ML20197A991

Responds to NRC Re Violations Noted in Insp Rept 99900404/97-02.Corrective Actions:Observation Provided to W as FSER Open Item 440.753F in

ML20197A991
Person / Time
Site:	05200003
Issue date:	02/27/1998
From:	Mcintyre B WESTINGHOUSE ELECTRIC COMPANY, DIV OF CBS CORP.
To:	Black S NRC OFFICE OF INFORMATION RESOURCES MANAGEMENT (IRM)
References
REF-QA-99900404 99900404-97-02, 99900404-97-2, NSD-NRC-98-55-8, NUDOCS 9803100075
Download: ML20197A991 (24)

Text

_ _ _ _ _ _ _ _ _ _ _ _ _ - - _ _ _

' Westinghout,a Electric Company  % 333 reormwsmass NSD-NRC 98-55 80 DCP/NRCl264 Docket No.: 52-003 February 27,1998 Document Control Desk U. S. Nuclear Regulatory Commission Wash!ngton, D. C. 20$$$

l Attention: Suzanne C. Illack i

l

Subject:

Inspection Report No. 99900404/97-02: Reply to a Notice of No.vonformance

Reference:

Letter, S. C. Illack to N. J. Liparulo, "NRC Inspection Report No. 99900404/97-02", Dated January 28,1998 This letter documents the Westinghouse response to an NRC inspection of the AF600 project conducted from November 17 through November 21,1997. This inspection focused on design calculations and computer codes supporting the SS AR Chapter 15 transient analysis and the containment analysis portion of SSAR Chapter 6. Two nonconformances and an unresolved item were identified in the Reference.

Nonconformance 99900404/97-02-01 identified concerns with w.e oesign calculations conformance to the quality assurance requirements for design analysis and verification. A description of the corrective action, actions to prevent recurrence and a resolution schedule for each of these concerns is contained in Attachment 1. Five of the concerns were addressed as FSER Open items where additional information was provided to supplement the design calculation. Two calculation notes were revised to provide justification for the engineering judgment exercised in the analysis. The impact of the revised calculaticns did not extend beyond the document involved. Two additions were made to the SSAR to include a discussion of the operator actions during a coolant inventory increase event and to modify the technical specifications to preclude potential boron dilution events. To prevent recurrence of the type of concerns cited in the nonconformance, WCAP-12601, the AP600 Program Operating Procedures, has been expanded to include requirerrents for design analysis documents to provide enhanced documentation of engineering judgment as well as adequate information to indicate how verifiers comments are resolved. /T> <j i 9803100075 900227 j f l ,

^ "

. ,. S " 5" 88R l{Hl.HM. 1%El.E i' l  ;

I SN h% ,

~a , y' ,

m , ,

4 t

r .

- Nonconformance 99900404/97-02-02 identified two failures to follow procedures in the use of x WCOBRAfrRAC and GOTillC. Corrective action request TARS) have been initiated and are g discussed in detail in Attachment 1. Neither of these failures had an impact on the SSAR. Following h,; the inspction, Westinghouse documentation was locr'ed that resolved the error repoits from NAl on GOTillC with 10CFR 21 implications. The error reports were dispositioned in accordance with our E

procedure WP 4.19.3, Errer Reporting and Resolution. NAl errors not involving 10CFR 21 did require error resolution as part of the corrective action and this also has been completed. None of the errors

% reported had an impact on the AP600 contal nent analysis. In addition to the corrective action defined for AP600, Westinghcuse Qudity Systemt .c performing a self assessment on our disposition of enor

_ reports from other third party software suppliers. To prevent recurrence, the AP600 Project addresseo y this issue by providing additional training to the code responsible engineer. The training reinforced the requin;ments of Westinghouse procedure WP ;.19 3.

h Unresolved item 99900404/97-0243 identified tech tical questions on six additional calculation r A detailed response to each of these concerns is also contained in Attachment 1. Additional infon in the design calculation could have assisted in the review and understanding of the calculation doct. mentation. In response to severni conceras on input to the containment analys!.4, corrections were made to the WGOTitlC material contained in calculatbn i100-SOC- 001 and a reanalysis was performed. The revised containment peak pressure analysis (which will included in SS A' subsection 6.2.1) showed e slight reduction in the peak pressure.

The raference letter also requested that Westinghouse " assess the adequacy of the AP600 QA design review process and the ime,,rity of the AP600 design." As documented in Attachment 1, a large majority of concerns identified by the NRC were related ta tne level of documentation provided in the calculation notes. These hve been resolved by providing additional information. In addition, WCAP-11601, the AP600 Program Operating Procedures, has been expanded to include requirements for design analysis documentation to provide enhanced doct mentation of enginecrms judgment and verifiers comments resolution. The three SSAR modifications resulting trom our correctWe actions to the first nonconformance and the unresolved item are not expected to change the conclusions reached in the SSAR review.

To further confirm the integrity of the design review process, Wertinghouse has initiated a Design Assessment Review (DAR) to evaluate an additional sample of ca uhtions that support the Chapter 15 analysis and the containment analysis section of Chapter 6. The review plan ano vn in Attachment 2 was discussed with the NRC in a meeting on February 11,1998. The DAR team will issue their final report by April 3,1998.

The NRC also indicated their concern that a large number of discrepsrues were noted by 'he inspection team "in such a small sample of the total population" To address this concern, a cause analysis was performed, under the leadership of the Westinghouse Quality Systems Department. Details and conclusions of the cause analysis are included as Attachment 3. The cause aaalysis assessment did not indicate a trend adverse to quality.

The focus of t'a root cause assessn ent was to establish the fundamental reasons for the lack of documentation by the Westinghouse authors and veiifiers. Interviews wm conducted with autho s and verifiers of a sample of calculation notes that had been reviewed by the MC during the inspection. The interviev s were conducted withM the framework of a root cause evaluation by using a combination of the "5 Whys" technique and enginee:i.1gjudgment. This assessment concluded that the root cause of the 2 M24H dos-

]

___ m - -

s lack of documentation supporting Westinghousejudgments is that the analyses were basically being prepared for internal review and verification by other qualified We.tinghouse personnel. That the verifier had sufficient informatic n to enable verification to be completed in accordance with the governing procedures, as evider ced by the verifiers signature, ind; cates that the author provided sufficient documentation detail. In tesponse to the assessment, the AP600 project has supplemented the AP600 general procedures for design malysis to include requirements for enhanced documentation ,

Beyond efforts specifically related to AP600, an Error Free Engineering Initiative had been established s by the Nuclear Service Division (NSD) as part of its ongoing quality improvement efforts. NSD is the division of Westinghouse responsible for safety analysis. This self identified initiative addresses six key areas; calculation note quality, design review and planning, software quality, corrective action, quality impmvement, and engineering quality cost. Results of the AP600 interactions with the NRC were provided for consideration in the Error Free Enginuring Initiative.

Responses to FSER Open Iteras noted in Attachment I and the corrective actions describeu in the contents of this tri.asmittal provide the required support for closing the nonconformances and the unresolved issue from the NRC irspection. Results of the DAR will be available in early Apri: and any necessary corrective actions will be implemented. Based on oir evalnation of the results of th' b, November 1997 NRC inspection observations, we believe that the quality of the design calcuh ions conforms to the requirements of 10CFP.50 Appendit B. We are expanding our procedure to provide an increased depth of documentation for the purpose of accommodating a third party review Any AP600 analysis performed from this point forward will be documented in accordance with the res! sed procedure.

E i

Please contact me or Bob Tupper on 412-374-5219 if you have any questions concerning this transmittal.

.F-Brian A. hiclntyre, Manager Advanced Plant Safety and Licensing Attachment 1: Disposition of Nonconformances and Unresolved Item n Attachment 2: ' P600 NRC Design Assurance Review Plan Attachment 3: Cause Analysis of NRC Findings on Calculation Notes E

cc: T. R. Quay, NRC/NRR/DRPM R.L. Pettis, NRC/NP.R/HQMB/DZH R. A. Gramm. NRC/NRR/HQMB/DI. 'H j J. D. Peralta, NRC/NRR/HQMB/DRCH

_ N. J. Liparulo, Westinghouse, NSD M. Mutyala, Westinghouse, QPS S. D. .Rupprecht, NSD o

21& d. A

}

a um

Attachment 1: Disposition of Nonconformances and Unresolved Item from NRC Inspection 99900404/97 02 ,

Several of the NRC concerns identified during the final inspection were identified as FSER Open items and Wertinghouse responses were provided prior to the release of the NRC Inspection keport. Rese have been identified in the Westinghouse response to tiec NRC observations noted below.

Nonconformance 99900404/97 024)1 NRC Observation 1:

SSAR GSC 189, "AP600 SSAR Inadvertent ECCS Analysis," Revision 2, assumed the need of operator actions in the analyses ofincreased RCS inventory events, butfailed to address:

\

The availability of(I) unambiguous alarms or indicationsfor increased RCS inventory events, and (2) clear 3 prc.ceduralinstructions to operators to take appropriate actions within the time frame assumed in the analyses.

10 CFR 50.36 "Techr.ical Specsfications," 9$0.36(cx2)(iiXC).

inspections. Tests, Analyses, and Acceptance Criteria (ITAAC) to verify the capacity of the system including the y RV head vent valves that would be used by operators to prevent pressurizer ovedllfrom oscurring as assumed in the analyses.

StalacJ Safety Analysis Report (SSAR) text providing detail of the analyses that credited the requisite operator actions.

Steps taken to correct this item:

his ut servation was provided to Westinghouse as FSER Open item 440.753F in a letter December 17,1998 Westinghouse provided a response to FSER 01440.753F in letter DCP/NRCl209 dated Jruary 9,1998. SSAR Section 15.5 has been revised to include a more complete discussion of the use of operator actions to mitigate the consequences of the increase in reactor coolant inventory events. In response to FRR 01440.785F, a technical specification for the reactor vessel head vent valves has been incorporated. The T CS ITAAC has been modified to ddress the required capacity to accommodate overfill events.

Steps taken to pa vent recurrence:

None required. His observatim represents a difference in approach for Westinghouse and the NRC regard.ng what should be inciuded in the SSAk. The requested information has been edded to the SSAR, Techrical Specifications E and ITA AC.

Date corrective action and preventative measure will be completed:

a Completed January 9,1998.

NRC Observation 2:

SSAR GSC 188, "AP600 Baron Dilution Analysis." Revision 0, relied on the boron muing testing data de:umented in EGG LOFT SE67 (Project Na P 394) t' ~~tablish the required RCS circulationflow rate of 10(

gym in TS 3.4.9 and to support its complete baron mixing model assumed in the baron dilution analyses (SSAR 15.4.6). However, despite signtficant differences between the AP600 design and the testfacility configuratson and testi'.g conditions discussed in EGG LOFT-5867, SSAR GCS-I88 failed to reconcile the applicabilitr of the boron mhing testung data to the AP600 design or to validate ths ccmplete boron mhing model assumed in the boron dilution analysis.

awa i m . _ _ . . ..

~

Nonconformance 99900404/97 0241 (Continued)

- Steps taken to correct this item:

This caservation was provided to Westinghouse as FSER Open item 440.754F in a letter December 17,1998.

Westinghouse provided a response to FSER O! 440.754F in letter DCP/NRCl248 dated February 6,1998. Technical Specification 3.4.9 has been modified to preclude potential boron dilution events when the ree.ctor coolant pumps are not running by isolating the demineralized water isolation valves. SSAR subsecdon 15.4.6 is modified to reflect the change to the technical sp-cification.

Steps taken to prevent recurrence:

None required, the requested info mation has been added to the SSAR and Technical Specifications.

Date corrective action and preventative measure will be cosapleted:

Completed February 6,1998, i

l NRC Observation 3:

l

. SEC APS.4838 CO. " Software Design Specifications ofAP600 NOTRUMP User Externals Cycle 2." Revision O. dated September 9,1995, contained a statement that numerous errors in code parameters were reviewed as insignificant and would be corrected in a later code version however, no basis was given to support this conclusion.

3 Steps taken to correct this item:

Revision 0 of the e dculation was issued to suppon a preliminary verification and validation (V&V) repon as document. tion of progress to date with the clear understanding it would be revised for the final V&V cffort. The author and the verifier were familiar with the code progress and possessed the engineering competence needed to assess the impact of the documented errors as insignificant for the purpose of supporting preliminary NOTRUMP V&V. In accordance with the quality assurance procedures for code vaSdation, the errors ware !isted in the code error report. All of the errors reported in Revision 0 were included as corrections in the modifications to the code reported in the subsequent revision of the code. No action is required to correct this item.

Steps taken to prevent recurrence:

None required. The items identified were previously co:Tected by the AP600 design process and represented only work in progress.

Date corrective action and preventative measure will be completed:

None required.

NRC Observation 4:

SEC-APS-4d37 CO. " Software Change Specificntion ofNOTRUMP Cycle 32." Revision O. dass I September 9 IU$. contained a statement that " ..the author doesn't know enough about the subject (void propagation) to determine the impact of the reviewer's comments." No evidence existed to support resolution of the reviewer's comment.

2Mhu h 2

Nonc aformance 90900404/97 0241(Continued) ,'

Steps taken to correct this item:

his cel-lation was revised to docu.nent the resolation of the teviewers comment during the NRC inspection. The impact did not extend beyond SEC-APS-4837 CO. The revised calculation was reviewed with the inspector and there are no opei issues hssociated with this calcula*'on.

Steps taken to prevent recurrence: "

WC/.P 12601 has beer (.xpanded to include .equirements for calc ation nc.es to provide enhanced documentation -

of engineeringjudgment as well as adeqc.e information to indicate how verifiers comments are resolved.

Date corrective action and preventative measure will be completed:

Completed February 27,1997.

NRC Observation 5:

SEC-APS 4746-CO, "ECOBRA/ TRAC hmg Term Cooling," provided no basisfor concluding that

"... variations in the imtial conditions are expected to have relt.'ively unimportant e[fects on the analysis results," and "the results ofchangir.g ICilP is noticeable but n' - rge..."

Steps ta';en to correct this item:

his observation wr.s provided to Westinghouse as FSER Open Item 440.757F in a letter December 17.1998.

- Westinghouse provided a response a FSER 01440.755 in letter DCP/NRCl212 dated January 13,1998.

r Calculation FEC-APS-4746-C0 documents the analysis simulating the separate effect tests performed for the CMT.

The comment cited deals with initiali nog the liquid temperature on the WCOBRA/ TRAC cell maleling the short length of pipe immediately above the CMT to contain hot water. The cell has a volume whir.:is only 1.5% of the uppermost cell :n the CMT component to which it connects. The CMT component temperatures in the

}.

simulations are initiated at the test values, which are approxiraately room temperature. The impact of initiating the small pipe cell as hot water increases the net temperature value of the CMT entrance region (the entrance pipe cell and the CMT componen* top cell combined) at the start of the prcblem by only a small amount,1.8 to 5.4 T. No corrective auion is required.

Steps taken to prevent recurrence:

WCAP 12601 has Men expanded to include requirements for calculation notes to provide enhanced documentation of engineenng judgment as well as adequate information to indicate how verifiers comments are resolved.

Date corrective action and preventative measure will be completed:

No corrective action is required. Revision 23 of WCAP 12601 was issued February 27,1998.

L te m 3

Nonconformance 99900404/97 0241 (Continued)

• NRC Observation 6:

LTCT T2C-417. "WCOBRA/ TRAC Geometricalin mt Datafor the OSU Testing." Revision 0, and LTCT-T2C-418, "OSU LTC Comparisons with WCOBRA/TRA C " Revision 1. '

e LTCT T2C-417. Pages 180-181 (Figures 6 and 7) acknowledged tl efailure topt DP vs (flow l however, the basis provided was that " ..despite thefailure to match, cverall agreement is reasonable." This unquantified anomaly was used as input to calculation LTCT T2C 418.

• LTCT-T2C 418. On page 16, a bias of 0.2 psia was applied to the atmospheric pressure ao compensatefor the disparity in DP vs (powf in calculation LTCT T2C-417. However, the calculation did not provide an explanationfor the use of this bias.

Steps taken to correct this item:

During the verification of LTCT-LC-417, the verifier stated on page 179 of the cale note that "Despite the failure to fh, oserall the ECOBRA/ TRAC prediction of surrp flow behavior seems reasonable." This is reasonable because

Figures 6 and 7 consider a range in [Dowl' of enly C 4 gpm', versus a test range of 9-6" gpm', Since a bias in pretsure is to be applied to improve the ECOBRA/ TRAC test prediction as ncted in p.175, the fact that sump

[ flow]* does monotonicall) herease with a P decrease is adequate agreement for the 0-4 gpm' range in [ flow]*,

With regard to LTCT T2C-418, Rev 2 of the calculation note was issued during the rudit and the basis for the 0.2 psi offset was better documented. It w as review sd with the NRC Inspector dunng the audit.

Steps taken to prevent recurrence:

WCAP-12601 has been expanded to include requirements for calculation notes to provide enhanced documentation of engineering judgment as well as adequate information to indicate how verifiers comments are resolved.

Date corrective action u d preventative measure will be completed:

Completed February 27,1998.

NRC Observation 7:

SSAR GSC-356, "Two inch Breck LOCA, LTC," Revision 0, presented a solution of DP vs. pow in which the author observed that a hermonic oscillation was built in to the soluuw , and therefore, he proposed to take the average value. However, th' inspection team could not determine if the average value was equal to the asymptotic solution had the oscillation not been present. The calculation also did not cddress the impact of oscillation in the asymptotic solutien. the impact of the oscillation on theflow resistance, and the presence of the oscillation in the vesselpow, DP, and vessel collapsed liquid level solutions.

Steps taken to correct this item:

his observation was provided to Westinghouse as FSER Open Item 440.755F in a letter December 17,1998.

Westinghouse provided a response to FSER O! 440.755F in letter DCP/NRCl227 dated January 23,1998. The oscillation occurred in a run as a standalone problem specifically established to characterize the DVI piping resistance between the sump and reactor vesseljunctions. He harmonic oscillation is believed to have been mus 4

Nonconformance 9990040497 02 01 (Continued) introduced by the way in which one or more of the FILL and BREAK components employed was connected to the DVI network. De standalone problem was repeated and the oscillations disappeared. The asymptotic Ocw rate in the new calculation is equivalent to the average value of the solutien obtained when the oscillations were present. ne SSAk cases were reviewed and it was determined that harmonic oscillations comparable to those in SSAR GSC 356 did not exist. Consequently, there is no impact on the SSAR and corrective action is not required, Steps taken to prevent recurrence:

WCAP 12601 has been expanded to include requirement'. for calculation notes to provide enhanced documentation of engineering judgment as well as adequate information to indicate how verifiers comments are resolved.

Date corrective action and preventative measure will be completed: I No corrective action is required. Revision 23 of WCAP 12601 was issued February 27,1998.

NRC Observation 8:

SSAR.GSC 377, "SBLOCA Long Term Cooling," identified discrepancies which included a calculationfor negative ireverst) DVIJlow with no correspondirg physical explanation provided, a two sided open break which did not agree with a two-inch pipe break assumed in the calculation, and discrepancies related to initial conditions assumedfor leakage through ADS 13.

Steps taken to correct this item:

t Dis observation was provided to Westinghouse as FSER Open item 440.756F in a letter December 17,1998 Westinghouse provided a response to FSER 01440.756F in letter DCP/NRCl236 dated January 29,1998. The ADS Stage 1-3 mass discharge referenced occurs during the sump injection phase of two two-inch cold leg break AP600 -

long-te m cooling scenarios. ECOBRA/ TRAC is initialized for these cases with the pressurker empty, which is ,

consistent with the long-term cooling test simulations. During the initial, pre-steady-state portion of the anal)ses, ECOBRA/ TRAC predicts ligt.id to enter the pressurizer, which is consistent with the fact that "the ECOBRA/ TRAC calculation is initially a transient until the mass redistributions occur." The code overrides whatever initial conditions are input to start a window mode calculation to define the solution for the boundary condiuons specified.

The ECOBRA/ TRAC mass balances for the scenarios referenced show that these cases are valid. Considering the two-inch cold leg break window of SSAR GSC-377 reponed in SSAR subsection 15.6.5.4C.3.5, the average total injection flow rate into the vessel through the DVI lines is 76 lbm/second. The average total flow through the ADS Stage 4 flow paths from the hot legs is equal to the injection now. Therefore, the venting capatility necessary for adequate core cooling is provided by ADS Stage 4 operation alone. He mass balance between the DVI and ADS How rates is the same for the other SSAR-GSC-377 scenario referenced.

When liquid is pres ~ent in the pressurizer, the discharge of same through ADS Stages 1-3 is to be expected. Mass is discharged through ADS Stages 13 at an average rate of about 4 lbm/second between 3300 and 4000 seconds of the subsection 15.6.5.4C.3.5 case. The pressurizer behavior is effectively independent of the reactor vessel durir; this time interval because the pressurizer receives no inletf low through the surge line; the now through ADS Stages 1-3 depletes the pressurizer mass inventory.

• '" 5 l

l u

4 Nonconformance 99900404/97 0241(Continued)

The magnitude of flow between the IRWST and/or the sump and the reactor vessel through the DVI lines during long term cooling is determined by the pressure balance that exists. As described in SSAR subsections 15.6.5.4C.3.4 and 5, the upper plenum pressure rises when the liciuid level is high in the upper plenum and the hot legs. More liquid (and less steam)is vented through the ADS Stage 4 flow paths, When the pressure in the reactor sessel upper plenum temporarily increases, injection flow from the IRWST/ sump will diminish until the vessel pressure has been reduced by the venting of higher quality liuid through the ADS Stage 4 flow paths. Check valves adjacen: to the IRWST isolation valves do not permit the flow to reverse direction; flow back it.to the IRWST and/or the sump is possible only if one (or more) of these check valves is assumed to remain open under a negative pressure gradient. For conservatism, the IRWST line check valves are as:umed to remain open in the SSAR long-term cooling analyses so that the potential reduction in core collapsed liquid level due to negative flow in the DVI lines is considered. The AP(/X) design provides adequate venting capacity in the ADS Stage 4 flove paths to assure that the necessary flow from IRWST and/or sump occurs during long-term cooling under conservativa Appendix X assumptions. The l AP600 SSAR long-term cooling analysis subsection demonstrates that the performance of the AP600 complies with ,

the 10CFR$0.46 requirements for long-term core cooling. No carrective action is required.

Steps taken to prevent recurrence:

WCAP-12601 has been expanded to include requirements for calculation no,es to provide enhanced documentation W -ngineering judgment as well as adequate information to indicate how verifiers comments are resolved.

Date corrective action sind preventative measure will be completed:

No corrective action is required. Revision 23 of WCAP 12601 was issued February 27,1998.

f

'I Ahlt es O

r

- Nonconformance 99900404/97 02-02 NL~' Observation It JYCOBRA/ TRAC code error reportfor MOD 7A, Revision I, listed an error affecting timestep control which was not evaluatedfor the specylc case of the AP600 design. In addition, codefailures identified in AP600 calculations were not reported and tracked in Westinghouse's error tracking system.

Steps taken to correct this item:

Westinghouse issued CAR 98-1014 to document the corrective action associated with this finding. As part of the 10CFR50.46 annual report review for 1997, the " error affecting time step control" was evaluated generically for ECOBRA/ TRAC best estimate LOCA applications. De peak clad temperature impact of using analysis restarts was generically identified to be zero *F for best estimate LOCA applications.

De " code failures" mentioned above refer to aborted n:ns. Only those code failures that are identified as cede errors are tracked in the E error tracking system. De code aborts noted in the NRC inspection report are not considered -

errors.

Stepr taken to prevent recurrence:

He error affecting time step has been evaluated for ECOBRA/ TRAC restarts. No further action is required.

Date corrective s ction and preventative measure will be completed:

Ev sluation of the restarts is documented in internal letter SAE-LIS-98-054 dated February 12,1998.

NRC Observation 2:

Over 100 code errors associated with GOTHIC (after Version 4.0) were identified to Westinghouse by the deviloper, Numerical Applications, Inc. (NAl). NAl stated to Westinghouse Inat some of the errors could afect safety determinations and may be reportable under 10 CFR Part 21, Westinghouse could not provide

. documentation to support the review and disposition of these code errors.

Steps taken to correct this item:

Westinghouse issued CAR 97 1332 to document the corrective action associated with this findin:. He error r: port disposition for the NAI notices, that may be reportable under 10 CFR 21, were completed and documented in Westinghouse internal letter NTD-NSA-CRA 94-304, dated December 7,1994. His could not be located at the time of the inspection because of engineers working off site during that week. Here were additional error reports, without 10 CFR 21 implications, that had not been dispositioned at the time of the inspection. All of the error reports were evaluated, dispositioned and documented in internal letter SAE-CRA-97 345 dated December 23,1997 which was issued to the Westinghouse GOTHIC users.

k usn y

m. _ . . _ m. ___ _ . . _ _ _ . _ _ . _

Ni<.Warmance 99900404/97 02 02 (Continued)

Users of the c(vie were requested to;

1. Review the errors in GOTHIC version 4.0 as reported by NAl and confirm the disposaion of those errors.

2. On a return receipt acknowledgment form attached to the letter to citiier A. Document their concurrence with the error disposition, or B. Identify their reason (s) for not concurring with the disposition.

, All authorized users have completed their review and returned their acknowledgment forms. All authorized users concur with the evaluation and disposition of errors in GOTHIC version 4.0 that indicated the errors did not impact

} the containment analysis being performed at Westinghouse.

Steps taken 11 prevent recurrence:

The AP600 Project addressed this issue by providing additional training to the code responsible engineer. The

. training reinforced the requirements of Westinghouse procedure WP 4.19.3 " Software Error Reporting and Resolution", and WP 4.19.4 " External C'mputer Codes".

In addition, the Westinghouse Quality Systems department has initiated a self assessment to determine how the error reports from other third party suppliers are dispositioned.

Date corrective action and preventative measure will be completed:

, The AP600 corrective action has been completed. 'The Quality Systerr: rif assessment will be completed by

March 16,1998.

t 1

I i

h 4

4 4

4 24Niiks 8 1

4

-1 Unrecolved item 99900404/97 02-03 NRC Observation 1:

The main change made to GOTHIC by Westinghouse I make EGOTHIC is thefilm heat transfer package used in modeling the AP600 containment sheil which Westinghouse refers to as the clime model wicich is a large and complicated model .:hange. The inspector reviewed CN CRA 93 219-RO which is the design specification of the chne model. The beginning of the document describes what is called o complete and correct mathematical and physical model of the film energy transport but the equations are not mathematically and physically complete and correct. A complete and correct description would start out with mass, momentum and energy balances on thefilm and then show what terms can be neglected to obtain thefinal mathematical model. Several terms are obviously missingfrom the equations including condeusation and evaporation temts which appear about 60 pages after the original

" complete" model equations are discussed. There are also terms missing that depend on the time rate ofchange of the film thickness that resultfmm the application of Leibni:'s rule to the integral balance equa:ionsfor a moving boundary problem, These missing te*ms may be negligible if thefilm thickness is changing slowly, but the assumptions incorporated into the complete egaations should be clearly stated in the documentation.

In addition, related to Equation 8. an artificial thermal capacitance equal to half the thermal capacitance of thefilm is adh i to the thermal capacitance of the wall node adiacent to thefilmfor numerical stability reasons. Adding this artificial te:m introdu-es an error in energy conservation. The term should either be removedfrom the equations or justification provided that the error introduced is negligible.

Westinghouse Response:

His observation was provided to Westinghouse as FSER Open item 480 lll4F in a letter December 17.1998.

Westinghou: e provided a response to FSER O1480.1114 Fin letter DCP/NRCl221 cated January 20.1998. A complete description cf the governing equation derivation in a single location with assumptions stated has be:n provided in WCAP 14407. Revision 1. The calculation note CN-CRA-93-219-R0 is written for a target audience thst has access to the domentation for previous versions of WGOTillC's, versions 1.0 and 1.1. Hus, the calcolation note provides only the modifications necessary to previously derived governing eqaations. The following provides additional information regardiag several statements in the observation, and a clarification of the numerical method implemented for the liquid film.

De film equa' ion in the calculation note is the energy equation, derived to include terms for convective energy transport and energy storage in the film - the two terms being added to the film equation for the upgrade to

-EGOTHIC version 1.2. Condensation and evaporation terms are in the equation, represented as "Q(Tsurf)", which represents thi. sum of condensation / evaporation, radiation, and convection, which are functions of the surface temperature of the film,Tsurf.

The liquid film is assumed to be a constant thickness during a time step, since its thickness varies slowly with time.

External to containment, the film flow rate is applied as a user inpw and is known to vary relatively slowly with time, relative to the time step size. Interral to containment, the film flow rate results from condensation on the cooler shell surface. For the small time steps used in the code, filn' thickness can a: in be assumed to be constant during a tim:

step. Derefore, there is no need to treat the liquid film as a moving boundary.

He numerical method chosen is justifiable and introduces no errors, as follows. A numerical method is used for solution of the temperature at the discontinuity (where the film and solid surface meet) to provide numencal stability.

The technique employed replaces the finite difference form of the bc,undary condition between two different materials (the heat flux into the boundary is equal to the heat flux out of tk boundary). Instead, a control volume spanning % layer into the film and % layer into the solid is definea, and a control volume equation is derived. The control volume equation is correct as derived, and results in an explicit representation of the energy storage in the

, mt.m 9

Unresolved item 99900404/97 02 03 (Continued) water and solid in the control volume, which models the damping of the boundary temperature during rapid trannents, giving increased numerical stability, ne value of the boundary tempers ure is solved by using the control volume eqt.ation, which includes the energy storage in the adjacent % layers. Temperatures for nodes away from boundaries ase solved by using the finite difference governing equation, avhich accounts for energy storage in the %

layers adjacent to eac6: node.

He term " artificial" in the calculation note is used to identify the difference between the traditional finite difference boundary condiaon used between dissimilar materials (Equation 2, repeated below) and the finite volume equation (Equation 8, repeated t,elow) which replace., the governing equs; ion and boundary r.onditioa for the boundary node.

TLu - T' a _ T,,, - TLa

-u su l

u wit &pk

~

k &nw~dTL, TL, - Tha TLu - T,,,

P-u C,,a 3 + p ,,,, p C,,,, 4 ,, = k_a 4 pu 8

, &p An apprcach based on the boundary condition specified in eqcation 2 would use the governing finite difference equanon where the energy stcrage term for % layers adjacent to the boundary appear. An approach based on equation 8 replaces the govermng equati a and its boundary condition at the interface with equation 8 where the energy storage term is explicitly included. Either appmach is mathematically and physically correct, and the one .

implemented in WGOTillC provides for numerical stability. Hus, the approach does not represent an error.

NRC Observation 2:

CN CRA 95 089," Validation and Verification of INOtR Small Internal-Use Computer Prograr " Rev. O and I.

The document disclosed that the author's response to the reviewer comment concerning an incorrect valuefor an area as used in the calculation was not considered to be ofsuficient consequence to warrant a code revision. No spectfic evaluution to support the conclusion was provided and the inspector could not assess the impact without recourse to the origutator. It was also noted by the author that ifother changes werefound to oe necessary then this error should be corrected.

Westinghouse Response:

This concern was originally identified at a Westinghouse inspection readiness review. He document wa: revised prior to the November inspection to provide the additional documentation requested by the inspector. The revised calculation Gowed that there was no error in the origind document. Westinghouse believes there are no more open or unresolved issues associated with this calculation.

2i42 "

10

.3 ,

d Unresolyed Item 99900404/97 02 03 (Continued)

' NRC Observation 3:

CN CRA-94147, " Phase 2/3 Large Scale Test Lumped ParameterYGOTillC Base Case Deck. " Revisions O and I, The review of this document disclosed that the reviewer's comments and the author's responses u ere imbedded in the document. In general, the autho,'s responses werefound to include sufficient details to assess the responses. The mspection team noted that errors in the model werefound aft;r the computer analyses had been performed. In one case, the author's response indicated that a k lossfac*or was likely acceptable if other specific conditions were met. There was no statement as to the expectation of the condition being met or vertfled. Westinghouse Response: Westinghouse reviewed both revisions of the calculation CN CRA-94-147 and established that the verifiers comments expanded the applicability c/ithe K factor over i greater range of Reynolds numbers and no further evaluation was required. Westinghouse believes there are ro more open issues associated with this calculation. NRC Observation 4: & CN TA 96153 AP60C "Steamline Break Mass and Energy," Revision 0. The review disclosed that errors existed in the analyses that were identified after the computer runs had been completed. The inspection team determined that some errors would be conservative and that some would be non-conservative. The author's assessment was that there was no impact associated with the errors. It was also noted that the specific errors did not occur in the limited analyses that support desi.en certification. The inspection also disclosed that the computer program used to calculate the SSAR mass and energy releasesfor the steamline breaks is LOFT 4AP Version I.8 and that the values presented in the SSAP are consistent with this supporting calculation. Westinghouse Response: Westmghouse issued CAR 98-1019 to document the corrective action associated with this finding. Included in CN-TA 96-153 are the following verifier comments: Comment i Two data are wrong in the hot zero power deck without impact on the final result (table in paragraph 2.5.2.1.1.) Pnmary average temperature : 551.3 is implemented in the deck and not 551.5 *F Initml enthalpy of the CMT balance line : 543.56 is used in the runs and not 548.40 BTU /lbm. Mm.ent 2 The manual actuation of the PRHR for the case 30% power and a break size of 0.1 ft* is wrong (see table in paragraph 2.5.2.1.2). The PRHR should be actuated at 77.5 seconds. During the transient the PRHR n injecting 7 seconds sooner at 70.5 seconds. This does not have an impact on the final result. Commenh Some errors were found in the feedwater flow calculation table for the full DER The were corrected in the tables but not in the input data. Comment 4 The energy flow rate for the 30% power level is wrong. It should be : 1193.2

• 9636.85 =

11498689 BTU /sec. The value u<cd in this analysis is higher (1150a664 BTU /sec) so that it is conservative for the energy result. No rerun is necessary. w:n lg

Unresolted item 99900404/97 02-02 (Continued) Each of the verifier's comments is discussed below. It should also be noted that the limiting cases of CN-TA-96-153 have been reanalyzed (for other purposes) and are documented in CN-CRA 97-13. Ccmment 1 Two issues are identified; a) It was desired that a value of 551.5 7 and not 551.3 7 be used as the average RCS temperature for hot zero power analyses. The analyses inadvertently used a value 0.2 T lower than was desired. It should be noted that the AP600 design no load average RCS temperature is 545" Current steamline break methodology as applied to licensed operating plants, performs the HZP analyses at the design no load temperature. Uncertainties are not applied to the initial RCS temperature at HZP conditions as they are for at power cases, in the course of performing the APbOO analyses several times, the no load operating temperature has been modifiH Rerefore, to minimize possible reanalysis if the design no load temperature was changed from 545 T, the at power temperature uncertainties ( 7.0 to +6.5 T) were also applied to he HZP average RCS temperature. Using this band around the design no load temperature is expected to prevent or minimize ! reanalysis if the no load temperature is changed. He analyses used an uncertainty cf -6.3 7 instead of the +6.5 7 as was desired. b) initial enthalpy cf the CMT balance line : 543 56 BTY/lbm is used in the runs and not 548.40 BTU /Lm. He steamline break analyses is performed at several power levels (102%,70% 30% and 0%). When setting the initial conditions 'nr each power level, the analyst must also manually set the fluid temperatures in the unisolated portions of the CMT connection lines which may be affected by the change in the initial , reac*or coolant temperature. His is done iteratively by making a short LOFTRAN run and observing the calculated initial cold leg temperature and enthalpy. In the fina! run the enthalpy (temperature) of the unisolated portions of the CMT connections lines shNid be set based on the reactor coolan: mperature. The analyses used the slightly colder 30% power value for enthalpy instead of tne 0% power value. His input affects the initial starting CMT recirculation flow. However, once the CMTs begin injecting, the lines are quickly swept clear of the original Duid within a few seconos and is replaced by Guid from the RCS and the CMT and has no further impact on the results. The volume of the unisolatable nodes in question are only e few tenths of a percent of the CMT loop overall volume. His was caused by a typographical error and an oversight of the analyst. The analyst derived the appropriate temperature but did not update the inpu* w the code. For the reasons stated above the impact on the results is inecusequential. Comment 2 , The PRHR is actuated 7 seconds earlier than it should have been denng the 0.1 ft' steam 1:ne break from 30% power case. As documented on page 143 of CN-TA-96-153, during the 0.1 ft' steam line break from 30% power case, the steam generators dry out, and the releas ; art terminated after 9194 seconds. During the event the integrated releases are 229800. Ibm and 273700000. BTU. The PRHR has no ir, pact on the mass and energy releases if the reactor returns to power and only a small benefit if the reactor does not return critical. If the reactor is critical the core power 2*2 " 12 l _ 0

    ~

4 Unresolved Item 99900404/97 02 03 (Continued)

      . increases to supply the heat load of the PRHR and the energy transfer to the faulted steam generator is not          a significantly impacted. If the reactor is not critical the PRHR will remove a small portion of the energy from the RCS which would otherwise be transferred to the steam generator and released from the break, Since the PRHR heat transfer area is - 4300 ft' and the steam generator is 75000 ft', the steam generator dominates the event.

This integrated energy of the PRHR over 7 seconds is scry small when compared to tue energy releasec. through the break during the event and has an insignificant impact on the results. Based on this evaluation, the impact on the results is inconsequential. Comment 3 For the purposes of steam line break mass and energy release calculations, it is conservative to use high values of feedwater flow to the faulted steara generator. Evc.y pound of water entering the steam generator eventually becomes a pound of steam in the containment. During a steam line break, the steam flow from the steam generators increase and the steam generators depressurize. De feedwater control system uses a two element controller which tnes to match feedwater flow to steam flow and to maintain programmed steam generator level. To conservatively approrimate the performance of the feedwater control system it is assumed that the control system fully opens the feedwater control valve on the faulted steam generator in an ettempt to match the steam flow from the SG. With the feedwater control valve fully open the feedwater flow is controlled by Ue head.Dow characteristics of the feedwater pumps. The feedwater pumps are not explicitly modeled in LOFTRAN and feedwater flow must be calculated manually and entered as a table mto the code. Tables of feedwater How as a function of.; team generator back pressure are supplied by balance of plant designers. As input LOFTRAN accepts table of feedwater flow as a function of time. An iterative process is used to define the feedwater flow. A test LOFTRAN run is made with estimre feedwater How rates to calculate the time dependent steam n~or pressures. Using the steam generator pres.ures calculated by LOFTRAN, new feedwater Oow rates are ..J and input to LOFTRAN again. During the course of the review, three arithmetic errors were noted in the hand calculations for feedwater flow. One of the errors resu'ted in feedwater flow being 1% conservatively higher. The other two c.rithmetic errors affected the hand calculated feedwater flow after the time at which main feedwater is isolated. The comment resulted from an arithmetical error in the conservative direction and had no impact on the results. Comment 4 Depending upon the break size and location, the LOFTkAN model does not simulate the blowdown of the main steam piping. To account for the piping blowdown, hand calculations are performed. The LOFTRAN calculated steam llows from the exits of the steam generators are adjusted by the results of the hand calculations to properly account for the piping blowdown. In calculating the reverse flow from the pipmg, an anthmetical error resulted in an energy flow of 11504664 BTU /see being used instead of 11498689 BTU!sec. This value is +0.05% high in the conservative direction (i.e. it increases the energy release to containment). S'"

• 13

'Y v + ~

Unresolved Item 99900404/97 02 03 (Continued) 4 NRC Observation 5: 1100-SOC 001, " Containment Volumes and Heat Sinks." Revisions 0 through 4 s The review of Revisions 2 through 4 disclosed that the development of the containment volumes and heat sinks were d:veloped and updated based on the nuciear island generalarrangement drawings starting with Revision 6 and ending with Revision 8. Rev. 0 nas a preliminary scoping document that was completely revised in Rev. 2. No . specific review comments were identified however each page contained a sign of block with the author and verifier (reviever's) signatures. Minor deviations were identified such as use of " estimated" and " assumed" dimensions for non-crtuc : components and diagrams without units or dimensions clearly identitled. With respect to SSAR 6.2.1.3 and 6.2.1.4, Westinghouse needs to:

a. Evaluate the significance of the insulation on thefree volume used to determine the peak containment pressure. ,

Provide adequateJustification, as appropriate, that thefree volume is conservative. '

b. Evaluate the significance of the insulation on theflow path characterizations used to determine the peak containment pressure, includingflow arcas andform losses, for both paths connecting below operating deck compartments as well asflow p 2ths connecting below operating detk regions to above operating deck regions. Assess the efectsfor each of the 4 LOCA phases as well as the MSLR.

1 With respect to SSAR 6.2.1.2, Westinghouse needs to:

c. Evaluate the significance of the insulation on theflow path characterizations used to determine the diferential pressures across subcompartment walls, includingflow areas andform losses,for both paths connecting below operating deck compartments as well asflow paths connecting below operating deck regions to above operating deck regions.

J. With respect to theflooding issues, Westinghouse needs to: Evaluate the signQcance of the insulation on compartmentflooding, address both timing and levels, as well as which compartments would be affected. Westinghouse Response:

         - This observation was provided to Westinghouse as FSER Open Item 448.lil2F in a letter December 17,1998.-

Westinghouse provided a response to FSER OI 480.1112F in letter DCP/NRCl236 dated 1/29/98. Calculation Note 1100-SOC-001 Rev. 6. " Containment Volume. and Heat Sinks," issued January 8,1998, documents the basis for neglecting the insulation on piping and equipment when calculating free volume for containment pressure calculations, and shows that the free volume calculation is acceptable. Westinghouse has also confirmed that insulation was considered in calculations of flow path charactenstics for containment pressure and subcompartment differential pressure input.- Response to (a) The presence of metallic reflective insulation can be neglected for the calculatiot of free volurre. In addition, the free volumes used in peak pressure calculations have been ecnfirmed to be acceptable (Calculation Note i 100-SOC. 001), consequently no change to the evaluation model volume input is required. P

             !e    %
                                                                                ,    s                        .

4 Unresolved Item 99900404/97 02 03 (Centinued) Response to (b) The presence of insulation was considered for Dow path calculations in the prior versions (Revs 0-5) of Calculation Note 1100-SOC-001, no related flow path changes in the evaluation model inputs are necessary. Response to (c) De geometry used in the calculations to determine the differential pressures across subcompartment walls accounts for the presence of insulation. Response to (d) ne presence of metallic reficctive insulation can be neglected for calculation of free volumes used in the evaluation of flooding levels within compartments and in the identification of which compartments may be subjected to flooding since the effective value is small relative to the Dooded value. NRC Observation 6: CN-CDBT 92 233, "AP600 EGOTHIC Input Deck Development," Revisions 2 and 3. { The review disclosed that errors existed in the ana!yses that were identified after the computer runs had been completed. The specific errors identsfied by the reviewer in Revis..n 3, dated May 22,1997, concerning errors in areas and k-loss fac >rs, were determined to have negligible impact on the analyses and therefore reanalyses were not performed. No discussion un how this conclusion was reached was provided. As a result. Westinghouse needs to:

a. Providejustufication that the impact of these errors and incorrect loss coef.clents are conservative, or tk2t the cumulative impact of known errors would nm result in a change in the pressure calculation greater than 0.2 psig. the lossfactors selected were considered to be applicable to " natural circulation" but,for theflow paths in question. the lossfactors should not have been applied. Assess the efectsfor each of the 4 LOCA phases as well as the MSLB.

b. Providejusttficationfor allowing knov:n errors to remain in the licensing anasyses that support design certtfication. Include the supporting knowledge base employed by Westinghouse that is used to assess errors to determine that, in consideration of the 0.2 psig (H of1%) margin in the calculated allowance to the design i

pressure, known errors have a negligible impact (for example only conservative errors remain, or that the cumulative impact of known errors would not result in a change in the pressure calculation greater than 0.2 psig). Consideration should be given to both accumulation of errors as well as the impact of errors in consideration oj the diferent phenomena and characteri:ationsfor each of thefour LOCA phases and the MSLB. Westinghouse Response: his observation was provided to Westinghouse as FSER Open Item 480.ll l3F in a letter Decemtser 17,1998. Westinghouse provided a response to FSER O! 480.1113F in letter DCP/NRCl236 dated 1/29/98. For the containment pressure Design Basis Analyses using WGOTHIC, Westinghouse has cor ected the flow area and loss coefficient values. Known errors have been corrected in the WGOTHIC evr.luation model used to calculate the peak containment pressure. The revised results, to be incorporated into SS AR subsection 6.2.1, Rev. 21, show a lower peak pressure than previously calculated.

me 4.

i$ j l

l l ' ( l Attachment 2: AP600 NRC Design Assurance Review Plan 1 B' k e k

              !! AIM 14%

                                                                                                                          /     .\

( )

                                                                                                                           'Q,,/

9 DCP/DCP0343 From New Plant Projects Division I WIN 284-5390 Date February 10,1998 Subject AP600 NRC Desien Assurance Review PIAD To D. N. Alsing EC W5-12 E. H. Novendstern - EC E4 26 J. A. Gresham EC E4-26 R.S.Orr EC E3-05

• M. Mahlab - EC E3-07A R. B. Tupper - EC E3-08 P. R. Mandava - EC E3-05 R. P. Vijuk - EC E3-06 B. A. McIntyre - EC E3-09 J. W. Winters - EC E3-08 cc W. E. Cummins - ECE3-07 M. Mutyala - EC W5 10 S. D. Rupprecht - EC E4-16 he NRC AP600 Final Inspection identified concerns about the effectiveness of Westinghouse's review of the design calculations. To respond to NRC coacerns,it is necessary to take additional steps to establish that the Westinghouse design documentation meets the 10CFR50 Appendix B, design verification and quality asautance requirements. A Design Assurance Review is being initiated to assess a sample of AP600 documentation that suppert SSAR Chapter 6 containment analysis and Chapter 15, Accident Analysis.

A Review Team, composed of technical experts independent of the calculations being reviewed, has been established to perform the review. Dr. William LaPay will serve as the team leader. He technical experts will perform an independent review oi a sample of AP600 calcula' ion notes for accuracy and documented verification. This will include a review of documented calc.ilations and interviews with the author and/or verifier as needed. Enclosed herewith is an outline of the Design Assurance Review Plan. The Team will issue a final report at the end of theit review. Jim Winters is responsible to manage the interface with the Team and the completion of corrective actions resuiJng from the Team's review. Please provide the Review Team your full cooperation and support. Origmal Signed By, R. M. Vijuk, Manager AP600 Projects ,

        /At'achment mwa

Westinghouse Raylew and Verification of Desig;n Documentation AP600 PROJECT Design Assurance Reslew Plan OR.IECTIVE: Assess the adequacy of the AP600 the safety analysis documents supporting the SSAR Chapter 15 and the por ion of Chapter 6 on containment analysis to satisfy the design certification provisions of Appendix B in10 CFR Part 50. B)CUS OF REVIEW: Satety related calculations that form the basis for Chapter 15 and the portions of Chapters 6 related to containment analysis of the SSAR. AREAS TO PE REVIEWED: A representative sample of AP600 r .!culations will be selected. Three distinct areas will be evaluatea which include the plant configuration input to the analysis, the analysis itself, and V&V, REVIEWERS: Leader: William LaPay LOCA (Large and Small Break) Mitch Nissley, Tim Andreycheck, Non-LOCA Analysis Steve Love Containment Systems Bob Jakub, Niclr Trikouros (EPRI/GPU) Long Term Cooling John Spaargaren, Dan Golden Radiological Bob Lutz, Stan Anderson Team Support Ted Batt (V&V), Jeff Himler (Mechanical), QA Bob Tupper (AP600 Interface) REVIEW PROCESS:

1. Develop a sample selection based on the matrix shown on Figure 1.

2. Establish the review criteria

3. Perform Calculation Reviews 4 Characterize and document review results PLANiSCHEDULE:

Selection of team /sar ,' . .s February 9,18 Review Start ' ebruary 10,1998 Review Complete March 20,1998 Project Comment on Review March 27,1998 issue Assessment Report April 3,1998 2'* 2

4 CORRECTIVE ACTIONS: Complete by April 30. ,1998 Figure 1 Calculation Sample Areas Analysis Area input Analysis Code V&V Large Break LOC \ X Small Break LOCA X X X Von LOCA X i Containment X X X Radiological X Long Term Cooling X X h IIA 2df des

                                                                                                   }

Attachment 3: Summary of Assessment to Determine Root Cause of NRC Concern on Documentation Detail An AP600 self assessment was performed to determine the root cause for the NRC concern with the level of documentation supporting engineeringjudgments in calculations reviewed during the NRC Inspecuan documented in Report No. 9990(M(M/97-02 dated January 26,1998. The NRC Inspection Report cites examples of design calculations which were considered by the inspection team to contain discrepancies or errors without documentation of an adequate evaluation by Westinghouse. The focus of this assessment was to establish the root cause for the lack of documentation by the Westinghouse authors and/or verifiers. Interviews were conducted with eleven authors and/or verifiers of approximately 14 calculation notes that had been reviewed by the NRC during the inspection. The interviews were conducted within the framework of a root cause evaluation using a combination of the "5 Whys" technique and engineering judgment. Observations from 'he interviews with the authors and/or verifiers included:

• The authors and/or verifiers were farmliar with the requirements for design analyses documentation detail as 3 required by ASME NQA-1 and the QMS:

                          "They (design analyses) shall be ssliciently detailed as to purpose, method, assumptions, design input, references, and unit; such that a person technically qualified in the subject can review and understand the analysesand venfy the adequacy of the results without recourse to the origint or". (ASME NQA 1)
                           " Design analysis documents are legible, reproducible, and describe the purpose method, assumptions, design input. and references such that the analysis can be reviewed and venfled by a person technically qualified in the subject without recourse to the preparer'. (QMS) e The authors prepare their analyses with the understanding that the " person technically quahfied in the subject" is another member in their functional area experienced with the Westinghouse design and methodology. Both                   '

authors and venfiers considered that most of the judgments reflected in the cale notes were intuitively obvious to those familiar with the work product. Many authors and/or venfiers agreed that in retrospect, given that this was a development project and subject to third party reviews, more documentation detail could have been provided in some cases. Some authors and/or ve"fiers reported that because of the developmental nature of the program, the continuing flow of changes, and the knowledge that analyses were going to be revised to accommodate future changes, the correction of " insignificant" errors was documen'ed for future revisions of the document, while significant errors, if any, were corrected. This assessment concluded that the rom cause of the lack of documentation supporting Westinghouse judgments is that the analyses were basically bemg prepared for internal review and verification by other qualified Westinghouse personnel. That the senfier had sufficient information to enable verification to be completed in accordance with the governing procedures as evidenced by the venfiers signature, indicates that the author provided sufficien. documentation de' ail. Some of the authors and verifiers were also influenced by the knowledge that as a development project, future revisions would afford an opportunity to clean up"insignificent" errors. mn s

   .                                                                                                                         t Recommendation for management consideration as a result of this assessment:

In c d:t to further assure that documentation detail provided in Westinghouse analyses is sufficient for the purpose intended, it is recommended that assignment of future work include guidance in terms of the audience for which the work is being prepared, if other than the internal Westinghouse verifier. This should be accomplished in the design planning phase. WCAP 12601 would be an appropriate documentation of the guidance for the AP600 Project.

                                                                                                                             .'1 JIA24f es
                                                                        }
                                  *

• ____J}}