Resolution of Generic Safety Issues: Item B-59: (N-1) Loop Operation in BWRs and PWRs (Rev. 1) ( NUREG-0933, Main Report with Supplements 1–34 )
The majority of the presently operating BWRs and PWRs are designed to operate with less than full reactor coolant flow. If a PWR RCP or a BWR recirculation pump becomes inoperative, the flow provided by the remaining (N-1) loops is sufficient for steady state operation at a power level less than full power. Although the FSARs for the licensed BWRs and PWRs present (N-1) loop calculations showing allowable power and protective system trip set-points, the NRC staff has disallowed this mode of operation for most plants primarily due to insufficient analyses. At present, BWR and PWR licensees have TS which require shutdown within a fairly short time if one of the reactor coolant loops becomes inoperable. This issue was originally identified in NUREG-0471.3
Allowing (N-1) loop operation gives utility operators more flexibility in deciding whether to shut down a plant or let it operate at a reduced power level.
In this issue, (N-1) loop operation is restricted to operation during a single RCP failure. When fixing an out-of-service pump becomes a major task, it is not expected that the pumps will be repaired while the plant is on-line. By continuing operation in the (N-1) mode, the repair work may be postponed until a scheduled refueling time.
In connection with MPA E-05, a SER773 was completed in July 1984 for the request by Beaver Valley Unit No. 1 (BV-1) for (N-1) loop operation. Based on this SER, it is expected that BV-1 will be authorized to operate with (N-1) loops when the TS are revised and updated appropriately in the near future.
The SER for BV-1 represents the resolution of this issue for PWRs. For PWRs, DL is expected to close out MPA E-05 on the basis that there are no other active PWR applications for (N-1) operation and, further, that none is expected in the foreseeable future.774 On the other hand, MPA E-04 covers 10 licensing actions on BWR submittals for (N-1) or single loop operation (SLO) for 7 licensees.
The staff has reviewed the requests and submittals from the BWR licensees and has approved them such that (N-1) loop operation for BWRs would be authorized if the licensees submit the appropriate TS changes. The question of potential thermal-hydraulic instability problems during SLO for BWRs and how restrictive the TS changes would have to be was raised775,776 by the staff, but this issue was resolved in Item B-19. However, in an effort to resolve certain plant-specific concerns about thermal-hydraulic instability in the Browns Ferry plant, TVA completed tests at Browns Ferry on February 9, 1985, and those concerns have been resolved. The tests demonstrated that TS based on GE SIL-380, which have been proposed for some BWRs and approved by the staff, are unlikely to result in any limitation on the achievable power level in SLO. They also indicated (pending verification by data analyses) that SLO is not significantly less stable than two-loop operation under similar power/flow operating conditions.
Permanent SLO has been approved for Peach Bottom Unit 3, Quad Cities Units 1 and 2, and Dresden Units 2 and 3 and will soon be approved for Duane Arnold. The staff expects to approve permanent SLO for the SLO applicants when appropriate TS changes have been submitted.
In the event that a loop becomes inoperative in an operating plant, it is not always feasible to place it back in service by the repair of the failed pump while the plant is on-line. The plant operation with the (N-1) loops, however, will not differ from operation with all loops, except for the requirement to operate at a decreased power level for the lower flow condition and with corresponding instrument/control set-point limitations. The accident sequences would be essentially the same as with all loops in operation and there will be no change in accident initiator frequencies. Moreover, the loss of a loop because of pump malfunction would not impair the function of the ECCS and the other on-demand systems should an accident initiator arise. There had been some concern that operation with one loop out of service could result in thermal-hydraulic instabilities and possibly core damage at low flow conditions as well as with jet pump vibration at high flow conditions that could lead to damage to the reactor internals; but this matter has been adequately resolved. Therefore, the resolution would affect public risk or ORE only slightly and might reduce risk because power and fission product levels would be smaller than at full power. The purpose of this change is to reduce the impact on licensees.
The purpose of this task is to develop a set of acceptance criteria, review guidelines, and TS changes for the (N-1) loop authorization requests. This set of criteria, guidelines, and TS changes will encompass accident scenarios (both LOCAs and non-LOCAs) to be analyzed by the licensees, computer models acceptable to NRC for these analyses and acceptable input parameters in terms of reactor operating conditions (such as allowance for uncertainties in power level and fluid measurement). This has already been accomplished for PWRs by virtue of the completion of the SER773 for BV-1. In addition, the BWR analyses have been reviewed and accepted by the staff and the appropriate generic TS changes to allow SLO for BWRs have been identified.
The analysis will be limited here to (N-1) loop operation for BWRs since the issue is essentially inactive for PWRs. The estimates provided below were based on calculations performed by PNL.64
When operating a nuclear plant at a power level proportional to a reduced number of loops, the safety margins are somewhat increased from those at full power, but this increased margin is not regarded as contributing to a significant reduction in risk. Therefore, any potential risk reduction associated with this issue is perceived to be negligible. Moreover, no additional ORE is anticipated for this issue inasmuch as major loop repair is likely to be done during scheduled downtimes.
Industry Cost: To estimate the cost to industry, it is assumed that the amount of work performed on BV-1 by the licensee to analyze plant performance will be comparable to that required for BWR plants.772 BV-1 analyzed a (N-1) loop large break LOCA and 12 non-LOCAs. Accidents involving the partial loss of forced reactor coolant flow, startup of an inactive reactor coolant loop, single RCP locked rotor, and complete loss of forced reactor coolant flow were analyzed in the original FSAR. Therefore, they were not reanalyzed. This leads to 13 transient scenarios to be analyzed.
Using a resource requirement of 5 man-wks and 15 computer hours for each case leads to a total of 65 man-wks and 195 computer hours. Another 30 man-wks/plant are allowed for preparing TS changes, modifying and upgrading procedures and/or systems, and familiarizing operations staff with upgrades. Using the industry rate of $2,270/man-wk and an estimated computer cost of $1,000/hr, the total implementation cost is estimated to be approximately $420,000/plant for BWRs and PWRs. In addition, the plant-specific tests run by TVA at the Browns Ferry plant on the weekend of February 9, 1985, required operation at reduced power ranging from 50% to 65% for about 6 hours. On this basis it is estimated that the cost to conduct the test, obtain replacement power, and reduce the data will not exceed $150,000.
The labor and analysis required for operation and maintenance of the resolution of this issue by the licensees is estimated to be negligible.
With the implementation of (N-1) loop operation, plant downtime can be reduced. The results of EPRI NP-1138431 and EPRI NP-2094114 indicate that the main contributor to (N-1) loop operation are pump seal failures. For Oconee 1, 88% of pump failure events are due to pump seal problems and 99% of pump maintenance time is on seal fixes. Since the non-seal failures only contribute 1% of the total maintenance time in the Oconee case, we ignore them for this analysis and use pump seal failure probability as the probability of losing one loop and operating under (N-1) loop conditions. While the failure frequency of PWR pump seals that contribute significantly to core-melt frequency is only 0.02/RY,366 seal failures that result in the loss of one loop are estimated to be at a rate of 0.5/RY for both PWRs and BWRs. Some of these are LOCAs or would become LOCAs, but these can be isolated by the BWR recirculation loop valves.
If a plant is base-loaded, it is more economical to shut down and repair a seal that fails more than 20 days before the end of a 540-day cycle than continue with one-out-of-two loop reduced power operation. But all plants are not run at full power. Also one loop operation allows flexibility in shutting down to make repairs. Therefore, it is assumed that out of the 0.50/RY events, at best only one-third of the events will be continued as (N-1) loop operation, i.e., 0.17/RY.
The savings in terms of the avoided outage is estimated113 to be 10 days (average extra outage time per pump seal failure). Therefore, the savings from avoidance of outage per reactor-year is given by:
(10 days)(1 loop/2 loops)($300,000/day)(0.17/RY) = $255,000/RY
It is noted that the licensee implementation phase covering the analysis and evaluations of the potential accident sequences have already been submitted to the NRC in many cases for BWRs. Therefore, it will be assumed that the remaining plant analyses and/or re-analysis that may be required by the NRC staff will affect one-half of the total number of reactors. Also, it will be assumed that the cost savings resulting from the avoidance of outages with (N-1) loop operation is the same each year for the industry. Further, assuming an average reactor lifetime of 28 years and 44 BWRs, the total industry costs are estimated to be as follows:
|Implementation [(1/2)($420,000)(44 BWRs)]||=$ 9,240,000|
|TVA Test at Browns Ferry||=150,000|
|Operation and Maintenance||=0|
|Outage Avoidance [($255,000/RY)(44 BWRs)]||=11,220,000/yr|
For BWRs, the present worth (PW) of the annual savings from outage avoidance over the average reactor lifetime of 28 years at a real discount rate of 5% is:
PW = ($11.22M)(0.05)-1[1 - (1 + 0.05)-28] = $168.3M
NRC Cost: The cost to NRC of developing a set of acceptance criteria and review guidelines and TS changes for the issue are negligible inasmuch as these have already been identified. Some additional effort will be required to revise SRP Chapter 15 to reflect the criteria needed to review (N-1) loop operation. The revision to the SRP may require approximately 4 man-weeks. In addition, NRC labor to support SER implementation should be minimal at about l man-wk/plant. The total costs to the NRC are:
|Revision to the SRP [(4 man-wk)($2,270/man-wk)]||=$ 9,080|
|Implementation[(1 man-wk/plant)($2,270/man-wk)(44 plants)]||=99,800|
|Development of Resolution||=0|
Thus, the total NRC costs are approximately $110,000.
It is concluded that this is a Regulatory Impact issue which has been resolved for BWRs and PWRs. For PWRs, the issue has been resolved on the basis of the BV-1 SER (MPA E-05) and, for BWRs, the issue has been resolved on the basis of Item B-19, the plant-specific tests at Browns Ferry, and the review of licensee submittals under MPA E-04.