ML20126E290
| ML20126E290 | |
| Person / Time | |
|---|---|
| Issue date: | 06/07/1985 |
| From: | Thompson H Office of Nuclear Reactor Regulation |
| To: | Harold Denton Office of Nuclear Reactor Regulation |
| References | |
| REF-GTECI-B-56, REF-GTECI-EL, TASK-B-56, TASK-OR GL-84-15, NUDOCS 8506170037 | |
| Download: ML20126E290 (3) | |
Text
JUN 0 71985 MEMORANDUM FOR:
Harold Denton, Director Office of Nuclear Reactor Regulation FROM:
Hugh L. Thompson, Jr., Director Division of Licensing
SUBJECT:
EMERGENCY DIESEL GENERATOR RELIABILITY---
MEETING WITH MANUFACTURERS On April 30, 1985, the NRC met with the manufacturers of emergency diesel generators (EDGs) for nuclear application to gain authorative information regarding assessing and assuring high reliability for the EDGs. The staff provided clarification on the intent of Generic Letter 84-15. Nuclear utility representatives attendino the meeting provided input to the discussions.
As you requested, J.T. Beard of ORAB/DL has developed the enclosed summary of the meeting based upon his direct participation and a review of the trans-cript. The single most significant point seems to be that EDG reliability might be improved if the utilities were motivated to form owner / user groups to establish and maintain effective communications with each designer /
manufacturer.
J.T. also suggests that, in the on-going B-56 program, we consider deleting the present accelerated routine testing reouirements (Reg. Guide 1.108) in favor of routine testing on a fixed monthly basis plus a failure-avoidance type of EDG reliability enhancement program at each plant. Such a program should focus on: early identification of potential problem areas (performance data trend analysis), root cause analysis of failures to avoid recurrence of similar failures, and evidence of effective corrective actions.
The NRC program for improving EDG reliability (Generic Issue B-56) is being managed by Les Rubenstein, DSI. We understand that DSI plans to have Brookhaven National Laboratory review and develop a coordinated set of recomendations based on the information from this meeting and information from other sources, such as the responses to Generic Letter 84-15.
[OrigbM Signed by
- }lugh L The, w v,-
Hugh L. Thomason, Jr., Director Division of.icensing CONTACT:
DISTRIBUTION J.T. Beard, ORAB NRC POR MSrinivasan P. [M '
E X27465 JHannon EMershoff H,t3cc k
(, @e drugs k
FBrenneman RHernan cc:
L. Rubenstein ossi RSingh
,I G. Zech Baranowsky 0.CvoMd ORAB:DL SL:
A L
AD DD/D G r
~
JTBea rd.'cl RWessman 01 Gp DCr ch ield TML a
4 /o4/85 4/f/85
/ /5 6/6/85
/
/
D/D HTh on 8506170037 850607 Q/}/85 PDR 9ttCI GELB-5
SUMMARY
OF EDG MANUFACTURERS MEETING---APRIL 30, 1985 EDG manufacturer's customer is the architect-engineer, not the end user (thelicensee).
.~
No effective communications link exists between manufacturer and end user, except in crisis situations.
EDG manufacturers do not get feedback on failures and are not staffed to assess reliability in a quantitative fashion.
Code requirements may reduce reliability of design, e.g. not allowing the use of components for which there are thousands of hours of successful experience.
EDG designs are basically highly.eliable; the problem is on the people side of equation.
1 Increasing the number of test starts and accumulating statistics is not a l
good gauge of reliability---it may be detrimental.
Best reliability gauge may be trending of performance parameters, (e.g cylinder pressure); trending can also identify potential problems before they cause failures.
Preventative maintenance needs to be improved.
Manufacturers maintenance instructions are often written presuming too much prior knowledge, i.e. for experienced engineers vs. novice mechanics.
Schedule for preventative maintenance activities should be upgraded to reflect the high-start, low-hours nuclear application.
High turnover of maintenance personnel may contribute to problems.
Need better training of maintenance personnel.
Routine testing should be monthly via start to idle speed, slow increase to full speed, load slowly over several minutes.
For relatively slow speed engines (450-600 rpm), an intermediate idle speed start may not be desirable; go straight to full speed.
Fast-start, fast-load tests should be conducted periodically. Average frequency suggested by manufacturers is twice/ year.
Every test start should be followed by load run at 75-100% for at least one hour, preferably two hours, with data taken for trend analysis.
_2_
When failures do occur, there should be a root-cause analysis to assure effective and comprehensive repairs and to avoid recurrence; the analysis should involve engineers who know what design considerations are.
Good diagnostic equipment may promote better troubleshooting and repair actions.
It is important to pay close attention to details, e.g. sources of water in air start system.
Tech. Spec. outage allowances of 3 or 7 days may not be sufficient for good maintenance in all cases.
The number of experienced mechanics may limit repairs to single-shift operation at some plants.
~ --,_,
4
.