ML18102B467
| ML18102B467 | |
| Person / Time | |
|---|---|
| Site: | Salem  |
| Issue date: | 07/24/1997 |
| From: | Dawn Powell Public Service Enterprise Group |
| To: | NRC OFFICE OF INFORMATION RESOURCES MANAGEMENT (IRM) |
| References | |
| GL-95-07, GL-95-7, LR-N970462, NUDOCS 9708040124 | |
| Download: ML18102B467 (14) | |
Text
-~ e
.o.ps~G Public Service Electric and Gas Company P.O. Box 236 Hancocks Bridge, New Jersey 08038-0236 Nuclear Business Unit **
- LR-N970462 United States Nuclear Regulatory Commission Document Control Desk Washington, DC 20555 RESPONSE TO REQUEST FOR ADDITIONAL INFORMATION GENERIC LETTER 95 11 PRESSURE LOCKING AND THERMAL BINDING OF SAFETY RELATED POWER OPERATED GATE VALVES" SALEM GENERATING STATION UNITS 1 AND 2 FACILITY OPERATING LICENSES DPR-70 AND DPR-75 DOCKET NOS. 50-272 AND 50-311 Gentlemen:
This letter provides the response of Public Service Electric and Gas Company (PSE&G) to the NRC letter dated June 10, 1997 requesting additional information concerning PSE&G's response to Generic Letter 95-07. By telecon on July 2, 1997 between my I staff and the Salem Project Manager, a two week extension was provided for this 30 day response.
Should you have any questions or comments on this transmittal, please contact us.
Sincerely, D. R. Powell Manager -
Licensing and Regulation Attachment Illllll 111111111111111111111111111111111
- '&&5618*
The power is in your hands.
95-2168 REV. 6/94
Otll 2 41997 Document Control Desk LR-N970462 c Mr. Hubert J. Miller, Administrator - Region I U. S. Nuclear Regulatory Commission 475 Allendale Road King of Prussia, PA 19406 Mr. L. N. Olshan, Licensing Project Manager -Salem U. S. Nuclear Regulatory Commission One White Flint North 11555 Rockville Pike Mail Stop 14E21 Rockville, MD 20852 Ms. M. Evans (X24)
USNRC Senior Resident Inspector Hope Creek Generating Station Mr. K. Tosch, Manager, IV Bureau of Nuclear Engineering 33 Arctic Parkway CN 415 Trenton, NJ 08625
JUL 241997*
LR-N970462 Response to. Request for Additional .. Information Gene~ic L~t~er 95-07 .
The technical information provided in this submittal is true and has been validated.
date date 7 clate
LR-N970462 ATTACHMENT RESPONSE TO REQUEST FOR ADDITIONAL INFORMATION GENERIC LETTER 95-07 SALEM GENERATING STATION UNITS 1 AND 2 DOCKET NOS. 50-272 AND 50-311 FACILITY OPERATING LICENSES DPR-70 AND DPR-75 In response to Generic Letter 95-07 (GL), PSE&G provided information to the NRC for Salem Generating Station in letters dated February 13, August 7, and August 30, 1996. In a letter dated June 10, 1997, the NRC requested that further information be provided so that the review of Salem Station for Generic Letter 95-07 could be completed. The following provides this information.
QUESTION 1:
The August 30 submittal stated that calculations were used to demonstrate that the pressurizer power operated relief valve block valves, PR6,7, would open during pressure locking and thermal binding conditions. Describe how these calculations were used to determine if the valves are susceptible to pressure locking or thermal binding. Identify any other valves where these calculations are used as long term pressure locking or thermal binding corrective action.
RESPONSE 1:
As discussed in PSE&G's 180 day response to GL 95-07 dated February 13, 1996, susceptibility to pressure locking and thermal binding was performed using a screening approach. The criteria used in the screening and the results of the screening are described in this submittal. No additional analyses were used to determine the susceptibility of a valve to pressure locking or thermal binding other than the application of the screening criteria.
The screening resulted in the determination that 14 valves were susceptible to pressure locking and 8 valves were susceptible to thermal binding. Since the February 13, 1996 submittal, an additional 6 valves have been determined to be susceptible to pressure locking. PSE&G submittal dated August 30, 1996 described an evaluation of the four PORV Block Valves (1PR6, 1PR7, 2PR6 and 2PR7) for susceptibility to pressure locking during a SGTR event. In response to concerns documented in NRC Special Inspection Report 311/97-11 dated May 22, 1997, the RHR to hot leg recirculation isolation valves (1RH26 and 2RH26) were reinstated in the GL 89-10 Motor Operated Valve Program, screened
ATTACHMENT 1 2
- LR-N970462 for susceptibility to pressure locking and determined to be susceptible. Therefore, 20 valves have been determined to pote.ntially be. susceptible. to. pressur:e, -locking and 8 valves were determined to be susceptible to thermal binding as summarized below:
Pressure Locking Thermal Binding 11CS2, 12CS2, 21CS2 and 22CS2 1PR6, 1PR7, 2PR6 and 2PR7 1PR6, 1PR7, 2PR6 and 2PR7 11CC16, 12CC16, 21CC16 and 1RH26 and 2RH26 22CC16 lSJl, 1SJ2, 2SJ1 and 2SJ2 2SJ12 and 2SJ13 11SJ113, 12SJ113, 21SJ113 and 22SJ113 For the valves susceptible to pressure locking, modifications were performed to the ten valves in the safety injection (SJ) systems to preclude pressure locking. With regard to the four containment spray valves (CS), the surveillance test procedures were revised to preclude pressure locking of these valves. The remaining valves listed (1PR6, 1PR7, 2PR6, 2PR7, 1RH26 and 2RH26) included further evaluation as a long term corrective action.
Pressure Locking Analyses were used to demonstrate that the required thrust for the PR6 and PR7 valves is insensitive to pressure locking.
During a steam generator tube rupture event, the effect of stem ejection force tending to help the valve open offsets the increased disc loads associated with bonnet pressurization conditions. PSE&G has determined that these valves are challenged by potential thermal binding loads, as discussed below. Therefore, the operability and set-up adequacy of the PR6 and PR7 valves is not dependent on pressure locking calculations.
Calculations do establish that pressure locking conditions do not increase the required stem thrust for these valves and that pressure locking does not pose a limiting condition.
Since the February 13, 1996 submittal, the RH26 valves were screened for susceptibility to pressure locking and thermal binding. As a result of their having been determined to be potentially susceptible to pressure locking, further evaluations were performed. The pressure source is assumed to be leakage past two reactor coolant check valves. Testing performed for the Westinghouse Owners Group on valves similar to those installed at Salem indicate that the bonnet cavity, if pressurized, will relieve pressure prior to the time the valve is required to open (14 hours1.62037e-4 days <br />0.00389 hours <br />2.314815e-5 weeks <br />5.327e-6 months <br /> post-accident). The residual bonnet pressure is small, and calculations are used to show that with this small pressurization the stem thrust is within the capability of the actuator to assure the opening function. Therefore, these valves will not pressure lock.
ATTACHMENT 1 3
- LR-N970462 To summarize, pressure locking calculations are used to show that the pressure locking contribution to the required thrust to open is_small._for,,the PR6, ,.PR7 .. and RH26 .valves. - These valves are*
assured to perform their design basis opening function.
Thermal Binding Calculations provide the basis to justify the operability of valves identified previously as being potentially susceptible to thermal binding. The calculations determine the required opening thrust under thermal binding conditions. The maximum closure thrust is an input to these calculations. As closure thrust increases, the required thrust to open, or unwedge, the valve also increases. Accordingly, proper control of the maximum closure thrust is a method which can ensure that the required opening thrust will not exceed the capability of the actuator.
The calculations are used as a basis to define a maximum closure thrust such that the valve will open under its required conditions, including considerations for the increase in the unwedging thrust due to conditions conducive to thermal binding.
For the PR6 and PR7 valves, in addition to this evaluation a control circuit modification was implemented to ensure closure of the valve with the full design basis capability of the motor operator, yet provide a relatively low closure force as compared to limiting maximum closure thrust as determined by the calculation. For the CC16 valves, the evaluation showed no motor control circuit modification was required. The PR6, PR7 and CC16 valves have been set up to ensure that the maximum closure thrust is not exceeded and their respective actuators are assured of having the capability to operate the valves under postulated accident conditions.
QUESTION 2:
On April 9, 1997, a public meeting was conducted to discuss the Commonwealth Edison and Entergy Operations, Inc. pressure locking thrust prediction methodologies presented in GL 95-07 submittals.
The minutes of the public meeting were issued on April 25, 1997 and placed in the Public Document Room. The Commonwealth Edison and Entergy Operations, Inc. methodologies that predicts the thrust required to open pressure locked flexible-wedge gate valves, validation testing of the analytical method, enhancements to the Commonwealth Edison pressure locking methodology, and pressure locking tests sponsored by the NRC conducted by Idaho National Engineering and Environmental Laboratory were discussed during the meeting. The minutes of this public meeting are an example of the type of information requested by the NRC in order to review and approve pressure locking and thermal binding thrust prediction methodologies.
ATTACHMENT 1 4
<LR-N970462 In order for the NRC to review and approve your pressure locking and thermal binding thrust prediction methodologies, please prov.ide,.,the following .information:
QUESTION 2.A:
Describe the pressure locking and thermal binding thrust prediction methodologies and provide the test procedure/results that validated the methodologies. Include any information that will help evaluate if your valves are similar to the test valves.
RESPONSE 2.A:
Pressure Locking The method used in the PSE&G pressure locking calculations was developed by MPR Associates, Inc. It is described in report MPR-1693, "Evaluation of Salem Valves for Pressure Locking and Thermal Binding", Revision 1. This document, which contains proprietary information, was reviewed by NRC personnel onsite during NRC Combined Inspection Report 272/97-03 and 311/97-03. The pressure locking methodology is also described in a paper presented at the Fourth NRC/ASME Pump and Valve Symposium (see Damerell, P. S., et al, "Effects of Pressure and Temperature on Gate Valve Unwedging, NUREG/CP-0152, pp. 3C-49 to 3C-71) .
The pressure locking thrust prediction methodology considers the mechanisms whereby changes in bonnet pressure, upstream pressure and downstream pressure (after the valve is seated) can affect the required unwedging thrust. An approach is used whereby all of the forces acting on the disc are considered. The appropriate areas upon which the pressures act are addressed and the flexibilities of the disc and body are evaluated. The method calculates the disc-to-seat contact forces produced during wedging and the changes in these forces due to bonnet and piping pressure changes. The upper bound (default) friction coefficients from the EPRI gate valve model are used in the calculations (Reference EPRI Report TR-103229). Use of this approach is conservative, since the unwedging thrust (even without pressure locking) is overpredicted. In the EPRI Program this conservatism was established (Reference TR-103237, Section 5)
Based on the descriptions of the Commonwealth Edison (ComEd) and Entergy methods in the public meeting report, the MPR method used by PSE&G is very similar to the ComEd method, and has some of the improvements referred to in the presentation by Kalsi Engineering (MPR originally worked on the development of the ComEd methodology) .
MPR independently completed the methodology, building on the same base of work, but under contract to PSE&G. The PSE&G/MPR method includes body flexibility in the model, whereas the ComEd method did not. Body flexibility is the key area addressed in the improvements to the ComEd methodology suggested by Kalsi Engineering.
ATTACHMENT 1 5
- LR-N970462 The MPR method was validated by comparing predicted thrust to measured thrust from several tests in the EPRI MOV Test Program .
.*. Specif.ically,, .eight _(8) ... tests on . .six (.6 ).. different valves were identified where the bonnet pressure was elevated relative to the piping pressure during an opening stroke of the valves. The six (6) valves are:
- 3-inch Anchor Darling 300-lb flexible wedge gate valve
- 18-inch Anchor Darling 300-lb flexible wedge gate valve
- 3-inch Borg-Warner 1500-lb flexible wedge gate valve
- 4-inch Aloyco 150-lb split wedge gate valve
- 6-inch Velan 900-lb flexible wedge gate valve
- 3-inch Westinghouse 1500-lb flexible wedge gate valve The predicted thrust by the MPR method bounded the measured thrust in all cases. The measured bonnet pressures (above system pressure) during the tests ranged from 50 psi to 1600 psi.
The pressure locking methodology was applied to two (2) sets of valves at Salem. The PR6 and PR7 valves are 3-inch, 1500-lb, Velan flexible wedge gate valves. The RH26 valves are 12-inch, 1500-lb Velan flexible wedge gate valves. These valves are similar to the valves for which the data was obtained.
Thermal Binding Information describing the thermal binding methodology is contained in MPR-1693, "Evaluation of Salem Valves for Pressure Locking and Thermal Binding", Revision 1. This proprietary document was reviewed by NRC personnel onsite during NRC Combined Inspection Report 272/97-03 and 311/97-03. The thermal binding methodology is also described in a paper presented at the Fourth NRC/ASME Pump and Valve Symposium (see Damerell, P. S., et al, "Effects of Pressure and Temperature on Gate Valve Unwedging", NUREG/CP-0152, pp. 3C-49 to 3C-71) .
The thermal binding thrust prediction methodology considers the mechanisms which can increase the disc-to-seat contact and friction loads, including differential expansion of the stem, disc, seat rings and body, and changes in the disc-to-seat friction coefficient while the valve is closed. These load increases are added to the loads produced during disk wedging. As with the pressure locking method, the bounding (default) friction coefficients from the EPRI gate valve model are used, which makes the method conservative.
Even for conditions without thermal binding, the unwedging thrust is overpredicted. (Reference EPRI Report TR-103237, Section 5). The methodology was validated by comparing thrust predictions to data for several tests conducted in the EPRI MOV Program. Specifically, five (5) valves were identified which had test conditions where the temperature decreased while the valve was in the closed position.
The five (5) valves are:
ATTACHMENT 1 6
- tR-N970462
- 6-inch Anchor Darling 900-lb flexible wedge gate valve
- , 6::.in61i B*arg:::*warner* 1500'-lh flexible wedge gate valve
- 2-1/2-inch Velan 1500-lb flexible wedge gate valve
- 6-inch Velan 900-lb flexible wedge gate valve
- 6-inch Walworth 900-lb flexible wedge gate valve The predicted thrust bounded the measured thrust in all cases. The measured temperature decreases for the valves during the tests ranged from 64°F to 243°F.
The thermal binding methodology was applied to two (2) sets of valves at Salem. The PR6 and PR7 valves are 3-inch, 1500-lb, Velan flexible wedge gate valves. The CC16 valves are 12-inch, Crane solid wedge gate valves. These Salem valves are similar to the valves for which the data was obtained. The model differentiates between solid and flexible wedge gate valves. Although there are no solid wedge gate valves in the test group, the flexible wedge gate valves in the test group are among the stiffest of those tested (i.e., most like solid wedge gate valves). Further, the contribution of thermal binding to required thrust was found to be most strongly influenced by friction coefficient changes, which depend on the disc and seat materials and the temperature change, but are not sensitive to disc stiffness. The disc and seat material for the Salem valves and the test valves is Stellite 6. The design basis temperature decreases for the Salem valves are 208°F (PR6 and PR7 valves) and 85°F (CC16 valves), which are within the range covered by the data.
QUESTION 2.B:
Results from pressure locking testing sponsored by the NRC performed by Idaho national Engineering and Environmental Laboratory (INEEL) on a double disk and a flexible wedge gate valve have been placed in the Public Document Room. Please discuss if your pressure locking thrust prediction methodology accurately predicted the results of these pressure locking tests.
It would be helpful if you discussed if your pressure locking thrust prediction methodology accurately predicted the results of pressure locking tests performed by Commonwealth Edison that were discussed during the April 9, 1997 public meeting.
RESPONSE 2.B:
At the time the pressure locking susceptibility evaluation was performed and before the valve and procedure modifications were completed to preclude pressure locking, required thrust and operability calculations were performed for Salem valves. As previously mentioned, the pressure locking thrust calculation
ATTACHMENT 1 7
'LR-N970462 method was successfully validated against data from six (6) valves in the EPRI MOV Program. These were the only data available. at c.the. time*.the work was performed, and were considered to be valid and applicable tests. The INEEL and ComEd data became available after the Salem GL 95-07 work was completed. As discussed in the response to Question 1, modifications or procedure changes were made to preclude pressure locking for all but six (6) Salem valves. For these six (6) valves, pressure locking is determined to be a small or negligible influence.
Because the PSE&G model was fully validated against test data and because pressure locking calculations have a small influence on the evaluations of operability for Salem valves, we believe there is little benefit in further investigation of additional data to validate our pressure locking methodology.
QUESTION 2.C:
NRC Inspection Report 50-311/96-18 identified several concerns with your thermal binding thrust prediction methodology. These concerns involved validation of the methodology and valve body and disk temperature gradient assumptions. Please discuss how you resolved these concerns.
RESPONSE 2.C:
Note that the NRC concerns were identified in NRC Combined Inspection Report 272/97-03 and 311/97-03. The NRC identified two concerns in regard to the MPR analytical method used to determine a maximum inertial thrust limit below which thermal binding is not a concern for valves 21CC16, 22CC16, 2PR6 and 2PR7: (1) the test data did not completely validate the model to determine the susceptibility to thermal binding, and (2) the MPR method may not adequately consider transient or steady state temperature gradients in the valve body or valve disc.
(1) Test data and model validation There are some aspects of the thermal binding model which are not validated by the data which were used. The data validated one key phenomenon of the model (disc-to-seat friction changes) .
For the PR6 and PR7 valves, the thermal binding thrust calculation is only affected by three phenomena: stem heating after insertion (stem growth), disc-to-seat friction changes, and stem-to-body differential expansion. Disc-to-body differential expansion has no influence on the analysis because they are of identical materials.
Further discussion is included under "Transient and Steady State Temperature Gradients" below that supports this determination.
Stem-to-body differential expansion has a small influence on the result and can be ignored. The other two phenomena have comparable effects on the thrust, but only one (disc-to-seat friction changes) was addressed by the validation. The predicted effect of the other
ATTACHMENT 1 8
<LR-N9'70462 phenomenon (stem heating after insertion) is considered to be acceptable because a bounding approach was used. Specifically, the full ,temperature *difference between the ,process fluid and.the ambient was applied to the stem. The valve stem on a hot system is typically observed to be well above the ambient temperature, which means that the actual heatup after stem insertion is less than that used in the calculation. Further, the stiffness used to determine the stem force change from the expansion due to heatup was validated against data from the EPRI Program, including data from Valve #13, which is similar to the PR6 and PR7 valves. The measured flexibility for EPRI Valve #13 (3.0xlOE-6 in/lbf) is about 50%
greater than the value used in the model (2.0xlOE-6 in/lbf), which means that the thermal binding thrust is overpredicted by a similar amount. Although this one model feature was not validated against valve testing which included this phenomena, the results bound the effects which could occur in the valve.
For the CC16 valves, the thermal binding thrust calculation is affected by only one phenomenon which is stem heating by insertion.
The scenario that would create the analyzed conditions involved an inadvertent closure of a CC16 valve when the RHR heat exchanger is in service for shutdown cooling or post-accident cooldown and the process fluid temperature decreases 85°F to an ambient temperature of 100°F. It is expected that if a CC16 valve were inadvertently closed during such operation, it would be reopened before any appreciable heat loss would occur. Thus, additional opening force requirements attributable to thermal binding are not likely. The calculation performed for this bounding thermal binding condition from stem heating by insertion resulted in an increase to the stem thrust by only 11%. Further, this calculation is considered to be bounding for the same reasons as discussed in the previous paragraph.
Specifically, the full temperature difference between the process fluid and the ambient was applied to the stem, although the actual heatup would be less. Further, the stiffness used to determine the stem force change was validated against data from the EPRI Program, including data from Valve #31, which is similar to the CC16 valves.
The measured flexibility for Valve #31 (2.33x10E-6 in/lbf) is about 17% greater than the value used in the model (2.0xlOE-6 in/lbf),
which means that the thermal binding thrust is overpredicted by at least a similar amount. Although this one model feature was not validated against valve testing which included this phenomena, the results bound the effects which could occur in the valve.
(2) Transient and Steady State Temperature Gradients The method does not consider transient or steady state temperature gradients in the valve disc or body. Although such gradients can exist in some situations and could affect required thrust, the scenarios for the PR6 and PR7 valves and for the CC16 valves do not include these circumstances. Specifically, these valves are involved in cooldown scenarios in which the cooldown is very gradual to a uniform temperature surrounding the valve (CC16 valves) or is
ATTACHMENT 1 9
~tR-N970462 driven by the depressurization of the steam space on one side of the valve (PR6 and PR7 valves) . In the first case, negligible thermal gradients .. would be. expected, since. the. cooldown. is by natural convection and the system has "anti-sweat" insulation which prolongs the cooldown. In the second case, the cooldown is driven by the steam temperature reduction in the upstream piping. The steam is in direct contact with the valve disc. Thermal gradients, if they occur, would be expected to result in a bulk valve disc temperature which is less than the valve body temperature. This would tend to reduce thermal binding effects. Accordingly, neglecting thermal gradients is appropriate for the PR6 and PR7 valves.
QUESTION 2.D:
Discuss the recommended margin between actuator capability and the calculated thrust value when using your pressure locking and thermal binding thrust prediction methodologies and if applicable, explain the basis and application requirements for the individual elements of this margin.
RESPONSE 2.D:
The thermal binding calculations are performed to determine a maximum allowable closure thrust which will give zero margin for opening under thermal binding conditions. The valves are then set up to provide a margin against this value. The margin for the four (4) Salem Unit 2 valves affected by the thermal binding analyses ranges from 37% to 64% and accounts for potential diagnostic error.
Implementation of the control circuit modification for the Unit 1 PR6 and PR7 valves to convert from torque to limit (position) control is scheduled for completion prior to Salem Unit 1 restart and adequate margin results are expected. Diagnostic testing of the Unit 1 CC16 valves is also scheduled to enhance performance parameters and is also expected to achieve adequate margin results.
QUESTION 2.E:
Recent pressure locking testing sponsored by the NRC performed by Idaho National Engineering and Environmental Laboratory indicated that the force required to unwedge a valve can significantly deviate as the valve is repeatedly stroked closed and then reopened. Discuss how this variance in unwedging force is accounted for in your pressure locking and thermal binding thrust prediction methodologies.
RESPONSE 2.E:
The variance in unwedging force is related to changes in the disc-to-seat friction coefficient as the valve is stroked. We account for this variance by using the bounding friction
ATTACHMENT 1 10
-LR-N9'704 62 coefficients from the EPRI gate valve method (EPRI Report TR-103229). We additionally consider a bounding value for normal unwedging thrust . to which-- the calculated increment .of additional thrust is added. Our method determines bounding values of unwedging thrust and the validations against test data affirm this conclusion.
QUESTION 3:
The August 7 submittal stated that the valves that were closed during surveillances would be evaluated for susceptibility to thermal binding and pressure locking prior to starting up the units. Please provide the schedule for submitting the results of this review to the NRC.
RESPONSE 3:
PSE&G identified the population of normally open, safety related power-operated gate valves that were previously screened out as not being susceptible to pressure locking or thermal binding based on only having a safety function to close. From this population those valves that are surveillance tested in modes where the system provides no automatic safety function were screened out as they do not defeat the capability of the safety system at the time the surveillance test is conducted.
The valves in the remaining population were further screened for those which the surveillance is limited to a stroke time test of short duration. Due to the short duration the valve is in the closed position, pressure locking and thermal binding are not an issue of concern. These valves meet the requirements of Option 1 from the Generic Letter 95-07 workshops.
The valves in the remaining population were screened for those that are required to open to accomplish the safety function of the system when the surveillance is conducted. For these valves the applicable Technical Specifications are followed for the train/system while the valve is closed.
Two valves remained to be considered. Procedures for performing the inservice testing of the charging pumps require the BIT inlet isolation valves (SJ4 and SJ5) to be closed. If closed under testing conditions, these valves are potentially susceptible to pressure locking. Salem Unit 2 procedures have been revised to maintain these valves open during this pump surveillance since the outlet isolation valves (SJ12 and SJ13) provide the isolation function and have been modified to preclude pressure locking.
Similar procedure changes will be completed for Salem Unit 1 prior to Salem Unit 1 entering Mode 6, Refueling.
On the basis of the reviews described above, PSE&G has determined that the subject valves are either not susceptible to pressure
ATTACHMENT 1
- LR-N970462 11 locking and thermal binding or appropriate technical specification guidance is considered for the train/system while the valve .is .. closed . . PSE&G's .evaluations .and results are .being validated and final results will be provided in 30 days.