ML20245B510

Pressurizer Code Safety Valve Reliability

ML20245B510
Person / Time
Site:	Diablo Canyon
Issue date:	03/27/1987
From:	NRC OFFICE FOR ANALYSIS & EVALUATION OF OPERATIONAL DATA (AEOD)
To:
Shared Package
ML20245A607	List: ML20245A605 ML20245B510
References
TASK-AE, TASK-T704 AEOD-T704, NUDOCS 8904260141
Download: ML20245B510 (7)

Text

- _ _ _ _ _ _ - _ - _ . ____ -

' AE00 TECHNICAL REVIEW REPORT 2 PLANT: Diablo Canyon 1 TR REPORT NO.: AE0D/T 704 DOCKET NO : 50/275 DATE: March 27, 1987 LICENSEE: Pacific Gas and Electric Company EVALUATOR /CDNTACT: M. Wegner NSSS/AE: Westinghouse / Pacific Gas and Electric Company

SUBJECT:

PRESSURIZER CODE SAFETY VALVE RELIABILITY

SUMMARY

Upon testing their pressurizer Code 2 safety valves with the unit at hot standby, Diablo Canyon 1 personnel found their lift setpoints to be above their technical specifications (TS) limits. The tec' method was questioned and its proper use was determined. The valves were correctly reset. A data search indicated that a total of 34 pressurizer Code safety valves at 17 plants had setpoint drift, leakage, misadjusted ring settings, or maintenance / instal-lation problems since January 1,1983. These problems could lead to inadver-tent reactor scrams, overpressurization of the reactor coolant system (RCS), or degradation of the reactor coolant pressure boundary. These problems appear to be generic to all safety / relief valves. It is suggested that AE00 initiate a case study to determine the extent of the problems and to assess the adequacy of present efforts toward increasing safety valve reliability.

DISCUSSION

1. Description of Deficiency On August 8, 1986 at Diablo Canyon 1 with the unit in hot standby at 525 F and 2106 psig, a surveillance test to determine lift setpoint of the pressurizer Code safety valves was performed (Reference 1). The plant TS require a lift setpoint of 248511% psig. The initial test results indi-cated that safety valve RCS-1-8010A lifted at 2747.8 psig, 9.5% above the TS limit of 2485+1% psig; safety valve RCS-1-8010B lifted at 3028.0 psig, 20.6% above the TS limit; and safety valve RCS-1-8010C lifted at 2661.0 psig, 6.0% above the TS limit.

The test method used valve stem displacement as indicated by hydraulic pressure on the test rig to predict the lift setpoint. Because the upstream loop seal was not drained, the licensee claimed the test was inaccurate. The licensee stated that, "when the loop was clearing during the first test, steam entered the valve seat area, and a larger valve stem displacement occurred," resulting in an inaccurately high predicted lift setpoint.

I 1 This document supports ongoing AE0D and NRC activities and does not represent  !

2 the position or requirements of the responsible NRC program office.

American Society of Mechanical Engineers Boiler and Pressure Vescel Code 8904260141 870327 PDR ADOCK 05000275 S PDC l

Test results for the second test indicated that safety valve RCS-1-8010A lifted at 2550 psig, 1.8% above the TS limit; safety valve RCS-1-8010B lifted at 2494.0 psig, which is within TS limits; and safety valve RCS-1-80100 lifted at 2337.6 psig, 0.9% below the TS limit of 2485-1%

psig.

The licensee attributed the root cause of the pressurizer Code safety valve lift setpoint being outside the TS limits to setpoint drift.

A search previously conducted to support the issuance of Inspection and Enforcement (IE) information notice (IN) 86-92 showed that five events involving pressurizer safety valve problems occurred with the reactor in power operation. These events, together with two events which occurred quite recently, are as follows.

Sequoyah 1 (Licensee Event Report (LER)84-031, Reference 2)

On May 5, 1984, with the plant at 547 F and 2235 psig, pressurizer relief valve 1-SRV-68-565 lifted. The technical specifications requirement for the setpoint is 248511% psig. The premature lifting of the valve was attributed to leaking according to the following rationale: The spring force calculations to maintain the relief valve closed were based on the cross sectional area of the seat with no leakage; with leakage, the area was increased allowing the valve to lift at a lower pressure.

Sequoyah 2 (LER 84-013, Reference 3)

On August 20, 1984, with the plant at 578 F and 2235 psig and the pressurizer relief valve 2-SRV-68-653 leaking slightly, one of the pressurizer relief tank rupture discs ruptured prematurely. When the relief valve was tested, it exhibited gross leakage at 2300 psig and its setpoint could not be determined.

Sequoyah 2 (LER 84-020, Reference 4)

On December 16, 1984, with the plant at 540 F and 1970 psig, pressur-izer safety valve 2-PCV-68-563 was found to be leaking. Actions taken to reseat the valve resulted in a reactor scram ar.d safety injection.

Byron 1 (LER 86/023, Reference 5)

On July 16, 1986, while in startup at 520 F and 1700 psig, excessive leakage of pressurizer safety valve 1RY80100 prevented power ascr.n-sion above 1750 psig. Leakage occurred because the disc insert had not been installed. The valve which had the disc insert missing had been installed on the pressurizer because it was stored next to a reworked valve. Paper work for the defective valve had been signed off following the rework of the other valve, resulting in the release and subsequent installation of the defective valve.

McGuire 1 (LER 86-016, Reference 6)

On September 2, 1986, with the unit in hot standby at 530 F and 2375 psig, pressurizer relief valve INC-1 lifted and did not resrat until i

RCS pressure reached 1800 psig. Relief setpoint for the valve is 248511% psig per technical specifications and reseating shculd have occurred when RCS pressure decreased about 10%. The satpoint spring was not properly seated, causing the premature lifting. The extended l

l

{

_ - - - - - - - - _ . - - - - - a

3 blowdown was caused by the' incorrect setting of the blowdown ring.

Valve-1NC-2 also had its blowdown ring incorrectly set.

'ANO 2 (LER 86/012, Reference 7).

On 09/25/86, during a post-scram transient with RCS pressure 2400 psia, acoustic monitoring devices indicated flow from pressurizer safety valve 2PSV-4633. Both 2PSV-4633 and 2PSV-4634 were tested to-determine the lift setpoint which were 2425 psia and 2450 psia respectively. Required setpoint was 250011% psia..

Calvert Cliffs 1 (LER not received yet, Reference 8)

On 03/11/87, during power operations, the pressurizer quench tank rupture disc ruptured. Indications were. that one or both of the safety valves were leaking. Both were tested for the lift setpoint.

RC-200, with a required setpoint of 250011% psia, lifted at 2470 psia. RC-201, with a lift setpoint of 256511% psia, lifted at 2705 psia.

The same search also identified pressurizer safety valve problems discov-cred during testing in situ, on the bench, or in outside laboratories.

These problens are gWen in Table 1.

2. Corrective Actions by Licensees Pacific Gas and Electric Company, the licensee for Diablo Canyon 1, sent a spare pressurizer safety valve to Wyle Laboratory to verify the accuracy i of their testing method and, as a result, decided to reset the pressurizer Code safety valve RCS-1-8010-C down to be within TS limits. Procedures are being changed by the licensee to require the loop seal to be drained prior to setpoint testing, to perform the tests immediately after the draining, and to use the updated information from the Wyle tests to calculate lift-setpoint.

Commonwealth Edison Company, the licensee for Byron 1, initiated a variety of corrective actions aimed at strengthening their maintenance of pressur-

-izer Code safety valves by improving control over equipment being reworked, closing out work requests expeditiously, upgrading quality assurance control of hold tags, properly tagging items in stores, and designating an area for contaminated equipment in stores.

Duke Power Company, The licensee for McGuire 1, replaced pressurizer safety valve 1NC-1, adjusted pressurizer safety valve INC-2, and disqual-ified all_ employee training and qualification records for employees previously qualified to work on pressurizer safety valves, and committed to retrain all personnel involved in the testing of pressurizer safety valves. Additional corrective actions planned were to send valve INC-1 to t

Wyle Laboratory for further testing; to establish the correlation, if any, between hot and cold testing and incorporate the results into the mainte-nance and testing procedures; to record the as-found and as-left ring settings; and to inspect the Unit 2 pressurizer safety valves.

The remaining licensees reported the removal and replacement of defective valves with spares, or the rework and replacement of the defective valves and retesting. The licensee for Ft. Calhoun reported only repeated testing. Of the series of tests, only the initial test was outside TS

' 4 limits. For the 1985 event at Calverc Cliffs, Baltimore Gas and Electric Company reported verifying the accuracy of the "hydroset" test rig.

3. Safety Significance The pressurizer Code safety valves are designed to relieve pressure from the pressurizer, hence the RCS, in the cue of an overpressure event to prevent the rupturing of the RCS. Such a postulated rupture of the RCS-would lead to a loss of reactor coolantiwhich if unmitigated could lead-to a core disruption, or core melt. In the case of a high lift setpoint, both the temperature and pressure of the RCS would be elevated over and above those associated with the TS limits. In the case of a low lift setpoint, the pressurizer Code safety valve lifting with the reactor at power would precipitate a transient. Leaking or stuck open pressurizer Code safety valves constitute a breach of the pressure boundary and may lead to a small-break loss-of-coolant accident (LOCA).

4. Generic Implications Of the 34 valves involved in the incidents investigated (including the Diablo Canyon svent), 24 are Crosby valves (Crosby HB-86-BP). The five Dresser valves involved in the remaining events are physically and mechan-ically similar to the Crosby valves. The main steam safety valves identi-fied in IE information notice 86-56 are also physically and mechanically similar to the Crosby valves and exhibit many of the same problems. The Target Rock safety / relief valves used on boiling water reactors (BWRs) are neither physically nor mechanically similar, but exhibit similar problems.

(See Rderence 24)

FINDINGS 1.

At Diablo Canyon, the root cause of the pressurizer Code safety valve lift setpoint being outside TS limits was attributed to setpoint drift.

2.

A data search indicated that a total of 34 pressurizer Code safety valves at 17 plants had problems involving setpoint drift, leakage, misadjusted ring settings, or maintenance / installation deficiencies.

3.

The observed operational problems involving setpoint drift, leakage, misadjusted ring settings, or maintenance / installation deficiencies are generic to all pressurizer Code safety valves, main steam safety valves, and dual action (safety / relief) valves.

4. A leaking or stuck open pressurizer Code safety valve constitutes a breach of RCS and may lead to a small break LOCA. Pressurizer Code safety valves with significant setpoint drift problems can degrade safety margins, and precipitate a transient.

CONCLUSIONS l

A significant number of pressurizer Code safety valves have exhibited problems involving setpoint drift, leakage, misadjusted ring settings, and maintenance /

installation deficiencies. These problems were identified either during power

)

I I

- _ _ _ _ _ - - _ _ _ _ _ _ _ - _ _ _ _ - _ - - _ _ _ _ _ _ - _ _ a

operation or during testing. These problems appear to be generic to all pressurizer Code safety valves, main steam safety valves, and safety / relief valves.

Because of the generic implications of the problems associated with the pressurizer Code safety valves, their repetitive occurrences, and their safety significance, AE00 should initiate a case study of the operating experience involving all pressurizer and main steam safety valves and safety / relief valves over an extended period of time. The study should be directed to assess the true extent of the problems, the root cause, and the adequacy of corrective (

actions.

l REFERENCES

1. LER 50/275-86/018

2. LER 50/327-84/031

3. LER 50/328-84/013 .

4. LER 50/328-84/020

5. LER 50/454-86/023

6. LER 50/369-86/016

7. LER 50/368-86/012

8. 10 CFR 50.72 1 hour1.157407e-5 days <br />2.777778e-4 hours <br />1.653439e-6 weeks <br />3.805e-7 months <br /> notification, Docket 50/317

9. LER 50/313-83/001

10. LER 50/327-83/001

11. LER 50/285-83/001

12. LER 50/272-83/008

13. LER 50/311-83/002

14. LER 50/255-83/064

15. LER 50/336-83/027

16. LER 50/328-83/127

17. LER 50/328-83/165

18. LER 50/395-84/019

19. LER 50/029-84/011

20. LER 50/334-84/019

21. LER 50/318-85/010

22. LER 50/029-85/004

23. LER 50/389-86/007

24. List of Relevant IE Information Notices IN 82-41 Failure of Safety / Relief Valves to Open at a BWR IN 83-39 Failure of Safety / Relief Valves to Open at a BWR - Interim Report IN 83-82 Failure of Safety / Relief Valves to Open at a BWR - Final I Report J IN 86-05 Main Steam Safety Valve Test Failures and Ring Setting Adjustments -

IN 86-05 Supplement 1 IN 86-12 Target Rock Two-Stage SRV Setpoint Drift IN 86-56 Reliability of Main Steam Safety Valves l IN 86-92 Pressurizer Safety Valve Reliability I

{

l I

d

. - Table 1. PRESSURIZER SAFETY VALVE PROBLEMS DISCOVERED DURING TESTING Plant Valve Mfr Comments Valve No. Setpoint* Lift Setpoint*

ANO 1 (Reference 9) Crosby Testing at Wyle Labs PSV-1001 250011% 2470 Sequoyah 1 (Reference 10) Crosby Testing 1-RV-68-565 248511% 2380 Ft. Calhoun (Reference 11) Crosby Testing at Lab; repetitive testing until several lifts were within TS limits; no rework.

RC-141 254511% 2577 Salem 1 (Reference 12) Crosby Testing at Wyle Labs; all valves were leaking.

IPR 3 2485 1% 2564 IPR 4 248511% 2512 1PR5 2484i1% 2546 Salem 2 (Reference 13) Crosby Testing at Wyle Labs; all valves were leaking.

2PR3 248511% 2416 2PR4 248511% 2400 2PR5 248511% 2454 Palisades (Reference 14) Crosby Testing RV-1039 2539-2591 2680 Millstone 2 (Reference 15) Unknown Testing at Wyle Labs 2-RC-201 Unknown <TS j

Sequoyah 2 (Reference 16) Crosby Testing; all valves were leaking.

2-RV-68-563 248511% <TS 2-RV-68-564 248511% <TS I

2-RV-68-565 248511% <TS '

Sequoyah 2 (Reference 17) Crosby Testing in situ; lift point indeterminate due to leaking caused l by loss of loop seal. l 2-VLV-68-563 2485t1% Indeterminate  !

Summer (Reference 18) Crosby Testing; XVR-8010-A was leaking XVR-8010-A Unknown Unknown XVR-8010-C 248511% 2590 Yankee-Rowe (Reference 19) Dresser Testing; Valve setting was wrong when installed SV-182 2560+0-3% 2567 1

Both setpoint and lift setpoint were measured using consistent units; e. g.,

psig or psia.

l t

L,______m _ . _ . . _ _ - - - - - - - - - - - - - - - - - - - - - - " '

F, ,.

Table 1. (Continued)

Plsat Valve Mfr Comments Valve No. Setpoint* Lift Setpoint*

Beaver Valley 1 (Ref. 20)

Unknown Testing at Wyle Labs B 248511% +2%

Calvert Cliffs 2 (Ref. 21) Dresser Testing in situ; hydroset test rig 2-RC-201-RV calibration was verified.

256511% 2619 Yankee-Rowe (Reference 22) Dresser Testing at Lab PR-SV-181 2485+0-3 2603 St. Lucie 2 (Reference 23) Crosby Testing on bench; V-1201 had a hole in the bellows and the internals were corroded; V-1202 was leaking so badly that its lift setpoint could not be determined V-1200 2515fl% 2554 V-1201 250011% 2893 V-1202 251511% Indeterminate i

• Both setpoint and lift setpoint were measured with consistent units; e. g.,

psig or psia.

1 l

I

___ _ - _ _ - _ - - - - - - - - - - - - - - - - - - - - - - - - - - - ---- - - - - - - - - - - - - - -- - - - - - - - - --- - --