ML20245E927
| ML20245E927 | |
| Person / Time | |
|---|---|
| Site: | Oconee  |
| Issue date: | 08/31/1988 |
| From: | Fineman C, Nalezny C EG&G IDAHO, INC. |
| To: | NRC |
| Shared Package | |
| ML16152A830 | List: |
| References | |
| CON-FIN-A-6492, RTR-NUREG-0737, RTR-NUREG-737, TASK-2.D.1, TASK-TM EGG-NTA-8227, NUDOCS 8902020192 | |
| Download: ML20245E927 (28) | |
Text
(
. 1 EGG-NTA-8227 i
TECHNICAL EVALUATION REPORT -
TMI ACTION--NUREG-0737 (II.D.1)
RELIEF AND SAFETY VALVE TESTING OCONEE - UNITS 1. 2 AND 3 DOCKET NOS. 50-269, 50-270, AND 50-287 C. P. Fineman C. L. Nalezny August 1988 Idaho National Engineering Laboratory EG&G Idaho, Inc.
. Idaho Falls, Idaho 83415 Prepared for the U.S. Nuclear Regulatory Comission Washington, D. C. 20555 Under DOE Contract No. DE-AC07-761001570 FIN No. A6492 I
4 ABSTRACT Light water reactors have experienced a namber of occurrences of improper performance of safety and relief valves installed in the primary coolant system. As a result, the authors of NUREG-0578 (TMI-2 Lessons Learned Task Force Status Report and Short-Term Reconnendations) and subsequently NUREG-0737 (Clarification of TMI Action Plan Requirements) recommended that programs be developed and completed which would reevaluate the functional performance capabilities of Pressurized Water Reactor (PWR) safety, relief, and block valves and which would verify the integrity of the piping systems for normal, transient,-and accident conditions. This report documents the review of these programs by the Nuclear Regulatory Commission (NRC) and their consultant, EG&G Idaho, Inc. $$ecifically,, this report documents the review of the Oconee Units 1, 2, and 3 Licensee response to the requirements of NUREG-0578 and NUREG-0737. This review found the Licenseo has not provided an acceptable response and, thus, has not reconfirmed that General Design criteria 14, 15, and 30 of Appendix A to 10 CFR 50 were met.
FIN No. A6492--Evaluation of OR Licensing Actions-NUREG-0737, !!.D.1 11 l
1 1
CONTENTS i
ABSTRACT .............................................................. ii 1.
INTRODUCTION ..................................................... I 1.1 Background ................................................. 1 1.2 General Design Criteria and NUREG Requirements ............. 1 2.
PWR OWNER'S GROUP RELIEF AND SAFETY VALVE PROGRAM ................4 3.
PLANT SPECIFIC SUBMITTAL .......................................... 6 4.
REVIEW AND EVALUATION ............................................ 7 4.1 Valves Tested .............................................. 7 4.2 Test Conditions ............................................ 8 4.3 Operability ................................................ 11 4.4 Piping and Support Evaluation .............................. 17 5.
EVALUATION
SUMMARY
..............'................................. 21-5.1 NUREG-0737 Items Fully Resolved ............................ 21 5.2 NUREG-0737 Items Not Fully Resolved ........................ 22 6.
REFERENCES ....................................................... 24 4
1 1
i I
l i
l l
iii
TECHNICAL EVALUATION rep 0RT TMI ACT!0N--NUREG-0737 (II.D.1)
RELIEF AND SAFETY VALVE TESTING OCONEE UNITS 1. 2. AND 3 DOCKET NO. 50-269. 50-270, AND 50-287
- 1. INTRODUCTION 1.1 Background .
Light water reactor experience has included a number of instances of improper performance of relief and safety valves installed in the primary coolant systems. There were instances of valves opening below set pressure,
, valves opening above set pressure, and valves failing to open or reseat.
From these past instances of improper valve performance, it is not known whether they occurred because of a limited qualification of the valve or because of basic unreliability of the valve design. It is known that the failure of a power operated relief valve (PORV) to resent was a significant contributor to the Three Mile Island (TMI-2) sequence of events. These facts led the task force which prepared NUREG-0578 (Reference 1) and, subsequently, HUREG-0737 (Reference 2) to recommend that programs be developed and executed which would reexamine the functional performance capabilities of Pressurized Water Reactor (PWR) safety, relief, and block valves and which would verify the integrity of the piping systems for normal, transient, and accident conditions. These programs were deemed necessary to reconfirm that the General Design Criteria 14, 15, and 30 of Appendix A to part 50 of the Code of Federal Regulations, 10 CFR, are indeed satisfied.
1.2 General Design Criteria and NUREG Requirements General Design Criteria 14, 15, and 30 require that (1) the reactor primary coolant pressure boundary be designed, fabricated, and tested so as to have extremely low probability of abnormal leakage, (2) the reactor coolant system and associated auxiliary, control, and protection systems be 1
l
o' ,
- 9
. =
designed with sufficient margin to assure that the design conditions are not i
exceeded during normal operation or anticipated transient events, and (3) the components which are part of the reactor coolant pressure boundary shall be constructed to the highest quality standards practical.
i To reconfirm the integrity of overpressure protection systems and thereby assure that the General Design Criter'.a are met, the NUREG-0578 position was issued as a requirement in a letter dated September 13, 1979, by the Division.of Licensing (DL), Office of Nuclear Reactor Regulation (NRR), to ALL OPERATING NUCLEAR POWER PLANTS. This requirement has since been incorporated as Item II.D.1 of NUREG-0737, Clarification of TMI Action Plan Requirements, which was issued for implementation on October 31, 1980.
As stated'in the NUREG reports, each pressurized water reactor Licensee or Applicant shall:
- 1. Conduct testing to qualify reactor coolant systehi ;ief ar.d safety valves under expected, operating conditions for design basis transients and accidents.
- 2. Determine valve expected operating conditions through the use of analyses of accidents and anticipated operational occurrences referenced in Regulatory Guide 1.70, Rev. 2.
- 3. Choose the single failures such that the dynamic forces on the safety and relief valves are maximized.
- 4. Use the highest test pressure predicted by conventional safety analysis procedures.
- 5. Include in the relief and safety valve qualification program the qualification of the associated control circuitry.
2 i
l:
. 6. Provide test data for Nuclear Regulatory Connission (NRC) staff
-review and evaluation, including criteria for success or failure of valves tested.
7.
Submit a correlation or other evidence to substantiate that the valves tested in a generic test program demonstrate the functionability of as-installed primary relief and saf'ety valves.
This correlation must show that the test conditions used are equivalent to expected operating and accident conditions as prescribed in the Final Safety Analysis Report (FSAR). The effect of as-built relief and safety valve discharge piping on valve operability must be considered.
- 8. Qualify the plant specific safety and relief valve piping and supports by comparing to test data and/or performing appropriate analysis.
i 3
_ _ - - _ _ _ _ _ _ _ _ _ - _ _ _ _ _ _ _ _ _ _ _ _ _ - = _ _ .
. 1
- 2. PWR 0WNER'S GROUP RELIEF AND SAFETY VALVE PROGRAM In response to the NUREG requirements previously listed, a group of '
utilities with PWRs requested the assistance of the Electric Power Research Institute (EPRI) in developing and implementing a generic test program for
~
pressurizer safety valves, power operated relief valves, block valves, and associated piping systems. Duke Power Co., the owner of Oconee, Units 1. 2, and 3, was one of the utilities sponsoring the EPRI Valve Test Program. The results of the program, which are contained in a series of reports, were transmitted to the NRC by Reference 3. The applicability of these reports f
is discussed melow. i EPRI developed a plan (Reference 4) for testing PWR safety, relief, and block valves under conditions which 'cound actual plant operating conditions. EPRI, through the valve manufacturers, identified the valves used in the overpressure protection systems of the participating utilities and representative valves were selecte'd for testing. These valves included a sufficient number of the variable characteristics so that their testing would adequately demonstrate the performance of the valvts used by utilities (Reference 5). EPRI, through the Nuclear Steam Supply System (NSSS) vendors, evaluated the FSARs of the participating utilities and arrived at a test matrix which bounded the plant transients for wt.ich over pressure protection would be required (Reference 6).
EPRI contracted with Babceck & Wilcox (88W) to produce a report on the inlet fluid conditions for pressurizer safety and relief valves in B&W designed plants (Reference 7). Since the Oconee units were designed by B&W, this report is relevant to this evaluation.
l Several test series were sponsored by EPRI. PORVs and block valves were tested at the Duke Power Company Marshall Steam Station located in Terrell, North Carolina. Additional PORV tests were conducted at the Wyle .
Laboratories Test Facility located in Norco, California. Safety relief '
r 4
___m______ _ _ . -_ _ . - . _ .. _ . . _ _ . . _ _ _ _ . _ . _ . _ . . . . . -
e #
valves (SRVs) were tested at the Combustion Engineering Company, Kressinger Development Laboratory, which is located in Windsor, Connecticut. The results of the relief and safety valve tests are reported in Reference 8.
The results of the block valve tests are reported in Reference 9.
The primary objective of the EPR1/C-E Valve Test Program was to test each of the various types of primary system safety valves used in PWRs for the full range of fluid conditions under which they may be required to operate. The conditions selected for test (based on analysis) were limited to steam, subcooled water, and steam to water transition. Additional 4 objectives were to_(1) obtain valve capacity data, (2) assess hydraulic and structural effects of associated piping on valve operability, and (3) obtain piping r'esponse data that could ultimately be used for verifying analytical piping models.
Transmittal of the test results meets the requirements of Item 6 of Section 1.2 to provide test data to the NRC.
5
. 1
- 3. PLANT SPECIFIC SUBMITTAL i i
An assessment of the adequacy of the overpressure protectdca system was submitted by Duke on July 1, 1982 (Reference 10). An assessment of the .
Pressurizer Safety and Relief Valve Piping was transmitted January 21, 1983 (Reference 11). A request for additional information (Reference 12) was submitted to Duke by the NRC on June 3, 1985. Duke responded to this request on October 1, 1985 (Reference 13). A second request for information was sent to Duke on April 22, 1987 (Reference 14). Duke Power's response to
, this second request for information was provided in three parts dated August 6, 1987 (Reference 15), January 15, 1988 (Reference 16), and February' 10, 1988 (Reference 17).
The response of the overpressure protection system to Anticipated Transients Without Scram (ATWS) and the operation of the system during feed and bleed decay heat removal are not considered in this review. Neither the Licensee nor the NRC have evaluated the performance of the system for these events.
I i
9 6
- 4. Rt.,1EW AND EVALUATION 4.1 Valves Tested Oconee 1, 2, and 3 utilize two safety valves, one PORV, and one block valve in the overpressure protection system of each unit. The safety valves are Dresser Model 31739A. The PORVs are Dresser Model 31533VX-30. The block valves are Westinghouse 03000GM88FNE000 (3GM88) 3 in, gate valves with Limitorque SB-00-15 actuators. Both the safety valves and the PORV/ block valve combination are mounted directly on a pressurizer nozzle.
The Dresser 31739A safety valves used at Oconee were one of the valves f
tested by EpRI. The valve was tested on a short inlet configuration which represents the Oconee installation without loop seals.
The PORVs installed at Oconee are the Dresser 31533VX-30-1 (dash 2 internals) design with a bore diameter of 1-3/32 in. The test vaive was a dash 2 design with e bore size of 1-5/16. The dash 2 design resulted from a need to improve the seat tightness and included modifications to the internals, the body, and the inlet flange. The body and flange n' modifications were not of a nature that would affect operability. The Oconee valves have incorporated the changes to the internals of the dash 2 design. The difference in pose diameter will only affect capacity and not operability. Also, the'PORVs at Oconee were modified to include heavier springs under the main and pilot disks to ensure closure at low pressure
.(less than 100 psig). At full system pressure the spring force is small relative to the force from the system pressure, so that valve operability is not effected. The test valve is, therefore, considered an adequate representation of the in-plant valve.
The Oconee 1, 2, and 3 block valves are Westinghouse 3GM88 3 in. motor operated gate valves with Limitorque 58-00-15 operators. This combination is identical to one tested by EPRI. The block valve was tested by EPRI in a l
f l
7
f 4
I horizontal configuration but is installed at Oconee 1, 2, and 3 in a q
vertical configuration. Duke Power stated (in Reference 13) that the only {
installation restriction given by the valve manufacturer was that the valve I not be installed with the stem below horizontal. This is to prevent .
lubriant from draining into the limit switch enclosure of the Limitorque operator. The test results are therefore considered to be applicable to the plant specific application.
Based on the above, the valves tested are considered to be applicable to the in-ple.t valves at Oconee 1, 2, and 3 and to have fulfilled that part of the criteria of items 1 and 7 as identified in Section 1.2 regarding applicability of test valves.
4.2 Test Conditions The valve inlet G G conditions that bound the overpressure transients for B&W designed pWR plants are identified in Reference 7. The transients considered in this report include FSAR, extended high pressure injection (HpI), and low temperature overpressurization events. Reference 7 addresses those transients listed in Regulatory Guide 1.70, Rev. 2, which potentially challenge the p0RV or safety valves in BAW plants. The conditions in the report that are applicable to Oconee 1, 2, and 3 are those identified for B&W 177-FA plants.
For the SRVs only steam discharge was calculated for FSAR type transients. The peak pressure was 2677 psia and the maximum pressurization rate was 175 psi /s. According to Reference 13, a maximum backpressure of 578 psia is developed at the SRV cutlet. Since Oconee 1, 2, and 3 do not have loop seals upstream of the SRVs, testing of the Dresser safety valves with the short inlet piping is applicable.
Seven steam tests with a short inlat pipe were performed with the 31739A valve which had a peak pressure of 2703 psia and a peak l
l pressurization rate of 333 psi /s. Tests with backpressure as high as l 866 psia were run. These tests used ring settings of -48 (upper), -40 and l -60 (raiddle), and +11 (lower). Duke power stated the safety valves at 8
1 s
Oconee 1, 2, and 3 will be adjusted to have ring settings of -48 (upper),
-50 (middle) and 43 (lower). The test ring setting are representative of I the plant ring settings. These conditions bound those expected at Oconee.
For er. tended HP1 events the safety valve will initially open on steam with transition to subcooled water calculated. A peak pressure'of 2515 psia j
was calculated with liquid temperatures ranging from 400 to 6450F. A peak {
liquid surge rate of 11,500 lbm/ min will occur. Pressurization rates from 0 to 65 psi /s are expected.
For the 31739A valve, testing included a steam to water transition test at 2489. psia and saturated conditions. Three water tests at pressures
- nging from 2389 to 2749 psia and with water temperatures of 414 to 6080F were run. The transition and water tests were run with pressurization rates from 1.8 to 3.2 psi /s. Although these represent the lower and of the range of pressurization rates calculated for B&W plants, they are adequate to demonstrate valve operability at Ocone's based on the test results discussed in Section 4.3. These conditions are sufficiently close to the conservatively selected bounding conditions, to adequately demonstrate valve performance.
For the PORV, FSAR events result only in steam discharge. Although Reference 7 indicated the PORV should be tested at a peak pressure higher than the opening set point, 2450 psia, the valve opens quickly enough that the increase.in pressure during the opening cycle is minimal., Additionally, the peak pressure listed in Reference 7 was based on an analysis in which the PORV was assumed to be inoperable. Testing with saturated steam at set pressure is, therefore, considered adequate. The Dresser PORV is a pilot operated valve and the backpressure developed at the outlet is of potential importance to valve operability. The ability of the valve to operate at backpressure at least as high as those expected in service should be demonstrated. The expected backpressure for the PORV was not explicitly given by Duke in Reference 13 but it was stated that the maximum PORV backpressure was bounded by the safety valve backpressure of 578 psia. This is reasonable because the PORV discharge pipe routing is similar to the 9
. + j 4
safety valves. The PORV rated flow, 119,000 lb/h, is 37% of the rated flow of the safety valve, 323,000 lbm/h. The 4 in:h discharge pipe of the PORV has approximately 44% the flow area of the 6 inch pipe for the safety valves. Testing of the valve (Reference 8) included numerous steam tests with opening pressures close to the Oconee set pressure and back pressures ~
as high as 760 psia, which adequately bounds the expected PORV conditions.
For extended HPI events the initial opening of the PORV will be orr steam but subcooled liquid could follow. HP1 events can, therefore, result in steam to water transition and water discharge (400 to 6500F) at a maximum pressure of 2500 psia (Reference 7). A steam to water transition test and liquid tests with temperatures ranging from 447 to 6470F and pressures of approximately 2500 psia were included in the test series. The tests were run using the same discharge pipe orifice which developed backpressure as high as 450 to 500 psia for the steam tests so that the expected backpressure was adequately represented. The Hp! events are, therefore, adequately represented by the tests.
The p0RV is used for low temperature overpressure protection. For low temperature overpressurization events the valve is required to open on steam at 565 psia. Reference 7 indicates transition and water flow will not occur at Oconee 1, 2, and 3 during low temperature overpressurization events.
Dr.ereing on steam is considered to be adequately represented by the full pressure steam tests discussed above. In addition, a PORV operability test is performed on steam at 60 psia prior to each plant start-up and, after maintenance, the PORV is tested at Wyle Labs at 65, 500, and 2200 psia.
Documentation supporting these statements by Duke was provided in Reference 15.
For the block valve only full pressure steam, 2435 psia, tests were '
performed (Reference 9). The block valve, however, is required to open and '
close over a range of steam and water conditions. Friction testing done by Westinghouse (Reference 18) on stellite coated specimens similar to those used in their gate valves showed that in the span of 21 test cycles (as applied in the tests at the Marshall steam station) the thrust required to 10
1 i
4 cycle the valve would be similar for steam and water tests. In addition, the required' torque to open or close the valve depends almost entirely on l the differential pressure across the valve disk and is rather insensitive to !
the momentum loading. Therefore, the torque requirement is nearly the same for water or steam and nearly independent of the flow. The full pressure
' steam tests, therefore, are adequate to demonstrate operability of the valve for low pressure steam and the required water conditions. l j
The test sequences and analyses described above, demonstrating that the test conditions bounded the conditions for the plant valves, verify that
, Items 2 and 4 of Section 1.2 were met, in that conditions for the operational occurrences were determined and the highest predicted pressures were chosen for the test. The part of Item 7, which requires showing that tne test conditions are equivalent to conditions prescribed in the FSAR, was also met.
4.3 Valve operability 1
As' discussed in the previous section the safety valves, PORVs, an'd block valves at Oconee 1, 2, and 3 are required to operata over a range of steam, steam-to-water transition, and subcooled water fluid conditions. The safety valves and PORVs were tested for the range of required conditions in the EPRI test program. .The block valves were tested with full pressure steam, the results of which apply also to liouid flow.
The safety valve ring settings at Oconee 1, 2, and 3 were determined-through a B&W Owners Group program. The program sponsored Continuum Dynamics, Inc. (COI) to determine optimum ring settings for six B&W plants, including Oconee 1, 2, and 3. CD1 used the valve dynamic simulation code COUPLE (References 19 and 20). COUPLE was validated against EPRI data. The optimum ring settings for Oconee were determined to be -48, -50, +8 (upper, middle, lower). The current ring settings are -48, -40, +8. Duke Power Co. I stated the ring settings would be changed at the next scheduled outage for each unit.
I 11 9
Oconee 1, 2, and 3 are B&W 177-FA plants which use Dresser 31739A safety valves. Three Mile Island, Unit 1 (TMI-1), is also a B&W 177-FA plant which uses the Dresser 31739A safety valve. In both TMI-1 and Oconee 1, 2, and 3 the safety valves are mounted directly on a pressurizer ,
nozzle. Based on Reference 7, the same inlet conditions are exper.ted at both TMI-1 and Oconee 1, 2, and 3. As discussed in Reference 21, the same ring settings will be used at both plants (the ring settings at TMI-1 were also determined by CDI using COUPLE). Reference 21 reviewed the EPRI test data and concluded the TMI-1 valves will operate properly with the -48, -50,
+8 ring settings for all expected valve inlet fluid conditions. Since Oconee 1, 2, and 3 use the same valves and ring settings, have the same expected valve inlet fluid conditions as TMI-1, and are installed in a similar manner, it is concluded the Oconee valves will also operate properly, i.e., stably, achieve rated lift and flow, and close with an acceptable blowdown.
Blowdowns for tne Dresser 31739A safety valve tested by EPRI ranged from 4.9% to 19.1% so that the measured blowdown generally exceeded the design blowdown of 5%. A B&W analysis (Reference 22) showed blowdown. up to 20% does not impede natural circulation due to hot leg voiding. Therefore, having the observed blowdown exceed the design blowdown is considered acceptable. .
The maximum bending moment applied to the discharge flange of the test valve was 20.144 ft-lb. Valve operability was not impaired by the application of this moments. The maximum expected moment at the outlet of the safety valves in the plant is 8,431 ft-lb. While the EPRI tests bound the plant condition as provided by Duke Power in Reference 13, the load combinations used by Duke Power (as provided in Reference 13) did not combine the valve and seismic loads. However, because of the margin between )
1 the maximum calculated bending moment without combining the valve and l seismic loads and the maximum moment in the EPRI tests (approximately 12,000 ft-lb), demonstration of valve operability with respect to the bending moment is considered adequate.
1 12
ml .
{
For t'he test performance to be a valid demonstration of plant safety I
valve stability, the test inlet piping must have a pressure difference at least as great as the plant. The plant valves are mounted directly on a pressurizer nozzle and thus have the minimum pressure. drop possible. The test piping included a venturi and a reducing flange and, therefore, a higher pressute difference. '
I During the 4140F water test the 31739A valve was stable but only achieved partial lift. The valve did not pass enough flow to prevent the test pressure from accumulating. Duke, in Reference 13, pointed out that the flow the 317M valve passed, in conjunction with the fact the plants have two valves each, was more than required for the steam line break (which results in the 4000F liquid flow case). The test valve passed I 9032 lbm/ min at a pressure of 2735 psia even though the valve was only open a small fraction of total lift. The maximum liquid surge rate on 4000F water is 6805 lbm/ min at 2515 psia.
Under conditions typical of the feedwater line break (FWLB), 2515. psia, and water flow at temperatures of 602 and 6400F, the required flows at 602 and 6400F are 10,400 lbm/ min and 11,520 lbm/ min, respectively. The 31739A test valve on the short inlet configuration passed 8633 lbm/ min (6030F ) I and 10,444 lbm/ min (6470F) at pressures of 2405 and 2346 psia, respectively. Oconee l' 2, and 3 have sufficient relief capacity at these conditions, however, because the the measured flows were at pressures itss than the 2515 psia used to determine the maximum surge rate, and two valves are installed at each plant.
A concern was raised that the water flow tests for the Dresser 31739A valve only included tests with pressurization rates less than 3.2 psi /s. f The B&W inlet conditions report (Reference 7) stated the upper bound for the pressuriza. sn rate was 65 psi /s. For each set of inlet conditions where the 31739A n1ve was required to pass water, the data above shows that the -
two valves in each plant are able to pass flows equal to or greater than the maximum calculated liquid surge rates. Therefore, even if the pressurization rate was 65 psi /s prior to valve opening for any of the required inlet conditions, the system pressure would stop increasing or 13 i l
. l
. =
I begin to decrease once the valve opened bec.*use, in each case, the safety valve flow would be equal to or greater than the pressurizer insurge. This indicates the 31739A valves at Oconee 1, 2, and 3 would be able to perform ;
their safety function prcperly even if they opened on water and the system pressurization rate was 65 psi /s.
Based on test results discussed above, and assuming Duke Pcwer Co.
changes the valve ring setting as dis:ussed above, demonstration of safety valve operability is considered adequate.
The Dresser PORV opened and closed on demand for all nonloop seal tests. Inspection of the valve after testing at tha Marshall Steam Station showed the bellows had several welds partially fa.il. The failure did not affect valve performance and the manufacturer concluded the failure did not have a potential impact on valve performance. The bellows was replaced and did not fail during any of b e additional test series.
A bending moment of 2125 ft-1b wts induced on the discharge flange of the test valve without impairing operability, The maximum bending moments calculated for the the Oconee PORVs were provided in Reference 17. The maximum bending moment calculated for the the Oconee 2 and 3 PORVs was 829 ft-1b. For Oconee 1,the maximum predicted bending moment was 2121 ft-lb. Thus, the expected plant bending moments are less than those applied in the tests. However, the load combinations used to determine the plant bending moments did not include a seismic load. If the seismic load was included, the bending moment for the Unit 1 PORV would probably exceed the test moment. For Units 2 and 3 the plant moments with a seismic load included would probably.not exceed the test moment because of the margin between the plant and test moments (approximately 1300 ft-1b). Therefore, the demonstration of operability of the PORVs for Units 2 and 3 under the '
maximum expected bending moment is considered adequate. Operability of the PORV for Unit i under the maximum expected bending moment has stut been shown -
satisfactorily.
14
,c q 4
(
As mentioned in'Section 4.2 Duke Power Co. provided data on PORV tests performed with low pressure steam (60 to 500 psia) at the plant and at Wyle I Labs. The valve opened and closed properly with these low pressure steam conditions.
The Oconee PORVs are pilot operated valves that use systes pressure to hold the disk tight against the seat. At one point Dresser Indu'stries recomended the block valve be closed at system pressures below 1000 psig to avoid steam wirecutting of the PORV disk and seat. A later recommendation by Oresser was to install heavier springs under the main and pilot disks to ensure closure at low pressures. Oconee modified its PORVs to include the heavier springs and, therefore, the PORVs will be available for cold overpressure protection.
The p0RVs at Oconee 1, 2, and 3 are set to close at 2415 psia. The maximum closing pressure in the EPRI tests was only 2360 psia. Duke Power stated that in recent tests at Wyle Labs with closing pressures of greater than 2415 psia the PORV closed without any problems. This data was provided for review in Reference 15. This demonstrates the ability of the Dresser PORV to close at pressures representative of the plant closing setpoint.
Based on the valve performance during EPRI tests, under the full range ;
of expected inlet conditions, and based on the modification using the I heavier springs, the demonstration of relief valve operability fer Units 2 and 3 is considered adequate. For the Unit 1 PORV, because the load combinations did not combine seismic and valve discharge loads and because the calculated bending moment without the seismic load approximately equals the maximum tested bending moment, operability of the Unit 1 PORV under the maximum expected bending moment was not demonstrated.
NUREG-0737, Item II.D.1, requires qualification of associated control circuitry as part of the safety / relief valve qualification. However, 4 meeting the licensing requirements of 10 CFR 50.49 for this electrical j equipment is considered satisfactory and specific testing per the NUREG-0737 requirement is not required. The Licensee disagrees with this interpretation of NUREG-0737 because the PORV and the PORV block valve are 15 l
w -__ . _ _ _
4 I
~
. 1 riot required to perform a safety fucction for any Oconee ' design basis j
accident. 'However, based on Duke Power's response to a question on control i
circuitry qualification in Reference 15, it is clear that operating procedures do not specifically prohibit use of the PORVs.in accident-mitigation. Failure to qualify either the PORV control circuitry or the PORV block valve for a harsh environment indicates that it'cannot be assured
- that the PORV can be closed or isolated via the block valve if the PORV opens arid fails to close. In the Licensee's response in Reference 15, reference was made to a Duke Power submittal dated October 1, 1986 on a NRC Safety System Functional Inspection Report dated August 1, 1986. Both the NRC report and the Licensee's response were reviewed and were found to be related specifically to the feed and bleed mode of cooling the primary system. NUREG-0737, Item II.D.1, is concerned with more than feed and bleed. Therefore, it is concluded that the Licensee has not met the PORV control circuitry qualification requirements of NUREG-0737, Item II.D.1.
The PORV block valve must be capable of closing over a range of steam
~
and water conditions. As described in Section 4.2, high pr. essure steam tests are adequate to bound operation over the full range of inlet conditions. The test valve / operator combination was the same as at the plant. The test valve was cycled successfully at full steam pressure with full flow. To open and close successfully, a torque of 175 ft-lb was required (Reference 9). For the Limitorque 58-00-15 operator tested, this required a torque switchI setting of 3.75, which is less than the maximum setting. The plant valve operator is set to close on limit so that the full stall torque of the operator is available to close the valve. Therefore, the tests are considered to adequately demonstrate acceptable valve operation.
The presentation above, demonstrating that the valves operated satisfactorily, verifies that the portion of Item 1 of Section 1.2 that
- requires conducting tests to qualify the valves and that part of Item 7 requiring that the effect of discharge piping on operability be considered -
were met for the safety valves and Unit 2 and 3 PORVs. However, Item 5 was 16
)
r- - _ - - - - - - _ _ - - - _ _ _ - - - - - - - - -
4 -
not met for the PORV control circuitry and that part of Item 7 requiring that the effect of discharge piping on operability be considered was not met-for the Unit 1 PORV.
4.4 Pipino and Support Evaluation In the piping and support evaluation, the safety / relief valve piping between the valve discharge flanges and the pressurizer relief tank and the pipe supports were analyzed for the requirements of the ANSI 1967 USA $s B31.1 and ANSI 1968 USAS B31.7 codes.
Duke Power provided the load combinations used in the Oconee analysis in Reference 13. Because the load combinations used in the Oconee analysis were not consistent with those suggested in the EPRI plant specific submittal guide (Reference 23), Duke Power was asked to discuss the basis for the load combinations used. Duke Power's response was provided ir.
Reference 16. In its response, Duke power agreed that the lead combinations used in the Oconee analysis were not consistent with those provided in the the EPRI guide. Instead, Duke Power stated the load combinations were specified in the analysis specification. From this response, however, it is not clear what is meant by the analysis specification. It is possible the analysis-specification refers to the system design basis or simply to the specification for the NUREG-0737, Item !!.D.1, analyses. For meeting the NUREG-0737 requirements, use of the load combinations in Reference 23 or the plant FSAR are the only acceptable bases. Because the load combinations used for Oconee are not consistent with the EPRI guide and it is not clear whether the load combinations are based on the Oconee FSAR, it is concluded that the load combinations used in the analysis do not meet the NUREG requirements.
4 The transient conditions analyzed included simultaneous lift of the safety valves on steam without PORV operation, lifting of the PORV alone on steam, and simultaneous lift of both safety valves on water following steam and water vsnting through the PORV. All cases were analyzed a,ssuming a conservative (t./ comparison to data in Reference 7) surge line flow rate of 220 lbm/s. Peak pressure and pressurization rate in the thermal-hydraulic
{
17
. j
_ _ _ _ - __ _ - - )
analysis were not controlled by input but were calculated as a result of the surge flow. This resulted in a peak pressure of 2515 psia and a i pressurization rate of 100 psi /s. These parameters are less than the peak pressure and pressurization rate of 2677 psia and 178 psi /s from Reference 7. However, they considered adequate because the peak pressure is conservative for the opening of the PORV and because the flow rat'e used in the safety valve analysis was representative of the f*#ow rates measured in the EPRI tests as discussed below. The forces generated from these conditions bound those from all other conditions expected at the plant.
The thermal-hydraulic analysis was performed with the program RELAPS/M001. The RELAPS verification work in Reference 24 showed it a suitable tool for calculation of valve discharge transients. Output from RELAP5 was used as input to the Control Data Corp.'s (CDC) computer program REPIPE to calculate the fluid forces on the system. Verification of REPIPE was reviewed as part of the D. C. Cook, Units 1 and 2, (Reference 25) submittal and the code was found to ade'quately calculate pipe forces due to valve discharge.
A RELAPS model for the safety valve and PORV piping from the valve discha m to the relief tank was developed. In the piping model, the key parameters of time step size, choked flow locations, and valve opening times were reviewed and found to be acceptable. A time step size of 1.0x10-5 s was used in the RELAPS analysis. To account for uncertainties in the safety valve flow rate, a conservative factor of 1.22 was included in the maximum rated valve mass flow rate for the SRVs. The conservative valve flow rates used in the analysis acceptably account for 10% ASME derating and potential error in the safety valve flow rate. The PORV rated flow was used even though PORV flows as great as 1.14% of rated were measured in the EPRI tests. This considered adequate, however, because the PORV does not have a '
loop seal at the valve inlet. Thus, the valve actuation loads are only a small part of the overall piping load. Valve opening times were 15 ms for -
the safety valves and 50 ms for the PORV. These opening times are representative of those measured in the EPRI tests.
)
l 18
__-____._____._______m__ _ _ _ _
l',
In Re'forence 16, the Licensee stated that the node size was approximately two feet. Based on'the studies in Reference 24, where the piping forces were undere3timated by 30% for a steam discharge transient when the node size was increased from 0.5 to 1.0 ft, the nodalization used
~
for the. 0conee. thermal-hydraulic analysis'is too coarse to ensure the maximum piping forces were conservatively calculated.
- The structural analysis was performed using Impell Corp.'s, computer program SUPERpIPE. SUPERPIPE was verified by comparing computer solutions from SUPERPIPE for a series of benchmark problems to that obtained from ASME sample problems or other computer programs such as PISOL1A, PISOL3A, NUPIPE, ADLPIPE, PIPESD, and EDSGAP. In addition, the benchmark problems in NUREG/CR-1677 were evaluated using SUPERPIPE and the results transmitted to tne NRC. This verification is considered adequate.
The key parameters of integration time step, cutoff frequencies, lumped mass spacing, and damping used in the' structural analysis were reviewed.
Time steps of 0.001 s were.used in the analysis and are considered adequate to capture an accurate total response. The Licensee stated the Oconee analysis was done using a direct integration method. The Licensee also stated that with this method a cutoff frequency is not input because high frequency effects are not truncated as with the modal superposition method.
Modal damping was stated to be 0.5% between 10 and 20 Hz. Because the direct integration analysis technique does not involve modal damping, this is not consistent with the Licensee's statement that direct integration was used. The Licensee also provided information on the lumped mass spacing used in the structural model that indicates the frequencies analyzed would I be limited to approximately 20 Hz. Therefore, the Oconee analysis did not adequately account for the effects of the higher frequency modes on the system structural response.
19
+
The results of the piping and support analysis identified a number of locations where support modifications were needed to relieve overstressed locations. To correct the piping overstress conditich,~tevaral supports were added or relocated, some variable or constant spri.1; supports were -
changed to rigid, and some welded pipe attachments were replaced' to reduce stresses to'within allowable limits. With the modifications to the supports, all stresses and loads are within their allowables. Duke Power stated all the necessary modifications to the support system were made.
The analysis discussed above, demonstrating that a bounding case was chosen for the piping configuration, verifies Itera 3 of Section 1.2 was met. However, because of inadequacies in the thermal-hydraulic end structural analyses, the analysis of the piping and support system does not !
verify that Item 8 of Section 1.2 was met.
4 l
I I
l l
l l
20
- 5. EVALUATION SUNiARY The Licensee for Oconee 1, 2, and 3 has not provided an acceptable response to the requirements of NUREG-0737, and thus has not reconfirmed that General Design Criteria 14, 15, and 30 of Appendix A to 10 CFR 50 were met with regard to the safety valves and PORV. The rationale for this conclusion is given below.
5.1 NUREG-0737 Items Fully Resolved Based on the following information provided by the Licensee, the requirements of Item II.D.1 of NUREG-0737 were partially met. This includes Items 1-4, 6, and Item 7 in Paragraph 1.2 for the safety valves and the Oconee 2 and 3 PORVs. Only that part of Item 7 that requires the valves tested in a generic test program demonstrate the operability the plant PORV was met for the Oconee 1 PORV.
The Licensee participated in the development and execution of an acceptable relief and safety valve test program to qualify the operability of prototypical valves and to demonstrate that their operation wou'd not invalidate the integrity of the associated equipment and piping. The subsequent tests were successfully completed under operating conditions which, by analysis, bound the most probable maximum forces expected from anticipated design basis events. The test results showed that the valves tested functioned correctly and safely for all steam and water discharge events specified in the test program that were applicable to Oconee 1, 2, and 3 and that the pressure boundary component design criteria were not Qxceeded. Analysis and review of both the test results and the Licensee justifications indicated the performance of the test valves and piping can be extended to the in-plant valves and piping.
Therefore, the prototypical tests and the successful performance of the valves and associated components demonstrated that this equipment was constructed in accordance with high quality standards, meeting General Design Criterion No. 30.
21
.c i
5.2 NUREG-0737 Items Not Fully Resolved Based on the information provided by the Licensee, the following requirements of Item !!.D.1 of NUREG-0737 as identified in Paragraph 1.2 were not met. ,
Item S: e The Lic~nsee's submittals have not demonstrated the PORV control circuits meet the qualification requirements of NUREG-0737, Item II.D.I. This is because neither the PORV control circuits or the PORV block valves are environmentally qualified. Therefore, it cannot be ensured that the PORV can be closed or isolated via the block valve if the 90RV. opens and fails to close in response to an accident or transient.
Item 7: Those parts of Item 7 that require consideration of the effect '
of ss-built discharge piping on PORV operability and that the test conditions used are equivalent to expected operating and accident conditions were not met. The maximum bending moment on the Oconee.1 PORV was approximately equal to the tested moment without including a seismic load. If the seismic load was included the maximum bending l
moment on the Oconee 1 PORV would exceed the test bending moment. I Thus, operability of the Oconee 1 PORV with the maximum expected applied moment could not be assured.
Item 8: Item 8, which requires qualification of the piping and supports, was not met. This is because the nodalization in the thermal-hydraulic analysis was too coarse to ensure the piping forces were conservatively calculated. Based on the information provided by the Licensee on the piping analysis, the analysis did not include the effects of the higher frequencies associated with valve discharge. -
Also, the load combinations did not combine the valve discharge and seismic loads. .
l k l
22 l
! l
. l c
1 i
Therefore, the Licensee has not demonstrated by testing and analysis I that the reactor primary coolant pressure boundary will have a low probability of abnormal leakage (General Design Criterion No. 14) and that the reactor primary coolant pressure boundary and its associated components (valves, piping, and supports) were designed with sufficient margin such
(
that design conditions are not exceeded during relief / safety valve events (General Design Criterion No. 15).
(
i I
0 i
23 1
i I
- 6. REFERENCES
- 1. TMI-Lessons learned Task Force Status Report and Short-Term Recommendations, NUREG-0578, July 1979,
- 2. ' Clarification of TMI Action Plan 'xecuirements, NUREG-0737, November ~
1980. *
- 3. R. C. Youngdahl 1tr. to H. D. Denton, Submittal of PWR Valve Test Report, EPRI NP-2623-SR, December 1982.
4.
EPRI 1980.
Plan for Performance Testino of PWR Safety and Relief Valves, July
- 5. EPRI PWR Safety and Relief Valve Test Program Valve Selection / Justification Report, EPRI NP-2292, December 1982.
- 6. EPRI PWR Safety and Relief Valve Test Pecoram Test Condition Justification Report, EPRI NP-2460, December 1982.
- 7. Valve Inlet Fluid Conditions for Pressurizer Safety and Relief Valves for B&W 177-FA and 205-FA Plants, EPRI NP-2352, December 1982.
- 8. EPRI PWR Safety and Relief Test Procram Safety and Relief Valve Test Report, EPRI NP-2628-SR, December 1982.
- 9. EPRI/ Marshall Electric Motor Operated Block Valve, EPRI NP-2514-LD, July 1982.
- 10. Letter W. 0. Parker, Duke Power Co., to H. R. Denton, NRC, " Response to NUREG-0737/NUREG-0660 TMI Action Plan Rgmts," July 1, 1982.
- 11. Letter H. B. Tucker, Duke Power Co., to H. R. Denton, NRC, "0R Submittal: TMI Action Plan Rgmt NUREG-0737 and NUREG-0660,"
January 21, 1983.
- 12. Letter J. F. Stolz, NRC, to H. B. Tucker, Duke Power Co., "NUREG-0737, Item II.D.1, Performance Testing of Relief and Safety Valves,"
June 3, 1985.
- 13. Letter H. 8. Tucker, Duke Power Co., to H. B. Denton, NRC, "Oconee Nuclear Station, Docket Nos. 50-269, -270, -287," October 1, 1985.
- 14. Letter B. J. Youngblood, NRC, to H. B. Tucker, Duke Power Co.,
"NUREG-0737, Item II.D.1, Performance Testing of Relief and Safety Valves," April 22, 1987.
- 15. Letter H. B. Tucker, Duke Power Co., to USNRC Document Control Desk, "Oconee Nuclear Station, Docket Nos. 50-269, -270 -287, Performance Testing of Relief and Safety Valves, Request for Additional Information," August 6, 1987.
24 i
- _ _ _ _ _ - _ _ _ _ - }
16.
Letter H. B. Tucker, Duke Power Co., to USSRC Document Control Desk, "0conee Nuclear Station, Docket Nos. 50-269 -270 -287, Performance Testing of Relief and Safety Valves, Request for Additional Information," January 15, 1988.
17 Letter H. 8. Tucker, Duke Power Co., to USNRC Document Control Desk, "Oconee Nuclear St? tion, Docket Nos. 50-269, -270 -287 Performance Testing of Relief nad Safety Valves, Request for Additional.
Information," February 10, 1988,
- 18. EPRI Summary Report: Westinghouse Gate Valve Closure Testino Program, Engineering Memorandum 5683, Revision 1, March 31, 1982.
- 19. " Coupled Valve Dynamic Model, for Isentropic, Two-Phase and Subcooled Discharge, Technical Description," Continuum Dynamics, Inc., Tech Note 83-6, prepared for participating PWR Utilities and EPRI, May 1983.
- 20. " Safety Valve Dynamic Analyses for Dresser Industries' 31739A and 31759A Valves," Continuum Dynamics, Inc., Report No. 83-4, Rev. 1, prepared for B&W, December 1983.
- 21. C. L. Nalezny, Supplement to Technical Evaluation Report TMI Action NUREG-0737 (11.D.1) Relief and Safety Valve Testino. Three Mile Island Unit 1 Docket No. 50-289, EGG-RST-6593 Supplement, June 1985.
- 22. Pressurizer Safety Valve,taximum Allowable 810wdown, B&W Report 77-113-5671-00, Asgust 1982.
- 23. EPRI PWR Safety and Relief Valve Test Program Guide for Application of Valve Test Program Results to Plant-Specific Evaluations. Revision 22 Interim Report, July 1982.
24 Application of RELAP5/M001 for Calculation of Safety and Relief Valve Discharge Piping Hydrodynamic Loads, EPRI-2479 December 1982.
I
- 25. C. Y. Yuan, C. L. Nalezny, and C. P. Fineman, Technical Evaluation Report. TM1 Action NUREG-0737 (!!.D.1). Donald C. Cook. Units land 2.
Docket Nos. 50-315 and 50-316. EGG-WTA-7881, October 1987.
25
- __ _