ML20135G966
| ML20135G966 | |
| Person / Time | |
|---|---|
| Site: | Indian Point  |
| Issue date: | 05/14/2020 |
| From: | Christopher Hunter NRC/RES/DRA/PRB |
| To: | |
| Littlejohn J (301) 415-0428 | |
| References | |
| LER 247/1992-007 | |
| Download: ML20135G966 (6) | |
Text
B-11 B.5 LER Number 247/92-007 Event
Description:
Reactor Trip and Auxiliary Feedwater Pump Problems Date of Event:
April 13, 1992 Plant:
Indian Point 2 B.5.1 Summary Indian Point 2 was operaiing at 100% power on April 13, 1992 when errors in returning a condenser hotwell to service after maintenance resulted in misleading hotwell level indication. Consequently, plant operators reduced hotwell level too far, resulting in insufficient suction supply to the condensate system and the main feedwater (MFW) pumps.
When the MFW pumps began to experience symptoms associated with cavitation, operators recognized the problem and opened a condenser makeup valve in a 12 inch supply line from the condensate storage tank (CST). MFW pump suction was restored, but the plant tripped a short time later on high steam generator (SG) level. Both motor-driven auxiliary feedwater pumps (MDAFWPs) received auto-start signals; one started and tripped repeatedly and the other did not start.
Investigation suggested that the auxiliary feedwater (AFW) pumps failed to successfully auto-start because of low pressure in their suction supply, which was provided from the same 12 inch header supplying the hotwell. The conditional core damage probability estimated for this event is 3.6 x 10-6. The relative significance of this event compared to other potential events at Indian Point 2 is shown in Fig. B.3.
IER 247/92-007 118-7 1E-6 lB-S TIP Prc=uuo cutff LOFW+
IPA 113-3 1E-2 360 h AFW Fig. B.3. Relative event significance of LER 247192-007 compared with other potential events at Indian Point 2.
LER NO: 247/92-007
B-12 B.5.2 Event Description Indian Point 2 was operating at 100% power when the MFW pumps began experiencing high vibration levels, low suction pressures, and speed variations. As operators attempted to identify the cause, they stepped reactor power down to 25% in an effort to maintain SG levels. It was then recognized that a low hotwell level was causing insufficient condensate supply to the MFW pump suction header.
Valve LCV-1 128 was opened to refill the hotwell via a 12 inch line from the CST and MFW pump performance immediately began to improve. A short time later, high SG levels resulted in a reactor trip. MDAFWP 21 auto-started but immediately tripped. It subsequently restarted and tripped five additional times.
Similar cycling was noted with control logic circuitry for the MFW pump 21 as well. MDAFWP 23 should have started but did not. The turbine-driven auxiliary. feedwater pump (TDAFWP) was not demanded. A short time after the trip, LCV-1 128 was closed and an attempt was made to manually start the MDAFWPs. This attempt was successful.
B.5.3 Additional Event-Related Information Prior to the event, condenser hotwell 22B outlet valve CS-1-3 was isolated during tagout of the circulating water side of that condenser. Later, the tagout was lifted but CS-1-3 remained closed. This resulted in a false high level indication, and operators reduced hotwell makeup to compensate. Low hotwell level resulted, causing the condensate and feedwater system perturbations described.
Operators opened LCV-1128 to quickly make up water to the condenser.
This allowed the MFW system to promptly recover SG level; level in one SG increased sufficiently to result in a high SG level turbine and reactor trip. The normal suction supply for all AFW pumps at Indian Point is from the same line which was used to supply the hotwell. It is believed that the high flow rate to the condenser which existed during this event resulted in a low pressure in the AFW supply piping. In turn, it is thought that this caused AFW pump suction pressure switches to prevent successful auto-start of the pumps. It is unclear why cycling of the main feed pump control logic circuitry was observed.
B.5.4 Modeling Assumptions Seventy-four seconds after the reactor trip, operators isolated the condensate makeup to the hotwell and apparently restored the AFW system to operability. Had they failed to do so, or delayed in doing so, it is possible that repeated start attempts could have resulted in damage to the AFW pumps. At Indian Point 2 a high SG level turbine trip and reactor trip result in a trip of the MFW pumps as well. It was reported that one MFW pump experienced control logic failures after the unit trip. The other feed pump was assumed to have tripped but recoverable.
This event was modeled as a reactor trip with a recoverable loss of MFW and reduced availability of AFW. MDAFWP 21 started and tripped six times in approximately one minute. Multiple starts of a large electric motor within a short period of time may cause its circuit breaker to trip. Motor winding damage is also possible. While it is not known whether MDAFWP 21 experienced any motor winding damage during the event, the motor clearly operated in a manner inconsistent with good practices and it is possible that the manufacturer's recommended duty cycle was exceeded. Therefore, it is considered inappropriate to credit MDAFWP 21 as being fully available at its usual level of reliability during the balance of the event. In addition, operation of the TDAFWP at Indian Point requires manual intervention to align pump output to a steam generator.
LER NO: 247/92-007
B-13 It is believed that MDAFWP 23 did not auto-start during the event because of the low pressure experienced at its suction. Further, it is also believed that this condition would have cleared without operator intervention before the steam generator inventory was depleted.
As there is no reason to question the pump's ability to perform its required function, MDAFWP 23 is credited as being fully available during the event. The AFW system model for this event consists therefore of one MDAFWP and one TDAFWP recoverable (or available with manual intervention) and one MDAFWP fully available.
Because cues existed to indicate the need to isolate LCV-1128, and because manual alignment of the TDAFWP was a proceduralized action, AFW recovery was assigned to ASP recovery class "R4" (Reference Vol. 17, Section A. 1.3 of this report). This recovery class is appropriate when "the failure appeared recoverable in the required period from the control room and was considered routine or procedurally based." The nonrecovery likelihood for this class is 0.04.
In event of complete AFW failure, it may be possible at Indian Point to rapidly depressurize the plant secondary side to 400 psig and supply the steam generators with the condensate pumps. While limited information is available concerning the thermal hydraulics, reactor physics, human factors, and other issues related to this approach, an effort has been made to credit this strategy. As time to implement this strategy could be limited and operator burden could be significant, the nonrecovery for this event is assigned from class "R3", "The failure appeared recoverable in the required period from the control room, but recovery was not routine or involved substantial operator burden." Component failures are assumed to be negligible in comparison with the operator nonrecovery probability. The nonrecovery probability for this class, 0.12, was incorporated by adjusting the AFW nonrecovery probability.
B.5.5 Analysis Results The conditional probability of core damage estimated for this event is 3.6 x 10-6. The two dominant core damage sequences, highlighted on the following event tree in Fig. B.4, are associated with failures of MFW, AFW, and feed-and-bleed cooling. This event has been analyzed based on the information available in the referenced LER.
LER NO: 247/92-007
B-14 A
RT A
WW I"ORv/" IO"" I HK F[
PO ThA~8RT #3
~
SRV 15kV m
W OPENI ME EMD
__COAL JRESEAT N 13 OK OK 11 CD 12 CD OK OK OK 13 C
14 CD OK OK 16 CW ME
~
17 C
1s ATWS Fig. B.4. Dominant core damage sequences for LER 247/92-007 LER NO: 247/92-007
B-15 LER NO: 247/92-007
B-16 LER NO: 247/92-007