ML20091D795
| ML20091D795 | |
| Person / Time | |
|---|---|
| Site: | Yankee Rowe |
| Issue date: | 03/31/1992 |
| From: | Bozarth D, Cheverton R, Dickson T, Merkle J, Minarick J, Nanstad R, Simonen F, Ward L, Kevin Williams Battelle Memorial Institute, PACIFIC NORTHWEST NATION, IDAHO NATIONAL ENGINEERING & ENVIRONMENTAL LABORATORY, OAK RIDGE NATIONAL LABORATORY, SCIENCE APPLICATIONS INTERNATIONAL CORP. (FORMERLY |
| To: | NRC OFFICE OF NUCLEAR REGULATORY RESEARCH (RES) |
| References | |
| CON-FIN-B-0119, CON-FIN-B-119 NUREG-CR-5799, ORNL-TM-11982, NUDOCS 9204130054 | |
| Download: ML20091D795 (149) | |
Text
_
4
- NUREG/CR-5799 ORNL/TM-11982 Review 0:? Reac:or Pressure Vessel Evaluation Recor' :for
' Yankee Rowe Nuc~ ear Power Sta: ion TYAEC No.1735)3
\\
1
, : Prepared by R. D. Chever*on T. L Dickson. J. G. Merkle R. K. Nanstad
- Oak Ridge National Laboratory ll Prepared for
,.U.S. Nuclear Regulatory Commission
$k kD KO O29 p
_ -. - -. ~ _ - - - -. ~, - -
e' AVAti,ASILITY NOTICE AvaJaMty of Reference Matends C'ted in N3C Put*catons Most documents cited in NRC putnaJons wG be evahabie from one of the following sources:
1, The NRC Pubhc Document Roorn, 2120 L Street, NW.. Lower Lever, Washington. DC 20555 2.
The Suoerintendent of Documen's, U S. Government Pnnting Offtec. P.O. Box 370b2, WasNngton, DC 20013-7082 ~
3.
The Nationa! TechnicalInformation Service, Springfield, VA 22161
.1 Although the listing that fo!!ows represen:s the majorny of cocuments cited in NRC pubhcations, it is net.
Intended to be exhaustive.
Referenced documents avaRab;e for inspection and copying for a fee trom the NRC Public Occument Room include NRC correspondence and internal NRC rnemoranda, f4RC bunetins,. circutars, informatlon notices, inspection and investigation notices, keenses er9nt reports: vendor reports and correspondence; Commis-slon papers: and appiicant and licensee cocements and correspondence.
The following documents in the NUREG series are avaHable for purchase from the GPO Sa;es Program, formal NRC s'aff and contractor reports. NRC-spensored conference proceedings, international agreement reports, gra pubfications, and NRC booklets and brochures. Also avaHabie are regJatory guides. NRC regulations \\n the Code 01 Federal Regutecnu and Nuclesr Regu;atory Commiwon Issuancea
. Documents avaliste from the National TechrJeg information Service include NUREG-series reports and technk:alreports prepared by other Feceraf agencies and reports prepa ed by the Atomic Energy Commis-sion, forerunner agency to the Nuciear RegJatory Comrnsslan Documents ava!!abie from pubuc and special techn! cal hbraries inctude all open hierature items. such as books, }oumai articies, and transactions, Federaf Regnfer notices, Federal and State legislatkin, and con-x
. grossional reports can usua!!y t'e obtained from these libraries
- Docurnents guert as tneses, dissertations; foreign reports and translations, and ron-NRC cc ference pro-coedings are ava!!able for purchase from the organization sponsorrg ths-pubhcation crted Sngle e. pies of NRC draft reports are ava!Iable tree. to the extont of suppry, uper wTitten request to the J
Office of Admin!strat;cn, Distribution and Maa SeMees tecton. US Nuclear RegWatory Commimen, Washington; DC 2Cf 55.
Copies of industry codes and standards used in a substaative manner in the NRC regu!rtory pros:ess are maintained at the NRC Ubrary,7920 Norto:k Avenue. Bethesda; Maryland, for use by the putMc Coaes and standards are usualty copyrighted and may be purchtsea from the o,ginating organization or, if they a4 o American National Standards, fros.1 the American Nat onal Stancards Institute.1430 Bre'id.vay, Neo VM NY 100181 1
Dic. CLAIMER NOT:CE Ths repart was prepared as an account of work sponsored by an agercy of ine Un.ted Statm Govemnwnt.
Neither the Unmi Gtates Oovememnt nor any agercy therect, or any of tner empicyees.makes any wa rant 6 expressed or implied, or aS3urv4 any kga! 1 abihty of responsibMy for any third pays use, or the resuits of l
Such ute, U* viy iruarrnation, apparatus, prcdact or p'ocess d"wimed in this report, cr represents that its use by such th/d carty ww!d not infr:nge povately owned nghia.
j l
o 2-I" t-i o
NUREG/CR-5799 ORNUTM-11982 RF Review of Reactor Pressure Vessel Evaluation Report for Yankee Rowe Nuclear Power Station (YAEC No.}1735) i.
Manuscript Completed: October 1991 Date Published: March 1992
.:\\
i Prepared by R. D. Cheverton, T. L Dickson, J G. Merkle, R. K. Nanstad j
Contributing Authors i
l D. P. Boi.arth', J. W Minaricki,' K. A. Williams 2, F. A. Simonen3 I. W. Ward 4 l
Under Centract to:
. (hk Ridge National laboratory.
l Operated by Martin Marietta Energy Systems,Inc.
Oak Ridge National l2boratory Oak Ridge,'IN 37831-6285 i
i Prepared for.
Division of Engineering -
Omcc of Nuclear Regulatory Research
- U.S. Nuclear Regulatory Commission Washington, DC 20555
- NRC FIN B0119
. Under Contract No. DE-AC05-840R21400 1 Science Applications International Corporation, Oak Ridge,TN 37836 rScience Applications Internat;0nal Corporation, Albuquerque, NM 87106 l-3 Pacific Northwest lateratory, Richland, W A 99352 l
didaho Nuclear fingineering Laboratories, Rockville, MD 20852 I
Abstract The YarAce Atomic Electric Company has perfctmed Also, the two codes have a few dissimilar peripheral an Integrated Pressurired Thermal Shock (IPTS)-type featusrs. Asi& from these differences, VISA Il and 7
evrJustion of the Yankee Rowe reactdr pressure vessel OCA P ve very similar. With enors contcted and an in axordance with the 1% Rule (10 CFR 50.61) and adjustment made for the differences in the treatment of U.S. Regulatory Guide 1.154, Upon receipt of We cor-frac'_ure-toughness distnbution through the w all, the respondmg dwument (YAEC 1715), the Nuclear tevo codes yield essentially the same value of the coadi-Regulatory Commission requested that the 02k Ridgr tional probability of failure.
Nadonal Laboratory (ORNL) review tlkt VAEC dxu-ment and perforrn an in&p:ndent proiubihsuc factur>
De ORNL irdependent evaluadon indicated RTNOT mechanics analysis The ORNL review included a h+
values considerably greater than those corresponding to tailed comparison of the Pacific Northwest Laboratory the FTS-Rule screening critena and a firquency of tad.
(INL) and th: ORNL probabilistic fracture-raxhanics ure substantially greater than that corresgundtng to the cois (VIS A-Il and OCA P, resgrctively). The revicw
" primary amptance entencv in Reg. Guide 1.154.
i&ndfied minor errors that were subsequently corrected Time constraints, however, prevented as rigorous a and one significant difference in philosophy with regard treaunent as the situadon deserves. Thus, these results to the variation of fracture toughneu through the wall.
are very vehminary.
h t
iii NUREG/CR-5799
E Contents hac Abstract
. ill J
......................m.....e..
Preface m...........
vii
.....m..m
~.....................
m........m
- Foreword ~........., ~..
lx 1-Introduction -
l
.....~..~ m. ~. -..
. - - -.. - - -.m m. -~~
i 3
1 2-Seope of Review
...............................................,.............................. ~.........
3 PTS Transients and Their Frequencies (Appendices A and B).....
3-3 4 PTS Transient Thermal / Hydraulics (Appendices A and B)..
5 ' Radiation Effects (Appendix C)..............,.,
4 4
.l 5.1 Increase in RTNDT ~ --~....o....
.. ".... u. ~ -.. ~ ~.
4 5.2 Decrease in the Upper. Shelf Energy.....
6 I'TS Fracure Mecleics (Appendices D anJ E) 5
' 6.1 Comparison of the PNL and YAEC Venions of VISA.Il(Appendix E) 5-
.........,,,............m
- 6.2 Comp.vison of OCA.P and the PNL Version of VISA Il......
5 6.3 ORNL OCA.P Analysis of Yankee Rowe (Appendix D) 6 7z ORNL Estimation of Frequency of Failure for Yankee Rowe -
7
~ 8 Conclusions 7
.m......
.m....
9 Referenecs 8
......,.... -....... ~.........,,............,,......,,,..
Apperdix A: SAIC Review of YAEC No.1735 A,I
- Appiix B
- : Technical Evaluation Report-Review of the YAEC Thermal 11ydraulic Accident Sequence Analyses for Assessemnt of Pressurim! Thermal Shock for the -
- Yankee Rowe Nuclear Power Statkm (INEL Review of YAEC No.1735)L.......-.........
H.1 -
Appendit C: ORNL Review of YAEC No.1735 - Radiation Effects on RTNDT arid Charpy.
Upper.Shell Energy C.1
- Appenda D: ORNL Review of YAEC ho.1735 -IMabilistic Fracture Mechanics D. i 1
E.1 --
Apperwlix E:- PNL Review of YAEC No.1735 A
)-.
l v
NUREO/CR 5799 i
I
'I
~, - - _ _.... _.. _ _.. _ _., __,._
..m_
A 3-Preface "Ihis report was originally submitted to the United tionia request was rnade by the NRC to mmimite -
- States Nuclear Regulatory Commission (USNRC) on changes, including editing and graphics. For this -
- November 5,1990, as a draf t of a letter report. More reason, the report does not meet the usual high
. recently, there was an urgent need on the part of the -
star.dards of the Oak Ridge National Laboratory nor NRC for the document to te published as soon as pos-does it fully conform to the NRC NUREO report -
, sine as a NUREG report, and to expedite publica-format.
-l 8
l t
a 4
?
^
y A
h r
l vii NUREO/CR-5799 4
a
+
n -
-,,wr---
r-
l FOREWORD The work reponed hese was performed at Oak Ridge 9.
T, R. Mager, Weannghouse Electric Corpora-National Laberatary under the Heavy-Seenon Stcci tion, PWR Systems Division, Piusburgh, Pa.,
T echnology (llSST) Program, W. E. Pennell, Program Post.frradiation Testing of 2T Conyuct Ten.
Manager. ne program is sponsored by the Office of '
sion Specimens, WCAP-7561, August 1970.
Nuclear Regulatory Research of the U.S. Nuclear Reg 10.
T. R. Mager, Wesunghouse Electric Comora-ulatory Commission (NRC). hc techmcal :nonitor br bon, PWR Systems Division, Pausburgh, Pa.,
the NRC is M. E. Maytir!d.
Fracture Toughness Characterization Study
- I
.nis report is designated HSST Report i16. Prior and 7S I r I h0 future reports m this series are listed below.
i 11.
T. R. Mager, Wesunghouse Electric Corpora L
S. Yukawa, Genetui Electric Company, tion, PWR Systems Division, Pittsburgh, Pa.,
Schenectady, N.Y,, Ewduation of Periodic Notth Preparation in Cocquct Tension Proof Testing and %' arm Prestressing Proce.
Specimens, WCAP 1579, November 1910.
duresfor Nuc/rar Reactor Vessels. HSSTP.
12.
N. Levy and P. V. Marcal, Brown University, IL1, July 1,1969+
Providence R.L Three Dimensional Elastic-2.
L W, Lmchel, Martin Marietta Corporation, Plastic Stress and Strain Analysisfor Frac.
Denver, Colo., The Effect of Testing Vari-ture Afechanics, Phase I: Simple flawed ables on the Transition Temperature in Steel.
Specimens,11SSTP-TR.12. December l970.
MCR-69 189, November 20,1969-13, W. O. Shabbits, Westinghouse Electric Cor-3.
P. N. Randall, TRW Systems Group, potation, PWR Systems Division, Pittsburgh, Redondo Beach, Calif., Gross Strain hica sure Pa., Dynamic Fracture Toughness Properties of fracture Toughness of Steels, IlSSTP.
ofIleavy Section A533 Grade B Class i Steel TR-3, November 1,1969.
Plate, WCAP-7623, December 1970.
4 C. Visser, S. E. Gabrielse, and W, VanBuren, 14, i'. N. Randall,TRW Systems Group..
Westinghouse Electric Corporation. PWR Redondo Beach, Calif., Gross Strain Crack Systems Division.Pittsburgh, Pa., A Two-Tolerance ofA 533-B Steel, IISSTP TR 14, May (,197 L Dimensional Elastic-Plastic Analysis of Frac-
- 15.
- 11. T, Corten and R. It Sailors, University of ture Test Specimens. WCAP-7368, Octoter 1969.
Illinois, Urbana, Ill., Relationship Between 5.
T. R. Mayer and F, 0. Thomas, Westing.
Atate rial Fracture Toughness Using Fracture house Electric Corporation, PWR Systems Afechanics and Transition Temperature Division Pittsburgh, Pa., Evaluation by Tests T&AM Report 346, August I,1971.
Linear Elastic Fracture Alechanics ofRadia-16.
T. R. Mager and V. J. McLaughlin, Westing-tion Damage to Pressure i essel Steels, house Electric Corpor:nion, "WR Systems
% CAP-7328 (Rev.), October 1%9.
Division, Pittsburgh, Pa., The Efect of an 6.
W. O. Shabbits. W, H. Pryle, and E. T.
Environment of High Temperature Primary Wessel, Westinghouse Electric Corporation, Grade Nuclear Reactor %'ater on the Fartgue PWR Systems Division, Pittsburgh, Pa.,
Crack Growth Characteristics ofA333 Grade ficavy-Section Fracture Toughness Proper-B Class 1 Plate and %'eldment Afaterial,
- ties ofA533 Grade B Class 1 Steel Plate and WCAP-7176.Detober 1911,
. Submerged Arc %'eldment. WCAP-7-414, 17, N. Levy and P. V, Marcal, Brown University, December 1%9, Providence,R.L Three-DimensionalElastic.
7.
F, J. Loss, Naval Research Laboratory, Plastic Stress and Strain Analysisfor Frac.
Washington, D.C., Dynamic Tear Test Inves-tare Afechanics Phase 11:ImprovedAfodel-ligatioru of the Fracture Toughness of Thick-ling, HSSTP-TR-17, November 1971.
Section Steel, NRL-7056, May 14,1970' 18.
S. C. Grigory, Southwest Research Institute, 8.
P. B. Crosley and E. J. Ripling, Materials San Antonio, Tet, Tests of 6 in.-Thick.
Research Laboratory, Inc., Glenwood, Ill, Flawed Tensile Specimens, First Technical Crack Arrest Fracture Toughness of A533 Summary Report, Longitudinal Specimens Grade B Class 1 Pressure Vessel Stect, NumbersIthrough 7, HSSTP-TR-l8 lune HSSTP TR 8, March 1970.
1972.
i ix NUREG/CR 5799
19.
P. N. Randall, TRW Systems Group, 23.
P. V. Marcal, P. M. Stuart, and R. S. Dettes, Redonda Be,xh> Calif., Efects ofStrain Brown University, Providence, R.I., Elmtic -
Gradients on the Gross Strain Crack Toler.
Plastic Behavior of a Longitudinal Semi-ance ofA333 B 5 cel, HSSTP TR.19. June Elliptic Crack in a Thick Pressure Vessel.
'\\
= 15,1972.
HSSTP-TR 28, June 1973.
20.
S. C. /3rigory. Southw est Research Institute, 29 W. J. Steliman, R. G. Berggren, and T. N.
San Antonio, Tex., Tests of 6-lach-Thick
' Jones, Oak Ridge Natl. Lab., Oak Ridge, Fittn ed Ter sile Specimens, Second Technical Tenn., ORNL Characteriration of fleavy.
s Summary Report, Transverse Specimens Section Steel Technology Program Plates of, I Numbers 8 through 10. Welded Specimens 02 and 03, USNRC Report NUREGICR-4092 Nwnbers 1I through 13, HSSTP-TR-20, June (ORN1/fM 9491), April 1985.
1972.
30.
Canceled.
-21.
L A. James and J. A. Williams, Hanford 31.
J. A. Williams, Hanford Engineering Devel'.
Engineering Development Laboratory Richland, Wash., Heavy Section Steel' Tech-opment Laboratory, Richland, Wash., The nology Program Technical Report No,21, Irradiation and Temperature Depender.ce of Tensile and Fracture Pmperties o g533, y,,g, g, (;,,, y S,,,, pga, FAST.bf The liffect of Temperature and Neutron Irra -
diatwn Upon Ihe Fatigue Crack Propagation BehavserofASThi A?33 GradeB Classi y,fg,,,,,y g 9g.3 y y 33.y$,g, y,y3.
Steel, HEDL-TME 72-132, September 1972.
32.
J. M. Steichen and'J. A, Williams, Hanford I
"E "##" 8
- I P""'"' b*
- 7'
- 22.. : S. C. Grigory Southwest Research Institute, Richland," Wash.,lligh Stram Rate Tensile San Antonso, Tex.,. rests of 6-inch 4 hick
- Flawed Tenssle Specimens. Third Techmcal Properties ofirradiated ASTbf A533 Grade B Summary Repor t, Longitudinal Specimen s Class 1 Pre:sure VesselSteel, July 1973.
Numbers 14 through 16, Unpas ed Specimen 33.
P. C Riccardella and 1. L. Swediow, West-
' Number 17. HSSTP-TR-22, October 1972.
inghouse Electric Corporation, Pittsburgh.
23;-
S. C Grigory, Southwest Research Institute, Pa., A Combined Analytical.Esperimental
- San Antonio, Tex. Tests of 6-inch Thick Fradwe Study of the Tw Leading TMs Tensile Specimens, F ourth Techmcal Sum-of 57,,,,,,p,,,,,, y,,,,,,, (,,,,,,,m; a,g mary Report, Tests of1-inch-Thick Flawed Equivalent Energy), WCAP-8224, October
- 933'
. Tensile Specimensfor Sise Efect Evaluation, HSSTP-TR-23, June 1973.
34 R. J. Podlasek and R. J, Eiber, Battelle 24.
S. P. Ying and S. C. Grigory, Southwest Columbus Laboratories, Columbus, Ohio.
Research Institute, San Antonio, Tex., Tests Mna Repon n Inmtigation ofAfode #1 Cmc twion in Reactor Pipsng..
of 6-Inch-Thick Tensile Specimens, Ftfth j
TechnicalSwrunaryReport, Acoustic Emis.
D*** b ' 4' b*
sion h!onitoring of One-Inch and Six inch-
- 35.. T. R. Maget, J D, Landes, D. M. Moon, and
, Thick Tensile Specimens, HSSTP TR-24, V. J. McLaughlin, Westinghouse Electric November 1972. -
Corporation, Pittsburgh, Pa., Interim Report i25.
R. W, Dctby, J. G. Markle, G. C. Robinson, on the Efect ofLow Frequencies on the Fa-G D. Whitman, and F.J. Witt Oak Ridge Que m m
amdvisdes ofA533 Natl Lab., Oak Ridge, Tenn., Test of 6-Inch-Gmde B Qass I Hatein an Endmnment of
- Thick Pressure Vessels: Serles 1:Intermedi.
mg empaature mary Gmde Nudear R tor Watu, WCM256, &crh~
= are Test Vessels V-1 and V-2, ORNL-4895, February 1974; 36.
J. A. WEams, Hanfom Enghenng &veb 126.. W.J. Stelzman and R. G. Berggren Oak
- Ridge Natl. Lab., Oak Ridge, Tenn., Radia-epment Laboratory, Richland, Wash.. The tion Strengthening and Embrittlement in Irm atedFmnureToughnessofASM A5, Gmde B, Cass i Stal Afrasured with
. lleavy Section Steel Plates and Welds.
- b*#
"###" # " b
- ORNL 4871* June 1973'-
P men. HEDL-TMF 75-10, January 1975.
F 27.
P. B. Cros!cy and E. J. Rip-hng, Matenals 37.
R. H. Bryan, J. G. Merkle, M. N. Raftenberg*
Research Laboratory,Inc., Glenwood Ill.,
Crack Arrest in an increwing K-Field' G. C. Robinson, and J. E. Smith, Oak Ridge HSSTP-TR-27 Janua.y 1973.
Natl Lab.. Oak Ridge, Tenn., Test of 6-inch.
Thid Presswe Vus& Serid led ate Test Vessels V-3. V-4. and V-6, ORNL-5059, November 1975.
l-l NUREG/CR 5799 x
i 38.
T. R. Mager, S. E. Yanichko, and L. R, Ridge, Tenn., Significance of Reheat Cracks Singer. Westinghouw Electric Corporauon, to the Ir.tegrury of Pressure Vesselsfor Light.
Pittsburgh, Pa. Fracture foughness Charac-Water Reactors, USNRC Relut serization ofilSSTintermediate Pressure ORNUNUREG-15, July 1977.
Ves3el Alaterial. WCAP.8456, Deceniber 41 G. C. S,uth and P. P. Iloiz, Oak Ridge Natl.
I#'
lab.. Oak Ridge, Tcnn, Repair Weld InduccJ 39.
J. G. Merkle, G. D. Whiunan, and R. H.
Residual Stresses in ThickMalled Steel Pres.
Bryan, Oak Ridge Natt. Lab., Oak Ridge, sure Vessels, USNRC Report NUREG/CR-Tenn., An Evaluation of the //SST Pregram 0093 (ORNUNUREG/l M 1531 Jt.nc 1978.
Intermediate Pressure Vessel Tests in Terms 49.
P. P. Itoli and S. W. Wiuner, Oak Ridge of Light-Water-Reactor Pressure Vessel Natl. Lab.. Oak Ridge, Tenn,llalf BeaJ Safety, ORNLJTM 5090, November 1975 tiemper) Repair Weldingfor IISST Vessels, 40.
J. C. Merkle, G. C. Robinson, P. P.11012, J.
USNRC Relon NUREGER-0113 E. Smith, and R,11. Bryan. Oak Ridge Natl.
(ORNUNUREG/TM 177), June 1978.
Lab., Oak Ridge, Tenn., Test of 6.In -Thick 50.
G. C. Smith, P. P Holz, and W. L Stelzman, Pressure Vessels. Series 3: Intermediate less Oak Ridge Natl. Lab., Oak Ridge, Tenn.,
dessel V 7, USNRC Report ORNU
(,,cg g,,,,,,,,,,3 g,,,,, 3.,,,, of x,,,;,,
NUREG.1, August 1976.
pg,,,g 5,,,q ygog,; y,,,,,,, y,,,,y,,
41.
J. A. Daudson, L. J. Ceschini, R. P. Shogan, USNRC Report NUREG!CR-0126 and G. V. Rao, Westinghouse Electnc Corpo-(ORNUNUR EGfrM-196), October 1978.
ration Pittsburgh Pa., The Irradiated Dy-5L R. H. Bryan, P. P. Holi, J. G. Merkle, G. C.
narruc Fracture Toughness of ASTA! A333.
Smith, L E. Smnth, rnd W. L Steltman, Oak Grade 8, Class i Steel Plate and Submerged Ridge Natt. Lsb.. Oak Ridge, Tenn., Test of Pc Weldment. WCAY-8775. 0ctober 1976, TM Pn'm Vm& Suies.*:Imr-42.
R. D. Cheverton, Oak Ridge Natt Lab., Oak med4 ate Test Vessel V-7D USNRC Repon Ridge, Tenn., Pressure Vessel Fracture NUREG/CR 0309 (ORNUNUREG.38),
Studies Pertaining to a PWR LOCA ECC Octobet 1978.
Thermal Shock: Experiments 7'SE-1 and
$2.
R. D. Cheverton, S. K. Iskander, and S. E.
TSE-2. USNRC Repon ORNL/NURFG/
Bolt, Oak Ridge Natt. Lab., Oak Ridge, TM 31, SeMember 1976' Tenn., Applicanility ofLEFAf to the Analysis 43.
J, G. Merkle, G. C. Robinson, P. P. Holt, and of PWR Vessels Under LOCA ECC Thermal J. E. Smith, Oak Ridge Natt. Lab., Oak Shock Conditions. USNRC Report Ridge,Tenn., Test of 6-in.-Thick Pressure NUREG/CR-0107 (ORNUNUREG-40),
Vesseis. Series t: Intermediate Test Vessels October 1978.
V-5 and V 9 with Inside Nonic Corner 53.
R. H. Bryan, D. A. Canonico, P. P, llolz, Cracks. USNRC Report ORNL/NUREG.,,,
S. K. Iskander, J. G. Merkle, L E. Smith, and August 1977.
W. J. Stelzman, Oak Ridge Nati. Lab., Oak 44 J. A. Williams Hanford Enginecting Devel.
Ridge Tenn., Test of 6-in.-Thick Pressure opment Laboratory, kichland, Wash.. The Vessels. Series 3: Intermediate Test Vessel Ductile Fracture Toughness ofIleavy Section V-8, USNRC Repart NUREGICR-0675 SteelPlate, USNRC Report NUREG/CR-(ORNU NUREG-iS), December 1979.
0859 September 1979.
54.
R. D. Chevertnn and S. K. Iskander, Opk 45.
R. H. Bryan, T. M. Cate P. P. Holz, T. A.
Ridge Natl Lab., Oak Ridge, Tenn.. Applica-King, J. O. Merkle, G. C. Robmson, G. C.
tion ofStatic and Dynamic Crack Arrest Smith, L E, Smith, and G. D. Whitman. Oak Theory to TSE~l. USNRC Report Ridge Natl. Lab., Oak Ridge, Tenn.,7est of NUREG/CR 0167(ORNUNUREG 57), June 6-in.-Thick Pressure Vesselt Series 3:
1979.
Intermediate Test Vessel V-7A Under Sus-55.
J. A. Wdliams, Hanford Engineering Devel-tained Loa 4ng, USNRC Report opment Latmrate:y, Richland, Wash., Tensile ORNUNUREG-9, February 1978.
p,0p,,,,,,,,7,,,g,,,,3,,3 y,,,,,,,,;,3 46.
R. D. Cheretton and S. E. Bolt, Oak Ridge Welds of A333 Steel Plaa and A508 Ferg-Natl. Id., Oak Ridge TennJressure Ves-ings, USNRC Report NUREG/CR 1158 selFracture Studies Pertaining to a PWR (ORNUSubf79 5091712), July 1979.
LOCA -ECC Thermal Shock: Experiments TSE-3 and TSE-I and Update of TSE-l and TSE-2 Analysis, USNRC Report ORNU NUREG-22, Decemter 1977.
47.
D. A. Canonico Oak Ridge Natl. Lab., Oak
$6.
K. W. Carlson and J. A. Williams,llanford M.
B. R. Bau. S. N. Aduri J. W. Bryson, and K.
Engineering Development Laboratory, Kathircsan, Oak Ridge Nad. Lab., Oak Richland, Wash.. The Effect of Crack Length Rsdge, Tenn., OR FLA W: A Firiite Element and Stde Grooves on the Ductile Fracture Pregramfor Direct Evaluation of K Factors Toughness Properties of ASTAf A333 Steel, for User-Defined Flaws in Flate. Cylanders.
USNRC Report NUREGICR-l171 and PressureNessel Nozzle Corners, (ORNUSub/79-5(917/3). October 1979.
USNRC Report NUREG/CR 2494 57.
P, P, Holz, Oak Ridge Natt Lab., Oak Ridge, (ORNUCSD/FM-165), Apnl1982.
Tenn., Flaw Preparationsfor //SST Program 6%
B. R. Bass and J. W. Bryson, Oak Ridge Natt.
VesselFracture Afechanics Testing; Afechan.
Lab.. Oak Ridge Tenn., ORA! GEN.3D: A ical-Cyclic Pumping and Electron Beam Finite Element Alesh Generatorfor 3-Dimen-Weld flydrogen Charge Cracking Schemes, sional Crack Geometries, USNRC Report USNRC Report NUREG/CR-1274 NUREG'CR 2997, Vol.1 (ORNL/rM.
(ORNUNUREG/rM 369), May 1980.
8527/VI), December 1982.
58.
S. K.1 stander, Computer Sciences Div.,
66.
B. R. Bass and J. W. Bry son, Oak Ridge Natl Union Carbide Corp. Nuclear Div., Oak Lab., Oak Ridge, Tenn., ORVIRT: A Finite
=
Ridg?.Tenn., Two Finite Liement Techniques Element Programfor Energy Release Rate for Computing Afode i Stress intensity Fac-Calculationsfor 2-Dimensional and tors in 1 wo-or Three-Damensional Problem 3,
- 3. Dimensional Crack Afodels, USNRC USNRC Report NUREG/CR-1499 Report NUREG/CR-2997, Vol. 2 (ORNUNUREG/CSD/IM 14), February (ORN1/FM 8$27/V2), February 1983.
I' 67.
R. D. Cheverton, S. K. I kander, and D. G.
59.
P. B. Crosley and E. J. Ripling. Materials Ball, Oak Ridge Nad. Lab., Oak Ridge, Research Laboratory, Glenwood, llt., Devel.
Tenn., PWR Pressure Vesselintegrity During opment of a Standard Testfor Alcasuring Kla Overcooling Accidents: A Parametric Analy-with a Afodified Compact Specimen. USNRC sis, USNRC heport NUREG/CR 2895 Report NUREG/CR.2294 (ORNUSub/81-(ORNL/FM 7931), February 1983.
7755/1), August 1981.
fe.
D. G. Ball, R. D. Cheverton, J. B. Drake, and 60.
S. N. Atluri, B. R. Ban, J. W. Bryson, and K.
S. K. Iskander, Oak Ridge Natl. Lab., Oak Kathiresan, Computer Sciences Div., Oak Ridge,Tenn., OCA-II, A Codefor Calculat.
Ridge Gascous Diffusion Plant, Oak Ridge, ing Behavior of 2 D and 3 D Surface Flaws Tenn..NOZ-FLAW: A Fini:e Element Pro.
in a Pressure Vessel Subjected to Tempera.
gramfor Direct Evaluation of Stress intensity sure and Pressure Transients, USNRC Report s
Factorsfor Presswe VesselNottle-Corner NUR6G/ CR 3491 (ORNL-5934), February Flaws, USNRC Report NUREG/CR-1843 19M.
(ORNUNUREG/CSD/TM-18), March 1981.
69.
A. Sauter, R. D. Cheverton, arWI S. K.
61.
A. Shukla, W. L. Fourner, and G. R. Irwin, Iskander Oak Ridge Natl. Lab., Oak Ridge, Universuy of Marylar.d Couege Park, Md.,
Tenn., Afo6fictulon of OCA-Ifor Application Study of Energ, Loss and its Atechanisna in to a Reactor Pressure vessel with Ciadding flomalite 100 Deiny Crack Propagau and on the Inner Surface, USNRC Report Arrest. USNRC Repoit NUREG/CR4. a NUREG/CR 1155 (ORNI/TM 8649), May (ORNUSub/79-7778/1), August 1981.
W 62.
S. K. Iskander, R. D. Chevet an, and D. G.
70.
R. D Ocverton and D. G. Ball, Marun Ball. Oak Ridge Natl Lao., Oak Ridge, Manetta Energy Systems,Inc., Oak Ridge Tenn.,0CA 1, A Codefor Cc.lculating the Nad. Lab, Oak Ridge, Tenn,0CA P. A Behmior of Flaws on the Inner Surface of a Deterninistic ard Probabilistic Fracture-Pressure VesselSubjected to Temperature Atechanics Codtfor Application to Pressure and Pressure Transients, USNRC Report Vessels. USNRC Report NUREGICR 3618 NUREG/CR-2113 (ORNL/NUREG-84),
(ORNL 5991), May 1984.
August 1981.
71.
J. G. Merkle, Martin Marietta Energy Sys-63.
R. J. Sanford, R. Chona, W. L. Foumey, nn t tems, Inc., Oak Ridge Natl. Lab., Oak Ridge, G, R. Irwin, University of Maryland, College Tenn.. An Esamination of the Site Effects ad Park, Md., A Photoelastic Study of th Influ-Data Scatter Chserved in Small Speconen ence of Non-Sirgular Stresses in Fracture Cleavage Fracture Toughness Test:ng.
Test Speciment, USNRC Report NUREG/
USNRC Report NUREG/CR.3672 CR-2179 (ORNUSub/7%7778/2), August (ORNUFM-9088), April 1984, 198L NUREG/CR-5799 xii
72.
C. E. Pugh et al., Martin Marietta Energy 79.
R. D. Cheverton, D. G. Ball, S. E. Bolt, S. K.
. Systems, Inc., Oak RUge Natt. Lab., Oak Iskander, and R. K. Nanstad, Martin Marietta Ridge, Tenn., IIcavy-Sertion Steel Technol.
Energy Systems,Inc., Oak Ridge Natl. Lab.,
ogy Progrwre--FiveJear Plan FY 1983-Oak Ridge, Tenn. Pressure Vessel Fracture 1987, USNRC Report NURECXR-3595 Studies Pertaining to the PWR Therm.s!-
(ORNt/TM 9008), April 19M.
Shod Issue: Experiment TSC 7, USNRC Rep rt NUREG/CR4XM (ORNL-6177),
71 D. G Ball, H. R. Bass, J. W. brpon, R. D.
August 1985.
Chevetton, and J. B. Drake, Manin Marietta Energy S stems,Inc., Oak Ridge Natl. Lnb.,
80.
R. IL Bryan, B. R. Bass, S. E. Bolt, J. W.
Oak Ridge Tenn., Sness Intensity Factor Bryson, L G. Merkle, R K. Nanstad, and influence Coeficienttfor Surface flaws in G. C. Robinson, Martin Marietta Energy Pressure Vessels USNRC Report Systems, Itr., Oak Ridge Natl.125., Oak NUREGNR.3723 (ORNLESD/TM 216),
Ridge, Tenn., Test of 6 in.-Thick Pressure February 1985.
Vessels. Series 3 Intewdiate Test Vessel L'.SA-Tearing Behavior of Low Upper-She(f 74.
W. R. Corwin, R. G. Berggren, and R. K.
Af aterial, USNRC Report NUREG/CR-4760 Nanstad, Martin Marietta Energy Systems, (ORN1<6187), May 1987.
inc., Oak Ridge Nati. Lab., Oak Ridge, Tenn., Charpy Toughness and Teraile Prop-85.
R. D. Cheverton and D. G. Ball, Martm erties ofNeutron irradiated Stainless Steel Manetta Encrgy Systems, hrc., Oak Ridge Submerged. Arc Weld Cladding Overlay, Natl. Lab., Oak Ridge,Tenn., A Parametric USNRC Report NUREGER-3927 Study of PWR Pressure VesselIntegrity (ORNi]lW9309), September 19M.
During Overcooling Accidents, Considering Both 2-D and 3-D Flaws USNRC Report 75.
C. W Schwanz, R. Chona. W. L. Fourney, NUREG/CR4325 (ORN1 JIM 9682),
and G. R. Irwin, University of Maryland, August 1985.
College Park, Md. SAAfCR: A Two.Dimen-sional Dynamic Finite Element Codefor the 82.
E C. Rodabaugh, E. C. Rodabaugh Associ-Stress Analysis of Moving Gacks, USNRC ates, Inc., Itilhard, Ohio, Comments on the Report NUREG/CR 3891 (ORN1/Sub/79 Leak Before Break Conceptfor Nuclear 7778/3), Novembee 1984.
Pow er Plant Piping Syecms, USNRC Report NUREG/CR4305 (ORN1/Sub/82-22252/3).
76.
W. R. Corwin G. C. Robinson, R. K.
^ " 8 "'
Nanstad, J. G. Mirkle, R. G. Berggren, G. M.
Goodwin, R. L. Swain, and T. D. Owings, 83.
L W. Bryson, Manin Marietta Energy Sys-Martin Marietta Energy Systems, Inc., Oak tems, Inc., Oak Ridge Natt. Lab., Oak Ridge, Ridge Natl. Lab.. Oak Ridge, Tenn., Effects Tenn.. ORVIRT.PC: A 2.D Finite Element of Stainien Steel Weld Overlay Cladding on Fracture Analysis Programfor a Aticrocom-the Structural Integrity of Flawed Steel Plates puter USNRC R: port NUREG/CR 4367 in Bendmg, Series 1. USNRC Report (ORNL-6208), October 1985.
NUREG/CR4015 (ORNL/TM-9390), April 84.
D. G. Ball and R. D. Cheverton, Martin 1985.
Marietta Energy Systems,Inc., Oak Ridge 77 R. it. Bryan, B. R. Bass, S. E. Bolt, J. W.
Natl. Lab., Oak Ridge,Tenn., Adaptation of Bryson, D. P. Edmonds, R. W. McCulloch, OCA.P. A Probabilistic Fracture 3fechanics L G. Merkle, R. K. Nanstad, G. C. Robinson, Code, to a Personal Computer, USNRC K. R. Thoms, and G. D. Whitman, Manin Report NUREG/CR4468 (ORN1JCSD/TM-Mariena Laugy Systems,Inc., Oak Ridge 233), January 1986.
Nati. Lab, Oak Ridge, Tenn., Prevs,
,i 83 L W. Bryson and B. R. Bass, Manin Marietta Thermal. Shock Test of 6.in.-Thica
- usure En< gy Systems,Inc., Oak Ridge Natl. Lab.,
Vessels. PTSE-l: Investigazion o! Warm O'., Ridge, Tenn., ORMGEN.I'C: A Aficro.
Prestressing and Upper Shelf Arrest-romputer Programfor Automatic ofesh Gen-USNRC Report NUREGER 4106(ORNL' s ation of 2-D Crack Geometries, USNRC 6135), April 19M5.
Report NUREG/CR4475 (ORN1 6250),
78.
R. D. Cheverton, D. G. Ball, S. E. Bolt, S. K.
March 1986.
Iskander, and R. K. Nanstad, Martin Maricua 86.
G. D. Whitman, Manin Marietta Energy Sys-Energy Systems, lac., Oak Ridge Natl. Lab.,
tems, Inc., Oak Ride Natl. Lab., Oak Ridge, Oak Ridge, Term., Pressure Vessel Fracture Tenn., Historical Summary of the Heavy-Studies Pertaimng to the PWR Thermal-Section Stes. Tec knology Program and Some Shock Isne: Experirnents TSE-5. TSE-SA, and TSE-6, USNRC Report NUREG/CR-4249 (ORNL-6163), June 1985.
xiii NUREG/CR-5799 1
1
Related Activities in Light Ware Reactor Wssels. PTSE-2: Investigation ofia Pressure Yessel Safety Research, USNRC Tearing Pesistance and Warm Prestressing, L Report NUREG/CR-4489 (ORNL-6259)J USNRC Relur.NUREGCR.488 (ORNL-March 19Sh.
6377) Ncember 1987.
87, 1 C. Inversini and J. W. Bryson, Manin R
J,11. Giovanola and R.'W. Klopp, SRI later-Marietta Energy Systems,Inc.,0.k Ridge.
national. Menlo Park, Cahf., Viscoplastic
- Natl. Lab., Oak Ridge,Tenn.;ORPLOT PC:
Stress.Stram Characterization of A533B
- A Gral> hic Utilaryfor ORAIGEN.PC and "
Class I Steel, USNRC Report NUREGiCR-ORVIRT.PC; USNRC Report NUREG/CR-5066 (ORN1/Suh/87-S A 193/1), September-4633 (ORNL-6291), June 1986.
I989.
83, J. L M)Gowan, R. K. Nanstad, and K. R.
95.
L. F. Miller et alI, Martin Manetta Energy
= Boms, Martin Marietta Energy Systems, Systems, Inc., Oak Ridge Natt Lab-, Oak Inc., Oak Ridge Natl. Lab., Oak Ridge,.
Ridge,Tenn., Neutron Exposure Parameters Tenn., Characterization ofirradtated Cury for1he Afetallurgical Test Spectmem in the reu-Practice Welds and A333 Grade B Class hlth Heavy Section Steel Technology irra-1 Platefor Nuclear Pressure Vessel Service.
dtation Series Capsules. USNRC Relut USNRC Report NUREG/CR-4880 NUREG/ CR-5019 (ORNL/FM-10582),
(ORNI/TM-10387), July 1988.
March 198f
~'
89; K. V. Cook and R. W-McClung, Manin 96 Canceled.
Marietta Energy Systetas, Inc., Oak RidgI 97.
D.1 Naus,L Keeney Walker,and B R.
Nad lab., Oak Ridge,Tenn., Flaw Dens' Bass, Manin Marietta Energy Systems, Inc.,
- Examinations ofa Clad Boiling Wa:er Rec.
Oak gidge Natl Lab..Dak Ridge, Tenn..
4 tor Pressure Yessel Segment, USNBC Report fligh-Temperature Crack-Arrest Erhavior in NUREG/CR-4860(ORNL/TM 10364), April 152 mm-Tnick SEN Wide Plates of Quenched
- 1987, and Tempered A $33 Grade B Steel, USNRC
- 90.
D. L Naus, B. R. Bass, C. E. Pugh, R. K.
Report NUREG/CR-5330 (ORN1/TM-
, Nanstad L G. Merkle; W. R. Corwiniand 11083), April 1989.
G. C. Robinson, Manin Marietta Energy 98.
K. V. Cook, R. A. Cunningham, Jr., and R.-
. Systems, Inc., Oak Ridge Natt Lab., Oak W. McClung, Manin Marietta Energy Sys-Ridge, Tenn., Crack-Arrest Behavior in SEN tems, Inc., Oak Ridge Natt Lab., Oak Ridge, Wide Plaies of Quenched and Tempered t
Tenn Detection and Characteri:atian ofin-
- A 533 Grade B Steel Tested UnderNon' dications in Segments of Reactor Pressure isothermalConditions.USNRC Repo"
. Pessels, USNRC Report NUREG/CR-5322 NUREG/CR 4930(ORNL-6388), August (ORNUTM-1107_2), August 1989J
-1987.,
91, D. B. Barker, R. Chona, W. L. Fourney, and-99.
R. D. Cheverton, W. T. Pennell, G. C.
Robinson,and R. K. Nanstad, Martin G. R. Irwin. University of Maryland, College Marietta Energy Systems, Inc., Oak Ridge
. Park, Md.;A Report on the Round Robin Natl. Lab.. oak Ridy, Tenn.. fmpact of Program Conducted to Evalaate the Pro-Radiation Embrittlement on Integrity ofPres-posed AS7M Standard Test Methodfor sure vessel Supportsfor Two PWR Plants.
Determining the Plane Strain Crack Arrest NUREG/CR-$320 (ORNLffM-10966),
- Fracture Toughness. K a, of Fernuc Mate-l pwy 19gg, rial.USNRC Repon NUREG/CR-4966
- (ORNI/Subr/9 7778/4), January 1988.-
00-S. N.Naus. L Keency-Walker, B. R. Bass, D.J Iskander, R. L Ficids, R. deWitt, and 92; W.1L Bamford, Westinghouse Electric Cor-S. R. iw III, Manin Marietta Energy poration, Pittsburgh, Pa A Sutrunary of Systems,Inc.,6k Ridge Natl 12b, Oak e
l Environmentally Assisted Crack Growth.
Ridge,Tenn.,5fK Wide-Plate Crack; Arrest
- Studies Performed at Westinghouse Electric Tests Utilizing A G3 Grade B Class l'
^ Corporation Under Fundingfrom the Heavy-Ataterial: IVP-CC Test Serses. USNRC -
Section Steel Technology Program. USNRC Report NURECtN-5408 (ORNLffM-
- ReportNUREG/CR-5020(ORNL/Sub/82" 11269). Novemter 1989.
21598/1), May 1983.
101 D. L Naus, L Kc mey. Waker, B. R. Bass, 93; R. IL Bryan, B. R. Bass, S. E. Bolt, L W.
S. K. Iskandet. :. L Fields, R. deWitt.and i-
. Bryson, W. R. Corwin, L G. Merkle, R. K.
S. R. Low Ill Martin Marietta Energy Nanstad and G. C. Robinson, Martin Systems, Inc.. Oak Ridge Natl.12b., Oak Marietta Energy Systems. Inc., Oak Ridge Ridze, Tennsligh Temperature Crack-Natl, Lab., Oak Ridge, Tenn., Pressurize l' artist Test.: ilsing 152-nan-Thick SEN Wide Thermal-Shock Test of 6-in.-Thick Pressure pgage4j,e,y Upper-ShelfBase Afaterial:
L-NUREG/CR 5799 xiv i
Teus WP-?1 und WP 2.6, USNRC Reprt (ORN1/Sut(19 7778C/5) prer,'utd at NUREGER.5450(ORNL/rM.1135?),
Unhersiiy of Maryland for M.utin Maaricita Fettuary 1990.
Energy Systems,Inc., Oak Ridge Nati, Lab,
b"'*I## I99I' in C
led iH.
pedm Aomht, Unhushy 103.
11 J. Naus 1, Keeney Wa!rer,11. R. Bass, U""'I'"*"'# A*"# #" ##gN@nt mRC Rerun NMpm G. C. Rndinson, S. K. hkander, D. J.
Alexander, R. J. Fields, R. deWitt, S. R, Low l
C. W. Schwaru, and I. llijohanwat. Martin '
@RNWWWM, prepmM at E.n-Marietta Energy Systems Inc., Oak Ridge su) of Maryland for Martin Manetta Energy Natl. lab., Oak Ridge,Tenn., Crack Artest Systems,irg. Oak Ridge Natl. lab, Oak i
behavior in SEN Wide Plates tflow Urper-
"E ""' W I" Y*#'O Shelf Base Metal Trued Under Nonisoti.cr.
Il2.
James W, Dally, WiIham L. Foamey, and mal Conditions: WP-2 3 erin, USNRC George B. trwin, Unhersuy of Marytand, Repon NURE0lCR 545l (ORNL 65 tit),
lower.BoundIs!!iation Toughness with a August 1990.
Mod $rd Charry Sperimen, USNRC Repmt 108.
T L. Di:kson, k. D. Chevenon, and D. K, NGE05M3 (ORMuW.MW),
Shum, Martin Marietta Energy Systems,Inc.,
preparW at Unnnisty oWar%and for Oak Ridge Nail. Lab., Oak Ridge Tenn.,
Martin Manetta Energy Systemn. Inc., Oak inclusion of Unstable Ductile Tearing and Ridte Nall. Lab. 0aL Ridge,Tenn.
U"'*k' ' 99 ' '
ExtralvlatedCrack Arren Toughness Data in PWR Vesselintryisy Assessment, USNRC 113.
S. K. lsk ander,0. C. Robinson, W. R.
Report NUREGER 5473 (ORN1/f M.
Corwin, D. C. Oland, D. J. Alexander, and a
11450), May 1990.
K. V. Cook, Martin Manetta Energy Systerns, 105.
T. J, Theiss, Martin Marictn Energy Systems.
Inc.. Oak Ridge Natl. lab., Oak Ridge, Inc., Oak Ridge Natl. lab., Oak Ridge, Tenn., OperuntaMnulu yTon to lian.
Tenn.. Recommendationsfor the Shallow.
tigate Flaw Behavior of Mechanically Crack Fracture Toughness Testing Task LoadWlainins SteMlad Pima, USNRC Within the /ISSTProgram. USNRC Report Iteport NUREGER 578'l(to te publishet NUREG/CR 5554 (ORNL/TM.ll509),
114.
S. T. Rolfe, University of Kansas,"Interpre.
September 1990.
tive Report on the Application of Shallow.
106.
J. G. Merkle, Martin Marietta Energy Sys-Maw POD Test Dma to the Strwtural tems,Inc., Oak Ridge Natl. Lab., Oak Ridge, Margm Assessment of Ret tor prenure Tenn.. An Overview of the Low Upper Shelf Venets wnh I' laws,,, USNRC Report NUREGXR S767 ORN1JSubMSI161011) t Toughness Siety Margin issue, USNRC Report NUREGER $$52(ORNL/rM.
3 D'S I99h r'
11314), August 6,1990.
I15.
D. R.
I'abe, Martin Marietta Ener.y Syn.
107.
D. K. M Shum, J. G. Merkle,J. Keeney, Ips. K 04 Ridge Natl. Lab.."Compari-Walker, and B. R. Bass, Martin Marietta suf Wriull and hic Analysis of Transition Energy Systems,Inc., Oak Ridge Nati. Lab.,
Range Fracture Twghness Data," USNRC Dak Ridge, Tenn., AnalyticalStudies cf
. Tort NUREGXR-5788 ( V rL/TM-Transverse Strain Efecis on fracture 11959)Junuary 1992.
Toughnessfor Circurr(crentially Oriented 116.
nis teport.
Cracks, USNRC Report NUREGRR 5592 (ORN1/TM 11581), April 1991.
l
\\te.
J. D. Landes, Extrapolation of the J.R Curve b:
for Predicting Reactor VesselIntegrity, C
USNRC Report NUREGER 5650 (ORNL/Sub/89.99732/1) January 1992.
109.
J. Keency Walker, H. R. Bass, and J. D, landes, Martin Marietta Energy Systems, l'
inc., Oak Ridge Natl.1.ab., Oak Ridge, 1
Tenn., An Invntigation c(Crack-Tip Stresp Field Criteriafor Predicting Cleavage Crack initiation, USNRC Report NUREGCR 5651 i
(ORN1/fM.I1692) Septembei 1991, 110.
O. R. Irwin, University of Maryland,"Use of hickness Reduction to Estimatc Values of K," USNRC Report NUREG/CR 5697 r
- Y NURi GER 5799
1 liitroductioni in early August 19A), the Nuclear Regulatory Com-R. K. NanstaA ORNL: es aluation of (1) f racture-l mission (NRC) requested that Oak Ridge Nabonal toughnen curs es for unitradiated rnatenal. (2) cal.
Lateratory (ORNL) coordmate anJ parucipate in a re.
culation of RTNDT. and (3) surs ollance program.
view of a seport entakd, Reactor Preswe Vmel fva!uarion Arperrfor ) Wee Nuclear Power l'lant, D. L Selb), ORNL: consultation with regard to delb Y AFC No.1735, July 1990,2 w hich was prepared by niuc if pistulated ITS transients ard estimation die Yankee Atomi; Elatric Company (YALC). 'lhe of their frequenues.
review was to cos er primarily the pressurited therms!-
shock (PTS) analpas deunbed in Sect. 6 and the D. A. Ito/anh, S AIC:d evaluation of the methodology upp:r4helf-energy analpis descrited in Sect 3. The ior esumaung rne.an s alues of the calculated fre-request aho indicated that Idaho National Engmeenng quency of Iallure.
Latoratory (INEL) would provide thermal /hydrauhc input, and P3c fic Nonhwest Laboratory (PNL) w ould J. W. Minarick, S AIC: esaluation of completenen of provide 6put with reg rd to the VIS A code,3 w hich hst of postulated FTS transients and review of was us.st by Y AEC for the probabilistic fracture-esumated frequencies.
rnechnics analysis.
K A. Williams, S AIC: (1) evahation of the thermal /
The NRC request also r.gcified a c ompicUon date of hydruuhe mixing analyses for transients involving September 17,1990, a deliverable m the fann of a drafI stagnauon in one or more loops and (2) evaluauon letter report on that date, and a planning mecung on of the appropriateness of RE*RAN5 for the pres-August 9,1990. This rnecting was held as scheduled, sunred. thermal 4 hock (PTS) (nmsient 'hermal/
but because of prior commitments and delays in estab-hydraulic analyses.
lishing a sub;ontract und in obtaining neccuary infor-madon from YAEC, the completion date for the draft F. A. Simonen, PNL: (1) companson of the YAEC was eventually changed to November $,1990. A final versbn of VISA Il and PNL's version,(2) evalua-repon w.s to le issued on un unspecified date, and this tion of input to the fracture mechanrs analyv s, publication constitutes t! e final report.
(3) panicipauon in the comparison of OCA P and VIS A II, and (4) evaluation of vessel insprtion The team members contnbuting directly to the ORNL p,ogram, coordmated effon are indicated below along wPh their inter ded areas of respon%bility.
L W. Ward, INEL: (1) cvaluation of edequacy of moochng used for RETRAN analyses,(2) com-R. D. Che s erton, ORNL: (1) coordmation of the panson of RETRAN and version of RELAP 56 efforts of the 10 resiewers;(2) collection and used for NRC/ORNL IPTS studies of II. B, transmittal to the NRC (Pat Sears) of all quesdons Robinson p' ant,7 and (3) consultadon with regard for the utihty; (3) contnbutions to the probabilis-to definition of postulated transients.
6: fracture mechanics analysis tes iew, and (4) preparation of a leuer repan and transmutal of INEL and specific members of S AIC and ORNL were same to the NRC by November 5,1990, that con-mcludej because of their earlier involvement in the de-t;nns the contnbutions of each of the 10 reviewers.
velopment of the imegrated preuunted thermal 4hotk (IPl S) methodology.'
T. L. Dickson, ORNL: (1) check input to the frrture.
mechames analpes; (2) ins estigate the salidity of To a large extent.k adequacy ard accuracy of the d
the probabilhtic frxture rucchnics codes OCA i YAEC evaluation were judged on the basis of tk and VIS A II;3 (3) evaluate appropriateness of Haw methodology devekped as a pan of the NRC/ORNL IITS study,7 he NRC PTS Rule (K1CFR50.61),8 and densuy. Haw-size distribution function, flaw aspect t
ratios, vessel region division, and strespintensity.
the NRC Regulatory Gmde 1.154,9 which identifies f actor influence coc!Ticients used in the Yankee acceptable IPTS-evaluation methidologies and a target Rowe VIS A ll calculation, and (4) cakulate the masimum pennissible value ("pnmary acceptance cri-condiaanal probabihty and frequency of fadure for terion") of the calculated fn quency of vessel f ailure the Yankee Rowe sessel using OCA.P.
(through wall-cracking).
J. G. Merkle. ORN L: (1) evaluation of the methodol-Primary sources of infomiation pertaining to the re-ogy used for incluJw fnu.ture toughness upper view wcre the YAEC report (Y AEC No.1735);
shelf and (2) consultation on ra:hatinndimage and Regularary Guide IS9, Rev. 2 (radiauondimage corre-fracture mechania iwues, lations); Ret Guide 1.134; KYTR50.61; radiation-damage evaluadons performed by G. R. Odette
%m Ae mm iraemel conw.uon i
N11 REG /CR 5799 l
(University of Cahfomia) and A. L. Iber, Jr. (NRC),
appendttes),a summary and discussion of the findings, w' o were net specified as memters of the alove team; and indications of information that is klieved neces-the Yankee Rowe e nergency geradng pnwedures sary for a more thorough review. Some of this infor-(EPOS); and Refs. 3,4, and 7.
madon was requested earlier but has not yet ken received This report consututes a compilation of the contribu-tions made t y each of the team members (included as 2 Scope of Review the scope of tir revicw inclucc5 a review of *all" arulpis, or equivalent, to obtain a "mean* value of the aspects of the PTS evaluation, upper shelf energy con-frequency for comparison with the value conesponding siderations, low tem;ierature over pressurizadon to the *p imary acceptance criterion" in Reg. Guide (LTOP), and vessel inspection. The PTS evaluation 1.1R Each of these iterns was consideied in the includes: (1) postuladon of FTS transients and estima-resicw.
Uon of their frequencies of occurrence;(2) thermal /
hydraulic analyses to otnain the downcomer coolant
%c xqc of the reuew also included an independent temperature, primary system pressure and vesselinnu.
calculadon by ORNL of the frequency of vessel failure, surfxe fluid-film heat transfer coef ficient, each as a For this analysis, best estimate inputs were used to ob-function of time in tie tramient;(3) rudiation induced tain a kst estimate of the conditional probability and increaw in the reference nil ductility transition temlera-frequency of vessel failure. %e inputs were kst esti-ture (RTNDT) for the vessel plate and weld material mates in the sense that in ORNL's opinion they rep-(this requires knowledge of the vessel fast rcutron flu.
resented the rnost likely values bawl on all data ences, operating temperatures and c hemistry (Cu and available to ORN1.at the time of the independent Ni)1;(4) a probabilistic fracture mechanics analysis to analysis. This approach is consistent with that used in determine the conditional probability of vessel iallure the NRC/ORNL ll'TS studies,7 w hich provided an for each transient beheved to be a significant contribu-NRC accepted probabilistic methodology for evaluating tar to the frequency of failure;(5) a summauon of the PWR pressure sesselintegrity. As additional plant-frequencies of failure for exh tramient to obtain the specific data are obtained,it is likely that the test-overall frequercy of failure; and (6) an uncertainty esumates will change also.
3 PTS Transients and Their Frequencies (Appendices A and 11)
Questions of particular concern with regard to this sub-De YAEC report identifies a small break LOCA ject matter are (1) have the actual dominant transients (SBLOCA 7) as the dominant transient and assigns a been postulated, and (2) are the esumated frequenues of frequency of occurrerv;c to this transient of-5 x occurrence of the transients that are suspected of teing 104/yr. Asindicated in Appendix A,the reviewers dominant realistic or at least conservadve?
suggest a more realistic value of 1 - 2 x 10-3, which is considered to be a mean value. If other LOCAs are The consensus of the reviewers is that insufficient in-included to account for their contribudon in a conseru-formation was available to make an accurate judgement tive n'anner, assuming that SBLOCA 7 :rpresents the with regard to the selection of transients. Even so,if most severe of the ! OCAs, the effective frequency is consideration of a single transient or category of tran-increased to ~4 x 10 3 sients indicates an excessively high frequerry of fail-ure, than consideration of other trandents may hot be nstessary. He reviewers followed this line of thinking in addition to making numerous comments, sugges-tions, and estimates regarding definition of transients and their frequencies (Appendices A and D).
l NUREG/CR 5799 2
- ~
~
- - -. ~ _. ~,
. _ - _., ~... -. _. _.. - - -
1 4 PTS Transictit Tlierinal/llyciraulics (Appentlices A unct II)
'ihe questan of paub' ular umeern regarding the Y Al C te mm tvere than im e uc Widi reprd to tem-c derrnal' hydraulic analy sis is whethet thit PTS tranuet.
ivrmure et presswr, the severity of the trarment u desented by the calculated primary-system pressure, more 1.kely to le peate Cun Icf.s. 'lhe bes;-trans!cr diu rcomer uolunt temter^ture, and vessel inner-utflu tent, un the other hand. is more hLely ta te less, s.urf a';c heat-transfer coef ficient is likely to te mote arid dus wonld tend to redare the severity; howeset, severe than indicated. Tic teviewers teliese that the ti.e.c4 on sen iouty studms in Ref 7,it is believed transient desenbed in the Yankee re;vrt is a test esti-diat the trJ A tion attributed to the best-uansf er acifi.
mate, but the octual trantierit is rnuch more likely to t ient u eulJ n::t tvc upnf karit.
s o
i Nt'RI:G A'IM7W
i 5 Radiation Effects (Appendix C) 5.1 Inercase in RTNDT
- PN'" "' te"""""" '" * *"dth'*'***rhe*-
istry, were used in the atove scheme to calculate tie irwrease in RTNDT caused by radiation damage here are two values of RTNDT of particular interest (ARTNDT) in the ORNL probabilistic fracture-with regard to 10 CIT,50.61 and Reg. GiJde 1.134.
mechanics knalysis of Yankee Rowe (Appendix D).
For 10CFR$0.61, a + 2a (two r,tandu ; deviations)
(Since the time diat these calculatie,s were performed, value is needed for compuison with de FTS screening an uplated set of flucreces locame available but were criteria. For Reg. Guide 1.134, a mean value and a dis.
not irgluded herein. The most recent values are tribution are needed for use in a probabilistic fracture.
slightly less th.sn dose used h this study.)
mechanics analysis.
ORNL and YAEC estimates for 10CFR50.61 +2a 5.2 Decrease in tlic Upper Shelf values, minus the 20 are given in Table C.2 of gn p pN, Appendis C for 1990. Assuming 20 a 60*F, it is apparent that all values exceed the screening criteria, which are 2707 for axial flaws and 300*F for circum-Here are two specihmem with regard to upper-ferential flaws. As required by Reg. Guide 1.99, shelf energy. One is whether the vessel satisfies de Rev. 2, the copper concentrauon in the welds was lowmpperyshelf analysis for 1 vels A,11, and C kud-4 assumed to le 0.35 wi% tecause measurements are not ing c ndiuons in accordance with criteria recommended available. Based on the 11R 3 weld clemicalsomposi-by the ASME Section X1 Working Group on riaw tion data, the concentration of nickel was assumed to be Evaluauon. He other penains to de selecuon of 0'7 wt%~,
upper she f fmeture toughness values for the probabih,s.
t tie fracture 'nechanics analysis.
- Dest estimate
- values of RT DT for the upper axial N
Time did not permit a review of the calculated stress.
weld were obtamed using Reg. Guide i 99, Rev. 2, intensity f actor (K ) values corresponding u kud levels I
with an addition of 44 F in the ORNL analyses to account fa a lower irradiation tem trature. (Reg.
A, B, and C; however, the J R curves used for compar.
Guide 1.99, Rev. 2,is based on an irradiation tempera-ison with the K1 values were reviewed. As mdicated m ture of 550*F, while the irradiation temperature for the Appendix C, ORNL believes there is adequate margin Yankee vessel is -506*F. De lower temperature re-f r exh of the kuding levels, assuming, of course, diat the K values are correct.
1 sults in a greater damage rate, everything e.lse icing equal.) As indicated in Appendix C, an irradiation-temperature correction factor of IT/l'F is believed to An appropriate upper shelf frrture toughness value for be an appropriate best estimate fcr the materials, flu.
use in the probabilistic fracture mechanics analysis was ences, and temperaturcs of mierest.
estimated by ORNL to be -140 ksik for the upper axial weld ( Appendix C). De YAEC report used a Appendices C and D indicate that in de absence of f---
specific data for the Yankee welds the test estimate of value of 200 ksiyin.8, which was also used for the the Cu concentration in the wekts is 0.29 wt%, and 0".NL IPTS studies.7 ORNL sensitivity studies that 1e = 0.07 wt%. Based on the BR 38 data, the associated with this review indicate that the effect on best estimate of the Ni concentration is 0.7 wt% ;
the conditional probability of vessel failure (P(FTE)] of Ap;cndix D also indicates that the test-estimate fast-the difTerence between 140 and 200 ksid is neutron fluence for the inner surface of the upret axial weld is 1.24 x 1019n/cm2. His fluence, with an
- ** N#8 "
'The hR 3 md Yankee Rowe reactor preuun venels were fabricated by the same manufacturcr at the same lunc and with similar rnalerials.
f
6 PTS Fracture Mechanics (Appendices D and E) 6.1 Comparison of Ihe PNL and revic* to* ccas and compare one ag:unst the 'ther.
YAEC Versions of VISA II
$$ t,7j#d "' ' '"'*'
'**I'""""
"h*
(Appendix E)
Reviews o toth codes were perfonned by Dow chart-r Four categories of dilhretres can be considered:
ing, down to a fine level of detail, the probabilis-tic methcAlology, details, input, and crrors. Formal docu.
methodologies, and by making comparison calcula-mentation of the Yariec Rowe version of the code was tions for wall temperatures, stresses, and K values, 1
not available for teview, and this pevented a compre.
The temperature and stiess comparisons, w hic h in-hensive comparison of basic methodology. A phone volved comparison with independent, validated, corn-conveisanon, however, rescaled no differences in basic mercialcodes, indicated that the VISA ll and OCA P methodology, although there were three differences m subroutines are valid, detail; the YAEC version it cluded (1) residual stresses in the welds,(2) a enore accurate representation of Reg.
GWde 1.99, Rev... and (3) a somewhat different set OCA P and VIS A Il toth use influence co;f ficient and of K1 values corresponding to pressure loading.
Supers siti n techniques for calculatug K1 values, but the details are differrnt. he OCAA procedure is mo,e With regard to K calculations,it is not clear whether accurate, but me differences sorm Jiy are not signifb the K innuence coefficients in the YAEC version cor-tant and are believed not to te s2nificant for the respond to R/w* = 7 (appropriate for Yankee) or R/w =
10 (built into the PR version of VIS A ll). A corn-parison in Appendix E indicates that this difference he deuuled companson of the two codes, by means of now charting, revealed three errors in VIS A II, two of would not affect imtiation and arrest of shallow flaws dich were almost trivial but one of which results in (alw < 0.$), w hich are the ones of primary concern.
an esce ssive number of stable arrests and thus in an underrstirnate of P(FiE). These enors have been cor.
he input parameters for the i.ankee cal:ulations were reviewed item by item for consistency with Reg. GWde rected in the PR version but presumably were not cor-1.154 and PR's recommendations for application of rected in the YAEC version. One comparison calcula-tion (Appendix D) indicates that correcting the errces VISA II.3 While several details of the Yankee inputt inescases P(FIE) by a factor of -10, but for the specific differed from those used in prior NRC studies, sensi-Yankee analysis, the difference is telieved to le less.
tivity calculations indicate that these dif ferences should not have a major impact on calculated failure probabil-After the alove corrections were made to VISA II, ities. Of course, inputs for prnsures, temperatures, OCA P and VISA Il were compared by using toth to i
and radiation induced embrittlement do have very sig.
calculate the " Rancho Seco" transient (Fig. D,1, nificant impacts. These inputs are discussed in Appendix D) with R/w = 10 "All" input was the Appendix D-same, aad only one region of the vessel, containing a single flaw, was considered. De temperature and stress he ORNL review of the PR version of VIS A !! re-distributiorts agreed very well, and the K1 values agreed vcaled three ermn, and it is assumed that these errors also txisted in the i AEC version of VIS A II. Dese reaenably well, particularly for alwd < 0.5. The number of initial initiations agreed within 12%, and errors are discussed further in Sect. 6.2.
the values of P(FIE) were within 3(fr. The number of arrests for OCA P were three tinics greater than for VIS A-II,lecause of the difference in K! values for 6.2 ComparISOD Of OCA P and a/w > 0.5. but were a factor of ~10 less than the the PNL Version of VISA il number of initiations, in which case the difference in arrests has very little cffect on P(FIE). Dus, the tentative conclusion is that OCA P and the corrected Prior to this review, the VISA !! and OCA P codes had PR version of VIS A-Il agree well, and both appear not been redewed in detail since 1984. Because toth s alid with respect to w hat they were intended to do. It codes are teing used by utilities and others for evaluat.
is irnportant to remember, however, that there are ing vessel integrity, and especially because VIS A-II is choices to be made in important input /modeling being used in connection with the Yankte hfe-exten, parameters that can result in significantly different sion studies,it was prudent at this time to carefully va'aes of P(FiE). Flaw density,its uncertainty,and
- Ratio of sesiel rdm to u au thd nus.
8Crad deph4au thidrus 5
NUREG/CR 57W
surface length of flaw at arrest are three of the more im-Again with reference to Fig. D 9 (Appendia D), these portant choices.
values should te increasud by a factor of -1.7 to in-clude tre residual stress. To convert these values from
' test estimates" to mean values, they must le rnulti-6.3 0RNL 0CA.P Analysis of Plied by 45, the ratio of mean to tesi-estimate fiaw i,ankce Rowe (Appendix D) density given in fief. 7. He best estimate flaw density is t flaw /m3 (Ref. 7), and a ftaw density or 45 flaws /mi correspands to ~1 flaw in the Yankee ves-sel upper weld. [lf there were more than I flaw in the
'lhe ORNL OCA P analysis of Yankee Rowe used the region calculated.OCA P might overestimate P(F1E) neutron fluences that conespond to 1990 and tie because of doub!c counting.]
region definitions and volumes given in Ref.10.
P(FIE) was calculated for the upper asial w cid only, and if the same flaw density is assumed for all regions of the Cu and Ni concentrations for this region (Cu "
the vessel, and if an approsimate correction is made for 0.29,10 = 0.07; N = 0.7) were best-estimate values double counting, the contribution to P(FIE) of regions taken from Ref.11. The number of flaws correspond-other than the upper axial weld will at least double the ing to a mean value of P(FIE) was essentially the same value otuined for tie latter region, as that used in the YAEC analysis (one flaw in the region). A uniform tensile stress of 6 ksi was included Applying a!! of the above factors yicids to simulate a residial suess, and the transient calculahd P(F1E) (mean, w/o repressurintion)>5.0 x was the SBLOCA 7 transient described in Ref. 2-10 4 x 1.7 x 43 x"t = 8 x 10 2 With reference to Fig. D.9 (Appendix D),
P(FIE)(mean, w/repressuriation) >l.2 x 10-3 x P(FIE) (base case)* = 5.0 x 10-4, and 1.7 x 45 x 2 = 2 x 101 P(FIE)(w/repressurir.ation)6 = 1.2 x 10 3
\\
k 3
Tppet 315al, we)J Wily, na ndJual stas 1 flawho,
w npeuuraauan 3
'lteper stial weld only, no residual stnis. I flasho.
p* sawizadan to 1550 pai at 20 enin UREGNR-5799 6
7 ORNL Estimation of Frequency of Failure for Yankee Rowe ne fxqwney L f failure of Ahc vecciis calculated as tourwis all othe-r similar LOCAs. As indicated in follows:
Aprendix A, loth values are considered to te reason-able mean values.
q(F) =
Ig (I)1 P (B)lP (FIE).
Using the lower of the two, g
ij g
9(F)(SDLOCA 7),a 2 x 10-3 x 8 x 10 2, 2 x 104/yr.
wlm _
9(F) = totcl frequency of fatture (failures / reactor yr),
With represwbation as descriled above, j
9(1) = iJtlator frequency for ith ransient, (F)(SBIDCA 7R) > 2 x 10-3 x 2 x 10 3 =
t 3
Pj(II) = branch pmbability for jth ranch,ith 4 x 10 /yr.
i b
4 tramient, and j
%cra values are substantially greater than the value of 5 x 104 failusedyr :cIerred to in Reg. Guide 1.134 as
= gi(i)flP j((B) = frequency of the FIS pi(E) i transient event).
the "pnmary acceptance enterion" for the ITS mean ficquexy of vessel faihire.
For the SBLOCA 7 transient,9(E) = 9(I) = 2 x 10-3/yr,or 4 x 10-3 If the transient conservatively j
8 Conclusions Values of RTNDT calculated for Yankee Rowe in ne IWO mean frequency of failure calculated by accordance with the rules in 10CFR50.61 for compari.
ORNL. for the Yankee Rowe vessel is >2 x 104/yr and son with the FTS Rule screening criteria are substan.
thus exceeds the value corresponding to the " primary tially greater than the screening criteria values.
acceptance criterion" in Regula-lory Guide 1.154 ($ x 10 6/ reactor yr). As stated in the Reg. Guide, however,
%c PNL version of VISA Il and the OCA P protu.
this does not necessarily mean that the vessel is unsafe bilistic fracture mechanics codes,which are referenced to operate.
in Reg, Guide !.154, are in good agreement and are valid for their intended purpose. (During this teview, There are many unanswered questions regarding details an error was found in VIS A Il that has subsequently of the YAEC IPTS typc7 evaluation of vesselinteg-teen corrected. De above statement p: stains to the rity. It seerns 'mhkely, however, that answers will corrected venion.)
substantially aher the atove es imated values of 9(F).
f I
7 NUREGER 5799 l
l l
a
9 References
- i. L. C. Shao, NRC, to W. D. Adams, ORO, "FY
- 7. D. L. Selby et al., Pressurized 7 hermal Shock 1940 Noder.r Regulatory Researth Order No.
Evaluation of the ll.B. Robinson Nuclear Tower (490476 for Oak Ridge Natianal latomtoiy, Plant, NUREG/CR-4183 (ORN1/TM 9567),
August 2,1990.8 htartin hiarietta Energy Systerns, Ire., Oak Ridge National Lab., Oak Ridge, TN
'a Yarlec Atornic Electne Co., Reactor Pressure (September 1985).6 Yessel En.'uatic n reportfor l'adre Nuclear Power Plant, Y AEC No.1735 (July 1990).a
- 8. Code of federalRegulations Title lo, Part 50, Section 50.61 and Appendix G.6
- 3. F. A, Simonen et al, VISA.Il - A Computer Codefor Predicting the Probability cf Reactor 9.
U.S. Nuclear Regulatory Commission, Pressure Ves:el Tailure, NUREGXR-M86 Regulatory Guide 1.154, " Format and Content (Phb577.4), Pacific Northwest Lab., Richland, of Plant Specific twssurized Thermal Stock WA (March 1986).6 Safety Analysis Re[ orts for Presurited Wa'.cr Reactors" 6 E'
4.
R. D. Cheveston and D. G. llall, OCA P, A Dettardsistic and Pmbabilistic Fracture
- 10. Written communication from Yankee Atomic Mechanics Codefor Application to Pressure E1cctric Company to htr. William Russell of Vessels, NUREG/CR 3618 (ORhb5991),
United States Nuclear Regulatory Commission Union Cartdde Corp., Nuclear Division, Oak (September 28,1990).*
Ridge National Lab., Oak Ridge, TN (May 1984tb 11.
K. E. hkore and A. S. lleller, Bd W 177-FA Reactor Vessel Beltline Weld Chemistry Study,
- 5. J. A. McFadden et al., RETRAN-02-A -
IIAW 1799,ll&W Owner's Group Materials Programfor Tramient Thermal.llydraulic Committee (July 1983).a Analysis of Complex Fluid Flow Systems, Volume 1: Theory andNumerics (Revisson 4),
Ef'RI NP-1850 CCMA-A, November 1988.c
- 6. C. D. Fletcher et al., RELAPS Thermal.
Hydraulic Analyses ofPressurized Thermal Shock in Sequencesfor the ll. B. Robinson Unit 2 Prenuri:ed Water Reactor, NUREGXR.
3977, Septemter 1984.6 1
6 8Avs aNe from NRC I'bhc Doewnent Rann for a fee.
6Available for purchase from GPO Sales Pngram 8 AvailaNe fnxn NRC Public Ibcurnent Room for a fe.
cAvailable for purchase frra crgantuuon sponsoring pubbcaum bAvailable for pearchase frurn GPO Sales Program.
cited, and/or from authors and/or recipients (doeurnented letters).
NUREGER 5799 8
F.
i Appendix A i
SAIC RevicW of YAEC No.1735 i
Contents Page d
A.1 Review of Statistical Issues in Pressurized Thermal Shock Evaluadon Report YAEC 1735 (D.P.Bozarth).........................................................................................................
A.3 2
Summary A.3 Comments..................................................................................................................
A.3 References A.6 A.2 Dermal. Hydraulic Behavior in Small. Break LOCAs of Significance to Presrerized Dermal Shock (PTS) with Consideration of the Ymkee Nuclear Power Station - SAIC Report No. 916501 (Final)(K. A. Williams)
A.7 Exeeutive Summary A.7 Preface........................................................................................................................
A.7 1 Intmdue tion and Background.........................................................................................
A.8 2 S cope an d Objec ti v e s..................................................................................................
A.8 3 Oeneric Dermal. Hydraulic Behavior in SBl.OCA and Controlling Phenomena......................
A.9 4 Parametdc Evaluadons of Downcomer Fluid and Vessel Walt Dermal Transient Behavior.......
A.11 4.1 Methodology A.11 4.2 Parametric Evaluntions A.11 5 Yankee Nuclear Power Station Audit Calculation.............................................................
A.13 6 Comparbons Between Yankee H. B. Robinson, and Calvert Cliffs.....................................
A.15 7 Summary Ind C,,nc1usions A.16 -
8 Recomme nda tions................................................................................................A.17 9 References A.18 Attachment A to SAIC Report No. 91 6501 ( Appendix A.2)...............................................
A.31 Attachment B to SAIC Report No.91 -6501 ( Appendi x A.2)......................................................
A.34 A3 Review of Accident Sequence Identification and Quantification in the Yankee Rowe Pressurized -
Dermal S hock Analy sb (J. W. Minarick)..................................................................
A.39 G ene ral Com me n ts.........................................................................................................
A 39 Inidating Event and Accident Sequenced Selection A.39 Human Reliabilit y Considitadons..............................................................................
A.40 Frequency and Branch Probability Estimation..........................................................,.
A.41 i
Credit for Operators Actions and Alternate Procedural Actions............
A.42 l4 l
A.I NUREG/CR.5799
Appendix A.1 Review of Statistical Issues in Pressurized Therinal Shock Evaluation Report YAEC 1735 D. P. Bontrth SAIC 70S S. lilinois Ave.
Oak Ridge, TN 37830 hmarY' From the fact that the lower "two sigma" limit curve shown is a straight line parallel to the fined line, it is clear that standard regression techniques ne following comments are those concems raised were n I used to estimate the lower confidence by statistical issues within the sulject report. It is recognized that not all of the comments contained O'".0 for a prediction from the regression line.
Das is (lear since the confidence interval for a lin.
herem are of critical importance to the overall conclu.
sions reached in the submittal. Ilowever,it is this re-canegrmonis gedrancJ viewer's position that statistical treatment of enginect-ing data should always le in accordance w ith accepted ne data fmm this figure w ere estimated and a re-gression line and lower 95 % confidence level were suuisucal medods.
estimsted. The regression line is quite strnitar to Overall, the methodology used in the report is an (Mt pmvi&d in the figure, indicating that the data applicable approach to the evaluation of PTS risk, were read from the figure uccurately. The lawer However, several major limitations prevent the report confidence mterval obtamed fro 6n the regression at from providing the necessary basis to make a final de-35 and 57 ft lbs are ~90 and 200 in. Ib/in 2, respec-terraination on the validity of the conclusions drawn tively. Dese values are 40% and 20% lower than the values ud therein. Principal among these is sirnply a lack of sufficient detail throughout from which an independent While n. appears that this analysis is not of major peer review can be accomplished. Examples are the lack of supporting data for correlations; justification significance to the overall l'TS evaluation, the use for assumptions, such as distributions for fracture rne.
of pmper statistical methodology through :Ut an chanics input parameters and the lack of sufficient de-analysts is important.
scription of the naw density distribution chosen.
nese shoncomings are more specifically enumerated
- 2. Pp. 5 21, Fig. 5 9 and Appendtn B. Anhenius Relation in the individual comments that follow.
Data for the correlation that are stated to be in A further major lacking with respect to the NRL' Reg, Gidde 1.154 is aay discussion whatsoever nf an effect Ap;wh b.ae tot provided. Appendix B refer.
ences Sect. $A2 instead of 5.3.2. It ts not clear of uncertainties in data and engineenng calculations on the final results, not any data on the sensitivity of the what the purgme of de conclauon 1s. If it is to vessel failure probabilities to these parameters. Rese show equivalences of slopes, then the resolu,ng re-two areas art specifically enumerated as being required gression skyes and ertimated variances in slopes to be addressed in the Reg. Guide. Further discussion should te reported.
of this point is also made in an individual comment.
- 3. Pp. 6-57. Small LOCA Initiating Event Frequency Comments Insulficient discussion of the nayestan update pro.
cedare used to estmiate the SblOCA event fre-
- 1. Pp. 3-7 and Fig. 3 Jic Correlation with Cv querry is presented to allow for adequate review.
While the reduction in frequency owing to the up.
Data for tuth transverse (T L) and longitudinal date is not scally numerically significant (20%), the (L-T) welds are plotted and fit with a imear regreS-actml prior distribution used armi the uptate tech.
sion line v. Cv in Fig. 3 5. However, the data nique sindo be sufficiendy descrited.
points are not identified as to which are T-L and w hich are L T. Consequently,it is not possible to his comment addresses only the lack of infonn-judge whether a single regression 1;ae is appropnate tion for the methcdoloey used for this estimate.
for both sets of data.
A.3 NUREGfR-5799 I
1
~
Other comments will address the apphcability of single flaw must te at least two orders if magni-the data base used to esdmate the actual initiating tude less than that of the asial welds foi their con.
event frequency.
tributions to the total vessel failure probability to tv equivalent.
- 4. pp. 6-Sh. Initiating Event Frequencies The 11.11. Robinson analysis utilised the initM Calculation of mean frequencies from the c.ssumed flaw-density distnbutice also quoted as the basis for tegnormal distributions is correctly (kine although the YNpS subminal, The interpretation of that the mearVmedian ratio for the error factor of 30 is reference is significantly different tetween the two higher than necessary. His revicwer calculated that analyses, however. The ll.11 Robinson interpreta, correction.o le 8.48 rather than 9.06. Ilowevet.
tim was that the value of one flaw per cubic rneter the use of the geometric and/or arithmet c mean is the most probable value h>r the llaw density and presented in many places is not clear. It appears that the actual flaw desaitnould te much larger (or that the mean from the assumed distribution should smaller) than this. For this reason, a right.
to used. Clarification of this guint is needed, truncated lograrmal distribution was uscd therein to descrite the initial flaw density. The mean (average) flaw densny under that model was and K a Curves
-.4fvin-3. The particular form of the distribution
- 5. Pp. 6-208. Kle l
chosen was not intended to be the only possible in-Discussion of the applicability of t'ese data to the terpretation. Ilowever,it was intended all available Yankee vessel and the uncertainties inherent in uti-information applicable to a particular vessel be lizing the "mean" values is required.
caref ully considered in specifying a jusdfiable flaw demity, in view of the essenually bncar nature of
- 5. Pp. 6 208. Flaw Density Distnbution the vessel failure probability on the initial flaw density, the discrepancy in the stated assumptions The initial number of flaws la a region of interest and a justification for the limiting flaw density dis-directly influences the probability of vessel failure tribution used should le provided. In gurticular, by introducing initiation sites for crack propaga-since from Tables 6.7 5 and 6.7 8 it appears that tion. The initial flaw density in the YNPS analy*
only the lower plate was considered in the vessel sis is stated to be 1 flaw /m3, and it is also stated failure probability, demonstration that the other that the number of flaws in the total irradiated weld areas are not signi0 cant is necessary.
and plate materialis five. His obviously implies a total irradiated material volume of 5 rn ne
- 7. Pp. 6-209. Normal Distributbns analysis then assumes that one flaw exists in ed of the (coincidentally) five vessel regions. De
%e truncation of Quence values in tie fracture second assumption implies that the volume of each mechanics simulations at the one sigma values region is one cubic meter. Irradiated volumes for seems unjustified. Other truncations fe' material each of the five regions are not provided in the re properties are at least three standard deviations such port but they have been obtained through MMES
- that tie effect of the truncation is not significant, c
- he volumes for the regions and the effective flaw density based on the assignmerit m one 11aw per
- 8. Pp. 6 209. Results of Analysis region are as follows:
The net result of the analysis is presented as being Volume Effective Flaw representative of a "mean value" estunate, This es.
Ession unh Densitv n/rtgiat0 timate may be more accurately classined as a mean conditional on the p:uticular values of the ther.
Upper Plate 3.51 0.28 mal. hydraulic toundary conditions and particular Lower Plate 130-0.77 input distributions used in the fracture mechanics Circum. Weld 0.085 11.8 calculations Theestimation of these parametersis Upper Axial Weld 0.018 55.6 consequendy ofimportance. In the 11. B.
Lower Axial Weld - ' 0.0068 147.0 Robinson analysis and reflected in Reg. Guide 1.154 it was recognized that uricertaimy is inherent he widely differing volumes cause a marked bias in the estimates of the parameters owing to limita-in the relative importance of the various regions to tions in available data and calculational techniques the oserall probability of vessel failure. An owrall as well as the effects of other necessary engineering
- unbiased estimate of total vessel failure probability approximations (such as binning of thenna! hy-is not possible since conditional probabilities for draulie transients, for example). Tie technique tec-all regions are not provided. However,it is clear ommended therein is an uncertainty analysis of the that the probability for failure conditional on a effect of the significant parameters on the estima-tion of the overall frequency of failure. His was accomplisted for the 11. B. Robinson analysis by Cl-enon.
Rike Nanonal latxrcury.real use of a hkote Carlo simulation, and the technique NUREGNR 5799 AA v-wen-n
,wn
.,,n-m.-~-r-m,.-m-ewv.r e -. r e, n n e r,,,e a
-a-,.---m
--~,
,-n.rv,,nw---~
a --,- -, -
g
is recommended in Reg. Guide 1.1.4 for several Owing to this effect and the incluOn of the pre-rearons mo3t sirificant of w hich are the comple4-viously mentioned mean flaw density of 45 m-3 ity of the analysis and the extreme nonhrcarity of Tic inclusion of uncertainnes for estimates of the the fracture mechanics inodel results ta varianons means of tae sigmficant fracture mechanics van-in input values.
ables. ircutron fluence, and thermal hydraulic tumding condiuons contritmted a factor of roughly Estnaation of system perfont.ance for nonlmear five out of that total for the distributions used proble.ns c known to te biased by exclusion of the therein. A smaller factor is due to the combinatior nonlinear terns.2 In small systems, systems in of etent tm < quence frequencies and tranch proba-w hich the non.incarity is not estrerne, or, il the bilitier, but the major effect is due to the nonlinear-uncertairies in parameters are small. the effcet of ity of the fracture mechanics results. Jusblication the nonlinearitia rnay be neghgible. For a PTS fe; not incl.iding uncertainties in these sigt.ificant analysis, none at the above condidons are met in paramete.rs for this a vlysis is necessary or an esti-the 11. B. Rob.nson analysis, the net i ffect was to mation of this effect on the presented results is raise the estimate of the mean by a factor of ~250 requirrd D
A.5 NUREG/CR 57W
References i
- 1. N. R. Draper and H. Smith Applied Rrgression
- 2. G.11. llahn and S. S. Shapiro, Statistical Models Analysis, John Wiley & Sons, Inc., New York, in Engineering, John Wiley & Sons,Inc., New 1966, p. 22 24
- York,1%7, p. 228ff a d
Available inxn puhtic and special technxal Ltwaries.
Availatale frwn putAic snd special technical htwanet i
' NUREGER-5799 A.6
Appendis A.2 Thermal flydraulic lleliavior in Small llreak LOCAs of Significance to Pressurized Thermal Simek (PTS) with Consideration of the hnkee Rowe Nuclear Power Station SAIC Report No. 916501 (Final) i MCC11iiVP Summary Analyses conducted with the presentindependent e
metledo ogy cod a review of the REMIX ca!c Ocneral umsideration has ken given to smalbbreak manual indicates that REMIX has been applied by LOCA thermal hydraulics of significance to PTS, with YAEC as the cole developers intended. The current emphasis on the potential to proceed to stagnation of analysis is in excellent agretment with the YAEC primary kop natural circulation flow, injection of cold REMIX cakulation. It was concluded that this
(-120T) safety makeup water into a
- stagnant" down, transtent should Ic comidered as a best estimate
- comer region will produce a rapid, perhaps ses cre, cwb resuh owing to the cxtensive REMIX assessmert ing of the pressure venel wall. An independent ana.
basis. Ilowever,it was aho concluded that any sig-lytical procedure w as des cloped to quantify the transient nificantly less severe cooldown is unhkely for this thennal response of the downcomer rnised rnean fluid scenano with primary k)op Dow stagnation. Thus, temperature and wull heat transfer in a manner similar the REMIX rnned-mean downcomer Guid tempera-to the REMIX code. Calculations speedic to the tun is an upper bound but represents the best-Yankee Nuclear Power Station plant were performed to estimate fur expected behavior. While die report evaluate the YAEC submittal using the REMIX code.
YAEC-1735 did nc,t provide complete thermal.
The followiny conchnions and recommeadations can h>draube de tails, these has e subsequently been pro-r te made based upm the present independent l'f $
vided by letter to the USNRC and considered in the therraabhydraulic c valuations.
present evaluations in this final report (but not in previous draft versons).
Small break (1 1/2 to 3 in.) LOCA transients should te expectf d to procecd to pnmary coolant
+ Yankee plant sp cine dessign features could le Dow stagnation. YAEC has correctly considered inigertant to the Huid mixing process, especi4Jiy to such a limiting scenario for the Yankee plant. A the Guid behavior adjacent to the pressure vessel I
flow stagrmtion scenano should u considered for all _
wall. The appropriateness of REMIX assumptions PTS evaluations, includmg the 11. B. Robinson and foi the YNPS geometry may need to be further con-Uderd Calvert Cliffs plants;it uppears that this may not have ken adequate!y adiesmi for these plants.
Preface This repirt was previously distributed vice in draft wndit ons on safety injecuan flow, and REMIX details i
form, including Rev.1. These two sersions con $ ered such as dow neomer mned.mean fluid temperature.
id only the limited inforrnation contained in the report YAEC-1735 on the SitLOCA thermabhydraulics. In This new infonnation necessitated re-analysis by S AIC June and July 1991, Yankee Atomic provided complete as well as minor modi 0 cations of the conclusions and detaih of their REMIX cvaluatian (inclu od as Attach-recoramendanons. This FINAL version is being pro-rnents A and D of this reparo. The new YAFCinfor-uded af ter con ideration of this complete information, mation pertained to primary niohmt system geornetry, initial Nnt corahtiom at flow stagnation,loundary
?
A,7 NUREG/CR-5799 w.~,..w--m
-. ~
..--r-
-w.
m m m----n--.
-x--
,w
..w.-.-
e,
-v.
---.---w-%v s-m
.~w-n-
1 Introduction and llackground Extensive safety assessmem rescarth. loth expaimen-fact that for a range of small (~ 1 u 3 in.) breaks, pn-tal and ar.alytical, was conducted during the past dcrade mary coolant kiop flow stagnation can occur t.t signifo l
on the pressurized thermal s,hmk (ITS) issue. This cant prcssure (~ 800 psi) accompmied by an extended work resulted in rule making,10CFR50.61, " Fracture period of r.afety inja tion (SI) of cold (- 120"F) water i
toughness requirements for protection agailut pressur-tato the downcomer region. %c YAEC predicted ized thermal shock." The U.S. Nuclear Regulatory downcomer pressure and fluid temierature are shown by Commission (NRC) has received a submittal under this Figs. A.1 and A.2, respectively. He downcomer rule for the Yankcc Nuclear Power Station (YNPS),
cooldown rate (0.4 to 1,3'F/s) for Yankee is signifi-
" Pressure Vessel Evaluation Report," Yan! ec Atomic cantly greatcr Otan that considered
- prototypical" from Electric Company (YAEC) Report No.1735 July the H. IL Robinson and Calvert Chifs baseline l'fS 1990.1 neir evaluation considered the individur.1 studies.2.3 nis large cooldown rate,in concert with "ITS risk" from a spectrum of hypothetical accident YNPS materials and neutron fluence, has rrbed con.
initiators and concluded that the dominant ever t is a cerns over a large through the-wall-crack probability small-break loss of coolant accident (SBLOCA) -
for this SDLOCA scenario. %ere is additional con.
(-1 1/2 in. diameter break). %e SBLOCA sequence cern over the specifie YNPS temperature history of tcing risk significant is in agreement with conclusions Fig. A.2, arising due to the calculated dichotomy in for the NRC's assessment of similar
- baseline plants "
downcomer cooling rates before and after 200 s-reduc.
H. B. Robinson and Calvert Cliffs, which previously tion in couldown by a factor of three. The trancient re-underwent signiGeant evaluation. SBLOCAs tend to sults of these figures required a switch from the system be risk dominant because of the potential for severe simulation (PliTR AN) to the loop downcomer ernpiri-(mpid) temperature cooldow n of vessel materials while cal model(REMIX) at 150 s.
at signincant pressure. His situation occurs >iue to the 2 Scope und Objectives in order to independendy evaluate tie calculated behav.
ITS signincance and then la pravide " audit calcula-ior in the YNPS, as well as to qualitatively consider tions" for Yankee as well as comparisons with H. B.
differences between H.B. Robinson and Calvert Cliffs Robinson and Calvert Cliffs. There are three specific behavior from Yankee, the present work has been per-objectives addressed in the following acctions. First, to formed; initial consideration was pmvided in
- Review provide a narrative of qualitative thermal hydraulic tran-of YAEC 1735 Reactor Pressure VesselEvaluation sient behavior leading to flow stagnation and to identify Report."4 The scope of the present work is to provide -
plant-speciGc gurameters potinent to cooldown tchav.
a qualitative description of SBLOCA dr f ior, Second, to quantify the cooldown potential of hydraalics behavier including controlling phenomena Yankee relative to H. B. Robinscm and Calvert Cliffs, and then to provide quantitative comparisons on including a fornmlation of bounding downcomer cool-cooldown potentialin YNPS relative to the earlier down behavior. %ird, to evaluate the Yankee behavior baseline plant studies. This work draws from insights during the stagnant loop flow regime when fluid-fluid gained from previous evaluations 2.3.5-9 as well as per-mixing dominates the thermal response. It is the intent forrns new, independent calculations for YNPS to help of the author to provide a review useful to those with explain plant-specific behavior and to clarify expected limited thermal hydraulics backgrotmd to help them deviation from those earUer studies.
comprehend generic plant response and to provide plant-specific perspectives to aid in evaluation of this first--
The overall objectives are to provide a narrative describ-
" PTS plant submittal" to the NRC.
ing generic thermal-hydraulic behavior in SBLOCAs of 5
l l.
l l
-NUREG/CR 5799 A.8 u
..u
~
t 3 Gerieric ThermabHydratille llehavior in SilLOCA and Controlling Plieriomerla Extr.rtsive reactor systems analyses with modere 11 the break size is large, flow stagnation will occur but thermal hydraulic computer codes, TRAC 3 and te accompanied by depressurization to low pressure.
RELAP5,8 have identified smellheak LOCAs as im-Dus, there is a spectrum of break sites with a lower portant scenarios with reyect to pressurited thermal and upper lirait that may te expected to envelope shock.D These thermal-bydraube analpes revealed SBLOCAs of special PTS significance. A simple pro-that there may be special concern for SDLOCAs which cedure has teen developed by this author and professur Theofanous7 o determine the minimum break size that result in stagnation of the primary coolant flow while t
at significant pressure. Such a scenario could result in can produce primary loop flow stagnation on a plant.
severe overcooling and pressurized transients owing to specific basis. This
- mapping" is possible after realir.
sustained periods of cold (12019 safety injection water ing that intenuption of natural circulation occurs due )
into downcomer w ater that has been holated from core a tweak in the primary circuit's liquid continuity, This and steam generator (reverse) heat sources. The fol-will occur if the primary system sustains a loss of 10 wing narrative presents a dncription for ' generic" liquid arising from the break flow exceeding the water 1
SBLOCA transients with particular attention on con-inflow from the safety injection pumps: Since both of trolling phenomena. This is intended to provide the these boundary flows depend on primary pressum, addi-reader with a qualitative " picture" of transient thermal-tional consideration must te given to the transient hydraulk behavior in such a risk dominant PTS thermal hydrau!!c tchavior for the small-b-cak LOCA, scenario.
An overall description is now presented for tle system -
Pressurized water reactors are designed to ensure that transient thermal. hydraulic response foi a S tiLOCA core heat removal capability is maintained in the event with a break size that leads to stagnation while at sig-that pumpmg capacity is lost, that is, to ensure natural nificant pressure. The scenana of particular sigaifi-cbculation of the primary coolant. Driving forces that cance for PTS is an accident initiation whtle the plant sustah the natural circulation are differential pressure is at so called
- hot zero power
- with a break size of
- heads" arising from the cooling of water in steam gen-typically 1 1/2 to 3 in. diameter, %e pnmary system erators located atove hotter water in the core. The den-pressure will typically fall rapidly but then stabilize for sity differenc4s and clevation changes can drive a signif-an extended penod, the behavior for Yankte shown in let.ht flow of primary coolant water, in consideration Fig. A.1 is a typical response, %l3 shows the ex-of I'TS scenarios, this natural ciretdation has a tw>
pected SBLOCA " pressure signature"; an initial pres-fold beneficial cf tect on mitigating the overcooling sure of over 2000 psia with a decrease to 700 m 900 transient First, the circulating water maimains the psia over 3 to 6 min. He important featu;c is that the t esoci v.ull with heat from the core as well as " reverse" pressure " holds" at a significant value for an extaded heat transfer from the steam generator secondaries. The period. His is a consequence of prinuey wahr flash-second effect is to promote mixing of the cold safety ing at its saturation temperature while being augmented
- injection water with the entire primary system water by reverse heat transfer fro:n the rteam generator sec.
mass. Thus, natural circulation can greatly mitigate ondary side to the primary v1ter. This kat transfer overcooling, If natural circulation is interrupted, the maintains the stagnant fluid in the steam generator stagnate configuration loses these twc beneficial effccts primaries at the saturation temperattue and therchire and eignificantly more severe overcooling will result.
maintains the pressure via toiling. He ensuing steam will then form a " void" region at the top of the system An interruption of natural circulation will occur if there (U tubes) and interrupt ratural circulation. His pies-is a " break" in coolant fluid stream continuity,i c., a sure plateau value can casily be computtd on a pbmt-void region forms and interrupts the siphon t.ffect. It is specific basis;it is simply the saturation pressure cor-possible for a ' void" (steam) to form after the blow-responding to the liquid temp:rature of the steam gen-down from a smalbbreak LOCA; this told r.ormally crator secondary (shell side). For PWRs this is typi-accumulates in the highest region of the system, for cally in the range of 800 to 1000 pda whib at hot zero example, the U tubes of the steam generators. De power. Knowing this plateau prr.tsure will then allow primary coolant circulation will remain stagnated, one to compute a plantapecific minimum break size in thereby setting the stage for an overcooling transient a SBLOCA that will rause flow st4 nation. The tucak unless this steam void is collapsed by condensation or outflow (QBreak) can be approximated by system repressurization.
/
V" For a SDLOCA scenario to be of extrerne PTS signifi-3 cance theni there must be both a flow stagnation and a G'"" ~ h '
A "de (A D l
significant primary system pressure. If the break size TCgj is small, the system will remain pressurned, but steam voiding will not occur and neither will flow stagnation.
A.9 NUREG/CR-5799
_,--,,,%.,w.#,-.-,,p,.
...mw,,,.
,,%+,.
,.p.-,%~,,,3
,m,p,.,_,.,,,,,.--w,,%_mee,.ni,w-n.,,,,,,,m.-w%
-._.,-.m_,_um,,._,.,..,_,,,_o,,.--
.,,..n..,,,, _ _
w here hig is the latent heat of essporntion. T is the within the downcc,mer segion with gcrhaps strong water temperature, Cpl is the specific heat, and A h the coupthig to warmer wates in the hiops and lower minimura break area. ne primary synem liquid vol.
plenum. Tie first phenomenon is the " flow splitting" umetric inflow can be computed from the plant speciDe of injected liPI water; some iraction of the cold injec-bigh pressure-injecuen (Hpl) head-Dow curvet esalu.
tion water may flow away from the sessel mitigating ated at this pressure. His will a!!ow for a calculation the downcorner cooldown. De second phenomeron is of the minimum becak area in Eq. (A.f) that results in an entrained, backflow of relatively wannet water from a primary system net liquid volume loss, and ult].
the downcomer region into the top of the cold leg.
mntely flow $tagnation.
His effectivelf wanns the inflowing water to the downcomer, also mitigating the cooldown. The third, Downcomer Duid temperature respc:ue is directly con.
ard by far the most important pheromenon, is fluid-trolled by the primary systems' coolant flow behavior.
fluid mixing between the dow neomer flows and the Prior to primary depressurir.ation and subsequent tut.
lower plenum water.
ing, natural circulation with heat sources tends to miti.
gate cooldown from injection of the cold lip! water.
However, almost immediately after flow stagnatim In summary, this section has provided a generie description of thermal hydmulic behavior for a limiting occurs, the downcomer fluid, temperature begins a rapid SBLOCA. Indeed,it has shown that for a specific decrease. nc YNPS behavmr of hg. A.2 is,qualita.
range of tweak sizes, flow stagnation can occur at pres.
tively representative of this effect, that is, mmamal cooling carly on, and then rapid cooling af ter stagna.
sure producing a severe overcoolmg. He transient p
wim a npid blowdown to a pressure plateau tion. Indeed, this Dgure shows very rapid ( 1.5 F/s) coch,ng immediately after stagnation and then reduced controlled by reverse heat transfer from the steam gen.
crator Mries' liquid. We minimum break area i
cooling due to wall heat transfer and warm fluid mix.
mg. As will be discussed in the fcIlowing sections' that produces this tchavior on a plant specific basis can I
the YNPS cooldown is greater than that of the casily be estimated bawd on a talance tetwcen break baseline Calvert Cliffs and it B. Robinson behavm, r.
ou'. flow and safety injection inflow. During this early period, there is significant ratural circulation loop flow nelong term SBIDCA thermal-hydraulic tchavior is m ganng downcomer cmMown. Akt Dow stagna; tion occurs, the cooldown can tecome severe due to controlled by hot and cold fluid-fluid mixing with the e
I 889. h the entire primary water mass. He stagnant f both heat sources and the bulk convecdve mix-absence of bulk loop circulation. Figure A.3 (taken ing wit from Ref. 8) conceptually tilustrates the flow behavior in the downcomer and cold leg regions. There are three d wntomer co ld wn rate is controlled by three phe-n mena f Duid-fluid mixing. Ilowever,the cooldown key phenomena of iruponance to the thermal hydraulic during this long term stagnant regime can te lounded cooldown behavior during this stagnation period. The downcomer cooldown is essentially controlled by the Ngh a simple energy balance, as down m me fob lowing secu,on.
inflow of a cold stream from the loops and mixing 4
s NUREG/CR-5799 A.10
4 Paranictric Evaluations of Downcomer Fluid and Vessel Wall Thermal Transient llehavior l
Esaluat ons of N pressure vesw! wall fracture me-wliety M h the mass of Guid in the snixing region and
.]
chanics are eksel coupird to the transient thermal-A hydraulic behavi in the downcomer. In particular,11 v w all is the waji heat transfer. If wall heat transf er is necessary to determine the pessure vessel wall tem-is igtered, that h adiabatic, this eqution can te inte-perature respom.c to the boundary conditions of the st1 gluted to give face heat transfer crefficient and fluid ternperatu c.
T
-la'i Daring the early puicd of %:urti circulation within the T.
4 primary coolant system, RELAP5, Rl!TRAN, or
- (" - - - - M = c (A3)
- Tup,
. TRAC
- systems codes" are traditionally employed.
flowever, once loop stagnation ecem s, these codes are inappropriate due to the inability to conectly repesent where Tis h the initial Duid temperatwe. Ilowever,
" stagnant" mixinF of cold and warm water regions; the u rcalistically evaluate reactor behavior for l'TS now behavior is dvainated by complex turbulent mix.
os, h wall heat transfer must be evaluated and ing driven by buoyancy rather than momeritum effects.
Eg. (A.1) integrated numerically, using to appropriate his flow behanor has teco the subject of extensive tmtial aid boundary conditions.
experimental t.no aralytical studies. As a result of these sudies, the REMIX computer code hia tren I
developed to evaluate downcomer response to safety 4.1.2 Wall IIcat Transfer injection of cold water into a 'stegnant" systeta.9 Yankee Atomic employed REMIX to quanttfy the YNPS behavior, as shown by Fig. A.2. Ire order to Quant;fication of the temporal response of Qwall povide an independent calculational audit of hse requires calculation of the vessel widl heat diff usion results, the present analyses have teen coeducted, The tchavim as well as the smface heat transfer coefficient, following sections ounine tic methods,p:aametric For the present work, the w>dl heat transfer has been studies, and Yankee evaluatioa with the pesent rnodel.
evaluated by solving a one<limensional,linite-differ-ence model subjected to a uniform initial temperature,
"" "'h"b ti' b "ad"'Y c*dti a ""he **'ca" $"'f8"-
4.1 Methodalogy and a known (tansient)intemal heat flus boundary
'0nd on at the intemal downcomer fluid face (i.e.,
The present study focused on parametric cvaluations of hA the mixed mean fluid temperature in the vessci down.
comer region. De rest 4ts are then expccted to be com-The above model of coupled wall and fluid transient i
parable with the correrponding REMIX value, Tm-thermal response has teen numerically implemented The present work did not attempt to predict REMIX-into a small PC computer program. This program simulated safety injection backflow and detailed mixing computes a mixed-mean downcomer Guid temperature with various Guid regirnes, but rather trrated these phe-subjecd to input k ndary and initial conditions. The nomena parametncally to obtain an enveloping tran-cale has been used to parametrically evahmte the mflu-sient irsponse. Fwthennore, the current model does ence on Guid and w all temperature transient of control-not address the plummg cffcct (treated by REMIX) that I ng parameters, including calculates colder temperatures in the plume below tie cold leg penetnitions; that is, mixing region 4 of fluid mass participating in the " stagnant" mixing Fig. A 3, problem, such as the cold legs, inlet annulus, I
""
- I"**'P""*8 "*
d ""'""? ion flow rate into the downcomer; and 4.1.1 Fluid Thermal Energy Balanee safety m)cct The " mixing cup temperature (T IX) with a fluid M
region (control volume) due to instantaneous mixing of 4a Paramtric Eva1tiatlons an ircoming (colder) fluid stream of flow rate m lipt and temp:rature T m is given by This section presents the results of parameter analyses i
using the above traasient thermal model with condo tions similar to hnse of the Yankee plant. Table A.1 r' m,(T"" - T* ) + b 65'8 " I " ** *8 * ' # "
- P^'"* "'i' '.*5.
M n
(A.2)
=
ton. Spainc YNPS resulu are given in Sect dt C
P3 A.11 NUREG/CR 5709 r
.m._,-._----..---
1 4.2.1 Ileat Transfer Coefficient
'entrainment" of waaner watert Figure A.6 illustrates the parametric effect of redxmg inflow rate to the Figure A.4 shows tia calculated natural cmve: tion heat
" fluid mixirg volume.' 11 can be concluded that the uansfes errKeicat as a futetion of temperature differener cooling is signincantly reduced only if at least one hrlf (AT)letween the wall and the fluid. As shown, the of the llP1 water dows away from the vessel.
value also depends on the water temperature (transpon propcrties). For the practical range of plant condition 4,2.3 MIXitig Water Volutne dunng the cooldown, the heat transfer coelucient varies between about 100 and 500 Blu/h.ft2/F, liigher coe(fi.
~11e present analysis for the mixed-mean fluid tempera-cients will tend to mitigate the downcomer water ture (as v.t'l as REMIX) assumes ifnt a single (large) i coeldown; however, this represents the most severe volume p rticipates in the hot cold fluid mixing pro-thermal shock to the pressure vessel wall and thus is cess. Figure A.7 ilhistrates the parametric effect of 6
- conservative" froin a safety perspective Figure A.$
varying the mixing volume assuming baseliac parame-illustrates the parametric effect of wall heat transfer coef-ters fer other variaHec, For the Yankee plant, the in.
ficient on temperature of the mixed mean downcomer cluded volumes represent the following regions of the water. This calculatbn used the tmeline values" of primary system:
Table A.1,11 can te concluded that for wall heat transfer coefficicnts greater than 4M) Blu/h fg20F. a
- conduction '
200 ft3+ Inlet anmGts below top of cold legs and -
16nited" process is governing. That is, the wrjl surface dowrr,mer region
+
is in thermal equilitrium with the fluid and heat transfet 333 ft3-Alm. tegions plus cold legs letween is limited by heat diffusion from the vessel wall material injection poir,t and vessel itself, A value of 400 Blu/h ft eF was selected for the 800 ft3-Atuve regions alus low.-r plenum.
2 NRC's 11._13. Robinson PTS evaluation 2 and is used as the present basehne in subsequent calculations.
Figure A.7 demonstrates that the mixed mean fluid temperature is strongly dependent on the assumed fluid 4.2.2 Safety Injecilon Flow' Rate regions participating in the mixing process. For the Yankee plant, it quantifies the substantially mitigating 3
Flow distribution of safety injection thigh pressure in-effect ofincluding the lower plenum volume (467 ft )
jection (liPI)) water in the cold legs is qualitatively in the mixing process.
illustrated by Fig. A.3. Flew tchavior is controlled by buoyancy effects, that is, by Froude numter similarity enteria. it.ls likely that some fr'sction of the IIPI water will
- backflow
- away from the vessel, tending to miti-gate the cooldown, at least durmg the initial rapid cool.
t ing regime. REMIX has an empirical model that de-termines the backflow fraction (and corresponding r
l l
l L
NUREG/CR 5799 A.12 mi y- +v a ws m-r,me -em >,e e-m,v,,m n-;me.e.
g m
w>
r--r-s a
e nxa e-w w
,w o se -rm.se e
~
+&m-em e*we~*e%~
,'tv-n-----
e m
er
--,t-e
-m-n,.-oe
-vme
5 Yankee Nuclear Power Station Audit Calculation 1
ne analytical rnodel discussed in the pevious section c) ne transient w as initiated at 150 s reactor tune has teen used to formulate a Yard ce Nuclear Pow er with the fluid at 476T.
Station plant 4pedfic evaluation. The objectim was to evaluate the reasonableness of the YAEC REMIX re-f) Safety injection flow was at 120T (tmehne) an:t sults using the independent calculational tool of the pre.
parametrically evaluated at 1707.
vious chapter, Specincally, this effort was to evaluate i
the YAEC dow rcomer Duld mixed mean te nperature g) Comparisons were raade with YAEC REMIX Tm (l.c., that corresponding with Tm f REMIX). As noted 5alues; the mixed mean Ould tempenture was i
o in the palace, detailed infonnation on the REMIX taken from the YAEC output listing.
model, input, and calculational results were provided by
- YAEC in a lettet report June 26,1991 (Attachment A)
S AIC's new results are compared with the YAEC value and through a telefax on July 5.1991 (Attachment B).
in Fig. AA Execlient agreement exists letween the
-. His information and teleconferences with YAEC and YAEC REMIX toults and the $AIC simplined model.
NRC staff have creatly clarified theit assumptions and ne deviation at 1200 s is less than 20T and likely results for the SitLOCA scenario. Essed upon this in-occurs due to Y AEC's correct inclusion of heat flow formation, the following is now known:
frca the core region,
^
a) Primary coolant flow stagnadon was calculated ne effect of preheating the safety injection water to (by RETRAN) to occur at 150 s with the system 1707 ts quantified in Fig. A.9 by comparison to the downcomer water at 4767. The REMIX cakula-YAEC REMIX at 1207.
i tion began at this time.
Consideration of the YAEC results and the present b) De downcomes fluid temperature presented in independent analysis leads to the following conclu.
YAEC 1735 is the REMIX value Tjump (a plume sions. The early time period (0-150 s) rooldown is temperatum) not the warmer mixed mean tempera.
realistic and consistent with expected fluid tchavior ture, which has now also teen provided (Curve 2 during the uansition to primary coolant system bulk of the YAEC 6/2ti/91 Leuer, Attachment A).
fisw ctagnation. 'fhe dramatic decrease in cooldown at 200 s, shown in the.eport YAEC 1715, is due to inac-c) REMIX values for Yankee geometry, materials, curate plotting of REMIX results. The long-term initial conditions, and detailed output are now cooldown under stagnant er didons (after 150 s) has available (see Attachment B).
been correcdy simulated by Y AEC with tM REMIX code. REMIX has been shown to te in very good SAIC's computer rode that calculates the mixed-mean agzeement with a wide range of experimental data in an downcomer fluid temperature was used with considera-extensive assessment project.10 The YAEC results tion of this new REMIX information. The following (Fig. A.8) must therefore be considered as best estimate
+
changes were made from the previous calculations:
results for die mixed mean downcomes fluid tempera.
ture. Fluid temperatums Iclow the cold legs (plume a) The system metal mass was expanded to also in.
region) are lower than these valuest the YAEC REMIX clude the lower plenum region and the (double.
values for this regi(m are those given in report YAEC-sided) thermal shield.
1735 (also Fig. A.2 of this report). There is no reason i
to expect that any less severe downcomer fluid tempera-b) Metal thermal conductivities and diffusivities were ture trmtsient could occur in this SULOCA scenarm imed upon the REMIX YAEC values, with de given initial and boundary conditions (e.g.,
tripped main coolant pump.s). De only known omis-c) Mixing volumes were compared to YAEC values sion of a heat souce from the REMIX calculations is and found to be in nearly exact agmement with that from hot water in the barrel baffle region. Thus,it is used in the preslous analysis; however, the total corgluded that YAEC (R EMIX) calculated downcomer S1 flow is now injected into the total REMIX fluid fluid temperatures are both 3 test-esthnate and likely an volume.
upper boand value for anticipated thermal, hydraulic be-havios. This result teing simultaneously an upter 4 The YAEC REMIX model includes the flow of lound and a best-estimate needs clarification. It is heat from the eue region; this is not included in besbestimate owing to the validity of REMIX per the the SAIC model thereby poducing a slighdy extenuve asseument basir.10 It is an upper bound greater cooldown.
in that there are only minimal dermal or Guid phe-nomenon that would kad to any less severe cooldown for this transient. Indeed, this is the essence of the i
A.13 NUREG/CR.5799 l-1 f
n--
.., n
.%.=-.-,
n
.c, w,
+
w.
,-,.n.n,
-~
' REM 1X phenomenological assumptions and it is well plant it is stwt 3 in. This much narmwer gap may.
validated thmegh rt>mpansons to extensive eagtimen-reduce lowei plenum water m!xing and influence der.
tal dat.t mal-hydraulic beheior in the downcomer. Anothcr concern is the influence of the Yankec geornetric details Finally,it is ncted that certain Yankee plant.s;aific in the downcomer inlet annulus region. As shown by design features may serve to influence die fluid fluid Fig. A.10, the up;er core support terrel has an outer miting process relative to that expecial of larger diameter significantly s.maller than the dowrcorr.cr plants, e.g., Cahett Cliffs. These unique fea. ores are region. This Yankee feature could also af fcct the down.
revealed by the ptcasure vessel restical crou wction of flow of colder water in the downcomer and the vessel Fig. A.10. For the Calvert Cliffs plant, the down.
wall cooldown, comer gap is roughly 10 in., w hile for the Yankee i
b I
L a
?
i 1
l k
is e
r t
T ll
)
l
. NUREG/CR-57991 A,14 l
,_.-.,,_..._._-.1.-1
6 Comparisons Between Yankee. II, IL Robinson, anti Cahert Cliffs This section will [vovide quantitative comprisons
' $cen Yankee,11 B. Robinson, and Calvert Cl.ffs dTM=
T*'*-T >)'"E '
(A 4)
.ne dowrcomer cool: lown rates sh<rtly after flo*
dl M
stagnadon. As discussed previously, the most severe cwitng transient is expected to occur in a SBLOCA scenano leading to cornplete stagnation. This is postu-H the wtal Gu}d ruiting volume is based upon that of lated to occur w hen the break liquid outflow exceeds the de cold legs, mlet annulus, and dow ncomer, the initial cwling rates for the three plants are given in lable safety injecdon inflow. The minimaru break area and diameter (cr this condition have teen computed for the Al @ geomenic plant values are wly approximate Yankee,11. B. Robinson, and Calvert Chffs plants with me intended papose w iUustrate We relatsse using Eq. (A.1). The comparison is made in cooldown potential between plants. Further more, the Table A.2; note that all three plants have a similar hypothetical calculation assumes all plants stagnate at brcnk size of about 1 1/2 in. His is because all three 480 F primary circuit water temperature. Au thiec plana han nearly de same HPl cripachy, about 1.5 plants have similar safety injection (IIPI) capacity ard h3/s at -130*F. Iloweves, tle volume of Yankee (as would have similar break outflows, near 100 lbm/s or Wumated by the downtomer volume)is significantly 3
1.5 ft /s. Breaks of this size and greater would result smanu Wan de een wy plants ud b the NRC s,
I IS study. Thus, the }mitial cooling rate in Yankee is in primary liquid les els dropping, and subsequent flow stagnation. However,if the break is toolarge 1.2 F/s, m M. B. Robinson 0.6 f/s, and in Calvert (24 in.), then depressurir.ation will likely o: cur before Clifts,0.2 F/s. This serves to quahuttively illustrate a severo cooldown can occur.
that i ankee should te expected to undergo a more sevue Wennal shcek than the other two plants.
in ieder to provide a useful, albeit incomplete, perspec-Hoevu, a is not clear mat the l'rS evahtations far tise on the relative PTS cooldown potential between
- 11. B. Robmson2 or Calvert Chifs3 uwt thermal tran-Yankee, H. B. Robinson, and Calvert Cliffs, the initial s nu as mue as walkcur under flow stagnation.
cooldown rates are computed he largest nLtc of cool-nsu en tune s ken available to thoroughly ing will occur shortly after flow stagnation, that is,
#"II8 ' "'
^
before large volumetric mixing and wall hem transfer taiacd in Refs. 2 and 3.
l<come important. De mixed-mean fluid temperature is then giv:n by Eq. (A.3). We can compute the iniual cooling rate to be the time derivative of this equation, namely l
A.15 gggnq/cg.$7w l
7 Summary and Conclusions General consideration has been given to small break b) After flow stagnation occurs (34 min), the initial E
- ISCA thermal hydraulics of significance to l'fS, with downcemer Guid cooling can be very vvere, over emphasis on the potential to proceed to stagnation of 1.5*8',s. Ilowever, the cocidown will soon be e
primary kiop rutural circulauon flow. For break sites moderated due to mixing with hot w ater in the in the range of I 1/2 to 3 in. in diameter, liquid out-cold legs, inlet annulus, downcomer, and lower flow will typically execed the safety injection capacity plenum. The specific cooldown behavior is con-(at 700 psi) producing a " break" in liquid continuity trolled by fluid fluid miting phenornena in these
' and thus interrupt neural circulation. Injection of cold primary system regions. REhtlX has been shown (1207) safety makeup water into a " stagnant" down-to have ternarkable predictive capabihty for such comer region will produce a rapid, perhaps severe, eml-I'TS behavior,10 ing of the pressure vessel wall. An independent ana-lytical nrocedure was develoint to quantify the transient -
thennal response of the downeomer fluid temperature c) Yankee Atomic Electric Compmy has employed and wall test transfer. His model predicts a mixed-the REhilX code to determine the transient mean downcomer fluid temperature in a manner similar thennal-hydraulic behavior in the downcomer to the REhilX code. This model was used to paramet-regi n. SAIC has deveksped an mdependent tool rically evtluate (le induence on fluid temperature of to evaluate the mixed-mean downcomer fluid tem-variations in wall heat transfer, safety injection flow, perature. Comparison tetween the two codes rates, and water volumes participating in the miting show execilent agrectnent. It is concluded that process. Calculations specific to the Yankee Nuclear YAEC has correctly applied REhtlX to evaluate Power Station plant were performed to evaluate the the SHLOCA sequence. llence, the results po-YAEC submittal using the REh11X code. Considera.
vided by YAEC should le considered test estimate tion was then given to the potential for both less snd va!ues. De mixed.mean fluid temperature of Fig.
more severe cooldown transients. ne following con.
A.8 slould te considered as toth test estimate and clusions and recommendations can te made based upan as an upper lound. Ilowever, the ongmal down-the present independent PTS thermal-hydraulic comer fluid thermal response of YAEC 1735 (Fig.
evaluations.
A.2) represetus the colder plume region below the cold legs.
a) Small-break (1 1/2 to 3 in.) LOCA transients should be expected to proceed to primary coolant 4 The Yankee plant has design features and opera-flow stagnation. YAEC has correctly considered donal characteristics that are unique. These in-such a limiting scenario for the Yankee planL.
' Clude a " recessed # upper barrel, e narrow down-llowever, the sequence frer}uency analysis should comer 4hermal shield region, as a large safety in-only consider smau break initiators ir; this range;
. jection flow for a "smalr' plant The potential that is, smaller and larger breaks will have either a trnpact,if any, on REh11X hydrodynamic and phe-much less severe downcomer & U cooklown c.:
nomenologicci modeling and assumptions should will depressurize to a low pressure. A flo v stag, be addressed, llowcer, the REhilX code has been nation scenario should be considered for all PTS well assessed and must te considered as represent-evaluations, including the 11. B. Rebinson and ing test-estimate behavior.
Calvert Cliffs plants; it appears that this may not have teen adequately addressed tused upon cursory review of Refs.2 and 3.
- NUREG/CR 5799 A.16 i'
._..._.m
.. _. _ _.. ~.. _. - _ _ _ _.-
I 8-Reconiniendations I
Y AEC las provided comrsete & tails from tie phtnts nent to le reviev,wl to ensure that adequate con-f SBLOCA REMIX calculation, tr(imbng plume tem-sideration was given to f!ow stagnation scenarios.
peraturrs in the downcomer segbrts telow tk cold -
legs. The fracture inochanics technical c2ncrts should A unsible need crists for research on the influcree of ensure that appropriate fluid temperaturer. ha"e teen plant s;ccifs features on the PI'S thennal hydraulic used in their evaluations, behavior and the inherent assu:nskms of RiiMIX.
Specifically. (o!&r) plants with narrow downcomet Therrral hydraulic SilLOCA trarnients used in the l'i'S gaps and large llPI flows could pose unique considera-studh
'n the 11.11. Robinson and Calvert Clif fs bons not covered in the awesunent liasis of REMIX, I
k s
9 9
A.17 NUREG/CR-$799
..w.,
,, - + - - -..
. - ---.,- -,~,. - -
-.. -.,.----.- ----, - ~ -- -- ---
~- _... -.- -
-. -.. ~. -.... ~ - - -..
9 Heferences 1, Reactor Frence Vessel Evalu,2 tion Report, 1 turd Int. Cort. on Reactor Thertrva flydraulics.
- Yankee Atomic Electric Company Regr.>rt No.
Newport, RI, Oct. 1518,1985, Peper 13E, 1735, July 1990.a Nuclear Sciccce and Engineer.ing,Y108, Issue i
No.2, pp.198 207 June 1991.8
- 2. Pressurized Tiermal Shock Evaluation of the
!!. B. RoNnson Unit 2 Nuclear Power Plant, 7.
T,0. Theofanous, J. l 12 Chance, and K. A.
NUREO/CR-4183, March 1985.6 Williams, The Thermalllydraulics ofSalOCAs l
Relatht to Pressurited Therrrn! Shock, 3.
A. D. Spriggs et al., TRAC.rri Ar. clysis of NUREGSR $13),Octobet 1988.6 Peter.tial Presswited Thermal Shock Transients at Cahert Clyys, Unit 1, NUREGICR.4109, 8.
C. D. Inctcher et al., RELAPS Thermal flydraulic February 1985.b Analyses ofPressurized Thermal Shotk in Sequences for Ihe ll. B. Robinson Unit 2 4.
D. P, Botarth, K. A. Williams, and J. W.
Pressurfied Water Reactor NUREO!CR.3977, Minarick," Review of YAEC 1735. Reactor Serlember 1984.8 Pressure Vess! Evaluation Repert," S AIC letter to R. D. Cheverton, Oak Ridge National.
9.
K. lyer,11. P. Nourtakhsh, and T. O. %eofanous, 1sboratory, November 2,1990!
REMIX: A computer Programfor 1erveratwe Transients Due to liigh Presswe injection After S.
K. lycr and T,0. ncotanous,
- Decay of Interr uption of Natwal Circulation, NUREGXR.
Buoyancy.DrivenStratifiedlayern.ith A;pli.
3701, May 1986.6 cations to Pressurized netmal Shock Reactor Predictions," Proceedings of198$ N/ITC,
- 10. T, O. Theofanous and 11. Yan A Unyied Aug 4 7,1985, Denver, CO, pp. 358 371 (also
' Interpretation of One-fifth to Full. Scale Therrrw!
NUREGICR 3100, May 1984).*
Mixing Experiments Related to Pressurized Thermal Shock, NUREO/CR.5677. Februsy 6.
K. lyer and T. O. Theofanous, *Fkuling Limited 1991.6 Thermal Mixing: De Case of liigh Fr injection,"
dAvailable from NRC Pubhc Document Renn for a fee.
bAvailable for parsAase frwn GPO Sales Program.
8
'Available for purchase from aganizathm spmsoring pubbcatim Available from publu: and special iethrucal hisancs.
b cited. and/or frorn authors and/or re.apents (doeurnented leucts),
Available for purchese frun GPO Sales Prograrn.
NUREG/CR 5799 A.18
{
I
. - ~.
.,,w,
.-,--,..-,-..<,--wm--~.e.w--ws
.~._,
e,.-mr
-.vv---
m-v
-,.e
.r
-e..
+v-
-- -.__._-.._ - _ _m.._.z.--_
Table A.1 Thermaland liydrault l'eameters Used f< r tie braneme livalurtions lhgh lYxau e inje4 tion TernperatureJF 130 liigh heuure injection Flow. ItVs '
45,67.901 j
initial Wall Ternperatuse at Flow Stagnadon. *F 480 Pressum Vessel Wall Arca,It2 460 Downcaner Water Ternperature at Stagrution, 'F 4(O Fluid Volurnes Participating in
- Stagnant Mixing fil 200,333,8(0*
Wall llcal Transfer Coeffkient, IlthfikF
.O,4(O*,105
)
i
'llest tsu. mate si tmt d study.
?
l
-b h
I t-i l
A.19 NURiiG/CH 5799
- .- ~._.. _ _., _. _ _ _ _ _ -.. _. ~. -. _... _. - _.. -. ~.... -. ~.. _ -. _ _ _. - -. - _, -.. ~. - -... - - - _. - _..... _. ~. -
Table A.2 PTS Comparison lietwut YNPS, it. B. Itobinson, and Calvert Cliffs PWRS j
1 Plant Yankee Rowe it D. Robinen Calvert Cliffs 1
PLANT PARAMETERS' flot Zero Power (MWt) 0.5 8.3 9.4 3
Downcomer Volume (f1 )
86 I84 706 I
f Cold leg Flow Area (ft'.
l.6 4.1 4.6 Coldleg Diameter (ft) 1.5 2.3 2.5 Saturation Pressure in S. G. (psia) 750 1088 911 3
IIPI Flow (at above P)(f1 h) 1.5 1.6 1.9 Assumed Dowrcomer Temp at Stagnation (*F) 480 480 480 IIPl Water Temp (*Fj 130 130 130
-i i
CALCUl ATED PARAMETFJtS Minimum Ilteak Area for Stagnation (in.2) 1.2 1.3 1.6 Minimum Ilreak Diameter for Stagnation (in.).
1.2 1.3 1.4 initial Downcomer Cooldown with Perfect lip 1 Mixing (*F/s) 1.2 0.6 0.4
'Nundu:n are omroxirnate and used to alusume relane valaes l= tween plants.
I t
I l
l I
L L
L L
l NUREG/CR-5799 A.20
~
1 PTS EVALUATON SMALL BREAK LOCA PUMP SUOTON BREAK CASE K78 WmCUT LOSS OF OFFSITE POWER 2500 i.
4 geoo.
k' }
1500 -- -
1 4
6 imo
%f 500 D~ -
t 200 400 600 SOC 1000 1200 1400 1600 TrME (SECONDS)
Fig. A.I.
YNPS SBLOCA downeomer pressure from YAEC-1735.
l l
A.21 NUREG/CR-5799
PTS EVALUATON SMALL BREAK LOCA PUMP SUCTON BREAK CASE HZP WinCUT OFFSTIE POWER e:c 5~-
a l,
\\s (Y
un.
1
- v
-w W
Q IM 0
200 400 400 000 1000 1200 1400 1600 Tu (SECONDS)
Fig. A.2.
YNPS SBLOCA downcomer ter; perature from YAEC 1735.
NUREG/CR-5799 A.22
c 3
"M Pump WRt T
l WR4 ea wo MR$
Ws.WT d--
4 /
a
% Ugt ~ cs : O y
. m: : r.: %
y Cold Leg N
-- 1 Di[U V
T*
'd Ct NT
+-T-U#
t i
=
i Desmcomer 1
a 'l
-g 0,
- i. I
. T,,
3 l
k Loop Seal T
Tm, f
e
.j t = ad v
Tu Lower Plenium Fig. A.3.
Conceptual dermition of flow regimes in the cold leg and downcommer irgions due to HPl water (Theofanous et al).
A.23 NUREG/CR-5799
l Heat Trans. Coet. (Eh-f^2 F) 500
~ ~
/
400
/
/
/
/
"~"
300
/
/
/
-[ -
Water Temp.130 F 100
.- 212 F Il
- - 48c r r
i 0
0 50 100 150 200 250 300 350 Delta T (F)
Nu = 0.1(Gr Pr)"(1/3)
Fig. A.4.
Natural convection wall heat transfer coefficient vs differential fluid to w all -
temperature.
I l
NUREG/CR-5799 A.24
Temperature (F) 500i 4
A5abatic
- --- h. 400 bt.-t'2 F 400
- - + - -
h - 10^5 ht>-f'2-F
+
\\
\\
\\
\\
g-
"r-
~!-
300
'N 3.. s.
)
'?-) g m.s's 200
' h-i e -
~.~?.-__
~-
" "
- seen =.., :
N i
i i
i 100 0
200 400 600 800 1000 1200 Time (sec)
~
HPI Flow - 90 lbWsec Fig..A.S.
Downcomer fluid temperature parametric influence of walliv.at transfer coefficient.
1 A.25 NUREG/CR-5799
,,, Temperature (F) i HPl flow = 90 tm's
"-" hPI flow = 67 tm's 400 t-
-r-HPl flow = 45 tm's N
.N.s
' N si 300 2'
h ~ q -
t
's
~,.
- % %un g
' '- ~ ~ ww 200 100 O
200 403 600 800 1000 1200 Time (sec) h = 400 btuMr-ft*2-F Fig. A.6.
Downcomer fluid temperature pammetric influence of HPl flow rate.
NUREGER-5799 A.26
Temperature (F) 500 Mix Vol. 200 h*3 333 fta3 600 tt*3 400 m
N N
N.N N,
N s 4
N 300
's*,
's N
s s
s s
s '
~,,,,,
200
~__
100' 0
200 400 600 800 1000 1200-Time (sec) h = 400 btu.hr/tta2 F. HPl-9010m4oc Fig. A.7.
Downcomer fluid tem [cmture parameterinfluence of water mixing volume.
l' L
l A.27 NUREG/CR-5799
~
i Temperature (F) l 4
-- siao.- os enra 500
- YAEC 4oo 300
~s 4em g'~~
~.
^200
~
100 0
0-200 400 600-800 1000 1200_
_1400 Time (sec)
- Tin = 120 F i
j Fig. A.8.
Comparison tetween the mixc4mcan fit 8d temperatures in the SBLOCA sequence.
l:
l s-i-
l P
- /
NUREG/CR-5799 A.28
- - -,, _ ~
l Temperature (F) 600
-- Si fkw - 96 tms 500 YAEC
's i
400
' 's s
?00
%- ~~~,
200 100 0
O 200 400 600 800 1000 1200 1400 Time (sec)
Tnn - 170 F
~
Fig. A.9.
Parametric influence on the mixed mean fluid temperatun: of 170'F safety injection water.
l A.29 NUREG/CR-5799
ir.s nn
=>ma. a,o.ua g g)T J'
'O.AACN ACwp & CDO PLME
/
i' m mcc NwnN u. / a' K, I
~
>s,tu.nto,vr o
i
- r&=
g'tW =,:a e sc <eo
- 7 l
-ex: nacec.o m Plat 4j ccac=5 g.2 se msoa *NA,-sci ooxrntrmos/M u,c.Dcm na
+q 4,
l N
Na_-
w* m =c e' _\\
itK SJ890RT PLNE
__ y _
z w'
x wmmm i
W Nsm m
- m uL J PN tW1.D CEPCST 8hD
~
. /y.
AC10R S.P8CRT GL65CT & LUG
,\\y I
5' TM"la uL SHELD $88CIP PN
,b TWPwi. EHC.D CfJCCR w
- ', e/
d N
"L W 8CR.9c.L h
d N
d
\\
- s
' )
- s\\
s mm J l
4 mei l
gk" -
G M
l 1
e fh7 y
T(j' 4
i 1
m c o r u n:
"LACCNG 700 PCAD s
Fig. A.10.
YNPS pressurc vessel venical cross section.
NUREGER-5799 A30
l Attachment A to SAIC Report No.91-650) (Appendix A.2)
Letter from J. D. Ilaseltine (YAEC) to Ur. 'C, BYR 91-082, June 26,1991 YANKEE ATOMIC ELECTRIC COMPANY y
580 Main Street, kotton, Massachnetts 017401393 n.9be kIS' me United States Nuclete Regulatory Co'Jnisiten Occusent Control Otsk Washington, DC 10555
Reference:
(e)
Licente he OPR 3 (Docket ho, 5019)
(b) U$NRC Letter, P. $sars to 6. Papanic, dated June 20, 1991
Subject:
Request for ACdittenai inferration Concerning REMIX Calculations (TAC 8 535)
Dear Sirsi Enclosac is our response to the ir.fermatien requested in Rtference (2).
We trust you will find this inforttatica. satisfactory.
If you naec additional information, please feel free tc contact us.
Very truly yours',
YAhKEt ATCHI ELICTRIC COMPANY John D. Hasaltine Project Director GP/mst /tlp/C7t\\10 Enclosures c:
USHRC Rtgion 1 USNR0 Resident inspector. YkPS
- 6. Elliet (NRC, NRR)
W. Russ011 (hRC, NRR)
A31 NUREG/CR 5799
F Attachmer,t (Enclosure for HYR 91-082)
In our July 1990 submittal, the limiting small break De attached figure provides three curves:
LOCA (1 5/16 in. pump suction) analysis conserva-tively assumed loop stagnation. Under these condi-
- 1) he original downcomer temperature response for tions, the REhi1X code was used to predict the tem.
the limiting small break LOCA based on the upper perature distribution of the injection plume in the plume temperature near the cold leg nozzle.
downcomer region. The application of REhflX to the
- 2) De downcorner temperattae response based on Yankee ECCS design is conservative. Based on the in-mixed mean temperature as a result of the thermal-jection velocities consistent whh the Yankee ECCS shield and core barrel geometry, design, more complete mixing would occur in the cold
- 3) ne downcomer temperature response based on leg than predicted from REMIX, nis would result in mixed-mean temperature without crediting the a warmer plume temperature than reported in the July lower plenum volume.
1990 submittal.
He attxhed figure shows that accounting for the Due to tne unique geometry of the Yankee vessel, the unique geometry of the Yankee thermal-shield and core mixing volume in the REhflX model included the barrel is equivalent to crediting the lower plenum as a lower plenum, ne lower plenum acts as a mixing mixing volume. Rus,if we were not to credit the volume in the Yankee vessel because of the thermal lower plenum as a mixing volume, conservatisms in-shield and core barrel geometry. ne thermal shield is cluded in our July 1990 submittal would offset the im-relatively close to the mactor vessel wall (-2-in. gap).
pact resulting in a similar downcomer temperature As a result, the plume emanating from the cold leg
- response, would be contained between the thennat shield and core barrel. This plume would also pass through the core It shuld also be pointed out that the application of barrel region, herefore, before reaching the vessel REMIX is conservative, and the assumption of stagna-wall the plume would mix with fluid in the lower tion leads to a conservative vessel temperature re-plenum region. Thus, the vessel wall under these con-sponse. Based on the higher injection velocities con.
ditions would see a temperature closer to the mixed-sistent with the Yankee ECCS design, more complete mean temperature calculated with REM:X.
mixing would occur in the cold leg than predicted with REMIX resalting irt a warmer plume temperature than The results reported in our July 1990 submittal for the reported in our July 1990 submittal.
downcomer temperature were conservatively based on the upper plume temperature predicted from REMIX Because of the unique geometry of the Yankee vessel and not the mixed mean temperature. In response to thermal-shield and core barrel region the lower plenum your recent request (6/2fy91) we have evaluated the im-volume should be included in the mixing volume in pxt of not crediting the lower plenum mixing volume the RFMIX calculation. Even if the lower plenum was in the REMIX calculation. De results of our evalua-not credited in the REMIX calculation, based on the
' tion are presented in the attached figure, above stated conservations, the vessel temperature response reported in our July 1990 submittal remains bounding.
NUREG/CR 5799 A.52
=
?l l
[
g
],
W
_] _j!
'_ pJ Jj w
~I g
e t g*
31 1
/1 E
g
'y If p
I f
h.\\
g L
g3
. l j
7 n
I /,/
.n y
Ehg
/ j
-g h
/
C3
/
/
._)
,y
/,
t
?
./ r
.,,/
'f
, f
,,,f s.-
y
'/J"
-a g
g (d)BWTL1W9JW WWM O N Tamparatura - HCy guadu traak '
E3*.00A Downcoas:
l A.33 NUREG/CR-5799
Attachment B to SAIC Report No. 91-6501 (Appendix A.2)
In response to our teleconference on July 5,1991, we Answen 0.89 ft3 per loop, and 120*r.
provide the following:
Question:
What was the thermal shield h.al transfer Question:
When did the REMIX cale-lation start?
an a?
Answen The REMIX calculation started at 150 s-2 or both sides of the thennat Answen 153 ft f shield, this typrescnts a one quarter seg-Question:
What was the initial temperature assurred ment of the tlame.1 shield.
in the REMIX calculation?
Answen The initial tempenture was 476 F.
Question:
What v as the volume assumed in REMIX?
Question:
What w:re the total SI flow rate and tha 3
SI water temperature used in tne REMIA.
TM 5dum = 2M ft, and mixing 3
calculation?
vohane = 203 ft.
NUREG/CR-5799 A.34
RW CONDITION 8 C31 - 8310CA NO SVMtDN a +s ohn 4Weid !?. a 9.J094 77+*3)
"R80 =
476.00 TRP1
- OJA0 AQEP! =
.00t+00 SQWPI =
.393400
!! MEN 8 IONS PCR M1XIN0 CCMPUTATICN8 VOL =
384.10 Voix = -
303.00 D! =
.188 DCL =
1.344 BCL =
31.600 WD =
.371 MMPUTATIONAL PARAMIT2R$
TIN =
150.00 TMAX =
4000.00 tLT =
30.00 TIM #A =
3000.000TIMPR =
2ATIC =
.50 SITA =
.80 1DGNS QNS AND PROPERTIES PCR REAT TRAN8 PEA DELCL =
.18t+00 CELDC =
.84E+00 D2LTS =
.13t+00 DILLP =
.333,00 LP =
.182+00 DELIA =
.182+00 CIIc h
.813-0100CC =
.005-03 ACLR =
168.000 ADCX =
$1.300 AT88 =
133.000 ALFE =
44.000 APH e
.000 ALSE,=
101.000ACAH =
317.000 ALCL =
.11r-03 ALDC =
.112-03 ALTS =
.11E 03.ALLP =
.113 03 ALP =.
.47E-04 ALZJ =
.115-03 ALCS =
.473-04ALDCC =
472'04 AKCL =. 473-03 AXDC =
.473-03 AKT3 =
.4/C.03 AXLP =
.672*C3 AKP =
.342-03. AXLS =
.47E-03 ARC 3 =
.3st-03M N.C =
.3st-03 NCL =
.14E+00 HN =
142*00 MLP =
.14E+00 EP '
.14E+00 XIJ =
.142+00 BC5 =
.14E+00 EC =
.55E-03 To*
$0.0 f0018 AND CIAD PAPJLw.ITina N = S1 XP3 = 31 KP3 = 31 2DCCL = 1 %CCDC = 3 IDCTS = 1 IDCLP = 3 ZDCP = 1 IDCIA = 1 IDCC3 = 1 TIWS =
.3300 AIING =.7473-05 ALFINS =.3113 0$
i i
A.35 NUREG/CR-5790
l i
T!XI
' T3t TI TC T.TUXP t[tc UC UN TR HFI 200.000 449 117 503.390 313.40s 394.7s4
.140 1.388
.344 13.47 390.000, 438 364 476.444 398 373 378.908
.140 1.333
.338 13.40 300.000 406.453 455.354
'388.041 380 003
.140 1.177
.333 14.08 330.000 388.897 438.304 373.383 344.373
.140 1 143
.338 14.78 400.000 373.113 418.391 353.733 331.316
.140 1.093
.334 15.43 450.000 358.334 403.444 353.500 319.091
.140 1.048
.330 16.09' B00.000 348.064 385.383 345.080 308.107
.140 1.030
.318 18.73'
$10 000 333.046 374.705 337.333 397.770
.140 1.007'
.313 17 40 800.000 333 050 363.090 330.370 388.711
.140
.973
.31c 14.03 830.000 311 970 351.865 334.183 300.343
.140
.857
.307 18.654 700.000 303.715 343.033 318.439 372.810
.140
.930
.304 19.359 750.000 394.304 033.518 313.073 385.781
.140
.505
.201 16.882 800.000 386.369 334.895 308.375 139.488
.140
.981
.199 30.43:
850.000 378.148 316.443 303.963 353.484
.140
.888
.197 31.01t 900.000 373 487 304.493 199.794 347.431
.140
.555
.184 31.631 98&.000 386.337 303.817 194.348 343,006
.140
.833
.193 23.13!
1000.000 360,865 395.405 193.645 337.559
.140
.833
.190 33.73t 1050.000 355.401 389.814 189.687 337 787
.140
.811
.189 23.345 1100.000 380.040 384.783 184.560 339 435
.140 791
.387 33.742 1150.000- 246.038 379.439 184 011 336.088
.140 780
.186 34.357 1200.000 -341 887 378.073 181.541-333.783
.140
.770
.184 24.1J7 1350 000 338.001 371.134 179.731 319.949
.140
.748
.183 28.169 1300.000 334.414 367.137 177.310 318.955
.140
.781
.181 08.830 1330.000 331.004 383.173 175.488 314.058
.140 741
.190 26.093 1400.000 337.993 359.685 173.888 311.810
.140
.733
.279 36.530 1450.000 235.119 388.937 173.373 309.480
.140
.738
.178 36.444 1500.000 333.447 313.859 170.485
$07.031
.140
.718
.177 27.313 1550 000 319.943 350.434 189.030 306.033
.140
.707
.176 37.683 1800.000 317.449 344.088 167.857 303.149
.140
.70s
.278 34.071 1880.000 315.494 345.745 16A.418 201.434
.140
.894
.174 38.401 1700.000 213.487 343.697 155.433 199.981
.140
.683
.173 38.708 1750.000 211.019 341.833 184.348 198.305
.140
.683
.173 39.038 1800.000 209.8s9 339.473 183.338 196.938
.140
.474
.173 29.335 1450.000 208.339 337.878 183.475 158.819
.;40
.874
.171 29.659 1900.000 306.716 335.*13 141.549 194.005
.140
.673
.170 30.03s 1950.000 305 194 334.208 160.851 193.134
.140-
.4E6
.170 30.314 2000.000
- 965 333.320 180.036 191.874
.140
.445
.189 30.335 2050.000 aw. 24a 331.886 159.483 191.330
.140
.661
.189 30.663 3100.000 301.537 330.306 154.781 190.308
.140
.601
.148 30.905 3180.000
'300.456 339.046 158.318 189.434
.140
.451
.188 31.113 3300 000 199 445 337.185 187.888 188.47s
.140
.444
.168 31.341 2280.000 198.487 336.518 157.087 187.836
.140
.443
.181 31.57S 3300.000 197.389 335.088 186.933 186.590
.140
.443
.idd 31.844 3350.000 198.746 334.035 158.104 183.989
.140
.643
.166 33 047 3400.000 195.054 333.373 183 708 185.548
.140
.638
.365 33.135 1450 000 198.310 333.477 188.350 184.847
.140
.836
.145 33.349 3500.000 194.810 231.473 184.988 184.317
.140
.635
.165 33.509 3350.000 193.8$3 331.385 154.854 184.138
.140
.635
.145 33.884 3400.000 193.333 330.143 134.384 103.536
.140
.635
.155 33.595 3650.000 193.548 319.801 153.737 133.718
.140
.633
.164 33.838 3700.000 193.098 318.858 153 434 183.335
.340
.833
.264 33.080-3750.000 191 579 318.171 153.189 181.700
.140
.433
.344 33 335 3800.000 191.031 218.293 153.335 181.851
.140
.633
.164 33.199 3850.000 190.637 317.s49 152.748 181.336
.140
.414
.384 33.331 3900.000 190.190 337.101 153.871 180.964
.140
.613
.163 13.481 3950.000 189.777 314.543 152 359 180.600
.140
.515
.183 23.873 3000.000 189.387 216.016 353.153 180.388
.140
.416
.183 33.685 l
l l
\\
l NUREG/CR-5799 A.36
l C1NTt1LMNE TEMPDATL*R18 AT DCWCciam thcAT!cNs, TIMS
- 3000.000 E31GHT TRCH CL CIWTRZ TtKPERATL*R2 1.344 188.307 3.484 180.887 8.888 184.038 31.887 188.881 31.C88 187.589 38.048 184.838 D!nW3IONUS# S'MIC.D'388 TEXPt1AWR3 301.884
.000 303.885
.c04 305.393
.008 307.090
.013 334.513
.113 387.983
.380 368.483
.400 285 380
.888 387.713
.704 3C5.411
.333 308.c88 1.00o A37 NUREG/CR-5799
~
,m.,
TDG
.TM TB TC 7.TL%F RC UC UE rR EFI 3C86.000 138.317 318.1t1 181.810 178.488
.140
.618
.163 33.88 3100 000 188.667 314.470 181.133 173.807
.140
. sis
.ist 3 3.8 3.
L AkD 2 DO NOT DAACRIT A RCCT stop - Program termir.ated.
l NUREGSR-5799 A.38
.e
Appendix A.3 Iteview of Accident Sequence Identification and Quantification in the Yankee llowe Pressurized Therinal Shock Analysis J. W. Minarici, S AIC 708 South Illinois Avenue, E101 Oak Ridge. TN 37830 The following comments have been developed based 4.
Here appear to be inconsistencies between the on a review of tne Yarikee Rowe (YNPS)ITS analy-cunent operating gocedures and the sequences sis submittal," Reactor Pressure Vessel Evaluation -
modeled in the MS analysis. For example, Repvt," YAEC No.1735, July 1990. This review except for bleed and feed coolmg following failure made use of the Yankee Rowe Updated FS AR, emer-of the safety injection pumps, the charging pumps gency operating piecedures currendy applicable to the are assumed in the analysis to te tripped as a plant, the MS analysis of 11. B Robinson performed result of the safety injecdon signal asso ;iated with by ORNL for NRC (NUREG/CR-4183), and data de-most initiating events and accident sequences veloped in the NUREG ll50 program. Major ques-which were analyred, Because of this, MCS re-tions exist conceming analysis assumpuons, the com-pressurization was limited in the analysis to the pleteness and appropriateness of the YNPS accident Si pump shutoff head. Ilowever, restart of the sequences, and the estimated frequencies for modeled charging pumps (which could pessuriac the MCS sequences.
to the pimary rehef valve setpoint) and reener-guation of the pessurint heaters are specific po-General Comments redural steps f ollowing terminadon of SI and in a situation where all steam generators blow dowrt 1.
De lirnited ckieumentation of the bases for screen.
ing accident inidators and quandfying the sequence Initiatine Event and Accident
,plit fractions pevents detailed review and verifi-Seauence Seleetion cation. Numerous accident sequence split frac-tions are justified by references to system event 1.
In the reported review of YNPS PR A es ent and trees and fault trees. Rese trees are not provided.
faut trees by the PTS analysts on pp. 6-26, it is Also, the frequencies of support states on which not clear what cnteria were usd to idendfy poten-various split fractions are conditioned are not pro-tial overcooling initiating events and at :ident vided; therefore, reproduedon of the sequence fre-sequences based on modals tb m pesumably quencies is not possible.
devchyed to address the potential for core damage.
Related to this, Table 6.5.2.1-4 was derived based 2,
ne overall resolution of initiating event selec4n on a review of the categorization in NUREG/CR-for sequence development was significantly ca cr 3862, " Development of Transient initiating Event than the resolution used in the MS analysis of Frequencies for Use in Probabilistic Risk Assess-EL B. Robinson. For example, the YNPS analysis ments " llowever, the focus of this reference was considered only two full-size steam line breaks and to suppirt development of core damage PRAs by only one small.bmak LOCA. Other events wre identifying transient events which caused scrams.
screened out based on frequency or consequence.
Because of this, transients listed in NUREG/
CR-3862 may not adequately bound all transient 3.
No attempt is made to systematically bound the classes with the potential for overcooling.
PTS risks stemming from initiating events and accident sequences that are screened out on the 2.
Section 6.6 states that only those events resuhing grounds of frequency or consequence. There is in a cooldown from hot soaked conditions with a often ambiguity as to whether a given initiating rate in excess of 200'F/h. and a relatively high event is to be conservatively grouped with another MCS pressure were considered capable of p> sing a initiator for which the anticipated consequence is MS concem. While this is consistent with the l
more severe, or whether it is being excluded from YNPS Critical Safety Function Status Tree F-0.4, further consideration. For the initiating events lhTEGRITY (which oocs not recognize an immi.
l-and accident sequences that are explicidy screened nent MS condition for MCS cold leg tempera-j' out, no attempt is made to determine their aggre-tures >280cF), it may be a nonconservative i
gate contribution to risk. Considemtion of rcid-threshold for transient evaluation. Replacement of
(
ual PTS risk played an important role in the the cooldown rate screening criterion (with, for ORNL analyses of Oconce, Calvert Cliffs, and ex ample, a criterion incorporating cooldown H. B. Robinson.
A.39 NUREG/CR 5799
magnitude) could have sigmficant impact on the to be intuitive for an operator to attempt to YNPS ITS risk profile.
isolate a LOCA.
Since tmnsients with less severe cool downs are b.
A cold water accident may be possible if tha far more frequent, their ITS usk could dominate valves on an isolated loop are suddenly even if the probability of through-wall crack is coened (operator enor or spunous valve oper-several orders of magnitude less. Exclusion of all anon). Such an event could result in asym.
less severe transients from funher analysis should metric cooling of the reactor vessel, combined be camfully justified.
with a rapid increase in reactivity, is this type of accident possible at YNPS and,if so, 3.
Consistent with the above comment, consequence what are its ITS consequences?
and Inquerwy arguments are used to screen out fm-f+r consideration ITS sequences that are
- 5. There is apparently no consideration (pp. 5-41) of imtiated by stud-open secondary side relief valves sequences initiated by loss of main feedwater fol.
and small main stcae line breaks. It should be lowed by actuation of cold emergency feedwater noted that:
(EFW). '!hese sequences were addressed in the Robinson study, a.
In the NRC's H. B. Robinson PTS study, the inidating event frequency associated with 6.
While loss of control air is discussed in the report, stuck open secondary vah es and small secon-it is not specifically addressed. The updated dary side ruptures is relativeh high (2E-2/yr).
FSAR notes that the charging pump fluid drive If there are arguments to rede e the correspon-speed is controlled by a pneumatic signal tased on ding YNPS frequencies relative to the pressurizer level. What is the charging pump Robinson number to a degree that warrants speed on loss of air? If the charging pumps fail to the exclusion of this initiating event, they high-speed then this initiator may require addi-need to be provided in greater detail.
tional scrutiny since the main feed control valves fail-as-is on loss of air.
b.
In the YNPS ITS Study, a rationale for ex-ciuding sequences involving stuck open sec-Human Rellabilitv Considerations ondary valves (Sect. 6.6.3.1.1) is that they result in cooldown rates which are less than 1.
While the HEP curves used in the YNPS analysis the screening criterion of 200 F/h. However, appear to be fairly conservative, their application the Robinson study indicates a significant in the main steam line break (MSLB) and small.
magnitude of cooldown for such sequences.
break LOCA event trees appears to have generated in fact, the class of sequences that is ITS optimistic HRA estimates. For example, wL at risk-dominant at Robinson involves stuck knowledge of the detailed application, it cannot be open secondary valves following reactor trip.
detennined what degree of credit has been given in the HEP estimates by using the seven modifying Such sequences account for a total through wall crack factors listed on pp. 6-83. The following are frequency of ~ IE-8/yr on that plant. The ruuonale some specific concerns related to the HEP -
given for excluding plant trip as an initiator in the estimates:
YNPS Study does not address the issue of the potential for the consequent sticking open of steam dump valves a.
On pp. 6-108 and 6-125, the EEP estimates or secondary side safety valves if these are challenged.
in the IE-7 to IE-5 range for cooldown con-Numerous stuck open secondary side safeties have been trol and system realignment for recirculation historically observed in the industry.
seem low. For comparison, the NUREG.
I150 analyses generally avoid the use of fail-4 Yankee Rowe is one of the few commercial ute probabilities less than IE-3 for any single nuclear plants in the United States with main operator action.
coolant isolation valves. Operation of these valves may affectITS sequences for the plant.
b.
On pp. 6-163, where the llEP estimates are For example:
discussed in more detail, the steps leading to the IE-7 probability estimate for failure of a.
Successful isolation of a small break LOCA the operator to control cooldown (with feed-in a main coolant loop could result in repres-water isolation successful and no SG blow-surization to the norinal plant opermirg pres-down) are not given. For the scenarios in sure. While closure of the loop valves to iso-which one or more SGs blow down, the IEP late a break is not addressed in the operating derivations provided on pp. 6-163 reveal that, procedures, closure of the PORV or its block effectively, operator error has not been valve to isolate a transient induced LOCA is accounted for in model'ng recovery from the included in several procedures. It would seem failure to isolate affected SGs. Hardware fail-ures therefore dominate. 'l o assumption is NURbG/CR-5799 A.40
=
~~
. - - _ ~.... ~.
that the hardware failure gobability of 3E-3 action, if the recovery llEP is set to IE-1, is significandy higher than operator error-this gives: OY1 = IE-3 and'OY2 s IE-3.
probability. His requires justification.
To OY3, a screening human error probability of IE-2 should be added to give OY3 =
c, ne probabilities anached to OYO-OYC are 2.2E-2.
conditional 5 the occurrence of a gevious operator enor (feedwater isolation). As such, d.
OYO OYC are cooldown control even:s 1E 2 is low, For comparison, in the 11.11.
conditioned on fa; lure to isolate feedwater.
Robinson FTS analysis, no credit is given for De assignment of a failure probability IE-1 AIM control if there has been failure to iso-to exh event would reflact the gencral late the affected SGs. In general, tic assign-NUREG 1150 approach of giving limited ment of operator error probabilities condi.
credit for operator actions following earlier tioned on the occurrence of a previous error in operator ermrs in the same sequercc.
an accident sequence should be conservative.
Foi comparison, the approach afopted in NUREO-1150 is not to give credit for second Erfquency and Branch Probabilily and sutmequent errors (in aggrega:e) of more Estimation than a factor of 0,1, i.e. cut sets with n ulti-pie ermrs are generally assigned IIEps of no less than IE-4 L On pp. 6 58, in the characterization of small break LOCA frequencies,[unitioning the pipe break fre-In general, it appears that in the YNPS PTS analysis, quency between the <1 in. and I in. to 2 in.
the IIEP pobability/ time correlation may have been ranges assu nes that all srnall breaks are effec.
applied to individual opentor acGoru, in each sequence tively guillotine, excluding scena:ios involving and the resultant probabilities then multiplied together,.
small breaks in larger piping, his assumption is ne HEP curves are more appropriately applied to the unjustified. A more appropriate treatment would combination of actions tequired to provide a given be to retain the Bayesian updated WASH 1400 '
function within a single sequence (e.g. feedwater isola-pipe rupture frequencies without the use of scale-tion and conuol). De detailed YNPS HRA calcula-dowa argumeats.
tions would need to be reviewed to assess the appropri-ateness of the HEP curve apphcation, Also, on pp. 6-171 it is stated that small-break LOCAs are " limiting" loss of coulant accidents Proposed screening requantification for HRA values:
from a IYrS consequence perspative. This does not preclude the possibility of significant risk a.-
01 - Failure of opemtor to isolate feedwater -
contribution from sequences initiated by larger.
after trip. Replace probability of 1.7E-4 (or LOCAs. If larger LOCa sequences are not to be 1.3E 4 as stated on pp. 6-162) by 1E 2, a
- considered expdcitly, their frequencies should be number reflecting typical assumptions for the
. conservatively added to the small brc^ LOCA failure probability associated with rule-based seqtence frequencies.=
actions in the NUREG-1150 study. Also, in-cease FI (feedwater isolation) failure poha-Pmiosed screening requantification: Replace the bility in the smal.1-break LOCA evenr. tree by 5.24E 4/yr small break LOCA initiating event two orders of egnitude, frequency with a YNPS WASil 1400 update value of 2.1E-3. Altermtely, a value of IE-3/yr for b,
OY - Failure to coatrol cooldown. He basis small-break LOCA, as utilized in NUREG il50 :
for the probability of OYO (tailure to control analyses (see Table 8.2-4 of NUREG/ CR-4550, cooldown given successful SG isolation) is -
Vol.1, Rev.1, " Analysis of Core Damage not provided. Typica! rule based actions are Frequency: Internal Events Methodology") could assigned failure probabilides in the range also he employed. Use of either of these values -
2E-3 to SE 2 in NUREG-1150. Without would be reasonable, considering the uncertainties knowledge of procedures and specific actions, associated with the estimates, the recommended screening value for OYO is IE-2.
Also, unless medium-and large-break LOCAs have been explicitly considered in the analysis, the i
c.
For events OY1 - OY3 (failure to control frequencies for these initiators should he added to L
cooldown given multiple SG blowdown), a the revised small break LOCA frequency to tound j.
human e Tor probability should to be added to larger LOCA contributions. Based on NUREG-each recovery failure probability. Since re-1150 data, medium and large LOCAs have a total covery is conditicaed on previous occur.rence frequency of 1.5E-3/yr, of an error of commission in OY2 and OY3, the approach nuopted in NUREO-1150 allows 2.
Partitioning pipe rupture probability uniformly limited credit for success of a sub-sequent among pipe sections that has no dependence on A.4 i NUREG/CR-5799
pipe size (pp. 6-59) seems inappropriate. A better sicw of the referenced boiler feed pernp uip fault assumpuon would tv: that cf a tmiform volume tree, evaluadon of the appropriateness of the density c,f inidal cracks in pipework wends. This probability used is not possible. Nevertheless, would imply a partitioning of rupture frequency this number seems low based on plant experience, that is dependent on pipe siic. De effect of the frequerry distnbution arguments used in the Proposed scnening n'quanaficadon: Increase of YNPS study is to scale down the pipe rupture fre-the HF0 probability by at least ont order of. nag-quencies in criucal pipe section.. (Based on a nitude (assuming one failure of feed pump auto footr>ote on pp. 6-7Ei, it appears that these k)ca.
uip m pbnt lifetime) would be appropriate in the tiondipendent scalc-down erguments were not MSLB cvent tree. Increase of the F1(feedwater uhimately used. This reqnites confirmation.)
isolation) failure prolubility in the small-break LOC A event (ree (in audition to the increae sug-3.
Vanous split fraedons in the steam line breal gested under Human Reliabihty Corcidera4 ions) event tree (pp. ti-108) are condi600x1 on events would also be appropriate, that are neither dcfined previou tly in the event tree nor characterized as support statcs (e.g., event 6.
Event G20*GNO (pp. 6-107) is the blowdown of a CNX which is condidoned on DC availabihty, single SG given a MSLB downstream of the
~
and events GN/C2 condidoned on nonreturn valve NRV, with the NRV actuanon train available.
(NRV) actuation train availability). In general, One NRV failure to close occarred m June 1982.
tne heavy rehance of event tree quanuficadon on if the NRVs are not tested monthly, then the plant fault tree and event tree models, w hich are NRV failure protubility used in the analysis not provided, allows only broad spht fraction (5.8SE-3) may be low. ht addition, auto-closure quantification checks. The absence of support of the NRVs has vuly recently twe1 implemented.
state frequency data pteeludes checks on the Does prior t-i g provide confidence in the relia-sequenec frequencies.
bility of fast closure as assumed in the analysis?
4 From the perspecuve of hardware rehabihty Credit for Ooerator Actions and (pp. 6-108 and 6-12.5), probability assumptions AIternate Procedural Aetions regarding failure to isolate / control feedwater (01 MSLB event in;e, FI in sma!!-break LOCA ever.*-
1.
It is not clear that exchiding credit for various trec) seem low given plant. specific experience.
acuons/ systems available to provide feedwater (1s For example, among the last 10 years of LERs are idenufied in Table 6.5.3.21)in MSLB sequences two events involving loss of feedwater control is a conservative assumption ir. the context of (event date 11/27/80, and LER No. 86 012-00).
irrS rid Proposed screening requantificadon: While the 2.
On pp. 6120, exclusion of operator depressuriza-implications of these LERs for event tree quanufi-ticm of the vesselin the small-break LOCA event catica requires more detailed systca s/ procedure 3 tree to permit LPSI injection is not necessarily knowledge, replacement of the OI/FI probability conservative relative to P fS poten0at In general, as desenbed under Human Relbbilityfonsidera-discussion of each item for which " credit is not uom should bound any modified hardware reliabil-taken" relative to PTS (vs core damage potential) ity estimates.
is warranted.
5.
Failure to automatically trip the boiler feed pumps 3.
Emergency feedwater actuation / control (pp. 6-120) in the MSLB cvent tree (pp. 6-98 and 6-108)is is not modeled in the small break LOCA event assigned a probability of 2.8E 3 coeditioned on tree. Sirce loss of feedwater control is a potential DC power being available. Accordmg to route to overcooling, the rational for this exclu-Table 6.5.3.2-1, no credit is given in this number sion should be provided.
for manual actions to trip the feed pumps.
Among the last 10 years of LERs is the report of a failure of the fced pumps to auto-trip. Without NUREG/CR-5799 A.42
[
I Appendix 11 r.
Technical Evaluation Report I
Review of the YAEC Thermal lIydraulic Accident Sequence Analyses for Assessment of Pressurized Thermal Shock for the Yankee Nuclear Power Station Contents
- hac B.1 Introduedan B3 B.2 Scope of INEL Revicw....
B3 B.2.1 Materials Needed to Complete Transient Thermal Hydraulic Review B.4 B.2.2 INEL Request for Additional Information B.5 B3 Review Eindings and Discussion of Major Cor.: erns B.5 B3.1 Judification for the Limiting Small Break LOCA.
B.5 B.3.2 Isolation of a Small Break LOCA B.5 B.33 Treatment of Pressurizer Noncouihbrium Thermodynamics B.5 B.4 - Conclusion..
B.6 Attachment A to Appendix B - Request for Additional Information..
B.7 General Questions...
B.7 Loss-of Coolant Accident B.7 Steam Generator Tube Rupture B.8 Main Steam Line Break D.9
. Feedwater -
B.9 Steam Generator Blowdown B.9 r.
B.1 NUREG/CR-5799
Review of the YAEC Thermal llydraulie Accident Sequence Analyses for Assessment of Pressurized Thermal Shock for the Yankee Nuclear Power Station Leonard W. Ward Craig Kulberg IL 1 Introduetion
- 3) The maximum break size wht:h can be isolated was not presented nor discussed in the report.
This Technical Evaluation Report presents a review of Since the Emergency Opecting Procedures do not pre.
the thermal hydrau!ic analyses performed oy the Yankee vent the coerator from iso:ating the break, the response Atomic Electric Company (YAEC) to address pressur, f r this esent should also be meluded for PTS ized thermal shock for the Yankee Nudear lher evaluanom.
Station. De therrnal hydraulic accidern analysis events were dtscuswd argi presented in the YAEC Reput B sed on the above concems, the thennal hydraulie No.1735, entitled " Reactor Pressure Vessel Evaluation analyses pnmed in the Y AEC report are not accept Report / dated July 1990. The thermal hydraulic analy.
able fa use in assuring tac worst case has been idenu-ses included the following events:
fied for l'TS evaluations of the Y mkee Nuclear Power Stadon. Resoludon of the above concems in addition
- 1) Main steam line break (5 cases) htsining rerivynses to a request for additional in.
t
- 2) Excessive feed flow (2 caseM fami tion regardmg the other events in the repxt
- 3) Small break LOCA (1 case) would be reeded to complete the NS review for the Yankee Nuclear Power Station In reviewing the aNve events, only the reactor coolant system pressure and downcomer temperature responses He desion of the scope of the Idaho National were provided. As suct, here was insufficient infor.
Engmeenng Laboratory (INEL) redew is described in mation reganiing the other key primas y and secondary Sect. D.2, A discussion of the major concems regard-system transient response parameters to permit a thor.
ing the review (.f the thermal hydrauhc analyses con-ough and proper review A review of the information tained m, the YAEC report is presented in Sect. B.3, provided in the report, however, identified several major while the conclusions are given in Sect. B,4.
concerns which will require resolution. %cse concems Attachment A presents a Request fe,r Additional include the following:
info-rnadon (RAr) which would be needed to complete this review. Under normal review circumstsaces, the
- 1) The prese.rir.er nonequilibrium model used in the Techmcal Evaluation Report would be wntten upon analyses did not properly account for the heat receipt of the responses to the RAl. Ilowever,due to transfer governing the thermal conditions in this the limited review schedule and the unavailability of regton during refill and repressurization. As a con _
sequence, the approach used in modeling the ther, ume to regpond to the questions, the R AI is included as part of this evaluation.
mal behavior in the pressurizer will tend to over-predict heat removal from the pressurizer fluid and underprecct peak pressure during rerdi. De effeci IL2 Scope of INEL Review of this nonconservative pressurizer modchng tech.
nique en those events which experience pressuriza-De scope of the INT'. effort consists of reviewing ine tion needs to be evaluated to demonstrate that the accident analyses contained in the YAEC document approcch does not adversely affect the results nor No.1735 entitled "Reac tor Pressure Vessel Evaluation change the conclusions presented in the report.
Report." He details of the review are summarized below.
- 2) No justification was provided to demonstrare that the small break LOCA presented in the report is De thermal hydraulic portion of the YAEC Report ik worst case for Pressurized Thermal Shock No 1735 describing tbc accident analyses contained in (MS) considaations. The analyses of a spectrum Sect. 6.6 enti' led "nermal. Hydraulic Analyses for of breaks needs to be evaluated and dtscussed to Representative Sequences" was reviewed. Revdatory demonstrate that the minimum temperature and Guide 1J51 entitled " Format and Content of Plant-maximum pessure ressmse for the break pre-Specific Pressurized Hermal shock Safety Analysis sented in the report bounds that for a spectrum of Reports for Pressuriecd Water Reactors" was used as break sizes.
guidance for the review.
B.3 NUREG/CR 5799
l Re sequences or accidents presented in the Y AEC nermal hydra.,4c accident safety analysis sections of Repc:No.1735 were reviewed to determine the tech-the Yankee FSAR desenbing the results of the LOCA, nical adequacy and acceptabibly of the analyses. He ourcooling, and steam generator tule rupture event (
review addressed the following major areas:
Rese results should be applicable to the current plant cycle.
- 1) ne RETRAN methodology - he methods wcre Detailed system descriptions for the following:
reviewed to assure the code used in the transient analyses properly trea's the thermal and hydraulic a.
ECCS (high and low pressure safety injection behavior for application to PTS evmts, pumps, Ecumulators) and any other injection sys-tems which can deliver coolant to reactor coolant
- 2) Input model--- A limited review of the nodal system. Provide head ss flow curves for injecdon model was performed to evaluate the adequacy of systems and descri ion of accumulators (cover gas.
the iiput model. Sensinvities of the input model pressure, elevadon head, and t.mk liquid such as nodalization of key reactor system coolant inventory).
components, various input information including wall tacoolant heat transfer coefficients, and ini.
b.
Pressurizer, spray and heater systems, geometry of tial conditiont were evaluated.
the internals, and the level control system.
- 3) Transients - Re transients presented in Sect. 6.6 c.
Steam generatcr secondary safety relief valve pres-of the Y AEC report were reviewed for their tech-sures and capacities, ADV steam flow capxity and nical adequacy with consideratica of items 1 and 2 rated condinons, main / auxiliary / emergency feed-
- above, water system flows, and secondary inventory at 100%,50% power, and HZP.
- 4) Completeness - The thermal hydraulic accident e.alyses were reviewed to assure that the limiting Because only RCS pressme and downcomer tempera-it msient had been identified with justification pro-ture were pmvided for each of the transients presented vided to demonstrate that the worst case produced in Sect. 6.6 of the YAEC report, the inforn.ation is in-the mininium temperature and maximum pressure sufficient for performmg a thorough review of the condition. Of particular importance is that the thermal and hydraulic system response to these events.
worst case initial conditions and appropriate To facilitate the proper teview of the transients, please operator actions and equipment / system responses provide the following plot information f or exh of the have teen properly xcounted for in the spectrum transients (includmg all cases for each event) discussed of transient events. As such, L review of the key in Sect. 6.6:
systems / equipment and operator actions from the appropriate Emergency Operattng Procedures was a.
steam generator pressure and liquid mass; afso performed, b.
feedwater mass flow rate and secondary break mass flow rate; This ef^ ort does not include a review of the methods c.
total SI mass flow, break mass flow, and quality and models used to compute the mixing of the fluid in (include PORV flow and quality if appropriate);
the injection section and downcornr regions of the d.
pressurizer two-phase level, steam temperature, wssel.
liquid tempemture, and wall temperature; e.
upper head am! upper plenum void fraction and Following the initial review of YAEC Report fluid temperatures:
No.1735, additional supplemental information was f.
discharge leg and hot-leg maes flow rates, quali.
requested in order to complete the review of the acci-ties, and temperatures; dent analyses. The request for this supplementary in-g.
core inlet, average, and outlet temperatures; formation w as transmitted to the YAEC and mcluded.
h.
core inlet / outlet mass flow rates and qualities;
- i. RETR AN parameters used as input to the mixing B,2.1 Matcrials Needed to calculanons if not included in above plots; Cornplete Transient
- j. please pmvide RCS pressure and downcomer tem-Thermal Ilydraulic perature for those cases w here the information was Review not provided in the report.
The Y AEC reports identified as Refs. 6.6.5,6.6.6, and Several statements were made regardmg " previous" or 6.6.7 in YAEC Report No.1735 dated July 1990.
- past" analyses in Sect. 6.6, but no references were provided. Please provide the documents describing the YAEC Yankee Emergency Operating Pracedures that previous analyses. Rese include refeances to pevious address LOCA, overcooling events, and steam generator of Past analyses on tube rupture.
NUREG/CR 5799 B.4
- pp. 6171,last paragraph regarding LOCA analyses.
cooled down to shutdown cooling conditions. The
- pp G172,last paragraph in regards to the SGTR operator procedures instrtet the operators to initiate a event.
cooldown to RllR conditions during a small break
-- pp. 6177, third paragraph regarding second MSLil LOCA. During the cooljown the RCS willicGli ad caac.
reprenurize quickly to a pressure where ECC injection
- pp. 6180, first paragraph for sixth MSLB case.
flow into the RCS equals the flow out de break.
Thus, as break size decreases, the refill will occur Of the above information requested,only the following ec t.cr in time and produce higher pressures after repres-
' information was reviewed for this review:
surization. The larger break sizes will senti and repres.
surize at lower tem [rratures but will repressurize to ac topical report describing the RETRAN methoda!-
lower pressures than that for the smaller breaks. An i
ogy; analysis of the spectrurn of breaks which experience b.
the Yankee Emergency Operating Procedures; and refill and repressurizauon is expected to prtxtuce the c.
a YAEC submittal for the Reanalysis of the Main RCS pressure responses illustrated in Fig.11.1, An Steam Line Rupture Event - Cycle XVI, dated evaluation of these break conditions is identified for the June 10,1983.
I'pS evaluation for tho.se breaks that refill and repres-surue. Also, performing a cooldown will increase Using the above materials and the results of the analy-ECC flow into the RCS and result in potentially lower e, of the thermal hydraulic events presented in the downcomer temperatures than that for the 15/14in.
Y AEC report,a Request for Additionalinformation break presented in the report.
(RAI) needed to complete the review effort is listed below %is information is normally evaluated prior to 11. 3. 2 Isolation of a Small issuance of a H.R; however,in view of the schedular lireak LOCA constraints and the limited time within which the -
Utility can respond to such requests, the itemized list De possibility of a small break occurring that can bc
- of questions is therefore contained in this report. De isolated during the event was also not discussed. De RAI is presented in the following section, maximum break size that can be isolated was not pre-sented nor discussed m the report. His worst break 11. 2. 2 INEL Request for that can be isolated needs to be compared to the limit.
Additional Information ing small break LOCA that results in refill and repres-surization of the RCS from item 1) above to assure die From a review of Sect. 6.6 of YAEC Report worst break has tren analyzed. Also discuss the poten-No.1735, dated July 1990, additional information tial for the ECC and charging systems to pressurne the was identified that is needed to compete the assessment RCS should the RCS becone refilled with ECC water of the thermal hydraulic events contained in the report.
after isolation.
The RAl is listed in Attachment A.
11. 3. 3 Treatment of Pressurizer B.3 Review Findings and Noucquilibrium Thermo-
-Discussion -of Major _
dynamics
- CcncernF s anTRAN treatment of the pressurizer during insurges following refill of the RCS includes a two-2 With consideration to the questions discussed in region representation of the pressurizer. The upper Attachment A, the major issues regarding this review region contains steam while the lower region accom-include:
modates the liquid. He RETRAN code allows one to model heat transfer between (1) the steam and the upper a) justification for the limiting s nall break LOCA; walls of the pressuruct and (2) between the upper b) isolation of a small break LOCA; and steam and lower liquid regions. The YAEC modeled c) treatment of pressurizer nonequilibruim the heat transfer between the upper steam and lower thermodynamics.
liquid regions only using a heat transfer coefficient of 2
50 Bru/h ft.*F. Because this method may not be rep-
. De above major issues are discussed in detail below.
resentative of the actual heat transfer mechanisms that l
occur in the pressurizer during insurges, justification
- S.3.1 Justification for the that this epproach bounds the actual tchavior in the Limiting Small Break pressurizer is needed. During insurges the pressur-izer LOCA will accumulate liquid thereby compressing the upper i
steam region which superheats, he dominant rnecha-Insufficient information wac presented in the report to nism that controls peak pressure during insurges is justify that the 1-5/l&in. break is the most limiting therefore the picssuri;cr wall surface area in contact break for PTS considerations. Furthermore, for break with the steam and the temperature difference between sizes -2 in. and smaller, the RCS will need to be B.5 NUREG/CR 5799
the walls and steam, Because the steam is nearly stag-above events, there was insufficient information regas, nant, the heat transfer coefficient is c spected to be ding the other key primary and secondary system tran-2 about 5-10 BttVh ft #F. Because the surface of the sient response parameters to permit a thorotgh and liquid in contact with the steam quk kly saturates, a proper review. A review of the information provided in
]
thermal layer or barrier is created which insulates the the YAEC report, iwwever, identined seveml major i
upper steam region from the lower region containing concerns which will require resolution. These concems
)
the liquid. After several feet of liquid accumulates in included the following the pressurtzer, mixing near the surface becomes dimin-ished and the upper steam region can be cortsidered to
- 1) The pressurizer nonequilibrium model used in the be thermally insulated from the liquid for the remainder analyses did not properly account for the heat of theinsurge. in view of these considemtions,the transfer governing the thermal conditions in this YAEC method of modeling the heat transfer between ngion during refill and repressurizauon of the the steam and liquid regions may be noncon-servadve.
RCS As a consequence, the approach used in Funbermore, modeling the lower liquid region es a modeling the thermal behavior in the pressurizer single region presupposes perfect mixing in this region may tend to overpredict heat removal from,he which also artificially lowers the liquid temperature av pressuriier steam region and undergadict Fak fluid is added during the insurge. As such,the use of a pressure during refill. The effect of this noncon-rather high heat transfer coefficient between the steam servadve pressurizer modeling technique on those and liquid mgions, coupled with an artificially low events w hich experience pn ssurization needs to be mixed mean temperature for the liquid, could result in evaluated to demonstrate that the appmach does not lower peak pressures calculate for the PTS transients adversely affect the results nor change the conclu-that experience refill. A more appropriate model would sions presented in the report.
include a three region pmssurizer consisting of Iwo lower liquid regions and an apper su:am region. In
- 2) The justificadon was insufficient to demonstrate view of the YAEC modeling techniques, justification that the small break LOCA presented in the report that the heat transfer coefficient of SO Btuih-ft2/F and is the worst case for l'fS considerations. The use of two regions bounds the actual or expected behaw analyses of a spectrum of breaks needs to be pro-ior needs to be provided.
vided to demonstra e that the minimum tempera.
ture and maximum pressure response for the break Lastly, the upper head region should also be modeled pmsented in the report bounds that for a spectrum as a nonequilibrium region to properly treat the refill of break sizes.
and repressurization prtress.
- 3) De maximum break size which can be isolated Based on IM above concerns, the thermal hydraulic was not presentec nor discussed in the report.
analyses presented in the YAEC report are not accep.
Since the Emergency Operating Procedures do not tab!c for use in assuring the worst case has been identi.
prevent the operator from isolating the break, the -
fled for PTS of the Yankee Nuclear Power Station, response for this event should also be included for Resolution of the above concerns in addition to obtain.
PTS evaluations. The operation of the ECC and ing responses to the Request for AdditionalInforma.
charging systems following isolation should also tion, presented in Attachment A regaro.g ail of the be discussed in regard to the potential for additional events presented in the report, would be needed to com-pressurization of the RCS.
plete the FTS review for the Yankee Nuclear Power
- Station, Based on the above concerns, the themtal hydr,tulie analyses presented in the YAEC report are not axep-B.4 Conclusiori table fw use in asswing the wwst case has been identi-fled for PTS of the Yank. ; Nuclear Power Station.
es udm oMe ah cmcems, b aMm to &
A zeview of the thermal-hydraulic analyses presented in t in' g responses to & Request fm Agmal m
[
the YAEC Report No.1735 was performed to evaluate mation mga6g be du evems m k p, u, the technical approach used as a basis to address Pres-needed to complete the FTS review for the Yankee surized Thennal Shock for the Yankee Nuclear Power ea wer Stadon.
l Station. The thermal / hydraulic analyses included the following events:
- 1) main steam line break (5 cases);
- 2) excessive feed flow (2 cases);
- 3) small break LOCA (1 case).
Because only the reactor coolant system pressure and downcomer temperature res;unses were pmvided for the I
NUREG/CR-5799 B.6 Iu
Attachment A to App ndi 11 Request for Additional Information Gencral QuestlonS thennal behavim of the pressurizer for those events which circrience refill of the RCS. The following in.
formanmis needed:
On pp. 6-167, two criteria for evaluating cool-dow n events are identified in regard to critenn 1), provide a) please provide benchmarks justifying the abibty of justification that transients with a cooldown rate less than 200T/h need not be considered for FTS evaloa.
that model to pedict pressunzer noncquilibrium tions. For example, a cooldown rate slightly <200T/h behavior during hquid insurges and outsurges.
may result m a higher pessure/ low temperature com.
Both sepante cfl'ects and integral tests should be bination that is more hmiting than that for the cases provided. Comparisons to plant data should also k Fovided if availaNe.
w hich are strictly limited to a cooldo un of 200 F/h or 7
Show the effect cf the use of the interfacial heat more. His condition could occur following isolation 2
transfer coefficient of 50 Btu /h ft.T and the of, or refill of, the Reactor Cmlant System (RCS) fol, lowing a small break, single k)wer liquid region representation en peak pressure predictions for a) above. insurge tran-Picase provide the " previous analyses and engineering sients with a range of inlet temperatures and hquid simulations' ider.tified in item 3) on pp. 6-169 mventories similar to that expected for the hkee plant should be provided.
During discussions between the INEL (L Ward), the NRC (M. Mayfield), r.nd the YAEC (P. Bergeron), the Plear desenbe the nonequihbrium thermodyremic thermal hydraulic analyses used the RETRAN non, modeling of the remainder of the reactor coolant sys-equilibrium two-region model in the pressurizer. A tem other than the pressurizer? Was the upgr head of heat transfer coefficient of 50 Btu /h ft T was used to the textor vessel modeled assuming nonequilibrium 2
model heat transfer between the upper steam region snd thumodynam{cs? If not,explam why nonequilibrium the lower liquid region. No heat transfer was modchi thermodynamtes is not important to the repressurir.a.
between the pressuriter walls and upper steam region.
tion process when many of the transients can develop During insurges of liquid into the pressurizer, the hauid a steam bubble in this region following refii) of the RCS.
will compress the steam causing the steam to super.
heat. The pressurizer walls in contact with the steam Desenbe the RIITRAN nonequilibrium modeling of the will act as a heat sink and influence the peak pressure achieved during the insurge. Because the steam is fluid in the loop piping during injection. While perfee.
mi ing of the ECC mjection acts to enhance depresur-h basically stagnant, heat transfea coe ficients between r
the steam and preuurizer walls is of the order of zation tarough condensation in the injection section, 2
the addition of cold, relatively unmixed ECC fluid 5-10 Btu /h ft +T, Because the steam ts basically stag-w hich inters the core region may reduce boiling and nant, heat transfer between the steam and hquid regions have a more significant effect on depressurizmion (fur-will cause a saturated layer to develop at the steam-ther increasing ECC flow) and minimum temperature.
liquid interface, the steam region wili quickly become Wadfy the RETRAN caculated parameters used in the insulated from the lower liquid region. As such, there REMIX code and the EPRI muung model.
is very httle or no heat transfer between the liquid and steam regions. He YAEC approach is therefore con-Aec, dent sidered nonconservative since tne model will have a LOSS-0f-COof an1 i
tendency to overpedict heat mmoval from the upper steam region which will result in tower peak pressures Provide justification that the t-5/16-in. break is the computed during events where the RCS refills. Also, limiting break size for PTS evaluations. Since larger the use of ringle lower liquid region further acts to break sizes can result in lower temperatures, provide reduce peak pressure since any fluid entering this region the results of larger break sizes to show that combina-with be perfectly mited throughout the liquid region tions of minimum temperature and maximum pres regardless of the amount of liquid in this region. This sure for these larger breat sizes are bounded by the single region repesentation of the lower liquid region 1-5/16-in. break. Smaller breaks which requin cool-therefore will minimize the lower liquid region temper-down m shutdown cooling conditions and which will ature and further enhance the heat removed from the experience repressurization during the event should also upper steam region w hich produces lower peak pres-be discussed. Since the EOPs do not identify when sures. This modeling technique is considen:d incorrect, uo!-down should be iniuated if RCS pressure is high More importantly, because of the nnaconservative (i.e. >h35 psig), the earliest time into the event that nature of the approach, additional justification is needed the operators would initiate a cooldown should be to assure the use of this appoach bounds the expected B.7 NUREG/CR-5799
assumed and emhinu should te at the mattmum What is the temperature of the !?CC and charging water I
alkm able mte.
injected by the pimps and accumulators used in the LOCA analyses? What is that minimum allowable What is k masimum break sire that com te isolated ternperature of the ECC water source? What a the wid a hat is de mini um temperature that could be minimurn ternperature of the charging flow?
achieved fw this t<cak? If such a small break LOCA is isolated just prior to sc'ill,is there sufficient tare for Please povide judcadon thSi the IIZP condition is the operaurs to throttle ECC and charging now to pre-the worst initial cor,iition for LOCA ITS evaluationt vent RC'; pressure from retuming to full power opent.
Please esplain w hy all other m(des of operadon are not ing pres.sare?
more hmiting for ITS considerauons?
What operator acuons are assumed in the LOCA analy-Please explain the methat used to cool die plant to ses? Iktw do these actions affect minimurn if.mpera-Shutdown cooling condidons following a small break ture and maximurn pressure achieved during LOCAs?
LOCA with hot water in that pressurircr aiki a bubble and hot water in the upper head? Does the p>tential for What systems can operate following a LOCA to mini-RCS pressure Ithavior impant ITS as k operator mire RCS ternperature and maximlie itCS pessure?
auemps to reduce RCS pressure to Sherdown cooling Are let down ard ausiliary emergency riray systems conditions by throttling ECC flow while also main-available -mi could the pressurizer heaters actuate upon taining the minimum subcooling, recovery s! pressurlier level upon refill of the RCS by k ECCS and/or charging pump 37 Provide justif; cation that the suction leg break location is the worst kratio. for this break? Inclu:Sc breakr in Provide pl a u the following for t e 15/16-in, break:
Inc hot leg piping in the justificadon.
h a) penurtur level; Figute E6 5 shows the temperature decreasing at the b) hot and cold leg two phase levels, flow rates, end of the analysit. Please provide k ternainder of the qualiides, and temperntures; analysis showing the time at which the temperature c) total ECC mass flow rate, break mass flow rate, reaches a minimum, and quality; da steam genentar prenures and levels; Steam Generator Tube Riipture c) upper head two phase level and fluid temperatures; and O core void fraaion, Please provide the analyses (referred to on pp. 6 112) thatjustifies thst me most severe cooldown for the fust Figure 6 6-4 picsents downcomer pressure for tic 10 min of a tu!e rupture event occurs for a single guil-15/16-in. break. If the operator initiated a cooklown lotine tube rupture.
using the steam generntors at could the ircreased ECC addiu,15 min into the event,Please explain why pressure stabilizes at 1250 psia.
on result ?n lower down-comer temperatures than that presented in Fig. 6.6 5 and then tmon refill of tne RCS, could system pressure Was the tuk rupture analysis carried out to the estab.
increase above that shown at the end of the pessure lishment of cutdown cooling? In gusticular, the plant plot of Fig. 6.6-47 The analysis should te presented must le cocal to shutdown cooling cmditions for out to the time refill occurs and where the break flow long term heat removal Please demonstrate tnat dur-equilibrates with injection flow, lhe results of the ing the cooldown the opentar is able to maintam sub-altitional treaks requested above should to be carried cooling margin and not repressurire the RCS at the out for this refilled condition.
low ternperatures necessary to initiate shutdown cool-ng. What pecautions are taken to pcvent inadvertert Please descr4the wall to. coolant heat transfer model repressurization early in the event and late in the event i
sned for the primary system. Identify tk regions that when low temperature conditions are met for entry into were modeled, the wallwoolant heat transfer coeffi-shutdown cooling.
cients, and the wall nodalization used for the conduc-tion solut on.
Please descrite the initial corxhtions for the tube rup-tun analysis.
What is the earliest time the operators would initiate a O eninN of Secon(lar} S.VStclu E
cooldown of the RCS following those small break LOCAs where heat removal is needed? Plese descrite Sienm Valves the method for cooidown of the RCS following a small break LOCA, and the precautions taken by the ogerator What are theinitialconditio:4 for the secondary valve to pcvent overpressurization r f de system when the opening transients? Identify all control systems that
{
RCS has teen cookd to shutdown cooling entry are acuve during these events. Also identily the ofera-conditions.
tar actions for each event.
NUREG/CR $799 BE
Why was the addition of feedwatu Mecludal for these eqmpment failures and initial conhtions test estimate events 7 or are they (omhk red worst case acmptiom.
Provide the tmis Icr assuming the NftV closes fot the List k minirnurn kmperature and masimum pressure opening of a single high set htSSV but faib m the for each raw that was used for l'TS evaluation. What other vahe ckisure events? Whr is the minimum was the worst case? For example,cu 6;ncluded a tenqvrature achieved if the NRY does not close, with minimum temperature of 140*F with pressures of 1550 i
l end without feedwater addition? Please justify that and 1000 psia idenufied. What pressure was used in omission of this event or providc the trsults of the
& l'IS cvaluation? Was cme 6 carried out through
- analysis, rehlt and repressuritabon of the RCS? What is the site of the IDCA? What cordidons are needed for the pres-For each of these cases de cook'own rates were cited as succ to remain r 1550 psia for this case? What is the a maximum *cx;rcted* cooldaw n rate; are these cryi-impact on this esent of closure of the PORV when the neaing judgen.ents or are these conclusions based on downcomer is at its minimum temperature? What calculations with the RETRAN code? Please explain, usumptions were inade in regard to charging system operauon?
What condiuons are recessary for the op:rators to trip the mair < niant pumps? What is k irnpact on kse llow was the pressurutt level control system mateled?
events (arg the abovt requested event involving non.
closure of the NRV)if k operator trips the main On pp. 6-177, w hat does
- minimal" feed mean?
coolant pumps?
The minimum tem;rrature hv case 4 is based on the
- M alli S(Cain I'lHe Ilr0H k unptying he nin nwr hot w H ahn which h1CS temperature would legin to increme. What actions would to raluired to prevent the hot well from empty-Plcme provide ik following plot information for each ing or MCS temperature to increase at 7,5 min and if of the stearn line break cases:
suc h conditions are possible, what minimurn tempera-ai steam generator pressure and liquid mass; b) feedwater mas now rate and break mass How rate Case $ shows k temperature in Fig. 6.616 decreasing (irchde primary break imormation for LOCA and at k end of the plot and it was stated to continue to opened PORV):
deercase thereafter, Either carry out the analysis untd c) Si now; tem;rrature begins to incrtase or identtfy the minimum 4 press %cr twcehase level, steam temperature.
temperature with the EPRI mixing motel?
bquit. tempe4ature, and wall temperature; c) upper head and upper plenum void frut tion and FeedWflter fluid temp:ramres; f) cold and hot leg loop mass flow rates, qualities, a d temperatures; Please pmvide the information requested under hiain Steamline !!reak.
g) core inlet, average, and outlet temperature; h) core inlet and outlet mass now rate; i) - IEIRAN parameters used as input to k mising St08In GCnerator liiowdoWn calculations if not inchded in above plots; ard j) please provide RCS pressure e.nd downcomer tem-please provide the informabon requested under hiain perature for those cases where the information was -
Steamhne Break.
not provided in the report.
Case 1 assumed a guillotine break of the 24 in. steam line What discharge coefficient was used for case !?
What break size and discNttge coefncient were assumed l'or the other cases? How was the break region nalal-ized? What critical flow model is included in REYRAN and how does k code model break flow that is not critical now?
Please provide a list of operator actions assumed for each of the events.
Provide jualification for not assuming additional NRY valve failures for case I when 2 and 4 NRV vah es were assumed in the other cases? Are tre choice of H.9 NUREG/CR 5799
zli 8
Q RCS Pressure (psia) e-2500
~
"8 A - Break Stre A16A 2* A3* A 4
- A5 A1-1.3125 In SBLOCA Froan MEG Repost 1730
-i 2000 AS 1500
- c..
..).w A4
<i~
1000 33 600 w....
r i
s O
l Time (Seconds)
Fig. B.I SBLOCA - Expected RCS pressure respnse.
1 m
i i
Appendix C l
ORNL lleview of YAEC 1735 Radiation Effects on RTNDT and Char py Upper-Shelf Energy Contents l' arc C.1Intraluetion..................................................................................................................
C.3 C.2 Com posi tion of unde 80 we lo s.....................................................................................
C.3 C.4 C.3 RTNDT Considerations...................................................................................
C.4 C.3.1 Summary of RTNDT Estimates....................................................................
- 0. 3. 2 O rai n S ize Ef fec ts..............................................................................................
C.6 C. 3. 3 Temperat ure Effoc ts.............................................................................................
C,6 C.7 C.3.4 Nickei Elfects....................................................................................................
C.3.5 Summary of Metallurgical and Temperature Effccts on RTNDT........................................
C.7 C.8 Ci Charpy Upper Shelf Energy Considerations C.4.1 Sommary of Upper S helf Energy Estimate s..........................................................s........
C,8 C.4.2 Low Upper. Shelf Energy Effects on Fracture Toughness......,..............................
C.9 C.10 C. 5 S ummary of Rad ation Ef fec ts.............................................................................................
C.I I C.6 References.....................................................................
i C.1 NUREO/CR.5799 i
l
~~
i Appendix C ORNL Review of YAEC 1735 Radiation Effects on RTNDT and Charpy Upper Shelf Energy J. G. Merkle R. K. Nanstad those reported for the BR3 reactor vessel. The jushfi-C,. I lIltrodtiell0n cadm is dua the vessels were fabricated atout the same time and would likely have similar chemistry.
The Yankee Atomic Electric Co. (YAEC) repostl Although that is acceptable for nickel because of includes detailed informadon regarding the materials, known specifications for nickel additions, that justili-fluence estimates, surveillance data, and operating in-cadon should be rejected fw copper locause the copper fortnation as well as their analysis of the cunent ar" content in the welds is a somewhat uncontrolled com-projected RTN1yr and Charpy upper shelf enc 4gy for b nation of that from the steel used to draw the welding cwh nmterial. Sule.umtlal diffciences existed between wire tself and that from the copper coating. It was not the original Y AEC estimates of R1Ntyr and those of an ciert 91 controlled by material specification.1hc the NRC stafI aml consultants. For the plates, the dif-copper content of Linde 80 welds can te quae variable ferences resuhed primarily from the YAEC assertion as shoun by a series of such welds fatricated by timi the A 302 grade 11 plates are rearse grained and, Babcock and Wilcox (B&W). The copper contents for therefore, not sensitive to irnxhation temperature in the the welds in that study varied frorn 0.1510 0.49%
range frem 550 to $0(rF, and that the cmrse gndn win an overall m na of 0.29% and a star dard deviation inlcrostructure also mitigates the potentiMiy embrit-of 0.07%).2 Recent chemical analyses of samples tling effects of nLLet on the lower plate. Regarding from the Midland Unit I reactor vessel have revealed the welds, the chemical compositions are unknown and coppes variations fiom 0.21 to 0.46% with an overa'i YAEC assumed that the copper content (0.18%) and mean of 0.29 wt% arul a standard devhtion of 0.071 nickel content (0.70%) are the same as those of similar and all the welds were fabricated with the same heat o' welds in the Belgian BR3 reator veuel fabricated by weld wire and not of welding nux 81he vanation in Babcock and Wilcox (B&W)in the same time frame as copper, then, can be very large even within one the Yankee vessel. Since Ref. I was issued, discu" wue/Oux combination. The following guidance is sions between Y AEC and NRC have led to conver-provided in Regulatory Guide 1.99, Rev. 2:
gence of the two organirathos' esumates of inadiated RT DT values, in Tables 1 and 2 " weight trrcent coppe'"
N areweight percent nickel" are Oc Lut 1here are many factors contributing to the uncertaintics CFUmalt values for the material, which will regarding the fracture toughness of the Yankee reactor nonnally le the mean of the measured vessel. Among these are the relatively low operaung values for a plate or forging or for wcld temperature (~500T), a snght amount of surveillance samples made with the weld wire heat data, effects of grain site and nlcket content, and lack of numter that matches the critical vessel chemical composition data. Each of these will be dis-weld. If such values are not available, the cussed. 'Ihe two toughness parameters of interest rela' tipper limiting values given in the material tive la the pessurized thesmal shock (FI'S) evaluation specifications to which the vessel was built are de reference temperature (RlhTJT) and the Charpy may be used. If mit available, totutmitix upper-shelf energy. The relationship t< tween Ic%
estimates (mean plus mtitarlitrd.dtyil Charpy upper-shelf energy and fracture toughness is ugn)lused on generic data may be used if aho discussed.
justification is provided if there is no in-fonnation available,0.35% copper and C.2 Composition of Linde 80 1.0x nickel should be assumed.
S Wcids The above guidance was the basis for using the generic The compositions of the Linde 80 welds in de Yankee data for Linde 80 welds, discussed above, to establish vessel are not knen. It is known that they were fab-ricated with coppercated wire and Linde 80 welding aNar. stand. It K., RCAbC. D b, &nd bW am. R. L. Variarwu in flux. The YAEC proposalis to assume a copper con.
nmau cwwat cmm /w thestu w u vaa n < ann tent of 0.18% and nickel content of 0.70%, the same as vessel I-Urr<*timates fo:
weld, f = 2.0$'8and ARTNDT = 359"F(Odette's value de plate materials were of f w o types, the first itycor-rounded up to 300T). Odette's ARTNDT results9 potating chemistry factors calculated on the basis of g g, erally ! css than the preliminary NRC values were gen Lrown plate chenustries, ami the second incorporuung gg chemistry factors calculated by 0 e method of least c reumferential weld have exceeded the PTS screenhg squares from the Yankee surveillance data following the 1
entena.
prtredare described in Replatory Guide 1.99. The ilR3 surveillance data were not considered in this cal' Following the receipt of Odette's estirvates. Ibser's culation. The surveillance based calculations used two gg g
sets of fluence values, different by a factor of two, sets of estimates were closer together lliser's6 Unal because of a YAEC claim that enors had occurred m 1
ARTNDT estimates for plate material werc insed on the original lluence calculations. The estimates for the welds were made on tle tusin of calculated chemistry only the 11R3 surveillance data to avoid the contreversy factors for two chemistries, the Regulatory Guide 1.99 about i AEC surveillance capsule Cuence accuracy.
default chemistry and the ilR3 weld chernistry clairned Um mulupheadve adjmtment for trradiation tempera.
by YAEC to represent the Yankee vesscl weldt im was replaced with un addiuve adjusunent based on i FfF, and the high ruckel content of the lower plate, Upper shelf drop esumates were also made by the Regulatory Guide 1.99 procedure with no adjustment relative to that for the upper plate survedlance slas-for irradiadon temperature, auurning that compensa-mens, was accounted for by adding 70'F to the upper.
plate correladon for ART UT. The reference irradiation N
tion is provided6 by the used1 R curves measured at 500*F. In contrast to the YAEC cetimates in Ref.1 temperature was lowcred from 511'F to 5(FF, same-the preliminary NRC estimams indicated that most what arbitrarily, thus raising the irradiation temperature if not all tic near beltline material RTNDT values adjustment by ll'F. (Time and lluence weighted exceeded the 10CFR50 FTS screening criteria. The average cold leg temperatures based on Tables 2.1 and YAEC estimates of Charpy upper shelf energy values 2.3 of Ref, I produce reference temperatures of 507.1 less than 50 ft-lb were also confirmed.
and 5(M.8'F, re@cetively, the combmed average of which is 50ffF.) Recogninng that a concave down-Odette's estimates 9 of ARTNDT were tmed on a study ward Regulatory G4.ide 1.99 fluence funedon curve prtduces higher ARTNDT estimates than a straight of available data for irradiation temperatures near line on a log-log plot, for fluences m the range of 500 F, nickel effects and a log-log plot of both the interest (see Fig.C3),lliser made both types of esu,,
Yankee Rowe and the IIR3 surveillance data, the latter mates. The latter was based on a linear least squares fit adjusted for irradiadon temperature effects. Odette's nI g% paper to We Gvc BR3 smemance s#
ARTNDT estimate for the upper plate was based on a men results for Duences exceeding 1019 rt cm2 (see r
/
linear interpolation (on log-log paper) between the tw T@ 5.7 of Rd R with an irradiation temperature YAEC surveillance points, using the originally adjustment to the data before fitting.6 'Ite resulting reported fluences,xcording to shift equation was ARTNDT
- 184 57IO34I9. 'F,a (j)
ART DT = 172.16193160,
-(3)
N where f = 4 x 10-19 n/cm2. Using ( = 2.3,b the constants in which are close to those in Eq. (1).
ARTNDT w 2WF, The ARTNDT estimate for the 1he revised ARTNDT estimates for the welds were lower plate was obtained by addmg a +80 F nickel made by the Regulatory Guide 1.99, Revision 2, pro-adjustment to the value for the upper plate, ignoring eedum (w three chemistries, Regulatory Guide 1.99 the differences in nuence between the upper and lower default (0.35% Cu,1.0% Ni), BR3 (0.18% Cu,0.7%
plates,toobtain ARTNDT = 325'F. The ARTNDT Ni), and *best estimate" (0.15% Co,0.7% Ni), the estimates for the asial and circumferential welds were latter chemistry corresponding to the 1(CFR50.61 obtained from a geretic upper bound Regs.'atory Guide abest estimate" values of ART tg. Chemistry and N
- 1.99 type curve for the 5(WF irradiadon data examined,'
Ouence factors were determined from Regulatory Guide accadmg to 1.99. An irradiation temperature adjustment of 50'F was added to the calculated shift. The values labeled
" test estimate" could more accurately be termed a -
8 th correlaum is shghuy dirferent from that stown in Ret 9 prudent estimate,' the conservatism in which provides tecause the one sh<m in Ref. 9 was rined to DR3 as weit as the YAEC si.miUance data. Ahhaugh li) (1) does not sclually syyar.n Ref. 9, it is ennuent with the arproach recommended tvy Odette.
aAs rnentimed later. A more accurate set ter Quencies than une OAs mentimed later a more acwrite set of Quencies than these tecome avaihbel after these calculations were made; they are Lewme availatel after thche calculatkw were made; they are inchuled chewhere in th i report. -
included elsem bere in this returt.
1
^
C.5 NUREG/CR-5799 -
a z
a 1
an incentive for the rtility to make copper-content assert ht a coarse grain microstructure results in an measurements for the Yankee vessel welds? No increased sensitivity to neutrun radiation.
mention is made of the depth in the vessel wall a' which the ARTNtyr values are being calculated, but One of the references they cite is that of Gordon and presumably it is the inside surface.
Klepfer.12 which concluded that coarse fenite grains in iliser's6 and Odette's9 ARTNUT estimates were trans.
ferritic steels ethibit greater irradiation induced shifts mitted to NRC-NRR, w hich selected a combination of due to longer difIudon paths to detect sinks. Likewise, the two sets of estimates for transtnittal to the utihty Nichols and llatriesl3 showed a smillar result. The as the staf f estimates,10 De original peak fluences Gadon and Mepfer wk, however, was performed and licensee estimates of RTNDT as well as the NRC with almost pre territe gram steels and, as stated by staff estimates of RT DT are shown in Table C.I. The Gordon and Klepfer, as substructure develwment N
uninadiated RTNDT values are from Ref. I and lliserp occun in de fonn o@ ark, Mne, meep,id et the assumptions used in their model luome mval t
he NRC ARTNDT values for the plate are Odette. 9 lccause the damaging defects no longer have a rela-s, while those for the welds are Iliser's6 *best estimate" tively direct diffusion path to a ferrite f ernte boundary.
values, with Odette's higher value for the circumferen-As shown in the Yant.ee report, the Yankee phte tial weld included as a precaution. The large dispanty microstructure is largely bainitic; thus, the Gordon-between the NRC and YAEC estimates is evident-Klepfer modci, even if it is correct, may not te appli.
cable to the Yankee case. On the other hard, Approxirnately a month after rnceiving the NRC staff Hawthomeld observed no cf feet of grain site on transi-estimatr the utility transmitted back to NRC revised tion temperature shift for A $33 grade B cbss 1 steel, i
1990 fluence values and RTNDr estirnates.ll %cse Likewise, Hostons and Wotton15 stated diat there were i
revised estimates are shown in Table C.2, which also no differences in quenched and tempered stects tocauw shows a comparison letween ARTNDT calculations of the finet carbide distribution inhertnt in me quenched performed at ORNL by the samt methods chosen by structure. he Yankee plates are quenched and tem-tie NRC staff for t!c preparation of Table C.! and the pered. Recent work by Amayevl6 on chromium-revised YAEC submittal. Table C.2 demonstrates that molybdenum stects reprted no ddlerences between fire the utility has accepted the NRC's basis and methods and coarse grains on the Charpy shift. Finally, for calculating ARTNUT values and, therefore, !$at Trudeau,17 for a 3.25% Ni steel, showed less shift for there is no longer a controversy about surveillance the coarse grain than the fine grain steel, specimen fluences, irradiation temperature effects, or nickel effects.
Dere are other papers in the literature which attempt to examine the effects os grain size on embrittlement.
T!c RTNDT s alues given in Tables C.1 and C.2 do not ne problem is that there are many confounding param.
include the margin terms discussed in Regubtory Gurde eters involved other than the site of the priar austenite 1.99 and 10CFR50, Para. 50.61. The utility applied a grains. De dislocation structure, precipitate stmeture, margin of 567 to the RTNUT estimate for weld metal etc. all conuibute to the rnobility of defects in the
. (see Table 5.9, p. 5-28, of Ref.1) but no margin was microstructure, and these are affected by the fabrication considered for plate. Hiser6 used margins of 347 for process, heat treatment, and chemistry. Re effects of plate and 56T for weld metal, apparently by doubling grain size on embrittlement are, in other words, very 1
the values of c3 n Regulatory Guide 1.99, but did not uncenain and lacking consensus, i
elaborate on the source of these numbers. The values transmitted by NRRio to the utility (see Table C.1)
C.3.3 Temperature Effects did not include margins.
The effects of irradiation temperature on embrittlement C.3.2 Grain Size Effects have been extensively studied. In a general sense, it is agreed that for ferritic low. alloy sta;1s hardening and De YAEC cport offered considerable discussion embrittlement increase with decreasing irradiation regarding the effects of microstructure on sensitivity to
_ temperaturc, at least within a certain_ temperature range.
irradiation. Based on the relatively high austenitizin8 Dis effect has been shown for many steels including temperatarea used for the Yankee pbtes (1750 to A 302 grade B.18 In the range from about 400 to
'200T), they assert that the plates have a relatively 600T, there is considerable scaner even for a given coarse austenite grain si/c and that their assertion is material, indicating a high degree of sensitivity to irra.
supported by BR3 microstructural analyses showing diation temperature in that approximate temperature relatively coarse prior austenile grains. Their assertion range. Dere are insufficient data for the Yankee plates, of relauvely coarse prior-austenite grains being present and none for the welds, with which to ascertain the in the microstructure is likely correct. They further effects of irradiation temperature on those specific materials.
8 hier, Jr.. A. L,NRC. personal communicadon to L G. Merkie-
%cre are many references which could te cited regard.
1 ORNI, Cctober 4.1Wo.
ing irradiation temperature effects.1iiser disc ussed
- !!iser, Jr, A 1., drsh d Ret 6, unamied NUREG/CR-5799 C.6
. a.-
n-w some important ones in his rnemorandum: Stallman19 estimate of the RTNlrg for Oc Yankee veswl seems on A 533 gnde B clan 1 (ilSST Plate 02), Odette20 reasonable and not overly conservative.
on haw and wcld metals, Saulet (unreIcrenced), Farry (unreferenced) on Lirde 80 welds, and low c21 on Lirde C.3.4 Nickel Eifects h0 welds. Odette obsemd a range of irmdiat on tem-perature effects with ddferent materials with l' incrraw Nic kel ha long toen identilbd as a potential " bad in transition temperature shift for exh 1* decreme in actor
- in irradiation emteittlernent of various steels.
irradiation temperature stated as a representative value-Ba'.ed on Se analysea of surveillance data from corn-11 should be noted, in fxt, that observations were noted mercial light + ater reactors, nickel plays a prominent in which emteitdement increased with increasing role in the estimates of embritdement in Regulatory ternpcrature, and the authors emphasite the synergkms Guidc L99 (Rev. 2). Odette ar d Lucas20,25 olserved of other variables such as flu % fluence, and comimi-dat nickel can have a strong elIcet on the transition tion. Stallman rdso observed an average deperdence of temlerature sh It in stects with copper, and that some l' shift inerme per l' decrease in irradiation tempera-data suggest an independent elfect of nkLet at high flu-ture. Saulet's analysis expressed the effect as a ratio, ences. ney aho observed contradictory results, but such dat a shiff at 550"F would be mulupbed by 1.45 the predominant otwervations led them to conclude to estimate the shift at 500*F. Using the Saulet that, for pressure vessel steels in general, nickel en-rnedal, a shift of 100 F at $50^F would be estimated hances embrittlement. As discussed in lliser's memor-as 145*F at $00'F. Using the representative value of andum, llawthomc26.27 reported significant cf fccts of 1* per & gree of inadiation tempermure simply adds nickel on Iwo pairs of plates (copper content was 50'F to the shift at 550'F. For a fluence of 2.16 x 016% in one guir and 0,28% in the other) from split 1019 neutrons /cm2 (>l MeV), the YAEC estimated melts where copper and all other elements were kept shift of 180'F for the upper plate would become 260'F constant, w hile nickel was inemased from 0.21 to using the satio method and 230*F using the additive 0.67% for cach pair. At 2.5 x 1019 nkm2 (>l MeV).
methcd.
the higher. nickel-content plates exhhited temperature shifts of 23% (0.16% Cu) and 44% (0.28% Cu) greater For the Linde 80 weld case, Fabry ottair ed a ratio of than those for the low nickel plates.
1.40 for Linde 80 weids irradiated in DR3, whi'e lowe's analysis of the ilSST Linde 80 welds &tcr-In other studies, Williams et al.2e bserved that nickel o
mined an increase of about 0.7* in the shift for l' de-tended to mitigate the temperature degradence, but the crease in irradiation temgeratmc. Analyzing the same studies were conducted with welds having nickel con.
IISST data, Nanstad and Berggren22 6btained an aver.
tents of alcut 0.3% or less and atout 1.6%. Studies age value of atout 0.5'F. For a fluerce of 1.93 x reported by Maricchiolo, Milclla, and Fini29 also indi, 1019 neutronskm2 (> l MeV), the Y AEC estimated cate a mitigating cffect of increased nickel, ahhough
[
shift of 203*F for the teltline welds would be increased the pn ponderance of their data were for nickel tn<opper
- by values ranging from 25 to 84*F using the various rarms from atout 5 to 25; while Fisher and Buswell30 methods dese-ited above, see enhamed sensitivity with increased nickel dependent on the copper and nickelcontents.
A couple of other pertinent sttdies are those of Williams et al.23 and Ahlf et al.24 For relatively llath niette and Lucas, and Williams et al. emptusize high fluences, the Williams study showed temperature that the cf fects of nickel are not very well understood, dependencies,in the manner discussed above, of 0.5 ne often-inentioned syncrgism of copper and nickel is and 1.0"FfF for two different materials. The Ahlf confoun&d by cifects of other elemenu.md hec.t ueat-study trported dependencies of 0.5,0.9, and 2.15'FrF, ments which may affcct the precipitttion kinetics of far an average of about L2*FrF, for three &fferrnt the copgrr as well as the matriutamage component of 4
materials.
embrittlement. Ahhough there are observations to the contrary, the evidence to support the YAEC claim of in summary, the effects of inadiation temperature are no nickel effect for the lower plate is minimal. Fur.
dependent on many variables and, although there are thetmorr. observations of significant enhancement of specific instances of contradiction, the bulk of the embrittlement from increased nickel make consideration studies reported in the literature indict e higher ernbrit-of a nickel adjusunent the prudent choice. Using 6f.
Llement with lower irradiation temperature in the tem-ferent methods,IIiser and Odette recommended the adde perautre and lluence ranges applicable to the Yankee tion of 70 and 80 F, respectively, to the upper plate situation. All the atove referenced studies involved shift to account for the higher nickel in the lower plate, radiation exposures in the range of 1019 nkm2
(>l MeV). De use of an empirical correlation such as C.3,5 Suminary of Metallurgical and one degree increase in shift for one degree decreate in Teinperature Effects on RTNDT irradiation temperature is certainly not a scientifically satisfying approach, but it is a prudent approach which
%c YAEC regort on the Yankee reactor vessel embrit-is substantiated with a body of rescarth. Based on the tiement presents extensive discussions regardir g the information cited, use of that value to make a best cf fccts of irmdiation temperature, nickel content, and ;
C.7 NUREG/CR 5799 n M
grain sire on neutron embntilement of the vessel Since both the utihty and the NRC utilued Fig. 2 of pistes. Their claim that the probable coarse grain site Regubtory Guide LW for esumating upper shelf of the plates mitigates the cifects of lower inadianon duips, and the irradiation tempeniture for the data base ternperature and higher nicket contei t is not substan-of that figure is 550"F,it is adiisable to coruider the tiated with sufficient cvidence. The confounding elfects elfect of irradiauon temg ature on this estimate.
of so many variables demands prudent choices in cases Ilise,6 noted that," lower irradtahan temperature tends like this where informadon is so sparse. The YAEC to result in greater radiation sensithity (i.e., greater claims may turn out to be correct, but the informadon shilts ard shelf dreps)" but also that. *the Regubtory avaibble at this time is inadequate to allow their use.
Guide is thought to be comervative for irradiation at
%e bases used by the NRC staff for shift estirnates are SyrF; the degree of conservatism is probably sulfi-reasonable under the circumstances and not overly cient to account for the Yankee Rowe operating tem-conservative, perature of 50TF." Informadon regarding the effect of irradia00n temperature on the Charpy t pper-shelf
'""" i"P"* N "#^h""'"# Wren 22 anaiyyd me U )I)0r-SheIf C*4 CharI)8 I
HSST low uggmshelf welds and determmed an effect Energy COH5lderall0HS of akiut -0.022 h4tM meaning that the upper shelf energy dxrcases 0.022 it-lb for each ore degree C.4.1 Summary of Upper Shetf Fahnt heit decrease in irradmtion iemieuture at a flu-Energy Estimates ence of about 8 x 10th n/cm2 &l MeV). For a 50"F decreme in temperature, the decreme in upgr. shelf he utility's estimates of Charpy V notch upper shelf eiergy is about 1.1 It-lb. For an upper-shelf energy of impxt energy at the end of plant life are given on ahmt 40 it-lb, that amount of change is certainly not pp. b26 of Ref. I and then repeated in less detail on substanual.
pp. b4, b5, and b7 of the same reference. %ese estimates are stated as follow 5: (pp. 5-26) "The predie.
De Yankee Rowe survedlance program puduced lions for plate longitudinal Charpy V-notch upper shelf upper shelf-drop data as well as transition temlerature energy are tused on data from the current BR3/ YAEC shift data.5 It should te notedd that of the five Yankee test program on surveillance capsule specimens at steel upper shelf values listed in Table 2 of Ref. 4,
!!R3. Rese data are shown in Fig. bl0." (pp. 3-5) only two are measurements. The others, denoted by "The measured upper shell energy of the Yankee pbte the approximation symbol (-),an estimates. (Dese material (bT) at a Duence associated with the year data were listed in NUREG 0569 without distinguish.
2020 is 5' It-lb. Therefore, using SRP 5.3.2 to ob-ing between experimental data and esumatesfI Data tain the transe (T L) direction, results in an upper for two of these rspecimens were used by Steele and Serpan32.33 o develop a graphical correlation between shelf er. cgy of 35 ft-lb.* [MTED 5-2, attached to t
SRP 5.3.2, prescribes a muldplying iactor of 0.65 for percent upper shelf drop and increase in Charpy estimating transverse direction upper shelf values from V notch 30-ft-lb temperature. Als plot, with the re-longitudinal direction upper shell values.) (pp. 5-26) maining data and estimates from Table 2 of Ref. 5 "The predicted upper shelf energy for weld metal is 40 added, is shown in Fig. C.4. Alr.o shown in Fig. C.4 ft-lb in the year 2020. It is based on an mitial upper is Hiser's6 esumate of percent shelf drop and ARTNDT shelf energy of 70 ft-lb and use of Reg. Guule 139, for the upper plate. The upper shelf drop, from Rev. 2, and BR3 chemistry to predict the drop in upper Table C.3, is 32.8%, and the ARTNDT value, from shelf energy. De validity of 40 ft-lb is also corrobo-Eq. (4), for f = 2.3,is 224'F. Ihscr's estimates are rated by data from the B&W Owners' Group presented consistent with the two Yankee surveillance data points at the May 24,1990, ACRS meenng in West Palm and the two additional estimates for Yankee matenal.
Beach, Florida, which showed that upper shelf energy The data for the ASTM correladon monitor material all for their Linde40 welds were above 40 ft lb for plot above the Yankee survedlance data. Odette's fluences out to and beyond 2 x 1019 n/cm2.-
estimate of ARTNDT or the upper plate was 245'F, f
which would shift the estimating point in Fig. C.4 Re NRC6 made calculadons for the individual reactor 23.F to the right, still presersing a consistent trerd vessel near beltline materials using Fig. 2 of Regula-with the other Yankee data.
tory Guide 1.99 and the same fluence values used to estimate ARTNDT. The NRC results are summarized Acklitional upper shelf drop data for ASTM correlation in Table C.3. It can be seen that the NRC 1990 esti-monitor material specimens w cre compiled by mates for plate are leu than the unlity's EOL estimate-NRL.34.35 Unitradiated upger shelf valucs ranged The NRC 1990 estimates for weld metal, assuming from 71 to 86 ft-lb in the longitudinal direct;on, and 0.35% copper, are close to the utility's EOL estimate.
45 to 46 ft-lb in the transverse direction tmdiated None of the foregoing upper shelf CVN estimates con-upper shelf values seemed to approach lower limits sidered through wall fluence attenuadon, although it is depending on irradiation temperature and spe;imen permitted by 10CFR50, Appendix G, Sect. V, to do so.
91ner, Jr.. A. L, NRC. penonal nanunicaatm to L G. Merkle, ORNL. 0cMer 11.1990 NUREG/CR 5799 C.8
orientation. For irradianon at $50'F, the lower limits would be for a vessel with a R/w ratio of 10. Thus,
'tppearrd to te 63 f t-lb for the longitudinal direction the utihty calculated adequate nmrgins on die up;rr and atout 41 ft-lb for die transver.c direcdon. For shell even diough u;qwshelf energies were esumated irradiation at temperatures less than 30TF, tte corre-to be as low as 40 ft -lb. This result was anticipated sponding lower limits were 44 h-lb and 18 ft-lb, in a previous NRC analysis)1 The NRC did not Clearly, irradiation tempe.rature atu! onentauon are resicw the utilitis upper shelf armlysis in detail. In important salubles. The estimated 1990 values for this evaluation, the udhty's calculations of apphed K!
Yankee plate m Table C.3 are all Setween the lo*cr due to preuure arnt thern.at loadmg invc not beca hmits for the corre ponding orientations given in checkul, but the choices of representadvc J-R curve:,
Ref. 35. Thus the estimaung praedurn m Regula-for base metal and wcld metal have been reviewed, tory Guide L99 apparendy dc, contain c x>uf.h ct nser-Additionally, the choices of upper.shell toughness sausta to justify applicauon to a vessel operating at values approprtate for use in ITS aralyses have been temgeratures totweca 500 and $$P F, examined.1his subject was tot discussed by Oc util-ity in Ref.1. Apparently, YAEC used ine ASME The revision of the fluentes for the Yankee vessel madnunn value of h.c n 200 bd as an upper-j given in Rc f.1 I required a recalculation of Oc upper shelf dnos. The pnmlure for estimating upper shelf shelf toughness, without questioning whether or not i
drups requires readmg and internotaung values from d'is value actually conesymds to the Charpy upper-Fig, 2 of Regulatory (;uide L99, which is a log log shelf energies estunated.
plot of percent shelf drop vs fluence, with copp r con-tent as a parametet,1his procedure is sornew hat Ic.
In Ref. I it was noted 0 at site effects have been dious, but since a9 the curves in the figure are straight observed in J.R curves measured by lliser aid Tctrell36 knct Oc pocedure is casily reduced to the apphrathm for transversely oriented (T-L) specunens of uninaJmted of simple equatums. The equanon of the upper toural A 302 grade B stccl. Additionally,as shown in Fig.
t utvc is C.5, such J.R curves can lose all shqc, approxhing constant values of J. Consequently, a pncedure was ACVNUX ) = 42,39;0.1502, (a) developed in Ref. I for estimating the J.R curves for irmdiated low ugqct-shelf A 302 grade il plate. The.
The equation of the lower curves for base metal is pnwedure umsists of devchping mean and mean 20 conciations tetwcen Charpy upp r. shelf impact energy and J c, as shown is Fig C.6, and then, based on ACVN(%) = (100 Cu 4 9)fn2m, (5) l Fig. C.5, assuming diat the up;cr bound constant level and the equation of the lower curves for wekt metal is of J for any tuse metal J.R curvc is 1.3 times Jje (sec.
pp. 3 6 and 3-7 of Ref.1). In Ref. 6, IUser developed ACVN(%) a (100 Cu + 14)fo 2W (6) mean and mean -2a correlations letweca Jo.1, corre-spmding to Aa = 0.10 in., and CVN, and these cor.
1he intersection of the lower curves with the upper relations are shown in Fig. C.6. The convergence of curve securs at conclation curves for Jo,3 and J1c for CVN approach-ing 15 ft -Ib in Fig. C.6 is further indication of the f = 142.39/Altl.55, (7) flattening out of kiw upper shelf J.R curves for A 302 grade B luse metal.
where A is Oc rnultiplying Iactor in Eqs. (5) and (6'.)
The recalcul.ned 1990 Charry V notch ulper shelf The conciathms in Fig. C.6 have the folk) wing impact energics are shown in Tabic C.4, AH changes equathms:
are reductions, but the only signihcant change from Table C 3 is for the upper asial weld,locause of the J c (mean) = 160 4 4.20 CVN, (H) large change in fluence.
C,4,2 Low Upper Shelf Energy Effects on Fracture Toughness Joj (mean) = 108 + 11,75 CVN, (10)
Low upper. shelf impact energy m reactor pressure mi vessel steels and wekls has the offat of lowering Oc Jg. (4a) = -162 4 11.75 CVN, (1I) margin between strength in Oc peser ec of flaws arvi i.
apphed huds. In Chap. 3 of Ref,1,0c utility per-where j is in in..lb/in) and CVN is in ft-lb.
fortned low upper. shelf analyses hv 1.cvels A,!!,and C hochng cornhthms acauding to procedures [mp ral by For estimating the J R curves for Linde 80 wcld metal.
the AS ME Scciion XI Working Group on Flaw de utility used a correlation, devehped by lliser 37 e l
Evaluation.' Itecause the rano of inside radius to wall twecn Oc parameters of a power law representathm of a thicknen (R/w) for the Yankee vessel is 6 83, includ' J.R curve, mg the Odckncu of the clathhng, the sucues due to pressure are roughly severuy percent of what they C.9 NURF.G/CR-5799 i
J = CIAa/kl,
(12) and 11551 Program studies. Repfatory (Juide 1.09, D
Rev. 2 allows the use of convrvadve estimates based a9d the Charpy ugyr-shelf impai energy. The coef, on geretic data (mean + stan&rd devianon). A copper ficients in d e correlation used in Ref. I are given in content of 0.35 wt% (mean of 0.29% plus standard Table C.2 of Ref. 37. Decause 14 (12)is a power devianon) was &termined for de Yankee welds, tmed law, tic estimated J R curve will not level off as did on the ll&W gerenc data.
the base tnetal J l curve shown in Fig C.S.
Pverticless there are J R curves for Linde 80 weld The Y AEC report auened that the plates have a rela-metal that display the tendency to flatten out. Such an tively coarse austenite grain site, which is likely, with example, corresponding to CVN = 39 It-lb, is shown a resultant it cremed sensiuvity to neutron radiadon and in Fig. C.7, which is hom Fig. C 50 in Ref. 37. The which mitigates the effects of the lower inadiation asymptotic upper level of Jmas for specimen temperature and nickel conter t. In surnmary, references WHA 121, from Fig. C.7, is about 600 in. lb/in.2 were cited which showed there are many confounding parameteciinvolved other than the size of the prior De Yankee Rowe esumate of Jman for A 302.!! plate austenite grains. The dislocanon structure, precipitate g
structure, etc., all contribute to the mobihty of defects in the microstructure and these are affected by the fab-Jmax = l 3 J e.
(13) r cadon pmcess, heat ucatment and chemistry. Tie l
cifetts of gram stic on embrittlement are, m other For purposes of estimating the upper shelf toughness w rds, very uncertain and lacking consensus.
tPpropriate for a ITS analysis values of Jmax (mean)
The effects of irradiation temperature are dependent on ran te converted to Kc by the equation rnany variables and, although there are specific tn-stances of contradiction, the bulk of the studies reported 2
K r1EJrnax(mean)/(19 )}l/2, (14) in the literature indicate higher erntrittlement with g
lower irmdiation temperature in the temperature and Applying Eys. (8) through (14) to the uppes-shelf fluence ranges applicable to the Yankee situahon. This l
Charpy impact crergies esumated by the utility and effect has teen shown for many steels including A 302 by NRC6 (prior to the fluence revision) gives de grade 11 and for Linde 80 welds. %c we of an empiri-values of J and K shown in Table C.5.
cal conclation such as one degree increase in shift for one degree decreme in irradiation temperature is cer-Figure C.8 shows the J.R curve for irradiated Linde 80 tainly not a scientifically satisfying appmach, but it is weld specimen W8A-121 imm Fig. C.7 compared to a pudent approach which is substantiated with a tody the J R curve for the unitradiated 6T A 302 grade B of rescarth. Based on the information cited, use of that specimen from Fig. C.5, plus the Jmax values Inn valuc seems reasonable ar.d not overly conservative for Table C.5 for A 302 grade B plate, based on the NRC the exposure conditions of the Yankee vessel.
1990 estimates of CVN. From Eq. (15), the value of corresponding to K = 200 ksd is Although there are observations to the contrary, the J
C evidence to support the YAEC claim of no nickel effect max 1213 in. lb/in.2 CIcarly, KC = 200 ksid is not for the lower plate is minimal based on de analyses an appropriate uppes-shelf toughness value for ITS of surveillance data from commerc,allight-water reac-i tors, nickel plays a prominent role in the estimates of analysis for the near teltline mal rials in the Yankee
- "I "
" '#I
- .W I' Rowe vessel, As indicated in Table C.5, values of Further more, the cited observadons of significant 141,126, and 113 ksik are more appmpriate for enhancement of embrittlement from ircreated nickel the welds, upper plate, and lower plate, respectively, make consideration of a nickel adjustment the prudent The sensitivity of P(FIE) to inclusion of lower values choice. Using different methods, lliser and Odette is discussed in Sect. D.4.2.
recomrnended the addition of 70 and 80"F, respectively, to the upper plate shift to account for the higher nickel C,5 Summary of Radiation i" * ' "" P ***-
l EffeetS
%c YAEC claim that the probable coarse grain size of the plates midgates the effects of lower irradiation tem-l There are many factors contributing to the uncertainties perature and higher nickel content is not substantiated
[
regarding the fracture toughness o the Yankee reactor with sufficient evidence. %c confounding ef fects of r
vessel. Among these are the relatively low operating so many variables dernands prudent choices in cases j
temperature (~500*F), only a small amount of surveil.
like this where information is so sparse ne YAEC lance data. effects of grain size and nickel content, and claims may tum out to be correct, but the information i
lack of chemical composition data.
available at this time is inahquate to allow their use.
He bases used by the NRC staff 'or shift estimates are De copper content of welds fatricated with copper-reasonable under the circumstances and not overly coated wire can te quite variable, as shown by B&W conservative.
NUREG/CR 5799 C.10 l
I l
u.
=.
Using avaitable drop-weight and Charpy impact data tm Ductitay Performance rf Vesul Stceland Yankee surveillance material and with the application Afatism Service I'luence Detert unedfreu of NRC Branch Technical Position MTED 5 2, the ini-Eyiosure During Cores ll, ///, and D', NRL tial RTNDT values for the Yaniec plates were esti-Report 6616. U.S. Nav al Rescanh Laboratory.
l' mated by the NRC ar I acceptal by YAEC. The NRC Washington. D.C., September 29, 1967.i' and YAEC estimatcs for the wclds were idendcal.
Although vast dif fererces initially existed tetween the 6,
A. L lliser, Jr., NRC, ~5ummary of Frt.zture i
YAEC and NRC staff e aimates of the RTNDT shifts Toughness Esttraates for Irradiated Y arikce Rowe J
for all the vessel materials, discussions between Y AFC Vessel Materials,* letter to C. Y Cheng, NRC, and NRC have led to convergence of the two organira.
with attachment, August 30,1990 8 tions' estimates, and indicate that the Ir13 screening criteria have leen ewmled.
7.
Branch Technical P(.sition - MT11115 2, t'
Frxture Toughness Requirements, pp. 5.3.2 13
%e NRC esdmates for upper-shelf energies were to 5.3.218 in U.S. Nuclear Regulatory Com.
somewhat kiwer than those of Y AEC cnd are imed on mission Standard Review Plan. NUREO 0800, those in Regulatory Guide 1.99, Rev. 2, w ith no con-Rev.1. U.S. Nuclear Regulatory Comrpinion,
-i sideration for the lower irradiation temperature of the Washington, D. C., July 1981.6
' Yankee venel because it was concluded by NRC that the Guide contains sufficient convrvatisms with 8
G. Papanic, Yankee Atomic Electric Company, respect to the specific conditions of Yankee. For rea.
letter to E. McKenna, U.S. Nuclear Regulatory 1ons cited in this report, however, the utihty calculated Cornminion, February 4,1987 8 adequate margins of stress on the upper shelf to com-penute for those dif ferrnces. 'Ihe arulyses at ORNL, 9
G. R. Odette, University of CahDrnia at Santa bowever, reganting frxture toughness and 1 R curves, Barbara, letter to A. Tatwda, NRC, July 30, indicates the utility's use of the ASME rnamiruum 1990, with attachment entitled, *19%I Shift Esumates for k Yankee Rowe VeneW value of K = 200 ksidn5 as an upper-slelf fracture e
toughness is too high for the low upper shelf materials 10.
T. E. Murley, NRC-NRR, letter to A. C.
i m the Yankee vessel-Kadak, Yanice Atcenic Electric Company, i
DaLet No. 54029 August 31,1990, with C.6 RefCrCilCCS attxhment entitled,"Safray Assessment of Yankee Rowe Vessel.*8 1.
Reactor Pressure Vessel Emluation Retwt, YAEC No.1735, Yankee Asimic Electsie 11, T. E. Murley, NRC-NRR, letter to R. F.
Company, !! alton, Massxhusetts, July 199Rd Fraley, ACRS, Odober 9,1990, with attach-ment: J. D. Itasettine, Yankee Atomic Ehetric 2.
K. IL Moore and A. S. lleller Bd W 177 FA Cornpany, letter to W. Runcil NRC-NRR, Reactoe Vesul Behline Weld Chemistry Study, Docket No. 59-29, Septen. et 28,1990.4 B AW.1799, liabcock & Wilcos, Lynchburg, Virginia, July 1983,a 12s. G. M. Gordon and it IL Klepfer, '"Ihc Engineering Significance cf Ferrite Grain Site 3.
C. Z. Serpan, Jr., II. E. Watson, J. R.
on the Radiation Sensitivity of Pressure vessel llawthorne, and L E. Steele, Tunice Reactor Steels l' pp. 484,6 in Egeris of Radiation on Pressure VesselSurveillance: Emluation of Structural Aletals, ACTM STP 426, American Specimens Eyesed During the Second Core, Socicly for Testing and Materials,1967f
- NRL Report 6179. U.S. Naval Research Laloratory, Washington, D.C., November 24, 13.
R.-W. Nichols and D. R. liarries, *llrittle 1964.6
- Fracture and Irradiation Effccts in Ferritic j
Pressure Vessel Steels," p.162 in Radiation 4.
L E. Steele et al., Irradiation Iflects on Reactor iflects on Afetals and Neutron Dosimetry, Structural Afaterials. Quarterly I'rogre,ss ASTM STP 341, A merican St-icty for Testing Report-1 Afay-J/ July 1966, NRL Memor.
and Matcrials, Jamory 19637 andum Repost 1719 U.S. Naval Research Laboratory, Washington, D,C., August 15, 14.
J. R. llaw thorne, " Demonstration of Improved 19666 Radiation Embrittlement Resistarice of A533-B Steel Through Control of Selected Residual 5.
C. Z. Serp.m, Jr., and J.'R. llawthome, ranAce Elements," p. 96 in ASTAf4TP 48J. American Reactor Pressure VesselSurwillance: Notch Society for Testing aml h1aterials,197l?
I
=
dAvadaMe inun NRC f%bhc Ihument Ran for a tee-
- AvaileNe frurn NRCINNic thwnent Ram for a fee.
0AveibMe from Natualinhnicnt Inrormeuon Servke,
~
6Availahte inun Nammi Tnhnicat informatum Servwe,
' AvadaNe Inn p,hhc and spmal tubrutal htwarnes C.1I NUREOCR4799 l
-.___._,.____-_.-_.._.,.~._.,__.,.-,,u.-__
~. - _, _ _.. _. _
o
-1:
15.
R. R. Ilodens and 11. l.. Wotton, "I he Ellat -
RPV Sted Embnttkment on Irradiation Tem.
of Fast Neuuon irradiation on the Meetuuncal perature athl Neutravi lingnisure,* in l'ror red 2nr3
' I nqtrties of Stunc Qw:skik'd arkl Tenipered of the 4th ASIhi l URAIUhl $ mposium on 3
Stech, p.142 in AX/ Af Str 4M, American
- cartor />vsimetry. Gaithetsburg, Md., IVM2p Swiety for Testing and Materials,1971 a 25.
G. R. Odette and G. E.1.ucas, *lrradiation Em.
16.
A. D. Amayev et al.,
- Radiation Stabahty of hrallement of Reactor Picuure Vessel Stecle VVER.l(XO Reactor Venct Mutenak,* Report Mechanisms, Modeh, and Data Correlations,"
-l of Working Gniup 3 of the U.S.S R ALS. Jomt pp. 2tM I in Radation End>riulement of l
Cterdinating Committee on Civilun Nuclear NurIrar Reactor Presswr Vr.urtStrets: An l
Reuetor St.fety (JCCCNRS), Moscow, Internatio.nal Review IScrond VolumeJ. ASTM-U.S.S.R., pp. 25 29 June 1990
S il' 909, L. E. Steele, Ed., American Swiety for Testing and Materiah, Philadelphia,19M6.8 17.
L. P. Trudeau, Radiarmn EITects on kcar tor Structural Afascrials, AEC Monoe.taph Series, 26.
J. R. lias thorne, Steel Impurity Element Amenean Soekety fof Metals, Roun an and Ifccts on Posnrradmtion i'roperties Recovery Littlefield, New York,19M.a by Annraling, NUREGNR 5388, August 1989.a
! 8.
L. E. Steele, Neutoon irradiation End>rittlement of Reactor Picssure Vessel Strels, p.123. Tech.
27.
J. R. llawthorne, " Status of Know Ldge of i
nical Retorts Series No.163, International Radiation limbrittlement in U.S. Reactor Pres.
Atomic Errrgy Agency, Vienna,1975 a sure Vessel Stech," in pp.100-15 in Radiatwn Endurittlernent and Survedlance cf Nuctrar Reau IV.
F. W. Stallnmn, Curve Fining and Uncertainly for Prr.uwr Vr.urls: An Internatiomd Study.
Analpis of Charry impact Data NUR*iGNR-ASTM-S11' 819, L. ii. Steele, Ed., American 2408, Oak Ridge National 14 oratory, January Society for Testing and Mate:iah 19g3 a 1982/
28.
T. J. Williams, P R. Hutch, C. A. English, and 20.
G. R. Odette and G. E. Lucas, Irradiation P. II. N. de la cour Ray, *The Eflat of Irradia.
Enduritdement of LWR Pressure Vessel Strels, tion Dose Rate and Tempercure, and Copper and NP-6114. Electric Pos er Research Instuute, Nackel Content, on the Irnaliation Shift of low
' January 1989.4 Alloy Stect Submerged Are Wckh," in Prorredi-ngs of the 7hird InternationalSyraposium vn
?1, A. L, inse, *An Evaluation of Linde 80 Sub.
the Environmental Degradation of Alattrials in merged Arc Weld Metal Charpy Data Inahated Nur/rar Power Systr.ra - Water Reactors, G.
in the llSST Program " AS7Af 57P IIM6, Vol.
J. Theus and J. R. Weeks, Eds., TMS 2, American Society for Testing and Materials, AIMFJANS/NACE,1988.a 1990.a 29 C. Maricchiolo, P. P. Milella, and A. Pini, 22.
R. K. Nanstad and R. G. Berggien, yTrcts of
- Pn diction of Reference Transition Temperature Irradiation on Low Upper-Shelf Weldt, ficacy.
Inctrase Due to Neutam Irradiation Exposure,"
Section Steriferadiation Program,lriadiation pp. 96-10$ in Radiation Embrullement of Series 2 and 3.NUREGXR.$606(ORN1JTM.
Nuclear Reactor Pressure VessetSteels: An 11804), July 1991.c Internatiorud Review (Second Volunw), ASTM-STP909, L. E. Stecle Ed., American Society 23.
T. J. Williams et al, "The influence of Copper, for Testing and Materials, Philadelphia,1986.a Nickel and Irradiation Temperature on the Irra-diatam Shift of 1.ow Alloy Steels," pp. 393-99 30.
S. It Fisher and J. T. Iluswed *A hici for m Proceedmgs of the SecondInternational Sym-PWR Pressure Vcuel Embrittlement: Int.J.
posiwn on Environmer tal Degrafation of h1 ate, Pressure Yesseh and Piping 27, pp.91-135 rials in Nuclear Power Splems - Water Rear-(l987).a tors. TMS AIMF/ANS/NACE, Monterey, Calif., Septembet 1985.a 31.
K,0. lloge. Evaluation ofI:ir Integrity of SL'P Reactor Vessels, NUREG 0569, U.S. Nuclear 24 J. Ahlf, D Bellmann, F. J. Schmitt, and W.
Regulatory Commission, Washington, D.C.,
Sgulthoff,' investigation on the Dependence of December 1979 6 d As tilable inun NRC Pubbc Ihument Rwn for a fee, 6Avadable fran Natkmal Techrual Informancri Service
'Avadable frorn pbhc and spenal technic.al htranet-JAvadable for pun;haec fnen organisatusi stunonng phhcation
" Avadable frwn NRC l'ubbc !)rumeni Rmen for a fee.
caed. an.iior fnen suuvri and/or rmpeenu (documented leurrs).
bAvadable fran Naminal Tuhmt al Information Service.
N1.'REGER-5799 C.12 o
-,-,-e---<,-.+
, -w w-
-.m
,.-w
-n--,.--..--m
,m
...-n-r
- -, - - - - - ~ - ~
-,.~-c,--.
---,e.a.,
+ne-
M.
L E. Steele and C. "L Serpan, Proceduresfor 35.
J. R. llawthorne
- Trends in Charpy V Shelf laterpreting Ihe StructuralImlications of Cnctgy lkgradathm and V\\cld Surogth inctrase Radiation Dansage Surwillance Results on of Nwutron-Embsittled Prct.sute Vessel Stects,*
Nuclear Pressure Yessels, NRL keport 7358, Nuclear Engineering and Design. IJ, pp.
U.S. Naval Researc h latwatory, Washington, 427 446 (1970).b D.C., Decemter 30,1971.8 36.
A. L. I!iser end J.11. Terrell. Site Effects on
- 33. L E. Steele Neutron Irradiation Embrittlernent 1 R Curvesfor A 302 8 Plate, NUREGICR-ofReactor Pressure Vessel Steels, Intcmational
$265 (MEA-2320), Materials Engineering Atomic Erstgy Agency. Technical Report Series Associates. Lanharn, Md., January 19895 h(
No.163, Vienna,1975.6 37.
R. M. Gamble. A. Zahoor A. Iliser,11. A.
B 34.
J, R. Ilawthome, Trends in Cha~py V Shey Ernst, and E. T. Pollitz Evaluation of Upper-i Encegy Degradation and lictd Strength increase SheV Toughness Requirernentsfor Practor cfNeutron EmbrittledPerssure VesselSteels.
Pressure Wssels EPRI NP-6790 SL Vol. 2 NRL Report 7011. U.S. Naval Research Draf t, Electric Power Research Institute, Palo Laboratory, Washington, D.C., December 22, Alto, Calif., October 1989.d 1969.a 6Available inwn Nadimal lechmcal Informauan Service.
~
'Available (nwn pbbc and spcual tochmcal htwartes.
a Avadable fne NRC l'obhc !)xim.ent Rasu for a fee.
JAvailabic for purthace from organisanan spanionns pbbrauon IASatlabic from Nanonal Technical Informanon Servite.
tited, and/or frurn audors and/or recipients (doeurnented leuers).
C.13 NUREG/CR 5799
2C:n mO 73?
'G E
Table C.1 Licensec and staff estimates of RTNDT or the YNPS f
beltline materials in 1990, pner to Septemtw 19')0 1990 Ihurnhant increase in stfererxx Reference YNPS onginal sta%
temperature resulting temperature RTNDT teltiine peak fluences temper =ure frtxn irradiatini m 199tf ma:enal (x 1019 n/cm2)
(T:)
m r'F)
Staff Liansee Staff lernwe Staff Licensce estimate estimate estimate estimate estimate estimate Upper plate 2.3 30 10 245 180 275 IW
.o Lower plaic 2.05 30 10 325t 173 355t i33 Axial welds **
0.38 10 10 216 131 226 141 Ortumferential 20.5 10 10 320-360 219 330 370 229 wekl**
- Does m4 mdade
- margin' tenn.
- NRC used Cu 35%. Nr47% YAEC used Cu-418 %, N,47%
19 2
N T ased m a noence of 23 x 10 nam rather than t%e careca value of 2 05 = IO B
ee e
t s a 5
5 0 8
3 w
nm 3
5 0
9 3
ei 2 3 3 2
3 t
t N
s eT Uc c R *0 nee9m 9
r r eu1 f
t ea n Rri 8
p L *te 7
m Na 6 6 U 8
3 9
6 5
(
Rm 2 3 3 2
3 e
3 t
Oits 3
0 e
99 1
n i
ls e
a t
ir g
a 5
5 0 3
8 e
e n t
ci*
m 5 2 9 3
2 a
nt n b i 2 3 2 2
3 l
t m
e uo s
- e asi I
t e
f e a n
eri cdm i
r l
n r a lt e
i r
e e r
s e
i b
n s
m S e yo L *t Pi e
e c
mf Na 6 6 0 3
3 r
r NL c
5 2 9 8
2 Rm 2 3 2 2
3 n
I e Y m Oi t
t e o s
h r
e f
tr s o n f oi Tsi Dv Ne mi e
r T0 ta R9 m
0 0 0 0
0 9
3 3 1
1 1
f t
t o1 s
n t
r e
e eb m at lu r
I s e u
u A a m A
rm a
s r
s e
me kf e
a re p
i t
t p
s e ne m n
d eS U
r L *e e
a t
L n f
No Nta 0 0 0 0
0 m
r a
Rd Rm 1
3 1
1 1
c Os Oi C
e ts R
d a e
N nb y
a b
e e
s n
e
)
t c
s2 s
i e
L e m c
h 2
0 d n /c n
1 4
1 u
e e n 9 s u 6 3 2 2
3 C
9dfl9 2 2 1
2 x.
1 1
c 1
ek 0 i
i r
b a
a e1 T'
p x-t m
t-7
(
e e 0 t
n =
"m o F sd r e a s a %
mb5
~
s 3 l
'e e
t u
t 0
a d
a el n
l m=
a S ni l
c e
Pi r e e la a
e l
xt x
e inti C t
i r
s t
l e
a a i
Nl e
t t
p p as at f
x a
a t
eL Ye d
rd mk b m i
Nd r
r rl el ue e e ee e
eR w p w pw ww c w D *O aa r
p o
pU oU L
C
- t i
l
[*
hxm&aM 3 i l'
,Il!:I!l li!9l?
n' i i '
,ii!4 fjI l
ia a
Table C.3 NRC estimates of Charpy upper shelf energies for the YNPS beldine materials in 1990, prior to September 1990 Original Initial Original ihrnce energy Drop 1990 energy Material (x 1019 n/cm2)
(ft.lb)
(4)
(ft lb)
Upper plate i
L 2.3 76 32.8 51.1 T
2.3 49.4 32.8 33.2 1.ower plate L
2.05 76 34.0
-50.7 T
2.05 49.4 34.0 32.6 Upper axial weld 0.35 Cu 0.38 70.2 37.0 44.2 0.18 Cu 0.38 70.2 25.5
.52.3 Circumferential weld r
0.35 Cu 2.05 70.2 47.0 37.2 O.18 Cu 2 05 70.2 37.6 43.8 L
~
L i
NUREG/CR 5799 C.16
._~_ - ---
Table C.4 ORNL estimates of Charpy unri shelf energies for the YNPS teltline materials in 19/0, tused on Septemler 1990 tevisions fron. Licensee Revised initial Mcvised fluence energy Drop 1990 energy Material (x 1019 n/cm?)
(ft !b)
(%)
(fi.lb)
Upper plate L
2.6 76 33.9
$0.2 T
2.6 49.4 33.9 32.7 l
1/rwer pbte L
2.31 76 35.3 49.2 T
2.31' 49.4
'a5,3 32.0 Upper axial wcld 0.35 Cu.
1.24 70.2 43.8 39.5 0.18 Cu 1.24 70.2 33.7 46.3 Lower a.xial weld 4
0.35 Cu 1.20 70.2 43.6 39.6 0.?E Cu 1.20 70.2 33.4 46.8 Circumferential weld 0.35 Cu 2.31 70.2 48.1 36.5.
0.18 Cu 2.31 70.2 39.1 42.8 I
i.
t I
a C.17 NUREO/CR-5799
=
4 i
)
i.
. l I
l z.
h tn O
l'
?3
- t
- o -
u Table C.5 Swrunary of Charpy upper-shelf energws and fractwe toughnesws w8 j..
for the YNPS behlme materials in 1990, prior to September l'M).
\\
I Charpy V-notch Mean igper. shelf K e for ITS analysis (ft Ib)
-2o R curve l
'YR NRC YR NRC ORNL YR NRC ORNL (2020)
(1990)
(2020)
(1990)
CO20)
(1W9 l_inde 80 we!d 4
40 44 52(Axial)
MEA arreladon for S;rximen %7A.121 Jmax = 600
}
I -
CVN = 40 ft4b CVN = 39 in.-:b/m2 l
37 44 (Circum-Kmai =200 Kmax = 14I i
ferential) ksi Yirt ksaVin.
i A 302 riate flent :2Ja D J e = 245 h-Rim 2 jo.1 = 289 in.Wm.2 3,,,. 43 g 57 31 (Uner)
I 4
in.-Ib(n2
{
b' SO(lower)
Jmad = 320 in Wei.2 Km.,=200 Kmax = 126 l
"8 ksi Vin.
EsiYin.
t A Xt2 niate !narsverse) i J.1 = 149 in -thin 2 Imat=383 1
J c = 150 in Wm.2 35 33 (Upper) 0 i
in. W m.2
}
33 (lower)
Jmad = 195 in.-Ib/n2 Kmsm= 200 Kman = ll3 i
ksi4:n.
ksi inet i
f 5
}
I f
e h
r r
i -
t
R P V W EL D & P L A.T E L_0_C A TlO NS_
D'if'N~ TLJ 7'
3;; 3
/\\
~
r s
s
/-
l 2' 6'
\\ /
_I 9 5/3'
--4 L.
/ l l
u 4
I l
g, 3 4' 10'
[j~
ucttte surtt covas t
i
..... 3 7 7/B'
- ti
/!
, \\,
,i l, 33, l
7'10'
/l creta surLL
\\/
I counst ycen 73
/\\
l i !
./
,e i
!/
\\1 c.........j I
7 7/8' 7' 'O' towra sHELL counsa j
i
\\l l
_]
\\\\
I g
3 7/8' 1
to' 5' Fig. C.I.
Schematic drawing showing krations of plates and welds in the Yankee rea:n>r w>scl. Scurce:.
. Reactor Prc.tsure Ve.sselEvaluation Report, YAEC No.1735, YanLee Atomic Elecuic Company, Etohon, Massachusetts, July 1990.
C.19 NUREG/CR 5799
Y.. l t
.i
)
i!
120 F.,
Yankee Rowe Plate oh-
+
i'
~ 100 -
. Surveillance Plate -
Vessel Piate_
3_
= - Westinghouse -
MillTests (B&W)
+
j.
g
+-NRL 80 -
v
- 4 g
3 CD u
so -
+
C
?
- - 50 ft-lb
'/
- =,.
o i
b 40 -
t e) m I
c
- +/
?
Q
?!
0 20 +4
[
s =.
l o
E /
o l
4
-200 o:
200 400 1
l.
Test Temperature ( F) 0 Fig. C.2.
Charpy-V impxt data for the Yankee Rowe epper plate, fron: the xtual vessel pb:.: and from t
the surveilbnce plate. "Ihc illustrated curve is an approrimate lower tumd to the surveillance pbte data. Source:
i I!iser.Jr., A. L, NRC. " Summary of Fracture Toughness EMimates for Irradiated Yankee Rowe Vewel MatercIs." !ct
[
i ter to C. Y. Cheng. NRC, with attachment, August 30,1990.
t i
h i
i s
4
-e n
w
-u+
m ev
+
w c
-n
800 i
r i
i 4/<
~O v"
A00
/ ' A.n' t
_ 200 F * ~N
/
1'O 0 BR3 + 1F/F 100 -
e BR3 x Ct o
D YRP-1x$1 nom.
~
50 A YRP 2x4tnam.
2 25 O.1 1
10 10C 2
$t(10'9 n/cm, E>1 MeV)
Fig. C.3.
Yankm Rowe Upper Plate (YRP) and tuijusted BR3 shift data vs fluence. Source: Hiser, Jr.,
A. L, NRC, ' Summary of Fracture Toughness Estimates fcr Imdiated Yankee Rowe Vessel Materials," letter to C. Y.
Cheng.NRC, with attachment, August 30,1990.
C.21 NUREG/CR 5799
?,
?
9 a
e Ny o
r 4
y q y Dt
[y r 3 y'3 o
g o
, u.
3 2% 0 44
'e O
-A -
.!1 8
- F..,f,#
y.
C3
-8 1 j.31 O
i ;11
' M Li
-3 t
.n j i :-
0 4
}l.,53 es
@5
-s i fil 0
4 g
g g;pg O
O
-3 i l iP 0
O O
0 q
70 O
O O
Ed,L g
-8r ve n
E 4
e JS y
4E i
d.
G
.!,f e
-s e
l ' Ji 8
a 3
s O
g
~"
o 2je.
J t,g
< a y~,3 i, 5
1-4 -o.
' j S li3 f Tv 8
- Si a ?<].p a g a a
.d -
hL J
<~
S a
E.s initEbkn e
n*k'6 2
u t
rit.oss 4 :;
004
=
$ bl p
i
-a e
(%l W 4X@16 d1 >$ '3 1
NUREG/CR 5799 C.22
I i
i l
1 1400 i
~
f
. I 1200 1000 N<
800 c
r C4 h.
W soo n-400 200
]
0 O.00 0.05 0.10 0.15 0.20 0.25 Crack Extension ( 6 a), Jnch hs. C$
14nwgral maring resw.we curve fcr a TL<nened t,T-CT gmmen of uninadiau.d AM2 B sect t
l-(Refs, I and 36).
l l
l l
C.23 NUREG/CR 5799 y
es e
a Ne
-g
,r<.
e--g----
yp.*
e y
Ts e er aw+.M*-m7 m--=-
ap mwta w
wewe-g-
r+e.e+---
mt.-m-ewe 4
--r*a w as w
e-s
-t e+ wm e--
e
cc N
+
6 3
r y u
I cf b
3 3
n aeg O
s
\\
\\
15s
-U T
O E
7 L
o
\\
"rF 2.
%8 N
E
~"
_,M.C O
\\
o 8u t
s'
=
t.'
Z$3 g
N
'k
\\
h -
O N
N i
\\
r-ng
- - -\\
_\\,m.
jjj A
y O
\\-
o x
i x
s k'* j$
O
\\
\\
h W
o
\\'
,> 19 *.
a N
o o
g
- n.
o s
x
'N T
v g30 N
H \\(H 3
y o
O.
\\
i
.Eg
-~
~
\\
n
\\
8 ~ lj R..
N
,te;
\\
m5
.A
- g..
x o
1 x
..g y q I
N an P O
u.$
i O
O O
O O
O O
O O
O O
k
- y m.}.y u
O o
O O
O O
O
-n
-o w - O
. O, -
a t
w n
n e
142-1 t
gv !/ql ul
- Clr L'
u NUREG/CR-5799'-
C.24 i
i i,-,.,r
-,,v~,-.,--,-..-vi v.,-,ru v~
w,i--*sw,-,---,-w,w.m.
,-,-~.--w.r,-v-,,--.w
,e-,w.---,
,ci,.
,,,ev~,,--
1
DIlLTA a (th) 0 0.Uk V,lu U.lb U.20 U.'S 150 i
~ ';
BD0 t.inde 80 Held (WSR) 0.51-CI, itrad6a,J g ?88'C 'gef p,,_.
}
j. /, c : /], -~
,Seremtw N8 4 -! /.i /..
(g13'p
~
IY 100
/8 l / hj$
^
/
288'C Only r7 i
j, 'A N.
[
lh
?
di i
$ 75
-i I
l 400
- /y q
~
22
=
,x II 50 f
l 4
_ u ff l
Large Delta a 20 Small Delta a Il l
~
,h
,I
- i.,
,I 0
1.0 2.0 3.0 4.8 5.0 6.0 7,0
[ELTH a beeil Dg. C.7.
hnwgral nr.ng rewswc da he a0 51 ct gi smwn.d grained tak W):utimerged an weld mnal (Ref. 37)
CIS NUREG/CR-5799
1lI I
s E
S E-a
- A s.
n'#'s ig 4
n F
i v
s s
1 n
f s
O 3
ev g
E r
6 4 3 5
uc 2
R.
~
~
~
0 Je ht a
s i
w
)
3 0
1
(
4 i,
.i4i i2
[
- I Ij 2
d A
0 na
)
8 h
(
c a
A n
4 i
4
).
E n
A 5
a w
t 1
r ii! !~
i'
!ji' f
3 A
0 de 4
(
n
!xu o
c A
l i
a s
c n
s e
e a
0 t
u l
x a
1 E
v
.t 1{I*
0 x
+
A k.
a m
J f
C o
mu a
5 n
i A,
0 pa 0
mo C
A S
A X.
0 C
L k
_A:4 e
0 g7 0
0 0
0 0
0 0
0 FC i
0 0
C 0
0 0
0 4
2 0
8 6
4 2
d 1
1 1
na 5 -C 7 5.2 c~
2 5
hN n DNG 8 9o@
.a l
l l
Appendix' D il ORNL Review of YAEC No.1735 Probabilistic Fracture Mechanics i
Contents
. h LM L
i y
D. I Intrat uction......................................
D3
,T D3 D.2 Scope.............
t DJ Comparison e VISA-Il and OCA P....................................
DJ r
D3.1 Comparison of Deterministic Methodologies....
D3 D3.2 Comparison of Probabilistic Methodologies......
DA D33 Comparison of VISA.Il(Baseline version) and OCA-P Solutions DA 4
DJ,4 Details of ORNL Suggested Corrections to VIS A.Il Probabilutic Code.......
D.5 D.4 OCA.P Applied to Yankee Rowe........
D.6 D.4.I Deterministic Analysis................................
D.6
' D.4.2 Probabilistic Analysis....
D.7 D.5 Flaw Density Considerations..........
D.7 D.6 Discussion of Specific Features in the YAEC Analysis...............
D.8 D.6.1 Number of Subregions Considered in the Beltline Region D.8 D.6.2 Flaw Density -
D.8 D.63 Flaw Configuration D.8
- D.6.4 - VISA-II Code Errors......
D.9 References D.9 4
'(
)
D.1 NUREG/CR-5799
I Appendix D ORNL Review of YAEC No.1735 Probabilistic Fracture Mechanics T. L Dickson R. D. Cheverton D.1 InirodUClion (ITS) loading 'a e t des perform a thermal analysis.
linear-clastic sm uysis, ar.d a linear clasde frac-ture-mechanics U. M) analysis; however, the two Nucler.r Regulatory Guide 1.154 (Ref.1) specifies that cod:s use different analytical methods. Knon funda-OCA-P (developed at Oak Ridge National tr>boratory) mental differences utilized in the d rministic aspects 2
and VIS A-Il3 (developed by the U.S. NRC and Pacific
" # #U #
"## "S Northwest laboratories) are acceptable codes for per-fmning the probabibstic fracture-mechanics analysis I) nermal analysis:
portion of the plant specific safety analysis that may be OCA P uses a general one-dimensional finite-performed for any nuclear plant that desires to operate element mettal. VISA-Il uses a closed-form tryond the pressunzed-thennal-shock (ITS) screening solution based on a slab-geometry fannulation.
critena.4 Yaakee Atomic Electric Company (Y AEC) has performed such an analysis for Yankee Rowe using OCA-P allows a point-by-point desenption of the a modified vctsion of VIS A II. His report reviews the thermal hydraulic boundary conditions, t.c., the YAEC analy sis and inchides an N.edependent" ORNL downcomer coolant tempemture -time history, analysis. He review is supplemented by an addiuonal which is input into the thermal analysis. VIS A ll study by Simonen (Ref. 5 and Appendix E).
fits a polynomial or an exponential (user selected) to five user-input data (Ume, temperature) points D.2 Scope used to describe the downcomer cmlant tempera-ture-time history. As a result, VIS A-Il has a ne original scope of work for the ORNL review of more limited, though usually adequate, capability the proPilistic frxture-mechanies analyses of Yankee for accurately roodehng the thermal-hydrauhc Rowe (a_ efined in the August 9,1990, initial boundary condidons.
Yankee review meeting) was to perform a compreten-sive comparison of the " baseline" VISA-Il and the OCA-P allows for accurate time-dependent rrnic!-
OCA-P probabilistic Imeture mechanics codes. The ing of the convective heat trar'sfer coefficient.
original scope was later expanded to include an inde-V:S A-Il is limited to a single value for a given pendent probabilistic fracture mechanics analysis of the analysis.
Yankee Rowe vessel when subjected to the YAEC-defired small-break kiss-of-coolant FTS imnsient
- 2) Stress analysis:
ISBLOCA 7), using the OCA P code. The iesults of OCA-P uses a general one-dimensional finite-these efforts are discussed in Sect 3 and 4, respec-element method. VISA ll uses a closed form,
- tively, one-dimensional solution technique.
De scope also included a discussion of the flaw-den-OCA P allows a point-by-point description of the sity treatment in the Y AEC and ORNL analynes (Sect.
pressure-time loading history, which is input to D.5) and a brief discussion of some other specific fea-the stress analysis. VISA-H fits a polynomial or 4
tures h the YAEC analysis (Sect. D.6).
an exponential (user selected) to five user-input data (time, pressure) points used to describe the D..
Comparison of VISA-II and pressure-time history. As a resul:, VISA ll has a nue Umited c pabihty fm a curately menng O C' A-P pressure-time histories. This could be significant in cases involving complex pressure-time histor.
D.3.1 Comparison of ies such as those corresponding to transients in-DeiermiDiSIiC volving repressurization.
Methodologies
- 3) Fracture. mechanics analysis:
Both codes perform a linear-clastic fracture-VIS A-Il and OCA-P are capable c' performing a mechanics (LEFM) analysis using stress-intensity deterministic fracture-mechanics analysis of a reactor factor (KI) influence coefficients and superposition pressure vessel subjected to pressurized-thermal-shock techniques to calculate K values. However, I
D.3 NUREG/CR-5799
VIS A-Il uses a 4th. order polynomial to fit the stress distribution, and Kj innuence coef0cients VISA ll and OCA P foth suthastically simubte the are calcubted for each of the terms. OCA P uses same paramet rs: fast neutron fluence at the inner sur-a relatively brge number of influenec coef0cients face of the vessel, RTNDTO, ARTNDT. K c. K a. the l
to obtain a more accurate value of KI. Esen so, concentradons dcopger and nickel, and the size of the for rnost cases, the diffesence in Kg is small, assumed Ibw.
It should be noted that the influence coef0cients in VIS A-fi uses NRC-derived mean fracture toughness the baseline version of VISA !! apply speci0cally curves, w hereas OCA P allows the user the option of to a vessel that has a rado of vessel radius to wall using the NRC-dnived curves or a set derived by thickness (R/w ratio) of 10. The R/w ratio for ORNL. The latter eat was udliicd in the Integrated-Yankee is -7, In applying OCA-P to the Yankee Pressurized Thermal-Shock (IPTS) studies.7 Rowe vessel, influence coefficients were derived (using a finite-element technique) for the specific To our knowledge, no catensive comparison of the reactor vessel geometry, details of the probabilistic methodologies utilized by OCA P and VISA II had been performed prior to this D.3.2 Comparison 0[
eff n. Pers onelat Pacific Northwest Laluratories Probabilistic performed a comparison of the conditional probabili-ties of failure calculated by visa-li and oCA-P in Meihodologies 1984.8 The conclusion at that time was that v!SA-Il appeared to calcubte conditiored probabibties of failure Esumation of the risk of vessel failure is carried ou; lower than those calculated by OCA P by approxi.
by means of probabilistic methods to account for the mately a factor of 6. It was concluded at that time that uncertainties in a number of critical parameters. The this difference was due to the fact that OCA-P included basic philosophical approaches used in VIS A Il and the stresses in the cladding w hereas VIS A !! did not OCA P are essentially identical. The models are based (the present version of VISA Il does have the capabil-on Monte Carlo techniques; that is, many vesscis are ity to include stresses in the cbdding).
simubted, and each is subjected to a deterministic frac-t.re mechanics analysis to detennine whether the ves-D.3.3 Comparison uf VISA-II Sd *"II""'
(llaseline Version) and Each vessel is defined by randomly selected values d 0CA-P several parameters that are judged to have significant uncertainties associated with them, and a detcrministic The purpose of comparing OCA P and the basetine analysis is performed for each vessel to determine if it version of VISA-Il was to examine their validity, and will fail when subjected to a specific PTS transient.
to facilitate this effon, the VIS A-!! code was kstalled In each deterministic analysis, it is assumed that each at ORNL. The Rancho-Seco PTS transient (Fig. D.1) region of the vessel being analyzed contains one Daw.
and a vessel radius to-wall thickness ratio (R/w) of 10 He calculated probability of failure for a specific ves-were chosen for the comparison (this value is consis-sel region, based on one flaw in the region and referred tent with the stress-inten+ity factor influence coeffi-to as the unadjusted value, is equal to the number of chnts utilized by the baseline version of VISA-ll); the vessels that fait divided by the total number of vessels initial downcomer-water and vessel temperatures were siruulated T5.e probability of failure based on the inadvertently assumed to be 590 instead of 550 F(for
" actual
- nr.mber of fbws in the region and referred to the purpose of comparing the solutions, this is of no as the aQusted valve,is obtained by mukiplying the significance); the potential benefits of warm-prettess-I unadit.ited probability of failure by the number of ing were not included in the analyses; and the preser-flaws that are assumed to exist in that region, The vice-inspection option in the flaw-size distribution total probability is obtained by adding the adjusted function was not included, probabilities for each of the regions. If the total num-ber of flaws in critical regions of the vessel is not too D.3.3.1 Deterrninistic solutions much grean.: than tmity (limiting value depends on the value of the probability), double counting is not a OCA P thermal-response and stress-analysis solutions nroblem; otherwise, a correction must be made for were previously successfully vahdated against the gen-hble counting (more than one flaw resulting in fail-erabpurpose, finite-elernent thermal and stress analysis
,se of the vessel).
codes ADINA Tand ADfNA,respectively. VISA Il thermal-response and stress analysis solutions were These failure probabilities are referred to as conditional previously successfully validated against the general-probabilities of failure [P(FIE)] because the PTS tran-purpose finite-element ANSYS code. Figures D.2, sient (event) is assumed to occur; the term " failure" D.3, and DA show the comparisons of de thermal-refers to full penetration of the vessel wall by the response solutions, the hoop, tress solutions, and the pmpagatmg flaw.
NUREG/CR-5799 D.4
Q stress-intensity factor solutions (for longitudinal infi-As can he concluded from the arme tabulation, the
. nite-length flaws), *espectively, for the Rancho-Seco ORNL specified VISA ll code mahfications dramati-transient. In each of the three cases, the solutions of cally decreased the number of stable crack arrests pre.
VISA-ll and OCAf agree reasonably well, although dicted by VIS A II, and this significantly increased the the VISA Il K1 values do not reflect the expected de-number of failures and thus P(FIE). P(FIE) calculated crease in K1 for deep Lws9 under the specific pressure by OCA P and the modified VIS A41 are nearly identi-and thermal loading conditions (Fig. D.4) (this latter cc.l; hosvever, the mochfied VISA Il predicted a smaller discrepancy is not a factor for most cases analyzed),
number of initiatians and arrests than OCA-P, which mdicates there is still some difference in the OCA P D.3.3.2 Probabilistic solutions and ViS A-Il methodologies. It is suspected that a con-tributing factor to the difference is the method used to After demonstrating that the basic engineering mechan, implement the flaw size distribution function in the ics (heat transfer, stress analysis, and fracture mechan, two codes. The VISA IIand OCA-P analyses define ics analysis) solutions of VISA Il and OCA-P appeared nine possibic initial flaw depths distnbuted according to the htarshall distribution function,10 which is used to be in reasonably good agreement, the probabilistic solutions of VISA-Il and OCA P were compared.
for the YAEC and ORNL analyses. De nine depths Initial attempts to achieve reasonable agreement were utilized by OCA P ranged from 0.08 to 2.08 in.,
not successful. OCA-P was predicting values of whereas the nine depths utiliaed by VISA II ranged P(FIE) higher than those for VIS A II by a factor of ~8.
from 0.125 to 3.5 in. Derefore, the initial crack fhis is consistent with the results observed i. the depth mesh used by VIS A ll is more heavily biased 1984 comparison of the VIS A Il and OCA-P prob,.
toward deeper flaws. It is expected that this would bilistic solsons 8 result in fewer initiations because of the lower values of ART ITf and lower thermal stresses associated with N
OCA-P was enhanced to print out a more detailed deeper flaws, it is necessary to determine the proper event summary (number of initiations, reinitiations, number and size of initial flaw depths by means of crack arrests, and stable terminating crack arrests) to convergence studies, and this was done for the OCA-P facihtate a more rigorous comparison of the proba, analyses.
bilistic solutioas. An examiration of the event sum-D,3,3.3 Summary of compar.1;n of maries indicated that the two Mes were predicting the pmbability of crack initiation to be approximately VISA II/OCA-P solutions equal; however, VIS A-Il was predicting lower values of P(FIE) as a result of predicting significantly more OCA P and the baseline version of VIS A-Il produce stable crack arrests than OCA-P.
nearly the same deterministic solutk)n for the Rancho Seco PTS event.
- An examination of OCA Psnd VIS A-Il by flow chart-ing down to a fairly fime level of detail was performed
' Duce errors were discovered in the probabilistic por-at ORNL. This examination reven'ed three areas m the tion of the VIS A-II code, one of which results in sig-VIS A-Il code that were thought to be the cause of the nificantly lower values of P(FIE) Upon correcting discrepancy bctween the two probabilistic solutions.
these errors, VISA-ll and OCA-P rodtred similar I
Corrections to VIS A Il appeared to be in order and probabilistic solutions, although as noted atove, there were dhcussed and coordinated with Fred Simonen at is still some difference in the OCA-P and VIS A U Pacific Northwest Laboratories. After ORNL made probabilistic methodologies.
these corrections to VIS A II, the prob 6ilistic solu-tions of VISA H and OCA-P agreed considerab!y bet-D.3.4 DeiailS of ORNL Sug-ter. %c following tabulation illustrates the proba.
geSted CDTTCC110DS i0 bilistic solutions of OCA-P, baseline VIS A II, and VISA-U (with the threc ORNL suggested corrections)
VISA-II ProbabiliStie for 100,000 trials for the Rancho-Seco PTS tranient:
Code
- 1) ne flags for flaw initiation (INITI A) and arrest N
(I ARRST) initialization were moved inside the of simuMng a ngw kw %mpt % b OCA-P VISA U 1 U
- E' 8'**b
'C*
" '"'##3 Number ofInitiations 3926 3711 3422 Number of Stable Arrests 488 3275 144 resgMe acamh sat M k numM Number of Failures 3438 436 3278 2) o e om al stress in the remaining Probabihty of Iruuation ligament to check for plastic instability was mods-ied (the sixth line below statement 500 in the bility of Failure main pr gram) t include the crack depth (a),i.e.,
PmE) 0.034 0.0043 0.033 stress = P * (R + a)by where:
D.5 NUREG/CR-5799
P = pressure at time, t, each crack depth incremeru w hen ads ancing the crack simulated crack depth, tip thruyh the thi:Lness of the vessel wall (in a =
R = inner vessel radius,and 0.25-in. screments) checking for crack arrest. This w=
vessel wall thickness.
randon. ness enhances the probabihty (relative to the methodology utilized in OCA P) of crack arrest be-Inclusion of crack depth (a) in die calculation of cause it increases the chances of a very high crack-pressure stresses results in a larger stress in the arrest fracture-toughness value at at Icast one of the remaining ligament and thus in more failures 0.25-in. increment check points for arrest. This is re-caused by plashe instabihty.
flected in the consideraNy higher number of stable crack arrests predicted by VIS A-II.
- 3) The call to subroutine ARTNDT (the thirteenth line below staternent 500 in the main program)
In reality th-re is variability of copper and nickel con-was deleted. This modificahon reconciled a sub-centrations through the wall; however, the approach tle yet fundamental difference in the VIS A-Il and adopted by both VIS A-Il and OCA-P assumes that OCA-P probabilistic methodologies and dramat-copper and nickel, for a specifie vessel, havc no van-ically reduced the number of stab!c crack arrests ability through the wall. Therefore, to be consistent 2
2 predicted by VISA II. The significance of this with this assumption,(o RTNDTO + g ARTNDT)W code modification is evident with an under-stand-ERRTN should also be assumed to be constant ing of how the salue of RT DT s calcula-ted.
through the wall for a specific vessel. The appmach N
i it is calctrlated by both VIS A Il and OCA P as utilized by VIS A-Il is equivalent to depending on in-follows:
homogenities in the wall to enhance the probability of crack arrest. His is a nonconservative approach, 3
RT DT = RTHDIO + ARTNDT + (0kTNDTo +
N Dele t q, se specified call to subroutine ARINDT in c1RTNDT)
- ERRTN, VIS A ll described above results m VIS A Il calculating a value of ERRTN once per simulated vessel. His is consistent with the methodology utilized in OCA P.
e imp ct of the first two code changes on the proba-RTNDT = Value of RTNUT adjusted for radiauon, bahstic soh, tion of \\ ISA !! was detectable but was not imbrittlement, significant w ith regard to ihe calculated value of PJ!E). De result of the third modification dramati-RTNUT0= Initial (um.rradiaica value of RT N N
cally d: creased the number of stable crack arrests and (User specified in input data) thus increased the number of failures and P(FIE).
ART UT = increase in RTNDT uc to radiatim D.4 OCA-P A d
N li EE ed to Yankce (is a function of fluence attenuated to the particular crack depth; copper, Rowe and nickel),
~
URTNUf = 10 uncertainty for the specified value 0
of RTNDTO, ne OCA-P code was used to perform an independent analysis of the Yankee Rowe vessel with the plant aRn'UT = 10 uncertainty in the correlation used suWected to the YAEC-specified SBLOCA7 FTS c
event (Fig. A-1). Based on recent data from YAEC to calculate ARTNDT' and the ORNL evaluation of RTNDT acktressed else-
~
w here in this document (Appendit C), the upper axial ERRTN is a nurnber between -3 and +3 that is we.J ~as selected for a detailed analysis of the condi-
+
obtained from a normal distribution having a tional probabihty of failure. Other transients and ves-mer.a of zero and a la of 1. ne product of sel regions also contribute to the overall frequency of 2
ERRTN and (0 RTNUTo + 0 ARTNUT)1/2 is th:failure; however,considering the preliminary nature of this study and the hmited time for its completion, it uncertainty in (RTNDTo + ARTNDT)-
was sufficient to conduct a detailed analysis of what were believed to be the dominant transient and region.
A fundamental differersce ia the twehne VIS A-Il and the OCA-P probabilistic methodologies is that OCA-P Input data used in the ORNL OC A-P and the YAEC calculates a value of ERRTN once for cach simulated VIS A-II heat transfer and stress analyses for Yankee vessel and uses this value throughout the wall thick.
Rowe are specified in Table D.1 (note that the ORNL ness when checking for either crack initiation or arrest.
analysis included cladding as a discrete region but the VISA-II calculates a value of ERRTN when checkmg YAEC analyds did not).
for crack initiation and then recalculates ERRTN for NUREG)CR-5790 D6 l
l l
l The thermal resp nse solutions predicted by VISA ll Value of 140 ksig ili. but P(11E) was not and 1R (the finite-element Octmal code used in con-
"8'"UC "d I npaced by this lower value Also,in Y
junction with OCA-P) are illustrated in Fig. D.5. As acmdance w e! Regulau g GuMe IW, k potennal indicated, the VlS 4 !! temperatures are a httle lower in knc6u of warmtqtremng wm not udu&d in the the base material. 'lhis is because the VISA-Il analy-anahn W an hus did assume a preservice inspec-sis did not inchxle the cladding, and because it was Lion as formulated by t e Marshall flaw nondetection based on slab geometry.
function.
Figure D.6 illustrates the SBLOCA 7 pressure tran-An Ws were pafmned for the SBLOCA7 transient sient. VIS A-II uses a polynomial to fit five points in desented m the i AEC irpirtl2 (no repressurization) time, w hereas OCA-P allows a rnore accurate point, and also for a case involvmg repressurtzauon to by-point description of the transient. This accounts 1M at a tune of 20 min, considering only the for the diffcsence indicated. Also illustrated in Fig.
upper axial weld in detail. The results of these analy-D.6 is the SBLOCA7 transient with repres-surization ses are presened in % y as a funcnon of k nm to 1.55 ksi (maximum head of safety injec6on sys-copp r mncenu.~oon, considering two values of the tem)11 at an assumed time of 20 min.
copper standant devtation (10 = 0.025 and 0.07). The actual weld chemistry is not known for the Yan-Lee J
Figure D.7 shows the hoop-stress soludons predicted we vessd, but a W esmak of RN wept pn-by VISA-11 and OCA P for the Yankee vessel when cent cp&l Oc =pper (mean with la = 0 07) and 0.7 weght pe subjected to the SBLOCA7 tnmsient. Both of the cent m
- 0) was &duced from Ref.1.,
OCA P hoop-suess solutions illustrated in Fig. D.7 included the 0.109-in, cladding; the VIS A-Il solutions
. e wrespon&ng,,be+ennate,, value of P(FIE) did not. The higher of the two OCA-P solutions in-with repressunzauon ano residual stresses is 2 x 10.
3 cludes a weld residual tensile stress of -6 ksi.
A rough estimate of the corresponding mean value of ORNL also compared baseline VIS A-II and OCA P Kg VE) p x 10 2, which was obtained by multiply-8
- """* ** "N "F""
values, even though the VIS A Il values correspnl to Y'
"N K innuence coefficients for R/w = 10, and those for 11e IPTS study foi ll.ts kobmson? (the corresrxeting 1
OCA P correspond to the actual Yankee geometry value in the YKr:C Yankee Rowe analysis was (R/w - 7). As mdicated in hg. D.8, for a/w < 0.5, 55 flaws /m3 for the upper weld).
the agreement is very pxxi. Since most initial crack initianons conespond to a/w < 0 5, a comparison of
'Ihe other regions of the vcssel(plate and other welds)
VISA-n/ Yankee and OCA-P/ Yankee is meardagful.
w 11 contribute to P(FIE); however, a more sophisti-cated analysis is required becauw of doubic<ountins D.4.2 Prolia1)ilistic Analysis probleros (more than one naw per vessei) introduced by the specific OCA-P methodology An appropriate input data and correlations used in the OCA P proba-analysis to account for all regions has not been per-bdistic fracture-mechanics analyses for Yankee Rowe fonned yet.
are persented in Table D.2, while Table DJ indicates differences between the input and correlations used in D.5 Flaw-Density the ORNL OC A-P analysis and the Y AEC VIS A-11 ConSIderat.10nS analysis.
The increase in RTNDr is a function of fluence The corlditional probability of vessel failure is directly (auanuated to the sp'cific wal!-depth kradon) and the proporuonal to the number of flaws in critical regions of the vessel, provided that the total number of flaws concentrations of copper and nickel. OCA-P was mod-in critical regions is one or less. With more than one ified to exactly irprtKluce ART tyr predicted in the fl w, a direct proportionality may fail tecause only N
" weld table" of Regulatory Guide 1.99, Rev. 2, plus a one flaw can result in failure riowever, with more low temperature operation correction factor of 44'F, th in one flaw the chances of failure tend to increase which is based on an irradiation-temperature correcdon becma k chames of having a entical flaw sue are factor of 1 F additional increase per l#F irradtadon increased, but the increase is not proponional to the temperature below 55(f F and operating data included in number of flaws.
g j g, in the li'IS study, the " test estimate" of the flaw den-and K c fracture toughness The ORNL mean Kla i
7 curves utilized in the IPTS studies were used in 1 Daw /m3. This flaw density was assumed OCA P f.or the Yankee analysis. The maximum value appropriate for all regiona (weld and plate) because it is of toughness at which crack arrest cotdd occur was spe' tdieved that the existence of shallow surface flaws is cified to be 200 ksib,in accordance with Ref. 7.
most likely associated w ith the cladding pncess and Additionai sensitivity analyses were performed using a attack of the cladding. There is, of course, a large D7 NUREG/CR-5799
uncertainty with regard to the surface density of lower axial welds, and circumferential weld). De plate shallow flaws, one reason being that they are and axial. weld regions were further sulxiivided longitu-extremely difficult to detect. flecause of the very large dinally to take advantage of the decrease in fluence shallow-flaw surface densit es *known" to exist in the toward the ends of the core. Assuming that the initial Sequoyah and Loviisa 0 vessels and the large axially oriented flaws are short enough to fall within
- uncertainties, a log normal distribution was assumed the height of a subregion, this procedure provides an for the IpTS studies. De most probable value was I accurate account of the potential for initial initiation of flaw /m3, the 84th percentile (+10) was 100 flawVm3, axially oriented fla"s. However, once initiated, the and the distribu. tion was truncated at the 94th Daw extends in surtace length beyond the borders of percentile (500 dawdm3). The correspondmg mean the specific subregion, and thus a higher fluence must value was 45 DawVm3, (It is of interest to note that be used for arrest and reinitiation. YAEC did not in-YAEC assumed essentially the same flaw density for corporate the latter feature, and thus initial initiations the upper weld but much lower densities for the plate tend to be treated accurately, but arrest and reinitiation regions.)
tend to be treated noncomervatively. De & gree of nonconservatism is negligible for initial flaws near More.ecently,0aw-density data have been obtained midheight of the core, where the neutron flux is a from sections of the Hope Creek and Midland vessels.
maximum and flat. For Daws near the end of the core, The corresponding surface densities were 6 and the error can be substantial.
7 flaws /m2 (Ref,13), while the surft e density corre-sponding to 45 flaws /m3 is 1I flaws /mJ Ifit is Division of the plate regions azimuthally to take assumed that the Hope Creek and Midland values are the advantage of the a7imuthal variation in flux could also most probable, and that a log-normal distribution with a be considered but was not. Instead, the flaw in the substantial standard deviation is reasonable, the mean plate was always assumed to be at peak flux in the values are substaatially greater than i1 Gaws/m2 azimuthal direction. This is a conservative approach.
Thus,it appears that 45 flaws /m3 is not necessarily a conservative mean value, D.6.2 Flaw Density Considering the volume of the Yankee Rowe upper axial weld and a flaw density of 45 flaws /m3, the De slaw density assumed by Yankee for the upper axial weld was 55 naws/m3 and for the plate about a number of flaws per weM r -1, m w;hich case there are factor of 200 less. The 5alue of 55 flaws /m3 is nearly no problems with double countmg, if only that weld the same as the mean value used in the ORNL IIrrS contributes sigmficantly to P(F1E).
studies 7 for all regions. ORNL believes, as re.ntioned in Sect. D.5, that surface flaws are most lihely the It appears that the upper axial weld is not the only resti: of the cladding process and/or sorne type of sigmficant contributor to P(FIE). As shown in Tabig 3 attack, such as stress corrosion cracking, in which case of Ref.14, the value of RTNDT or the upper plate is surface Daws are probably just as likely over base-f about the same as that for the upper axial weld metal as over welds. Thus, ORNL believes that
_ (~300 F). Assuming the high-fluence region of the
- higher Oaw densities should be considered for the plate upper plate to be substantially broader (azimuthally) regions.
than the weld region and assuming the Oaw densities for the two regions to be the same (for reasons men.
D 6.3 Flaw Confi uratton N
tioned above) the contribution of the plate region would be substantially greater than that of the weld.
Under these conditions there a more than one flaw Reference 11 states that infinite-length flaws were used total in all regions of concern, and, thus, P(FIE) is no for the initial initiating events and for subsequent longer directly proportional u the number of flaws.
events in the welds and upper plate, while a 47 in.-
P(FIE) wilt however, be substantially greater than -
long semiciliptical flaw was used for subsequent P(FfE) for the weld alone.
events in the lower plate. De YAEC VISA 0 input data sets indicate that 6/1 semiciliptial flaws were used D.6 Discussion of S Iecific f 'i"i" ""id*d " '*t** and f r subsequent events I
(arrest and reuutiation) 47-m.. long semielhpucal flaws Features in the YAEC were used for the lower plate and 94 in. flaws for the Analysis upper plate. The ORNL IPTS studies 7 considered l
both infinite-length and finite-length Daws for subse-l quent events, and the results indicated little differance D.6J Number of Subregions in the calculated value of P(FIE) for the document tran.
Considered in Beltline sients, which were high pressure. For low-pressure Region transients the effect w as much larger; however, ORNL has not conducted a similar comparison for Yankee He YAEC approach was to divide tne beltline region into five subregions (upper and iner plate, upper and NUREOK'R-5799 D.8
l L
D.6,4 VISA-II Code Errors 7.
D L Selby ei al.,cressurired ThermalShoct -
l Evaluatk>n of the li.B. Robinson Nuclear fower Plan NUREGER-4183 (ORNI/TM.
9M7h Man {m.;'arietta Energy Systems, Inc.,
it is assumed that the version of VISA ll used by l
- YAEC to perform the Yankee Rowe anal) sis contained the three coding errors discessed in Sect. D.3.4.
Oak Ridge Nauona, ' ab., Oak Ridge, TN, L
l Therefore,it is suspected that the results of the YAEC September 1985."
analysis world under predict P(FLE) because of the ten.
8.
Personal written comrnunication from F. A.
dency to over predict the number of stable crack
. Simonen of Pacific Northwest Laboratories to
- arrests.
D. G. Ball of Oak Ridge National laboratory, Marth 29,1984.6 i
RC(Crel1CCS 9,
- 1. 0, Meskle, An Approximate Stress intensity Factor Solutionfor a Deep Inside Surface 1,
U.S. Nucicat Regulatory Commission, longitudinal Crad in a Cylinder Under Regulatory Guide 1.154, " Format and Content Thermalleading,ORNI/NUREG/TM-3, of Plant. Specific Pressurized 1hermal Shock April,1976.a l
Safety Analysis Reports for Pressurized Water Reactors"."
10.
W. Marshall,"An Assessment of the Integrity
-I of PWR Pressure Vessels," Octobes 1976.'
2, R. D, Cheverton and D. G. Ball, OCA.P. A Deterministic and Probabilistic Fracture 11, YanLee Atomic Electric Company, Reactor hiechanics Codefor Application to Pressure Pressure Vessel Evaluation Reportfor Yankee Vessels, NUREG/CR 3618 (ORNL-5991).
Nuclear ronr Station, YAEC 1735, July Union Carbide Corp., Nuclear Division, Oak 9,1990,c Ridge National Lab., Oak Ridge, TN, May 1984.a 12.
K. E. Moore and A. S. lleller, B& W 177-FA Reactor Vessel Beltline Weld Chemistry Study, 3.
F. A. Simonen et al., VISA II - A Computer BAW 1799, B & W Ow cr's Group Materials Codefor Predicting the Probability ofReactor Committcc, July 1983.c Pressure Vessel Failure, NUREC/CR-4486 (PNL 5775), Pacific Northwest Lab., Richland, 13.
W. E. Pennell,"fleavy Section Steel WA (March 1986).a Technology Program Overview," presented at the 18th Water Reactor Safetf nformation I
4 Code of Federal Regulations, Title 10, Part 50 Meeting, Rockville, MD, October 22 24, Section 50.61 and Appendix G.a 1990,b 5.
Letter from F. A. Simonen, Pacific Northwesst 14.
Written communication frem Yankee Atona laboratories, to R. D. Cheverton, Oak Ridge Electric Company to Mr. William Russell of National Laboratory, " Review of Yankee United States Nuclear Regulatory Commission, Atomic PTS Report," October 29,1990.6 -
September 28,1990 e 6.
Mantred Geib, Verification of OCA-P and VISA 11 on Behalfof Strains and Stresses Induced During flDR-TEhfB Thermal ofixing Tests, Battelle-Institute, Frankfort, Germany.b 4
8Available for purchase from G PO Sales Program.
6 "Available for purchaw frum G PO Sales Program.
Available from National Technical Information Service.
bAvailable from Natumal Technical tnformation Service.
'Available from NRC Pubhc Document Rcumn for a fee.
D.9 NUREGER-5799
Table D 1. Input data used in the ORNL OCA-P and the YAEC VISA-Il thennal and stress analyses of Yankee Rowe Vessel dimensions:
Vessel Inner Radius = 54.5 in.
Wall thickness = 7.875 in.
Cladding thickness = 0.019 in.
Cladding propertiesa,b:
Modulus of Elasticity (E) = 27,000 ksi Poisson's ratio (v) = 0.3 Thermal expansion coefficient (a lad) = 9.9E-6/F c
Thermal Conductivity (k) = 10 BTU /hr-ft F Specific Heat (cp) = 0.12 BTU /lb F Density (p) = 488 lb/ft3 Base metal properties,c,d a
Modulus of Elasticity (E) = 28000 ksi Poisson's ratio (v) = 0.3 Thermal expansion coefficient (atuse) = 7.85E-6/F Thermal Conductivity (k) = 24 BTU /hr-ft-F Specific Heat (cp) = 0.12 BTU /lb-F Density (p) = 488 lb/ft3 Temperature Vesselinitial temperature = 515 F Waterinitial temperature = 515 F Coefficient of convective heat transfer =504 BTU /hr ft2*F aNo temperature dependence of material properties included in analyses.
bThe YAEC analysis did not include cladding in either the thermal or stress analysis.
cVISA-II requires an input value for E* abase /(1-v) rather than input for each of the individual parameters. The YAEC analysis used E* abase /(1-v) = 0.312. Using the OCA-P input values for E, abase, and v yields a value for E* abase /(1-v) of 0.314. This difference is not significant.
k dThe Thermal Diffusivity pc, of the base metal used by YAEC was 0.953 in.2/ min. For OCA-P it was 0.982 in.2/ min. This difference is not significant.
NUREG/CR-5799 D.10
I L
o Table D.2'. Correlations and values of parameters used in OCA.P probabilistic frrture-mechanic analysis of Yankee Rowe Volume of weld = 0.63 ft3 Flow stress = 80.0 ksi Flaw Data:
Haw density = 1 flaw /m3 (0.03 flaws /ft )
3 Number of crack increments to be used for initial crack depth = 9 Size of first crack depth increment = 0.169 in.
Extreme dimension of deepest crack depth increment = 2.25 in.
Marshall flaw size distribution function used Marshall flaw nondetection function used (simulates preservice inspection and repair)
Haws were assumed to be axially oriented and infinitly long Fracture Toughness Data:
Kic and Kh mean curves same as those used in the original WrS studies, i.e.,
K, mean = 1.25* ASME lower bound Kh curve i
Kie mean = 1.43* ASME lower bound Kic curve Maximum K = 200 Ksi 6,140.ksi 6a h
Kg standard deviation = 0.15 h standard deviation = 0.10 R.
Kle t
Kjetruncation = 130 K truncation =130 h
RTNDT Data:
RT uro = +0 F N
RT urostandard deviation = 17*F N
ARTNUrocalculated by Regulatory Guide 1.99, Rev. 2 (Welds) with an additional 44*F added as a correction factor for tud low temperature operation of the Yankee plant (44 F = 550 - 506 F)
ART ur truncation = 30 N
Huence at inner vessel wall = 1.24E+19 n/cm2 Huence standard deviation (fraction of mean) = 0.3 Huence variability truncation = 30 Mean copper content = various values Mean nickel = 0.7 wt% -
Copper standard deviation = 0.025 and 0.07 wt%
Nickel standard deviation = 0.0%
aUsed 140 ksi 6 for sensitivity study; however, this did not significantly impact the calculated conditional probabilities of failure.
D.!1 NUREG/CR.5799
Table D.3 VISA-ll/YAEC probabilistic fracture-mechanics analysis input data that were different from those used in the ORNl;OCA-P analysis The YAEC Analysis:
- 1) used the NRC mean Kic and Kia fracture toughness curves;
-2) - did not simulate a preservice inspection;
- 3) used ART m values specified by NRC; N
- 4) used Regulatory Guide 1.99, Rev. 2, to calculate ARTNDT; however, no correction factor for low-temperature operation was included;
- 5) assumed rero variability for RTNDTo (la = 0);
- 6) assumed to = 28 F for ARTNDT;
- 7) used a flow stress of 75.6 ksi; m
- 8) used la = 10% of mean for inner surface fluence;
- 9) truncated variability of fluence at lo;
- 10) assumed flaws were axially oriented, and were semielliptical with aspect ratio equal 6/1 for initial initiation and 47-in. long for arrest and reinitiation (lower plate and axial wid);
- 11) did not treat cladding as a discrete region; and
' 12) - used flaw-depth increments greater than those used by ORNL (may not be converged).
NUREGER-5799 D.12
l
-600 2.5
.i i
i i
i i
i 2.0
- 500 C
j 1.5
~
e
- 4i)0 3
2 m
o" a
1.0 E
e e
O-
- 300 e
0.5 s
Pressure lemperature 200 0.0 0
10 20 30 40 50 60 70 80 90 100 Transient time (minutes)
Figure D.I.
Approximate Rancho-Seco PTS event.
D.13 NUREG/CR-5799
600
- ;- a
~
580 f
f j,
560
.f
/
3 540 Eo CL
/
E 520
{
f 500 oo
_ o.
1R e
VISA 11 480 85 86 87 88 89 90 91 92 93 94 95 A
Vesselinner radius R (inches)
Vessel outer radius Figure D.2.
VIS A Il and IR predicted temperature distributions at time = 10 min for Rancho-Seco thennat transient (PJw -10).
NUREG/CR-5799 o,34 i
50
,j,
+--
VISA ll l
C OCA P l
40
)
Q
=
u)E 30 g
u>
m w
CH*
20 cl.
O O
%%%p 10 g
85 86 87 88 89 90 91 92 93 94 95 R (Inches)
Figure D.3.
VISA-I! and OCA-P predicted hoop stress distnbudans at dme = 10 min. for Rancho-Seco PTS event (R/w ~10).
l D 15 NUREG/CR 5799
800 c
p~g
/ (J 600
- c u
//
~
3 o
4 3-400 L'7 Tn n.
b
/
s m
OCA P
--+----
Vl3A 11 g
85 86 87 88 89 90 91 92 93 94 95 R (inches)
Figure D.4, VIS A-II and OCA P predicted stress intensity facters at time = 10 min. for Rarcho-Seco PTS event (R/w -10).
i NUREG/CR-5799 D.16
1 l-l-
550
-m-tr
- ,,r C
4 7,$h
}
s 450
~
y,9 o
u s.
3
/
M
/
wo 400 a.
i Eo I--
350 VISA ll w/o cladding 1 R w/o cladding 1R w/ cladding 300 i
I i
i 54' 55 56 57 58 59 60
.1 62 63-R (inches)
Figure D.5.
VIS A-II and IR predicted temperature distributions at time = 10 min. for Yankee Rows.
SBLOCA7 thermal transient (R/w ~7). -
l' l
l l-1.
D.17 NUREG/CR-5799
1 i
i 1
..i-av-.-,-,..*,,.----...n...-w.w.*.
c..--
.-,,m
~.,
Re-pressurization
[
to 1,55 ksi at time : 20 min f
=
=
=
j C
h 4
VISA 11 1.0
- s
@m DV
~
A
(
OCA P 0.0 0
20 40 60 80 100 time (min)
Figure D.6.
SBLOCA7 pressme transients.
l NUREG/CR-5799 D.18
60 g
g i
VlSA11 50 Y.
OC A-P (no residual stress) e OCA P (with 6 ksi residual tensile stress) g 40
?
1 s
's E
30 J
s m
s m
l 20
\\
tn
~,
a 10 N
s 0
x
-10 i
i i
-20 i
54 55 56 57 58 59 60 61 62 63 R_(inches)
VIS A-Il ud OCA-P pedicted hoop stress distr:butions at time = 10 min. for Ya Figure D.7.
SBLOCA7 PTS event (R/w -7).
l l
i l'
O.19 NUREG/CR-5799
i 60 g
g j
VISA 11 i
50 Y.
OC A-P (no r.esidual stress) g OCA P (with 6 ksi residual tensile stress) 40
\\
W
\\
6 30 1\\,
m m
\\'
w 20
'N E
N r
M s,
a 10 0
I i
-10
.l
-20 54 55 56 57 58 59 60 61 62 63 R (inches)
VIS A.Il and OCA-P pedicted hoop stress distnbutions at time = 10 min. fo Figure D.7.-
SBLOCA7 PTS event (R/w ~7).
i D.19 NUREG/CR-5799
800 c
VISA 11 KI
_5 OCA P KI
'.3
.E 600 lue E
o Q
Lt.
400 b
'Ec22 200 m
m 8
25 0
54 55 56 57 58 59 60 61 62 63 R(inches)
Figure D.8.
VIS A-!! and OCA P predicted stress intcasity factors :d time = 10 min. for Yankee Rowe SBLOCA7 PTS event (Pf.v ~7).
NUREGtCR-5799 D.20 i
i Best Estimate Chemistry; g
[0AW 1799puly,1983))
5 thckelConstant 07%(allpoints) 0-Copper Unrettainty; o 1 sigma = 0.025 Copper = 0.29;(1 sigma a 0.07) e 01 slefu s 0 070 s
Nicket 3 0.7%
g including Residua! stress : 6 ks1
.po
-o SBLOCA7: With Re.pressurizallon to 1.55 hsi at time a 20 rhinutes 3
5,coj NRC Mean Kk and Kla S
o o
E y 4 SDLOCA7:(No Be pressurization)
~
[
?
3 0001 O.15 0.20 0.25 0.30 0.35 0.40
% Copper Figure D.9. Results of OCA P probabihstic fracture mechanics analyses for Yarlee Rowe upper axial wcld subjected to I'TS event SBLOCA7.
l l
D.21 NUREG/CR 5799
Appendix E I
PNL Review of YAEC No.1735 F. A. Simonen Contents Em E.3 Review of Yankee Atomic PTS Repon E.3 PNL Conclusions E.4 Comparison of Yankee Rowe and PNL Versions of VISA.Il.
E.4 Basis for Comparison E.4 Numerical Compari.;ons,...........
E.5 Differences in Codes E.6 Residual Stresses E.6 Reg. Guide 1.99 Rev. 2 Shift.................
E.7 K. Solution for Pressme...........
s E.7 Check of Fractare Mechanics inpat..
E.8 Observation on Crack Arrest Behau E.9 Input for Clad Stress.........
..s E.9 Modeling of Fluence Gradiets.
E.9 Residual Stresses E.10 Input for Standard Deviath on aawcc.
E.10 Input for Standard Demtice os RTNDT Standard Devlations on CopMr and iackel E.Ii E.I1 1nput for Upper Shelf Fracture Toughness.....
E.11 Inputs for DeterminMic Parameters............
E.12 Input for Flaw length Before Initiation..
E.12 Input for Raw Length After Initiation........
E.12 mput for Flaw Size Distribution E.13 input for Flaw Density E.13 Polynomial Approximation of Transient 1
E.1 NUREG/CR.5799
Comparison of the OCA-P and VISA.ll Comptact Codes E.14 Programraing Errors E.14 Plastic instability Calculation E.14 Simulation of Shift E.15 Temperatures and Thermal Stresses E.16 Evalaation of Yankee inspection Program E.16 Attachment # 1,.........
E.19 Attachment #2 E.20 Attachment #3 E.23 Attachment #4 E.24 Attachment #5 E.25 Attachment #6 E.26 NUREGER-5799 E.2
OBaHelle October 29, 1990 pg, g,g g, Battelle 1 odes aid p o se, m Mr.
R.
D. Cheverton Rx Wr.d. w*ro m52 Pressure Vessel Technology Section Tele.hme d*3 7 5-2 0 8 7 Engineering Yechnology Division Oak Ridge National Laboratory P.O. Box 2009 Oak Ridja, Tennessee 37831-8047
Dear Dick:
REVIEW OF YANKEE ATOMIC PTS REPORT This letter is my-input for your review of the document " Reactor Pressure Vessel Evaluation Report for Yankee Nuclear Power Station", YAEC No. 1735, July 9, 1990.
My comments cover the following areas as described in your letter dated August 16, 1990, to Mr. M. E. Mayfield at NRC 1.
Comparison of Yankee Rowe's (Ron Gamble's) version of VISA-II and the PNL version.
J 2.
Check input to the fracture-mechanics analyres.
3.
Participation in the comparison of OCA-P and VISA-II, 4.
Evaluation of the vessel inspection program, My review of the Yankee report was performed from the standpoint of compliance with Regulatory Guide 1.154
- Format and Content of Plant Specific Pressurized Thermal Shock Safety Analysis Reports for Pressurized Water Reactors *,
PNL CONCLUSIONS Some details of the PNL and Yankee versions of the VISA-II code ere found to be somewhat different.
However, the two codes are expected to give similar predictions of vessel f ailure The probabilities except with respect to residual stresses.
Yankee version takes a more censervative approach to residual stresses than required by Reg. Guide 1.154, and therefore was found to-predict slightly higher values for vessel failure probabilities.
The input parameters for the Yankee calculations were reviewed l
item by item for consistency with Reg. Guide 1.154 and PNL's l
recommendations (NUREG/CR-4486) for application of VISA-II.
l While several details of the Yankee inputs differed from those used in prior NRC studies, sensitivity calculations indicate that_
these differences should not have a major impact on calculated failure probabilities.
Inputs for pressures, temperatures and irradiation induced embrittlement do have very significant E.3 NUREGER-5799
impacts - on calculated failure probabilities, but these parameters w re outside the scope of FNL's review.
PNL participated with CRNL in offorts to compare the VISA-II and OCA-P codes.
The codes were fcund to give similar predictions, except in a detailed aspect of simulating the shift in RTndt for purposes of predicting the arrest of a growing crack.
Both codes make reasonable assumptions for the arrest simulations, and give approximately the same numerical results if input parameters are assigned in a manner consistent with assumptions stated in the user's manual for the reepective codes.
The chapter in the Yankee report on NDE plans was reviewed by_
PNL.
It is noted tat the-Yankee report does not take credit for NDE as a factor in calculating vessel failure probabilities, and therefore Reg. Guide 1.154 does not call for discussion of inspection programs.
Nevertheless, this chapter does provide interesting and useful information of preliminary _ plans by Yankee for future inspection of the reactor _ pressure vessel.
COMPARISON OF YANKEE ROWE AND PNL VERSIONS OF VISA-II Section 5 of Reg. Guide 1.154 states that calculations chould be performed with a probabilistic fracture mechanics code such as OCA-P or VISA-II.
The Yankee Rowe evaluation was performed with a modified version of the VISA-II code, and therefore one part of FNL's review was to compare the Yankee Rowe version and PNL versions of VISA-II.
FNL's objective was to assure that the modified code still complied with the_regairements of Reg. Guide 1.154, and to assure-that any-code changes.did not introduce unacceptable unconservatisms into calculated failure probabilities.
Basis for Comparison - Formal documentation of the Yankee Rowe version of the code was not available for PNL's review.
Computer outputs from the-Yankee Rowe version of the code did permit PNL to make some limited benchmark numerical comparisons.
The review waschased on 1) numerical results from an example output file and as cited in rather limited detail in the Yankee Rowe report, and L
- 2) informal " word of mouth" reports of the types of changes that
[
were made in the Yankee Rowe version of the code.
L Numerical Comparisons - PNL was provided with a copy of the input-i data used in the Yankee Rowe-calculations for the case "SBLOCA 7, LOWER' PLATE,~ MAX MEAN DATNDT=315".- Calculations were performed with PNL's version of VISA-II for this set of data.
There-were no predicted failures for 500,000 simulated vessels, and this l
result agreed with the results presented in the Yankee Rowe l
report.
However, for this case-the Yankee report gave no more NUREG/CR 5799 E.4 l
,_,.--...v.
details or ciscussion of actual numerical results.
Rather significantly, PNL's calculations gave nearly 1,000 flaw initiations for the 500,000 s imula tions, all of which became arrested cracks without any through wall penetrations.
Such a i
trend is not'rentioned in the Yankee Rowe report.
During October 1990, PNL received input and output files for a ccmputerLrun made_with:the Novetech version of the coda.
Attachment il compares crack initiation and vessel failure probabilities for the Novetech and PNL versions of-VISA-II _ The Novetech version predicts high-probabilities, with the higher probabilities ~ attributed to the inclusion of residual stressen.
Further comparisons of'the two versions beyond the results of Attachment #1 could not be made,. since PNL was not provided the deterministic output of the Novetech version.
A full comparison would require that FNL have a copy of the-Novetech code, so that more extensive benchmark calculations can be performed.
Differences in Codes - Prior to the Yankee Rowe review Dr. Fred Simonen of PNL (one of the code developers) and Mr. Ron Gamble of Novetech Corporation (a user of the code) had engaged in phone discussions of detailed aspects of the VISA-II code.
These discussions _had occurred on several occasions over a-t me_ period of a year or more.
During a discussion on August 6, 1990,- Mr.
Gamble described a number-of changes to VISA-II.
It has since
'become-known that Novetech performed the probabilistic fract'ure mechanics calculations on behalf of Yankee Rowe.
A subsequent j
phone discussion between Simonen and Gamble at Novetech on October 18, 1990 furt+:
flarified aspects of Novetech's version F
of Visa-II.
Notes from the August 6th phone discussion indicate the following modifications:
1.
Inclusion of a velding residual stress of 8.ksi tension at the -inner. vessel surf ace and -becoming compressive at
-the mid wall ^f the vessel.
The distribution-was said l.
to be-consistent with data published by Paris.
2.
The shift in RTndt was re.saed to exactly reproduce the numbers in the tables of Reg, Guide 1.99 Rev.
2.
The final version of this guide was published after the VISA-II code was issued, and the numbers in the final-version of 1.99 differ sightly from the numbers upon which-VISA-II vere based.
r E,5 NUREG/CR.5799
N 3.
The crack tip stress intensity factor solution for internal pressure loading was replaced with a recent solution due to Dr. Zahoor of Novetech.
This' list of changes is generally consistent _with subsequent statements made by NRC staff during the course of this review effort.
]
Lacking full _ details of the Yankee Rowo version of VISA-II, we l
can only offer some qualified comments.
In particular it will be assumed _that all.the modifications as described by Novetech were correctly implemented into VISA-II with no coding errors.
Residual Stresses - The inclusion of residual stresses was not stated as a requirement in Reg. Guide 1.154.
During the development of VISA-II the possible presence of residual stresses was recognized.
However, there is generally little information regarding the levels and distributions of such stresses for-a given vessel,-although levels of residual stresses in the welds of reactor vessels are believed to be small relative to the thermal stresses during PTS events.
In the overall balance between conservative and unconservative assumptions a decision was made to neglect residual stresses in the VISA-II code.
Inclusion of_a modest level of residual stresses as in the Yankee Rowe calculations would increase the number of initiation events, but~should contributed little to the noted tendency for the these initiated cracks to arrest before becoming through wall cracks.
The residual stresses were assumed to be approximated by a cosine
_ function.
PNL was referred _to a solution in tl. "Tada Flacture Handbook" for details of the crack tip stress intensity factor solution by Novetech.
This handbook gave a polynomial function
.for the solution, which differed from the trigonometric type of function _ described by Novetech.
Nevertheless, the two functions-give the same general trend for stress. intensity factors.
They agree for small ID surface flaws, and both predict small values of stress intensity factor _for deep flaws-that extend to the_mid-wall of the vessel.
For the_ Yankee evaluat' ion, we suggest that residual stresses be neglected because: 1) the calculations would then be fully consistent with Reg. Guide 1.154, and 2) this would avoid concerns that inclusion of residual stresses in the Yankee
-calculations contributed to the large number of predicted crack arrest events.
Reg. Guide 1.99 Rev. 2 Shift - The recoding of the RTndt shift equation was not considered to be an important consideration in the review of the Yankee Rowe calculations.
PNL has found its l
. NUREG/CR-5799 E.6
calculated shift values to adequately reproduce numbers tabulated in Reg. Guidt 1.19 Rev.
2.
Nevertheless, further precision in the calculaticn as done for the Yankee Rowe evaluatica $s certainly an acceptable change to the code.
K-solution for Pressure - The stress intensity factor solution for pressure loading in PNL's version of VISA-II has been checked for accuracy, and has been found to give acceptable results for the probabilistic fracture mechanics calculations.
Nevertheless, further precision in this part of the calculation is certainly an acceptable code revision.
It is understandable that Novetech would make use of equations frcm their own recent re2earch.
CHECK OF FRACTURE MECHANICS INPUT PNL was provided with computer files of the input data for certain of the calculations described in the Yankee Rowe report.
Each item of this input for VISA-II was reviewed for consistency with the guidelines given in Reg. Guide 1.154 and in NUREG/4486 (the user document for VISA-II).
The main focus of this review was the case titled " YANKEE, SBLOCA 7, LOWER PLATE,.VJsX ME AN DRTNDT-315".
Part of PNL's review consisted of performing calculations with PNL's versicn of VISA-II using the Yankee input.
Some 38 variations of the baseline case " YANKEE, SBLOCA 7, LOWER PLATE, MAX MEAN DRTNDT-315" were evaluated to establish if Yankee's values of vessel failure probabilities were particularly sensitive to choices of input parameters.
These calculations are listed in Attachment #2 along with the calculated values of crack initiation and vessel failure probabilities.
It should be noted that the levels of RTndt estimated for the Yankee vessel are quite high.
For this reason we have concluded that predicted crack initiation and arrest events are to a large extent governed by the materials fracture toughness in the lower shelf regime.
As such, the failure probabilities appear to be rather incensitive to parameters that govern the shift in RTndt (i.e. fluence, chemistry, errors in shift predictions, etc.).
On the other hand, parameters that govern the applied level of stress intensity factor (pressure levcl, crack length, etc.) have more significant effects on failure probabilities.
Our comments en fracture mechanics inputs will compare the Yankee inputs to those that have been recommended and used by PUL for the VISA-II ccmputer code (NUREG/CR-4466).
It is assumed that CRNL will in a similar manner ccmpara Yankee inputs with those used for OCA-P in the IPTS study.
E.7 NUREG!CR-57W
M2geryatlu' AM_REA9.kJr.regLDhaylpt - A ve ry ctr! ning t rend was seen in the malculation for tbc baneline case (case 1 of Attachment #2).
While the calculated failure probability (lenn than 1.0E-05) agreed with the renult reported by Yankee, the output from VISA-Il chowed a rather high probability of crack initiation (1.05E-03).
However, cr ack ar* Mt was predicted to accur for all the initAated cracks, and henco no vessel failuret were predicted.
keasons for thin unucual trend were unught.
The 38 VISA-17 calculations of Attachment P2 point to certain factorn that contribute ntrongly to the very large nunber et arrect events:
1 The presnure during the critical parto of the anal 1 break LOCA transient for Yankee Rowe remaine at a relatively low level of 670 pai.
Evidently reprencurization behavior as predicted in other PTS studies lo not predicted to occur for the Yankw Howe plant.
Caces 3-5 of Attachrent #2 nhow that on repressurization to only 1000 psi renultn in a noticeable increase in the calculated vencel fallure probability.
A substantial (but typical) reprecourization to 2000 poi given a relative;y high failure probability of 7.990-03.
It in recormended that the Yankee accident scenarios be clonely examined, to determine if possibilitico for reprecourization have been overlooked.
2.
The Yankee calculationc assume that lengthw.ce growth of flaws in the lower plate will not exceed the 47 inch dimension of thin plate.
Case 11 of Attachment #2 chows that cracks no longer tend to arrest, if the initiated flaws are permitted to grow co an eccentially infinite length.
Cason 16-18 addrecs flaws of varioun
- lengthr, a..a. Lows an progressive increase in failure probability as the initiated flaws are permitted to become longer.
Cace 16 (failure probability of 3.80E-05) is of particular interest, since the final length of the flaw is the height of the beltline regian of th vessel.
3.
The Yankee calculations assumed an upper shelf fracture toughneca (for both initiation and arrect) equal to 200 ksi 6.
Cases 12,13,30,31 and 32 address lower valueu of this upper shelf toughness.
It is seen that the Yankee value of 200 must be reduced to 70 kaiVIIi before the terSency for cracks to arrent is reduced.
A toughners of such a low level in not considered to be a credible assumption.
4.
Many vther facters are addr e tt s ed in Attachment 62, and none of theso were found to reduce the strong trend for crsck arrest.
Scmo inpact on initiation probabilities can be noted.
Hcuever, wo did not detennino if any of these factors in ecmbination could sollectively give a substantial irpact on crack arrest behavior.
Input _for Clad Stross - The Yankee calculations neglected clad stresses as a factor that can prcmoto crack growth.
Evidently the clad offects were considered to be insignificant becauso tno clad is relatively thin, and because mechanical interactions were minimal due to the " stitch" process used to bond the clad to the bascmetal.
Information available to PHL indicated that the " stitch" attachment actually bonds clad to a rather significant fraction of the inner surface of tho vessel.
Also the veld areas are clad in a conventional manner with a wold deposit.
Thoroforo interaction of clad with the basemetal is probably'substanti'a1.
In Caso #8 clad stresses woro modoled as part of the VISA-II calculation.
However, the results show little change in the calculated probability of crack initiation, and it is concluded that the Yankoo calculations woro reasonablo in neglecting clad stresses.
Modeling of Fluence Gradient - The Yankoo calculations mado uso cf a ieature in VISA-II that permits simulation of the spacial
~
variations in neutron fluence over the inner surface of a vossol.
Reg. Guido 1.154 would permit this approach, which accounts for the fact that the peck surface fluence may exist only over small fraction of the overall surface area of a given plate or weld.
In Case 6 of Attach ent #2, the baselino calculation was performed but with tno conservativo assumption that the peak fluence existed over the entiro surfaco of the lower plate.
The resulting probability of crack initiation increased by a factor of about 10.
This was about the expected change based on the fraction of the plato exposed to the peak levels of fluence.
Besidual stresses - The Yankee calculations included e contribution of welding residual stresses to the crack tip stress intensity factor.
A cosino function approximated the distribution of residual stress through the vessel wall.
There was a tensile stress of 8.0 ksi at both the inner and outer surfaces of the vessai, and a ccmpressivo stress at the mid wall location.
The tensile residual stress at the inner surfaco increases the probability of initiating flaws.
It is estimated that the E.9 NUREG/CR.5N9
residual stress was rcughly equivalent to an increased pressuro of about 1.0 Ysi.
Case 4 of Attachment #2 suggests an increase in the crack 1..itiation probability by a factor of about 3.0, Gi.*en crack initiation, the residual stress should have little effect on crack arrest events.
Input for sf;ndard Deviation on Fluence -
The Yankoo calculations assumo a stan3ard deviation on fluenco of 0.1 of the n.can fluence valuo, as compared to the 0.3 valuo suggested in NOREG/CR-4486.
Tho results for Case 14 indicates that 0.1 versus 0.3 has only a small ofiect on calculated failure probabilities (factor of about 10 po; cont).
It is believed that the Yankeo Report bases a lower value of 0.1 on the foct that the fluenco lovols aro rather well establishod for the Yankee vessel.
Hevover it should be noted that VISA-Il usos the uncertainty in fluence levols in largw measure to represent the uncertainty in predictions of the si:if t 9quation.
In this rogard, imornved knowledge of fluences for the Yankoo vossol is not rolove.cw to the uncertal.itios in the fluence levels for the surveillanco opectmens which formed tho 'sasis of the RTndt shift corrolation.
Nevertholoss, tho Yankee calculations are perhaps consistent since they apply an error term to the shift equation ao an alternativo to the 0.3 sigma value for fluer.co uncertainty.
Input for Standard Deviation on RTndt - The Yankoo calculations use inputs that difier somewhat from thoso suggested in the VISA-II user document (NUREG/CR-4486):
Standard Deviation. d.egree F Initial Value of RTndt Shift Yankee Calculations 10.0 17.0 NUREG/CR-4486 17.0 0.0 (for Plate)
Case 1 versus Case 33 of Attachment #2 compares calculated failure probabilities for the two approaches.
The difference is only about 15 percent for initiation probability.
This trend is consistent with our belief that failures are governed by fracture toughness on the lower shel'f, and are therefore insensitive to calculated levels of shift.
Evidently the Yankee report follows certain recommendations of Reg. Guide 1.99 Rev. 2.
These recommendations address the I
NUREG/CR-57W E.10
niculatica cf shift fer purpercs of detorzinistic fracture acch,es calculations, withcut regard to the probabilistic appr u6 used in the VISA-II cede.
'n'a noto here that VISA-II variability in shift through the imposed uncertainties add
+
if-o, copper centent, etc., and therefore adding an error term to the chift equation introducon excocsivo " noise" into the calculations.
An input of zero error in the shift equation would Le nero censistent with the recomrendations of NUREG/CR-4486.
Standard Deviations on Copper and Nickel -
The Yankee inputs differ senewhat from t!.e valuer suggested in NUREG/CR-4486 as follows:
Standard Deviationt
'h t it Copper Nickel Yankee Calculations 0.017 0.05 NUREG/CR-4466 0.025 0.00 case 1 versus Case 34 of Attachment #2 compares calculated failure probabilities for the two approaches.
The difference is only about 10 percent for initiation probability.
This trend is again consistent with the fact that failures are governed by lower shelf fracturo toughness, and are therefore insensitive to calculated levels of shift.
Input for Upper Shelf Fracture Teu3 nos h
calculations assuned an upper shelf val.s - Tho Yankeoue of 200 ksi\\/in for both the initiation and arrest toughnesses.
These values are consistent with prior applications of the VISA-II code.
Cases 12,13,23-32 address the sensitivity of the calculated f ailure probabilities to decreases in upper shelf toughness values (as low as 50 ksi\\/in).
Unless the toughness is decreased to a value less than 100 there is little impact on the failure probability.
Only at a toughness of 70 ksi\\/in (Case 31) did we predict that a significant fraction of the initiated cracks penetrato thu vessel wall without arrest.
Inputs for Deterministic Parametern - The table below compares inputa used in the Yankee calculations with the corresponding inputs reccmmended in NUREG/CR-4486.
There are no significant cifferences between the two sets of parameters.
E 11 NUREG/CR $799
NUPEG/CR-4486 Yankoo Thermal Diffusivity 0.982 0.983 Fluid / Vessel Film Coefficient 400 504 Constant for Fluence Attenuation 0.24 0.24 E alpha (1 - nu) 0.320 0.312 Wann Prestres; no i
no ISI no no Flaw Length Before Initiation infinito 6:1 Flaw Length After Initiation infinito length of eo l d..
Lccation of Flaws ID Surfaco ID Surfaco Input for Flaw Length Before Initiation - The Yankoo input specifies that the flaw aspect ratio beforo initiation is Gil, but also providec a tabular description of the (Marshall) flaw sizo distribution which gives an infinito length to the flaws.
These are conflicting inputs.
However, examination of the outpat of VISA-II shows that the later specification of infinito flaw length governed in the calculations.
Input for Flaw Longth aft ~r Initiation - 1ho Yankee calculations Fjecifies that _ the flaw extends in length only to the entire height of the plato or weld of concern.
Cases 16-18 show that failuro probabilities will increase if the flaw is: permitted to grow beyond the confines of the plato or vold.
The assumption of finito flaw length is a possible unconservativo feature of the Yankte calculations.
Input for Flaw Size Distribution - VISA-II uses the Octavia flaw size distribution as the default selection, but recognizes uncertainties in this aspect of the probabilistic model by suggesting that the Marshall distribution is also a suitable selection.
The PTS screening limit was based in part on VISA calculations which used the Octavia distribution, whereas the subsequent IPTS calculations used the Marshall distribution.
The use of the Marshall distribution in the Yankee calculations is consistent with Reg. Guide 1.154.
Case 9 of Attachment #2 shcws that the Octavia distribution will give a similar but NUREG/CR-577)
E.12
semcwhat 1cwer probability of crack initiation (by a factor of about 2) than the Marshall dis ribution.
Input for Plew Density - The Yankee report follows the IPTS assumption of eno flaw per cubic meter of vessel material to estimate a total of five flaws for the beltline region of the vessel.
It then assumes that five t'aws are approximately equivalent to assuming eno flaw in each of the plates and welds of the vessel beltline.
An implicatien of this assumption is that flaws are more likely (on a per unit volume basis) in Welds than in basemotal.
That is, tho Yankoo calculations imply that there is far more than one flaw per cubic meter of weld metal.
Wo believe that this assumptien is plausible, and does not conflict with Reg. Guide 1.154.
The documentation for the VISA-II code does not abko specific reference to a given number of flaws por cubic meter.
The original Octavia distribution is believed to have assumed a total of one flaw for the six axial wolds in the beltli60 region Bf a reactor pressuro vessel, for consistency, it would reasonable to assume there is also one flaw for the circumferential wolds of the vessel boltline.
The discussion of 11UREG/CR-4486 suggests that flaws are less frequent in basemetal (i.e. flaws per unit volume of metal).
Hence, a total of one or two flaws in the basemetal of the beltlino is a logical extrapolation of the VISA-II approach.
In conclusion, the VISA-II documents would suggest an assurptica of some 3-4 flaws in the beltlino region of the Yankeo vessel.
In summary, tho Yankee calculations assumo somewhat moro flaws than the prior studios referenced in Reg. Guido 1.154.
Thus the Yankee predictions of vessel failuro probabilities are conservativo in this regard, n
. Polynomial Ap_proxination of Transient - The VISA-II code approximates pressure and temperature transients with polynomials that are fit through five points of the transient, for the small break LOCA both the temperature and pressure have rapid changes during the first five minutes of the 100 minute transient.
Therefore the polyncmialc give a relatively poor approximation during the critical early part of the transient.
for Case 15 (Attachment #2) the calculations focused only on the early part of the transient, and as a result the polynomial approximation was much improved for this more limited timo period.
This inproved calculation gave a slightly lower prcbability of crack initiation (1.28C-03) than for the baselino calculation (1.65E-03).
Thus this aspect of the Yankee calculations is somewhat conservativo.
5 It l 3 NURiiG/CR 5799
I coy.FARISON OF THE OCA-P AND VISA-II. CCMPUTER CODES A major part of FNL's review consisted of interactions between fred Simonen at FNL and Terry Dicksen at ORNL in a cocperative effort to ccmpare the OCA-P and VISA-II codes.
PNL's latest versien of VISA-II (i.e. the version that was placed in the Argcnne code center) was sent to ORNL on Auguet 14th along with other data and documentation.
Similarly CRNL sent a copy of the OCA-P code to PNL along with documentation.
The VISA-II code was
?
installed and extensively exercised at ORNL.
Installaticn of 1
OCA-P at PNL was found to be a mero involved effort, which was beyond the scopo of the short term review project Both VISA-II and OCA-P are specifically mentioned in Reg. Guide 1.154 as examples of prcbabilistic fracture mechanics codes that aro ccnsidered suitable for uso in PTS evaluations.
During the devolcpment of VISA-II there were semo limited benchmark calculations that ccmpared results from the codes-(letter frem F.
A. Simonen of PNL to D.
G. Ball at ORNL dated March 29,1984).
The tua codes were found to give generally the same results for vessel failuro probabilities.
Numerical differences in the 1984 study were attributed to different assumptions regarding flaw sizo distributions, simulation of fracture toughness, etc.
Both codes have sinco undergono further development, and have been extensively r..odi fied.
The benchmark comparisons as discussed here are fcr the latest versions of the code, and address a wider range of input variables than considered in the 1984 calculations.
Programming Errors - Some minor programming type errors in VISA-II were found and corrected as part of the benchmarking activity.
These are described in Attachments 2 and 4 to this letter.
The reccmmended corrections of Attachment #3 prevents the generation of an excessively long output summary tablo, which is semo cases caused the calculations to abort due to a stack ovbrflew condition.
Attachment #4 addresses a concern (noted previously by some users of the code) whero the tabulation of initiation and arrest events appeared to give inconsistent totals, with occasional cases where the calculated crack depth for an arrested crack was smaller than the initial deptn of the crack (physically impossible).
Our recent review shows that VISA-II was double counting the number of initiation events for certain unusual combinations of simulation parameters, and giving other associated inconsistencies in the output table.
Plastic Instability Calculation - To be more correct and to be
'onsistent with OCA-P we have made a change in VISA-II to account or the pressure acting on the crack faces during the prediction Nt' REG /CR 57W E.14 e
cf plastic instability.
This char.ge has little effect en calculat ed f ailure prcbabilit ies.
After correcting the " minor" programming Sinulation of Shift errors in VISA-II, there were stLtl rather c'gnificant and unexplainable differences in calculated failure probabilities when comparing CCA-P and VISA-II.
These particular numerical comparisons were fer the Rancho Ecco transient.
Recults of the cc:rpariscns are summa rized in Attachment #5.
Case #1 versus Caso #2 indicates the apparent lack of agreement between OCA-P and VISA-II that sas first noted by CKNL.
Case #1 versus Caso #3 indicates the very gcod agreement that was eventually achieved once a basic difference in the probabilistic assumptions and logic of the two codes was identified.
Cace #1 versus Case #4 indicates the rather reasonable agreement (within a factor of about 2) when OCA-P and VISA-II are each applied using assumptions and consistent inputs as reccmmended in their respective uscrs manuals.
For Caces #1 and #2 the two codes actually agree quite well in their predictions of the probability of initiating a crack.
However, VISA-II appears to predict a much greater trend for cracks to arrcat, once they do initiate.
After a careful look at each code it was determined that VISA-11 resimulates the randem errce in RTndt for each small increment o.
- rack depth during the simulation of crack growth and arrest events.
In contrast CCA-P simulates the error in RTndt cnly cnce fcc each crack and uses this same error term for each advance of the crack as it predicts if the growing crac'- will aires t.
Cnce VISA-II was reprogrammed to match the assumption used in OCA-P the very good agreement of Case #1 versus Case r3 resulted.
The crack arrest calculations of the VISA-II and OCA-P codes follow scmewhat differen*. philosophical apprcaches for simulating the variability of RTndt.
We belicve that one approach is not inherently more correct than the other.
Both approaches appear to be reasonable.
Furtheinore the predicted failure probabilities frem the two codec agree within a factor of about two.
The Yankee Rowe calculations have been reviewed frcm the standpoint of simulating variability in shift.
We fcund that Yankce's application of the VICA-II code and the selection of input parameters were not entirely censistent with roccmmendations given in the user document for the code.
E 15 NlIREG/CR-5N9
However, sensitivity calculatiens shcwed that tro Yankee selection of inputs had only a modest impact on predicted failure prcbabilities.
B y ergtr_res and_ Thermal Stresses - Another part of the benchmark effort addressed the deterministic calculations of temperatures and thermal stresses.
Solutions were compared for the Yankee Transient *SBLOCA 7, LCUER PLATE, MAX MEAN DRTNDT=315".
ORNL generated results using the OCA-P and the ADINA general purpose finite element code.
PNL generated solutions using VISA-II and the ANSYS general purpose fini te clement code.
Attachment #6 gives FNL's results that show rather good agreement between the temperature and stress calculations of the VISA-II and ANSYS codes.
With the ANSYS code it was possible to model the cylindrical goonotry of a vesuel, and the calculations were performed icr R/t 6.916, 10 and 1000.
VISA-II uses flat plate a
solutions to approximate the temperatures and thenn31 stresses in a vessel, and the numerical results of Attachment 66 are in fact identical for the cases of R/t 6.916 and R/t =_.100.
The.,ANSYS a
results show a relatively omall effect of R/t on the temperature and thermal stress solutions.
As expected, the VISA-II flat plate solutiens agree hest with the AUSYS solutions for the largest value cf R/t (=1000).
In conclusion, we have further validated the temperature and thermal stress calculations in the VISA-II code.
For this purpose we used water temperatures for one of the critical transients fren the Yankee Roue report.
We belicve that VISA-II as applied in the Yankee Rcwe evaluations provides an accurate method for calculating temperatures and thermal stresses.
EVALUATION OF YANKEl' IM PECTIC'N PROGRAM In Chapter 7 (titled "Reectcr Vessel Inspection") first briefly describes the fabrication history, preservico inspections, and inservice inspecticns to date.
The chapter then concludes with a much longer discussion of plans for future inspecticns with particu)ar attention to the beltline region of the vessel.
The PDL review has addressed this Chapter of the report.
Section 8.3 of Reg. Guide 1.154 gives brief note to inservice inspection as an optional part of a plant specific analysis of PTS risk.
Discussion of ISI is required in the FTS risk evaluation cnly if state-of-the-art nondestructive examinations (NDE) are used as a basis for decreasing any ccnservatism in the flaw density value used in the analyses.
The Yankee vessel evaluation does not decrease any conservatisms using NDE as the basis, and in this context the discussLon of Chapter 7 is not NUREG/CR-5799 E.16
ersential to the report and cculd be deleted.
1:evertheless, thi-and contributes to a Chapter does provide useful information, more coaplete understanding of the status of activities by Yar.)xe Atcmic in the area of vessel integrity.
Inspection of the beltline of the Yankee vessel is made dif ficult by access prcblems, and by the partial bonding of the cladding to the basemetal.
Yankee appears to be making a systematic effort to find soluticns to these prcb1 cms, and scme of the proposed approaches ara encouraging.
It should be noted that these o f ** s are being driven by an NRC requirement that the beltline
. Yankee vessel be inspected by 1993.
It should be noted c
for that this requirement is only indirectly related to concerns pressurized thermal shock.
A PNL expert en 3DE technology (Dr.
S.R. Doctor) has reviewed the Tne information centent of Chapter 7 of the Yahkee Report.
contained in Chupter 7 provides some very good background on and type of both PSI an d fabrication history, and on the extent that has been conducted on the Yankee Rowe vessel in thB ISI Those inspecticr.s were condacted to the standards of the past.
day, but in many cases does not providt, the inspection offectiveness that is necded for the flaws of concern to the PTS The chapter then discusses alternative means of access to issue.
both the inside and the outcide surfaces of the vessel while listing the advantages and disadvantages of each access alternative.
It would appear that the mosc likely possibility is to conduct the inspection frcm the inside.
For the belt line fits within a 2 inch region this will require a scanner that annulus between the thernal shield and the vessel wall, Scannere will need to be developed to operate with this physical l'i n a lly, the chapter contains an overview of some constraint.
preliminary research condected to adurocn the inspection of thy and stitch cladding using s combina+. ion of eddy current ultrasonic methods.
The work ccnducted on these techniques shows the inspection is not a hcpeless case.
However, until that it is not techniques and procedures have been ' ally developed, possible to ccmment an the offcctiveness of the proposed inspection.r,ethods.
Ultimately, blind testing of the proposed methods will needed to determine technique reliability, and then the measured inspection sensitivity will need to be compared t o the sizes of flaws that are impcrtant to the PTS issue.
E.17 NUREG/CR.$799
Fitase call c.a at 533-375-2CS7 11 ycu have any questions cr C C L".O n t S.
Sincerely, h($
iLGP(M F. A.
Simonen Theoretical and Applied Mechanics Group Energy Sciences Department
/fas cci ME Mayfield - USNRC SR Doctor l
NUREG/CR 5799 E.18
ATTACHMENT $1 Case 1 Case i Caso 2 Case 2 PARAMETER VISA-II VISA-II VISA-II VISA-II (Novetech)
(PNL)
(NOVETECH)
(PNL)
- initiations 56 349 164 669 5
- arrests 56 349 156 669 l
- stable 6
299 114 665 arrests
- failures 50 50 50 4
probability
.048
.0099
.051
.0067 initiation probability
.043
.0012
.015
.00004 failure a
1.
Total number of vessels simulated was 100,000. for all cases 2.
Caso 1 is for upper plate, max mean DRTNDT=248 3.
Case 2 is for lower axial weld, max mean DRTNDT=288 4.
PNL's version of VISA-II is the original VISA-II with only the corrections for counting of initiation events and considerction of crack depth for plastic instability anslysis.
E.19 NUREGER 5799
ATTACHMEliT #2 VISA-II SE!1SITIVITY DTUDIES Baseline cace lower plate - SDLOCA Yankee input file LP-315.I!J YAfJEEE, SBLOCA 7, LOWEli PLATE, MAX MEA!1 DitTIJDT-315 INITIATIO14 FAILURE C._ABX Y_hFl&TLQN_.RQli_Jh8ILLl40 CASE PEQDADlklD PROEADILITY 1
lione (baseline) 1.85E-03 0.00 2
Precsure = const. = G70 psi 1.98E-03 0.00
'.)
Min. pressure = 1000 psi 3.22E-03 2.00E-05 4
Min. pressure = 1500 psi 4.96E-03 1.87E-03 5
Min. pressure = 2000 poi 1.04E-02 7.98E-03 6
All material at max. fluence 2.14E-02 0.00 7
flaw aspect ratio = 999.0 9.34E-03
.00 3
Clad stress 1.25E-03 0.00 9
Octavia flaw distribution 0.72E-03 0.00 10 Threshold flaw size 1.92E-03 0.00 11 Infinite flaw lenoth 1.83E-03 1.36E-3 after initiation 12 Upper shelf toughness = 150 1.85E-03 0.00 13 Upper shelf toughness = 100 1.89E-03 0.00 14 Fluence sigma = 0.30 2.04E-03 0.00 15 Better polynomial fit of 1.28E-03 0.00 carly transient 16 Flaw length after initiation 1.85E-03 3.8E-05 full height of beltline (122.6) 17 Flaw length after initiation 1.84E-03 2.0E 06
= 90 inch NUREG/CR-5799 E.20
ATTACHMENT #2 (continued)
VISA-II SENSITIVITY STUDIES Baseline case lower plate - F.Bl OC A Yankee input file LP-315.l!J Y A!JKEL, SBLOCA 7, LOWER PLATE, MAX 14EA!J DRT!iDT-316 INITIATION PAILUBE CASE YAR_IAIl0M EBOM_DADEldRil_ CASE PROEhPIM TX PRODA. DIM TY 18 Flaw length after initiation 1.90E-03 1.00E-04
= 160 inch 19 Kia truncated at 1.94E-03 0.00 99 standard deviations 20 Kia truncated at 1.86E-03 0.00 99 standard deviations 21 Kic standard deviation 1.15E-02 0.00 0.3 of mean 22 Kia standard deviation 1.85L-03 0.00 0.2 of mean 23 Zero sigma on Kic & Kia 1.32E-04 0.00 Uoper shcif toughness = 100 24 Zero sigma on Kic & Kia 3.5CE-02 3.500-02 Upper shelf toughness = 56 25 Zero sigma on Kic & Kia 1.34E-04 0.00 Upper shelf toughness = 75 26 Zero signa on Kic & Kia 2.17E-03
?.07E-03 Upper shelf toughness = 65 27 Sigma on Kic = 0.0 1.34E-04 0.00 Sigma on Kia = 0.1 of mean Upper shelf toughness = 75 E.21 NUREG/CR&n
l A?tACHMEllT 12 (continued)
VISA-II SENSITIVITY STUDIES Baselino caso lower plato - SBLOCA Yankeo input file LP-315.I!1 Y Al;K E E, SBLOCA 7, LOWER PLATE, MAX MEAN DRTliDT-315 INITIATIO11 FAILURE CASE V,ARIATION FROM RASELT!1E CASE PROBABI LI T_Y P ROBABI L1':"I 28 Sigma on Kic = 0.15 of mean 4.13E-03 1.11E-(4 Sigma on Xia = 0.0 Upper shelf toughness = 75 29 Upper shelf Kic & Kia = 75 3.96E-03 6.85E-05 30 Upper shelf Kic & Kia = G5 0.84E-03 7.37E-03 31 Upper shelf Kic & Kia = 70 5.82E-03 1.19,E-03 32 Uprar shelf Kic & Kia = GO 2.77E-03 0.00 33 Sigma en Initial RTndt = 17 2.13E-03 0.00 Sigma on shift = 0 34 Sigma on copper = 0.025 1.76E-03 0.00 Sigma ca nickel = 0.0 35 Rerun of caso #1 1.97E-03 0.00 36 Sigma on Initial RTndt 17 3.57E-03 4.00E-OS
=
Sigma on shift = 0 Min pressure a 1000 psi 37 Sigma on Initial RTndt = 17 4.67E-03 3.74E-03 Sigma on shift = 0 Min pressure = 1500 psi 38 Sigma or. Initial RTndt 17 2.25E-03 0.00
=
Sigma on shift =0 Sigma on copper = 0.025 Sigma on nickel = 0.0 Sigma on fluenco = 0.30 of mean NUREG/CR-5799 E.22
isTTACl!ML:;T &3 V I S A--I I 110TICE OF CC REECTIC11 Date September 17, 1990
Subject:
Output of Summary Tablo Problem:
When a large number of initiaticns and arrests occur the program can ter::,inato dce to stack overflow.
This is due to a coding error for the parameter list of the call to subroutine WRITEP.
Also, in cert ain cases the su:r nary table giving examplos of initiatien and arrest eventa can greatly exceed the intended list of 50 examples.
This is occurs when a large fraction of the flaw initiations arrest without vessel fallure.
Correction:
The logic has been changed to terminate t.hu tablo when 50 initiations occur, rather than after 50 vessel failures.
The following changes to the-Fortran coding are recensended in the main program Before 550 WRPSUlITOT) = WRPSUM(ITOT) + WRP(ITOT)
IF (11F. uE. 5 0 ) C ALL WRITEF (1 F,1;11F,111 )
580 WRPSUM(ITOT) = WRPSUM(ITOT) + WRP(ITOT)
IF (liF. LE. 5 0 ) CALL WRITEP (11F,1;11F,111)
After:
550 WRPSUM(ITOT) = WRPSUM(ITOT) + WRP(ITOT)
I F ( 211. LE. 5 0 ) CALL WRITEP 580 WRPSUM(ITOT) = WRPSUM(ITOT) + WRP(ITOT)
I F (111. LE. 5 0 ) CALL 'n..ITEP Effect:
All prior calculations should be correct.
An output table may be longer than desired.
Scue calculations may abort before failure probability aalcolations are complete.
o nt a c t :
F.A.
Simonen (509-375-2087)
Pacific 1;orthwest Laboratory PO Box 999, Richland, WA. 99352 E 23 NUREGER-5799 l
ATTACliMENT # 4 VISA-II NOTICE OF CORRECTION Date:
Septcrier 17, 1990
Subject:
Tabulation of Initiation and Arrest Events Froblem:
The tabulation cemetimes indicates arrested ficw depths that are less than the initial depth of the flaw.
Also the table of summary statistics has inconsistencies in the numbers of arrests and arrested nonfailures.
Correction: The ingic has been changed to reset flags, which ca cc ce rtain counters to correctly tabulate summary e tatir.t.c s.
The following changes to the Fortran i
coding aro recommended in the main program:
E9foro:
C SET FLAGS FOR FLAW INITIATION AND ARREST,.
INITIA = 0 7ARRST = 0 C
RETURNS HERE TO SIMULATE NEXT FLAW B0 CCNTINUE After:
C RETURNS HERE TO SIMULATE NEXT FLAW 80 CCNTINUE C
SET FLAGS FOR FLAW INITIATION AND ARREST INITIA = 0 IARRST = 0 Effect:
Frior calculated probabilities of failure should be correct.
However, probabilities of flaw initiation may be slightly overestimated.
Such errors are greatest when the specified number of flaws per vessel is much greater than one, and when large fractions of initiations arrest before vessel fracture occurs.
Contact:
F.A.
Simonen (509-375-2087)
Facific Northwest Laboratory FO Box 999, Richland, WA. 99352 NUREO/CR 5799 E.24
l ATTACRME"T #5 Caso 1 Case 2 Caco 3 Case 4 PARAMETER OCA-P VISA-II VISA-II VISA-II (original)
(modified)
(original)
- initiatiens 3926 3711 3422 2745
- arrests 983 3104 474 1747
- stable 488 3275 144 1383 arrests
- failures 3438 436 3278 1362 prcbability
.039
.037
.Q34
.027 initiation probability
.034
.0043
.033
.014 failure 1.
Total number of vessels simulated was 100,000 for all cases Case 2 is for original VISA-II which included error in 2.
counting of initiation events.
3.
Case 3 used a.evised and corrected VISA-II.
The only revision was to simulate the error in RTndt only once per vessel, without repeated simulation of this error with each advance of the crack depth during the crack arrest calculations.
Corrections included the error in counting of initiation events and consideration of crack depth for plastic instability calculation.
Case 4 used the orig aal VISA-Il with only the corrections 4.
for counting of initiation events ar.d consideration of crack The repeated depth for plastic instability analysis.
simulation of errm; in RTndt was retained from the original However, the error in shift in RTndt was set equal VISA-II.
to zero in accordance with the reccmmendation and standard practices used in prior applications of VISA-II, 5.
Cases 1-4 are based on a standard deviation of 24F in the shift in RTndt to estimate the error in this parameter.
This practice has been customary for applications of OCA-P, but not for applications of VISA-II.
The two codes make and a different assumptions to simulate the error in shift, standard deviation of zero (versus 24F for OCA-P) iscode.
appropriate to the assumptions made in the VISA-II E.25 NUREG/CR-5799
ATTACHME!1T # 6 ID WALL TEMPERATURES,
'F
- Time, VISA-II AllSYS min.
R/T=6.916 R/T=100 R/T=6.916_
R/T=10 R/Ta1000 10 357 357 364 364 363 20 260 260 267 267 266 30 214 214 219 219
,218 40 197 197 201 200 200 50 192 192 195 195 154 60 188 188 190 190 189 7 ')
179 179 181 181 180 80 166 166 167 167 167 90 153 153 155 154 154 100 154 154 155 155 155 NUREG/CR-5799 E.26
~
ATTACRMENT $6 (Continued)
CD WALL TEMPERATURES,
'T ANSYS Tir.e, VISA-II rni n.
R,/T=6.916 R/T=100_
R/T=6.916_
R/T=10 R/T=1000 10 509 509 510 509 509 20 466 466 470 470 468 30 405 405
^12 410 407 40 346 346 354 353 348 50 300 300 30C 306 -
301-60 266 266 273 271 267 70 241 241 248 246 241 80 222 222 227 225 222 90 205 205 209 207 204 100 189 189 193 192 189 E.27 NUREG/CR 5799
ATTACliMENT #6 (Continued)
ID llOCP TilEFyJ4 STRESS, kai
- Tiro, VISA-II ANSYS min.
It /T= 6. 916 R/T=100 R/T=6.916 R/T=10 R/T=1000 10 34.6 34.6 34.1 34.0 33.6 20 43.5 43.5 44.1 43.7 42.8
{
30 39.0 39.0 40.5 40.1 38.9 40 30.1 30.1 31.8 31.3 30.0 50 21.6 21.6 23.2 22.7
- 21. 5,,
60 15.7 15.7 17.0 16.6 15.5 70 12.7 12.7 13.8 13.4 12.5 80 11.6 11.6 12.5 12.2 11.3 90 10.6 10.6 11.4 11.1 10.3 100 6.8 6.8 7.6 7.4 6.7 NUREGK'R-5799 E.28
NUlti'G CR 5799 O R N L/Tht.1198 2 Dist. RF Internal Distributing
- 1. B. P. Bass 36.
D. J. Naas
- 2. J. W. Bryson 37.
W. E. Pennell
- 3. E. W. Carver
- 38. C. B. Oland 4-13. R. D. Cheverton 39.
C. E. Pugh 14 J. A. Chnard
- 40. G. C. Robinson
- 15. J. hl Corum
- 41. D. K. hi. Shum
- 16. W. R. Corwin
- 42. T. J. Theiss 17-21. T. L L)ickson
- 43. E. W. Whitfield
- 22. J. E. Jorcs Jr.
- 44. ORNL Patent Section
- 23. 1. Keerry. Walker
- 45. Central Research Library
- 24. W. J. hicAfee
- 46. Document Referetre Section
- 25. D. E. hicCate 47-48. Latoratory Records 26-30. J. G. hierkle
- 49. 12txratory Records (RC) 3135. R. K. Nanstad Ihternal Distrlhullan
- 50. L. C. Shao, Director, Division of Enginecting, U.S. Nuclear Regulatory Comminion, Washmgum, DC 20555
- 51. C. Z. Serpan, Jr., Division of Enginecting U.S. Nuclear Regulatory Commission, Washington, DC 20555 52-53 hl. E. hiayfield, Division of Engineering. U.S. Nuclear Regulatory Commission, Washington, DC 20555
- 54. A. Tateada, Division of Engineering, U.S. Nuclear Regulatory Commission, Washington, DC 20555
- 55. D. P. Bozarth, Science Applications international Corporation,708 S.1111nois Ave., Oak Ridge,TN 37830.
- 56. J. W. hiinarick, Science Applications international Corporation,708 S. Illmois Ave-, Oak Ridge, TN 37830.
- 57. F. A. Simonen, Pacific Northwest Lato,atories, P. O, Dox 999, Richland, WA 99352
- 58. L. W. Ward, Idaho Nuclear Engineering Laboratories. EG&G Idaho, Inc.,11428 Rockville Pike, Suite 410 Rockville, hiD 20852
- 59. K. A. Williams. Scienec Applications International Coqoration,2109 Air Park Rd., SE, Albuqacrque, Nhi 87106
- 60. Office of Assistant hianager for Energy Researth and Development,IX)E-OR, Oak Ridge, TN 37831 61-42. Office of Scientific and Technical Information P. O. Box 62, Oak Ridge,TN 37831 63-312. Given distribution as shown in category RF (NTIS-10)
o a Nucti An se cut aioni couwssion i o mm Nmt H 1"CII.'*E '$,M**
l.ac: n - m L'*bu BIBLIOGFIAPHIC DATA SHEET w.ua NUREG/CR-5799 a.,w w,
,w. r,,.,,
ORNL / T'i-11982
- 2. mit no wuo Review of Reactor Pressure Vessel Evaluation Report for Yankee Rowe Nuclear Power Station (YALC No.1735) 3 outairosteuw win
{1992 4.,
.=
thrch 4 liN oH GR ANI NUV0 tit B0119 6 I v 8 6 0F 91t PoH1
$ AU1 HOh N R. D. Cheverton, T. L. Dickson, J. G. Merkle, R. K Nanstad Technical
>ttamucovtuto" * ~ o->
Contributing Authors l
l 2
D. P. Bozarth, J. W. Minarick, K. A. Williams,
F. A. Sinonen, L W. Ward" 3
e r i r.e o 9 w No o a c. ANiz A t iuN - N Au e AN o AD D H t 55 tn'
- ac am o...u.a. o,n,,., a.,
. e 5 *. *., a.,
,.,, rem, a.-< -
4, a.,.
u <.-i..< > c...
rerse ned orce,Wg eM s I
'Sdence Apphtauons Internatuma! Corporatuin. On RWge.1N 378%
Oak Ridge Netiona1 Leboratory
'Saence Apphentuim Internat> mal Coquration. AHiuquerque. NM SMi Oak Ridge, TN 37831-6285
%ahc Nonbe st IWoratory. Rdland, W A WW
- 1daho N uclear I;rgtncenng i Muratones. RMvine, MI) 20A52 sm - me e. ". + s,
- ec N - re=, w *ar*.a v i we, a e ne,wr co-a,s e U LPoNWRING oRGANIZ AlloN - N AM6 AND ADDHtSS (d Mr v.s.
s a a,.ne..a Division of Engineering Office of Nuclear Regulatory Research U.S. Nuclear Regulatory Commission Washington, DC 20555 10 $UPPLEMt N1 ARY NOTis The Yankee Atomic Electric C: apany has perf ormed an Integrated Pres-
- ~'
surized Thermal Shock (IPTS)-type evaluation of the Yankee Rowe reactor pressure vessel in accordance with the PTS Rule (10 CFR 50.61) and U. S. Regulatory Guide 1.154.
The Oak Rictge National Laboratory (ORNL) reviewed the YAEC docun.ent and perforced an inde-pendent probabilistic f racture-mechanics analysis. The review included a comparison of the Pacific Northwest Laboratory (PNL) and the ORSL probabilistic f racture-mechanics codes (VISA-II and OCA-P, respectively). The review identified minor errors and one signific. ant difference in philosophy. Also, the two codes have a f ew dissimilar periph-eral features. Aside from these differences, VISA-ll and OCA-P are very similar and with errors corrected and when adjusted for the difference in the treatment of fracture toughness distribution through the wall, yield essentially the name value of the con-ditional probability of failure. The ORNL independent evaluation indicated FT values screeningcriterhanda considerably greater than those corresponding to the PTS-Rule f requency of f ailure substantially greater than that corresponding to the " primary acceptance criterion" in U. S. Regulatory Guide 1.154.
Time constraints, however, pre-vented as rigorous a treatment as the situation deserves.
Thus, these results are very preliminary.
' u n *i m.o n itartu m u x t v woHorot scwe r oH s ro., ~.
r..
.w.w m.-.u n -,-,m
~,.wn,
Unlimi ted pressurized thermal shock fracture mechanics
- i. wo. u*m ith PWR pressure vessels Yankee Rowe reactor vessel w.,
Unclassified Unclassified
- ift NUMSiH ot FAGl$
16 PHict NHC F ONU 3D fJ 89)
THIS DOCUMENT WAS PRINTED USING RECYCLED PAPER
a UNITED STATES uu = >ote, o m.au NUCLEAN REGULATORY COMMISSION N5'*"jg~'5r**
WASHINGTON, D.C. 20E55-rimwi, c n -
s OFF401At BUSINC$s MNALTY FOR PRIVATE USL $M
.l i
?
i s
t t
t
+
5
- I a
1
'h 4
9 i
F n
?
l I
a
_,.g
.ry-M W< P N1'-'Y*"v w-
"i-'1rM'Ts*'Ovt'+FwPY"'
T
- D W
1'WD'W' I
Y
" ' "