ML17174B136
| ML17174B136 | |
| Person / Time | |
|---|---|
| Site: | Dresden  |
| Issue date: | 02/10/1992 |
| From: | Burgess B NRC OFFICE OF INSPECTION & ENFORCEMENT (IE REGION III) |
| To: | |
| Shared Package | |
| ML17174B135 | List: |
| References | |
| 50-237-91-36, NUDOCS 9202180033 | |
| Download: ML17174B136 (8) | |
See also: IR 05000237/1991036
Text
"
U.S. NUCLEAR REGULATORY COMMISSION
REGION III
Report No.
50~237/91036(DRP)
Docket No.
50-237
Licensee:
Commonwealth Edison Company
Facility Name:
Dresden Nuclear Power Station, Unit 2
Inspection A~: Dresden'Site, ~orris, IL
License No.
Inspection Conducted: October 20, 1991, through January 14, 1992
Inspectors:
W. Rogers
M.
Peck
M.
Huber
Approved By:
&~ ~
fut.
B. L. Burgess, Chief,
Projects Section lB
Inspection Summary
Inspection from October 20. 1991 through January 14. 1992 (Report No.
50-237/91036CDRP}.
Areas Inspected: Special safety inspection conducted by resident and regional
inspectors concerning the 2A recirculation pump discharge valve failure to
close on August 7, 1991.
Results: *The low pressure coolant injection system was inoperable for
approximately seven months.
Several weaknesses were identified in the site.
management control system:
Limited root cause evaluation
Inadequate* documentation of qualification requirements for
personnel performing a special process
Use of a non-safety consultant for safety-related work
Inadequate test acceptance criteria
Lack of independence in analyzing VOTES traces
Inaccurate Licensee Event Report information
Several additional weakness were identified in the corporate engineering
management control system concerning:
Failure to perform a safety-related engineering evaluation under
rigorous management control
Failure of erigineering personnel to recognize and document a
9202180033 920210
"* .
ADOCK. 0500.0327 ____ _
O
1.
- 2.
DETAILS
Persons Contacted
Commonwealth Edison Company
- C. Schroeder, Station Manager
- J. Kotowski, Production Superintendent
- D. Van Pelt, Assistant Superintendent - Maintenance
- K. Kociuba, Nuclear Quality Programs Superintendent
W. Kapinus, Motor Operated Valve Coordinator
- D. Pool, Corporate Motor Operated Valve (MOV) Group
- R. Radtke, Regulatory Assurance Supervisor
- W. Morgan, BWR Nuclear Operations Supervisor
- R. Kyrouac, Nuclear Quality Programs Supervisor
- E. Zebus, Project Manager
.
- L. Gerner, Technical Superintendent
- R. Meadows, Maintenance Staff Supervisor
- G. Smith, Assistant Superintendent, Operations
- D. Booth, Master Electrician
- R. Ungeran, Corporate MOY Administrator
- Denotes those attending the exit interview conducted on
January 14, 1992 .
The inspectors also talked with and interviewed several other licensee
employees including members of the technical and engineering staffs.
Background
a.
System Operability Requirements
At Dresden, the low pressure emergency core cooling system is
comprised of the two core spray trains and the low pressure
- coolant injection (LPCI) system. Technical Specification (TS)
3.5.A requires LPCI to be operable during power operation to
ensure adequate heat removal capability in t~e event of a loss of
coolant accident (LOCA).
The TS limiting condition for operation
prohibits continued reactor operation beyond seven days if LPCI is
For LPCI to be operable, the system must be able to determine in
which reactor recirculation (RR) loop the LOCA occurred, isolate
the faulted loop, and inject water into the reactor vessel via the
non-faulted RR loop .. Isolation of a faulted loop occurs when .the
associated RR discharge valve closes.
b.
VOTES Testing Methods
The licensee tests motor operated valves using the Liberty
Technologies Valve Operation Test Evaluation System (VOTES).
2
,,
3.
After recording the VOTES data, maintenance technicians select a
zero reference point from the latter of two transition points in a
closed-to-open-to-closed trace. The transition points occur when
all stem compression and tension is relieved as the stem nut turns
freely through the thread clearance.
The VOTES computer software
then uses data collected by a strain gage, located on the valve
yoke, to calculate valve thrust by referencing the force
coordinate events to the zero refer*ence point. Based upon the
calculated thrusts a torque switch setting is then established.
Reactor Recirculation Valve Failure
In December 1990, during the 12th refueling outage on Unit 2, the
licensee performed VOTES testing on the 2A reactor recirculation loop
discharge valve, 2-202-5A.
The technician was unable to determine the
zero reference point due to anomalies in the VOTES trace.
One of the
technicians sent the digital test data to a recognized VOTES industry
expert, employed with Babcock and Wilcox Nuclear Technologies {BWNT},
and requested guidance in identifying the zero reference point.
On December 14, 1990, the recognized industry expert telephoned the
technician and assisted in identification of the zero reference point.
This zero reference point was entered into the VOTES analysis program
and the valve thrusts were calculated.
Fro~ this calculation, the valve
- operator torque switch setting was reduced from 2.5 to 1.0.
On August 6, 1991, the 2A reactor recirculation pump tripped on over
excitation due to a failed resistor in the motor generator voltage
regulator circuit. After replacing the resistor~ operations personnel
attempted to restart the pump.
In accordance with the procedure, the
operator tried to close the loop discharge valve prior to restarting the
pump.
However, the valve would not close beyond the intermediate
position.
Investigation showed that the valve motor was prematurely
tripping due to an improper torque switch setting. The incorrect torque
switch setting corresponded to a valve operator thrust between 19,000
- and 20,000 pounds.
The required thrust window, based on a design
differential pressure of 200 pounds per square inch, was between 48,280
and 56,517 pounds.
On August 10, 1991, the torque switch was reset to 3.25, based on a
motor current trace, and the valve was declared operable.
On August 20, 1991, as part of the corrective action for the wrong
torque switch setting, the corporate nuclear engineering department
(NED} motor operated valve (MOV} group performed a reanalysis of the 39
remaining safety-related VOTES tests from the D2Rl2 outage.
The
reanalysis concluded that all the other valves were within their design
thrust windows.
3
I,
4.
Event Contributors
a.
Cause of the Event and. Root Cause Evaluation
The licensee determined the cause of the event to be a
misinterpretation of the zero reference point on the force versus
time VOTES trace. The zero reference point identification was
hampered by the unusual characteristics of the trace and the
limitation of the VOTES computer software {Version I.I) to expand
the force axis portion of the trace. The licensee event report
(LER) indicated the newer edition of the software {Version 2.I)
,provided enhanced resolution as well as other features to ensure
proper zeroing interpretations.
The inspector concluded the root cause of the event was a
limitation in the VOTES process.
The determination of the zero
reference point for the discharge valve was consistent with the
VOTES methodology, as described in the formal VOTES training
program and the VOTES manual.
The reactor recirculation discharge
valve yoke exhibited a bending moment when the valve was cycled.
The bending effect resulted in abnormal running loads and a
discontinuity between the first and second transition points of
approximately 50,000 pounds force {lbf) compared to 200 lbf in a
typical trace. The VOTES testing program, including the
methodology, training, and software, was not equipped to properly
evaluate the reactor recirculation discharge valve data, due to
the bending yoke effect. Both versions of the VOTES software
provided expansion capability of the time axis.
The expansion
feature ~llowed for clear identification of both transition points
on the valve trace; however, expansion of the time axis may not
have identified the torsional bending yoke effect to the
uninformed individual. Therefore, the use of version I.I instead
of version 2.I did not contribute to the misidentification of the
zero reference point.
b.
Failure to Document Personnel Qualification
Nuclear Operations Directive NOD-MA.I, "Guidelines For Motor-
Operated Valve {MOV) Testing, Maintenance'and Evaluation,
September I990," delineated the special training requirements for
the Station Motor Operated Valve {MOV) Coordinator, the Corporate
MOV Coordinator, and engineering personnel performing work for the
Corp~rate MOV Coordinator.
The.inspector determined that the licensee had not maintained
records documenting station and corporate personnel qualification
to NOD-MA.I requirements.
The Nuclear Quality Programs group had
previously identified this concern for corporate personnel as
documented in .Audit Report CE-9I-04, November IS, I99I.
The inspector confirmed, through interview, that personnel
performing VOTES analysis were qualified and that the lack of
4
as-left torque switch sittings, or the valve thrust windows.
(2)
The MOV coordinator's concerns about selecting the zero
point on the valve and use of industry expert services were
not documented.
For example, no notations were made in the
work package nor was a condition adverse to quality record
initiated.
(3)
No independent analysis or evaluation was required by the
test procedure. Additionally, no independent verification
of conformance to the thrust windows was required.
The inspector determined through interview that the technicians
performing the testing ~nderstood management's expectations that
the valve torque switch settings would be set consistent with the
design thrust windows.
Therefore, the inadequacies in the test
procedure were not a contributor in the event.
5.
Corporate Engineering Involvement
a.
Reanalysis of the VOTES Traces by NED
The corporate motor operated valve (MOV) group completed a
reanalysis of the 39 safety related VOTES tests performed during
the outage on August 20, 1991.
The MOV Program Administrator
i~sued a letter to the station concluding that all the valves,
except the reactor recirculation discharge valve, were within the
currerit design thrust window.
The inspector determined that five of the re-evaluated valves were
actually outside the design thrust window.
Two of the five valves
had been identified during the outage and had been found
acceptable, as documented in an engineering transmittal (CHRON No.
161468, January 15, 1991, to E. Eenigenburg).
Based upon
engineering judgement, NED MOV personnel determined that the
- operability of the other three valves was not affected by the
degraded torque switch settings. The engineers involved failed to
document either the reanalysis effort or the discrepancies between
the plant configuration and design documentation.
Further, the
NED organization did not communicate to the station that
discrepancies were found or how they were resolved.
b.
MOV Design Thrust Window Data
The licensee contracted with Bechtel Corporation to provide MOV
thrust windows for worst ca~e differential pressure and minimum
- terminal voltage.
The NED MOV group adjusted the Bechtel thrust
window values as a function of torque switch tolerance, valve
lubrication history and measured equipment uncertainty. The
adjusted thrust values were sent to the site via letter. Station
personnel used the letter to determine the acceptable valve thrust
in *the VOTES analysis.
6
c.
documentation did not contribute to the incorrect torque switch
setting.
Use of a Non-Safety-Related Consulting Service
Prior to the outage, the licensee contracted with Babcock and
Wilcox Nuclear Technologies (BWNT} to provide consulting services
for VOTES data analysis (Purchase Order Number 333188}.
In
accordance with the contract, BWNT was to issue a formal analysis
report within five days of receipt of valve data.
The purchase
order was issued as non-safety related.
Under this purchase order, a site technician sent the reactor
recirculation discharge valve digital test data to BWNT for
analysis, as described in section 3 above.
Assisted by the BWNT
recognized industry expert, the technician selected a zero point
for the trace.
When the BWNT industry expert participated in
selecting the zero reference point for this safety related valve,
he performed safety related work.
During the review of the valve trace, the BWNT industry expert
determined a problem existed with the valve. This observation was
not communicated to the site during the zero reference point
dialogue.
As stated in the contract, the industry expert wrote a
VOTES summary report.
The report concluded the valve might not
shut off flow under actual operating conditions and recommended
that the valve be disassembled and deficiencies corrected.
However, BWNT failed to forward the summary report to the
licensee.
The site MOV coordinator failed to recognize that the
summary report was not received.
The inspector concluded that the procurement of the consultant
services under a non-safety purchase order was not a contributing
factor to the event. All practical aspects of a safety related
program were met.
The expert was qualified and the analysis was
documented in accordance with BWNT's QA program.
The report was
placed into the vendor's QA document control program.
Procuring
the consultant services under a non-safety related purchase order
had no bearing on the consultant's failure to submit the report to
the licensee.
d.
Test Procedure Quality
The inspector reviewed the work package (WR 091331} and testing
procedures for VOTES testing of the reactor recirculation
discharge valve.
The irispector identified the fol~owing
weaknesses associated with the documentation of the work
activities:
(l}
Neither the work package nor the testing procedure
delineated any quantitative or qualitative acceptance
criteria related to the VOTES diagnostic, the as-found or
5
6.
The inspector requested to review the NED adjustment calculations.
However, these calculations were not available since they were not
retained under a QA program.
Licensee Event Report Followup
(Open) LER-237/91-023, 2A Retirculatiori Pump Discharge Motor-Operated
Valve Failure to Close.
The inspectors identified that the LER
contained incorrect and misleading information.
The licensee committed*
to revise the LER, correcting inaccuracies, and to review the LER
- information gathering process.
7.
IO CFR 21 Evaluation Followup
The station procedure for review and processing significant conditions
adverse to quality, Dresden Administrative Procedure (OAP) 2-8, provides
three Part *21 deviation screenings.
The first review is performed by
the Operation Engineer, the second by the deviation report coordinator
and a final review by the on-site review group.
When a deviation is
identified through the screening, the information is forwarded to the
corporate Part 21 coordinator.
The coordinator then ensures that an
evaluation for a defect, under IO CFR Part 21, is completed.
For the reactor recirculation discharge valve incorrect torque switch
setting, all three screenings concluded that a Part 21 evaluation was
not required;
However, the LER concluded that li~itations of the VOTES
computer software contributed to the improper zero reference point
identification.
The inspector inquired why a Part 21 evaluation was not performed beyond
the initial screenings. During the inspection, the licensee began a
10 CFR Part 21 evaluation of the VOTES process relating to the bending
valve yoke phenomenon.
8.
Analysis of Inspection Findings
- ---
Several weaknesses associated with the VOTES testing program were
identified in the site management control system:
The management control system did not ensure an adequate
root cause evaluation was completed following the failure of
the RR discharge valve to close.
The management control system did not ensure the
qualification requirements, as outlined in NOD-MA.I, were
. documented for personnel performing VOTES testing.
The management control system did not ensure consultant
services used for safety-related work were procured on a
safety related purchase order.
7
The management control system did not ensure the VOTES testing
procedure/work package delineated quantitative or qualitative
acceptance criteria.
Given an emergent technology, the management control system did
not establish a commensurate level of review of the VOTES
analysis~
The management control system did not ensure the abnormal
VOTES trace was identified and documented as a condition
adverse to quality.
The management control system did not ensure that LER 91-023
contained factual information.
Several additional weaknesses w~re identified in the NED Management
Control System:
The management control system did not ensure the re-evaluation of
the remaining Unit 2 MOVs was performed and documented under
rigorous management controls.
The management control system did not ensure engineering personnel
recognized or documented the failure of three Unit 2 MOVs to meet
the design specification as a condition adverse to quality.
The original unresolved item, 50~237/91022-0S(DRP), issued on this
matter will remain open.
Closure of this item is dependent upon
completion of the licensee's 10 CFR 21 evaluation of the VOTES process
and further NRC review of the licensee's condition adverse to quality
system and licensee calculation controls.
An unresolved item is a
matter requiring more information to ascertain whether it is an
acceptable item, an open item, a deviation, or a violation.
9.
Exit Interview
The inspectors met'with licensee representatives (denoted in
paragraph l) during the inspection period and at the conclusion of the
inspection period on January 14, 1992.
The inspectors summarized the
scope and results of the inspection and discussed the likely content of
this inspection report .. The licensee acknowledged the information and
did not indicate that any of the information disclosed during the
inspection could be considered proprietary in nature.
8