ML18065A933

Submits Comments on Draft 1982-83 Accident Sequence Precursor Rept

ML18065A933
Person / Time
Site:	Palisades
Issue date:	09/26/1996
From:	Bordine T CONSUMERS ENERGY CO. (FORMERLY CONSUMERS POWER CO.)
To:	NRC OFFICE OF INFORMATION RESOURCES MANAGEMENT (IRM)
References
NUDOCS 9610040126
Download: ML18065A933 (7)

Text

consumers Power Palisades Nuclear Plant: 27780 Blue Star Memorial Highway, Covert, Ml 49043 September 26, 1996 U.S. Nuclear Regulatory Commission Document Control Desk Washington, DC 20555 DOCKET 50-255 - LICENSE DPR PALISADES PLANT Thomas C. Bordlne Manager. licensing COMMENTS ON DRAFT 1982-83 ACCIDENT SEQUENCE PRECURSOR REPORT In a letter dated April 30, 1996, the NRC issued draft excerpts from its 1982-~3 Accident Sequence Precursor Report for comment. Under the Accident Sequence Precursor (ASP) review program, the NRC evaluates plant events and conditions to establish the potential increase in risk that resulted. Two Palisades events from 1982 were identified for possible inclusion in the report. The events were:

1)

Licensee Event Report (LER) No. 255/82-002.

On January 6, 1982, the normal monthly surveillance test disclosed that the controllers for Auxiliary Feedwater control valves (CV-0727 and CV-0749) were not properly adjusted, the flow from one valve was erratic, and the other valve opened 15 minutes late. On January 3, 1982, a plant trip on startup occurred as the result of a loss of condenser vacuum. Using the assumption that the two events were coincident, the ASP program estimate of the conditional core damage frequency*(CDF) for this event is 5.0E-05: - *

2)

Licensee Event Report (LER) No. 255/82-024, 255/82-025, 255/82-044.

On August 19, 1982, a design error was identified which disclosed that during a Design Basis Event, if either emergency diesel generator failed to start or run, then the service water pumps operating on the running diesel generator could trip due to runout conditions. On August 27, 1982, a second design error was 9610040126 960926 PDR ADOCK 05000255 S

PDR A alS ENERGY COMPANY

2 identified that revealed, given a Loss of Offsite Power (LOOP) and subsequent operation of the Normal Shutdown Sequencer (NSD), the Design Basis Accident (OBA) sequencer would not operate if a safety injection signal were generated more than 55 seconds after the LOOP. On November 30, 1982, a third design deficiency was identified that indicated that Motor Control Center (MCC) 1 and MCC 2 feeder breakers could potentially overload following a Loss of Coolant Accident (LOCA) if the station batteries were discharged or the hydrogen recombiners were placed on line. The draft report postulates an increased CDF of 3.0E-04 for this combination of conditions.

Consumers Power Company has reviewed the draft report and has concluded that the calculated.CDFs greatly overstate the actual risks from the events. This appears to be a result of inyalid assumptions being used by the analyst due to insufficient information being available from the associated LERs. Attachment 1 to this letter provides Consumers Power Company's comments on the draft report and includes additional information on the events.

SUMMARY

OF COMMITMENTS This letter contains no new commitments and no revisions to existing commitments.

Thomas C. Berdine Manager, Licensing CC Administrator, Region Ill, USNRC Project Manager, NRR, USNRC NRC Resident Inspector - Palisades Attachment

ATTACHMENT 1 CONSUMERS POWER COMPANY PALISADES PLANT DOCKET 50-255 COMMENTS ON DRAFT 1982-83 ACCIDENT SEQUENCE PRECURSOR REPORT 4 Pages

A.

ACCIDENT SEQUENCE PRECURSOR REPORT EVENT

1.

Licensee Event Report (LER) No.255182-002.

On January 6, 1982, the normal monthly surveillance test disclosed that the controllers for Auxiliary Feedwater control valves (CV-0727 and CV-0749) were not properly adjusted, the flow from one valve was erratic, and the other valve opened 15 minutes late. On January 3, 1982, a plant trip on startup occurred as the result of a loss of condenser vacuum. Using the assumption that the two events were coincident, the ASP program estimate of the conditional core damage frequency (GDF) for this event is 5.0E-05.

Consumers Power Company's Comments:

For Licensee Event Report (LER) No.255182-002, the analysis in'the Accident Sequence Precursor (ASP) program appears to assume that the conditions identified in the surveillance test of the 6th of January were present three days prior on the 3rd of January when the plant tripped. Additionally, the AFW system was assumed to be failed because the LER made no mention of its operation and the analysis assumed that this was an ATWS condition because, as indicated in the LER, the reactor remained critical. In hindsight, while the information provided was apparently adequate for the time, it is very limited and misleading when examined by itself years later. In actuality, this probably is a non-.event. The operators were in control throughout the event. The plant had a cycling relief valve on the hogging air ejector and leaks from several failed rupture disks on the LP turbine. The operators were reducing power so the reactor could be maintained critical with the main turbine isolated for repairs. In fact, the operators were successful in that at 21 :03 on January 3, 1982, the main generator tripped, but, by design, there was no turbine trip and, therefore, no need for a reactor trip. The plant continued to operate successfully on normal power conversion (Main Feedwater and Main Condenser). Under these conditions, there would be no demand on the AFW system (steam generator levels were adequate and being maintained) and, therefore, no need to mention its performance. The plant was operated in this condition until 05:45 on January 4, 1982, when all of the leaks had been identified. The turbine and main feed pumps were tripped by the operators to allow isolation of the turbine and condenser for repairs. At this point, we. assume the plant was operating with AFW in service and atmospheric dumps cycling to remove heat while the Main Steam*lsolationValves*were closed. Additionally; the AFW system would have had to operate successfully prior to the event, since it would have been necessary to allow the startup to have proceeded to this point. Given the added information, this should not be considered an ATWS event and the AFW system should not be considered failed during the event.

. With respect to the surveillance test failure, it would be appropriate to analyze this as a normal plant trip with the AFW system degra~ed. However, the Technical Specification determination of operability should not be confused with the real ability of the system to perform its function. The system was capable of delivering adequate flow to the steam generator(s). By definition, the system can be successful if it provides 150 gpm to each steam generator or 300 gpm to one steam generator. Under normal circumstances, the mass of water in the steam generators is adequate to allow substantial time for the operators to compensate for malfunctions in system response. The limiting event for this condition would be a complete loss of main feedwater event (which is not the event that occurred). The sequence of events, as described above, was taken from Unit Outage Report 82-002 from the Plant Operations/Maintenance Superintendent to the Plant Manager.

B.

ACCIDENT SEQUENCE PRECURSOR REPORT EVENT

2.

Licensee Event Report (LER) No.255182-024, 255182-025, 255182-044.

On August 19, 1982, a design error was identified which disclosed that during a Design Basis Event, if either emergency diesel generator were to fail to start or run, then the service water pumps operating on the running diesel generator

• could trip due to runout conditions. On August 27, 1982, a second design error was identified that revealed that, given a Loss of Offsite.Power (LOOP) and

• *subsequent operation of the Norma/Shutdown Sequencer (NSD), the OBA.

sequencer would not operate if a safety injection signal were generated more than 55 seconds after the LOOP. On November 30, 1982, a third design deficiency was identified that indicated that Motor Control Center (MCC) 1 and MCC 2 feeder breakers could potentially overload following*a LOCA (Loss of Coolant Accident) if the station batteries were discharged or the hydrogen recombiners were placed on line. The draft report postulates an increased GDF of 3. OE-04 for this combination of conditions.

Consumers Power Company's Comments:

This event involves three instances over a short period of time in which design deficiencies were identified that had the potential to impact risk of operation.

First, on August 19, 1982, a design error was discovered that indicated, given a Loss of Coolant Accident (LOCA) with a concurrent Loss of Off site Power (LOOP) and failure of one emergency diesel generator (EOG), running service water pumps might be susceptible to failure due to runout. On August 27, 1982, another design error was discovered that indicated that following a LOOP and subsequent operation of the Normal Shutdown sequencer, the OBA sequencer would not operate if a safety injection signal (SIS) was received more than 55 seconds after the LOOP. On November 30, 1982, another design error was

t

. discovered which indicated that Motor Control Center (MCC) 1 and MCC 2 feeder breakers could potentially trip on overload following a LOCA if the station batteries were discharged or the hydrogen recombiners were placed on line.

The underlying issue with the first deficiency is that in a OBA scenario when the injection systems transfer to recirculation from the containment sump on low Safety Injection and Refueling Water (SIRW) tank level, the service water valves (CV-0823 and CV-0826) to the Component Cooling Water (CCW) heat exchangers receive an open signal (at the time the valves could go to a full open position). It was postulated that if the valves were to go full open and all three servic;;e water pumps were not operating (assumed due to one diesel generator failure), then the operating service water pump(s) would fail due to runout conditions. The failure of the operating service water pump fails the operating diesel generator and results in a station blackout condition. As noted in their analysis, the condition could occur without a LOCA, given that instrument air compressors are load shed on a LOOP and not sequenced back on to a safety bus. The analysis assumes the instrument air system decays in 2.6 minutes and that it is not likely that the operators would have manually aligned power to the instrument air compressors before the 2.6 minutes expires. It was then assumed that the CVs would transfer to their full open position. Based on discussions with the system engineer, the basis for time of 2.6 minutes is not clear but probably close enough that the difference would not be significant. In addition, since these specific valves have BETTIS actuators which have a history of leaking by their seals, it would be difficult to make a case that the valves would not open. The corrective action for this event was to install hardstops on the CVs to limit their opening position to prevent runout and adjust the discharge valve on P-7B to limit its flow to less than assumed runout flow. Subsequently, the analysis assumptions were modified. The new assumptions were; if EOG 1-1 failed 1.eaving two pumps in service, runout conditions would not occur, but if EOG 1-2 failed leaving only P-7B in service, P-7B could trip due to runout conditions. Therefore, the P-7B discharge valve was set to a position determined to limit flow below runout conditions.

The central issue for the second event is the assumption that if less than three service water pumps are operating after a LOOP, then the operating pumps will fail due to runout when the service water valves (air operated) to the CCW heat exchangers fail open on loss of instrument air (compressors are load shed on loss of offsite power). The failure of operating service water (SW) pumps fails the operating diesel generator(s) and places tt:ie plant in a station blackout condition. In the draft analysis, it was assumed that the failure of EOG 1-1 would leave two SW pumps operable and was not a condition that would result in

/

runout failure. The analysis results were adjusted to include a 0.05 probability of failure of EOG 1-2 (fails two SW pumps and leaves one available but it trips on runout).

_ The Palisades evaluation of the event concluded that the pumps could operate down to very low discharge pressures without tripping on a runout condition.

The evaluation indicates that the pumps would not trip nor experience any damage. These conclusions were substantiated by the manufacturer (verbal discussions). Testing being done at the time also supported these conclusions.

The final recommended actions were to reopen the discharge valve for SW pump P-78 and set the SW valves to limit flow to the CCW heat exchangers to assure adequate flow to the engineered safeguards room coolers and other essential loads; i.e., balance system flows. If emergency power (EP) is not assumed to fail due to SW pump runout conditions, then the resulting probability of core damage could be two orders of magnitude less than indicated in the analysis (using the recent ASP branch probability for random failure of emergency power (SE-04). Therefore, if the assumption of pump failure on runout is excluded, the scenarios become much less significant.

- - ------------- --