Jack Lewin-Hydraulic Gates and Valves in Free Surface Flow and Submerged Outlets-T. Telford (1995).pdf

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 ß 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

3.

Gates in free surface flow Radial gates Radial automatic gates Other water operated gates Vertical-lift gates Rolling-weir gates Flap gates Fuse gates Barrier and barrage gates Drum and sector gates Bear-trap gates Gates in submerged outlets Intake gates Control and guard gates Emergency closure gates, maintenance gates and stoplogs Summary of types of gate

4 4 9 15 17 22 23 33 35 48 50 51 51 53 59 63

Valves

71

Sluice valves Butterfly valves Cavitation in valves Hollow-cone valves and hoods Hollow-jet valves Needle valves Pressure-reducing valves Sphere valves Matching terminal discharge valves and guard valves Summary of types of valve

71 71 75 78 84 85 85 87 89 90

Contents

4.

5.

93

Trashracks and screens in submerged intakes Trashracks and screens in culverts and river courses Screen instrumentation Screen raking Debris

93 96 96 97 97

Structural considerations

101

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

101 103 107 107 109

Operating machinery

115

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

116 121 126

Detail design aspects

127

Seals Guide and load rollers Trunnion assembly Trunnion bearing failure and lubrication Trunnion mounting for radial gates Limit switches Ropes Chains

127 137 140 142 144 145 147 148

8.

Embedded parts

151

9.

Hydraulic considerations pertaining to gates

157

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

158 166 171 171 171 172 172

Gate vibration

185

Types of gate vibration The vibrating system Excitation frequencies Added mass

185 186 188 191

6.

7.

10.

iv

Trashracks, screens and debris

11.

12.

13.

14.

15.

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

192 194

199 200 201 201 201 202 202

Control systems and operation

207

Control objectives Operating rules and systems Telemetry Factors in the choice of automatic gate control systems Fall-back system and standby facilities Instrumentation

207 209 216 217 218 223

Hazard and reliability of hydraulic gates and valves

227

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

227 232 234 235

Ice Formation

247

Ice

247

Earthquake effects on gates

251

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

252 254 255 255 256 257 257 261 261

Materials and protection

265

Materials Steel corrosion and painting Cathodic protection

265 266 268

Contents

196

v

Contents

16.

17.

18.

vi

Model studies

271

Froude scale models Two-phase flow problems Two- and three-dimensional models to Froude scale Models for investigating vibration problems

271 272 272 273

Environmental considerations

277

Gated river control structures Barriers Barrages

277 281 281

Maintenance and operation of gate installations

285

Appendix: Calculation of hydrostatic load on radial gates

289

Index

295

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 KÏnz 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 valves

Note on units 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 conditions, 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


Gates in free surface flow 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) Hydrostatic forces acting on a radial gate Figure 2.3. Lifting and lowering forces acting on a radial gate

5


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


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 conditions 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


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


Figure 2.14. Radial gate using floats (11.5 m radius, 15 m width)

12

the inlet weir is drowned and the gate operates according to the ratio R1/R2 as 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 à' 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

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

Types of gates

Trunnions Most gate manufacturers use phosphor bronze bushed bearings. Selflubricating 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


environmental conditions require a less obtrusive appearance, as at Pulteney 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: ú ú ú ú

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: ú ú

ú ú

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 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.

Types of gates

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


Figure 2.17. Automatic crest gate (after Townshend7)

16

Control of the gate response to inflow into the float chamber is by a penstock 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

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) (b) (c) (d) (e)

the requirement for gate slots possibility of trash jamming in the wheels overhead structure wheels have to rotate underwater 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


18

lifted, the counterbalance is reeved 2:1 so that it travels for only half the distance. 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 overand 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.

Figure 2.19. Types of vertical-lift gate

19


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


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 à' 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 è' 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


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 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

Types of gates

Figure 2.25. Different versions of bottom-hinged flap gates operated by hydraulic cylinders

25


Figure 2.26. Recess in sluiceway to permit complete retraction of flap gate under flood conditions

26

þ3 mm along both axes over the distance of gate movement. While this saves the 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.

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


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


design and spacing of flow dividers is critical.3 The dividers must project beyond 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.9–14 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


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 oleopneumatic 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


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, 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: ú

Types of gates

the additional loads imposed by fuse gates.

The main criteria for selection of fuse gates on all dams are: ú

ú

ú

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


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 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.

Types of gates

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

37


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


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 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

Types of gates

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 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 ÿ11.85 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


Figure 2.48. A tophinged 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. Verticallyhinged 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


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 crosssection of the floating sectors (after Ieperen23)

47


leaf can be lowered to provide overflow from the upstream reservoir and the 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


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 à' 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 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 à' 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.

Types of gates

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


Figure 2.57. Intake gate, servomotor 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


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 crossradius 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


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


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


The term stoplog derives from the time when it was general practice to isolate 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 à', 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 è' on the stoplog is positioned on the sill beam or on top of another section. The rod is displaced upwards and permits disengagement 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


Stoplog guide channels 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 à', 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 railmounted 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

Types of gates

Summary of types of gate Type 1. Radial gates motorised

2. Radial automatic gates

Main application Advantages Gates in open channels No unbalanced forces Absence of gate slots Low hoisting force Mechanically simple Bearings out of the water Can be fitted with overflow section Some inspection possible with gate in service Sluice No outside source of installations power required River control Absence of machinery Low maintenance Sluice installations River control Spillways Barrages

3. Radial gates float operated

Spillways

No outside source of power required Absence of machinery

4. Automatic crest gates

Limited application at spillways

No outside source of power required Absence of machinery Low maintenance

5. Automatic scour Gates

Limited application at low level outlets of low height dams

No outside source of power required Absence of machinery Low maintenance

Disadvantages Extended flume walls High concentrated loads Increased fabrication complexity

Wide piers to accommodate displacers Counterbalance visually intrusive Can malfunction due to incorrect design Can malfunction due to blockage of inlet or control system Wide piers to accommodate counterbalance and floats Can malfunction due to rope system Can malfunction due to blockage of inlet or control system Requires special type of spillway Coarse level control compared with 2. Limited height of gate compared with 1. and 2. Limited height of opening Limited upstream head Free discharge only Coarse control

63


Type

Main application Advantages

6. Vertical-lift gates

Sluice installations River control Old installations: ú barrages ú spillways

7. Vertical-lift gates hook type

River flow control Barrages where ebb tide exposes gates

8. Free rolling gates

Shut-off of turbine intakes

9. Rolling-weir River flow gates control

64

10. Flap gates bottomhinged

Tidal barrages Sluice installations River control

11. Flap gates top-hinged

Tidal outlets

Disadvantages

Gate slots required Load rollers under water can jam due to debris High hoisting load unless counterbalanced Overhead support structure visually intrusive Gate slots required Load rollers under water can jam due to debris 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 No longer Very reliable Capable of operating at manufactured differential heads up to Guide slots required High hoisting load 30 m Can be manufactured Complex to fabricate very wide Vulnerable side seals Debris can become lodged in rack Complete separation of Requires extensive side staunchings for saline and fresh water Overflow to clear debris side sealing or very accurately constructed No visually intrusive pier walls overhead structure Hinge bearings not Can in some cases be easily accessible and engineered to open permanently under gravity in an immersed emergency Cannot control water No outside source of level power required Automatic operation Will not entirely exclude tidal water if Simple construction downstream water Low maintenance level rises above sill Gate slam can occur Can be fitted with overflow sections Short piers Wide span gates can be engineered to provide good navigation openings Up and over gates can reduce height of supporting structure Greater independent control of overflow and discharge under gate than 6. with a flap section for overflow

Type 12. Fuse gates


Disadvantages

Limited application to spillways

Operates as fixed weir until it tilts Gates have to be recovered after tilting and may require repair or replacement Use may depend on frequency of flood events Requires piers Permits some flow upstream Complex to fabricate Complex machinery

13. Rising sector Storm surge gates barriers

14. Buoyant gates bottomhinged

Storm surge barriers Tidal barriers

15. Vertical-lift gates for navigation channels

Storm surge barriers

16. Vertical-lift Storm surge gates for non- barriers navigation channels

Automatic operation Simple construction Easy addition to height of a dam Particularly applicable in an unsophisticated environment Unobstructed navigation passage Not visually intrusive Can be raised for maintenance and inspection without placing stoplogs Unobstructed navigation passage without piers No structure above navigation passage bed level Excellent from visual considerations

High clearance and large span can be achieved No underground passages required Gravity closure in an emergency Conventional hoist machinery Can be maintained in high level position subject to safety requirements No underground passages required Gravity closure in an emergency Conventional hoist machinery Can be maintained when elevated above waterway

Types of gates

Capable of withstanding only limited differential head Gates move independently under wave action resulting in leakage between gates Requires detachable hinges Gates have to be interchanged for maintenance High overhead structure Gates are normally counterbalanced Large gates require lifting chains of the type where each pin can be lubricated

Lock required Separate overhead structure Gates are normally counterbalanced Large gates require lifting chains of the type where each pin can be lubricated

65


Type 17. Mitre gates

Main application Advantages Storm surge barriers

Disadvantages

Unobstructed navigation passage Excellent from visual considerations

18. Vertically Storm surge hinged sector barriers gates

19. Drum and sector gates

Spillways

20. Bear-trap gates

Water control in logging rivers

Width of navigation more limited than some other gates Cannot accept reverse thrust Opening has to be effected when water levels are nearly equal Heavy mitre thrust Cannot close against high flow Requires wide recess Unobstructed in banks to navigation passage Can be opened or closed accommodate gates on opening against flow Can be opened at differential head Can be constructed with wider opening than mitre gates Excellent from visual considerations No outside source of Complex gates power required Requires extensive civil engineering works Requires zero downstream level Control system critical Can silt up Not favoured Seals critical No outside source of Control system critical power required Can silt up Clears debris Provides unobstructed Rarely used flow

Gates in submerged outlets 1. Vertical-lift intake gates servo-motor operated

66

Control and emergency closure

Reliable control gate Good load distribution in the slide version Damped

Gate slots required Load rollers or slides operate under water Requires stem connections between servo-motor and gate Possible cavitation problems Slow operation to raise to maintenance position Requires air admission

Type 2. Vertical-lift intake gates rope operated


Disadvantages

Bulkhead gate

Cannot be used as a control gate Cannot be used as an emergency closure gate Requires balanced head for operation Guide slots required Possible cavitation problems Requires bypass system Wide gate slots required 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 Requires chamber to retract High concentrated load Lintel seal critical Requires dewatering of tunnel to carry out maintenance Requires air admission Low natural frequency of vibration due to rope suspension and low friction Possible vibration problems Large gates require counterbalance to reduce hoisting forces Gate slots required Possible cavitation problems Requires bonnet for withdrawal Requires air admission

Can be load roller or slide gate Does not require air admission

3. Caterpillar or Control and coaster gates emergency closure

Control gate for very high heads

4. Radial intake Control and gates emergency closure Intake gate

Absence of gate slots Requires no load rollers or slides

5. Cylinder gates

Intake gate

Capable of controlling intake flow and large openings

6. Slide gates

Control gates in conduit Back-up gate for a control gate

Reliable control gate or emergency closure gate Inherently damped due to sliding friction

Types of gates

67


Type


Disadvantages

7. Jet-flow gates

Control gates in conduit for high head application

Gate slots required Requires bonnet for withdrawal Requires air admission

8. Radial gates

Control gates in conduit

9. Ringfollower gates

Back-up gate for terminal discharge gate or valve

Reliable control gate at high heads Inherently damped due to sliding friction Can be circular or rectangular Absence of gate slots, rollers or slides Requires lower hoist force compared with 6. and 7.

Can control flow Provides unobstructed flow Does not require transition section from circular to rectangular duct

Requires chamber to retract High concentrated load Lintel seal critical Requires dewatering of tunnel to carry out maintenance Requires air admission Large overall height approximately three times that of fluid way Requires regular flushing Drain connection 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 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.

Types of gates

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, hollowjet 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 nonlinear 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


Figure 3.1. Solid disc and through-flow butterfly valves

72

There are two types of butterfly valve: the solid-disc valve, sometimes 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.

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


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 servomotor 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 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: Kv

Valves

H V 2 =2g

where H V g

total head loss velocity of flow in the valve 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


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:

76

cavitation index

hd ÿ hv hu ÿ hd

or

hu ÿ hv hu ÿ hd

or where hd hv hu v

hu ÿ hv v2 =2g

Valves

head downstream of the valve vapour head at the inlet temperature head upstream of the valve 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: Vir or Vcr C1 Vr where Vir Vcr C1 Vr

hu ÿ hv 50

0:39

incipient cavitation critical cavitation a correction factor velocity Vir from Fig. 3.6 for incipient cavitation or the reference velocity Vcr from Fig. 3.6 for critical cavitation

C1 is taken from Fig. 3.7.

Table 3.1. Vapour pressure of pure water

Vapour pressure 2

Temperature C

N/m

0 5 10 15 20 25 30

610 870 1230 1700 2330 3160 4230

Head of water: mm 62 89 125 174 238 323 433

77


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 where Q Cd A

g H

3:1

discharge discharge coefficient (approximately 0.85) area of the valve based on the inside diameter of the valve body gravitational constant 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


Figure 3.8. Hollow-cone valve

feed energy into the elastic system. Possible modes of vibration for a hollowcone 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

Valves

Table 3.2 Values of C in equation (3.2)

N Ts/Tv C

4 1.00 2.22

5 1.00 2.35

6 0.50 1.98

6 0.90 2.40

6 1.00 2.48

6 1.20 2.53

Q=CDTv p Eg=e where Q C D Tv E g e K N

6 2.00 2.75

3:2 discharge a dimensionless coefficient depending on K and N valve diameter vane thickness Young's modulus gravitational constant mass per unit volume (of the material of the valve) ratio of shell thickness to vane thickness Ts/Tv 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 f C

Eg . Tv D2 e

3:3

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


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 hollowcone 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


Figure 3.13. Hollow-jet valve

84

Hollow-jet 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.

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 andcrosssection 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


Figure 3.15. Interiordifferential 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. Pressurereducing 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


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


downstream of the butterfly valve. If the section of pipe upstream of the valve 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

90

Type


Disadvantages

1.

Sluice valves

Controlling flow Low cost at low velocities Simple Closure and Reliable opening of flow

Unsupported valve blade during raising and lowering Eddy shedding from blade tip

2.

Butterfly valves

Closure device in Relatively low pressure conduit loss coefficient Available in large sizes Capable of working at high heads Closure by gravity can be arranged

Normally opened under balanced conditions Possibility of blade flutter Possibility of eddy shedding from blade tips

3.

Terminal Hollow-cone valves discharge (Howell^Bunger valves)

Very efficient energy dissipation Simple construction Relatively low cost Can be operated electromechanically or by oil hydraulics Good discharge coefficient Available in large sizes Least flow obstruction of any terminal discharge valve

Seal of sliding sleeve may leak Can trap debris but much less than 4., 5. and 6.

Type


Disadvantages

4.

Hollow-jet valves

Terminal discharge

Dissipates energy Can be arranged to discharge into a stilling basin at an angle

Less efficient energy dissipator than 4. Lower coefficient of discharge than 4. Greater cost than 4. Fluid passages can become blocked Internal moving parts Inspection and servicing requires removal of valve

5.

Needle valves

Terminal discharge

Dissipates energy Can be used as an in-line pressure reducing valve

Less efficient energy dissipator than 4. Low coefficient of discharge Greater cost than 4. Fluid passages can become blocked Internal moving parts Inspection and servicing requires removal of valve

6.

Pressurereducing valves (perforated cylinder type)

Pressure control in closed pipes

Pressure control

Orifices in perforated cylinder can be blocked by debris Internal moving parts Inspection and servicing requires removal of valve

7.

Sphere valves (rotary valves)

Shut-off control in high pressure conduits

Greater cost than 2. Low loss coefficient Shuts droptight Manufactured in large sizes Capable of working at high heads Can be supplied with a maintenance seal

Valves

91


92

References 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 flowinduced 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.

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


Fixed screen installations are used principally to protect intakes at concrete 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: ú

ú

Figure 4.2. Removable screen installation

94

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.

ú ú

Head loss across the screen. Vibration.


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: fn 2 where fn

r

EIg m mw L3

E I g m mw

L

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) Young's modulus moment of inertia of screen bar gravitational constant mass of screen bar added mass of water; that is the mass of water vibrating with the bar 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 or mw

m b 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 S V d

forcing frequency Strouhal number approach velocity thickness of screen bar 95


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: ú

ú ú

ú

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 screen should be restricted to measuring a limited range of head. The accuracy of 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 þ0.75% 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 buildup 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


Figure 4.3. Screen raking machinery

98

At flood diversion channels which are controlled by gates, and which incorporate a fixed weir alongside, a boom may be fixed across the flood channel 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


where i

Qni Rn

load factors that account for variability in the corresponding loads 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 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)

1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11.

12. 13. 14. 15.

102

Loading

Factor i or f

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

1.2 1.6 1.6 1.4 1.4 1.2 1.2 1.2 1.2 1.2 1.6 1.0 1.4 1.0 1.0 1.2

When considering the combination of loads under service conditions, the following are deemed not to act simultaneously (numbering corresponds with loading factors given in Table 5.1): ú

ú

ú

ú

ú


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


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


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


Figure 5.4. Wide span radial gate with four arms per side

106

Where the span of a radial gate is wide in relation to the height, the 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 counterbalanced 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

can also occur on the lower horizontal tie member of the gate arms of vertically stiffened skin plates. When a gate operates with the downstream side partially submerged the sheared flow inthestilling basinresults instrong recirculation which impinges onthedownstream 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


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


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


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: ÿ 1 qb2 =t 2 where q t

intensity of load (hydrostatic pressure), and plate thickness.

At the centre of the plate 2 qb2 =t 2 110


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 2

0.3078 0.1386 0.0138

0.3834 0.1794 0.0188

0.4356 0.2094 0.0226

0.4680 0.2286 0.0251

0.4872 0.2406 0.0267

0.4974 0.2472 0.0277

0.5000 0.2500 0.0284

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


Table 5.3. Variables in stress equations of plates (three edges fixed)

a/b

0.25

0.50

0.75

1.0

1.5

2.0

3.0

1 2 3

1

2

0.020 0.016 0.031 0.114 0.125

0.081 0.066 0.126 0.230 0.248

0.173 0.148 0.286 0.341 0.371

0.321 0.259 0.511 0.457 0.510

0.727 0.484 1.073 0.673 0.859

1.226 0.605 1.568 0.845 1.212

2.105 0.519 1.982 1.012 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 (Poisson's ratio 0.3)

L/B

V1

V2

1 20 10 6.67 5 4 3.33 2.86 2.5 2.22 2 1.67 1.43 1.25 1

1 0.984 0.938 0.867 0.783 0.697 0.616 0.545 0.484 0.433 0.391 0.325 0.276 0.241 0.195

1 0.861 0.753 0.660 0.580 0.512 0.453 0.404 0.363 0.324 0.295 0.250 0.215 0.190 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: e

p

x2 y2 ÿ 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 S

I d A

shear force static moment about the centroid of the section of part of the cross-section between the point concerned and the extreme fibres moment of inertia web plate thickness 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


114

3. US Army Corps of Engineers (2000): Design of Spillway Tainter Gates, EM1110-227023, 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.

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 servomotor, 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


In large radial gates, the arrangement of shafts connecting mechanical drives 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 fullload 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


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 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 servocontrols. 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.

Operating machinery

Figure 6.3. Hoist rope attachment to radial gates

119


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 flow

Operating machinery

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: ú

ú ú ú

ú

ú ú

ú ú

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:

ú

ú

ú ú

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


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 considerations 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: ú ú

ú

ú

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

123

Operating machinery

Figure 6.7. Hydraulic circuit for a power pack operating two radial gates


124 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 are: ú ú

ú

ú

Operating machinery

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-ofbalance cylinder forces. Slide gates and roller gates, when controlling flow or opening and closing under unbalanced hydraulic conditions, must be operated by a hydraulic servomotor. 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) (b) (c) (d)

Inlet strainers for the tank. Suction and delivery filters with 5 m apertures. Two manually operated pumps. Directional control valves which can be manually operated in addition to 125


(e) (f)

(g) (h)

electrical actuation by solenoids. If several valves have to be actuated 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. 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. 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. Flexible hoses which are rated 50% above the system working pressure. 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


adequate contact pressure or may extrude through the gap under the 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


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


site adjustment. Clamping of the seal mounting angle and the skin plate should 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


Figure 7.3. Sill seal for a high head, vertical-lift gate

The types of gate shown in Fig. 7.4 are: (a) (b) (c) (d) (e) (f) (g) (h) (i) (k) (l) (m)

vertical-lift gate radial gate submergible radial gate bottom-hinged flap gate mitre gate drum gate double leaf vertical-lift gate vertical-lift gate with flap bear-trap gate cylinder gate vertical-lift tunnel gate culvert valve (reverse tainter gate)

Figure 7.4. Gate boundaries requiring sealing

131


Sealing radial gates 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


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


The US Army Corps of Engineers' Manual `Design of spillway tainter gates' 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. Bottomhinged 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


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: ÿ0:28a 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


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


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


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


Operational condition

Contact surfaces

Gates (not frequently operated)

Gates and locks (frequently operated)

Rolling motions between nonhardened contact surfaces

Rail-wheel 1

1.85 B cylindrical 2.00 B

1.6 B cylindrical

Roller-wheel or roller-axle 2

1.9 B

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: above 300 to 2000 revolutions under load per year: above 2000 to 20,000 revolutions under load per year: above 20,000 revolutions under load per year:

by 10% by 15% by 30% 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 Ò' rings. For maximum reliability it is necessary to ensure that seals are lubricated. Some gate manufacturers provide separate grease passages to each seal. 139


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 à' 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


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


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, selflubricating 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


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: ú

ú

ú ú

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


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


Figure 7.17. Trunnion assembly with a spherical selfaligning, 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


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 backup 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. 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 10ÿ6 per C, the same as for all steels. With long falls, thermal elongation and shortening of ropes 147


Table 7.2. Apparent modulus of elasticity of wire ropes8

Type of rope 6 stranded ropes ^ fibre core simple construction (e.g. 6 7) 6 stranded ropes ^ steel core simple construction (e.g. 6 7) 6 stranded ropes ^ fibre core compound construction (e.g. 6 19, 6 36) 6 stranded ropes steel core compound construction (e.g. 6 19, 6 36) Multistrand non-rotating construction (e.g. 17 7)

E: N/mm2 103 62.0 68.7 49.0 59.0 42.0

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


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


150

References 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.

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


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


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 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.

Embedded parts

References 1. 2.

DIN 19704 (1976): Hydraulic steel structures; criteria for design and calculation. 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


Other hydraulic considerations complete the chapter: pressure coefficients 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 where Q Cd Go W g H

9:1

discharge coefficient of discharge gate opening (denoted b in Fig. 9.5) gate width gravitational constant 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 programmed into a computer for automatic control of gates or for calculating 158


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


Figure 9.3. US Corps of Engineers' chart for the coefficient of discharge under submerged flow conditions

160

that the experimental data show that different gate lip designs (even a minor modification) can result in a ÿ7% 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.

For radial gates on spillway crests, the discharge through a partially open gate can be computed using the same basic orifice equation: p Q CA 2gH where C A g H


9:2

coefficient of discharge area of opening gravitational constant 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 nonsubmerged 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


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


Figure 9.5. Coefficient of discharge map for free and submerged flow under a verticallift 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 ÿ y ÿ 0:85 F expW R gR where p g R Ho y F

9:3

mean pressure at the point examined on the gate surface water density gravitational constant radius of gate curvature heights of the energy head above the gate hinge vertical distance above gate hinge v2/(gR) where v mean horizontal velocity of flow over the gate edge 163


Figure 9.6. Diagram of water flow over a hinged gate

W

W W Y

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 ÿ y and ho, calculated from equations (9.4) and (9.5) (9.4) 1:22Y 0:7 for < =2 and < =2 (9.5) ÿ7:1F0:5 Y 0:4 in the remaining cases (9.6) Ho ÿ y=R=ho =R ÿ 1

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 L m g q L p

upstream head distance from hinge to the centre of gravity of the flap mass per unit width of flap gate gravitational constant discharge per unit width of flap gate depth of flap 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


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


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

h ÿ dd

where h dd

upstream head 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 ÿ1:15V p exp L g d where L V g d

9:7

loss of head in feet velocity of flow through the gate in feet per second gravitational constant 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


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

167


Figure 9.9. Diagram of tunnel gate under submerged flow conditions (after Naudascher et al.23)

F T ÿ B BdV 2 =2 where F T B B d V

9:8

downpull force downpull coefficient at the top of the gate downpull coefficient at the bottom of the gate width of gate depth of gate density of water 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 it is favourable from considerations of downpull force, although one model 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 verticallift 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


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 oversize the servo-motor accordingly.


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


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 number Fr 0^0.5. Intakes at existing stations and model studies within these 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 smÿ1/2. 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


others, although damage to stainless steel occurred downstream of the 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


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: s where hi hv v g

h i ÿ hv v2 =2g

9:10 head vapour head flow velocity 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


for a gate slot is: 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. The low pressure conditions on the downstream edge of the gate slot can be 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 where Hd C Hv


9:12 pressure difference from reference pressure pressure coefficient 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


given in Fig. 9.19 therefore show only the improvement which can be achieved 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 crosssectional 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


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


Proximity of two gates 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 ÿ 11:06 where Fr V g y

9:13

ratio of air flow to water p flow Froude number V= gy flow velocity at the vena contracta gravitational constant 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 well as entrance and exit air flow losses. These cause subatmospheric pressure 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 ÿ10.33 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 Q

air demand 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 ÿ 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


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, Karlsruhe, paper C5. 10. US Army Corps of Engineers: Tainter gates in open channels discharge coefficients (free 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


184

33. Ables, J H (1979): Vortex problem at intake Lower St Anthony Falls Lock and Dam, 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): SchuÞstrahlzusammenfÏhrung 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 flowinduced 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 fÏr Schwingungssichers TalsperrenverschlÏsse (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 BelÏftung des Abflusses in TiefauslÌssen (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.

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 shearlayer 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


The vibrating system The equation of motion for the simplest form of vibrating system with linear components is: m

d2 y dy c ky Ft 2 dt dt

where m y t c k F

10:1

mass displacement time damping (viscous) spring rigidity 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: m mw where cw kw F

d2 y dy c cw k kw y F 2 dt dt

10:2

added mass damping added mass rigidity 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

Figure 10.1. The vibrating system

186

10:4

Gate vibration

The system is unstable or negatively damped when: 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 approximation, added damping cw and added mass rigidity kw are neglected:

C p <1 2m mw k=m mw

10:7

Kolkman2 gives an explanation of added damping and added rigidity. The added mass coefficient Cm is given by: Cm

mw D2 L

where D

L

10:8 fluid density gate depth or characteristic body dimension (gate immersion) 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: fr

1 p k kw =m mw 2

10:9

187


The resonance frequency fr must be a factor higher than the excitation 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: TR

1 1 ÿ f =fr 2

10:10

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 L

V

excitation frequency 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) a representative flow velocity at the gate V

where He

p

2gHe

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 f

188

2gHe 7L

10:11

The vortex trail will spring from the upstream edge of a flat bottomed gate, 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.

Gate vibration

Figure 10.2. Strouhal number for the horizontal excitation of a culvert gate (after Abelev4)

189


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 where V L

10:12 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) length of conduit upstream from the gate

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

190

1 p gE=12s 2

where E

s

10:13

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) 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))

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


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) (ii) (iii) (iv)

192

Establish the total suspended mass (m). Establish the spring rigidity of the suspension system (K). Determine the added mass (mw). 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.

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


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 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

Gate vibration

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


unsuitability for this purpose. Diaphragm seals on bottom-hinged gates are also 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


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


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 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.

Gate vibration

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


The model of the Split Yard Creek control structure in Queensland, 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


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 TiefschÏtzen, 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

9. 10. 11. 12. 13. 14. 15. 16. 17. 18. 19. 20. 21. 22. 23. 24. 25. 26. 27. 28. 29. 30. 31. 32.

Naudascher, E and Rockwell, D (1994) Flow-induced vibrations, an engineering guide, A A Balkema, pp. 343^344. 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. Hart, E D; Hite, J E (1979): Barkley Dam gate vibrations, 19th I.A.H.R. Congress, Karlsruhe, paper C15. 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. Wendel, K (1950): Hydrodynamische Massen und Hydrodynamische Massentraggeheits-momente, Jahrbuch der Schiffsbautechnischer Gesellschaft, 44, 207^ 55. 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. Derunz, J A; Geers, T L (1978): Added mass computation by the boundary integral method, Int. Journ. Numerical Methods Engng., 12, 531^49. Hardwick, J D (1969): Periodic vibrations in model sluice gates, PhD thesis, Imperial College of Science and Technology, London. 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. 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. Kolkman, P A (1988): A simple scheme for calculating the added mass of hydraulic gates, Journ. Fluids and Struct., Vol. 2, pp. 339^353. Petrikat, K (1979): Seal vibration, 19th I.A.H.R. Congress, Karlsruhe, paper C14. Lewin, J (1983): Vibration of hydraulic gates, Journ. I.W.E.S., 37, 165^179. Krummet, R (1965): Swingungsverhalten von Verschlussorsganen im Stahlwasserbau, Forschung in Ingenieurwesen, Bd. 31, No. 5. Mitchell, W R (1979): Vibration due to leakage through a reverse radial gate, 19th I.A.H.R. Congress, Karlsruhe, paper C17. Nielson, F M; Pickett, E B (1979): Corps of Engineers experience with flow induced vibrations, 19th I.A.H.R. Congress, Karlsruhe, paper C3. 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. Lewin, J (1980): Hydraulic gates, Journ. I.W.E.S., 34, No. 2, p. 237. Bruce, B A; Crow, D A (1978): Hydroelastic model studies of the Pershore Mill sluice gates, B.H.R.A., report RR 1485, Jul. Merkle, T (1979): Hydraulically induced vibrations in a bear-trap weir, 19th I.A.H.R. Congress, Karlsruhe, paper C19. Schmidgall, T (1972): Spillway gate vibrations on Arkansas River Dams, Proc. A.S.C.E., Journ. Hydr. Div., HY1. Petrikat, K (1976): Structure vibrations of segment gates, 8th I.A.H.R. Congress, Leningrad. Murphy, T E (1963): Model and prototype observations of gate oscillations, 10th I.A.H.R. Congress, London. Pennino, B J (1981): Prediction of flow induced forces and vibration, Water Power and Dam Construction, Feb., p. 19. Vrijer, A (1979): Stability of vertically movable gates, 19th I.A.H.R. Congress, Karlsruhe, paper C5.

Gate vibration

205


206

33. Hampton, I G; Lesleighter, E J (1980): Effect of gate shape on closure loading, 7th 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.

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


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 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


type but can operate mechanically without external power, with electrical 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


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 òffset' 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


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 ÿ 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 A surface area of reservoir or reach of river


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 x1ÿx2 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


During a flood, the inflow will rise and subsequently fall. Any difference between the demanded level and the measured level will ú ú

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: ú

ú

ú

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


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: rate of fall in level

outflow rate ÿ inflow rate q 2 ÿ q1 A surface area of reservoir or reach

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


K1 n where n

q2 max 100

level tolerance.

Using the Kotmale reservoir7,8 as an example:

then

q2max A n K1

5560 m3/s (probable maximum flood) 6.63 km2 (surface area of the reservoir) say 20 mm (level tolerance) 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: ú ú ú

216

availability of mains supply availability of transformers condition and availability of standby generating plant

ú ú ú

gate availability aggregated faults in any one motor circuit or gate equipment 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: ú ú

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


ongoing training and the support of an effective management structure. 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 start-up on mains failure, is common. The generating set is either connected 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-enginedriven 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 dieselengine generator set

219


Figure 11.9. Standby diesel-enginedriven 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


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


in a subsequent failure to clear the fluidway. Uncontrolled opening of a spillway 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 òpen' 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 Rhoªne.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 policy at dam outlets, has been reviewed and tabulated by Combelles and 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 þ0.25%. 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


discharge for most valves is not constant throughout the range of valve 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 þ0.1%. 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 via the pneumatic tube and a service unit to the receiver, which applies a force 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 þ0.25%. 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


226

6. Di Stefano, J J; Stubbard, A R; Williams, I J (1976): Feedback and control systems, 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 Rhoªne River, France.

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 approximately 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


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


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. The pin connections of the ropes to the gates had seized, causing the wire rope to 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


Table 12.1. Summary of published information (not comprehensive) on major failures of spillway gate installations

Failure

Type of failure

Dam failure due to gates Malfunction of gates failing to open (7 failures) (fatalities)19 Power supply20 Power supply18 Power supply (fatalities)9 Vibration of outlet gate21 Power failure19, 22

Country Russia Romania India Spain India Romania

Structural failures of spillway gates (5 failures)

Trunnion friction7 Ice loading/brittle fracture23 Vibration4 Vibration12 Debris accumulation3

USA

Hoist failures

Damaged hoist Hoist failure/motor stalled Hoist rope failure/jammed gate Hoist chain failure Hoist chain failure Limit switch failure Hoist rope sheave seizure Hoist rope failure Hoist failure Hoist rope failure/gate jammed Hoist failure/gate jammed Brake failure/runaway gate Brake failure/runaway gate

Hungary USA South Africa

Overtopping of dam and erosion Condensation of electrical contacts Inadvertent opening

Spain

Ice formation Ice formation Ice formation

Canada Sweden Norway

Controls

Seals leakage

Russia Japan UK Sweden

Spain Canada Sweden Canada Canada Australia Pakistan Canada UK Sweden

Slovenia France

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


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 gates or valves

Silting of inlet Silting of inlet Cavitation Structural collapse of gate Structural collapse of gate Serious vibration (9 cases) Cavitation of discharge valve Cavitation of discharge valve Cavitation of discharge valves

Spain Romania Pakistan Romania Romania Romania New Zealand Australia Turkey

233


When the discharge from a tunnel gate does not result in supercritical flow, 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 ú ú ú ú

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

Radial gates Sector gates Vertical-lift gates, roller type Vertical-lift gates, sliding type Needles Stoplogs

234

Number of gates in survey

Fault frequency per 10 years

Fault frequency % per gate per year

362 107 590

235 125 770

6.5 10.0 13.1

2418

73

0.3

944 433

11 44

0.1 1.0

ú

ú ú

ú ú ú ú ú ú ú ú ú

Norway; however, such problems must exist in other parts of the world subject to severe winter weather (extreme environmental factors) 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


Figure 12.5. Sandkite collision with a pier of the Thames Barrier, October 1997

236

Even a relatively small gate opening will cause a surge wave to travel up a 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

order spares, as was the case at hydropower plants in a tropical country. Stockkeeping of spare parts for gates and auxiliary equipment is absent in many cases. 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


southern Africa are causing loss of key trained personnel in rural areas where 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: À 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, Òrchard', 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 questions which help to analyse the system. Examples are failure modes and 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


240

assessment based on the systematic application of engineering judgement are 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.

Reliability indices 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 10ÿ4 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: ú

ú ú


probability of not closing due to human or technical errors less than 10ÿ3 on demand probability of collapse less than 10ÿ6 in any year probability of not opening due to human or technical errors less than 10ÿ4 on demand. For the Seven Mile Dam in British Columbia, reliability analysis35 resulted in:

ú

ú

ú

probability of failure of spillway gates to open due to environmental hazards 9.68 10ÿ6 probability of failure of spillway gates to open due to electrical or mechanical failures 2.07 10ÿ7 probability of power supply unavailability to the spillway gates 2.07 10ÿ7

A good industrial system standard is one failure in 10ÿ4 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 10ÿ4 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 10ÿ4 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 10ÿ5 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


structural and mechanical plant which have some similarity. The available data 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 several gates, would permit the discharge of a flood less than the probable 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 steepsided 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 ÒR' gates (such as gates 3, 4, 5 and 20). The ÀND' 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


244

References 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 taintergate 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.

23. Freishist, A R; Rozina, I D; Rakhmanov, A L (1976): Operating experience gained with flat hydraulic gates under winter conditions. Gidrotekhnischeskoe Stroitelo'stvo, 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: ú ú

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: ú

ú

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: ú ú

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


by debris (timber, ice and other foreign objects) or lateral loading due to 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 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.

Ice formation

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 ú ú ú ú

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


To prevent ice from forming immediately upstream at the top of gates, air 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: ú

ú

ú

ú

ú

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, Ú (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), Gidrotekhnicheskoe 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 ú

ú

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


Hinks and Gosschalk3 quote examples of the amplification of earthquake 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


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 w H y

= = = =

hydrodynamic pressure at depth y from the water surface horizontal ground acceleration in units of g unit weight of water depth of water at the location of the structure depth from the water surface

The assumptions of the Westergaard approach are: ú

ú ú ú ú

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


face cannot apply. However, the Westergaard formula is extensively used when 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 þ10 Hz, so that a structure with a natural frequency much higher than this tends to respond as a rigid body. The 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 counterweight ^ which is nearly twice the mass of the gate ^ will occur. The wire ropes 255


will act as a spring due to their relatively low modulus of elasticity (see Chapter 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 which will act at the bolted connection of the servo-motor and the bonnet. 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: ú

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


ú

ú

ú

ú

ú

ú

ú

ú

ú

258

A different drive system or design of gate will throw up different problems 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 onsite 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

ú

ú

ú

subsequently be reset. This can be true of protection systems where a manual reset is often considered a reasonable precaution. Limit switches may be 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


Figure 14.3. Arrangement of spring-loaded transverse guide wheel of a tunnel gate to absorb an earthquake shock

260

An alternative design in which the piers and the sluiceway floor are 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.

Control buildings Control buildings must be designed to withstand earthquakes. Suspended light fittings should not be used and emergency light fittings should be selfcontained. 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


full spillway flow within one hour of earthquake'. In practice this may always be 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 Engineering, IV, No. 1, Feb. 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


Stainless steels are often welded to low carbon structural steels. To prevent 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 likely to occur at crevices, conjunctions of dissimilar metals, areas where debris 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


Most gates are assembled with some bolted connections. The mating surfaces 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 high current yield ensures maximum current distribution. The addition of 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 g d

flow velocity gravitational constant depth/length

Thus at a model scale of 1 in S: Flow Qm Qp =S2:5 Velocity Vm Vp =S0:5

271


Time Tm Tp =S0:5 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 timeconsuming. 273


Figure 16.2. Model of hydroelastic similarity of the gates for the storm surge barrier of the Eastern Scheldt

274

The scaling criteria which must be satisfied in constructing a hydroelastic 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

UK the rising sector gates of the Thames Barrier were the subject of 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:

Model studies

ÿ 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


276

8. Haszpra, O (1979): Modelling hydroelastic vibrations, Pitman Publishing. 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.

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


Highlands, fish ladders and fish lifts are used. Since the flow in fish passes is 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 upstream of gates will occur during most of the year when river flow is low. To reduce or eliminate this discharge, gate overflow is required. The bottomhinged 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: ú ú

ú ú

ú ú

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


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 clash with environmental considerations are difficult to achieve.


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 Bottomhinged 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


is effected by raising the lower leaf of the gate, causing discharge under the gate. 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 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


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 well as shackles and pins for hoisting attachments. Where ropes and chains are 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: ú ú ú ú ú

ú

ú ú ú

Maintenance and operation of gate installations

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


gearboxes should be regular to prevent the build-up of water and other 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 a e b c d GO TH TV TR

a g

gate width height of gate centre of pressure of hydrostatic thrust on the gate skin plate upstream water level downstream water level height of trunnion gate opening horizontal hydrostatic thrust on the gate skin plate vertical hydrostatic thrust on the gate skin plate resultant of horizontal and vertical hydrostatic thrust passing through the trunnions angle of TR with respect to the horizontal 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


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 e 1 3 2 Z1 Z2

horizontal hydrostatic thrust 12 b2 gW kN b/3 cosÿ1 d=R cosÿ1 d ÿ b=R 3 ÿ 1 R sin 1 R sin 3

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

290

Z3 Z2 ÿ Z1 area 1 A1 1=2Z3 b 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 ÿ R2 cos2 =2sin2 =2 vertical hydrostatic thrust A Tv p1 A2 gW kN TR resultant hydrostatic thrust TH2 Tv2 ) kN tanÿ1 TV =TH

Calculation of hydrostatic load on radial gates

Gate in the open position, downstream water level above gate lip (drowned discharge) The analysis assumes hydrostatic pressures. TH2 e2

resultant horizontal pressure centre of pressure of horizontal force on gate measured from sill

1 1 b ÿ Go 2 Wg ÿ c ÿ Go 2 Wg 2 2 1 2 Wgb ÿ Go ÿ c ÿ Go 2 2 1 b ÿ Go 1 c ÿ Go e2 ÿ Go TH2 b ÿ Go 2 Wg ÿ c ÿ Go 2 Wg 3 3 2 2 " # Wg b ÿ Go 3 c ÿ Go 3 ÿ e2 Go 2TH2 3 3

TH2

Z1 Z2 Z3 Z4 Z5

R sin 4 R sin 6 Z p2 ÿ2 Z1 R ÿ d ÿ c2 Z4 ÿ Z1

Figure A.3. Forces due to water pressure on a gate in the open position with drowned discharge

291


area 1 4

6

5

A1

Z3 Z5 b ÿ c 2

cosÿ1 d ÿ c=R ÿ1 d ÿ b ÿ1 d ÿ c ÿ b ÿ c cos cos R R 6 ÿ 4

area of sector ABC R2 5 =360 R2 5 5 5 ÿ R2 cos sin 360 2 2

area 2

A2

TV2 TR2

vertical hydrostatic thrust A kN p1 2 A2 gW 2 resultant hydrostatic thrust TH2 TV2 kN

tanÿ1

TV2 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 TH3

e3 TH3 e3

area 3

Figure A.4. Gate load due to water pressure under overflow conditions

292

1 2 1 b Wg ÿ b1 ÿ a2 Wg 2 1 2 b1 1 2 b1 ÿ a 1 2 b Wg ÿ b1 ÿ a Wg 3 2 1 3 2 Wg 3 b ÿ b1 ÿ a3 6TH3 1 A3 Z3 ÿ b1 ÿ a

TV3

vertical hydrostatic thrust A1 A2 A3 gW kN

TR3

resultant hydrostatic thrust

3

=

tanÿ1

For c > 0 TH4

e4TH4

p

2 2 TH3 TV3 kN

Calculation of hydrostatic load on radial gates

TV3 TH3

1 2 1 1 1 b1 Wg ÿ b1 ÿ a2 Wg ÿ c2 Wg TH3 ÿ c2 Wg 2 2 2 2 b1 1 2 b1 ÿ a 1 c 1 2 2 b Wg ÿ b1 ÿ a Wg ÿ c Wg 3 2 1 3 2 3 2

b31 Wg b1 ÿ a3 Wg c3 Wg ÿ ÿ 6 6 6

e4

Wg 3 b ÿ b1 ÿ a3 ÿ c3 6TH4 1

TV4 TR4

vertical hydrostatic thrust A 3 kN p1 2 A2 A 2 resultant hydrostatic thrust THA TVA kN

4

tanÿ1

TV4 TH4

The calculations are an overestimate which ignores the velocity head due to flow over the gate. 2 v To refine calculations b2 b1 ÿ 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 ú the correction is approximate ú the overflow is not curtailed ú 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 overflow head b1 ÿ a W gate width velocity of overflow V1 velocityphead hV Q 0:57W g b ÿ a3=2 V1

Q Wb ÿ a 293


Figure A.5. Gate load due to water pressure for a submergible gate

hV

V12 2g

effective hydrostatic head on gate b1 ÿ hV This value is now substituted for b1 A more accurate formula for Q is p 0:150b ÿ a Q 0:564 1 W g b ÿ a 0:0013=2 a For calculation of the approach velocity, see Ackers et al.2

Submergible gate ú

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, 285±286 added mass, gate vibration, 191±194 air demand, conduits, 180±182, 201 air supply pipes, submerged outlet gates, 51 articulated vertical lift gates, 20 automatic control methods, 209±216 automatic control systems, choosing, 217±218 automatic crest and scour gates, 16±17, 63 automatic tilting gates, 27, 29 Barking Creek Barrier, 41, 43 Barkley Dam, Kentucky, 189±190 barrier and barrage gates, 35±48 bottom-hinged buoyant gates, 39±41 bottom-hinged flap gates, 47 caisson gates, 48 environmental considerations, 281±282 flap gates, 41±44 hook-type double leaf gates, 47±48, 64 large span vertical±lift gates, 41 lock gates, 44±47 pointing gates, 48 rising sector gates, 35±39 tidal power barrages, 48, 64 battery powered standby systems, 220±221 bear-trap gates, 50±51, 66 bed protection, 46 Bondi scheme, 30±31 bottom-hinged buoyant gates, 39±41, 65 bottom-hinged flap gates, 23±30, 64 debris, 26±27, 97±99 disadvantage, 25 gravity standby systems, 222 nappe oscillation, 29±30 oil hydraulic operation, 120 seals, 134±137, 138 storm surge protection, 47 venting, 29±30 versions, 25 bottom outlet tunnel gates, earthquakes, 256±257

bottom outlet valves, 89 bottom outlets, failures, 232±234 bubbler devices, 224±225 butterfly valves, 71±78, 90 bottom outlet valves, 89 cavitation, 75±78 closure types, 73±75 loss coefficients, 75±76 seal arrangements, 71±73 caisson gates, 48 Cardiff Bay Barrage, 47±48 cascade controls, 209±210 caterpillar gates, 51, 53, 54, 67 cathodic protection, 268±269 cavitation conduits, 173±174 gate slots, 174±178 Tarbela Dam, Pakistan, 233 valves, 75±78 chains, 148±149 circuits, oil hydraulic operation, 122±126 circular-orifice gates, 56, 58 closed-loop systems, control, 211 coaster gates, 51, 53, 54, 67 composite construction, vertical-lift gates, 109±113 computer assisted control methods, 209, 210±216 computer program, radial gates, 14±15 conduits air demand, 180±182, 201 cavitation, 173±174 erosion, 173±174 gate conduits, 178±179 gate slots, 174±178 hydraulic considerations, gates, 172±182 proximity, two gates, 180 trajectory of jets, 179 vibration, 180 vorticity, 172±173, 188±190, 201 control and guard gates, submerged outlets, 53±59

295


control buildings, earthquakes and, 261 control systems and operation, 207±226 choosing, 217±218 control objectives, 207±208 failures, 235 fall-back systems, 218±223 instrumentation, 223±225 operating rules and systems, 209±216 standby facilities, 218±223 telemetry, 216±217 control weir, radial gates, 11 corrosion, steel, 266±268 counterbalance, radial gates, 13±14 cranes, emergency closure gates, 60, 62 culvert gates, vibration, 189 culvert valves, radial gates as, 8 culverts, screens, 96 cylinder gates, 52±53, 67 cylinder intake gates, 56, 67 damping, vibration, 186±187 debris, 93±99 bottom-hinged flap gates, 26±27, 97±99 environmental considerations, 278±279 floating booms, 99 overflow gates, 97 radial gates, 107, 108 underflow gates, 97 design criteria, gates, 101±103 design standards, 101±103 detail design aspects, 127±150 diesel generators, standby, 219±220 discharge characteristics fuse gates, 35 radial gates, 158±162 discharge measurement, 223±224 displacers/displacer chambers, radial gates, 11±13 downpull forces, hydraulic, 166±171 downstream offsets, gate slots, 177±178 downstream roller and turbulent flow, 197 downstream water level control, radial gates, 13, 14 drowned flow rectangular flap gates, 165±166, 168 stage±discharge relationship, 165±166 drum and sector gates, 48±50, 66 earthquakes analysis methods, 254±255 design loading, 102±103 effects on gates, 251±263 sample event tree, 261±262 spillway gate installations, 252±254 Eastern Scheldt Storm Surge Barrier, 274 elastomeric gate leaf, flap gates, 30±31 electrodes, water level measurement, 224

296

electromechanical drives, 116±120 embedded parts, 151±155 emergency closure gates, 59±62 empirical stage±discharge relationship, 166 environmental considerations, 277±283 equivalent-static analysis, 254±255 erosion, conduits, 173±174 excitation frequencies, gate vibration, 188±190 failure modes and effects analysis (FMEA), 239 failure modes, effects and criticality analyses (FMECA), 239 failures bottom outlets, 232±234 control systems, 235 fault frequency, gate type, 234±235 hollow-cone valves, 79±81 screens, 96 spillway gate installations, 227±232, 242±243 trunnion bearings, 142±144 fall-back systems, control systems and operation, 218±223 Fault Tree Manager, 238, 240 fish-belly flap gates, 24 fixed roller gates, 17±19 flap gates, 23±33 bottom-hinged see bottom-hinged flap gates elastomeric gate leaf, 30±31 storm surge protection, 41±44 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, 196±199 flow oscillation, 171±172 flow over gates, 162±164 flow reattachment, gate lip, 199±200 flow under gates, 158±162 Flowgate Projects, 16±17 FMEA see failure modes and effects analysis FMECA see failure modes, effects and criticality analyses Folsom dam, Sacramento, 142±144, 227, 229, 230 free flow, stage±discharge relationship, 164±165

free-rolling gates, 21±22, 23, 64 free shear layer, vibration, 200, 201±202 free surface flow gates, 4±51 hydraulic operation, 120±126 Froude scale models, 271±273 fuse gates, 33±35, 65 gate arms arrangement, conventional, 7 constructional features, 7±9 radial gates, 103±107 gate conduits, 178±179 gate design guidelines, 197±199 gate dogging device, access, 286 gate guide rollers, vertical-lift gates, 151±155 gate installations, maintenance and operation, 285±288 gate lips flow reattachment, 199±200 hydraulic downpull forces, 199±200 gate position measurement, 225 gate slots cavitation, 174±178 conduits, 174±178 downstream offsets, 177±178 gate vibration, 185±205 added mass, 191±194 damping, 186±187 excitation frequencies, 188±190 flow attachment, 196±199 flow reattachment, 199±200 hydraulic downpull forces, 199±200 preliminary check, 192±194 pressure waves, 190 seal leakage, 194±196 slack, gate components, 202 two-phase flow, 202±203 types, 185 unstable flow, 200±201 vibrating system, 186±188 vortex trails, 188±190, 201 gated river control structures, environmental considerations, 277±281 gates design criteria, 101±103 free surface flow, 4±51 hydraulic considerations, 157±184 structural considerations, 101±114 submerged outlets, 51±62 types, 3±69 `Gibb' gates, 221±222 gimbal mountings, hydraulic cylinders, 122 grappling beams, stoplog, 61, 62 gravity, 10, 12, 13, 115 standby systems, 221±222

guard valves, matching terminal discharge valves, 89 guide rollers, 137±140 guide wheels, preloading, vertical-lift gates, 203, 204

Index

hazard and operability study (HAZOP), 239, 240 hazard, hydraulic gates, 227±245 high head gates, 151±155 hoist rope attachments, radial gates, 119±120 hoist speed, 124, 126 hollow-cone valves, 78±83, 90 bottom outlet valves, 89 hoods, 82±83 point of flow attachment, shifting, 81±83 submerged, 82±83 vane failure, 79±81 hollow-jet valves, 84±85, 90 hook-type double leaf gates, 21, 47±48, 64 hook-type gates, 18, 20, 21, 64 Howell and Bunger valves see hollow-cone valves hydraulic considerations, gates, 157±184 conduits, 172±182 flow over gates, 162±164 flow under gates, 158±162 hydraulic downpull forces, 166±171 hysteresis effect, gate discharge, 172 limited ponded-up water, 171 reflux downstream, 171±172 stage±discharge relationship, 164±166 three-dimensional flow entry, sluiceways, 171, 172 hydraulic operation see oil hydraulic operation hydrostatic load calculation, radial gates, 289±294 hysteresis effect, gate discharge, 172 ice formation, 247±250 inlet system, radial gates, 10±11 inspection, gate installations, 286±288 instability causes, radial gates, 14 instrumentation, control systems, 223±225 intake gates radial, 55, 67 submerged outlets, 51±53, 66 jamming, radial gates, 107, 109 jet-flow gates, 56, 58, 68 King George V lock, storm surge protection, 41±44 Kotmale dam, Sri Lanka, 93 Kotri Barrage, Pakistan, 3, 41, 42

297


large span vertical±lift gates, 41 Legadadi Dam, Ethiopia, 222 level control, 210±211 limit switches, 145±147 limited ponded-up water, 171 lintel seals, leakage vibration, 196 load and resistance factor design (LRFD), 101±103 load factors, 101±103 load rollers, 137±140 local control methods, 209 lock gates, storm surge protection, 44±47 locomobiles, sector lock gates, 45, 47 LRFD see load and resistance factor design lubrication chains, 149 trunnions, 142±145 wire ropes, 148 machinery, operating, 115±126 maintenance, gate installations, 285±288 maintenance gates, 59±62 manual control methods, 209 materials and protection, 265±269 mitre gates, 44±47, 66 model studies, 271±276 motor drives, 116±120 Mrica Hydroelectric Project, Indonesia, 179 nappe oscillation, bottom±hinged flap gates, 29±30 navigation clearance, 41, 42±44 needle valves, 85±86, 86, 91 New Waterway, Rotterdam, 45, 46, 47 oil hydraulic operation advantages, 121 disadvantages, 121 features, 125±126 free surface flow gates, 120±126 hydraulic circuits, 122±126 leakage, 120, 125 operating machinery, 115±126 operating rules and systems, 209±216 operation, gate installations, 285±288 painting, protection and, 266±268 PD see proportional plus derivative control Pershore Mill gates, 196±197 PI see proportional plus integral control PID see proportional integral derivative control piers, dividing, radial gates, 14 PLC see programmable logic control point of flow attachment, shifting, 81±83, 196±199

298

pointing gates, 48 PRA see probabilistic risk analysis pressure methods, water level measurement, 224±225 pressure-reducing valves, 85±87, 87, 91 pressure waves, gate vibration, 190 probabilistic risk analysis (PRA), 239 programmable logic control (PLC), 218, 237 proportional control, 211 proportional integral derivative (PID) control, 210±216 proportional plus derivative (PD) control, 211±212 proportional plus integral (PI) control, 211±212 protection, materials and, 265±269 proximity switches, 147 pseudo-static analysis, 254±255 radial automatic gates, 9±15, 63 radial gates, 4±15, 56±58 advantages, 4 arms arrangement, 7 computer program, 14±15 control weir, 11 counterbalance, 13±14 as culvert valve, 8 debris, 107, 108 disadvantages, 4 discharge characteristics, 158±162 displacers/displacer chambers, 11±13 distribution of pressure head, 6 downstream water level control, 13, 14 electromechanical drives, 116±120 embedded parts, 151±155 float operated, 63 floats, 12 gate arms, 103±107 hoist rope attachments, 119±120 hydraulic forces, 5±6 hydrostatic load calculation, 289±294 inlet system, 10±11 instability causes, 14 jamming, 107, 109 malfunction causes, 14 motorised, 63 oil hydraulic operation, 120±126 piers, dividing, 14 roller faces, 151±155 seals, 132±134 side guide rollers, 137±140 side seal contacts, 151±155, 259 stiffening, 103±107 structural design, 103±107 trunnions, 13, 103±104, 142, 144±145, 146

two-phase flow, 202±203 types, 5 vibration, 189±190 water level control, 9±10 wide span, 106±107 radial intake gates, 55, 67 Randenigala Project, Sri Lanka, 179 rectangular flap gates, drowned flow, 165±166, 168 reflux downstream, 171±172 reliability, hydraulic gates, 227±245 reliability indices, 241±242 remote control methods, 209 reservoirs, control objectives, 207±208 ring-follower gates, 58±59, 68 Riprap bed protection, 46 rising sector gates, 35±39, 65, 66 risk assessment, 235±243 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, 151±155 rolling-weir gates, 22, 24, 64, 125 ropes, 119±120, 147±148 rotary limit switches, 145±146 rotary valves see sphere valves Rotterdam, storm surge protection, 45, 46 San Roque Dam, Philippines, 178±179 screens, 93±99 in culverts, 96 design criteria, 94±95 failures, 96 instrumentation, 96±97 raking, 97±98 in river courses, 96 vibration, 95±96 seals, 127±137 bottom-hinged flap gates, 134±137, 138 radial gates, 132±134 seal leakage, 194±196 seal shapes, 128 vertical-lift gates, 131, 134, 135, 136 Selsyn motor drives, 117±118 Severn Estuary, tidal power barrage, 48 Severn Tidal Project Study, 30±31, 48±50 side guide rollers, radial gates, 137±140 side sealing heating, 248 leakage vibration, 196 radial gates, 151±155, 259 sill beams, 151±155 sill seals, leakage, 195±196 slack, vibration and, 204 sliding gates, 48, 57, 67, 125 sliding paths, vertical-lift gates, 151±155

sluice valves, 71, 90 sluiceways, three-dimensional flow entry, 171, 172 sphere valves, 87±88, 91 spillway gate installations earthquakes, 252±254 events, 227±232 failures, 227±232 fault tree, 242±243 vibration, 228±230, 231 stage±discharge relationship, top-hinged flap gates, 164±166 standby facilities, 218±223 stiffening, 196±197, 198 radial gates, 103±107 vertical-lift gates, 107±113 still seals, embedded parts, 151±155 Stoney-roller gates, 17±19, 139 stoplogs, 59±62, 161 grappling beams, 61, 62 stoplog guide channels, 62 storm surge protection, 65, 66 flap gates, 41±44 lock gates, 44±47 stress analysis, 109±113 Strouhal numbers, gate vibration, 188±190 structural considerations, gates, 101±114 structural design radial gates, 103±107 vertical-lift gates, 107±109, 198 submerged outlets, gates, 51±62, 66 surge control, 215±216 systems analysis, earthquakes, 257±260

Index

Tarbela Dam, Pakistan, cavitation, 233 telemetry, control systems, 216±217 terminal discharge valves, matching guard valves, 89 terminology, notes on, 2 Thames Barrier, rising sector gate, 35±39 three-dimensional flow entry, sluiceways, 171, 172 three-dimensional models, 272±273 thyristor-controlled drives, 118±119 tidal power barrages, 48, 64 tilting gates, seals, 134±137 Tokyo City, storm surge protection, 44±45 top-hinged flap gates, 30±33, 64 adjustable pivot lugs, 32 elastomeric gate leaf, 30±31 hydraulically cushioned, 33 stage±discharge relationship, 164±166 Torrumbarry Weir, Australia, 189±190 trashracks see screens trunnions bearing failures, 142±144 lubrication, 142±145

299


mountings, 142, 144±145, 146 radial gates, 13, 103±104, 142, 144±145, 146 trunnion assembly, 140±142, 143, 144 tunnel gates, submerged flow conditions, 166, 168±169 tunnel lining sections, high head gates, 151±155 turbulent flow, 196±199 two-dimensional models, 272±273 two-phase flow gate vibration, 202±203 problems, 272 types of gates, 3±69 summary, 63±68 units, notes on, 2 unstable flow, gate vibration, 200±201 valve position measurement, 225 Venice Barrier gates, 39±41 venting bottom-hinged flap gates, 29±30 overflow gates, 201 see also air demand vertical-lift gates, 17±22, 64, 65, 67 advantages, 17 composite construction, 109±113 disadvantages, 17 embedded parts, 151±155 free-rolling gates, 21±22, 23 gate guide rollers, 151±155

300

guide wheels, preloading, 203, 204 load rollers, 138±140 long span, 18 overflow sections, 18, 20 reinforcing methods, 107, 110 seals, 131, 134, 135, 136 sliding paths, 151±155 stiffening, 107±113 structural design, 107±109, 198 vibration, 189±190 vertical-lift spillway gate installations, earthquakes, 255±256 vertically hinged sector lock gates, storm surge protection, 44±47, 66 vibration bottom-hinged flap gates, 29±30 conduits, 180 cylinder gates, 53 free shear layer, 200, 201±202 gate see gate vibration hollow-cone valves, 79±83 model studies, 273±275 screens, 95±96 seal leakage, 194±196 spillway gate installations, 228±230, 231 Victoria Dam, Sri Lanka, 54, 115, 209±210, 221±222 vortex trails, gate vibration, 188±190, 201 vorticity, conduits, 172±173 water level measurement, 223, 224±225

Jack Lewin-Hydraulic Gates and Valves in Free Surface Flow and Submerged Outlets-T. Telford (1995).pdf

Recommend Documents