ML20202F300
| ML20202F300 | |
| Person / Time | |
|---|---|
| Issue date: | 07/01/1986 |
| From: | Advisory Committee on Reactor Safeguards |
| To: | |
| References | |
| ACRS-T-1530, NUDOCS 8607150131 | |
| Download: ML20202F300 (429) | |
Text
-
c OllG NAL
~
d O UNITED STATES NUCLEAR REGULATORY COMMISSION IN THE MATTER OF: DOCKET NO:
ADVISORY COMMITTEE ON REACTOR SAFEGUARDS SUBCOMMITTEE ON METAL COMPONENTS LOCATION: COLUMBUS, OHIO PAGES: 1- 208 DATE: TUESDAY, JULY 1, 1986 r . . . -.m.,,
si 0 h!;t ; d.: . J. m 11 O!;i ,' i. . 3IC~.7DiIce (
ACE-FEDERAL REPORTERS, INC.
O O!Yicial Reporters 444 North Capitol Street Washington, D.C. 20001 (202)347-3700 8607150131 e60701 PDR ACRS PDR T-1539 NATIONWIDE COVERACE s--
' 1 -
273134 .
O 2 3
4 MEETING OF Tile ACRS METAL COMPONENTS SUBCOMMITTEE Si 6
7' i BATTELLE COLUMBUS DIVISION 8l i
9 505 KING AVENUE 10 COLUMBUS, OIII O ll!
12
~
13 14 i
i 15i i 16 17 18 JULY 1, 1986 l
19i 20 4 21 1
22 23 24 25 l
{
._.7, ..__.m.,_m._m..._ , ,__, e. _m,_,_,.,,,-,_,
2 1 Present:
9b 2
3 P.G. Shewmon, ACRS 4 Harold Etherington, ACRS 5 W. Kerr, ACRS 6 Al Igne, ACRS 7I M. Bender, ACRS Consultant 8 E. Rodabaugh, ACRS Consultant 9i J. Hutchinson, ACRS Consultant 10 11( ;
12 Gery Wilkowski, Battelle 13 Mike Mayfield, NRC Research 14 Guy Arlotto, NRC Research 15 16l t 17 18 i
19 i
20' 1
21 '
22 1
23 24 25 l i
1 e,--. . -, , -, ,. . . _ . . _ _ . . _ _ . . .--_.--.,,_-,,,,,,,,,,-,---,,n,--,- -.--...,,-,,--,n,-,,,-,-,,. . .--. . ,~w,n,.,--_,.
-- ~
3 g 1, INDEX l
2 Subject _ _ Speaker_
Page No.
3,I Degraded Piping Program Overview - (G. Wilkowski) 5 i
j Limit-Load Analysis and 5 Plastic-Zone Screening
, Criterion - (P. Scott) 36 6
- Significance of Material 7, Behavior at LWR l Conditions -
(M. Landow) 59 8
l Fracture Evaluations of Weld 9 Overlay Repaired Pipe - (P. Scott) 79 I
10 Introductory Comments on Elastic-
! Plastic Fracture Mechanics 11! Analyses - (J. Ahmad) 93 t
i 12 Load and Load Versus Displacement l Predictions for Through-Wall O 13 Cracked Pipe - (F. Bud Brust) 109 14 Complex-Cracked Pipe Evaluations - (G. Kramer) 122 15
[ Finite Element Analysis 16 Verifications of Estimation l Schemes - (V. Papaspyropoulos) 134 17 l l Closing Comments and 18 Discussion -
(G. Wilkowski) 144 i
. 19 Evaluation of the Ductile Fracture of Carbon and 20 Stainless Steel Welds -
(R. Hays) 148 21 Piping Fracture Mechanics 22 Data Base - (A. Hiser) 184 23
(~ 24 25
PUBLIC NOTICE BY THE UNITED STATES NUCLEAR REGULATORY COMMISSIONERS' ADVISORY COMMITTEE ON REACTOR SAFEGUARDS TUESDAX, JULY 1, 1986 The contents of this stenographic transcript of the proceedings of the United States Nuclear Regulatory Commission's Advisory Committee on Reactor Safeguards4 (ACRS), as reported herein, is an uncorrected record of the discussions recorded at the meeting held on the above date.
No member of the ACRS Staff and no participant at
() this meeting accepts any responsibility for errors or inaccuracies of statement or data contained in this transcript.
4 i 4
1 MR. SHEWMON: This is a meeting of the 2 ACRS Subcommittee on Metal Components. I'm Paul 3 Shewmon, chairman of the subcommittee. The other ACRS members in attendance are Harold Etherington, )
4 '
5 Bill Kerr on my right, and beyond that there's ACRS I
{ 6 consultants Bender, Rodabaugh and Hutchinson. The i
I 7 subcommittee today will review the research program 3 8 on degraded piping being performed at Battelle as '
- 9 well as some work at Argonne and possibly other labs.
i
- 10 Al Igne on my left is the ACRS member for the 11 meeting.
12 The rules for participation have been
) 13 announced as'part of the notice of this meeting that
! 14 was published in the Federal Register on June 24th.
I i 15 We do have a recorder over here, and it's requested i
16 that each person speaking identify himself or i 17 herself and speak loudly enough so that she can be 18 readily heard.
I i 19 We've received no written comments or i
- 20 requests for time to make oral statements from
) 21 members of the public. I don't see any other people 22 wishing to make comments, so we'll now proceed with
, 23 the meeting. And I'll call on Gery Wilkowski, g 24 project manager, Degraded Piping Program of Battelle, 25 to begin.
n-,e-a .,-r .-. -._n-_ .. ~. p ,., ,-w.w, ..,,_.,,,mm,..n, ,...,,m ,,_n,e n___.,,- ..,g.,n_,,,---..e ,,mw .y,_ae.-,-m , c ,. w -.e mw,.
i 5
4 1 MR. WILKOWSKI: Good morning, gentlemen.
- 2 My name is Gery Wilkowski with Battelle. During 3 this morning e plan to give you a number of brief i
! 4 presentations, which hopefully will discuss a little f 5 bit of technical aspects of the program, try to hit i 6 more towards what the results and the significance 7 f the results are.
l 8 The agenda that you see up here is in the ,
9 handout that you have. I'll start off with an
! 10 initial introduction on the Degraded Piping Program, 11 and then afterwards a number of our staff members 12 will talk about specific technical topics, and at
, () 13 the end then I'll have some closing comments and 14 time for some additional discussions.
15 The Degraded Piping Program is a three 16 year program. We're currently just past two years i 17 of the program, and we have a large number of people 18 that do work on the program. There's myself, Dr.
i 19 John Kiefner who is the deputy manager, a number of ,
j 20 key people, technical people, and also we've had one
! 21 visiting scientist from MPA Stuttgart in West 22 Germany, who was here for one year in time. He has i
23 since returned back to Germany.
, 24 The objectives of the Degraded Piping
, 25 Program are to verify, improve and develop flaw
-.m-- - - ~ - - ._,,,,,._.,,,__._.,,% w._ ,,_c, .
, , . , _ , , . _ _ . . . , _ _ , , , _ . . .,,_...,,_,.,.r ....,-.,.m...%. m. r_,., ,,.e..~ -
6 s 1 assessment analysis for leak-before-break
~# 2 methodology. The evaluations of the methodology 3 that are underaay involves experimental assessments t
In this case the pipes 4lI of the actual failure modes.
1 5l that we are testing, most of them have b6cn procured 6f from cancelled nuclear power plants, so they are i
7l pedigreed materials. We are doing tests at 550 8 degrees fahrenheit, LWR conditions. So we will also I
9l assess what actually happens to the failure of these I
10 pipes as opposed to what one might predict in the 11l experiments.
I One of the next steps is the simplest 12l (p) 13l types of analysis to predict the strength of cracked 14 piping is a limit-load analysis. So in this case 15 we're conducting a number of experiments and 16 evaluating whether the limit-load analyses, which 17 are used in many of the failure criteria in the la nuclear industry, indeed are conservative analyses.
19: To evaluate whether the limit-load analyses are 20 appropriate, we've developed a screening criteria to 21 show when it might tend to overpredict the loads or 22 underpredict the loads.
23 In the course of the program we've also
(g 24 done a significant amount of material LJ 25 characterization at LWR conditions. We find some
7 1 unusual behavior in some of the welds and an effect 2 which we believe might be due to dynamic 3 strain-aging where the crack tends to exhibit local l 4 instabilities and jumps a significant amount in both 5 the laboratory specimen tests as well as in the full
- 6 scale pipe tests.
7 Another evaluation that you'll hear about 1
8 later is the prototypical weld overlay repaired I
9 fracture experiments. In this case we're comparing 10 the experimental results to the ASMEIWB-3640 11 analysis procedure.
l i
- 12 Another aspect is evaluating l
i
) 13 elastic-plastic fracture mechanics, estimation j 14 schemes. In evaluating fracture behavior of piping 15 it's a very complicated type of procedure, and we're 16 looking at ways to predict not only the loads at 17 crack initiation and the maximum load, but also how l
2 18 to predict the load versus displacement behavior of 19 the cracked pipes. Predicting displacements is much l 20 more difficult to do.
21 And finally, many of the estimation i
22 schemes that are ongoing and exist for these types 23 of analysis of loads, loads versus displacement, are 1
24 indeed just that, they're estimates of what might j 25 actually happen. So we're doing a number of j
_ . . _ . - . - _ _ . . . _ . . . . . - - . . . , . , - . _ , , _ , _ . _ , _ . , . _ , , - - . _ . . , . - - _ . _ . , . ._ . . , _m... , . - . . . . . . - . _ . . , - . . . , .
8 3 1 detailed finite element analysis to verify these
'NJ 2 estimation schemes.
3 Within that work the scope of the program 4 essentially is outlined in the next two view graphs, 5 that is, first of all, we're involving experiments 6 on circumferential1y cracked pipe, not actually 7 cracked pipe or helically cracked pipe, and the flaw 8 geometrics under evaluation are simple through-wall 9 crack, and idealized through-wall cracks, internal 10 surface cracks in the pipe, and finally a complex 11 crack, where a complex crack is a case of a very 12 long surface crack that might have grown and I
(~N y) 131 partially penetrated through tne wall thickness, so 14 it looks like the Duane Arnold safe end flaw.
15 MR. ETHERINGTON: That's one of the same 16 length on inside and outside.
17 MR. WILKOWSKI: That's right. The 18 materials that are evaluated, most of them we 19 procured from cancelled power plants. There's 20 austenitic materials, which would be wrought 304, 21 316, or Inconel 600; carbon steels, A106, SA333; as 22 well --
as I forgot the A516 Grade 70. There's some 23 of that material as well. Austenitic welds, both
'\ 24 TIG welds and flux welds, and the carbon steel welds.
(V 25 MR. SHEWMON: Are those austenitic welds
9 1 all 304 against 304 and 316 on 316 --
\
2 MR. WILKOWSKI: Yes, that's right.
3 MR. SHEWMON: -- or is there some' cast 4 stainless steel on some of those welds too?
5 MR. WILKOWSKI: We do have some cast 6 stainless steel in the program that has been sent to 7 Argonne for aging, and in that case those are actual 8 cast stainless steel welds, because the welds were 9 fabricated by a vendor, and we have the whole length 10 of the piping system.
11 MR. SHEWMON: You said, you have some cast 12 stainless steel and you're talking about welds.
) 13 MR. WILKOWSKI: And the piping. The whole 14 piping system was removed from service, and we 15 bought not only the pipe itself but welds that were 16 fabricated by the vendor.
! 17 MR. SHEWMON: Okay.
i i 18 MR. WILKOWSKI: We do have a couple 19 different carbon steel submerged arc welds that have 20 been evaluated in the program.
i 21 The types of loadings that we've leaked at 22 are, first, looking at the case of pure bending and 23 then getting pressure induced axial tension and then 24 looking at combirations of pressure and bending.
25 We are not in this program looking at
10 1 torsional loads, for instance. That has been i
A 2 outside the scope of what we're doing.
3 The pipe sizes have ranged from a 42 inch 4, diameter down to four inches in diameter. The i
i 5 thicknesses have ranged from a quarter inch up to .
1 I
6 three inches in thicknkss.
- 7 MR. BENDER
- Excuse me. There are no t
i 8 dynamic loads in this.
I 9 MR. WILKOWSKI: That's right. This is all i 10 quasi static loading. Tomorrow morning you'll hear 11 a little bit about another program called the IPIRG l 12 program, in which the scope of that program is aimed I
() 13 more towards dynamic loading. This is developing 14 the technology that's needed for that IPIRG program l 15 to assess what happens under the dynamic loadings.
16 MR. BENDER: Okay. Thank you.
17 MR. WILKOWSKI: To continue on, most of l 18 the pipe tests are conducted at 550 degrees j 19 fahrenheit. Tensile tests are conducted at room 20 temperature, 300 and 550, to see what the effects of 4
21 temperature are as well as the fracture toughness 22 specimens, called the J-R curve test in this case.
23 They are conducted at 300 and 550 F, which are the 24 upper and lower bound temperatures generally for
(
25 high energy piping systems.
l 11 l
I 1 The analysis efforts involve looking at 4
2 simple limit-load, which is a net-section-collapse l 3 analysis, developing a plastic zone screening i
4 criteria for the net-section-collapse analysis.
5 21astic-plastic fracture estimation schemes involve 6 two different types. One is an Eta factor analysis 1
- 7 in which you can calculate the toughness of the 8 material from the pipe test. However, that type of 9 analysis you cannot predict loads or displacementa i
10 that might occur in the piping --
plant piping i
l 11 system, but you can calculate the fracture 12 resistance from that test.
) 13 The other types of analyses, which are 14 mainly what we're concentrating on, are predictive 15 analyses where you can predict the load displacements 16 or crack growth, or you could predict the J-R curve 17 if you don't have the J-R curve.
i 18 Finally, finite element analyses involve 19 verifying the J-estimation schemes that are commonly I
20 .used, looking at geometry effects, especially in the 21 case where we have welds in a specimen, where there 22 is a mismatching in the strength between the weld 23 and the base metal, and also assessing other 24 advanced fracture mechanics parameters in case the 25 types of parameters we're using now do not work.
i
)
4 12 1 MR. SHEUMON: Not working would mean what, 2 or how would you know if a system wasn't working?
3 MR. WILKOWSKI: Mainly by comparison 4 between the estimation scheme predictions and the 5 full scale tests that we're conducting as well as t
6 tests from other organizations. So we are always 7 making comparisons between them. Can we predict the 8 amounts of displacement that are occurring in the 9 tests? Can we predict the amount of crack growth 10 that's occurring between initiation and maximum load?
11 MR. SHEWMON: You don't have enough 12 disposable parameters to shield that or to hide it
) 13 any way, just fit it?
14 MR. WILKOWSKI: There are a number of
- 15 parameters. However, only one of the parameters has i
, 16 really been developed significantly for use in 17 estimation schemes that allow predictions to be made ,
i l 18 for plant piping. For instance, most of the work 19 that we're doing uses the deformation J integral 20 parameter.
21 However, there are some incremental I i
22 plasticity parameters that exist, but they are very 23 complicated parameters. They always require the use 24 of a finite element analysis. There are no simple 25 estimation schemes that exist for those parameters
J 13
- - 1 right now. So we're trying to make an assessment of 2 current analysis procedures and doing some i 3 improvements to them. We're not doing a major 4 development of new analysis procedures.
5 MR. BENDER: I would like to ask a i
6 question about the materials that have been selected 4
7 for analysis. These bracket what systems?
I j 8 MR. WILKOWSKI: They're pretty y 9 representative of most of the primary systems.
1 10 MR. BENDER: Primary means?
i 11 MR. WILKOWSKI: Primary and secondary i
12 systems.
( 13 MR. BENDER: And secondary. So it's 14 primary coolant, the steam systems and BWR and PWR 15 systems.
16 MR. WILKOWSKI: That's right. They're I
17 very typical.
i i 18 MR. BENDER: And they're intended to
! 19 display the behavior of straight pipe as opposed to i
20 bends.
21 MR. WILKOWSKI: Yeah, there could i-22 potentially be some interactions between, for
- 23 instance, if you had a crack right in the nozzle
[\ 24 where the nozzle could stiffen the piping system as
' (LJ 25 opposed to a straight section of pipe where you
,_,__...m m..7_ . . _ . . , . . - _.___.-....___,._.,___..,,_,.m.,_,,._., , , . . . , . , ..__-m.. . . _ . , , . _ , _ . , , _ , . ,
i
'f
?
14 1 would have a crack, where you have more flexibility i-
! 2 to the piping.
3 MR. BENDER: Okay. Thank you.
4 MR. WILKOWSKI: Some of the interactions i 5 that we've considered within the program have 6 involved both regulatory, industrial and
{
7 international policies and how we might impact upon i
8 these. For instance, within the NRC regulatory 1
9 evaluations, there's the BWR pipe cracking, l
3 10 assessment of flaws in BWRs, in which there are two 4
I 11 different NUREG reports that make assessments of how 12 to treat flaws in BWR piping.
i j 1. 3 The results of a program would also impact i
l 14 upon the removal of' pipe whip restraints and jet 15 impingement shields, mainly from the point of
- i 16 verification of the analysis procedures that are ,
i 17
~
being used.
18 As far as industrial plans, what we're
- 19 doing right now strongly impacts the IWB-3640 I
20 austenitic piping criteria. The ASME is also 21 developing a future criteria for carbon steel piping 22 as well. Many foreign governments also have i 23 leak-before-break policy. For instance, the 4
a 24 Japanese have an austenitic piping criteria for BWRs.
25 They're currently working on a carbon steel piping .
?
15
- 1 criteria which --
2 MR. SHEWMON: Do you know what the 3 Japanese use for leak-before-break in BWRs? It 4 interests me, because in this country the NRC staff 5l has talked as if they wouldn't allow it, though i
6 possibly in these -- when they think they have 7 corrosion under control, they would be willing to.
Bj MR. WILKOWSKI: They did go in and change i
out all of their piping systems to what they 9{
1 10l considered to be a material that would be not lit susceptible to stress corrosion cracking. I'm not i
l familiar with all of the details right at this 12l (s
( ) 13 moment of what they've done, but I could get you 14' some more information on it if you want it.
15 MR. SHEWMON: Well, I'm curious. Go ahead.
16 MR. WILKOWSKI: Okay. The carbon steel 17 criteria is one that they're currently developing, 18 and they hope to have that in place within the next 19 three years. The West Germans also have a basis 20' safety approach for leak-before-break.
21 Interestingly, they have applied that 22 right now only to BWRs as opposed to PWRs. It's 23 only been applied to BWRs where they have replaced f'N 24 the piping system with carbon steels. They do have
'J, 25 plans to apply it for the newer --
I think it's l
16 1 convoy series of PWRs that Craftwork Union is (s,
2 building.
3 MR. RODABAUGH: Having applied this 4 leak-before-break, what's the following consequence?
5 For instance, do they not postulate breaks anymore?
6 Is that the end point of their policy?
7 21R . WILKOWSKI: Of the West German policy?
8 MR. RODABAUGH: Yeah.
9 MR. WILKOWSKI: I believe so. I don't 10 know as much detail about it as I would, but, yes, 11l that's the impact of it. They do have a criteria as 12l far as flow from the leakage of about ten percent in
(~N i
() 13 l the cross sectional area that they use.
I 14j MR. RODABAUGH: That's far from a 15 guillotine break.
i 16! MR. WILKOWSKI: Far from the guillotine 17 break, that's true.
MR. BENDER: Isn't it true that the 18 l. I i !
19' postulation is to give them some basis for i
20 establishing loading for structural evaluation 21 purposes?
22 MR. WILKOWSKI: Yes.
23 MR. ETHERINGTON: If industry has plans to
('\
s) 24 specify crack limits, doesn't that to some extent 25 negate the leak-before-break proposition?
17 s 1 MR. WILKOWSKI: In which way?
2 MR. ETHERINGTON: Well, on a 3 leak-before-break you can presumably just sit tight 4 until the pipe leaks.
5 MR. WILKOWSKI: Yes. But if they have --
6 MR. ETtiERINGTON: Without even worrying 7 about the crack size.
8' MR. WILKOWSKI: Um-hmm. Well, for 9 instance, IWB-3640 type of approach really is 10 ,
designed to insure that you don't even get to a leak, 11! that the crack will be totally safe, because they do 1 21 have significant safety factors on the applied 13 stresses.
14 MR. RODABAUGH: I think Harold in my mind 15 has raised another point. Is your work showing that i
16 I it's impossible to have a break before leak? Are 17 there not conditions where you can have a break 18 before a leak?
19 MR. WILKOWSKI: There are extreme 20 conditions where that occurred, but probably not for 21 conditions that would exist in a plant piping system; 22 that is, experimentally we could artificially create 23 those conditions by making very large flaws, having
"\ 24 extremely large loads relative to the loads that are (J 25 in the plant piping systems.
l 18
,_ 1 MR. BENDER: We have those cases where it
( l
/ 2 occurred in the fossil power steam plants where 3 breaks did occur, and I don't know whether they were 4 preceding leaks or flaws or not.
5 MR. SHEJMON: There were not. Nothing 6l detectable, nothing visible.
I 7' MR. WILKOWSKI: Um-hmm.
8 MR. SHEWMON: They were also under creep 9' conditions.
10 MR. WILKOWSKI: The failure mechanism is 11 so much different.
I 12 MR. BENDER: Are you arguing that that (m) 13 mechanism could not exist here? Is that what you're 14 saying?
15 MR. WILKOWSK1: Nell, I'm not necessarily 16 an expert in the creep fatigue area, but from what 17 I've heard, that's been the strong implication is 18 that it's a mechanism that might not exist. We are 19 evaluating -- one concern is that's a mechanism that 20 they thought would not cause the failure, and we are 21 going through and evaluating different types of 22 failure mechanisms that we see that we did not 23 anticipate and making sure that we understand those
("N 24 mechanisms within our program. But we have not 25 experienced that type of mechanism at all.
19 1 MR. BENDER: Well, I'm not trying to say 2 yea or nay, you know, what I'm trying to be sure of 1 3 is that we have a viable argument if we're going to e
4 say it can't happen here, and I think it has to be i
t j 5 more than "they said that it won't happen here."
6 MR. WILKOWSKI: Yes.
7 MR. BENDER: Somebody is going to have to 8 get that story together.
9 MR. SHEWMON: Let me point out that though 10 it may seem logical once you have leak-before-break 11 to extend it to what lia r o ld suggests --
12 MR. ETHERINGTON: I recognize the two I
) 13 levels are --
14 MR. SHEWMON: --
the NRC has not allowed i
i 15 that yet and, in fact, the ACRS'did not suggest that 16 in their letter under Revision GBC4, although there
. 17 were a couple of dissents by Ward and Shewmon, '
1 18 saying that they thought logically they should apply 19 it to leak-before-break. . So that is still, you ,
i 20 know, we're still in a belts and suspenders approach, 21 which is perhaps the best.
i
-1 22 MR. ETilE RI NGTON : I agree.
23 MR. SII E W M O N : Go ahead.
('\ 24 MR. WILKOWSKI: Okay. Some other points i %)
25 is that we know that your committee has been t
l
_ . . _ . _ _ _ , _ ~ . _ . . . _ _ . . - _ _ _ . , . _ - _ . . - _ _ . - . . . _ , . _ _ . - - _ . . . _ , _ _ .
t i 20 i
j 1 concerned by thermal aging of centrifugally-cast 2 stainless steel and the effects of plugging on leak i 3 rates. These are concerns that-we're trying to 4 incorporate into our work. For instance, in all of l
]
5 the pipe experiments we conduct, we take detailed 6 measurements of the crack openings, so that those l 7 data could be used to assess leakage area prediction
]
8 analyses.
9 Also, we're interacting with Argonne on 10 thermal aging of some centrifugally-cast stainless i 11 steel pipes that we've received, and Bill Shack will 12 talk about that a little bit later on as well.
) 13 There are a number of other related pipe 14 fracture programs. These are more for your benefit i
15 just to let you know that what we're doing in this
- 16 program, we're not just narrowly focusing on what l i 17 Battelle is doing. We've tried to incorporate i
18 experimental results from a number of past programs l l
19 going back to a critical program that Bob Eiber did 20 here at Battelle for the old AEC, which was really 1
21 the basis of what the West Germans are doing in 22 their basis safety program.
i i 23 Some of the oldar EPRI work that we've i
24 done here, for instance, was the foundation for
! 25 what's being used in the ASME IWB-3640 code for i
, . . , ,-,-._...-,-n-- .
, , - -- -,,n,._,,, . , , _ . _ . , , . . , , ~ . ,
, , . . , , , . _ _,,,,,.,,,w,,,mmw_,v,, , _ . , . . , , , . , , , n,
21 1 evaluation of austenitic piping. We've also 2 incorporated a lot of this data into some of the 3 screening criteria and limit-load analysic that 4 we've done and you'll see a little bit later on 5 today.
6 The next view graph here just shows some 7 current programs that are ongoing. Of course, 8 there's the programs at David Taylor. We work with 9 them fairly close in coordinating analysis of their 10 experiments and insuring that there is no 11 duplication of efforts.
12 There are some programs, current programs
) 13 going on at MPA in West Germany, of which we have 14 some of the data for evaluating our criteria, and 15 also we're coordinating our program with some of the 16 Japanese programs on carbon steel piping where they 17 might slightly overlap in their test matrix compared 18 to what we're doing, but they look at points that 19 are a little bit different.
20 One of the questions is going to be after 21 we're all done with this program where are we going 22 to be after three years? And some of these -
1 23 responses here are somewhat speculative in that 24 we're only in the second year of t h'e program. So 25 I've tried to make some assessments of some critical i
l
22 1 aspects of the program after three years for you.
2 First of all, an important point is what 3 type of limitations exist on the 4j net-section-collapse analysis that are used in not i
51 only the ASME code, NUREG-0313 criteria for i
- 6) evaluating flaws in BWR piping, but it's also used i
7l in the Japanese criteria and in the West German 8l basis safety approach. ,
9; We are experimentally documenting when you 10l can use the net-section-collapse analysis and get 11 reasonable agreement. However, that's a simple 12; strength of materials approach. It doesn't account 13j for the toughness of the material. If you get to ,
I 14} large diameter pipe, as you'll see later on, even if 15 it's very tough in its large diameter you could fail 1
l 16l at stresses below that predictive l I l 17l net-section-collapse. So it's not only toughness, ,
I l 18: but it's the size of pipe.
19 And to account for that we've developed a 20 screening criteria. The screening criteria is a l
21 relatively simple approach which says when you can 22 use net-section-collapse, and for instance, in the 23 NUREG-1061 Volume 3 approach, you can either do a 24 detailed elastic-plastic fracture mechanics analysis l 25 or if the stresses are less than one-third of the i
23 l net-section-collapse stresses, then it says you can 7~ )
! i
'# 2' use the net-section-collapse approach. You don't 3 have to do the detailed fracture mechanics. So this 4 criteria is a good way of assessing whether that 5 approach indeed is a good valid approach to use.
6! This will also impact, for instance, the 7 Japanese leak-before-break criteria for stainless 8l steel. They're very interested in some of the large 9 diameter pipe tests that we've conducted because 10 most of their facilities only have capabilities to I
11 test small diameter piping system and what we're 12i seeing is with larger diameter pipe the failure
) 13 loads will be less than that predicted oy the 14 net-section-collapse approach.
15f MR. SHEWHON: What is the Japanese 16 interaction? Is there a formal agreement on this or --
17 MR. WILKOWSKI: They will be joining this 18 International Pipe Integrity Research Group, the
- 19. IPIRG group, which will be discussed tomorrow in the 20 morning. So that would be a formal interaction.
21 We've have a lot of informal interactions with them.
22' We have some analytical round robin: which they l
23 contribute to very significantly. We have many
'}
v 24 visitors from over there all the time. And one of )
25 our staff members is --
Mich Nakagaki is also
.24 1 working with Professor Yagawa over at the University 2 of Tokyo, so we have pretty close contacts there.
3 MR. ETHERINGTON: You mentioned the
( 4 net-section analysis can be used if the stress is 5 less than one-third of the ultimate; is that right?
'i 6 MR. WILKOWSKI: If the applied stress is 7 one-third of what the net-section-collapse. stress 8 would say the failure would be.
- 9 MR. ETHERINGTON: The applied stress, that
- 10 means the local -- no, that would be the design 11 stress.
- 12 MR. WILKOWSKI
- The bending stresses 1
() 13 under --
l 14 MR. ETHERINGTON: The designed membrane 15 stresses, is that it?
16 MR. WILKOWSKI: No, under faulted 17 conditions.
18 MR. ETHERINGTON: Under faulted conditions.
19 MR. WILKOWSKI: So they say if you've got 20 a safety factor of three between the 21 net-section-collapse predicted load for the cracked I
[ 22 pipe versus the applied load, then you don't need a 23 detailed elastic-plastic fracture mechanics analysis.
l 24 That's a significant margin.
25 MR. ETHERINGTON: The faulted stress will i
, .-.,_..,_._.....,.,-,.,,.~.,-___,,.,_,,.,,--.~._..,,.,,.,,_,,,_..,_---_,,_..,,___-._.,-._r ,-,,_,,,,_r,_,m._..
25 1 usually be more than a third, won't it, in ordinary O 2 design?
3 MR. WILKOWSKI: Well, it's relative to the
] 4 net-section-collapse predicted failure stress for l 5 the flaw size that you're looking at, and that flaw 1
6 size is relative to whether you'll detect leakage 7 under normal operating conditions.
', 8 MR. ETHERINGTON: Um-hmm.
9 MR. WILKOWSKI: Relative to the pipe 10 fracture data base, we believe that probably for 11 through-wall cracked pipe that we'll have a very 12 sufficient data base. We'll have evaluated weld
() 13 overlay repairs and the fracture behavior both in 14 small diameter pipe, this is wc k that you'll hear l
I 15 today, versus very large diameter pipe.
j 16 We'll find out that from the surface crack I
17 data that we see that the R/t ratio of the pipe is
- 18 very important, even in the limit-load analysis, 4
19 that is, as you get to thinner wall pipe it tends to 20 flex quite a bit and ovalize, such that the 21 load-carrying capacity decreases.
22 MR. RODABAUGHz Gery, you said just a l l
, 23 minute ago that it was the size effect like a 42 i l
4 24 inch test. -
l 25 MR. WILKOWSKI: Yes, for the through-wall E- .._ _ ____ . _ _ . _ _ _ _ .
26 ;
, 7 1 crack behavior. There is the size effect as well on 2 both.
3 MR. RODABAUGH: Is it size effect or is it 4 an R/t?
5 MR. WILKOWSKI: For the through-wali 6 cracked pipe we don't see an R/t because we've had
- 7. some very high R/t pipe that still is predictable l
l 8 just based on the size effect, whereas for the 9 surface crack case we see both.
10 MR. RODABAUGH: What bothers me is if 11 you're talking about a size effect, then I don't see 12 that any of the ASME formulations are correct
) 13 because they in no ways f.nvolve an absolute size.
14 MR. WILKOWSKI: The recent modific: tions 15 to the IWB-3640 code for low toughness flux 16 austenitic welds does have a size effect in there 17' such that as the diameter increases the allowable 18 flaw sizes decrease. It has a streus magnification 19 factor, which is a direct function of the diameter 20 of the pipe. But it doesn't have it for the wrought 21 stainless steel.
22 MR. RODABAUGH: It's already been 23 published.
24 MR. WILKOWSKI: That has already been 25 published and I think it's been incorporated
27
- 1 formally now.
- ~ 2 MR. BENDER: Could I ask about the weld 3 overlay approach. What is it you're doing?
4 MR. WILKOWSKI: We will have a detailed 5 presentation on that, but very briefly what we're 6 doing is we're putting in flaws -- fatigue cracks 7 entirely through the wall thickness of a pipe and 8 halfway around the circumference, this being an 9 estimate from some of the NRC licensed people to be 10 a worst case condition. And then we're sending that 11 fatigue cracked pipe to Nutech. They are putting an 12- overlay on the pipe, and then we're testing that
) 13 overlaid pipe under pressure and bending to failure 14 and comparing it to the code.
I 15 MR. BENDER: Okay.
16 MR. SHEWMON: Now, there are various 17 amounts of overlay one can put on and depending on 18 the philosophy of whether you take credit for the 19 piping or don't take credit, which are you going to 20 do in that.
21 MR. WILKOWSKI: What we've done is we've 22 put on more of a standard overlay which is two I 23 layers. We wanted to make sure that in the
(~N 24 experiments indeed we could fail the pipe as opposed
\_) ,
25 to putting on a -- I can't think of the name, but 1
_. - .- __ .- = , . - - - , - . - .--
28
,, 1 the initial overlay that was used was -- the 2 thickness of the weld metal was exactly the same 3 thickness of the pipe, and in that type of case we 4 probably just bend the pipe around the overlay no 5 matter what type of flaw was in the pipe. So we 6 tried to carefully design the experiment so that 7 indeed we would get failure, and then we're making 8 an assessment of the failure criteria.
9i MR. SIIEWMON : Okay. -
10 MR. WILKOWSKI: Okay. Back to the surface lli cracked pipe. We find out that both the R/t ratio 12l of the pipe is important as well as the pipe size.
I
(%
() 13 We've developed an ovalization correction to account 14l for the stiffness of the pipe. So far that's only 15 been an ampirical thing. It's probably a function 16l of different crack sizes as well, and there might be some initial evalu tions that need to be done or 17l 18 additional evaluations that need to be done.
19 Ue initially had plans for conducting two 20 thermal-aged pipe tests. Those thermal-aged 21 centrifugal pipe tests will probably not be done.
22 Because the materials are still aging. They're not -
23 going to experience sufficient aging during the ,
f% 24 course of our program, so that's something that will J 25 have to be done at a later time.
-ei, ... ., m 4w - y - ,4m9.+ -
29 1 de do see that we have very little carbon
\" 2 steel weld data. Some of the carbon steel welds are 3 very low in toughness compared to anything else that 4 we've evaluated. So you'll see a little bit about 5 that later on.
I 6 We have not done anything with bimetal 7 welds, for instance, the safe end of a cold-leg, 8 where the cold-leg might be welded onto the pump 9 where there's a bimetal interface there. And, of 10 course, as I mentioned before, we have not d o r.e 11 anything with torsional loads.
I 12 MR. BENDER: Is ovality an issue or is it
('N q_)
i just one of those minor perturbations that you 13!
l 14' encounter?
l 15l MR. WILKOWSKI: I think it's probably a i
16' minor perturbation. The pipes that we've received 17 from service, they've'had tremendous wall thickness 19 variations, tremendous ovality differences, yet we t
19l see that they're falling within the scatter of the 20: data, for instance, from the screening criteria.
21 MR. BENDER: Tremendous implies very large.
22 Is it really tremendous or --
23 MR. WILKOWSKI: Well, in one case, yes, it 24 was very large.
25 MR. BENDER: Like what?
8 30 1 MR. WILKOWSKI: Greg, do you remember on hN'J 2 the stainless steel six inch diameter pipe what the l 3 wall thickness variations were? Was it about 15, 20 t
4fpercent?
I MR. KRAMER: 0.53, 0.63.
5l 6 MR. BENDER: Then it didn't make all that 7 much difference, is that what you're saying?
8 MR. WILKOWSKI: Well, as long as we i
i 9' accounted for it in the analysis it didn't make that 10l much difference.
11' MR. BENDER: How about in the ovality?
12 MR. WILKOWSKI: The ovality? We have f'N x) 13!
taken the measurements, but haven't gone back and '
i 14 tried to correlate whether the initial ovality 15 affected the results.
16 MR. KRAMER: The initial ovalization was 17 not too bad in most of the tests. I mean a couple 18 of the seam welds you see tremendous initial ,
19 ovalization.
20 MR. WILKOWSKI: Probably seamless pipe is 21 probably pretty good in its ovality because it's 22 extruded through a die. But what you'll see is the 23 hole is not matched with the center of the pipe lots
'\ 24 of times.
{J 25 MR. BENDER: Most of the ovality comes
, 31 1 from -- the ovality often comes from weld shrinkage.
2 MR. WILKOWSKI: Yes. .That's w h a^t Mr.
3 Kramer said is that we've seen more of a problem 4 with the case of seam welded pipe.
i 5 MR. RODABAUGH: Gery, this is a list of 6 things that might be needed. Getting back to Mike's i 7 comment, how about elbows, branch connections --
8 MR. WILKOWSKI: That's the second page.
1 j 9) MR. RODABAUGH: That's the next page.
l 10 Okay. Go backwards.
j 11 MR. WILKOWSKI: I wrote this page for you,
! 12 Everett.
) 13 MR. RODABAUGH: Okay. Good.
14 MR. WILKOWSKI: I thought it was on here.
15 But you're right, we haven't done things as far as j 16 looking at what happens when the crack is close to 1
j 17 the nozzle. I was intending to include that because
! 18 those are additional effects that could affect the i
I 19 results.
i I
- 20 As far as the material property data base,
! 21 this is something that we're coordinating with the I l l 22 Materials Engineering Associates, who is developing l
- 1 1 l j 23 a computerized data base. We've encountered an
}
l 24 unusual behavior which we think is attributed to i
25 dynamic strain-aging. We see that there are effects 1
I i 32 I
- 1 of the weld size relative to the specimen size on
- 2 the J-resistance cit r v e , the fracture toughness l 3 evaluations of the material. And we do see that i
! 4 there is for low toughness welds, which could be an i
5 austenitic or carbon steel, the type of variability j
! 1
- 1 6 that you might experience could be significant in i 7 evaluating the load-carrying capacity of the pipe.
8 For elastic-plastic fracture mechanics, we
) 9 generally see that the General Electric /EPRI l 10 analysis for through-wall cracks is conservative 11 compared to the experimental data. The method 12 developed by Paris is not general enough in that it
,( j 13 does not incorporate the strain hardening 14 characteristics of the material. The method i
) 15 developed by --
initially developed by Ray Klecker 1
16 within Licensing appears to be the most accurate of 17 the analyses that we've seen for through-wall cracks.
18 We are currently developing a finite
. 19 length surface crack analysis. This is a very I
}
20 difficult problem, and we've done a few comparisons 21 of experimental data to this analysis, so this is 22 ongoing work.
23 For combined pressure and bending, this 4
l 24 work is currently underway, so we'll have an
- 25 evaluation of that type of loading conditions on the i l
1
33
_ 1 estimation schemes. We'll have looked at the ASME
\ t
\ 2 Section XI analysis and have verified that approach.
3 And we're also looking at methods to 4 extract J-resistance curves, the fracture toughness 5; curves for large amounts of crack growth. In many 6 cases what you have is small specimen machined from 7 pipes, because of the curvature of the pipes you 8l need to make predictions for large amounts of crack 9 growth. So that's a difficult problem to evaluate.
10 Finally, in the elastic-plastic fracture 1 11 mechanics we'll be looking at predicting loads and l
12l load versus displacement relationships. We see that 13 when we try to predict displacements as well as I
l 14l loads, it's a much more difficult problem. We find 15 that --
we think at this point we'll probably have 16 an analysis that for through-wall cracked pipe will 17 be sufficient for engineering applications. We'll 18 have developed an instability analysis for a surface 19, cracked pipe. That is, if a surface crack exists in 20 a piping system and you put it under a load, can you 21 predict if there is an instability how far that 22 crack would jump and what would be the resulting 23 leakage area from that.
f\ 24 Finally, looking at the complex crack LJ 25 analysis, we find that using existing analyses that
I i 34
,, I we need some sort of empirical correction to the O 2 analysis based on experimental data.
3 And then we've looked at the case of 1
i l 4 evaluating welds in pipes, that is, most of the i
! 5 analysis assumed that it's a homogeneous material.
i 6 That it all has the same strength and hardening l
7 characteristics and how you account for the a
8 different strain hardening between the weld and the t
l 9 pipe and still get accurate predictions.
I i
10 MR. BENDER: Can I go back to the question 11 about extrapolating the J-R curves. What is the 1
l 12 application where we need to do that?
a i
() 13 MR. WILKOWSKI: When you're doing a l 14 leak-before-break analysis of the idealized i
l 15 through-wall crack, a simple through-wall crack. In 16 some cases if the loads are high enough, you might i 17 get crack initiation and some crack growth under a j, 18 faulted condition, whereas under normal conditions i 19 you might just have some leakage. What you want to l
20 do is insure that if you do have the crack growth j i l
! 21 that you can reasonably predict it, and to do so you !
! i 1
22 need a resistance curve for the material that I i
4 j 23 accounts for the crack growth. The crack growth I
24 from initiation to maximum load, for instance, can j 25 be very significant.
l i
l
35 1 MR. BENDER: Have we had any cases where fm ~
L') 2 that's had to be done or is it just one of those 3 things that we want to cover the waterfront?
4 MR. WILKOWSKI: Yes, there are some cases 5 where Licensing has had t. 3o beyond crack 6 initiation to make an eva_aation of 7 leak-before-break behavior, that crack initiation 81 alone wasn't enough to insure it.
I l
9' MR. BENDER: Okay. ;
1 1
10 MR. SHEWMON: Has anything of that sort 11 arisen out of this Beaver Valley question yet or --
12 there's a WHIPJET project there, as you know, to see
) 13 what they can do in the secondary.
14 MR. WILKOWSKI: There was some talk of 15 trying to do a fair amount of analysis for that.
16 The program at this time I think is questionable as 17 to whether it's going to continue onwards for 18 evaluating that type of technology.
19 MR. BENDER: S. large crack would be how 20 much, what fraction of the --
I guess it's the 21 length of crack you're talking about.
22 MR. WILKOWSKI: Yes. Maybe 20 percent of 23 the circumference would be a very large crack.
g 24 MR. SHEWMON: What's a BCD surface crack?
L 25 What's a BCD? Right in the middle.
)
36 s 1 MR. WILKOWSKI: Oh, this is Battelle
' 2 Columbus Divisions.
3 (Laughter) 4 MR. WILKOWSKI: It's a modifier to the 5 analysis.
1 6 MR. S H E W M O 11 : Good.
7 MR. WILKOWSKI: Okay. Do we have any 8 other questions?
i 9! If not, I would like to introduce the next 10 speaker. This is Paul Scott. Paul will talk a 11 little bit about the pipe experimentt that are 12t conducted. He'll show you a couple of view graphs
) 13 of some of the pipe fracture tests and also discuss 14 how these results compare to the 15 net-section-collapse analysis and simple screening 16 criteria to predict when the loads if the 17 net-section-collapae does not work.
i 18 MR. SCOTT: My name is Paul Scott. I'm !
19 from Battelle Columbus. Before we start looking at 20 some individual specific results and some 21 experimental efforts, Gery thought it would be wise 22 to give you a brief overview of the experimental 23 effort in general.
, 24 As part of the Degraded Piping Program in i 25 the first three years there's a total of 61
.. - - - - _ - - . - . _ _ ~ --. - - - - - - - -
37 gs 1 experiments planned. Of those 61, approximately 45 2 have been conducted to date. There's three 3 different flaw geometries we're looking at, 4- through-wall flaws, surface flaws and complex cracks.
5 Three basic types of materials, austenitics, carbon 6l steels and welds. Within the austenitics we've got 7 predominantly stainless, although there have been a 8 few tests on some Inconel.
9 MR. ET tlE RI N G TO N : I recognize you can't do 10: everything, but why did you exclude the low alloy 11 steels?
12 MR. WILKOWSKI: Which low alloy steels in
) 13 particular?
14 MR. ETHERINGTON: Well, the SA533 Grade B.
15 MR. SCOTT: 533 is included.
16 MR. SHEWMON: That's called a carbon steel.
17 MR. ETHERINGTON: Oh, that's called a 18 carbon steel. Okay.
19 dR. SHEdMON: That's BCD nomenclature.
20 (Laughter) 21 MR. SCOTT: Within the through-wall cracks .
l 22 we tested pipes from four to 42 inches. The surface 23 cracks and the complex, we've concentrated on pipes 24 in the 6 to 16 inch range.
25 MR. HUTCHINSON: Can you describe very
38 1 briefly for me again how you determine the complex 2 crack?
3 MR. SCOTT: Okay. The next view graph 4 kind of shows you a picture. A picture is worth a 5 thousand words. The complex crack would be a long 6 internal surface crack that's broken through the 7 wall in a local area, similar to the Duane Arnold 8 crack. v 1
9l In addition to the three materials, the 10 three loading technicians and the different pipe 11 diameters, we're testing four different loading 12 conditions; pure bending, pure pressure, combined
) 13 pressure and bending, and then compliant bending.
14 In addition to the 61 experiments that 15 we're conducting as part of the Degraded Piping 16 Program, we're also evaluating different experiments 17 conducted at different programs. We're looking at 18 data from two programs conducted at David Taylor, 19 ene on A106 Grade d carbon steel pipe and one on 20 stainless steel TIG welds.
21 We're looking at past data from a Battelle 22 program t1 2 we conducted for EPRI on 304 stainicas.
23 we're looking at some carbon steel data generated by 24 a gentleman named Reynolds on some exial membrane 25 stress tests, some of Bob Eiber's data for the AEC,
39 1 and we're looking at a little bit of data generated 2 by Jim Joyce on aluminum pipe. Here we're just 3 comparing the screening criteria.
4 Within the Degraded Piping Program and 5 also within the other programs, predominantly the 6 pipe diameters have been in the small four to ten 7 inch range, and in the Degraded Piping Program we've l
Bj also started to look at some of the larger diameters i
9' pipes.
10 MR. SHSWMON: Would you tell me what a l
11l compliant bend is?
12 MR. SCOTT: Compliant bending where we are i
7-( 13 trying to simulate the excess compliance that you 14 might have from a long piping system, we'll put a 15 series of springs in with the load train to simulate 16 the elastic energy that's stored in the piping 17 system.
18 MR. SHEWHON: Okay.
19 MR. ETHERINGTON: Would you elaborate a 20 little further? Compliant is what? Force per unit --
21a MR. SCOTT The compliance would be the 1
22 displacement for unit force.
23 MR. ETHERINGTON: The other way. Okay.
)
'N 24 MR. SCOTT: Yes. This is a view graph of '
(\.J 25 the first pipe fracture experiment we conducted as i l
)
l
40
, 1 far as the Degraded Piping Program. This is a four NJ 2 inch SA333 Grade 6 carbon steel pipe that had a 3 through-wall crack 36 percent of the pipe 4 circumference in length. When we initially did the 5 test we were quite surprised that we saw that the 6 fracture grew out of the circumferential plane.
I 7! Okay. When we ran the C(T) specimen test, the 8 initial notch was in the same orientation as the 9 through-wall crack, we saw that it also gret out of 10 the plane of the circumferential --
out of the 11 circumferential plane.
12I subsequent to this, we ran some fracture 4
) 13 tests for some C(T) specimens from this material in 14 different orientations, and we found that indeed the 15 load toughness direction was basically in this 1
16' orientation.
17 11 R . SHEWMON: Load toughness. You're 18 suggesting that is an anisotropy .n the 19 elastic-plastic properties of the piping or the 20 materials?
21 !!R . SCOTT: Of the materials, yes.
22 MR. S H E W!i O N : Do you have any idea why it 23 comes in 45 degrees or how you could predict it?
24 MR. SCOTT: Well, we think it's due to the 25 the way that the material is manufactured in -- the I
41 1 seamless pipe the way it's manufactured, that
"# 2 there's a cold working taking place in that 3 direction.
4 MR. SHEWMON: But this is annealed pipe.
5 It's some memory that survives --
at least I assume 4
6 it's annealed pipe.
7 MR. WILKOWSKI: No, not necessarily.
8 There's two processes that occur in the fabrication 9 of seamless pipe. It's typically the hot forming 10l process will also cause a helical flow of the 11 material, okay, in which the inclusions then are 12; built into forming in that direction, and that's the
) 13 major thing that affects the toughness going in the 14 helical direction.
15' Secondly, after the pipe has been 16l completely fabricated, there's a final straightening 17 process. The straightening of the process also 18; involves the cross rolling in the helical pattern as i
19 well.
20 MR. ETHERINGTON: Is the Mannesmann 21 process still used for making pipe?
22 MR. WILKOWSKI: Yes, quite largely, and 23 it's mainly in the Mannesmann process that this 24 occurs.
25 MR. SHEWMON: This was one of the carbon
42 =
1 steels. 1 2 MR. WILKOWSKI: Yes, that's right.
! I 3l MR. SHEWMON: Do you know what the sulfur 1 I j 4 was?
5 MR. WILKOWSKI: Not off the top of my head, 6, but we have the data. f I
7l MR. SHEWMON: I would be interested in 1
l 8l learning.
9 MR. HUTCHINSON: So do you look at the 10 microstructure?
11 MR. WILKOWSK1: Yes, we do a chem analysis t
well as tensile properties. In this particular 12fas
) 13 case since it was the first time that we saw a crack 14 growth at significant angle, we did a fair amount of 15 investigation to look at the inclusions relative to 16 the orientation in the pipe.
17l MR. SCOTT: This is a photograph showing 18 the 28 inch through-wall crack carbon steel 19 experiment that we ran. Basically, this is just 20 showing our experimental set-up. Basically these 21 are the rams and the wire ropes that used to 22 restrain the applied load to put the pipe section in 23 between the rams in the four point bending.
24 This test was interesting in two facts.
25 In one we again saw that the fracture grew out of
43 1 the circumferential plane, much like in the four 2 inch test, but also we saw scveral minor
! 3 instabilities during this test, in which there would i
i 4 be an instantaneous drop in load, and associated f
5 with that instantaneous drop in load there would be 6 rapid crack growth.
l i
- 7 In some of these load drops we saw crack 1
! 8 growth as much as an inch and five-eighths at both
! 9 crack tips. This similar behavior was seen in the I
l 10 C(T) specimens for this material, and it's been i 11 postulated that this might be due to a dynamic 12 strain-aging effect. Mark Landow will speak a
() 13 little bit more to this when he talks.
, 14 In this presentation we'll look at the 15 l limit-load criteria, the net-section-collapse 16 analysis, and look at the validation of that 17 criteria by experimental data. For the through-wall 18 cracks we'll look at the effect of pipe diameter, j 19' and for the surface crack we'll look at the effect i l 1
i 20 of the R/t ratio. As Gery mentioned several times,
- i
! 21 we'll also look at the plastic zone screening 22 criteria which was developed to assess when this a 23 net-section-collapse analysis is appropriate. We'll r 1
24 look at the analysis both for the through-wall crack i
f 25 and for the surface crack.
i
! 44 1 The significance of this work is that in i
2 NUREG 0313 and in the IWB-3640 austenitic piping 3 criteria, that this net-section-collapse criteria is
, 4 embodied in these criteria. We have seen that for i
5 large diameter wrought stainless steel pipe that the i
! 6 net-section-collapse analysis may not be appropriate.
7 This is a view graph of the maximun 8 experimental stress divided by the j
9 net-section-collapse stress as a function of pipe J
10 diameter. This is for all of the experiments
- 11 conducted as part of this program as well as some of i .
I 12 I the other programs we've analyzed. It has the 1
() 13 through-wall crack, surface crack under various j 14 loading conditions, and also the complex crack data l 15 for the different materials. In this particular 16 case the flow stress has been defined as 3 S of M 17 from the code.
i 18 Basically, there's a general trend that 19 you see that as the pipe diameter increases, that 20 the maximum --
the net-section-collapse stress tends
]
21 to overpredict the maximum experimental stress.
. 22 MR. RODABAUGH: Paul, according to your 1
i 23 table on your progress report of 4-27 the largest 24 pipe has a diameter to thickness ratio of either 150 l 25 or 67. So I'm still curious, how do you know that's i
45
.g 1 a diameter effect and not a radius to thickness
'"J 2 ratio effect?
3 MR. WILKOWSKI: You can show that on the 4 screen.
5 MR. RODAbAUGH: I know I have to keep 6 harping on this because I think in the long run this 7 will influence how people evaluate piping, and I 8 would like to make sure that it is a diameter effect.
9 In fact, it might even be a test set-up effect or a 10 material effect. But do you have that plotted 11 against radius to thickness ratio?
12 MR. SCOTT: We do, but only for the s) 13 surface cracks, and those large R/t ratios you're 14 talking about are for through-wall cracks.
15 MR. RODABAUGH: Well, this table --
it 16 seems to me I could claim it's an R/t effect as well 17 as a diameter effect, and I was looking for the 18 other graph.
l t) MR. WILKOWSKI: That's one reason why 20 we're conducting some additional experiments with 21 lower R/t ratica. We just conducted a few in which 22 not all of them are presented within the program 23 results today.
f'h
~%)
24 MR. RODABAUGH: This would be big pipe.
25 MR. WILKOWSKI: Big pipe.
46
.s 1 MR. RODABAUGH: But also very thick.
2 MR. WILFOWSKI: Also heavy wall, 36 3 diameter, three inch thick pipe.
4 MR. RODABAUGH: Okay. Thanks.
5 MR. SCOTT: This is a similar view graph, 6 except the only difference between this view graph 7 and the last view graph is that the flow stress here 8 has been defined as 1.15 times the average of the l
l
\
9l yield in the ultimate strength of the material. I i
10 l1 This is a graph exclusively --
just for i
the through-wall crack experiments. Basically the --
11l l
12' it's again -- the maximum experimental load to the
) 13 predicted net-section-collapse load is a function of 14 the mean type of pipe diameter.
15 Again, you see as the pipe diameter 16 increases, the net-section-collapse stress tends to 17 overpredict the maximum experimental stress. These 18 three particular experiments down here are 19 relatively small diameter tests that also do not 20 reach the net-section-collapse load. These 21 particular experiments are for some very low 22 toughness or relatively low toughness materials.
23 The two diamonds are for some submerged arc weld 24 tests where the JIc would maybe be 600 inch-pounds 25 per inch square. This square is for carbon steel
47 1 where the JIc was approximately 1200 inch-pounds per
(
2 inch square.
3 MR. RODABAUGH: .J e l l , you're I think 4 saying the same thing that I'm trying to say, that 5! those three points, the two diamonds and the square, 6l that are at low diameters have some other reason for 7 beitig down low.
8 MR. SCOTT: Right. Okay. And we looked 9 at that --
1 10l MR. RODABAUGH: You haven't convinced me 11 that it's a diameter effect.
12 MR. SCOTT: Okay. Embodied in the D)
( 13 net-section-collapse analysis is the assumption that 14 the remaining ligament reaches fully plastic 15 conditions. Okay. In order to assess when this -
16 assumption is satisfied, we developed the plastic 17 zones screening character criteria. Essentially 18 what we did is compared the plastic zone size to the 19 remaining tensile ligaments from the distance from 20 the initial crack tip to the neutral bending axis.
21 Okay. If this plastic zone is greater 22 than the tensile, remaining tensile ligament, then l
1 23 the dimensionless parameter is obviously greater 24 than one, and we see that net-section-collapse
(~1) x_
25 conditions should be satisfied, and we see that the l
48 1 experimental load of crack initiatior. is greater b
' 2 than the predicted limit-load.
3 Notice, we're using the experimental load 4 of crack initiation, and the reason for that is the 5 J we are using is the J from a C(T) specimen at 6 crack initiation. Okay.
7 We also drew a trend curve in, which would 8 kind of show the data based on maximum load. We see 9' that when this dimensionless parameter is less than 10 one, you have contained plasticity, the fully 11 plastic conditions are not being satisfied, and we 12 see that this experimental load of crack initiation t
() 13 can be significantly less than the predicted 14 net-section-collapse load. In fact, for this one 15 stainless steel submerged arc weld, we only reached --
16 the load at crack initiation was only approximately 17 30 percent of tne predicted net-section-collapse 18 stress.
19 Now, this is the surface crack data, and 20 here we are plotting R/2 ratio. Okay. We are 21 plotting the maximum experimental load to the 22 predicted net-section-collapse load as a function of 23 the R/t ratio. The line here is a least squares 24 fifth through the solid data points. Now, why we 25 chose the solid data points and disregarded tne open
49 g3 1 data points is that the solid data points had a ts, 2 plastic zone parameter greater than one. In other 3 words, these points we are reaching -- we predict 4 that we are reaching fully plastic conditions and 5 that we're not being influenced by the toughness of 6 the material.
7 Based on this line we generated an 8 empirical correction term that accounted for the R/t i
9 ratio, such that at large R/t ratio where the pipe 10 becomes more flexible, we would expect that you 11 would reach -- you might not reach the 12 net-section-collapse conditions.
13 MR. SHEWMON: Before you leave that, tell 14 me how the industry should protect itself from the 15 curves that didn't develop a fully plastic condition, 16 that is, some of these points lie appreciably below 17 that, and you say, well, I don't need to take 18 account -- I'm sorry. We don't take account of that 19 with this model because it didn't go fully plastic, 20 yet those are welds which are out in the field. Is 21 this then to be a specification on how the welds are 22 made or what?
23 MR. SCOTT: Well, in the next -- well, I 24 account for them. What I'm doing is I'm just --
I'm 25 not accounting for them in developing this empirical
50 1 correction factor, okay? And the reason is because 7_
(')
2 they're being influenced by probably their low 3 toughness.
4 MR. S il E W M O N : I understand that.
l 5 MR. SCOTT: But when I go to the next 6 slide -- l 7 MR. HUTCHINSON: Before you do, I didn't 8 understand what you suggested was causing the line l
9 to have a negative slope. Is that the ovalization 10 effect?
11 MR. SCOTT: Yes. As you go to a large R/t 12 ratio, the pipe becomes more flexible and it begins
, /'N
() 13 to ovalize on you.
14 MR. HUTCHINSON: So in other words, what 15 you're referring to as the net-section-collapse load 16 does not account for that?
17 MR. SCOTT: Right. Okay. Back to your 18 question --
19, MR. SHEWMON: What you have is fine for
{
20 models, but what I'm interested in is how do we 21 protect ourselves from welds that aren't very tough?
22 MR. SCOTT: Okay. Maybe --
23 MR. WILKOWSKI: I think he could ansner 24 that with the next few view graphs.
{,.
25 MR. SHEWMON: Fine.
51
,s 1 tiR . SCOTT: Yeah. Here what we did is we
(#l
- 2 said, all right, now we have our empirical 3 correction term for our R/t ratio. Okay. What we 4 did --
since some of these tests were combined 5 pressure and bend tests, okay? We plotted not only 6 the stress at crack initiation but also the axial 7 membrane stress divided by this correction tera 8 times the net-section-collapse stress, plus the 9 axial membrane stress. Okay? And plotted that 10! ratio as a function of this dimensionless parameter, 11land again this is a similar dimensionless parameter l 12l that we saw in the previous view graph.
(n) 13 This again is just the plastic zone size, 14 3 whereas this is the distance from the center of the 15 crack to the neutral axis. Okay.
16 Now, as we get to large R/t ratios, then 17 this number would become small and so we would 18' essentially artificially decrease the 19 net-section-collapse stress. Okay? Does that 20 answer your question?
21 t1R . SHEWMON: No.
22 MR. SCOTT: Okay.
23 MR. SHEWMON: You didn't say anything f'\ 24 about specifications which is what I was really
'. )
25 looking for, but that's not your business. So what
52 1 I'm asking is in a sense irrelevant to your talk.
b V 2 What as I understand you've said is that with the 3 analysis you have you can fit both data points where 4 the joint goes fully plastic and where it does not 5 go fully plastic.
6 MR. SCOTT: Right. And we also in the 7 surface crack case also take into account the 8 ovalization that might occur due to the extreme R/t 9 ratios.
10 MR. SHEWMON: Okay. Thank you.
11 MR. SCOTT: Okay.
12 MR. RODABAUGH: Paul, if I could -- I
) 13 think there's a better answer in the IWB-3640, is 14 there not? This point that you were pointing at 15 that was done at a ratio of .3 with respect to the 16 l>ordinary net-section-collapac.
17 MR. SCOTT: Yes.
18; dR. RODABAUGH: Because I think it's a 19 submerged arc weld or something like that.
20 MR. SCOTT: Hight.
21 MR. RODABAUGH: Is it not true that in the 22 present IUB-3640 there are penalty factors on 23 submerged arc weld?
24 HR. SCOTT: Yes, just based on the 25 diameter, okay? That particular diameter was like
53 1 1. --
it was about a 16 inch diameter. And I'm not (s s
)
2 exactly sure what the stress multiplier that would l
3 result for that, but I imagine to be about 1.35.
4 MR. WILKOWSKI: 1.5. Essentially what the 5 IW3-3640 analysis procedure does is it says for 6 certain welds that have a typical toughness, that
- 7. they will derate the load-carrying capacity as a 8 function of the pipe diameter. So what they're 9 doing is comething that'c very similar to what you 10 could show from this work, in that here we've got a 11 certain tcughness, the J, and we've got a pipe size 12 accounted for. And both of those parameters are
(' accounted for in the IWB-3640 flux weld criteria.
13ll 14 We've also compared all of our flux weld 15 data to that criteria and fuund ont that that 16i criteria predicted that the loads would be below 17 what the experimental load was, so it was a safe 18 criteria to use.
19 MR. SHEWMON: All right. Okay. Thank you.
1 20 MR. SCOTT: Yes. We also tested some --
21 in addition to the 6 and the 16, we also tested a 22 section of a weld removed from service from Nine 23 Mile Point, and I'm not sure if that data is on 24 these figures or not.
25 MR. WILKOWSKI: That's a through-wall
54 1 crack.
O
\' 2 MR. SCOTT: Yes, a through-wall crack.
3 This is a view graph for through-wall 4 cracks showing the possible significance of the 5 screening criteria. What we did here is we assumed 6l a 20 percent through-wall crack in a pipe, and then 7 took typical values of J, the flow stress for 8 different materials for a typical system, 9 considering a BWR recirculation line made out of 316 10 stainless steel. In this case typical value of J 11 may be 4,000 inch-pounds per inch square, flow 12 stress a little over 60 ksi.
l (~% s ) 13 For a 28 inch diameter pipe with this 20 14 percent through-wall crack, we would calculate the 15 plastic zone size t. o remaining ligament ratio to be i
16! about .29. Going back to our scrcening criteria 17 j based on our trend curve for the maximum load to the I
I 18 net-section-colltpse load, we would see that the 19 maximum stress to the not-section-collapse stress 20 would only be about 71 percent.
21 Go based on this you can see that ~ there's 22 a possibility even for the --
based on the screening 23 criteria we predict that even for the large --
even 24 for the stainless steel, the wrought stainless steel, 25 high toughness stainless steel, that we may not
55 7_
1 reach net-section-collapse conditions. Okay.
2 This is a similar view graph, except in 3 this case we are looking at the significance of the 4 surface crack screening criteria.
5 MR. SHEWMON: That first --
second line of 6l data says we don't have to wait for the stuff to age, I
7 it's bad enough when we put it in service almost; is 8 that right?
9 MR. SCOTT: Down here for the unaged?
10 MR. SHZWMON: No, up at the top where you 11 go down by an order of magnitude between tne base 12 metal and its weld.
) 13 tiR . WILKOWSKI: Tha weld data, Paul.
14 MR. SCOTT: Oh, okay. Yeah, but this has -+
15 this is an empirical --
this does not include the 1
16: credit that you would rcccive or -- not that you 17 would receive, but the credit that would be taken 18 away through the load-carrying capacity due to the o
19 flux weld criteria that's now embodied in IWB-3640.
20 MR. SHEWMON: No, I'm just looking over --
21 you look at JIc and draw the same conclusion. You 22 still got an order of magnitude.
23 MR. WILKOWSKI Yes, indeed the toughness 24 is lower, and the ASME code has accounted for that N
25 in their flux weld criteria.
56 1 MR. SCOTT: Right.
('s 1
2 MR. SHEWMON: I wish they could improve 3 the material instead of accounting for it.
4 MR. RODABAUGH I would say the cast 5 stainless was up there around 1500; is that right?
6 They ought to be able to get as good as :ast 7 stainless. Is that bottom line --
yeah, that's cast.
8 MR. SCOTT: Yes, this would be cast 9' stainless. This would be unaged and this would be 10 aged, lli MR. RODABAUGH: Thank you.
12 MR. SHACK: Some of the aged cast
) 13 stainless might well get worse than that. I don't 14 think there's any effort to claim that that's the 15 worst case cast stainless.
16l MR. SCOTT: These were jus'. typical values.
17 okaf. Quickly, as far as the significance of the I
i8: screening criteria, basically within NUREG-0313 and 19, the ASME Section XI that embodied in them, those 20 documents, are the fact that net-section-collapse 21 thould ~oe sufficient for austenitic high toughness 22 materials, although we're seeing for large diameter 1
23 this may not be the case.
24 Also of note is in this West German --
25 they claim that if you can get the toughness of the
l l
1 57 l
, 1 material to be -- the Charpy Plateau Energy be ;
1 2 greater than 35 foot-pounds or 50 Joules that the
- 3 net-section-collapse should work irregardless of 4 pipe size or strength.
- 5 The basis for that argument is this data,
{ 6 but their data is all taken from specimens, C(T) 7 specimens or center crack panels or whatever, and 8 small diameter components. When you plot their data i
l 9 you get the 50 foot-pounds, and then you do reach 10 net-section-collapse conditions, but if you plot our 11,l through-wall crack data and some of our larger i
12 diameter surface crack data, then we see that their
() 13 criteria leaves a little bit to be desired.
- 14. Any questions?
15 MR. RODABAUGH The extra points you added 16 are wrought --
l 17 MR. SCOTT: Some of these are wrought.
18 MR. RODABAUGH: Open coin is pipe tests?
l 19' MR. SCOTT: I'm not sure which are which.
i l
20 I believe most of this data here is wrought, I think i
21 because we just assumed --
22 MR. WILKOWSKI: That would be stainless 23 steel.
24 Ma. SCOTT: Yes, this is stainlesu steel 25 hecause we assumed a Charpy Energy of about 210
53 1 foot-pounds for the stainless steel.
O 2 MR. RODABAUGH But the base material, not 3 the --
4 MR. SCOTT: This is base material.
5 MR. RODABAUGH: The test is on the base 6 material, right.
7 MR. SCOTT Yes.
8 tiR . KERR: This is an irrelevant comment.
9 North of the tiason-Dixon line there is no such word 10 as irregardless.
11 MR. SCOTT: Okay. I'm from West Virginia, 12i and that's south of the Mason-Dixon line.
13 (Laughter) 14 MR. HUTCHINSON: Are we going to hear 15 anything later about attempts to calculate the 16 effect of the ovalization on the methods? Is 17 anybody going to comment?
18 MR. WILKOWSKI: Yes. That's something 19 that we haven't done within the scope of the program 20 so far. Are there any further questions for Mr.
21 Scott?
22 MR. S il E W H O N : tio . I was just wondering 23 whether now would be a nice time to take a break or 24 after the next speaker.
(
25 MR. WILKOWSKI: After the next speaker is t
4 59 i 1 what we had planned if that's appropriate with you.
2 MR. S II E W M O N : Am I privy to your schedule i
3 too?
4 4 4 .IR . WILKOWSKI:
No, you're the ruler.
5 MR. S il E W M O N : Okay. Go ahead.
l
] 6 MR. LANDOW: My name is Mark Landow. I'm f 7 with Battelle Columbus. And this morning I'm going 1
l 8 to talk about some of the significant mechanical
, 9 behavior at light water reactor operating conditions, i
) 10 materials there. This task of the Degraded Piping i
! ll i Program deals with characterizing the materials used
}
! 12; in the pipe fracture tests.
() 13 Three areas I'm going to discuss this 14 morning are typical tests which are used to 15 characterize the pipe materials, the mechanical f 16 properties of weld metal in nuclear piping materialu, 17 and observed instabilities of carbon steel piping i
i 18 materials tested at 550 fahrenheit.
[ 19 In characterizing piping materials, first t
j 20 we do the chemical analyais to make sure that the l 21 material is within the specification, and then we 22 machine a number of mechanical property tests, 23 tensile tests, Charpy tests out of the ferritic 1 i 24 materials and various fracture toughness tests, the 25 compact type, the bend type and the full width facu l l
i 4
1
i 60 1 notch type.
\- 2 The crack orientations in this view graph 3 also indicate the crack orientations used in the
.4 piping tests. The crack to toughness test is used
! 5 to measure the crack growth resistance of the f I i
6 material. And as Paul Scott mentioned, I brought a l 7 specimen that indeed the orientation can have an j
- 8 effect in fracture toughness test specimens. This 1 i i
I 9' is a plate that we tested without side grooves in, I 10 and the crack bent essentially at 90 degrees to the 1
l 11 plane of the crack.
I I
. 12' What we do --
again, about all we can
() 13 calculate off from this is an initiation value. So 14 we can't get any crack growth resistance curves. Go i
- 15 what we do is sidegroove the specimen slightly to 4
f 16 force the crack Lo grow in the plane of the -- tne 1
! 17 crack plane.
l Id ;1d . Sil L W M O N : Does that migration reflect 1
f 19 considerable anistropy in the properties?
l 20 MR. L A tJ D O W s Apparently it does, and in 4
i 21 one series of tests that Paul had mentioned, we did 22 machine specimens at different orientations in the ;
, I l 23 pipe, and when we found the weakest direction, l
) 1 4
- 24 direction of weakeut toughness, the crack did run l 25 straight across the plane of the crack without side t
i l
l l
_ . . _ ~ -- . _ .. .- - . -.
l 4 61 l 1
1 grooves.
l 2 t1R . BLUDER: I would like to go back to a i
)
3 question Dr. Shewmon asked earlier. How does this 4 test work relate to what's being done for deaver 5 Valley? Does anybody know?
6 MR. WILKOWSKI I don't know all of t'.
- details of what's being done for Beaver Valley otner 7l 8 than one of the things that they're interested is in i
9 evaluating the removal of pipe restraints for some 1
10 of the carbon steels. In the carbon steel piping 11 cases there is concern over the toughness of the 12 material both in the base metal as well as in the 1
) 13 weld metal. And some of the results that we're 14 showing shows that the toughness of the carbon steel 15 welds is considerably lower, so that you would have 16 to take that into account.
j
! 17 MR. BENDER: I'm not surprised at that.
18 What I'm trying to understand is if I look at the i
) 19 Beaver Valley test program, can I judge anything 20 about its adequacy from what you're telling we about i
l 21 this one, because that's really what I would like to 22 know. You don't have to answer the question now, 23 because I suspect I wouldn't get an answer that i
i 24 would have been thought about. But, in fact, if 1 l 25 could use it for that purpose, it would be serving i 1 l
l
62 1 its most value to me.
O 2 MR. WILKOWSKI: Certainly, I think if you 3 knew a lot of the details of our program you could 4 answer positively to that.
5 MR. BENDER: I would like to know if the 6 Beaver Valley program is any good or not. I don't
- 7) really care whether yours is any good unless it will 8 tell me that.
9 (Laughter) 10 MR. SHEWMON: You talk about a Beaver 11 Valley test program. I would guess they're doing 12l analyses.
() 13 AR. BENDUR They're doing some testing to 14 get information, and I'm not sure what it's supposed ,
15 to tell me. They've selected a group of different 16[ size pipes, and they're putting -- not specimens of 17 this sort, and I'm not sure what they're doing.
18 MR. ETHERINGTON: What is the material?
19 MR. BENDER: It's AlO6 pipe I think, some 20 of it is, but I'm not really sure what they're doing.
21 MR. WILKOWSKI: Yes. There were plans to 22 conduct a number of fracture toughness tests as well 23 au some pipe testu, and the implication from our 24 work is that you should look at the low toughnesa 25 weldu, and indeed they were planning to do some 1
63 1 evaluations of their weld procedures.
O 2 MR. BENDER: Well, you understand the 3 question I'm asking. It doesn't have to do with 4I whether -- I'm not trying to judge their test 5 program right now, but somebody needs to --
6 MR. WILKOWSKI: Um-hmm.
7 MR. LANDOW Okay. The fracture mechanics r
8' tests, they're conducted at temperatures and under 9, conditions that will simulate the pipe tests, l
10{ temperature of 550 fahrenheit loading rate which lli bring the maximum load at about the same time that 12; we would axpect it in the piping test, and we also
() 13 i
machine them with notch acuities that are similat to 14l the piping teste as it is a sharp machine notch.
15 vie also test some with fatigue cracks to 16i compare to --
tor comparison with other material 17 behavior, with other materials.
ld Next, I'll talk about some of the effecta 19 of weld metal properties on the fracture behavior.
20 There's increasing interest in weld metal proporties 21 in nuclear piping because welds are extensively used.
22 When stress corrosion cracking occurs, they're 23 usually asucciated in areas and vicinities of welds, 24 and there's growing evidence that weld metals 25 duposited by flux rae t h o d s have poorer fracture
_ _ . ~
I 64 1 resistance than do base metals.
2 We have tested several welds in the 3 Degraded Piping Program, submerged arc welds from a 4 Type 304 stainless, as deposited and then annealed; 5 Tungsten inert gas weld in Type 304 austenitic 6 steels; and some submerged arc welds in SA-516 Grade 7 70 carbon steel.
First, we'll look at the tensile 8f 1
9l properties of welds of 304 stainless plate and also 10 a submerged arc weld in that plate and a submerged i
11 are weld after annealing. The base metal has a 4
12 I fairly low yiele strength, cnd then typically it has I
()
i 13 large elongation. The weld metal has a much --
1 i 14 these two cutves up here, have much higher yield i
I 15 strength. The ultimate str .gth is about the same l
l 161 as tha base metal, out the ductility is much less.
17 Ar.d after annealing, the yield strength of the weld i 10 metal approaches that of the base metal. The 3 19 ultimate strength in the same and the --
20 MR. SilEWMou s In A45 the ASTM --
21 MR. LANDOW: No, that's a plate
{
22 specification. That's a specinen specification that '
l l
23 we have.
24 MR. Sil E Wit 0N : Can you tell me what the 25 material in?
l
)
i 4
. . .- - - ~. - .-_ . . __ -. - _ -- _ __.__-_ __ -
65 i
1 MR. LANDOW It's a 304 stainless steci
- 2 plate.
l 3 tiR . SHEWMou: Okay. Thank you.
!, 4 MR. LANDOW: There's just one of the 5 material specifications within the group Degraded ;
6 Piping Program.
7 And after annealing, some of the ductility I
8 is improved over the as-welded material, but not a 9 whole lot better. There's a brief summary of the 10 proporties, tensile properties. Again, we can sec 11 the trend --
in yield strength within yield while 12 approaching that of the as-received base material,
) 13 and the tonsile strengths are all relatively the 14 same.
l
{
15 !!R . Gi! EWriO N : What was the anneal, what 1
16 was the temperature the anneal?
i j 17 MR. LANDOW: I believe it was 1900, 1950 l
18j for half an hour or 20 minutes.
19 t1R. GilEUMOU : It's an honest to gosh t
20 anneal then.
21 MR. LAUDOW: Yes, you, right.
22 MR. SilEWMON : Okay.
J 23 t1R . LANDOW: And water punching.
] I
- 24 Next, we'll look at Lhe fracture toughnesu l 25 properties of first the base material compared to a ,
I L l
4
)
I 66 i 1 sul me r g ed arc weld. And this is the same material 2 as in the previous view graph. This we have --
3 we're plotting moditied J here instead of
. 4- deformation J. We can see the sharp decrease in the 5 initiation value and also the resistance curve, the 6 slope of the base metal is much higher than that of 7 the weld materials, the submerged arc weld.
I 8 MR. SHEWMON: Are both of the botton 1 l 1 9i curves subarc wel.ds?
) 10 MR. LANDOW Yes, there are two specimens.
I lli MR. GHEWMON: Okay.
I 12l MR. LANDOW: They are two different
() 13 specimens showing reproducibility also.
l 14 1(ext, we'll show some Tungsten inert gas i
15; weld or inert gas welds compared to the base metal, 16 and also we've put in a submerged arc weld from the 17' previous slide to show the difference. In this case l 18 we see the TIG welds have a much higher initiation 19 value than even the base metal, and also the 20 resistance curves after initiation are roughly the 21 same between the baso metal and the TIG welds.
I 22 I would aluo point out these 2.5 T
! 23 specimens have a much higher slope. These are 24 typical uize specimens that are used for fracture
{
l 25 toughness testing, and it may point out that small 2
~~e I
67 1 size specimens may not really tell us what the flow --
"' 2 the crack resistance characteristics of the material 3 really are.
4 MR. SHEWMON: Is it well established why 5 there's this large difference microstructurally? Is 6 it grain size or a lot of inclusions or --
between 7; the TIG weld and subarc?
l dj MR. LANDOW: Well, the subarc weld is a 9 flux weld.
10 MR. SHddMON: That doesn't answer my l
11 question as a metallurgist. Now, it may be a 12 correlation you find useful. You don't know why the
() 13 I toughness is low, except that it's a flux weld, is 14 that your answer?
15 MR. LANDOW: I believe that's -- wo 16 haven't really looked into it much further than that.
17 MR. ETHERINGQUM: Are the production welds 18 with -- is TIG used jukt for the root pass or all 19 the way through? ,
20 MR. LANDOW: In the larger welds it's used i
i 21 as a TIG pass.
22 MR. ETHERINGTON: Just the root pass.
23 MR. LANDOW: Just the root pass, yes.
24 Somo of the amaller ones, 1 --
25 MR. WILKOWSKI: Some of the smaller ones
68 1 will use completely a TIG weld.
2 MR. ETHERINGTON: Here we're referring to
- 3 TIG all the way through, are we?
4 MR. LANDOW Yes, this is TIG all the way 5 through.
6 Next, we'll show a view graph of the 7 submerged arc welds in ferritic steel compared to 8! the base metal property and here we have --
we can 9 show --
it shows the far lower initiation. This was 10 one of the lowest initiation values in the program, 11; about 250 inch-pounds por inch squared for the l
12 l submerged arc weld, and also the effect of the side
() 13 grooves on the fracture properties. The initiation 14~ is aboat the same, but the resistance curves are 15 shallower for the sidegrooved specimens in both 16 cases.
l l 17 nd. ETHERIUGTON: Grade 70 is no longer l
18 sanctioned by the ASME, is it, for pipes?
19' !! R . WILKOWSKI: It's still in the code, I 20 believe.
i 21 tih . E ? ll E R I N G T O ll What?
{
22 MR. WILKOWSKI: I believe it's still in 23 the code.
24 :t a . L ill E R I N G TOti It'a in the ASME. It's 25 utill in the section 2, but I didn't recall that it
- e. m -- w --., -
- , - - - v- - - - . . - - - - , ,y.,._y_,-,..c,,my-,-
y ,,_,. -,_.,7.,,--g..,.,,.,.gr3---->-_,,.-4m,q g,%w--y-==.--e.,.- .--*e-.- +
69 1 was in Section 3 sanctioned it. Well, anyhow --
0 2 MR. RODABAUGH It was there the last time 3 I looked, Harold.
4 MR. ET tlE RI NGTO N : In Section 3?
5 MR. RODABAUGII: Yes.
6 MR. LANDOWt Next we're going to uhow 7 briefly some observation of instabilities in carbon 8 steel at 550 degrees. Again, as Paul showed in his 9 pipe tests, we also see in the compact test some 10 areas where there's rapid reduction in load, 11 followed by an area of stable crack growth and 12 a zie t h e r rapid reduction load, which is a dynamic
) 13 instability in the test specimen.
14 MR. SilEW MON : Excuse me. That is due to 15 nacroscopic tearing due to inclusions or weak areas 16 or --
17 Md. LANDOW: We've looked at those under 18 the SEM, and these areas of instability are ductile 19 fracture. As far as I know there weren't any 20 notable increase or decrease in inclusion -- changes 21 in inclusion content.
22 ;1R . ;illC WMON : There's another question.
23 MR. liU T C ii I N S O N : One can't tell by looking 24 at this curve whether these straight uteep sections 25 are actually dynamic. Is the crack running
> 70 4
l 1 dynanically in those?
i 2 i1R . LANDOW: There's an audible pop in the l i
i 3 test.
4 MR. HUTCHINSON: So it's jumping ahead.
5 MR. LANDOW: It's jumping ahead.
6 f1R . WILKOWSKI: Potential drop 7 measurements of the crack growth show dynamic change.
8 MR. LMIDOW s It's an audible pop and they ---
4 9 MR. HUTCHINSON: So the load-line 4
! 10 displacement is measured over just a portion of the l 11 specimen.
J t t
.i 12 HR. WILK0WSKI: This is a utandard compact 13 tension test.
j 14 tiR . 11U T Cif 1 N S O N : Okay. And the fact that i
3 15 you're instantaneous or nearly instantaneously able , l 16 te get additional load-line displaccuent is because 17 the specimen is in a compliant situation?
I
- 18 HR. WILKOWSKI No, it's a low compliant i
q 19 situation, really low.
I l 20 AIR . HUTCHINSOd Where does this extra l
i 21 load-line displacement come from? I mean is it i
l 22 accommodated oy t.h e machine?
l 23 MR. LANDOW: Some of it probably i t, ,
24 accommodated by the curves because the pins aren't 25 very tight fitting, uo that they can.
I
4
- 71 1 MR. WILKOWSKI: I'm sorry.
1 2 MR. HUTCHINSON: Maybe I'm missing 3 s o :n e t h i n g . If you're measuring displacement across )
4 the grips of the machine and if we pretend that the
- 5 nachine is a rigid machine, then you're not going to - -
6l in an instantaneous event with the machine not --
7 (Indicating)
I i 8! MR. WIL KOWSKI: Okay. As that crack jumps 1
9- you change the compliance of the specimen somewhat, I
10, so that there is some effects there. Within the low i i 11 compliance that exist in the test machine there is i 12 .a o a a compliance there.
l 13 ;tR. IlU T C ll I N J O N : 60 that reflects that?
14 11R . WILKOWSKI: Yes.
l f
15 MR. IiUTCHINSout Okay.
16 MR. KRAMER: Mark, in that particular 17 graph, that is not necessarily dynamically produced.
18 ttR. LANDOW: That's true too. The data 19 acquisition system is going slower, so we .n a y not 20 catch the two points opposite side to the crack
- 1 2L oither --
or the instabilities.
22 MR. S!!EWMO N : Could you tell me what l
2a t e m p e r a t. u r e that test is performed at?
24 MR. LANDOU: 550 fahrenholt to ABC.
N 25 Ma. 361 E W M o u s okay.
l .. .
i 1
i
. _ . - .- . _ . -. . _ _ . ~_ _ _ . _- - - -_
4 I 72 1 MR. LANDOW: We think that this phenomena O 1 may be caused by dynamic strain-aging, which is I
3 accompanied by a high rate of work hardening. The J 1 4 stress / strain curves tend to be serrated, though not 5 at 550, but we do see that at 300.
6 I'll just go on here. At 300 degrees in a I 7 tensile test we do see a serrated yielding going on i
8 or serrated stress / strain behavior. However, we l
'l 9 don't see instabilities in the fracture toughness f 10 . test. At 500 we don't see the serrated behavior.
i l 11 liowever, we do see the instabilities. I'll show 12 that. Il e r e ' s 300 fracture toughness test. We even 13 see the serrations in the fracture toughness test.
14 These are two on loadings that were done during the i
15 test, and you saw the one at --
again, we. don't see i
16 any instabilities at 300 fahrenheit, but we do at 17 550.
i 18 And we feel that one of the --
if dynamic j 19 strain-aging is in effect why we don't see the 20 serrations may be due to the behavior of the 21 material as shown in this figure, where during 'the i
l 22 tensile tests at 550 fahrenheit we're in an area 1
23 here, where at 300 we're over in here,'but due to
/~\ 24 the higher rates of deformation at the notch tip in V
25 the fracture toughness specimens we actually_ may at L
4
i i-i 73 1 300 be in the machine up here where at the notch tip 2 doesn't see serrations but where the fracture -- at 3 550 we may be in the reach and up here we might see 1 4 dynamic instabilities.
. 5 In conclusion here, the significance of J
! 6 what we're seeing is that there is the material i
7 anisotropy which affects the crack growth behavior 4
j 8 in the piping materials.
i 9 Solution annealing of the stainless steel 10 welds reduces the yield strength but does not change 11 the fracture toughness significantly. Pipe tests of 12 the solution annealed piped weld had load-carrying
() 13 capacities reduced in direct proportion to the flow
- 14 strength, which is the average of the ultimate and l 15 the yield.
16 The stainless steel TIG welds had greater 17 toughness than the base metal or the flux welds, and i
l 18 the carbon steel flux welds had the lowest toughness 19 observed in this program.
20 MR. ETHERINGTON: Would you clarify for me 21 a little dynamic strain-aging. The strain-aging 22 occurs suddenly, but is that because of the rapidity
- 23 of the load or you just reach a load where
~
24 strain-aging occurs?
25 MR. LANDOW: Well, strain-aging occurs
l 74 1 over a region of temperature and strain rate.
2 MR. ETHERINGTON: Yes, I understand that. ,
! 3 It is strain rate. But when you have a drop of a i
~
4 beam in a tensile test, would you call that dynamic ,
f 5 loading? I'm talking about the old time.
! 6 MR. LANDOW That's not --
that's not a i
7 strain-aging effect. That's --
i
, 8 MR. ETHERINGTON: I thought it was, but r 3
h 1
9 not dynamic.
I j 10 MR. SHEWMON: He's talking about a yield I
11 drop in the tensile test at room temperature.
. 12 HR. LANDOW: Yeah, that's due to i
) 13 ~ dislocations or pinning of dislocations which are j 14 suddenly released. The drop is a dynamic one.
j j 15 MR. ETHERINGTON: It only occurs in, well, i 16 typically, for example, in Rimsfields, doesn't it?
17 MR. WILKOWSKI: Okay. That would be the
- 18 data that you have there, Mark, where you see it.
l 19 MR. ETHERINGTON: I always understood that i,
20 was strain-aging effect.
i l 21 MR. SHEWMON: Yes, but if you raise the '
1 i
i 22 temperature at which you do the test, then you can --
J 23 if you do the test you're talking about, then you ;
24 don't recover the yield point for a long time at ;
, 25 room temperature, but if you do the test hot, say, ,
1 4
1
75 l
1 at a couple hundred degrees, see, then you CD) 2 continually get recovery and yielding, and I think 3 he's suggesting that that's what he's seeing here.
4 MR. LANDOW: Yes, in these tests in most 5 of the steel pipe materials that we've tested the 6 ultimate tensile strength at 550 is greater than 7 that at room temperature.
8 MR. ETHERINGTON: Yes.
9 M F. . LANDOW: Which is an effect of the 10 dynamic strain-aging.
11 MR. ETHERINGTON: But is that dynamic?
12 MR. LANDOW: In tensile test it is, yes.
) 13 You're dynamically testing the specimen. It's not 14 a --
15 MR. ETHERINGTON: Well, you've got -- well, 16 brittle hardening, for example, would be about 500 17 degrees too, or 450. That's not a dynamic effect at 18 all, is it, that's a temperature effect.
19 MR. LANDOW: Right.
20 MR. ETHERINGTON: So this is a dynamic 21 effect then. ,
22 MR. LANDOW: Yes, this is --
yeah.
23 MR. WILKOWSKI: It's dynamic, but still a 24 relatively slow strain rate.
25 MR. LANDOW: It's a strain rate --
it's
76
_s 1 influenced by the strain rate.
2 MR. ETHERINGTON: Strain rate and not just 3 strain.
4 MR. LANDOW: Right. Strain rate, not just 5 strain. And temperature.
6 MR. ETHERINGTON: Okay.
7 MR. LANDOW: Just one last summary thing.
8 de see serrations from strain-aging occurring in our i
9' carbon steel tensile tests at 300 and not 550. We 10l see dynamic crack jumps in carbon steel fracture 11 tests at 550 but not at 300.
I 12 We believe that strain-aging is to be the
(,)
/
13 cause of these crack jumps at 550 due to the high 14 tip strain rates, and that dynamic strain-aging may 15 affect the fracture behavior of carbon steel piping 16 under seismic loading rates.
17 Are there any other questions?
18 MR. HUTCHINSON: I don't know if you 19 looked at the segments of the crack surface where 20 these dynamic jumps have occurred. Have you looked 21 at those just in an optical microscope?
22 MR. LANDOW: We've looked at them in the 23 SEM. They are ductile.
1 /~N 24 MR. HUTCHINSON: And they look pretty much NJ 25 the same as everywhere else?
77 l
1 MR. LANDOW: Yeah.
7
( )
1 # BENDER: I'm trying to understand the 2 MR.
3 application of this information. Presuming that 4 it's meaningful --
and I can't absorb it in the Slj limited time I've got, so I don't know whether it is i
6i or not. But the loading postulate starts with 7! something that says there's a crack there, there's a 8 sustained load from pressure and something else, 9 maybe it's thermal or whatever, and now we're going 10 to impose a seismic load on it. And it's the 11 seismic load that is the driving force, I guess, for 12 the crack moving and because it's going to be
(^x '
() 13 something that's going to be put on and taken off 14 very quickly.
15 What fraction of the sustained load that's 16 there all the time has to be seismic for it to make 17 any difference? Do I know anything about that from 18 this information?
19 tlR . WILKOWSKI: I think the implication 20 from this work -- this dynamic crack jump behavior 21 that we have observed, and that's all it is right 22 now is some observations and trying to understand 23 why it's occurring, but a potential implication is fN 24 that if the crack tip strain rate is higher, then v
25 the fracture behavior could be different than what
73
,, I we would predict using quasi static C(T) results.
- x
% )
2 So if the strain rate at a crack tip is 1
3 higher because of some seismic loading, that might 4 be important, what is the strain rate as opposed to is the magnitude of the strain.
5lwhat
)
6: MR. BENDER: I think you and I are on the 1
7 same wave length. I'm trying to understand what 8 strain rate I would be looking for before I would 9' start to think about whether it was important or not 10 and how I might try to judge that. Is it some I
i 11f fraction of the sustained load or some other way 12l which I would use to judge whether the seismic i
r" (N) 13' loading is going to make all that much difference?
l 14 MR. WILKOWSKI: I guess I can't answer the 15 point now. It's a point under investigation that we 16 started in this program and that would be continued 17 in the IPIRG program that we'll discuss a little bit 18 tomorrow.
19 MR. BENDER: We'll get back to it, I 20 suppose. In some parts of the country the seismic 21 loading ;s low and in some other parts of the e +
22 Jountry the seismic loading is high, and I would 23 like to know when it is I am to start worrying about
~
24 it, and I'll just leave it there right now.
{V')
25 MR. SHEWMON: Do you want Oak Ridge's
79 7, 1 opinion or somebody else?
\") 2 MR. BENDER: Well, I would rather have a 3 good stress analyst's opinion, never mind a 4 physicist who wants to argue stress analysis.
5 MR. WILKOWSKI: Are there any other 6 questions?
7 MR. SdEWMON: Not if there's.a break.
8 MR. WILKOWSKI: Yes, there's a break.
9 (Laughter) 10 MR. SHEWMON: Okay. Ten minute break.
11 (Short recess taken.)
12' MR. W I L KO W'S K I : Paul Scott will talk about p)
( 13 the work on weld overlay repairs.
14 MR. SCOTT: Paul Scott from Battelle 15 Columbus. As Gery indicated, we'll be talking in 16 these few minutes about the weld overlay experiments 17 that we've conducted to date and the one experiment 18 we have planned in the future.
19 To date we've conducted three weld overlay 20 tests on six inch ciameter pipe at 550 degrees F.
21 We've evaluated the results of those tests in light 22 of the net-section-collapse analysis and the 23 IWB-3640 criteria.
's 24 The significance of the results to date
{%
25 are the weld overlay pipes failed well below the
80
. 1 net-section-collapse, possibly due to residual N 2 stresses in the weld.
3 The weld overlay failure loads were 4 slightly below the ASME Section XI IWB-3640 source 5 equations if no safety factor is used in the ASME s
6 analysis. The weld overlay pipe failed slightly 7' above the ASME Section XI IUB-3640 table values if 8, no safety factor was used in the analysis.
i 9 For the larger diameter pipe, contained 10 plasticity may lower these failure loads with 11 respect to the net-section-collapse even farther.
12 And finally, we'll look at different ways
) 13 that can be used to predict the net-section-collapse 14 stress under pressure and bending loading.
15 MR. SHEWMON: Would you define contained 16 plasticity?
17 MR. SCOTT: That's where the plastic zone 18 size would be less than the remaining tensile 19 ligament. In other words, you don't reach a fully 20 plastic condition.
21 MR. Sil E W M O N : Okay.
22 MR. BENDER: Excuse me. Can you tell us 23 what the variables are that de ought to be concerned ;
l L
'N 24 about with weld overlays?
{d 25 MR. SCOTT: What the variables --
l
81 1 MR. BENDER: Yes.
(m
')
2 MR. SCOTT: Well, obviously --
3 MR. BENDER: The amount of overlay, the l
4 length of the crack that's being overlaid, what are 5 the things that we ought to be thinking about?
6 MR. SCOTT: Well, offhand I would say that 7 the size of the crack and'the thickness of the 8 overlay would all come into play based on how they 9 fit into the net-section-collapse of the IWB-3640 10, analysis.
t 11 -
MR. WILKOWSKI: The size of the pipe.
12 MR. SCOTT: Yeah, the size of the pipe.
() 13 3R. BENDER: Well, you're doing one test.
14 I don't know how many tests you're going to do.
15 MR. SCOTT: .le've done three right now.
16 MR. BENDER: And I'm trying to figure out 17 what the tests will bracket. What variables will 18 they bracket? What are they intending to test, just 19 to show that you can do an analysis which is 20 predictable?
l 21 MR. SCOTT: Basically that's what we're 1 22 trying to do right now is to verify or evaluate 23 those analysis procedures. I don't think we're 24 trying to bracket anything with our analysis or with (Ns) 25 our experiments.
i
82 1 MR. BENDER: I'm obviously only asking i,s>
s i
# 2 about the application, and so the thing I keep 3 asking myself is what kind of things exist out there 4 that I'm going to evaluate with this procedure. If I
i 51 I just put an overlay on a piece of pipe, do I have 6 to assume that there's a certain crack geometry l
7 under it? Do I have to assume that the overlay has 8 some dimensional relationship to the original 9 thickness of the pipe in order to have some 10 understanding of whether the analysis applies or not?
11l That's what I'm trying to find out.
12' MR. WILKOWSKI: Excuse me. Most overlays f%
() 13 are typically designed based upon what the flaw size 14 is determined to be by ultrasonic testing with some 15 appropriate aafety factors, and they base that 16 overlay on the anticipated crack and the 17 net-section-collapse analysis. So what we're trying 18 to do is to evaluate tha net-section-collapse 19 analysis for what we considered one of the worst 20 case types of cracks that might occur.
21 tiR . BENDER: This is where you assume 22 there's a crack --
a worst case crack exists and 23 somebody has already pat the overlay on or are they 24 deciding how much overlay to put on?
{v')
25 MR. WILKOWSKI: It's a case where they
83 7.s 1 have put an overlay on a pipe and stress corrosion
\ l
'# 2 cracking has continued until the crack has reached 3 the overlay. So it's coupletely through the wall of 4 the original pipe. And then the stress corrosion 5 crack has been arrested 'o y the weld metal that's 61 being used in the overlay process.
7 MR. BENDER: Well, that's enough for now, 8 I guess. Thank you.
9 MR. SCOTT: Okay. As far as future plans, 10 as indicated, there's one 16 inch diameter pipe 11 which will be tested to assess the plastic-zone i
12 ! screening criteria and the IUB-3640 analysis. As
, ) . 13 mentioned, all the previous tests to date have been 14 on six inch diameter pipe.
15 MR. RODABAUGH: Six inch Schedule A.
16 II R . SCOTT: Six inch Schedule 120 is what 17 the tests have been conducted today.
18 MR. RODABAUGH: And the 16 inch will be --
19 MR. SCOTT: Schedule 100.
20 MR. RODABAUGH: That's the one inch wall.
21 MR. SCOTT: One inch wall, 1.03. This is 1
22 just a schematic of the repaired pipe section.
23 Essentially what we did is we started with an 24 initial EDM surface flaw, then loaded the thing, and 25 three point bending in fatigue, and grew this
___.J
84 gy 1 surface flaw until it grew to the wall until such 2 that we had a through-wall crack, fatigue sharp l
3 through-wall crack halfway around the pipe 4 circumference.
5 Then we shipped the pipe to t!utech 6 i Engineering in California and they put a two-layer
(
7 weld overlay on top of this cracked pipe section.
Si Now, the thickness of the overlay was about .31 9' inches, such that the D/t ratio or the wall I
i 10fthickness to the combined wall thickness ratio was 11 about .65. So the D/t ratio for each of the tests 12l was about .65 and the 2 C over pi D or the pipe
) 13 length was about 50 percent.
14! This just shows the test conditions for 15 one of the three weld overlay tests. The other two 16 tests were very similar. The only thing that I
17!
I changed was the internal pipe pressure. We varied i
18j the internal pipe pressure from test to test, but i
f 19l basically with very small differences. The 20 diameters, thicknesses, all stayed the same.
21 For this particular case the maximum 22 bending stress was almost 20,000 p.s.i.
23 The flow stress for the base material, base metal
( 24 material was 50.85 ksi. That's the 3 S of M value.
25 And the flow stress of the weld metel as measured
85
,m 1; was 57.9 <si based on our flow stress definition of
( i !
2 1.15 times the yield in the ultimate strength. l 3 MR. RODABAUGH: Why are you using in this 4' sort of comparison 3 S of M other than in your i
5 following definition of flow stress for the weld 6 metal. 3 S of :4 might be way off compared to your l
7' standard --
8 MR. SCOTT: Well, for somebody in the code '
9 that's probably all they have --
that's making an 10 evaluation using the code, that's probably all they 11 have is 3 S of M value.
12 :1R . RODABAUGH: You surely know the yield h 13 strength and ultimate strength of the base material:
14 don't you?
15 :1R . SCOTT: Yes.
16 MR. RODABAUGH: I'm just curious. It 17 seems like here is an unfair --
or comparisons on 18 two different bases for some strange reason.
19 MR. SCOTT: Well, we're not really trying 20' to compare the results for the two different 21 analyses. We're just going to present the results f i
22 for the two different analyses. But it's just two i
i 23 different ways to evaluate tnis. We thought this is
(
'd
, 24 probably technically the most correct, but this is 25 what somebody in the field would have to work with.
I 86 gs 1 MR. RODABAUGH: Why would it be --
what's
\s 2 resisting the bending and pressure is a combination 3 of base metal and weld metal on one side. Why would 4l one or the other be more accurate? Why would you i
5 use weld metal in preference to base metal?
6 3 MR. WILKOWSKI: Well, technically that's a i
7, good question, but in the actual design of the weld i
4 8l overlays, we'll use 3 S of M of the base metal, so 9l that's why we used that.
I 10f MR. RODABAUGH: Because in this big table 11, with your latest progress report I noticed you use 3 12 S of M quite a bit. I'm wondering why you didn't "A : !
(N 13 use your standard flow stress definition.
i MR. MAYFIELD: Everett, we have asked 14l r
15 Battelle to include 3 S of M as well as their 16{ standard definition because we're getting some i
t 17; questions and some interest from the NRR staff as to, i
18l well, if we don't have actual property we use 3 S of i
19l M, the code tells us to use 3 S of M, so to compare 1
20 the experimental results to 3 S ot A. l l
21 MR. RODABAUGH: That's why. I would like 22 to see the 3 S of M comparison, but I also like to 23 see the flow stress comparison. And it's on the
(} 24 table here, for example. I don't have it.
25 MR. WILKOWSKI: Okay.
.- - _ - _ , -- . _. _- -~ . . - .
87 l
,cx 1 MR. RODABAUGil: It's just a comment.
2 MR. SCOTT: In this table we are looking 3! at different ways to calculate the limit stress 4 based on net-section-collapse, the IWB --
IWB-3640 5 equations using two different formulations for the 6 membrane stress, the PR over 2 T and also the thick 7 wall equations, and two different definitions of l
8l flow stress.
I 9 Of interest, we see that if you took the 10 maximum bending stress and added it to the membrane 11 stress, took that and divided it by the collapse 12 stress plus the membrane stress, we see that in each
) 13 l
case that the net-section --
or the collapse stress 14i overpredicted the experimental stress.
15l MR. SHEWMON: Tell me again what you just l
16l told me. Those numbers being less than one says the I ^
i 17l acceptance procedure is conservative?
l 18i MR. SCOTT: Well, it would say that the 19l acceptance procedure would be inappropriate -- it i
I 20 would be -- well, nonconservative I guess is about 21 the best way to say it.
22 MR. SHEWMON: Okay.
23 (Laughter)
SCOTT:
( 24 AR. Yes. That's what it's saying.
25 MR. SHEWMON: Thank you.
88 1 MR. WILKOWSKI: But that's only for the 7x
\ )
2 net-section-collapse analysis.
3f MR. SCOTT: Yes. Okay. In the code i
41 they've got built-in safety factors, so it's not i
5j nonconservative with respect to the code. It's just 6! not as conservative as the code would lead you --
I 7 MR. SHEWMON: You would like to have that 8 number be at least constant, so that it times the 9 safety factor would be reliable; is that right?
10' MR. WILKOWSKI: Yes.
11 MR. RODABAUGH: What's the presumed safety 12i t' actor for faulted conditions?
13 MR. WILKOWSKI: 1.39.
1 14, MR. RO D AB AUGil : So if I take .74 times 15 1.39, I have a safety factor of less than one.
16 MR. WILKOWSKI: I think we need to see a 17 few more of Paul's view graphs because this is based l
18' on net-section-collapse analysis, whereas the code 19, uses a simplification of the net-section-collapse 20' formulas, and in doing so they add some additional 21 conservatism in there, which makes things look 22 better.
23 MR. RODABAUGH: Right at the moment it
(} 24 25 looks like we've got a safety factor of less than one though for faulted conditions.
89 1 MR. WILKOWSKI: Yes, for faulted Is i N .%') !
2l conditions.
l 3! MR. SCOTT: Okay. This is a --
that last i
4 view graph was just for one experiment. This is for 5 all three experiments. You'll note that for 6; experiment 3 we've got two different data points.
7, Essentially at this pressure level we couldn't get 8 the specimen to fail, so we had to raise the 9 pressure level to this pressure level to get it to 10 i fail, although we were looking at the crack opening 11 displacements at these two tests, which neither one 12 of them did fail, it looked like comparing them to
('s (m) 13 the crack opening displacements of these two tnat i
14l did fail that they were on the verge of failure.
15} Also, we did section this specimen here I
16 and the crack had initiated, but if you plot --
this 17 ! is -- the bending stress divided oy the flow stress 18l is a function of the membrane stress divided by the 19 flow stress. This would be net-section-collapse, no 20 safety factor. So, you see, you do fall below 21 net-section-collapse, no safety f a ~c t o r , for both 22 definitions of flow stress that we're using.
23 If you start, now compare just for the 3 S f'\ 24 of M value of flow stresses, all we're presenting
\j 25 here, this would be the curve for the IWB-3640
1 90
-s 1 source equations. As Gery said, they do kind of --
I
%N) 2, with the membrane stress, they do --
based on the 3 source equations are a little bit more conservative 4l than net-sect ion-collapse .
5 This would be the IWB-3641 table. Note 6l that in IUB-3641 for a flaw of 50 percent or greater, 7 they assume a 360 degree flaw. So they built in for 8' a flaw of 50 percent of the pipe circumference in 9l length or greater, they built in an extra safety 1
10' factor based on the fact that they assume you have a 11 360 degree flaw in their analysis.
12, And you can see that for -- if you use the rx k)m 13l table for this particular flaw geometry, that you 14l fall above your conservative respect to the tables.
I 15l Here, this is the IWB-3641 tables with the safety l
16j factor of 2.78 built into it.
t 17! MR. BENDER: Excuse me. I would like to 18 follow the point just a little bit further because 19 my memory is poor. What is the NRC postulating as i
20 their limiting length of flaw now? Is it two-thirds 21 of the circumference?
22 MR. 11 A Y F I E L D : 0313 essentially adopts 23 IWB-3640.
lf) l
\~/
24 25 MR. BENDER: Well, that says you do an analysis based on a 360 degree length of flaw, but l
l _ _
91 s 1 that obviously doesn't have -- if it really existed,
'} '
2i you wouldn't have any conservatism in it. What i
3! you're operating on is the basis that you're 4, analyzing on the basis of totally --
a flaw that's i
5i all the way around the circumference, but you really 6' think the flaw is not going to be that long. That's 7 what the real conservatism is, as I understand you.
8' And I'm trying to figure out what length 9! of flaw I might be thinking of in order to be i !
10 comfortable with the analytical procedure that l ,
11 exists.
i 12 MR. WILKOUSKI: I don't think there is any
) 13 safety factor associated with limiting the length of 14j the flaw. In many cases when they're doing the
- 15 ultrasonic inspection, the inspectors are in there, 16 and they see intermittent stress corrosion cracking 17 at many locations, and they might actually claim 18 that's a 360 aegree crack, but I haven't seen any i
19 cases where people have limited it because of the
] 20 length of the flaw.
21 MR. MAYFIELD: This 2.78 - maybe I don't 22 fully understand your question, but the 2.78 margin 23 is on the loads that are postulated in the analysis.
24
( MR. BENDER: Well, I'll learn more about 25 it some other time. It's not the place to argue y e - - - N 7'"
92 s 1 about that.
2l1 MR. SCOTT: Okay. This is my final view i
3 graph. Basically if you took the maximum 4l experimental stress as a function of the 5l net-section-collapse stress, these are for our three 6f tests that we have run as a function of the oatside l
7! pipe diameter. If you took the screening criteria 8l and used the assumptions, BWR conditions, Schedule i
9! 80 pipe and a two layer weld overlay, that you would i
10 predict that as the pipe diameter increases that you 11 would get a failure curve that looks something like 12: this; that you might expect based on this that you f'N) s 13l would be even somewhat lower than that.
14 Okay. For the one test that we have i
15 planned in the future, we are planning to run a 16 16; inca Schedule 100 pipe with a three layer weld 17l overlay, and why we're using a three layer weld 18 overlay in this case is it gets you the same R/t 19I ratio as our six inch test.
20 Here we would" expect it to be a little bit 21 higher than this curve, so we might expect based on 22 these results that we might see failures somewhere 23 in this area of the 60 to 75 percent range. And
(~\ 74 that's one of the reasons we're doing this larger d
25 diameter test to see if this contained plasticity )
1
b l
93 1 will lower the failure stress even lower than what 2I we saw back here.
i 3 MR. ETHERINGTON: Are the weld overlays 4I machined?
5, MR. SCOTT: Are they machined?
l 6i MR. ETHERINGTCN: Yes, machined.
I 7j MR. SCOTT: When they --
l 8l MR. SHEWMON: To remove surface roughness.
I 9! MR. SCOTT: No, no.
I lof I MR. SHEWMON: I assume that's what you 11 meant.
12 MR. ETHERINGTON: Yes.
) 13 MR. SHEWMON: It's usually done if anybody
- 14 wishes to ever do a UT inspection on it or try, for i
15 example.
16 t1R . SCOTT: Um-hmm.
17 MR. WILKOWSKI: Okay, i
lot dR. SCOTT: Tnank you.
1 l 19 MR. WILKOWSKI: The next speaker is Jalees 1 20 Ahmad, who will talk a little bit on the 1 l l 21 introductions on elastic-plastic fracture mechanics. l l
l 22 MR. AHMAD: He asked me to speed up about 23 three minutes. I don't know how many I had to start
(} 24 25 with, but I'll try to speed up as much as I can.
Again, my name is Jaless Ahmad. My role l
, l i
. _ _ _ ~ _ - _ - . - _ ,
94 s 1 in the program is to coordinate and to some extent F
2 work on the elastic-plastic fracture mechanics j 3 analyses.
4 To refocus here a little bit, we have been i
5! talking mostly about experiments. One of the main i
6 program objectives here is to develop, improve and 7' verify predictive fracture mechanics analyses for 8 nuclear piping.
9 You saw some analyses discussed during the 10! predominantly experimental portion of the 11 presentation, and those were like
- 12. net-section-collapse or limit-load type analyses.
) 13 it e r e I'm going to talk about a step further into the 14 more elaborate type of analyses.
15l '
MR. RO D AB AUGil : Before you leave that, 16! what's TNT?
17 MR. WILKOWSKI: That's explosive.
18! MR. AHMAD: To quantify the energy.
19i MR. W1LKOWSKI: Quantify what type of 20 fracture experiments we can do. They have to be 21 relatively low energy release.
22 MR. RODABAUGH: Okay. Thank you.
l 23 MR. AHMAD: The overall program then on 24 the analysis side is described here.
( We have 25 laucratory specimen which include simple stress j
95
_ 1 strength tests and fracture property tests. We have
%~) l 2{ some full scale pipe experiments and advanced l
3 elastic-plastic fracture mechanics techniques, and 4lthe idea is to develop some simple predictive l
5 analysis methods.
6 MR. SHEWMON: Now, these advanced I
7j techniques do not take account of change in shape, 8i like ovalization though; is that right?
9f MR. AHMAD: They could. They could in i
10 principle. For example, if one were to perform i 11 elastic-plastic finite element analysis of these l
12 five experiments they could be taken into account,
) 13 large deformations and stuff. But the overall 14j purpose here is to come up with simple engineering i
15 estimation schemes, learning from more detailed 16 finite element analysis what goes on, but the main 17 product has to be in, for example, a simple PC 18 computer code form or equation form that can be used 19 in practice. The two types of analyses, you heard i 20 quite a bit about the limit-load of 21 net-section-collapse analysis, and by now I think we 22 all have a pretty good picture that in all cases 23 that we have looked at net-section-collapse analysis
\ 24 is not always satisfactory.
[V So the idea is to 25 perform net-section-collapse analyses, go through a
. - - _ _ _ __ -. - - _ _ . _ _ ~_ _ _ _ _ - - _ _ . - - _ _
96 x 1 screening criterion, and again the screening s
u) 2 criterion, just to refresh your memory, there's a 3l view graph that Paul Scott showed you earlier, which I 1 4l basically tells you when you can and when you cannot 1 5 use net-section-collapse analysis. I 6 Well, when you cannot use i
7l net-section-collapse analyses, only then we drop 8i down to this more sophisticated type analysis using 9i elastic-plastic fracture mechanics.
10' MR. ETHERINGTON: Now, does this apply to 11 ferritic steels only?
1 12l MR. AHMAD: No, sir, it's applicable to
) 13 all kinds of steels.
I t
14l! MR. ETHERINGTON: What?
l i i
15j MR. AHMAD: It's applied to all kinds.
16!t fir . ETHERINGTON: All kinds?
l 17'j MR. AHMAD: Yes. The main product of the i
18I analysis effort, and I'll emphasize again, is to l
t 19l find and validate predictive engineering analysis
)
i 20 methods, okay? The purpose is not to do necessarily 21 very detailed finite element analysis in cases which 22 cannot be used in practice. And again the product 23 is going to be a user-friendly computer code for 24 engineering analysis and such a code already exists.
25 It's continually being updated and refined, and also
97 i
t e3 1l an accompanying handbook of solutions that can be s
N) 2. used.
l 3! MR. RODABAUGH: Jalees, looking at this 4i curve, picture you just shoued us, I notice first 5 that it has a J sub I as one of the parameters which, 6, is beginning to get into your field. I also notice, t
7' if I understand this, that you nave unusually good --
l 8' what I would call adequately good agreement between 9 test and this parameter. So ray question with 10 respect to what you're saying --
lli MR. AHMAD: dell, maybe you can clarify 12 that plot a little bit because that's no analysis
/~N.
\m) 13 involved in the -- okay. Those curves are basically 14 ten curves combining those data , points. All the 15 information on this plot is experimental essentially.
i l
16! Okay. Put in this form. You're right about J sub I I
i 17 1 being a fracture parameter. But there's no way to 18l predict this curve, for example.
19 MR. RODABAUGHz Well, you could do it 20 purely empirically by fitting this curve on this 21 parameter.
22 MR. AHMAD: You could conceivably require 23 a large number of experir7nts, especially in this I\ 24 region where the slope is extremely steep and small
'd 25 perturbations in J sub I, for example, would throw l
1 98 l gw, 1 you quite a bit off fron, that curve. So there are
\u 2ll problems associated with that.
I 3l , dR. RODABAUGd: Okay. So this is not a 4 satisfactory predictive method, and now you're going 5 on to show how to get to a satisfactory method.
l 6' MR. AHMAD: That's exactly right.
i 7! MR. SHEWMON: Would you remind me what P i
8l is, P max, PL, PI?
I 9l MR. AHMAD: Jkay. Tnis is one of the 10' plots that I borrowed from Paul Scott, and P sub I 11 here is essentially the loaded initiation. P max is 12 the maximum experimental load.
) 13 MR. SHEWMON: Fine.
i 14l MR. AHMAD: And analysis was --
1 i
15i MR. SHUWMON: That's enough.
16: MR. AHMAD: --
done two ways.
17l I will take you through the overall
' l 1
18i approach, analysis approach in this program, and 19 it's kind of a busy view graph. Also, it doesn't I
20 come out too clearly, but that says generate small 21 scale fracture experimental data, the kind of data 22 that you get from compact tension specimens and such.
23 Analysis plays a dual role. First, you've (h
%)
24 got to analyze the small scale specimen data which 25 comes out essentially as a load displacement versus
~
l
99 x 1 crackling record that you've got to analyze in order 2! to generate, for example, J-R curve. So there is
)
3, what we call a generative or interpretative type of '
I 4l analysis required there to interpret the data to l 1 5; establish a J-R curve. ,
l 6 Okay. Once a J-R curve is developed, then )
l 7lthe t
other kind of analysis that's being performed is 8 the predictive elastic-plastic fracture mechanics i
9 analysis, in which this J-R curve is now fed as an to. input, and you predict the loads and displacements !
11 at crack initiation, its instability. And as Gery 12, mentioned earlier, it's important for us to predict 13 the entire load displacement curve from the i
14t leak-before-break assessment point of view.
I 15 I so in the predictive analysis then, J-R I
i 16l curve is an input and the predictions are here.
i 17lThose predictions are then compared with the pipe i
16' experimental data. Okay?
19' MR. RODABAUGH: What does EPFM mean?
20 HK. AHii A D : Elastic-plastic fracture 21 mechanics.
22 MR. RODABAUGH Thank you.
23 MR. AHMAD: Using small specimen data, l 1
24 predicting pipe behavior, comparing it with full
(
25 scale pipe experiments. Then you look at those
100
- 1 comparisons, whether it is good or not. If not, you
\m/ We 2 go back, develop or modify the analysis methods.
3 start with some existing analysis method and try to 4 improve upon them if the predictions are not always j 5 good.
6 That development and modification process i
7* is assisted by some detail finite element analysis.
I i
8j Okay. Those are the more expensive types of I
I 9 analyses, which are not used in the analysis of i
10l every experiment, but only as an aid to improving i
11! our analysis techniques.
I 12iI At the same time those developments and 13 modification methods also feed into analyzing small 14' specimen data.
15 In case the predictions are not very good, 16l you need either to develop and in some cases --
and l
l 17l that depends upon the position of the program 18j manager upon the significance of the finding or the 19' problem --
that you simply then document and point 20 it out as an area which needs further attention.
21 So that's the overall approach. And again, 22 what you have is a product here --
is a document and 23 l a computer code for predictive analysis.
(s_
24 MR. RODABAUGH: You have over here on my 25 left, what's that as compared to the clastic-plastic l
101 e 1 finite --
(~%)
m')
2: MR. AHMAD: Finite element analysis 3l versus --
4 MR. RODABAUGH: Perform EPFM.
5I MR. AHMAD: Tuis is all engineering simple 6! analysis here.
7 MR. RODABAUGH: Including elastic-plastic -~
8' MR. AHMAD: Right. It's c:omething like a 9' J-estimation scheme. Do you remember the --
10j MR. RODABAUGH: All right. Then the 11l finite element analysis up in this block?
12!t MR. AHMAD: The finite element anajysis is
) 13 used essentially as an aid to developing those 14 simple formulas or to verify their accuracy.
15l MR. ETHERINGTON: 'le
. l l , in an austenitic i
16 ! steel do you ever reach a prediction of instability 17 based on the J-R curve?
18 MR. AHMAD: In compliant tests, yes. We 19 can predict instability using J-R curve. Is that 20 the question?
21 MR. SHEWMON: It will do for now. Go 22 ahead.
23 MR. AHMAD: Okay. Just to refresh our f% '
24 memory, and maybe this will answer your question a
(%s) 25 little bit in more detail, the technology used here
102 1 is a J-based elastic-plastic fracture mechanics (w
s N
'j 2 analysis, and this view graph I put just to remind 3 you what we are doing, and also to point out certain 4, issues that come out of it. 1 5; Looking at basically J equals JIc, plane I
6! strain conditions exist as a criterion for i
7l initiation, that doesn't always happen, especially l
8j in tougner materials, and, for example, in 304 9! stainless steels we are doing full thickness l
10 specimen tests, and using J equals J sub I as an i
11 initiation condition.
I 12l Stable growth, that occurs as J equals the 13 J material or the J-R curve, and then the i
14 instability condition is defined as dj/da on the 151 slope of the applied J versus A greater than dj 16 material da which is also being formalized in terms 4
17, of the tearing modulus T-MAT, and on the right t
18l you'll simply see a graphical representation of i
19l those materials.
205 MR. ETilE R I N G TO tJ : That gets back to my 21 previous question. In austenitic steels, those two 22 curves don't cross, normally at least, the bottom 23 right-hand curve.
1
[N x_)
24 MR. WILKOWSKI: Well, they can under 25 certain conditions.
i l 103
! I l! MR. AHi1 A D : Under certain conditions they 2 can.
3; MR. ETHERINGTON: Under which conditions?
4 MR. AHMAD: Under certain conditions they 5 can.
6 :1R. ETHERINGTON: What are those 7 conditions?
8 MR. A H :1A D : Well, the condition is 9 formalized right here, if that ever happens.
10 MR. ETil E R I dGTO U : Yes, all right. Okay.
11 MR. AHMAD: If your dj/da --
12 11R . ETHERINGTod: I understand that, but I h 13 didn't think that was a condition at which thcy 14 would intersect.
15 MR. AHMAD: Under certain conditions they 16 will.
17 MR. ETHERINGTON: Nell, I'll accept your 18 statement.
19 11 R . All M A D : Okay. In that technology, of 20 course, the elastic-plastic fracture mechanics, the 21 inharent assumption here is that the J-R curve that 22 is developed using the small specimen data is a 23 material property. Okay? 30 you're using the J-R 1
(; 24 curve as a material property from going from a small V J 25 specimen to predicting pipe behavior.
I
104 1 Okay. In this program, as you've already 2 seen, we are looking at basically three types of 3 flawed geometries, predominantly under pipe under 4 four point bending, in some cases under combined 5; bending and internal pressure. You have a i
6l through-wall crack case that's being locxed at. You 7! have the circumferential surface crack case and 8 third would be calling as a complex crack where we 9 simulate the condition that you have a very long 10 surface flaw which has broken through the part of 11l the thickness and is a combination of surface flaw 12l and through-wall crack.
) 13 Now, going trom small specimen to i
14l preaiction of flaw behavior, where flaws look like 15 that, there are some issues which need to be 16j aadressed, and these are being addressed in the 17l program as well. First being the constraint effect, 18 and the simple example of that would be, for 19l example, the plane --
the plane strain coastraint i
20 used in the small specimen data not being satisfied 21 in the pipe, and that is primarily being addressed 22 by using full thickness pipe specimens.
! 23 Application of cracks in welds, whether
( 24 this technology basically developed originally from 25 monolithic materials can be extended to application
105
/,T lj of cracks when the cracks are existing in welds and
! I
~
2' you have more than one material.
3 The geometry dependence of the J-R curve, 4 going from small specimen to large specimen, whether 5 you have the same J-R curve behavior, that is being 6 addressed.
7 Large amount of crack growth, and that 8 question was bought up earlier, is important because 9 from small specimen data you get a J-R curve which 10is only a fraction of an inch in dimensions of crack 11 extension, while in pipe experiments that we're 12 trying to analyze you can get very large amounts of h 13 crack growth. And to analyze those, you need data 14i for large amounts of crack growth.
15 .i e l l , when the crack growth is large 16 there's a question whether the deformation theory 17 base J analysis is valid or not. We are looking at 18' that by performing some finite element analysis as 19 well, which are going to be discussed a little later, 20 !
but besides these, there's two issues about the 21 availability of analysis methods. Even if the 22 theory was perfect, well, we have simple analysis 23 methods available to analyze all these kinds of
( 24 flawed geometries and loading conditions. In many
)
25 cases they do exist. In some they don't, but in
106 1 cases that analysis methods have been proved, we 2 have taken it by performing some detailed finite 3 element analyses and direct comparison with 4l experiments to check the accuracy of those analysis 5 methods before we propose the engineering analysis 6I techniques in the form of a code or a handbook.
l 7- MR. BENDER: Excuse me. Are you satisfied 8 that you're going to get enough data to establish l
9I the accuracy of the analysis methods?
I l
10l t1R . AHMAD: For the cases that are 11l included in the program, yes, sir. That program is 1
12l designed so that the analysis methods that are being
) 13 looked at have sufficient information from the 14l experinental part of the program.
i 15l MR. BENDER: Well, how do I know that?
16l Can I find a document that tells me?
i Because I'm 17l kind of interested in that section. Is there I
18I something that says why these tests are being done i
19 for the purpose of proving the analytical method, 20 its accuracy or whatever you want to show it to be?
21 MR. WILKOWSKI: There are statements to 22 that effect in many of our reports.
23 MR. BENDER: There are statements to that
( 24 effect concerning the program. I'm interested in 25 something that says what experiments are being done
107 :
i 1 to prove what part of the analytical method and why.
(3
"~J l
2 And I guess I don't see that yet. And I'll take on I
l 3 faith that you're getting enough data, but I would 4 like to see something that explains it better.
i 5l MR. WILKOWSKI: Okay. Perhaps we can do 6 something to that effect.
7 MR. BENDER: Okay.
8 MR. AHMAD: Okay. I'll skip a few things 9; here and just show you somewhat of an explanation of l
101 the kind of things we are doing here, and this will i
11 then continue on with the following speakers. At i
l 12l the start of this program, for example, there was no
) 13 satisfactory simple formula or an estimation scheme 14 available for analyzing pipe with a part-through 15 crack, which is not 360 degrees all the way around 16 the pipe, and this being a three-dimensional problem 17 essentially and requiring nonlinear analysis.
18 We have right now developed a procedure i
19 which uses the available information, in this 20 particular case solution for an edge crack panel for 21 which a simple estimation scheme in the old GB and 22 EPRI handbook does exist, and using that simple 23 solution available in the handbook, then try to 24 develop and estimation scheme for the surface crack
, N 25 pipe.
_ . _ _ _ _ _ . _ _ . _ _ . ~ - _ _
108 g~.., 1 And I won't have time to go through all
(
2 1
the details of this method, but what comes out of 3 that analysis procedure is again a simple formula i
4 accompanied by tables that can be used to predict Sj crack initiation, load and load at maximum --
in the i
f 6: maximum load in a surface crack pipe experiment.
7:
Some of the detailed --
and again in the l l 8; course of an hour or so, we cannot possibly present l !
9! all the analysis efforts that are going on in overy l
- 10. detail. So what we have done today is selected a 11 few key topics that will, we hope, give you a flavor 12 of the kind of things that are going on.
) 13i The first of those is an assessment of 14l through-wall crack pipes in bending, using some f 15i existing J-estimation methods which were used in the i
16l analyses, and again if you remember the approach, [
i i 17j certain modifications or improvements in those l 18 methods were made, and that is to predict both the f
19 loads at initiation, maximum load, and also the l
i r
20 load-line, the load displacement predictions. That j i
21 will be presented by Dr. Bud Brust. I 22 Following that, there will be some l t
23 discussion by Mr. Kramer on the assessment of
( 24 complex cracked pipes. He will describe both the 25 experimental part of it and some of the analyses
109 gm 1 efforts, and there we've deviated a little bit from
'm ) 2l the general J-based clastic-plastic fracture i
l l
3l mechanics and used some CTOD estimation methods as 4: well.
5' And following that, we'll hear a little 6l bit about our different finite element analysis 7l efforts which look, for example, at applicability of 8! the J-based approach to large amounts of crack.
9; With that, do you have any more questions?
10 MR. SHEWMON: Thank you.
lli MR. BRUST: My name is Bud Brust from 12 Battelle Columbus. I'll be discussing some of the
[~)
Nsj i
13! details of the general approach that Jalees just 14i described, and I'll be discussing through-wall 15; cracks in pipes.
I I
16l In effect we use the results from the l
17l material characterizations that were discussed by 18 Mark Landow earlier, and we compare using our 19 i analysis procedure to the full scale experiments 20 that are carried out, and they were discussed 21 carlier by Paul Scott.
22 So the purpose is to see how good these 23 estimation schemes perform when compared to
( 24 experimental data. Again, I'm talking about 25 through-wall cracks in pipes subjected to bending i
110 1 loads.
2 To repeat what Jalees just mentioned, we 3 use material property data in the form of a 4 J-resistance curve and material stress / strain data, i
5' and those are written --
the estimation schemes are 6 4 assumed such that the material property data can be 7I characterized by a Ramberg-Osgood curve.
8j And I would just show this briefly. This 9; is a slide that Mark showed earlier to show --
we 10l get the J-R curve, the resistance curve, from a 11 compact tension specimen cut from a piece of pipe, t
i 12j and the Ramberg-Osgood data is obtained from a
) 13 tensile specimen cut from that same pipe.
i 145 So given stress / strain data and J-R curve 15 data, we use a J-estimation scheme to see how well i
16l they perform. The J-estimation schemes can predict I
17l crack initiation, maximum load and load versus I
18i displacement behavior, and I note that these are 19l general approaches. It's not Just applicable to one 20: given material property under a certain type of 21 condition as an experiment is limited.
22 I'll first describe a few of the methods 23 and then show some of the results. One of the 24 methods is a compilation of numerical solutions,
(%)'\
25 so-called GE handbook approach, and we note right
111 gm 1 off the bat that shell elements were used to produce 2 these results, and they tend to underpredict loads 3land displacements. They tend to produce too stiff l
4! solutions. And we also tend to get poor displacements ,
t 5i but these are underpredictions, that is, they tend 6! to give conservative predictions.
7l h second method is Paris method which
! l 8i interpolates between a known linear elastic solution 9 3 and a known fully plastic solution in some fashion.
l.
10l It's independent of material property behavior.
i lli That's a problem of it.
i 12 The LBB-NRC method takes the Paris method
('s s m) 13, and modified it to account for material hardening, I.
14j and this gives an improvement on predictive results.
15! Also, within the course of this program 16 we've developed a couple of estimation methods, 17i which one is actually based on the GE results, but 18l just interpreted in a different way, and another 19 method is effectively a different procedure to l l
20 estimate the reduced compliance in a pipe due to '
21 plastic behavior that occurs at a crack tip. )
22 Tne Paris method assumes a plastic zone 23 size to estimate this reduced compliance. This is a 5 24 separate method.
s_
1 25 Jumping to results, these are some results 1
I
112 I fs 1 for stainless steel pipe. What we have here is 2 experimental load civided by predicted load versus 3 pipe diameter again. And note, first of all, that j 4 there's a range for the GE method, for example, and 5l this is because a Ramberg-Osgood equation isn't i 1 6l really sufficient to characterize the stress / strain 7 behavior for certain materials.
I i
8l One can choose to fit the data over a low
! 9. strain range, over a high strain range or average it 10! throughout the full strain range. And this gives 4
11l the difference in results using one type of 12{ stress / strain fit versus another type of
) 13!
stress / strain fit. Again, all these numbers are l
14i above this dash line here. Points above this dash 15 line here in effect give cafety factor. Points i 16 above the line are conservative.
17l GE method gives underpredictions of --
18 this is for maximun load. This is for utainless i
j steel. The Paris method also gives good predictions, 19ll 20 and the LBB-NRC method also gives good predictions 2
21 for the stainless steel pipes that we've looked.at.
22 MR. SHEWMON: Now, if we had the ,tainless i 23 steel welds, that would not be true, and when 24 somebody does a code analysis do they --
and I don't l
25 mean INB-3640 or whatever it is, because that's only l
113
,-s 11 tor overlays; isn't it?
(' ~'/
s 2 BI R . WILKOWSKI: No.
3 iiR . BRUST: We'll get into that in a 4 moment. That is, what we do for a welded pipe and 5 aow good those predictions are given the approach 6 that we've chosen. We'll be right there in a minute.
7 1' h e s e are some results for some of the 8 carbon steel pipes, and I note that there's not a 9 range as there was for the stainless steel pipes, 10 because carbon steel pipes tend to give a fairly 11 good prediction, Ramberg-Osgood prediction. The I
12 stress / strain data tends to fit well with a h 13 Hamberg-Osgood curve.
14 lsg a i n , tuin is experimental load divided 15 by predictive load versus diameter. So above'this 16, line is in effect safety factor. We can see that 17' all these points for all methods are above the line, la except for this particular pipe here.
19 And what happened here was the experiment d 20 this was an cight inch diameter pipe. The 21 experiment was performed --
it was under four point 22 bending, and it was performed with a center span 23 that was only 1.5 diameter. It was small compared --
!7 ,j 24 in the Degraded Piping Program we've always used y) 25 four pipe diameters for the center span.
114 4 1 And the point is that apparently end 2 effects caused the maximum load to be reduced. In 3;! effect, this pipe night be considered to be a crack 4, neer a stiff section of a structure, a crack in a 5! pipe near a juncture to a pressure vessel, for 6l example. And the point is that perhaps the i
j 7l estimation schemes need to be improved for that sort i i 8l of situation.
i 9 For welded pipes, some of the ones that i !
- 10! we've looked at, what is done or what --
we looked 4
! 11l at a number of possibilities. What was chosen is to t
12 use the pipe material stress / strain data and the
) 13 weld metal J-R curve for the analysia. Of course, 14j it's really a bimetallic material, and to really do i
15fthe analysis correctly, one would have to go to a j 16 finite element approach.
I 17 dut using estimation scheme, we find that l
1Sl we get conservative results for all three methods, 19: except for the Paris method was a little low here.
20 So the method, with certain assumptions, can be used i
21 and give good predictions for welded pipes.
i 22 MR. SilEWMO N : If one looked at the bottom 23 line of that triangle you could say you quit just
(
24 before you got to where the big pipes are in the 4
25 field. Do you have any comment on the extrapolation
-*e- T--- --- -- , --m- - -
-g i-- --
I l 115 l
-s l' of that to larger diameters?
ks~~-l !,
2 i1R . BRUST: ies. That's a good point.
3 This again was -- this range here is due to material 4 property characterization. .ow, if we're going to 5 choose a material -- well, if we're going to -- this 6 was for the strain range for the Ramberg-Osgood data 7 at hign strains, and so the idea is that if we're 8 going to use this in a code we'll choose the 9 Ramberg-Osgood fit that gives the most conservative 10 prediction. 1. c a , that's a good point. Now, if we 11 used the low strain fit, we might end up up here for 12 tne larger diameter pipes. We don't know.
h 13 MR. WILKOWSKI: We do have one large 14 -l i a m e t e r pipe test we recently conducted with the 15 welds. So we'll be evaluating that in the future.
16 MR. ARLOTTO: llow big?
17 MR. WILKOWSKI: Twenty-eight inch diameter la pipe t ria t was removed from the Nine Mile Point plant.
19 b1 R . BRUST: So those are load predictions, 20 naximua load predictions, t 21 MR. ETilE RI N G T O U : Are the load predictions 22 on these various models supposed to incorporate a 23 margin of safety of some kind or not?
('
') 24 :1R . BRUST: We're looking --
we haven't 25 included any factor of safety.
w---.-.-.--__-. I
i 116 1 t1R . ETIIERINGTON: No, I meant on the
, [N
%') 2 GE/EPRI, for example, on the NRC.
3, MR. BRUST: Are you asking is there a
- I 4 factor of safety already included in that method?
I MR. ET LIE RI N GTO N : Yes. That's what I'm 5l 6! asking.
7 MR. BRUST: I don't believe so.
8 MR. WILKOWSKI: That's just the inherent i
4 9 safety factor in the antlysis.
10 MR. ETIIE RI NGTON : Um-hmm.
11! MR. WILKOWSKI: One point from this is 12j you'll notice that the GE/EPRI analysis method seems r's I
- (,) 13 to have the highest margin of safety, and it was i 6 14l that procedure that was used for developing the flux 15 weld criteria in the IJS-3640 approach for flux 16l welds. So that has some inhertnt margin of safety l
17- in it to begin with, i
18f MR. BRUST: If we briefly summarize the 19l load predictions, we note that the GE method tends 20 to give underestimates of the experimental load.
21 The Paris method is independent of stress / strain 22 behavior of material properties. It only depends on 23 flow stress. And we noticed that the LBB-NRC method 24 seemed to be the most accurate and reasonable.
(
25 MR. BENDER: Before you get off that, the
-,,--r- m-, - - , _ - - - e -o,
117 f,s 1 question that I think continues to pop up is should 2 we be using the EPRI approach or should we be using i
i 3 the NRC approach or some other approach? Do we know 4 enough yet to know which one we should be using?
5i MR. BRUST: For the pipes that we've
- 6) looked at we've always gotten conservative results 7' for both methods. Do we know enough yet? That's a 8; difficult question.
I 9f MR. BENDER: Well, what's the criteria for 10! knowing?
11 MR. BRUST: Well, we found that all of our i
12 i predictions were conservative, and that's one good 13 criteria.
I 14l MR. BENDER: As Dr. Snowmon said, I've got 15' to extrapolate out to some pipe sizes that are a bit 16 larger than these things.
17!!
MR. BRUST: For the welded pipes.
l 18- MR. BENDER: Yes. bo the question is when I
19 I start to extrapolate, where will I be? Will I 20 have a basis for knowing that? Will I be better off 21 or worse off with one of these computational methods?
22 MR. BRUST: Well, we'll give you one more 23 data point.
24 MR. WILKOWSKI: We have a couple very 25 large diameter prototypical tests that are planned
l 1
118 1 to try to insure that if there are extrapolation 3
D Q 2 errors that they would be insignificant. We're 3 looking at very large diameter pipe that has welds 4 in it, both for carbon steels and ferritic steels.
1 i 5j And those experiments are ongoing, 30 we'll be I
6',
updating this work as time goes on.
7j MR. BENDER: And from the standpoint of I
8! the eventual application, I'm not sure that it's
[
9( important for me to be exactly right. I would 10 3 really be looking for conservatism. If the GE 11! method is conservative, will it get ma --
too i
12 i conservative, will it get me in trouble?
I
' 13 MR. WILKOWSKI: That's a concern certainly l 14; that the industry has is that it's so conservative I
15l that they feel that it might be inappropriate at I, I r
- 16! times.
I
! MR. BENDER: So we have to look at what
! 17l, 18} the "at times" issues are?
l 19 MR. WILKOWSKI Yes.
20 MR. BENDER: Okay. Thank you.
21 MR. BRUST: I'll very briefly cover j
22 displacements and to start off, just note that the 23 reason displacements can be important is for 24 modeling a correct piping system, a complex system 25 that has a section of cracked pipe in it, one can
119
,- 1 use an -- estimate that section, that cracked
~ )
2 section of pipe with the load displacement 3 relationship produced from one of these methods anc 4 do a full scale simplified analysis.
5 So the load displacement relationships are 6 important for that purpose and also for a simplified 7 instability criterion. Whenever a surface crack 8 pops through to become a through-wall crack, a lot 9 amount of energy is released, and we've been looking 10 at a very simplified method, which is based on 11 knowing the load displacement relationship of a 12 through-wall cracked pipe to determine whether or h 13 not instability will occur.
14, I'll just briefly show some of the 15 predictive results and then move on. This is load 16 versus displacement. The circles are experiments. l l
17 And we've got five methods, and we note that, as we 18 noted earlier, the GE method tends to underpredict 19 the experiment. The Paris and LBB-NRC methods are a 20 little better, and then some of the methods that 21 we've developed here seem to give better predictions, 22 and I don't want to --
there's also --
that wasn't a 23 bad prediction. These are typically what we obtain.
f]
j 24 ..e tenu to underpredict the experimental load 25 displacement relationship using these J-estimation
120 1 schemes.
7s
'N ) l So to summarine, again the GE method 2t l
3' continued to underpredict the particular experiment.
4 All of the methods are accurate for small 5 displacements as one would expect, for small amounts 6l of crack growth. And some of the modified methods 7' seem to give the most accurate predictions.
8j And to conclude, I would like to note that 9
i J tearing theory we all know deteriorates under I
10; certain conditions after a certain amount of crack i
11l growth and under certain nonproportional loading 12 cases that the crack starts from my experience 13 throughout its lifetime. Go the idea is to use an 14; engineering estimation scheme based on the J 15 integral as long as we get --
as long as these i
16i methods give us cons rvative predictions.
17l In the meantime we continue to search for i
18; estimation type parameters that can be useful in l
19 engineering judgments. There are a number of them, 20 and one that we've been looking at here is another 21 type of integral that we know gives good predictions 22 in conjunction with finite element analyacs, but 23 there is no estimation scheme available yet for 24 those.
25 11R . SHEWMON: When you talk about the
-- . - . - .- - -- . .- .. = - . . - _ . . - .
i l 121 1
1 theory deteriorating, is this at such large crack O 2 opening displacements that the pipe would nave 3l leaked badly, but you're concerned about whether it 4 would tear unstably or what?
I i
a 5 MR. BRUST: It's even more fundamental I
6l than that. The estimation schemes are based on 7! being able to --
a Ramberg-Osgood fit of the 1 1
! 8 stress / strain data, not so-called deformation theory
{ 9t plasticity applying. Whenever the crack grows, what l
10l really happens is certain unloading at the crack tip.
11! So deformation theory does not apply.
12l Flow theory of plasticity applies, and
) 13 hence, that in effect is the problem. The theory 14,i itself la no longer valid. J integral, for example, l
15l becomes path-dependent once crack growth occurs.
16 MR. ETHERINGTON: Can you say in a few 17 words what the T integral is?
18 MR. BRUST: Oh, it's something quite 19 similar to the J integral.
20 MR. ETHERINGTON: And the difference?
21 MR. BRUST: A correction term to account 22 for crack growth and nonproportionality.
23 MR. dTHERINGTON: Okay.
24 MR. SHEWHOW The last time 1 saw 25 something like that with a star on it, I think it
._ ~ _ _ _ .
122 I was a C, and it had creep in it. Does this have
,O 2, creep in it?
I 3! MR. BRUST: No, but a rate form of this, I
4lif we take --
delta T, delta T star divided by an l 5ll increment of time, then it becomes a rate parameter, i
- 1 6 which C star is. T star is not related to C star,
! t j
7l but one can also use it in creep situations, 8! MR. HUTCHINSON: In the test cracks in the i
9 welding material in the pipes, does the plastic zone l
10l grow in size considerably larger than the width of llI the welding zone?
1 i 12' !! R . WILKOWSKI: Yes.
13 MR. HUTCHINSON: Several times the width?
I .
[ 14' MR. WILKOWSKI: Yes, yes, many times.
l l
, 15! MR. HUTCHINSON: Okay.
1 l
l 16! MR. BRUST: Any more questions?
17 MR. WILKOWSKI: Okay. Let's sec. Our t
i 18; next speaker is Greg Kramer.
I 19{ MR. KRAMER: My name is Greg Kramer from l
20 Battelle Columbus. As we've defined earlier, the 21 complex crack is a combination of a through-wall and
! 22 a surface crack. There are two prime motivations i
23 for going ahead with an evaluation of complex cracks.
(} 24 25 The first in that this is a very severe flaw
- geometry that nas been found in service. As Gery
123 1 mentioned, the Duane Arnold flaw was essentially a O 2 complex crack, a long circumferential crack that i
3f gradually grew through the wall thickness.
l 41 The second important point was that an 5i understanding of the behavior of complex cracks will 6 also help us in being able to predict the stability 7 or instability of long surface crack pipes. As a 1
8 surface crack grows through the thickness, it 9 essentially becomes a complex crack, and the amount I
10 of crack jump is dependent upon the behavior of the 11, complex crack.
12! So maybe we can jump right to the results.
) 13i The next section just kind of summarizes the major i
14! results of this task so far. The first point is I
15; that limited crack instabilities did occur in --
l 16j well, let me back up befc*re I mention the results.
171 There were six pipe fracture tests i
18 conducted in this subtask. Two were on A10; carbon 19; steel, two were on Inconel 600, and two were on Type 20 304 stainicas steel. We have a test matrix of six 21 tests, two of three different materials.
22 In the carbon stool test we saw limited 23 crack instabilities as we've seen in other ferritic 24 materials. These occurred once past uaximum load.
25 The A106 Grade B pipes failed below
124 1 net-section-collapse predicted loads, even though
\ .
2 they were fairly small diameter. The screening 3 criteria that has been shown before shows that most 4 of the small diameter pipes fall in the fully 5 plastic zone of that screening criteria, thus they 6l should meet net-section-collapse criteria.
7 The J-R curves or the J-M, modified l
8l resistance curves from these complex cracked pipes i
9 were lower, significantly lower than 20 percent 10 sidegrooved C(T) specimens. So in trying to compare 11 snall scale specimens to pipe tests, there's some 12' work that needs to be done.
) 13 The J-estimation schemes do not account at i
14; this time for radial crack driving force. We are t
is! casentially using current schemes to analyze the 16fcomplex crack as a through-wall crack essentially.
17l There is no crack drivi.ng force accounted for in the i
18' radial direction or due to the internal surface 19 crack. It's just --
we're analyzing it currently as 20 a through-wall crack growing around the 21 circumference.
22 To try to correct for this problem, we've 23 tried to coming up with some empirical factors or 24 cmpirical corrections to help make the through-wall 25 crack --
complex crack predictions a little more
i i
125 1 realistic.
2 The final point is that the J-estimation 3, schemes that have been used up to this point, such '
I 1
4 that Dr. Brust mentioned in his talk, have worked 5;i very well up to maximum load in these complex crack 6! tests, but once we get past maximum load they l
7: overestimate the load substantially.
4 8i Currently we're looking at some other l
9l techniques to handle this behavior past maximum load, 10 but up to this point we've gotten very good results i
lli with some of the current J-estimation schemes.
I l
12l The future plans for this subtask are to
) 13 look at further refinements, improvement to the 14}J-estimation schemes for complex cracks. He are 15 currently writing a topical report which will be 16 finished shortly, and some larger diameter 17! instability tests are planned. These will be done 18 on 16 inch diameter 304 stainless steel pipe.
19 I'll skip over this real quick. This is a 20 comparison of the Duane Arnold flaw to the complex 21 crack geometry that we're using in this experiment.
22 As you can see, this was a very severe flaw, 50 to 23 80 percent of the wall thickness in depth.
24
( MR. S tic WMO ti Your machine cracks are how 25 wide?
126 1 MR. KRAMER: The internal surface cracks 2 or the through-wall cracks?
3 dR. 311CWMON : Well, you had a ran ch i n e 4l crack going all the way around that, and I asked i
5! what the width of it was.
i 6l MR. KRAMER: The internal surface crack 7 that goes 360 degree's around the circumference is 8l about an 8th of an inch. Through-wall crack is 9i approximately about the same thickness, I guess.
l
- 10) A post test examination of the fracture 11l surface --
this is a fracture surface of the 304 i
12 stainless steel material. Now, I've kind of
) 13
]
oriented this 90 degrees from the last page. So the i
14l through-wall crack is really over on the left-hand 15{ side of the page. But the two significant points
+
16l from the fracture surface that we see are the crack 17l growth is very nonlinear through tne thickness. You 18! get substantially more crack growth on the inner 19 ! surface than on the outer surface, and this suggests 20 . that there is a significant amount of constraint, a 21 significant amount of contribution due to the 22 internal surface crack.
23 These marks were produced on the fracture
(} 24 25 surface by a number of unloadings during the pipe fracture test. So nonuniform crack growth and
127
, fs 1 significant effect of the surface crack.
[
- 2 This is a comparison of low displacement t
l 3 record from four different pipe experiments. These
! 4 are all the same diameter, six inch diameter, all 3 5l the same wall thickness.
I We're looking at two 1 .
6l stainless steel and two carbon steel pipe tests. .
As t
7 you can see, the stainless steel has a slightly I
8 nigher load-carrying capacity, much lower ductility i
j 9l present. In the case of the two stainless steel 10 tests, we have two different D/t ratios. In the 1 t 11 case of the carbon steel test we have approximately i i J !
) 12j the same D/t ratios. So this carbon steel curve 4
) 13 should correspond to this stainless steel curve.
I
] 14 l This carbon steel curve should correspond to this i
15l stainless steel curve. The small X's designate the l
16( initiation loads.
i 17: ;1R . Sli B W M O N : You say the stainless steel
- 18 has lower ductility?
s l
19 11K . K R A!!E R : I'm sorry, the carbon steel 4 i a
20- has significantly lower ductility than the stainless.
21 MR. Sti E W M O N : All right.
22 MR. KRAMER: Also of significance is the --
j 23 again, you can see the limited crack instabilities 24
( that occurrud in the carbon steel, the A106 Grade B.
25 This was present in both cases. Loud crack pops,
- n,., . , , - . , , - , - - - - - - - . - - , - - - - .,,,..--.__.,-_-- -,,., -, ,,.
~,--,---- -
128 cm 1 jumps in the EP signal and decreases in the load.
2 This is a moment rotation diagram, very
- 3. similar to the load displacement diagram presented 4 carlier. In this case all three curves of 304 S stainless steel pipe, all the same size, same wall 6! thickness. The upper curve is a through-wall crack 7 ! with essentially no surface cracks. The D/t is zero.
8l The next two curves are complex crack 9l tests where the D/t is increasing.
What we see in i l 10! this is --
the significant point about this curve is 1 1
11 that the complex crack significantly reduces the 12 tearing resistance of the material. The area under
(")3 s s 13i this curve is significantly lower than the area l
14 under the tnrough-wall crack curve, such that when i
15 we get into the analysis using the J-estimation i
16l schemes we should see a much lower tearing 17 resistance for the complex cracks as compared to the 18,! through-wall cracks.
I 19f MR. ETilE RI N GTON : What is this steel?
20 MR. KRAMER: I'm sorry?
21 MR. ETHERINGTON: What is this steel?
22 MR. SilEWMON : That curve, what's the 23 material?
(} 24 25 MR.
stainless steel.
KRAMER: Oh, all three are 304
. - ~ _ . . _ - . _ _ , , - . - - - -
l l
l l
129 ;
1 MR. ETilE RI N G TO N : Okay.
vx'l 2 MR. RODABAUGH: If you had a surface crack 3 on this same picture, would it ched any light? I i
4! Would it be in between the through-wall and the i
f 5l complex?
6! MR. KRAMER: No, surface crack -- well, I
7 depending on the length of the surface crack it al would probably be much higher off.
9! MR. WILKOWSKI: It's a function of the f
i 10l depth of the surface crack as well, so you could get i
11' anything in between, either it's greater or slightly 12 less.
) 131 i
MR. KRAMER: Yes, the surface crack curve i
14' could be substantially higher or somewhere in 15l between.
l 16i MR. IiUTCHINSON: But, for example, if you 17 took that middle curve and eliminated the fully i
18, through crack, any idea?
i 19 MR. WILKOWSKI: It would have --
20 11R . HUTCHIllSON: It would be way up?
21 MR. WILKOWSKI: Yes, right.
22 MR. KRAMER: Yes. A quick comparison of 23 the experimental results compared to limit-load I'N 24 using net-section-collapse analysis shows that in D
l 25 general --
okay. In this figure I'm showing the t
. _ ~-
130 1 maximum load compared to the net-section-collapse 2t load, versus the six different pipe tests.
l predicted 3 The first two tests, although it's not 4j marked, the first two are stainless steel. The next i
5l three -- well, the next two tests are the Inconel 6! material. One of the tests is presented here twice 7 for two different analyses. And then the final two 8,l curves are for the carbon steel. ,
i 9! What we see from this is that the carbon 10l steel typically doesn't meet limit-load conditions I
in these tests. There are three different flow 11l 12l stress criteria being evaluated here, the average of
) 13 yield ultimate 1.15 times the average and 3 S of M.
I 14j So the carbon steel using the material property I
15 values is not reaching the net-section-collapse 16 limit-load.
17l From the pipe fracture test we were able 18 to use the Eta factor technique to produce a J-R 19 curve from the pipe fracture experiment. In this 20 case this is from two of the Inconel pipe tests. It 21 shows --
the one curve is for a D/t of .34. The 22 lower curve is for a D over 2 of .61.
23 What we see in comparing the J-R curves is
(} 24 25 that there is a constant ratio all the way up to J-R curve. This lower curve here is consistently 65
131 w 1 percent of the upper curve, and this trend was found
's) 2l in all three materials. Regardless of the material, J we found this consistent trend due to the difference i
4 in thickness.
i 5f The other point to note on this figure is 1
6! the dotted line at this point is the J-R curve from 7I the small scale C(T) specimen test. So we're 8 comparing the two pipe fracture tests to the C(T) 9 specimen. We're seeing that -- trying to 10' extrapolate the C(T) specimen in this particular I
11' case would greatly overpredict the J-R curve of the 12, pipe test. So that there's not a real good
) 13 agreement at this time between the small scale 14fspecimen test and the pipe fracture test.
l 15! MR. HUTCHINSON: Yeah, but you wouldn't 16 expect that.
17 MR. KRAMER: bo.
18 MR. WILKOWSKI: That's because the 19 estimation schemes that exist all neglect the radial 20 crack driving force component.
21 ilR . KRAMER: Right.
22 MR. HUTCHINSON: Yes, but just the ;
23 geometry is totally different; isn't it? In the ;
1
( 24 small specimen you're talking about running a crack l
25 straight ahaad, running from face-to-face. In this ;
l
132 1 case you've got a specimen that's already partly 2 cracked, a deep crack along --
I mean I suppose if l
3' you wanted to try to make a correlation you could do 4 the same thing with a small specimen, just --
I'm Sl not i
1 suggesting you would, but --
6i MR. WILKOWSKI: Well, the C(T) specimens 7l' are sidegrooved specimens, but again the crack is 8,
t always perpendicular to the loading, the principal 9f loading.
10l tiR . HUTCHINSON: Yes.
I lli MR. KRAMER: From the results of the 12l complex crack experiments, we have compared the
) 13l complex crack test to the through-wall to some i
- 14. specific through-wall crack pipe tests, and what we 15: find is that we can come up with a factor or a 165 correction relating the J-R curve of the complex 17fcrack to the J-R curve of the through-wall crack as 18 a function of the D/t.
19 . On this graph are plotted four 20 experimental data points. The line in between the 21 two --
between the solid points illustrates the 22 scatter in the data. This shows that even for a i
23 very shallow D/t ratio we see significant drop in ,
l f'N
'w) 24 the ratio of the complex crack J-R curve to the l
25 through-wall crack J-R curve.
I 133 i
,- 1; So this is being worked on right now, and N)
2 i we're pursuing this, anc this may help us in coming 3 up with an cupirical correction for the complex 4, crack data.
5 And then quickly, if I can show Dr.
6 Brust's results using the different J-estimation 7 schemes compared to the complex crack, load versus 8 displacement, the dotted points are the experimental 9 data. vle see that up to maxinua load -- uost of the 10 estimation schemes did a pretty good job of 11 predicting the load displacement history. The 12 GE/EPRI data is the most conservative.
h 13 Once you get past maximum load, the curves 14 are greatly overestimated. But then again, we're 15 trying to use a very small scale C(T) specimen J-R l 16 curve to predict the results of the pipe test. So 17 we're trying to extrapolate from a very small amount 18 of crack growth to large crack growths in the pipe 19 test.
20 And to try to correct this, we are 21 currently looking at a new approach using a crack 22 tip opening angle. Tnis seems to be showing much 23 better agreement on this unloading portion of the
[]
x,i 24 curve. This procedure is being modified at this 25 point and being looked at further.
l
134 gs 1 So limited crack instabilities have 2 occurred. Carbon steel has failed below 3 net-section-collapse predicted loads. J-R curves 4 were lower for the pipe tests than for small scale i
5l specimen tests. Radial crack driv?ug force is not 6! taken into account in this investigation, and we see 1
7' good predictions up to maximum load and looking at I
8l predictions past maximum load, ways of correcting 9l the current analysis.
10 Are there any questions? Okay. I believe i
11{ Victoria Papaspyropoulos is next.
l 12:
i
- 4S . PAPASPYROPOULOS: I'm Victoria
) 13 t
Papaspyropoulos, and I'm with Battelle also. I'm i
14i just very briefly going to talk about some of the 1
15 finite element analysis work.
16 ,
All finite element analysis in the l
17 Degraded Piping Program are being performed for two l
l 18 reasons, and that Dr. Ahmad explained to you earlier.
19l The first reason is to assess the accuracy for i
20 J-estimation schemes which are currently used, and 21 the second reason is to address some specific issues 22 in many of these topics.
23 We see in this view graph and in the f'N V
24 second view graph is a list of the kind o. analysis 25 we are doing using the finite element method. And
135 gw 1 this includes compact tension specimens such as 1T, 2 3T and 10T of size 304 stainless steel. They also i
3! include welded C(T) specimens with cracks in both 4 the TIG and the submerged arc welds.
5 Also they include full width face notch 6fspecimen which Mark Landow talked to you about 7 earlier this morning, and then we also have some 8! three-dimensional geometries, such as pipes which 9l don't contain any welds, but also pipes with welds 10 in the TIG at d in the submerged arc.
I 11 And also we note here that there are some 12 ! items that have stars on them, such as this one and
) 13 that one, the previous plot as well as. And what i
14! this indicates are problems that were included in l
15 analysis round robins which we are involved in.
16 The first round robin was already
! 1 17' completed and took place last year, and it involved 18 the 10T compact tension specimen as well as a 16 l I
19 inch pipe with a through-wall crack, and I'll 20 briefly show some.results of the round robin a 21 little bit later.
22 Overall, the significance of the finite 23 element analysis work can be summarized in this view
(} 24 25 graph, and it's focused first to assess the accuracy of the estimation schemes, then to assess the
i 136 x 1 applicability of small specimen J-R curves to large 2' amounts of crack growth, assess geometry dependence 3 of J-R curves, and lastly to assess the effect of 4 weld size relative to specimen size on fracture 5: nochanics parameters.
6: Now, as I mentioned, we did several 7! compact tension specimen geometries, and one of them I
i 8; is seen here, which happens to be for the 3T compact t
9l tension specimen. The main objective of the l
10l analysis of the compact tension specimens was to 11l assess the J tearing theory for extrapolating small 4
1 12l specimen J-R curves to large specimen J-R curves, 13 but the main purpose of the finite element analysis i
14l was just to assess the validity of such estimation 15j schemes.
t l 16! 'J h a t we do in the finite element analysis, 17: we are trying to simulate experimentally described i
18j applied crack extension versus the displacement i
19j record, and follow this particular record in our 20j analysis to simulate crack growth. And a typical 21 record like this is seen here for the 3T specimen.
22 During the analysis we calculate the load 23 and the J, but, of course, the overall objective is
( 24 to obtain the J-R curve. dhy this is done and the 25 calculation, in other words, the load displacement t
137 1 curve, is in order to assess the model inaccuracy by Cy uJ T
2 directly comparing it with an experimental 3 displacement record. And such co rapa r i s on is seen 4, here for the 3T specimen.
5 Now, considering large deformations in i
6 Type 304 stainless steel and amount of crack growth 1
7 of almost three inches, this comparison is quite 8 reasonable.
9i Next I just show a typical J-R curve for i
l 10 one of the three specimens, and this is for the 3T i
11! specimen. We see J-R curves computed using the J 12l integral contour with two different contours, onc
) 13; near the crack tip and one away from the crack tip, s
i 14 and also we see on the same plot, the J computed by 15 an estimation scheme.
16l Now, as 1 mentioned earlier, we had a 17 round robin being organized, and there were two i
18l problems in it, one of which was the 10T compact i
19! tension specimen, and the results in terms of the 20 load displacement curve as shown here for a total of 21 six different participants, but they are shown here 22 with the X's. iie can see that really this is not 23 very significant and overall most of the solutions
( 24 are in really good agreement with the chart.
25 The key result of the J-R curve as
138 gx 1 obtained from all the different participants is 2 shown here. And we also have on the same plot the 3 J-resistance curve obtained using two different 4l definitions of J. And this is shown with the dashes I
5 and the solid line here.
6! There's something a little bit interesting 7 about this curve, and it is that the Battelle 8fresults together with the participant 2 and 4 show i
9{ quite good agreement with the J-R curve that was 10l obtained using the J sub T definition of J, and I 11' i
guess there is a good reason for it because 12 participant 5 or, I'm sorry, participant 3 here did
) 13 not have modeling, and since he did not take into 14j account the history of things in the crack growth 1
15l process, he agrees better with the J-deformation l
1GI theory than J.
17j aoving really quickly to the next view 18> graph, what this is showing is that we did analysis I
i 19l on C(T) specimen geometries, and what we've have 20 done here is plotted all of them together on the 21 same plot and we've compared them with a J sub M 22 estimation scheme results, and we can see the same 23 trend, but the far field finite element J-resistance
(\
x_)
24 curves do agree quite well with the J-resistance 25 curves obtained by J-M estimation schemes.
139 fx 1 The second problem in the round robin was N
2 the pipe. In four point bending it involved the 3 through-wall crack. And again, I'll just show you 4 quickly the comparison in terms of load displacement 5 i of the different participants. Four people solved 6, this problem, and we see that the crack initiation, t
7 most of them are in very good agreement, but i
8j afterwards all the solutions are under particular i
9i load.
I 10l Pretty much the same is true in terms of l
11i the J-resistance curve, with crack initiation most 12f solutions are very similar, but then quickly after s ) 13l some amount of crack growth they deviate from each 1
14l other, and they tend to be lower than the J l
15 estimation scheme, J-R curve.
16 ti o w , moving on to another important issue 17l being studied here, the analysis of cracks in welds, l
18! The main objective in analyzing cracks in welds is t
19 to determine whether the available engineering 20 estimation schemes can be used to study this 21 important class of problems. As an example here we 22 see the finite element mesh of submerged are weld, 23 10T compact tension specimen. What is shown as a fh N) 24 shaded area here on the bottom is representing the 25 weld material, and the dashed lines represent the
l 140
[s '
1 different contours that were used in order to obtain NJ' 2 the J integral.
3 The key result from this analysis was this 4! set of J-R curves, and why it is important is i
S' ,
because it's showing that tne J-R curve obtained 6l using the contour integral and the virtual crack l
7l extension integral, J is in very, very good 8l agreement with the J obtained by estimation scheme.
9; Additionally, we see on the same plot the i
10l resistance curve obtained using some other fracture 11 parameters that Dr. Brust talked earlier, namely, 12l the TP star and the J whien are also evaluated
{N A )
m 13 from -- one hundred percent from the finite element 14i results.
15! We also analyzed welded pipes, and we i
16: analyzed the experiments of welded pipes from both 17 Battelle and David Taylor. We won't have time to go i
18l into describing any of the results, so I guess I'll 19{ just show you a typical result, which is quite 20 interesting, that we obtained from the analysis of 21 the welded compact tension specimens. And what it 22 shows here is that the initiation toughness seems to 23 depend on the ratio of the height of the weld
(} 24 25 material relative to the size of the base metal.
For example, in this p l o t- we have
141 m, 1 submerged arc weld data and TIG weld data, and for
%~)
2j. the case of the submerged crc weld data we see I
3! initiation toughness seems to decrease and then l
4l level off with increasing d/h where this is the 5 height of the weld, and H is height of the specimen, 6lwhereas in the case of the TIG welded data the 7l initiation toughness increases, and then it drops i
8l off dramatically for the smaller specimen.
9 MR. SHEWMON: Now, to vary D, did you vary 10ll the heat input to the weld? Is that it?
11 MS. PAPASPYROPOULOS: You mean in this 12 analysis? No.
O)
(s 13j MR. WILKOWSKI: There are different size i
14! specimens all with the same weld, so that H is
- 15. changing and D is staying constant.
I i
16l AR. SHEWMON: Okay.
I 17l MR. HUTCHINSON: And J here is measured by l
18l the usual technique area under the curve and so I
I 19 forth.
20 MR. WILKOUSKI: It's done both ways by the 21 finite element, which agreed fairly well with --
the 22 usual Eta factor analysis on there was good 23 agreement between the two, so these results appeared 24
( to be reasonable.
25 MR. HUTCHINSON: But the results we saw
142 1
,, 11 earlier took the contour outside the -- essentially
- i l
~
'u] I .
l I
2l outside the weld.
l 3 1 dR. WILKOWSKI: Tnat's right.
4! MR. HUTCHINSON: Have you tried to look to S see if there's any correlation with the calculated 6 value of J inside the weld?
7 MR. WILKOWSKI: Is that you or Mich? Mich 8 can answer.
9 MR. NAKAGAKI: You said weld -- my nane is 10 Mich Nakagaki. He asked -- answering to your 11 question --
12 .3. S H E ii M O N : The rest of us are
,n
() 13 interested, so speak up if you would.
14 MR. NAKAGAKI. I'm trying to explain the 15 correlation between the J, the contour, inside a 16' weld and outside a weld, and as you expect, up to 17' initiation it is good given the weld, and the best 18 interface there exist and across the tip, since this 19 is parallel due to crack growth and is valid.
20; r1R . HUTCHINSON: So you're saying that 21 prior to any crack growth --
22
- MR. uAKAGAKI: Yes. ,
l 23 MR. HUTCHINSON: --
the J values you l
("'5) 24 calculate on contours inside the weld are very close l 25 to those calculated --
t-
143 gg, 1 MR. NAKAGAKI: Essentially that is, yes.
t i i
'V 2 MR. HUTCHINSON: So that certainly 3 couldn't explain this.
4 MR. NAKAGAKI: No, not this.
I 5 MR. HUTCHINSON: Thank you.
6' MS. PAPASPYROPOULOS: I have two more view 7] graphs, what we've shown as problems for the second 1
8l Battelle/NRC round robin which is going to be held 9lat the end of this month in Chicago. And this i
10 involves the full width face notch specimen which we lli saw earlier, and is used to develop J-R curves for 12! the through thickness direction. And it also I
r" (N) 13; involves a three-dimensional geometry, which is a 14! five and four point bending, contained in a long i
15l internal surface crack. And again as I said, the I
16l results will be presented for this problem at the 1
17l end of this month in Chicago at the conference.
k 18j MR. WILKOWSKI: The objective --
19 MR. RODABAUGd: On that sketch, that i
20 picture you just got there, as part of your round 21 robin problem will you describe in detail the ,
22 saddles, the axial?
23 MS. PAPASPYROPOULOS: You mean if in our 24 finite element model we're going to include (v}
25 geometrical details of saddles?
144 rx 1 MR. RODABAUGH: That's my question, yes.
No 2 MS. PAPASPYROPOULOS: I don't know what i
l 3, the progress of that plotting in such detail is j 4 going to be. You might want to ask --
5 MR. WILKOWSKI: I think we've tried to 6! discuss as much experimental details so that they I
7l could properly do the boundary conditions and the 8 finite element analysis. They've asked a few more 9j questions since then.
10 MR. RODABAUGH: Well, I raised the i
11' question because I've been doing some speculation 12 j that part of the results in the experiments depend 13 upon the details of those saddles and, of course, 14; the spacing between the saddles.
15! k1 R . WILKOWSKI: Yes. Of course, that's --
1 16i MR. RODABAUGU: So I would hope that the 17l finite element analysis would duplicate the saddle 18j conditions that you want them to.
I 19f MR. WILKOWSKI: Yes. Okay. And the i
20 ! ob]ective of this surface crack round robin was to 21 provide us with guidance to the finite element or 22 the J-estimation scheme that Jalees Ahmad had talked 23 about earlier.
(V5 24 That pretty well wraps up the technical 25 discussions that we had.
145 1 f1R . SilEWMON : Can you tell me who the (g
2 groups were that took part in the round robin?
3 dR. WILKOWSKI: Jalees, do you recall the' 4 different groups?
i 5 MR. AHMAD: All the solutions that we got 6' were outside this country. We were the only ones 7i from here. But we had solutions from liPA, Stuttgart.
8 ! We had two or three solutions from Japan, University i
9! of Tokyo, and -- who else was there?
t 10' MR. WILKOWSKI: Kawasaki.
11 MR. AHMAD: We had solutions from GRS I
12 Germany.
(~h
%%j 13 3R. 3HEWMON: Is GRS down in Freiburg or i
14l who is GRS?
15 MR. WILKOWSKI: That's Risodi who is 16 located in Bonn or Cologne. I can't remember.
17, MR. AHMAD: And we had solutions from VTT 18 Finland and --
l 19 MR. WILKOWSKI: France.
20 MR. AHMAD: France, LCEA Carderache.
21 IIR . SHEWMON: Okay. Thank you.
l 22 MR. WILKOWSKI: I just want to make one 23 clarification point. There were a couple questions i
( 24 about the instability of stainless steel, and the !
25 comments that we made was that we could have the i
146 es. 1 instability in stainless steel, but those are 2 relative to the types of test conditions. In our 3 tests we have fairly large crack sizes. We put on 4 fairly large loads or have large elastic compliances, 5 large amounts of stored elastic energy, so that i
6; we're forcing an instability to occur so that we can 7fthen verify the instability analysis, that indeed dl that it will work for stainless steel.
i 9 That's not to say that instability will 10 j occur in a plant piping system. It's just under our i
11 test conditions, okay?
12 MR. SHEWMON: Have you found the sulfur
) 13 1
content of the steel 106 yet?
14! MR. WILKOWSKI: The sulfur content of the l
15 steel was .015 percent.
16 MR. SHEWMON: The Japanese will tell you 17 they can't find stuff with that high sulfur to do 1
18 tests on anymore, because they can't confirm them.
19 MR. WILKOWSKI: Probably will say that.
20 MR. BENDER: Is that measured or spec?
21 MR. WILKOWSKI: That's a spectrographic 22 analysis. We also have the mill analysis too.
23 Okay. And within your handout I have a
/'N 24 list of other aspects of the Degraded Piping Program u) \
l 25 that we did not have time to cover, also a list of '
i
8 i
i
! 147 ;
i
] 1 topical reports and other things that are ;
! l
]
2 deliverable to the program, so you can look at that I i
]
3 j within your leisure.
) ;
j 4[ And if there aren't any other questions,
- 5l then we're pretty close to lunch.
i -
6! MR. SPEWMON: We're adjourned for lunch.
l 7; - - - - -
1 i i
l 8l Thereupon, the luncheon recess 9 was taken at 12:05 o' clock p.m.
i 10f - - - - -
i 11 i
i .
j 12; i
13 t i j
14l i
{ 15
! 17 '
l f
i 181 4
l l 19 l
20 21 l
i 22 f
i 23 l
24 25
i 1
148 l
i gx 1 Tuesday Afternoon Session l s 2 July 1, 1986 l 3lI 1:10 o' clock p.m.
i 4il I
5l MR. HAYS: My name is Rich Hays. I work 6 ,for David Taylor Naval Snip Research and Development 7 Center, Bethesda, Maryland. The title we are ,
l 8 conducting for the NRC is fracture mechanics 9 evaluation of LWR alloys. My coworkers are Michael 10 .
Vassilaros and Dr. John Gudas.
11 To briefly touch on the objectives of our I 12 l progran, the first and foremost objective is the i
f%)
s 13 ! piping materials fracture characterization, that 14 includes what Battelle is doing, conduction of i
15 l testing of mechanical properties of material, l
16 , general factor toughneau of both compact laboratory-i 17 ; type specimens and full-scale pipes and also an I
18 lavaluation of the tearing instability behavior of i
19 lthe materials.
i 20 Some of the otner objectives are to 21 support the ASME and their code flaw evaluation 22 development for piping, also to develop specifically 23 a new fracture mechanics analysis for pipe geometry, f% 24 q} tuat is a modified J interval of the pipe geometry, 25 and also to then look and assess the geometry effect
149
~x 1 of fracture analysis, that is to look at the more J 2 detailed study at the JM for compact specimens as 3 ,
well as the use at compact specimens to model pipe 4 behavior.
l 5! tid . RODABAUGH: Pipe geometry as you are l
6lusing the term in the previous there, it's , straight 7 pipe?
8 Mn. HAYS: Still straight piping, yes.
9 MR. RODABAUGH: Is any part of your lo t program directed to anything other than use of 11 l straight pipe?
12 '
.iR . HAYS: Not at the present time. He
/"'N k)w 13 are looking at both welds and base metal but only 14 , straight runs of pipe. We have in the past looked 15 at several piping materials doing a laboratory lo fracture mechanics analysis, developing JR curves i
17 which will then be fed into the data base. Some of i
18 l those materials are AlOo steel and the DTA weld of 19'that material, type 304 stainless steel and the GTA 20 sela of that material.
21 MR. SHEWMON: Does Grade 70 mean the 22 ultimate is 70 KSl?
23 MR. HAYS: I don't believe so. Just a
( 24 materials specification.
25 i1R . R O D A B A U G il : Grade 70 corresponds to
150 rx 1 the specified model strength.
2 , MR. il AY S : Mininum. And specific piping l
3 !research, we have looked at four different types of l
4 jfull-scale piping. Eight inch diameter A106 Grade B 5 l steel base metal pipe. We are currently looking at 6 l inch diameter welded A106 steel, four inch 304 7 stainless steel and four inch diameter welded-type 8 stainless steel.
i 9 MR. SIIEWMO d : If we can stop there for a 10 ; minute. I don't care who answers, but the 160's, is I
11 .that the carbon steel that are steam lines but not I
12 ! primary system, is that --
13 21 a . 11 AY F I E L D : That's not always true.
14 The large diameter, 516 Grade 70 is a 106 Grade C I
15 ;and tends to be stainless steal glad and used cold 16 l leg and hot leg crossover in B & W plants, l
17 l LIR . SiiCWMON: But that's not the 106?
18 ! MR. MAYFIELD: Not that eight inch.
I i
19 j f1R . 311E Wi10 N : The 106 which is more a 20 carbon steel lower alloy is not used as a primary 21 piping. From wnat you said isn't in disagreement 22 with that, is it, or isn't it?
23 MR. 11 AYF I E L D : Not that diameter it's not r%
24 used as a primary piping. The larger diameter,
(}
25 heavy wall stuff, 36, 40 inch. It's 106 Grade C of
l 1
151 I
1 the 516 Grade 70.
MR. RODABAUGH: The secondary system, is l
i 2l 3 lthere anything except 106 used?
! I t
j 4 i i
MR. MAYFIELD: In some plants we used Slo 1
l 5 , Grade 70, for which I don't know what it ends up as 6
{wnen it's rolled and welded. Do you have it?
7 MR. HAYS: That's 160.
I i
a dR. ZiA Y F I E L D : I can't tell you off hand j l
} .
It would be essentially we f l 9jwhat a spec would be. --
l 10 !are talning, they all boil down to essentially 106 l
4 i
i i
11 ! Grade B kind of materials, and then the Grada C has !
l !
v ,
j 12 i a little bit higher ultiaate strength. ;
13 I! R . ETHERINGTON: It is a seamless. 106 I i :
j 14 ; is a seamless type. That's not made in sizes more {
i j
15 'i than 24 i nc he s , I believe. Anything over 24 inches !
I l i
- lo is a welded plate. l l
! i
] 11 l MR. t1AYFIdLD: The B & W plant showed ,
I t l
l l
13! Grade 106 I C.
l
- 19 MR. ETHERINGTOJ
- That's not seamless, i l
i 1
l 20 though.
l 21 IIR . S H E rh40d : At one time you could get ,
22 seamless 30 inch pipe.
, 23 MR. LIAYFIELD A piece of 106 Grade C cane 1
i out or Case Western that was seamless, and it's 36 l 24 25 inen, and it had -- it was from the mill. The mill i
._ ~-_ _ ____ _ _ _ _ _ _ -. _ _ - , _ . _
i l
152 1 certificate identifies this as A106 Grade C.
(xT
'%s' 2 MR. 11 A Y S : What I would lika to ao here is 3 just take you through an overview of each of these 4 ;four specific areas. Begin with the eight inch l
5 l diame ter A106. Objective here was to investigate r
6 the fracture toughness, first of all, develop a JR i
7 '
curve for the A106 steel and then look at the 8 l comparison of the pipe resistance curves with l
9 compact laboratory-type specimens.
10 This is a schematic similar to the one i
j ll ' Mark showed earlier showing the orientation of our I
12 coupact specimen and Charpy specimens as well as the
[^h
\_/ 13 circumferential crack in the pipe.
3 I
14 l We alsa did Tensile specimens oriented I
15 Iwith the long axis, long axis of the pipe. This 16 jaat snows che mi c a l composition on the Tensile 17 fmechanical properties typical of A106 Grade B.
ld Charpy evaluation we did, which turns out i
19 to be important now because of the new flaw 20 avaluation procedure for ferritic piping being 21 proposed before in Section 11 which includes a 22 provision for the use of five times the upper shelf 23 Charpy energy as a J initiation or J1C value to put
[h ss' 24 into the code in the case of not having specifie 25 material data on a given pipe.
, - - _ - . - _ . - - . . . . . - - -. . - ~ . -._ _. .
l l 153
\
l lid . S H C W 10N : Five times the upper energy? I i
2 MR. HAYS: That's correct.
., 1
! 3i Md. SudWMou: Five times a hundred and ten .
< l'
] 4 in this case?
5 ;1d . HAYS: That would be correct. That is 6 l an input as J1C value if you did not have a specific f !
I j 7 ; property acasured for that material.
! i
(
I 3! MR. SHEWMON: That would be down around I !
9 l the 6v0?
l 10 MR. HAYJ: Yes. 550 in this case. For !
11 this investigation we use only the simple --
what we
.l t
4 ,
i, 12 {have been calling the simple through-wall crack, i i i i
- 13 which is the geometry on the left.
i '
14 As I say, we did both JR carve t
] 15 characterization and a tearing instability i 16 evaluation. Tnis is a schematic of our pipe set-up.
I 17 Four point bending with the upper and lower load
- 18 points, ra e a s u r e the load line displacement, crack i
f 19 mouth opening displacement, displacement of the I
i 20 t crosshead, and we varied the compliance of our test.
i 21 Ue were using the insertion of Belleville I
l 22 susner-type springs, i
j 2 .1 By inserting different amounts of springs i
24 in different number of stages, we could vary the l 26 compliance of the test rig.
1 t
d I
154 1 MR. RODABAUGH: As I understand this, is
\,_)
2 this an unpressurized t e s t. ?
3 MR. HAYS: Yes. This particular set of 4 ; tests were don 2 at 125 degrees F, merely to insure l
5 ! upper shelf behavior and unpressurized. He do not 6 have the facilities to perform pressurized --
7 combined pressure bending tests.
8l '1R . RODABAUGH: I'm not trying to do more I
9 than just understand. Without the pressure loading --
10 let ne put if the other way. The pressure loading i
11 'would have what kind of effect on the data? What 12 !might it introduce that you don't see hare? As for O
(_ 13 l example, change the stiffness of a pipe.
14 MR. HAYS: That would change the stiffness.
i 15 It would also affect some of the analyses, 16 specifically tna lialt load type analysis wnan you 17 have to calculate a change in the neutral axis, but 16 with the circumferential crack like this, the 19 additional loading may not be real significant duc 20 to the pressures.
21 MR. RO D A B AUGif : So the assumption you are 22 making is somewhat internal loading and doesn't make 23 that uuca difference. Is that what you are saying?
(~
s 24 MR. HAYS: Our reasoning for not going to N
25 a pressurized system was that we felt we needed to
+ -y w w- ry
155 1 understand what was going on and be able to analyze
{~N
'N 2 tna circumferential crack without the complications, 3 the added complications of pressure before moving 4 onto tne more complex problen.
5 MR. RODABAUGH: I see. That is good l
6 janougn answer for right now. I'm trying to i
7 , understand.
8! aR. HAYS: This is a photograph of our i
9 l test facilities.
I As I pointed out, we have the t
10 l stacks of Belleville springs here, heater tapes i
l 11 l wrapped around the test section to insure the l
12 ; temperature. This instruaentation is for the load i
(s~/S 13 line displaceuent measurements.
14 ,
Taroughout this presentation I'll be i
15 referring to both the elastic unloading compliance lo metnoa of estimating cruck length through a test and 17 the DC potential drop or the EP method. This is a 18 common experimental detail, but I would like to go
}
19 ;through it so you will be able to understand some of I
20 i the successive view graphs.
21 During a test, during our test records, we 22 go up and load, perform a series of small clastic 23 unloadings, neasure the stiffness of the pipe, O
g i 24 relate that to a calibration curve and get discrete
%d 25 crack length throughout the test. So the elastic
s -
156 rw 1 unloading compliance technique gives you only
( '
2 discrete crack length estimates throughout the test.
I 3-! The electric potential-type method works 4 l by feeding a constant ' current across the specimen l
5 j test section, neasuring the chunge in potential 6 lthroughout the test as the crack grows, picking off l
7 a crack initiation by need and the potential versus 8 l crack opening displacement curve, then relating 9 succeeding potential to a calibration curve and I
'1 ! coming up with crack extensions.
11 Again, these tuo techniques, we use them 12 bcth together during all the tests, and they come
(\
ss 13 j frca different -- they are different but we can 14 Istill use them together. We use them to back each i
15 iother up.
lb We used what has been referred to a couple i
17 ; of times earlier today as a Eta factor approach to 16 get an estimate of the J interval during the test.
19 For this portion, for this investigation, we looked 1
20 at the one by Zahoor, which is a function of the l i
21 actual load in displacements and thus accounts for 22 material hardening during the test. We have also i 23 looked at one by TADA coworkers which assumes the
( 24 elastic perfectly plastic behavior and does not 25 account for the hardening during the test.
l l
i
157 rw 1 We also --
they also have similar L
2 expressions for the tearing modules, the applied l
3 tearing modules. For the A106 steel, it appeared 4 lthat j the assumption of elastic perfectly plastic i
1 5 I behavior was not such a bad one because you can see I
t 6 ithat the curves are very, very close to each other.
l 7 l Ana nere's the first instance of where you would see 8lthe difference in crack length estimation techniques.
i 9 Bioving onto the results of the series of 10 , pipe tests, a series of eight pipe tests were 11 ' conducted. Tnese are the results from the plastic 12 unloading compliance results technique. You can see 13 a fairly wide range of scatter there with J 14 ! initiation value around 4000 inch pounds por square i
15 linen. Tne same pipes using the DC potential drop l
16 technique. Sca tte r is greatly reduced and perhaps a l
17 lslightly hig he r estimate of tne J initiation value.
t 18 As I indicated, we looked at the 19 applicability of using laboratory specimens to model 20 this pipe behavior. We looked at one-half T, 1 T 21 and 2 T plan cpecimens cut from the wall of the pipe, 22 and in this case you can see that the range of the 23 nalf T specimens fell nicely within the range of the
\ 24 pipe tests.
(G 25 We run into a problem here that Gery
158 1 talked about earlier, which was there is a very v
2 limited amount of crack growth available on such a 3 small laboratory-type specimen so we can feel 4 ! comfor taole with modeling the initiation point with 5 ,these small specimens, but to say much about 6 jexpended crack growth is very difficult.
7 i MR. RODABAUGH: What is the wall thickness?
i 8 21 R . II A Y S : This was about 540,000ths.
9 It's a Schedule 80.
10 f .1 R . ROD A B AU Gil : Thickness of pipe, do you I
11 i have to consider how thick?
t 12 l :1R . MAYFIELD: If you think of the I
s) 13 diameter, it would run 3, 3 and a half i nc he s . Ilo t 14 legs will run up four and a half inches thick.
15 MR. RODABAUGH: Okay, thank you.
16 MR. II AY S : These are similar results for l
17 the 1 T plan specimens and kind of border on the le high side there, and the 2 T plan specimons we had l
19 ,l to flatten after taking the blanks from the wall of 2d the pipe due to the curvature of the pipe, and witu t 21 that did was induce a prestrain effect at the crack 22 cross section, resulting in lower resistance curves
- 23 from the 2 T specimens.
(} 24 So, what we said was that essentially you 25 must use the same thickness specimen, laboratory-t
159 1 type specimen, and a specimen that has the same 2 history for use in modeling the pipe R curve 3 behavior.
4 The next couple are just some numbers 5 which we will allow you to look at at your leisure.
6 Julie's module had a very similar view graph today 7 l explaining the tearing instability approach.
f ajEssentially it says you have a material tearing 9 modulus which comes from the normalized slope of the i
10 JR curve and then an applied curve whien is I
11 jstructural in nature. And the conditions.for 12 j instability are that applitary modules succeed in (s's) 13 jmaterial tearing modules.
I 14 l In our case, we can vary the applied l
15 l tearing modules through the insertion of Belleville i
i 16 I springs. I think that goes back to your question 17lwhether or not we can induce instability in type 304 16 stainless steel, and indeed we have done that. But 19lthat was a very, very compliant system.
20 To feed this data into the material 21 tearing modulus-type approach, we had to fit the 22 data to poacr our curve to get a nice smooth 23 function for the JT space. And here are the results.
24 Again, you want to remember the conditions for 25 unstable crack growth, that is that the applied
a -..
160 s 1 tearing modulus exceeded the material tearing T
2 modulus.
3 In this case we had not a very compliant 4
system at all. The material was higher than the I
5 l applied T throughout the test and stable crack 6 ' extensions resulted. This shows the two 7 i formulations for it applied. It looks as if the l
8 iTyler approach would have predicted an instability 9 {right at the end of the test, but tnere was none 10 ,during this-test, and we did have stable crack i
11 extensions.
12 : This shows that the Tyler approach of
, N) 13 ! assuming an elastic perfectly plastic behavior might i
14 j be somewhat more conservative. Again it does not 15 l account for material hardening during the test.
16 l For pipe 13, we did nave unstable cracking i
17 l extensions. This curve should continue up here in a l
ld l smooth fashion, and at the end of these curves is 19 where we did produce instability.. So, both 20 techniques predicted instability, but to be exact, 21 this should have produced an instability right at 22 this point, that the difference may be due to some 23 of the JR curve smoothing that we did.
l 24 We weren't looking at the local tearing x i 25 modulus but a smooth tearing modulus. And similar
161 1 results for pipe 15.
N.
2 MR. HUTCHIUSON: Where was the 3 ; instability tnere, at the end of curve?
l 4 l MR. HAYS: Because we wanted to put the i
5 luriting in on this view graph, you should continue t
6 lthis curve up to w he re the applied curves end. And I
7 that's where the instability occurred.
8I Moving onto the GTA welded A106 steel pipe, 9 these are circumferentially welded. This is a use i
10 l task which we are just undertaking at this point. 1 l
11 ; don't have any data to show you as of yet.
12 j MR. SHEWMON: The material on the last set t
13 i of pipes was what?
I 14 I MR. HAYS: Is 106 Grade B.
15 a2. SHEWAON: And this is what grade?
16 MR. HAYS: A welded 106 Grade B. A 17 !circumferentially welded pipe with notches machined i
i 18 { i n the circumferential weld-and fatigue cracks grown l
19 from the tips of the machine notch. Objectives here i
20 l are similar to the objectives we had for the A106.
21 The baca metal pipes, first of all, we feel we have I 22 to go in and really characterize the R curve 23 benavior and coupare that to what we got for the 24 This is a gas-tungston are automatic
( base metal.
25 process, so it should exhibit fairly good toughness
)
l l
{
162 ps 1 we feel.
I' 2 ,
.ie also want to go in and look at the 1
3 tearing instability behavior, and look at the use of 4
lsmall scale of laboratory-type specimens to model 5 !the pipe R curve behavior, and really try in this '
l 6 case to go out and take our coupact specimens to l 7 i large amounts of crack extensions. ;
i !
ol MR. RODABAUGH: Is gas-tungston are used ,
i i I
- 9 I on carbon steel piping much?
4 l
j j 10 3R. 11 AY S : 1 think that is probably not L
- t l 11 Ithe case. These pipes were welded for us several !
l 12 years ago, and this was going to be our first I i - ,
l l % 13 attempt at producing R curves on the welded pipe. f i
14 lAnd fron un experimentalist standpoint for i 1 !
l i 15 l understanding what the processes were and developing
< 4
{
16 l analyses, we wanted to try to have as homogeneous 1
- 17 ,
weld as possible.
} l3 l :1R . S ii E VlM O N : Is the most homogeneous part I
i
- 19lfor modeling would not have been a weld at all, is i . I 1 1 4
20 lthat correct?
j 2L MR. HAYS: That's right. That was the
! 22 next step beyond the base metal. We felt we j 23 understood what was going on with the base metal.
l O) j (u/ 24 Lie want then to move onto welds which was of more t
j 25 concern. So, that was the emphasize to going to the 4
I k
i
163 1 GTA weld. We have a total of six pipe tests planned, (m
V 2 all of shich will be conducted at 550 degrees F.
3 Again, eight inch diameter Sc he d u le 80.
4l ,
We believe using the simple crack i
5l geometries, we will test some in the non-compliant i
6 i and soae in the compliant modules.
7l MR. RODABAUGH: The number 8.60 is not a i
8 i common pipa diameter.
l 9 MR. 11 AY S : That's an outside dia me te r or 10 an eight inch.
11 HR. l' O D A B A U G il : The outside dianeter is i
12 '
8.65. I'm not sure w he t he r you have measured the
/~N l 13 surface or whether --
14 l MR. HAYS: That's an average of some 15 i measurements, outside diameter measurements, that we 16 did. Moving onto the four inch diameter type 304 17 stainless steel type, these are base metal pipes, i
18 jagain Sc he d u le 80 with about 340,000ths wall I
13 I thickness. Tnese were again circumferentially l
20 I cracked, in this case, however we put blunt notches 21 in them. Tne machine blunts notches in them, and 22 that's a 50 milliroot radius.
23 The purpose of this series of tests was 24 for inputs to a J calibration curve for development
(
25 of the JM analysis being performed by liugo-Carnst.
I
164 gss 1 Also a secondary aspect of this work was to feed b 2 this into perhaps a screening criteria or use thase 3 pipes as data points for development of the code 4 flaw evaluation procedure.
t 5! This group of pipes is significant because I
6 it's --
the control here or tne variable here was 7 lthe initial crack angle or the initial crack length.
l i
8 ; We range from about ou degrees to about 140 degrees I
9 initial crack angle.
10 We t he n took all of them except one to 11 ; maximum load. So, we had a series of crack angle i.
12 versus maximum load. Initial crack angle. That 13 fits very well into verifying the code flaw 14 l evaluation procedure.
I 15 l MR. RODABAUGH: You also have the crack i
i 16 ! initiation modes?
I t
17 MR. HAYS: I don't have those, no. These i
18 iwere not as fully instrumented as the other ones.
19 l Jo we sere unaole to pick those off exactly.
1 But in 4
20 all cases, as you see, we did have some crack 21 extensions during tne test, and we only took t he m to 22 maximun load. So crack extensions did occur prior
- 2) to maximum load in the blunt notch base metal plate.
24 RODABAUGH:
(O} MR. I guess what I'm striving 25 at, when you view the ASi1E code guidelines, IWB 3640,
4 165 rs 1 are those related to initiation loads or maximum 2 loads?
3 . IIR . 11 AYS : Maximum loads as far as I'm l
4 Iaware.
5 Ma. RO D A B AU Gil : I guess with everything i
6 '
going through my mind, as one starts this process 7 with an elastic piping system analysis, when you get 8 to maximum loads you're deformations are way be yond l l
9! elastic response. So, I'm wondering if somewhere 10 along the road we haven't lost --
really if we are 11 not mixing apples and oranges and using elastic 12 analysis in conjunction with plastic maximum load 13 criteria. Well, I don't c::pect a response. Just a 14 concern of mine.
15 ! IIR . liA Y S : Well, one would certainly feel 4
16 i tnat certainly a type 340 stainless steel is very 17 lductiule material, would be fully plastic at maximum 18 j load and is going to be fully plastic. These were 13 ! snail pipes and rather large initial crack, and this 20 also fits into Gory's scraening criteria for the use 21 of limit load. You can uoe just what we expected 22 here, maximum load being inversely proportional to 23 the initial crackling. ,
I
(} 24 MR. RODABAUGH: Tnis graph that you have l 25 right there illus tra te s my concern. You see the 1
l
160
<g 1 clastic initial slope is more or less a
~) 2 representation of plastic response.
3 MR. HAYS: Yes.
1 4 MR. RODABAUGH: And for an elastic piping 5 cystem analysis to be generally regarded, you would i
6 llike to know that you are talking about loads that l
7 lare -- well the code definition is to take the i
8 l elastic slope and double it, say 15,000 on that i
9 l scale there, that would be a load limit that would i
}
10 i keep your elastic piping system analysis sensible, i
11 llet's say. aow look at your maximum load and notice
, 12 j it's up to 23,000. So, I'm a little bit concerned 13 in the overall evutuation process wnether we are
- 14lreally --
elastic piping system analysis is really t
i 15 l feeding in the right data to evaluate the cracks.
16 MR. BENDER: What might happen, how might 17 it influence or something?
18 l MR. RODABAUGH: You would get a large 19 deformation in rotation at some point in the piping 20 system. Now, you are probably going to be 21 conservative with respect to that particular 22 location. Remember this piping just wonders all 2J over the place. So, at other points elsewhere
( 24 because you have this large de fo rma t ion, which is 25 not any part of the elastic value analysis, suddenly
l LUI--
1 over here you have other estimated loads.
[h sjm 2 l. MR. BENDER: You get some displacement i n 1
3 other places?
t 4i ,
MR. RODABAUGH: You are going to get 5l unexpected loads in other places.
That's my 1
6 { underlying concern.
7- MR. GENDER: Unless there's a crack there, i
8lwould it make any difference? Where are we going?
9I I'u expressing total ignorance.
10 MR. RODABAUGH: Yes, there would have to 11 ! be a crack at two locations or approximately.
12! Location A you are looking at is one where you are 13 going to apply this critoria very high in that loada
{
14!were really applied. Then the load that you 15 calculated for another potential crack location on 16 the piping system analysis might be quite 17 unconservative. . think I understand.
18 MR. HAYS: I think you also have to recall i
19 ; that this procedure does try to predict the load 20 l carrying capacity, but then there is also that two 21 and three-quartars safety margin on the load. So 22 for that, if you can indeed predict that maximum 2J load cloucly, then the application of that safety I 24 margin may bring you down where you are still in O 2b good shai y-3-=*--*-y*+9wap 9- -y=-amp-e fw=e-tei*-T**pe ePw e mt rwa my-mn--
1 MR. BENDER: For the purpose of just 2 d i s p l a~y i n g the margin, this isn't too bad, but I 3 think Uverett's point may still be valid. You 4 probably have to look at how far you go and where S
lthe allowable load could go before you could 6 influence the piping system somewhere else. I think i
7 that's the point he' trying to make.
I 8 j MR. IIA Y S : As a displacement argument as 9l opposed to a load. That's true. These do go to a 10 large displacement before t he y do hit maximum load.
i 11 .
f1R . RO D A B A U GII : 1 think this is be yond i
12 I what you are directly trying to address in these i
13 prograus, is a step beyond wnere the y are trying to 14 l apply out in the Beaver Valley II, for example.
t 15 MR. IIA Y S : I think it shows the importance i
- 16 of some of these things that we saw earlier today, 17 I some of the estimation s c he me s where we are going in i
18 iand predicting real loads and displacements.
i 19 , We applied the next section collapse-type 20 l limit loads to these pipes. In this case we had I
21 only a simple circumferential crack so this A bar 22 term which takes into account the A over T effect i
23 goes away, and this unows that basically the inputa 24 to the equation are pipe size, some geometry, which f'h A/ 25 is simply crack related, and then the flow stress, t
l
I i
1 69 .
(
l which is a measured material property or an
) 2 estimated material property.
3 These show the results for that next ;
l(
4 1section collapse-type analysis. In this case I l
5 I plotted this inversely from Battelle. I have the l t
6 l limit load over the maximum experimental load. What 7 lthis shows is that -- and the differences between
! i 8 l these two curves are the flow stress, the upper l 9 l curve being the 3 S of M estimate flow stress and l
< j r 10Itha lower curve being a measured flow stress, i
, i
' l 11 ;
average of the yicid. l i
12 i ilha t tnis indicates is that, as Gery has i
13 : chown earlier, that the application of this s l l 14 ! teenniqua does appear to be a function of the
(
[
15 ' initial crack or the applicability of this technique. l
! I i i 16 It does appear to be a function of that crack length.
17 ! So in concluding this segement, I would lallike to say that the formulation of the JM for the 19l pipe geometry is continuing. It's due to be out j 20 very shortly. I nave some of the reports back on my i
21 desk. And that, again, the accuracy of the limit 22 load expression is a function of the initial crack j i
23 length. I 24 :lo v i n g~ onto the four inch diameter GTA .
(:) 25 welded 340 stainless steel pipe, again four inch '
l i
i i
l
- 17U 1 S c he d u le 80 pipe, circumferential wald, notched ir s 2 the weld metal in the center line of the weld metal 3 and fatigue cracked prior to testing. Our i
4 objectives here again --
1 5 :1R . RuDABAUGH: Let me tie this weld in 6 lwith what we heard this morning from Battelle. T he y 7 were talking about a gas weld.
b MR. HAYS: The y are similar properties --
s 1 9 l or excuse me, the y are the same in this case.
10 MR. RODABAUGH: Same as TIG?
I l
, 11 1 MR. 11 AY S : TIG is perhaps an older '
i 12 l expression for what t he y are calling GTA at this I
i 13 point. We wanted to characterize the JR curve and
{
{ 14 jalso investigate the use of compact specimens to i
15 lmodel fracture to u g hn.s s a , and also the third l
16 l objective was to look at the tearing instability i I 4
17 luehavior of this pipe, which I'm not prepared to
) !
18 li talk about today but which tests we have performed.
I h i 19 l This is a shot of our test facility, again, i !
20 showing the four inch diameter pipe. This is simply 2L an insulating cover. We are heating it to 550 with
- 22 internal resistance-type strip heaters. In this i
23 cuse we nad to move some of our instrument machines 24 away from the pipe due to the high temperatures.
25 Thia is one I think we are all familiar i
171 l
now after earlier this uorning. We looked in 1 lwith 9 2 this case at both the simple and tne conplex crack 3 y e o me t r'i e s . The pipe welded configuration was 4 Jouble V butt welo with 125,000 band, automatic 5 gas-tungston are with type 3080 weld, stainless 6 steel filler uetal and fairly low heat input. Very 7 nice weld.
l 3 :1R . S riL WHUN : In the J03U the ferrite is 9 Kept constant with the carbon as well, is that it?
lu l i .i . tla (S : 'J h e carbon is e x t r e zae l y low.
'l "' h a t ' s the L designation. I can't really say much
>ut tue ferrite content. I have a ferrite gauge, i
g 13 but ay measurements would be influenced by the based 14 uaterial on each side. It was o fairly thin weld, 15 so I could not get a ferrite number on that.
lu .in . SdLun0N: You could always resort to 17 cetallurgy if a e , se fai!n.
1J '4 d . li A Y S : Okay. Thic gives you an idea 19 of wnat the Tensile properties were. Interestingly 20 here, wa nave a weld uctal was about twice --
hau l
21 twice the yield strength as the base metal, similar la ultimate strengtn. 't h e flow stress here again is 2J just the averaje of tne yield ultimate strength.
,e 21 uova auctility in cota the wcld of the base metal au L -J 25 shown ooth in the longation and reduction area.
i 172 1 We encountered some of the ovalization O
' 2 ; that Gery and some others had talked about earlier 3 this mcrning. And harping back to the elastic i i 4 l compliance schematic 1 showed you today earlier, we 5 lhave to relate our ueasured pipe stiffnesses to a l
6 calibration curve. This one in particular, which we 7 lused for tne eight inch diameter AlO6 pipe was 8 developed by Joyce, and once we encountered this 9 ovalization, we were having a hard time masking our 10 i initial and final cracklings to initial and final 11 ' measured crack length ofter the test, and we had to 12 develop this compliance calibration curve, which
() 13 shows with tne ovalization which occurs, makes the 14 l pipe look like that. Difficult to see here today.
15 This is the machine notch fatigue crack e x te nd ing r
i 16 ;here, and then the test on e i t he r side, the test I
17 uraa.
1d l Fi R . IlU T C H I N S O N : I can't tell how that's l
19 I ovalized.
I 20 ha. liA Y S : Okay. This is the major 21 diameter, and this is the minor diameter. So it's l 22 ovatizing, the measured diameter is becoming the 23 vertical diameter of the pipe.
24 MR. H U T C li I n S a u : it's just the opposite in 25 a completely -- what happens in a completely bent J
173 1 test.
2 lid . i!A Y S : In a flattening-type test.
3 ;in . dUTCHINSON: Yes.
i 4 1 ..d. ii A YS : If you had no crack there at 5 all, you might expect to get flattenin . This 6 produces un effect of utiffening of the pipe, which 7 led to that flattening of that compliance.
O MR. R O D A B A U G tI : 1 still don't understand 9 wny Joyce's results differ from yo ur s with respect 10 to ovalization.
11 ilk . liAY S : Because his was developed on 12 aluminum pipes with fairly long crack lengths and h 13 thay uion't experience ovalization. .i e didn't 14 'really run into this problem the A106 pipes at all, la but once we got into the stainless steel --
10 :1R . RODABAUGH: Ordinarily low grada l ~i struss analyslu would say that the amount of 16 ovalization that you get in these tests would depend 19 cn are anglu of those saddles on ooth top anu bottoa ;
1 l
20 and the length between the saddles. Have you found '
i 21 l uny evidance at all that your test rig is 22 influencing your results?
2J MR. huYb: .. 'O Were Concerned about that, fm s 24 and because of our concern, Mr. Ila y f ie ld had us run
\- )
23 a comparicon teut witia uattelle C o l u ;a b u s . We use 12
174 ;
l 1 inch diameter on 12 inch inside spacing in between l 4
i
(h
- 2i our upper load points, which in this case is about 3 three diameters. Our loads did come out somewhat -- .
1 i I 4i our maximum obtained loads come out somewhat smaller '
l 5 than Battelle's, however when we looked at l 6' ovalization as a function of distance from the ,
7 character cross-section, we were both getting about i
8 the same amount of ovalization, I think, and also 1
9' the ovalization died out within about a diameter on 1
i 10 either side of crack cross-section.
11 MR. SHEWMON: Is it clear why for the same 12 geometry that shows up in stainless but not ferrity
) 13 for aluminum, at least if not, ferrity versus i
14, stainless?
15 MR. HAYS: I think it's due to the high l
l 16 toughness. You get gross plasticity during these I
17 tests due to the high toughness. Plasticity that's i
18 not associated with the crack at all, and I think 19, that shows in the diameter or the change in diameter i
. 20 is a function of the distance from the crons-section.
21 MR. SHEWON: And the yield here was 22 one-third or half what the yield was in the ferrity, I 23 is that right?
24 MR. HAYS: Yes, approximately.
i 25 MR. SHEWMON: You had 70 for a number. I i
1 l
175 l
1 don't know whether that was yield or ultimate.
I' l
( )\
2 MR. HAYS: Well, I would have to go-back l
l 3 and check that. Another thing we have to remenber i
i 4, in this was that we have the weld in there which has 5l an extremely high yield strength in comparison to l
6 the base metal surrounding it, and induces lots of 7 plasticity'in the surrounding base metal before you l
8 can get any crack going on in the weldment.
9 Looking at some of the results, this is 10 from elastic compliance from the first pipe that we 11 did with an initial crack length of about a hundred 12 degrees.
() 13 The fill point here is a measure, if you 14 will, of the quality of the data. That's the final 15 measured crack length after every test. We break 16 them open and measure the initial and final crack I
17 length. In this case, we over-predicted the final 18 crack length by a small amount.
L9 I should also say about this pipe that i
20 both crack tips stayed in the weld metal, and we had 21 an initiation around 4500 inch pounds per square 22 inch.
23 Second pipe we tested, we got results from 24 both elastic compliance and potential drop
(
25 measurements, with a hundred 40 degrees initial j
176 1 crack length. In this case, both of the crack tips Ih
'L/ 2 grew out of the weld metal and into the toe of the i
3 ' we ld . Again, with elastic compliance technique, we 4 ,
over-predicted the final crack length. The DC I
5; potential drop is tied to the elastic points so t
! 6i there was no way we could under-predict.
7l i MR. ETHERINGTON: What is a potential drop 8 i technique ?
l 9 MR. HAYS: The potential drop technique is 10 where you set a constant current field'across your i
11 j specimen, across the crack cross-section. As the 12, crack begins to grow, you measure the potential i
l
() 13 i
across the crack face and you relate the change of 14l potential to a change in crack length.
15 i In this case, it should be because of the i
16l change in geometry that you get the change in all i
17l the plasticity. This technique should work better 18 l for a stainless steel.
Elastic compliance depends 19 on being able to measure that change in stiffness, 20 and when you are getting ovalization and plasticity 21 effects, it's not as sensitive. It does not work 22 quite as well.
23 Ma. RODABAUGH We are looking at the weld 24 property, right? TIC, or let's see, you call it
{
25 something else.
1 1
177 1l MR. HAYS: GTA weld. This is welded pipe.
2 MR. RODABAUGH: This is way up there.
i 3 This weld metal is almost as good as a base metal?
4 MR. HAYS: It's very tough.
5 HR. RODABAUGH: Okay.
6 MR. HAYS: Again, we have results for 900 7 from elastic compliance and potential drop. In this 8 case we under-predicted the final crack length a 9 little bit. If you were to correct this curve, it 10 might -- the triangles here might shift over a L1 little bit and be in better agreement with the 12 potential drop curve. In this case, also both crack lll 13 tips grew out of the weld metal and into the toe of 14 the weld.
15 And I have a shot of that right here. I 16 believe this is pipe 900 after some considerable 17 fracture at post-test fracture. Again, the machine 18 notch, a little bit of the fatigue crack, and then l ') tne fracture grew out of the weld itself and into 20 the toe of the weld. This is just a comparison of l
21 ! the simple crack results using elastic compliance. l 22 The high one there was the last one I showed, 900, 2J and that may be a result of the crack growing out fq 24 into the base metal.
%)
25 The results found from potential drop for
.. - - - . .- __ . _ _ . .- . . . . . _ - - . .. . = - . ._. . - .
4 1
! 178 l
l 1 the simple cracks. Fairly good agreement at crack i
l 2 initiation, but again we attribute the higher R 3 icurves to the fact that the cracks did grow out into 4 the ~oase metal.
,I l
i 5 . Moving onto the complex crack, Gregg i
6 I pointed out a couple of things that you should I
l 7 j remember when you are looking at a complex crack 8 results, that is that there is presently no way o f f
9l handling those results or thoce crack geometries, I
j I
10 j the radial comp one force for those cracks. So, we l
) 11 i had to use the same analysis for the complex crack
} i j 12 , as we did for the circumferential or the simple ,
) 13lthrough-wall crack. That showed as a great l 14 reduction in the initiation as well as the apparent l 15l tearing modules on the slope of resistance curve.
4 d
j 16 In this case we had about a 40 percent i
1 17 through-wall crack, interior through-wall crack, and l
j l 18 that kept the loads low. Kept the plasticity down i
): 19 and gave us pretty good agreement between our two ;
i l
A l 20 crack destination techniques.
) 21 Similar results for pipe 600, which was a 1
i 22 complex crack. In this case only 25 percent l
, 23 through-wall crack and a slightly lower or smaller j
j 24 initial crack.
1
! 25 MR. SilE W10N : If you are going to get 1
I
~ ~~
179 l
l 1 through all this by two o' clock, you are going to 2 have -- going to have go. You might skip a few of 3 the compliance curves.
4 MR. HAYS: Okay. Well, to move on a 5 little bit quicker, this simply shows the difference 6 between the simple and the c o .n p l e x crack. Grea'tly 7 reduced initial J crack initiation as well as 3 reduction in the tearing modulus or the apparent 9 tearing mcdulus.
10 Moving onto the compact specimens, in this 11 case we used sizes ranging from 3 T and 1 T cut from 12 a plate as opposed to the pipe because of the high e
w . ,i 13 curvature of the small diameter of the pipe.
il the specimens were cut with the crack problem 15 promulgating in the direction of the weld, using 16 approximately the same weld procedure.
17 Comparing the pipe results to the compact ld re sult s for the pipe results, again we are unable to 19 get much crack extensions due to the high 20 displacements required in stainless steel, but they ,
i 21 i do seem to follow or at least fall within the band i
22 of the simple crack pipe data. In fact agree very ,
1 23 well at J initiation.
r- 24 MR. ETtiERINGTON: What wall thickness are m
25 we dealing with here?
180 1 MR. HAYS: This was about 3.4 inches. So, O 2 concluding, we characterized the R curve behavior 3 for the simple wall --
simple circumferential wall 4 through cracks. The J 1evel at initiation was about 5l6400ths inch pounds per square inch. We also saw 6lthe increased crack restraint due to the crack i
7l produced a much lower resistance curve.
l 8l We also saw that ovalization in plasticity i
9lgave un some experimental difficulties, and that 10 l here the good agreement between the pipe and the I
11 ' compact specimen bodes well for enabling us to use i i 12 these types of fracture mechanic-type specimens to
) 13 model pipe, the J level initiation for pipes. Again,,
14'i we need to go back and look at large crack 15 extensions.
10 Finally, I would like to go back, I would i
17 ' like to look at some work that we did for Section 11 18 Committee on pipe flaw evaluation concerning the use 19 l o f the proper flow stress for circumferential cracks I
20 on stainless steel pipe. This is a rather difficult 21 thing to look at here.
22 What I did was get all the Tensile data I' 23 could for type 304 stainless steel base metal. I 24 then calculated the flow stress as an average of the
(
25 yield and ultimate strength at 550. This is all 550 f
l 181 1 data. These bars show the minimum average and the
[
2, maximum reported. Then Section 3 allows you to 1
i !
3 calculate S of M three different ways. That's i
i i 4l one-third the ultimate at room temperature.
b MR. ETHERINGTON: They are yield strength l
6 at room temperature?
i 7 MR. HAYS: The yield strength at room i
8, temperature.
9' MR. ETHERINGTON: Whichever is the lower?
I 10 MR. HAYS: Whichever is the lower, correct.
i 11 What I did was multiply that by three to give me a l
12' flow stress from the 3 S of M. I also then just for
() 13 interest looked at what if I used the one-third the 14 ultimate at 550, the measured ultimate and 15 seven-tenths of the measured ultimate at 550. But 16 what this graph intends to point out is at no time l
17 does the 3 S of M flow stress go lower than even the 1E maximum of the measured flow stresses. Then I
- 19 looked at our base metal and weld metal type 304 i
20 stainless steel pipes.
m 21 In the limit load using the measured flow I i
I 22 stress being the boxes, and the 3 S of M flow stress
! 23 being the squares, that shows that the boxes or the 24 measured flow stress gives you a fairly accurate or
{
25 very accurate prediction of the load carrying
l 182 i
l' capacities of the load pipe whilo the 3 S of M flow 2, stress may lead to some over-prediction of the load 3 4 carrying capability.
i l
4 We apply the same kind of results to the 5 welded pipe. We felt kind of comfortable in that 6 since we did see the cracks going out of the welds i
7 and into the base metal. The results are somewhat l
t 8 more skattered, but what it does show is that there 9 were a couple of times using the 3 S of M of the 10, base metal that you could get some rather large over-11 predictions. '
12 So, in conclusion, I would just like to
) 13 say some words about the other research that we are l
i 14 currently conducting, as opposed to what we report, 15 in other things as well as pipes. We are currently 1 6' conducting investigation of the upper transition I
17; fracture toughness of 5533 B steel, and we are also 18 evaluating the geometry dependence and independence i
19 on the current JM formulation for laboratory-type l l
20 specimens, and the idea of the JM is to be the i
21 geometry independent type of analysis, and we are ,
l 22 just going in to verify that and critically look at 1
23 that in some controlled situation.
I 24 MR. S HEWMON : Is upper transition fracture
, {v'q l 25 toughness the same as saying what's J 1 for the l
183 l! upper shelf energy, or what does this mean?
2 MR. HAYS: That is in the upper transition 3 region.
4 MR. S HEWMON : What is the upper transition 5 region?
6 MR. HAYS: It's kind of a loose term for l
l 7 after you turn the knee and start to come down into 8 transition.
9 MR. S HEWMON : On a Charpy?
10 MR. HAYS: On a Charpy or however you want 11 to define it, DT or a dynamic bend bar and these are 12 going to be fit into some of the wide plate tests em 13 t
) that are being done.
14 MR. SHEWMON: Okay.
15 MR. HUTC HI NSON : Can I ask you one quick 16 question because it is confusing me a little bit.
17 On both with respect to your presentation and 18 evaluation this morning for the complex crack pipe, 19 when you define J there -- I'm not quite sure if I'm 20 asking a sensible question -- but do you divide out 20 by the full thickness or do you divided out by the 22 remaining thickness?
l 23 MR. HA YS : I have been using it --
and I
(^)
L. )
24 can't speak for anybody else --
but I have been l
25 using the net thickness. That way I do account a
184 l! little bit for the internal crack as I would do in a fi !
2 side crack tension.
4 MR. HAYS: In pure fracture mechanics, you i
j 5 might think of J as the measurement of the material 6' toughness, but in this case we are using more a l
7 relative measure of a material slash geometry or 8 crack geometry type. So, as validity in that you 9 can compare the two types of cracks and the l
10 toughness of the two, but it's not a pure material
- \
1 1; toughness.
12 MR. S HEWMON : Other questions? Thank you.
) 13 MR. HISER: I'm Allen Hiser with Materials 14 Engineering Associates, and I'm here to tell you 15 about the piping mechanics data base that we have i
1 6, given the acronym PIFRAC. I'm going to talk about i I 1
17 three topics fairly briefly. The first being the I
18 implementation of the data base. Some recent 19 results on the factor toughness of aged cast 2 stainless and just a brief summary of some other 21 experimental results.
22 There has been a lot of discussion this l 23 morning about JT curves and structural integrity i
24 assessments. Just to show you where the material l 2$ properties fit in, on the left-hand side we have the 1
185 1 material fracture toughness. In this case the JR
) 2. curve. The JR curve obviously is something that l
3 would be in the data base. If one converts that l
4 into say a JT curve, then the JR curve feeds in l
5 right here, and as well the Tensile properties feed i
6 into the applied toughness curve or applied loading 7 curve.
i 8 Now, the goal of this work is to provide 9 in a computerized fashion all of the available data 10 that could be used and collect analyses or any other l
11 analyses of a piping systems. The key part is that l
12 this is in a computerized form for ease of access.
1
) 13 I
The approach we followed is outlined here.
14 It's a five step approach, and in the first -- well, 15 in this case the first two steps have been completed i
16 in that we have formulated the data base and we have l
17 completed a survey of the FSAR's from operating 18 plants, data collection, procurement of materials l
19 and e s t ;.bli sh me n t of a data base are currently 1
20 ongoing.
21 In terms of the types of information that 22 we have in the data base is briefly outlined here.
23 The key types of information about each material are 24 the chemistry Tensile properties, Charpy data if
(
25 it's available, and the JR curve data. The critical
186 l! perimeters that help to define the toughness and
) 2, properties of the material include the type of the i
3l material and the size of the pipe in terms of the 4, diameter and the wall thickness and the temperature l
5' and the orientation of the flaws are also very 6 critical.
7 I guess one thing that I want to emphasize 8 is that in the data base we have sufficient room for 9 any type of characterization information that 10 anybody would have on the materials and in materials 11; of chemistry and things like that. That just may i
12 provide more keys to find'ng out what data would be
() 13 most appropriate for each application.
14 Now, this survey of the FSAR's resulted in 15 our construction of this table, and this basically l 16 gives the materials that we found to be most 17; commonly used in the current operating plants. As 18 you can see, we have some carbon steel, cast i
19 stainless, stainless steel and various weld metals.
l t t
20 And also given here are the diameters and the wall 21 thicknesses that were most prevalent and just a 2d brief -- some brief words on the applications.
23 MR. RODABAUGH: Looks like you are 24 concentrating on what we call code class one piping.
25 Was that intentional?
187
- 1. MR. MAYFIELD: It was initially, yes. You
)
2 have to have a starting place, so we started with a l
3 code class one principally so we would be consistent 4 with what's in 1061 and the limited scope i 5 modification in GDC 4.
I 6 MR. HISER: Mike is rather familiar with 7, this because he's the one that actually did the 8 survey. Little bit familiar. First step that we 9 did then was to convert that table into a matrix.
10 And basically this matrix is intended to cover the 11 chief types of information that we are looking for 12 for each material. On the left-hand side we have f's 1
() 13 each of the materials repeated with the range of 14 diameters and wall thicknasses. The chief material
(
15 property types we are looking for are Charpy, l
16 Tensile and JR curve with the appropriate l
17, orientations according to ASTM 399, and with the JR 18 curves, we also want our JR curve data to overcome i
I 19 some fairly broad temperature ranges.
l 20 The emphasis is primarily on room 21 temperature and 550, but we find around 300 degrees 21 we get a minimum in behavior and we also want to try 23 to characterize that. In this case, the boxes
( 24 filled-in areas represent cases where we have 25 performed testing and have the data in-house, while i
N -
_ -_-__--,-------J
188
_x 1 the X's represent cases where testing has been
- 1 '
2 completed but we don't have the data in the data 3 base as of yet.
4 Just a brief rundown on the data that we 5 have found available in the testing community with 6 the facility or laboratory that has the data. On 1
7 the left-hand side, just a brief description of the l 8 kind of materials that are available and just sort 9 of a guesstimate of the number of R curves. Don't 10 have any number next to Battelle because their bank 11 of data is growing literally day by day.
12 In the case of numbers with two stars next h 13 to it, that represents vendor data, and we are in 14 touch with, in this case, B & W and GE to try to 15 obtain this data, and I think probably will be 16 successful in each case.
17 MR. SHEWMON: What does one star mean?
18 MR. HISER: One star means it's pretty old 19 data. Probably is not of the quality of the current 20 data that's being generate.
I I
2d MR. SHEWMON: Now is that Westinghouse l 2 2, stuff? I it raw, or the cast they make reactors out l
23 of?
(' 24 MR. HISER: I think that's just raw. I
(_,I 25 think it's just --
l l
l l
f 189 1
1; MR. RODAB AUG H: Their piping is usually 2 cast stainless steel used in the primary.
3 MR. HISER: That's 304 data. I believe 4 that's about five years old or ten years old.
5 MR. MAYFIELD: Excuse me, on the 6 Westinghouse data, Bamford has some old casting 7 steel data that is in the public literature and is 8 accessible.
9 MR. SHEWMON: Is that included under 10 or 10 24 here?
11 MR. HISER: I think that might be under 12 the 10. If it is cast stainless, we would do our f 13 best to incorporate that in here, no matter how 14 primitive the test methods may have been back then.
15 We would try to patch something together.
16 MR. SHEWMON: It might be worthwhile not 17 to average it in in the raw material. I don't know 18 what your data base -- how it is, if it is set up so 19 it maintains its discreteness.
20 MR. HISER: Yes, everything is kept 21 ordered and everything has identifier numbers and i
2 things like that. So there really is no mixing of l
23 data unless the user chooses to do that. If you
("'
24 want a listing of all the raw data, then you would l
25 of course get that.
I
190 1 1 I guess two major points down here, some I
2l work that we have done for Dupont includes dynamic 3 JR curves on 304 stainless, which is different from i
4' the remaining data which tends to be just from i
5' statistic determinations, and the head data goes up 6 to I believe about a thousand degrees.
7, MR. SHEWMON: What do the numbers in the
, 8, last column mean?
9 MR. HISER: That's guesstimates on the 10 number of R curves.
II I MR. S HEWMON : Oh, I thought we are talking i
12, about a ten inch a minute ago and I got confused.
() 13 MR. HISER: No, that's just number of R 14 curves, so there aren't too many'out there. Now in 15[ terms of material procurement, this is just a i
l 16 listing of the materials that we have procured and i
17; in large part in conjunction with Battelle and with t
18 Argonne. Again, just some carbon steel, cast I
i l
19 stainless, austenitic stainless, Inconel and A-516.
20 These materials we have pretty much 21 completed the characterizations of, except for the 22 A-516 Grade 70, which shows up on the next list, 23 which I've called future materials. These are 24 materials that we either have in-house and are in ,
~A 25 the process of characterizing or that we expect to
191 obtain in the next few months.
4 2 I guess one of the key ones there is a 3 bimetallic end weld with a 516 Grade 70 on one end 4 and SA-182 on the other end. Now, in terms of the 5 computerized form of this data base, there were ,
1 l
6 three main requirements that we were looking for.
7 The first ic that catch-all phrase called 8 user friendly. What we have driven for is a menu 9 driven system that minimizes the knowledge and 10 interaction that the user has to give the data base 11 to pull out the required data.
12 On-line accessability was obviously one of
( ,, 13 the key requirements, and a third is processing 14 capability. The key part there is if there are any 15 modifications in J interval theory or anything, we 16 can incorporate those changes into the data that we 17 have for Rambert Ozgood curve fitting, things like 18 that. We were able to do that with this data base.
19 MR. S HEWMON : Can somebody plug in now or 20 next year or when?
2 MR. HISER: About two months. Month and a 22 half. The file structure for the data base is 23 outlined here. One of the key parts -- well, this 1
I
(' ; 24 reference file basically contains the chemistry and
~ ~J 25 heat number and material specification information
192 l' for each heated material. It also includes 2 reference information to reports that contain the 3 data and things like that.
4 The left-hand side we have the Tensile 5 data. The file that we have called ten prop l 6 contains basically specimen dimensions and basic 7 information like that on each material.
8 Ten data is essentially a tabular form of 9 yield strength and ultimate strength, reduction area, 10 longation. basic things like that for each specimen, 11 and ten tab contains the tabulated stress strain 12 data, and what we are hoping to get there for as ll 13 many of the specimens as possible is both 14 engineering and true stress strain. This way we can 15 always go back and curve Rambert Ozgood equations or 16 any other type of equation that one wants to use.
17 In the ten data file, we will also have 18 Rambert Ozgood perimeters for each test. So that 19 should be easy analysis use of the data base 20 significantly.
2h This column headed by impact file contains l
22 specimen dimensions in various information like that l
23 on the Charpy data or any other impact type of tests 24 that was run with the tabulated energy temperature
(~'
x,, )
2 %, and lateral extension numbers contained in that file.
L__ _
e 193 1 A third file that we have is called KlC.
7,~'T t I
. N)
' m 2
I If there is any KlC data available for the material, l
3 it would be listed in there. The heart of the data 4 base is really on the right-hand side. In the file 5 called JR summary we have the vital specimen 6 dimensions and a summary of J1C T average and plus 7 stress valuen and crack length, things like that for 8 each test.
9 And underneath there is a file called JR 10 data. This file contains the measured load 11 displacement on character length for each test. So 12 in this case we are probably looking at about 50 to f^
( )m 13 a hundred points for each test.
t 14 MR. ETHERINGTON: It's a complete curve?
15 MR. HISER: Pretty much. It's a complete 16 curve in terms of what one gets from an unloading I
1] compliance test. It would give you the discrete I
18 points for the R curve.
19 But one of the key aspects here is that i
20 since we have all of the caw data for the test, if a 1
21 new fracture perimeter is developed for the piping, j 22 we can go back and recalculate all the data in terms l
23 of that perimeter. So that is really the key aspect 24 to this.
25 That file also currently will have J and i
194 1: Delta A pairs which one would plot for an R curve. ;
2 Currently we plan on having deformation J and j 3 modified J as the two J formulas that we use on a 4 routine basis.
5 The file that is called here other data 6 contains dimensions and justifications for 7 non-scandard specimens. Typically we expect these 8 data to be either from compacts or possibly bend 9 bars which are very standard according to the ASTM.
10 As an example, with the full-width face 11 notch tension specimen that Battelle uses, that can 12 be fed into the data base with the critical lll 13 dimensions stuck in that file.
14 One other aspect that we consider to be 15 important is contained in the file called JR System, 16 and basically that is a brief description of the 17 test system and test procedure that were used in 18 determining the JR curves. Possibly if we have any 19 bias in any data we might be able to tie it together 20 using that file.
2L This is just a sample of what one of the 22 files looks like. In this case this is the header l
23 file, the reference information file. You can see r' ' . 24 we have room for the authors and basic reference
<)
25 information for the material, including a reference
195 1 to a report that the work may have been published in.
f%
1 i
2 The key part here, though, is really the chemistry l
3 information and the manufacturer, heat number, 4 product form and specification. This is just a i
5 typical sample of the structure of one of the data i
6 files.
7 So, I guess to end the data base i
8 implementation part of my talk, just to show this 9, flow chart, this is the way that we see the data 10 base working. Data base has many data types with l
11 many different data from many different materials.
i 12 Sources of data include Battelle, Argonne, NSRDC,
) 13 t
hopefully every premium vendors will also contribute 1
1 14 data. Users, of course, are primarily the NRC, but
)
15 also we expect some vendor use.
3 16 MR. HU TC HI NS ON : Is there a long-term 17 commitment to support this?
18 MR. ARLOTTO: Depends on the Congress of 19 the United States, l
i 20 MR. HISER: If I had my way, obviously. I 21 haven't been elected yet.
l 22 MR. RO DAB AUG H: What about the utilities, 1
23 do they not have interest in this?
24 MR.
( HISER: I would expect so, yes. I 25 sort of lump them together.
l l
t
196 s li MR. RODAB AUG H: I think between the
'l 2j utilities and NRC, the vendors ought to be licensees, l
3 and if they are properly motivated, they would work l
4 hard to keep the data base alive.
5 MR. HISER: I always sort of thought of 6 vendors as licensees, but I guess that really isn't i
the case.
8 MR. RODAB AUG H: They are very much.
9 MR. HISER: To shift gears a little bit 10 and talk about cast stainless and specifically aged 11 cast stainless --
12 MR. SHEWMON: As much as I love cast
) 13 stainless, let me interrupt you for a minute. What 1 4, recorded machinery do you have this'on or do you 15 visualize something that will go on a hard disk, 20 l
16 or 40 meg? Could you put it on an AT?
17, MR. HISER: Right now the use of this 18, that we expect is just for a user to download to his i
19 own system for whatever internal routines the user 20 would have. We have this on our mainframe computer 21 at MEA, and we are in the process of purchasing I l
2@ think about a 67 megabyte hard disk that would be 23 dedicated to the system. We expect not to use all
(' 24 67 megabytes for many years, but that disc would be Nu j) <
i 25 dedicated to this.
197 h
MR. S HEWMON : Okay, and the format is b'N j
'd 2. something which is DB2 or DB3, or what sort of 3 software do they need to go in and interpret or read
- 1 1 1 4 these?
i 5 MR. HISER: The software is all in our 6 system. When a user requires -- i l
7 MR. SHEWMON: You say they download and l 8 do what they want to with it, so there has -- they 9 have to take the software or be able to buy it or l
10 something. .
l l
11 MR. HISER: No, the actual idea is that l 12 they download, just download the numeric data into i
I
() 13 i
their own system, into a file that they create on 14 their own disk; and if you wanted to do a JT l
15 analysis, then you would have to have your own 16 software that would read these JR curve data and did I
17 whatever process you wanted to do. This is simply 18 just like a file cabinet. All we are going to do is l
l 19 hand you a file that has the data that you say you 20 need. What you did with it after that is up to your 21 ingenuity.
22 So, with the aged cast stainless, there 22 are --
first of all, with the unaged cast stainless, l
(~}
%/
24 there are some experimental problems. Here is all 25 the data that we were able to generate on a 1 T
198 l! compact specimen using about six-tenths of an inch 2 of load line displacement.
3 Now for those that aren't too familiar l 1
4' with experimental JR curve determinations, it's '
5 quite a bit of displacement, but dispite all that, 6 we only got about one-eighth of an inch of crack 7 growth in this 1 T specimen.
8 These lines that are drawn here are the 9 blunting line and the exclusion lines which one 10 would calculate using the ASTM 813 procedures. They 11 are based upon the average of the yield and ultimate
! 12 for the material.
l 13 According to E813-81 we don't have JRC in 14 this specimen. The specimen really hasn't initiated.
15 We have a problem with that. How do we compare the 16 effect of aging and things like that when we really 17 don't have crack initiation? That isn't anything 18 that we have --
we aren't the first ones to see.
19 That has been seen many times in the past by people 20 trying to test regular austenitic stainless.
l 2L What we have projected doing right now is 2q just fitting the initial part of the curve to a l
23 straight line and coming up with, I guess, you would 24 call a match flow stress in that case. So the next 2d couple of curves I show with the aging data are not
199
. 1 based upon the actual measured Tensile properties.
s 2 They are with some sort of a match flow stress.
I 3, Some of the testing that we have done l
4 recently for Argonne has been to evaluate the 5 Tensile property of aged cast stainless. There will 6 be considerable discussion of this tomorrow by Dr.
7 Chopra.
8 MR. SHEWMON: I was at Argonne last week.
9 I asked them where the tests were, and they said 10 they had sent them to a long time ago but hadn't 11 heard back. Did a surge just come or was this one 12 sample?
() 13 MR. HISER: This is everything we had.
14 MR. S HEWMON : Has it been shipped to them i
15 yet?
l 16 MR. HISER: It's's sitting right in my t
17 briefcase.
18 MR. SHEWMON: It's close anyway?
1 MR. HISER: Yes. I guess at this point i
2d I'll bring up some of the experimental problems with 21 testing Tensiles of cast stainless. Due to the 22 small size of specimen that was available and due to 21 the very large grains in the Tensiles, it's hard to 24 come up with good strain measurements on the
{
25 Tensiles. We fussed with that quite awhile.
- - - - ~ ~
k 200 1: Finally we came up with a procedure we were happy 2 with and have cince completed the Tensiles.
i 3 MR. SHEWMON: You mean it's like single
( 4 crystal or something?
i 5 MR. HISER: Almost. The gauge diameter 6 ovalization and what we have termed orange peeling 7 of the Tensiles is just phenominal. I should have 8 brought several samples with me. As I mentioned to 9 Bill Shack earlier, it's hard to measure the 10 elongation of the specimen after the test because 11 the two pieces don't fit together. It is just 12 really atrocious from an experimental standpoint.
h 13 This data will be written up and submitted 14 to Argonne and to our contract manager within about 15 two or three months. I just wanted to show briefly 16 a summary of some of the data, and what I have here 17 is data for three of the heats that were supplied to 18 us by Argonne, and in parentheses I have the percent 19 fahrenheit that they have determined. And basically 20 for the unaged condition I have the actual measured 1
2 yield and ultimate values.
l 22 These are in terms of NPA's.
23 And underneath the aging conditions here (x3
% ,)
24 would be 350 degrees C for 10,000 hours0 days <br />0 hours <br />0 weeks <br />0 months <br />, and I think 25 there is one mistake here. This should be 400 l
i_._
i
\
201 1 degrees for 10,000 hours0 days <br />0 hours <br />0 weeks <br />0 months <br />. I won't say it's perfect i l 2 otherwise, but I know there is that one mistake.
3, In general the aging increases the yield I
4 and ultimate fairly substantially in some cases. I l
5 guess 25 percent increase in the yield is the 1
I 6 greatest. Here we have one anomaly where there is i
j 7 actually a slight decrease in the yielded ultimate.
l 8 MR. ETHERINGTON: Is there any way you can l
! 9 reasonably extrapolate back to PWR temperature?
J !
- 10 MR. HISER
- I'll leave that up to Dr.
l l l 11 Chopra tomorrow.
- ?
i 12 MR. S HAC K : The tests are performed at the 1 ;
l ) 13 right temperature. The aging has to be extrapolated 4
l l- 14, and we will discuss that in detail tomorrow.
f 1
l 15 MR. HISER: All of these tests were I
16 performed at room temperature or at 550. So, they I
i 17 should be:(--
l 18 MR. RODAB AUG H: You don't have any tests I
19 at 550 fahrenheit?
l 20 MR. HISER: Yes, 550.
t 21 MR. RODAB AUG H: A whole column, fine.
l .
22 MR. HISER: These yielded ultimate.
2 MR. RODABAUG H: I'm sorry, aging. .Did you
( 24 do any aging at 290? .
25 MR. HISER: No. Dr. Chopra will discuss
/
202 li that tomorrow.
2, MR. S HAC K : Aging at 290 is a 20 to 40 3 year phenomina. We need long-term support.
4 MR. HISER: This data Argonne has had for 5 awhile but just not the appropriate Tensile data.
6 In this case we used guesstimate flow stress values 7 of about 650 MPA's, which is about a hundred KSI 8 which is approximately a fracture of two greater 9 than the actual measured property. So if we have 10 the actual properties, the curves would be down 11 somewhere like this. It would be about a factor of 12 two lower. But we just draw these in just to give 1ll> 13 you e 11tt1e dit of form to your data.
14, In the case of this centrifically cast 15 bite with the ferrite number of 5.6, the aging at 16 350 grease for both 3000 on 10,000 hours0 days <br />0 hours <br />0 weeks <br />0 months <br /> gave us 17 data that was right in line with the unaged data.
18 So it had absolutely no effect.
19 Aging at 400 degrees for 10,000 hours0 days <br />0 hours <br />0 weeks <br />0 months <br /> did 1
20 give about a 25 percent decrease in JR curve level. '
2I That's at room temperature. At 550, almost l
22 identical conclusions. l 23 The 350 degrec aging just didn't seem to
("y 24 have any effect on the toughness. Slightly
~J 25 different heated material, this one has ferrite l 1
h i
j 203 l
1; number of 24, and we can see that aging at 350 in h 2; this case gave us, I guess, about a 25 or 30 percent 3 decrease in the JR curve temperature. But in this 4 case with the 400 degree aging, looks like we have 5 about a 75 or 80 percent decrease in JR curve level.
6 That was at room temperature. At 550, the decreases 7 are slightly less. And as of right now that's all I 8 have to say about aged cast stainless. I 9 just want to briefly touch on a few topics. Just a 10 couple of comparisons with RPV steel behavior. In 11 the testing of RPV steels, a 357 B, a 3302 and their 12 welds, we found the use of side grooves helped h 13 substantially in containing a crack growth to within 14 a single plane, and just in general, improved the 15 test technique.
16 One of the other aspects was that it 17 tended to give us slightly lower JR curves. Now, in 18 the case of ferritic pipe steels, we see a similar 19 behavior. With A106 if you side groove our specimen, 20 you get a better test for one thing, but you also 1
21 get a slight decrease of the JR curve.
l 22 In the case of the stainless steel in the 23 top 0 percent and the 20 percent side grooves, don't i
I
('; 24 give us any difference in the R curve. In terms of
<J
~
l 25 fracture specimen, we end up with a straighter crack, u
\
r 204 i
- 1) which is better from our experimental technique.
/, N '
2 But in a case like the bot tom here , which is a 3 stainless weld, side grooving gives us progressively 4' lower JR curve behavior. So that's a little bit 5 different from the RPV steel where we consistently 6 saw a lowering in the JR curve witn increased side 7 grooving.
8 Now, one topic that Rich just talked about 9 was flattening specimens. Since many pipes tend to 10 be relatively small diameter with relatively thin 11 wall, it's very difficult to come up with a plainer 12 specimen that has a large enough ligament that you l 13 can measure the JR curve for a large amount of crack 14 growth.
15 One study that we did was to take a carbon 16 pipe and a stainless pipe and flatten blanks for 17 specimens and then measure the JR curves from the 18 specimens machined from those blanks and specimens 19 machined from virgin material. In the case of both 20 the stainless and the carbon steel, the virgin 21' material specimen R curves lie above those from the 22 flattened blanks.
23 In this case, we used 1 T- CT and 2 T-CT 24 blanks and half the specimens we applied the stress f~')s
%~
l 25 relief to to try to recover the residual stresses I
205 1 that were induced in the flattening process.
h 2 Unfortunately, we don't seem to get any 3 benefit to the stress relief. So I guess the idea 4 here is that you really can't flatten the pipe to 5 get extended amounts of crack growth.
6 MR. RODAB AUG H : There is lots of piping 7 common to nuclear power plants that have cold work 8 which is equivalent to flattened, unless something 9 else is going on here. If there weren't, by the 10 time you push and pull a piping system into place, 11 you got cold working on its strength. What I seem 12 to see here is a tremendous difference between 13 flattened and non-flattened.
14 MR. HISER: I think that's correct. I 15 think that for those cases you really need to 16 account for plastic work that has been put into the 17 specimen.
18 MR. RODABAUGHz And you were just saying 19 annealing did what to it?
20 MR. HISER: It didn't do anything. In l
21 each case in this lower band, the open symbols are l
22 the specimens that were stress released.
23 MR. RODAB AUG H: You don't anneal on the 6, l
,n s 24 do you?
% -] l 25 MR. HISER: Well, we did try to recover i
L_ _ ___
T 206 l' the residual.
2 MR. S HEWMON : What does annealing mean for 3 carbon steel? What does that mean, what temperature?
4 MR. HISER: I don't exact)j remember. I 5 think it was about 750 or 900, something.
6 MR. SHEWMGH: That's pretty low.
7 MR. RODAB AUG H: Did you try to anneal the 8 A106?
9 MR. HISER: We did the same thing in the i
10 stainless.
11 MR. RODAB AUG H: This time you did anneal 12 at 1900?
l 13 MR. HISE: No, no, down at the same 14 temperature.
15 MR. RODABAUGH: You are talking the stress 16 related?
17 MR. HISER: Yes, stress related. Just a 18 couple quick view graphs. Just leave you with this ,
19 last one. The effected term of the R curve is also 20 dramatically different from the RPV steels. On the 21' bottom we have one of the so-called lower-upper 22 shelf welds that MEA tested about five years ago, i
l 23 and as you can see in this case, increasing the test
/~' '. ,
24 temperature from 75 degrees C to 550 brought a L )
25 progressive lowering of the JR curve.
207 1 Now with A106 C, if I can find room 4 2 temperature, if we increase to 121 C, 232 C, 288, 3 343, the R curve reaches a minimum somewhere about 4 450 degrees. Dramatically different from the RPV 5 scale. Just to show that here using J1C as a !
6 perimeter, whereas a lower-upper shelf showed 7 essentially linear behavior with increasing 8 temperature. The A106 shows a minimum somewhere in j 9 the 450 range and then dramatic increase in 10 toughness.
11 MR. SHEWMON: What this RTV steel that you l
12 are talking about?
{
l 13 MR. HI S E R : RPV, reactor pressure vessel.
14 MR. SHEWMON: I've heard of them.
15 MR. HI S E R : One question that was raised 16 last summer at a piping integrity review group 17 meeting had to do with the variability of stress 18 strain curves. What we did to check this out was to 19 take again a carbon steel and a stainless and we ran 20 multiple tests of multiple specimens chosen through 2 the wall thickness. In addition, we used two 22 different size specimens to see if there was any 22 effect of the specimen size.
f") 24 As you can see for the stainless steel, we
'xs l 25 essentially had no gradient through the thickness t_
l
208 1: and at each temperature pretty good agreement 2 between the two specimen sizes. For the A106 there 3 appears to be a slight gradient through the wall 4 with the minimum properties in the middle. We also 5 have a stress strain curves, but I didn't bring 6 those today. These will also be coming out in a 7 report within about three months.
8 MR. RODAB AUG H: These specimens are in the 9 circumferential direction with respect to the pipe?
10 MR. HISER: These are axial. The axial 11 Tensile represents a circumferential crack plane.
12 Just a quick summary on the status of the
) 13 data base. The FSAR survey is complete, the data 14 base format has been established on a computer, the 15 query software that is required to drive the data 16 base is near completion, and data at collection 17 material procurement and testing at our lab are 18 continuing.
19 MR. S HEWMON : Any questions? Thank you 20 very much.
21 - - - - -
22 Thereupon, the proceedings were 23 adjourned at 2:50 o' clock p.m.
< 24 - - - - -
'n) l 25
1 CERTIFICATE 2, STATE OF O HIO :
I 3 COUNTY OF FRANKLIN : SS.
l6 4 We , Scott H. Gamertsfelder, and Barbara Leonard 5; Rogers, RPRs and Notaries Public in and for the 6 State of Ohio duly commissioned and qualified, do i
7, hereby certify that the foregoing is a true and I
8 correct transcript of the proceedings had in the i
9 aforementioned cause and was completed without i
10 adjournment.
11l We do further certify that We are not a 12 relative, counsel or attorney of either party herein, ll) 13 or otherwise interested in the outcome of this II action.
15 IN WITNESS W HE REO F , We have hereunto set our ld hand and affixed our seal of office at Columbus, 17 I
Ohio, o6 t t
a N day of JQJ U i
f , 1986.
18
- l BARBARAjLEONARD ROGERS, Notary Ph611c -
State of 19 Ohio. My Commiseion expires June 14, 1989.
I -
f orM- :*"' , -
21 SCOTT N. (AMERTSFELdER, Notary Public -
State of Ohio. My Commission expires August 1, 1987.
21 23 CD 25 t
I
_ . - -. _ ..___ _ _ _ . ._ _ ._ _ _~
M M M M M M M M M M M M Mp M M[ {M
{ OTHER RELATED PIPE FRACTURE PROGRAMS h
e AEC/BATTELLE (ElBER)
AXIAL CRACKS AND CIRCUMFERENTIAL CS PRESSURE TEST (550 F) e AEC/GE (REYNOLDS)
AXIAL AND CIRCUMFERENTIAL CRACKS (70 F) e EPRl/BATTELLE (NP-192)
CIRCUMFERENTIAL IWC IN SS PIPE UNDER PRESSURE AND BENDING (70 F) e EPRl/BATTELLE (NP-2347)
CIRCUMFERENTIAL TWC, SC, CC IN SS PIPE IN BENDING (70 F) l e MPA/BMI(PHXNOMEN0LOGISCHEBEHXLTERBERSTVERSUCHE)
AXIAL CRACKS AND CIRCUMFERENTIAL CS PRESSURE TEST (550 F) e WESTINGHOUSE (0WNER'S GROUP)
THERMAL-AGED 4-INCH DIAMETER IWC PIPE IN PRESSURE AND BEND (400 F) e NUPEC/MITI (YAGAWA)
CIRCUMFERENTIAL TWC AND SC IN 4-INCH DIAMETER SS PIPE IN TENSION AND PRESSURE (550 F) l OBallelle
s HITACHI (HASAGAWA)
CIRCUMFERENTIAL IWC AND SC IN 2-INCH DIAMETER SS PIPE IN TENSION (70 F) e DTNSRDC/NRC (NUREG/CR-3740)
CIRCUMFERENTI AL IWC IN 8-INCH DI AMETER A106 B PIPE IN BENDING (125 F) l e KWU CIRCUMFERENTIAL IWC IN }2-INCH DIAMETER CS PIPE IN BENDING (550 F) e DTNSRDC/NRC CIRCUMFERENTIAL IWC AND SC IN 4-INCH DIAMETER SS PIPE IN BENDING (550 F) e MPA/ UTILITIES CIRCUMFERENTIAL IWC AND SC IN 16-INCH DIAMETER CS PIPE IN PRESSURE AND BENDING (70 F) e MPA/BMI l CIRCUMFERENTIAL CRACKED PIPE IN PRESSURE AND EXPLOSIVE BENDING (
e JAERI/STA l
CIRCUMFERENTIAL CRACKED 6-INCH TO 16-INCH DIAMETER SS PIPE IN BENDING (70 F) j l
e NUPEC/MITI (ASADA)
CARBON STEEL PIPE PROGRAM (550 F) i I
l C4Battelle i C j l
- - - - - - a------ -:
t
] p 3
i STATUS OF DP II AFTER 3 YEARS (CONTINUED)
(4) MATERIAL PROPERTY DATA BASE (COORDINATED WITH MEA)
- ENCOUNTERED DYNAMIC STRAIN-AGING
- EFFECT OF WELD SIZE TO SPECIMEN SIZE ON J-R CURVE
- VARIABILITY OF AUSTENITIC SAW TOUGHNESS (5) ELASTIC-PLASTIC FRACTURE MECHANICS ANALYSES TO PREDICT LOADS
- G.E./EPRI TWC ANALYSIS CONSERVATIVE
- PARIS METHOD NOT GENERAL EN0 UGH
- NRC.llR kETHOD MOST ACCURATE FOR IWC
- BCD Sum V CRACK ANALYSES DEVELOPED Al4D EVALUATED
, - COMBilt,^ PR13SURE AND BENDING LOAD ANALYSES EVALUATED
- ASME SECTION XI ANALYSIS VERIFIED
- METHODS TO EXTRAPOLATE J-R CURVES FOR LARGE CRACK GROWTH ASSESSED (6) ELASTIC-PLASTIC FRACTURE dECHANICS ANALYSES TO PREDICT LOAD-DISPLACEMENT RELATIONSHIP
- THROUGH-WALL CRACK PIPE ANALYSIS SUFFICIENT FOR ENGINEERING APPLICATIONS l - INSTABILITY OF SURFACE CRACK GROWTH ANALYSIS DEVELOPED
- COMPLEX CRACKED PIPE ANALYSES NEEDS EMPIRICAL CORRECTION
- WELDED PIPE METHODOLOGY DEVELOPED.
OBattelle N ]
STATUS OF DP3 11 AFTER 3 YEARS (1) LIMITATIONS ON NET-SECTION-COLLAPSE ANALYSIS
- NUREG-0313; NUREG-1061. VOLUME 1.
(2) SCREENING CRITERIA DEVELOPED l - NUREG-1061, VOLUME 3
- GDC li ANALYSIS
- ASME SECTION XI AUSTENITIC AND CARBON STEEL PIPE CRITERIA
- W. GERMAN AND JAPANESE LBB CRITERIA.
- (3) PIPE FRACTURE DATA BASE
- THROUGH-WALL CRACKED PIPE DATA BASE SUFFICIENT
- WELD OVERLAYS EVALUATED
~
- SURFACE CRACK DATA SHOW PIPE R/t IS IMPORTANT. ADDITIONAL DATA NEED FOR DIFFERENT CRACK SIZES
- THERMAL-AGED PIPE TESTS WILL NOT BE CONDUCTED
- CARBON STEEL WELD DATA NEEDED
- BIMETAL WELD DATA NEEDED
- DATA ON TORSIONAL LOAD INTERACTIONS NEEDED.
1 OBattelle
mm um suu uns e sus um aus g aus um um num um um ep e
( h '
NUMBER OF EXPERIMENTS (PIPE DIAMETERS)
THROUGH-WALL CRACKS SURFACE CRACKS COMPLEX CRACKS 4 13 8 AUSTENITIC (6-42 INCHES) (6-16 INCHES) (6-16 INCHES) 7 11 4 CARBON STEELS (4-42 INCHES) (6-16 INCHES) (6-10 INCHES) 6 8 WELDS (4-37 INCHES) (6-16 INCHES)
OBaHelley
TEST MATRIX
SUMMARY
THROUGH-WALL CRACK SURFACE CRACK COMPLEX CRACK v ss v
BEADING BENDING BENDING PRESSURE PRESSURE PRESSURE PRESS & BEND PRESS & BEND COMPUANT BEND COMPLIANT BEND MATERIALS EVALUATED:
CARBON STEEL: A106 GR.B PREDOMINANT STAINLESS STEEL: T(PE 304 PREDOMINANT INCONEL 600 UNIQUE EXPERIMENTS:
WELD OVERLAY REPAIRS SUBMERGED-ARC WELDS COLD LEG PIPE OBallelle
(_
- ---- a------- _) -
l t
e GER t M
I T a CS DS B
M R
S4 NT3 0 )E 2
4 O D C EY
,P O M 8( J 6T , 7 0 I C 3 1
RE M A E/ PA C
D /R RDE 8 SCB l 2
M N T
BE D
(
A S T E M A 4 H 2 C D
N R
I E
M H T
R E
O 6 T g ' 1E M
A I
DD A
/
/ L RG A GN /
M N _
EI DIPf
/ 8M I
O P N _
M .
6 M _
~ ' -
M 4 W I M -
2 0 5 0 5 0 5 5 0 4 3 3 2 2 1 M
l s 0z :_$gL a a_O 8c L 5n@Z M
- . + - --- . . _ - _ _ . . - - m -
_am .-
- _ _
l l
e o.
= E O
c:: as (U E CQ A
'4#
l
( . e w ,. . ,
n :.; e n p,i " ^ L -
,*6
'.q .,?.ff?h? g'41y'Q.; $f '~ ~ '
.g .v4 -93 ,. . . . . E E
.- V,. .2:.c:
' ' W ;. .. .
g" . y U
. . . . . -.,,,t... . ..
E
,. \M . ,
- ' ~ /
,p_ . ':J n .
l
}'v. , . .
I ?, l?' .l, n *. ..
A, .'.. ,
, y. ,_
?. .
, L f. ' ..
'lM?3di.p.$l:~'-.
l I
i .
)
i y'4
.[
J
] ,.[. ~.h' c.-
- c
- h s
, 'k - -
'.r
- - ~
- i. e d i af 1; >A
, 'q )
l f<N -
1 g
' ~
6 e Set up pipe bend test facility prior to conducting Experiment 4111-4 on 42-inch (1.07 m) diameter, X65 carbon steel line pipe.
The overall specimen length was approximately 60 feet (18.3 m).
l
! l I
SA-6/85-F2.1.8 f Battelle
I i
i i
i ng ? , . , .y f , L y ' .qwy ./
l -
t ._ _ n;r l ",
h a ! g. K l?
"? j ;
mi;,hNda V4 c; r j 1 L
y l
rae3BBB=s Post-test photograph of Experiment 4111-2 28-inch (711 m) diameter carbon steel specimen apparatus in bend facility.
OBattelle
% - - a . . s
M M M M M M g M M M M M p
{ h SCOPE OF EFFORTS IN DEGRADED PIPING PROGRAki (CONTINUED) e TEST TEMPERATURES
- PIPE TESTS AT 550 F (288 C)
- TENSILE TESTS AT ROOM, 300 F, 550 F (149, 288 C)
- J-R CURVE TESTS AT 300 F, 550 F (149, 288 C) e ANALYTICAL EFFORTS
- NET-SEC TION-COLL APSE (NSC) ANALYSIS
- DEVELOPMENT OF PLASTIC-ZONE SCREENING-CRITERIA FOR NSC
- ELASTIC-PLASTIC FRACTURE J-ESTIMATION SCHEMES
- 1. e-FACTOR ANALYSIS TO CALCULATE J-R CURVES
- 2. PREDICTIVE ANALYSES TO PREDICT LOADS-DISPLACEMENT-CRACK GROWTH, OR J-R CURVE
- FINITE ELEMENT ANALYSES VERIFY J-ESTIMATION SCHEMES
, EVALUATE GEOMETRY EFFECTS ASSESS ADVANCED FRACTURE MECHANICS PARAMETERS C4Ballelle N Y
INTERACTIONS WITH REGULATORY / INDUSTRIAL / INTERNATIONAL POLICIES (G. WILK0WSKI) e CURRENT NRC REGULATORY EVALUATIONS
- BWR PIPE CRACKING NUREG-1061, VOLUME 1 NUREG-0313. REv. 2.
- PIPE WHIP RESTRAINT AND JET IMPINGEMENT REMOVAL NUREG-1061, VOLUME 3 GDC-4 BROAD SCOPE PROVISIONS.
e FUTURE NRC REGULATORY POLICY PLANS e INDUSTRY PLANS
- ASME IWB-3640 AUSTENITIC PIPING
- ASME FUTURE CARBON STEEL PIPING.
e FOREIGN GOVERNMENT LBB POLICY
- JAPANESE AUSTENITIC PIPING CRITERIA FOR BWR'S
- FUTURE JAPANESE CARBON STEEL PIPING CRITER' A
- W. GERMAN BASIS SAFETY APPROACH.
e TECHN3 LOGY TRANSFER WITH NRC-NRR
- IMPROVEMENTS INCORPORATED INTO NRC.LBB ANALYSIS
- NINE MILE POINT DOWNCOMER PROBLEM i - LEAKAGE AREA ASSESSMENTS.
i e ACRS CONCERNS
- THERMAL-AGING 0F CENTRIFUGALLY-CAST STAINLESS STEEL PIPING
- PLUGGING ON LEAK RATES.
e INTERACTION WITH OTHER NRC CONTRACTORS ggBallelle w
- DTNSRDC, MEA, ARGONNE, BATTELLE NORTHWEST.
l l
i n
i M M M M M M M M
]M M M M M M mp M LORD-LINE DISPLRCEMENT, mm 150000 I I I i l I I 600 125000 -
500 Z j g 100000 -
d
[
C o -
400 ]
1 A Q W 75000 - W 300 E n_ O_
[ [
_1 50000 -
J I
200 g f
o o r H l 25000 -
100 i
l 0 O O 3 6 9 12 15 i
- LORD-LINE DISPLRCEMENT, inches j Total ap 4111-2 28-inch
[ plied load (711 as nm)adiameter functioncarbon of load-line displacement steel pipe specimen]. for Experimentgpg l
I SA-6/86-F2.1.2 %#
1 REVIEW 0F SPECIFIC TECHNICAL TOPICS (1) e REVIEW 0F LIMIT-LOAD CRITERION AND VALIDATION BY EXPERIMENTAL DATA (P. SCOTT)
- PIPE SIZE EFFECTS ON THROUGH-WALL CRACKED PIPE
- PIPE R/t EFFECT ON SURFACE CRACKED PIPE.
- (2) e A PLASTIC-ZONE SCREENING CRITERIA WAS DEVELOPED TO ASSESS WHEN NET-SECTION I
COLLAPSE IS NONCONSERVATIVE (P. SCOTT)
- IHROUGH-WALL CRACKED PIPE DATA
- SURFACE CRACKED PIPE DATA.
e SIGNIFICANCE (P. SCOTT)
- LARGE DIAMETER WROUGHT STAINLESS STEEL PIPE MAY FAIL BELOW NET-SECTION-COLLAPSE PREDICTED LOADS NUREG-0313 REV. 2 ASME IWB-3640 AUSTENITIC PIPING CRITERIA JAPANESE BWR PIPING LBB CRITERIA W. GERMAN CRITERIA 0F CHARPY ENERGY GREATER THAN 50 JOULES IS INSUFFICIENT l
C4Battelle N.
a - )
i OUTSIDE PIPE DIRMETER, mm i
- O 200 400 600 800 1000 1209 y 1.5 i i i i 1 W
in F B C
1.25 F g -E B W
o- N (f l E h E y 1 C
g u ,
Z C o y C
[ .75 -
E D U W
W J I A I
z
.5 -
C c ,,.......
(n St I m 't= ;;;3" w.-x '"::; '
cg .25 -
.c Ce k to Ing 5 Surf..c. J s.. .. Cr.ch-Bendtag c,. + .i & Pr...ur.
r.n.... e a g ,
c i.. c, .
a w x
E O la 20 30 40 50 OUTSIDE PIPE DIRMETER, inches o
f = 3Sm '
SA-12/85-F4.1 h CBallelle
OUTSIDE PIPE DIRMETER, mm y
- 0 200 400 600 800 1006 1200 a1 I I I I I I w
E l--
C g 1.25 -
W F
- a. F x F -
J pd J
O g _E j U E E
'R C Z A O
Cy pT
[ .75 -
B U U y h A w i A
W J E y
d .5 -
J I Z
W N.
(
G Statntoes Steel W Carmen Statnisse kbeerged Incanel y Steel Steel Rre Mold 685 g .25 -
H Throgh-68 ell Crack A E I W krfece Crack-Bendtng 8 F krf ace Crack-Bendtng & Peseewre J
, krface Crack-Antal Tonalen C G y Ceeples Crack D H K x 0 l
r 0 10 20 30 40 50 1
) OUTSIDE PIPE DIRMETER, inches of =1.15(ay + u)/2 -
SA-12/85-F4.2 OBallelle 1
I
l I
l Mean Pipe Diameter, mm I O 250 500 750 1000 1250 1.2 i i i i l
Ol0 m 5 0
81.0 I 07 60 J
3 0.8 -
9 o
O'l 04 N O O2 u
l g 0.6 -
d3 03 J
E I E E
F
'g 0.4 R
j O Carbon Steel data g O Stainless Steel data 1 - 0.2 -
O Stainless Steel SAW data I I I I 0.0 O 10 20 30 40 50 g Mean Pipe Diameter, inches I
I (iN 1 4 s,4.a . 4 0Banele
)
^
I 1
I.2 ,
~
8 l Experimental trend curves / Pmax/PL 8
l L P;/Pt g i.o _ - - - - - - - -
'i # yl , _ _ __. __. .__ 75 m l -
e / i ' 's
.E- / , 7 O,g 3 0.8 - / 8 '
' & / 9 I i
' s g l Fully plastic conditions satisfied
/
.E /A l
' -[ 12 15 Test Temperature g Oli Through-wall 125F 550F I y / 14 crack Ambient (52C) (288C) l I ~ I 17 l Stainless Steel data e o O.4 3 i g Carbon Steel dato a o o l Aluminum data l 8 f A
! 3 ,/f,3 ll (w 4a}O Stainless Steel SAW 3 j data o M O.2 -
l l A - - I '"'.
E i oxis 1 i
.n i i I I I I I I I I I O
O O.5 1.0 2 2.0 2.5 3.0 3.5 4.0 4.5 5.0
! EJ i (2wa,z)/(r-a 4 )o I
i SA-6/86-F2.1.1 CBattelle
- a -
gii;
l.25 B Least squares fit through the data points O'
3 4
,3 which have a value of thescreening e
a criteria dimensionless parameter greater h 07 3 than 1.0 (Solid points >10) y iOO - # 90'3
.5 1
E I40 5
! 12 Os E O.75 -
O i
E
-j n O' o
I
- E 3 0.50
]
Zo O Stainless steel l
g 0 Carbon steel O Stainless steel SAW O
j to O.25 -
i a
. 2 I I I l O.00 d
O 5 10 15 20 Mean Pipe Radius / Pipe Wall Thickness
)
i l SA-12/85-F2.2.3 i C4BaHelle l; ]
CIRCUMFERENTIAL SURFACE CRACKED PIPE DATA i
UNDER BENDING O'
, 8 IS O "4 og , 03 013 10 O2 g l4
' Rp 2 r , 0.75 - 1 + d/t(a) i U i2 6" 2r 4Rteff
+g rg c5 b* '7 Experimental 2a
> 0" trend curve l 1 0.50 -
\ _
'l 1J _ ,
j j N'eutral Axis e
< J based on CT
, specimen g5I O.25 -
O Stainless Steel Data at 550F(288C)
+ Fully Plastic e Storiess Steel Datact 72F( 22C)
Conditions Satisfied O Carbon Steel Data at 550F(288C)
O Stainless Steel Submerged Arc Wek!
Dojo at 550F(288C O O.5 1.0 1.5 2. 0 2.5 3.0 3.5
[ EJ; ) [, + (d/t)a ~ Ri 2 P\ '"
(2re,2) 21 4Rie
_( r> _ Battelle
. .%----- a------ %-
{ h SIGNIFICANCE OF TilROUGil-WALL CRACK SCREENING CRITERION J, Plastic i Pipe Zone " Max Flow Stress, Diaraeter, System / Material 2 2 Size / o in-lb/in (kJ/m ) ksi (MPa) inches (mm) Ligament NSC BWR Recirculation Line 31655 (Base) 4000 (0.70) 60.9 (420) 28 (711 0.29 0.71 308 SAW 550 (0.10) 65.9 (454) 28 (711)) 0.03 <0.30(d)
PWR Cold Leg A516 Gr. 70 1500 (0.26) 64.7 (446) 37 0.08 <0.45 A106 Gr. B (940) 600 (0.11) 66.0 (455) 24 (610) 0.05 <0.35 PWR Steamline A516 Gr. 70 1500 (0.26) 64.7 (446) 28 (711) 0.10 0.53 CF-8M unaged 3000 (0.53) 53.5 (369) 4.5 1.79 0.95 CF-8M aged (114) 1500 (0.26) 57.9 (399)- 4.5 (114) 0.76 0.90(b)
, PWR llot Leg CF-8M unaged 3000 (0.53) 53.5 30 0.28 0.69 CF-BM aged (369) (762) 1500 (0.26) 57.9 (399) 30 (762) 0.12 0.52(c)
SA-12/85-T2.1.2 (a) agax (308SAW)/oMay (316 Base)=0.46 (b) oMax (aged)/oMax tunaged)=1.02 (c) oMax (aged)/oMax (unaged)=0.82 OBaHelle
SIGNIFICANCE OF SURFACE CRACK SCREENING CRITERION J Pipe Plastic Ic' Flow Stress, Zone Diameter, O MAX 2 2 Material R/t in-lb/in (kJ/m ) ksi (MPa) in (m) Size / y Ligament NSC BWR Recirculation Line 316 SS (Base) 12 4000 (0.70) 60.9 (420) 28 (711) 0.19 0.75 308 SAW 12 550 (0.10) 65.9 (454) 28 (711) 0.02 <0.25(a)
PWR Cold Leg A516 Gr. 70 4.8 1500 (0.26) 64.7 (446) 37 (940) 0.05 <0.65 A106 Gr. B 6.2 600 (0.11) 66.0 (455) 24 (610) 0.03 #0.37 PWR Steamline A516 Gr. 70 14.6 1500 (0.26) 64.7 (446) 28 (711) 0.06 0.57 PWR Hot Leg CF-8M unaged 5 3000 (0.53) 53.5 (369) 30 (762) 0.18 CF-8M aged 5 1500 (0.26) 57.9 (399) 30 (762) 0.08 0.85(b) 0.73 (a) a MAX(weld)/oMAX(base) = 0.36 (b) oMAX(aged)/oMAX(unaged) = 0.93.
QBaHelle
i SIGNIFICANCE OF SCREENING CRITERIA l
s NUREG-0313 IMPLIES NET-SECTION-COLLAPSE WORKS FOR ALL WROUGHT PIPE.
i e IMPACTS ON ASME SECTION XI FLAW EVALUATION CRITERIA FOR AUSTENITIC PIPING.
i e JAPANESE.BWR AUSTENITIC PIPE CRITERIA USES NET-SECTION COLLAPSE.
i I
o W. GERMAN WORK SHOWED THAT LIMIT-LOAD SHOULD BE APPLICABLE TO PIPES WITH CHARPY PLATEAU ENERGY GREATER THAN 35 FT-LBS (50 JOULES) IRREGARDLESS OF l PIPE SIZE OR STRENGTH.
1
- CBaHelle i
N ]
0 l
l e
5 2
t t
e
' 3 * , . _
_J l
a l
A B l
5 n, D 0
Q 1, . -
0 2
e s' T e T
e p e ip -
st Pi n n P e e mnEC ic oTS .
p e m3 3 p oP p
- .8 l S C Do O
- = ae
. . 0 51 M
s l
e u
o j o
y m g r -
- e 1 .( 0 n 0 E no 1 y
p
. . 3 . * - r l - * , : i.: a .
- a
=
h C
_ - -M oi'3 i,I Ij
.8 .* */ -A
, ! /
, * ** /
_ : / .
- $ *e * .
I = . .= .= = g. . .
' y. =4.e $i. .
3 0 2 3 O
d d a a o o l
l nt u
- i s m i
x i M
a L -
) : ,l1ll .l'. ;l
- t
! lji: jl ,l
i
( h i
i l
! SIGNIFICANCE OF MATERIAL BEHAVIOR I
1 AT LWR OPERATING CONDITIONS l
I BY 1
M. P. LANDOW AND C. W. MARSCHALL l
i.
l i
QBattelle
! ( )
i l
- e TYPICAL PIPE TEST MATERIAL l
e l MECHANICAL PROPERTIES OF WELD METAL IN NUCLEAR PIPING I
I e OBSERVED INSTABILITIES OF CARBON STEEL PIPING l MATERIALS AT 550 F (288 C) l l
\
l fl0Ballelle N- .) -
_- _ - - - .-____-_--_-_ - a - - - _
" M M M M M M m m m m e e e e M[ []
{ f\ h
/ \
// N l
//
/
\
N N
/l/ A
~
~~-
/ / E i R
\ , D //
\
//
l \
B C ,
ll
- \
\
O 7 ,,
0 //
l \ /
x s /
x .
/
'~
Specimen Types:
l A Full-width-face-notch (tension) - FWFN(T) l 8 Compact (tension) - C(T)
! C Charpy V-notch-CVN 0 Tension - T p*
l E Bend - SE(B)
- Ballelle l
l t
70000
,. - 's 450
',/
\
1 60000 -
l'
- , \ -
400 l//
i .- \
- .- , / \
i 50000 i' l / /
/
\ -
350 l i // / \
I 300 m
[/ /
' * , n.
g 40000 / 550F I
// /
/ (288C) -
250 ui 4
a to J'/ /
' g; m
l 30000 -/ -
200 $
Base Metal 20000 l- -
A45SW-1 Weld Metal A45SW-2 A45WA-1 -
100 W Id Metal Annealed A45HA-2 10000 -
50 I I I I I I I I I 0 O O 0.05 0.1 0.15 0.2 0.25 0.3 0.35 0.4 0.45 0.5 STRAIN. in/in OBaHelle C-a
M M M M M M M M M M M M M M M M l h EFFECT OF WELD METAL PROPERTIES ON FRACTURE BEllAVIOR INCREASING INTEREST IN WELD METAL PROPERTIES IN NUCLEAR PIPING 0 WELDS ARE USED EXTENSIVELY IN REACTOR COOLANT PIPING SYSTEMS I WilEN STRESS-CORROSION CRACKS OCCUR, TilEY GENERALLY ARE ASSOCIATED WITil WELDS 8 GROWING EVIDENCE TilAT WELD METALS DEPOSITED BY A METil0D TilAT EMPLOYS FLUX llAVE MUCil POORER FRACTURE RESISTANCE TilAN DO BASE METALS WE IIAVE TESTED SEVERAL WELDS IN DEGRADED PIPING PROGRAM 8 SUBMERGED ARC WELD (FldX TYPE) IN TYPE 304 AUSTENITIC STEEL (1) AS DEPOSITED (2) ANNEALED 8 TUNGSTEN INERT GAS WELD (TIG, NO FLUX) IN TYPE 304 AUSTENITIC STEEL 8 SUBMERGEDARCWELD(FLUXTYPE)INASMESA-516, GR. 70 CARBON STEEL aBallelle W
u
TEllSILE PROPERTIES OF BASE METAL AND SUBMERGED-ARC WELD METAL IN TYPE 304 SS PLATE DP2-A45 TESTED AT 550 F (288 C)
ELONG. IN 0.2% OFFSET TEllSILE 0 . 5 111. AREA SPEC. YlELD STR., STREllGTil, (12.7 tin), REDUCT10ll, NO. C0llDIT10N Ksi (MPA) KSi (MPA) % %
Af15-1 BASE METAL; AS 24.5 68.9 117(^) 79.0 i RECEIVED l Af15-2 BASE METAL; AS 21.1 67.6 f17. 5(^) 78.9 RECEIVED A45SW-1 AS-WELDED '19 . 1 68.1 30 46.0 AliSSW-2 AS-WELDED 115.0 67.1 33 112.fi Af15WA-1 WELDED & ANNEALED 26.5 66.7 36 48.2 Af15WA-2 WELDED & ANNEALED 30.0 68.2 33 f15.8 (A) ELONGATION IN 1 INCll (25. flan)
C4Battelle
(
-- - - - - - - a------- -
l ~
CRACK EXTENSION. mm 0 2.5 5 7.5 to 12.5 15 17.5 30000 i i i I I I I t
5000 25000 - A.15-37 (Base Metal) 4000 20000 -
E Type 304 Stainless Steel C
- iT Compact Specimens
) - 3000,e n 550 F (288 C) N,
. ~ 2 c: 15000 c E i
2000 j ' 10000 -
{
A45W-2 *
- i ....
~; ;.... -f4sw_1 -
2000 5000 -
(Submerged Arc Wald)
. .p. Z' i l l 1 I I 0 0.p 0 0.1 0.2 0.3 0.4 0.5 0.6 CRACK EXTENSION, inch QBallelle N
l CRACK EXTENSION, mm 0 2.5 5 7.5 10 12.5 15 17.5 12000 , , , , i i g O
2000 SU8 MERGED-ARC WELD D 0
10000 - IN TYPE 304 SS PLATE DP2-A45 0 iT C (T) O TESTED AT 550F (288C) O g X- 1500 O o X 8000 0
- m -
O o x E O o x 0 X 5 O a x u
) O o x (
n a o X >
T 6000 -
O o x 0
1000 E O o X
0 x ,i 5
OX ox O
k 0 X 3
2 4000 - O gx l O %
0 % x A45W-2 A's-weldad Fatigue Crack g _
O g A45W-7 As-welded Sharp Notch 2000 -
S O A45WA-3 Annealed Sharp Notch e7
'I ' ' ' I I I 0
i.' O.1 0.2 0.3 0.4 n.5 na nh CRACK EXTENSION. inches C4Battelle
]
a s
CRACK EXTENSION, mm 0 5 10 15 20 25 30 35 40 45 50 55 40 , , , , , , , , , , , 7 A46-1 (3T) 35 -
(TIG) -
6
/
/
30 -
f
/ -
5
/
c '
g 5 '
A46HM-2 C (TIG)(0.5T) A46-2 (1. 5T) ,' A46-10 (3T) -
4 "E (TIG) ,' N
.E H / (base) 20 / I m '
/ .
o .- '
/ k 33
/ /
~
x 15 -
- A46-4 (1. 5T)
,.- (base) 2 10
./- f
'/. A46BM-2 (0.5T)
~
/[ (base)
- g. A45H-2 -
i 5
(SAW) 0 I I ' ' ' ' ' ' ' ' o 0 0.2 0.4 0.6 0.8 i 1.2 1.4 1.6 1.8 2 2.2 CRACK EXTENSION, inches QBattelle l
k
i i
CRACK EXTENSION. mm 0 2.5 5 7.5 10 12.5 15 17.5 -
, 20000 i i i i i 3500 0 F40-37 Baen Metal 0% SG O 18000 -
O F40-38 Base Metal 20% SG O A F40W2-55 Submerged Arc Wald OX SG O
- 16000 -
0 F40W2-54 Submerged Arc Weld 20% SG O U
A516 Gr70 -
2 00 I 14000 -
iT Compact Specimene 0 "r Fatigue Precracked 0
0 C 12000 -
se e a 0F (288C)
O
} O -
2000=e n a n 7 10000 -
O 6c- O A 0 O@
" O 1500 7 8000 -
0 0 O 0
, k o 0 1 7 O 0 0 8000 -
u G l a 0gO -
1000 i O 00 i
4000 -
O O
00 o0 g O AAA -
500 O
2000. gA A,fo o0 g
pek s a n i e n l 9
'i.l 0.1 0.2 0.3 0.4 0.5 0.6 0.7 CRACK EXTENSION, inch CBallelle N-
.. . a - )
mm mM NNM M C] m m m m m m m p m
{ h OBSERVATIO.'lS OF FRACTURE INSTABILITIES IN CARBON STEEL AT 550 F AND POSSIBLE ROLE OF STRAIN AGING f
QBalleIIe
- N J
l l
=
O N 'peol
- =
(Q o
o o
o o
o o
o o
o o @
8 m
8 N
8 N
8 8
8 m o m, #4 9p 4
i l i I I I '
o tn
- N l o
w -
E G $
_ N. ,
y ou O C C H E E 1 m e CD 8 E m -y _ m-1 l
@4 9S ow m o o
- d. m e Q 5 5 i J g J
t
.o e oe O O J J N
tn O,
o 1 I I i o e o o o o o o o o o o o o o o o o o o o o o o s e in v m N
- sqt ' peal I
{ l J
mm m m m m m m g m m m m m m up m
( h STRAIN AGING STRAIN AGING IS THE CHANGE IN PROPERTIES OF A METAL DUE TO THE INTERACTION OF DISLOCATIONS AND INTERSTITIAL SOLUTE ATOMS (PRIMARILY NITROGEN AND CARBON) DURING OR AFTER PLASTIC DEFORMATION 9 STATIC STRAIN AGING--PROPERTY CHANGES OCCUR AFTER PLASTIC DEFORMATION 4 DYNAMIC STRAIN AGING--PROPERTY CHANGES OCCUR I DURING PLASTIC DEFORMATION l
OBaHelle
STATIC STRAIN AGING Aged ged 7
g a.
i Prestrain ELONGATION EFFECTS OF STATIC STRAIN AGING 4 INCREASE IN YIELD STRESS AFTER AGING 8 RETURN OF YIELD POINT ELONGATION (LUDERS STRAIN)
. I INCREASE IN THE ULTIMATE TENSILE STRESS 8 DECREASE IN TOTAL ELONGATION 8 REDUCED FRACTURE TOUGilNESS OBallelle
- - - - - - a------ -
NN mhmM bmgM M M M M M Mp M
{ h CilARACTERISTICS OF DYNAMIC STRAIN AGING 8 ACCOMPANIED BY A IIIGil RATE OF WORK llARDENING e Tile STRESS / STRAIN CURVE IS SERRATED 8 GREATER STRENGTilENING TilAN STATIC STRAIN AGING 8 DEPENDS ON BOTil TEMPERATURE AND STRAIN RATE e EFFECTS ON FRACTURE RESISTANCE ARE NOT WELL ESTABLISilED l
l OBallelle
E{ a 5
)-
0 0
6 0
0 5
0 0
4 0
0 3 2 0
0 0 0
1 0 7
1 6
A l
l e
- 9 A,
s l
l e
n T
a r a C
t e
B Q .
0 M.
0 e 6 i l
s c.
l C.
C W 0 d 5 0 n 5 a i
L C.. -
C(
C 0 -
0 s 4
- 0 G 1 N X I
C 5 G
- 0 7
A 0 l .
J =
N i I
% n, A ,
o R n T C i o is n _
S
- 0 t
a t e
C 5 g n 2 n i I '
o d M l E
e n
A l i a
N r Y ts C
D
- l e
0 e 0 ts 2
n o
i lr
- a c
C a
- f 0 o 5 s 1
e v
ru c
0 n
o 2 C i 0 _
- t a
1 0 g 0 n 1
_ 1 l
o 5 c 1 -
s A s e
tr S
C 0 5 1
2 1 1 .
)
- _ - ~ - - -
- - e8 r
0 0 0 0 0 u7 9 7 # 5 4 0
3 0
2 1 0 ig9
)- I" i l
1
(
(
I
.p..,_. .__...r.. . - . ...
.a .. ._-..t.. .n4 _._ , . . . .
}... .
.__ -. ....9....l.. . . . ...1 . .. . ..
'- ~ ~ " . . . . _" ~ - * - ---- ~
__ Specimen F29-3 ~ ~ f ~ b -~ ' _ . _ . .'.
_. . J ___._. . .
.4 -_ _--
Tested at 300 F (149 C) _ -__. ....._. . _ . -. _ _ _ .. - . _ _ ...._.t..-..~ .._.p .. _.__.....
...p._____.-- .
-g _...
.. . _ _ _ ; ..% , , , , -~ - - . - . . - . .
.i . .. . n., --- -
t 5000
--T' ~~~
~
~-- "I-e - "I > . . <. , .*- . . . -
--~~" -~
~
a _ .g.. . . .. . . _ _
...._+
j ,
. . _ . . j_ __
.{ _ . . . . . .. . . _ . . . _ _ . . . .
.__q__.
r
. .p( . . . . ..._.., .
i _ . . . . . _ . ..
_ . .9__ _ g.,._..
. . . _ . . A: . . .. . . _ . . . ._. _ . _ - _ . .._. . . . . . _.4_.. ._. _
Yi. .
... w.
l.-_
. . . . .. .s. .
._. h.'.. ... ... ._.4._._
- i. . ._ ,
.t 1000
.g... p_...._ ._,___
.)...
.__.~ _. _ _ _ . .
, ..l..
__ . _L __ _ . = . - _ . . __ _ . __ _ ._. _. . _ _ _ _ . l 1 . . _ . . . . - ._ . . _ . .
i .. . 4.. c .._ _.. ._ _ .
. . . _ . . __ I._ ___ . . _ _.___ . _._ _ _ _ _ . __. . . _ _ . .
.. . ._ . _ _ . __ p__ _ _ . . _ _ _ _ _ _ _ . - ._ . _ _ _ . .
3000 _ _ _ _ _ . ._ . . _ - . . .
. ._. .. . 1 _ _.....T__.-.
..__1.._._.. .. _ _ .. . _ _ .
____4...._
.. t ...__ _ _ . . _ , . . . ._
l._.
... t . . .
. . . _ . ._.p... .
_ __ j __ . _
~ O ~-12 ~=" O..'T
.._u. . . . ._ . .. . _. _. _ . . . ,_ . .
~~M
.f.2500 ' 10.M- - 0 - %" ~ ' 16 " ' 1
---". .X._ _
.. t '. ... ._ _ _. __'.
.l. __
Stra.(n. .. ._ ___..______ ..___ ._....n._ _
.'..._. .___g... _ . . . _g. _ _ ._ _ ' _A_ . ._ . . . _ .
p t . .. .. _ . . __ _ . . _ .. . . . ._
r
~ _ _ _ _ ...___ +__ _._ __
.. . __. _._ ._ ..,. _. ..._....4___. _..
- . .. . ..,r.. .. ._ ._ .
g }. .._A g
}
..._.1...
..{....._.
. .., _ . ./%
. f. . ._. .
4...
...._4... ._.._}_. . . . .,. _.
,20nn ww ._a_. . . _ _ . . _ _ _ _ . ._.
- v. .
__ ..i, ..
_ ..j. . . _.
. . _.7.__
y j , . - . . . .. ..... ._ . . _ . . . _ . . . . . . . _ __ . . . _. .
c e .A..g . _
x _. _ . _.p
_a .. __. _ . _ . . . _ . . . .
, .[.
__ . .T. w. . _. n . . . _ .. ... . . _ . _ . .
.}. .
. ._. __ .f.. .._
-. _ , _ .~~ . ____. __ _.
'3
~
.! ~ ~
1500
., , . _ _ .,j .
__t,.'--.. . . . . . .i, - _ . . . .. _ [.
-.6
.(.
..}.. . . v. p. . . _ g . . . . , . . . . . .. . .. . . . _
.. .L.
. .l. . _ . . . .. _. ._ . . . .
.2. _ _ . . _. . _ _ . . _. .__._u -_ . _ . I _. . . _. _ _ . . . _ . _
4 ._
.}. .- _. . . . . . . ._. . .
.. _ . . }_-
. ,1000 ._ .. . . __ . .._ ._. . _...__. . . . _ _ .__.g___. _ . _ _ . _ _ . . . . .
NN .._W __ . m' .N,_.M_ . _ . .__
_.._.9.__. . _ . _., . ._ _.g _. ..,._.. . . .
n_ ._ .._..
. . .._ f __ _ _ _ . 1. . . . {!._. . _ . . _
_ ll
- ._1___..!_ _ _._ __. .... .._. ;; . _ _ _ _ . . _ _ .._ ___.9....
-T 4---:~ .
. . . _ .T_:.t:g.9} __
u , _ _ . _ .._
J , _ ___ l
. . . _ . j.__....__ . - . . . _ .. _. .__ .__ ._ _ _ _ .
._..-t-- - _ _ _ . . _ _ _ _. ._
_J- .
4.. ___ ._.
- ' ' - - ~~~
~~ '-~
l -
.-._4_0'0
~
0.002 0.004 0.006 0.010 0.012
(
'0 .' 00 8'-
str 4n OBattelle I
l
.. . _ ..l... _____. ..
Specimen F29-5 l J Tested at 550 F: -
1 (288 C) . .'... ... . . . . . _ _ _ . .., ... . . . _ . .. . . . _ . _ _ _ . . . . . . _ . . .
l 1
2400- - - - - - - - - _ - ~ - - - - - - - - - - - - - - - - - . - - - - l l
2100_. - - - _ . . . . . _ _ . - - - . . _ - - - - . . - . _ - - - - - - . _ . . - . . . .
li 1800 - - . - - .- - - - - - - - - - - - - - - - - - - - -
I i
I \
m - 1500 -- -
=.
.o o
m O
_J 1200 - -- - - - - - - - . - - . - - - - - - - - - - -- --- - - - - - - - - - - - - . . - . .
I i
l-900 . ...... _ . - - __ = _ . . . . _ . . ._. . . .
I ,
. .: . 4 :
. .V , -
- = * - * - .
- 600 --
4 r- -....,' --
r
. .....g.
. _ . . . - - - 6000
_._".r.
.,.. .- ,[.- . _
7...-.-._-- _
H..
.____4,..
.i . . . . . 1 . , _ .. .
4
' g .
8 300 -
- - - 3000 m
.o 7 4
- s. -
... ;.. ... . e
....,_-.__.4..4,.,_... _ ,,
. _ _ m o
-q_-.- ..,
.. , . . . .. . . . . . _ _a 0- 0 c m r so.ioio arn -
0 0.004 0.008 0.012 0.112 0.212 gp3Balfelle w
STRAIN
46 3 e
=
O g ". C0 a
I m.
vfe O
E I - a .
O l ii o _ N.3 I ?u.
p e
N o m3 i
U- o -
9$
= ca.
o_
I Q i o
~s !
o I I
y 9 ) l I I o O I o o O g
$ $ $ O R i i e a a S g sat paa, It
e
{ <
0 -
= l m, b b
cc y A g Yef l,
_ E d
5' m 5 s:
E I cu _ NS 7 5$ l ol i N o _ 98 yta O oE 4 cu E l
U-a - SE.
l o
_ 1 E O
l m
I I
I I I I I i ,
S $ $ $ $ $ $ $ $ $
$ $ e $ S e S $ a sqt puol MI.
l:
l l
l e
e n ,
l l
e i io.
t i
eC a
ni ol sk s e o B nL o it o
a r
r e
Wl C. i l
l l
l i
O s d w f n a oa r e ,
G c a c 4 ndM e a 5 l 3 s k e 2 e r ah p
c N 'f t
i 2 iY o 3 lt n S r N lio oci s O
I 0 f
K s i T I d 3 e .
m A i r S. e r R uA p q(
O ER I 8 e r
ei NS 2 e *C i l
C 00
- 1 i m05 t t
i
. 1 S g2i E 6 e
- R N 1 2 K r n w U O ) i T
I T T
, Tod
/ f s r e A A 4 0
/ 1
(- de ,
u R 0 R 1 2 00 i 8 E 5 1 l
R )
i lc6 P E 1 ( n9 M S e e1 E 2 t u T 1 a q r
t 0 i 2 .li 0 I inle ig r 2 a et y r s p S 0 t N i s o 0 O 2 r Cc a
5 2
1 I T
c%
i 3 s, l c A 8 s 0. n i R
l 1 c0 i
r a
R T fay n l
OE G - ~
NS . oD 4 - 7 s 3 2 4 6 1 e n ).
- ~ - 1 1 v
r ir ok 0 0 0 1 ut ao 1 1 1 e c cY r
$ 5 E s<E mE<tI*
u u wlo w g
oies 7 = ;
Ff DN l l 1ll
SIGNIFICANCE e MATERIAL ANISOTROPY CAN SIGNIFICANTLY EFFECT CRACK-GROWTH BEHAVIOR.
e SOLUTION ANNEALING 0F AUSTENITIC STAINLESS STEEL WELDS REDUCES THE l
YIELD STRENGlH BUT DOES NOT CHANGE FRACTURE TOUGHNESS SIGNIFICANTLY.
PIPE TEST RESULTS SHOW SOLUTION ANNEALED PIPE WELD HAD LOAD-CARRYING CAPACITY REDUCED IN DIRECT PROP 0RTION TO CHANGE IN FLOW STRENGTH (AVERAGE OF YIELD AND ULTIMATE).
e STAINLESS STEEL TIG WELDS HAD TOUGHNESS GREATER THAN BASE METAL OR FLUX WELDS, e CARBON STEEL SUBMERGED ARC WELD (SAW) HAD LOWEST TOUGHNESS OBSERVED IN PROGRAM. TOUGHNESS WAS HALF 0F STAINLESS STEEL SAW.
!/$Battelle
--k_---- a------ -)-
M M M M M M M M gM M M M M M Mp m
( h e SERRATIONS FROM STRAIN-AGING OCCURRED IN OUR CARBON STEEL TENSILE TESTS AT 300 F BUT NOT 550 F.
e DYNAMIC CRACK JUMPS OCCUR IN CARBON STEEL C(T) FRACTURE TESTS AT 550 F BUT NOT 300 F.
e DYNAMIC STRAIN AGING IS BELIEVED TO BE THE CAUSE OF THE CRACK JUMPS AT 550 F DUE TO THE HIGH CRACK-TIP-STRAIN RATE.
e DYNAMIC STRAIN-AGING MAY EFFECT FRACTURE BEHAVIOR OF CARBON STEEL PIPING AT SEISMIC LOADING RATES.
OBallelle.
REVIEW 0F SPECIFIC TECHNICAL TOPICS (3) e FRACTURE EVALUATIONS OF WELD OVERLAY REPAIRED PIPE (P. SCOTT)
- CONDUCTED THREE WOR TESTS ON 6-INCH-DIAMETER PIPE AT 550 F
- EVLAUATION OF NET-SECTION COLLAPSE AND IWB-3640 CRITERIA.
e SIGNIFICANCE
- WOR PIPE FAILED WELL BELOW NSC, POSSIBLY DUE TO RESIDUAL STRESSES IN WELD
- WOR PIPE FAILURE LOADS WERE SLIGHTLY BELOW ASME SECTION XI IWB-3640 SOURCE EQUATIONS (NO SAFETY FACTOR USED IN ASME ANALYSIS)
- WOR PIPE FAILED SLIGHTLY ABOVE THE ASME SECTION XI IWB-3640 TABLE VALUES (NO SAFETY FACTOR USED IN ASME ANALYSIS)
. - FOR LARGER DIAMETER PIPE, CONTAINED PLASTICITY MAY LOWER FAILURE LOADS
- MANY DIFFERENT WAYS EXIST TO PREDICT NSC UNDER PRESSURE AND BENDING.
e FUTURE PLANS l
- ONE 16-INCH-DIAMETER PIPE WITH WOR TO BE TESTED TO ASSESS PLASTIC-ZONE SCREENING CRITERI0ll AllD IWB-3640 ANALYSES.
OBattelle C j
, a
fikmN) 6} inch O.D. *A Internal I (168 mm) Surface Crack
[_ ---_)
I g ( ____
Weld 07erlay j _- _
Pipe Thickness I Repair M ~0.562 inch (14.3 mm) p Thickness of Weld Overlay ~0.314 inch (8.0 mm) a
, Internal Surface i Crack f ,
= = Clip gage location I g i * = d-c EP locations
/
\
x /
\~_ ,e SECTION A-A l
OBattelle
((
g sa e,es.e2.e.1 y
I
6-INCH (152 MM) NOMINAL DIAMETER STAINLESS STEEL WELD-OVERLAY-REPAIR (WOR) EXPERIMENT TEST CONDITIONS Outside Diameter: 6.625 inch (168 mm)
Pipe Wall Thickness: 0.566 inch (14.4 mm)
Overlay Thickness: 0.325 inch (8.25 mm)
Effective Thickness: 0.891 inch (22.6 mm)
Inside Radius: 2.746 inch (69.8 mm)
Effective Mean Radius: 3.192 inch (81.1 mm)
Internal Pressure: 4750 psi (32.75 MPa) 2c/7rD: 0.50 e
a/t: 0.635
! Test Temperature: 550 F (288 C)
Base Metal Material: SA376-Tp304 Stainless Steel Flow Stress (3 Sm): 50.85 ksi (351 MPa)
Flow Stress; weld metal (1.15(a y +ou )/2): 57.9 ksi (400 MPa)
" max : 19,680 psi (136 MPa)
(
t OBallelle) i M 'M M M M M M M M M M M M M Mb M
AGENDA FOR DEGRADED PIPING PROGRAM PRESENTATIONS AT ACRS MEETING JULY l-2, 1986 8:45 - 9:10 AM DEGRADED PIPING PROGRAM OVERVIEW G. WILK0WSKI 9:10 - 9:35 AM LIMIT-LOAD ANALYSIS AND PLASTIC-ZONE P. SCOTT SCREENING CRITERION 9:35 - 9:55 AM SIGNIFICANCE OF MATERIAL BEHAVIOR AT M. LANDOW LWR CONDITIONS 9:55 - 10:10 AM BREAK 10:10 - 10:30 AM FRACTURE EVALUATIONS OF WELD OVERLAY P. SCOTT REPAIREn PIPE 10:30 - 10:45 AM INTRODUCTORY COMMENTS ON ELASTIC-PLASTIC J. AHMAD FRACTURE MECHANICS ANALYSES 10:45 - 11:10 AM LOAD AND LOAD VERSUS DISPLACEMENT F. (Buo) BRUST PREDICTIONS FOR THROUGH-WALL CRACKED PIPE 11:10 - 11:25 AM COMPLEX-CRACKED PIPE EVALUATIONS G. KRAMER 11:25 - 11:45 AM FINITE ELEMENT ANALYSIS VERIFICATIONS V. PAPASPYROPOULOS OF ESTIMATION SCHEMES l 11:45 - 12:00 N00N CLOSING C0tiMENTS AND DISCUSSION G. WILK0WSKI C4Ballelle ,
NRC DEGRADED PIPING PROGRAM - PHASE II (DP3 II)
PROGRAM MANAGER: GERY M. WILK0WSKI DEPUTY MANAGER: DR. J0llN KIEFNER TASK LEADERS: JALEES AllMAD, 6 REG KRAMER, CilARLES MARSCHALL, PAUL SCOTT, V. (PASU) PASUPATHI KEY ENGINEERING STAFF: DICK BARNES, F. (bud) BauST, NU GHADIALI, DAVE GUERRIER!, GERALD KULil0WVICK, MARK LANDOW, BILL MAXEY, MICii NAKAGAKI, RICK OLSON, VICTORIA PAPASPYROPOULOS KEY SUPPORT STAFF: ELISA ALEXANDER, BRENDA BLANTON, B0s GERTLER, PAUL HELD, CilUCK MIELE, PAUL MINCER, DENNIS RIDER, JOE RYAN, DALE Sil0EMAKER, JIM WOOD CONSULTANTS: D. BROEK, J. PAN, C. POPELAR, G. WORKMAN VISITING SCIENTIST: K-H. HERTER (MPA - W. GERMANY) 4-85 To 11-86 OBattelle m _ - . _ _ - - - - - --- -
3 DP 11 PROGRAM OBJECTIVES (G. WILK0WSKI)
VERIFY, IMPROVE, AND DEVELOP FLAW ASSESSMENT ANALYSIS FOR LEAK-BEFORE-BREAK METHODOLOGY.
EVALUATION OF METHODOLOGY
- EXPERIMENTAL ASSESSMENT OF ACTUAL FAILURE MODE.
- LIMIT-LOAD ANALYSES OF CRACKED PIPE.
- SCREENING CRITERIA TO ASSESS LIMIT-LOAD ANALYSES.
- SIGNIFICANCE OF MATERIAL BEHAVIOR AT LWR CONDITIONS, l.E., WELDS, DYNAMIC STRAIN-AGING.
- FAILURE EVALUATIONS OF PROTOTYPICAL WELD OVERLAY REPAIRED PIPE.
- ELASTIC-PLASTIC FRACTURE MECHANIC PREDICTIONS OF LOADS AT INITIATION AND THE MAXIMUM LOAD, AND LOAD-DISPLACEMENT RELATIONSHIPS.
- FINITE ELEMENT ANALYSIS VERIFICATION OF ESTIMATION SCHEMES.
TECHNOLOGIES INVOLVED
- MATERIAL CHARACTERIZATION.
- FULL-SCALE PIPE FRACTURE EXPERIMENTS.
- FRACTURE MECHANICS ANALYSIS ESTIMATION SCHEMES FINITE ELEMENT ANALYSIS.
QBattelle
SCOPE OF EFFORTS IN DEGRADED PIPING PROGRAM e CIRCUMFERENTIAL CRACKED PIPE
- SIMPLE THROUGH-WALL CRACKS
- INTERNAL SURFACE CRACKS
- COMPLEX CRACKS e TYPICAL MATERIALS
- AUSTENITIC STEELS TYPE 304, TYPE 316, INCONEL 600
- CARBON STEELS A106 B, SA333 GR 6
- AUSTENITIC WELDS TIG, SMAW, SAW, SAW-SOLUTION ANNEALED
- CARBON STEEL WELD SAW(2)
- CENTRIFUGALLY CAST STAINLESS STEEL (CF8M) e TYPES OF LOADING
- PURE BENDING
- PRESSURE INDUCED AXIAL TENSION
- PRESSURE AND BENDING e PIP 5 SIZES
- DIAMETER FROM 42 TO 4 INCHES (1,067 TO 102 MM)
OBallelle
(
M- M
- THICKNESS FROM 0.25 TO 3 INCHES (6.35 TO 76.2 MM)
M M M M m mm m
^ -
( )
EQUATION RADIUS USED TO USED TO FLOW BASIC CALCULATE CALCULATE STRESS A
CASE FOMULATION MEMBRANE MEMBRANE DEFINITION NUMBER USED STRESS STRESS USED L(ksi) U+UA L U L(Case 6)/UL 6 Net-Section-Collapse Thin-wall inside 1.5(Uy + gu)/2 34.37 0.648 1.00 7 Net-Section-Collapse Thin-wall inside 3S m 29.20 0.739 1.18 8 IWR-3640 Source Eqns Thin-wall inside 3 S, 27.87 0.767 1.23 9 IWB-3640 Source Eqns Thin-wall Outside: 3 Sm 24.58 0.857 1.40 10 IWB-3640 Source Eqns Thick-wall inside 3 Sm 29.23 0.731 1.18
( -- - -- -_ ._- - - _ - - ..
OBallelley
I 08 Note: d/t =0.65 and 2c/wD=0.5 Net-section collapse i p=ll50 psig (7.93 MPa) 0.7 m (No safety factor) 2 p=4750 psig (32.75 MPa)
N
% 3a p=3750 psig (25.8 MPa)
N g 3e p=5700 psig (39.2 MPa)
N 06 - N C=l.15(ey + "u)/2
, , , s
'i>
' s O =3 s m N
N s
0.5 - c)
J. N.
N
@ 'N 4 0~4 -
< B @
O @
Q3 -
8 a2 -
ai I I I
o O o.1 0.2 0.3 P,/ef OBattelle SA-12/85-F2.5.4
0.8 f
Note: d/t=0.65 and 2c/wD=0.5 (3 p=ll50 psig (7.93 MPa) 0.7 q ion h p=4750 psig (32.75 MPa) l N g (so) p=3750 psig (25.8 MPa)
I N @ p=5700 psig (39.2 MPa)
N 0.6 -
N %,
N IWB-3640 source s quations e
I o.5 -
N N
IWB-3641 Table s (Safety factor =l.0) N
(;~a) g\
0.4 -
g sb o.3 -
p 0.2 -
IWB-3641 Table i normal conditions o.1 -
(Safety factor =2.78) i e L l2 a, Pm/#
f OBaHelle g(
ss.12,es.,2.s.s I )'
I l --- - - - -
f 1.25 h Trend curve based on assumptions of BWR
$ conditions, Schedule 80 pipe, and a two-E I.OO layer weld overlay repair a
e Pm = 0.04(3Sm) 8 e g X 16-inch Schedule 100 pipe with g Pm= 0.17 (3 S three-layer WOR
$ O.75 I; Pm = 0.20(3Sm) 2 N
E si O.50 -
0 15 e
.5 t
E O.25 -
w E
a O
- = Experimental 2 g go I I I I I O 5 10 15 20 25 30 Outside Pipe Diameter I
I
)
l l
SA-12/85-F2.5.6 i S%eBanelle
e u p num num aus e uns um rg mum e e e um sum uma p sum I
PROGRAM OBJECTIVES
! Develop, improve, and verify predictive fracture i mechanics analyses for nuclear piping. The program j integrates, advanced elastic-plastic fracture mechanics l
material property testing full-scale pipe fracture experiments (<25 lbs TNT) to develop accurate engineering estimation schemes.
1 i
OBallelle k- -- -- --
boratory Specim i Material Property Evaluations Limit-Load Analyses Full-Scale Simple Pipe Experiments 4
Predictive Elastic--Plastic j
Fracture Mechanics Advanced Estimation i
Elastic-Plastic Schemes racture Mechanic OBallelle
f h l
( )
l l
1 i
- NET-SECTION -
l COLLAPSE ANALYSIS SCREENING CRITERION a
ELASTIC-PLASTIC FRACTURE MECHANICS i
ANALYSIS i
I i
I l OBattelle 4
1
1.2 ,
8 l Experimental trend curvesJ'Pmax/PL J e ,[ P;/Pt
- N 1.0 - ! ----g----- 0
~ ~ ~ ~ ~ ~ ~ ~ '
] #
- /
. / Og E O.8 - l 8 0 l
k / 9 lgFully plastic conditions satisfied
_g /
E / l Test Temperature 5 ~/12 ,3 l 125F 550F j f Oli ) Through-woll Ambient (52C) (288C) u '# CM j
f2 '7 Stainless Steel dato e O E ot4 Carbon Steel dato a m O 3 2.
g ,
Aluminum dato A
, J ,3 u
n (,4_,)D Stainless Steel SAW
'5 L I dato D c A-q O.2 - - l '"'.
E oxes
! l i I I I I I I I I O
I, O O.5 10 2 2.0 2.5 3.0 3.5 4.0 4.5 5.0 (2ra,2)/('4 )D Ratio of load at crack initiation to predicted net-section-collapse load as a function of a dimensionless plastic-zone size parameter.
i i
SA-6/86-F2.1.1
m h ey p l l
H e
a B
m Q m
m EI S VS T LY I
C A I
m DN EA R
PG N
S m T DIRS E
R O T E A ED O F N F IDI L GTH m E AN E VEM S
a I
S Y
L A
N A
m F O
T C /\
m U D
R O E R T G P U m PI N MR OE E C
m YG L N IN D
N E
E R K m I ROS FFI O
O S B Y D RE L m EDA N SON A UCA H e
m e y{
m u
m
~
I APPROACH f GENERATE SMALL SCALE PERFORM EPFM GENERATE FRACTURE = ANALYSIS OF SMALL ~ STRESS-STRAIN EXPERIMENTAL DATA l
SPECIMEN DATA DATA (GENERATIVE / INTERPRETIVE)
I V
FINITE J-R CURVE g DEVELOP / MODIFY
)
ELEMENT ANALYSIS METHODS ANALYSIS i j y y y l
PERFORM EPFM PERFORM NSC g ANALYSIS OF PIPE ANALYSIS m (PREDICTIVE) u u l PREDICT LOAD, PREDICT DISPLACEMENT AT INITIATION, FAILURE LOAD h
INSTABILIT(,
h DURING CRACK l
. GROWTH A
g V U GOOD COMPARE WITH g PREDICTION PIPE EXPERIMENT m NO DATA 7
DOCUMENT YES AND g
^
- COMPUTER CODE l; OBaHelle k
i
( D J-BASED ELASTIC-PLASTIC FRACT!!RE MECHANICS METHODOLOGY INITIATION:
- ------- - - - - - - - - - - - ~ ~
J =J (PLANE STRAIN) ,
IC ;
J =J i ,g ______ _____________
STABLE GROWTH: do s, .
Ao (a) Ma,erial J-Resis,once Curve J(P,aa)=J MAT (^ ) J .
J"' ' '
EJ-R CURVE '
. Applied
'e' 4 cu,ve, J
',~
o, [,",[ ~ ~ ; UNSTABLE INSTABILITY: '.'el%f,,,-
u.,e, .
STABLE ("' a 3**'
dJ dJ 3rc -- -- ~~---------- -
MAT da dJ E
T>TMAT m.,,-
0Ballelley i s, r,_,. ,e 1.s, s. ,,,,
n ". f
- T ,_
n t po r
,1 1
k E S
O R, i
E U b o S R E S
I V a R
- U SCH T S
_ D D S
_ LRW O D E O HO WJR T H G E T NF ME I
OK S M
- C CR YI S A
T R NC E
L AY L
= t C C DF NA E
F NO E
AN JET F A FO ENPT OF TN DUO Y O
.\-\ NOI Y M T Y .
0
\' M ka IATR c
I LC I
BA A R
/'/' C SILM NPOR OPE A G LU A C I
E g
M l l
e l
l e
M a B
M 2 k O
C c a
M N C r
l l
a M W -
e h l
c e g ir at fc ak c u t n M gm y io nh u rar i a o r
g ISSC lp h k n m T c i
r a a o r M T e C
- C e
m h e)
' e g t
]
l l
iC t
c(
u 2 S t
s y t o
e C
D k c
a in g
N D u
M C r s a N t n
d e r e e a c c
a L o n N m e
M I c f \ a S
t r
u i
m u
m h
gC\
g% l i
D p
s M r o
x a \ C 2
h S
M \\ g C
M
\
\g \ g \
lI !
- Il l M '
d a
o M L 5 A
F 5
8
/
4 M
M M
l i l
{ - T<
i n\ IM,$/2 a
t M,$/2
[f y i _ga _ _ __, _ __
l t k --- - - - - - - - - - - - - - - )
U (a) Surface-cracked pipe under pure bending b
a
+- t l' _ _ _ _ _ __ -
0"
,Rj Neutral Plane l
(b) Cross-section containing internal surface crack cr(R) c(R) cr b
. - a a
=
l: L y: L -
(c) Plane at r < 6, L = ct I
Geometry and loading of a pipe with an internal circumferential surface crack.
SA-6/86-F2.2.1 * *Ballelle
mM M M M M M M
{mM M M M M Mp M l h TOPICS TO BE COVERED TODAY e ASSESSMENT OF TWC PIPES IN BENDING BY
- EXISTING J-ESTIMATION METHODS
- MODIFIED J-ESTIMATION METHODS (INCLUDING LOAD-DISPLACEMENT PREDICTIONS) e ASSESSMENT OF COMPLEX CRACKED PIPES IN BENDING BY J/CTOD ESTIMATION METHODS e SELECTED FINITE ELEMENT ANALYSIS RESULTS
( - - -
$3BaHelley
I
{ ANALYSIS PURPOSE: PREDICT THE STRUCTURAL RESPONSE OF FLAWED PIPING SYSTEMS. ANALYSIS GENERAL, AS OPPOSED TO EXPERIMENTS WHICH ARE APPLICABLE ONLY TO THE CONDITION PRESENT FOR CURRENT EXPERIMENT
! METHOD: J-ESTIMATION ANALYSIS METHOD f
MATERIAL PROPERTY DATA 8 J-RESISTANCE CURVES 6 RAMBERG-0SGOOD CONSTANTS (STRESS-STRAIN CONSTANTS)
WITH MATERIAL PROPERTY DATA DEFINED, THE J-ESTIMATION TECHNIQUE PREDICTS O CRACK INITIATION 3 MAXIMUM LOAD l 8 LOAD VERSUS DISPLACEMENT BEHAVIOR l VERY GENERAL - CAN PREDICT THE BEHAVIOR OF ANY SIZE PIPE g g,
I QBallelle I
i
V l J-ANALYSIS METHODS D
I
$ GE/EPRI COMPILATION OF NUMERICAL SOLUTIONS DEVELOPED VIA THE FINITE ELEMENT METHOD g
PROBLEM: GE USED SHELL ELEMENTS - PERFORMANCE PROBLEMS
- TENDS TO SIGNIFICANTLY UNDERPREDICT LOADS
- POOR DISPLACEMENTS g 8 PARIS INTERPOLATES BETWEEN THE LINEAR ELASTIC SOLUTION (KNOWN) AND THE FULLY PLASTIC SOLUTION (KNOWN)
PROBLEM: INDEPENDENT OF MATERIAL HARDENING p
8 LBB-NRC PARIS METHOD MODIFIED TO ACCOUNT FOR MATERI-AL
HARDENING l
8 LBB-BCL TWO COMPANION TECHNIQUES IMPROVE DISPLACEMENTS I (A) BASED ON NUMERICAL SOLUTIONS (B) AN ALTERNATIVE TO PLASTIC-ZONE SIZE TO
! ESTIMATE THE REDUCED COMPLIANCE DUE TO l
THE PRESENCE OF THE CRACK.
l I
I I
OBallelle (N
i )
s
0
)_
5 e
0 5
2 1
I I R
P C
R l
l e -
E N. H
/
E G
B B
L B a -
]
0
/
/ l d
a I
0 4
O -
0 ^/ e o teL I
0 1
'/
/
Smsu s
e em e
l i ia t
nx aM S t s m05 m7 I
a i 0 e 3 hc i
n r r, t
e t e
e e a
m m i
a i a
D D e 0 p e
~
p 0 s I
i 2 i P 05 n P a
P 0
5 I
- i 0 1
2 J-0 8 6 4
0 8
6 4
2 g
O -
2 1 1 1 1 0 0 0 0
]3. Dj . 3ce@tos N-
. : 1!1' ,I )
ll p 0 5
0 5
2 1
- l l
e t
e M t
- a
- B M
0 0 ;
l e
eL d
a o
I 0
4 O 0 - t M 1 Smnu
- om M - ~- -
bi r x aa _
~-
~
CM s t
M # a 0 e m05 3 hc I
m7
, - i n
r r, M t e
e
. t e
e C
g m I
/ -
g,-
a R R m i
S P i a
D I E N. - D e R / B e p 0 A E B - 0 p M i P 05
- P G L I 2 i P
M -
M -
0 5 ,
~ 1 0 1
2 M -
M - e M - - - - 0 0 '
6 4 2 0 8 6 4 2 0 2 1 1 1 1 0 0 0 0 J 3 ] ;t l Q " 3Ej&x" M
M l
e 5
5 2
e e
l 7 I l
6 H
e B
a -
0 d
a l
2 0 O e 0 I eo pL 5 I R
P C
R i
Pm du E N. em B
i
/ dx E
G B
L l
eM a e
_ W t s m57 m3 I
a I 5 1 h i
e c
n r r, e
t e t e
e e
i m
a p m a
D i D
e p 0 4
e 0 p i
P 5 I S I 1 i m
2 ' I R P u A
P e
5 2 I 1 5
1
- - - - - 0 0 8 6 4 2 0 8 6 4 2 0 2 1 1 1 1 1 0 0 0 0 i
j3 y5]"-]3 35E.los .
i m
u
- !! - ,I .l
I
% D g
SUMMARY
- LOAD PREDICTIONS GE/EPRI METHOD I
g SOLUTIONS T00 STIFF - PREDICTED LOADS T00 LOW COMPARED T00 EXPERIMENTAL DATA PARIS - INDEPENDENT OF STRESS STRAIN BEHAVIOR l LBB-NRC - MOST ACCURATE AND REASONABLE D
I I
I I
I I
CBallelle (N
I
)
I L
i I
DISPLACEMENTS IMPORTANCE l e NONLINEAR SPRING FINITE ELEMENT MODEL FOR g COMPLEX PIPING SYSTEMS E l
I e SIMPLIFIED INSTABILITY CRITERION L
I I
I I
I I
l OBallelle l
)l I
3 omO, OH'iam i afS l
l e
8 6
8 5
8 4
8 3
8 2
8 1 g 5 t t
e
- - - - - - 4 a B
l 4
a 8 B t n
CLI 2 4 i
8 i
- e L 1
- m R C i
r N. C B. B.
o .' ~
e p i xE aBB B s
r BB B 5
~
,~
EGPLL L i 3 s
- e m o ,
- - - h m 8 o -
- c i
8 O - - n
- , - i 3
i
- T -
N E
T M N
- E 0 N E C ' 5 M A i
- L P
8 6
i
.\ 2 E C
R S .\ L I o P D '
( . 2 S I
E N
,N D I O ,%
E L 8 i -
N N
- 4 5 I D * .
L R .,
1 O o D L R O
' . 1 L
8 i 2
. 5
'f .
l ,
8 - - - ,
,8 5 2 S G 3 g 1 1 2
B F
6 8
a x c @ n_ a hj M n. a .E y y H /
2 M
lll1 ,ll,l
LORD-LINE DISPLACEMENT, mm
-l 0 100 200 300 400 140 i g i -
SOE cP o O Experimental o
120 - /g o GE
\
j' o -
Paris -
500 H
'l -
0 LBB.NRC z 3
o 1
l M 100 -
J
, , g. ------ LBB.BCL I a
[ \g,'. o - - - LBB.BCL2 -
400
- 80 -
[
\! ,
J O
a LJ f, (.T -
300 O
H
_I 60 c.
o
' \% O_
[
I 0: %,
l O _j i _j .
200 m 40 f g g H o 100 20 - _
- )
0 O O 4 8 12 16 i
LORD-LINE DISPLRCEMENT, inches fleBaHelle Q-2/86-FB7 -
j
~
p-g c ==
N LORD-LINE DISPLRCEMENT, mm h 8 10 20 30 40 50 GB 14 i i ; i ; i _
0 0 0 13 -
o o 0 o 12 - , f ,[-
5'.- o r.. / ,g O ',
o- gg l (n 11 . ,
\ - d ) Z O. - ' s
\. '
Y 1B .l ,
& S - jf .
40 h 5
_J g _s' \. .
\, 3 o \ o a La
' 7 g .
30 d
N, .
]'
3 \ Q-6 O Experimental S S ?
T 5 -
GE _j
_J 20 c
[ 4 - .
Par.is s o LBB.NRC H H 3o -
LBB.BCL I -
2 - - - - - LBB.BCL2 1
0 O O .25 .5 .75 1 1.25 1.5 1.75 2 2.25 2.5 LORD-LINE DISPLRCEMENT, inches TIG WELDED 4.5" DIA. PIPE OBallelle N -
]
I
{
SUMMARY
- DISPLACEMENT GE/EPRI AND PARIS - UNDER PREDICT COMPARED TO EXPERIMENT LBB + NRC - ACCURATE FOR SMALL DISPLACEMENTS LBB-BCL METHODS MOST ACCURATE
-J-ANALYSIS METHODS - THEORY DETERIORATES AFTER CRACK GROWTH /
NONPROP0RTIONAL LOADING OCCURS
- USE ENGINEERING J-ANALYSIS TECHNIQUES AS LONG AS CONSERVATIVE RESULTS ARE PREDICTED
- CONTINUE TO SEARCH FOR ENGINEERING TECHNIQUES BASED ON MORE THEORETICALLY SOUND ELASTIC-PLASTIC FRACTURE l
PARAMETER + T*-INTEGRAL l
l I'
I OBallelle
-l
M M i] p
{ h REVIEW 0F SPECIFIC TECHNICAL TOPICS (6) e COMPLEX-CRACKED PIPE EV4LUATION
- SEVERE FLAW GEOMETRY FOUND IN SERVICE
- TO PREDICT STABILITY OF LONG SURFACE-CRACKED PIPE.
e SIGNIFICANCE OF RESULTS TO DATE
- LIMITED CRACK INSTABILITIES OCCURRED IN A106 GRADE B CARBON-STEEL PIPE AS WELL AS C(T) SPECIMENS
- A106 GRADE B PIPES FAILED BELOW NSC PREDICTED LOADS EVEN THOUGH THEY WERE SMALL DIAMETER
- J -R g CURVES FROM COMPLEX-CRACKED PIPE WERE LOWER THAN SIDEGR00VED C(T) CURVES
- J-ESTIMATION SCHEMES DO NOT ACCOUNT FOR RADIAL CRACK DRIVING FORCE.
AN EMPIRICAL CORRECTION WAS DEVELOPED TO MAKE REALISTIC PREDICTIONS.
- J-ESTIMATION SCHEMES PREDICT LOAD-DISPLACEMENT WELL UP TO MAXIMUM LOAD, BUT OVERESTIMATE LOADS PAST MAXIMUM LOAD.
e FUTURE PLANS
- IMPROVEMENTS TO J-ESTIMATION SCHEME PREDICTIONS UNDERWAY
- TOPICAL REPORT BEING WRITTEN INCH-DIAMETER COMPLIANT CC PIPE TESTS UNDERWAY.
C4Ballelle
I'
( >J
=
e I'
O I N'
6 4%
jk
~/ f,
/
' ' ; I
,l l / !
l 's l
\ \
- j;e!
._ R rk I.
m ,
G
- a. .
!I b
n - ss l m :x v
s$$
i! l
% c' .i 8
- 3 d
1
. E
,E
l ll M
)
p M
l l
e t
t e
M a
/j ,
j I lf
, B O
J
- 1 e
M t r
u a
,+*
r e
p
) m " _
M [/[j N
k
(
t t
e n
4I b ) ) ) ) e h c m #.W.
d, 3 0 )6 3 8 )9 b m
n m 4 076(43 1 i M / L o
o (1 ( ( (
1
(
yt r a o
0.
2 ii 5 0
000000 t d 00000 0 6,5.0,2,8 8 mr u e e "
le /b/
i p 52740,8 ot p 1 1 1 e c 0 /
M ,, A 22 l
go wf r
5e L
e 4 '
) ,
a
/ ,'
t ots o l f et e ..
T - c lats i 30 o M '
,% o i o it inP 0e f
r u
s cg m.
- ,'k .
N I 1 1
r e
g -
f e
c a
r u
d f
P e
o r
i23456 I M
U i 2 0
I n
n s,
3-p^sI 5
r e
n
- 0. i 0 1
i
__~
M 3 n I 3
k c
oaso W,
M o r aw a '
c ,. W- %
l la m' w
M f e
c a
h g
u o
e c
m.
r u h r
, f a
s T r r
. u s
M t e
u r
e O + <
t u
O
%' I M
M _
M M
I ll)i')j;Ai ,1j- .l ; <
l l !
j
,2
l l
e l
l e
Z# p$" CdJ_E 3
$ a B
4 0
5 1
0 0
1 0
5 5
2 C
0 n I o 6 i t
a i
t -
i n
I k
0 c ,
2 5
i a s r
C l e x e h m t e c m
l e S n e i t s S s T e ,
N 0 s l T
4 i
s n N E e i 5 M l a ,
E E n t 1 M
i S C a E R S t 43 06 C
L 3 R P 41 0 L S 03 e=
pt P 3
S I 0 i 0 y/
3 g-D e- Td I pt y/ D E Td N ,
1 E N
I L
/ I L
D 0 -
R 2 i
D O R L B.1 O r3 G L 0
S -
Bt , 5 t /
Ad a
l i
B. 4 r6 G
0 6 =
/
9t 1 /
Rd -
0 8 0 0 0 0 0 0 0 g 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 5 0 5 0 5 5 4 3 3 2 2 1 s
tD~ @8- cd 3_EC 3
y
(
I1l l l 4 ili j!!II , '
) 1I :1 1 1 3
l C
l l
l e
a B
Q 0
4 n
o ~
it k ia c s a
r 7
it n c e 3 ik e l )
l r 0 co o0 0 g
= r w5 '
3 e D c - 5 d r
i f h0
/ ot g c n u=
ot 2i o r f )
P h el C T( O
(
a t n
t n
e e K m C im) i) r A e
r 3 e11 0 R p 9 p2 ' 2 x 3 x eO C e0 kc g =
k =
c g O
a g g o r e T r
ct e c (t x
( x E e1 le2 U lp - p - D m1 3 m13 oI o1 C4 N C 4 O 01 I T
R T
O -
R O
0 0 0 0 B 0 0 0 0 4 3 2 1 g$ * - . sE
_ $gE oZ E$ .
l i
e i
t c
a B
4 C
.: - 6 3 .
_ 1 .
_ 1
_ 4
- : .:- 5 .
_ 3 1
l _ .
1R
- _ 4EB m _ _
S
_ u 3
_ U l
4N T
3N1E 1u4i
_ R ta _
E a 2 . P d D .: : : ...: :.: : . .
X d
/ .. :
AO A . ;. 4E a o
) - - : : . . . :
U
+ / L .
3 E 1P m u l
Y
(
5
/
/
E S
1P 4
I i
m 1 Y P A D a E M x
1 L L : : - : K O .: - ...: .-
3C .
C . :: : : . . : . . : -. . . :: .:- A B
- - R N - 3 1
C O -
I T _- 1 4XE 2 C _
L
/
) E P U S , . : :
u
+ T ,..: :: : : :: ... .~.. . : . . . . : :: .. :- .
.: : .. .- - : : : .: 2O C
(Y E N 3
_ 1 1
/ 4
_ D A _
O_
i _
1 u ,
3 u 1 u
i 1
x 4 A
u . - - - -
2 6 4 2 o.
- o. a. o 1 i
(
! .I j , l !l
CRACK EXTENSION, mm e 2e 40 Se Be les 12e 14e N
4eese g i ; i i g i
. . . 304 SS (d/t = 0.63) 36 M -
Sees
- INCONEL 600(d/t = 0.61) -
3200s - .- AIO6 Gr. B (d/t = 0.64) sees 2000s -
% 24000 - < _ 4,,, ,
N / $
S 2000s - p 2
/ $
16000 -
/
b /
1200s - / - 2000 e"
~
/
r' /
- less Aeee 7 - - - -
./
. . . e
,P , .
2 3 4 5 6 e 1 CRACK EXTENSION, Inches _
C4Battelle
\, )
l l
e e
m N b x. "
l l
e 0 0 0 0
0 0
0 0
0 0 a 0 0 0 0 0 0 0 e B C-0 4 3 2 1 6 5 5
- - - 4
- - - e 1 p
_ 0 i 2I 6 P 1
4 0 d 3 - de t ev 0 / ko
- d co t arn
/ rGe - 4 d C- m 0
0 1
I xdo et o eI lSp S
p%
m s .
o0T e .
C2C h
. c m 0 I
n .
m 0 I
. . 3 N N O O I I S S N N E 0 i E
T 6 T X X E
K C
R 2
E K
C R
g R R C 0 C 4
,1 0 i 2
- - - ' .E5y 0 0
0 0 0 0 0 0 0 0 0 0 0 '
0 0 0 0 0 4 6 8 0 2 1 4 3 2 Nc N D. - c~ 'm i i I 1 ! ;' :a Il j!i < 'II )
10 g 09 1 g a DTNSRDC 08
\
^ BCL
~\ A BCL
\
07 - \g
\
06 \
g o
-y ,M p
05 -
\
g ^
g 04 -
\ II N
j 03 -
IN ,
i \
! 02 -
s's 01 -
1 i
g l l l l l l l l l
' 0 01 02 03 04 05 06 0.7 08 09 l0 d/t I
i l
1 I
l ff#Battelle ,
! \s 4
- I e
E J
NM '0801 03Ildd8 18101
&m -
= = .
=
=
v
~ ., Q ,5 i ji i i s #4 i 1
- %9 m
/
5 g_
- l. l . ,
/. / l
/ j
/ / . ns 'sa E ~ g
-g _ / O -
w l l.
5 / /-
o t
l 5"
e
- / / ,
~W d ! l 6 5 l e
= -
/ /
o
- 9 a
M w 7 / O w
l, -
s" /. se i sa 3 3 /o of i wofazm sig - -43 l g _/o f l (
N g o il!. , il g ii
-l l
. I .mh' 'M I i _o i ,
N N N 89l 'QU07 03IlddB 19101 2
ii l p
l C
l _
t t
e Z# d$g odEQC -
0 a B
4 _
0 8
0 6
0 4
0 2
3 C
M 0 7 i h c '
a o
r . -
t n p 5 e p 0 ma 2 6 i ir e A s p O e x T o h m E C m c n
T 0
5 i
o - 2 i N ,
E T M N E E C M A O E L 0 i C 4 5 A P
S o 1 L I P D n S o I E D N 0 i 0
I 3 E L N
- I 1
D L R o -
O D L 0 0 R 2
o O o L o
o 0 5 o
0 i 1
0 o
0 0
0 0 - J o- l 5
0 0 0 0 0 0 0 0 0 0 0 0 0 5 0 5 2 1 1
.o- gCO ala nC OW q
l ll lill l ,'
ELASTIC-PLASTIC FINITE ELEMENT ANALYSIS EFFORTS i
LABORATORY SPECIMEN ANALYSE (ALL AT 550 F) e 1-INCH THICK STAINLESS STEEL 1T, 3T, AND 10T C(T) SPECIMENS 4
e 0.35-INCH THICK STAINLESS STEEL 1/2 T AND 3T C(T)
]
SPECIMENS WITH CRACK IN TIG WELD
~
e 1-INCH THICK STAINLESS STEEL 1T AND 10T C(T) SPECIMENS WITH CRACK IN SAW e
- A106 GRADE B CARBON STEEL FWFN(T) SPECIMEN 1
- ROUND-ROBINS CONDUCTED TO ENSURE FINITE ELEMENT ANALYSIS ACCURACY 1
i OBallelle i
e - e-- -
M M M M M M M M M M M M M Mp M
]M PIPE ANALYSES e 16-INCH DIAMETER STAINLESS STEEL PIPE WITH TWC (AT RT) e 4-INCH DIAMETER STAINLESS STEEL PIPE WITH 1WC IN TIG WELD (AT 550 F) e 16-INCH DIAMETER STAINLESS STEEL PIPE WITH TWC IN SAW (AT 550 F)
- 16-INCH DIAMETER A106 GRADE B CARBON STEEL PIPE WITH INTERNAL SURFACE CRACK (AT 550 F)
,
- ROUND-ROBINS CONDUCTED TO ENSURE FINITE ELEMENT ANALYSIS
' ACCURACY i
i
( - -
OBallelley
l
( .
SIGNIFICANCE OF FINITE ELEMENT ANALYSES e TO ASSESS THE ACCURACY OF ESTIMATION SCHEMES e ASSESS THE APPLICABILIT( OF SMALL-SPECIMEN J-R CURVES l TO LARGE CRACK GROWTH ANALYSIS e ASSESS GEOMETRY DEPENDENCE OF J-R CURVES e ASSESS EFFECT OF WELD SIZE RELATIVE TO SPECIMEN SIZE ON FRACTURE MECHANICS PARAMETERS OBaHelle
Nota: 1 inch = 25.4 etc.
5"dia m.
'2 (
a
, 1 f?.;V* ^ry_
" " i a.s" =
n rm t I g
- s* -
c'
- 20" 25, }
(a) ,
l l l 1
l l
l f%
[ T l )
i
\
- N ~
s ;*'
1 I , t; . ,
'l Total I Cracp Crack Growth 5 'T)
( C4BaHelle M-12/84-F13 I
l l
e t
t e
a B
EE~oq 4 0
8 1
0 6
1 0
4 1
0 2
1 0
0 1
0 8
O 6
O 4
0 2 O 9
C .
0 - - - - - - - -
4 l 2
t - .
0 8 nnn 2 l eie 2 mam eti csc 0 a l 4 p e
0 l p0 s 2 7 s3 i k 0 dec pi 8 l d yh 1
eTt i
0 6 l T) l h p0 m m6 1 c n
p1 m a
m i d4
,0 ,
se t l t uv5 4 5 n so2 n1 a
e e ro(
er m0 m v gh e2 l e - c c1 c nen a 4l p a odi i i -
lp ss1 is0 I s nn D1 0 i
eol D t ne x e eet 0 I 3 h s 8 kt c s
- ars roe
- 0 I Cfl 6 2 0 I 4
I 0 I 2
- - - - - - O 7 6 5 4 3 2 1 O
55 " oq k_
i i! , - ', !' , i!ii1 lllI:j!: jii1i!!;l!.I !
Load-t.tne Displacement, mm O IO 40 60 40 4 i i i e a i -
- 15 0 l
30 -
~
os J'ox-x"%"N o
o xg n - d o ,o x f
$ of
- / o o
kx - ICO 3
g C */ o
- 4 io
- 20 -
- /
x o c
\
- o J
i o l
_s o x - :
D 7
o o
o I -
50 10 b i;
x Finite element l i
o Experiment i
~
l i I I I O O.O 1.0 2.0 3.0 Load-Line Displacement, inches Comparison of FEM predicted to experimental load versus .
load-line displacement for nonside-grooved Type 304 stainless steel 3T compact tension specimen tested at 550 F (288 C) (Specimen A45-50).
M-12/84-F9 l
1 l
l
1 Crack Extension, rnm O 10 20 30 40 '
i i F8 3T-CT 40 . O
/,
o l 6
a ,o o o o
30 -
o 0 o/*o o/
g u E e o o -
4 g 2
I 20 -
o 0 C
/ '.
5 l
9 o
/ Ng" -x - x g *
- o
- /
o a
/x/
2 10 -
x J ( Near Field) o J (For Field) o J, (Experimental Octo)
- g. I I I I I I I o 0.0 0.2 0. 4 0.6 0.8 1.0 l.2 1.4 1.6 Crack Extension, inches J-resistance curve for 3T nonside-grooved Type 304 gg stainless steel compact tension specimen. wBaHelle l M-12/84-F10 g g
}
DISPLRCEMENT, mm 0 25 50 75 100 125 150 175 200 225 100 ; ; ; i ; ; ; ; ;
90 - Experiment -
400 Finite Element Results by:
80 - X x X Battelle -
350
% gyyggo X X X Participant 2 U
70 - o g a Participant 3 3 -
300 o a o Participant 4 3
60 - o Participant 5 N O X -
250 $
X 3 50 0 0 X -
200 C a
4 O E O o g
g o
40 g J ,X d<a o 150 30 -
X j' o a 20 -L E 10 L 50
)
o 0' O O 1 2 3 4 5 6 7 8 9 DISPLRCEMENT, inches Comparison of the finite element analysis results for the CT specimen (Problem A) with experimental data.
OBallelle k
l t
Aa, mm 0 25 50 75 100 125 150 175 200 140 I I I I I I A 24 l 1 JD t
J g 120 - Finite Element Results by. % ,
X Bottelle O f'
20 0
1 X Porticipont 2 I 100 - a Porticipant 3 0 ,-
X
! cu o Participont 4 M ,-,' X
.E o Participant 5 0 ,-,' -
16
\ O '
/
- E / X 80 -
e
' , -,y i
- O
\
l [
o O W
,X 12 Z
E O '
,-'X - .
m 60 -
o y ',- A e,
cg
~
M , - X OD a
. O /A a O -
8 h 40 - O y'/ o s D O s' O
- O O
20 o -
4 a
j t O
'O l 3 0 O O 1 2 3 4 5 6 7 8 Aa, inches Comparison of the .1 (far-field) versus crack growth curves computed by finite element analysis with the J and J resistance curves.
D M C$Battelle
(: a g .
CRRCK EXTENSION, mm 0 20 40 60 80 100 120 140 120000 , , , , , , , 3 z 20000 l
A A A 100000 -
A A A
A A A A -
15000 l 80000 - ,%
N AA
.h A N
! N g E l
f 60000 - d N i
10000 .se .
C A j
- A A *
- .A m 40000 -
e
.* o o j
j f /m'o D^ u -
5000 20000 o FE #M IT O G 3T O E 1 . , , pet . A A ,
- O O
- O 1 2 3 4 5 6 i
j CRRCK EXTENSION, inches 1
Comparison of J-Resistance curves by FEM with the Jg-Resistance curves i for the nonside-grooved IT, 3T and 10T,1-inch thick (25.4 mm) compact tension specimens of Type 304 stainless steel tested at 550 F (288 C).
- C4Battelle
- \ -
}
l l
Z -
- 540 oches (13.7 m) 16 inen(406.4 mm) diame ter 1.03 6nch (26.2mm) -
L Type 30 4
~64 inches (1.62m) Wire rope stainiass steel ,,
,, O ,,
s .0 is
, stop block Saddle d
/ Y,' 8 d ' C p Pret ensioned were rope
"-=
of Swivel base crack 48 inch (1.28 m) diometer by 0.72-icch (18.3*nml tN.ek p,pe l i
t D =
D= 16" (406.4 mm) t= 1.03" (26.16 mm) 2ao/7rD = 0.3675"
-* t -
/'77'N \
! 200 Crack The apparatus and cracked pipe geometry for 16-inch (405.4 mm) outside diameter pipe bending experiment.
l l
e t
t e
n a
f a f
a $i' B s
C t
n e
m l
e e
. n i
o
_ s
' n e -
t
.:' x e
k c
I, a i
r -
f ( \ c -
l a -
.' u t
r i -
I, I v _
f[ k \ h -
t i
w i'
h s
I I e
\
s m fr t
n e
P I
I
/
yN N\(N \
x xl I
l e
m e
e x
1
\
.i D
M
, 1
/
/ CC
/r i\
l\
.(
\
xl t i
i n
.4 I L
y!. -
F il- -
.44
- ' i i' l! i
- l
i DISPLRCEMENT, mm 0 200 400 600 000 1000 60
) ; , , , ,
i -
250 i
l 50 - On O i 9 a O X A D 0 200 X -
0 X X o a 40 -
i o
~
Xo o X
/ -
150 3 30 -
g 4
x A cc o
J E a o Experiment 100 J 20 -8 ,
Finite Element Results by:
K X Battelle 10 -a Participant I -
50 e o Participant 2 j ^ Participant 3 0 !! ' ' '
0
- O 5 la 15 20 25 30 35 40 l DISPLACEMENT, inches i
Comparison of the finite element analysis results for Problem B with experimental data.
- gy,BaHelle w
1 L
a - )
M p
\
ENr 'm j l
l e
0 0
2 6
1 2
1 8 4 O t t
e M 5 1
i - - a - _
6 a
e B
4 a h c
m e b
- y s i s
, C s t l s 5
n u 23 ' 5 y s I M 2 1
o ia R t
e t t n n n t l
a n
a mt n l e pap ap a i i i t
e n eiitc c i
A t l
c e M o X s e m t t
t r r r a a a a i
t m
e
- le l J E BPPP e 0
0 i e '
4 e M 1 it n XooA t
i n
iF i f
o y b
X s m e d e
h m
g a A
5 7
i
^
3 i c
n a
t u
p m
oe cme A hh t c ws M o rn go i
kt 0
5 i X '
2 ca am M o ri ct s
se
^ u -
sJ r
M 8 ea v
X y 5 i a ' 1 Jb 2 ed M a h n Ta a
o M
0 - - - -
f O M 0 2
1 0
0 1
0 8
0 6
0 4
0 2
0 "c5fI5 3 ,*
Q l
l l
e a =
,- - o d
l l
e d
e a
a ssE .8m y B
l e
w
.b g d g m I Y kl d
c r
a le f
= e g
r 8:'I ,l,: ll;li .i e
m b
u ll;l s
_l,: l .i In r?
Is T 0
,;! . , 1 Is r .
P? W :.
_ . on Iz f e r, W i II si m
r h c
. Wy_ t e ap W >
> ps Wu _
l n ao f _ > ri gs
_ Y-y_ - en t e
_ _ / i s nt a.
_ i l l ,
t 5
rc
,5 ua 7 _ op 3 _ 2 t m 2
'Il l!;! !-
6 no oc O
l c
l
=
o d e ne 1il .l, l ' l l ! o at k s c h -
a ss Ii r es N s C me t n ni ea l
mt es 1lP - / 4 l
e4 0
e3 t
ie np i y t Ft n
io p
e g
< - o G
h li ,1 )1 > ,' ;;
M M M M M M M M M M M M M M M l
CRACK EXTENSION, mm e 2e 38 40 Se se 30000 i I I I I I sees O CONTOUR J ON l's M VCE J BETWEEN Ps AND Ps l
+ Jo
.j o IAT$ ON P3 -
4ees (
e b T ON Ps 2 y asses -
p g
/ ^
h
! # i d -
2ees l i'-
Z
/ j z
~
~
1988 0 5 0 i
e '
e e .'5 1 1.5 2 2.5 CRRCK EXTENSION, inches Calculated integri.1 parameter R curves for 10T submerged-arc welded type 304 stainless-steel compact tension specimen at 550 F (288 C),
fl&BaHelle k
l T
h
) -
l l
e t
t e -
a B
(\ Q C
D R
S l N
~) / \ T .
f k D2 er h e tb m
f u oN rt en t e rm ai
} I ur qe a
/ [ p ex nE o
e f p oi p
g nl i e l e et ds 4 o
_ /
P r
mss 3 ,
/ I ( \I t e i
n nl 3 en mi 9 ]l 1
ea l t m es
[0 m } I d9 / l \ ed t e le4 id W( }
/ [ x \
I nl ie Fw t
s n
e 4
1
}
1 I f-f I f; I
7-[\Ige mI I
t n
o r
- F m
le e 6# k c
a 3
- J r F E C 5 C 8 V /
6 A
S -
L -
_ l;) !!j1 i il;( , !11lilii11Ii(;!
li
l m
) ]
i-E m 2 I l l
e d
e d
Gde* W T t
e G dT A t m
I l TeC W
Il Te W* S C a t
t
, B m O
t n n n m e m
e m
M E
e m
o i
r i
r F, i r
T e I w
e p M p e p M .
x E x i
x E 3 s J s m E, i
F, i
E, E
C E,
i F,
j 0
h e
n g
J J Ji V J J u o
e O A a 5 O t m
n i
o
'
- O t a
m i t
i n
i n
o m I 2 e z
m s
0H/
i s
D n e
i m
d l
c e e p
W W s M G A o I S t T
f e z
M " H1 T
i s
- 5.
- 1 d
l 1
e O w M h f o
t c
e M P
~
p ,O f
f E
t b 9 t T
M 5 9
6 3
7 2
M 0 0
h^ O F
0 0 O 6 8
0, p /
0 5 6 1 -
A
= s. ' 6 S
( N I !
\
{
i i
1 i :
! THE SECOND BATTELLE/NRC FRACTURE MECHANICS ANALYSIS ROUND-ROBIN l
i I
I
}
ff#Battelle 3 _.
ll f
g g l l
e l
e g l a
B t
g C g
g
" h y n
g = )
b i e
m 2 :
a ( c
/
8 o
w u e p
a s g ;
=
)
N F
W F
) (
)
r m m )m d m )m )m mm 4 m h e
c 3 9 m5 9 t 0.
8 8850(84 2. 8 o n
M 1
( ( 2( d p
e h h( h c c ch c ni n h
/ y c a
in i nci f
M 1
03lO nf 5
- 7. 5. i O3
= = =
Q0 2 G o
/
, h t
d i
w-WBhd o c l u
M )
l f
u a
( e h
, t M
M M
n n
m d
mf t G
o y
r e
m o
e illlll \
Internal Surf ace Crack P/2 P/2 Schedule 100 Carbon S,s.i e...
, y I a ' e i a bly '
Load ceil h
!! !;\ Rom N Strorgn:t ll 6J ll LJ
( )
Min -
Y D = 15.85 in.(402.59 mm) t= 1.04 irt (26.416 mm)
{
wD = 0.532 l
df, = 0.662 l I, l
I X
%d v
t
".-- D/2 Apparatus and cracked pipe geometry for the surface cracked pipe bending experiment.
CeBanelle t >
J
M M M M M M M M M M M M M M Mp M
[]
{ h STATUS OF DP3 11 AFTER 3 YEARS (1) LIMITATIONS ON NET-SECTION-COLLAPSE ANALYSIS
- NUREG-0313; NUREG-1061. VOLUME 1.
(2) SCREENING CRITERIA DEVELOPED
- NUREG-1061, VOLUME 3
- GDC Il ANALYSIS
- ASME SECTION XI AUSTENITIC AND CARBON STEEL PIPE CRITERIA
- W. GERMAN AND JAPANESE LBB CRITERIA.
(3) PIPE FRACTURE DATA BASE
- THROUGH-WALL CRACKED PIPE DATA BASE SUFFICIENT
- WELD OVERLAYS EVALUATED
- SURFACE CRACK DATA SHOW PIPE R/t IS IMPORTANT. ADDITIONAL. DATA NEED FOR DIFFERENT CRACK SIZES
- THERMAL-AGED PIPE TESTS WILL NOT BE CONDUCTED
- CARBON STEEL WELD DATA NEEDED
- BIMETAL WELD DATA NEEDED
- DATA ON TORSIONAL LOAD INTERACTIONS NEEDED.
C4BaHelle
3 STATUS OF DP 1I AFTER 3 YEARS (CONTINUED)
(fi) MATERIAL PROPERTY DATA BASE (COORDINATED WITH MEA)
- ENCOUNTERED DYNAMIC STRAIN-AGING
- EFFECT OF WELD SIZE TO SPECIMEN SIZE ON J-R CURVE
- VARI ABIL T TY OF AUSTENITIC SAW TOUGHNESS (5) ELASTIC-PLASTIC FRACTURE MECHANICS ANALYSES TO PREDICT LOADS
- G.E./EPRI TWC ANALYSIS CONSERVATIVE
- PARIS METHOD NOT GENERAL EN0 UGH
- NRC.LBB METHOD MOST ACCURATE FOR IWC
- BCD SURFACE CRACK ANALYSES DEVELOPED AND EVALUATED j
- COMBINED PRESSURE AND BENDING LOAD ANALYSES EVALUATED I
- ASME SECTION XI ANALYSIS VERIFIED
- METHODS TO EXTRAPOLATE J-R CURVES FOR LARGE CRACK GROWTH ASSESSED (6)
ELASTIC-PLASTIC FRACTURE MECHANICS ANALYSES TO PREDICT LOAD-DISPLACE RELATIONSHIP i
- THROUGH-WALL CRACK PIPE ANALYSIS SUFFICIENT FOR ENGINEERING APPLICAT
' - INSTABILITY OF SURFACE CRACK GROWTH ANALYSIS DEVELOPED
- COMPLEX CRACKED PIPE ANALYSES NEEDS EMPIRICAL CORRECTION
- WELDED PIPE METHODOLOGY DEVELOPED.
OBallelle e,,, .
M N M -
p
{ h ASPECTS NOT COVERED e PROCUREMENT OF PIPE REMOVED FROM SERVICE e NOTCH ACUITY EFFECTS e HAZ AND FUSION LINE CRACK GROWTH IN SS WELDS e GE0 METRY EFFECTS ON J-R CURVES e ANISOTROPY EFFECTS ON TOUGHNESS e CHARPY VERSUS J CORRELATIONS l IC e EVALUATION OF CRACKED PIPE UNDER COMBINED PRESSURE AND BENDING I
e RESULTS OF CIRCUMFERENTIALLY CRACKED PIPE UNDER PRESSURE (AXIAL MEMBRANE)
LOADING e DEVELOPMENT OF A FINITE LENGTH SURFACE-CRACKED PIPE J-ESTIMATION SCHEME e EVERY BALANCE APPROACH TO PREDICT LENGTH OF CRACK JUMP e APPROACH TO PREDICT CRACK STABILITY UNDER DYNAMIC LOADING (SEISMIC, PRESSURE RELIEF VALVE BLOWDOWN, ETC.)
e "NRCPIPE", A J-ESTIMATION SCHEME COMPUTER CODE.
fl*Ballelle 4
3 DP Il DELIVERABLES e MONTHLY REPORTS e SEMIANNUAL REP.0RTS e FINAL REPORT e TOPICAL REPORTS (10) e PIPE FRACTURE DATA RECORD BOOKS e MATERIAL FROPERTY DATA RECORD BOOKS
- COORDINATE WITH MEA DATA BASE e J-ESTIMATION SCHEME COMPUTER CODE e TECHNICAL NOTES e PIRG MEETINGS, WRSRIM MEETING AND REPORT, BRANCH CONTRACTOR REPORTS e VIDE 0 TAPE e PHOTOGRAPHS, SLIDES, ASSISTANCE IN NRC STAFF REPORTS e TECHNOLOGY TRANSFER TO NRC-NRR.
OBaHelle N .)
a - .
mM M M M M M M M M M M M M Mp M TOPICAL REPORTS (1) NRC.LBB ANALYSIS METHOD NUREG/CR-4572 (MAY 1986)
(2) ASSESSMENT OF TWC PIPE NUREG/CR-4574 (JULY 1986)
(3) PREDICTIONS OF J-R CURVES WITH LARGE CRACK GROWTH NUREG/CR-4575 FROM SMALL SPECIMEN DATA (JULY 1986)
(4) RESULTS OF FINITE ELEMENT ROUND-ROBIN ON A C(T) NUREG/CR-4573 AND TWC PlPE (JULY 1986)
(5) ASSESSMENT OF COMPLEX-CRACKED PIPE NUREG/CR-4687 (JULY 1986)
(S) FRACTURE MECHANICS ASSESSMENT OF STAINLESS STEEL (OCTOBER 1986)
TIG WELDS (7) DEVELOPMENT OF BCL.LBB ANALYSIS AND VERIFICATION (NOVEMBER 1986)
FOR COMBINED PRESSURE AND BENDING (8) ASSESSMENT OF SUBMERGED ARC WELDED PIPE (DECEMBER 1986)
(9) FRACTURE EVALUATIONS OF WELD OVERLAY REPAIRED PIPE (JANUARY 1987)
(10) ASSESSMENT OF FINITE LENGTH SURFACE CRACKED PIPE (FEBRUARY 1987)
C4Battelle
o o o r R A : ~~ .RE VE -AN:: S E V A _ _ A -~ :: : \
- = _WR A__:YS
- R. A. HAYS, V . G. V ASS
- : L A ROS, J. 3
. GUJAS DAVID TAYLOR NAVAL SHIP RESEARCH i
AND DEVELOPMENT CENTER BETHESDA, MD i
i 4
e e e E B J E :~~ :: V ES:
- PIPING MATERIALS FRACTURE CHARACTERIZATION a) MECHANICAL PROPERTIES b) J-INTEGRAL FRACTURE TOUGHNESS c) TEARING INSTABILITY BEHAVIOR
.., ;rrsw- V - .. .._
m ~~
o o o
- 9J E :~~ :: V ES (con ):
- SUPPORT ASME CODE FLAW EVALUATION J
DEVELOPMENT
- DEVELOP FRACTURE MECHANICS ANALYSES i FOR' PIPE GEOMETRIES i
- ASSESS GEOMETRY EFFECTS ON FRACTURE l
i j ANALYSES i ,
i .
i i
~
_ _ _ _ _ l_I 11_ . . _ _.
o o o v A ~~ E R :: A _ S :: \ V E S-~ :: ] A ~~ E E
- A106 STEEL AND GTA WELD
- TYPE 304 STAINLESS STEEL AND GTA WELD
- CF8A STAINLESS STEEL AND GTA WELD
- A516 Gr. 70
- A533B STEEL
~~~~' ~~~'
l O O O c
i 3:: 3:: \] RESEAR :
I
- 8 in. A106 STEEL BASE METAL
- 8 i n. WELDED A106 STEEL i *4 i n. TYPE 304 STAINLESS STEEL i
- 4 i n. WELDED TYPE 304 STAINLESS STEEL
.i l
l _ _ _ _ _ _ _ _ . _ _ . . . . . _
O Lo O
~
1 W LLJ n i H LL] O Z
O < l ,
H LLJ n LLJ l
1 (D U l Z !
s 03 O l
. 1'
O O O
~
i l
l l RESEARCH OBJECTIVE i
i I
f INVESTIGATE J-INTEGRAL FRACTURE TOUGHNESS OF l ASTM A106 STEEL IN BOTH PIPE AND COMPACT i
- SPECIMEN GEOMETRIES l
I l
l
O O O 9
h 7 o COMPACT ! [, ~ g SPECIMEN ./ / \
\ l \
l _ CHARPY "V" k
' SPECIMEN \ !
/ \s /
x__s /
m
/ \
CIRCUMFERENTIAL CRACK PIPE SECTION i
i
,i O ..
t Chemical Composition (weight percent)
C Mn P S Si Fe 0.23 0.81 0.0062 0.013 0.164 REM i
Mechanical Properties Spec Yield Strength Ultimate Tensile Elongation Reduction in Area
- kai (MPa) Strength kai (MPa) % in 2-inches % (L = 4D) 1 41.5 (286) 71.2 (491) 40 65 2 41.7 (287) 72.2 (498) 37 65 3 41.5 (286) 71.1 (490) 37 64 4 _41.6 (287) 70.7 (487) 33 ~64 5 42.0 (289) 70.7 (487) 39 64 O 6 7
41 5 (2=6) 38.4 (264) 72.7 <>oi) 71.3 (491) 41 38 65 64 8 38.0 (262) 70.4 (485) 46 64 9 39.6 (273) 72.0 (496) 42 64 10 i2.9 (296) 72.9 (503) 34 65 11 43.0 (296) 73.0 (403) 37 65 12 42.1 (290) 72.3 (498) 33 65 Average 41.1 (287) 71.7 (495) 38 64.5 O
TEMP ( C)
- 120 - 100 60 20 0 20 40 60 80 100 130 i , , , , , , , , , , i 120 -
- 160 5
110 -
g U -
1@
100 -
O 90 - -
120 U
80 -
A FILLED POINTS ~
ARE AVERAGES g y 70 -
D -
80 2 60 -
z U U
@ 50 -
a -
60 40 -
30 -
a e
O 20 -
U 10 -
E E -
0 0 -
I I I I I I I i l
- 200 - 150 - 100 - 50 0 50 100 150 200 TEMP ( F)
O -
e
l TEMP ( C)
- 120 - 100 60 20 0 20 40 60 80 100 130 i i i , , , , , , , , ,
120 - -
160 5
110 -
g 0 -
1@
100 -
O 90 - -
120 U
80 -
A FILLED POINTS -
- 9 70 - ARE AVERAGES g g z
-@ g 2 U U y 50 -
g -
60 40 -
30 -
a _
q O
20 -
~
U 10 -
B B _o o
I I I I I I I I I
l - 200 - 150 - 100 - 50 0 50 100 150 200 TEMP ( F)
- 9 _
e
\ ,
l l
26! 20 N
- R'mi Rm ,
a
/ \
t i
l SIMPLE CRACK ; COMPLEX CRACK i
I 2 e = CIRCUMFERENTIAL CRACK LENGTHj 2 e = CIRCUMFERENTIAL CRACK LENGTH j RM = MEAN RADIUS I RM = MEAN RADIUS t = PIPE WALL THICKNESS '
t = PIPE WALL THICKNESS a = RADIAL CRACK LENGTH l
j l
i e o e 1
l
O l
l N
I O T
I S
I P
I l
11 l i i4; l I
- S T
l
_ G -
NR N N EA I R
O MB ITET E
T R P T S C RR \ - A E L N D E
L UO L
P L E N
FSP O M EAP - T R E A DEU E G CNE MS R A ONI LT P CH I
- A
.;,gT D S E C
@ D LA I
l : -
E 1
B G E FM ,
I A C H N NIP EI C
E H
D A
k l!
A PL F ODO O M HAT L L T TON S ULE E
T OEM
@ (',i=
f MIPE K PC A
- C=L 2 2
d E Ad PS LS R I L
I G C D VN = L S EI LR 1
d A LP T _
ES O B T -
=
3 d
i e
j' E
8 !l l,
- 1 h '
P I -
P - l 3 _
I 6
- N -
I O
T I
S I
I P
O= -
=
w- - +
0 SINGLE SPECIMEN METHOD MULTIPLE UNLOADING
/////////m a l 9
DISPLACEMENT J
J/2af O #O 1 y
/ Aa
- =
DTNSRDC
- D.C. POTENTIAL DROP CRACK LENGTH ESTIMATION TECHNIQUE LOAD u
' V V 3
l DIFF AMP l
l i
}r POTENTIAL DROP 4
CRACK INITIATION CRACK EXTENSION
- L JL
]
I
/
COD a/W i
i e - - - - - - - - _ _ - - _ . - _ _ .
.e -
8 --
FORMULATIONS FOR J AND Tapptimo IN A PIPE ZAHOOR AND COWORKERS TADA AND COWORKERS Jz = f (ACTUAL LOAD AND JT = f (ASSUMED FLOW DISPLACEMENTS) STRESS AND MEASURED BEND ANGLE Tapptigo = f (Jz, Ku AND TAPPLIED = f (JT , Ky )
1 MEASURED HARDENING OF PIPE MATERIAL)
Ku = TOTAL SYSTEM COMPLIANCE,
= TEST MACHING STIFFNESS +
FIXTURE STIFFNESS +
SPRING STIFFNESS O
9 ._
CRACK EXTENSION, (mm)
-2.50.0 2.5 5.0 7.5 10.0 12.5 15.0 17.5 20.0 22.5 25.0 I I 1 1 I I I I I I i 2500 14000 - ,
SO I 12000 -
y Xo 2000 a Zg -
- E 10000 -
y o m
" n 0 -
1500 m, E
.d ".5
- 8000 --
o O 3 n 9 0 0 6000 -
0O J-INTEGRAL CRACK LENGTH - 1000 Z30 EXPRESSION MEASUREMENT TECHNIQUE A J-TADA ELASTIC COMPLIANCE V J-ZAHOOR ELASTIC COMPLIANCE O J-ZAHOOR D.C. POTENTIAL DROP - g 2000 -
FILLED POINTS ARE FROM FINAL OPTICAL CRACK LENGTH MEASUREMENTS i i i 1 i i i I 0 i
- 0.1 0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0 CRACK EXTENSION, (in.)
um u-JrR CURVES FROM P INCH DIAMETER A106 STEEL PIPE USING UNLOADING COMPLIANCE p i i i i i i i i i iae a -
14000 - ,a A U A V
. a o U^ -
12000 -
, pa 8 10000
^ o -
oQ v 3 o
~ in 8000 -
,a O h -
m v ^^#
, 6000 - fV9 a , g O -
o e:PE,i 9 ,a g O PIPE 9 _
g _
V PIPE 7 Q
9 7 a PIPE 3 E PIPE 15 m ~
2000 -
A PIPE 12 m '
0
-0.1 0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0 CRACK EXTENSION, INCHES O O O o -- i
CRACK EXTE'NSION, (mm)
-2.5 0.0 2.5 5.0. . 7.5 - i10.0 ;12.5. 115.0 i l7.5 20.0 22.5 i25.0 I I I I I I I I I QI 14000 _
l2500 D9 W
12000 - l O 9 O O
% ^
O 2000 l
0 e o 6 m
10000 -
a "t
og 9 3 AO 2"J O A
- 1500 in.-Ib 8000 -
O g
in.2 sf O m2 6000 -
@O V PIPE 7 O l PIPE 8 O Q ' pips ,o 4000 -
O PIPE 11 9 PIPE 13 A PIPE 15 -
500 2000 -
> TEST TEMPERATURE 125'F(52*C) 0 1 I I I I I I I I
- 0.1 0.0 0.1 0.2 - i 0.3 0.4 i 0.5 0.6 0.7 :0.8 ' O.9 1.0
! CRACK EXTENSION, (in.)
O _ _ _ - _._.
O O
..7
.. . s CRACK EXTENSION mm 7.5 10.0 12.5 15.0 17.5 20.0 0.0 2.5 5.0
$g l l l l l l I
l O -
2000
, U 6 301 - 1/2 T COMPACT SPEC O o O V 303 - 1/2 T COMPACT SPEC _
10000 -
I O o i
O l
Ej O -
1500 1 8000 - U! g FILLED POINTS ARE i
' 2O/ O POST TEST OPTICAL
$I O O CRACK LENGTH e u MEASUREMENTS ] "E
.y ,d 6000 - 3/ -
1000
- m g3 O
1 O PIPE TEST 7 m -
O O PIPE TEST 11 l - 500 O PIPE TEST 12 O
2000 -
i i 0 0 C4 0.0 0.1 I
0.2 I
0.3 i i 0.4 i
0.5 0.6 0.7 0.8
- 0.1 CRACK EXTENSION -(in.)
- t
. ~
& }
O O
>' $ v + % f
CRACK EXTENSION mm 10.0 12.5 15.0 17.5 20.0 0.0 2.5 5.0 7.5 3g O t 2000 l o O o O 10000 - f O O
y[ O 1500 a 0 M - O EI O o FILLED POINTS ARE o N. 2' A O POST TEST OPTICAL n
- 3 CRACK LENGTH 3 E d .5 6000 - # MEASUREMENTS 1000 O ,
7 p Q PIPE TEST 7 O PIPE TEST 11 O
O PIPE TEST 12 _
2000 - 6100 1T PLAN COMPACT SPEC Q 103 - 1T PLAN COMPACT SPEC i i I i i 1 0 i 0.3 0.4 0.5 0.6 0.7 0.8
- 0.1 0.0 0.1 0.2 CRACK EXTENSION (in.)
l
-m -m i n l
t 1
l l
CRACK EXTENSION, (mm)
-2.50.0 2.5 5.0 7.5 10.0 12.5 15.0 17.5 i 200A 22.5 i o 25.0 r
i i i i i i i f -
2500 D
14000 - ./ o ASTM A106 STEEL q O
^
O O O 12000 -
E 2T COMPACT #2a{ O pg A 2T COMPACT #1al NO SIDE GROOVES q _
/ O o o 6
ou e 10000 -
/ 80 e a o 00 - 1500 2 "E
)s-g g.a O
..Ax Ej QS /
6000 -
s of O W G V PlPE 7
_ 4000 a Pipe a
- f po Q PIPE 10 g _
0 PIPE 11 9 PIPE 13 -
g a PIPE 15 i i i i .
i i i 0 i i l
0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0
- 0.1 0.0 0.1 0.2 CRACK EXTENSION, (in.)
O JI nitiation Final Crack in.lb/in.2 (kJ/m_) 2 Extension Measurements Elastic Optically Pipe Elastic Compliance DCPD Compliance Measured %
Test
- (Figure 22) (Figure 23) in. (mm) in. (mm) Error l 1.101 (28.0) 1.41 (35.8) -28 3 2947 (516) -
3197 (560) 0.576 (14.6) 0.757 (19.2) -31 7 4397(77h) 2940 (515) 0.738 (18.7) 0.803 (20.4) -9 8' 3985 (697) 2530 (443) 0.733 (18.6) 0.995 (25.3) -35 10' -
11' 2042 (357) 2349 (411) 1.120 (28.4) 1.017 (25.8) +9 1.150 (29.2) 1.427 (36.2) -24 12 2550 (446) -
- 2873 (503) - + -
13
- + -
14' 3880 (679) -
4125 (722) -- + -
15 2496 (437)
Average 3185 (557) 3002 (525) -
l
.l + - No optical measurement l '
i
,O l
\- - - - - - _ .__ __ _ __ _ _ _ _ _ _ _ _
e Tearing Crack Extension Crack Extension 0.5-in. Range for Modulus I J Initiation Predicted Measured Specimen (13 mm) Thick Test 2 J Initiation Error Number Compact Spec. T emp. in-1b/in2(kJ/m ) inch (mm) inch (mm)
Calc. )
Ceometry l inch (mm) 263 0.049 (1.24) .066 (1.68) -50 RT 2073 (363) 0.048 (1.22) 202 1/2T 285 0.093 (2.36) .119 (3.02) -28 RT 2684 (470) 0.051 (1 30) Not 204 1/2T --
0.196 ( 4.90) measured 0.061 (1.55) 222 300 1/27 RT 3962 (694)
.272 (6.91) -30 0.057 (1.45) 230 0.209 (5.31) 301 1/2T RT 3220 (660)
.175 (4.45) -15 0.058 (1.47) 260 0.152 (3.86) 303 1/2T RT 2623 (459)
.224 (5 69) -47 0.074 (1.88) 355 0.152 (3.86) 100 1T Plan RT 3606 (632)
.132 (3.35) -37 0.075 (1.91) 396 0.096 (2.44)
IT Flan RT 2543 (445) 102 .186 ( 4.7 2) -12 0.078 (1.98) 352 0.163 (4.14)
IT Plan RT 3149 (551) 103 .128 ( 3.25) -14 0.071 (1.80) 340 0.149 (3.78)
IT Plan RT 3299 (578) 104 RT = Room Temperature
~G g .,
t 0.5 in. Thick Crack Extension Specimen Compact Spec. Test J Initiation Range for Tearing Modulus Number Geometry Temperature in.lb/in.2 (kJ/m2 ) J Initiation Calc.
inch (aus) l I
la-FD 2T Flan IT 1590 (278) 0.072 (1.83) 291 2a-FD 2T Flan Er 2197 (385) 0.086 (2.18) 234 7a-FD 2T Plan 125'F 2234 (391) 0.073 (1.85) 299 Sa-PD 2T Flan 125'F 2860 (500) 0.078 (1.98) 257 l
RT = toon Temperature Crack Extension Crack Extension 0.5 in. Thick Specimen Compact Spec. Test J Initiation Range for Tearing Z
i Modulus Predicted Measured Number Geometry Temp. in.lb/in.2 (kJ/m2 ) J Initiation inch (mm) Error Calc. inch (mm) i inch (mm) 0.526 (13.4) 0.563 (14.2) -7 la 2T Plan Er 2428 (425) 0.073 (1.85) 349 273 0.482 (12.2) 0.568 (14.3) -17 2a 2T Plan Er 2777 (486) 0.061 (1.55) 0.076 (1.93) !!3 0.485 (12.3) 0.506 (12.9) -4 7a 2T Flan 125'F 3661 (641) 0.056 (1.42) 290 0.476 (12.1) 0.478 (12.1) -0.5 Sa 2T Plan 125*F 2316 (405)
RI = toon Temperature i e e -
e
.t
!i
!! s i!
I l
APPLIED i
I, i
lf MATERIAL x
N (udsTABLE)
N ll J \
1 i
(ST/ABLE)
- i
~ --_-___.__--________
- IC T
Cotter, Chang, and Zahoor 1982 4
e O
i l __
CRACK EXTENSION, (mm)
- 2.5 0.0 2.5 5.0 7.5 10.0 12.5 15.0 17.5 20.0 22.5 25.0 I I I I I I I 1 I 8 34 12 14000 -
- 2500
/ '
3,,
,- 15 E '
12000 -
' - '1T y ,J' ,-
2000 m -
!/
10000 -
- - PIPE TEST o a - '
- m ' - NUMBER
- ~
d ~.5 , ,' -
1500 .x E 8000 -
,' ' m
/
6000 -
l/ , , - / /
1000
/ /
/ /
4000 -
f -
500 l 2000 - /
0 / POWER i i LAWi APPROXIMATION i i i i i TO ELASTIC COMP i
- 0.1 0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0 CRACK EXTENSION, (in.)
- - = - -
e.
G -
9
/
CRACK EXTENSION, (mm)
- 2.5 0.0 2.5 5.0 7.5 10.0 12.5 15.0 17.5 20.0 22.5 25.0 I I I I I I I I I 110 ' i 14000 -
/,/ 8 -
2500 f
7 11
/
12000 / . 13 -
I
/,/ ' ,
' 15- 2000 su 2"
/
/- / #g.- PIPE TEST 10000 -
3/ // 'g 4 NUMBER
- 9 m E / /#
.E .d 8000 -
j C
/ /[5-1500 2 "E a g..-
s*
6000 -
l// -
1000
//./
j 500 2000 -
POWER LAW APPROXIMATION TO D.C. POTENTIAL DROP DATA 0 / i i i i i i i i 1
- 0.1 0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0 CRACK EXTENSION, (in.)
O- 0 0
, i 3500 20000 , , i i i STABLE CRACK EXTENSION 3000 15000 - -
2500 I
2000 m 2 3
$ E 10000 -
\
's,
' JIT MATERIAL 1500
',,%~ EL
~~ ASTIC COMPLIANCE
_ , 1000 5000 -
IAL Drop
,TAPPLIED. TADA
- 500
, e TAPPLIED, ZAHOOR 0 J li i i i e i 375 O
75 125 175 225 275 325
-25 25 T
Ree
- 7 8- G g
/
20000 , , , , , i i 3500 STABLE CRACK EXTENSION _ 3000 TAPPLIED, ZAHOOR 15000 -
2500 1 g _
\
\
\
a
'g _
2000 s ,
a ~E
.N 5 10000 -
's
'N
~
l JIT MATERIAL
's
%' N -
1000 5000 -
N?*TENTIAt o
~~~Jee__
ELASTIC 500 COMpEIANCE TAPPLIED. TAD A 0 i t i i i i 0
-25 25 75 125 175 225 275 325 375 T
? IPG ID O .- .O O
e 20000 i , , i -3500 i i i 3000 UNSTABLE CRACK EXTENSION TAPPLiED, ZAHOOR 15000 -
2500 TApptigo, TADA -
2000 y a g \ 3 "E E
10000 -
1500 f
JIT MATERIAL 5000 -
- 500
' i i i i ' ' 0 0 i
-25 25 75 125 175 225 275 325 375 T
l IPE lb O -- -_
O _ __ __
e
~
d k 20000 , , , , , , , 3500 UNSTABLE CRACK EXTENSION 3000 TAPPLIED, ZAHOOR 15000 -
250?
TAPPLIED, TADA -
2000 2
i 5 10000 -
"E 1500 J/T MATERIAL -
1000 5000 -
= 500 i i l i i ' '
0 0
-25 25 75 125 175 225 275 325 375 T
Ree #6 0- O O
S t
m , .. -_ _
m_ - ,
_._ g 8 .. a w ..
- - J 7 _._
-_a ,
.. - muummuuuS lus i. .. -
BJE ::VES:
- CHARACTERIZE J-R CURVE BEHAVIOR OF 8-INCH OIAMETER GTA WELDED A106 STEEL PIPE
- INVESTIGATE TEARING INSTABILITY PERFORMANCE
- MODEL PIPE J-R CURVE USING COMPACT TENSION SPECIMENS
- e. -- _ -
O - - - - -
G
GTA WELDED A106 STEEL PIPE TEST MATRIX NUMBER TEST OUTSIDE OF TEM PERATU RE DIAMETER CRACK TEST SPECIMENS ( F) (in.) GEOMETRY CONDITION 3 550 8.60 SIMPLE NON-COM PLI ANT I 3 550 8.60 SIMPLE COMPLIANT I
}
i i
).
I
I m
/ _
V g __
/_
S ~~ A :: \ _ E S S S ~~ E E _ 3:: 3E
- . - _ .O. . . - - .
O O
- BlE ~~ :: V ES:
- DEVELOP J CALIBRATION CURVES FOR Jm ANALYSIS
- VALIDATE ASME CODE FLAW EVALUATION PROCEDURE
_-- - . - l- --_ - . __ _ _: _ _ _ . ._ _ . _ - - - - _ -
TYPE 304 STAINLESS STEEL BLUNT NOTCH PIPE TEST MATRIX TEST INITIAL FINAL MOMENT ARM MAXIMUM SPECIMEN TEMPERATURE CRACK ANGLE CRACK ANGLE LENGTH LOAD ID ( F) 29,(deg) 29,(deg) (in.) (Ib)
J GG K-100 550 57.7 79.1 15 23,400 GG K-200 550 78.6 81.8 15 -
GG K-300 550 98.0 116.2 15 16,980 GG K-400 550 119.2 131.5 15 14,180 GG K-500 550 138.9 144.8 15 12,800 i
i O O O
y . T h n N O g e d 8 -
5 - a W
- 5 a :
$ $ 8 g v v n g ,
i w
a
= 8 m :
= 8i -
5 :
E S r <
i g . =
5 g e5%$^ -
i g A
!g. h - "
g !
O o g i 1
5 '
5 s -
5 ;
R l s o
4 .
a.
Ne a
. _ S 3 3 Ss . .
M 4 g . . ,
8 a - * .
o a :
g -
o!
a
~
or m m m m m m n -
8 8 8 8 8 8 8 8 8 8 . 5 8 N & & 5 53 8 8 8 .
0]
Jo<o -c Jmm !
i
CRACKED 2 e.
R.
l .-
a- t p
a a
fifff1I LIMIT MOMENT Mp Mp = 400 R2 t Up (e, H, P)
NO CRACK CLOSURE Up = (1 - 3) (COSa - 1/2 SIN e) + n/2 P SINa CRACK CLOSURE Up = (1 - 5) 2a COS a I SIN e + " P SIN a 1-a 2 2 -
i O O O
l
- $ O a < _ i W
m 5 R
'_ _ 3 -
G
- to 8 a < 5 -
E w $ ^
EI m
el illi a 8s 8
o 8d k
3 Si!
5 g
2 a 4 s
i I I I I I 3 g g
- " + e
- g a g = g xowd/*Iid 8 O
-- -,,---,,-------w_ -
2: \ _ _ S :: : \ S:
- FORMULATION OF Jm FOR PIPE GEOMETRY CONTINUING
- ACCURACY OF LIMIT LOAD EXPRESSION FUNCTION OF INITIAL CRACK LENGTH
4 8* em e --same m een- % Posee
-=== - m--. - -
EEW WWD m M heWeme 6 h M mmmmmme U 1 .
em 6 m u - -- .. -
0- _ _ _ _ _ _
O - - . - .
O
OBJECTIVES:
! 1. CHARACTERIZE ELASTIC-PLASTIC FRACTURE PROPERTIES OF FOUR INCH DIAMETER WELDED TYPE 304 STAINLESS STEEL PIPE AT 550 F l 2. INVESTIGATE USE OF COMPACT i
SPECIMENS TO MODEL FRACTURE l TOUGHNESS AND RESISTANCE CURVES OF FULL SCALE PIPING
\ ** .
' il- .E *
'~ . I bf-' ' (E . + . ,
. * ,- - [, . _
.-...J)h t., q'., d Y . - ' .
h L
' ~ SY e N Qf.fl
, , cf.- . .
s g.L
'*t
' * .gg * " ,
_d - , _
'~
'.*?..
,,g
- V # h . -'; h f
,. , ? .
- 2. .,-.
. .n p h
c 3,. , y, '
l e f .'F. g ;*,, . '.. ... ._, ; . ',. .{ll gjhg, ff ..y ;9 3 p ,
,.py _
. , L : ':g _
,;.Q:1f l l .l.lI ..' e '. :f: j' ] .(. .. .' - -; ': ;- .
y,L'.l ,
- $ =
.4su s.1 . ---.
= -
u -
g 8 - . - ...- .
.- .r -e .
. + . l-.
N. ,g. [4y.. e z .
,[. ' .yg'{,(..'4. rh._ ".[..n.,'[
.. .\ . ,.
- 1
- h. , [.q .ef,g. -. .*.:. L
. g, ..,. ; ., ,5() k.t:. /..#.~-- -: >. -~.-1.,.
- 4. 1 n ,m -
. - 3. , s ,', ~ . .%. s. ,g.{ . @ [ :<. , , ( * ,. . . .
.c , , . .g
- t. sd.
y % .- - % - ~'
- = , W k.:% .: =-
.; c g.-g f f
.'g J, . . ,;
4l.A;G '?, .,:@' '; a >t q;Qg;j,9, yeh, p.s -
.g
. '"-. ; 3 q: p l % -
- e. ~...3.
. en, g:., - .. . d*e
- ;9: C, r, .. .
%- *:l ., 4 4 - ' . q *W=rd--- -
+ .,
E I '\, . . . ' , *,,..' ,g {r . Ey.~
I r
v ,:, =
_.,en,
.t k1 Q . j~ ,. t
.'.'l*'.
'w 7',
u-,;_., r
,.f , ), I ", ' , ',
'.~3. .s-
. , '$.s -
g.
- +[ M * , '. _ .) #h ,f, '. %.
- y + q 3, p ,
Y.'.
W '. .)
s_,
- nt,*
]p_ _ ~~
' ;- ? ;. ~.
-* "?-;N ,
m q.
y g' gg g'k'..y 4
E'*.','.,k*
i-j q 4 l,*
,, r ;-
'k,'-'.#"-
h ^4%
7 N { ': .*
f * . >,:g' .'- : _
- ~ , . &W' a. + .? .? '
,-(.
y., _Q,, -
. . . . .::e,, -v -- . - . . .; ~ .
. > $ y_ .. .y . m w ; . :. . . - - ;. , s
.? (p .f, . ..
' 4
, W'.w;., .,. 1 ,' ' .. ' j.> ..c 'e 4 T5 =
r j j. ' .< ',. .
+
.t j' ~.s ; y,7g\ '&'T. ,.'9 y._'
4 . , - 6, . ,
4 M-; ; . ..q ~. ;:; -
yn. a. 9 .
,~,
. ' ., y-
.y L.
+--- '4 .. ' y>s,. , mr
%* < , . y,h,i } .O',' . - "- - -. Y- ; ,
g;N _ p
- - 'M: i s .'f g
[_
.M - . -[. *
[ , -
\, ' ' + - 8 - = '
"'s " , .: :
} 't -l 'h ,. .* -
. 4 6, - WS[ " 2, b * *% .3. '-
-h , k S. '. " '* {' ' g h/- 33 J,. s. . .: ' ,
' ' - . L
-g. ,,T 2,t f g " _
3 .
. 4 ? .*. . s s . . ' ' , * .. ,g r "'sc* i:,
.}f .
- 4- . /* } - ,, * ,'. s R_ .
% 'f ' ,-
-k ' ': '
.. j
.'t',
~
-.4 ', - , . [.- "[if ).. ;g f
%,b,&."g ?vuf p;fG;$Vg..W
%* ;Q y'.g ..;, je . %P*' - .-
b 4 ,.
7 t- ~ . , , v, -
..6., . + . at
- -}' ".
- :; ~; t . c:. ._.v.* m; .,, , u'yw gN fy , v Wi. , . . t.. , -. ,' +' . y - ., y , 4. ;
-' . m%a
? .
- t <-g.r; 4
. r. M:: '&, . ? .M.fu c 11, *t ' .? $ %. V
',5
' 4.l. + : W. i., ,.: ,r; ... . u4;dl-l .- [j.p ? N, '?-v N.53
- s. - - 'J ,a : % % ? a.s e .*
1 4l >: 3f.+
>D r g
', g .. 64 ,, ,, .- . 4 -4.. ;, .1.l ;g ; ,.4 ,.' .
l' h 's
- f. -
g['. I D .-h
, r;.O w - b- ) . '-",p ..;.
g, i
j . . .,
I
.'Q n.
i'
,.4
'f ' .. ..: :,
% , a ".4
..,.',y
. p, e
.M'"
'E .
g,w ;4 5' i ! i.%g ,h. ~
~'
% Q:s b Q; Y,-? ? ',. L, , Qdg, ;*
,' .t - W kh Q. -; . v w.
- .,. s -
- a)n . #
s,r.,, y - . s5
. (I.f:.* "
u.
.u .a 4
.'.,4 p#,- 'sp".' e. vv e, ..'
q.w.g : .- A; a
- R.:k,b g%} - .;?. ., .)my. ; 2 -.
j u ,. T gets & 'st Q i:*
g'
- .: .7, C _
- { p. )l1
@w[M ~o.; . j ..p1;n.
. . w y g , - q-1 .h. c.r.,;',,s,'g-' ' ,. ?y ~ ,. .i, , j. .
3*,../. ,.
.m.
,- c- -
m 4-',. ..).*h
- . .g e."b , s.-. , y
- ' ' - - '* .. p - .
9>g.y -
i ,.+
f - **, tf,.,~p, .:..,.,_..'w , , , .s s y.
s V. . . ,
q, a....
,, y
<s.- . .,. . .,,'/tr .g-4 $ - .
- t *,,,,-; V Y -
..h
.,{ MP
_4,..- ?
'Q
^
?
T'....
- L
_ 'j . ' . .
- A .'
- , g '. , J ..g*
_ #, e ..' t
.{ am-
- = 4
- (t . a ,[ -Ms f-~' z ^;g Qo ,
'..v .t j_.' '_ . $ . ": '
,,Q o
- &p r;Q f 'r%p b ne . .
n
~
~'
e'm&#
w%%
M W 31b3 d:;v[- % d' q. ..eR' : ' e
- Me .&
Q , .n:'
y.. a. t3.y . %g. y;y : , .
- ' ;D1. _, Qa% f ,;; Q J ._ ; G
. n,
- s y
- . ? - pm
.. p.
'e..x. :.. :;G; , Q. 2. ~ ~ 1.
p.
,,_ q. 3. .ys:-Q,
. . s. 3 . , - . 4 S%
' f G)giL >
a.
c fq&f gM E
L m-
.r ,
. g ,.c . . . s
~ G J'.; g%. f .; f [.. Q,,,
,: .*.;; : 'o;.~-;p..n. yn 9 p . f
^ .
>? 1~gi .p.%
q .i ?g::3;;. ,..wyc -t 2- t y '3 -p xm -q g3.* y. . p. y_y u.y..;;4 3. . >
. p .y. . : q j.n >.:. .O . o. g
.g . - ; ' .
h(.[.i;*.Q9,,'g.-.h_.L..;fv
< ,.:- - 1 - G . . .g , p . .s, ~. y .. :. .:..q ; .o y * .;.; s-, --
' h' y j,\ (v,. . y,hij.' y. ;pf.),,.,7, ,y -; h b Q ,.f.-v [.I.; ~ jb '. h. .by.i. }q,'[. :
+s -- -<
I -
% . - n ..
>u , .s > -
.c :.' -
7 g
[,. ?. ; , 9 j. . . v ', y* -a, ,y rip,4y .M , 2.Sf.,N
{. . . .! 'l " ^ ~ V - - a%, g .
._.p,..y ? p. -- ; v: .,~.'o.4 :..Agmg.;,; :,,e.~p. Q. .: e...
%s. -e..,,
v.7>.
a , , s. .* . m, . .
.'s<b. ;< m..s -.yv, r.. c9., .? . ; e .., u.,, ..,m m <::
y w- .<+.. ..t
.y m
- v. m n
.3 y.- . g 9'..
>A .
- ,M. ~m, ,.4'p ..
I.,
r mt :a,p q,
c s y w
p
?, ,Q:
A . %g.- . w;w ,s,.., _q,y %::(
.x 3,,7l%,m. J.SQ
- s. -. ',;f.l>* n Q.9.Q.t,, -g )f4:: ! .;- m:g,n
_ .- .Q:.)}; 'q.,
3
..g.%
s e - .y . .:. , . .m# ;.p- a :..
- ,, u m..,j, m, Q.
- v. - e-,, e n 4. , <.i,, t.
.e. .
=
-k. f. . 1.e'. ' ,, '. .
hf; 's'-
y %glkp b $,l-f - ), ;-::J. .
El f ' 6':;O*ir$' f 1.,h
'$ $8}.S.',7Q': h "h Q4i h . _!h l . Q - Y.NN.' * ? $l'I~ - -
Mv f *Qjr;ry ' .- ,
WDQ ..,
2.H.W. R fi; N W
3 f,0 f. Y. .hi G' ~.
W W[lh.0 F y& - A 2diRDQ
_Q9h Mw.nawwFs@Q@g'h:,j?F. Wh.4 m %w li; w m
)
0.337!
m 0.110 1
PIPE WELD CONFIGURATION l WELD PROCESS: AUTOMATIC GAS-TUNGSTEN ARC
! FILLER METAL: TYPE 308L STAINLESS STEEL i
l HEAT INPUT: 30 kJ/in.
i i
i l
) . . . .
I I
TENSILE PROPERTIES OF TYPE 304 STAINLESS STEEL BASE METAL AND WELD AT 550 F YlELD ULTIMATE FLOW NG oR STRENGTH STRENGTH STRESS N _ REk (ksi) (ksi) (ksi)
WELD 44.3 65.4 54.9 33.0 74.3 42.2 64.3 53.3 30.0 72.9 AVG 43.3 64.8 54.1 31.5 73.6 BASE 22.2 67.3 44.8 39.0 70.8 22.8 68.8 45.8 40.5 70.8 l
l AVG 22.5 68.1 45.3 39.8 70.8 l
i O O __
O
i
. 4r w
i;i ' s - , , r [M* .)
2pf* : n y- l, h
n : . , n W ;3; ..! Mu
+ .. < .
6 & > $* g yF.. D
, - z. &'.f.v. kY ?v. * *. p ,
i '
[ k f<' h ^gg 3"
t l ' < , 'v *gf g- ff."5
a i ed %.
=
. . . .'-2,-@,yM, , ;3 f.74u #
.];- Q ', *, h e.,, _' . w s a-st.,,
e.y 7 , ~v ?*
7%
,- ,%' M.-W{?bl ~ ,p
.-% q?
-) &% N W.
, ;-; . ',', 2 ' ~
1
- 4; *
,,,4,4 < ~: N ' y+TN - w-[%.:'
)
. ~ ,g. .
,..,m-S 'O m v 4
ng .P i
i ,g,% 3 '-
-., m
- 1.,g . ,h,gs y t
- w 3. ,. /
u ,v., e.cA w.~ %,9 A.
- $% w' m p ib, *c
- }.o- -
'p ~ .
%4L ea ww -
se.A..n 2
, ,g ,
.b ?r 3
& N w
J7<c
~.
O' R ?$.. x b.;m, f%. -
- 4 bgc:
g sy
- ~,x . y
- v =, ~r.y
- s
_g j (- : . M_ . v W 7 N.?
gx ,,,,. ( *.C
. .3.,) 4 >g ' [a ~ n ' ,. t / Mj;h, .
.s' -%
g - 7e
. ce v
- GPi,3.u G%
9 q-d${WW 5 i u <f, Y ig y
?}6 cfnpW 'W A g%
yMW -^
n.n .<ys,;g 3 Jn pie /.d ~
c .s o s , h "m" .,1
- , g.FO'W
. . , , * ) h,I.. jg/ [ N *(f(E'$e
. ( .' M' . . - - . . ~s '#
$. = , *~ .
s 4
,' %@-U, L"b - ,
, bl% f. w
.g.
V , 8 - (gO,
^
b.,7,' [ -._
w -' ,. . ),
N,, f. . ,.
. ,== <'. b.;f;'- '
y es-3 "
W,Q^
s **
3
,, o
, w~w
, h '4 * k
- k W
t w$ 4 %
.J.n . .s
- l,,,,,, +~ v,.
,.u _O Q.k G ;q h At 7 [pMl1. # :. .:' " iJ.,,-..y'-
- .'*k yje. ,
y'
{ N n .
-
- f i ,.[ g y~pr m. e . ,
N
&-oL: . ,
,;; - (
y . "h W ' 'p , 'f >
ns c .
y, ,# a.fi.:
q ,
.\.,.
%th's( ,m['- '*< Y Tg'c,'A
' * . r
.w
'8 fW4 j,' _
- y?
en - . ,
a mfd
, ' ~' *my . , ,;E~
'e,.,., .a
- f V:
+ >
'dQA * ' Q* ' $$-'.N' ,)
YQ v4 ei t aM .'
paa' 4 ??
4t sgw ' I W ryg ;.,7. *g V3
%Q 5 .r.
)hWDl4 n w& M na v
l
- f(&1<bN -- b s.O}4% -r , .s {;w, p ,.
Rl
'--
- d ,4 g . Mg M* .*
., , ,N e y ,.* Al~'
~L.
- s. . v M..Vm' wo 4+ % ,g:
-s e. u ,
. m M, .. ,.
W qp%,.?i:js'ir4.. . . .j+W
@mmmm 3. n . w m m ,h p
- g ge % m,7 y
s
). ' '
- y N O N,.y:v. m$mgt g gy;4 p-i
.~*
- . f.j ~ 1 m.& hy'd2lf&e e a - nY hl$N C
-r -
,4 7 y ,a _
~. >- .n ,
f g g. ;p gqy&i A &;g:s ug Qq(@ W Kig:CG" 4
h f-?f'? m% $AF/.
&)
, , - ., ~
w , 3,.g% -+ us
- n. 4 .,m$h f 3 .
.a.. s. ~ va %p%, _.9
.Q
- m. w#.
ggggN
- v. -g r %,(4.3 p" }U
?
?
- w. ~j wj-u. ewg.c;%n > ~ wn s -
E,,N-O g ..
+ .,
.~A , c- - - ,,
- ~
r tbY Y 'e ,
,. . 'g..
l A M 3 i'<- , "'
. . - - - -,_-_,-e,. , . - - - - . _ _ - , - . . - , - . - - . --------.-.----e,.
7 0
K C
A R K 6 C C 8 X A 0 E R L C P E M L 6 O P 0 C M I
S 5
0 5
0 2
8 9 5 1
4 W
/
E 0 a C g Y
O J F 4 1
0 1
W
/Wa/
La ' 5 AL 0 A
mN NI a I F 3
++ 0 6D I 5
2 f
0
- - - - - 2 O
3 5 2 5 1 5 0 0 2 1 0 0 0 0
0 0 0 a
u0242 zouONa$ OZ l1,l '1 !
! l .
(mm) 0 5 10 15 20 25 30 36 40 N I 3500 I I I I I I I I 18000 -
3000 16000 -
~
14000 -
12000 -
J(igp)
I -
( "')
1500 8000 -
6 Ao 6000 o O 6 6 -
1000 GAM-100 O SIMPLE CRACK,20= 100*
4000 -
O 6 ELASTIC COMPLlANCE o -
500 2000 -
0 ^ I ' ' ' ' ' ' I 0.00 0.20 0.40 0.00 0.80 1.00 1.20 1.40 1.60 CRACK EXTENSION (in.)
e 9 9
J-R CURVED F3R WELDED TYPE 304 STAINLESS STEEL PIPE 20000 18000 - !
16000 -
14000 -
7 5
I D
12000 -
D E .$ D
.N l 10000 - I D a b A b 8000 - d a 3
6000 -
gb U O GAM-200 4000 - b SIMPLE CRACK,20=139
. A! ELASTIC COMPLIANCE O POTENTIAL DROP 2000 -
l 0 ' ' ' ' ' ' ' '
O.00 0.20 0.40 0.60 > 0.80 1.00 1.20 1.40 1.60
.i CRACK EXTENSION (in.)-
J-R CURVED FER WELDED TYPE 304 STAINLESS STEEL PIPE 20000' 18000 -
B 16000 -
I A A 14000 -
A D jd 12000 - D
.d E 6 O
, 10000 -
D 8000 -
6000 -
GAM-900 SIMPLE CRACK,20=107
[
6 h ELASTIC COMPLIANCE 4000 -
O POTENTIAL DROP 2000 -
x A . .
0 . . . . . -
0.00 0.20 tI 0.40 . 0.60 1 0.80 ; 1.00 1.20 1.40 1.60 I
)! CRACK EXTENSION (in.) ;
g ,.
J-R CURVED FbR WELDED TYPE 54 STAINLESS STEEL PIP 18000 -
16000 -
O $
14000 o
12000 -
E .5
.5 E 10000 -
o a D 8000 -
D D A OO A O 6000 -
O g A
M SIMPLE CRACK, ELASTIC COMPL.
4000 -
0A A; GAM-100, 20 =100
!DO # "I O GAM-900,' 20=107 3
2000 -
- ^
O ' ' ' ' ' ' ' '
O.00 0.20 0.40 ,
0.60 0.80 1.00 1.20 1.40 1.60 CRACK EXTENSION (in.) i
- g. G #
.<-.--a - h --- .a a h a-- _.aw. - . . - - . - . . . - - - - - - - - - - - - - - - - - - - - -
j rv e.4.mn v:-N.m... ,--- - --
?, * ..c,'.
ss., - ~ 4' N
- sc,' ng,%yf ew, *,m vye_k.m -fy W .4 *. **G vg,mgx N,', cxy~m$ > s'j*}- ",. QA 1. . .' . A. 4- 21
.~>,3
- . i s .
ss -c
.. ..i ..
.\g.
- g. ' j.eUf@.1 W c W[%y; , , , , ,.9 -
, ,,g4 t*,M$/ ',,,g 9, -
' :@',f
,~ y. s p,c,..p 'L* f . '; ,
-. f j/
WQjq,p y '3,,' N,,%.
- q -
'F@*/MWMN&[>43.M W(wn@.g'. s UNs.'.,e i _ C# M ' ~ <
~A
.mn. i@ Wa4i am,;ve, ..v 5*[-
h h ^[4 9 Y h g A;p \ cQg u .s _M ' @e.s< n "7h Jky,. , ,. . + .: , . o,e ,e w, i, w. ~ w
' ,;y'Nyo if '@e.$e$,~y[w c u g: .
$fI'73 ' ' . 6"hi(#Q. I
- W
.+d'q i Q m bA-de'!Mifit?* : .js x W v t. *.
, -+ .
T ',' '
eq .94 - -a ,.v v
- ,[,44' e'.4 N. * * ' jg;s s.
, ,;^ '
I
/Q: ' ['y; k~,\ . . S i ;(/ M[# .,,' '
h .. .ly.t y.s
. fw.
g[y$- /
} Q:-l$ '
I
$b t.
'f< :&. . i, - , ,y,% , \w$. . ' ., ., .
',7
}-
-d f V ,;M j me .
I -
< * .;g
. _. geh ~ . <
x / N 4
, s h. s i
' x f k k l
- 9 -
~. kn .
,N
' tu cr. -
i
^
%g
- k. y A. -
0 I l
. .t , ,
O 6 -
s !
l
.., i
,s i v ,
.3- .
gj. '
.x v. .
/
s 4 _* .
'g
.-u , ,+ , w , <
. ,h' [ .;
, ; p i,.,'
q&. . . a
% * ',. e t.
~
- > fy 3 .I ~ . '
g .
Y' e G, < #j
\
[
+
':y e .&* . *e . .s
. , 1
. . e yk
+t <
,$ @p g .A-
)a,}. e.- rgw ?r ,;g,4 ' VWg
- f. ,
- r. ..
- j c.
gygtoys*). . .t.c t :.
=
d # g, .
t
-E "
e' VL*p* w'
- p\
y ,,[',n= *
- . \'y 1 w' 3.( - *w* %
4
- '(,
- 4.. c.*
c' fe' .,'[(
~ '
+ - - . ~
a.,_ . - _-.- [ -
1 t'
J-R CURVED F3R WELDED TYPE 304 STAINLESS STEEL PIPE 20000 4
I l
18000 - i 4
. i o
16000 -
i 14000 - 0 ,
7 I D 12000 o D j
7 f D
!. 8_.d_ o j,yg 10000 - I D
, oo l 8000 -
l "
6000 -
]9' SIMPLE CRACK, POTENTIAL DROP O G AM-200, 20 =139 1
{' O G AM-900, 20 =107 1
4000 -
u 0 0.60 - 0.80 1.00 1.20 1.40 1.60 0.00 0.20 0.40 l
' CRACK EXTENSION (in.)
ll,.
1 J-R CURVES FOR WELDED TYPE 304 STAINLESS STEEL PIPE 10000 t
9000 _ i GAM-400 8000 -
COMPLEX CRACK 1 20=154 , a/t = 0.38 7000 -
1 o ELASTIC COMPLIANCE V POTENTIAL DROP 1
.55E r
4000 - 1 Y
V 3000 '- Y
- 9 1
V o 2000 - V L t
ov 9 l dV 1000 -
0 ' ' ' ' ' ' ' '
O.00 0.40 i 0.80 1.20 1.40 1.60 0.20 l ! 0.60 l, 1.00j
, CRACK EXTENSION (in.) -
- e. .
O. - - - - -
O
i 10000 I J-R CURVES"FOR WELDED TYPE 304 STAINLESS STEEL PIPE t
9000 -
I f
8000 -
i 7000 -
r 6000 -
t 5
D
.I$ 'c
~
.5 g 5000 -
I D 7 D i
4000 - I 0 a l
D 3000 -
D i
D GAM-600 2000 - t COMPLEX CRACK D 20=128c, a/t = 0.256 D
D POTENTIAL DROP 1000 -
D 0 ' ' ' ' ' '
0.00 ,
0.20 j
- 0.40 l !.0.80 : 0.80 , 1.00 1.20 1.40 1.60 i CRACK EXTENSION (in.) i g _. O
J-R CURVES FER WELDED TYPE 304 STAINLECS STEEL PIPE 10000' i
M- COMPLEX CRACK t
POTENTIAL DROP
~
v GAM-400, 20=154 a/t = 0.38 D GAM-600, 20=128 a/t = 0.256 i
7000'- t 5
6000l- '
D D
.s if 5000 - 1 D
i D
4000'-
D Y y
3000 f-D Y l D V V V 2000 -
y U D
V0 1000 -
0 - - - - ' ' ' '
O.00 0.20 t 0.40 0.60i 0.80 ~ 1.00' 1.20 1.40 1.60 CRACK XTENSION (in.fi
- - - - - - - - - - - - - - - - - . _ _ . _ _ _ = - -- - - - - -
= -
J-R CURVES FOR WELDED TYPE 304 STAINLESS STEEL PIPE I
10000 -
l 9000 - I t
8000 -
3 A A 7000 -
r b
l b
6000 -
ELASTIC COMPLIANCE t
3 a SIMPLE CRACK, GAM-100
.E.5 6 20=100 y g 5000 - '
3 V. COMPLEX CRACK, GAM-400
, 20=154 , a/t = 0.38 4000 - 1 0
4 3000 0 V VV V V
'A y V 2000 -
p y V V
Y 1000<- p 0 0.b l 0.b , 0.b l 0.80 1.00 1.b 1.40 1.60
. CRACK EXTENSION (in.) ;
9 O O
J-R CURVED FOR WELDED TYPE 304 STAINLESS STEEL PIPE 18000 t
16000 -
, r 14000 -
1 5 D
- D 12000l j D
POTENTIAL DROP E .5 1 -
D D SIMPLE CRACK, GAM-200
,N l I D 20=139 g 0 COMPLEX CRACK, GAM-600 8000 -
20=128 , a/t = 0.256 0
6000 -
9 D Q D I D 4000 - O D D
i D h-D 2000 -
0 D
DU Ol - - .
0.00: -0.20! 0.40: 'O.60! ' O.80' 1.00 1.20 1.40 1.60 CRACK EXTENSION (in.); :
~
LOAD VS L@AD LINE DISPLACEMENT FER WELDED TYPE 304 STAINLESS STEEL PIPE 24000 22000 -
20000 -
18000 -
s SIMPLE CR CK, GAM-900' 16000 -
i 20=107
~
_ 14000 --
o 12000 -
<t O
10000 -
COMPLEX CRACK, GAM-400 8000 20=154 , a/t = 0.38 6000 4000 2000 -
l ' ' ' ' '
0 O.000 0.500 , 1.000 j 1.500 ; 2.000' 2.500 3.000 1 , LOAD LINE DISPLACEMENT (in.)
i .
~~
1 COMPACT SPECIMEN GEOMETRIES PLAN SIZE NO. OF SPECIMENS W (in.) B (in.) a (in.) a/W B/b 3T ] 6.00 0.34 3.90 0.65 0.162 2T 2 4.00 0.34 2.60 0.65 0.242 1T . 2.00 0.34 1.30 0.65 0.486 N N
~ bi : a i nl: l
~
'\
i n4 ,
\ ....
'Bl Wl- r O- O O
304 STAINLESS STEEL. PLATE WELDMENT e PREPARATION: 2 INCH THICK DOUBLE-VEE BUTT WELDMENT e PROCESS: HOT-WIRE AUTOMATIC GAS TUNGSTEN ARC WELD (HWAGTA)
- FILLER METAL: SFA-5.9, ER 308L e HEAT INPUT: 21.7 KJ/IN.
Ars/v
/- /
/
+---6.0 I N .: r-l O, - , . . . . . .. , $ .,. ... g
i 0
@ 0 0 0 0 0 g m 5 1
0 1 5 2 2
- _ - 0
_ _a 0
2 A 0 0 T 5
I A -
D 8 1
h E
C NN 0 -
N A NN I AALL AA -
LLL PP 6 0
4 I
+o PPP MTT22 O 1 1 TT
,1 2
' 1 0
o C1 -222- - - 4 .
CPPPP
' 1 T GGGG I )
0 . _
SGGGG n g i
(
O A6o L
E '
2 1
N O
0 3
I O0 0 I S
N
) E
(
m m b oo '
0 1
0 T
X E
g K 8 C
' 0 A 0
bm 0 RC os I
2 6
' 0 .
n 0 o ' 0 4
0 1
I O 0 2
' 0 ._
0 0 _
0 O _
0 I .
0 0
0 0
0 0
0 0
0 "oa 0 0
0 8
g 0
0 0
2 0
4 2 0 1 1 1
)
a (g
J l l i
- J-R CURVES FER WELDED TYPE 304 STAINLESS STEEL C@MPACT SPECIMENS
'i D
i
'" ~
/
12000 10000 -
[ D D D D B
pD D D 8 m .
D
- 9 .E tr 8000 -
Qp AoI D D D DD Y U DQg@
M00 -
g6 8
A POTENTIAL DROP DATA
^
AIGGP- 1,1T PLAN 0:GGP- 2,1T PLAN 2000 - Q GG P-21, 2T PLAN O iGGP-22,2T PLAN l olGGP-32, 3T PLAN 0
0.000 0.020 0.040 0.060 ' O.080 0.100 0.120 0.140 0.160 , 0,180 0.200 0.220 i
- ' CRACK EXTENSION (in.)
~
- J-R CURVES FOR TYPE 304 STAINLESS STEEL 20000 t
18000 - I t
16000 -
D E 14000 -
I D
1 12000 -
t D5 U
3 g 7
o- 10000 - 1 E *
, A A A m _ b g O
g Ib 6000 -
4 >D A O A .
8 ELASTIC COMPL!ANCE DATA 4000 -
A PIPE, GAM-200 A O PIPE, GAM-900 l g _
' Q1T PLAN COMPACT, GGP- 2 0 l2T PLAN COMPACT, GGP-21 0, "# ' ' ' ' ' ' ' '
0.00 0.20 0.40 0.60 ' O.80 1.00 1.20 1.40 1.60
!: CRACK EXTENSION (in.)
O. -
O
J-R CURVE 3 FOR TYPE 304 STAINLESS STEEL 20000 t
18000 - 1 5
f 16000'- U i
14000 -
D A
i A
12000 -
p D A m . b
.c C
}- o- 10000 - I D' O P
8000 - i 6000 -
4000 - , POTENTIAL DROP DATA A PIPE GAM-200 Ci PIPE GAM-900 l 2000 -
Cl1T PLAN COMPACT, GGP- 2 0 i2T PLAN COMPACT, GGP-21 0 ' ' ' ' ' - -
0.00 0.20 0.40 0.60 0.80 1.00 1.20 1.40 1.60 i
i ~ CRACK EXTENSION (in.) '
O- . ._. . -._ _ -
O . . . -
O
WELDED TYPE 304 STAINLESS STEEL PIPE AND COMPACT SPECIMEN FRACTURE TOUGHNESS RESULTS SPECIMEN SPECIMEN CRACK GEOM./ J-INIT, EC J-INIT, PD ID TYPE SPEC. PLAN (in.-Ib/in.2) (in.-lb/in.2)
GAM-100 4654 -
GAM-200 SIMPLE 5501 6086 GAM-900 9117 6575 PIPE AVERAGE 6424 6330 GAM-400 COMPLEX 1596 1559 GAM-600 -
1097 AVERAGE 1328 GGP-32 3T -
4886 GG P-21 2T 7951 5850 GG P-22 2T 7074 6406 GGP-11
^
1T 4727 4600 G G P-12 1T 6197 5526 AVERAGE 6487 5453
l l
CONCLUSIONS (Continued)
- CRACK INITIATION FOR COMPACT SPECIMENS OCCURRED AT AN AVERAGE J LEVEL OF 6000 I N.-LB /I N.2
- GOOD AGREEMENT BETWEEN THE PIPE AND COMPACT SPECIMENS USING THE D.C.
POTENTIAL DROP TECHNIO.UE INDICATES THAT CRACK INITIATION MEASUREMENTS ON LABORATORY SIZED SPECIMENS MAY BE APPLICABLE TO PIPE GEOMETRIES FOR WELDED TYPE 304 STAINLESS STEEL
l l CONCLUSIONS
- CRACK INITIATION FOR CIRCUMFERENTIALLY WELDED TYPE 304 STAINLESS STEEL PIPE WITH SIMPLE CIRCUMFERENTIAL THROUGH-WALL CRACKS OCCURED AT AN AVERAGE J LEVEL OF 6400 IN.-LB/IN.2
- INCREASED CRACK TIP CONSTRAINT DUE TO INTERNAL NOTCH IN COMPLEX CRACK GEOMETRY REDUCED THE J LEVEL AT CRACK INITIATION BY APPROXIMATELY A FACTOR OF 4 AS COMPARED WITH THE SIMPLE CRACK
- PRESENCE OF THE COMPLEX CRACK ALSO SIGNIFICANTLY REDUCED
- APPARENT TEARING RESISTANCE OF THE PIPE
!
- OVAllZATION OF THE CRACK CROSS-SECTION AND LARGE AMOUNTS l OF PLASTICITY IN THE NEAR-CRACK REGION LEAD TO UNCERTAINTY IN
, ELASTIC COMPLIANCE AND D.C. POTENTIAL DROP CRACK LENGTH i ESTIMATION TECHNIQUES RESPECTIVELY 9 O e -_
(<4 .
h o
t f
l _ , _ _ . . .. _ -
S __ .. _ \ X.. _
_ W S__R_SS _
_ R g -
y___ _
S- A :: \ _ ESS S- = E _
3:: 3:: \]
O O O
'k . -
TYPE 304 STAINLESS STEEL, 550 F FLOW STRESS FROM 35m 90 -
l 80 - j8Pec) 2.7YS eas) esso *F
- YS pec)
(Y3+uTS)/2 es!o5 f I
. 44.14 73.6 66.02 60.00 50.76 60.11 G MINIMUM Q AVERAGE E MAXIMUM e :. _ _ _ _ __
O O
f ae.
~
TYPE 304 STAINLESS STEEL PIPE, 550 F BASE METAL 30 -
29 -
28 - +
[ 27 26 25 -
J 24 -
h 23 -
i? 22 -
@ 21 -
]g 20 -
aC 19 -
Ub g 18 - o o 17 - ,
O 16 -
g 15 -
14 - ,
f g 13 -
12 - a 11 -
10 , , , , , , , , , , , , , , , , , , ,
10 12 14 16 18 20 22 24 26 28 30 (Thousands O MEAS. FLOW STRESS MEASURED MAXIMUM) +
LOAD (Ib) .
3Sm FLOW STRESS e------ _ ---- --- _-
G 9
c - - ~
rg.py gy .
TYPE 304 STAINLESS STEEL PIPE, 550 F
~
CIRCUMFERENTIALLY WELDED 22
+
21 - ,
o 20 - i O 8
' 1 9 ..-
t- o +
2 M 18 - O O + i i ]}g 17 -
po 06 16 -
i D o 15 -
n
! I 14 - +
m i
E '
13 -
D 12 18 18 12 14 20 22 sonde
.MEASUR MAXIMUM) LOAD 1 l D MEAS. FLOW STRESS + 35 STRESS i
1 Q. _. _ - _ _ _ - _ . - - _ _ .
'Q' Q
(
- ~~ - E R : \ ]::: \ ] RESEAR -
1
- UPPER TRANSITION FRACTURE TOUGHNESS OF A5338 STEEL
- EVALUATE GEOMETRY DEPENDENCE / INDEPENDENCE OF CURRENT Jm FORMULATION
_O_ :_ .
O O
O 3 3 \G rRAC RE V EC- AN CS JA~~A 3AS E (3 rRAC) by 4 O ~ A. _. - SER MATERIALS ENGINEERING ASSOCIATES, INC.
9700-B PALMER HIGHWAY LANHAM, MD 20706 (301) 577-9490 PRESENTED AT: ;
ACRS METAL COMPONENT SUBCOMMITTEE O JULY 1,1986 <
i i ;
- O 1
i i
~~O 3 CS I
i e DATA BASE IMPLEMENTAILON.
l 4 ;
i e AGED CAST STAINLESS iO i
j e OTHER EXPERIMENTAL RESULTS 1
1 i
.e>
i l
)
1 l
iO
)
l i
ii ! li! l Ill
. t J
a l
a , m de ev E ir v e J ir L er -
lpu B t J t -
pc A ac
_ - A T M
- S -
N U
k E
- c
- a L r B cth
~
w d o A ~ f T oor dd e S - t g
r u- - e s t n cy at i
- O rl ti f
b -
ota s
t n ei s
n -
O J
/
y t
r e -
p o
r p - H T
l a - W i
r
- O e R t
a G M -
K C
- A R
C i
c i
J J
- ! i
- i ' ! l
~
O R ES EARC- GOA_S Formulate comprehensive computerized numeric data base, for use in postulated accident analyses.
A33ROAC-O FORMULATION OF DATA BASE SURVEY OF PIPING USED -
COLLECTION OF DATA AVAllABLE PROCUREMENT OF MATERIALS l ESTABLISHMENT OF DATA BASE
-Computer System O -Testing l
l 1
o e
JA- A BAS E S-R C m RE e Jata ~~yaes
) -Chemistry I -Tensile O -Charpy
-J-R Curve
~~
e est 3 a ra me':ers
-Material Type / Size
-Temperature i
-Orientation lO t
i
O O O t
List of Commonly Used Piping Haterials to be Included in the Fracture Toughness Dara Base Material Nominal Wall Typical Application / Comments Specification Diameter Thickness (in.) (in.)
SA-516 Cr. 70 30 to 42 2-1/2 to 4-1/2 S t r'ai ght Pipe and Elbows in Hain Coolant Loop / Plate Material formed into Product SA-106 Cr. C 30 to 42 2-1/2 to 3 Straight Pipe in Hain Coolant Loop / Seamless Pipe SA-351 Cr. CF8H 30 to 32 2 to 2-1/2 Straight Pipe and Safe-ends in Main Coolant Loop / Cent. Cast 316 SA-351 Cr. CF8H 30 to 32 2-1/2 to 3 Elbows in Main Coolan Loop / Cont. Cast 316 SA-351 Cr. CF8A 30 to 32 2 to 2-1/2 Straight Pipe and Safe-ends in Main Coolant Loop Cont. Cast 304 S A-3 51 C r . CF8A 30 to 32 2-1/2 to 3 Elbows in Main Coolant Loop / Cent. Cast 304 SA-351 Cr. CF8H 12 Sch. 160 Surge Line and Branch Pipe Safe-ends / Cent. Cast 316 SA-182 F316 30 to 32 2-1/2 to 3 Safe-end in main coolant loop / Forging 316 SA-182 F316 12 Sch. 160 45* and 90* Branch Nozzle in Main Coolant Loop / Forging 316 SA-182 F316 4 to 6 Sch. 80 to 160 Branch Nozzles in Main Coolant Loop / Forging 316 SA-182 F304 30 to 32 2-1/2 to 3 Sife-end in Main Coolant Loop / Forging 304 SA-182 F304 12 Sch. 160 4 5* to 90* Branch Nozzle in Main Coolant Loop / Forging 304 SA-182 F304 4 to 6 Sch. 80 to 160 B ranch Nozzles in Main Coolant Loop / Forging 304 SA-376 Type 316 12 Sch. 160 Sarge Line and Branch Line Safe-ends / Seamless Pipe 316 SA-376 Type 316 4 to 6 Sch. 80 to 160 Auxiliary Piping / Seamless Pipe 316 SA-376 Type 304 4 to 12 Sch. 80 to 120 BWR Piping / Seamless Pipe 304 SA-376 Type 304 20 to 24 Sch. 160 BWR Piping / Seamless Pipe 304 SA-358 Cr. 304 C1.1 4 to 12 Sch. 80 to 120 BWR Piping / Welded Pipe 304 SA-106 Cr. B 4 to 12 Sch. 80 to 120 BWR Piping / Seamless Pipe-Carbon Steel SA-105 Cr. 2 4 to 12 Sch. 80 to 160 Branch Pipe Nozzles / Carbon Steel Forging Weld: SA-516 Cr. 70 30 to 42 2-1/2 to 4-1/2 Pipe to Pipe or Pipe to Elbow Weld in Main Coolant Loop / Field to SA-516 Cr. 70 Weld Weld: SA-106 Cr. C 30 to 42 2-1/2 to 3 Pipe to Elbow Weld in Hain Coolant Loop / Field Weld to SA-516 Cr. 70 Weld: SA-106 Cr. C 30 to 42 2-1/2 to 3 Pipe to Safe-end Dissimilar Metal Weld / Field Weld to SA-351 CF8H Weld: SA-106 Cr. C 30 to 42 2-1/2 to 3 Pipe to Safe-end Dissimilar Metal Weld / Field Weld to SA-182 F316 Weld: SA-376 Type 316 4 to 12 Sch. 80 to 160 Pipe to Safe-end (Ford ng)i Weld / Field Weld to SA-182 F316 Weld: SA-376 Type 316 4 to 12 Sch. 80 to 160 Pipe to Safe-end (Cent. Cast Pipe) Weld / Field Weld to SA-351 Cr. CF8H
1 O O O i
l I
CHARPY ENSILE I J-R CURVE DATA DIA. T CURVES CURVES 300 *F e 200*F 550*F t 50*F 650*F
- 50*F MATERIAL , l Room TEMP Un] On) CL LC CR LR CL LC CR LR CL LC CR LR CL LC CR LR CL LC CR LR CtlLCCR LR s < ,, 4 ss=lOb ( F A le [ 19 l) 4, 9.
sa-ins ,.r.
s,- i o s . . 2
( in to 42 -"
. .. i:
-1
- u; i ii
. i i
_ E_ ,
i u - m <,. 70 m ie e s-o l l l l l l l u- m n i,, mo u n-i l l l l l l l sa- m e n ., u ssi ... u s, w o, 3:
m .. c n-3 2-n X _ _' E [ME XX__
- l r
u-isi <,.<rsa m to i: 2- n l l l s s- is i c.. . < , s4 n to e n-3
--~M g g g ,
s s- 3 s i < , . < , 3., o s,s iso l l sa- m 3,- n ,, a s,s iso l l
.,os ""o-s s- i ,, 1 , m l __ __
s.-n.,1,n
%A ]7h Tvpe lo4
- w. . .n IRTf?_
20 to 2- Scs 160
__ E____ __ __
SA-lW tr. 104 eI. I 4 to 12 l[N$g' l l uu.. s3- q .. . m te.
30 ,, ,, ,,,,,,, j l x 3ca c: ,f ':c" ' "' w . ., e ,, - 3 I l
- Unii',; D ,4 ' l' io to . n-3 l l to io to 62 n-3 l l 1 gpl4,spr)a
- ugggg , nr. i r. tm to . ,o u y 97 l l
- e w t;ntsp r
- L- . io o e irc l C>< X X ionso t m e sa so l l l l l l l l l
l l
I l
l
1 o
AVA_A3_1 JA A GE STAINLESS 16**
l '
W STAINLESS 10*
W STAINLESS & WELD 24 B&W A106-C & WELDS 50**
O .
NSRDC A106-C, WELD,304, 50 CF8A,A516 BCL SA 376-304,A533-6 ?
A358-316L,A358-304, A106B & WELDS DUPONT/ MEA 304, WELDS,HAZ 78 HEDL STAINLESS & WELDS ?
O l l
O V A~~ER A_S 3 ROCU R EJ SA 106 Gr. C 34"x3.4" SA 106 Gr. B 6" Sch. 40 6" Sch.120 6" XXS SA 351 CF8A 30"x1.4" (304 Similar)
SA 351 CF8 35"x2.5" b SA 351 CF3 36.6"x2.9" Weld-Cast Stainless 23.6"x2.4" SA 182 F 304 20"x1.2" SA 376 TP 304 6" Sch. 40 6" Sch.120 Inconel 6" Sch. 80 - -
l l
i O l l
i J
\ -
3 l FUTURE MATERIALS .
i l
! SA516 GR. 70 37"x3.5" j SA516 GR. 70 52"x4.5" BIMETALLIC SAFE-END WELD 37"x3.5" l
! SA155 -
STAINLESS AND CARBON WELDS (CONTAMINATED) .
4 SA106-B/ WELD l
347 PIPE / WELD
)
4
z
!O :
1 t
i
- CO V L ~~r R Z
_ rJ
_ JA~~A 3AS r i
4 ;
i-i i
Rec uiremen':s i
e User friendly
- e On-line accessibility to
! NRC / Contractors / Vendors
!O r
i e Processing Capability !
t
<O l
O O O -
i JA~~A _E S~~ RUC~~U R E REFERENCE i
TENPROP IMPACT Kic JRSUM TENDATA CVDATA NDTDATA JRDATA OTHDATA l
i TENTAB JRSYS 4
4
P11~RAC 1
'A . Reference Inforrastion (0) RCF t ____
- 1) TITLE
~ _-________-__--_--__- ---_-_-_--_----------------_-_-_-------_-----_---_-----_
(2) AUT10RS (3) AU TilORG AIFILIATIONS (4) RCI'CRCNCE --_-___-____-_--_________
(5) SPONGURING ACCNCY (6) CONTRACT NUMDER _--____--______-_--___--__-_-_.----_________---_--
(7) DATC ._______
(0) RCMARKG ___-__-___--___--______________-_-____--------__-
.FMI PlFRAC 2 D. Material Identiftcation (20) MATCRIAL TYPC__-________--_-_--__---_____---______-__--________
(21) GPCCIFICATION ________-________-___________--_-____________-----
(22) PRODUCl FORM _ _ _ _ - - _ _ _ - _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ - - - - _ _ _ _ _ _ - - - _ _ - _ _ _ - -
(23) MANUFACTURCR _ - - _ _ - _ _ _ _ _ _ - _ _ _ - _ - _ _ _ _ _ - - - - _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ - - -
(24) llCAT HUMDER _______-_---______--__---
Composition (Wt. 7. )
(25) C ___-_- (36) Al --_-__ '
(26) Hn ____-- (37) Co __--__
gy (27) P ______ (30) H -__-_-
(w- ) (23) G -_____ (3V) As ______
(29) Si ______ (40) Gn -_-___
(30) Hi ____-- (ii) Zr __--__
(31) Cr -----_ (42) Nb ____-_
(32) .Mo ______ (43) Ob -_____
(33) Cu _-___- (44) Zn ______
(34) t1 --____ (45) N _-..___
( . ,~, - ) ft ______
(46) ANALYSIS CODE (L/P) _________-__-____-__----_
. psi PIFRAC 2 continued (47) T HERMO--M CCll AN IC AI. PROCCGUING HISTOR1 (40) ASTM GRAIN SIZC -_
(49) 7. IERR1TC FOR STAINLEGG GTECL ____
(50) LOCATION TilRU Till CKNEGG _-__ ___ ___
(51) HARDNEGG/GCALE _____________-_ -__-_
__----___6_____ ____-
__-____h___m___ ___--
EXTRA 1 _----_-___-________-
2 __-___-____-________
TRA 3
+XTRA XIRA 4 PTP)
l
~
O' o o .
V EC- AN CS
~
3 3 NG TRAC ~~ RE j
I
, SOURCES DATA BASE l~ BCL DATA T(PES MATERIALS ARGONNE > J-R Curve Carbon l
= M EA NSRDC A Charpy-V Stainless Welds EPRI -
Tensile Chemistry Supplem. ental VENDORS j
Testing 3 i
USERS -
l j- 1f
~
vExooas \RC i
i I
O O O TENSILE PROPERTIES OF AGED CAST STAINLESS i
25*C 290 C AGING COND. YIELD ULTIMATE YlELD ULTIMATE I (17.1%)
UNAGED (251.0) (571.5) (168.8) (399.6) 350/10000 +19% +11% +11% +5%
P1(24.1 %)
L UNAGED (247.0) (582.5) (157.5) (434.0) 350/10000 +14% +9% +13% +5%
- 400/10000 +16% +16% +3% +12%
C UNAGED (243.5) (582.5) (153.5) - (423.5) 350/10000 +11% +5% +17% +7%
) 400/10000 +17% +13% +9% +19%
P2 (15.6%)
l L UNAGED (212.5) (549.5) (1 41.0) (396.0) i
, 350/3000 +17% +8% +14% +5%
- 350/10000 +25% +11% +9% +6%
400/10000 +11% +12% +4% +9%
C UNAGED (230.0) (548.6) (159.5) (397.0) 350/3000 +10% +10?j -4% +1%
~
l 350/10000 -- --
. -2% +1%
350/10000 +2% +10% +1% +13%
I DELTA a (In.)
l 0 .100 .200 .300 .400 10000 i i i i I 50e00 RGED CRST STRINLESS STEEL 800e -
(25 C L.-C 20% 637.8 MPa) 40002 6000 - ! CD txp 5 5 / / oo ?
f$
0 h / [B o
/ 0
- 4020 -
% o
/ o O MATERIAL CODE P2T -
20000 -
!co 0 0 UNAGED 2000 - A RGED 3000 PRS. t 350 CC o8 / 0 AGED 10000 HRS. t 350 SC
~
IEEUU
@g / o AGED 10000 HRS. t 400 OC g / i i .
0 0 2.5 5.0 7.5 10.0 12.5 DELTA a (mm)
O DELTA a (In.)
120 .300 .300 .400 10200 ' i i ,
I
- RGED CAST STRINLESS STEEL 50000 8000 -
(288
- C L-C-20'4 SG 689.5 MPa)
MATERIAL CODE P2T i,
40000 O UNRGED g
6000 -
A AGED 3000 HRS. t 350 8C n AGED 10000 HRS. t 350 OC
(.q i
2 / / o AGED 10000 HOS. t 400 8C 30000 ;g
" 4000 ! ! N b*
$ @ * =
/ / to s% co 0 o - 20000 *,
gg 7
! 2000 -
!g 0 o 0
0 10000 kb
- /
0
! ' i i i l 0 2.5 5.0 7.5 10.0 12.$
! DELTA a (mm) 6 g
i i DELTA a (In.)
! 0 .190 .289 .300 .400 .500
- 7500 i i i i i RGED CRST STRINLESS STEEL -
40see
]C 4
6000 -
(25'C, L-C, 29% SG, 689.5 fra) o 4
O O g -
30000
) . O o g 4500 -
/
/ O O I a / / o D s
a ~
D _MATERIfL C0K PIT c' i
" 3000 - / O f D
~O UNAGED C o
j
[o O
O/ D D AGED 10000 FRS. t 350*C 0 AGED 10000 FRS. t 400'C I
O 7o -
10000
} 1500 o0 0 0 a /
@ 0 l
0 0 0mo 0 l 0 0o i ddP
- g / i i i i i a
4 0 2.5 5.0 7.5 10.0 12.5 15.0
} DELTA a (mm) 1 1
1 I
i DELTA a (In.)
, 0 .100 .220 .300 .400 .b20
- 7500 i , i i i i
1 -
40000 RGED CRST STRIllLESS STEEL 6000 -
(pgg oC L-C 20 % SG 689.5 MPa)
HATERIAL CODE P!T -
32000 O UNAGED 1
O AGED 10000 HRS. t 350 OC C 4500 -
g / / 0 AGED 10000 HRS. t 400 eC - 24000 l l O h O O 3000 - I O O a M O 00 u
/ / o o 0 *
! / / 00 ,
= o 0 1500 - C o0 0
/ g m0 o -
6000 j doh j
0 0
1 0 2 '.5 5.0 7.5 10.0 12.5 15.0 i DELTA a (re) l 1
1
-. . _ _ - _ - - - . _ . - - - _ - . . _ . _ - _ ~ _ _ . . . . - ._ -_ .. - - - -
O' O O -
~_.-
TELTR a (In.)
.18 .28 .30 .48 5898 i i j i
! P2TILC -
2sse8 R451CF3, 20% SG, 22 C /
4ae8 -
l 14% DELTR FERRITE /
twee
/
3* ~
/ -
3
/ / -
tsel- -
2008 -
u
/ / k
/ / -
isseB"
." / /
I* ~
f f - sogg
/ / .
g:. !! I I i_ i 8 2.8 4.8 S.I' 8.8 18.0 12.8 IELTA a (m) a
. _ . ~-
j ElTR a (In.)
8 .82 .94 .96 .58 .1 .12 .14 .18 .10 388 i i i e a i i i i
- - 16888 A.
EI* ~
OA O l -
148N a -
12 M 2M -
Op l Oe A EPW-1,2,3 (2R S.C.1 tum **d g -
0 I N T.O (R 5.C.) y 1
4 d
ism - 4[a 2 <
/ - gm e
e, i
I
~
/ 5 ggg .
l -
6810 b/ / . ,,,,
m . / /
l l -
20a8 g /t i / i a i e i t g 0 .5 1 1.5 2 2.5 3 3.: 4 4.5 5 l
ELTR (as)
EFFECT OF SIDE GROOVES I
DufR a(In.)
8 .85 .18 .15 .28 .25 sees . . i e i RISI TYPE 304 STAINLESS STEEL _
i3 .
24ea -
/ a2,* *t, I eA
- te m
,ises -
/ ,% '., " , a j.
jo: "."
, j 1 2e ,
sm; 4
M "
l
'tzea -
/ g "k "/o. " . y sees 3
/ y a "/e ASTM E399 (SR) .394T-CT WELD f , ;. " f a - 0% Stde Groove (2) see - ;o / o - 10% Side Groove (2) . jees i
/ n - 20% Stde Groove (2) i g, i/ ' ' '
s
! e 1.5 3.8 4.5 8.8 7.5 cttTR e feel i
g l
l ETA e Itn.)
8 .82 .84 .M .88 .18
>= , . . a. a N !$80
,AIES Cr. B Pige, a a , a 1
248 . 19-en Thtch, 23'C, 284 SG a a af A aaa a
f
- "0 -
1208 r , % v 198 -
- o y "g sA , a
- 4g f
[ 64 o o Do,fl * -
f 0 "
f asi o
o, 0, ##)
128 - l a*'a o o, ,, -
aa o ,o , ,
,, a O_,8 o
- 8 + A - 8.394f-CT
,e -
40s 88 ~ g v- IT-CT, Flattened only e- IT-CT, Flettmd and SRR e- IT-CT, Flatted only o- 21-CT, Flatted and 9tR g
g si! i i t! e t g 8 .5 1.8 1.5 2.8 2.5 E TR e (mm) i i
f
! EFF ECT OF PIPE FLRTTENING
!O
.I E TA e (In.)
0 .18 .28 .38 .48 .!8 3522 3 , , i 22228 v' o
'8 2tc8 -
f ,
' o, * - 1s22a
^] s o o +
/ va n 0
! _2ica f '/ , ,' o "
ime;.
l
~
A e
/.4,/., e,o' ..o+ 'a *
- 5 1
1428 - ( + R 376-324 Staintes:~PfEe -
em "
- 9 3~A~Ni~i?171 W ,"fft $f ~~
A - 8.394T-CT 728 - / ' - IT-CT, Fletteced Only n .- It.CT, Flettened Onlyand SPR
+ - fi-Cf. Flattened Only
- 4220
/
/ 0 " 21-CT, Flettened end son
,e / , ., , , ,
8 2.5 5.8 7.5 18.8 12.5 ETA e (ne) n
! O.
- e
l
! O O O i
l !
i i
i TEMPERATURE ( F) 1 50 150 250 350 450 550 650 750 250 , i i , , i i jl A B6 Gr. C 3 ice ISO 200 -
1 1
{ G -
120 i
~150 - d
~ ,
4 G !
g b i cr U :
>- O -
80 i & -
j @
o 100 S !
i a E o
i j 1
50 -
40 l
A A - Inside Diameter Filled Symbols are L-R am - mid-Thickness !'
Open Symbols are C-R o e -- Outside Diameter
! 0 -
0 0 50 100 150 200 250 300 350 400 j TEMPERfiTURE ( C) i i
I i i I
i
i 2
a O 2880 9 .18
.20 BELTR e (te.)
.30 48 g , ,
343'c a 20*C y pl21'C a * - 10000 l ,a 1688 -- !* g, '
, ' g 2ge C e 0 0 fa v v,, e o a '.
0 & , hy 4 o 7580 ,
1288 - e O A vy
- e a q t
/ A y ' , e a 5
$ 5 , o D 232'C I
- g e 4a A 4 $ D D o a -
5m 5 800 - 4 ,o -
eg a "
,fg 4 ,A!06 Gr. C Pige (L-R) 4gg . IT-CT, 281 SG -
2500 8
t A - 28 C a - 232'C 0 - 343'C v - 121'C + - 268'C g / t L- ' L. i g
8 2.5 5.8 7.5 18.8 12.5 E LTR e (me) l 4
- O l
l 4
- EllR a (fa.)
8 .B2 .e4 ,gG
$g .08 .lg
- i 4 .
.!!W M ED , ,
saa - e.57-ci, let sc /,
F!tLfD SYMfmLS flPC ImnnintrD a / '
i a / , 20PO
! A s
,383 -
, a ,
T
.].a
~ a y g / . .- / o i5 ;.
1 -
/ .
o ./
1
' O O
g OO
- 1900 "
f.a, ,
^
. a
.o S e OU O III
'a ,' "(DO e e - 29u C (81W-121) -
o - 288* C (Elu-lee,it5) 500
, a - 75* C (81W-le2)
,lr / ,
v = lil' C ($1W-143)
, ,/ , ,
8 .5 f.8
. 1.5 2.0 2.5 KLTR a N )
\
- I i
i
'T T[w05mE ('F)
('v;, 100 i
8 lab t
2ag 3eg 480 T
5m i
ses i
?se i
3me 580 -
A105 Gr. C PI E C_
~
li-CT, 28% SG 4M ~
a-N A - C-R - tage R
) 3PJ!
=
?
} -
1500 '-
o 4 / O
% f ism -
n
- o f
jeg .-
~
b A
.-* $00
.g_ ' L '
B
! p,a
! e igg PRO 388 4M TDfTmitM (*C)
O>
s v
ttmenrupt i en reo 288 - aco <ce rea saa see a
_H_S_S_I W_e _l d__ G_ l.W__
_ icea
!!8 . o IILLCD S"80LS v M IRvnDIATED
.. gea
~ ,~
~,, *
> ics i e.
ggpg . 121 %e n -
- Ec0 h
? $
8 _ ,oa 5 I e sa - * '
n e . 41.Cf a a a .. g,st.cf e v .. e.g t Cf 73, 0 e .. g,$t.cf 8 '
- 9
.J -
g 183 100 30c 400 f[t1TRAftFC f a c) ,
i o
O~ O O Ta b t ., le-1 A. .r up 11 Tensi.I<* Resutts (550*F)
ID/0D 0 . 2 % Y i e l.1 T ggg n u E Fracture c 0.505 0.226 0.505 0.226 0.505 0.226 0.505 0.226 0.505 0.226 0.505 0.226 Diam Dian Dian nian Dia, Diam D i .m Diam Dian Diam Diam Diam (ksi) (ksi) (106 psi) (ksi) i;A 182-304 00 18.69 17.48 56.82 57.15 2.429 2.306 1.902 3.947 21.96 21.06 127.3 142.8 1/2n 18.94 19.64 56.98 57.64 2.616 2.633 1.660 3.813 26.89 26.36 141.4 134.9 ID 18.47 20.09 56.49 57.92 2.270 2.632 4.413 3.923_ _
27.12 26.88 125.2 143.0 u
Average 13.70 19.07 56.77 57.57 2.438 2.124 3.992 3.894 25.32 24.77 131.3 140.2 i
l A 106-C on 48,85 53.20 92.16 94.15 4.782 5.41.5 0.8001 0.6720 27.83 23.72 161.9 166.3 1/4D -----
43.20 - - - - - -
93.05 - - -
4.809 ------ 0.5486 ----
28.21 -----
149.8 1/20 41.75 41.10 83.29 98.60 4.730 4.647 0.5365 0.5216 27.86 27.16 143.0 150.5 3/4D - - - - -
45.75 - - - - -
91.87 -----
4.718 ------ 0.6854 - - - - - -
28.18 -----
151.1 ID 51.33 55.87 92.81 91.99. 5.095 5.773 0.7905 0.6468 27.51 27.07 161.3 169.3 Average 47.31 47.82 91.09 92.13 4.869 5.072 0.7090 0.6149 27.73 26.87 155.4 157.4 l
f
~
p\
~/
1000 ' ' i i R1 6-C_
(PIPE 32 in. 0.D. x 3.25 in. WALL) 125 800 -
100 2 6 U ="
- n. 500 -
m 5 lLTIMATE -
75 W
w
. $ 400 -
a g $
m m g A -
50 m
D A 0.2% YIELD 200 -
A 0.505 in. -
25 0 0.22G in.
i i i i 1 p' 0 03 1/4 1/2 3/4 ID THICKNESS LOCRTION A
V 800 , , , , ,
5R 182-304_. _ go (PIPE 20 in. 0.D. SCIIEDIR.E 100) 500 -
70 ULTIMATE 60 400 -
x E 2 ,.,,
~
r[
L -
50 .x m 300 -
m E]
40 gj Bi W.
200
~
0
- 0.2% YIELD g W j -
20 100 -
A 0.505 in, a 0.226 in. -
10 g i i i i i g CD 1/4 1/2 3/4 TD THICKNESS 10 CATION O
O l
!O i
i STATUS /
SUMMARY
I l
i e FSAR SURVEY COMPLETE
{ e DATA BASE FORMAT ESTABLISHED ON- COMPUTER--
i
)Q e QUERY SOFTWARE NEARLY COMPLETE e DATA COLLECTION, MATERIAL PROCUREMENT, j AND TESTING ARE CONTINUING I l I
4 1
I l
l 1
lO i '
i