ML20042B462
| ML20042B462 | |
| Person / Time | |
|---|---|
| Site: | Perry  |
| Issue date: | 03/31/1982 |
| From: | GILBERT/COMMONWEALTH, INC. (FORMERLY GILBERT ASSOCIAT |
| To: | |
| Shared Package | |
| ML20042B461 | List: |
| References | |
| 2388, NUDOCS 8203250337 | |
| Download: ML20042B462 (200) | |
Text
{{#Wiki_filter:March, 1982 GAI Rsport No. 2388 PE12Y STAINLESS STEEL CLADDING EVALUATION REPORT Perry Nuclear Power Plant Units 1 and 2 - The Cleveland Electric Illuminating Company Prepared by: - Gilbert Associates, Inc. P. O. Box 1498 Reading, Pennsylvania 19603 8203250337 820315 PDR ADOCK 05000440 S PDR Gdbert/Com r.n=eelta
TABLE OF CONTENTS Sdetion Item Title Pg 1.0 SYNOPSIS 1-1
2.0 DESCRIPTION
OF PROBLEM AND 1 INVESTIGATIONS UNDERTAKEN 3.0
SUMMARY
AND RESOLUTIONS 3-1
4.0 REFERENCES
4-1 Attachments 2-1 SUPPRESSION POOL CROSS SECTION - PERRY NUCLEAR POWER PLANT . Appendix 1 ANALYSIS OF CRACKS IN STAINLESS STEEL CLADDING 11 FRACTURE AND FATIGUE EVALUATIONS III- STRENGTH OF STRUCTURES IV. EVALUATION OF POTENTIAL CORROSION PROBLEMS IN THE SUPPRESSION POOL GeartIConunenessith
1.0 SYNOPSIS Liquid penetrant exaninations of welds in the containment vessels ani drywell vent structures in the region of the suppression pool at Perry Nuclear Power Plant Units 1 and 2 revealed a nunber of indications in the stainless steel cladding adjacent to the weld seams. Because of the recurring nature of these indications, an investigation was conducted to determine the extent, nature and significance of these indications. Metallographic examination disclosed that the indications werecaused by intergranular fissure,s in the heat affected zones of the stainless steel cladding. Although the stainless macerial was found to have a large amount of carbide precipitated at the grain boundaries as a result of heat treatment following cladding, evidence indicated that the fissuring resulted from hot cracking during welding rather than by corrosion. Analysis using the design loads indicated that the fissures would not propagate to a critical size by fatigue nor would they initiate fracture in the structures. The potential for corrosion of the sensitized stainless was found to be small because of the relatively low temperature in the suppression pool; however, crevice attack on the underlying carbon steel at unrepaired fissures was considered a possibility. l As a result of this investigation, the following repair plan was devised: , 1) Coat weld areas with a. history of fissures in the l l drywell vent structure and containment vessel with l ramset/ common u
thermally sprayed aluminum with an epoxy top coat.
- 2) Monitor the effect of the suppression pool environment
~
on the materials by the use of samples placed in the pool. The findings of the investigation indicate that this plan will result in structures which will fulfill their specified function with no reduction in integrity or performance. l l l i Gdbert/Commoneselth t I 1-2
t-
2.0 DESCRIPTION
OF THE PROBLEM The containment vessels of Perry Nuclear Power Plant Units 1 and 2 are pressure retaining structures each composed of a free standing steel cylinder with an ellipsoidal dome, secured to a steel lined reinforced concrete foundation mat. The free standing portion of the containment vessel is supported by and anchored into the foundation mat, and is designed, fabricated and erected in accordance with the requirements of ASME Code Section III for Class MC components (1974 Code inicuding the Summer 1974 Addenda). . The lower 18'-6" of the vessel forms the outside of the suppression pool (120 feet diameter); the inside of the suppression pool is formed by the vent region of the drywell wall. (See attachment 2-1). The vent structure is an approximately 5' thick right circular cylinder composed of two concentric cylinders fabricated of 1" thick plate of SA-516 Grade 70 clad with a nominal 10% stainless steel. The upper 12'-5" from elevation 588'-7" to elevation 601'-0" is clad with stainless steel conforming to SA-240 Type 304L material. i The lower 13'-7" is clad with stainless steel conforming to SA-240 Type 304 material. The area between the two drywell steel cylinders is filled with concrete. The lower 23'-0" of the containment vessel to elevation 597'-10" and the mat liner are also fabricated of stainless steel clad plate. The materials of the lower 13'-6" is _ lh" clad steel piste of SA-516 Grade 70 with a nominal 10% cladding of SA-240 Type 304 stainless steel. The material for the next 9'-6" Geert/Commoneenth 2-1 L
is also 1-1/2" clad plate but in thie case the 10% nominal thickness of stainless steel is of SA-240 Type 304L. The foundation mat is lined with a h" thick clad plate of SA-516 Grade 70 with a nominal 10% of SA-240 Type 304 cladding. The stainless clad plate described above was manufactured by the roll cladding process by Lukens Steel Company and conforms to ASME SA-264. Following the roll bonding operation, the material was heat treated as follows:
- a. Annealing at 2000-20500F, holding 15 minutes per inch and.'
air cooling, then
- b. Normalizing at 1625-1675 F, holding hour per inch (minimum) and air cooling.
During installation of the drywell vent structures, liquid penetrant examinations (PT) revealed indications in the stainless steel cladding material adjacent to some butt wells Similar indications were also revealed adjacent to the fillet weld attaching a doubler plate to the bottom of the containment vessel. All welds in the h" thick floor plate receive 100% PT inspection and have shown no signs of fissuring to date. It was reported (and documented on Field Question No. 9072, I dated 9/16/80 Ref.1) that attempts to remove these indications by grinding and weld repair sometimes resulted in new indications adjacent to the weld repair,_ again in the stainless cladding . Gstart/Commanuseth 2-2 L
material. The Regias III office of the Nuclear Regulatory Commission was notified of this condition by telephone on October 6 and November 6,1980, and by letter dated November 12, 1980 (Reference 2). An investigation into the extent, cause and significance of these indications was begun at that time. An analysis of the occurrence of the penetrant indications disclosed that they were found with greatest frequency in stainless steel material with carbon contents near the high end of the range allowed by the specification; that is, the , indications were usually in material with carbon contents of between 0.06% and 0.08%. Additional PT inspections siso confirmed that fissures were limited to the heat affected zones of welds and to the Type 304 material. No fissures have been found away from welds or in the Type 304L material. Because of the recurring nature of these indications adjacent to welds, four nonconformance reports were issued which described these incidents (References 3 through 6). On 14 May 1981, Newport News Industrial Corporation, the contractor responsible for fabrication and erection of the containment vessel, drywell structure and foundation mat liner, issued a report (Ref. 7) describing their investigation into the occurrence of the indications. The report's conclusion included the following items:
- 1. In the material NNIC investigated, the stainless steel clad layer was highly sensitized, that is, the're was a high Gdbert/Commonesep 2-3
degree of carbide precipitation in the grain beundaries of the material,
~
- 2. The material contained intergranular fissures associated with the heat affected zones of welds in two of the samples studied and
- 3. The censitized condition of the stainless clad layer-was not detrimentn1 to the ultimate tensile strength of the cladding or of the composite plate.
Since the NNIC report was not able to pinpoint the cause, severity , or consequences of the fissuring phenomena, Aptech Engineering Services was requested to assist in the investigation of these issues. Specifically, the issues that were addressed and the associated Appendices where they are discussed are as follows:
- 1. Cause determination - to develop a complete understanding of the problem, what is the cause of the fissuring?
(Appendix I)
- 2. Potential for fracture / fatigue - is there a potential for undetected fissures to grow to a critical size through fatigue or to initiate a fracture of the plate? (Appendix II)
I 3. Strength of structures - does the presence of known or undetected fissures reduce the strength of the containment vessel and drywell vent structure to an unacceptable level? (Appendix III)
- 4. Potential for corrosion - does the sensitized structure in the stainless steel or possible undetected fissures present a problem from a long term corrosion aspect? (Appendix IV)
Gdbert/Commoneenth 2-4
1 h I e DRYWELL WALL [h~:; i .. : EL. 670-6 ..
. . . . . . . . . . J, ,,i,..%. ',:.
L .L v: - fs*v. 4;:
? 2'-2' CONTAIMMENT -
VESSEL ,
=
== =
i'. s" SHIELD
=
T a BLDG. ,n otn m 2g 9 3 jI -
. mj OLUMN J.
8 *-
- $m (TYR l'l
@ .,?
Eu . PLACES) .d _4
- f. -
" gw EL. 60l' 0" EL. 599'-9"
"- "=
EL. 598'- 4" - I-- l[I I
.l.'
, ( .. .l h *-WEIR WALL EL. 597'-ICfd 12'5" 304LCl.ADg b h ,
' V
-i DRYWELL VENT STRUCTURE l
f,. EL 588' 4" - p 3 1 " SST LINER
" EL. 583'-G" EL. 588'-7"M o '
rDOUBLER o :
' FILLET WELD 4 U 9
.D d T
[ cLEAK CHASES g _E CTYP.) "o- - h.
/
~
4v.5 - h, , Y9 l'!, - - ( EL. 514'-10")5@.4 y lp x ,, 1h yy. N:!kD.,gd" 7 y-n.-s l I DOUBLER I" C C PLATE / d FLOOR 504 CLAD SUPPRESSIOh! POOL CROSS SECTION PERRY NUCLEAR POWER PLANT ATTACHMENT 2-1
- 3. 0 RESOLUTION AND SUW1ARY 3.1 Weld Itepair .
Liquid penetrant inspections have revealed indications in the heat affected zones of the clad stainless steel. In some cases minor grinding removed the discontineities. In others, weld repair was performed af ter removal of the indications. it was reported that af ter some of these weld repairs, indications would be found in the heat affected zones of the repairs. Various modifications of heat input and pass sequence were evaluated and while these procedural modifications seemed to improve the success of repairs, they were not completely e f fective. Extensive investigations were initiated to evaluate l this problem.
- During the course of these investigations it was demonstrated that
- 1) the pressence of sensitized material did not have a deleterious effect on the mechanical properties (strength and ductility) of the stainless material or the composite plate,
- 2) that fissures in the stainless steel layer will not cause failure of the structures through fatigue or fracture, and
- 3) that the carbon steel portion of the clad plate furnishes sufficient strength without the stainless cladding for the drywell vent structure and containment vessel.
Geert /Commanuseth 3-1 l
- 4) that the floor plate is unaffected since inspections have revealed no fissuring. .
For the drywell vent structure and containment vessel, the stainless clad layer is only required as a corrosion barrier to protect the underlying carbon steel from the suppression pool environment. To ensure this function in the weld and heat afffected zone, a coating system will be provided as discussed in Section 3.2 below. No weld repair of fissures in the drywell vent structure or containment vessel will be required. No fissures have been found to date in the floor plate. If fissures are found in the floor plate, they will be weld repaired. 3.2 Coating Application Because the possibility of crevice corrosion could not be ruled out (Appendix IV), welds and associated heat affected zones at which fissures have been shown to exist will be coated with a high reliability coating system as described below. This coating system will be applied in bands along weld seams. The I coating will extend a minimum cf 4 inches from the weld fusion lines each side of the weld to ensure that all fissures have been covered. Fissures have not been found more than 2 inches from the fusion lines. The coating system that has been selected consists of thermally sprayed aluminum top coated with epoxy. Similar systems have i provided excellant service in condensate storage tanks and other l environments similar to the suppression pool. In addition Catert /Conunomseep 3-2
they have been subjected to DBA testing at Oak Ridge National Laboratories (ORNL) with no evidence of any signs of failure.
~
The ability of this system to provide protection to carbon steel under fissured stainless steel cladding is currently being verified in tests bein conducted under the suprevision of Aptech Engineering Services. Because of the outstanding history of good performance from this service, no problems with confirming its ability to fully protect the carbon steel is anticipated.
- 3. 3.3 Inservice Monitoring A plan for the placement of corrosion monitoring coupons'in the suppression pool and periodic evaluation of these coupons is being developed by Aptech Engineering Services. These coupons will represent the sensitized stainless steel cladded to carbon steel and the fissured stainless steel cladded plate protected by the coating system described above.
3.4 sum!ARY I Following the discovery of liquid penetrant indications adjacent to come welds in the stainless steel portion of clad plate which make up the suppression pool structures of the Perry Nuclear Power Plant Units 1 and 2, a study was performed to j determine the cause and the effects of these indications on 1 the safety and performance of the structures, and additionally, to determine any remedial actions needed to resolve questions of reduced integrity of the affected structures. The study concluded that the indications were the result of intergranular fissures, probably resulting from hot cracking which occurred l Geerucomamou 3-3
during wel, ding. While corrosive attack.on the stainless steel in service was concluded to be very remote, attack on the underlying carbon steel through fissures in the, stainless steel was found to be a possibility. To prevent attack on the carbon steel, a coating consisting of thermally sprayed aluminum top-coated with epoxy will be applied to all weld areas at which fissuring has been found or is suspected. Since the presence of fissures in the stainless steel was shown to have no detrimental effect on the design load carrying capacity of the affected areas, this additional corrosion protection will result in the strength, performance, and safety characteristics originally intended for these structures. Gdbert /Commoneesith 3-4
4.0 References
- 1. Field Question 9072 dated March 16, 1980. .
- 2. Letter dated November 12, 1980, to James G. Keppler, Office of Inspection and Enforcement, U.S. Nuclear Regulatory Commisiion, from D.R. Davidson, Vice President, The Cleveland Electric Illuminating Company, Docket Nos. 50-440; 50-441.
- 3. Nonconformance Report No. 17-149 on the Unit #1 Containment Vessel Doubler Plate Region.
- 4. Nonconformance Report No.17-138 on the Unit #2 Containment .
Vessel Doubler Plate Region.
- 5. Nonconformance Report No. 96-471 on the Unit #1 Drywell Vent Structure Region.
- 6. Nonconformance Report No. 96-769 on the Unit #2 Drywell Vent Structure Region.
- 7. " Report on Stainless Steel Clad Plate Material Used in Construction of the Perry Nuclear Power Plant, Units #1 and
#2", Revision A, May 14, 1981.
l Gdbert /Commoneestth ( . _ . - - , ,,
I I APPENDIX I l i i i f I I l l
AES 8106262-1 Final Report RPTECH engineering servicer,Inc enoiseeRiuo CO~SuL1 ANTS 795 SAN ANTONIO ROAD . PALO ALTO . CALIFORNIA 94303 (415)858 2863 APPENDIX 1 ANALYSIS OF CRACKS IN STAINLESS STEEL CLADDING AT THE PERRY NUCLEAR POWER PLANT Prepared by Terry W. Rettig APTECH Engineering Services, Inc. 795 San Antonio Road Palo Alto, California 94303 l [ Prepared for l Gilbert Associates, Inc. Post Office Box 1498 Reading, Pennsylvania 19603 ATTN: Paul B. Gudikunst, Project Manager l February 1982 Services in Mechanical and M etallurgical engineering, Welding, Corrosion, Fracture Mechanics, Stress Analysis
- I
VERIFICATION RECORD SHEET -
Title:
Analysis of Cracks in Stainless Steel Cladding at the Perry Nuclear Power Plant (AES 8106262-1) Originated by: ( ( n ( I 5-9 '52 Terry W. Rettig 9 Approved by: 'y[ [ 3-f N v i Geoffrey r. gan I ;
- f . p,
, Verified by:
j 'l D RogerH.Richm/n Quality Assurance Review by: - ", Russell C. Cipolla Quality Assurance Approval: _ 9h f//at IZ - Edwin R. Pejack O -r m
-- n , . - - - - ,
1 Q 9 INTRODUCTION AND BACKGROUND APTECH Engineering Services was asked to conduct a limited failure analysis investigation into the problem of fissures detected in the containment and dry-well vent structure of the Perry Nuclear Power Plant. The fissures were detected in the heat affected zones (HAZ) of some of the welds by dye penetrant inspec-tion. A more detailed examination had previously been performed by Newport News Industrial Corporation, and therefore, APTECH restricted the investigation to a limited metallurgical examination of boat sanples removed from the structure and previously mounted, polished, and examined by Newport News. Part of the suppression pool at Perry Nuclear was fabricated from a composite plate comprised of Type 304 stainless steel roll bonded onto carbon steel base plate with a thin nickel interface to promote good bonding. The composite was heat treated as follows after roll bonding:
- Anneal at 20000-20500F, holding one-quarter hour per inch, then air cool eNormalize at 1625 0 -16750F, holding one-half hour per inch, then air cool OBSERVATIONS Metallographic examination of the submitted mounted boat samples in the "as- ;
received" condition confirmed the presence of cracks in Boat Sample No. 5 (containment) and in Boat Sample No. 6 (drywell vent structure). Cracking was not detected in any other boat samples. The grain boundaries in all the samples appeared to be severely sensitized, that is, extensive intergranular chromium carbide precipitation was observed. Figure 1 is a photomicrograph taken in the HAZ of the weld from Boat Sample No. 6 in the as-received condition (etchant 10% oxalic, electrolytic). The weld fusion line (not shown) is just to the left of the cracks. Some of the cracks can be seen to extend the full thickness of the cladding, and at one location, a crack penetrates into the carbon steel. Smaller cracks are also present adjacent to the clad interface and do not appear to be interconnected [_
~
2 to the through-clad cracks. The contour of these cracks was almost parallel to the contour of the weld fusion line. In this plane of examination, it could not be determined if these cracks are separate or merely branch extensions of
~
the through-clad cracks. It is important to distinguish this' feature because if the small cracks are not connected to the surface, the possibility of stress corrosion cracking (SCC) can be ruled out. Figure 2 is a composite photomicrograph after repolishing and then etching electrolytically in 10% oxalic acid, showing the crack pattern in a different plane from that in Fig.1. The network is still visible, but in this plane the cracks appear to be even more independent of the large cracks. The cracks are also more open near the clad interface than in the middle of the clad. In this plane, no penetration into the base metal was observed. If the cracks were initiated at the clad interface, it was thought that other regions of Boat Sample No. 6 might contain only subsurface cracks. A dye pene-trant examination was performed on the remaining piece of Boat Sample No. 6 to select a location for sectioning away from any surface indication. A'small surface indication was observed near one end of the boat sample. A section was taken in a " clean" region that was about one-half inch away from the dye pene-trant indication and examined metallographically. A subsurface crack about 0.015 inch long was observed to extend from the clad interface into the cladding (Fig. 3). No other cracks were found in this section. The observation of an isolated crack not connected to the surface and entirely separate from the large, through-clad cracks, confirms subsurface cracking and suggests that the through-clad cracks originated at or near the interface and propagated toward the surface. The subsurface crack is strong evidence that SCC was not the cause. This is supported in the Newport News report, which mentions that corro-sion products could not be found in the fissures (1_). DISCUSSION Numerous other examples of intergranular cracks unrelated to corrosion in austenitic stainless steel can be found in the literature. Peckner (2) showed intergranular fissuring in tubing made from 18-8 type material that was a result of bending in the hot-short range. In Gleeble tests performed by L. .n -
3 Yamaguchi, et al., (3_) hot cracks were observed to propagate intergranularly in Type 3C.' stainless steel . It was suggested that columbium compounds with phosphorus and silicon precipitate at grain boundaries. Kobayashi, et al., (4) demonstrated that intergranular cracks could be induced at a strain-of less than one percent in the HAZ of Type 316 stainless steel welded with a heat inpct of 12 KJ/cm (TIG) in a varestraint test. Moreover, they found that the resistance to cracking is improved with Nuclear Grade (low carbon) 316 stainless steel. Newport News showed that the cladding is sensitized, and other work has shown that the residual stresses in austenitic cladding on carbon steel are high (5_). These two conditions are consistent with hot shortness in the clad layer. It is then probable that intergranular cracks are induced, during subsequent weld-ing operations, by the scenario outlined below or by some similar mechanism. Damage is believed to occur as a weldment is being completed and the final weld passes are made adjacent to the stainless steel clad layer. In the heat-up portion of the final weld cycle, the cladding and the adjacent layer of base plate both expand. Even though the clad expands more than the base plate, stresses in both parts are relieved by plastic flow at high temperatures. During the cooling cycle, the base metal will cool faster than the cladding due to better heat transfer and much larger mass. As a result, it shrinks to approxi-mately final dimension and becomes stronger while the cladding is still hot. The cladding now tries to contract as it cools and is constrained by the base metal . This effect would be accentuated by the interface where the constraint is most pronounced. The result is that the clad is pulled during cooling while still at a relatively high temperature and cracking occurs at the interface. Hot shortness or hot cracking is a condition characterized by greatly reduced ductility in certain elevated temperature ranges. Austenitic stainless steels have two such ranges: 1) in the temperature regime of carbide precipitation at grain boundaries, and 2) at the even higher temperatures of delta ferrite formation (2). Both of these potential problems are time dependent and would be less likely to occur during welding of stainless steel that is not sensitized. However, the clad has already been sensitized as a result of the cladding pro-cess and, thus, is probably subject to cracking under the right conditions of strain rate and temperature. f s hue
,w
4
SUMMARY
On the basis of this limited investigation, it is concluded that the observed intergranular cracks were associated with welding a material that was already
~
sensitized and, thus, brittle at certain elevated temperatures. Stresses arising from differential thermal expansion during welding were the specific cause of the cracks. The situation may have been aggravated by the use of high arc energies. I a 0 O +
5 REFERENCES
- 1. Cobb, G. M., " Report on Stainless Steel Clad Plate Material Used in Con-struction of Perry Nuclear Plants, Units 1 and 2, Revision A," Newport News Industrial Corporation, Newport News, VA (1981).
- 2. Peckner, D., and I. M. Bernstein, eds. , Handbook of Stainless Steel, McGraw-Hill Book Company, New York (1977J, Pp. 30-7 to 30-8.
- 3. Yamaguchi, S., F. Kurosawa, and S. Abe, " Hot Ductility and Grain-Boundary Precipitates in Austenitic Stainless Steels," The Fundamental Research Laboratories, Nippon Steel Corporation.
- 4. Kobayashi, T., et al., " Stress Corrosion Cracking Resistance, Mechanical Properties and Weldability of Nuclear Grade 316 Stainless Steel," Seminar on Countermeasures for BWR Pipe Cracking, Session 6, The Electric Power Research Institute (January 1980).
- 5. Kupka, I. , P. Mrkous, and C. S. S. R. Skoda, "Some Remarks on the Analysis of Stress-Corrosion Cracking of Austenitic Stainless Steel Cladding," In Stress Corrosion Cracking Problems in Primary Pressure Systems, Interna-tional Atomic Energy Agency Report No. IWG-RRPC-78/2, Palo Alto, CA (1976),
Pp. 138-144. r r 1 f J -
I
- h
~.. )Ed
~
yy' % ..,e . =7; y . : v; f (f. $'[*Q.'f[gG . db . -
. w,, .
r w
. -e s
'; , --, nr~.( 2,: sv;,,y, rv -- n. .;. . : .
,, .s
,, 'm L .? efGW~'".%
L '.,:g Q.z w
.. y c(
<I h C '('3.1 M..M g :s;<,%# '"*
FIGURE 1 -?hl.(g'.t J I 12- ' ' Composite photomicrograph from boat sample 3',.' %. j.: I-4t, ' .$.. s, . ~" C' ?)r4,. '.' r% . . .,;-/@, .. ' {,7? $- No. 6 showing the cracking network in the clad layer. NQ ,;"m_ g 4 MAG.: 100x M I' , D ;' ,' - A 7i P ~ w+ g a: ;
~
4 , > . o e o , Q:c.. -
=3,j.
.s ..
.a G,,
- y. : .
i, ci~',k j U " [4,./ ., ( , g ,,. ~ ' '
- s. m .+ 2. .1
.- . x
'"'-K, j//p .' ', .
gf,P ,
;'r y - '-
n%.93 .o' .
- Y,.g._ g. ;c- A . - ,;s A,e .
- 5. t-wLe4 \\ /K' , , 'g h[? .f
>.~ : ; '^
'$.y -\
L~ )';' ~f < '.y.k ,-G Au . / \ ,-Q e .
. - f, y-1.-g 4 , ,-
l ,Ty.- 'l.a , :r 3 t
- i9
- . ,w4; X y '
y (6 y:. p',dy, a . c ~: - i $ . + s(
- .r , , i:.
,;\-~. %<g, ;itg. U
' ;* /, . .
y 70 / i ,., - S/. N- .- +#- . .x .gd . %. ff- . '-
'w s .. 1-)
r,
*Q. n
+1 }., a:.. N,Yy~y+16
, : ' ' p;}(b:g .i ' *a - -
.w.;.: .
.c w;;h% ,n'a . :
w). ;- y;, r. bkq.. \ v3.f ;;f8 c. y
- 9: m-
.j~g ,7
, e hY.;-' X .';-Y ^
9.
. c . .
t,O-
, # g v.~< s g,- 3 .
~; .
*! I'~'W***- -
t __
,4 wn s' . , f >>u w '* e p 3 ; # r e A
))mg,},d.f,
' j ~
, e4. e, 7 '- , ,,. a.
, ,1 ;
h. h
' c c.
= u,R>:)u4'k. aWo r
?i 5
(. :..$_ O_ L'kh?av-y*y
- QQychn'n.f.)J-i:,.&s.g.
1s ,
' );,
,:. , ~
"f
'z;.4,
} h' ,
; j9f, .~ i p.. .
ef.lN Aq,@*"L'-
/ '.,! a:' .
,b ..g?
E . N 6 e )
.h %
ot 0 N x5 m 0 eo 0d l r 1 e pf c m : u at .d sn Ge e Ar f i td an oe bp e M( Q fd on n i G. oe ib g eo rt t r na ee rp
. ,* 2 ep g y, E f a f' , <:,0+:.).@$.A.mT4
\ T R f
*p C '. ' 3,.ou: 2 U i s Q
G dk L
. I c
%-,;.L,.. N::. ;(d. ' ,h.m i-
).z ~(.
F aa r n .. , mc y c;.
' s o g ., o , ,
re f c
'.*1-...i.
a
'g
~
. f..', . , , ...g ,s9 > ;:, . - ,. hf prs s'q y y,>y.-
auk
- 3. ' .
y &' .:;s% f
,h. '*. 2 c r
,. .,yL,- ,7,l'r f rsc gb a our rsc y& f['y, y},2,flt .~
~7 Q
c i ed
- r. ,.: * .
?
- n.W mh e Q
N QQ ! ;f . .. + ,9%.N1
, . :: t ott oee c
., ,q4 f;
% ': . '. .f9-5 h nn pan epc l o
[ * ,. ., .. ~. Q%. ,.%r,*,
- t ise e~j f. ~n.t' , 'r f- sic n'
Mmz oh a ptf m onu CI s r l , t
][ ,
- r i
' ,- ^-
A'
*[
- Y _
f' - -
'h .
h
i I !s a i w , 9 J I 3 .
. \ .
( .
,.,-e .;-l 'l .
.b *
- 7R,,"-yq ::
*; 7 - {;. :<a , .,
t
. ,..2 w.2 , .
ql , . r.,, -l.W ,.. ,, ._ f/ ' {' ; _
- '.; ,e ,
- c. . :. .. , . . . . . . -
1 I 1 f, fx.~FWLp@t: m.m TV55lY[?en? w .c.w A zVf eauxsa i FIGURE 3 i Photomicrograph showing a subsurface crack that was i located approximately one-half inch from any surface indication. MAG.: 100x . l l N b i i e 5 r 1
- - . . , ,, . - , - . ~ , - . . . , - , . , _ - , , , , , - - , , , , . - - , , - - - - - - , -..,n_, , , . - , . - - - . . . _ _ , _ _ _ , - . - . - _ . , - - - - . . - - -
J - APPENDIX 11 l l l i
' " ~ -
^ ' - - - - - - - ~ - - ._m _ _ _ _ _ _ _ ,
AES-81-11-88 APTECH engineering services,inc e~olueeRino COusuL1AurS 795 SAN ANTONIO ROAD . PALO ALTO . CALIFORNIA 94303 (415)858 2863 APPENDIX II THE SIGilIFICANCE OF SENSITIZED STAINLESS STEEL l'ATERIAL If4 DRYWELL VENT AND CONTAINfENT STRUCTURES IN THE PERRY NUCLEAR POWER PLANT
- FRACTURE AND FATIGUE EVALUATIONS -
I by: Geoffrey R. Egan Warren P. ficNaughton Jeffrey D. Byron Alan E. Ellis Prepared for: Gilbert Associates, Inc. I Post Office Box 1498 Reading, PA 19603' Attn: Paul B. Gudikunst Project f4anager l' Novenber 1981 4 Services in Mechanical and Metallurgical engineering, Welding. Corrosion, Fracture Mechanics, Stress Analysis
VERIFICATION RECORD SHEET
Title:
The Significance of Sensitized Stainless Steel Material in Drywell Vent and Containment Structures in the Perry Nuclear Power Plant - Fracture and Fatigue Evaluations
- Originated by: bk #
h Warren P. McNaughton
~
?
Approved by: /h Geoffrey R / Egan Verified by: 1 th/Wf ieffrfLl3 rover Quality Assurance Review by: ,, ~. 2/r 7/.t,1 Russell C. Cipolla l l Quality Assurance Approval: I /Z ID l Edwin R. Pejack O
+=
i TABLE OF CONTENTS _ Section Title Pace SYN 0PSIS i
1.0 INTRODUCTION
1-1 2.0 ANALYSIS METHODS 2-1 2.1 Fracture Mechanics Background 2-1 2.1.1 Linear Elastic Fracture Mechanics (LEFM) 2-1 2.1.2 Elastic-Plastic Fracture Mechanics (EPFM) 2-5 2.1.3 Limit Load Analysis 2-6 2.1.4 Summary of Fracture Mechanics Background 2-7 2.2 Fatigue Loading 2-8 2.3 Evaluation of Fissures - Flaw Modeling 2-9 2.3.1 Flaw Orientations 2-9 2.3.2 Flaw Distributions 2-12 References 2-15 3.0 ANALYSIS OF STRESSES 3-1 3.1 Drywell Vent Structure Primary Stresses 3-1 3.1.1 Vent Region 3-1 3.2 Containment / Doubler Plate Primary Stresses 3-4 3.3 Drywell Vent Structure Residual Stresses 3-6 3.3.1 Transverse Stress Distribution - On the Plate Surface 3-8 3.3.2 Longitudinal Stress Distribution - On the Plate 3-8 Surface 3.3.3 Transverse Stress Distribution - Through Thickness 3-8 Variation 3.3.4 Longitudinal Stress - Through-Thickness Variation 3-9 3.4 Containment / Doubler Plate Residual Stresses 3-9 3.5 Cyclic Stresses 3-9 3.6 Combined Stresses: Summary 3-9 i 3.7 Grinding and Bonding Residual Stresses 3-13 References 3-18 l 4.0 FATIGUE GROWTH RATES 4-1 4.1 Introduction 4-1 4.2 SA516 Grade 70 4-2 4.2.1 Effect of Environment on Fatigue Growth Rate 4-4 4.2.2 Note on AK Threshold Effects 4-6 4.2.3 Temperature Effects 4-6 4.2.4 Geometry Effect 4-6 4.2.5 Effects of Composition 4-7 4.2.6 Cycle Characteristics 4-7 4.3 Fatigue Growth in Type 304 Stainless Steel 4-9 4.3.1 Effect of Sensitization on Fatigue Growth Rate 4-11 4.3.2 Effect of Environment 4-11 g
ii (TABLE OF CONTEPRS - Continued) Title Page Section _ 4.3.3 Effect of AK Range 4-13 4.3.4 Effect of Temperature 4-13 4.3.5 Effect of Geometry 4-13 4.3.6 Effect of Compositional Variation 4-16 4.3.7 Effects of Cycle Parameters 4-16 4.4 Fatigue Growth in Stainless Steel Weldments 4-17 4.5 Fatigue Crack Growth in Carbon Steel Weldments 4-20 4.6 Summary of Fatigue Growth Rate Data 4-20 References 4-24 5.0 CORROSION ASPECTS 5-1 6-1 6.0 FRACTURE TOUGHNESS 6-1 6.1 Intrnduction 6.2 Background 6-3 6.3 Toughness Data for Drywell Vent Structure 6-7 6.4 Toughness Data for Containment Materials 6-8 6.5 Crack Opening Displacemcnt (C0D) Values 6-8 Oth'er Material Properties 6-12 6.6 6-13 References 7-1 7.0 RESULTS OF ANALYSIS 7-1 7.1 Introduction Limit Load Analysis 7-2 7.2 7-6 7.3 Drywell Vent Seam Region Drywell Vent Structure - Vertical Welds at Plate Seams 7-10 7.4 7-10 7.5 Containment Doubler Plate Region Elastic-Plastic Fracture Mechanics (EPFM) Results 7-16 7.6 7-29 References
SUMMARY
AND CONCLUSIONS 8-1 8.0 A-1 APPENDIX A: Controlled Documents Calculation of Drywell Vent Seam and Plate B-1 APPENDIX B: Edge Stresses APPENDIX C: Containment Stress Summary C-1 APPENDIX D: Residual Stresses Due to Welding D-1 APPENDIX E: Statistical Evaluation of Drywell Vent E-1 Material Properties F-1 APPENDIX F: Statistical Evaluation of Containment Material Properties
SYt40PSIS This report sunnarizes the results of f racture mechanics and fatigue evaluations which were performed for the Perry fluclear Power Plant Project. Their purpose was to determine the significance of sensitized stainless steel clad material which comprises the drywell vent structure, containment and suppression pool liners. Fissures have occurre<4 in the heat affected zone of the welded clad naterial in both the drywell vent and containment structures. The effect of defects of a size equal to a worst case postulated fissure has been assessed. Stress data were supplied by Gilbert Associates for both the drywell vent and containment. Both cyclic and steady state stresses were evaluated. Postulated flaw dinensions were compiled from field and laboratory data. Material properties of interest, such as strength, toughness and crack growth rates were compiled from Certified liaterial Test Reports and compared to generic data for the same material obtained f rom the open literature. It is concluded that crack growth by fatigue is snall over the design plant lifetime, even assuning conservative stresses, large initial flaws and worst-case propagation rates. The applied stress intensity values reached during such growth are nuch less than the critical value to cause fracture. However, because of the potential for crack growth by nechanisms other than l fatigue (discussed in supplenental reports), the fracture behavior of much l larger defects was investigated, including assuned through-wall flaws. It is shown that relatively large through-wall flaws can be tolerated before critical fracture conditions are reached.
1-1
1.0 INTRODUCTION
The drywell vent and containnent structures of Perry Nuclear Power P-lant (PNPP) are constructed of stainless-steel-clad material. The clad, classified ASME SA-264, consists of a 10% layer of SA240, Type 304 stainless steel roll clad to SA516 Grade 70 carbon steel. After cladding, the plates are annealed and normalized. The clad plates are field joined to one another by a two step weld procedure which joins the carbon steel portions followed by a two pass overlay of stainless naterial. During inspection of the vertical and vent seans in the drywell and doubler plate to containment wall joints, " fissures" were detected by liquid penetrant inspection techniques. The indications, located in both units, were oriented both transverse and parallel to the weld in each geometry. The finding led to issuance of Nonconformance Reports 96-471, 96-769, and 96-493 for the drywell vent, Units 1 and 2 respectively and Report numbers 17-149 and 17-138 for the containment in Units 1 and 2. (These docunents are Aptech Engineering Services Controlled Docunent Numbers 22, 23, 93, 20 and 21 respectively. Future references to input provided by Gilbert Associates, Inc. during the course of the work will be by the appropriate controlled document (CD) number. The complete listing of these docunents is provided in Appendix A.) A concern is that if unrepaired, or incompletely repaired, the fissures could lead to early structural failure. This report addresses the potential for defect growth by a fatigue mechanism and the potential for subsequent failure
- by fracture. The results can be conbined with evaluations of other potential l f ailure mechanisms to characterize the significance of the fissures.
The remainder of this report consists of seven sections. Section 2.0 outlines the analysis methods that have been used to evaluate the fissures. The next four sections introduce and discuss input to the analytical model; they are the evaluation of stresses - both applied and residual stresses (Section 3.0), i fatigue growth data (Section 4.0), a note on corrosion behavior (Section 5.0) l and characterization of f racture toughness (Section 6.0). Section 7.0 provides the results of the work performed. Conclusions and recomendations are provided in Section 8.0. l l
i 1 2-1 2.0 AfaLYSIS 11ETHODS The following sections discuss those aspects of fracture nechanics And fatigue theory which were used in the analysis of the present problem. A presentation of general fracture rwchanics background (2.1) is followed by a discussion of nethods of analysis to assess fatigue growth (2.2) and details of flaw noceling (2.3). 2.1 Fracture rechanics Backcround The f ailure behavior of structures under nonotonic (slowly increasing) loadinc can be classified into three regines in which a specific type of f ailure node is appropriate. These three regines cover brittle f racture, ductile fracture and plastic collapse. The disciplines required to assess these regines are:
- 1) Linear Elastic Fracture Mechanics (LEFit) - The structure f ails in a brittle nanner and, on a macro scale, the load to failure occurs within noninally elastic loadina.
- 2) Elastic-Plastic Fracture 11echanics (EPFf:) - The structure f ails in a ductile nanner, and significant stable crack extension by tearing nay precede ultinate failure.
- 3) Fully Plastic Instability (Linit Load Theory) - The failure event is characterized by large deflections and plastic strains associated with ultinate strength collapse.
A diacran that shows the relationship between critical or f ailure stress and flaw size for the three f ailure modes is given in Fig. 2-1. The shape anc position of the f ailure locus will depend on the f racture toughness (K ) and i strength properties (og ) of the naterial, as well as the structural ceonetry and type of loading. r 2.1.1 Linear Elastic Fracture Mechanics (LEFP.) i l The principles of linear elastic fracture nechanics (LEFM) are applied to [ assess quantitatively the conditions for brittle fracture. Brittle
\
\
\
Limit Load
,P /
N g N r Llastic-Plastic (EPFil)
'?
.]
c
/
N[ ' g Fracture (LEFli) N u N
/ f Limit load C n r lleg s
\
/
/ N
<// Frac ture e\ N
'm 0
?z As\MhN 2a/l g
fion-Dimensional flaw Depth, 2a/t Figure 2 Schematic Showing the Relationship Detween failure Stress and flaw size for Two Limiting failure flodes
2-3 fracture consists of two separate events: (1) the initiation of a crack, and (2) the subsequent propagation of the crack to conplete f ailure. Each of these events, initiation and subseauent propagation, has_ different characteristics. For ferritic structural steels of the SA516 type and carbon nanganese weld metal of the E7018 class, the resista1ce to a propacating fracture is usually lower than the resistan'ce to fracture initiation under slowly applied loads. This is because steels of this type are sensitive to loading rate; the high loading rates associated wit'h a running track lead to higher yield strength and, hence,' lower values of fracture toughness. In constant load situations, therefore, continued crack propagation is expected once the fracture has initiated. For this Jreason, no attenst is made to evaluate the characteristics of the propagating crack af ter it has initiated, and the criterion thf f racture initiation is used as the definition of f ailure in the fracture-analyses. Fracture initiation occurs at a de ,en the crack driving force exceeds the naterial's inherent .
.ance to crack initiation, or fracture touchness. The crack driving force is a function of the stresses acting on the defect and the aeonetry of the defect. The stresses which act on the defect include both prinary (applied) stresses anc secondary (internal) stresses. Exanples of secondary stresses are residual stresses and thernal stresses that are in equilibrian across the section. The nanner in which a structure will fail will be determined by the interaction of the defect geonetry, loading, and material. toughness.
In linear elastic f racture nechanics, the nost useful paraneter for characterizing the behavior of cracks is the stress intensity f actor L7, which describes the magnittde of singular stresses ahead of a crack in a linear elastic body loaded in tension. For loading normal to the crack plane (l' ode 1), fracture initiation occurs when the applied stress intensity f actor, X;, equals or exceeds sone critica! value, which is called the fracture toughness of the material. The applied stress intensity f actor can be written in the forn: K;
= C . 'T (2.1)
2-4 where c is the acting stress, a is the characteristic flaw dinension, and C is a paraneter which accounts for the flaw shape, structural geonetry, and the type of loading. In general, C is a function of a and-in many cases nust be evaluated nunerically. Fracture will occur under quasi-static loading when, . Y,I > YIC (2.2) (i.e., when the applied stress intensity factor eauals or exceeds the static fracture toughness, Klc). This neans that the occurrence of fracture is controlled by: (1) the stress level. (2) the flaw size, and (2) the f racture toughness. For snall flaws, low stresses and high touchness, the applied K will not reach Kg , and fracture will not occur. These relationships are relevant for naterial properties determined under plane strain, linear elastic conditions. To deternine the significance of the fissuring found to date, it is necessary to know the naterial f racture toughness, acting stress level and distribution of expected defect sizes and shapes. This distribution will change depertina on the decision to repair. This issue is addressed in later sections. Knowing any two of these paraneters, one can solve for the third. For exanple, the critical flaw size to cause f ailure is calculated from: 1 2
/K Ic T a#
= 11 C5 \ #c / (2.3) if both the toughness (Kg) and the stress level (c )c are known.
Conversely, the critical applied stress as a function of crack depth can be computed fron,
, IC
'O c (2.4)
Although these conditions most eppropriately describe the behavior of low toughness, high strength naterials where little ductility precedes
-e - -a.-- , - - - , - . . - - - - r - -
2-5 fracture, the use of K as a toughness neasure for either SA516 steel or k E70?8 weld netal ensures a conservative estinate of critical flaw size for brittle fracture, since no account is taken of the increased toughness which results f rom post-yield (transitional) behavior. Incorporating transitional behavior with more pre-f racture ductility gives increased toughness levels and decreases the susceptibility of the structure to a given sized flaw. The tenperature dependence of toughness properties neans that at anbient or higher temperatures, both SA516 steel and E7018 weld netal are above their lower shelf values on a fracture enercy versus tenperature curve. This in turn inplies that the use of standard elastic fracture nechanics will ba conservbtiVP. Elastic-plastic crack opening displacenent (COD) concepts have been used as a check on structural integrity. Elastic-plastic fracture nechanics
^
concepts are discussed in the next section. 2.1.2 Elastic-Plastic Fracture liechanics (EPFli) The basic principles of EPFf' have been developed over several years (2-1, 2-2,2-3,2-4) and one national standard exists for crack opening displacenent (C00) testing (2-5). This nethod requires critical COD values which were not available for the actual field material. A review of the literature (notably 2-6) was nade to check the appropriate naterial characteristics. One of the best nethodologies for EPFf: evaluation, the British Draft Standard nethod. utilizes crack opening displacement (COD) concepts. In principle, the critical condition is reached when the applied K; or C0D (6) reaches the resistance level of toughness necessary to cause fracture (K g or 6c). The C0D nethod is conplately compatible with the LEFri approach (2-7) and can be used in place of the K nethod. For applied stresses well below yield, Se a
= Y J 6 r loc-esec ('7 c y-) (2.5)
' 2-6 where 6 is the developed COD;y e andyo are the yield strain and stress respectively; o is the applied stress; and ,a_ is the half crack length of a center-cracked plate model. It can be seen from Egn. (2.5) that as o approaches e,, the developed C00 becones infinite. This only occurs for the elastic-perfectly-plastic material behavior that was assuned for the developnent of Egn. (2.5). For materials that work harden,'the ' relationship between C0D and applied strain (for stresses above yield) has been deternined by analytical, nunerical, and experimental methods (2-8,2-9).
As in LETI', once the stresses and material properties have been characterized, it is possible to determine the allowable flaw size to prevent fracture initiation. It is then possible to fietermine the expected nargin of safety-between the flaws which may be in the structure and those necessary to cause failure. The elastic-plastic analysis discussed in this section and the linit load analysis discussed in the following section, were also perfornec. 1 2.1.3 Linit Load Analysis As the size of a critical flaw increases, a regine is entered in which increasing material touchness no longer can prevent initiation of a craci under nonotonic loading. The initiation criterion betones independent of toughness and now becomes a function of the strength properties of the naterial and the remaining liganent of naterial. In this regine, a lir:it load or plastic collapse analysis describes the governing failure node. t For linit load analysis, the critical stress to cause failure is calculated fron an interaction relation connon in the analysis of steel structures. This relationship between the applied nenbrane load (P) and f bending nonent (fi) at f ailure in a bean or plate with a rectangular cross-section is: I
2-7 (h)2 k'1
+
t u (2.6) where P and M are the applied loads, andgP and gM are the limiting values of P and it. The magnitude of gP and fig are functions,of crack length, 2a, flaw peonetry and naterial properties. The linit load of P isg deternined fron the geonetry of the section and
" the naterial properties. After the reduction in area due to the flaw is accounted, the linit load can be expressed in terns of a .linit stress and the cennetric variables. The limit stress is normally the naterial yield strength when the naterial behavior is assuned to be elastic-perfectly plastic. However, for naterials which exhibit significant strain hardening, o gcould be sonewhere between yield and ultinate, and the appropriate value to use should be deternined by tests.
For this analysis, we use a flow stress which is the average of the yield and ultinate strengths, i.e.: ot * (c y+ C uts (2*7) where eq is the flow stress, o the y specified nininun yield stress and the specified mininun ultinate strength. uts ( ' Once the linit conditions have been calculated, Eau. (2.6) and the expressions for applied nenbrane stress as a function of pressure and applied nonent can be used to determine the failure condition. l A limit load evaluation has been made in conjunction with LEFil nethods in the present case. 2.1.4 Sunnary of fracture nechanics background The f ailure behavior of structures under nonotonic loading can be classified into three regines. Of these linear elastic fracture M
- y- .-.g-
2-8 nechanics has been deternined to be most applicsble to the current naterial and service conditions. Bounding studies based on elastic-plastic and plastic limit load analyses have also been perforned. 2.2 Faticue Loadinc 2.2.1 Analysis Method The preceding discussion addressed the case of nonotonic loading. In the present case, there are a snall number of cyclic loads which nay occur on both the drywell vent and containnent structures. This section discusses the way in which these loads can be evaluated in the light of the previous discussion. Fatigue evaluation, based on f racture nechanics, assuries that initial flaws are present of size a$ and that the lifetime of a component is that reauired for a crack to grow from the initial size, a$, to the critical size, a . Crack crowth rate data nay be correlated to the crack tip c stress intensity f actor range for the given load cycle in the followinc forn; da/dN = f(AK) (2.8) l where da/oN is the crack growth per load cycle. By integrating Eqn. 2.6 l with the appropriate conponent stress field to calculate K, the nunber of cycles, N. (residual life) for a crack to grow fron a$ to a is corputed from ,, da N= da/dN a 4 (2.9) The final flaw size expected at the enc of the design life, a f, can be l deternined by integrating Eqn. 2.9, using the appropriate stress distribution to calculate K, and the number of total design cycles N froo ( l i a f da N - =0 o da/dN (2.10) d a i 1
2-9 where Eqn. 2.10 is a transcendental expression involving af and must be solved by an iterative process. 2.2.2 Crack Growth Rate Representation . Many empirical relations to express da/dN behavior have been proposed; the earliest and most well known is the Paris rule (2-10) which takes the form, da/dM = CLK" (2.11) where C and n are constants determined fro' the data, and AM is the range of applied stress intensity factor computed from the minimum and maxinun stress in the cycle: AK = K ~K ( 'I ) nax min The advantage of the Paris relation is that it is ' simple in form and it fits experimental data well in the middle range of LK. A disadvantage of the relationship is that it does not directly account for mean stress effects (P-ratio effect where R = Koin I"nax) which can accelerate fatigue crack prcpagation. However, these effects are accounted for in the choice of experinental data used in the modeling procedure. 2.3 Evaluation of Fissures - Flaw !!odeline 2.3.1 Flaw Orientations 1 Fissures have been located with orientations transverse and parallel to the weld seams in both the drywell vent and containment structures. ' Identification of flaw orientation in this report has been made with respect to the weld directions - that is transverse or longitudinal relative to the weld. The details of interest are shown in Figures 2-2 and 2-3 for the drywell and containment respectively. Although no fissures were found in the weld around the drywell vent proper, analysis of this region was also performed as it represents the highest stress region. Figure 3-1 shows the details of this area.
2-10 Stainless Steel Carbon Steel tieldment Weld [ -. SA516 Gr. 70 [ g i
\SA240 Type 304 ~
Longitudinal I Flaus e I I i I Transverse i Flaws l l l
\ ;l /
Figure 2-2 Schematic Representation of Drywell Vertical Seam Flaw Orientation s-L
2-11 (Clad) Containment Transverse Flaws Longitudinal Flaws Fillet Weld p
/
(Clad) Doubler Plate
/
Figure 2-3 Schematic Representation of Containment Flaw Orientation (. e r-
+
2-12 2.3.2 Flaw Distributions Fissures have been associated with the heat affected zones created between the stainless steel weldment and stainless steel portion of the clad plate. Therefore the defects tend to be initially of short extent in the transverse direction and potentially long in the longitudinal direction. The transverse cracks in both structures are therefore nodeled as initially semicircular in shape, e.g. with crack aspect ratio (length to depth) equal to 2. The longitudinal flaws are conservatively nodeled to have initial aspect ratios equal to 10 and 50 in each locatien. The first ratio may be representative of long single defects, the latter I of linked defects. In practice, both gave essentially the sane results. The defects are surface connected and were thus modeled as surface semi-elliptical flaws. Figure 2-4 illustrates this model. It contains three degree of freedon growth, that is, independent growth is possible in the through wall direction (Degree of Freedon -1) and at the ends (Degree of Freedom II or III). In addition, an analysis of a buried defect has been performed. This would sinulate a repair which had closed the surface connection but not repaired the entire depth of the defect. This model is shown in Figure 2-5. One of the netallurgical samples (No. 6) had fissures which " penetrated through the entire stainless clad layer, ar.: at one point, penetrated the internediary nickel flash between the stainless cladding and carbon steel base material" (CD-12). For this reason, the worst case initial flaw depth (through-wall dinension) was chosen to be equal in extent to the ! full stainless clad layer. This is 10% of the total thickness for both j the drywell vent and containment plates. The initial flaw depths were thus taken to be equal to .1" (10% of 1") and .15" (10% of 1.5") respectively. 4 The following two sections discuss inputs to this model, stress and crack-growth-rate information respectively, i Ih w
2-13 lF:
- 30!- . IL;::*!Ca. 5Jr A;i CA;;.
. .. (; ..e . :
(
- i, *
. t I
h I s i I lj y., , 3
~
I f
=
'\\., l 11 .
l ,\
". g I
I I o l I I l t l t e
=
1 .* \
,..... v 4 , .
i g 1 5
.C a.x i._. g .....,
f.... s.. a;;I. rt:.- **.t FA R"I'Et , F*; .***. f i v.::s1 l ::e. '.. :er IF: 225
- . - er of ;e; ces of Frect: . IC:F 3 (raci Frcn; $P.a;L -- Se a.Elli;tical Crau 0' erin; "-*e -- Pode 1 Finite didth Ef fe*ts w ho l
Variatie Thickness Ef f t:ts f *" NO l Figure 2-4 Schematic of Surface Elliptical l Flaw Model l l 1 i' 1 1 -- - - . _ . . - ___
2-14 IF1
- 300 - S#.: . I;;;r :::. C G:n
-*::*. ;i?vi*:.
4 t o , li 4 Rg ll4 l l I I l
. l
. t 1 l i e l
l ! l 's-3- l l
!f
=. I N n -
l 4 l N s l I l i _ _ _ _,N% .\ i i - l _. { i e 3 I t i
) ,
t l I i l f ,l
, ,,4 ,
+
t- - j
* : : ._._~ . : . : ' ' *.
- !:*.: 5 it*:"I*E: *: ::. !.* :--
- ::ri r.:e n :.. :e* FI :::
- er :' *e rets , c' Free:: . I ::
Ora:e Fr nt $*.!:t -- ElliO!1;&l tra:e ::e.ir.; "::e .. I'c ce l Firtte aict9 Effe::s = ft: l variaDie 'ni:. ess [ffe::t IJH f.: i l l Figure 2-5 Schematic of Buried Flaw Model L i 1 k b*
2-15 Section 2 REFERENCES 2-1 Wells, A.A., " Notched Bar Tests, Fracture Mechanics and Brittle Strengths of Welded Structures," Houdremont Lecture 1964, British Welding Journal, No.1 (January 1965). 2-2 Sumpter, J.D.G. and C.E. Turner, " Fracture Analysis in Areas of High Nominal Strain," Proceedings Second International Conference on Pressure Vessel Technology, San Antonio, TX (October 1973). 2-3 Egan, G.R., "The Application of Fracture Toughness Data to the Assessment of Press;re Vessel Integrity", Proceedings Second International Confyyce on Pressure Vessel Technology, San Antonio, TX (Octwer 1973). 2-4 Burdekin, F.M. and M.G. Dawes, " Practical Use of Linear Elastic and General Yielding Fracture Mechanics With Particular Reference to Pressure Vessels", Conference on Practical Application of Fracture Mechanics to Pressure Vessel Design, Institution of Mechanical Engineers, London, UK (1971). 2-5 British Standards Institute, " Methods for Crack Opening Displacement (COD) Testing", (1972). 2-6 Stuber, A., J. Wellman and S. Rolfe, " Eighth Progress Report on Application of the C0D Test Method to the Fracture-Resistant Design of Pressure Vessels", for the Subcommittee on Effective Utilization of Yield Strength of the PVRC, (February 1980). 2-7 Egan, G.R. , " Compatibility of Linear Elastic (Kye) and General Yielding (COD) Fracture Mechanics", Engineering Francture Mechanics, Vol. 15,(1973). 2-8 Merkle, J. , " Analytical Applications of the J Integral", ASTM STP 536, American Society for Testing and Materials, (1972). 2-9 Hayes, D.J. and Turner, C.E., "An Application of Finite Element Techniques to Post-Yield Analysis of Proposed Standard Three-Point Bend Fracture Test Pieces", International Journal of Fracture, Vol. 10, (1974). 2-10 Paris, P.C. , M.P. Gomez and W.D. Anderson, "A Rational Analytic Theory of Fatigue," The Trend in Engineering, Vol .13, No.1 (January 1961). l 0 4 l-1 1 I 5 L.
3-1 2.0 ANALYSIS OF STRESSES The models which have been established in Section 2.0 require several kinds of information. This section discusses sources of stress which are present in the drywell vent and containnent structure. The discussion includes both primary (" applied", " service") stresses and secondary (" residual") strasses. Further, both structures are subjected to a limited spectrum of cyclic loading. The details of stresses for the drywell vent and containment structures are found in the sections to follow. Additional details are found in Appendices B and C respectively. 3.1 Drywell Vent Structure Prinary Stresses The appropriate stresses for analysis of the drywell vent structure have been obtained fron unit stress calculations (CD-34). Figure 3-1 indicates the two regions which were examined around the vent sean and at the vertical plate edge. In both locations, three elements were checked to determine the most highly stressed areas. These are marked 1, 2 and 3 around the vent and A, B, and C at the vertical sean. Highest stresses were at 2 and C respectively so these values are used at all locations. 3.1.1 Vent Region flinor fissuring has been found in the vent region and has been weld repaired (CD-91); however the stresses here are the highest for any location in the drywell. Therefore a buried flaw model was used to sinulate a subsurface fissure. Unit stresses were taken for elements 1, 2, and 3 as marked. Component stresses xo and a were calculated for both the SRV activation and S.S.E. cases. 0.B.E. stresses were conservatively assumed to be equal to the S.S.E. stresses. The stresses were found to be (for location 2): Component Maanitude (ksi) Load Combination i D+P SRV x o W D+P 3py y ) w.
y u 3-2 X
'N Vent Seam Areas 1, 2. 3 1 4
l Plate Edge Areas A, B, C l Figure 3-1 Drywell Vent Structure Showing Areas Analyzed (Taken from CD-34) L
3-3 Load Combination Component Magnitude (ksi) D + SSE + P 3py o x 30.5 D + SSE + P SRV y 8.1, Appendix B outlines the derivation of these values. The SRV stresses were applied for a minimum of 16,740 stress cycles (per CD-18). Similarly, cyclic stresses equivalent to the SSE magnitude were applied for a minimum of 60 cycles corresponding to the total number of 0.B.E. and S.S.E. stresses over the plant design life. Note that these stress derivations will be conservative in that:
- 1) The maximum value of stress along the seams for any location is taken as representative of all fissure locations (or potential fissure locations). Obviously, the stress situation will be this maximum value only if fissures happen to coincide with this one location.
- 2) The 0.B.E. stresses are assumed equal to S.S.E. values when, in fact, they are lower.
- 3) For most cases, substantially more cycles are considered than actually applied (see Results - Section 7.0).
In a similar manner, the stresses for the worst location (No. C) vertical seam were found to be: Load Combination Component Magnitude (ksi) D+P SRV 0.6 x D+P SRV y D + SSE + P3py o 4.9 x D + SSE + P3py o 20.5 y 0 e 9
3-4 3.2 Containment / Doubler Plate Primary Stresses i Details of the primary stresses for the containment / doubler plate region were derived from two calculation forms, the doubler plate analysis by NNIC (CD-31) and the Containment Shell Analysis by NNIC (CD-17). Fig. 3-2 (CD-31) shows the detail of interest at elevation 575' - 10". A composite sheet of all the stress cases of interest was made for this elevation and is shown in Appendix C. The majority of stress cases for this region result in compressive stresses. Since tensile residual stresses were anticipated, the consideration of these compressive stresses was necessary as cyclic crack growth could still occur (as the resultant cycling is then in the purely tensile regime). Since fatigue crack growth has been shown to be a function of AK level (see Section 2.3), the maximum stress difference is the appropriate parameter. By taking the maximum ranges of component stresses, the worst case (largest aK) value for o g (appropriate to longitudinal defects, see Fig. 2-2) is found for case 12 long term LOCA. The range of stresses including all unit stresses is -454 psi to +4462 psi for case 12. This range bounds all other cases both in AK and in maximum s+,ress and this maximum range was applied to all appropriate stress cycles at outlined in Section 3.5. This procedure results in a conservative analysis. In a similar manner, the worst (largest) stress range for T (appropriate to transverse flaws, see Figure 2-2) was found to be Case 7 (normal operating) where the range was -3065 psi to -7668 or ao=4603 psi. However, Case 12 (short term LOCA) has a slightly less compressive stress level -2762 psi to -6586 psi (ao = 3824 psi). A case bounding both of these was established so both stress level and ao were considered. In summary, the following worst case stress ranges were used for fatigue evaluations for the cantainment/ doubler plate: Stress Range (Ksi) Flaw Orientation 4x Longitudinal -0.5 to +5.0 5.5 Ksi Transverse -2.5 to -7.0 4.5 Ksi N. 1 I m
3-5 __ Elev. 575'-10" Containment Shell Doubler Plate (Inside Surface) (Inside Surface) Containment Shell Doubler Plate (Outsice Surface) (Outside Surface) N Base Liner Concrete Elev. Foundation Mat
- - 574'-10.187" I
Figure 3-2 Containment / Doubler Plate Stress Locations i
I 3-6 3.3 Drywell Vent Structure Residual Stresses The region where the fissuring occurs could be subject to several sou,rces of residual stresses. Section 3.6 discesses two of these potential sources - grinding residual stresses and bonding residual stresses. As will be shown, the potential effect of these is either compressive, small or relaxed out by assumptions about initial flaw size. Therefore, the only residual stress type which will be used in the analysis is that due to welding. The drywell vent fissuring is associated with the heat affected zones of the vertical seams between plates (see Fig. 2-2). These welds are single-side groove welds. A literature review was performed to characterize the resulting distribution of residual stresses. This aistribution will have components transverse to the weld and longitudinal or parallel to the weld. There will be a decreasing level of stress as the flaw location moves away from the weld region. There will also be a change in stress through the thickness of the welded joint. Traditional fitness for service evaluations have assumed uniform yield level stresses for all regions. While this assumption is conservative, it is in f act overly conservative in many instances. In the present analysis, actual bounding distributions are used to determine stress levels. For reference, Fig. 3-3 shows a typical shape of the stress di striou:. i on. The literature review was organized to screen data not appropriate to the present study. Residual stresses have been found to be a complex function of a number of variables and thus an effort was made here to obtain data relevant to the present geometry, material, postweld treatment, etc. All the references cited are for carbon steel materials (most for A212-B i or SA516 Grade 70), and weld metals of a type similar to the field case j (generally E7018 - one of the permissible electrodes CD-8). The thickness range chosen was 1" - 1 5/8". None of the literature data cited were for plates post-weld treated, either thermally or mechanically. Where geometry differences occurred (as between single and double sided joints), the effects I were considered. Appendix D illustrates the residual stress distributions used in the present study. l l 4
3-7 Comoression Tension m (Tension
" ,~
/
(a) (b) Compression 8 Tension I i i I l (c) J i Figure 3-3 Schematic of Tvoical Assumed Residual Stress Distributions in Plates Without Fixed Ends for a Single-V Groove Weld: (a) longitudinal, (b) transverse, and (c) through thickness. L =f
3-8 3.3.1 Transverse Stress Distribution - On the Plate Surface Fig. D-1 illustrates the resultant surface, transverse residual stress distribution. It has a magnitude of 0.6 cy at the weld centerline decreasing to zero at a distance 6 inches from the centerline. This is conservative to the data shown from Nordell and Hall (3-1). This is also conservative in that it is appropriate to transverse stresses in the mid-plate weld region (that is away from the ends of a plate). As the defect origin moves to the top or bottom of a plate, the transverse stresses become less tensile (see Fig. 3-3 for reference). 3.3.2 Longitudinal Stress Distribution - On the Plate Surface Fig. D-2 shows the distribution that has been developed for the decay of residual stresses with distance for the longitudinal direction. Data from Nordell and Hall (3-1) and Greene (3-2) are shown. Note that no account has been taken of beneficial compressive stresses beyond 4" from the weld centerline in the present analysis. 3.3.3 Transverse Stress Distribution - Through Thickness Variation Fig. D-3 shows the manner in which transverse residual stresses decay through the tnickness. The work by Leggatt and Kamath (3-3) is particularly relevant here. They ran investigations of both notched and unnotched specimens. The stress relaxation which does occur on high r surface tensile residual stresses due to the notch is relevant in the i present case in which fissuring has occurred. These data support the fact that taking full residual stresses in the presence of a crack or notch is unduly conservative. Thus, the distribution used in this analysis bounds all the notched specimen data and most of the unnotched l L data. The points above the line correspond to unnotched data and/or data inappropriate to the present analysis. L
3-9 3.3.4 Longitudinal Stress - Through-thickness Variation The data for the through-thickness variation of longitudinal stress are shown in Fig. D-4. Other appropriate references are listed at the end of this section (3-4 through 3-6). 3.4 Containment / Doubler Plate Residual Stresses The typical assumed weld residual stress distribution used for fillet weld stress is shown in Fig. 3-4. Note that this distribution provides substantially more compressive stress area, particularly through thickness, than that developed for the butt weld above. There is a dearth of data strictly applicable to the geometry (primarily thicknesses of plates) for this j oint. Therefore, the distribution developed above for the drywell has been applied to the containment as well. This should provide a conservative estimate of the effect of weld residual stress in this joint. 3.5 Cyclic stresses The crywell vent and containment are subject to the same stress spectrum. The events which could cause cyclic stress are shown in Table 3-1 with the number of cycles to be considered over the plant design life as previously outlined. l Note that for the containment, the worst case stresses were applied to all 16,800 cycles. That is, safety relief valve actuation stresses and OBE stresses were considered to be as severe as SSE stresses. t For the drywell vent, the cyclic stresses were applied as SRV stresses for i 16,740 cycles and SSE stresses for 60 cycles as discussed. I 3.6 Combined Stresses: Summary The total stress that is considered for the evaluation of a defect, consists l .
w,B s s e r t S t c l e a f u e d D . i s e en o Ro T i dt l eu a tb c ci i i r p dt y es T ri PD N #
/ -
j
~
v- -
/
~
W / - n n o o i i s s s n e e r T p m o C 2 w1 WESC n o* 's82C.0, e?e*r5: FU E* N2a 027[C.8 2 '5 " A F " 7 1I !
1 3-11 ! I
- Table 3-1
SUMMARY
OF CYCLIC STRESSES
# Cycles Per Total Number Sources # of Occurrences Occurrence of Cycles SRV Activation 1860 9 16,740 OBE 5 10 50 SSE 1 10 10 1
i (fromCD-18) 4 1 1 1 l i F e t
3-12 of the primary and secondary components. Both sets of stresses have been chosen to bound the expected stress state conservatively. For analytical purposes, they have been superposed with elastic-perfectly plastic material behavior. Table 3-2 lists all flaws, stresses and locations modeled.~ Typical total stresses are shown in Figures 3-5 through 3-7. These are provided to illustrate the types of behavior modeled and consist of: Stress Flaw Figure Structure Location Type Orientation Flaw Type 3-5 Drywell Vent Vertical Seam SRV Transverse Surface 3-6 Drywell Vent Vertical Seam SSE Longitudinal Buried 3-7 Containment Doubler SSE Longitudinal Surface Note that each figure contains a through wall distribution (part a) and a surface distribution (part b). Also each figure represents the original residual stress distribution, the cyclic load level and relaxation level. 3.7 Grinding and Bonding Residual Stresses Residual stresses induced by grinding operations and the bonding process have not been considered in this analysis. In the worst case of grinding stresses, - large surface tensile stresses are possible. Their magnitude and distribution is a function of many parameters. A literature review of the state-of-knowledge of this subject was completed (see for example, references 3-7 l through 3-11). The item of greatest relevance is that the effects are generally confined to very near the surface. Figure 3-8 illustrates that the high tensile stresses are in effect only to a depth of 10 mils or less. Thus for the present study, for assumed defects which extend through the clad to a depth of 0.1" in the drywell and 0.15" in the containment, the effect of grinding stresses will be negligible. The residual stress state due only to bonding stresses would be expected to place the stainless portion in tension with corresponding compression in the carbon steel. Since the fissure is modeled to extend through the clad, the tensile stresses in the clad will be relaxed while no account of rasidual compression is taken. l
i Table 3-2 SIM1ARY OF ALL CASES EXAMINED FLAW TYPE FLAW ORIENTATION STRESS TYPES RUN THROUGH RErilON Na BURIED SURFACE WALL 1RANSVERSE LONGITUDINAL SRV SSE STRUCTURE Drywell Vent Vertical W1 X X X Seam W2 X X X W3 X X X W4 X X X W6 X X X W7 X X X W8 X X X W9 X X X W10 X X X W12 X X X , Vent Seam W13 X X X U W14 X X X W15 X X X WIS X X X W17 X X X W18 X X X Containment Doubler W22 X X X Plate W21 X X X 1 W19 X X . X W20 X X X 09A X X X
I I 3-14 l I (a) Ist SRV Activation Initial Residual i 80 - Stress Distribution l 60 - 'N-O 40 - Shakedown (Unloaded After
.0
~
1st SRV Activation) 20 - 0 E o : : : : M 0.2 0.4 0.6 0.8 1.C Inner Surface Position Through Wall, (inches) (b) I 80 -
. / nitial Residual Stress Distribution 1st SRV Activation 60 -N-3 40 -
d.
- 20 - Shakedown M
$ 0 , ; ; ,
1 2 3 4 5 Distance from Weld Centerline, (inches) Figure 3-5 Combined Stress Distribution for Drywell Vent Edge, Transverse Flaw (Residual Stress + SRV Loads) (a) Through-thickness distribution, and (b) Surface Distribution i
3-15
,)
l 80 - . - - Residual Stress Level S.S.E. Cyclic Load 60 - 3 g 40 -
~ ~ - -
N 20 -
%'%s '%
E % 3 0 :
. . '~m.
\ 0.2 0.4 0.'6 0.8 ' s M N
inner surface s Position Through Wall, (inches) (b)
-- Residual Stress Level 80 -
S.S.E. Cyclic Load 60 - O 40 - f
. 20 -
0 E O M Distance from Weld Centerline, (inches) (constant - 1.3") Figure 3-6 Combined Stress Distribution for Longitudinal Flaw in Drywell Vent Seam (Residual Stress + S.S.E. Stresses) (a)through-thicknessdistribution,and (b) surface distribution
3-16 (a)
-- Residual Stress Only 80 - rn r Range of Cyclic Stresses N
O w 40'* \ N d N
. 20 -
0 ; :
.3 .6 .9 1.2 1.5 Inner Surface Position Through Wall, (inches) *
(b.)
, -- Residual Stress Only 80 -
77777 Range of Cyclic Stresses 60 -'s
's O 40 - 's I
's N
- 20 - s m s N *
,E D . .' . .
2 4 6 8 Distance from Weld Centerline, (inches) Figure 3-7 Combined Stress Distribution for Containment / Doubler Plate Region, Transverse Flaw (a) Through-thickness Distribution, and (b) Surface Distribution. l l l. i ._. . . . . . - ._.
3-17
\
\ --- Inconel 718 80 - -
\ AISI 4340
\ -
g Type 304
\
\
60 "
\
N 's N \
\
g 40 - - g
\
k> \\ \
*, 20 - -
\\
e M s-a e a I e , m 0 E ( 6 's $ 10 D -N
's'~ ___,-----
-40 Depth, mils Figure 3-8 Typical Distributions for Grinding Residual Stresses
3-18 Section
3.0 REFERENCES
3-1 Nordell, W.J. and W.J. Hall, "Two Stage Fracturing in Welded Mild Steel Plates," Welding Research Supplement, (March 1965), p. 124-5 to 134-5. 3-2 Greene, T.W., " Evaluation of the Effect of Residual Stresses," Welding Research Supplement, (May 1949), p. 193-5 to 104-5. 3-3 Leggatt, K.H. and M.S. Kamath, " Residual Stresses in 25 En Thick Welded and Locally Compressed Specimens in the As-Welded and Locally Compressed States," E45/1981, The Welding Institute, Research Report, (June 1981). 3-4 Rosenthal, D. and J.T. Norton, "A Method of Measuring Triaxial Residual Stresses in Plates," Welding Research Supplement, (May 1945), p. 295-s to 307-s. 3-5 Parlane, A.J.A., "The Determination of Residual Stress: A Review of Contemporary Measurement Techniques," Paper 8, Residual Stresses in Welded Construction and Their Effects, The Welding Institute, (1977). 3-6 Takahaski, E., K. Ivai and K. Satoh, "A Method of Measuring Triaxial Stresses in Heavy Section Butt Weldments," Trans. Japan Welding Society, Vol. 10, No. 1, (1979). 3-7 Chrenko, R.M., " Residual Stress Studies in Austenitic and Ferritic Steels," International Conference on Residual Stresses in Welded Construction and Their Effects, The Welding Institute, (November 15-17, 1977). 3-8 Kuniya, J., et al., " Effects of Surface Finishing on Stress Corrosion Cracking of Austenitic Stainless Steels in High Temperature Water," Materials for Energy Systems, Vol. 1, (March 1980). 3-9 General Electric Company, " Studies on AISI Types - 304, 304L and 347 j Stainless Steels for BWR Application," NED0-10985-1, 3, 5, EPRI Contract RP-449-2, First, Third and Fifth Quarterly Progress Reports. 3-10 Giannuzzi, A.J., " Evaluation of Near-Term BWR Piping Remedies," NEDC-21463-3 Third Semi-Annual Progress Report, (April-September 1977), L EPRI Contract RP 701-1. l 3-11 Field, M. and W.P. Koster, " Surface Integrity in Grinding", in New Developments in Grinding Proceedings of the International Grinding Conference, Carnegie Press, Pittsburg, PA., April 18-20, 1972. l i I
- a. . . _ _ __. __ _
4-1 4.0 FATIGUE GROWTH RATES 4.1 Introduction , In order to assess the extent of crack growth that could occur over the plant design life, a fatigue evaluation was performed. This evaluation combined the cyclic stresses (Section 3.0 - Stresses) with the appropriate crack propagation rates (this section) to obtain expected growth (see Section 7.0 - Results). The purpose of this section is to assess propagation rates for defect growth ~ by a fatigue mechanism. The clad construction requires consideration of the various materials - Type 304 stainless steel, SA516 Grade 70 carbon steel, stainless steel weldments and carbon steel weldments. The following sections assess these materials: Section Material Classification 4.2 SA516 Grade 70 4.3 Type 304 Stainless Steel 4.4 Stainless steel weldments (such as 308, 308L, etc.) 4.5 Carbon steel weldments For each material classification, references have been drawn together to assess a conservative bound (fastest) on potential crack growth. The growth rates chosen for the analysis must reflect the appropriate parameters including:
- 1) Sensitization (in the 304 cladding portion)
- 2) Exposure to Suppression Pool Water Chemistry
- 3) AK level effects
- 4) Temperature (4 150*F)
- 5) Geometry: Nominally 10% clad plate with single vee butt weld
- 6) Material composition: Nominal AISI Specification
- 7) Cycle Characteristics: Number of cycles, ao levels, R ratio, waveform, frequency etc.
L.-
.-- , - ~ . - . ~, , -
4-2 Note that no data are available for the exact condition in effect. However, sufficient studies have been performed to tllow an evaluation of bounding considerations. Each of the seven considerations is discussed in the sections to follow. Unless otherwise stated, the following unit conventions are in effect: e AK (alternating stress intensity), ksi /ili e T (temperature), 'F e da/dN (crack growth rate), inches / cycle e f (frequency), cycles / minute 4.2 SA516 Grade 70 As the fissuring has been conservatively taken as extending to the carbon steel material, the fatigue growth rate of a sharp crack in the SA516 Grade 70 has been examined. Fatigue data on 516 Grade 70 found in the literature are plotted in Figure 4-
- 1. The scatterband (4-1) was derived 'from tests en nine plain carbon and low alloy steels similar to and including A212-B (precursor to A516 Grade 70). The data points (4-2, 4-3) are for A-212 Grade B, the equivalent to A516 Grade 70.
4 scatterband of data for SA516 Grade 70 (4-4) is also shown. The upper bound (4-5) represents behavior limits of A516 Grade 60. All but two of the data points relating to this upper bound fell above the scatterband of Fig. 4-1. This is the cause of the knee of the line. The ASME Boiler and Pressitre Vessel Code Reference Curves (4-6) for both air and water environments are also shown in Fig. 4-1. The air curve is an upper bound on fatigue crack growth data measured on SA533 Grade B, SA508 Class 1 and SA508 Class 3 steels for use "if data from the actual product form are not available". The curve can be used as a bound for steels similar to A516 Grade
- 70. The other curve, for water environment, drawn to the left of all other data, is shown on Fig. 4-1. It should be noted that this line is a bounding curve for behavior corresponding to the following conditions:
- 1) Sound material
- 2) Air environment
- 3) AK range shown on Fig. 4-1
-' W - - - ---v--y - - - - - ---
w-- e
-m --r-M - - - - m
4-3 Stress intensity factor range, AK (MN/m3 /2) 4 6 10 20 40 60 80 100
-3 _, , , , , , , ,
10 ,,l , , ,l _ _ o (4-2) a (4-3) R=0.1 -3 e 10
- e (4-3) R=0.67 3 a- T 10~# -
e a T : 8 _ R g - Upper Bound ' g x u (4-5) e o _ /
-. e 3e _
e 8 / Ic ' .5 j - a : 5
.:: 3 - _ m _- , =
a _ E R -
-o .g p _
m
- : a _ o I .
2 _ _e _ _ 10-b 5 5C _ $ L b, Baseline _
=
a (ASME Sectio t, M XI Water - 2 5 10-6 -- Environmen ) - u
~
_ < Scatterband
. (4-1) _
10-6 10-7 i e i i i I e I i i I i It-4 6 8 10 20 40 60 80 100 Stress intensity factor range, AK (ksi /iR) ! Figure 4-1 Fatigue Crack Growth Rate Data for SA516 Grade 70
4-4
- 4) Temperature 75'F
- 5) Variable geometries
- 6) Nominal ASTM A516 Grade 70 material
- 7) Cycle characteristics: R 0 to .25, f = 6 to 600 cpm, sinusoidal waveform For convenience, the da/dN vs AK baseline of Fig. 4-1 can be expressed as:
jt = 3.8 x 10-10 3g3.726 cf dN The effect on propagation rate of changes in these conditions is discussed below. 4.2.1 Effect of Environment on Fatigue Growth Rate
- Research has shown for many materials, that although aqueous environments accelerate growth by fatigue mechanisms, the environmental effect is minimized at low temperatures (Ref, 4-7, 4-8). Thus a conservative upper bound for low temperatures (such as are at issue here) would be data obtained in high temperature aqueous environments. Fig. 4-2 shows the upper bound for A516 in 500*F water (4-9). Furthermore, water chemistry has been shown not to be a significant factor at temperatures less than 150 F (4-10). This reference points out that fatigue crack growth behavior of medium strength carbon steel is very similar in any water environment. Thus the baseline shown in Fig. 4-1 which was chosen to be very conservative for air environments is also conservative for aqueous
, environments at icw temperatures.
- This is not an assessment of corrosion, or corrosion assisted fatigue.
See Section 5.0 - Corrosion Considerations of this report. For the .f discussion in this section and others to follow in Section 4.0 the issue of concern is only a fatigue mechanism. t be
4-5 Stress intensity factor range, AK (MN/m3 /2) 4 6 10 20 40 60 83 100 10-3 , , , , , ,
,,l , , , , , ,
,,l ,
10-3 T 10~# - E 5 - b S R -
. o g -
_._ 10 ' .!! e E T e 3
= - ,
c . s 1" ' , 5 - E o 3 4:. Baseline r
?
x - : Upper Bound at _ 10~' 5 T _ 500"F (4-9) : ! e - u o o - a t 4 - e 5 10-6 -
-6 10 10-7 , , , ,,l , , , , , , , , ,
4 6 8 10 20 40 60 80 100 Stress intensity factor range, AK (ksi /In) Figure 4-2 Comparison of Fatigue Growth Rates for Baseline and data at 500 F
4-6 4.2.2 Note on AK Threshold Effects The literature indicates that the threshold value for SA516 Grade 70 in aqueous environment has been estimated to be anywhere from 1 to 12 ksi
,Mi(4-12,4-13). The threshold is thought to be a function of stress ratio, R, with low threshold values for highest R ratios. The threshold is that value of aK below which fatigue growth will not occur despite
, continued cyclic loading. The baseline curve whicn is shown in Fig. 4-1 is thus conservative, even at low AK, as it does not assume a threshold but continued growth down to applied AK = 0. 4.2.3 Temperature Effects Although the temperature can range up to 150*F (CD-32)for maximum normal operating pool, (continuous safety relief valve blowdown without bubble pressure loads), the expected operating temperature will generally be lower. For example, maximum pool temperature with safety relief valve operation is 130*F. For European steel StS2-Nb (approximately equivalent to ASTM A516 Grade 70), each 16*F temperature increase has been shown to have the same effect as a 2% increase in AK (4-14). Thus for 130*F (a 55*F increase), would have about the same effect as a 6% increase in AK. Multiplying by the slope of the baseline (slope = 3.726) produces a change of 24% in (da/dN). As the closest data point below the baseline in Fig. 4-1 would require an increase of 50% in (da/dN) to reach the baseline, the 24% shift due to temperature effects, does not warrant changing the baseline. Further, as shown in Section 4.2.1 on environment, even at 500*F in an aqueous environment, the baseline provides a bounding value. 4.2.4 Geometry effect As (da/dN) is a materials parameter, the relationship between (da/dN) and AK is not considered to be dependent on the test sample geometry. This is confirmed by the fact that almost all of the data representing several different geometries of Fig. 4-1 fall inside the scatter band. It has t-- - - - - - - y
4-7 also been shown that size and thickness have no affect on fatigue rates in A516 Grade 60 (4-5). 4.2.5 Effects of Composition The effect of compositional variations within the ASTM specification on fatigue rates should be accounted for in the scatterband of Fig. 4-1. Further it has been shown for example, that a reduction in the number of MnS inclusions in 516 Grade 70 does not significantly change da/dN (4-4). Therefore, the effect of composition is incorporated in the baseline curve used.
! 4.2.6 Cycle Characteristics The effect of stress cycle characteristics on crack growth rate can be evaluated by considering' these factors:
- 1) Frequency
- 2) Haveform (Sinusoidal, Sawtooth, etc.)
li 3) Stress Ratio (omin/ cmax) II 4.2.6.1 Frequency and Waveform Effects
- The baseline of Fig. 4-1 includes data for the range of frequencies . 6 to 600 cpm. It has been reported (4-11) that growth rate
[. increases with decreasing frequency, reaching a maximum acceleration for frequency between 0.1 and 1.0 cycles per minute. This effect has been included in the high R ratio data shown in Figure 4-3. ] The literature has shown that there are effects on (da/dN) due to waveform (4-9). Neither ramp nor hold time significantly increase and may decrease fatigue growth rates with respect to sinusoidal or
- sawtooth loading (4-11, 4-15).
I w-i e-
l 4-8 Stress intensity factor range, AK (MN/m ! ) 4 6 10 20 40 60 80 100 10-3 3, , , , , , , , ,
, , , l ,,l ,
'~
10-3 Winter 1980 Section XI Water Environment - R > 0.65 . p" 4
/ s.
T -
/ _ 4 s
b .
,- e g - -_ w -
f; e
- E v
=
Baseline
% . %s ,% b-5 __
O _ e
$E _
R>0.5 " / $ (4 11) / ig-5 2 5 - T g ._ R < 0.5 - E g (4-11) _ g g - e u . u o 10- -
-6
_ 10 10- i e i e il i i e I i i i1-4 6 8 10 20 40 60 80 100 l Stress intensity factor range, AK (ksi Mi) Figure 4-3 Effect of R-ratio on Fatigue Growth Rates in SA516 Grade 70
.m-..-
l l 4-9 4.2.6.2 Effect of Stress Ratio, R Fatigue rates at high R ratios have been observed above the baseline of Fig. 4-1 (4-9). In order to account for this, several new ASf4E reference curves have been proposed which would be valid for R>.S. Two of these are shown in Fig. 4-3. The proposed new Section XI (4-10) curves are based on data from A516 Grade 70 and other similar materials. Both statistical and graphical curves for the range R>0.5 and R<0.5 are provided (4-11). For comparison, the baseline curve of Fig. 4-1 and the present Section XI curve (which has been shown to be a conservative bound to all conditions) are shown. Taking a combination of the work by Bamford (4-11), the new proposed Section XI curve and the baseline of Fig. 4-1 (present section XI), the most conservative combined curve over the given AK range is: g = 3.8 x 10-10 6g3.726 AK<4.864 dN da = 4.4 x 10-13 3g8 .0 11>aK)4.864 dN l da = 3.16 x 10-6 g .4 3g)31 dN This will be taken as the representative curve for A516 Grade 70 in the sections that follow. It should be noted that the upperbound and ASitE curves of Fig. 4-1, 4-3 are based on data at high temperatures and in aggressive aqueous environments. The conditions in question are low temperatures and a non-aggressive aqueous environment. 4.3 Fatigue Growth in Type 304 Stainless Steel There are many published studies of (da/dN) in annealed 304 stainless. Fig. 4-4 shows a scatterband compiled from eight published studies (4-7). Also shown in Fig. 4-4 are data f rom three additional studies (4-3, 4-16, 4-18). All of these data fall below the upperbound of the scatterband. Thus this baseline represents a conservative estimate of (da/dN) for annealed 304. It should be noted that this line is representative of behavior for the following conditions:
4-10 Stress intensity factor range, AK (Mil /m3 /2) 4 6 10 20 40 60 80 100 10-3 ,, , , ,
, , , ,) , , , ,
,,l
~
_ a (4-18) -
- 0 (4-3) 10" T
~
~
10-4 -- e u 3 - p K -
.t 4 - 2 E'
- r. -
10_4 5 - a ' : 5
.5 - m (4-16) v =
p -
" pr . E s -
e 30-0 a - 2 m E t0*
~
5-a 2 - C 5 -
/g e 10-"
g 5 - ~ b 2
~ -
=
t, o jo Baseline r 10-6 _ ,
- 3
- 4 Scatterband (4-7)
- -6 10
~
10-7 i e i ieI i I r I 1 i i1-1 4 6 8 10 20 4u 60 80 100 Stress intensity factor range, AK (ksi ,'in) Figure 4-4 Fatigue Crack Growth Rate for Type 304 Stainless Steel
4-11
- 1) Unsensitized Material
- 2) Air environment
- 3) AK range shown
- 4) Room temperature (T = 75*F)
- 5) Variable geometries
- 6) Nominal AISI 304 compositions
- 7) Sinusoidal or sawtooth waveform with frequency = 2 to 600 cpm, and R=0.
Any changes in the crack growth behavior because of differences between these conditions and those present for the material in question will be discussed below in the order that they appear here. For convenience, the da/dN vs AK plot of Fig. 4-4 can be mathematically modeled by a Paris type of equation. That is: cijt = 9 x 10-10 6g2.941 dN 4.3.1 Effect of Sensitization on Fatigue Growth Rate Studies have been published on the effect of F.ensitization on crack growth rates in 304. Fig. 4-5 shows data from (da/dN) vs AK tests in sensitized 304 (620*C - 4 hrs) (4-7). These data fall well below the baseline of Fig. 4-4. This is consistent with the fact that fatigue in 304 is usually transgranular and thus grain boundary carbides would have little effect. It has been concluded that sensitization has little effect on fatigue crack propagation rates in 304 in air or water (4-9). Based on these studies, the baseline curve established in Fig. 4-4 l appears to be conservative even for sensitized 304. 4.3.2 Effect of Environment The environmental conditions are assumed to be water at or below 150*F 5 Sm i_
4-12 Stress intensity factor range, AK (MN/m3 /2) 4 6 10 20 40 60 80 100 10-3 ,_ , , ,
,,,l , , , , , , ,; _
10-3 T
~
n 10*# -- s - - r K - 3 4 - 13 .; y g .-
.c _ e g -
3
~
0 _ _
= = - =
R -
- 3
-8 I f' __
- : oH00 - J 3 _ g A$r @75*F200*C %'
2 - Air @ 75 F 5 h 5 - 8 u Ba seh.ne _ E en l
. x "u - E 2 _ b 10-6 _
\ 5 10 6
~
l 10
~
' ' I I I i i I i 1l~
4 6 8 10 20 40 60 80 100 Stress intensity factor range, AK (ksi /f5) Figure 4-5 Effect of Sensitization on Fatigue j Growth Rate in Type 304 1
4-13 with water chemistry specifications as discussed previously. Fig. 4-6 shows published test results for (da/dN) for 304 in salt solutions and humid air environments (4-7, 4-20). These data all fall below the baseline. The salt solution of Fig. 4-6 provides a more aggressive environment than that in question here. Studies indicate that environmental effects are negligible at low temperatures such as are present in this case. It has also been shown that 304 exhibits no increase in fatigue crack propagation rates when subjected to water environments (4-21). Based on these studies, the baseline curve appears to be conservative for exposure to the suppression pool water chemistry in question. 4.3.3 Effect of AK Range As with the discussion for SA516, threshold effects for low applied stress intensity (aK) values have been established. The potential for threshold or zero growth has not been taken into account in the derivation of the baseline curve of Fig. 4-4. It has been conservatively ignored. The data presently fall well within the range of aK values presented in that figure. 4.3.4 Effect of Temperature The change in (da/dN) with a change from room temperature (75*F) to any l other temperature even to 150*F is negligible. This is confirmed by the data plotted in Fig. 4-7. These data fall below the baseline, even though they were obtained at 600*F. The line of Fig. 4-7 represents the upperbound of data from this reference, (4-7). Thus the baseline appears to ba conservative for the relatively low temperatures in question. I ! 4.3.5 Effect of Geometry The scatterband of Fig. 4-4 was compiled from 10 separate geometries of i
4-14 Stress intensity factor range, AK (MN/m3 /2) 4 6 la 20 40 60 80 100 10-3 ._ , ,
, , ,, l , , , , , , ,,l _
10-3 T
,_ 10-# .
3 .: - t
? -
3 4 -
. E 10 :
-[ -
_ e M e.: 0 _- 3
= . G e -o _
m 10 ._. 2 _ . 5-
.c _ _ 10-" 5 Tc . _/
Baseline -
- u E -
t 4 2 5 ..-6
+ Scatterband - "
[ (4-7,4-20) - 10-6 10-7 i - i i e iI i i i i I i i l-4 6 8 10 20 40 60 83 100 Stress intensity factor range, AK (ksi Si) Figure 4-6 Effect of Environment on Fatigue Crack Growth in Type 304
4-15 Stress intensity factor range, AK (MN/m3 /2) 4 6 10 20 40 60 80 100 10 ' I I i I ' i ' i i iil i'l - 10 3
~
T u
- 10~4 - s D
e : _
~
C
,u -
u -. o s s g - 10 ~ - 5c -
- 5 3.
= - r s
m , , :- _ Uo o o : - e' h - s. C -
-# c
.= _ - 10 a
*;' 3
~
8 Baseline : - b, _ a t= 4 dpper Bound of data _ y to 600*F (4-7) _ 23 10~ ~ 10-6
-7 , , ,,,; , , , , , , ii .
10 4 6 8 10 20 40 60 80 100 Stress intensity factor r6nge, AK (ksi /iii) Figure 4-7 Effect of Temperature on Fatigue Crack Growth Rate in Type 304 t L
4-16 test specimens (4-7). This study showed no significant relationship between (da/dN) and specimen geometry. This is fundamental to the fracture mechanics approach to fatigue growth as the geometry is accounted in the determination of AK. 4.3.6 Effect of Compositional Variation The effects of compositional variation (within the ASTM specification) such as would appear in heat to heat variations, has been studied and published (4-7, 4-22). The research indicates that there is no significant effect on (da/dN) in 304 due to heat to heat variation or melt practices, d 4.3.7 Effects of Cycle Parameters The type of cyclic loeding (cycle characteristics) consists of three aspects:
- 1) Frequency
- 2) Waveform (sinusoidal, sawtooth, etc.)
- 3) Stress Ratio (cmin/amax)
The ao level (omax/cmin) is accounted for in the AK level section. Each of the above factors is. discussed in detail below. 4.3.7.1 Frequency The effects of frequency on (da/dN) in 304 have been shown to be insignificant in air or water environments at temperatures below 550 F (4-7, 4-21). 4.3.7.2 Waveform The effect of waveform on (da/dN) in 304 has been shown to be r,- - e,- - -
4-17 insignificant for sinusoid, sawtooth (4-7, 4-22) and square waves (4-7,4-22,4-23). Ttus, although the time history of the structural response is not fully characterized, the baseline represents a bounding case. 4.3.7.3 R-Ratio The general trend of increasing (da/dN) with increasing R ratio is well documented in the literature. The effect of R ratios has been shown to be less significant at low temperatures (4-7). Thus a conservative estimate may be obtained for (da/dN) at low temperatures by considering the effect of R ratio on (da/dN) at high temperatures. An upperbound for the (da/dN) of 304 in 1000*F air and R = .750 is shown in Fig. 4-8 along with the baseline of Fig. 4.4. This upperbound should conservatively estimate (da/dN) for 304 under the conditions in question, i.e. those listed in Section 4.3. For convenience, the upperbound in Fig. 4-8 may be expressed as: 2 (da/dN) = 3.52 x 10-83g .08 (AK)5) 10.3 (da/dN) = 6.3 x 10~I4aK (AK<5) 4.4 Fatigue Growth in Stainless Steel Weldments Data are available on (da/dN) in 308, 308L, 309, 309L stainless steel weld f metals. In ten published studies, only three showed that stainless steel weld metals had higher (da/dN) values when tested under the same conditions as the base metal (4-7). In these three exceptions, the difference between the' fatigue growth rates in the weld and base metals was not significant. Fig. 4-9 shows data for 308 weld metal at 77 and 800*F (4-3, 4-24). The data for 800*F fall slightly above the baseline for 304 parent material (Fig. 4-4), but l the data for 77'F fall well below the baseline. The available literature l indicates that stainless steel weld metals generally exhibit (da/dN) values ! equal to or lower than stainless base metals. It has also been shown that l if
4-18 Stress intensity factor range, AK (MN/m3 /2) 4 6 10 20 40 60 30 100 10-3 ,, , ,
,,,l , , , , , , ,, ,
10-3
~
T 5
~
_ 10-4 -
$x
~
5
- f, O -
E g - -- 10 4 - c . c E - 8 C _
= . %
Ao -5 N = 3.52 x 10-8 3g2.08 , 10 __ dN o
- (4-7) . J 3 : %'
2 _ g _ __ 10-5 3 - e Baseline - E En _
.x M e U - "
10-6 __
-6
' d ~14 10*3 - 10 f = 6.3 x 10 AK (4-7) -
10~ i e ii1 e i e i I i 11-4 6 8 10 20 40 60 80 100 Stress intensity factor range. AK (ksi /iR) Figure 4-8 Conservative Upper Bound Curves for Type 304 e L.
! 4-19 Stress intensity factor range, AK (MN/m /2) 4 6 10 20 40 60 80 100 10-3 ,, , , , , , , , , , , , ,, 10-3 l
~
T 5
~
_ 10~4 - Scatterband - a
* (4-3,4-22,4-24) -
u x
- f.
E - E g - 10 4 -- 5 -
- 5 3
E - R
-5
- 4 10 5 5 ~
5 2 - ' g _ __ 10-5 5 5 E
- cn
_ Type 304 - t= Baseline _ g u U 10~" --
~
l : -
-6 10
-7 i i e iiI 10 i i e 1 I i Il-4 6 8 10 20 40 60 80 100 Stress intensity factor range, AK (ksi @
Figure 4-9 Fatigue Crack Growth Rates - Stainless Steel Weldments m
4-20 welding process has no effect on (da/dN) (4-7, 4-3). Based on the available literature, the analysis of (da/dN) in 304 base metal (Section 4.3) will conservatively predict (da/dN) in the stainless steel weld metals as well. 4.5 Fatigue Crack Growth in Carbon Steel Weldments Data available in the literature on fatigue growth rates were collected for E7018 weld metal. Fig. 4-10 shows an upperbeund for (da/dN) data in an E7018 butt weld deposit which was stress relieved (4-25). Only the data shown on Fig. 4-10 fell near the upper bound curve. Fig. 4-10 also shows an upper bound from three other types of similar weld metals with and without stress relief (4-26). No (da/dN) data for E7018 or similar weld metals were found above the baseline in Fig. 4-10. The literature states that weld metals for joining steels such as A516 Grade 70 exhibit lower fatigue growth rates than the base metals (4-19). Residual stresses may increcse (da/dN) but if these stresses are included in estimating crack growth rates, the data indicates that the A516 Grade 70 curves of Section 4.2 will conservatively predict (da/dN) for weldments. 1 1 4.6 Sumary of Fatigue Growth Rate Data l The available literature data for four classifications of material have been reviewed. These are SA516 Grade 70 and associated weld metals, as well as Type 304 stainless steel and weld oetals. Conservative bounding curves [ have been found and the effect of various conditions checked against the L baseline curves to assure that their use will provide conservative results. The (da/dN) curves for the relevant conditions may be sumarized as: For AISI 304 and Stainless Steel Weld Metal: 10.3 AK<5 da_ = 6.3 x 10-143g dN
-8 2 da = 3.52 x 10 3g .08 AK)5 dN L -
m , - _ - .---,,,y - y-.-,- ---, , ,,,--m.-- = = - ~ ~ ' ~'w*" ~ = ~* * ~ "
4-21 Stress intensity factor range, AK (MN/m / ) 4 6 10 20 40 60 80 100 10-3 _
,,l 10-3
~
~
T
-4
~
E 10 - D T -
~
t z 3 y -
-4 E 3 -- 10 y 2 - ,
E E - O 0 -
- Upper Bound ~
(4-25) ig q % D -5 - _- Upper Bound Baseline (4-26)
- J J - :
%s.
%u SA516 Grade 7
_ 10-5 5 - E - b b, -
- x a - E M u u - "
u 10'0 -
- 10-0 10-7 i e i iiil i i , i i i ii 4 6 8 10 20 40 60 80 100 Stress intensity factor range, AK (ksi 66)
Figure 4-10 Fatigue Crack Growth Rates - Carbon Steel Welaments 4
+
4-22 For A516 Grade 70 and AWS E7018: fjti = 3.8 x 10-10 6g3.7'46 ggg4,9 , dN jLt = 4.40 x 10-13 3g8 .0 11>AK)4.9 dN jfjt = 3.16 x 10-6 ag1 .4 3g)33 dN These curves are shown in Fig. 4-11. Note that the SA516 curve chosen will bound both the base metal and weld metal. Similiarly the Type 304 baseline will account for both the base plate and weldment. / 1 e
!b
,, , , , . , _ . , _ , , _ _ m.__ __
4-23 Stress intensity factor range, AK (MN/m /2) 4 6 10 20 40 60 80 -100 10-3 i i i iil i i i i i i
- iiI
_ b = 3.16 x 10-6 3gl .4 ] 10
, 10~4 - $ 5 -
c - z
-13 8.0 h=4.4x10 -- 10-4 x
C _ 3 z z
" ~
-5 R
E 10 _, q 3
~
B 2 : 2 5 - __ 10-5 3 - h - h
$u bdN= 3.8 x 10-10 gg3.726 _
u - u 10-6 --
- --. 10-6
~
10-7 , , , , ,; , , , , ,. , , , , 4 6 8 10 20 40 60 80 100 Stress intensity factor range, AK (ksi /iB) Figure 4-11 Summary nf Fatigue Crack Growth Rates Used for Analytical Models, SA516 Grade 70 . l-L.
f 4-24 Section 4 REFERENCES 4-1 Crooker, T.W. and E.A. Lange. "The Influence of Yield Strength on Fatigue Design Procedures for Structural Steels," Conference on Fatigue of Welded Structures, The Welding Institute, Abington, Great Britain, (July 6-9,1970). 4-2 Battelle Columbus Laboratories, Structural Alloys Handbook, "Abl5, A516, Steel," (1981). 4-3 Rocketdyne North American Rockwell, Fracture Mechanics Data Bank, Fatigue Crack Growth in Structural Alloys, Volume 1, (July 1973). 4-4 Wilson, A.D., "The Influence of Inclusions on the Toughness and Fatigue Properties A516-70 Steel," Journal of Engineering Materials and Technology, Vol. 101, (July 1979), p. 265-274. 4-5 Sullivan, A.M. and T.W. Crooker. "The Effect of Specimen Thickness on Fatigue Crack Growth Rate of A.'iCr 0 Pressure Vessel Steel," Journal - of Pressure Vessel Technology, M y 1977), p. 248-252. 4-6 American Society of Mechanica'. rngineers, Boiler and Pressure Vessel Code, Section XI, " Codes for In'arvice Inspection of Nuclear Power Plant Components," Section XI. 4-7 James, L.A., " Fatigue Crack Propagation in Austenitic Stainless Steels," Atomic Energy Review, Vol. 14, No. 1, (1976). 4-8 Suzuki, M., H. Takahashi, T. Shoji, T. Kondo and H. Nakajima, "The Environment Enhanced Crack Growth Effects in Stru::tural Steels for Water Cooled Nuclear Reactors," Institute of Mechanical Engineers Conference on the Effect of Corrosion on Fatigue, London, (May 1977). 4-9 Cullen, W.H. and K. Torronen, "A Review of Fatigue Growth of Pressure Vessel and Piping Steels in High Temperature Pressurized Reactor Grade i Water," NUREG/CR-1576, United States Nuclear Regulatory Comission, l Washington, D.C. , (1910) . 4-10 American Society of Mechanical Engineers, Boiler and Pressure Vessel Code Section XI, (Winter 1980 Addenda). 4-11 Bamford, W.H., " Application of Corrosion Fatigue Crack Growth Rate Data to Integrity Analyses of Nuclear Reactor Vessels," American Society of Mechanical Engineers, Paper No. 79-PVP-116, (1979). 4-12 Barsom, J.M., " Fatigue Behavior of Pressure Vessel Steels," Welding Research Council Bulletin, No. 194, (May 1974). 4-13 Garwood, S.J., " Fatigue Crack Growth Treshold Determination," The Welding Institute Bulletin, (September 1979), p. 262-5. L
4-25 4-14 Nibbering, J.J.W. and A.W. La11eman, " Low Cycle Fatigue Problems in Shipbuilding," Conference on Fatigue of Welded Structures, The Welding Institute, Abington, Great Britain, (July 6-9,1970). 4 '5 Richards, K.G., " Fatigue Strength of Welded Structures," The Welding Institute, Abington, England, (May 1969). 4-16 Shshinian, P., H.H. Smith and J.R. Hawthorne, " Fatigue Crack Propa- ~ gation in Stainless Steel Weldments at High Temperature," Welding Research Supplement, (November 1972), p. 527-s to 532-s. 4-17 Shahinian, P., H.H. Smith and H.E. Wilson, " Fatigue Crack Growth Characteristics of Several Austenitic Stainless Steels at High Temperature," Fatigue at Elevated Temperatures, ASTM STP 520, American Society for Testing and Materials, (1973), p. 387-400. 4-18 Bathias, C. and R.M. Pelloux, " Fatigue Crack Propagation in Martensitic and Austenitic Steels," Metallurgical Transactions, Volume 4, (May 1973), p. 1265-1273. 4-19 Hale, D.A., C.W. Jewett and J.N. Kass, " Fatigue Crack Growth Behavior of Four Structural Alloys in High Temperature, High Purity, Oxygenated Water," Paper No. 79-PVP-104, American Society of Mechanical Engineers, (1979). 4-20 Shahinian, P., W.E. Watson and H.H. Smith, " Fatigue Crack Growth in Selected Alloys for Reactor Applications," Journal of Materials, JMLSA, Vol. 7, No. 4 (December 1972), p. 527-535. 4-21 Bamford, W.H., " Fatigue Crack Growth of Stainless Steel Piping in a Pressurized Water Reactor Environment," ASME Paper No. 77-PVP-34, American Society of Mechanical Engineers, (1977). 4-22 James, L.A., "Effect of Heat-to-Heat and Melt Practice Variations Upon Fati5ue Crack Growth in Two Austenitic Steels," Properties of Austenitic Stainless Steels and Their Weld Metals (Influence of Slight Chemistry Variations), ASTM STP 679, American Society for TestingandMaterials,(1979). 4-23 James, L.A., " Hold-Time Effects on the Elevated Temperature Fatigue Crack Propagation of Type 304 Stainless Steel," Nuclear Technology, Vo. 16, (December 1972), p. 521-529. 4-24 James L.A., " Crack Propagation Behavior in Type 304 Stainless Steel Weldments of Elevated Temperatures," Welding Research Supplement, (April 1973), p. 173s-179s. 4-25 Maddox, S.J., " Fatigue Crack Propagation in Weld Metal and Heat Affected Zone Material," The Welding Institute, Abington, Great Britain, (1969). 4-26 Seeley, R.R. , L. Katz and J.R.M. Smith, " Fatigue Crack Growth in Low Alloy Steel Submerged Arc Weld Metals," Fatigue Testing of Weldments, ASTM STP 648, American Society for Testing and Materials, (1978),
- p. 261-284.
w.
5-1 5.0 CORROSION ASPECTS This report assesses fatigue and fracture considerations involved w th sensitized cladding on structural plate. Although some affects of en';ironment on fatigue crack growth were considered in the previous section, the effects of general corrosion or corrosion fatigue will be addressed more completely in other reports. For the purposes of this report, it should suffice to comment briefly on two aspects of corrosion influence on the fracture and fatigue behavior. The fracture analysis (Section 7.0) assumes a sharp defect in all evaluations of critical defect size. Although this may be a good assumption for the initial fissures which occur along grain boundaries, the effect of general corrosion may be to blunt the crack tip. Such crack tip blunting (which can also occur as a result of plastic zone formation at the crack tip) has been shown to lead to higher vr. lues of fracture toughness and thus larger allowable defects for a given stress condition. The fatigue analysis relies upon cyclic stress intensity (aK) as the rate-governing condition. This is known to be appropriate f'- fatigue. However, growth by a stress corrosion mechanism has been shown to be related to the maximum stress intensity, Kmax, and corrosion may in fact be unrelated to either absolute or cyclic stress level. If growth by corrosion and related mechanisms can be estimated or measured, then the additional effect of fatigue crack growth can be assessed by the methods outlined in the previous sections. l l 'k (' .1 'L
6-1 6.0 FRACTURE TOUGHNESS 6.1 Introduction The final two model inputs to be discussed are the material properties
, fracture toughness and strength. As discussed in Section 2.0, the applied stress intensity is compared to a critical value which is defined as the fracture toughness. Thus, to determine allowable flaw sizes, the fracture toughness must be characterized. Although no direct measurements of fracture toughness were performed in the course of this work, inference about the level of fracture resistance inherent in the material can be made by reference to the Charpy impact values which are available. This dist.ussion is presented in Section 6.2. These data are discussed in Sections 6.3 and 6.4 below for the drywell and containment materials. Section 6.5 analyzes typical crack opening displacement values to be used in the elastic-plastic fracture mechanics evaluation. Section 6.6 addresses the yield and ultimate strength values to be used in the limit load assessment. Certified Material Test Reports (CMTR's) for the drywell (CD-13 and CD-16) and the containment (CD-27 and CD-35) were analyzed to detemine Charpy (CVN), yield strength, tensile j strength and lateral expansion (LE) data. Analyses and tabulations of this
! information are given in Appendices E and F. In each case, standard statistical methods were used to arrive at the mean and standard deviation. Also given are the 90% and 95% occurrence levels at the 95% confidence levels for each data type. (These figures give the percentage of a large population which can be expected to meet or exceed a given value. For example, if the 95% occurrence level for CVN is 50 at a 95% confidence level, then one can say with 95% confidence that there is a 95% chance that a randomly selected sample from a large population will meet or exceed CVN = 50). Also shown are cumulative probability graphs for each data type. A sumary of these results is shown in Table 6-1. All tests were performed on the clad msterial in the thicknesses used in the l field (Drywell = 1" and containment = 1.5"). All CVN tests and LE tests were taken at 0*F whereas all yield strength tests were taken at room temperature ( 75*F). Comparison of field data with the APTECH toughness data base (6-1) was also made as a check.
6-2 Table 6-1 Sutt1ARY OF CMTR PROPERTIES DATA Devwell Vent .- One-Sided Highest Lowest Sample Standard Tolerance Property Value Value fiean Deviation Limit! Reference Charpy V-Notch 81.0 23.3 49.07 13.72 22.13 Table E-2 (ft.1b.) Fig. E-1 Yield Strength 2 65.5 45.7 52.54 4.25 44.19 Table E-3 (ksi) Fig. E-2 Yield Strength: 65.5 45.7 53.44 4.24 61.77 Tab'e E-4 (ksi) Fig. E-3 Tensile St ength 84.5 72.4 77.58 3.10 71.49" Table E-5 (ksi) Fig. E-4 Lateral Expansion 0.096 0.023 0.048 0.016 0.0164 Table E-6 (inches) Fig. E-5 Containment One-Sided Highest Lowest Sample Standard Tolerance Property Value Value f.iean Deviation Limit! Reference Charpy V-Notch 109.33 22.33 53.86 16.55 30.98 Table F-2 (ft.lb.) Fig. F-1 Yield Strength 2 60.10 41.00 48.11 3.86 40.44 Table F-3 (ksi) Fig. F-2 Yield Strength 3 60.10 41.00 49.41 4.22 51.80 Table F-4 (ksi) Tensile Strength 83.90 72.80 77.10 2.46 72.22 Table F-5 (ksi) Fig. F-3 ! Lateral Expansion 0.0903 0.0157 0.0573 0.0160 0.0255 Table F-6 Fig. F-4 i (inches) l l t 2At 95% confidence, the 95% level occurrence level 2 , Lower Bound 8 L Upper Bound l " Lower Limit l l l l
6-3
6.2 Background
To use the analyses described in Section 2.0, it is necessary to have the appropriate value of material fracture toughness in terms of the critical plane strain stress intensity factor, KIc. These data are not normally available for pressure vessel steels, such as SA516 Grade 70, in the temperature range of interest. For example, a literature data base compiled at APTECH (6-1) indicates that valid K Ic data for SA516 Grade 70 as defined by AETM E399 (6-2) are generally only available in reasonable thicknesses (that is 1" or 1.5") at temperatures of -150*F and below. Typical data are shown in Table 6-2 and Fig. 6-1. It is possible to infer information about the relative toughness of the present material from available CMTR's. There are several correlations that have been proposed to relate Charpy energy to K Ic values. These include two empirical relationships proposed and verified by Barsom and Rolfe (6-3). The relationship for the transition temperature regime is: Ic = 2 (CVN)3/2 (6.1) E where K = Plane strain fracture toughness (psi /in) Ic E = Young's modulus (psi) CVN = Charpy V-notch energy (ft lbs) The corresponding relationship for the upper shelf regime is: 1 IK Ic = _5- CVN (6.2) fy ) "y 20 where o, = Material yield strength (ksi) l K = Plane strain fracture toughness (ksi) Ic
i 6-4 f i Table 5-2 TYPICAL K IC DATA FOR SA516 W DE 70 f (Valid to ASTM E399) Temperature Thickness Range of Values (ksi / inches)_ No. of Data Points l 21 - 45 14
-250 2 37 - 72 23
-150 2 M
i Source: APTECH Data Base (6-1) 1 .i 4 a w c,- - - - <.- ,- w - -- - - - - . . - -+-,--,e - - - - - - - -,i-,,ew c-- .,----w
- --e-- -- - - - ~ , -
e
<..C V A _ J : S _ Ok A5,.6-70 S__ -.E
,,, , , ,- i,.. . . . , , , , ,_
+ _
+ _
u 60 -
*+
~
+
z _ y
- 40
+ *+
u) = + - M - w
+ .o.
x _
+ -
+
20 0 '' ' ' ' ' I'I' ' ' '
-300 -200 -10 0 0 100 200 TEMPERATURE, DEGREES F t e
! O) (. ) Figure 6-1 Typical Valid KIC Data For SA516-Grade 70 l
6-6 Barsom and Rolie found that at 80*F, the upper shelf correlation was , appropriate for all material they tested. All their tests were with material of f eld i strength greater than 100 ksi, although they claim that equation 6.2 is valid for materials with yield strength less than 100 ksi if dynamic yield strength is used instead of static yield strength. Another comon correlation, due to Sailors and Corten, which was developed for A533B and A51.7F (6-4), is (6.3) K Ic = 15.5 (CVN)0.5 where K yc
= ksi T5 CVN = ft lb Pisarski (6-5) who revinwed and verified by experiment ten correlations including those listed above, found that good predictions can be obtained for For lower strength steels, the high strength steels (cy > 113 Ksi).
correlations tend to be generally conservative with the degree of conservatism increasing with decreasing yield strength. Thus, either equation (6.2) or (6.3) should provide conservative estimates of critical fracture toughness. As a check, relations between critical crack opening displacement value and K are also available from Barsom and Rolfe (6-6) and Egan (6-7), and take l yg the form: O c
- IEIc) (6.4) c ys (o,3)2 where 6 = Critical crack opening displacement (in.)
c c ys
= Yield strain (in/in)
K = Critical fracture toughness (ksi/iii) Ic o = Yield strength (ksi) y3
6-7 A further evaluation of crack opening displacement values is found in Section 6.5. The next two sections analyze data on the clad material in the drywell vent and containment structeres respectively. - 6.3 Toughness Data for Drywell Vent Structure The details of the reduction of toughness data for the drywell vent structure are given in Appendix E. This Section will sumarize the methods and results presented there. The available data were provided by Gilbert Associates in the form of Certified Material Test Reports (CMTR) represented by (CD-16 and CD-13) . These reports were summarized according to quality control number (see Appendix E, Table E-1). The total number of quality control numbers found on field sheets was 80. The number of these values found on CMTR's was
- 69. For the balance, the lowest occurrence was taken for a given quality control number as outlined in the procedure listed in Appendix E. For ehch distribution, the mean, standard deviation and cumulative median rank were calculated. The 95% confidence level, 90% and 95% occurrence levels were then calculated. Table E-2 shows the cumulative Charpy values. Figure E-1 demonstrates the cumulative probability (median ranks) for the drywell vent Charpy values. The values calculated are sumarized in Table 6-1.
The 95% occurrence level, 95% confidence interval value of 22.13 ft.lbs. is used in equations (6.1) and (6.3) to arrive at estimates for the value of KIc of 79.0 ksi /in and 72.9 ksi /f5 respectively. For the results discussed in Section 7.0, the lower of these is used. Note that this will be conservative for the following reasons:
- 1) The correlation itself is conservative for materials in this yield strength regime.
- 2) The value for K Ic is calculated at 0*F (the temperature at which the Charpy data exist). At 70*F, the actual toughness will be higher.
- 3) A value of toughness (at the 95%/95% level) has bee, calculated and it is lower than any of the field test points.
6-8 6.4 Toughness Data for Containment Materials The toughness data for the containment plates were reduced in a similar manner. Appendix F oitlines the details of this analysis. The appropriate values are presented in Table 6-1. The 95% confidence level, 95% occurrence level Charpy value is 30.98 ft. lbs. Using the lowest occurrence of 22.33 ft lbs. will give KIc values from equations (6.1) and (6.3) of 79.6 ksi M and 73.2 ksi E respectively. The lower of these two values is used for the evaluations discussed in Section 7.0. These results are subject to the same inherent conservatisms in effect for the drywell vent as discussed above. Typical data from (6-1) are shown in Figure 6-2, along with the range of CMTR data in this project. As shown, the data are typical for SA51'6 Grsde 70 Charpy V-notch values. 6.5 Crack Opening Displacement (C00) Values The clad material will (particularly in the hypothesized fissured state) derive the bulk of its fracture resistance from the SA516 Grade 70 portion. Literature data are available (6-8) to provide typical C0D values. These were then used in two ways; first as a check in the derivation of K Ic values and second as direct input to the EPFM analysis. A check on derivation of the K Ic value used can be provided by equation (6.4). From the data in Table 6-3 and Fig. 6-3, the lowest C0D value data at 70 F is
.033" (note that O'F appears to be near to the upper shelf value so that there is a relatively constant lower bound between O' and 70*F). The fact that the lowest experimental value at 70*F is slightly lower than at O'F, is most likely due to random experimental error. For this value of COD, and for cys "
0.4%, o value is calculated as: y = 65.5 ksi, the resulting KIc K Ic = 103.9 ksi /lii Thus the value of 71 ksi Mtaken in Section 6.3 is confirmed as conservative.
CV \ V A _UES 20. . A5: 6-70 S~~ E _ 150
, T Project Range 22.3-131
+
+ .
100
+ '-
+
m ~ i
-> + _~ l-
+ :.:
l-- :: LL _ , _
$ _ g
- :: a Z == ==
U 3 =b 50 s a t' EE + $ + +
=
,q p
[! lb - Q.
+H+ + D + U + + -
+ + + -HW $ f,i +
g , , , , I , ,+, , I
, , , , I , , , ,
-200 -100 0 100 200 TEMPERATURE, DEGREES F Figure 6-2 Data Base CVN Values for SA516-Grade 70 Showing Project CMTR Range I e
6-10 Table 6-3 TYPICAL LITERATURE C0D VALUES Test Temperature (FO ) Thickness (inches) Range of Values (in._[ No.-of Data Points
.5 .001 .009 6
-160 1.0 .004 .005 3
.003 .006 2 1.5
.5 .004 .036 6
-110 1.6 .006 .009 2
.5 .032 .044 2
-105
.006 .016 4 1.0
.5 .009 .061 6
- 80 1.0 .012 .017 4 1.6 .009 .011 2
- 60 1.0 .027 .050 2
.5 .035 .071 7
- 40 1.0 .064 .079 4 1.6 .023 .111 3
.070 .073 2
- 20 1.0
.5 .035 .092 6 0
1.0 .077 .103 4
.117 .119 2 1.6
.5 .033 .081 6 70 1.0 .087 .119 4
.088 .108 2 1.6 l
(
I __. i C0J VA _UES _- O H A 5:.6-7 0 S __ _: _. 0.125 - 4 + -
+ 4
+
0.100 -
+
+ _
+ _
+ _.
- +
0.075 -
+ +
+ +
2 _ 4 + _
~ + + _ m
+
+*+
~
O 0.050 - o -
+ -
4
+ + + + _
+
$ +
0.025 + + - 4 -
+t !!
4 :: - 55 M ' ' ' ' l ' ' ' ' ' ' ' ' 0.000 0 100 200
-200 -10 0 TEMPERATURE, DEGREES F Figure 6-3 Typical C00 Values for SA516-Grade 70
6-12 The value of 0.033" C0D was also applied in the EPFM analysis, the results of which are presented in Section 7.0. 6.6 Other Material Properties Appendix E and F also address the evaluation of yield strength, ultimate strength and lateral expansion for the drywell vent and containment materials Their respective respectively. These values are summarized in Table 6-1. cumulative probability distributions are plotted in Appendix E and F. l
6-13 Section 6 REFERENCES 6-1 APTECH Engineering Services Fracture Toughness Data Base, Compiled for Electric Power Research Institute, Project No. 1757-2,(1981). 6-2 American Society for Testing and Materials Standard E-399, " Test for Plane-Strain Fracture Toughness of Metallic Materials," (1980). 6-3 Barsom, J.M. and S.T. Rolfe " Correlations Between Kg and Charpy V-Notch Test Results in the Transition Temperature Ranges," ASTM l STP 466, (1970), p. 281-302. 6-4 Sailors, R.H. and H.T. Corten, " Relationship Between Material Fracture , Toughness Using Fracture Mechanics and Transition Temperature Tests," ASTM STP 514,(1973),p.164-191. 6-5 Pisarski, H.G., "A Review of Correlations Relating Charpy Energy to K The Welding Institute Research Bulletin, (December 1978), p. 362-367.IC' 6-6 Rolfe, S.T. and J.M. Barse:n, Fracture and Fatigue Control in Structures: Applications of Fracture Mechanics, Prentice-Hall, (1977). 6-7 Egan, G.R., " Compatibility of Linear Elastic (K Ir) and General Yielding (COD) Fracture Mechanics," Engineering Fracture Mechanics (1973), Vol. 5,
- p. 167-185.
6-8 Stuber, A., J. Wellman and S. Rolfe, " Eighth Progress Report on Appli-cation of the C0D Test Method to the Fracture Resistant Design of Pressure Vessels," For the Subcomittee on Effective Utilization of Yield Strength of the Pressure Vessel Research Comittee, University of Kansas Center for Research, Inc., (February 1980). i
7-1 7.0 RESULTS OF ANALYSIS 7.1 Introduction For purposes of discussion, the regions of interest have been divided into three sections. The drywell vent structure is considered in two regions. One area, corresponding to a vertical seam, and the other is the most highly stressed region associated with a weld. These two regions were described in the section on drywell vent stresses, and are termed vent and edge regions respectively for discussion. The third region is the containment in the doubler plate region. The cases that have been analyzed are outlined in Table 3-2. They fall into two broad categories according to initial defect location. These are surface flaws as represent unrepaired surface connected fissures and buried flaws as represent incomplete repairs of fissures which are not surface connected. The following approaches for flaw growth have been followed.
- 1) For buried defects: Design fatigue cycles are applied. In no case did this cause the flaw to grow sufficiently to extend to the near surface. However, continued cyclic loading was applied until the flaw would have broken through to the free surface. The flaw was then modeled as a surface connected three-degree-of-freedom flaw.
l 2) Surface-connected flaws were subjected to the design fatigue loading. In no case did this cause the flaw to grow through wall during the proposed lifetime. However, growth was modeled to through wall behavior.
- 3) At some point before through-wall growth by fatigue, a failure of the remaining ligament in the cross-section would occur, causing a through-wall crack. The net section failure was modeled by a limit load criterion. It should be noted that this criterion determines the stress level required to cause failure of the uncracked ligament between the flaw a. d the free surface. This means that the flaw o
7-2 becomes a through-thickness flaw, which does not necessarily mean that the structure would fail by limit load. The structure is so large that even for a through-thickness flaw several inches in length, the loss of area is such a small percentage of the total area that the redistributed average stress increases very little. However, this does not rule out a failure of the structure by brittle fracture of a through-thickness flaw.
- 4) The final step was to assess the potential for fracture from postulated through-wall cracks.
It should be emphasized that for all assumptions about buried flaws, growth by fatigue over the plant design life never resulted in growth to the free surface. Similarly for a surf ace flaw, growth,never resulted in a through-wall flaw over plant design life. Performing the analysis to failure is provided as background to assist in the assessment of potential growth by other mechanisms such as corrosion or corrosion fatigue. Section 7.2 discusses the results of the limit load evaluations. Sections 7.3, 7.4 and 7.5 assess the three regions of interest: drywel-1 vent cutout area, drywell edge (vertical weld), and containment doubler plate detail, respectively. Section '.6 discusses a check on these results, using elastic-plastic fracture mechanics (C0D theory). 7.2 Limit Load Analysis As discussed in Section 2.0, limit load provides a bounding method of analysis for structural failure. It is based on two theorems. The first gives rise to lower bound solutions and states that a structure will not fail if the applied forces can be balanced by a re.listribution of stress such that the induced stresses do not exceed the yield or flow stress. The second theorem, which is an upper bound theorem, states that the structure will collapse when the rate of external work done by the applied forces exceeds the rate of internal plastic work for any collapse mechanism. As the first theorem provides lower bound results, it will be used in this analysis.
7-3 Many solutions have been devtiloped to calculate the critical stress for various geometries and loading conditions using the lower bound theorem. Several of these are reported in (7-1). For a center-cracked plate under uniform tension (Figure 7-la): og = c,(1-2a/t) (7,1) and for a single edge-cracked plate in tension (Figure 7.lb): o = f.\] at(1-a/t)[1+1n 1-a/2t for a/t 3 0.884 (7.2a) C (1-a/tj and o = 2.571cg (1-a/t) for a/t 2 0.884. (7.2b) g These solutions are for infinitely long flaws, which provide overly pessimistic solutions. The effect of flaw aspect ratio may be considered by replacing a with: a(1 -(1+j2f pt2 ) -1) 2 2 1-a(1+1 /2t )d/t) (7.3) For the present case, o, has been taken to be: og = 1/2 (cy + u) II*4) where cy and are the minimum values of yield strength and tensile strength u for any CMTR in either structure. This value is: og = 1/2 (41.0 + 72.4) = 56.7 ksi (7.5) The critical flaw size to cause plastic collapse of the uncracked ligament can be determined as a function of the applied stress by combining Equations 7.2 and 7.3 for a rurface flaw and Equations 7.1 and 7.3 for a buried flaw. Figure 7.2 is a plot of the critical flaw size for a surface flaw. Since the l postulated buried defect is near to the free surface, it is properly 1 considered in the limit load analysis for a surface crack.
= t =- _= t =
~
i 2a* _a.
?
^ -
- a. Center-cracked Plate b. Single Edge-cracked Plate i
Figure 7-1 Flaw Geometries for Limit Load Analysis
, , 7-5, , ,
a a a 3 60 - - _ 50 - - -- Aspect ratio = 2 40 - m
$ Aspect ratio = 10 E 30 - -
b m i E 20 - 10 - I I t t i i I i 0 .2 .4 .6 .8 1.0 l Non-dimensional flaw depth, ac/t Figure 7-2 Limit Load Failure - Critical Flaw Size for Surface Flaw
7-6 l l In summary, the limit load analysis indicates the following critical flaw ! sizes: l Defect SSE Critical Structure Region Orientation Stress (ksi) Depth (in) Drywell Vent Seam Transverse 30.4 .83 Longitudinal 8.1 .98 Plate Edge Transverse 20.5 .86 Longitudinal 7.9 .98 Containment Doubler Plate Transverse - - Longitudinal 4.5 1.48 7.3 Drywell Vent Seam Region Minor fissuring has been found in the region around the vent cutout and this has been repaired. Therefore, the initial modeling was as a buried defect, shown in Figure 7-3. The results are presented in Table 7-1. As is evidenced, there is little growth of the flaw over the design life of the
- plant. As noted in footnotes 2 and 3, the transverse buried defect would not even break through to the surface until the equivalent of over 1,000,000 SRV cycles or 12000 cyclic loads equivalent in magnitude to S.S.E. stresses. The longitudinal buried defect has even longer lifetime before it becomes a surface connected flaw. The limit load analysis indicated that the surface connected flaw would then have to grow to .83" in depth by some mechanism before it would pop through-wall. Table 7-2 indicates the margin against fracture for these cases. The applied K value for a defect after it has been subjected to the full range of plant cyclic stresses is approximately 13 ksi E substantially less than the critical value of 71.0 ksi /In computed in Section 6.0. The tolerance to a through-wall failure is noted in Table 7-2 to be approximately 3.2 inches and greater than 20 inches for the transverse and longitudinal orientations respectively.
I In summary, buried defects (of worst dimension) will not grow even to the nearest free surface with the cyclic stress levels and number of cycles expected. Substantial margin exists before a surface connected defect is reached, further margin before growth by fatigue reaches a critical limit
7-7 0 Fissure through to Carbon Steel Clad I
/ 6
\ v 2/3t' y t' j \ l
\ Buried Defect used in Analysis (I
I l
/ Carbon Steel t' = thickness of clad Figure 7-3 Bounding Buried Defect Sizing
Table 7-1 . Summary of Results Region: Drywell Vent Seam Fatigue Results Buried Defect Stress Type _ Growth (in.) Cycles I Flaw Orientation
.0002 16,740 Transverse SRV SSE .0001 60_
Total 2 .0003 16,800
< .0001 16,740 y Longitudinal SRV SSE < .0001 60 Total 2 < .0002 16,800 I Design number of cycles from CD-18.
2 To break through to surface would take > 1,000,000 SRV cycles or approximately 12,000 SSE cycles or some combination. Subsequent growth would be modeled as a surface flaw. , 3 To break through to surface would take > 1,000,000 SRV cycles or SSE cycles. 8
Table 7-2 Sunn,ary of Results Region: Drywell Vent Seam I Fracture Results K Length Through Wall IC Flaw Orientation K Over Total Design Cycles K IC K I Crack Where KIC " E a i 13.7 % 71.0 5.2 3.2" Transverse 12.6 % 71.0 5.6 > 20" Longitudinal w I Assumes a through wall defect. 2 See section 6.0 on Fracture Toughness. I
7-10 level, and the structure's tolerance to a through-wall defect has been calculated. 7.4 Drywell Vent Structure - Vertical Welds at Plate Seams l This drywell vent structure location has fissures detected by dye penetrant i methods. The analytical methods are similar to those outlined in Sections 7.1 and 7.3 above. Tables 7-3 and 7-4 outline the results for botih surface connected and buried flaw evaluations. Buried flaws were taken to be 2/3 the clad depth (see Figure 7-3). For both initial assumptions, the number of cycles to reach a critical defect (as defined by the limit load evaluation of Section 7.2) far exceed the expected number of design cycles. Table 7-5 illustrates the margin against fracture for a surface or buried defect which had been subjected to the full extent of design cycles. These values are 25.8 ksi /in and 10.9 ksi /Iii for flaws with transverse orientation compared with the critical value of 71.0 ksi /Iii. The length of through wall defect to reach an applied Kg equal to the critical toughness level is also presented. In this region, as with the more highly stressed vent cutout, neither buried nor surface connected flaw reaches a critical value within the expected design cyclic loading. A substantial margin against fracture exists for all flaw orientations and locations. . 7.5 Containment Doubler Plate Region The containment doubler plate region had indications of fissuring as well. Again these were modeled for both buried and surface defects and the growth
. compared to the limit load critical values. The results for fatigue are given in Table 7-6 and for fracture considerations in Table 7-7. As noted in the section on stresses, the SSE level of cyclic loading was applied over the entire range of expected cycles including the effect of SRV loading. Actual SRV level stresses were not accounted for. Had this been done, the values for amount of flaw growth would have been correspondingly smaller. As with the 1 other cases outlined above, neither buried nor surface flaws grow to a critical size over the plant design life by a fatigue mechanism.
Table 7-3 Sunnary of Results Region: Drywell Vent Structure Vertical Seams Fatigue Results 4
! Fissured (Surface Connected) Material Cycles I 2
Flaw Orientation Stress Type Growth Factor on Flaw Size Transverse SRV .0004 16,740 SSE .0003 60 8 5 Total .0007 16,800 '. 007 l-Longitudinal SRV <.0001 16,740 SSE < .0001 60
.98 = 9.7 Total < .00 02 16,800 .1002 I Design from CD-18.
2 critical flaw from limit load Factor on flaw size = maximum flaw growth + initial flaw (assumed) a l i l l l
1 Table 7-4 Summary of Results Region: Drywell Vent Structure Vertic-1 Seams Fatigue Results Buried Defect Cycles I Stress Type Growth Flaw Orientation
< .001 16,740 SRV 4 Transverse SSE < .001 60 }l ro 2 < .001 16,800 Total
- < .001 16,740 Longitudinal SRV
< .001 60 SSE 16,800 Total 2 < .001 Design Cycles from CD-18 '
2 Surface breakthrough would occur in >1,000,000 SRV cycles or >100,000 SSE cycles I l l
=
yG n e h W l l a W CI hK 2 2 a g u= 7 0 7 0 o 2 2 r , hK % > % > T h t g n e L C 8 3 3 3 I y K 2 3 5 6 2 0 C 0 0 0 I 1 1 K 1 1 7 7 7 7 s t l u 5 s
- e 7 R e f
e f l o i L b a y n T r g 8 8 4 3 a i s 5 1 3 1 m u e 2 2 1 1 S D t a s g s K s m a e e n h S g u l a o c t T c i t l l e e r n a a f r e o e n e n e u V i s i s i d t t r d r d c e a e u e u l a r t v t v t l r u n s i s i a F t e n g n g n w n c i a n a u r r o r o h g o r O T L T L t I u 0 S o s r 6 t h l t n l e l u o w s a i t y e s c r R e e D e p m S e y e u
- r c d s e n u T a e s e o t w f i A S i
g c a r r a u u e r l F S B I 2 R F
uj; B 2 2 s s e e l l c c y y C C 1 0 1 1 0 1 0 0 0 0 0 0 0 0 8, 0 0 8, 6 6 , n 1 < < 1 n r r e e v ) v o d o e h t h s t c t t w e w l o n o u r n r 6 s G o G
- e C 7 R f, e e c l o a b f a y r T r u a S m
m I t 3 u n o n S o N i o i t
, t a l d a l a
n t a e r t n e n o n e n u e s i i e s r i d s i r d g l a i r e u s r e u 5 6 e O v t i O v t - R i F s i 2 r s i g w n g 2 e e w n n a a n e t a a n r o r e t a a l r o o l F T L u r l M F T l r g u g P ) i i i a f r d p f e e e e l t R e e b c s e u e s s o n l a , l D n i l e ,
/ o t e v l t C r d e e n a o l d o e e m m c P
( s m i n t s f a wa s e w r t r a a l c l t u u e f t l n s S f s f o e ( e e C R D c E d d a S e
- e e d f S i r
n u r e r o g u i u t u i i s r S A B g t s u e a i B 1 2 3 R F F l l :
__ 7_ Table 7-7 Surnary of Results Region: Containment / Doubler Plate Region i Fracture Results i Fissured Material g IC Kg After 16,800 Cycles K Flaw Orientation (ksi G ) KIC (ksi 5 ) I ; 3 24.7 71.0 2.87 l Transverse 26.8 71.0 2.65 [ Longitudinal * , i Through-Wall Model Flaw Orientation Length Where K, = KIC a Transverse note 1 Longitudinal > 20 t 1 Applied Stresses are totally in compression , i
- 1 i
1
7-16 7.6 Elastic-Plastic Fracture Mechanics (EPFM) Results To assure that ;he linear elastic fracture mechanics analysis was conservative, ar. EPFH analysis was completed utilizing the concepts outlined in Section 2.1.2. The basis for this analysis is the " Draft British Standard of the Rules for the Derivation of Acceptance Levels for Defects in Fusion Welded Joints" (7-2). Since a rigorous elastic-plastic analysis is very complex compared to LEFM, the Draft Standard has simplified the procedure by using a semi-empirical design curve set in its current form by Burdekin and Dawes(7-3). Experimental work was performed by Burdekin and Stone (7-4). For a defect in a uniform stress field, the relationships between C0D and applied strain are: C = (e/ey ) , for 0 < e/e, < 3.5 (7.6) -
~
o = e/e - 0.25, for 0.5 < e'c y y <2 where & is the non-dimensional C00, i.e.,
&= 6 2ney i (7*7)
The C0D design curve, shown in Figure 7-4, relates the non-dimensional COD, 4, to the ratio of applied strain to yield strain. The applied strain is taken as the local strain which would exist in the vicinity of the crack if the crack itself were not present. It has been shown (7-5,) that this design curve f l 1s conservative, and thus the allowable flaw size, I, will be smaller than the critical flaw size, acrit* The design curve is based on a through-thickness defect (i = through-thickness crack half-length). Although not rigorously justified, it has been suggested ! that the effect of crack shape was the same for contained yielding problems as { for linear elastic problems (7-6). Thus, a pseudo-elastic-plastic solution to part-through cracks can be developed. For a through flaw with the geometry shown in Figure 7-Sa, the LEFM expression is: a.
I .t ; il ,: niwu
)
5 7 - f 3 e R m o r f ( e v
- r u
C g n i - s e D e s r 0 u 0 e Y 0 n I 2 c g / o s e 4 e -
- 0 7 0 e 0 r C
~ u g i F _ y s
. J t j 1
! I1-!,ill, l'
- i. , I
l 4 h=t * = t- H , I w v_ -u m-w i gh ; hl _ l s- 2a -m. lI d 2C r i 2C J'__ -.
- a + + 2a . L r co a) Through Flaw b) Surface Flaw c) Buried Flaw 4
Figure 7-5 Defect Geometry l 1 l I
7-19 I (7.8) For a surface crack with the geometry shown in Figure 7-5b, the LEFM-expression !s: MM K g
= TS oEa"
$2 (7*9) where M is the magnification factor due to finite thickness effects, M 3 is T
the free surface magnification factor, and2 4 is the elliptic integral of the second kind. It can M seen that E
- b"S )
surface flaw through-thickness flaw K (/ I
$2 (7.10)
From this relationship, solving for an equivalent through-thickness crack size gives: f"l'S k*h\#2/ (7.11) a were taken from a survey by Maddox (7-7), and The values of "T"S = f {a are shown in Fihbre 7-6 ' It can be similarly shown that for a buried elliptical crack with the geometry shown in Figure 7-Sc: (Mo Mr )2 a .a I j t t ( 0 2/ (7.12) where M g is the magnification factor at a point due to the nearest free surface and thr is the magnification factor at that point due to the more remote free surface. These values were derived from the finite element work by Shah and Kobayashi (7-8), and Hg and Mn are shown in Figures 7-7 and 7-8 respectively. The elliptic integral, 4 , is plotted in Figure 7-9. For a/c = 2 0; Mgr1, was derived from Fedderson's relationship (7-9): r 1/2 H = 1 sec "a (7.13) t (
'1
7-20 so . . . . , , ,
? >: _
- o. ,)
4o . .
.l >
2C~ NM s y %c,o, O l 2o - beecy
%c c3
/ lt -
* %~9g
/ l %,e , e e i e i i e i o os 02 03 as 05 os 07 os 0/g l Figure 7-6 Variation of (M3h/4) with Crack Depth for Various Shapes (from Ref. (7-7)) l l l I l lL l
MAGNIFICATION FACTOR FOR NEAREST FREE SURFACE 2.0 i i i i i i i i i i i i- i i i i j iiii a/c .. 1.8 0.10 - _ 1.6 M - 0.20 _ ~ s M b
~
1.4 _ O.40 O.60 _ f 0.80 - 1.2 -
/l00- _
/
/ 1 1.0
' ' ' ' "-'" '~i ' ' ' I' .
0 0.2 0.4 0.6 0.8 1 o/h m Figure 7-7 m
,I oo 1
- 0 - 0 0 0 '
c 0 1 2 4 6 8.0 0
/ O o1 E , 0 0 0 0 '
C A F ' R , U I 8 S 0 E , E ' R i F , T S i E 6, H 0 T R A F n1
~ h 8
7
~
/
a e R
" r u
O 5 g
=
i F 4 F R O
- - 0 i
T C i A F , N , O 2 I T 0 A , I C i F I , N G , A _- M - - - - - _ - . O 3 5 0 5 0 0 0 2 1 1 1 1 1 1 1 gs 7
, __ _ _ - _ _ _y,--. , ,, - __
f ELLIPTIC INTEGRAL OF THE SECOND KIND
. . 1.6 , , , , , , , , , . , , , , , .
~
1.5 ] _ 1.4
~
j _ i 1.3 i Of s u [ bl i 1.2 _ T t _ i 1.1 - l _
~
, , l , , , , l , , , , , ,, , l , , , ,~
3,g , i 0 0.2 0.4 0.6 0.8 1 1 a/c Figure 7-9 m i
7-24 Thus, combining the C0D design curve with Equations 7.7, 7.11 and 7.12, a relationship can be developed between allowable flaw size and allcwable strain levels for each crack configuration. The Draft Standard contains criteria which require that when a crack tip approaches a free surface, the flaw should be reassessed as a surface-connected flaw (a buried flaw is recategorized as a surface flaw and a surface flaw is recategorized as a through-thickness flaw). Using the buried flaw model shown in Figure 7-3, the flaw is close enough to the free surface to warrant a recategorization as a surface flaw. Thus, only surface-connected and through-thickness flaws will be evaluated here. The critical value for crack opening displacement was taken to be 0.033" as outlined in Section 6.5. The yield strain is assumed to be 0.4%. The calculated allowable flaw sizes for surface defects are shown in Figure 7-10 for the 1.0 inch thick drywell vent region and Figure 7-11 for the 1.5 inch thick containment doubler plate region. The theoretical development used is valid to strain ratios (e/ey) up to 2.5 but are plotted beyond that point to demonstrate that at flaw depths of approximately 10% of the wall, a very large margin to fracture exists. The results are shown for a variety of aspect ratios. For a/2c = 0.0, an infinitely long flaw is modeled. The draft standard (7-2) provides a recategorization process at a depth of 50% of the wall thickness. The flaw
- should then be considered as a through flaw.
l Figure 7-12 shows the results of the elastic-plastic through-flaw analysis. The length of a critical flaw as a fection of strain is given. For through flaws, residual strains are not included. For the present case, maximum applied strains to yield strain.are less than .5 so that flaws greater than 10" are allowed previous to fracture. Most l relevant cases have lower strains and correspondingly larger predicted allowable flaw lengths. Thus, the LEFM analysis (Tables 7-2, 7-5 and 7-7) is conservative when compared to the EPFM results. l Two notes are of interest here. First, the toughness used was taken from generic data and not plant specific information. Thus the flaw values
7-25 should only be used to gauge trie reasonableness of the t.EFM approach % was their intent. Second, this EPFM approach has an inherent minimum factor of safety of 2 on flaw size. 1 I I O v. L
;._u - -- . . _ . . _
7 ._. . STRAIN vs. CRITICAL FLAW SIZE (S"RFACE FLAW) 5 i i i i , , , , ,
\ ,a/2c-3.4 i i , i ii, .
4 -
~
a/2c-0.3
~
3 _ i' S a/2c-0.2 ~ e IR'1Nb ID - k 2 _ 1 a /2c-0.1 _ i 1 a/2c-0.0 0 O 0.2 0.4 0.6 0.8 1 FLAW DEPTH, INCHES N Figure 7-10 Strain versus critical flaw size for drywell vent structure (EPFM) r.
I STRAIN vs. CRITICAL FLKW SIZE (SURFACE FLAW) 5 . . .i ,i, ,i i , , , , ,,,, , , , , -
~
a/2c-0.5 [ 4 - _ a/2c-0.4 _ 3 - i _ _ g - a/2c-0.3 _
) ~
IE"1Nb if* - E 2 - _ a/2c-0.2 _ a/2c-0.1 1 _ n/2c-0.0 - 0 O 0.25 0.5 0.75 1 1.25 .5 1 FLAW DEPTil, INCHES m Figure 7-11 Strain versus critical flaw size for Containment / Doubler Plate (EPFM)
STRAIN vs THROUGH-THICKNESS CRITICAL FLAW SIZE 2.5 , , , i , , , , , , . . . , , i . , , l _ 6c -0.033 in. , 2.0 1.5 - j - g - N G)
?
n 1.0
- 0.5 -
0.0 ' ' ' ' ' ' ' ' ' ' ' ' ' ' ' O 2 4 6 8 iO FLAW LENGTH, INCHES Figure 7-12 Strain versus through-thickness flaw m (EPFM) l
7-29 Section
7.0 REFERENCES
7-1 Chell, G.G. , " Elastic-Plastic Fracture Mechanics," Central Electricity Research Laboratories, Report No. RD/L/R 200F, (August 1979). 7-2 British Standards Institute. " Draft Standard Rules for the Derivation of Acceptable Levels for Defects in Fusion Welded Joints," WEE /37, February,1976. 7-3 Burdekin, F.M. and Dawes, M.G. , " Practical Use of Linear Elastic and Yielding Fracture Mechanics with Particular Reference to Pressure Vessels", Institute of Mechanical Engineers C5/71. 7-4 Burdekin, F.M. and Stone, D.E.W., "The Crack Opening Displacement Approach to Fracture Mechanics in Yielding Materials", Journal of Strain Analysis, Vol. 1, No. 2, (1966), p. 194. 7-5 Harrison, J.D. et al, "The C0D Approach and Its Application to Welded Structures", The Welding Institute Research Report 55/1978/E, January (1978). 7-6 Dawes, M.G., " Fracture Control in High Yield Strength Weidments", Welding Journal Research Supplement, Vol. 53, (1974), p. 369s. 7-7 Maddox, S.J., "An Analysis of Fatigue Cracks in Fillet Welded Joints", 4 International Journal of Fracture, Vol.11. No. 2, April (1975), pp. 221-243. 7-8 Shah, R.C. and Kobayashi, A.S., " Stress Intensity Factors for an Elliptical Crack Approaching the Surface in a Semi-infinite Solid", . International Journal of Fracture Mechanics, Vol. 9. No. 2, (1973),
- p. 133.
7-9 Fedderson, C.E., " Discussion to Plane Strain Fracture Toughness Testing", ASTM STP 410,(1967),p.77. i 6 t
8-1 8.0
SUMMARY
AND CONCLUSIONS e Evaluation of the potential for defect growth by a fatigue mechanism and subsequent fracture has been assessed for three regions of interest. These are: Drywell Vent Seam Drywell Vent Structure Vertical Plate Seam Containment Structure Doubler Plate o Conserva^.ive (fastest) fatigue growth rates and conservative (lowest) fracture toughness levels were combined with design stress levels in all regions. , e The results indicate that should buried flaws exist in any of the three regions of a size equal to 2/3 of the clad thickness, they will not grow by a fatigue mechanism to the nearest (inside) free surface over the design life of the plant. e For all three locations, should surface connected defects equal to the depth of the clad exist, they will not grow by fatigue to a critical size (as defined by either limit load or fracture criteria) for the postulated design lifetime. e The margin against fracture as provided by the difference between expected applied stress intensity, K and conservative estimate of the fracture toughness, K Ic has been evaluated for each location and flaw orientation. e The results of this analysis can be used in conjunction with evaluations of corrosion and stress corrosion cracking aspects of the regions of interest. The combined effort will allow an evaluation of the significance of the sensitized stainless steel in the Perry Nuclear Power Plant.
A-1 APPENDIX A CONTROLLED DOCUf1ENTS 1 i e i e O f l 1 l,. I
A-2 DOCUMENT CONTROL NUMBER DESCRIPTION . la Maps of the Containment Vessels (4 pages) Units 1 and 2 lb Maps of Drywell Vent Structures (8 pages) Units 1 and 2 2 GAI Drawing B-301-734, Revision E, " Quencher Arrangement Design Envelope" 3 NNIC Drawing 249746, Revision 0, " Quencher Support Details and Arrangement" 4 Detailed Welding Procedure Specification No. WPS 8.1.4.1A, Revision 2 for welding quenchers to the stainless clad base plate by General Electric 5 Welding Procedure Specification No. WPS 1.2.3-001, Revision 1 for welding floor liner plate by NNIC 6 Welding Procedure Specification No. WPS 1.1-O'14 Revision 3 for welding the containment by NNIC 7 Welding Procedure Specification No. WPS 1/8.1-001, Revision 1 for welding the containment by NNIC 8 Welding Procedure Specification No. WPS 1.1-018, Revision 2 for welding the drywell vent structure by NNIC (. I 9 Welding Procedure Specification No. WPS 1/8.1-002, [ Revision 1 for welding the drywell vent structure by NNIC. I 10 Welding Procedure Specification No. WPS 1.3-001, l Revision 2 for welding the drywell vent structure by NNIC 11 Project D - Test 140 - Stainless Clad Steel Bond '. Integrity Tests Report 12 " Report on Stainless Steel Clad Plate Material Used in Construction of the Perry Nuclear Power Plant, Units #1 & #2 Rev. A, May 14, 1981, including original photos 13 Material Test Certificates for the Drywell Vent Structure, Units #1 and #2 (24 pages)
A-3 DOCUMENT CONTROL NUMBER DESCRIPTION - 14 NNIC Drawing 249632, Revision A, "Shell Doubler
- Detail - Upper Containment Vessel" 15 NNIC Drawing 249629, Revision D, "Shell Embedment Containment Vessel" 16 Typical section of drywell vent structure showing stresses 17 Section 6.04, " Containment Shell" (pp 381-392 of NNIC stress report) showing stresses in the lower portions of shell where fissuring is presently defined 18 Estimate of fatigue cycles 19 Cover letter for documents PY-STR-1365, contains water chemistry, etc.
20 Non-conformance Report No. 17-149 21 Non-conformance Report No. 17-138 22 Non-conformance Report No. 96-471 D 23 Non-conformance Report No. 96-769, missing page 17 of 18 I 24 Drawing NNIC 249813, Rev. C, "Drywell Vent - Structure Segnent Details" 25 Drawing NNIC 249802, Rev. D, "Drywell Vent Structure Segment Assembly" 26 Drawing NNIC 249801, Rev. E, "Drywell Vent Structure Segment Assembly" 27 Containment CMTR's, Units 1 and 2 1 28 Floor liner CMTR's and summary sheets 29 NNIC Drawing 249640, Revision E, " Floor Plate Details Clad Steel Floor Area - Units 1 and 2 - Containment Vessel" 30 Drywell Vent Structure Model, Pages 3:21.2-40 and 3:21-40.1 h
A-4 DOCUMENT CONTROL NUMBER DESCRIPTION - 31 Doubler Plate Analysis, Section 6.0.3, pages 370-380 32 Containment Vessel Specification, pages 44, 45 and 46 of SP-660, Temperature Information 33 Cover Letter for Metallurgical Samples Received (Original with Samples) 34 Drywell Vent Unit Stresses Agenda Meeting 11-3-81 35 Page 17 of 18 NRC 96-769 37 GE WPS 8-2.4.6A, Automatic Welding of Quenchers 38 Quenchers Base Plate CMTR, 76NNI319 39 Computer Run WOI 40 Computer Run WO2 3 41 Computer Run WO3 42 Computer Run WO4 43 Computer Run WO6 44 Computer Run WO7 45 Computer Run WO8
- 46 Computer Run WO9 47 Computer Run W10 48 Computer Run W12 49 Computer Run W13 50 Computer Run W14 l
51 Computer Run W15' 52 Computer Run W16
- l. .
L
A-5 DOCUMENT CONTROL NUMBER DESCRIPTION , 53 Computer Run W17 54 Computer Run W18 55 Computer Run W22 56 Computer Run W21 e 57 Computer Run W20 58 Computer Run W19 59 Computer Run 09A 60 SA516 Gr. 70 Data Base CVN Data 1 61 SA516 Gr. 70 Data Base C0D Data . 62 SA516 Gr. 70 Data Base KIc Data 63 Magnification Factors, Elliptic Integral Plots
, 64 WEE 37 Strain versus Through Flaw 2
65 WEE 37 EPFM t=1.0" 66 WEE 37 EPFM t=1.5"
- 67 Drywell Vent Seam Stress Calculations 9 Pages Plus Figure 68 Drywell Vent Edge Stress Calculations l
9 Pages Plus Figure [ 69 Containment Stress Figures 70 Computation of Limit Load Values 71 Residual Stress Data L 72 Crack Growth Data i j 73 Statistical evaluation of crack growth data l-74 Grinding Residual Stresses 75 Mill Sheet Properties - Drywell and Containment - Methods and Results f
~
L
r A-6 DOCUMENT CONTROL NUMBER DESCRIPTION 76 Final Report - Fracture and Faticue (Draft) 77 Final Report - Corrosion Aspects (Draft) 78 Preliminary Report - Fracture and Fatigue To: P. Gudikunst 79 Letter: P oposed GE Automatic Welding Procedure, Quencher Support T-stiffeners. November 9,1981 80 Letter: Recommendations for Repair Welding of Fissured, Sensitized Stainless Steel. November 9, 1981. 81 Letter: Review of Quencher Support T-stiffeners Weld Procedures 82 Letter: Corrosion Aspects of Fissures in Sensitized Stainless Steel Clad Containment Liners. October 6, 1981. 83 Letter: Crevice Corrosion Aspects of Containment Plates. September 16, 1981 3 84 Letter: Project Quality Assurance Procedure. Rev. 0 7-24-81. s 85 Minutes Meeting, November 3, 1981. 86 Distribution of Gilbert QA Requirements, Sent September 8, 1981. 87 Purchase Order 196451 from GAI dated 8-28-81 (w/ QARequirements) 88 GAI/QAD Evaluation of AES Proposal dated August 27, 1981 89 Memo to G. Egan, From E. Pejack
Subject:
GAI Audit of AES QA System, dated September 28, 1981. 90 Cover letter - receipt comments on draft comments 91 Draft Report on Fracture and Fatigue w/ comments from GAI 92 Draft Report on Corrosion w/ comments from GAI 93 NCR NR96-493 1 M
B-1 APPENDIX B CALCULATION OF DRYWELL VENT SEAM AND PLATE EDGE STRESSES i I r I I l 1
B-2 Method The drywell vent seam and plate edge stresses were calculated from (CD-34) , unit stresses. They were evaluated for six element locations--three in the vent seam, three in the vertical seam. The worst location in each group was used in the fracture and fatigue calculations. Normal operating load combina-tions were derived from: B-1 S = D + L + G + To + PSRV + TSRV + Ro B-2 S = D + L + G + F,q, + To + PSRV + TSRV + Ro B-3 S = D + L + G + F,q, + To + PSRV + TSRV + Ro
. where D = dead load To = 100 F L = live load T stresses from SRV G = hydrostatic pressure included p SRV SRVJ
() Activation in LOCA (Pa) Rg = pipe reaction For which (1) L is conservatively taken as zero. l (2) G is self-balancing. (3) TSRV ignored where it reduces other effects. } (4) Ro = 0. Thus the equations become: S=D+P B-4 SRV 1 S = D + Feqo + P SRV B-5
. S = D + Fegs + P B-6 SRV Using the outside face stresses (which are higher than inside face) and +1 PSRV stresses or +19 P stresses, whichever gives the highest value (+1 PSRV IU"
, SRV e and o , +19 P x SRV
#0" "XY), the following load combinations are obtained (ksi):
B-3 SRV (ox) = 0.56 SSE(ox)= 4.40 SRV (cy) = 5.04 SSE (cy) = 14.21 SRV (oxy) = -0.48 SSE (oxy) = 3.87 The appropriate unit stresses are: Location Source ox gy, g Unit x 1.12 .01 .01 A Unit y .01 1.10 .02 Unit xy .01 .05 1.06 Unit x 1.12 .15 .95 B Unit y .01 1.44 .08 Unit xy .03 .58 1.01 Unit x 1.13 .32 .02 C Unit y .01 1.58 .02 Unit xy .01 .14 .91 Unit x 2.42 .53 .47 1 Unit y .37 .12 .43 Unit xy .87 .57 2.37 Unit x .50 .08 .19 2 Unit y 2.48 .51 .19 f -2.07 ! Unit xy .67 .51 l .31 3 Unit x 2.53 .65 Unit y .49 .03 .2 L Unit xy .92 .60 2.49 1 The resulting component stresses (ksi) are thus found to be: SRV(cy) SSE(ox) SSE(cy) Location SRV(ox) 1.9 30.5 8.1 (2) 12.5
.6 7.9 4.9 20.5 (C)
Locations 1 and 3 were less than 2. Similiarly, A and B were found to be lower than C. Care must be taken to apply the component stresses in the appropriate ( orientation. L ._ i
C-1 6 APPENDIX C CONTAINMENT STRESS
SUMMARY
f a f I l e e 4 e l' a
c __ - ; . . - - ;. . _ _ - 2-___--...-- - .- _ . . _ -.- _- _ . . .; . - . . , 1 1 Summary of Containment Stresses - Selected Bounding Cases Sumary of Contatament Stresses - SelecteJ Bounding Cases Combination Stress StressesTotalf5f} D.L* L.L* T T R G G* F F
,_e' p yF F* P r P P.
_Num6er Component o a o no f _e_g a 5 s MJs. 6 33, - M
$ - Nomal og -892 518 -97 - - -21 - *l10'. -20 - - - 4% -42 - 1777 -2539 3766 ,
Operating -826 218 o -238 I?? -5910 -
- 339 - f381 56 - - - 223 1706 - -3321 -7530 4209 t
-190 -1429 7 - Nomal og -892 518 -97 - - -21 - - - +1786 -40 - 465 -42 - 1937 1239 5176 Operating -826 218 o -238 122 -5910 - - 339 - - -
1581 112 - 223 1706 - -3065 -7668 4603 L
-190 -1423 m
e 12 - Short o, -892 518 -97 - - - - - - 11726 -40 1035 - - -59 2480 -2665 5145 N Tem LOCA -826 170 og -23R I?? -M10 - - - - - - 1581 112 1295 - - 1276 -2762 -6W W
-190 -1074 12 - Long o, -892 518 - -323 - -11 - - - +1786 -40 3424 - - - 4462 -414 4916 Tem LOCA -826 o -238 122 - -IR482 - 223 - - - 1581 112 1643 - - -
-16039 -17513 1974 1
-190 Worst case taken as c, = -0.5 to +5.0 kst. Mg = 5.5 kst og
= -2.5 to -7.0 ksi. Mt " "*
Notes: 1) Stresses for Elevatton 575'-10"
- 2) Data from CD-17 e
- 3) Note definedthatina figure here nomal 2-3. to longitudinal defect. og nomal to transverse defect as
D-1 APPENRIX D RESIDUAL STRESSES DUE TO WELDING L
D-2 Introduction Figures D-1 through D-4 illustrate the results of a literature review to charac-terize the state of residual stress due to welding. The materials examined were low carbon and carbon manganese steels in the range 1" 5/8" thick. The illustrated distributions are normalized to a percentage of yield and to percentage through wall. Where double sided weld data were given, the associated single side distribution was derived and is presented. The data shown along with the bounding curve are: Figure Stress Orientation Distribution D-1 Transverse (to weld) Surface D-2 Longitudinal Surface D-3 Transverse Through Thickness D-4 Longitudinal Through Thickness
- L.
I I I 100 1 1 i 1 i i 80--- - - 3
.$ 60_ .
o 40__ Di ,
/ stribution used e
8 20- - S "
! 7 : . ,,
g 0; - -u------ E ?" g - -~- E M - - 3 - 8
-100 f i n 3
- 9. weld 1 2 3 4 .i 6 Distance from weld centerline, inches key x - Nordell & Hall (_3-1) e 0 - Nordell & Hall (3-1)
Figure D-1. Butt Weld Transverse Residual Stress Distribution on Surface Away from the Weld.
100 l l {- l l j _ 80-- 60- - a
-{ Distribution used l ,
40- - a E 20 --
- a B p o
=
g, 0 -
' o I :
2 h e a,, - l E a - o 40 --
*2 - -
4 -
"m E .go - .-
I_
-100 l l _ _ h--
4
- -- b 1 1 2 3 l
weld Distance from weld centerline, inches key x,o - Nordell & Itall (3-L) o - Greene ( 3-2) Figure D-2. Butt Weld Longitudinal Residual ', tress Distribution on Surface Away from Weld.
I 100 ( ! ! 80 - - o 60 - - ^ Distribution used
-~
g
^
40 "- , x [ -
, o w'
~
20 - - x --
$# x *-
O x 0 '
; 'E ?,
u c' E - n -- E o e a o b, v o AP 3 m (
- 7,-
-JL-g -
-100 ' ' ' '
O 2'O 4'O d0 80 100 Position, % of wall thickness kt x - Nordell & Itall (hl) * - Leggat & Kamath (1-2.) a - Legatt 8 Kamath (.3.-Z) o - Nordell a llall (1 1) n - Nordell & fla11 (2 .1.) x - Rosenthal & Norton (3.-1) Figure D-3. Butt Weld Transverse Through Thickness Residual Stress Distribution. l l
100 y h "
" c o
- L
~ ' ~ ~
g Distribution used
- a 9
. 60- -
x j w a 40- - - o e
- si g 20-y - * - -
B E 0 8' ? cn e m -20 - - I E
'm~
a c'
-80 - - -
-100 i i i f i i _
0 26 46 5'O 66 8'O' 100 Position, % of wall thickness , key x,0,0
- Nordell & Hall (3-1)
- Rosenthal & Norton (2_4)
Figure D-4. Butt Weld Longitudinal Through Thickness Residual Stress Distribution.
E-1 APPENDIX E STATISTICAL EVALUATION OF DRYWELL VENT MATERIAL PROPERTIES b
E-2 INTRODUCTION A statistical analysis of material properties for the drywell vent structure was made. The results of this analysis were used in. conjunction with the linear elastic, elastic-plastic, and limit load fracture mechanics evaluations. The following properties were evaluated: Charpy V-notch, y1ald strength (a lower bound for limit load evaluations and an upper bound for residual stress characterization), tensile strength, and lateral expansion. The method outlined below indicates what evaluations were done and how account was taken for missing information. STATISTICAL METHOD (1) Take all values of CVN, yield strength, and LE that appear on CMTR sheets (CD-13). (2) To account for missing CMTR's, we assume worst properties. Thus, for LE and CVN, add (to the values taken in (1) above), the lowest value for a given quality control number, the number of times that number is found in the maps (CD-16) but not in the CMTR's (CD-13). Subtract the highest value for a given quality control number, the number of times
- that number is found in the CMTR's but not in the maps. Add the lowest value of any quality control number, the number of times illegible or missing quality control numbers appear in the maps.
- (3) For yield strength lower bound, follow (2) above. For yield strength upper bound, follow (2) above except add the highest values and subtract the lowest values (in all cases).
(4) Compute mean, standard deviation, and median rank cumulative probability for each distribution. (5) Compute mean plus three standard deviations. Any data greater than this value are outliers and should be neglected. ! (6) Compute tolerance limits at 95% confidence level, and 90%, 95% and 99%. (All of these will be lower limits, except the yield strength upper bound which will be an upper limit.) 1
E-3 NOMENCLATURE l i a- Mean of sample s = Standard deviation of sample n = Number of values in sampling y = Confidence level P = Occurrence level k = Number such that the event that a given value is greater than x + ks has a probability y at occurrence level P, or Pr{Pr(i ( x + Ks) >, P} = y In summary, the properties are final to the following statistical values: ONE-SIDED HIGHEST LOWEST SAMPLE STANDARD TOLERANCE PROPERTY VALUE VALUE MEAN DEVIATION LIMIT REFERENCE Charpy V-Notch 81.0 23.3 49.07 13.72 22.13 Table E-2 (ft. lb) Fig. E-1 2 Yield Strength 65.5 45.7 52.54 4.25 44.19 Table E-3 (ksi) Fig. E-2 3 Yield Strength 65.5 45.7 53.44 4.24 61.77 Table E-4 (ksi) Fig. E-3 4 Tensile Strength 84.5 72.4 77.58 3.10 71.49 Table E-5 (ksi) Fig. E-4 Lateral Expansion 0.096 0.023 0.048 0.016 0.0164 Table E-6 (inches) Fig. E-5 I At 95% confidence, the 95% occurrence level 2 Lower bound 3 Upper bound 4 Lower Limit l l
Table E-1 PNPP DRYWELL CMTR STATISTICS NUMBER OF TIMES tlUMBER OF TIMES NUf1BER OF TIMES QUALITY NUMBER OF TIMES FOUND ON DRYWELL FOUND ON MAPS FOUND ON CMTR'S CONTROL FOUND ON DRYWELL MAPS DUT NOT ON CMTR'S BUT NOT ON MAPS NUMBER CMTR'S 5 1 76NNI478 4 6 6 76HNI479 1 76NNI480 1 3 1 76NNI481 2 2 2 76NNI482 4 4 76HNI483 2 2 76NNI484 2 2 76HNI485 2 3 1 76NNI486 2 2 76HNI487 m 4 4 76NNI488 1. 1 76NNI489 . 1 5 6 1 76NNI490 5 3 76NNI491 2 1 76NNI492 3 2 4 6 2 76NNI493 4 4 76NNI494 1 2 1 76NNI495 2 5 2 76NNI496 3 2 2 76NNI497 2 2 76NNI498 1 2 76NNI499 1 1 0 1 BLANK 1 1 77NNI141 77NNI142 1 I i 2 2 77NNI203 1 77NNI293 1 1 77NNI294 1 1 76NNI604 1 2 2 76NNI605 80 13 4 TOTAL 69
E-5 Table E-2 CHARPY V-NOTCH VALUES CHARPY V-NOTCH VALUE NUMBER OF TIMES CUMULATIVE (ft-lcs) OCCURRING PROBABIL'ITY (%) 23.3 3 0.87-3.36 30.7 1 4.60 31.3 2 5.85-7.09 32.7 4 8.33-12.06 34.0 2 13.31-14.55 36.7 2 15.80-17.04 37.3 6 18.28-24.50 39.3 2 25.75-26.99 39.7 2 28.23-29.48 40.0 1 30.72 42.3 2 31.97-33.21 42.7 1 34.45 43.0 2 35.70-36.94 44.7 6 38.18-44.40 45.3 2 45.65-46.89 48.0 2 48.13-49.38 49.3 2 50.62-51.87 50.0 1 53.11 51.3 1 54.35 52.0 1 55.60 53.7 4 56.84-60.57 54.3 2 61.82-63.06 55.3 1 64.30 57.3 2 65.55-66.79 57.7 2 68.03-69.28 58.0 1 70.52 58.7 3 71.77-74.25 60.0 2 75.50-76.74 60.3 2 77.99-79.23 60.7 2 80.47-81.72 65.0 2 82.96-84.20 69.3 2 85.45-86.69 71.3 3 87.94-90.42 73.7 1 91.67 78.0 1 92.91 81.0 2 94.15-95.40 131.0 3 96.64-99.13
- Average of three tests NOTE: High Value = 131.0 ft. lbs. Mean = 52.14 ft. lbs.
Low Value = 23.3 ft. lbs. n = 80 values Standard Deviation = 20.65 ft. lbs. Ignoring outliers at 131 ft lbs.: High Value = 81.0 ft. lbs. Mean = 49.07 ft. lbs. Low Value = 23.3 f5. lbs, n = 77 Standard Deviation = 13.72 ft. lbs. At 95% confidence: Occurrence Lower Tolerance Limit 90% 27.68 ft. lbs. 95% 22.13 ft lbs.
E-6 Table E-3 YIELD STRENGTH LOWER B0UND VALUES VALUES NUMBER OF tit 1ES CUMULATIVE KSI x 10 0CCURRING PROBABILITY (%) 457 5 0.87-5.85 467 4 7.09-10.82 468 3 12.06-14.55 476 1 15.80 481 1 17.04 482 2 18.28-19.53 486 2 20.77-22.01 489 2 23.26-24.50 491 1 25.75 492 1 26.99 495 1 28.23 496 3 29.48-31.97 506 2 33.21-34.45 507 2 35.70-36.94 515 5 38.18-43.16 523 2 44.40-45.65 531 2 46.89-48.13 532 4 49.38-53.11 534 2 54.35-55.60 536 3 56.84-59.33 538 2 60.57-61.82 541 4 63.06-66.79 548 2 68.03-69.28 551 2 70.52-71.77 553 3 73.01-75.50 556 1 76.74 558 2 77.99-79.23 i 560 2 80.47-81.72 564 1 82.96 571 2 84.20-85.45 572 1 86.69 578 2 87.94 89.18 579 2 90.42-91.67 582 1 92.91 593 2 94.15-95.40 609 2 96.64-97.89 655 1 99.13 NOTE: High value = 65.5 ksi Standard Deviation = 4.25 ksi Low value = 45.7 ksi n = 80 values Mean = 52.5 ksi i At 95% confidence: Occurrence' Lower Tolerance Limit 90% 45.91 ksi 95% 44.19 ksi i u
E-7 Table E-4 YIELD STRENGTH UPPER BOUND VALUES VALUES NUMBER OF TIMES CUMULATIVE KSI x 10 OCCURRING PROBABILITY (%) ~ 457 2 0.87-2.11 467 1 3.36 468 1 4.60 476 4 5.85-9.58 481 1 10.82 482 1 1E.06 486 2 13.31-14.55 489 2 15.80-17.04 491 1 18.28 492 1 19.53 495 3 20.77-23.26 496 2 24.50-25.75 506 2 26.99-28.23 507 2 29.48-30.72 515 5 31.97-36.94 523 2 38.18-39.43 531 1 40.67 532 4 41.92-45.65 534 2 46.89-48.13 536 2 49.38-50.62 538 2 51.87-53.11 541 4 54.35-58.08 548 3 59.33-61.82 551 4 63.06-66.79 68.03-70.52 553 3 556 1 71.77 558 2 73.01-74.25 560 2 75.50-76.74 564 2 77.99-79.23 571 2 80.47-81.72 572 1 82.96 578 2 84.20-85.45 579 2 86.69-87.94 582 2 89.18-90.42 593 2 91.67-92.91 609 3 94.15-96.64 655 2 97.89-99.13 NOTE: High Value = 65.5 ksi Mean = 53.4 ksi Low Value = 45.7 ksi n = 80 values Standard Deviation = 4.24 ksi At 95% confidence: Occurrence Upper Tolerance Limit 90% 60.05 ksi 95% 61.77 ksi
E-8 Table E-5 TENSILE STRENGTH VALUES NUMBER OF CUMULATIVE VALUE (KSI) OCCURRENCES PROBABILITY (%) 72.4 3 0.87-3.36 72.5 3 4.60-7.09 73.1 2 8.33-9.58 74.0 3 10.82-13.31 74.6 2 14.55-15.80 74.8 7 17.04-24.50 74.9 2 25.75-26.99 75.0 2 28.23-29.48 75.2 2 30.72-31.97 75.9 2 33.21-34.45 76.1 3 35.70-38.18 76.7 4 39.43-43.16 76.8 1 44.40 77.3 2 45.65-46.89 77.4 1 48.13 77.5 1 49.38 77.6 4 50.62-54.35 77.7 2 55.60-56.84 78.2 2 58.08-59.33 78.5 1 60.57 78.9 2 61.82-63.06 79.1 2 64.30-65.55 79.3 2 66.79-68.03 79.8 3 69.28-71.77 80.0 2 73.01-74.25 80.1 2 75.50-76.74 80.3 2 77.99-79.23 80.5 1 80.47 81.3 7 81.72-89.18 81.6 2 90.42-91.67 82.4 2 92.91-94.15 82.8 2 95.40-96.64 l- 84.0 1 97.89 84.5 1 99.13 i NOTE: n = 80 x = 77.58 ksi s= 3.10 ksi For 95% confidence interval: 95% occurrence k= 1.964; 71.49 90% occurrence k= 1.559; 72.75 a
+ - - --
E-9 Table E-6 LATERAL EXPANSION VALUES NUMBER OF TIMES CUMULATIVE LE VALUE * (mils) OCCURRING PROBABILITY (%) 23.0 3 0.87-3.36 27.3 2 4.60-5.85 30.0 1 7.09 32.0 6 8.33-14.55 32.3 2 15.80-17.04 32.7 3 18.28-20.77 35.0 1 22.01 35.2 2 23 26-24.50 37.0 3 25.75-28.23 38.3 2 29.48-30.72 38.7 2 31.97-33.21 39.0 2 34.45-35.70 40.0 2 36.94-38.18 41.0 1 39.43 42.7 4 40.67-44.40 43.3 2 45.65-46.89 46.0 2 48.13-49.38 47.0 2 50.62-51.87 48.7 1 53.11 50.0 1 54.35 50.3 8 55.60-64.30 53.3 1 65.55 54.3 5 66.79-71.77 55.7 2 73.01-74.25 56.0 2 15.50-76.74 57.3 2 77.99-79.23 58.0 2 80.47-81.72 61.3 2 82.96-84.20 l 67.0 2 85.45-86.69 l 68.3 3 87.94-90.42 70.7 1 91.67 75.7 1 92.91 77.0 2 94.15-95.40 96.3 3 96.64-99.13
- Average of three tests.
NOTE: High Value = 0.097 inches
- Low Value t 0.023 inches l
Standard Deviation = 0.0162 inches Mean = 0.0482 inches n = 80 At 95% confidence: Occurrence Lower Tolerance Limit
.90 0.0229 inches
.95 0.0164 inches
- L
E-10 e k d I i l-- I "---'I l l @
- - R d -
g 8 m._ - g a s _ 8 g . 3 -
.2 -
e
- n. . -
- s. - -
5
$8 - -
2 m cn - _ m R
~
- ~
g - _ g e h- L l__ _ . l . . . . _1 I I I I 5, en o 20 30 40 50 60 70 80 90 Charpy V-Notch Value, ft. lbs. Figure E-1 Cumulative Probability Distribution Charpy V-Notch Impact Values, Drywell Vent Structure f l[.
O E-11 m o _ g i j 3- -- q .--g 3 , g m, cn d - m g
~ _
_ m cn 8 s g - _ g' _ _ u-a - _ W - _ s - _ _m m e 8 o 5 - - m m - _ m en m - - - cn - _
^
d cn _ o cn m g i i 1.. _. I I I I I 5 46 48 50 52 54 56 58 60 Yield Strength, ksi Figure E-2 Cumulative Probability Distribution Lower Bound Yield Strength, Drywell Vent Structure i 'i
!?
il . i
E-12 8 -?
, __ y .q _, _ . . ; q g o
m cn g
~ - -
g i C - - g a a
, f - -
i - - 5 . . as a
- s
[ 8 - - o cc en - -
~
f t 8 - m - g - 5 en
, h- l 1 .. l . . . _ . . I I I I I E cn o 46 48 50 52 54 56 58 60 }
- i Yield Strength, ksi i
Figure E-3 Cumulative Probability Distribution Upper Bound Yield Strength, Drywell Vent Structure I i
E-13
. Ch cn. -
- o. . .
. , .- -.. m
. - 2 c -
m
-r -
E
~._ -
g
- s _
a w y - 2 . g . 5.
- a. ._ -
g _ - 3, . - E a8 - - a 40 E - -
~
l g _ _ - o > cn - d._ cn g m cn - g 1..... i. . 1 .l _.__.1- I I o, 3 cn o 72 74 76 78 80 82 84 86 0 Tensile Strength, ksi e !i d l l Figure E-4 Cumulative Probability Distribution A Tensile Strength, Drywell Vent Structure
?
\- P 0 L
E-14 m
- e o d d i I i i 1 I I e
- - m.
en o m
~ -
g C - 8 e a .
.o _
a
$u .
- n. _
W _ y _ a . m - o 8 b - co _ m m - g -
* ~
g -
- d i
4 en ~ h- .1 - __ L __ _ l . _..l ___..l._.._.. L __.__I 1 cn o 20 30 40 50 60 70 80 90 i Lateral Expansion, mils Figure E-5 Cumulative Probability Distribution Lateral Expansion, Drywell Vent Structure e w - - -- - - - . .
F-1 { l APPENDIX F STATISTICAL EVALUATION OF CONTAINMENT MATERIAL PROPERTIES o I i L.
F-2 STATISTICAL METHOD _ The methods used for evaluating the containment material properties are , similar to those used in Appendix E for the drywell vent materials. ONE-SIDED HIGHEST LOWEST SAMPLE STANDARD TOLERANgE PROPERTY VALUE VALUE MEAN DEVIATION LIMIT REFERENCE
- Charpy V-Notch 109.33 22.33 63.86 16.55 30.98 Table F-2 (ft. lb) Fig. F-1 Yield Strength 2 60.10 41.0 48.11 3.86 40.44 Table F-3 (ksi) Fig. F-2 Yield Strength 3 60.10 41.0 49.41 4.22 57.80 Table F-4 (ksi)
Tensile Strength 83.90 72.80 77.10 2.46 72.22 Table F-5 (ksi) Fig. F-3 Lateral Expansion 0.0903 0.0157 0.0573 0.0160 0.0255 Table F-6 (inches) Fig. F-4 1 At 95% confidence, 95% occurrence 2 Lower Bound 3 Upper Bound a
}!
t n 1
F-3 Table F-1
~
PNPP CONTAINMENT Cf1TR STATISTICS QUALITY NUMBER OF NUf1BER OF NUMBER OF NUf1BER OF CONTROL TIMES FOUND TIMES ON C0ft- TIMES ON MAPS TIMES ON CMTR NUMBER ON CMTRs TAlflMENT MAPS BUT NOT CMTRs BJT NOT MAPS 75NNI006 0 1 1 75!1NI444 2 2 75NNI445 2 2 75NNI446 1 1 75NNI447 1 2 1 75NNI448 1 1 75NNI449 2 2 75NNI450 1 0 1 75NNI451 1 2 1 75NNI452 1 1 75NNI468 2 2 75NNI469 4 4 75NNI470 1 2 1 75NNI471 2 2 75NNI4i2 1 1 75NNI473 4 4 75NNI474 1 1 75NflI475 1 1 75NNI476 1 1 75NN1477 0 1 1 75NN1478 0 1 1 75NNI479 0 1 1 75NNI480 1 1 75NN1481 2 2 75NNI482 3 3 75NNI483 1 1 75NNI484 1 1 75NNI485 1 1 75NNI486 1 1 75NNI487 1 0 1 75NNI488 3 3 75NNI489 1 1 75NNI490 1 1 75NNI493 1 1 75NNI494 1 1 75NNI495 1 0 1 ! 75NNI502 1 1 l- 75NNI503 1 1 75NNI504 2 1 1 75NNI520 1 1 76NNI006 5 4 1 76NNI007 1 1 ,' 76NNI008 1 1 I m
F-4 Table F-1 PNPP CONTAINMENT CMTR STATISTICS (Continued) - QUALITY NUMBER OF NUf1BER OF NUMBER OF NUMBER OF CONTROL TIf1ES FOUND TIMES ON CON- TIMES ON MAPS- TIMES ON CMTR NUMBER ON CMTRs TAlflMENT t'.APS BUT NOT CMTRs BUT NOT MAPS 76NNID09 1 1 76NNIO10 1 1 76NNIO11 1 l' 76NNIO22 1 1 76NNIO62 1 1 76NNI154 1 1 76NNI185 1 1 76NNI204 2 2 69 71 7 5 f f f 4
,,,- , n. , , _ . . - -.--.-.. -. ., - - . . , . - - -
~ , , . ..- - - -
Table F-2 CHARPY V-NOTCH VALUES CHARPY V-NOTCH
- NUMPER OF CUMULATIVE CHARPY V-NOTCH
- NUMBER OF CUMULATIVE (ft. lb) OCCURRENCES PROBABILITY (%) (ft. lb) OCCURRENCES PROBABILITY (%)
22.33 5 0.98-6.58 68.00 1 51.40 30.67 1 7.98 68.33 1 52.80 33.33 1 9.38 68.67 1 54.20 36.00 1 10.78 69.00 1 55.60 46.00 1 12.18 69.67 1 57.00 47.67 1 13.59 70.00 2 58.40-59.80 51.33 1- 14.99 70.33 1 61.20 52.67 1 16.39 70.67 1 62.61 56.00 1 17.79 71.00 1 64.01 56.33 1 19.19 71.33 1 65.41 57.00 1 20.59 71.67 1 66.81 57.33 1 21.99 72.00 1 68.21 57.67 1 23.39 72.67 1 69.61 58.67 1 24.79 73.33 1 71.01 7' 59.00 1 26.19 74.33 3 72.41-75.21 "' 59.33 1 27.59 74.67 2 76.61-78.01 60.67 2 28.99-30.39 75.00 1 79.41 61.00 2 31.79-33.19 75.67 1 80.81 63.33 1 34.59 76.33 3 82.21-85.01 64.00 1 35.99 77.33 1 86.41 64.33 2 37.39-38.80 77.67 4 87.82-92.02 64.67 1 40.20 78.67 1 93.42 65.00 1 41.60 79.33 1 94.82 65.33 1 43.00 81.00 1 96.22 66.00 2 44.40-45.80 91.00 1 97.62 66.67 2 47.20-48.60 109.33 1 99.02 67.67 1 50.00
- Average of three tes'.s.
NOTE: n = 71 Highest Average Value = 109.83 ft.lbs. x = 63.86 ft.lbs. Lowest Average Value = 22.33 ft.lbs. s = 16.55 ft.lbs. 95% confidence interval (one-sided tolerance limit) 90% occurrence = 37.73 ft.lbs. k = 1.579 95% occurrence = 30.98 ft.lbs. k = 1.987
Table F-3 LOWER BOUND YlELD STRENGTH VALUES NUMBER OF CUMULATIVr YlELD STRENGTH NUMBER OF CUMULATIVE YlELD STRENGTH (ksi) OCCURRENCES _P.ROBABILITY (%) (ksi) OCCURRENCES PROBABILITY (%) 41.0 5 0.98-6.58 48.6 1 55.60 42.1 2 7.98-9.38 48.9 1 57.00 44.0 1 10.78 49.1 2 58.40-59.80 44.4 2 12.18-13.59 49.5 3 61.20-64.01 44.5 1 14.99 49.6 1 65.41 44.9 1 16.39 49.7 - 1 66.81 45.0 1 17.79 49.8 1 68.21 45.1 1 19.19 50.0 1 69.61 45.2 1 20.59 50.2 2 71.01-72.41 45.7 1 21.99 50.4 1 73.81 45.8 2 23.39-24.79 50.6 1 75.21 46.0 1 26.19 50.8 1 76.61 7 46.1 1 27.59 50.9 2 78.01-79.41 o' 46.2 2 28.99-30.39 51.0 1 80.81 46.3 3 31.79-34.59 51.3 2 82.21-83.61 46.5 1 35.99 51.8 3 85.01-87.82 46.7 1 37.39 51.9 1 89.22 46.8 1 38.80 52.6 1 90.62 46.9 1 40.20 52.8 1 92.02 47.0 2 41.60-43.00 52.9 1 93.42 47.2 1 44.40 53.5 1 94.82 47.3 1 45.80 57.5 1 96.22 47.5 1 47.20 58.0 1 97.62 47.6 1 48,.60 60.1 1 99.02 47.8 2 50.00-51.40 48.5 2 52.80-54.20 , NOTE: n = 71 Highest Value = 60.1 ksi x = 48.11 Lowest Value = 41.0 ksi s = 3.86 95% confidence interval (one-sided tolerance limit) 90% occurrence = 42.01 ksi ) Lower Bound 95% occurrence = 40.44 ksi )
r__ _ . . Table F-4 UPPER BOUND YlELD STRENGTH VALUES YlELD STRENGTH NUMBER OF CUMULATIVE YIELD STRENGTH NUMBER OF CUMULATIVE (ksi) OCCURRENCES PROBABILITY (%) (ksi) OCCURRENCES PROBABILITY (%) 42.1 2 0.98-2.38 48.6 1 47.20 44.0 1 3.78 48.9 1 48.60 44.4 1 5.18 49.1 2 50.00-51.40 44.5 1 6.58 49.5 3 52.80-55.60 44.9 1 7.98 49.6 1 57.00 45.0 1 9.38 49.7 1 58.40 45.1 1 10.78 49.8 1 59.80 45.2 1 12.18 50.0 1 61.20 45.7 1 13.59 50.2 2 62.61-64.01 45.8 2 14.99-16.39 50.4 2 65.41-66.81 46.0 1 17.79 50.5 1 68.21 46.1 1 19.19 50.6 1 69.61 ? 46.2 2 20.59-21.99 50.8 1 71.01 46.3 3 23.39-26.19 50.9 2 72.41-73.81 46.5 1 27.59 51.0 1 75.21 46.7 1 28.99 51.3 2 76.61-78.01 46.8 1 30.39 51.8 3 79.41-82.21 46.9 1 31.79 51.9 1 83.61 47.0 2 33.19-34.59 52.6 1 85.01 47.2 1 35.99 52.8 1 86.41 47.3 1 37.39 52.9 1 87.82 47.5 1 38.80 53.5 1 89.22 47.6 1 40.20 57.5 1 90.62 47.8 2 41.60-43.00 58.0 1 92.02 48.5 2 44.40-45.80 60.1 5 93.42-99.02 NOTE: n = 71 Highest Value = 60.1 ksi i = 49.41 Lowest Value = 41.0 ksi s= 4.22 95% confidence level (one-sided tolerance limit) , 90% occurrence = 56.07 ) Upper Bound ' 95% occurrence = 57.80 ) 4
F-8 Table F-5 TENSILE STRENGTH VALUES (ksi) TENSILE STRENGTH NUMBER OF CUMULATIVE (ksi) OCCURRENCES PROBABILITY (%) 72.8 5 0.98-6.58 73.3 1 7.98 73.8 1 9.38 73.9 1 10.78 74.0 1 12.18 74.4 1 13.59 74.6 1 14.99 74.9 1 16.39 75.2 2 17.79-19.19 75.5 1 20.59 75.6 1 21.99 75.8 5 23.39-28.99
., 76.0 2 30.39-31.79 76.1 2 33.19-34.59 +
76.2 2 35.99-37.39 76.3 2 38.80-40.20 76.5 3 41.60-44.40 76.6 2 45.80-47.20 76.7 2 48.60-50.00 76.8 1 51.40 76.9 2 52.80-54.20 77.2 3 55.60-58.40 77.3 1 59.80 77.5 2 61.20-62.61 77.7 2 64.01-65.41 77.9 1 66.81 78.0 2 68.21-69.61 78.1 2 71.01-72.41 78.2 1 73.81 78.4 1 75.21 l 78.8 2 76.61-78.01 79.4 2 79.41-80.81 79.6 1 82.21 79.7 1 83.61 80.0 1 85.01 ( 80.2 2 86.41-87.82 80.5 1 89.22 80.9 1 90.62 L 81.3 1 92.02 81.5 2 93.42-94.82 81.6 1 96.22 l- 81.8 1 97.62 l 83.9 1 99.02 I NOTE: n = 71 Highest Value = 83.9 ksi l i = 77.10 Lowest Value = 72.8 s = 2.46 ksi 95% confidence level (one-sided tolerance limit) L 90% occurrence = 73.22 ksi 95% occurrence = 72.22 ksi
g -- - . -. ~ .. - - . . .. , Table F-6 LATERAL EXPANSION AVERAGE VALUE NUMBER OF CUMULATIVE AVERAGE VALUE NUMBER OF CUMULATIVE (inches) OCCURRENCES PROBABILITY (%) (inches) OCCURRENCES PROBABILITY (%) 0.0157 5 0.98-6.58 0.0607 2 47.20-48.60 0.0267 1 7.98 0.0610 2 50.00-51.40 0.0307 1 9.38 0.0613 1 52.80 0.0337 1 10.78 0.0617 1 54.20 0.0417 1 12.18 0.0620 1 55.60 0.0420 1 13.59 0.0623 2 57.00-58.40 0.0437 1 14.99 0.0627 1 59.80 0.0463 1 16.39 0.0633 3 61.20-64.01 0.0490 1 17.79 0.0640 3 65.41-68.21 0.0493 1 19.19 0.0653 2 69.61-71.01 0.0500 1 20.59 0.0657 1 72.41 7' O.0507 1 21.99 0.0660 1 73.81 O.0510 ' 1 23.39 0.0673 1 75.21 0.0517 1 24.79 0.0677 1 76.61 0.0523 1 26.19 0.0680 2 78.01-79.41 0.0527 1 27.59 0.0683 1 80.81 0.0530 1 28.99 0.0687 1 82.21 0.0533 1 30.39 0.0700 1 83.61 0.0537 2 31.79-33.19 0.0710 1 85.01 0.0567 1 34.59 0.0737 1 86.41 0.0577 2 35.99-37.39 0.0743 2 87.82-89.22 0.0580 2 38.80-40.20 0.0747 2 90.62-92.02 0.0600 3 41.60-44.40 0.0750 3 93.42-96.22 0.0603 1 45.80 ' 0.0833 1 97.62 O.0903 1 99,02 NOTE: n = 71 Highest Value = 0.0903 inches x = 0.0573 Lowest Value = 0.0157 inches s = 0.0160 inches 95% confidence level (one-sided tolerance limit) 95% occurrence = 0.0320 inches 95% occurrence = 0.0255 inches
F-10 e o * . g , __, , ,__..., .g. ,___.. g
- - m o
m
~
m g
~ -
g 5 - 8 va g . -
.o . _ .3
- O -
k.
- a. _ _
W - 3 a g - . o
$8 - -
E m m - __ m
* ~
g - - d m m - g J . __ ... l _ i L_._.I I I I o e o ! 30 40 50 60 70 80 90 100 Charpy V-Notch Value, ft. lbs. I i '. Figure F-1 Cumulative Probability Distribution Charpy V-Notch, Containment f t$
F-11 m o e. d l l l-*-~~l -~~'-l- l 1
-g g -
~
g g
~ -
g C o - 8 a b 2 . 2 . E
- o. . .
g . s, - - f .8 - o e m - -
~
g - - - g- -
- g i
t m 9- 1 l . J . . . ... . L .l_ l i I E m o 43 45 47 49 51 53 55 57 l 1 Yield Strength, ksi 't !! Figure F-2 Cumulative Probability Distribution Yield Strength, Containment i: 1 6 f i t j lW. i .
- - - , - - - - . - - - - - . - y---n - . - , -n. , , - - -
O F-12 o E. d l I I i ' -'I I I $ 8 d -
- g
~ -
E C - 8 u C g _ g . u
- n. _
g . 3 . . 8 - - ce m - m j E - - - , m - - l s '-
- d m
l g- 1 I _- .L. _ . 1 _I I I 1 5 73.8 75.8 77.8 79.8 Tensile Strength, ksi i
; Figure F-3 Cumulative Probability Distribution Tensile Strength, Containment I
il i .f 4
F-13 o ?. d I i 7 ~- ' i '-'~l i 1 's
~ ~
d -
?
g
~ _
g o -
- e m
e l ; - a 4 _ _ g - u -
- n. -
y .
'a - .
E a 8 - - e i ao m - - ~ g - _ - m. m ,_ _- o . m _ cn
~
2- 1 l_ ___ l . .. . L_. _. I I t- I m o l 20 30 40 50 60 70 80 90 Lateral Expansion, mils Figure F-4 Cumulative Probability Distribution Lateral Expansion, Containment l 1\,
J 4 mA s,.m--4------An K --m - --m. n --se- m- J-na - --m- - - - - A.-J mM A--me.-rg - - . - ._m .an----_.m-----a 4 sA4A,. ,s,sa.1 .-4s-- _---- - r-4 w-1 1 i e i I APPENDIX III f 6 I l 6 I I } l. w,----,-r--y,--- --- -
i , APPENDIX III THE SIGNIFICANCE OF SENSITIZED STAINLESS STEEL MATERIAL ON THE STRENGTH OF THE DRYWELL VENT AND CONTAINMENT STRUCTURES I IN THE PERRY NUCLEAR POWER PLANT I m Geert/Commonmoelth
1.0 Introduction The lower 13'-7" of the drywell vent structures, lower 13'-6" of the containment vessel, and floor plate are clad with a 10% layer of SA-240 Type 304 stainless steel. This report will address the significance of the fissured and sensitized stainless steel cladding on the strength of these structures. The design calculations for the containment vessels and floor plates are based on the total plate thickness including cladding while the calculations for the drywell vent structures are not. A , brief discussion of the concern and reso*ution for each structure is presented below. 2.0 Drvwell Vent St ructure The fracture and fatigue evaluations of Appendix D address the potential for defect growth by a f atigue mechanism and the potential for subsequent failure by fracture. Appendix IV concludes that crack growth by fatigue is small over the design plant lifetime and that fracture will not occur. Based on this conclusion and that the design calculations of the vent structure only used the carbon steel thickness, no weld repair of fissures on the drywell vent structure will be required. However, because of the potential for crevice corrosion as discussed in Appendix IV all welds and the heat affected zones in the drywell vent structure. exposed to the suppression pool environment will be coated as discussed in Section 3 of the report. The remedial steps taken will eliminate any potential adverse effects on the strength of the drywell vent structures. GeertICommoneesRh 1I1-1
3.0 Basemat Floor Plate It has been shown by testing (Reference 7) that the floor plate is sensitized, but not to the extent of samples from the containment or drywell vent structure. To date PT inspections of floor plate welds have shown no signs of fissuring. n11 full penetration floor plate welds are 100% PT inspected. In addition all floor plate welds are covered by leak chase channels / angles and pressurized to 25 psig. Full penetration welds will not be exposed to the environment of the suppression pool. Mechanical tests were made on the sensitized material (Reference 7). Mechanical tests were made of carbon steel only, full thickness composite, and stainless steel cladding only. The ultimate tensile strengths of all specimens exceeded the minimum specified ultimate tensile strengths. If, during completion of final inspections of the floor plate, any fissures are foand they will be weld repaired. No remedial steps on the floor plate are required. The strength performance of the floor plate is unaffected by the conditions defined. 4.0 containment Vessel PT inspections of the following regions have disclosed varying degrees of fissuring:
- 1) immediately above the fillet weld connecting the doubler '
plate to the containment at the base of containment (Attachment 2-1) I
- 2) adjacent to fillet welds attaching several attachment plates to the containment vessel and less frequently Gdbert /Commonweeth III-2
~
t
- 3) along horizontal and vertical weld seams of the shell plates Although the cladding was used in the original design calculations of the containment vessel, with the addition of the concrete in the lower 23'-6" of the annulus between the shield building and containment vessel this cladding is not needed for strength. Design calculations for the final containment stress report will demonstrate the adequcy of the containment without the cladding for the filled annulus condition.
Therefore, similar to the drywell vent structure no weld repair of fissures in the containment will be required. However, all weld seams and the heat affected zones in the containment vessel will be coated as discussed in Section 3 of the report because of the potential for crevice corrosion. The remedial steps taken will eliminate any potential adverse effects on the strength of the containment vessels. i o 1 ( l i Gdbert/Commoneesth III-3
_ __ a - .n -. . - . - -. m,. a e APPENDIX IV e - - . . _., _._ ~ ~ ' w-w- wv. ,_ _ _ _
l AES 8106262-:2 Final Report APMG engineering iervices,Inc eso,meeR,noCOssuL1Am1S 795 SAN ANTONIO ROAD . PALO ALTO . CAllFORNIA 94303 (415)858 2863 l APPENDIX IV EVALUATI0tl 0F POTENTIAL CORROS10ti - PROBLEfts Ill TiiE SUPPRESS 10tl P0OL AT PERRY flVCLEAR POWER PLAtlT Prepared by Terry W. Rettig Aptech Engineering Services, Inc. 795 San Antonio Road Palo Alto, California 94303 Prepared for Gilbert Associates, Inc. Post Office Box 1498 Reading, Pennsylvania 19603 Attn: Paul B. Gudikunst, Project fianager f4 arch 1982 i 6 Servicesin Mechanicaland Metallurgical Engineering, Welding, Corrosion Fracture Mechanics, Stress Analysis
VERIFICATION RECORD SHEET ,
Title:
Evaluation of Potential Corrosion Problems in the Suppression Pool at Perry Nuclear Power Plant (AES 8106262-2) Originated by: ,- 3-s.ca Terry W. Rettig
., ?
Approved by: L-v e'k ; .g 4 2. Geoffrey R Egan
. l Verified by: l[ ; M -
RogerH.Rickman I l' Quality Assurance Review by:
.w
/,teil [ [,hf.[
_s /,/.e: Russell C. Cipolla Quality Assurance Approval: Edwin R. Pejack l k.
TABLE OF CONTENTS . Section Title Page 1.0
SUMMARY
1
2.0 INTRODUCTION
AND OBJECTIVE 3 2.1 Crevice Corrosion 3 2.2 Stress Corrosion Cracking of Sensitized Stainless 4 Steel 2.3 General Corrosion of Low Alloy and Carbon Steels 5' 2.4 Method of Approach and Scope 5
3.0 BACKGROUND
7 3.1 Environment of the Suppression Pool 7 3.2 Residual Stresses in Cladding 8 3.3 Material and Geometry 8 4.0 POTENTIAL CORROSION MECHANISf1S 10 4.1 Corrosion of Carbon and Low Alloy Steel 10 4.1.1 General Corrosion 10 4.1.2 Crevice Corrosion 13 4.1.3 Corrosion of Carbon and Low Alloy Steel - Summary 14 4.2 Stress Corrosion Cracking of Sensitized Stainless 14 Steel 4.2.1 Stress Corrosion Cracking of Type 304 Stainless 24 Steel - Summary 5.0 CORRECTIVE ACTIONS 27 5.1 Shot Peening 27 5.2 Coatings and Cathodic Protection 28 5.3 pH Control 29 6.0 DISCUSSION AtlD
SUMMARY
31 REFERENCES 35 i 4 l.
1 1.0
SUMMARY
The suppression pool at Perry fluclear I is constructed of A516 Gr 70 steel plates that have been clad with 304 stainless steel. During dye penetrant inspection of the welds, several cracks were discovered. A review of the f abrication procedures for the plate deternined that the stainless steel was sensitized and contained significant residual tensile stresses. A literature review was conducted to determine if the operating environment of the suppression pool and the condition of the material would result in a stress corrosion cracking problem during the design life of the plant. The results of the review showed that at saturation levels of oxygen and for the expected temperature a cracking problem does not exist. If other contaminants, such as borates, thiosulfates, or halogens, increased in concentration, a small possibility of cracking exists, although it would take years to initiate. A literature search was also conducted to determine if corrosion of the underlying A516 Gr 70 steel could be a problem should a crack that penetrated the stainless steel not be detected during inspection. The results indicate that the crevice produced by the crack could result in significant corrosion rates. However, it is expected that the cracks would exist only at regions rear the welds because of problems associated with welding. ' The results of other work showed that fracture of the suppression pool will not occur and that the only possible mode of failure would be a leak (1). However, crack growth by fatigue to a leak condition will not occur in the design life of the plant. From the corrosion point of view, the action to be considered to ensure the integrity of the plant is whether to take preventive measures now or corrective actions later, if required. This decision is a matter of economics provided an inspection and nonitoring schene is implemented. Several potential repair schemes are addressed including coatings and metal spraying, shot peening, and pH control. Several nonitoring schemes are .t 6 -l L
._ . _ _ . . _ , . . _ _ _ . _.. _ ._ _ ._ . _= _ _ . _. . _ _ _. _ . _.
i l' 2 J suggested. It is concluded that the most reliable approach to preventing corrosion of the base material is to use a flame sprayed aluminized coating adjacent to weld regions where fissuring has occurred. To provide, further coating integrity, a sealing coat of epoxy is also proposed. i I i
- t. .
i I l .s J i i i I
. . , , , , , - _ _ _ _ . _ . - - - , - . - . . _ _ _ . . .-- __- , _ . . _ . ~ _ . _ . . - . . . . _ . . _ _ _ _
3
2.0 INTRODUCTION
AND OBJECTIVE APTECH Engineering Services was requested to perform a f racture mechanics crack growth analysis of the materials in the suppression pool to determine the impact of weld heat affected zone (HAZ) fissuring on plant integrity (1). The information obtained was used to determine the nargin against failure for the plant design life. Additional factors that need to be considered are: (1)the effect of crevice corrosion of the A516 steel on the integrity of the structure, and (2) the effect of the environment oil the integrity of the stainless steel cladding from a corrosion or stress corrosion cracking view. If the corrosion rate and its effect on crack geonetry, such as increased rate of penetration or blistering, can be determined, it can be factored into the crack growth evaluation. This would provide additional insight into whether the plant can continue to operate over the design life without repair of the containment. The general objective for this work was as follows: To evaluate the potential for corrosion in the suppression pool over the design life of the plant. This neans that the specific nechanisms to be covered will include those discussed below. 2.1 Crevice Corrosion Crevice corrosion is characterized by a geonetrical configuration in which the cathode reactant (usually dissolved oxygen) can readily access the metal surface outside the crevice by convection and diffusion. Access to the stagnant area within the crevice is much more difficult and will only be achieved by diffusion. This results in a difference in concentration of the cathode reactant at the two surfaces. Although differential aeration is important in the mechanism, it is really more complex, partly as a result of the acid formation within the crevice. The cathodic reduction of oxygen on the f reely exposed surface will result in anodic dissolution of the metal within the crevice. The excess of positive charges (metal ions) formed will result in the nigration of chloride and hydroxide ions f rom the bulk solution to maintain charge balance. Hydrolysis then occurs with the formation of an acid solution in the crevice. Frequently, there is a long incubation time for attack to initiate, but once initiated, the attack is usually autocatalytic, proceeding at W
4 a rate that increases with time. The incubation time to commence crevice corrosion and rate of penetration are very complicated for the situation existing in the suppression pool. Some of the variables that influence crevice
~
attack are listed below: e Depth and width of crevice or fissure e Environment chemistry and local effects e Environnent in the crack and its change with time e Ratio of external area to crevice area e Change in passivity of cracked cladding e, Effect of nickel on alloy modification e Time e Temperature e Cycling of crack opening displacement e Diffusion of aggressive ions into the crevice e Corrosion product accumulation in the crack 2.2 Stress Corrosion Cracking of Sensitized Stainless Steel When exposed to an aggressive environment, the grain boundary of a metal is usually attacked slightly more rapidly than the grain faces. This effect is usually of little or no consequence in most applications. Ilowever, under certain conditions, grain interfaces are very reactive and intergranular attack results. It is usually caused by impurities at the grain boundaries or by enrichment or depletion of one of the alloying elenents. Numerous failures of Type 304 stainless steel have nccurred because of intergranular stress corrosion cracking (IGSCC). Failures have been experienced in environments where the alloy should exhibit excellent corrosion resistance. The reason most often cited for these failures is based on depletion of chromium O _4.
5 in the grain boundary area. If the chromium in a steel is effectively lowered below about 12%, the relatively poor corrosion resistance of carbon steel is approached. In the temperature range of approximately 900* to 1500*F, Cr C 23 6 precipitates out of solid solution if the carbon content is above about 0.02%. The chromium is removed from solid solution, and the result is metal with lowered chromium content in the area adjacent to the grain boundaries. This narrow zone is corroded because it does not possess sufficient corrosion resistance for many applications. 2.3 General Corrosion of Low Alloy and Carbon Steels The general corrosion of metals in neutral waters is usually a result of the action of oxygen. The source of oxygen is usually the oxygen dissolved from the air. The most familiar corrosion of this type is perhaps the rusting of iron when exposed to water: , 4Fe 6H O + 30 2 2 4Fe(0H)3 The above equation shows the reaction of iron with water and oxygen to form ferric hydroxide. 2.4 Method of Approach and Scope The importance of each of these mechanisms depends on a range of parameters including stress level, environment, ter.perature, material condition, and the presence of any pre-existing cracks. Each of these aspects is outlined in detail in the sections that follow. . To evaluate the situation with respect to corrosion, a literature search was initiated to obtain data to determine whether or not a corrosion problem exists l with the present materials / environment combination. This report presents the results of the literature review and addresses the possibility of corrosion attack during the life of the plant. The report also outlines the important variables that affect each potential corrosion mechanism. It was concluded that for all three corrosion mechanisms and the known
. conditions of the suppression pool, oxygen was the most important variable.
a m O
6 Recorrnendations are also made to ensure the integrity of the suppression pool and to monitor for potential corrosion cracking problems in the event they DCCur. f e I l l i h lL __ _ -. _ D
7
3.0 BACKGROUND
The suppression pool at Perry fluclear is constructed of A516 Grade 70 steel clad with Type 304 stainless steel with a nickel plate interface to promote good bonding. After cladding, the composite was annealed at 2000 F to 2050*F, air cooled, then normalized at 1625*F to 1675*F and air cooled. The clad plates are then joined in the field by welding. During inspection by liquid penetrant, several cracks were detected in the vertical weld seans. The potential for corrosion may be considered as the reaction between a susceptible material and an environment. Stress corrosion cracking cnntains these essential parameters in addition to a suitable stress. Each of these parameters as they exist in the suppression pool is described below. 3.1 Environnent of the Suppression Pool The water characteristics in the suppression pool are as follows (2): Conductivity 5 unho/cn at 20 C (68 F) Cl' 1 0.5 ppm Suspended solids 11 ppn . Heavy metals f 0.1 ppa pH 0 25*C (77"F) 5.3 to 7.5 Although there are no controls on 02 c ntent, it is expected that the 0 c2 ntent will be approximately 8 ppa which represents the saturation level of water at a water / air interface at 100*F. Much of the analysis that follows, therefore, is based on the assumption that oxygen levels will be nuch Inwer than 8 ppm. The temperature in the suppression pool during normal operation is expected to be about 90*F. Short-tern transients above this level during postulated accidents will not be significant. l O t l -
8 3.2 Residual Stresses in Claddino Corrosion cracking mechanisms will be influenced by stress levels. The total stresses in the clad will consist of the following: e Service stresses - to include dead weight, pressure, seismic, etc. e Working stresses - from grinding, etc. e Residual stresses - f rom welding, cladding, and heat treatment It has been determined that the service and working stresses are low compared to the residual stresses from cladding and welding (1). Kupka (3) found that as a result of different chenical composition, mechanical properties, and coefficients of thermal expansion between stainless steel weld cladding and a steel base metal, sharp stress peaks will occur in very small a reas. Residual stress will exist in both the longitudinal and transverse directions. After annealing, the longitudinal stresses in one specinen were measured to be 54 ksi and the transverse stresses at 29 ksi for the cladding that'was applied using manual arc welding. Kupka referred to a paper (4) where it was determined that the state of stress was found to be biaxial both prior to and after thermal treatment and in both cases reached the yield point of the cladding. Minor variations in the residual stresses in the suppression pool lining may occur because of differences in cladding technique and the nickel layer, but the above shows that significant residual tensile stresses will be present. It is clear, therefore, that the residual stresses in the cladding will be nuch higher than service-induced stresses. 3.3 Material and Geometry The stainless steel clad plate used in the lower portion of the suppression pool is a composite plate consisting of several materials and was developed to provide the benefits of expensive naterials at lower costs. In this application, 304 stainless steel and A516 structural steel have been mill rolled under heat and pressure until they are integrally bonded over the interface. To pronote good bonding, an electrolytic coating nf nickel is applied between the i .L
9 alloys. During the hnnding process, slight diffusion of the nickel occurs in the base metals, thus locally altering the composition of the alloys adjacent to the interface. As a result of subsequent heat treatments, the stainless steel clad has become sensitized (5), and contains a continuous network 6f intergranular carbides. The welding process has resulted in HAZs that are more sensitized than the parent metal. The crevices or cracks produced as a result of welding are very tight and a have a width to depth ratio of the order of 1:100. Material in similar sensitized condition has been known to crack in BWR
- systems, as will be discussed later.
1 I l l k O
s 10 4.0 POTErlTIAL CORROS10f4 fECHANISHS The approach that was taken to evaluate whether there is a potential crevice corrosion or stress corrosion cracking problem was to review the IJterature and to review situations that are similar or slightly nore severe than those in the suppression pool. 4.1 Corrosion of Carbon and Low Alloy Steel A review of the literature was conducted to obtain information about the general corrosion rate of carbon steel as well as the crevice corrosion problems that may exist under the cladding. The structural support beams rising out of the suppression pool are fabricated from carbon steel and are unprotected. The decision to protect these members is dependent on the corrosion rate that could be expected, in addition, there was concern that corrosion products could accumulate or settle in the water and cause additional corrosion problems in the cladding. As a result of these concerns, the decision was taken to protect the columns with a Type 304L wrap. 4.1.1 General Corrosion Dissolved oxygen is usually the controlling factor in the corrosion of steel by water. Figure 1 shows the relationships between oxygen content, temoerature, and corrosion rate for a low carbon steel in a tap water (6). The corrosion rate goes up dranatically with an increase in hnth oxygen and temperature. At 90*F with 5 ppn oxygen, the corrosion rate is 80 nils per year (mpy) and increases to 200 apy at 120 F. These rates were the highest that were obtained fron the literature survey. These rates are higher than the corrosion rates reported in the Metals Handbook (7) for oxygenated water (Fig. 2). At 104*F and 5 ppn oxygen, the corrosion rate is approximately 20 mpy over the pH range of four to eight. The corrosion rate also doubled when the temperature was increased from 72*F to 104*F. Uhlig (8) reported that in aerated distilled water, the corrosion rate was 8 mpy at 75 F and, in fact, can he higher than in seawater. Likewise, the corrosion rate in water with a pH less than seven i. i.
11 i h i i Corrosion rote (mpy) 300 4 izo r ! 250 - 1 200-so r l': / 50 - O ' ' I 2 3 4 56 7 8 9 10 l p pm oxygen 4 Figure 1 - Corrosion Rate for a Low Carbon Steel In a Tap Water (6). i a a f l 6
.. ..r..,._....,_-_. . , _. . , _. _ , . . _ ,,._,,m _m, , _ , _ . . _ __ _ , , . _ _ _ _ , . , . _ , _ _ . .g___
12
}# 1 I
I j. CCg HC'
! 150 ,
4l l 1
! 198 4$
i ~, A i la [ ~ m 2g 4 i h
* *, , a s a s s' '
la 3" 2 r, ' ;- , E ss i i g % 103 . _ _- 4 f
} 53 Aj.....
,'w i l{
17 g i 0 e ' 72 C g i I ' i it is . . r i.1 Corrosion rates are normalized to a solution containing 1 mL 02 per litre of water. To estinate corrosion rates at other concentrations, multiply values derived from this graph by the oxygen concentration in mL/L. Figure 2 - Corrosion of Steel (a) in a Saturated Solution of Carbon Dioxide in Water and in Dilute Hydro-Chloric Acid and (b) in Aerated Water Q). L j
13 can be greater than 8 mpy, while in high pH water, the rate is less than 2 mpy (9). This seems to be in general agreenent with Uhlig (8), who reported that the corrosion rate of mild steel imersed in still seawater saturated with air ranges from 2 npy to 6 npy. However, he pointed out that pitting Inay produce a penetration rate ten tines the corrosion rate. He also reported that the corrosion rate in distilled water with only oxygen present increases about two-fold between 60*C (140*F) and 90 C (194*F). Differences in oxygen concentration will accelerate corrosion and pitting attack. This type of attack is found at the waterline of surfaces exposed partly to air and partly to water. Matsudaira (_10) found that in pure water with dissolved oxygen, a passive film is forned, but in stagnant water, small pits are formed and spread gradually showing general corrosion. Freshwater tests conducted at Gatun Lake in the Panama Canal Zone (11_) showed corrosion rates ranging from 8 npy af ter the first-year exposure down to 3 npy after eicht years of exposure. The water was of good quality with an average total dissolved solids of 113 ppm, chlorides 7 ppn, and a pH of 7.5. The average temperature was approximately 81*F and the average conductivity was 110 unhos. 4.1.2 Crevice Corrosion The complexity of the crevice that would exist in the suppression pool is perhaps best illustrated using the materials of its construction. The surface of the suppression pool is Type 304 that is in a passive state. In the crevice, which would soon be depleted of oxygen, the stainless steel nay lose its passivity and become active. Once active, it would have a potential approaching that of the structural steel. As a result, it would begin to corrode because it would become anodic with respect to the surface. This would tend to open the crevice and could result in increased diffusion of aggressive ions. At the sama time, the tip of crevice is exposed to a nickel layer under which exists a structural steel, the surface of which has been alloyed. As a result, it may be more appropriate to consider the tip of the crevice located in a shallow nickel alloy steel with a decreasing nickel content. It is expected that the tip exposed to alloy steel would be anodic to the stainless steel and would, thus, exacerbate the situation. Another possibility is that the corrosion, if it occurred, could run into a barrier at the nickel stainless steel interface and j, propagate horizontally. I L
14 Forgeson (11) reported that millscale on unalloyed low carbon steel, which is continuously inmersed in seawater, will intensify localized corrosion. He attributed this to the cathodic behavior of the scale in relation to the underlying steel. After eight years of exposure in Pacific seawater, three perforations occurred in 0.263-inch thick steel; a penetration rate, which was three times greater than the penetration rate that occurred in Steel that had been pickled prior to exposure. However, in similar tests conducted in Gatum Lake, the millscaled steel performed as well as the pickled steel and pitted to only 50 nils af ter eight years. The nillscale on the carbon steel is somewhat analogous to the cladding, with the clad functioning as the more cathodic naterial. 4.1.3 Corrosion of Carbon and Low Alloy Steel - Surnary These data suggest that the corrosion rate in carbon or low alloy steel can vary from a few opy to well over 100 mpy; the rate being extremely sensitive to the environnent. It was, therefore, prudent to protect the carbon or low alloy steel from the environment with the stainless steel cladding. If no corrosion allowance is provided on unprotected carbon and low alloy steels, the corrosion of the steel could affect the integrity of the structure. In addition, the corrosion products formed could affect the quality of the water as well as deposit on the stainless steel components producing additional crevices and sites for pitting. The rate of crevice attack due to the breach in the cladding cannot be quantified and, therefore, must be assuaed to be at least as rapid as plain exposed steel or twice the appropriate attack rate produced by the presence nf millscale, because of the higher temperature. 4.2 Stress Corrosion Cracking of Sensitized Stainless Steel Weld related cracking in the stainless steel clad suppression pool liner has
, raised concerns about the prospects for long term stress corrosion' cracking (SCC) throughout the suppression pool. This is postulated because the high degree of material sensitization conbined with a stagnant aqueous environment and significant residual tensile stresses correlates with previous BWR pipe cracking experiences.
i Le 9
15 At present, no means exist by which SCC thresholds for the constant effects of these parameters can be predicted; however, field experience in BWR systems and laboratory studies provide useful comparisons. Studies of BUR cracking experience have identified certain correlations in material and environmental factors. Some of the real or potential problems influencing IGSCC of BWR austenitic stainless steel piping are shown in Tables 1 and 2 (H). The cyclic stresses are of minor importance, but the residual stresses from fabrication are believed to have a major impact. Sensitization, b' oth furnace and weld induced as well as oxygen, are also believed to have a cajor effect on IGSCC. However, just how important most of these factors are is not really known. Some relevant features associated with SCC incidents in BWR piping are as follows: o Higher carbon contents favor SCC. o Various wrought product forms of 304 stainless steel have experienced cracking (forgings, plate, pipe), o Duplex (5% ferrite) stainless steels (castings, weld metal, etc.) appear to be immune to SCC. o Welding procedures (especially heat input) are key carameters and vary widely; failure probabilities are higher in field welds as opposed to shop welds, o Correlations with combined thermal and service stresses indicate higher failure probabilities with increasing stress and/or restraint. o Impurities or contaminants both in the form of non-metallic inclusions in the steel and aqueous contaminants have a significant role in SCC. o Surface treatments, such as machining o.r grinding, can accelerate crack nucleation by both mechanical and metallurgical effects. The surface degradation caused by chemical processing, such as pickling, can accelerate occluded cell incubation. Pickling of previously sensitized plate greatly enhances this. At Three Mile Island I IGSCC initiated in Type 304 spent fuel piping at the HAZs of circumferential butt welds (H). Seven leaks were observed as well as 35 L
p : _. - .u - Table 1 REAL OR POTENTIAL PROBLEMS INFLUENCING INTERGRANULAR STRESS CORR 0SSION CRACKING (IGSCC) 0F BWR AUSTENITIC STAINLESS STEEL PIPING (PRIllARILY TYPE 304) (12) PROBABLE INPACT ESTIMATE OF STATUS PARTIALLY PARAMETER OR FACTOR MAJOR filNOR KNOWN KN0 Hit UNKNOWN RESOLVED UNRESOLVED Stress (Strain) X X X Residual-Weld X X X Residual-Machining, Etc. X X X Thermal X X X Cyclic X X X Combined X- X X Low Cycle Fatigue X X Sensitization Furnace X X X Held X X X Effect of Alloy Com- X X X position Effects C, N X X X Environment , Oxygen X X X Crevice Effects ? X X Surface Contaminants X X X Crud Holdup of Ions ? X X \ l
17 Table 2 , PARAMETERS CONTRIBUTING TO STRESS CORR 0SION CRACKING IN NUCLEAR POWER SYSTEMS AS IDENTIFIED FROM AVAILABLE CASE HISTORIES (12) 18/8 AUSTENITIC 304, 316 MATERIAL Ferrite Percent S Carbon Effect - FABRICATION Cold Work S Cleanliness S Heat Treatment Solution - Sensitizing S Welding Heat Input S Pickling S DESIGN Stress S (Weld-Residual Thermally Induced External Loads) Crevices (Concentrator) S Galvanic Coupling - Cyclic Loading S OPERATIONAL Temperature H Local Heat Flux S Neutron Irradiation M pH M+S Residuals Oxygen S Chlorides S Fluorides C KEY: S = Signif; cant M = Minor C = Controversial i mese
- + ,
18 additional indications. Ambient temperature water with boric acid (100*F max), stagnant conditions, and traces of chloride were thought to be responsible for the failure. It was believed that the degree of sensitization was_the rate controlling step with respect to crack initiation. Cracking has been reported in barated water lines at operating BWRs, but the cracking could not be duplicated in laboratory tests using horic acid at 100 F. However, cracking could be induced by the addition of a few ppm of sodium thiosulfate, a contaminant that could enter the boric acid line from a leak in the core spray system (14_). 1 Crevice corrosion tests conducted with a Russian equivalent of Type 304 under rubber in a 0.5N Nacl showed that the average rate of attack was 1.3 run/ year (55 mpy) (.15). Bryant (_16,) reported an occurrence of intergranular attack on furnace sensitized wrought 304 by fluoride at 70*F. He concluded that fluxes and their by-products from shielded metal arc and submerged arc welds are the prime source of fluoride contaminants. Oyxgen greater than 30 ppb was necessary for attack initiation. He also reported that the combination of crevices and contaminated oxides overlaying a sensitized substrate pose a distinct prot'len as the stagnation
' prevents renoval of the fluoride ion and release to the environment.
l
- Bush (1_2_)
2 noted that care should be taken in the pickling of sensitized y stainless steels with HNO3 - HF solutions as grain boundary attack to depths greater than 0.015 inch is known to occur. This attack will result in crevices and concentration of inpurities. He also reported instances of SCC in sensitized stainless steel at very low concentration of chloride ion concentration which can initiate SCC, but this concentration is influenced by 1 stress levels and oxygen concentration. In his review of SCC in the nuclear industry, Bush concluded that proof that IGSCC can occur during shutdown is not established and remains an open issue. The review of the frequencies of SCC L indicates a greater frequency occurs in BWRs af ter five years of operation and failures occur periodically before startup. Specific causes were cited in the report, and a select few am described in the following: 4 L
19 e Arkansas Nuclear 1 developed a leak in a shop-fabricated stainless steel spool piece during hydrotesting. The crack was located 0.5 inch from a shop weld and was attributed to IGSCC. e The Shippingport Atomic Power Station (PWR) experienced SCC of a sensitized Type 304 stainless steel canopy seal. Cracking was attributed to oxygen entrapped under the seal during filling operations. o Nunerous cracks have been reported for components operating at higher temperatures. Figure 3 shows the frequency of SCC incidents as a function of the years of operation for nuclear power plants in the United States. In a review of SCC in sensitized Type 304 stainless steel in oxygenated pure water at elevated temperatures, Szklarska-Snialowska and Cragnolino (17) showed that at 288*C (550*F), the time to failure decreased with increasing oxygen concentration and at 48 ksi the failure time was approximately 40 hours for ogygen in the range of 2 - 10 ppn. For the sane oxygen concentration and 35 , ksi, the f ailure time was more than doubled. Andresen reported an incident of SCC using strain rate tests of sensitized stainless steel in pure water at room temperature with an oxygen content of 2 ppn but emphasized the severity of the i test (18). Large defects on the metal surface, like crevices and notches, have very pronounced effects of time to failure. Vermilyea also stated that crevices accelerate cracking (19). Battelle (20, 21) conducted a study to determine the influence of oxygen, hydrogen peroxide, and tenperature on the susceptibility of furnace sensitized 304 stainless steel to SCC. The results obtained in constant extension rate screening studies (4 x 10'0 sec - ) showed that a critical temperature of at t least 250'F was necessary for cracking in oxygenated (8 ppm) high purity water. 1 . This is consistent with corrosion fatigue tests, in which the load was cycled between 0.1 and 1.5 times yield in 185 F water with 8 ppm oxygen, that failed to produce any cracking after 167 hours (22). h L e
20 1.2 OTHER
. 17J BWR -
1.D f E] PWR ~! fi5 'l DW - ' g
-a at -
er ' 9
$$ 'i 25 -
g; a6 '
%2 my v i} -
- )
cd - 0 I , o m, r-gs
.. b l * , .
59 F1 . *L !h. *
.}
il i~ i,,1 29 nr -
. '.- 11 il
.V[Lii tr I1< j<
, ! ).t:
- m y[ .
il :3 .! g _, I' Y!' n 1
. h;f lf-e e e s ,
e i [ E,* 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 p CARS Of Or[RAfl0N
- Figure 3 - Stress Corrosion Cracking Incidents - Limited to United i
States Data (An incident is defined as the discovery of one or more cases of stress corrosion cracking at a given time in the same alloy. Fuel cladding failures are not
- i. included)(12). -
(. l-I f 4 e i s M-e
21 l l The test was terminated and the coupons pulled to failure. No evidence of SCC was observed on the fracture surface. Constant load tests in these temperature ranges were suspended because cracking could not he induced in a reasonable
~
time, suggesting that the Constant Extension Rate Test (CERT) may be conservative for statically loaded components. The accelerating effect of the CERT was also demonstrated by Kaikawa, et al. (H). In oxygenated water at 288 C (550 F), a CEl!T coupon strained at 4.2 x
~1 10~ 0 sec with an applied stress of 0 - 38 ksi failed in ten hours, while a constant load specimen stressed at 44 ksi took 400 hours to fail. The fracture surface of the CERT coupon was 1007, intergranular and that of the constant load coupon only 507,intergranular.
Andresen and Ford (2_4) _4 reported that crack propagation was measured at tenperatures as low as 50"C (122 F) with some cracking occurring in laboratory tests at roon temperature, llowever, they pointed out that field cracking has not been observed in BUR application at low tenperature (less than 212 F). They pointed out the complex condition that exists during startup of a BWR when the water goes f rom a low temperature and high 20 to a high temperature and low 0
- 2 This is a competing effect; the susceptibility of SCC increases as the temperature is raised, but the oxidizing condition decreases as the 0 is 2
lowered. This work also showed that the effect of impurities, such as chlorides, are more pronounced in oxidizing environments and that at low electrode potentials (low oxygen) crack growth rates for stainless steel in chlorides are similar to those observed in pure water. Deaeration, which lowers the potential, therefore, can play a significant role in reducing the i deleterious effects of impurities, such as chlorides. 1 Indig @) used constant strain rate techniques to demonstrate the importance of deaeration during BUR startup. Deaeration to 0.2 ppm 02 at 30 C (86 F) and I maintaining that concentration up to 290*C (552*F) resulted in a dramatic l reduction in SCC compared to a sta' tup with 8 ppm2 0 finishing at 0.2 ppn et
! 290*C(552 F). Indig also observed that weld sensitized material was resistant to IGSCC during straining electrode tests at 300 F and 225 F. This was in contrast to the work of Ford and Povich, who found severe IGSCC at these 't I
I
22 temperatures. However, the latter investigators used furnace sensitized material. The straining electrode experinents were repeated on the welded specimen after the samples were given a low temperature sensitization treatment, and the results were in agreement with the Ford-Povich results, again suggesting that furnace sensitized material is nore susceptible to cracking than as-welded material. Table 3 shows the results of tests (15) to deternine the amount of intergranular attack for several different austenitic stainless steels, in every case, the nunber of coupons attacked was nuch greater fnr furnace sensitized coupons than for welded sensitized coupons. Indig also found that the resistance to crack initiation at 400 F and above improved as the applied stress was lowered. Sanples stressed at 30 ksi showed much greater resistance to IGSCC than samples stressed to 40 and 50 ksi and suggested that perhaps below 30 ksi complete resistance to IGSCC would occur in his tests using a simulated aerated startup. Stachle, et al. (26) showed that in oxygenated water, the tenperature at which stress corrosion cracks can be observed, is dependent on the strain rate. At 8 ppm 0 , the threshold for the observation of SCC is approxinately 200 C 7
-1 (392*F) at a strain rate of 10-5 sec . At a strain rate of 10-0 sec , the threshold is between 200*C (392 F) and 150*C (302 F). They concluded that since the rates of crack growth are highly temperature sensitive, long term (hence, very slow strain rates) tests are needed to observe cracking as the temperature drops, and that at 150 C (302"F), a strain rate of 10-7 sec ~1 or lower is needed to observe SCC. This suggests that the standard CERTs may be inadequate to establish the threshold temperature at which SCC will nnt occur and that extremely slow strain rates may be required to observe SCC (if it occurs) at temperatures expected in the suppression pool. It was also determined that the crack depth was proportional to the applied stress.
White and Berry (27_) reported that General Electric was able to produce cracking
-7 at 121*C (250 F) in water containing 8 ppm2 0 using a strain rate of 6.7 x 10 sec . The measured crack depths were 20 mils. White could not duplicate the results but found that the extrapolation of his data for cracking of furnace sensitized material showed that the lower temperature limit for SCC at 8 ppm l'
23 Table 3 COMPARATIVE SEVERITY OF llELDING AND HEAT TREATMENT ON INTERGRANULAR CORROSION (15)
. NUMBER OF RACKS SHOWING INTERGRANULAR ATTACK 3 AISI TYPE WELDED HEAT TREATED 2 316L 0 11 316 1 15 317 6 22-302 15 19 304 1 10 304L 0 5 321 6 12 347 1 13
' Racks (24 specimens exposed to a number of actual service i
environments for times ranging from 30 to 1600 days). 2 Specimens heat treated for one to four hours at 593 to 6770C. i l. i i l I l I 1
. i
-..r -
,m _ , _ , _ _ . , . . . _ - , -
24 02is about 121*C. A round-robin test was also conducted on special specimens from pipe sections sensitized for 20 hours at 1085'F, which were provided by General Electric. No evidence of SCC was observed after tests in oxygenated water (8 ppm) at 250*F (121*C) and a strain rate of 1 x 10-6 sec _1 ~ Ford and Silverman (2_8) reported IGSCC in high purity water down to 50*C (122'F) in a CERT provided the oxygen was in this region of 1 - 4 ppm. They aise provided a curve summarizing the work of others in producing stress corrosion at various temperatures and oxygen concentration (Fig. 4). Figure 5showstheresultsofotherworkconductedbyFordandSilverman(g)on the SCC rate in water with 1.5 ppa oxygen at 95*C (203*F). To obtain a crack growth rate of 10-8 cm/sec, an initial crack size of 8 x 10~2 cm and a stress level of 340 MN m-2 (49.3 ksi) is required, which is abnve what they estimate for residual stresses. They also assume that 8 x 10-2 cm is the inspection resolution limit. Using the static K ISCC curve, a crack depth of 2 x 10 -1 cm will result in a crack growth rate less than 10-8 cm/sec-1 This rate would be expected to be lower at suppression pool operating temperatures. 4.2.1 Stress Corrosion Cracking of Type 304 Stainless Steel - Summary The results of this literature review show that SCC in the suppression pool is very unlikely. Stress corrosion cracking has not been demonstrated in the laboratory for the conditions expected in the suppressicn pool. The CERTs that are used to define the threshold conditions for SCC do not take into account the time to incubate and initiate a nacrocrack. The use of these test data, therefore, to predict in-plant performance is conservative. 0 4 o f n
25 h h b I O 0,O F .o og. 9,05 g,% b a [n
.s . 3, llG M.C 1
o ,; y, 8!!i lo 0 99 20 e .o n ,e 75 .. Lo
! @._oi;
. 0 06 p 3 3j
$ : 'i :
a _ _ . c o. . 2." *1., , (sr.co Gat cI- v < C Tsiodem s 'l # .'I 4') e s -,-, I I I i f s
- 50 10 0 850 200 250 300 TEMPERATURE 'C .f f Figure 4 - Variation of Stress Corrosion Susceptibility of Sensitized I
304 Stainless Steel in BWR Quality Water for Various Oxygen / Temperature Combinations. Note: (*) Denotes no F cracking observed; (X) denotes IGSCC cracking observed at l an average propagation rate given by subscript number in. I cm s-1 x 10-7, L = 2.1 x 10-/ s-1 Hatched area denotes uncertainty of exact position of boundary line (28). s 1 3. i h J ty . e
-. - - ~
26 vg . . 77 T'T T {*~T r a I i TTT{ a
, y , i v iv j PROIAcal 04 8 Lit >i7%m sec4 Soo _ .
Ei sttISI' llc 3 g E 400 - D' 1I.".Eg_ E300 m i E ! LCA0mG Taut y AESt0ual 5 aq$ ~
= 200 516tSS S0 secs 0 $ socs 100- l B e 60-8 cms (8 i 8'
lA55uutD 4DT htSOLUTION Livif
' ' ' d ' ' '"' ' ' ' "'
0
2 10 go e ioO gol a ChAtaCEPin.cm
,g . . . ...; . i ii .g i i . . ...;
PROS A01.fiO*4 EATE si7km sec'l
$00
.uiscetSTAit:1 7
1400 - 3 a SffRESS_ U l g0 - { SUPERlWPOSED
~
g ; 1014 VIBRATIONS
! EISIDUAL R509 g 200 R = 0.7 -
i STRESS R =0 3 1 I iOO - I 8 U <lT ce sec'8 1
. lASSuut0 hDT EESOLUTION LIMIT 0
' ' ' 'I ' ' '"I ' ' ' 'I
2 10- 10~ ' 10 0
10' b (RACK DEPTH, tm l} Figure 5 - Stress / Crack Depth Design Diagram for Sensitized 304 Stainless Steel With Boundary Lines Between Subcritical Crack Propagation .. and Propagation at Rates Less Than 10-8 cm sec-1 for (a) Various lI Loading Times at 95
- 30C in Water /1.5 ppm Oxygen, and (b) Various R Values for Superimposed Vibrations of 10 Hz Fre-quency (Note that the important assumption that the cracked region is under plastic constraint has been made) (29,).
4 e
- 6, 27 5.0 CORRECTIVE ACTIONS Several options are available f or preventive and/or corrective measures in the suppression pool, should it become necessary. One of the concerns that has been expressed is how long will a fix last. If it does not last the design life of the plant, provisions must be made to repair the structure where a preventive measure was taken. If the cladding or environment are not altered and do not present any problems over the design life of the plant, any intended cure may becorie the problem. It would seem prudent, therefore, that the preventive measure taken should provide two benefits to be feasible. The first would be to forestall any potential problens in the suppression pool, and the second is to have a life much longer than the time in which corrosion of the carbon steel under the cladding may be a concern. Sone of the corrective actions that may provide these benefits have been reviewed and are described in the following sections.
5.1 Shot Peening Shot peening is a cold working process that is used to produce a compressive surface stress in a part. The surface of the part is bonbarded with particles with each particle acting as a small hanner. Cast steel shot is the nost connonly used material, but stainless steel, ceranic, and glass beads are also used, especially when control of contamination night be necessary. Nearly all fatigue and corrosion failures originate at the surface of a part, and, f or that reason, the surface compressive layer induced during shot peening
, can be of benefit in preventing crack initiation. Shot peening has been shown to increase the SCC resistance of 7075 T-6 aluminun in 3.5% Nacl fran 250 hours (as nachined) to greater than 15,000 hours (30). Moller (j31) has suggested shot peening as a preventive neasure to minimize chloride SCC in austenitic stainless. Intergranular cracking resistance of austenitic stainless steels has been shown to be improved by shot peening prior to sensitization by breaking up surface grains and forming nucleation sites for carbide precipitation below the surface (32). On the other hand, Povich (33) conducted CERTs in which he found that sensitized and shot peened Type 304 cracked nore readily than unpeened specimens in high temperature oxygenated water but warned that CERTs apply excessive stress and strain, and extrapolation to practical situations l
28 should be done with caution. Shot peening did appear to be beneficial where the level of strain was less than the amount needed to cause crack initiation. Dynamic strains of the nagnitude occurring in CERT tests would not occur in the suppression pool. It has been recommended that shot peening of stainless steel be performed with stainless steel shot or glass beads to minimize stress corrosion problems (34). Peckner and Bernstein (_35) have also recommended shot blasting adjacent to welds in stainless steel to apply compressive stresses to prevent SCC. Friske (36) reported on tests he conducted to determine if shot peening will protect sensitized 304 stainless steel from SCC. After sensitizing at 1000 F for 100 hours, one set of specimens was peened to an Alnen intensity of 0.008A prior to the SCC test in flgCl ; the other was tested in the sensitized condition. The 2 sensitized and peened specimen survived 1004 hours while the unpeened specimen f ailed in two hourr. Several other successful uses of shot peening sensitized welds have been described (H). 5.2 Coatiegs and Cathodic Protection Coatings and cathodic protection are nentioned here as mitigating measures that should be considered, but the practicality and suitability of their use in the wetwell environment should be evaluated. Protective coatings provide a barrier film between the metal and the electrolyte. There are many protective coatings available, several of which appear suited for this application. Coal tar epoxy, coal tar enamel, vinyl, and urethane are but a few of the coatings that exhibit excellent resistance to acids, alkalis, and water, as well as other environnents. They can be used with many types of inhibitors and primer coatings. Thermal spraying is also used to form a coating by depositing metallic or non-metallic materials. One of the major uses is to provide coatings resistant to salt water. These processes include electric arc spraying, flane spraying, and plasma spraying. Of the . three, plasma spraying produces coatings that are generally superior, and the coating density can be 95%. Almost any metal can be sprayed. Aluminum sprayed coatings on steel have been exposed for over 20 years in very severe atmospheric conditions and have given perfect protection (38). In a report of
29 - a 19-year study of corrosion tests of flane sprayed steel coupons immersed in seawater in Texas and North Carolina, aluminun sprayed coupons showed almost perfect protection (39). In the very few instances where blisters were observed, the underlying steel was completely protected by the sacrificial protection of the aluminun. Vinyl-sealed panels afforded the optimum resistance. The aluminum sprayed coating was only 0.003 inch thick. Zinc Sprayed coatings showed less protection. Cathodic protection is a very effective means of corrosion control. It is achieved by supplying electrons to the metal structure to be protected by sacrificial galvanic coupling or through the use of a power supply. The high resistivity of the water may require a large nunber of anodes to protect the structure adequately. However, if cathodic protection was used in conjunction with coatings, it would serve as a back-up in the event a break occurred in the coating. This would be especially attractive if only the weld regions need to be coated and protected. The anodes or electrodes could then be lined up with the weld regions and would significantly reduce the size of the cathodic protection system required. 5.3 pH Control Increasing the pH in the wetwell could significantly reduce the propensity for pitting attack as well as reduce the general corrosion rate of both low alloy and stainless steels. In the case of pitting, an increase in the bulk pH of the solution will increase the pitting resistance, and a similar improvement with increasing pH has been reported for chloride cracking resistance (_15). The corrosicn of stainless steel is increased by lowering the pH and raising the temperatu re. Dissolved oxygen, carbon dioxide, and low pH all accelerate the attack of steel in freshwater. Holler (31) has found that low pH solutions in the presence of halogens should be avoided. In tests with 47. Nacl at 90 C (194 F), he found that strong alkaline solutions resist pitting and at pH 10.5 pitting ceases altogether, even with crevices present. Uhlig (8) also found that alkaline chloride solutions pit less or not at all compared with neutral or acid chlorides. Staehle (40) L t
?
30 found that in tests in a circulating autoclave, increasing the pH to 10.9 significantly increased the time to f ailure for 304 stainless steel. Cheng(3) found that riaintaining pH close to 10.5 in high temperature water is beneficial for both iroi and stainless steel. DePaul (4_2,)2 found that the corrosion rate of stainless steel was reduced by a factor of two at pH 10 compared to the rates in neutral water. Kaikawa, et al (22) showed that in tests conducted with constant load coupons in oxygenated water, all specimens failed after 50 hours at a pH of five,100 hours at a pH of six, but no f ailures were experienced after 1000 hours with a pH of seven or nine. For the general case, the potential -pH (Pourbaix)diagramsgivenbyShrier(3)indicatethatthecorrosionofferrous rietals ceases (or is drastically reduced) in the pH range 9.5 to 12.5. I l.
31 6.0 DISCUSSION AND SUt: MARY The results of the literature search have shown that IGSCC can occur in pure water at elevated temperatures if oxygen is present, even in small amounts. At lower temperatures, the amount of oxygen that can be tolerated increases. In
, the present case, cracking of the suppression pool liner is not expected because the operating temperature (90*F) is below the temperature threshold for SCC.
The majority of the test data that could be located was focused on the cracking problens that occurred in the BWR primary systens. These data are not considered to be directly relevant to the suppression pool problem, and if they are used to assess the potential for cracking, the result will be conservative for the following reasons: Loading - Almost all the testing has been performed under dynamic loading conditions which are appropriate for primary coolant systems because of the pressure and thermal cycles that result from a startup. However, the loads in the suppression pool are for the most part static, resulting fron fabrication and dead weight loads. Dynanic strains occur only periodically and are small when compared to the dynamic strains in primary circuit piping. A major part of the life of a stress corrosion crack is spent in initiation, and the constant extension rate technique is an accelerated
, test and is not designed to measure initiation times. Data from constant load or constant deflection tests would be more relevant to the situation as it exists in the suppression pool, but these tests would not provide data in a reasonable time. Constant extension rate techniques do provide some information for comparative purposes provided the extension rate is sufficiently high. As noted earlier, cracking could be observed at lower temperatures and concentrations of oxygen by lowering the strain rate, but the strain rate would be so low as to be impractical for the current problem. In summary, the data from the CERT ; are conservative for material that is subject to constant load or deflection.
I 4 L D
32 Temperature - Almost all the tests conducted have shown a threshold temperature above the temperature experienced in the suppression pool. Only one test has been reported showing racking at less 120*F, and the
~
details of the tests are not known. It is impnrtant to note that other work does not support this. The majority of the work indicates that a threshold temperature almost 100 F higher than exists in the suppression pool is necessary to produce cracking in the CERTs. It may be even higher in constant load or deflection tests. Contaninants - The literature strongly suggests that oxygen is perhaps the critical contaninant that induces liiSCC. Hydrogen peroxide was added to the water in several laboratory tests because of its presence as a result of radiolytic decomposition in NJRs. As the temperature was lowered, nore oxidant was required to induce cracking. Chlorides were found to accelerate cracking, but the presence of oxygen is a requirement. In addition to oxygen and chlorides, contaminants, such as sodium thiosulfate and chlorides,'are aggressive in their ability to induce IGSCC, and the inmunity of the stainless steel to SCC will only be naintained if these species are excluded from the pool. The sensitization that exists in the weld HAZ in the suppression pool is more severe than the rest of the cladding in that the welds have seen additional thermal cycles from welding. However, similar conditions have been simulated in
! many of the tests referred to earlier, and the threshold temperature for cracking is still above that expected in the suppression pool.
The stresses that exist in the cladding are near yield level, both in the weld region and in areas renote from the weld (due to differential thernal expansion during cooling from the heat treatment tenperatures). This is not a desirable situation, but there is some consolation in that the applied or dynanic loads 3 are infrequent, and they will not add significantly to the possibility of crack initiation or rate of propagation. The driving force for initiation of SCC will cone from the residual stresses, and these will relieve or redistribute as a crack propagates, making further propagation nore difficult. t
T i, 33 The results of the investigation show that there is no risk of IGSCC in the suppression pool even in sensitized material at the operating temperature of 90"F. If aggressive contaninants are permitted to accumulate, a possibility of cracking must be assumed. However, if it should occur, it would probably take many years to initiate, and the expected growth rate would be very low. If cracking should initiate, indicating a more aggressive environment than expected, the problem then focusses on whether the underlying carbon steel would be attacked at a rate sufficient to corrode through during the life of the plant. Since the environnent has now become aggressive enough to attack the stainless steel, it should be assumed that it is aggressive enough to corrode the underlying carbon steel, especially at the tip of a crevice. The structural analysis has shown that fracture would not occur, and the structure would, therefore, corrode through the wall until it leaked. In the weld regions, however, cracks have already been detected, and it would seem prudent to assune that the plant will go into service with some isolated cracks still present in the HAZ of the welds, some of which nay have penetrated to the carbor steel or very nearly so. The decision is whether preventive measures should he taken now or should oction ) be taken only if corrosion of the carbon steel occurs. This decision seems to be one of economics rather than of safety considerations provided a means of surveillance is established to ensure that corrosion of carbon steel can be detected. A reconnended approach would be the following: Welds - Initial or additional cracking would probably occur first, if at all, in the HAZ because of the additional sensitization resulting from the welding processes. Preventive action to be considered could include several possibilities. A flame sprayed coating of aluminum, sealea with a vinyl or other suitt. ole sealant, appears to offer a protective barrier that shows potentia' to last the life of the plant. An added benefit would be the sacrificial protection offered by the aluminum coating in the event of a holiday in the coating. i n
34 Cladding Renote Fron Welds - No inmediate action is proposed. Corrosion Coupons - Corrosion coupons could be placed at strategic locations in the suppression pool and exanined periodically. The coupons should be designed to , ensure that they would show evidence of an aggressive environnent prior to attack of the cladding or, of the carbon steel. Coupons could consist of portable compact tension specimens, U-bends, and double U-bends with and without graphite wool to simulate crevices. They should be treated and stressed in a nanner that is more severe than the service material. Similar coupons that have been protected by coating could also be exposed. Inspection and ibnitoring - Insitu inspection of the suppression pool to exanine for evidence of rusting and cracking should be considered. An inspection progran should be invokved whether or not preventive measures are taken. i l' 1 l l i l. L
[ - 35 REFERENCES
- 1. Egan, G. R., et al., "The Significance of Sensitized Stainless Steel flaterial in Drywell Vent and Containment Structures in the Perry Nuclear Power Plant - Fracture and Fatigue Evaluations," APTECH Draft Report No.
AES-81-11-88 (November 1981).
- 2. Letter from Gilbert /Conrnonwealth to APTECH, dated August 25, 1981 (Ref. PY-STR-1365).
. 3. r;upka, I., et al., "Some Remarks on the Analysis of Stress Corrosion Cracking of Austenitic Stainless Steel Cladding," In Stress Corrosion Cracking Problems in Primary Pressure Systems, International Working Group on Reliability of Reactor Pressure Components, International Atomic Energy Agency Report No. IWG-RRPC-78/2 (flarch 29-31, 1976), Pp. 138-144.
- 4. Vinckier, P., "A Review of Underclad Cracking in Pressure Vessel Components,"
Welding Research Council Bulletin 197.
- 5. Cobb, G. ft., " Report on Stainless Steel Clad Plate Material Used in Con-struction of Perry Nuclear Power Plant Units 1 and 2," Revision A, Newport News, VA, Newport News Industrial Corporation (May 14,1981).
- 6. Godard, H. P. , " Corrosion of Metals by Waters," Materials Performance (May 1979), Pp. 21-27.
- 7. American Society for Metals, Metals Handbook, Volume 1, 9th Ed. , tietals Park, OH, ASM, c1978, Pp. 734.
- 8. Uhlig, H., Corrosion Handbook, 1948.
/- 9. Nelson, G. A., The Corrosion Data Survey, Shell Development Corporation, (; 1960.
- 10. Matsudaira, Corrosion 37(5)(May 1981).
I 11. Forgeson, B. W., et al., " Corrosion of Metals in Tropical Environments, l Parts 1-5," Corrosion, Vol.14 (February 1958 and September 1958), Vol. 16 (flarch 1960 and October 1960), and Vol. 17 (July 1961).
- 12. Bush, S. H., " Stress Corrosion in Nuclear Systems," In Stress Corrosion Cracking Problems in Primary Pressure Systems, International Working Group i'- on Reliability of Reactor Pressure Components, International Atomic Energy
., Agency Report' No. IWG-RRPC-78/2 (March 29-31,1976), Pp.115-130.
- l. 13. General Public Uti'lities, " Investigation of Intergranular Stress Corrosion l Cracking at Three Mile Island, Unit 1" (January 4, 1980).
- 14. Private communication, flichael Fox, Electric l'ower Research Institute.
e e i.-
F~
- 36
- 15. Sedriks, A. J., Corrosion of Stainless Steels, New York, John Wiley and Sons, 1979.
- 16. Bryant, P. E. C., " Fluoride Induced Intergranular Corrosion of Sensitized Austenitic and Austoenoferritic Stainless Steels, "In Stres_s Corrosion Cracking Problems in Primary Pressure Systen.s, International Working Grouo on Reliability of Reactor Pressure Components, International Atomic Enerny Agency Report No. IWG-RRPC-78/2 (March 29-31,1976), Pp.105-115.
- 17. Szklarska-Smialowska, Z. and G. Cragnolino, " Stress Corrosion Cracking of Sensitized Type 304 Stainless Steel in Oxygenated Pure Water at Elevated Temperatures (Review)," Corrosion 36(12):653-665 (December 1980).
- 18. Andresen, P. L., "Effect of Simulated BWR Type Environments on the SCC of Carbon Steel and Stainless Steel," BWR Water Chemistry Projects Review Meeting, Electric Power Research Institute (flay 13,1980).
- 19. Vermilyea, D. A., Corrosion 29:442(1973).
- 20. White, E. L. and W. E. Berry, "The Influence of Cyclic Load and Environ-mental Effects on Stress Corrosion Cracking of Sensitized Stainless Steel,"
Final Report No. EPRI NP-1991, Columbus, OH, Battelle Columbus Laboratories, EPRI RP311-3 (August 1981).
- 21. White, E. L. and W. E. Berry, "The Influence of Combined Environmental Effects on Stress-Corrosion Cracking of Welded Stainless Steel Piping, Progress Report for October - December 1978, Columbus, OH, Battelle Columbus Laboratories, EPRI RP311-3 (February 15,1979).
- 22. White, E.L., W.E. Berry, and W.K. Boyd, "The Influence of Combined Environmental Effects on Stress-Corrosion Cracking of Welded Stainless Steel Piping", Progress Report for October-December 1977, Battelle Columbus Laboratories, Columbus, OH,RP311-3(February 15,1978).
23. Kaikawa, et al., "Deaerated Startup Operation as a Countermeasure for BI'R Pipe Cracking," Seminar on Countermeasures for BWR Pipe Cracking, Session 5, Electric Power Research Institute (January 22-24,1980). 24, Andresen, P. L. and F. P. Ford, " Technical Considerations for Startup Deaeration Procedures to Mininize Stress Corrosion Cracking of 304 Stairo less Steel Piping," Seminar on Countermeasures for BWR Pipe Cracking, Session 2, Electric Power Research Institute (January 22-24,1980).
- 25. Indig, M. E., " Controlled Potential Simulation of Deaerated and Nondeaerated
' Startup," Seminar on Countermeasures for BWR Pipe Cracking, Session 5, Electric Power Research Institute (January 22-24, 1980), Pp. 1-34
- 26. Staehle, R. W. , et al., " Corrosion and Corrosion Cracking of Materials for
' Water Cooled Reactors," Progress Report for July 1,1976, to December 31, 1976, Fontana Corrosion Center, The Ohio State University, RSFD 4152-Al-3.
- 27. White, E. L. and W. E. Berry, "The Influence of Combined Environmental Effects on Stress Corrosion Cracking of Welded Stainless Steel Piping,"
- t Columbus, OH, Battelle Columbus Laboratories, EPRI RP311-3 (April 1979).
if-m O
s q 37 28. Ford, F.P., and M. Silverman, "The Prediction of Stress Corrosion Cracking of Sensitized 304 Stainless Steel in 0.1M NASO 4, " Corrosion 36(10:558-5G5 (October 1980). I i
- 29. Ford, F.P., and M. Silverman, "Effect of Loading Rate on Environmentally l Controlled Cracking of Sensitized 304 Stainless Steel in High Purity Water",
Corrosion, Vol. 36 (11), 597-603 (November 1980).
- 30. Metal Improvement Company, Shot-Peening Applications, 4th Ed.
- 31. Fblier, G. W., " Misuse of Stainless Steel," Western States Corrosion i
Seminar, Proceedings (May 3-5,1977), Paper No. 32/1.
- 32. Johnson, J. R. and J. J. Daley, " Shot-Peening Takes Expanding Role,"
Power (February 1975), Pp. 68-70.
- 33. Povich, M. J., " Constant Strain Rate Studies on Shot Peened Type 304 Stainless Steel in High Temperature Aqueous Environments," Corrosion 34(5):162-169 (May 1978).
- 34. Campbell, J. E. , " Shot Peening for Improved Fatigue Properties and Stress
?
Corrosion Cracking Resistance," Metals and Ceramics Information Center (MCIC-71-20) (December 1971).
~
- 35. Peckner, D. and 1. M. Bernstein, Handbook of Stainless Steels, New York, McGraw-Hill Book Company, 1977, p. 16-84.
- 36. Friske, W. H. , " Shot Peening to Prevent the Corrosion of Austenitic Stain-less Steels" (Al-75-52) (September 15,1975).
- 37. Private communication, Paul field, Metal Improvement Company.
- 38. Shreir, L. L. , Ed. , Corrosion, Volume 2, Boston, Newnes-Butterworths, 1977, p. 13:84.
1 i 39. American Welding Society, Committee on Thermal Spraying, Corrosion Tests of Flame-Sprayed Coated Steel,19-Year Report, Miami, FL, AWS, 1979
, (AWS C2.14-74),
- 40. Staehle, R.W., Progress Report for October 1,1970, to June 30, 1971 The Ohio State University Research Foundation, Columbus, Ohio, Contract AT(11-1)-2069.
- 41. Cheng,C.F., Corrosion (20)341T(1964).
i
- j. 42. DePaul , D.J. , Corrosion 13:75(1957).
- 43. Shreir, L.L., Ed. , Corrosion, Volume 1, Boston, Newnes-Butterworths, 1977, p. 1:59.
- L
.w
- - - . - - ,}}