ML20058P522
| ML20058P522 | |
| Person / Time | |
|---|---|
| Site: | Susquehanna, Diablo Canyon  |
| Issue date: | 02/03/1990 |
| From: | NRC OFFICE FOR ANALYSIS & EVALUATION OF OPERATIONAL DATA (AEOD) |
| To: | |
| Shared Package | |
| ML20058P504 | List: |
| References | |
| NUDOCS 9008170144 | |
| Download: ML20058P522 (6) | |
Text
.
ACCIDENT SEQUENCE PRECURSOR PROGRAM EVENT ANALYSIS Preliminary i
This analysis was based on information.which is preliminary in nature and subject to revision.
L E R N o.:
387/PNO I 90 8 l
I Event Desciption:
Loss of RPS bus "B" causes loss of shutdown cooling.
Date of Event:
February 3,1990 Plant:
Susquehanna Unit 1 Summary:
A ground fault resulted in the loss of Reactor Protectic: System (RPS) bus "B". The RPS I
buses provide power to the isolation control system and the de-energizatior. of RPS bus "B" resulted in the isolation of a shutdown cooling suction supply valve causing a loss of shutdown
- cooling, t
Core cooling was maintained using safety relief valves, control rod drive pumps, and the suppression pool cooling mode of RHR. The conditional probability e subsequent core r
j damage associated with the event is conservatively estimated to be 4.1 x 10.5. The relative
[
l significance of this event compared to other postulated events at Susquehanna is shown below:
LER No. 387/PNO 190-8 IE-8 IE 7 IE 6 IE 5.
IE-4 IE.3 IE-2 I
I I
I I
i i
o a
Trip LOFW HPC1/RCIC Unavail(360 hrs) l 1
Event
Description:
Susquehanna Unit One was sno down on February 1,1990 for maintenance. Two days later h
-[
9008170144 999g33 gDR ADOCK 05000323 i
i a test of the alternate power supply to RPS bus "B" was conducted. When the normal power o
supply was de-energized, the alternate supply failed to close in on the bus. Manual closure of the normal and alternate supplies was attempted but they immediately tripped open.
Subsequent investigation revealed that an insulator on a circuit breaker fed by RPS bus "B" had developed a crack which resulted in a ground fault on the bus.
The RPS system provides power to the isolation control system. Loss of power on RPS bus "B" resulted in isolation of certain valves controlled by this system incitJh ; a Residual Heat Removal (RHR) System shutdown cooling suction supply valve. With shutdown cooling lost, reactor water temperature began to rise. Operators stemmed the coolant temperature rise at 252 F. and 31 psia by opening three Safety Relief Valves (SRVs) and pmviding makeup fmm the control rod drive system. Suppression pool cooling was used to remove the heat transferred to the suppression pool by tl. SRVs.
De defective circuit breaker was replaced and RPS bus"B" was re-energized.' This permitted a retum to shutdown cooling.
Event Related Plant Information:
nere are two RPS buses for each unit, each of which is normally fed by a high inertia motor-generator (MG) set. The MG sets are fed from 480 volt boards and generate 120 Volts AC (VAC). There are alternate 120 VAC sources which may fecd the RPS buses but these are interlocked to ensure that no more than one bus is fed from its alternate supply at any given time. Intef icks also exist to ensure that the MG sets are not paralleled with the alternate supplies.
In addition to providing power for the tractor protection system, the RPS buses pmvide power to the isolation control system. This system operates valves as required to isolate the reactor vessel and/or primary contai. ment to conserve coolant inventory and prevent the release of radioactive materials. Loss of power to RPS bus "B" de-energizes all"B" train logic in the l
isolation control system and results in the isolation of the Reactor Water Cleanup System and the shutdown cooling supply w the RHR system.
4 The normal shutdown cooling flow path is from Recirculation loop B, through inboard and outboard containment isolation valves, to either of two pairs of RHR pumps and thence to either of two RHR heat exchangers which are cooled by RHR service water. The cooled water is then retumed to the reactor vessel through either "A" or "B" recirculation loop. Isolation of either shutdown cooling supply valve renders the shutdown cooling system inoperable and other means of decay heat removal must be pmvided when the unit is s'aut down.
l Susquehanna's " Loss of RHR Shutdown Cooling Mode" procedure required that operators provide for natural or forced reactor coolant circulation within the vessel, that they demonstrate operability of alternate methods of decay heat removal using RHR/ Low Pressure Coolant injection (LPCI) system and core spray system, and that they reestablish cooling by one of several methods. These methods involve control rod drive cooling, condensate transfer through keepful and/or shutdown cooling ilush, condensate, RHR, core spray, RWCU recirculation / letdown, and SRV blowdown.
{
ASP M neiling Assumptions and Approach:
An event tree model of sequences in core damage given a total loss of shutdown cooling was developed considering the pob. ial unavailability of mitigating features described in Susquehanna procedure ON-14r001, Rev. 7, " Loss of RHR Shutdown Cooling Mode."
This event tree, shown in Fig.1, addresses RPV makeup via the control rod drive, condensate, core spray, or LPCI systems. Heat removal and letdown is via the SRVs (to the suppression pool) or the RWCU system. If the SRVs are used, then sappression pooling cooling is also assumed required. If the RWCU system is utilized, then the model assumes that the condensate system is required for makeup. Both this assumption and the requirement for short term suppression pool cooling are most likely conservative, considering the shutdown decay heat levels which exist during cold shutdown.
Additional conservatism exists in that not all makeup / letdown combinations identified in ON-149 001 are included in the model.
Fig.1 includes the following core damage sequences:
Srquence Descriotion 1
Successful use of the SRVs and SP cooling for heat removal, but failure to provided RPV makeup via the CRD, condensate, core spray and LPCI systems.
2 Failure of SP cooling following successful opening of the SRVs. RWCU is successful but makeup via the condensate system fails.
3 Failure of SP cooling following successful opening of the SRVs. RWCU fails to provide letdown / heat removal.
4 Similar to sequence 2 except the SRVs fail to open.
5 Similar to sequence 3 except the SRVs fail to open.
n o
, 4 0-The following branch probability values were utilized with the event tree.
I Branch Enihim Probability Failure of at least three SRVs to open. The SRV/ ADS failure 4.4 x 104 probability estimated from precursor data was utilized, with'a non-recovery estimate of 0.12, to take into account the long time period i
available for repair.
Failure of SP Cooling. A failure probability of 2 x 10 3 was _
2.4 x 104 j
assumed (this value is also used in the current at power ASP models), with a non recovery estimate of 0.12.
Unavailability of RWCU. A failure probability of 0.05 was 5.0 x 10 2 assumed. RWCU is isolated upon loss of RPS bus "B." The containmem isolation signal must be unblocked or the RWCU isolation valve manually opened. The failure probability for this j
was assumed to be 0.04. In addition, a failure probability of 0.01
/
for W RWCU system itself was also assumed.
j i
Unavailability of CRD Cooling.
- 1.0 x 10 2 l
i Unn 211 '"N of makeup via the condensate system. A failure 1.0 x 10 2 proom ty.,. v.01 was assumed.
U...ailability of e spray.
- 1.0 x 10-3 U
.*+ ~ '1.PCI Given SP cooling success, failure of 2.0 x 10 3 i.
Avu rd
, to oper. the series RPV injection valves in bd b,..ae probability of 0.002 was assumed.
- value ebently used wib current ASP models.
Failure % implement the loss of SDC procedure has not been specifically addressed in Figure 1. Based on the long estimated time to core uncovery (~16 hours based on simplified hand calculations) and the two year operator training cycle at Susquehanna, the likelihood of -
operator error.is low, and equipment failure is assumed to dominate the core damage probability estimate.
I Analysis Results:
Based on the model described above, a core damage probability of 4.1 x 10 5 is estimated.
Because of the long response times associated with shutdown related events, and the potenti')
for system unavailabilities during shutdown because of allowed maintenance, the uncertainty in the core damage probability estimate is high.
3
(
l l
l l
l
\\
l
.,l l -
h b
i Loss of SRVs M
CR0 Core LPCI N
Segwnce -
SDC Open.
(SP
.RWCU Cooling COND sp,,y stale
, No.
Cooling) i oK oK oK l-ok l
CD i
l ok CD 2
CD 3--
oK CD 4
CD 5
l l
l
\\'
Fig.1. Event Tree for Loss of Shutdown Cooling at Susquehanna 1 -
4
,