ML20199F206

Summary of 971112 Meeting W/Owners Group Members,Nuclear Energy Institute & Industry Officials Re Review of Responses to GL 92-01,Rev 1,Supplement 1.List of Meeting Attendees, NRC Presentation Viewgraphs & Addl Staff Guidance Encl

ML20199F206
Person / Time
Issue date:	11/19/1997
From:	Wichman K NRC (Affiliation Not Assigned)
To:	Sullivan E NRC (Affiliation Not Assigned)
References
PROJECT-689 GL-92-01, GL-92-1, NUDOCS 9711240146
Download: ML20199F206 (63)

Text

. _ _ _ _ _ __. _ _ . _ _ ._ _ _ _.__ _ _ _ _ _ _ _ _

j#%4 I -

' ' p , UNITE 3 STATES _

4

.T NUCLEAR REGULATORY COMMISSION-L 5 L wAsumerou, o.c. mem emn o

%%go 1November 19, 1997 MEMORANDUM TO: Edmund 1 Sullivan, Acting Chief I

- Materials and Chemical Engineering Branch

,. Division of Engineering FROM: 1 Keith R. Wichman, Chief.._

qJ_ Section A =

Materials and Chemical Engineering Branch L Division of Engineering

SUBJECT:

• MEETING

SUMMARY

FOR NOVEMBER 12,1997 MEETING WITH OWNERS GROUP REPRESENTATIVES AND NEl REGARDING -

REVIEW OF RESPONSES TO GENERIC LETTER 92-01, REVISION 1, SUPPLEMENT 1 RESPONSES On _ Wednesday, November 12,1997, several members of the U.S. Nuclear Regulatory Commission (NRC) staff participated in a public meeting with owners group members, a Nuclear Energy institute (NEl) representative and other industry officials at the NRC's

Rockville, MD, Headquarters to discuss the status of the NRC staff's review of responses to Generic Letter (GL) 92-01, Revision 1, Supplement 1. Attachment 1 is a list of the meeting -

participants. Attachment 2 is the NRC staffs' presentation view graphs. Attachment 3 contains additional staff guidance on issues discussed at the meeting.

L The issues discussed at the meeting included 1) the determination of weld wire heat

• U best-estimate chemistries,2) the evaluation and use of surveillance data,3) a brief discussion

• on variability in plant-specific initial reference temperature values for reactor pressure vessel (RPV) welds from the same weld wire heat, and 4) industry maintenance of the reactor vessel integrity database (RVID). Attachment 2 is the detailed presentation material, most of which prompted significant discussion among the participants.

The following issues were raised during the meeting:

- Exclusion of data in Ma* ==*imate chemistrv determination J

The Combustion Engineering owners Group (CEOG) submitted a report in July 1997 which >

described their efforts to obtain best-estimate chemistry values for each heat of weld wire in CE

. fabricated RPVs. In addition, the staff conducted an inspection of_Babco:k & Wilcox

. _ (Framatome) weld fabrication records in May 1997. Data were excluded for some of the weld wire heats based upon inconsistency with expected norms. The staff believes that exclusion of i data should be based on a statistical test for outliers as well as a physical reason for believing the data to be atypical '

1. <

In response to this statement, a CEOG member noted that a copper (Cu) measurement may appear to have been taken from the heat-affected zone (HAZ) of a material, however, this could i

- be a judgement call since it would not be possible to go back and sample the material. The ll...ll lllllllllll1.1

~

mp M*ns TJRL PM W W CBUER COPY

i :1 .

+ j p'  ::

a l

guidance from staff was that a licensee could, for example, submit a letter which indicates . l' that a physical basis for exclusion of data would not be possible, it should be noted that -

inclusion of the data that were excluded in the CEOG report does not, in general, have a ~ -

significant impact on the CE weld wire heats.

Best fit hne through surveillance data Consistent with the equations in Regulatory Guide (RG) 1.99, Revision 2 and 10 CFR 50.61 (pressurized thermal shock rule), the NRC presenter stated that zero is the appropriate -

y intercept for a best fit line through surveillance data (plot of ART,ev as a function of fluence l factor).' There were industry oMicials that disagreed very strongly with this conclusion. See Attachment 3 for further details.

Non-conservative table chemistry factor r

The NRC presenter described an example where the surveillance data are not credible, and the '

RG 1.99, Revision 2 table chemistry factor (CF) is non-conservative (i.e. the measured values from the surveillance data are greater than the projected mean plus two standard deviation values from the RG 1.99, Revision 2 tables). The staff guidance in this case was to use the more conservative CF from the surveillance data with the full margin term. An owners group member stated that he would consider other information (e.g. the correlation mnnitor material) before deciding which CF to use. There was also an inquiry as to the regulatory need to use the CF determined from non-credible surveillance data when the Table CF is non-conservative.

See Attachment 3 for further staff guidance.

Irradiation environment adiuntments With regard to temperature adjustments between surveillance specimens and the vessel being assessed, the NRC presenter stated that studies have shown that for temperatures near 550*F, a 1 *F decrease in irradiation temperature will resuit in approximately a 1'F increase in ART,e7 A owners group member asked for the references for these studies. The references are included in Attachment 3.

Initial reference temnerature issue for weids There was a very brief discussion on the initial reference temperature (RT,ory) issue for welds.

The NRC presenter highlighted examples of widely varying plant-specific RT, erg values within the same weld wire heat (e.g. 80*F difference in values), The presenter also highlighted the number of plant specific and generic values that are currently in the RVID. The staff noted that t

it may pursue significant plant-specific RT,.ru issues on a case by case basis Evaluation of-- - -

plant soecific RT,,,, values was characte-ized as an item which could lead to consideration of rule changes.

t-r

.. - - ,- + r ma g-a m e- 1 =s -- %-My- = am-1 mg+m 4ye-sy +*pa-W y. yk m y-+i- >=w

i i i g .

3 l Industry maintenance of the RVID The staff noted that the roftware for the RVID has been transferred to the Microsoft Access" forrnat, and the database application is in the trouble-shooting and maintenance phase.: -

Enhancements have been made to the RVID as a result of feedback from the industry training ,

session that was held at Headquarters in May 1997. The plant data will be updated after r licensees have assessed, where applicable, the issues that were presented during this meeting.

There was some discussion on the fact that the RVID uses the calculational procedures in RG 1.99, Revision 2 to evaluate docketed input values. The new version of the RVID has sections called " input data" and 'ca!culated data" which are clearly delineated. The staff suggested that the industry consider the pros and cons of maintaining an auditable version of ..  ;

the database.

Letters to licensees and owners groups in the interest of closing out GL 92-01, Revision 1, Supplement 1, the staff will issue letters to each licensee whose vessel material or surveillance program material best-estimate chemistries may be affected by the information from the CEOG and Framatome activities.

The letters will ask licensees to review the information, including those issues raised by the staff regarding them, and submit to the staff any necessary revisions to their facilities' licensing-basis information. These letters will also include a request for licensees to consider surveillance data -

issues in updating RPV integrity evaluations. The staff also plans to issue letters to the CEOG and Frematome, as necessary, regarding resolution of the issues with data exclusion, outliers, and generic values.

- The staff closed the meeting by re stating that the intent of GL 92-01, Revision 1, Supplement 1 was to be consistent with the existing regulatory frame-workc However, the staff presented -

examples which addrcss cases where detailed guidance is not outlined in the regulations. The staff also pointed out that these issues need to be resolved before looking ahead to advanced approaches that are currently being developed.

Attachments: 1, Meeting Participants

2. View graphs

3. Additional guidance on issues discussed at the meetir.g cc: See attached page a

.+

8 J

P

,. > ,s-,+ .. , , ,. r+ - - -

't 3

-1 Industry mainianance of the RVID The staff noted that the software for the RVID has been transferred to the Microsoft Access I j . format, and the database application is in the trout;le-shooting and maintenance phase.

Enhancements have been made to the RVID as a result of feedback from the industry training

- session that was held at Headquarters in May 1997. The plant data will be updated after.

licensees have assessed, where applicable, the issues that wsrs presented during this meeting.

There was some discussion on the fact that the RVID uses the calculational procedures in RG 1.99, Revision 2 to evaluate docketed input values. The new version of the RVID has sections called " input data' and " calculated data' which are clearly delineated. The staff suggested that the industry consider the pros and cons of maintaining an auditable version of the database.

Letters to licensees and owners groups in the interest of closing out GL 92-01, Revision 1, Supplement 1, the staff will issue letters to each licensee whose vessel material or surveillance program material best-estimate chemistries may be affected by the information from the CEOG and Framatome activities.

The letters will ask licensees to review the information, including those issues raised by the staff regarding them, and submit to the staff any necessary revisions to their facilities' licensing-basis information. These letters will also include a request for licensees to consider surveillance data issues in updating RPV integrity evaluations. The staff also plans to issue letters to the CEOG and Framatome, as necessary, regarding resolution of the issues with data exclusion, outliers, cod generic values.

N &.Qff closed the meeting by re-stating that the intent of GL 92-01, Revision 1, Supplement 1 was to be consistent with the existing regulatory frame-work. However, the staff presented examples which address cases where detailed guidance is not outlined in the regulations. The 2

staff also pointed out that these issues need to be resolved before looking ahead to advanced approaches that are currently being developed.

Attachments: - 1. Meeting Participants

, 2. View graphs -

3; Additional guidance on issues discussed at the meeting cc: See attached page "

. Distribution: See attached page DOCUMENT NAME: G1EE\MTGSUM.WPD To receWe a copy of this document, inecate in the box C= Copy w/o att9, ghment/endoeure E= Copy with attachment /endoeure N = No copy / l\ .

EMCB_ l E, EMCB lE EMQB$l 7 lE '

AteeM BElliot h>f KRWietGhan 11( M T 11/F/97 11#Yl97 1

wrv. -e s, -r.s.-,<,# -,y *r---,---w.- --w--w----w,e w-e w*.s.5<- * - - - - ,w -

e-o-=--,,-

w.

cc:

1 Project Number 689 Nuclear Energy institute -

Kurt Cozens NEl 17761. - Street, NW Suite 400 Washington D.C. 20006 3708 DlSTRIBUTION:

( W 100 FILE CENTER >

SJCollins -

FMiraglia BWSheron GLainas JStrosnider CCralg >

SMagruder JWilson JBirmingham MMalloy MMitchell .

KKarwoski MKhanna SSheng JMedoff MMayfield EHackett fDb I I Pre)-Ln?

m  ; 0: .

,b .

i

-O MEETING PARTICIPANTS l

MEETING TO DISCUSS STAFF REVIEW OF RESPONSES TO GENERIC LETTER 92-01, REVISION 1, SUPPLEMENT j NOVEMBER 12,1997 bl6Mfi ORGANIZATION Ted Sullivan NRC/NRR/DE l Bob Carter EPRI Bill Server ATI Consulting i Tom Spry Commonwealth Edison T. Satyan Sharma AEP Sam Binger WOG Projects Steve Cc"*rd FPL Steve Byme ABB/CE Bob Hardies BGE Lee Abramso". NRC/RES/PRAB George L. Lehmann GPU Nuc. ear Dana Covill BWRVIP Assessment, GPU Nuclear Kurt Cozens NEl Stan T. Rosinski EPRI Keith Wichman NRC/NRR/DE Cayetano Santos NRC/RES/DET Edward C. Wenzinger Sr. NUS Information Services BnJee A. Bishop Westinghouse NSD Thomas J. Laubham Westinghouse NSD Ed Terek Westinghouse ATTACHMENT 1

i i l i

l l

MEETING PARTICIPANTS (ATTACHMENT 1 - CONTINUED)

NAME QfMLtilZATION _

Terry Sides Southern Nuclear  ;

1 Chades A. Tomes WPSC Roy Waterman NSP H. F. Conrad NRC/NRR/DE Meena Khanna NRC/NRR/DE i Michael Mayfield NRC/REWDET John R. Harrell VA Power William Brady Duke Energy Don Howell _

Framatome Technologies Matthew DeVan Framatome Technologies James Pfefferle Wisconsin Electric Power Company Deann Raleigh SEARCH Group Bechtel Alan R. Stalker Carolina Power & Light Company l

Simon Sheng NRC/NRR/DE James Medoff NRC/NRRIDE Barry J. Ellio' NRC/NRR/DE Ken Karwoski NRC/NRR/DE Andrea Lee NRC/NRR/DE Matthew A. Mitchell 'JP,C/NAR/DE

. - - --- . - - . . . a. - - ..n; - a u - - -.. --- ------ - - - .-- -- -- . .a h

l l

l 1

NRC INDUSTRY MEETING on 1 STATUS OF GL 92-01, SUPPLEMENT 1 NOVEMBER 12,1997 ROCKVILLE, MD. l 1

l 1

ATTACHMENT 2

l I

PURPOSE OF MEETING

• TO DISCUSS:

- STATUS OF GL 92-01, SUPP 1 AND

- ACTIONS NECESSARY TO CLOSE OUT THE GL

-ISSUES RELATED TO:

Best estimate chemistry Use of surveillance data Initial RTndt

-SCHEDULE

- RVID l

l

4 STATUS OF GL 92-01. SUPPLEMENT 1 CE OWNERS GROUP e SUBMITTED GENERIC REPORT IN JULY 1997 e LICENSEES COMMITTED TO ASSESS IMPACT OF REPORT Responses outstanding B&W OWNERS GROUP e RESPONDED THAT NO ADDITIONAL ACTIONS WERE NECESSARY e NRC RESPONDED THAT LICENSEES NEED TO ASSESS USING RATIO PROCEDURE No acceptable siternative proposed Assessments still outstanding e FRAMATOME INSPECTION FINDINGS f -

Reported to licensees in letter from Framatome Report to be issued soon Findings need to be assessed by licensees BWR OWNERS GROUP 1

e COMMITTED TO GENERIC REPORT IN DECEMBER 1997 Licensees will need to assess generic report b

SCHEDULE -

e 12/31/97: STAFF ISSUE RAls TO CEOG AND B&WOG LICENSEES e 02/15/98: STAFF ISSUE RAls TO BWROG LICENSEES e 03/31/98: LICENSEES RESPOND TO RAls e 06/30/S8: STAFF ISSUE FINAL CLOSEOUT LETTER TO LICENSEES

• 12/31/98: STAFF ISSUE REVISED NUREG 1511 AND RVID 4

i l

l

-h

MATERIALS ISSUES AFFECTING RPV INTEGRITY ASSESSMENTS ISSUE 1.) THE DETERMINATION OF WELD WIRE HEAT BEST-ESTIMATE CHEMISTRIES THE EVALUATION AND PROCESSING OF DATA

• CALCULATION OF A BEST-ESTIMATE AND THE USE OF GENEP. :

VALUES lSSUE 2.) THE EVALUATION AND USE OF SURVEILLANCE DATA THE ASSESSMENT OF CREDIBILITY NECESSARY DATA ADJilSTMENTS (lRRADIATION TEMPERATURE, MATERIAL CHEMISTRY)

THE USE OF IRRADIATED DATA FROM OTHER VESSELS ISSUF 3.) VARIABILITY IN PLANT-SPECIFIC INITIAL REFERENCE TEMPERATURE FOR RPV WELDS FROM THE SAME WELD WIRE HEAT

• OBSERVED VARIABILITY LONG TERM ASSESSMENT ACTIVITIES Q.

STAFF lSSUES REGARDING THE DETERMINATION OF WELD CHEMISTRY THE EXCLUSION OF DATA THE SELECTION OF A HEAT-SPECIFIC BEST-ESTIMATE METHODOLOGY THE USE OF GENERIC VALUES I  ;

1 3

REVIEW OF BEST-ESTIMATE CHEMISTRY METHODOLOGIES SIMPLE MEAN GROUP OR SAMPLE MEAN (MEAN-OF-THE44EANS)

COIL-WElGHTF.D MEAN GENERIC VALUE FOR A CLASS OF MATERIAL (COPPER COATED WIRES, NICKEL ADDITION WELDS, ETC.)

TANDEM-WIRE WELO DECONVOLUTION OTHERS i

CONCLUSIONS OF THE COMBUSTION ENGINEERING REPORT AND FRAMATOME INSPECTION COMBUSTION ENGINEERING ADDITIONAL DATA RELEVANT TO THE DETERMINATION OF BEST-ESTIMATE CHEMISTRIES WAS INTEGRATED INTO AND EVALUATED DURING THE REVIEW PROCESS SEVERAL METHODOLOGIES FOR THE DETEFd?.iNATION OF BEST-ESTIMATE CHEMISTRIES WERE CONSIDERED GIVEN THE AVAILABLE DATA. FOR THE PURPOSES OF THE REPORT, A METHOD WAS CHOSEN TO REPRESENT THE "BEST-ESTIMATE" FOR A PARTICULAR HEAT.

• lN SOME CASES, GENERIC VALUES WERE CHOSEN FOR EITHER COPPER, NICKEL, OR BOTH.

FRAMATOMEi ADDITIONAL DATA RELEVANT TO THE DETERMINATION OF BEST-ESTIMATE CHEMISTRIES WERE EVALUATED DURING THE REVIEW PROCESS IT WAS DETERMINED THAT THE BEST-ESTIMATE VALUES WHICH HAD BEEN USED PREVIOUSLY TO CHARACTERIZE SPECIFIC WELD-WIRE HEATS WERE ACCEPTABLE.

O

i RESULTS OF THE STAFF'S PRELIMINARY REVIEW

• lN GENERAL THE STAFF HAS CONCLUDED THAT THE i

ANALYSES PRESENTED BY CEOG AND FRAMATOME '

PROVIDE ACCURATE OR CONSERVATIVE METHODOLOGIES FOR THE DETERMINATION OF RPV

WELD BEST-ESTlMATE CHEMISTRIES

• HOWEVER, THE STAFF DOES HAVE SPECIFIC QUESTIONS TO BE ADDRESSED WITH THE CEOG AND FRAMATOME REGARDING SOME ASPECTS OF THE EVALUATIONS (DATA EXCLUSION, USE OF GENERIC VALUES)

• THE STAFF HAS DEVELOPED THE FOLLOWING GUIDANCE FOR LICENSEES ON THE USE OF INFORMATION FROM THE CEOG AND FRAMATOME l ACTIVITIES IN THE EVALUATION OF THEIR FACILITIES i

j l

l l

1

_ _ _ . ~ . . _ _ _ _ _ _ - _ . _ _ _ . . _ _ . . . _ _ _ . . . _ . . . _ - _ . _ . - . . _ _ _ _ _ , . _ - _ . . . . _ _ _ . _ _ _ _ . _ - - . _ _ _ _ _ _ - - - - . _ . ~ . _ _ , . - _ _ _ - -

ACCEPTABLE METHODS OR REASONS FOR EXCLUDING DATA A. IF THERE EXISTS AN IDENTIFIED ANQ RECORDED DEFICIENCY IN A DATAPOINT

• A UUPLICATE RECORD

• AN UNTRACEABLE RECORD A RECORD WHICH IDENTIFIES AN ATYPICAL CONDITION OR SAMPLE LOCATION B. IF A DATAPOINT IS IDENTIFIED AS A STATISTICAL OUTLIER M A PHYSICAL BASIS EXISTS FOR BELIEVING THE DATAPOINT TO BE ATYPICAL ALL DATA NOT EXCLUDED IN (A.) SHOULD BE USED AS THE DATASET A PRIORI EXCLUSION OF SOME DATA BASED UPON "lNCONSISTENCY" WITH EXPECTED NORMS SHOULD NOT BE USED BEFORE AN ANALYSIS FOR STATISTICAL OUTLIERS IS CONDUCTED TESTS FOR STATISTICAL OUTLIERS SHOULD BE MORE STATISTICALLY RIGOROUS THAN CHAUVENET'S CRITERION IF A DATAPOINTIS SCREENED AS AN OUTLIER, ADDITIONAL PHYSICAL EVIDENCE SHOULD BE AVAILABLE BEFORE THE DATAPOINT IS REMOVED FROM THE DATASET r

---

_ _. . . - - _ . - - . . . ~ .

@TAFF GUIDANCE ON DETERMINING BEST-ESTIMATE CHEMISTRIES IN ACCORDANCE WITH THE CURRENT REGULATORY

FRAMEWORK OF 10 CFR 50.61, THE GulDANCE IN REGULATORY GUIDE 1.99 REVISION 2, AND PREVIOUS STAFF EVALUATIONS:

. A HEAT-SPECIFIC CHEMISTRY DETERMINATION CAN BE MADE WITH 1 OR MORE VALID DATAPOINTS FOR THE WELD WIRE HEAT

• lF ONLY A LIMITED AMOUNT OF DATA ARE AVAILABLE, THE STAFF WILL CONSIDER AS PART OF ITS REVIEW PROCESS THE IMPACT OF USING GEN !RIC COPPER AND/OR NICKEL VALUES FOR THE SUBJECT CLASS OF MATERIAL 4

• lF NO VAllD DATA ARE AVAILABLE TO DEFINE EITHER THE COPPER OR NICKEL CONTENT OF A WELD WIRE HEAT, A GENERIC MEAN VALUE FOR THE MATERIAL CLASS PLUS ONE STANDARD DEVIATION SHOULD BE USED SURVElLLANCE WELD CHEMISTRY SHOULD CONTINUE TO BE BASED ON THE CHEMISTRY DATA FOR THAT SPECIFIC WELD RATHER THAN THE HEAT BEST-ESTIMATE CHEMISTRY FINALLY, THE STAFF HAS DETERMINED THAT WHEN SUFFICIENT DATA EXISTS FROM MORE THAN ONE SOURCE (WELD) COIL-WEIGHTED AND MEAN-OF-THE-MEANS ARE EXPECTED TO PROVIDE A MORE RELIABLE ESTIMATION OF THE HEAT-SPECIFIC CHEMISTRY l

l i

y

, .._._..--._ ---.__,,,___ ,..._ ,._ --.. ..-, _ ..,_. ._ ., -._. ---- _.. , . _ , . _ - . , ~ . . . . . . , - - _ _

PROPOSED STAFF actions l

• THE STAFF PLANS TO ISSUE LETTERS TO THE CEOG AND FRAMATOME, AS NECESSARY, REGARDING RESOLUTION OF THE ISSUES WITH DATA EXCLUSION, OUTLIERS, AND GENERIC VALUES AS NOTED PREVIOUSLY  ;

THE STAFF WILL ISSUE LETTERS TO EACH LICENSEE WHOSE VESSEL MATERIAL OR SURVElLLANCE PROGRAM MATERIAL BEST-t ESTIMATE CHEMISTRIES MAY BE AFFECTED BY THE INForIAATION FROM THE CEOG AND FRAMATOME ACTIVITIES. THE LETTERS WILL ASK LICENSEES TO REVIEW THE INFORMATION, INCLUDING THOSE ISSUES RAISED BY THE STAFF REGARDING THEM, AND SUBMIT TO THE STAFF ANY NECESSARY REVISIONS TO THEIR FACILITIES' LICENSING-BASIS INFORMATION. THESE LETTERS WILL ALSO INCLUDE A REQUEST FOR LICENSEES TO CONSIDER SURVEILLANCE DATA ISSUES IN UPDATING RPV INTEGRITY EVALUATIONS (TO BE DISCUSSED IN MORE DETAIL LATER).

LICENSEES MAY CONCLUDE, BASED ON THE NEW DATA, THAT THEIR CURRENT RPV INTEGRITY EVALUATIONS (PTS, P-T LIMITS)

ARE NON-CONS 8ERVATIVE. lF SO, THE STAFF EXPECTS THAT THE LICENSEES WILL PURSUE THE REQUIRED LICENSING CHANGES ON AN EXPEDITIOUS BASIS, IN PARALLEL WITH THEIR RESPONSES TO THE STAFF'S LETTER ON THESE SUBJECTS.

2 1

t i

Regulatorv impact of Changing Best-Estimate Chemistries Without involving Surveillance Data Case #1 The limiting material remains the same but the current RT,rs is nonconservative.

Facility: XXXXXX Surveillance Weld Hea'ti XXXXX [ Heat is present in vesse!, but does not affect limiting material)

RVID Evaluation of Limiting Materials CEOG Evaluation of Limiting Materials Heat Cu Ni CF RT,, Heat Cu Ni CF RT,,

(orientation) (orientation)

Cire. Weld 0.170 0.920 197.8 244.9 Cire. Weld 0.192 1.038 224.2 276.3 Note: All CFs are based on RG 1.99 Rev. 2 Table determinations.

Case #2 A new material becoms:' limiting and the RTpys value for the vesselincreases.

Facility: YYYYYY Surveillance Weld Heat: YYYYYY [ Heat not present in vessel]

RVID Evalestion of Limiting Materials CEOG Evaluation of Limiting Materials Heat Cu Ni CF RT,, Heat Cu Ni CF RT,,

(orientation) (orientation)

Axial Weld 0.210 1.000 229.0 263.7 AxialWeld 0.203 1.018 226.8 281.3 1 1 Avial Weld 0.190 0.970 215.7 248.9 AxialWeld 0.219 0.996 231.1 266.0 2 2 Note' All CFs are based on RG 1.99 Rev. 2 Table determinations.

H

l l

EVALUATION AND USE OF SURVEILLANCE DATA l

Basic Methodology l Assessing Credibility Data from i source 1 Pata from more than 1 source Effects of Credibility Credible Data

! Non credible Data ,

Example demonstrating a non conservative Table CF Adjustments to Surveillance Data Chemical composition adjustments irradiation environment (temperature) adjustments Summary of 5 Cases Illustrating issues Related to Assessing Surveillance Data i

---

- - - - - - - - .e----wm

ASSE3 SING SURVEILLANCE DATA Methodology described in Regulatory Guide (RG) 1.99, Revision 2 and 10 CFR 50.61 Recent issues related to assessing surveillance data Appropriate y intercept for best fit line through surveillance data (plot of ARTum as a function of fluence factor)

Using a chemistry factor determined from the surveillance data rather than the generic tables when the surveillance data are not credible Correcting for chemical composition (ratio procedure)

Correcting for irradiation environment (temperature)

Appropriate chemical composition to assign to surveillance capsules from a source (i.e., mean value for all capsules from that source)

Appropriate parameters to normalize surveillance data to when assessing credibility (i.e., mean of surveillance data) and determining the chemistry factor (i.e., best estimate of vessel) l s .

. _ . _ _ . - .- _ _ . . . - . _ . . . . _ - , -- .~. - , . , . - . . .. .,. . --.

a METHODOLOGY FOR ASSESSING RTuor/RT,r. ,

Basic equation for at. .sssing reference temperature l RTwor = RTwortu) + M + ARTwor i

ARTwo, accout.ts for effects of irradiation ARTuor is calculated from the following equation:

ARTwor = (CF)*FF where:CF = chemistry factor '

FF = fluence factor = f (* ** 0  :

2 methods for determining CF (described in RG 1.99, R2)

Position 1.1: used when less than 2 surveillance data points are available or when surveillance data are not credible. In this case, CF is determined from generic tables.

Position 2.1: used when 2 or more credible surveillance data points are available. CF is determined from best fit line of surveillance data (ARTwor versus FF).

a

i ASSESSING CREDIBILITY  ;

Credibility Criteria from RG 1.99, Revision 2

" the use of surveillance data from a given reactor (in place of the calculative procedures given in this gulde) requires considerable engineering judgment to evaluate the credibility of the data and assign suitable margins."

"When surveillance data from the reactor in auestion become available, the weight given to them relative to the information in this guide will depend on the credibility of the surveillance data as judged by the following criteria:"

1. Materials in the capsules should be those Judged limiting ,

2. Scatter in the Charpy curves should be small enough to permit the determination of Tu and upper shelf energy unambiguously

3. When there are 2 or more sets of data from one reactor, the scatter of ARTm values about a best fit line should be less than 28'F for welds and 17'F for base metal. Even if the fluence range is large (2 or more orders of magnitude), the scatter should not exceed twice those values

4. The irradiation temperature of the Charpy specimens should match the vessel wall temperature at the cladding / base metal interface within 125'F

5. The surveillance data from the correlation monitor material in the

, capsule should fall within the scatter band of the data base for that material

/"

ASSESSING CREDlBILITY (cont'd)

Cases to consider Data from one source  !

Data from plant whose vessel is being assessed i Data from another source but not from the plant whose vessel is being asseesed Data from many sources (e.g., plant is part of an integrated

program)

Data from plant whose vessel is being assessed and data from other plants Data from other sources but not from the plant whose vessel l Is being assessed '

Credibility evaluation when data are from one source i Since chemical composition and Irradiation environment of the surveillance data are essentially the same, determine best fit line ,

relating the measured (unadjusted) ARTuor to the FF  :

I Y-intercept of best fit line is zero since from of equation has no y-intercept term (i.e., ARTuor = (CF)*FF)

Data are credible (subject to other criteria)If the difference between the measured (unadjusted) and predicted values (from best fit line) of ARTuor for all of the data points are within 28'F for welds and 17'F for base metal After credibliity determination, >

determine CF by adjusting the data to account for differences in the chemical composition and irradiation environment between the capsules and the vessel (normalize data to the best estimate chemical composition and temperature of the l Vessel being assessed)

)

determine a best fit line through this data

l l

ASSESSING CREDIBILITY (cent'd)  ;

Credibility evaluation when data are from many sources  !

" Adjust" measured values of ARTwo, to account for Irradiation environment and chemical composition differences - discussed  ;

below i For credibility determination, normalize data to the mean chemical composition and temperature of the surveillanca specimens Plot adjusted ARTuor as a function of FF and determine best fit line (least squa'res regression)

Y intercept of best fit line is zero since from of equation has no -

y-intercept term (i.e., ARTwo, = (CF)*FF)

Data are credible (subject to other criteria) if the difference between

" adjusted" and predicted values (from best fit line) of ARTwor for all of the data points are within 28'F for welds and 17'F for base metal After credibility determination, determine CF by adjusting the measured data to account for differences in the chemical composition and irradiation environment between the capsules and the vessel (normalize data to the best estimate chemical composition and temperature of the vessel being assessed) determine a best fit line through this data For credibility analysis, normalize surveillance data to the mean chemical composition and temperature of the surveillance specimena '

For determination of the CF, normalize surveillance data to the best estimate chemical composition and temperature of the resral being assessed

/7

1 EFFECTS OF CREDlBILITY  :

i Credible Data Use CF from surveillance data (i.e., best fit line with data normalized to the best estimate chemical composition and temperature of the vessel being assessed)

Use reduced value of o3i n calculation of margin term, M M= Margin =2*lo*u+da ou(sometimes referred to as o,) is the standard deviation for RTwortu) 0 3 is the standard deviation for ARTuor For welds,0 3= 14'F when data are credible (but need not exceed % of the mean value of ARTuor For base rnetal, o3 = 8.5'F when data are credible (but need not exceed % of the mean value of ARTwo, i

N

0 4

EFFECTS OF CREDlBILITY (cont'd)

Non credible data Use CF from generic tables in Regulations unless the CF determined i from the surveillance data is significantly greater than that from the j '

generic tabin indicating that the table CF is non-conservative for this material Use " full" value of c6 In calculating of margin term, M (even if using

the CF determined from the surveillance data in the case when the l tables are non-conservative)

For welds, c3 = 28'F when data are not credible (but need not 1

exceed % of the mean value of ARTuor For base metal,06= 17'F when data are not credible (but need not exceed % of the mean value of ARTuor 4

l Determining if the Table CF is non conservative Use CF from table (based on surveillance specimen chemical

composition) to determine predicted ARTwor Determine difference between predicted and measured ARTwor if difference exceeds 2 standard deviations (56*F for welds,34*F for base metal), the Table CF is considered non conservative i

c

{

4

. j i

EFFECTS OF CREDIBILITY (cont'd) i Non conservative Table CF Example  ;

Data are from plant whose vessel is being assessed (T,,,,, = Te.,,,i.)  !

- no adjustment for temperature differences needed j Base metal - no adjustment for chemical composition differences  ;

needed i Chemistry 0.20% Cu,0.54% NI - CFr.w,v ien,. = 141.8'F Since base metal, CFr..,y,,,,ien = CF,,w,...en .

CF determined from best fit line = CF. .o = 159.9'F Capsule Fluence Measured Prod;cted ART , Measured .  ;

Factor (FF) ART , from Table Predicted (CF,. *F0 ART ,

1L 0.703' 128.1 99.7 28.4 1T 0.703 137.8 99.7 88 1 2L 0.895 118.9 124.9 8.0 2T 0.896 131.8 126.9 4.9 3L 1.018 147.7 144.0 3.7 ST 1.018 179.9 144.0 38.9 Since the scatter from several of the data points exceeds 20 (34'F),

the Table CF is not conservative.

1 The portion of the margin term, M, attributed to 03 is non conservative in this case T..e CF from the surveillance data, CF.., o,,,,(i.e., from the best fit line with data normalized to the best estimate chemical composition and temperature of the vessel being assessed) should be used in assessments of RTuoy. In this case, no normalization was required (base metal and T,,,,, = Tc.,,,i ).

/f

l'

CHEMICAL COMPOSITION ADJUSTMENTS To account for chemistry differences between the surveillance specimen weld and the vessel weld, an adjustment to the data must be performed Wold data are adjusted through the ratio procedure Adjust measured values of ARTuor from welds for chemical composition differences per the ratio procedure as follows

Retto Adjusted bMT,,7 = ('**CFC **' '* **F'**

• )

• bM Tuo r, ,,, ,,,,,,,

resse, surv. chem.

P IRRADIATION ENVIRONMENT ADJUSTMENTS Irradiation temperature and fluence (or fluence factor) are first order environmental variables in assessing irradiation damage Other variables are believed to have less significant contributions To account for differences in temperature between surveillance specimens and vessel, an adjustment to the data must be performed Studios have shown that for temperatures near 550*F, a 1'F decrease in irradiation temperature will result in approximately a 1 *F increase in ARTuor For capsules with Irradiation temperature of Tc.,,,,, and a plant with an irradiation temperature of Tri ni, an adjustment to normalize ARTuor, . ,,, to Tp ,is made as follows:

Temp. adjusted ARTuor = ARTuor,,n ,,, + 1.0 * (Tc.,,,, - Tri ni)

/

_ __ . _ _ . _ , . . ~ . _ . ~ . _ . . .- . . _ _ _ - - -

ADJUSTMENTS TO DATA Adjustments to the data for both Irradiation environment and chemical composition may be necessary, in this case, adjust ARTwor,,,, ,,, for either Irradiation temperature or for chemical composition.

Use the resultant value (temp adjusted or ratlo adjusted ARTuor) and make the other adjustment (i.e., for chemical composition or irrad'ition environment)

The final value is the " ratio and temperature" adjusted ARTwor The fewer the adjustments to the measured values of ARTuoy, the less likely errors will be Introduced into the measurements Data from the plant being assessed,if available,is most representative (least adjustments)

All Irradiation environment variables are essentially identical in this case When data from plant being assessed are not available, data from other plants should be considered. Data from plants with the same NSSS vendor as the plant being assessed may bo more representative since t!,e magnitude of the temperature adjustment may be lower.

L

. t

SUMMARY

OF 5 CASES Case 1 Simple, straightforward RG 1.99, Revlulon 2 case where the surveillance data are determined to be credible Surveillance data (weld) are available from the plant being assessed but not from any other source Data are determined to be credibli bssed on best fit line relating measured (i.e., unadjusted) ARTwo versus FF For CF determination, No temperature adjustment necessary Correction for chemical composition per ratio procedure CF is slope of best fit line relating " ratio and temperature" adjusted ARTwot (normalized to best estimate chemical composition and temperature of vessel being assessed) to FF.

Case 2 Simple, straightforward RG 1.99, Revision 2 case where the survalliance data are determined to be non-credible Surveillance data (base metal) are available from the plant being assessed but not from any other source Data are determined to be non-credible based on best fit line relating measured (i.e., unadjusted) ARTwor to FF For CF determination, No temperature adjustment necessary Correction for chemical composition not necessary since base >

metal (ratio procedure applies to welds)

.o

1

SUMMARY

OF 5 CASES (cont'd) l Case 3 Similar to Case 2 (non credible surveillance data) except that the  !

Table CF is determined to be non conservative Surveillance data (base metal) are available from the plant being assessed but not from any other source Data are determined to be non credible based on best fit line relating measured (i.e., unadjusted) ARTwor to FF  ;

No temperature adjustment necessary Correction for chemical composition not necessary since base metal (ratio procedure applies to welds)

Since the CF from the surveillance data was greater than that from  !

the Table in RG 1.99, Revision 2, an assessment of whether the difference was significant was performed The assessment Indicated that the difference was significant ,

and that the CF determined from the surveillance data should be used in RTuoy assessments If this were a weld (Instead of a base metal), must compare CF from surveillance data (normalized to the best estimate chemical composition and temperature of the vessel being assessed) to the CF in the Tables I

~ , - - -

SUMMARY

OF 5 CASES (cont'd)

Case 4 Surveillance data available from plant whose vessel is being assessed and from other sources  !

Surveillance data (weld) are available from Cie plant being assessed i and from other sources Surveillance capsule chemistry is average from all chemistry analyses for that material from that source resulting in one set of ,

chemistry values for all specimens removed from a given source  :

(i.e., all Plant "x" weld specim7ns are assigned the same chemical i composition) 4

^

Credibility determination i Data from the plant being assessed is the most appropriate i

, Least amount of adjustments (irradiation environment is the same)

Data were determined to be credible based on best fit line relating measured (i.e., unadjusted) ARTuor to FF.

Also considered all of the data together For credibility determination, data are normalized to the

mean chemical composition and temperature of the surveillance specimens Data were

'm credible based on best fit line reladng

" ratio and temperature" adjusted ARTuor to FF

~

CF is based on surveillance data from plant being assessed.

CF is determined from best fit line with data normalized to the best estimate chemical composition and temperature of the vessel being assessed.

No temperature adjustment necessary (T,w = Tc.,,,i )

Correction for chemicc composition per ratio procedure

I

SUMMARY

OF 5 CASES (cont'd)

Case 5 Surveillance data not available from plant whose vessel is being I assessed but are available from other sources Surveillance data (weld) are only available from sources other than the plant whose vesselis being assessed Surveillance capsule chemistry is average from all chemistry analyses for that material from that source resulting in one set of chemistry values for all specimens removed from a given source (i.e., all Plant "x" weld specimens are assigned the same chemical composition)

Credibility Determination Data requiring the least adjustments are typically the most appropriate.

The data from the same NSSS 'tandor has the closest  !

Irradiation environment to the plant whose vessel is being

, assessed Data were determined to be credible based on best fit line relating measured (i.e., unsojusted) ARTuor to FF. Used measured values since all data are from one source Also considered all of the data together.

For credibility determination, data are normalized to the mean chemical compe-ition and temperature of the

. surveillance specimana.

A Data were determined to be credible based on best fit line relating " ratio and temperature" adjusted ARTuor to U CF based on surveillance data from Plant "a" CF is determined from best fit line with data normalized to the best est: mate chemical composition and temperature of the vessel being assessed.

Temperature adjustment necessa;y Correction for chemical composition per ratio procedure

?

SUMMARY

AND FUTURE ACTIONS Licenses assessments:

Y Intercept for best fit line should go through the origin (i.e., with a FF of zero, ARTwor = 0)

Corrections for chemical composition and irradiation environment should be made Chemistry of surveillance welds should be mean value determ!ned from all chemical analyses corresponding to that surveillance weld (i.e., all Plant "x" weld specimens should have the same chemistry values which is the mean of all chemical analysis obtained from that serveillance weld)

For determination of credibility, surveillance data should be normalized to the mean chomical composition and temperature of the sirveillance speQlmfRA in the case where all surveillance data are from one source, the normalized values would be the measured (unadjusted) values of ARTuor Determination of appropriate CF and margin (e.g., using a CF determined from non-credible surveillance data when the generic Tables are non-conservative)

For determination of CF, surveillance data should be normalized to the best estimate chemical composition and temperature of the vessel being assessed l

s

CASE 1 - SURVEILLANCE DATA AVAILABLE FROM PLANT BUT NO OTHER SOURCE - CREDIBLE DATA Surveillance data are available from plant whose vessel is being assessed (and not from any other source)

Best estimate for heat Weld rnetal 0.283% Cu, 0.755% Ni - CFr.w,v ics , = 212.2*F Credibility assessment Since all data from one source, plot measured ARTuor versus FF and determine best fit line Slope of best fit line = 191.3*F Capsule Cu Ni irradiation Fluence Fluence Measured Predicted (Measured .

Temperature (x10") Factor ARTwo, ARTwo, from Predicted)

(Tc .) (FF) best fit line ARTuo, Plant x .1 0.2284 0.7391 632.0 0.669 0.837 176.0 140.2 14.8 Plant x . 2 0.2294 0.7391 632.0 2.070 1.198 236.0 229.2 6.8 Plant x . 3 0.2294 0.7391 632.0 2.890 1.282 230.0 246.2 16.2 where predicted ARTuo, = (Slope..)*(Fluence Factor)

Data are credible since scatter is less than 28'F for all surveillance specimens i

J

CASE 1 - SURVEILLANCE DATA AVAILABLE FROM PLANT BUT NO OTHER SOURCE - CREDIBLE DATA (cont'd)

Determination of CF No temperature adjustments are necessary since Tc.,,,a = T,..,

(capsules were irradiated at plant being assessed)

Adjust measured values of ARTuor for chemical composition differences as follows (normalize data to best estimate of vessel being assessed):

Ratio Adjusted bRT '** '" "' ***'

)

• bR Tu o r. oroesures uor = ( cF,,,,,, ,,,, e,,,,

C Fr,3w, ,,,, e,,,, = 19 2.5 ' F Determine best fit line relating adjusted (" ratio and temperature" adjusted) ARTuor to FF. The slope of this best fit line is the CF,,,, o, .

Since no temperature adjustments were required in this case the ratio adjusted ARTuor is the same as the " ratio and temperature" adjusted ARTuor C F ,y, o,i, = 210.9 ' F Effects of data analysis technique Assume: RTworcui = -56.0'F; M = 44.0*F; FF = 1.307 RTwor.wo n.u. = -56.0 + 44.0 + (191.3*1.307) = 238.0'F RTuor,n.,3. = -56.0 + 44.0 + (210.9*1.307) = 263.6'F

CASE 2 - SURVEILLANCE DATA AVAILABLE FROM PLANT BUT NO OTHER SOURCE - NON-CREDIBLE DATA AND TABLE CF IS CONSERVATIVE Gurveillance data are available from plant whose vessel is being assessed (and not any other source)

Best estimate for heat Base metal 0.14% Cu,0.55% NI - CFr.,i ,y,,,,ies = 97.8'F Credibility assessment Since base metal and all data are from one source, plot measured ARTwor versus FF and determine best fit line Slope of best fit line = 95.9'F Capsule Cu NI Irradiation Fluence Fluence Measured Predicted (Measured .

Temperature (x10") Factor ARTuo, ART o, from Predicted)

(Te.,..) (FF) best fit line ARTue, il 0.14 0.66 N/A 0.680 0.848 86.0 81.3 3.7 1T 0.14 0.66 N/A 0.680 0.848 65.0 81.3 26.3 2L 0.14 0.66  !. A 1.69 1.144 106.0 109.8 4.8 2T 0.14 0.66 N/A 1.69 1.144 90.0 109.8 19.8 3L 0.14 0.66 N/A 2.96 1.287 135.0 123.4 11.6 3T 0.14 0.66 N!A 2.96 1.287 106.0 123.4 18.4 4L 0.14 0.66 N/A 3.82 1.346 165.0 129.1 26.9 4T 0.14 0.66 N/A 3.82 1.346 146.0 129.1 16.9 where predicted ARTuo, = (Slopem)'(Fluence Factor)

Data are not credible since scatter is greater than 17'F for several surveillance specimens l

l 1

I

i CASE 2 - SURVEILLANCE DATA AVAILABLE FROM PLANT l' BUT NO OTHER SOURCE - NON-CREDIBLE DATA AND TABLE CF IS CONSERVATIVE (cont'd)  !

I Determination of CF l

Since data are not credible, the CF should be determined from the  !

Tables.

To ensure that the Table CF is conservative, determine the CF from the surveillance data as follows No temperature adjustments are necessary since Tc.,.. = T,.n, (capsules were irradiated in plant being assessed) i No chemical composition adjustments are necessary since material is base metal (ratio procedure applies to welds)

Determine best fit line relating measured ARTwo, to FF. The slope of this best fit line is the CF.,n,, o,,, which was determined to be 95.9*F.

Since no chemical composition or temperature adjustments were required in this case, the measured ARTwor is the same as the " ratio and temperature" adjusted ARTwor Since the CF from the Table is greater than the CF calculated from the surveillance data, the Table CF is conservative and should be used in assessments of RTuor Effects of data analysis technique Assume: RTworcui = 15.0*F; FF = 1.35 i R T uor,A . w o,c w o . = 15.0 + 17.0 + (95.9*1.35) = 161.5*F RTuor, r., u .cw w o = 15.0 + 34.0 + (97.8*1.35) = 181.0*F J

- - . . . . . . . . - _ _ . , . , _ - , _ - - . _ , - - - ~ _ . . - - _ . . . . _ _ . - , . - . - - - ~ . . -

I CASE 3 - NON-CREDlBLE SURVEILLANCE DATA AND TABLE CF IS NON-CONSERVATIVE  ;

7 Surveillance data are available from plant whose vessel is being assessed (and not from any other source)

I

Best estimate for heat Base metal O.20% Cu,0.54% NI - CFr.w,vuem = 141.8'F Credibility assessment Since base metal and all data are from one source, plot measured ARTuor versus FF and determine best fit line

Slope of best fit line = 159.9'F Capsule Cu Ni irradiation Fluence Fluence Measured Predicted (Measured .

Temperature (x10*) Factor ART,, ART ,from Predicted)

(Tc ) (FF) best fit line ART ,

Plant x .1L 0.20 0.64 M 7.0 0.340 0.703 128.1 112.4 15.7 Plant x .1T 0.20 0.54 547.0 0.340 0.703 137.8 112.4 25.4 I Plant x . 2L 0.20 0.64 547.0 0.688 0.896 118.9 143.1 34.2 Plant x . 2T 0.20 0.64 547.C 0.488 0.896 131.8 143.1 11.3 Plant x . 3L 0.20 0.64 547.0 1.0;t 1.016 147.7 182.4 14.7 Plant x . 3T 0.20 0.54 647.0 1.0 W 1.014 179.9 142.4 17,5 i

where predicted ART,,,, = (Slope..)*(Fluence Factor)

Data are not credible since scatter is greater than 17'F for several surveillance specimens i

1

. . _ _ _ _ _ _ . . . _ _ _ _ _ . _ _ _ _ _ _ . - _ _ _ _ . _ _ . _ . , _ _ . _ _ - . . ~ . _ . . - . . _ , _ . . . _ _ _ - . . . , . . _ . . - . . . . . . . . . , .

CASE 3 - NON-CRED(BLE SURVEILLANCE DATA AND TABLE CF IS NON-CONSERVATIVE (cont'd)

Determination of CF Since data are not credible, the CF should be determined from the Tables To ensure that the Table CF is conservative, determine the CF from the surveillance data as follows No temperature adjustments are necessary since Tc.,,,w = Tri n, (capsules were irradiated in plant being assessed)

No chemical composition adjustments are necessary since materialis base metal (ratio procedure applies to welds)

Determine best fit line relating measured ARTwor to FF. The slope of this best fit line is the CF.,,,, o,. which was determined to be 159.9'F Since no chemical composition or temperature adjustments were required in this case, the measured ARTwor is the same as the " ratio and temperature" adjusted ARTwor The difference between the CF from the surveillance data und the CF from the Table is significant (see earlier example comparing Table predicted ARTwor to measured ARTwor); therefore, the CF from the surveillance data (CF ,,,, o ) should be used in assessments of RTwor even though the data are not credible Effects of data analysis technique Assume: RTwortui = 27.0*F; M = 34.0*F; FF = 1.293 RTuor,er,,,,,,,,,o,,,= 27.0 + 34.0 + (159.9*1.293) = 267.8'F RTwor,er,,, y,6i, = 27.0 + 34.0 + (141.8*1.293) = 244.3*F

o CASE 4 - SURVEILLANCE DATA AVAILABLE FROM PLANT AND OTHER SOURCES .

Surveillance data Capsule NSSS Cu NI irradiation Fluence Fluence Measured Temperature Ratio, Designation Vender Temperature (x10") Factor ART,,,, Adjusted Temperature (T%) (FF) (660*F) Adjusted ARTuo, (660*F)

ARTwo, Plant a .1 B&W 0.37 0.70 666.0 0.779 0.930 214.0 220.0 196J Plant b .1 B&W 0.33 0.67 666.0 0.107 0.431 124.0 130.0 126.0 Plant b . 2 B&W 0.33 0.67 666.0 0.866 0.960 203.0 209.0 202.6 Plant c .1 B&W 0.33 0.67 666.0 0830 0.948 182.0 168.0 182.2 Plant c . 2 B&W 0.33 0.67 666.0 0.968 0.991 222.0 228.0 221.0 Plant x - 1 We:t. 0.24 0.66 636.0 0.281 0.663 166.0 161.0 172.1 Plant x . 2 West. 0.24 3.44 636.0 1.940 1.181 240.0 226.0 267.6 l ll Normalization for credibilltv determination (when all data are being used)

Data normalized to mean chemical composition (i.e., copper and nickel) of surveillance specimens Cu = 0.31%

Ni = 0.67%

Data normalized to mean temperature of surveillance specimens Tu-ur. = 550*F

CASE 4 - SURVEILLANCE DATA AVAILABLE FROM-PLANT AND OTHER SOURCES (cont'd)

Assume the following for Plant "x" (the plant whose vessel !s being assessed)

Wold heat 299L44 is in the surveillance program and in the vessel T, , = 536

• F Surveillance date N heat 299L44 is also available from other sources Best estimate for iient Wold metal 0.34% Cu,0.68% NI - CF1,u,,v, ,c%,. = 220.6*F Credibility assessment - Plant "x" data only The data inost representative for Plant "x" is that from Plant "x" since the irradiation environment of the surveillance capsules and the vessel are very similar. This data requires the least adjustment (e.g., no temperature correction)

Plant "x" data should be examined independently to determine credibility Since all data are from one source (Plant "x"), plot measured ARTum-versus FF and determine best fit line Slope of best fit line = 214.8'F Capsule Cu NI Irradiation Fluence l Fluence Measured Predicted (Measured -

Temperature (x10") Factor ARTue, ARTwo, from Predicted)

(T w ,) (FF) best fit line ARTuo, Plant x - 1 0.24 0.66 636.0 0.281 0.663 166.0 140.3 24.7 Plant x - 2 0.24 0.66 636.0 1.940 1.181 240.0 263.6 -13.6 l ,

where predicted ART ,= (Slope .)*(Fluence Factor)

Data are credible since sce**er is less than 28*F for all surveillance specimens

d CASE 4 - SURVEILLANCE DATA AVAILABLE FROM PLANT AND OTHER SOURCES (cont'd)

Determination of CF - Plant "x" data only No temperature adjustments are necessary since Tc.,,o = Tp o.

Adjust measured values of ARTwor for chemical composition differences as follows (normalize data to best estimate of vessei being assessed):

Ratio Adjusted ARTuor * ( CF '"')

• AR Tuor, mouvred rosse. sun. chem.

CFr,m, oy, . = 182.9'F Determine best fit line relating adjusted (" ratio and temperature" adjusted) ARTuor to FF. The slope of this best fit line is the CF oy, o, .

Since no tempe,ature adjustments were required in this case the ratio adjusted ARTuor is the same as the " ratio and temperature" adjusted ARTuor CFio n , o ,,, = 2 5 9 .0

• F

!\

w

CASE 4 - SURVElLLANCE DATA AVAILABLE FROM PLANT AND OTHER SOURCES (cont'd)

- Credibility assessment - All data The data from all sources should also be considered

, Since data are from multiple soumes, must adjust data for chemical composition and irradiation environment differences and then plot the  !

" ratio and temperature" adjusted ARTuor values versus FF and determine best fit line For credibility determination, data are normalized to the mean chemical composition and temperature of the surveillance specimens Slope of best fit litie = 218.4*F Capeule Cu NI Irradiation Fluence Fluence Ratio and Predicted (Adjusted Temperature (x10") Factor Temperature ART,,,, from Predicted)

(T%) (FF) (860) beet fit line ARTuor Adjusted ARTe,i Plant a 1 0.37 0.70 666.0 0.7791 0.930 196.0 203.1 7.1 Plant b 1 0.33 0.67 666.0 0.107 0.431 126.0 94.1  : 31.9 Plant b 2 0.33 0.67 668.0 0.866 0.960 202.6 209.6 7.1 Plant c .1 0.33 0.67 666.0 0.630 0.948 182.2 207.0 24.6 Plant c . 2 0.33 0.67 666.0 0.966 0.691 221.0 218.4 4.6 Plant x 1 0.24 0.84 636.0 0.281 0.863 172.1 142.8 1 29.4 Plant x 2 0.24 0.66 638.0 1.940 l 1.181 267.6 266.0 -0.4 where predicted ART,,o, = (Slope J(Fluence Factor) 4 Data are not crealble since scatter is greater than 28*F for several surveillance specimens 8

t 7

3O

CASE 4 - SURVEILLANCE DATA AVAILABLE FROM '

-PLANT AND OTHER SOURCES (cont'd)

Determination of CF - All Data if data were credible, the CF would be determined as follows Must make irradiation temperature and chemical composition adjustments sincs the irradiation temperature and chemistry differ between the capsules and the plant being assessed For capsules with Tc.,,, greater than 536*F (i.e., T,..),

must increase ARTuor,,,,,u by 1.0*F for each degree difference in irradiation temperature to get the temperature adjusted ARTuoy (i.e., ARTuoy,r )

To obtain the " ratio and temperature" adjusted ARTuor, apply the ratio procedure as follows:

CF i RattolTemperature Adjusted hRTuor * ( gpreese. v ..: ca.m.)

• uor, ARTr swasted f-Determine best fit line relating adjusted (" ratio and temperature" adjusted) ARTwor to FF. The slope of this best fit line is the-C F.., o, . C F... o,,, = 247.2

• F Appropriate CF Data from the plant being assessed were evaluated and the data were credible When.all of the data were evalurled, the data were determined to be not credible l Since the data from the plant being assessed is the most appropriate L and requires the least amount of adjustment, the CF determined from

( evaluation of the Plant "x" data is the most appropriate.

i CF.., o,,, = 259.0 *F v

l -- .. - . . - . . . ---- -- . . . . . . . - . - . _ , .. . . . - . , .-.

CASE 4 - SURVEILLANCE DATA AVAILABLE FROM PLANT AND OTHER SOURCES (cont'd)

Effects of data analysis technique (Ratio procedure and using plant-specific data)

Previous analyses Ratio procedure not applied, temperature correction to data from other sources not made, All data used RTwortui = -7.0

• F; M = 49.8; CF = 217.0*F; FF = 0.8745 RTuor,,,,,io,, = -7.0 + 49.8 + (217.0

• 0.8745) = 232.6 *F Current analyses Ratio procedure applied, No temperature correction necessary, Only Plant "x" data used RTwor,curr.ni = -7.0 + 49.8 + (259.0

• 0.8745) = 269.2*F G

5*

4-CASE 5 - SURVElLLANCE MTA NOT AVAILABLE FROM PLANT BUT AVAILABLE FROM OTHER SOURCES Surveillance data Capsule N888 Cu NI irradiation ' Fluence Fluence Measured Temperature Ratio, Designation Vendor Temperature (x10") Factor ART m Adjusted Temperature (T%) (FF) (647'F) Adjusted ARTun (647'F)

ARTun Plant a 1 West. 0.23 0.62 642.0 0.602 0.808 110.0 106.0 104.0 Plant a . 2 West. 0.23 0.62 642.0 0.829 0.947 165.0 160.0 168.4 Plant a . 3 West. 0.23 0.62 642.0 2.380 1.234 166.0 160.0 168.4 Plant a 4 West. 0.23 0.62 642.0 2.420 1.238 180.0 175.0 173.3 Plant b 1 B&W 0.22 0.68 666.0 0.610 0.812 148.0 167.0 163.4 Plant b . 2 B&W 0.22 0.68 566.0 1.670 1.141 168.0 177.0 184.2 Normalization for credibilltv determination (when all data are being used)

Data normalized to mean chemical composition (i.e., copper and nickel) of surveillance soecimens Cu = 0.23%

Ni = 0.61%

Data normalized to mean temperature of surveillance specimens knn.m. = 547'F I

I i

CASE 5 - SURVElLLANCE DATA NOT AVAILABLE FROM PLANT BUT AVAILABLE FROM OTHER SOURCES (cont'd)

Assume the following for Plant "y" (the plant whose vessel is being assessed)

Westinghouse is NSSS vendor Wold heat 72445 is not in the surveillance program but is in the vessel ,

Tpam = 536*F Surveillance data for heat 72445 are only available from other sources Best estimate for heat Weld metal l 0.22% Cu,0.58% Ni - CFr,m,v,. icum, = 164.0*F Credibility assessment - Plant "a" data only if the irradiation environment for Plant "a" is Judged to be closer to Plant "y" than Plant "b", the Pla;.t "a" data should be used in assessing the integrity of the vessel Magnitude of te .iperature adjustment ,'s lower for Plant "a" than Plant "o", and the data are from the sarae NSSS vendor Since all data are from one source, perfona credibility analysir similar to Case 1 (i.e., dotermine best fit ;ine through measured ARTuor values) l l

CASE 5 - SURVEILLANCE DATA NOT AVAILABLE FROM PLAN -BUT AVAILABLE FROM OTHER SOURCES (cont'd)

Credibility assessment - Plant "a" data only (cont'd)- ,

Slope of best fit line = 145.8 Capsule Cu Ni Irradiatloa Fluence Fluence Measured Predicted (Measured -

Temperature (x10") Factor ARTwe ART , from Predicted)

(T%) (FF) best fit line ARTc Plant a .1 0.23 0.82 842.0 0.502 0.808 110.0 117.8 7.8 Plant a . 2 0.23 0.42 642.0 0.829 0.947 165.0 138.2 26.8 Plant a 3 0.23 0.62 542.0 2.380 1.234 165.0 179.9 14.9 Plant a 4 0.23 0.42 542.0 2.420 1.238 180.0 180.6 4.6 whce predicted ART, y = 'f "'

)*(Fluence Factor) 1 Datsi are credible since scatter is less than 28'F for all surveillance

specimens Determination of CF - Plant "a" data only Must make irradiation temperature and chemical composition adjustments since the irradiation temperature and chemistry differ between the capsules and the plant being assessed For capsules with Tc.,,,,, greater than 536*F (i.e., Tri ni), must increase ARTuor,,,,,,,,,,, by 1.0*F for each degree difference in irradiation temperature to get the temperature adjusted ARTuor I

(i.e., ARTuor, y ,,,,,,,,)

To obtain the " ratio and temperature" cdjusted ARTuor, apply the i ratio procedure as follows:

1 RatlolTemperature Adjusted bRT C F * '** *** *' **

uo 7, 7 ,,y,,,,

i . uor " ( CFr .as., sm. ca. . )

• bRT l

4

. , . , . . . . .. . , . , - . . . . . ~ - , - . - . - - . - . .

CASE 5 - SURVEILLANCE DATA NOT AVAILABLE FROM PLANT BUT AVAILABLE FROM OTHER SOURCES (cont'd)

Determination of CF - Plant "a" dat9 only (cont'd)

Surveillance data showing irradiation environment and chemical composition adjustments - normalind to best estimate of191331 Capsule Ns58 Cu Ni irradiation Fluence Fluence Measured Temperature Ratlo, Designation Vendor Temperature (x10") Factor ART,,,, Adjusted Temperature (Tw) (FF) (636*F) Adjusted

- ART.,,, ART ,,,

Plant a .1 West. 0.23 0.62 642.0 0.602 0.808 110.0 116.0 110.3 Plant a . 2 West. 0.23 0.62 642.0 0.829 0.947 166.0 171.0 162.7 Plant a . 3 West. 0.23 0.62 642.0 2.380 1.234 166.0 171.0 162.7 Plant a . 4 West. 0.23 0.62 642.0 2.420 1.238 180.0 186.0 176.9l Determine best fit line relating adjusted (" ratio and temperature" adjusted) ARTuor to FF. The slope of this best fit line is the CF.,,,o, .

l CF,oy, o,,, = 144,0.p i

t

, - . + , , - - - , . , - . . , . . - - - . . - - - - -

CASE 5 - SURVEILLANCE DATA NOT AVAILABLE FROM PLANT BUT AVAILABLE FROM OTHER SOURCES (cont'd)

If it had been determined that all data were applicable to Plant "y", credibility and the CF would have been evaluated as follows:

Credibility Determination - All data Plot " ratio and temperature" adjusted ARTwor values (normalized to the mean chemica; composition and temperature of the surveillance specimens) versus FF for all data (not just Plant "a") and determine best fit line i

Slope of best fit line = 150.3*F Capsule Cu NI irradiation Fluence Fluence Ratio, Predicted (Measured -

Temperature (x10") Factor Temperature ART,,or from Adjusted)

(T%) (FF) Adjusted best fit line ART,,or ART,,,,

Plant = .1 0.23 0.62 642.0 C.302 0.808 104.0 121.4 17.4 Plant a - 2 0.23 0.42 642.0 0.829 0.947 168.4 142.4 16.0 4- Plant a 3 0.23 0.62 642.0 2.380 1.234 168.4 186.4 27.0 Plant a 4 0.23 0.82 642.0 2.4 n 1.238 173.3 186.1 12.8 Plant b - 1 0.22 0.68 668.0 0 ' su 0.812 163.4 122.1 ~ 41.3 Plant b .2 0.22 0.68 666.0 1.670 1.141 184.2 171.6 12.7 Even tliough only 5 of the 6 data points are credible (i.e., the scatter is less than 28'F), the data set was considered credible r

Determination of CF - All data Must normalize data to best estimate chemical c.omposition and temperature of vessel being assessed and determine best fit line through the " ratio and temperature" adjusted ARTuor values as a function of FF CFem, o,. = 154.4*F 1

, . - - ~ - ., , . . --

w-------- - - - - - - - - - - - - - - - - - m - s.-a _.i-:# .e m. -e .a-. .= ata am.. as, a i..a. u ;uem- _.e,ww_c.m+asa>4e.- , 4_a_. a w_a,_.mm-a e 9

CASE 5 - SURVEILLANCE DATA NOT AVAILABLE FROM PLANT BUT AVAILABLE FROM OTHER SOURCES (cont'd)

, Effects of data analyses technique (Ratio procedure, use of similar data, use of alldata)

Previous analyses Retio procedure not applied, Temperature correction to data frorn other sources not made, All data evaluated RTworlui = -5.0* F; M = 48.3; CF = 149.2*F; FF = 1.354 RTwor,% = -5.0 + 48.3 + (149.2

• 1.354) = 245.3*F Current analyses - Using only Plant "a" data Ratio procedure applied and temperature correction made RTuor,,on , = -5.0 + 48.3 + (144.0

• 1.354) = 238.3*F Current analyses - Using all data Ratio procedure applied and temperature correction made RTuor,,un,n, = -5.0 + 48.3 + (154.4

• 1.354) = 252.4*F

Conclusion:

CF determined from Plant "a" data is the most appropriate Least amount of adjustments to data - same NSSS vendor Environment at Plant "a" is closer to plant being assessed (Plant "y")

b___________. . . _ _ _ . . . - _ _. _ -.. -

?

SUMMARY

OF INITIAL REFERENCE TEMPERATURE DETERMINATION

.i - METHODS FOR WEl,QS

1. PLANT SPECIFIC METHODS A. . ASME Code - NB-2331

.1 B. MTEB 5-2 i

C. BWR Projected Value (GE Methodology) 1

)

II. GENERIC o

t l'

6 e

$N -

EXAMPLES 'OF WIDELY VARYING PLANT SPECIFIC INITIAL REFERENCE TEMPERATURE VALUES FOR WELDS WELD WIRE HEAT PLANT SPECIFIC RTworcu> VALUE 1P4218 -50 1P4218 10 31401 -40 31401 10 31401 40 3P4000 -50 3P4000 10 '

442002 -30 442002 10 1

_ . . _ . _ . ..A

o QUANTITY OF GENERIC AND PLANT SPECIFIC RT,ermVALUES FOR WELDS CURRENTLY IN THE REACTOR VESSEL INTEGRITY DATABASE (RVID)*

• Only those weld wire heats with more than one occurrence, and more than one RT, eye value were included.

WELD WlRE # OF # OF WELDS # OF WELDS # OF HEAT NUMBER WELDS WITH GENERIC WITH PLANT DIFFERENT WITH THIS RT,orm SPECIFIC PLANT HEAT RT,erm SPECIFIC VALUES OF RT,erm 09M057/C109A27A 3 0 3 2 10137 7 2 5 3 1248 3 2 1 N/A (ONLY ONE VALUE) 13253 5 2 3 3 13253/12008 7 5 2 2 1P3571 4 3 1 N/A 1P4218 2 0 2 2 21935 5 2 3 2 21935/12008 4 2 2 0 25295 2 1 1 N/A 27204/12008 5 1 4 3 299L44 5 4 1 N/A 305414 3 1 2 0 305424 3 1 2 2 31401 3 0 3 3 33A277 8 3 5 3 348009 7 6 1 N/A

e OUANTITY OF GENERIC AND PLANT SPECIFIC RTamVALUES CURRENTLY IN THE REACTOR VESSEL INTEGRITY DATABASE (RVID)

(CONTINUED)

WELD WIRE HEAT # OF # OF WELDS # OF WELDS # OF NUMBER WELDS WITH GENERIC WITH PLANT DIFFERENT WITH RTen SPECIFIC PLANT THIS RTen SPECIFIC l HEAT VALUES OF RTom 3P4000 2 0 2 2 3P4955/0951 2 0 2 2 406L44 6 5 1 N/A 4278 3 2 1 N/A 442002 2 0 2 2 442011 4 0 4 2 494K2351/L307A27AD 2 0 2 2 4P4784 2 0 2 2 4P6052 6 0 6 2 4P7216 2 0 2 2 SP62148 3 0 3 2 SP6756 5 0 5 2 SP6771 2 0 2 2 ,

6329637 2 1 1 N/A 71249 8 3 5 2 i

821T44 4 3 1 N/A 83637 6 0 6 2

'. QUANTITY OF GENERIC AND PLANT SPECIFIC RTagVALUES CURRENTLY IN THE REACTOR VESSEL INTEGRITY DATABASE (RVID)

(CONTINUED)

WELD WIRE HEAT # OF # OF WELDS # OF WELDS # OF NUMBER WELDS WITH GENERIC WITH PLANT DIFFERENT WITH RTag SPECIFIC PLANT THIS RT., SPECIFIC HEAT VALUES OF RTeu 83642 3- 0 3 2 83650 2 0 2 2 86054B 10 8 2 2 87005 2 0 2 2 89476 3- 0 3 2 895075 3 0 3 3 89833 3 0 3 2 90099 3 2 1 N/A 90136 4 2 2 0 D53040 4 0 4 2 Mll B-4 8 0 8 5 S3986 7 0 7 4 W5214 7 5 2 0

c L.

!' CONCLUSIONS AND FUTURE ACTIONS

• STAFF IS AWARE OF INITIAL REFERENCE TEMPERATURE ISSUE (i.e. VERY LARGE VARIATIONS WITHIN A HEAT)

STAFF WILL ASSESS ISSUE ON A PLANT SPECIFIC BASIS FOR REVIEW OF. RESPONSES TO GL 92-01, REVISION 1, SUPPLEMENT 1

• . STAFF MAY PURSUE SIGNIFICANT PLANT-SPECIFIC ISSUES ON

. A CASE-BY-CASE BASIS

i ADDITIONAL GUIDANCE ON ISSUES DISCUSSED AT NOVEMBER 12.1997 Best fit Ir.a through surveillance data Question: Should the scatter of the surveillance data about the least squares fit of the surveillance

- data be determined using a linear equation in which the increase in transition temperature (A) has a value of zero at a fluence factor (FF) of zero or should the scatter of the surveillance data about the least squares fit of the surveillance data be determined using a linear equation in which a can

. have a nonzero value when FF is zero? -

Answer The general linear equation describing the telationship between the chemistry factor, CF, A and FF is a = (CF)(FF) + b, where CF is the slope of the line and b is the value of A when FF is zero. Based on this general equation, from Reference 1 the least squares estimate of the chemistry factor from surveillance data (FF,,A,) is given by .

CF = - EFF, A,-(EFF)(EA,)/n (Eq 1)

EFF,' - (EFF,)'/n -

Eq.- 1, however, is not the formula for CF used in RG 1.99, Revision 2, or 10 CFR 50.61.

Credibility Criterion 3 in RG 1.99, Revision.2 indicates that the scatter of the surveillance data should be calculated using the best fit line described in Regulatory Position 2.1. Regulatory j Position 2,1 in RG 1.99, Revision 2, indicates that the CF is determined by multiplying each  !

adjusted A, by its corresponding FF, i summing the products, and dividing by the sum of the squares of the FF,. This results in Eq. 2:

CF= EFF, A, (Eq. 2)

EFF,8 By minimizing the sum of the squares of the +....oons of data from the linear equation A .= (CF) FF, the chemistry factor CF is given by Eq. 2. To minimize the sum of the squares of the deviations from a linear equation, the sum of the squares of the deviations are differentiated and set to zero. The resultant equation is solved for the chemistry factor.

Hence, the reason for using Eq. 2 instead of Eq.1 to determine CF is that, implicit in the equations contained in RG 1.99, Revision 2, the least squares fit of the surveillance data should be performed using a linear equation in which A has a value of zero when the FF is zero. Since Credibility Criterion 3 specifies that Eq. 2 should be used to determine the scatter of the surveillance data, the scatter of the iurveillance data must be fit to a linear equation with b equal to zero to satisfy RG 1,99, Revision 2, Reference 1: " Applied Regression Analysis," N. R. Draper, and H. Smith,1966 a

9

,,. ,r -

. - _ _ __ ._ . _ . - -- .-_.._ _ . - - _ . . ~ - - _ . . . . . _ -

r.

o Non-conservative inhia chemistry factor

- A question was asked regarding the regulatory need to use the chemistry factor determined from the surveillance data in the case when the data are non-credible and the Table CF is non conservative. If the surveillance data indicate that a material experiences more embrittlement than would be predicted from the generic correlations, such information should be factored into a licensee's assessment of RTa. This is consistent with 10 CFR 50.61 Paragraph (c)(2) which states, in part, that "to verify that RT for each vessel beltline material is a bounding value for the specific reactor vessel, licensees shall consider plant specific information that could affect the level of embrittlement." In addition, it appears that this methodology would improve the securacy of the RTn, value significantly (paragraph (c)(3) of 10 CFR 50.61). Although a specific methodology was presented during the meeting, the staff recognizes that other methodologies may provide equally accepteble results. These methods would require staff review and appioval.

Irradiation environment adiustments There was a question regarding references for the 1 *F adjustment when accounting for irradiation environment adjustments. The references are the Electric Power Research Institute (EPRI) report

- NP-6114, and The American Society for Testing and Materials (ASTM) report STP-1046 Vol. 2.

1

_ . _ - _ _ _ _ - _ _ _ - - - .- - - -- --