ML20245G976
| ML20245G976 | |
| Person / Time | |
|---|---|
| Issue date: | 06/12/1989 |
| From: | Sniezek J Office of Nuclear Reactor Regulation |
| To: | Jordan E Committee To Review Generic Requirements |
| Shared Package | |
| ML19306D150 | List: |
| References | |
| TASK-051, TASK-51, TASK-OR NUDOCS 8906290339 | |
| Download: ML20245G976 (3) | |
Text
h'. h[* a rrag h UNITED STATES E
o
, ,E NUCLEAR REGULATORY COMMISSION WASHINGTON, D. C. 20555 g ;) n 4 k
\, ...../ ;W ] f11E JUN 1 2 1989 MEMORANDUM FOR: Edward L. Jordan, Chairman Comittee to Review Generic Requirements FROM: James H. Sniezek, Deputy Director Office of. Nuclear Reactor Regulation
SUBJECT:
NRC STAFF POSITION ON RISK OF SERVICE WATER SYSTEM PROBLEMS The Committee to Review Generic Requirements (CRGR) met on May 10, 1989 to-consider among other items our proposed generic letter regarding service water system (SWS) problems affecting safety-related equipment. The staff proposed that this generic communication be issued on the basis that it is needed to ensure that licensees and applicants maintain compliance with General Design Criteria 44, 45, and 46 in 10 CFR Part 50, Appendix A. During the meeting, we also discussed compliance with 10 CFR Part 50, Appendix B as a further basis for issuing the generic letter. Reference to Section XI, " Test Control," will be added to the generic letter, as agreed.
Even though we do not consider the proposed action to be a backfite the staff i has estimated the costs expected in implementing recommended Actions I, II, and N III and the risk avoidance represented by these actions. In estimating the U safety benefit to be realized in implementing the proposed generic letter, we had the benefit of several estimates of the risk and safety significance.
p' ;'
g2 Although these assessments contained considerable uncertainty, HRR has conclud g Q ed that the benefit to be derived from the proposed action is sufficient to u justify the costs involved. As you requested at the May 10, 1989 meeting, NRR is herewith addressing the basis for this conclusion.
NRR worked closely with RES, AEOD, and Region II in developing the proposed generic lette. . Region 11 submitted a draft bulletin in February 1988. RES published " Technical Findings Document for Generic Issue S1: Improving the Reliability of Open-Cycle Service Water Systems," NUREG/CR-5210 in March 1988.
AE0D published " Operating Experience Feedback Report - Service Water System Failures and Degradations," NUREG-1275, Vol. 3 in November 1988. RES published "Value/ Impact Analysis for Generic Issue 51...." NUREG/CR-5234 in February 1989. The RES documents were developed by their centractor, Pacific Northwest Laboratory (PNL).
CONTACT: C. Vernon Hodge r ! R 492-1169 '
/ nC g%f. <,a y W - __
.c 0 Edward L. Jordan JUN 1 2 1989 The PNL study was limited to consideration of the degradation of open-cycle SWS due to fouling from biofouling, sediment, and corrosion products. Action I of the proposed generic letter is based on this study as reported in NUREG/CR-5234. In a memo dated March 31, 1989, PNL stated, and the staff agreed, that results of their study should not be used as the basis for either support or rejection of the proposed generic letter other than for Action I. ,
The approach taken by PNL and its results are summarized as follows:
The base-case core daalage frequency (CDF) (i.e., the frequency of core damage sequences that contain failure (s) of SWS functions prior to implementation of Action I of the proposed generic letter) and the adjusted-case CDF (i.e. after implementation of Action I) were calculated using the Reactor Safety Study Methodology Applications Program (RSSMAP). Only the portion of the total CDF that contains functional failure of the SWS, or its components, is included in the base-case and adjusted-case CDFs. Grand Gulf was taken to represent a boiling watcr reactor (BWR) and Oconee 3 was taken to represent a pressurized water reactor (PWR). The resulting best estimate of the CDF reduction from Action I of the proposed generic letter was 1.6E-6/ plant-yr for the BWR and .7 E-6/ plant-yr for the PWR.
A number of plant-specifi: probability risk assessments (PRAs) have shown a relatively high CDF resulting froEn failures of SWS. An Oconee analysis spomored by the Electric Power Research Institute (EPRI) indicated that a loss of the low pressure SWS is the most important initiating event, yielding an estimated CDF of 1.3E-5/ plant-yr. A similar analysis for Crystal River sponsored by +h1 B&W Owners Group estimated a CDF of 1.2 E-5/ plant-yr for a loss of SWS init1 aced event. An NRC review if a Byron PRA sponsored by the licensee estimated a CDF of 7E-4/ plant-yr for a two train SWS with one train out of service and a failure of the redundant train being the dominant sequence.
ihe AE0D service water study also assessed the CDF for a loss of NS based on operatin; experience. Twelve events were identified in the study that involved SWS failures in 650 plant-years. Based on this initiating frequency, the estimated CDF ranges from 1E-3 to IE-5, allowing for uncertainty in the potential recovery from the event before actual core damage occurs.
In addition to the 12 operating events involving a complete loss of SWS, the AEOD study identified 264 eventt, that involved degradation of SWS. Service water impacts all core damage sequences either directly because of component seal or bearing oil cooling or indirectly because of room cooling that affects component operability, particularly electrical components. Thus, degradation of the service water function increases the failure probability of most of the systems needed to mitigate core damage sequences and so increases the total CDF.
L3 I
~
4 I
'A i
( _
Edward L. Jordan M 121989 None of these studies appear to include consideration that individual trains, or portions of trains, may have undetected, degraded heat transfer and/or flow l capability for extended periods of time. Such situations have been identified by regional inspections.
In summary, estimates of the postulated CDF associated with the functional failure (s) of SWS range widely. Plant-specific PRAs cited above, which include a large variety of vulnerabilities of SWS, indicate that failure of these ;
systems may. result in CDFs ranging from 1E-5/ plant-yr to 7E-4/ plant yr. This
.{
represents a much larger potential for reduction in CDF than shown by the PNL i study, which considered only SWS failures from biofouling, sediment, and I
corrosion products.
The proposed generic letter recommends five licensee actions to improve the major service water and component water system deficiencies and degradations observed from operating experience and plant inspections. Based on the PRAs discussed above, the staff would expect that these recommended actions, if properly implemented by licensees, would significantly reduce the current risk.
qyo mes H. Sniezek, Deputy Director Of ice of Nuclear Reactor Regulation
Enclosures:
As stated 4
_ . - _ _ _ _ _ _ _ _ _.__.__.__.____..-_______.__________m_.________