Resolution of Generic Safety Issues: Issue 151: Reliability of Anticipated Transient Without Scram Recirculatlon Pump Trip in BWRs (Rev. 2) ( NUREG-0933, Main Report with Supplements 1–34 )
This issue was identified in a DSIR/RES memorandum1329 which addressed the concern for the reliability of breakers used to trip the recirculation pumps at high pressure or low water level signals during ATWS mitigation in BWRs. A staff review of BWRs that experienced failures of breakers in the recirculation pump trip (RPT) system was documented in AEOD/E804.1328
If a plant transient requiring a reactor scram occurs and the scram function does not occur, then an ATWS event exists. To lessen the effects of an ATWS event, negative reactivity must be added to the reactor core by tripping the recirculation pumps. Negative reactivity is added as a result of the ensuing steam voiding in the core area as the core flow decreases, thereby decreasing the power generation and limiting the power or pressure disturbance.
Plants equipped with GE AKF-25 circuit breakers have experienced failures of the field breakers in the RPT system which were caused by binding of the trip latch mechanism and misadjustment of the breakers' mechanical linkage. GE issued a service information letter which attributed the circuit breaker failures to misadjustment or lubrication problems and suggested corrective actions and maintenance practices to improve the breakers' performance. In addition, Information Notice No. 87-121322 was issued by the NRC to all BWR licensees to alert them of potential problems with these circuit breakers.
A RPT breaker is included in the design of plants to automatically trip the recirculation pump on high vessel pressure or low reactor water level during an ATWS event. If the RPT breaker fails to trip on demand, the reactor could experience continued power generation resulting in high suppression pool temperature. This issue affected BWRs only.
A possible solution was based on the corrective actions implemented at Pilgrim 1 and involved installing a redundant ATWS trip signal that would interrupt current to the recirculation pumps. Specifically, a new trip coil initiated by an ATWS signal was installed in each recirculation pump motor-generator set drive motor. During an ATWS event, both the recirculation pump field breaker and the motor-generator set supply breaker would receive trip signals, if either high vessel pressure or low reactor water level was reached. Thus, the reliability of the RPT system would be increased and the potential for reaching an unacceptable suppression pool temperature during an ATWS event would be diminished.
Several designs are used in accomplishing the RPT function; however, the GE AKF-25 breaker is used primarily in BWR/3 and BWR/4 designs. There were 6 BWR/3 and 19 BWR/4 plants affected by this issue with average remaining lives of 15.4 and 30 years, respectively. Pilgrim 1 (BWR/3) was excluded as an affected plant because it had already implemented the proposed solution.
The issue affected a plant's ability to render the reactor subcritical followinq an ATWS event. Since the reactor subcriticality analysis64 of Grand Gulf 1 (BWR/6) is analogous to the WASH-140016 analysis of Peach Bottom (BWR/3), the Grand Gulf 1 analysis was used to quantify the influence of the solution on accident frequency and consequence. Thus, the accident sequence affected by this issue was a scram followed by a failure to render the reactor subcritical and was depicted as T23C; the transient-initiating event (T23) had a frequency of 7 events/RY.64
Failure of reactor subcriticality (Event C) has been probabilistically modelled as the product of the following: (1) failure of the RPS; and (2) failure of the RPT or failure of the operator to take the appropriate actions to shut down the reactor, given RPS failure.64 From WASH-1400,16 the failure rate of the RPS was given as 7.7 x 10(6)/demand. The operator error, which was estimated to be 0.1, dominated Item 2 above. The RPT circuit breaker failure rate was given as 3 x 10(3)/demand.
To derive the base case value for event C, the RPT failure rate was modified to 5 x 10(2)/demand, which reflected the lower reliability of the GE AKF-25 circuit breaker.1328 Therefore, the base case value for event C was approximately (7.7 x 10(6))[0.1 + (5 x 10(2))]/demand or 1.16 x 10(6)/demand.
Installation of a redundant ATWS RPT signal was assumed for the adjusted case value of Event C. Therefore, the event (recirculation pump fails to trip) required failure of both RPT subsystems. Assuming that the RPT subsystems are independent and using the GE AKF-25 circuit breaker reliability value, the RPT failure frequency was (5 x 10(2))(5 x 10(2))/demand or 2.5 x 10(3)/demand. This value did not take credit for potential increases in reliability that could result from improved test and maintenance programs, or from changing to a more reliable RPT circuit breaker. The estimate also did not consider the potential decrease in RPT system reliability due to common cause failure mechanisms. Thus, the adjusted case value of Event C was about (7.7 x 10(6))[0.1 + (2.5 x 10(3))]/demand or 7.9 x 10(7)/demand.
Therefore, the T23C accident sequence frequency was 8.1 x 10(6)/RY for the base case and 5.5 x 10(6)/RY for the adjusted case. The total reduction in accident frequency was 2.6 x 10(6)/RY.
Accident sequence T23C falls into the BWR-2 release category (7.1 x 10() manrem/event).64 The total public dose associated with the base case and adjusted case was 57.4 man-rem/RY and 39.2 man-rem/RY, respectively. Thus, the estimated public dose reduction from implementing the possible solution was 18.2 man-rem/RY and the total risk reduction for 25 reactors with an average remaining life of 26.5 years was 12,000 man-rem.
Industry Cost: The cost to implement the possible solution would vary from plant to plant. The following Pilgrim 1 actual costs were used to estimate the industry cost: (1) engineering = $390,000; (2) labor = $66,000; (3) hardware = $10,000. Thus, the implementation cost was estimated to be about $466,000/plant for a total industry cost of about $11M (excluding Pilgrim).
It was expected that the installed redundant RPT subsystem would only be operated for testing purposes and operation costs were thus negligible. In addition, the testing and maintenance procedures for the redundant RPT subsystem would be very similar to existing RPT subsystems and, therefore, should require very little additional development time. Thus, testing and maintenance were each estimated to require 0.5 man-week/RY, resulting in a plant cost of $2,270/RY and a total industry cost of about $1.5M. Thus, the total industry implementation, operation, and maintenance cost was $12.5M.
NRC Cost: Development of the solution was estimated to require one man-year of contractor labor, at a cost of $100,000/man-year, to complete an evaluation of the solution and any potential alternatives (e.g., enhanced test/maintenance or replacing the degraded RPT breakers with more reliable models). This study would also need to include a preliminary review of plant designs to determine the technical feasibility of the proposed modifications. Development of the solution would also include issuing an NRC generic letter to the affected plants, which was estimated to cost about $11,000.961
Review of the proposed plant modifications was estimated to take 5 man-weeks/plant for a total NRC review cost of $280,000 for the 25 affected plants. Thus, the total NRC cost for development and review was $390,000.
Total Cost: The total industry and NRC cost associated with the possible solution was approximately $13M.
Based on an estimated public risk reduction of 12,000 man-rem and a cost of $13M associated with the possible solution, the value/impact score was given by:
The scram frequency estimate used in the above calculations was considerably greater than that reflected in performance indicator reports published prior to the above evaluation which was completed in August 1991. In addition, the RPS failure rate was originally developed for WASH-140016 prior to the ATWS rulemaking and was also quite outdated. As a result, the RPS reliability goal from the ATWS rulemaking proceedings was utilized as a conservative value and the risk reduction calculations were repeated using: (1) a scram frequency of 3.1/RY, derived from data in the 1988 AEOD Annual Report and Part 1 of the Third Quarter 1990 AEOD report, "Performance Indicators for Operating Commercial Nuclear Power Reactors"; and (2) an RPS failure rate of 3 x 101453/demand from the ATWS rulemaking proceeding.704 These estimates resulted in a potential risk reduction of 20,800 man-rem and a value/impact score of 1,600 man-rem/$M.
Based on the potential public risk reduction associated with the possible solution and the other considerations above, the issue was given a medium priority ranking (See Appendix C) and resolution was pursued.
Following identification of this issue, NRC Information Notice No. 92-061453 was issued to alert licensees to the importance of maintaining the reliability of equipment required by regulations but not addressed in plant TS. In resolving the issue, the staff conducted a more detailed calculation of the risk reduction potential and concluded that the issue posed a low risk. Thus, the issue was RESOLVED and no new requirements were established.1439 In an RES evaluation,1564 it was concluded that consideration of a 20-year license renewal period did not affect the resolution.