ML20083C405
| ML20083C405 | |
| Person / Time | |
|---|---|
| Site: | Brunswick  |
| Issue date: | 09/20/1991 |
| From: | Spencer J CAROLINA POWER & LIGHT CO. |
| To: | NRC OFFICE OF INFORMATION RESOURCES MANAGEMENT (IRM) |
| References | |
| NUDOCS 9109260184 | |
| Download: ML20083C405 (17) | |
Text
,
l l
Cp&it Carolina Power & Light Company Ismmensraanmusamen:a Brunswick Nuclear Project P. O. Box 10429 Southport, N.C.
28461-0429
$EP i' O kW FILE:
B09-13510C 10CFR2.201 U.S. Nuclear Regulatory Conunission ATTN:
Document Control Desk Washington, D. C. 20555 BRUNSWICK STEAM ELECTRIC PIRIT UNITS 1 AND 2 DOCKET NOS. 50-325 AND 50-324 LICENSE NOS. DPR-71 AND DPR-62 REPLY TO A_tiOTICE OF V101.ATIOf!
Centlemen:
The Brunswick Steam Electric Plant (3SEP) has received NRC Inspection Report 50-325/91-09 and 50-324/91-09 and finds that it does not contain information of a proprietary nature. This report included a No' '.c Of Violation.
Enclosed is Carolina Power & Light Company's response to that Notice of Violatioti.
Very truly yours.
$$]cv J-(&v i
J,. W. Spence General Managet Brunswick Nt/ i 'r Project SFT/
Enclosure ec:
Mr. S. D. Ebneter Mr. N. B. Le BSEP NRC Resident Office i
kD$$4sjgggg
/
Pf)g
ATTACllMENT 1 j
DRUNSWICK STEAM ELECTRIC PLANT UNITS 1 AND 2 DOCKET NOS. 50-325 AND 50-324 LICENSE NOS. DPR-71 AND DPR-62 REPLY TO A NOTICE OF VIQIJG'lQ1{
ylOIATION 10 CFR Part 50, Appendix B,
Critorion XVI, Correctivo Actions, requires in part, that measures be established to assure that conditions adverse to quality, such as failure, malfunctions, deficiencies, deviations defectivo material and equipment,.and nonconformancos are promptly identified and corrected.
i Contrary to the above, measures were not adequately catablished to promptly identify and correct deficiencies with the Emergency Diosol Generators (EDG) and Electrical Distributinn System ( E['9 )
equipment.as evidenced by the following exampics:
1.
On April 27, 1991, the MG #2 output breaker failed to shut on demand during a
routino test.-
Based on
' manufacturer information it was determined that the failure resulted from improper initial assembly of the breaker.
The licensco was informed by the nanufacturer that the affected mechanism was not visible for inspection, without extensive breaker disassembly and was not of a potential generic concern.
Ilowever, during field inspection by the NRC on May 3, 1991, it was noted on a nimilar breaker that the mechanism in question was visually accessible without disassembly.
It was also noted that this breaker exhibited the same type of defect as the failed breaker.
While certain corrective actions were
- taken, adequato correct >ve measures were not complete.
P.
During an EDG #3 test run on April 14, 1991, an air leak was noted by the NRC on the flange where the Jet Assist piping attaches to the turbocharger.
This leak had a trouble ticket attached dated December 24, 1990.
Measures to promptly identify and correct this item had not boon accomplished by the licensee.
It was also noted that a formal evaluation of the leak's affect on the operability of the Jet Assist feature had not boon adequately accomp} i Aed as of May 3, 1991.
The 1icensee had also not determined the cause of the leak or its possiblo response iluring a seismic event.
3.
The EDG #3 ECCS load test for Unit 2 f ailed due to a slow start on April 15, 1991.
There were previous tests that also faijed due to slow starts.
Corrective actions to correcu these p. oblems had been focused on tuning the air start system.
After this failure, adjustment of the air
e,
- e ATTACilMENT 1.
2 start system did not correct the problem.
l'u rthe r ovaluation of the protlom identified the degradation of the EDG shaft driven fuel oil pump.
This was masked by the test methods usod to dotormino oporability of tho EDGs.
The defectivo equipment was not promptly identified or corrected.
4.
Tho lubo oil cooler on the EDG #4 would not maintain the lubo oil temperature below the operating limit.
Similar problems were noted on the other EDG's.
In August 1990, a technical support memo was written, which doloted all the work tickets related to those problems. The problems with the lubo oil coolers were not promptly corrected.
It was also noted that an evaluation of the oil coolor's affect on the operability of the EDG had not boon performed as of April 11, 1991.
S.
The NRC team witnessed the EDG #3 test run on April 23, 1991, and noted that tho #6L cylinder was tapping loudly and the delta T on the turbocharger intercooler at 100%
load indicated 0 degrees F.
Those items voro noted to the operator by one of the NRC inspectors observing the performance of tilo test. These items were not documented following the performance of the test and potential advorno conditions were not identified or corrected.
6.
Along with the generator brush f allere on EDG #3 on March 18, 1991, sovoral problems, i.e., brushes oscillating on coastdown,10 loud knocks f rom eng3 no and fluctuations of 20-30 amps woro noted by the licensco.
Those problems were not properly evaluated or
'm cause identified prior to returning the EDG to operab; statun.
7.
The licensoo had identifiod that the relays for the Jet Assist feature on the EDG's were not in their preventive maintenance progran for calibration in February 1989.
They had planned to have the prococuros developed to calibrate the relays in September 1991.
This item indicaten that measures were not established to promptly identify and correct deficiencios to the calibration of components on a safety feature.
This feature of tho EDG had not been tested or maintained since about 1974.
This is a Severity Level IV violation (Supplomont 1).
Retorence:
Report Nos. 50-325/91-09 and 50-3?4/91-09
l ATTACHMENT 1 3
RESPONSE
I.
Admissipp of the Violation CP&L maintains the position presented during the August 2,
1991 Enforcement Conference and denies the applicability of examples #1, #3 and #6 identified above to the violation as stated. A summary of CP&L's position is provided in Attachment 2 for each of these examples.
Howaver, CP&L admits that the violation c.s stated for the remainder of the examples did occur.
Raasons for the violation, corrective actions past and future to ensure compliance are discussed below.
Additionally, CP&L is concerned about the management controls that permitted these conditions and agrees that actions to improve the effectiveness of the work control processes are necessary to prevent recurrence.
II.
Reason foJ the ViolM ion The incident noted in example 2 occurred as a result of personnel error.
The incident is discussed below.
On 12/24/90 during a test run of Emergency Diesel Generator (EDG)
- 3, the flange where the jet asnist piping attaches to the turbocharger was noted leaking air. The Technical Support sub-unit manager observed the leak and determined the leak was of a small magnitude and the probability of Icakage increase was minimal.
The cause of the leak at this time was considered gasket degradation.
Additionally, other diesel operating parameters noted during that run of the diesel indicated diesel performance was not degraded as a result of the leak.
The Technical Support sub-unit manager generated a corrective maintenance work order (WRJO) to support prompt identification of the concern. Diesel operability assessment was documented on the WRJO identifying no operability concorps. The priority assigned to the WRJO was consistent with the requirements of site procedures governing WRJO prioritization.
In order to minimize EDG out of service time, corrective maintenance for problems not affecting operability is consolidated and performed during planned system outages.
a ATTACHMENT 1 4
On 4/14/91 following NHC observanco of a subsequent EDG
- 3 test run, maintenanco activitieu were initiated to correct the problem. During flango disancembly a noction of the flange ganket was noted missing and one of the six helicolla partially backed out of its flange mounting bolt hole.
Inspection rf the holicoil revonled that the helicoil threads woro galled.
The galled threads woro repaired and the gasket replaced.
During thread inspection and
- repair, maintenanco personnel noted ruinalignment of the pipe to fle.nge and initiated another WRJO to support pipe to flange ron.1 ig nme nt.
The pipo misalignment WRJO was connicored to be a non-operability concern and acceptable for %ork ac scheduling allowed.
On 4/15/91 during an EDG #3 tent run, post maintenance touting revealed the Jet Ano: at flango leak repair et fort did not fully correct the air leak.
The leak, though utill present, appeared no more sigr tricant than before the repair ef fort and the original op( ability accenomont basis was considcred acceptablo.
E.DG #3 was declared operable at this timo.
Maintenance initiated a requent to the Technical Support group to support evaluation of the misaligned pipo problem.
On 4/k./91 enginotring ovaluation (EER) 91-0144 was initiated to addreas the repair of the misaligned flange piping.
On 5/3/91 EER 91-0144 was approved.
On 5/4/91 Maintenance implemented the corrective actions prescribed within EER 91-0144 for replacement of the helicoil identified an backed out during initial repair efforts and flange pipo misalignment.
On 5/5/91 post maintos.ance testing verified operah111ty of the Jet Air Assist System and EDG #3 was restored to an operable condition.
During the initial repair effort, pipe misalignment causing one of the six flange mounting bolta to pull away from the flango was noted; however, Technical Support personnel supporting the repair effort did not recognize the coicmic implications at thic time.
Connequently, a scismic evaluation supporting Jet Ansist operability with five of the six flange mounting bolts functional was not addressed prior to returning the diesel to serv ice on 4/15/91. On 5/16/91 EER 91-0190 was prepared to evaluato the acceptability of the five of six bolt condition.
This ovaluation determined that.
operability concern existed an a result of the five of alx holt condition.
An investigation to determine the cause of the seinmic issue identified that Tt ch.aical Support personnel, though aware of the need fo:' a seismic evaluation in such circumstances, inadvertantly overlooked the seismic isnue due to the number of !; sues being addressed during the ongoing dual unit outs.go.
)
~
g a
f ATTACHMENT 1 5
The incident noted in example 4 occurred as a result of personnel orror.
This incident is discussed below.
Operating procedures have historically allowed operator manual control of the diosol lubo oil tamporaturo control valves to maintain temperatures within acceptablo limits and require maintenanco work ordor (WRJO) gonoration upon initiation of manual control.
On 7/41/90 - the Maintenance organization generated a Technical Support Memorandum (TSM) requesting ongineering ovaluation of lubo oil temperaturc control operation in an effort to possibly reduce tho number of WRJOs generated. On 4/6/91 l
a WRJO was initiated identifying the failure of the-EDG
- 4 lubo oil control valve to maintain temperature loss than the normal operating range upper limit of 170 degrees Fahrenheit.
On 4/8/91 Maintenanco planning personnel doloted the existing WRJOs related to lubo oil i
temperaturo control valvo concerns based on initiation of the Technical Support Memorandum referenced above.
On 4/19/91 an engineering assessment documented within TSM 91-0361 addressed acceptable lubo oil temperaturo ranges and confirmed the vendor specifiod maximur contitiuous lubo oil temperature limit of 187 degrecs Fahrenheit. O n 4/23/91 Maintenanco completed replacement of the EDC #4 lubo oil temperature control valvo poppet valvo assembly.
On 6/25/91 the Emergency Diosol Generator Operating Procedure OP-39 was revised to identify proper acceptable lubo oil operating ranges and appropriate oporator actions.
On 8/1/91 Maintenanco Policy Notico (MPN)91-008 was generated describing the propor methodology for deleting WRJOs based on completed enginocring evaluation.
CP&L recognizes that more timely action should have boon taken to address the lubo oil totuperature control valve concern and that manual control. capability is only acceptable as an interim measure.
The procedural controls fo; ensuring WRJO initiation upon exercise of manual control woro established to support timely corrective action.
Failure to pursue timely correction of the lubo oil temperature control valves and a lack of urgency in rostoring equipment to the original design is the root cause of this incident.
i
=. * -
- mm -- a.
u
.-A--. A
.-m
ATTACHMENT 1 6
Additionally, personnel involved with the planning of correctivo maintenance work orders failed to recognito that WRJO deletion based on an unronolved TSM in not an adequate work control practicc. Consequently, the lube oil temperature control valvo related WRJOn woro doloted without proper resolution.
Finally, Cp&L considers the actions taken by Operations on 4/6/91 to datormino the Japact to dionol operability adequato.
Operationa personnel considered the inability to manually control lubo oil temperaturo to a valuo less t.han the maximum operating limit insignificant sinco the annunciator cotpoint of 100 degroes Fahrenhoit had not boon reached.
Consequently, the potential for diouol equipment degradation reculting from high lubo oli temperature did not exist.
The incident noted in examplo #5 occurred no a result of personnel error.
The incident in discussed below.
On - 4/23/91 during performance of EDG
- 2 test
- run, Maintenanco pornonnel observed and documented 6L cylinder knocking.
The knocking initially observed subsided during the run.
Review of the diesel generator log maintained by the Maintenance Sub-unit Manager identified that the cylindor noise condition was not considered a significant problem at that timo and did not rondor the diosol inoperable.
Since the knocking subsided during the ran, the Maintenanco Manager suspected the knocking was the renult of vil having been drained from the system during recont Maintenance activities.
Additional log notoa indicato-that the noino would be rochocked on the next run of-the diosol and a WRJO initiated if nooded.
On 4/25/91 during - the next run of the diosol, the 6L cylinder noise was again observed by Maintenance.
The dociolon to repair the suspect lifter was mado and a corrective maintenance work order generated.
The 6L cylinder was replaced on 4/28/91.
On 4/23/91 during the EDG #2 test run referenced above, the NRC obnorved turbocharger intercoolar delta T of 0 degrees Fahrenheit at 100% load condition.
Efforts to identify who the NRC informed of the turbocharger condition have proved unsuccessful and consequently, why the condition was not documonted is not teta11y understood.
Turbocharger intercoolor-delta T
was r'onitored during the next runs of EDG #2 and #4 occurring 4/26 and 4/28/91 respectively. The observed temperatures were noted consistent with original factory acceptance tcat data.
On 4/30/91 a Technical Support Memorandum (TSM) 91-0369 was generated to evaluato intercaoler
ATTACl! MENT 1 7
performance and provido verification of all four diesel turbocharger intercooler operabilitica.
Actions to document and correct the EDG #2 6L, nylinder noise problem are considered timely and adequato and no further corrective actions are required.
- Ilowever, Operations pornonnel involved with performance of the 4/23/91 EDG #2 test run failed to identify and pursue correction of the turbocharger intercooler delta T
concern.
Though subcequent testing and evaluation verified acceptable intercooler performance, moacuren to assess the potential adverno' condition were not taken immediately.
The incident noted in examplo #7 occurred an a result of 1
personnel error.
The incident in discussed below, Up until February of 1989 the diosol generator Jot Anaint relays were not specifically identified as safety related and woro not being calibrated on a periodic bacia.
However, como of the relays had been calibrated an a ronult of corrective maintenance activit.co.
In February of 1989 the subject relaya woro specifically identified ao safety related.
In March of 1990 Maintenanco aclf-initiated a
Nonconformanco Report (NCR)
S-90-011 identifyjng the nood to ensure that all safety related components were included in the preventive maintenance program.
In October of 1990 in an ef fort to resolve the concerns addressed by NCR S-90-011, Maintenanco completed review of preventive maintenance requirements fo;' cafety related components. As a result of this review, the need to include the Jet Assist relays in the proventive maintenance program was identified and a calibration scheduled for February of 1991.
In February of 1991 in preparation for the scheduled testing of the Jet Aaalst relays, Maintenance pornonnel determined that the generic proceduro specified for calibration of the Jet Accist relaya was inappropriato for testing those specific relays and submitted a procedure reviador request to the Maintenanco Procedures group.
The duo dato for completion of the procedure revision was consistent with the next refuel outage.
On 5/1/91 due to concerna identified by the EDSFI, a special process procedure was developed to test the Jot Air Assist 1X LOCA logic for all diesels. On S/2/91 test results identified that some l
of the Jet Analat relay setpoints were not within i
calibration tolerances.
The relays were then calibrated withjn acceptable calibration limits.
On 5/3/91 an engineerina ovaluation (EER) 91-0151 was generated to ovaluate operability of the Jet Issaist feature and justify acceptance of the 2X LOCA logic until procedures l
ATTACilMENT 1 0
could be entabliahod and performed during the next Unit 1-and Unit 2 refuel outagos.
The ovaluation verified that although como of the relays woro out of calibration toleranco, the EDGa would have performed their intended safety function.
In summary, controls to oneuro Maintenanco awareness of component safety classification changes woro not adequato until resolution of the concerna addressed within NCR S-90-011 in October of 1990.
NCR correctivo actions included the notablishment of controls within existing procedures for informing Maintenanco of component safety
- clannificatlun changos requiring addition or dolotion to the proventivo maintenanco program.
Ilowevor, additional controla are nooded to onouro timely componont
)
calibration onco a nood for proventivo maintenance of anfoty related componenta la nooded.
CP&L considera the failuro to recognize the nood for and implomont auch controla the root cauno of this incident.
III. Corrective Action:LHhlglLilave Boon Taken Examplo #2 Efforts to resolve the Jot Assist flango leak including i
replacement of tho flange gaoket and helicoil, repair of the misaligned piping and ovaluation of soiamic concerno as related to operability have boon completed and EDG #3 reatored to an operable condition. Technical Support hau completed a review of the solomic lanuo addressed within Adverso Condition Report (ACR)91-243 and has datormined that-training is nooded __to reinforce the nood for Technical Support personnel to stay alert to situations impacting equipment solamic analyala and n'milar j aauos.
This training was completed on 9/3/91.
Examplc #4 Efforta to resolve the EDG
- 4 lubo oil temperature control valvo problema including engincoring nonosoment of acceptable temperaturo-rangen, arecification of lubo oil continuous maximum temperature limit, revision of-operating proceduros, and the development of specific interim guidance governing the WRJO dolotion procosc have boon completod.
Those changea woro implomonted to increano confidence in tho ability of the dicaolo to perform properly without manual intervention. Exporlonco since implementation of the aforemontioned actions has provided the evidence that EDGs #1 and #2 lubo oil-temperaturo contral_ _la accomplished without manual intervention.
Additional correctivo maintenance la l
f n
y
,,,..-vv-e--
,,c-,-- - -,,,,.. _, _,...,,,,,
a
e I.TTACHMENT 1 9
required to provide hands off control of EDGs #3 and #4 as identified within_Soction IV. of Attachment 1.
Example #5 Efforts to evaluate performance of EDG #2 turbocharger intercooler have been completed.
Diesel test runs performed following the incident verified proper turbocharger performance _ and no additional corrective actions were required.
i Example #7 l
Testing of the single Loss of Coolant Accident (LOCA) Jet I
Assist logic and calibration of the Jet Assist logic relays has been completed for all four diesels.
Additionally, testing of EDGs #3 and #4. to verify the capability of the Jet Assist' function to maintLin frequency within acceptable limits under emergency loud conditions has been performed satisfactorily.
Similar testing of-EDGs #1 and #2 is scheduled to occur during the next Unit 1 outage.
An engineering evaluation was generated to assess acceptability of the 2X LOCA logic until procedures can be developed and performed.
A review of the preventive maintenance program to support identification of safety related components which have not been calibrated has been completed.
The results of this review indicate 'that of the safety related components included within the preventive maintenance program, twelve (12) have not been calibrated tr aate.
An assessment has been performed-to determ,i rc.: _ the operability of the equipment and systems affected by those uncalibrated components.
Based on the results of this assessment no operability concerns exist.
These components -will be calibrated as required by the preventive maintenance schedule.
IV.
Corrective Actions To Be Taken and Date of Full Compliance Example #2 CP&L considers the actions taken to date to resolve the Jat Assist flange leak adequate for ensuring diesel operability and ensuring sensitivity to the need for future ceiamic evaluation.
ATTACHMENT 1 10 Example #4 CP&L considers the actions taken to date adequate for ensuring reliable auton:atic op3 ration of EDGs #1 and #2 lube oil temperature control systems within normal operating limits.
- However, corrective maintenance involving disassembly, inspection, rebuild and repair if needed and calibration of EDG #3 jacket water intercooler temperature control valvo, jacket water heat exchanger to sump bypass valve, lube oil temperature control valve and EDG #4 Jacket water intercooler temperature control valve and jacket water heat exchanger to sump bypass valve is needed to provide hands off operation.
EDGs #3 and #4 lube oil and jacket water temperature control valve rebuilds shall be completed by 12/4/91 pending part availability.
Additionally,
. Maintenance procedure Corrective Maintenance (Automated Maintenance Management System)
OMMM-003 will be revised by 1-15-92 to support incorporation of Wluo deletion criteria established within the interim measure MPN 91-008.
Example #5 CP&L considers the established programmatic controls and Cporator training adequate for ensuring prompt identification end correction of adverse conditions.
However, reinforcement of the need to identify, document and pursue correction of potential adverse conditions is needed.
Training of Operations personnel including a review of the event and the need to address all concerns whether witnessed by the individual or relayed by other personnel shall oc implemented by 1/15/92.
Example #7 CP&L considers the established controls for cr.suring identification of safety related component classification changes adequate.
- However, additional controls are needed to ensure timely calibration of components which have been upgraded from non-safety to-a safety related classification.
The Equipment Data Base System (EDBS) group-will review the preventive maintenance prcgram calibration history during the component upgrade process for the purpose of identifying components needing evaluation of calibration requirements.
EDBS review of existing controls and implementation of additional controls if needed will be completed by 1/15/92.
.g ATTACHMENT 1 11 In summary, Cp&L concurs with the NRC concern regarding tho nood for management controls designed to improve offectivonuss of the work control process.
Brunswick management directivos, soveral which woro previously implementod, woro dovt loped for the purpose of improving paraonnel attention to
- detail, human rollability, owner-hip and communication skills.
Increased supervisory personnel in-fiold timo is an examplo of a recont directivo developod by management in which bonofits to increased personnel accountability, improved timoliness of corrective actions and prompt identification of adverso conditions are expected.
Recognizing that a mature 'culturo' in which essentially each employeo understands, adopts and employs the philosophics necessary to consistently achievo performanco excellence requires management reinforcement over time, Brunswick Nuclear plant management has and continues to reinforco existing human performanco improvement directivos and actions, and monitor their results.
'fhrough continued monitoring of pcraonnol performance and reinforcement of existing work control practicos and management directivos established to support improved human performance, UNp management bolloves that prompt identification and timelinoso of correctivo actions has and will continue to improvo.
l Cp&L considers the established controls adequate for cnsuring compliance with 10CFRSO, Appendix D, Cr)torion XVI.
The correctivo measures addressed above and the reinforcement of existing management controls will I
support continued improver.ent of timely correctivo action and ensuro continued compliance.
l i
i i
a ATTACllMEllT 2 1
I.
Ilanin_Isr_ Din puting _th e_A pplicalti lity_oLExampir_R_to 1
the Violation On 4/27/91 with Unito 1 and 2 in uhutdown, the EDG #2 output breaker charging upring failed to chargo during testing.
A WitTO was promptly initiated.to correct the prohlom.
Following initial attempta to identify and resolvo the cause of the problem, the broakor van replaced. On 4/29/91, the plant-Nuclear Sa foty_ Committee (pHSC) directed Technical Support to_ identify the root cauco of-the problem.
The vendor was contacted and the details of the event discusnod. The vendor nunpocted the holding pawl anaombly was either worn or brokon and
-indicated the - failuro apponrod to be an inolated cano based on che sympto.no exhibited and the vondor's experience. On 4/30/91-'Iochnical Support decided to ahlp the failed breakor to tho vendor's repair facility for repair and root cause datormination. Tho System Engineer was also cent to tho vendor < repair facility to witnosn the dianunombly and -inspection of the breaker.
On 5/2/91, _following dianacombly and inopoction _ of the failed ~ breakor, tho vondor connidorod tho failuro moctanium to be an isolated caso.-
Concluniona of the root caune wore lator confirmod and documented by the vendor. final analyais report on 5/6/91.
As dofined within the vondor report, tho au found damago to the charging apring ratchot accombly wan cauuod by misalignment of the holding pawl annombly during original assembly of the bronker.
On 5/3/91 Technical Support reviewed the nood for breaker inopoctions to _assons whethor thoro was a generic L
concern.
Ilaced on thin review the docinion was made to proceed with a check of the cloning apring charga indication in an effort to provido a-prompt indication of the operability statuu of the breakorn. A final dociolon on what additional inopoctionn, including _ pawl annombly_
inopoction, may have boon nococcary to renolvo tho innue was not made as the ayatom engineer had not yet returned from the factory with the neconnary information regarding what additional-Inspections woro appropriato and foanible. The charging npring. Indications of the anfoty related and 11 non-nafety related 4KV breakora were inspected and verified to have proper indication.
,m
ATTACHMENT 2 2
Later that same day the NRC Senior Resident inspector suggested that inspections of a 4KV breaker was needed.
The breaker selected for incpection was a test and ground breaker which was already partially disassembled.
With the selected breaker partially disassembled, a complete inspection of the pawl assembly was performed.
The test and ground breaker pawl assembly was found misaligned.
Misalignment of the test and ground breaker triggered a second inspection involving a sample of safety related breakers. Tha scope of this inspection included about 90 percent of the pawl assembly since complete inspection was not possible without breaker disassembly.
- However, enough of the assembly could be inspected to support an accurate determination of the pawl assembly condition.
Based on discussions with the vendor which indicated no past experience with other failures of this type and the confidence gained by the results of inspections performed, plant management considered the disassembly problem a
non-generic concern.
- However, it was determined that additional inspections would continue.
On 5/5/31 Engineering Evaluation 91-0155 was approved documenting the results of the vendor failure analysis and breaker incpection.
The evaluation implemented the corrective actions required to ensure continued breaker inspections and review of the final vendor-failure analysis report for applicability to 10CFR21 reporting requirements. On 5/6/91 Units 1 ard 2 startup comnienced.
As of 6/6/91, 16 additional safety related breakers were inspected and no 6dditional assembly problems were identified.
In su.ama ry,
Cp&L c.
4idars the actions initiated including corrective r a i..tenance, breaker replacement, coordination or
'endor inspection and failure analysis, and charging spring checks were prompt and effective in providing the confidence that the problem was an isolated occurrence.
The test and ground breaker is not a safety related breaker, was partially disassembled and not representative of safety related breakers installed in the plant. All inspections of breaker pawl assemblies to date have noted no deficiencies and support initial vendor and Technical Support preliminary failure analyses.
Nonetheless, additional inspections of 4KV safety and non-safety related breakers will be performed during future preventive maintenance activities.
ATTACllMl'NT 2 3
II.
ihulgLJar_Dininiting_t htLA pRlEAkility_.pL l;x mmln_j Ltp the ViolAtb n CP&b maintains the position prosented during the August 2,1991 Enforcement Conference and considern the cauno of the 1991 EDG #3 load tent slow responac timo to be unrelated to the clow start timon incurred in 1988 and 1989.
The cauno of the slow start timo conditionn woro correctly and promptly identified in each of the three occurrences. A discunnien of the evento during the 1988, 1989 and 1991 tento in provided below.
In Januuy of 1988 EDG #3 start time of 12.02 seconds prompted troublo ahooting offortu which determined the root cauno of the ulow start time to be air start distributor minalignment.
Following adjustmont of tho air start distributor, an acceptable start time of 8.93 neconda wan verified.
Further inventigation determined that air distributor rotary valvo components are subject to wear.
Procurement of the rotary valvo componenta wan initiated in May of 1988.
In September of 1989 during dienol testing, a ntart time of 10.78 acconda was verified.
Following air atart distributor adjuntment, a special procedure wan performed to verify diesel ntart time.
The special proceduro required de-energizing E-bun 3 and rendering EDG #3 electric driven fuel oil pump inoperable prior to dienol atart. The acceptable responne timen verified following both the 1980 and 1989 repairs indicato the caune of the alow start weis identifled and corrected and that the shaft driven fuel oil pump wan not having a detectable affect on the start timo.
On 4/7/91 during performance of the EDG #3 load tout, an unacceptablo etart time of 10.725 secondo wan verified.
A correctivo maintenance work order wrn initiated to promptly identify and correct the advs. ne condition.
On 4/10/91, ef forts to adjust the air utart dintributor were completed.
Concurrently, monitoring instrumentation was inntalled prior to the diesel run for evaluation of starting system performance. The data gathered f rom thin run verified that the air utart distributor adjuntment did not improve start timen within acceptablo limitu.
Additional testing concentrated in the fuel oil nyntem was initiated and pinpointed a problem with the shaft driven fuel oil pump.
On 4/20/91, following fuel oil pump replacement, EDG #3 start time was rotested and verifled acceptable at 9.315 neconda.
Vendor failure analysin determined that chatt driven fuel oil pump rollef valvo setpoint inaccuracy caused the degradation of pump performance.
ATTACllMENT 2 4
In.May of 1991 the air distributor rotary valve parts were received.
Current plans include replacement of the EDG #3 air distributor rotary valve components during the next diesel outage.
In summary, CP&L considers the actions initiated to
-resolve the EDG #3 start times in 1988, 1989 and 1991 including adjustments to the air start distributor, diagnostic
- testing, shaft driven fuel oil pump replacement and coordination of vendor failure analysis prompt and adequate.
1991 air distributor adjustment was considered the most promising of corrective measures due to past experience.
11oweve r,
concurrent with air distributor adjustments, efforts to install monitoring instrumentation in support of identifying other causes were implemented.
Finally, as reported in voluntary LER 2-91-007, the electric driven fuel oil pump could mask potential problems with the shaft driven fuel oil pump.
liowever, monthly load test data since 1988 does not j
support a
trend indicative of pump failure and consequently cannot be used to determine the starting point of pump degradation.
Therefore, the results of EDG
- 3 fuel oil system diagnostic testing performed in April of 1991 provided the first indication of a fuel oil system problem, to which prompt measures were taken to correct the root cause. CP&L considers the actions taken to be adequate, timely and connistent with the intent of 10CFR50, Appendix B, Criterion XVI.
III. Basis for Discutina the Apolicability of Example #6 to the Violation On 3/18/91 during perforuance of Special Procedure SP 90-048 ' Emergency Diesel Generator Fuel Oil Consumption' on EDG #3, abnormal observations including knocking sounds, 20-30 amp output fluctuations, collector ring arcing and brush oscillation were made by Operations and Technical Support personnel.
EDG #3 was declared inoperable and a corrective maintenance work order initiated to support repair of the generator brush rigging.
Additionally, in accordance with the Corrective Action Program requirements an Adverse Condition Report ACR 91-140 was generated by Operations.
The WRJO and ACR did not specifically identify all of these observations as they were considered to be symptoms related to the main problem of brush arcing. Tecunical Support determined on 3/19/91 that the knocking and output fluctuations were related to improper fuel oil injection caused by impending failure of the brush rigging transmitting oscillating load signals to the engine governor.
~ '
8 5
ATTI.CilMEliT 2 5
Furthermore, Technical Support concluded that the cause of the engino knocking and oscillations was understood and limited to the EDG #3 brush failure.
The failuro modo for the brush rigging was _ identified as an electrical short across the brush rigging positivo and negative bus bars due to carbon buildup.
Additional inspections of EDG #1,
- 2 and #4 brush rigging were performed.
WitTOs were initiated to correct the concerns identified on the other diosols.
On 3/20/91 a test run of EDG #3 was performod during which diesel performance was observed by the same individuals involved in the initial 3/18/91 observations so they could assure all of the abnormalities they had previously identified had boon addressed.
lione of the abnormal conditions observed on 3/18/91 occurred during this test run indicating that the root cause determination had boon properly assessed.
d w
l l
n n.~,.,,..
e.
,v
,,4 w,-- - - - -
--