ML17250B357
| ML17250B357 | |
| Person / Time | |
|---|---|
| Issue date: | 10/18/1989 |
| From: | Mike Williams NRC OFFICE FOR ANALYSIS & EVALUATION OF OPERATIONAL DATA (AEOD) |
| To: | Novak T NRC OFFICE FOR ANALYSIS & EVALUATION OF OPERATIONAL DATA (AEOD) |
| References | |
| NUDOCS 8910310146 | |
| Download: ML17250B357 (4) | |
Text
pI'~0 fy p t~
n C
O 0 Q ly
++*++
i UNITED STATES NUCLEAR REGULATORY COMMISSION WASHINGTON, D. C. 20555 OCT $ 8 8 MEMORANDUM FOR:
Thomas H. Novak, Director Division of Safety Programs Office for Analysis and Evaluation of Oper ationa1 Data FROM:
SUBJECT:
Mark H. Williams, Chief Trends and Patterns Analysis Branch Division of Safety Programs Office for Analysis and Evaluation of Operational Data
SUMMARY
OF OCTOBER 13, 1989 MEETING OF MAINTENANCE INDICATOR DEMONSTRATION PROJECT On October 13, 1989, representatives from the Nuclear Management Resources Council (NUMARC), the Institute for Nuclear Power Operations
( INPO), and six nuclear uti lities met with It~ to discuss the staff's Maintenance Effectiveness Indicator (MEI).
This was the second meeting of the Maintenance Indicator Demonstration Project and was held in the Maryland National Bank Building, 7735 Old Georgetown
- Road, Bethesda, Maryland.
A list of attendees is attached.
Because of travel considerations for some industry participants, the length of the meeting was limited to three hours.
Consequently, the industry presented only highlights of the results of their review of the MEI data (which they had received from the staff at the previous meeting on September 12, 1989).
With the goal of reaching consensus on the formulation and use of the maintenance effectiveness indicator, the purpose of this meeting was to obtain the initial results of the independent utility reviews of the indicator.
It appears that two central issues emerged from the independent utility reviews:
(1)
Differences still exist on what constitutes the definition of "maintenance."
Most of the utility industry still characterizes nuclear power plant
~
maintenance, and thus what equipment failures are related to maintenance, primarily in terms of corrective work performed by the maintenance personnel at the plant, and thus, only failures that directly result from errors by those individuals in performing corrective maintenance are identified by the utility to be caused by maintenance.
Failures due, for
- example, to hardened or contaminated lubricant tend to be characterized as due to normal wearout or unknown causes.
The NRC staff found the indicator to characterize "maintenance effectiveness" based upon a root AEOD/S804B, January 1989.
1/
cause analysis of the failure narratives for about 4000 NPRDS failure records.
- However, the definition utilized by the staff, which is specified in AEOD/S8048, although not as general as the Commission's definition in the proposed maintenance rule and policy statement, is much more general than the utility definition.
Therefore, the utility analysis of failure narratives characterized the constitution of the indicator to be due primarily to component wearout and unknown reasons, rather than maintenance.
Resolution of this issue will be pursued.
(2)
Failure to completely report to the NPRDS or over-reporting may affect the indicator.
One utility reported that station practices, i.e.,
how aggressively they historically reported component failures, appeared to, be a factor in the frequency of indications.
Another utility was concerned that reporting beyond what is required by the NPRDS might penalize preventive maintenance (PM) programs.*
The potential impact of under-reporting will likely necessitate the'involve-ment of each utility's management to ensure that all of the failures that are within the NPRDS program scope are reported.
It was noted that the implementation of the NPRDS failure reporting guidelines and the INPO review to ensure the correct categorization of the failure (incipient, degraded, or immediate) before inclusion in the database should minimize any over-reporting problems.
The indicator calculations uses only the top two classes of failures reported to the NPRDS.
Incipient failures, those that should be expected to increase in number s as a result of a good preventive maintenance
- program, are not included in the calculation.
Since this issue is related to NPRDS reporting, resolution will be pursued with INPO.
Other minor issues were also identified for resolution.
One was the tendency of the algorithm to place an indicating flag in a month immediately following a month with a significantly high component failure discovery count.
This situation can be resolved with additional software programming.
Another minor issue was that it is desirable for any maintenance indicator to detect repetitive failures across systems.
As a possible resolution of this issue, the indicator calcu-lation could be supplemented using CFAR.
This matter wi 11 be pursued further.
Prior to adjournment, all parties agreed that further work was appropriate to pursue resolution of these issues.
The next meeting was suggested for December 1989.
OriRinal $igned py:
hark H. Qadi>a~
Mark H. Williams, Chief Trends and Patterns Analysis Branch Division of Safety Programs Office for Analysis and Evaluation of Operational Data
Enclosure:
Attendance List
ATTENDANCE LIST OCTOBER 13, I989 MEETING MAINTENANCE INDICATOR DEMONSTRATION PROJECT Tom Tipton Malt Smith Marren Hall Gary Fader Jay Mells Paul Kuhel Bill Angle Ron Moomaw Tom Dente Mike Rodin Mayne Hallman Stuart Lindsey Tom Mar low Tom Laats Howard Stromberg Tom Novak Mark Williams Bob Dennig Don Hickman Tom Wolf Mar c Harper Larry Bell Neill Thomasson Pat O'Reilly ORGANIZATION NUNRC HUYiARC NUMARC IHPO INPO Commonwealth Edison System Energy Resources Incorporated System Energy Resources Incorporated Hortheast Utilities Southern California Edison Duke Power Duke Power Rochester Gas and Electric EGKG Idaho, Inc.
EGKG Idaho, Inc.
HRC/AEOD NRC/AEOD NRC/AFOD NRC/AEOD NRC/AEOD HRC/AEOD NRC/AEOD NRC/AEOD NRC/AEOD