ML17265A129
| ML17265A129 | |
| Person / Time | |
|---|---|
| Site: | Ginna  |
| Issue date: | 12/24/1997 |
| From: | NRC OFFICE OF INSPECTION & ENFORCEMENT (IE REGION I) |
| To: | |
| Shared Package | |
| ML17264B152 | List: |
| References | |
| 50-244-97-13, GL-89-10, NUDOCS 9801050169 | |
| Download: ML17265A129 (20) | |
See also: IR 05000244/1997013
Text
U.S. NUCLEAR REGULATORYCOMMISSION
REGION I
License No.
Report No.
50-244/97-1 3
Docket No.
50-244
Licensee:
Rochester Gas and Electric Corporation (RGRE)
Facility Name:
R. E. Ginna Nuclear Power Plant
Location:
1503 Lake Road
Ontario, New York 14519
Inspection Period:
October 27-31 and November 5-7, 1997
Inspectors:
Kenneth Kolaczyk, Systems Engineering Branch, DRS
Doug Dempsey, Systems Engineering Branch, DRS
Lois James, Systems Engineering Branch, DRS
Mark Holbrook, Contractor, INEL
Approved by:
Eugene M. Kelly, Chief
Systems Engineering Branch
Dive~i n of Reactor Safety
980i050i69 97i224
ADOCK 05000244
8
EXECUTIVE SUMMARY
R. E. Ginna Nuclear Power Plant
NRC Inspection Report 50-244/97-13
This special inspection reviewed the status of the Ginna Motor Operated Valve (MOV)
program for the purpose of determining if Rochester Gas and Electric (RGSE) had met their
commitments under Generic Letter (GL) 89-10
~
Substantial improvements were evident.
Self assessments
and independent reviews were
utilized to develop significant enhancements
in MOV design and testing.
Program
documents and procedures were rewritten, test data reexamined, revised assumptions
developed and new diagnostic test equipment procured.
The quality of design calculations
was generally good, and degraded voltage and weak link analyses were redone.
However,
the NRC was unable to reach closure regarding the GL 89-10 baseline program because of
the following:
~
Calculations (performed by a vendor) had not been finalized and accepted under the
Ginna Station Quality Assurance
(QA) program and had not received formal RGRE
review and approval. The failure to approve the vendor's calculations is a poor
engineering practice with respect to configuration control and was a violation of 10 CFR 50 Appendix B Criterion Vll, "Control of Purchased
Material Equipment and
Services" (VIO 97-13-02).
~
Input assumptions were used in several instances without adequate validation,
resulting in incorrect design calculations, which under-estimated
the thrust
requirements for three MOVs. The failure to adequately validate MOV design inputs
was a violation of 10 CFR 50, Appendix B, Criterion III, "Design Control"
(VIO 97-13-03).
The acceptability of administrative controls which govern the use of a programmable
database,
SMARTBOOK, for safety related calculational input was left unresolved
(UNR 97-13-01).
TABLEOF CONTENTS
EXECUTIVE SUMMARY
PAGE
~
~
~
~
~
~
~
II
III. Engineering
E1
Motor-Operated Valve Program Review (Tl 2515/109)
E1.1
MOV Sizing and Switch Settings .........
~
E1.2
Grouping Criteria
E1.3
Valve Factor Selection ............
E1 4
PORV Block Valves
E1.5
Reactor Coolant Pump Seal Water Return Valve
E1.6
Tracking and Trending
E8
Miscellaneous Engineering Issues....... ~....
E8.1
(Closed) Follow-up Item 50-244/96-08-03.
E8.2
(Closed) Follow-up Item 50-244/95-06-02.
E8.3
(Closed) VIO 50-244/96-08-01:...
~ ~,
.
E8.4
(Closed) VIO 50-244/96-08-02 ~....
~ .,
E8.5
(Closed) Follow-up Item 50-244/95-06-06
~
E8.7
(Closed) Follow-up Item 50-244/95-06-07
E8.8
(Closed) Unresolved Item 50-244/95-06-09
E9
Review of Updated Final Safety Analysis Report (UFSAR)
V. Management Meetings
X1
Exit Meeting Summary ...
.
1
~
~
~
~
~
~
~
1
.
1.5
~
~
~ 6.7
~
~
~ 9
~
~
~
~
1 1
. 11
~
~
~
~ 12
~
~
~
~
~
1 2
~ . 13
. 13
~
~
~
~
~ 13
. 14
~
~
~
'
~
~
~
1 4
... 14
... 14
Re ort Details
III. En ineerin
E1
Motor-Operated Valve Program Review (Tl 2515/109)
Backcaround
On June 28, 1989, the NRC issued Generic Letter (GL) 89-10, "Safety-
Related Motor-Operated Valve Testing and Surveillance," which requested
licensees to establish
a program to ensure that switch settings for safety-
related motor-operated valves (MOVs) were selected,
set, and maintained
properly.
NRC inspections at Ginna have been conducted based on guidance
contained in NRC Temporary Instruction (Tl) 2515/109, "Inspection
Requirements for Generic Letter 89-10."
During an inspection performed in July 1996, the NRC determined the MOV
program could not be closed, principally because
unfounded engineering
assumptions were used in calculations for the Residual Heat Removal (RHR) core
deluge valves, and because the tracking and trending program was not being
effectively used.
The inadequate verification of design assumptions for the core
deluge valves was discussed
during a November 1996 enforcement conference.
RGKE subsequently committed to a number of corrective actions and program
enhancements.
These commitments were described in a January 13, 1997, letter
to the NRC.
The purpose of this current inspection was to examine the actions implemented by
RGSE to address the problems identified during the July 1996 inspection and to
determine if the related actions were sufficient to warrant "closure" of the NRC
staff's review of the Ginna GL 89-10 MOV program.
E1.1
MOV Sizin
and Switch Settin
s
a.
Ins ection Sco
e
The inspectors reviewed the test results and engineering evaluations for the seven
MOVs listed below. The review consisted of examining calculational inputs
associated
with: (1) valve factor, which correlates differential pressure to the stem-
thrust requirement; (2) stem friction coefficient, which affects the conversion of
actuator output torque to valve-stem thrust; and (3) rate of loading or load sensitive
behavior, which reflects the change (usually a loss) in deliverable stem thrust under
dynamic conditions as compared with the available thrust measured under static
conditions.
RCS-MOV-313
RCS-MOV-51 5/51 6
CCW-MOV-81 3
CS-MOV-860A
MS-MOV-3504A
SW-MOV-4664
Reactor Coolant Pump Seal Water Return Isolation
Pressurizer Relief Stop (Block) Valves
Component Cooling Water Supply
Train 1A Containment Spray Pump Discharge
Turbine Driven Auxiliary Feedwater Steam Admission
Turbine Building Service Water Isolation 1A2
The inspectors also reviewed MOV program documents,
Engineering Work Request
(EWR) 5111, "The Motor-Operated Valve Qualification Program Plan" and
Engineering Work Request (EWR) 5080, "Design Analysis Ginna Station GL 89-10
MOVs."
Observations
and Findin s
General
Following the July 1996 inspection and subsequent
enforcement conference,
program documents and procedures were rewritten, test data reexamined, and
revised assumptions
and calculations developed.
Based on recent industry
information contained in NRC Information Notice (IN) 96-48, "Motor-Operated Valve
Performance Issues," and NUREG/CR-6478, "Motor-Operated Valve (MOV) Actuator
Motor and Gearbox Testing," RGSE also revised their methods for calculating motor
actuator capability in the closing direction by discontinuing the Use of run
efficiencies and replacing them with more conservative pullout efficiencies.
This
action required the adjustment of several torque switch settings during the most
recent refueling outage.
New dynamic test methods to require the use of on-line
pressure transmitters (where possible) were implemented.
This improved method
provided a differential pressure
(DP) measurement
over the complete valve stroke
and increased the accuracy and confidence level of MOV performance
determinations.
However, RG&E did not adequately assess
MOV design guidance or control design
information in all cases.
Specifically, input assumptions
were used for certain
double disc gate valves in unapproved applications without adequate justification,
and errors were detected
in several design calculations.
Thrust calculations that
were performed using the Electric Power Research Institute's MOV Performance
Prediction Program (PPP) had not been formally approved and accepted by RGRE,
as required by the Ginna Quality Assurance
(QA) manual.
These issues are
discussed
in detail in Sections E1.3, 4 and 5.
Process Controls and Guidance
Program Plan EWR 5111 is the guidance document for the Ginna MOV program,
describing the assumptions
and the methods used to establish and set torque
switch settings.
Program document EWR 5080, identified the MOVs included in the
GL 89-10 program and described the maximum DP each MOV must overcome to
perform its safety function.
MOV design inputs were consolidated in a computerized (Microsoft Access)
database
program called the "SMARTBOOK." The program used design inputs and
valve specifications to calculate required thrust values using the standard industry
equations,
including adjustments for load sensitive behavior and degradation of
actuator capability. These values were reproduced onto data sheets, which RG&E
used to establish the MOV switch settings.
A series of "lower tier" procedures,
including M-64.1.2, "MOVATsTesting of Motor Operated Valves," outlined the
methodology for setting the torque switch settings.
The SMARTBOOKprogram simplified the storage and retrieval of MOV data and the
calculation of thrust values.
However the program was not procured as safety-
related and it was not apparent how RG&E verified that the outputs were correct.
Ginna staff indicated they performed hand calculations to verify the correctness of
the program outputs independently.
However, none of these verifications or
independent calculations were documented,
nor was it evident how the information
in the program was controlled.
Finally, aside from a description of the program in
EWR 5111, there was no formalized procedural guidance for how the program
should be used or how its associated
calculations should be checked and verified.
Load Sensitive Behavior
RG&E used a 15% margin in thrust calculations to accommodate the effects of load
sensitive behavior.
The margin was applied as a "bias" value and was based upon
a statistical analysis of dynamic tests performed on eleven Ginna globe and gate
valves of various sizes and pressure classes.
The statistical analysis indicated the
15% margin was appropriate for the valve population at a 97% "confidence level."
The inspectors reviewed the load sensitive behavior study and statistical analysis
and determined the 15% margin bounded the majority of the test data.
Including
the margin as a bias value resulted in an additional conservatism when compared
with alternative methods used in the industry.
Stem Friction Coefficient
RG&E assumed
a value of 0.20 for a stem friction coefficient. This value was
based upon a statistical analysis of Ginna gate and globe valve dynamic and static
test data.
The inspectors reviewed the test data and noted only one valve, MOV
738A, had a stem friction coefficient that exceeded
the assumed
value under
dynamic conditions and that difference was not significant.
Further, there was
limited variation in the data.
Therefore, the inspectors determined the stem friction
coefficient assumption was appropriate.
Control of Purchased
Services
Following the July 1996 MOV inspection, RG&E used
a contractor to assist in the
enhancements
to the Ginna MOV program.
calculations were developed using the SMARTBOOKprogram.
RG&E used the
revised calculations as the basis for establishing MOV operability.
Although the quality of the revised engineering calculations and program documents
was good, RGSE did not followthe requirements of the Ginna Quality Assurance
(QA) program and engineering procedures before using the vendor supplied
engineering calculations to establish MOV design basis capability.
Specifically,
Engineering Procedure
(EP) 3-P-154 "Review and Approval of Vendor Drawings,
Design and Manufacturing Technical Documents," states that vendor-supplied
calculations shall not be used to establish
a basis for operability of safety-related
equipment unless formal, final approval of the calculations has been obtained in
accordance with EP-3-P-154.
The approval shall be documented using a
memorandum incorporated into each vendor calculation.
However, as of the time of
this inspection, RGRE had not completed its formal final approval of the calculations
of record for all Ginna GL 89-10 MOVs.
Conclusions
Assumptions for load sensitive behavior and stem friction coefficient were
technically acceptable.
RGSE used the outputs from the SMARTBOOKprogram to change or validate the
switch settings of safety-related
MOVs. However, RG5E had not formally
controlled the data contained in the program or validated the program outputs in a
documented manner.
This issue was partially addressed
as part of a
recommendation from a QA audit of the GL 89-10 program conducted in
August 1997. This would otherwise be acceptable
provided that each MOV
calculation receives effective independent verification and approval.
The
acceptability of the current administrative requirements for use of the program was
unresolved at the end of the inspection.
(UNR 97-13-01)
The inspectors considered the use of unapproved calculations to be a poor
engineering practice and a potential problem for configuration control, particularly
since the torque switch settings for a number of MOVs were adjusted during the
recent outage based on these calculations.
10 CFR Part 50, Appendix B, Criterion
Vll, "Control of Purchased
Material, Equipment, and Services," requires, in part, that
"Measures shall be established to assure that purchased
.
~
. services conform to
Documentary evidence that material conforms to the
procurement requirements shall be available at the nuclear power plant site... prior
to the use of such material.
This documentary evidence shall be retained at the
nuclear power plant site... and shall be sufficient to identify the specific
requirements,
such as codes, standards,
or specifications, met by the purchased
material." The failure to review the vendor supplied calculations and document they
conformed to the procurement requirements before using them as a basis to
establish MOV operability was a violation (VIO 97-13-02) of 10 CFR 50 Appendix B
Criterion Vll, "Control of Purchased
Material, Equipment and Services."
Grou in
Criteria
Ins ection Sco
e
There are 60 valves in the Ginna GL 89-10 program population.
RGRE used the
standard
industry (Limitorque) equation to establish MOV switch settings.
The
switch settings of thirty-one of the valves were verified through dynamic tests.
The
remaining valves that could not be dynamically tested were placed in one of
fourteen valve groups based on valve type (e.g, flex-wedge, double disc, globe
etc.).
If a sufficient number of valves were available using the first criteria, then
valve size, type, and ANSI pressure class rating were used to subdivide the valve
groups further.
Observations
and Findin s
RGSE applied the highest valve factor obtained through dynamic tests to the
nondynamically tested valves in the respective group.
However, using the highest
valve factor (from a test), and inferring that factor would represent the performance
of the remaining valves in the grouped population, may not be conservative without
further analysis.
Specifically, for groups in which the data are scattered,
applying
the highest valve factor may not encompass the statistical performance variation of
the valves.
This issue was discussed
in NRC Information Notice (IN) 97-07,
"Problems Identified During Generic Letter 89-10 Closeout Inspections," dated
March 6, 1997, which noted that "..
~ some licensees
have selected
a valve factor
based on a sample of tests that does not accommodate
reasonable
variation in the
valve factor for other MOVs in the group."
RGRE's grouping criteria contained valves with wide ranges of sizes, pressure
classes,
and manufacturers.
The inspectors questioned the rationale for two
groups: (1) Group C contained seven Crane flex-wedge gate valves and two. Borg
Warner flex-wedge gate valves, and (2) Group K contained seven butterfly valves
from two different valve manufacturers.
To assess
the significance of the valve
factor variation, and in response to the inspectors'uestions,
RG&E reassessed
the
valve factors using a mean plus two standard deviation statistical method for each
group.
The results indicated the assumed valve factor bounded the majority of the
test data.
Conclusions
In groups B, D, and E, the assumed valve factor did not bound the upper limitof the
data scatter.
Although only a few valves were affected and the inspectors did not
identify any operability concerns, the bounding nature of this assumption needs to
be re-evaluated by RGS.E (IFI 97-13-04) during the periodic verification program
under their GL 96-05 commitments.
E1.3
Valve Factor Selection
a.
Ins ection Sco
e
The inspectors evaluated valve factors used in groups that had no available or
meaningful in-plant test results.
RGRE applied valve factors that were based on
friction coefficients obtained from the Electric Power Research Institute's (EPRI's)
"separate effects "friction testing performed in support of the MOV Performance
Prediction Program.
b.
Observations
and Findin s
Friction coefficients were used in the standard industry equation to develop a
predicted thrust value for 21 separate
valves in groups A, C, H, J, M, and N.
However, this method was technically incorrect in some instances,
since the EPRI
friction factors were obtained under controlled conditions that do not account for
changes
in the performance of installed valves, and which could reasonably
be
expected.
Specifically, unlike actual valve factors obtained from in-plant tests, the
friction coefficients do not account for possible valve guide wear or bending, disc
tipping, the force required to wedge certain valves or other valve-specific
performance characteristics.
NRC technical concerns regarding selective use of the EPRI data were outlined on
page 2 of the February 20, 1997, safety evaluation report (SER) supplement that
approved use of the EPRI program.
In the SER, the NRC indicated:
"The NRC staff has reviewed the EPRI methodology as a complete package
in that certain nonconservative
assumptions
in the models are compensated
by other conservative assumptions
in the analytical formulas.
Selective use
of test data or methods from the EPRI program may result in underpredicting
the thrust or torque required to operate gate, globe or butterfly valves."
Based on a review of Ginna program documents and selected thrust calculations,
the inspectors found th=-,: ."IGRE did not address the NRC comments in the SER
regarding selective use of coefficients from the EPRI program.
C.
Conclusions
Although data from the EPRI program were used in the above instances
in a
technically incorrect manner, this application did not appreciably impact
functionality since most valves appeared to have adequate
design margin.
The use
of EPRI coefficients in an approved methodology for the 21 MOVs associated with
the groups discussed
above willneed to be re-evaluated
by the NRC (IFI 97-13-05)
for purposes of GL 89-10 program closure.
Notable exceptions with respect to
margin included the pressurizer power operated relief (PORV) block valves, and the
reactor coolant seal water return valve, which are discussed
in the following
sections.
PORV Block Valves
Ins ection Sco
e
The inspectors reviewed Altran Calculation No. 96190-C-28, dated June 1997 for
the PORV block valves.
This calculation established the minimum required design
basis thrust and torque limits for purposes of diagnostic setup and testing of the
valves.
The calculation referenced
use of the EPRI Topical Report (TR)-103232,
dated November 1994, associated with use of a 0.42 valve factor assumption.
The
inspectors also evaluated the design application of these valves, particularly with
respect to leak tightness, as described below.
Observations
and Findin s
~Back round
The PORV block valves, (MOVs 515/516) are three-inch Anchor Darling double disc
gate valves designed to achieve
a leak tight seal when the upstream and
downstream parallel discs are "hard seated" against their respective seat.
This
design is intended to force the discs apart by the sliding action of angled upper and
lower disc wedges as the disc assembly strikes the bottom of the valve. This
condition is called "wedging" and typically requires greater force than just achieving
full seat overlap (also referred to as flow isolation or blockage) for this type of
valve. The amount of flow through the valve at the full seat overlap position will
depend on the DP loading, the disc area, and the contact stress.
Significant
additional thrust can be required to spread the disc wedges sufficiently to meet
leakage limitations and can be affected by the orientation of the valve discs.
Specifically, valves with the upper wedge located downstream of flow (non-
preferred direction) can require much more thrust to achieve a hard seat wedged
disc position.
Past calculations assumed
a 0.5 valve factor, but the technical basis for this value
was questioned during the July 1996 inspection.
Since the block valves could not
receive a dynamic test, the thrust requirements for these valves are currently based
on a 0.42 disc friction coefficient obtained from the so-called EPRI "separate
effects" testing.
The PORV block valve thrust is controlled in the close direction by
a torque switch. The switch settings were established
based upon Ginna design
document EWR 5080, which indicated the block valves had two safety-related
functions; isolate a stuck open PORV, and open to depressurize the reactor coolant
system.
Chapter 5.4.10 of the Ginna Updated Final Safety Analysis Report
(UFSAR) indicated the design leakage
limitfor the block valves was within the
capacity of one charging pump (approximately 60 gpm).
Because of the relative small design margin, and the potential that the lab-developed
EPRI friction coefficient (OA2) would not approximate actual valve performance
characteristics,
the inspectors requested
RGSE to reevaluate thrust margins using
the more appropriate NRC-approved
EPRI "hand calculation" method.
Two values
were calculated:
(a) thrust required to achieve full seat overlap and (b) thrust
required to achieve wedging.
Since the licensee was unaware of the orientation of
the block valve discs, photograph X-rays were taken of both valves on October 31,
1997.
The photographs revealed valve 516 was oriented in the non-preferred
direction.
Prior to this inspection, the block valves were set to approximately 11,000 pounds
(Ibs) force.
By the inspector's independent calculations, this was above the EPRI-
predicted required thrust for flow isolation (10, 200 Ibs) but less than what would
be predicted to achieve hard seat contact or full wedging.
While there are no
specific design basis leakage criteria for the PORV block valves, other than being
less than the charging pump's capacity, a setup to only achieve the flow blockage
condition does not take full advantage of the wedging design for these particular
valves.
The Ginna Technical Specification Bases 3.4.11 define the licensing design
basis for the PORV block valve closure as "...terminating the RCS depressurization
and coolant inventory. As stated in the NRC's SER for the EPRI PPM, the
inspectors did not consider torque switch setup for thrusts predicted at or near flow
isolation to be sufficiently justified with respect to design function. The term
isolation is more aptly described
as "blockage," but does not necessarily ensure any
measure of leak tightness.
loss" associated with a stuck-open PORV which is, in
effect, a small break LOCA.
In response to the inspectors'oncerns,
the torque switch setting for these valves
was changed to increase thrust output to 13,843 and 14,652 Ibs respectively,
which was equivalent to better than 80% wedging.
The analytical work to support
the thrust increase was outlined in calculations 19703245 and 19703246
performed on November 3, 1997. Additional increases
in thrust output were
precluded by the weak link structural limits. The inspectors independently
calculated
a predicted thrust, using the program value for rate of loading effects,
and determined the torque switch setting for the block valves was adequate.
However, the inspectors noted the revised calculations contained an error in that
RGSE failed to provide an allowance for rate of loading when determining MOV
target thrust.
Conclusions
Although the past (prior to November 1997) switch settings may not have achieved
complete wedging (and leak tightness) under all differential pressures,
the
inspectors concluded partial wedging and flow blockage would likely have occurred
under design conditions,
RGRE was unaware that the thrust required to achieve
wedging was dependent
on disc orientation.
Further, the valves may not prevent
flow under all differential pressure conditions when the torque switch settings are
set to achieve only full seat overlap.
The current PORV block valve switch settings were established
on
November 3, 1997, in part, based upon incorrectly performed calculations that did
not adequately consider a value for rate of loading.
Criterion III, "Design Control," requires, in part, that "~ .. measures
shall be
established to assure that applicable regulatory requirements
and design basis for
structures, systems,
and components
are correctly translated into specifications,
drawings, and procedures."
The failure to apply a correct value for rate of loading
in the November 3, 1997, calculations and the use of incorrect equations to
calculate the required thrust for the PORV block valves was the first example of a
violation (VIO 97-13-03) of 10 CFR 50, Appendix B, Criterion III, "Design Control ~"
E1.5
Reactor Coolant Pum
Seal Water Return Valve
a.
Ins ection Sco
e
The inspectors reviewed the thrust calculations and design requirements for the
reactor coolant pump seal water return valve, (MOV 313),
a 3-inch, Aloyco split-
wedge containment isolation valve.
The disk assembly consists of two separable
disk halves connected by a ball and socket joint. The wedging action is similar to
that of the Anchor-Darling double disk gate valve in that significantly more thrust is
required to wedge the valve disks than to achieve primary flow blockage.
Similar to
the Anchor Darling valves, thrust can be affected by orientation of the valve discs in
preferred or non preferred directions.
Also valve leak tightness may not be assured
under all DP conditions unless the valve discs are wedged closed.
b.
Observations and Findin s
RGRE had previously dynamically tested MOV 313 at 62% of design basis DP, but
was unable to derive a valve factor from the test data.
Accordingly, a thrust was
calculated using a 0.65 valve factor obtained from a friction coefficient (EPRI
separate effects testing) corrected for wedge angle.
The valve factor was applied in
the standard industry equation, in Altran Calculation 96190-C-84approved
for
release
(by the vendor) on October 10, 1997. The minimum required closing thrust,
including load sensitive behavior, was recorded to be 2,048 lbs. The current switch
setting (based upon this aforementioned calculation of record) is 2,400 lbs. The
inspectors considered the predicted thrust requirement to be low and incorrect,
without proper consideration for leak tightness, and inconsistent with the EPRI
methodology appropriate for the assumed valve factor.
10
Because of the general NRC concern that using the EPRI friction coefficients in the
standard industry equation would not accurately predict the required thrust, the
inspectors questioned the thrust requirements for MOV 313.
Since the valve's
wedge orientation was unknown, the predicted thrust values were independently
calculated by the NRC assuming the valve discs were in the nonpreferred direction.
Using the appropriate
EPRI equations associated with the 0.65 valve factor, and
including load sensitive behavior, the current design setup for MOV 313 was
determined to be as follows:
Flow blockage:
2140 Ibs
Current Switch Setting
2400 Ibs
Wedging
6603 Ibs
Although these thrust numbers indicate that the valve would achieve flow blockage
with the current switch setting, the valve would apparently not achieve complete
wedging.
Therefore, absent further analysis, leak tightness (the design function of
the valve) may not be assured.
It should be noted that thrust requirements using
EPRI methods for the wedged condition exceed by a factor of three those for "flow
isolation."
Use of the standard industry equation is essentially equivalent to the
EPRI hand calculation for flow isolation, with the difference of a "torque reaction
factor" that amounts to approximately 4-5% higher thrust.
Using the standard
equation would, therefore, significantly under-estimate the thrust for a containment
isolation valve.
The need to ensure that sufficient wedging is achieved to meet
valve-specific leak tightness requirements was outlined in the February 20, 1997,
safety evaluation report (SER) supplement which evaluated the EPRI test data for
Aloyco valves.
In the SER supplement, the NRC indicated "...model users will also
need to justify that leakage limits are satisfied for the specific valves where only
flow blockage is achieved."
Based upon an independent review of the past dynamic test results for MOV 313,
and discussions with Ginna staff, the inspectors found that the "loads" under
dynamic conditions were dominated by packing.
Although no appreciable dynamic
effects were evident (the test was done at approximately 90 psid) to derive a valve
factor, the successful stroke under test coupled with low delta-P conditions and
high margin form a qualitative case for functionality. Therefore, notwithstanding
the incorrect design basis calculations, operability was not in question.
Although
the predicted thrust for a fullywedged condition would exceed'weak
link structural
limits, the existing setup has successfully passed Appendix J containment leak rate
testing.
While the inspectors did not verify past leak rate test results for MOV 313,
the above arguments were considered sufficient basis for operability; pending either
correct use of the EPRI methodology or better technical justification for use of the
0.65 valve factor. As of the end of this inspection, the licensee disagreed with the
inspectors'echnical
conclusions and proposed violation of Appendix B, Criterion III.
RGRE staff consider the Altran calculations in question to be correct and adequate.
11
C.
Conclusions
Since MOV 313 would only develop enough thrust to achieve primary flow
blockage, additional design consideration and analysis should have been performed
to ensure the valve would perform its containment isolation function under design
basis leak tightness conditions.
Incorrect equations were used in Altran Calculation
96190-C-84to predict the minimum required thrust, resulting in an underestimate of
thrust requirements and an incorrect design.
10 CFR 50, Appendix B, Criterion III, "Design Control," requires, in part, that "...
measures
shall be established to assure that applicable regulatory requirements and
design basis for structures, systems,
and components
are correctly translated into
specifications, drawings, and procedures."
The failure to perform the required
analysis to ensure MOV 313 would remain leak tight under design conditions was
the second example of a violation (VIO 97-13-03) of 10 CFR 50, Appendix B,
Criterion III,'Design Control."
E1.6
Trackin
and Trendin
aO
Ins ection Sco
e
The inspectors reviewed Attachment J, "Assessment
and Feedback Criteria," of
EWR 5111 which contained the guidance for the MOV tracking and trending
program.
Items to be monitored included MOV running load, stem friction
coefficient, motor current and voltage and torque switch settings.
Some of this
data had been placed into the SMARTBOOKcomputerized database
for analysis.
b.
Observations
and Conclusions
The scope of the trending program appeared to be adequate to detect meaningful
trends in valve performance.
At the time of the inspection. RGSE was still in the
process of developing
a procedure to implement the tracking and trending program.
As such, the program had not been fullydeveloped.
The development of a tracking
and trending procedure was scheduled to be completed by January 31, 1998. The
inspector determined that date was appropriate for program closure.
E8
IVliscellaneous Engineering Issues
E8.1
Closed
Follow-u
Item 50-244 96-08-03:pressure
isolation valve concerns.
This
item was opened to track NRC analysis of the acceptability of the RHR core deluge
piping configuration.
The RHR core deluge piping consists of two parallel six-inch
lines that branch off the RHR cold leg return piping in the containment structure and
connect into the reactor vessel head.
Each deluge line contains one normally closed
motor-operated deluge valve, MOV-852A/B, and an associated
swing check valve,
853A/B located between the reactor vessel head and the respective MOV. When a
~ safety injection actuator signal occurs, the core deluge valves are designed to open
and allow RHR system water to discharge through swing check valves and into the
reactor vessel.
The design pressure of the piping between the MOV and the reactor
12
vessel is 2485 psi; the design pressure of the upstream piping is 600 psi. The
inspectors were concerned if the swing check valve failed, the possibility of an inner
system loss of coolant accident (ISLOCA) event could exist if the deluge valves
were opened when reactor coolant system pressure was greater than the 600 psi
piping.
Independent examination of the RHR deluge piping configuration conducted by the
Office of Nuclear Reactor Regulation (NRR) concluded the current piping
configuration was adequate.
This conclusion was based,
in part, on the fact that
the upstream check valves 853A/B receive periodic leak checks to verify their
integrity. The inspectors reviewed the results of recent leakage test performed on
the check valves and verified the leakage test results for the past five years were
consistently below the Technical Specification leakage limit of 5 gpm.
Based upon
the minimal check valve leakage, and the determination reached by NRR that the
current configuration was acceptable, this item is closed.
Closed
Follow-u
Item 50-244 95-06-02: margin justification for valve factor.
This item was opened to identify the fact that RGRE would need to justify certain
valve factor assumptions.
Although the overall quality of the MOV program was
improved, it was not evident that the design verification effort for the GL 89-10
program was adequately performed in all instances.
Specifically, as discussed
in
sections E.1.3 - E.1
~ 5 of this report, valve factors for the PORV block and reactor
coolant seal water return valve MOV 313 were inadequately verified. Therefore,
this item is closed and will be tracked as part of VIO 97-13-02 and 03.
Closed
VIO 50-244 96-08-01:design control. This violation concerned
RGKE's
failure to establish adequate
design control measures
in the GL 89-10 program.
As
a result, RGSE did not adequately assure the RHR core deluge valves would operate
under design conditions.
Corrective action included: contracting with a vendor to
reanalyze existing MOV program data and verifying the remaining valves were
operable, revising the design control process, reestablishing participation in industry
forums, and instituting third party reviews of the MOV program.
As discussed
in Section E1.1, MOV program documents appeared
up-to-date and
reflect the most recent industry practices.
The revised design control process
as
outlined in Engineering Procedure
(EP) 3-S-125, "Design Verification and Technical
Review" appeared to improve the rigor of the design review process by instituting a
design review checklist and clarifying the responsibilities of the independent
reviewer.
Additional oversight of the MOV program was evident.
Several
independent reviews of the program had been completed and the Quality Assurance
department was developing a schedule to periodically audit the MOV program.
The
corrective action from this item was therefore considered to be adequate,
and the
item is closed.
Specific issues associated with design control will be followed under
VIO 97-13-02 and 03.
13
Closed
VIO 50-244 96-08-02:ineffective corrective action.
This violation
concerned RG&E's failure to ensure the RHR core deluge valves had adequate
capability to operate under design basis conditions despite evidence the valves had
limited design margin.
The marginal performance of the valves had been previously
noted in a March -April 1995 NRC MOV inspection report and during testing of the
valves during the spring 1996 refuel outage.
Immediate actions consisted of
shutting down the Ginna Station in August 1996 and installing larger actuators on
MOVs 852A/B. Additional action included revising the Commitment and Action
Tracking System (CATS) to require the explicit identification and resolution of each
action item contained in NRC inspection reports and other docketed
correspondence.
Previously, one CATS item was assigned to an inspection report.
The revised CATS process appeared to adequately track commitments,
observations,
and issues identified in NRC correspondence.
The inspectors
determined RG&E has implemented adequate corrective actions.
Therefore, this
issue is closed.
Closed
Follow-u
Item 50-244 95-06-06: periodic verification of design basis
assumptions.
This item was opened to track development of the periodic
verification program.
Although the periodic verification program had not been
completed, Ginna MOV engineers indicated the program would generally follow the
testing guidance developed by the joint owner's group.
RG&E did not include a
specific margin to account for valve degradation.
However, Attachment G, "Switch
Setting and Static Testing Criteria," of EWR 5080, did specify that it is desirable to
set the torque switch to maintain a minimum 10% margin above the target thrust to
account for the effects of stem lubrication degradation.
If this could not be
achieved, then the attachment required that the stem lubrication frequency be
increased.
This approach appeared to be reasonable.
To ensure MOVs are tested in their as-found condition when necessary
during the
periodic verification program, testing and preventive maintenance activities are
currently coordinated through informal, (i.e., word-of-mouth) communications.
Specifically, when as-found data on a valve is required, the MOV engineer notifies
the MOV maintenance
engineer not to overhaul the valve before testing is
completed.
This method appeared to be working appropriately, and revised
procedures were under consideration to require maintenance
personnel to contact
the MOV engineer before working on a valve.
The inspectors concluded the
planned periodic verification and MOV testing programs were appropriate.
Accordingly this item is closed.
Final acceptance
of the Ginna periodic verification
program will be reviewed as part of RG&E's response to GL 96-05, "Periodic
Verification of Design -Basis Capability of Safety-Related Motor-Operated Valves."
Closed
Follow-u
Item 50-244 95-06-07: post-maintenance
testing.
This item
was opened to document RG&E did not currently perform a valve thrust verification
test (either static or dynamic), following packing adjustment, if gland nut torque
remains below the original diagnostic baseline value.
The inspectors considered this
position to be technically indeterminate
as well as contrary to industry practice
since packing adjustment may effect valve performance.
Following the August
1996 MOV inspection, RG&E stopped adjusting valve packing if a subsequent thrust
verification would not be performed.
A study was then commenced to determine
what amount of packing adjustment, if any, could be allowed before a thrust
verification was necessary.
The study included an analysis of Ginna valve
performance characteristics
before and after packing gland adjustment.
Although
the study had not been completed at the time of the inspection, RG&E believes the
results will indicate a thrust verification is not required, if gland nut torque remains
below the original diagnostic test value.
The inspector concluded RG&E's approach
was adequate to address this issue.
Therefore, this item is closed.
E8.8
Closed
Unresolved Item 50-244 95-06-09:Pressure
Locking/Thermal Binding
(PLTB) of Gate Valves.
This item was opened to track the status of RG&E's
corrective actions for gate valves determined to be susceptible to pressure locking.
In a letter dated February 16, 1996, RG&E described the process used to evaluate
valves for susceptibility to PLTB, and the results of the evaluation.
Although
several valves met the initial screening criteria for susceptibility to PLTB, valve
operability was confirmed in a subsequent
RG&E reanalysis completed in
November 1996. The actions taken by RG&E to address
PLTB at Ginna are
currently under review by the Office of Nuclear Reactor Regulation (NRR) ~
Therefore, this item is closed.
E9
Review of Updated Final Safety Analysis Report (UFSAR)
The inspectors verified that the PORV design criteria described in Section 5.4 of the
Ginna UFSAR were consistent with design assumptions
and calculational results
used in the GL 89-10 program.
V. Mana ement IVleetln s
X1
Exit Meeting Summary
RG&E was informed of the scope and purpose of this inspection at an entrance meeting on
October 27, 1997.
The findings were discussed with RG&E representatives
during the
inspection, and were formally presented
during two meetings on October 31 and
November 7, 1997 at the Gir<<ia site.
A final exit was held by telephone on
November 18, 1997.
RG&E disagreed with the characterization of the design control
findings identified during this inspection.
15
PARTIALLIST OF PERSONS CONTACTED
Rochester Gas 5, Electric
R. Mecredy
B. Flynn
F. Maciuska
R. Marchionda
M. Farnan
J. Smith
J. Widay
G. Wrobel
T. Marlow
K. Muller
T. Alexander
D. Kuhn
M. Lilley
M. Zweille
Vice President, Nuclear Operations
Primary Systems Engineering Manager
Operations Training Manager
Production Superintendent
Equipment Diagnostic Coordinator
Maintenance Superintendent
Plant Manager
Nuclear Safety 5 Licensing Manager
Dept Manager NES
Mechanical Engineer
Nuclear Assurance Manager
Quality Assurance Analyst
Quality Assurance Manager
Senior Engineer
NRC
P. Drysdale, Senior Resident Inspector
C. Osterholtz, Resident Inspector
INSPECTION PROCEDURES USED
Tl 2515/109 (Part 3), Inspection Requirements for Generic Letter 89-10 "Safety-Related
Motor-Operated Valve Testing and Surveillance"
LIST OF ITEMS OPENED, CLOSED, AND DISCUSSED
Closed
50-244/95-06-02
50-244/95-06-06
50-244/95-06-07
50-244/95-06-09
50-244/96-08-01
50-244/96-08-02
50-244/96-08-03
Margin justification for valve factor
Periodic verification
IFI
Post-Maintenance
Testing
Pressure
Locking and Thermal Binding
Failure to verify design inputs for 852A/B
Inadequate corrective action
IFI
Pressure isolation valve concerns
16
~Qened
50-244/97-1 3-01
50-244/97-1 3-02
50-244/97-1 3-03
50-244/97-1 3-04
50-244/97-1 3-05
Design Control for SMARTBOOKProgram
Failure to approve and accept vendor calculations
Inadequate verification of design assumptions for MOV
313 and RCS 515/516
IFI
Grouping Criteria for Groups B,D,E
IFI
Valve factors for Groups A,C,H,J,M,N
LIST OF ACRONYMS USED
DP
GL
INEL
IR
PLTB
RV
RGRE
Sl
TS
Tl
differential pressure
Engineering Procedure
Electric Power Research Institute
Engineering Work Request
Generic Letter
Idaho National Engineering Laboratory
Inspection Report
Loss of Coolant Accident
Motor-Operated Valve
Power Operated Relief block Valves
pressure locking or thermal binding
Quality Assurance
Reactor Safety Study
relief valve
Rochester Gas and Electric
safety injection
technical specifications
Temporary Instruction