ML042940294
| ML042940294 | |
| Person / Time | |
|---|---|
| Site: | Boiling Water Reactor Owners Group, PROJ0694, PROJ0693 |
| Issue date: | 10/20/2004 |
| From: | Macon W NRC/NRR/DLPM/LPD4 |
| To: | Crawford H, Putnam K, Schiffley F BWR Owners Group, Westinghouse Owners Group |
| Macon W, NRR/DLPM,415-3965 | |
| References | |
| TAC MC2346, TAC MC2347, TAC MC2348 | |
| Download: ML042940294 (12) | |
Text
October 20, 2004 Mr. Howard C. Crawford, Chairman B&WOG Executive Committee AmerGen Energy Company Route 441 South P.O. Box 480 Middletown, PA 17057-0480 Mr. Kenneth S. Putnam, Chairman BWR Owners Group Nuclear Management Company Duane Arnold Energy Center 3277 DAEC Road Palo, IA 52324 Mr. Frederick P. Schiffley, II, Chairman Westinghouse Owners Group Exelon Nuclear Engineering Design Cornerstone II at Cantera 4300 Winfield Road Warrenville, IL 60555
SUBJECT:
REQUEST FOR ADDITIONAL INFORMATION - MPR-2524, REVISION 0, "JOINT OWNERS GROUP (JOG) MOTOR OPERATED VALVE PERIODIC VERIFICATION PROGRAM
SUMMARY
" (TAC NOS. MC2346, MC2347 AND MC2348)
Gentlemen:
By letter dated February 27, 2004, the JOG submitted Topical Report (TR) MPR-2524, Revision 0, "Joint Owners' Group (JOG) Motor Operated Valve Periodic Verification Program Summary," for NRC review and approval. The NRC staff has performed an initial review and audit of the TR and finds that it needs additional information to complete its review.
Therefore, it is requested that you respond to the enclosed request for additional information by February 15, 2005, for the NRC staff to complete its review. The response date was proposed by Mr. Chad Smith representing the JOG via e-mail on October 12, 2004, and agreed to by the
Multiple Addressees NRC staff. Except for a couple of typographical corrections, the enclosed questions are unchanged from those sent by e-mail to Mr. Smith on September 28, 2004.
Sincerely,
/RA/
William A. Macon, Jr., Project Manager, Section 2 Project Directorate III Division of Licensing Project Management Office of Nuclear Reactor Regulation Project Nos. 691, 693, and 694
Enclosure:
Request for Additional Information cc w/encl: See next page
ML042940294 NRR- 088 OFFICE PDIV-2/PM PDIV-2/LA PDIV-2/SC NAME WMacon EPeyton RGramm DATE 10/18/04 10/18/04 10/20/04 DOCUMENT NAME: C:\\ORPCheckout\\FileNET\\ML042940294.wpd
Joint Owners Group cc:
Mr. Charles B. Brinkman, Director Washington Operations Westinghouse Electric Company 12300 Twinbrook Parkway, Suite 330 Rockville, MD 20852 Mr. Gordon C. Bischoff, Manager Owners Group Program Management Office P. O. Box 355 Pittsburgh, PA 15230-0355 Mr. David J. Firth Manager, B&W Owners Group Services Framatome ANP 3315 Old Forest Road Lynchburg, VA 24501 Mr. J. J. Kelly, Manager B&W Owners Group Services Framatome ANP 3315 Old Forest Road Lynchburg, VA 24501 Mr. Thomas G. Hurst GE Nuclear Energy M/C 782 3901 Castle Hayne Road Wilmington, NC 28402 Mr. Paul Damerell MPR Associates, Inc.
320 King Street Alexandria, VA 22314-3230 Mr. Todd Spears MPR Associates 320 King Street Alexandria, VA 22314-3230
REQUEST FOR ADDITIONAL INFORMATION JOINT OWNERS GROUP (JOG) PROGRAM ON MOTOR-OPERATED VALVE PERIODIC VERIFICATION By letter dated February 27, 2004, the JOG submitted Topical Report (TR) MPR-2524, Revision 0, "Joint Owners Group (JOG) Motor Operated Valve Periodic Verification Program Summary," for NRC review and approval. The NRC staff has performed an initial review and audit of the TR and finds that the following additional information is needed to complete its review.
1.
On page 7-55, the TR states that applications of the JOG Motor-Operated Valve (MOV)
Periodic Verification (PV) Program to unbalanced disk globe valves operating with non-flashing water above 150°F are covered by extension. As indicated in Discussion Item 3 provided in the summary of the NRC public meeting with the JOG on October 1 and 2, 2003 (dated October 14, 2003), the basis for assigning valves to particular classes in the JOG long-term MOV PV recommendations should be specified.
2.
As indicated in Discussion Item 9 provided in the summary of the October 2003 public meeting, the change in valve friction coefficient over the entire 5-year test program should be considered in addition to friction coefficient changes from one stroke to another.
3.
On pages 7-14 and 15, the TR discusses the application of Criteria 4.1 and 4.2 in determining whether gate valves have a "qualifying basis" in justifying that the required thrust is not subject to degradation. These criteria are intended to allow a licensee to demonstrate that the required thrust for a particular gate valve is stable. Figure 7-2 of the TR indicates that variation might occur in the coefficient of friction for gate valves considered to have stable behavior. The JOG is requested to explain the manner in which potential variation in "stable" valve performance is accommodated in the evaluation of required thrust/torque for gate, globe, and butterfly valves, as applicable.
4.
In Section 7, the TR allows licensees to use judgement in implementing the JOG program (e.g., see discussion of engineering judgement on page 7-2; evaluation of Class D valves on pages 7-2 and 58; extension of the Electric Power Research Institute (EPRI) MOV performance prediction methodology (PPM) on pages 7-5, 28, 44 and 54; evaluation of data to determine whether required thrust is controlled by disk-to-seat friction on page 7-12; applicability of differential pressure data on page 7-13; determination of valve strokes to achieve reliable friction plateau on page 7-14; justification of the results under plant specific conditions on pages 7-14 and 34; justification for valve grouping on pages 7-15 and 34; and evaluation of balanced disk globe valves on page 7-48). The JOG is requested to discuss the basis for allowing judgement without specific implementation guidance in the long-term MOV PV program for gate, globe, and butterfly valves.
5.
On page 7-2, the TR defines Class A valves as those valves not susceptible to degradation, as supported directly by testing performed in the JOG program or other suitable basis (e.g., the EPRI MOV PPM). The JOG is requested to clarify that the suitable bases intended by this definition are those specified in the TR.
6.
On page 7-4, the TR states that for the majority of gate valves tested in the JOG program, there was no observed service-related degradation of required thrust due to differential pressure stroking. The JOG is requested to clarify the potential for service-related degradation for the remaining gate valves in the JOG program (e.g., whether the potential for degradation is dependent on valve type, maintenance, or performance characteristics).
7.
On page 7-9, the TR states that valves that have a design-basis function to operate under differential pressure conditions, but do not stroke against differential pressure in-service can have a rating that may lead to Class A categorization (which involves minimal margin requirements with extended static test intervals). A similar discussion is provided for globe valves on pages 7-46 and 54. Appendix B to the TR allows special, infrequent valve strokes (deliberate or inadvertent) under differential pressure conditions during a scenario that is not expected to be repeated to not be counted as differential pressure stroking. The JOG is requested to discuss the potential for increased thrust or torque requirements in the performance of gate, globe, or butterfly valves, as applicable, as a result of inadvertent differential pressure strokes, including the basis for the guidance provided in Appendix B to the TR.
8.
Table 7-1 of the TR shows a matrix of recommended testing intervals for Class A, B, and C valves as they relate to the plant safety risk and margin. In a risk-informed matrix such as this, one would expect that the impact on plant safety associated with the periodic verification testing program is expected to be constant along the matrix diagonals. Therefore, one would expect the test periodicity to also be the same along the matrix diagonals. This is true in Table 7-1 with the exception of the test interval for the combination of medium risk and medium margin MOVs. Please justify the recommended eight year test frequency for this case rather than the six year test frequency recommended for MOVs categorized as low risk and low margin, or categorized as high risk and high margin.
9.
On page 7-32, the TR indicates that there were no valves tested in compressible flow (such as steam or air) in the JOG program, but notes that there are butterfly valves in nuclear power plant service that operate in air or nitrogen service. The TR suggests that testing performed by the Idaho National Engineering and Environmental Laboratory (INEEL) under NRC sponsorship shows bearing friction coefficients in air that were similar to that in treated water. The JOG concluded that the application of the JOG program results for valves in treated water to valves in air, steam or nitrogen service is reasonable. The INEEL air testing results were for refurbished valves that were tested only in the closing direction. No age-related data was obtained for those tests. Please discuss the applicability of the INEEL and EPRI test results to aged butterfly valves tested in the JOG program. Explain why the treated water butterfly test results can be expected to bound the results for all butterfly valves and bearing material combinations in air/nitrogen at less than 150°F.
10.
The TR states that the observed decrease in bearing friction coefficient for butterfly valve B11.1 (Figure 4-3) between the baseline and second test is a result of an unusually low static unseating torque (133 ft-lbs) in the baseline test compared to the measured static unseating torque in the second and third tests (213 to 224.5 ft-lbs).
The JOG also indicates that the unseating behavior is more consistent across the three tests as seen in the overlaid dynamic torque traces. The differential pressure (DP) torque overlay traces do show a small decrease in DP unseating torque between each test; however, the JOG states that the DP torque for all three tests is almost identical at the beginning of the stroke. The JOG concluded that, based on the DP torque overlay plot, the change in bearing friction from baseline to second test is attributable to the unusually low static seating torque in the baseline test and that the bearing friction behavior was therefore stable across the three tests. Since it was not possible to discern the dynamic unseating torque values from the torque overlay traces in Figure 4-3, the JOG provided the dynamic and corresponding static unseating torque values for all B11.1 valve test series during the audit. When compared, this data showed that the behavior of the dynamic unseating torque was consistent with the behavior of the corresponding static unseating torque values. Like the static test data, the baseline dynamic unseating torque was also significantly lower than the second and third tests series. It is not clear that the dynamic torque data alone suggests that the observed coefficient of friction (COF) change between the baseline and second test was not valid.
Also, it is not clear from the discussion in the TR that other potential reasons for the observed increases in static unseating torque were addressed. Please provide additional discussion regarding the interpretation of these results, the potential causes considered and the evaluations performed by the JOG to support the conclusion regarding the validity of the baseline test data.
11.
The TR states that the interim TR identified the "accumulation of particles in the bearing" as a potential degradation mechanism. For all butterfly valves tested in untreated water systems, the TR states that significant variations in the COF between tests were observed and that, for several cases, the observed changes in COF cannot be completely attributed to measurement uncertainty. The TR states that these variations are due to the effects of untreated water. During the audit, the JOG referred to untreated water effects as dirt/particles in the bearing. The TR also concluded that the thrust variations observed in balanced disk globe valves tested in untreated water were particulates interfering with disk motion. However, the TR does not discuss these observations and conclusion relative to the potential for particle accumulation degradation mechanisms. Please discuss the nature and cause of the untreated water effects observed and the technical basis that supports the JOG conclusion that particle accumulation in the bearing is not a potential degradation mechanism in butterfly and balanced globe valves. Discuss the relevancy of this conclusion with respect to variations in the state of untreated water systems from plant-to-plant and describe any limitations associated with this conclusion.
12.
With regard to untreated water systems, are the JOG results applicable to all safety-related raw water systems in boiling water reactor (BWR) and pressurized water reactor (PWR) nuclear power plants? For example, please discuss the relevancy of the test results, conclusions, and recommendations for gate, butterfly, and globe valves in the TR for both service water systems and salt water systems?
13.
The TR shows that the COF associated with butterfly valves having hub seals in untreated water did not vary significantly. In the summary of the October 2003 public meeting, the NRC staff stated that it would be beneficial for the JOG to compare its results to operating experience from nuclear power plants that have experienced significant variations in the performance of their butterfly valves. Please discuss the manner in which this issue should be addressed by JOG participants.
14.
In the summary of the October 2003 public meeting, the NRC staff stated that the JOG applied deterministic approaches to bound 95 percent of the test data. The NRC staff indicated that this approach is generally consistent with the analysis of test data during other MOV activities. However, the staff recommended that the JOG TR apply deterministic approaches where the applicability of the statistical approach is not readily apparent and that the JOG use statistical approaches to confirm the deterministic approaches where practical. The staff stated that it would be beneficial to compare the deterministic 95 percent bound test value for bearing friction coefficient to a value determined from the data set using a statistical approach (such as the mean of the data plus two standard deviations). Please address these comments and compare a deterministic bound of 95 percent of the test data with a statistical mean + 2-sigma approach.
15.
The TR states that bearing friction behavior of the bronze bearing butterfly valves with hub seals tested in untreated water was relatively constant over the test series and the variations in COF were small at less than 0.05. The TR also states that this behavior is consistent with that observed for the bronze bearing butterfly valves tested in treated water. The NRC staff notes that a review of the bronze bearing test data shows the median change in COF is approximately a factor of three greater than the median change in COF between tests. Although not addressed in the TR, some of the measured friction coefficients observed between the treated water tests change by more than 20 percent. If the changes between the baseline and second tests are considered for JOG valve B11.1, the variation in COF would be much higher. In addition, the magnitude of the bearing friction coefficients for valves tested in treated water is consistent with the COF values observed for butterfly valves without hub seals tested in untreated water. In this case, the median change in friction coefficient for the valves tested in treated water was approximately a factor of three less than the changes observed in valves without hub seals tested in untreated water. Please discuss how the observed COF changes for the treated water test valves compared with the estimated test instrument uncertainties.
16.
The TR states in Table 7-7 that the bearing friction coefficient threshold of 0.39 for bronze bearing butterfly valves bounds 95 percent of the measured test data for the bronze bearing butterfly valves. The JOG states that this approach is slightly more conservative (and more suitable) than simply bounding 95 percent of the results for bronze bearings in untreated water, without hub seals, where the data were more limited. A statistical review of the test data for the bronze bearing butterfly valves indicates that the mean COF and standard deviation for valves tested in treated water were slightly more conservative (mean + 2-sigma = 0.44) than for valves without hub seals tested in untreated water (mean + 2-sigma = 0.42). These values are approximately a factor of three greater than the results for hub seal valves tested in untreated water. The recommended threshold COF of 0.39 is equal to the 95 percentile COF for the no-hub-seal bronze bearing butterfly valves tested in untreated water and is slightly less than the 95 percentile for the treated water valves. The TR states that, for several valves tested in untreated water, the large variations in COF were beyond that which could be attributed to instrument uncertainties and could result from untreated water effects. Most of these valves were stroked numerous times under DP conditions between tests. Therefore, since only three test series were completed for each valve and considering the varying nature of the state of untreated water systems in different nuclear power plants, there exist added uncertainties regarding the true range of friction coefficients for the valves tested. Unlike the threshold values for other butterfly valve bearing materials, the report does not appear to account for these additional uncertainties when establishing the threshold values for the bronze bearing valves.
Please discuss the justification for the proposed bronze butterfly valve threshold value in Table 7-7 relative to these concerns.
17.
Please describe the differences in the proposed threshold values in Table 7-7 with the default friction coefficient assumptions that would be applied if the EPRI MOV PPM method is used.
18.
The TR states that for 300 series stainless steel bearings, data were obtained for only one valve and the results showed variations. For this material, a threshold of 0.60, which is 20 percent greater than the maximum measured value (0.5), was selected.
The TR states that the higher value was chosen to provide margin and to be consistent with typical maximum values of metal-to-metal friction. Please identify the source of the typical maximum metal-to-metal friction coefficient values used. Please discuss the applicability of the testing conditions in which these values were obtained. How do these values compare with friction coefficients for gate valve test results for 300 stainless steel and 17-4PH disk-to-guide testing results?
19.
In Section 3, the TR discusses extending the gate valve test results that were obtained in water and steam to air/nitrogen based on separate effects testing performed during the EPRI program or by others. These earlier test programs were not aging programs and any conclusions from that work need to be presented and evaluated in order to verify that the results can be extended to air/nitrogen. The JOG is requested to discuss the justification for extending the water and steam gate valve test results by others to air/nitrogen service conditions.
20.
In Section 3, the TR presents threshold coefficient of friction values for valves with non-Stellite seat and disk material. The friction values are much higher than either the unaged or aged frictional values for Stellite. The TR does not discuss whether these values are typical of the frictional values of unaged non-Stellite surfaces. If the results were typical of unaged surfaces, then the results might not support the presence of a corrosion film affecting the friction between the surfaces. However, if the values for an unaged surface were lower, then the presence of corrosion would be expected. The JOG is requested to discuss the basis for the recommendation in light of any unaged friction data for these materials.
21.
In Section 3, the TR states that for some valves that require hard seating, the thrust that needs to be applied could be affected by the direction of flow through the valve. For instance, the test results for both the Anchor/Darling double-disk gate valves and the Aloyco split wedge gate valves were shown to be sensitive to the direction of flow through the valve. The flow direction through these valves may be important in estimating the thrust demands of such valves, unless worst-case performance characteristics are used for both directions. The JOG is requested to clarify whether the data used to generate the threshold friction value included all the flow data for the flow direction sensitive valves, or just the flow data from the worst case flow direction. If all the flow data was used, the JOG is also requested to discuss the justification for using all the data instead of data from only the flow sensitive direction.
22.
In Section 3, the TR states that valves in hot water systems behave in an identical manner to valves in cold water systems. For the purpose of determining a threshold friction, the results for the cold water valves were combined with the results for the hot water valves. If the valves do behave in an identical manner, the threshold friction for just the cold water valves should be about the same as the threshold friction for just the hot water valves. The JOG is requested to discuss whether the threshold friction values for the cold water valves are about the same as the hot water valves, or whether there is a bias in the proposed threshold friction.
23.
In Section 3, the TR states that the one valve that was tested in a hot water system had higher valve factors that were repeatable between tests. The test results for this valve were reviewed and the friction for this valve was not only very high, but also not representative of the cold water valve test results. The JOG is requested to discuss whether the very high, although apparently repeatable, friction for this valve is expected to be typical of all hot water valves. The JOG is also requested to discuss the decision to include the test results from a valve with such high friction with the results from cold water valves that do not exhibit the same type of frictional behavior. In addition, the JOG is requested to discuss the basis for extending the results to all hot water valves considering that the high friction value for the one hot water valve will be eliminated from the threshold friction calculation and not be represented in the resulting threshold friction value. In other words, should the threshold friction from the resulting cold water valve testing be used for the hot water valves?
24.
In Section 3, the TR includes information as to the number of valves that were part of each group. For instance, the group that consisted of gate valves tested in steam service contained 11 valves. The TR also states that the test results that were used to develop the threshold friction for the steam service valves consisted of 37 tests (successful first and second tests at one of the three test intervals). However, it appears that the majority of these 37 tests came from about one third of the 11 valves.
In essence, the threshold friction for steam valves was based on the results from only a few valves. Considering that a subset of the total steam valve test population was actually used to develop the final recommendations, the JOG is requested to discuss whether the resulting recommendations contain any limitations or biases that would compromise the conclusions.
25.
In Section 3, the TR includes information as to locations during the valve stroke that were used to evaluate the response of a valve. Some of these locations might include unwanted transient effects. For instance, unwanted transient effects are most likely present at flow isolation and at flow initiation. At these two points, the pressures and differential pressures are nearly stable; however, the thrust is still undergoing rapid change. With the pressure effects steady, it is not clear why the thrust should be undergoing such rapid change; although one explanation would be that the disk to seat friction has not stabilized. The data packages indicate that the friction values at these two points are typically 5 to 15 percent lower than at initial wedging or just after cracking. The JOG is requested to discuss the stability of the friction at flow isolation and at flow initiation, and whether using the friction results from these two locations could bias the threshold friction.
26.
In Section 3, the TR includes information on age-related degradation mechanisms and how the test results were used to evaluate such degradation mechanisms. However, the evaluation did not discuss other evidence of age-related degradation mechanisms.
The Stellite aging testing that was performed by INEEL concluded that three trends would support the presence of age-related degradation. These trends include: (1) the no to low stroke friction values as a group are higher than the high stroke friction values, (2) the friction consistently decreases from the first stroke of a test to the second stroke, and (3) the friction consistently increases between the last test of a test period and the first test of the subsequent test period. These trends become less pronounced with increased valve strokes between tests. If age-related degradation were not present: (1) the difference between the no/low stroke friction values and the high stroke friction values would be random, (2) the friction values between the first and the second stroke of a test would be random, and (3) the friction values between the last test of a test period and the first test of the subsequent test period would be random. The test results at the initial wedging position for the steam valves show evidence of such age-related degradation. The JOG is requested to discuss the implication of the observed age-related degradation mechanism and any effect such a degradation mechanism might have on the recommendations presented in Section 7 of the TR.
27.
On page 7-12, the TR discusses disk-to-body guide materials for those valves whose maximum opening thrust is controlled by the guides. However, it is not clear how this thrust controlling mechanism will be evaluated. For instance, a valve that is susceptible to the disk-to-guide interaction affecting the maximum thrust might never actually be tested if the valve is evaluated based on testing of other valves in a group. Even if the specific valve was tested, testing at less than design-basis conditions might not reveal that the guides are influencing the thrust. The JOG is requested to discuss how this thrust controlling mechanism will be evaluated for individual plant valves.
28.
On page 7-15, the TR discusses testing up to two valves from a group and then applying the results to all valves in that group. The JOG is requested to discuss whether there should be a limit to the number of valves in a group, or whether a minimum percentage of valves in a group should be tested.
29.
In several locations, the TR indicates that engineering judgement was used to extend the application of the JOG program recommendations. For example, see selection of 150°F as temperature threshold on pages 7-32 and 47; performance of butterfly valves with Stellite bearings in untreated water with a hub seal on page 7-33; selection of data points and valves for determining appropriate bearing friction coefficient on page 7-35; extension to disk-to-guide material combinations not tested in the JOG program on page 7-45; applications with non-flashing water above 150°F on page 7-55; selection of 86 ft/sec for the maximum globe valve flow velocity on page 7-55; definition and counting of DP strokes in Appendix B to the TR; and determination of disk-to-seat COF allowances in Appendix E to the TR. The JOG is requested to discuss the basis for its use of engineering judgement in the specific applications in the TR.
30.
In selecting the threshold values for friction coefficients for JOG program valves, the TR uses a 95 percent bounding approach. Please discuss the need to consider information provided by those valves that demonstrated performance characteristics not bounded by the assumed threshold value during plant-specific application of the JOG program.