Water Power and Dam Construction. Issue June 2010

Подождите немного. Документ загружается.

WWW.WATERPOWERMAGAZINE.COM JUNE 2010 41

PUMPED STORAGE

OVATION SYSTEM AND SCADA

The Ovation system incorporates state-of-the-art networks, con-

trollers, I/O, databases, and workstations with the latest software

to produce the most efcient control technology. Ovation provides

a multitude of options specialised for hydroelectric plants, includ-

ing hardware and software for local unit control, emergency hydro-

electric units, LV auxiliaries, metering systems, and HV substations.

Ovation provides state-of-the-art programming features, alarming

capabilities, engineering tools, and data collection within its standard

software applications.

The Emerson system manages 2,500 I/O points and monitors and

controls all major equipment and processes, including the Turboatom

hydraulic turbine, Electrotyazhmash pump turbine generator, lubri-

cation and cooling, drain and leakage systems, temperature monitor-

ing systems, auxiliary equipment and main electrical system. Ovation

also interfaces to the turbine governor, vibration system and excita-

tion system.

The Ovation system provides basic control schemes for each

unit, including the pump and turbine operations. The system also

provides all transition regimes as well as providing control for an

emergency shutdown.

The system management level consists of workstations for use by

the operators and executive and technical staff. There are also work-

stations for engineering management, system database and data his-

torian, and anti-virus protection.

The Ovation system also controls the 330kV substation. Unifying con-

trol under Ovation results in integrated control, which not only enhances

unit-wide compatibility, but also contributes to improved unit stability,

responsiveness and operating efciencies; tighter overall control of plant

operations; and a more streamlined view of key plant parameters.

Structurally, the Ovation system consists of operation level equip-

ment (network equipment and workstations) and control level equip-

ment (input/output modules and controllers). To provide staggered

commissioning of the various devices in the pumped storage plant,

the management system is separated into three independent subsys-

tems with individual controllers. These comprise a common station

level control subsystem, unit level control subsystem, and 330kV

switchgear control subsystem.

Each subsystem includes a dedicated Ovation controller, a local

user station and control panels with input/output modules. The

system uses a powerful Ovation controller, working in a multi-task

regime. The controller executes simple or complex modulating and

sequential control strategies, performs data acquisition functions,

and interfaces to various I/O sub-systems. Each controller has the

capability to originate up to 32,000 I/O points. Ovation optimises

multi-network communication trafc and the global management

system oversees the dedicated controllers and high-reliability periph-

eral input/output modules.

All components of the Ovation system have a redundant connection

to the main 100-Mbit/second Fast Ethernet network. This network

ensures that all data sources broadcast their data on the network and

the receivers (for example, user stations or other controllers) “see” all

available data from all these sources. Ovation’s SCADA architecture

provides tools to transfer important data between generating sites and

the UkrHydroEnergo (UHE) and National Dispatch Centres.

Workstations for operations, technical and executive staff provide

information on the current condition of the equipment and are designed

for direct process management. Engineers can use the stations for system

conguration, introduction of applied programmes, algorithm introduc-

tion, and changes to control sheets and workstation screens. They are

password protected to restrict access to authorised personnel only, and

anti-virus software is installed to protect against virus attacks.

The Ovation Process Historian provides mass storage and retrieval

of process data, alarms and operator actions for the control system.

Information on the status of all computer and intelligent eld devices

is stored in a database and the historian server allows up to 20,000

measurable and accounting parameters to be scanned and archived.

The historian is optimised to organise vast amounts of real-time proc-

ess data and present meaningful information to all users, including

operators, engineers, and maintenance personnel.

SUMMARY

“We have found Emerson to be a reliable partner,” said Yuri

Bondarenko, chairman of the board, Enpaselectro JSC, the prima-

ry contractor for supply and installation of a protection and con-

trol system for Dnestrovskaya Unit 1. “Based on their ability to

accomplish this comprehensive scope of work, combined with their

advanced technologies, we hope to continue our successful relation-

ship with Emerson.”

Using Emerson’s PlantWeb digital plant architecture with the

Ovation expert control system at Dnestrovskaya has created a simple

and easily understandable hardware and software platform for the

operator interface.

The use of controllers, information networks and operator station

functions, and applications within a truly distributed architecture, has

greatly increased the reliability of the control system, and allowed the

diagnostic systems to simplify system service as far as possible.

The pumped storage plant’s control system integrates all the tech-

nological processes of an electrical power station, provides staff with

accurate information, and guarantees continuous centralised access to

all technological parameters through the Ovation workstation. This

approach allows up-to-date and accurate decisions to be taken during

operation, and quick reaction to events arising within the boundaries

of any management mechanism.

For further information contact: Jackie Laurent, Emerson

Process Management. Jackie.Laurent@Emerson.com.

www.EmersonProcess.eu

Ovation operator workstation with graphic display showing Unit Control

Emerson’s Ovation system monitors and controls all major equipment and

processes at the Dnestrovskaya pumped storage power plant (PSPP)

IWP& DC

Walo Bertschinger AG

Dam and Landfill Construction

P.O. Box 1155, CH-8021 Zurich

Phone +41 44 745 23 11

Fax +41 44 740 31 40

asphalt.lining@walo.ch

www.walo.ch

Asphalt Lining: Rehabilitation of the Bituminous Sealing after 55 Years in Service

Brandstatt Compensation Reservoir, Austria

ZDO,QVHUDW%UDQGVWDWW''LQGG 8KU



603



448



1562

info@geokon.com

www.geokon.com

Vibrating Wire Piezometers

for Dam Monitoring

$Reliability: Excellent long-term stability

$Quality: ISO 9001:2008

$Accuracy: NIST traceable

$Excellent Customer Service: Contact us!

Why Geokon?

www.geokon.com/dams

WWW.WATERPOWERMAGAZINE.COM JUNE 2010 43

PROJECT DEVELOPMENT

ONSTRUCTION of the Bengoh Dam is progressing to

schedule. With the objective of securing sufcient water

supply for Kuching city and its surrounding areas, the

project comes at a time when there is a real risk that

present water resources are unable to meet the demands of a grow-

ing population, tourism and economic and agricultural develop-

ment. When completed, the 63.2m high structure will be the second

RCC dam in Malaysia after the Kinta dam near Ipoh, Perak.

Kuching is the state capital of Sarawak, Malaysia, located on the

north coast of western Borneo near the mouth of the Sarawak river,

where it forms a delta with the South China Sea.

In 2003, UK consultant Halcrow and Malaysia’s KTA (Sarawak)

Sdn Bhd were appointed by Malaysia’s Public Works Department

(JKR) to progress the development of Bengoh Dam. Initially they

were to undertake an assessment study, and subsequently develop the

outline and detailed designs of the structure – with the scope to meet

water demand requirements until 2030.

Tony Morison, an associate director with Halcrow, explains the

urgent need: “Kuching relies heavily on the Sarawak river. It’s an inte-

gral part and key feature of the city; provides water to a population of

over half a million people as well as to industry and agriculture; and

enables the movement of people and goods throughout the region.”

He continues: “But its natural resource is gradually being depleted

and without proper management, there may come a day when scar-

city of water will pose a problem for the state capital. Water supply

in Sarawak is a government function, managed by the Kuching Water

Board (KWB) and JKR. Bengoh Dam is a critical element of their

master plan that aims to help secure water supplies for the next 20

years. The population of Kuching has grown by leaps and bounds,

and it will not be long before the demand for water will outstretch

the stable natural supply during dry periods.”

Greater Kuching had a population of over 500,000 in 2003, when

water resource studies for the Bengoh Dam project were carried out,

and this is expected to grow to over one million by 2030. In addition,

increased tourism and planned industrial developments in Kuching

are expected to create additional requirements for treated water. The

water demand is growing fast, increasing at over 8% year on year, as

individual water consumption increases and more outlying areas are

connected to the treated water system.

Bengoh Dam will ensure that adequate raw water is supplied to the

Kuching Water Board’s Batu Kitang water treatment plant in times

of low river ows, to meet the water requirements of Kuching and

surrounding areas such as Samarahan, Asajaya, Sadong Jaya, Tasik

Biru-Lundu-Sematan, Tapah-Beratok-Siburan and Serian.

Construction of the dam started in 2007, at a site 40km south

of Kuching and about 2km from Kampung Bengoh accessible from

Kuching by bitumen road. The terrain provides a signicant reservoir

area with sufcient volume upstream for a relatively small dam.

Securing a major role

for Bengoh dam

Water demand is increasing year on year in the Malaysian city of Kuching. The fear is that

this will soon outstrip natural supplies during dry periods. In order to prevent this from

happening, Bengoh Dam is set to play a critical role in a master plan to help secure future

water supplies for Malaysia. With dam construction well underway, Tony Morison, Roland

K W Ling and Danien Ak Rangu give an insight into this important project

44 JUNE 2010 INTERNATIONAL WATER POWER & DAM CONSTRUCTION

PROJECT DEVELOPMENT

PROJECT STATUS

Tony Morison provides a project update: “At the present time the

RCC placing is complete to the underside of the spillway crest, with

the left abutment block still to complete. Work on the spillway crest,

crest bridge and equipment room, all of which will be conventional

mass or reinforced concrete, is in hand, for completion later this year.

Current works in progress include the concrete spillway side collector

channels and training walls; with the grout curtain being drilled from

the dam upstream heel.”

Morison goes on to add that drilling of the dam and foundation

drainage (from the dam galleries) has not yet started; whilst the

upstream intake shaft reinforced concrete is about 80% complete.

Valves and pipe work will be installed when the river diversion is

closed and the reservoir starts to impound.

“The critical path for lling the reservoir will be resettlement of the

upstream villages,” Morison says, “plus clearing of vegetation from

the reservoir area as required by the environmental mitigation plan.”

KUCHING MASTER PLAN

Between 1983 and 2000, a number of studies were carried out to

develop a master plan to ensure future water supplies for Kuching and

the surrounding area, with the aim of securing them beyond 2005.

The main water source is located at Batu Kitang. Abstractions here

have progressively increased and the capacity of the water treatment

works extended to meet increased demand, based on the natural

yield of the river. However, recognition was given to the fact that

this is a limited source. Alternatives considered to supplement the

yield included: additional supply from other river catchments; limited

local potential for groundwater development; provision of regulation

storage behind a weir in the Sg Sarawak Kiri river channel; develop-

ment of an emergency pumped intake on the Sg Sarawak Kanan to

supply the Batu Kitang WTW; and construction of a regulating res-

ervoir in the Sarawak Kiri catchment. (The study into this considered

a number of different sites and recommended the Sg Bengoh as a

suitable site for a river regulation storage reservoir.)

The ndings of studies into the above alternatives were incorpo-

rated into the Kuching Water Supply Master Plan Study. As a result,

the recommended plan to supplement water resources included the

development of a storage weir on the river at the Batu Kitang intake

by 2005, and construction of Bengoh dam by 2010.

DAM TYPE AND LOCATION OPTIONS

Options for RCC gravity, earthll, clay core rockll and concrete

face rockll dams were considered on four different dam sites within

the Bengoh gorge. Two of these sites were located where spillways,

outlets and diversions could be constructed at bends in the river,

more suited to embankment dams. The cost of the access road to the

dam was taken into account in assessing the options. The furthest

upstream site offered a lower and slightly cheaper dam. However,

this would incur signicantly higher access road costs plus some delay

while this was constructed.

The most cost-effective option at the site is an RCC gravity dam –

with a lowest foundation level 23m asl, the spillway crest level 80m

asl and dam crest level of 86.2m - giving a maximum dam height of

63.2m and a dam 267m long.

SUITABLE GEOLOGY

The gorge section of the Bengoh valley cuts through the lower Kayan

formation, consisting of interbedded and cross-bedded sedimentary

sandstones, pebble conglomerates and mudstones. Around the dam

sites these dip in an upstream direction at angles of between about 45

and 650 skewed about 450 from the selected dam axis. The thickness

of individual horizons varies from a few millimetres to tens of metres.

The sandstone is the hardest of the layers and forms the steep sections

of the river. The mudstone or soft shale layers tend to weather quickly

and control the sliding properties of the rock mass. The valley sides

are weathered to depths of typically 20m. Where unweathered, the

predominantly sandstone areas form a suitable foundation for most

types of dam.

Geological investigations consisted of surface mapping, borehole

investigations and laboratory testing at two sites, including some 12

boreholes to a depth of up to 80m at the selected dam site. Geological

assessments conrmed the suitability of the selected site for a concrete

gravity dam and provide data for the dam foundation design.

DESIGN CONSIDERATIONS

The Bengoh dam is designed as a gravity section to be stable under

normal, extreme oods and seismic loads. Seismic records show the

area to be of very low seismicity. A nominal 0.1g peak ground accel-

eration was adopted for the dam design but proved not to be the

critical design criterion.

The dam body has a vertical upstream face with the upper part at

a slope of 0.78:1, and the lower at 0.9:1 to reduce tensile loads at the

foundation contact.

The non-spillway section was made 8m wide for ease of RCC con-

struction. Stresses and deection of the section were checked by nite

element analysis using ANSYS. Consolidation grouting is specied at

the upstream heel and an inclined grout curtain is included to about

40m below foundation level beneath the upstream face of the dam.

A drainage curtain in the foundation and dam body is connected to

a system of galleries in the dam section.

The total volume of the dam and associated structures is about

160,000m

of mass concrete and RCC, with an additional 15,000m

of reinforced concrete in the diversion, outlet structures and spillway.

The selected aggregate for the RCC is Bau limestone. A new quarry

and aggregate plant were set up within 3km of the dam site to provide

and process the aggregate for dam construction. Trial mixes for the

RCC using cement and yash from sources in Kuching were used to

develop an RCC design mix containing 60kg of cement, 120kg of

yash, 0.6% of water reducing admixture and a water cementitious

ration of around 0.5. This has a modied VeBe time of between 8 and

WWW.WATERPOWERMAGAZINE.COM JUNE 2010 45

PROJECT DEVELOPMENT

Figure 8 – Part of embankment with

upstream wave protection being placed

15 seconds, little segregation and an average compressive strength

of about 35 MPa (N/mm

) at the specied 91 days. This is signi-

cantly more than the specied 15 MPa characteristic strength but has

not been changed. Reducing the cement content and increasing the

yash content to maintain the cementitious content provides little

cost saving and thermal aspects are within specication.

Initial set time is between 10 to 14 hours, and nal set between 24

and 36 hours, allowing most layers to be placed with hot or warm

joints. The maximum recorded temperature rise from the 30˚C speci-

ed maximum placing temperature is about 14˚C. The average RCC

density is around 2450kg/m

TOPOGRAPHIC SURVEY

Detailed topographic surveys were carried out for the dam and access

road sites. In addition, contour surveys at full supply level (FSL) and

maximum design ood level for the reservoir were carried out to

dene the reservoir area, together with a series of selected cross-sec-

tions within the reservoir area. This was done as the detailed sur-

veys of the dam site showed that the contours on the 1:10,000 maps

tended to be above actual ground level by over 10m in places. It is

thought that the published mapping was based on aerial photogra-

phy, and the difference reects the density of the vegetation cover.

The dam site and reservoir area surveys were done by ground survey

techniques and required cutting sight lines through the dense second-

ary vegetation.

The survey showed the FSL surface area of the reservoir at 80m asl

to be 10.2km

, an increase of about 15% on the surface area from

the map contours. The storage volume is thought to be increased

by more than this, although insufcient survey work was carried

out to conrm the exact volume. This will be done after clearance

of vegetation from the reservoir area. It was decided not to reduce

the FSL of the reservoir, but to leave the expected additional storage

volume as a margin for environmental releases and transmission

losses in the river between the dam site and Batu Kitang. FSL was

therefore maintained at 80m asl.

ENVIRONMENTAL REQUIREMENTS

An EIA was carried out by the University of Malaysia Sarawak

Reporting in March 2007. Environmental approval for the project

was given in February 2008, subject to conditions on:

r Dam design and safety.

r Soil erosion and sedimentation.

r Control of air and water pollution.

r Waste management.

r Protection of biological resources.

r Reservoir preparation.

r Socio-economic, health and safety aspects.

r Environmental monitoring.

r Audit and site clearance.

Biomass clearance and disposal and erosion control before reservoir

lling were particular requirements, as were the preparation of a

dambreak impact assessment and emergency response plan for the

area downstream of the dam.

RIVER DIVERSION

The river diversion as originally designed takes the form of twin box

culverts each 4m wide by 4.5m high, located on a rock shelf on the

dam left abutment. These can pass the 1:10 year ood of 365m

/sec

with an upstream cofferdam level of 42m asl, about 15m high. The

outlet pipework is to be installed in the diversion culverts when the

river diversion is closed.

During construction, a third similar opening was added in the dam

body to prevent backwater ooding of the existing villages (kam-

pongs) upstream of the dam, up to the 1:100 year ood once the dam

body construction rises above cofferdam level.

46 JUNE 2010 INTERNATIONAL WATER POWER & DAM CONSTRUCTION

PROJECT DEVELOPMENT

SPILLWAY AND OUTLETS

An overow spillway on the dam crest with a stepped downstream

face was adopted, designed to pass the probable maximum ood

(PMF) without overtopping the dam crest. The PMF peak inow

was determined as 2420m

/sec. Routing this over the 62m long crest,

the routed outow was 1800m

/sec with a ood rise of 5.8m. The

stepped spillway is expected to dissipate around 75% of the head at

the 1:10 year ow, reducing to about 45% at PMF.

The draw-off system is designed to discharge up to 21m

sec down

to a minimum operating level of 55m asl, from a multi-level intake

to the river downstream of the dam. The intake takes the form of a

dry well with inlets at ve levels feeding a single 1.6m diameter pipe

stack above one of the diversion culverts. The pipe discharges to

the downstream culvert through a hooded xed-jet cone (Howell-

Bunger) valve. Each inlet is controlled by a gate valve with provision

for an upstream maintenance gate. A lifting D-shaped trash screen

is placed in the gate track of the operating intakes.

In addition to the draw-off, a 1.6m bottom outlet is included in

the second diversion culvert, discharging with a similar arrangement

to the drawoff.

When acting together, the draw-off and bottom outlet are capable

of drawing down the reservoir by about 0.4m/day from FSL, and

References

PS Consultant, Esa Jurutera Perunding & Montgomery-Hosking Pty Ltd.

Kuching Water Supply Master Plan 1983.

Jurutera Jasa (Sarawak) Sdn Bhd & SMHB Sdn Bhd. Hydrological Study of Sg

Sarawak Kiri Catchment and Feasibility Study of a Submersible Weir across

the Sg Sarawak Kiri. 1996.

Jurutera Jasa (Sarawak) Sdn Bhd & SMHB Sdn Bhd. Feasibility Study for a

Multi-Purpose Dam on the Sg Sarawak Kiri. 1996.

Chin & Associates & GHD Pty Ltd. Study of Raw Water Sources for Kuching

City and Surrounding Area.1996.

Jurutera Jasa (Sarawak) Sdn Bhd & SMHB Sdn Bhd. Kuching Water Supply

Master Plan Study (1995-200). Final Report. Nov 1997.

PS Consultant & Montgomery Watson. Study on the Safe Yield of KWB’s Raw

Water Source at Batu Kitang. 2000.

KTA (Sarawak) Sdn Bhd. Assessment Study Final Report. Final Report. Dec 2003.

KTA (Sarawak) Sdn Bhd. Preliminary Design Report. April 2004

KTA (Sarawak) Sdn Bhd. Interim Design Report. Dec 2004

CTTC & UNIMAS. Environmental Impact Assessment. March 2007.

BHR Group. Bengoh Dam, Sarawak. Physical Model Study Final Report Oct 2008.

KTA (Sarawak) Sdn Bhd. Final Design Report. Draft. June 2009

draining 90% of the reservoir volume to minimum operating level in

45 days against the highest (January) average monthly inow.

SIGNIFICANT ISSUES

Resettlement was recognised as a signicant issue for the project and

is likely to be the controlling factor in the programme for reservoir

lling. Four villages (kampongs) were within the catchment area.

However, three of these would all lose low-level agricultural land to

the reservoir. The fourth kampong is remote from the reservoir and

largely unaffected.

Over 1100 people need to be resettled, which is the responsibility

of the Sarawak government. Extensive discussions and negotiations

on resettlement and compensation have been going on and a resettle-

ment site has been agreed on land in an adjacent catchment. Work is

in progress on construction of access roads and housing.

The authors are: Tony Morison, Associate Director,

Halcrow Group, UK; Roland K W Ling, Director, KTA

(Sarawak) Sdn Bhd, Sarawak, Malaysia; and

Danien Ak Rangu, Assistant Director, Jabatan Kerja

Raya, Sarawak, Malaysia.

IWP& DC

Progress updates

In August 2007 a turnkey contract for project implementation at Bengoh

dam was awarded by JKR to Naim Cendera Sdn Bhd (NCSB). KTA (Sarawak)

are providing engineering and construction supervision with assistance

from Halcrow.

Work started on site on the bridge and access road in September 2007.

NCSB awarded the dam construction contract to Sinohydro, and clearing

of the dam site started in March 2008. Construction of the river diversion

was delayed by unseasonal oods in August and September 2008, and

development of the quarry was delayed by a legal challenge over access

routes. However these problems have been successfully overcome and RCC

placement started in June 2009.

The dam construction was originally scheduled and more or less on track

to complete so as to start impounding by December 2010. This would have

required casting bottom outlet and draw-off pipework and valves into the river

diversion culverts (a dry-season activity).

The dam body RCC is already complete and nal concreting of the spillway

crest and bridge (which provides access for installation of the draw-

off valves in the intake) is in progress for completion by about August

2010. Valve and pipework installation, except in the diversion, should be

completed in the autumn.

Resettlement has now been agreed, but the delay in resettlement of the

upstream villages means that it is no longer possible to close the diversion

as originally scheduled. This means that there will be at least a one year,

and possibly longer, delay in closing the diversion to start impounding. This

is beyond the control of the dam construction team and no clear date has

been conrmed.

Unfortunately it means that there will be an enforced indeterminate break in

completion of the dam and testing of the operating equipment. The clearing of

vegetation from the reservoir area has likewise been delayed, as there is no

point doing this more than one year before impounding.

7RJHWKHUZHFDQRIIHULQYHVWRUVFRPSOHWHVROXWLRQVUDQJLQJIURP

WKHGHYHORSPHQWDQGGHVLJQWRWKHGHOLYHU\RIK\GURHOHFWULF

SRZHUSODQWHTXLSPHQWRQDWXUQNH\EDVLV

6WURQJHUWRJHWKHU

(XURSHDQUHOLDELOLW\:RUOGZLGHÁH[LELOLW\

ZZZOLWRVWURMSRZHUHX

/,726752-32:(5²/LWRVWURM3RZHU

/LWRVWURMVNDFHVWD

/MXEOMDQD

6ORYHQLD

7HO

)D[

(PDLOLQIR#OLWRVWURMSRZHUHX

ZZZOLWRVWURMSRZHUHX

/,726752-32:(5²/LWRVWURM+\GUR

3DFLÀTXH(VW

%URPRQW4XHEHF

&DQDGD-/-

7HO

)D[

(PDLOLQIR#OLWRK\GURFD

/,726752-32:(5þ.'%ODQVNR(QJLQHHULQJ

þDSNRYD

%ODQVNR

7KH&]HFK5HSXEOLF

7HO

)D[

(PDLOLQIR#FEHQJF]

ZZZFEHHX

/,726752-32:(5+\GUDXOLF5HVHDUFK&HQWHU%ODQVNR

+RUQt/KRWD

%ODQVNR

7KH&]HFK5HSXEOLF

7HO

)D[

(PDLOUDGRPLUKDYOLFHN#KUFEODQVNRF]

/LWRVWURMB,PDJHB2JODVB[B3UHVVLQGG 

Rehabilitation of Hydro Power Plants

Our customer is an international company active in the rehabilitation and modernization of hydro power plants. Its high

professionalism and know-how is acknowledged in many countries in Europe, Asia, South America, as well as in Africa.

Due to promotion of the current holder of this position, we are now seeking a new Head of Sales Department.

Leading a team of about ten based in Switzerland, the successful candidate will be in charge of project sales worldwide.

Working closely with top management, you will develop marketing and sales strategies at Country, Customer, and Project

levels. On a nancial level, you will be responsible for controlling costs and risks and will be accountable for the gross

Margin. With good listening skills, the Head of Sales will liaise with colleagues internally and externally throughout the

whole sales cycle from the creation of the oer up to the negotiation and signature of the contract. You will actively

contribute to market development through frequent contact with customers as well as involvement in workshops and

seminars.

We are seeking a strong leader with a technical background as an Engineer and a successful track record in the sales of large

investment projects at international level. Beside strong sales skills, a rigorous approach to the management of costs, risks

and issues is required. Previous experience in the eld of Hydro power plants will provide candidates with a good under-

standing of such complex projects as well as a good knowledge of this specic market. A good communicator, you will be

uent in English and French and have a good understanding of German. Some command of Spanish would be also an asset.

Should you require additional information, please call us or send us your dossier. Your application will be treated with the

utmost discretion.

The contact person is Michel Abellan, address mail abellan@ems.ch

Avenue d’Ouchy 18 · CH-1006 Lausanne · Phone +41 (0)21.613.70.00

· emslausanne@ems.ch · www.ems.ch

Basel · Bern · Lausanne · Zürich

48 JUNE 2010 INTERNATIONAL WATER POWER & DAM CONSTRUCTION

PROJECT DEVELOPMENT

OLLOWING the discovery that FEH (IOH, 1999) 1 in

10000 year rainfall exceeded FSR (NERC, 1975) probable

maximum precipitation (PMP) at certain sites in the UK

(MacDonald and Scott, 2000; Babtie Group, 2000), a new

assessment of extreme rainfall has recently been published (Defra, In

Press). The new results show a reduction in 24 hour 1 in 10000 year

rainfall of up to nearly 200mm at some sites. For 6-hour 1 in 10000

year rainfall the reductions are often in excess of 150mm. Smaller

reductions are suggested for 1000 year and 100 year rainfall depths.

In this paper I report on several concerns about the revised meth-

odology used in the Defra (In Press) report, and provide a specic

test of the revised rainfall estimates. The results show that the depth

of ood producing rainfall has been substantially underestimated.

Furthermore, several notable rainfall records have been omitted from

the analysis, including the largest recorded rainfall in the UK. Another

problem is the inadequate raingauge network in certain areas, a short-

coming that has not been compensated for in the analysis.

New estimates of PMP are not presented in the Defra report, which

could, given a suitable rarity, have allowed the merging of T-year

rainfall estimates with estimates of PMP. I also present a review of

estimates of PMP and how they can be merged with standard rainfall

frequency analysis. The result is then applied in the estimation of

ood rarity for seven well documented extreme oods in the UK, via

a joint probability model based upon effective rainfall (rainfall–soil

moisture decit, SMD) and SMD. The implications for dam safety are

described. The revised estimates of 2–10

year rainfall will form part

of a New Guide to Flood Estimation in England and Wales.

THE STA TION YEAR METHOD OF RAINFALL

ANALYSIS AND THE ERGODIC HYPOTHESIS

The agglomeration of rainfall records from either all raingauge sites

(NERC, 1975), or those gauges that are assessed as being independ-

ent, (Dales and Reed, 1989), are examples of sampling in space as a

substitute for sampling in time. For example 50 independent records

each with 30 years’ observations produces 1500 station years’ of data.

With a large data set the possibility of making estimates of rainfall

with high return periods becomes attractive. Substitution of space

for time is a variation on the ergodic hypothesis – a term borrowed

from physical mechanics. It is a potentially very powerful way of

providing answers to difcult questions especially where long time

periods are involved. For example the degradation of a marine cliff in

South Wales through time that was progressively protected from ero-

sion by an advancing sandbar, was demonstrated by Savigear (1952).

Similarly, Ruhe (1952) showed the evolution of a drainage network

in central US because it was possible to date the age of different areas

since the ice retreated, based on the age of till deposits over the last

50000 years: as time progressed drainage density increased rapidly

at rst and then slowed down. As the ice retreated from Glacier Bay

in Alaska during the 19th century, changes in the soil and vegetation

were identied by Crocker and Major (1955). In all of these studies

the time dimension is clearly apparent, although the inuence of dif-

ferences in solid and drift geology should be taken into account.

In the case of rainfall analysis the time dimension is the time span

over which the records were gathered. This length of time is about

100 years: any extension of the estimates of rainfall to higher return

periods cannot be made unless a method of extrapolation is used or

estimates of the upper limit of rainfall depth – PMP – are made. The

latter approach also requires an estimate of the rarity of PMP. Thus the

ergodic hypothesis cannot be applied to rainfall frequency analysis. In

none of Emil Gumbel’s papers (Gumbel, 1941a, b) or his classic mono-

graph (Gumbel, 1958) are predictions made of events whose return

period exceeds 1000 years. But the problem facing the dam engineer is

that events can and do take place whose rarity is much greater. Hence

the problem reduces to producing a method that gives realistic esti-

mates of rainfall depth up to the probable maximum. Before this is

considered further the question of missing data must be considered.

OMISSION OF HISTORIC STORM DATA

One of the objectives of the Defra (In Press) report was to assemble

all the available data on extreme events since this would give the best

chance of estimating the depth of very rare storms. As part of this

process (Svensson et al, 2009) have produced a useful list of events,

noting at the same time that ve events have exceeded FSR PMP.

These are listed in Table 1.

The lower part of Table 1 shows the corrected depth and duration

where appropriate for three of the events. In the case of the Bruton

storm the contemporary newspaper reports all suggest a storm dura-

tion of about 12 hours; the Hardy Monument reading was collected

by Ian Forbes (Clark, 2005a); while the Horncastle event was a simple

mistake in copying the storm depth at Horncastle whose duration was

Rain, ﬂoods, and dam safety

A new assessment of extreme rainfall has recently been published by DEFRA in the UK. In this

paper, Colin Clark raises some concerns about the revised methodology used in the report

Land

Raingauges

0 50 100 150

100

Height (m)

% land and raingauges

Figure 1. Percentage distribution of land

elevation and raingauge elevation

WWW.WATERPOWERMAGAZINE.COM JUNE 2010 49

PROJECT DEVELOPMENT

six hours and the lower depth at Ravesby Reservoir whose duration

was 2.8 hours. More serious is the complete omission of observations

made during seven other events. Table 2 gives the relevant data. The

most important and best documented storm event in Table 2 is the

Martinstown event (Bleasdale, 1956; Clark, 2005a).

THE NATIONAL RAINGAUGE NETWORK AND A

COMPENSATION TECHNIQUE

In spite of the dense raingauge network of about 4000 stations in the

UK, there are areas that do not have the recommended raingauge

density as per Shaw (1989), and Raghunath, (2006) of one gauge

per 26km

-2

and one gauge per 37km

respectively. For example in

an area of 971km

in East Yorkshire only 6% of the gauges cover

16% of the land area (Figure 1). This is equivalent to a gauge density

of one per 156km

. This particular area is above 100m OD where

the highest storm rainfall is likely to occur. A good example of this is

the Louth storm of 29 May 1920 (Arellano and Clark, 2004), where

Figure 2 shows the depth area relations for this storm, a result that

implies a rainfall of 232mm covering an area of 10km

. If this took

place in about three hours then the new FEH rainfall estimates in

Defra (In Press) give it a return period in excess of 10

years.

In order to compensate for an inadequate raingauge network in

India Rakhecha and Clark (1999) carried out depth area analysis to

estimate rainfall at a scale of 10km

. The results were then applied

in the WMO (1986) moisture maximisation procedure. This method

is one of several international standards for estimating PMP. The

resulting estimates of PMP (Rakhecha and Clark, 1999, 2000, 2002)

have been criticised as being too high (Dhar and Nandargi, 2000).

However, the previous standard estimates of PMP (IITM, 1989) were

exceeded by 24cm during a storm over western India in 2005 (Kumar,

et. al. 2008), so that the more recent estimates of PMP are now being

used for dam safety assessments in Gujarat (RMSI, personal commu-

nication). These ndings suggest that depth area analysis can help to

make up for a lack of raingauges in certain circumstances, especially

where the gauge density is less than 30km

gauge

-1

The likely underestimation of maximum rainfall leads us to a con-

sideration of estimates of PMP, together with a simple procedure for

merging the results with T-year rainfall estimates.

PROBABLE MAXIMUM PRECIPITATION

Although the concept of PMP has fallen out of favour with some

dam engineers in the UK (Brown and Gosden, 2004), this should not

detract from the two near miss dam incidents at Ulley and Boltby in

recent years. The event at Ulley was only caused by about 100mm

rainfall in just below 24 hours, while the Boltby incident was caused

by rainfall well below estimates of PMP (Clark, 2008a) for this area.

Therefore it was with some surprise that the Defra (In Press) report

did not include a reassessment of PMP, especially as the confusion

with the FEH 1 in 10000 year rainfall was partly as a result of PMP

not being reassessed, and more particularly as some preliminary work

on PMP for the Defra project had already been completed (Dent,

2008). The greatest uncertainty surrounds the depth of the greatest

rainfall that can be reasonably be expected, hence the term probable

maximum precipitation. Therefore as many methods and as much

evidence as possible should be applied in order to make a realistic

estimate of PMP. Similarly: “It was agreed that a combination of

techniques (storm maximisation, frequency analysis, hydrodynamic

modelling), and the use of local data are required to minimise uncer-

tainty.” (Svensson, et. al. 2005).

To date the present author has used the following methods and

evidence to estimate PMP.

Y = -109.81Log x +342.14

10 100 1000

200

150

100

Area (km2)

Rainfall (mm)

Table 1: Rainfall events deemed to have

exceeded FSR PMP

Date Rainfall (mm) Duration (hours) Location

28/6/1917* 243 8 Bruton, Somerset

18/8/1924 238 5 Cannington, Somerset

18/7/1955* 280 15 Martinstown, Dorset

7/10/1960* 184 3 Horncastle, Lincolnshire

19/5/1989 193 2 Halifax, Yorkshire

Corrections to the above made in this paper*

28/6/1917 243 ~12 Bruton, Somerset

18/7/1955 356 15 Hardy Mon. Dorset

7/10/1960 159 2.8 Revesby R. Lincolnshire

Table 2: Storm data omitted in Svensson et al 2009

Date Rainfall Duration area Location

(mm) (hours) (km

)

11/6/1682* +500 0.25 <0.1 Oxford

3/7/1893* +250 2 12 Cheviot Hills

26/8/1912 209 22 1 Sprowston, Norfolk

17/9/1913 163 14 30 Doncaster, Yorkshire

3/11/1931 244 36 1 Trecastle, Brecon

29/5/1920 153 3-3.5 55 Hallington, Lincolnshire

18/7/1955 356** 15 6 Hardy Mon. Dorset

18/7/1955 343 15 10 Goldcombe Fm. Dorset

18/7/1955 305 15 20 Winterbourne Steepleton,Dorset

* Harrison, (1682), date is New Style Calendar. Not listed in British Rainfall.

* Clark, (2005b) ** Also noted in Table 1

Figure 2. Depth area relationship for

the Louth storm, 29 May 1920

50 JUNE 2010 INTERNATIONAL WATER POWER & DAM CONSTRUCTION

PROJECT DEVELOPMENT

r 1. Depth duration data from SW England (Clark, 1995), yielding

24-hour PMP of about 500mm.

r 2. Endpoint of sedimentation equation applied to rainfall data in

SW England. This gave estimates of 24 hour PMP ranging from

400-550mm.

r 3. Storm maximisation using the WMO (1986) technique for his-

toric storms, which gave estimates between 292mm for Inverness

and 594mm for Seathwaite (Clark, 2002a). The same technique

has been applied to the Louth storm of 1920 (Arellano and Clark,

2004) and the Cheviot Hills storm of 1893 (Clark, 2005b).

r 4. Frequency analysis of peaks over a threshold data using the modi-

ed Gumbel scale of Rakhecha and Clark (1999). Britain was divid-

ed into three regions: south, midlands, and Scotland. The results

varied according to annual average rainfall as well as into three

partially overlapping groups, from about 300mm in the drier parts

of Scotland, increasing to about 375mm in wetter areas. For the

middle and southern areas the estimate of 24 hour PMP varied from

about 320mm in drier areas to 590mm where the annual average

rainfall was in excess of 2000mm.

r 5. An analysis of worldwide estimates of PMP in relation to lat-

itude (Clark, 1997) showed that in the southern part of Britain

one-day PMP is about 500mm. It was later pointed out (Hardaker,

personal communication) that some African countries have much

lower PMP than what latitude alone would predict. This point is

accepted and the simple relationship between latitude and PMP is

to be regarded as an upper limit when atmospheric moisture is not

a limiting factor.

r 6. Estimates of very extreme rainfall, possibly getting close to PMP

have been obtained from reverse engineering of some historic oods

(Clark, 1999, 2003). The results for the 1768 ood at Bruton sug-

gest a 1 hour PMP in the range 136-190mm, while the 1898 ood

at Windemere suggested a depth of about 260mm in 24 hours over

a catchment area of 187km

. By using the FEH (IOH, 1999) area

reduction factor the point value was estimated as about 315mm.

These six approaches have produced estimates that are considerably

higher than those in the FSR/FEH. The implication from the Defra (In

Press) study is that the FSR (NERC, 1975) estimates of PMP should

remain. This is in spite of a recommendation (KBR, 2003) that a new

study of PMP be made, following the report to Defra by Sir David

Cox (Cox, 2003). Furthermore, Dent (2008) produced estimates of

PMP based on historic storms and the WMO method that are in line

with those described above, and in some cases are higher.

Estimates of the return period of PMP vary from about 1 in 60,000

years (Koutsoyiannis, 2004); to 1 in 250,000 years, (Austin et. al.

1995); and 1 in 800,000 years, (Morris, 2007). From an analysis

of historic oods on the upper Brue in Somerset (Clark, 1999), and

frequency analyses of one-day Indian rainfall (Rakhecha & Clark,

1999), a return period of 10

years has been chosen because in both

cases it was consistent with estimates of historic oods. If a value of

around 8 x 10

years had been chosen as suggested by Morris (2007)

then very little change in the estimate of PMP or indeed PMF would

be produced in both cases.

In order to obtain a 2-year to PMP rainfall frequency curve the esti-

mate of the 2-year D-hour rainfall can be obtained from local data for

a given site. Alternatively, the FEH CD-ROM can be used since the

estimates are for a common event and many rainguage stations were

used in their estimation. On the modied Gumbel scale the 2-year

event y = 0.1891. For the PMP event taken to have a rarity of 10

years, y = 9.382. Using the Log of rainfall depth, R, and the values of

y, a regression equation which links frequency and rainfall depth is

established. For example if the 2-year 24 hour rainfall is 40mm and

the 24 hour PMP is 475mm then the rainfall frequency equation:

Log R = 0.11689y + 1.5799 (1)

Where y = [( -ln ln (1 – 1/T - 3.3842) * 1.09348 * T

–0.046518

] + 3.3842.

This gives the depth of rainfall from a return period 2- 10

years.

RAINFALL FREQUENCY AND FLOOD FREQUENCY

Because of the uncertainty of the estimates of extreme rainfall, it is

essential that the results be tested for their credibility. Defra (In Press)

do not provide any robust test; this may be because of a perceived

lack of a standard against which the results can be assessed, but more

likely because of a lack of focus on the end result, namely extreme

oods. One method that was used in a general assessment of Indian

PMP was to use the existing estimates (IITM, 1989) in order to esti-

mate the probable maximum ood (Rakhecha and Clark, 1999). For

sites that have a well-documented ood history, the rarity of the ood

resulting from a measured storm can be estimated and then compared

with the result from the ood frequency curve.

-2.0 -1.0 0.0 1.0 2.0 3.0 4.0 5.0 6.0 7.0 8.0 9.0

1.01 1.1 1.5 2 5 10 20 50 100 500 10

Discharge (cumecs)

Return period (years)













Return period of ﬂood (years)

10 100 1000 10000 100000

(a)

Ratio of Column 5:4 (Table 3)

Return period of ﬂood (years)

10 100 1000 10000 100000

(b)

Ratio of Column 6:4

(Table 3)

Figure 4. Changes in the ratio of estimated ﬂood return period to recorded return

period of ﬂoods using: (a) Defra (In Press) rainfall; (b) New Guide rainfall

Figure 3. Estimates of ﬂood frequency for the Withycombe Brook, Exmouth, UK