ML20059C259
| ML20059C259 | |
| Person / Time | |
|---|---|
| Issue date: | 12/30/1993 |
| From: | NRC |
| To: | NRC |
| Shared Package | |
| ML20059C261 | List: |
| References | |
| AEOD-S93-06, AEOD-S93-6, NUDOCS 9401050061 | |
| Download: ML20059C259 (1) | |
Text
q.
i-p.* ~n UNITED STATES a
=!
NUCLEAR REGULATORY COMMISSION Sg.....f W ASHINGTON, D.C. 20555 December 30, 1993 h
Note to: Brian K. Grimes and Distribution
Reference:
AE00/S93-06 Potter & Brumfield Model MDR Rotary Relay failures The above referenced report was forwarded on December 14, 1993. However, it was discovered that several pages were missing from some of the copies.
Will you please check your copy of the report, and if necessary, insert the attached pages.
Attachments: As stated vistrioution:
PDR KRaglin, TTC DCD/ Central' File MTaylor, EDO j
ROAB R/F KNaidu, NRR DSP R/F WTRussell, NRR RSpence GCwalina, NRR Glanik CGrimes., NRR JRosenthal CBerlinger, NRR VBenaroya JBirmingham, NRR EJordan AChaffee, NRR Dross RNorrholm, NRR LSpessard Plewis, INPO SRubin DQueener, NOAC PBaranowsky VChexal, EPRI O
0300~3 f0 9
< 9 9401050061 932230 PDR MISC 9401050062 PDR ROAR ROA nRP /
nn
- vp& *,,
- m e
f **c
'o
- *a
- %)*
UNITED STATES 3-il(N, 3!
NUCLEAR REGULATORY COMMISSION a
y d, /
wassinoron. o.c. osswooi DEC 141993 MEMORANDUM FOR: Brian K. Grimes, Director Division of Operating Reactor Support Office of Nuclear Reactor Regulation FROM:
Gary M. Holahan, Director Division of Safety Programs Office for Analysis and Evaluation of Operations Data
SUBJECT:
POTTER & BRUMFIELD MODEL MDR ROTARY RELAY FAILURES AE00 has recently completed a study of the operating experience associated with Potter & Brumfield (P&B) model MDR rotary relay failures and their potential safety implications. This Special Study Report, 593-06, is enclosed for your information and use, and should be of particular interest to staff and licensee personnel who are involved with Information Notice (IN) 92-04,
" Potter & Brumfield Model MDR Rotary Relay failures." The study found that over 124 P&B MDR relay failures, due to specific failure mechanisms, have occurred in various nuclear power plants (NPPs) between 1984 and 1992 in reactor protection, emergency core cooling, and engineered safety feature systems and caused a wide range of results. About 1/3 of these relay failures occurred in 10, multiple-relay, simultaneous-f ailure events. Five of these events involved simultaneous failures of redundant components, which defeated the single f ailure assumption relied on in nuclear power plant design.
Failures were often not detected until relay operation was tested or demanded and some MDR relays failed to reset af ter testing leaving a believed operable system inoperable. A number of failures were ncnrecoverable, because of specific relay function.
All MDR relays were constructed of the same materials, making each subject to the same failure mechanisms.
Similar failures occurred in ac, dc, latching, and non-latching relays. Most of the failures occurred in normally energized relays, but about 30 percent occurred in normally de-energized relays. While the timing of the relay failures is affected by a number of variables, the failure mechanisms are caused by several specific material or application problems.
P&B has instituted a series of design improvements which address these problems. However, P&B has taken exception to 10 CFR 21-reporting, has not issued a Part 21 report, or made any recommendations to MDR users. The NRC issued a violation to P&B in 1992 for failing to evaluate deviations or informing licensees of the deviations, as required by 10 CFR 21.
This study suggests that a supplement-to NRC IN 92-04 be issued to inform all commercial NPP licensees of the additional MDR relay common-cause failure mechanisms identified since the IN was initially issued.
It also notes that an increase in reliability and a reduction in challenges to safety-related systems could be effected by replacing MDR relays, subject to the dependent failure mechanisms identified in this study, that are relied upon to actuate
~
or operate safety-related systems.
?.;M95pdq# - +@?cm$gqfiggiM&1:3@@rh S ~ @ W:Miggg: y that P&B had ceased to comply with -10 CFR 21 since the GE purchase order did not mention that compliance with 10 CFR 21 was required. P&B had adequately informed-other plants that the requirements of 10 CFR 21 would no longer be accepted, when purchase orders required compliance with 10 CFR 21."
3.11 Waterford Unit 3 On December 8,1991, a P&B MDR-7034 rotary relay used as the EFAS-2 Actuation Relay K-112, which controlled the emergency feedwater turbine steam valve EFAS-2 actuation, failed to drop out of the actuated position ~during a surveillance test. The licensee removed the relay and found the shaft would not rotate either electrically or manually." The MDR-7034 relay had been in service at 32 V dc (28 V de design) in a horizontal shaft configuration.
The relay was hand carried to P&B for disassembly and inspection on December 17, 1991. P&B found varnish outgassing deposits readily visible on the top bell bearing surface, top and bottom spaces, top and bottom shock plates, and rotor assembly bearing surfaces and shims, consistent with those found in similar, previously analyzed MDR failures.
CE also sent HRL one MDR relay for failure analysis. HRL used Energy Dispersive Spectroscopy to identify the foreign material as typical corrosion and coil outgassing i
products, including, chlorine, copper, sulfur, zinc, aluminum, carbon, and oxygen on the lowe. bushing. In addition to these elements, iron, chrome, titanium, calcium, phosphorous, silicon, and magnesium were found on the bearing surfaces on the rotor shaft.
i CE concluded that:
Past analyses on similar failures of these relays has shown that over life, material used in coil construction outgasses due to elevated temperatures.
The outgassed materials (moisture, chlorine, sulfur, etc.) then corrosively attack the metallic components of the relays. The corrosion of the by-products then combine to cffectively "pendrate" the bushings surface and prevent the operatiori of the relay.
It should be noted that some chemical contaminants may have occurred during manufacturing or disassembly of the relay prior to the failure analysis by HRL Therefore, the chemical contaminants given in this analysis may i
include _ elements not resulting directly from the outgassing.
CE recommended that it would be useful to examine additional relays which operate at different temperatures.
On December 17,1991, Waterford 3 personnel scanned all eight ESFAS cabinets with an infrared thermal imaging system, when the plant was at 100 percent power and the MDR relays.were in their normal energization states. The operational and spare relays were 9401050070 931230 PDR MISC 19
'9401050061 PDR
yhggg p p g.
- g p g g y p g yy g s p g g qr g g yg sqr mounted horizontally in two vertical columns of up to 9 relays with wires wrapped around them. In each cabinet, the hottest relays were in the upper half; the hottest relays found ranged from 147* F to 152 F.
Four additional operational MDRielays (-7034 [two), -7033, and -7032) were removed and evaluated by CE and HRL to determine the effect of air temperature and voltage.
Two high-temperature and two low-temperature relays were selected based on the thermographs provided by Entergy Operations. Selection of the relays by using thermal imaging failed to reveal a correlation of relay damage with temperature.
Three of the relays did not meet original electrical performance specifications. An MDR-7034 and an MDR-7032 relay were the most severely degraded in performance.
Both required greater than 24 V to latch and less than 3 V to drop out, yet the MDR-
+
7034 had the most foreign material in the motor cavity while the MDR-7032 had the least in the 4 relays. The contacts on the MDR-7032 reopened between the initial closure and latching.
Inspection of all 4 relays found outgassing products and scoring on the bearing surfaces.
One relay's contacts discoloration was found to have been caused by sulfur, an outgassing product. The MDR-7033 relay had a moist, paste-like foreign substance rather than the dry dust noted in the other relays. Scanning electron microscope examination / energy dispersive spectrometry found carbon, oxygen, sodium, calcium, potassium, zinc, silicon, sulfur, chlorine, copper, iron, and chrome in varying amounts on the shaft bearing surfaces.
On May 1,1992, CE concluded that this and similar analyses showed a tendency for the coil varnish to outgas over the life of the relay. They could not determine which environment or electrical conditions would be more likely to exhibit outgassing that could eventually result in failures of MDR relays. CE recommended that Waterford 3 change to newer MDR relays, which alleviate the problem of outgassing and consider increasing the surveillance testing frequency for those MDR relays that could be tested without interruption of service.
On October 2,1992, an event occurred at Waterford Unit 3 that shows the potential affect of an MDR relay failure on multiple systems. While Waterford Unit 3 was in refueling, an improperly placed electrical jumper, during replacement of an MDR relay, resulted in the de-energization of 10 other ESFAS actuation relays, which affected the component cooling water, chilled water and safety injection systems.2 Voltage checks made just after the event showed some voltage on the terminal board downstream of the break in the circuit, but apparently not enough to prevent the relays from de-energizing.
Part of the alligator clip was attached no more than one eighth of an inch out of position onto the terminal lug insulator.
i 20
i en a,,
I pm nc<
h O
i)
( lf' t
(
S UNITED STATES 3
- :j NUCLEAR REGULATORY COMMISSION
%Q,, f y' s, msmorou. o.c. mss.am DEC 141933 MEMORANDUM FOR: Brian K. Grimes, Director Division of Operating Reactor Support Office of Nuclear Reactor Regulation FROM:
Gary M. Holahan, Director Division of Safety Programs Office for Analysis and Evaluation of Operations Data
SUBJECT:
POTTER & BRUMFIELD MODEL MDR ROTARY RELAY FAILURES i
AE00 has recently completed a study of the operating experience associated with Potter & Brumfield (P&B) model MDR rotary relay failures and their potential safety implications.
This Special Study Report, S93-06, is enclosed for your information and use, and should be of particular interest to staff and licensee personnel who are involved with Information Notice (IN) 92-04,
" Potter & Brumfield Model MDR Rotary Relay Failures." The study found that over 124 P&B DR relay failures, due to specific failure mechanisms, have occurred in various nuclear power plants (NPPs) between 1984 and 1992 in reactor protection, emergency core cooling, and engineered safety feature systems and caused a wide range of results. About 1/3 of these relay failures occurred in 10, multiple-relay, simultanaous-failure events.
Five of these events involved simultaneous failures of redundant components, which defeated the single failure assumption relied on in nuclear power plant design.
Failures were often not detected until relay operation was tested or demanded and some MDR relays failed to reset af ter testing leaving a believed operable system inoperable. A number of failures were nonrecoverable, because of specific relay function.
All MDR relays were constructed of the same materials, making each subject to the same failure mechanisms.
Similar failures occurred in ac, dc, latching.
and non-latching relays. Most of the failures occurred in normally energized relays, but about 30 percent occurred in normally de-energized relays. While the timing of the relay failures is affected by a number of variables, the failure mechanisms are caused by several specific material or application problems.
P&B has instituted a series of design improvements which address these problems. However, P&B has taken exception to 10 CFR 21 reporting, has not issued a Part 21 report, or made any recommendations to MDR users.
The NRC issued a violation to P&B in 1992 for failing to evaluate deviations or informing licensees of the deviations, as required by 10 CFR 21.
This study suggests that a supplement to NRC IN 92-04 be issued to inform all commercial NPP licensees of the additional MDR relay common-cause failure mechanisms identified since the IN was initially issued.
It also notes that an increase in reliability and a reduction in challenges to safety-related systems could be effected by replacing MDR relays, subject to the dependent failure mechanisms identified in this study, that are relied upon to actuate or operate safety-related systems.
~
\\[ h CY-8 [.
~
j
~
Brian Grimes.
If you have any questions regarding the enclosed study, please contact Bob Spence on 492-8609.
Please contact me if AE0D can provide any additional assistance.
Original signed by Jack E. Rosenthal for:
Gary M. Holahan, Director Division of Safety Programs Office for Analysis and Evaluation of Operational Data
Enclosure:
As stated Distribution:
PDR KRaglin, TTC DCD/ Central File MTaylor, EDO ROAB R/F KNaidu, NRR DSP R/F WTRussell, NRR RSpence GCwalina, NRR Glanik CGrimes., NRR JRosenthal CBerlinger, NRR VBenaroya JBirmingham, NRR EJordan AChaffee, NRR Dross RNorrholm, NRR LSpessard Plewis, INP0 SRubin DQueener, NOAC PBaranowsky VChexal, EPRI ROAB R0" DSP S
00' RSpence:mmk Gla..ik JRosenthal VB aroya GHolahan 12/y'/93 12/4/93
/ /93 12//f/93 12hy/93 f