Resolution of Generic Safety Issues: Issue 113: Dynamic Qualification Testing of Large Bore Hydraulic Snubbers (Rev. 2) ( NUREG-0933, Main Report with Supplements 1–34 )
This issue was raised1014 in March 1985 to address the staff's concern that there were no requirements for dynamic qualification testing or dynamic surveillance testing of large bore hydraulic snubbers (> 50 kips load rating). The safety concern was the integrity of the steam generator lower support structures (SGLS) when subjected to a seismic event.1015 However, the issue was applicable to all LWRs with components, structures, and supports that rely on LBHS for seismic restraint and other dynamic loads such as high energy line breaks and water hammers.
In the absence of the restraint to the steam generators provided by the LBHS, the steam generator support structures (SGS) might fail. Failure of the SGS might subsequently result in rupture of the primary system piping (large-break LOCA), the main steam lines (MSLB), and the feedwater (normal and auxiliary) piping lines. Such failures could result in a core-melt from the loss of all means of core cooling and could pose a significant risk to the public. Other dynamic load events could further increase the safety significance of this issue but, this limited analysis focused primarily on the seismic concern raised.
The staff suggested1014 a number of tests or alternative tests to provide adequate assurance of the operability of the LBHS when subjected to a seismic event. The test options primarily focused on dynamic cyclic testing, to assure operability of the snubber control valves when subject to cyclic loads, and the determination (or correlation) of the snubber system spring rate when subject to cyclic loads. As previously stated, the resolution may effect all operating plants (BWRs and PWRs) that use LBHS as seismic restraint devices.
Initiating Frequency: As stated above, the initiating event evaluated in this analysis focused on the potential seismic-induced movement of steam generators in PWRs. The probability of failure of the SGS is 0.051016 for a peak ground acceleration of 0.5g, which is approximately three times that of an SSE. The SGS failure probability corresponds to the failure probability of 0.05 for hydraulic snubbers from all design causes.1017 Thus, given a failure of the LBHS in the SGS, the SGS under a 3 SSE loading was assumed to have a failure probability of 1. Assuming the failure probability is proportional to the load, the failure probability of SGS and LBHS subject to a SSE loading is therefore 0.017.
The function of a LBHS during an earthquake is to lock-up and to resist motion of the steam generator. Failure to lock-up in either the compression or tension stroke will result in loss of snubber restraint (soft snubber strain rate under repetitive input loading). For this analysis, the inertial time lag (rocking) of the steam generator during its inertia-induced motion in one direction (snubber compression mode on one side and tension on the other side of the steam generator) combined with failure of the snubber to lock-up and resist the inertia-induced movement of the steam generator was considered. Thus, only one-half of the seismic input frequency was involved in the relative motion (rocking) between the steam generator and the snubber rigid attached wall.
Failure of the snubber from sticking of the control valve accounts for approximately 1% of the tested snubber failures.1017 Assuming that the control valves are as likely to stick open (failure to lock-up) as to stick closed (failure to unlock), the LBHS failure to lock-up is (0.017)(0.01)(0.5) = 8.5 x 10-5/demand.
The strong ground motion of the SSE was assumed to contain an input frequency of 33 cycles/second over a duration of 10 seconds. Considering only one-half the input frequency as discussed above, the snubbers could experience (0.5)(33)(10) = 165 demands. The probability of the LBHS and SGS failure, given an SSE, would therefore be (8 x 10-5)(165) = 0.014. A possible conservatism in this assumption is that all the LBHS in the SGS ganged (grouped LBHS arrangements) sets of snubbers are assumed to fail as one composite LBHS failure. Such a common mode failure was assumed to be representative of a generic design defect that results from the absence of adequate dynamic testing programs.
Core-Melt Frequency: Given an SSE event with a return period of 2 x 10-4/RY (References 1018 and 1019), and the conditional failure probability of a SGS/SSE as (1.4 x 10-2), the core-melt frequency was estimated to be (2 x 10-4/RY)(1.4 x 10-2) = 2.8 x 10-6/RY.
Containment Failure Frequency: For this analysis, the containment failure probability was assumed to be 1 due to overpressurization from the high energy released into the containment from the piping failures or containment bypass by way of the ruptured steam lines.
Based on the above frequency estimates, the probability of a large release from a core-melt caused by dynamic (cyclic) failure of the LBHS during a SSE is (2 x 10-4/RY)(1.4 x 10-2)(1) = 2.8 x 10-6/RY. The public dose within a 50-mile radius of the plant, with a surrounding uniform population density of 340 persons per square mile, no evacuation, and meteorology typical of the Braidwood site is (2.8 x 10-6/RY)(5.1 x 106 man-rem) = 14.3 man-rem/RY. Assuming an average remaining plant life of 30 years, the potential public risk was approximately 430 man-rem/reactor.
The cost of the proposed solution(s) will be highly dependent on the option selected to verify the dynamic capability of the LBHS in operating plants, i.e., a snubber vendor qualification of snubber types and/or in-plant tests that augment the current TS functional test requirements. For operating plants, the cost will be highly dependent on the state-of-art of test equipment, the number of snubbers tested per plant, the surveillance frequency of the tests, the existence of or lack of prior qualification tests (snubber vendor-specific), the vintage and distribution of various vintage LBHS in the plants, and replacement power costs (should the LBHS tests result in extended plant outage time).
The expected large variations in all the above elements necessary to arrive at a realistic cost estimate for this issue clearly indicated that the costs used in this analysis must be regarded as very rough estimates.
Vendor Qualification Tests: The average cost for snubber qualification tests (including dynamic testing) was estimated to be $100,000 per snubber type.1017 This cost may be significantly higher per snubber type for the smaller population LBHS, but insignificant on an average per-plant basis when compared to other industry costs. Further functional (in-plant) tests of the LBHS that might augment the existing TS requirements, given an adequate vendor qualification testing program (including dynamic testing), may be lower than the in-plant tests cost estimated in this analysis.
In-Plant Testing: The annual testing cost for hydraulic snubbers was estimated to be approximately $1000/snubber. Assuming the snubber population ranged from 500 to 1000 snubbers/plant, and 15% of the snubbers were LBHS,1017 each plant may have approximately 75 to 150 LBHS.1020 Based on the existing TS functional testing criteria,1020 it was assumed that 20% to 25% of the LBHS will be tested per refueling outage (approximately every 1.5 years). This amounted to 11 to 23 LBHS on an annual basis.
The existing test requirements for LBHS were estimated to cost approximately ($1000)(0.15)(75 to 150) or $11,000 to $23,000 per RY or, on an average, approximately $17,000/RY. Estimating that a dynamic testing requirement (including setup, tests, and equipment leasing) would increase the existing LBHS test cost by 50% to 100% yielded an increased cost of approximately $8,500/RY to $17,000/RY. The present worth costs, at a 5% discount rate over 30 years, ranged from $131,000 to $262,000 per plant. These costs would be attributed to an in-plant dynamic testing requirement for LBHS only.
Replacement Power: Cost factors related to a hydraulic snubber test program according to the TS were cited in NUREG/CR-4279.1017 The TS snubbers testing phase resulted in extending the plant outage time by approximately 3 days. Assuming the existing TS functional test surveillance requirements are linear with respect to the LBHS population (15%), the outage extensions due to the current LBHS testing may be extended an additional 0.45 day. If it is assumed that the outage is extended only one additional hour per LBHS tested (11 to 23 LBHS), the outage extension ranges from one-half day to one day per year, which is consistent with the above estimate. Therefore, the estimated replacement power cost of $500,000/day yields an annual replacement power cost of $250,000 to $500,000/plant-year. Based on a 5% real discount rate, the present worth replacement power cost over 30 years may be $3.85M to $7.7M per plant.
NRC Cost: The estimated NRC cost for this issue ranged from $50,000 to $100,000, including technical assistance contractor costs. The effort would likely involve a review of the LBHS used in industry, determination of the need (risk reduction) for additional LBHS test requirements, discussions with snubber vendors, development of acceptable testing requirements, and possible preparation of additional TS requirements.
Total Cost: The NRC cost ($50,000 to $100,000) would be insignificant when compared to the industry costs. A per plant cost for vendor qualification of each snubber type would likely be distributed over the total population of the tested snubber type and not significantly affect total industry costs. In addition, some LBHS types may have existing and adequate testing programs. The present worth of surveillance testing costs ($131,000 to $262,000) and replacement power cost ($3.85M to $7.7M) yielded a cost that ranged from approximately $4M to $8M per plant. Therefore, the average cost to implement a dynamic testing requirement in operating plants for the LBHS was estimated at approximately $6M/plant, if plant outage time is extended because of the additional tests. If the LBHS test can be done within normal plant outages (refuelings), the total cost would be approximately $200,000/plant.
Based on a public risk reduction of 430 man-rem/reactor and an estimated average cost of $6M/reactor (including replacement power costs), the value/impact score was given by:
If replacement power costs are not involved, the value/impact score would be:
(1) The uncertainties in this analysis were large for the risk estimates. The risk for this issue was estimated to result from the absence of a LBHS dynamic qualification test requirement, or a LBHS dynamic functional surveillance test program in the TS. The risk estimates center on failure of the SGS due to a common mode failure of the LBHS in the event of a SSE seismic excitation. In estimating the probability of failure of the LBHS, it was assumed that the LBHS control valves may fail open during seismic excitation, e.g., due to cyclic (frequency) loading, and thereby result in free motion (lack of snubber restraint) and lateral movement of the steam generators. Lateral movement of the steam generators (absent LBHS restraint) was assumed to fail all the SGS and result in massive piping failures in the primary and secondary piping runs.
(2) The existing TS functional test requirements provided some unknown amount of assurance relative to cyclic operability to the LBHS, but the strain rate (k) during repetitive loadings was not assured and therefore may be even more uncertain. The analysis did not treat the above uncertainties in the estimates because of the lack of supportive data. Previous qualification testing (if dynamic tests were performed) might negate the need for further strain rate (dynamic) testing since it could be inferred that the operability and repetitive loading capacity for a given design was confirmed by the qualification tests and only operability need be verified in subsequent tests.1021
In the absence of any correlation between the LBHS operability and strain rate capacity when subject to dynamic loading, the assumption that control valves may stick open during the cyclic loading, and remain open, assumed a generic design defect might exist. This inherent assumption in the analysis may overestimate the probability of LBHS failure by one to two orders of magnitude. Thus, the risk estimates may be high for the analyzed SGS structure failure scenario.
(3) The LBHS are also used to support other components in nuclear power plants. The consequences (plant damage states) from other dynamic loads or other accident scenarios that might be similar to the seismic-induced LBHS failures are assumed to be dominated by the seismic-induced SGS structure failure scenario. A more detailed analysis that examines all the potential plant damage states that may result from LBHS failures could be considered in the final resolution.
(4) The resolution and implementation of the NRC Piping Review Committee recommendations (see Issue 119) may result in removal of some of the piping snubbers. Thus, more flexible piping systems may result in higher nozzle loads to the steam generators and other support structures during seismic events.1018 In such cases, the reliability of the remaining smaller population of snubbers (including LBHS) may become more critical. As a further example, the GDC-4 limited scope rule would allow removal of many snubbers (including LBHS) from PWR reactor coolant system piping and large components, such as steam generators, subject to NRC approval of demonstrated acceptability of the licensees' requests. Additionally, the proposed GDC-4 broad scope rule change, if approved, will expand the scope of the limited scope rule to all high energy piping systems in all nuclear power plants. Thus, the determination of the need for dynamic qualification testing of the LBHS should also consider the potential impacts of Issue 119, and specifically the GDC-4 rule changes.
(5) Some of the uncertainties in the cost estimates were due to the variations in snubber vendors' qualification programs, from plant to plant design differences, whether or not plant refueling outage times would be extended due to the additional testing, and other elements discussed above. Therefore, a wide uncertainty band for costs existed and no precise generic cost estimate seemed appropriate for this issue, i.e., the costs may vary significantly from plant to plant. As previously stated, the cost estimates should be regarded as very rough estimates. The value/impact ratio of this issue will be strongly influenced by any outage extension that may or may not result from additional LBHS testing.
The purpose of this issue was to assess the need for an NRC requirement for dynamic qualification testing of LBHS in operating plants. The limited assessments provided in this analysis should only be considered as rough baseline risk, cost, and value/impact estimates. Further and more detailed analyses may show either higher or lower values. However, this analysis identified that a broader and more complete evaluation was needed to resolve the issue. Based on the estimates determined in this analysis, the potential need for higher reliability LBHS (pipe snubber removal and optimization programs were being pursued), and the observed failures of LBHS in operating plants,1015 this issue was given a high priority ranking (See Appendix C). Work being done by RES in the Nuclear Plant Aging Research,1022 the resolution of Issue 119, and the effects of the GDC-4 rule changes were to be considered in the resolution of this issue.
In resolving the issue, the staff concluded that there were few cost-beneficial changes to existing requirements that would result in improved LBHS reliability and reduced risk. However, it was recommended that a Regulatory Guide be developed for future plants. Thus, this issue was RESOLVED and no new requirements were established.1450 In an RES evaluation,1564 it was concluded that consideration of a 20-year license renewal period did not affect the resolution.