ML20046C853

Discusses Staggered Surveillance Testing to Address Uncertainty in Plant Risk Caused by Maint Related Common Cause Failures

ML20046C853
Person / Time
Issue date:	06/24/1993
From:	Israel S NRC OFFICE FOR ANALYSIS & EVALUATION OF OPERATIONAL DATA (AEOD)
To:	Rosenthal J NRC OFFICE FOR ANALYSIS & EVALUATION OF OPERATIONAL DATA (AEOD)
References
TASK-AE, TASK-E92-02, TASK-E92-2 AEOD-E92-02, AEOD-E92-2, NUDOCS 9308120309
Download: ML20046C853 (3)

Text

~

O O

'o s

UNITED STATES g

NUCLEAR REGULATORY COMMISSION n

E WASHINGTON, D. C. 20555 JUN 2 41993 MEMC" ANDUM FOR

J. E. Rosenthal, Chief Reactor Operations Analysis Branch Division of Safety Programs Office for Analysis and Evaluation t

of Operational Data FROM:

Sanford L Israel Reactor Systems SectionlV_ and B&W Reactor Operations Analysis Branch Division of Safety Programs Office for Analysis and Evaluation of Operational Data

SUBJECT:

STAGGERED SURVEILLANCE TESTING TO ADDRESS UNCERTAINTY IN PLANT RISK CAUSED BY MAINTENANCE RELATED COMMON-CAUSE FAILURES At the conclusion of my common-cause failure study (AEOD/E92-02), T. Novak directed that additional consideration be given to staggered surveillance testing as a remedy for maintenance related common-cause failures. Subsequently, I visited four sites (7 units) and explored their concerns on staggered surveillance testing.

Staggered train testing is used at all plants to some degree because of existing requirements in the technical specifications, or national equipment code requirements embraced by the regulations. A review of testing practices at four plants indicates that it may be easily implemented for certain situations, but may produce hardships in other i

cases. There appears to be a number of variables (discussed below) that impinge on the i

flexibility and degree of implementation of staggered testing that might be easily achieved.

Maintenance scheduling practice is a big factor. One site plans maintenance only about 1 month in advance and takes full advantage of the schedule slippage allowed by the J

i U technical specifications. It is difficult to maintain a staggered test schedule when chronic e

test slippage is commonplace. Other sites schedule on an 84 day quarter (4-week I

and 12-week quarters), with no slippage routinely allowed, so the testing schedule is the i

same from one quarter to the next. Other sites have 5 year maintenance schedules that 1

include major preventive maintenance activities. One plant has alternating 2-week (!V/ M !

f periods for working on train A, then on train B. This later situation automatically introduces at least a 2-week stagger in the testing program.

lg//

g, j (###

9308120309 930624 g

PDR ORG NEXD n u l u is 20 EI CEi!1TER COPY

/sM '

'1,Jsyc @d

~

Jack E. Rosenthal

{') g w

Distribution of maintenance resources at a site can also impact flexibility. One 3-unit site does not use dedicated personnel for each unit, while another 3-unit site does. Thus, it may be more complicated to juggle resources among three units than for a single unit.

The test frequency also has an impact. Tests or checks with intervals of a week or less appear to be handled by standing orders in the control room. These tests are normally performed sequentially. Tests performed at refueling outages may also have constraints because of the inability to perform the test while the plant is at power. The reasons for these constraints include dose burdens, pctential for a reactor trip, and plant configuration. Fluid system fouling tests have intervals on the order of a refueling cycle.

These tests may also be impacted by the need for adequate temperature differentials that occur only in the summer. At some plants, heat exchanger fouling is determined by opening up the equipment and performing preventive maintenance, not by testing. Many of the maintenance related common-cause failures cited in AEOD/E92-02 were concerned with fluid system fouling, which may be less amenable to staggered testing or maintenance.

One licensee uses system-level test procedures for instrumentation tests. Separate train test procedures would have to be developed to implement staggered testing for each train at this site. One licensee likes staggered testing because it allows different personnel to become experienced with the test activities. Another licensee has the opposite view that sequential testing was better because he could then develop experts for these activities (although this goal is not precluded by staggering the tests).

Generally, all the licensees contacted indicated that staggering tests having intervals 1-month to 6-months, would not be intolerable, if not already being done.

I do not advocate any new generic initiatives at this time because staggered testing appears to impact only a portion of the potential common-cause failure situations noted in AEOD/E92-02. An information notice has been issued on the findings of this study and about 10 to 20 requests have been received for the AEOD report.

It remains to be seen what action the licensees may initiate on their own with regard to uncertainty in the perceived risk at their plants. A voluntary, best-effort, approach would eliminate the NRC staff having to review exceptions that would inevitably be proposed by each licensee. However, the results of this study could also be considered on a case by case basis as part of the NRC staff interactions with the licensees on their IPE results or as part of corrective action programs for poorly performing licensees.

OdsNI Signed by-Sanford L Israel Reactor Systems SectionlV and B&W Reactor Operations Analysis Branch Division of Safety Programs Office for Analysis and Evaluation of Operational Data nf' Distribution: See attached list.

n ROAB Cd ROAIQt[f C

AB WJ nds JNosenthal Sisrael:mmk 6/7//93 6/y/93 6/#p'93

ck,y g

/

w Jack E. Rosenthal Distribution:

PDR EJordan KBrockman Central File -

Dross PBaranowsky ROAB R/F GHolahan KRaglin, TTC SIsrael VBenaroya AChaffee, NRR WJones LSpessard RSavio, ACRS JRosenthal SRubin MTaylor, EDO