ML20091K691

Part 21 Rept Re Programming Error in Detector Code,Supplied by Shanstrom Nuclear Assoc,Inc.Initially Reported on 840523. Software Enhanced.Nmfm Procedure 7 Will Be Revised by 841231.LER Will Be Submitted

ML20091K691
Person / Time
Site:	Cook
Issue date:	05/25/1984
From:	Alexich M INDIANA MICHIGAN POWER CO. (FORMERLY INDIANA & MICHIG
To:	Deyoung R NRC OFFICE OF INSPECTION & ENFORCEMENT (IE)
References
REF-PT21-84 NMFM-84-0238, NMFM-84-238, NUDOCS 8406070171
Download: ML20091K691 (7)

Text

_ ._ _ _- _.

. ;u v

INDIANA & MICHIGAN ELECTRIC COMPANY P.O. BOX 16631 COLUMBUS, OHIO 43216 May 25,1984 NMFM 84-0238 o

Donald C. Cook Nuclear Plant Unit Nos. 1 and 2 Docket Nos. 50-315 and 50-316 License Nos. DPR-58 and DPR-74 Mr. Richard C. DeYoung, Director Office of Inspection and Enforcement U.S. Nuolear Regulatory Commission 1717 H Street Washington, D.C. 20555

Subject:

Programming Error in the DETETOR Code Supplied by Shanstrom Nuclear Associates, Incorporated

Dear Mr. DeYoung:

This letter is to confirm the telephone conversation of May 23, 1984 between Amerloan Electric Power Service Corporation and Mr. D. Wigginton, Project Manager, NRC, reganiing notification made pursuant to Title 10 CFR Part 21.

In the process of modifying our code DETECTOR, which is supplied by Shanstrom Nuolear Assooistes, Inc., an error was found in the coding. Our review showed that the coding errer did not constitute a significant safety problem in its application at the Donald C. Cook Nuclear Plant. 1;evertheless, notification has been made to Shanstros Nuclear Associates, Inc. and the NRC.

It is our understanding that other users of the code have been notified.

Additional details on this event are included in the enclosed attachment.

If you require further information please call J.M. Cleveland (614/223-2050) of my staff.

Very truly yours, g6070171e40525 le 8 ADOCK 05000 Vice President 00: Harold R. Denton, Director, NRC D. Wigginton - NRC J. Keppler - NRC, Region III John E. Dolan W.O. Smith, Jr R.C. Callen G. Charnoff E.R. Swanson, NRC Resident Inspector - Bridgman d R.T. Shanstrom 8 j

's .

o usa, v4. m.6u

", * , ' = = = .

UCENSEE EVENT REPORT (LER)

' a ".58 " * *

..=urv a ni Donald C. Cook Nuclear Plant Unit 1 o Is i o io io f 31115 t lopl I

,1,b. 6 6 Discovery of Error in DETECTOR Code

.. , u n . ., . . . .ci v , . . . ..s, .. .

.. . , . . u.... Ca., seven ..

"' M 8 8.' w e , so.as

'",*M ' ' 6 D., v...

h sse.

as ,se S., v. .. , ,,,

o f sto toto t st 0l 5 2 l2 84 8h

~

ll 0 l0 l l 8l4 .-.,-., ,n.

oisioisio, , ,

,,s . , .,n. ..,ver.....w......n.......i,

==.= n. .4 aH nai ..

M.3 4eHit W.794sHaiM f LFt tet

. Ske teH1H. _

et.e l) 0i0 m .awHum _

m.m= = _

w n. aH.= _ L, .,y =,g 30'8EUIO N"I "I NO 88II"#UN'"N8 38048

.p _

nun:H n.

- s.. winH. .'amHsH.

.f 36sie.nla.1 13 eH2Hel N e isH1Het 488. ash. Co.v.C, 9 v.t. L.. f t.t n t.. . . ..

.. 'a"'"'

Michael A. Saum American Electric Power Service Coronratinn 6fl f4 2f Pil t_19101R 17

. ,s . n . u.. . .. . . c _m ., e m u. . .. .c . ... . .. ,... . . . , n si 8

C.W.. 8, ., . as C.es ...f y [ "

C.w.. . fe.as CO. 3..., "[.[ T. , '[ '

^ "

! i e ! I t i t I i i I ! t ,

! I I i t i !

f f I i i i f .,.6 ... i....

. u .. .. 6 . . . , . .. . cu . n ,,,,,,,

g...

' " "u... , , ,

J , u ,,, - -. ,u,ces, ... ... u re, 7e

- n .,

.u, . .C, ,- . u .a - . .. -. ., - ~. :. -

This is submitted as a voluntary LER During the process of modifyinq the DETECTOR code, which analyzes raw flux map data to determine compliance with Power Distribution Technical Snecifications, an error was discovered in the calculational logic. This error was present in DETECTOR version 23, which was used in analyzing the first 47 flux maps taken during Unit 1 Cycle 8. These chanqes were made in August,1983 in accordarice with Nuclear Materials and Fuel Manaqement (NMFM) Drocedure No. 7 Chinqes to the DETECTOR Code. Testinq of this version of DETECTOR, which was carried out at the time the chances were made, did not indicate that this error was present. All 47 flux maps were reviewed and it was determined that no Technical Specifications were violated. Discussion of this conclusion appears in the LER text.

s e a 3 p p '- 6==

• T.7 e= m.

A-1 j

wa .ueu.= ne w een, r 2.

mm

• • ' UCENSEE EVENT REPORT (LER) TEXT CONTINUATION mu em se um e=

e...a u. .

ean ==

• 6...... . i. 7

"- " t'.M 7 ' l'At Donald C. Cook Nuclear Plant Unit 1 o' oisteiolo13I115 814 -

1I -

010 012 I

- ~ . . . , ,,,,,

see attache'd unformatted page for suggested text.

? (

9 i

i en f

e

,/

5 i ,-

.e

- +

t

?

e t s I

e h

a V

f a

k m

A-2 ,

DETETOR CODING ERROR - LER tl*;

3 W

t'

-In August of 1983, modifications were made to the DETETOR code to allow comparison to Technical Specification parameters which varied with Fuel Type.

n These modifications were made by Shanstrom Nuclear Associates, who in fact, was

, y'{ - the original author of the code.

The modified code was tested by making runs on old data sets, was debugged and put into production for Unit 1 Cycle 8. The changes to DETECTOR were carried out in accordance with IBGPM Procedure No. 7, Changea to the DETETOR y fdada.

8 Dinaowery of Error An effort was begun in May of 1984 to modifg the DETETOR code in house to incorporate the ability to suaitor a modified Fan Technical Specification required fy Unit.2 Cycle 5. The modifications Involyp incorporating into the code two F An 1 pits,onerelatedtoDNB(thecurrentF A limit), and a new, LOCA relateE Fg limit. During this process, an ermr as discovered in the logic of the vaj in which DETECTOR compares measured 3 F ,to the Technical

, Specification F' limit. This logio erme first occurr5d in the August,1983 version of DET OR and thus was present in the analysis of the first 47 flux maps taken for Unit 1 Cycle 8.

Mature of Error c The DETECTOR code requires that the input data include Technical Specification limits for each fuel type. With the August 1983 modification to the DETETOR code, it was intended that the relative power of each fuel pin (assemblage) be compared to the limit appropriate for its fuel type. However, an error was made in the coding such that the relative power of each pin was always compared to the limits of the last fuel type in the input data set.

I Tgerefore, the DETETOR output would not indicate the correct margin between T5H and its Technical Specification limit for the first fuel type.

1 It should be noted, that the error affected only one page in the DETETOR l output. Review of other pages could potentially lead to identifying i

discrepancies in the data. The specific error was that a transfer was made to the wrong line of code, i

i b

4 a,i.,-_._.____..___._,._.__..,_,.._._._._.,,..,.._____.__

. 23 pact on Unit 1 Cycle 8 The coding ermr in DETECTOR did not cause a Technical Specification violation during Unit 1 Cycle 8 operation. To justify this statement, one must look at the input going into DETECTOR for Unit 1 Cycle 8 flux map analysis.

There were two sets of Technical Specifications which were applicable for Unit 1 Cycle 8. Technical Specification set 1 was applicable to Exxon Nuclear Company (ENC) fabricated fuel, which applied to once and twice burned fuel assemblies present in the core. Technical Specification set 2 was applicable toWestinghousefuelwhichwasfreshatthestartogUnit1 Cycle 8. The corresponding Technical Specifications limits for Fg input g into DETECTOR were:

Technical Specification Set 1:

FAH(1) 1 1.45[1 + 0.2 (1-P)]

Technical Specification Set 2:

FAH(2) i 1.49[1 + 0 3 (1-P)]

where P is the ratio of actual thermal power to rated thermal power (RTP).  ;

In all cases DETECTOR compared F to the Technical Specification limit for Technical Specification set 2 (WNtinghouse) regardless of whether the FfH was associated with an ENC (Technical Specification set 1) or a Westinghouse '

( echnical Specification set 2) fuel assembly. Thus if a An greater than 1.45[1 + 0.2(1-P)] occurred in an ENC fuel assembly it might not have been indicated as a violation of the Technical Specification limit by DETECTOR.

3 To verify that this did not ocogr, Flux Maps 1 - 47 for Unit 1 Cycle 8 were analyzed to detgrainefor whether ENC fuelany F 3h =for ENC fuel was greater than 1.45 (the with agst limiting FAu 1.0). No maps wgre identified where Fgg (ENC) was greater than the Technical Specification FAH limit for ENC fuel, and therefore there were no Technical Specification violations.

Once satisfied that no Technical Specification violations had occurred, the possibility that the most limiting Technical Specification margin edit did not contain completely accurate inforgation was investigated. Specifically, the possibility existed that an ENC Fan was closer to its Technical Specification limit than the most limiting TechnIgal Specification margins printed out for the Westinghouse fuel. Since the F3 for ENC fuel would be compared to the Westinghouse limit, which is higher Uhan the ENC limit, this ENC fuel assembly (or pin) might not be included in the most limiting Technical Specification margins edit.

This in fact did occur on two flux maps, 108-04 and 108-05. However, these maps were taken at BOC, < 50% RTP, with the Technical Specification margin for the most limiting pins approximately equal to 0.20. Therefgre, the fact that ENC fuel assemblages were not listed on the most limiting Fou edits does not appear on the basis of engineering judgement to be significalit.

One should gote also that from a core analysis of the Unit 1 Cycle 8 core, the hot spots FAH and Fg G,0 will occur in hesh N1 assemmes once equilibrium HFP core coilditions are reached. This was confirmed by the  ;

analysis of all Unit 1 Cycle 8 flux maps.

i i

i . Possible Impact on Unit 2 Crale 85

-It is difficult.to postulate whether the error would have been discovered if the Unit 2 Cycle 5 Technical Spoo fications had not required modification to include the addition of LOCA based A limi a ions. If we assme hat ne error would not have been discovered,Hwe can look sg the two cases and see the potential outcome. In either case the applicable FAH Tech Spec Limits for the two difflerent fuel types are:

N 7

Exxon Fuel: F H I 1.49 [1.0 + 0.2 (1-P)]

Westinghouse Fuel: Fe8 1 1.48 [1.0 + 0.2 (1-P)]

Case 1 In this case Exxon Fuel would be assigned to Technical Specification set 1 and Westinghouse Fuel to Technical Specification set 2. One should note that j the Unit 2 Cycle 5 core consists of one region (twice burned) of Westinghouse '

fgel and 2 regions '(once burned and fresh) of ENC fuel. In this case, the peak j F g ,ification limit for Westinghouse Fuel.oosurring Sp5c However, in the the ENC fuel, would have been comp FAu Techniogi Specification limit for Westinghouse is more conservative tHan the Fa Technical Specification limit for ENC, therefore this would not have b en a problem. Furthermore, it is believed that this problem would have been ipntified immediately upon analysis of the most limiting pins on the

, FaH 1 west Technical Specification margin edit.

Came 2 In this case Westinghouse Fuel would be assigned to Technical Specification set 1 and ENC Fuel to Technical Specification set 2. This case is similar to what actually occurred in Unit 1 Cycle 8 in that the fresh fuel Technical

! Specifications were input as the second Technical Specification set. The fresh l fuel Technical Specification limit would be applied to all fuel. This is a nonconservative comparison for the Westinghouse fuel. However, since the Westinghouse fuel is twice burned and consequengly operates at low power, it is highly unlikely that this fuel would reach an Fa as ig as a own Unn or H

l the marginally higher Exxon limit.

l

.-------,,,-,.,,,--,,.,.,,,---.-,,---,----,,_.,,.,_,-,,.n -

. - - , - , , , , . , , _ ,..,,,ma.,w-,,_.,n,,,nn,,.-_.,...w-.w-,-n._--,,,..,m,,.yww. - _ , - , , , -

Corrective Action The coding error will be corrected in conjunction with the other DETH: TOR modifications being made for Unit 2 Cycle 5.

The two flux maps that indicated the wrong most limiting pins on F Unit 1 Cycle 8 maps 108-04 and 108-05, will be remn with the corrected DETb,OR version.

AEPSC has changed their SOURCE library disk file management system on the corporation computer system from SOURCE to LIBRARIAN. LIBRARIAN offers a much more thorough method of maintaining an accurate audit trail of changes made to a program than previously existed with SOURCE. It is believed that this software enhancement, will reduos the possibility of future code modifications being in error.

It was determined prior to this event that the procedure which contmla changes to the DETECTOR code, NMFM Procedure No. 7, Chanaan to the DEIECIgjl f.ede_, should be revised to assure that not only are test cases m n, but that an independent line by line review of the coding is performed. This procedure will be revised by December 31, 1984.

. _-. . - - - - .