is
pointed out that to prevent gate vibration there must be a sharp point of flow
separation.
The seal is set to deflect 2^3 mm when the gate seats on the sill. The sill seal
should be located downstream of the weir plate; upstream it can cause flow
separation. This favours the arrangement of the side seal downstream of the skin
plate, as shown on the lower section of Fig.7.5(a). The seals are then in the same
plane and a leakage path at the gate corners can be eliminated.
The sill seal can be used to take up only limited variation between the weir
plate and the sill. Excessive projection of the seal below the weir plate causes it to
deflect and leakage to occur. The angle at the sill should not be more acute than
45, otherwise lip deflection will result in sealing difficulties.
In practice, local leakage can occur due to dimensional variations and the seal
clamping arrangement shown in Fig. 7.6(b) has been employed. The upper
adjustment screws are used to increase the pressure on the lip of the seal, forcing
it into contact with the sill.

133
 Hydraulic gates and The US Army Corps of Engineers' Manual `Design of spillway tainter gates'
valves recommends the seal configuration shown in Fig. 7.6(c). If this is used, the angle
of the lip should be checked when the gate is in the open position and is still
controlling the discharge. Under these conditions flow reattachment ^ a cause
of vibration ^ can occur. This is discussed in Chapter 10. Vibration has occurred
in gates with a lip (Fig. 7.6 (c)). Gates with a sill seal designed by the US Bureau
of Reclamation (Fig. 7.6 (a)), have not been subject to vibration.
The sill seal can be abutted to the side seals in Fig. 7.5(a). Preferably the two
seals should be vulcanised together. When used in conjunction with any other
side-seal configuration, such as Fig. 7.5(b), a leakage gap is created between the
seals. To avoid this a block seal is introduced at the junction.
On large spillway gates, especially in tropical countries, there can be an
appreciable temperature difference between the upstream face of the weir plate
which is in contact with the water and the downstream face which may be
exposed to the sun. The resulting curvature of the plate can create a leakage
gap in the middle of the sill. Some gate manufacturers claim that leakage can
be avoided by mounting the seal upstream of the weir plate, and that the effect
of the turbulent hydraulic conditions created by mounting the seal in this
manner can be overcome by stipulating that the gate is never opened by less than
100 mm. This avoids vibration of the gate due to severe pressure fluctuations at
low flows. However, the seal should not be placed on the upstream face.
Stipulating a minimum opening of the gate does not necessarily compensate
for the turbulent hydraulic conditions caused by mounting the seal upstream
of the weir plate.
The sill seal and the side seal contact plates should be manufactured from
stainless steel and this is also the practice for side guide roller contact plates.

Sealing vertical-lift gates

Side sealing Side seals are usually of the musical note type, arranged as shown in
Fig. 7.7.

Sill sealing Figure 7.3 shows a detail of a sill seal for a vertical-lift gate.

Lintel sealing For gates in conduit, a double-stem centre-bulb section is used to

seal at the lintel. Using a seal of similar profile to the side seal and arranging it in
the same plane simplifies the detail of upper corners. Figure 7.8 shows a detail of
a typical lintel seal. By admitting the upstream head of water to the underside of
the seal, the pressure deflects the seal towards the contact plate. A similar seal
arrangement is used at the lintel when radial gates are located in a conduit or
operate as culvert valves.

Sealing bottom-hinged flap or tilting gates

Side sealing Figure 7.9 shows different arrangements of side seals. Bottom-
hinged flap gates are frequently used at tidal barrages because they can prevent
the flow of estuarial salt water into the fresh water river course. In this
application they may have to resist pressure in either direction. The seal
arrangements of Fig. 7.9(a) and (b) will effect this.

134
 Detail design aspects

Figure 7.7. Side seal for a

vertical-lift gate

Hinge sealing Figure 7.10 shows the arrangement of a hinge seal to

withstand only upstream head. Sealing at the piers is by abutting the seal
to the sluice wall. To make an effective seal requires accurate assembly which
is not always attained. The transition between the side and sill seals also
presents problems.

Side-seal friction The relationship between the coefficient of sliding friction and
the water pressure decreases linearly with water pressure for elastomeric seals
operating on a wet surface. For SH65 seals the relationship in the range of
0.2^1.5 N/mm2 pressure is:

 0:28a 0:86

where  coefficient of friction and a applied pressure N/mm2. For PTFE

faced seals  is reasonably constant at a value of 0.1 throughout the range. The
size of the area in contact has no influence on the coefficient of friction.

135
 Hydraulic gates and
valves

Figure 7.8. Lintel seal for a high

head, vertical-lift gate

The seal friction equals twice the length of the seal (on one side of the gate) 
average pressure (a)  coefficient of friction ()  width of the seal in contact
with the seal plate. p
Seal contact plates should have a machined face and a surface finish of 1:6
or better.
Side seals are designed to deflect 3 mm on assembly to ensure that they
remain in contact with their contact plates due to dimensional variations and
thermal contraction. For calculating hoisting forces due to seal friction, the
force to deflect a seal adds to the friction force.1
If calculations for gate vibration are carried out (see Chapter 10), the seal
friction is a damping force. It would be prudent to omit the additional force
due to seal deflection.

136
 Detail design aspects

Figure 7.9. Side seal

arrangements for bottom-hinged
flap gates

Guide and load rollers

Side guide rollers for radial gates
Figure 7.11 shows a side guide roller. Smaller gates of aspect area up to 40 m2 are
sufficiently rigid not to require side guide rollers. However, in the event of rope

137
 Hydraulic gates and
valves

Figure 7.10. Hinge seal for

bottom-hinged flap gates

failure or jamming of a gate due to other causes, the provision of guide rollers
would be justified. Larger gates are fitted with two guide rollers per side.

Load rollers for vertical-lift gates

Load roller pressures are specified in DIN 19704,2 where clause 7.3.6 lists
permissible Hertz pressures.

Figure 7.11. Side guide roller

for radial gates

138
Table 7.1. Permissible Hertz2 pressures 1 and 2 Detail design aspects

Operational Contact surfaces Gates (not Gates and locks

condition frequently (frequently operated)
operated)
Rolling motions Rail-wheel 1 1.85 B 1.6 B cylindrical
between non- cylindrical
hardened contact Roller-wheel or 2.00 B 1.9 B
surfaces roller-axle 2

The two materials in contact with one another have different ultimate tensile
strengths and B is the lesser of the two values. For hardened contact faces, the
stresses may be increased according to the hardness of the material.
The above values apply to rolling components temporarily immersed in
water. For rollers permanently immersed in water and temporarily exposed to
heavy water flows, the Hertz2 stresses should be reduced:
for 0 to 300 revolutions under load per year: by 10%
above 300 to 2000 revolutions under load per year: by 15%
above 2000 to 20,000 revolutions under load per year: by 30%
above 20,000 revolutions under load per year: by 40%
For spherical revolving surfaces (crowned rolling face) with a diameter ratio
of 15:1, the permissible Hertz2 pressure between wheel and rail may be
increased by 50%. The diameter ratio is twice the radius of the crown of the load
roller divided by the diameter of the load roller.
The values given in Table 7.1 should be halved for Stoney rollers, due to the
uneven load distribution which occurs in Stoney-roller trains. This does not
apply to caterpillar rollers because they are not set in a fixed train.
These criteria lead to very high stresses, particularly if manganese or nickel
manganese steels are used. USA and UK practice is to use lower Hertz pressures
(0.7^ 0.8 of these values).
Bearings are either bronze alloy bushes or bushes with lubricant inserts as
shown in Fig. 7.16, used because of their established performance in underwater
conditions. This applies whether the bearing is sealed or not, because the seal is
liable to break down after 10^15 years due to ageing or wear. Graphite
containing lubricants should not be used in conjunction with stainless steel as
this causes electrolytic action which is accelerated underwater.
If a plain bush is used the roller is often crowned to ensure that it will
centralise and that the pressure distribution is symmetrical. On vertical-lift gates
crowning of the load rollers compensates for deflection of the gate, which
would otherwise cause excessive contact pressure on one side of the roller.
With self-aligning bearings of the angular contact type, the rollers have
parallel faces and the articulation of the bearing ensures centring of the roller.
The design of the sealing system and the back-up to prevent ingress of water
through any bolted face is critical. This is achieved by `O' rings. For maximum
reliability it is necessary to ensure that seals are lubricated. Some gate
manufacturers provide separate grease passages to each seal.

139
 Hydraulic gates and
valves

Figure 7.12. Load rollers with

bronze bush and roller bearing

The difference in the coefficient of friction between a bronze bush and a roller
bearing (0.1 as compared with 0.0018) makes an appreciable difference to the
lifting force required, as the following example illustrates (see Fig. 7.12).
With a bronze bush:

H1 10 000  0:1  200=600 333 kN

With roller bearing:

H2 10 000  0:0018  300=600 9 kN

Where there are a number of rollers mounted on each side of a gate,

adjustment to ensure even contact and appropriate load distribution is
important and can be provided by making the shaft `a' in Fig. 7.13 eccentric.
The roller shafts are rotated on assembly of the gate so that all the rollers are
accurately aligned. When this process is complete, the rollers are locked in
position.

Trunnion assembly
The trunnion assembly consists of the bearing housing or pedestal which is
bolted to the trunnion beam, which is anchored to the pier. A trunnion spindle
or axle and a bushing or bearing complete the assembly.
The gate load is transmitted either via the trunnion beam to the pier nosing
or, in some smaller gates, directly into a recess in the sluice wall (Fig. 7.14).
Spherical bearings can take the form of roller bearings (Figs 7.13 and 7.15), or
slide bearings formed by an inner or outer sphere (Fig. 7.16).

140
 Detail design aspects

Figure 7.13. Roller assembly

with eccentric shaft

Spherical bearings are more expensive than bushed bearings because of their
size and additional parts. However, they compensate for a degree of misalignment
of the gate arms, construction tolerances and thermal changes in dimension. Such
misalignments must be very small because they can result in problems with the
tracking of a radial gate. When bushed bearings are used, even a slight
misalignment of the gate arms in the horizontal plane causes non-uniform pressure
distribution on the trunnion shaft and the bush. The coefficient of friction of a
conventional lubricated bronze bearing material varies between 0.2 for starting
conditions to 0.08^0.1 for permissible bearing pressures of 200^300 bar, although
in practice designers tend to limit bearing pressures to 70^80% of the permissible
values. For self-lubricating bronze or alloy bearings the starting coefficient of
friction is stated to be 0.1, and under running conditions 0.08.
Self-aligning double row roller bearings, such as the ones shown in Figs 7.13
and 7.15, have a coefficient of friction of 0.0018, although a higher value is often
assumed for design purposes.

141
 Hydraulic gates and
valves

Figure 7.14. Gate trunnion

with bushed bearing

The friction at the interface of the bearing and shaft in Fig. 7.16, or the two
faces moving relative to one another in a self-aligning bearing, causes a bending
moment on the gate arms. From this point of view, roller bearings are the
preferred choice. On the other hand, roller bearings require maintenance, which
is not the case for self-lubricating bearings.
Figure 7.17 shows a trunnion assembly with a spherical self-aligning, self-
lubricating bearing. The lubrication is provided by solid lubricant inserts, as
shown in Fig. 7.16.
The trunnion shaft is of austenitic stainless steel, the same as the material of
the bearing's inner ring. A bronze sleeve is interposed between the trunnion
shaft and the inner ring to prevent galling of the stainless steels, which could
prevent disassembly. The axial centre holes in the trunnion shaft permit
application of hydraulic pressure should it become necessary to disassemble
the bearing during the lifetime of the gate.

Trunnion bearing failure and lubrication

Spillway gate 3 of the Folsom Dam on the American River in Sacramento
County, California, collapsed in July 1995,3 causing the release of an
uncontrolled flow of 1130 m3/s. This was due to corrosion on the loaded side
of the steel trunnion pins which had increased friction over a period of time,4
initiating the failure of a strut brace of the gate arms.
The Folsom Dam was designed and constructed between 1948 and 1956.
At that time, vertical moments in the struts caused by trunnion friction were
not considered or specified in contemporaneous standards. This

142
 Detail design aspects

Figure 7.15. Gate trunnion

with self-aligning roller bearing

subsequently became a mandatory requirement.5,6 This event drew attention

to inadequate design of bushed trunnion bearings and structural design
deficiencies.
The trunnion shafts were of low-alloy steel. Rust occurred only on the loaded
side of the shaft because the lubricant film was thinnest there. A thin lubrication
film on the loaded side is the result of high trunnion loads and the extremely low
rotational speed. Due to the thin film and small spaces, water can enter between
the pin and bushing by capillary action. Another cause of corrosion could be a
galvanic reaction when the metal-to-metal contact occurs with water or another
contaminant acting as an electrolyte.
Some of the conclusions of the report were:
Adding grease while the gates are in motion may reduce friction by as much
as 40%. The only practical method for adding grease while a gate is in motion
is to install an automatic system for each trunnion.
Frequency of greasing is important. The greasing system(s) should be
operated whenever the associated hoist motor is running. Also, the system(s)
should be operated monthly or weekly, depending on reservoir level.
Selection of the correct grease is critical (see Chapter 18 on maintenance).
Weather protection covers should be installed on the trunnions.

143
 Hydraulic gates and
valves

Figure 7.16. Gate trunnion

with spherical plain bearing

If self-lubricating bushes are used, or self-aligning bearings based on

lubricant inserts as shown in Fig. 7.16, the above precautions are not necessary
provided the bearing manufacturer can demonstrate a long record of
satisfactory underwater operation.

Trunnion mounting for radial gates

Trunnion assemblies are sometimes recessed into the piers (Fig. 7.14). When the
loads are considerable it becomes difficult to distribute the loads to the anchors.
The most frequent type of gate anchorage system is a trunnion beam at the
end of a pier which is anchored by bolts or tendons extending into the concrete
pier (Fig. 5.5). Anchorage systems may be either prestressed or non-prestressed.
The prestressed system uses high strength preloaded components. Prestressing
reduces the load deflection and results in a more rigid anchorage system.
Figure 7.18 shows post-tensioned anchorage systems where tendons, which
may be high strength low-alloy steel bars or strands, are enclosed in ducts. The
tendons are post-tensioned at the trunnion beam. The embedded length of

144
 Detail design aspects

Figure 7.17. Trunnion

assembly with a spherical self-
aligning, self-lubricating
bearing

tendons is typically 10^15 m. The ducts are formed using either galvanised steel
tube or polyethylene, and are arranged to prevent moisture entering the duct.
Prestressing cables for this application are encapsulated in oil to permit
restressing after relaxation.

Limit switches
Limit switches are one of the most vulnerable components of the gate hoist.
Problems range from icing over to corrosion of contacts, failure to actuate,

145
 Hydraulic gates and
valves

Figure 7.18. Post-tensioned

anchorage system

breaking of limit switch arms and loss of calibration. Since limit switches
control overhoisting, failure can have catastrophic consequences (see Chapter
12).
Rotary (geared) limit switches are frequently used, especially in tunnel gate
installations where the gate has to be lowered a long distance. Preferably they
should act only as a back-up to position limit switches. Rotary limit switches are
difficult to calibrate due to stretch of wire ropes and thermal changes. Use of
pre-stretched ropes avoids the need to adjust switches due to the initial set of
ropes. The effect of thermal changes can be compensated by setting the gate in
the closed position in winter and in the fully open position in summer.
Because of the importance and vulnerability of limit switches they should be
provided with a back-up switch, and the electrical circuit should be arranged so
that the function of the limit switches can be tested at the gate control cubicle.
One arrangement is to make the operating limit switch resetting and the back-
up limit switch non-resetting to alert the operator that the primary switch has
failed. However, this can inhibit gate operation at a critical time. It also provides
no indication that the back-up limit switch has failed, due perhaps to icing or
corrosion, while the operating limit switch is still functional.
Electrical testing of a limit switch will only reveal that the electrical circuit is
functioning. Failure is often caused by a spring fracture, breaking of the arm or
loss of the roller follower which will not show up on an electrical test.

146
 Where possible, vane operated magnetic proximity switches should be used. Detail design aspects
They are totally enclosed with sealed contacts and have no mechanical moving
parts. These switches provide a much higher degree of reliability than can be
achieved with cam or lever operated limit switches. They can be supplied
suitable for immersion in up to 100 m of water. The maintenance of accurate
working clearances is important when proximity switches are used. This is
difficult to attain at rope operated gates.

Ropes
The usual practice is to provide two ropes per side for hoisting and to attach
them to the gate by means of a load compensating arm. The arm is arranged so
that in the event that one rope fails, hoisting can continue with the other.
Fibre core ropes provide increased flexibility compared with steel core ropes;
however, the strength of fibre core ropes is less than that of steel core ropes of
equivalent size. Increased flexibility permits a smaller diameter rope drum, but
this advantage is cancelled out if strength requirements result in the selection of
a greater diameter rope. At gates, sections of hoist wire ropes are either
immersed for long periods in water or, at best, subjected to frequent splashing.
The core becomes a zone of moisture accumulation and decomposition. Fibre
core wire ropes should not be used on hydraulic gates.
Stainless steel wire ropes have outstanding corrosion resistance. For larger
sizes of rope (greater than 15 mm) there is an appreciable cost difference. If
replacement of ropes is difficult and likely to create operational problems, this
becomes a factor in the selection of stainless steel ropes. Another consideration
is fatigue life, which is lower for stainless steel ropes. If the gates are frequently
operated or if they pass over sheaves, creating additional bending stresses, high
tensile steel wire ropes should be considered.
Spearman7 recorded electrolytic corrosion of stainless steel wire ropes on
spillway gates where the ropes were located upstream of the gate skin plate. It was
not mentioned whether the contact area between the ropes and the skin plate was
lined with stainless steel. In the absence of lining, the chafing between the ropes and
the skin plate causes removal of paint on the skin plate and early corrosion.
Galvanised wire ropes can provide a reasonable level of corrosion protection.
Some high tensile strength galvanised wire ropes have a lower breaking load
than ungalvanised ropes. The reduction can be in the region of 3^5% for similar
size ropes. This is due to a reduction in the diameter of the rope strands to
compensate for the added galvanising zinc. The reduction in breaking load does
not apply to all ropes.
Ropes are subject to elongation due to settlement of the wires in the strands
and the strands in the rope. When using geared limit switches of the type that
measure a gate's distance of travel by the rotation of a hoisting drum, the limit
switches have to be reset after the initial constructional extension of the rope has
occurred. In most cases it is advantageous to use ropes which have been
prestressed.
The coefficient of linear expansion of wire ropes is 12.5  106 per C, the
same as for all steels. With long falls, thermal elongation and shortening of ropes

147
 Hydraulic gates and Table 7.2. Apparent modulus of elasticity of wire ropes8
valves
Type of rope E:
N/mm2  103
6 stranded ropes ^ fibre core simple construction 62.0
(e.g. 6  7)
6 stranded ropes ^ steel core simple construction 68.7
(e.g. 6  7)
6 stranded ropes ^ fibre core compound construction 49.0
(e.g. 6  19, 6  36)
6 stranded ropes steel core compound construction 59.0
(e.g. 6  19, 6  36)
Multistrand non-rotating construction 42.0
(e.g. 17  7)

between summer and winter conditions may require seasonal adjustment of

geared limit switches.
Lubrication of wire ropes is important for maintaining their lifespan. This
applies to high tensile steel wire ropes as well as stainless steel ropes. A dry rope
unaffected by corrosion, but subject to bend fatigue due to wrapping around a
hoisting drum, is likely to achieve only 30% of the fatigue life of a lubricated
rope (Bridon Ropes8). Wire ropes should be supplied pregreased by the
manufacturer.
The modulus of elasticity of a rope is much less than that of the steel of its
individual wires because of the helical winding of wires making up a strand and
the helix of the strands. It varies according to the construction of the rope. Table
7.2 gives some values.
Generally in gate hoisting applications, the factor of safety on the rope
breaking strength is 5; this is considered a `normally loaded' rope.
If gate vibration occurs (see Chapter 10) it is likely to cause deterioration in
wire ropes. If vibration continues for some time it can initiate wire fractures
because the rope absorbs the vibration. The fractures may be internal and so
not revealed by visual inspection, unlike fatigue fractures of external strands
due to repeated bending or chafing.

Chains
Two types of chain are used for hoisting gates.
In the roller chain shown in Fig. 7.19(a), known as a `Galle' chain, the link
pins rotate in the chain links and the link pins slide in the sprocket teeth during
hoisting.
In Fig. 7.19(b) the link pin carries a bush. With this type of chain there is no
sliding of the pin relative to the teeth of the sprocket, rather movement occurs
between the bush and the pin. The chain absorbs less power and can be supplied
with grease nipples in each pin so that all rotating faces can be lubricated. The
chain in Fig. 7.19(a) is usually lubricated by drip feed.

148
 Detail design aspects

Figure 7.19. Hoisting chains

(by courtesy of Renold plc)

Difficulties have been experienced with chains of type (a) due to high friction
at the pin face which bears on the link. This has resulted in chains failing to fully
articulate. This is less likely to happen with chains of types (b) and (c). However,
both types have suffered from corrosion, particularly where chains are used
upstream of the skin plate of a radial gate and remain submerged in water for
long periods. Service lubrication of the joints and pins of both chain types is
only partially successful and will not prevent corrosion.
Corroded chain links will develop kinks at the joints. Even under load, some
chains have not straightened out. Kinks change the effective length of a chain
and the gate will not be raised evenly.

149
 Hydraulic gates and References
valves
1. Semperit: Gate seals catalogue, 3rd edition, Semperit Technische Producte, Wien,
Austria.
2. DIN 19704 (1976): Hydraulic steel structures: criteria for design and calculation.
3. Bureau of Reclamation (1996): Forensic report, spillway gate 3 failure Folsom Dam, US
Bureau of Reclamation, Mid-Pacific Regional Office, Sacramento, California, Nov.
4. Koltuniuk, R; Todd, R (1996): Determination of trunnion friction coefficient from tests on
reinforced spillway radial gate, Folsom Dam, Ca., Bureau of Reclamation, Technical
Service Center, Denver, Colorado, Jun.
5. US Army Corps of Engineers (1993): Design of hydraulic steel structures, EM 1110^2^
2105, Mar.
6. US Army Corps of Engineers (2000): Design of spillway tainter gates, EM 1110^2^2^
702, Jan.
7. Spearman, P C (1967): Design and development of radial spillway gates in New
Zealand, New Zealand Engineering, Feb.
8. Bridon Ropes: Steel wire ropes and fittings, publication 1304. Bridon Ropes,
Doncaster, South Yorkshire.

150
 8
Embedded parts

Sill beams, side-seal contact and roller faces on radial gates, gate guide roller and
sliding paths for vertical-lift gates and tunnel lining sections of high head gates
have to be embedded in concrete. They must be rigidly secured and accurately
aligned. The practice is to provide cut-outs in the primary concrete and means of
fixing alignment screws. The embedded parts are then accurately set up and
secondary concrete is cast around them.
To illustrate the sequence of erection, an example of an embedded sill beam is
shown in Fig. 8.1.
The pads (1) for the adjusting studs are cast into the primary concrete. The
adjusting studs (2) are then welded to the pads. This is followed by positioning
the sill beam (3), aligning and levelling by adjusting the nuts on the studs. The
final operation is to cast the secondary concrete. Adjusting studs should not be
less than 15 mm in diameter. They should not be assumed to tie in the primary
and secondary stage concrete; separate reinforcement should be provided to
carry out this function. Dovetailing the first stage blockouts on the sides is
advantageous.
Best practice is to machine the top flange of the sill beam and to line it with a
stainless steel sill-seal contact plate. The plate is either welded to the beam, or in
some cases screwed to the beam so that it can be renewed. If this practice is
adopted, insulation against electrolytic corrosion between the carbon steel and
the stainless steel is advisable.

Figure 8.1. Embedded parts for

sill seals

151
 Hydraulic gates and
valves

Figure 8.2. Side seal contact

face and guide roller path for a
radial gate

Figure 8.2 is an example of the embedded side-seal contact face for a radial
gate and the roller path for the gate side guide rollers. (1) is a rail for the gate
transverse guide roller, (2) is the roller path and (3) is the side-seal contact plate.
Figure 8.3 is an illustration of the gate slot of a high-head vertical-lift roller
gate. (1) is a rail for the gate transverse guide roller, (2) is the roller path and (3) is
the side-seal contact plate.
The method of providing a rigid fixing and alignment of the embedded parts
of a lintel seal for a high-head slide gate is shown in Fig. 8.4.
It is highly desirable for all faces in contact with water to be of stainless steel.
The corrosion resistance of stainless steels depends on the alloying content of
chromium and nickel. Austenitic stainless steels with a chromium content of
15% or greater and nickel of 10% or greater have the best corrosion resistance
of the three groups of stainless steel (see Chapter 15). Unprotected low carbon
steels should not be closer than 75 mm to a water face.
Steel linings in gate slots or tunnel inverts can be repainted when stoplogs or
bulkhead gates have been positioned and the section has been dewatered. This
excludes steel linings for stoplog slots, which cannot be refurbished throughout
the existence of the structure unless the reservoir or river reach is drained down.
In practice the application of stainless steel to faces in contact with water is often
confined to seal contact and sliding faces.
The design criteria for thrust faces of embedded parts, sill beam and slide or
roller paths are empirical. The distribution of load is assumed to be effected by
the lower flange of the beams (Fig. 8.5(b)). The dimensions of the distribution
cross-section are given in DIN 19704.1 In a gate slot, the minimum distance
from the outer edge of the concrete should, as a rule, be not less than 150 mm.
The design of the beam is conventionally based on that of a beam on an elastic
foundation with the modulus of concrete having a value of C = 200 N/mm3.
The usual checks apply, such as the compressive stress of the concrete below

152
 Figure 8.3. Gate slot of a high head vertical-lift roller gate

Embedded parts
153
 Hydraulic gates and
valves

Figure 8.4. Lintel seal of a high

head slide gate

Figure 8.5. Concrete bearing

area of embedded parts

154
the transmission area and the shear stress in the beam. Some designers advocate a Embedded parts
corrosion allowance for all embedded parts of carbon steel. If the usual design
practice for hydraulic equipment is followed (that is, derating the permissible
working stresses) this can be accommodated within such an allowance.
At high head gates the jet emitted under the gate at small openings can
cavitate and become attached to the bottom of the conduit. It is therefore
common practice to line the invert for 2^4 m downstream of the sill, and to line
upstream between one-half and two-thirds of the downstream length.
In the situation where a hydraulic jump downstream of a gate is contained
within a concrete tunnel, considerable erosion to the invert can occur due to
recirculation of debris within the jump.2 Under such conditions it may be
necessary to extend the invert lining to protect the concrete.

References
1. DIN 19704 (1976): Hydraulic steel structures; criteria for design and calculation.
2. Lewin, J; Whiting, J R (1986): Gates and valves in reservoir low level outlets;
learning from experience, BNCOLD/IWES Conference on Reservoirs, Edinburgh,
Sept., p.77.

155
 9
Hydraulic considerations
pertaining to gates

The first section in this chapter outlines the basic data required to determine the
stage^discharge characteristics of gates. The discharge coefficients for radial
gates used in the equations are not directly comparable because the definition
of energy head varies. In some cases it may be the head upstream of the gate, in
others the head to the middle of the gate opening, and in drowned discharge
both the true energy head, the difference between upstream and downstream
water levels, and the downstream head enter the equation. Little information
has been published on the stage^discharge relationship of top-hinged flap gates.
Available data have been included in this chapter.
The next section deals with hydraulic downpull forces on vertical-lift gates.
This hydrodynamic effect is usually ignored when designing gates in open
channels, because under these conditions it is low and is absorbed by the margin
of hoisting force provided in gate installations. It becomes important for high
head gates. Because of the number of variables involved in determining
hydraulic downpull, calculations must be considered approximate. All research
in this field has been carried out on models representing vertical-lift gates,
although hydraulic downpull forces also act on radial gates.
Later sections draw attention to instability in a reach of a watercourse which
can be caused by the operation of a gate when there is limited ponded up water.
Problems can also arise when there is a change from 3D to 2D flow. This is the
condition of flood flow from a reservoir into a sluiceway. This type of problem
can be resolved by a physical model study. The occurrence of reflux at a
multigate installation and observed flow oscillations are described, as well as
the hysteresis effect of gate discharge during raising and lowering.
The final section begins with a discussion of vorticity at intakes. The
introduction of air into a conduit can cause severe pressure fluctuations at
control gates. Awareness that free vortices can occur should prompt
reconsideration of the design of an intake.
Cavitation and erosion are factors whenever flow velocities of the order of
13^15 m/s are reached or exceeded. Cavitation can affect gate slots and the invert
of tunnels. Information relevant to gates is included in this section and is
complementary to cavitation in valves which was discussed in Chapter 3.

157
 Hydraulic gates and Other hydraulic considerations complete the chapter: pressure coefficients
valves for gate slots, confluence of jets created by gates in parallel conduits, proximity
of two gates in parallel and air demand.

Flow under and over gates

Flow under gates
The coefficient of discharge of a radial gate installed in an open channel
watercourse to control water level and flow rate varies with the gate geometry,
the opening, upstream and downstream water levels. For submerged discharge
it is in the range 0.3^0.6 and for free discharge is in the range 0.5^0.7. Rouse1
gives a graph based on Metzler2 for one value of the gate radius to height of
pivot above the sill. This is reproduced by Lewin.3 Chow4 gives discharge
coefficients for radial gates where the radius of the gate is the denominator in
the non-dimensional functions, whereas Metzler2 uses the height of the pivot
above the gate sill.
For a flat vertical sluice gate Franke and Valentin5 developed a discharge
formula for free flow by measuring the pressure at the floor directly below the
gate lip, and relating this value to the geometry of the jet. The pressure can be
determined analytically and Franke and Valentin developed an expression for
the free discharge case. Young and Fellerman6 extended this for the general case
of submerged flow. In many cases a direct solution can be obtained from the
expression for the general case; in others, a solution has to be obtained by trial
and error.
Should the floor level drop appreciably downstream of the sluice gate, the
equation derived by Young and Fellerman cannot be applied and direct pressure
measurements are then required. Another limitation arises when the jet efflux at
the gate opening attains a subcritical value.
The general equation for discharge through an underflow gate can be
expressed as:
p
Q Cd  Go  W 2gH 9:1

where Q discharge
Cd coefficient of discharge
Go gate opening (denoted b in Fig. 9.5)
W gate width
g gravitational constant
H upstream water head

The variables affecting the discharge characteristics of a radial gate are shown
in Fig. 9.1.
An example of the coefficient of discharge map for free and submerged flow
under a radial gate based on Metzler2 is shown in Fig. 9.2.
Buyalski7 used similar maps to derive discharge algorithms which can be pro-
grammed into a computer for automatic control of gates or for calculating

158
 Hydraulic
considerations
pertaining to gates

Figure 9.1. Variables

affecting the discharge
characteristics under a radial
gate

discharge based on measurement of water levels and gate attitude, converted to

gate opening. The algorithms derived by Buyalski were based on experiments
carried out with gates having a lip seal of hard rubber rectangular section. The seals
were mounted upstream of the gate skin plate, whereas the preferred practice is to
locate the seal downstream of the skin plate. An upstream seal causes flow
disturbance and is not in accordance with the rule suggested by Lewin8 and Vrijer9
stating that flow separation should be arranged at the extreme downstream edge of
a gate to achieve flow conditions which are as steady as possible. Buyalski7 claims

Figure 9.2. Coefficient of discharge map for free and submerged flow under a radial gate for r/a = 1.5

159
 Hydraulic gates and that the experimental data show that different gate lip designs (even a minor
valves modification) can result in a 7% to +12% difference in the coefficient of
discharge Cd. The different seal configurations investigated were the rectangular
hard rubber section, a seal of musical note shape and no seal, that is, metal edge
contact. A seal of musical note shape should not be used as a lip seal as it can lead to
gate vibration.
The US Corps of Engineers' Hydraulic Design Criteria10 include graphs for
free discharge for ratios a=R 0.1, 0.5 and 0.9, where a is the height of the gate
pivot above the sluiceway floor and R is the radius of curvature of the skin plate.
The graphs are based on Toch,11 Metzler2 and Gentilini.12 However, the
geometry of many gates is outside the range of these ratios. The charts are useful
because they incorporate an ancillary graph to give adjustment factors when the
gate sill is raised above the floor of the channel.
The chart produced by the US Corps of Engineers13 for the coefficient of
submerged discharge is independent of the a/R ratio and is plotted for the ratio
of raised sill downstream submergence over gate opening. It appears that the
height of the sill above the approach bed is not an important factor in
submerged flow controlled by gates. One of the graphs (sheet 320^8) is
reproduced in Fig. 9.3.

Figure 9.3. US Corps of

Engineers' chart for the
coefficient of discharge under
submerged flow conditions

160
 For radial gates on spillway crests, the discharge through a partially open gate Hydraulic
can be computed using the same basic orifice equation: considerations
pertaining to gates
p
Q CA 2gH 9:2

where C coefficient of discharge

A area of opening
g gravitational constant
H head to the centre of opening
The coefficient is primarily dependent upon the characteristics of flow lines
approaching and leaving the orifice. In turn, these flowlines depend on the
shape of the crest, the radius of the gate and the location of the gate pivot.
The Hydraulic design criteria14 plot average discharge coefficients from
model and prototype data for several crest shapes and gate designs for non-
submerged flow. On the chart, the discharge coefficient is plotted as a function
of the angle  formed by the tangent of the gate lip and the tangent to the crest
curve at the nearest point of the crest curve. This angle is a function of the major
geometric factors affecting the flow lines of the discharge.
Figure 9.4 gives suggested design values for discharge coefficients of
0.67^0.73 for  from 50 to 110.
The sill of radial gates on spillway crests is usually located downstream of the
crest axis. Provision is made for placing stoplogs upstream of the gates.
Positioning the gate sill and the sill for the stoplogs close to the crest axis reduces
the overall height of the gates and the stoplogs in relation to the reservoir retention
level. A practice favoured by gate designers is thus to make the gap between
stoplogs and gate just sufficient for work to be carried out within the space with
the stoplogs located upstream of the crest axis and the gates downstream. Limited
test results suggest that within the normal practical dimensions of location of the
gate sill there is no effect on the discharge coefficient, but the crest pressure will be
affected.15 Slight negative pressures occur on the spillway crest for a Go/Hd ratio
of 0.4 or with the gate seat located on the crest axis. Crest pressures derived from
the charts in reference 15 are positive for all other Go/Hd ratios and gate seats
downstream of the crest axis.
The discharge coefficients in references 10, 13 and 14 are based principally on
tests with several bays in operation, and it is suggested that discharge
coefficients for a single bay would be lower because of side contraction. Limited
experimental data16 indicate that provided the gate piers project at least half a
bay width upstream of the gate sill and the approach channel is sensibly straight,
each sluiceway operates as if it was independent of the adjoining bays. The
requirement for projection of the piers is usually met because of the practice of
locating a bridge upstream of the gates for access purposes and for mounting a
gantry crane for handling of stoplogs.
To compute the discharge through each bay, the pier flow contraction
coefficient17 must also be considered.
The equation of discharge through a vertical-lift gate is the same as equation
9.1 for a radial gate, and using the same notation as Fig. 9.1 the variables
affecting the discharge characteristics are shown in Fig. 9.5 based on Rouse.1

161
 Hydraulic gates and
valves

Figure 9.4. US Corps of

Engineers' chart for the
coefficient of discharge for radial
gates on spillway crests for gate
lip angle from 50 to 110

Flow over gates

Radial gates have been designed to be overtopped. The top of the gate then acts
as a sharp crested weir. If the nappe impacts on transverse structure stiffener
beams downstream of the skin plate, it can result in gate vibration. When a radial
gate with an adjustable overflow section discharges over the gate, the flow
conditions are those of a broad crested weir and the shape of the crest is designed
to avoid negative pressure and flow separation. This also applies to hook-type
gates. At low elevation, bottom-hinged flap gates are subject to the flow
conditions of a broad crested weir which change with increasing elevation of
the gate to those of a sharp crested weir.
Drum, sector and bear-trap gates are all subject to varying discharge
coefficients throughout their raising and lowering motions.
The major load acting on an overflow gate or bottom-hinged section of a
gate is the hydrodynamic water pressure. The pressure is dependent on the
velocity distribution of the flowing water, the shape of the flow boundaries

162
 Hydraulic
considerations
pertaining to gates

Figure 9.5. Coefficient of

discharge map for free and
submerged flow under a vertical-
lift gate
and the separation of the water stream from, or its adhesion to, the gate surface.
The hydrodynamic pressure is of a pulsing character due to random velocity
pulsations,18,19 the instability of the flow separation point from the gate, and
water level oscillations induced by wave motion or a change in flow conditions.
Distribution of mean and fluctuating pressures for different operating
conditions is usually determined by a model study.20
An analytical method of determining the mean value of local hydrodynamic
pressure and a mathematical model of pressure pulsation for a non-submerged
bottom-hinged gate with circular curvature has been given by Rogala and
Winter.21
Reference 21 derives an equation to determine the hydrodynamic pressure at
an arbitrarily chosen point on an overflow hinged gate which is neither
submerged nor fully aerated. The pressure at a point depends on the geometric
parameters of the gate, its position and the flow discharge. The equation is:

p Ho y
0:85  F  expW 9:3
gR R

where p mean pressure at the point examined on the gate surface

 water density
g gravitational constant
R radius of gate curvature
Ho heights of the energy head above the gate hinge
y vertical distance above gate hinge
F v2/(gR) where v mean horizontal velocity of flow over the
gate edge

163
 Hydraulic gates and
valves

Figure 9.6. Diagram of water

flow over a hinged gate

W an exponent dependent on the gate inclination angle  and

on the angle co-ordinate  of the point examined on the gate,
as well as Ho y and ho, calculated from equations (9.4) and
(9.5)
W 1:22Y 0:7 for  < =2 and  < =2  (9.4)
W 7:1F0:5 Y 0:4 in the remaining cases (9.5)
Y Ho y=R=ho =R 1 (9.6)

Figure 9.6 shows the flow of water over a hinged gate.

Stage^discharge relationship of a top-hinged flap gate

A theoretical treatment of the stage^discharge relationship of a rectangular flap
gate was established by Pethick and Harrison.22 This was derived from two
different theoretical concepts which yielded sensibly similar results.

(a) Free flow

For free flow the flap gate relationship was expressed as a plot of three
dimensionless parameters (Fig. 9.7).
where h upstream head
L distance from hinge to the centre of gravity of the flap
m mass per unit width of flap gate
g gravitational constant
q discharge per unit width of flap gate
L depth of flap
p density of fluid
For a given gate, the mass parameter 2m/pL2 is constant and so the stage^
discharge relationship is read from the ordinate in Fig. 9.7 and h/L values on the
appropriate vertical line. For design purposes the mass parameter can be

164
 Hydraulic
considerations
pertaining to gates

Figure 9.7. Rectangular flap

gate ^ free flow (after Pethick
and Harrison22)

determined if the values q, h and L are known. In practice, h is likely to be

variable and design will have to be carried out on a trial and error basis.
Pethick and Harrison22 have pointed out that the Froude number upstream
of the gate:
p
q= gh3
is unity at the point where the h/L curve intersects the vertical axis.

(b) Drowned flow

Drowned flow leads to a four-parameter relationship shown in Fig. 9.8. It is
plotted for h/L values of 0.6 and 0.8.
The theoretical analysis does not give a solution over a small range close to
the free flow limit. Visual observation of model tests suggest that there is a
discontinuity due to unstable flow in this range.

165
 Hydraulic gates and The figures show that throughout the drowned flow regime for a constant
valves upstream water level, discharge reduces very rapidly as tailwater increases;
h dd is the effective head across the gate and, as expected, q is proportional to
p
h dd

where h upstream head

dd tailwater head.

(c) Empirical stage^discharge relationship

Tests carried out on circular flap valves at the State University of Iowa, USA to
establish loss of head through flap valves derived the following empirical
formula:
 
4V 2 1:15V
L  exp p 9:7
g d

where L loss of head in feet

V velocity of flow through the gate in feet per second
g gravitational constant
d diameter of the outlet in feet

It is assumed that this formula applies only to free discharge, although this was
not stated.

Hydraulic downpull forces

Hydraulic downpull forces under a gate are due to the reduction in pressure
caused by discharge under the gate. For gates in open channels this is the main
factor affecting downpull and is a function of the gate geometry. Approach flow
effects can slightly modify the downpull effect.
In tunnel gates and particularly in high head gates, downpull is significantly
affected not only by the geometry of the gate bottom, but also by the rate of flow
passing over the top of the gate through the gate well, which can exert a major
effect on the magnitude of the downpull.
There are two states of flow at a high level gate: free flow, in which the space
downstream of the gate is filled with air; and submerged flow, in which that
space is submerged and pressurised (Fig. 9.9).
The flow conditions for a typical arrangement of a gate partially withdrawn
into a well are also illustrated in Fig. 9.9.
The primary part of the downpull stems from the difference between the
integrated distributions of piezometric head on the top of the gate and the
bottom surface, and may be expressed as:

166
 Hydraulic
considerations
pertaining to gates

Figure 9.8. Rectangular flap gate ^ drowned flow (after Pethick and Harrison22)

167
 Hydraulic gates and
valves

Figure 9.9. Diagram of tunnel gate

under submerged flow conditions
(after Naudascher et al.23)

F T B BdV 2 =2 9:8

where F downpull force

T downpull coefficient at the top of the gate
B downpull coefficient at the bottom of the gate
B width of gate
d depth of gate
 density of water
V velocity in the contracted section of the jet

The geometric parameters which affect downpull at the bottom of the gate
are:
(i) the ratio of gate opening to tunnel height which can be expressed as y=y0 ,
the percentage opening of the gate
(ii) the angle of gate bottom 
(iii) the ratio of the radius r to the depth of the gate r/d
(iv) the ratio of the projection of the gate lip e to the depth of the gate t/d
(v) the relative conduit height yo/d.

168
 It is usual to make the angle  45 for stiffness of construction and because Hydraulic
it is favourable from considerations of downpull force, although one model considerations
pertaining to gates
study25 showed that a reduction in downpull of 5% can be achieved by
increasing  to 50.
A radius r at the transition is important and prevents flow separation at the
lower end of the vertical face. Design practice is to make e a minimum.
Further downpull over and above F is due to the pressure difference acting
on the horizontal projection of the top seal of the gate or the projection of the
extended skin plate.
When the gate lip approaches the tunnel ceiling, the downpull may become
negative, that is it may transform into an upthrust. Under these conditions it
could inhibit safe gate closure.
Figures 9.10 and 9.11 show downpull conditions for tunnel-type vertical-
lift gates versus gate openings. The graphs incorporate the results of two
different investigations. When using these coefficients an estimate of the
contracted jet issuing from beneath the gate must be made. The discharge
coefficient of the jet varies with gate opening and gate geometry and is
superimposed on Fig. 9.10.
Figure 9.12 illustrates the dependence of the bottom downpull coefficient on
gate geometry, and Fig. 9.13 shows its dependence on relative conduit height.
Naudascher et al.23 have concluded that gate slots do not affect hydraulic
downpull forces.
At intake gates the approach flow conditions can cause strong variations of
downpull and discharge coefficients.25 Piers at the intake, and trashrack grids a
short distance upstream of the gate, can cause flow separation or alter the
turbulence characteristics of the flow regime near the gate compared with free
stream turbulence.
Since there are several variables involved in hydraulic downpull forces, the
values of downpull coefficients shown in the graphs should be considered
indicative as opposed to actual design values. It may sometimes be expedient

Figure 9.10. Resultant

downpull and discharge
coefficients versus gate opening
(after Weaver and Martin24)

169
 Hydraulic gates and
valves

Figure 9.11. Top and bottom

downpull coefficients versus gate
opening (after Weaver and
Martin24)

Figure 9.12. Dependence of

bottom downpull coefficient on
gate geometry for a ratio of
conduit height to depth of the
gate, yo =d 6 (after Thang
and Naudascher25)

Figure 9.13. Dependence of

bottom downpull coefficient on
relative conduit height y=d
(after Thang and
Naudascher25)

170
and cost-effective to assume subatmospheric pressure at the gate bottom and Hydraulic
oversize the servo-motor accordingly. considerations
pertaining to gates

Limited ponded-up water

Opening a gate which controls a limited reach of a river will send a wave upstream.
When this wave is reflected it can, in turn, cause a disturbance of the gate. The
motion can amplify and cause serious instability, illustrated in Fig. 9.14. An
instance of this type of instability in a radial automatic gate is cited by Lewin.8
A change in the discharge under or over a gate will similarly cause a wave to travel
upstream, and if it is reflected it will register a false increase in water level. This can
actuate the control system and initiate a further opening of the gate or gates.

Three-dimensional flow entry into sluiceways

Where the flood flow from a reservoir into sluiceways is not trained by an
approach channel, cross-flow will occur (Fig. 9.15). This can result in eddy
shedding at the piers and turbulent conditions at the gate face.
Kolkman26 quotes an example of self-exciting wave oscillations in the
upstream basin of a sluice, experienced during a model study carried out in the
Delft Laboratory. When, for instance, only six openings out of ten discharged
while the others (especially the outer ones) were closed by gates, a transverse
wave oscillation occurred, resulting in a wave amplitude in the prototype of
2 m near the closed gates. The transverse flow component related to the water
level oscillations interacted, most probably, at the point where the main flow
separated from the side walls.

Reflux downstream at a pier and flow oscillation

In multigate operation, the drowned discharge from one gate when the
adjoining gate is closed can reflux into the quiescent stilling basin and cause a
periodic disturbance. This can act on submerged structural members
downstream of the gate and can result in gate oscillations. In an installation of
two radial automatic gates, the pier between the gates extended 15.75 m
downstream of the sill and had to be extended by an additional 8 m to prevent
gate oscillation under the conditions previously described.
Flow oscillation has occurred when the water surface differential between
pool and tailwater level is relatively small and the flow is controlled primarily

Figure 9.14. Wave action due

to limited ponded-up water

171
 Hydraulic gates and
valves

Figure 9.15. Cross-flow due to

change from three-dimensional
to two-dimensional flow

by the tailwater.27 It caused bouncing of radial gates due to surges of flow which
moved back into the gate bay and struck the bottom girder of the gate. The fluid
flow was strong enough to lift the gates, causing the bouncing phenomenon.
Extension of the piers would have reduced the load on the gates. Flow
oscillation has also been noted in other model studies.28

Hysteresis effect of gate discharge during

hoisting/lowering
When flow under a gate becomes detached from the gate lip it accelerates and
drops away from the gate. To regain control of the flow by the gate it has to be
lowered further. Contact of the gate lip with the face of the water will initially
have no effect. A slight further immersion will cause an afflux at the gate, and an
increase in upstream water level higher than the level prior to the upstream flow
becoming detached. Where this sequence of events affects level measurement
for a control system, it can result in unstable operation.

Hydraulic considerations pertaining to gates in

conduit
Vorticity at intakes
When air is introduced into a conduit, severe pressure fluctuations can occur at a
control gate due to the build-up of stagnated air under high pressure at the
conduit crown upstream of the gate. The formation of free vortices at intakes
must therefore be avoided.
Two factors principally determine the formation of vorticity at an intake:
submergence and circulation of the approach flow. Circulation is the primary
parameter influencing submergence.29,30 Gulliver et al.31 have suggested that, as
a first approximation prior to a model study, two design parameters should be

172
used: the dimensionless submergence S=D 0.7^4 (Fig. 9.16) and Froude Hydraulic
number Fr 0^0.5. Intakes at existing stations and model studies within these considerations
pertaining to gates
limits experienced no vortex problems. It is presumed that the data to support
this were obtained under reasonably straight approach flow conditions.
Gordon32 gives the minimum submergence of the top of the gate opening as
determined by the formula:

S C  v  D1=2 9:9

where C is a constant of suggested value 0.3 for symmetrical approach flow. The
units of C are sm1/2. The formula was derived from observations of 29
prototype intakes.
Intake screens will, by streaming the flow, permit lower submergence of
inlets before the onset of formation of free vortices.33,34
Antivortex devices have been used, such as long approach channel walls over
the intake, or the hooded inlet developed by Song35 and Blaisdell and Donelly.36

Cavitation and erosion

The causes of cavitation and its effect on engineering structures have been well
documented. The immense damage to the tunnels at Tarbela, Pakistan in 1974
has been attributed to cavitation (Fig. 12.4). With high velocities and the
potential for sheared flow adjacent to conduit boundaries, regions of low
pressure can be set up with pressures close to that of incipient cavitation. Small
surface irregularities can be sufficient to drop the pressure to a level which will
initiate cavitation.
Considerable cavitation damage was reported by Wagner37 due to high
velocity flow up to 41 m/s. This was due to poor alignment of the liner joints,
projecting joint welds and minor ridges and depressions in the paint coating.
Offsets as little as 0.8 mm into the flow produced marked damage and the degree
of damage increased with larger offsets. Depressed surface offsets of 6 mm
produced paint removal and minor pitting. Laboratory studies were conducted
by Ball38 to establish the velocity^pressure relationship for incipient cavitation
at offsets with rounded corners and sloping surfaces protruding into the flow.
These may be used as guidelines for establishing tolerances for surface
irregularities of linings downstream of gates. The greater resistance of stainless
and high nickel steels to cavitation damage is documented by Wagner37 and

Figure 9.16. Vortex

formation at a submerged
intake

173
 Hydraulic gates and others, although damage to stainless steel occurred downstream of the
valves regulating gates at the low level outlet at the Dartmouth Dam in Australia.39
Cavitation conditions can arise at pier nosings where piers divide several gate
passages, especially under asymmetric gate operating conditions. Anastassi40
gives an equation for an elliptically shaped pier nosing to reduce pressure
fluctuations for symmetrical flow, and a modified equation for a small elliptical
shape for asymmetric flow conditions. The paper also notes that the taper of the
pier must be gradual to prevent flow separation.
Serious cavitation can be caused by high pressure flow through small gaps
at seals and at gates which are just closing or opening. Tests have shown that
gaps of less than 0.1 mm are safe for short periods, whereas gaps of more than
2 mm can cause serious erosion, apart from the possibility of inducing gate
vibration. Since cavitation damage is time dependent, high head gates should
not be kept in operation at small openings. Minimum openings should be not
less than 100 mm. This will also minimise erosion downstream of the sill (see
Chapter 10).

Gate slots
Vertical-lift gates of the roller or slide type require recessed slots in abutments or
piers for the movement of the gate guide rollers or slides. The flow of water
across the slots causes flow separation at the upstream edge of the slot and
reattachment on the downstream side. Eddies are set up within the slots and
vortices are formed. Under conditions of high velocity flow cavitation can
occur within gate slots.
Flow conditions due to gate slots are influenced by the upstream and
downstream edge shape and the cavity depth to length ratio (Fig. 9.17).
Radiusing the upstream edge increases the flow into the cavity hence
increasing energy dissipation, so this should be avoided. A radius on the
downstream edge reduces energy dissipation. A single, stable vortex forms in
cavities where the d/w ratio is close to unity. This results in low losses. Between
d/w ratios of 0.2 and 0.8 circulation is unstable, with periodic disturbances
influencing the main flow. Loss coefficients for sharp edged gate slots have a
minimum value of about 0.01 with d/w ratios of 0.5; this rises to 0.03 for
d=w 2:5. (The loss coefficient is defined as the ratio of head loss V2/2g where
V is the mean velocity.)
The flow past the gate slot results in a reduction in pressure on the conduit
wall immediately downstream from the slot. Cavitation can occur within the
slot or downstream from the slot when high velocity flow occurs and there is
insufficient pressure in the region of the slot.
Cavitation in valves was discussed and varying intensities of cavitation were
differentiated in Chapter 3.
Incipient cavitation is the onset of the phenomenon and usually occurs
intermittently over a restricted area. Noise is slight and there is no damage
except at isolated local conditions, such as a step.
The next stage is critical cavitation where noise and vibration are acceptable
and damage will occur only after long periods of operation. This is usually
adopted as a design criterion in gate and valve installations.

174
 Hydraulic
considerations
pertaining to gates

Figure 9.17. Flow in gate slots

A further stage is incipient damage when pitting occurs after short periods of
operation and is accompanied by a high noise level.
Choking cavitation occurs when the outlet pressure is lowered to vapour
pressure. At this stage the flow is unaffected by the downstream pressure, and
flow and pressure loss relationships no longer apply. Close to choking, noise,
vibration and damage due to pitting are at a maximum.
Specialist literature should be consulted for super cavitation, the stage
beyond choking cavitation.
Cavitation of gate slots was investigated by Ball41 and Galperin.42 May43
reviews cavitation in hydraulic structures and deals extensively with cavitation
due to gate slots.
The cavitation parameter s of a slot is given by:

h i hv
s 9:10
v2 =2g

where hi head
hv vapour head
v flow velocity
g gravitational constant

Cavitation can be initiated by decreasing hi or increasing v. Therefore the lower

the cavitation parameter, the greater the intensity of cavitation.
If incipient cavitation is the design criterion and if the incipient cavitation
parameter for a gate slot si is known, the flow velocity at the gate slot can be
calculated. A greater velocity will move the condition into the region of
incipient damage cavitation.
In gate slots there are a number of geometric factors which affect the incipient
cavitation parameter. These can be combined to yield an overall value of si
using equation (9.11). Figure 9.18 shows the factors C1, C2, C3 and Kis and their
dependence on the geometry of a gate slot. The incipient cavitation parameter

175
 Hydraulic gates and for a gate slot is:
valves
si C1 C2 C3 Kis 9:11
where Kis is the value of incipient cavitation at the upstream or the downstream
edge of the gate slot.
In the graph of C3,  is the thickness of the boundary layer which can be
calculated from the boundary layer equation for smooth turbulent flow.43 Since
gate slots on high head gates are long (w) in relation to the boundary layer
thickness (), using a value for C3 of 1.4 will result in safe designs.
The results are applicable to a fully open gate and when flow is approximately
two-dimensional. The latter condition may not apply to an intake gate.
Galperin42 also gives data for vertical-lift gates which are partially open.
Typical values of si for gates discharging under submerged conditions can vary
between 1.0 at a gate opening of 35% and 2.5 at 90% open. For free discharge the
range is 0.3^1.0.

Figure 9.18. Factors for

incipient cavitation parameters
of gate slots (after May43)

176
 Another important variable is the conduit geometry downstream of the slot. Hydraulic
The low pressure conditions on the downstream edge of the gate slot can be considerations
pertaining to gates
improved to some degree by offsetting the downstream edge of the slot and
returning gradually to the original conduit wall alignment. Figure 9.19 shows
pressure coefficients for gate slots with and without downstream offsets and with
downstream rounded corners. The coefficients were computed using the equation:

Hd CHv 9:12

where Hd pressure difference from reference pressure

C pressure coefficient
Hv conduit velocity head at reference pressure

Figure 9.17 illustrated flow patterns in the plane of the slot. In addition,
forced vorticity occurs in the vertical direction resulting in a complex 3D flow
when the gate is in the open or partially open position. The pressure coefficients

Figure 9.19. Pressure

coefficients for gate slots with
and without downstream offsets

177
 Hydraulic gates and given in Fig. 9.19 therefore show only the improvement which can be achieved
valves by providing a downstream offset. There are likely to be significant resultant
hydraulic forces on the gate rail tending to lift it from its mounting, due to high
stagnation pressures which can develop near the downstream edge of the slot
where the gate rail is fastened.
Ball41 showed that deflectors at the upstream edges of slots produce an
ejector action which lowers pressures at the slot far below the reference pressure
and will induce cavitation. A very large deflector which causes a heavy
contraction can be used successfully, and is the basis of the design of jet-flow
gates (Fig. 2.64).
Some of the conclusions in the paper by Ball41 can be used as a guide for the
design of gate slots. Offset corners of slots and a variable rate of convergence are
most desirable from hydraulic considerations. Arcs used in this design should
have radii in the range of about 100^250 times the offset of the downstream
corner. Ellipses can also be used with excellent results.
The upstream corners of the gate slots should not be rounded or notched as
both are detrimental to pressure distribution.
The widening of slots permits more expansion of the jet into the slot, tending
to increase the contraction at the downstream corner. However, pressure
conditions are acceptable for a wide range of slot width-to-depth ratios in
designs using offset corners with converging walls. This is particularly true
for the 24:1 convergence and the long radius curved convergence.
Sharp downstream corners of gate slots should always be offset away from
the flow. The offset of the downstream corner of a gate slot should be small
and related to the slot width. Within reasonable limits, this offset is not
critical.
Abrupt offsets into the flow and irregularities in flow surfaces are particularly
troublesome. Offsets of less than 3 mm will cause damage. It is extremely
important to provide smooth continuous surfaces downstream from gates
operating under high heads.

Gate conduits
The investigation of the bottom outlet of the San Roque Dam in the
Philippines40 demonstrated severe turbulent flow separation upstream of the
control gate installation, of the type illustrated in Fig. 9.20. This was of a
periodic nature causing peak pressure surges. The geometry of the approach
section of the tunnel and the transition to the conduits containing the gates
affected the pressures in the gate chamber.
The paper also drew attention to the desirability of the conduit's cross-
sectional area at the point of gate discharge being less than that of the approach
tunnel, to avoid subatmospheric pressures which could limit the opening of the
control gate.
Where two or more gates are to be installed in parallel it is necessary to
consider effects due to the confluence of the jets downstream and potential
combinations of the jets downstream and of asymmetrical flow. Problems can
result from flow separation, unstable flow, excessive bulking, oblique flow and
cross waves.

178
 Hydraulic
considerations
pertaining to gates

Figure 9.20. Bottom outlet of

the San Roque Dam, showing
flow separation and turbulence
within the chamber upstream of
the gate fluidways (after
Anastassi40)

Koch,44 in the model study of the bottom outlet at the Randenigala Project in
Sri Lanka, found that a downstream length of 8 m was insufficient for the
dividing wall. With velocities up to 43.2 m/s, flow was separating from the
curved face of the dividing wall. In order to guard against low pressures which
were likely to result in cavitation, it was necessary to extend and taper the wall by
35 m and incorporate facilities for air entrainment.
The bottom outlet of the Mrica Hydroelectric Project on the Island of Java,
Indonesia45 has twin conduits, each housing a control and emergency closure
gate of the slide type. The dividing wall extends 8.8 m downstream of the
control gate's sill with no physical flow separation beyond the wall. The
maximum jet velocity is 33.55 m/s.

Trajectory of jets due to floor offsets

The model study of the drawdown culvert control structure for Mrica45 showed
that the deflectors of the aeration slots in the invert downstream of the gates
caused the trajectory of the jet leaving the step to be thrown up to strike the
tunnel roof. Omitting the deflectors and modifying the step led to an acceptable
trajectory accompanied by a small decrease in the volume of entrained air.

179
 Hydraulic gates and Proximity of two gates
valves
The proximity of two gates in a conduit can cause vibration of the downstream
gate when the control gate is in an intermediate position and the guard gate is
lowered. The jet from the guard gate gives rise to alternating forces on the
control gate, and in some combinations of gate positions there can be turbulent
recirculation of flow between the gates.
Nielson and Pickett46 recorded vibration due to this cause and Petrikat47
mentions curing a similar problem by using jet dissipators which broke up the
discharge jet from the guard gate. Naudascher48 also highlights the danger of
vibration due to two consecutively positioned gates in conduit. These problems
are generally transitory, however, if considered acceptable they must still be
confined to a level which does not damage the structure.

Air demand
Flow under a gate
When a vertical-lift gate in a conduit is opened and the downstream section of
the conduit contains no water, a demand for air arises due to entrainment of air
in the issuing jet.
The total air demand consists of two different parts: air entrainment in the
water flow as bubbles or larger air pockets in the air/water transition region, and
air flowing above the transition zone because of the drag of the flowing mixture.
At initial gate openings, the issuing jet is accompanied by spray which entrains a
high proportion of air.
To explain differences in air demand, flow has been classified.49 The total air
demand for free surface flow in conduits is not normally at a maximum when
gates are fully open. Often two maxima exist; one for very small gate openings,
when spray flow occurs at 4^8% of gate opening, and a second one usually larger
than the first when the gate opening is between 40^70%.
If a hydraulic jump occurs further air entrainment will take place. Kalinske
and Robertson50 expressed this in terms of the ratio of air flow to water flow. Air
entrainment without jumps has been investigated by a number of
researchers.49,51
The suggested design assumption52 is:

 0:03Fr 11:06 9:13

where  ratio of air flow to water

p flow
Fr Froude number V= gy
V flow velocity at the vena contracta
g gravitational constant
y water depth at the vena contracta

The contraction coefficient for a gate with a 45 lip is 0.8. The above formula
results in significantly more conservative volumes of air than those arising from
the investigation of Kalinske and Robertson.50
The air admission pipes should be designed for velocities of not more than

180
40 m/s to prevent excessive pressure loss due to flow resistance in the ducts as Hydraulic
well as entrance and exit air flow losses. These cause subatmospheric pressure considerations
pertaining to gates
conditions in the water conduit.
Air flow losses can be calculated from data in the CIBSE Guide.53

Flow over a gate

The requirement to vent the nappe of an overflow gate was stated in Chapter 2.
Air is entrained in the falling water and reduces the pressure under the gate
unless it is vented. The subatmospheric conditions resulting from air evacuation
cause the nappe to oscillate and the water level to fluctuate. This can lead to
severe gate vibration.
Venting is effected by flow dividers which break up the nappe locally and
admit air through the openings created by divided flow. Ducts leading from
pier level to the underside of a gate (Fig. 2.27) are another method of venting.
In most cases both means of admitting air have to be used together.
In the extreme case pressure under the gate can reach 10.33 m gauge. This
would exert a negative pressure of 10.33 Mpa/m2 on the underside of the gate.
Air entrainment is proportional to the velocity head of the nappe. The air
demand can be expressed as:

QA Q

where QA air demand

Q flow over the gate
 a coefficient depending on height of fall of the nappe, h4,
depth of the nappe, h3, and the Froude number Fr of the
nappe

The depth of the nappe is approximately 0.6h2, where h2 is the actual head
above the gate lip. Therefore:
 
v2
h3 0:6 h1 1
2g
where v1 is the velocity of the approach flow to the gate and h1 is the energy head
above the lip of the gate.
The Froude number of the nappe is:
Fr v2 =gh3
where v2 is the velocity of the overflow.
Fig. 9.21 shows the coefficient of air demand against the ratio h4/h3 for two
ranges of Froude numbers.
The air ducts can be sized in a similar manner to those required to satisfy the
air demand for underflow gates. Since ducts for overflow gates are usually short
compared with those for underflow gates, the duct entry and exit losses will be
more significant when the friction loss of the air supply system is calculated.
They should therefore be considered.

181
 Hydraulic gates and
valves

Figure 9.21. Coefficient of air

demand for an overflow gate

The ducts should be arranged so that the air supply in any position of
overflow of the gate is not blocked by the downstream water level, and that
the air is admitted under the gate. In order to achieve this, the outlets of the
air supply ducts are staggered as shown in Fig. 2.27. It may even be necessary
to stagger the termination of the air supply pipes relative to one another in
opposite sluiceway walls. It is usual practice to screen the outlets of the vent
ducts. The screens must be set back from the face of the sluiceway so that
they do not damage the side seal of the gate when it moves over the duct
outlets.

References
1. Rouse, H (1949) editor: Engineering hydraulics, Proc. 4th Hydr. Conference, Iowa
Institute of Hydraulic Research, John Wiley, p. 540.
2. Metzler, D E (1948): A model study of tainter gate operation, MS thesis, State
University of Iowa, in Proc. 4th Hydr. Conference, Iowa Institute of Hydraulic
Research, editor Rouse, H.
3. Lewin, J (1980): Hydraulic gates, Journ. I.W.E.S., 34, No. 3, p. 237.
4. Chow, V T (1959): Open channel hydraulics, McGraw-Hill.
5. Franke, P G; Valentin, F (1969): The determination of discharge below gates in case
of variable tailwater conditions, Journ. Hydr. Res., 7, No. 4.
6. Young, L R; Fellerman, L (1971): Toome sluices calibration tests, B.H.R.A., report RR
1105, Jul.
7. Buyalski, C P (1983): Canal radial gate discharge, algorithms and their use, Proc.
Speciality Conf. on Advances in Irrigation and Drainage: Surviving external pressures,
Jackson, USA, July, editors Borelli, J; Hasfurther, V R; Burman, R D, New York,
USA, A.S.C.E., pp. 538^545.
8. Lewin, J (1983): Vibration of hydraulic gates, Journ. I.W.E.S., 37, 165^182.

182
 9. Vrijer, A (1979): Stability of vertically movable gates, 19th I.A.H.R. Congress, Hydraulic
Karlsruhe, paper C5. considerations
10. US Army Corps of Engineers: Tainter gates in open channels discharge coefficients (free pertaining to gates
flow), Hydraulic Design Criteria, sheets 3204 to 3207.
11. Toch, A (1952): The effect of a lip angle upon flow under a tainter gate, Masters thesis, State
University of Iowa, Feb.
12. Gentilini, L B (1947): Flow under inclined or radial sluice gates ^ technical and
experimental results, La Houille Blanche, 2, p.145.
13. US Army Corps of Engineers: Tainter gates in open channels ^ discharge coefficients
(submerged flow), hydraulic design criteria, sheets 320-8 to 320-8/1.
14. US Army Corps of Engineers: Tainter gates on spillway crests ^ discharge coefficients,
Hydraulic design criteria, sheets 311-1 to 311-5.
15. US Army Corps of Engineers: Tainter gates on spillway crests ^ crest pressures, hydraulic
design criteria, sheets 311-6.
16. Milan, D; Habraken, P (1984): Kotmale, report on spillway radial gate, model tests,
Neyrpic Thermohydraulic and Hydroelasticity Laboratory. (unpublished).
17. US Army Corps of Engineers: Gated overflow spillways, pier contraction coefficients,
hydraulic design criteria, sheets 111-5, 111-6.
18. Naudascher, E; Locher, F A (1974): Flow-induced forces on protruding walls, Proc.
A.S.C.E., Journ. Hydr. Div., Vol. 100, HY2, paper 10347, Feb.
19. White, F M (1979): Fluid mechanics, McGraw-Hill, New York, USA.
20. Muskatirovic, J (1984): Analysis of dynamic pressures acting on overflow gates,
I.A.H.R. Symposium on scale effects in modelling hydraulic structures, Essingen am
Neckar, Germany, Sept.
21. Rogala, R; Winter, J (1985): Hydrodynamic pressures acting upon hinged-arc
gates, Proc. A.S.C.E., Journ. Hydr. Engineering, 111, No. 4, Apr.
22. Pethick, R W; Harrison, A J H (1981): The theoretical treatment of the hydraulics
of rectangular flap gates, 19th I.A.H.R. Congress, Karlsruhe, subject B (c), paper 12.
23. Naudascher, E; Rao, P V; Richter, A; Vargas, P; Wonik, G (1986): Prediction and
control of downpull on tunnel gates, Proc. A.S.C.E., Journ. Hydr. Engineering, 112,
No. 5, May.
24. Weaver, D S; Martin, W W (1980): Hydraulic model study for the design of the
Wreck Cove control gates, Canadian Journ. of Civ. Eng., 7 No. 2.
25. Thang, N D; Naudascher, E (1983): Approach-flow effects on downpull of gates,
Proc. A.S.C.E., Journ. Hydr. Eng., 109, No. 11, Nov.
26. Kolkman, P A (1984): Phenomena of self excitation, in Developments in hydraulic
engineering ^2, editor Novak, P, Elsevier Applied Science Publishers.
27. Hite J E; Pickering, G A (1983): Barkley Dam spillway tainter gate and emergency
bulkheads, Cumberland River, Kentucky; hydraulic model investigation, US Army Engineer
Waterways Experiment Station, Vicksburg, Miss., technical report HL-83-12,
Aug.
28. Grace, J L (1964): Spillway for typical low-level navigation dam, Arkansas River,
Arkansas; hydraulic model investigation, US Army Engineer Waterways Experiment
Station, Vicksburg, Miss., technical report 2-655, Sept.
29. Daggett, L L; Daggett K G H (1974): Similitude in free-surface vortex formations,
Proc. A.S.C.E., Journ. Hydr. Div., 100, Nov.
30. Anwar, H O; Weller, J A; Amphlett, M B (1978): Similarity of free-vortex at
horizontal intake, Journ. Hydr. Research, 16, No. 2.
31. Gulliver, J S; Rindels, A J; Lindblom, K C (1986): Designing Intakes to Avoid
Free-Surface Vortices, Water Power and Dam Construction, Sept, pp. 24^28.
32. Gordon, J L (1970): Vortices at intakes, Water Power, Apr.

183
 Hydraulic gates and 33. Ables, J H (1979): Vortex problem at intake Lower St Anthony Falls Lock and Dam,
valves Mississippi River, Minneapolis, Minnesota, US Army Engineer Waterways Experiment
Station, USA, technical report HL-79-9, May.
34. Ziegler, E R (1976): Hydraulic model vortex study Grand Coulee third powerplant,
Engineering Research Centre, Bureau of Reclamation, Denver, Colorado, USA,
Feb.
35. Song, C C S (1974): Hydraulic model tests for Mayfield Power Plant, University of
Minnesota, St Anthony Falls Hydraulic Laboratory, project report No. 148, Apr.
36. Blaisdell, F W; Donnelly, C A (1958): Hydraulics of closed conduit spillways: part X, the
hood inlet, Agricultural Research Service, St Anthony Falls Hydraulic Laboratory,
technical paper 20, Series B.
37. Wagner, W E (1967): Glen Canyon Dam diversion tunnel outlets, Proc. A.S.C.E.,
Journ. Hydr. Div., 93, HY6, Nov, pp.113^134.
38. Ball, J W (1963): Construction finishes and high-velocity flow, Proc. A.S.C.E.,
Journ. Constr. Div., 89, (CO2).
39. Dickson, R S; Murley, K A (1983): Dartmouth Dam low level outlet aeration
ramps, Ancold Magazine.
40. Anastassi, G (1983): Besondere Aspekte der Gestaltung von Grundablassen in
Stollen (Design of high-pressure tunnel outlets), Wasserwirtschaft, 73, 12.
41. Ball, J W (1959): Hydraulic characteristics of gate slots, Proc. A.S.C.E., Journ. Hydr.
Div., 85, HY10, 81^114.
42. Galperin, R (1971): Hydraulic structures operation under cavitation conditions,
14th I.A.H.R. Congress, Paris, Vol. 5, pp.45^48.
43. May, R W P (1987): Cavitation in hydraulic structures:occurrence and prevention, Hydraulic
Research, Wallingford, report SR79.
44. Koch, H J (1982): Schustrahlzusammenfhrung bei einem Grundablass mit
Nemeneinanderliegenden Segmentschutzen (Confluence of two jets created by
two parallel segment gates of a bottom outlet), Wasserwirtschaft, 72, 3.
45. Bruce, B A; Crow, D A (1984): Mrica Hydroelectric Project: hydraulic model study of the
culvert control structure, B.H.R.A., report RR2325.
46. Nielson, F M; Pickett, E B (1979): Corps of Engineers experiences with flow-
induced vibrations, 19th I.A.H.R. Congress, Karlsruhe, paper C3.
47. Petrikat, K (1979): Seal vibration, 19th I.A.H.R. Congress, Karlsruhe, paper C14.
48. Naudascher, E (1972): Entwurfskriterien fr Schwingungssichers Talsperren-
verschlsse (Design criteria for avoiding vibration of high head gates),
Wasserwirtschaft 62, 112.
49. Sharma, H R (1973): Air demand for high head gated conduits, University of Trondheim,
Oct.
50. Kalinske, F; Robertson, R A (1943): Closed conduit flow: Symposium on
entrainment of air in flowing water, A.S.C.E., Transactions, paper 2205.
51. Wunderlich, W (1961): Beitrag zur Belftung des Abflusses in Tiefauslssen
(Commentary on air demand in conduit gates), Technische Hochschule, Karlsruhe.
52. US Army Corps of Engineers: Air demand, regulated outlet works, Hydraulic Design
Criteria, Sheet 050-1.
53. Chartered Institute of Building Services Engineers: Guide, C4-48 and C4-49, Figure
C4.3 Air flow in round ducts, Figure C4.4 Air flow in rectangular ducts.

184
 10
Gate vibration

Gate vibration, when it occurs, can be a serious problem. It can result in

structural damage or restrict operation at certain gate openings. In some cases
vibration of a gate will occur under specific hydraulic conditions which may
only become manifest years after commissioning of the installation. Even when
these conditions have been identified it may not be easy to reproduce them so
that they can be investigated. Apparent steady-state conditions may be subject
to a minor hydraulic disturbance which overcomes the damping forces acting
on the gate and initiates an unstable motion, giving rise to oscillations of
increasing amplitude.
This chapter is intended as an introduction to the subject and offers some
guidance on design features which will prevent vibration.
Many gates incorporate elements which are likely to result in vibration but
continue to operate satisfactorily. One possibility is that disturbing forces are of
low magnitude and are damped out; this can be the case with small gates in river
courses. Also, since gates are designed for long return period events, it may be
that the conditions which could cause vibration have not yet occurred. It does
not follow that gates of similar design will be equally satisfactory at a higher
head or scaled up in size.
Most of the research papers on gates deal with vibration problems. Vibration
is perhaps the most frequent cause of malfunction of gates.

Types of gate vibration

Gate vibrations can be classified into three types.1
(a) Extraneously induced excitation which is caused by a pulsation in flow or
pressure which is not an intrinsic part of the vibrating system (the gate).
(b) Instability induced excitation which is brought about by unstable flow.
Examples are vortex shedding from the lip of a gate and alternating shear-
layer reattachment underneath a gate.
(c) Movement-induced excitation of the vibrating structure. In this situation
the flow will induce a force which tends to enhance the movement of the
gate.

185
 Hydraulic gates and The vibrating system
valves
The equation of motion for the simplest form of vibrating system with linear
components is:

d2 y dy
m 2
c ky Ft 10:1
dt dt

where m mass
y displacement
t time
c damping (viscous)
k spring rigidity
F impressed force

(see Fig. 10.1)

The total mass of a submerged gate or a gate under free discharge conditions
is made up of its mass (m) and the hydrodynamic mass or added mass of water
vibrating with the gate (mw). Similarly there is a hydrodynamic component of
damping (cw) and rigidity (kw).
For a system in water, equation (10.1) can be represented by:

d2 y dy
m mw 2
c cw k kw y F 10:2
dt dt

where cw added mass damping

kw added mass rigidity
F all hydrodynamic forces

The system is stable or positively damped when:

c cw > 0 10:3

As a first approximation if added mass damping is neglected:

c>0 10:4

Figure 10.1. The vibrating

system

186
The system is unstable or negatively damped when: Gate vibration

c cw < 0 10:5

The critical damping coefficient Cc is given by:

p
Cc 2m!n 2m mw k kw =m mw 10:6

where !n natural frequency of oscillation of the gate in water and the damping
ratio  is:

c

Cc

When damping is less than critical  < 1 and oscillation occurs with
diminishing amplitude (stable or positively damped). If, as a first approxi-
mation, added damping cw and added mass rigidity kw are neglected:

C
 p <1 10:7
2m mw k=m mw

Kolkman2 gives an explanation of added damping and added rigidity.

The added mass coefficient Cm is given by:
mw
Cm 10:8
D2 L

where  fluid density

D gate depth or characteristic body dimension (gate
immersion)
L spanwise width of gate

In calculations, damping c is usually assumed to be constant friction

damping. In slide gates friction between the gate and the downstream bearing
face, and also the seal friction, provide damping. In roller gates it is the roller
bearing friction and rolling resistance as well as the seal friction. Where
upstream reaction pads have been provided in order to eliminate transverse
movement of the gate within the gate slots, they will contribute to the damping
forces. However, the assumption of constant friction is not necessarily valid
under conditions of gate vibration.
To prevent vibration, the dominant excitation frequencies should be well
away from resonance frequencies, given by the equation:

1 p
fr k kw =m mw 10:9
2

187
 Hydraulic gates and The resonance frequency fr must be a factor higher than the excitation
valves frequency f due to the flow velocity or of a reflected pressure wave. Kolkman3
advocates striving for at least a factor of three. At the condition when the
excitation frequency is equal or close to the resonance (natural) frequency, the
displacement amplitude for the vibrating system increases very rapidly and may
result in failure of the gate suspension system. The transmissibility ratio, TR, or
magnification factor, is given by:

1
TR 10:10
1 f =fr 2

The transmissibility ratio should be negative to prevent excitation of the

gate, that is the frequency ratio should be greater than unity. The range between
transmissibility ratios of unity and zero is sometimes called the isolation range,
with the percentage of isolation expressed between these limits. It is desirable to
produce a design with a high percentage of isolation.
With a frequency ratio of 3 as recommended by Kolkman,3 the
transmissibility ratio is 0.125 and the percentage isolation 87.5. At a ratio of 2,
TR is 0.333 and at 1.5, TR is 0.800.

Excitation frequencies
Two possible sources of disturbing frequencies are the vortex trail shed from
the bottom edge of a partly open gate and the pressure waves that travel
upstream in a conduit to the reservoir and are reflected back. The frequency
of a vortex trail in the case of flow induced vibration can be defined by the
Strouhal number:

S f L=V

where f excitation frequency

L a representative length of the flow geometry (in the case of a
tunnel gate it is the width, or twice the projection of the gate
into the conduit)
V a representative flow velocity at the gate
p
V 2gHe

where He energy head at the bottom of the gate

The Strouhal number of a flat plate is approximately 1/7. The excitation

frequency of a vortex shed from a gate may therefore be estimated as:
p
2gHe
f 10:11
7L

188
 The vortex trail will spring from the upstream edge of a flat bottomed gate, Gate vibration
causing pressure pulsations at the bottom of the gate. Where the gate has a 45
lip or a larger angle, the vortex trail springs from the downstream edge,
eliminating bottom pulsations. The use of flat bottomed vertical-lift gates to
control flow is strongly discouraged.
Abelev4,5 has presented a number of studies to establish the dominant
Strouhal numbers S and the excitation coefficient C1. The S values for
horizontal excitation of a culvert gate are shown in Fig. 10.2.
For a flat bottomed gate, S numbers for vertical excitation were given by
Naudascher6 and are reproduced in Fig. 10.3.
The Strouhal numbers of vertical-lift gates for culverts and of the flat
bottomed type in Figs 10.2 and 10.3 are within a range 0.4^3.0 and 0.18^
0.30 respectively. Martin et al.7 reported a Strouhal number of around 0.33
for a fixed vertical gate model with extended lip. In a laboratory study by
Kanne8 on gates elastically suspended in the vertical direction, values of
0.3^0.5 were measured for a gate opening of three times the gate thickness.
There is little information available on Strouhal numbers for radial gates
either in model or prototype scales. However, at least two investigations have
been carried out. The radial gates at the Barkley Dam on the Cumberland
River, Kentucky, were subject to severe vibration.9,10 The gates were
15.2 m high and 16.8 m wide. Vibration occurred within a range of gate
opening of 0.75^2.75 m. The vibration frequency was 30 Hz. At the
Torrumbarry Weir on the River Murray in Victoria, Australia, radial gates
6.25 m high and 11 m wide vibrated11 when the gates were between 3^4 m
open. This occurred at a vibration frequency of 10 Hz. In both cases the lips
of the gates were submerged by high tailwater levels when high frequency
vibration occurred. The Strouhal number for the Barkley and Torrumbarry
prototype gates was about 0.1.
The removal of sill seals from the Barkley gates during field testing largely
eliminated the excitation.9 At Torrumbarry, spoilers were fitted to the bottom
of the gates to break up the regularity of eddy shedding.11 This was successful in
preventing gate vibration.

Figure 10.2. Strouhal number

for the horizontal excitation of
a culvert gate (after Abelev4)

189
 Hydraulic gates and
valves

Figure 10.3. Strouhal number

for vertical excitation of a flat
bottomed gate (after
Naudascher6)

However, it remains unclear why high flexural vibration of radial gates with
high tailwater levels is a relatively rare phenomenon, because the lip geometry of
the gates at Barkley and Torrumbarry is typical of many radial gates.
A list of Strouhal numbers for various gate configurations is required to
enable quantitative analyses of transmission ratios to be carried out for most
gates.
The frequency of a reflected pressure wave is given by:

f V=4L 10:12

where V velocity of the pressure wave (the value of V ranges from

1400 m/s for a relatively inelastic conduit to 1000 m/s for a
relatively elastic pipe)
L length of conduit upstream from the gate

The natural frequency of free vertical oscillation of a suspended gate is:

1 p
fr gE=12s 10:13
2

where E modulus of elasticity of the suspension (for rigid suspension

rods this is 200 kN/mm2, but for wire ropes which stretch
under load due to the spiral winding of the strand E, it is
usually taken as 70 kN/mm2 , see Table 7.2)
s length of the suspension
 unit stress in the suspension (in calculating the unit stress,
the mass of the gate and the added mass must be used (m +
mw))

190
Added mass Gate vibration

Figure 10.4 illustrates the simplest case of added mass. In this case, the added
mass is the total mass of water above the piston; oscillation of the mass of the
piston m forces the mass of water above to move with the same velocity. When a
submerged body oscillates with small amplitude in a stagnant fluid, some of the
fluid will oscillate in phase with the vibration of the body, but the further away
the fluid is from the body the smaller its velocity compared with the body. The
added mass mw has the dimension of a virtual volume of fluid.
Computational methods have been established for determining mw. They
assume stagnant fluid conditions with potential flow induced by a body in
harmonic vibration. The investigations were carried out by Wendel,12
Zienkiewicz and Nath13 as well as Derunz and Geers.14
Added mass coefficients Cm (see equation (10.8)) have been established
experimentally.15^18
The experimental work of Hardwick15 and Thang17 showed that the
amplitude of vibration has little effect on Cm, but Thang found an appreciable
effect at frequencies between 9^12 Hz and some effect at vibration frequencies
greater than 23 Hz.
A gate stiffened by girders is analagous to the condition in Fig. 10.4. The
added mass and Cm will be considerably greater than that of a closed type gate.
The design of the gate bottom also has an appreciable effect on Cm, as shown
in Fig. 10.9.

Figure 10.4. Added mass

Figure 10.5. Variation of

added mass coefficient with
submergence for a flat bottomed
gate (after Hardwick15)

191
 Hydraulic gates and
valves

Figure 10.6. Effect of

vibration behaviour on added
mass coefficient Cm (after
Thang 17)

Preliminary check on gate vibration

The sequence for carrying out a preliminary check on whether a gate is liable to
vibration is as follows:
(i) Establish the total suspended mass (m).
(ii) Establish the spring rigidity of the suspension system (K).
(iii) Determine the added mass (mw).
(iv) Calculate the damping forces due to slide or roller friction, roller bearing
friction and seal friction. In all cases the sliding friction rather than the
static friction should be taken, because a disturbance may initiate gate
movement and the friction forces must then damp out the movement.

192
 Gate vibration

Figure 10.7. Effect of

submergence and lateral
confinement by gate slots and side
walls on added mass coefficient Cm
(after Thang17)

(v) Determine the critical damping coefficient Cc from equation (10.6) and
the damping ratio .
(vi) Calculate the resonance frequency fr from equation (10.9).
(vii) If the gate is in conduit and the gate lip is 45 or greater, calculate the
frequency f of the reflected pressure wave from equation (10.12).
(viii) Establish the transmissibility ratio TR from equation (10.10).
Other dynamic forces which can cause gate vibration are wave action,
cavitation, two-phase flow and water column separation.
Kolkman3 has suggested that vibration due to unsteady flow probably arises
via a mechanism involving a fluctuation discharge coefficient, induced by the
added mass flow of the vibrating gate. These conditions cannot be analysed
theoretically and are described later in this chapter.

193
 Hydraulic gates and
valves

Figure 10.8. Effect of gate opening

on added mass coefficient Cm (after
Thang17)

Vibration due to seal leakage

This is probably the most frequent cause of gate vibration. The mechanism of
self-excitation due to seal leakage has been explained by Petrikat19 and Lewin,20
although explanation of the hydrodynamic effect differs in the two papers.
Petrikat mentions a case of vibration caused by the top seal for a low level

Figure 10.9. Effect of gate bottom on

added mass coefficient Cm (after
Thang17)

194
vertical-lift gate at the Bharani Dam, and Krummet21 discusses a similar case of Gate vibration
vibration of a radial gate at a bottom outlet due to leakage of the top seal. Other
examples of vibration caused by seal leakage have been given by Kolkman3 and
Mitchell.22

Sill seals
Sill seals should be of rectangular shape and should be moulded in a moderately
hard elastomer (Shore A hardness 65) for gates in open channels and medium
head gates. For high head gates a Shore hardness of 75 is more appropriate.
Bottom seals can be metal to metal in order to overcome seal induced
vibration,19,23,24 but perfect sealing with such an arrangement is difficult to
effect with large and heavy gates21 and sometimes also with smaller ones.10
Under no circumstances should elastomeric seals project more than 5 mm below
the faceplate of a gate, and in high head gates the projection should be no more
than is required to effect a seal, about 3 mm. Wide block seals of timber or other
materials are not suitable,24,25 because they can shift the point of flow
attachment.26 Musical note shape seals have been used as sill seals,19,23 as
evidenced from comparatively recent model investigations, despite their

Figure 10.10. Arrangement

of suitable and unsuitable seals

195
 Hydraulic gates and unsuitability for this purpose. Diaphragm seals on bottom-hinged gates are also
valves vulnerable to vibration. A seal of this type failed on a bear-trap weir and was
replaced by a sliding seal.27 Figure 10.10 shows some arrangements of seals
which are unsuitable, alongside the correct configuration.
Slight vibration initiated by leakage from the sill seal can usually be identified
by vibration of skin panels and small amplitude ripples upstream of the gate.
Severe vibration can cause high amplitude movement of a gate and may be
attended by loud noise.

Side seals
Leakage past the side seals of gates in open channels rarely causes vibration of a
gate as a single unit, but can initiate flexing of local structural members,28
sometimes severely.21, 25 It mostly results in seal flutter, which can be very noisy.
Using two seals one after the other to suppress the jet from a leaking primary seal
is an unacceptable solution, because vibration can still be initiated by the
primary seal. The junction between sill and lintel seals and side seals often
presents design problems. Although special moulds are available for transition
sections, these require rigid attachment. It is difficult to assess the incidence of
vibration due to leakage at corners; it is probably high. The likelihood of corner
leakage can be reduced by arranging side and horizontal seals in the same plane.

Lintel seals
Vibration of medium and high head gates due to flow past or impinging on
lintel seals or protruding lips has been recorded by Petrikat19,29 and Krummet.21
Based on these and similar cases, centre bulb seals should be used in preference
to musical note shape seals (see Fig. 10.11). Flow past the opening created at the
lintel once the seal is no longer in contact can induce vibration.20
In emergency closure gates and draft tube inlet gates, when gates are used for
initial filling of the tunnel, it is highly desirable if not essential for the seal to
remain in contact with embedded parts of the lintel structure for the degree of
opening required to fill the tunnel. Since leakage through narrow gaps leads to
vibration,2 any design of an upstream sealing gate which aims at a rapid increase
of the lintel seal gap after opening, will lead to a cantilever mounting of the seal2
and be subject to excitation due to impingement of the flow.

Flow attachment, shifting of the point of

attachment and turbulent flow
Structural stiffening members
Structural stiffening members on radial and vertical-lift gates are frequently
placed too low, so that intermittent flow reattachment occurs at the flange of
the stiffening member. Figure 10.12(a) illustrates the design of the bottom
section of a radial automatic gate which suffered severe vibration problems
due to this cause. Figure 10.12(b) is a section through one of the Pershore Mill
gates on the Avon in Worcestershire,20 flat-bottomed gates where flow
attachment and vibration were predictable (this shape is the most unstable

196
 Gate vibration

Figure 10.11. (a) A lintel

seal arrangement which caused
vibration before baffle plate was
fitted 29 (b) Preferred
arrangement

because a free shear layer lies close to the bottom of the gate3). Figure 10.12(c) is
the configuration of the bottom section of a diversion tunnel gate which would
lead to flow reattachment problems because the structural stiffening member is
placed too low.

Roller and turbulent flow downstream of a gate

When a gate operates under drowned discharge conditions, an unsteady roller
occurs and conditions of high turbulence arise.30 If the roller acts on submerged
structural members, whether these are part of the gate skin plate stiffening or the
arms of a radial gate, vibration is likely to occur. This is not so much a case of
flow reattachment as the hydrodynamic action of turbulent flow. Where
structural members are submerged, the potential for flow induced vibration is
minimised when the member has a blunt trailing edge.31

Gate design guidelines

In designing a gate to avoid vibration the following guidelines can be stated:
(a) No structural member upstream or downstream of the control point should
protrude into a line at 45 from the point of flow control, and upstream
preferably at 60,2, 10 as shown in Fig. 10.13.
(b) It is better to arrange for the vortex trail to be shed from the extreme
downstream edge of a gate in order to achieve flow conditions that are as
steady as possible.32
(c) A sharp cut-off point should be provided at the lip.28, 32
Where a gate is situated near the crest of a weir and there is clear discharge
downstream, hydrodynamic excitation of gate arms or of projecting
structural stiffening members downstream of the skin plate does not arise.
In tunnel gates, radial gates or vertical-lift gates in open channels, subject to

197
 Hydraulic gates and
valves

Figure 10.12. Arrangement

of unsuitable structural
stiffening members at the
bottom section of gates

downstream drowned conditions, hydraulic considerations should override

structural priorities to dispose members in the most economical manner (Fig.
10.14).
No investigation appears to have been carried out into the effect on gates
when the downstream discharge is in the region of an oscillating hydraulic

198
 Gate vibration

Figure 10.13. Arrangement

of structural members at the
bottom of a vertical-lift gate

Figure 10.14. Preferred

hydraulic structural
arrangement of a gate subject to
drowned discharge

jump. This is likely to cause problems where structural members are located low
on the skin plate.

Hydraulic downpull forces and flow

reattachment at the gate lip
Vibration due to elastic deflection caused by hydrodynamic downpull forces can
occur in hydraulic servo-motor operated or rope suspended gates. In servo-motor
operated gates, the problem can arise due to a long operating stem and the
compressibility of the hydraulic fluid can also contribute to it. Elastic deflection
due to ropes can be reduced by substituting chain suspension, but this can cause
difficulties in the layout of mechanical components where gates have to be hoisted
through a considerable height. Chapter 9 gives some information on hydraulic
downpull forces. In the absence of a hydraulic model study, a first approximation
of investigating whether a gate is likely to vibrate due to hydraulic downpull
forces can be made by following the procedure set out earlier in this chapter.
The following guidelines will reduce downpull forces and conditions of
separated flow and reattachment:
(a) Gate lips should have a sharp cut-off point28, 32 (see also previous section on
gate design guidelines).

199
 Hydraulic gates and
valves

Figure 10.15. (a) Separated

flow (b) Possible shear layer
deflection of entrapped fluid

(b) Gate lips should be as narrow as possible.

(c) Gate lips should project as far as possible below the body of the gate.10, 33
(d) The angle of inclination of the upstream face of the gate should be at least
45 (see also previous section on gate design guidelines).
(e) The system should be as rigid as possible.
Conditions of flow are illustrated in Fig. 10.15 which shows separated flow at
the bottom of a gate and also possible shear layer deflection of entrapped fluid
based on Martin et al.7

Unstable flow through small openings

Pressure fluctuations which in turn cause discharge fluctuations, and
thereby exert a force on the lower edge of a gate, can initiate gate vibration
at small openings. This is a self-exciting phenomenon which occurs at high
velocity flows under small gate openings or at leakage gaps at lintel seals
when a gate is raised. It can be triggered by a vertical movement of the
gate, which is then translated into a momentary pressure change which
can reinforce the initial gate movement. The inertia of the water under flow
conditions contributes by causing a pressure rise in the conduit. The gate

200
movement will then either be amplified under resonance conditions or Gate vibration
damped out.
Vibration at small gate openings has been investigated by Kolkman2,3 and
Vrijer.32 Kolkman suggests that the width of the leakage gap should be at least
1.5 times (preferably 2 times or more) the width of the gate edge. This criterion
should also be applied to minimum gate openings of tunnel gates for cracking
open to fill downstream sections of the tunnel. If the resultant flow is
unacceptably high as a consequence, a bypass system for filling should be
provided. Experience suggests that gates will often remain stable at gate
openings below the minimum recommended values, probably because the
damping forces have been underestimated. In Chapter 9 it was stated that
minimum gate opening should not be less than 100 mm. If Kolkman's criterion
for minimum leakage gap results in a greater opening, this should be considered
the minimum.

Flow over and under the leaves of a gate

Instability due to vortex trails and/or flow separation where a gate leaf or a
stoplog has to be lowered into flowing water has been investigated by Brown34
and Grzywienski.35 Vortex action is a real threat if the depth of flow over the
gate exceeds 0.45 of the gate height for a flat sharp crested gate leaf. The depth of
underflow is less critical. When sections of the gate or stoplogs can be connected
together to avoid complete immersion, instability can be eliminated. Excitation
can be disturbed by introducing another jet into the wake, for instance, through
a controlled opening in the skin plate.

Vibration of overflow gates due to inadequate

venting
The means to effect venting of the nappe of overflow gates is shown in Fig. 2.27.
The underside of the nappe entrains air and causes a subatmospheric pressure
in the absence of venting or an inadequate air supply. If the air is exhausted it
results in collapse of the nappe, when it will suddenly attach itself to the
underside of the gate with a violent impact.
In an inadequately vented overflow gate, the level of water on the inside of
the nappe fluctuates due to inertia, causing variations in pressure which result in
horizontal movements of the nappe. These in turn cause fluctuations of
discharge, since a reduction in pressure increases the discharge over the gate.
This sequence of events can result in severe gate vibration which may be
transmitted to the civil engineering structure.

Vibration due to a free shear layer

In addition to the possible shear layer deflection of water trapped under a gate,
as shown in Fig. 10.15, vibration due to a free shear layer has occurred.

201
 Hydraulic gates and The model of the Split Yard Creek control structure in Queensland,
valves Australia36 indicated a well-defined shear layer at the gate shaft opening. This
was subject to apparent periodic oscillations. Two control gates were located
upstream of the shaft and in the fully open position were subject to flow induced
excitation in the form of beats, because one frequency component of the
excitation was close to the natural frequency of the gate assembly. The problem
was cured by a combination of leading and trailing edge ramps at the
intersection of the shaft and the tunnel.

Two-phase flow
Where air can be introduced into a conduit, severe pressure fluctuations can
occur at the control gate due to the build-up of stagnated air under high pressure
at the conduit crown upstream of the gate. The air, which is uniformly
distributed at the head race tunnel, accumulates and forms air pockets due to
the relatively low velocity of flow in the conduit and the long distance upstream
from the gate. The air pockets stagnate at the upstream side of the skin plate until
they are partially drawn under the gate. When the pressurised air is released, it
reaches atmospheric value almost instantaneously, with explosive force.
A particularly severe problem of this type was noted in a model study by
Rouve and Traut,37 as shown in Fig. 10.16. Discussing the paper, Maurice Kenn
of Imperial College, London pointed out that air entraining water flows are
notoriously difficult to model, except perhaps when tested with full scale
velocities. Because of scaling problems, pressure fluctuations in prototypes
may prove less severe than those suggested by model tests.
Nielson and Pickett23 have recorded severe vibration of a reverse radial gate
which was attributed to the collapse of large vapour cavities near the gate. The
gate acted as a control valve for a high lift lock with a maximum differential head
of 28.1 m. This type of problem can only be solved by venting upstream as well
as downstream of the gate.
Singh et al.38 have reported the dislocation of a bulkhead on a tower type
intake due to air compression. Air entraining vortices had formed at the intake
under some lower reservoir levels, and the subsequent operation of the
emergency gate 200 m downstream of the portal caused surges which increased
the pressure on the trapped air, driving it up the intake in an air/water spout
which dislocated the bulkhead.

Slack in gate components

When a gate opens or closes, the inertia of the water creates regions with
increased or decreased pressure. Excitation can also be brought about by flow
velocity fluctuations at constant gate openings which will vary across the
opening.
Vertical-lift roller gates can be subject to hydrodynamic pressure conditions
so that the uppermost wheel or wheels are on the point of unloading. Kolkman3
gives an example of this type where a strong rotational vibration occurred,
centred on the lower wheel shaft.

202
 Gate vibration

Figure 10.16. Two-phase flow below a radial gate (after Rouve amd Traut37)

Excessive slack in mechanical gate components such as guide wheels, pivots

and hoist chains should be avoided. Where clearances are essential it is
important that the component be preloaded. An example of preloading of the
guide wheels of a surge shaft gate is shown in Fig. 10.17.

203
 Hydraulic gates and
valves

Figure 10.17. Preloading of the

guide wheels of a vertical-lift
gate

References
1. Naudascher, E (1979): On identification and preliminary assessment of sources of
flow induced vibration, 19th I.A.H.R. Congress, Karlsruhe, paper C1.
2. Kolkman, P A (1984): Vibration of hydraulic structures, in Developments in hydraulic
engineering 2, editor Novak, P, Elsevier Applied Science Publishers.
3. Kolkman, P A (1979): Development of vibration free gate design, Delft Hydraulics
Laboratory, publ. 219.
4. Abelev, A S (1979): Investigation of the total pulsating hydrodynamic load acting
on bottom outlet sliding gates and its scale modelling, 8th I.A.H.R. Congress,
Montreal, paper 10A1.
5. Abelev, A S (1963): Pulsations of hydrodynamic loads acting on bottom gates of
hydraulic structures and their calculating methods, 10th I.A.H.R Congress, London,
paper 3.21.
6. Naudascher, E (1964): Hydrodynamische und Hydro-elastiche Beanspruchung
von Tiefschtzen, Der Stahlbau, Nos 7 and 9.
7. Martin, W W; Naudascher, E; Pradmanabham, M (1975): Fluid dynamic
excitation, involving flow instability, Proc. A.S.C.E., Journ. Hydr. Div, 101, HY6,
Jun.
8. Kanne, S (1989): Vibration of a vertical-lift gate with variable bottom geometry (in
German), Diploma thesis, University of Karlsruhe, Germany, referred to in

204
 Naudascher, E and Rockwell, D (1994) Flow-induced vibrations, an engineering guide, A Gate vibration
A Balkema, pp. 343^344.
9. Hart, E D; Hite, J E (1979): Gate vibration tests, Barkley Dam, Cumberland River,
Kentucky; technical report HL-79, US Army Corps of Engineer Waterways
Experimental Station, Vicksburg, Missouri, May.
10. Hart, E D; Hite, J E (1979): Barkley Dam gate vibrations, 19th I.A.H.R. Congress,
Karlsruhe, paper C15.
11. Hardwick, J D; Attari, J; Lewin, J (2000): Flow-induced vibration of Torrumbarry
Weir gates, Proc. 7th Int. Conference on flow-induced vibration, Lucerne, Switzerland,
Jun., editors Ziada, S and Staubli, T, Balkema, 219^224.
12. Wendel, K (1950): Hydrodynamische Massen und Hydrodynamische
Massentraggeheits-momente, Jahrbuch der Schiffsbautechnischer Gesellschaft, 44, 207^
55.
13. Zienkiewicz, O C; Nath, B (1964): Analogue procedure for determination of virtual
mass, Proc. A.S.C.E., Journ. Hydr. Div., HY5, Sept., p. 69.
14. Derunz, J A; Geers, T L (1978): Added mass computation by the boundary integral
method, Int. Journ. Numerical Methods Engng., 12, 531^49.
15. Hardwick, J D (1969): Periodic vibrations in model sluice gates, PhD thesis, Imperial
College of Science and Technology, London.
16. Hardwick, J D; Ken, M J; Mee, W T (1979): Gate vibration at El Chocon
Hydropower Scheme, Argentina, 19th I.A.H.R. Congress, Karlsruhe, paper C7.
17. Thang, N D (1982): Added mass behaviour and its characteristics at sluice gates, Int.
Conference On flow induced vibrations in fluid engineering, B.H.R.A., Reading, England,
Sept., paper A2.
18. Kolkman, P A (1988): A simple scheme for calculating the added mass of hydraulic
gates, Journ. Fluids and Struct., Vol. 2, pp. 339^353.
19. Petrikat, K (1979): Seal vibration, 19th I.A.H.R. Congress, Karlsruhe, paper C14.
20. Lewin, J (1983): Vibration of hydraulic gates, Journ. I.W.E.S., 37, 165^179.
21. Krummet, R (1965): Swingungsverhalten von Verschlussorsganen im
Stahlwasserbau, Forschung in Ingenieurwesen, Bd. 31, No. 5.
22. Mitchell, W R (1979): Vibration due to leakage through a reverse radial gate, 19th
I.A.H.R. Congress, Karlsruhe, paper C17.
23. Nielson, F M; Pickett, E B (1979): Corps of Engineers experience with flow
induced vibrations, 19th I.A.H.R. Congress, Karlsruhe, paper C3.
24. U.S. Waterways Experimental Station (1956): Vibration and pressure cell tests, flood
control intake gates Fort Randall Dam, Missouri River, South Dakota, technical report
No. 2435, Vicksburg, Mississippi, Jun.
25. Lewin, J (1980): Hydraulic gates, Journ. I.W.E.S., 34, No. 2, p. 237.
26. Bruce, B A; Crow, D A (1978): Hydroelastic model studies of the Pershore Mill sluice gates,
B.H.R.A., report RR 1485, Jul.
27. Merkle, T (1979): Hydraulically induced vibrations in a bear-trap weir, 19th
I.A.H.R. Congress, Karlsruhe, paper C19.
28. Schmidgall, T (1972): Spillway gate vibrations on Arkansas River Dams, Proc.
A.S.C.E., Journ. Hydr. Div., HY1.
29. Petrikat, K (1976): Structure vibrations of segment gates, 8th I.A.H.R. Congress,
Leningrad.
30. Murphy, T E (1963): Model and prototype observations of gate oscillations, 10th
I.A.H.R. Congress, London.
31. Pennino, B J (1981): Prediction of flow induced forces and vibration, Water Power
and Dam Construction, Feb., p. 19.
32. Vrijer, A (1979): Stability of vertically movable gates, 19th I.A.H.R. Congress,
Karlsruhe, paper C5.

205
 Hydraulic gates and 33. Hampton, I G; Lesleighter, E J (1980): Effect of gate shape on closure loading, 7th
valves Australasian hydraulics and fluid mechanics Conference, Brisbane, Aug.
34. Brown, F R (1961): Fluctuation of control gates, 9th I.A.H.R. Congress,
Dubrovnik, pp. 258^269.
35. Grzywienski, A (1963): The effect of turbulent flow on multi-section vertical lift
gates, 10th I.A.H.R. Congress, London.
36. Chang, H T; Hampton, I G (1980): Experiences in flow induced gate vibrations,
Int. Conference on water resources development, Taipei, Taiwan, May.
37. Rouve , G; Traut, F J (1979): Vibrations due to two-phase flow below a Tainter
gate, 19th I.A.H.R. Congress, Karlslruhe, paper C10.
38. Singh, S; Sakhuja, V S; Paul, T C (1982): Some lessons from hydro and aero-elastic
vibrations problems, Int. Conference on flow induced vibrations in fluid engineering,
Reading, Sept., paper B1.

206
 11
Control systems and
operation

This chapter deals with control objectives, operating rules and systems,
telemetry, fall-back systems and standby facilities, as well as instrumentation.
In control systems, the trend is for electromechanical controls to be replaced by
electronic closed-loop control systems. A contributory factor is the increasing
complexity of operating rules which may require flood attenuation in the discharge
from a spillway to a river course; or, in barrages, a different mode of gate operation
when dealing with a flood compared with high tides or storm surges. Appropriate
responses can be programmed into electronic controllers.
Where different responses to different conditions are required, operator
training and practice present problems because there is little or no experience
of rare or extreme events. Training may have to be carried out on a simulator.
Invariably, automatic gate control systems are backed by manually operated
standby controls. Safety and reliability engineering requires that operating and
standby systems are independent of one another and are functionally different.
The practice is for an electronic controller to be backed by a second electronic
controller with automatic changeover from one to the other in the event of
malfunction.

Control objectives
In rivers
The operation of gates in rivers is designed to maintain the upstream water level
for navigation or for water abstraction and to pass flood flow. Gates may be
used to lower the water level for construction purposes such as bank
consolidation or for channel improvement works; such tasks are carried out
under the control of an operator. Automatic control is mainly focused on
upstream level control. Where flood warning systems are in operation the water
level in a reach or in a reservoir may be lowered in anticipation of a flood.

In reservoirs
Safety of the dam must be ensured by preventing the reservoir level from rising
to within a few metres of the crest, as overtopping could destroy fill dams or
other types of dam.

207
 Hydraulic gates and
valves

Figure 11.1. Attenuation of

outflow and reduction in the rate
of outflow compared with inflow

For hydroelectric and many irrigation projects it is required that the

optimum operating level be reached at the end of the flood to maximise
generation and storage for irrigation release.
Flood routing may also be an objective, in which circumstances some of the
following requirements may apply:1,2
(a) The maximum flood discharge must not be increased and preferably should
be significantly attenuated (Fig. 11.1).
(b) The rate of flow must not be increased.
(c) The flood propagation rate should be attenuated (Fig. 11.2).
(d) Storage capacity must be provided in expectation of a flood or snowmelt.
(e) Bank stability must be maintained in the reservoir by avoiding rapid rise or
fall of water level, and downstream of the reservoir by limiting channel
velocities.

Figure 11.2. Attenuation of

flood propagation rate

208
Operating rules and systems Control systems and
operation
Manual methods
Local control of motorised operation
This takes the form of increasing the outflow in steps. At its simplest level gate
openings may be determined by the operator based on experience in responding
to a signal indicating river or reservoir level.
To reduce reliance on operator judgement, curves may be used which show
in 30^60 min intervals the increase in gate opening as a function of river or
reservoir level and rate of change of level during the interval.3 The charts can
be designed so that, for a given rate of change of level, a specified discharge rate
is not exceeded in order to limit downstream flooding.

Remote control of motorised operation

This may be located remote from the spillway, barrage or weir, or may be carried
out from a centre controlling a number of dams. In all cases alternative or
standby operation of the gates is provided at the spillway or barrage.

Computer assisted control

Extensive input data may require the use of a computer to determine how gates
should be operated. This is the case when a hydrometeorological model is used,
or when the complexity of interpreting operation charts is likely to lead to
errors.
For example, data from raingauge stations, river gauging, inflow into the
reservoir, reservoir level, meteorological information and gate opening may
be fed into a computer, which then, by reference to the available flood storage,
executes flood routing calculations and prints out the operating instructions.
These are then carried out by personnel (River Medway flood prevention
scheme4). The data may be keyed in manually, or the computer may operate
on signals received direct from the gauging stations and control instruments.

Automatic methods
Cascade controls
Cascade control is mostly applied to the control of reservoirs and will therefore
be discussed in that context. The distance between the retention level and the
maximum reservoir level is divided into a series of steps, each corresponding to
a gate opening. On reaching a specific water level, the gate hoist motion is
started and the gates open in sequence to their predetermined height, controlled
by limit switches. A frequent refinement is to provide alternative limit switches,
permitting greater gate openings if one of several spillway gates is out of
operation due to maintenance or malfunction.
Cascade control is generally used in conjunction with power actuation of
gates, either by electric motor driven winches or by oil hydraulic cylinders,
but there are exceptions.
At the Victoria Dam in Sri Lanka counterbalanced radial spillway gates were
devised,5 which can open under gravity. The control system is of the cascade

209
 Hydraulic gates and type but can operate mechanically without external power, with electrical
valves controls as a standby.
Eight gates, each 12.5 m wide, open in pairs in four stages at 0.7, 2.5, 4.7 and
9.35 m. Opening is actuated by floats, which over a rise of upstream water level
of 0.64 m operate oil hydraulic poppet valves. These direct the oil from the
piston side of the hoist cylinders to the tank, permitting the gates to open under
gravity. When the gates have reached their appropriate opening step, an
actuator on the gate closes another poppet valve which stops the flow of oil
and locks the cylinder in position. Other floats are connected over pulleys to
electric limit switches which energise the relays for solenoid operation of oil
hydraulic, directional control valves. This provides a standby system for
opening the gates. Closure of the gates is by oil hydraulic cylinders, supplied
by electric motor driven, oil hydraulic power packs.

Level control
Electromechanical This is actuated by a predetermined rise in water level above
the retention level, which initiates opening of the gates in steps and in sequence.
A water level control band is set. When the upper limit of the control band is
reached, the opening motion is started and continues in steps until the level falls
below the upper limit, when the motion stops. Closure of gates commences
when the lower limit of the control band is reached. The raising and lowering
motions are interrupted by a dwell period to prevent hunting.
An ultimate upper water level limit switch initiates an alarm signal, and some
control systems are designed so that the gate hoist dwell period is cancelled
during the time when the uppermost level is reached or exceeded.

Computer controlled (feedback control system6) This moves the gates in turn to a set
point after determining the desired outflow with reference to the measured
inflow. Control instructions are issued by a computer which is programmed
with the strategy for maintaining upstream water level.
In proportional integral derivative (PID) control, the value at which
upstream water level is to be maintained is compared with the value transmitted
by water level sensors. The difference between the two signals, the error, is
computed at fixed intervals and is used to operate the proportion and
integration algorithm. The proportional term causes an immediate and
longer-term corrective action, and as long as the error persists the integral term
will increase or decrease continuously so as to open or close the spillway gates.
Feedback control systems6 compare the actual value of a variable with its
desired value and take the necessary corrective action. In gate control the
variable is usually water level, although the rate of change of water level may
be used as an additional control parameter. It is characteristic of a good
feedback-based closed-loop system that it maintains the desired level, the set
point, and corrects for any variations with a minimum of oscillation. The gate
should respond so that a small error results in a small opening and a larger error
in a larger gate opening. The actual as compared with the desired level should be
closely tracked by the system.
A block diagram of a closed-loop system is shown in Fig. 11.3

210
 Control systems and
operation

Figure 11.3. Block diagram of

a closed-loop system

Criteria of a successful closed-loop control system

(a) How well the system reduces the error signal to zero or almost zero.
(b) The final difference between the measured value and the set point, called the
`offset' in control terminology.
(c) The speed with which a system responds or restores agreement (in gate
operation speed of response is not an important factor).
(d) The system should be stable, that is, free from large and violent oscillations.

Modes of control
(a) Proportional control ^ magnitude oriented.
(b) Proportional plus integral (PI) control ^ magnitude and error time duration
oriented.
(c) Proportional plus derivative (PD) control ^ magnitude and error rate of
change oriented.
(d) Proportional plus integral plus derivative (PID) control ^ magnitude,
error time duration and error rate of change oriented.

Proportional control
In proportional control, the output is proportional to the error signal. When
there is an increase in water level due to increased inflow, the controller will
actuate raising of the gate to compensate for the increase in flow. Since gate
opening is proportional to the error signal the new opening can only be
maintained if there is a permanent error, thus proportional systems tend to have
a permanent error, the offset.

Proportional plus integral (PI) control

To overcome the offset problem, the time integral to the error signal
(magnitude of the error signal multiplied by duration of the error) is used to
determine the new gate opening. The proportional term positions the gate in
proportion to the error signal, i.e. the increase in water level, and the integral
term senses the offset which remains and continues motion in the same direction
until the offset is reduced (Fig. 11.4).

211
 Hydraulic gates and
valves

Figure 11.4. Proportional

plus integral control

Proportional plus integral plus derivative (PID) control

The derivative term measures the rate of change and causes the system to react
more rapidly (Fig. 11.5). PID control is most frequently used in gate position
control.

Application of PID control

Symbols: x1 required upstream water level
x2 actual water level
xe error signal x1 x2
Ts sampling interval
Vp output of proportional term algorithm
Vi output of integral term algorithm
Vt Vp + Vi
q nominal demanded flow rate
 gate angle, which is corrected in order to be proportional to
gate opening
Tgs interval between gate error signals
K1 proportional constant
K2 integral constant

Figure 11.5. Comparison of

different control modes

212
 n level tolerance (of water level), also known as dead band Control systems and
A surface area of reservoir or reach of river operation

The value x1 of the upstream water level to be maintained, is compared with the
value x2 transmitted by a water level sensor. The sensor may be a float, an
electrode or an ultrasonic device, pressure transducer or bubbler device. The
difference between the two signals xe x1x2 is computed every Ts seconds
and is fed into the PI algorithm. The output Vt of this algorithm is

V p K1 x e proportional term

V i V p K2 x e T s integral term

Vt Vp Vi

An upper and lower limit are placed on Vi (maximum and zero flow,
respectively).
The difference xe between the retention level and the actual water level
requires the gate to rise and discharge the increased inflow into the reach or
reservoir, and a level difference xe will correspond to a specific gate opening.
The gate angle is not directly related to gate opening, and the coefficient of
discharge for flow under a gate varies with gate opening. The computer must
therefore be programmed to convert the required flow rate into a set of
demanded gate angles using a polynomial fit. This may require inbuilt logical
hysteresis. The computer determines the required outflow rate from the signal
Vt and converts it to the new gate angle .
For free discharge conditions, the upstream water level and gate angle are
input to the control program. For gates in barrages where the discharge can be
submerged, the downstream water level is also measured and transmitted to the
computer to enable the required outflow to be calculated and converted to
demanded gate angle. From gate control considerations where upstream level
has to be maintained, knowledge of the exact relationship between gate angle,
i.e. gate opening, and discharge under or over a gate is not essential. It is only
required if the data are logged to obtain a record of the flood hydrograph.
In multigate installations, each sluiceway will normally behave as if it is
independent of the adjoining sluiceways, provided the approach to the gates is
sensibly straight. Thus the computer can be programmed to calculate the gate
opening , depending on how many gates are operational.
The value of the demanded gate angles will now be held for the rest of the
current sampling period Ts. The demand angles are compared with the
measured gate angles and generate gate error signals every Tgs seconds. Tgs must
be considerably shorter than Ts, because the dynamics of the gate loop are very
much faster than the response of the reservoir or river reach.
If any of the error signals exceed a threshold, the gate hoist motors are started
up and close or open the gates to adjust flow. This adjusts the upstream water
level until the difference between the actual and required levels is eliminated.

213
 Hydraulic gates and During a flood, the inflow will rise and subsequently fall. Any difference
valves between the demanded level and the measured level will
cause an immediate corrective action due to the proportional term
cause a longer-term corrective action to eliminate any difference between x1
and x2 (the error).
If there is any difference between x1 and x2, the integral term will increase or
decrease continuously causing the gates to open or close. Hence there is no
requirement within the control system to have an accurate relationship between
actual and demanded outflow rates. If there is inconsistency between the two,
the integral action will compensate for it. Ultimately, after a change of inflow to
the reservoir or the reach, the flow over or under the gates will balance to
eliminate any difference between x1 and x2.
To facilitate design of the control system and the parameters Ts, Tgs, K1 and
K2, the variables mentioned in the following paragraphs must be fixed:
The maximum allowable change in water level above retention for a given
percentage change in the inflow and the time at which this maximum should
occur (Fig. 11.6).
The time for the transient to settle within a specified proportion of the peak
(Fig. 11.6).
The maximum permitted variation in level, shown in Fig. 11.6. The time at
which it should occur together with a time at which this should decay to a
given tolerance band, shown by lines a and c in Fig. 11.7.
In practice this is sometimes approached differently. When a gate is operated
by an electromechanical hoist, it is advisable to limit the number of motor starts
per hour to ensure long life. If, for instance, this is fixed at four starts per hour, a
check is required that gate opening is compatible with the rate of change of
upstream level at the steepest part of the design hydrograph, and that the change
in water level above retention during the rest period of the motor is acceptable.
This should be the case when all gates in a multigate sluice installation are in
operation. It may be critical when one gate out of a two- or three-gate sluice is

Figure 11.6. Illustration of

upstream level variation due to
increased inflow

214
 Control systems and
operation

Figure 11.7. Illustration of

upstream level variation due to a
change in demand level

out of action due to maintenance or a defect. The gate controller can be

programmed to override any restriction on the frequency of operation when
one or several gates are not available, or when the increase in upstream water
level exceeds a critical value.
When gates are operated by oil hydraulic servo-motors, frequency of motor
starting is less important because motors driving oil hydraulic pumps are started
with the pump offloaded.
Any gate installation must be operated so as to minimise surges upstream and
downstream. These could affect navigation, fishermen and other river users,
and could also be important if there is a sluice installation under automatic
control in the downstream reach. Where the possibility of surges exists,
transients in upstream level measurement must be filtered out. The widely
adopted gate hoisting rate of 300 mm per minute does not prevent surges in
rivers unless the hoisting time is limited.
For a given change in flow rate through the sluices, or a given change in
demanded level, there will be a time limit for the system to respond, depending
on the flow discharge due to gate opening and the upstream head, or the
difference between upstream and downstream head. The speed of response is
also influenced by the rate of inflow. To a first approximation, the rate of fall
in level is given by:

outflow rate inflow rate q 2 q1

rate of fall in level
surface area of reservoir or reach A

The speed of response can be made higher for small signals by increasing K1
and K2. Making K2 larger makes the system faster but less damped, i.e. more and
more oscillatory with progressively larger overshoots and undershoots.
A usual and conservative limit for a maximum outflow q2 is:

215
 Hydraulic gates and q2 max
K1 n
valves 100
where n level tolerance.
Using the Kotmale reservoir7,8 as an example:
q2max 5560 m3/s (probable maximum flood)
A 6.63 km2 (surface area of the reservoir)
n say 20 mm (level tolerance)
then K1 2780 m2/s

In order to obtain adequate damping of the oscillations, it is advisable to

place a limit on K2 given by:

K2  K12 =A

For Kotmale, this would result in:

K2 1:16 m2 =s2

The intersample interval Ts must be short enough not to destabilise the

system. It is suggested that there should be at least 20 samples per cycle of
transient oscillation. The cycle time T of the system will be of the order of:
p
T 2 A=K2

Hence for Kotmale:

T 15021 s

Therefore it is inadvisable to use an intersample interval Ts greater than

15021=20s 750s:

Other control parameters can be used apart from upstream level, such as
constant downstream level which is common for irrigation channels.

Telemetry
Automatically operated sluices and spillway gates are in most cases remotely
supervised by telemetry at a central monitoring station. Usually the station
supervises several installations, sometimes carrying out different functions. Key
information is displayed on a VDU and recorded on an event printer. For a sluice
gate installation, in addition to essential data such as upstream and downstream
water levels and gate positions, other data will be transmitted such as:
availability of mains supply
availability of transformers
condition and availability of standby generating plant

216
 gate availability Control systems and
aggregated faults in any one motor circuit or gate equipment operation
valve position (if valves are part of the installation).
If the electrical switchgear and standby generators are located in a control
building, additional data may include:
intruder alarm status
fire alarm status.
In general, telemetry is confined to information and data transmission. If
these indicate a malfunction or a potentially dangerous condition, personnel
are sent to the site to deal with it. In a few cases, instructions and commands
are sent to the gate installation by telemetry.
Where operating or override control of an automatic system is via telemetry,
such as in a large barrage, the telemetry lines are duplicated. Good reliability
requires that an emergency control system is independent and generically
different from the primary controls. This is accomplished by hard-wiring the
emergency controls and locating equipment for this method of operating gates
and associated plant in a separate room from that of the primary control system.
This prevents a common cause of faults, such as a fire, putting the installation
out of operation.
Where the gate installation and the supervisory station are some distance
apart, two dedicated telephone lines form the link. Some geographically isolated
sites use short wave radio transmission. Operating experience of telemetry
transmission over distance suggests that the link is the weakest part of the
control system.

Factors in the choice of automatic gate control

systems
The choice of gate control systems has to take into account social environment,
available skills and technical back-up. Electronic control systems have to be
replaced at substantially shorter time intervals than mechanical and
electromechanical plant. Replacement of components becomes impossible after
a few years because of the rapid changes and developments in electronics.
In developing countries, the concept and practice of maintenance which
includes the systematic operational testing of an installation can be variable or
even completely absent. Problems can include lack of awareness by management
of the requirements for ongoing operational reliability of control systems, of the
technical skills required to maintain it, and financial or budgeting restraints.
Many dams in developing countries are financed by international aid or regional
development organisations. Parallel funding of maintenance during the lifetime
of the structure is seldom provided.
Where conditions for continuous reliable operation of automatic control
systems are absent or in doubt, it is questionable whether they should be part
of the initial installation. Even in developed countries, and at hydro plants
where an efficient technical back-up organisation exists, it cannot be assumed
that operating staff will carry out the correct action in an emergency without

217
 Hydraulic gates and ongoing training and the support of an effective management structure.
valves Emergency situations are rare events and are very stressful. Typically they
include loss of communications, power supply and failure of automatic gate
systems.
Should flood routing be required in order to limit downstream inundation,
an operator controlling the gates may have to perform actions which are
counter-intuitive at some stages of the flood rise. During the early rise of the
hydrograph, the initial opening of gates is followed by successive closures when
the bank full capacity of the river has been reached. Closure of gates is the next
step to maintain the flow discharge while the reservoir level rises, until the
retention level of the reservoir is reached, when the direction of gate operation
has to change again.
When flood routing has to be carried out, and also when a river is controlled
by a cascade of gated structures, automatic programmable logic control (PLC)
operation based on tested algorithms is advantageous. Incorrect operation of a
gated structure controlling a river can result in an amplification of the flood
flow.
The usual practice is to duplicate PLC controlled operation, to introduce
automatic self-checking and to arrange for standby equipment to take control
automatically in the event of failure. Operating staff must be trained and capable
of carrying out the control of spillway gates or a river barrier. Detailed simple
operating rules should be available and should not require the operator to carry
out calculations or exercise judgement. The rules should be applicable
irrespective of the size of flood flow, and actions should depend on easily
observed data such as upstream water levels (sometimes also downstream water
levels) and gate openings. PLC operation depends on transmission of data from
instrumentation. Where possible, instrumentation should be backed up by
simple visual measuring devices, such as water level gauges and gate position
indicators where a pointer moves across a graduated scale.
At automatically controlled gate installations, the provision of event
recorders is of value. A record of the system's performance permits
improvements to be made to the control system and/or the control program.
In the event of a failure, it can assist in identifying the cause. At manually
controlled gate installations, event recorders can be used to assess whether
operating staff carry out their duties correctly.

Fall-back system and standby facilities

A survey of existing practice reveals wide variations. In nearly all cases at least
one stage of redundancy is introduced for critical equipment, and in many cases
two stages are provided. For a safety critical control structure, such as spillway
gates, barriers or river barrages, two stages of redundancy are strongly
recommended.
Electrical supply is usually via two independent feeds or a ring main. With
high voltage mains service, two transformers would complete such an
installation, each transformer feeding a section of the busbar which is divided
by an air circuit breaker.

218
 Diesel engine driven standby generating plant, with or without automatic Control systems and
start-up on mains failure, is common. The generating set is either connected operation
permanently to the busbars via an interlocked circuit breaker or is of the
portable type with plug-in facility.
The probability of failure of a diesel alternator set to start and run for two
hours per demand is 0.043 per demand. Assuming that there is 1 demand in 2
years, the frequency of failure is approximately 1 in 46 years. This is not an
acceptable level of reliability for emergency equipment. For adequate security
two standby generator sets are therefore required; the probability of
simultaneous failure of two sets is 1 in 540 years. This assumes that there is no
common cause failure, that is, an event or circumstance which affects both
generating sets. Diesel fuel can be such a cause. Difficulties have been
experienced with waxing during winter weather in cold climates, and with fuel
stratification due to long storage and bacterial growth in tanks. On this account
the US Bureau of Reclamation favours the use of low pressure gas engines for
standby generating plant at spillway gate installations.
Petrol engines, used intermittently, are subject to more severe problems than
those experienced with diesel engines.
While standby generation ensures a power supply in the event of mains
failure, it does not provide for other electrical failures in an emergency such as
a motor burn out, failure of the busbar or motor starter. Some operating
authorities consider that in a multigated installation, in the event of a flood,
there is sufficient time to exchange motors if a failure occurs. Others provide a
spare motor as a standby. It is possible to arrange for the standby generating set
either to plug into the busbar or to bypass the busbar and the individual motor
starters, as shown in Fig. 11.8.
Standby equipment for servo-motors takes the form of a portable or mobile
power pack which can be connected to the operating cylinder or cylinders by
flexible high pressure hoses and self-sealing couplings. For high reliability the
portable power pack should have two diesel engines, each driving an oil pump.
Other emergency equipment for driving gate hoists includes diesel-engine-
driven hydrostatic transmissions which can be coupled to each gate in turn,
either to the hoist gearbox or the hoist motor extension shaft. This is shown
diagrammatically in Fig. 11.9.
Figure 11.10 shows a diagram of a compressed air storage system with
permanent air motors at each gate.

Figure 11.8. Standby diesel-

engine generator set

219
 Hydraulic gates and
valves

Figure 11.9. Standby diesel-engine-

driven hydrostatic transmission

Battery powered emergency drive systems are another means of making

stored power instantly available. They can take the form of permanent DC
motors (Fig. 11.11). The alternative is to interpose an inverter to utilise existing
AC motors and starting equipment (Fig. 11.12).
A battery powered emergency drive system has a frequency of failure of the
order of 1 in 800 demands.

Figure 11.10. Schematic

diagram of compressed-air
storage system with permanent
air motors at each gate

220
 Control systems and
operation

Figure 11.11. Schematic

diagram of battery power with
DC motor drive

Figure 11.12. Schematic

diagram of battery power with
AC motor drive

Gates requiring no external power in an emergency

Two types of spillway gate can be opened under gravity.

(a) `Gibb' gates, Victoria Dam, Sri Lanka5 The spillway gates of the Victoria
Dam in Sri Lanka are counterbalanced so that they open under gravity, and
closure is effected by oil hydraulic cylinders (see section on cascade controls
earlier in the chapter). So far they appear to be the only gates which operate
automatically under gravity.
There are a few spillway installations where gates are counterbalanced to
open and close by oil hydraulic cylinders, for example the radial gates at the
Pueblo Viejo Dam in Guatemala.9 Gate opening is activated by pressing a single
button. To remain closed, the hydraulic circuit must be active. The system is
designed so that any major malfunction results in opening of the gates. In case
of gate malfunction or blockage in the closed position, gate overflow will result

221
 Hydraulic gates and in a subsequent failure to clear the fluidway. Uncontrolled opening of a spillway
valves gate can result in a risk to people and property downstream of a dam;
presumably there was no such danger at this site. Furthermore, it is difficult to
design gates to collapse under a specific overload.

(b) Bottom-hinged flap gates These have been used as spillway gates, specifically
at the Legadadi Dam in Ethiopia. Emergency lowering can be effected by
manually setting a directional control valve to port, which will also direct the
annulus of the gate operating cylinder to tank and vent the cylinder side. The
pressure of the reservoir water then lowers the gate.
An Australian hydroelectric authority employs automatic float operated
counterweighted gates at spillways, similar to those shown in Fig. 2.11. The
gates can be raised in the event of malfunction by shutting the displacer chamber
outlet pipe.
The provision of manual winding of gates operated by electromechanical
hoists is almost universal, although it is used as a last resort to open larger gates
because the time required can vary from six to 12 hours.
On oil hydraulic power packs for servo-motors, a hand pump is provided to
operate the gate in an emergency. As with electromechanical drive systems, it
takes a long time to effect gate opening. This can be improved to a limited extent
by fitting two pumps.
It is general practice to arrange hydraulic circuits so that in an emergency
gates can be lowered (or in some cases opened) by gravity through manual
operation of a control valve.

Alternative means of control

In all automatic control systems, means are provided to revert to manual control
by an operator pressing `open' and `close' buttons if the automatic controls fail
to operate. These also serve to actuate the gates for testing, servicing and
maintenance. The practice of testing spillway gates varies at different dams
and with different operators.10
Central computer operation for a cascade of dams or barrages is backed up by
a local computer at each barrage, for example on the River Rhone.10 At another
project the spillway gate control computer can, in the event of failure, transfer
control functions to a second computer at the power station. At Kotmale Dam8
an electromechanical level control system takes over if the computer control
fails to maintain maximum retention level.
Computer controlled systems are almost always designed to be self-checking.
At Kotmale8 an independent checking and warning system coexists with that
incorporated in the computer.
A difficulty arises in training operators to manually control spillway gates
which are normally operated automatically. Their experience of dealing with
flood events is limited and intermittent. Where a computer is provided such
training could be by simulation, although at present such applications appear
to have been confined to testing control strategies (Kotmale Dam) rather than
operator training.

222
 The practice of operating flood release from reservoirs, as well as staffing Control systems and
policy at dam outlets, has been reviewed and tabulated by Combelles and operation
Tinland.10

Instrumentation
The two main parameters which must be measured in order to control gates and
valves are water level, which may include both upstream and downstream water
levels, and gate or valve opening. In some cases direct flow measurement is also
required, although this is more frequently deduced from calibration curves of
water level and gate or valve opening.

Water level measurement

Water level can be measured by electrodes, bubbler devices, ultrasonic sensors,
pressure transducers or float gauges. The latter appear to be the best choice for
limited distances of approximately 10^15 m. Pressure transducers tend to be
more reliable for measurement of greater depth, but less accurate than precision
bubbler devices which can be accurate to 0.25%. At least two instruments are
used for reservoir level measurement. Their results are averaged and compared,
and if they deviate by more than a predetermined amount a warning is registered
in the control room. The current trend is to use three instruments on a `voting'
basis, that is, the three signals are compared and if one deviates from the other
two its reading is ignored. Again, this actuates a warning signal.

Discharge from a gate

Discharge from a gate cannot be measured directly. It is usually inferred from
measurement of upstream water level and gate opening.
Water level should be measured in a stilling well to avoid false readings
which would result from the velocity head in the sluiceway or its approaches.
If means are provided for isolating the stilling well from the sluiceway
approaches, the instruments can be calibrated and checked independently of
the water level to be measured. A gauge board should be provided for checking
the instruments, or if this is not practical, a piezo-electric water level gauge.
It is current practice to provide three pressure transmitters arranged on a
`voting' basis. However, provision of three instruments of the float
displacement kind requires appreciable space and increases cost. For reliable
measurement at least two instruments should be used.
Discharge under gates is determined using the relationship of gate opening
and head above the weir crest. This can be established from a model study. Most
spillway sluiceways operate as if they are independent of adjoining sluiceways,
provided there is no significant superelevation. Thus there is no need to run the
model for various combinations of gates and gate openings which arise if one
gate is out of operation during a flood, due to maintenance, or if a gate fails to
open due to malfunction.
To determine discharge over a gate, the total head above the weir crest and
the weir coefficient must be known. For a valve, the head upstream of the valve
and the valve rating curve are the measurement parameters. The coefficient of

223
 Hydraulic gates and discharge for most valves is not constant throughout the range of valve
valves openings and a rating curve is therefore required. However, in small closed
conduits direct measurement of flow is usually possible.

Water level measurement instruments

Electrodes
Electrodes operate as on/off devices. A number of electrodes can be used to
provide cascade control where a gate or gates open in steps depending on water
level. Each step corresponds to the setting of a limit switch.
For level control two electrodes are used, set apart by the dead band, the
upper and the lower limit of control. The upper electrode initiates raising of
the gate in steps when it becomes submerged. De-energising of the electrode
causes the hoist motion to stop. When the water level exposes the lower
electrode, closure of the gate is started in steps.
The upper electrode is usually duplicated, one acting as a standby to the
other. An additional electrode is sometimes provided at higher level to warn
when a danger point has been reached.
Electrodes can sometimes be energised by dripping water, and can be
protected against such inadvertent activation by a sheath of nylon or PVC.

Level gauges of the float actuation type

The float is linked by a tape to the instrument which displays level on a circular
dial, usually with two hands, like a clock. The float is steadied by guide wires.
The measuring tape actuates a sprocket wheel which operates the hands through
reduction gearing. Float level gauges are available with electrical analogue or
digitally measured value transmission, and with a series of contacts for
signalling high, low or intermediate levels. The measuring range can be up to
30 m. While the accuracy of the mechanical reading is about 2 mm the
electrical transmission, accuracy and bias depends on the range of level the
instrument has to cover.

Pressure transmitters
Pressure transmitters are used for depth measurement. They are encapsulated
integrated silicon strain gauge bridges. They are available for a range of
pressures from zero up to 500 bar and even higher. For water level
measurement, vented gauges are used with a conventional 4^20 mA range.
Linearity and hysteresis are obtainable to 0.1%. For good accuracy, it is
advantageous to select a pressure transmitter which only just exceeds the
required range of water level. Most pressure transmitters have a high overload
capacity. For high reliability, three pressure transmitters are arranged on a
`voting' basis.

Precision pressure balances or bubbler devices

The instrument supplies compressed air through a pneumatic tube to the point
where the water level is to be measured. A small stream of air is allowed to
bubble into the water at the discharge nozzle. The air pressure at the nozzle is

224
proportional to the depth of the water at that point. This pressure is transmitted Control systems and
via the pneumatic tube and a service unit to the receiver, which applies a force operation
proportional to the value to be measured to the balance beam system. When the
equilibrium of the balance is disturbed by a change in the measured value, the
displaced beam triggers a control contact. The servo-motor is actuated and
moves a travelling mass until the equilibrium of the balance is re-established.
The servo-motor and the travelling mass are connected via gearing to a digital
display counter and analogue or digital switching and transmission units.
Bubbler devices can be accurate to 0.25%.
Devices are arranged so that a blocked measuring nozzle can be cleaned by
the application of full compressor pressure. Operationally, bubbler devices
require more frequent checking and maintenance than float level gauges or
pressure transmitters. They are frequently employed to measure head loss across
a screen, or where a wide range of water level has to be measured accurately. In
practice bubbler devices are often found to be non-operational due to lack of
maintenance or incorrect setting by inexperienced staff.

Gate or valve position measurement

This is carried out by indicating transducers. These can measure angle of
rotation or linear movement. Angular transducers convert angular deflection
into a load-independent analogue signal. The input shaft is coupled to a
reduction gear which drives the transducer via a friction clutch. A visual angle
of position indicator is frequently incorporated. Some instruments are available
with an integral second angular transducer to give greater accuracy over a
limited range. When the range of displacement of one transducer is exceeded,
the second transducer is coupled in and rotates to the end of the required range.
Limit switches or changeover contacts form part of some instruments.
Angular transducers are also used to measure linear movement, converting
linear displacement to angular motion by passing a cable over a pulley or by
linkage.

References
1. Lewin, J (1985): The control of spillway gates during floods, 2nd Int. Conference
on hydraulic aspects of flood and flood control, Cambridge, B.H.R.A., Fluid
Engineering.
2. Lewin, J; Denham, H (1983): An adaptive control system for flood routing
through a reservoir, 1st Int. Conference on hydraulic aspects of flood and flood control,
London, B.H.R.A., Fluid Engineering.
3. Anon (1976): Flood control by reservoirs, Chapter 6, Spillway operation, Section 6,
Considerations for spillway operation, Hydrologic Engineering Centre, US Army
Corps of Engineers, Feb.
4. Evans, T E; Halifax, P J; Floyd, D S (1983): A real time computer operated model
developed for the River Medway Flood Storage Scheme, 1st Int. Conference on
hydraulic aspects of flood and flood control, London, B.H.R.A., Fluid Engineering.
5. Back, P A A; Wilden, D L (1988): Automatic flood routing at Victoria Dam, Sri
Lanka, Commission Internationale des Grands Barrages, 16th Congress, San Francisco,
Q63, R52.

225
 Hydraulic gates and 6. Di Stefano, J J; Stubbard, A R; Williams, I J (1976): Feedback and control systems,
valves McGraw-Hill.
7. Gosschalk, E M; Longman, A D (1985): Sri Lanka's Kotmale Hydro Project,
International Water Power and Dam Construction, Mar.
8. Lewin, J (1987): The spillway gates and bottom outlet of Kotmale Dam,
International Water Power and Dam Construction, Aug.
9. Bremen, R; Martinez, R E (2000): Installation of gates on the spillway of the Pueblo
Vieja Dam, ICOLD 20th Congress, Beijing, Q.79^R.15, pp. 209^225.
10. Combelles, J; Tinland, J M (1984): Operation of hydraulic structures of dams,
Commission Internationale des Grands Barrages, Bulletin 49 Appendix 1,
Caderousse on the Rhone River, France.

226
 12
Hazard and reliability of
hydraulic gates

Dam safety has been examined and an extensive technical literature exists on the
subject. Statistics of dam failures have been collected and analysed.1
Corresponding investigations into the hazard and reliability of reservoir
appurtenances are more recent. There is a greater awareness that the integrity
of a dam installation includes the reliability of gates controlling flood release and
the facility to empty a reservoir if a fault develops.
In an analysis of causes of embankment incidents and failures, according to
USCOLD (USNRC, 1983),2 2% of 240 dams experienced malfunction of gates.
Since the publication of this analysis a few catastrophic events have been
recorded involving spillway gates and bottom outlets, and a number which
demonstrated risk.
The most serious risk is posed by common cause failures, that is failures
which affect the operation of a total system. These may consist of failure of the
mains supply and back-up system, failure of central control systems, fire,
explosion, an aircraft crash or ship collision in the case of barriers, or a natural
disaster such as an earthquake.

Events at spillway gate installations

The results of a study carried out in Sweden3 indicated that serious incidents or
breakdowns caused by spillway gates were rare.
In 1967 a spillway gate 12 m high and 9 m wide on the Wachi Dam in Japan
collapsed suddenly.4 It was swept downstream. The cause was dynamic
instability induced by eccentricity of the trunnion bearings.5,6 (See Fig. 12.1.)
Eccentric trunnion bearings are deliberately introduced on some large radial
gates to reduce the hoisting effort. When this is practised, it is advisable to check
that the damping forces compensate for possible dynamic instability.
On 17 July 1995, spillway gate No. 3 of the Folsom Dam on the
American River in California collapsed7 and released a flow of approxi-
mately 1130 m3/s to the Lower American River. The gate was 15.24 m high
and 12.8 m wide. The failure occurred when the reservoir was nearly full
(Figs 12.2 and 12.3).

227
 Hydraulic gates and
valves

Figure 12.1. Spillway gate

vibrations leading to the collapse
of a gate (after Ishii et al.5)

Corrosion on the steel trunnion pins had increased trunnion friction over
time. Collapse occurred when a strut brace in one of the radial arms sheared at
its connection (discussed in Chapter 7).
The possibility of failure had existed for some time and could have been
predicted, however, this was the first actual case. Was it recognised that the

228
 Hazard and reliability
of hydraulic gates

Figure 12.2. Discharge at the

Folsom Dam due to the collapse
of a spillway gate (after US
Bureau of Reclamation7)

design of the lubrication system and the choice of materials for the bearing
system were vulnerable? Was this not modified because of the number of similar
installations which appeared to have operated satisfactorily?
In 1992 a spillway gate malfunctioned at the Tarbela Dam, Pakistan,8 when it
became stuck during a lowering operation. It collapsed, breaking two hoist ropes,
damaging the gate and the weir. The gate was 28.6 m high and 15.2 m wide. Over a

229
 Hydraulic gates and
valves

Figure 12.3. Collapsed

spillway gate at the Folsom
Dam (after US Bureau of
Reclamation7)

long period, the clearance between the side-sealing plates on the piers (the seal
contact plates) and the clamping bar securing the rubber seal on the gate had
deteriorated. The cause of the dimensional change was not reliably established.
An instance of failure of a dam due to inability to open the spillway gates
occurred in Spain.9
In 1988 at the Seton Dam in Canada10 the wire lifting ropes for the 8.1 m wide
by 10.3 m high radial gate broke during a scheduled operation, allowing the gate

230
^ which was about half open ^ to fall on the sill, causing some structural damage. Hazard and reliability
The pin connections of the ropes to the gates had seized, causing the wire rope to of hydraulic gates
bend acutely: first one side snapped, allowing the gate to twist and jam, then the
second snapped. After repair, the same gate was damaged again in 1989 when
local frost expansion packed shut the operating contacts on the hoist motor. It
wound up the gate to beyond its stops until the motor fuse was blown, but not
before structural damage was caused. (Why did the overload protection fail to
come into operation?)
The Canadian Terzaghi Dam10 has vertical-lift gates 7.6 m wide by 10.7 m
high moved by screw-stem hoists. In 1994 one was damaged when the
downward power drive was not stopped in time and the gate was forced on
the sill. During a dewatering test in 1995, one of the gates jammed in the fully
open position. This was due to congealed lubricant and wind-blown dirt on the
exposed long screws.
A similar event occurred at Bray Weir on the Thames, where the open screws
operating double leaf vertical-lift gates jammed due to congealed lubricant,
dead insects and wind blown dust. In this case the problem was aggravated by
a high lead angle of the screws due to a four-start thread. This increased the
screw friction.
A spillway gate of a Swedish dam collapsed3 due to debris accumulation.
Also in Sweden, a serious breakdown occurred during the remote control of a
sector gate3 due to the gate passing the upper limit switch. The bolts on the gate
bearings sheared, causing the gate to break loose and to move down the
spillway.
At the Jackson Meadows Dam in California, the trunnions of the three radial
spillway gates became displaced.11 The dam, completed in 1995, has gates 9.1 m
wide by 4.6 m high. The displacement varied under load and temperature
conditions and was the result of failure to post-tension the trunnion assembly
anchor bolts. Part of the displacement was due to elastic deflection of the
trunnion assembly.
An example of serious gate vibration occurred at one of the spillway gates at
Dundreggan near Loch Ness in Scotland.12 Vibration at low gate opening
caused numerous fatigue cracks at stress concentrations. The cause was flow
reattachment at the gate lip. The gate had been installed during the 1950s and
had previously been operated at larger gate openings.
It should be a matter of course, on changing operational procedures, that the
possible technical consequences are investigated; also that gate vibration should
be reported as soon as it is noticed, even if it is confined to a limited range.
At a large multigate sluice installation, self-exciting wave oscillations
occurred in the upstream basin when six openings discharged while four others
were closed by gates.13
Some operational failures of gates due to severe winter conditions are
mentioned in Chapter 13 on ice formation, and while two examples of gate
vibration are included in the selection above, Chapter 10 on gate vibration is
more representative of the problem as a whole.

231
 Hydraulic gates and Table 12.1. Summary of published information (not comprehensive) on major failures of spillway gate
valves installations

Failure Type of failure Country

Dam failure due to gates Malfunction of gates Russia
failing to open (7 failures) (fatalities)19
Power supply20 Romania
Power supply18 India
Power supply (fatalities)9 Spain
Vibration of outlet gate21 India
Power failure19, 22 Romania
Structural failures of Trunnion friction7 USA
spillway gates (5 failures) Ice loading/brittle
fracture23 Russia
Vibration4 Japan
Vibration12 UK
Debris accumulation3 Sweden
Hoist failures Damaged hoist Hungary
Hoist failure/motor stalled USA
Hoist rope failure/jammed South Africa
gate
Hoist chain failure Spain
Hoist chain failure Canada
Limit switch failure Sweden
Hoist rope sheave seizure Canada
Hoist rope failure Canada
Hoist failure Australia
Hoist rope failure/gate Pakistan
jammed
Hoist failure/gate jammed Canada
Brake failure/runaway gate UK
Brake failure/runaway gate Sweden
Controls Overtopping of dam and Spain
erosion
Condensation of electrical Slovenia
contacts
Inadvertent opening France
Seals leakage Ice formation Canada
Ice formation Sweden
Ice formation Norway

Incidents and failures of bottom outlets

There are a number of research papers on hydrodynamic problems which have
occurred at bottom outlets. Because bottom outlets experience high velocity
flow compared with spillway gates, hydrodynamic problems are more frequent.
Bottom outlets have failed to open due to silting. At the Barasona
Reservoir in Spain,14 the silt extended to a depth of 20 m adjacent to the

232
 Hazard and reliability
of hydraulic gates

Figure 12.4. Collapse of No.

2 diversion tunnel of the Tarbela
Dam, caused by cavitation

dam and had completely blocked the outlet. The problem became dangerous
following a major storm in 1993.
A number of bottom outlets are never, or rarely, exercised. A survey of
reservoir appurtenances at dams in Indonesia identified some bottom outlets
which had not been operated since impounding of the reservoirs. These are
not isolated cases. Similar situations were noted in Sweden.3 Seals under
high pressure are subject to contact welding over time. Gates and bottom
outlets which have not been regularly moved may be difficult or impossible
to raise.
In 1974, the diversion tunnel of the Tarbela Dam on the River Indus in
Pakistan collapsed during the construction phase due to cavitation damage.
The cause was the sticking of a control gate in a partially open position. This
was one of the most destructive failures of a tunnel gate.15

Table 12.2. Summary of published information (not comprehensive) on major failures of reservoir bottom
outlets

Failure Type of failure Country

Bottom outlet control Silting of inlet Spain
gates or valves Silting of inlet Romania
Cavitation Pakistan
Structural collapse of gate Romania
Structural collapse of gate Romania
Serious vibration (9 cases) Romania
Cavitation of discharge valve New Zealand
Cavitation of discharge valve Australia
Cavitation of discharge valves Turkey

233
 Hydraulic gates and When the discharge from a tunnel gate does not result in supercritical flow,
valves the possibility of creating highly sheared flow is present and the Tarbela
incident illustrated its destructive power.
A comprehensive survey of the operation of bottom outlets at 50 large dams
was carried out in Romania.16 While it may not be representative of experience
in other countries, significant deterioration, incidents and failures were
recorded. Damage had occurred at 38 gate installations. 60% of the incidents
and failures were due to vibration problems, including two structural failures
which occurred after 8 and 20 years' operation. Four instances of intake
clogging made the bottom outlets unavailable and nine vibration problems
were classified as `serious'.

Fault frequency by gate type

Lagerholm3 published the results of a questionnaire sent out to major hydro
power and water regulation authorities in Sweden. A breakdown of the fault
frequency per 10 years by type of gate is shown in Table 12.3.
There is very little published information on failures of river control gates.
Experience suggests that the incidence of gate vibration is higher than at
reservoir control gates.
The operation of barrages beyond their working life presents a considerable
risk. For example, large barrages were constructed on the Indian subcontinent
in the interwar years. Failure of one of the gates of the Sukkur Barrage on the
River Indus in Pakistan resulted in replacement of all the gates. At the Kotri
Barrage, also on the Indus, the possibility of imminent failure of several gates
was prevented by fabricating new gates.

Frequent operational problems or deficiencies

power supplies (operating machinery, control systems and operation)
limit switch function (detail design aspects)
trunnion bearing problems (detail design aspects)
ice problems ^ examples of operational problems have been recorded in
Northern Europe, Canada, some states of the USA, Russia, Sweden and

Table 12.3. Fault frequency per 10 years by type of gate (after Lagerholm, 19663)

Type of gate Number of gates Fault frequency Fault frequency

in survey per 10 years % per gate per
year
Radial gates 362 235 6.5
Sector gates 107 125 10.0
Vertical-lift gates, 590 770 13.1
roller type
Vertical-lift gates, 2418 73 0.3
sliding type
Needles 944 11 0.1
Stoplogs 433 44 1.0

234
 Norway; however, such problems must exist in other parts of the world Hazard and reliability
subject to severe winter weather (extreme environmental factors) of hydraulic gates
seal leakage, which can cause gate vibration and in winter can result in
freezing of gates (see also sections on detail design aspects, extreme
environmental factors and gate vibration)
failure of heating systems
failure of hoisting systems ^ ropes, winding screws, wedging of gates, motor
overloads, brake failures, chains
gate vibration
discharge valve cavitation
cavitation at the bottom outlet of high head dams
silting of the intake of bottom outlets
lack of regular exercise of bottom outlets (also mentioned by Lagerholm3)
floating debris in extreme floods
electrical cable fractures
control system malfunction
instrumentation.

Control system failures

Incidents involving self-induced operation of gates under automatic control
have occurred. Rajar and Rryzanowski17 have recorded the self-induced
opening of spillway gates on the Mavcice Dam in Slovenia. Two radial gates
20 m high and 13.5 m wide opened, discharging at a rate of 1192 m3/s,
equivalent to a 50-year return period flood. Other incidents of uncontrolled
gate openings have occurred but have not been publicised because they have
not resulted in loss of life or damage.
A number of gate designers and reservoir operators specify that automatic
gate control systems are backed up by hard-wired electrical circuits which
inhibit the duration over which gates can operate or the distance travelled
following a command to move a gate.

Risk assessment of gated hydraulic structures

There is sufficient evidence of failures and malfunction of reservoir
appurtenances for spillway and bottom outlet gates and valves to be included
in the risk assessment of a dam.
Various detailed hazard and reliability assessments of estuarial flood barriers
have been carried out. Until recently, more detailed assessments of reliability
were carried out on barriers than on reservoir flood control structures. This
may be because a barrier is constructed to prevent flooding, while the function
of a reservoir may be electricity generation, water supply or flood storage and
the hazard is perceived as a consequential risk. However, this is changing as a
result of awareness of the risk of reservoir systems, not just of a dam.
A risk assessment of a gate installation should include the total system. In the
case of a reservoir it starts at the inflow characteristics. Where the reservoir is
one of several on the same river, the operation of a control structure upstream
can have an important effect.

235
 Hydraulic gates and Even a relatively small gate opening will cause a surge wave to travel up a
valves reservoir or a reach of river. On reflection it can affect the water level recorder
and in an automatic control system the reflected wave can initiate opening of a
gate. When the ponded-up water is limited it can cause severe instability, as
described in Chapter 9.
Some methods of gate operation can be a danger to river users downstream of
a dam or barrage and may even cause local flooding. Fatalities have occurred due
to excessive continuous opening of gates at the onset of a flood.
Vandalism can result in similar problems. Shaw and Hakin24 record that
vandals gained access in July 1998 to the Windsor Dam near Ladysmith in
South Africa. While there is security at the dam it is not permanently staffed.
The vandals penetrated to the control panel and raised the 12 m wide by 8 m
deep main sluice gate by 500 mm. A large quantity of water was released into
the River Klip at a time when flow is usually at its lowest. Six children playing in
the river downstream of the dam were caught unawares by the sudden flow rise
and were washed away. Two children were rescued, but four died.
The determination of reliability must include potential shortcomings due to
design, operation, maintenance, operator training, inspection and supervision.
Common cause failures predominate, affecting the whole installation.
A ship collision occurred at the Thames Barrier in October 1997. The vessel
Sandkite collided with one of the piers and partially sank (Fig. 12.5). It
discharged its cargo of sand and gravel onto one of the rising sector gates;
fortunately there was no significant structural damage. If the gate had been
rendered temporarily inoperable, it could not have been closed when a tidal
surge was expected. The potential for such an accident exists at barriers with
intermediate piers.
Maintenance can be deficient, of a variable standard at different stations, or
completely neglected because there is no maintenance budget or authority to

Figure 12.5. Sandkite

collision with a pier of the
Thames Barrier, October 1997

236
order spares, as was the case at hydropower plants in a tropical country. Stock- Hazard and reliability
keeping of spare parts for gates and auxiliary equipment is absent in many cases. of hydraulic gates
Many reservoir control structures are 30^40 years old; spare parts are no longer
obtainable even for mass produced articles. A reliability assessment should
include a parallel contingency plan covering how replacement or substitution
of potentially vulnerable parts or components can be carried out.
Marking and documentation of spillway gates and auxiliary systems is often
indistinct or defective.
In a reliability assessment, factoring the age and condition of the plant is
difficult and depends on judgement. To reduce the uncertainty, recording of
comprehensive and systematic data from maintenance and failures should be
practised. This can provide information for an analysis of wear, degradation
due to ageing of plant or machines, and incipient failures where a large number
of similar components are used. Limit switches, oil hydraulic valves, seals and
electrical relays are examples of devices which can be present in tens or hundreds
in gate control and hoist systems.
In a high-level assessment of failure probabilities human factors are likely to
dominate, such as availability of operating staff, events which prevent staff from
reaching their control station, misreading of instruments, incorrect action,
failure of communication and failure to follow laid down practice.
A reliability assessment considers all the factors which can cause
unavailability of safety critical hydraulic structures. It can also be expanded to
quantify the contribution of each event or action to failure of the top event,
which is the safe passage of a flood or, in the case of a tidal barrier, prevention
of the flood.
The operational and physical condition of gate systems often varies during
the lifetime of the installation. This can apply to spillway gates, and even more
to tidal barriers. Developments and encroachments on the river downstream of
a dam can affect the consequences of operational procedures at the spillway.
At tidal defence barriers, the rise in water level due to global warming will
cause higher and more frequent storm surges. These will have wide ranging
consequences.
Risks which were of a low order when a structure was first commissioned can
become more significant. For instance, at the time of carrying out a risk
assessment of the Thames Barrier in 1988,25 the probability of an aircraft crash
affecting any of the barriers was considered very small and therefore did not
warrant further consideration. Since then, the London City Airport has started
operation. Statistics indicate that the areas for peak probability of aircraft crash
now include some of the barriers on the Thames.
Spillway gate and barrier control systems are sometimes upgraded. For
example, at the spillway gates at Dundreggan12 automatic controls were
introduced; at the Thames Barrier the original relay-based control system was
replaced by a programmable logic control (PLC) system because of problems
with the previous installation. Changes can also comprise alterations in working
practice, manning, training and the retirement of experienced operating staff.
A problem of this kind affecting spillway gate operating personnel in rural
South Africa has been reported by Shaw and Hakin.24 As a consequence of the
increasing influence of HIV and AIDS, negative population growth rates across

237
 Hydraulic gates and southern Africa are causing loss of key trained personnel in rural areas where
valves they are difficult to replace, and where the provision of adequately skilled
back-up personnel is simply not possible.
Reliability of mechanical equipment can be affected by the consequences of
alkali aggregate reaction of the civil engineering works. Notable examples are
the Kariba Dam and Owen Falls Dam. After symptomatic repairs, these dams
are now monitored through management programmes. On a smaller scale, in a
rural environment in Africa, repair before the problem reaches a critical point is
not always realistic.24
Reliability assessments must, therefore, be regarded as a time dependent
overview of the causes of unavailability of gated structures. They should be
updated at intervals to take into acount management, operational,
environmental, technical and hydraulic changes.

Techniques for analysing risk

There are a number of definitions of risk. The simplest one is `the likelihood of
occurrence of adverse consequences'.26 For the purpose of quantifying risk, the
definition by BC Hydro27 is more useful: `A measure of the probability and
severity of an adverse effect to health, property, or the environment. Risk is
estimated by the mathematical expectation of the consequences of an adverse
event occurring (i.e. the product of `probability  consequence').'
Risk analysis must by definition include probabilistic events, although they
may sometimes be implicit. Risk assessment is a combination of art, judgement
and science, in that order, constrained in a formalised process.28
The most detailed methods used in risk analysis are fault trees and event
trees. Fault trees allow the diagrammatic presentation of components that
may lead to failure of a system element. A general failure event ^ the event
to be analysed ^ is at the top of the fault tree, the remainder of which is formed
by specific events which can potentially lead to the failure. Analysis of the
fault tree results in determination of minimal cut sets, the minimal
combination of events which cannot be reduced in number and whose
occurrence cause the top event. Calculation of the probability of occurrence
for each minimal cut set is carried out from the probabilities of the basic
events. An example of a simplified fault tree for spillway gate failure is
reproduced at the end of this chapter.
An event tree represents all the possible sequences of events which could
result from a given initiating event. Unlike a fault tree, it works from the specific
to the general, tracing how failure sequences propagate. Branching is limited to
`yes' or `no' at each system response. There are similarities with operational logic
diagrams. An example of an event tree for a seismic event on a dam is included in
Chapter 14.
For analysing complex systems, computer programs have been developed to
handle elaborate schematic structures. The one best known in the UK is AEA
Technology's Fault Tree Manager. A previous version, `Orchard', was used in a
reliability assessment of the Thames Tidal Defences25 and the barriers for the
flood prevention of the City of Venice.32 Hoyland and Rausand29 discuss other
programs.

238
 Other techniques used to identify failure modes consist of structured Hazard and reliability
questions which help to analyse the system. Examples are failure modes and of hydraulic gates
effects analysis (FMEA), and failure modes, effects and criticality analyses
(FMECA). BS 5760: Part 5: 199130 describes these as methods of reliability
analysis intended to identify failures which have consequences affecting the
functioning of a system within the limits of a given application, thus enabling
priorities for action to be set. Hazard and operability study (HAZOP) is another
analytical tool which concentrates on identifying deviations from design and
operating conditions. These techniques use worksheets which are filled out
during analysis of a system to document a qualitative assessment.
In dam engineering, a probabilistic risk analysis (PRA) is used as a basis for
making decisions when selecting among different remedial actions, and to
determine priorities. Access to previously collected statistics is helpful but is
not essential for PRA. When assessing gates and valves, an analysis of service
records is a useful guide. This is also used when assigning failure probabilities to
fault and event tree branches.
Reliability assessments based on fault trees have been carried out for the
Thames, Barking Creek and other storm surge barriers comprising the Thames
Tidal Defences,25 also two reliability assessments of the design of the barriers for
the flood defence of the City of Venice.31,32
The Rykswaterstaat has carried out similar assessments on barriers for flood
protection in The Netherlands. Some results of a risk assessment of the New
Waterway Storm Surge Barrier were given in papers by Ieperen33 and Janssen et
al.34 At spillway gate installations a fault tree reliability assessment was carried out
as part of the deficiency investigations for the Seven Mile Dam in British
Columbia.35 Thiswasamajorundertaking thedocumentation oftheinvestigation
of reliability for normal conditions is extensive, comprising three volumes.
Lagerholm3 mentions that fault tree analysis has been performed in Sweden
on different types of spillway gate functions. The wording of his paper suggests
that these were not total system assessments.
The construction of fault and/or event trees and the production of minimal cut
sets, together with the computational work required, involves considerable man
hours. This type of analysis is considered justified in special cases such as the Seven
Mile Dam, where the operation and reliability of the spillway and drainage systems
are crucial to the safety of the dam, or the Folsom Dam where collapse of a spillway
gate has resulted in consideration of a fault tree assessment of the spillway system.
In most spillway systems there should be adequate redundancy so that the
malfunction of a gate does not result in a serious risk. If redundancy is provided,
overall system reliability depends more on common cause failures, that is, on an
event which affects the total installation. However, redundancy of gates is rarely
provided for an extreme event.
Even failure modes, effects and criticality analyses (FMECAs) and hazard
and operability studies (HAZOPs) can involve much technical manpower.
They are usually carried out by a team of engineers and technicians familiar with
an installation, and can result in lengthy evaluation of specific elements of the
control structures.
Where the operator of several dams requires an initial hazard assessment of a
number of reservoir appurtenances of different design and age, methods of

239
 Hydraulic gates and assessment based on the systematic application of engineering judgement are
valves sometimes used. In Norway, a simplified risk analysis is being applied to dam
safety. Scottish and Southern Energy36 uses a similar approach to determine
priorities for maintenance and improvement of spillway gate and reservoir
bottom outlet structures.
Fault trees and minimal cut sets are important tools for assessing the
reliability of a total installation and for quantifying the contribution of
subsystems and major components to the failure of the top event. They are
not, as a rule, extended to include details such as limit switches, an important
vulnerable element of gate hoists, local leakages of seals which can cause gate
vibration, freezing up of side seals at spillway gates in winter, and so on. Unless
data pertaining to operational problems over an extended period of time are
available, it is difficult to assign failure probabilities to these and similar
elements. This does not apply to the electrical supply and distribution systems
of spillway gates and bottom outlets. General and detailed statistical
information is available to assign a failure probability to each element, and the
result will more accurately reflect the failure probability than the parallel
assessment of gates and their mechanical features.
Fault tree reliability assessments are a valuable tool to determine the overall
integrity of an installation in relation to the risk to the dam and reservoir. The
inclusion in the risk assessment of management, operator training, operational
procedures, communication, possible malicious action and failure of advance
warning systems results in a comprehensive assessment. In barriers, ship
collision is an important risk factor and, more remotely, an aircraft crash.
Operationally, engineering assessments are required when the main objective
is to determine the adequacy of maintenance, elimination or improvement of
features or elements which are vulnerable, and reduction in the probability of
failures which can put a gate out of operation. A structured assessment system
based on engineering judgement is probably the best means of achieving this.
A reasonable record of experience is available relating to design features of
gates and valve systems which are likely to result in operational problems, or are
indicators of risk. Instead of structured, generalised questions which are the
basis of HAZOP, more specific charts could be devised for carrying out
reliability assessments, perhaps taking the form of diagrams and description of
design features, or assigning a number to the condition of a component. The
sum of the numbers would form an index of priorities with individual high
numbers drawing attention to areas of urgent action. It would not form a
probabilistic index, but if well constructed could be part of a risk assessment.
Reservoir control appurtenances are designed for extreme events and few
have been subjected to exceptional loading. However, hydraulic conditions
which cause gate vibration, while not necessarily extreme events, may not occur
for years after commissioning. Such conditions, combined with structural,
mechanical and electrical deterioration, can cause a risk and hazard because they
are the coincident event of a number of probabilities. In a formal probabilistic
investigation they may not show up, because this assumes that at each demand
the structural, mechanical and electrical condition remains the same and that no
deterioration has occurred. To factor wear and deterioration is difficult and
quantifying it depends on judgement, subject to wide latitude.

240
Reliability indices Hazard and reliability
of hydraulic gates
The probabilistic reliability derived from a fault tree analysis can be expressed as
failure per demand (in the case of a spillway gate, this is the opening of the gate).
For the Thames Barrier, this was determined at 1.55  104 per gate per
demand.25 Expressed differently, there is a chance that a single gate will fail to
close on one in 560 closure demands, and that two of the ten gates will fail to
close at one full closure in approximately 6000 closure demands. This reliability
assessment was carried out before control of the Thames Barrier was upgraded
by replacing the relay systems with programmable logic controllers (PLCs).
In the hazard and reliability study of the Flood Prevention Scheme for the
City of Venice, failure was defined as flooding of Venice more than 280 mm
above Venice datum. The design resulted in 1 event in 800 years.32
For the New Waterway Storm Surge Barrier in The Netherlands,34 the
derived reliability targets were:
probability of not closing due to human or technical errors less than 103 on
demand
probability of collapse less than 106 in any year
probability of not opening due to human or technical errors less than 104 on
demand.
For the Seven Mile Dam in British Columbia, reliability analysis35 resulted in:
probability of failure of spillway gates to open due to environmental hazards
9.68  106
probability of failure of spillway gates to open due to electrical or mechanical
failures 2.07  107
probability of power supply unavailability to the spillway gates 2.07  107
A good industrial system standard is one failure in 104 per demand. The
reliability of a spillway gate installation depends on whether all the gates can
pass the probable maximum flood (PMF) or the half PMF. The usual standard
adopted in multigate spillway systems is that a thousand year return period
flood can be passed with one gate out of operation. A failure rate of 104 per
gate per demand would appear to be adequate under these conditions. If the
gates cannot pass the PMF a lower failure rate would be appropriate. This would
depend on the hazard resulting from a gate failing to open under flood
conditions. Some spillway gates, especially older ones, would not qualify for a
failure rate of 104 per demand let alone a more severe criterion.
Bottom outlets frequently consist of a single operating gate with a back-up
gate or discharge valve backed by a butterfly valve or a gate. Two or more
parallel fluidways are less frequent. It is suggested that a failure rate of the order
of 105 would be appropriate where only one gate with a back-up gate is
provided. Whether greater reliability is required than for a spillway gate
installation depends on the risk associated with failure of the bottom outlet.
Failure probability ratings for electrical services, both for details and systems,
are available and are statistically valid. These include standby generating plant.
For spillway gates, their hoisting machinery and control systems, failure
probabilities have to be assessed from service records, known incidents or

241
 Hydraulic gates and structural and mechanical plant which have some similarity. The available data
valves will probably be of low statistical validity. The selection of a failure probability
for each item of a fault tree branch will therefore involve a significant element of
engineering judgement. Such judgement, whether exercised by an individual or
collectively, depends on experience.
Problems encountered with bottom outlets often stem from the interaction of
structural and mechanical aspects with hydrodynamics. Assignment of failure
probabilities to fault tree branches when investigating bottom outlets is therefore
even more dependent on judgement and knowledge of theory and practice.
Some technical papers record failure events which resulted in serious hazards.
This would not have been highlighted in a conventionally constructed fault tree
and would have resulted in a low failure probability. Integrating experience and
knowledge on a wide scale may identify areas where hazards resulting from a
rare combination of factors would result in a different construction of a fault
tree, or simply indicate the need for remedial action.

Fault tree for spillway gate failure

The fault tree investigates possible failures which contribute to causing the top
event (see Fig. 12.6). In this case it is `Spillway gate fails to open when required
to discharge flood inflow to the reservoir'. In practice, a spillway gate

Figure 12.6. Fault tree for spillway gate failure

242
installation will comprise a number of gates, and failure of a single gate, or even Hazard and reliability
several gates, would permit the discharge of a flood less than the probable of hydraulic gates
maximum or the design flood. Since interest is in a complete failure of the
spillway capacity, the analysis is likely to be dominated by events with the
potential to affect all of a set of multiple gates. Random independent failures
of multiple components are clearly possible but the probability of occurrence
is likely to be of a much lower order than common cause events, that is, an event
which affects the total installation ^ for instance, failure of the mains supply.
While multiple independent failures have not been shown in the fault tree, they
would be included in a real analysis.
The increase in inflow to a reservoir during a flood varies depending on the
geography of the catchment area feeding the river or rivers discharging into the
reservoir. Snowmelt can usually be predicted well in advance, whereas a steep-
sided reservoir in a mountainous area would exhibit a very rapid increase in the
rate of inflow, which could be as short as a few hours. A fault tree would usually
omit any failure mode which can be rectified within the period between the
onset of a flood and the time when multiple gates are required to open to
maintain the reservoir level.
The impression created by the fault tree shown here might be that the events
causing failure are fairly obvious. In practice, it would be developed and
expanded to ensure that all events are considered, not just structural, mechanical
and electrical ones. The reliability of an installation depends as much on human
factors, such as the actions of operators, the organisation, the standard of
maintenance, training and communication as on other events. External events,
particularly when they can cause a common cause failure, would be included,
such as a lightning strike, an earthquake, exceptional wave action due to a
landslip into the reservoir or a storm of rare intensity.
The structure of the tree also allows for specific areas where redundancy
provides higher reliability, or where combinations of failures could cause
unexpected results.
The gates of the fault tree, which link the progression from a lower to a
higher event, indicate when a failure can be caused by either of several different
occurrences, so-called `OR' gates (such as gates 3, 4, 5 and 20). The `AND' gates
(7 and 10) require that all the immediate lower events occur to cause failure and
the next higher event.
When a fault tree has been fully developed, it may raise concerns which result
in design changes. Using data on the reliability of components, the fault tree top
events can be quantified. This provides a relative ranking of the identified
failure modes, as well as an estimate of the absolute failure probability. Most
useful is the insight gained into potential vulnerabilities of the system. Even
an approximate quantification can support concerns which are raised by good
engineering judgement and experience.

243
 Hydraulic gates and References
valves
1. ICOLD (1995): Dam Failures: A Statistical Analysis, Bulletin 99, International
Comission on Large Dams, Paris.
2. USNCR (1983): Safety of existing dams: evaluation and improvement, US National
Research Council, National Academy Press, Washington DC, USA.
3. Lagerholm, S (1996): Safety and Reliability of Spillway Gates, ICOLD Symposium,
repair and upgrading of dams, Stockholm, Jun.
4. Yano, K (1968): On the event of the gate destruction of the Wachi Dam, Disaster Prevention
Research Institute, Annals of Kyoto University, Japan, 11-B, 1^17.
5. Ishii, N; Imachi, K; Hirose, A (1968): Instability of elastically suspended tainter-
gate system caused by surface waves on the reservoir of a dam, Am. Soc. Mech, Eng.,
Fluids Eng. Div., Joint applied mechanics, fluids engineering and bioengineering conference,
New Haven, Conn, Jun., paper No. 77-FE-25.
6. Ishii, N; Imachi, K; Hirose, A (1979): Dynamic instability of tainter gates, 19th
I.A.H.R. Congress, Karlsruhe, Paper C9.
7. Bureau of Reclamation (1996): Forensic report of spillway gate 3 failure, Folsom Dam,
Bureau of Reclamation, Mid-Pacific Regional Office, Sacramento, Cal, USA,
Nov.
8. Khan, K A; Siddique, N A (1994): Malfunction of a spillway gate at Tarbela after
27 years of normal operation, ICOLD, 16th Congress, Durban, Q71, R27, pp. 411^
428.
9. Water World (1982): Overtopped Spanish Dam collapses as spillway gates stay shut, 5, No.
11, p. 8.
10. Watson, M A (1997): Spillway gates: Will they open safely, ICOLD, 19th Congress,
Florence.
11. McManus, R A (1999): Measurement of tainter gate trunnion displacements,
Jackson Meadows Dam, California, ASCE Conference Hydro's Future, Las Vegas,
Nevada, Sep.
12. Noble, M; Lewin, J (2000): Three cases of gate vibration, British Dam Society, 11th
Conference, Bath, Jun., in Proceedings, editor Tedd, P. Thomas Telford.
13. Kolkman, P A (1984): Vibration of hydraulic structures, in Developments in hydraulic
engineering 2, editor P Novak, Elsevier, p. 46.
14. Romeo, R (1996): Drawdown of the Barasona Reservoir, report by the author at the
Symposium on repair and upgrading of dams, Stockholm, Jun. Hydropower and
Dams, 1996, Issue 5, pp. 65^74.
15. Kenn, M J; Garrod, A D (1981): Cavitation damage and the Tarbela Tunnel
collapse of 1974, Proc ICE, Part 1, Vol. 70, Feb., pp. 65^89.
16. Ionescu, S et al. (1994): Damage and remedial work during operation of several
bottom outlets, ICOLD, 16th Congress, Durban, Q71, R7, pp. 79^90.
17. Rajar, R; Rryzanowski, A (1994): Self-induced opening of spillway gates on the
Mavcice Dam ^ Slovenia, ICOLD, 16th Congress, Durban Q71, R8, pp. 97^112.
18. Narayana Murty, T V S (1979): Failure of the Machhu-II Dam, Indian Journal of
Power and River Valley Development, Mar., 54^67.
19. Reynolds, P; Hindley, M (1994): Double dam flood failures, Water Power and Dam
Construction, 46, No. 9, p. 2.
20. Utillas, J L; Gamo, A; Soriano, A (1992): Reconstruction of the Tous Dam, Water
Power and Dam Construction, 44, No. 9, pp. 55^65.
21. Sagar, B T A; Tullis, J P (1970): Problems with recent high-head gate installations,
Proc. of international hydraulic research symposium, Stockholm, Sweden, paper F1.
22. Diacon, A; Stematiu, D; Mircca, N (1992): An analysis of the Belci Dam failure,
Water Power and Dam Construction, 44, No. 9, pp. 67^72.

244
23. Freishist, A R; Rozina, I D; Rakhmanov, A L (1976): Operating experience gained Hazard and reliability
with flat hydraulic gates under winter conditions. Gidrotekhnischeskoe Stroitelo'stvo, of hydraulic gates
USSR, No. 4, 348^353; English Translation, Plenum Publishing, New York.
24. Shaw, Q H W; Hakin, W D (2000): Factors of influence on the selection of surface
spillway control systems in developing countries, ICOLD, 20th Congress, Beijing,
Q79, R18, pp. 259^274.
25. UKAEA (1987): A reliability assessment of the Thames tidal defences, Safety and
Reliability Directorate, SRS/ASG/31447. (unpublished).
26. McCann, M W et al. (1985): Preliminary safety evaluation of existing dams, Department
of Civil Engineering, Stanford University, Vol 1, report 69.
27. BC Hydro (1993): Guidelines for consequence-based dam safety evaluations and improvements
(interim), BC Hydro, Burnaby, BC, Canada.
28. Bivins, W S (1984): Risk analysis in dam safety programmes, Proc. Water for resource
development Conference, Coeur d'Alene, Idaho, Aug, editor Shreiber, D L ASCE,
New York, pp. 115^119.
29. Hoyland, A; Rausand, M (1994): System reliability theory, models and statistical methods,
John Wiley & Sons, New York.
30. British Standards Institution (1991): BS 5760: Part 5, Guide to failure modes, effects and
criticality analysis (FMEA and FMECA).
31. AEA Technology (1989): Reliability study for the Venice flood defences, Part 1: reliability
assessment of gate performance; Part 2: hazard assessment, Safety and Reliability
Directorate, SRS/ASG/31466, Nov.
32. Lewin, J (1993): System reliability assessment of the definitive design of the Venice flood
defences. (unpublished).
33. Ieperen, A van (1984): Design of the new waterway storm surge barrier in The
Netherlands, Hydropower and Dams, May, pp. 66^72.
34. Janssen, J P F M; Jorissen, R E; Ieperen, A van; Kouwenhoven, B J (1994): The
design and construction of the New Waterway Storm Surge Barrier in The
Netherlands (Technical and Constructual Implications), ICOLD, 16th Congress,
Durban, C15, p.877^900.
35. Klohn-Crippen Integ & Northwest Hydraulic Consultants (1996): Seven Mile Dam
deficiency investigations; spillway and drainage systems, reliability for normal conditions, for
BC Hydro, Task C8 ^ Part 1A, Vol 1; Vol 2 ^ Appendices A^L, Vol 3 Appendices
M^0, Sept.
36. Sandilands, N M; Noble, M (1998): A programme of risk assessments for flood gates
on hydroelectric reservoirs, Proc. 10th British Dam Society Conference, Bangor,
Wales, Sept. in The prospect for reservoirs in the 21st century, editor Tedd, P, Thomas
Telford, pp. 27^38.

245
 13
Ice formation

Ice
Criteria for the design of gates under ice conditions in Northern Europe are
given in DIN 19704,1 where different empirical rules apply to inland and
estuarial conditions.

Empirical criteria for inland conditions

In the design of skin plates and their associated stiffening members, as well as for
the main girder, it is assumed that the formal triangular hydrostatic distribution
of water pressure at a depth of 1 m is replaced by:
an even surface pressure of 30 kN/m2 where the ice formation is 300 mm
an even surface pressure of 20 kN/m2 in waters with moderate ice formation
up to 300 mm thick

Empirical criteria for estuarial conditions

In the design of skin plates and their associated stiffening members, the
following loads should be assumed over and above the hydrostatic loads within
0.5 m above and below water level:
an even surface pressure of 100 kN/m2 when severe ice formation is present
and ice displacement occurs
in conditions of moderate ice formation, an even surface pressure of 30 kN/m2
should be used.
For main girders, additional loads should be applied to the node points level
with the water surface amounting to:
a distributed load of 350 kN/m in severe ice conditions
a distributed load of 100 kN/m in conditions of moderate ice formation.
The design of underflow gates should be checked for a distributed load of
30 kN/m at the gate lip. Ice forming within the gate structure should also be
taken into account.
The US Army Corps of Engineers, Engineer Manual, Design of spillway
Tainter gates2 (radial gates) specifies an ice impact load to account for impact

247
 Hydraulic gates and by debris (timber, ice and other foreign objects) or lateral loading due to
valves thermal expansion of ice sheets. This is a distributed load of 73 kN/m acting
in the downstream direction along the width of the gate at the upstream water
level.
Starosolsky3 gives extensive information on ice formation and provides
some data which can be used to arrive at a rational basis of design. Otsubo4
and Johansson5 give examples of the effect of severe ice formation on gates
and some means that have been employed to mitigate it.
Under winter conditions of light frost, the side seals of gates can freeze to
their contact face. An attempt to operate a gate under these conditions could
result in tearing of the seal.
Where there is an operational risk of side seals freezing to their contact face,
either because moisture is trapped between the seal and its contact face or
because there is leakage past the seal, the side staunching has to be heated when
air frost occurs. Lintel seals will also have to be protected against freezing.
Heating of side-seal contact plates for radial gates under frost conditions is
extended throughout the length of travel of the side seals. It is effected by
electrical resistance cables, usually of the mineral insulated, stainless steel
sheathed kind. Figure 13.1 shows such an arrangement. Heating cables have
to be insulated to prevent undue conduction of heat to the flume walls. An
alternative method involves the circulation of hot oil.
Heating provision was made at the Tees Barrage in the UK where it was
introduced for some of the bottom-hinged barrage gates in their closed
position. The seal contact plates of the large vertical-lift gate of the Barking
Barrier on the Thames also incorporate provision for heating.

Figure 13.1. Heating cables for side

seal contact face

248
 In England an approximate heating load of 1000 W/m2 is considered Ice formation
adequate when the gate may have to be moved under conditions of air frost.
For central European conditions an approximate heating load for side seal
contact faces is 1.5 kW/m2. Under severe frost when the upstream and the
downstream water is subject to ice formation to some depth it is not practical
to move a gate and is unlikely to be required.

Operating experience under severe winter conditions

The build-up of ice due to leakage past side seals of radial gates has frequently
occurred. Watson6 gave an example at the Peace Canyon Dam in Canada. In
1979, during the first winter of operation, the six spillway radial gates (15.2 m
wide by 12.6 m high) were found to leak a small amount of water through the
upper 3 m of the rubber side seals, where the head was insufficient to properly
seat the seals. Cold winter temperatures caused the leaked water to freeze, with a
build-up of ice which bridged between the gate and the side walls of the conduit,
potentially impairing safe gate operation. A stainless steel spring plate was
retrofitted to the upper portion of the seals, providing enough precompression
to make a tight seal.
Similar problems have been reported at the Gavins Point Dam.7 The spillway
has 14 radial gates 12.19 m wide by 9.14 m high. Ice forms in the downstream
corners of the gates between the gates and adjacent pier walls. As a result, the
gates are often not operable during the winter months. Side-seal heaters were
originally fitted which proved inadequate. They were replaced by higher
capacity side-seal heaters which added little, if any benefit, and eventually
experienced electrical failure.
It appears that the record of side-seal heating systems is variable. A range of
faults have been reported:8
monitoring of heating equipment is insufficient and not safe or reliable
faults in heating systems have caused freezing
heater cores have burnt out
heating elements have a short lifetime.
The operational record of mineral insulated heating cables is significantly better
than that of self-regulating heat cables. In practice, even and consistent contact
pressure of side seals is not always achieved. The build-up fabricating tolerances
can cause problems. The conventional design parameter for precompression of
side seals is 3 mm interference. The thermal contraction of skin plate assembly
can also be a factor in severe winter conditions. If a gate has been fabricated at
15 C ambient temperature and the outdoor temperature in winter falls 25 C
below the temperature during fabrication, this results in a contraction which
would amount to 3.5 mm in the case of a spillway gate at the Gavins Point Dam.
The hardness of elastomers increases at lower temperatures and flexibility
decreases. At the Gavins Point Dam, J-type side seals of a softer compound were
tried. They reduced leakage but did not provide a watertight seal.
To avoid side-seal leakage during severe winter weather requires greater
dimensional accuracy. The method of mounting side seals shown in Figs 7.2
and 7.5(c) permits some site adjustment and should ensure better sealing.

249
 Hydraulic gates and To prevent ice from forming immediately upstream at the top of gates, air
valves bubbler nozzles can be embedded upstream of a gate. They provide a constant
emission of air supplied by compressors through nozzles set in the concrete
invert. This causes relatively warm water at the bottom to circulate upwards,
preventing ice from forming at the water surface. If the nozzles are not close
enough to the bottom of the curved gate skin plate, ice may form at the bottom
sill. In addition, leakage past the sill seal can create ice which bonds the bottom
of gates to the sill beam.
Air bubbler systems are discussed in the US Army Corps of Engineers
Manual, EM 1110-8-1 (FR)9 and Haynes et al.10
Other problems which have occurred during severe winter conditions have
included:
Absence of a spillway gate building for machinery and control equipment,
allowing hoist machinery to be covered in ice and snow.
An uninsulated spillway gate building, which caused oxidation of electrical
contacts resulting in failure.
Attachments of the hydraulic cylinder operating a bottom-hinged flap gate
located in a pit which had not been drained, causing the piston clevis to freeze
in position.8
Structural failure of a spillway gate occurred on the River Svir in Russia, due
to ice loading and brittle fracture.11
At radial gates with parallel arms, ice has formed between the arms and the
sluice walls, forming a bond.

References
1. DIN 19704 (1976): Hydraulic steel structures: criteria for design and calculation.
2. US Army Corps of Engineers (2000): Engineer Manual 1110^2^2702, Design of
spillway tainter gates, Dept. of the Army, Washington DC, 1st Jan.
3. Starosolsky, (1985): Developments in hydraulic engineering ^ 3, editor Novak, P,
Elsevier Science Publishers, pp. 175^219.
4. Otsubo, K (1959): Ice problems of gates at hydro-electric plant in northern districts
of Japan, 8th I.A.H.R. Congress, Montreal, Vol. III, pp. 2-S1-1 to 2-S1-2.
5. Johansson, H (l959): Ice problems relating to dam gates, 8th I.A.H.R. Congress,
Montreal, Vol. lll, pp. 27-S1-1 to 27-S1-3.
6. Watson, M A (1997): Spillway gates: will they open safely, ICOLD 19th Congress,
Florence, Q74, R4.
7. Bockerman, R W; Wagner, P A (1999): Solutions to spillway tainter gate problems,
A.S.C.E. Conference, hydro's future, Las Vegas, Nevada, Sept.
8. Lagerholm, S (1996): Safety and reliability of spillway gates, ICOLD Symposium,
Repair and Upgrading of Dams, Stockholm, Jun.
9. US Army Corps of Engineers: Engineer manual EM 1110-8-1(FR), winter navigation
on inland waterways, Departmen. of the Army, Washington DC.
10. Haynes, F D; Hachnel, R; Clark, C; Zabilansky, L (1997): Ice control techniques for
corps projects, Technical Report REMR-HY-14, US Army Engineer Waterways
Experiment Station, Vicksburg, MS.
11. Freishist, A R; Rozina, I D; Rakhmanova, A L (1976): Operating experience gained
with flat hydraulic gates under winter conditions (English translation), Gidro-
tekhnicheskoe Stroital'stvo, USSR (No. 4 pp. 348^358), Plenum Publishing, New York.

250
 14
Earthquake effects on gates

The performance and safety of dams during earthquakes worldwide has been
remarkably good.1 Nevertheless, the failure of a dam can have such serious
consequences that earthquake safety evaluation of existing dams and of new
constructions is a general requirement. Following an earthquake, the release
of reservoir water can be a critical control function if the dam has been damaged
by the seismic motion. Therefore, spillway gate installations and bottom outlets
need to remain operational after an earthquake and should be included in the
seismic analysis. As a New Zealand engineer expressed it, `When the big one
hits, the likely scenario is that massive power load will be dropped and spilling
will quickly be necessary to prevent dam overtopping and serious damage to
generating facilities'.2
A review of incidents at dams which have been exposed to seismic events3
shows that dam performance is on the whole good. However, there are a
number of events in which dams have suffered significant structural damage.
In some of these cases the dam has subsequently failed entirely, although this
has rarely happened at the time of the earthquake; most failures have occurred a
few hours later (up to 24 hours after the earthquake itself).
After a dam has experienced a significant seismic event, there is likely to be an
urgent need for the water level in the reservoir to be lowered quickly, both to
reduce pressure on the potentially damaged and weakened structure and to
alleviate the consequences should the dam fail at a later time. Spillway gates
and bottom outlets will be used for this purpose, with the spillway gates
providing the greater initial capacity for level reduction.
There are two primary considerations in the seismic evaluation of any
structure:4
The selection of the seismic design events, the maximum design earthquake
(MDE) and the operating basis earthquake (OBE).
The selection of the analysis method.
The evaluation of the MDE and the OBE are outside the scope of this book.
Guidance can be obtained from references 4, 5 and 6, and for the United Kingdom
from references 1, 5 and 7. Charles et al.1 give information on other regions.
When considering a spillway gate installation, the ground acceleration at the
toe of a dam during an earthquake has to be reassessed at the crest of the dam.
The shock can be considerably amplified over and above that at the dam base,
and the spillway gates will be subject to this amplified acceleration.

251
 Hydraulic gates and Hinks and Gosschalk3 quote examples of the amplification of earthquake
valves acceleration at two rockfill dams in Mexico, which had accelerographs on the
dam crest and on the rock near to the dam. The acceleration on the crest of the La
Villita Dam was between 9 and 22 times greater than that measured on rock at
the right abutment, depending on whether the transverse, longitudinal or
vertical acceleration is considered. At the El Infiernillo Dam, which is 2.5 times
the height of La Villita, the amplification was significantly less at between 2.7
and 4.8 times. The difference between the two measurements is probably due to
the different substrata of the dams. La Villita has deep alluvial deposits in the
valley beneath the dam and so the acceleration at the dam base is likely to have
been considerably greater than that measured at the rock abutment. The total
amplification is probably the result of two combined factors: firstly the
transition from rock to alluvial deposits, and secondly the height of the dam.
The type of strengthening that may be required to cope with these enhanced
acceleration levels is exemplified by three dams in New Zealand. Here the
spillways have vertical-lift gates operated by winches positioned on slabs of
reinforced concrete structural frames above the spillway deck level. The winch
support frames were identified as being at structural risk during an earthquake.2
The problem is similar to that of freestanding outlet towers. The operator of the
dams initiated checking of the effects induced by earthquakes on mechanical and
electromechanical structures such as gates, penstocks, headgate servo-motors,
transformers, high voltage switchgear and control panels. Scottish and
Southern Energy carried out similar investigations at hydro plants in Scotland.8
Seismically induced vibration may not be the only way in which control
structures are damaged during an earthquake. It is possible for seismic activity
to set up seiches in a reservoir. Hinks and Gosschalk3 give the example of the
35 m high Hebgen Embankment Dam in Montana following the earthquake on
17 August 1959.9 The reservoir was full at the time of the earthquake. The waves
created in the reservoir overtopped the dam by 1 m uniformly over the crest.
This overtopping was repeated four times and in each case lasted for about ten
minutes. The effects of such an event could cause serious overloading of
spillway gates and could result in the collapse of the arms of radial gates. Serious
overtopping of spillway gates may also occur when the reservoir is suddenly
affected by a landslide. The triggering of landslides by earthquakes is a common
occurrence in hilly areas,10 and may cause destructive waves in addition to the
increase in reservoir water level.

Spillway gate installations

During an earthquake, vibration of a spillway gate in contact with the reservoir
water will cause a fore and aft motion of a certain volume of water, the added
mass. This is similar to the effect which occurs under conditions of gate
vibration, described in Chapter 10.
Westergaard11 suggested that inertia forces of the added mass can be
approximated by a parabola (Fig. 14.1), which expresses the hydrodynamic
pressure distribution as:

252
 Earthquake effects on
gates

Figure 14.1. Added mass due to

Westergaard11 and total pressure on
a spillway gate due to a seismic event

7 p
Py w Hy
8
where Py hydrodynamic pressure at depth y from the water surface
 = horizontal ground acceleration in units of g
w = unit weight of water
H = depth of water at the location of the structure
y = depth from the water surface

The assumptions of the Westergaard approach are:

The structure can be idealised as a 2D rigid monolith with a vertical upstream
face.
The reservoir extends infinitely in the upstream direction.
The water is incompressible.
Surface waves are not taken into consideration.
Horizontal ground acceleration is considered to act only in the upstream^
downstream direction.
The assumption of a vertical upstream face is not valid for a radial gate, nor is
the spillway gate installation part of a rigid monolithic structure. Moreover for
earth embankments or rock filled dams, the assumption of a vertical upstream

253
 Hydraulic gates and face cannot apply. However, the Westergaard formula is extensively used when
valves seismic evaluation of gate installations is carried out and is considered a
sufficient approximation.
In 1999 the Izmit earthquake in Turkey, close to the Yuvacik Dam, caused
very high wave motion in the reservoir. At the time, the reservoir level was
below the spillway weir and the gates were not subjected to hydrostatic and
added mass loading, so the combined effect of the additional impact of wave
motion was not experienced.
Kolkman12 has developed a relatively simple method for approximating the
hydrodynamic mass for hydraulic gates of various configurations. It is based on
2D flow, without wave radiation, and can be implemented using a spreadsheet
for solving the potential flow problem. It was developed for investigating gate
vibration but its use in computing hydrodynamic mass, due to seismic motion,
is equally valid.

Methods of analysis
Equivalent^static or pseudo^static method
The equivalent^static method represents the seismic forces as an equivalent^
static force equal to the effective mass of the structure multiplied by the peak
acceleration of the input response spectrum, thus allowing for dynamic
amplification of the input acceleration.8
Computation involves the hydrostatic forces on the gate due to the reservoir
water level, plus the load imposed by the added mass of water, multiplied by the
peak acceleration of the input response spectrum . Additional forces are the
amplification exerted at the trunnions, the gate sill and the side guide rollers in
the case of a radial gate.
The vertical acceleration is often assumed to be 0.6 of the horizontal
acceleration. This ratio is likely to increase within the vicinity of the epicentre.
It is considered excessive to assume that peak vertical and peak horizontal
accelerations will be in phase, so a phase separation should be assumed. In radial
gates the amplification of forces at the sill beam and trunnions, due to vertical
acceleration, need not be added to the horizontal effects.
Structural damping for steel of 3% and 5% is recommended in connection
with the OBE and the MDE, respectively.4 Friction damping will occur at the
side seals and the trunnions of a radial gate. For side seals this can be estimated
from the data given in Chapter 7.
An equivalent^static analysis is a 2D model and does not represent important
3D effects in the gate. It is advisable to confine equivalent^static analyses to
preliminary assessments. Dynamic finite element 3D analyses are recommended
for gates with complex structural arrangements, for which 2D modelling would
be inadequate.8
ICOLD (1999)4 suggests that a dynamic analysis should be considered if the
structure has a fundamental frequency of less than 33 Hz. This would apply to
most gated structures. However, the response spectra at 5% damping show
natural frequencies for hard sites between about 5^12 Hz. For soft sites, the
values are between 3^8 Hz.13 Most of the energy in an earthquake ground motion

254
is concentrated in the frequency band up to 10 Hz, so that a structure with a Earthquake effects on
natural frequency much higher than this tends to respond as a rigid body. The gates
range of natural frequencies between 3^12 Hz includes most spillway gates and
suggests that dynamic analysis should be used if a refined analysis is required.
A dynamic study of a gate using finite element analysis involves a
considerable technical input. ICOLD (1999)4 suggests that it results in
construction cost savings for new gates. The output of a refined 3D finite
element model of the radial gate at Kilmorack Dam8 was compared with an
equivalent^static analysis. Overall, the equivalent^static analysis tended to yield
conservative results for response parameters compared with the time history
analysis of the 3D model, although axial forces in the lower gate arms were
slightly underestimated, and significant 3D effects due to curvature of the
stiffened skin plate assembly could not be represented. Given these limitations,
the equivalent^static analysis of simple 2D models appears to be well suited to
the preliminary seismic assessment of gates. Adopting a multiplication factor
for the peak acceleration from the input response spectrum as advocated in the
ASCE recommendation14 may avoid underestimating seismic responses and
render the equivalent^static method suitable for a main analysis. It is likely to
provide a conservative estimate and will indicate whether the structural capacity
of the gate will be exceeded. If this is the case, dynamic analysis of a 3D model
may be more appropriate.

Allowable stresses
US manuals and codes permit a 33% increase in allowable stresses when
combining earthquake loading with the normal dead and live loads. ICOLD
(1999)4 suggests that allowable stress values up to 90% of the material strength
may be appropriate because earthquakes impose an extreme loading. Since
permanent distortion due to yielding or exceeding the proof stress could make
a gate inoperable, the material strength should, in this instance, be the yield
point stress or an appropriate proof stress. ICOLD (1999)4 does not justify the
suggested allowable stress values.

Vertical-lift spillway gate installations

The loading conditions of these gates can be derived in a similar manner to those
of radial gates. The special feature of vertical-lift gate installations is the
overhead gate lifting structure. Fig. 2.19 shows a typical structure of this type
at a river barrage. There are a few examples of reinforced concrete lifting gantry
structures. Williams2 describes the strengthening of these structures at Maraetai,
Whakamaru and Ohakuri Power Station dams in New Zealand. Examination
showed that there was insufficient flexural and lateral reinforcement in the
gantry columns for the structure to remain serviceable following earthquakes
greater than about a 150-year event.
Older gate installations tend to be counterbalanced. Under earthquake
conditions, a considerable amplification of the suspended mass of the counter-
weight ^ which is nearly twice the mass of the gate ^ will occur. The wire ropes

255
 Hydraulic gates and will act as a spring due to their relatively low modulus of elasticity (see Chapter
valves 7). The integrity of the overhead structure is at risk, and by extension that of the
gate. The practice of partially counterbalancing radial gates was used on some
older structures to reduce the hoisting effort. If these are assessed under
earthquake conditions the possible failure of the counterbalance ropes should
be investigated, since fracture may prevent raising of the gates.
In most cases, hoisting gantries and their access ladders will require a
dynamic assessment. This should also include the anchorage of hoist motors,
gear boxes and hoist drive supports, as well as the vibration of long transmission
shafts and their resulting stresses.
Lighting columns on elevated gantries are vulnerable during an earthquake
and may cause consequential damage on failure. Area lighting and emergency
lighting can also cause damage. Lights on gantries should be located at platform
level.
Spillway gate installations include a bridge linking the piers and abutments.
This should also be part of a seismic assessment.
Overhead hoisting machinery is either mounted on the bridge, or sometimes
on a separate structure spanning the piers. The integrity of the structures, and
especially their anchorages, as well as the bolted machinery connections to the
structure, must be considered under earthquake conditions.
Many spillway gate installations include a gantry crane for handling
stoplogs. These are liable to topple over or jump rails during an earthquake.
To prevent this, rail keep plates are fitted. During the 1987 earthquake at the
Matahina Dam on the North Island of New Zealand, many rail keep plates,
fitted to prevent toppling of the transformers, sheared. The dynamic effect of
a swaying gantry crane and any measures designed to prevent toppling over
must be fully considered.

Bottom outlet tunnel gates

High transient hydrodynamic pressures can be generated by earthquakes,
causing shaking of the water in the tunnel. This can develop pressures well in
excess of the hydrostatic pressure. Surge shafts are constructed in turbine
penstocks to protect the system after turbine rejection. They also damp
hydrodynamic pressures due to surges caused by earthquakes. ICOLD (1999)4
states that when a surge shaft is provided `there is little need for analysis of
hydrodynamic pressure due to an earthquake'.
An evaluation of the transient hydrodynamic pressures in a tunnel without a
surge shaft can be carried out by the discrete Fourier transform technique.15 The
analysis indicates that water in the tunnel develops a dynamic pressure which is
proportional to the length of long straight sections and to the input acceleration.16
Some analyses show that even for moderate earthquakes, hydrodynamic pressures
can be several times larger than the hydrostatic pressure, and cavitation may occur.
Because of the short duration of an earthquake, cavitation damage ^ which is time
dependent ^ will not be significant.
In tunnel gates of the vertical-lift type, the gate is withdrawn on opening
into a bonnet. Operation is by a servo-motor which is bolted to the bonnet

256
closure section. The cylinder structure is vulnerable to earthquake forces Earthquake effects on
which will act at the bolted connection of the servo-motor and the bonnet. gates
Depending on the natural period of vibration of the cylinder, dynamic
amplification can occur at the top of the cylinder causing amplified bending
stresses and high tensile stresses at the holding down bolts. One such case
was identified at the diversion tunnel gate of the Ohakuri Power Station17 in
New Zealand. Three symmetrically arranged tension ties were designed
which anchored the brace to secured points at ground level. They were
positioned to support the cylinder at the centre of gravity of load
distribution.

Operating machinery under earthquake

conditions
At new spillway and bottom outlet installations, seismic qualification of
equipment can be made a requirement. It significantly increases the cost of
machinery and ancillary equipment and restricts choice of components. For
areas of low seismic risk and hazard it is probably too onerous. The risk
classification of dams given in ICOLD Bulletin 725 can be used as a guide.
Seismic qualification of machinery and electrical equipment for gates appears
unwarranted for dams in category I and possibly category II.
Seismic qualification is the process of integrating experimental and finite
element analysis to investigate the paraseismic design of safety plant
equipment.18 It is a requirement in the nuclear power industry and a large
amount of test and analysis data has accumulated. By using these data on seismic
characteristics and performance of safety related systems, benefits in cost and
time can be achieved in the design of operating equipment for gates in areas of
high seismic risk.19
Many spillway gate and bottom outlet installations are of considerable age,
some dating back to the 1950s. In these cases, the options for retrofitting or
improvement to enhance their capacity to remain operational following an
earthquake are limited.

Systems analysis
A systems analysis20 to determine the impact of failed components on the ability
of safety critical spillway gates and bottom outlets to withstand an earthquake
without major consequence would include the following, in addition to the
issues already discussed:
The analysis will be dominated by events with the potential to affect all of a
set of multiple gates. Random independent failure of multiple components is
clearly possible but its probability of occurrence is likely to be of a much
lower order than common cause events. Multiple independent failures have
to be included. The potential for equipment to be out of action because of
maintenance is frequently an important contributor to the failure of intended
redundancy.

257
 Hydraulic gates and A different drive system or design of gate will throw up different problems
valves for consideration; rope supported gates, for example, may be particularly
vulnerable to seismic motion in a part open position.
The gate drive mechanism will require a mains electrical supply for normal
operation. There will, however, be a standby system for actuation of the
normal drive, in the form of either an auxiliary power supply or manual
operation, or both. Failure of the drive system itself obviously leads to failure
even if the standby system functions correctly. It must be determined
whether the manual system has the capability to open the gates within a
reasonable timescale before this arrangement can be factored into a reliability
analysis. Similarly the arrangement of many diesel auxiliary power supplies
makes them vulnerable to vibration-induced failure through features such as
high-level fuel tanks and rigid piping to the diesel engine.
In most significant earthquakes it must be expected that the off-site power
supply will fail because of the vulnerability of overhead lines. Because it is
assumed that the turbines will be tripped by the earthquake the normal on-
site electrical power supply will not be available. Many installations include a
battery standby supply. Insufficient structural resistance of the battery rack
to horizontal earthquake motion, or batteries which have not been properly
secured to the rack, can cause failure of the supply.
A variety of components may be damaged if they are insecurely anchored and
are overturned by seismic vibration.
Hydraulic cylinders are vulnerable to low frequency vibration that can
damage the hydraulic seals. Gimbal mounting of cylinders operating radial
gates, shown in Fig. 6.6, reduces the natural frequency of the cylinder and the
amplitude of forced vibration.
Hydraulic pipework to the drive cylinders may also be vulnerable to
vibration, particularly if there are poorly supported long pipe runs.
Seismic vertical acceleration may be accentuated if gates are in a partially open
position; out-of-phase motion could result in large relative accelerations
between the gate and the trunnions mounted on the piers. In these
circumstances, hydraulic cylinders could act as dampers to alleviate the
situation. Alternatively the hydraulic cylinders could be effectively stiff,
resulting in overpressurisation of the hydraulic system. The pressure
transient may be too short for the pressure relief system to react adequately.
(If the gate is closed, the gate sill may prevent overloading of the cylinders or
gate supports.)
Overhead structures are a feature of vertical-lift gates and have already been
discussed, but there may also be a gantry crane, or overhead service gantries.
Consideration should be given to whether failure of the overhead structure
could result in failure of several gates, either through some direct effect or via
a vulnerable component, cable or pipework common to all the gates.
In a typical control installation there are likely to be tens, or perhaps
hundreds, of relays and switches. These are potentially vulnerable to
spurious intermittent opening or closing due to seismic vibration. In many
cases relay chatter will cause no net effect because the relays will return to
their original state when vibration ceases. In other cases, the making or
opening of the relay or switch may set in the new condition and will not

258
 subsequently be reset. This can be true of protection systems where a manual Earthquake effects on
reset is often considered a reasonable precaution. Limit switches may be gates
vulnerable to spurious operation, with overtravel switches effectively
preventing operation of the gates. The full range of spurious effects
generated by relay chatter or switch operation will depend on the exact details
of the control system design, but they are capable of causing operator
confusion or even failure of the controls.
Cable trays which contain cabling for multiple systems are often poorly
supported. Vibration-induced failure is possible and may have wide ranging
effects on apparently independent systems.
Wiring failures can be serious where there is little segregation of cable runs
for different systems and where cable runs cross joints in the concrete
structures. Overhead cable runs can be vulnerable to gantry collapse.
Cubicles housing control relays and other control equipment can topple if
they are not adequately anchored, and such a failure can have widespread
effects on multiple systems.
The transverse movement and possible displacement of sluiceway piers is a
potential consequence of an earthquake. Radial gates, due to their close tolerance
to effect sealing at the piers, can become wedged or suffer local buckling. In these
conditions, the side-seal mounting is so arranged that the seal mounting section
will deflect or buckle under severe impact, as shown in Fig. 14.2. It is more usual to
arrange the seal mounting bracket upstream of the skin plate to prevent debris
accumulation in the pocket formed between the seal mounting plate and the skin
plate. The practical difficulty which arises from the arrangement in Fig. 14.2 is the
bridging of side and sill seals at the junction between the seals. This is done by a
rubber block, which can be a source of leakage.
Displacement of piers at the location of trunnions is a more difficult
problem. Self-aligning trunnion bearings will compensate for misalignment of
the axis of the two trunnions, but the gate arms will permit only limited
deflection.
Increasing the shear resistance at the floor joints of the crest structure can
prevent or reduce the potential for deflection of the sluiceway walls. This can
comprise increasing longitudinal reinforcement bar size, or forming concrete
shear keys at the joints.

Figure 14.2. Side seal mounting of a

radial gate to provide a collapse zone
in the event of transverse movement of
a sluiceway pier due to an earthquake

259
 Hydraulic gates and An alternative design in which the piers and the sluiceway floor are
valves monolithic may prevent jamming of a gate. An arrangement of this kind was
considered for the Torrumbarry Weir on the River Murray in Australia.
Electrical cables and oil hydraulic pipes which cross the movement joint
between two sections of a pier have to permit movement.
Rope sheaves and rope drums require substantial rope guards to prevent
hoist ropes from jumping grooves. In vertical-lift roller or slide gates,
horizontal accelerating forces will be applied through the transverse guide
slippers or guide wheels. Figure 14.3 shows an arrangement of spring-loaded
transverse guide wheels to absorb shock. This arrangement is also used on
stoplogs and draft tube gates when they are placed under balanced pressure, in
order to locate them close to the sealing faces for good initial sealing.

Figure 14.3. Arrangement of

spring-loaded transverse guide
wheel of a tunnel gate to absorb
an earthquake shock

260
Control buildings Earthquake effects on
gates
Control buildings must be designed to withstand earthquakes. Suspended light
fittings should not be used and emergency light fittings should be self-
contained. Oil hydraulic pipework passing through walls should be rubber
sleeved. Electrical trunking and any switchgear which is wall mounted should
be rigidly secured. Holding down bolts of electrical control cabinets and oil
hydraulic power packs must be substantial and, where they pass through sheet
metal, the area must be reinforced and the load distributed. Fire-fighting
equipment should be available.
Transformer oil tanks are vulnerable to seismic shocks, as are the day tanks of
standby generating plants. Rail mounted transformers have toppled in
earthquakes even when substantial rail keep plates were fitted.

Sample event tree for a seismic event on a dam

The value of this type of analysis is to ensure that potentially important
vulnerabilities are addressed at the design stage of a project, and that a clear case
is made for resolving each of the concerns. Quantification permits issues to be
prioritised and decisions can be made on risk levels (see Fig. 14.4).
The following notes apply to specific headings on the event tree.

Damage to dam The extent of damage will depend on the magnitude of the
seismic event. Since the demand on the control systems will depend on the level
of damage, or perhaps the extent to which this is obvious immediately after the
earthquake, a family of event trees for different categories of dam damage may
be relevant for a given earthquake. The analysis will need to reflect the operating
procedures determined by the dam owners.

Failure of spillway gates Depending on the extent of damage to the dam, there
will be a requirement for water drawdown over a specific period of time. For the
present example it is assumed that this requires operation of all the spillway
capacity, initiated within one hour after the earthquake, followed by continuing
operation of the bottom outlet. The full specification for the top event of the
relevant fault tree will therefore be `Failure of the system as designed to initiate

Figure 14.4. Event tree for

seismic event on dam (after
Ballard and Lewin20)

261
 Hydraulic gates and full spillway flow within one hour of earthquake'. In practice this may always be
valves the requirement for the more severe categories of seismic event because of the
difficulty of determining the real extent of dam damage. However, the potential
for downstream damage due to operating the spillways to their full extent will
be an important factor in deciding what action to recommend.

Operator failure to recover While the spillway gates may fail to respond in the
manner intended because of control or other failures, operators may be able to
recover the situation in time by various planned or ad hoc actions. The extent to
which this is possible will depend on the time available, but also on other factors
such as whether the dam is normally manned, whether the operators are
practised in fault finding and recovery and whether advice is available on a
communication link which is still operating.

Unintended operation of bottom outlet Operator action or equipment malfunction

may lead to either the spillway gate or the bottom outlet opening when it is not
required. The potential for downstream damage as a result of such opening
means these events must be considered in the analysis. For the present example
it will be assumed that operators will follow clear operation procedures and will
not initiate opening of either spillway gates or bottom outlets unnecessarily. In
the case of equipment malfunction it will be similarly assumed that the operators
would quickly recognise any unintended operation of the spillway gates because
these are immediately visible to them. However, it is less clear that they would
see and recognise control indication of bottom outlet initiation or that the
outfall of the bottom outlet would be visible. This event is therefore retained
within the event tree.

Failure of bottom outlet Depending on the extent of damage to the dam, it may
be necessary to open the bottom outlet to provide continued lowering of the
reservoir water level. Failure to open could lead to dam failure despite successful
operation of the spillway gates, although the lowering of water levels resulting
from that operation could mitigate the consequences of any subsequent dam
failure.

Event sequence consequences The analysis is driven and limited by consideration of

events that concern the dam owners. This example is limited to those events
with the potential to cause fatalities. In practice, the owner may be interested
in a wider range of consequences such as damage to generating capacity etc.
Balanced against this interest is the greater complexity that would be required
in the event trees and the fault trees.

References
1. Charles, J A; Abbiss, C P; Gosschalk, E M; Hinks, J L (1991): An engineering guide to
seismic risk to dams in the United Kingdom, Report, Building Research Establishment.
2. Williams, I S (1996): After the earthquake ^ ensuring that the spillway can be
operated when it really counts, 7th Hydro Power Engineering Exchange, Hamilton,
New Zealand, Oct.

262
 3. Hinks, J L and Gosschalk, E M (1993): Dams and earthquakes ^ a review, Dam Earthquake effects on
Engineering, IV, No. 1, Feb. gates
4. ICOLD (1999): Guidelines for earthquake design and evaluation of structures appurtenant to
dams, Committee on Seismic Aspects of Dam Design, draft CIRC 1540, Apr.
5. ICOLD (1989): Selecting seismic parameters for large dams, Bulletin 72, Paris.
6. Irving, J (1985): Earthquake hazard in Britain, Proc. of Conference on earthquakes
engineering in Britain, University of East Anglia, Apr., Thomas Telford, London,
pp. 261^277.
7. Alderson, M A H G (1982): Seismic design criteria and their applicability to major hazard
plant within the United Kingdom, report SRD R246, Warrington, United Kingdom
Atomic Energy Authority, Safety and Reliability Directorate.
8. Daniell, W; Taylor, C (1999): Seismic study of Kilmorack Dam radial gate, report for
Scottish and Southern Energy, University of Bristol, Earthquake Engineering
Research Centre.
9. Sherard, J L; Woodward, R J; Gizienski, S F; Clevenger, W A (1963): Earth and
earth-rock dams, John Wiley & Sons, New York.
10. Skipp, B O (1980): Earthquake vulnerability of superficial materials ^ landsliding of natural
and manmade slopes, Lecture Notes, Course on Assessment of Earthquakes Risk,
Geological Society of London.
11. Westergaard (1931): Water pressure on dams during earthquakes, Am. Soc. Civ.
Engrs., Transactions, paper 1835, Nov., pp. 418^472.
12. Kolkman, P A (1988): A simple scheme for calculating the added mass of hydraulic
gates, Jour.Fluids and Struct., Vol. 2, pp. 339^353.
13. Principia Mechanica Ltd (1982): Ground motion specification, Report ET 17 for British
Nuclear Fuels Ltd.
14. ASCE Standard 4^86 (1986): Seismic analysis of safety related nuclear structures and
commentary on standard for seismic analysis of safety related nuclear structures, Sept.
15. Kojic, S B; Trifunac, M D (1988): Transient pressures in hydrotechnical tunnels
during earthquakes, Earthquake engng. and struct. dyn, Vol. 16, pp. 523^539.
16. Zienkiewicz, O C (1963): Hydrodynamic pressures due to earthquakes, Water
Power, 15 Sept.
17. Power Engineering Works Consultancy Services (1994): Diversion gate lifting
cylinder, Ohakuri Power Station, ECNZ Waikato Hydro Group.
18. Aguiard, P; Boutillon, J P; Bonnecase, D; Dorey, P; Maurin, N; Panet, M (1992):
Model testing and calculation integration applied to the component seismic
qualification, Seminar Seismic and Environmental Qualification of Equipment, I Mech E,
London, Oct., pp. 55^59.
19. Ziadlourad, F; Bye, A D (1992): The use of a UK seismic and environmental
database approach to reduce qualification costs and timescales, Seminar Seismic and
Environmental Qualification of Equipment, I Mech E, London, Oct., pp. 67^73.
20. Ballard, G M; Lewin, J (1998): Should reservoir control systems and structures be
designed to withstand the dynamic effects of earthquakes? Proc. 10th British Dam
Society Conference, Bangor, Wales, Sept. in The prospect for reservoirs in the 21st
century, editor Tedd, P, Thomas Telford, pp. 52^65.

263
 15
Materials and protection

Materials
Most gates are fabricated in low carbon structural steels. Except in high head
gates, the higher tensile stress grades are rarely employed because deflection
often becomes the critical design parameter so that sealing faces do not open
under load. In gates which are subject to ice formation, low temperature
structural steels are used to prevent brittle fracture.
In fluidways and valves where high velocity flow is experienced, and where
there is a risk of cavitation, stainless steel is selected. The same material is used
for linings downstream of gates in high head tunnels where the boundary layers
have not developed sufficiently to protect the walls from high velocity flows.
Seal contact faces and guide roller paths are also constructed in stainless steel.
The degree of corrosion resistance of stainless steels is determined by the total
amount of nickel and chromium components. Thus high nickel and chromium
austenitic steels are generally the best choice.
Operational experience of the use of austenitic stainless steels for shafts, pins
and other parts suggests that the chromium nickel molybdenum steels
(European Norm 316 group) give better performance than the unstabilised
austenitic chromium nickel steels (European Norm 304 group).
Stainless steels in contact with one another are subject to galling, the chafing
process which causes seizure of two parts moving relative to one another. Two
stainless steel components which are physically assembled in contact with one
another, and which may have to be disassembled or replaced at some stage,
require a sleeve of a different material such as phosphor or aluminium bronze
to be interposed. Figure 7.17 shows a stainless steel self-aligning trunnion
bearing mounted on a stainless steel shaft with an interposed bronze sleeve.
Operationally there is no movement between the parts, but the absence of the
sleeve might make it impossible to disassemble them. The same figure shows
passages in the shaft to enable the use of high pressure oil hydraulic fluid to ease
separation when replacement is required.
Galling can also occur at stainless steel bolts and nuts. At a number of gate
installations where austenitic stainless steel nuts and bolts have been used,
selecting bolts of the unstabilised steel grade and nuts of chromium nickel
molybdenum steel has averted the problem.

265
 Hydraulic gates and Stainless steels are often welded to low carbon structural steels. To prevent
valves carbon migration at the welds, stabilised austenitic chromium nickel steels must
be used. The steel is stabilised by the addition of titanium. Under some
conditions, unstabilised stainless steels can be welded to low carbon steels using
special welding rods with molybdenum and/or titanium additions.
Stainless steel clad carbon steels are used to achieve cost reductions. The layer
of stainless steel is rolled on top of the carbon steel to produce an integral plate.
It is desirable that the cladding be at least 1 mm thick. Welding of a clad steel to a
carbon steel is possible, but requires a special technique. The protection of cut
faces and bolt holes in clad steels presents problems. The cost advantage of clad
steel does not always justify its use.
A limited number of cases of electrolytic underwater corrosion have been
reported due to the proximity of stainless and carbon steels but, in general,
operating experience in inland waters has been favourable.
Large cathode (stainless steel) to anode (carbon steel) area ratios should be
avoided. Surfaces of both metals should be painted. If only the anode metal is
painted and there is a small defect in the coating, the cathode to anode area ratio
will be very large and rapid corrosion will occur.
In saline waters, nickel^copper alloy (Monel metal) is used, and also in fresh
water where parts are in contact with brass or bronze. Nickel^copper alloy is
available in plate and sheet form, rod and bar, and is more expensive than
stainless steel. It is closer to the brass and bronze alloys in the galvanic series
and its corrosion resistance, especially in saline waters, is superior to that of
stainless steels.
Guide rollers, shafts and pins which are permanently or occasionally
submerged in water, are selected in high chromium ferritic stainless steel, in free
machining austenitic chromium^nickel steel or in nickel^copper alloy. Their
use in connection with bearing materials of leaded bronze is common.
Martensitic stainless steels can be suitable for some applications but should be
used with caution for heavily loaded rotating parts, as some have a relatively low
fatigue limit.
Bolts and nuts, particularly where they may have to be removed for
maintenance or replacement, should be in stainless steel or Monel. Bronze or
brass bolting is not suitable.
The usual range of engineering materials from ductile iron castings to alloy
steel castings, from medium carbon to high tensile alloy steels, is used. Their
selection and application are comparable to that found in general engineering
practice, except that factors of safety are generally higher. Grey iron castings are
not, as a rule, used for stress carrying components in gate installations because
they are liable to brittle fracture.

Steel corrosion and painting

The corrosion protection of hydraulic structures, particularly gates, is of the
utmost importance to the operating authority. To carry out maintenance
painting of a gate the sluiceway has to be stoplogged and pumped dry, and in
many cases a protective shelter has to be provided over the area to be painted.

266
 Certain areas of gates are more susceptible to corrosion than others. It is most Materials and
likely to occur at crevices, conjunctions of dissimilar metals, areas where debris protection
or mud can accumulate, and particularly where water can pond up. Other areas
prone to corrosion, due to the difficulty of applying an adequate protective
coating, are sharp corners, locations of bolts, drain holes and areas which are
difficult to access. Erosion occurs at radial gates where hoisting ropes are
located upstream of the skin plate assembly. To avoid damage to the protective
coating of the skin plate, a stainless steel plate is welded along the contact face.
The water line upstream of radial and vertical-lift gates in free surface flow is
another area vulnerable to corrosion.
The use of closed sections as main structural members, and also as stiffener
members of skin plate assemblies, offers advantages such as reduced paint areas,
reduction of pockets or surfaces where water or silt can accumulate, structural
efficiency and appearance. Figure 5.7(b) illustrates the problem of ensuring that
the underside of the flanges of the T-section stiffener beams of a skin plate are
adequately painted in the first instance, and that the areas can be prepared for
repainting.
To prevent moisture penetration and condensation before the sections are
sealed by welding, box section members or other enclosed sections are filled
with dry inert gas or a vapour phase inhibitor. Bolt holes in enclosed sections
should be avoided or, where essential, a structural sleeve should be provided to
maintain the moisture protection of the enclosed areas. Welds must be
continuous; 100% testing to avoid discontinuities and pinholes is advisable.
The selection of the most appropriate paint system is difficult. There are a
number of books1,2 and papers3,4 which give guidance, as well as British
Standard 5493.5 The sequence of selection in the BS is given by a table, but to
an engineer who is not a specialist in the subject, it does not assist in choosing
from many combinations of acceptable systems.
A frequent practice is to blast clean to Swedish Standard Sa21/2 and to apply a
wash coat and repeat applications of epoxy coal tar to 350 m minimum dry film
thickness. When applying the epoxy coal tar coating under normal ambient
temperatures and normal curing, a polyamide cured epoxy is used. Under low
ambient temperatures, when quick curing is necessary, an isocyanate cured
epoxy coal tar is applied.
Epoxy coal tar paints are available in a restricted range of colours. This can
present a difficulty when painting spillway gates in the tropics. It is desirable to
reduce the solar heat gain on the downstream side of gates by using a white or
aluminium colour coating. The differential expansion of a spillway gate due to
the heating of one face by the sun, and a lower temperature due to reservoir
water on the upstream side, can result in leakage at the sill seal.
Until some years ago, metal sprayed and thick paint coatings were widely
used on bridges. There is no literature on the effectiveness and durability of this
protective treatment for gates.
The large radial automatic gate at the Pulteney sluices in the City of Bath was
zinc sprayed to 250 m without subsequent painting. The protection was
effective for approximately 25 years. The difficulty of ensuring adequate and
consistent thickness of a sprayed coat at corners and sharp re-entrant angles
militates against repetition of this treatment.

267
 Hydraulic gates and Most gates are assembled with some bolted connections. The mating surfaces
valves must be protected from corrosion, even if high strength friction grip bolts are
used. The bolts have to be derated due to contact with a painted face. Joints which
have been coated and are assembled with friction grip bolts have also to be derated.
The CIRIA6,7 and Transport and Road Research Laboratory8 reports offer
guidance. Some gate designers favour seal welding of bolted connections.
High head gates often incorporate enclosed sections. The delta configuration
of the lip of a slide or roller gate is an example. It is possible for oxygen and
moisture to penetrate such enclosed sections. Moisture penetrates by differences
in vapour pressure, and as a result of changes in temperature it can condense and
cause local corrosion. Bolt holes in enclosed sections should be avoided, and
welds, which must be continuous, should be tested to avoid discontinuities
and pinholes.
Paint selection, preparation for painting, the testing of painted surfaces and
repair of damaged coatings constitute a specialist subject. The preceding notes
mention some factors pertaining to hydraulic structures and provide references
to literature for further reading.

Cathodic protection
Cathodic protection is the technique of reducing the corrosion rate of an
immersed metallic structure by making the steady-state or corrosion electrical
potential of the metal more electronegative. The thermodynamic considerations
have been dealt with by Shrier,9 Pearson10 and others.
When two dissimilar metals are electrically connected and immersed in an
electrolyte, which can be fresh or salt water, a current flows through the
electrolyte and the metal so that anions enter the solution from the anode, and
at the same time electrons move from the anode to the cathode via the metallic
connection. The level of corrosion protection depends on the amount of current
flowing, which in turn depends on the electromagnetic force (e.m.f.) and
various ohmic and non-ohmic resistances in the circuit.
The e.m.f. may be provided by a metal which is more electronegative than the
metal to be protected (sacrificial protection) or by an external e.m.f. and an
auxiliary anode (impressed current protection).
Cathodic protection of gates in sea water or estuarial locations is mainly by
the use of sacrificial anodes of zinc,11 magnesium or aluminium. Impressed
current cathodic protection is used when corrosion conditions are severe and
where inspection and remedial work during the lifetime of the structure are
impossible or impractical.
Cathodic protection is not effective in the splash zone of a gate, and in a tidal
application will afford only reduced protection in the upper tidal zones.
The current density required for steel for adequate cathodic protection in
moving fresh water is 55^65 mA/m2. In stilling basins where the water is highly
turbulent and can contain dissolved oxygen, the range is 55^165 mA/m2. In sea
water the current density for cathodic protection is within the range
55^300 mA/m2, whereas in highly polluted estuarine water 600^2000 mA/m2
may be required.

268
 Magnesium is probably the most widely used sacrificial anode material, as the Materials and
high current yield ensures maximum current distribution. The addition of protection
aluminium to magnesium reduces self-corrosion, but minor alloying elements
such as copper, nickel and iron can significantly increase this tendency and
counteract the efficiency of magnesium as a sacrificial anode. Alloying elements
are therefore controlled within limits in magnesium anodes.
Current output is related to the composition of the anodes, surface area and
shape, while the working life is dependent on the ratio of surface area to weight
together with the current demand of water at the gate location.
Although the principles of cathodic protection are essentially simple, its
practical application to the protection of steel structures, such as gates,
immersed in water appears to be more of an art than a science.
Cathodic protection applied to a structure, particularly when applied only to
elements of a structure, can present a danger to adjacent unprotected structures
or parts.
A further application for cathodic protection is the prevention of cavitation
damage. It requires high current densities in order that the hydrogen freely
evolved from the protected metal can act as a gas cushion between the collapsing
vapour cavities and the metal surfaces. This makes it impractical to protect more
than a limited surface area. Where cavitation cannot be avoided, such as in the
area downstream of a high head tunnel gate, it is more economical to provide a
replacement liner.
BS 7361: Part 1 Cathodic protection is a detailed guide for the design of
sacrificial and impressed current protective systems.
Paint coatings used on gates subject to cathodic protection must be
compatible. This should be checked with the paint manufacturer.

References
1. Hudson, J C (1940): The corrosion of iron and steel, Chapman and Hall.
2. Evans, L I R (1960): The corrosion and oxidation of metals, Edward Arnold Ltd, Chapter
13.
3. CIRIA (1982): Painting steelwork, editor Haigh, I P.
4. HMSO (1971): Report of the Committee on Corrosion and Protection.
5. British Standard 5493:1977, Code of practice for protective coating of iron and steel
structures against corrosion. (Note: this standard has been proposed for obsolescence
and has been partially replaced by BS EN ISO 12944 Parts 1 to 8.)
6. CIRIA (1969): Protection of steel faying surfaces, editor Day, K J, interim research
report.
7. CIRIA (1980): Design guidance notes for friction grip bolted connections, editor Cheal, B D,
technical note 98.
8. Black, W; Moss, D S (1968): High strength friction grip bolts ^ slip factors and protected
faying surfaces, Transport and Road Research Laboratory, report LR 153.
9. Shrier, L L (1963): Corrosion, section 11, Cathodic protection, George Newness Ltd.
10. Pearson, J M (1955): Fundamentals of cathodic protection, in Section Vll,
corrosion protection, The corrosion handbook, editor Uhlig, H H, John Wiley and
Sons Inc./Chapman and Hall Ltd.
11. Day, K J (1977): Protective treatment, in Proc. I.C.E. Conference Thames Barrier
design, 5 Oct., paper 16.

269
 16
Model studies

Numerous model studies of gates and gate installations have been carried out.
Many relate to a particular project and were undertaken to obtain specific
numerical results. These have limited validity and only show certain
relationships of observed hydraulic parameters within the range of the
experiments undertaken. Some studies have explored problems which are not
specific to an individual installation; extrapolation from data gathered during
empirical investigations has led to the formulation of general guidelines for
other similar cases of flow. They can therefore supplement experience and
enhance understanding of fluid dynamic behaviour.

Froude scale models

The vast majority of models are to Froude scale. This represents the condition of
dynamic similarity for flow in a model and prototype exclusively governed by
gravity. It cannot be used to determine other forces such as frictional resistance
of a viscous liquid, capillary forces, the forces of volumetric elasticity and
cavitation phenomena.
To obtain similarity between model and prototype for flow conditions where
inertia and gravitational forces are dominant, the Froude number Fn of the
model and the prototype must be the same.
p
Fn V= gd

where V flow velocity

g gravitational constant
d depth/length

Thus at a model scale of 1 in S:

Flow Qm Qp =S2:5

Velocity Vm Vp =S0:5

271
 Hydraulic gates and Time Tm Tp =S0:5
valves
where the suffixes m denotes model and p denotes prototype conditions

For viscous forces it can be shown by dimensional analysis that the Reynolds
number Re in the model and prototype should be the same. Both the Reynolds
and the Froude numbers for a model and prototype cannot be made equal. Any
difference in the Reynolds number is not of great import as long as both model
and prototype have high values (above 100 000) and similar roughness to
diameter ratios. Under these conditions the head loss is a common function of
the square of the velocity in both model and prototype. If the reduced Reynolds
number of the model approaches the point of transition of turbulent to laminar
flow, the laminar flow could occur in the model but turbulent flow would occur
in the prototype. This must be avoided, consequently a minimum operable
Reynolds number has to be chosen. Increasing the velocity to improve
Reynolds number correlation is a technique often used to test for safety margins
that have been eroded by these non-scale effects. A full discussion of the theory
of similarity has been given by Novak and Ca belka.1

Two-phase flow problems

A number of hydrodynamic problems encountered in overflow and tunnel gates
involve two-phase flow. Air may be entrained at intakes to conduits or drop shafts
and, due to reduced pressure caused by high velocity flow or flow transition
phenomena, is liberated in the tunnel. In steady flow in open channels, air
entrainment depends on flow velocity and generally occurs at velocities of about
6 m/s and higher. A frequent cause of air entrainment in hydraulic structures is
steady flow transition, such as a hydraulic jump in closed conduits, the transition
phenomenon of a jet into pressure flow. The same effect occurs in overflow gates
due to the suction effect on the inside of the nappe. Measurement of two-phase
flow is possible and is discussed by Novak and Ca belka (Section 4.71), but the
quantitative results are only valid under prototype conditions.

Two- and three-dimensional models to Froude

scale
Hydraulic models of gates and gate installations can be divided into three
categories, although the division is arbitrary and categories overlap to some
extent.
A 2D model which can be constructed in a flume is intended for the study of
gate characteristics, and to verify the design of an associated stilling basin or a
weir crest. The objectives of such a model may range from determining the
discharge characteristics of a radial gate, or the hydraulic downpull forces acting
on a vertical-lift gate in a tunnel, to studying the interaction between an
operating gate and a guard gate in a conduit.
An example, Fig. 16.1, is the model study of the spillway gates of the
Kotmale Dam in Sri Lanka.2 The objectives were to determine the gate

272
 Model studies

Figure 16.1. Model study of

the spillway gates of the
Kotmale Dam, Sri Lanka

characteristics throughout the range of gate openings, but particularly at small

gate apertures. These were required in order to program the gate discharge at
the onset of a flood to increase downstream river flow gradually, safeguarding
river users. Another requirement was gate operation to attenuate return period
floods of up to 100 years. The model was also used to measure subatmospheric
pressures on the upper part of the spillway, to determine pressure variations
across the weir crest and below.
The second category of model is the 3D approach flow, which includes a
significant section of the downstream geometry. An example of this type is the
model of the River Medway flood relief scheme,3 where the pattern of flow from
the storage reservoir into the sluiceways was part of the investigation. The
model of the tunnel gates in the bottom outlet of the Mrica Hydropower
Scheme4 is another example. The conduits from the two gates split the approach
channel and reunite downstream of the gate installation. The hydraulics of
splitting the flow, operating one gate only, or two gates with different
discharge, and uniting the flow downstream of the gates were the major reasons
for the model study.

Models for investigating vibration problems

The third category of model study investigates existing or potential
problems of gate vibration or cavitation. Some problems of flow-induced
structural vibration can be investigated by reproducing in the model a single
degree or multiple degrees of freedom.5,6 However, most studies require
that the model be constructed with an overall reproduction of elasticity.
This provides a check on the design but necessitates the use of special
plastics. The design and construction of the model is expensive and time-
consuming.

273
 Hydraulic gates and The scaling criteria which must be satisfied in constructing a hydroelastic
valves model are given in Kolkman (Appendix C)7, and Haszpra.8
Most of the models of hydroelastic similarity for studying potential gate
vibration problems were investigated at Delft Hydraulics, such as the Hagestein
visor gates,9 the radial gates of the Haringvliet sluices and the vertical-lift gates
of the storm surge barrier across the Eastern Scheldt, shown in Fig. 16.2. In the

Figure 16.2. Model of

hydroelastic similarity of the
gates for the storm surge barrier
of the Eastern Scheldt

274
UK the rising sector gates of the Thames Barrier were the subject of Model studies
hydrodynamic load and vibration studies.10,11 Studies to compare model and
prototype results.12,13 have been undertaken. These, together with the success
of designs which have been tested using models of hydroelastic similitude,
justify confidence in vibration models.
Some actual or potential vibration problems can be investigated by
constructing a model only to Froude scale. This requires observation of the flow
conditions, coupled with the experience to judge whether these are likely to
cause gate vibration. A similar model strategy can be employed to study possible
conditions of flow separation and reattachment at bottom sections of gates.
Cavitation can be an important cause of dynamic load as well as causing
significant loss of material. Gates have to be designed to be cavitation free, or
to be subjected only temporarily to a low degree of cavitation. To satisfy the
theoretical requirement, model research must be such that model and prototype
vapour pressures occur at equivalent locations. The relationship between the
pressure at any other location in the flow and the pressure in the critical location
is given by pV2, where p is the density of water and V is the reference velocity.
Therefore the criterion is that the Thoma number:

 pvapour
V 2

is correctly reproduced, where  water pressure and pvapour vapour

pressure of air.
This cannot easily be done and requires a special test facility,14 such as the one
developed by the Snowy Mountains Engineering Corporation in its Fluid
Mechanics Laboratory.14 Small-scale models have been used successfully at
Imperial College, London, to indicate likely patterns of cavitation and even of
cavitation erosion for the elements of a large structure.15, 16

References
1. Novak, P; Ca belka, J (1981): Models in hydraulic engineering ^ physical principles and
design applications, Pitman Advanced Publishing Program.
2. Milan, D; Habraken, P (1984): Kotmale, report on spillway radial gates, model tests,
General Technical Services, Lyon (unpublished).
3. Palmer, M H (1979): Hydraulic model study of the River Medway flood relief scheme control
structure, B.H.R.A., report RR 1572.
4. Bruce, B A; Crow, D A (1984): Mrica Hydroelectric Project: hydraulic model study of the
drawdown culvert control structure, B.H.R.A., report RR 2325, Nov.
5. Abelev, A S (1959): Investigations of the total pulsating hydrodynamic load acting
on bottom outlet sliding gates and its scale modelling, 8th I.A.H.R. Congress,
Montreal, paper A10.
6. Abelev, A S (1963): Pulsations of hydrodynamic loads acting on bottom gates of
hydraulic structures and their calculation methods, 10th I.A.H.R. Congress,
London.
7. Kolkman, P A (1976): Flow-induced gate vibrations, Delft Hydraulics Laboratory,
publication 164.

275
 Hydraulic gates and 8. Haszpra, O (1979): Modelling hydroelastic vibrations, Pitman Publishing.
valves 9. Kolkman, P A (1959): Vibration tests in a model of a weir with elastic similarity on
Froude scale, 8th I.A.H.R. Congress, Montreal, paper A29.
10. Crow, D A; King, R; Prosser, H J (1977): Hydraulic model studies of the rising
sector gate; hydrodynamic loads and vibration studies, Int. Conference on Thames
Barrier Design, London, Oct.
11. Hardwick, J D (1977): Hydraulic model studies of the rising sector gate conducted
at Imperial College, Int. Conference on Thames Barrier Design, London, Oct., I.C.E.,
London, 1978.
12. Geleedst, M; Kolkman, P A (1963): Comparison of measurements on the prototype
and the elastically similar model of the Hagestein Weir, 10th I.A.H.R. Congress,
London, paper 3.21.
13. Geleedst, M; Kolkman, P A (1965): Comparative vibration measurements on the
prototype and the elastically similar model of the Hagestein Weir under flow
conditions, 11th I.A.H.R. Congress, Leningrad, paper 4.7.
14 Lesleighter, E J; Harrison, R D (1981): Development of a cavitation test facility,
Conference, Institution of Engineers, Australia, Canberra, Mar.
15. Kenn, M J; Garrod, A D (1981): Cavitation damage and the Tarbela Tunnel collapse of
1974, Proc. Instn Civ. Engrs, Part 1, 70, Feb: Discussion Proc. Proc. Instn Civ. Engrs,
Part 1, 1982, 70, Nov.
16. Kenn, M J (1983): Cavitation and cavitation damage in concrete structures, Proc.
6th Int. Conference on erosion by liquid and solid impact, Cambridge, Sept.

276
 17
Environmental
considerations

Environmental considerations for spillway gate installations differ from those

which apply at gated river control structures. The former are part of a
prominent structure, a dam and the reservoir formed by the dam. The
environmental impact involves the whole system, including changes to the river
feeding the reservoir and the watercourse receiving the discharge from the
spillway. These wider aspects are outside the scope of this book.
Some hydraulic effects due to river and estuary control structures also have
an environmental effect and are briefly outlined in this chapter.
In the UK, rivers were controlled initially for commercial navigation but are
now more frequently treated as an amenity. The position is different in Central
and Eastern Europe, where river transport plays an important role.

Gated river control structures

The main environmental considerations at gated river control structures are:

Navigation
This requires the inclusion of one or more locks. When locks are located
alongside the control structure, extended piers are required upstream and
downstream of the lock to prevent cross-flow and to provide quiescent water
conditions when river craft enter and leave a lock. At some structures, mooring
basins are provided for boats and vessels waiting to enter a lock. These
structures have an environmental impact which has to be balanced against
navigation requirements.

Flood management and maintenance of river levels

River levels are often lowered in anticipation of a flood to permit flood routing.
This has a limited, temporary effect.

Fish passes
At many rivers, fish passes for migratory fish are an important part of the
control structure. Criteria for the the design of fish passes is a specialist subject
and have been given by Beach.1 At hydroelectric stations in rivers in the Scottish

277
 Hydraulic gates and Highlands, fish ladders and fish lifts are used. Since the flow in fish passes is
valves continuous, even if all gates at a weir are closed because of low river flow, fish
passes are frequently located at or near the centre of a river barrage or weir.
Migrating fish are attracted to falling water or the turbulent conditions in a
stilling basin due to the discharge of gates. As a result, priority of operation
during low river flow is often given to those gates adjoining a fish pass.

Avoidance of erosion of the river bed

Avoidance of river bed erosion due to discharge under or over gates requires
dissipation of the energy of the discharge by a stilling basin, and sometimes
armouring of the river bed downstream of the stilling basin.

Debris passage and removal

Debris accumulation at gates can be very unsightly, as for example at the Holme
Sluices on the River Trent (Fig. 17.1).
An undershoot gate, that is a conventional radial gate or a vertical-lift gate,
will not discharge floating debris until it is at least 30^40% open. Such openings

Figure 17.1. Debris upstream

of vertical-lift gates at Holme
Sluices, Colwick, Nottingham

278
are associated with high river flows or flood events. Thus debris accumulation Environmental
upstream of gates will occur during most of the year when river flow is low. To considerations
reduce or eliminate this discharge, gate overflow is required. The bottom-
hinged flap gate (see Figs 2.25^2.27), is effective in this respect.
Radial gates with overflow section (Fig. 2.1(b)) and vertical-lift gates with
overflow section (Fig. 2.20), will discharge upstream floating debris at low
river flows. Submergible radial gates (Fig. 2.1(c)), operate in the same way.
Accumulation of floating debris and logs of wood on open structural
members stiffening the skin plate of a gate can be avoided by adopting the
construction used for the radial gates of the Torrumbarry Weir on the River
Murray in Australia mentioned in Chapter 5.
Not all debris will clear in a stilling basin. It becomes trapped by recirculation
due to sheared flow, and by the induced hydraulic jump caused by kicker plates
or energy dissipating blocks at the end of the stilling basin. The bypass system
illustrated in Fig. 2.27 can assist in clearing floating debris from a stilling basin
when gates are shut. When the span of a gate exceeds 3^4 m it will leave dead
areas where debris will remain. Floating oil or beer cans in a stilling basin can
generate noise by repeated impact on the downstream side of gates.
Since most rivers are controlled by a succession of weirs, clearance of debris
at a gate structure results in its accumulation at the next downstream weir. The
removal of debris from rivers aided by debris booms has met with only limited
success.

Adapting structures to the environment

Certain design features can assist in blending a river control structure into the
surrounding environment, making it more aesthetically pleasing:
least prominence
the minimum number of gates and therefore piers which ensure a safety
margin if one or several gates are out of operation due to maintenance or a
defect
simple, unobtrusive operating machinery
generally, the use of closed sections (compared with angles or beams) will
result in an economical structure which is pleasing, and easy to maintain
avoidance of unnecessary bracing members
careful detailing of a bridge, overhead structure if required, and hand
railings.
Least prominence of gates and their structures is achieved by selecting
bottom-hinged flap gates. The operating cylinder of bottom-hinged flap gates
can, in many cases, be arranged in a near horizontal position, in order not to
break the skyline. Radial gates are more prominent, particularly when they are
elevated during high river flows or a flood. Their appearance is often spoilt by
clumsy overhead hoist machinery. Operation of radial gates by oil hydraulic
cylinders can be designed so that the hoist machinery has little visual impact.
Vertical-lift gates require a high overhead structure. Even the double leaf
vertical-lift gate has to be suspended from a high structure in comparison with
bottom-hinged flap gates or radial gates. Masonry structures for vertical-lift

279
 Hydraulic gates and
valves

Figure 17.2. Vertical-lift

gate with overflow flap.
Retention 2.55 m above sill
level, span 6.85 m

Figure 17.3. Radial gate with

overflow flap. Retention
2.55 m above sill level, span
6.85 m

gates built early in the 20th century are often considered visually pleasing and
well integrated with the environment, perhaps because masonry does not have
the same industrial associations as steel.
Figures 17.2^17.4 show three different types of gate of the same aspect area
which have been designed to provide overflow to clear debris and to minimise
the visual impact of the sluice structure.
Vandalism is an important consideration in designing river control
structures and protecting existing installations. This requires high fences.

Figure 17.4. Bottom-hinged

flap gate. Retention 2.55 m
above sill level, span 6.85 m

280
Sensitive solutions which effectively impede access to river barrages but do not Environmental
clash with environmental considerations are difficult to achieve. considerations

Barriers
Closing and opening of barriers
Closing of barriers can cause surge waves upriver, because on closing a barrier
the natural flow of the river is stopped. If tidal flow upriver is arrested, a surge
wave downriver can be initiated. Similarly, on opening a barrier against a slight
differential head a surge wave is started. The time taken to close or open a barrier
is another critical element. Operational management of a barrier must take into
account the effect on shipping, pleasure craft and, in some cases, people who are
fishing. The method of barrier closure may have to be investigated for different
surges to determine the reflected wave when the barrier is closed.

Alteration to river characteristics due to a barrier

This is often studied in a physical model and sometimes by a mathematical
model. Investigations have been carried out into the behaviour of rivers before,
during and after barrier construction. Other examinations may include the
effect on water level, salinity, currents and sedimentation.

Bed protection
Bed protection adjoining the barrier may have to be investigated to take into
account the possibility of failure of a gate to close, causing a scouring action.

Navigation and shipping

Navigation and shipping considerations are an important element during the
planning stage of a barrier.2 The location of a barrier in the river is an important
factor because it determines approach flow. The width of gates is another
important consideration (Thames Barrier) and in the case of vertical-lift gates
the clearance for the passage of shipping (Hull Tidal Barrier and Barking Creek
Barrier). At the barriers protecting the navigation passages into the Venice
Lagoon, it was a requirement that there must be no piers in the waterway; this
must also have been the major consideration during design of the Storm Surge
Barrier in the New Waterway in South Holland.

Barrages
Siltation
Siltation can be a problem at barrages. Tidal rivers carry sediment which may be
flushed out by the tides. When fluvial transport rates are a consideration, a
morphological model study may be necessary at the design stage.3 Bottom-
hinged flap gates, which permit only overflow, are selected at barrages to
prevent ingress of saline water. If the low tide recedes from the barrage during
part of the tidal cycle, hook-type vertical-lift gates (Fig. 2.21) enable flushing of
silt to be carried out from the ponded-up watercourse or an upstream bay. This

281
 Hydraulic gates and is effected by raising the lower leaf of the gate, causing discharge under the gate.
valves Wave action may have to be considered at some barrages.

River flood flow at barrages

One of the difficult problems at barrages is the release of river flows during
extreme floods.4 Evacuation problems occur when on rising tide the water level
in the estuary will be higher than the upstream impounded water level swelled
by the flood. In an area like Cardiff Bay, flood water can be stored for a limited
time. When this is not possible, or the retention level of the impounded water
must not be exceeded, the upstream water level may have to be lowered in
anticipation of the flood. Prediction of the flood hydrograph is then required
to carry out a flood routing operation.

Groundwater effects at barrages

Barrages are likely to have groundwater effects which have to be evaluated at the
planning stage.

Water quality
The main function of a barrage is to improve the amenity of the area. This
cannot be achieved if the quality of the impounded water is not acceptable.
The first requirement is to reroute discharges into the river upstream of the
barrage. If saline water is allowed to penetrate the barrage, it can form a stagnant
layer at bed level and can create anaerobic conditions commonly associated with
unpleasant smells.
Under prolonged dry weather conditions and low river flows, dissolved
oxygen levels may be depleted, requiring oxygen injection.
Creating a fresh water reach of river or a freshwater lake can provide
conditions for algal growth during some summer conditions.5

Intertidal mud banks

Migrating birds feed at mudflats which are lost when a barrage is constructed.
About 1% of the migrating bird population of Western Europe require
mudflats. At the Cardiff Bay Barrage it was suggested that alternative feeding
grounds be established to provide a suitable environment for migrating birds
that might otherwise disappear from the area. However, fresh water would
attract new species of animal and plant life to an estuary.

Fish migration
Fish passes are incorporated into barrages.6 Usually migrating fish move from
sea water through brackish water to fresh water. At barrages, the migratory fish
move directly from sea water to fresh water and vice versa. What effect this has on
the fish population is not yet fully established.
It is clear that stratification due to presence of some saline water and the
resulting anaerobic conditions would be harmful to fish and fauna.

282
References Environmental
considerations
1. Beach, M H (1984): Fish pass design ^ criteria for the design and approval of fish passes and
other structures to facilitate the passage of migratory fish in rivers. Lowestoft: MAFF, Fish.
Res. Tech. Rep (MAFF Direct Fish. Res Lowestoft), No. 78.
2. McCallum, I R (1996): Navigation aspects of barrage design, Int. Conference barrages,
engineering design and environmental impacts, Cardiff, Sept. in Barrages, editors Burt, N;
Watts, J, John Wiley and Sons, 1996, pp. 465^478.
3. Han, Z C; Shou, W B; Shao, Y Q (1996): Comparison of siltation between
prediction and field data in downstream of tidal barrage, Int. Conference on barrages
pp. 129^137.
4. H R Wallingford (1994): Tees Barrier Weir: investigation of operating rules, Report 2943,
Jan.
5. Reynolds, C S (1994): The threat of algal blooms in proposed estuarine barrages,
models, predictions, risks, Int. Conference on barrages, pp. 83^89.
6. Gough, P J (1994): Potential impact of estuarine barrages on migratory fish in
England and Wales, Int. Conference on barrages, pp. 73^81.

283
 18
Maintenance and operation
of gate installations

The operational reliability of electromechanical operating systems depends on

regular, systematic maintenance and test operation of gates. In a reliability
assessment it is assumed that the probability of failure on demand increases with
time from the last test operation.
Detailed inspection and regular maintenance routines for gates are specified
by the gate manufacturer in the maintenance manual. For electrical and
mechanical elements such as power supply, electrical distribution, gate hoists
and instrumentation, they will vary according to the type of gate and the
detailed design.
There are some requirements which will affect the design of gate
installations, such as access and safety of personnel who have to carry out
maintenance operations.

Access
Spillway gate installations and river control structures require stoplogs to
permit temporary closure of the fluidway for emergency situations, gate
maintenance or repair. The slots for placing stoplogs or bulkheads should
provide for adequate working space and the erection of scaffolding between
the gate and the closure structure. At most spillway gates only upstream
stoplogs are required, unlike river control gates where the downstream water
level is frequently above the sill and a second set of stoplogs may be necessary.
At large gates where the discharge under a gate can be at high velocity,
stoplog or bulkhead slots can cause disturbance of the approach flow, such as
eddies and vorticity. Where this is a factor, stoplog slots can be provided with
withdrawable masking plates to produce a smooth face at the pier or abutment.
Gates require inspection and access to parts which have to be serviced or
replaced. On large radial gates, a walkway along the uppermost gate arms
should be provided together with access ladders for inspection for corrosion
of structural members, rope or chain anchorages and seals. It is important that
welds for attachments and brackets supporting access ladders and handrailing
are not located so as to cause stress concentrations at the gate structure.
Gate inspection and side-seal replacement is often carried out with radial
gates in the elevated position. For safety reasons, radial gates should be dogged

285
 Hydraulic gates and
valves

Figure 18.1. Gate dogging

device

when maintenance, repair or repainting is carried out on an open gate. Dogging

is effected by providing a block on each side of a gate to prevent the gate from
being lowered. It often takes the form shown in Fig. 18.1.

Inspection
In practice, gate structures rarely need to be inspected unless there are signs of
distress or significant corrosion. The recommended procedure is to develop an
inspection plan for each different design of gate. For radial gates this is set out in
the US Corps of Engineers specific requirements for inspection of hydraulic
steel structures.1 It is not economical to carry out a detailed inspection of all
the gates of a spillway gate installation or a barrage. It should be limited to
critical areas, that is, the elements which are fracture critical and whose failure
would cause collapse or render the gate inoperative.
Fracture critical locations, or areas susceptible to weld-related cracking, may
include trunnion weldments, trunnion beams, gate arms (particularly at their
attachments to the skin plate assembly), and welds attaching lifting brackets.
Intersecting welds are also sometimes subject to stress concentration.
Visual examination is the first method used to inspect all critical elements. If
cracks are suspected, non-destructive test methods should be used, such as dye
penetrant, magnetic particle or ultrasonics.
Inspection for corrosion should be part of the regular maintenance
procedure. Before there is a noticeable breakdown of the paint system, local
corrosion may have started at crevices, the junction of dissimilar metals, seal
clamping plates, location of bolts, drain holes, sharp corners or edges.
Guidelines to quantifying corrosion damage are given in Greimann et al.2
Kumar and Odeh3 provide information on the corrosive behaviour of stainless
steels under conditions experienced at gates.

286
 Regular inspection of ropes and lifting chains is a general requirement, as Maintenance and
well as shackles and pins for hoisting attachments. Where ropes and chains are operation of gate
installations
anchored upstream of the skin plate they are immersed for long periods in water.
Even where ropes and chains are attached on the downstream side of gates they
are often subject to partial immersion and/or splashing. Chains are subject to
corrosion at their links, causing failure to articulate over the lifting sprockets
resulting in uneven hoisting of a gate. Severe loads due to racking of a gate
can result from corrosion of chain pins.
Debris can lodge between the gate arms of a radial gate and the pier near the
trunnion, at side seals and structural members on the downstream side of gates.
At vertical-lift gates there can be problems due to debris in gate slots,
particularly at the closely spaced rollers of Stoney-roller gates. Debris shields
are often provided to prevent penetration of floating matter into gate slots. At
some gates, debris deflection shields are also fitted at side seals. Deflector plates
should preferably be fabricated from abrasion resistant, lightweight material
such as ultrahigh-molecular-weight polyethylene.
One of the most important components of radial gates requiring regular
lubrication is the trunnions, unless these are of the self-lubricating type (see
Chapter 7). The collapse of a spillway gate at the Folsom Dam in California
due to corrosion on the steel trunnion pins was mentioned in Chapter 12. The
excessive friction which developed as a result of corrosion caused flexure
stresses in the gate arms which resulted in shearing of a strut brace. The
recommendations of the investigation into the collapse of the gate4 included
adding grease while the gates are in motion, and that greasing should be carried
out whenever the gate is raised or lowered. Small electric motor operated piston
pumps, which supply automatically metered fluid grease wherever the hoist
motor is started, satisfy this requirement. The report also stressed that the
selection of the correct grease is critical.
Grease for trunnion lubrication must satisfy the following requirements:
prevention of rust
resistant to water washout
must be soft to be pumped into clearance of the load zone
mineral-base oil must be used
must have antiwear, antiscuff properties; the recommended grease must
contain soluble antiwear additives, such as sulphur and phosphorous
compounds but no molybdenum disulphide or polytetrafluoroethylene
must have good adherence properties, that is, contain the chemical additives
iso-butylene or polyethelene
non-corrodible towards bronze, low copper corrosion properties
long stable life
will not separate in storage.

Routines for the maintenance of electromechanical and oil hydraulic systems,

and for electrical supply and distribution, are well established. Because
environmental conditions at most gate installations are more onerous than at
most engineering installations, maintenance operations should be carried out
based on time intervals and not on usage. For instance, changes of oil in hoist

287
 Hydraulic gates and gearboxes should be regular to prevent the build-up of water and other
valves contaminants in the oil. Similarly, moving parts of the hoist, such as bearings
and couplings, should be lubricated regularly to protect surfaces from corrosion
and expel contaminants.
Chapter 12 identified a number of elements of gate installations which have
frequently caused failures, such as limit switches, heating installations, winding
screws, brake failures, ropes and others. These should be regularly inspected,
tested where possible and maintained to a high standard.
If gate vibration is noted, even if confined to a restricted range of gate
opening, it should be reported5 immediately. The cause should be established
and, if possible, vibration should be eliminated.
Test opening of gates should be carried out regularly. (This is sometimes
omitted at spillway gates because of loss of reservoir water.)
If opening of gates is carried out too rapidly it can cause a surge wave to travel
down river endangering river users. If the extent of test opening and closing a
gate, or the rate of opening is not acceptable because of loss of water and
possible danger to river users, the placing of stoplogs should be considered.
Gates at bottom outlets of reservoirs are sometimes not test operated because
of the high velocity of discharge. Since most terminal discharge gates and valves
are backed by a second gate or valve, these can be shut to permit the primary gate
or valve to be test operated under no flow conditions. This is an acceptable
practice for regular testing but is not a substitute for testing a gate under load
at less frequent intervals.

References
1. US Army Corps of Engineers: Responsibility for hydraulic structures, Specfication
ER1112-02-8157.
2. Greimann, L; Stecker, J; Rens, K (1990): Management system for mitre lock gates,
technical report REMR-OM-08, US Army Engineer Waterways Experiment
Station, Vicksburg, MS.
3. Kumar, A; Odeh, A A (1989): Mechanical properties and corrosion behaviours of stainless
steel for locks, dams and hydroelectric plant applications, technical report REMR-EM-6,
US Army Engineer Waterways Experiment Station, Vicksburg, MS.
4. Bureau of Reclamation (1996): Forensic report on spillway gate 3 failure, Folsom Dam,
Bureau of Reclamation, Mid-Pacific Region, Sacramento, California, Nov.
5. Noble, M; Lewin, J (2000): Three cases of gate vibration, Proc. of the Biennial
Conference of the British Dam Society, Bath, Jun., in Dams 2000, Thomas Telford,
editor Tedd, P, pp. 95^108.

288
Appendix
Calculation of hydrostatic
load on radial gates

Gate loads due to water pressure

W gate width
a height of gate
e centre of pressure of hydrostatic thrust on the gate skin plate
b upstream water level
c downstream water level
d height of trunnion
GO gate opening
TH horizontal hydrostatic thrust on the gate skin plate
TV vertical hydrostatic thrust on the gate skin plate
TR resultant of horizontal and vertical hydrostatic thrust passing
through the trunnions
a angle of TR with respect to the horizontal
g gravitational constant

Linear units are in metres, forces are in kN.

Note: The vertical force is the water displaced by the gate. The horizontal
and the vertical component of hydrostatic thrust pass through the centroid
of the area in question, the resultant in turn passing through the axis of the
trunnions.
The ratio d/R (Fig. A.1) should be equal to or less than 0.707 to ensure that
the sill seal is at an angle of at least 45 to the sill beam. Angles less than 45 can
cause leakage at the sill, which can result in gate vibration.
The force diagrams are valid for gates in the closed position. The forces
are lower when there is discharge under the gate. A more precise
determination of at least the horizontal component under discharge
conditions requires an estimation of the discharge characteristics of the gate,
see Hunter Rouse.1

289
 Hydraulic gates and
valves

Figure A.1. Gate load due to

water pressure

Gate shut, no downstream water level

Go O C O

This is the most severe loading condition. In gates in river installations when
there is always a downstream water level above the sill it can occur due to
inadvertent drainage of the downstream reach or testing of the gate behind
stoplogs.

TH horizontal hydrostatic thrust 12 b2 gW kN

e b/3
1 cos1 d=R
3 cos1 d b=R
2 3 1
Z1 R sin 1
Z2 R sin 3

Figure A.2. Forces due to water pressure on a closed gate

290
Z3 Z2 Z1 Calculation of
area 1 A1 1=2Z3 b hydrostatic load on
radial gates
distance AB chord length 2R sin2 =2
area of triangle ABC 1=2R cos2 =2  2R sin2 =2
R2 cos2 =2sin2 =2
area of sector ABC R2 2 =360
area 2 A2 R2 2 =360 R2 cos2 =2sin2 =2
Tv vertical hydrostatic thrust A
p1 A2 gW kN
TR resultant hydrostatic thrust TH2 Tv2 ) kN
 tan1 TV =TH

Gate in the open position, downstream water

level above gate lip (drowned discharge)
The analysis assumes hydrostatic pressures.

TH2 resultant horizontal pressure

e2 centre of pressure of horizontal force on gate measured from sill
1 1
TH2 b Go 2 Wg c Go 2 Wg
2 2
1
Wgb Go c Go 2
2
2    
1 b Go 1 c Go
e2 Go TH2 b Go 2 Wg  c Go 2 Wg 
2 3 2 3
" #
Wg b Go 3 c Go 3
e2 Go
2TH2 3 3
Z1 R sin 4
Z2 R sin 6
Z3 Z
p2 2 Z1
Z4 R d c2
Z5 Z4 Z1

Figure A.3. Forces due to

water pressure on a gate in the
open position with drowned
discharge

291
 Hydraulic gates and Z3 Z5
valves
area 1 A1 b c
2
4 cos1 d c=R
   
1 d c b c 1 d b
6 cos cos
R R
5 6 4

area of sector ABC R2 5 =360

R2 5 5 5
area 2 A2 R2 cos sin
360 2 2
TV2 vertical hydrostatic thrust A
p1 2 A2 gW kN
2
TR2 resultant hydrostatic thrust TH2 TV2 kN
TV2
 tan1
TH2

Overflow gate, overflow not curtailed

Overflow and flow under the gate should never occur together except for the
short time when the gate is raised, see Fig. A.4.

For c 0
1 2 1
TH3 b Wg b1 a2 Wg
2 1 2    
b1 1 2 b1 a 1 2
e3 TH3  b Wg  b1 a Wg
3 2 1 3 2

Wg 3
e3 b b1 a3
6TH3 1
area 3 A3 Z3 b1 a

Figure A.4. Gate load due to

water pressure under overflow
conditions

292
TV3 vertical hydrostatic thrust A1 A2 A3 gW kN Calculation of
hydrostatic load on
p 2 2 radial gates
TR3 resultant hydrostatic thrust TH3 TV3 kN
TV3
3 = tan1
TH3
For c > 0
1 2 1 1 1
TH4 b1 Wg b1 a2 Wg c2 Wg TH3 c2 Wg
2 2 2 2
      
b1 1 2 b1 a 1 2 c 1 2
e4TH4  b Wg  b1 a Wg  c Wg
3 2 1 3 2 3 2

b31 Wg b1 a3 Wg c3 Wg

6 6 6

Wg 3
e4 b b1 a3 c3
6TH4 1
TV4 vertical hydrostatic thrust A
p1 2 A2 A 3 kN
2
TR4 resultant hydrostatic thrust THA TVA kN
TV4
4 tan1
TH4
The calculations are an overestimate which ignores the velocity head due to flow
over the gate.
 2
v
To refine calculations b2 b1
2g
where b1 energy head
v velocity of the approach flow to the gate

Correction of the hydrostatic load on a gate subject to overflow

the correction is approximate

the overflow is not curtailed

it is assumed that the piers or abutments project upstream and that the

approach flow is therefore not subject to side contraction.

Notation as Fig. A.2.
Q overflow
b1 a overflow head
W gate width
V1 velocity of overflow
hV velocityphead
Q 0:57W g  b a3=2
Q
V1
Wb a

293
 Hydraulic gates and
valves

Figure A.5. Gate load due to

water pressure for a
submergible gate
V12
hV
2g
effective hydrostatic head on gate b1 hV
This value is now substituted for b1

A more accurate formula for Q is

 
0:150b a p
Q 0:564 1 W g  b a 0:0013=2
a
For calculation of the approach velocity, see Ackers et al.2

Submergible gate
Use as condition 3. The forces on the gate below the sill are due to the
downstream level and balance.

References
1. Rouse, H (1950): Engineering hydraulics, Chapter VIII, John Wiley and Sons, New
York.
2. Ackers, P; White, W R; Perkins, J A; Harrison A J M (1978): Weirs and flumes for
flow measurement, John Wiley and Sons, p. 47.

294
Index
access, gate installations, 285286
added mass, gate vibration, 191194
air demand, conduits, 180182, 201
air supply pipes, submerged outlet gates,
51
articulated vertical lift gates, 20
automatic control methods, 209216
automatic control systems, choosing,
217218
automatic crest and scour gates, 1617, 63
automatic tilting gates, 27, 29
Barking Creek Barrier, 41, 43
Barkley Dam, Kentucky, 189190
barrier and barrage gates, 3548
bottom-hinged buoyant gates, 3941
bottom-hinged flap gates, 47
caisson gates, 48
environmental considerations, 281282
flap gates, 4144
hook-type double leaf gates, 4748, 64
large span verticallift gates, 41
lock gates, 4447
pointing gates, 48
rising sector gates, 3539
tidal power barrages, 48, 64
battery powered standby systems, 220221
bear-trap gates, 5051, 66
bed protection, 46
Bondi scheme, 3031
bottom-hinged buoyant gates, 3941, 65
bottom-hinged flap gates, 2330, 64
debris, 2627, 9799
disadvantage, 25
gravity standby systems, 222
nappe oscillation, 2930
oil hydraulic operation, 120
seals, 134137, 138
storm surge protection, 47
venting, 2930
versions, 25
bottom outlet tunnel gates, earthquakes,
256257
bottom outlet valves, 89
bottom outlets, failures, 232234
bubbler devices, 224225
butterfly valves, 7178, 90
bottom outlet valves, 89
cavitation, 7578
closure types, 7375
loss coefficients, 7576
seal arrangements, 7173
caisson gates, 48
Cardiff Bay Barrage, 4748
cascade controls, 209210
caterpillar gates, 51, 53, 54, 67
cathodic protection, 268269
cavitation
conduits, 173174
gate slots, 174178
Tarbela Dam, Pakistan, 233
valves, 7578
chains, 148149
circuits, oil hydraulic operation, 122126
circular-orifice gates, 56, 58
closed-loop systems, control, 211
coaster gates, 51, 53, 54, 67
composite construction, vertical-lift gates,
109113
computer assisted control methods, 209,
210216
computer program, radial gates, 1415
conduits
air demand, 180182, 201
cavitation, 173174
erosion, 173174
gate conduits, 178179
gate slots, 174178
hydraulic considerations, gates, 172182
proximity, two gates, 180
trajectory of jets, 179
vibration, 180
vorticity, 172173, 188190, 201
control and guard gates, submerged
outlets, 5359
295
 Hydraulic gates and control buildings, earthquakes and, 261
valves control systems and operation, 207226
choosing, 217218
control objectives, 207208
failures, 235
fall-back systems, 218223
instrumentation, 223225
operating rules and systems, 209216
standby facilities, 218223
telemetry, 216217
control weir, radial gates, 11
corrosion, steel, 266268
counterbalance, radial gates, 1314
cranes, emergency closure gates, 60, 62
culvert gates, vibration, 189
culvert valves, radial gates as, 8
culverts, screens, 96
cylinder gates, 5253, 67
cylinder intake gates, 56, 67
damping, vibration, 186187
debris, 9399
bottom-hinged flap gates, 2627, 9799
environmental considerations, 278279
floating booms, 99
overflow gates, 97
radial gates, 107, 108
underflow gates, 97
design criteria, gates, 101103
design standards, 101103
detail design aspects, 127150
diesel generators, standby, 219220
discharge characteristics
fuse gates, 35
radial gates, 158162
discharge measurement, 223224
displacers/displacer chambers, radial gates,
1113
downpull forces, hydraulic, 166171
downstream offsets, gate slots, 177178
downstream roller and turbulent flow, 197
downstream water level control, radial
gates, 13, 14
drowned flow
rectangular flap gates, 165166, 168
stagedischarge relationship, 165166
drum and sector gates, 4850, 66
earthquakes
analysis methods, 254255
design loading, 102103
effects on gates, 251263
sample event tree, 261262
spillway gate installations, 252254
Eastern Scheldt Storm Surge Barrier, 274
elastomeric gate leaf, flap gates, 3031
electrodes, water level measurement, 224
296
electromechanical drives, 116120
embedded parts, 151155
emergency closure gates, 5962
empirical stagedischarge relationship, 166
environmental considerations, 277283
equivalent-static analysis, 254255
erosion, conduits, 173174
excitation frequencies, gate vibration,
188190
failure modes and effects analysis (FMEA),
239
failure modes, effects and criticality
analyses (FMECA), 239
failures
bottom outlets, 232234
control systems, 235
fault frequency, gate type, 234235
hollow-cone valves, 7981
screens, 96
spillway gate installations, 227232,
242243
trunnion bearings, 142144
fall-back systems, control systems and
operation, 218223
Fault Tree Manager, 238, 240
fish-belly flap gates, 24
fixed roller gates, 1719
flap gates, 2333
bottom-hinged see bottom-hinged flap
gates
elastomeric gate leaf, 3031
storm surge protection, 4144
top-hinged see top-hinged flap gates
flat bottomed gates
added mass, 190
vibration, 190
flawed design, 1
float actuation, water level measurement,
224
floating booms, debris, 99
floats, radial gates, 12
flood routing, 208, 218
flow attachment, gate vibration, 196199
flow oscillation, 171172
flow over gates, 162164
flow reattachment, gate lip, 199200
flow under gates, 158162
Flowgate Projects, 1617
FMEA see failure modes and effects
analysis
FMECA see failure modes, effects and
criticality analyses
Folsom dam, Sacramento, 142144, 227,
229, 230
free flow, stagedischarge relationship,
164165
free-rolling gates, 2122, 23, 64
free shear layer, vibration, 200, 201202
free surface flow gates, 451
hydraulic operation, 120126
Froude scale models, 271273
fuse gates, 3335, 65
gate arms
arrangement, conventional, 7
constructional features, 79
radial gates, 103107
gate conduits, 178179
gate design guidelines, 197199
gate dogging device, access, 286
gate guide rollers, vertical-lift gates,
151155
gate installations, maintenance and
operation, 285288
gate lips
flow reattachment, 199200
hydraulic downpull forces, 199200
gate position measurement, 225
gate slots
cavitation, 174178
conduits, 174178
downstream offsets, 177178
gate vibration, 185205
added mass, 191194
damping, 186187
excitation frequencies, 188190
flow attachment, 196199
flow reattachment, 199200
hydraulic downpull forces, 199200
preliminary check, 192194
pressure waves, 190
seal leakage, 194196
slack, gate components, 202
two-phase flow, 202203
types, 185
unstable flow, 200201
vibrating system, 186188
vortex trails, 188190, 201
gated river control structures,
environmental considerations,
277281
gates
design criteria, 101103
free surface flow, 451
hydraulic considerations, 157184
structural considerations, 101114
submerged outlets, 5162
types, 369
`Gibb' gates, 221222
gimbal mountings, hydraulic cylinders, 122
grappling beams, stoplog, 61, 62
gravity, 10, 12, 13, 115
standby systems, 221222
guard valves, matching terminal discharge Index
valves, 89
guide rollers, 137140
guide wheels, preloading, vertical-lift
gates, 203, 204
hazard and operability study (HAZOP),
239, 240
hazard, hydraulic gates, 227245
high head gates, 151155
hoist rope attachments, radial gates,
119120
hoist speed, 124, 126
hollow-cone valves, 7883, 90
bottom outlet valves, 89
hoods, 8283
point of flow attachment, shifting, 8183
submerged, 8283
vane failure, 7981
hollow-jet valves, 8485, 90
hook-type double leaf gates, 21, 4748, 64
hook-type gates, 18, 20, 21, 64
Howell and Bunger valves see hollow-cone
valves
hydraulic considerations, gates, 157184
conduits, 172182
flow over gates, 162164
flow under gates, 158162
hydraulic downpull forces, 166171
hysteresis effect, gate discharge, 172
limited ponded-up water, 171
reflux downstream, 171172
stagedischarge relationship, 164166
three-dimensional flow entry,
sluiceways, 171, 172
hydraulic operation see oil hydraulic
operation
hydrostatic load calculation, radial gates,
289294
hysteresis effect, gate discharge, 172
ice formation, 247250
inlet system, radial gates, 1011
inspection, gate installations, 286288
instability causes, radial gates, 14
instrumentation, control systems, 223225
intake gates
radial, 55, 67
submerged outlets, 5153, 66
jamming, radial gates, 107, 109
jet-flow gates, 56, 58, 68
King George V lock, storm surge
protection, 4144
Kotmale dam, Sri Lanka, 93
Kotri Barrage, Pakistan, 3, 41, 42
297
 Hydraulic gates and large span verticallift gates, 41 pointing gates, 48
valves Legadadi Dam, Ethiopia, 222 PRA see probabilistic risk analysis
level control, 210211 pressure methods, water level
limit switches, 145147 measurement, 224225
limited ponded-up water, 171 pressure-reducing valves, 8587, 87, 91
lintel seals, leakage vibration, 196 pressure waves, gate vibration, 190
load and resistance factor design (LRFD), probabilistic risk analysis (PRA), 239
101103 programmable logic control (PLC), 218,
load factors, 101103 237
load rollers, 137140 proportional control, 211
local control methods, 209 proportional integral derivative (PID)
lock gates, storm surge protection, 4447 control, 210216
locomobiles, sector lock gates, 45, 47 proportional plus derivative (PD) control,
LRFD see load and resistance factor design 211212
lubrication proportional plus integral (PI) control,
chains, 149 211212
trunnions, 142145 protection, materials and, 265269
wire ropes, 148 proximity switches, 147
pseudo-static analysis, 254255
machinery, operating, 115126
maintenance, gate installations, 285288 radial automatic gates, 915, 63
maintenance gates, 5962 radial gates, 415, 5658
manual control methods, 209 advantages, 4
materials and protection, 265269 arms arrangement, 7
mitre gates, 4447, 66 computer program, 1415
model studies, 271276 control weir, 11
motor drives, 116120 counterbalance, 1314
Mrica Hydroelectric Project, Indonesia, as culvert valve, 8
179 debris, 107, 108
disadvantages, 4
nappe oscillation, bottomhinged flap discharge characteristics, 158162
gates, 2930 displacers/displacer chambers, 1113
navigation clearance, 41, 4244 distribution of pressure head, 6
needle valves, 8586, 86, 91 downstream water level control, 13, 14
New Waterway, Rotterdam, 45, 46, 47 electromechanical drives, 116120
embedded parts, 151155
oil hydraulic operation float operated, 63
advantages, 121 floats, 12
disadvantages, 121 gate arms, 103107
features, 125126 hoist rope attachments, 119120
free surface flow gates, 120126 hydraulic forces, 56
hydraulic circuits, 122126 hydrostatic load calculation, 289294
leakage, 120, 125 inlet system, 1011
operating machinery, 115126 instability causes, 14
operating rules and systems, 209216 jamming, 107, 109
operation, gate installations, 285288 malfunction causes, 14
motorised, 63
painting, protection and, 266268 oil hydraulic operation, 120126
PD see proportional plus derivative control piers, dividing, 14
Pershore Mill gates, 196197 roller faces, 151155
PI see proportional plus integral control seals, 132134
PID see proportional integral derivative side guide rollers, 137140
control side seal contacts, 151155, 259
piers, dividing, radial gates, 14 stiffening, 103107
PLC see programmable logic control structural design, 103107
point of flow attachment, shifting, 8183, trunnions, 13, 103104, 142, 144145,
196199 146

298
 two-phase flow, 202203
types, 5
vibration, 189190
water level control, 910
wide span, 106107
radial intake gates, 55, 67
Randenigala Project, Sri Lanka, 179
rectangular flap gates, drowned flow,
165166, 168
reflux downstream, 171172
reliability, hydraulic gates, 227245
reliability indices, 241242
remote control methods, 209
reservoirs, control objectives, 207208
ring-follower gates, 5859, 68
Riprap bed protection, 46
rising sector gates, 3539, 65, 66
risk assessment, 235243
River Hull Tidal Surge Barrier, 41, 42, 42
River Hunte, storm surge protection, 44
rivers, control objectives, 207
roller and turbulent flow, downstream, 197
roller faces, radial gates, 151155
rolling-weir gates, 22, 24, 64, 125
ropes, 119120, 147148
rotary limit switches, 145146
rotary valves see sphere valves
Rotterdam, storm surge protection, 45, 46
San Roque Dam, Philippines, 178179
screens, 9399
in culverts, 96
design criteria, 9495
failures, 96
instrumentation, 9697
raking, 9798
in river courses, 96
vibration, 9596
seals, 127137
bottom-hinged flap gates, 134137, 138
radial gates, 132134
seal leakage, 194196
seal shapes, 128
vertical-lift gates, 131, 134, 135, 136
Selsyn motor drives, 117118
Severn Estuary, tidal power barrage, 48
Severn Tidal Project Study, 3031, 4850
side guide rollers, radial gates, 137140
side sealing
heating, 248
leakage vibration, 196
radial gates, 151155, 259
sill beams, 151155
sill seals, leakage, 195196
slack, vibration and, 204
sliding gates, 48, 57, 67, 125
sliding paths, vertical-lift gates, 151155
sluice valves, 71, 90 Index
sluiceways, three-dimensional flow entry,
171, 172
sphere valves, 8788, 91
spillway gate installations
earthquakes, 252254
events, 227232
failures, 227232
fault tree, 242243
vibration, 228230, 231
stagedischarge relationship, top-hinged
flap gates, 164166
standby facilities, 218223
stiffening, 196197, 198
radial gates, 103107
vertical-lift gates, 107113
still seals, embedded parts, 151155
Stoney-roller gates, 1719, 139
stoplogs, 5962, 161
grappling beams, 61, 62
stoplog guide channels, 62
storm surge protection, 65, 66
flap gates, 4144
lock gates, 4447
stress analysis, 109113
Strouhal numbers, gate vibration, 188190
structural considerations, gates, 101114
structural design
radial gates, 103107
vertical-lift gates, 107109, 198
submerged outlets, gates, 5162, 66
surge control, 215216
systems analysis, earthquakes, 257260
Tarbela Dam, Pakistan, cavitation, 233
telemetry, control systems, 216217
terminal discharge valves, matching guard
valves, 89
terminology, notes on, 2
Thames Barrier, rising sector gate, 3539
three-dimensional flow entry, sluiceways,
171, 172
three-dimensional models, 272273
thyristor-controlled drives, 118119
tidal power barrages, 48, 64
tilting gates, seals, 134137
Tokyo City, storm surge protection, 4445
top-hinged flap gates, 3033, 64
adjustable pivot lugs, 32
elastomeric gate leaf, 3031
hydraulically cushioned, 33
stagedischarge relationship, 164166
Torrumbarry Weir, Australia, 189190
trashracks see screens
trunnions
bearing failures, 142144
lubrication, 142145
299
 Hydraulic gates and mountings, 142, 144145, 146
valves radial gates, 13, 103104, 142, 144145,
146
trunnion assembly, 140142, 143, 144
tunnel gates, submerged flow conditions,
166, 168169
tunnel lining sections, high head gates,
151155
turbulent flow, 196199
two-dimensional models, 272273
two-phase flow
gate vibration, 202203
problems, 272
types of gates, 369
summary, 6368
units, notes on, 2
unstable flow, gate vibration, 200201
valve position measurement, 225
Venice Barrier gates, 3941
venting
bottom-hinged flap gates, 2930
overflow gates, 201
see also air demand
vertical-lift gates, 1722, 64, 65, 67
advantages, 17
composite construction, 109113
disadvantages, 17
embedded parts, 151155
free-rolling gates, 2122, 23
gate guide rollers, 151155
300
guide wheels, preloading, 203, 204
load rollers, 138140
long span, 18
overflow sections, 18, 20
reinforcing methods, 107, 110
seals, 131, 134, 135, 136
sliding paths, 151155
stiffening, 107113
structural design, 107109, 198
vibration, 189190
vertical-lift spillway gate installations,
earthquakes, 255256
vertically hinged sector lock gates, storm
surge protection, 4447, 66
vibration
bottom-hinged flap gates, 2930
conduits, 180
cylinder gates, 53
free shear layer, 200, 201202
gate see gate vibration
hollow-cone valves, 7983
model studies, 273275
screens, 9596
seal leakage, 194196
spillway gate installations, 228230, 231
Victoria Dam, Sri Lanka, 54, 115, 209210,
221222
vortex trails, gate vibration, 188190, 201
vorticity, conduits, 172173
water level measurement, 223, 224225