ML20099L274
| ML20099L274 | |
| Person / Time | |
|---|---|
| Site: | Vogtle  |
| Issue date: | 09/06/1995 |
| From: | GEORGIA POWER CO. |
| To: | |
| References | |
| OLA-3-A-171S, NUDOCS 9512200270 | |
| Download: ML20099L274 (9) | |
Text
.+e.
4 a
.e 2
..-u be-A Aa
.m
-J-
,_.4 a.
J 4
4__.s 1A4 m4 A
-_A
+-- - _ i..
n-n
~
GPC EXHlITF li l'/1-5 WEB 8 EX. c.19 DOCKETED USNRC TITLE Mg
/- pg -
g '95 gy,[3 :/e x,vde.
OFFICE OF SECRETARY DOCKEflNG & SERVICE
/
BRANCH
/
Mgr. Engr. Supt.
Date Mgr. HP/ Chem.
Date d Y&
/ Cfnl9n
/
"gigr. y ch. Sd>t.
~Date -
Mgr. Maint.
Date i
e e
/
/
Mgr. Admin.
Data Mgr. Ops.
Date 4
/
/
l Mgr. Trng.
Date Mgr. Out./ Plan Date 3
/
/
i Asst. General Mgr., - Support Date Asst. General Mgr. - Operations Date g
y+dckdasa-90-B4
&l2ll90 PRB Mtg. No.
Please review, sign where appropriate and return to I 4d/1 by If there are any questions, please call TI~
at extension
.7 / 05.
s I
)k NUCLEAR REGULATORY COW 488 TON Docket No. 5424/425 0LA-3 EXHl81TNO. 0 E 'IT-O f 5 '
j in #* maner of Geornia power co. et at ye unho 1 a 2
(
0 8 toff pil Applicant O lntervonor O r,.w
! (
I
.'@k8'n#W BW O Rejected Reporter SD c.,De di-o' 4C Wimes.
Q,W
(
~.
n-~
q,
%. 4. i.].9$ 2. U,Dj:ly5 e-u.:..se:s. 3.f'.-
u.
9512200270 950906 e'r+ :
92 PRCUECT 057419 PDR ADOCK 05000424; PDR o
. Joy'
s et[
LER 1-90-6 LOSS OF OFFSITE POWER LEADS TO SITE AREA EMERGENCY EVENT DATE: 3-20-90 ABSTRACT on 3-20-90, Unit 1 was in a refueling outage and Unit 2
was operating at 100% power.
At 0820 CST, the driver of a fuel truck in the switchyard backed into a
support for the phase "C"
insulator for the Unit 1 Reserve Auxiliary Transformer (RAT) 1A.
The-insulator and line fell, causing a phase to ground fault.
Both Unit 1
RAT 1A and Unit 2 RAT 28 High Side and Low Side breakers
- tripped, causing a loss of offsite power condition (LOSP). Unit 1 Diesel Generator (DG) 1A and Unit 2 DG28
- started, but DG1A tripped, causing a loss of residual heat removal (RHR) to the reactor cora since the Unit 1 Train 8 RAT and DG were out of service for maintenance.
A Site Area Emergency (SAE) was declared and the site Emergency Plan was implemented. The Reactor Coolant System heated up to 136 degrees F from 90 degree F before the DG was emergency started at 0856 CST and RHR was restore 3.
The initial notifications wara not made within the required 15 minutes due to the loss of power to the Emergency Notification Network (ENN).
At 0915 CST, the SAE was downgraded to an Alert after onsite power was restored.
The direct cause of this series of events was a cognitive personnel error.
The truck driver failed to use proper backing procedures and hit a support, causing the phase to ground fault and LOSP.
The most probable cause of the DG1A trip was the intermittent actuation of the DG Jacket water temperatur.
switches.
Corrective actions include strengthening policies for control of
- vehicles, extensive testing of the DG and replacement of suspect DG tamparature switches, and improvements in the ENN system.
9 1
92 PRaECT 057420 L
_... ~ _. - _ _ ___ ____
d '
- A.. REQUIREMENT FOR REPORT i
' This event '1s reportable pers a) 10 CFR50.73 (a)(2)(1v), because 3'
an unplanned Engineered Safety Feature (ESF) actuation occurred s
1 when the - ESF Actuation System Sequencer started, and b) Technical l
Specification-d.8.1.1.3, because a valid diesel generator failure occurred.
Additionally, this report serves as a summary of the l
}
Site Area Emergency event.
I B.
UNIT STATUS AT TIME OF EVENT l
d Unit 1.was in Mode 6 (Refueling) at 0% rated thermal power.
The reactor had been shutdown since 2-23-90 for a 45 day scheduled j '
refueling outage.
The reactor core reload had been completed, j
l the initial tensioning of the reactor vessel head studs was
- complete, and the outage team was awaiting permission from the l
control room to begin the final tensioning.
Reactor Coolant System (RCS) level was being maintained at mid-loop with the Train A
Residual Heat Removal (RHR) pump in service for decay l
heat removal.
The temperature of the RCS was being na1ntained at j
approximately 90 degrees F.
l 4
l Due to the refueling outage maintenance activities in
- progress, some equipment was out of service and several systems were in i
abnormal configurations.
The Train 8 01esel Generator (DG18) was out of service for a required 36 month maintenance inspection.
l The Train 8
Reserve Auxiliary Transformer (RAT) had been comoved from service for an oil change. The Train B Class 1E 4160 Volt switchgear, 18A03, was being powered from the Train A
RAT 1A through its alternate supply breaker.
All non-1E switchgear I
was being powered from the Unit Auxiliary Transformers (UAT) by I
l backfeed from the switchyard.
All Steam Generator (S/G) nozzle dams had been removed, but only S/G's 1 and a had their primary manways secured.
Maintenance personnel were in the process of costoring the primary manways on S/G's 2 and 3.
RCS level was i
being maintained at mid-loop for valve repairs and the S/G manway restorations.
In addition, the pressurizer manway was l
removed to provide a RCS vent path.
5 l.
C.
DESCRIPTION OF EVENT i
On March 20, 1990, at approximately 0817 CST, a truck driver with security escort entered the protected area in a
fuel truck.
Although not a member of the plant operating staff, the driver j
was a Georgia power Company employee belonging to a service group used to perform various plant services.
The driver checked the 1
welding machine that was in the area and found that it did not i
need fuel.
He returned to the fuel truck and was in the process i
l of backing out of the area when he hit a support holding the
{
phase "C" insulator for the RAT 1A.
The insulator and line fell causing a
phase to ground fault, and the transformer breakers
,(
tripped.
j
- (
l i
i 2
92 PROJECT 057421 i
i
3 1
{
a
'At 0820 CST, both Unit 1 RAT 1A and tho Unit 2 RAT 28 High Sido and Low Side breakers tripped causing a loss of offsite. power
-condition (LOSP) to the Unit 1 Train A Class 1E 4160 volt bus
- 1AA02, the Unit 2 Train 8 Class 1E bus 28A03, and the 480 volt l
4 busses supplied by 1AA02 and 28A03.
The Unit 1 Train B Class 1E i
4160 volt buss 18A03 also lost power since RAT 1A was feeding l
both Trains of Class 1E 4160 volt busses.
The loss of power caused the associated ESF Actuation System Sequencers to send a
start signal to one Unit 1 and one Unit 2 Diesel Generators. DG1A and-DG28 started and sequenced the loads to their respective 3
busses.
Further description of the Unit 2 response to this event is provided in LER 50-4 25/1990-002.
One minute and twenty seconds after the DG1A engine started and sequenced the loads to the Class 1E bus, the engine tripped.
This again caused an undervoltage (UV) condition to class 1E bus 1AA02.
The UV signal is a maintained signal at the sequencer.
- However, since DG1A was coasting down from the
- trip, the shutdown logic did not allow the DG fuel racks or starting air i
j solenoids to open and start the engine.
This properly caused the engine starting logic to lockup, a condition that existed until the UV signal was reset.
For this
- reason, DG1A did not automatically re-start after it tripped.
After the trip, operators were dispatched to the engine control 4
panel to investigate the cause of the trip.
According to the
- operator, several annunciators were lit.
The operator briefly reviewed several instrument read-outs and detected no immediate l(
problem. In order to restore emergency power, the operator reset the annunciators without delaying to evaluate or record the annunciators that were present.
During this
- time, a
Shif t Supervisor (SS) and a Plant Equipment Operator (PEO) went to the l'
sequencer panel to determine if any problems were present on the 1A sequencer.
The SS pushed the UV reset button, then reset the sequencer by deenergizing and energiring the power supply to the l
sequencer.
This caused the DG air start solenoid to energize for another 5
seconds which caused the engine to start.
This l'
happened 19 minutes after the DG tripped the first time.
The engine started and the sequencer sequenced the available loads as designed.
After 1 minute and 10 seconds, the breaker and the engine tripped a second time.
It did not automatically re-start
[
due to the starting logic being blocked as described above.
By j
this
- time, operators, a
maintenance foreman and the diesel i
generator vendor representative were in the DG room.
The initial report was that the Jacket water pressure trip was the cause of the trip.
This report was discounted because the maintenance s
j foreman and vendor representative observed that the Jacket water pressure at the gauge was about 12-13 PSTG.
The trip setpoint is 6 PSIG and the alarm setpoint is 8 PSIG.
- Also, the control room observed a lube oil sensor malfunction alarm.
Fifteen minutes after the second DG1A trip, DG1A was started from the engine control panel using the emergency start breakglass button.
The engine started and loads were manually loaded.
When l
k.
'the DG 1s started in the emergency mode, all the trips except j
3 92 PRCMECT 057422 i
- four cro bypsened.
- However, all alarms will bo onnunciated.
During the amargency run, no trip alarms wara noticed by. the personnel either at the control room or at the angina control panel.
The only alarms noted by the control room operator assignad for DG operation were lube oil pressure sensor malfunction and fuel oil level High/ Low alarm, neither of which would have tripped the diasal.
At 1040 CST, RAT 18 was energized to supply powan to 4160 volt bus 18A03.
OG1A supplied power to 4160 volt bus 1AA02 until 1157 CST, at which time 1AA02 buss was tied to RAT 18.
A Site Area Emergency was declared at 0840 CST, due to a loss of all offsite and onsite AC power for more than 15 minutes.
The Emergency Director signed the notification form used to inform l
offsite government agencias of the amargency at 0848 CST.
The shift clark attempted to initiate offsita notification utilizing the primary ENN in the control room but found it inoparable due to the loss of power.
The shift clerk than went to the back-up ENN and initiated notification after cell call on this system at 0857 CST.
Due to the loss of power, which rendered the primary Emergency Notification Network (ENN) inoperable, and some mis-communication, the initial notification was not received by all agencias until 0935 CST.
f The Emergency Director instructed personnel to complate various j
tasks for restoring containment and RCS integrity. All work was accomplished and maintenance personnel exited containment by l
~
1050 CST.
i The SAE was downgraded to an Alert Emergency at 0915 CST after restoration of cora cooling and one train of alactrical power. By
~1200
- CST, plant conditions had stabilized with both trains of alactrical power being supplied from an off-site source (RAT 18)
After discussions with the NRC and local government agencias, the i
amargency was terminated at 1247 CST and all agencias were notified by 1256 CST.
l D.
CAUSE OF EVENT Direct Cause:
- 1) The direct cause of the loss of offsita class 1E AC power was the fuel truck hitting a pela supporting a 230kV line for RAT 1A.
This was a cognitiva personnel acror on the part of the truck driver. There wara no unusual characteristics of the work location that directly contributed to this personnel arror.
- DG1A, to start and load the LOSP loads on buss 1AA02.
1 4
92 PROJECT 057423
/*
- 3) Tho direct causo of the failuro of tha primsry ENN cystcm in the control room was the loss of electrical power to Unit 1.
The primary ENN in the control room is powered from Unit 1
Class 1E AC power.
Therefore, when Unit 1 lost Class 1E AC i
alactrical power, the primary ENN in the control room did not work.
Root Caussa
- 1) The truck driver met all current site training and qualification requirements, including holding a
Class 2
Georgia driver's 11cansa.
- However, site safety rules, which t
require a
flagman for backing vehicles when viewing is impaired, were violated.
P
- 2) The root causa for the failure of OG1A has not baan conclusively datarmined. There is no record of the trips that were annunciated after the first trip because the annunciators were rasat before the condition was fully 4
evaluated. Therefore, the cause of the first trip can only be postulated, but it was most likely the same as that which caused the second trip.
The second trip occurred at the and of the timed sequence of the group 2 block logic.
This logic allows the OG to achiava operating conditions before the trips become active.
The block logic timed out and the trip occurred at about 70 seconds.
The annunciators observed at the second trip included Jacket water high temperatura along with other trips.
In conducting an investigation, the trip conditions that were observed on the second OG trip on 3 i 90 could be duplicated by venting 2 out of 3 Jacket water temperature
- sensors, simulating a tripped condition.
The simulation duplicated both the annunciators and the 70 sec.
trip time.
The most likely cause of the OG trips was j
intermittant actuation of Jacket water camparatura switches.
t Following the 3-20-90
- avant, all thraa Jacket water tamparatura
- switches, which have a design setpoint of 200 i
degrams F,
wara bench tasted.
Switch TS-19110 was found to have a satpoint of 197 degrees F, which was approximately 6
degrams below its previous setting. Switch TS-19111 was found to have a setpoint of 199 degrees F, which was approximately
)
the same as the original setting.
Switch TS-19112 was found to have a satpoint of 186 degrams F, which was approximately 17 degrees F below the previous setting and was re-adjusted.
Switch TS-19112 also had a small leak which was Judged to be acceptable to support diagnostic engine tests and was eminstalled.
The switches were recalibrated with the l
manufacturer's assistance to ensure a consistant calibration technique.
During the subsequent test run of the OG on 3-30-90, one of the switches (TS-19111) tripped and would not rasat.
This appeared to be an intermittent failure because it
('
subsequently mechanically rasat.
This switch and the leaking N.
switch (TS-19112) were replaced with new switches.
All i
5 92 PROJECT 057424
t oubscquent testing wse conducked with no 3dditional problems.
A test of the Jacket water system tamparature transient during engine starts was conducted.
The purpose of this test was to determine the actual Jacket water temperature at the switch locations with the engine in a normal standby lineup, and than followed by a series of starts without air telling the engine to cap 11cate the starts of 3-20-90.
The test showed that Jacket water temperature at the switch location decreased from a
standby temperature of 163 degrams F
to approximately 156 degrees F and remained steady.
Numerous sensor calibrations (including Jackat watar tamparatures),
special pneumatic leak testing, and multiple engine starts and runs were performed under various conditions.
After the 3-20-90 avant, the control systems of both enginas were subjected to a comprehensive test program.
Additionally, the Jacket water high temperature switches were i which found the switches sent to an independant laboratory, set at tamparatures ranging fqom 162 degrams F to 195 degrees F rather than the 200 degree F setting that was required. The calibration technique was changed and new switches were calibrated and installed DG1A on 5-23-90.
- However, another failure occurred (Sea Technical Specification Special Report 1-90-4.).
These switches were also sent to the independant laboratory, which found the settings to be from 164 degr.nas F to 169 degrams F.
Subsequent to this
- testing, the onsite calibration procedure was again revised to provide a
technique that is consistant with the actual. operating conditions that the switches _ experience.
Switches were calibrated using this new technique, installed and found to operate within the expected parameters.
Based on the above facts, it is concluded that tha Jacket water high tamparature switches were the most probable cause of both trips on 3-20-90.
E.
ANALYSIS OF EVENT The loss of offsite power to Class 1E bus 18A03 and the failura of DG1A to start and operata successfully, coupled with DG1B and RAT 18 being out of service for maintenance, resulted in Unit 1 being without AC power to both Class 1E busses.
With both class i
1E busses deanargized, the RHR System could not perform its required safety function.
Based on a noted cate of rise in the RCS tamparature of 46 degrees F in 36 minutes, the RCS water would not have been expected to begin boiling until approximately 1 hour1.157407e-5 days <br />2.777778e-4 hours <br />1.653439e-6 weeks <br />3.805e-7 months <br /> and 36 minutes after the beginning of the event.
Using more conservative assumptions and methods, but the same actual time of the avant, the calculated worst case time to boiling was found to be approximately 1 hour1.157407e-5 days <br />2.777778e-4 hours <br />1.653439e-6 weeks <br />3.805e-7 months <br /> and 11 minutes, and time to core uncovering was found to be approximately 11 hours1.273148e-4 days <br />0.00306 hours <br />1.818783e-5 weeks <br />4.1855e-6 months <br /> and 5 minutes.
(
This assumed no gravity feed from the RWST.
\\...
6 92 PRCMECT 057425
h/
4
- Rostoration of RHR cnd clesuro of the containment equipmhnA hatch
- were completed well within the estimated 1 hour1.157407e-5 days <br />2.777778e-4 hours <br />1.653439e-6 weeks <br />3.805e-7 months <br /> and 36 minutes
.1
'for' the projected. onset'of boiling in the RCS.
A review of information obtained from the Process and Effluent Radiation Monitoring System (PERMS) and grab sample analysis indicated all, I
normal values.
As a
result of~this
- event, no increase in radioactive releases to either the containment or the environment l
]
occurred.
' Additional systems were either available or could have been made
(
ava i la ble to ensure the continued safe operation of the plants j
returned to service approximately 2 hours2.314815e-5 days <br />5.555556e-4 hours <br />3.306878e-6 weeks <br />7.61e-7 months <br /> into the event.
~
.2) Offsite power was available to non-1E equipment through i
the generator step-up transformers which were being i
used to "back-feed" the Unit Auxiliary.
Transformers j
(UAT) and supply the non-1E busses. Provided that the phase to ground fault was cleared, Class 1E busses 1AA02 and 18A03 could have been powered by feeding through non-1E bus 1NA01.
L.
t i
- 3) The Refueling Water Storage Tank could have been used to manually establish gr.avity feed to the RCS to ma'intain a
i I
supply of cooling water to the reactor.
Consequently, neither plant safety nor the health and sa(ety of a
the public was adversely affected by this event.
i i
4 F. CORRECTIVE ACTIONS l
- 1) A management policy on control and operation of vehicles has been established.
- 2) Temporary barricades have been erected with signs which
}
directs authorization for control of switchyard traffic to j
the SS.
- 3) The Loss of'Off-site Power (LOSP) diesel start and trip logic l
has been modified on both Unit 1 and Unit 2
so that an automatic " emergency" start will occur upon LOSP.
Therefore, non-essential diesel engine trips are blocked upon LOSP.
j-Additionally, high Jacket water temperature has been deleted as a trip signal in the emergency start mode.
- 4) The DG1A test f requency was increased to 3 times a week until 4-20-90 when the test frequency was changed to once every 7
2 days in accordance with Technical Specification Table 4.8--
1.
This frequency will be continued until 7
consecutive valid tests are completed with no more than one valid failure in the last 20 valid tests.
Up to and including the
{'
two valid f ailures of the ' 3-20-90 avant, there were a total L
of four valid failures in 4+ valid tests of DG1A.
l bl 7
92 PRCMECT 057426 i
i J
!k. *
,*'. - p ' M}t.. r ' ',*:..* e.".i"
,$,:l $:.hd~~.
^
- ~.'
M
.t i ir.;.s.@ W <:r
.s -.-
- 5) Tha DG18 Jockot water temperaturo cwitchna w?k y.-oro ecplaccd cn d..'-Q-2,.
we r e c a l i b e'a t'a d 'u's'i n g. a['m'o r e 'appropriats'"}
,5,u
~
the new switches technique prior to their' i n s ta l l a t i on,.,,. J a c ke t tamparatura switches in t he other DG 's wi1:1,ba.re-cali bratedwater.;/
. *I.. ' 'If; I,y.%fe%',,' *
~3..
and/or replaced in turn.
... fit
_ ~
- 6) A back-up ENN system powered from,the".;AT&T system, which previously existed and was operational. for South Carolina
- agencias, has baan extended.to includa,Gaorgia local and stata agencias.
instructions,have baan,given
'to Emergency Directors and Communicators concerning use of',the amargancy communication systems.
-' ~~TyTE S*?.'
.~
- t.
- t a.e
.r,
e
.. 4 y
~
.m s..?
G.
ADDITIONAL INFORMATION
.t ' j
..r.
~
1.
Failed Componentsa Jacket Water High Tamparature' Switches manufactured by California Controls Company.
Model # A-3500-W3 s.
4' 2.
Previous Similar Events:
,. g,.?
Nona c
. : c.
.'."..4.,..
Energy industry identification Syst,em Code s '.Q'<, a.,..
m
~
3.
.. ~,..,,,
.f.f*.
Removal System - BP
.li",.
Lu be Oi l Sys tem '.I A "'. [*-[? L.,
Residual Heat g
l' Diasal Canarator Diasal Generator Starting Air 3 ystem - LC.;ip.
S Diasal Canarator Cooling Water ?Sy' stem
.L8 ?l4
~
System - EK' M"-
Dissal Generator Power Supply'Q, [.*,' fir.2 h. -
Safety inJaction System - BQ
.$.* '.. e -
7,,i t-.
~
13.8 kV Power System - EA
/
~ ' '
4160 volt non-1E power system '- EA -
' y[
i..-
4160 volt Class 1E power system.- EB 2
Chemical and Volume Control System - C8 Containment Building - NH 480 volt Class 1E Power System - ED Engineered Safety Features Actuation System - JE Radiation Monitoring System - IL
.,.,j,
.s
- ?'.~
s s
8 92 PFKMECT 057427
,