Resolution of Generic Safety Issues: Issue 95: Loss of Effective Volume for Containment Recirculation Spray ( NUREG-0933, Main Report with Supplements 1–34 )
This concern was raised by an NRC resident inspector who questioned the practice of leaving the refueling canal drain valve in the closed position during operations at H. B. Robinson Unit 2.1219 A subsequent investigation by the licensee revealed that W had intended the refueling canal drain valve to be open during operation, but that operation with the valve closed would not adversely impact safety due to the large volume of water available from the Refueling Water Storage Tank (RWST) relative to the maximum volume of water which could be entrapped in the refueling canal. Nevertheless, the licensee decided to operate the plant with the valve open and to revise plant procedures accordingly.
Because of the potential generic implications of this matter, Task Interface Agreement (TIA) 83-144 was established1220 to review the containment sump configuration for certain plants to determine whether unacceptable entrapment of water could occur during recirculation. The TIA specifically made two requests:
(1) A review of H. B. Robinson Unit 2 to determine if the plant configuration and licensee's corrective action (operation with the drain valve open) were adequate to prevent unacceptable water entrapment. The licensee's response was taken under staff review and subsequently found acceptable.1220
(2) A reexamination of those PWRs that were not reviewed under SRP11 6.2.2, "Containment Heat Removal Systems," to determine if water entrapment could occur such that recirculation heat removal may be inadequate. Twenty-seven PWRs were identified that needed to be surveyed to determine if entrapment of containment recirculation water could occur.1219
The principal concern of this issue was the potential for water entrapment in the refueling canal of certain operating PWRs. If the refueling canal drains in a PWR dry containment are closed during plant operation and a LOCA occurs, that fraction of the containment spray which falls into the refueling canal would be prevented from returning to the containment emergency sump. Eventually, the entire volume of the refueling canal would be filled with water which would then not be available for the recirculation mode of containment and reactor cooling.
The system failures of concern in the recirculation mode are the containment spray recirculation and emergency coolant recirculation systems. Following a postulated LOCA and after the injection phase of containment spray is complete, it is likely that containment spray flow could be discontinued while maintaining containment pressure reduction with the containment fan cooler units and returning all of the recirculated water to the core. Since the containment fan coolers may be sufficient to maintain containment pressure following spray injection, the safety concern is the potential loss of emergency coolant recirculation. If the volume of water held up in the refueling canal is greater than the margin of excess water allotted in the design of the containment sump and recirculation systems, containment cooling and/or reactor cooling could be lost for some LOCA events. If there is not enough water for effective emergency coolant recirculation, then core-melt may ensue.
It was believed that resolution of this issue could involve the following actions:
(1) Review containment sump and recirculation system (including refueling canal drain valve position while the plant is at power). Evaluate the recirculation system to determine if water entrapment could occur such that recirculation heat removal may be inadequate.
(2) If necessary, change procedures such that the refueling canal drain valve is in the open position while the reactor is operating to assure that the most water which can be available is available for recirculation.
To establish the priority of this issue, the potential reduction in core-melt frequency was quantified as the result of the affected plants revising their current operating procedures due to implementation of the proposed resolution. It was believed that, by changing the operating procedures of the affected plants, the risk will be reduced due to the expected improvement in the reliability of the emergency coolant recirculation and containment spray recirculation systems.
The Oconee 3 PRA studies were used as a representative PWR to estimate the increase in the reliability of containment spray recirculation and emergency coolant recirculation systems. In performing this analysis, it was assumed that resolution of this issue will serve to identify a need for procedural changes at approximately one-half of those PWRs that were not reviewed under SRP11 6.2.2. (SRP 6.2.2 requires that the potential for and effects of recirculation water holdup be specifically addressed in plant FSARs and/or TS). The remaining one-half are assumed to already operate with the refueling canal drain valve in the open position. A reduction in public risk at those plants which are currently operating with the refueling canal drain valve in the closed position will result from an increase in the reliability of the containment spray recirculation and emergency coolant recirculation systems.
In performing the analysis, it was assumed that the issue affected 26 operating PWRs which were licensed prior to 1975 and hence not reviewed under SRP11 6.2.2. However, it was assumed that one-half of these plants operated with the refueling canal drain valve in the open position and were not required to modify their procedures. Therefore, 13 plants will realize a reduction in risk by implementing the issue resolution. At the time of this evaluation, the average remaining life of the affected operated PWRs was 24 years.
It was estimated that the value for the unavailability of sufficient water for recirculation due to water entrapment was no worse than the unavailability of the low pressure injection system and the high pressure recirculation system. The parameters affected are H (Emergency Coolant Recirculation System Failure) and F (Containment Spray Recirculation System Failure); F is a dependent failure.
For those plants operating with the refueling canal drain valve in the closed position, the base case values are 1.07 x 10-2/demand for H and 0.4/demand for F, given H. For those plants that revise their operating procedures, the adjusted case becomes 7.7 x 10-3/demand for H and the dependent failure F remains at 0.4/demand, given event H. The reduction in failure probability of H is thus 3 x 10-3/demand. This value was estimated on the basis of human error (procedural) and was added to the Oconee RSSMAP value of H to determine the base case probability of H (0.0107/demand), because the determination of the value of H in the PRA does not consider inadequate water supply as a cause of ECCS failure.
The estimated base case core-melt frequency for those plants operating with the refueling canal drain valve in the closed position was 4.3 x 10-5/RY and the adjusted case core-melt frequency for those plants implementing procedural changes was 3.1 x 10-5/RY. The estimated resultant accident frequency reduction was 1.2 x 10-5/RY.
The base case public risk was 113.4 man-rem/RY and the adjusted case affected public risk was 80.1 man-rem/RY. The resultant public risk reduction was 33.3 man-rem/RY. Therefore, the total potential risk reduction associated with this issue was 1.2 x 104 man-rem for all affected reactors.
Industry Cost: Implementation of the proposed resolution was assumed to affect the 26 PWRs that were licensed prior to 1975 and hence not reviewed under SRP11 6.2.2. Of these 26 plants, it was assumed that 13 required change and the remaining 13 did not. It was assumed that all 26 plants will have to evaluate their existing containment recirculation systems and review normal valve position and operating procedures. Two man-weeks/plant were estimated for the evaluation and review. In addition, all 26 plants must respond to the NRC request for system analysis. Two man-weeks/plant were estimated for plant response to the NRC. For those plants that have to implement the proposed resolution, 3 man-weeks/plant were estimated for TS revisions and 3 man-weeks/plant for administrative changes and modifications. Therefore, it was estimated to take 10 man-weeks/plant for those plants requiring change and 4 man-weeks/plant for those plants not requiring change. At $2270/man-week, the plant implementation cost was $22,700/plant for plants requiring change and $9,080/plant for plants not requiring change. The implementation cost for the 13 plants requiring change was $295,100 and $118,040 for those plants not requiring change, for a total implementation cost of $413,140. No additional plant maintenance and operating costs were expected from implementing the proposed resolution.
NRC Cost: It was estimated that resolution of this issue would involve a review by the responsible technical branches to determine which of the 26 affected plants had a vulnerability to water entrapment. It was assumed that a review of 2 man-weeks/plant would be needed to verify each licensee's analysis of the matter of water entrapment. In addition, the cost analysis anticipated that the NRC would issue a bulletin requesting that licensees review their containment sumps and recirculation systems, as outlined in the proposed resolution.
The NRC cost for safety issue resolution (SIR) development was estimated to be $26,000 for the bulletin and $120,000 for review of licensee evaluations, for a total of $140,000. Resolution of this issue assumed procedural changes at 13 of the 26 plants and one man-week/plant for support of the SIR implementation. At $2270/man-week, the NRC cost for support of SIR implementation was estimated to be $30,000. The total NRC cost for development and support of SIR implementation was $170,000.
Total Cost: The total estimated industry and NRC cost associated with the possible solution to this issue was $0.58M.
Based on a potential risk reduction of 1.2 x 104 man-rem and a cost of $0.58M, the value/impact score is given by:
(1) The implementation of the proposed resolution was assumed not to involve any labor in radiation zones because implementation would consist of modifications to plant procedures.
(2) The occupational dose increase for SIR operation and maintenance was anticipated to be 0.01 man-rem/RY. Plant operation with the refueling canal drain valve in the open position has no effect on occupational exposure. The SIR would increase occupational dose by the amount of exposure required to close and open the valve once per fuel cycle.
(3) The core-melt frequency reduction of 1.2 x 10-5/RY results in avoidance of occupational exposure associated with core-melt cleanup operations (20,000 man-rem/core-melt).64 The accident avoidance dose over the remaining plant life is (24 years)(1.2 x 10-5/RY)(20,000 man-rem/core-melt) or 5.8 man-rem/plant.
(4) The present worth cost of a core-melt accident is estimated at $1.65 billion considering cleanup and replacement power cost over a ten-year period.64 Based on this value and the resultant core-melt frequency reduction associated with implementation of the proposed resolution, the accident avoidance cost was estimated to be about (24 years)(1.2 x 10-5/RY)
($1,560M/core-melt) or $0.475M/plant.
(5) Since the issue was limited to a relatively small number of older PWRs and the configuration of the refueling canal and drains vary significantly between these plants, the ultimate resolution of the issue would require
a plant-specific analysis for each of the affected plants. Therefore, it was recommended that the affected licensees be sent a 10 CFR 50.54(f) letter requesting certification that:
(a) During power operation, each plant operates with refueling water canal drains open so that water from the canal drains to the containment sump following a LOCA and that the drains are sized so that the return flow to the containment sump is adequate to maintain the required net positive head to the recirculation pumps, or
(b) The ECCS analysis for each plant, performed as required by 10 CFR 50.46, considered the amount of water that could accumulate in the refueling canal and that recirculation flow can be maintained without the water accumulated in the canal.
This suggested approach would have been consistent with existing regulations and would have eliminated the necessity of attempting to perform generic risk analyses on what appeared to be a very plant-specific problem.
Based on the above the value/impact score, this issue as given a high priority ranking and was later RESOLVED by NRR with the issuance of Information Notice No. 90-191296; no new requirements were established.