experience with use of c57.139 ltc dga...
TRANSCRIPT
Experience With Use of C57.139 LTC DGA GuideIEEE TRANSFORMER COMMITTEE MEETINGS JOHN PRUENTEOCTOBER 23, 2011 DIR. ENG. TECHNICAL SUPPORT
© SPX Transformer Solutions, Inc. 2012
Outline of Presentation
• Background of UZD DGA Study
• Summary of Data Reviewed
• Results of Study to Date
• Suggestions for Revisions / Additions to C57.139 Guide for Consideration
2
UZD DGA Study ‐ Background• Began with request from Midwest utility in March 2010• Too many “false positives” with code 4 DGA results from oil lab.• Utility needed a better way to analyze and use DGA data as a predictive
maintenance tool.• Grown to include 5 utilities located in different regions around the United
States:– Northeast – North Central – South Central– North Western– South Western
• Utility selection criteria– Have 50+ UZD’s in service– Sample LTC for DGA annually (minimum)– Have historical DGA database to work with
3
UZD DGA Study Objectives• Determine calculated DGA limits for each utility population based on the
C57.139 guide applied to the historical data.• Verify limits by correlating DGA diagnosis to maintenance inspection “as
found “ condition.• Develop SPX Waukesha endorsed universal UZD DGA criteria for use by
UZD owners (if possible).• Develop a more consistent approach and analysis method (algorithm) for
UZD DGA. • Investigate “false positives” and identify a way to recognize them at the
analysis stage.• Identify most effective next steps to be taken when a UZD is outside the
norms.• Identify which maintenance activities are most critical for the UZD model.• Learn as much as possible about operational performance of the UZD
model LTC.
4
Results to Date
• Represents 12% of Population (approx. 6500 units sold)
5
Utility Geographic Region # of UZD's # Samples Oldest Sample
A North Central US 134 780 Apr‐95
B South Central US 265 1988 Jan‐00
C Northeastern US 57 444 Jan‐91
D Northwestern US 212 2197 Apr‐95
E Southwestern US 116 930 Nov‐02
784 6339
Table 1 : Summary of UZD Populations by Utility in Study
Results to Date
6
Maintenance Item Utility A Utility B Utility C Utility D Utility E
# of Operations N/A N/A50K / 100K***
or 12 yrs50K**
Time Based 3 yrs
DGA ResultCode 4 or 4*
TASABased on Internal
CriteriaBased on
Internal CriteriaCode 3, 4, or 4*
TASABased on
Internal Criteria
Oil Sample Rate Annual* Annual* Annual* Annual* Annual*
Infra Red Scan Annual* Bi‐Annual* As Needed Annual* As Needed
Oil Replaced at Maintenance YesReuse (Large Vol)Replace (Small
Vol)yes
≈50% based on oil condition
Yes
Oil Filter Operation 24/7Based on LTC Operation
Periodic(Time Based)
24/7Periodic
(Time Based)
Oil Filter Cartridge Replacement
Pressure Differential
Pressure Differential or
AnnualAnnual
Pressure Differential
Pressure Differential
Operate LTC Through Neutral If Possible6 Months if Possible
If Possible3,6,9, or 12
months based on Model
QuarterlyIf Possible
Table 2: Comparison of LTC Maintenance Criteria by Utility in Study* ‐More frequent if problem suspected or critical unit** ‐ 75K if temperature differential monitor installed*** ‐ 100K for vacuum type
Results to Date
7
Utility A Utility B Utility C Utility D Utility EEquipment Population Size : 134 Equipment Population Size : 265 Equipment Population Size : 57 Equipment Population Size : 212 Equipment Population Size : 116
Variable NAlert Limit (95%)
Alarm Limit (99%)
U1 U2 Variable NAlert Limit (95%)
Alarm Limit (99%)
U1 U2 Variable NAlert Limit (95%)
Alarm Limit (99%)
U1 U2 Variable NAlert Limit (95%)
Alarm Limit (99%)
U1 U2 Variable NAlert Limit (95%)
Alarm Limit (99%)
U1 U2
h2 497 1375 1952 h2 1509 502 842 h2 399 271 390 h2 1770 697 1001 h2 930 331 492
ch4 456 750 1106 ch4 1303 134 231 ch4 379 103 148 ch4 1527 266 381 ch4 930 156 229
c2h6 375 552 816 c2h6 616 35 123 c2h6 109 84 121 c2h6 1148 97 137 c2h6 930 37 56
c2h4 506 2540 3758 c2h4 1466 479 790 c2h4 396 448 639 c2h4 1774 767 1095 c2h4 930 575 860
c2h2 506 6272 9479 c2h2 1573 3724 5874 c2h2 406 2496 3513 c2h2 1817 4171 5979 c2h2 930 5850 8700
THG 367 4194 6083 THG 588 1121 1542 THG 108 1565 2218 THG 1104 1232 1704 THG 878 761 1123
water 505 58 77 water 1967 63 91 water 432 40 55 water 1827 56 76 water 833 46 61
fluidtempc n/a n/a n/a fluidtempc 717 61 82 fluidtempc n/a n/a n/a fluidtempc 1589 53 68 fluidtempc 912 61 79
c2h4/c2h2 525 0.78 1.34 0.88 1.34 c2h4/c2h2 1374 0.19 0.22 0.21 0.26 c2h4/c2h2 358 0.22 0.27 0.21 0.26 c2h4/c2h2 1688 0.36 0.49 0.44 0.63 c2h4/c2h2 909 0.123 0.143 0.136 0.173
n2/o2 544 5.07 9.08 6.46 8.99 n2/o2 1717 4.95 5.86 5.21 6.48 n2/o2 412 2.62 2.81 5.1 7.01 n2/o2 1850 17.77 29.35 18.3 27.6 n2/o2 530 3.69 5.31 3.56 4.38
THG/C2H2 368 1.14 1.97 THG/C2H2 547 0.35 0.44 THG/C2H2 93 0.52 0.69 0.87 1.22 THG/C2H2 1088 0.771 1.133 0.73 1.05 THG/C2H2 130 0.221 0.264 0.26 0.34
Table 2: Comparison of Calculated Limits for UZD Populations by Utility
• No universal limits
• Some population limits extremely low , should not be applied
Results to Date
• Wide variation in limits between utilities– Differences in maintenance practices
• Length of maintenance cycle
• Replace vs. reuse oil
• Use of on line filtration
• Desiccant breather maintenance
– Transformer loading
– Frequency of operation
8
Results to Date
9
• False Positive “Maverick” Units Identified– Ethylene / Acetylene ratio outside calculated limits (>1)
– First noticed at Utility A (< 7% of population)
– Indicative of “Severe Contact Heating” classical DGA diagnosis
– IR Scans while in service and internal inspections revealed no issues
– Further research needed to identify source of elevated Ethylene
– Trend using Duval Triangle 2² to verify stability
“Maverick” Unit “Normal” Unit Unit in “Fault Evolution”
² ‐ Source IEEE Magazine Nov / Dec 2008 – Vol. 24, No.6
UZD Best Practices
10
• For units flagged outside the DGA norms established – Retest oil to verify trend and rate of change
– Check Duval Triangle 2² trend
– IR scan to check LTC tank temp and look for localized heating
– Check N2/O2 to verify breather open (ratio should be < 4)
– Check for recent change in loading and / or frequency of operation
– Review past maintenance history for previous problems (repeatingproblem)
• After performing maintenance– Replace old oil with new , clean, filtered oil
– Sample after oil filling and before energization (benchmark)
– Re‐sample again within one week of energization to verify no issues
² ‐ Source IEEE Magazine Nov / Dec 2008 – Vol. 24, No.6
Summary
11
• Universal DGA limits for UZD LTC’s not possible
• C57.139 Guide works well for given utility / LTC model population
• Application of U1 / U2 limits good start for alert and alarm levels
• Application 95% / 99% limits for key ratios worked well for identifying units outside the norm and with real problems
• N2/O2 ratio on free breathing LTC’s works well for identifying breathers in need of maintenance
• Identified “Maverick” gassing pattern for UZD’s and method of recognizing it
• Use of Duval Triangle 2² for trending LTC’s headed towards failure very useful
² ‐ Source IEEE Magazine Nov / Dec 2008 – Vol. 24, No.6
Suggested Changes
12
• Add suggestion on what to do when calculated limits are too low– Suggest using C2H4/C2H2 boundary limits of Duval Triangle 2:
• <.35 – normal and in “N” Zone continue operation• > .35 and <1.2 – increase sample frequency / trend (Quarterly)• >1.2 and < 4.0 ‐ increase sample frequency / trend ( Monthly / Weekly)• >4.0 – Consider outage and inspection
– And/or apply Stenestram Ratio¹ limits for UZD, UZE, UZF, or UZG models
• Encourage utilities to correlate calculated limits to LTC inspection observations for refinement of initial limits
¹ ‐ Source ABB Info No. IZSC000498‐AAA en, dated 6‐21‐2004² ‐ Source IEEE Magazine Nov / Dec 2008 – Vol. 24, No.6
Duval Triangle 2 With Suggested C2H4 / C2H2 Limits²
Suggested Changes • Add use of Duval Triangle 2 for trending , possibly include case study example in Annex C
• Consider adding mention of importance of N2/O2 ratio for Free breathing LTC’s or breath through a desiccant (dryer).
– Ratio > 8 indicate plugged air path with high degree of confidence
• Add recommendation on how often limits should be recalculated and updated.– Suggest every 5 years or after any major maintenance initiatives involving multiple units of a given
model are completed
• Add recommendation for replacing oil in LTC when maintenance performed– Due to length of time between maintenance intervals– Filtering used oil does not remove gasses, can cloud after maintenance benchmark
• Recommend “sanity” limits for individual gasses set at 5 – 10 X U2 calculated limit to catch cases where :
– Breather is plugged and gas concentrations increase significantly– Excessive LTC operations due to faulty controller– Mechanical binding causing slow operation and longer than normal arc duration
13