Resolution of Generic Safety Issues: Issue 78: Monitoring of Fatigue Transient Limits for Reactor Coolant System (Rev. 3) ( NUREG-0933, Main Report with Supplements 1–34 )
STS for newer OLs require licensees to keep account of the number of transient occurrences to ensure that transient limits, based on design assumptions, are not exceeded. However, a number of older plants for which detailed fatigue analyses were performed on ASME Section III, Code Class 1 (RCPB) components did not have TS requirements for monitoring actual transient occurrences. These transients could significantly affect the fatigue life of the RCS. This issue was originated as MPA B-70, but a letter to licensees was never issued for the collection of information.528,529
Repeated thermal cycling of RCS components produces some degree of fatigue degradation of the component materials which could lead to failure, thereby increasing the likelihood of a LOCA.
A possible solution was to require the affected plants to implement TS to monitor plant transients and to verify that the design life of all ASME Section III, Code Class 1 components had not been exceeded. Plants that had experienced transient events that exceeded design limits may find it necessary to perform fatigue analyses to determine the number of remaining thermal cycles before fatigue limits are exceeded.
The reference plants selected for this analysis were Oconee-3 for PWRs and Grand Gulf-1 for BWRs. It was estimated64 that 30 PWRs and 16 BWRs with average remaining operating lives of 27.7 years and 25.2 years, respectively, were affected by this issue. These estimates were based on the original 40-year operating license.
The affected parameters in the Grand Gulf 1 and Oconee 3 PRAs were determined to be the LOCA-initiating event frequencies.64 The original values of these parameters are shown below:
|Oconee 3||Grand Gulf 1|
|S1 : Large LOCA = 1.0 x 10-4/RY||S : LOCA = 1.4 x 10-3/RY|
|S2 : Medium LOCA = 4.0 x 10-4/RY|
|S3 : Small LOCA = 1.3 x 10-3/RY|
The values for these parameters were developed assuming that the fatigue life for RCS components was as designed. Potential events such as overcooling or transients that may result in RCS cooldown rates in excess of limits that could reduce the design life of RCS components were not taken into account. Therefore, these values also represented the adjusted case values of affected parameters as they were derived assuming: (1) the design limits would not be exceeded; and (2) the number of allowable thermal cycles was the same as that used in the design. These values also represented plant conditions representative of the improvement they would achieve through resolution of the issue.
The base case values of the affected parameters were developed in consideration of the actual fatigue degradation that may be experienced by RCS components. The degradation had been shown to reduce the lifetime of RCS components which, in turn, will increase the frequency of RCS piping failures. This increase in frequency was assumed to be inversely proportional to the reduction in the number of allowable cooldown cycles with respect to the as-designed number of cycles that a plant may experience.
A fatigue analysis1366 was performed in support of the restart of Rancho Seco-1 after several transients were experienced during which the normal cooldown rate was exceeded. This analysis examined the limiting fatigue usage factors in the reactor coolant piping, pressurizer, CRD mechanisms, and the RCP casings. It was concluded that the allowable number of cooldowns should be reduced from 240 to 235. Taking the remaining plant life into consideration, it was estimated that Rancho Seco-1 could experience an additional 2 transients during which the normal rate would be exceeded up through the midpoint of remaining lifetime. This would further reduce the number of cooldowns from 235 to 233 and represented an overall reduction of about 3% of the number of allowable cooldowns. The inverse ratio, which represents the reduction in the frequency of RCS pipe breaks, is a factor 1.03. This factor was multiplied by the adjusted case values given above to estimate the base case values of the affected parameters. These values are shown below:
|Oconee 3||Grand Gulf 1|
|S1 : Large LOCA = 1.05 x 10-4/RY||S : LOCA = 1.47 x 10-3/RY|
|S2 : Medium LOCA = 4.20 x 10-4/RY|
|S3 : Small LOCA = 1.37 x 10-3/RY|
The probability of reactor vessel failure as a result of thermal cycle fatigue was not included in the base case or adjusted case values because it was believed that the frequency of reactor vessel failure was so small relative to the frequencies of LOCA sequences that it would be insignificant. This conclusion was supported by: (1) data in WASH-140016; (2) the fact that it was not examined in the Rancho Seco-1 restart; and (3) the resolution of Issue 79. Using the affected accident sequences, i.e., all minimum cut sets initiated by LOCAs, and the base case and adjusted case LOCA frequencies developed above, the potential reduction in core-melt frequency was estimated to be 6.3 x 10-7/RY for PWRs and 10-7/RY for BWRs.
The potential risk reduction from the possible solution was estimated to be 1.26 man-rem/RY for PWRs and 0.72 man-rem/RY for BWRs. When applied to the affected population of 30 PWRs and 16 BWRs over their estimated remaining lives, a total potential public risk reduction of 1,300 man-rem was estimated.
Industry Cost: Implementation of the possible solution would require preparation of an initial report, TS development, and performance of fatigue analyses, as appropriate. For the 46 affected plants, the total implementation cost was estimated64 to be $4.6M. Operation and maintenance of the possible solution would involve weekly monitoring of fatigue transient limits, estimated64 to cost $910,000.
NRC Cost: Development of a solution was estimated to cost $91,000, including development of a model TS and preparation of a letter to licensees. Support for implementation of the solution was estimated to cost $630,000, including review and approval of license amendments and TS and the preparation of safety evaluations. Support for operation and maintenance associated with the possible solution was estimated to cost $910,000, resulting in an overall total NRC cost of $1.6M.
Total Cost: The total industry and NRC cost associated with the possible solution was estimated to be $7.1M.
Based on an estimated public risk reduction of 1300 man-rem and a cost of $7.1M associated with the possible solution, the value/impact score was given by:
(1) The assumed solution, consistent with the narrowly specified scope for this issue, involved only the monitoring and reporting requirements and, perhaps, evaluations of thermal cycle fatigue limits against the original licensing basis ASME Code criteria. No work in radiation zones would be required for implementation, operation, or maintenance of the possible solution.
(2) Assuming that the affected plants will have their licenses extended for an additional 20 years and that these plants would be operational 75% of this time, the average remaining operating life would be 42.7 years for PWRs and 40.2 years for BWRs. The possible risk reduction, including overcooling events for this extended period, would be 2,790 man-rem. The total industry cost for implementation, operation, and maintenance of the possible solution would be $6M; the NRC cost over this same period would be $2.1M. Thus, the total industry and NRC cost would be $8.1M. The value/impact score for life extension would be 344 man-rem/$M.
(3) In June 1992, NRR requested1452 that the issue be expanded to include those early vintage plants whose RCS components were not designed to ASME Section III and did not receive a fatigue analysis. Expansion of the scope of this issue to include recognition that many early vintage RCS components were not specifically evaluated for fatigue effects would result in an increased, though unquantified, probability of failure and also would increase the uncertainty of the failure probability.
(4) The staff expressed concerns regarding the adequacy of the fatigue curves used in ASME Section III in taking into account environmental effects. At the time of the initial evaluation of this issue in July 1992, data indicated that existing Code fatigue curves may have had less margin than that originally intended when considering the effects of fatigue induced by the operating environment. The net result of the reduced conservatism in the fatigue design curves in ASME Section III could be an increased probability of component (including reactor vessel) failure by an unquantified amount.
Taking the above considerations into account would tend to increase the priority ranking of the issue.
Based on the potential public risk reduction and the uncertainties discussed above, the issue was given a medium priority ranking (See Appendix C). The issue was expected to provide technical information for the resolution of Issue 166 and, therefore, did not require separate consideration of a 20-year license renewal period.1564
The portion of this issue that addressed transient monitoring was integrated into the Fatigue Action Plan (FAP) which was reported complete in SECY-95-245.547 Thus, the staff's efforts were redirected towards evaluating the risk associated with Issue 166. As a substantive element of the FAP, the data gathered in the resolution of Issue 78 provided essential technical information for the resolution of Issue 166.
In resolving Issue 166, the staff obtained available records of transient monitoring from the licensees of the 7 plants selected for the FAP sample. On the basis of the FAP sample and considering the available transient monitoring records and the conservatism identified in the component analyses, it was concluded that there was no significant safety concern since the CLB had not been exceeded at operating plants.629 Thus, Issue 78 was RESOLVED with no new requirements.