Resolution of Generic Safety Issues: Issue 146: Support Flexibility of Equipment and Components (Rev. 2) ( NUREG-0933, Main Report with Supplements 1–34 )
This issue was identified1477 by NRR when concerns were expressed that the seismic loading on equipment and pipe-mounted components may have been underestimated. These concerns could be divided into two sub-issues:
Sub-Issue 1: On-site tests from the Lotung soil-structure interaction study and dynamic tests performed on concrete expansion anchor bolts1478,1480 indicated that, in certain cases, some of the existing regulatory guidelines and industry practice on seismic design regarding the stiffness of anchorage in equipment and piping might potentially lead to underestimation of their seismic loadings.
Sub-Issue 2: Substantial relaxation of anchorage bolt torque could occur after years of operation causing reduction in equipment and piping natural frequency; this could lead to higher seismic loading on equipment and piping than originally estimated.
Sub-Issue 1: SRP11 Sections 3.7.2, 3.9.2, and 3.10 provide guidance and criteria for seismic analysis of seismic subsystems (equipment and piping) when connected to the supporting seismic systems (buildings and structures). In order to simplify analytical effort, the common industry practice is to assume rigid and fixed attachments between the seismic subsystems and the supporting seismic systems. This means that the anchorage of the seismic subsystems (anchor base plate, anchor bolts or through-bolts, concrete and reinforcement in the area) is assumed to have rigid body characteristics. Therefore, the influence of the stiffness of anchor base plate and anchor bolts on the dynamic response is neglected. For a large number of anchoring systems, especially for those seismic subsystems having only a small mass as compared to that of the supporting seismic systems, this assumption is appropriate. However, in some cases, particularly in heavy equipment, this assumption can potentially lead to underestimation of seismic loadings.
Sub-Issue 2: After years of operation, relaxation of anchorage bolt torque could occur, causing reduction in equipment and piping natural frequency. Since the seismic loading on equipment and piping is determined on the basis of the response spectrum curve, this shift of seismic subsystem natural frequency could potentially increase the seismic response of equipment and piping.
In both the above-mentioned cases, the possible underestimation of seismic loadings could result in malfunction or failure of safety-related equipment, especially the heavy ones, thereby potentially leading to increased CDF and risk to the public.
Tests and collection of seismic data indicated that piping systems designed under existing Code provisions appeared to be quite rugged against seismic excitation. Therefore, the concern of this issue was limited to equipment anchorage only, especially those for heavy equipment. Possible solutions to the two sub-issues were as follows:
(1) Develop guidelines for the re-analysis of equipment anchorage.
(2) For safety-related seismic subsystems, especially heavy equipment, perform plant-specific re-analysis of anchorage considering the stiffness of anchor base plates and anchor bolts on the dynamic response, the local response of the supporting building structure, etc. A bounding study may be needed as this analysis most likely will involve some sort of detailed finite element analysis and could be very costly.
(3) Modify equipment hardware, if required as a result of re-analysis.
(1) The Seismic Qualification Utility Group (SQUG), with NRC participation, developed detailed plant-specific walk-down review guidelines (review of expansion anchor bolts being part of this) for the resolution of Issue A-46. The IPEEE will also utilize the SQUG review guidelines in its plant-specific walk-down reviews. Thus, this sub-issue was addressed in the resolution of Issue A-46 and the implementation of the IPEEE.
(2) For new designs, NRC was expected to conduct research on anchor bolt capacity and dynamic behavior for individual and groups of anchor bolts. When definitive recommendations were available, suitable changes to the SRP11 were to be implemented.
Separate analyses were made for operating and future plants.
As mentioned above, tests performed and collection of seismic data indicated that piping systems designed under existing Code provisions appeared to be quite rugged against seismic excitation. Therefore, the concern of this issue was limited to equipment anchorage only, especially those for heavy equipment. It was assumed that this issue affected all 111 nuclear plants in operation at the end of 1991 (74 PWRs and 37 BWRs). These plants were assumed to have an average remaining life of 24 years.
It was assumed that none of the 111 plants had done a re-analysis of the anchorage stiffness of seismic subsystems. Further, it was conservatively assumed that, as a result of re-analysis, half of the 111 plants would require some degree of hardware modification.
The source of quantitative risk information was a study performed1065 to identify seismic risk-sensitive components in nuclear power plants during and after a seismic event. This study used PRA methodology to expand seismic risk-sensitivity analyses by accounting for seismicity and component fragility data taken from existing nuclear power plant PRAs. This modified PRA analysis consisted of event trees with seismic-induced initiating events (such as seismic-induced small LOCA, transient, LOSP, etc.), and seismic as well as non-seismic related failures in the core damage sequences. It was estimated that for the base case, the affected annual CDF from seismic events was about 4.2 x 10-6/RY for a hypothetical hybrid PWR plant and 9.1 x 10-5/RY for a hypothetical hybrid BWR plant. However, based on the findings of the more up-to-date NUREG-1150,1081 this estimation1065 for a PWR plant was about one order of magnitude too low and, therefore, was revised to 4 x 10-5/RY for this analysis. In addition, the estimated annual CDF from seismic events for the hypothetical hybrid BWR plant used in NUREG/CR-33571065 was atypical due to seismic inadequacy in the plant-specific building structures and, in view of NUREG-1150,1081 was also reduced to 4 x 10-5/RY. Therefore, applying these corrections resulted in an annual CDF estimate of about 4 x 10-5/RY for both PWR and BWR plants.
Because of the above-mentioned atypical seismic inadequacy in building structures used for the hypothetical hybrid BWR plant in NUREG/CR-3357,1065 it was assumed that the results of that study for a PWR will be applicable and representative of all operating plants. The rest of this evaluation was based on this assumption.
In NUREG/CR-3357,1065 it was stated that, for the hypothetical hybrid PWR plant studied, building structural failures contributed 38% to the annual CDF from seismic events; the remaining 62% was contributed from failures of equipment (60%) and piping (2%).
To estimate the risk reduction achievable, the adjusted case of NUREG/CR-33571065 assumed upgrades ("hardening") to various equipment and piping systems such that there was a factor of 5 increase in the median peak ground acceleration (the level of peak ground acceleration at which a component has a 50% probability of failure) for these equipment and piping systems. It was assumed that all the upgrades to the equipment were done to the equipment anchorages where the anchorages were modified to be rigid.
To calculate the changes in CDF that could result from implementation of the possible solution, the contribution from equipment to base case seismic CDF (i.e., base case CDF from seismic events) needed to be calculated first. The contribution to the base case seismic CDF from equipment was (4 x 10-5/RY)(0.60) = 2.4 x 10-5/RY.
The reductions in CDF that result from implementation of the solution were obtained by multiplying the above numbers by a reduction factor of 0.28 to CDF due to equipment upgrade.1065 Therefore, the CDF reduction was given as follows:
|CDF||= (contribution to base case seismic CDF from equipment) x (reduction factor to CDF due to equipment anchorage upgrade)|
|= (2.4 x 10-5/RY)(0.28)|
|= 6.8 x 10-6/RY|
The public risk reduction associated with the solution was obtained from the product of the change in CDF and public dose consequences. Because the core damage sequences encompass all release categories, a weighted average consequence value was developed to approximate the consequences of a core damage accident. The weighted average consequence value of 4.9 x 106 man-rem/event was calculated for all affected release categories using the dose consequences from NUREG/CR-2800.64 Thus, the potential public risk reduction was 34 man-rem/RY.
It was conservatively assumed that after evaluation, half of the 111 plants would need some sort of hardware modification to equipment anchorage. Thus, the total estimated public risk reduction resulting from implementation of the possible solution was (34)(55)(24) man-rem or 45,000 man-rem.
Industry Cost: Implementing the possible solution at each plant would be done in two parts: (1) perform an equipment anchorage evaluation to assess the effects of actual anchorage stiffness on equipment seismic response; and (2) modify the anchorage if needed.
It was assumed that 2 man-years at a cost of $2,270/man-week would be needed to perform the re-analysis at each plant. Further, it was assumed that other costs for re-analysis (computer costs for finite element analysis, retrieval of original equipment anchorage design information, etc.) would be $100,000. This resulted in a cost of $340,000/plant and a total of $38M for all 111 plants.
For the 55 plants that required some degree of anchorage hardware modifications, the cost/plant was based on the estimate of equipment support modification in Table 1 of NUREG/CR-3875.1064 This cost of support modification included two parts: labor (including engineering time) and material. For heavy equipment such as large valves, large pumps (both horizontal and vertical), and diesel generators, the average cost ranged from $2,000 to $50,000. A total cost of $500,000/plant was assumed. A 10.08 adjustment factor was applied for labor productivity effects for work in radiation zones and congested areas, manageability, and access/handling difficulties.961 Assuming of the hardware modification would encounter these difficulties resulted in a cost of $2M/plant or $111M for the 55 affected plants.
Therefore, the total plant implementation costs were $149M and included $38M for the analyses and $111M for the hardware modifications and associated labor.
NRC Cost: NRC implementation of the solution could be quite extensive. The work scope needed to develop the equipment anchorage re-analysis screening criteria and/or re-analysis guidelines could include the following: (1) relevant literature search and evaluation of the potential impact; (2) actual study of major subsystems of selected plants, including reviewing the models, assumptions, and calculations; (3) if needed, some small-scale experiments to confirm the results of the above studies; and (4) recommendations to improve the existing conditions in operating plants.
Items 1, 2, and 4 were estimated to require approximately 2 man-years of labor to develop, review, and approve. At a rate of $100,000/man-year, the estimated cost was $0.20M.
Item 3 was estimated to require 1 man-year of labor to prepare the test procedures, QA activities, and analysis of test results. Labor costs were estimated at $100,000/year, for a total program cost of $0.35M. A generic letter directed to all operating plants would be required. It was estimated961 that the labor requirement to issue a generic letter was 6 man-weeks. At a rate of $2,270/man-week, this cost would be $0.01M.
NRC reviews of licensee submittals in response to the generic letter were assumed to be 5 man-weeks/plant. At $2,270/man-week, the total cost for 111 plants was $1.3M. Thus, the total NRC cost is $(0.20 + 0.35 + O.O1 + 1.3)M or $1.86M.
Total Cost: The total industry and NRC cost for implementation of the possible solution was estimated to be $(149 + 1.86)M or approximately $151M.
Based on a potential public risk reduction of 45,000 man-rem and a cost of $151M, the impact/value ratio was given by:
(1) Using the assumptions in "Industry Cost," it was estimated that about 30,000 man-hours/plant would be required for hardware modification in radiation zones. Assuming a dose rate of 0.025 rem/hr,64 the total occupational dose for 55 plants was estimated to be (55 plants x 0.025 rem/hr x 30,000 man-hour/plant) or 41,000 man-rem.
(2) The occupational dose reduction due to accident avoidance was calculated from the reduction in CDF multiplied by the assumed accident dose of 19,860 man-rem.64 The possible solution would reduce the CDF in 55 plants that have an occupational dose rate reduction of (19,860 man-rem x 6.8 x 10-6/RY) or 0.14 man-rem/RY. With the 55 plants having an average remaining life of 24 years, this resulted in a best estimate total occupational dose reduction due to accident avoidance of 185 man-rem.
(3) The accident avoidance cost savings were estimated to be the product of the CDF reduction (6.8 x 10-6/RY), the estimated cost of a core-melt accident ($1.65 Billion), and 55 plants with an average remaining life of 24 years. This resulted in a total cost of $14.8M.
It was assumed that there would be 10 future plants and half of them would need some sort of hardware modification; future plants were defined as plants seeking licenses (CP/OL) or combined licenses (per 10 CFR 52). Furthermore, the usual assumption of 40-year plant life was assumed for both types of plants.
Just as in the case of the backfit analysis for operating plants,
|Reduction in CDF||= 6.8 x 10-6/RY|
|Public Risk Reduction||= 34 man-rem/RY|
For 5 plants with an average remaining life of 40 years, the total public risk reduction was (34)(5)(40)man-rem or 6,800 man-rem.
Industry Cost: Assuming that the cost/plant was the same ($340,000) to perform the seismic analysis during the design stage to consider the effects of equipment anchorage stiffness, the total cost for 10 plants was estimated to be $3.4M.
NRC Cost: It was assumed that the NRC costs would be $0.2M to develop evaluation guidelines, $0.35M for a test program, and $0.5M to update relevant Regulatory Guides and SRP11 Sections. Thus, the total NRC cost was estimated to be $(0.2 + 0.35 + 0.5)M or $1.05M.
Total Cost: The total industry and NRC cost associated with the possible solution was $(3.4 + 1.05)M or approximately $4.45M.
Based on a potential public risk reduction of 6,800 man-rem and a cost of $4.45M, the impact/value ratio was given by:
(1) The equipment seismic experience data base collected by SQUG in the resolution of Issue A-46 could provide some clue regarding the performance of equipment under seismic loadings. This data base ("Summary of the Seismic Adequacy of Twenty Classes of Equipment Required for the Safe Shutdown of Nuclear Plants," EQE, Inc., February 1987) was collected from over 100 industrial facilities (fossil and hydroelectric power plants, electric distribution stations, petrochemical plants, and other large industrial facilities) in the strong motion areas of 13 earthquakes that have occurred in the U.S. and Latin America since 1971. It indicated that heavy equipment (e.g., large pumps, valves, large diesel generators, etc.) generally performed well under earthquake loadings, in spite of the fact that seismic loadings were not considered in the original designs for most of the plants in this data base.
(2) For U.S. nuclear power plants, seismic loadings were required by the NRC to be considered in the original design, even though the local anchorage stiffness effects on equipment were not routinely evaluated. Furthermore, for existing OL/CP plants, the accident avoidance cost savings and occupational dose reduction due to accident avoidance were not significant, when compared to the cost and doses used in the impact/value score, and the occupational dose increase (41,000 man-rem) due to implementation of the possible solution was nearly the same as the best estimate public risk (45,000 man-rem) that could be reduced.
(3) For future plants, the total reduction in public risk would be less than that for existing plants since it was assumed there would be fewer future plants than the existing plants. However, for future plants, the impact/value consideration would be more favorable since equipment anchorage stiffness effects could be evaluated in the design and analysis stage and backfit would not be required. In addition, the ORE to forward-fit future plants was limited to operation and maintenance and should be minimal.
Based on the impact/value assessment and total public risk reduction potential and using the revised guidelines of SECY-93-108,1479 this issue had a borderline medium priority ranking for operating plants and a medium priority ranking for future plants. However, the seismic PRA model used in NUREG/CR-33571065 (which was issued in 1983 without the benefit of the SQUG equipment seismic experience data base) had inherent large uncertainties associated with it. Furthermore, as discussed above, other factors such as the large occupational dose increase due to implementation of the resolution needed to be considered. When all these factors were considered together with the impact/value assessment and total reduction in public risk, the priority ranking was judged to be low for operating plants and medium for future plants.
For operating plants, no further action was needed on Sub-Issue 1 and Sub-Issue 2 was covered in the implementation of the resolution of Issue A-46 as well as in the IPEEE Program.1222 For future plants, the following recommendations were made:
(1) Sub-Issue 1: There was a need to reflect the uncertainty in anchorage stiffness assumption for heavy equipment. Pertinent sections of the SRP11 should be revised to adequately treat anchorage stiffness assumptions.
(2) Sub-Issue 2: NRC was conducting research on anchor bolt capacity and dynamic behavior for individual and groups of anchor bolts. When definitive recommendations were available, appropriate changes to the SRP11 should be made.
In an RES evaluation,1564 it was concluded that consideration of a 20-year license renewal period did not change the priority of the issue.