ML15239A255

Technical Review Rept on 850711 Failure of Main Steam Safety Valves (Mssvs) to Properly Reseat

ML15239A255
Person / Time
Site:	Oconee
Issue date:	06/17/1986
From:	Freeman R NRC OFFICE FOR ANALYSIS & EVALUATION OF OPERATIONAL DATA (AEOD)
To:
Shared Package
ML15239A254	List: ML15239A253 ML15239A255
References
REF-GTECI-A-49, REF-GTECI-RV, TASK-A-49, TASK-AE, TASK-OR, TASK-T605 AEOD-T605, NUDOCS 8607230058
Download: ML15239A255 (16)

Text

AEOD TECHNICAL REVIEW REPORT*

UNIT:

Oconee, Unit 2 TR REPORT NO.: AEOD/T 605 DOCKET NO:

50-270 DATE: June 17, 1986 LICENSEE:

Duke Power Company EVALUATOR/CONTACT: R. Freeman NSSS/AE:

Babcock and Wilcox, Duke Power Company/Bechtel

SUBJECT:

FAILURE OF MAIN STEAM SAFETY VALVES (MSSVs) TO PROPERLY RESEAT EVENT DATE: July 11, 1985

SUMMARY

On July 11, 1985 with Oconee Unit 2 at 94% power, while troubleshooting the electro-hydraulic control (EHC) system, a spurious signal was generated causing steam isolation to the turbine and resulting in a reactor trip.

Following the reactor trip, two main steam safety valves (MSSVs) that actuated to reduce main steam pressure failed to fully reseat at their proper setpoint.

To reseat the MSSVs, main steam pressure was reduced to approximately 990 psig by use of the turbine bypass valves (TBVs). The spurious signal generation was caused primarily by inadequate procedures on the use of EHC test points. A contributing factor to the spurious signal generation was the inherent sensiti vity of the EHC system to electrical noise. The failure of the MSSVs to reseat was caused by improper blowdown ring settings.

The primary safety concern was that MSSVs which fail to reseat could result in an uncontrolled steam generator blowdown. Such a transient could cause severe overcooling of the reactor coolant system (RCS) and could subject primary system components to high thermal and mechanical stresses. A combination of high ther mal and mechanical stresses on the reactor vessel has the potential to lead to significant crack growth in areas subjected to neutron induced embrittlement.

To address the above safety concern, AEOD conducted an evaluation to review the safety significance of the Oconee event and to identify any associated generic safety implications and their impact on plant safety.

The results of this investigation indicate that the safety implications of the Oconee event were minimal because no RCS overcooling occurred. Furthermore, even though the Oconee event and other operating experience shows that the poten tial for MSSVs failing to reseat and initiating an overcooling transient is higher for Babcock and Wilcox (B&W) designed plants than for other pressurized water reactors, the likelihood of such an overcooling to challenge the structural integrity of the pressure vessel is small. The corrective actions taken by the licensee were judged to be appropriate and no further action by this office is deemed necessary.

8607230058 860617 PDR ADOCK 05000270 S

PDR This document supports ongoing AEOD and NRC activities and does not represent the position or requirements of the responsible NRC program office.

-2 DISCUSSION

1. Event Description On July 11, 1985 with Oconee Unit 2 at 94% power, the licensee was conducting an investigation in order to determine the cause and to repair, as necessary, observed abnormal main steam header pressure oscillations. The initial investigation indicated that the EHC system was initiating main turbine control valve movement that caused the main steam header oscillations.

Troubleshooting within the EHC system cabinets was then conducted using test probes from a six channel analog recorder. When the fifth probe was inserted into the EHC system cabinet, a spurious output signal was generated. The EHC system interpreted the noise as a high rate of change in turbine speed and closed the main turbine control valves and the low pressure turbine intercept valves. This resulted in an immediate increase in RCS pressure and temperature due to the primary-to-secondary power mismatch caused by the isolation of main steam to the turbine. Approximately seven seconds after isolating main steam to the turbine, the reactor tripped on high RCS pressure. The maximum RCS pressure and average temperature attained during the event was 2319 psig and 600oF respectively. The pressurizer relief valves were not challenged.

Following the reactor trip, two MSSVs which actuated to reduce main steam pressure failed to fully reseat at their proper setpoint. To reseat the MSSVs, the licensee reduced main steam pressure to approximately 990 psig by use of the TBVs. To maintain pressurizer level, the licensee manually initiated high pressure injection for approximately three minutes to increase RCS water inventory. Aside from the two MSSVs failing to reseat, no other abnormal system responses were identified and no RCS overcooling occurred.

The licensee attributed the root cause of the spurious signal generation to a lack of documented guidance on the use of EHC test points. A contributing factor to the spurious signal generation was the inherent sensitivity of the EHC to electrical noise. The failure of the MSSVs to reseat was caused by improper blowdown ring settings. The licensee has instituted a program to rebuild four MSSVs on each refueling outage.

2. Pressurized Thermal Shock Considerations At first it was postulated that the most severe thermal shock a pressurized water reactor (PWR) pressure vessel would be required-to withstand would occur during a large-break loss-of-coolant accident (LOCA). In this situation, room temperature emergency core coolant would flood the reactor vessel within a few minutes and rapidly cool the vessel wall. The resulting temperature gradient across the pressure vessel wall would cause high thermal stresses. However, the presence of pressure stresses during such an event was not considered because during a large-break LOCA, the primary system pressure is expected to remain at low pressure. However, the occurrence of a non-LOCA-type event at the Rancho Seco nuclear plant on March 20, 1978 showed that during some types of overcooling transients the rapid cooldown could be accompanied by repressurization. As a result of the Rancho Seco event, the U.S. nuclear industry and the Nuclear Regulatory Commission are currently conducting evaluations of the capability of PWR pressure vessels to withstand severe thermal shocks without compromising their integrity.

-3 PWRs are susceptible to certain types of hypothetical accidents that under certain circumstances, including operation of the reactor beyond a critical time in its life, could result in failure of the pressure vessel as a result of extensive propagation of crack-like defects in the pressure vessel wall.

Events of particular concern are those that result in rapid cooling of the inner surface of the pressure vessel wall, and also involve substantial primary system pressure. This concern resulted in pressurized thermal shock (PTS) being designated as an unresolved safety issue and task action plan A-49 was established to evaluate the safety significance of PTS of PWR pressure vessels. On the basis of preliminary analyses, the NRC concluded that no event having significant probability of occurring could cause a PWR vessel to fail today or within the immediate future (Ref. 2).

However, as PWR vessels are irradiated, a few vessels could eventually become susceptible to PTS events.

To address the PTS issue, the NRC published a proposed rule (which was subsequently approved) that: (1) establishes a screening criterion on the reference temperature for nil-ductility transition, (2) requires licensees to accomplish reasonably practicable flux reductions to avoid exceeding the screening criterion, and (3) requires plants that cannot stay below the screening criterion to submit a plant-specific safety analysis to determine what, if any, modifications are necessary if continued operation beyond the screening limit is allowed (Ref. 3).

In addition, the NRC organized a PTS research project to help confirm the technical bases for the proposed PTS rule and to aid in the development of guidance for the licensee plant-specific PTS analyses, and the development of acceptance criteria for proposed corrective measures. The research project consists of PTS pilot analysis for three PWRs:

Oconee Unit 1, designed by B&W; Calvert Cliffs Unit 1, designed by Combustion Engineering; and H. B. Robinson Unit 2, designed by Westinghouse. A comparison of the three plant-specific analyses of PTS is reported in Reference 4.

A review of Reference 4 indicates that failure of one TBV or MSSV to close resulting in blowdown-of one steam generator was the most likely initiator for an overcooling event at Oconee. The probability of this initiator occurring at Oconee was 7.0 x 1OE-1 per reactor year. This value reflects the observation that at Oconee at least some of the MSSVs lift on each reactor trip. (In con trast, at both Calvert Cliffs and H. B. Robinson, the nuclear steam supply system (NSSS) is designed to preclude the lifting of MSSVs following a reactor trip.)

Although failure of a TBV or MSSV was the most likely initiator for an overcool ing event at Oconee, the probability for this accident sequence to cause a through-the-wall crack (TWC) in the pressure vessel is very low. This is because this accident sequence results in a relatively warm RCS temperature, approximately 328 0F for the Oconee plant. The reference temperature for nil ductility transition temperature for the Oconee pressure vessel at 32 effective full power years was determined to be approximately 270 0 F. The best estimate probability for this accident sequence to cause a TWC in the pressure vessel was determined to be 7.0 x 1OE-8 per reactor year.

A more severe case was analyzed where one or two TBVs or MSSVs were assumed to fail open. The failures were to occur in such a manner that both steam genera tors blow down. This accident sequence is a more severe transient which results in a relatively cold RCS temperature, approximately 198 0F for the Oconee plant.

-4 However, the probability for this initiator occurring at Oconee was determined to be 2.2 x 1OE-4 per reactor year, three orders of magnitude lower compared to overcooling transient initiated by the failure of one TBV or MSSV to close resulting in blowdown of one steam generator. The overall best estimate prob ability for this accident sequence to cause a TWC in the pressure vessel was determined to be 4.4 x 1OE-7 per reactor year. Thus, the analyses show that failure of a MSSV to reseat and cause an overcooling transient is highly probable, however, the actual safety significance of such an event from a PTS standpoint is very low.

The severity of the cooldown transient is highly dependent upon the available water in the steam generators to convert to steam. Thus, to reduce the severity of the transient, timely operator action must be taken to identify the main steam line with the failed open MSSV and isolate feedwater to the affected steam generator. An MSSV actuation and failure to reseat may not be immediately apparent to the operators because most plants do not provide remote MSSV position indication in the control room. This may increase the severity of the cooldown transient due to the delay in initiating feedwater isolation. On the other hand, because the B&W once-through steam generators (OTSGs) maintain a relatively low water inventory by design, the severity of the cooldown transient would be minimal, and not result in relatively low RCS temperatures provided appropriate operator action is taken. For PWRs which employ steam generators that maintain a relatively large inventory of water by design, (e.g., Westing house and Combustion Engineering), the NSSS is designed to preclude the lifting of MSSVS following most reactor trips. Thus the probability of an overcooling transient initiated by a failed open MSSV is less because the number of challenges to the values are reduced.

For Oconee, the worst case overcooling accident analyzed in the Updated Safety Analysis Report (USAR) is the double-ended rupture of a thirty-four inch main steam line from rated power conditions with offsite power available (Ref. 5).

However, the USAR main steam line break analysis is concerned with the reactor's return to criticality due to extremely rapid cooling of the RCS and does not bound the lower temperature which can be reached by the RCS and pressure vessel. Thus, the review of Oconee's USAR indicates that the analyses are not as applicable as the generic PTS analyses. This appears to be a general situation applicable to most PWRs in the United States for the reason stated in SECY-82-465:

Such analyses [Safety Analysis Reports (SARs) in support of licensee applications] tend not to be of much help in evaluations of PTS.

Many of the assumptions in such analyses were developed and accepted for licensing purposes without regard to PTS concerns. While SAR analyses appear to be appropriately conservative for calculations of reactor core thermal performance, PTS evaluations are most usefully performed using best estimate calculations of pressure and tempera ture behavior. In addition, some potential event sequences that are not generally analyzed in detail in SARs, because their consequences for core cooling are bounded by the design-basis event analyses, can be of greater significance for PTS evaluations.

-5

3. OPERATING EXPERIENCE: Overcooling Transients A review of operating experience found three events where one or more MSSVs failed to fully reseat following actuation and resulted in an excessive plant cooldown. On April 23, 1978, with Three Mile Island Unit 2 at 30% power, a reactor trip occurred due to a noise spike on the power range detector (Ref.

6).

When the reactor tripped, the turbine tripped causing a very rapid pres sure increase in the steam generators and resulted in lifting four of the six MSSVs on the train B steam generator and one MSSV on the train A steam generator.

The four train B MSSVs and the one train A MSSV failed to properly reseat. The operator started to manually cutback on feedwater demand; however, because the operator failed to recognize that the feed pump was in manual control and because the response of the feedwater control valves were slow, too much water was fed into the steam generators. The safety valves failing to fully reseat at their proper pressure coupled with overfeeding of the steam generators caused a rapid depressurization and cooldown of the RCS. The RCS average temperature dropped from 583oF to 464 0F in three minutes. The RCS cooldown caused the primary system water inventory to contract and pressurizer level dropped below the minimum indicated level range. Due to the rapid depressurization of the RCS, safety injection automatically occurred approximately one minute after the reactor trip. Pressurizer level was restored two minutes into the event as a result of safety injection and several of the train B MSSVs reseating which reduced the RCS cooldown rate. Approximately two and one half minutes into the event, feedwater flow to the steam generators was terminated, mitigating the cooldown transient.

On December 27, 1978, with Arkansas Unit 2 at 15% power, a MSSV lifted and failed to properly reseat due to steam dump valve failures (Ref. 7).

The failed MSSV caused the RCS temperature to drop 107 0F within 52 minutes, exceeding the 100'F/hour RCS cooldown limits. The relief valve was reseated approximately one hour after initial actuation.

On March 2, 1984, with Davis-Besse at 99% power, a main steam isolation valve (MSIV) inadvertently closed isolating the steam side of steam generator No. 2 (Ref. 8).

This caused an increase in feedwater to the other steam generator which resulted in overcooling the corresponding primary side loop in the RCS.

The reactor power level increased because of the negative moderator temperature coefficient and caused the reactor to trip on high power level approximately thirteen seconds after the MSIV closed. Following the reactor trip, one of the MSSVs which actuated to reduce main steam pressure failed to fully reseat. This caused an excessive RCS cooldown rate and by procedure, the effected steam generator was allowed to boil dry. The MSSV failure was caused by a failed cotter pin that secures the release nut in place at the top of the valve stem. The failed cotter pin allowed the release nut to spin down the valve stem while the valve was open. The release nut contacted the manual lifting device and prevented the valve from closing. A similar event occurred at St. Lucie Unit 2 was an MSSV stuck partially open following a reactor trip due to the release nut preventing the valve from closing (Ref. 9).

The partially failed open MSSV in the St. Lucie event, however, did not result in any RCS overcooling. Due to the similiarities of both events, IE issued an information notice notifying licensees of a potentially significant problem pertaining to failure of safety/relief valves caused by failed cotter pins (Ref. 10).

-6 An event occurred at a foreign facility involving a main steam safety relief valve (MSRV) which inadvertently opened during a plant startup. For approximately twenty-five minutes steam was released through the MSRV into the atmosphere causing a blowdown of the steam generator until it boiled dry. The average primary system temperature in the loop associated with the isolated steam generator dropped from 446 0 F to 284oF in twenty minutes. The opening of the MSRV was due to a maintenance error when the cable connectors to the solenoids of two pilot valves controlling the MSRV were inadvertently inter changed during the refueling outage. When power to the solenoids was switched on during the plant startup, the MSRV opened. When the power supply was disconnected, the MSRV remained open because the piston of one of the pilot valves stuck in its open position.

4. Operating Experience: Failures of MSRVs or MSSVs The operating experience was also reviewed to determine if problems with reseating of MSRVs and/or MSSVs was generic to all nuclear power plants.

Twenty-six events were reported between January 1980 and January 1986 in which MSSVs and/or MSRVs failed to properly reseat following a plant trip, a spurious actuation, or during valve testing. The majority of the events, sixteen out of the identified twenty-six, occurred at B&W designed plants. The operational data also shows that failure of one or more MSSVs to properly reseat has occurred repeatedly at the Oconee Nuclear Generating Station. A brief descrip tion of each of the twenty-six events can be found in the appendix at the end of this report.

Data searches conducted using the Nuclear Plant Reliability Data System (NPRDS) identified 165 safety/relief valve failures during the January 1984 to November 1985 reporting period. Approximately 46% of the reported valve failures were attributed to out of tolerance lift settings, 30% due to valve leakage, 9% to reseating, and the remaining 15% due to other miscellaneous valve related failures. Most of the safety/relief valve failures were attributed to setpoint drift and normal wear of the valve disk and seat. The cause of setpoint drift is not fully understood; however, review of the operating data indicates that the lack of a standard method for testing and setting safety/relief valve setpoints, coupled with the tight tolerance band on the lift setpoints, could be contributing factors for the large number of reported valve failures.

Setpoint drift could lead to premature lifting of safety/relief valves and increase the number of challenges to the valves. The problem of safety/relief valve setpoint drift is an identified generic problem and is currently under investigation by industry and vendor groups.

Table 1 presents the B&W reactor scram experience from 1980 through 1984. As shown in this table, most reactor scrams were caused by the high RCS pressure and loss of turbine reactor trips. In most cases, events involving reactor scrams on high RCS pressure would also involve lifting of several MSSVs to prevent overpressurization of the steam generators. Thus, nearly 35% of all the reactor scrams that occurred during the four year period would usually result in challenges to the MSSVs. In addition, most events involving reactor scrams on loss of turbine, in general, also challenge the MSSVs, particularly if the plant is operating at any significant power level.

This is primarily due to the limited turbine bypass capability, coupled with the relatively fast

-7 response of the B&W steam supply system, which results in steam pressure being initially controlled by the MSSVs following a turbine trip. This observation is further supported by the recent restart test program for Three Mile Island (TMI) Unit 1:

During the reactor trip on January 2, 1986, the MSSVs again con trolled steam pressure. Prior to the trip, television cameras and personnel were stationed on the intermediate building roof to monitor safety valve performance. The safety valves were again reseated by manually lowering the steam header setpoint after the third lift.

The licensee will monitor the video tapes and perform a technical evaluation. The NRC TMI-1 Restart Staff concurs with the licensee evaluation that the relifting and reseating of the safety valves on a trip does not preclude continued safe operation of the plant. The TMI-1 Restart Staff is not satisfied that the number of challenges to the safety valves has been minimized. A plan for corrective actions will be finalized by the end of the eddy current outage in April 1986.

Currently, under the B&W Design Generic Review Study, the B&W Owners Group is investigating ways to minimize the number of challenges to MSSVs and to improve their reliability.

Table 1 B&W Reactor Scrams - 1980 Through 1984 Loss of Loss of High RCS Power/

Power/Flow/

Other Year Turbine MFPs Pressure Pumps Imbalance Causes Total 1980 10 0

18 0

3 3

34 1981 13 3-14 6

0 9

45 1982 15 1

12 4

3 7

42 1983 12 3

14 1

6 8

44 1984

_4 4

7 1

1 3

20 Total 54 11 65 12 13 30 185 29%

6%

35%

6%

7%

17%

-8 Findings and Conclusions The results of this investigation indicate that the safety implications of the July 11, 1985 event were minimal because no RCS overcooling occurred. The spurious signal generation was caused primarily by inadequate procedures on the use of EHC test points. A contributing factor to the spurious signal generation was the inherent sensitivity of the EHC system to electrical noise. The failure of the MSSVs to reseat was caused by improper blowdown ring settings.

Review of available PTS analyses indicate that failure of one TBV or MSSV to close resulting in blowdown of one steam generator was the most likely initiator for an overcooling event at Oconee. The frequency for this accident sequence in the Oconee PTS analysis is very high because, at Oconee, at least some of the MSSVs lift on each reactor trip. However, the probability for this accident sequence to cause a TWC in the pressure vessel is very low.

The severity of a cooldown transient caused by a failed open MSSV is highly dependent on the available water inventory in the steam generator. Thus, to reduce the severity of the transient, timely operator action must be taken to identify the main steam line with the failed open MSSV and isolate feedwater to the affected steam generator. An MSSV actuation and failure to reseat may not be immediately apparent to the operators because most plants do not provide remote MSSV position indication. However, because the B&W OTSGs maintain a relatively low water inventory by design, the severity of the cooldown transient would be minimal and should not result in low RCS temperatures provided appro priate operator action is taken.

Review of Oconee's USAR indicates that the analyses are not as applicable as the generic PTS analyses with regard to PTS concerns. This appears to be a eneral situation applicable to most PWRs in the United States that such analyses Safety Analysis Reports (SARs) in support of licensee applications) were devel oped and accepted for licensing purposes without regard to PTS concerns. While SAR analyses appear to be appropriately conservative for calculations of core thermal performance, PTS evaluations are more usefully performed using best estimate calculations of pressure and temperature.

A review of operating experience had identified a number of events where one or more main steam safety/relief valves failed to properly reseat and ini tiated an overcooling transient. Based on the events reviewed, the severity of the cooldown transient is highly dependent upon appropriate operator action.

Operating experience also indicates that reseating problems with main steam safety/relief valves repeatedly occurred at the Oconee Nuclear Generating Sta tion and, in general, predominantly occurred at B&W facilities. This is because the B&W design does not preclude lifting of MSSVs following a reactor trip.

Thus, the number of demands is higher for B&W plants and the number of failures for MSSVs would be expected to be higher. Currently, under the B&W Design Generic Review Study, the B&W Owner's Group is investigating ways to minimize the number of challenges to MSSVs and to improve their reliability.

Data searches for safety/relief valve failures identified a significant number of reported failures. The majority of failures were due to out of tolerance lift setpoints caused by setpoint drift. The cause of setpoint drift is not fully understood; however, a review of the operational data indicates that the

-9 lack of a standard method for testing and setting safety/relief valves, coupled with the tight tolerance band on the lift setpoints required by most technical specifications, may be the major contributing factors for the large number of reported valve failures. Setpoint drift could lead to premature lifting of safety/relief valves and increases the number of challenges to the valves. The problem of setpoint drift of main steam safety/relief valves is an identified generic problem and is currently under investigation by the industry and vendor groups. The corrective actions taken by the licensee were judged to be appropriate and no further action by this office is deemed necessary.

-10 REFERENCES

1. Duke Power Company, Licensee Event Report 85-006, Oconee Unit 2, Docket No. 50-270, dated August 12, 1985.

2. Letter from W. J. Dircks, NRC, to the Commissioners, "Pressurized Thermal Shock," SECY-82-465, dated November 23, 1982.

3. Letter from W. J. Dircks, NRC, to the Commissioners, "Proposed Pressurized Thermal Shock Rule," SECY-83-288, dated July 15, 1983.

4. Memorandum from C. Johnson, NRC, to R. Woods, NRC,

Subject:

ORNL Letter Report on PTS, dated January 3, 1986.

5. Duke Power Company, Updated Safety Analysis Report, Oconee Unit 1, 2, and 3, Docket Nos. 50-269, 50-270, 50-289, dated July 1982.

6. Metropolitan Edison Company, Licensee Event Report 78-033, Three Mile Island Unit 2, Docket No. 50-320, dated May 8, 1978.

7. Arkansas Power and Light Company, Licensee Event Report 78-029, Arkansas Unit 2, Docket No. 50-368, dated January 25, 1979.

8. Toledo Edison Company, Licensee Event Report 84-003, Davis-Besse, Docket No. 50-346, dated March 30, 1984.

9. Florida Power and Light Company, Licensee Event Report 84-004, St. Lucie Unit 2, Docket No. 50-389, dated March 9, 1984.

10. U.S. Nuclear Regulatory Commission, Inspection and Enforcement Information Notice No. 84-33,

Subject:

Main Steam Safety Valve Failures Caused by Failed Cotter Pins, dated April 20, 1984.

11. Memorandum from W. F. Kane, NRC, to T. E. Murley, NRC,

Subject:

TMI-1 Restart:

NRC Hold Point for Startup After Completion of Restart Test Program, dated January 10, 1986.

Appendix A Failure of Main Steam Safety Valves to Reseat From 1980 to 1985 Plant Event Date Docket #

LER #

Description of Event Turkey Point 2 2/15/80 50-251 80-003 At hot shutdown, while testing a feedwater control valve, a rapid change in steam generator level caused lifting of one MSSV. Binding between the lifting gear and the release nut prevented the valve from reseating.

Pilgrim 1 10/7/80 50-293 80-069 At 90% power, a spurious actuation of a MSSV occurred resulting in an uncon trolled blowdown of the reactor coolant system. The nitrogen supply pressure to the pneumatic control solenoid was too high preventing the operator from closing the MSRV from the control panel.

Farley 2 5/25/81 50-364 81-002 During testing, an MSSV was declared inoperable when it opened and did not reseat. The cause of the failure could not be determined.

Davis-Besse 6/24/81 50-346 81-037 Following a reactor trip, an MSSV opened and failed to properly reseat. The MSSV had a bent spindle which prevented it from fully closing.

Salem 2 1/14/82 50-311 82-004 During an attempt to bring the plant to a stable operating condition following a rapid load rejection, an MSSV opened and failed to reseat after

the steam dump system was inadvertently disarmed. the lifting disk associated with the manual lifting arm had related approximately two full turns down the valve stem and was jamming valve travel.

Salem 2 7/6/82 50-311 82-059 During routine warm up of the main steam lines, an MSSV lifted below its setpoint and failed to reseat. The MSSV manual actuator spindle nut was found jammed between the spindle and the actuator cap, and the spindle threads were damaged.

San Onofre 2 7/19/82 50-361 82-063 At hot shutdown, a MSSV was declared inoperable when its setpoint was found below its required setting.

the valve also failed to seat properly and was gagged closed.

Calvert Cliffs 8/9/83 50-318 83-043 Following a reactor trip, a MSSV opened and failed to reseat. The valve failed to reseat due to a worn locking pin allowing rotation of the blowdown ring.

St. Lucie 2 2/9/84 50-389 84-004 During a unit trip, one MSSV opened and did not fully reclose. The MSSV failure was due to a broken cotter pin which allowed the release nut to spin down and contact the manual lifting device and prevented the valve from closing.

Davis-Besse 3/2/84 50-346 84-003 At 99% power, a main steam isolation valve went closed and caused a feed water and a reactor coolant system transient that led to a reactor trip. After the trip, one MSSV failed

to reseat and caused an excessive reactor coolant system cooldown rate and by procedure the steam generator was allowed to boil dry.

the MSSV failure was due to a broken cotter pin which allowed the release nut to spin down and contact the manual lifting device and prevented the valve from closing.

Crystal river 3 4/26/84 50-302 84-010 At 97% power, the "Y" non-nuclear instrumentation power supply failed causing erroneous signals to be sent to the integrated control system and resulted in a reactor trip. One atmospheric dump valve and several MSSVs failed to fully reseat following the reactor trip. The MSSV failures were due to setpoint drift.

Oconee 1 5/12/84 50-269 84-002 At 100% power, a reactor trip occurred due to the failure of the key selector switch and relay for the reactor coolant system hot leg tempera ture indication. This caused the integrated control system to reduce feedwater flow and the unit tripped on high reactor coolant system pressure. Two MSSVs did not reseat properly following the unit trip.

Trojan 9/26/84 50-344 84-017 At 50 power, a manual turbine runback was initiated to avoid an automatic trip of the operating main feedwater pump from low suction pressure.

Because the control system was in manual, the reactor coolant system temperature increased from the resulting mismatch between turbine

demand and reactor power. This caused an increase in steam generator pressure and lifted several MSSVs. One MSSV failed to reseat following the reduc tion in steam pressure. The MSSV failure was due to dirt or debris between the valve disk and guide area.

Oconee 1 12/2/84 50-269 84-006 At 43% power, a reactor trip occurred due to opening of the generator field breaker. During the trip recovery phase, two MSSVs did not reseat properly. The MSSV failures were attributed to setpoint drift.

Oconee 1 12/3/84 50-269 84-007 At 57% power, a reactor trip occurred due to a loss of main feedwater.

During the trip recovery phase, two MSSVs did not reseat properly. the cause of the MSSV failures is not fully known.

Oconee 1 1/22/85 50-269 85-002 At 100% power, a reactor trip occurred due to a inadvertent closure of all six intercept valves to the high pressure turbine. During the trip recovery two MSSVs failed to reseat properly. The cause of the MSSV failures is not fully known.

Oconee 4/11/85 50-269 85-005 At 100% power, a reactor trip occurred due to the closure of the low pressure turbine intercept valves and the main turbine control valves.

During the trip recovery two MSSVs failed to reseat properly. The cause of the MSSV failures is not fully known.

Oconee 1 4/11/85 50-269 85-006 At 17% power, a reactor trip occurred due to a decrease in feedwater flow caused by steam pressure oscilla tions. After the trip, two MSSVs did not reseat properly.

Duane Arnold 4/11/85 50-331 85-017 During a refueling outage, six MSSVs and two safety valves were removed and tested prior to maintenance on the valves. While testing the setpoint pressure for valve lifting, four of the valves were found to lift at a lower pressure than the minimum allowable pressure. One other valve did not reseat after initial actuation.

Oconee 1 4/25/85 50-269 85-007 At 94% power, a reactor trip occurred due to a loss of main feedwater when both main feed pumps tripped on high discharge pressure. During the trip recovery, four MSSVs failed to reseat properly.

Oconee 2 4/26/85 50-270 85-005 At 75% power, a reactor trip occurred due to an inadvertent closure of the low pressure turbine intercept valves and the main steam turbine control valves. During the trip recovery, main steam pressure had to be dropped in order to reseat a partially stuck open MSSV.

Oconee 2 7/11/85 50-270 85-006 At 94% power, a reactor trip occurred due to an inadvertent closure of the low pressure turbine intercept valves and the main steam turbine control valves. During the trip recovery, main steam pressure had to be reduced to reseat two partially stuck open MSSVs.

Crystal River 3 10/26/85 50-302 85-023 At 95% power, a manual turbine/reactor trip occurred due to erroneous control board indication caused by an inverter failure. After the trip, one MSSV failed to reseat properly and required lowering of the main steam pressure to 990 psig to assist in reseating the valve.

Crystal River 3 12/3/85 50-302 85-028 At 93% power, a fault occurred on the B 6900 volt unit auxiliary bus resulting in the loss of the D and B reactor coolant pumps. A reactor trip occurred due to the loss of coolant flow. After the trip, one MSSV failed to reseat properly and required lowering of the main steam pressure to assist in reseating the valve.

Crystal River 3 1/1/86 50-302 86-001 At 93% power, a reactor trip occurred caused by a reactor power to reactor coolant system flow imbalance. After the trip, three MSSVs failed to properly reseat.

Three Mile Island 1 1/4/86 50-289 86-002 At 20% power, a reactor trio occurred due to high moisture separator drain tank level which caused a turbine trip. during the trip recovery, main steam header pressure had to be reduced to reseat a partially stuck open MSSV.