Resolution of Generic Safety Issues: Issue 29: Bolting Degradation or Failure in Nuclear Power Plants (Rev. 2) ( NUREG-0933, Main Report with Supplements 1–34 )
Prior to 1981, the number of bolting-related incidents reported by licensees was on the increase. A large number of these were related to primary pressure boundary applications and major component support structures. As a result, there was concern for the integrity of the primary pressure boundary in operating plants and the reliability of the component support structures following a LOCA or earthquake. This issue was identified by the ACRS.1384
There are numerous bolting applications in nuclear power plants the most crucial of which are those constituting an integral part of the primary pressure boundary such as closure studs and bolts on reactor vessels, reactor coolant pumps, and steam generators. Failure of these bolts or studs could result in the loss of reactor coolant that could jeopardize the safe operation of the plants. Other bolting applications, such as component support and embedded anchor bolts or studs, are essential for withstanding transient loads created during abnormal or accident conditions. A report summarizing bolting failure experience was issued by DL/NRR.184
At the time of this evaluation in 1983, there had been a total of 44 reported bolting incidents most of which were discovered either during refueling outages or scheduled ISI or maintenance/repair outages. These incidents had no immediate impact on public health and safety since they had not resulted in any accident. However, degradation or failure of such studs and bolts constitutes a reduction in the integrity of the primary pressure boundary.
Concern was compounded by the fact that there was no reliable NDE method to detect the cracking or degradation of bolts or studs resulting from the principal modes of failure: stress corrosion, fatigue, erosion corrosion, and boric acid corrosion. Visual examination was the only reliable method to discover degradation by boric acid corrosion or erosion corrosion. In almost all cases, this required disassembly of the component in order to inspect the bolts or studs. If there is no clear evidence of boric acid leakage to the surroundings, bolting degradation by boric acid corrosion can potentially be undetected until the bolts or studs completely fail. Under the existing ISI program, visual inspection of bolts was not a mandatory requirement and UT inspection was not required on pressure-retaining bolts or studs with diameters less than 2 inches. A major accident such as a LOCA could conceivably occur due to undetected extensive bolting failure of the primary pressure boundary.
Because bolting has a wide range of application in nuclear power plants, there was no simple solution to the problem. Therefore, in order to minimize the potential bolting problems in new power plants, improvements in one or all of the following five areas could be recommended: design, materials, fabrication, installation, and ISI. For this analysis, the focus was placed on improving the efficiency and adequacy of ISI programs.
Based on a review of the 44 bolting incidents reported by PWR licensees up to the time of this evaluation, the principal causes of bolting failure or degradation were classified as stress corrosion, fatigue, boric acid corrosion, erosion corrosion, and other types. Nineteen of the 44 incidents were identified as resulting from stress corrosion, the most common cause of bolting failure. Boric acid corrosion was the second most common cause of bolting failure or degradation reported. Twelve resulted from boric acid corrosion. The remaining 13 incidents were either fatigue, erosion corrosion, or other types. No bolting failures in BWRs were reported.
Sixteen of the 44 reported incidents were related to primary pressure boundary bolting applications such as various closure studs in reactor vessels, pressurizers, steam generators, and hold-down bolts in various types of valves. Thirteen incidents related to component support structures, such as the column support or embedded anchor bolts or studs of steam generators, reactor vessels, reactor coolant pumps, and piping restraints, were reported. Although failure of such bolts or studs will not normally impair the normal operation of a plant, extensive failure of such bolts or studs could cause component damage or multiple piping failure under abnormal or accident conditions such as a LOCA or an earthquake. The 44 bolts or stud failures occurred in about 350 RY of experience. Thus, the frequency of corrision-initiated events was 44/350 event/RY or 1.3 x 10-1 event/RY.
Based on experience, there was a good chance that the corrosion would be discovered and the studs replaced before failure occurred. However, it was conservatively assumed that 10% of the bolts or studs will not be discovered before they failed and will result in a small break LOCA (S2). Therefore, the frequency (F) of corrosion-initiated events was estimated to be 1.3 x 10-2 S2 event/RY.
Twenty-nine of the reported incidents or 66% had a direct potential for causing a large-break LOCA due to bolting or stud failure in restaints for large piping, component supports, or steam generator manways when these hold-down devices have degraded to the point that they will not provide the necessary support following a water hammer or seismic event. However, even though the actual determination was complex, the S2 event was believed to be the most limiting.
An S2 event can result in a wide spectrum of consequences, depending on whether or not the engineered safety features are required to function or whether they do function. Using WASH-140016 S2 sequences with the frequency estimated above, the release was determined to be 3 x 104 Curies/RY.
The total whole-body man-rem dose was obtained by using the CRAC Code64 for the particular release category. A uniform population density of 340 people per square mile (average for U.S. domestic sites) and a typical (midwest plain) meteorology were assumed. Therefore, the estimated public risk was 3.5 x 103 man-rem/RY. For 43 plants with an average remaining lifetime of 30 years, the potential risk reduction was 4.5 x 105 man-rem.
Industry Cost: The proposed fix could vary considerably depending on the type and depth of solution sought. However, the most probable fix was to visually inspect bolts or studs using an improved UT inspection technique and a more frequent inspection schedule. This represented an increase in surveillance and would require extra effort during each plant refueling outage. Because of the wide variety of uses of studs and bolts for safety functions in nuclear plants, the actual cost would vary greatly.
Based on the information64 provided, an interim and simple fix would be to inspect studs and bolts only on components that had been opened for inspection or maintenance during a refueling outage. This would require a minor increase in surveillance and would not require an extension of outage time. It was estimated that 3 man-weeks/plant of extra effort would be required per 18-month refueling schedule and 40 man-weeks/plant to allow for administrative overhead. At $100,000/staff-year, the cost (C) of the increased surveillance over the 30-year life of a plant was given by:
If, however, each plant was required to inspect 10% of the bolts or studs in primary system components per refueling outage, whether open for inspection or not, then based on an 18-month refueling schedule, each plant would have inspected 200% of its bolts or studs over its 30-year lifetime. (This inspection frequency should detect any bolt degradation that might occur.) This would represent an increase in surveillance and would require an extension of outage time by 1.5 working days. At $300,000/day for replacement power, the total cost (C) over the plant life was given by:
NRC Cost: The NRC cost was negligible in comparison to the industry cost.
(1) For inspecting bolts on disassembled components only, the value/impact score was given by:
(2) For inspecting 10% of the bolts, the value/impact score was given by:
The uncertainties in the estimates of accident frequencies and consequences were such that, if they lowered the value/impact score by an order of magnitude, the score would still be above a threshold that would warrant resolution of the issue. As the cost estimates increase for specific solutions (particularly when plant shutdown or extended shutdowns are required), the value/impact scores decrease and could affect the priority ranking. If the cost estimates associated with inspecting 10% of the bolts are off by a large factor, the potential risk reduction would still be high.
(1) A secured reactor primary system pressure boundary, which depends on the integrity of the system's piping and components, is an integral part of the defense-in-depth concept embodied in the design of nuclear power plants to protect against a core-melt. Also, some safety system functions rely on a secured pressure boundary to prevent or mitigate the consequences of an event. Accordingly, inspection of lO% of the bolts per refueling outage (200% over the lifetime of the plant) provides assurance that the primary system pressure boundary will not be breached by failed bolts or studs.
(2) When the averted costs of cleanup following a LOCA are considered, the value/ impact scores calculated above become more favorable. It was estimated that the averted occupational dose of inspection versus reduction due to accident dose would fall between a PWR-8 or PWR-9 event and a PWR-l to PWR-7 event. This represented an averted dose between 2,400 to 8,000 man-rem.
Based on the above value/impact scores, this issue was given a high priority ranking (See Appendix C). In resolving the issue, the staff took into consideration previous actions taken by the NRC to address the concerns related to threaded fasteners: Bulletin Nos. 82-02,1129 87-02,1389 and 89-021388; Information Notice Nos. 86-25,1393 89-22,1390 89-56,1391 and 89-701392; and Generic Letter Nos. 87-O21387 and 88-05.1386
The staff's regulatory analysis, NUREG-1445,1398 proved to be inconclusive regarding a mandatory program on safety-related bolting for operating plants. The staff's technical findings were documented in NUREG-13391395 which endorsed the recommendations of independent studies performed by the industry Joint Task Group on Bolting. This group was set up by AIF, EPRI, and the Materials Properties Council and its studies resulted in EPRI NP-5769, "Good Bolting Practices," and three EPRI video training tapes on "Pressure Boundary Bolting Problems."
The staff concluded that leakage of bolted pressure joints was possible, but catastrophic RCPB joint failure that could lead to significant accident sequences was highly unlikely. This conclusion was based on: (1) operating experience; (2) actions taken through bulletins, generic letters, and information notices; and (3) proposed industry actions. Generic Letter No. 91-171385 was issued to licensees to: (1) implement the industry bolting integrity program, as presented in the EPRI report and video tapes; and (2) continue actions in accordance with commitments made in response to NRC generic letters and bulletins. Thus, this issue was RESOLVED and no new requirements were established. However, in order to improve the review of future plants and significant modifications to operating plants, the staff recommended that a new SRP11 Section be developed to codify existing guidance and industry recommendations.1394 In an RES evaluation,1564 it was concluded that consideration of a 20-year license renewal period did not affect the resolution.