2003 ancold - spillway gate reliability - barker et al (2)

8/12/2019 2003 ANCOLD - Spillway Gate Reliability - BARKER Et Al (2)

1/15

Spillway Gate Reliability and Handling of Risk for Radial and Drum Gates

NZSOLD/ANCOLD 2003 Conference on Dams Page 1

SPILLWAY GATE RELIABILITY AND HANDLING OF RISK FOR

RADIAL AND DRUM GATES

M. Barker1, B. Vivian2, J. Matthews3and P. Oliver4

ABSTRACTThis paper discusses reliability issues of the fourteen 3.85m high by 7.89m wide radial gates atGlenmaggie Dam in Victoria and the twin 3.6m high by 16.5m wide drum gates at Little Nerang Damin Queensland. The Glenmaggie dam radial gates are manually controlled using electrically driven

(mains and diesel generator power supply) hoist motors with a petrol driven hydraulic pack for use inthe event of complete electrical power supply failure. A detailed fault tree analysis was developed forthe spillway gate reliability of the Glenmaggie Dam gates as part of the risk assessment for the dam,

which was being completed at the time of publishing the paper. Each of the identified components ofthe spillway gates, including human error in operation was used to evaluate the probability of failure

of a single gate or multiple gates for inclusion in the event tree to estimate the risk and assist theevaluation of the requirement for remedial works. The Little Nerang drum gates are fully automatic

hydraulically operated gates with independent operating mechanics and a common override system inthe event of automatic system failure. Drum gates are uncommon on dams and the system operation is

discussed together with an assessment of the reliability and measures taken for handling operatingrisks during floods for the dam, which has some stability concerns.

Key Words: Spillway, Gates, Fault Tree, Failure, Sensitivity, Human Response

1. INTRODUCTION

1.1 Glenmaggie Dam

The Glenmaggie Dam is a concrete gravity

dam with a maximum height of 37m. The damis owned and operated by Southern RuralWater Authority of Victoria and is usedprimarily for irrigation water supply.

A Design Review of Glenmaggie Dam carriedout in 1999 (SMEC 1999) indicated thatGlenmaggie Dam:

withstands the PMP design flood (SatisfiesANCOLD Criteria);

cracks right through at the dam-foundation

interface under the Maximum DesignEarthquake (1:10,000 AEP) and so wasjudged as not passing the ANCOLD criteriafor earthquake resistance.

Subsequently, Southern Rural Water (SRW)commissioned Gutteridge Haskins & DaveyPty Ltd (GHD) to carry out a staged review ofGlenmaggie Dam, which included thefollowing:

Part 1 A Seismic Capacity Evaluation usingtime history with updated seismicground motion data for the SeismologyResearch Centre. This part of the studyfound that despite sustaining damage,

the dam withstood the MDE.Part 2 A limited scope risk assessment to

review the overall risk profile of thedam under earthquake and floodloading to confirm the level of risk.

A structural analysis was carried out for floodand seismic loading of the spillway gates forinclusion in the Part 2 risk analysis and faulttrees were developed for the spillway gateoperation to determine the system reliability asinput to the risk analysis.

The spillway operation and maintenanceprocedures and test intervals were used toderive failure probabilities for components andthe response procedures to flood events werereviewed and incorporated in the fault trees.

1.2 L ittle Nerang Dam

The Little Nerang Dam is a concrete gravity

1. Principal Engineer Dams, GHD; 2. Principal Engineer Mechanical & Electrical, GHD; 3. Dam Safety

Engineer, Southern Rural Water; 4. Asset Management Team Leader, Dams and Treatment Service DeliveryBranch, Gold Coast Water


2/15


3/15



abutment sections of 79.55m AHD. TheProbable Maximum Flood with an inflow of11,700m3/s results in a reservoir level of 84.8mAHD, assuming all gates are fully open. Atthis level, the structure will be completelyovertopped but stability analyses have shownthat it should withstand this loading.

Hoist and Controls

Each radial gate has its own electric motorcentrally located above the gate, which iscoupled to a worm reduction gear box anddrive shaft. Further spur gear reductions drivetwin rope drums from which the lifting ropesare attached to the gates (Photo 3). Thrusterbrakes are used to maintain the gate position

for each gate operation.

The gates are operated from the structure usingthe relevant switch board. Should the powersupplies fail, an emergency petrol drivenhydraulic pump can be connected through adog clutch to the drive shaft to open each gate(Photo 4).

Power Supply

The electric power can be supplied from thenormal incoming grid supply or a 20kVA

standby diesel generator. Only two of thegates can be operated at any one time using thediesel generator supply

The standby generator is located in a buildingon the right bank near the dam crest level and a

single power supply cable is taken from thediesel generator switchboard to the mainswitchboard, which results in a common causefailure mode for the power supply atGlenmaggie Dam (ie a fault at the mainswitchboard, diesel switchboard or powersupply cable will disable both AC powersupplies).

O & M Procedures and Response to Inflows

Southern Rural Water has a program of regularinspection and maintenance of all gates andtheir hoist equipment. This includes oiling,greasing and testing of components atprescribed intervals. Qualified electricalpersonnel also carry out 3 monthly inspectionsand testing. Any faulty mechanical orelectrical components are immediatelyreplaced or appropriately serviced. The gatesare exercised monthly (if not used for floodoperation). Ageing components such as limitswitches are replaced before they become aproblem. The gates are repainted or patch

painted as is determined necessary fromregular inspections.

Photo 2

Photo 3

Photo4


4/15



During a flood, the gates are operated via a setprocedure. Inflows and rainfall can bemonitored remotely and inflows are computedbefore gate movement is determined. Thefrequency of this depends on the magnitude ofthe storm. In general the gates are opened in0.5m increments and although only one gate isoperated at a time, several gates may bepartially open at once. The central 8 gates areopened prior to the outer 6 gates to minimisedownstream erosion.

2.2 L it tle Nerang Dam

The Little Nerang Dam has a capacity of8400 ML and was constructed between 1954 to1961. The dam is a concrete gravity structure

of 175m crest length with a central ogee gatedspillway section having twin 3.6m high by16.5m wide rotating drum gates for floodcontrol. The outlet works are located in theright bank block adjacent to the spillway andcomprise a multilevel dry tower offtakeleading to a single outlet conduit. A drainagegallery has been provided at 3.1m from theupstream face with drain holes at 6m centresextending into the foundation and concretesection above.

Gates

The two gates are of identical form and size,and consist of watertight steel drums, whichfloat on the water in the gate chambers rotatingabout a hinge on their upstream edge, as shownon Figure 1. The gates, which lower todischarge, are ideal for passing floating debrisover the spillway without delay or blockage intimes of flood. The position of the gates isdetermined by water pressure in the drum

chambers. If the automatic control gear isfunctioning, the gates will be controlledentirely by the height of water in the dam.

With the exception of the emergency controlhydraulic system the gates are notinterconnected and work entirelyindependently. The interconnection of thehydraulics provides an added level of operatingredundancy to each gate and theinterconnecting valves are kept open at alltimes.

Seals are provided on the ends of the upstreamand downstream skin faces of each gate to

prevent water leaking between the gate and itsadjacent end abutments or piers. Seals are alsoprovided along the downstream gate seats, andbear against the downstream drum skin of eachgate to prevent water in the gate chambersfrom escaping.

Figure 1 Little Nerang Section of Drum Gate

Controls

The gate can be operated using three methodsof control:

Automatic Control

Manual Control

Emergency Control

The controls, which operate each gate undernormal conditions and for manual control, arelocated in Control Rooms formed within theconcrete structure immediately adjacent to theouter ends of each gate. The controls, whichoperate the gates under emergency conditions,are located in a Control Room formed within

the concrete structure of the centre pierbetween the two gates.

Access to all three Control Rooms is gainedvia vertical shafts and a horizontal galleryformed within the concrete structure below thelevel of the bottom of the drum chambers.

Automatic Control (Figure 2)

As long as the dam water level is up to the topof the spillway ogee crest (RL 168.02m) or

higher, the gates will float at their maximumheight with top edges at RL 171.6m.


5/15



Figure 2 Automatic Control when Water Level is Above RL171.6m

Photo 5 Little Neran Dam - Larner Johnson


6/15



If the water level continues to rise, it will flowover the top of the gates and the floats in thefloat chambers will begin to rise.

Through the associated rigging, shownschematically on Figure 2, the floats will begin

to open the Larner-Johnson discharge valves(Figure 3 & Photo 5). By the time the waterlevel has reached 180mm over the top of thegates (RL 171.78m) the Larner-Johnson valvesshould be fully open and sufficient water willbe discharged from the drum chambers tocause the gates to move downward todischarge more water over their top edges.

Figure 3 Larner-Johnson Flow-valve

The downward movement of the gates willtend to lower the dam level by increasing thedischarge over their top edges, at the sametime the movement will be reflected throughthe control rigging to slightly close theLarner-Johnson valves and reduce the rate ofdischarge of water from the drum chambers. Ifas a result of these combined effects, the damwater level is reduced to RL 171.60m or lower,the Larner-Johnson valves will re-close fullyand the gates will resume their position atmaximum height.

Manual Control

Manual lowering of one or both gates isachieved by closing the interconnecting valvefor the emergency system and the valve on the375mm inlet line to the drum chamber on the

gate to be lowered and then opening theLarner-Johnson valve and discharging thewater from the drum chamber. The gateposition can be held at any position by closingthe Larner Johnson valve, however, this cannotbe relied upon to keep a gate indefinitely in afixed position because leakage into and out ofthe drum chamber will influence the gate levelover a long period.

Emergency Control (Figure 4)

If the dam level continues to rise to theemergency control intake level atRL 172.21m, water will flow into this inletand into the emergency control bucket. Thebucket contains a drain hole in the bottom,allowing water to flow into the bucket welland then drain away to the downstream face ofthe dam. When the water level reaches172.25m, the inflow will exceed the outflowand the bucket eventually fills. When thisoccurs, the weight of the bucket and its

contents will overcome the counterweight andopen the pilot valve. This will discharge thewater from above the 457mm float valve(Figure 5) and allow the float valve to open.

Figure 4 Emergency Control When the Water Level is above RL 172.21m


7/15



Figure 5 Schematic of the Operating Principle of the Emergency Valve

The valve on opening drains water from thedrum gate chamber, which reduces the upliftpressure on the drum gate lowering the gate ata faster rate than normal by:

Acting in addition to normal operation;and

Creation of a powerful hydraulicsuction as the water falls down througha long pipe exiting just under the lip ofthe ski-jump at the bottom of thespillway. This suction evacuates thewater from the gate chamber muchquicker than it can enter, thus lowering

the gate regardless of normaloperation.

When this additional discharge causes the damlevel to recede below RL 172.21m water canno longer enter the emergency control intake,hence the bucket will empty and thecounterweight will close the pilot valve. Thisin turn will cause the float valve to close andcut off the discharge of the drum chamberwater, following which the gates will thenreturn to their highest level.

3. SPILLWAY GATE FAULT

TREE ANALYSIS

The spillway gate reliability for theGlenmaggie Dam gates was evaluated usingfault trees required to capture the various pathsto failure of a single gate or multiple gates.

A fault tree is defined as a systemsengineering method for representing the

logical combinations of various system states

and possible causes which can contribute to aspecified event (called the top event).

Methodology

For the purpose of the Glenmaggie Dam gatestudy, the top event for the fault trees wasdefined as failure of a single gate or multiplegates when considering failures of componentscommon to all gates e.g. AC Power supply.

The fault trees were structured to examine boththe independent component failures and thefailures of components common to all gates.The independence of gate operation for

multiple gates will provide increasingredundancy, which greatly reduces theestimated probability of a failure andsubsequent loss of discharge capacity. Forcommon cause failures, a larger number ofgates will generally offer a limited advantage.

In order to comply with the basic requirementsof the fault tree process, as defined above, thefault trees were developed using the followingsteps:

(a) Definition of the spillway gate system into

components and sub-components asshown on Table 1, abstracted for theelectrical system;

(b) Site visit to examine and confirm thedefinition of the system and discuss theoperation and maintenance procedureswith the site personnel as well asdetermine the human response to gateoperation failure;

(c) Perform a Failure Modes and EffectsAnalysis (FMEA) for the components of

the gate system as defined after steps (a)and (b);


8/15



Table 1 Abstract of Component Definitionfor Glenmaggie Gate ElectricalSystem

(d) Development of the single gate fault treeusing immediate cause concept for theelectrical and mechanical components,and structural failure of the gates orsupports (Figure 6).

Figure 6 Fault Tree Basis for Top Event

It should be noted that the overallprobability of failure is not simply anaddition of the contributing branches butmust take into account the mathematics ofBoolean algebra to calculate thecombination of the events (and/orpossibilities must be taken into account).

(e) Each of these areas was then broken downsystematically into basic events, which

could either be a single component eg.wire rope failure or an event for whichfailure rate data is available e.g. dieselgenerator. Figures 7 and 8 show part ofthe breakdown for motive power failure.

A critical issue in the fault treedevelopment is the inclusion of commoncause failures for components where twoor more systems are dependent on acommon component, failure of whichcould lead to failure of both systems e.g. acommon power supply cable for the mainsand diesel generator to the dam (seeFigure 7).

Figure 7 Fault Tree Details Abstract 1

Figure 8 Fault Tree Details Abstract 2

3

Components No Sub Components No ID No

Mains Power 1 Transformers 1 3.1.1

Switching 2 3.1.2

Power lines and Poles 3 3.1.3

Cables 4 3.1.4

Standby Diesel Generator 2 Diesel Motor 1 3.2.1

Alternator 2 3.2.2

Switching 3 3.2.3

Fuel 4 3.2.4Cables 5 3.2.5

Battery 6 3.2.6

Emergency hydraulic pump 3 Petrol Motor 1 3.3.1

Fuel 2 3.3.2

Hydraulic Lines 3 3.3.3

Gate brake 4 3.3.4

Operator 5 3.3.5

Battery 6 3.3.6

Main Motor 4 Power Supply 1 3.4.1

Bearings 2 3.4.2

Shaft Coupling 3 3.4.3

Windings 4 3.4.4Control 5 3.4.5

Switching x 8 6 3.4.6

Thrustor Brake 5 Power Supply 1 3.5.1

Bearings 2 3.5.2

Adjustment Mechanism 3 3.5.3

Brake Shoes 4 3.5.4

Windings 5 3.5.5

Electrical

SPILLWAY GATES


9/15



(f) Evaluation of available failure rate andhuman error data and use of establishedmathematical methods of analysis toestimate the failure likelihood ofequipment on the basis of the type of theevent in the Fault Tree and the operatingexperience at Glenmaggie dam for thefollowing:

Dormant components e.g. hydraulichose;

Continuous running components;

Running failure probability of adormant component, given asuccessful start;

Probability of occurrence of an event;

Combined Dormant Failure andRunning Failure.

Human error was included in thefollowing areas of the gate operation forwhich typical estimates of human errorare shown on the table below.

Individual controls at the gates;

Back-up diesel generator switching;

Operation of the hydraulic backup

system.Human Error Factor Prob.

Manual Operate Valve/switch, left in wrongposition

-Maintenance. Shutdown, orcalibration 0.085

- Routine operation 0.015

- Routine Testing 0.015

-Normal Operation (Includesoperation of hydraulic backup) 0.01

Table 2 Glenmaggie Dam Human

Error Probability Estimates(Draft Results)

(g) Quantification of the fault trees;

(h) The analysis of multiple gate failures wasachieved by developing the fault tree withgates using common cause components toobtain the probability of 1, 2 3 gatesfailing etc, as shown on Figure 9.

Analysis Results

The results from the fault tree analysisincluded:

Overall likelihood of system failure;

Cut sets for failure combinations of events,which is the minimal number of events thatcould lead to system failure. These showedthat there are several sets with only two

events leading to system failure;

Fault tree gate failure probabilities;

Importance rating for events.

The results obtained for single and multiplegate failure, which were provisional at the timeof publication, are shown on Tables 3 and 4.

Figure 9 Fault Tree for Multiple Gates

No of Gates Failing Probability of

Failure

1 0.125

2 0.037

3 0.0193

4 0.006865 0.00681

6 or more 0.00681

Table 3 Abstract of Combined FailureProbabilities for Glenmaggie DamGates (Draft Results)

6 or More

Gates fail

1 to 6

Gates Fail

No Gates

Fail0.00681 0.20178 0.79141

Table 4 Glenmaggie Dam Gate FailureProbabilities (Draft Results)


10/15



Discussion of Fault Tree Analysis Results

A breakdown of the main system contributionsto the overall failure likelihood is shown inTable 5, which indicates that the electrical andhydraulic motive power and the mechanical

hoisting equipment are the major contributorsto failure of a gate to operate.

System and Subsystem P

(Failure)

%

Total

All Motive Power 4.31E-02 34%

Hoisting Mechanism 7.12E-02 57%

Support Structures 1.10E-02 9%

Gate Structure 4.99E-03 4%Table 5 Glenmaggie Dam Risk Analysis

Analysis of Sub System Failures forSpillway Gates (Draft Results)

Hydrological Risk Analysis

The spillway gate reliability results wereused in a risk analysis, which includedseismic and flood events. The followingfailure pathways were developed for theflood events based on the results of thestructural and hydrological analysis and aFMECA for the dam system andsubsystems as follows:

Flood event;

Debris present;

Spillway blockage due to debris;

Spillway gates fail to operate;

Spillway gates structural failure;

Resulting reservoir level

o Non-overflow crest section failuremode:-

Overtopping

Downstream erosion - section

instability

breach or no breach

No downstream erosion - section

instability

breach or no breach

No Overtopping

Section instability, breachor no breach

o Overflow section failure mode

Section instability

Breach or no breach

The risk analysis included an evaluation of theerosion on the abutments due to overtopping aswell as @Risk simulations using friction andcohesion distributions to evaluate the stabilityof the concrete spillway and non overflowsections. The results of the risk analysis aresummarised in Table 6 for one Flood Range(F2) and shown on Figures 10 and 11.

F1

SpillF1

AbutF2

SpillF2

AbutF3

SpillF3

AbutF4

SpillF4

AbutF5

SpillF5

Abut

Debris/Gate Fail - Mech/Elec

Debris/Gate Fail - Mech Elec / Structural

Debris/Gates OK

Debris/Erosion

No Debris/Gate Fail - Mech Elec / Structural

No Debris/Erosion

No Debris/Gates OKNo Debris/Gate Fail - Mech/Elec

0%

10%

20%

30%

40%

50%

60%

70%

80%

90%

100%

Figure 10 Percentage Failure Contributionsfor Flood Ranges and FailurePathways

It can be seen that cases involving spillwaygate failure make a 10 to 20% contribution tothe probability of failure due to hydrologicalevents.

1.0E-06

1.0E-05

1.0E-04

1.0E-03

1 10 100 1000

Potential Loss of Life (N)

FProbabilityoffailureperdam

yearwithexpectedlossof

lifeN

Seismic & Flood Graham Total

Seismic & Flood Graham Incremental

Flood Graham Total

Seismic Graham Total

Figure 11 Glenmaggie Dam Societal Risk


11/15



Failure Mode Total

Gates OK Erosion Gates OK Erosion

Mech/Elec Mech Elec /

Structural

Mech/Elec Mech Elec /

Structural

F2.1 (Spillway 3560 to 5120) : Debris : 6 or More Gates Fail Closed : 7.44E-10 7.36E-08 0.00E+00 0.00E+00 2.63E-08 2.60E-06 0.00E+00 0.00E+00 2.70E-06

F2.1 (Spil lway 3560 to 5120) : Debris : 1 to 6 Gates Fail Closed : 2.20E-08 2.18E-06 0.00E+00 0.00E+00 7.79E-07 7.71E-05 0.00E+00 0.00E+00 8.01E-05

F2.1 (Spillway 3560 to 5120) : Debris : All Gates Open : 0.00E+00 0.00E+00 8.64E-06 0.00E+00 0.00E+00 0.00E+00 3.06E-04 0.00E+00 3.14E-04

F2.1 (Spillway 3560 to 5120) : No Debris : 6 or More Gates Fail Closed : 0.00E+00 0.00E+00 0.00E+00 0.00E+00 4.96E-08 1.75E-06 0.00E+00 0.00E+00 1.80E-06

F2.1 (Spillway 3560 to 5120) : No Debris : 1 to 6 Gates Fail Closed : 0.00E+00 0.00E+00 0.00E+00 0.00E+00 1.47E-06 5.19E-05 0.00E+00 0.00E+00 5.34E-05

F2.1 (Spil lway 3560 to 5120) : No Debris : Al l Gates Open : 0.00E+00 0.00E+00 0.00E+00 0.00E+00 0.00E+00 0.00E+00 5.76E-04 0.00E+00 5.76E-04

Total Spillway 2.28E-08 2.26E-06 8.64E-06 0.00E+00 2.32E-06 1.33E-04 8.82E-04 0.00E+00 1.03E-03

Percentage Sub-total Spillway 0.0% 0.2% 0.8% 0.0% 0.2% 13.0% 85.7% 0.0% 100%

F2.2 Abutment 3560 to 5120 : Debris : 6 or More Gates Fail Closed : 7.44E-10 7.36E-08 0.00E+00 0.00E+00 2.63E-08 2.60E-06 0.00E+00 0.00E+00 2.70E-06

F2.2 Abutment 3560 to 5120 : Debris : 1 to 6 Gates Fail Closed : 2.20E-08 2.18E-06 0.00E+00 0.00E+00 7.79E-07 7.71E-05 0.00E+00 0.00E+00 8.01E-05

F2.2 Abutment 3560 to 5120 : Debris : All Gates Open : 0.00E+00 0.00E+00 8.64E-06 0.00E+00 0.00E+00 0.00E+00 3.06E-04 0.00E+00 3.14E-04

F2.2 Abutment 3560 to 5120 : No Debris : 6 or More Gates Fail Closed : 0.00E+00 0.00E+00 0.00E+00 0.00E+00 4.96E-08 1.75E-06 0.00E+00 5.44E-07 2.35E-06

F2.2 Abutment 3560 to 5120 : No Debris : 1 to 6 Gates Fail Closed : 0.00E+00 0.00E+00 0.00E+00 0.00E+00 1.47E-06 5.19E-05 0.00E+00 1.61E-05 6.95E-05

F2.2 Abutment 3560 to 5120 : No Debris : All Gates Open : 0.00E+00 0.00E+00 0.00E+00 0.00E+00 0.00E+00 0.00E+00 2.06E-04 0.00E+00 2.06E-04

Total Abutment 2.28E-08 2.26E-06 8.64E-06 0.00E+00 2.32E-06 1.33E-04 5.11E-04 1.67E-05 6.75E-04

Percentage Sub-total Abutment 0.0% 0.3% 1.3% 0.0% 0.3% 19.8% 75.8% 2.5% 100%

Debris Blockage No Debris or No Debris Blockage

Gates FailGates Fail

Failure Probability

Table 6 Glenmaggie Dam Risk Analysis - Results Summary (Draft) for Flood Range (F2)

The results of the risk analysis show thatspillway gate failure has a small contributionto the overall risk of dam failure which isbelow the ANCOLD tolerable limit for societalrisk. The results of the study have helped toidentify the critical components of the systemand will also assist in any future decisionsregarding remedial works for the spillway

gates. At present Southern Rural Water has noplans for undertaking remedial works to thespillway gate system, apart from routineupgrading of equipment to current standards.

The study also showed the importance of thegate operator in the correct functioning of anyspillway gate system. Redundancy in gateoperating systems was also shown to beimportant.

4. LITTLE NERANG DAMSPILLWAY DRUM GATE

RISK ANALYSIS

The methodology used to evaluate the risk tolife for the drum gate operation andcomparison with the ANCOLD acceptancecriteria was carried out using the followingtasks:

Evaluation of uplift data for foundation

piezometers and Critical Stability Levelsfor estimation of failure probabilities;

Development of spillway gate dischargerating data;

Floodrouting of floods from 1 in 2 AEP tothe PMF and estimation of Critical StabilityFlood exceedence probabilities;

Estimation of Population at Risk and Lossof Life;

Evaluation of Hazard Rating andDeterministic Fall Back PositionAcceptable Flood Capacity;

Risk to Life Estimation, including gatefailure estimation;

Comparison of Risk with ANCOLDCriteria

Uplift Data

Analysis of the available uplift data for the

foundation piezometers has shown that theuplift is generally at or below the 33% designline for all blocks except block 11 adjacent tothe spillway for which the uplift follows amodified design line of 44% residual head atthe drains. This revised uplift distribution wasused for analysis of block 11 at the foundationwhile the 33% design line was adopted for theremaining nonoverflow and spillway blocksat the foundation and a 25% residual head wasused for the drains within the concrete section.

The data indicates that there is no trend ofincreasing foundation pressures evident in thedata base, however, a 50% residual uplift head


12/15



was also used for analysis of Block 11 for asensitivity analysis.

The resulting analyses indicated that Block 11was the critical block Reservoir levels atwhich the shear friction factor was equal to 1.0

were determined, as shown on Table 7, forvarious residual drain uplift factors.

Description Critical Level

(m AHD)

Foundation (RL 136.0m)

33% uplift 174.25

44% uplift 173.75

50% uplift 173.45

Dam Top - Change of Slope RL 167.2m

25% uplift 173.81Table 7 Little Nerang Dam Critical StabilityLevels for Block 11

Spillway Rating Curves and Floodrouting

Spillway rating curves were developed for gateoperating scenarios for which flood routingwas completed using floods ranging from a 1in 2 AEP event to the PMF as follows.

Both gates locked down (Case I)

Both gates operating normally (Case II)

Both gates suffer malfunction and remain inup position (Case III)

One gate operating normally and one gatesuffering malfunction (Case IV)

The reservoir level flood exceedence data isshown on Figure 12 together with the criticalstability levels.

169

170

171

172

173

174

175

176

1 10 100 1,000 10,000 100,000 1,000,000 10,000,000

Flood AEP (1 in X)

Res.

Level(m

AHD)

Water Level Gates Down

Water Level Gates Working

Water Level Gates Failed Up

1 gate working and other gate failed up position

Block 11 (33% uplift) Foundation 136m - Crack Instability Level

Dam Top - Change of Slope RL 167.2m -



Figure 12 Little Nerang Dam Flood Levelversus AEP

Population at Risk (PAR) and Loss of Life

(LOL) Estimation

The population at risk is located within 3.5kmfrom the dam and comprises 6 houses and ayouth rehabilitation centre, which ceased to

operate and was purchased by the Gold CoastCity Council to reduce the PAR as a riskreduction measure.

The estimated total and incremental PAR forthe gate operating Cases are shown on Table 8,together with an estimate of the total PAR thatcould be present after a warning is given. Itwas assessed that the Total PAR would beappropriate for the Sunny Day failure eventsi.e. seismic failure.

Gate

Operation

Case

Total

Breach

PAR

Total

Breach

PAR

afterWarning

Incremental

Breach to No

Breach

Prior to purchase of Rehabilitation Centre

1 & II 56 22 46

III & IV 56 22 46After purchase of Rehabilitation Centre

1 & II 18 6 3

III & IV 18 9 6

Table 8 Little Nerang Dam Population atRisk for Gate Operation Cases

Loss of life estimates were made using theUSBR (Graham 1999) methodology assuminga high severity of flooding. The appropriateloss of life factor is 0.75 of the PAR for whichthe LOL estimates were made, as shown onTable 9. No loss of life was assumed for floodevents with no dam failure.

Gate

Operation

Case

Total

LOL

Total

LOL

after

Warning

Incremental

Breach to

No Breach

LOL

Prior to purchase of Rehabilitation Centre

1 & II 44 17 34

III & IV 44 17 34After purchase of Rehabilitation Centre

1 & II 14 5 3

III & IV 14 7 5Table 9 Little Nerang Dam Potential Loss

of Life for Dam Failure with GateOperation Cases


13/15



Hazard Rating and Fallback Position Flood

Based on the incremental PAR of 6 afterremoval of the PAR from the rehabilitationcentre and the medium severity of damage andloss assessed for Little Nerang Dam, the

Hazard Category of the dam is Significant. Asthere is a potential for loss of life, thesignificant rating is required to be increased toa High C (ANCOLD 2000b).

The hazard rating for the PAR of 46 would beHigh C.

With the hazard rating of Significant and thePAR of 6, the required fallback flood capacityAEP would be about 1 in 6000 with nopotential for loss of life. As the possible lossof life is recognised for the High C rating, the

required AEP for the Acceptable FloodCapacity (AFC) is 1 in 10,000.

Using the incremental PAR of 46 prior toremoval of the rehabilitation centre, the AEPfor the AFC would be about 1 in 50,000.

Spillway Gate Reliability

During the evaluation of the spillway gatereliability in 1995, a number of dam ownersworldwide were contacted and responses

received from the following dam owners withdrum gates: USBR, Greater VancouverRegional District (GVRD); Scottish HydroElectric (SHE); Ministry of Public Works(Spain) (MPWS); Sydney Water Board(SWB). Two experts were also approached fordata: Dr Nelson Pinto (Brazil) (NP) andProf J. Lewin (UK) (JL).

The responses and comments are shown onTable 10 for the events, which resulted in gatesfailing to lower due to mechanical causes.

The USBR experience has been that withaggressive maintenance including replacementof seals and hoses, the drum gates areextremely reliable. The automatic floatsystems have been replaced in a number oftheir dams by manual operation or electricalsensor operation.

Based on the responses and comments onTable 10, the probability of failure for theLittle Nerang Drum gates was taken to be 1 in50.

The estimated AEP of the flood for each of thespillway gate operating scenarios and

combined gate and flood exceedencefrequency is shown on Table 11, whichindicates that the worst case scenario is forfailure of both gates with a combined AEP of 1in 17,500.

Contact No ofDams

Ageyrs

Data and Comment

USBR 9 60to70

Extremely reliable1 malfunction:Worn seat seal

GVRD 1 40 1 malfunction:Worn control valve

SHE 2 44 3 malfunctions :Automatic valvesAnalysis 1:67 pergate

MPWS 1 35 None

SWB 1 35 1 malfunction:Worn O ring oncontrol valve

NP - 1 in 100 operationsJL - 1 in 500 operations

Table 10 Drum Gate Malfunction Data

It is clear from Table 11 that the dam was notable to pass the fallback position flood prior tothe removal of the PAR from the rehabilitation

centre but currently meets the ANCOLDfallback position AFC.

Gate

Case

AEP of

Failure

Flood(1 in X)

Gate

Failure

Probability

Gate &

Flood Freq.

(1 in X)

I 250,000 1 250,000

II 250,000 1 250,000

III 7 1/2500 17,500

IV 3000 1/50 150,000

Table 11 Little Nerang Dam Block 11 Floodand Dam Failure Frequency Data

(Critical Level with 44% Uplift)

Given the AEP of 1 in 7 for the flood reachingthe critical stability level and the required AEPfor the spillway design flood of 1 in 10,000implies that the spillway gate failure for eachgate could be as low as (7*1/10,000)0.5 or2.6E-2 (1 in 38) per gate before there is aconcern over the gate reliability. This failurerate is clearly in the order of the historicalfailure rate shown on Table 10.

Notwithstanding this, a further level ofredundancy in the spillway gate operation is


14/15



being evaluated using a manually operatedbypass for the Larner-Johnson valve.Furthermore, the maintenance work performedfor the gates at Little Nerang Dam and the useof the Operation and Maintenance Manual andan Emergency Action Plan for all floodconditions are considered appropriate forensuring that the assumed gate reliability isachieved.

Risk to Life and Comparison with

ANCOLD Criteria

The data for gate failure probabilities, damfailure frequencies and potential loss of lifewas used to develop a simplified event tree foranalysis of the flood failure events with the

following scenarios:

Both gates operating;

One gate failure to operate;

Both gates failure to operate.

The probability of dam failure at the criticallevel was assumed to be 0.5 and the analyseswere performed for the total LOL with andwithout warning and the incremental LOL forthe 44% and 50% uplift critical failure levelsfor Block 11.

Structural analysis of the dam for seismicloading has shown that the dam meets thecurrent ANCOLD Criteria and the annualfailure frequency was assumed to be 1E-5 inthe risk analysis.

The results of the analysis using the 44%residual uplift critical level and the total LOLafter warning the PAR are shown on Table 12.

Using the total annual failure frequency (F)shown in Table 12 and an exposure factor of

0.5 for the individuals in the downstream floodaffected area, the Individual Risk wascalculated to be 2.2E-5. A sensitivity analysisperformed using the data for the 50% residualuplift gave an individual risk of 5E-5. Theseindicate that the individual risk is below theANCOLD tolerable limit of 1E-4 for anexisting dam (ANCOLD 1998).

Societal risk FN plots were developed usingthe data shown on Table 12 for each upliftcritical level and LOL estimates forcomparison against the ANCOLD Criteria, asshown on Figure 13.

Description FailureFreq.

LOL F

Seismic Failure 1.0E-05 14 1.0E-05

Both Gates Fail 2.9E-05 7 3.9E-05

1 Gate Fails 3.3E-06 7 4.2E-05

Both GatesOperate 2.0E-06 5 4.4E-05

Table 12 Little Nerang Dam Risk AnalysisResults with 44% Residual UpliftCritical Level and Total LOL afterWarning

1.0E-08

1.0E-07

1.0E-06

1.0E-05

1.0E-04

1.0E-03

1.0E-02

1 10 100 1000 10000

N, Number of Fatalities due to dam failure

F,

Probabilito

ffailure

erdam

earw

ithex

ectedlossoflife>=

Incremental LOL 44%Uplift

Incremental LOL 50%Uplift

TotalLOLwithWarning44%Uplift

TotalLOLwithWarning50%Uplift

TotalLOL44%Uplift

TotalLOL50%Uplift

Total LOL pre-removal of Rehabilitation Centre PAR 44%Uplift

TotalLOLnowarning-pre-removalofRehabCentrePAR44%Uplift

ANCOLD Limit of

Tolerability ExistingALARP

Little Nerang Dam

50%Uplift

LittleNerangDam

44% Uplift

Figure 13 Little Nerang Dam Societal RiskData

Figure 13 shows that the societal risk wasabove the tolerable limit for the conditionsprior to removal of the PAR from therehabilitation centre and for the conditionwhere the uplift is allowed to increase to the50% residual value in the Block 11 drains.

Clearly both of these conditions were notacceptable and provided strong justification for(a) removing the PAR from the flood affected

zone and (b) ensuring that the monitoring ofthe uplift is maintained and necessary actiontaken to improve the drainage efficiencyshould the monitoring indicate an increase inuplift towards the 50% residual level.

Operation, Maintenance and Monitoring

Gold Coast Water have also taken steps toinstitute an intrusive operation andmaintenance programme and to develop aflood monitoring procedure as follows.


15/15



Inspection and Maintenance Programme

Six Monthly

Routine inspection of control and floatchambers, checking for operation, sludgeremoval, opening of outlets and inlets,

counterweight operation, valve operationincluding Larner-Johnson valve andbutterfly valves.

Annually

Lubrication and cleaning of all valvespindles, screws pulley pins, turnbucklesand wire cables. General maintenance ofdoors, access ladders etc.

2 Yearly

Inspections of the gate structures, hinges,seals, dogging devices and drum chambers.

Flood Monitoring Procedure

A simplified flood monitoring procedure hasbeen developed based on the recorded depthsof rainfall and dam water level. The focus ofthe monitoring is to provide a user friendlysystem for the dam operators withmeasurements taken from the on site rain andwater level gauges. The procedure was

developed using the data obtained from routingof the floods with AEPs from 1 in 2 to the 1 in50,000 AEP events with the various gateoperating scenarios and start reservoir levels todevelop a series of graphs similar to thatshown on Figure 14.

The flood monitoring procedure involves thefollowing steps:

Recording of reservoir level, gate status,rainfall and time for all storm events;

Calculation of incremental and cumulativerainfall intensity;

Estimation of AEP for the recorded rainfall;

Comparison of the plotted water levelversus time after start of the storm eventapplicable to the estimated AEP with thecalculated data similar to Figure 14.

This comparison is used to determine (a) thelikelihood of reaching the flood warning leveland (b) likelihood of reaching the criticalstability level to be used for warning thedownstream PAR.

Little Nerang Dam Water Level Versus Time After Start of Storm EventCASE III (Both gates failed up) - 1 in 5 year ARI

171

171.5

172

172.5

173

173.5

174

0 2 4 6 8 10 12 14 16

Time After Start of Storm Event (h ours)

WaterLe

vel(mAHD)

1 hour storm event

2 hour storm event

3 hour storm event

4.5 hour storm event

6 hour storm event

9 hour storm event

12 hour storm event

18 hour storm event

24 hour storm event

DANGER: Evacuate Downstream

DAMINSTABILITYPOSSIBLE

STARTWATERLEVEL= 171.6mAHD

Figure 14 Little Nerang Dam: 1 in 5 AEPEvent; Case III Gate Operation;Water Level versus Time

5. CONCLUSIONS

The application of a detailed fault tree analysisfor the radial gated Glenmaggie Dam spillwayhas been used in a risk assessment to assist inthe evaluation of the requirements for remedialworks and provide guidance for operation ofthe gated systems.

The drum gate operating system of LittleNerang Dam has been evaluated in some detailto provide the required operation andmaintenance procedures necessary to ensurethe risk is kept within the acceptable tolerablelimits. The historical failure rate for thesegates has been used in a simplified riskanalysis performed for the gates to determinethe societal risk. As a result of the risk analysesperformed over the last few years, measureswere taken by the dam owner to reduce thepopulation at risk by purchasing downstreamproperty as a means of reducing the risk to life.

ACKNOWLEDGEMENTS

The authors would like to thank the Southern

Rural Water and Gold Coast Water forpermission to produce this paper and GHD forthe encouragement and support provided.

REFERENCES

1. ANCOLD 1998, Position Paper onRevised Criteria for Acceptable Riskto Life

2. ANCOLD 2000a, Guidelines on Selectionof Acceptable Flood Capacity for Dams

3. ANCOLD 2000b, Guidelines on

Assessment of the Consequences of DamFailure

2003 ancold - spillway gate reliability - barker et al (2)

Documents