ML17333A100
| ML17333A100 | |
| Person / Time | |
|---|---|
| Site: | River Bend  |
| Issue date: | 05/16/1989 |
| From: | Callan L, Constable G NRC OFFICE OF INSPECTION & ENFORCEMENT (IE REGION IV) |
| To: | |
| Shared Package | |
| ML17333A101 | List: |
| References | |
| 50-458-89-20, NUDOCS 8905220106 | |
| Download: ML17333A100 (22) | |
See also: IR 05000458/1989020
Text
APPENDIX
U.S.
NUCLEAR REGULATORY COMMISSION
REGION IV
AUGMENTED INSPECTION TEAM
NRC Inspection Report:
50-458/89-20
Docket:
50-458
Licensee:
Gulf States Utilities (GSU)
P.O.
Box 220
St. Francisville, Louisiana
70775
Facility Name:
River Bend Station
(RBS)
Inspection At:
RBS, St. Francisville, Louisiana
Inspection Conducted:, April 21-24,
1989
Operating License:
Team Members:
H." F. Bundy, Reactor Inspector,
RIV
D.
R. Lasher, Electrical Engineer,
W. F. Smith, Senior Resident
Inspector,
RIV
C.
D. Sellers, Materials Engineer,
Assisting
Personnel:
E. Ford, Senior Resident
Inspector
W. Paulson,
Project Manager,
Team Leader:
onsta
e,
ie,
eactor
rogects
Section
C, Division of Reactor Projects
te
Approved:
L. J.
an, Director, Division o
Reactor
Pro ects
Da e
ppgSg2(>10/>
890516
PVP
ALeCK O.'=0004
8
9
'I
0
0
OETAILS
1.0
Persons
Contacted
2.0
J.
E. Booker, Manager,
River Bend Oversight
J.
L. Burton, Supervisor,
Independent
Safety Engineering
Group
J.
W. Cook,
Lead Environmental Analyst, Nuclear Licensing
T.
C. Grouse,
Manager, Quality Assurance
(QA)
R. T. Davey, Senior Enginee~,
Design Engineering
J.
C. Deddens,
Senior Vice President,
River Bend Nuclear Group
D.
R. Derbonne, Assistant Plant Manager,
Maintenance
L. A. England, Director, Nuclear Licensing
R.
G. Finkenaur, Electrical Engineer
A. 0. Fredieu, Supervisor,
Operations
P.
E. Freehill, Outage
Manager
J.
R. Hamilton, Director, Design Engineering
D. 0. Hartz,
Duty- Operations
Supervisor
G.
K. Henry, Director,
QA Operations
R. J. King, Supervisor,
Nuclear Licensing
, J.
D. Mullen, Mechanical
Maintenance
Foreman
T. F. Plunkett, Plant Manager
A. Soni, Supervisor,
Environmental Qualification and Specification
K. E. Suhrke,
Manager, Project
Management
R. J.
Vachon, Senior Compliance Analyst
J.
Venable, Assistant Operations
Supervisor
D. Zemel, Mechanical
Maintenance
Supervisor
The
NRC inspectors
also interviewed additional
licensee
personnel
during
the inspection.
General
2.1
Descri tion of Event
At about
11:45 p.m.
on April 19,
1989,
a freeze
seal failed on
a 6-inch
service water line.
The freeze
seal
had
been established
to allow
inspection
and repair work on manual isolation valves to
a safety-related
auxiliary building cooler.
The bonnet of the manually operated
valve was
off the valve and the service water system
was in operation at the time of,
the event.
The failure of the freeze
seal
resulted in flooding portions of
the auxiliary building. Approximately 15,000 ga'lions of service water was
discharged
through the disassembled
valve covering portions of the 141-foot
level of the auxiliary building with water.
A portion of the water flowed
through holes in the floor under safety-related
480 Yac motor control
centers
(MCCs) onto nonsafety-related
cabinets
on the 114-foot level
containing disconnect
links and
a 13.8 kY/480 Vac transformer.
The
cabinets
were not designed to shed the water leading to an electrical
fireball that damaged
the cabinet
and components.
A 13.8
kV supply breaker
opened
aeenergizing
that cabinet
and two others
causing the loss of the
0
operating residual
heat
removal
(RHR) system,
normal spent fuel cooling,
and normal lighting in the auxiliary building, control building, and the
reactorbuilding.
The operators
isolated the service water system in
15 minutes
and restarted
RHR in 17 minutes.
No increase. in reactor
temperature
was observed.
Backup spent fuel cooling (service water) was
available but was not immediately needed.
Temperature
in the spent fuel
pool rose to 123'F at which time normal cooling water was restored
and
temperature
was returned to normal.
The
NRC AIT arrived on site on April 21, 1989.
Prior to that time, the
NRC
senior resident inspector
(SRI) monitored the licensee's
preliminary
response
to the event.
The AIT was composed of a materials engineer,
an
electrical engineer,
an SRI,
a Region
IV inspector,
a team leader assisted
by the
NRR Project Manager,
and the
NRC resident inspectors
stationed at
River Bend.
The AIT reviewed the event,
inspected
the affected
equipment
and
interviewed various managers,
operators,
engineers,
and technicians
involved.
2.2
AIT Tasks
Region IV, in consultation with NRR, formed an AIT on April 20,
1989.
The
AIT tasks
were specified in a charter letter to ter.
G. L. Constable
from
Nr. L. J. Callan.
These tasks
were:
2.2.1
2.2.2
2.2.3
Perform
a thorough review of the sequence
of events
leading to and
recovery from the event.
Review the operator
response
to the event.
Review the licensee's
use
and control of freeze. seals
including
contingency
measures.
2.2.4
Pcview electrical
system vulnerability and response
to this event.
2.2.5
Review the licensee's
plans for recovery including restoration
from
switchgear water damage.
2.2.6
Review licensee's
management
control of outage activities with regard
to this event, including any potential tie with previous
outage-related
problems.
3Jl
~At 7 I
3.1
Se uence of Events
The following sequence
of events
was established
by the
NRC inspectors's
a
result of review of the control
room (CR) log and interviews with licensee
personnel.
Note;
All times are best estimate,
Central Daylight Time.
A ril 18,
1989
-12 midnight
~Ail 19,
1999
6 a.m.
9 a.m.
6 p.m.
11:48 p.m.
The maintenance
foreman briefed operations
personnel
on
installation of freeze seals to allow repair work on Standby
Service Water Isolation Valves
1SWP*525 and
1SWP*524 for
Auxiliary Building Unit Cooler HVR*UC11B.
Installation of freeze seals
was in progress.
Night shift
provided turnover to day shift.
Freeze
seals
were declared
established
and verified by
opening unit cooler vent and drain valves.
Work began
on
valve repairs.
Work was turned over to night shift.
The failure of the unit cooler inlet line freeze
seal
was
detected
by a maintenance
mechanic,
assigned
duty as the
freeze
seal
watch,
who heard
a loud noise
and observed
the
flow of service water from the open
body of the inlet
isolation valve.
11:50 p,m.
The control operating
foremen
(COF) in the control
room (CR)
received
a telephone call advising him that there
was
leakage
past
a freeze
seal in the auxiliary building (AB)
on the 141-foot elevation.
11:51
p.m.
11:52 p.m.
11:53 p.m.
11:54 p.m.
The
COF instructed the
AB senior nuclear equipment
operator
(SNEO) to assess
the severity of the leakage
and
report damage.
The
AB SNEO arrived at the 114-foot elevation of the east
side of the
AB and observed water on the floor.
He
irmediately went through the tunnel
on the 123-foot
elevation to the west side
and observed water falling from
the overhead.
He then proceeded
to the 141-foot elevation.
The
AB SNEO arrived at the
AB 141-foot elevation
and
observed water flowing across
the floor.
AA column of water
approximately
6 feet high was observed
flowing out of the
open bonnet of a valve near
AB Unit Cooler HVR*UC11A or B,
which are adjacent to each other.
(It turned out to be the
inlet isolation valve for Unit Cooler HVR"UCllB.)
He
assisted
maintenance
personnel,
who were trying to install
the bonnet
on the open valve body.
A nuclear control operator
(NCO) received
an emergency call
from a firewatch at the
AB 141-foot elevation regarding
the
lt
ll:55 p.m.
ll:56 p.m.
11:58 p.m.
ll:59 p.m.
A ril 20,
1989
12 midni ght
flooding and notified the
COF.
At the COF's direction, the
NCOs attempted to contact the
AB operator,
who did not
respond.
The
NCOs noted that all AB safety-related
pumps were
operating in the crescent
area.
Also, an alarm
(H13P808)
was received indicating
a ground fault condition on
1 of
20 nonsafety-related
load distribution centers
(LDCs).
The
NCO initiated actions to identify the
LDC which was
grounded
and obtain additional
SNEOs to assist in the
while the
COF briefed the shift supervisor
(SS)
on the
recent events.
The
COF informed the
SS that he thought the
leak was
on
a Division II standby service water
(SSW) line
to an
AB unit cooler, but that
he was not certain with
regard to the division.
Alarms and other indications received in the
CR indicated
the following:
A partial loss of CR AC lighting occurred.
Backup
lighting was energized.
Reactor Protection
System
(RPS)
B normal motor
generator
(MG) set power was lost.
Division II containment isolation valves closed
as
a
result of the loss of power.
Residual
heat
removal
(RHR), which was operating in the
shutdown cooling. mode,
was lost as expected
when the
containment isolation valves closed.
Operators
entered
Abnormal Operating
Procedure
(AOP)-0010 to
respond to loss of RPS
B power by transferring this bus to
the alternate
source.
From
CR indications, the
NCOs
determined that the most probable
cause of the power-related
events
was
a trip of Breaker
NPSACB016, which at the time
was the sole feeder to LDCs 1A, 1B, IC, 1D, 1S,
and
1T.
An
NCO was dispatched
to the normal switchgear
(NSG) building
to check relay targets
on NPSACB016.
The
NCO at the
NSG building reported that all phases
on
Breaker
NPSACB016
showed instantaneous
over current.
The
AB SNEO called the
CR to report the power failure and was
asked to confirm that the leak was
on Division II SSW
piping.
The
AB SNEO reported that line tracing would be
required to confirm the division.
12:Ol a.m.
12:03 a.m.
12:04 a.m.
12:05 a.m.
The
COF received
a report from the scene that
smoke
was
emanating
from electrical
equipment in the
AB at the
114-foot level, elevation west.
The fire brigade leader
was
dispatched
to assess
the situation.
The
AB SNEO called back
to report that he thought the leak was from SSW Division II
piping, but that he could not be absolutely
sure without
tracing the line further.
The
COF dispatched
additional operators
to the
AB to confirm
the source of the leak.
The fire brigade leader called the
CR to report that Transformer
1NJS-LDC lA was not on fire,
but that cover plates
were buckled outward and
smoke was
still coming from the disconnect cubicle.
He reported that
the fire brigade would not be needed
because
the switchgear
was deenergized
and that he would proceed to the 141-foot
elevation to assist in controlling the water leak.
The
SS and
COF discussed
options for removing
SSW systems
from service
and other event information.
The
NCO verified
all isolations
occurred in accordance
with design
requirements
and restored
systems
to meet operational
requirements.
The
SS and
COF decided to isolate
and
remove Division II of
the
SSW system from service without positive confirmation of
the leak source.
12:06 a.m.
The
NCO isolated
the Division II SSW system in accordance
v ith the COF's directions.
12:08 a.m.
The
COF received reports from AB that th'e leakage
was
diminishing and that the bonnet
on the leaking
SSW valve was
nearly in place.
12:15 a.m.
The
RHR shutdown cooling was restored
by placing the
Division I
RHR System back in service.
1:50 a.m.
The water level in the
AB crescent
area
dropped to less than
floor level
as
a result of all four sump
pumps
running
continuously.
A ril 21,
1989
6:03 a.m.
Restored
power to component cooling water
(CCW)
pumps
and
reestablished
cooling water to spent fuel cooling heat
exchangers.
(Haximum temperature
in the spent fuel'ool was
123'F at this time.)
The delay in restoring
power to the
CCW pumps
was because all three
pumps received
power from
the
damaged
13.8
kV load center.
-7-
Review of 0 erator
Res
onse to the Event
On April 21 and 22, 1989, the team reviewed documentation
provided by the
licensee
and conducted
interviews with operations
personnel directly
involved with the loss of freeze
seal incident.
The interviews included,
among others,
the Senior Nuclear Equipment Operator
(SNEO); the Nuclear
Control Operator
(NCO), who was
a licensed
Reactor Operator;
the Control
Operating
Foreman
(COF), who was
a licensed
Senior Reactor Operator;
and
the Assistant
Operations
Supervisor.
The Shift Supervisor
(SS)
who had the
watch during the incident was unavailable for an interview (absent
due to
illness in family).
The team was able to obtain
a good insight as to the
operators'esponses
and the circumstances
to which they were subjected
while being called upon to act.
In the interest of brevity and clarity, the information provided in this
section of the report was expressed
as fact or observations
of the team
members;
however,
much of the information was based
on the viewpoint of
licensee
personnel
being interviewed and,
as such,
was not observed
directly by the team.
The first indication that
a problem existed
was at approximately
11:50 p.m.
on April 19, 1989.
The
NCO had received
several calls, from all of 'the
levels below the Auxiliary Building (AB) 141-foot elevation in the
Auxiliary Building (AB), reporting the presence
of water coming from the
overhead.
At the
same time, the
COF, who was in the back of the control
room, received
a call from the individual attending the freeze
seals
on
the 6-inch standby service water lines
on the
AB 141-foot elevation.
The
indivioual stated,
in a calm, matter-of-fact fashion, that
a freeze
seal
was leaking.
I(e did not identify the lines, nor did it appear that there
was
any need for emergency action.
Therefore, at the outset,
there
was
no
sense of urgency to the problem.
The
COF was in the back of the control
room because
he was also fulfilling
the du'ties of, the administrative
COF in his absence.
Normally, for the
refueling outage,
the licensee
assigned
four senior reactor operators
on
shift.
In addition to the shift supervisor
and the
COF,
a second
COF was
provided to perform administrative tasks,
such
as review of clearances.
A
duty operations
supervisor
was also provided to oversee
operations
department activity.
During this event, the administrative
COF was not
available
on shift.
At approximately
11:51 p.m., the
COF dispatched
an
SNEO to the
AB to
investigate.
It was not yet evident in the control
room what was
happening.
Upon arrival at the east
AB 114-foot elevation,
which is physically
separated
from the west side, the
SNEO found water on the floor.
He then
went back through the passageway
and proceeded
to the west
AB 114-foot
elevation where
he found water falling from the overhead,
particularly in
the vicinity of the 13.8 kV/480V load aistribution center
(LDC).
By
11:53 p.m., the
SNEO arrived at the leak site
on the
AB 141-foot elevation
and found the tlaintenance
Foreman
and
a mechanic struggling to install the
on
a valve.
The floor was flooded with 2 to 3 inches of water and
a
6-foot high column of water was flowing from the valve body.
The
SNEO,
immediately climbed up the scaffo1d to assist without first informing the
control
room of his findings.
At approximately ll:54 p.m., the control
room received
a call
on the
emergency
channel of the plant paging system (Gaitronics) from a fire
watch on the
AB 141-foot elevation informing them that flooding existed.
Attempts to contact the
SNEO that had been dispatched
were unsuccessful.
He did not hear the calls.
For approximately the next
5 minutes,
a number of electrical
problems
occurred,
as evidenced
by telephone calls
and control
room annunciators
(among the multitude of those already lighted due to 'the refueling outage),
but none provided any good clues
as to where the flooding was coming from.
A ground fault alarm was received indicating
a problem on
1 of 20 possible
nonsafety-related
LDCs.
Lighting went out in the AB, reactor building,
and control building.
The reactor protection
system
(RPS)
B motor generator
output was lost,
a number of containment isolation motor operated
valves
closed,
and shutdown cooling was lost.
Reactor Plant
Component
Cooling
Pump
B lost power, which resulted in a loss of the heat sink for spent fuel
pool cooling.
At this point, discussions
continued
between
the
NCO,
COF, and
to evaluate
the problem and
come
up with actions to recover from
the electrical
losses
and stop the flooding.
By process of elimination,
the
SSW system
was the only viable source of flooding.
was
on Division I, and since it was not yet known which division of SSW was
leaking, they determined that Division II should
be secured first.
Abnormal Operating
Procedure
"Loss of One
RPS Bus," was entered
to recover from the loss of
8 power so that the isolations
could be
cleared
and shutdown cooling restored.
The operators
monitored reactor
temperature,
which did not change appreciably.
During the next
5 minutes, confirmation was received that there
was
no fire
at the
LDC, but it was visually damaged.
An operator at the
AB 141-foot
elevation reported that he thought the
SSW leak was from Division II but
could.not
be sure without tracing the line further.
Meanwhile, the control
room
w
NCO was restoring isolated
systems that .were needed for safety.
Spent
f
1
1 temperatures
were being monitored,
and
no rise was seen.
The
ue
poo
to the
t rs considered
the alternative of supplying
SSW directly
component cooling side of the fuel pool coolers,
but chose
not to until
necessary,
due to chemistry considerations.
At 12:06 a.m., the
NCO isolated the Division II SSW loop and secured
the
pumps
as instructed
by the
COF, although it had not been confirmed that
Division II was the source.
Preparations
were being
made to isolate
and
secure Division I if required,
but reports
came to the control
room from
the
AB that leakage
was diminishing.
It was
now possible to finish
installing the valve bonnet,
which had been unachievable until
SSW flow
through the valve body, was reduced.
Normal shutdown cooling was restored within 17 minutes following the loss
of the
B bus.
The team considered that to be excellent time for
restoration of shutdown cooling, under the circumstances.
Considering the
large volume of water in the reactor vessel
due to refueling activities,
the negligible temperature
rise,
and extended
shutdown time, the 17-minute
interruption
appeared
to be of minor safety significance.
Technical
Specifications
allow RHR to be stopped for 2 hours2.314815e-5 days <br />5.555556e-4 hours <br />3.306878e-6 weeks <br />7.61e-7 months <br /> in an 8-hour period
under the circumstances
that existed at the time of the event.
The team concluded that operator response
to the event
was adequate,
but
might have
been
more timely had the following weaknesses
not existed:
Operations
apparently
considered
freeze
seals to be comparable to a
closed valve and were insensitive to the potential for failure because
of human error,'nd to the possible
consequences.
As a result, there
was
no special visibility given to freeze
seal
status for control
room
operators
through logs or contingency plans that might be needed
in
the event of failure.
Nor did turnover, verbal or written,
seem to
address
the freeze
seals that were installed
on
SSW Division II.
NEOs and
SNEOs apparently
v ere not adequately trained to ensure
the
control
room was informed of the exact nature of a casualty before
becoming involved in the recovery.
At the time of the. event, the watch section
was short
one
COF than
had
been normally provided for this refueling outage.
This placed
additional
burdens
on the remaining
COF in dealing with the event.
CHP-9186,
Revision 4, "Freeze Seal,"
Precaution
7.3
NOTE states
that
the freeze
seal
should
be tagged
as
a system boundary
as if it were
'any other type of isolation.
This was not done for the
SSW freeze
seals.'f this note
had
been
complied with, the level of awareness
on
the part of the control
room might have
been greater.
3.3
Control
and use of Freeze
Seals
Freeze
seals or ice plugs are routinely used in nuclear reactor fluid and
t systems
to drain or isolate
components
which for various
reasons
cannot
be conveniently valved out.
Basically,
a freeze
seal is produce
y
d b
chilling the outside of the pipe, usually with liquid nitrogen supplied to
a jacket surrounding
the pipe.
Eventually, the water at the inner surface
of the chilled pipe freezes
and the ice/water interface
grows towards the
center of the pipe and also along its axis.
Obviously, the above
description is rather oversimplified and there. exists certain
problem areas
about which care must
be taken,
both for the establishment
of an effective
freeze
seal
and the assurance
that the freeze
seal will be adequately
maintained for the expected
duration of the repair.
It is essential,
-10-
therefore, that written procedures
be established
and that the procedures
be followed.
It also is obvious that adequate
training in the exercise of
the procedure
be provided.
At River Bend, freeze seals
are produced
by both outside contractor
personnel
and plant maintenance
personnel.
Each organization
has its own
freeze seal
procedure.
The River Bend procedure,
Corrective Maintenance
Procedure
CNP 9186, "Freeze Seal" Revision 4, dated
November 8, 1988,
permits
a freeze
seal contractor to use his own procedure.
Indeed,
much
of the in-house
knowledge
on freeze
seal
production
was gleaned
from
observation of the freeze
seal contractor during the first refueling
outage.
A memo from operations
gA to the gA manager
dated April 21, 1989,
"Review
of Freeze
Technology Inc.," states that freeze
seal contractor personnel
are all trained by the
company
owner on an on-the-job type of training.
There was
no formal training or qualified personnel list for River Bend
maintenance
personnel.
The procedures
have
some notable differences.
The freeze
seal contractor,
Freeze
Technology International Inc., "Freeze
Plug Procedures,"
dated
September
15,
1987, requires installation of a temperature
measuring
.device into a sleeve in the chamber.
The River Bend procedure is
ambiguous in that it states
a resistor temperature
sensor
probe
be taped
to the pipe,surface,
but shows
a sketch with a thermocouple
protruding
from the chamber.
The freeze
seal contractor requires
recording of
temperature
every
5 minutes during establishment
of the freeze plug and
every
15 minutes while the plug is being held.
The freeze
seal
contractor
utilizes
a manifolded boot and controls flow by having liquid nitrogen
dropping from a vent, whereas
the River Bend procedure
vents
gaseous
In addition to temperature
measurement,
there are other
indications of freeze
seal conditions.
These are frosting of the pipe at
each
end of the boot and observation of water flow downstream
from the
freeze seal.
The freeze
seal contractor prohibits multiple seals
from a single nitrogen
bottle, but no such prohibition was stated in the River Bend procedure.
Other than
NDE of the pipe area in the vicinity of the freeze
seal prior to
placement of the freeze
seal
boot,
no other gC/gA involvement was required
by either procedure.
There
was
a gA review of the maintenance
work
package,
but no in-process
inspection
by gC was required.
In the incident of April 19, 1989, there were
no temperature
measuring
devices
used to monitor temperature.
Additionally, two freeze
seals
were
produced
from the
same nitrogen bottle.
Nitrogen flow was controlled by
observation of the plume at the vent,
and
some valve manipulation
was
required to produce
equal
plumes from both vents
and to maintain roughly
uniform plume size for the duration of the freeze.
Temperature
indication
was estimated
by the axial length of
rosting
on the pipe on either end of the
boot.
Because frost is more readily initiated or maintained
on empty pipe
-11-
than
one full of water, the more visible outlet ends of the freeze might
not have given any indication of loss of freeze.
This is speculative,
as the maintenance
individual on duty at the time the seal
was lost
stated that he did move around for observation of the frost from other
locations.
Following the loss of freeze
seal
event,
a revised maintenance
work
order
(NWO) package
was prepared to produce
seals in the
same lines but at
different locations.
The
new location chosen
was further away from the
valves (i.e., further upstream
on the inlet and further downstream
on the
outlet).
This was
done to permit a borescope
examination of the site of
the previous plugs.
Additionally, there were other changes.
These were:
1) the seals
would be produced
on a vertical run of pipe rather than
horizontal; 2) there would be
no flow in the line; and 3)
a freeze
seal
contractor,
Freeze
Technology International
Inc. (FTI), would produce the
, actual
seals.
FTI selected
the sites
and marked the area to be used.
The
paint was
removed from the marked area to permit a surface
examination of
the freeze
seal location.
The
gC inspector
performed
a magnetic particle
(dry) examination.
Six linear indications were found.
All were spirally
aligned at approximately
30 degrees
from the axis.
Two were
on the inlet
pipe and were tandem about
1 inch and 1/2 inch long with about 1/2 inch
between.
The other four were parallel
and
one was estimated
at 8 inches.
Because of the linear indications, the package
had to be modified to
require grinding out of the indications
and to require ultrasonic wall
thickness
measurements.
Light flapper wheel grinding failed to remove all
of the indications.
Ultrasonic wall thickness
measurements
were required
by the modified
NWO package to precede additional grinding so wall
thickness
measurements
were made.
Random spot measurement,
including the
ground areas,
showed
some locations with a wall thickness of less
than
.220 inch in several
locations.
Ninimum required wall thickness
was
.245 inch.
An engineering
analysis
and more maintenance
work package
modifications would be required prior to resumption of the work.
Observation-of the freeze
seal
work by the
NRC AIT member
was terminated at
this point.
3.4
Electrical
S stem Yulnerabilit
and
Res
onse
The team assignment
in this area
was to review the vulnerability of the
plant electrical
system to this type of event
and to review the manner in
which the electrical
system
responded
to this particular event.
During this inspection,
the team reviewed the available information on the
event
and its consequences
including the sequence
of events,
documents,
control
room logs, relevant procedures,
event description,
event location,
and drawings of affected
systems
ana equipment.
The team also inspected
the location of the event
and surrounding areas.
The team's specific work
activities included:
inspection of the damaged
13.8
kV load center;
-12-
verification of the path by which floodwater from the leak reached
the
damaged
load center;
reconstruction
and validation of the electrical portion of the event,
including the route by which the loss of the 13.8
kV bus caused
loss
of RPS
Bus
B and subsequent
closure of Division II Containment
Isolation Valves;
review for existence of potential
common cause
events
(see section 3.7
below); and
review to identify any ties between the current event
and any previous
events.
Inspection of the AB, in the vicinity of the leak,
showed extensive
flooding occurred
on the 141-foot level with subsequent
flow of water to
lower levels of the
AB through floor openings
and
down cable
and pipe
chases.
Also,
some electrical
equipment
on the 141-foot level
was wetted
by splashing
and spraying caused,
in part,
by operators
attempting to
replace
the valve bonnet to stop the leakage of water.
Electrical
equipment
on the next lower floor of the
AB -(114-foot level) was affected
mainly by water falling from openings in the floor above
and
down cable
and pipe chases.
Some electrical
equipment
on the 95-foot and 70-foot
levels were wetted slightly by water draining
down from the upper levels
and from water backing
up in floor drains
on these levels.
The immediate
effect of water flooding on electrical
equipment
was severe
damage to
LDC 1NJS-LDC 1A/B.
This was caused
by water from the 141-foot level
flowing through openings in the floor under
ViCCs 2J and 2L onto the top
of the load center
below causing
ground faults in the load center.
This, in turn, resulted in burning out the windings of step-down
Transformer
1NJS-X1A and
an electrical explosion in the adjacent
13.8
kV
manual
disconnect
switch bay.
Switchgear
1NPS-SWGIA Breaker
16 then opened
resulting in loss of power to Load Centers
1NJS-LDC 1C/D and 1S/T,
As a
result,
power to
Bus
B was lost giving a half scram
and closing the
Division II Containment Isolation Valves which resulted in loss of shutdown
cooling to the reactor core
and loss of cooling to the spent fuel pool.
Normal lighting to the reactor, control, and auxiliary buildings
was also
lost.
No safety-related
equipment
was
damaged
as
a result of this event.
The control
room operators,
upon observing the loss of power to the
Bus B, implemented
Abnormal Operating
Procedure
"Loss of One
RPS Bus," to restore
power to the bus, reset
the containment isolation,
arid
reopen the isolation valves to restore
shutdown cooling and spent fuel
pool cooling.
These actions
were taken in accordance
with preplanned
procedures
established
to handle the consequences
of this event.
The licensee
acted to restore lighting to the affected buildings by
establishing
temporary feeds
from an existing 480-volt construction
loop to
the
YiCCs which feed the lighting panels in the affected buildings.
This
-13-
action was assigned
a high priority because
having adequate
lighting in the
reactor, control, and auxiliary building was essential
for the recovery
and
restoration activities in these buildings.
The licensee
informed the team that an event with similar results,
not
involving a freeze seal,
had occurred in 1986.
The team reviewed
LER 86-005, Revision 1, together with the licensee's
Condition
Report 86-028 which showed that, although the event
was caused
by a
different source,
the
same transformer
was affected
and the result was
a
plant scram from startup
(3 percent power).
The analysis
concluded that,
had the openings
under the 2J and 2L NCCs on the 141-foot level
been
sealed,
the event would have
been terminated with less serious
consequences.
Initially, an
NWR was written to seal
these
openings
but it
was later determined that sealing the bottom edges of these
MCCs was
undesirable
because it would allow water that might enter the
HCCs to
accumulate
there
and consequently
sealing of these
openings
was not done.
The potential for damage to these
MCCs, caused
by heat
and combustion
products entering through these
openings
from a fire in the load center
located immediately below on the 114-foot level, was apparently not
considered
(see Section 3.7 below).
There was
no discernible effect on safety-related
equipment other than
minor moisture accumulation in some of the 480-volt NCCs,
and in some pull
boxes
and junction boxes for Rosemount transmitters that were located in
the splash
area of the 141.-foot level in the AB,
No loss of safety-related
equipment occurred.
The licensee's
failure to either plug the openings
under the 2J
and '2L NCCs
on the 141-foot level or arrange to divert water
draining through
them away from the
LDC NJS-LDC lA/B was
a major contributing
factor to the severity of the event.
These
openings constitute
a valid
path
by which heat
and combustion products
could enter the two safety-related
MCCs from a fire in the nonsafety-related
load center
immediately below.
This could cause their loss or damage'and
should
be considered
by the
licensee.
In discussions
with licensee
personnel, it appears
they are
considering fire sealing the openings
and providing
a drain duct under
each
MCC that would divert water collected under the
I'CC away from the
load center.
This appears
to be
a reasonable
solution since it would
drain away any water that collected under the
NCC, divert the drainage
away from the load center,
and prevent heat
and combustion products
from
gaining direct access
to the 2J
and
2L NCCs from a fire in the
INJS-LDC 1A/B
load center.
The AIT considers
the licensee's
actions to restore
shutdown cooling and
lighting to the reactor, control,
and auxiliary buildings appropriate.
From the
NRC inspectors
review of LER 86-005, Revision 1, and the
licensee's
Condition Report 86-028,
we concluded that action to either plug
the openings or to provide means for diverting water that could flow
through the openings
and fall onto the load center
below should
have
been
taken
as part of the recovery from the earlier event
and chat,
had such
action been taken,
the consequences
of the current event would have
been
-14-
3.5
much less
severe.
We conclude that the licensee's
proposed action to seal
the openings
on the 141-foot level or provide diversion of the
drainage'way
from the load center
on the 114-foot level should
be pursued.
Licensee
Plans for Recover
The AIT reviewed the licensee's
plans for recovering from the effects of
this event'including restoration of water damaged
switchgear.
The team
interviewed plant staff members that were involved in the recovery
and
restoration efforts.
Recovery from damage to electrical
equipment included
a walkdown of the affected areas to ascertain
the type and extent of damage
to electrical
equipment including wetting of electrical
equipment.
This
initial inspection
included the 141-, 114-, 95-, and 70-foot levels of the
AB and was conducted
by a team from the
RBS environmental qualification (Eg)
group which identified electrical
equipment that could have
been affected
and performed
a visual inspection for water intrusion into the equipment
components.
This initial inspection
was completed
and identified
a group
of electrical
equipment that
needed to be opened
up and partially
dismantled to inspect the internals for moisture ingress
and
damage.
This
latter activity was projected to be completed
by April 25,
1989.
Some
'reas
on the
AB 95-foot level were wetted
and
became
contaminated
zones
around
some of the electrical
equipment located there.
Inspection of this
equipment, for moisture intrusion and damage, will be done
by an
MWO when
clearance
to work in the contaminated
zone
can
be obtained.
Several
pieces
of equipment, particularly breakers
from MCCs in the splash
zone near the
break,
were found to have moisture inside.
These
were disassembled,
dried,
cleaned with solvent,
reassembled,
tested,
and returned to service.
The
damaged
13.8
kV Load Center
1NJS-LOC 1A/B will have the disconnect
switch
bay refurbished,
the.
transformer will be replaced,
the control
equipment will be cleaned
and tested,
and the individual breakers will be
dried, cleaned,
and terminals
and connections will be inspected for
corrosion.
Moving parts will be cleaned,
checked for corrosion
damage,
and
lubricated
as needed.
This work is being done under currently active
MWO RO-56269.
Lighting panels
supplying the reactor, control,
and
auxiliary buildings are currently being fed from the construction
loop
source
pending completion of recovery actions to the circuit breaker panels
of Load Center
1NJS-LDC lA/B.
The team believes
these actions to be
adequate.
3.6
Mana ement Control of Outa
e
The team reviewed this event to determine if there
was any tie to previous
events
during this outage.
Prior to the team's arrival, the licensee
had
evaluated
the event including root cause.
The licensee
performed
a
management
oversight risk tree
(MORT) analysis,
and
a
MORT chart was
prepared that indicated
inadequate
training as the principal root cause.
The
NRC review of this event led to the conclusion that
GSU underestimated
the hazards of using freeze seals.
As
a result controls were not
established
that would have required adequate
preplanning,
including
-15-
3.7
training,'rior to the use of freeze seals.
The problems associated
with
this event are sufficiently different from earlier outage
problems that no
strong connection to previous events
were noted.
Review of possible
Common Mode Failure in Auxiliar Buildin
Floodin
The team inspected
the various paths
taken
by water that leaked out of the
SSW system
and on to the
AB 141-foot elevation floor.
Consideration
was
given to the potential of flooding to become
a
common mode failure through
which redundant
systems
could be disabled.
The team noted possible
paths
on the east
and west ends of the 141- and ll4-foot elevations
where
safety-related
motor control centers
(MCCs) for redundant trains of
equipment, were installed.
Three possible
common cause
events that could result from water leakage,
such
as from the current event,
were examined.
One involved water flowing
down from the 141-foot level causing
a loss of safety-related
MCCs of
redundant divisions located
on the 114-, 95-,
and 70-foot levels of
the AB.
Another involved water flowing through openings
on the east side
of the 141-foot level of the
AB onto safety-related
NCCs and switchgear
located
on the 114-foot level at the
same time as'water flowing through
openings
under the 2J and
2L MCCs on the 141-foot level initiates
a fire in
the
1NJS-LDC lA/B load center.
Heat and combustion products
from this fire
can flow up through these floor openings into the safety-related
MCCs 2J
and
2L causing
loss of these
YiCCs.
The last postulated
event involves
flooding of safety-related
electrical
equipment located
on the 95- and
70-foot levels of the
AB caused
by water leaking
down from breaks or
fire fighting activities on higher levels coupled with water backing
up
from overloaded floor drains
on these
lower 'levels.
This last postulated
event is potentially more serious
because
the three
RHR pumps,
the reactor
core isolation cooling pump, the low pressure
pump and their
associated
MCC cabinets,
and the
pump and its local control panel
are
located
on these
two levels, which are the lowest levels in the AB.
Mith regard to our postulation of common cause
events,
we believe the
probability of occurrence of the first postulated
event
(redundant
equipment lost due to flow paths of water) to be small given that the
involved have the conduits entering from the top sealed with Duxseal, the
tops of the
NCCs are
one piece with a lip covering the side panels
which
acts to shed water,
and the lowest components
in the
MCCs are located at
least
6 inches
above the floor level.
The probability of occurrence of the
second postulated
event (loss of equipment in one train due to water with a
- loss in the other train due to an electrical fire in nonsafety
equipment)
is believed to be greater than that of the first but still to, be small
b
f th difficulty of getting enough of the splashing water inside
the par'tially sealed safety-related
YiCC to cause
shorts or gground faults.
The probability of occurrence of the last postulated
event (inadequate
capacity of drain system)
appears
to be large
enough to warrant
consideration.
Given that these levels are the lowest levels in the AB,
they will collect water from breaks
on higher levels because
of the
apparently quite limited capacity of the
AB floor drains to deal with
-16-
either water from large pipe breaks
or extended 'fire fighting activities.
Also, there is
a large
amount of safety-related
equipment
located
on those
levels.,
4.0
The use of freeze plugs, with the system operating
and with the loop not
securely isolated
(by valving it off), resulted in an opening in each line
lai ger than the design basis
opening
used to size the floor drains in the
area.
NUREG 0800, the Standard
Review Plan
(SRP), Section 3.6. 1,
Appendix
B (BTP ASB 3-1), -defines the design basis critical crack size
as
being "taken to be 1/2 the pipe diameter in length
and 1/2 the wall
thickness
in width."
Using this definition, the critical crack size
can
be
calculated
as (6x0.5) x (0.5x0.5)
= 0.75 sq. in. (assuming
pipe wall
thickness
as 1/2 inch) which is assumed
to be the design basis for sizing
the floor drain (assuming this 6-inch line was the largest in the area).
Each valve body opening (bonnet
removed)
could have
a nominal
area of
($ )~ Pi
= 28. 2 sq.
in. which appears
to far exceed
the design basis for
sizing the floor drains in the area.
SRP Section 9.3.3 states
that the
floor drain system
should
have the capacity to carry away the
maximum
potential flooding that could occur
from normal operation,
maintenance,
testing,
and postulated
accidents
(pipe break,
tank ruptures,
etc.)
as
required
by General
Design Criteria 4.
The River
Bend Station
Updated Safety Analysis Report,
Section 9.3.3.3,
states,
"The floor drainage
systems
servicing buildings which house
equipment
needed for safe
shutdown
and .accident prevention or mitigation
have sufficient capacity to minimize water buildups that could hamper those
activities."
The licensee
was requested
to review the floor drainage
system design to assure
that there is a sufficient basis to support the
statement.
The licensee
committed to review the adequacy
of the floor
drain system.
Conclusions
and Findin
s
4.2
4.3
GSU did not provide adequate
control over freeze
seal
work.
Specifically,
the licensee's
freeze
seal
procedure
did not clearly specify that
a
temperature
measuring
device.was
required to monitor the temperature
and
hence
the integrity of the freeze seal.
In addition,
two freeze seals
were
being fed liquid nitrogen from the
same nitrogen bottle which can lead to
seal failure; however, if the temperature
of the seal
had been monitored,
the degradation
in freeze
seal integrity should
have
been detected.
Technicians
monitoring the freeze
seal
received
no formal training.
Operations,
engineering,
and management
personnel
appeared
to be unaware of
the potential for a freeze
seal failure.
A freeze
seal failure was not
viewed as credible; therefore,
neither preplanning
nor an awareness
of
freeze
seal
status
was evident
among operations
personnel.
This lack of
awareness
delayed operator
response;
however,
the leak was isolated in
15 minutes.
The operators
responded
well under the circumstances
which
included flooding,,
a potential electrical fire,
and
a loss of lighting.
-17-
t
4.4
The electrical
system performed
as expected
under
the circumstances.
No
damage to safety-related
electrical
equipment
has
been identified.
4.5
The failure of the freeze
seal
on a 6-inch line resulted in flooding that
significantly exceeded
the design capacity of the floor drain system.
This event raised the potential for a
common
mode failure of certain
safety-related
equipment
due to the resulting collection of water in the
lowest levels of the AB.