Ri k B d M i tRisk Based Maintenance ERC ScSc edu g o C cu ...

ERCPSRi k B d M i t ERCPSRisk Based Maintenance Scheduling of Circuit Sc edu g o C cu t

Breakers using C diti B d D tCondition-Based Data

Mladen KezunovicTexas A&M UniversityTexas A&M University

EPCC 10, Dublin, IrelandJune 14-18, 2009

ERCPSOutline

• Introduction• CB Monitoringg• Maintenance Quantification Model• Risk Based Maintenance Approachpp• Case Studies• SummarySummary

IAB Meeting, Dec. 4-5, 2008

ERCPSIntroduction Bus 15Introduction

0.35 Probability Between Limits is 0.94016

GBB1

Bus 14

L24

L23

L29

B1

B2

B4

B5

B7

0.05 0.1

0.15 0.2

0.25 0.3

Den

sity

BB2

B s 17

Bus 19

L28

L29

B3 B6 B8

L d

Condition Based Data• CB Control Circuit

Performance Indices• CB Failure

Risk Analysis• Probability• Consequence

System Maintenance• Risk Reduction

8 10 12 14 16 18 200 Critical Value

Lower Upper (msec) Bus 17Load

• CB Control Circuit• Signal Processing

• CB Failure• Bayesian

Approach

• Consequence• Risk

• Risk Reduction• Optimization

ERCPSCB Monitoring g

Over view of monitoring choices:

• Operating MechanismC t t T l ti M t- Contact Travel time Measurement

- Control Circuit Monitoring- Vibration Analysisy

• Contacts- Resistance Test- Temperature Monitoring

• Inspection of oil (oil circuit breakers)• Partial Discharge


ERCPSCB Monitoring: Data from CBMsCB Monitoring: Data from CBMs

Close Trip

ControlDC +

Close Initiate

Trip Initiate

52a52X/a

Control DC

52TC

CC

52

52Y/a52a

52Y/b52Y/b

X

Y CBM

PortableDC _ Devices


ERCPSCB Monitoring: Data from CBMsCB Monitoring: Data from CBMsWaveform abnormalities and signal parameters

EVENTEVENT EVENT DECRIPTIONEVENT DECRIPTION SIGNALSIGNAL

11 Trip or close operation is initiated (Trip or close Trip or close operation is initiated (Trip or close initiate signal changes from LOW to HIGH)initiate signal changes from LOW to HIGH) T1T1g g )g g )

22 Coil current picks upCoil current picks up T2T2

33 Coil current dips after saturationCoil current dips after saturation T3T3

44 Coil current drops offCoil current drops off T4T444 Coil current drops offCoil current drops off T4T4

55 B contact breaks or makes (a change of status B contact breaks or makes (a change of status from LOW to HIGH or vice versa)from LOW to HIGH or vice versa) T5T5

66 A contact breaks or makesA contact breaks or makes T6T6

77 Phase currents breaks or makesPhase currents breaks or makes T7T7

88 X coil current picks upX coil current picks up T8T8

99 X coil current drops offX coil current drops off T9T9pp

1010 Y coil current picks upY coil current picks up T10T10


ERCPSCB Monitoring: Data from CBMsCB Monitoring: Data from CBMs

Summary of Test Records During Closing Operation

Manufacturer and Type: GE VIB-15.5-20000-2

Summary of Test Records During Closing Operation of Circuit Breaker

Date T2 (sec) T3(sec) T4(sec) T5(sec) T6(sec)

2/12/2002 0.001215 0.010417 0.028993 0.056597 0.0668402/12/2002 0.000868 0.012500 0.032639 0.058160 0.0682292/12/2002 0.000868 0.012500 0.032639 0.058160 0.0682292/13/2002 0.001042 0.014236 0.048785 0.055903 0.0664932/13/2002 0.001736 0.011979 0.043229 0.052951 0.0661462/19/2002 0.001389 0.017361 0.037500 0.059896 0.0078132/21/2002 0.003819 0.004861 0.034375 0.056424 0.067535

6/11/2002 0.001736 0.011285 0.032292 0.063542 0.0729176/11/2002 0.000868 0.014236 0.031076 0.063021 0.0725696/11/2002 0.000694 0.010243 0.032465 0.060590 0.0708336/11/2002 0.000694 0.013889 0.032639 0.061458 0.0704866/11/2002 0.001042 0.011111 0.048958 0.057118 0.068056

ERCPSMaintenance Quantification ModelMaintenance Quantification Model

0.1

0.15

0.2

0.25

0.3

0.35Probability Between Limits is 0.94016

Den

sity

8 10 12 14 16 18 200

0.05

0.1

Critical Value Lower Upper (msec)

History of control circuit

Extract signal parameters

(T1-T10) and fit i i i

Define performance indices using

signals distribution to each parameter

parameter distributions

-4 -2 0 2 4 6 80

0.1

0.2

t1 (msec)0 5 10 15 20 25 30

0

0.05

0.1

t2 (msec)

0.04

0.06

0.1

Bayesian approach to update parameter

distribution

Monitored control

circuit data10 20 30 40 50 600

0.02

t3 (msec)45 50 55 60 65 700

0.05

t4 (msec)

55 60 65 70 75 800

0.05

0.1

t5 (msec)

distributioncircuit data

ERCPSAssessment of CB ConditionAssessment of CB Condition

P(t ) i d fi d th b bilit th t th t t f ll i• P(ti) is defined as the probability that the parameter ti falls in the predefined interval, and is given by

• As long as the parameter ‘ti’ falls in the specified interval, it is said that there is no violation with ‘t ’

)Pr()( iiii utltp ≤≤=

is said that there is no violation with ti .

0.3

0.35Probability Between Limits is 0.94016

0.1

0.15

0.2

0.25

Den

sity

pi

8 10 12 14 16 18 200

0.05

Critical Value Lower Upper (msec)

pi

ERCPSPerformance IndicesPerformance Indices

Performance of close/trip coil

Performance of Auxiliary contacts

Performance of breaker• Failure

• Coil• Free Travel

Time

• Contacts• Mechanism

Travel Time

Failure Probability Index

)()()(1)CC( tptptpp )()(1)AB( tt ∏−=6

)(1)Br( if tpp)()()(1)CC( 432 tptptpp f −=

)()(1)( 32 tptpFTp f −=

)()(1)AB( 65 tptpp f −=

)()(1)MT( 53 tptpp f −=

∏=2i

if

ERCPSConcept of Riskp

Event, E• Failure of a component

or group of components• Line, Bus bar, Breaker

Event Probability, p(E)• Control circuit data• Failure probability

Event Consequence, con(E)• Loss of Load (CCDF)

Risk(E)• p(E)*con(E)• Risk associated with

each eventRi k d tiindex

( )• Loss of Line (OPF)• Loss of Generator

(OPF)

• Risk reduction• Maintenance decisions


ERCPSOptimized problem formulation

xRMax

N

N

iiiΔ∑

=1

Where,

iorxCxcSTN

iii ∀=≤∑

=

10:0

i''b ki t ibd tiRi ki''breaker ofcost eMaintenanc:

breakers ofnumber Total : breakeronindex:

RcNi

i

Δbudget Total:

i''breaker maintaningby reduction Risk :C

RiΔ

This optimization problem is a standard Knap-sack problemThis optimization problem is a standard Knap sack problem and can be solved using dynamic programming techniques

ERCPSCase Studies

Category Case study # Details of the dataList of case studies

Maintenance Quantification Model

Case study I CB control circuit data during OPEN operation

Case study II CB control circuit data during CLOSE operation

Case study III Approximation to the BayesianCase study III Approximation to the Bayesian approach in case studies I & II

Risk based maintenance O ti i ti

Case study IV Bus 16 of IEEE Reliability Test S tOptimization System


ERCPSCase Study I: Open Operation

Th f f ti i f• The sequence of occurrence of timing of parameters during opening is: t2-t3-t6-t4-t5. Rename them as y1-y5 in that order

• y1 y2 and y3 can be treated as independenty1, y2 and y3 can be treated as independent.• y4=β0+β1y3+ε4

• y5 = β0 + β1y3 + β2y4+ ε5

Tolerance Limits for Open Operation

Event Lower Upper

Scatter plot analysis of timing parameters

Event Lower(msec)

Upper(msec)

t2 0 2t3 13.6 18.6

26 4 35 4t4 26.4 35.4t5 28.7 38.7t6 22.4 32.4


ERCPSCase Study I: Open Operation

Summary of Analysis for Open Operation

PerformanceIndex

Observations Maintenance required?

pf(TC) Abnormal behavior of Yestrip coil current.

pf(AB) Auxiliary contacts areoperating properly

No

pf(FT) Abnormal free traveltimes. Improper

Yeses. p ope

operation of trip latchmechanism

pf(MT) Abnormal mechanismtravel times. Improper

ti f ti

Yes

operation of operatingmechanism.

pf(Br) Improper operation ofbreaker as a whole

Yes

Performance indices for CB opening


ERCPSCase Study II: Close Operation

• The sequence of occurrence of timing of parameters during opening is: t2-t3-t4-t5-t6. Rename them as y1-y5 in that order

d b t t d i d d t• y1, y2, y3 and y4 can be treated as independent.• y5=β0+β1y4+ε5.

Tolerance Limits for Close Operation

Event Lower(msec)

Upper(msec)

Scatter plot analysis of timing parameters

t2 0 5.5t3 9.8 16.4t4 26 43.4t 49 9 67 5

y g t5 49.9 67.5t6 62 75.8


ERCPSCase Study II: Close Operation

Summary of Analysis for Close Operation

PerformanceIndex

Observations Maintenance required?

pf(CC) Abnormal behavior of Yesclose coil current.

pf(AB) Auxiliary contacts areoperating properly.

No

pf(FT) Abnormal free traveltimes Improper

Yestimes. Improperoperation of close latchmechanism.

pf(MT) Abnormal mechanismtravel times. Improper

Yes

operation of operatingmechanism.

pf(Br) Improper operation ofbreaker as a whole.

Yes

Performance indices for CB closingIAB Meeting, Dec. 4-5, 2008

ERCPSCase Study III: Comparison

CB opening

Comparison of index pf(Br) between Bayesian and Sequential Bayesian approaches

CB closing

ERCPSCase Study IV:

Risk Based System MaintenanceRisk Based System Maintenance

b i• IEEE 24 bus RTS is considered

• Generator = 155MW and LoadGBB1

Bus 15

L24

Generator 155MW and Load = 100MW

• 8 breakers (B1-B8)Bus 14

L23

B1

B2

B4

B5

B7

• Which breaker needs immediate attention?

• How to spend a fixed pool of

Bus 19

L29

B3 B6 B8

• How to spend a fixed pool of money towards the maintenance of these breakers?

BB2

Bus 17

L28

Load

Substation configuration of bus 16IAB Meeting, Dec. 4-5, 2008

ERCPSCase Study IV: List of EventsCase Study IV: List of Events

EventEvent## DefinitionDefinition EventEvent

## DefinitionDefinition EventEvent## DefinitionDefinition## ## ##

E1 Fault on BB1 E15 Fault on L28 E29 Fault on B2, B3 fails

E2 Fault on BB1, B1 fails E16 Fault on L28, B5 fails E30 Fault on B3

E3 Fault on BB1, B4 fails E17 Fault on L28, B6 fails E31 Fault on B3, B6 fails, , ,

E4 Fault on BB1, B7 fails E18 Fault on L29 E32 Fault on B3, B8 fails

E5 Fault on BB2 E19 Fault on L29, B2 fails E33 Fault on B4

E6 Fault on BB2, B3 fails E20 Fault on L29, B3 fails E34 Fault on B4, B5 fails

E7 Fault on BB2, B6 fails E21 Fault on G E35 Fault on B4, B7 fails

E8 Fault on BB2, B8 fails E22 Fault on G, B7 fails E36 Fault on B5

E9 Fault on L23 E23 Fault on G, B8 fails E37 Fault on B5, B6 fails

E10 Fault on L23, B1 fails E24 Fault on B1 E38 Fault on B6

E11 Fault on L23, B2 fails E25 Fault on B1, B2 fails E39 Fault on B6, B8 fails

E12 Fault on L24 E26 Fault on B1, B4 fails E40 Fault on B7

E13 F lt L24 B4 f il E27 F lt B1 B7 f il E41 F lt B7 B8 f ilE13 Fault on L24, B4 fails E27 Fault on B1, B7 fails E41 Fault on B7, B8 fails

E14 Fault on L24, B5 fails E28 Fault on B2 E42 Fault on B8


ERCPSCase Study IV: Event Risk

Risk curvesRisk associated with each

event and breaker


ERCPSCase St d IV: Risk Red ctionCase Study IV: Risk Reduction

Interesting to note that, the f i k d d b 16000

18000

)()()( EConEpERisk ×Δ=Δ

amount of risk reduced by maintaining B6 is less compared to B3 and B8 B3 and B8 should be given 6000

8000

10000

12000

14000

Ris

k R

educ

tion

B3 and B8 should be given priority based on the risk reduction levels 0

2000

4000

R

1 4 7 10 13 16 19 22 25 28 31 34 37 40

Event

For the test system under consideration, it can be concluded For the test system under consideration, it can be concluded th t b k B3 d B8 i t t f ll d b B6th t b k B3 d B8 i t t f ll d b B6that, breakers B3 and B8 are more important followed by B6 that, breakers B3 and B8 are more important followed by B6 and should be given priority in budget allocationand should be given priority in budget allocation

ERCPSSummary of Achievements

• A probabilistic methodology, ‘Maintenance Quantification Model’ is proposed and implemented

• An approximation to the Bayesian approach, called Sequential Bayesian approach is implemented

• Risk based system level maintenance strategy is proposed and implemented


ERCPSFinancial Support

Power Systems Engineering Research Center (Pserc), Project:

“Automated Integration of Condition Monitoring with anAutomated Integration of Condition Monitoring with an Optimized Maintenance Scheduler for Circuit Breakers and Power Transformers”.

Iowa State University: James D. McCalleyV t HVasant Honavar

Texas A&M University: Mladen KezunovicTexas A&M University: Mladen KezunovicChanan Singh


ERCPSPublications

• S Natti and M Kezunovic “Assessing Circuit Breaker Performance Using• S. Natti and M. Kezunovic, Assessing Circuit Breaker Performance Using Condition-Based Data and Bayesian Approach”, IEEE Trans. On Power Systems. (In Review).

• S. Natti and M. Kezunovic, “Risk-Based Decision Approach for Maintenance Scheduling Strategies for Transmission System Equipment Maintenance”, 10th Int. Conference on Probabilistic Methods Applied to Power Systems, Rincon, Puerto Ri M 2008Rico, May 2008.

• M. Kezunovic, E. Akleman, M. Knezev, O. Gonan and S. Natti, “Optimized Fault Location” IREP Symposium 2007 Charleston South Carolina August 2007Location , IREP Symposium 2007, Charleston, South Carolina, August 2007.

• S. Natti and M. Kezunovic, “Model for Quantifying the Effect of Circuit Breaker Maintenance Using Condition-Based Data” Power Tech 2007 LausanneMaintenance Using Condition Based Data , Power Tech 2007, Lausanne, Switzerland, July 2007.

ERCPS• S. Natti and M. Kezunovic, “Transmission System Equipment Maintenance: On-line

Use of Circuit Breaker Condition Data”, IEEE PES General Meeting, Tampa, Florida, June 2007.

• M. Kezunovic and S. Natti, “Risk-Based Maintenance Approach: A Case of Circuit B k C di i B d M i i ” 3 d I i l CIGRE W k hBreaker Condition Based Monitoring”, 3rd International CIGRE Workshop on Liberalization and Modernization of Power Systems, Irkutsk, Russia, August 2006.

• M Kezunovic and S Natti “Condition Monitoring and Diagnostics Using• M. Kezunovic and S. Natti, Condition Monitoring and Diagnostics Using Operational and Non-operational Data”, CMD 2006, Pusan, Korea, March 2006.

• S Natti M Kezunovic and C Singh “Sensitivity Analysis on ProbabilisticS. Natti, M. Kezunovic and C. Singh, Sensitivity Analysis on Probabilistic Maintenance Model of Circuit Breaker”, 9th Int. Conference on Probabilistic Methods Applied to Power Systems, Stockholm, Sweden, June 11-15, 2006.

• S. Natti, P. Jirutitijaroen, M. Kezunovic and C. Singh, “Circuit Breaker and Transformer Inspection and Maintenance: Probabilistic Models”, 8th Int. Conference on Probabilistic Methods Applied to Power Systems, Ames, Iowa, September 2004.

Ri k B d M i tRisk Based Maintenance ERC ScSc edu g o C cu ...

Documents