25
DoE Peer Review Nov 2010 1 Cable Diagnostic Focused Initiative National Electric Energy Testing Research Application Center (NEETRAC) PI: Rick Hartlein November 2010
DoE Peer Review Nov 20101
Cable Diagnostic Focused Initiative
National Electric Energy Testing Research Application Center (NEETRAC)
PI: Rick Hartlein
November 2010
DoE Peer Review Nov 20102
• Underground cable system infrastructure is complex and aging.
• Failures are increasing• If not addressed then old
infrastructure will not support future operation of the grid.
• Not enough money / manufacturing capacity to simply replace because they are old.
• Need diagnostic tools to prioritize Active Asset Management.
• Some tools are available, but there is significant mistrust and commercialism that has limited their effective deployment.
Ca
ble
Fa
ilu
re
s p
er Y
ea
r
20052000199519901985198019751970
1000
800
600
400
200
0
Why do we need diagnostics?
DoE Peer Review Nov 20103
CDFI TeamNEETRACJorge AltamiranoTim AndrewsYamille del Valle*Bryan DavantStacy ElledgeBarry Fairley
Nigel Hampton (Co-PI)Rick Hartlein (PI)Thomas ParkerJoshua Perkel*Dean Williams
Georgia Tech - ECEMiroslav BegovicRon HarleyJ.C. Hernandez*Salman Mohagheghi
IREQJean-Francois Drapeau
*PhD supported by CDFI
DoE Peer Review Nov 20104
NEETRAC Members
Non NEETRACMembers Supporters
Dept of Energy
Diagnostic Providers
CDFI
CDFIPartners
• 13 Electric Utilities
• 5 Manufacturers
• 6 Diagnostic Providers
• DOE: $1,700,000
• Cost Chare: $1,275,000
• Total: $2,975,000
DoE Peer Review Nov 20105
Participants
American Electric Power HV TechnologiesAmeren Hydro QuebecCablewise / Utilx IMCORPCenterPoint Energy NRECAConsolidated Edison Oncor (TXU)Cooper Power Systems PacifiCorp (added mid 2005)Duke Power Company Pacific Gas & Electric (added Jan 06)Exelon (Commonwealth Edison & PECO) PEPCOFirst Energy PrysmianFlorida Power & Light Public Service Electric & GasGeorgia Tech Southern California EdisonGRESCO Southern CompanyHDW Electronics SouthwireHigh Voltage, Inc. Tyco/RaychemHV Diagnostics
DoE Peer Review Nov 20106
CDFI Activities
CDFI
Analysis Lab Studies
Field Studies Dissemination
DoE Peer Review Nov 20107
CDFI Activities
FieldStudies
Georgia Power XLPE
Jkt & UnJkt24 Conductor Miles
DukeXLPE & Paper
Jkt & UnJkt29 Conductor Miles
Offline PD (0.1Hz)Offline PD (60Hz)
Tan δMonitored Withstand
Offline PD (0.1Hz)Tan δ
Monitored Withstand
Charlotte * 2CincinnatiClemson
Morresville
EvansMacon
Roswell * 3
Alabama Power Paper & XLPE
Jkt & UnJkt7 Conductor Miles
Offline PD (0.1Hz)Tan δ
Monitored Withstand
BirminghamMontgomery
DoE Peer Review Nov 20108
Diagnostic Data Obtained from Many Sources
PPL*
PGE
PEPCO
NETA*
ONCOR
Keyspan* InterMountain*
FPL
Duke
Com Ed
ConEd
Ameren
Alabama Power
AEP
Utility Data
* Provided by Non Participating Companies
DoE Peer Review Nov 20109
Significant Data Gathered
Data Type Technique Laboratory[Conductor miles]
Field[Conductor miles]
Diagnostic
DC Withstand - 78,105
Monitored Withstand 1.8 260
PD Offline 4.8 490
PD Online 5 262
Tan δ 4.3 640
VLF Withstand 4.6 9,900
IRC 0.3 -
Service Performance ALL 89,000
DoE Peer Review Nov 201010
Diagnostic Testing Program(Approach is Important! - SAGE)
Failures [#]
Time
Selection
Action
Generation
Evaluation
Decreasing Failures
Increasing Failures
DoE Peer Review Nov 201011
Log Time (Days)
Log
Cum
ulat
ive
Failu
res
200015001000750
3000
2000
1500
1000
750
500
250
100
FAILURETEST
Type
Cable System Phases - Actual Case
Pre-Diagnostic ProgramProgram Start Up
Full Program
Benefits apparent but are often subtle.
DoE Peer Review Nov 201012
Interpreting Diagnostic Data – What we believed to be true was wrong!
(Partial Discharge Example)
5 years after test, red = failed, green = no failure
Partial Discharge Diagnostic Features
Sim
ilari
ty L
evel
[%
]
Nw [p
ulses
/cycle
]
Mean E
nergy
RatioD
Neg.
Mean P
hase
[deg
]
Neg.
Phas
e Ran
ge [d
eg]
Neg.
Mean E
nergy [
pC*k
V]
Neg.
Max En
ergy
[pC*
kV]
Pos.
Max En
ergy
[pC*
kV]
Pos.
Mean E
nergy [
pC*k
V]
Pos.
Qmea
n [pC
]
Neg.
Qmea
n [pC
]
Neg.
Qmax
[pC]
Pos.
Qmax
[pC]
Pos.
Mean P
hase
[deg
]
Pos.
Phas
e Ran
ge [d
eg]
15.18
43.45
71.73
100.00
50 % Similarity Level
1 2 3 4 5 6 7
3a 3b
Cluster Variable Analysis
DoE Peer Review Nov 201013
Interpreting Diagnostic Data(Tan δ)
Tip Up (1e-3)
Tan
Del
ta (
1e-3
)
-1010-1-
-
150
6
0
Unfilled Polyolefin Insulations
No Action
Further Study
Action Required
DoE Peer Review Nov 201014
Defining Accuracy:Ability to Predict Failures
No Action Required Action Required
Year12345
DoE Peer Review Nov 201015
Accuracy – Failures over Time
Time[Years]2 4 6 108
Accuracy[%]
100
0
No Action Required Accuracy
Action Required Accuracy?
• System Changes• Additional Aging• Increased Load
DoE Peer Review Nov 201016
Overall AccuracyNo Action AccuracyAction Accuracy
100
80
60
40
20
0
Dia
gnos
tic
Acc
urac
y [%
]
Overall AccuracyNo Action AccuracyAction Accuracy
100
80
60
40
20
0
Dia
gnos
tic
Acc
urac
y [%
]All Accuracies
Overall AccuracyNo Action AccuracyAction Accuracy
100
80
60
40
20
0
Dia
gnos
tic
Acc
urac
y [%
]
Overall AccuracyNo Action AccuracyAction Accuracy
100
80
60
40
20
0
Dia
gnos
tic
Acc
urac
y [%
]
Overall AccuracyNo Action AccuracyAction Accuracy
100
80
60
40
20
0
Dia
gnos
tic
Acc
urac
y [%
]
DoE Peer Review Nov 201017
Accuracy – Probabilistic Approach (Partial Discharge Example)
DoE Peer Review Nov 201018
Accuracy – Probabilistic ApproachTan δ Example
101FOT
40
30
20
10
5
3
2
1
Elasped Time between test and failure in service at May 09 (Month)
Serv
ice
Failu
res
[% o
f Te
sted
]
24
32
10
1.9ACTION REQUIREDFURTHER STUDYNO ACTION
Action
Classes based on CDFI data
DoE Peer Review Nov 201019
VLF Withstand – Effectiveness & Application Time
100010010
70
60
50
40
30
20
10
5
3
2
1
T ime to Failure [Days since T est]
Se
rvic
e F
ail
ure
s [%
of
Te
sts]
2 Y
ears
32.5%
17.0%
PassFOT
ResultTest
100010010
70
60
50
40
30
20
10
5
3
2
1
T ime to Failure [Days since T est]
Se
rvic
e F
ail
ure
s [%
of
Te
sts]
2 Y
ears
28.3%
3.2%
PassFOT
ResultTest
2.5 Uo, 15 min 1.8 Uo, 30 minutes
DoE Peer Review Nov 201020
Dissemination1. First practical utility implementations of Monitored Withstand Diagnostics in the USA; Chris L Fletcher, Nigel Hampton, Jean Carlos
Hernandez, Jeff Hesse, Michael G Pearman, Joshua Perkel, C Tim Wall, Walter Zenger; submitted to International Conference on Insulated Power Cables JICABLE11, Versailles France, June 2011; Abstract # 9
2. Challenges associated with the interpretation of dielectric loss data from power cable system measurements; J. Perkel, J.C. Hernández, R. N. Hampton, J. F. Drapeau, J. Densley; submitted to International Conference on Insulated Power Cables JICABLE11, Versailles France, June 2011; Abstract # 6
3. Application Of Artificial Intelligence To The Problem Of Selecting The Appropriate Diagnostic For Cable Systems; Yamille Del Valle, Nigel Hampton; submitted to International Conference on Insulated Power Cables JICABLE11, Versailles France, June 2011; Abstract # 3
4. Cable Fleet Management; RN Hampton, M Olearczyk, J Perkel, N Weisenfeld; IEEE Spectrum; Nov 20105. Experience of Withstand Testing of Cable Systems in the USA; Hampton, R.N..Perkel. J., Hernandez, J.C., Begovic, M., Hans, J., Riley, R.,
Tyschenko, P., Doherty, F., Murray, G., Hong, L., Pearman, M.G., Fletcher, C.L., and Linte, G.C.; CIGRE 2010, Paper No. B1-3036. Characterization of Ageing for MV Power Cables Using Low Frequency Tan-delta Diagnostic Measurements; JC. Hernandez-Mejia, RG.
Harley, RN Hampton, RA Hartlein; IEEE Transactions on Dielectrics and Electrical Insulation, Vol. 16, Issue 3, pp. 862-870, June 2009. 7. Determining Routes for the Analysis of Partial Discharge Signals Derived from the Field; Hernández-Mejía, J.C.; Perkel, J.; Harley, R.;
Begovic, M.; Hampton, N.; and Hartlein, R.; IEEE Trans. on Dielectrics and Electrical Insulation, December 2008, pp. 1517-1525.8. Correlation between Tan δ Diagnostic Measurements and Breakdown Performance at VLF for MV XLPE Cables; Hernández-Mejía, J.C.;
Perkel, J.; Harley, R.; Hampton, N.; and Hartlein, R.; IEEE Trans. on Dielectrics and Electrical Insulation, February 2009, pp. 162-1709. Some Considerations on the Selection of Optimum Location, Timing, and Technique, for Diagnostic Tests, RA Hartlein, RN Hampton & J
Perkel; IEEE Power Engineering Society (PES) General Meeting Panel Session Pittsburg 200810. Characterization of Aging in Medium Voltage Power Cables Using Low Frequency Tan-delta Diagnostics Features R.N. Hampton, R. Harley,
R. Hartlein & J.C. Hernandez; IEEE Transactions in Power Delivery; submitted 11. Validation of the accuracy of practical diagnostic tests for power equipment; M. Begovic, RN. Hampton*, R. Hartlein, J.C. Hernandez-Mejia,
and J Perkel; CIGRE 2008 Paris Study Committee D1 Paper 205 12. On Distribution Asset Management: Development of Replacement Strategies; Miroslav Begovic, Joshua Perkel, Nigel Hampton, Rick Hartlein;
IEEE PES PowerAfrica 2007 Conference and Exposition; Johannesburg, South Africa, 16-20 July 2007 13. Practical Issues Regarding The Use Of Dielectric Measurements To Diagnose The Service Health Of MV Cables; R.N. Hampton, R. Harley,
R. Hartlein & J.C. Hernandez; International Conference on Insulated Power Cables; JICABLE07, Versailles France, June 2007 14. Validating Cable “Diagnostic Tests”; M Begovic, RN Hampton, R Hartlein, J Perkel; International Conference on Insulated Power Cables;
JICABLE07, Versailles France, June 2007 • Periodic Update Meetings throughout the project• Regional Meetings – San Ramon, CA, Atlanta, GA, Columbus, OH, New York, New York, IEEE Education Session, St. Petersburg,
FL2009/2010
DoE Peer Review Nov 201021
CDFI - At the Beginning
• For many utilities, the usefulness of diagnostic testing was unclear.
• The focus was on the technique, not the approach.
• The economic benefits were not well defined.
• There was almost no independently collated and analyzed data.
• There were no independent tools for evaluating diagnostic effectiveness.
DoE Peer Review Nov 201022
What We Now Know (1)
1. Diagnostics can work – they tell you many useful things, but not everything.
2. Diagnostics do not work in all situations.3. Diagnostics have great difficulty definitively determining
the longevity of individual devices. 4. Utilities HAVE to act on ALL replacement & repair
recommendations to get improved reliability.5. The performance of a diagnostic program depends on:
• Where you use the diagnostic• When you use the diagnostic• What diagnostic you use• What you do afterwards
DoE Peer Review Nov 201023
6. Quantitative analysis is complex BUT is needed to clearly see benefits.
7. Diagnostic data require skilled interpretation to establish how to act.
8. No one diagnostic is likely to provide the detailed data required for accurate diagnoses.
9. Large quantities of field data are needed to establish the accuracy/limitations of different diagnostic technologies.
10. Important to have correct expectations – diagnostics are useful but not perfect!
What We Now Know (2)
DoE Peer Review Nov 201024
Reflections
• Approach to data analysis established in CDFI• Standards upgraded (IEEE 400 series)• Many questions answered, gaps remain:
– Defining the Benefits– Identifying anomalies that lead to failure
• Answers will come with continued analysis of field test data (Diagnostic tests with circuit performance monitoring).
• The potential value of continued analysis is high• New approaches appear promising
– Monitored withstand (HV withstand + tan δ or partial discharge)– Combined diagnostics (simultaneous tan δ and partial discharge)– New technologies (oscillating wave, cosine VLF withstand)
DoE Peer Review Nov 201025
CDFI Phase 1 / CDFI Phase 2
Element CDFI Focus, Phase I
CDFI Focus, Phase II*
Voltage Level MV MV & some HV
Test Type Condition AssessmentCondition Assessment &
Commissioning / Recommissioning
Cable Service Aged Service Aged & Laboratory Testing of Service Aged
Diagnostics Currently in use in USCurrently in use in US &
those that might reasonably be used
Data Utility Distribution System Distribution, Industrial & Transmission
Lab Studies Field Aged Cable Cable & Accessories*Approved in July of 2010
