ML20135G969
| ML20135G969 | |
| Person / Time | |
|---|---|
| Site: | LaSalle  |
| Issue date: | 05/14/2020 |
| From: | Christopher Hunter NRC/RES/DRA/PRB |
| To: | |
| Littlejohn J (301) 415-0428 | |
| References | |
| LER 1993-015-00 | |
| Download: ML20135G969 (7) | |
Text
A. 10-1 A. 10 LER No. 373/93-015 Event
Description:
Scram and Loss-of-offsite Power Date of Event: September 14, 1993 Plant: LaSalle I A. 10. 1 Summary LaSalle I was operating at 100% power on September 14, 1993, when a fault occurred in the buswork associated with the system auxiliary transformer (SAT). The resulting electrical system perturbations caused the loss of one main feed pump and a reactor scram on low vessel level. When the main generator separated from the grid, the unit auxiliary transformer (UAT) was no longer able to provide power to plant auxiliaries, and a loss-of-offsite power (LOOP) ensued. The conditional core damage probability estimated for this event is 1.3 x 10- . The relative significance of this event compared to other postulated events at LaSalle I is shown in Fig, A. 10. 1.
LER 373/93-015 1E-7 1E-6 1E-5 1E-4 1E-3 I1E-2 II I I IL LOFW &HPCS I L-.360 hEP 1LL LOOP I ~360 hiHPCS & RCIC L "
-Precursor Cutoff Fig. A. 10. 1 Relative event significance of LER 373/93-015 compared with other potential events at LaSalle 1 A. 10. 2 Event Description Water leaking into the Unit I SAT bus duct caused a severe electrical fault at LaSalle 1 while the plant was at 100% power. SAT output voltage dropped sharply, and loads fed from the SAT transferred to the UAT.
The voltage reduction caused the 1B turbine-driven reactor feed pump control circuitry to reduce pump flow to zero. Shortly thereafter, the reactor scrammed on low vessel level.
Reactor makeup after the scram was initially supplied by the motor-driven reactor feed pump, but the vessel overfilled, resulting in feed pump and main turbine/generator trips. Once the main generator separated from the grid and the UAT was deenergized, the plant experienced a LOOP. The emergency diesel generators (EDGs) started and loaded to supply emergency buses, and the high-pressure core spray (HPCS) diesel started. Safety/relief valves (SRVs) were operated to reduce pressure by relieving steam to the suppression pool. Suppression pool cooling (SPC) was initiated, and reactor core isolation cooling (RCIC) was aligned for vessel makeup. After about 75 min, offsite ac power was restored to Unit 1 by connecting Unit 1 buses to Unit 2. Late in the event, one SRV failed to operate on demand. When reactor pressure decreased to 500 LER No. 373/93-0 15
A. 10-2 psig, the low-pressure core spray (LPCS) system was aligned to provide makeup, and the reactor was then placed in shutdown cooling (SDC).
During the event, ambiguous position indication was observed for several SRVs. An investigation determined that one valve was misaligned and did not fully open. The other ambiguous position indications were attributed to miscalibrated position indicators. The safety parameter display system (SPDS) also lost power during the event and was unavailable; however, redundant control room instrumentation and the process computer remained available.
A. 10.3 Additional Event-Related Information In the event that residual heat removal (RHR) fails, the containment can be vented to remove decay heat and prevent overpressurization. To achieve this, the operator manually vents the suppression pool or the drywell. The steaming that will occur in the suppression pool may fail any injection source [such as low-pressure coolant injection (LPCI)] that draws from the suppression pool. Therefore, the feed operation associated with venting must come from an injection system that operates at low pressure and that has a source of water other than the suppression pool.
A. 10. 4 Modeling Assumptions This event was modeled as a plant-centered LOOP, in accordance with the LOOP classification scheme of NUREG- 103 2, Evaluation of Station Blackout Accidents at Nuclear Power Plants. The short-term (within the first 30 min) LOOP nonrecovery probability was assumed to be 1.0, given that the SAT was tripped and Jocked out by the fault. The probability of not recovering ac power before battery depletion following the postulated failure of the EDGs was determined using the models described in Revised LOOP Recovery andPWR SealLOCA Models, ORNL/NRC/LTR-89-1 1, August 1989. These models are based on the results of the data distributions contained in NUREG- 1032. Because offsite power sources remained available throughout the event and procedures were in place to line up these sources, it was assumed that the overall recovery would be similar to a normal LOOP. That is, the probability of not recovering ac power at the point of battery depletion was assumed to be essentially equal to a normal LOOP, although the probability of not recovering ac power at 30 min was assumed to be 1.0 because of the problems with the SAT. Therefore, the long-term (before battery depletion) nonrecovery was revised to be the overall (combined short- and long-term) nonrecovery probability (5.5 x 10-2)
The SRV failure that occurred was attributed to leakage of nitrogen control air from a nonsafety accumulator over a 2-h period. Because the automatic depressurization system (ADS) accumulator for the SRV was unaffected, ADS was not impacted, and no changes were made to its failure probability included in the accident sequence precursor (ASP) models. Also, because additional indications were available, the SRV indicator and SPDS malfunctions were not considered to impact any failure probabilities included in the ASP models.
The current ASP event trees for LaSalle do not model the potential use of RCIC to provide reactor pressure vessel makeup in the event of a single stuck-open SRV. However, the use of RCIC for this purpose was included in NUREG- 1150 and the utility's individual plant examination. To address this, the conditional probabilities for the applicable sequences (sequences 50-55 and 69 in the event tree in Fig. A. 10.2) were reduced by the probability of failing to successfully use RCIC for this purpose. This is the probability that either RCIC fails, two or more SRVs fail to close given one or more fails to close, or RHR long-term cooling (SDC and SPC) fails given RCIC is successful and only one SRV fails open. This probability can be approximated by LER No. 373/93-015
A. 10-3 p(RCIC) + p(2 or more valves fail open II or more valves fail open) .
This approximation assumes that sequences involving RCIC success avoid core damage if RHR is also successful. Because the probability of RHR failure is very small relative to the probability of failing RCIC, this approximation is valid. The failure probability for RCIC during this event was estimated at 0.042, the nominal ASP value. A value of 0.027 was estimated for p(2 or more valves fail open I I or more valves fail open), based on an estimated probability of 0.0015 for two or more SRVs stuck open (see NUREG/CR-45 50, Vol. 1, Rev. 1, Analysis of Core Damage Frequency: Internal Events Methodology, January 1990, p. 6-10) and an estimated probability of 0.056 for one or more SRVs stuck open (developed as described in NUREG/CR-4674, Vol. 1, Precursors to PotentialCore Damage Accidents: 1985 A Status Report, Appendix C). The multiplier used to adjust the conditional probability for sequences 50-55 and 69 to account for this potential use of RCIC to mitigate the effects of a single stuck-open SRV is, therefore, 0.042 [p(RCIC)] + 0.027 [p(2 or more SRVs fail open I I or more SRVs fail open)] = 0.069.
The existing ASP model was also modified to include the potential use of containment venting for decay heat removal in the event that both RHR!SPC and RHR/ SDC fail. This was done by revising the conditional probability for sequence failure of both RJIR cooling modes (sequences 40-44, 47, and equipment) to also include failure to vent the containment. The probability of failing to vent was assumed to be dominated by human error. A probability of 0. 01 was utilized for sequences in which the injection source operates at low pressure and has a source of water separate from the suppression pool. For sequences in which the injection source takes suction from the suppression pool (such as LPCS or LPCI), an alternate injection source, the control rod drive (CRD) pumps or essential service water (SW) (RHRSW in the ASP models), must be aligned for injection following venting. Venting is considered much less reliable in such cases; an operator error probability of 0.5 was utilized (see NRR Daily Events Evaluation Manual, 1-275-03-336-01, January 31, 1992).
The conditional probability for sequence 40 was revised from 5.6 x 10-6 to 5.6 x 10-8 to reflect the potential use of containment venting (although HPCS provides injection success in this sequence, the CRD pumps, which do not take suction from the suppression pool, are also assumed to be available for injection). The conditional probability for sequence 55 was revised from 4.8 x 10-6 to 5.3 x 10-7 to reflect the potential use of RCIC following a single stuck-open relief valve. Other sequences that were potentially impacted by these two changes had calculated probabilities below 4.0 x 10-6 (unmodified) and had minimal effect on the core damage probability estimated for the event. The probability values for these sequences, which are not shown on the calculation sheets, were not revised.
During the event, reactor protection system (RPS) motor generator set B tripped on a motor fault. Because an alternate supply to RPS bus B could not be reenergized from the Unit 2 feed or from the EDGs, certain primary containment isolations could not be immediately reset, including one affecting SDC. It was necessary to arrange a temporary power feed to RPS bus B before SDC could be placed in service. It was assumed that the failure to deal successfully with the RPS failure and recover SDC was bounded by the existing operator error value for the SDC branch, because substantial time was available for alignment of SDC.
A. 10. 5 Analysis Results The conditional probability of subsequent core damage for this event is estimated to be 1.3 x 10 The dominant core damage sequence, highlighted on the event tree presented in Fig. A. 10.2, involves the plant-centered LOOP, a postulated failure of emergency power, and failure to recover emergency power before battery depletion.
LER No. 373/93-0 15
A.I10-4 Consideration of RCIC for makeup after a single stuck-open relief valve and containment venting, following the postulated loss of RHR shutdown and SPC have little effect on' the core damage probability estimated for the event.
LER No. 373/93-015
A. 10-5 IIoo (LNG)
SHT RCSOLD HMIAPS IMODE)
(H) PC (SP MODE) OTHIER SEQ END NO6 STATE OK 40 Cot 06 41 CS 92 Co 06 OP 44 CO OK 4 COK Q0 Co M4-7 CO 1
OK 49 Co OK C
00 OK 51 CD OK 52 CD OK 53 53 CO 54 Co OK OK 5R CO OP S,
0 S. CO.
0 55 CO OP OP OK Al CE) 62 CID 63 CSD 54 Co R ATWS AS CO OK OK PA Co 67 CD 0K OK 66 DE 69 CO OK 06 A2 Co Al CD 97 ATh5S Fig. A. 10.2 Dominant core damage sequence for LER 373/93-0 15 LER No. 373/93-015
A. 10-6 CONDITIONAL CORE DAMAGE PROBABILITY CALCULATIONS Event Identifier: 373/93-015 Event
Description:
Scram and Loss-of-Offsite Power Event Date: September 14, 1993 Plant: LaSalle 1 INITIATING EVENT NONRECOVERABLE INITIATING EVENT PROBABILITIES LOOP 1.OE+OO SEQUENCE CONDITIONAL PROBABILITY SUMS End State/Initiator Probabili ty CD LOOP 1.4E-0401)
Total 1.4E-04(l)
ATWS LOOP 3.OE-05 Total 3.OE-05 SEQUENCE CONDITIONAL PROBABILITIES (PROBABILITY ORDER)
Sequence End State Prob N Rec**
83 LOOP emerg.power -rx.shutdown/ep EP.REC CD 1.3E-04 8.OE-01 40 LOOP -emerg.power -rx.shutdown srv.chatt/Loop.-scram -srv.ctose CD 5.6E-06(2) 1.2E-01
-hpci rhr(sdc) rhr(spcoot)/rhr(sdc) 55 LOOP -emerg.power -rx.shutdown srv.chatt/loop.-scram srv.ctose CD 4.8E-06(2) 2.4E-01 hpci srv.ads 98 LOOP -emerg.power rx.shutdown ATWS 3.OE-05 1.OE+00
- nonrecovery credit for edited case SEQUENCE CONDITIONAL PROBABILITIES (SEQUENCE ORDER)
Sequence End State Prob N Rec**
40 LOOP -emerg.power -rx.shutdown srv.chatt/Loop.-scram -srv.ctose CD 5.6E-06(2) 1.2E-01
-hpci rhr(sdc) rhr(spcooL)/rhr(sdc) 55 LOOP -emerg.power -rx.shutdown srv.chatt/Loop.-scram srv.ctose CD 4.8E-06(2) 2.4E-01 hpci srv.ads 98 LOOP -emerg.power rx.shutdown ATWS 3.OE-05 1.OE+00 83 LOOP emerg.power -rx.shutdown/ep EP.REC CD 1.3E-04 8.OE-01
- nonrecovery credit for edited case SEQUENCE MODEL: e:\asp\modets\bwrcseal .cmp BRANCH MODEL: e:\asp\modets\tasatte.SL1 PROBABILITY FILE: e:\asp\modets\BWR_CSL1 .PRO No Recovery Limit BRANCH FREQUENCIES/PROBABILITIES Branch System Non-Rec cv Opr Fai l trans 7.4E-05 1.OE+O0 LOOP 1.6E-05 > 1.6E-05 5.3E-01 > 1.OE+00 LER No. 373/93-015
A. 10-7 Branch Model: INITOR Initiator Freq: 1.6E-05 Ioca 3.3E-06 5.OE-01 rx. shutdown 3.OE-05 1.OE+00 rx .shutdown/ep 3.5E-04 1.OE+00 pcsl trans 1.7E-01 1.OE+00 srv.chaLL/trans. -scram 1.OE+00 1.OE+00 srv.chaLt/Loop.-scram 1.OE+00 1.OE+00 srv.c Lose 5.6E-02 1.OE+00 emerg .power 2.9E-03 8.OE-01 EP .REC 1.7E-01 > 5.5E-02 1.OE+00 Branch Model: 1.OF.1 Train 1 Cond Prob: 1.7E-01 > 5.5E-02(3) fw/pcs. trans 4.6E-01 3.4E-01 fw/pcs. loca 1.OE+OO 3.4E-01 hpc i 2.OE-02 3.4E-01 rcic 6.OE-02 7.OE-01 crd 1.OE-02 1.OE+00 1 O0E-02 srv.ads 3.7E-03 7.1E-01 1.OE-02 Ipcs 2.OE-02 3.4E-01 Lpci (rhr)/Lpcs 6.OE-04 7.1E-01 rh r(sdc) 2.3E-02 3.4E-01 1.OE-03 rhr(sdc)/- lpci 2.OE-02 3.4E-01 1.OE-03 rhr(sdc)/lpci 1.OE+00 1.OE+00 1.OE-03 rhr(spcooL )/rhr(sdc) 2.OE-03 3.4E-01 rhr(spcooL)/- Lpci .rhr(sdc) 2.OE-03 3.4E-01 rhr(spcoot)/Lpci .rhr(sdc) 9.3E-02 1.OE+00 rhrsw 2.OE-02 3.4E-01 2.OE-03
- branch model file
- forced Notes:
- 1. See Analysis Results for the core damage probability estimated for this event after consideration of additional mitigating features.
- 2. See Modeling Assumptions for the revised conditional probability for this sequence.
- 3. Value modified to account for overall AC power recovery. See Modeling Assumptions for the development of this value.
LER No. 373/93-015