ML19261C487
| ML19261C487 | |
| Person / Time | |
|---|---|
| Site: | Millstone, Hatch, Monticello, Dresden, Peach Bottom, Browns Ferry, Nine Mile Point, Fermi, Oyster Creek, Hope Creek, Cooper, Pilgrim, Brunswick, Vermont Yankee, Duane Arnold, Quad Cities, FitzPatrick |
| Issue date: | 02/27/1979 |
| From: | Office of Nuclear Reactor Regulation |
| To: | |
| Shared Package | |
| ML19261C478 | List: |
| References | |
| TAC-07939, TAC-07940, TAC-7939, TAC-7940, NUDOCS 7903230011 | |
| Download: ML19261C487 (12) | |
Text
.
SAFETY EVALUATION REPORT OF THE INTERIM ASSESSMENTS OF MULTIPLE-CONSECUTIVE SAFETY-RELIEF VALVE ACTUATIONS IN.
MARK I CONTAINMENT PLANTS 790323 cro f (
INTRODUCTION On October 6,1977, the General clectric Company (GE) informed the staff of a design deficiency in the safety-relief valve (SRV) control system for the BWR-6 / Mark III system design.
In a letter dated October 11,1977 (G. Sherwood,' GE, to N. Mosley, NRC) and during a meeting between representatives from the Mark I Owners' Group, GE, and the NRC staff on October 27, 1977, the implications of this design deficiency to the operating BWR facilities were discussed. Subsequently, the staff requested that each utility formally submit their basis for continued plant operation by November 1,1977.
In letters dated March 20, 1978, the staf f presented the assumptions and criteria that were to be used for interim plant-specific assessments of the affected operating BWR facilities, until this issue could ultimately be resolved as part of the Mark I Containment Long-Term Program (LTP).
The design deficiency identified by GE concerns the potential for multiple-consecutive SRV actuations following a reactor isolation transient event.
Isolation of the primary system will cause a pressure rise within the reactor vessel. When the pressure reaches the setpoints of the SRVs, the valves will open and discharge steam into the suppression pool, thereby counteracting and eventually reducing the primary system pressure. When the system pressure drops to approximately 800 to 900 psia, the valves will automatically reclose. However, the decay heat produced by by the core will cause the pressure to rise again resulting in repeated SRV actuations. Consecutive SRV actuations, referred to as " hot pops " cause a higher loading on the suppression chamber (torus) and its support structures due to an increase in the length of the water leg and internal energy of the airspace in the SRV discharge line, as compared to that normally existing prior to an SRV actuation. Multiple SRV actuations also cause a higher net loading on the torus due to a reinforcing of multiple sources to specific locations on the torus.
During the Mark I Containment Short-Term Program (STP), operating experience and in-plant test data from the Quad Cities plant indicated that the loads were sufficiently low that only fatigue cycling need be considered. On this basis, SRV discharge loads were classified as secondary loads and excluded from a more detailed con-sideration in the STP (Ref.1). However, this conclusion did not consider the potential for multiple-consecutive SRV actuations. The design deficiency identified by General Electric resulted from a more detailed transient analysis indicating that several SRVs would experience consecutive actuations following a design bisis reactor isolation transient (e.g., closure of all main steamline isolation valves).
In our meeting on October 27, 1977, the Mark 1 Owners and GE presented the results of a generic assessment of the effects of mul tiple-consecutive SRV actuations. The Mark 1 Owners considered the numoer of valves credicted by the a"alysis to actuate con-secutively to be overly conservative in comparison to operating experience. Therefore, the generic assessment was based on what the Owners' Group considered to be a more realistic estimate of the number of valves which would experience consecLtive actuations.
The resulting structural response was based on more recent data obtained from the Monticello in-plant SRV discharge tests (Ref. 2).
The staf f concluded that the generic assessment did not provide an adequate basis for interim resolution of this issue, since it did not consider the plant SRV configurational differences and there was subjective judgment involved in the application of the Monticello test results. Therefore, on March 20, 1978, letters were sent to each of the Mark 1 Owners requesting that they perform an interim, plant-specific assessment. These letters specified the criteria to be used to perform the interim assessment and indicated that the structural response should be compared to a limiting strength ratio of 0.5, in accordance with the structural acceptance criteria for the Mark 1 STP. Continued operation was permitted, based on past operating experience of transient isr sation events, during the period while the assess-ments were.ing performed.
Table i summarizes the licensees' responses and the corrective action taken where it was necessary to satisfy the acceptance criteria for the multiple-consecutive SRV discharge event.
EVALUATION Each plant-unique structural analysis of the response of the torus and its support system to the loads associated with multiple-consecutive SRV actuations was performed in accordance with the following staff criteria included in the March 20, 1978, letters:
"(1) The number of valves which experience subsequent actuatio-shall be determined f rom a plant-unique assessment of the transient which reflects the valves groupings and the SRV setpoints in the facility's Technical Specifications.
Variations in +he SRV setpoints may be accounted for, provided all
- the setpoints are distributed in a manner dictated by ac$ual SRV performance testing. Plants with similar SRV discharge arrangements may De grouped for this assessment, provided their similarity is demonstrated.
"( Although discussions are currently being held between GE and the staf f regarding the transient analysis models used to predict the SRV response sequence, we conclude that the current models are acceptable for this interim assessment.
The ultimate resolution of this issue in the Long-Term Program will require the use of transient analysis models which resolve staff concerns regarding the current models).
"(2) The plant specific variations to the hydrodynamic charac-teristics of the SRV discharge line configurations shall be accounted for by the use of a correction factor derived from the SRV discharge analytical model. This factor shall be based on average line conaitions for those lines pre-dicted to subsequently actuate, as compared to the Monticello
' Bay D' discharge conditions. The basis for averaging shall be described and justified.
"(3) All available peak structural response data for single SRV discharge events, with approximately the same distances between the discharge point and a point on the structure, should be averaged to obtain the expected values of peak structural response at that point as a function of its distance from the discharging SRV. Certain data may be omitted if it can be demonstrated that such data are inconsistent and should not be considered.
"(4) For structures excited primarily by the overall moments of the torus (e.g., the suction header, the torus support columns, the ring header,.etc.), the absolute sum of the structural responses to single SRV actuations shall be used to determine the ef fects of the same valves actuating simultaneously.
"(5) The consecutive valve actuation factors shall be determined from the Monticello data, or any other available test data, by considering the peak structural responses for an appro-priate set of gauges for all consecutive valve actuation tests.
For a given set of gauges, the mean plus one standard deviation of all peak structural responses for each gauge shall be compu-ted. These values, in conjunction with the appropriate cold pipe condition structural responses, shall be utilized to compute a set of consecutive actuation factors. These consecutive valve actuation factors shall be averaged to determine one consecutive valve actuation factor which is applicable to the area (s) of the structure for which this set of gauges is appropriate. Certain data may be omitted if it can be demonstrated that such data are inappropriate and should not be considered.
4-
"(6)
If the results of this assessment indicate that the limiting strength ratio for either the torus shell or the torus support system is greater than 0.5, corrective measures should be promptly instituted to reduce the limiting strength ratio (s) to less than 0.5.
This action may consist of reassigning SRV setpoints, reducing the SRV setpoints, or other measures.
If it is determined that corrective measures are necessary for the facility, the submittal should describe proposed corrective measures, including the associated schedule for their completion."
These criteria were developed by the staff from a detailed review of the Monticello test data and with consideration for the uncer-tainties associated with the models used to predict the number of SRVs which will consecutively actuate. We conclude that these criteria will provide a reasonable estimate of the structural response of the torus and its support system to a multiple-consecutive SRV discharge event.
In addition, there are existing conservatisms in the models that are used to predict the number of valves that consecutively actuate and have not been altered.
Because the Monticello test data used for the individual plant assessments is proprietary, GE submitted a generic proprietary report that detailed the methods used to develop plant-specific hydrodynamic and structural correction factors (Ref. 3).
This report was subsequently amended (Refs. 4,5) to correct the analysis results and to incorporate comments by the staff pertaining to the SRV discharge line parameters used for certain analyses.
Hydrodynamic multipliers were generated from computer codes developed by GE to predict the peak pressure loads on the torus resulting from SRV discharge. These multipliers were generated to account for SRV discharge configurational differences between all of the individual plants and Monticello. The ratio of the maximum positive pressures was used to adjust the Monticello torus shell response. This factor was corrected for attenuation to produce a multiplier for the Monticello support column response. The staff has concluded that these computer models do not conservatively predict the peak pressure loads. However, because the results of these models are raticed, we find that the application of these models for this interim assessment is acceptable.
Similar correction factors were derived directly from the Monticello test data to account for structural variations between the facilities. These correction factors were developed and apolied in accordance with criteria 3 through 5 above. The resulting SRV-related stresses were combined with the seismic and dead weight loads, and the combined response was then compared to the structural acceptance criteria for the base case analysis in the Mark I STP (Ref. 6).
In a number of cases, plant-specific in-plant test results were used to supplement the generic evaluation technique. The results of relief valve discharge tests performed at Peach Bottom, Pilgrim, Hatch, and Millstone were compared with the results of applying the GE generic evaluation technique.
In all cases, the GE multipliers predicted loads on tne torus support system that were greater than the loads measured during the tests. However, in two cases, Millstone and Peach Bottom, the tests produced tcrus shell stresses greater than the stresses predicted from the GE evaluation technique.
The average midbay torus shell at Peach Bottom (measured by strain gauge 2) due to SRV discharge tests of valves "C" and "D" was 80 percent higher than the stress predicted by the GE multipliers.
The Peach Bottcm facility differs from the Monticello facility in that Peach Bottom has saddle supports and the Peach Bottom relief valve discharges are not located along the centerline of the torus. These effects create higher local bending stresses in the Peach Bottom torus shell than the stresses that occurred in Monticello.
The midbay torus shell stress in the Millstone facility (measured by strain gauge R22) for the test of valve "A" was 60 percent higher higher than the stress predicted by the GE multipliers. Examination of the stress components shows that the largest component is in the longitudinal direction, thus indicating a high proportion of bending stress which may not have been considered a primary stress for this evaluation. Nevertheless, we performed an evaluation of the torus shell stresses for those plants using the GE evaluation technique assuming an increase of 80 percent to account for the worst set of measured test data. The results of this evaluation showed that all shell stresses for these plants are within the limiting strength ratio of 0.5.
Prior to the assessment in this report, quencher SRV discharge devices had been installed in the Oyster Creek facility. All other facilities utilize ramshead discharge devices, with the exception of Monticello where quencher devices have recently been installed.
In May 1978, the Jersey Central Power and Light Company submitted the results of in-plant tests of the quencher discharge loads for the Oyster Creek f acili ty.
Specific tests performed to address multiple-consecutive SRV actuations show that the structural response is within the limits of the ASME Code.
SUMMARY
The results of the plant-specific interim assessments of multiple-consecutive SRV discharge loads on the torus and its support structures show that all plants are within the limiting strength ratio of 0.5 for the combination of weight, seismic, and SRV actuation loads, are in tenformance with the acceptance criteria specified for the Mark I Containment Short-Term Program, and are subject to the corrective action specified in Table 1.
In addition both Oyster Creek and Nine Mile Point meet the ASME Code limits for column stability and torus shell primary membrane stresses.
In some cases, we did not agree with certain plant-specific assump-tions used for the interim assessments.
In those cases where the differences were significant, the affected licensees were directed to revise and resubmit their analyses. For the remainder, we performed analyses that showed the differences not to be significant and the limiting strength ratio to be within the acceptarce criteria; thus, no action was taken.
In those cases where corrective action was necessary, that action has been completed or will be completed prior to plant startup. Millstone had committed to install column braces by December 1978 to reduce the limiting strength ratio to less than 0.5.
The Millstone licensee informed us on December 8,1978, that this modification has been completed.
The results of these interim assessments demonstrate that each of the operating Mark I BWR facilities, with the necessary corrective action, can accommodate a multiple-consecutive SRV discharge event with sufficient margin to assure the functional performance of the torus and its support st uctures.
Oi. this basis, we conclude that continued operation of these f acilities is acceptable until this issue is ultimately resolved as part of the Mark I Long-Term Program.
TABLE 1 MARI I MULTIPLE-CONSEClTTIVE SRV INTERIM ASSESSMENT Submittal Number Number MCA*
Corrective Plant Name Dates of SRVs in MCA*
Limiting Component Action Browns Ferry 1-3 06-06-78 11 7
Saddlas None 08-17-78 Brunswick 1-2 05-15-78 11 11 N/A None 09-29-78 CooIer Station 06-13-78 8
8 Column / torus weld None 06-29-78 Dresden 2-3 06-30-78 5
i Column base pin None Duane Arnold 07-25-78 5
2 Column /torusweld Stagger SRV setpoints FitzPatrick 07-07-78 11 2
Column Stagger SRV setpoints 07-31-78 08-18-78 08-25-78 09-28-78 11-14-78 Hatch 1-2 05-24-78 11 4 & 11 Outercolumn(Unit 1) None Torus shell (Unit 2) 07-27-78 08-01-78 08-07-78 Millstone 1 06-05-78 6
6 Column Strengthen 31ni tine, r
support columns 07-31-78 10-04-78 10-23-78 12-08-78 Monticello 05-24-78 8
8 Torus shell none Nine Mile Point 05-26-78 6
6 Column None 07-26-78
TAELE 1 (continued)
Sutraittal Number Number MCP Correct?ve Plant Name Dat es of SRVs in MCA*
Limiting Component Action __
Cyster Creek 06-27-78 5
2 Column None Peach Bottom 07-03-7t 11 8
Outer column None Pilgrim 06-05-78 4
4 column / torus weld None 07-21-78 Qumi Cities 1-2 06-30-78 5
1 column /torun weld None Vermont Yankee 05-24-78 4
1 Columr./ torus weld staeger say setrointo
- Multiple-consecutive ac *.uation (number of valves predicted to consecutively actuate)
REFERENCES 1.
U. S. Nuclear Regulatory Commission, " Mark I Containment Short Term Program Safety Evaluation Report," USNRC Report NUREG 0408, December 1977.
2.
General Electric Company, " Final Report, In-Plant Safety / Relief Valve Discharge Load Test - Monticello Plant,"
GE Proprietary Report NEDC-21581-P, August 1977.
3.
Droprietary letter from L. J. Sobon, GE, to V. Stello, Jr., NRC.
Subject:
" Mark I Containment Program Multiple Consecutive S/RV Actuation Evaluation, Task 7.1.3,"
July 21,1978.
4.
Letter from L. J. Sobon, GE, to V. Stello, Jr., NRC,
Subject:
" Mark I Containment Program Multiple Conse-cutive S/RV Actuation Evaluation," August 14, 1978.
5.
Letter from L. J. Sobon, GE, ;o C. I. Grimes, NRC,
Subject:
" Mark I Containment Program Multiple Conse-cutive S/RV Actuation Evaluation," November 13, 1978.
6.
NUTECH Company, " Description of Short Term Program, Plant Unique Torus Support Systems and Attached Pipinn Analysis,"
NUTECH Report MK1-02-012, Revision 2, June 1976.
Mr. I. R. Finfrock, Jr.
CC G. F. Trowbridge, Esquire Shaw, Pittman, Potts and Trowbridge 1800 M Street, N. W.
Washington, D. C.
20036 GPU Service Corporation ATTN:
Mr. E. G. Wallace Licensing Manager 260 Cherry Hill Road Parsippany, New Jersey 07054 Anthony Z. Roisman Natural Resources Defense Council 917 15th Street, N. W.
Washington, D. C.
20005 Steven P. Russo, Esquire 248 Washington Street P. O. Box 1060 Toms River, New Jersey 08753 Joseph W. Ferraro, Jr., Esquire Deputy Attorney General State of New Jersey Department of Law and Public Safety 1100 Raymond Boulevard Newark, New Jersey 07012 Ocean County Library Brick Township Branch 401 Chambers Bridge Road Brick Town, New Jersey 08723 e
e