ML20045B221
| ML20045B221 | |
| Person / Time | |
|---|---|
| Issue date: | 02/19/1991 |
| From: | Curtiss NRC COMMISSION (OCM) |
| To: | Carr, Remick, Rogers NRC COMMISSION (OCM) |
| Shared Package | |
| ML20042D089 | List:
|
| References | |
| FRN-57FR14514, REF-GTECI-B-56, REF-GTECI-EL, RULE-PR-50, TASK-B-56, TASK-OR AE06-1-059, AE6-1-59, NUDOCS 9306160431 | |
| Download: ML20045B221 (4) | |
Text
{{#Wiki_filter:-. _ - _ - _ _ _ _ _ - ' WO $ - -{- / a , FEB-20 *91 16:11 1 D: U5mC-EFO-17 TEL NO:492-0162 1:109 PO2 e ~ " 'h - UNITE Ai o !T-NUCLEAR REGULATORY COMMISSION n E WASHINGTON. D C. 20555 Q e c V February 19, 1991 We" pg OFFICE OF 1HC COPJ C-91-001 "~~iJ COMMt$100f4 C R Y CD ' A MEMORANDUM FOR: Chairman Carr Commissioner Rogers Commissioner Remick FROM: Commissioner Curtiss
SUBJECT:
DIESEL GENERATOR RELIABILITY In my vote on SECY-90-340, " Resolution of Gcncric Safety Isaue B-56, Diesel Generator Reliability", I recommended for your consideration an approach to the resolution of GSI B-56 that would rely upon a results-oriented, performance-based approach to this issue, rather than the programmatic approach originally recommended by the staff in SECY-90-340. Since that time, I have had an opportunity to discuss the performance-based approach in some detail with the imC staff. Those discussions, which your respective staffs have attended, have resultml in a~ number of useful insights and suggestions, from both a technical and a legal percpoetive. of particular interest, I think it is fair to say that all parties now have a better appreciation of just exactly how a performance-based approach to the resolution of GSI B-56, with a focus on the quantitative result that the agency necks to achieve under the station blackout (SBO) rule, would work. Additionally, I believe that all partics agree that if a performance-based approach to the resolution of GSI_ B-56 is to be employed, there are two key principles that will be central to the success of such an approach. Those two principles are as follows: (1) the approach must be sufficiently rigorous to give the imC, as well as the licensee, sufficient "early warning"'of a potential deterioration in diesel generator performanect and (2) prior to taking enforcement action, there must be a sound statistical basis for concluding that diesel generator performance has deteriorated below the reliability levels assumed in the station blackout (SBO) cubmittalc. On the basis of the foregoing discuccions with the staff, I continue to believe that, with some refinement, a performanco-based approach along the lines that I originally recommended in my vote on SECY-90-340 is the preferred approach to the resolution of GSI B-56. Accordingly, I would now propoco for your consideration the following approach: 9306160431 930422 9 PDR PR 4 50 57FR14514 PDR 0
u4 g - f, FED-20 '91 16:11. I D:lG 4RC-EDO-17 TEL tJO:492-016,2 M109 PO3 -2 The staff chould proceed to implement a performance-baced approach to the resolution of GSI B-56, in lieu of~the approach originally recommended in SECY-90-340.I The performance-based approach would consist of the-following principal olements: 1) Trigger values would be cctablished for the purpoco of determining whether the reliability of a licensee's diesel generators has fallen below the levolc accumed in the SBo submittals.2 Specifically, I would propose that we adopt a " double excendance" trigger value of 5 failures out of 50 starts and 8 failuroc out of 100 starts for licensees required to maintain a reliability 1cvel of 0.95; and 4 failurce out of 50 starts and 5 failuroc out of 100 starts for licenseen required to maintain a reliability lovel of 0.975. These values, which would be applied on a unit basis (i.e., to the aggregate of all diesels at a given unit rather than to each individual dicscl), would minimi=e the probability of " false alarms" and, if exceeded, would provide a firm statistical basic for concluding that the reliability levels assumed in the SBO submittals have not been met. Additionally, the values would be consistent with the trigger values established in NUMARC Initiative SA, the document endorsed by the NRC ctaff. Failure to meet the specified trigger values would serve as the basis 'or taking enforcementaction.jncludingth. .sssiblity of regulatory action,
- For thoco licenaces who have already committed to a diesel generator rollability program in the context of the submittals required pursuant to the station blackout rule, staff should promptly advice those licensees that the Commission has considered the need for such commitments and, in view of the approach that the commission has endorsed for the resolution of GSI B-56, those commitments no longer need be considered as legally binding.
2 For moct licenseen, the cubmittala required under the station blackout rule assume that emergency diesel generator reliability will be maintained at a level of 0.95, measured on a unit-wide basis. For a small number of licensees, the submittals accune a reliability level of 0.975. 3 The exact nature of the regulatory response, including the specific level of enforcement action to be taken, is a matter that I would leave to the discretion of the staff, with the understanding that the agency's response in such a situation would be commensurate with the safety significance of the
11109 POG FED-20 '91 16:13 I D: U9 EC-EDD-17 TEL NO:492-0162 _3-In addition to establishing a trigger value for 2) of determining whether the reliability the purpoca of a licenccc'c diccc1 generators han fallen below the level accumed in a licensco's SDO submittal, I early warning" n would propose that a ceparato trigger value be established for the purpose of detecting potential diesel generator degradation at an early stage. Thic trigger value would consist of a requirement that licensccc -- (i) rcport diesel failurcs to the NRC when such failures reach 3 failures out of 20 starts, meacured on either a per unit or a por diccc1 basist and (ii) undertake accelerated testing, as well as cubmit a report to NRC, in the event that diccci failures reach 4 out of 25 failurec, neasured on an individual dicsol basis.' Unlike the trigger values established in (1), above which, if not met, provide clear statistical evidence that the underlying reliability levels have not been achieved -- failure to meet the "carly warning" trigger values would not serve as a basis for taking regulatory action.5 These trigger valuce would nerve strictly as an "carly warning" mechanism to flag a potential decline in diesel generator reliability. In implementing t c foregoing approach, the staff should e seek to take maximum advantage, to the extent legally violation and concistent with the principles established in the General Statement of Policy on Enforcement, 10 CFR Part 2, App. C. The central point here is that the regulatory response, whatever it night be, would be based upon a failure to meet the rather reliability level assumed in a licensee's SBO submittal, than for failure to comply with an NRC-imposed programmatic requirement specifying the details of how the reliability objective is to be mot. ' In the event that an individual diesel generator fails on four or more starts out of 25 tests, I would propose that we adopt the approach to accolcrated testing that is taken in NUMARC Initiative SA. Under this approach, the licensee must demonstrate restored diccel generator performance by conducting with the seven consecutive failure-free start and load-run tests, testing to be conducted at a frequency of no less than 24 hours and no more than seven days between cach demand. I would note that licenacco are currently required under 5 10 CFR Part 50, Appendix B, to undertake root cause analyses and corrective actions for significant conditions adverse to quality. In vicw of the safety significance of emergency diesel generators, this provision, in my judgment, would clearly apply whencvor a diesel generator fails to start.
- -- FEB-20 ' 91 16:13 ID:UStRC-ETO-17 TEL ttO:492-0162 M109 POS s permicsible, of the provisions of the existing SBO rule, as well as the, existing reporting requirements contained in 10 CFR Part 50.72 and 50.73.
If modifications to th.c oxicting rules are required to provide the necescary legal underpinning for the approach outlined above, staff should proceed promptly to develop proposed changes to the oxicting rules for the Commission'a consideration. 2e licenscos that Finally, staff should proceed to ad' todifications to individual technical specifications to address the issue of excessive accelerated testing for " problem diesels", as well as the issue of excessive reporting for EDG failures, will be entertained frca this point forward, without licensees first having to commit to the approach set forth in Regulatory Guido 1.9. In my view, the foregoing approach would reprocent a significant advancement in the application of performance-based principles to our regulatory process. Indeed, if this approach is adopted, it will represent perhaps the most important application of the principles of performance-based regulation to date by this agency. Beyond the inherent benefits of such an approach -- including the flexibility that licensees would have to meet the objectives of our SBo rule in whatever manner they might individually elect -- I firmly believe that the use of such an approach here will serve as a valuabic opportunity to demonstrate our commitment to minimizing prescriptive regulation, where appropriate, while at the same time achieving the health and safety objectives that we, as an agency, are scoking. SECY, please track. cc: SECY EDO OGC l . _ _ _ - - - - _ _ _ _ _ _ - _ _ _ - _ _ _ _ - _ _ _}}