ML20059J606
| ML20059J606 | |
| Person / Time | |
|---|---|
| Site: | Fermi  |
| Issue date: | 10/31/1993 |
| From: | Hill S, Spence R, Steinke W IDAHO NATIONAL ENGINEERING & ENVIRONMENTAL LABORATORY |
| To: | NRC OFFICE FOR ANALYSIS & EVALUATION OF OPERATIONAL DATA (AEOD) |
| Shared Package | |
| ML20059J601 | List: |
| References | |
| NUDOCS 9311120285 | |
| Download: ML20059J606 (28) | |
Text
'
_ 's .
TRIP REPORT:
ONSITE ANALYSIS OF. ;
THE HUMAN FACTORS OF AN EVENT AT ENRICO FERMI 2 j ON AUGUST 13,1993 i
SPURIOUS REACTOR SCRAM WITH LOSS OF CONDENSER VACUUM i
Robert A. Spence Susan G. Hill William F. Steinke David A. Prawdzik Published October 1993 t
Idaho National Engineering Laboratory EG&G Idaho, Inc.
P.O. Box 1625 ,
Idaho Falls,ID 83415 !
i Prepared for the ,
Office for Analysis and Evaluation of Operational Data U.S. Nuclear Regulaton Commission Washington, DC 20555 Under DOE-ID Contract No. DE-AC07-761D01570 9311120285 931104
-PDR P ADOCK 05000338 l PDR
O ~
k EXECUTIVE
SUMMARY
At 9:46 a.m., August 13,1993, a reactor building nuclear power plant operator was !
removing tape residue from a pressure instrument calibration / vent valve connected to a common reactor pressure vessel instrument reference leg. The operator's action slightly ;
unseated the valve causing a spurious reactor pressure vessel high water level signal. !
The unit was at 93.5 percent when the ensuing turbine trip, reactor feed pump trip, and reactor scram occurred. Subsequently, reactor pressure vessel water level decreased, ,
which initiated high pressure coolant injection and reactor core isolation cooling, tripped the reactor recirculation pumps, and initiated group isolations including the reactor water cleanup system.
The control room crew did not address annunciators associated with the gland seal steam system, which was in manual control. As a result, condenser vacuum was lost 16 minutes ,
into the event and the main steam isolation valves closed. Without the main condenser !
available, the crew maintained pressure control by cycling safety relief valves. The crew cycled nine safety relief valves during a 1-% hour period. A leaking fuel assembly caused ,
radiation levels in the drywell/ torus areas to increase as coolant was discharged to the torus through the safety relief valves. The nuclear shift supervisor classified the event as an Unusual Event according to the emergency plan at 10:15 a.m. because of high ,
pressure coulant injection system flow into the reactor pressure vessel.
A post scram evaluation was conducted by the plant staff, which resulted in several plant
+
deviation event reports to follow up on significant equipment problems that occurred during the recovery. As part of this evaluation, they are considering procedure changes for restoring reactor water cleanup and controlling control rod drive flow during scram recovery to minimize thermal stratification in future events. Recovery procedures have already been changed to address the revised partially-automatic operational configuration of the gland seal steam system.
In November 1992, a loss of feedwater scram occurred at the beginning of the operating cycle that can be compared and contrasted with this event. All control panel actions were performed by the two nuclear supervising operators assigned to the control room ;
compared to the August event where up to 9 licensed operators were used. The November crew's performance differed from the August 1993 response for feedwater control, and recognition of alarms on the condenser in time to maintain vacuum, though these actions were not procedurally driven. The end result of the November 1992 event was that the reactor pressure vessel cooldown rate was not exceeded, the condenser remained available, and eliminated the need to cycle safety relief valves to maintain reactor pressure.
The human performance analysis of the August 1993 event focused on control room operations and the operators response to the spurious reactor pressure vessel Level 8 ;
signal actuations and subsequent main steam isolation valve closures. The analysis reinforced the importance of simulator fidelity to actual operations affected by human >
performance; operator simulator training reflected the gland seal steam system in its 1
iii
automatic mode, as it was originally designed to operate, but not its existing manual I
operating mode.
Command, control, and co'mmunication could have been better. Control board activities were not clearly assigned by the responsible control room nuclear supervising operator, ,
but were assumed by the first operator to arrive at the panel. The result was that the '
monitoring of all the panels was not effective, purticularly the balance of plant controls and displays, including the gland seal steam system and the main condenser vacuum.
A comparison of this event with the November 1992 loss of feedwater demonstrates the importance of obtaining both positive and negative human performance information in a lessons-learned approach. The scope of the licensee's November 1992 human performance investigation could have been extended to include an evaluation of " things that went right". Lessons could have been learned regarding the need to manually operate the gland seal steam system to prevent loss of condenser vacuum during reactor and turbine trips, throttle control rod drive flow, and restore the reactor water cleanup -
early. These could have been disseminated to training and operations, and thereby minimized some of the complications in the August 1993 event. .
L T
l 9
i I
i
$V
+ p ACKNOWLEDGMENTS ,
t We appreciate the Enrico' Fermi Unit 2 staff's cooperation in freely providing the necessary information and scheduling interviews to analyze the human factors of the operating event. We thank the Unit 2 operators who were on duty during the event and the three crew members from the November 1992 event for their cooperation during the >
interviews.
i t
l I
e L
I v
a CONTENTS ,
EXECUTIVE SUM MAR . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . iii ACKNOWLEDG MENTS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . v .;
ACRONYMS...................................................... ix INTRO D U CTI ON . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1 1
1.1 Pu rpos e . . . . . . . . . . . . . . . . -. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . I i 1.2 Scope ................................................... 1 1.3 Onsite Analysis Team . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1 ,
r i
- 2. DESCRIPTION OF THE EVENT ANALYSIS . . . . . . . . . . . . . . . . . . . . . . . . 2 ~
2.1 Ba ckgrou n d . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2 2.2 November 1992 Loss of Feedwater Scram . . . . . . . . . . . . . . . . . . . . . . . . 5 -
2.3 Time Line of August 13,1993 Event . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5 i
2.4 Analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8 2.4.1 Training . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . 9
- 2.4.2 Teamwork, Command, Control, and Communications . . . . . . . . . . 10 2.4.3 Shift Technical Advisor . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11 2.4.4 Proce dures . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11 2.4.5 S t ress . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 12 2.4.6 Human Machine Interface . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13 ,
14
- 3. LESSONS LEARNED / HUMAN PERFORMANCE STUDY . . . . . . . . . . . .
FIGURES .
t Figure 1 Enrico Fermi 2 control room staffing . . . . . . . . . . . . . . . . . . . . . . . . . . . . 16 a
Figure 2 Sketch of enrico Fermi 2 control room . . . . . . . . . . . . . . . . . . . . . . . . . . 17 :
Figure 3 Enrico Fermi 2 gland seal steam supply system . . . . . . . . . . . . . . . . . . . . 18 Figure 4 Enrico Fermi 2 reactor vessel level . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 20 APPENDICES Appendix A Description of Gland Seal System Operation Appendix B Thermal Stratification ..
vil r
r
W ACRONYMS
^
. AEOD Analysis and Evaluation of Operational Data BWR boiling water reactor CRD control rod drive CRNSO control room nuclear supervising operator DSP Division of Safety Programs EDG emergency diesel generator EOP emergency operating procedure .
HPCI high pressure coolant injection HPES human performance enhancement system INEL Idaho National Engineering I;tboratory ,
LPCI low pressure coolant injection MSIV main steam isolation valve NASS nuclear assistant shift supervisor NPPO nuclear power plant operator NRC Nuclear Regulatory Commission NSO nuclear supervising operator NSS nuclear shift supervisor PCRMS primary containment radiadon monitoring system RCIC reactor core isolation cooling RFP reactor feed pump s RHR residual heat removal ROAB Reactor Operations Analysis Branch RPS reactor protection system RPV reactor pressure vessel RWCU reactor water cleanup ,
SBFW standby feed water SRV safety relief valve STA shift technical advisor 1 6
iX
1 INTRODUCTION -
1.1 Purpose .
The Office for Analysis and Evaluation of Operational Data (AEOD) of the U.S.
Nuclear Regulatory Commission has a program to study human performance during !
operating events. As part of this program, AEOD formed a team to conduct an onsite analysis of the event that occurred at the Enrico Fermi Unit 2 during the day shift on:
August 13, 1993. This report documents the human factors analysis performed as part of t
the study. The Idaho National Engineering Laboratory (INEL) provided program assistance.
At 9:46 a.m., a reactor building nuclear power plant operator (NPPO) was removing tape - ,
residue from an instrument calibration / vent valve. The valve, which was located on a- !
line attached to a common instrument reference leg, unseated, causing a spurious reactor pressure vessel (RPV) high level signal. Unit 2 was at 93.5 percent when the turbice i tripped, reactor feed pump (RFP) tripped, and reactor scrammed. Subsequently, RPV water level decreased to where high pressure coolant injection (HPCI) and reactor core isolation cooling (RCIC) automatically initiated, reactor recirculation pumps tripped and group isolations occurred including reactor water cleanup (RWCU). RPV water level then increased and Level 8 was reached about 3 minutes later. Because of a pressure controller problem, the gland seal steam supply to the main turbine seals was in a manual valve alignment, in which the low power or shutdown gland seal steam supply was isolated to the turbine seals. As a result, condenser vacuum was lost and the main steam isolation valves (MSIVs) closed. The crew cycled nine safety relief valves (SRVs) to control pressure during the next 1% hours. RPV cooldown rate limits were exceeded during the transient and thermal stratification developed, which delayed restart of the recirculation pumps until 8:00 p.m. that evening.
i 1.2 Scope The human factors analysis focused on the factors that influenced the performance of operations activities throughout this event. The analysis was based on data derived from interviews with control room operators, the plant's Post Scram Data and Evaluation Report, and a review of procedures and training material. The plant's training staff also .
reproduced the event on the plant simulator for the onsite analysis team.
A similar event occurred in November 1992. Post scram data and interviews with control f room crew members on duty at that time were used to analyze differences in human performance between the two events.
1.3 Onsite Analysis Team The onsite analysis team visited the Enrico Fermi Power Station Unit 2 from August 18 to August 20,1993, and was composed of the following members:
1
m Robert A. Spence, NRC/AEOD/DSP/ROAB (team leader)
Susan G. Hill, INEL/EG&G Idaho, Inc.
William F. Steinke, INEL/EG&G Idaho, Inc.
David A. Prawdzik, INEL/EG&G Idaho, Inc.
- 2. DESCRIPTION OF THE EVENT ANALYSIS
2.1 Background
The Enrico Fermi Power Station, located in Monroe County, Michigan, on the western shore of Lake Erie, is owned and operated by the Detroit Edison Company. The Unit 2 reactor is a General Electric single-cycle, forced-circulation boiling water reactor (BWR)/4 class with a pressure-suppression Mark I containment. The original design rating was 3,293 MWt and has since been upgraded to 3,430 MWt. Unit 2 has been in commercial operation since January 1988. Enrico Fermi Unit 1, a liquid metal fast breeder reactor, has been shut down since 1972 after 6 years of service.
Unit 2 was at 93.5 percent and had been at power for about 113 days before the event occurred. The dayshift crew members (7 a.m. to 3 p.m.) were on their third day of a 6-day shift rotation. Control room personnel consisted of a nuclear assistant shift supervisor (NASS), a control room nuclear supervising operator (CRNSO), and an additional licensed reactor operator referred to as the panel 603 nuclear supervising ;
operator (NSO) (see Figure 1). The nuclear shift supervisor (NSS) and an additional licensed reactor operator, the shift tagging NSO, were in the NSS office adjacent to the control room discussing clearance work at the time of the event.
At 9:46 a.m., a reactor building NPPO was removing tape and its residue from instrument valve B-21-R004B, " Reactor Vessel Division 2 Pressure Indicator." The calibration / vent valve, which was on a line connected to a common RPV instrument reference leg on instrument rack H21-P005, unseated, causing a spurious RPV high-level 8 signal (214 inches above the top of fuel) on the "B" and "D" channels. The NPPO noticed a fluctuation on the pressure indicator located directly above the valve and subsequently heard loud noises in the plant. He immediately called the control room ,
from the nearest phone and was told that a reactor scram had occurred and he proceeded with his immediate actions. He did not inform the control room crew members of the instrument valve movement at that time.
In the control room, the CRNSO was at the ventilation section of the control boards acknowledging annunciators caused by surveillance testing when he saw multiple alarms ;
illuminated associated with the trip of the feedwater pumps, turbine, and reactor. He immediately proceeded to the feedwater section. The unit responded as designed to the RPV level 8 signal by tripping the turbine and initiating a reactor scram from turbine control valve fast closure. The reactor feedwater pumps also tripped from the high-level signal, resulting in a loss of feedwater. The second reactor operator, the 603 panel NSO, went to the full core display area on panel 603 (see Figure 2) and placed the reactor 2 1 I
mode switch in shutdown. As the RPV water level decreased, RPV Level 3 (173 in.)
actuated a second reactor protection system (RPS) scram within a few seconds; After ~
assessing plant conditions, the NASS positioned himself at the NSO desk and entered the emergency operating procedures (EOPs) based on RPV water level decreasing below 12 vel 3. The tagging NSO ran into the control room area from the NSS office and ,
proceeded to panel 601 to start the standby feedwater (SBFW) pumps. Thirty seconds into the event, RPV water level decreased to Level 2 (110.5 in.), tripping tbc recirculation pumps, actuating the HPCI and RCIC systems, and causing various primary .
containment isolations, including the RWCU. As the tagging NSO started the SBFW pumps, the RCIC system (panel 601) and the HPCI system (panel 602) automatically started. The RPV water level decreased to 105 in. and then began recovering.
A number of failures occurred in the initial stages of recovery that diverted or consumed operator attention. A failed pressure instrument caused the low pressure coolant injection (LPCI) loop select logic to select to "B" loop rather that default to "A" loop as the operators expected. The Division-I torus /drywell pressure recorder locked up. The NASS, after assessing the recorder failure, directed the panel operators to use the Division-II recorder on panel 602. About 100 indicator light bulbs on the control panels ,
failed.The full core display did not have full-in, green-light indication for 6 control rods ,
when the 603 panel operator checked the full core display. He used the process computer control rod display to verify individual control rod positions. An amber light for containment isolation groups 14/16 did not light when expected on a group isolation mimic display board; panel operators had to use open/ closed indicator lights to verify correct positioning of each valve. Shortly after the scram, the CRNSO and tagging NSO smelled smoke. Both operators directed extra personnel in the control room to search for the source of the smoke behind the panels.
Full flow from the HPCI and RCIC systems combined with 1200 to 1300 gpm from the SBFW system refilled the RPV to level 8 in about 3 minutes. Injection flow was ,
automatically terminated by RPV I2 vel 8 trips. The NASS directed the actions of the control room operators to regain and control RPV water level between 173 and 214 in.
in accordance with the RPV Control EOP (29.100.01). The 603 panel NSO implemented the reactor scram abnormal procedure 20.000.21 at the NASS's direction.
A design problem, which had persisted since startup, caused the operators to use tiie manual regulator bypass valves on panel 8M to supply steam to the turbine seals. _ Above 75 percent power the turbine was self-sealing such that the gland seal steam system valve lineup had both regulating and dump control valves isolated with only the manual dump bypass valve F605 open [see Appendix A and Figure 3(c)). Thus, the gland seal steam-supply to the main turbine was not available when the turbine tripped. A white, low gland seal steam pressure annunciator was located on panel 8M to alert the operators of abnormal system pressure, but none of the operators noticed this alarm. About 4 minutes after the reactor trip, a yellow, condenser high pressure annunciator alarmed on panel 804. In the absence of sealing steam, condenser vacuum decreased and caused the main steam bypass valves to close 16 minutes after the turbine trip, and the MSIVs to ,
close 1 minute later.
3 .
b n w ,
The NASS prepared to use SRVs for pressure control by directing the tagging NSO at -i panel 601 to place Division I residual heat removal (RHR) in the torus cooling mode. !
From 10:06 a.m. until 11:49 a.m., nine SRVs were individually cycled by the tagging NSO, maintaining a pressure band between 900 and 1050 psig. The SBFW system was '
restarted to supply 500 gpm to maintain RPV water level while cycling the SRVs. After cycling SRV C, proper closed indication lamps did not light on the control board, but the -
operators verified that the closed position was indicated on the Emergency Response '
Information System computer.
After assessing plant conditions and conferring with the shift technical advisor (STA), the :
NSS declared an Unusual Event due to HPCI injection into the RPV at 10:15 a.m.. ;
Both the NSS and STA had been in the control room since the start of the event. After j
i completing his immediate scram actions, the reactor building NPPO went to the control room and reported his actions regarding the instrument calibration / vent valve to a relief NSS. The NPPO took the relief NSS to the valve on the instrument rack and described -
what he did. The relief NSS informed the on-duty NSS of the NPPO's actions about 10 to 15 minutes after the reactor trip.
i Efforts to restart the recirculation pumps were delayed by the amount of RPV thermal stratification that resulted from the transient. At 10:02 a.m., the 603 panel operator reduced control rod drive (CRD) flow to a minimum of 15 gpm. RWCU Dow was ,
reestablished at 11:25 a.m.,1% hours after the scram. With bottom head temperature on l the RWCU system available, the operators determined that a AT in excess of 145*F was i
present and that the RPV cooldown rate had exceeded 100 / hour during the transient and subsequent SRV cycling. In an effort to induce natural circulation and reduce the RPV temperature differential, the operators raised RPV water level above its normal ,
band to 230 inches. The tagging NSO at panel 601 suggested overriding the SBFW' discharge valve auto closure by holding down the open pushbutton of the SBFW discharge valve until the desired level was achieved. The NASS concurred.
Additional licensed operators from training and the relief crew stood behind control room barriers ready to assist if called upon. The NASS used nine licensed operators in addressing equipment problems that occurred after the reactor scram. An RFP seal tank overflowed after the condenser vacuum was lost. When the vacuum drag no longer drained the tank, a drain pump did not automatically cycle on when high level was reached, which caused the tank to overflow through the tank vent. Emergency diesel generator (EDG) 11 low air pressure annunciated during the recovery, causing the NASS to enter the technical specification limiting condition of operation; maintenance had 3 signed on a clearance for the EDG at the beginning of the shift for work on the air start >
system. A balance-of-plant battery charger 2PC-1 AC input breaker tripped requiring the crew to place a spare charger in service. A small steam leak in the turbine building
- was reported and attributed to a SBFW drain valve. Operators stopped the leak by
' tightening the valve and a pipe cap. When the primary containment radiation monitoring system (PCRMS) was restored at 11:31 a.m., an extra operator determined that radiation levels in the drywell were elevated from 300 to 4800 counts per minute.
The NSS attributed the increase to a known leaking fuel assembly and subsequent SRV ;
cycling. In the interview, the NSS stated he had considered the possibility of upgrading t
4 f
i to an Alert classification if accurate radiation readings were not available for the ,
drywell/ torus area.
By 11:49 a.m., the operators had reopened all MSIVs and established normal pressure control. Seven of the 8 MSIVs open indicator lamps did not work.. The operators l
returned systems to standby or normal status over the next several hours as effarts . '
continued to reduce the RPV AT to less than 145 F to meet conditions for restarting a reactor recirculation pump. At 1:15 p.m., the NASS assessed plant conditions as stable i and exited the EOPs. The NSS terminated the Unusual Event at 7:10 p.m. The reactor recirculation pumps were placed in service about I hour later.
l The plant staff conducted a post scram evaluation, which resulted in plant deviation >
event reports to follow up on significant problems that occurred during the recovery. A corrective action taken to avoid reoccurrence of the initiating cause of the transient was >
posting cautionary signs on some sensitive instrumentation panels in the plant.
2.2 November 1992 less of Feedwater Scram In November 1992, a loss of feedwater scram occurred at the beginning of the operating ;
cycle. All control panel actions were performed by the two NSOs assigned to the control room. That crew's performance differed from the August 1993 response for feedwater control, recognition of turbine panel alarms in time to maintain condenser vacuum, and l timing for restoration of RWCU.
Operators knew immediately that feedwater had been lost by the sequence of the annunciators. The 603 panel operator manually actuated the scram with the reactor mode switch. The operators throttled HPCI and RCIC systems as RPV level was restored to the normal band of about 196 inches. Even though the condenser high pressure annunciator had alarmed before the crew was cognizant of the loss of gland seal steam pressure, they were able to recover vacuum with the aid of the mechanical vacuum pump and gland seal steam. The operators reestablished RWCU within the first hour at the direction of the NASS to minimize the amount of RPV thermal stratification.
The November 1992 event occurred on the evening shift and control room panel actions were performed by only the CRNSO and the 603 panel NSO. They did not encounter the distractions the other crew had with burned-out indicator lamps, the smell of smoke, and additional plant staff in the control room during the recovery.
2.3 Time Line of August 13,1993 Event ,
The following sequence of events was developed from interviews with the on-duty control room personnel, copies of the control room logs, and plant computer pritatouts. All times are Eastern Daylight Time.
5
J TIME EVENTS i a.m.
9:46:14 The reactor building NPPO unseated an instrument calibration / vent valve which caused the following:
- Reactor water level high turbine trip .
Reactor scram on turbine control valve fast closure .
RFP trip
173.4 in. due to void collapse on increase RPV pressure and loss of ;
feedwater. This actuated a second RPS trip.
- Division I torus /drywell pressure recorder locked up.
9:46 31 603 panel NSO placed the reactor mode switch in shutdown, with the concurrence of the NASS, and identified about 70 multiple lamps, not -
illuminated on the full-c re display including six rod insertion indications completely dark. -
9:46 + Tagging Center NSO entered the controls area from the NSS office and initiated SBFW, with the concurrence of the NASS. Open indication on SBFW discharge valve failed (burned out lamp).
9:46:45
- RPV level 2 setpoint was reached and caused the following actuations: 4
- R.eactor recirculation pumps tripped.
- LPCI loop selected "B" loop.
- HPCI/RCIC auto initiated and injected into the RPV.
- RWCU isolated with other group 10 and 11 systems.
9:47:02 RPV level 2 (110.8 in.) cleared (lowest level reached was 105 in.). ;
9:48 + Control room crew smelled smoke, and the CRNSO and Tagging NSO began an investigation.
9:48:27 RPV Level 3 alarms cleared.
6
h 9:49:11 RPV Level 8 reached and tripped HPCI/RCIC turbines.
9:50:08 Main generator output breaker tripped on reverse power.
i 9:50:12 Condenser high pressure annunciator alarmed (gland seal steam supply in ' i manual control resulted in gland sealing steam loss following the turbine trip).
Not acknowledged by crew.
9:50:55 SBFW discharge valve auto closed on a non-safety related RPV high level '
signal.
t 9:55:00 While verifying group isolations, operators noticed amber indicator expected for group 14 and 16 was not illuminated. j 1
10:02 Operators reset alternate rod insertion and scram. 603 panel NSO reduced CRD flow to 15 gpm.
10:02:28 Bypass steam system low fault (low condenser vacuum) signal closed the ;
bypass valves. l 10:03:26 Group 1 MSIV isolation on low condenser vacuum. j 10:05 Division 1 RHR placed in torus cooling mode. l 10:06- SRV "F" cycled for pressure control. Nine valves cycled between 10:06 and 11:46 to equalize torus heat loading (sequence: F,J,M,E,D,L,R,C,P).
10:14 SBFW was started at 500 gpm to maintain RPV water level.
10:15 NSS discussed emergency plan classification with STA and declared an Unusual Event because of the HPCI injection to "he RPV. .
f 10:30 + RFP seal tank overflowed because drain pump failed to start automatically.
10:33 HPCI and RCIC systems returned to standby.
10:44 NASS directed a relief shift NSO to start recording torus water temperatures and panel operator to place Division II RHR system in torus cooling mode.
11:13 NASS directed 603 panel NSO to return RWCU to service (to verify temperature AT <145 F)in preparation for restarting recirculation pumps.
(The NSS intentionally gave RWCU restart a low priority.). An instrument technician rebooted and returned Division-I torus /drywell pressure recorder toservice. !
l 7
i
r 11:25 NASS determined technical specification 3.4.6.1 RPV cooldown rate limit of 100 F/ hour had been exceeded (bottom head temperature was available after RWCU flow was established).
11:31
- NASS authorized 601 panel operator to override SBFW Level 8 auto isolation by holding down open pushbutton to increase RPV water level to 230 in. to establish maximum natural circulation and reduce thermal stratification.
- PCRMS was restored, and higher than normal radiation levels were measured in the drywell area.
11:34
- NPPO reported a steam leak in the turbine building basement from a SBFW drain valve and stopped the leak by tightening valve.
- BOP battery charger 2PCI tripped; NASS directed operators to place 2PCI battery on spare charger.
11:46 Operators reopened first MSIV.
11:49 All MSIVs open (open indication on 7 of 8 not illuminated due to burnt out lamps).
11:50 EDG 11 low air pressure annunciator received. Entered technical specification 3.8.1.1. Portions of the air start system had been isolated at the beginning of day shift.
p.m.
1:15 Plant conditions were considered stable by the NASS and the EOPs were exited. Efforts continued to decrease RPV AT to restart recirculation pumps. ,
b 7:10 NSS exited Unusual Event classification.
8:04 Reactor recirculation pump "A" restarted.
8:24 Reactor recirculation pump "B" restarted.
2.4 Analysis Human performance aspects of this event are identified and discussed in the following sections. The analysis focuses on the actions taken by the control room staff in response ,
to the spurious RPV Level 8 signal, the subsequent turbine and reactor trips, and the MSIV isolation. However, other human performance aspects outside of the control ;
room are also discussed.
8 j
l4 m.- - -
2.4.1 Training ;
The gland seal steam system had been operated manually for approximately 4 years. ,
However, the simulator training that operators received modeled the gland seal steam -
system as it was originally designed to operate in automatic mode. The valve lineup of .
the gland seal steam system in manual operation mode was " common knowledge" among -
the operators but there was no abnormal or emergency transient training that reflected the existing manual operation lineup. The operators did not seem to have had an awareness of the impact of the gland seal steam system lineup and its loss following a turbine trip. No simulator training demonstrated the effects of operadng the gland seal steam system in manual after a turbine trip. (Additional information on the gland seal steam system is presented in Appendix A and in Figure 3 [a-d].)
Simulator training was conducted with the minimum complement of two operators (ROs) and a control room supervisor (SRO). The licensee's training philosophy was that it is most appropriate to train the minimum number of operators so that, in cases with only two operators, they were trained and have the skills needed to handle the event. The usefulness of this training approach was demonstrated in the November 1992 event, which took place during the evening shift when there were two ROs, the NASS (SRO),
and the STA.
However, in the August 1993 event, when more than two ROs participated in the control ,
room board activities, without multiple-operator training, some operational ineffectiveness resulted, as discussed in the command, control, and communication section below. This may suggest that at least some training with more than the minimum -
complement of operators may help maintain a more orderly command structure and avoid some confusion during an actual emergency.
Operators recalled that thermal stratification was briefly discussed in classroom training.
However, simulator scenarios did not last long enough to reach the operational implications of thermal stratification; training scenarios were reported to be last about 1/2 hour to I hour. As a result, the operators interviewed had not develop a sensitivity to the effects of emergency core cooling flow, RPV water level, and restoration of RWCU on RPV thermal stratification. The fact that the simulator model did not have the capability to produce thermal stratification contributed to their lack of sensitivity.
(Additional information on thermal stratification is presented in Appendix B.)
SBFW, HPCI, and RCIC were used at maximum flow and allowed to automatically trip the system and stop injection flow to the RPV. CRD flow was throttled at 17 minutes ,
after the reactor trip. In contrast, the November 1992 crew throttled flow as the level increased and controlled the level within the normal band using the flow controller tape setting. During the interviews, it was established that the latter procedure is the method taught in training. Higher RPV level (i.e., more cold makeup water) may have contributed to the amount of thermal stratification developed and the excessive RPV cooldown rate.
9
The RWCU was restored some 30 to 60 minutes earlier in the November 1992 event.
From the interviews, this was attributed to the knowledge gained by the NASS in writing the RWCU recovery procedures. Emergency or recovery procedures, as written, did not prioritize the reinitiation of RWCU flow, nor did simulator training stress it because of .
its lesser importance in major events. As demonstrated by this event, the RWCU system is important to monitoring RPV conditions for thermal stratification and cooldown. A delay in restarting this system from uncomplicated scrams, when forced flow is lost, may also significantly contribute to the development of thermal stratification. It should be noted that the probability of the loss of forced flow is significant considering the number of automatic trips and the non-safety grade electric power supply for the' reactor recirculation pumps.
Simulator fidelity to actual operations is important to good operator performance. The gland seal steam system operation was not accurately reflected in simulator training. '
The question of the most appropriate number of operators to train as a crew in -
simulator emergency scenarios is important. The current philosophy of two-operator ,
crew training assumes the minimum complement as a " worst case," but then operators do !
not receive exposure to operating with more personnel at the panels. The licensee also :
identified that the simulator did not properly model the reactor coolant inventory ;
decrease for a simulation of this event, and has initiated work to model the reactor t coolant inventory decrease more accurately. >
2.4.2 Teamwork, Command, Control, and Communications :
The shift supervision command structure was observed overall; the NASS retained direct control of control room operations and the NSS maintaining the overall supervisory position. The operators clearly turned to the NASS for immediate direction. The NASS became the procedure reader once the emergency procedures were entered. However, i two related issues of command, control, and communication warrant further discussion.
The first issue is the number of control panel operators on shift and how they coordinated their emergency operation roles. In this event, three reactor operators were on shift who were near the control panels after the turbine and reactor tripped. The 603 i panel reactor operator and the CRNSO were in the control room " horseshoe" area (see l Figure 2). The third, tagging NSO was in the NSS's office when the first alarms were i
heard. The CRNSO, who had RO responsibility for the control room operation, and the '
tagging operator both reached the control panel areas where water level controls were located at approximately the same time. The tagging NSO took the SBFW controls (on panels 601 and 602). It was apparent that control board activities were not clearly -
assigned by the responsible RO (i.e., the CRNSO), but were partially assumed by the l first RO (i.e., the tagging NSO) at the 601/602 panel. The CRNSO then coordinated a search for a smokey, acrid smell that might have indicated a fire, although no fire or
- other smoke source was ever definitely identified. Although all three operators were l licensed, qualified, and on shift, there was a moment when suddenly the CRNSO was not carrying out the emergency activities for which he had trained. This can be characterized as a disruption of expected activity, where some uncertainty then existed in i 10
the mind of the CRNSO as to what he should be doing. Most of the actions he thought to take were being done automatically or by someone else. This disruption of activity contributed to ineffective monitoring of all panels, particularly the 804/805 panels where the balance of plant controls and displays such as the gland seal steam system and the
~
main condenser vacuum were located.
The second issue is that of effective use of available personnel. After the reactor scram' l occurred, operations (e.g., the relief crew) and management personnel arrived at the control room. The NASS assigned operators to carry out specific tasks. However, there is a question of the most effective span of control for a supervisor. Individuals will differ in their span of control ability. There will be a trade-off between span of supervisory control and workload of limited numbers of personnel individually carrying out needed tasks. The more personnel available, the more tasks can be accomplished in a shorter amount of time, but have to be accomplished in a controlled manner.
The tagging NSO requested quiet at one point because the noise level in the control room prevented fully hearing a radio communication. Noise has been identified in past events as a contributing factor to poor communication.
It is interesting to compare the command, control, and communication of this event to that of the November 1992 event. The November 1992 event occurred at about 8:30 p.m. in the middle of the evening shift when two ROs, the STA, and the NASS were in t the control room. All other personnel that were normally present during the day were not present at that time. These were the conditions for which the crews had trained, and this contributed to monitoring and actions being carried out in an effective manner.
2.43 Shift Technical Advisor The STA was present in the control room at the beginning of the August 1993 event.
During the early moments of the event, the STA preferred to check indications on the control board instead of reading plant information from the computerized safety parameter display system. The NASS directed the STA to verify all level 2 and 3 actuations had occurred properly as part of his early assessment. Later, he was involved in technical discussions during the recovery with the NASS and NSS, particularly those concerned with thermal stratification during plant recovery. The NSS directed the STA to independently verify the emergency plan classification before it was implemented. .
The STA is assigned to and trained with the crew.
2.4.4 Procedures Operators used the procedures as written. Several specific areas were not addressed in ,
the procedures, which, if they had been, could have assisted the operators in plant recovery. ,
i 11 ,
The manual operation of the gland seal steam system was not addressed in plant '
procedures used after a reactor or turbine trip. Either simulator training or alarm ;
monitoring could have acted as a barrier to have overcome this deficiency, but did not.
5 As a corrective action in response to this event, the reactor scram and the turbine trip abnormal procedures have been temporarily revised, until the gland seal steam system is !
modified to operate fully automatically, to address maintenance of gland seal pressure. ;
An additional step in these revised procedures is intended to alert the operators to the j
potential loss of the gland seal steam system and require a specific action to verify gland seal steam pressure. In retrospect, if this procedural step had been present in the i abnormal procedures for this event, it is probable that the direction to check gland seal ;
steam pressure would have identified the problem and prevented the loss of main condenser vacuum and MSIV closure. This temporary procedural revision will ensure l proper operational attention to the manual gland seal steam system configuration. l i
In this case, the gland seal steam system will still be able to be operated in manual after !
the system is properly automated. A gland seal steam pressure verification step may still be valuable for several reasons. The operators have little time to act to retain condenser -
vacuum and it was been shown that the white (less important) " low gland steam pressure" I
alarm has been overlooked twice during actual events.
There were no plant-specific or generic BWR procedures that addressed minimizing )
RPV thermal stratification with a prioritized establishment of natural circulation restart !
of RWCU, and throttling of CRD pumps following a scram with loss of forced circulation. Despite the existence of multiple industry documents on thermal ,
stratification including GE SILS 251 (1977),203 Supplement (1984), AND 430 (1985); an industry group document (1992); and NRC IN 93-62 (1993), plant procedures had not
(
been modified at the time of the study.
l The NASS identified several automatic actuations that were not on the list of RPV Level 2 automatic actuations that was used to ensure their restoration. A complete list l of all equipment that was automatically actuated would assist the operators and l supervisor in ensuring that all equipment was restored to its normal (or desired) !
configuration during plant recovery. This deficiency prompts a number of unanswered questions about why this incomplete equipment actuation list was not fed back from :
previous reactor trip restorations or simulator drills, or identified in post trip reviews. l Was the list not used in simulator drills? Is feedback 'of such issues encouraged? Is the ,
scope of post trip reviews sufficiently broad to cover such human performance issues? !
2.4.5 Stress i
- All operators reported an elevated response, stress, to the unexpected alarms and turbine ;
and reactor trips. For most operators, this was their first reactor trip at high power. The i unexpectedness of the event and uncertainty of what had caused the trip kept the l operators' stress levels elevated, particularly during the first few minutes of the event. l t
12
.- r
4 This may have contributed to some initial ineffective operator response, especially in the turbine panel area. j
. A number of d istracti ons 'h t at occurred during the event were reported. These- !
distractions added to the activity level of the operators as well as the number of things of I
which the NASS needed to remain cognizant. Although the distractions themselves did -
not appear to detract from operator performance, the distractions required additional task activity, including using extra personnel to check the distractions, which increased the demand on the supervisor's tracking activities and the operators' using alternative - i indications. Such distractions and additional activity most probably contributed to j operator stress. ,
2.4.6 Human Machine Interface The control boards appeared to adhere to good human factors engineering principles.
The general approach had been to design the control boards as mimics, with flow lines connecting the equipment representations, valves, and indicators. 'Ihe mimic flow lines had directional arrows imbedded within the lines to aid in the representation of the plant. ,
Distractions that occurred during the event included about 100 light bulbs on the control boards being burned out. The operators reported that this necessitated the use of other indications several times to discern the information needed. For example, an indicator lamp for a valve being open was burned out, but the operator could verify flow existed ,
through the associated line. After SRV C was cycled, the closed indicator lamps did not ,
light on the control board but the closed position was indicated on the Emergency :
Response Information System Computer. Although the number of ourned out control board lamps were not reported as anything other than as frustrating distractions, this- '
occurrence illustrates the importance of multiple indicators and positive indication of equipment status. The repeated use of alternative indications takes some additional time and interrupts normal operational actions. ;
Another reported distraction was the lock-up of the Division I recorder for post event ,
drywell pressure monitoring instrumentation (wide and narrow range). Because the digital display showed nonsense characters, the operators knew that accurate and reliable data were not available from this display and strip chart recorder. Data would have been available if the operators had known how to reset the recorder (data on the recorder are primarily used for post-event analysis). Later, an instrument and control technician showed the operators that the reset button was underneath the paper strip.
Reset buttons can be positioned in readily visible and accessible places and operators can be informed of how to reset such control room equipment, if it is determined that it s would be detrimental for the operators to lose access to this information.
As the event continued, thermal stratification development within the RPV was a concern. However, instrumentation for evaluating actions related to cooldown rate and thermal stratification were not available. Temperature indication was available for the j i
l 13
)
.i top of the vessel, but once the RWCU was isolated, the operators did not have '
instrumentation for monitoring the vessel bottom head temperature. Only after the fact were determinations made of cooldown rate and thermal stratification. 1 The apparent initiating action for this event was the inadvertent movement of a valve handle located at an instrument rack within the plant. This instrument rack contained ;
sensitive instruments, and a yellow metal railing had been placed in front of the instrument rack to keep people and extraneous items away from the rack. In this case, the equipment operator's attention was drawn to a pressure display because he was interested in maintaining knowledge of current plant status. Close to the display was the valve handle with duct tape residue attached to it. In an attempt to carry out housekeeping duties, the equipment operator removed the tape from the handle. The yellow railing did not serve as a sufficient warning not to touch the sensitive equipment.
As a corrective action, the licensee has installed a caution sign on the instrument rack that reads:
' CAUTION: DO NOT TOUCH THIS RACK CONTAINS SENSITIVE ^
REACTOR INSTRUMENTATION FOR DIV 1 AND DIV 2. NSS APPROVAL '
REQUIRED FOR ANY ACTIVITY."
Licensees at other facilities have surrounded sensitive instrumentation with locked cages !
and other barriers to eliminate reactor scrams. .
- 3. LESSONS LEARNED / HUMAN PERFORMANCE STUDY A comparison of this event with the prior loss of feedwater event of November 1992 suggests the importance of obtaining both positive and negative human performance information in a lessons-learned approach. The licensee performed a human performance enhancement system (HPES) study on the November 1992 event.- The focus of that study was the event leading up to the reactor scram; the study did not incorporate lessons learned regarding positive control room recovery actions. The operators' statements taken after the November 1992 event did not mention the "near-miss" of loss of main condenser vacuum. Neither the operators nor the post-scram evaluation team identified this as a human performance lesson learned. If the time line of the study had been extended, to include an evaluation of" things that went right,"
lessons learned regarding the importance of gland seal steam system and.the potential loss of condenser vacuum during reactor and turbine trips, and priority given to early restoration of RWCU may have been disseminated to training and operations, and probably would have prevented some aspects of the August 1993 event.
A more effective licensee post-trip review and human performance investigation of the .
November 1992 event could have yielded important lessons learned. A sensitivity to l human performance issues where personnel performed well could have identified those 14
[;
- 7.- . ,
near misses. The result could have been identification of good personnel actions to mitigate those potential incidents and human performance areas where improvements -
could have been achieved.in training, procedures, human-machine interface, etc.
i f
f 5
5 5
I P
W 15
Nuclear Shift Supervisor SR0 - 8 years Nuclear Assistant Shift Technical Shift Supervisor Advisor - 2 years SRO - 3 years Non-Licensed Control Room Nuclear Panel 603 Nuclear Supervising Operator Supervising Operator R0 - 1 year R0 - 6 years Tagging Center Nuclear Supervising Operator R0 - 4 years Note: The years given indicate the number of years the individual has held the level of license indicated.
Figuar 1 Enrico Fermi 2 control room stainng i
16
I Oland seal steam fut core dsplay \ 603 system alarm MSIV A HPCI 602 U 804 j RNCU e condonser I vacwm atami i
.I i
RCIC SRv 601 805 l SBFW !
NSO Desk !
l 811 ;
4 I '
l 810 807 809 808 816 817 i
Note: During the first 15 minutes, the attention of the crew was directed primarily towards panels 601,602, and 603.
Figure 2 Sketch of enrico Fermi 2 control room (Not to scale) 17
4
, r ,________..., ,
s I '!
' j forrcom GuwO sttaan 8 8
' sssicM ,
s' A e !
1r i I gg 3 t F450 1 -, e cIG = j3 _s____, ;
- , U
'~~J g PCP-KOOS e
g co.ireou.cc
?
j M-l><3 FGol F601
'N F431 X )4 FGod F455 #f I wtnte h-3 cast cuMP ouwp Imota
. re,- iso e eso vatug Hextr.es ce v.we vat.ve wur.
i M kJ LJ V3 VM F605 FSOS
- i ouw czcw EwASS BwpAss 5 vawc VALVE' (a) Low power operation as designed 6
,_ q~_ _
-_-...)
8 3 ;
Ifrf0M GLAMO STtano 8 A' sssics i B' L L I sh t YNt gg l $
N=4t 27 .
C448 g; .!__.. ,
YWJ U e
{ PCP-KOOS '
} CoarT Rou.r.E s
j MFGol
-[> d FSOE
'N F431 FGod F455 #f E I4MntR S.npsg.=j fzt22 DUMP Ouep INAM '
= TEsP 15 0 M 850 ~ VAtyg Hf.KTER.3 te Vauvat VALVE vau/E. .
M I kJ LJ ;
FR V3 i F605 FCOS i esco DUMP i EMASS BM PASS I VAtut vAtvc [
t (b) liigh power operation as designed j L
F
-Figure 3 Enrico Fermi 2 gland seal steam supply system ,
t 18 f et m
.+
e '
i.
y 3--- ..-... 3 8 1
' l forreos atAwo s,tua 8 3
susic.M 0 kJ l YNI gg i F4'A t 72 8 fe 3 ~'~~~~1 ,
g Pcp-goos o coercou.ce O kJ k.J kJ kJ k Y V3 Vm V3 V3
{I FGQi FSOE mette samt-J F431 emo t FGod oump F453 oown 8f r h KTra
. we iso e eso v4ent kexters r te Vatyt Vata /c VAu/C.
M FGO3 Fths esce ouw EwASS BW5%ss va w e. v uc (c) liigh power operation (pre-event) p )----.......)
l 8 l
l forrcom ctAwosssicx sw 8 I
_'s i
8 A e i
,r I '
I Ud I F4*A l 72 8 q $
32 PCP*KOOs o co.rreotwe O k kJ kl m f y FN Vm Vm 6c
' ?
4 FG01 FGot F431 FGod F453 f I Wstte e m-J EcG 2 Dump ounp FNKItR I
. %p e50 M tSo VAud; h& TEES !
te Vatut VAWE 6 4 i VGO3 YGOS eene ouwp EwASS Bstws3
- va w e. vAvt i
(d) liigh operation (postcent) i Figure 3 Enrico Fermi 2 gland seal steam supply system (cont l) 19 i
. m I
I r
- p STEAas DevE R
^
t
-4 4 4 f \ s ii 4
-\. UZ'.?R*:51 Water Level ACACTGR WATER LEVEk L
STE Au stPAmatom 2 l iqg s
i n-FES Dw AT8 8 w '
fTYPtCAL o el JET euano N.EELE :
/
t' ~ W U / !//ti r ti!s t!/ / f tttit titif ?l e.
'h/ - \h/ f ~~'
... .. . na- ,,w n t- m ., :
co uso,,o., .ca o
$2( - fig',afylf,;'oa
=<ciacco..suct.o ~
{ ~
qI
> L xe ....
] ,.
TJ."oi'T!*.?
i N/ .
A coo ousmo
! I e
v
~ cus::::::.:.". ~~- ,
- "M%E"J."?!"
Figure 4 Enrico Fermi 2 reactor vessel level ;
2.0 ,
?
k.
+
-O APPENDIX A DESCRII'rION OF GLAND SEAL SYSTEM OPERATION l
7 9
i e
e
+
4-DESCRIPTION OF GLAND SEAL SYSTEM OPERATION The turbine gland seal steam system serves two distinct functions. At low power conditions, the system prevents air in-leakage to the turbine cylinders and steam valve glands. At high power (>75 percent), the system prevents steam leakage to the atmosphere from the turbine. ,
Pressure inside the turbine casing during low flow conditions is about the same as the vacuum in the main condenser. These conditions require a sealing mechanism to -
prevent air flow into the turbine casing and condenser while ejector pumps remove air to maintain the vacuum. Sealing steam is supplied from the manifold through two pressure regulating valves (see Figure 3 (a)). These " reg" valves open sequentially in response to a controller to supply enough steam to maintain a positive pressure of 2 psi on the packing supply line.
Above 75 percent power, the packing line pressure is maintained by a steam supply taken from the high pressure turbine inlet loop piping. This supply is applied inside the turbine casing and flows out of the turbine in the packing line. As pressure increases with higher steam flow in the turbine, excessive packing line pressure is regulated by a dump valve [see Figure 3 (b)]. A controller modulates open the dump valve to maintain '
packing line pressure.
At Fermi Unit 2, a design problem was identified where the single gland seal steam controller was supposed to operate three different size valves. A single satisfactory proportional band could not be found to obtain the desired response from all three valves to satisfy low and high power operation. Because of the mismatch of the two control modes / proportional bands, operators manually controlled packing line pressure by operation of the control bypass motor-operated valves F603 and F605, with the control valves isolated [see Figure 3 (c)].
On restart following the event, the gland seal steam system was modified by selecting a controller proportional band allowing automatic response of the regulating valves. At high power, the regulating valves are closed but will open if packing line pressure decreases. Packing line pressure still has to be maintained by manual operation of valve F605 [see Figure 3 (d)].
i i
A-1
a
.c
,i APPENDIX B ,
TIIERMAL STRATIFICATION-b i
i
-h k
THERMAL STRATIFICATION G.E. believes that thermal stratification occurs in boding water reactors following a reactor scram with loss of forced circulation and isolation of the RWCU system. The term " thermal stratification" is used to describe the concentrating of cold water in the :
lower regions of the RPV when only natural circulation flow is available. As early as 1977, General Electric distributed information concerning control of bottom head ,
temperatures and stresses associated with reactor pressure vessel ATs. G.E. believes that natural circulation flow is insufficient to mix the relatively cool CRD, HPCI, and RCIC injection flow with water already in the reactor pressure vessel. CRD flow may be the primary contributor to thermal stratification. Large ATs between the core exit and the bottom head area in excess of the 145* F prohibit the restart of recirculation pumps. In-this condition, the only mechanism available to establish flow in the bottom head is a 2 inch line going to the RWCU system (see Figure 4). Temperature of the RPV bottom head is inferred from the water temperature measured in the RWCU drain line. This measurement is valid only when RWCU flow is established.
Automatic isolation of the RWCU in many water level induced scrams removes a means of avoiding thermal stratification, as well as the only available bottom head temperature indication. By design at Fermi Unit 2, an RPV level 2 signal trips the reactor ,
recirculation pumps and isolates groups 10 and 11 which contains the RWCU isolation [
valves. CRD flow following a scram also temporarily increases, further aggravating the -
situation before the scram is reset, although most of this repressurization flow does not find its way into the RPV.
The magnitude and duration of thermal stratification is dependent on operator action.
The operator can reduce CRD flow after the scram has been reset. RWCU system can .
be restored to reestablish flow in the bottom head and regain temperature indication.
('Diis is addressed by Steps 12 and 19 of the Fermi 2 Reactor Scram Abnormal Procedure.)
A comparison of the two events found that the measured RPV AT decreased much more rapidly during the August 1993 event. During the November 1992 event, the RWCU was restored within the first hour after the reactor trip and the RPV AT was measured at i 223 F. The crew throttled HPCI/RCIC flow to minimize cooldown. With the RWCU' suction lines to the RPV and recirculation loops open at 3 hours3.472222e-5 days <br />8.333333e-4 hours <br />4.960317e-6 weeks <br />1.1415e-6 months <br /> into tSe event, the RPV AT was measured at 285* F. The RPV water level was raised above Level 8 to ensure natural circulation 6 hours6.944444e-5 days <br />0.00167 hours <br />9.920635e-6 weeks <br />2.283e-6 months <br /> after the reactor trip. ,
During the . August 1993 event, HPCI/RCIC injected 15,000 gallons in 2 minutes after the reactor trip and SBFW injected about 52,000 gallons in 1-3/4 hours using automatic trips to maintain RPV level. The RWCU was restored about 1-% hours after the reactor trip. The RPV water level was raised above Level 8 to ensure natural circulation 2 hours2.314815e-5 days <br />5.555556e-4 hours <br />3.306878e-6 weeks <br />7.61e-7 months <br /> after the reactor trip. With the RWCU suction lines in the same configuration as the November 1992 event, the AT was initially measured at 251* F at 1-3/4 hours B-1
.4 . ,
w;
- (initially higher than in the November 1992 event) and at 180" F at 3-3/4 hours into the ;
event (substantially lower than in the November event). ,
The licensee has requeste'd that General Electric and Westinghouse analyze the ,
mechanism, timing, and extent of RPV thermal stratification using data from these j events.
i I
6
.7 P
5 i
i
?
J .L .
i f
I t
I s
B-2 l i