ML20072N549
| ML20072N549 | |
| Person / Time | |
|---|---|
| Site: | Columbia  |
| Issue date: | 03/28/1983 |
| From: | Chu A, Michejda O, Studnicka Z BURNS & ROE CO. |
| To: | |
| Shared Package | |
| ML20072N544 | List: |
| References | |
| REF-PT21-83 NUDOCS 8307150362 | |
| Download: ML20072N549 (17) | |
Text
. _ .
,a l
EVALUATION OF ADLPIPE ERROR IN TEE CLDtENT RELATING TO DEFECT AND NON-COMPLIANCE EVALUATION REPORT NO. 83-01 Washington Public Power Supply System l
Nuclear Power Plant #2 March 28, 1983 l
1 l
I Prepared by: {W
__A. Chu Reviewed by: 2 k
---~StIEl'nl~cYh Z.
Approved by: - -'
D.Illche3da W
S
,s ' r
. i INTRODUCTION This study is in response to a letter from B. C. Hanley of DIS /ADLPIPE to H. Boyd of United Information Services, dated 9/28/82, stating that for Class 2, 1972 analysis the section modulus used for reduced outlet branches was actually that of the run and not that of the branch. The letter was included as an attachment in a memo, #CS9182 Ref. (1) to O. Michejda, dated 12/2/82. After an inquiry to DIS /ADLPIPE, Burns & Roe was informed on 12/20/82 that an error definitely exists in the in-house version of ADLPIPE, Ref. (2). The Piping Engineering Department immediately formulated a cor-rectio$procedureandstartedtoreviewallthecalculations in the Washington Public Power Supply System Nuclear Power Plant #2 project. The applicable codes for the project are ASME Secti6n III, 1971 edition through 1973 Winter addenda and ANSI B31.1 1973 edition, Ref. (3). WNP-2 project used the ADLPIPE program containing the above mentioned error with some few exceptions where the 1974 ASME Class 2 code was used which is more conservative than the 1972 Code.
As stated in Ref. (1), the calculations would contain an error if the following conditions occured concurrently:
- 1) The piping system had reduced branches,
- 2) The fittings were modeled with TEE, 1T, 2T or BRANCH cards,
. 3) ASME Class 2, 1972 code was used.
l l
1
The review of all calculations showed incorrect stresses at sixty seven nodal points. A procedure was prepared to assess the effects of the error on the calculated stresses at the identified locations.
PROCEDU_RES, Correction factors for stress evaluation were calculated from the dimensions of the fittings to acc.ount for the incorrect calculation of the branch section modulus in accordance with Article NC3652.4 of AS?tE B&PV Code Section III (Ref. 4)
Z r rg't R Correction Factor =
2 Z rt b e Where:
r = n minal mean cross-section radius of pipe connected to j R
the run portion of the tee i rb = nominal mean cross-section radius of pipe connected to the branch portion of the tee Z = section modulus of the run portion of the tee Zb = section modulus of the branch portion of.the tee t
R
= thickness of oipe connected to the run portion of the tee tg = effective branch wall thickness, lesser of t g or i x 't b
2
t = thickness of pipe connected to the branch portion of the b
tee i = stress intensification factor for teo The erroneous ADLPIPE stresses were multiplied by the above correction factors. Later the procedure was confirmed by a letter from I. W. Dingwell of DIS /ADLPIPE to T. Butler of Burns & Roe, dated 1/10/83 (Attachment A) . ,
The results including ADLPIPE output stress, corrected stress and allow-able stress are summarized in the Calculation No. 8.90.38 Book o. 0.90.16.
The ADLPIPE stresses for equations 8 and 9 consist of two terns; stress due to pressure and stress due t'o moment. Only the stress due to moment was affected by the incorrect section modulus in the ADLPIPE program. In the process of correcting the ADLPIPE stresses, correction factors were applied to the printed stresses as indicated in Attachment A. This results a conservative stress values. If overstressed condition was found, the printed stress was broken down into two parts and l then stress due to moment only was multiplied by the correction factor resulting in actual stress values.
1
{
l C_ONCLU_SI_O_N All corrected stresses were compared with the allowable limits.
3 1
l l
No overstressed condition was found. The original calcula-tions are being revised to correct the stress values at the affected locations.
i l
4
REFERENCES
- 1. Memo, T. Butler to O. Michejda/A. Dajani, dated 12/2/82, CS9182, UIS System ADLPIPE Errors.
- 2. Memo, T. Butler to O. Michejda/P. Chan/H. Brandmaier, dated 12/20/02, CSlC982, Possible Tee Errors in ADLPIPE.
- 3. Piping Design Guide for Washington Public Power Supply System Nuclear Project No. 2.
- 4. ASME B&PV Code, Division I,Section III, 1972 Winter-Addenda.
4 e
W 9
e O
4 5
- L/iJ/.il )Li'l 1t., InC.
p., ;. s _. .. . e. . .
. . : .e
. . . ' ' n.:
ATTACHMENT A -
Janua ry 10, 1983 Via Federal Express Mr. Thomas Butier BURt!S & ROE, INC.
650 Winters Avenue Paramus, NJ 07652
Dear Mr. Butler:
The purpo'se of this letter is to provide you with information con-cerning a forthcoming change in ADLPIPE, and to advise you that if previous revisions of ADLPIPE were used in the design of a nuclear reactor under specific circumstances, an evaluation under 10 CFR Part 21 may be required.
The May '197t release introduced into ADLPIPE the requirements of ASME Section III (Nuclear Power Plant Components) for Class 2 piping (Subsection :E,1972 Winter Addenda). tiew editions of Subsection lC
' were published by the ASME in 197a and 1977 The requirements of
.here codes were implemented into ADLPIPE in April 1975 and
-: ovember 1951 respectively. l i
An error has recently been discovered in the May 1974 release of ADLPIPE and all subs.equent versions related to reduced outlet branch tee fittings whose stress was evaluated under Class 2,1972 Winter Addenda. This error does not affect calculations unde- the 1974 and 1977 editions of Subsection flC.
For evaluations under the' Class 2,1972 Winter Addenda, the calcula-tion of the section modulus of the reduced outlet branch (Paragraph t!C.<3652.4(c)) incorrectly substituted the run section modulus for the branch section modulus.
It is suggested that all Class 2,1972 Winter Addenda stress calcula-
.tions for the reduced outlet branch be reviewed. The following factor -
should be applied to the printed stress at the intersection of the run and reduced outlet portions of a tee: . .
OEVELOPER OF DEEIGN INFORMATION SYSTEMS An A5500iatec Techno!o;;iesincorpo.ated Company ATT. A, PAGE l'
. ' DihADLFiPE,Inc.
e Janua ry 10, 1983
,]'
ATTACHMENT A Mr. Thccas Butler EURNS & ROE, IHC. -
(a) for cases where tr ' Itb factor = "2 2
'b (b) -for cases wherert > it b 2 t factor = "
r b
2
- it b where: .-
r = ncminal mean cross-section radius of pipe connected to the
, , . run portion of the tee l I
r
- . -b = ncminal mean cross-section radius of pipe connected to the branch (reduced outlat) portion of the tee .
t r= thickness of pipe connected to the run portion of the tee t b= thickness of pipe connected to the branch portion of the tee i = stress intensification factor for tee (printed out in Class 2 Strass Peport).
The above factors apply for reduced outlets for TEE,1T, 2T and BRANCH
, connections.
- This correction will be made in a future change to ADLPIPE.
If the AOL?!?E code was used in the design of a facility licensed by the United States Nuclear Regulatory Commission (NRC) under circumstances where the foregoing change is pertinent, we request that the user perform an evaluation to ascertain whether such use could create a " substantial safety .
hazard," as defined in 10 CFR Part 21. If a substantial safety hazard could be created, we further request that the NRC be notified in accordance with the requirements of Part 21 and that we be advised of such notification.
Very truly yours,
] ,
.* ! : .,:an. i u S'. . p . l l
I. W. Dingwell President ATT. A, PAGE 2
6di c. a ra sp a t va g a hJ iLb n!J d COPIES TO:
. BURNS and ROE, Inc. oaTe .i ./ n / n 3 .i i . , , i ,, ,, , , , , . r
<!b To E. R. Kummerle FROM O. fliche j da CUBJECT Poten tially Reportable De ficiencies No. 82-12 and 33-1 The errors in ADLPIPE program identified by the Potentially Reportable Deficiencies:
No. fll-1.2 - Bellows Skew Card and No. 83 Branch Tee Section flodulus have been evaluated for all the nuclear prhjects within Burns and Roe, Inc. Two reports, one for "each deficiency, have been submitted previously for Nashington Public Power Supply System Project, ICIP- 2 .
The'.other projects have been evaluated and the results submitted in documents attached to this meno. The table below relates the copies of the attached documents to the tuo reports and the projects.
.EST. - 8_2_ _12, R_ep ., _8_3 - 1, Cooper Station Att. 1 Att. 1 Oyster Creek Att. 2 Att. 3 Tf1I II Att. 2 Att. 3 and 4 Phil. Nuc. Plant Att. 2 Att. 3 PSET.G Att. 5 Att. 5 CRBRP Att. 6 Att. 7
/Ibf'14f0 -
.v ~, .?-
O. ttichejda ~ ~ ; ~ .:-.: :n; .;.
Ot1/ha s attachments E y,cc pac 3,ggn7 rnainttR1NG AND DCSIGNI DIVISION APR 81933
--l
_ _ l .. _
l q
c \
m.~~;~;~
t
\\
,' , ATTA 'llf ti:UT I COP 133 TO:
l 3 URNS and ROE, Inc. oara 3/ir./83 m ai>et
' PI' lievised 3/31/83 db TO O. Michejda PROM J. Glaser j
i 3UCJCCT Nebraska Public Power District Cooper Nuclear Station i Brownville, Nebraska Errors Discovered in "ADLPIPE" Calculations
- 1. Skewed Bellows (82-12)
- 2. Reduced outlet tee fitting (83-1)
The 2 "ADLPIPE" errors recently discovered had no effect on the subject nuclear plant. "ADLPIPE" was never used during the design and licensing of the subject plant. The BRI Computer Program was used for " thermal" calculations and EDS was contracted to
- perform the seismic calculations utilizing their own
- program.
) , Investigation also showed that during the NRC i 79-02 and 79-14 review of the pipe supports in the subj ec ti, plan t , the 2 "ADLPIPE" errors had no effect on the results. The skewed " bellows" and " flexible"
~
input cards (subject error #1) and the affected versions of ADLPIPE (subject error #2) were never used in the review calculations.
s
.f !
f ,.
/cd .
a t
e F
9 l
i
s
^T * ' n .a r e a b bJ 2 4 2 L G D 0 0 W J ,
CC PICS TO:
J
- A.D. Dam E3URL'JG and RGE, Inc. o ^ T c 3 /31/03 j n. cagliardo T. 13u tle r J. Sunol T
O. Michejda L.h. Zuchowski F H o r41
'Pf(2) 10*02A J. Kiven j 75.50 SUDJcCT W.O. 3344 s Westinghouse Electric Corporation Philippine Nuclear Power Plant Unit No.1 Error in ADLPIPE coding-Bellows-preceded by Skew card Reference Letter: Dingwell(Disadipipe Inc.) to Butler (B&R) dated October 22, 1982.
The reference letter states that a problem exists with t.he Bellows card. in the IBM version of ADLPIPE only.
Until the time when ADLPIPE was adapted to the Burns and
~
Roo machine, which is a rather recent development, the CDC versi8n of ADLPIPE was used at Burns and Roe either through CDC - Minneapolis or through The CDC machine at Grumman.
This period covers all work done for TMI-2, and those cal,culations are not affected by this problem.
Work done more recently for The Philippines Project and for Oyster Crock has been reviewed and no use of the in-valid coding has been-found.
. Kiven JK/kr I
Lwul L u d V L w L. A L J L' 4> ud e J col'IES TO: )
.. UUNNC CMd CCC,Inc. oayn 3/1/03 ADDam PGagliardo j To O. Michejda Tuutler JSudol F R O M4 J. Kiven LAZuchowski pf (2) 10.02A CUBJECT *
- 75.50 Westinghouse Electric Corporation Philippine Nuclear Power Plant Unit No. 1 TEE Errors in ADLPIPE Ref. 1. Memo T.B. to O. Michejda/P. Chan H. Brandmaier dated 12/20/82, Same Subject Ref. 2. Letter B. Hanley, (DIS /ADLPIPE) Inc. to Tom Butler, (B&R) dated 1/10/83
)
References 1 and 2 state a tee calculation error exists in ADLPIPE for'the calculation version based on ASME Section III Subsection NC for Class II piping as issued in the 1972 Winter Addenda and brought out by ADLPIPE in May 1974. However, other calculation options exist in ADLBIPE for Class II piping which are correctly coded.
A review of ADLPIPE stress calculations for projects under my cognizance has been made in accordance with your request to determine whether our analysis has ever used the Class II 1972 option. It appears that this version was never used and no errors have been introduced. The project calculations reviewed were TMI-2, Oyster Creek and Philippines Unit I.
In TMI-2, considering the very large number of calculations, a 10% sampling was made with the effective conclusion as stated.
In the other projects 100% of the calculations were checked.
e d
p
- J. Kiven JK/dm
(
e e+3 I
O e k'eE,lau d O h'N N b b q COk*lCGTO:
- n A.S. Dam GURNS and FEOE, Inc. onys 4/5fe3 1 n, nsaipota M. Maher T. 11ut'J er TO O. Michejda / R. Gagliardo i pf (2 )
FROM J. Kiven C U C.J E C T W.O. 3775-04 G.P.U. Nuclear Corp.
Three Mile Island Unit #2 TEE Errors in ADLPIPE
Reference:
Memo Kiven to Michejda same Subject, dated 3/1/83 The reference memo states a review of 10% of the approximately 700 TMI-2 pipe stress calculations indicated that the incurr-ectly coded ASME class II Winter 1972 option was not used.
The option that was used referenced the B31.1 code which is equivalent to ASME Section III Subsection NC which covers Class II piping.
However in those piping configurations to be analyzed that were p.artly Nuclear. Class I and partly Nuclear Class II, the B31.1 option was not available. In these cases the coding had to reference the 1972 Asnc code where the validity of the procedure for calculating stresses in Tees is.ih question.
A study of appropriate flow diagrams determined that a maximum of 6 " mixed" Class runs existed.
A detailed review of existing analyses for the 6 configurations indicated that all,of the designs are acceptable. .
O V
~ /f ( s ? n /i?^)
// J. Kiven JK/kr 9
. lua s=, AW8 U tt J4 N D Le hgg u rnemu.:n :,
COPIES TO:
BURNS and ROE,Inc. DATE 3/31/83 JDel/>oper DGoldner To O. Michejda pr FROM D. Goldner GUBJECT
- Computer Program This is to inform you that errors in the ADL Pipe Computer Program have not affected Burns and Roe calculations for PSE&G's Salem 1 and 2 piping systems.
This is due to the fact that these calculations were performed using the PIPDYN II computer p'rogram. This program was developed by Franklin Institute for PSE&G's use on the Salem projects.
iff(,&W DG/gg D. Goldner 9
l 1
1
m ,. ......v.
QA-364-82
,' BU,8NS and ROE, Inc. carc 11/4/a2 ncy3cy commtetcc:
. M. C. Ascher To Messrs. K. A. Roe (via 0.0.P), S. Baron,
- I* C '#i "
W. H. Young . A. Dajani FROM E. J. Corrigan [. p. l p,'
O U BJ E CT 10CRF21/10CFR50. 55 ( e) Evaluation
- Licensing:
of ADLPIPE Program Proolem . A. Lageraaen (WPPSS WNP-2, REFERENCE': BRD-L-2.3, Reporting of Defects and Woodbury)
Noncompliances . W. Conn (WPPSS WNP-2, Richlani
. F. Patti (all other BR Mucle-Projects)
- W. P. Rausch O '
- pf: 1.1, 2.8,6.6,
. 101.11
'. sc
- W/ Attach. 1 only
.~.
' Attachments 1 and 2 are provided as required by the referenced procedure.
The attachments describe the evaluation that was performed by the Breeder Reactor Division for the Clinch River Breeder Reactor Plant to determine re-portability under the law for the ADLPIPE Program problem.
em
)) The attachments reflect the decision that the pro-blem described therein is not reportable under either
f 10CFR21 or 10CFR50. 55 (e) .
EJC:ic Attachment 1 - 10CFR21/10CFR50. 55 (e) Evaluation Report dated 11/4/82 Attachment 2 - 10CFR21/10CFR50. 55(e) Evaluation Review Committee Meeting Minutes
- s. *
- 1 O' i m. i , a w b m . , . .. . , . i o , . . e. . . . . :. . . . i
' Burns and i * , Inc.
700 Kinder 6arack Road
- Oradell, fl. J. 07649 PROJECT: Clinch River Breeder Reactor Plant EVALUATI0tl REPORT # BRD-82-1 WORK ORDER: 3067 & 3070 DATE: flovember 4 1982 SPECIFIC AREA 0F C0tlCERil:
IBM Version of ADLPIPE Computer Prngram - 2C, 202 and 2012 OUTSIDE C0tlTRACTOR IflVOLVED:: DIS /ADLPIPE, IfiC., 55 Wheeler Street, Cambridge, Ma. 02138
- 1. DESCRIPTI0tt 0F POTErlTI AL DEFECT OR fl0tlCOMPLI AtlCE: Discrepancy between piping support loads calculated using the ADLPIPE IBM version of the program (2C, 202 and 2gl2) when compared to the CDC version of the program when spring stiffness 3 10 were used in modeling a skewed support as a bellow element.
- 2. DATE Afl0 METHOD OF DISCOVERY: BR Corp. QA meeting report received 10/26/82.
- 3. AtlALYSIS OF SAFETY IMPLICATI0ti: All stress calculations in safety related systems p were reviewed to determine whether the subject program had been utilized. Only four calculations were identified as having used the program, f4cne of the calcula-tions had proceeded to the approval stage for incorporat, ion into the design of a facility, component or activity related to the project. Further, none of the cal-culations included the conditions (stiffness values) which could have resulted in a safety hazard.
The defect was not the result of a failure to comply with any applicable rule, regulVtion, order or license issued by the fluclear Regulatory Commission.
Accordingly, the defect is not reportable by the criteria of 10CFR21. Based on the licensing phase of construction (i.e. non-safety related site preparation activities) a potential 10CFR50.55(e) deficiency does not exist.
4 OTHER FACILITIES WHICH MAY BE AFFECTED: Program Users being identified by outside contractor DIS /ADLPIPE, Inc.
- 5. CORRECTIVE ACTI0ft (TAKEft/ PROPOSED):
DIS /ADLPIPE, Inc. has issued a corrected program (Change 469) and is notifying users of the IBM version of the program of the potential problem.
m
- 6. REQUIRED REVIEW AttD IrlDIVIDUAL DETERMIrlATI0ti: Eng. Study Sionature/Date Reportable flot Reportable Required Project Manager 'MC' d h 'Yb, X QA Manager [ b ,- m $s/n #
X Ass,t. P.M. Engr. & Des. Ser. - ?2TN%' )b 6 Cog. Section Manager Lic. & Environmental Mgr.
C e4 U [
8'$N.w </$b h I I
Initiating Employee flot Jdpli[able
- 7. EVALUAT10tt BY BRD VICE-PRESIDEllT: flot required.
- 8. ACT10ti TAKEft: fio Burns and Roe,.Inc. action required on Clinch River Project.
$. . w n/V/C2 i3: 3o QA Manager p' Date Time 1
l
' .' 'IVIE MORANDUM sA-oao-sa COPIES TO:
' BURNS and ROE,Inc. o,7g 4f7fg3 3 Archer l OMichejda TO A. T. Dajani, ECorrigan Wf-(,%, M PTimar PROM W. H. Gibton Pf 1.1 sc cusJacT ADLPIPE TEE Errors I called Dr. P". W. Chan on Friday, April-1, 1983 and asked him whether ,1 review of CRBRP stress analyses was performed relative to the ADLPIPE reduced branch TEE problem. He replied that such a review i.ad been performed and that no analyses were made using the 1972 Version (Section III, Winter Addenda 1972). As a counter check I reviewed some of our existing calculations and found this to be. true. .
CRBRP therefore has no further required action on 'this point.
WHh/lv -
O I
l i
i l
l l
l