ML16148A603
| ML16148A603 | |
| Person / Time | |
|---|---|
| Site: | Oconee  |
| Issue date: | 01/14/1992 |
| From: | Belisle G, Binoy Desai, Harmon P, Poertner W NRC OFFICE OF INSPECTION & ENFORCEMENT (IE REGION II) |
| To: | |
| Shared Package | |
| ML16148A602 | List: |
| References | |
| 50-269-91-34, 50-270-91-34, 50-287-91-34, NUDOCS 9202140232 | |
| Download: ML16148A603 (20) | |
See also: IR 05000269/1991034
Text
UNITED STATES
NUCLEAR REGULATORY COMMISSION
.
REGION H
a
101 MARIETTA STREETN.W.
&1
ATLANTA, GEORGIA 30323
Report Nos. 50-269/91-34, 50-270/91-34 and 50-287/91-34
Licensee:
Duke Power Company
422 South Church Street
Charlotte, NC 28242
Docket Nos.: 50-269, 50-270 and 50-287
License Nos.: DPR-38, DPR-47 and
Facility Name:
Oconee Nuclear Station
Inspection Conducte
November 23 - December 21, 1991
Inspectors:
.
/. /4
P. E. Harmo ,
enio
es den
spector
Date Signed
de
Bp
esaDate
Signed
W. K. Poertne/
Resid itIns
c or
Date Signed
Accompanying Inspectors: J. J. Lenahan, Reactor Inspector, Region II (RH),
Division of Reactor Safety (DRS)
L. S. Mellen, Reactor Engineer, RH, DRS
R. B. Shortridge, Radiation Specialist, RII,
Division of Radiation Safety and Safeguards (DRSS)
E. D. Testa,
Em
gency Preparedness Specialist, RII,
Approved by: G.A./1te,
a
Signed
Division of Reactor Pr jects
SUMMARY
Scope:
This inspection was conducted to follow-up on the primary Reactor
Coolant System (RCS)
leakage that occurred on Unit 3 on November 23
and 24,
and the resulting loss of approximately 87,000 gallons of
reactor coolant water to the reactor building.
Results: Two violations were identified, concerning the failure to follow
procedures (paragraph 2.c) and inadequate installation and inspection
procedures to ensure proper installation of a compression fitting
(paragraph 3).
9202140232 920114
PDR ADOCK 05000269
Q
2
The evolution was well controlled by the Technical Support Center
(TSC).
Numerous
problems were encountered and solved in a
conservative, safety-conscious manner.
During the actual plant
shutdown, operatcr actions led to an inadvertent Reactor Protection
System actuatio
which had no direct consequences.
Operators
exhibited sound judgement, knowledge of plant, safety consciousness,
and an overall appropriate conduct of responsibility with one
exception noted above.
The licensees radiological response was
prudent and effective.
Emergency response activities were
satisfactory.
REPORT DETAILS
1. Persons Contacted
- H. Barron, Station Manager
- J. Davis, Quality Assurance Manager
- G. Rothenberger, Superintendent, Integrated Scheduling
- R. Sweigart, Superintendent, Operations
C. Little, Superintendent, Instrument and Electrical(I&E)
W. Foster, Superintendent, Mechanical Maintenance
Other licensee employees contacted included technicians, operators,
mechanics, security force members, and staff engineers.
NRC Resident Inspectors:
- P. Harmon
W. Poertner
- B. Desai
- Attended exit interview
2. Event Description
a. Event Overview
Prior to the event, the unit was at 100 percent power with RCS
temperature and pressure being controlled at 580 degrees F and
2155 psig respectively.
The core was at 232 effective full power
days burnup (complete cycle being 410 effective full power days).
One of the three high pressure injection (HPI) pumps was in service
maintaining pressurizer level.
The other two HPI pumps were in
standby with both trains of.the HPI system operable. Both trains of
the low pressure injection (LPI)
system were operable and in the
standby.mode.
Containment integrity was maintained and the reactor
building spray and the reactor building cooling systems were
operable. All offsite power supplies were operable.
At the time of the event there were indications of defective fuel
pins and the licensee had been monitoring the Iodine(I) concentration
in the reactor coolant.
Based on concentration of 1-131, 132, 133,
134 and 135, the licensee had projected that. there were potentially
eight fuel pins with small defects.
Reactor coolant activity was
higher than normal as a result of the defective fuel pins.
The Inadequate Core. Cooling Monitor (ICCM)
instrumentation was
installed at Oconee as part of a post Three Mile Island (TMI)
modification.
This modification had been completed on Unit 3 during
the refueling outage that ended in March 1987. A portion of the ICCM
includes the "A" and the "B" hot leg level indications. The hardware
for hot leg level instrumentation consists of two impulse lines (one
2
for each loop) that taps into the high point vent line located at the
apex of each hot leg.
This line feeds the low pressure side of a
differential pressure (dp) cell.
The high pressure impulse. line taps
into the decay heat drop line.
A RCS pressure transmitter serving
the ICCM subcooling margin monitor as well as providing control room
indication also taps into the impulse line feeding the high pressure
side of the "A" hot leg level dp cell.
Prior to the event, the
control room RCS pressure indication (off ICCM)
was reading the
normal value of 2155 psig.
Both trains of ICCM were .operable;
however, the hot leg level indications were not valid (normal) due to
forced flow conditions.
At approximately 1:41 a.m.
on November 23,
1991, operators in the
Unit 3 control room received alarms indicating problems with the ICCM
"A" train as well as fire alarms in the reactor building.
The
operators noted that the letdown storage tank (LDST)
level and
pressurizer level were decreasing.
The operators also noted that
reactor coolant makeup through the pressurizer level control valve
and reactor building normal
sump level were increasing.
With
symptoms of an RCS leak, the operators entered the Abnormal Procedure
for excessive RCS leakage.
The RCS leak was estimated by the
operators at approximately 70 gpm. AT 2:14 a.m., an ALERT emergency
classification was declared and per the emergency plan, the Technical
Support Center (TSC),
the Operations Support Center (OSC)
and the
Crisis Management Center (CMC) were activated. A controlled shutdown
and a subsequent depressurization of the unit was initiated.
Based
on the fault on the ICCM "A" train, as well as the failure to zero of
a wide range
RCS' pressure gage,
the source of the leakage was
suspected to be the hot leg level portion of the Reactor Vessel Level
Indication System (RVLIS)
instrumentation connection located at the
apex of the "A" hot leg (candy cane). . The RCS was completely
depressurized at approximately 5:00 p.m.
on November 24, at which
time the leak was stopped.
After purging the reactor building,
personnel entered the building and positively identified the source
of the leak.
A
compression fitting, 3/4 inch in diameter,
associated with the RVLIS hot leg level instrumentation, had failed.
Approximately 87,000 gallons of reactor coolant leaked from the
failed compression fitting within a time frame of approximately
40 hours4.62963e-4 days <br />0.0111 hours <br />6.613757e-5 weeks <br />1.522e-5 months <br />.
The RCS leakage was caused by the failure of an improperly installed
compression.fitting. Some difficulty was experienced in cooling down
and fully depressurizing the RCS to stop the unisolable leak.
Containment atmosphere samples detected radioactive iodine and noble
gas concentration as high as 1380 times the Maximum Permissible
Concentrations (MPC).
Radiation dose to the public as well as to
plant personnel as a result of the accident was minimal.
3
After the reactor shutdown was completed and the reactor containment
was purged, the licensee walked down all compression fittings in the
Unit 3 Reactor Building and found a high percentage had not been
properly made up.
All accessible fittings were either tightened to
manufacturers'
specifications, or analyzed as acceptable for those
which could not be brought into conformance.
b.
Sequence of Events
Date/Time
Event
November 23, 1991
0100
Units 1, 2, and 3 operating at 100%.
0120
RCS leakage calculation performed for the last hour
indicated 0.18 gpm.
0141
Operators received alarm indicating problem with ICCM
"A" train and also received fire alarms in the reactor
building.
0143
Operators noted Letdown Storage Tank (LDST) level
decreasing, Pressurizer. level decreasing,
reactor
coolant makeup increasing, and reactor building normal
sump level increasing.
Operators referred to
AP/3/A/1700/02, Excessive RCS Leakage.
0155
Radiation Instrument Alarm (RIA)-47 (Reactor Building
Particulate) alarmed high.
0203
Operators initiated a controlled shutdown of the unit
at 15 MW/min.
Reactor coolant leak estimated to be
70 gpm.
0214
Emergency classification of ALERT declared based on
RCS leak greater than 50
gpm.
Site Assembly
commenced. Leak estimated at 60-70 gpm.
0217
With unit power at 78%,
Integrated Control System
(ICS)
Asymmetric
Rod indication resulted in unit
runback to 60%.
Operators placed SG/Rx Master and Rx
control in Manual and inserted control rods to reduce
power.
0218
Site Assembly requested
0221
NRC, State, and Local Counties notified.
4
0231
OSC Activated
0239
Received reactor building High Iodine alarm on RIA-48.
0250
Site Assembly Complete
0300
Unit at 48%, control rods being manually inserted to
achieve 1% per minute decrease.
0305
Technical Support Center (TSC) established.
0325
RCS leak estimated to be stable at 60-70 .gpm.
0327
Reactor trip occurred at approximately 35% power due
to a divergent steam pressure oscillation.
0330
Crisis Management Center (CMC) established.
0339
Reactor building sample results indicated Iodine at
2 MPC, and Noble Gas at 407 MPC.
0530
RCS at 1735 psig and 535 degrees F, Hot Shutdown
Conditions.
0629
RCS Chemistry results received; boron concentration
933 ppm.
0635
Operators placed RPS channels A, B, C, D in shutdown
bypass.
0638
Operators reset control rod drive (CRD) breakers. Turbine
Bypass Valves (TBV)
opened due to shift in Main Steam
header bias setpoint, initiating an inadvertent rapid
cooldown of the RCS.
Main Steam and reactor coolant
pressure began increasing from about 1620 psi when
operators manually closed turbine bypass valves.
0639
Turbine Bypass Valve Controller placed in Automatic.
0641
RPS actuation as reactor coolant pressure increased to
1700 psi (no control rod movement as rods were already
at the bottom).
Operators subsequently reset the CRD
breakers and withdrew group one rods to 50% per
procedure. Cooldown and depressurization continued.
1427
Unit cooldown reached 250 degrees F.
1550
LPI pumps A and B started for decay heat removal.
1718
Unit cooldown reached 200 degrees F (Cold Shutdown).
5
1720
ALERT terminated and Crisis Management Center
deactivated (leak less than 10 gpm).
2054
Reactor Coolant Pump 3B1 (final pump) secured.
2156
Pressurizer cooldown initiated.
November 24, 1991
0002
5,000 gallons. pumped from reactor building to the
miscellaneous waste.holdup tank for sampling,
processing, and eventual release.
0800
Leak estimated to be 5 to 10 gpm.
1700
Leak stopped after unit completely depressurized.
Final depressurization involved opening pressurizer
relief valves to allow colder RCS loop water to enter
the pressurizer and collapse the -remaining steam
bubble.
High point RCS hot leg vents were opened,
allowing higher elevation hot leg water to drain down
and enter the pressurizer. Although the "A" hot leg
level dropped as expected, the "B" loop hot leg hung
up because the vent arrangement did not provide
effective venting into the RCS.
The "A" hot leg was
effectively vented inward by the failed instrument
tube.
NOTE:
On Novembver 24-25, Emergency Notification
Updates and Event News Releases were given as required
by established procedural controls.
November 25, 1991
0100
Reactor Building Purge initiated.
1300
Damage assessment team entered reactor building and
identified failed RVLIS compression fitting.
11/27/91
Completed pumping water from reactor building
basement for processing.
11/28/91
Reactor building decontamination activities in
progress.
11/30/91
RVLIS valve and tubing sent to B&W for analysis.
Commenced inspections of valves, penetrations,
instruments, etc. in reactor building.
6
12/08/91
All inspections in the reactor building
complete. Unit 3 in hot shutdown with RCS at 536
degrees F and 2190 psig. Startup activities in
progress. While pulling Rod Group 5 to achieve
criticality, Rod 6 of Group 5 dropped into the core.
Shutdown (cooldown and depressurization) was
commenced.
12/09/91
A reactor building entry confirmed that the stator
rod 6 of group 5 was bad. Remeggering of.control rod
drive stators initiated.
12/11/91
26 of 69 stators did not meet meggering acceptance
criteria.
12/12/91
RCS cooldown and depressurization initiated to drain
RCS to 200 inches on the pressurizer to enable work on
12/13/91
Reactor vessel head vented, 29 stators pulled
for troubleshooting and drying.
12/14/91
RCS fill begun, 4 stators replaced, all 29 stators
dried, 24 connections replaced.
All stators
remeggered.
12/15/91
Small RCS leak noticed during a building walkdown.
The leak was identified as a crack in the decay heat
drop line.
12/16/91
Shutdown.Cooldown commenced to isolate and repair
leak.
c. Event Details
Resident inspectors arrived on site shortly after the initiation of
the event.
Control room activities as well as TSC/OSC activities
during recovery were monitored on a "round the clock" basis by the
resident staff. In addition, inspectors interviewed plant personnel,
reviewed logs, attended meetings, as well as observed portions of
activities in the reactor building associated with the event.
The
following information was gathered from the above inspector
activities.
Prior to the event, on the morning of November 23,
1991, Units 1,
2 and 3 were operating at 100 percent power. Unit 3 was staffed with
one unit supervisor (SRO), one control room supervisor (SRO) and two
reactor operators.
A RCS
leakage calculation performed at
approximately 1:20 a.m. on November 23, indicated a RCS leakage of
approximately 0.18 gpm which was
well within the Technical
Specification (TS) limitations.
7
At 1:41 a.m.,
operators in the Unit 3 control
room received a
statalarm indicating problems with the ICCM "A" train and also
received statalarms indicating a fire in the Reactor Building.
Shortly thereafter, the operators noticed that the vertical panel
board indicated that the letdown storage tank (LDST) level as well as
the pressurizer level were decreasing and that the reactor coolant
makeup and reactor building normal sump level were increasing. LDST
level decrease is attributed to increased makeup flow as the high
pressure injection pump draws suction from the LDST during normal
operation.
The operators immediately recognized the symptoms of
excessive RCS leakage and referred to the abnormal procedure for
excessive RCS leakage, AP/3/A/1700/02.
Per the instructions of
AP/3/A/1700/02, efforts were initiated to determine the leak size.
At approximately 1:55 a.m.,
Radiation Instrument Alarm (RIA)-47,
which monitors reactor building particulate activity, alarmed high.
By 2:03 a.m.,
the RCS leak rate calculation was completed and the
leak rate was estimated to be 70 gpm.
This leakage exceeded the
TS 3.1.6 limits of unidentified leakage greater than 1 gpm as well as
total RCS leakage greater than 10 gpm. A unit shutdown was initiated
with a load reduction of 15 MW/min. Operators also isolated letdown
to minimize further inventory loss.
At this time, there was no
conjecture as to the location of the leak.
At 2:14 a.m.,
an ALERT emergency classification was declared, per
Enclosure 4.1.1 of Emergency Classification Procedure RP/0/1000/01,
based on an unisolable RCS leak greater than 50 gpm and subcooling
margin greater than 0 degrees F. .A site assembly was also commenced
per the -requirements of RP/0/A/1000/03 which was the applicable
emergency procedure while in the Alert emergency classification.
Notifications were made to activate the TSC, OSC and the CMC.
During the controlled shutdown, while at 78 percent power, the unit.
experienced an ICS runback to 60 percent power due to an asymmetric
rod caused by indication of a safety rod in Group 3 not at its out
limit. However, all rods were verified by the operators to be at
their required positions.
The cause of the runback was later
determined to be a failed out limit indication and not an actual
dropped rod.
The runback was terminated at 60 percent power by
taking the Reactor Diamond Control to manual.
The operators
continued to reduce power by manually inserting the control rods. By
3:05 a.m.,
the TSC and the OSC had been fully activated and the
Station Manager
had assumed responsibility as
the emergency
coordinator. The unit was at 42 percent power and manual shutdown
was continuing at the rate of 1 percent per minute.
Field
monitoring teams were dispatched to identify any changes in the
offsite dose rates.
8
At 3:24 a.m., with the unit at 35 percent power, operators placed the
3B Main Feedwater pump (MFWP) in manual in anticipation of securing
the pump.
This resulted in a divergent steam pressure oscillation
due to an inherent. instability of turbine pulser control at low
power.
The slower response of the feedwater system in manual
initiated the divergent steam pressure oscillation.
As the speed of
the pump was reduced, the discharge feedwater pressure reached a low
enough pressure to trip the Reactor Protection System (RPS)
trip
bistable for the 3B MFWP on RPS channels A and D.
To dampen the
divergent steam pressure oscillation, operators tried to increase the
speed of the 3B MFWP. This caused the speed of the 3A MFWP (which
was in auto) to decrease.
The 3A MFWP reached the low discharge
pressure setpoint of approximately 800 psig, tripping the RPS
channels A and'.D trip bistables.
Since the bistable trip caused by
the 3B MFWP was already locked in, the two MFWPs were now perceived
by the .RPS to be tripped.
Therefore, the RPS actuated on loss of
both MFWPs and the unit tripped. The post trip response of the plant
was normal with reactor coolant leak still in progress.
All control
rods were inserted into the core.
Both 4 KV and.7 KV electrical
power supplies transferred to the startup source. Unit 3 stabilized
at hot shutdown. No Engineered Safeguards (ES) or pressurizer relief
valve actuations occurred.
Both MFWPs continued to pump into the
steam generators, removing decay heat through the turbine bypass
valves and into the condenser. "A" MFWP was later secured by the
operators.
The licensee's review of the operators'
actions
regarding the feedwater pump shutdown and control of the ICS
concluded that operator familiarity with various combinations of
auto/manual station controls should be increased.
Increased training
in this area is planned by the training staff.
By 3:30 a.m., the Crisis Management
Center
(CMC)
had been
established.
The unit was stable at hot shutdown. RCS leakage at
this time was estimated at approximately 130 gpm. This leak rate was
suspected to be in error on the high side due to the inherent large
uncertainties, associated with leak rate calculations during transient
conditions.
Reactor building sample results indicated that the
concentration of radioactive Iodine in the building was two times the
maximum permissible concentration (MPC) and noble gas was 407 times
MPC.
Field monitoring teams did not detect any offsite increase in
activity. Operators noted through the reactor building video camera
that a significant amount of steam was rising from the "A" cavity.
The steam was condensing on virtually all visible walls, handrails
and equipment.
No significant rise in reactor building pressure was
noted.
A boron concentration sample of the RCS,
requested earlier by the
TSC,
indicated the RCS concentration to be 579 ppm.
At 4:45 a.m.,
cooldown of RCS to 532 degrees F was begun. Further boration of the
RCS was initiated to ensure adequate long-term shutdown margin.
Contents of the bleed hold up tanks, as well as the concentrated
.9
boric acid mix tank were constantly pumped into the LDST to keep up
with the inventory loss.
By the controlling procedure for unit shutdown, OP/3/A/1102/10, at
approximately 5:30 a.m.,
the ES system was bypassed to enable
lowering RCS. pressure to below the ES high pressure setpoint of
1750 psig.
Further cooldown was in progress by controlling secondary
pressure using the ICS turbine header pressure setpoint control with
the turbine bypass valves in the automatic mode of operation.
Shutdown Procedure OP/3/A/1102/10, Enclosure 4.2, step 2.3, requires
the,%turbine bypass valves to be placed in Manual.
The CRO, after
consulting with the CRSRO,
decided to leave the TBVs in Auto
and
control the TBVs by manually lowering the demand signal.
The
operator and the CRSRO believed this method of control to be
equivalent to placing the controller in Manual.
The method chosen
was based on a desire to reduce operator duty and attention during
the transient.
However, the consequences of leaving the controls in
Auto were not properly analyzed or recognized.. *When the reactor trip
breakers are reset, the ICS automatically removes a bias that
increases the setpoint by 125 psig.
If the TBVs are in auto, the
removal of the bias results in the TBS-sensing a pressure error of
125 psig. At approximately 6:33 a.m., the RPS was placed in shutdown
bypass to allow the RPS system not to actuate while RCS pressure was
reduced below the low pressure trip setpoint. An over-pressure trip
setpoint of 1710 psig RCS pressure is automatically reinstated by
this evolution to prevent inadvertent re-pressurization when the
reactor is reset.
At 6:38 a.m., the control rod breakers were reset in preparation for
partially withdrawing one group of control rods as a standby source
of negative reactivity. When the CRD breakers were reset, the ICS
removed the 125 psig steam header pressure automatic bias applied to
the turbine header pressure setpoint following a reactor trip.
As a
result the turbine bypass valves opened to achieve the new lowered
turbine header setpoint. This created a cooling transient on the RCS
and the RCS pressure dropped to approximately 1620 psig.
The
operator took the turbine bypass valves to Manual and started driving
them closed creating a reduction in heat removal from the primary
side. RCS pressure and.temperature started increasing. RCS pressure
reached the over-pressure setpoint and tripped the RPS.
The CRD
breakers opened. However, all control rods were already in the core
so no other consequences occurred.
RCS pressure continued to
increase to approximately 1720 psig when the operator took the. bypass
valve back into Manual and reopened them to stabilize pressure.
Subsequently, the operators reset the CRD breakers, withdrew one
group of control rods to 50 percent in accordance with the shutdown
procedure and continued the cooldown.
10
During the transient, the operator and his supervisor chose to not
perform a step required by the relevant procedure.
While performing
steps out of sequence or changing the procedures is allowed after
careful analysis and review, the transient imposed by the procedure
deviation indicates that full cognizance of the ICS functions and the
rationale for the step requiring placing the TBVs in Manual were not
adequately considered. This is a violation for failure to follow
procedures, VIO 287/91-34-01: Failure -to Follow Procedures..
At approximately 7:30 a.m.,
the RCS was at 1500 psig and 480
degrees F. The operators noted that intermediate nuclear range
instrument NI-4 had failed low.
Samples obtained from the reactor
building indicated Noble gas concentrations of 1023 MPC and Iodine
and particulate concentrations of 970 MPC.
-No offsite dose
indications were found. RCS leakage was approximately 50 gpm.
Cooldown continued and at about 3:30 p.m.,
Low Pressure Injection
(LPI) pumps A and B were aligned to remove decay heat. At 5:18 p.m.,
cold shutdown was reached and subsequently the Alert was terminated.
An
RCS leak rate of 10-15 gpm still existed.
The leakage was
suspected to be in the A hot leg RVLIS instrumentation as the CR
indication had failed low.
HP personnel were sampling the reactor
building every four hours.
Unit 3 reactor building samples taken at
9:15 p.m.
showed Iodine concentrations at 1386 MPC.
Noble gas
concentrations at 1416 MPC,
particulate concentrations at .84 MPC
and tritium concentrations at less than 1 MPC.
By 2:00 a.m., on November 24, 1991, operators had started pumping the
reactor building sump water into the waste holdup tanks located in
the Auxiliary Building for processing. The airborne Iodine activity
in the reactor building was estimated at 1382 MPC.
By 8:00 a.m.,
the RCS leak was still estimated at 5 to 10 gpm, and the pressurizer
was at saturation temperature and at 30 psig.
This pressure was
being maintained by the temperature and saturated conditions in the
pressurizer. Although all pressurizer heaters had been turned off,
the hot pressurizer continued to hold pressure above atmospheric.
The normal cooldown process requires personnel entry into the reactor
building to align valves to establish flow through the pressurizer to
cool it. However,
the airborne activity in the reactor building
precluded the normal cooldown path. The dose rate to the skin from
the airborne concentration of noble gas in the reactor building was
estimated at 321 .mrem/hour. A special procedure was written to vent
the pressurizer and the hot legs to reduce the hot leg levels.
This
involved opening the pressurizer relief valve to vent the pressurizer
to the Quench Tank, allowing the colder water from the loops to enter
the pressurizer. To enable the inrush of cooler water to enter.the
pressurizer and cool it down,
the high point vents on the hot legs
were opened to drop the loop levels.
The "A" hot leg dropped since
the hot legs are higher than the pressurizer, and an effective vent
into the RCS existed through the break. The
"B" loop did not drop,
however,
because an adequate vent is not available without a
containment entry.
Kerotest valves in the vent line act as
stop-checks and prevent flow from atmosphere into the loops, and the
vent line configuration can allow loop seals to develop.
Hot leg
venting problems are a known phenomenon at Oconee.
The hot legs are
very difficult to vent inwards,
and pressurized nitrogen into the
vents is routinely used to positively vent the hot legs as the levels
are dropped.
As soon as the "A" hot leg level was dropped, the leak
stopped.
By 5:00 p.m.,
November 24,
1991, the unit was completely.
depressurized and the leak stopped.
Approximately 87,000 gallons. of contaminated water had accumulated in
the reactor building basement.
Plans were established to remove,
treat, and discharge the radioactive water as well as to perform a
filtered/controlled release of the radioactive concentrations of the
reactor building air.
On November 25, at 1:00.a.m.,
the reactor building purge was started
to reduce the activity for building entry.
At 1:00 p.m., airborne Iodine activity was reduced to 27 MPC.
An
inspection team entered the reactor building and identified the
location of the leak. The inspectors witnessed the containment entry
on the video monitor in the control room. The leak was confirmed to
be from a 3/4 inch diameter instrument line that had pulled out of a
compression fitting just downstream of the instrument root valves for
the A hot leg RVLIS.
Within this area, one violation was identified.
3. Engineering Review
The configuration of the instrument line was as follows:
A 1 inch
diameter pipe is attached to the apex of the hot leg pipe.
At about
12 inches above the hot leg there is a tee junction.
One portion of the
tee goes to the high point vent system and the other portion, the line of
interest, is the line associated with the RVLIS instrumentation.
Downstream of the tee is a 3/8 inch diameter restrictor followed by a
3/4 inch isolation valve.
A 3/4 inch adapter is butt welded to this
isolation valve.
Then a 4 inch piece of stainless steel tubing is
attached to this adapter using a Parker/Hannifin compression fitting. A
series of tubing reducers were used to transition from the 3/4 inch tubing
to the 3/8 inch tubing which resulted in a total of six compression joints
per instrument line. At the end of the 3/8 inch tubing is a dp cell.
The
3/4 inch tubing was found to have been pulled out of the 3/4 inch tubing
adaptor at the valve, causing the RCS leakage.
The leakage was limited by
the 3/8 inch ID restrictor located just downstream of the tee and upstream
of the root valve. The root valve, fittings and tubing were subsequently
removed from the system for inspection and analysis.
12
Analysis of the tubing/fitting concluded that the tube pulled out of the
compression fitting due to inadequate engagement of the ferrule to the
tube. This conclusion was based on:
-
The gap measurement between the nut and the fitting was found to be
0.182 inch versus a nominal gap of 0.153 inches specified by the
manufacturer.
The manufacturer stated that the nominal valve did not
constitute acceptance criterion or specifications,
but rather
provided indication of proper installment.
-
Though the ferrule had been installed in the correct orientation,
visual examination of the tubing identified that the ferrule had not
been compressed onto the tubing.
An analysis performed by B&W indicated that, based on the shallow axial
score marks,
the ferrule had not bitten deeply into the tube.
In
addition, the crimping of the tube made by the ferrule was much less than
expected. The inside diameter of the tube was not appreciably deformed,
confirming improper crimping or swaging.
The outside diameter of the
tubing met specification requirements.
Based on the above,
the B&W
analysis concluded that improper makeup of the. fitting joint caused the
tube fitting to detach.
The ferrule had not been compressed enough to
hold the tube in place and allowed system pressure to force the tube out
of the ferrule and fitting. The licensee concluded that the compression
fitting nut was not tightened enough.
Oconee uses both Swagelock and Parker/Hannifin compression fittings.
Both manufacturers provide installation instructions which specify that
fittings should be installed "finger tight," then tightened an additional
number of turns depending on the tubing's outside diameter.
Scribing the
nut to positively indicate the: number of turns past finger tight is
recommended.
Swagelok provides specific acceptance criteria for the gap
between the nut and the fitting and provides a go/no-go inspection gage to
assure that fittings were properly tightened.
While installation
technicians are trained in this requirement, procedures for installation
di.d not contain instructions to reinforce the requirement.
The failed
fitting and the high number of other deficiencies, discussed below,
indicated that skill of the craft was inadequate as a means of ensuring
the manufacturers'
requirements were met in this area.
The installation
procedure, TN/3/A/32401/00/AK1, consists of a single sentence:
"Install
3/4 inch SS tubing, 3/4 inch x 1/2 inch union,
1/2 inch SS tubing, 1/2
inch x 3/8 inch union and 3/8 inch SS tubing between the root valve at the
Reactor Vessel Head and the 3/8 inch compression fitting between the tee
connection bellows sensor xxxx (bellows number specific to instrument).
Reference drawings xxxx."
Skill-of-the-craft as to how to actually put
fittings, unions tubing and ferrule together was implicit.
Manufacturers'
requirements,
recommendations and precautions were not specified in the
installation procedures.
After the single sentence installation
13
procedure, a sign-off space is provided for the individual performing the
installation (craft), a second space for an independent verifier (IV),
and
a third space for Quality Assurance (QA).
The completed procedures
reviewed by the inspector had initials in each -of the required spaces,
but
there are no requirements or acceptance criteria to any of the three
parties signing each space what is expected for that signature.
inspection procedure QAE2,
Instrumentation Installation, Modification and
Maintenance Inspections,
Rev.
14,
governs the inspection.by the QA
individual for the case at issue.
The specific instructions to the QA
inspector consist of a single item in Section 5.7.c. requiring "For QA-1
and QA-3, verify that all fittings are tight."
"Tight" is a relative term
and could mean anything past finger tight.
10 CFR 50, Appendix B, Criterion V, Instructions, Procedures and Drawings
requires, in part, that for activities affecting quality, "instructions,
procedures
and drawings shall include appropriate quantitative or
qualitative acceptance criteria for determining that important activities
have been satisfactorily accomplished.
10 CFR 50, Appendix B, Criterion
X, Inspection, requires,
in part, that "If
inspection of processed
material or products is impossible or disadvantageous, indirect control by
monitoring processing methods, equipment, and personnel shall be provided.
Both inspection and process monitoring shall be provided when control is
inadequate without both."
Attributes of these requirements are not
contained in the records of the installation of the fittings for the RVLIS
modifications.
The inadequate procedure which resulted in an improper installation of a
compression fitting is identified as VIO 287/91-34-02: Failure in Quality
Assurance Process to Detect an Improper Installation of a Compression
Fitting.
The root valve tubing and associated fittings on both the "A" and "B" hot
legs RVLIS instrumentation were replaced with a modified design.
The
licensee inspected a sample of other Parker and Swagelok fittings located
in the reactor building. Approximately ten percent of the fittings were
found to be out of nominal range.
Consequently the licensee performed an
inspection of all similar fittings on the RCS and HPI system.
This
inspection included 455 fittings (191 Swagelok and 264 Parker/Hannifin).
Of these, twenty eight percent were found out of nominal range.
The
licensee attempted to tighten all of these fittings; however, 23 fittings
were left out of range as they could not be further tightened into nominal
range without use of excessive force.
The licensee performed an
.
engineering evaluation and concluded that it was acceptable to leave the
23 fittings out of range.
The licensee also inspected a broad range of components in the reactor
building due to exposure to humid atmosphere during the event.
These
included:
-
4 of 11 Environmentally Qualified (EQ)
transmitters were opened and
visually inspected. No evidence of water intrusion was found.
14
-
8 of 8 Target Rock EQ solenoid valves were successfully cycled and
tested.
26Limitorque limit switch housings were inspected. No moisture was
noted.
-
All 22 Fire detectors were checked..
One was found bad and was
replaced.
-
5 electrical penetration junction boxes were opened. No moisture was
evident.
-
2 Pressurizer heater junction boxes were inspected and 3 cables were
meggered.
No defects were found.
-
Sample of Incore instruments were checked.
No problems were
identified.
-
5 resistance temperature detectors on RCS were checked and were found
to respond satisfactorily.
-
The power operated relief valve acoustic leak monitor was visually
inspected.
-
2 of 4 reactor coolant pumps
(RCP)
were Hi Pot tested.
Oil was
changed on 1A2 RCP when.water was found in the upper oil pot.
In addition, 20 of 69 CRDs were meggered. The initial resistance values
obtained were within the tolerance limits. However, during startup, just
prior to criticality, a control rod in group 6 dropped into the core.
As
a result, the licensee initiated shutdown and decided to remegger all 69
CRDs.
Of these,
29 CRDs did not meet the acceptance criteria.
The
dropped rod's stator indicated a phase to phase fault.
It is postulated
that the heatup caused the intruded moisture to vaporize and affect the
resistances in the CRD mechanisms.
The 29 CRDs were purged with nitrogen
and remeggered.
Four new CRD stators were installed and 24 cable
connectors replaced.
Startup activities were initiated.
During a reactor building tour, a
small leak from the decay heat drop line was noted. The unit was cooled
and depressurized in order to accommodate work to change out portions of
piping in.the decay heat drop line. Details of the leak will be provided
in a future report.
At the end of this inspection period, the unit was in cold shutdown with
repairs on the decay heat dropline in progress.
Within this area, one violation was identified.
15
4.
Operational Review
Upon receiving indications of a fire in the reactor building as well as
indications of increased makeup to the RCS, increase in reactor building
sump levels, and decreasing pressurizer levels, operators in the Unit 3
control room correctly diagnosed the problem as being excessive RCS
leakage.
They began a rapid controlled shutdown, and correctly declared
an ALERT emergency classification.
Some difficulties were experienced at
35 percent power and a reactor trip occurred.
The trip -was due to
divergent steam pressure oscillations and was unavoidable due to the short
period of the oscillations.
Another difficulty experienced during
shutdown was caused by the failure of operators to adhere to a step in the
shutdown procedure. The operators exhibited sound judgement, knowledge of
plant arrangement and safety consciousness.
With one exception noted
above, all procedures were followed.
Communications between the control
room and TSC and within the control room were carried out satisfactorily.
Emergency as well as operating procedures were adequate to diagnose,
respond to and mitigate the event.
Within this area, no violations or deviations were identified.
5. Radiological Review.
The failure of the compression fitting on Unit 3 resulted in approximately
87,000 gallons of reactor coolant leaking from the instrument line.
The
leakage was contained within the reactor building. Activity levels in the
reactor building due to Iodine and noble gases were also significantly
elevated following the event.
High dose rates as well as contamination
levels were produced in portions of the reactor building. The licensee's
response during the event was both prudent and effective. The licensee
correctly assessed the root cause of the event and took steps to ensure
plant personnel's and the public's radiological safety.
Reactor.coolant from the leak through the instrument line had accumulated
in the basement of the reactor building. This was gradually pumped to the
Miscellaneous Waste Holdup tanks
(MWHUT)
located in the auxiliary
building. The MWHUT has a capacity of approximately 15,000 gallons. From
the MWHUT,
the reactor coolant was pumped to the Feeder tanks located in
the Radwaste facility for processing. From the Feeder tanks, the coolant
was channeled through a set of filter/demineralizers to reduce the
activity levels to below regulatory requirements. The processed coolant
was then held up in waste monitor tanks where it was sampled.
After
ensuring acceptable levels of radioactivity, the contents of the waste
monitor tanks were released to Lake Hartwell.
The licensee started
transferring the effluent from the reactor building at 11:58 p.m.,
on
November 24, 1991. The releases were completed by 5:15 p.m.,
November 28,
1991.
16
Airborne radioactivity concentrations were high, particularly Iodine, due
to the leaking fuel pins as well as the unit trip from 35 percent power.
The reactor building purge system was used to lower airborne activity in.
the reactor building. The reactor building purge system is a once through
ventilation system which filters reactor building air through High
Efficiency particulate air and carbon adsorber filters prior to release
through the unit vent. The purge.exhaust system was.started at 1:08 p.m.,
on November 25, 1991.
Concentrations of Iodine and noble gas ranged
downward over the 25th and 26th of November. During this release, neither
the licensee's nor the. state of South Carolina's environmental survey
teams detected any increase of radioactivity in the airborne environment.
On November 26,
1991, with Iodine concentrations between 7.12 and 12.55
MPC, fourteen licensee personnel made entry into the unit 3 reactor
building. The survey results obtained by Health Physics were as follows:
RADIATION DOSE RATES (mR/hr)
Area
High Contact (1")
General Area (18")
1st floor
100
5-8
2nd floor
60
5-15
3rd floor
5-20
4th floor
700 10-300
"A" cavity
-at leak
450
150-200
-West stairs
150-200
"A" cooler
120
10-60
"D" cooler
160
"C" cooler
120
CONTAMINATION LEVELS
1st floor -
10,000 cpm/100 cm2 up to 2 mR/hr gamma;.09 MRAD Beta
2nd floor -
2,000 cpm up to 66 mRrad Beta
3rd floor - 19.6 mrad Beta up to 135 mrad Beta; 6 mR/hr gamma
4th floor -
5-40 mR gamma; 2,440 mrad Beta
"A" cavity (at leak) -
180 mR gamma; 6,420 mrad Beta
17
All personnel wore a full set of protective clothing, wet suits, .and self
contained breathing apparatus (SCBAs) during the entry.. Upon exiting at
5:10 p.m. on November 26, no personnel were found to be contaminated after
removing the protective clothing.
The licensee performed decontamination of the reactor building in
conjunction with repair of the leak and investigation for damage to.
components.
The target levels of 5,000 to 15,000 disintegrations per
minute per 100 centimeters square (dpm/100 cm2) were not achieved. After
using high pressure spray equipment in various areas,
smearable
contamination levels ranged as high as 1 to 10 mrad Beta on the floors and
up to 100 mrad Beta in the area of the leak.
During heatup and
pressurization on December 2, 1991,
Iodine levels began to increase.
Iodine concentration increased from 12 times MPC. to 45 times MPC on
December 4, 1991.
On December 5, 1991, just prior to criticality the unit
experienced a rod drop caused by a problem with a stator.
During
cooldown, Iodine and noble gas levels decreased to less than 1 times MPC
upon entry.
The licensee calculated the dose to public due to the releases.
The
numbers indicate the dose that a person would get if the person were to be
present.at the site boundry (approximately 1 mile) for the duration of the
release.
.
SOURCE/CURIES
DOSE TO WHOLE BODY
DOSE TO THYROID
Liquid .0305 Ci gross
.00139 mrem
.0122 mren
.0165 Ci Tritium (.015% of annual limit) (.0407% of annual
limit)
.293 Ci Noble Gas
Noble Gas 672 Ci
.00218 mrad
N/A
(.00727% of annual limit)
Iodine Gas
.0004 mren
(estimated)
(.114 percent of annual
limit)
Within this area, no violations or deviations were identified.
6. Emergency Preparedness Review
An ALERT emergency classification was declared as a result of the RCS
leak. A site assembly was performed; the TSC, OSC and the CMC were also
activated.
The inspector reviewed the emergency response activities
associated with the recognition, declaration, initial and followup
notifications, and activation of the emergency response facilities.
The
licensee used approved procedures and emergency response plan.
Trained
18
personnel made timely recognition, classification, initial notification,
and followup notifications.
Site assembly.was completed in a timely
manner and a field team was appropriately dispatched. Cellular telephone
compensated for radio communication difficulties with the field team.
The
licensee's emergency organization responded and performed in a
satisfactory manner.
Within this area, no violations or deviations were identified.
7. Exit Interview
The inspection scope and findings were summarized on December 23,
1991,.
with those persons indicated in paragraph 1 above.
The inspectors
described the areas inspected and discussed in detail the inspection
findings listed below.
No dissenting comments were received from the
licensee.
The licensee did not identify as proprietary any of the
material provided to or reviewed by the inspectors during this inspection.
Item Number
Description/Reference Paragraph
VIO 287/91-34-01
Failure to Follow Procedures (paragraph
2.c)
VIO 287/91-34-02
Inadequate Installation and Inspection
Procedures (paragraph 3)