Containment Atmosphere Control System (CAC)

Summary of Issues

Washington Public Power Supply System

October 7, 1992

9211230052 921102PDR ADOCK 05000397P PDR

CONTENTS

1.0 Introduction

2.0 Description of thc CAC System

3.0 CAC Issues

3.1 Summary

3.2 Design

3.2.1 Control

3.2.2 Capacity

3.2.3 Catalyst

3.2.4 Drawings

3.3 Modilications

3.3.1 RHR Interface

3.3.2 Loop Seal Drains

3.3.3 Blower Overload BISI Indication

3.3.4 Pushbutton

3.3.5 Blower Seal Replacement

10

12

12

12

3.4 Maintenance and Operation 13

3.4.1 CAC Material Condition

3A.2 Surveillance Programfl'eating

13

14

3.4.3 Operating Procedures

4.0 Conclusions on CAC Operability

15

16

5.0 Root Cause

6.0 Safety Significance

7.0 Open Items

8.0 Similar System Reviews

17

18

19

20

9.0 Reportability Evaluation Timeliness

Appendix A: Chronology of CAC Events

21

22

3.

1.0 IntroductionThis report is a summary of recent Containment AtmosphereControl (CAC) activities at WNP-2. It is a compilation ofsever'al Supply System identified reportable events thatoccurred in 1991 and early 1992. It also includes NRC issuedNotices of Violation associated with the CAC System. Thecompilation of these events in one place provides increasedvisibility of the solutions to past problems, lessonslearned, and the status of corrective actions associated"with the system. As a result of these extensive activitieswe bel'ieve the CAC system was operable prior to plantstartup on March 18, 1992 and will remain operational asneeded to support WNP-2 power operation.

Nine man-months were spent by Supply System and Consultantson a review of the CAC system. The Supply System CAC SSFITeam consisted of a group of six individuals from variousdisciplines. The response to the, concerns uncovered by theSSFI was prepared by a team of ten Supply System Engineeringand Consultant (Bechtel and United Catalyst) personnel. Inaddition, the WNP-2 Plant Technical staff had a team of fiveindividuals involved in CAC testing and management ofmodifications to the system.

2.0 Description of the CAC SystemI

At WNP-2 the CAC System includes redundant catalytichydrogen/oxygen recombiners provided to control oxygenconcentrations in the Primary Containment during degradedpost LOCA conditions. This is achieved by recombination ofthe oxygen with hydrogen. It is expected that hydrogen willbe available in sufficient amounts to not limit the removalof oxygen. The recombiner subsystems (A and B) are locatedadjacent to the Primary Containment in the Reactor Building(Secondary Containment). Each redundant subsystem consistsof a wet scrubber, blower, electric heater, catalyst vessel,after cooler, moisture separator and associatedinstrumentation, valves and piping. A constant speed bloweris used to draw the atmosphere from the Primary Containment,process it through the equipment and return it back to theContainment. The amount of recombination is controlled bythe amount of recycle flow that is directed back through theunit. The amount of recycle flow is controlled by RecycleFlow Control Valve, CAC-FCV-6A/B. As the amount of recycleflow is increased the rate of recombination decreases. IfCAC-FCV-6A/B is fully closed the system functions withsingle pass flow through the unit resulting in maximumrecombination from the containment.

Each CAC train drains water through a loop seal to the

1

Suppression Pool., The drain is required to dispose-of thewet scrubber water flow of up to 10 gpm plus any steamcondensed from the gas stream. Zn addition, wat'er collectedin the moisture separator due to the recombination ofhydrogen and oxygen is also drained to the Suppression Pool.

The CAC trains are provided with Class 1E power and containequipment that is environmentally qualified.

3.0 CAC Issues

3. 1 Summary

Appendix A to this summary provides a chronology of CACrelated events. This chronology begins with a numberof precursors to events discovered during that pastyear that were caused by actions taken during the laterphases of plant construction. For most chronologicalitems a reference to a Licensing Event Report (LER) orInspection Report (IR) is given. These documentsshould be consulted for additional detail.On December 2, 1991 Plant Management requested aTe'chnical Review of the CAC System by the NuclearSafety Assurance Division. This review was in responseto problems discovered with the recycle flow controlloop. This review identified a number of additionalconcerns and resulted in further reviews by the SupplySystem Engineering organization supported by BechtelPower Corporation., In addition, a recognized expertfrom United Catalyst was employed to review selectedaspects of the design.

As a result of this scrutiny and associated testing theplant was forced to shutdown on February 25, 1992 whenproblems were discovered with the CAC drain piping.Design modifications to correct these problems wereimplemented during early March. Reviews by varioussupport organizations were used to support. plantstartup on March 18, 1992.

During the R-7 refueling outage of 1992 a number of CACrelated corrective actions were implemented. TheTechnical Assessment SSFI Final Report was issued andthe Bechtel evaluation of the CAC System was reviewed;

The following discussion is a topical summary ofactivities associated with the CAC System.

3.2 Design

The SSFI Team followed the standard approach forperforming the inspection. The system regulatory andcode requirements were identified. The CAC designbasis was established by reviewing the Licensing BasisDocuments and this was used as a baseline for theremainder of the review.

In addition, Bechtel was tasked to "Perform a technicalreview of the design of the CAC system to establish itsability to perform its intended function." The Bechtel

evaluation was subsequently reviewed and verified byEngineering.

Control

On August 7, 1991 a contract engineer,performing a setpoint calculation review,discovered that incorrect ContainmentAtmospheric Control (CAC) Recycle FlowControl controllers (CAC-FC-67A/B) wereinstalled for both divisions in the controlroom. The plant design and operatingprocedures required these instruments to beused in the automatic mode of operation tocontrol recombiner recycle flow. If theseincorrect controllers had been used in theautomatic mode, they would not havecontrolled recycle flow which could haveresulted in a reduced recombination rate orpossible system shutdown due to excessiverecombination.

Corrective action was taken to change plantprocedures. requiring operation of theseinstruments in the manual mode. This allowedplant operators to control recycle flow from .

the control room by manually positioning theRecycle Flow Control Valve (CAC-FCV-6A/B).

The SSFI team questioned the manual methodbeing used to control=the CAC recycle flow.They believed that manual control based onrecombiner exit temperature was not accept-able. The instantaneous reaction rate asso-ciated with recombination could lead to highrecombiner exit temperatures resulting insystem shutdown at the trip setpoint of=1150'F. This concern was verified by theconsultants (Bechtel and United Catalyst).Due to the transient nature of Post-LOCAoperation, oxygen or hydrogen fluctuationscould come from a) localized pockets ofunmixed gases, b) over compensating recycleflow control by the operator or c) increasing

'crubberdehumidification capacity fromresetting water flow. These fluctuationswould be especially critical if therecombiners were operated near their uppertemperature limits. As a result of this con-cern, Plant Procedures were changed tosimplify operation of the CAC system. Thischange was performed with an appropriate

50.59 review and analysis.. PPM 2.3.3A/B,System Operating Procedures for ContainmentAtmospheric Control (Division I/II) werechanged to require the Plant Operators toplace the CAC units on a. constant 60 percentrecycle flow under post-LOCA conditions. TheEmergency Operating Procedures (EOPs) providefor early startup of the CAC System at 0.5percent. hydrogen by volume. This low hydro-gen concentration combined with high recycleflow will provide a large margin between theoperating point and the high temperaturetrip.There is one open issue on the subject ofcontrol that needs to be resolved.

a ~ The SSFI and Bechtel Teams bothrecommend testing be performed tovalidate the accuracy of the CAC floworifices. .Duplicate flow or'ifices havebeen manufactured and test data for theorifices is expected to be obtained byDecember 1992.

Capacity

The oxygen removal capacity, flow rate, andscrubber heat transfer capacity were- reviewedby Bechtel. The evaluation verified adequateperformance of the system.

Catalyst

The SSFI review questioned the operability ofthe catalyst. These questions involvedpotential halogen poisoning during postulatedaccident conditions; possible water damagedue to wetting during past testing; andadequacy of the surveillance test todetermine catalyst operability. The consul-tants reviewed past data on halogen testingand the system design. The CAC system isdesigned with preheaters and operates withthe feed temperature at 500'F. At thistemperature halogen poisoning is not aconcern because these gases are driven fromthe catalyst bed. Additional margin isprovided by the fact that the surveillancetesting is performed at. a lower temperature(less than 300 degrees F).

The consultants also re-reviewed past testdata to determine the condition of theexisting catalyst. Heat and mass balanceswere performed to estimate the 'amount ofconversion in different sections of thecatalyst bed for seven years of test data.This analysis showed that the bulk of thereaction was occurring in the top half of thecatalyst bed as it should for a healthyconfiguration. The performance of the twounits was slightly different but withinexpected parameters. " The analysis did showthat the "A" bed performance was lessefficient when compared to the "B". However,variation in the process variables andinstrumentation inaccuracies on the two sub-system would also account for this dif-ference.

On March 11 and March 12, 1992,the normal 18month surveillance procedures for the

. catalyst, PPM 7.4.6.6.1.4/5, CAC-HR-1 A/BFunctional Test and Visual Examination, wereperformed as part of the plant restarteffort. During this test the "A" catalystachieved a 102'F per percent hydrogentemperature rise. Thus, it did not meet theTechnical Specification required 120'F perpercent hydrogen.

The test results and associated analysis madeit evident that the acceptance criterion ofthe Technical Specification was verydependent on the analytical methods of'alcu-lating the input parameters and measurement .

of performance indicators; hydrogenconcentration, process flows and tempera-tures. Depending on the data, the tem-perature rise could vary significantly. The,review had led to refinements in inputparameter determination. As a result, duringthis review the temperature rise acceptancecriterion required by the Technical Specifi-cations was identified as suspect. Testresults showed variations are primarily dueto instrument uncertainty and the inabilityto set initial conditions and input parame-ters. Consideration was given to otherfactors such as heat removed by the gas flow,heat capacity of the catalyst bed and thevessel, losses through vessel insulation,supports, and piping, time lag, heat loss

caused by temperature sensors, anduncertainties in flow determination duringtesting. This led to a conclusion that thepresent, temperature rise acceptance criterionwas not an accurate reflection of catalystoperability.The Bechtel consultants concluded that sam-pling of the inlet and effluent gases for

'hydrogen concentration would be a more directindication of catalyst efficiency. This wasconfirmed with the assistance of industryexperts in catalyst bed design and operation.'n

appropriate acceptance criterion would be„

the sampling of the effluent gas stream forhydrogen concentration. for a defined inputvolume percent of hydrogen. A sample retain-ing less than 25-parts per million by volume

'ppmV) hydrogen after passing through thecatalyst bed would indicate acceptable recom-biner operation for a feed of at leastone percent hydrogen by volume. In responseto this recommendation, the normal 18'monthsurveillance procedures for the catalyst,PPMs 7.4.6.6.1.4/5, CAC-HR-1 A/B FunctionalTest and Visual Examination, were modified torequire sampling of hydrogen at both theinlet and outlet of the recombiner. Inaddition, a second set of procedures werewritten and performed, PPM 8.3.246/247, CAC-HR-1A/B Two Percent Hydrogen Test, to injecttwo percent hydrogen by volume in the inletgas stream. A comparison of the temperatureprofiles for the one and two percent testsallows a prediction of catalyst health andability to process higher percentages ofhydrogen. The sampling performed on March 11and 12 showed 6 ppm volume hydrogen concen-tration in the effluent stream of bothrecombiner trains with the "A" train showingslightly better performance. It was thesetest results that led the Supply System toconclude that the catalyst„efficiency wasacceptable and capable of performing itsintended function.A Technical Specification amendment has beenrequested to surveillance requirement4.6.6.1.b.3. The requirement checks effluenthydrogen concentration as follows: "Verifyduring a recombiner system functional testthat, upon introduction of at least 1% by

volume hydrogen into the catalyst bedpreheated to a temperature not to exceed 300degrees F, the effluent stream has a hydrogenconcentration of less than 25 ppm by volume."The NRC requested additional temperatureprofile data be collected in the surveillanceto allow trending of catalyst bed efficiency.The temperature profile data collectionrequirement has been added.

Drawings

On January 23, 1992 CAC Circulating Fan (CAC-FN-1B) failed to start during the performanceof a routine surveillance test. Plantoperators discovered the fan did not operatebecause- of tripped overloads. It was thendetermined that Train B of the ContainmentAtmosphere Control (CAC) System had beeninop'erable longer than the 30 days allowed bythe Technical Specifications. PlantTechnical conducted an investigation thatrevealed the overloads most likely trippedduring the performance of a special one-timetest (PPM 8.3.229TP) to verify flow controlvalve CAC-FCV-6B could be controlled manuallyfrom the control room. The test resulted incycling the fan motor on and off three times

, under low flow conditions, causing theoverloads to trip. The investigation alsofound a discrepancy between the elementarydiagram and Electrical Wiring Diagram (EWD)which lead Plant Operators to erroneouslyconclude the overloads were not trippedfollowing the performance of the special one-time test. This event is described in LER92-003 and the following corrective actionswere identified to correct the drawingproblems:

(a) Perform an as-built inspection to verifythe field wiring on each CAC skid (LER92-003). This inspection was completedprior to the end of the R-7 refuelingoutage.

(b) Update applicable drawings to reflectthe actual field wiring configuration ofthe CAC system (LER 92-003). Thisaction will be completed by November 1,1992.

(c) Upgrade the as-built correctedelementary drawing for .the CAC skid toTop Tier status (LER 92-003). Thisaction was completed prior to the'nd ofthe R-7 refueling outage.

(d) Perform a review of drawings for "similarsafety related equipment/systems andcorrect any=discrepancies. Fromcriteria previously established thefollowing systems will be considered inthe review; main steam leakage control,control room chillers, acousticmonitors, process-post accidentsampling, and process radiation monitors(LER 92-003). These reviews are ongoingand will be completed by December 31,1992-

3.3 Modifications

Five plant modifications have been or will beimplemented to address issues identified during the CACreview.

3.3.1 RHR Interface f

The Licensing Basis Documents have arequirement for simultaneous operation of CACand Containment Suppression Pool Coolingduring Emergency Operations that could not bemet during plant operation. This situationwas discovered simultaneously by the SSFIteam and the System Engineer. The reviewdiscovered a precaution in the plantprocedures stating the isolation valve (RHR-V.-134A/B) in the drain line from the CACsystem to the RHR system must be closed toprevent flooding of the CAC system during'testing of the RHR system. Flooding wouldoccur because the drain line for the CACequipment was connected through a loop sealto the RHR line that returns flow to thesuppression pool. The back-pressure in thisline with RHR running would be approximately90 psig which would flood the CAC equipmentunless the interface isolation valves (RHR-V-134A/B) were closed. During CAC Systemoperation the RHR-V-134A/B valves must beopen to allow the scrubber (CAC-AW-1A/B) andthe recombination water to drain to prevent

flooding the CAC units. Furtherinvestigation showed this interaction betweenCAC and RHR would also occur'hen RHR wasoperated in the suppression pool cooling modeduring postulated accident conditions.

Further evaluation showed restrictingorifices (RHR-RO-8A/B and RHR-RO-9A/B) wereinstalled in the RHR returns to theSuppression Pool during the later phases ofconstruction by the ArchitectEngineer/Construction Contractor. Designchange PED 215-M-N002 was written in April1983 to move the restricting orifices becauseof excessive cavitation and vibration. Thechange moved the orifices from a pointupstream of the CAC drain .line to a pointdownstream of the CAC drain. Calculation5.17.26 requires a RHR pump head of 267 ft ata flow rate 'of 7450 gpm through the RHR heatexchanger. When the orifice, sized to thisrequirement, was moved downstream of the CACdrain line, the two systems could not beoperated at the same time without floodingthe CAC system.

An Urgent Plant Modification (92-0056) wasinitiated to correct the interface problembetween CAC and RHR. The drain line for each.recombiner was rerouted from the RHR returnline to the suppression pool using pipingthat is part of the deactivated steamcondensing mode of RHR. This change preventswater from backing up from RHR into the CACequipment (LER 92-007). This work wascompleted and the system was tested prior toplant restart on March 18, 1992.

Loop Seal Drains

Concurrently with the investigation of theRHR drain interface issue, additional actualflooding problems were realized. As a resultof an SSFI concern that a full integrated CACsystem test had never been performed, aspecial test, PPM 8.3.230TP, "CAC-HR-1BRecycle Flow Verification from Drywell," wasbeing performed on February 25, 1991. Thistest required operation of the 'B'ecombiner(CAC-HR-1B) with the CAC scrubber (CAC-AW-1B)in operation. The scrubber drains throughthe loop seal to the RHR system and is the

10

main source of water from CAC as a result of=system operation. During the performance ofthis test erratic cyclic operation of the CACsystem air flow was noted and the 'B'nitblower (CAC-FN-1B) and motor. were wetted.The wetting resulted when the, scrubber'rainwater backed up'into the CAC units. Furtherinvestigation and testing of the

'A'ecombinerunit on February 27,. 1992, showedthe drain problem was also present on theredundant unit.A review of the Preoperational Test Summary(PT 22.0-A) showed that drainage problemswere experienced on CAC during testing inNovember 1983. On November 28, 1983 the testsummary remarks note that after scrubber flowwas established the temperatures droppedrapidly due to the fan pumping water. It wasnoted that no water was draining and the fan,CAC-FN-1B,. tripped on overload. The remarksnote that prior to tripping cyclic fillingand draining was occurring in the moistureseparator.

The SSFI- review of the design documentsassociated with CAC showed two design changes(S215-M-7001 and S215-M-7097) were issued inJanuary 1984 to eliminate water carryover tothe blowers of the CAC units. These designchanges were implemented by PlantModification Record 84-304 in March 1984.Changes were made in'the piping associatedwith the CAC drains but no record could befound where a functional test was performedto verify the drains worked correctly.An Urgent Plant Modification (92-0057) wasinitiated to correct the CAC drainageproblem. The inlet side of the loop seal wasvented to the incoming line from primarycontainment. This change provided a path forthe gas trapped in the loop seal to escapewhen scrubber flow b'egins to flow into theloop seal., It also precluded the formationof a vacuum that could'inder drainage fromthe scrubber.

Following implementation of the plantmodification, the two test procedures, PPM8.3.248TP and PPM 8.3.230TP, CAC-HR-1A/BRecycle Flow Verification from Drywell, were

11

again performed to verify proper operation ofthe. CAC system. These tests confirmed theproper functioning of the manual recycle flowcontroller from the control room and properdraining of the system with the CAC scrubberin operation (LER 92-007). This testing wascompleted prior to plant restart on March 18,1992.

Blower Overload BISI IndicationDuring the January 23, 1992 event describedin section 3.2.4 above, the Plant Operatorsdiscovered the CAC blower motor overloadswere tripped, but the BISI did not indicateCAC was out-of-service. The CAC systemdesign was changed to modify the controlpower and BISI circuits to provide positiveindication of system power alignment (LER 92-003). This work was completed prior to theend of the R-7 refueling outage.

Pushbutton

PER 292-231 was written when it wasdiscovered the CAC test pushbutton couldcreate a primary containment bypass leakagepath if a hot short were to occur. Thisissue was evaluated to be not reportable asthe amount of bypass leakage was below theanalyzed value. A temporary modification iscurrently in place to mitigate this designdeficiency. PMR 92-0085 was issued on March18, 1992 to replace the pushbutton with amaintained contact switch and a second switchin series.

Blower Seal Replacement

PER 292-480 was written when it wasdiscovered that the installed CAC blowerlabyrinth seals contained teflon which wasnot environmentally qualified and coulddeteriorate when exposed to high radiation.This would lead to degraded blowerperformance.

New seals were designed, manufactured andinstalled. Subsequent testing showed theblower performance was satisfactory. with thenew seals. Both CAC blowers were modifiedunder PMR 92-0162 prior to the end of the R-7

12

refueling outage.

3.4 Maintenance and Operation

3.4 ' CAC Material Condition

On September 11, 1991 one train of CAC wasfound to be inoperable due to the loss oflubricating oil causing the blower shaft toseize. The oil had leaked out due to aloose drain plug.

In October 1991 an NRC Inspector notedproblems with the material condition of theCAC skids (missing or loose nuts on hangersand uncertainty about the environmentalqualification of CAC-PT-1A). TechnicalSpecification 4.6.6.1.b.4 requires a visualexamination of abnormal conditions everyeighteen months. The Supply Systemacknowledges the existing surveillance didnot provide physical verification of fastenertightness. Further Supply System reviewsidentified additional problems. As a resultof these two findings on material conditionthe following corrective actions werecompleted or initiated:(a) Lockwires have been installed on the

drain plugs of the Recombiner blowers(LER 92 025 1g IR 91 44)

(b) The appropriate documentation has beenchanged to require the drain plugs ineach recombiner blower to be properlyinstalled and lockwired (LER 92-025-1,IR 91-44).

(c) A Maintenance and Operations Bulletinwas issued to appraise craft and systemengineers of the necessity to ensurepositive locking ability for criticalpipe plugs in vibrating equipment (LER92-025-1/ IR 91-44).

(d) Procedures were reviewed to evaluate howfasteners and pipe supports were ad-dressed. The review (SS2-PE-92-0047)recommended that enhancements be made totwo procedures. Plant procedures PPMs10.2.10 and 10.2.29 were deviated toprovide improved guidance on instal-

13

lation of pipe fasteners and supports(IR 91-44).

(e) An evaluation, including a review ofrecommendations in the manufacture'sliterature was performed on the CACblowers to determine the appropriateinterval for shaft rotation to preventshaft damage due to system vibration (IR91-44). The evaluation was completedJune 26, 1992, and concluded there wasno specific need for rotating theblowers more frequently than what isrequired to satisfy other existingsurveillance requirements.

(f) Plant Procedures PPMs 7.4.6.6.1.4/5 wererevised to more clearly'efine therequired visual examinations with theSystem Engineer Program Weekly FieldWalkdown results.

Surveillance Program/Testing

Various surveillance tests are performed onequipment associated with the CAC system.This includes an 18-month surveillance (PPM4.6.6.1.b.1) which requires performing aCHANNEL CALIBRATION of all recombineroperating instrumentation and controlcircuits. Plant Procedures PPMs7.4.6.6.1.3C/D, H2 Recombiner 1A/B FlowInstrumentation Channel Calibration, are thechannel calibration procedures. However, onNovember 21, 1991, it was discovered thatthis surveillance had not tested theoperation of the Recycle Flow Control Valve,CAC-FCV-6 A/B, from the Remote Master FlowController, CAC-FC-67 A/B in the manual modeof operation. A trouble shooting plan wasformulated and implemented that demonstratedoperation of the CAC-FCV-6 A/B from thecontrol room.

As a result of the deficiencies found in theSurveillance procedures, the followingcorrective actions were identified:(a) Plant Procedures PPMs 7.4.6.6.1.3C/D

were to be revised to incorporate a testof the CAC-FC-67A/B to CAC-FCV-6A/Binstrument control loop (LER 91-029-1).

14

Since the CAC system has 'to beoperational to obtain a meaningful testthis surveillance requirement wasinstead incorporated into PlantProcedures PPM 7.4.6.6.1.1 and7.4.6.6.1.2, Preheater OperabilityTests. This item was completed prior tothe end of the R-,7 refueling outage.

(b) CAC surveillance procedures were. validated and verified to assure all

Technical Specification functions arebeing performed (LER 91-029-1). Thisitem was completed prior to the end ofthe R-7 refueling outage.

(c) The process and direction used for.system walkdowns has been strengthenedand walkdowns on the CAC system areperformed on a weekly basis. This itemwas considered closed as of May 1, 1992.

(d) The CAC system was run every 30 daysuntil the SSFI was complete and itsrecommendations were evaluated (LER 92-003, IR 91.-44). This item wasconsidered closed as of May 1, 1992.

(e) The plant procedures for verifyingrecycle flow from the drywell will beperformed on a periodic basis.Performance of these integrated testsprovides additional assurance allaspects of the CAC system functionproperly.

Operating Procedures

For the recycle flow controller problem,plant System Operating Procedures, PPMs2.3.3A/B, Containment Atmospheric Control,were revised to require operation of CAC withCAC-FC-67A/B in the manual mode. The CACoperating procedures were validated and veri-fied (LER 91-029-1). These actions werecompleted prior to startup following the R-7refueling outage (LER 91-029-1, IR 91-44).

15

4.0 - Conclusions on CAC OperabilityNine man-months were spent by Supply System and Consultants

*

on a review of the CAC system. The Supply System CAC SSFITeam consisted of a group of six individuals from variousdisciplines. The response to the concerns uncovered by theSSFI was'prepared by a team of ten Supply System Engineeringand Consultant (Bechtel and United Catalyst) personnel. Inaddition, the WNP-2 Plant Technical staff had a team of fiveindividuals involved in CAC testing and 'management ofmodifications to the system.

As discussed above, Bechtel was given a scope of work thatincluded an overall review of the CAC system. In addition,the extensive focus by all the SSFI team members resulted indesign reviews, system testing, and system walkdowns, whichprovided confidence that the system was operational prior tothe March 18, 1992 startup. Specifically, quantitativeanalysis of the catalyst performance based on effluenthydrogen concentration, full flow testing of the recycleratio control using the drywell atmosphere and maximumscrubber flow, and verification of the catalyst hightemperature shutdown safety function all provided confidencethat the system was fully capable of performing its intendedsafety function.

During the R-7 refueling outage outstanding correctiveactions were implemented to provide further assurance of CAC.operability. This included an as-built inspection of thefield wiring on each CAC skid. In addition, the variousreports associated with the CAC review were finalized.Following resolution of the blower seal (Teflon) issue, theCAC system was declared operable following successful post-modification testing and surveillance testing prior torestart at the end of the R-7 refueling outage.

16

5.0 Root Causes

The dominant features in the root cause analyses performedto date on significant CAC issues have been less thanadequate design analysis and/or design review.

Inadequate design analysis has resulted in systeminteraction problems (CAC-RHR, Pushbutton), unworkable drainconfiguration, incorrect controller selection, faultycontrol methods (operator controlled recycle) and inaccuratemonitoring of equipment condition. Design review failuresin the form of inadequate testing precluded discovery of theerrors from the analysis/design process and allowedfundamental design deficiencies from the construction phaseof the project to continue.

The root cause of the unqualified Teflon seals beinginstalled in the CAC blowers was less than adequate workpractices and procedures. Corrective actions have beenimplemented..

17

6.0 Safety Significance

We believe the events associated with CAC had safetysignificance because the operability of. both, divisions ofone of the WNP-2 Engineered Safety Features was impacted.However, the actual safety significance is mitigated by thevery- low probability of the need for the system.

With an inerted containment, combustible gas control isbased on the amount of oxygen in the containment. In thedesign basis analysis, the probability of oxygen productionfollowing a large LOCA event has been assessed by the

Supply'ystem

PRA effort. A large LOCA can be due to pipe break ordue to'ailure of sufficient SRV's to open when needed. Onthe basis of generic failure rate for pipes and plant-specific failure rate for SRV's, the probability of a largeLOCA has been determined. A large LOCA event can be miti-gated as long as any one of the ECCS systems operate. Plant-specific system unavailabilities have also been determined.The oxygen production (or a LOCA without coolant injection)probability can be assessed by multiplying the LOCAprobability by the ECCS System unavailabilities. Thisnumber is approximately lE-7/year. The NRC's proposedsafety goal is 1E-4/year for core damage and 1E-6/year forcontainment failure resulting in a large release to theenvironment (SECY-91-270). Although oxygen production doesnot mean there is a subsequent burn or containment failure,the oxygen production probability already meets the safetygoal for containment failure.Severe accident studies done in support of Individual PlantEvaluation (IPE) work has shown that the amount of oxygengenerated and surviving as a gas in a degraded core scenariois very small. — With an inerted containment the oxygen isnot likely to exceed the flammability limits precluding theneed for the recombiners. In addition, the initial workbeing done on the WNP-2 IPE concludes that CAC is anegligible contributor. toward accident mitigation. For coremelt accidents, the exothermic metal water reaction, oncestarted, generates large quantities of hydrogen. Sincethere is a limited amount of oxygen that can be reacted theCAC system is of marginal value. Hydrogen combustion is, nota credible containment failure mechanism for an inertedcontainment (Reference: FAI/91-110, Deflagration andDetonation of. Hydrogen, July 1991).

18

7.0 Open Items

Many CAC issues have been closed as noted in the discussionabove. However, the following open items are still in theprocess of resolution:

7.1 The SSFI and Bechtel Teams both recommend testing beperformed to validate the accuracy of the CAC floworifices. Duplicate flow orifices have beenmanufactured and test data is expected to be obtainedby December 1992.

7.2 Applicable drawings will be updated to reflect actualfield wiring configuration of the CAC system (LER 92-003). This will be completed by November 1, 1992.

7.3 An engineering evaluation and drawing review will beperformed on four similar safety relatedequipment/systems. The evaluations and reviews areongoing and will be completed by December 31, 1992.

7.4 The Supply System SSFI team issued a total of 25Quality Finding Reports (QFRs), which mainly identifiedprogrammatic concerns and deficiencies in documentationassociated with the CAC system. Only one of the QFRsinvolved an operability issue. This issue isassociated with the CAC flow orifices. A Basis forContinued Operation, which was approved by POC,addresses this issue. Currently the QFRs are beingdispositioned by the responsible organizations.

19

8.0 Similar System Reviews

A concern was raised when it was discovered- that a designchange had been issued (February 1982),. but not implemented, .to correct an identified problem with the recycle flowcontrollers.The design change process in place during constructiondepended on contractors to implement changes that wereissued by the Architect-Engineer. It is concluded, based onthe turnover process put in place at the end ofconstruction, that the failure to implement the designchange was an isolated occurrence. The construction designchange process in place when this event occurred wascompletely changed when the plant went into operation.Therefore, no further corrective'ction is warranted (LER91-029-1, IR 91-44).

A more general action being taken in response to the prob-lems associated with the CAC System involves the identifica-tion of other systems that may have characteristics similarto those which are believed to have contributed to the CACsituation. This evaluation considered seven'riteria: 1)Late additions to plant design, 2) Not fully testable, 3)Not fully designed by Burns and Roe, 4) Not fully designedby General Electric, 5) Importance of the system to plantorganizations, and 6) Current status of system deficiencies.The results of this evaluation recommended four systems forconsideration of additional reviews: Control Room Chillers,Post Accident Sampling, Main Steam Leakage Control andProcess Radiation Monitors. An'ngineering evaluation willbe performed for each system which will include a drawingreview.'his will include a limited team inspection. Theseevaluations and inspections are ongoing and will becompleted by December 31, 1992 (LER 91-029-1, LER 92-003, IR91-44).

20

9.0 Reportability Evaluation Timeliness (IR 91-44)

An additional issue associated with the reporting of the CACflow controller problem, was the timeliness of reporting.Corrective actions have been taken and planned to deal withthis concern as follows:

(a) Action was taken to reduce the backlog of items requir-'ing a reportability evaluation to less than ten. Agoal has been established to maintain the number ofitems at a reasonably low level.

(b) A survey was performed of the process/methods used byother utilities to manage reportable items.Information has been obtained from fourteen utilitieson the process used to evaluate reportability and theaverage backlog. See item (d) below.

(c) The responsibility for root cause analyses in responseto Notices of Violation has been transferred from theCompliance group (those responsible for performingreportability evaluations) to the Operating EventAnalysis & Resolutions group. This has freed upresources to perform reportability evaluations.

(d) An independent assessment was performed on theReportability Evaluation process. This evaluationlooked at resources verses=tasks and determined howimprovements can=be made in the overall process. The'evaluation took into account information obtained fromother utilities. The assessment was completed July 1,1992. The evaluation recommendations are beingevaluated for implementation.

21

r r

Appendix A

Chronology of CAC Events

June 2, 1981,

February 11, 1982

April 1983

April'983

November 1983

December 1983

Startup Problem Report identifies the wrongrecycle flow controllers have been installed(See August 7, 1991 below) (LER 91-029-1).

Design Change (PED 218-I-3923) issued toprocure and install the correct ratiosetpoint controller. (See August 7, 1991below.) (LER 91-029-1)

A system lineup test was performed on thewrong controller. The test was limited to afunctional test of the wrong device. Theprocurement process failed to procure thecorrect flow control device for recycle flowcontrol. (See August 7, 1991 below.) (LER 91-029-1)

Design Change (PED 215-M-N002) written tomove the RHR restricting orifices in the RHRreturn line to the suppression pool. Thiswas later realized to precluded simultaneousoperation of CAC and RHR Suppression PoolCooling mode of operation. (See February 25,1992 below.) (LER 92-007)

During the CAC Preoperational Test drainageproblems were experienced. Cyclic fillingand draining occurred followed by blowerflooding. A design change was initiated toincrease the size of the loop seal to preventflooding; (See February 25, 1992 below.) (LER92-007).

The Preoperational test performed on the CACsystem. The test did not discover the factthat the incorrect flow control device wasinstalled in the control room nor was thescrubber retested to determine if thedraining problem was solved. (See August 7,1991 below.) (LER 91-029-1)

February 13, 1984 „An event occurred where both hydrogenrecombiner fan (CAC-FN-1A/B) motors trippedon electrical overload during preoperationaltesting at 18 psig containment pressure (LER84-013) .

22

1984-1990 Routine surveillance testing was performed onthe CAC system.

May 24, 1990 The CAC 'A'kid was flooded when stroke timetesting was performed on the CAC drain valve,RHR-V-134A, simultaneously with the operation.of RHR 'A'n the Suppression. Pool Coolingmode of operation (LER 92-007).

December 8, 1990 During the CAC 'A'urveillance test, theblower oil drain plug vibrated loose causingCAC A" to be inoperable during four monthsof power operation. (See September 11, 1991below.) (LER 91-025-1)

August 7, 1991

August 29, 1991

September 11, 1991

A contract engineer working on the setpointevaluation program discovered the wronginstrument was installed for CAC recycle flowcontrol (LER 91-029-1).

CAC operating procedures revised to controlCAC recycle flow in the manual mode prior toplant startup. Plant operators were directedto monitor recombiner catalyst temperatureand drywell pressure to maintain minimumrecycle ratio (maximum recombination) (LER91-029-1).

CAC 'A'lower found seized with oil drainplug lying at the bottom of the blower-motorassembly container (LER 91-025-1).

October 1991 The NRC Inspector noted problems with thematerial condition of the CAC skids (missingor loose nuts on hangers and uncertaintyabout the environmental qualification of CAC-PT-1A) (IR 91-44).

October 31, 1991 Reportability Evaluation completed thatconcluded the problem identified with therecycle flow controller on August 7, 1991 wasreportable (LER 91-029-1).

November 21, 1991 A review of the survei'llance proceduresdiscovered. the remote manual control of the'recycle valve from the control room had neverbeen tested. Movement of recycle valve wasdemonstrated (LER 91-029-1).

23

November 27, 1991

December 2, 1991

December 17, 1991

December 20, 1991

January 23, 1992

January 31, 1992

Further evaluation found a Safety Evaluationwas not performed for the change to the CACOperating procedures performed on August 29,1991. The Safety Evaluation required adedicated operator to be. placed at therecombiner panel in the Control Room withinsix hours following a LOCA (LER 91-029-1).

Plant Management requested a TechnicalAssessment (SSFI) be performed on. the CACSystem by the Nuclear Safety AssuranceDivision (LER 91-029-0).

A recycle flow verification test (PPM8.3.229TP) was performed to verify manualcontrol on Train. B. This test resulted incyclic operation of the CAC blower whichcaused an undetected trip of the phase A andC'overloads. (See January 23, 1992 below.)(LER 92-003)

Enforcement Conference regarding CAC systemproblems.

During the performance of the routine CACpreheater operability test the CAC 'B'lowerfailed to start due to tripped overloads (LER92-003) .

Supply System SSFI team presents preliminaryconcerns. Bechtel contracted to evaluate thepreliminary concerns raised by the SSFIreview.

February 6, 1992 Notices of Violation and Civil Penalty issued(IR 91-44).

February 6, 1992

February 12, 1992

February 25, 1992

Plant Operations Committee (POC) approvedTest Procedure TP 8.3.230, CAC-HR-1B RecycleFlow Verification From Drywell, for useduring Mode 1 operation. (See March 30, 1992below.) (IR 92-03)

A review of surveillance procedures performedto show operability of the catalyst sinceplant startup showed three tests did not meetTechnical Specification requirements for flowand temperature rise (LER 91-029-1).

WNP-2 is required to shutdown because ofrequirement to have the capability to operateCAC and Suppression Pool Cooling mode of RHR

24

February 25, 1992

simultaneously. Operating both systemsduring accident conditions would result inflooding of CAC (LER 92-007).

Actual flooding of CAC 'B'ccurred when atest was performed to prove CAC scrubberoperation in response to an SSFI concern.The problem was traced to incorrect ventingof the drain loop seal (LER 92-007).

February 27., 1992

March 12, 1992

March 13, 1992

Further evaluation showed the CAC scrubberdrainage problem was also present in CAC

'A'LER92-007).

During functional testing of the CAC'A'atalyst,Technical Specification temperature

rise requirements were not met. Measurementof the hydrogen concentration to verifycatalyst performance, as recommended by theCatalyst Consultant, met performance criteria(LER 91-029-1).

WNP-2 'requested relief from the TechnicalSpecification requirement to verify catalystoperation by measuring temperature rise.

March 15, 1992 A request was made to change the TechnicalSpecifications to require verification ofcatalyst performance by measuring hydrogenconcentration (LER 91-029-1).

March 15, 1992 Operating procedures were revised to requirea constant 60 percent recycle flow under postLOCA conditions. This change was made inresponse to the SSFI concerns (PER 292-229)about the manual method being used forcontrol the CAC system (LER 91-029-1).

March 18, 1992

March 30, 1992

CAC Operational — Plant Startup

NRC issued a Notice of Violation (NOV) forapproving the use of Test Procedure 8.3.230during power operation. Performance of thetest would have been a violation of theTechnical Specification limit on PrimaryContainment Bypass Leakage. (See February 6,1992 above.) (IR 92-03)

May 14, 1992 CAC blower labyrinth seals were found tocontain Teflon which was 'not environmentallyqualified (LER 92-022).

25

May 29, 1992

June 11, 1992

June 24, 1992

July 1, 1992

Supply System Technical Memorandum TM-2022was issued conveying the Bechtel CAC SystemReview report.CAC System SSFI final report was issued.

Completed installing modified blower seals.

CAC operational following post-maintenancetesting and surveillance testing.

26