Resolution of Generic Safety Issues: Issue 106: Piping and the Use of Highly Combustible Gases in Vital Areas (Rev. 2) ( NUREG-0933, Main Report with Supplements 1–34 )
Combustible gases such as H2, propane, and acetylene are used during normal operations of nuclear power plants in limited quantities and for relatively short periods of time. H2, the most prevalent of these gases in nuclear plants, is used as a coolant for electric generators in both BWRs and PWRs, for the control of reactor water chemistry and waste gas disposal in PWRs, and in the volume control tank (VCT) which is usually located in the auxiliary building of PWRs. It is stored as high pressure gas in vessels and is supplied to the various systems in the auxiliary building through standard piping, usually 3/4-inch in diameter. As a result, H2 piping is field-run and its location is plant-specific.
The concern in this issue is that leaks or breaks in the H2 piping and supply system could result in the accumulation of a combustible or explosive mixture of air and H2 within the auxiliary building. Inasmuch as the auxiliary building is a safety-related structure which houses most of the components of the safety-related systems of a plant, the accumulation of combustible or explosive mixtures of gas represents a threat to plant safety by virtue of the potential disablement of safety-related equipment, in the event that the combustible gases are inadvertently ignited. H2 detectors can signal the presence and accumulation of gas, but these are not qualified as safety-grade equipment and do not have an emergency power source. Thus, they are not regarded as sufficient protection against the development of H2 leakage and subsequent uncontrolled combustion or explosion. This issue primarily affected operating reactors licensed prior to the issuance of SRP11 Section 9.5-1, "Fire Protection," which addresses the safe use of combustible gases on site.
This issue was identified in NUREG-070544 and is related to Issue 136, "Storage and Use of Large Quantities of Cryogenic Combustibles on Site." Whereas Issue 106 addressed the normal process system use of relatively small amounts of combustible gases on site, Issue 136 addressed the considerably greater hazards of much greater amounts of combustible materials introduced by new needs at sites, i.e., solid waste processing and BWR H2 water chemistry control and the unique hazards associated with the transport and storage of large quantities of combustibles on site in a cryogenic liquid state.
The auxiliary building is a safety-related structure housing safety-related system components. Inasmuch as the most frequently used combustible gas (H2) is piped into this building for use in the VCT, there is the potential for leakage and the inadvertent ignition of the gas. The ensuing combustion or explosion can cause damage or failure of safety-related equipment, thereby contributing to a possibly significant increase in the core-melt probability of the plant.
In the event of a piping system break or large leak, large releases of combustible gas and the accumulation of combustible or explosive mixtures in air can be prevented by the installation of excess-flow check valves located close to the source of the combustible gas. SRP11 Section 9.5-1 recommends the use of excess-flow check valves. Other measures are needed to reduce the frequency of, or cause of, combustible gas accumulation accidents from such events as valve malfunctions or leaks, connection or fitting leaks, operations errors, material failures, etc. Plants licensed in accordance with the guidelines of SRP11 Section 9.5-1 were assumed to be not affected by this issue; however, the backfitting of excess-flow check valves at all plants not licensed in accordance with SRP11 9.5-1 plants was assumed.
Excess-flow check valves were an effective "fix" for piping system breaks, but other fixes, such as installation or upgrading of H2 detection systems, design changes, procedural changes, etc., will be required for other types of accidental releases. The risk and cost analyses performed for the installation of excess-flow check valves as a fix were extrapolated to develop a proper perspective.
It was assumed that, of all the combustible gases routinely used in a nuclear power plant, the most significant safety concern was associated with the use of H2. Unlike most other gases used in small quantities at nuclear power plants, H2 is used almost continuously while most other gases are used intermittently and, most likely, in the presence of trained personnel such as during welding operations. H2 leaks could continue unnoticed as a result of leaks or pipe breaks that go undetected for a sufficient time to accumulate a combustible mixture. It was assumed that H2 detectors were either not provided (as was the case in an event at Vogtle)1031 or were inoperative. In addition, it was assumed that operating plants licensed prior to the issuance of SRP11 Section 9.5-1 did not have excess-flow check valves in place. This latter assumption was a conservative element in this analysis because it was likely that some of the plants licensed prior to SRP11 Section 9.5-1 may have already had excess-flow check valves in place.
The auxiliary building is a safety-related structure that contains most of the components of the safety systems of a plant. However, the design of this structure and the location of safety-related components within the structure are plant-specific. In addition, location of the H2 source and, in particular, the H2 field-run piping layout are also plant-specific. In view of this, it was not possible to identify a particular damage scenario that represented a bounding sequence for the purposes of a generic analysis. Therefore, a reasonable but not necessarily bounding damage scenario was assumed in order to evaluate this issue. This scenario entailed the assumption of an H2 piping system leak or break, the accumulation of a combustible mixture within a room or space containing safety-related equipment, an ignition source, and damage contained within that room or space.
A PNL analysis64 of this issue based on a pipe break was extrapolated to estimate the frequency of all events that might result in the release and accumulation of combustible gases in the auxiliary building. It was assumed that the pipe break frequency (for ¾" pipe) may be obtained from WASH-1400,16 but that the probability of the accumulation of a combustible mixture, the probability of the availability of an ignition source, and the probability of total demolition of the safety-related redundant equipment are 1 in each instance. This latter assumption was conservative.
The scenario that was selected as a reasonable one for this analysis was the loss of both RHR heat exchangers (complete loss of heat sink). Resolution of this issue would affect operating plants using H2 and not already in compliance with SRP11 Section 9.5-1 with respect to H2 gas piping. Specifically, resolution was anticipated to include all operating PWRs. Therefore, the number of affected plants was 47 PWRs with an average remaining lifetime of about 27.7 years.
H2 piping is standard piping generally thought to be ¾" in diameter. Based on the results of WASH-140016 (Tables III 2-1, 2-2), the pipe break frequency for piping less than 3" in diameter was 10-9/hour per section. In general, it was assumed that the H2 piping in nuclear plants is comprised of about 25 sections. With 8,760 hours/year and an assumed plant utilization factor of 70%, the frequency of H2 release due to pipe break (fp) was estimated to be:
(10-9)(25 sections)(8760 hours/year)(0.7) = (1.5)(10-4) pipe breaks/RY
A review of 96 H2 accidents by NASA1030 indicated that about 52% could have been attributed to causes that relate to use of H2 in a gaseous state and about 48% to causes that relate to the use of H2 in its liquid (cryogenic) state; only about 2% was attributed to piping breaks. Therefore, it was assumed that an H2 accident from a gaseous state cause was (52% ÷ 2%) or 26 times more likely to occur than an H2 accident due to a pipe break. The probability (P) of an H2 release (leak) was given by:
|P||= (26)(Probability of Pipe Break)|
|= (26)(1.5 x 10-4)|
|= 3.9 x 10-3|
The probability of failing both RHR heat exchangers, f(RHR), is the product of the probability of an H2 leak, the probability of obtaining a combustible mixture, the probability of ignition, and the probability of being in the blast zone. Thus, f(RHR) = (3.9 x 10-3)(1)(1)(1) = 3.9 x 10-3.
When both trains of the RHR system are inoperable, the TS require plants to proceed to the hot shutdown condition within 12 hours; this requirement was modeled in this analysis as a T3 (PWR) transient. Therefore, an H2 explosion was modeled as an additional initiating transient (with a frequency of 3.9 x 10-3 /year as calculated above). All other initiating transients and LOCA parameters were scaled by (12/8,760 hours/RY) in order to model the occurrence of other random initiators during the 12 hours that the reactor is proceeding to hot shutdown. Finally, using the Oconee 3 PRA as representative of all PWRs, the RHR heat exchangers were modeled as inoperative by setting their representative system unavailabilities to 1. The Oconee PRA was then altered to incorporate the modified initiating event frequencies and the RHR systems unavailability in the affected minimal cut sets for each affected accident sequence. All affected Boolean equations were solved to calculate new core-melt frequencies for all containment failure modes and the affected core-melt frequencies were summed for each of the 7 distinct PWR core-melt categories. Public risk was then determined by summing the products of core-melt frequency and their respective release category dose factor for each release category.
The analysis was repeated for a period of 96 hours (4 days) as an approximation of the time necessary to achieve cold shutdown by alternate means such as feed-and-bleed. Based on these details, the following core-melt frequency results were calculated:
|PWR Base Case||= 5.46 x 10-6/RY|
|Adjusted Case||= 2.26 x 10-7/RY|
|Reduction in Core-melt Frequency||= 5.20 x 10-6/RY|
For the time required to come to hot shutdown (12 hours), the results of a PNL analysis64 indicated that resolution of this issue would result in a risk reduction of 8.8 man-rem/RY for PWRs. The total public risk reduction was estimated to be approximately 11,500 man-rem. The estimated occupational risk reduction due to accident avoidance was approximately 135 man-rem.
Industry Cost: The cost of installing excess-flow check valves in the H2 lines outside of the safety-related areas was estimated.64 It was assumed that these valves would be installed during scheduled reactor shutdown periods so that there would be no additional power replacement cost incurred. Based on two vendor quotations, the average cost of one excess-flow valve was approximately $870. The costs for the implementation, maintenance, and operation of the excess-flow check valve "fix" was detailed as follows:
|Hardware Design and Review||= 2 days|
|Procurement||= 1 day|
|Pre-Installation Check||= 0.5 hour/valve|
|Installation||= 2 days/valve [1 man-day/valve (welder),1 man-day/valve (fitter)]|
|Post-Installation Check||= 1.5 hours/valve|
|Documentation||= 0.5 day|
|Total Labor Time (PWR)||= 3.5 days + 2.25 days/valve|
|= 8 days|
|Labor Cost (PWR)||= (8 days)($2,270/man-week)/(5 days/man-week)|
|Valve Cost (PWR)||= (2)($870)|
The total implementation cost/plant was $(3,632 + 1,740) or $5,372; for all 47 affected plants, this cost was approximately $255,000.
Operation and maintenance would include a semi-annual check of the installation to ensure that the valve shaft was not "frozen" and replacement of the valve diaphragm as needed; the frequency of this replacement would depend upon the valve environment. For this analysis, the diaphragm was assumed to require replacement every 7 years with an associated labor requirement of 0.5 man-day.
(a) Labor for Maintenance and Operation
|Semi-annual Check||= 2 hours/valves|
|Diaphragm replacement required (average plant life/7-year replacements) over the remaining lifetime at 4 hours per replacement.|
|Labor (PWR)||= (2 hours/valve)(2 valves)(2 checks/year)|
|+ (27.7 years/7 years)(4 hours/valve)(2 valves)/(27.7 years)|
|= 8 hours/year(checks) + 1.14 hours/year (average diaphragm replacement)|
|= 9.14 hours/RY|
|Labor Cost||= (9.14 hours/RY)($2,270/man-week)/(40 hours/man-week)|
Thus, the operation and maintenance cost was ($519/RY)(47 plants)(27.7 years) or $675,686 and the total industry cost for the resolution of this issue was approximately $(255,000 + 675,000) or $930,000.
NRC Cost: Development of the implementation of the solution, including the formulation of guidelines and documentation requirements, and review and inspection of the final installation were estimated to cost (4 man-weeks) ($2,270/man-week) or $9,080. Implementation costs were estimated to be (0.6 man-week/plant)($2,270/man-week) or $1,362/plant; for the 47 affected plants, this cost was $64,000. Review and inspection of plant operation and maintenance activities were estimated to be (0.5 day/plant-test)(2 tests/year) or 1 day/RY; at a daily rate of [($2,270/man-week)/(5 days/man-week)], this cost was $454/RY. For 47 plants with an average remaining life of 27.7 years, the operation and maintenance cost was estimated to be (47 reactors)(27.7 years)($454/RY) or $591,000. Thus, the total NRC cost was estimated to be $(9,080 + 64,000 + 591,000) or $664,000.
Total Cost: The total industry and NRC cost associated with the possible solution was estimated to be $(930,000 + 656,000) or approximately $1.5M. Installation of excess-flow check valves was considered a satisfactory "fix" for the possibility of sudden accumulation of combustible or explosive mixtures of H2 resulting from a piping system break, but it was not a solution for H2 accidents arising from slow leaks in valves or fittings, purging errors, material degradation problems, contamination, etc. Other "fixes" are required to reduce or preclude H2 accidents from scenarios other than pipe breaks. These other "fixes" would include the installation or upgrading of existing H2 detection and alarm systems, completion of combustible gas system design reviews and modifications to plant design and hardware, operating procedure reviews and modifications, improved preventive maintenance programs, and major modifications to the auxiliary building ventilation system. For an assumed population of 47 plants, the total industry cost of these other "fixes" would be very much more than the costs estimated for the very restrictive "fix" (excess-flow check valves) estimated by PNL. Therefore, the cost of the other "fixes" was assumed to be an order of magnitude greater than that calculated for the installation of excess-flow check valves. Thus, the total estimate for complete resolution of the issue was at least $15M.
Based on a potential public risk reduction of 11,500 man-rem a cost of $15M for a possible solution, the priority score was given by:
Based on the potential reduction in core-melt frequency of 5.2 x 10-6/RY for PWRs, the cost savings from accident avoidance was ($1.65 billion)(5.2 x 10-6/RY) (47 reactors)(27.7 years) or approximately $11.4M.64
For the potential reduction in both public risk and core-melt frequency, a medium priority was appropriate, regardless of potential cost, unless the value/impact score was greater than 3,000 man-rem/$M, in which case, a high priority would be appropriate. For the value/impact score to be $3,000 man-rem/$M, the total industry resolution cost could not exceed about $3.8M. Based on the relatively detailed estimate of the cost of excess-flow check valve installation (~$1.6M), which was only a small portion of the total "fix," and extrapolation, it appeared very unlikely that this issue could be completely resolved for less than $4M. Therefore, the issue was given a medium priority (See Appendix C).
The staff's technical findings and regulatory analysis were reported in NUREG/CR-57591544 and NUREG-1364,1545 respectively. The issue was RESOLVED with no new requirements1546 and Generic Letter 93-061547 was issued to inform OLs and CPs. However, RES recommended that the SRP11 be revised to: (1) include the information contained in Generic Letter 93-061547; and (2) reference EPRI NP-5283-SR-A,1548 "Guidelines for Permanent BWR Hydrogen Water Chemistry Installations - 1987 Revision"; the latter recommendation was consistent with the RES recommendation made with the resolution of Issue 136. Consideration of a 20-year license renewal period would not affect the resolution of Issue 106.