Resolution of Generic Safety Issues: Issue 70: PORV and Block Valve Reliability (Rev. 3) ( NUREG-0933, Main Report with Supplements 1–35 )
In order to export or print, this web site requires you to enter text displayed in the distorted image on the left, which is known as Captcha, into the text field below the image.
DESCRIPTION
Historical Background
PORVs and block valves were originally designed as non-safety components in the reactor pressure control system for use only when plants are in operation. The block valves were installed because of expected leakage from the PORVs. Neither the PORVs nor the block valves were required to safely shut down a plant or mitigate the consequences of accidents. RSB determined in 1983 that PORVs are relied upon to mitigate a design basis SGTR. However, the acceptability of relying on non-safety grade PORVs to mitigate a design basis accident (e.g., an SGTR) was questioned by RSB.575 This issue was identified in a DL memorandum515 and addressed the assessment of the need for improving the reliability of PORVs and block valves.
In most plants, the low temperature overpressure protection (Ltop) system is designed to use PORVs. For this mode of operation, the valves are typically set to open at 500 psig rather than the high pressure (~2300 psig) setpoint used at power. Westinghouse and some CE-designed plants use redundant PORVs for Ltop concerns. These plants are brought to a water solid condition during shutdown. In contrast, B&W owners use a single PORV and the gas (steam or nitrogen) space in the pressurizer functions as the primary Ltop system. The PORV and associated actuation circuitry function as a backup, should an operator fail to terminate a low temperature overpressure challenge before compression of the gas space. Ltop systems, as specified in SRP11 Section 5.2.2, are to be single failure proof, testable, designed to quality standards, and operable from emergency power. Full implementation of IEEE Standard 279 to withstand an SSE is not specified, but the OBE is. At the time this issue was evaluated, Ltop system requirements were being implemented under MPA B-04.578
NUREG-0737,98 Item II.D.1, set forth functional requirements for both PORVs and block valves. All plants were required to demonstrate the functionability of these valves for all expected flow conditions during operating and accident conditions. It was further required that the block valves be capable of closing to ensure that a stuck-open relief valve can be isolated, thereby terminating a small LOCA. In response to the Item II.D.1 requirements, PORVs were tested extensively by EPRI583 and the results reported to the staff. Only limited block valve testing had been performed in the EPRI program at the time this issue was evaluated. Reports describing the test program results were submitted to the staff for review. Most plants requested exemptions to the specified completion date for Item II.D.1 in order to obtain additional time for the required evaluation of piping associated with safety valves, PORVs, and block valves.
When PORVs are used for high point vents in some plants under Item II.B.1 of NUREG-0737,98 both PORVs and block valves are required to meet seismic and environmental requirements for safety-related equipment. There were no TS requirements that these components be operational when the plants were at power. Continued operation at power with inoperable PORVs and block valves is permitted by the TS if the block valves are closed and power to the block valves is removed. At the time this issue was evaluated, many of the plants operated with the PORVs blocked.
In the safety evaluations572,573,574 for Item II.K.3(2) it was determined that an automatic PORV isolation system was not necessary. However, the possible need to improve PORV reliability was recognized. The Item II.K.3(2) conclusion was predicated on the absence of the need to reduce the PORV-SBLOCA frequency by this single modification. Therefore, the scope of the evaluation of Item II.K.3(2) was limited in that it did not consider, or combine, the automatic PORV isolation system as a subset with other measures that could be taken to improve overall PORV/block-valve system reliability. The analysis provided herein also indicated that a singular modification, like that reviewed in Item II.K.3(2), or a reliability program to improve PORV/block-valve system reliability, without considering improvements to the control element (automation vs. manual), would yield small benefits. The broader program reviewed herein included an automated PORV isolation system as a possible part of an overall PORV/block-valve system reliability improvement. Therefore, this analysis expanded the evaluation of Item II.K.3(2) by including additional means to improve the PORV/block-valve system reliability and assessing all of the modes of risk reduction (and costs) that would likely result from these improvements.
Safety Significance
PORVs and block valves are used in various modes of plant operation. On a number of occasions, these valves were stuck open when required to operate. Such malfunctions have led to significant plant transients and aggravated others. Most notable are the events at TMI-2 in 1979 and Crystal River 3 in 1980. The failure of a PORV block valve to completely close aggravated recovery from a reactor coolant system leak at H. B. Robinson in 1981. In January 1982, a malfunction of the PORV aggravated recovery from an SGTR event at Ginna.581 Some accident sequences and transients can be mitigated by using the PORVs and HPI pumps for pressure and coolant inventory control. This mode of operation is known as feed-and-bleed, or bleed-and-feed, depending on the HPI capability of the injection pumps and system design. In these situations, the PORVs could experience multiple openings and closures. Life cycle testing, which is not now required, could provide better assurance that the PORVs can withstand this type of operation.
In other cases, when the PORVs have leakage problems and are blocked, this could cause the safety valves to be challenged and, if they stick open, could result in an unisolable SBLOCA. Also, during startup and shutdown operations, many of the plants use the PORVs as part of the Ltop system. The Ltop systems that rely on lifting the PORVs to reduce reactor pressure, in accordance with established pressure/temperature limits, have experienced several events where the Ltop systems have been inoperable.577,591,592 As a result of the inoperable Ltop systems, the potential for brittle fracture of the reactor pressure vessel is increased. Even though the Ltop system discussed above is a separate issue, the common cause failures are inoperable PORVs due to PORV leakage; maintenance errors disabling the Ltop systems; procedural deficiencies; and inadequate inspection or surveillance of the gas (nitrogen/air) supply that provides the PORV opening force in certain plants.
Possible Solution
Resolution of this issue could involve specification of the PORV/block-valve combination to some or all of the requirements associated with safety grade systems, better initial qualifications for the valves, and specified maintenance and testing requirements. The need for an automatic actuating circuit for the motor-operated block valve with manual override and reset capabilities was evaluated and determined unnecessary as a singular requirement.572,573,574 Any program initiated to improve PORV/block-valve system reliability should take into consideration the Item II.K.3(2) conclusions and the basis and criteria for the conclusions. This analysis includes implementation of an automatic PORV isolation system to gauge the potential benefits this modification could provide if combined with other reliability improvement measures. Broader and more specific ideas relating to possible resolutions are discussed below.
PRIORITY DETERMINATION
To establish the priority of this issue, the potential reductions in the SBLOCA frequency that could result from certain improvements in the PORV/block-valve system were quantified. The change in the SBLOCA frequency and resulting risk reduction were assumed to estimate the baseline benefits for this issue. The effects of transients, other systems (e.g., Ltop), functions (e.g., feed-and-bleed), ATWS events, and safety grade PORV/block-valves on the overall PORV/block-valve reliability are additional unquantified benefits that were considered in arriving at the priority ranking.
Frequency Estimate
PORV Challenge Frequency: The PORV challenge frequencies are based on operating data572,574 for W and CE plants for the post-TMI period from April 1, 1980, to March 31, 1983. These data included PORV challenges for seven CE plants and 28 W plants. Examination of these data showed 18 of the plants (51%) with no PORV challenges (lifts). Since approximately 55% of the plants operate with closed block valves,572,574,586 the 18 plants with no PORV challenges were assumed to be within the null set (55%) of plants that operate with closed block valves. PORV challenges at low power, as identified in the data, were also eliminated. Graph plots of the remaining 37 PORV lifts for the 17 plants in the active set indicated that approximately 50% of the plants account for 69% of the PORV lifts over the 3-year period covered. Therefore, the PORV lift frequency for 50% of the plants in the active set is 1/RY. The data573 related to B&W plants were not used because they were considerably lower than that shown for CE and W plants. Therefore, this analysis is applicable only to W and CE plants.
PORV/Block-Valve Failure Frequency: The PORV-SBLOCA (given that the block valves are open) requires actuation (challenge) of the PORV, failure of the PORV to close, and failure to close the block valves. The failure frequency of the PORV to close, given that it has opened, is 0.02/demand.366 The failure frequency of the block valves to function is estimated at 0.005/demand.346 Potential reduction in the failure probability of the PORV to close, given that it has opened, is highly judgmental. Since the "weak spot" in the PORV appears to be the control systems576 and, for valves in general, malfunctions of valve operators and control equipment have occurred about twice as often as malfunctions of the valves proper,590 a PORV failure probability of 0.01 was assumed to be obtainable. The improvement was assumed to result from improved maintenance, testing, surveillance, and proper matching of the valve operators with the valve body as a valve assembly. A similar reduction to 0.003 in the block valve failure to close probability was also assumed obtainable.
Operator Error Frequency: Failure of an operator to close the block valve increases the chance for a SBLOCA through the PORV flow path, given that the PORV sticks open. Based on operating experience,76 the pre-TMI operator error rate was approximately 0.29 which is in agreement with the WASH-140016 (Table III, 6-1) estimate of 0.2 to 0.3 for plant operators under very high stress levels where dangerous activities are occurring rapidly. Since the TMI-2 event, valve position indicators and more emphasis on operator training should reduce the chance of operator error. From the analyses performed,572,574 an operator error rate (HEP) of 0.05 was assumed based on an 83% improvement in operator performance (reliability) as a result of TMI-2 improvements and the increased emphasis on operator training.
Automatic Actuation Block-Valve Failure Frequency: Because the plant operator is "in fact" a control element that affects overall reliability for the PORV/block-valve system, automatic actuation of the block valves to replace the manual operator action was considered in this analysis. The failure rate for an automated block valve (MOV) was assumed to be 0.002/demand based on the assumption of one failure in 486 tests366 even though no failures were observed in these tests. Therefore, the 0.002/demand failure rate for automatic actuation should be conservative.
PORV-SBLOCA Frequency: For the purpose of this analysis, the PORVs and safety valves were assumed to be normally closed and the block valves were assumed to be normally open.
The base case PORV/block-valve SBLOCA frequency of 1.1 x 10-3/RY with the operator controlling closure of the block valves is the product of the PORV challenge frequency (1.0), the probability that PORV sticks open (0.02), the probability that the operator will not close the block valve or that the block valve malfunctions by failing to close (0.05 + 0.005).
Improved valve maintenance, testing, surveillance, qualification and specification, and automated block valves were considered as combined improvements. They are estimated to reduce the PORV/block-valve SBLOCA frequency to 5 x 10-5/RY. The reduced SBLOCA is the product of the PORV challenge frequency (1.0), the probability that the PORV sticks open (0.01), and the probability that the automatic actuation circuitry of the block valve fails or that the block valve malfunctions by failing to close (0.002 + 0.003).
The potential 95% reduction in the PORV/block-valve SBLOCA is therefore estimated to be 1.05 x 10-3/RY. It seems likely that improvements in PORV/block-valve reliability could also reduce the PORV actuations that have resulted from low power lifts, spurious signals, maintenance, and testing that are not specifically reduced in the above analysis.
For comparison, the above generic PORV/SBLOCA frequency of 1.1 x 10-3/RY is approximately one-third the PORV-SBLOCA frequency (3.1 x 10-3/RY) calculations572,574 for the CE and W plants. Also, it is estimated588,589 that a PORV-SBLOCA has a frequency of 5.1 x 10-3/RY. Based on the above, the PORV-SBLOCA might range from 1 x 10-3 to 5 x 10-3/RY. The above values are all within the uncertainties inherent in this generic analysis.
Transient Frequencies: The PORV challenges (lifts) in the database resulted from plant operating experiences that included transients, spurious actuations, and one SGTR event. Therefore, high pressure transients that challenge the PORVs most frequently, and the potential for a SBLOCA from such events, are included in the SBLOCA section of this analysis. Less frequent high pressure transients that might occur, and that are not inclusive in the 3-year database, are expected to provide secondary effects that are in addition to the SBLOCA base-line effects. These secondary effects should not significantly alter the results, but were factors considered in the conclusion of this analysis.
Ltop Events: Ltop events and low power actuations of the PORVs were eliminated from the calculated PORV lift frequency of 1.0/RY used to estimate the SBLOCA frequency. The four such events that were evident from the 3-year database could have been attributed in part to inadequate PORV/block-valve system reliability programs. Ltop events that result from failure to place the block valves in the open position to allow pressure relief through the PORVs, as stated earlier, is a separate issue (MPA B-04).578 However, resolution of the Ltop concerns (event frequencies) would likely be a subset of the overall measures to improve PORV/block-valve reliability addressed in this issue. In recognition of possible double counting of the benefits attributed to resolution of the Ltop issue, credit was not taken for the potential benefits that may be more appropriately obtained within the more definitive resolution of the Ltop issue. Nevertheless, coordination of the resolutions of the various operations for which the PORV/block-valve systems are used was considered an important element that was factored into the evaluation.
Feed-and-Bleed: Feed-and-bleed operation was evaluated in Issue A-45. Any need to require improved PORV/block-valve reliability related to specific needs in the feed-and-bleed mode of operation were developed as part of the resolution of Issue A-45.
ATWS Frequency: Reliability improvement benefits to the existing PORV/block-valve systems with respect to ATWS events were evaluated based on three types of pressure-relieving scenarios: (1) success, (2) partial success, and (3) failure. Success is the probability that the block valve, PORV, and SVs open (function) to relieve pressure build-up in the reactor. Success, however, does not indicate that sufficient relief capacity is available through the existing valves. Partial success is that either the PORV or SVs open to relieve pressure build-up in the reactor. As in the success case, partial success does not mean that the existing valves have sufficient relief capacity to mitigate high reactor pressures that might result in exceeding a 3200 psi stress level. Failure is the probability that no relief is obtained through either the PORV/block-valve system or the SVs.
ATWS Success: The probability of ATWS success (Case 1) before PORV/block-valve reliability improvements is 0.441A and is the product of the ATWS frequency (A~10-5/RY), the probability that the plant is operating with the block valves open (0.45), the probability that the PORVs open (0.99), and the probability that the SVs open (0.99). Assuming that PORV reliability can be improved by 50%, the probability of opening the PORVs is increased to (0.995). The resulting increase in an ATWS success (as defined above) is (2 x 10-3)(A)/RY. Assuming that the ATWS event leads to core-melt, the potential reduction in core-melt frequency is 2 x 10-8/RY.
Partial ATWS Success: Partial ATWS success (Case 2) involves the sum of three event sequences, given that an ATWS occurs (A~10-5/RY):
(1) Case 2a
Block valves are open (0.45) PORVs open (0.99) SVs fail to open (0.01)
The probability for Case 2a before PORV/block-valve reliability improvements is (4.45 x 10-3)(A). Assuming a 50% improvement in PORV reliability (PORV opens = 0.995), the resulting increase in partial ATWS success for Case 2a is (3 x 10-5)(A)/RY. Assuming the ATWS leads to core-melt, the potential reduction in core-melt frequency for Case 2a is 3 x 10-10/RY.
(2) Case 2b
Block valves are open (0.45) PORVs fail to open (0.01) SVs open (0.99)
The probability for Case 2b before PORV/block-valve reliability improvements is (4.46 x 10-3)(A). Assuming a 50% improvement in PORV reliability (PORV fails to open = 0.005), the resulting increase in partial ATWS success for Case 2b is (2.23 x 10-3)(A). Assuming the ATWS leads to core-melt, the potential reduction in core-melt frequency for Case 2b is 2.2 x 10-6/RY.
(3) Case 2c
Block valves are closed (0.55) SVs open (0.99)
For Case 2c, plants operating with closed block valves, it was assumed that the operator does not or cannot open the block valves in time to relieve the primary system pressure. In this case, reliability improvements to the PORV/block-valve system will not affect the partial success resulting from the opening of the SVs.
Failure From ATWS: Failure to relieve system pressure from the PORVs and SVs following an ATWS event can occur whether the plants operate with the block valves open (Case 3a) or with the block valves closed (Case 3b).
(1) Case 3a
Block valves open (0.45) PORVs fail to open (0.01) SVs fail to open (0.01)
The probability for Case 3a before PORV/block-valve reliability improvements is (4.5 x 10-4)(A). Assuming a 50% improvement in PORV reliability (PORV fails to open 0.005), the resulting decrease in failure for Case 3a is (2.3 x 10-4)(A). Assuming the ATWS leads to core-melt, the potential reduction in core-melt frequency for Case 3a is 2.3 x 10-9/RY.
(2) Case 3b
Block valves closed (0.55) SVs fail to open (0.01)
For Case 3b, with the plants operating with closed block valves, reliability improvements to the PORV/block-valve system have no affect.
Combined ATWS Frequency: The combined reduction in core-melt frequency for the above ATWS cases was determined to be approximately 4.4 x 10-8/RY. Based on this estimate, improvements that increase the reliability of the PORV/block-valve systems by as much as 50% will not provide significant benefits toward reducing core-melt or consequences that would result from an ATWS event. This estimate addressed only the effects of improved PORV/block-valve reliability to provide some relief in system pressure following an ATWS. The analysis did not address the adequacy of existing valves to provide sufficient relief capacity to mitigate an ATWS.
Safety Grade PORV/Block Valves: A design basis event (an SGTR) occurred at the Ginna plant within the 3-year period that was used to determine the PORV lift frequency (1/RY). Other than the stated RSB position that the PORV/block-valve system should be safety grade, no evidence had been revealed that indicated that safety grade components were more reliable than control grade components (PORV/block-valve systems). In the absence of such evidence, no benefit (risk reduction) could be quantified for the proposed method of improving PORV/block-valve system reliability. It was recommended that any subsequent staff information developed in the resolution of Issue 70 be coordinated with, and factored into, the need for safety grade PORV/block-valve systems or individual components of the PORV/block-valve systems.
Core-Melt Frequency: The WASH-140016 median core-melt frequency (1.5 x 10-5/RY) for a SBLOCA (S2=10-3/RY) is dominated by failures in the emergency core cooling injection system (S2D) and the emergency core cooling recirculation system (S2H), resulting in a Category 7 type release (basemat melt-through). A more representative (although conservatively biased) value of the probability of failure of the HPI(D) systems was considered587 to be 3 x 10-3/demand, as opposed to the WASH-140016 value of 9 x 10-3/demand. Therefore, the WASH-140016 S2D sequence frequency of 9 x 10-6/RY was replaced by the value of 3 x 10-6/RY. In addition, RRAB proposed that, for the WASH-140016 S2H sequence frequency of 6 x 10-6/RY also involving a Category 7 type release, the operator may not need to go to recirculation, or the time available for corrective action (close the block valve) will most likely be hours instead of minutes. Therefore, the HEP=0.05 in the PORV-SBLOCA (discussed above) should be reduced to HEP~0.005 for the S2H (WASH-1400)16 sequence. The adjusted (reduced) S2H sequence frequency is therefore 6 x 10-7/RY. The result is a core-melt frequency of 4 x 10-6/RY that is dominated by the Category 7 type release. Ratioing the potential reduction in PORV-SBLOCA frequency (1.05 x 10-3/RY) to the WASH-140016 SBLOCA frequency (S2=10-3/RY) yields an estimated potential reduction in core-melt frequency of (1.05)(4 x 10-6/RY) = 4.2 x 10-6/RY.
Consequence Estimate
The risk reduction attributed to improvements in the PORV/block-valve reliability is shown below in Table 3.70-1. Column 1 lists the WASH-140016 dominant release categories; Column 2 lists the modified WASH-140016 S2 (small-break) core-melt frequency for the specific release category. The modifications delete the S2C sequences and containment failure modes from steam explosions, in accordance with RRAB recommendations. The RRAB position that an S2 (SBLOCA) with a failure of the HPI(D) system may not lead to core-melt, if aggressive cooldown via the steam generators (secondary-side) is used, could possibly be considered on a plant-specific basis. RSB evaluated aggressive cooldown as a means to mitigate very small LOCA scenarios and to improve methods for decay heat removal under Issue A-45. However, recovery credit for aggressive cooldown (elimination of the S2D core-melt sequence) was not considered generically representative because improved decay heat removal methods had not been established at the time of this evaluation.
| 1 | 2* | 3 | 4 |
|---|---|---|---|
| Table 3.70-1 | |||
| RELEASE CATEGORY | MODIFIED WASH-140016 (S2~10-3/RY) CORE-MELT FREQUENCY (RY-1) | DOSE (man-rem) | RISK (2)(3) (man-rem/RY) |
| 2 | 2 x 10-9 | 4.8 x 106 | ~ |
| 3 | 2 x 10-7 | 5.4 x 106 | 1.1 |
| 4 | ~ | 2.7 x 106 | ~ |
| 5 | 1 x 10-8 | 1.0 x 106 | ~ |
| 6 | 1 x 10-8 | 1.4 x 105 | ~ |
| 7 | 4 x 10-6 | 2.3 x 103 | ~ |
* Not including the 10% from adjacent categories used to smooth the data.
Ratioing a reduction in the PORV-SBLOCA of 1.05 x 10-3/RY to the WASH-140016 SBLOCA (S2=10-3) and remaining plant life of 27 years, the potential public risk reduction from improving the reliability of the PORV/block-valve systems is (1.05)(1.1)(27) man-rem/reactor = 31 man-rem/reactor.
The analysis provided herein provides the baseline potential risk reduction. Consideration of Ltop, feed-and-bleed operations, ATWS events, and safety grade PORV/block-valve systems are not expected to significantly alter the potential risk reduction of this issue, but these unquantified secondary effects are additional qualitative factors considered below.
Cost Estimate
Industry Implementation Cost (Operating PWRS): PNL estimated64 that valve backfit labor costs are $27,200/plant based on 12 man-wk/plant and $2,270/man-wk. This includes management review, QA control, licensing review, and engineering for the backfit. Material requirements are two safety-grade PORVs and two instrumented (for automatic actuation) block valves, each costing $25,000. Incremental material costs such as piping, supports, hardware, etc., beyond those associated with initial installation of the safety-grade PORVs and instrumented block valves at a plant are estimated at $50,000. The cost for the safety analysis is estimated at $50,000/plant. A Class III License Amendment for the valve upgrade is placed at $4,000. The implementation cost is therefore estimated to be $237,200/plant.
Industry Maintenance Cost (All PWRs): Additional annual maintenance and testing is estimated at (0.5 man-wk/RY)($2,270/man-wk) = $1,140/RY. The present worth of this cost with a 4% discount rate585 over 27 years is $17,860/plant.
Industy Implementation Cost (Planned PWRs): For new plants, an incremental effort above the analysis required for relief valves is estimated at $5,000/plant. Assuming that new valves at $25,000/each will be required, the forward-fit implementation costs are $105,000/plant.
NRC Cost (Operating PWRs): NRC costs will most likely involve plant-specific reviews, generic studies necessary to establish reliability and performance goals for the PORV and block valves, and preparation of a Regulatory Guide. The generic studies and preparation of a Regulatory Guide were estimated to require 3 man-years of effort ($300,000). This cost was assumed distributed over 47 operating plants and 48 planned plants for an NRC cost of $3,100/plant. The plant-specific reviews (which include review, SER preparation, and technical specification changes) were estimated to require 1.5 man-months ($12,500). Thus, the total NRC cost for plants affected by a backfit was estimated to be $15,600/plant.
NRC Cost (Planned PWRs): The NRC reviews were assumed to be part of the normal licensing process. However, as stated above, the NRC costs associated with the generic studies and development of a Regulatory Guide ($3,100) were distributed over both operating and planned reactors.
Value/Impact Assessment
Based on a public risk reduction of 31 man-rem/reactor and a cost of $0.27M/reactor for operating plants only, the value/impact score is given by
Other Considerations
(1) Occupational Risk Change
(a) Implementation ORE: PNL estimated that replacement of the existing PORV/block-valve system with a safety grade (or equivalent) system would require 96 man-hours/plant. The radiation field in the region of the pressurizer was estimated at 0.2 R/hr (EPRI NP-1139, page 3-26).584 The implementation dose was therefore estimated to be 19.2 man-rem/plant.
(b) Maintenance ORE: The maintenance ORE for additional annual testing was estimated by PNL at 4 man-hrs/plant in the 0.2 R/hr field. Over a 27 year period, this results in an ORE of 21.6 man-rem/plant.
Based on information in EPRI NP-1138,431 the PORV/block-valve maintenance on two PWRs over 6 years required approximately 50 man-hrs/RY. In a 0.2 R/hr radiation field, this amounts to approximately 10 man-rem/RY. The EPRI report (page 4-103)431 also concluded that plants with the least maintenance were those that contracted outside specialty vendors and/or manufacturers to perform maintenance and adjustments. This indicated a general need for additional training in maintenance procedures for those plants that perform their own maintenance. The extent of improvement (maintenance reduction) that could be attributed to improved maintenance procedures was difficult to judge. However, assuming only a 10% improvement, the potential ORE reduction is 1 man-rem. Over a remaining plant life of 27 years, this amounts to 27 man-rem/RY. Therefore, it is estimated that the ORE resulting from the additional annual testing described above can be offset by the ORE reduction resulting from improved maintenance methods.
(c) Outage Avoidance ORE: An estimate of the potential ORE reduction that can be attributed to improved reliability of the PORV/block-valve system and from potential outage avoidance cannot be quantified due to incompleteness in the information reviewed. However, an example of such an event involved a rupture of the rupture disc in a pressurizer relief tank that resulted from improper seating of a PORV (EPRI-1139).584 The tank repair time for this event required approximately 16 hours, of which, 3 hours were classified as critical path time (lost power production). The radiation field and ORE were not given. Such events were not believed to be frequent but have occurred at other plants. Improved PORV/block-valve reliability may reduce similar occurrences and thereby reduce the ORE resulting from such repairs.
(d) Accident Avoidance ORE: The reduction in core-melt frequency of 4.2 x 10-6/RY results in avoidance of ORE associated with core-melt cleanup operations (20,000 man-rem/core-melt).64 The accident avoidance dose over a remaining plant life of 27 years is [(27) (4.2 x 10-6)(2 x 104)] = 2 man-rem/plant.
The combined implementation and maintenance ORE expected to result from upgrading the PORV/block-valve system is 20 man-rem/plant. However, this ORE may be offset by less frequent outages and repairs that result from improved PORV/block-valve system reliability. For purposes of comparison, the median annual collective dose (ORE) for PWRs was approximately 400 man-rem/RY.482
(2) Outage Avoidance Cost: Most of the repairs made to PWR RCS relief valves do not require reactor shutdown. However, as reported in EPRI NP-2092,114 an average of 115 effective full power hours (EFPHs) per outage with an event frequency of 0.11/RY was mainly attributed to the PORVs and block valves. It was assumed that improved reliability of the PORV/block-valve system can also reduce the outage frequency. If it is further assumed that the outage frequency reduction is proportion to only 30% of the potential reduction in the PORV/block-valve SBLOCA frequency [(0.3) (1.05/1.1) = 0.29], the reduction in outage frequency is
where EFPD = effective full power day.
Based on a replacement power cost of $0.3M/day,64 the potential reduction in outage frequency results in a replacement power cost savings of $45,000/RY. Assuming a 4% discount rate585 over 27 years yields a present worth cost savings of $0.76M/reactor. If higher replacement power cost is used, the cost savings would be proportionately greater.
(3) Accident Avoidance Cost (On-site)
The present worth cost of a core-melt accident is estimated64 at $1.65 billion considering cleanup and replacement power cost over a 10-year period. The present worth of accident avoidance at each plant is [(4.2 x 10-6/RY)($1,650M)(27 RY)] = $0.2M.
(4) As a result of some of the TMI Action Plan items, various requirements and options have been installed in operating reactors. The options (changes) were directed primarily at reducing the PORV/block-valve challenges. The extent of each plant's mitigating features is beyond the scope of this evaluation. However, to adequately assess which method, or methods, would provide optimum improvements to the reliability of the PORV/block-valve systems, the plant-specific status of related TMI Action Plan items should be determined. As previously discussed, there were a number of open issues that were related to the resolution of this issue at the time of this evaluation. It was believed that these open issues could have had an effect on the safety mission of the PORV/block-valve assembly and thus affect the appropriate reliability/qualification objectives.
At the time of this evaluation, the following items were pending: (1) Item II.D.1 was still ongoing with the outcome uncertain; (2) a feed-and-bleed mission with or without seismic qualification was not yet determined (Issue A-45); (3) the ATWS rule was not yet complete; and (4) the severe accident research program could conclude that it is desirable to depressurize the reactor vessel during pressurized core coolant boil-off (station blackout, etc.) in order to enable the accumulators to dump and thus buy time to restore HPI or to avoid the effects of pressurized vessel melt-through such as direct heating, missiles, pressure spike, subsequent accumulator dump, etc. Also, the need for safety grade PORV/block-valves beyond the safety features that some plants already had was not yet determined. If replacement valves that meet all the requirements of a safety grade PORV/block-valve system were not needed, the value/impact (benefit/cost) ratio for this issue could be significantly improved with a lesser pedigree of safety grade PORV/block-valve systems.
CONCLUSION
In general, the PORV/block-valve systems were believed to be primarily for operational flexibility in pressure control and not required to safely shut down a reactor. However, these valves are sometimes used to mitigate certain design basis accidents (e.g., SGTR), transients, Ltop events, reduce safety valve challenges, and potentially to help mitigate the affects of an ATWS. SBLOCAs through this system and resulting challenges to safety systems appeared to be of sufficient frequency that, based on the above evaluation, improved reliability of the PORV/block-valve system might yield a potential public risk reduction of 31 man-rem/reactor at a cost of $0.27M/reactor. The resultant value/impact score of 115 man-rem/$M and the potential reduction in core-melt frequency of approximately 4 x 10-6/RY indicated that a medium priority ranking was appropriate for this issue. Replacement of the existing PORV/block valves that meet all the requirements of a safety grade PORV/block-valve system may not be needed and reliability could be improved with less costly modifications. Therefore, subject to the above considerations, the value/impact score for this issue could be higher but could not exceed the medium ranking.
The outage avoidance cost savings of $0.76M/reactor and the accident avoidance cost savings of $0.2M/plant were not included in the value/impact assessment. Potential cost savings through outage and accident avoidance were estimated at approximately $1M/reactor. These potential cost savings were nearly four times greater than the estimated implementation costs and could have provided an additional industry incentive for resolution of this issue. Likewise, potential increases in the ORE resulting from expanded and improved maintenance testing and surveillance procedures could be offset by ORE reductions brought about by improved PORV/block-valve system reliability.
Conversely, development of aggressive cooldown capabilities through the resolution of Issue A-45 could have reduced the potential for core-melt from a PORV-SBLOCA. This capability would have reduced the value/impact ratio.
For plants that have not yet commenced operation, the value/impact ratio could be considerably greater. This is because forward-fit costs would be less than the backfit cost. The magnitude of the value/impact ratio was dependent on the existing licensing/construction status and the extent to which upgraded PORV/block-valve systems already existed or were already planned in the plant designs. The CE plants that did not have PORVs were outside the scope of this issue and were addressed separately in Issue 84, "CE PORVs."
Because of large uncertainties inherent in this limited assessment, this issue could have been assigned a medium priority ranking on a base-line risk reduction and other qualitative considerations, even though the estimated base-line risk reduction was not significant. Further and more careful analysis could show greater potential risk reduction, although this would not be expected to occur for all, or even most, medium priority issues. In this case, analysis previously performed under TMI Action Plan Item II.K.3(2) concluded that automating the closure of the block valve would not reduce SBLOCA frequency significantly. (See References 572, 573, and 574.) Thus, the II.K.3(2) analysis was considered completed and the results of this analysis were comparable in that both estimated similar values of SBLOCA frequency. However, this issue identified the possible need for both a broader and more specific resolution.
It was believed that resolution of this issue might entail imposition of some or all of the attributes of safety grade qualification such as (a) redundancy for selected design basis challenges; (b) N-stamp; (c) seismic Category I qualification; (d) environmental qualification; (e) TS on operability and/or the normal alignment of the block valves; and (f) QA pedigree.
Resolution might also entail deterministic or probabilistic reliability qualification for one or a variety of missions for the PORV and block valve, beyond that comtemplated in NUREG-0737,98 Item II.D.1 (e.g., feed-and-bleed). Resolution might entail particular component reliability monitoring, surveillance, and follow-up in service with corrective action for instances of below-average reliability performance. Resolution might entail systems analysis to identify common causes of PORV or block valve failures under circumstances in which their operability was important, perhaps leading to altered power supplies for valve actuations. Resolution might entail special qualification or analysis for water hammer in PORV or code safety valve discharge lines, enlarged flow capacity, or the replacement of relief valves with fast-acting control valves.
This issue was given a medium priority ranking to coordinate efforts to look at the details of the PORV/block-valve situation (e.g., outcome of Item II.D.1, look at data, coordinate new information on, or incentives for, a broader safety mission for PORVs from each of several ongoing programs), and assess, on a schedule tied to related programs, the adequacy of existing PORV/block-valve requirements. The resolution of this issue was later combined with that of Issue 94 so that one set of requirements could be issued to licensees.1287
In resolving this issue, the staff issued Generic Letter 90-061290 which required TS revisions at PWRs with PORVs and block valves. This resolution represented new staff positions for some licensees and CP holders and was considered a backfit. The regulatory analysis for the issue was reported in NUREG-1316.1293 In addition, revisions1294 to SRP Sections 3.2.2, 5.2.2, and 5.4.7 were forwarded to NRR for incorporation into NUREG-0800.11 The Commission was informed of the staff's resolution in SECY-90-153.1295 Thus, this issue was RESOLVED and new requirements were established.1292
REFERENCES
|
