ML20147E422

Forwards Summary Rept, Evaluation of Candidate LSSC Check Valves for Risk Based IST Extension at Palo Verde Units 1,2 & 3

ML20147E422
Person / Time
Site:	Palo Verde
Issue date:	11/11/1996
From:	Poole A OAK RIDGE NATIONAL LABORATORY
To:	Jackson J NRC
Shared Package
ML20147E427	List: ML20147E422 ML20147E430 ML20149H819
References
FOIA-97-177 NUDOCS 9703030064
Download: ML20147E422 (2)

Text

--

l CAK RIDGE NATIONAL LABORATORY

$4ANAGED SV LOCKHEED MARDN ENENGY RESEARCH CORPORATION PHONE: (423) 574-0734 FoR THE U S DEPARTMENT of ENERGY FAX: (423) 676-0493 l POST onCE DOX 2009 OAK RIDGE TN 3703M030 November 11,1996 Jerry E. Jackson l

U. S. Nuclear Regulatory Commission '

MS T10 E10 Washington, D.C. 20555

Dear Mr. Jackson:

i As we discussed in the October 8,1996 meetMg with NRC personnel, selected Palo Verde check valves listed in Appendix D to 13-NS-COS, Rev. O, have been reviewed relative to failure history.

The time period covered was from 1986 through 1995.

Palo Verde has requested IST extension from their current Code requirements (usually quarterly) to an interval of 6 years on approximately 228 " low risk significant" check valves (76 valve applications were listed in the submittal; it is assumed that all three units affected). In an effort to provide information needed to evaluate potential candidate risk based inservice test (RBIST) check valves at Palo Verde for extended IST intervals, Oak Ridge National Laboratory (ORNL) has done a brief review of the available NPRDS failure records and performance data for the valves in question. The results of this review are provided in the attached summary report.

The most significant findings of this study resulted from a brief review of the 106 raw NPRDS failure records listed for check valves at Palo Verde Units 1,2,3 during the time period 1986-1995.

Of the 106 raw NPRDS failure records,55 were associated with the candidate valves. Seventeen percent of the valve applications listed in the relief request submittal have experienced repeat failures. Some valves had as many as seven repeat failures (all three units considered). It is important to note that at least 16 of the failed valves listed in Table 10 of the attachment had failed cr degraded intemals caused at least in part by some age-related failure mechanism such as " wear" cr " cyclic fatigue." These types of failure causes need to be considered when evaluating whether to extend inservice testing intervals. Of the 55 failure records associated with the deferral candidate valves,11 involved extemal leakage, while a characterization of the remainder according to extent cfdegradation resulted in 29 (66%) moderate and 15 (34%) significant failures. These results are comparable to those found industry-wide during previous ORNL studies for check valve failures occurring during 1991 and 1992.

The system of service for candidate IST deferral should also be considered. At least 75 percent of the candidate " low safety significance" valves for deferral come from either AFW, Diesel Starting Air, Containment Isolation, CCW, Main Steam, or RHR systems, which in previous ORNL studies i have been shown to have some of the highest relative failure rates by system for significant failures

(in terms of component degradation).

~ !Eringing eScienceio Bfe

. h0hhh'TW)]

Y _

. l 6

-. Jerry E. Jackson November 11,1996 Page 2  ;

j The number of repeat failures and type of failures listed in NPRDS (see page 7, 8, and 9 of the

, attachment) certainly seems to indicate that age-related failure mechanisms are present in the

' following systems:

• CVCS Diesel Starting Air Main Steam

• RHR The recently provided Draft NUREG/CR-6508 " Component Unavailability Versus Inservice Test (IST)lnterval: Evaluations of Component Aging Effects With Applications to Check Valves," has shown that unmitigated component aging can significantly increase the unavailability and risk due to decreased testing. The Palo Verde submittal has not addressed aging-related effects on the risk analysis completed. I Although this examination was rather cursory in nature, ample evidence exists to question the technical validity of extending the inspection interval for the requested check valves.

Unavailabilities of all check valves in applications susceptible to aging should be simultaneously increased by the appropriate factor to cover the simultaneous effects of aging. This s*hould be completed to show that the impact on risk remains low even for unmitigated aging.

We hope that this information will be usefal to you. Should you need additional information we would be glad to provide further assistance. More failure data information on check valves will be ,

provided later by special Letter Report.

Sincerely, a8 Por&

A. B. Poole ABP:jke.

Attachments cc: enc: P. L. Campbell, NRC F. Grubelich, NRC J. Colaccino, NRC W. E. Vesely, SAIC K. L. McElhaney J. P. Vora, NRC D. C. Fischer, NRC R. H. Wessman, NRC W. C. Gleaves, NRC

&