IR 05000352/1986023
| ML20207G854 | |
| Person / Time | |
|---|---|
| Site: | Limerick  |
| Issue date: | 12/22/1986 |
| From: | Gallo R NRC OFFICE OF INSPECTION & ENFORCEMENT (IE REGION I) |
| To: | |
| Shared Package | |
| ML20207G828 | List: |
| References | |
| 50-352-86-23, NUDOCS 8701070413 | |
| Download: ML20207G854 (31) | |
Text
.
.
U.S. NUCLEAR REGULATORY COMMISSION
REGION I
Report No. 86-23 Docket No. 50-352 License No. NPF-39 Licensee: Philadelphia Electric Company 2301 Market Street Philadelphia, Pennsylvania 19101 Facility:
Limerick Generating Station, Unit 1 Inspection Period:
September 16 - November 26, 1986 Inspectors:
E. M. Kelly, Senior Resident Inspector S. D. Kucharski, Resident Inspector A. E. Finkel, Lead Reactor Engineer k
22hg Approved by:
Robert M. Gallo, Chief, Projects Section 2A Date Summary: Routine daytime and backshift inspections (285 hours0.0033 days <br />0.0792 hours <br />4.712302e-4 weeks <br />1.084425e-4 months <br />) of Unit 1 by the resident inspectors and Region I Specialists consisting of:
followup on outstanding items; walkdown of the RHR system using PRA guidance; plant tours including fire protection measures; maintenance and surveillance observations; and review of LERs and periodic reports.
The inspection covered a feedwater minimum flow valve failure on October 24 and a noise-induced half-scram signal on November 13.
Independent inspections were conducted that evaluated the RPS power supply breakers, the recirculation pump trip breakers, and the potential use of HGA relays in class 1E circuts. Several meetings were held onsite during the period, addressing QATTS findings and the Limerick PRA update, and these are documented herein.
No violations were identified.
Unresolved items were initiated for: isolation of loop A drywell spray; effect of long-term nitrogen bottle supplies on ADS; and completion of a safety evaluation for RCIC injection valve 1F013.
The inspectors noted strong engineering support within the Maintenance Division for the HPCI steam valve packing problem, and within the Field Engineering Group I
!
8701070413 861223
~
PDR ADOCK 05000352 O
!
{
,
,
-.
-
-
.
.
for the RPS power supply and APRM noise issues.
I&C technicians continue to exhibit proper response to unplanned testing conditions by immediately stopping work and informing control room supervision.
Control of noise, unnecessary traffic, and alarms in the main control room were significantly improved during the inspection period, and the number of reportable events have continued to be low. Reactor operation at near 100% capacity factor has reflected the low incidence of personnel errors and equipment failures experienced over the past 3-4 months.
l f
l l
l
!
-
-_
.
.m-..--
-. _,-
, _. _. _, - _ _ _ -,-
- - _,. _ _ - _.. _ _..
-
.. -.
-...
. - -
- - -
.
.
,
DETAILS 1.0 Principals Contacted Philadelphia Electric Company J. Corcoran, Engineer-In-Charge, Field QA R. Diederich, Supervising Engineer, PRA Group J. Doering, Superintendent of Operations R. Dubiel, Senior Health Physicist P. Duca, Technical Engineer J. Franz, Station Manager J. Milito, Field Engineering Supervisor R. Moore, Superintendent, QA Division W. Volmer, Mechanical Engineering Also during this inspection period, the inspectors discussed plant status and operations with other supervisors and engineers in the PECO, Bechtel and General Electric organizations.
2.0 Followup on Unresolved Items 2.1 (Closed) Unresolved Item (80-17-05); Cutting of Non-Designated Field Adjusted Ends of Jet Barrier Steel.
The inspector reviewed the following documents related to the instal-lation of jet barrier steel piece 3788-1:
-
Bechtel drawings C-972, C-973 and C-975.
Applied Engineering Co. drawing NS-3412-37 (8031-C43EE-52-3)
-
-
PECO Finding Report N-220 Bechtel QC Inspection Report C-972-C-63-1
-
-
Bechtel Specification C-63, " Erection of Structural Steel" Bechtel Field Change Report (FCR) C-7578F
-
-
Bechtel Field Memoranda Engineering (FME) 4135 The installation specification was revised to allow the field to cut structural steel without engineering approval provided that minimum edge distance for AISC criteria was met and engineering required minimum dimensions were maintained. The CRD jet barrier steel and connections were inspected by QC and found in accordance with engineering and AISC requirements. Based upon the specification re-vision, the subsequent QC acceptance of the steel installation and the inspector review of the documents listed above this item is closed.
s v
s
, - -
--
,
- -, - - -
-e
--
,
.-
2.2 (Closed) Unresolved Item (83-09-01); Justification for Use of ASCO Valves in a High Temperature / Humid Environment.
The inspector reviewed the Environmental Qualification Review Records (EQRR) of ASCO solenoid operated air valves for the following systems:
-
Equipment and Floor Drains - Liquid Radwaste Collection Emergency Service Water
-
-
-
Core Spray Containment Atmosphere Control
-
Reactor Core Isolation Cooling
-
Muclear Boiler (Inboard isolation valve for Main steam line Sam-
-
ple line)
-
Reactor Recirculation High Pressure Cooling Injection
-
The inspector found these valves to be environmentally qualified ac-cording to the specifications set by the licensee. Based on the in-formation for the above systems contained in EQRR Nos. 77, 78, 80, 81, 82, 83, and 124, the valves were appropriately qualified for high temperature and humidity (each Revision 20 dated April 1, 1986) and this item is closed.
2.3 (Closed) Unresolved Item (86-03-03); Addition of Nitregen Backup for Toxic Gas Analyzers to Daily Surveillance Log.
The inspector reviewed surveillance test procedure ST-6-107-590-1, Daily Surveillance Log, Opcon 1, 2 & 3, Revision 14, September 23, 1986.
The licensee has incorporated into the daily round procedure the recording of the backup nitrogen bottle pressure and the reduced pressure reading to the header for the backup nitrogen bottles for the toxic gas analyzers. Also, added to the procedure is a require-ment to changeout the bottle when the pressure drops below 1000 psi.
These readings are taken daily. This item is closed.
2.4 (Closed) Followup Item (82-16-01); Control of PECO Construction Field Office Memorandum (CFOM)
The inspector reviewed the PECO computer data base for the CF0M log.
The log identified which CF0M required a response, the Bechtel re-sponse document identification, and the current action complete
+
-
.
.
.
.
- qe
'
t status for the CFOM. The inspector interviewed PEC0 personn'al re-garding the development and maintenance of the log. The inspector found the log satisfactorily maintained. This item is closed.
'
2.5 (Closed) Unresolved Item (86-04-02); RPS Power Supply Breaker Coil
'
,
L Operations
!
,
This item is addressed in Detail 5.4.
'
\\
3.0 Review of Plant Operations
T
'
3.1 Summary of Events
'
.. (c
'
The plant operated at full rated power throughout the inspection
/,"
,
period.
'
s 3.1.1 Reactor Availability
~~
'
'
,
Unit 1 was in its 137th day of contindous power operaticd
,
,
at the end of the inspection period. Plant activities were'
)
observed to be carefully considered and implemented through-
out the inspection period.
The plant was maintained at an average capacity factor of 98.5% over the inspection period,
.
'
and is consistent with reactor operation since startup on
'
July 13, 1986. Core burnup as of November 19 was approxi-t z
mately 6570 MW-days per metric ton'of uranium ors 31G effec-
'
f
_
tive full power days (EFPD) of Unit 1 first cycle coces S operation.
>
s 3.1.2 Refueling Outage Planning
'i.
.
Initial refueling is scheduled to last 11 weeks and expect-
^!
'
ed to begin in May 1987. An end-of-cycle coastdown will s
commence at a target burnup of 371 EFPD. _ Because of the
.
high capacity factor experienced, a Unit'l license amend-
'
ment is being prepared for reactor coastdown at increased l core flow and decreased feedwater heating prior to shusting down.
The refueling will involve a full core offload 'and the replacement of 292 (or 38%) of the 764 fuel bundles.
Proposed major outage work will involve:
,
630 corrective maintenance items and:450 preventative l
-
.
maintenance items
'
<,
,
70 to 80 modifications, including a SGTS tie-in on the
-
refueling floor and a condenser cathodic protection system
'
'
-
turbine rotor inspections and main g'enerator i
maintenance
,
b
- i'
c'
M i
.
'~'
.)G.;@ %s Y g(
,,9<I
,,
- s
.1
'
. p'
5,
r s
r
~.
,
e.7 pp-
!
\\\\
'
' t a
-
,
.ys u
.+c s
,,
M,f
[
)
/.)
h h
y y
y a
,
A
,
,
!
s
.,
,
s
.
.
.Y
'everhaul of all four emergency diesel engines and 20
-
'd-coritrol rod drives.
,
4,
.
,
\\
h
'
^
Pland C emistry 3. L. 3 v
e--
r a-s
!
\\
,0 Reactor water conductivity was maintained in the range of l
0.112 to 0.195 micrombos per centimeter (umho/cm), averag-
.
'
j.;
'ing about 0.15 umho/cm during the inspection period.
Con-T
-
eductivity has trended downward since July 1986 and has
'
remained in a narrow range about 0.12 umho/cm over the last
,
,
half of the inspection period. A condanser tube leak was
,
experienced at the beginning of the inspection period,
-
estimatad to be 40 gallons per day, which eventually
'
splugged itself. The leak resulted in a decreased cleanup demineralizer bed lifetime during September - October 1986.
^
,; g
,Fue14ntegrity; monitoring of primary coolant gross radicio-
"
-- dines 'and,offgas activities have indicated no degradation.
,
.k,
'
,
,1
'
Chemical treatment with Retanone was performed on October 16
'
<
%
"'
.
for spray ponc' fish removal, and normal operations of the
'
ESW and NR service water systems were restricted through
,
, Octcber 2h
.
,
-
,
,
,
~
3.1.4 L Other Events
'
..
.
>.
I
)' Unidentified drywell leakage was at 1.5 gpm at the end of
'
'
_
tti irispection period, and has exhibited an increasing trendfof 0.5 gpm Isr month during the period. At this b
' fats',' the k chnical Specification limit of 5.0 gpm_would be redched'in May 1987.
'
<
,
-
a
,
'/
Turbine stop valve' testing on October 12 uncoverad failures
-),
in th6 test, circuitry'which were corrected and are dis-
's
\\,
cussed further in Detail 6.3.
A packing replacement was
>
-
q
/*
performed fx the outboard HPCI steam supply valve during l ( y'
' a th6 period.0ctooer 17-23 causing HPCI to be inoperable for
,
'!;.
6-1/2 days. JMaintenance and LLRT for the valve are dis-
.
,
?'.
!!!
cussed in Detail 7.2.
Discussed in detail are: a feedwater s
"
t transient which ' occurred on October 24 that caused a reactor
vesW " level dicrease due to the "A" feed pump minimum flow a
R'
valve failing open; a RCIC injection valve packing leak
'
.
'
=found on November 3; and, a half scram caused by noise in-s
'
'
duced on two APRM channels on November 13.
'
3.
s.
(
3.1.5 NRC Soecialist Inspectors During the inspection period, the following NRC specialist
+
+
inspections were conducted:
,
J
%
,i
.
.
I
h,
',f '
s t
y
,W
,
[/
'
..
..
-__
_
..
.
.
-._
__
_._.-.
__.
_
-
_
_ _ _..
._
!,
.
/
,
Inspection No.
Subject
'86-19 Security
'
86-20 Maintenance
-
86-21 Radiological Controls 86-22 Training
'86-24 License Exams
'
86-25 Security 86-26 Chemistry
,
3.2 Operational Safety Verification-3.2.1 Control Room Activities The inspector toured the control room daily to verify prop-
'
er manning, access control, adherence to approved proce-dures, and compliance with LCOs. The inspector reviewed shift superintendent, control room supervisor, and operator logs covering the entire inspection period. Instrumentation and recorder traces were observed and the status of control room annunciators was reviewed. Nuclear instrument panels and other reactor protective systems were also examined.'
.
Effluent monitors were reviewed for indications of abnormal
)
releases; none were evident. Panel indications for
onsite/offsite emergency power sources were examined for
'
automatic operability.
Also during the inspection period, control room activities
,
were observed for the following attributes:
Operator Professionalism
-
Licensed operators were businesslike and had a profes-sional appearance, exhibited a sense of pride, and
were knowledgeable of plant status.
,
-
Noise Control Conduct of business with shift management was in an orderly manner, with little interference from routine duties. Although Unit 2 Control Room construction was underway during the inspection period, noise levels were kept to a minimum.
Control Room Access
-
The licensee had implemented new access control mea-sures at the beginning of the inspection period. These included signs to preclude routine entry during shift turnovers and entry for other than official business,
,
the controls were judged to be effective during the period observed.
I
!
.
.
-
.
. -
-
__
.
-.
. -
.
..
. ~.
.
.
.
. -
-
.
W
Annunciating Alarms
-
The number of nuisance and routinely alarming condi-tions was significantly reduced and maintained at a low number over the inspection period.
Less than ap-proximately ten alarms on safety-related panels were consistently present, and the licensee continues to track and aggressively pursue solutions towards unnec-essary alarming conditions.
-
Control Room Appearance Overall appearance of the control room was clean, and orderly, with appropriate procedures / manuals filed where needed.
3.2.2 Security and Radiological Controls During entry to and egress from the protected area and vi-tal island, the inspector observed access control, security boundary integrity, search activities, escorting and badging, and availability of radiation monitoring equipment including portal monitors.
No unacceptable conditions were found.
3.2.3 Physical Plant Conditions Sampling reviews were made of equipment trouble tags, night orders, and the temporary circuit alteration (TCA) and LC0 tracking 1 cgs. The inspector also observed shift turnovers during the period. The operations activities were observed for conformance with the applicable procedures and require-ments; no unacceptable conditions were noted.
One condition was found where 55 gallo, drums of charcoal were stored in a high risk fire zone at Reactor Enclosure elevation 217 for over a month. A fire watch was provided until the drums were removed. The inspector questioned the lack of a written safety evaluation for the storage, al-though he concluded that the condition did not represent a significant change to the fire hazards report.
3.3 Station Tours The inspection toured accessible areas of the plant throughout this inspection period, including: the Unit I reactor and turbine-auxiliary enclosures; the main control and auxiliary equip-ment rooms; battery, emergency switchgear and cable spreading rooms, and the plant site perimeter. During these tours, observa-tions were made of equipment condition, fire hazards, fire protec-tion, adherence to procedure, radiological controls and conditions,
-
.
_
._
.
.
housekeeping, security, tagging of equipment, ongoing maintenance and surveillance and availability of redundant equipment.
No unacce,et-able conditions were identified.
3.4 System Walkdowns 3.4.1 Engineered Safeguards Features (ESF) System Walkdown The inspector performed a detailed walkdown of portions of the RHR system in order to independently verify system op-erability.
The walkdown included verifications of the fol-lowing items:
Review of RHR technical specifications, the FSAR, sys-
--
,
tem operating procedures and P&ID's
,
--
Inspection of RHR equipment condition
--
System check-off-list (COL) and operating procedures consistent with plant drawings
--
Valves, breakers, and switches properly aligned, in-cluding appropriate locking devices Instrumentation properly valved-in and operable
--
--
Satisfactory control room switches, indicators and controls Within the scope of the inspection, no unacceptable condi-tions were noted.
3.4.2 PRA-Based System Inspection The inspector performed selected system walkdowns utilizing methods prescribed in a study prepared for the NRC by Brookhaven National Laboratory using the Limerick Probabi-listic Risk Assessment (PRA). The study entitled PRA-Based System Inspection Plan dated February 1986, provides in-spection guidance by prioritizing plant safety systems with respect to their importance to risk. The study contains an abbreviated version of the licensee's system checklists which contain components that are considered to have a high risk factor as determined by the PRA.
The inspector verified the proper configuration of the fol-lowing RHR low pressure injection (LPCI) system components on several occasions during the inspection period:
..
_ - _.
. _ - _
.
i
.
-
RHR heat exchanger Service Water, inlet valves HV-51-IF014A and B RHR heat exchanger Service Water, outlet valves
-
HV-51-IF068A and B RHR heat exchanger Service Water, bypass valve
-
HV-C-51-1F0488 LPCI injection valves HV-51-1F017A, B, C and D
-
No unacceptable conditions were noted.
3.5 QATTS Trending An onsite meeting was held on August 5, 1986 with licensee represen-tatives from the Electric Production Quality Assurance Organization to discuss their use of the QA Trending and Tracking System (QATTS).
The meeting was in response to an NRC SALP report recommendation.
The QATTS is designed by and unique to PECo. An auditor typically enters individual findings into QATTS. There have been a total of over 20,000 open items entered to-date in the following_five categories:
OIL 1 Significant NCRs
-
-
OIL 2 Deficiencies
-
OIL 3 NRC
-
-
OIL 5 Correspondence, Bulletins, LERs There are approximately 800 items in QATTS currently open and requir-ing verification. The QATTS was described as being used not just as a QA tool, but also as a QC data base for the statusing of licensing, radiation protection, and maintenance items. All PECo organizations are both initiators and users of QATTS.
The inspector reviewed data presented in a recent QATTS report and discussed the meaning of those findings with licensee representatives. Trends dating back to Unit I licensing were displayed in a number of functional areas and were observed to generally compare with similar Peach Bottom Station findings.
3.6 PRA Update A meeting was held onsite on October 17, 1986 with licensee PRA group representatives to discuss the recent update of the Limerick Unit 1 Probabilistic Risk Assessment (PRA) Revision 4, September 1986.
The licensee decided to update the " Level 1" portion of the PRA, re-calculating a new core damage frequency due to internal events
-._ -
-
_
-
.
-
_
. - _ _ _
- - _
.
.
only. The update consisted of modifying system fault trees to re-flect the as-built system designs, and revising the accident sequence event trees to include consideration of the emergency operating TRIP procedures, updated station battery life estimate and a changed MSIV closure setpoint. The result of the update was a reduction by a fac-tor of 3 in the calculated core damage frequency.
In the original PRA, the accidence sequence contributing the greatest percentage to the core damage frequency was a loss of offsite power followed by a loss of the diesel generators or a Station Blackout),
which in turn fails all emergency coolant injection. The frequency of this sequence was reduced because of:
a reassessment of station battery life from four to eight hours.
-
Station Blackout Procedure E-1 procedure which instructs
-
operators to initiate alternate room cooling for the high pressure pumps, and align the high pressure pump suction to the CST to avoid damage to the pumps when the suppression pool temperature reaches 165 degrees F.
Consideration of the TRIP procedures and increased confidence in op-erations depressurizing the reactor resulted in an order of magnitude reduction in certain events.
Updated initiating event frequencies were generally lower than originally assumed. Also a lowering of the MSIV closure setpoint from -38" to -129" water level in the reactor vessel reduced the postulated frequency of MSIV closures.
During the meeting, licensee representatives summarized the following from the updated PRA, offering insight into the revised core damage frequency for Unit 1:
40% of the total core damage frequency is due tc transient
-
initiated sequences where a failure of long-term RHR produces containment failure and subsequent core damage. Containment venting was not included in this update.
25% of the total core damage frequency is due to loss of offsite
-
power initiating event followed by a failure of all on-site AC power (Station Blackout); and, 20% is attributable to ATWS sequences.
The initiating events whose sequences are the dominant contri-
-
butors to the core damage frequency are turbine trips and loss of offsite power.
-
Loss of coolant accident sequences make a very small contribu-tion to core damage frequency.
The licensee plans to use the current PRA as an analytical tool for
evaluating items which might affect reactor safety. This would
.
.
include performing cost / benefit analysis on design changes and modi-fications, evaluating technical specifications, and providing a pri-oritization method for an integrated living schedule for licensing actions. A proposed technical specification change to RPS instrument calibration frequencies and out-of-service times is being prepared using the Limerick PRA as supportive bases, and is expected for sub-mittal to the NRC in 1987.
4.0 Onsite Followup of Events The inspector performed onsite followup of the following events that oc-curred during the inspection period. The events were evaluated for proper notification to the NRC, reactor safety significance, licensee efforts to identify cause and propose effective corrective action, and verification of proper system design response.
4.1 Feedwater Pump Minimum Flow Valve Failure During the load reductions for control rod scram testing, the A reac-tor feedwater pump (RFPT) minimum flow valve failed open on October 24 due to a loss of speed signal.
Reactor water level dropped from a normal level of 34 inches to 20 inches during this event. The low water level reactor scram setpoint is at 12.5 inches. A reactor re-circulation pump runback to 75% speed occurred, as designed, but re-actor power remained unchanged since the reactor was at 2100 MWt or 64% power at the time. Operators were in the process of removing the C feedwater pump from service when the loss of speed signal to the A pump minimum flow valve caused it to fail wide open. The operator took immediate manual control of the running feedwater pump and that action considerably lessened the vessel level drop.
Automatic minimum flow valve control is only required when the ultra-sonic minimum flow system is operating, and has been inoperable due to reliability problems.
Valve control has been manual since the recent May-June 1986 outage during which manual / auto stations were installed in the control room. The minimum flow valve automatic function was disabled on the A&C pumps. The B minimum flow valve is manually blocked due to a leak.
Further evaluation of minimum flow control is currently being done by licensee engineering under modifi-cation 84-0116. Also, the ISEG is preparing an event report.
4.2 HPCI Steam Supply Valve Failure The outboard isolation valve HV-55-1F003 in the steam supply to the HPCI turbine was discovered to have its packing dislodged on October 17. Although the normally open valve had further opened to its full backseat position - sealing itself and preventing any steam leakage - the HPCI system was removed from service at 11:56 a.m. on October 17 to repair the valve since the plant entered a 4-hour LCO
'.
. - - -
..
'
.
for containment isolation purposes upon discovery and evaluation of the failure. An ENS call informing the NRC of the HPCI system inoperability-was subsequently made at 9:11 p.m. on October 17. The 1F003 valve repairs were successfully completed and HPCI returned to service on October 23, 1986.
Inspection in Detail 7.2.
The inspector noted that the ADS gas supply bottles were isolated at the time of the HPCI system inoperability, and questioned the licensee as to their effect on ADS operability. The bottles are de-signed as a long-term supply of actuating power for the SRV's which are an alternate mode of long-term decay heat removal. The effect of gas bottle availability on the ADS function is unresolved (86-23-01)
pending further evaluation.
4.3 RCIC Injection Valve Packing Leak On November 3, 1986 a steam leak was discovered on RCIC valve HV-49-IF013, located in the outboard MSIV Room. The leak on the nor-mally closed valve was successfully stopped by opening and electri-cally backseating the valve. The operability of the valve was subsequently demonstrated by a stroke time / exercise test from the backseated position.
To maintain the valve in this position, the following RCIC system configuration changes were made:
HV-49-1F013 was opened and electrically backseated via TCA
-
- 781.
-
Upstream isolation valve HV-49-1F012 was closed.
RCIC full flow test valve HV-49-1F022 was blocked closed
-
with the feed removed.
Manual supply valve 49-1F061 from Condensate Transfer was
-
blocked closed.
Three concerns were created by the above system change: RCIC
.
oerability, RCIC isolation capability, and high energy line break (HELB) analysis.
RCIC operability and isolation capability were im-mediately determined to be unaffected and a supportive safety evalua-tion associated with TCA #781 documented those conclusions.
The leak was initially suspected to be from the valve's pressure seal, and torquing the bolts did not stop the leak. HV-49-1F013 remained open and backseated with the outboard injection valve 1F012 closed. A firewatch was posted to monitor the line between the 1F013 and IF012 valves due to the concern regarding a potential high energy line break. A HELB analysis was being performed by Mechanical Engineering but was not available as of the end of this inspection.
The full safety evaluation of this temporary modification to the RCIC pump discharge valves is therefore unresolved (86-23-02).
__.
-.
-
--_. _ - - -
-.
--
.
.
4.4 Noise Induced on APRM Channels 4.4.1 Background A half scram signal was received on the A2 RPS Channel at 3:02 p.m. on November 13. An automatic sequence-of-events computer printout showed APRM upscale alarms on the C and D APRM channels that feed the A2 and 82 RPS channels respec-tively, suggesting that a full scram logic should have been satisfied. Control room operators immediately verified that no reactor power changes or unusual recorder traces on the four other APRM channels had been experienced.
Plant management and engineering support were promptly in-volved, and evaluation of transient computer ERFDS print-outs showed noise spikes on the C and D APRM channels of an approximate 100 millisecond duration. The flow-biased trip setpoints of those channels were known to be 114 and 115%,
and the transient data indicated peaks of 125 and 113%.
The C APRM channel clearly exceeded its trip setpoint for a long enough duration to allow the associated K14 scram re-lays to de-energize and cause the half-scram condition.
However, the D APRM channel did trip as shown by the pro-cess computer and local APRM upscale alarms.
Because the ERFDS sample frequency is every 50 milliseconds, the licensee concluded that the duration of the signal on D APRM channel was not long enough to allow the associated K14 scram relay contacts to open, although apparently high enough (above the 115% trip setpoint) to begin the logic for other interposing relays such as K12.
4.4.2 Immediate Actions Surveillance testing was performed on the evening of Novem-ber 13 to verify APRM Channel D operaoility, and an RPS channel B2 response time test was also performed that showed the K140 scram relay properly responding within 35-40 milliseconds of an APRM upscale condition. There-fore, a full scram would've occurred if the "D" channel upscale signal duration were approximately 10 milliseconds longer.
Unit I was at 100% power at the time of the event.
The resident inspector witnessed the half-scram signal and subsequent licensee actions. A Nuclear Review Board Meeting was convened on November 14 to review the event, and was l
attended by the Senior Resident Inspector. The NRB con-cluded that the RPS response was reasonable and proper, but recommended monitoring APRM channels during the following weekend to detect any (if other) further noise. No abnor-mal signals were subsequently found.
The origin of the noise spike has not been ascertained, although welding or radio transmissions in the area are
- -
.
-
-
.
.
.
.
suspected to be potential causes and have been suspended pending further investigation by the licensee. The indepen-dent safety engineering group (ISEG) was directed to perform an event report addressing the cause and effect of the noise-induced problems.
Limerick Unit 2 activities underway at the time of the noise spikes may have included an activity termed "PVC-search" whereby a high frequency signal is introduced onto conductors inside PVC conduit prior to core-boring into concrete walls. No coincident activities were identified as of the end of the inspection period.
4.4.3 Independent Evaluation The inspector discussed the RPS logic and APRM cable rout-ing with Field Engineers and ISEG representatives, walked-down the conduit containing the LPRM cable that feeds the C and D APRM channels, and interviewed the STA and other personnel involved in the November 13 noise tran-sient. An element of commonality found was that the LPRM cable feeding the C and D APRM channels is run in the same electrical conduit.
LPRM signals are transmitted on 75-ohm coaxially shielded cable that is connected once at drywell penetration JX-100C and then not again until terminal cabi-net 10C787 in the auxiliary equipment room. Grounding problems or introduction of noise on these cables would be likely to occur at either of the two connection points.
Intervening cable runs are completely enclosed in conduit i
and well shielded. The C APRM channel is normally fed by 21 LPRM detectors; however, two (2A-48-49 and 3A-32-33)
were bypassed at the time of the event.
The D APRM channel is normally fed by 22 LPRM detectors and all were in ser-vice on November 13. Two of the 19 active detectors (4A-16-17 and 5A-48-17) for the C APRM, the channel that remained upscale long enough to cause the half-scram, were found by operators to have individual upscale trip alarms.
The individual LPRM trip units are set at 100% of full scale but at different currents corresponding to that de-tector's normal fullpower flux exposure. Normal LPRM cable signals are small currents in the 500-900 microampere range.
The RPS relay responses during the noise event were as ex-
'
pected and per design.
The K14 scram relay is a General Electric CR1050 contactor with three main contacts and six auxiliary contacts. There are eight K14 relays, used exclu-sively in the RPS, and these were preoperationally timed in May 1984. The newly installed relays opened at between 3.8-5.5 milliseconds following coil de energization.
These same relays are currently installed and will seal-in when actuated fully. The interposing K12 relay is a small AC non-latching MDR series device manufactured by Potter &
_ - - - -.- -
-_
..
-
.
. -...
...- -
-, -.
.
.
Brumfield. The K12 is specified to de-energize and open in 5-18 mi"iseconds, but does not seal-in or latch. While each o.'. ' K-12 interposing relays in the C and D APRM scram logic were recorded to have actuated (i.e. contacts open as shown on the process computer Sequence-of-Events log), only the C APRM (or RPS 12 channel) signal was present long enough to keep the associated K12 relay de-energized long enough to drop-out the K14 C&G scram re-lays. The B2 RPS logic channel fed by the D APRM signal experienced an upscale signal for a time short enough to open and re-close the associated K12 relay contacts without de-energizing the K14 D&H coils. Therefore, assuming that the original K14 operation time of approximately 5 milli-seconds is still representative (i.e., no slowing with normal aging and wear), then the upscale signal that existed on the D APRM channel was less than 5 milliseconds in duration.
Response times on the order of 5-10 milliseconds are a frac-tion of the total RPS response ti nes subsequently proven for the D APRM channel (35-40 millise.conds) and required by the technical specifications (90 milliseconds).
The inspector also reviewed computer records of plant pa-rameters, control rod positions and fuel thermal limits for times prior to and after the APRM noise was detected.
No appreciable changes were recorded, and normal variations about the following plant parameters were seen in the 30 seconds preceding and following the event Reactor Power (A APRM)
98-7 - 99.2%
A Feedwater Flow Steady 0 4.8 M1b/hr B Feedwater Flow Steady 0 4.6 M1b/hr Reactor Pressure 1010.6 - 1011.7 psig Reactor Water Level 35.2 - 35.6 inches Total Steam Flow Steady 0 15.4 M1b/hr Feedwater Temperature Steady 0 429 deg. F Total Jet Pump Flow 98.2 - 98.4 M1b/hr The inspector also reviewed fuel thermal limits from P-1 computer edits taken at 3 and 7 p.m. on November 13 and found all limits at the twelve most-limiting core locations to be within specifications and not indicative of any un-usual fuel conditions.
4.4.4 Conclusions The inspector concluded that no actual transients were ex-perienced which would have required a reactor scram. Also, the RPS functioned as expected and within appropriate re-sponse times.
Finally, the inspector concluded that, while an ISEG investigation was exploring the potential causes of
.-
.
the noise, the initiating event was most likely not expect-ed to be found or duplicated.
5.0 Licensee Reports 5.1 In-Office Review of Licensee Event Reports The inspector reviewed Unit 1 LERs submitted to the NRC Region I of-fice to verify that details of the event were clearly reported, in-cluding the accuracy of description of the cause an adequacy of corrective action.
The inspector determined whether further informa-tion was required from the licensee, whether generic implications were involved, and whether the event warranted on-site followup. The following LERs were reviewed:
LER Number Report Date Subject 85-102 August 20 Excessive Leakage of Containment Spray (Note a)
Valve, HV-51-1F016A 86-042 October 10 Immediate Surveillance Test Due to Personnel Error and Procedural Deficiency 86-043 September 25 Toxic Gas Detection System Operating (Note a)
in Nonconservative Condition 86-044 October 15 Vinyl Chloride Toxic Gas Alarm Setpoint Error 86-045 October 15 NSSSS Isolations Due to Personnel Error and Equipment Problems (86-046 October 23 Main Control Room Chlorine Isolations l
& Emergency Fresh Air System Actuations86-047 October 31 RCIC Steam Supply Isolation during testing
,86-048 November 3 RWCU Isolation during troubleshooting (Note a)86-049 November 7 RWCU Isolation during testing (Note a)
Notes:
a: Addressed in Detail 5.2 of this report.
l i
_-
- _ _.
-_ _-
_
.-
.-
. - _ - _ _ _ _ - _ _ _ _.
- _..
.
__
5.2 Onsite Followup of Licensee Event Reports For those LERs selected for onsite followup as noted in Section 5.1, the inspector verified the reporting requirements of 10 CFR 50.73 and technical specifications had been met, that appropriate corrective action had been taken, that the event was reviewed by the licensee, and that continued operation of the facility was conducted in accor-dance with technical specification limits.
5.2.1 LER 85-102: Excessive Leakage of Containment Spray Valve, HV-51-1F016A As previously discussed in Inspection Report 86-03, a local leak rate test was performed in December 1985 on penetration 39A for the "A"-loop drywell spray header. The penetration failed to hold pressure when the normal method of performing the test was used. As-found leakage was in excess of 60,000 standard cubic centimeters per minute (sccm). Outboard
>
isolation valve HV-51-1F016A, was manually tightened shut, the test was re performed, and penetration leakage was reduced to 372.4 SCCM. A supervisory block was applied to the hand switch for the 16A valve in the control room, its'
motor control center was de-energized, and the manual hand-wheel for the valve was secured in the full closed position.
Although technical specifications do not require operability of drywell sprays, closing the 16A valve rendered the "A" loop of containment spray inoperable for the subsequent five months of plant operation.
During the 6-week outage in May-June 1986, maintenance was performed on the 16A valve operator that included cutting a slot in the torque limit sleeve of the spring pack used to sense motor torque. This was to allow the release of any excess grease that may have accumulated and was suspected as a cause of the valve's failure to hold pressure. The torque switch on the valve operator was adjusted to allow approximately 90% of the maximum possible torque to close the valve. However, on June 4, 1986 the penetration was re-tested and failed to achieve test pressure because of leakage. The valve was again manually closed and the "A" loop of containment spray remained inoperable. The licensee plans to repair the valve during the refueling outage
scheduled to begin in May 1987.
The inspector questioned the allowable out-of-service time with one containment spray loop inoperable, and the risk of losing the other loop coincident with an accident beyond design basis. This will be followed as item (50-352/86-23-01).
r
-
,- - -. - - - -.
-
--
,
--.
- -
e,-----
- -,.
- ---- - -
,
.
.
5.2.2 LER 86-048: NSSSS Isolation Due to Improper Valve Sequence On October 3, 1986 at 11:20 a.m., a Nuclear Steam Supply Shutoff System (NSSSS) isolation of the Reactor Water Cleanup (RWCU) system occurred. During the removal of the equipment block on the 'A'
filter demineralizer, a plant operator opened the 'A' filter demineralizer discharge iso-lation valve out of sequence.
The system was blocked to install pressure gauges on the
'A'
and 'B'
Clean-up Filter Demineralizer outlets to trou-bleshoot the system because of flow problems that occurred i
since the last outage. After the pressure gages were in-stalled a nonlicensed operator was verbally instructed by a test engineer on returning the system to operation. The operator opened outlet valve 31A to the filter demineraliz-er, prior to closing vent valves 1027A and 1028A, creating a backflow from the 'B' filter demineralizer through the vent line.
The backflow was greater than 54.9 gpm and ex-isted for longer than 45 seconds which caused a RWCU system isolation.
The operator realized that the isolation oc-curred and immediately corrected the problem. The isola-tion was reset and the system was returned to service within one hour.
The inspector reviewed the licensee's corrective action of counseling the personnel involved on the importance of fol-lowing the equipment removal sequence and in recognizing an RWCU valving configuration which could result in a high differential flow condition. The inspector discussed the lack of written instructions which also contributed to the isolation. The inspector had no further questions, and identified no violations.
5.2.3 LER 86-043; Toxic Gas During the week of July 29, 1986, one of the two control room toxic gas monitors was declared inoperable after it erroneously indicated a high level of vinyl chloride. An inspection performed by the manufacturer's field service representative indicated that the high reading was caused by the presence of water vapor in the gas sample.
Both analyzers were subsequently declared inoperable pending further investigation.
The detector assemblies were removed from the plant and returned to the manufactur-er for diagnostic testing on August 6, 1986. The testing consisted of as-found tests to determine the analyzer's error attributable to the water vapor and carbon dioxide content of a sample.
The analyzers attempt to dynamically compensate for the water vapor and carbon dioxide contained in a sample.
Errors remain, however, due to the
.
.
non-linearity of the response of the analyzer to various concentrations of the compounds.
One analyzer was found to be slightly overcompensated for water vapor and carbon dioxide. This caused the analyzer to read zero or slightly negative for all ranges of water i
vapor and carbon dioxide. The other analyzer was found to be undercompensated for water vapor and carbon dioxide.
This caused the analyzer to falsely indicate an elevated vinyl chloride level. The manufacturer re-adjusted the water vapor and the carbon dioxide correction factors of the two anaryzers; both were made slightly undercompensated.
(As hig'n as four to five PPM during periods of high humidity and hi.gh carbon dioxide levels.) It was thought, however, that t'his undercompensation would also restrict non-conser-vative. errors during periods of low humidity and low carbon
,
dioxide levels to acceptable values. However, following L
licensee review of the test data, it was determined that the non-conservative error exhibited by the vinyl chloride
,
channel was still unacceptable. The analyzers were re-turned to the manufacturer on August 20, 1986 for recalibration. An alternate wavelength was used for the vinyl chloride channel. Analyzer performance was found to be improved to the point that carbon dioxide compensation is no longer required, and was therefore removed from the analyzers. While humidity compensation is still present, the errors from this compensation were concluded to be tol-erated within present setpoints.
Both detectors were re-turned to service on August 29.
The inspector discussed the calibration problems with plant management and I&C supervisors. A full-time I&C engineer was dedicated to technical difficulties associated with the toxic gas analyzers until the problem was solved.
Satis-factory performance has been evident through the inspection period, and the inspector had no further questions.
5.2.4 LER's86-047 and 049; RCIC and RWCU Isolations Caused by Testing The inspector reviewed the details associated with the above events and discussed these with responsible I&C su-pervision and test personnel. Although both events were of minimal consequences, I&C technicians involved in the iso-lations immediately recognized that an off-normal test con-dition had occurred and stopped testing to inform control room supervision.
I&C supervision also exhibited a tenden-cy to pursue root cause beyond immediate but not necessar-ily complete explanations.
In the case of the LER 86-049 RWCU isolation on October 9, an unrelated but loose relay was disturbed by flex conduit which was moved as a result
.
.
of the technician's attempt to place test leads at a limit-ed access location. This cause was identified after
-,
attempts to explain what seemed to be a personnel error were replaced by the licensee's decision.to recreate the event (in the presence of the Senior Resident Inspector) on October 21. The true cause was then determined to be a design problem and appropriate corrective action consisting of pursuing the need for a design modification was ini-tiated.
The inspector had no further questions.
5.3 Review of Periodic and Special Reports Periodic or special reports submitted by the licensee were reviewed by the inspector. The reports were reviewed to determine that the report included the required information, that test results and/or supporting information were consistent with design predictions and performance specifications, and whether any information in the report should be classified as an abnormal occurrence.
The following reports were reviewed:
Monthly operating report for September 1986
--
Monthly operating report for October 1986
--
Seismic Monitor Inoperability, dated September 29, 1986
--
These reports were found acceptable.
5.4 Part 21 Report - RPS Power Supply Breakers An evaluation of a potential defect with the Westinghouse Type LBB breakers used in the power supply to the RPS was completed on June 9, 1986 and a Part 21 Report was submitted to the NRC on June 16, 1986.
The licensee also made an entry of this generic problem on the INP0 Nuclear Network on June 10, 1986.
The inspector reviewed the Part 21 report, evaluated surveillance testing of the breakers, observed modifi-cations to the breaker design, and discussed the corrective actions with licensee Field Engineers, 5.4.1 Description On one occasion at Linerick and several instances at Peach Bottom, the shunt trip coils of the reactor protection sys-tem (RPS) power supply breakers were found to be damaged af ter apparent normal operation of the breakers. Licensee investigation determined that the failures were caused by the combination of misoperation of the breaker in the pres-ence of a minor manufacturing defec.
.
i After a successful trip of the breaker, the shunt trip coil can be damaged if attempts are made to reclose the breaker with a trip signal present. This damage will occur provid-ed that the breaker also has a manufacturing defect known as " insufficient spin-over". The spin-over process is per-formed to remove excess metal from the bottom of a rivet on the breaker handle post. The presence of the extra metal can cause binding between_the breaker handle post and breaker operating mechanism. Attempts to close an unaffect-ed (i.e. without insufficient spin-over) breaker with a trip signal still present apparently does not cause damage to the shunt trip coils.
5.4.2 New Breaker Installation and Modifications The four Unit 1 RPS supply breakers were replaced with new breakers. Utilizing a test procedure provided by Westinghouse, each of the replacement breakers was tested by licensee Field Engineering (prior to installation) to determine if the breakers had the spin-over binding prob-lem. The breakers successfully passed the test.
Additional modifications were made to the breaker shunt trip circuitry. A circuit interlock was added to each of the RPS breaker shunt trip circuits by wiring +he.second unused main contact of each breaker in series with its shunt trip coil and auxiliary contact. The additional in-terlock contact will open the shunt trip coil circuit when-ever the breaker is opened or tripped, assuring that the shunt trip coil is not energized beyond its momentary rat-ing. Neon indicating lamps which are used to determine if a shunt trip signal is present will be replaced with multi-light emitting diede lamps during the first refueling out-age currently scheduled for May 1987. These lamps have better visibility and will make it easier to determine if the trip signal is present.
Westinghouse informed the licensee that the breaker manu-facturing drawings had been revised to change the amount of spin-over on the handle post. Westinghouse is also now testing these breakers with the operating handle in the closed position to ensure that binding, which could cause shunt trip coil failures in this adverse operating condi-tion, does not exis6.
5.4.3 Modification Package 490 The inspector reviewed modification package (MDCP) number 490 to replace under and over-voltage relays and modify the breaker shunt trip circuits using a spare main contact as an additional circuit interlock. The protective relays and
.
.
circuit breakers were verified to be procured as Class 1E items.
The two poles of the breaker are electrically iso-
-lated and are physically separated by two barriers.
The safety evaluation for MDCP-490 described the modifica-tion as a replacement of the existing relay, with a new version of the relay and wiring the spare main contact of each breaker in series with its shunt trip coil. No exter-nal rework was performed, and there was concluded to be no effect on bus loading, or seismic / dynamic and environmental qualification of the relays. The new protective relays were installed in June 1986 but acceptance testing for the modification showed the relays to be sensitive to noise in the form of harmonics present in the inverter output being monitored. The methods of performing the 18-month calibra-tion ST and the acceptance test for MDCP-490 were changed to require use of the actual source or inverter output.
5.4.4 Field Engineering Field Report The inspector reviewed the Field Engineering Group report that updated the status of MDCP 490 and discussed actions taken as a result of the modification acceptance test being unsatisfactory. Currently, the A RPS channel supply is being fed from the normal UPS inverter; but, due to vari-ability in waveforms, the B channel is fed from the alter-nate TSC inverter source.
Satisfactory results have been obtained from July 1986 testing of the new undervoltage 27N relay with the harmonic filter module installed. This relay demonstrated a consis-tent response to the inverter, TSC and utility output volt-ages.
Therefore, the licensee concluded that established Engineering and Research calibration tests, procedures, and operational ST's should continue to be the basis for verifi-cation of the proper response and operation of the new under and over-voltage relays.
Revision 3 to MDCP-490 was being prepared during the inspec-tion period to procure additional under and over voltage relays to be installed at a future date with harmonic fil-ter modules.
5.4.5 Conclusions The inspector concluded that the Westinghouse circuit breakers in the RPS supply power were adequate to reliably trip at voltages and frequencies that might otherwise damage the scram solenoids.
The inspector found that the trouble-shooting and expertise demonstrated by the licensee Field
.
.
Engineering was effective and added to the assurance of a reliable power source to the RPS.
No unacceptable conditions were noted.
6.0 Surveillance Activities 6.1 Test Observations The inspector observed the performance of and/or reviewed the results of the following tests:
ST-6-092-311-1; D11 Diesel Generator Operability Test Run, con-
--
ducted October 14, 1986
--
ST-1-LLR-651-1; Suppression Pool Leakage Rate Test Spray, con-ducted May 31, 1986 ST-2-042-660-1; Reactor Vessel Water Level Channel Functional
--
Test, conducted October 9, 1986
--
ST-3-107-790, Control Rod Scram Timing, conducted on October 26, 1986 The tests were observed to determine that test procedures conformed to technical specification requirements; proper administrative con-trols and tagouts were obtained prior to testing; testing was per-formed by qualified personnel in accordance with approved procedures and calibrated instrumentation; test data and results were accurate and in accordance with technical specifications; and equipment was properly returned to service following testing.
No unacceptable conditions were noted.
6.2 Turbine Stop Valve Testing During performance of surveillance ST-6-001-760-1, on October 13, 1986, the main stop valve /CIV exercise test, the four main stop valves failed to move upon depression of their respective test buttons.
The problem was traced to a fault in the cable which provides a 24 VDC supply to the test pushbuttons. The stop valves were successfully stroked later on October 13 under a temporary procedure change to the ST. As a short term solution, a Temporary Circuit Alteration (TCA) was installed on October 17 to allow further stop valve testing.
Cable repairs were deferred until the next outage.
The failures were with an open connection in the two Canon-type con-nectors between the test panel in the main control and auxiliary equipment rooms. The test was completed within its required interval using the TCA to install a temporary pushbutton. The inspector ob-served portions of the troubleshooting and the subsequently success-ful stroking of all four stop valves during October 13-14. Also, the
.
-
.
inspector reviewed logic drawings and discussed the problem with licensee test engineers, verifying that the open connection was limited to the test circuitry only and did not affect the ability of the valves to perform their safety function.
No violations were identified.
7.0 Maintenance 7.1 Target Rock SRVs Setpoint drifts were experienced with main steam. safety relief valves (SRVs) tested at Wyle Laboratory during the periad of May 12-29, 1986 with Unit 1 shut down and in the process of an outage. The valves had been installed since initial fuel load in October 1984 and for a total of 161 effective full power days of Unit 1 operation. Ten of the 14 valves failed to initially lift within 1% of the set pressures speci-fied by technical specifications. The test failures were similar to other two-stage Target Rock SRV setpoint drift problems described in NRC Information Notice 86-12.
NRC Region I Inspection Report Nos.
50-352/86-09 and 86-11 have documented the results of the Limerick test failures as well as the storage and installation of replacement SRVs.
With respect to the reportability under 10 CFR 50.73, the licensee concluded that the SRV test failures did not meet the criterion of a condition prohibited by the Limerick Technical Specifications, spe-cifically because no violation of the technical specification oc-curred. The licensee analysis of the Limerick SRV test failures also concluded that other criteria of Part 50.73 were not met. Specifical-ly, the licensee evaluated the safety implications of operating Unit 1 at the higher than expected SRV opening pressures with respect to the following:
-
pressure margin to the ASME Code limit of 1375 psig fuel thermal margins (MCPR)
-
ECCS performance
-
drywell temperature response
-
-
HPCI and RCIC performance The licensee's Plant Operating Review Committee (PORC) met during the first week in June 1986 to review a GE analysis of the SRV setpoint drift. The GE assessment was documented on May 28, 1986, and in-volved a bounding analysis assuming all 14 SRVs opening at 10% above their nominal setpoints.
The PORC concurred with the GE assessment that peak vessel pressure for the most severe postulated overpressure transient was below the upset limit of 1375 psig, and that a drift of up to 10% above nominal had no safety impact on plant performance.
The licensee's evaluation was completed approximately 20 days after being informed of the failures by Wyle Labs. The licensee concluded that the setpoint drift did not meet the 10 CFR 50.73 reportability
.
.
requirements in that the magnitude of the setpoint drift did not prevent:
safe reactor shutd)wn
-
residual heat remo al
-
control of radioactive material releases
-
accident mitigation
-
nor was the plant determined to be in an unanalyzed condition.
The inspector had no further questions and concluded that no viola-tions of Part 73 reporting requirements had occurred.
7.2 HPCI Steam Supply Valve On October 17, 1986, at 10:00 a.m. shift supervision was notified by a plant operator that he had observed debris on the floor in the area of the HPCI steam supply outboard isolation valve HV-55-IF003. Upon investigation it was discovered that the valve's packing was dis-lodged and the packing follower gland clamp was broken off of the valve. Once all of the packing was dislodged, the valve back-seated itself; therefore, no steam leakage occurred.
7.2.1 Immediate Actions The HPCI inboard isolation valve HV-55-IF002 was de-energized in the closed position and HPCI declared inop-erable at 11:56 a.m. in compliance with technical specifi-cations. A Maintenance Request Form (MRF) 8606188 was issued to work on the valve. Once the valve operator was removed it was discovered that the follower gland had seized to the valve stem and heavy scoring was noted on the valve stem.
7.2.2 Suspected Cause The licensee suspects that either a foreign object was lodged between the packing and valve stem or there was a defect in the packing. A question arose because scoring of the stem occurred in the vertical plane and the yoke legs are in the horizontal plane.
The static weight of the overhung operator combined with system operation may have been sufficient to permit contact of the stem and gland.
Based on the Seismic Analysis Report the valve was analyzed and qualified in its worst position; that in which it is presently installed. The valve was qualified to 5.8 g of acceleration and met the operability requirements. The piping system is rigid and did not allow excessive valve operator movement.
_,
- - - - -. - -
.
_-
.
.
7.2.3 Repairs and LLRT On October 22 the repairs on the valve were completed and at 9:45 p.m. a local leak rate test (LLRT) was performed on the valve in the presence of the resident inspector. The test was conducted in accordance with an approved procedure with acceptable results.
The inspector observed that the test personnel were familiar with the test equipment and knowledgeable in the performance of the test. The recorded leakage for the valve was 58.1 sccm which cause the total for the penetration to increase from 191.83 sccm to 249.93 sccm still within the acceptable limit. On October 23, 1986 the HPCI system was declared operable at 9:00 p.m.
7.2.4 Conclusions Maintenance performed on the HPCI valve was well planned and tnoroughly evaluated.
Strong engineering support with-in the Maintenance Division and with corporate engineering personnel was evident.
No unacceptable conditions were identified.
8.0 Recirculation - Pump Trip Breakers The design of the recirculation pump trip (RPT) utilized to mitigate an anticipated transient without scram (ATWS) at Limerick was reviewed for breaker operability and reliability.
8.1 Design The Unit 1 RPT is an energized-to-operate system with logic arranged so that two out-of-two taken once will trip the 4160 VAC breaker. The breaker has a dual trip coil that trips the MG set generator on high pressure at 1193 psig or low reactor water level at minus 38 inches.
The trip circuits are powered from the Division 1 & 2 125VDC power source.
The level and pressure sensors are different from those used in the reactor protection system; however, they are environmentally quali-fied and maintained on the licensee's Q-list. The logic and control for RPT can be tested on-line. The breakers are tested when the re-actor is shutdown or during a refueling outage. The selection of trip i
set points, training instructions and the procedures are such that inadvertent actuations will be minimized.
Station Trip Procedure T-101 provides the necessary guidance for op-erators in the event the recirculating pump fails to trip on the high pressure or low water level trip signals.
The emergency operating procedures instruct the operator to manually scram the reactor and trip the recirculating pumps. The transient resulting from tripping of the recirculation pumps is analyzed in the FSAR, and consequences
.
.
of these transients are well within limiting transient analyses. The RPT system design at Limerick 1 is consistent with applicable re-quirements of 10 CFR 50.62 (ATWS Rule). This design and operating procedure is based on guidance developed by General Electric and ap-proved by the NRC.
8.2 Use of GE-AK-F-2-25 Breakers The licensee does not use the GE-AK-F-2-25 breakers for the RPT func-tion. This type of GE breaker is used on Unit 1, but no reports of failure have been made since the turnover of Unit 1 to the operating
staff. During the preoperational testing phase, the GE-AK-F-2-25 breakers experienced problems, but were corrected before turnover.
The data for that phase of the test program is not part of the CHAMPS system and was not readily available at the time of the inspection.
8.3 Surveillance The RPT surveillance test requirement is listed in the CHAMPS system and is performed per procedures as required in letters to the NRC dated March 31, 1980 and October 17, 1985. The frequency of system testing is once every refueling cycle. The surveillance of the in-strumentation and logic systems is performed quarterly. There have been no RPT failures of sensors, transmitters or breakers in this system to-date.
8.4 Maintenance The licensee has a formal maintenance program which includes both corrective / preventive maintenance. The scheduling, documentation and history of the safety related systems re documented in the Component History and Maintenance Program System (CHAMPS).
Component hi. tory, corrective and preventive maintenance and quality control results are documented in the CHAMPS.
The licensee has in-stituted the requirements for trending data within the CHAMPS and information-supplied by vendors. The program for trending was start-ed approximately 3 to 4 months ago.
Preventive maintenance is performed on the components of the RPT sys-tem during the refueling cycle.
8.5 Design Reliability The single failure criteria for the RPT is satisfied in that the sen-sors, transmitters, sensing lines, power supplies, cable and trip coils are redundant. The ATWS breaker and trip coil of the Brown Boveri 5HK250 breaker actuates on level and pressure signals from four trip ch:nnels.
-
-
.,
. - -
,. _
.
-
- - -. - - _ -.
- - _ _ _ _ _ - _ _ _ _,
..
.
An approved quality a surance program in conformance with 10 CFR 50, Appendix B was appliea to the RPT design including the 4160 volt breakers. The reactor recirculation pump and associated motor generator set are not safety related.
All components used to trip the recirculation pumps are independent and separate from components that provide the reactor protection function. Although the sensors for both the RPT and the RPS are lo-cated on the same instrumentation racks, there are redundant racks with the sensors arranged so that a failure of a rack or of an in-strument sensing line will not prevent a RPT.
The failure rate for the RPT system is defined in the Limerick Probabilistic Risk Assess-ment as 1.0E(-4) per demand.
No violations were identified.
9.0 General Electric HGA Relays 9.1 Background The resident inspector reviewed the application of GE Type HGA relays utilized in safety-related applications at Limerick. The concern is for HGA relay seismic qualification in Class 1E electrical systems, specifically contact chatter previously experienced during seismic testing at Wyle test labs.
9.2 Design Controls The inspector discussed this potential problem with PECO Field Engi-neering. The licensee had been notified by the Limerick Architect Engineer (Bechtel Power Corp.) in March 1973 that seismic analysis of auxiliary relays showed a better seismic capability for types other than HGA. GE test results obtained in late 1972 showed the following seismic qualification limits:
HMA Relay 3.0 g HGA Relay 0.9 g Although the licensee agreed (at that time) to the use of a smaller type HMA relay to meet seismic Class I requirements, the two pole HGA relay was retained for non-safety related applications.
Eventu-ally, another type (HFA) relay was also utilized for larger contact applications, since the HFA relay incorporates six contacts.
9.3 Findings The inspector independently reviewed electrical schematics and in-spected selected Unit 1 safety related panels and cabinets to verify that GE type HGA relays were not installed.
ECCS panels for all four
o
.
Divisions of RHR and Core Spray, located in the Auxiliary Equipment Room, were observed to have HMA relays for the pump start /stop cir-cuitry and an HFA relay for the automatic initiation signals associ-ated with the diesel generator /ESW/4kV bus logic. The back panel cabinets were opened and verified to have no HGA relays installed.
The auxiliary relays used for internal logic circuitry are an Agastat Model EGPD which is smaller and uses four contacts (versus the two-pole, larger HGA).
The inspector also walked down all Unit 2 panels and cabinets in the Auxiliary Equipment Room and found no GE HGA relays installed in Class IE cabinets or safety related circuits.
The RPS cabinets for Unit 2 (similar in design to Unit 1) employ Potter-Brumfield Series MDR relays.
The inspector reviewed electrical schematic diagrams for the ESW (drawings E321) and RHRSW (drawings E-361) systems which depicted the use of Agastat Model EGPO coils for breaker and logic circuitry. The 4kV safeguards switchgear panel 10A115 for emergency diesel generator Dil, ESW and RHRSW systems, and the 201 Safeguards bus were opened and inspccted. No GE relays of any type were found, and the in-stalled Agastat coils were verified to be in place.
Finally, the inspector observed the installed equipment in local diesel generator cubicle panels IAC-514 and 1ETB-AG501 for the D11 engine. Auxiliary and protective relaying employed were of the Agastat and Square-D type, and no GE HGA devices were found.
9.4 Limerick Seismic Design The Limerick site Safe Shutdown Earthquake (SSE) is defined to be
"zero point" accelerations of 0.15 g horizontally and 0.10 g verti-cally. Based on ground response spectra and Control Enclosure ampli-fication and damping, the required response of the Class 1E relays installed in the auxiliary equipment room at elevation 283 ft, was determined by testing. These data were provided by licensee Mechan-ical Engineering representatives from the Limerick Dynamic Control Room Panel Seismic (SQRT) report, and compared against GE fragility limits as follows:
Acceleration (in g's)
Required HFA Relay HMA Relay front-to-back 3.09 4.9 4.0 side-to-side 1.48 11.0 11.0 vertical 1.01 7.5 7.5 As seen from these data, the HMA and HFA relays installed in ECCS auxiliary equipment room panels on Unit I have been verified by test to be capable of withstanding the expected seismic effects at their installed location o
..
9.5 Conclusions The inspector concluded that the seismic qualification questions as-sociated with GE type HGA relays were identifir.d early (1973) in the design of Unit 1, and appropriate controls were exercised to preclude their installation in Class IE safety-related electrical systems. No violations were found, nor were any additional concerns identified by the inspector.
10.0 Exit Meeting The NRC resident inspector discussed the issued in this report throughout the inspection period, and summarized the findings at an exit meeting held with Mr. John Franz ana others of his staff on November 18 and again on November 26, 1986. At these meetings, the licensee's representatives indi-cated that the items discussed in this report did not involve proprietary information.
l