ML20154H011
| ML20154H011 | |
| Person / Time | |
|---|---|
| Site: | LaSalle  |
| Issue date: | 05/11/1988 |
| From: | Azab B, Ron Kopriva, Phillips L, Ring M, Shemanski P NRC OFFICE OF INSPECTION & ENFORCEMENT (IE REGION III) |
| To: | |
| Shared Package | |
| ML20154H008 | List: |
| References | |
| REF-GTECI-B-19, REF-GTECI-TH, TASK-B-19, TASK-OR 50-373-88-08, 50-373-88-8, 50-374-88-08, 50-374-88-8, GL-86-02, GL-86-2, NUDOCS 8805250195 | |
| Download: ML20154H011 (30) | |
See also: IR 05000373/1988008
Text
.
- _ _ _ _ _ _ _
<t
'
y
. . .
1:
U.S. NUCLEAR REGULATORY COMMISSION
REGION III
-
Report No. 50-373/88008; 50-374/88008
Docket No. 50-373; 50-374
Licensee:
Commonwealth' Edison Company
P. O. Box 767
Chicago, IL 60690
Facility Name:
LaSalle County Station, Units 1 and 2
.
Inspection At:
LaSalle Site, Marseilles, IL
Inspection Conducted: March 16 through 24, 1988
,
Inspectors:
NRC Augmented Inspection Team
Team Leader:
M. A. Ring
~
) /%
Date
Team Members:
R.A.KoprivD
4
5
'l
Date
L. E. Pn11l'i
C, h/
5/ 'kE
/
Date
P. Shemansk
?
- 47
Date
B. A. AzaD7h b 6
t/M
j
ate
b. J .
. $. ~
Approved By:
W. L. Forney, Ch'ief W
- /e/r .
Reactor Projects Branch 1
Date
Inspection Summary
Inspection on March 16 through 24, 1988 (Report No. 50-373/88008(DRP);
50-374/88008(DRP))
Areas Inspected:
Special Augmented Inspection Team (AIT) inspection conducted
in response to the dual recirculation pump trip and subsequent core power
8805250195 880516
ADOCK 05000373
O
-- __ _
_ - _ _ _ _ _ _ - _ _ _ _ _ _ _ _ _ _ - _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
. _ _ _ - _ _ _ _ _ _ _ .
,
t
..
4
oscillations resulting in a reactor trip on March 9, 1988, at LaSalle, Unit 2.
>
The review included root cause determination, safety significance, performance
of operators and equipment, adequacy of procedures, effects on the reactor,
reporting actions and potential generic implications.
Results: No violations or deviations were identified; however, the licensee.
has committed to procedure and Technical Specification changes as well as
further study in the areas of inherent shutdown mechanisms, instrumentation
'c'apability and uncertainties in the decay ratio calculations.
The licensee's
interim report, as required by the CAL, is included as attachment 5 to this
report.
!
,
i
(
,
I
I
t
2
.
_ _ .
_ _ _ _ _ _ _ _ _ - _
?
o
Augmented Inspection Team Report
Page*No.
I.
Introduction
1
A.
Synopsis of Event
1
B.
AIT Formation
1
C.
AIT Charter
2
D.
Persons Contacted
2
II. Description - Dual Recirculation Pump Trip Event
of March 9, 1988
3
A.
Narrative Description
3
B.
Sequence of Events
5
III. Investigative Efforts
7
A.
Synopsis of AIT Activities
7-
B.
Core Nuclear and Thermal Hydraulic Performance
8
1.
Core Performance
8
2.
Chemistry Results
10
3.
Equipment Performance
11
a.
Recirculation Pumps & Flow Control Valves
11
b.
12
,
,
c.
Power to Flow Scram
12
d.
13
e.
ATWS System
14
l
~
C.
Operator Performance
14
!
D.
Procedure Adequacy and Training
16
1.
Technical Specifications
16
2.
16
3.
Abnormal Procedures and Training
17
\\
.-_.
. - .
- . -
- - . .
c:
).
. , :.
Page No.
4.
Simulator Training
" 17
E.
Reporting
17
!
1.
Reporting Sequence
18
2.
Reporting Evaluation
20
IV. AIT Concerns and Recommendations
21
A.
Concerns
21
1.
Decay Ratio
21
2.
Technical Specifications
21
3.
Instrumentation
22
4.
Oscillation Characteristics
23
5.
Additional Concerns
24
B.
Recommendations
24
V.
AIT Conclusions
24
VI. Exit
25
,.
_ - _ _ _
_ _ _ _
._
. _ _ _ - _ _ _ _
_ _ _ _ _ _ _ _ _ _ _
_ _ _ _ _ _ . - _ _ _ _ _ . . . _ _ . _ _ _ _ _ _ _ - _ _ _ _ _ _ - _ _ _ _ _ _ _ _ _ _
- 9
(,
.-
4
Attachments
.
Attachment'No.
Descriotion
~
1
Confirmatory Action Letter (CAL-RIII-88-03)-
2
Augmented Inspection Team (AIT) Charter
3
Figures
Figure 1
BWR Power to Flow Map
Figure 2
Decay Ratio
Figure 3
Startrec Traces - Beginning of Event
Figure 4
Startree Traces - Oscillations
4.
Technical Specifications
5
Commonwealth Edison Company (Ceco)
Response to CAL item 5, dated
April 15, 1988
6
LER 88-003-00
7.
..
.
I.
INTRODUCTION
A.
Synopsis of Event
-
On Wednesday, March 9, 1988, around 5:30 p.m. CST, the LaSalle Unit
2 reactor was operating at steady state conditions at approximately
84% power.
Instrument Maintenance Department personnel were in the
process of performing an instrument surveillance when a valving
error produced a pressure pulse which actuated the instrumentation
which causes a trip of both recirculation (RR) pumps in order to
decrease power in the event of an Anticipated Transient Without
Scram (ATWS). Both RR pumps tripped causing a flow and power
decrease.
Control rods remained in the high power (99% Flow Control
Line (FCL)) position. As a result of the rapid power decrease caused
by the trip of the RR pumps, the feedwater heater level control
system was unable to control level in the feedwater heaters and
began isolating extraction steam from the heaters.
This resulted in
cooler feedwater being supplied to the reactor. Approximately five
minutes after the RR pump trip, operators observed the Average Power
Range Monitor (APRM) indication in the control room to be oscillating
between 25% and 50% power every 2 to 3 seconds. Approximately
seven minutes after the RR pump trip, as operators were attempting
to restore forced flow and making preparations to scram, the reactor
automatically scrammed on high neutron flux as seen by the APRMs.
At 6:32 p.m. CST, the licensee notified the NRC of the RR pump trip,
the loss of feedwater heating, and the resultant scram.
B.
AIT Fomation
At the time of the event on March 9, 1988, the Resident Inspector
assigned to LaSalle was offsite attending the Resident Seminar and
the Senior Resident Inspector position for LaSalle was vacant due
to a recent promotion. The initial licensee report on the event
did not discuss the flux oscillations but indicated that the event
was still being investigated. Upon further investigation and
appreciation of the magnitude of the oscillations a Region III
morning report update of the event was issued on March 15, 1988.
On March 16, 1988, an Augmented Inspection Team (AIT) was formed
which included three Region III individuals; M. A. Ring, Chief,
Reactor Projects Section 18 and Team Leader, R. A. Kopriva, LaSalle
Resident Inspector, and B. A. Azab, Reactor Safety Inspector, and
two NRR individuals; L. E. Phillips, Senior Nuclear Engineer, and
<
P. Shemanski, LaSalle Project Manager. All of the AIT members had
arrived onsite by the morning of March 17, 1988. Concurrent with
the AIT activities, Region III issued a Confirmatory Action Letter
,
(CAL-RIII-88-03) which was received by the licensee on March 17,
'
1988, and is included as Attachment 1 to this report.
The CAL
confirmed certain actions to be taken by the licensee in support of
the AIT and established conditions to be met prior to the restart of
LaSalle, Unit 2.
t
t
1
,
e
._-
. _ _
_. ..
.
.
'
C.
AIT Charter
'
On March 17, 1988, a draft charter for the AIT was formulated and
transmitted to the AIT onsite (Attachment 2 to this report).
The
general areas to be investigated were:
Sequence of events
Core performance during the event
Operator performance
Procedure adequacy
Reactor effects
Reporting
D.
Persons Contacted
Commonwealth Edison Company
- G. J. Diederich, Station Manager
- N. Kalivianakis, General Manager, BWR Operations
- D, Galle, Vice President, BWR Operations
- H. E. Bliss, Manager of Nuclear Licensing
- W. R. Huntington, Services Superintendent
- T. Rausch, Nuclear Fuel Services
- W. F. Naughton, Nuclear Fuel Services Manager
- M. Turbak, Assistant Licensing Manager
- J. Bitel, Manager Nuclear Safety
- R. J. Raguse, Production Training Supervisor
- T. Shaffer, Training Supervisor
- R. O. Armitage, lead License Instructor
- K. W. Peterman, Nuclear Fue? 5+cvices
- L. H. Lauterbach, Onsite Nuclear Safety Supervisor
- H. McLain, Onsite Nuclear Safety
- W. S. Marcus, Engineering-Site Supervisor
- J. C. Renwick, Productica Superintendent
- J.
A. Miller, Technical Staff
- M. H. Richter, Assistant Technical Staff Supervisor
- D. A. Brown, Quality Assurance Superintendent
- P. F. Manning, Assistant Superintendent - Technical Services
- T. A. Hammerich, Technical 5taff Supervisor
- A. C. Settles, Regulatory Assurance
B. S. Westphal, Operating Engineer
R. W. Stobert, Director of Quality Assurance Operations
J. A. Silady, Nuclear Licensing
M. Wagner, Dresden Nuclear Group
M. G. Santic, Master Instrument Engineer
L. W. Raney, Nuclear Safety Braidwood
R. Weidner, Production Training
J. Dedin, Production Training
R. Graham, Nuclear Station Operator
E. McVey, Technical Staff
2
-
.
-.
.- -
_.
-
.
_.
.
-_
'
\\
t
.-
t
General Electric Company (GE)
H. Pfefferlen, licensing
-
G. A. Watford, Engineering
- Denotes those attending the exit meeting on Marcil 24, 1988.
In addition, several other members of the LaSalle staff were contacted by
the AIT.
II.
DESCRIPTION - DUAL RECIRCULATION PUMP TRIP EVENT OF MARCH 9,1988
A.
Narrative Description
On Wednesday, March 9, 1988, around 5:32 p.m. CST, the LaSalle Unit 2
reactor was operating at steady state conditions at approxir.iately
84% power with 76% rated core flow using both recirculation (RR)
pumps and with the control rods withdrawn to the 99% flow control
line.
Feedwater temperature was 402 F.
LaSalle Unit I was
operating at power in steady state conditions and was unaffected by
the subsequent events on Unit 2.
Instrument Maintenance Technicians
(IMs) were in the process of performing a surveillance test on Wide
Range level instrument 2B21-N0378B to check the Reactor Core
Isolation Cooling (RCIC) initiation function at -50 inches reactor
level.
The IMs were stationed at the instrument rack and in the
control room and had received permission from the appropriate
operations personnel to perform the surveillance.
The IM at the
instrument rack had correctly isolated and equalized the instrument
(2821-NO378B) in accordance with the functional test procedure,
LIS-NB-404.
The next action was to open the test / vent valves,
however, instead the IM technician opened the isolation valves to the
variable and reference legs to the instrument.
Since the equalizing
valve was still open, a pressure equalization occurred between the
variable and reference legs for this instrument and all the other
instruments which share the same reference leg. At the time of the
valving error feedwater level control was selected to channel B
which takes input from an instrument which utilizes the same
reference leg as 2821-NO37B8.
The equalization produced by the
valving error resulted in a high "indicated" level to feedwater
level control, causing the operating feedwater pumps (A turbine
driven reactor feedwater pump - TDRFP, and the motor driven reactor
l
feedwater pump - MDRFP) to begin reducing flow.
The IMs realized a
valving error had been made and attempted to correct the error by
'
I
shutting the reference and variable leg isolation valves.
This
action caused a pressure pulse on the reference leg of all the
instruments which share the same reference leg at that instrument
rack.
Increasing pressure on the reference leg caused the level
instruments to indicate low reactor vessel level. The key instru-
ments which were affected by this pulse were the ATWS RR pump trip
switches 2B21-N036C and 2B21-N0360, which are designed to trip the A
and B RR pumps to off.
Both RR pumps did, in fact, trip off.
Instrument 2B21-N024B which provides a reactor protection system
l
l
'
3
.
l
-
.-
. _ _
.
-
. .. . . - -
-.
. - -
-
.
.
(RPS) channel B1 low level 1/2 scram was also affected and resulted
in.a 1/2 scram signal and the associated alarm.
'
The trip of the RR pumps resulted in a large and rapid power
reduction (approximately 45%) as a result of the large flow
reduction (to natural circulation conditions).
The control rods
remained in their pre-RR pump trip position on the 99% control line
(see Attachment No. 3-BWR Power to Flow Map). This region of the
BWR Power to Flow Map was known to be susceptible to instabilities
in some BWRs. As a result of the large drop in power, a large drop
in steam flow occurred causing large changes in extraction steam
flow and extraction steam pressure.
Extraction steam supplies the
heating to the feedwater heaters. The changes in extraction steam
caused severe perturbations in the feedwater heater level control
system due to water flashing to steam from lower shell pressures,
reductions in shell side input from reduced steam flow, and changes
in condensing rate. The feedwater heater level control was unable
to react fast enough to control the large load reduction and tripped
the extraction steam input to the heaters in order to prevent
induction of water into the main turbine.
The st. curing of steam
heating to the feedwater heaters resulted in cooler feedwater being
supplied to the reactor (approximately 45'F decrease in 4 minutes)
which is the equivalent of a positive reactivity addition. This
resulted in an increased power to flow ratio which further reduced
the margin to instability.
At this point in the event, the operators in the control room were
primarily concerned with attempting to restore feedwater heaters.
The operators had correctly determined that an ATWS event had not
occurred but that an instrument problem had resulted in the loss of
both RR pumps.
The loss of feedwater heating was r.ot unexpected for
the large power drop caused by the RR pump trip.
The operators also
realized that the reactor was operating in a region of the power to
flow map where instability was possible. Between 4 and 5 minutes into
the event, the Average Power Range Monitor (APRM) indications were
observed by the operators to be oscillating between 25% power and
50% power every 2 to 3 seconds and the Local Power Range Monitor
(LPRM) down scale alarms began to annunicate and clear.
(Later
examination of the STARTREC, Startup Transient Recorder, (a high
speed, multi-channel recording system installed for startup testing
which starts recording when selected parameters exceed predetermined
limits) showed the oscillations to be much larger than the operators
were able to see).
The APRM indications confirmed the onset of
'
instabilities and the operators attempted to restart a RR pump in
order to increase flow to leave the instability region. Attempt 3 to
start a RR pump were unsuccessful and the shift commenced preparations
l
to manually scram the reactor. About 7 minutes into the event and
l
before the shift was able to perform a manual scram, the reactor
automatically scrammed on high neutron flux as seen by the APRMs.
The scram shutdown the reactor as designed and recovery from the
!
scram proceeded normally.
Some m r r equipment problems occurred
i
during recovery and subsequent - shutdown, however, these were
judged by the AIT to have no ef.a t on the event and will not be
r
l
discussed further in this report.
Tne licensee informed the NRC
at 6:32 p.m. CST of the RR pump loss, feedwater heating loss and
'
resultant scram.
4
l
l
'
.
.-
.
-,
.
_
.
l
,
,
.
B.
Sequence of Events
Times and sequences of events in the previous narrative description
.
were derived by the AIT from a combination of interviews and hard
data.
The following sequence of events represents a compilation
of information by the AIT taken from the alarm printer, the Startrec
recording system and interviews with licensee personnel. Times are
given in 24 hour2.777778e-4 days <br />0.00667 hours <br />3.968254e-5 weeks <br />9.132e-6 months <br /> clock time (17:32 equals 5:32 p.m.) and are all
'
Central Standard Time.
SEQUENCE OF EVENTS FOR MARCH 9, 1988
LASALLE UNIT 2 INSTABILITY EVENT
Initial Conditions
84% Reactor Power (930 MWe)
o
Steady State Conditions
99% Flow Control Line
6
76% Rated Core Flow (82 x 10 lb/hr)
-
Feedwater Temperature = 402 F
LIS-NB-404 in process (Surveillance that tests RCIC initiation
at -50" reactor water level .)
Event Summary
March 9, 1988
!
Time
Event
17:32 (approximately)
Instrument Maintenance Technician valved
in the variable and reference legs of a
differential pressure switch with the
equalizing valve open; initiating a
pressure equalization between the two legs
and a high "indicated" reactor level.
17:32i33
High Reactor Water Level Alarm initiated.
STARTREC (Startup Transient Recorder)
initiated on increase in narrow range
level and ran for programmed 1 minute
duration.
Instrument Maintenance Technician corrected
valving error by isolating reference leg
from variable leg which resulted in a low
"indicated" level spike cauting other
instrumentaticn to actuate.
5
.
l
,
.
,
,
17:32:49
2A/2B ATWS alarm initiated a trip of both
Reactor Recirculation (RR) pumps and power
',,
and flow began coasting down to natural
circulation levels.
Division 2 Reactor
low Level Alarm initiated.
2A ATWS cleared.
17:32:50
Half scram-on +12.5" reactor water level
initiated,
s
2B AiVS cleared.
<
Division 2 Reactor lo Level Alarm cleared.
Reactor Water Level 3 Alarm (+12.5") initiated.
17:32:51
Reactor Water Level 3 Alarm cleared.
Reactor Water Level Hi Channel B trip.
17:32:52
Reactor Water Level Hi Channel B was
manually reset.
Half Primary Containment Isolation System
(PCIS) level trip was manually reset.
Nuclear Station Operator (NS0) saw that
B narrow range reactor water level
indicator was approximstely 30" and rising
while A and C were steady at approximately
40".
17:33:10
First feedwater heater high level alarm
annunciated.
17:33:20
First feedwater heater isolates.
heaters continue to isolate for duration
of event.
Unit 2 NSO reviewed feedwater heater
situation and planned to reopen extraction
steam valves after valves fully closed to
regain feedwater heating.
Shift foreman discharged to local heater
controllers to aid in reestablishing feed-
water heating.
17:36(approximately)
Shift engineer entered control room.
Operators observed APRMS oscillating
between 25% and 50% power with an
approximate 2-3 second period.
6
. - .
.
_ _ _ -
h-
..
F
Operators attempted to restart the RR
pumps per abnormal operating procedure,
LOA-RR-07.
~
17:36:55
B RR flow control valve locked up.
Equipment Operator discharged to reset the
lockouts on flow control valves.
17:37:21
First LPRM downscale alarm annunciated. The
LPRM downscale alarms continued to flash
and clear, on a 2 second period.
17:37:51
First LPRM Hi Alarm annunciated.
17:38:20
The A flow control valve was reset.
STARTREC initiated on increase in narrow
range level and ran for 1 minute.
NSO ramped the A flow control valve to
minimum position.
17:39(approximately)
NS0 attempted restart of 2A RR pump twice,
but was unsuccessful.
Shift Engineer directed a manual scram to
be initiated.
'
17:39:19
STARTREC terminated its second 1 minute run.
17:39:23
Channel A neutron APRM trip.
17:39:25
Channel B neutron APRM trip.
Reactor scrammed on 118% neutron flux.
{
NOTE:
Designates that ne Hathaway Recorder time was available for the
event.
However, the event is listed in the approximate sequence
in which it occurred.
III. INVESTIGATIVE EFFORTS
A.
- ynopsis of AIT Activities
The AIT members had all arrived onsite on March 17,1988.
~1 were
thoroughly briefed on the event by the licensee and Gen'-r) Ilectric
personnel in a meeting at the site at 1:00 p.m. hours c: Aoch 17.
The team was provided with pertinent instrument records d the
event (including Sequence of Events data), and with documentation
comprising the safety evaluation by the licensee. The latter
7
.
-
included correspondence between the licenses and the reactor vendor
(GE),
In response to a team request, the 1.censee also provided a
written description of the operator response and assessment'of the
event as it occurred.
The inspection team had various meetings with the LaSalle plant
management and staff and with GE supporting staff during the
assessment of the event.
Subsequent to telephone conferences with
Headquarters and Region III offices, the plant was allowed to
restart at approximately 10:45 p.m. on March 17.
Operation was to
be under operating procedures which had been modified by a standing
order to require manual scram after trip of the recirculation pumps
in Operating Modes 1 or 2.
The AIT investigation continued with interviews of control room
operating personnel and a walkdown of the control room response to
the event.
The AIT documented several issues and concerns which
were presented to the licensee at a preliminary exit meeting on
March 18. The AIT concluded onsite activities with an exit meeting
on March 24, 1988.
B.
Core Nuclear and Thermal Hydraulic Performance
1.
Core Performance
In general, the AIT confirmed the adequacy of the assessment
of core performance performed by the licensee and the reactor
vendor.
Several concerns and questions, however, were
developed by the AIT and these are discussed in the Concerns
and Recommendations portion of this report (paragraph IV).
The following paragraphs provide a discussion of core
performance during the event.
Following the trip of the RR pumps, Core Thermal Power (CTP)
decreased and stabilized within about 30 seconds at about 40%.
The APRMs showed stable indications (the APRMS read neutron
flux as distinguished from CTP, however, both were stable at
this point). As feedwater temperature decreased, CTP increased
slightly to 43%. At approximately 4.8 minutes after the RR
trip, the APRMs began oscillating and the LPRM down scale
alarms were received. At 5.8 minutes after the RR pump trip,
the STARTREC system initiated its second recording for the
designed 1 minute period and stopped about 8 seconds before
the full scram.
STARTREC information is not available to the operators in the
control room at the time of recording, so from the APRM
recorders in the control room, the operators believed the
oscillations were approximately 25% power (neutron flux) in
magnitude (between 25% and E0% power) every 2 te 3 seconds.
Analysis of the STARTREC traces showed APRM peak to peak
l
i
8
_
- _ _ _
- ______ _
.
.
oscillations ranging from 20% to about 95% power.
Extrapola-
tion of the traces to the time of the scram leads the AIT
to believe the oscillations were at least 100% peak to peak
,
when the scram occurred.
The oscillation frequency was
'
approximately 0.45 hz.
The approximately 2 second period of
the oscillation is driven by core flow hydraulic conditions,
primarily the time it takes a void perturbation to travel the
length of the fuel . The APRMs measure neutron flux and during
reactivity changes the neutron flux leads the fuel cladding
heat flux by the thermal time constant of the fuel
pellet / pellet-clad-gap arrangement.
This time constant is
pproximately 6-7 seconds and acts to filter neutron flux
spikes. Consequently, the heat flux oscillations for this
event are estimated to be less than 10% of the neutron flux
oscillations cycling around an average CTP of about 45% during
the final minute of STARTREC recording.
Based on LPRM alarm
signals at 87% APRM power and LPRM readings after pump trip,
the AIT estimates that the peaking factor increased from 2.11
price to the event to a value of 2.65 at the time of the APRM
alann. This corresponds to a peak neutron flux level of 312i
(relative to rated core average) at the 118% APRM scram level.
However, because of the smaller changes in heat flux, the 13.4
Kw/ft fuel design limit was not exceeded and no core damage was
indicated by chemistry results.
One of the more important considerations in this type of event
is whether the LPRM swings are in phase with the APRMs or not.
The reason for this concern'is that the core protection actions
are actuated by the APRM signals, so, if some LPRMs were
oscillating out of phase with the core average, the effect
would be to lower the APRM signals that these LPRMs feed.
Consequently, the effectiveness of the APRMs as a protective
system would be less and local areas of the core would be
undergoing oscillations of much greater magnitude than
indicated by the APRMs. Analysis of the alarm printout of the
LPRM downscale and LPRM hi alarms, the "clean" sinusoidal wave
shape, and the in phase APRM traces from STARTREC by the
licensee and the reactor vendor (GE) determined that the LPRMs
and APRMs were in phase with each other.
The AIT verified this
analysis and concurred with the licensee's interpretation.
This type of oscillation is less severe with lower power peaks
to trip than would occur with regional oscillations which have
becn observed in foreign reactors. Generic analyses performed
during the resolution of Generic Issue B-19 bound the LaSalle
Unit 2 instability and demonstrated that the fuel thermal or
mechanical limits were not exceeded during the event.
General Electric's evaluation of the LaSalle event concluded
that the frequency and magnitude of oscillations which Unit 2
experienced were consistent with the characteristics cbserved
during stability testing and operation at other BWRs. GE
further concluded that the event was bounded by the generic
'
9
_ _ _ _ _ _ _
.
.
'
,
analysis and that no fuel thermal or mechanical limits were
exceeded during the event. While the AIT did not agre.e or
disagree with the GE conclusions at the time of exit, there
were several questions and concerns relating to unexpected
aspects of the event which required further study. These
included the following:
(1)
failure to predict the susceptibility of the Unit 2 core
to thermal hydraulic oscillations based on the calculated
decay ratio.
(2) magnitude of the oscillations compared to previous events
with similar behavior,
(3) potential for out of phase regional oscillations of much
larger magnitude and the lack of a defined boundary based
on inherent shutdown mechanisms,
(4)' questions about the adequacy of instrumentation for
detection, suppression, and evaluation of limit cycle
neutron flux oscillations, and
(5) questions about the adequacy of technical specifications
and procedures for detection and suppression of neutron
flux oscillations.
These issues are discussed in more detail in paragraph IV,
"Concerns and Recommendations".
2.
Chemistry Results
Following the scram on March 9,1988, the LaSalle Station
Chemistry Department took post shutdown samples of the reactor
coolant water in order to determine if there were any
indications of fuel damage. Analysis of the iodine result
fromthissampleindicatednoabnormalities(between2x10~3
5
and 1x1f microcuries per gram for iodine 131 through 135
ThereactorwaterdoseequivalentI-131waslessthan2x10}4
microcuries per gram as compared to a Technical Specification
limit of 0.2 microcuries per gram.
This data, as well as the
past two months sample data for both Units 1 and 2, was made
available to the AIT for verification by the licensee. At a
result of the CAL, the licensee developed an increased
frequency sampling program of reactor water and off gas which
was implemented following restart of Unit 2.
The results of
this sampling are documented in Attachment 5 and show no
indications of fuel damage or abnormalities from readings prior
to the event.
10
.
..
3.
Equipment Performance
The following paragraphs summarize the AIT's conclusicas
regarding the performance of specific systems or pieces of
equipment during the event. By the CAL, the licensee was also
requested to address equipment performance and that assessment
is included in Attachment 5.
(a) Recirculation Pumps and Flow Control Valves
A trip of both RR pumps is a designed feature of the
LaSalle plant in order to cause a power reduction in the
event of an ATWS, as indicated by a loss of reactor level
without an associated scram.
The pressure pulse on the
reference leg of the ATWS switches appeared to be
sufficient to provide this indication and consequently
the trip of both pumps occurred as would be expected.
During the RR pump trip recovery, the operators attempted
to restart the RR pumps and were unsuccessful. At the
time of the exit, the RR pump start failure was believed
to be due to the failure to satisfy one of the pump start
interlocks.
Exactly which interlock was not satisfied was
not conclusively determined, however, likely candidates
include the recirculation flow control valve not fully
in minimum position and the hi speed start permissive
(depends on feedwater flow). The operators have no
indication available in the control room to determine
which DD
n"=
- + - ' ' " " '
follow
Was CC
belies
equipn
y
[
,#
Y l(f f$ g()
Q
fti
p
believ
hf6
Q
were a
C
noted
RR pum
recogq
by try
prepar.
prior
RR pumi
into tl
,
NSO wat
i
,
positic...
..m
o
iv.
vas not reset Decause there was an
l
l
abnormal signal alarm which would have required additional
operator actions.
FCV lockout was not fully investigated
4
by the AIT.
i
11
_
_ _ _ _
,
.
f
l
b.
Shortly after the RR pump trip, the feedwater heater level
control system began isolating extraction steam from the
heaters in order to protect the turbine from water
induction. The resultant cooler feedwater which was
supplied to the reactor caused the reactor to become even
more susceptible to instability. The AIT reviewed the
feedwater and feedwater heater performance and concluded
that the feedwater heater actions were as expected and had
been seen before for similar large rapid drops in load.
A more complete description is provided in Attachment 5.
This observation, however, negated an initial licensee
statement that two abnormal conditions, operation at high
power in natural circulation and reduced feedwater heating,
were necessary to put the reactor into the unstable
region, in that, feedwater heating would be expected to
be lost every time both RR pumps were lost from high
power. The AIT believed the licensee efforts in investi-
gating the performance of the feedwater system were
appropriate,
c.
Power to Flow Scram (APRM flow-biased trip)
Since the reactor ends up at lower power following a RR
pump trip, and the reactor eventually scrammed on high
power in terms of APRM flux, concern was expressed over
the operation of the power to flow or APRM flow-biased
trip.
The APRM flow-biased trip is designed to protect against
spurious scrams due to transient events that cause spikes
in neutron flux. APRM neutron flux correlates to thermal
power level during steady state conditions. However,
during power increase events the APRMs will over-predict
themal power because the neutron flux leads the reactor
heat flux due to the thermal time constant of the fuel and
cladding.
It should be noted that no credit is taken for
the flow-biased scram in the transient analysis; only the
120% high flux scram is credited.
The APRM flow-biased trip circuitry receives the APRM
neutron flux signal and filters it through an R-C circuit
with a 6 second time constant. This R-C circuit which
is known as the "themal power monitor" essentially
integrates the APRM neutron flux signal over the past 6
seconds and develops an APRM reading which simulates
thermal power by :ccounting for the lag in response of the
fuel cladding heat flux to neutron flux variations.
The
output of the thermal power monitor is then compared to
the flow-biased scram setpoint which is calculated by the
equation:
S=0.66 WD + 51%. W is the driving flow which
D
I
t
12
_ _ _ _ _ _ _ - _ _ _ _ _ _ _ _ _ _ _ _ _ -
_
.,
.
.
that at 6.9 minutes following the RR pump trip, RPS APRM
channels A and B actuated to produce the full scram.
APRMs A, C, and E all indicated neutron monitoring trip
and are grouped as channel A in the logic scheme'. 'No
trip indications appeared for t.he D and F APRMs which is
believed to be due to the fact that the B APRM signal
completed the logic and actuated the scram.
The
licensee's review included a functional check of the APRM
trip setpoints following the event, a comparison utilizing
gain adjustment factors, and an evaluation of whether the
flux oscillations could have been missed by the scram
logic. The AIT concurred in the licensee's conclusion
that the Reactor Protection System performed as designed.
The licensee's investigation of RPS performance was
appropriate.
e.
Anticioated Transient Without Scram (ATWS)
The oni; ATWS mitigation equipment actuated by the event
was the dual RR pump trip actuated by the pressure pu'se
,
to the ATWS RR pump trip switches' reference leg.
T'.ie s e
functioned as designed to trip off the RR pumps on
indicated low reactor level. The operators spent the
first 20-30 seconds after the RR pump trip and associated
alarm confirming that no condition existed which required
a scram without a scram occurring, hence no ATWS. The
available recordings also confirmed no other ATWS
actuations, which is consistent with the personnel
interviews and the initial valving error, in~that, no
other ATWS actuations would be expected.
l
C.
Operator Performance
On March 17, 1988, as part of the AIT charter, members of the
l
inspection team interviewed the on shift operators and supervisors
l
who had been assigned to Unit 2 during the March 9, 1988, reactor
power oscillation event.
The licensee's view of operator actions
is included in Attachment 5.
,
l
l
Personnel in the control room at the time of the event included the
IM personnel performing the surveillance on a headset with the
i
technician at the instrument rack, an NSO at each unit, the Center
Desk NSO (who operates Unit 1 and Unit 2 common systems) and the
l
Shift Centrol Room Engineer (SCRE).
The first indication of a
i
problem was a hi level alarm (due to the initial equalization
valving error).
Th:s brought the Center Desk Operator (CD0) over
to the feedwater control station to assist the Unit 2 NSO. As more
alarms came in (low reactor level and 1/2 scram due to the second
valving error and resultant pressure pulse) the Unit 2 NSO stationed
himself at the RR panel. Based on available indication and the fact
that an instrument surveillance was in progress, the operators
suspected an instrument problem. The Unit 2 NSO determined that he
14
,
4
.
is zero when both reactor recirculation pumps are tripped.
Therefore, the scram setpoint (S) during the instability
event was 51%.
.
Although peak APRM readings reached 118%, the peak output
of the thermal power monitor during the instability event
was approximately 42% to 45% because of the effect of the
R-C circuit on the APRM signal. Therefore, the flow-
biased scram never reached its setpoint of 51% and the
reactor tripped on high neutron flux at 118%.
d.
Reactor Protection System (RPS)
From a review of the alarm printer data, the STARTREC
information and interviews with operating personnel, the
AIT verified that the only RPS actuation setpoints
exceeded were those associated with the initial valving
errors and related pressure pulsing of the reference leg
which started the event and the APRM hi flux scram which
ended the event.
The RPS instrumentation which shares the refarence leg
with 2B21-N037BB consists of Level 3 (+12.5 inches
setpoint), Level 2 (-50 inches setpoint), or Level 1
(-129 inches setpoint) differential pressure indicators
and straight pressure indicators.
The pressure pulse
caused by the valving error is believed to have caused the
instruments to see about -40 inches reactor level (or a
pulse equal to about 80 inches of water column).
This
pulse would equal about 2.9 psig which would be too small
to affect any of the pressure instruments significantly.
This corresponds to the alarm printout which indicates no
pressure instrument actuations.
The 40 inches level would
also be expected to be too small to have affected the
Level 1 (-129) or Level 2 (-50) instruments, but would
have been expected to actuate the Level 3 instruments.
No Level 1 actuations were indicated by the alarm printer.
However, the ATWS RR pump trip switches 2B21-N036C and
2B21-N036D which did trip the RR pumps are Level 2
switches.
This indicates the pulse was either larger than
able to be seen by the level transmitter feeding STARTREC
or the switch trip setpoints were conservatively high.
Consequently, other Level 2 actuations may or may not
have occurred depending on setpoint and the strength of
the pulse as seen by the instruments. All of the Level 3
switches actuated as would be expected, including the RPS
channel B1 low level alarm and 1/2 scram and the Automatic
Depressurization System confirmatory alarm. Both of these
are confirmed by the alarm printout.
With regard to the APRM hi flux scram, the licensee's On
Site Review (OSR) had concluded (from the alarm printout)
13
_
_
_ _ _ _ _ _ _ _ _ _ _
4
..
had lost RR flow and that ATWS alarms were indicated.
At that
point, efforts were concentrated on ensuring that a valid ATWS event
was not occurring.
,
The Station Control Room Engineer (SCRE) responded to the initial
alarms and, remaining cognizant of his position as the SCRE,
positioned himself such that he could clearly observe operator
actions and reactor parameters.
He then contacted the Shift Engineer
(SE).
l
By this point, the NS0s had determined a valid ATWS did not exist
I
and had directed the IM personnel to stop their surveillance.
Multiple feedwater heater alarms were occurring and many of the
heaters were isolating extraction steam due to the rapid power
reduction. A shift foreman had been sent to the local heater
control panels to help restore feedwater heating. At about the time
of the Shift Engineer's arrival in the control room, preparations
were being made to attempt to restart the RR pumps and power
oscillations of 25'. magnitude ranging from 25'4 to 50*. began to be
seen on the APRM recorders.
The SE, upon arriving in the control
room, quickly assessed the status of the unit, reviewed the
operators' actions, and assumed responsibility for efforts to
recover from the reactor power oscillations. At this point,
operators were uncertain of the specific actions they should be
taking once the power oscillations were observed. The procedures
and training appear to have been inadequate. The operators all
recognized that they were in the region of core instability, but
were uncertain of what actions should be taken.
The procedures
basically instruct the operator to leave the region of instability
without directing "how".
Thoughts included; (1) driving in control
rods in sequence, (2) using the "CRAM" array of control rods, (3)
starting a RR pump, and (4) scram the plant. Operators felt that
driving in control rods in sequence would have been too slow in that
the first rods to go in would be rods of little reactivity worth.
Use of the "CRAM" array of control rods was addressed only in the
licensee's procedures pertaining to a loss of feedwater heating
consequently operators were uncertain if they should use the "CRAM"
array in this condition of 2 RR pump trip and loss of feedwater
heating.
The operators did not want to scram the plant if they
did not have to.
Therefore, they decided to try and leave the
instability region by restarting a RR pump (either a flow increase
or a lowering of the rod line would get the reactor out of the
unstable region) which they may have accomplished if the RR pump
start sequence could have been satisfied.
It is also likely that
if a RR pump had been successfully started, the reactor may have
scrammed anyway on the resultant power increase.
As the operators
were unsuccessful in starting a RR pump and were preparing for a
manual scram, the reactor scrammed automatically.
The AIT evaluated the response of the individuals versus the
information they had available, and has the following observations.
15
,
.
..
_ _ - _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ - _ _ _ - _ _ _ _ _ _ _ _ _ _ - _ _ _
_____
N
.
'
With the exception of the personnel error by the instrument
mechanic (IM) that started the transient, the AIT has concluded
that the personnel on-shift at the time of the event took
prompt corrective action in accordance with the procedures
and training available to them.
The AIT believes that the
operators could have taken better corrective action by driving
in control rods, but that this direction was not provided by
procedures or training.
The procedures available to the operators at the time of the
event appear to be inadequate in that they don't specifically
address the necessary action to be taken in order to exit the
region of instability. General Electric Company's Service
Information Letter (SIL) No. 380, issued February 1984,
addressed the issue of core stability and the actions to be
taken if there were a RR pump (s) trip and the core entered the
region of instability. The GE SIL had not been incorporated
into the licensee's procedures or training.
This issue is
further discussed in paragraph III.D.
D.
Procedure Adequacy and Training
1.
Technical Specifications (see Attachment 4)
The LaSalle Unit 2 Technical Specification (TS) 3.4.1.1.b
contained the requirements for loss of both RR pumps or "no
reactor coolant system recirculation loops in operation".
The TS states "immediately initiate measures to place the unit
in at least HOT SHUTDOWN within the next 6 hours6.944444e-5 days <br />0.00167 hours <br />9.920635e-6 weeks <br />2.283e-6 months <br />." The
defined actions, however, were not prescriptive for the seven
minutes in which the operators had to act for this event.
The
TS for loss of a single RR pump 3.4.1.1.a. in step 3.4.1.1.a.2a) 2)
discusses increasing core flow.
While the reactor was not in
this condition (both RR loops were not operating versus one),
the operators were aware of this TS, the associated surveillance
requirements and the fact that increasing flow would get the
reactor out of the unstable region of the power to flow map.
These considerations in their own minds served to back up the
operators decision to restart RR pumps.
The NRR representa-
tives on the AIT were familiar with the development of the
LaSalle TS and indicated that part of the reason that LaSalle
did not have more specific TS was that the Unit 2 decay
factor was submitted as 0.60.
With this much margin to a Limit
Cycle Oscillation decay factor of 1.0, it was believed that
the probability of oscillations at LaSalle would be very low.
Since oscillations have now been observed and the decay factor
calculation is in question (see Concerns paragraph IV), the
AIT believes that the TS were inadequate.
2.
GE SIL 380 Revision 1 (Attachment 7)
In 1984, GE issued SIL 380 Revision 1 containing recommendations
regarding BWR Core Thermal Hydraulic Stability. This document
16
--
____ ____ _ _ __________
.,
.
.
discusses RR pump trip and recommends, "Immediately reduce power
by inserting control rods to or below the 80*4 rod line using
the plant's prescribed control rod shutdown insertion. sequence."
Further, the SIL recommends that the operation of restarting
recirculation pumps should be performed from below the 80* rod
4
line.
These recommendations were not incorporated into
LaSalle's TS or abnormal operating. procedures.
Like with
TS, part of the motivation to not incorporate these recommend-
ations into procedures was based on the margin to instability
provided by the calculated decay factor as well as a skeptical
attitude regarding the susceptibility of the LaSalle reactors
to instability.
3.
Abnormal Procedures and Training
The LaSalle abnormal operating procedures (LOAs) and the
associated surveillance procedures (LOS) and general procedures
(LGP) provided only general guidance regarding what to do in
the event of loss of single or two loop recirculation flow.
While guidance and training had been provided to all the
licensed operators on how to recognize the onset of
instability, there was little guidance provided on what to do
next, other than performance of the TS required surveillance
and to leave the region of likely instability.
The loss of
recirc flow procedures were directed at restarting RR pumps.
The AIT concluded that the operators reacted as their training
and procedures led them, in that, they recognized very well the
onset of instability, however, interviews indicated they were
confused and uncertain as to how to exit the region.
The AIT
believes that both the abnormal procedures and the training
were inadequate in this regard.
4.
Simulator Training
The licensee attempted to simulate LaSalle's instability event
on the simulator, but was unsuccessful.
The event could
possibly be simulated with the instructor interacting with the
computer, however to produce an accurate, detailed simulation
a new program must be written.
Currently, the Production
Training Software Group has the data from the actual event and
is reviewing possible changes.
The licensee stated that they
expect a permanent program will probably be available by
July 1988 and they will keep the resident informed on the
status of the changes.
E.
Reporting
As the nature and magnitude of the core power oscillations on
,
i
LaSalle Unit 2 became widely disseminated and understood such that
the AIT was formed on March 16, 1988, several questions were raised
'
regarding the adequacy of the licensee's reporting of this event.
These questions were in large part motivated by the fact that
17
.
.
-
.
..
.
. .
-
.
.
._-
-
.
_
.
.
the majority of the NRC staff did not learn of the core power
oscillations aspect of the event until issuance of the Region III
morning report update on March 15, 1988, even though the event had
occurred on the evening of March 9, 1988.
The AIT examined the
reporting aspects of this event through interviews, individual
recollections and available tape of phone conversations.
The
following summarizes the team's observations.
1.
Reporting Sequence
(a)
Following the trip of the RR pumps and the subsequent
scram at about 5:39 p.m. on March 9, 1988, the licensee's
initial investigation classified the event as reportable
to the NRC within 4 hours4.62963e-5 days <br />0.00111 hours <br />6.613757e-6 weeks <br />1.522e-6 months <br /> under the provisions of 10 CFR 50.72 due to actuation of the Reactor Protection System
(RPS).
The licensee actually notified the NRC within 1
hcur at about 6:32 p.m. on March 9,1988.
The initial
phone call communicated information on the RR pump loss
due to personnel error, loss of feedwater heating and
reactor trip on high APRM flux.
No mention was made of
the core power oscillations, however, the licensee's caller
stated, "we are still investigating all our alarm typers
and everything else." Com.ments were made to the AIT that
the licensee's caller sounded almost jovial over the
phone.
After listening to the tape of the 50.72 call,
this observation was explained to the AIT's satisfaction
in that the NRC Headquarter's Duty Officer (HDO) and the
licensee's caller recognized each other's voices from
previous association and exchanged pleasantries.
This
'
situation did not appear to detract from proper communica-
tion of information.
(b) The HD0 contacted the Regional Duty Officer (RDO) at 7:21
p.m. CST, who, in turn, notified the Cognizant Section
Chief (also AIT leader) and Branch Chief.
(c) At approximately 7:00 p.m. CST on March 9, 1988, the
licensee and the Resident Inspector (RI) succeeded in
establishing contact and the RI was briefed regarding the
event.
The RI was offsite attending the Region III
Resident Seminar (March 8-10).
The licensee informed the
RI that the APRMs had oscillated between 25-50% during
,
the event and that the reactor had scrammed on APRM high
flux at 118% setpoint.
At this point, the RI did not
appreciate a significant concern with the APRM oscillation
in that oscillations are a potentially expected phenomena
given the parameters of high reactor core power and low
reactor core flow.
1
!
!
'
18
-
- -_-.
,.
..
..
.
.
.
.
.
.
.
.
.. .. . ..
.
._
.'
.
.
(d) On March 10, 1988, at approximately 10:00 a.m., Region III
issued a morning report describing the event.
No mention
of oscillations was made in this report.
The high power
scram was attributed to a power surge caused by the cold
(e) At approximately 12:30 p.m. on March 10, 1988, the Region
III Section Chief contacted the LaSalle Station Manager
for the purpose of ensuring the information in a Preliminary
Notification (PN) was correct.
During this conversation,
the Station Manager pointed out that while the PN was
'-
substantially correct, the unit had experienced some
neutron flux oscillations during the event which were
being evaluated. The Region III Section Chief did not
include any information on the oscillations in the PN
based on the evaluation still being in progress.
(f) On Ma.ch 11, 1988, the RI returned to the site and
participated in several discussions with the licensee
involving the event including a meeting where the STARTREC
traces showing the oscillations (20%-95%) were made
available. The RI was informed that the licensee had
involved GE in the evaluation and that GE had indicated
the oscillations were within analysis bounds and provided
for by the APRM high flux scram. Also, on March 11, 1988,
the Station Manager contacted the Region III Section Chief
and provided additional information including an
explanation of the reactor's position on the power to
flow map, the effect of the loss of feedwater heating, and
the onset of flux oscillations.
During this call, the
Section Chief does not recall that the magnitude of the
oscillations was discussed, but, rather the Station Manager
indicated the oscillations were being evaluated and the
reactor vendor was involved.
The Station Manager indicated
that GE felt the oscillations were within analysis bounds,
that the core performed as expected for the conditions and
he offered to make a presentation on the event to Region
III if desired.
The Section Chief indicated this may be
a good idea but wanted to wait until the RI and possibly
Region III core physics inspectors had a chance to look
at the data.
Following his review of the available
information, the RI contacted the Section Chief on
March 11, 1988, and comunicated the oscillation infoma-
tion which indicated the event was more severe than
initially believed. At this point, the Region III Branch
Chief and Division Director were informed that flux
oscillations had occurred but were within analysis and a
more complete description would be available on Monday,
,
March 14 following the RI's further investigation.
The
NRR Projects Office was similarly informed.
19
.
.
__-____ _ _ _ ____
-
.
.
(g) On March 14, 1988, the RI was given a copy of CECO's
Potentially Significant Event (PSE) report and an advance
copy of a Rapid SIL generated by GE addressing the event.
Due to following up on additional information on the event
and the length and complexity of the morning report, the
RI was unable to get the report documented and transmitted
by the 10:00 a.m. deadline on March 14, 1988.
Since
Region III was waiting for further information from the
site, no additional actions were taken by the Region on
March 14, 1988.
'
(h) On March 15, 1988, the Region III morning report update
describing the event, including the oscillations, was
disseminated, resulting in the formation of the AIT on
March 16, 1988.
2.
Reporting Evaluation
In the AIT's view, the licensee correctly classified the event
on the initial call on March 9, 1988, as being 4 hour4.62963e-5 days <br />0.00111 hours <br />6.613757e-6 weeks <br />1.522e-6 months <br />
reportable based on the requirements of 10 CFR 50.72 for
an RPS actuation. While the initial call did not mention
oscillations, the caller stated that the information was still
being evaluated.
Consequently, the AIT believes the initial
call was satisfactory for as far as it went. However,10 CFR 50.72(c) requires the licensee to make followup notifications
in addition to initial notifications as follows:
(2) Immediately report (i) the results of ensuing
evaluations or assessments of plant conditions,
(ii) the effectiveness of response or protective
measures taken, and (iii) information related
to plant behavior that is not understood.
The licensee made no formal followup reports via the ENS
circuit (red phone) to the HDO.
The AIT believes that such a
followup call would have been appropriate since the occurrence
of oscillations and the nature and magnitude of the
oscillations would significantly add to the information
communicated on the initial call.
The AIT believes the
licensee felt that subsequent calls to the RI and the Region
III Section Chief would accomplish the intent of communicating
followup information. The AIT believes this to be incorrect
for two reasons.
First, the intent of 10 CFR 50.72 is to
l
describe the formal notification process which is to be
conducted over the ENS circuit and notification of other NRC
personnel (including the RI) does not relieve the licensee of
l
the responsibility to notify the NRC via ENS, Second, the
information communicated to the NRC in the subsequent calls was
-
insufficient to allow the NRC to appreciate the nature and
l
magnitude of the oscillations.
The NRC has a need for prompt,
pertinent information on this type event which is at least
'
equivalent to the information which the licensee furnished
'
to GE.
20
i
!
l
l
_ _ _ - _ _ _ _
l
.
l
IV. AIT CONCERNS AND RECOMMENDATIONS
A.
Concerns
-
The following paragraphs document concerns developed by the AIT
t
during the review of this event.
Several of these concerns were
discussed with the licensee prior to the exit and the licensee has
provided several responses in Attachment 5 to this report.
These
responses had not been completely evaluated at the time of the
writing of this report.
Other issues may require further study and
subsequent response.
1.
Decay Ratio
Decay ratio in a reactor is a measure of the response of the
neutron flux to a change or perturbation.
A decay ratio of
,
less than 1.0 indicates that the response to a perturbation
will decay to the steady state value (see Attachment 3, Figures
2, 3 and 4). A decay ratio of 1.0 represents the special
condition when the response to a perturbation will be a steady
state oscillation of constant magnitude (limit cycles).
GE predictive methods for determining BWR decay ratio were
approved with 20*4 uncertainty.
Predicted decay ratio for
LaSalle conditions was 0.60 (40*4 uncertainty) yet limit cycles
were observed.
The only obvious deviation from prediction
assumptions was in core water level and associated natural
circulation flow rate.
The 40% difference needs to be
explained. Have calculations been discredited as acceptable
evidence of core stability?
2.
Technical Specifications (TS)
BWR thermal hydraulic stability was the subject of Generic
Issue B-19.
Early BWRs were designed to maintain decay
ratios less than 0.5 (a decay ratio = 1.0 corresponds to an
undamped limit cycle oscillation).
Later core designs tended
,
towards decay ratios of 1.0 due to higher power density cores
'
and changing fuel design characteristics.
In the resolution of Generic Issue B-19, core designs which
were potentially unstable (DR = 1.0) under natural circulation
operating conditions were approved with the provision that
operating procedures and technical specifiestions would assure
that neutron flux oscillations indicative of core instability
would be readily detected and suppressed as required by GDC 12.
Licensees were informed by Generic Letter 86-02 that
such procedures and TS must be implemented for new reload
cores unless it could be demonstrated by approved calculation
methods that the core was stable throughout permissible
operating regions of the power / flow map.
Calculated core decay
ratios of less than 0.80 by General Electric methods were
approved as acceptable evidence of core stability,
j
21
-
i
.
.
LaSalle Unit 2 did not have fully implemented procedures and
TS in accordance with the B-19 resolution because the
calculated decay ratio for the current operating cycle is
0.60.
LaSalle Unit 2 did have TS and procedures for
stability surveillance under conditions of single loop
operation (SLO). but had declined suggestions by the NRC staff
that these should also apply to two loop operation because
higher decay ratios were certain for future reloads.
The
licensee has been informed that revised TS are required for
Unit 2 now that oscillations have actually been observed.
LaSalle Unit 1 is currently under review for reload. The
licensee has been informed that stability technical specifi-
cations will be required for that reactor even though the
calculated decay ratio is 0.75 (indicative that it is less
stable than Unit 2).
Since several BWRs have higher decay ratios than LaSalle, the
issue of decay ratio in paragraph IV.A and stability TS is
considered generic.
Improved standard technical specifications
relating to stability need to be developed.
Further, the
implementation status of stability TS on all BWRs should be
reviewed with particular attention to the adequacy of require-
ments for response to loss of RR pumps.
The criteria for
applicability of stability TS based on decay ratio calculations
and reactor type should be reexamined.
3.
Instrumentation
The following concerns with the available instrumentation at
LaSalle were developed by the AIT:
The slow pen response time and chart speed of the APRM
strip chart recorders in the control room reduced
indicated amplitude of oscillations.
Time delay relays for the LPRM Hi and APRM Hi alarms
delay recognition.
Running Average of APRM signal in Power / Flow circuit
delays or prevents reactor trip at lower power levels
with reactor oscillations.
If oscillations are regional in nature and LPRM signals
are out of phase, LPRM inputs to APRMs will tend to cancel
each other so that oscillations are not evident on the
j
!
Hi Speed Recording of APRM signals used for event evalua-
'
tion was triggered by Low Water Level signal
- not
always a condition of the instability event. This data
1
22
l
l
_ _ _ _ _ _ _
_ _ - - - - -
-
,
,
.. :
>
4
is needed to assess the nature and magnitude of neutron
flux' oscillations and the safety of restart after an
instability event.
-
LaSalle and some other BWRs do not have high speed data
recording instrumentation which can be' committed for
availability during plant operation.
4.
Oscillation Characteristics
Some characteristics of the LaSalle neutron flux oscillations
were atypical of previous events and have led to concerns about
the applicability of previous safety analyses.
The magnitude of
in-phase limit cycle oscillations previously observed on the
APRMs curing special stability tests and operating reactor
events were typically in the range of 5', to 15% (peak-to peak) of
rated power, and as high as 25%.
This compares to peak-to peak
values of about 100% at the time of the 118% neutron flux trip
for LaSalle.
The estimated value of local power at the time of trip was
greater than 310% and LPRM readings indicate that the core power
peak shifted and increa:ed by 25%.
Even though the fuel LHGR
limit of 13.4 kw/ft was not exceeded because of the thermal time
constant of the fuel, the increased power peaking was unexpected
based on Vermont Yankee stability tests, and was not factored
into the generic safety evaluation performed by GE during review
of the thermal hydraulic stability Generic Issue B-19.
The previous GE safety analyses considered several limiting
moderate frequency transients which were initiated while the
neutron flux was oscillating below the 120% scram setpoint, and
included a rod withdrawal error with the flux oscillating up to
the 120*. scram level . Additional analyses were performed to
evaluate the impact of oscillations that approached 300% of
rated neutron flux (e.g., regional oscillations) without scram
prior to rod insertion and termination of the event. All of
these analyses showed that significant fuel thermal margin
existed to safety limits. While there are several aspects of
these analyses which differ from LaSalle (initial power level
and amplitude of the oscillations; no change in bundle peaking
factors due to the event, etc.), the AIT agrees that they are
sufficiently representative and conservative to demonstrate that
no fuel thermal or mechanical limits were exceeded during the
event.
However, reliable detection and suppression provisions
are necessary to assure protection against future events which
could involve regional oscillations to higher power levels.
The licensee was also asked to review the impact of the event on
stability considerations addressed in the 1979 GE Generic ATWS
report, "Assessment of BWR Mitigation of ATWS" (NEDE-24222).
23
.
.
.
.
.
.
..
.
.
. _ - _ _ _ _ _ _ _ _ _ -
.
4
The report does specifically investigate the sensitivity and
potential impact of limit cycle neutron flux oscillations up to
500% of rated bundle power following recirculation pum'p trip.
It was concluded that oscillations of this magnitude would not
result in sufficient fuel clad temperature variation (130'F) to
affect fuel integrity.
It was further concluded that a loss of
clad integrity due to prolonged exposure to limit cycles was an
acceptable consequence in view of the importance of the
recirculation pump trip (RPT) to minimize the energy deposited
-in the suppression pool (thereby maintaining containment
pressure within limits) during an ATWS event.
In view of the large magnitude of the APRM oscillations in
LaSalle, the AIT believes that the ultimate power level without
scram is unknown, and that the 500% level assumed in the ATWS
investigation may not be bounding.
LPRM oscillation magnitudes
more than seven times those of the APRMs have betn observed in
the case of regional oscillations. The licensee reports that
the BWROG is discussing this issue (inherent power limits) and
the licensee will provide a status report on July 1, 1988.
5.
Additional Concerns
Several additional concerns were presented to the licensee in
the form of questions. These questions and the licensee's
response are contained in Attachment 5 to this report.
B.
Recommendations
The A" recommends that the concerns identified in items IV. A.1
+'
,i IV.A.5 of this report be examined by NRR for generic and
sie specific resolution.
In the interim, the AIT recommends
that revised stability TS as discussed in IV.A.2 be developed for
LaSalle Units 1 and 2 and the licensee be authorizied via letter to
modify interim operating procedures provided they remain consistent
with the new T.S.
The revised technical specifications and
procedures should incorporate the changes summarized in Attachment 5
(Appendix A, Item 3), which include immediate insertion of high worth
rods and observation of APRM/LPRM noise when no pumps are operating
and power is above the 80% Rod Control Line. The reactor is to be
tripped immediately whenever instability is suspected.
It is
expected that the time available (greater than 5 minutes) to
instability following a two pump trip transient is sufficient to
permit manual power reduction, avoiding the need for reactor trip
unless the core is unstable by a large margin. Proposed procedures
permit manual action for up to two minutes (prior to scram) to
reverse operating actions which may result in small margins of
instability when one or both pumps are operating.
V.
AIT CONCLUSIONS
The AIT finds that the core power oscillations observed on LaSalle Unit 2
on March 9, 1988, were initiated by a aersonnel error resulting in the
24
.
.
. _ _ _
_.
_
,_
.__
_.
__ __.
.
__
_
.
.
trip of both recirculation pumps compounded by the loss of steam heating
to the feedwater.
The licensee's evaluation of the event (including
assistance from the reactor vendor) is still continuing, however, -to date
is believed to have been thorough and proper.
No evidence of any fuel
damage has been noted related to the event.
Performance of equipment
during the event is believed to have been as designed for the conditions
encountered during the event. Operator performance, while it could have
been better, is believed to have been prompt, appropriately controlled and
within the bounds of the procedures and training available to them. Both
4
the procedures and training available to the operators are believed to
have been inadequate in that prudent recommendations of GE SIL 380 were
,
not incorporated and little direction was provided regarding what to do in
the event of instability.
Reporting is believed to have been inadequate
in that no formal followup report was made regarding the results of the
investigation and determination of flux oscillations.
The licensee's
response to NRC initiatives by shutting down, providing the AIT with
prompt and technically sound information and responding to the CAL is
believed to have been excellent.
Several concerns and recommendations on
'
this event remain to be followed as documented in paragraph IV.
On March 17, 1988, the AIT determined that the licensee had complied with
.
the restart provisioas of the CAL, includ ,g a requirement for a manual scram in the event of a loss of both recirculation pumps from an operating
condition, and the Region III Administrator's designee authorized restart
of LaSalle Unit 2 at 10:45 p.m.
The AIT and the licensee have agreed on
,
the essential features of revised technical specifications and operating
procedures to protect against this event during futur6 operation.
VI.
EXIT INTERVIEW
The inspectors met with licensee representatives (denoted in paragraph
I.D.) informally throughout the inspection period and at the conclusion
,
of the onsite inspection activities on March 24, 1988, and summarized
the scope and findings of the inspection activities.
The inspectors also discussed the likely informational content of the
inspection report with regard to documents or processes reviewed by the
inspectors during the inspection. While the inspectors did review some
proprietary material, none of the areas expected to be contained in the
report were identified by the licensee as proprietary.
The licensee
ackn1wledged the findings of the inspection
i
e
25
.
.
. .
.
. . .
.
_
.__
CONFIRETORY ACTION LETTER
.
l
,
CAL-RIII-88-03
Docket No. 50-374
Comonwealth Edison Company
ATTN: Mr. Cordell Reed
Senior Vice President
Post Office Box 767
Chicago, IL 60690
Gentlemen:
This letter confirms the telephone conversation between E. G. Greenman of
this office and you on March 17, 1988, related to the dual recirculation pump
trip and related core performance anomalies occurring at LaSalle Unit 2 on
March 9, 1988. With regard to this event, we understand that you will:
1.
Perform an evaluation of reactor performance during this event including
secondary systems, the reactor protection system, and ATWS systems.
2.
Perform an evaluation of operator performance during this event.
3.
Evaluate the adequacy of your Technical Specifications, operating
procedures, abnormal operating procedures, and emergency procedures with
respect to this event and vendor recommendations (GE SIL-380).
4.
Perfonn increased activity level sampling during Unit 2 startup to
verify no abnormalities.
5.
Submit to NRC Region III a formal report of your findings and conclusions
within 30 days of receipt of this letter.
Throughout this investigative effort, we understand that you will take those
actions necessary to ensure that complete documentary evidence of the
conditions being examined is maintained, and furnished to the NRC's Augmented
Inspection Team which was initiated on March 16, 1988.
Q&$WV
y
CONFIRMATORY ACTION LETTER
_
CONFIRMATORY ACTION LETTER
.
.
Comonwealth Edison Company
2
We further understand that startup of Unit 2 will occur after a detennination
has been made by the AIT Team Leader that equipment performance was satisfactory,
including verification of reactor coolant samples, and that interim procedural
changes are satisfactory including a requirement to trip the reactor if no
reactor recirculation loops are in operation, and the reactor is in operational
conditions 1 or 2, and with concurrence of the Regional Administrator or his
designee.
Issuance of this Confinnatory Action Letter does not preclude the
issuance of an order requiring implementation of the above commitments.
None of the actions specified herein should be construed to take precedence
over actions which you feel necessary to ensure plant and personnel safety.
Please advise us immediately if your understanding differs from that set
forth above.
Sincerely,
Orignial Signed by A. Bert Davis
A. Bert Davis
Regional Administrator
cc:
D. Butterfield, Nuclear
Licensing Manager
G. J. Diederich, Plant
Manager
DCD/DCB (RIDS)
Licensing Fee Management Branch
Resident Inspector, RIII
Richard Hubbard
J. W. McCaffrey, Chief, Public
l
Utilities Division
l
David Rosenblatt, Governor's
!
Office of Consumer Services
J. M. Taylor, DED0
l
T. E. Murley, NRR
l
E. L. Jordan, AEOD
'
J. Lieberman, OE
M. Johnson, ED0
W. Lanning, NRR
F. Miraglia, NRR
G. Holahan, NRR
D. Crutchfield, NRR
J. Partlow, NRR
J. Strasma, RIII
J. Goldberg, 0GC
0. Muller, NRR
CONFlitMATORY ACTION LETTER
,
l
l
_