ML16342D344
| ML16342D344 | |
| Person / Time | |
|---|---|
| Site: | Diablo Canyon  |
| Issue date: | 06/17/1996 |
| From: | Wong H NRC OFFICE OF INSPECTION & ENFORCEMENT (IE REGION IV) |
| To: | |
| Shared Package | |
| ML16342D343 | List: |
| References | |
| 50-275-96-12, 50-323-96-12, NUDOCS 9606260051 | |
| Download: ML16342D344 (14) | |
See also: IR 05000275/1996012
Text
ENCLOSURE
U.S.
NUCLEAR REGULATORY COMMISSION
REGION IV
Inspection Report:
50-275/96-12
50-323/96-12
Licenses:
Licensee:
DPR-82
Pacific
Gas
and Electric Company
77 Beale Street,
Room
1451
P.O.
Box 770000
San Francisco,
Facility Name:
Diablo Canyon Nuclear
Power Plant, .Units
1 and
2
Inspection At:
Diablo Canyon Site,
San Luis Obispo County, California
Inspection
Conducted: 'ay 13,
1996,
through
Hay 28,
1996
Inspectors:
D. Corporandy,
Project Engineer
P. Goldberg,
Reactor
Inspector
Approved:
on
C se
,
ctor
rogects
rane
C~gvlr c
Date
Ins ection
Summar
Areas
Ins ected
Units
1 and
2
Special,
announced
inspection of the
licensee's
actions in response
to the test results
obtained during the
April 1996 augmented
testing of the Unit
(MSSVs)
pressure
setpoints.
Results
Units
1 and
2
An Unresolved
Item was identified involving the licensee's
failure to
determine
the magnitude of AVK test equipment bias, correlation
factors
(CFs), for all
20 Unit
1 HSSVs during
1R7 as stated
in PG&E's
letter to the
NRC dated
November
1,
1996 (Section 1.3).
An apparent violation of 10 CFR Part 50, Appendix B, Criterion XVI, was
identified for the licensee's
failure to promptly identify and correct
out-of-tolerance
setpoints
on Unit
1
HSSVs following augmented
testing
of- the steam
1 HSSVs
on April 2,
1996 (Section 3.3).
9h0626005f 9606i7
ADQCK 05000275
8
,PDR
Two examples of an apparent violation of 10 CFR Part 50, Appendix B,
Criterion V, were identified for'the licensee's
failure to follow
procedures:
(1) to notify the Operations Shift Foreman of a deficient
condition, i.e., three of five HSSVs out-of-tolerance
high,
and (2) to
document
a prompt operability assessment
within 24 hour2.777778e-4 days <br />0.00667 hours <br />3.968254e-5 weeks <br />9.132e-6 months <br /> of
identification (Section 4).
NRC inspectors
identified two weaknesses
in the licensee's
HSSV
augmented
testing
program which may have contributed to the problems
encountered
during the April 1996- augmented testing:
(1) the
HSSV
augmented testing
program was not sufficiehtly formalized,
and
(2) the
licensee
did not sufficiently plan for the evaluation
process
and
actions to be taken, following the augmented testing
(Section 5).
Summar
of Ins ection Findin s:
Unresolved
Item 275/96012-01
was opened.
An apparent violation of 10 CFR Part 50, Appendix B, Criterion XVI, is
identified in Section 3.3 (Violation 275/96012-02).
Two examples of an apparent violation of 10 CFR Part 50, Appendix B,
Criterion V, are identified in Section
4 (Violation 275/96012-03).
Attachments:
Attachment
1 - Persons
Contacted
and Exit Heeting
Attachment
2 Acronyms
DETAILS
1
BACKGROUND
l. 1
Overview of A ril 1996
On April 2,
1996, the licensee
conducted
augmented
testing to determine
the
"as-found" pressure
setpoints of the five Unit 1,
Steam
1 HSSVs.
Testing
results
showed three of the five NSSVs,to
be out-of-tolerance
(high).
On
April ll, 1996,
the licensee
tested
the five Unit 1,
2 MSSVs and found
.
three of five HSSVs to be out-of-tolerance
(high).
On April 12,
1996,
preliminary results of the licensee's
evaluation of the test results following
the
2 testing determined that the out-of-tolerance
NSSVs would have
caused
pressure
to exceed
the
110 percent
design pressure
by
8 psi.
The licensee
reported this condition to the
NRC in accordance
with
The licensee did not test the
HSSVs
on the remaining leads until April 14,
1996.
At that time, five of five Lead
3 HSSVs
and
one of five Lead
4 HSSVs
were found to be out-of-tolerance
(high).
The licensee's
subsequent
evaluation of the condition indicated that the
would
have
exceeded
110 percent of design
pressure
by 27 psi.
The inspectors
reviewed these
events,
the available data at the time,
and the
licensee's
actions
based
on the test results.
1.2
Desi
n Bases of Diablo Can
on
Diablo Canyon
has five NSSVs
on each of four separate
main steam leads for a
total of 20 MSSVs per unit.
Each of the five NSSVs
on
a lead
has
a different
set pressure
varying from a low nominal set pressure
of 1065 psig with a
tolerance of -2 percent,
+3 percent to
a high set pressure
of 1115 psig with a
tolerance of +3 percent.
The primary function of the
HSSVs under
a design
basis
accident is to prevent
pressure
from exceeding
110 percent of design pressure
(1194 psig).
In addition,
another function of
the
HSSVs is to limit steam generator
pressure
to allow the auxiliary
(AFW) pump to maintain required flow greater
than
440 gpm.
The
licensee
has submitted
a license
amendment
request
to reduce the
AFW minimum
flow requirement
from 440
gpm to 410 gpm.
The flow is affected
by the lift
pressure
of the lowest lifting MSSV.
The nominal set pressure
of the lowest
set
MSSV is 1065 psig.
1.3
Histor
of Diablo Can
on
Out-of-Tolerance Setpoints:
Diablo Canyon
has
had
an history of HSSVs being
found outside of the Technical Specification
(TS) tolerance.
On February
9,
1994, for Unit 1,
and
on March 5,
1994, for Unit 2, while performing setpoint
testing
on the
NSSVs,
13 Unit
1 valves
and
15 Unit 2 valves did not meet the
TS setpoint tolerance of +1 percent.
The setpoints of the valves
were
out-of-tolerance,
both high and low.
Trevitest test equipment
had
been
used
to perform the tests.
The licensee's
operability evaluations
determined that
neither of the units would have
exceeded
their design basis with the
out-of-tolerance
NSSVs.
During the
1R6 outage
which followed the February
1994 testing,
the licensee
removed the Unit
1 NSSVs
and sent
them to the
Service
Center test facility where the valves were refurbished
and reset
using live steam.
Just prior to the Unit
1
1R7 refueling outage in September
1995, the licensee
performed preoutage
set pressure
testing
on the Unit
1 MSSVs.
The as-found
results of the tests
were that
19 of the
20 NSSVs exceeded
TS tolerances
with
the following detailed results:
12
NSSVs with initial set pressure
> 6 percent
Il
3
6 percent
3
1
3 percent.
Additional testing
was conducted
on the valves
and eight of the twenty
required 'adjustment.
The other out-of-tolerance
valve setpoints drifted into
the acceptable
range
as
more tests
were run.
Prior to this testing,
the
licensee
had
used Trevitest test equipment for MSSV in-situ testing.
Commencing with this testing,
the licensee
switched to AVK test equipment.
The licensee's
operability evaluation
determined that the out-of-toleranc'e
NSSVs would have caused
the steam generator
pressures
to exceed the
110 percent
design allowable.
In response
to the Unit
1 findings, the licensee initiated testing the
20
Unit 2 MSSVS.
On September
22-24,
1995, the licensee
completed testing
16 of
the
20 valves. 'he as-found results of the tests
were:
5 MSSVs with initial set pressure
> 3 percent
5
II
1 - 3 percent
5
<
1 percent
1
< -3 percent.
After the sixteenth valve had
been tested,
Unit 2 experienced
and four of the low setpoint
NSSVs lifted.
The valves were supposed
to open
at
1065 psig,
but all four. opened
low, between
1023 psig and
1050 psig.
Prior
to the plant trip, the licensee
had adjusted
Valve RV-3 by lowering the
setpoint,
since the as-found setpoint
had
been out-of-tolerance
on the high
side.
It was also noted that at the time of the reactor trip, RV-7, the valve
which lifted at
1023 psig,
was in the process of being adjusted,
and the
adjusting nut/locknut assembly
had not 'yet been retightened
when the reactor
trsp occurred.
The licensee
used the
AVK equipment to test the valves.
Individual
MSSV Setpoint Distribution:
After the plant trip, the licensee
performed
a series of tests
using the
AVK test
equipment to determine
valve
setpoints.
Selected
valves were set pressure
tested
10 to 20 times each.
The
licensee
concluded that
each valve had
a defined setpoint distribution in the
shape
of'
bell curve.
They also concluded that this distribution was valve
dependent
and could range outside of the allowable tolerance.
The licensee
stated that adjustments
to the valve setpoint
should only be made
once the
setpoint distribution curve was
known.
Once
known, the licensee
concluded
that the valve could
be adjusted
by moving the
mean setpoint value.
'License
Amendment to Increase
Allowable MSSV Setpoint Tolerance:
On
September
30,
1995, the licensee
submitted
License
Amendment
Request
95-06,
"Request for Emergency
Review and Approval of Change to TS 3.7. 1. 1,
Table 3.7-2-2
Increase
in Setpoint Tolerances for HSSVs."
The request
was
made to change
the set pressure
tolerance of the
HSSVs from +1 percent to
+3 percent
and -2 percent for the lowest set pressure
HSSVs
on each
steam lead
and
+3 percent for the remaining
HSSVs.
The licensee
made this request,
since
they concluded that each
HSSV had
a setpoint distribution specific to each
valve that might exceed
+1 percent,
and because
pressure
and
AFW flow were demonstrated
by analysis to remain within the original design
basis with the proposed
increased
setpoint tolerances.
In addition, the
licensee
stated that the liftdistribution would occur whether the valve was
set
on live steam or with an hydraulic lift device.
The licensee
based this
conclusion
on the tests results
from the September
12-29,
1995, testing.
As a.
condition of the
amendment,
the licensee
committed to perform augmented
testing of all the
HSSVs based
on
a test schedule starting
3 months after the
seventh refueling outages for the respective units with the initial augmented
testing to involve one steam lead at
a time.
Testing Methods:
Both Trevitest
and
AVK test
equipment
use,an
hydraulic lift
assist
methodology whereby the steam pressure
present
at testing is "assisted"
by an additional hydraulic force to lift the safety valve disk off its seat;
hence,
providing
an onset of steam flow through the safety valve.
The
principal difference
between
the Trevitest
and
AVK test methodology is that
Trevitest detects
the onset of safety valve opening
by relying on the test
technician to hear
a hiss or pop from the valve,
whereas
the
AVK test
equipment
uses
an acoustic
sensor to electronically detect
the onset of safety
valve opening.
AVK Correlation Factor
(CF):
The licensee
stated that in 1994 they had
conducted
a test
program testing the
HSSVs at the Westinghouse
Service Center
using both live steam
and the
AVK device.
The licensee
stated that they found
that there
was good correlation
between
the setpoints
measured
on live steam
and the
AVK device.
In
PGEE Letter DCL-95-241 of November
1,
1995,
the
licensee
committed to perform additional testing during the seventh refueling
outages for each unit to determine
the magnitude of the
AVK test equipment
bias
compared to live steam.
During the seventh refueling outages for each
unit, the licensee
measured
AVK test method liftdistributions to develop
between-live
steam
and
AVK testing.
Apparently,
due to equipment
problems at
the test facility and the licensee's
decision not to .delay the refueling
outage,
CFs were developed for only nine of the
20 Unit
1 HSSVs.
CFs were
developed for all five of Steam Generator
(SG)
1-1
HSSVs,
and four of five
SG 1-2 HSSVs.
have not yet been
developed for the
HSSVs of SGs
1-3 and
1-4.
CFs for the nine Unit
1 HSSVs averaged
about
1 percent with a maximum
of 1.4 percent.
The licensee
did not inform the
NRC of the decision not to determine
the
magnitude of AVK test equipment bias for all 20 Unit
1 HSSVs
as committed in
the November
1,
1995, letter until after restart
from 1R7.
This is identified
as
an Unresolved
Item pending the licensee's
review and discussion with the
NRC of the basis for the decision
(URI 50-275/96012-01).
1.4
Periodic
TS Surveillance Testin
Versus
Au mented
HSSV Testin
The licensee
uses
AVK test equipment for both periodic
TS surveillance testing
and augmented
HSSV testing to obtain
HSSV setpoint pressures.
As-found
setpoints
are required to be within the
amended
TS tolerances.
One
HSSV at
a
time is'ested.
Before proceeding
to the next
HSSV to be tested,
the tested
HSSV is returned to within +1 percent of its nominal setpoint.
.This is
demons6 ated
by achieving
two consecutive lifts within the required tolerance.
Uncorrected out-of-tolerance
HSSVs are subject to the actions required
by
TS 3/4.7. 1 "Turbine Cycle, Safety Valves."
All 20 HSSVs of a unit are tested
during the periodic
TS surveillance testing.
Under the licensee's
commitment to perform augmented testing of the
HSSVs
(refer to
PG&E Letter DCL-95-241 dated
November
1,
1995,
and
LER 1-96-003-00),
the Unit
1 augmented testing involves testing the
HSSVs of one steam
at
a time conducted
on
a staggered
basis at
3 month intervals.
The
TS
surveillance testing is performed
under licensee
Test Procedure
STP H-77,
"Safety and Relief Valve Testing," which requires verification that the valves
meet lift setpoint requirements
of the
ASHE Boiler and Pressure
Code,
Section
XI, 1977, with Addenda through
Summer of 1978.
ASHE Section
XI requires
testing of additional valves, if any valves in the sample set are found to be
out-of-tolerance.
1.5
Assum tions
on the Causes of HSSV Set Pressure Drift
The licensee
postulated
causes for the set pressure drift of the
HSSVs
as
follows:
Thermal bonding/Hiero-welding:
The
HSSVs have different materials for the
disc
and nozzle seats.
The licensee
postulated that,
since the nozzle
and
disc seats
were of different materials with different coefficients of thermal
expansion,
during valve heatup the valve seats
could gall.
The licensee
postulated that galling would provide
a mechanism for micro-welding and
thermal
bonding of the seats
which would cause
the first liftto be high.
In
April 1996,
the licensee
believed that this was the probable
cause of the high
initial,set pressures.
The licensee further postulated that subsequent lifts
would tend to exhibit lift pressures
which would drift to the initial set
pressure
assuming
the valves
remained
at temperature.
The licensee
also
considered
that time,interval
between lifts might be
a factor in the thermal
bonding of the valve seats.
Setting
on Live Steam Versus
AVK Hethod:
The licensee
stated that differences
were observed in lift pressures
depen'ding
on whether the lift was achieved
entirely due to steam pressure
or in part due to steam pressure
plus hydraulic
I'
lift assist
from the
AVK test equipment
used at Diablo Canyon.
During the
April Unit 2 outage,
the licensee
performed
HSSV set pressure
testing
on the
20 Unit 2 valves at
a test facility.
The licensee
noted that there were
differences
in measured
set pressures
between testing using only steam
pressure
to initiate and measure lift point and testing
using partial
steam
pressure
(approximately
90 percent of the lift point) plus hydraulic assist
from the
AVK test
equipment.
The licensee
developed
CFs for each valve and,
as noted
above,
found that
on average
the
CFs were
+1 percent with a maximum
of +1.4 percent.
Lift Distribution/Signature:
As mentioned in Section
1.3 of this report,
the
licensee
concluded
from additional testing in September
1995, that each valve
had
a defined setpoint distribution similar to
a bell curve.
The licensee
also concluded that this distribution was valve dependent
and could range
outside of the setpoint tolerance.
This was
one consideration
in the
licensee's
application for a
TS amendment to increase
the allowable
HSSV
out-of-tolerance
setpoint values.
It appears
that the licensee
recognized this
phenomenon
as
a result of
investigating out-of-tolerance
problems with Diablo Canyon's pressurizer
safety valves.
The pressurizer
safety valve design is similar to the
HSSV
design.
The licensee identified that under lo'ading the safety valve spring
experiences
minute buckling which apparently
caused
the safety valve setpoint
pressure
repeatability problems,
an industry-wide problem.
The licensee
sponsored
development of a prototype valve with a modified upper
spring washer
assembly
designed
to reduce the spring buckling and pivoting.
Preliminary results of the prototype testing
showed
a significant reduction in
liftdistribution (i.e., significant improvement in setpoint pressure
repeatability).
The valve manufacturer
acknowledged that the modification
would not affect the ability of the valve to lift, relieve its rated capacity,
and close (i.e., maintain its overpressure
protection function and reclose
after blowdown).
2
SE(UENCE
OF
EVENTS
On April 2,
1996, the licensee
performed
augmented testing
on the five Unit 1,
1
HSSVs
as specified
in" the licensee's
TS amendment
submittal.
All five
1
HSSVs
had
AVK CFs.
Upon completion of the
1 testing
(and
resetting of HSSV lift setpoints
as necessary),
the licensee
noted that the
as-found setpoints
of three of the five HSSVs were out-of-tolerance
(high).
Initially, the site engineering
personnel
responsible
for performing the test
decided not to document
a prompt operability assessment
(POA),
because
they
believed that the conditions
observed
on April 2 were enveloped
by conditions
already evaluated
in a previous operability evaluation
(OE 94-02,
Revision 4,
"Operability of HSSVs with Potentially High Initial Lift Setpoints").
According to the licensee,
although the decision
was
made not to document
a
POA, offsite engineering
was tasked with analyzing the as-found conditions
observed
on the five Lead
1
HSSVs
and projecting the results to the other
three Unit
1 leads.
In addition to projecting the out-of-tolerance
high lift
setpoints
to the
HSSVs
on the other leads,
the in-tolerance
HSSVs were
assumed
to be high at the maximum allowable
(3 percent
above nominal) setpoint,for
their respective
HSSVs.
Engineering
also modelled
an additional
3 percent of
nominal pressure
to account for pressure
accumulation during the initial 'lift.
The results of the analysis
showed that the maximum allowable steam generator
pressure,
110 percent of design pressure,
would not have
been
exceeded.
On April 4,
1996,
Engineering
conveyed the results of this analysis
as
an
update to the action request
associated
with the April 2 testing.
Engineering
did not document their evaluation of the adequacy of AFW flow.
However, in
response
to questioning
by the inspectors,
offsite engineering
explained that
they did not document
an evaluation of AFW flow adequacy
because
they did not
consider it to be
a problem.
At the time, they did not believe it to be
a
problem because
AFW flow is not credited in the accident
analyses
until
60 seconds after the beginning of the design
basis event.
AFW flow depends
on
the lift pressure
of the
HSSV with the lowest lift point.
For the first
60 seconds
an
HSSV would be expected
to cycl'e open several
times.
The April 2
testing
showed that once the initial lift of the low setpoint
HSSV was
achieved,
subsequent lifts demonstrated
a lower lift pressure
than the initial
lift pressure.
This data
was consistent with the licensee's
thermal
bonding/micro-welding assumption.
On April 8,
1996, the licensee
discussed
the
1 test results with NRC
personnel
(Region- IV and
NRR).
During the April 8 discussion,
the licensee
expressed
their intention to tes't the Unit
1
2 HSSVs.
It was noted that
four of five of the
2 HSSVs
had
AVK CFs.
Upon completion of the
2 testing
on April ll, 1996, the licensee
noted
that the as-found setpoints- of thre'e of the five HSSVs
had
been
out-of-tolerance
(high).
On April 12,
1996,
the licensee's
operability
evaluation
concluded that
110 percent
maximum allowable pressure
(Steam
Generator
1-2) would have
been
exceeded
as
a result of the out-of-tolerance
2 HSSVs.
The licensee
reported
the condition to the
NRC in accordance
with 10 CFR 50.72.
On April 12, the
NRC questioned
the licensee
as to the
potential for the
HSSVs
on the other two untested
leads to be out-of-tolerance
and whether the plant remained within its design basis.
The licensee
erroneously
responded
that the plant remained within design basis
based
on the
first ten
HSSVs having been reset to +1 percent.
The licensee
did not test the
HSSVs
on the other 'two 'leads until Sunday,
April 14,
1996.
The April 14 testing revealed that five of five HSSVs
on
3
1-3) were out-of-tolerance
(high)
and
one of five
HSSVs on Lead
4 was out-of-tolerance
(high).
According to the licenseee's
calculations,
the
3 results
would put Steam Generator
1-3 outside its
design basis.
The inspectors
also noted that at the conclusion of the April 11,
1996,
testing,
one
HSSV on Lead
2 was left at 1.2 percent out-of-tolerance,
and at
the conclusion of testing
on April 14,
1996,
one
HSSV on Lead
3 was left at
1.2 percent out-of-tolerance.
This was discussed
with the
NRC.
On April 21,
1996,
the two HSSVs were reset to within +1 percent
using the
AVK test
equipment.
3
UNIT
1
AUGMENTED TESTING
3. 1
Evaluation of Data Followin
the
A ril 2
1996
Testin
Evaluation of OE 94-02, Revision 4:
Following the April 2,
1996, testing of
the Unit
1 Lead
1 HSSVs, onsite engineering
decided not to document
a
POA,
because
they believed that the conditions
observed
on April 2 were enveloped
by conditions already evaluated
in
a previous operability evaluation
(OE 94-02,
Revision 4).
The inspectors
reviewed
OE 94-02,
Revision 4,
and the
April 2 test results.
The inspectors
observed that
OE 94-02,
Revision 4,
evaluated
AFW flow adequacy
based
on the
1065 psig nominal setpoint valve
lifting at
3 percent high.
The April 2 as-found lift setpoint of the Unit 1,
1
1065 psig valve was 7.5 percent high,
and therefore,
not enveloped
by
the
OE 94-02 evaluati'on
as
assumed
by the licensee.
The licensee
'acknowledged
the inspectors
observations.
Evaluation Considering
September
1995 Unit
1 Test Results:
The inspectors
noted that in addition to
OE 94-02,
information from the recent
September
1995
test results
was available.
The inspectors
compared
the April 2 Lead
1 test
results with the September
1995 data.
The inspectors
noted that the April 2,
1996, test results
were generally consistent
with the results of the September
1995 testing in that the as-found setpoints
were all above the nominal
setpoint.
The only exception
was
HSSV RV-6, which had
an as-found setpoint
1.9 percent
below the nominal setpoint.
This apparent
anomaly
may be
explained
by previously established
setpoint behavior characteristics,
namely,
the setpoint bell curve distribution and the
AVK CF for RV-6.
The inspectors
noted that the licensee's
evaluation of the September
1995 out-of-tolerance
HSSVs
was available at the time and that it demonstrated
that the
110 percent
design basis
pressure
would have
been
exceeded.
Since the
Unit
1 April 2,
1996, test results
were generally consistent with the Unit
1
September
1995 test results, it would seem prudent to have considered
the
September
1995 results
as well as the April 2,
1996, results
when evaluating
the potential for out-of-tolerance
conditions to exist
on the remaining
untested
Unit
1 HSSVs.
3.2
Res
onse to Results of A ril 11
1996
Testin
On April 11,
1996,
the licensee
tested
the Unit
1 Lead
2 HSSVs
and found three
of five Lead
2 HSSVs to be out-of-tolerance
(high).
The licensee
performed
a
POA which was completed
on April 12,
1996.
The
POA showed that the
out-of-tolerance
HSSVs would have
caused
the
110 percent design basis
pressure
to be exceeded.
The licensee
did not test the remaining ten Unit
1
HSSVs until April 14,
1996.
The inspectors
considered
that
on April 12,
1996, the licensee
had determined
that the out-of-tolerance
Unit
1
2 HSSVs would have placed the plant in a
-10-
condition outside of its design basis.
These conditions were steam lead
dependent.
In other words, the fact that the
2 HSSVs
had
been reset
had
no bearing
on the condition of the
MSSVs on steam
3 and 4. If the
HSSV
out-of-tolerance
conditions
on steam
3 or 4 were similar to those
found
on the'ead
2 HSSVs,
then their respective
SG would also
be outside of the
design basis.
The inspectors
considered that the test information available to the licensee
after completion of the
POA on April 12, provided
a reasonable
doubt about the
setpoint tolerances
on the remaining
10 Unit
1 HSSVs
and their ability to
. maintain their respective
within the design basis in the
event of a postulated
accident.
3.3
Conclusion
I
The inspectors
considered that the lack of a documented
POA contributed to the
licensee's
incomplete evaluation of the April 2 test results,
which resulted
in a missed opportunity to promptly identify and correct the out-of-tolerance
conditions which existed for HSSVs
on the untested
steam leads.
Consequently,
the licensee
did not recognize
the potential for the untested
HSSVs to place
the steam generators
outside the design basis
(110 percent of design
pressure),
or to possibly exceed
the
TS setpoint tolerances.
On April 2,
1996, the pressure lift setpoints
on three of five HSSVs
on main
steam
1 were identified by testing to exceed their allowable
TS
tolerances,
a condition adverse
to quality.
Failure to promptly identify the
- possible out-of-tolerance conditions
on the remaining Unit
1
HSSVs resulted in
placing the plant in a condition outside of 'its design basis.
The
HSSVs
on
2 were not tested until April 11,
1996,
when three of five
MSSVs were found to be out-of-tolerance.
The HSSVs
on main steam
3 and
4 were not tested until April 14,
1996,
when six .of ten
HSSVs were found to be
out-of-tolerance.
The operability evaluation,
which modelled the out-of-tolerance
HSSVs
showed
that the maximum allowable
SG pressures
on
SG 1-2 and 1-3, would have
been
exceeded
by 8 and
27 psi respectively
under design basis conditions.
Based
on the above,
the inspector identified an apparent violation of
10 CFR Part 50, Appendix B, Criterion XVI, which requires,
in part, that
conditions
adverse
to quality such
as deficiencies
are promptly identified and
corrected
(Violation 275/96012-02).
NRC Inspection
Manual, Part 9900,
"Operable/ Operability:
Ensuring the
Functional Capability of a System or Component," states
that timeliness of
corrective actions is determined
by the safety significance of the issue.
Specifically,
"The Allowed Outage
Times contained
in TS generally provide
reasonable
guidelines for safety significance."
Diablo Canyon's
TS allowed
outage
time for the
HSSVs is 6 hours6.944444e-5 days <br />0.00167 hours <br />9.920635e-6 weeks <br />2.283e-6 months <br />.
The inspectors
determined that the
licensee's
corrective actions to identify and restore
HSSV setpoints within TS
allowed tolerance
were untimely.
4
IMPLEMENTATION OF
PROCEDURE
FOR EVALUATING DEGRADED CONDITIONS
The inspectors
reviewed licensee
Procedure
OM7.IDS, Revision 2, "Operability
Evaluation"
and the licensee's
implementation of the procedure
during the
April 1996 augmented
testing of the Unit
1 HSSVs.
The inspectors
identified
two examples of an apparent violation for the licensee's
failure to follow
Procedure
OM7. IDS.
10,CFR Part 50, Appendix B, Criterion V,, requires that
activities affecting
quality shall
be prescribed
by documented
procedures
and shall
be accomplished
in accordance
with "those procedures.
Diablo Canyon Procedure
OM7. IDS,
Revision 2, Subsection
2.2.3,
requires that:
"For Degraded
Conditions
impacting Structure,
System
and
Component operabi.lity identified by physical
evidence at
DCPP,
the
POA should
be completed
and documented
during the
operating shift in which the physical
evidence
was identified.
In all cases,
the
POA shall
be completed
and documented within 24 hours2.777778e-4 days <br />0.00667 hours <br />3.968254e-5 weeks <br />9.132e-6 months <br /> following
identification of a Degraded Condition."
On April 2,
1996, licensee test engineers
identified three of five HSSVs
on
steam
1 of Unit
1 to be out-of-tolerance
(high), but as of May 14,
1996,
NRC inspectors
identified that the licensee
had not documented
a
POA of the
degraded
condition
(namely three of five NSSVs out-of-tolerance
high).
This
is considered
as
an example of an apparent violation for the failure to meet
the requirements
of 10
CFR Pa} t 50, Appendix 8, Criterion
V and Procedure
OH7.IDS (Violation 275/96012-03).
Diablo Canyon
Procedure
OH7. IDS, Revision 2, Subsection
4. 1 requires that
"The
individual
and his/her group supervisor identifying a Degraded
Condition or an
Issue
Needing Validation is responsible for:
Immediately notifying the Shift
Foreman, if the condition is an observed
physical
Degraded
Condition at the
plant that could adversely affect the Operability of a Structure,
System,
and
Component."
On April 2,
1996, licensee test engineers
identified three of five HSSVs
on
steam
1 of Unit
1 to be out-of-tolerance
(high), but did not notify the
Shift Foreman.
This is considered
as another
example of an apparent violation
for the failure to follow the requirements of 10 CFR Part 50, Appendix B,
Criterion
V and Procedure
OH7. IDS (Violation 275/96012-03).
The first example of the violation may have contributed to the delay in
testing the other
NSSVs following the April 2 Lead
1
NSSV testing.
The second
example,
in effect,
excluded Operations
from the process of responding
to the
NSSV deficiencies identified on April 2.
5
WEAKNESSES IN HSSV AUGMENTED TESTING
PROGRAM
Interviews of licensee
personnel
who performed the initial augmented testing
revealed that they had considered
the test to be important for gathering data,
but did not consider that it would require the
same actions in response
to
test results
as would be required with periodic surveillance testing.
The
-12-
inspectors
noted that one of the apparent violations, failure to inform the
Operations
Unit Shift Foreman of the degraded
condition (three of five HSSVs
out-of-tolerance)
on April 2,
1996, resulted
in part,
because
the individuals
who performed the augmented testing
had not considered
the test to be governed
by the licensee's
formal procedures.
The inspectors
did note that shortly
following the April 2 augmented testing,
the individuals performing the
HSSV
testing were
made
aware of the necessity
to follow the licensee's
formal
processes
for reporting
and evaluating test results.
The inspectors
observed
that the Shift Foreman's
log following the April ll, 1996, testing did contain
documentation
pertaining to the
HSSV as-found out-of-tolerance setpoints.
The inspectors
observed that the licensee
had not developed
an action plan
and
had not- documented
any guidelines or procedures
to be used
once the data from
the augmented testing
was obtained.
Interviews with licensee
personnel
revealed that
some of the delays
in testing after April 2 occurred
because
they were unsure
about which other
HSSVs should
be tested
and
how the as-found
test results
should
be evaluated
(e.g., there
was uncertainty about what would
be done with the as-found test results for the
HSSVs for which AVK CFs
had not
yet been developed).
The inspectors
concluded that two weaknesses
in the licensee's
HSSV augmented
testing program contributed to the problems
encountered
during the
implementation of the testing.
The licensee did not sufficiently formalize
the
HSSV augmented testing
program prior to its implementation,
and the
licensee
did not sufficiently plan for the evaluation
process
and actions to
be taken following the
augmented testing.
ATTACHNENT 1
1
PERSONS
CONTACTED
1. 1
Licensee
Personnel
- S. Allen, ES/EOP,
Engineering
Services
- J. Alviso, Regulatory Services,
NRC Engineering Assistant
- ¹D. Brosnan,
Acting Director, Regulatory Services
- W. Coley, Engineer,
Regulatory Services
- W. Crockett,
Manager,
Nuclear guality Services
- W. Fujimoto, Vice President,
Operations
and Plant Manager
¹S.
Furnis-Lawrence,
Engineer,
Nuclear guality
¹T. Grebel, Director, Licensing
and Design Basis
- ¹C. Groff, Director, Engineering
Services
- ¹C. Harbor, Regulatory Services,
NRC Interface
- C. Joyce,
Engineer,
Nuclear Performance
Monitoring
- D. Hiklush,
Manager,
Engineering Services
- D. Taggart, Director, Nuclear Performance
Honitoring/NSE
- B. Waltos, Director, Engineering
Services
1.2
NRC Personnel
- S. Boynton, Resident
Inspector
- ¹D. Corporandy,
Reactor
Engineer
- ¹P. Goldberg,
Reactor
Inspector
.¹H. Tschiltz, Senior Resident
Inspector
¹H.
Wong, Chief, Reactor Project
Branch
E
- Denotes those attending
the initial exit meeting
on Hay 17,
1996.
¹Denotes
those attending the telephone exit meeting
on Hay 28,
1996.
In addition to the personnel
listed above,
the inspectors
contacted
other
personnel
during this inspection.
2
EXIT MEETING
A preliminary exit meeting
was conducted
on Hay 17,
1996,
and
a final exit
meeting
was conducted
on Hay 28,
1996.
During these
meetings,
the inspectors
reviewed the scope
and findings of the report.
The licensee
acknowledged
the
inspection findings documented
in this report.
The licensee
did not i,dentify
as proprietary
any information provided to, or reviewed
by, the inspectors.
ATTACHNENT 2
AFM
HSSV
POA
TS
ACRONYHS
correlation factor
Diablo Canyon
Power Plant
prompt operability assessment
Technical Specifications