ML20202F003
| ML20202F003 | |
| Person / Time | |
|---|---|
| Issue date: | 07/02/1986 |
| From: | Advisory Committee on Reactor Safeguards |
| To: | |
| References | |
| ACRS-T-1530, NUDOCS 8607150060 | |
| Download: ML20202F003 (514) | |
Text
.
ORENAL
&f$1=/530 0
UNITED STATES NUCLEAR REGULATORY COMMISSION IN THE MATTER OF:
DOCKET NO:
ADVISORY COMMITTEE ON REACTOR SAFEGUARDS SUBCOMMITTEE ON METAL COMPONENTS O. .
LOCATION: COLUMBUS, OHIO PAGES: 210-485 DATE: WEDNESDAY, JULY 2, 1986 "p 7* r 3-- -
,,s m-
?,;
g ; .. - .
1 Jo Not Rerr. ave imm ACRS Office Q ACE-FEDERAL REPORTERS, INC.
OfficialReporters 444 North CapitolStreet Washington, D.C. 20001 e607150050 860702 (202) 347-3700 PDR ACRS T-1530 PDP NATIONWIDE COVERAGE
4 PUBLIC NOTICE BY THE UNITED STATES NUCLEAR REGULATORY COMMISSIONERS' ADVISORY COMMITTEE ON REACTOR SAFEGUARDS WEDNESDAY, JULY 2, 1986 l -
The contents of this stenographic transcript of the proceedings of the United States Nuclear Regulatory Commission's Advisory Committee on Reactor Safeguards (ACRS), as reported herein, is an uncorrected record of j the discussions recorded at the meeting held on the above 1
date.
4 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.
l l
c 4
)
- . . ~ . , - . _ ._
4 W .A -
210 1 .
O 2 l
3
~
l 4 MEETING OF THE ACRS METAL COMPONENTS SUBCOMMITTEE 5
I 6
i 1
7,
- 8 BATTELLE COLUMBUS DIVISION l
9 505 KING AVENUE 10 COLUMBUS, O HIO 11 i
15 13, i
14 f
15 I
16 17 18 JULY 2, 1986
. i 19 20 1 2L ;
22 ,
23 d 24 25
211 1, Presents
() 2, l
3 P.G. Shewmon, ACRS i
4' Harold Etherington, ACRS 5 W. Kerr, ACRS 4
6 Al Igne, ACRS 7 M. Bender, ACRS Consultant I
8 E. Rodabaugh, ACRS Consultant 9; J. Hutchinson, ACRS Consultant i
10 i
11:
i 12 Gery Wilkowski, Battelle
() 13 Mike Mayfield, NRC Research 14{ Guy Arlotto, NRC Research 15 16' l
17 18 i
1 9.
I 20 21 22 23
{) 24 25
l I 212 1 INDEX 2 Subject -
Speaker Page No.
3 International Piping Integrity Research Group Program -
(B. Saffell) 213 4
International Piping Integrity 5 Research Group Program -
(G. Wilkowski) 221 6 International Piping Integrity Research Group Program -
(B. Saffell) 261 l
7 Assessment of Leak Detection for 8 Nuclear Reactors -
(D. Kupperman) 266 9 Investigations of the Mechanisms of Thermal Aging of Cast Stainless 10 Steels -
(O. Chopra) 305 11 Environmentally Assisted Cracking of Stainless Steels in BWR 12 Environments -
(W. Shack) 305 O 13 l and Applications -
Oriented Fatigue Fatigue Crack Growth Studies 14 in LWR Materials - ( W . II . Cullen) 431 15 l
16 17 18 19 20 21 i
22 i
23 24
(])
25
i i
j 213 MR. S HEWMON : This is a continuation of
()
4 1
2 the meeting yesterday, so I don't have to read you
) 3 all your rights and privileges, I'm told. What
}
4 we're going to hear first about is the International 4
l l
5 Piping Integrity Research Group. Seems to me it was I 6 I in for some catchy phrase for it next year. But, l
7 Gery, are you doing this?
8 MR. WILKOWSKI: We're going to split it up.
1 9 MR. SHEWMON: Fine. Okay.
10 MR. SAFFELL: My name is Bernie Saffell,
\
) 11 and I'll be discussing the International Piping 12 Integrity Research Group, program. This is a program j
j
() 13 which will consist of the NRC and a consortium of l 14 foreign countries to participate in performing i
1 15 research aimed at enhancing the structural or not i
l 16 enhancing but quantifying, I guess, further the i
! 17 structural integrity of piping.
I l 18 In terms of what is IPIRG, what will be t
j 19 IPIRG, which is the acronym used for this group, it i
20 is an international program managed by the NRC for j 21 the performance of complex piping experiments. It's l 22 intended to provide a forum for reaching l
- 23 international consensus on new pipe break rules, f
{} 24 specifically replacement for the double-ended break l 25 criterion.
t i l f
214 7~ l Now, this is the near term. Obviously,
(_/
2 this group could address other issues should they be 3 i of interest to one country, vis-a-vis the United 4l States, with the agreement of the group as a whole.
5i The near term work will be performed by l
6!Battelle, that is, the initial experimental program 7 which Gery will discuss in detail, and this group is 8 composed of both industrial and regulatory 9forganizations from countries with nuclear power 10 1 programs, and the agreement is that the research 11 1 aspects of the program will commence when there are 12 i five participants in addition to the Nuclear
) 13lRegulatoryCommission.
14 ! The general objective of the group is to 15 develop, improve and verify engineering methods for 16l evaluating the structural integrity and performance 17fof nuclear power plant piping containing defects.
I 18 The emphasis here I would say is on the " verify,"
19!that is, we ate going to be developing data with 20 which to assess further the existing techniques for 21 evaluating piping structural integrity. And we see 22 the results of this program affecting a number of 23 aspects related to plant operation, design criteria 24 and the analysis techniques used to design these
(])
25 plants.
4
, 215 1 The specific IPIRG objectives -- and these,
() 2 as you'll see in the subsequent view graph, are 3 essentially translated into tasks -- are first to 4 develop an understanding of the response of high 5 energy, flawed piping systems to dynamic loading.
l 6 And this addresses the issue you raised yesterday, 7
7 Mr. Bender, where you expressed an interest in the i '
8 ; dynamic response of flawed piping'to dynamic loads, 9 vis-a-vis the Degraded Piping . Program, which has !
10 addressed the static or quasi-static loading to date.
11 ,
This is the next step.
]
12 And another contrast between this and the 13 Degraded Piping Program is that we do intend to test j 14 at least one representative piping system, flawed, 1
I
- 15 and the loads that will be applied will be typical f
16 of loads that a plant might see, such as a seismic 17 event or those associated with relief valve loads.
f I 18 We don't intend to duplicate any particular j 19 earthquake, but the frequency content, for example, :
i 20 of the loading we would intend to apply would be '
j i 21 characteristic of seismic or relief valve.
1 22 MR. S !!EWMON : What's the relief valve load?
- 23 MR. SAFFELL
- When the relief valve opens i
i 24 it causes a fluid transient, which can obviously
[])
! 25 shake then the piping system, applies a dynamic load i
L..__...__...._._...
4 i
216 1 to the piping system.
O 2 MR. S HEWMON : I would think when i t 3 stopped it would apply even a larger one. Is that I
J 4 also part of it or not?
5i MR. SAFFELL: That is a possibility. The
, l 6! load I'm thinking of is when the valve opens you i
l 7 then in many cases sometimes have a water slug which
! 8!then can impart fairly large loads at bends or area l
j 9 reductions or expansions.
j 10 Yes, sir, when the valve slams shut that
! 11 can also impart a load. I know a concern a few i !
! 12 years ago was related to sudden closing of check j
( 13 valves, both fron a valve damage standpoint as well j 14 as from the integrity of the remainder of the piping ;
4 i j 15 because of fluid transient set-up, yes, sir.
l 16 MR. BENDER: Could I ask a couple 17 questions about that. First, with regard to the 1
l 18 i seismic business, the point I made yesterday, and f 19 I ' .1 just repeat it- again, is that what you have to ,
t
! 20 worry about depends a lot on.what the initial i
i 21 scismic design criteria are. And so i f the seismic 22 loadings are modest, it may not make any difference, l 23 but when you get to the point where the seismic i
{} 24 loadings are a quarter or maybe a half or maybe 75
- 25 percent of the total pipe loading, then you need to 1
i.
'p7T-,,==rwwe-w w 1r v s--i-p',a.t=yiegiw-fm- 'z+w%e-
. - p-w M v 4=nwrm h7 -- o* ep e - yq- ,, e,- w t gp i rw + e y eav-**nw.,m+i--ee +wg ,M w m pp *w-pws w v y-eargyet~-e-----*Niterg g g-r e- + & m ? r -v +,etwee "'n w g e 1-'
l 217 1 know a lot more about them.
O 2 And you really don't have much of a story 3 lright now on how the change in -- how the change in 4 proportion of dynamic to static loadings affects the 5 piping integrity issue. There's a lot of arm waving, 6 Ibut there's not much quantitative value.
7 So it makes a lot of difference in this 8 research program as to how you go about assessing 9: the meaning of the thing, and I think the picture in i
I 10 the past has been to go from one extreme to the i
11 other, either to look at the worst events or to look 12!at the least important events, and it's in between
() 13fthat are the places where the controversy arise.
14 f With regard to the relief valve loading 15 thing, which Paul asked about, this is a place where 16, water hammer may be of most significance because of 17 just the point you made there, the water slugging.
18 But the mounting of the relief valve to accept the 19 end loads from the sudden change in forces 1
20 associated with opening the valve han been the thing i l
21 that's contributed most to pipe failures. There's I 22 several really cases where the valve has blown off l
23 the pipe because of those forces, and so l
{} 24 understanding what the loadings are out there on the 25 end when the valve opens has been an important i
218 1 engineering issue. And I think when people look at 2 pipe failure statistics, that's often the kind that 3 dominate the statistics. So being able to show that 4 the engineering practice has reduced the probability 5 of those things contributing to pipe failures I 6 think would make a difference if we eventually go to 7 a problem listing concept.
8 1 MR. SAFFELL: I appreciate your comments.
9 MR. S HEWMON : Let me stop and interrupt 10l things for a minute. I have a lot of difficulty 11 understanding how you can develop much force if you 12 have a pressure operating on a column of water and 13 ! all the sudden remove the pressure from one end so 14 that you accelerate this. It's a lot easier if i
15 somebody slams the gate and all the sudden puts that 16 inertial pressure built up on something. Can you 17 tell me --
18 MR. ETHERINGTON: The pressure wave goes 19 out and then comes back.
20 MR. S HEWMON : From the start?
21 MR. ETHERINGTON: Yeah.
t 22 MR. SHEWMON: But it's initially at rest?
23 It's not initially at rest.
24 MR. ETHERINGTON: It's initially at rest,
(])
25 yes.
-- . -- . _ _ . . ~ - - . - - -
219 1 MR. S HEWMON : Then there's no -- well, we O 2 can talk about it later.
3 MR. BENDER: Well, it's always been at 4 rest, but you don't know --
many times people don't 5 recognize the directional forces, where they're 6 coming from. The valve is usually a right angle 7 opening.
8 MR. S HEWMON : Yes.
9 MR. BENDER: So you get a big change in 10 direction of the force right where it spurts out of 11lthe valve, and that by itself changes the loading on 12 the pipe when you go f rom a static system to a 13 dynamic system.
14 MR. SHEWMON: What it's spurting into is 15 necessarily free of water or possibly free of fluid?
16 MR. BENDER: It's open air for the most 17 part.
18 MR. S HEWMON : Okay.
19 MR. ETHERINOTON: I thought we were 20 talking steam. If you're talking water, then, of i 21 . course, you can't get much --
l 1
22 MR. BENDER: It's steam we're talking j 23 about.
24 MR. S HEWMON : Okay. That helps. Thank
{])
25 you.
220 1 MR. SAFFELL: Okay. In addition, we will l O 2 be contributing to the MEA -- the data base that MEA 3 is developing for the NRC in terms of providing 4I information based on the material properties of the l
5' piping we test, aird we will also be documenting the 6lresults of our pipe fracture experiments in a form f
7, amenable for use by the members of the IPIRG.
i 8! Finally --
and you'll hear more about each 9j of these in detail later on during the morning --
a l
10 task initiated by the NRC as part of this program 11 addresses the issue of leak rate estimation models i l
. 12 and the ability to predict leakage rate, including 13 ! the effects of plugging.
I !
14 l And finally, there's a task where one of I
15l the objectives is to coordinate the results of the l
16 j program obviously and disseminate that information 17lto i
the members, and also through the program to 18 : address any regulatory and/or technical issues which 19i exist and may arise in the future.
i 20 ' So the remainder of the morning will in 21 essence be devoted to a discussion of the --
not the 22 remainder of the morning, the remainder of this hour 23 will address the discussion of the base program Q 24 tasks.
25 Task 1, and this is the major effort, is
221 7
1 the experimental program, and Gery Wilkowski in a 2 second will discuss that. It talks about the 3 experimental efforts. We will address the 4 analytical efforts which are comprised of pretest 5 ! design and post test analysis, material 6 characterization, the data base efforts, leak rate 7 estimation models, and then we'll say a few words 8 about the deliverables we expect to see out of this 9 I program and the information exchange seminars.
10 So with that, I'll turn it over to Gery.
11 MR. WILKOWSKI Within Task 1 I'll first 12 be discussing the experimental efforts and then
() 13,afterwards sumnarizing the analysis and finally the 14 l materials characterization efforts. I'm sure that 15lwhen I discuss some of the experimental efforts, 16 you're going to ask me some questions about the 17 analysis, and I probably should have blended them 18 together, but this is the quickest way to try to go 19 through this presentation.
20 Within Task 1 we've divided it into three l 21 subtasks. The first subtask essentially looks at 22 what happens to flawed piping systems under purely 23 inertial loading or a flawed pipe under purely 24 inertial loading.
{}
25 The second one is to look at what happens
. _ . . _ _ _ _ _ _ _ _ _ - . _ _ _ _ . ~ . ~ _ . - _ _ _ _ _ . _ _ . . _ .
222 i
i 1 to a crack in a pipe under dynamic displacement 2 controlled conditions,.in that inertial loads are a -
- j. 3 time dependent load control type of load, whereas a i 4 dynamic displacement control load is quite a bit i
j 5 different animal in regards to the crack stability
{ 6 behavior.
7 The last subtask, Subtask 1.3, then looks i
} 8 at a representative piping system that would have a i
9 combination of normal static loads, dynamic 10 displacement control loads as well as inertial loads.
{
- i 11 So what we're hoping from the first two subtasks is 12 we do that on some smaller diameter pipe under 13 relatively simple conditions to try to gain an 14 understanding of what's happening under the dynamic 15 loading condition.
j 16 For this first subtask the objective of
! 17 the subtask, as I said, is to perform experiments to i
^
18 assess analysis methodologies for circumferential1y i
i 19 cracked pipe under inertial loadings, and really in 4
i i 20 this program right now what we're going to be doing ;
21 is assessing current analysis methodologies.
22 For instance, if you do a code of analysis, J
t 23 you might do a modal analysis on a pipe system that i !
{} 24 is elastic. You then would predict some peak 25 stress. You take that peak stress and then you 1
l 1
223 I would apply that peak stress to elastic-plastic i (3
%J 2 fracture mechanics type of analysis, and there is 3 perhaps some discontinuity between reality and that 4 type of analysis as Everett referred to yesterday, 5;in that there are some drastic nonlinearities that 6 'are occurring in the experiments, but the analysis 7 is essentially linear elastic to determine the 8 stresses.
(
9l MR. ETHERINGTON: I have a difficulty 10l understanding what's meant by inertial loading. Is 11 that fluid inertia or --
12 MR. WILKOWSKI: No, this would be inertia O
(_/ 13;from the movement of the pipe.
14 , MR. ETHERINGTON: Okay.
i 15 l MR. RODABAUGH: I think it would help if I
16 you indicate how you're going to put on this 17 inertial loading.
18 MR. WILKOWSKI: I have a sketch of that.
19l One of points that I wanted to make is that there is 20 a difference betwoon displacement control and load 21 control loading as far as the crack stability. For 22 instance, in this graph here or view graph we see 23 two different test records for conter cracked
(~) 24 stainless steel pipe experiments that were conducted
%)
25 under displacement control versus load control and
224
- 1 loading.
2 You can see in the displacement control 3 cases that we could unload, you know, 50 cycles and 4 essentially come back to the same load displacement i
5l record. There wasn't much of a history effect.
1 6i Whereas under a load control case, once you're past I
7' the maximum load, then there's a very big difference 8 in the stability of the crack. You get a very large 9 amount of crack growth that occurs with very few 10 l cycles.
11 The experiments that are planned will 12 I involve both carbon steel and stainless steel piping, 13 six inch diameter. The carbon steel piping is a 14 very low toughness material, so that it will exhibit !
15 l the contained plasticity that we would see, for l
16 I instance, in a very large diameter pipe.
17 l We have six experiments planned. In a l l 18! minute I'll show you the test set-up for it. And I l 19 these involve through-wall cracks and surface cracks, !
20 internal surface cracks in the pipe. The objectives 21 of these are to assess what happens when the cyclic 22 loading occurs under mainly inertial loading, and 23 I'll show a little bit about some of the possible 24 effects in a little bit later on in another view
[]) --
25 graph.
225 7-1 The experimental set-up, for instance, is
\_/
2 lshown schematically here. Here we have a 3 l cantilevered length of pipe. There is a crack at 4 ! some fixed distance away from where it's being 5 supported, so that there aren't end effects. This 6lmight be, for instance, a distance of four diameters, 7lsome relatively safe distance. And then the piping 8 l system would be pressurized with water at LWR 9 conditions, and then excited so that the pipe wants 10 to whip at this end and cause the inertial loading 11 ' to occur.
12 The pipe will be instrumented to record 13 , the dynamic loads, the displacements as well as the 1
14 crack growth, which is a very difficult type of 15 i instrumentation process.
16 , MR. RODABAUGH So your computer control i
' l 17l actuator will be putting on, say, a time history l
18 more or less representative of some earthquakes?
19lThe opposite, of course, is to put on --
20 MR. WILKOWSKI: Yes, more or less 21 representative, in that the frequencies would be 22 within the typical frequency range that you would 23 expect to see the large amplitude loads from an 24 earthquake.
[}
25 MR. RODABAUGH: Of course, the other
226 1 possibility being to put on a sine sort of input and 2 what makes me wonder is unless your pipe length, 3 diameter, weights are just right, you'll be way out the frequency in most earthquakes or into it 4lof 5 MR.. WILKOWSKI ,
That is a problem. We 6fhave a fair amount of --
3 7 MR. RODABAUGH Let me put it this way.
l 8lAre you thinking of tuning the system to come 9 frequency in the earthquake range?
a I
In order to tune it, what 10i MR. WILKOWSKI:
l 11 happens is as you get plasticity, the stiffness of j 12 the system will change. The more plasticity you get, i
l 13lthe more the stiffness will change. So we have to I 14 do a fair amount of pretest nonlinear analysis 15 design in order to help select the amplitude that we 16 apply as well as to determine what frequency to 17 apply in order to get the experimental data that 18 we're interested in.
l I
19 In this case we're not interested in 20 simulating an earthquake event whatsoever. We want l l
21 to gain some knowledge about the fracture process. I 22 MR. RODABAUGH That's why I kind of 23 mentioned the sine wave input, which you could 24 control and follow your natural frequency. I'm only
(]) I 25 suggesting that you think of using sine waves rather
-,. _ r -- - ,-m. ,,, ,, -. - - , - , - - , r --r--, .-
4 1
227 1 than earthquake input.
()
1 2 MR. WILKOWSKI: We've considered three 3 different alternatives to the loading history here.
4 One is to keep the amplitude constant, do a sine 4
5 , wave, but gradually change the frequency of the sine l
6' wave, starting off with something that's relatively 7 'high and then decreasing the frequency as more 8 plasticity occurs. ,
9{ Another way would be to increase the i
10l amplitude gradually with the same frequency, or you l
11 can change both the frequency and the amplitude to 12 account for the change in the stiffness of the 13 piping system as the plasticity occurs.
! 14 l MR. BENDER: I'm not as clear as I would 15 like to be about the cyclic loadings or the logic
- 16 behind what you're going to put on. My perception 17lof the loaded condition that's of most concern is i
I 18 l one which involves the way in which the piping i
19 l system is mounted, and the concern really has to do 20 with how much displacement is associated with the 21 movement of the pipe and how that might add onto the 22 loadings that are already there. And I don't get 23 much of a feeling from the description here as to 24 whether that's the emphasis in this case.
(])
25 MR. WILKOWSKI: Yes. The emphasis here is a
-. -,, _ , - - - , . - _ , - - , . . - _ _ _ . - ~, -- , - - - - - - ---
228 1 really to look at just the inertial --
if you had a 7_
V 2 big piping system that was being shaken at some 3l anchor point, then up near the crack you would 4! experience some inertial loads. What we're trying l
5i to do here is to assess what happens when you have 6fjust the inertial load component with some fixed 7 pressure loading. As we go on further in other 8' subtasks, then we build onto that.
9' MR. BENDER: I'll accept your approach, 10; but I guess I'm not comfortable that that's an 11 ; important question. I think the more important 12l question is how much displacement you can get due to
()
13 ' these low frequency, relatively high amplitude 14fmovements of pipe that might compound the problen 15 associated with the steady stead conditions that you i
16! had tested previously. I'll just leave it there.
I 17 ' MR. WILKOWSKI: Yes. We have made this 18 task. This is a much smaller task compared to the 19I other tasks that is looking more towards the 20 question that you're interested in.
21 MR. SAFFELL: I think you'll see when Gery 22 gets to subtask 1.3 that --
I would like to think 23 that we're taking a step in addressing the concern 24 you have, in that we'll have a system anchored at a
(])
25 couple of points and shake it and be looking at the
4 I
229 1 displacement in between the anchor points.
2 MR. BENDER: There has been some study 3 work down out at Idaho. I never understood what it i
l 4 neant, but they started looking at some systems, and j
j 5 then the Germans have done some too. I would think I
6 you could take those into account.
. 7l MR. WILKOWSKI: We are also going to try 1
l J
coordinate the efforts within this program with
- 8fto i 9 {what's being done in the piping reliability program i
10 that's supported by NRC and EPRI.
l I
11 MR. ETHERINGTON: Does the literature show j l 12 much frequency dependence of fatigue life, I mean on
( 13 the standard laboratory specimens?
j 14 MR. WILKOWSKI: I'm not the fatigue expert.
l 15 Bill Cullen can -- he shakes his head yes.
l j 16 MR. ETHERINGTON: It does?
i 4
17 MR. WILKOWSKI: Yes.
j 18 MR. RODAB AUG H: Bill, I would say the 19 opposite.
I 20 MR. MAYFIELD: Let's try and be clear 21 where we're looking at the frequency effect. I 22 think your question pertains to during a seismic
- 23 event, and I suspect Bill's answer has more to do 24
{} with during the normal operation. In that case
- 25 there is indeed a very strong frequency effect. )
4 l
i )
)
. .-- _ . . . _ _ - . . . , .- . ._=.._ - . .. .. . . . . -
i ,
l 230 i
1 During the seismic event it's washed out to the j
k) 2 point where there really isn't much of a frequency a
i 3 effect.
4 .MR. WILKOWSKI: Yes, because of the time 1
5 duration in between there's no --
3 i
6 MR. ETHERINGTON: What is the normal
! 7 frequency effect? Does high frequency shorten the i 8 number of cycles to failure or what does it do?
9 MR. MAYFIELD: It's the other way. Well,
- 10 during normal operation the lower the frequency the 11 faster the crack growth.
I i
12 MR. ETHERINGTON: Yes.
13 HR. RODABAUGH But are you not --
14 MR. ETHERINGTON: What kind of frequency
- 15 range are we talking about?
16 MR. MAYFIELD: Normal operation?
I 17 MR. ETHERINGTON: No, I meant in-the --
18 MR. MAYFIELD: During the seismic event?
. 19 MR. ETHERINGTON: In the effect that's I
20 noticeable in normal testing.
I l 21 MR. MAYFIELD: Normal plant operation !
- 22 where we would see that --
)
4 l 23 MR. ETHERINGTON: No, no, I was getting-24 back to laboratory rotating beam. specimens, what
(])
f 25 range of frequency are we talking about when we're
. _ .- . _ _ _ _ _ . _ ~ , _ _ . _ _ _ _ _-. .- . _ - .
231 1 saying the low frequency gives you a shorter life?
2 MR. MAYFIELD: In the water environments
- 3 on the order of one cycle a minute.
4l MR. ETHERINGTON: One cycle a minute.
5l i MR. WILKOWSKI: Whereas in here we're 6 j talking about one to ten hertz type of loading.
~
7 MR. E THE R ING TON : So are we really talking 8 about an environmental effect mostly when we speak 9 of frequency dependence --
10 MR. MAYFIELD: Yes.
11 MR. ETHERINGTON: --
of fatigue life?
1 2 ;l MR. MAYFIELD: Yes.
() 13 MR. E THE RING TON : Okay.
14 MR. WILKOWSKI: Okay. In some of the 15 later experiments within this effort we'll also try 16 to measure the flow rate once that crack starts 17 opening from some of the surface crack tests. That 18 is, once that surface crack breaks through.
19 MR. KERR I'm sorry. You said you try to 20 measure the flow?
l 21 MR. WILKOWSKI: The flow rate through the 22 crack.
23 MR. KERR: The leak rate, in other words?
l 24 MR. WILKOWSKI: The leak rate. Okay. The
(])
, 25 next subtask, the objective of this one --
this is
l 232 1 also a relatively small subtarsk compared to Subtask 2 1.3, is to develop some basic data to assess what 3 happens to cracked pipe under relatively high cyclic 4l displacements.
5f The other point that I failed to mention 6 so far is that in the design of these experiments, 7 what we're trying to do is to design them so that 8' ductile tearing occurs within a few number of cycles, 9 that is, we're not growing the fatigue crack, so to 10 ; speak, a fatigue crack during the seismic event.
11 We're really at conditions where we're very close to 12! a failure occurring. So we're assessing 13 l elastic-plastic fracture behavior during a seismic 14lcvent and what the margins of safety might be during 15 that event.
i 16 MR. KERR: Are these experiments in which 17 you expect to know what the results will be or are 18 l you flying blind? l 19 MR. WILKOWSKI: Some of them we have good I 20 guesses of what the results will be, and I'll show 21 you an example in this particular case. Within this 22 subtask there are three things that will be 23 investigated. For instance, one, we'll look at the l 24 effect of loading rate alone. For instance, in the 25 Degraded Piping Program we do all the experiments
233 1 with increasing displacement but at very slow rates.
2 We'll have one experiment here that we do at a 3 relatively rapid rate that would be a comparable 4 rate to a seismic type of loading, and that way we 5 can see what the increase of the flow behavior and 6 fracture toughness is on the material.
7 Assuming that that gives us what we expect
- 8 l to see, that we can predict that from the existing 9 J-estimation schemes, then what we'll be doing is 10 going on and investigating what happens under cyclic 11 ! effects where there is a fair amount of plastic 12 tearing that's occurring during each cyclic events.
() 13 It's not really fatigue crack growth, so to speak.
14 So what we see here is that some materials can be 15 severely affected if one is to calculate a 16 J-resistance curve, for instance, under such type of 17 cyclic loading conditions.
18 Finally, this will be done under i
19 quasi-static conditions. And then we'll be doing 20 this under dynamic conditions to see what the 21 interactions is. For instance, under dynamic 22 loading you might expect-in lots of materials the 23 load-carrying capacity will increase, but the cyclic 24 effect could decrease some of the load-carrying
(])
25 capacity on certain materials. So you could have l
234 1 compensating effects between the two.
2 MR. ETHERINGTON: Are all these in hot 3 water environment?
l 4 MR. WILKOWSKI: Yes, they are, all at 550 5 F.
6 MR. RODABAUGH: On item 2 has there been l l 7{any pipe test yet to show the --
8l i MR. WILKOWSKI: Not a pipe test. Just a 1 9 material test. And there's only one material test i 10 .that_I'm aware,of. This is some results on pressure i
11 vessel steel where under monotonic loading they 12 would get this type of J-resistance curve. However, 13 under different type of cyclic loading, with the 14 cyclic loading if they were to calculate a pseudo 15 resistance curve, although this type of approach 16 would have a theoretician throwing his hands up 17 because you're not supposed to do cyclic unloading L 18 in a J-resistance analysis, but if you were to do 19 that type of analysis you would predict that this 20 would be the change in the resistance curve for the 21 naterial.
22 MR. RODABAUGHz Thirteen cycles, what's 23 the -- you'mean you're --
are you running a fatigue 24 test on your
(])
25 MR. WILKOWSKI: They're doing cyclic
l l
235 l
r~w 1 unl ading as the ductile tearing is occurring. So, V
2 for instance, if you were to have seismic event and 3lyou would have a whole series of loadings and 4 unloadings that would occur, you could have some i
5 iductile tearing occurring during each one of these i
6 loading events.
7 MR. SHEWMON: See, 13 doesn't do anything 8 as far as they ran it, and 75 begins to break it.
9 jSo he's showing that low cycle fatigue will drive a 10' crack through a pipe.
11 l MR. RODABAUGH: The 75 cycles.
12 ) MR. SHEWMON: Yes.
13 MR. WILKOWSKI: But it's much lower than 14ione would predict if you just did the fatigue i
15 analysis and then applied the monotonic type of 16l resistance.
17 MR. SHEWMON: How do you do a fatigue 18 analysis --
what on there talks about a da/dn for l
19 driving cracks at 75, 80 cycles?
20 MR. WILKOWSKI: Well, this doesn't talk 21 about the fatigue aspects as one would normally 22 apply it in a linear elastic fashion. Okay? This 23 only talks about what's happening during the plastic fl 24 loading cycles where you're getting ductile tearing.
\s 25 In most cases it would be near the end of the
236 1 fatigue da/dn delta K type of plot where you 2 frequently ignore that data.
3j To see the significance of what this 4fchanges in the resistance curve, we did a few 5icalculations looking at what would be the loads for 6 the case of a 28 inch diameter pipe. This is a 7 typical steam line. About one inch wall thickness 8 with these types of Ramberg-Osgood coefficients for 9 the stress / strain curve and had a through-wall crack 10 , of this particular size.
11 We then looked at two different analysis 12 methods to predict the loads at crack initiation and 13 the maximum loads to see how significant, for 14 , instance, this one example would predict the change l
15 j in the load-carrying capacity. '
i 16 And what we see is that if you use the 17 cyclic resistance curve as opposed to the monotonic I
18l resistance curve that at initiation these two 19 methods predict that the initiation load might 20 decrease by about 30 percent.
21 If you look at the maximum load predicted 22 by these two methods, it would predict that this 23 cyclic loading effect if you accounted for it in the 24 resistance curve would decrease the load-carrying
(])
25 capacity somewhere between 72 and 88 percent
237 1 depending which analysis you would like to use.
2 MR. ETHERINGTON: Was this conducted with 3 a progressively increasing cyclic load; is that 4 right?
i 5 MR. WILKOWSKI: Yes.
6; MR. E T HE RI NG TON : How many cycles with 7 each step and how long?
8i MR. WILKOWSKI: Well, that was just using 9 lthe data from this resistance curve, okay? Assuming i
10lthat you're applying a similar type of cyclic l
11 loading history as they applied to that particular 12! laboratory specimen.
() 13 MR. RODABAUGH: So the square series of 14ipoints involve 75 cycles during that crack I
15' prediction.
16 MR. WILKOWSKI Um-hmm. And you would 17 really have to dig into that particular technical 18 paper to find out are those cyclic unloadings 19l typical of the large amplitude loadings that you l
20 I might look at for piping. Okay. There's a lot of 21 potential parameters that could affect this. How 22 severe are the unloadings that they used in that 23 type of work.
24 MR. RODAB AUG H z Why do you talk about
{~ )
25 unloadings? Don't they have to load it back up 1
238
- 1 again?
2 MR. WILKOWSKI: Well, it's unloading and 3 reloading.
4 MR. RODABAUGH: Okay. Thank you.
5i MR. WILKOWSKI: Okay. So this next view 6i graph then shows the experimental test matrix that's i
7 l; involved here. As I said, the first experiment, 8j that would be just a dynamic monotonic loading to what the rate effects would be on a stainless 9fsee 10 pipe that's relatively high toughness, and we 11 already have from the Degraded Piping Program 12 similar data on that same pipe at low rates. So we 13 can see what just the rate effect is by itself.
i 14 i The next four experiments here look at 15 what happens when you do different type of cyclic 16 loading on two different materials. One is a 17 stainless steel which is very tough, and we suspect 18 in the stainless steel that there's going to be 19 almost no effect, whereas in the carbon steel we i
20 expect that there will be effect because it's a much 21 lower toughness material.
22 MR. I!U TC HINSON : These are going to be 23 laboratory specimen tests?
24
[]) MR. WILKOWSKI: These are pipe tests, six 25 inch diameter pipe, Schedule 120. This is the wall
_. - _ - - _-. _. ~.
239 4
1 thickness. They will be done at 550 fahrenheit.
(2) j 2 MR. HU TC HINSON : Are you going to do such j 3 tests on laboratory specimens as well?
4 MR. WILKOWSKI: The plan right now is to 5 first look at the data in the literature before we 6 ido any laboratory specimen testing, because there
- 7jcan be such a wide variety of different parameters
. 8 ithat you can apply in the tests, and we would also
- 9l like to look very closely at some of the data as how i !
. 10 lthe pipes unload and reload and how much --
when the i
3 11 crack reinitiatea. From the potential drop method I
12lyou can see when the crack reinitiation occurs after
( 13 an unloading event.
14 i MR. HU TC HI NSON : The one thing I think you l 15 ought to be on your toes about, you've identified i
l 16 this strain-aging effect --
i 17 MR. WILKOWSKI: Yes.
which obviously is 19 associated with the temperature range that we're 20 concerned about.
21 MR. WILKOWSKI That's right.
- 22 MR. HUTC HINSON
- And you may not find the 23 right data for that. Maybe it does exist for 24 laboratory specimens. So I have a feeling it would
(])
25 be very important to do some of these rate effect
4 I
. 240 l 1 tests on laboratory specimens at that temperature.
O 2 MR. WILKOWSKIs What I'm describing here 3 is, and Bernie as well, is the base program for the l 4 l IPIRG, whereas if additional members do join there's 1
5 some of these other aspects that we could evaluate l 6, that might be important to understanding the 1 i
)
^
7l material behavior or enhancing the analysis or l
l 8! understanding plugging effects better, for instance.
HR. SHEWHON: Let me interrupt for a 9l 10 minute. Guy, before you start a program of this 11 ; size you have to write something and explain to NRR t
i, 12 what you feel the need is for a program like this.
i
{s 13I Has such a document been britten and has NRR I
j 14 l responded?
15 MR. ARLOTTO: A document such as that has i
1 16 not been written as such, but we have had a great i
) 17 deal of interaction with them, including a 18j leak-before-break conference at which NRR made some 19 major presentations. And the only thing that I'can 20 say where we got official endorsement from NRR on 21 such a document is through the identification in its r
22 program, in our long range plan, and the contractual 23 arrangements were being made with Battelle and our 24 agreements with the international. They have been i
(])
25 involved with that. We do not have the specific
4 241 4
, 1 document you referred to.
]
O 2 MR. SHEWMON: What I'm missing in this is 1
j 3 a description of what's broken in the phrase "if i
4 it's not broke, don't fix it." We're talking about i
j 5 another program to do things, and it's not clear to l
me where we think the code that's been used in 6
7l inadequate or called into question or what it is I
8ithat makes us think we need another program to go l
9 I out and study these things longer. And as I look 10 through here, I don't think I'm going to get it, so 1
11 !I've interrupted to see if I've missed something. .
l !
, 12 1 MR. ARLOTTO: Okay. I think from a
( 13 strictly regulatory viewpoint, I will answer that 14 question as explicitly'as I can. From our strictly I 15 regulatory viewpoint we are, as you know, relooking i
j 16 at a great deal of the history associated with 17 determination of dynamic effects as well as others '
i
, 18 with respect to assumptions for licensing purposes
- I
! 19jassociated with pipe breaking. We have taken
! l 20 certain actions, as you know, in channel design
)
- 21 criteria for limited scope and broad scope rule.
22 The question that arose that we were i
23 concerned about was what do we do if we do eliminate
{} 24 the double-ended pipe break for licensing purposes?
25 We have a group of people in Germany who believe i
- m. --,#m-mi--mw -%%a-_-, ,w%-_aw_ _ - ,,.ma+a-+m - -
242
_. I that you could assume a ten percent break to replace 2 the double-ended pipe break. We went --
and I've 3 spoken with the Germans on several occasions, 4l including a trip, and so have our people, and there
- 5) is no scientific base for that ten percent 6 assumption.
7' So our next step in our regulatory process 8' which the particular people at NRR are interested in 9 is what do I replace the double-ended pipe break i
10 with, not only for dynamic effects, but is there a 11 ' justification for replacing it, replacing 12i double-ended pipe break for the purposes of ;
(3 s) 13 ; containment design, energy low cooling and equipment i
14 j qualification.
15 l MR. S HEWMON : Well, on the other hand, let l
16l me argue that you've already said what you replace i
17 it with when you've argued that any significant leak 18f or crack will leak at a measurable rate before it I
19 l gets to a dangerous size.
20 MR. ARLOTTO: That's true.
21 MR. S HEWMON : So there's no need for any 22 of this, and it's just a traditional belts and 23 suspenders approach that says we assume any large
(')
us 24 break like that. You're assuming that, though after 25 you've argued for the GDC-4 revision that you don't l
l 243 l 1 need any of that, now you've come back and said, 2 well, we've got it and maybe we need some. It seems
, 3 to me if you stay with what you argued with GDC-4 4 you're never going to find a --
I hate to say 5 rational. There must be some kinder or different i
6 word than that, but some, you know, calculable limit 7lon this. So I'm still somewhat lost on what it is l
8lyou're going to use if these things are supposed to i
, 9 l be detectable long before they get to any i
I size anyway.
10}! appreciable I
q 11 MR. ARLOTTO: Well, I guess in GDC-4 we i i 12lwere looking principally at the limitations
() 13 associated with dynamic effects. It's relatively --
14 ; w e felt that our research showed, and I don't know I
15lwhether I'm going to be able to answer your question l
16 completely, but I am going to give it a try. We 17 felt our research showed that the double-ended pipe 18 break was a low enough probability in certain 19 conditions, under certain conditions, as not to be 20 necessary to assume.
21 However, there's a big difference between 22 the kind of discharge we could get from a I
i 23 double-ended pipe break and zero. And is there --
's
' {]) 24 and with this research we don't know the answer. Is 25 there something that we could justify and say,
244 1 within certain bounds we could assume for regulatory CJ 2 purposes that a break equivalent to X should be 3 assumed. Now, maybe we will never get there. And 4 it makes a-significant difference in terms of what 5 would be required for particularly low pressure
- 6j injection on energy low cooling and what would be 4 1 7 required for equipment qualification.
8! MR. SHEWHON: Yes. It certainly makes a i 9 profound difference on qualifications. The
' 10 double-ended pipe break was, as we said, officially 11l nonmechanistic. I'm tempted to say irrational, but i !
i 12l I guess nonphysical is the kinder word. And if you
( 13 could back off from it --
but since initially it was 14 nonphysical, you're trying to get some physical 15 argument to justify something smaller.
16 MR. ARLOTTO: U m -h m m .
17 MR. SHEWMON: I wish you God speed I guess 18 is my reaction. Other comments on this?
19 MR. ETHERINGTON: As I understood from the 20 previous meeting, implementation of GDC-4 is 21 dependent on inspection, and if you find long cracks 22 it will not be acceptable as a criteria. Does that 23 still stand?
j.
{} 24 MR. S HEWMON : I didn't understand the 25 question. Would you state it again?
i
~ - -
4 245 1 MR. ETHERINGTON: I understood that, first O 2 of all, you do require inspection.
. 3 MR. SHEWMON: Yes.
j 4 MR. ETHERINGTON: Is that right?
5 MR. MAYFIELD: Yes.
I I 6l MR. ETHERINGTON: You can't just rely on 1 i l 7lGDC-4 without any inspection. And secondly, if you
! l a long crack during inspection, then the 8l find 9 l leak-bef ore-break doesn't apply. Is that right or j
! 10 l not?
i 11 I MR. ARLOTTO: Well, Harold, I think we're I
I 12 mixing two things. GDC-4 is trying to address the
- () 13 issue of what do you do before the fact, that is, 4
14 what is your assumption, on what basis can you 15 assume that the pipe isn't going to break in j
16 double-ended fashion? What you say obviously is an 17 operating condition and if we don't have the pipe d
18 with restraints in it already and we see that s
19 there's a long crack, we're going to have to do 20 something about repairing or replacing that piping.
21 So yes, GDC-4 requires looking into things J
22 like --
or having a capability for NDE, but it also 1
23 says you've got to meet other very special criteria 24 before the fact, like limitations on stress
(])
- 25 corrosion cracking, certain concerns about thermal J
?
4
, - ~ _.. _ _____ _ _ _ __ _ _. _ . - - _ - _ . - . . _ _ ._ . . . , _ . _
246 l
7, 1 fatigue, certain concerns about whether or not the 2 pipe can be subjected to water hammer and those kind 3 of things. Those are the kind of limitations before 4l: you could justify the application of l
5l leak-before-break to a piping system for dynamic 6l purposes.
7 MR. KERR: Well, your comment earlier I l 8! think implied that the elimination of the 9l double-ended pipe break was based on what I would 10l call probablistic considerations. You wouldn't say l
11 ' it was absolutely impossible, but the probability l
12: was low. I think implicit in that also then is that ;
13 part of what you want to find out in this program is 14! what is the probability of whatever the criterion 15jyou decide upon. I don't see probability in what 16 l has been described. I see something that says, is 17lit possible to achieve a certain kind of crack. But i
18i am I missing something?
19 MR. ARLOTTO: You're absolutely right.
I 20 And if my associates at Battelle would give me the 21 credit for it, this is exactly an issue that I 22 brought up when I reviewed this program a couple 23 months ago, that they've got to put it in a
() 24 regulatory perspective in terms of some probability 25 of bounding how big the crack would be in the end.
247 s 1 I think that that is a --
take yes for an answer.
2 It's something that I feel is deficient so far in 3 the scope of the program as presently envisioned l
4jor --
I don't know if they've corrected it yet or ,
l 5lnot, but it was one of my comments also when I 6 reviewed the program.
7 MR. SAFFELL: And Guy is correct, one of 8i the things that we're attempting to do is to relate 9fquantification of margins to probability. That will I
10 be a part of this program. F ra nk ly , to be candid, 11 ;our approach in that area is it's best to say still 12 under development.
( 13 We see, you know, the conduction of the i
14 ' experiments and we see relating that to i
15iquantification of margin, you know, in terms of the 16l development of criteria. The relationship of that 17 f to probability can be done, but it obviously depends 18i on things like distributions regarding material i
19 properties, regarding flaw detection, things like 20 that, and that's the link that still needs to be 21 made.
22 MR. BENDER: Mr. Chairman, I would like to 23 take a shot at trying to put some order in this 24 discussion. And there's some disconnects and maybe
{)
25 misunderstandings in what's been said.
4 248 i
l The original concept behind the 1
() 2 leak-before-break was that you may have cracks and 3 the presumption is that you will find the crack i l
4 before it progresses to a point where catastrophic 5 failures have occurred. And I think that's why Mr.
t
! 6 Etherington said, well, you have to say that the i
l 7 crack will only have a certain length. It has to be 8 short enough so that you can make that case.
- i i
9l Now, given that that's the presumption,
! 10l then the question really is. How long is short 1 I 11 enough? And it really relates to the question'of l
12' what kind of loading conditions can be imposed. And
] f) v 13 l if I understand this program, the intent is to start j
14 with some postulated flaw and introduce in it the 15 kind of loading phenomena that can cause crack 16 extension, presumably crack extension that's slow 17 enough or limited enough so that the inspection
- 18 program will catch i t before the crack grows to the 19 point where you're in trouble.
20 Dynamic effects i s one, mounting behavior l 21 is another. And the resistance of the restraint i
4 22 system is another. And I think all of these things 23 have to be addressed, and as soon as you take off 4
{} 24 the pipe restraints and all the mounting restraints, 25 then you're totally dependent on the piping itself i
J
^ ' ' -4 -Wvv- w-- gv-~ - ' - "g wP-y w g y y w--Tw e 9.w-y. -e-,- er-y- 4-&.e-- >ye-a p ,w evW g g e-7 ,e g--*,-e- + -'M-
'-m-4 -- - -
249 1 to provide the load resistance.
O 2 And so I think we kind of have to think 3 about what scenarios we're working with when we i
4 decide whether this program is going to tell us 5 anything. If that's illogical, then I'm sorry, but 6 that's the best I could do to put it in context.
7i MR. MAYFIELD: This program addresses one
. 8 end of what you described, and that is to look at '
9fthe maximum load-carrying capacity of the large j 10 hypothetical flaw. All right? There are other 11 l elements of our overall program that look at piping, I l 12jthat take it from some pre-existent flaw,
( 13 manufacturing defect, whatever, and look at the 14 progression of that flaw, its rate of growth and the 15 detection probability.
16 MR. BENDER: When Guy said --
and I 17 haven't had a chance to really think over what might 2
18 be implied by it --
that we're trying to come to a 19 position as to whether given that the Germans have 1
1 20 no basis for ten percent opening or whatever it is
- 21 they use for designing their restraint system, and,
) 22 therefore, some basis needs to be provided, what
! 23 should that basis be. I don't know whether it's 24
{) going to come out of this part of the program at all.
25 It might be to the extent that you would worry about 1
250 s 1 things like dynamic loads, whether the flaw could 2 , grow in some way.
3l i But I'm just unsure, but I know what --
4lwhat kind of criteria should be applied. It's l
5! convenient to have something like the Germans have 6 because then that says at least wa'll have some i 1 7i basis for saying that the restraint system will l 8 ,' carry some kind of load. Otherwise, you'll have <
I i
9 !nothing but dead load to deal with.
I 10 l MR. S HEWHON : If it doesn't fly apart then 11 an order of magnitude down is a nice round number, I
12 ; you know, it sort of -- it's less than one and you 13 don't want to assume one percent, so you assume it's 14 ! ten percent.
i 15 MR. BENDER: That's a good set of 16 ! engineering logic, but it wouldn't make a physicist 17 ! very happy.
18 ! MR. SHEWHON: Yeah, I can think of certain i
19 l members of the committee that will have trouble with 20 it.
21 Okay. Would you state again what you said 22 in the first sentence two minutes ago and said the 23 program will show --
what I have is response systems
(]) 24 to large flaws. That's not quite it.
25 MR. MAYFIELD: That's not quite it. The
.. - _. - .. . . ~.
l 251 l
- 1 program is going to look at the maximum 2 load-carrying capacity of the piping system that has 3 a large --
some hypothetical large flaw. We're I
4 looking in this program to establish failure margins.
i
- 5 So that really is the extreme condition. We've
! 6 assumed that for whatever reasons we've managed to 4
7llend up with a large flaw in the piping system, and I
8 we want to find out what's the maximum load that 9 that piping can withstand, that flawed piping or 10 cracked piping can withstand under a dynamic --
so 11 (under a quasi-static condition, our Degraded Piping ;
1 l 12 Program has addressed that point, and we've been
( 13 verifying, validating, whatever word you want to use, 14 engineering analyses to predict that maximum 3
15 load-carrying capacity, and we've looked at the l
16 margins inherent in those as well as to the kinds of i 17 margins we might like to. apply from a regulatory 18 sense.
19 The IPIRG program, Task 1 in particular, f
- 20 goes the next step and looks at the margin inherent i
4 21 in the analyses for dynamic loading. So it's the 22 next logical step we feel.
23 MR. S HEWMON : And this is something that's 24 way outside the code, and so the code doesn't
(]) --
it l 25 may give you materials and design procedures which
1 252 I
s 1 would put you in a reasonable position for this, but k
2 it doesn't tell you what the margins arer is that --
3 MR. MAYFIELD: That's correct. And what 4 we're looking at is what are the margins between 5lwhat the code would permit and what would actually I
6lcause failure.
7j HR. BENDER: Mike, I don't want to lose 8l the point that I was trying to make. I may not even i
9l have presented it. If you put a restraint system in i
10 there and introduced a certain kind of stiffness in 11 the system, and that may affect the loading that is 12 seen from these dynamic effects, and so you have to 13i have some restraint system to start with. And if 14lyou don't know what the basis for it is, it's going 15 to be awfully fuzzy as to how to use the results, 16 and I'll just stop there.
17 MR. MAYFIELD: I guess the only point I 18 was really trying to make is that we're trying to 19l validate analyses to make sure that we have a l
20 realistic way of predicting piping system failure 21 loads, for want of a better term. Then you have to 22 go back and look at the actual plant configuration, 23 the actual restraints in the system, and decide 24 whether or not realistically one can see that kind
(])
25 of deflection, those kinds of loads that would cause
.-e- + ,
253 1 failure.
O 2 If I might return to your initial question 3 of what kind of NRR support have we had here, one 4 other --
two other issues, one under the Piping 5 Review Committee in Volume 5,-their recommendations, 6 under the first research recommendhtion they said we 7 should complete the Degraded Piping Program, looking 8 into a variety of parameters. One of those was 9 eeismic or dynamic loading.. So this program 10 addresses that recommendation in part. Part of this 11 program addresses that recommendation.
12 Beyond that, when we went before the 13 Senior Contracts Review Board, we did have an NRR 14 representative present, and he in an unqualified 15 sense supported this program.
16 MR. SHEWMON: Okay. We've screwed up your 17 schedule, but I think it's been useful. Why don't 18 you do what you can in the remaining time.
19 MR. WILKOWSKTt 0 ::ay . My last view graph --
20 (Laughter) 21 MR. WILKOWSKI: I think it's probably 22 important to just briefly discuss the final.subtssk 23 within this one task here in that here we're trying 24
{) to assess what happens with a more realistic piping _
25 system.
254 1 I might say that what we have is a series 2 of experiments. In this case we're looking at a 3 larger diameter pipe, about 16 inch diameter, about 4 one inch wall thickness, but we've tried to select 5 different types of materials. Again, for instance, 6 here we have a carbon steel, for instance, an A1068 7 Grade B pipe that we think is susceptible to the 8l dynamic strain-aging that we talked a little bit 9 about yesterday. So we'll see how that phenomena 1 0 ;i behaves under a seismic loading condition.
11 The stainless steel is an extremely tough I 12 material, so that we think we know what's going to
()
k> 13l happen. This particular carbon steel weld is very 14l low in toughness, and we also see that it's 15 susceptible to dynamic strain-aging. So that if you 16 have an internal surface crack it's possible that 17 under almost linear elastic type of loads that that 18 surface crack could penetrate through the wall 19 thickness, and you would end up with leakage from 20 that surface crack under the simulated seismic 21 loading.
22 And, of course, then there's also a 23 stainless steel weld material where the flux welds 24 are fairly low in toughness.
(])
25 This is a stick model of what one possible
l 255 1 piping system is that we looked at where this has a O 2 certain amount of dimensions and a crack would be 3 located at some critical location. We've done some 4 finite element analyses, modal analyses on this 5 initial design and looked at what are the bending 6 loads, torsional loads, et cetera. That still needs 7 some further optimization so that we have types of 8 loadings that would be typical if there is a typical 9 nuclear piping system.
10 MR. RODABAUGH I see you do have your 11 input is F sub X equals P. Is that supposed to be 12 cosine T?
( 13 MR. WILKOWSKI: Yes, cosine T.
14 MR. RODABAUGHz So you'll have the input 15 in the form of the cosine or sine wave in the one 16 point.
17 MR. WILKOWSKI: At one point, yes.
18 MR. RODAB AUG H: And a couple other 19 supports.
20 MR. WILKOWSKI: That's right. And it's 21 fairly complex. So that, for instance, if we were 22 trying to do the sane thing with all of the bending 23 that occurs here, if we were putting some input here, 24 instead of a single end point input loading function
(])
25 here, if we were to simulate the same type of i
1 256 4
1 loading function up here, for instance, you would l ( 2 have to have several components or actuators. We 3 tried to simplify what type of loading system we had 4 for these experiments, so that it's relatively easy 5 to do experimentally as well as to analyze it.
6 This i s just a rough schematic again.
7 Here woulJ be that representative piping system 8 under some electrically isolated loading system that -
i 9 would be computer controlled. Also, accompanying l 10 with this would be a pressure vessel for supplying i 11 extra water for taking leakage measurements-once the 12 surface crack did break through the wall thickness 13 during the experiments.
14 I'm going to have to really breeze through 4 15 the analytical efforts that are involved here.
16 MR. BENDER: Let me just ask one thing.
}
17 MR. WILKOWSKI: Yes, sir.
- 18 MR. BENDER
- Sometime it will be useful l
19 for somebody to make a table that lists the 20 parameters that are being investigated with this 21 kind of system just so we really know what they are.
22 We've got so much stuff that you stated and it's 23 hard to absorb i t.
24 MR. WILKOWSKI Okay. As far as the
(])
j 25- analytical efforts, we're really not doing any major v-r - a - , ,- w, , , - - .,n--<w-- , . _ - - , ,- -
y ,,. , , - , , - - - . -,,e.-,---,-,w,.,-.n
257 1 analytical developments. It's more or less
()
l 2 assessing existing technology. We will for the 3 ! design of some of the more complicated experiments, 4 i such as in 1.3, have to develop a simple beam I
5 [ element that would simulate the plastic behavior of l
6: the crack location -- at the crack location so that l
7 we know how to develop the frequencies and i
8 anplitudes that are necessary in order to get the 9 crack to grow. And that's not a very simple problem.
10 As far as the post test analysis, we would 11!use current analysis procedures and try to define 12 what the margins of safety are based on what people 13lare using right now as far as a linear elastic 14janalysis of the stresses in the piping system, 15!taking a peak load and applying a static fracture 16l mechanics approach to it.
17 With that, I'll skip all the way to the 18 material characterization. This is a very similar 19 type of characterization between Subtask 1.1 and 1.2 20 in that we're selecting materials that have been 21 quasi-statically tested in the Degraded Piping 22 Program, and we've already conducted pipe 23 experiments on these under quasi-static loading, so
{} 24 that the testing that would be done within this 25 program is really doing tensile tests under strain
258 1 rates that are comparable to what you would see in 2 the piping system and also doing conventional 3 i compact tension specimens or bend bar specimens at 4 the similar type of loading rates to be experienced 5! in a pipe test, and also a nonconventional fracture 6ftoughness specimen that simulates the crack growth I
7j thtough the thickness as in the surface crack pipe 8l tests.
9l We would eventually be writing a topical 10l report on the effects of the loading rate on these 11; different materials.
12 , My last view graph then goes to the next O
s/ 13l task, which is development of data bases. There's 14l two subtasks within this. One is the data base on I
15l nuclear piping materials. The intent of this 16 particular subtask is really to build upon what MEA 17 is doing in their data base, that is, try to obtain !
18 material property data from the other IP'RG members 19 l as well as coordinating any material tests that we 20 develop and making sure that that's input into the 21 MEA data base.
22 There may also be cases where some of the 23 other IPIRG members would like to donate some 24 materials to the MEA data base for their testing
(])
25 within their own program. .
- 259 1 The final effort here is the subtask --
0 2 this is a data base on pipe fracture experiments.
! 3 This includes adding to a data base that we have 4 from the Degraded Piping Program, including the 5 experimental data from this IPIRG program as well as 6 going back to other critical experiments that people 7 have run in past programs or in current programs and I
8 l pulling together that data and putting it all in one 9 location, so if any of the IPIRG members or the NRC 10 wish to review that data with future analysis 11! methods, for instance, then they can do so in a 12l quick fashion.
13 And that completes --
that leaves Bernie l
14 l minus three minutes.
t 15 MR. BENDER: Well, let me use one of them.
16 MR. WILKOWSKI: Okay.
17 MR. BENDER: Who are the participants in 18lIPIRG now? Do we know?
19 MR. WILKOWSKI: Contractually we don't 20 have --
21 MR. S IIE W M O N : The last slide has a 22 potential list, I would guess, if we got to that. I 23 don't know.
24 MR. WILKOWSKI: Maybe Bernie can get to
{' )
25 that point.
i
, 260 4
1 MR. BENDER: If you're going to talk about O 2 it later I'll just not ask.
3 MR. WILKOWSKI: Yes.
4 MR. HU TC HI NSON : Can I make one quick I
5Icomment?
I 6 MR. WILKOWSKI Yes.
i 7! MR. HU TC HI NS ON : I know very little about t
8 the dynamic methods for analyzing piping, but from 9 what I have heard I know that a lot of them are 10 basically static methods in which you put in dynamic 11 factors. And the one thing I would hope, and I I
12 ! didn't hear you say it, and it's not obvious from 13 your presentation either, from the overview, is that 14!to the extent possible I think it would be very good l
15 to do from a fracture part, static and dynamic tests 16 side by side on essentially identically cracked 17 pipes to the extent that that's possible.
18 ' I mean I know you've got lots of static 19 data already. It would be a real shame if you used 20 different geometries and so forth for all your 21 dynamic tests, because I think probably the most 22 valuable thing that could come out of all of this i 23 would be statements about how given geometry of l
{} 24 cracked pipe behaves under dynamic, certain kind 25 of --
rate of loading as compared to how it behaved 4
261 1 under quasi-static loading.
(~')
2 MR. WILKOWSKI: Virtually every experiment 3 that we have in the IPIRG program, we have a 4 parallel quasi-static pipe test in the Degraded 5 Piping Program.
6! MR. HU TC HI NS ON : Same pipe.
i 7l MR. WILKOWSKI: Same material, same flaw 8 size, but just at lower rates. So you can just make 9fa direct comparison, and you don't need any analysis, i
10 t for instance, to make a quantitative statement.
11 , MR. HU TC HI N S ON : Okay. Good.
12 MR. SHEWMON: Bernie?
() 13 MR. SAFFELL: This was a view graph that 14 was presented earlier in your presentation. I'll 15 just take a minute to say that right now there are 16 no specific tasks identified within this Task 3, the 17 fracture of piping containing high energy fluids.
18{This is being addressed right now within the scope I
19i of Task 1, but it remains in the event that member 20 countries have concerns that require the 21 identification of specific subtasks within that task.
22 Task 4 is the resolution, as you can see, 23 of unresolved safety issues. That's in general, not
{] 24 strictly from the NRC Degraded Piping Program.
25 HR. WILKOWSKI: Bernie, that's not safety
1 262 4
1 issues.
()
2 MR. SAFFELL: What did I say? If I said 3 safety issues, I misspoke. That's unresolved issues 4 from the NRC Degraded Piping Program and other 5 issues as they may be identified by member countries.
! 6 The first task which will be pursued
- 7 within that relates to the evaluation and refinement i 8 of leak rate estimation models. The NRC is
[ 9 sponsoring the first --
the intent is for the NRC to 10 s!onsor the first year of this, and the scope of 11 this will involve the integration of existing crack 12 opening area and thermal hydraulic models to enable f
13 one to predict leakage flow rate through a variety i
14 of cracked pipe conditions, and the thing is here to i
! 15 get the interface between the thermal hydraulics and 16 the crack opening or the structural part of the
]
17 crack opening prediction integrated.
j 18 We will also do an evaluation through i
i 19 looking at the literature and surveying the member 20 countries into the effect of particulates on leak
]
21 rate.
i
- 22 In terms of the leak rate testing area, we 23 will prepare a test matrix based on the data
]
{} 24 required as opposed to the data which exist required 25 to verify the analytical model. In other words, i
263 I
1 we'll have a model. We know that leak rate data O 2 exists. We don't know whether it's complete for the 3 range. We'll need to assess this model, and so a 4 ; test matrix will come to fulfill any data needs.
I 5: Also, out of this one will come a determination or a I
6 ! recommendation regarding the need for experimental 7l evaluation of particulate plugging.
8 MR. KERR: What will you do with the 9 information on particulate plugging? Or what will 10!someone do with it? Actually perhaps that's a I
11jbetter question.
12 l MR. SAFFELL: Okay. I guess the intent of
() 13 this is to determine if that is an issue which 14 requires further work and the magnitude, I guess, of 15 the issue, you know, is it a significant effect on 16 the cracks and the openings.
17 HR. KERR: Even if it turned out to be a 18 significant influence, would you then assume that 19 particulates always existed in reactor coolant 20 systems? I doubt it.
21 MR. SAFFELL: No, I don't think we would.
22 In fact, that is one of the things we'll be looking 23 for is what is the likelihood -- I personally don't 24 have an appreciation for the particulates and the
(]) l 25 size of the particulates that would exist in, for 1
264 1 example, a reactor coolant system.
' O 2 MR. BENDER: You may be dealing with this 3i in the wrong context. The reason why this issue is l
r 4 l important is because we have to come to grips with 5; the matter of what size leak we can find, and what 6i size leak we can find depends a great deal on what's 7 interfering with flow. And so particulates get in 8 there and you have to postulate them. Then you'll 9fhave to assume that the opening isn't as big as it 10 ! would be without particulates. And so you've got 11 that question to deal with. And there are a lot of 12 others like that. So in trying to work on this os 13 program, the main interest is in trying to get an 14 understanding of how big a leak we can detect in 15 terms of the flaw size or crack, whatever it is 16 that's out there, that we're saying is the limiting l
17l size that we can tolerate. And we have to show the i
18 crack is big enough so we can find the leak before 19 it grows to catastrophic proportions. That's the 20 whole premise on which we're working.
21 MR. MAYFIELD: Indeed, you're right, and 22 this piece of the program is the first step or at 23 least we've seen this as the first step along that 24 path, and it will eventually link with the work, for
(]}
25 example, Dr. Kupperman is going to talk about, the i
j
265 1 leak detection.
b 2 MR. BENDER: All I know so far that's 3 l available is some work GE that has done. Kupperman l
4 =ay have some additional work.
5 MR. SAFFELL: Okay. The current status i
6 l regarding membership within the IPIRG is that EPRI i
7Ihas indicated they will join, and we have this list 8 of countries that we expect to join. The United i
9 ! Kingdom, France, Japan and Canada will join l
10 ' individually, and as will Taiwan and the Swiss and 11 j the Swedes will join together. These countries are 12iconsidering it, and Finland and Belgium have 13 declined to date.
14 f MR. S HEWMON : If one's population is under 15l30 million you get to come in as a group, is that it i
16 ' o r --
17 MR. SAFFELL: No.
18 MR. MAYFIELD: This has been kind of a 19l sticky point for us. Sweden, Switzerland and Taiwan 20 will all be half members for want of a better word 21 or better description.
22 MR. S HEWHON : Fine. Okay. There's at 23 least half an argument for that.
24 MR. MAYFIELD: Yes. Perhaps that's the
{])
25 most you can get is half an argument.
266 g 1 MR. BENDER: Try charging for kilowatt of b) 2 capacity.
3 MR. MAYFIELD: We've spent a lot of time 1
4i agonizing over how to handle that.
i 5' MR. S HEWMON : Is that the end of the IPIRG?
6; MR. WILKOWSKI: Yes.
t 7i MR. S HEWMON : Okay. Then I guess we shift i
8l over to ANL where we will be for several hours.
9l Dave?
i 10 ' MR. KUPPERMAN: My name is Dave Kupperman, 11 and I'm going to be discussing the Argonne program 12 on assessment of leak detection technology for !
O 13} nuclear reactors. This presentation will discuss 4
14lj the current practice in leak detection, the problems 15 that exits, a discussion of existing technology, 16 i motivation for and evaluation of acoustic leak I
i 17; detection, and some recommendations that we would 18 j like to make.
19 The current practice is driven by 20 Regulatory Guido 1.45 which recommends three methods 21 be applied for leak detection, the sump flow monitor, 22 airborne particulate radioactivity monitor, 23 condensate flow rate or airborne gaseous 24 radioactivity monitor. And these are
{])
25 recommendations, and utilities do not necessarily
.~.- - - -- . - . - - - - . - _ . - _ - --.. - - ~ ~ - ~ - -- ---- - -
i 1 267 2
1 follow these suggestions, and we've looked into what
() 2 utilities actually do.
3 MR. SHEWMON: Now, you have two up there 4 that are required, but you say in spite of being
- 5 4 required they aren't followed or which is --
6 MR. KUPPERMAN: Required doesn't mean that 7 there's a law. It'just means that the Nuclear 8 Regulatory Commission would like them to use these 1
9 l two methods for detecting leaks.
10 .
MR. SHEWMON: Okay.
s i
11 MR. KUPPERMAN: Very few reactors actually f 12 follow these guidelines -- very few utilities
(} 13 actually follow these guidelines.
}
j 14 We've look at about 74 plants, and we can i
4 15 say that at least one of the two required techniques 16 are used in all plants, but as I said, few'use three, i
17 and in general the. main method for detecting leaks i
i 18 is the sump pump monitor in those reactors which j 19 have one.
I 20 I should also point out that the permitted i 21 limit on unidentified leakage is different for PWRs i
f 22 than BWRs. For PWRs it's one gallon per minute and j 23 BWRs it's five gallons per minute. This is a result i
{} 24 simply of practical considerations in operating the
- 25 plant.
i
267 1 follow these suggestions, and we've looked into what 2 utilities actually do.
3l MR. SHEWMON: Now, you have two up there 4lthat are required, but you say in spite of being l
5irequired they aren't followed or which is --
t 6' MR. KUPPERMAN: Required doesn't mean that 7{there's a law. It just means that the Nuclear l l 8i Regulatory Commission would like them to use these 9 two methods for detecting leaks.
i 10 MR. SHEWMON: Okay.
11 I MR. KUPPERMAN: Very few reactors actually i
12 follow these guidelines --
very few utilities
() 13 actually follow these guidelines.
14 f We've look at about 74 plants, and we can I
15lsay that at least one of the two required techniques l
16 ' are used in all plants, but as I said, few'use three, 17 and in general the main method for detecting leaks 18 is the sump pump monitor in those reactors which 19 have one.
20 I should also point out that the permitted 21 limit on unidentified leakage is different for PWRs 22 than BWRs. For PWRs it's one gallon per minute and 23 BWR3 it's five gallons per minute. This is a result 24 simply of practical considerations in operating the
{])
25 plant. l l
268 1 The Duane Arnold incident suggests that at O 2 least in some cases the leak detection technology is 3 not adequate, and what we see here is a 360 degree 4 crack which penetrated the outer surface over about 5 an 80 degree angle giving a crack length of about 6! seven inches and a leak rate of about three and a 7, half gallons per minute.
8 First of all, this is about three times 9 less than one would have predicted f rom a crack of 10 ' that length, and second, they didn't even have to I '
11 ' shut the reactor down for a situation such as this.
12 Fortunately, someone decided this was an anomalous 13 situation and the reactor was shut down before the i
14 i pipe guillotined.
i 15 i There's considerable motivation in general i
16 for research on leak detection and for advancing i
17 leak detection technology, and in the report of the i
18l U. S. Nuclear Regulatory Commission Piping Review i
19l Committee, they specifically indicate that --
they 20 suggest that improved leak detection systems and 21 their development should be pursued and recommend 22 that improved leak detection systems under 23 development be completed in field testing.
24 MR. S HEWMON : Dave, I'm still confused by
(])
25 your use of the word required, but I've substituted ;
i
269 1 recommended, if that's what a reg guide does, and as 2 I understand it usually they do. There must be 3 ,something in the tech specs that says how many 4! gallons per minute they can be dumping on the floor?
I 5 MR. KUPPERMAN: Each tech spec has that in 6 there, and for most PWRs it's one gallon per minute, 7 unidentified.
l 8i MR. S HEWMON : Okay.
9i MR. KUPPERMAN: For BWRs it's five gallons 10 per minute.
11 MR. SHEWMON: So reg guides recommend.
12 MR. KUPPERMAN: Recommend.
13 MR. SHEWMON: What are the requirements?
14 1 The requirements are in the tech specs.
l 15 { MR. KUPPERMAN: The requirements are in 16 the tech specs.
17 MR. S HEWMON : They talk about unidentified 18l leakage is one gallon for PW and five for BW7 19 MR. KUPPERMAN: In general. There's some i
20 exceptions, but in general that's correct.
21 MR. SHEWMON: And how they do that then is 22 left up to the operator or the licensee, and the reg 23 guide suggests what would be accepted.
+
24 MR. KUPPERMAN: How to do it.
{])
25 MR. S HAC K : Doesn't the tech spec actually
270 1 specify the leak detection system for each reactor?
2 MR. KUPPERMAN: Yes, it says in the tech i
3; spec how they're going to detect leaks.
4! MR. SHACK: That's sort of agreed upon l
5lbetween the utility and the reg.
6! MR. BENDER: What the reg guide says is, 7l here's one way you can do it that they'll accept.
8 l If you want to use something else, you have to 9l propose something. And normally they take the reg 10! guide because that's all there is.
11 MR. KUPPERMAN: There are other ways to 12 detect leaks other than those three methods, and 13 I'll discuss, you know, some of those.
14 I just want to point out that there are i 15 other experimental programs on advancing leak i
16l detection technology in other countries, and 17lspecificallythere are significant programs in West 18! Germany and France. So there's a considerable 19 interest in advancing leak detection technology.
20 The reason --
there are several reasons .
21 but --
several of them have been listed here. There 22 are problems, and i t. is possible that you could have 23 a large crack with a low leak rate. And this could 24 be as was discussed because of corrosion fouling and
(])
25 uniform growth of a long crack before it actually l
271 1
1 penetrates the OD. !
O 2 I indicated the Duane Arnold incident, 5 3 suggesting that existing systems may not be adequate 4 in some cases. And simply tightening the current 5 leakage limits to improve the sensitivity could 6 result in unnecessary shutdowns because of the 7 ' inability to identify the source of the leak.
8; So the potential solutions or, you know, 9 the possible solutions to these problems are simply, l
10 not simply, but to improve the sensitivity and 11 reliability of the existing systems, and this can be 12 ! done through the installation of some more effective
() 13 leak detection systems, and specifically acoustic I
14 l monitors and moisture sensitive tape have the i
15 3 greatest potential for improving leak detection
! 16l technology.
i l 17 ; Utilities have not been idle in this area, i
7 18! and they have tried some relatively -- at least in i
19lour minds, relatively elementary systems in the i
- 20 ' areas of acoustics and moisture sensitive tape.
21 They've looked at acoustics. They've installed some 22 acoustic monitors on unrepaired welds that have some 23 crack indications, welds with overlays, and they 24 have been monitoring some valves. You can look at
(])
25 this slide a little more carefully in your hand-out
I i 272 1 to see some of the details of the way that's done.
O In addition, they have looked at moisture 2l
}
3 sensitive tape, and again they've put some of these i
4l tapes on weld overlays and unrepaired potentially l
5 cracked pipe. The moisture sensitive tape is 6j subject to false alarms, and there's already been l
7 one instance where an alarm has gone off from a leak 8I that was not near the tape. It was a large leak far
)
9' away from the valve.
10 MR. SHEWMON: Are the acoustic monitors 11 always looking for a path through metal? ,
i e l
12 < MR. KUPPERMAN: The acoustic --
13 ; MR. SHEWMON: Or never?
t 14 ' MR. KUPPERMAN: The acoustic wave that is 15l generated as a result of a leak can propagate 16 through the metal and/or through the water, and 17lactually in our experiments we can separate the 18 signal that goes through the metal from these.
19 ' MR. S HEWMON : But it's not through air.
20 MR. KUPPERMAN: It's not through air, 21 right.
22 MR. SHEWMON: Okay.
23 MR. KUPPERMAU: The current technology 24 does not provide the optimal sensitivity location
{)
25 capability and flow rate accuracy that are needed.
273 1 And we through our experiments, which I'll discuss,
() 2 have seen that acoustic monitoring techniques can
)
l 3 improve the leak detection capability, but the 4 existing systems, the ones that are used by the 5 utilities today, do not provide source 1
6 discrimination or leak rate information. So there l t
7lare considerable areas for improvement. i I
8 ll Moisture sensitive tape in general is <
l 9finherently limited. It does not provide any r 10l quantitative information. There's much more f
11linformation in acoustic signal than one can acquire l
12jfrom moisture sensitive tape. '
() 13 MR. BENDER: You said moisture sensitive 14 tapes are being used somewhere now.
15 MR. KUPPERMAN: There's several plants 16 that have tried -- have installed these, right.
17 MR. BENDER: What kind of systems are 18 those? Besides being moisture sensitive, what do 19 they consist of?
20 MR. KUPPERMAN: What is a moisture 21 sensitive tape?
l 22 MR. BENDER: Yeah.
I 23 MR. KUPPERMAN: It's literally a piece of l 24 tape that is attached to a tube. The tube goes
[)
25 through the insulation, and when there is a leak
274
~3 1 hopefully the water will find its way to the tube, V
2 drip down the tube and come in contact with the tape.
3i The tape has a couple of wires in it and the salt I
i 4l that conducts when it gets wet and shorts out the i
5 circuit and sets the alarm.
6l MR. BENDER: So the water is delivered to 7 a --
l 8' MR. KUPPERMAN: To the site. The water i
9' has to find its way to the site where the tape is i
10 ; located.
i 11 l MR. BENDER: That's enough about that.
I 12lThank you.
() 13 t I
MR. KUPPERMAN: Okay. Let me point out 14l that there's also economic incentive to promote 15l leak-before-break to other piping systemer and leak 16l detection technology, acoustic technology and even 17lmoisture sensitive tape can be used to do this.
l 18l Acoustic leak detection can also offer a low cost 1
19 margin of safety for seam welds, for example, in 20 some of these systems outside of containment.
21 So the technology that we're working on 22 does not necessarily have to be applied simply to 23 insido containment. It can be very useful in 24 outsido containment where there are no leak
(])
25 detection systems currently operating.
i f
275 1 Now, I would like to discuss the Argonne O 2 program which has consisted of a review of leak 3 detection technology employed in reactors and 4 considerable amount of laboratory and field work on 5fthe assessment of advanced leak detection technology.
6 Laboratory studies at Argonne have been 7 aimed toward assessing the adequacy of acoustic 8' methods to detect, locate and size leaks. One of 9 the unique aspects of our investigation has been the 10 fact that we have carried out these studies with 11 . field-induced cracks, sections of piping that have 12 been removed from reactors and brought to Argonne
] 13 and integrated into our leak detection system.
14 We've also been able to acquire acoustic 15 background data from existing reactors, a very key 16 element in the analysis of how sensitive acoustic
' l 17 l systems can be.
18 l And we have also provided help to 19i utilities already in the example where we have 20 reproduced a system that has been installed by a 21 utility, we have reproduced that system at Argonne 22 and established its sensitivity and systems dynamic 23 range, giving us a feeling of what is actually out 24 there right now.
(])
25 And we have put together a bread board
276 1 system of digital continuous acoustic monitoring 2 system that uses advance signal processing, and 3jwe've shown how this kind of technology can improve i
4i leak detection.
5! MR. S HEWMON : On the first one, do you 6lhave any evidence that the leak rate varies with 7 time or after you change the water chemistry, or 8fwill you get into the studies you've done on those?
9 MR. KUPPERMAN: The leak rate variations i
10 ' that we have studied have been the result of 11 : temperature changes, pressure changes and changes in 12 stress.
13 MR. S HEWMON : Changes in stress?
i 14 ' MR. KUPPERMAN: The loading of the pipe.
15 t MR. SHEWHON: No, but if there was 16! plugging and you then rattled it around and changed 17lthe water chemistry --
18 MR. KUPPERMAN: Plugging is very serious.
19 When we've got the cracks into Argonne, they do not l
20l leak. We have to clean them up.
21 MR. SHEWMON: Were they leaking before you 22 got them there?
23 MR. KUPPERMAN: Presumably.
24 MR. SHEWMON: I mean that's the way they
(])
25 were found presumably.
J 277 1 MR. KUPPERMAN: But I don't know what the 2 l loading was on them when they were leaking.
l 3j MR. S HEWMON : Okay.
4 MR. KUPPERMAN: But we do have to work the 5 fcrack to get it to leak. It doesn't leak when you I
6 just stick it in. Now, part of this could be due to 7 the welding of it into our system, and maybe the 8! crack is tighter in our lab than it was in the field.
9 It's hard to know that. But the leak rate grows as i
10lit is used, as it gets cleaned out, because we put 11 in clean water and some chemicals to clean it out.
12 ; MR. SHEWMON: Okay.
13 MR. KUPPERMAN: Also, I don't know if I l
i 14; mentioned, but we also built a facility to evaluate i
15'the moisture sensitive tape so we can compare the 16 two technologies. This is a photograph of a leaking 17 intergranular stress corrosion crack. This plume is 18 a result of a leak of only a few thousandths of a 1
19 gallon per minute from a crack opening here which is l 20 only a few millimeters long. And we studied leaks 21 of this type. At this point we have studied 22 relatively small leak rates, on the order of a few 23 thousandths of a gallon per minute to up to on the 24 order of a tenth of a gallon per minute for IGSCC
(])
25 and up to one gallon per minute for --
l
278 1 MR. SHEWMON: Is there anything special
(),
2 about that lighting aside from the fact that it's a 3l relatively intense source?
l 4{ MR. KUPPERMAN: Excuse me?
5! MR. S !!EWMON : It looks like a fairly 6 intense light. Is there any special wave length or 7i; do you need to show it up that clearly?
I 8 MR. KUPPERMAN: You can see the leak very 9l clearly by eye. There's no problem. It's not like I
- 7. 0 !a steam leak that is invisible. I mean it's --
11 ' MR. SHEWMON: This is a steam leakr isn't 12j it?
13 MR. KUPPERMAN: Well, it's coming out as a 14l mixture of steam and water.
15 MR. SHEWHON: Okay.
MR. KUPPERMAN: It's flashing in the crack 16l 17l and then it's a mixture.
i t
18i MR. S lfE W H O N : So it's solid water on the l
19I other side?
20 MR. KUPPERMAN: It's solid.
21 MR. S ilE W H O N : Okay.
22 MR. KUPPERMAN: So that's the kind of leak 23 that we have been looking at. And this is a
{} 24 photograph of the facility that has been used to 25 evaluate these leaks. It consists primarily of a
279 1 ten meter pipe run of ten inch Schedule 80 piping O 2 into which we can weld these pipe sections.
3 f What we see here is a hood that allows us 4 lto handle these radioactive pipe sections initially.
5 ' We need this hood after the pipe --
after leaks have 6lbeen run through it and the radioactive particulate 7fmatterhas been blown out, this can be removed, 8 lwhich allows us to do a better investigation.
9l This facility allows us to reproduce the 10 Itemperature and pressures that exist in the reactor, 11' and generate leaks that we assume are very similar
)
12!to what one would find in the field. We have a 13 hydraulic jack here that allows us to load the pipe 14 l s o w e can vary the flow rate, and we can do studies I
15 l with various types of insulations to see how that l
16 might affect the generation of the acoustic signal
- 17!and the propagation of the acoustic waves.
18 This is a photograph of the moisture 19 sensitive tape facility. It's much simpler and 20 there's a relatively simple system, electronic 21 system. As I indicated before, it simply detects 22 the presence of water on the tape, which is attached 23 to a --
it's a piece of tape attached to a tube i
24 which in inserted into the insulation. And we've
{])
25 l done some studies to see how sensitive it is. 4
i
, , 280 1 It is quite sensitive, but it's quite 2 important, you know, the geometry is very important, 3 and it's possible that you could have a leak where j 4 the fluid does not go to the tape and does not set i 5 off the alarm, and also as I indicated, it's not at 6 all quantitative. It's an on/off system, and once 7 it's on it takes a long time to dry. So if you have 8 a false alarm, your system is out of commission for 9 quite a while.
10 Now, I would like to talk about some of 11 the data that we've acquired at Argonne. As I 12 indicated, we have looked at intergranular stress 13 corrosion cracks. They are different in their 14 characteristics in terms of generating acoustic 15 waves than are other types of leak sources, and this 16 slide indicates to some extent that difference.
17 This is the amplitude of the acoustic signal as a 18 function of the leak rate.
19 This shows data from three IGSCC that were 20 welded into our pipe run. Despite the variation in 21 geometry and the frequency window of interest for us, ,
1 22 the data pretty much is the same. l 23 It's different and it's noisier than, i
24 let's say, a thermal fatigue crack or mechanical
[]}
i
] 25 fatigue crack. There are other differences, but j i
~ - _ _ - - . . . - , . . . . ,
281 1 this immediately shows that an IGSCC because of its O 2 tightness does have different characteristics than 3 other types of leak sources. And because the leak 4 rate is similar, even though the geometry changes, 5 we have some confidence that the acoustic signal can 6 be used to estimate the size of the leak.
7 Just another quick slide about the fact 8 that we've looked at valves and flanges, and that 9 data, although there's considerable more scatter, 10 does, you know, fall at least in the order of l
11, magnitude of the amplitude of the signals from the i
12! thermal fatigue cracks and mechanical fatigue cracks.
( 13 Other differences between IGSCC and other 14 types of cracks and valves and flanges can be seen 15 in the frequency spectrum. Here we've noticed that 16 the amplitude of the acoustic signal as a function 17 of frequency has a slightly different -- well, in 18 some cases a significantly different character.
19 The general trend is that the signal at 20 higher frequencies relative to lower frequencies is 21 larger for IGSCC than all other types of leak 22 sources that we've looked at. So by comparing the 23 signal in two different frequency windows, we have !
l 24 information regarding whether or not the signal is
{}
25 coming from an intergranular stress corrosion crack.
l l
l
I 282 1 There are many practical aspects of 2 evaluating acoustic leak detection systems. One of l
3 them is simply how does the insulation affect the 4 generation and the propagation of the acoustic wave.
5 One thing we've seen is that under certain
- I 6l circumstances with a flow rate on the order of a few l
j 7 hundredths of a gallon per minute, the fact that the 8 insulation is present results in a considerable ,
9 enhancement of the ultrasonic signal. This is the 10 result of the noise being generated when the fluid I
11 hits the insulation, and also it condenses and falls ;
l 12 back onto the hot pipe and fries, so to speak, and 13 generates considerable noise. But it's only I
14 significant for certain flow rates and under certain a
15 geometrical configurations.
16 We've also had a look at how acoustic 17 signal varies, you know, with the geometry, that is, 4
18 the relative position of the receiver and the crack, 19 let's say, and this slide just shows th a t we've 20 studied that, and that the relative circumferential 21 position of the crack compared to the transducer, 22 you know, it doesn't matter if the transducer and 23 the crack are along the same line or if the crack is 24 down below and the transducer is on top of the pipe.
[])
25 When you get a meter or so away, t h a t' variation is 4
+
e s- - - ,y- ,r ..y , . _ . ,. , - - - -, , 7 .. , , - , . , - - , , , - ~ yp .-.. - - , - , . . - - .
a A = ' = w =
_ a
= w' s-.p+r e _ , , - - __. a w- - 'Aa-' ~
283 1 insignificant. So we don't have to worry about that.
1
[
() 2 One advantage of a very sensitive leak l
3 detection system that uses acoustic waves is that 4 you can get quantitative information regarding the 5 flow rate when the flow rates are very small, and 6 this slide illustrates this point. These acoustic j 7 systems are sensitive enough to detect flow rates in 8 the orders of thousandths of a gallon per minute.
9 If you can catch the leak when it's very small, 10 through the variation in the amplitude of the signal 11,you can watch the crack grow. That is, the leak --
l 12 ' I should say you can watch the leak rate increase.
13 This shows a larger variation and signal 14 about the mean for a two thousandth of a gallon per i
15 minute leak compared to the same crack as we load it 16 and open it and increase the flow rate to about six 17 thousandth of a gallon per minute. So there's i
18 information in the acoustic signal that one really 19 would have a very difficult time acquiring by any 20 other technique.
21 MR. BENDER: Excuse me. Before you get 22 off that point, when you say you can do that 23 quantitative determination --
1 24 MR. KUPPERMAN: This illustrates that
(])
i 25 there's information that we should be able to. I'm I
l i
. ~ . -- .. . . . . _ - . _
~~ ..- . . _ _ . - - . . - - . _. -
4 284 1 not saying we can.
2 MR. BENDER: I'm thinking about the 3 uncertainties in it at the moment.
4 MR. KUPPERMAN: There are uncertainties.
5 MR. BENDER: And maybe before you get done 6 you can say what -- when you talk about quantifying, 7 if I can use that word, what is the measurement 8 capability that we ought to be having and how far is 9 it affected by the various factors that would be 10 considered, for example, the location, the types of 11 insulation that are around the pipe and a few things j
12 like that.
() 13 MR. KUPPERMAN: One of my slides indicates i 14 what the sensitivity of an acoustic system can be.
l 15 MR. BENDER: So if I'll wait, I'll hear it.
16 MR. KUPPERMAN: If I haven't answered all !
l 37 the questions with that slide, ask again. I 18 MR. BENDER: I don't know all the 19 questions.
20 MR. KUPPERMAN: We have a pretty good 21 feeling of what is out there in the field in those 22 few reactors that are monitoring pipes with overlay 23 or pipes that have indications that they might be 24 cracked. As I indicated, through the reproduction
{)
i 25 of the system that's used at Ita t c h , what we've
. . ~ - - . . -. . .. .- . - - . ~ . . - . - - - -
I
'T 285 1 learned is that the system that is out there at 2 Hatch in particular is quite sensitive.
'3 This is a trace of the RMS signal as a j 4 function of tine. This level is the no acoustic l
j 5 background noise, no flow through the crack. We
! 6 have data from Hatch regarding the background noise.
7 So we can electronically induce noise into the pipe 8 to simulate the background noise. So this is our
! 9 base level. And it's clear that even with the 10 elementary systen that they have, you can detect
- 11 leaks on the order of a few thousandths of a gallon 12 per minute. But the problem is --
13 MR. KERR How do you measure .002 gallons 1
j 14 per minute?
I
- 15 MR. KUPPERMAN
- We have a turbine. The 16 fluid goes through a turbine before it goes into the I 17 pipe. So we have a turbine flow meter type of [
18 device that's calibrated against absolute 19; measurements. This is before it goes into the pipe
! 20 we measure the flow rate.
21 MR. KERR: That's about two drops, right?
22 MR. KUPPERMAN: No, what you saw, that 23 photograph. No, it's considerable. I mean it's 24 noisy. It's naybe
({}
25 MR. KERR: No, I'm talking about the fluid I
1
.--,,,v-, .--.y. -.. .,...----y-
-.-m,-m,, -
. - - , , , , - . , , , , . - , . - - - ,~,.-m.-,. , - , , , , , e- - ,, - - , - -
l
?
286 l
1 volume, not the noise.
O i
2 MR. KUPPERMAN: It's small, yeah, it's 3 very small. The point about this system that's at 4 Hatch is that it saturates at .006 gallons per 5l minute. So any signal, whether it's from the crack 6 or whether it's from anywhere in the system, will 7 saturate the existing electronics and prevents any 8 quantitative information to be acquired. This is t
, 9 one of the limitations of the existing system, that 10 there's not enough dynamic range. The thing 4
1 4 11 saturates so quickly.
j j 12 ; MR. S HEWMON : Is that inherent in the 13 phenomena?
14 MR. KUPPERMAN: It's inherent in the 15 electronics that are used that you can buy cheaply 16 at this point. What is available and what is 17 inexpensive does not allow you to get information --
18 significant quantitative information. I mean that 19 sensor would be saturated by a valve leak far away.
20 And they wouldn't know is it from a valve far away 21 that's been on or is it actually from that area that 22 they're monitoring. They have no way of knowing 23 that easily.
24 MR. BENDER: Well, that's most of what the
(])
25 issue is, if you make the detection system too
._. _- . . . _ . _ . ,__ -._.~
4 f
287 1 sensitive then it's going to give you a lot of falseJ O 2 signals.
3 MR. KUPPERMAN: It's okay to have it 4 sensitive if you have the capability to locate the 5 source, the position, and that's what we have 6 addressed.
7 One of key elements in estimating the 8 sensitivity of an acoustic system is acquiring '
9 acoustic background data, and we've acquired it not 10 only from Hatch but also from Watts Bar. I don't 11 have a view graph, so I'll have to pass around this 12 photograph.
() 13 We have gone out to Watts Bar and acquired 14 a considerable amount of acoustic background data, 15 noise data. And with that information, with the 16 information regarding the signal amplitude versus 17 flow rate, the ultrasonic acoustic attenuation data 18 that we've acquired and all the work we've done with 19 insulation and so on, we end up with an ability to 20 make a prediction regarding the sensitivity of 21 acoustic systems that might be installed in the 22 field.
23 There's a considerable variation in the )
i
{} 24 acoustic background noise in reactor, not only have 25 we found this, but also the Germans report similar l l
l
4 288 1 results. And this means that when one tries to 2 estimate the leak rate one has to -- sensitivity, i
has to look essentially at different. background j
3lone 4 noises, and that's what this slide shows. This 5 indicates the signal to noise ratio as a function of
} 6 distance, the separation between the sensor and the i
i 7 source of the leak as a function of -- I'll use this 4
8 one, as a function of the leak rate going from .01 9 gallons per ninute to one gallon per minute, and 10 also this indicates the variation in the signal 11 under those conditions in which the leak could be 12 hitting reflective or metal insulation and enhancing 13 the signal.
14 So you can see that in low background 15 noise you have very sensitive -- you have the 16 capability to detect even at three or four. meters 17ll leaks on the order of a hundredth gallon per minute.
18 But in high background noise, even a leak l
19j of one gallon per minute might be difficult to 20 detect. So there's considerable variation, and it 21 depends where the leak is which will determine how e
- 22 small of a leak you can find.
l 23 MR. ETHERINGTON: Do you visualize this as 24 a permanent installation or
(])
t --
25 MR. KUPPERMAN: No, permanent installation, v-
,-, ,y. --- ,v -<-y %-f ,7 -,y y. , - - -- ,..-..,,r ,m ,- - -w- , -
.m , .m.- , -
-m e, y
289 1 right, with self-calibrating, self-checking 2 electronics and so on.
3 MR. ETHERINGTON: You're going to have to 4 have a lot of these sensors then.
j 5 MR. KUPPERMAN: On the order of maybe 20 l
6 to 40, something in that area.
7 MR. SHEWMON: Is that for one piping 8 system or for the whole plant?
9 MR. KUPPERMAN: No, for the whole plant, j i
10 20 to 40 I would say for the entire plant should be 11l sufficient to detect one gallon per minute. You 12 would need more sensors in areas where the I 13 background noise is high, the sensors would have to l
f 14 be closer together, but you could have them much 15 further apart in the area where the background noise 16 was low.
17 One key aspect of advancing leak detection 18 l technology is to identify the location of a leak, 19 and we have looked into advanced signal processing 20 i cross correlation technology where we get 21 information from two receivers that straddle the 22 source of the leak, the source of the acoustic noise, 23 and I'm not going to go into the details of this
(] 24 technology. This cross correlation technique allows 25 one to generate an amplitude-independent signal that
1 290 1 indicates the location of the source of the noise, 2 and this slide just shows some experimental data 3 that shows that as we --
we have a receiver on the 4 end of our pipe run, one on each end, and we're 5 moving an electronic, a simulated leak signal around l
6 the pipe, and you can see that the peak of this 7 correlation function tracks the movement of the 8 simulated leak source. So this concept allows one >
9 to got information about the source location.
10 We've also looked at --
11 MR. BENDER: Excuse me. To do this would '
12 you have to pre-calibrate the system or could you --
j 13 MR. KUPPERMAN: You do not have to 14 calibrate the system to do leak location by this 15 technique. You only have to have a signal above 16 background to do it.
17 MR. S HEWMON : And how does it work?
18 MR. KUPPERMAN: You correlate -- you i
] 19 capture a window in time of the acoustics. You just
- 20 literally try to match the two wave forms. If the 21 two wave forms --
if the leak is exactly in the j
22 middle of the two receivers, then the wave form at A
]
23 will look exactly the same as B. They'll just fit 24 over each other. HI f they're shifted, if the leak
(])
! 25 source is not in the center, if it's moved off to
. , - - - - - , - , ,4
,- , , --.,.,-.~a.,-. , . . - . , , . - , , - - . -
i 291 l 1 the side, then you will have to electronically 2 translate in time that wave form until it matches 1 3 what is received at the other one. You try to 4 match -- imagine that the leak just --
5 MR. SHEWMON: So the two signals are 6l coherent in time?
7 MR. KUPPERMAN: To a certain extent. They 8 look the same to a certain extent. They don't look 9 exactly the same, but what is similar is the 10 positions of these crossing points as the wave 11 oscillates. And you match them up, and you shift 12 the signals in time to match them up, and that time
() 13 shift tells you how far away from the middle point 14 it is.
15 Those are simulated leaks. We also looked 16 at the IGSCC, and it does show that even with a very 17,small leak rate in the order of a few thousandth 18 gallons per minute you can get correlation of peak.
19 We've done some investigations on a computer to try 20 to simulate these kinds of curves, just to see if we 21 can, you know, if we can build a model that will 22 simulate data from different sizes of pipes and so 23 on, we could obviously save a lot of time and can 24 avoid a lot of experimental work.
{])
25 And this just shows you that we can't --
i
. , - - , _ __ - . . , . , , , - . .-,n, , - - -
i 292 1 we understand enough about what. kind of signals that 2 we've generated by the leak so th a t we can model it 3 and reproduce in a general way the kinds of 4 correlation' functions that we get from IGSCC. So we j
5 are able now on an IBM PC to do some simulation 6 studies regarding this technology.
! 7 To go out into the field requires more 8 work, and one thing in particular is related to the 9l design of the sensor --
the system that collects the 10 acoustic signal, and we put together a relatively
- 11 simple system that we now have used in the field I
12 during the acquisition of acoustic background data 13 and some experiments that we did at Braidwood.
i 14 MR. KERR: What is the wave, juct a solid 15 piece of metal?
16 MR. KUPPERMAN: Yes, this is a solid rod 17 on the order of three milli me te r s in diameter, which 18 threads through a plate which is strapped down to l
4 19 the pipe, and we just put enough pressure on the
- 20 pipe so that it's in good contact.
21 Then there's a commercially available type 22 transducer which is attached to a plate which is 23 welded onto the rod, and this allows us to collect
{} 24 the acoustic data that we need.
25 The work at Argonne has led in i
1
293 1 collaboration with GARD, Incorporated, which is a 2 company near Chicago, a system which can fit into a 3 volume on the order of the size of a desk that can 4 be taken out to the field. This allows us to 5 capture and store information from the reactor, and e 6 ' we can bring it back to the laboratory and do the 7 analysis or you can do the analysis in the field.
8 But this is a system that allows us to carry out 9 experiments in the field.
10 And the most recent work that we have 11 ' carried out has been at Braidwood where we did some 1
12 ; experiments where we simulated leaks electronically, 3
f~J 13 and again I have a photograph but no transparency of 14 what it looks like. There's a probe attached to the 15 pipe. There's no insulation in this case. They're 16 under construction. Here's the electronic system.
17 There I am. And this is, you know, about 30, 40 18 feet away from where we detect the signal.
19 The most significant -- we acquired a lot 20 of data at Braidwood, this Commonwealth Edison plant, 21 but the most significant is that we have 22 demonstrated that we can acquire a cross correlation 23 peak which allows us to locate the source of the
(']
L 24 acoustic signal, this peak here, which is off center.
25 Here, this is the center here. This is off center.
294 1 This peak was acquired despite the fact that there 2 was a valve between the receivers.
3 Now, the conventional technology could not 4 have located this source of acoustic wave because 5 the presence of the valve would distort the 6 amplitude information, but this cross correlation 7 technique is independent of amplitude and allows us 8 to locate leaks even though there's some geometrical 9 thing in the way, like a valve or a T or something 10 like that. So this is a considerable accomplishment 11 in advancing the capability to locate leaks. l 12 ; Let me just summarize the work that we've 13 done by saying that the leak location capability has 14 been improved by advanced signal processing. We've 15 demonstrated that, specifically the use of this 16 cross correlation technique. Leak characterization 17 ;l has been improved by spectral analysis in general.
18 We've tested the computer based system under field 19 conditions. The system survived the rough 20 environment, and we were able to acquire some 21 significant data.
22 We are working --
we had been working at 23 Watts Bar to monitor that plant when and if it ever 24 goes on line with the ANL/ GARD two channel acoustic
(])
25 monitoring system. In collaboration with P and L,
=
~_
295 1 we have sensors there already installed. And that O 2 is on hold at the moment.
3 MR. KERR: What is the likelihood that 4 you'll detect a leak when one doesn't exist?
5 MR. KUPPERMAN: Okay. That's actually a 6 very good question, and it's hard to answer because 7'if you say I must detect a leak that's one 8 thousandths of a gallon per minute, that means the 9 sensitivity of the system is, you know, at its limit.
10 And signals that come into the transducer might look 11!like a leak. If you say I only have to detect two 12 gallons per minute, I would say it's very unlikely 13 that you would have a faire alarm.
14 MR. S HEWMON : How about a tenth of a 15 gallen a minute? -
l 16 l MR. KUPPERMAN: I don't know. I don't i ,
I l 17 know how to answer that question.
18 j MR. SHEWMON: See, people decide they can i
19 I live with a few gallons a minute. So you need an 20 order of magnitude on that with good reliability.
21 MR. KUPPERMAN: Well, it's the best system 22 that's around right now to minimize false alarms.
23 There's more information in acoustic signal than
{} 24 anything else that's around.
25 I think that, you know, you would have --
i l - - - . . . - - . ,-,.-.-- - -
296 1 it depends. It just depends on so many factors. I 2 don't know how to answer you. It depends on 3 background noise. If the background noise is low 4 and you're monitoring a specific --
if you think 5 that there's a problem at a specific weld, I would 6 say the probability that you would have a false 7 alarm is virtually zero.
8 First of all, there haven't been any, and 9 there are four reactors that are monitored. There 10 have been false alarms with the moisture sensitive l
11 tape.
12 MR. KERR: I don't understand why the fact 13 that you think there's a problem at a weld makes a 14 false alarm less likely.
15 MR. KUPPERMAN: Because you can straddle 16 the weld with two sensors and lock in on a certain i
17 spatial area.
18 HR. KERR: But if you're going to get a 19 false alarm, I'm assuming now that the alarm does 20 not come from the weld, it comes from something else.
21 MR. KUPPERMAN: The false alarm would have 22 to be the result of an acoustic wave hitting the i 23 receiver. Electronically I don't think you would I
l 24 have a problem with that. But the ability to locate
{])
25 through the correlation analysis tells you if the l
297 1 signal is between those two sensors which could be O 2 l one meter apart straddling the weld. If it's l
3 !between those two sensors, I would say it's a very 4 lhigh probability that it's a leak.
I 5; MR. S HEWMON : If somebody decided they 6 wanted to buy one of these things and put it on a l
7jsteam line, how would they go about it? I mean is 8 this sort of each a one of a kind? Can they buy it 9 someplace? Would they go to GARD and say, would you 10 design me this system?
11 MR. KUPPERMAN: I'm sure GARD with Argonne 12ltogether would --
specifically go to GARD and
(i j b> 13j probably with us, we would have a system put l
14ltogether. That system that I showed you is --
it's 15 considered a bread board. It looks like it's i
I 16 finished, but there's a lot of work that still has l 17 te be done, and I'm indicating this on this last 18 slide here. i 19 MR. BENDER: You told us you were going to 20 say what kind of --
I don't know whether the word 21 was uncertainties. I'll use it anyhow. What kind 22 of things needed to be investigated in order to get 23 to a position where you could say it was a good idea 24 or not a good idea to put 20 sensors on in a nuclear
(])
25 power plant.
4
298 1 MR. KUPPERMAN: What we need to do -- I 2 could indicate quite easily here. What really needs l 3 to be done is complete the software development.
1 4 The computer based system is not optimal. It's very i
5 difficult to work with. There's a considerable 6 amount of work there. That would improve the 7 reliability.
- 8 We have not carried out enough experiments j 9 with large leaks, and we're looking at very small 10 leaks, and we still are not sure how to extrapolate
, i f 11 i all this information to large leaks. We have some I
12 data --
[ () 13 MR. BENDER: What's the leak range we i 14 really want to investigate?
i 15 MR. S HEWMOL4 : You tell him.
16 MR. BENDER: No, he has to tell me because
, 17 I --
if he doesn't know --
1 18 MR. KUPPERMAN: I know that we can detect i
1 19 a hundredth of a gallon per minute. Now, you know, 20 one gallon per minute -- or I should say certainly 21 five gallons per minute in a BWR I think is much too 22 large.
23 MR. BENDER: When somebody tells me what 24 leak is associated with what size crack, then I'll
(])
25 be able to say what leak I want to detect, and then
.i
i 299 1
. 1 I'll have to deal with the question of what things 2 affect the leakage itself, the corrosion --
- 3 MR. KUPPERMAN
- If you knew that there was 4 a crack in a pipe, would you let the plant run? No.
i 5 So that means that even a leak that's a few 4
- 6 thousandths of a gallon per minute should be i
7 detected because that's enough to shut the plant
- 8 down.
9' MR. BENDER: Well, I don't know what the 10 measure of that size leak means in terms of that 4
11 crack. That's all.
12 MR. KUPPERMAN: Any leak in a crack is
() 13 enough to shut the plant down.
14 MR. MAYFIELD: I think the question he's 15 trying to come to is what size should we-detect, and 16 what size crack night come through the wall and then 3
I I
17lwe would want to detect that. I don't think there's l ,
{
10' any argument --
- 19 MR. BENDER
- I've got to deal with it this
) 20 way. I suspect there will be some sensitivity-l 21 that's too sensitive, that will give signals I don't l
l 22 want.
23 MR. KUPPERMAN: How can it be too 24 sensitive if it tells you you have a leak in a crack?
(])
25 MR. BENDER: Because it may detect leaks 1
4 n -
,-,---,-,.-,,,,--,---,r.,--.,,
a -n- , ,e ~ , . , - -- .,a- ,c- ,v -
>-w - - . , , - - - = , <,,~---<,e
= _ _ _ _ _ . _ _ . . . _ . _ . _ _ _ _ . _ _ _
t a
J 300 J
i 1 that don't exist.
)
2 MR. KUPPERMAN: Well, that's a problem of j 3 false alarms, and it's not clear at this point . what 4 the ultimate reliability of-the acoustic system is. '
f 5 I can't answer that question. There's no way that 6 question can be answered at this point. We~have no --
l 7 we don't have enough field data.
f 8 MR. BENDER: What am I going to do to get 1
! 9 the field data?
10 MR. KUPPERMAN: We are going to monitor j 11 plants. We were going to monitor plants. We don't 12 have the money to do this.
j CD 13 MR. BENDER: We'll have to have plants 14 that will eventually have some leaks in them, and if !
l 15 you're not going to have any leaks, then I've got to i; 16 decide --
}
i 17 MR. KUPPERMAN: No, that's not true l 18 regarding false alaras, because all plants have >
- 19 valve leaks and flange leaks, and those leaks would l
20 be detected throughoutLthe year.
I 21 MR. BENDER: .11 leave --
my professorial f l
22 friends always say that's an exercise for the l
\
23 students.
24 MR. KUPPERMAN: That information would be
[)
25 acquired when there's a valve leak and a flange leak.
- _ _ .- _ _ ~ . ___. __ __ - . . - _ _ _ _ _ - -
~ -
l l
l 301 1 We'll be acquiring that data in the field.
O 2 HR. BENDER: Okay. I've asked enough.
3 MR. HU TC HI NS ON : You expect your system to 4 able to discriminate between a valve leak --
5 MR. KUPPERMAN: And a crack, absolutely, 6 that's the whole objective. That's the whole 7 objective. And as we acquire data in the field, as 8 valves leak --
from reading the licensing reports 9 you've got little pipes cracking all the time, 10 you'll probably be able to pick one up_in a year or 11 two. You know, you'll learn how well it works. But 12 there's no way we can say now.
13 MR. HU TC HINS ON : But you're also saying --
14 I think you're saying now that if the leak turns out 15 to be too big,.maybe half a gallon a minute, the i-16 crack leak turns out to be that big,.that may 17lj confuse your system.
MR. KUPPERMAN: Not necessarily.
18 l 19 MR. KRAMER: What he's saying is he won't 20 know the difference between a tenth of a_ gallon and 21 three gallons perhaps.
22 MR. KUPPERMAN: Well, a tenth of a gallon l
23 and one gallon would be hard --
we're reasonably 24 confident you can get the order of magnitude of the
{])
25 leak rate, the order of magnitude at this point and
302 1 if the leak is very small and detected it would be O 2 even better.
3 MR. HUTCHINSON: You're not worried about 4 missing a leak of a gallon a minute?
5l MR. KUPPERMAN: It's possible in a high
! It's 6l background -- very high background noise area.
a 7 possible because you have to be within a couple of 8l meters.
9 MR. KERR: I thought from what you had 10 ; shown and some of your comments that these leaks did i
11 !
have characteristic frequency spectrum, so that you 12i could probably identify a leak, and you didn't have
() 13 much possibility of a false alarm. What you seem to 14 be telling me is that you don't have enough data yet 15 to be sure of that.
16 MR. KUPPERMAN: We don ' t have enough data i
17 to be sure of that, that's correct.
18 MR. SHEWMON: It seems to me that one of 19 the things that would push --
may push people to get 20 into this business would be your steam line argument, 21 which seems to me quite cogent because their people 22 are talking --
23 MR. KUPPERMAN: It's much simpler --
l 24 MR. SHEWMON: Let me finish. May I?
(])
25 MR. KUPPERMAN: Excuse me. I'm sorry.
1
_ _ . _ _ _ _ _ _ . _ . ~ . _ _ _ _ . _ _ . . _ _ . . - _ . _ _ . _ _ _ . . _ _ _ _ __ __ _._,_____ ..
1 i ,
303 4
! 1 MR. SHEWMON: Because their people do have l 2 this WHIPJET project, and there is a fair amount of 3 concern, and since I'm not talking to, I'm talking :
i 4 to the people behind you, I guess, there is a fair 5 amount of concern about whether or not there's any 6 way -- currently none of the current techniques will j 7 detect things out of containment. Yet. people are --
i 8 there's a fair motivation to not have to analyze for i I
9 arbitrary breaks out there. You know, there may be i
j 10 advantages inside, whether or not there's motivation j 11 to find leaking flanges and valves this way.
12 I do not share your idealism I guess with 1
() 13 regard to any operator is just going to be delighted
- j. 14 and shut his plant down right away if you show him 1
t l 15 there's any crack leaking through his pipe wall. I 16 suspect there may be some people who say, if it's 5'
17 not big enough to violate the tech specs and we're I
j 18 making kilowatts, I don't want to know about it.
I 19 When it gets dangerous, then soon enough. But 20 outside there is a motivation there, as I read it.
l 21 Are there other comments or' questions?
i 22 MR. RODABAUGH: Your little passing remark 23 about a crack in a small line,'if you had your 1
{} 24 sensor on a large line and a very small branch line 25 was coming off which is, as you said, where most
, l I
304 I
1 leaks occur, would your sensor pick that up and ;
l 2 identify the location?
I 3 MR. KUPPERMAN: That's a tough question to 4 answer. Certainly, if there's an acoustic path if 5 there's continuous metal from the source of the leak 6 to the sensor, we would detect the signal.
7 MR. RODABAUGH: You would hope to catch a 8 small branch line leak?
I 9 MR. KUPPERMAN: It's certainly possible.
10 I'm not that familiar with the configuration. I do 11 l want to point out that there's no funding to 12 continue work anywhere in the United States of any 13 significant size on acoustic or advanced technology, 14 none at this point.
15 MR. BENDER: Well, setting aside the fact 16 that you all aren't funded, the issue that's out 17 there is the one that Dr. Shewmon mentioned, namely, 18 Beaver Valley is making a case of some sort. If 19 they can make the case without the benefit of this 20 gadget, great. If they can't, somebody might have 21 to argue that it's a new gadget.
22 MR. SHEWMON: They aren't going.to make it i
23 on anything outside containment from anything I've 24 heard so far.
{])
25 MR. MAYFIELD: The last meeting I was at d
1 l
305 1
1 with Beaver Valley they were talking about an 2 infrared sensor to look for hot spots in the 3 insulation. They were also talking about operator 4 walk-down. We're not talking about using this kind 5 of system.
6 Let me take that a step further. There 7 were some qualifiers that, well, in some instances 8 we might have to do something more sophisticated and 9 it was left at that, but they were principally 10 talking about things other than this kind of 11 technology.
12 MR. KERR: They couldn't talk about using 13 this system because this system in the available 14 form didn't exist, as I recall.
15 MR. MAYFIELD: I mean the acoustic 16 technology. The specific system, no.
17 MR. SHEWMON: Okay. -Why don't we take a 18 short break at this point. We're running late.
19 (Short recess taken.)
20 MR. S HEWMON s. Okay. The next effort is 21 Omas Chopra from Argonne.
22 MR. CZDPRA: We are interested in 23 investigating the degradation of cast stainless 24 steel under the light water reactor conditions. We
[])
25 know that the toughness of these cast stainless
306 s 1 steels can decrease drastically under these light 2 water reactor temperatures. For example, the room 3l temperature Charpy impact energy can decrease from l
4iabout two to three hundred joules down to less than i
t 5 ! 50 joules in eight years at 300 degrees centigrade.
I 6fSothe objective of our investigation is to study l
7' the significance of this ombrittlement problem and 8 obtain the mechanical property data required for the l
- 9i safety analysis of reactor components.
I i
The scope of --
10 l I
11 j MR. SHEWMON: We'rc looking for a hand-out.
i 12 Do you have one?
13 HR. C HO P RA : They are all gone.
i 14 l MR. S HEWMON : Okay. Fine. We'll pay 15 attention to the screen.
16 MR. C HO P RA : The scope of the l 17 investigation is to characterize the microstructure 18 of long term in-service reactor components and low 19 temperature laboratory aged samples. This 20 information would give us an idea of the processes 21 which occur at the reactor temperatures, and also 22 see that the processes which occur in the laboratory 23 aged samples are, in fact, the same which happen at 24 the reactor temperatures. So that we want to make
{]} i 25 sure that we are looking at the same process. i i
I
l 4
307 1 Then to correlate this microstructural 2 changes with loss in toughness. That means we 3 identify the mechanism of embrittlement and to check 1
i r 4 the validity of extrapolating the experimental data l 5 to reactor conditions. l l
6 Also, another aspect of the program is to j 7 characterize the loss of fracture toughness in terms 8 of fracture mechanics parameters such as the JIc and 9 J-R curves, also to obtain the tensile properties of 10 these age samples and changes in the tearing models 11 of these materials.
12 MR. ETHERINGTON: On the first line, how 13 do you get low temperature laboratory aged specimens?
14 MR. C HO P RA : We were fortunate to get 15 samples from the Swiss company George Fischer. Most 16 of the available information is from their study.
17 They have aged samples up to eight years at 300 18 degrees centigrade.
- 19 MR. ETHERINGTON: Oh, um-hmm.
i i
l 20 MR. C HO P RA : And, in fact, there are now 21 longer aged samples available.
22 MR. ETHERINGTON: What temperature is low j 23 temperature?
{} 24 MR. C HO P RA : 300 C. And the last aspect 25 is to provide understanding of the effects of
308 1 compositional and metallurgical variables on the O 2 kinetics and degree of embrittlement.
3 By compositional parameters, I mean the 4 changes in the chrome, nickel, moly content, and 4
5 also interstitial elements such as carbon and i 6 nitrogen. Metallurgical variables are basically the 7 amount of ferrite and the morphology and
, 8 distribution of ferrite, which can change by a 9 difference in the cast process or the thickness of 10 these components.
11 This information gives us an idea to 12 define the parameters, or rate the regime of these
(*
13 parameters which would lead to embrittlement of 14 these materials.
15 Now, we have several cast materials. We J
< 16 got the aged sample-from the Swiss study, George 17 Fischer Company in Switzerland, five different heats, 18 and these samples were aged up to 70,000 hours0 days <br />0 hours <br />0 weeks <br />0 months <br /> at 19 these three temperatures,'300, 350 and 400 C. These 20 samples we are basically using for microstructure 21 studies.
, 22 Then we obtained the pump cover plate from 23 the KIB reactor in Germany, which was in service for
{} 24 12 years. The actual time at temperature was eight 25 years. And we are using this both for l
i 309 g3 1 microstructure studies as well as mechanical testing.
O 2 We got six large heats. They are static 3 cast slabs three inches thick, and they are 4 different compositions. They are three different 5 grades, the low carbons, you have three grades, and 6 the CF-8 which is the higher carbon, and the moly ,
7{containing grade CF-8M.
8 We also have six heats from reactor i
9 components. They are four different pipes of pump 10 cover casing -- pump casing --
no, I'm sorry, pipe 11 casing and impeller. The pipes are centrifuga11y 12l cast, and the other two are static cast components.
(3 ss 13 These are in the unaged condition.
14 The large heats and the reactor component 15l materials we are using for the fracture toughness 16 test and also to obtain the shift in transition 17 ' temperature, Charpy impact tests.
18 Then we have 19 small heats. They are in 19 the form of key blocks, and the composition again is 20 varied over a larger range to give us ferrite levels 21 between three and 30 percent. All these 22 compositions are within the ASTM specs.
23 MR. SHEWMON: Do you have any subarc welds?
24 MR. CHOPRA: No, we have no welds in this
(]
v 25 program at the moment.
310 1 MR. S HEWMO N : Do you know whether they 2 deteriorate from their initial miserable condition?
3 MR. SHACK: We do have welds. The welds 4 are being aged. Nothing has-been done with them.
5 They're cooking at the moment.
6 MR. S HEWMON : Okay. But they're in the 7 program.
i 8 MR. S HAC K : They're in the program. We 9 don't have enough welds perhaps, but we do have some l 10 weld material from a surge line.
1
- 11 MR. SHEWMON
- Fine. l 1
12 MR. C HO P RA : The question is to first make 13 sure what conditions would give us a large effect, 14 and then try to concentrate on those compositions,
, 15 those welds.
l 16 MR. SHEWMON: The purpose is also to try i
17 to predict what happens to components in plants, and 18 if welds are some'of the components in plants, then i 19 that's part of what you would like to predict.
20 MR. CHOPRA: Yes. The ASTM specs for the i
21 three grades are given in this view graph, the low 22 carbon which is .03 max carbon, the CF-8 is .08
't.
23 maximum carbon, and the moly containing has the two 24 percent moly, and the rest of the compositions are
(])
25 given in this table.
+, . - - - , .. --
. . , _ _ , , , _ _ , , - . _ , _ _,_,,__._,,,-__,,,,,mm- ,._ m,._.,_..-,,,_m.---,,,_,.
311 1 HR. SHEWMON: Somebody yesterday, maybe 2 Hays, had a 308 --
that was a 308L though, not --
3 MR. SHACK: Right. That's the weld metal.
4 l MR. C HO P RA : Weld metal.
I 5
MR. S HEWHON : Okay.
6! MR. CHOPRA: The CF-3 is the 304L wrought i
7l material, CF-8 is the 304 standard, and the 316 is 8 the CF-8M.
9 MR. SHEWMON: Okay.
10 ;I MR. C HO P RA : Now, if I plot these ASTM l
11 composition specs on a Schaeffler diagram which I 1
12 lgives the chrome and nickel equivalent, we have the r~g kJ 13 austenitic region here, complete ferritic steel here, 14land this is the duplex which is zero percent ferrite, l
15 hundred percent ferrite, and I plotted one line 16 which showed the ten percent ferrite. And these 17 blocks are the ASTM specs, and the circles show the 18 compositions which we are looking at, and you see 19 that they cover the ASTM specs, quite well.
1 20 The squares are the compositions which '
21 were used by the Swiss study, George Fischer people.
22 A lot of these countries are outside the ASTM specs, 23 and the other thing is that they cover a very narrow 24 range, and the existing correlations to predict the
{])
25 embrittlement problem is based on that data. Since
312 i
1 they cover such a narrow range of compositions, they 2 may not show us the effects of certain parameters 3 since they are covering a.very small range.
4 MR. SHEWMON: What's the hashed square 5 about?
6 MR. CHOPRA: The dotted -- yeah, this is 7 the CF-8M, the moly containing specs.
8 MR. SUEWMON: And the solid line?
9 MR. C HOPRA : Solid is the CF-3 and CF-8 10 grades.
l 11 l MR. ETHERINGTON: The shift of the diagram 12 is for weld deposit. Would you expect the cast pipe
( 13 to conform?
14 MR. CHOPRA: Using this you can predict 15 the amount of ferrite based on the composition in 16 any cast.
17 MR. ETHERINGTON: The rate of cooling is 18 much different; isn't it?
19 MR. C HO P RA : Yes, there are slight 20 variations, but that gives a difference in the 21 morphology of the ferrite.
22 MR. ETHERINGTON: I see. That's what I 23 was really asking ---
I was really asking the '
24 question. You're saying you can.
(])
I 313 l l
1 MR. ETHERINGTON: We were interested
(
()> 2 recently in some micrographs of centrifugally cast 3 pipe, showed a columnar dendritic structure on the 1
4 joutside and equi-ax on the inside. Do you know 5 anything about the casting process?
6: MR. C HO P RA : Is that the CF low carbon 7 grade?
l 8l MR. ETHERINGTON: Probably was. Is the 9 mold lined or not in the --
10 MR. C HO P RA : No, I think the shift in the 11l macrostructure, what people believe now, the study 12!at CEGB, they have noticed that in their cast which 13 they are looking at. What they believe is happening, 14 the low carbon grade are in a region where from the 15 melt you form gamma rather than delta.
16 So the formation of gamma right from the 17l melt changes the macrostructure as well as the i
18 morphology of ferrite, because in one case it's a ,
19 l liquid phase diffusion transformation, and in the l
20 other case it's a solid / solid, the delta changing to 21 gamma later on for most of the cast, whereas that 22 shift from the surface to center, the dendritic to a 23 more equi-ax structure, is because of the
/~} 24 composition of that casting.
\s 25 MR. ETHERINGTON: Is the mold lined or not
I 314 1 lined?
2 MR. C HO P RA : I don't know if --
3 MR. S HEWMON : This'is a centrifugal 4 casting, so it has to be something out there to -- I 5 don't know what you mean by lining, I guess.
6 MR. ETHERINGTON: Well, is there 7 refractory lining or not? Was it cast against the 8 chill?
9 MR. CHOPRA: I don't know. Most of the 4
10 time I think it's just against the chill.
i 11 MR. ETHERINGTON: See, that would account 12 for this kind of structure you get in an ingot.
) 13 MR. S HEWMON : Well, I guess I don't know 14 which end -- my impression was that the finer grain 15 size was on the inside.
l 16 MR. E T HE RI N G TON : It was.
17 MR. SHEWMON: Well, then the chill is on 18 the outside.
19 MR. ETHERINGTON: Yes.
20 MR. S HEWMON : So any chill effect would 21 produce the fine effect on the outside.
22 MR. ETHERINGTON: - Not in an ingot. In an 23 ingot you get-the same thing, the dendritic 24 structure, and then when it pulls away it's equi-ax j
(])
3 25 inside.
1 r,_-..= _ _ ~ . , . . . . , , _ . - . - . _ _ . _ , . , , _ . . - , . . _ . _ , . __...._.__.._.m.__.,,, . , _ , _ , - . . . . . . . . , _
i 315 1 MR. SHEWMON: Yes. I don't know. Okay.
2 MR. JASKE: On that question, 3 centrifugally castings are typically made, there's a 4 steel outer mold and then it's sand.
5 MR. ETHERINGTON: Sanded inside?
6 MR. JASKE: There's sand on the inside.
7 You inject the liquid molten metal with a nozzle in 8 the casting. The whole mold is spinning while this 9 process is going on. It's cooling from the outside 10 or solidifying from the outside in typically.
11 We've looked at ring sections of Folson 12 centrifugal castings, duplex stainless steels and 13 various kinds, and seen that sometimes you get the I
14 dendritic structure on the outside as you talked 15 about, equiax on the inside, and sometimes on the 16 same heat. The material cast in subsequent tubes, 17 you'll get the opposite effect. So it's not clear.
18 MR. CHOPRA: It's more complicated than --
19 MR. JASKE: It's a very complicated 20 phenomenon. It's not too well understood, but 21 typically you expect the columnar structure near the 22 outside diameter, an as cast structure.
23 MR. CHOPRA: Also, to give an idea of what
() 24 compositions the Framatome people and the EDF people 25 in France are looking at, they have published the
,~ -
~r - - - - , -
- , ., , - - , n-~ ,---w,-n., -
, - - - + -- ,~ - ,
r-
316 1 work. The circles and the squares give the 2 compositions studied by the Framatome and the EDF 3 people, and again they cover a very small area in 4 the ASTM specs. Now, if we combine our compositions 5 with the George Fischer Company results, I think we 6 can get good correlations and the effects of 7 different parameters on this embrittlement behavior.
8 The time and temperature of aging for 9 these cast materials, we are aging them at five 10 different temperatures from 290 to 450 and the times 11 range from 100 to more than 50,000 hours0 days <br />0 hours <br />0 weeks <br />0 months <br />.
12 We have completed aging up to 10,000 hours0 days <br />0 hours <br />0 weeks <br />0 months <br />.
13 In fact, now some of the samples have completed 14 21,000 hours0 days <br />0 hours <br />0 weeks <br />0 months <br />, and the large heats have completed 15 11,000 hours0 days <br />0 hours <br />0 weeks <br />0 months <br />.
16 These A, B and C are different mechanical 17 tests. A are the Charpy impact tests at room 18 temperature. B are Charpy tests to determine the 19 shift in transition temperature, and also the J-R 20 curve tests using a 1-T compact tension specimen, 21 both of the tests at room temperature as well as 290 22 C. The C condition are tests --
the J-R curve tests 23 using 2-T samples to study the side effect on the 24 fracture resistance behavior.
{])
25 Coning to the results, mi c r o s t r u c t.u r al
317 1 changes in these low temperature aged materials, the 2 Swiss materials, like I mentioned, they have been 3 aged up to 70,000 hours0 days <br />0 hours <br />0 weeks <br />0 months <br />. All of our samples we 4 examined using three different techniques, the 5 transmission electron microscopy, atom probe feed 6 line microscopy and small angle neutron scattering.
l 7 The atom probe work was done by Oxford University, 8 Professor George Smith, and the small angle neutron 9 scattering we have done both at ANL and at Oak Ridge.
10 Now, all these Swiss samples show 11 fluctuations in chrome content between 15 to 40 12 atomic percent in all the samples, even the samples
, 13 which were aged for 3,000 hours0 days <br />0 hours <br />0 weeks <br />0 months <br /> at 300 C.
14 Now, we observed two other phases by TEM 15 work, the G phase which is a phase rich in nickel j 16 and silicon, and the Type X phase which we do not i
17 know the structure or the composition of this phase.
18 It's an unknown phase. And we observed these two 19 phases in samples which were aged for 70,000 hours0 days <br />0 hours <br />0 weeks <br />0 months <br /> 20 at 300 C or for mote than 10,000 hours0 days <br />0 hours <br />0 weeks <br />0 months <br /> at 400 degree 21 centigrade.
22 Now, these samples also showed the mottled 23 structure of the alpha phase. The mottled structure I
24
[]) is a very diffuse image on the TEM plate. l
{
25 MR. S HEWHON : Are those in the gamma phase l i
i
318 l
1 or the alpha phase?
()
2 MR. C HO P RA : They are all in alpha phase. I 3 What I am giving here is only the ferrite phase. In ,
l 4 gamma phase we do not see any change in the 5 microstructure.
6 Now, the small angle neutron scattering 7 results give us the size of these phases which form 8 in the ferrite matrix, and the results correspond 9 with the size of the G phase which we see in 10 transmission electron microscopy. I'll show you the 11; results of these different examinations.
l 12 j Now, the moly containing steel also has r
13 another phase which we call Type ML. Still we do 14 not know the structure of this phase, and it's 15 always associated with the G phase. It seems to 16 cover the G phase.
17 ' The KRB pump cover plate material, which 18 was in service at temperature for eight years, also 19i shows the fluctuations in chrome content, and this 20 we see by the atom probe study. And this material 21 has alpha prime, G and the Type X phases which we 22 can see by TEM study. In addition to these phases, 23 we see grain boundary carbides or phase boundary 24 carbides at the alpha / gamma boundary.
(])
25 Now, these carbides may not be associated
319 1 with aging, and I'll show that actually the carbides 2 are present in the as cast material. All the high 3 carbon, the CF-8 grades, have carbides in the 4 initial casting.
5 And we have also examined the ANL heats, 6 four of these heats, and these were aged up to i
7l10,000 hours at 350, 400 and 450 C. We see all 8 these phases, alpha prime, G and Type X, and the i
9 ' high carbon grades also show the grain boundary 10 carbidos.
11 To give an example of what we see, the G 12 phase, this micrograph gives a bright and dark field
() 13 image of the Swiss heat 280 which was aged for about 14 70,000 hours0 days <br />0 hours <br />0 weeks <br />0 months <br /> at 400 C, and we see 'a fine 15 distribution of the G phase throughout the ferrite 16 matrix.
17 This micrograph gives the top micrograph 18 for the KRB pump cover material, and this shows the 19 alpha prime, the mottled structure. You see the 20 fine structure, and the larger contrast is not from 21 the microstructure in the material, but it's due to 22 the lack of use to polish the sample, so we can 23 actually get rid of this larger contrast. The fine 24 mottled structure, which quite often is called
(])
25 orange peel structure, is due to the alpha prime
320 g 1 phase.
V 2 The bottom micrograph is for the ANL heat 3 60 which was aged for 10,000 hours0 days <br />0 hours <br />0 weeks <br />0 months <br /> at 400 C, and in 4 l this micrograph we see the phase boundary here and 5 the austenite phase and the ferrite phase. We do 6 not see anything in the austenite and the mottled 7 structure of alpha prime phase in the ferrite. And 8 at the boundary, you cannot see it very clearly here.
9 But you see these two large carbides. They are the 10 M23/C6 carbides.
11 l Now, the results from the aton probe work, 12lthese two curves show the fluctuations in the chrome 13l content in the ferrite matrix. This is for again 14 the Swiss heat 280, aged for only 3,000 hours0 days <br />0 hours <br />0 weeks <br />0 months <br /> at 400 15 C, and we see these large fluctuations. This is raw 16 data. But the distance of these fluctuations is 17 about ten angstrom. The actual composition of alpha 18 prine phase is 78 percent chrome, and we do not see 19 that high concentrations. They are maximum about 45 20 atomic percent.
21 Also, same heat, aged for 70,000 hours0 days <br />0 hours <br />0 weeks <br />0 months <br /> at 22 300 C also shows these fluctuations. In this sample 23 aged 3,000 hours0 days <br />0 hours <br />0 weeks <br />0 months <br /> at 400 C by transmission electron 24 microscopy we do not see anything, yet the impact
(])
25 energy of this sample decreased more than 50 percent. !
l l
l
- - - . 4
321 1 MR. SHEWMON: What's plotted along the 2 horizontal axis?
3 MR. CHOPRA: Actually these are aisles.
4 HR. SHEWMON: What?
5 MR. C HO P RA : Aisles. What you receive --
l 6 actually you collect -- you remove layer by layer.
7 j And the ions collected, if you can work back to the 8 rate of --
9 MR. SHEWMON: So the vertical variation is 10 noise and the mean would then be a percent chrome?
11 j MR. C HO P RA : No, vertical is chrome, and 12lthis is the distance as we go into the sample. You
() 13 l basically drill a hole l
14 MR. S HEWHON : I'm with you. Go ahead.
15 MR. C HO P RA : So we do see that although 16 this sample did not show anything in transmission 17 electron microscopy, we do see a reduction in impact 18 energy which may be due to these fluctuations in 19 chrome content in the ferrite matrix.
i :
20 Now, the next figures give the same 21 results for the heat 280, aged for only 3,000 hours0 days <br />0 hours <br />0 weeks <br />0 months <br /> 22 at 300 C, and we still see the fluctuations in 23 chrome, and the bottom curve is for the KRB pump 24 cover plate material, and again we see those
(])
25 fluctuations in that sample. But this sample, the
322
,, 1 KRB material, does show the mottled structure of N]
2 alpha prime in TEM examination.
3 Observing these fluctuations that after 4i 3,000 hours0 days <br />0 hours <br />0 weeks <br />0 months <br /> at 300 C indicates that the formation of l
5! alpha prime is not by nucleation and growth but I
6 occurs by spinodal decomposition because then there 7 is no incubation period, and actually these 8 fluctuations would increase by growth of these 9 regions to form the final alpha prime phase.
10 MR. HU TC HINS ON : For someone who is an 6
11 expert in these areas, are these comparisons you're 12, making absolutely convincing? I mean for someone 13 who is like me who is not, you know, I look at these 1
14 l two curves and they look like -- if you had 15Iinterchanged them, I would have been equally likely 16lorunlikely to have accepted your argument.
17 l MR. CHOPRA: No. These are the raw data.
t 18 They have done more analysis to show a better 19 comparison from one curve to another because I think 20 at present these results are based on different 21 voltages applied on the sample and the rates of --
22 so once they convert them in distance versus atomic 23 percent chrome, you can then compare, and it would
(~ ; 24 give you an idea the distances over which you see
]
25 these fluctuations. And in most of the studies done
323 1 either at Oak Ridge or by Westinghouro at the 2 University of Pittsburgh, these fluctuations range 3 over about ten angstroms, and that's what people 4 believe is the frequency of these alpha prime phases.
5 The results from this small angle neutron ,
6 scattering, here I've plotted the frequency of --
7 the relative frequency of particles of different I
8 sizes. Now, the top curve is for 278 Swiss heat 9 aged for 10,000 hours0 days <br />0 hours <br />0 weeks <br />0 months <br /> at 400 C, and we see that the 10 peak is at a size of about 16 angstrom. After aging 11 70,000 hours0 days <br />0 hours <br />0 weeks <br />0 months <br />, the peak shifts to about 50 angstroms 12 which shows that these sizes correspond to the G
() 13 phase, indicating that with time the G phase grows 14 in size.
15 We also looked at the fracture surfaces of 16 these Charpy impact test specimens, and we see that 17 for room temperature tests, the ferrite phase fails 18 by cleavage in the samples which were embrittled.
19 And fracture occurs preferentially along the ferrite 20 phase. That means even though the material may have 21 20 percent ferrite, the fracture may have 60 or 80 22 percent cleavage, indicating that the fracture jumps 23 from one island of ferrite to another.
24 We conducted tests at the liquid nitrogen
(])
25 temperature of minus 196 degrees centigrade to see I
324 es 1 if there are any other fracture molds which may
\
2; influence the overall fracture behavior, because at 3 this temperature the ferrite phase even in the 4 unaged sample would fail by cleavage.
5 And we do see that, that all samples 6 unaged or aged the ferrite phase, failed by cleavage, 7lbut we also see that the high carbon grades, all 8 conditions, unaged or aged, and the moly containing l
9 grade in the aged condition only show alpha / gamma 10 boundary separation, which indicates that carbides 11 are present in the unaged material. And in some 12 cases since we see the aged sample showing
) 13 alpha / gamma boundary separation which indicates that 14! the carbides may grow after aging at the high 15 temperaturco, and in fact we see that samples which i
16 were aged at 450 C failed predominantly by grain 17 boundary separation. I'll show you some micrographs 18 of these fracture surfaces.
19 This shows three different grades. The 20 top one is the low carbon grade CF-3, middle one the 21 CF-8 grade, and the bottom is the CF-8M grade. Now, 22 in this we basically see a ductile failure of the 23 austenite and a few regions where you see cleavage.
24 MR. ETHERINGTON: The white is the ferriter
(])
25 is it?
__ ..,.-._~_-1 -- --
325 1 MR. C HOPRA : Ferrite at lower temperatures -~
,G V
2 no, this is --
all these three samples were aged for 3 about 10,000 hours0 days <br />0 hours <br />0 weeks <br />0 months <br /> at 400 C.
4 MR. S HEWMON : Can you tell ferrite from 5 austenite in that by the fracture surface?
6 MR. C HOP RA : Let me show the next slide 7 and then we can --
yeah, actually if we take x-ray 8 analysis, we can by the difference in the 9fcomposition we can make out, ferrite has about 26 10 percent chrome, and the austenite has about 18 i
11 percent chrome. So if we measure the chrome content 12 ! on these surfaces, we can make out whether the grain rm I (J 13lis ferrite or austenite.
i 14 Same thing with nickel, in the austenite 15 nickel is about nine percent and the other one has 16 lower nickel.
17 MR. S HEWMON : Okay. Go ahead.
18 MR. ETHERINGTON: Doesn't the cleavage 19 area show differently in this micrograph? Is the i
20 cleavage area white, is that what I'm really --
21 MR. SHEWMON: You can't see it here. On j l
22 the screen you can see a small area of cleavage, and
?3 then he can go in with a microprobe and do a 24 chemical analysis on that and tell the phase.
(]}
25 MR. ETHERINGTON: So the fact we have some i
326 1 white areas and some black, that doesn't tell you 2 anything at all?
3 MR. C HO P RA : No. In this view graph you 4 can't make out much, but on an actual sample when 5 you're looking at it, you can see the cleaved I
6 l regions which have a typical river pattern. In some 7 cases -- out here you can see it has a structure 8 which is termed as a herringbone cleavage. You can 9 see the cleaved areas quite easily on the screen.
10 ; They have these two typical structures.
l 11 MR. ETHERINGTON: What optical effect i
12 makes these wide snaky areas then?
() 13 MR. C HO P R A : It's not an optical effect.
14 Actually they are steps on the cleaved surface.
15 MR. ETHERINGTON: They are steps on the 16 cleaved surface?
17 MR. CHOPRA: Yes.
18 MR. ETHERINGTON: I see.
19 ; MR. C HO P RA : Anyway; the low carbon grade 20 shows very little cleavage, whereas the higher 21 carbon in this micrograph it's mostly cleavage 22 surface, and this sample had only 21 percent ferrite, 23 and the surface is nearly 90 percent cleavage. You 24 can hardly see small regions of ductile failures.
(])
25 MR. S HEWMON : Would you try to speed up a
I
[ 327 l little bit through the micrographs. We're running
()
! 2 late.
+
3 MR. CHOPRA: Okay. I'll skip some.
l 4 Now, the Charpy impact test results, we
! 5 have completed Charpy tests.on the KRB material and 6 16 heats of the inner heats, which were aged up to 7 10,000 hours0 days <br />0 hours <br />0 weeks <br />0 months <br />, and basically the results show that 8 the embrittlement depends on the ferrite content of .
9 the material and the concentrations of carbon and 10 nitrogen in the material and also the distribution 1
11 of ferrite in the duplex structure.
12 MR. BENDER: How about answering a
() 13 question for a nonmetallurgist. Is there anything 14 in here that's surprising?
15 MR. CHOPRA: Yes. These two --
in the 16 existing correlations people have not included the 17 effects of carbon and nitrogen and the distribution 18 of ferrite. The correlations which were from :
19 Framatome people are based on the George Fischer 20 study, and they do not include these two parameters.
21 MR. BENDER: What am I to judge from'this?
22 MR. CHOPRA: Okay. The existing thing 23 they say that ay steel with.about 15 to 20 percent 24 ferrite may have a potential for embrittlement or
(
25 loss in toughness. Anything less than that is no
328 I l
fs 1 problem. What we are seeing that in certain l )
x; 2 structures, even at ten percent ferrite steel, can 3, show large reduction in toughness.
4 MR. BENDER: Now, can I convert that into i
5l the cast stainless steel that we're using in these i
- 6. piping systems and say "therefore"?
7 MR. C HO P RA : Okay. Most of the existing l
8! pipes have ferrite levels from about eight to l 9! fifteen percent and some have higher, but I think 10 i ten to fifteen is --
a majority of the pipes have 11 that. Now, some of these results suggest that a 12 casting with ten percent ferrite may have as much ,
(3, l (J 13 reduction in toughness as what the existing 14l correlations predict for much higher ferrite.
15 ! MR. BENDER: Okay. And when I learn that, I
16l then I ask myself what should I be concerned about? ;
17 Should I be concerned that cracks that we've ignored 18! before we now should be concerned about, or should I 19 , be concerned as to whether cracks exist or what?
20 MR. C HO P R A : Okay. I think what we need 21 before --
we need to convert these offects in terms 22 of some fracture mechanic data, which can then be 23 used for --
so we need the information to do the (n) 24 analysis for the impact on --
l 25 MR. SHEWMON: Is there any way to j
L_ _ _ _ _ __
. . . .. . .. . - - . . _ . _ - ._ _.._ _ _. - _ _ _ _ . _ _ _ _ _ _ . ~ - _ _
1 i 329 1 determine whether or not the ferrite is distributed
()
i
! 2 in a more or less continuous way?
l t i 3 MR. C HO P RA : Yes. I'm going to come to l
1
[ ;
4 that. What we use --
anyway, I'll skip some of the -
1 l 5 results here. Yeah, one point I want to make. I've '
f f' 6 taken the existing information and converted all our 1
l 7 results for samples which were aged at these three 8 different temperatures, and I've plotted the impact i
s j 9 energy versus the aging parameter. You can 10 normalate the results based on the existing
] 11 correlations, and what we should see is that all ,
12 these results should follow one curve, but in our
}
i
() 13 study we --
the low temperature aged samples do not 14 follow the curve. The higher, 400, 450, follow one i
j 15 curve, and the low temperature data is on the -
1 1
4 16 different curve, which indicates that the existing 17 correlations are not accurate for a lot of the '
i
! 18 compositions.
I 1
i 19 MR. S IIE W M O N : Does it also say that low i
- l 20 temperature aging is not as bad as high temperature 21 aging? ,
! 22 MR. C I!O P RA : It says there are two s
23 different mechanisms we are looking at. We do know
{} 24 know what would happen to these. One assumption is
- 25 that, okay, the path might be different and we would i
l ,
! 330 1 end up at the same level. I don't think we know
)
i 2 that. If the mechanism is different, with this path '
i 3 you may go even lower.
l 4 MR. S HEWMON : This P is a temperature l 5 compensated time of some sort?
l I
6 MR. CHOPRA: What it is is the time like i
7 3,000 -- no, 3 is thousand, ten to the power three
! 8 hours at 400 C. So four would be 10,000 hours0 days <br />0 hours <br />0 weeks <br />0 months <br /> at 9 400 C. So you normalate everything as aging at --
10 it uses actuation energy for --
11 MR. SHEWHON: Thank you. That was what I 12 asked.
- ( 13 MR. C HO P R A : Yes.
t 14 MR. ETHERINGTON: I'm sorry that I missed i 15 what the parameter P was. Would you repeat that, 1
16 please?
l 17 MR. C HO P RA : It's ten to the power -- this 18 number is like ten to the power four, which means i
] 19 10,000 hours0 days <br />0 hours <br />0 weeks <br />0 months <br /> of aging at 400 degrees centigrade.
- 20 MR. SHEWHON
- It's a temperature i
21 compensated time, the sort that creep people have 22 been using for years.
23 MR. C HO P RA : These results are for the 4
24 moly containing. The bottom curves shows two i
(])
25 different heats. Both have ten percent ferrite, and
}
i
331 1 we see that one of them, which is a pipe, shows much
,,, 2 more reduction than the other one. The P-4 heat 3 contains higher nitrogen and also has a very 4 different morphology for the ferrite and the 5 distribution of ferrite. We do not know whether 6' it's the nitrogen which causes this difference in 7 lthe mechanical response or whether the ferrite l
8I morphology and distribution affects the overall 9 toughness.
10 To see the microstructure of these two, 11 I the left shows the ferrite islands in the P-4 heat, t
12fand the right column shows the ferrite islands in 13 the heat 63. Now, what we do is measure the mean 14 intercept between these ferrites, and I've plotted 15 this number here, which is a good way of coming with 16 a number for the distribution of ferrite.
17 In P-4 the mean intercept between ferrite 18 is 182 microns where in the other one it's 81. It i
19l seems -- well, any casting which has more than a 20 hundred micron mean intercept spacing, the ferrite 21 may be continuous, the islands may be continuous, 22 indicating that the fracture may propagate 23 preferentially along the ferrite and you may have
- s
{} 24 very low toughness for those steels. We need more 25 information to come up with a better correlation for
332 1 these --
2 MR. S HEWMON : You're saying if L bar is 3 larger for the same amount of ferrite, then there's 4 a better chance that the ferrite has a continuous 5, path through the matrix?
6 MR. C HO P RA : Yes. That's what we are 7 saying. And actually the few Swiss castings which 8 we have looked at, they all have a mean intercept of 9 over 150 micron, and that's why they show very low 10 impact energies in the aged condition.
I I
11! MR. ETHERINGTON: The ferrite is the dark I
12 material?
13 MR. C HO P RA : That's the small islands 14lwhichyou see. Yes, the dark islands which you see.
15 Now, the fracture toughness and tensile 16 tests, these were done at MEA, and I think the 17 results were presented yesterday. What basically wo 18 see is the effect of aging on JIc are similar to 19 what we see the effect on changes in Charpy impact 20 energy, which tells us that we have a good chance of 21 correlating the Charpy impact energy with the JIc 22 numbers, so that we can use all the Charpy impact 23 data to predict the results or the changes in the 24 J-R curves.
(^ )
25 MR. SHEWMON: On this thing where --
you
333
_ 1 showed this view graph where you showed the mean 2 intercept distance for the ferrite. What was the 3 ferrite content in volume percent on that?
4 MR. CHOPRA: Ten, both have ten.
5 MR. HU TC HI NS ON : All this fracture data 6 has been taken at room temperature, and it's clearly 7 dominated by cleavage.
10 MR. SHEWMON: Let him finish.
11 MR. HU TC HI NSON : Is that the concern with 12 the embrittlement at a room temperature situation?
( 13 MR. CHOPRA: No. We are actually now 14 doing the complete test up to 300 C. Now, if you 15 look at the results from Framatome, even the upper 16 shelf energy comes on quite a bit. In fact, they 17 see a flat curve, means no difference between room l 18 temperature or tests at 300 C, and they give a 19 number of about 15 joules for the upper shelf in the l
20 aged condition, the worst case in the aged condition. l
'l MR. HU TC HINS ON : So in other words, there 22 is no transition?
23 MR. CHOPRA: They see a --
yes, they see a 24 flat it's brittle even at high. temperatures.
(])
25 MR. IlU TC llINSON : And does the fracture
334 1 surface suggest the cleavage is still taking place?
2 MR. CHOPRA: Yes, at least in the Charpy 3 impact we do see cleavage.
4 MR. HUTCHINSON: At 300 C7 5 MR. C HOP RA : At high temperatures also, 6 yes.
7 MR. SHEWMON: And this is t r'l e only in the 8 aged, which is hardened a good deal by your 9 precipitation processes that you're studying.
10 MR. CHOPRA: Yes, right.
11 MR. SHEWMON: Okay. So if you get it hard 12 enough, it's brittle or cleaves there too.
( 13 MR. C ITO P RA : See, once the ferrite is 14 embrittled, then it's a question of how it's 15 distributed in the duplex structure, and that'would 16 control the overall fracture behavior and the 17 toughness. And that thing is not included in the 18 current correlations, the distribution of ferrite.
19 They only include the amount of ferrite.
20 MR. HUTCHINSON: But isn't cleavage at 300 21 C variable?
22 MR. CHOPRA: It is, but I think -- if we 23 look at --
the presence of alpha prime can do it.
24 There's very fine -- it is possible.
(])
25 Now, the mechanism -- the results indicate I
i n_'_1 '
335 v 1 that the low temperature embrittlement is caused by 2 the alpha prime precipitates and that the alpha 3 prime phase forms by spinodal decomposition.
4 Some of the results show that aging at 450 5 degree centigrade may be outside the miscibility gap 6 for this spinodal decomposition to occur, which 7 indica:es that we should not age samples at 450 and 8 use that data to predict what would happen at 9 reactor temperatures, because at lower temperatures 10 alpha prime would form by spinodal decomposition.
11 At present we are not sure for the role of 12 G phase and the Type X phases, what role they play 13 in the overall fracture behavior and controlling the 14 toughness of the material.
15 Also, we see that the presence of carbide 16 particles at the a ~ pha/ gamma boundaries can 17 influence the overall toughness, especially of the 18 high carbon grades, and after aging at high 19 temperatures, in some cases even aging at 400 20 degrees centigrade, which suggests that even 400 C 21 aging at this temperature may be high enough to 22 change the mechanism of the overall aging process, 23 and we should not extrapolate that data to reactor
. 24 temperatures.
{)
25 tiow, we did do some initial tests to
l 336 1 recover the toughness, and we aged the embrittled 2 samples, which had very low impact energy, at 550 C 4
3 for 30 minutes, and we have not measured the --
4 carried these things for mechanical testing; but 5lwhenwo measure the microhardness of the ferrite
- 6 phase, it goes down to the or'ginal i level of what we 7 see in the unaged sample, indicating that we can 8 reverse this aging process of formation of alpha 9l prime and recover the toughness by a short term t
I 10 aging of embrittled material -- short term aging at 11 < 550 can reverse the whole embrittlement process.
' 12 Now, we are aging --
we will heat treat 13 some of the mechanical test specimens and see if we i 14l get back the toughness. Actually this --
15 MR. SHEWMON: Why don't you skip that one i
16 and let's go on.
17 MR. CHOPRA: Okay. Let's get back to the 18 correlations. Okay. I've mentioned that the i
19 existing correlations may not represent '
20 embrittlement behavior over the complete temperature 21 range of 300 to 450. That means we should be 22 careful of the temperature we use for experimental i 23 studies. Definitely 450 C, aging at that 24 temperature is too high, and we are in a different
(])
25 mechanism which is not the same as what happened at
.(.
337 1 the reactor temperatures.
2 And extrapolation of this high temperature i
I 3 data to reactor conditions may not be valid. And at 4 present we do not know the influence of the
- 5 compositional and metallurgical variables on the ;
l 6 overall embrittlement behavior, and these effects 1
7 should be included in the correlations.
8 The two correlations which are mostly used t
9 are the ones from Framatome. This one is the l l
10 Framatome correlation here, which predicts the room i 11 temperature impact energy for samples which are aged ;
12 for 10,000 hours0 days <br />0 hours <br />0 weeks <br />0 months <br /> at 400 C. This in terms of reactor )
13 conditions means the lifetime for the cold leg, 14 means 40 years at 290 C. 30,000 at 400 C, they 15 believe, corresponds to the lifetime at hot leg 16 temperatures.
17 MR. SHEWMON: Would you explain what's 4
18 measured there?
19 MR. CHOPRA: Okay. This says impact 20 energy --
21 MR. SHEWMON: Charpy V notch?
, l 22 MR. CHOPRA: Yes.
23 MR. SHEWMON: What are the units?
(} 24 MR. CHOPRA: This is decajoules per i
25 centimeter squared. The reason I used that they 4
338 1lalways --
the correlations are in decajoules per
() 2 centimeter squared.
i 3l MR. SHEWMON: Could you convert that *a 4 old-fashioned English units?
5 MR. C HOP RA : Okay. That's a hundred i
6' joules per centimeter squared, so it would be --
7l MR. S HEUMON : Which is about a hundred i
8i foot-pounds.
9 MR. C HO P RA : -- in our samples two 10 contimeter -- two millimeters notch, which would 11' be --
divide the thing by .8, which would about what, 12 1 125 or so joules, which would mean about 80 13ffoot-pounds.
14 l MR. S HEWMON : Okay. Thank you.
15 i MR. C HOPRA : I've plotted the four 16 predicted curves for different chrome, moly, silicon 17 levels here 21 to 24. Basically what the results 18 show is our data above the predicted numbers means 19 the predicted values give --
the predictions are t
20 conservative. Since the predictions are based on 21 this George Fischer study, I plotted those results 22 also, and they agree well. These curves are based 23 on those results, but the point from this figure is 24 that this correlation does not include certain
(])
25 parameters, like I mentioned, the distribution of
339 1 ferrite, the amount of carbon or the amount of O 2 nitrogen. Once we include these and we have tried 3 some analysis and we can come up with better 4 predictions, we need to include those effects in the 5 correlations.
6 MR. ETHERINGTON: I would like to suggest 7 though for us reactionaries if you'd show 8 foot-pounds on the right-hand side.
- 9 MR. CHOPRA: Okay. Actually this is more 10 complicated. I have converted --
all their results 11lare U notch.
l The French use a five millimeter deep 12[U notch for their tests, and George Fischer, the
]
13 Swiss people, use a 2.5 millimeter U notch, and we 14 ' use a two millimeter V notch. And there are a lot
, 15 of correlations for these. So the reason I've used 16 their units is, you know, it gets very complicated 17 to convert everything.
18 MR. BENDER: For the convenience of those 19 who really don't care what the units are but would 20 like to know what parts of the curves are at 1
4 21 interest -- leave it up there for a minute.
22 MR. C ilO P RA : I have another one. I'll put 23 that one.
1 24 MR. BENDER: I don't really care.t I think
{)
25 what we would like to know, first of all, is what 1
. _ _ . - _ - - , . , , - , _ . _ . . _ . . - _ - . - ~ , , . - _ ___ _
1 340
} l part of the curve is of significance to us?
okay. That's 50 joules. d 4 That's what, 30 foot-pounds. Now, if we say that 30 i
i l 5 foot-pounds is the limit, then anything in this 1
6 region is of concern.
7 MR. BENDER: Okay. i 8 MR. CHOPRA: So we should avoid a pound l 9 position which is in that region. That means if I 10 take a .05 carbon steel, then the amount of ferrite I
11I I can tolerate in that steel would be 20, 20 percent. ,
12 l MR. BENDER: Fine. Fine. I think I --
O 13 having told us what the range is, now tell us what 14 it is that the cast stainless steels are that we're
, 15 using. Where would they fall on that curve?
l 16 MR. C HO P RA : Some of them in this region.
17 MR. BENDER: That's what you're telling us
- 18 that we have to worry about right now? Okay. Now,
- 19 I'm trying to make some case for being unconcerned 20 or not concerned about it. What would we have to do 21 to become unconcerned?
22 MR. C HO P RA : Well, get the information and j 23 carry out the analysis.
{} 24 MR. MAYFIELD: I think we have a couple 25 steps to go yet, and that's to make sure that wo
341 1 understand everything that's driving, exactly the O 2 degree of degradation, and look at the materials 3 ; that are in fact in-plant, contrast their conditions, l
4 : chemistry, the temperature experience and see are we 5
really in trouble, where are we in trouble, to what 6 , degree, what limitations must we impose.
7 MR. BENDER: Well, let me start the other ,
I I ~
8 jway. We originally started pressing for this 9lbecause we didn't have very many facts. Now, we've 10 I gotten some, I think, partial facts. And the 11 question is have we discovered something to be t
12 really concerned about, or is it we just don't have 13 enough information yet? This discussion is i
14 ; interesting, but I don't know whether I should be 15 concerned about it or not.
16 { MR. CHOPRA: No, I think we do not have 17 enough information, but a couple of things which 18lthese results do show is that low carbon grades are 19 more resistant to embrittlement.
20 MR. BENDER: No surprise. We always 21 thought that.
22 MR. C HO P RA : In fact, the French and the 23 British have an upper limit on carbon, .04. And a 24 lot of the castings which are used in our reactors
(]}
25 are the high carbon, and another factor'is that the
, 342 1 moly containing is worse than even the CF-8, and O 2 some pipings are CF-8M.
3 MR. BENDER: Well, I don't know for right 4 now to satisfy me, but -- not enough to know whether 5 I should be concerned or not.
f 6l MR. MAYFIELD: Well, my impression is that
{
7 we've learned enough that we're perhaps not as 8 concerned as we night have been, but we've also 4
9 learned enough to know that we can't dismiss it and i
10 that we need to know more.
11 l MR. DENDER: Okay.
12 l MR. MAYFIELD: Is that a fair answer to 1 13 your concern or your question?
14 MR. BENDER: For me that's good enough, 15 but eventually somebody will say what will this 1
j 16 program do about telling me enough to either create 17 concern or eliminate concern. I don't get much of a 18 message back yet from that, but that's your problem 19 , to determine and not mine.
20 MR. MAYFIELD: Okay.
21 MR. C HOP RA : The last view graph is the 22 future work. We are starting the Charpy test to 23 obtain the ductile transmission temperatures and to 24 conduct tensile tests and the J-R curve tests on the
(])
25 large experimental heats at six different i
i
. - - ,. ,. e... _ . . . , _ _ - _ . . - . . , . , , , , . . _ . , _ . ,-. - ,w, ,m,c- --- - ,-
i i
343 l 1 compositions and aged up to 10,000 hours0 days <br />0 hours <br />0 weeks <br />0 months <br /> at all O 2 those five different temperatures.
3 We also are conducting these mechanical 4 l tests on the KRB pump cover plate material to 5 compare the laboratory aged samples with the 6 in-service material, and we would try to develop 7 correlations that would include influence of 8 nitrogen, carbon and distribution and morphology of 9 ferrite and what effects they may have on the 10 overall fracture toughness.
11 ! And we'll continue the microstructural 12l studies on both the laboratory aged samples and the t
13 reactor in-service components.
14 MR. HU TC HI NSON : Are these tests going to 15 be at 5507 16 MR. SHEWMON: F you mean.
18 MR. C HO P RA : The J-R curves are both at 19 room temperature and 550. The tensile tests are at 20 both temperatures. Charpy tests we are going all 21 the way from minus 200 to 300 or 550 F.
22 MR. SHEWMON: You can run a ferrite meter 23 up against something. The original spec probably 24 tells you what the carbon content will. I suspect
[])
25 you won't find anything that tells you what the mean y,.9- , sy , g r...--.3 r-%v --+.,g - ,,
---,.,.7-.s.i 9 +
344 1 ferrite pacing is --
3 MR. S HEWMON : -- in anybody's specification.
4 Can you correlate this with whether it was static 5' cast or dynamically cast -- sorry, static cast or 6 centrifuga11y cast or any other parameter that says 7 what class would be more or less prone to that large 8 spacing?
I 9 MR. C HO P RA : I think the static cast 10 probably would be more prona. No, thick sections, 11 thick sections, anything with eight inch thickness 12 would have --
13 MR. S HEWMON : So you're saying it's the 14 cooling rate primarily?
15 MR. C HO P RA : Right. And I think there are 16 some studies where they have included the section 17 thickness and its effect on tensile properties, for 18 example. But again those are from the breeder 19 program, higher temperatures and so on. There were 20 some studies.
21 MR. SHEWMON: Now, the difference in 22 composition between welds and casts is what? Are 23 they usually lower carbon or different from this?
24 MR. C HO P R A : They are in the same range.
(])
25 MR. SHEWMON: Completely right --
one on
t t
345 r
1 top of the other?
2 O 2 MR. C HO P RA : On the lower -- they would 3
3 be --
the weld itself would be probably slightly
- 4 lower.
5 MR. S HEWMON : Lower in what?
i 6l MR. C HO PR A : Ferrite.
3 7 MR. S HEWMON : Okay.
8l MR. CHOPRA: They are normally three to l
9 l ten percent ferrite.
) !
4 i 10 l; MR. SHEWMON: So there's less ferrite and
< 11 more slag from what we would guess now at least; is 12 that it?
,~
14 MR. S HEWMON : Okay. Any other questions?
15 MR. ETHERINGTON: Does the --
maybe I 16 should be asking you, Paul. Does the P correlation 17 enable you to pick a time that corresponds to the 18 actual temperature of the operating plants?
19 MR. S HEWMON : Yes.
j 20 MR. C HO P RA : Yes. But the question is 21 whether the activation energies which we are using
! 22 to convert everything in P, are they accurate, do 23 they represent the actual process?
24 MR. ETHERINGTON: Can we interpret it to
(]}
) 25 actual conditions? And I think you said yes.
i
, - - - , - - c -
. - - . - - . , _ _ .. , . - - - - , , , . .,,mm. - , - , , , - - . - - . , - - - - ..v _ . . -
346
. 1 MR. SHEWMON: Yes. The other thing they 2 find though is that if you go to the same --
that 3 assumes you end up in the same final stage, and at
- 4. higher temperatures you use a different mechanism so 5 you don't get to the same stage.
6i MR. ETHERINGTON: Well, is there some kind 7 of an Arrhenius correlation? ,
l l
8! MR. CHOPRA: That's what it is actually, 9 it's Arrhenius extrapolatio.'.
10 MR. SHEWMON: But he doesn't like to use 11 that word for some reason.
12 MR. SHEWMON: Okay. Th a nk you.
() 13 ' MR. S HAC K : I'm talking about our program i
14l on environmentally assisted cracking of stainless t
15l steels in boiling water reactor environments and, of i
16; course, I'm representing a large number of my t
17 colleagues at Argonne in presenting this work and Ed 18 Rybicki who has been helping us.with some f i r ! >. e 19l element calculations for residual stress remedies 20 like IHSI and the weld overlay.
21 When I made my first presentation to this 22 committee on this problem, someone had said we were 23 sure that we were going to solve this problem 24 because it had been solved at least a half dozen
(])
25 times since he had been on the ACRS, and you'll be l
I !
i 347
- 1 glad to know we're still solving the problem.
2 MR. SHEWMON: Even though he's off the 3 ACRS.
- 4 MR. S HAC K
- Intergranular stress corrosion 5!I cracking has been the problem of main concern. It I
6! arises in sort of two situations that we're i
7 concerned about here. One is in the sensitized 8l stainless steels and the recirculation piping 9 systems, and that's the problem that's caused the 10 ; major concern and been the major focus of attention i
11lof!
all the utility action, the NRC action and a 12llarge part of our program.
, () 13 The proposed remedies for this problem 14 fall into three classes. One is simply to replace 15 the piping system with an alternative material, the 16 316 nuclear grade, which is a low carbon, nitrogen 17 strengthened 316 stainless steelt the TP347 is a 18 German variety of 347, which has all the usual j 19 benefits one associates with 347 but according to i
20'the Germans solves the weldability problems that has 21 given 347 such a terrible reputation in this country.
J 22 Another solution is to alter the water 23 chemistry. A boiling water reactor is different
{) 24 from a pressurized water reactor in the sense that l 25 it's an open cycle system. A pressurized water I
1 l' x 348 s 1 reactor operates conventionally with a hydrogen 2 overpressure, which keeps the dissolved oxygen which 3 l is produced by radiolysis in either reactor to a 4lvery low level. In the boiling water reactor people 5i have shown over the past few years that it's i
6 ! possible to reduce the dissolved oxygen content in 7 the BWR, if not to PWR levels, at least much, much 8i lower than it conventionally is by adding hydrogen.
9i The other solution -- a g a i n , the problem 10 has typically arisen in the vicinity of weldments, 1
11 partly because that's where the sensitization occurs 12l with the conventional stainless steel, but also
(- 13l because the typical normal welding practice produces i
14l very high tensile residual stresses on the inside 15I surface of the weldment in the heat effect. So i
16 another possible set of remedies is to go in and 17 modify that stress state in the vicinity of the weld.
18 You can solution heat treat the field or the shock i
19l welds.
20 In the field we have remedies such as 21 conduction heating stress improvement and MSIP, 22 which is mechanical stress improvement process, and 23 I'll talk a little bit about those, which give you 24 the possibility of altering that residual stress
(])
25 state.
349 1 Now, there's another possibility for
() 2 intergranular stress corrosion cracking which 3 involves nonsensitized stainless steels in high 1
4 jradiation areas. This is essentially a problem that 5 was known early on in light water reactors when i i 6l stainless steels were originally used as cladding I
7 imaterials and when they were subjected to high j 8 fluences which, of course, a fuel pin sees very soon 9iin its career, they became susceptible to this I
10l!irradiation I
assisted stress corrosion cracking.
11 Now, the solution for that problem was 12 simply to replace stainless clads with zircoloy that
[
13 gave you some neutronic benefits also, but 4
14 unfortunately we also have other stainless steel 4 15 l components in the reactor, and although they don't l
l 16 accumulate irradication damage nearly as rapidly as 17 fuel cladding does, as the reactor starts to l 18 approach significant parts.of its lifetime these 19 other things like top guide and the boiling water 20 reactor and some of the other components reach the 4
21 fluence levels at which we believe they're 22 susceptible to this irradiation assisted cracking.
23 Now, again, as I say, people were aware of 24 the problem with the stainless steel fuel cladding.
{])
25 They solved the problem ve ry quickly, so that the
350 I'
1 interaction between the irradiation, the environment l () 2 and the stresses is somewhat unclear for this 3 problem.
f 4 We have some proposed remedies for that.
I l 5 Impurity control of the stainless steel is one. The l 6 Lacross boiling water reactor, for example, still 3
{ 7 operates with a stainless steel clad.
l 8 MR. SHEWMON: I was going to ask. There i
9 are some people who found they can use it for life.
10 I thought Con-Yankee did for n good while or 11 i something too but --
4 1
5 12 f MR. SHACK: Right. What the people found i
Q 13 in those early investigations of the cladding was 14 that they seemed to tie the susceptibility to 15 precipitation of things like phosphorus and sulfur 4
16 to the grain boundaries.
i 17 So the solution that the people who still 18 successfully use a stainless steel clad is to use a 19 very low impurity or very tight impurity controls, j 20 especially on phosphorus and sulfur, on their i 21 stainless steel.
22 Now, although that was developed and i
i 23 implemented for the cladding materials, the reactor i
24 vendors, especially GE, have never bothered to
{])
l 25 impose such inpurity controls on the other stainless 4
d
,- w ,, , e .----y. . . , . , . , .r = r-.-+-n-w--= ' -
-w v-ea_ w vr - w 'r- ' ' " - - ' - + ' - - ' * - ' '
.- ~. - - . -- _ . - - . -
351
{
l 1 steel components that are in the core. So that
() 2 there are no real phosphorus and sulfur specs on,
- 3 for example, the top guide.
4 Another remedy or one would hope it's a 5 remedy is hydrogen water chemistry. Again, if that ,
6 can effectively change the corrosion or capability
+
7 of the environment even in the reactor core, it ,
i 8 becomes another possible remedy. But we obviously 9 understand much, much less about the irradiation 10 assisted cracking than we do about the cracking in 11lthe recirculation piping system.
12 MR. S HEWMON : Is it at all clear whether 13 this is something radiation does to the water or 14 whether it's displacement damage in the metal?
! 15 MR. S HAC K : It's both. It's not so much 16 displacement --
The displacement damage has another
. 17 bad consequence in the sense that not only does the 18 material become more susceptible to cracking, but it 19 becomes embrittled. You know, it's one thing to 20 have a crack in a very tough material; it's another i 21 thing to have the material lose toughness as it 22 develops a susceptibility to contracting. So the i
23 displacement damage is essentially reducing the 24 fracture toughness of the material.
4
{])
1 25 The segregation --
let's assume f or a i
i
, ~.,,..n. , . - , - - , , , , - , . - , , . , - , , , , , . . , , , , ~ - , ,.,_, ,,,, - ,-r-- - --..-,.g ..--
352 fs 1 moment though that we're not exactly sure. Let's U 2 assume that the segregation is the mechanism. The i
3' segregation at the same time is promoting the 4{ susceptibility to the stress corrosion cracking.
i 5- The radiolysis is also essentially I
6! producing your corrosion potential in the 7 environment. You know, it's the radiolysis that's 8j producing the oxygen that causes the problem both in i
9' the core and in the recirculation piping system, so 10 ! that --
r 11 MR. S HEWMON : Let me get on a separate 12lmore detailed issue. It's not clear to you if you
() 13 i do radiolysis and produce both hydrogen and water 14, and if it occurs in a nonsensitized --
sorry, 15 hydrogen and oxygen, if it occurs in a nonsensitized 16 material, then it's not obvious that it's still the 17lsame process that's occurring, but that's a fine 18l quibble we can argue about later. It's not clear to 1 I
l 19 i me that it's still the same kind of stress corrosion 1
I 20 cracking grain boundary. l 21 MR. S HAC K : No, it's a different kind of l
22 stress.
23 MR. SHEWMON: Then you may not need the
() 24 oxygen potential. It could be the hydrogen 25 potential.
I 353 1 MR. S HAC K : That's possible. The early 2 work on the stainless steel cladding seemed to 3 indicate that you needed an oxidizing environment, 4 but that work was all done 15 years ago, and it's 5 hardly conclusive.
6 MR. S HEWMON : Is it clear that th'is is I
7 stopped by a hydrogen overpotential, this radiation 8 enhanced cracking?
4 9 MR. S HAC K : Let's put it this way. The 10 problem is considerably less severe in PWRs than it
{
11lis in BWRs, which would certainly indicate that 12 lowering the dissolved oxygen is a helpful process.
() 13 MR. ETHERINGTON: There's no real concern 14 as to what may happen to the cladding; is there?
i 15 MR. S HAC K : Well, in virtually all the.
16 commercial power plants, except for Lacross, as far 17 as I know in this country, the cladding has been
! j i 18 replaced by zircoloy, which is immune from this l 3
19 problem.
20 MR. ETHERINGTON: All the way through? I 21 MR. RODABAUGH You're talking about two 22 different claddings, I think.
3 23 MR. S HAC K : Oh, cladding. Oh, pressure 24 vessel cladding. Oh. I think that the pressure 1
(])
25 vessel clad is probably far enough away. One has to I
- .. -. .- ~ ,, , - . . ~ ~_ . ~ . __ . _ . . _ . ~ . _ _ _ . . . . . . - . _
354 1 go back and examine all these materials. I believe O 2 that the pressure vessel cladding probably sees a
- 3 low enough fluence that this isn't a problem for 4 that particular stainless steel. It's the other 5 core components which are in a little tighter that i
i 6! become the more serious problems.
7 MR. ETHERINGTON: Even if the cladding 8 goes all to hell it doesn't really matter an awful I
9 lot; does it?
10 HR. S HAC K : No, and I think the concern is 11 more components like the top guide --
if the top 12 ; guide comes apart in a BWR, there's this rather 13 nasty distribution of fuel elements all over the 14 reactor core.
15 Well, as I mentioned, the intergranular 16 stress corrosion cracking problem is what concerned 17 us in the recirculation piping system up until now.
18 We have these remedies. The remedy that's probably ,
19 most strongly implemented in the field is simply to 20 replace the piping systems, especially with the 316 i
21 nuclear grade.
22 What we're showing in our laooratory 23 testing is that, although that does seem to 24 eliminate the intergranular stress corrosion i
(])
25 cracking problem, we still have a potential problem i
s
355 l 1 for transgranular stress corrosion cracking in these O 2 materials. We've demonstrated that for the 316 3 nuclear grade. We're now investigating the 347.
4 I'm guessing at the moment that the 347 will also be 5 susceptible to that.
6 MR. BENDER: Bill, you are going to say t
7 something about the tests that establish this or --
8 MR. SHACK: Yes. I'm trying to get the 9 conclusions up front and then we'll go to the detail.
10 MR. BENDER: I like what you're doing. I I
11ljust wanted to be sure you'll cover it.
12 l MR. S HAC K : You'll get the data. Now, we 13 can mitigate this transgranular stress corrosion i
14 l cracking problem by tight control on the impurities 15 in the coolant, residual stres's improvement. It can 16 be eliminated by hydrogen water chemistry.
17 And I think if we had to offer an opinion, 18 if you want the same resistance to stress corrosion 19 cracking in a boiling water reactor that you have in 20 a pressurized water reactor, you're not only going 21 to have to change the piping system, but you're i 22 going to have to implement a hydrogen water l
23 chemistry. Now, whether you need that much I
(} 24 resistance to stress corrosion cracking as a problem, 25 I won't attempt to answer. l i
356 1 MR. BENDER: You say you had to do both?
2 MR. SHACK: You would have to do both if l 3 you wanted an equivalent resistance. I won't answer 4 the question whether you do want equivalent 5' resistance.
6 I'm going do try to outline what my talk l 1
7 will cover. Again, I've been allotted two and a 8 half hours on the schedule. I have lot of material !
9 to cover. I'm going to try to organize it, and 10 you'll be relieved to know I did throw out a few 11 more view graphs last night, but let me try to l
12 review what I'm going to cover. If there's anything '
13 you're not interested in, you can tell me to skip it 14 and at least you'll know what's ahead if there's 15 some parts t h'a t are putting you to sleep.
16 I'm first going to cover the work on the 17 alternate' materials, talk a little bit about some of 18 the previous experience and some of the laboratory 19 tests that convinced people that this was a good 20 idea in the first place, a few words on the factor 21 of improvement, then look at our work on impurity 22 effects, both in either CERT tests and crack growth 23 tests that cause some concern about the performance
{} 24 of these materials.
25 MR. HU TC HI NS ON : What are CERT tests?
-.. _ - . _ _ _ _ . _.. _.- . - _ .. .-- - _ . . . . . _ _ _ .. - . .. . - ~ . - . . .
2 357 l
i i 1 MR. S IIA C K : Constant extension rate i
() 2 tensile tests.
4 3 Follow-on work, again we've raised some
'4 4 concerns here. We think we still have to do some
{
5 work in order to understand that problem a little
, 6 better, so I'll describe that f ollow-on work' that
{ 7 I we'll be doing.
i j 8 One of the things that we will find is 9 that the material looks especially poor in crack 10 propagation tests, so that once you have an existing i .
I j 11 crack in your material our tests showed that it f 12 really isn't any better than the material that's out I
v 13 there now. That makes us very concerned about j 14 initiation, that is, we don't want any cracks there.
i 1
15 Now, two possible sources for initiation 16 beyond the usual stress corrosion thing, surface i
) 17 cold work, and we think that water chemistry has an j 18 important role there.
19 If you go to -- well, even 360 nuclear j 20 grade has had some problems in the field with l
21 weldability and fabrications.that is, it doesn't t
y 22 really matter if a material is extraordinarily 4
1 23 resistant to stress corrosion cracking, if you can't 24 fabricate it into'a piping system, it's not a whole
{])
25 lot of good to you. 347 is, of course, notorious, l
f
- - . . -- --. , - . , _ ~ . , . - , , - - , - . - - - - . . - . , - . . . . - _ . - . - - . - . . . - . . . - , , - . . - . . - - -
358 s 1 at least among certain utilities in the country, 2 they wouldn't want to touch it with a ten foot pole.
3 l We have some cooperative efforts with EPRI, the New 4 York Power Authority and the University of Tennessee i
5!to look at some of these weldability problems.
6, I'll then go on to discuss some work on 7! weld overlays, which are certainly widely used in 8ithe field as a remedy or at least a temporary remedy 9 I for cracked piping. I'll talk a little bit about 10 MSIP which is mechanical stress improvement process.
11 .
It's a new technique to produce a beneficial 12 residual stress state on the inner surface of the O
\/ 13 l pipe that could be of interest to people with 316 14 nuclear grade piping systems. Then I would like to some future work that we're looking at on 15l discuss 16lferritic I
steels and the irradiated stainless steels.
17l And Gery Wilkowski has told me that you're i
18 i going to have steak for lunch and so I have to quit l
19i about 12:15.
There will be a small break in the 20 proceedings.
21 Just to go over on the alternative 22 materials. Significant nunbers of BWRs have now 23 changed over their piping systems. I think it's 24 about nine reactors. It's a littic bit difficult
({}
25 actually to V.eep up. At present all the utilities si
359 s 1 that have replaced materials have chosen 316 nuclear 2 grade as the replacement material.
3 Fitzpatrick is currently installing the 4 German 347 in the core spray and some other l
5 ! nonrecirculation system piping, partly to convince 6! themselves that they like working with 347. The 7 Germans are making an active promotion of their 347 8 i material in this country. So far as I know, 9lFitzpatrick is the only plant that's actually 10 installing, but other people are considering it.
11 MR. S HEWMON : What is TP347 chemically?
12 MR. SHACK: It's 347 with a lower carbon
() 13; level and tight restrictions on impurities, and an 14 i adjustment in the chrome / nickel level to give you i
15labout one to two percent ferrite. So it's like the l
16 347. It's an nicobium stabilized stainless steel.
17 MR. SHEWMON: How low is the carbon?
18 i MR. S HAC K : The TP347 I think is an 03 19 carbon --
it's equivalent to essentially an L grade.
20 It's not quite as low as the nuclear grade stainless 21 is in this country.
22 MR. SHEWMON: Okay.
23 MR. SHACK: The French are busy trying to
{} 24 interest utilities in the CF-3 piping. Everybody 25 sees this as a great opportunity. I'm not sure that
- ------,w --
l i
l
>- 360 l 1 they've managed to rouse up any serious interest at C:)
] 2 the moment.
) 3 MR. SHEWHON: And what's CF-3?
l j 4 HR. S HAC K : CF-3 is essentially a low 1
5 carbon cast stainless. It would be equivalent to ,
6 304 L. It would be what t h e' French and the British j 7 are basically using in their pressurized water I
! 8 reactors, but in these -- the relatively thin I
j 9 thicknesses, you know, a BWR piping system, the
- 10 thickest pipe is like an inch and a quarter. You 1
] 11 can actually make a very nice cast pipe, very fine 12 grained, and it's a nice piping system actually, you
( 13 know. Looking at it from a completely unbiased 14 point of view, it's a rather attractive sort of 15 material.
16 There was some --
the reason the utilities
} 17 are interested in these other materials is that i
18 there was some early difficulties in fabrication i .
i 19 with the 316 nuclear grade component. One of the t 20 things that people are doing when they're replacing 21 their piping systems, since you have to inspect 22 weldments there's a great incentive when you replace 1
23 the piping system to cut down the number of welds,- '
1 24 so that instead of having, say, a riser, an elbow
{])
25 and then an inlet to your reactor as in a typical
?
, _ . . . . , . - . . . . _ _ . . . _ . . . , _ _ . . . _ _ , , _ . , , _ , . . - . . . . . _ _ , . ._._.._._.._,,___,.-___,m._.m.m
361 1 BWR system, you essentially just bend one continuous O 2 riser piece and eliminate the elbow. So you've 3 saved two welds on each riser.
4 They had microcracking in these bent 5; components until they learned how to essentially do I
6 j it and tighten up some of the controls on the 316 i
7l nuclear grade composition. There have been some 8 l questions about its weldability which again people 1
9 have adjusted compositions to handle that, and we've 10 raised some questions about its stress corrosion 11 resistance in impurity environments.
12 ; MR. S HEWMON : Raised them where?
13 MR. SHACK: Pardon me?
14 MR. S HEWMON : Raised them where? What I'm 15! interested in is NRR is still meditating on whether I
16 they're going to let people off with only one of 17 these fixes or whether they're going to have to go i
18jwith two.
19 HR. S HAC K : Well, I think the people in 20 NRR are aware of our work. I would say it's hardly 21 definitive, you know, we had input to 1061 which 22 again the conclusion there was that 316 nuclear 23 grade was very good, they would like to see you do
(]) 24 something else, but they weren't going to require it.
25 EPRI has essentially taken much of our
362 1 work on impurities and incorporated that into the br_
2 owners' group recommendation for water chemistry 3! practice. Again, utilities would much rather have 4 !
recommendation than they would requirements.
5l MR. SHEWMON: Okay.
l 6 MR. S HAC K : Just going back to the nuclear 7' grade stainless steel, in the original qualification 8 tests that were set up for that, EPRI and GE ran a 9I series of pipe tests where they took four inch 10 diameter pipe and subjected them to full scale I
11 cyclic tensile loadings in reactor coolant type 12 environment; and whereas a 304 pipe weldment in that 13j kind of test would fail in typically a hundred hours, t
14fthat i
you had only one failure in 108 316 nuclear 15, grade weldments in 5,000 hours0 days <br />0 hours <br />0 weeks <br />0 months <br />. So it appears a 16 tremendous factor of improvement in that kind of a 17 test.
18 The 316 nuclear grade in Japanese reactors, 19! they started installing the material about 1978.
20 1 Prom all appearances it appears to be performing 21 very satisfactorily in those materials. By and 22 large low carbon materials, of which 316 nuclear 23 grade is one, have performed much, much better in 24 U. S. reactors than conventional materials.
{])
25 However, this factor of improvement you
363 4
1 get in the laboratory tests appears to be strongly O 2 dependent on the particular test that you use. It
! 3 looks terrific in this pipe test. In our ongoing 4 crack propagation tests we find the 316 nuclear 5 grade is not only not superior to the sensitized 6 stainless steels, it's actually inferior to
[
7 sensitized 304 stainless steel, even in high purity i
j 8 water. That's a disconcerting result. It was very ,
9 disconcerting to us.
10 MR. HU TC HINSON : So those are standard 11 fatigue tests?
12 l MR. S HAC K : I'll describe those tests in 13 more detail.
f j 15 MR. S HAC K : They're not standard fatigue l 16 tests. They're much more constant load rather than i
l 17 a fatigue type test.
i 18 MR. BENDER: Just a matter of time frame.
19 When was the ANL work done?
20 MR. SHACK: It's ongoing at the. moment.
) 21 MR. BENDER: And the announcement of this 22 disconcerting result occurred when?
i f 23 MR. S HAC K : I wrote a letter to GE in June.
I i
{} 24 MR. BENDER: Thank you. I just wanted to j 25 know that.
i
,,--,,..w,-+, ,..--n.--,~,---, - . . , , ,-m,-,,, ,n r.n a-,v ,c,,,,,,,n-- n. ,-n, ,,-,--.--<w-vn..-.., ,e m - ,,r+, ,--
l l
364 i
1 MR. S HEWMON : June. When in June? June 2 is -- this June, '86, '857 3; MR. S HAC K : No, June '86, early June '86.
4 Of course, our NRC sponsors have, you know, been 5 aware of the problem for some earlier months.
I
. 6t. MR. BENDER: But that says it just got ont i
7l on the street.
l 1 8 MR. S HAC K : Yes, this particular 9 difficulty. I mean this one is the most A
10 disconcerting of all. I would say now we've 11 reported earlier that 316 nuclear grade is 12 susceptible to environmentally enhanced cracking in 13 impurity environments. That's older work. I think
- 14 there's now general agreement among the utilities,
! 15 the vendors and EPRI that that's a real phenomenon 16 and you do have to have, you know, very careful i
17 controlling over your impurity levels even with the
~
18 316 nuclear grade piping system.
j 19 The fact that we can show inferiority in 20 high purity water I think is a much more 1
2 21 controversial point at this time. As I say, our j 22 conclusion is that to get full benefit from
]
l 23 installation of a 316 nuclear grade piping system, !
{} 24 it's going to be necessary to maintain a very high 25 water chemistry standard, and if you want to'be
i i 365 1 comparable to a PWR you'll have to go to a hydrogen 4
0 2 water chemistry.
3 MR. BENDER: Your term high purity water 4 does not mean it's water that has been enhanced by 5 hydrogen.
6 MR. S HAC K : No, no, that just means very 7 low conductivity levels but essentially no hydrogen 8 additions.
9 MR. S HEWMON : Don't dump resin into it.
10 MR. S HAC K : Don't dump resin into it.
11 Don't have leaking condensers.
12 The constant extension rate test, for 13 those who are unfamiliar with it, is basically a 14 very slow stain rate tensile test that's performed 4
15 in a reactor type environment. So it's a tensile 16 test in an autoclave at a strain rate that typically 17 goes from somewhere like two times ten to the minus 18 six to five times ten to the minus eight, depending 19 on the patience of the investigator.
20 The only thing that's really important in 21 this data is to note that as we raise the 22 conductivity, which in our tests is typically 23 controlled by adding sulfuric acid, and the source
(} 24 of the sulfates, as Professor Shewmon has mentioned,
, 25 is your resin intrusion from your water clean-up
366
~
l system. As you go-from this very low conductivity 2 water, the less than .2, to slightly higher 3 conductivities you get this transition in the f
4; failure mode from purely ductile failures to i
5 l transgranular stress corrosion cracking, and this l
- 6l occurs at sulfate levels down around 25 ppb.
i 7 MR. ETHERINGTON: Sometimes you hear-that 8 reduced forms of sulfur are bad actors and now it's 9 sulfates. Is that sulfate ordinarily ~a bad impurity 10l in --
11 MR. S HAC K : It's certainly a bad impurity i
- 12 in our tests, and most of the other people who have
() 13 tried it also find it a bad form. Exactly what form 14 that sulfur is at the temperatures that we're 15 working at we really wouldn't claim to know. You 16 know, we add sulfate at 25 degrees C.
17 MR. ETHERINGTON: It would hardly be 18 reduced anyhow; would it?
19 MR. S HAC K : It's unlikely. Just to give t
20 you some feel --
21 MR. HU TC HINS ON : Could we go back here a 22 second?
23 MR. S HAC K : Sure.
() 24 MR. HU TC HI NS ON : T/f is time to failure? [
25 MR. S HAC K : Yes.
4
l 367 l i
l l 1! MR. HUTC HI NSON : So even though the mode I
() 2 ' of fracture changes from ductile to transgranular, !
l 3 lthe strain is actually going up; is that right? Am 4fI reading that right? l l
l ,
! 5 MR. SHACK: Yes, basically that's correct.
i 6 Again, you're looking at very low crack propagation
! 7 , rates here, so that although we're getting l
8l transgranular failures, you really have only very 9 small transgranular cracks on the surface, and it's 10 not affecting the gross measures to failure like the l
11 time to failure too much.
12 MR. HU TC HINS ON : So would you say that the
() 13, time to failure or, in other words, the strain to 14 failure is essentially inuependent of the sulfate?
15lThere's not a hell of a lot of variation.
16 i MR. S HAC K : Yes, for these materials it's l
17lnot. I mean you can have materials which are more 18l susceptible to impurities than this. The only thing 19 that we're really looking at here, as I say, this is 20 a fairly gross measure of susceptibility. If these 21 showed dramatic changes, then you would really be 22 worried. What I'm really concerned with showing 23 here is the possibility of a stress corrosion crack
{} 24 and the crack growth rate or the more critical 25 measures.
4 368 F
4 1 MR. S HEWMON : When you -- this is still a 2 CERT test. Is this part transgranular and the rest 3j is ductile?
4 MR. S HAC K : It's ductile.
5 MR. SHEWMON: So this says that you find 1
6i some of both on the fracture surface.
i 7 MR. SHACK: Yes. All this is really 8 indicating is that we have some transgranular stress 9 corrosion cracking evident. There's a much larger 10 portion of the fracture surface that's ductile.
I 11 l This is actually fairly old data at this i
i 12 f point. I should point out.that the Reg Guide 156,
- 13 which recommends water chemistry for boiling water 14 reactors, and the technical specification which 15 actually specifies water chemistry that the boiling 16 water reactor must meet, requires that you have 17 essentially a conductivity less than one microsiemen l
18 per centimeter, is the conductivity that you must 19 meet in your reactor for steady state operation.
20 That would correspond roughly to a tenth of a ppm of 21 sulfate in the water as sulfuric acid. Essentially 22 the theoretical limit, if you nad no impurities in I
23 your water, at this temperature --
or at 25 degrees
{} 24 C would be a .05 microsiemen per centimeter 25 conductivity. So that sort of brackets the range of
. _ . . _ . _ _~ . _ _. ~ _.- - . _ _ _ _ _ . _ - _ _ .- ._ _
l
- l 369 l
) 1 conductivities that you might expect. This just 2 shows the number of reactors and their typical 3 average conductivity. And you can see --
4 unfortunately this includes both U. S. and foreign s
l 5 reactors. Unfortunately, the foreign reactors 6 dominate this portion of the curve. The U. S.
7 reactors dominate this portion. And I think --
4
- 8 MR. S HEWMON
- That's primarily the design j
9 of balance of plant?
j 10 MR. S HAC K : No, it's primarily operating i
i 11 practice -- well, it's partly the balance of plant.
i 12 The condenser design makes a tremendous difference.
13 All the Scandinavian reactors, for example, have 14 titanium condensers. It also reflects partly 15 operating practice. One of the worst sources for 16 impurity intrusion in reactors is your water 17 clean-up system and how you essentially clean up 18 your demineralizer beds. The Japanese are very, ,
19 very careful. Although they have the same GE design 20 essentially as the U. S. reactor, they're very 21 careful to specify their procedures and typically 22 have managed to avoid most of the resin intrusion 23 incidents that plaque U. E. reactors.
24 MR. BENDER: Bill, does the average mean
(])
j 25 anyth!.g? I mean obviously if it's low your chance 1
4 w ree -e---ye - - , ,# -
y , --c--- e -.ce . < - , . - ,m .,g,-,---, + .-m% =.emi , - , _ - * . , - e -ev---- v , e
370 1 of getting out of the limits are better, but --
O 2j MR. S HAC K : An average this high indicates i
3ja kind of a sloppy attitude towards water chemistry, 4l and a person that has an average this high probably 5 also has a considerable number of resin intrusion l 6l incidents that raise the thing up to significantly l !
l 7 l higher levels. By itself it's not a terribly i
8! meaningful number.
9 And again, there's been a tremendous 10 effort in the industry to improve these water 11 chemistry practices in the U. S. I think that if we 12 ; re-did this graph based on last year's data it would 13 look much more --
14 j MR. S HEWMON : Having to replace the piping 15l tends to get their attention apparently.
I 16 ; MR. BENDER: I gathered that, even to the 17 l extent of putting in titanium tubing in a few i
18! condensers.
19 ! MR. S HAC K : Most of the data I'm going to l
20 show you on impurities will concern itself with 21 sulfates, and one reason that we concern ourselves 22 with sulfates is that they're a breakdown product of 23 the resins that are used to clean up the reactor.
(} 24 So in many ways sulfates and chloridos are probably 25 two of the most common impurities that you can find
~
47 371 1 in a reactor coolant.
2 We've also taken a sort of Edisonian 3 approach to water chemistry studies here, we add a 4llittle bit of this and a little bit of that and see i
5 lhow it all comes out. If you do that, you start out 6'with the high purity water here, and you can see 7 that sulfur species in general are among the worst, 8l and they're all roughly equivalent.
9l MR. ETHERINGTON: Are these all equivalent l
10 i concentrations?
11 MR. S HAC K : These are all equivalent 12 concentrations, essentially a tenth of a ppm, and 13 that's a basic environment. We've done the same i
14 thing, in an acidic environment, and again the 15l sulfur species look the worst, and the different 16 i
! sulfur species seems to be roughly equivalent. So 17Iwe've focused on sulfate. It's a common impurity, 18j and it appears to be the most detrimental of these 19l common impurities.
20 MR. SHEWMON: Why are the growth rates two 21 orders of magnitude bigger here than they were back 22 on your table of the CERT tests?
23 MR. S HAC K : I think we're changing from (3 24 centimeters to meters.
C/
25 MR. SHEWHON: Well, you don't in the I
372 1 caption. You may be in the fact. Okay. Well --
2 MR. S HAC K : You have to check the strain 3, rates. Those were probably much lower strain rate i
i 4 tests, which will also lower the strain ratds on the 5l CERT table.
6 MR. S HEWMON : Onward.
7: MR. S HAC K : Probably just a few more words.
I 8 Again, you could probably go on studying water 9 chemistry by the Edisonian approach for as long as 10 Dr. Arlotto would be willing to fund us. We could
)
11' keep adding various sorts of impurities forever.
12 We tried to take a somewhat more I
<~
s> 13 i systematic approach to this, and at least f or a i
14 t large class of impurities we think we understand how 15 !
the impuritics affect the problem and how mixtures 16 I o f impurities, that is, if you take sulfate and you 17 mix it with chloride is that going to be worse than i
18 jsulfate alone or worse than chloride alone? We've 19 done a few Edisonian experiments of that nature, and 20 ' sulfate typically comes out the worst.
21 Just a quick view on one of the things 22 that the impurities might be doing to you, we 23 believe that the mechanism for the stress corrosion 24 cracking in here is anodic dissolution.
(]) So we have 25 essentially the stainless steel which is generically m
! l
- 373 '
\
1 represented as metal. That's Me. That's a new
- 2 element. It dissolves at the anode, becomes an ion 3 and we lose Z electrons. We have some electrons.
l 4 i Now, if we had a purely acidic environment, I
5 what you would probably do is you have a direct i
6l anodic dissolution mechanism here and a cathodic I
i 7 I reaction essentially would just be the hydrogen 8l reduction.
i 9! Since a BWR is not an acidic environment, 10 you go through a much more complicated series of
't 11! steps that earn these hydrolysis reactions, but we 12! can ignore all those. What you really need is an 1
13' anodic reaction and a cathodic reaction to maintain i
2 14 charge balance, you know, we're producing electrons.
15 We somehow have to remain electrically neutral.
16 The cathodic reaction that we commonly 17 find in a BWR is this oxygen reaction. Many of l
18lthese impurity species simply provide alternate
) 19 cathodic reactions, and so they work in much the I 20 same way as oxygen. It's one of the reasons why in t 21 your hydrogen water chemistry environment you have I 22 to maintain the tight impurity control. I f v >u 23 eliminate the oxygen and you still got sulfate, you
^
() 24 still have a cathodic reaction that can take place.
25 So that when you implement your hydrogen i
i l - -
. _ _ ~ _ _ _ _ . . _ _ . , _ . ~ . _ _ , _ - , _ . , _ . _ _ _ _ _,_ _ _ - _. ~ ._ . .... _ , .... - -
374 gs 1 water chemistry you're essentially trying to 2 clininate all the possible cathodic reactions that 3 ! could occur. And as I say, Tom Kassner has done a l
4 !
great deal of work in this, and he can show that, i
5' for example, the effects of these impurities are
- 6) roughly additive. So that if, you know, there's no 7' synergistic reactions if we have a .1 ppm of sulfate l
8 ! and we have .1 ppm of nitrate, essentially the 9 effect is additive, and we can understand how that 10 occurs, and using sulfate as our worst case impurity 11 seems quite reasonable.
12 Now, all impurities don't work quite as (msj 13 j neatly as that. Life is never that simple, but at i i
14 least this takes care of a large class of impurities. !
15 l To just get back now after that slight 16 digression to the effect of impurities on the 316 17 nuclear grade stainless steel, what I showed in the 18l CERT test graph, that there's a great deal of 19 difference between high purity water, but it doesn't '
20 ' take very much impurity to get you up over the hump.
21 You sort of go from no cracking at essentially the 22 very high purity water to, you know, 20 ppb gets you 23 up in the range where you're almost at the same 24 crack growth rate you get out at the 1 microsemen
(])
25 which corresponds to your .1 ppm.
" ~ ^
375 1l Again, this is your Reg Guide 156 limit is
() !
2 iright here. A typical reactor operator might be in I
i 3 ! here. He's going to have to stay very, very clean l
4 l in order the get himself much benefit, at least l
5 according to these tests.
6! The good news is you might have picked a 7 ll worsematerial. We looked at the effect of nitrogen 8l additions on the I
susceptibility of stainless steels 9!to this transgranular stress corrosion cracking. In 10 ' the U. S. the nuclear grade stainless steels are 11 ' limited to a .1 percent nitrogen addition. In Japan 12lthey let them go up to .12 percent. The LN 13 stainless steels, which are also now code acceptable 14 in the ASME, go out to --
is it .187 And we can see 15lthat as you get up to these higher nitrogen levels, 16 you're producing considerable increase in 17 susceptibility to your transgranular stress 18l corrosion cracking.
19 So we're very happy with the .1 percent 20 limit on the nitrogen, which is just enough nitrogen 21 to get your strength back up so you don't have to 22 redesign your piping system when you go from 304 the 23 nuclear grade stainless steel, but no higher seems a q{} 24 good idea to us.
25 MR. BENDER: Is LN supposed to have some
376 1 strength enhancement associated with it?
2 MR. S HAC K : Yes. The LN materials, you i
3; know, from a mechanical engineer's point of view 4lthey look terrific, you know, it's free strength.
i 5I MR. BENDER: And what you're saying is 6 that they --
7 MR. S HAC K : They pay a price.
8 MR. BENDER: At one time some people said 9 they needed the additional strength. Has that gone 10, away?
11 MR. S HAC K : As far as I can tell, people 12! have been able to meet it with the .10 nitrogen 13 limit, you know, they seem to --
in addition to the I
14 .10 nitrogen, they do add a strength specification, 15j but the materials that meet those things seem to 16' meet the strength specification.
17 We've got much less data for the German 18l347, which is here, but we see the same kind of 19 ! phenomenon happening, that is, in high purity water i
20 we get ductile failures or --
I should say this 21 doesn't show that. The 347 is also susceptible to 22 transgranular failures in the sulfate environment.
23 It does seen perhaps slightly more resistant than 24 the nuclear grade because we have to go to slightly
(])
25 slower strain rates to produce the transgranular
I 377 1 stress corrosion cracking.
2 So that the one times ten to the sixth 3 l level in most heats of 316 nuclear grade we already i
4 get transgranular stress corrosion cracking in this 5 , environment. We have to go to slower strain rates 6 to get that cracking in the 347. It indicates a 7 ;
benefit. Whether that benefit will be there as we -
I ~
\
8llook at more heats or we're simply looking at a heat 9 to heat variation isn't clear at the moment.
10 ' Just to make my next set of view graphs 11 more interpretable, we normally think of hydrogen 12 water chemistry in terms of lowering the oxygen in
() 13 the reactor coolant system. The corrosion scientist j 14 ; or corrosion engineer tends to think in terms of 15 corrosion potential. [It's slightly more general i a It's an easier te'r, m to measure, and in fact 16l term.
\
17 it's more general.
18 When we try to correlate the data produced I
19 l in different laboratories and in reactor tests, we 20 get the best agreement when we examine the 21 potentials rather than the oxygen level, and this 22 goes back to the fact that although oxygen is 23 probably the most important cathodic reaction we
{} 24 have to worry about, it's not necessarily the only 25 one that we have to worry about. And so the
i f
1 378 l 1 potential actually gives us a better bind sort of
- O 2 measure what we're interested i n rather than 4
3 focusing on the oxygen alone, but this just shows 4 that it can be correlated.
5 A typical BWR' operates at around 100 6 millivolts on the hydrogen scale, so this is your i
7 typical boiling water reactor with a conventional ,
8 environment. We go down to the hydrogen water l
! 9 chemistry, we're somewhere down around minus 500 l
10 millivolts.
4 11 MR. S HEWMON : Is that potential j 12 measurement made at room temperature?
13 MR. S HAC K : Yes.
I 14 MR. SHEWMON: So it's relatively easy to 15 do accurately.
16 MR. SHACK: It's relatively easy to do a
4 17 accurately.
I 18 MR. SHEWMON: How do they measure the 19 oxygen content?
{ 20 MR. S HAC K : With an orbisphere. Stick an 21 instrument in. You would have to ask Tom or Wes 22 that question. I don't really know the answer.
23 MR. SHEWMON: Okay.
i
() 24 MR. S HAC K : The orbisphere works on a j 25 permeable membrane exchange basis of some sort.
I
,--.,--,,-,.,-,,,n,--..+.,--n -n. . , , - . - . , , .-
y ..,,--,.e.~,e, ,,_,..,,a -
m ,-,m.,..r, ,,,,,s.--n.--.n..,,_m,,,,,,,.-.,,-,.,,w.-.w...u.,,,~,-.,n,,--,,-
l i 379 1 It's difficult, but not impossible.
. () 2 What we're showing here essentially is one
- 3 argument for the hydrogen / water chemistry as a 3
4 combination of lowering your oxygen level, which 5 , essentially takes your potential down and keeping l
l 6 'your water clean, i 7 Again, your Reg Guide 156 limit is here, 8land you obviously have a trade-off now between your 9 potential and your conductivity. In this case we're 1
10 looking at the worst case additive, the SO4 so that, 11lyou know, if your conductivity is due perhaps to l l 12l dissolved copper, this is conservative, but, you
() 13 know, without requiring the utility to analyze i
141 specifically for these very low levels of impurities, i
15 it's much simpler just to focus everything on
. 16 conductivity and you can sort of see a suggested li trade-off here between lowering your oxygen and l 18 keeping your water clean.
19 Again, even with the lower potential,
) 20 you're going to have to be careful to stay with i
j 21 relatively clean water, but as you lower your oxygen 22 you actually sort of build up a tolerance for 23 impurities, j
24 Well, I don't think I've said anything so i
(])
j 25 far that anybody would find particularly f
i l
, . ., . - , _ - , . . _ . - , , . , _ _ ,__,..,-.._...m. , , . . , . _ . , . ,- _ _ _ . , , , , , . , _ . , - - . , . ,
,,l
._-_._ _ .-___._ .-- -- __ . .-- _ _ - _ _------__-_-_ _ __ ~,
380 1 disconcerting, even at GE or EPRI. The next two 2 view graphs start through the more disconcerting i
3 data, but I'm running out of time. So we'll save ;
i 4 the disconcerting data for after lunch.
5l MR. S HEWMON : Okay. I'm ready to eat.
6 7 Thereupon, the luncheon recess 8 was taken at 12:25 o' clock p.m. j L
9 - - ---
10 11 12
() 13 I
14 l 15 l
16 17 18 19 20 21 22 23
{} 24 25
l 381 1 l
/ 1 Wednesday Afternoon Session
() 2 July 2, 1986 l
3 1:35 o' clock p.m.
4 MR. S II A C K : This is data for 5 environmentally assisted cracking for 340 stainless t
6 steel. This is a fracture mechanics-type crack li 7 flow-through test. So we have a 1 T, C T specimen ,
l 3 in an autoclave in an environment, and everything I l
9 talk about is for a nearly steady state load.
l 10 We do cycle slightly but the load ratio is i
- 1) .95. So what we have is a small ripple on top of a l l l
12 practically steady state. It's just everybody has l
( lhI their own records of what they think is probably i
14 most typical in the reactor situation.
I 15 Many people use the constant load test.
I 1$ In fact there are pressure temperature fluctuations l
17 so that we take the .95 that is perhaps a slightly 1 B more conservative but still representative loading l
j 1p history.
1 i
r i
2b I showed you this graph before based on l
21 the slow strain rate CERT test which indicated the 22 regions on susceptibility in a regular stress l
2 3 relation cracking or not cracking, depending on the 24 potential and the conductivity.
)
2 5 What we have wanted to do u.a s to replicate
m-e i
j 382 1 those slow strain rate or CERT results which some O 2 people have difficulties with since we are standing 3 there pulling this thing apart and imparting very I
4 high plastic strains, and it doesn't seem very i
5 prototypical of reactor load situations. So we i
- 6 wanted to do this under conditions which are much
~
7 more 'represen ta t ive and in this case would be 8 typically what one would expect to have for cracks 9 in a reactor.
I 10 We wanted to see as we moved through the
- 11 different water chemical conditions whether in fact 12 this graph that we developed based on the CERT data
- () 13 would also describe what we would see under 14 l presumably a more prototypic level of conditions.
1 15 So, we start out in condition one, and this is now 16 essentially .1 PP sulfate and normal BWR oxygen 17 environment, and sure enough, we get a crack growth i
18 rate, and as I show you later there is some inner 1
19 l granular stress cracks just as we expected.
{
- 20 So, the first experiment we did, we simply l 21 had to clean up the water. That is we stated with 22 the conventional DUR oxygen level but we lowered the 23 impurity content typical of what you might get from 24 very good practice i n a reactor.
)
25 So that's condition four, and you can see
,___-_...y -+,,,r- w--.- y-,.w,- ,--,w.-,- ,-,,,w- , j.y y- ,y ,y.
y, , , - .- .- y. -, ,my ,.y - w w,,y---m, + , - -
. . - - =-. ...
383 1 that two of the specimens have stopped cracking.
2 The one specimen has continued to crack. Now, 3 what's a little bit surprising here, we have three 4 specimens that are loaded in the series in the 5 autoclave. So we are all receiving exactly the same 6 load. They are all in exactly the same temperature, 7 all seeing exactly the same environment.
8 We have a solution annealed specimen, a i
9 lightly sensitized specimen and a more heavily 10 sensitized specimen. Into dirty conventional PWR 11 environment, all three specimens are cracking and 12 all three of them are cracking about the same rate.
() 13 So, we have even the solution annealed specimen is 14 cracking. We clean up the water. The two specimens 15 that stopped cracking are the sensitized specimen.
16 MR. MAYFIELD: You said PWR. You meant j
17 BWR.
l 18 MR. SHACK: BWR. Sorry. The specimen is 1
19 still cracking as the solution annealed specimen, 20 which is a rather disconcerting result. We then go 21 back just to make sure we haven't flubbed the whole 22 test. We come back to condition three, and sure 23 enough, we did get everything rolling again about 24 the same rate. We then go -- I'm sorry, go to two
)
25 first. Clean up the water. Then we took it down to j
, , - , - - , r, - - - , - - - , . - - ,,,,,,,,-----,,--,,,-,,e-
y_
a 384 1 a hydrogen water chemical environment. I'm getting 4
2 them mixed up. Don't worry about the order.
4 3 Let's say we go from four back to this d
4 environment. Again, everything starts rolling again .
5 It's not an artifact of the test. You know, 6 everything grows again. We come down to the 4
7 hydrogen water chemical environments, and first we
[
8 ! go to sort after a moderate-low oxygen level, and
)
9 again, we stop the growth and sensitized specimen is 10 dead.
11 We at least slow the solution annealed 12 l specimen down. We come back up again to this one
() 13 i
just to make sure that isn't an artifact of the test, i'
14 l and sure enough everything starts to grow again at I
I 15 ; about the same rate that it always does in that I )
i 16 ' environment. And we come back down now to an even 17 lower. This is a true hydrogen water chemical.
)
18 l Well, we stopped everything dead. So, finally we i
19 ; managed to get the solution annealed specimen to i
l 20 j stop growing.
21 ! Again, we go back up, we dirty up the 22 water a bit, and we get the results that we expect.
23 In fact we get more crack growth rate. Again, a
, r^g 24 littln bit of crack growth in the sensitized I \J i 25 specimen and most crack growth in the solution 1
. . _ _ _ . _ _ _ . _ . . _ _ _ ~ , . _ _ . , . - . _ _ . _ _ . . . _ _ _ _ _ . _ . _ . - . , _ , , _ . _ . . _ _ _ _ . , , _ , . . _ . _ . . , _ __ ___
i 385 1 annealed specimens. There is our first 2 disconcerting result. Everything we have normally 3 been led to believe in the solution annealed 4 specimen should have been fine, and you know, when 1
] 5 we saw a crack in the Ballentine ring, the corner of 6 the Ballentine ring, which we are supposed to worry l
7 about on this one, not this one. Fo r- those who 8 weren't stress crack experts, we will just let the l
9 Ballentine ring go.
l 10 Now, as you would expect, the crack growth 11 that we get in the solution annealed specimen --
you l 12 probably can't see, but it's much easier to see on
, () 13 the etched surface here. The crack is transgranular .
14 So we are not getting either granular stress 15 I cracking solution in the annealed specimen, but we 16 are getting transgranular crack growth.
17 Now, when we get crack growth in the 1
18 sensitized specimens as you would expect, it's 19 inner-granular. Whether it makes you feel a whole i
- 20 lot better than not having inner-granular stress i
21 crack, again the heavily sensitized specimen grows 22 right through inner-granularly also. It's important 23 to note that perhaps we look at this in a larger l 24 context. If we take all the crack growth rate data
)
25 that people have generated over the years, we use in
386 1 assessments for growth in cracks in reactors, I sort
' O 2 of have it all plotted here.
3 Here's NRR's assessment curve. That's the 4 analytical representation of this data to be used 5 for these assessments, and there is my crack growth 6 rate for the solution annealed 340. Now, it's even j 7 more disconcerting of course because even the new 8
revised Section 11 has curves, lower curves, for 9 lightly sensitized material and higher curves for 10 l photosensitized material, but nobody really has a l'
11 curve for the solution annealed material.
i 12 l MR. SHEWMON: Is it the NRC curves or
( 13 something like them are now in the ASME code Section 14 i 11? Is that the revised Section 11 you are talking i
15 l i about?
1 16 MR. SHACK: Yes. These lines shift up and 17 l down as we go through these gates. They send me a l
18 ! revision, and I object wholeheartedly to GE's l
l 19 j position. I'm not sure where they end up. I think 20 l the two curves -- there is,one about like there that 21 i is supposed to be for weld sensitized material, 22 which would correspond to this line, and this line 23 would correspond to photosensitized material.
24 MR. ETHERINGTON: On this diagram here, is t 25 there any new data?
387' 1 MR. SHACK: The only new data that I O 2 happen to have put on here is this one data point up 3 here. I have a revised version of this with all the 4 new data on it, but you start to see a black cloud 5 of data points.
6 MR. SHEWMON: So, your main point is that 7 the transgranular growth rate is consistent with the 8 sensitized?
9 MR. SHACK: Right, it's right in the 10 middle of all the crack growth that we have ever 11 been worried about. Well, that's disconcerting in 12 itself, of course, and of course our next step is to
() 13 run exactly the same kind of experiment with 360 14 nuclear grade. So, in this test we are comparing 15 360 nuclear grade and a 340 sensitized 340 specimen.
1 16 Again, we have two specimens loaded in series in the 17 same autoclaves.
18 So, we have exactly the same loads, 19 exactly the same environment, and we go through this 20 same sort of thing. We shift from water chemical to 21 water chemical. We start out innocently eno ug h in 22 our impurity environment, and sure enough, we get 23 crack growth, fairly rapid crock growth, in the 340, 24 and somewhat lower but still steady crack growth in 25 the 360 nuclear grade. ;
1
.. ,. . - . - . w. -- ._ . . . ,. . , _ . . - .- . , .-_ . , - . ,
1 388 1 We then switch to the high impurity
() 2 environment. So, this is a conventional BWR 3 environment, usual dissolved oxygen level, but we j 4 clean all the impurities out. The sensitized j 5 specimen stops dead in its tracks. The 360 nuclear i
6 grade doesn't even seem to change the water chemical .
i 7 Just to make sure that we haven't branched the 340 8 crack, and so we can get everything going again, we 9 go back to our impurity environment.
4
! 10 340 takes off again. 316 is just trucking 1
1 i i 11 right along. We then finally go to a hyd rog en water i
4 12 { chemical department, and we stop everything dead in i
() 13 t
its tracks. We now thought, well, we went to the l
14 I high impurity environment, and we had started in the l
15 impurity environment, and if we switch to the high i
16 impurity environment, the crack kept on growing.
j 17 We now have a crack that's stopped. Maybe 18 it won't start again in the high impurity 19 l environment. So, we go back to the high impurity 20 environment. The crack takes off on the 316. The
- i 21 l sensitized stainless steel keeps standing still.
22 It's all very disturbing at the moment.
23 What's different about our tests than j 24 everybody else's tests? Hell, one thing that's a )
)
25 little bit different is that we run our tests in not i _
4
389 1 a pure constant load but a t. .95. I would argue what 2 we are seeing here is not fatigue loading, although 3 we are cycling a little bit.
4 For one thing .95, the Delta K we are 5 talking about, is about 2 ksi. Almost anybody's 6 notions of a threshold for heat crack is less *;an 7 that, and in fact when we are in the right 8 environment, we can cycle this thing all day and it 9 doesn't want to grow. So, it's not the heat crack.
10 The cycling loading cou1d play a role. So, we run 11 our last stretch now in high purity water, but 12 instead of the .95 load, we go to a purely constant O la ioed.
14 Well, you can't see it so on this, but 15 this crack is now standing still. 340 is starting 16 to grow in a high impurity environment. So, in a 17 high impurity environment with a load, 340 stands 18 still and 316 nuclear grade grows. In a high purity i
19 environment with a constant load, the 316 nuclear 20 grade stands still and the sensitized 340 starts to
- 21 move. Obviously we are very confused at this point.
22 The major circumstances that I would like 23 to point out here is this portion of the curve i 24 whereas we go impurity to the high purity
)
25 environment, the crack keeps on going in the 316
l 390 1 nuclear grade.
O 2 And I want to show you the raw data before 3 because, you know, it's always better to see just 4 what these things look like as raw data rather than 5 after I have proved it here and clouded it up like 6 this. This curve basically tells you basically what 7 you saw in the previous graph. I won't go through 8 it. It just puts it in a flow chart form.
9 MR. SHEWMON: Tell me the condition of 10 these again. The 340 was sensitized and the 316 was 11 not?
12 l MR. SHACK: Is not. It's virtually I
( 13 l impossible under any realistic heat treatment to l
14 sensitize the 316 nuclear grade. If you soaked it a ;
15 l ,
thousand hours at 600 C you might sensitize it. Now ,
1 16 if we go back to the same usual batch of data that l
i 17 we have for crack growth in these environments, we 18 i see the 360 nuclear grade fits right in the middle I i 19 '
of all the other data that's been collected around.
20 Just to show that we did even worry a long i
21 time ago about this cycling load versus this 22 constant load sort of thing, and although it 23 sometimes seems to make a difference and sometimes 24 not to make a difference, there is a series of tests 25 l we ran probably two and a half years ago that i
'l f
.. .- ~ _ . ..
391 1 convinced us that we really didn't have to worry O 2 about the changes and the growth histories because 3 we ran this one through R ratios from 1 down to .5 4 and changed the frequencies around a little bit, and 5 it just didn't do much for the crack growth rates, 6 especially when we could compare them to .9 data 7 with the constant.
8 So, we were fairly confident it wasn't 9 going to make a big difference in our results. The 10 two sets of data were basically interchangeable. It 11 certainly doesn't seem to be the case now. But even 12 in this case, again, we were getting -- we run the
() 13 pure constant load and we do get inner-granular 14 growth. When we go to the load, although the crack 15 growth rates don't change very much, we do change 16 the load.
17 Well, I'll change the order of my 18 presentation slightly here. Where did we go from 19 here? What we need really are better definitions
- 20 for the conditions for which we get the crack growth 21 in the type 316 nuclear grade. So we are now 22 running a series of tests. I should have pointed i 23 out we really ran those tests that I've shown you at 24 basically 1 K. The only variations in K were those
)
25 that were sort of associated with the crack growth.
1 1
392 i
1 There might be threshold effects. It's O 2 possible that in fact if we went to a lower K we 1
3 would find a more rapid drop off in 316 nuclear 4 grade crack growth rate than we do with the 340. We
, 5 have seen that there are obviously effects here of R 6 ratio and frequency.
a 7 And, again, for the methodic point of view ,
8 we know these are all tied together in the crack tip 9 strain rate. Ne don't know enough about the 10 fracture mechanics to execute the crack tip strain i
1 11 l rate. So, we are more or less forced to investigate 12 ,
those things empirically at this point.
i ,
I (T l
(_j 13 i We still think there are purity ctfects.
l I l 14 Although our crack growth rate tests aren't showing 15 ; much of an impurity effect, we do see a difference 16 in the CERT testa, and it's possible that whatever 17 results we see on R ratio frequency threshold will 18 also be affected by impurity. So, we have to map 19 this out. It's just uncertain at the moment.
20 '
The question is do the German nuclear 21 grace material and the cast materials show the same 22 behavior, or do we have some material that's readily 23 available, readily usable, that is inherently 24 resistant to this crack propagation in BWR. My 25 guess is the TP 347 is not, but we will simply have
393 1 to run the test to find out.
O 2 MR. RODADAUGH: Bill, when you talk about 3 type 347, you are referring to something more 4 restrictive than the standard?
5 MR. SHACK: Yes, the only 347 that anybody 6 I I
is showing any interest in is the German. That's 7 why I keep showing it as the TP 347. TP stands for 8 something in German.
9 MR. RODABAUGH: It also stands for type, i 10 MR. SHACK: I don't mean type. I mean TP.
11 A German 347 with their specs. We have become very 12 concerned now about crack initiation. We certainly
() 13 don't seem to have much margin here in terms of 14 crack growth, yet there are all these other kinds of 15 tests, the pipe tests that I initially mentioned 16 with the 360 nuclear grade showed normal improvement l 17 over the sensitized stainless steel.
l 18 We don't really understand the difference
( 19 between initiation propagation very well. We have l 20 become very interested now in anything that might 21 start cracks. One of these is certain surface cold 22 work. We still have a little bit of work-up here.
l 23 One thing we do know about 316 nuclear grade is that l 24 it can be susceptible to cracking under bulk cold 25 work. Even GE will admit that. All the pipe tests
r l
394 1 .
work was basically done on four inch pipe, and the l 2 related URC research program at Battelle Northwest l
3 laboratory, they have been busy measuring 4 essentially the cold work associated with growth.
5 So, if you think about it, you take a 6 large diameter weldment rather than a small diameter l 7 , weldment, you're essentially cycling the inside 8 j surface with the 30 passes that you use to make the 9 weld, and the temperature actually drops off fairly 10 quickly after the first couple of passes, but still 11 sees strain cycles.
12 So, they are estimating that effectively
() 13 there is a great deal of cold work on the inside 14 surface to have that large diameter pipe simply from 1
l 15 i the welding.
16 MR. SHEUMON: Is that still transgranular 1
i 17 ' then?
l 18 MR. SHACK: That's granular in that case. '
19 j GE itself gets inner-granular cracking in cold work.
)
20 l They did their cold work on welding plates. We are i
21 interested in seeing if we can go back and look at 22 one of these large diameter weldments where in fact l
23 the cold work can be associated with the weldment 24 and will also give us similar effects. That's
)
25 another piece of work that has to be done.
.. - .-_. - - -- .~ __
395
(
l We have become -- again, we think that l 2 these impurities will play a role in initiation, but s
3 now that we are so concerned about propagation, we l
4 have to go back and define this data a little bit.
5 What we are trying to do here is look at some 6 defected specs.
7 one of the troubles with initiation is 8 that if you stop and test when the crack is 9 initiated and very tiny, how do you find the crack?
I 10 What we have tried to do is try to introduce defects 11 so we know where to look. That's one thing, to put 12 a small notch under an SEM. It's another thing to 13 try to do an SEM scan over the whole gauge section 14 of the specimen. Again, we still have the problem 15 of transient water chemicalu. How long does it have 16 to be under.
- 17 MR. BENDER: When you say weld cold work, 18 do you mean weld shrinkage? Is that the kind of l 19 cold work you are talking about?
4 i 20 MR. SHACK: No. I mean the plastic l
21 cycling that goes on on the inner surface simpl'y 22 because everytime you lay a weld bead down, you l
23 essentially have a thermal stress cycle laid on the 24 pipe. So, there is a cycling plastic strain the 25 inner surface goes through, and if you believe the
396 1 measurements and the element calculations -- Libicky O 2 has done element calculations for us, and if you 3 look at the equivalent on plastic strain, we get 4 something like 15 percent strain on that surface.
l 5 The measurements at Battelle Northwest 6 give you the bigger numbers, but even if you take 7 Libicky's 15 percent, that's a lot of cold work. I 8 should say semi-cold / warm work.
9 MR. BENDER: In the crack initiation, when l
I 10 does it start? And sometimes that period of 11 , incubation is two or three years. What does your l
12 l work tell us about that?
( 13 MR. SHACK: That's a hard -- again, two or 14 j three years is a lifetime in a laboratory test. We 15 l hope to be able to look at the initiation by l
16 ; comparing initiation in sensitized materials and 316 17 nuclear grade under accelerated conditions. We are 10 l going to have these defected specimens. Some of the l
19 ! defected specimens will be sensitized 340, some will 20 be 316 nuclear grade. We will take them out to very 21 small overall strains and see if in fact we initiate 22 the cracks sooner in the sensitized 340.
23 What we are looking at is that rather than 24 a absolute standard for initiation, which as I say 25 is a long term laboratory test, we are looking for i
397 1 comparison between the sensitized and the 316 2 nuclear grade.
3 NR. BENDER: I don't have any problems 4 with doing things in the short term, but I think 5 it's important not to ignore the fact that whatever 6 it is may take some time to develop, and you may not 7 see it in many of the laboratory tests.
8 MR. SHACK: Yes, that's true.
9 MR. SHEWMON: Before you go onto another 10 subject, seems to me a few years ago when GE was 11 putting in ferrite pipe to replace some of the 340 12 and stagnate lines, somebody told me that ferrite
() 13 pipe had a fair transgranular stress crack rate once 14 it initiated, but that has performed well, has it 15 not, or none of that has shown up yet? i 16 MR. SHACK: One of the things that we do i
17 know on the basis of the GE work about the carbon 18 steel pipe is that there is a very high threshold i
19 for that crack propagation. It's true that once the 20 crack initiates, it propagates at a good healthy 21 clip, but it doesn't initiate until you get it past 22 the threshold of about 32 ksi square root inches.
23 MR. SHEWMON: Now, the thing about this, 24 do you know anything about the stress for the 25 threshold here?
i 398 1 MR. S !! AC K : No, that was one of the things 2 that I point out here. That seems to be one of the 3 major questions that we have to answer with the 316 4 nuclear grade, will we find a similar threshold 5 effect. That's almost the next set of tests to be i
6l run an the 316 nuclear grade.
7 MR. SHEWMON: Any other questions? Onward .
8 MR. SHACK: Okay. One of the other things 9 that we are doing are essentially we are trying to 10 redo the pipe tests on a much smaller scale to see i
11 '
again if we can get these kinds of phenomena with 12 what seem to be more prototypical kind of loads l !
- f^)
s_ 13 i rather than at CERT test or the pipe tests.
14 j On the view graph here we are finding we 15 l1 are getting break before lead conditions in a 340 1
16 stainless steel piping system because we are !
17 initiating cracks on this diagram. Our pipe test 18 specimen is down here. There is a fitting up at the 19 end with a weld in it. We are getting cracks going 20 right through the weld metal. They are essentially 21 initiating from flaws in the weld.
22 What's interesting here is that under 23 these circumstances we essentially get break before m 24 leakage in the 340 stainless steel, which I think 25 goes back to confirm the wisdom of the NRC. When
399
)
1 you are having essentially crack system phenomena O 2 that can be controlled perhaps by residual stresses, 3 then it's very wise not to rely upon the leak before 4 break.
5 The thing that's most interesting here and 6 perhaps somewhat puzzeling is that we can account 7 I for the initiation of this crack by the notch that 8 we have on the inner surface, but the notch only 9 covers about 180 degrees of the inner surface, yet 10 we get crack growth that essentially goes around 360 i
11 degrees on the surface.
12 Now, one of the things that's very
() 13 different about our tests from the real world piping i
14 system is we have gone to a great deal of trouble to 15 eliminate all bending moments from this test. So, 16 you have the pure axial load. Again without any 1
17 bending component at all, of course we have driven 1
18 the crack through on the bent test side, but since 19 we went to a great deal of trouble to get pure axial 20 loading, I find it remarkable that we still even in 21 this case manage to grow the crack 360 degrees 22 around the circumference when we can only find an 23 initiating notch on the 180 degrees.
24 My only explanation for that is that the
)
25 crack fatigue like the stress cracking is something l
l
400 1 strongly influenced by residual stresses. If you O 2 may have the right kind of residual stress pattern, l.
3 you get a crack tha t grows through and then 4 essentially is slowed down by a residual stress 5 pattern on this through-wall growth permitted to 6 grow around the pipe giving you that uniform crack i
7! growth before the final failure.
8 MR. HUTCHINSON: If you are el im ina t i ng 9 j bend, then you are essentially forcing the crack I
10 j j around that side, is that right?
MR. SHACK: But even in the pure bending 11 l 12 case, if I initiate a crack without residual stress
() 13 I
but in a pure axial stress state without residual 14 j stresses, I still should grow through roughly a two i
15 to one basis. So, even in a pure axial mode, 16 without residual stresses, I should still pop 17 through with an elliptic crack.
18 : I have to have something thet slows the i
19 , through-wall growth down to let the circumferential l
20 growth catch up. Again, to get this uniform crack 21 growth, that bending stress has to be kept fairly 22 small. That is that through cracking can't get too 23 large before it grows around and evens everything 24 out or else it would have popped through before it 25 made that uniform growth.
401 1 This is just the way we are looking at O 2 surface cold work. We are preparing specimens with 3l different degrees of surface cold work. They will 4 be essentially clamped down and in between these 5 specimens and essentially put in an autoclave. We 6 will be looking with different degrees of surface 7 cold work for initiating cracks on 316 nuclear grade .
8 This is our defected CERT vessel. That is 9 a typical CERT specimen. What we have done in this 10 case is to drill some very small holes, and of 11 course this gives us a strain localization, but more 12 importantly it tells us where to look for the first
() 13 cracks in the specimen, and so instead of preparing 14 up usual CERT tests, this is done with failure that 15 will stop at perhaps one to two percent overall 16 strain, which perhaps will just initiate cracks, but 17 when we take the section through the small hole, it 18 is something we can examine very carefully under an 19 SEM in a reasonable amount of time and we will know.
20 where to look for the possible crack initiation.
21 Just a few quick words again about the.
22 fabrication. There were some initial problems with 23 microcracking in the 316 nuclear grade during hot
(~N 24 bending operations. People have now learned how to
\J 25 control those operations, and that problem seems to
l 402 1 be fairly good.
O 2 Looking through the data in the literature ,
3 you find out that the 316 and 316 nuclear grade are 4 more susceptible to weld pipe cracking than type 340 ,
5 at least on the basis of conventional hot ductile 6 tests that people use for these things, at least 7 some questions about the adequacy of this hot 8 ductility for the higher strength load for nuclear 9 piping systems, which is a question people have. ,
l 10 Again, 347 has a miserable welding history which 11 makes people very nervous about using it.
12 {i MR. RODABAUGH: This time you do mean type ;
()
i 13 f 347?
14 ! MR. C li AC K : This time I do mean type 347.
15 Again, we have now learned something about such are i
16 l welding. We know that weld toughness can be very 17 different from the base material toughness with 18 these new welding processes and configurations.
19 j Again probably should put in 316 and 340 t
k 20 4 welds here, too, but where you make these piping 21 replacements, typically there is some part of the 22 piping system that doesn't get replaced. You have 23 to come in and somewhere along the way having sort 24 of a dissimilar weld. In this case we are 25 interested in 347 to 340.
403 1 Just some data, if you look at one of the O 2 standard sort of tests with the resistance to hot i
I 3j cracking, all this is really showing the 316 is less 4 resistant to this cracking than 340. Although it's 5 less resistant, it may have perfectly adequate hot 6 ductility. All it indicates is that it's something 7 to think about, that you are paying a certain 8 penalty here, but the low carbon 316's don't seem to 9 be any different than the rest of the 316's. It's 10 something inherent with 316.
11 This is one case where people have made 12 changes in their specifications for 316 nuclear 13 grade. This is some data that's picked up on 14 solidification of weld metal. People look at this 15 sort of thing and come to some conclusions. It's 16 the effect of phosphorous and sulfur on essentially i
17 cracking, but it also applies to weld hot cracking. '
18 You find out the chrome wiggle equivalent i I
19 will say 1.6, but you can pull a fair rate of l
20 phosphorous and sulfur before you get weld hot l
l 21 cracking. With a chrome nickel equivalent of say 1. 2, ;
22 you need a good clean steel. If we happen to take a 23 look at this graph which Argonne said I shouldn't
(% 24 show to anybody, what you will find is that all the
%)
25 340's, your chrome nickel equivalent, are up 1.5,
i 404 1 1.7, 1.8.
O 2 MR. ETHERINGTON: What was the formula for 3 the chrome nickel equivalent?
4 MR. SHACK: It's standard --
5 MR. ETHERINGTON: Okay.
6 MR. SHACK: The numbers are at the top.
7 If you come down to the 316 nuclear grade, you find I
8 you are down 1.3, 1.1, 1.2. These are essentially 9 heats of 316 nuclear grade that were used in the 10 original GE test program. So, if you order 316 off 11 the shelf, this is where you end up. So, you have 12 l to keep very careful control of the phosphorous and I
() 13 f the sulfur, and people have now learned this in the 14 specs for the 316 nuclear grade.
15 They buy from Japan where they can get 16 ; very low phosphorous or sulfur, or else they specify l
17 i a chrome and nickel that moves them up a little bit i
18 higher. So, it seems to have solved the initial 19 ! difficulties people were having in welding this I
i 20 , material.
21 347 as I mentioned is having a problem.
22 We are working with EPRI and New York Power 23 Authority. This is a weld that was made down at I
r% 24 EPRI, and it was the first cross-section that we U
25 made of the new German 347 weld. It gave us a crack l
l _-_
405 1 right at the weld fusion line. It's sort of a 2 disconcerting way to start the whole experience, but 3 again we really don't like to have starter cracks in 1
4 our weldments. )
5 The general experience with the German 347 )
6 actually has been very good. The EPRI people think 7 it's a very weldable material, but people at the New 8 York Power Authority have made up some test welds, 9 and they think it's very weldable. Mr. Dean at the 10 University of Tennessee is doing hot ductility t 11 testing. He thinks it's a very good material. We 12 have essentially cooperative efforts from all them.
() 13 The NEC Center, the Power Authority and 14 the University of Tennessee are primarily focused on 15 the welding. We are primarily focused on the stress 16 cracking, but we are all sharing materials. So we 17 have sort of a consistent picture of what's going on.
18 We will be delivering some of these weldments to meg 19 for fracture toughness testing.
20 MR. Sil EWM O N : Do you know why that l
l 21 cracking did occur in the first run and whether it ;
22 occurred never again or in all of them?
23 MR. S il A C K : The EPRI didn't want to use 24 the German process for the welding. The Germans --
25 perhaps it will be clearer if I put up the slide
w ea 406 1 here. The Germans use a narrow gap welding process.
O 2 It's a very deep gap. No insert. It's very 3 different than U.S. practice, and they feel very 4 comfortable with it.
5 EPRI wanted to develop a procedure that 6 was a little more like the welding procedure that 7 the utilities have been using and all the welders 8 ,
have been using. So, when this weld was prepared, 9 that was probably the first or second 347 weld 10 l trying to develop their own U.S.-like' procedure. So ,
i 11 l we are looking here as an early initial effort to 12 develop a U.S.-type welding procedure. The German
() 13 l i
narrow gap welds were not only examined at Argonne, 14 ! they have been produced by KWU and Welding Services, i
15 !I Incorporated. New York Power Authority just picked I
16 l!
a welder and said, try the German process and see 17 how you like it. You know, do it their way.
18 MR. SHEWMON: Is it a full wire?
19 MR. SHACK: Yes, it's a full fire wire.
20 l It's a very narrow gap.
i 21 MR. SHEWMON: 308?
22 MR. SHACK: No, it's an actual 347 weld 23 metal. So, this is a whole 347 system. We now have 24 these weldments. We haven't done the metallurgical 25 sectioning, but the radiographs and UT and the dye
407 1 penetrant checks that were run on them, all these
' O 2 welds look very good. We can't find any good i
3 defects in any of them.
4 We will be checking these, and as I say, 5 are mostly now completing the metallograph. We 6 looked at the weld process, and we will be using 7 these as materials to look for stress crack testing i
8 and fracture toughness testing.
9 That gets me through pretty much of sdat I 10 wanted to say on the alternative m a t e .r i a l s . Perhaps 11 a little bit of work on weld normal lays and the 12 MSIP process. What we have been d o l ., q with the welal
() 13 overlays is taking some of the wel6 overlays that ,
l 14 1 have been formed on actual reactor components in the 15 field. That ic when plants have replaced their 16 piping system, we have taken the overlays that they 17 have removed in the process and brought them in for l 18 examination to see what they look like.
19 Thus far we have looked at two pipe to 20 elbow weldments, two 20 inch pipe to end cap 21 overlays. So, we have looked at four overlays thus 22 far. These were all reported as 360 degree 23 intermittent cracking. We seem to be batting about 24 500. One of the elbows and one of the ends caps was 25 complerely free of cracking when we examined it.
i 408 1 The other two had a number of short axial cracks and 2 ; rather deep circumferential cracks, but they were 3 all very short. Basically the longest crack we had 4 was like three centimeters and they tended to be 5 clustered in narrow regions.
6 Heavy post-weld grinding again. These 7 ;
people essentially cut notches. I'm not sure --
8l i whenever they were removing on the radiograph, they 9
didn't seem to care what they left, so long as as it l
10 l passed the radiograph. They basically ground i
11 ! notches into these weldments, which is probably what ,
12 the MEA guy was seeing when he was reporting these
() 13 cracks, the notches that the grinders left there.
14 l Interestingly, most of the cracks occurred 15 in the forge components, that is if you looked at i
16 the end cap to the pipe and elbow to the pipe, most 17 of the cracks were in the end cap and the elbow.
18 Most importantly, the cracks d id n ' t propagate during i
19 l the application of the overlay or in service 20 ft afterwards.
21 This is a picture of an overlay on the end 22 cap to a recirculation. So, this is like a 20 clear 23 inch line with the hemispheric end cap. The weld l
24 was actually under here, and they saw the defects 25 l and came back and puttered on this big band-aid.
409 1 Our evidence of the crack propagation was O 2 found mostly in the metallograph. We have looked at 3 the crack tips. The cracking, as we expect, is 4 inner-granular. It's blunted. We see no evidence 5 of any transgranular tearing, which is what we might 6 expect to happen during the application of the 7 overlay if those stresses were too high at the crack 8 tip, nor any evidence of a fine crack tip beyond the 9 blunt.
10 The blunted crack tip should mark the 11 application of the Ills I . Any crack beyond that 12 blunted region would indicate to us crack growth (O,f 13 that occurred after the application of the overlay.
l 14 Since we don't see that, I assume basically the 15 crack was blunted and has been arrested ever since.
16 The only crack that was of any slight concern was 17 this one axial crack which g ives some hints that 18 it's a growing axial underneath the overlay.
19 It becomes a little confusing because we 20 have lost the information on what the original weld 21 looked like. So, we are not clear whether this is 22 growing in the heat affected zone of the o r ig inal 23 weld or the heat affected zone now produced by the 24 overlay on the outside surface of the o r ig in al pipe.
(")%,
u- i i
25 But as I say, that was the only slightly j l
I l
n ~ - - - _ .- e ,-- -
410 1 disconcerting piece of information that we have O 2 gained about the overlays from the field examination .
3 We have these reported crack indications : )
)
4 along this inner surface. We actually did find a 5 crack along here. One of the things that they were 6 seeing was e s s e ra t i a l l y an undercut at the weld 7 fusion line obviously spotted with the UT. That's I
i 8 one reason they recalled this plate.
l 9 The major residual stresses on the overlay, i
10 [ again we are looking at the inside surface of the i
11 ! pipe now trying to see if in fact as we expect the l
12 l overlay managed to produce a fairly favorable
() 13 l residual stress state. They are not as quite as l 14 compressive in this configuration as they are in i
15 i most of our lab reporting mockups where we have 16 l worked with pipe to pipe welds.
l l l l 17 l The pipe to elbow perhaps has a little bit l
18 l more flexibility. We get much more asymmetry in the 19 results. The pipe to pipe test is almost purely
(
l 20 i isosymmetric in that one is somewhat more asymmetric . ,
i 21 But even the worst location here we still have a l
1 22 significant compressive residual stress.
23 MR. S II E W M O N : Before you go back to the 24 earlier. You talked about a couple of reasons for O
25 overhauls. One was the grooves and grinding and the
411 l
1 other was an undercut on the weld. Do you know O 2 whether EPRI or the NRC in their tests of samples 3 that they check people out or look for things like l
I 4 that or what happens --
l 5 MR. SHACK: I'm sure nobody deliberately )
6 :: t through this into their weldments. j 7 MR. SHEWMON: Perhaps that's unfortunate 8 if that's the way they are out in the field.
9 MR. SHACK: I guess I can't answer the 10 question as to exactly. They are making up test 11 flows, they are using segments from reactor 12 components, but exactly what those are, I simply j () 13 don't know. One thing I should point out, this was 14 in service for a year and a half. One of the 15 questions that people always have about residual 16 stress remedies, okay, I have produced favorable 17 residual stresses. Do they stay there? Does it go 18 through a cycle or does it relax?
19 Again, 18 months isn't a very long time, 20 but at least the initial results look favorable.
21 The pipe to end cap overlay looks a lot more like 22 what we expect of the mockup flow. It's more nearly 23 axial symmetric and everything is very compressive.
24 So, almost looks like our models.
(~))
x 25 I think I'll, in the interest of saving
e 412 1 time, skip over the rest of the resul ts of these O 2 overlays. Well, perhaps another one. I don't know 3 if I made this point before. Let me skip ahead to 4 one result that is perhaps of some interest. These i
5 are some through-wall stress measurements that we 6 have made based on mockups that were prepared by us 7 for Newtech.
8 What I'm trying to show here on the 9 measured through-wall residual stresses, that is 10 essentially blocks, and what I have done, because I l I 11 didn't see much variation or as much axial variation 12 under the overlay, I took all my through-wall stress
() 13 measurements, and I just sort of dumped them on one 14 f curve to give you some sort of an idea what the I i
i 15 i general data cloud looked like.
l \
16 The solid curve is an element calculation i l
17 of the residual stresses produced by a weld overlay.
18 The reason for showing this is, of course, this is i
19 ! the best distribution that we measure, is of no real 20 l interest. What you are really interested in are the 21 residual stresses at the point of the crack that 22 happens to be over that overlay.
23 We can't measure the residual stresses in
~
24 front of a crack. The only thing we can do is 25 execute them, and we have to execute them using
413 1 finite elements, and you are going to have to O 2 believe the results when you are done. So the best 3 we can do is at least to show when we don't have a 4 crack. The finite element program seems to give us 5 a reasonable indication of what residual stresses 6 look like. So this is essentially a validity check 7 for our finite element program.
8 MR. SHEWMON: I'm struck by how high the 9 stresses are that you find relative to the yield 10 stresses, which was half that, as I recall.
11 MR. SHACK: Right, that comes back to our 12 thing about the weld cold work that you find on the
() 13 surface of the large diameter pipe. This thing has 14 been cycled thro ug h many , many cycles of plastic 15 strain. GE has done some hardness measurements. We 16 haven't done the microhardness measurements.
17 MR. SHEWMON: But you are not surprised to 18 think that the yield would be up to 80,000 from 40 19 or 30 or something?
20 MR. SHACK: No. If you look at 340 stress 21 strain curve and you go out the 12 percent, it's a 22 very reasonable kind of stress to end up with. We 23 just accumulated a fair amount of plastic strain on 24 the surface.
25 MR. RODABAUGH: 80,000 is a little high to
n 414 1 go out 12 percent.
\ 2 t1 R . SII AC K : The 80,000 point, right, 1 3 might. Now, we have sort of got some confidence in 1
4 the program. We have gone back and looked at a weld S overlay over a pipe with a 50 percent t h ro ug h-wa ll 6 360 degree crack. We don't want to do any three 7 dimensional analysis here. So, all our cracks are 8 always 360 degrees, 50 percent through-wall.
l 9 We are now looking at the stress data and l 10 the crack opening displacement around that crack 11 I with an applied axial stress of 9 ksi on the pipe, i
12 l which is a fairly typical sort of axial mode in a
}
() 13 i
piping system. We see finite element result is that 14 ; we look at the crack tip. That is during the l
i 15 ' application.
! 1 16 The overlay, the crack type, has gone 1
l 17 tensile. It's been blunted out, and essentially as l l
18 we get unloaded, from the history, we closed the 19 ,
crack back up behind the crack tip, and now with the i
20 ,
application of the 9 ksi load, we have essentially 21 ! reopened the crack, but we haven't opened it all the 22 way out to the crack tip so that what we are doing l 23 essentially underload, opening the crack up that we 24 haven't managed to open the crack all the way.
)
25 This gives us a rather complicated stress
415 1 statement around the crack tip because of course we O 2 have a stress here. It goes compressive as we are 3 squeezing the thing together here, back up to 0 at l
4 the blunted region, and then in front of the crack 1 l
5 we have a compressive stress state, which again is 6 supposed to give you a warm feeling, but maybe the 7 crack won't go anywhere with those compressive 8 stresses on the front end of it.
9 Your luck runs out of course if you run 10 the stress up too high. If you go to 15 ksi, which 11 is not unknown in a piping system, it had been a 12 pretty high stress, but there are some joints where
() 13 it would go that high, and we see that we are now 14 back up to essentially tensile stresses in front to 15 have that crack tip. You essentially blocked 16 yourself quite a bit in terms of a favorable 17 residual stress state, even through they are a very, l
18 very severe crack. We are talking 360 degrees, 50 19 percent through-woll. So, this is a large crack 20 that we have postulated in the pipe.
'21 This is probably too technical. We will l l
22 skip the next few. One of the things we are 23 interested in is the weld overlay. In addition to 24 giving you a favorable residual stress state, as 25 always thought, the overlay material itself is
1 i
- 416 i
1 highly resistant to crack propagation. In fact, O 2 even if you had a crack that went all the way l !
l 3 through the overlay, the weld overlay material, then 4 it essentially arrests the crack.
5 If we are trying to measure the inherent 6 resistance to that weld overlay material of the 7 crack propagation with test specimen something like 8 this, what we have done is taken a pipe and l
9 photosensitized it and then lay in a weld overlay 10 ; over the top of the pipe. Then we have come through l
11 ! and cut the chunk over that pipe over layout and l
12 ;l then electronic beam welded some additional material
() 13 ;
over the front face and the back face so we get 14 I standard C T specimen th we can go off and test.
I !
15 Now we crank up and we drill a crack 16 through the sensitized material. It's supposed to 17 come out and hit the overlay, and we can see whether
)
18 it arrests or not. We actually first try to do this.
f 19 with what we might consider the right way. We make i
20 l a weld and we don't furnace sensitize the base metal .
l 21 f MR. HUTCHINSON: Aren't there tensile 22 stresses that develop in the overlay itself?
23 MR. S !! AC K : Yes, the overlay itself is in 24 tension, but it's insulated from the environment.
25 So, yes, everything has to stay in equilibrium.
J
417 1 Fortunately it would be nice occasionally to get rid O 2 of the equilibrium and have everything fresh.
3 MR. SHCWMON: Then I don't understand your 4 last remark saying that in connection with this --
5 MR. SII A C K : Oh, the crack would never get 6 the overlay. It would have to go through the 7 compressive stress field to get there. If we are 8 correct about compressive stress, that won't happen.
9 I would like to go on now to mechanical 10 stress improvement process. This is associated with 11 0'Donnell and Associates. It's another way of 12 inducing a favorable residual stress state on the
() 13 l inner-surface of a pipe weld. We are all familiar 14 with I ll S I . This is an alternative one. What you de 15 is you essentially put a big massive collar around 16 the pipe. It's a split ring device, and there is 17 such things as shims in it so you control the amount 18 of squeezing, and then you put the split ring over 19 it and then squeeze the split ring down on the pipe 20 and plastically deform the pipe. And this 21 particular deformation essentially induces favorable 22 residual stresses on the inner-surfaces of the pipe.
23 MR. E T li C R I N G T O N : That's an interesting 24 device. Is anybody showing any signs of interest in 25 this?
418 1 MR. SHACK: Yes. Yes, utilities would use O 2 it if they could get --
you know, it's one of these 3 l chicken and egg things. The utility wants credit 4 from the NRC, and NRC says it's an interesting idea, 5 but we need more data, and we are not going to give 6l you credit at the moment. Why don't you put it on 7 and we will look at it for awhile. So it goes 8 around.
l 9 Many people I think are planning to use iq 10 on when you replace the pipe there are always a few 11 joints that you can't get to. So, this would be 12 ,
their solution for those few joints that they can't
() 13 replace. They would go through a squeezing process 14 and hopefully get some threaded out.
15 , MR. SHEWMON: Is this the uniform 16 squeezing so it upsets out the ends?
17 MR. SHACK: It mostly bends the pipe. It 18 is relatively little extrusion. So you are 19 : basically bending the pipe and stretching it.
20 ! MR. ETHERINGTON: This is a radial 21 compression?
22 MR. S il AC K : Yes. I don't have any 23 application of the tool. These pictures give you 24 some idea of what a pipe looks like after it's been
)
25 MSIPed. You can see the marks on the surface where 1
419 1 the tool has been applied. This is sort of a waffic 2 like space that leaves a mark there, and you can see 3 the substantial squeeze-down on the pipe. Again, 4 this particular pipe that you are looking at has 5 been subjected to the maximum two percent strain.
6 So, it Coke bottles pretty good.
7 It's only applied to one side of the pipe.
8 So it can be used with complex geometries. That is 9 if you are coming up to a valve or some sort of a 10 fitting, you are only really applying this tool on 11 the pipe side and squeezing down on it. There is no 12 need for cooling of the inner-surface.
l () 13 l
This simplifies the scheduling, and I 14 l think O'Donnell claims it to be true, and I think 15 most people would agree that it's simply cheaper and 16 faster than IHSI. Instead of trying to bring 340 17 kilowatts of power up to the pipe, you are walking 18 in with a mechanical tool.
19 It's just a simpler sort of process.
20 There is no need for the cooling. You simply follcw 21 behind the welder, one weld behind him, squeezing 22 the pipe behind him. There are no drawbacks to it.
23 It certainly seems like a competitive process.
24 MR. SHEWMON: How many men and a boy does 25 it take to carrying this rig?
g-420 1 MR. SHACK: The 12 inch one isn't too bad.
O 2 I haven't seen a 28 inch one yet. I assume they 3 have some big husky guys and a crane. The 12 inch !
4 two men can easily handle. It's a fairly movable 5 sort of thing. The big one has got to be a hunk.
6 MR. RODABAUGH: Bill, these pictures here i
7 show you put a straight edge on the pipe. Is that 8 uniform?
9 MR. SHACK: Yes, that's uniform as you go 10 around.
11 MR. RODABAUGH: There is some property of 12 l the clamp that makes it uniform? l s 1
() 13 MR. SHACK: Makes it uniform around. It 14 . looks very much like the pipe walk that you saw i
15 ;
before. It's got that same sort of waffle tool in 16 it that sort of evens out some of the distribution 17 of the load, and it seems to make a fairly uniform ,
i 18 even squeeze.
I 19 l O'Donnell maken a big point of the fact i
i 20 l that there is a monotonic plastic flow with no l a rg d 21 tensile modes on the inner-surface like the reverse 22 plastic flow, what you get with the induction 23 heating process. Just how important that is to you,.
24 I'm not sure. He likes it as a selling point.
25 One of the things they did as sort of a I
421 1 verification was this mockup weldment that vermont-0 2 Yankee made up, and they took 12 inch pipes and they 3 made two welds on the pipe, and they welded on a ,
)
4 base plate on the bottom. They applied the MSIP 5 tool to squeeze this weld right here, then it went 1 6 down to J.A. Jones and the EPRI center, and they 7 filled this thing up with boiling magchloride and 8 cooked it for a day in boiling magchloride in 9 austenitic stainless steels.
10 If you have tensile stresses anywhere, 11 boiling magchloride will do a nice job of cracking 12 it. They did the magchloride test. They cracked
() 13 the hell out of this weld, which wasn't protected by 14 the MSIP. J.A. Jones could find no cracking in this 15 weld or this weld.
16 MR. HUTCHINSON: Why is it squeezed just 17 off the weld? Why not center it on the weld?
18 MR. SHACK: When you go through the 19 analysis, you find out that the biggest compressive 20 stresses are over here, a little bit off to the side ,
21 and in fact, they predicted they get slight tensile 22 stresses on the inner-surface right under the tools.
23 That's what their finite element analysis tells us.
24 After the magchloride tests, we got it and the O
25 residual test measurements.
I
422 1 We got strong compressive residual O
V stresses all over the inner-surface of the pipe, 2
3 which is consistent with the magchloride results.
4 They d id n' t find any cracking. After we sectioned 5 the pipe, we did some more metallograph and dye 6
penetrant checks, and we had short shallow cracks in 7 the region of the counter bore and a longer deeper 8 crack in the seam weld directly under the bore.
9 This caused a great deal of consternation 10 at the time. This crack was about four inches long 11 ;
and a quarter inch deep, and it's still puzzeling to 12 us why it wasn't seen in J.A. Jones.
() 13 f i
The metallograph of the cracks as we look 14 ! at them is characteristic of magchloride cracking.
I 15 Our judgment is that the magchloride --
or the MSIP 16 didn't cause the cracks. It was -- essentially wha t 17 we have done is we had a local s' tress riset that 18 produced the tensile region that MSIP d id n' t relieve 19 and that tensile region cracked with the magchloride .
20 I Wha t 's puzzeling to us and what wo can't 21 i exactly explain is why if it was a tensile region 22 here, why the J.A. Jones didn't see it and they 23 didn't dye penetrin check after the magchloride, and 24 } the only thing we can argue is there is a very :
25 localized tensile region and that essentially i i I
423 1 everything else is compression and closed up the O 2 crack and simply didn't allow the die penetrin to 3 get in and really couldn't see it until we relieved 4 all those inside surface stresses by sectioning the ,
5 pipe.
6 Now, let me skip over the results for the 7 , residual stress measurements. I would point out, i
8 what you need to look at if you want to look at them 1
9 yourself, the inside surface stress measurements are l 10 given one set of numbers. The through-wall stress 11 measurements are always described in terms of the 12 gauge numbers given or the outside surface.
i
() 13 I So, if you find something that doesn't 14 seam quite consistent between gauge 5 and gauge 8, l i
i 15 it's --
cr a gauge 5 and a g a ug e 5 are not l l
l 16 consistent is because we are reporting inner-surface i
! l 17 measurements. I've got this set of numbers and ,
l 18 reporting through-wall measurements and I've got 19 this set of numbers.
20 The important stresses to look at are 21 gauges 1 and 2, 8 and 9 which are in the heat 22 affected zone basically. If you look at the data in 23 the table, you will see they are very compressive on 24 the inner-surface. Contrary to the predictions of 25 the finite element code, even if gauges 4, 5 and 6,
424 1 i which are right under the tool, we get compressive
. I 2 stresses. They are not nearly as compressive as 4 l
3 these stresses are, but they are not tensile either, l
4 So, it's even more favorable than finite element 5 code.
6 Somewhat surprisingly, it's the gauges on 7 the far side of the weld that show the most 8 compressive stress. But I'll let you look over that 9 data for the 12 inch pipe.
10 MR. S iiEW M O N : Were the welds twice as 11 strong as the base metal like they were yesterday 12 with the TIG welds that lia y s was talking about in
()
13 ! his ovalization? Part of the rationale that he gave 14 l was the strength is very substantially d if f erent.
15 , MR. S il AC K : It's probably true, a l tho ug h 16 what we have here is almost a continued gradation of 17 properties in the weldment affected zone to the base 18 metal. I think it's true. It's like a 2 to 1 ratio 1 19 between the base metal and the weld, but 308 is 20 l almost not inherently stronger than 304.
21 The real difference is that everybody 22 always compares essentially unworked hardened 304 t o' 23 the 308 weld metal that they cut of the weld which 24 l has been cycled umpty-dump times. The gradation 25 would not be dramatic. The heat affected zone is
_J 4
425 4 1 almost as strong as the weld metal.
O 2 I'll skip over the rest to get to the 28 3 inch diameter weldment. I passed the pictures of 4 this weldment around to mention that you can see the 5 maximum plastic deformation permitted under the MSIP 6 specification. We did dye penetrim examples before 7 and after sectioning this thing. We got no cracks.
8 This one wasn't exposed to the mag chl o r id e .
9 The metallurgic examination confirmed that 10 the only observable dye penettim example indications 11 that we did see at all and d id n ' t look to us like 12 l cracks were in fact weren't cracks. They just
() 13 smeared some weldments over it and was drown out and 14 left a little groove undernerth the weldment. The 15 strain gauge results indicate strong compressive 16 stresses on the inner-surface.
17 Unlike the outer residual stress 18 techniques, IHSI, this one appears effective on the 19 large d iame te r pipes than on the smaller piping.
20 IHSI works like a charm on the 4 and 12 inch pipes.
21 Doesn't work quite as well on the bigger diameter 22 pipes. But this one with the basic root force 23 squeeze seems to do as good a job, and you can just 24 sort of see some of the data.
)
25 Again, highly compressive stresses at the
426 1 three that we have look at right in the vicinity of CNJ 2 the weld. As you go out, we in fact do see sl ig htly 3 tensile stresses in this case under the tool. Again.
4 they are really very low. So what you have done is 5 traded for untreated weld very high tensile stresses 6 in the heat affected zone for much more favorable 7 ! stress state in the heat affected zone and slightly 8 higher tensile stresses out there away from the heaq 9 affected zone.
10 The next graph shows you the hoop stress, i
11 -
and I'll skip over that. I've mentioned some of our 12 future work just in connection with the 316 nuclear
() 13 grade material. I've described that work which is 14 something very closely directed toward the kinds of 15 1 things that we have been doing, but we also have I.
16 ! some work that goes off in new directions that we 17 haven't really been doing before, and I want to 18 describe some of that.
19 i This work on the ferritic steel is just i
20 l getting under way. It's again answering someone l
1 21 else's question, and it's done in response to a 22 request from regulatory. The susceptibility to 23 stress cracking of these ferritic steels in high 24 purity oxygenated is well established, and we will 2% tell you a lot about that in the next talk.
427 1 We also know that impurities can play a O 2 critical role even in non-oxygenated environments.
3 As I pointed out before, one thing that impurities 4 can do is provide an alternative lethargic reaction.
5 If you don't have oxygen but you have sulfate or 6 something else, it can also be complied with a 7 little lethargic reaction when you drive the stress 8 cracks.
9 People seem to have found copper cracking 10 in the end union point steam generator. So it 11 wouldn't be unexpected to find that the impurities 12 play an important role in this cracking. But as
() 13 investigating a system a particular way, and we will 14 be going through that looking for what kind of 15 impurities could play an important role in that case .
16 The other area I talked about before was 17 the radiation assistance stress cracking. As I've 18 noted before, the plotting clearly exists that the 19 phenomena can happen. You have current failures 20 already on components like dry well tubes.
21 Now, a failure in dry well tube is not 22 particularly a serious problem. If you have a 23 failure in a top guide, then you have a serious 24 problem. We know that the radiation has an effect 25 both on the material and the environment. As I
= % - yr -
m' M --
p ,e,_ wey, , -- - ~-- -m e-- er-me s =w-e ~ n-r- -
9*-t1--wF
l 428 ,
l 1 mentioned before, you get this segregation that 2 might be the material effect. So it's not a classic l 3 resensitization problem, but it's a grain opened 4 boundary segregation problem.
5 It can also have essentially an important 6 role on the environment. We have lethargic 7 reactions, we are producing oxygen. We also produce 1
8l other specimens that were short-lived that play a 9! lethargic reaction like h yd r og e n chloroxide. The
! 10 radiation can also in addition to giving you I
11 alternate reactions can affect the rate of these l
12 reactions.
() 13 MR. SHEWMON: Ten to the 19th in the 14 l pressure vessel is relatively high and the few t ime s 15 l it has ended its life. And this stuff is set in a 16 i few times 10 to the 20th.
17 3 MR. SHACK: But I'm not sure. I always 18 see the 10 to the 19th for the pressure vessel with 19 that first half inch of the vessel getting --
it's 20 ,
underneath -- it's not underneath it by a whole lot.
21 l That estimate of five times 10 to the 20 comes 22 mostly out of the air.
23 MR. Sli EUM O N : To bring up a separate thing
, 24 on that, seems to me there were some bolts changed 25 out of PRW's perhaps more than here. Those probably; I
429 I were ferritic. Can you help me? I'm hazy at this 2 point. I don't think it was all austenitic was my j 3 point.
I 4 MR. Sil AC K : No, but those were all high i
5 strength bolts. I think that the problems with high
! 6 strength bolting materials, both ferritic and 7 austenitic, the A286 bolts are terrible. You know, 8 they are an austenitic bolting. The 718 high 9 strength materials have been a problem, and there 10 are some ferritic bolting. The ferritic bolting I 11 think.is mostly outside the environment. What you
! 12 get through there is leakage of t.ie environment that
() 13 gets into those high strength ferritic bolts.
14 I don't know of any high strength ferritic 15 , bolts inside the reactor. As far as I know that's i
16 austenitic materials, but I'm hardly an expert on 17 bolting.
18 Again, the work on the assisted cracking 19 is just getting under way. We are in the business 20 of trying to get some eradiated material to work 21 with. We are currently trying to characterize the 22 kind of environments me might get in terms of 23 radiation, and what we set up here is essentially an 24 autoclave and a hot cell with a cobalt 60 gamma
)
25 source, and we essentially measure the potentials
! 1
430 1 and the chemical as best we can in the autoclave in O 2 the intense environmen'.s, and we compare that with I
3 the measurement.
4 We make to the autoclave outside the hot 5 cell away from the gamma-field. Again the thought l 6 is here we are actually seeing, except for the 7 . effect of the radiation products, the same water 8 chemical in both these autoclaves in preparing the 9 chemical changes, the crack potentials. l l 1 10 l I guess I didn't leave a whole lot of the i
11 two and a half hours unused. There is no summary 12 l
because I gave you all the important information at 1
( 13 the beginning.
14 MR. SilEW M O N : Any questions? Why don't we l 15 go on in view of the hour and people leaving and we 16 will go back and get our coffees while you are 17 talking. Thank you, Bill.
t 18 ! I i 4 I !
19 : J l
20 l l
I 21 ,
l 22 23
() 24 25 !
I i i
- --- __ - . . _ __ . _ . __ _~ _ . _ , _ _ . _ _ _ _ _ - . . _ _ _ _ _ _
431 1
0 2 MR. CULLEN: I would like to spend some 3 time talking to you today about part of the program 4 that is NRC sponsored at Materials Engineering 5 Associates. We are at about the two and a half year 1
6 point in a four year program that is basically i 7 divided into three parts. There is the fracture 8 part, there is an eradiation sensitivity part, and ;
9 there is this fatigue part, and I will try to revieu 10 most of the aspect of the part during the next hour 11 or a little less.
l 12 The emphasis is decidely on applications,
() 13 and I would spend a few minutes now upfront showing 14 you the genesis of how we got to where we are over 15 the last 20 years or so that this kind of work has 16 basically be going on.
17 The whole business of being able to do 18 these kinds of tests in autoclave, in pressurized 19 high temperature water, began roughly in about 1965, 20 and there was a review that went on in the NRC, then 21 the ADC, and in '69 the first of the NRC contracts 22 to do this kind of work was given to Westinghouse.
23 Some of the information on the left-hand 24 side, which I'm not going to .ead off to you, can b 25 found in this historical retrieve that Grady WhitmaM
! 432 1 put out about two or three months ago. Really a O 2 nice piece of work on Grady's part. Over here on
.1 3 the r ig ht-ha nd side I describe some of the effects, 4 some of the conclusions that have fallen out of this 1
5 work as we have progressed through the first ten or 4
6 so years of just trying to figure out how to do 7 these at that time quite difficult tests that are 8 nowadays routine.
i 9! About ten years of just test methodology l 10 development followed by about ten years of 11 generating data. Just gobs and gobs of data.
1 12 During that ten year period we learned something
() 13 about load ratio effects that in fact fatique crack 14 growth rates do increase as one increases the load 15 i I ratio. About '79 PIB published a NUREG that 16 described some of the waive form effects that l
17 : linealized wave forms, triangular waves. Ramson I'
18 holes did not in PWR environments give anything of 19 the same kind of crack growth rates that autowava 20 once gave.
i 21 Later on in about 1980, Westinghouse with
- 22 some confirmation by ourselves, came up with the 23 sulfur chemical effect, that increasing the sulfur 24 content of pressure vessel steels yielded increases 1
25 in fatique crack growth rates again in PWR
433 1 environments. In '81 we began piping steel tests, O 2 and I will hone in, of course, for obvious reasons, i
3 on that data during the course of this presentation.
4 In 1982 we began to find some temperature 5 effects, effects of test temperature on fatique . .
I 6 crack growth rate results. In 1983 we were awarded 7 this applications program, which has some mile-8 stones that are shown up here. We began some stress 9 life tests which is one of the very first o b j ec t i v e s; .
~
i 10 Part-through crack tests began later on.
11 During this past year some viable amplitude tests l l i 12 began, and other tests will be initiated and
() 13 completed as we continue on in this project. I 14 The emphasis is on applications. Yes, we 15 are testing materials. We are using piping 16 materials. We are using some pressure vessel 17 materials, but we are not really over the course of 18 this project trying to define materials effects.
19 Much of the work, as you have seen over the last 20 couple of days, has been oriented in that direction, 21 '
but we are in applications.
22 The general trend of things for all the 23 laboratories in this area over the last few years in 24 shown schematically here. Tests of largely compact
)
25 fracture specimens are performed in autoclaves.
u
I 434 1 From those tests the data is plotted vise-a-vise the O 2 current ASME code format in Appendix A of Section 11 ,
3 and we are trying to show one of two things here.
4 Can wt find any steels which will provide us crack 5 growth rates that exceed the ASME boundaries, or 6 conversely, if we find a lot of steels that do 7 provide data that exceed those boundaries, then i
8 indeed the boundaries should be changed, and tnat in i
9{ fact was done in 1980. We then had enough data to l
10 show that the totally linear format of the reference 11 lines found in Appendix A should be changed, and we 12 have now what is generally regarded as this bilinear l
1
() 13 format.
14 But over the years we have used compact 15 specimens to generate data which is compared to the 16 ASME code. Now we are using other kinds of 1
1 17 specimens.' We are performing other kinds of tests, I I l 18 ! but again, the same general procedure. We will get I
19 l some data which we will compare against the existing' 20 l codes to try to find out whether or not the codes 1
21 are good enough as is, whether some change in the 22 way we process our data or use our data or apply our 23 data is called for.
24 MR. SHEWMON: Would you tell me where this i
25 is germane, because most of the primaries are l l
l
435
\
l stainless lines, isn't it? Does this come out of O 2 steam lines or where? Where could it be applied in 3 a PWR?
4 MR. CULLEN: Where'could this type of d a t a-5 be applied? I think the driving force behind this 6 is we are looking at load ranges which are typical 7 of the start-up modes, the sort of trip transients i 8 that one incurs as well as the normal day-to-day 9 power loadings and unloadings. That's the driving 10 force that says we should be performing these kind 11 of fatique crack growth rate tests. The steels up 12 until '81 were entirely pressure vessel steels, at
() 13 least in this program.
! 14 MR. SHEWMON: My point is the 15 environmental assisted part, and where does the i
16 environment see a non-stainless --
where does this 17 environment see a ferritic surface?
l 18 MR. CULLEN: Anything that penetrates the 19 pipe. The clad has been removed from consideration j 20 for our point of view. The assumption has been a l
I 21 quarter teeth through clad crack. What would happeq 22 in terms of fatique crack growth to that kind of 23 crack. That was the question that wac being asked.
24 That was the question that these sorts of projects 25 addressed.
f 436 l
1 MR. SHEWMON: Kind of a hypothetical, one, O 2 since no one has seen one of those in a real plant, 1
3 is that correct? )
4 MR. CULLEN: Exactly correct. Which is 5 why the emphasis since 1981 has been decidely away 6 from the pressure vessel steels and into the piping 7 steels.
8 MR. SHEWMON: Well, if I ask the same 9 questions on piping, where is stainless or where is 10 water against the ferritic material in the piping?
11 i MR. CULLEN: I think I w uld have to take 12 a look at these guys who are more into the design
() 13 than I am, but it's again --
14 MR. MAYFIELD: Everett may want to chime 15 ! in here, but as I understand it, there are some i
16 regions in the B & W and CE primary piping near the 17 well as an example where some of the ferritic 18 ; material may be exposed. There is also a region 19 ! where you may see a breach in that liner.
20 ! MR. SHEWMON: Now, GE put ferritic pipe in.
21 part of it, and I'm not sure where, but part of 22 their r e pl acem en t to get away from the grief they 23 were having with the stainless. ' '
i 24 l MR. SHACK: Now, the BWR people went in l
25 l and scraped the clad off of certain regions. They i
I __
. _ _ - - _ _ _ - - - - - -1
437 1 were getting thermal fatique cracks because they had O 2 the poor thermal expansion properties of the 3 stainless. They went in and cleaned this off.
4 There were actual regions in a BWR pressure vessel 5 where ferritic material does affect CE environment.
6 MR. SHEWMON: Okay.
7 MR. CULLEN: This is all PWR work. That's 8 the only environment that we run at MEA.
9 MR. RODABAUGH: Now, at 2000 psi 10 performance, you got the steam lines to worry about.
11 You have all the secondary systems. \
12 MR. SHEWMON: Well, you assert that this
() 13 is PWR and you use a hydrogen over pressure?
14 MR. CULLEN: The water is hydrogen 15 saturated, correct. Saturated with hydrogen.
16 Technically it's about 35 to 50 cc's. It's all 17 hydrogen. So, this is dissolved hydrogen, but there 18 is little or no dissolved o x yg e n . Less than 10 ppb 19 generally speaking.
20 Just to give the right perspective now, we 21 are not looking at sizemic loads. He heard a lot 22 about sizemic loads earlier today and yesterday.
23 These are normal operating transients that we are 24 trying to model in this program. The program is
)
25 divided into several subcategories, many of which b
I 438 1 you will hear about in the next few minutes. We are 2 phasing down the conventional Da/Dn Delta K test in 1
3 order to replace them with more application oriented 4 efforts.
4 5 Specifically there are some work on three 6 dimensional shapes, part-through cracks, there is 7 work on the affect of clad on the shape of these 8 pa r t-t h ro ug h cracks, but again these are through-9 clad cracks that we are worrying about. These 10 l efforts, we are using pressure vessel steels as the 11 test material, but I think the work is generic. It 12 l could just as well be applied to piping steels.
i
() 13 We are looking for shape effects rather 14 than specific material dependent crack growth rate 15 effects. There is some work done on mechanisms.
I 16 This is somewhat theoretical, this is somewhat 17 number crunching, but we are looking for the 18 l mechanistic reasons which drive environmentally l 19 assisted fatique crack growth.
I 20 '
So, some of the things that Bill alluded 21 , to this morning, we are worrying about ionnetic 22 specimens, ionnetic transport in and out of.the 23 crack, the effect of water chemical, the effect of i I j 24 hydraulysis.
l
! 25 MR. Gi!EWM O N : What effects of the water o
l 4
E _ _ _ _ _ . _ . _ _ . _ _ _ _ _ _ _ _ . _ _ _ _ . _ _ _ . _ _ _ . _ _ _ __. _ _._.___ _ _ _ . _ _ _ _ . . . _ _ _. - _ _ _ . . _ _ _ _ . _ _ . _ _ _ _ _ _ . . _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
439 1 chemical? Did you ever go to something tha t's J 2 l
appropriate for a steam line without lithium or 3 boron?
4 MR. CULLEN: No, we are not --
5 MR. SHEWMON: Did you ever do anything 6 without hyd rog en? I 7 MR. CULLEN: No, we are only asked to do 8 the idealized PWR environment. By effects of water 9 chemical, I'm worrying about the hydraulysis that 10 could happen under certain temperature conditions, 11 those kinds of things. Those are the only variables 12 that we are charged to worry about.
() 13 MR. SHEWMON: Mike, why is this tradition 14 so strong that this is limited only to PWR I
15 environment?
16 MR. MAYFIELD: The answer is I really 17 can't tell you other than Bill Shack has been with 18 the BWR stuff the way that Bill has structured the 19 MEA program, has been in traditionally and continues 1
1 20 to look at PWR system. l 21 MR. SHEWMON: And he will start looking at 22 ferritics? See, if there is a need for this sort of l 23 work? It seems to me there ought to be some 24 symmetry or more symmetry than I see here. I'm 25 curious as to why not.
r-440 1 MR. MAYFIELD: The point is well taken and O 2 I can't answer.
3 MR. C U L L E tJ : We have specifically asked 4 from time to time should we be worrying about water 5 chemical and purities and variables, and the answer 6 has been stick to the one chemical and tell us about.
7 materials or tell us about applications, but don't 8 try changing too many things all at once.
9 We will be worrying about varicale i
10 l amplitude which is often called cumulative damage, I
11 but up until now, we have basically used constant 12 amplitude fatique crack growth rate tests, and now
() 13 we will try to analyze and understand what happens 14 !i under variable amplitude loadings.
- i 15 There is some work on initiation of cracks!
16 from notches of various kinds. There is work on 17 stress life, and whether or not the ASME Section 3 18 code could be modified to account for environmental I
19 } effects on stress life curves. The whole idea is to i
20 ) provide some sort of predictive capability for 21 growtn or the Section 11 treatment of fatique cracks 22 or to provide some predictive capability for 23 initiation to the section 3 type of design 24 concideration. Merging those together, one could
)
25 try to generate some predictive capability for total
.l
441 1 life of the structure.
O 2 l
MR. E T ii E R I N G T O N : You have 3-D dimensional 3 shapes but you are staying with 1-D dimensional?
4 MR. CULLEN: Yes, that is correct. I'll 5 try to make that very clear in a view graph two or 6 three foils away from here. There was a question 7 this morning. Now, this particular slide is the 8 very last one in your handout, but because of a I 9 question that was raised this morning, I moved it up 10 here, and I'll give it again at the very end, but it 11 was asked of another presentor what is the coc5 12 relevance of your work.
() 13 I want to stress that as clearly as I can
\
, 14 that our code relevance is that we are doing work on 15 stress life curves using BWR environments and that 16 relates directly to the ASME Section 3.
l l 17 MR. S II EWM O N : The ASME that you are 18 getting to --
maybe if I shut up I would learn --
19 but it's basically mid-range for cycles, is that ,
20 right, sir? Ten to the 3, 10 to the 5, 10 to the 21 6th? It's not a high cycle?
22 MR. CULLEU: You're right in the range, 23 yes. Ten cube to 10 to the 7th or so is the range, 24 the life range, that we are working in.
25 MR. S !! EWM O N : So, you are getting an order
%- - _ _ . . -,. _. ~_ . _ _ _ . . , , , ..
442 1 of magnitude to approach the old 10 to the 6th. It O 2 used to be somewhat of a limit in the code, as I 3 recall. Couple of years ago there was some Japanese 4 man was saying why didn't we do any work above 10 to 5 the 6 because that's where the failures were and 6 that the code at that time, or whatever he looked at 7 for the code, quit at 10 to the 6th, and it grieved 8 me. Okay.
9 MR. CULLEN: The fatique crack growth work 10 in compact 10 specimens and part-through crack I
11 , specimens, as well as the work on variable am pl i t ud e ,
l I 12 loadings will relate to ASMU Section 11, l
() 13 specifically Appendix A. The micro-mechanisms work 14 in the models for total life that we are trying to l
15 generate would apply to both of those. We are 16 trying to keep in mind the relevance of this 1 I
17 research to the current issue of plant extensions 18 i believing that our choice of using PWR environments I f 19 ! and very low cycle long term fatique crack growth i
20 rate experiments relates rather directly to this 21 ! question of plant extensions and plant aging.
22 The life prediction method d e v e l o pm e n t 23 la work for both the fatique issue and fatique crack 24 l growth issue as well as the experiments to 25 development m o rl e l s for total life again relate to i
i _ _-
t
_ _ _ _ ._.--,-..c. . _
,. - , , , . ._ _ , . . _ . . . . . - . _ , . , _ _ , _ . ~.,, -_.s-- ,
443 1 the plant extensions question. The mechanisms, if O 2 we can come to some understanding of the mechanisms 3 for environmentally assisted growth, that provides 4 some predictive capability, saving us from doing 5 experiments that would last four years, and again 6 allowing some ability, hopefully, to execute 7 predicted life, and the test matrixes that we are 8 performing, carrying out at it E A , use reactor typical 9 materials, flaw geometries, and there is in the 10 program a structural element --
a set of structural 11 element tests, basically flaw pipe test under 12 elastic fatique loadings, which I will describe in a
() 13 few minutes, that for about the first time will 14 provide some indication of how part-through crack 15 grows under fatique in a pipe.
16 The general outline as I would like to 17 deal with part-through crack work first, variable la l amplitude loading studies second, stress life 19 testing of piping steels, and then fatique crack 20 growth rate work in piping stocls. So, the bias is 21 to put the piping steel work first. I will only 22 spend a few minutes on the first two items and the 22 bulk of the presentation on the last two.
24 In this task to examine the growth 25 characteristics of 3-dimensional cracks, the 1
1 1
! 444 l
l 1 1 approach is to conduct a series of tests in PWR 2 environments to correlate both 2-dimensional cracks, 3 those are cracks in the ordinary compact tension
)
4 specimen, with 3-d imen s io n a l cracks in a 5 part-through cracked panel.
6 To develop both the experimental and 7 analytical techniques to monitor the 3-dimensional i
8 crack extensions, we are using the direct current l
9 potential drop method that you have seen described I
10 in our presentations in the last two days in order 11 to monitor those cracks. ;
12 The crown jewel, if you will, of this l
() 13 particular sub-task is to construct what we are calling a structural element fatique test facility )
14 .!
15 f or a cyclic pressurization test facility to grow t
16 flaws in up to 12 inch diameter pipe sections in 17 both warm air or PWR environments. PWR environments 18 being the 550 degree f a h r e n 'h e i t , 2000 psi claddings.
19 ,
That facility has been designed. We are 20 putting together a bid package for it. As you will 21 l see in a couple slides from now, it's a rather 22 simple design concept, and we think we can get it in 23 place before the end of this calendar year.
24 Why is it that we are looking at the
)
25 comparison between 2-dimensional cracks and
i 445 1 3-dimensional cracks? Bearing in mind that the ASME O 2 code has been developed entirely for this type of i
3 specimen, virtually e n t i r e l ycp there are a couple of 4 reasons. One is that in this crack the environment 5 has access from three sides to get at the crack, 6 which is really different from the more structurally 7 similar crack where the environment only has access 8 from one side.
9 l
Another consideration is that this type of 10 specimen has both a mode one crack opening, but 11 there is also a bending moment at the crack tip 12 which one does not get in the more structurally
() 13 similar part-through crack, which is almost always 14 in pure tension. At least in the reactor wall. l 15 jl i
So, we are testing in this program really 16 three kinds of specimens. We have always tested thc 17 compact specimen on the left. He are now beginning I 18 to compare that work to the tests of part-through 19 crack specimens shown in the middle, and Mr.
20 Et he r i ng ton has asked a couple of times what is the 21 uni-axial loading on this type of specimen. It's a 22 purity axial load.
23 When we test these specimens, we put them 24 in very rigid grips to assure that the load will be 25 uni-axial during the whole course of the test. So,
446 1 as this crack tends to open more and more, we do not O 2 allow that specimen to bend. We just keep it in 3 pure tension. So, to better simulate an even larger 4 test geometry than what we are able to get, which is 5 about a one inch by four inch test specimen or one c
6 and a quarter by four in some places, the idea is to l I
7 provide a way of predicting what would happen to a 8 part-through crack flaw in a pipe or in something 9 that is really structurally similar, and there are 10 tests of this sort of thing also in the program.
11 MR. SHEWMON: How do you make these flaws 12 when you get to that?
() 13 l
MR. CULLEN: We begin with an electrically 14 < discharged machine notch which is then fatique l
15 l cracked additionally, then put into the environment l
16 l where the test begins. That's true in all three 17 cases. So, we are testing some of these compact l
18 l, specimens, but clearly we have been asked and are i
19 l decreasing the effort in that area, where we are 20 l bringing up to speed the efforts to testing PTC
, 21 panels and PTC pipe components.
23 MR. CULLEN: Part-through cut. I have 24 some view graphs showing crack growth factor
)
25 surfaces of some of these test results. This is a
447 1 pressure vessel steel, medium sulfur content. The O 2 sulfur of course doesn't have much effect on growth 3 rates in error, and these three slides are of error 4 results.
5 We are on the verge of beginning companion 6l tests in a PW R environment, but you can see that the 7 shape of the crack is roughly semi-circular, two to 8 one ratio, but when you test at high temperature, 9 there is some canoeing of the cracked front. l 10 Presumable the stress state of the surface here does 11 not allow really rapid growth.
12 MR. SHEWMON: The results are of mild l ) 13 asymmetry. Do you know what causes that?
14 MR. CULLEN: It is probably more related 15 to the way the f a tique pre-crack was able to 16 nucl ea te from the electric discharge notch. If you 17 take a look at this asymmetry, it is more pronounced i
18 right at the very beginning than it is at the end.
19 So, after this crack gets moving, it tends to 20 symmetrize itself --
that's not such a word -- but I 21 you understand what I'm trying to say.
22 MR. SHEUMON: Okay.
- 23 MR. CULLEN
- This is in unclad pressure i
24 vessel steel. Ilere is the companion test in clad I
O-25 pressure vessel steel showing that the effect of the
+- _
-r --
r i
448 1 cladding is reasonably significant. It does tend
/"%
sJ 2 somewhat to hold that fatique crack in and keep it 3 from opening along the outer clad surface so that 4 canoe shape effect does get to be a little more 5! pronounced.
6 What will be interesting is when we do geu 7 ,
this into the PWR environment, I have deliberately i
8I chosen a pressure vessel steel which is somewhat i
9l sensitive to environmentally assisted crack growth.
10 l And we know that ~ ---
and you will see later on --
L l
l 11 l that the cast stainless steels are generally not I
12 l very sensitive to fatigued crack growth in PWR
() 13 l
environments. So, we expect this canoe shape to get 14 l oven more pronounced when we begin these tests in a I
l l 15 PWR environment.
16 Now, that's a shape of a crack -- if 1 17 turned it around you could see it better -- at 550 18 degrees fahrenheit. Shape of a crack at 550 degrees j l
19 I :
fahrenheit. If you do the same test at room 20 ! temperature, that cladding which is under rather 21 strong tensile force really causes that leading edge 22 to unzip. So, now the canoe shape is completely-23 lost, and we are in a semi-eliptical shape. Let me rm 24 just flip back and forth so you get a good idea of V(
25 the difference there. That's at 550 fahrenheit.
_ _ ~ _ _ . _ -
.~ .- _
449 1 That's what happens at room temperature. Everything
() 2 else is the same. The same cyclic frequency, same 3 test load. So the comparison can be made one to one .
4 MR. SHEWMON: Almost looks like there is a 5 different structure on the base bottom of that and 6 the top. Is that just the way you prepared it or 7 what?
8 MR. CULLEN: Are you talking about what's
~
9 going on right here?
10 MR. SHEWMON: Yes.
11 MR. COLLEN: That's the clad. This is the 12 base plate.
O 13 na. suEwMoN: eine.
14 MR. CULLEN: You are absolutely right.
15 MR. RODABAUGH: Bill, do you have a 16 companion to yours without clad, hot and cold?
17 MR. CULLEN: Yes.
18 MR. RODABAUGH: Yes is all I'm interested 19 in.
20 MR. CULLEN: Let me show you the view 21 foils afterwards. This spread out is not as severe 22 at all. Okay, we are using direct current potential 23 drop to monitor these things. This is just some 24 indication of the goodness of fits of the electric 25 potential drop data to the visually monitored in 1
_, w -
m - - - - - , - - . . - - - - - - - , +
_ _ _ , ,-. ap - y - < - , , - , . - - -
i 450 f 1 this case from these kinds of specimens so that we O 2 can --
we are developing and are now able to use the 3 direct current potential drop data without getting 4 visual data which of course is a necessity for when 5 we go to the autoclave test. That's the whole point 6 of this experiment here.
t 7h When you differentiate that data and plot 8 crack growth rates versus the supplied cyclic stres 9 intensity factor, these are the results of the room 10 ,
temperature tests, but the data resides I
11 i approximately on the ASME Section 11 air reference l 12 j line as one would expect they would, showing at
() 13 l I
least the part-through crack geometry of and by 14 itself doesn't have any particular effect.
15 , These results agree with compact specimenI l \
16 I data under the same sort of test conditions. We 1 17 will now go on and perform these tests in a PWR 1
18 environment and see what happened at that case.
19 The structural element test facility will i
20 l look something like this. We are going to try to 21 assemble it from readily available components and 22 caps, pipe sections and pipe flanges go into that 23 the test device on the outside. The test vessel i
24 itself into which we will machine and pre-crack some 25 l flaws in this article right in here.
451 i
l l We will put PWR environment on the inside O 2 and pressurize that environment. On the outside we 3 will have single phase fluid which will be cyclicly 1
4 pressurized in order to apply the fatique loadings 5 to these flaws. So the load on the flaws really 6 comes from the PUR liquid that is on the inside and 7 to which the flaws are exposed, but it's by c h a ng i n g-8 the pressure on the outside that those loadings are j 9 realized. j 10 Features of this facility are that we will 1
11 be able to test sections of pipe up to 12 inches in l 12 diameter, about one inch, thick using PWR
() 13 environments and up to 650 fahrenheit.
14 Pressurization, as I said, is by the single pbnao 15 fluid, which is external to the test vessel with the 16 /PWR environment on the inside. We will have two or 17 more flaws per test vessel, beginning with two. If 18 we find that we can handle experimentally more than 19 that, we will try using DC potential drop crack 20 extensions monitoring.
21 tie x t aspect of this program that I would 22 like to describe is the work on --
I'm sorry, one 23 more slide here. The payoff for this particular 24 effort will be that we hope to be able to
)
25 characterize real flaws of part-through crack flaws L
l 452 1 in real structures, which for us at this point will 1
() 2 be up to a 12 inch diameter pipe in both air and 3 /PWR environments in order to help support us in 4 licensing decisions by NRR.
5 We will also try to evaluate the 6 ,
significance of clad induced residual stresses to 7 vessel and piping integrity via our work on clad 8 part-through specimens.
9 The second part of this is to very 10 beliefly describe our effort on evaluating and 11 understanding the mico-mechanisms of environmentally 12 assisted crack growth. The approach here is to
() 13 l review, select and evaluate some mechanistic process:
14 l and calculational model to go along with it, then to 15 i conduct some selected experiments to try to verify l
16 the accuracy or to improve the development of that 17 particular mechanistic process or.model, and we have 18 l invited some visiting scientists to MEA to help 19 l support us in this crea, specifically Giovanna l 20 Gabetta from CISE, Italy, who really knows and 21 understands a lot about straining rig models and i 22 crack tip straining models, and Hanna llanninen from J
J 23 Finland who is one of the experts in various i 24 mechanistic processes of the environmentally 25 assisted cracking. '
.i
- - - , . ,.-- ,, . , ,- .n., --,.-~_,-..,n, . , , - - , . - - . - - - , - , . - . . - . - . . . . , , , - . . - . - - . - , . - - . . - ~ , - , - , _ , . - -
l
. 453 4
1 Which one thing about environmentally l
)
2 assisted cracking, there are a number of processes 3 that you need to consider. There are the flushes of
- 4 the water and the metal ions in and out of the crack. )
5 One needs to worry about the electric-chemical 6 potential of both bulk environment and of the 7 environment in side' the crack tipped enclave.
- 8 One needs to worry about the strain rates 9 which are induced at the crack tip both by the j
10 cyclic load and by the fact that the crack itself is 11 advancing. That provides deformation and additional 12 strain at the crack tip.
i
() 13 Bill gave a much more extensive 14 description of various electric-chemical actions 15 which can occur at the crack tip. Giovanna is 16 currently working at MEA right now continuing
- 17 development of her strain rig model, and I was i
i 18 pleased that Bill put up a work in progress kind of J
- 19 view graph because that's exactly what this thing is , i I
20 and I'm sure many of you are saying that is the i 21 biggest God awful mens I ever saw. What does it 4
22 mean?
23 Giovanna is taking all of our MEA data on 24 fatique crack growth rates in PWR environments and 25 executing them according to the equasions of this t
r_
454 l
l 1 model which she has developed. The basic idea is to
() 2 take an environmental component of the fatique crack 3 growth rates which one obtains by subtracting away 4 the fatique crack growth rates you would obtain in l 5 ordinary air.
6 So, you just leave yourself with an l 1
7 environmental component and then plot that back 1
8 remaining component versus the cracked tip strain l l
! 9 rate using the equasions which Giovanna has 10 developed, and what one is looking for and what we l 11 ! are getting here, at least in some of this data, is l
12 l some sort of an upper limit to the velocity.
l l ( 13 !
If one can get an upper limit which is l
14 l independent of the single cycle strain rate, then
- )
15 you are able to invoke a crack component, affix l l
> l l 16 velocity crack component which can be added to the l 17 inert environment fatique crack growth rate, 18 allowing you to execute for any cyclic frequency the-19 ! environmentally assisted crack growth rate.
20 l Well, there are some tests results which 21 do not conform to the model. They are way down here 22 in environmental component. That basically means 23 that was the material that did not show any 24 appreciable environmentally assisted fatique crack 25 .
growth rate component, but we are getting a goodly i ,
1 I,
i i
455
}
. 1 number of tests which do seem to reside on some sort
() 2 of upper limit of the environmentally assisted
- 3 component when applied as a function of cracked tip 4 strain rate.
1
! 5 He will obviously be continuing to work on i 6 this and develop it more fully. The payoff that we
] 7 hope to get from this kind of thing --
8 MR. SHEWMON: Is the model more
+
9 dissolution or embrittlement?
I l
10 MR. CULLEN: That is something that right i
j 11 now almost comes down to a mechanical set of
- 12 arguments rather thar. a technical set of arguments.
}
1
() 13 There are two schools of thought. You will find l
j 14 people firmly camped out in both of them. There are 15 the hydrogen assisted crack growth group of people, 4
1 16 and I put myself over in that category. There are
- l 17 the active path dissolution group of peoplc, and a '
i 18 goodly number of impressive d ig n i ta r ie s reside in j 19 that camp as well. The dialog between the two is i 20 good.
21 MR. S ii E U M O N : Okay. I thought you were 22 going to say nonexistent.
}
l 23 MR. CULLCH: No, it is active, it is 24 enthusiastic. Both sides have really good 25 contentions to make. We don't have conclusions.
j __
i t 1
i
456 1 MR. SHEWMON: Okay, fine.
O 2 MR. CULLEN: I think what it comes down t o' 3 and over coffee people will say a little of both is 4 going on. Something has to pre-dominate. It may '
5 come down that this material bulk system in this one 1
6 crack propagates mostly by hydrogen assistance and a
7 that one is mostly by active path. Whether one can 8 generalize to one or the other, I don't know.
< 9 MR. SHEWMON: Is the environmental l 10 contribution more pronounced, which is 78 or the 550 11 F7 12 MR. COLLEN: .I know the answer to that l () 13 question. It is again dependent very much on the l 14 material environment system. I have some 1
15 temperature effects data towards the end of this 16 presentation that would really clearly address that.
17 Rather than jump there, let's wait a few minutes.
. 18 MR. SHEUMON: Fine. Sorry. Go ahead.
19 MR. CULLEN: If we have to come down to it ,
20 low temperature generally always provides a rather l 21 significant environmental effect. High temperature 22 is not always thero, but temperature is a very 23 important variabic.
24
{) We hope out of this mechanism work to 25 obtain a payoff where we will develop some ability l
i 457
, 1 to extrapolate this laboratory data which is O 2 ,
developed over three, four, six months, a year maybe-l' 3 to reactor typical that is four year long load time 4 regimes. We also hope to provide some sort of l
l 5 I predictive capability and improvement of these i
! l 6l calculational models. The whole idea is to provide l '
l 7 some support to the plant life extensions dialog 8 which is underway right now.
) 9 Crack fatique under realistic, that means 10 variable amplitude loading. Up until now virtually
- 11 ,
all of these tests have been run under constant load 12 amplitude, some minimum loads, some maximum loads, l () 13 just cyclea between them for the whole duration of i
, 14 ! the test. We do not know for these pressure vessel i
15 material and piping material, plus PWR bulk
- 16 situation, we do not know what will be the effect of 17 spectrum loading.
18 It's been something that the aircraft 19 industry and to a certain ex ent the marine industry, 20 ground transportation industries have been involved j
21 in for many, many years. We in this business have 22 not been aale to get to it simply because of the 1
t 23 t echnol og y d evelo pmen t that has been going on for 24 the last twenty years. Now, we are finally at the
)
- 25 point where we can begin to address these sorts of l -
- --.,---.---n--<---,--,c . , . , - , - - - - - . - -,.--,,--.,-.--,-,----n .
- v. , .,,,-g- ,-.--n- ,. .---., -. - ,- ----- --. ,----. -n--
l 458 4
1 things, the somewhat more difficult tests, and we 9 2 are doing that within the context of our program.
3 We are beginning easily, using simple 4 prominations of loads which are typical of pressure i
l 5 induced transient during the operating history of ,
6 PWR's. We are doing some experimental work, at the j l
7 same time we are trying to develop an analytical j l
8 model to account for these combined load l 1
9 indirections and developmental effects. ;
r 10 There are lots of models kicking around to i
11 just handle the load interactions. Many of you may!
1 12 know about Wheeler and Whillemborg, and all t h( e h 13 sorts of things, but models to handle both load plus i
14 bulk ar? not generally available. ;
15 What do I mean by simple combinations?
1 r, something like this where we will be cycling over a i 17 load ratio of 0.9. Hill was showing work at 0.95. l l
18 We are in the same general area. Occasionally put l l
19 in a complete unloading, as you might have to get 11!
l 20 you completely shut down the reactor, or cycling of i
21 load ratio .9 with a little overload and an under-22 load, typical of perhaps a turbine trip or something 23 like that. Or the symmetric opposite of that, an g 24 underload followed by an overload.
25 ,
Those things are different. They are ,
459 1 subtle differences when you just look at them, but O 2 there are measurably d if f erences when you do that 3 sort of testing. We are headed in the general 4 direction of brand line spectra. You will see an I 5 example of what we are doing towa rd s the end which 6 is pseudo-recognized. You will recognize it as that .
7 But we have got to begin with something, a bite that 8 we can chew. I think just to drive into something 9 completely at random would not allow that.
) 10 I don't think that I should go through all 1
11 of the next five or six graphu which are in your
- 12 handout. They are there for completeness and you 4
{
() 13 l can read the legends on them.
14 Here is as an example what happens to one a 15 of these load ratio 0.9 tests if, as I did in the j 16 very beginning, shows some overloads which is just a i
j 17 little too harsh, and what happened was that I was i
18 getting crack growth. This is a crack length versus i
19 cycles kind of a plot.
20 So, at the start of the test, this was the 21 constant amplitude of 0.9, got some crack extensions ,
i 22 and then hit it with an overload. That absolutely 23 arrested the crack, and nothing happened for about 4 24 million cycles. I finally gave up, increased the 25 amplitude of the constant amplitude portion. Still i
460 1 with a load ratio --
I'm sorry we went to a load O 2l ratio of 0.8 actually. That was enough driving I
3 force to get that crack going again. Went back to 4 the normal load range of 0.9, hit it with another 5 overload, froze the crack in its tracks again.
6 That's enough of that kind of thing. That's a 7 learning experience.
8 MR. SHEWMON: The overload does not 9 increase the crack length, it just blunts it and 10 stops if then?
11 MR. CULLEN: That's ight. The crack 12 arrests via overload. That's potentially useful.
i
() 13 Certainly some industries have found deliberate 14 overloading, proof testing of structure skimming 15 verf, very helpful. This is a helium gas bulk, this' 16 particular test. So whether that would work in a i
17 PWR environment is something that is upcoming in t h e.
18 , next couple of months of testing.
19 i Certainly we do have the crack closure 20 camp of people who have a better model and will 21 l explain it that way. I am also in the plasticity 22 school of thought. I'm trying to combine this 23 plasticity with what I feel about my hyd rog en 24 ansisted developmentally cracked growth, I haven't 25 yet been able to merge those two in my mind. In I
~
l ,
I l
461 1 fact, I don't even have a clear approach on how I'm i I
2 going to do that.
3 Here is one more example from a little 4 further down in the handout, just to show you that 5 if you do apply an underload as in the example from ,
I I
i' 6 those spectra earlier one, you can get some 7 enhancement, that is some increase in crack 8 extensions over by comparison with the constant 9 amplitude test given by the rectangles there.
10 Everett, this is an example of that 11 psuedo-random kind of spectrum. The International 12 Cyclic Crack Growth Rate Group, of which MEA is a '
() 13 member, Argonne is a member, the NRC itself is a 14 member, has suggested this sort of a spectrum as a 15 candidate for a round-robin set of tests to be 16 conducted by ICCGR members. In fact, we have run 17 this spectrum in helium gas environment a couple of 18 times.
19 We just began this week to conduct the 20 first of these tests in PWR environment. What it to 21 composed of, in one block which is essentially a 22 load ratio 0.2 load to model a start up, you then 23 add about 1800 cycles at load ratio 0.9 which is 24 intended to model the power loadings and unloadings, 25 and every 25 or so cycles you insert something with
(
O
n4, *m- ,
I 462 l 1 a load ratio of 0.7 which is intended to model the
' O 2 trip event. So, this is a three component block 3 loading scheme which includes in some ro ug h 4 proportion the kinds of load amplitude that one 5 would expect in a design spectrum.
l 6 What's the payoff of this? Well, we uant l
7 to know whether the linear damage calculation which 1
8l is currently suggested by some peo pl e , and in fact I l 9 think there is a section in the code itself about l 10 this, we want to really know whether that is going 1 ,
l 11 to be conservative or not. In other words, will l
j 12 l overloads benefit and guarantee that they will bc
- () 13 l l
conservative or paybe in the presence of an i l j L4 I environment the overloads will not benefit and that '
l 15 will not be a conservative approach to the 16 ] accumulative damage assessment.
l 17 ; MR. SHEWMON: Linear is number of cycles l
18 1 over life?
l 19 HR. COLLEN: Yes, somothing like a fatique i
20 crack growth analog to the Coff-Manson stress life.
i 21 Speaking of stress life, the approach in this )
l 12 particular sub-task is to test both smooth and 23 notched fatique specimens. Those are basically j p 24 smooth tensile fatique specimens, some of which will I v
! 25 '
have notches with specific notched factors of both a. I
-e- - _ _ - - - - -
1 j 463 1 106 carbon steel and class 1 welds in that same
(:) 2 carbon steel. These tests will be conducted in both<
i j 3 air and PWR water at PWR typical operation t
I 4 temperatures, and in addition to the smooth and 1
5 notched small specimen tests, we will be conducting i 6 testo for the purposes of comparison correlation on
{ 7 four inch diameter girth welded a 106 pipe, class 1
! 8 weld, A106 pipe.
I 9 The whole idea is to determine whether the 10 results of the axial fatique test, pure axial 11 fatique on this four inch pipe, whether the results 12 of those axial fatique tests on the pipe can be i
() 13 predicted in some way by the small specimen tests on i 14 essentially identical materials. The ideal would be 15 to show whether the design curves in Section 3 of l 16 the ASME code can indeed be used to predict the 1
l 17 desigr life of component such as this four inch pipe .
18 For the PWR tests that we will be running l
i 19 on this four inch diameter pipe, we will be l 20 internally pressurizing with PWR environments. That i
21 assures the pipe will be in just hot air. At this
]
22 point in time we have completed the base line.tosts --
1 23 MH. RODABAUGil: On the girth weld, you're 24 not going to put any defects in tha t?
)
1 25 MR. CULLEN: No. These are intended to bo
}
i I
E_______._________________________________________
i
{ 464 l l
1 1 field-typical welds. We are frankly having trouble ,
l
() 2 getting field-typical welds in pipe such that the l
3 test section is stripped. It's been a real problem i
! 4 for us to do that. The results of smooth specimens I 5 in both base metal and weld are shown on this slide, j 6 showing that both the weld on the smooth specimen ,
i i
7 generate roughly the same -- show generally the same ,
I l
8 ; performance. ,
9! i The reason for that basically is not !
j
! 10 because the weld and base plate are the same in the f
i 11 l terms of yield strength but because the veld is so j i ! l
) 12 l much higher in yield that when we test the specimen !
i 4
J
() 13 l '
i as a weld in it, rather than failing at the weld, 1 j
i 14 ! t the specimen fails back in the base plate a way from i i 15 l where the weld is. So, reall it is somewhat ironic, i
i 16 i those blue dots which are supposed to be -- in fact r I
17 are from specimens with welds in them really failed i
I j 18 ! in the base metal, so it is essentially the same
( l j 19 l type of cant as the red squares.
I 20 ,
Do you get what's happening now? The i ,
21 i tensile specimen has a weld in it, but instead of 22 falling in the weld, it's failing at the base metal. I
- 23 So, we are in fact getting the same test results as i
24 ,
just a pure base metal specimen. The question is 25 whether that's going to happen when we get into the
.f 4
l l
465 '
1 environment. Even though the yield strength of the O 2 weld is higher, what will happen when we put these 3 tests into an environment, which is where we are now 4 in the test program.
5 The notched specimen results are shown on
- 6. that slide where we have tested specimens with threo l
7 types of notches of varying degrees of sharpness, 8 pr ov id ing us with different notch factors and 9 applied them versus the companion test of smooth 10 specimens. Now, these data in order to execute the Il stress, we are using in this particular plot simply 12 { the net section. The net cross-sectional area, okay?
() 13 necause we found out that if you include notch 14 factors, that takes the colored data points and puto l
15 them way up here.
16 We don't totally understand what is going 1
17 on at this present time, but for now it just seems l 18 that the notch specimen data correlate with the 19 smooth specimen data if we f o rge t about the fact 20 th7t there is a notch in there at all. Just simply 21 go with the cross-sectional area of the specimen, 22 the remaining cross-section area.
23 Yes, it does tend to but the notch data 24 slightly below the smooth specimen data, but 25 certainly looku better than including the notch
466 1 factor and applying that data up where it belongs.
O 2 MR. S II E W M O N : These are displacement 3 controlled?
4 MR. CULLEN: Yes, yes. Well, everything h 5 to the left of about there is displacement 6 contrclied. There are only a few tests that we can 7
l run in load control and get away with it.
8 l At this point in the program, we have 9l completed the air test, we have completed the 10 l assembly of minature autoclaves which will include 11 these tensile s pec im en s and will be initiating those 12 l tests in the next few weeks. We have begun to welds I () 13 up these four inch diameter pipes and are trying to L 14 develop at the present time a way of gripping these 15 l
pipes which will allow the axial tensile, load to be l
16 applied and also contain the environment on the I
17 , inside.
I' 18 The last component which I would like to
- I 19 l chat about is the continuation of, shall we say, 20 l conventional environmentally assisted crack growth 21 rate tests. These are test results for compact 22 specimens, and as I indicated in the beginning of 23 the presentation, we have been asked to phase down
{) 24 these sorts of tests which are aimed at defining 25 materials properties in favor of more application l
\
t__________.._ ._
467 1 oriented tests, but we in the first couple of years O 2 of the program when we were running these sorts of 3 tests, what our objectives were to try to extend 4 these tests to a load ratio range of 0.85 which is 5 higher than the low ratios of 0.2 and 0.7 that we 6 had run most of our earlier tests at. Hig her and 7 more typical of what one finds in ordinary PWR l
8 operating scheme. l 9 We were looking at temperature ranges from 1
1 We were 10 200 fahrenheit up to 640 fahrenheit.
11 looking at orientation effects and we were looking 12 at material chemical effects. The whole thrust of
() 11 this research was to try to determine whether or not 14 there were some combinations of load ratio or test 15 frequency or in fact some other variables wnich %.
i 16 would provide data which resided o u t s id e of the ASMU 17 Section 11 reference curves.
18 I've shown on this one question marks as 19 to whether or not we can find some combination of 20 critical variables which will give us data residing 21 outside the reference curves due to either 22 circumcritical variables or due to the variabic 23 amplitude loading schemes.
)
24 Some of the materials, and especially the
)
25 piping matorials that were includes had i n this mm-____,-am -
_--e,--w--&e-.-- -mN vt,__--w-9m-9.--=miwemp',--u--+--rwiam.se9w--i--,a-.-.-,--ye-e wa -,--te'-**--m=q mm-'-M--*'-=wm"+tW " WW (W-W'e-*F-W*'-NPP-*'r'='
i 468 1 study, are shown here. There were carbon steels, O 2 A106 and A516 and cast stainless steel CF8A. We did 3 not in this p r o g r a ra have any aged capped stainless 4 steel at all.
5 Some of the critical variables which we 6 were taking a look at in low frequency load ratio, 7 and this slide is an older foil. Since that time we 8 l have conducted some tests at 0.85. We were looking {
9 ; at orientation st ud ie s , looking at temperature 10 l effects, and we are also looking at the fractograph 11 in che materials that are shown.
l 12 [ MR. S il EWM O N : Just for my general
( 13 education, which may be beyond hope here, but this 14 l CFA351 is an ASTM number like A106?
15 MR. CULLEN: That's correct.
16 MR. S !! EWM O N : And the CF85 is put out by 17 whom?
! 18 MR. CULLEN: The 351 is the chemical part 19 I of the definition. CF85, CF8N, they are all in the 20 351 category. In your handouts, I have accidentally 21 misplaced, mispositioned, this particular sl id e .
l 22 You will find it in your handouts four back from 23 where you are right now. I had to put it up front 24 here where it belongs. These are air environment or l 25 base line tests in A516 showing that the high 2
F-469 1 temperature test data reside very close to the room l
O 2 air data shown by the straight line.
3 That slide just simply gives us the air or 4 reasonable inert environment base line which is 5 needed for the environmentally assisted data shown 6 later. I have drawn in that base line on these 7 items in the handout which followed, and the base 8 line is not shown on my view foils, but there are 9 points that we need to show here.
10 This addresses the frequency effect that 11 fi r . Etherington was worrying about this morning, and 12 Mr. Rodabaugh. You were asking whether or not in a I
() 13 i PUR environment there is a frequency effect on the 14 fatique crack growth rates. Well, here it is.
1 15 These are test data for 1 hertz, and squarc3 are the 16 test data for 17 milihertz which is one cycle per 17 minute, showing indeed, as tii ke confirmed, that thio 18 particular piping steel there is a significant l
19 increase in crack growth rates on a per-cycle basis l
20 when you go to the slower wave form.
21 Similarly on the right-hand side we have 22 the simple kind of thing, 1 hertz data, 17 milibertz 23 data. You should note that there is a rather 24 straining V shape over here. Instead of growth rata 25 monotonically increasing as Delta K increases, they E
470 1 did not do that. These odd or non-monotonic shapes O 2 can be in general directly related to the 1
3 microstructural influence.
4 These particular steels are somewhat 5 loaded with magnesium sulphied inclusions, and in a i
6 PWR environment that magnesium sulphied inclusions 7 cause very active crack tip conditions. Generally 8 cause significant increase in fatique crack growth 9 rates. We were able to find in the cracktology of I
10 t i
this specimen large numbers of magnesium sulphied 11 inclusions when the crack was generating data in 12 this area, kind of mid-specimen, and we ran out of
() 13 magnesium sulphied inclusions and again got into ;
14 i more of them t o wa rd s the end of the test when the 15 I crack growth rates increased.
16 One point that needs to be made very 17 i strongly for there particular material is that the l
10 l growth rates are very micro-structurally dependent.
I 19 That is shown again on this slide here which also 20 gives you some idea of orientation offects. On the i
21 luft we have again the frequency dependency. One 22 l hertz data, one cycle per minuto data, and on the 23 right some idea of the dependency of this data on 1
24 microstructure to identical test specimens which l
1 25 ; physically resided in the plate one right next to L___._._.__.-__.__ _ __ _ _ . _ _
471 1 the other and yet they generated crack growth rates O 2 different by effectively eight or so simply because 3 of the way the inclusions happened to occur on the 4 crack plane of the two different specimens.
I 5 MR. S ii E W M O N : You say these were different?
6 MR. CULLEN: No, they were the same 7 orientation specimens that were physically r ig h t 8 next to one another. We tested them both using l
9 identical loading conditions, identical environments, 10 and yet the data, the crack growth rate data, are 11 different. When you look the fractograph, you find 12 just as it happened, just because of the position of
() 13 the specimen in the plate, magnesium sulphied 14 inclusions on one which gave rather high crack l
i l
.5 growth rates and significantly fewer magnesium j 16 sulphied inclusions just happened to be intersected l
17 by the cracking, giving rise to the difference in l
18 crack growth rates.
19 MR. S il EU M O N : What fraction of the area of 20 the surface was covered by sulphieds? l l
l 21 fi k . CULLEN: Small fractions of the area 22 to answer that question directly, magnesium sulphied 23 as a number ~to pull out of the air might only be ]
24 found on two or three percent, but the influence as 25 those sulphied specimenn dissolved in the water in l
l
,r 472 1 that sulphied spreads throughout the crack tip O 2 conclave, that is a poison to the passivation 3 processes which occurred at the crack tip and 4 allowed a large amount of the environmental assisted .
5 MR. SHEWMON: Or else injects much more 6 hydrogen depending upon your viewpoint.
l 7 ftR. CULLEN: Agreed. Okay, the l 8 conclusions from the tests on this particular kind 9 of steel is that there was a great variability in 10 l the fatique crack growth rates, perhaps due -- I c a n-11 even say definitely due to the microstructure, l I
12 j generally a high degree of environmental
() 13 susceptibility.
14 In other words, as we can see from the 15 ; position of the air environment data line to the PW R 16 developmentally assisted data, there is a rather 17 significant amount of environmental contribution 18 there. I did not bring the fractograph along. It 19 just doesn't display very well at meetings and view 20 graphs, but I have written NUREC's that are 21 available that show the fractography in some detail.
22 You can see in there that there are 23 features which are generally associated with 24 environmentally assisted growth. So, there is a
)
25 correlation between what one sees in the
473 1 fractography and what one sees in the fatique crack 2 growth.
3 I think we can go rather quickly through 4 some data on A106 Grade C. As it turns out this is 5 some air data on the left, just to allows us to 6 develop a base line. The base line has been plotted 7 on the following view graphs. Shows some data with 8 a load ratio of 0.2 and a load ratio of 0.7. For 9 the load ratio of 0.2, there is not much of an 10 environmental assistance, but one does find a 11 measurably amount of environmental assistance as you 12 increase the load ratio, although the air
() 13 e n v i r o nm e r. t tests were performed at a load ratio of 14 0.2.
15 In an air environment, you will have to 16 accept my word for it on the basis of other tests.
17 There is not much of a load ratio effect in the 18 Delta K range that we are talking about here. So, 19 relatively or nearly so, the same air default line, 20 air reference line can be used for the load ratio 0. 7 21 tests.
22 The message here, as one increases load 23 ratio in these constant amplitude tests, one 24 increases the amount of environmental contribution.
25 These are tests which were conducted at 550
474 1 fahrenheit. In your handout, I don't really need to O 2 show them, are test results for both lower 3 temperatures and higher temperatures, specifically 4 200 fahrenheit, 640 fahrenheit. And if you plot all 5 of those results together using trend lines, you 6 generally find this, the 93 degree centigrade or 200 7 fahrenheit data represented by the solid line, the 8 288 data by the bottom line like so, and the 338 or 9 640 fahrenheit data by this kind of dotted line.
10 Remember I said at the beginning of this 11 when you ask a question, perhaps the lower 12 temperatures, generally speaking, provide a bit more
() 13 of the environmental assistance. That is clearly 14 l!
true for certainly the higher Delta K regime in this 15 material. Down in here it's ver y dif ficul t to say 16 because the slopes are very steep. They are 17 anticipated as being somewhat affected test start-up 18 transients which it's not really valid to make 19 j conclusions in that section of the curve.
l l 20 i MR. RODABAUGH: Why don't you tell us why 21 those plots are such a scale that those slopes are 22 so steep?
23 MR. CULLEN: Why not stretch it out?
24 Because mathematicians like to see things on a one-25 to-one scale and because we have three orders of
____ - _ - ___- ___ _ 1
475 1 crack growth rates --
'r 2 MR. SiiEW M O N : You wouldn't want to make 3 the mathematicians unhappy Just for clarity, would 4 you?
5 MR. CULLEN: If you select a specific 6 value of cyclic stress intensity factor and then say 7 what are the crack growth rates at that particular 8 value and plot them as a function of inverse 9 temperature and point toward an Arrhenius treatment 10 of this kind of data, if you do that for a couple of 11 specific values of Delta K for that A106 data that 12 we saw a few minutes ago, you can see as the
() 13 temperature increases, and now we would have to go 14 from right to left in order to see that, as the 15 temperature increases, crack growth rates decrease.
16 Tha t 's a typical cold plate temperature, 17 that's a typical hot plate temperature. About 550 18 fahrenheit and about 640 fahrenheit. So, that's 19 relatively good news for the application of this 20 particular kind of data. As we get up into the 21 operating regime, crack growth rates tend to fall '
22 off in this particular kind of steel, perhaps I 23 should say even this heat of steel. We haven't 24 tested but two heats of this A106. So, I would be O I 25 very hesitant to absolutely generalize, but l
C r-I 476 1 certainly we have found through the work that we O 2 have done, crack growth rates decrease as 3 temperature increases.
4 As you can see in the dotted line which is 5 from some earlier work that I did on a 508 class 2, 6 there is a bit more have a complex temperature 7 effect.
8 MR. SHEWMON: Just for chewing over beer 9 some day, I guess when you get together with your 10 i experts, as I recall, the stress crack cracking data.
I.
11 ! at least for turbine steels has a significant l
12 j temperature increases with temperature. So, I guess
() 13 that's stress crack cracking and there isn't, but it 14 clearly goes the other way.
( 15 MR. CULLEN: Our stress crack data shows I
16 an enormous drop off at that 320 temperature. It l
17 j drops off the shelf. We figured that's another plus.
i l
18 I l with PWR. I think it is materials dependent. I l
l 19 l think it says so right here essentially. 508 is l l 20 l1 headed for a strong increase in fatique crack growth 21 rate, and if you say that there is a stress cracking.
22 component that is superimposed on a mechanical l
23 component, then it must be the stress cracking 24 .
component that is taking off here.
25 I have some stainless data which we will
477 1 see in a couple of minutes here which also shows --
O*
i 2l I well, it shows the temperature affect slightly i
3 j
increasing. So, very roughly, data on these kinds 4 of steels in air is a very shallow U. I've drawn it 5I much too severe. Very shallow U. The range is okay.
6 That is a conclusion that's only about four or five'
! I'm not sure entirely of the validity 7 j weeks old, and I
i 8 of that conclusion. More talk over beer about 9 whether that's true or not.
10 MR. RODABAUGil: When you say air, dry air --
11 MR. CULLEN: What is air?
12 MR. RODABAUGH: What is air.
O)
(, 13 MR. CULLEN: We don't pressure our air 14 environment at all to dry it out. Put the thing in 15 an oven and run the test. For the A106 then we havd i
16 ) these conclusions. There are relatively small 17 l environmental effects for the load ratio 0.2, but to 18 the load ratio 0.7 a relatively large environmental 19 l effcet. For the load ratio 0.2, there is a 20 relatively small frequency effect, and increase in i
21 the temperature results in decrease in growth rates.
22 Load ratio 0.7 increases the frequency, provides 23 higher crack growth rates.
- r~T 24 Now, this is not clear from the exact set
\ /
25 of data that you have, but again, if you take a look
l l
! 478 l I
l 1 1 at the NUREG's that I've put out on this sort of l'
/~T I
() 2 i i
thing, it's a little more obvious. This flys in the i
a 3l .
face of what we were saying earlier that lower i
4l frequencies give higher growth rates. That is true i
5l for a great many materials. This particular case at 6 this particular load ratio, I want to stress that.
7 '
This is a particular combination of critical areas.
8 lii g h e r frequencies give higher crack growth rates.
9 This is an area that we are just beginning to move 10 i into in this business of high load ratio, high test 11 frequency effects.
12 MR. SHEWMON: What's a high frequency for!
( 13 '
you?
14 MR. CULLEN: One hertz, five hertz, ten 15 hertz, twenty hertz, in that general area. The 1
16 ,
highest arguable reactor typical frequency would be i
]
l 17 about sixteen hertz, which is ro ug h l y a thousand CPM i i
18 which is sometimes invoked as a pump frequency. So, 19 there are reactor typical operating load 20 applications at around the sixteen hertz figure.
21 ,
They are very, very small. Small in amplitude, but 22 when we are talking about these very,high load
. 23 l ratijs, thar is exactly the regime that we are 1
?
24 looking at.
(J3 25 Let's wind up by taking a look at some of
479 1 the cast stainless steel effects. On the left again; O 2 we have air data. This data points are from some 3 MEA tests. The data line is from a compilation of 4 literature dated that is predominantly due to Lee 5 James' work on 340's, 316's and CF8's, 8M's and 8A's 6 that he has done at Westinghouse. So, if one takes 7 the roughly 300 degree centigrade data that Lee has 8 generated and put a straight line through it, it 9 comes out to be that for comparison, and it is this
( 10 straight line that I have plotted in the following 11 view graphs.
]
12 This one over here shows tests with a load
() 13 ratio of 0.2 for which there is virtually no 14 environmental effect whatsoever. Tests at a load 15 ratio of 0.7 for which the factor is immeasurably 16 environmental effect. Again, while this air line 17 has been developed for low load ratio test, there is 18 only a little load ratio effect in the inert 19 environment.
20 We have since for another customer 21 completed tests on CF8M at these same load ratios 22 plus a load ratio of 0.85. I will be able to 23 concl'ude that data in some presentations in the 24 future, but as yet it has not been presented to the 25 customer himself, but I can tell you that the data
s
,m i
480 1 essentially resides right on top of this data that 2 you see here, and the load ratio of 0.85 results are 3 right about here, residing just above the load ratio.
4 0.7, preserving the monotonic load ratio dependence 5 of this crack growth rate data on load ratio.
6 Again, we don't need to go into the next 7 view graphs in any detail, but it gives some results 8 for a growth rate result on this material at a load 9 ratio of 0.2 and a variety of temperatures and for 10 two orientations, LC and CL, showing first that i \
11 l there is very little environmental effect. Nothing 12 that you can really measure and that there is l
l l
() 13 virtually no orientation effect either. On both 14 li sides the LC and CL data reside just about the same 15 : regime.
t l
16 j MR. SHEWMON: The code has the 316 curve 17 like this and A316 in water or doesn't have either?
18 MR. CULLEN: Doesn't have either.-
19 ; MR. SHEUMON: But it does have these codes l
20 ; for ferrite, is that it?
21 MR. CULLEN: When you say code, you mean 22 Section ll?
23 MR. SHEWMON: I guess so, yes.
24 MR. CULLEN: No, there is nothing about 25 piping steels in the Section 11 Appendix A for
1 481 1 fatique crack growth.
' O 2 11 R . SiiEW M O N : Just pressure vessel steels.
3 MR. CULLEN: Yes. That is why I do not 4 show the Section 11 reference lines on these dots.
5 If you would like to use them, I suppose that's up 6 to the utility to present that to the NRR and try to 7 get away with it. Let me just say, though, that if 8 you d id plot the code reference lines, you would 9 find that this data resides well within the code l 10 reference lines, which I think is an interesting 11 point to make, but it is not really code relevant, 12 shall we say.
() 13 There is data in your handouts at 450 14 fahrenheit, 550 fahrenheit, 640 fahrenheit showing 15 in all cases very little environmental effect and 16 virtually no orientation dependence at all. If you 17 put trend lines on that data and then co-plot the 13 trend lines, you will see small increases in crack 19 growth rates as the temperature increases. Leaves 20 data at 450 fahrenheit with higher data at 550, 21 still a little liar at 640.
22 So, from the work on cast stainless steel, 23 we can see that there is little environmental effect 24 of load ratio at 0.2, for none there is a measurable 25 but still is not serious by comparison environmental
482 1 effect at higher load ratios. There is relatively O 2 little temperature dependence. This is another way 3 , of saying what I just said, a normal load ratio <
\ l 4 dependence. ;
5 As you increase the load ratio you get 6 reasonable increase in fatique crack growth rates.
7 Again, I did not bring the fractograph along, but 8 the fractograph on these is for environmentally -- l 9 for tests in a PWR environment appears to be very, 10 very brittle like in appearance. That really in my i
11 mind rings some sort of an alarm. l i
i 12 Why is it that we are in the crack growth
() 13 ,
i rate data seeing virtually no environmental 4
14 f component, no environmental assistance in the crack i
15 growth rate data but the fractograph is just totally 16 different. If you do these tests on cast stainless 17 steel in air you get predominantly ductile crack 18 growth rate. Do these things in an environment, 19 crack growth rates are the same, fractography is l 20 totally different. So, there is something 21 mechanically really different about what's going on, 22 but in terms of data, not much.
23 MR. SHACK: Did you ever look at it under i
24 a microprobe to see if it's following the ferrite?
25 MR. CULLEN: It is definitely not
483 1 following the ferrite, yes. In fact the crack O 2 orients itself very carefully and with a great deal 3 of forethought so that it intersects every Delta 4 ferrite plane as near perpendicular as it possibly 5 can. The crack will see some Delta ferrite in therg 6 and screw itself around until it cuts right t hr o ug h .
7 Very interesting.
8 MR. SHEWMON: That's for the environmental 9 or --
10 MR. CULLEN: Yes, environmental. And 11 again, just to close and repeat that I think the 12 various aspect of this work do relate directly to i
() 13 improvement of various setrions of the ASME code 14 which is really in the end what provides the driving 15 force for doing this work. And with that I conclude .
l 1
16 If there are anymore questions, I will try my best i 17 to answer them.
18 MR. SHEUMON: What is the mechanism by 19 which this would get into the code?
20 MR. CULLEN: I think all I can do is give
)
21 you a little bit of history how it worked a little 22 bit once before and probably would work that way 23 again. After a suitable amount o f da ta is generated 24 and plotted against the code as it currently exists, 25 if a modification of the code is called for, in l l
484 1 other words, the code just is in some other space O 2 from where the data exists, then a document is 3 prepared to the code committee and discussed. It's 4 really rather straight forward. It takes time.
! 5 Took a couple of years the last time in 1981.
6 MR. Il U TC H I N S O N : Back to the points you 7 made with respect to the compact tension versus 8 part-through test, you indicated that with the 9 center crack -- the compact tension the environment 10 has much more access to the crack. Is there direct 11 evidence that this is important? I mean do we know i
12 ,
that's important?
() 13 MR. CULLEU: There is mixed ev id ence.
14 l There are some test results which seem to indicate 1
15 that capturing that environment will allow changes 16 in the electric-chemical, will allow changes in the 17 ionic contents which could conceivable provide 18 changes in the environmental contributions.
19 MR. HUTCHIUSON: So, a more limited access l.
20 might control.
21 MR. CULLUM: Yes. I say there is mixed i
22 results because there was about a year and a half 23 ago an experiment where the environment was jetted 24 in to compact -- fracture compact tension space.
25 There was no change at all from the same situation
I 485 1 without the environment being jetted in. So, I O 2 don't know what to make of that in particular.
3 But we have lots of concerns overall in 4 this program with the clad there, if you do 5 penetrate the clad. The clad could be a fairly 6 tight crack. Even though it would be a through-crack, 7 it could still be pretty tight and would not allow 8 ingress and egress of the environment as you might 9 suspect.
10 So, these are the things that we are 11 trying to look at. The NRC has asked us really to 12 take a broad brush look at a number of potential
() 13 problem areas. It's okay if we don't really do ther 14 in depth, but try to use our background and pick a 15 few possibly critical variables and see if there are 16 areas where we ought to look hard over the neck two 17 or three years or something like that.
18 MR. SHEWMON: Thank you very much.
19 Closing comments?
20 - - - - -
21 Thereupon, the proceedings concluded 22 at 4:20 o' clock p.m.
23 - - - - -
24 25
1 CERTIFICATE 2 STATE OF OHIO :
3 COUNTY OF FRANKLIN : SS.
4 We, Scott N. Gamertsfelder, and Barbara Leonard 5l Rogers, RPRs and Notaries Public in and for the 6fState of Ohio duly commissioned and qualified, do 7 hereby certify that the foregoing is a true and i
8i correct transcript of the proceedings had in the 9; aforementioned cause and was completed without H 10 I adjournment.
11j We do further certify that We are not a 12 ! relative, counsel or attorney of either party herein, I
Q 13 or otherwise interested in the outcome of this l
14j action.
15! IN WITNESS WHEREOF, We have hereunto set our l l
)
16l hand and affixed our seal of office at Columbus, l 17 0 io, on th' d f f , 1986.
18 /
BARBARA LEONARD RO G E RS ,(/No t a r y Public - State of l 19 Ohio. My Commission expires June 14, 1989.
l l
N/}?)JI 1 YfA ]
21/~SdOTT'N. G 4E RTS FELI}jlR, Notary Public--'3 tate of
~
Ohio. l My. ommission expires August 1, 1987.
22 23 ]
O 25
)
_ _ - -N
Y' e
APPLICATIONS -
ORIENTED FATIGUE AND FATIGUE CRACK GROWTH STUDIES IN LWR MATERIALS WILLIAM H. CULLEN '
1 I
MATERIALS ENGINEERING ASSOCIATES, INC.
9700B M. L. KING HIGHWAY LANHAM, MARYLAND 20706 (301) 577-9490
20 YEARS OF ENVIRONMENTALLY - ASSISTED CRACKING RESEARCH IN LWR MATERIALS RESEARCH MILESTONES HISTORICAL MILESTONES 1987 STRUCTUR_A_L ELEMENTS TESTS _
(986 TOTAL LIFE TESTS 1985 SECOND lAEA SPECIALISTS MEETING 1985 VARIABLE AMPLIWDE TESTS BEGIN APPLICATIONS 1984 PART-THROUGH CRACK TESTS BEGIN 1983 EPRI BEGINS MECHANISMS RESEARCH 1983 is83 STRESS-tire TESTS BEGiN 1982 TEMPERATURE EFFECTS 1982 MEA CONTINUES NRC PROGRAM 1981 PIPING STEEL TESTS BEGIN AT MEA 1981 FIRST IAEA SPECIALISTS MEETING IN FRElBURG DATA 1980 SULFUR CHEMISTRY EFFECTS 1980 SWEDEN, FINLAND,lTALY OPERATIONAL GENERATlON 1979 waverORu errECTS 1977 ICCGR FOUNDED 1975 -is7T LOAD RATIO ErrECTS~ ~ - ~ ~ ~ ~ ~ ~ ~ -
1975 NRL BEGINS EAC WORK HARWELL ,CERL,KWU. C-L(now UNIREC)OPER AT-IONAL 1971 KONDO PRESENTATION AT WRSRIM GE OPERATIONAL METHODOLOGY 1969 AEC CONTRACT T_0_WESTINGH_0_USE_
1965
CYCLIC CRACK GROWTH CYCLIC STRESS INTENSITY, .w
~
[
As08-2 Sie c O R N
- Spec. 071-22 //
~
q' ~ IF mHa sinusoidal /l R = a2 t Q R l , ..-.
t ; ,
% 'Q ' l k ,k * -
o$ l k D s
CRACK ,/
- l GROWTH 8
- R R R R' RATE I 'g l/ ,
~ (IN./ CYCLE)
Q - ? 8
- q. ; (MM/ CYCLE) Surface (w.ier R<= .25),eFiaw lo/
s
( Q [s '** r !
e lSubsurf ace Fiaw (Au)
Q l 8*,8 k
Q: ,-. CYCLIC STRESS lf e
N INTENSITY, .w l l , ie
- NN .
a n.'i . . . . . n'.I' i ,,-s !/
..-a a . ...
i .. .
,e .a . ASME // CYCLIC STRESS w .W j
~
CODE lNIENSITY,
/l RULES i8 WATER REACTOR lieso) /'
ENVIRONMENT / ~
550*F. 2000 PSI o / *el *
..->- / /
j CRACK l
l GROWTH l l RATE l l ~
(MM/ CYCLE)
Sursace Fi..
(water R ( e ,2s),,e e l :
,(IN./CYC LE) e
- so *y e
l Subsurface Flaw -
8 s ,s (Air)
E ,
?
t e ,
e s
, l ,l , i.-*
I e ,
~~s na -'
so ' --
CYCLIC STRESS -T INTENSITY, i
1
TASK 2bu TASK 2c TASK 2d TASK 2euu TASK 2g TASK 2f if CONVENTIONAL THREE CLADDING GROWTH CUMULATIVE INITIATION da/dN vs AK DIMENSIONAL EFFECTS MECHANISM DAMAGE -
MECHANISM SHAPES STUDY RULES y y v TASK 20,2f TASK 2a PREDICTIVE CAPABILITY S-Ni or AXIAL LOAD for GROWTH S-Nf P!PE SECTION XI CURVES
\
PREDICTIVE CAPABILITY for INITIATION SECTION lli PREDICTIVE CAPABILITY for TOTAL LIFE OF STRUCTURE Subtask Interactions for Task 2: Environmentally-Assisted Crack Crowth In LWR Materials
i t
l RELEVANCE OF MEA RESEARCH TO PLANT LIFE EXTENSION ISSUE
) Addresses Environmental Effects i Low Cycle, Long Term Fatigue
- Life Prediction Method Development
) Fatigue and Fatigue Crack Growth Total Life Mechanisms of Environmentally-Assisted Cracking Reactor-Typical Test Matrices
! Materials -
Clad RPV, Piping Steels Flaw Geometries and Loading Schemes 1
Structural Element Fatigue Tests I
4
! AC RS 3 R ES E N ~~A~~ O \ OU-~_.NE 4
l
- 1. PART-THROUGH CRACK GROWTH STUDIES j a. Panel Specimens - Piping & RPV Steels j b. Structural Element Fatigue Facility .
l 2. VARIABLE AMPLITUDE LOADING STUDIES ,
i
- 3. STRESS-LIFE TESTING OF PIPING STEELS
- 4. FATIGUE CRACK GROWTH OF PIPING STEELS I
- a. Carbon Steels 4 b. Cast Stainless Steel 1
i f
1 i
4
GROWTH OF REALISTIC (3-D) CRACKS IN CLAD AND UNCLAD PRESSURE VESSEL STEEL IN PWR ENVIRONMENT Approach' Conduct Tests to Correlate 2-D and 3-D Growth Rates Develop Experimental and Analytical Techniques to Monitor 3-D Crack Extension Construct Cyclic Pressurization Test Facility to Grow Flaws in 12" Diameter Pipe Sections in Air and PWR Environments
! <,i,r. L , ; ,
ff f
- / S h E Y
\
N R T
I t
i T E E M d
f M
O E
O E
G
? G ls W-d 3 O I
L A
N S
N
._ E
- M
~.
I T
D N -
E S E E
- M N N R l t
i O E T i -
R I WT 7 V O T D N L A N E F P A L
- A N
O I
S y- A N
E M
I D
~g Y O
W R T f, T E
N I
- M
=
7 O E
S N
/
f/
l G E l
/ d 2
T T
A P
W
~ O L
F
~ ' -
'l ,' ! lll 1 l\_
- : l l I!
. > ,. !' l ! ! l 1 j N
f s#
g,
[\\ L
- / A
, ' E R S 7
E I
- - d 3
R T
E
' M O
E
- G K
C A
ll C
N O
I
/ S N
E M
I s L D
' A E
'f E D
E il I
H
- T
/ d 3 D N
A L
A N
O I
S N
N E E M M I I
C D E O P
OuO S d
2 W
T NN
^
-a - - _ - -----u . . - --- a m..o. ee, a -. _w- -, a a
Y%m, .enu.: 7 nww:.: -. n.?d ,1~ --
w p,9p,4hy'Nui',h<V )
Q.pe. : % -?
m.g w J e
Y ,
t g.
v .<w,
- . :o. .
+ ;
.m .
wa -
44A' .
fkh$1
~
i
.s-c --.-
i 5
I m
d e
j M
7' g i{..', {f '." -
- :p3 . _
,A I -
h CO N
l 1
W WD L
OI O- &
- Z1 <C 3_
U O T k
~
i g o -
N \
g W o
i t wX g _
i D _
6 elumm I
amme e _
. O O e_*
e '#,
1
,J
_ P ,
--m=== ,
,~' ..; . s _
4 K:
i'
?:2 1': :,s.. -
L'd
, . .:ly,,' O f' , :Ijgk $_ , __
7, - p -j.
e, .
- , @ l.
s.y t
_ . :z '~ * * .~ 4,e .y3%. . P .'s - -
,l = 7., ,
- e,
- 1.L2,6
- 7 . ., > ,
) 2: r -
y . .' ' .~ ,
f_
ej. .M- -
. , . , ,i4.. g ,l*, vj qp$,W , ,
1..
- ' g
., ,' ',, . A .' ., ; * ' . .'e 0r.*
l @
Ti ; , j? 'n;* : * *~,'
\ , . ,
,,C
& *( :. . ' ;c. ."-;h%
wt- 7;_ . . l ' v i '
~ s.g 3
~
C
~jh,*.c.;1 ,
.~ . ,' ' ;, .fG
~
t i
l l . )1 , i , !! , l l a .
i s iheb.
9 $b' ' .
. .. 8 6Af[
j$
- a. 'l 1
l l
j l
l L
7 N
l N 4-l E C F 6 _
M S 8 I
C2 EE 7 _
PX 1
S 5
. l l
f
- 1
(
l l
l l 4 l
l
- !)! -
!!l l\ l ll< >
SPECIMEN N2 XE2SC-l 550 F
+. -
-- c.,--,c-~--.w-=m} 'e.s. ggjih. ' _ r.
?"
C Q. ] ':, , f' . 's c., 7 : ,3f ". . ... . . . . .
.f." "V, ,
..a . ** E enl,..4.s '.
. a. .:.,; .. . ;. p
.;: ; v ;.n. ,. .g.. u.g
-: ..s. . , . .
_ ; . ;,y
. g_, .
ggz.:..
.~., :
.. 't. *~ . y.
y 1:2,
'. .: . ,.???r '
'. g 9 -
jg
,' [-i 9
. [,f 'c . . 3 i 8.'
e.,.;;
T i
. Qj .
.i
! E f.i'
,,' U.s' lll lll lll llIlllijiji 1lijililiji I
lji i ll l l l [ l i l 5 '
6' 7 8 j
CALCULATED FIT OF PART TilROUGil CRACK EXTENSION DATA - AIR TEST Experiment SURF 4 2.s n
c 2,.. o o ,
%# D
- 2. 4 <
U N
- . : - o
.C s 2 c) o C o o 3.s . o' J o
'o 3 ,
3.5 o D
U o rJ o
' 2.% - 0 U o P
O U
- 2. ;
2.2 n. 2 2.* 2.c :.2 Electric Potential i
FATIGUE CRACK GROWTH RATES VS. APPLIED CYCLIC STRESS INTENSITY FACTOR FOR SURFACE TRACE AND THROUGH THICKNESS DIRECTIONS OF A PTC APPLIED CYCLIC STRESS INTENSITY, ksi V in.
I 10 100
. . i i iiiig i . . .,iiig i i _ ,4
. A 533B Steel - 0.004*4S / -
50 I -
24 C (75 M f -
~
I 3 _. Room Rtr Environment /
10
- A *a*-dtraction I -
"O "c*-dtreetton j .
- R = 0.2 /
i l
e f e E . 3 Hz sinusoldal 5 *U y 7 -
10 g i e i .- c e
r / U -
N m 10 " - / . E
' 5 h
- Ie g g - -
o - I g i$ .
I b . ] 10
~
W 4 -
M 4
~
d
$^ 10 T h
~
l .
co l
~
/ 7 10 '#
l .
I '
' I'1 ' ' ' I '
10 ! 10 100 4 RPPLIED CYCLIC STRESS INTENSITY, hda 6 4
9 J
--n.. .-
Cyclic Pressurization Test Facility Test Vessel
~
PWR Environment fth "D
]
g 91..
_J
. _ ? A*
I[. @ <s,
,a _ g. ;- ,
v y- .
[ 7- 'g,*
, a.
', $1 1, (\ Elliptical 5 s' ,
o and Semicircular,oFlaws [/+
- {"p:.e p>
Q A /,
Pressurizing Fluid
EXPERIMENTAL FEATURES OF CYCLIC PRESSURIZATION TEST FACILITY Test Sections up to 12" Diameter j PWR Environments at 2500+ psi, 650 F
- Pressurization Via Single-Phase Fluid External to the Test Vessel
- Two or More Flaws Per Test Vessel D. C. Potential Drop Crack Extension Monitoring
}
GROWTH OF REALISTIC (3-D) CRACKS IN CLAD AND UNCLAD PRESSURE VESSEL STEEL
. IN PWR ENVIRONMENT Payoff Characterization of "Real" Flaws in "Real" Structures in Air and PWR Environments for Licensing Decisions Significance of Clad-Induced Residual Stresses to Vessel and Piping Integrity i
Y
_ _ _ _ _ _ _ _ e
MECHAXISMS OF ENVIRONMENTALLY-ASSISTED CRACK GROWTH (EAC)
Approach Select and Nvaluate a Mechanistic Process and Calculational Model Conduct Selected Experiments to Support and Verify Process and Model Meet Staff Needs with Visiting Scientists 1
i B ulk E
corr. Enclave .
corr. da = A , e. . n '
dN Inst.
inst.
j n
- H 2O z da m A e. n dN Ave. Ave.
j -
J+g u
Possible Reactions:
1 M+HO 2 =M++e-aq 3
4
, ; M + H,0 = MO + 2H + + 2e -
4 1
PRELIMINARY RESULTS OF ANALYSIS OF MEA DATA WITH GABETTA MODEL SINGLE CYCLE STRAIN RATE (1/sec)
~4 ~ ~
-1 8 10 10 10 ' 10
- 10 10
~I ' '" ' ' ' ' ' '"
10 "I '"I '"I
- 288 C (Sh0 F)
~
PWR water -
R508-2 STEEL ~
A 0714C
- CO23C :
10 ' '
l o CO224C :
~ o 07122C '
o R108fC v 1HT2BC -
-2 '
10 Fo FW15C o 1HT2RC A533B STEEL g :'o W72C5C 'd o o L8318C _,' :S ,
's
{ .o L831SC 7 l@ ^
$ v W72C2C e!
-3 _
10 : 4 o : CISE Base 1ine I'3
%o :
- p -
_c fi d -
g 10 ~ a b4 gg 10 : mm ,
A A
W
^
g -
o 3* . E g %o o E 5 *
,0 gk o 10 $ <
-s _
' o 8 8 o 10 v o
- o
- a _gg . ;
g' o l o o o -7 '
1 7 10
-s
, , ,,,,,,i en. , , , ,,,,i . - ---- _ _ ,,,i , ,,,,,,2 10 -s -4 -3 -2 -1 a 10 10 10 10 10 10 SINGLE CYCLE STRAIN RATE (1/sec)
MECHANISMS OF ENVIRONMENTALLY-ASSISTED CRACK GROWTH (EAC) l Payoff Ability to Extrapolate to Reactor-Typical Load-Time Regimes 4
i
); Predictive Capability and Improvement 1 of Calculational Models l
l i
CORROSION FATIGUE UNDER REALISTIC (VARIABLE AMPLITUDE) LOADING Approach Conduct Tests Using Simple Combinations of Loads Typical of Pressure-Induced Transients Develop Analytical Model to Account
! for Combined Load Interactions and Environmental Effects 1
i .
i i
1 J
I _ _ ___ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 0
i I
4
- SIMPLE VARIABLE AMPLITUDE WAVEFORMS .
Complete Unloading Overload / Underload 0
O Underload / Overload I b f ,
M M i >
I i
' C
Crack Length (mm)
- a : : :
b b
b b b
g Apply Over1ood 8
- F E' 2 %
e' I *E. ~c n* 0 w Q o ." * *_ O l
N
. . e E. .r f. E. $.
~ G*Yn,'
O
(
I
?
a-2 .
- .
- 1 2
- a s d a
o O
' M l
E , increase Lood Range, force Growth M l
, ', Resume Normal Lood Range CD Apply Overtood -
k Y
)l n -
~~
O -
Increase Lood Ronge. Force Growth d I Ol g N. 5 Resume Hermot Load Range -
O O M
'9 Appy Overlood 3
e - -
l -
_ O o
! l O j
3 increase Load Range. Force Growth
~
q J. , , g Resume Normal Lood Ronge - 0 l
~ .."*'***s W r
e
.. .N:
- ~,. M I - -
M e+
[ ..s.. g ,
I U
g I .
w.
O
, , , , , , , , , , p i- a r
.- ?
8 E
?
?
?
Crack Length (in.)
^E" 5 ocj xyLU 9 a 8 e 0 e 0 0 6
0 5
1 8 7 2
. e. s. 1 1 1 1 1 2 2 2 4
_ R 5 E '
T 3 A
. W '
g R
_ W A P
_ A D A N
A '
A ,
3 M A U
_ I L A E
l l A )
Y A '
R '
D 5 0
' 1 N A 2 I
S x T h S (
_ E T s O D e I l 2
c T '
_ A A y R
D A
/ C f
O 0 L D A o 1
1 C } 0 ,_5 r
e 1
1 ' A 1 b 1
F 0 m u
O D N g A N D O
S D I
_ R 0 A A ) g P r e , 1 M ' % e a. a D
. O 2 t a
c c(
_ C 1 0 W . D A u z 0
_ 0 R t H t i
= P e 5 D A H
S , r D 6 , o
_ . ) 1 4
)
0 , 5 l F - -
' ee 0 R
O R
e r
e D
t 5 H H t 0 S 5 a
- ( 9 . . W B c c (
3 C 8 e e 3 p p z 5 0 = S S H 0 B R 2 R D A 1
-, D
- - - - - - - - - L
- 8 8 8 8 8 8 8 8 8 8 8 8 8 8 6 4 2 0 e 6 4 2 8 8 6 5 5 5 5 5 4 4 4 4 4 3 3
]E " 5my_ . . fs'o _
- !1Il. ' , l 4 i ;
i' . ,
_ _ ._ _ _ . _ . _ _ _ _ _ . _ . _ _ . . _ _ _ _ ~ . _ . _ . _ _ _ _ . _ _ _ - _ _ . _ , . _ . . _ . _ . _. __,...m._ . .__. . _ _ _ _ . . _ _ . . .
i l
COMPARISON OF TEST WITil AND WITil0UT OVER/UNDERLOADS IN PWR ENVIRONMENT l
I i i i i 2.30
! 58.3 -
R 5333 Steel -
56.8 288 C (550 F) -
2.20
~
Reactor grade Water -
54.9 -
O Spec. H8R-IS R = 8.S O -
2.10 0
52.0
- S Pec. H8R-11. R = 0.S with IJL/OL .
g D -
] 1 Hz sinusoldal 0 -
2.00
- E 50.8 gu 5 D / -
{
48." -
D g }
J c
46.S a
Q B
/
/ -
1.00 n DU - x o
e c
44.0 -
BO ~...#./ -
1.70 0
=
s D ...;........'.* -
o 42.B -
g O
.....v- !
Q ;g -
t.60 l 49.B -
0 -
i
- ... Each Bar Denotes UL/0L ApplIcat1on 38.g
.y 0
. d . O . .D ............ .
1.50
- D -
36.8 f,B D , , , , , , , , , , ,
8 .25 .5 .75 1 1.25 1.5 1.75 2 2.25 2.5 Number of Cycles ( x 10)
-.. . _.- ~. . . ~ - . . - . . ~.. - - _ _ . _ . -- .. . _ . - - - - . _ - _ . _ _ ~ - . . - - . _ _ - ~ .-- .-
(
l r
l EFFECT OF UNDERLOADS ON FATIGUE CRACK GROWTH RATES IN HELIUM GAS ENVIRONMENT 58.0 R 533B Steel t 208 C (550 F) ,/ -
2.25 Hellum Ces Environment /
56.0 - C Spec. W8R-4, R=0.9 -
2.20
- Spec. WGR-12. R-0.S with Underloads 5 Hz sinusoldal -
.15 54.0 -
[
j'. -
2.10 l 52.e - g _ ,,g3
,.B n E u -
2.00 .
E 50.0 -
j, g c
1.95 [
/ g r Each Vertical Line Represents -
j P 40.0 -
/ 0 1.90 .c Location of an Underload Cycle p 01 f g l
h 46.0 ,
,[ 0 1.85
[
e J -
0 i.e6 a x /
1.75 m 44.0 -
M U
l o'
- 1. 7 ^3 e 8 U 42.a _
1.65 p
B -
1.60 40.0 -
g 1.55
- O 30.8 -
8 ~
8 58
,,.. 0 B -
1 45 36.e L' ' I I I I g g , g 9 .5 1 1.5 2 2.5 3 3.5 4 4.5 5 8
Cycle Count ( x 10 )
/
r
=
0 3 .
f ok .
c .
d o .
nl .
M EB .
U - .
R b .
T u .
C S .
E .
P .
S .
T .
S .
E .
T .
P .
U .
O .
R .
C .
E .
T .
A . k R
l i
. c T
W O
o R l C
K 2 .
B C f .
A ok .
R C d o c .
e -
C nl EB n I
L C b O -
Y u .
C S L
'T s N
I i -e 9
- l f
F ok e c
O . y0 c l C =
C d o c y
7 .
R I
nl 0 .
5 T EB C = 2 A b 2 R M u /
E l
l S 1 C .
s S .
e
- l
- c S.
y 0
. - C =
. R 5
2
=
l 1
l l
4 i CORROSION FATIGUE UNDER REALISTIC 1
I I (VARIABLE AMPLITUDE) LOADING Payoff An Assessment of Conservatism in Linear Damage Summation Approach 1
4
1 1
i DEVELOPMENT OF STRESS-LIFE CURVES
- FOR XUCLEAR-GRADE STEELS f
l IN LWR ENVIRONMENTS 1
i l Approach 4
l Test Smooth and Notched Fatigue Specimens of:
A 106 Gr. B Carbon Steel Class 1 Welds of Carbon Piping Steel i
i l Conduct Tests in Both Air and PWR Water l
J Conduct Tests on 4" Diam. Girth Welded Pipe l
i Axial Fatigue to Match Small Specimen Tests l Internal Pressurization During PWR Tests i
- O i
W E
o I
5 I t;
W l
- /
O /
d /
m ,
w W O / d m %I I >.
cc u
/ l
/
o /
- /
/:
O l$:
1 w/ *
/
/ 3C
&/ l
/ i>
b SS381S
Stress-life data for ASTM A-106 Gr. B carbon steel in air environments at 288'C (550*F). .
10 4 _ _
~
A-106 GR. B CARBON STEEL AIR ENVIRONMENT, 288'C (550*F) -
O SMOOTH SPECIMEN (BASE METAL) _
O SM00Til SPECIMEN (WELDMENT) 10 4 2 la 3 -
n n.
=
5 Z --
5 T T 5 -
ASME MEAN DATA LINE M 2 - FOR CARBON STEEL _ C I I a
c -
o e
e oO _
10 3
=
5 gg e Z A
" 5 z og -
ASME DESIGN CURVE 0 g _
i I
i 10 e 10 1 ! ' ' ' ' ' ' ' ' '
10' 10* 10* 10 10* 10* 10 Cycles To Failure i
Stress-life data for ASTM A-106 Gr. B carbon steel in air environments at 288 *C ( 550* F) . .
10
- _ _
- A-106 CR. B CARBON STEEI, -
AIR ENVIRONMENT, 288"C (550*F)
O SM00Til SPECIMEN -
A NOTCIIED, K g = 2 O NOTCHED, K = 3 -
t Q NOTCilED , K = 6 t
10 9 2 10 3 -
3 o.
e Z -
5
]
=
A -
a -
e ASME MEAN DATA LINE _
~
- a. FOR CARBON STEEL o.
e _ o e I 10 8 I i =
- i
- 10 = -
Z O 00 -
Ob
~
ASME DESIGN CURVE O A O O
la 1 10
- 10* 10* 10* 10* 10* 10 Cycles To Fatlure
4 i
4 i
i CONTINUATION OF ENVIRONMENTALLY-ASSISTED i
3
! FATIGUE CRACK GROWTH DATA DEVELOPMENT ,
I 1
l Approach i
Continue, But Phase Down, Use of Compact Specimens l
Extend Limits of Critical Variables 1
1 i- Load Ratio to 0.85 i
j Temperature from 200 F to 640 F
! Orientation Effects I
Material Chemistry Effects l
1 l - - - - - - - - - - - - _ - - - - -- . ._ _- _ _
4 1
a 3
i RPPLIED CYCLIC STRESS INTENSITY, ksi W .
1 10 100 i i i i i i
- . ii..; i i i viiii _3
/ / -
10
~
/ / -
~
/
} / ~
/
l /
i -2 _ /
- 10 -
/
/
~
I / ,
- /
l -
/ I
/ / o i l
i 2 ~
/ I 4T I E / / 7 10 g o - v.ci nie agitis m O I s
, e E
I / o -
~
j Q w/ O / -
1 g 10
-3 -- O /
'O
/
I - E
'g o , _
Bc 1 I
I ~
i.l @
I f
- ~
I l 5 5 I x
! m I _ -5 M J
$ I / -
10 m U
w I /
~
W a
1
- e 4 ~
~
~ I /
- C -4 Surface Flew -
e
<weser n,..ssil i 10 -
/ w i j : I /
\
I I
~ * ** ~
f / Subsurface "' Flew MtgitFrequency O I */
OO y;O
- y
-s i _
- 10 I
~
i i;i i il O I d -5 i i i i i i i i n iil i -
i 10 j i 10 100 i APPLIED CYCLIC STRESS INTENSITY, MPaVii 1
J !
i ASME SECTION XI REFERENCE LINES l
! l I
1 i . l
Steels Included in This Study Carbon Steels A 106 Gr. C -
Two Heats High Temperature Service A 516 Gr. 70 Medium Temperature Service
. Cast Stainless Steel ,
A 351-CF8A ,
High Temperature Service
i i
}
CRITICAL VARIABLES INVESTIGATED i
Frequency (1, .017 Hz) -
A 106, A 516, A 351
! Load Ratio (0.2, 0.7) -
A 106, A 516, A 351 I.
)1 Orientation -
A 516; A 106 & A 351 (in part) l i
I Temperature -
A 106, A 351 l Fractography -
A 516, A 351 l~
i I
t I
RPPLIED CYCLIC STRESS INTENSITY, kalViii.
i 1 10 100 i i i 1 i iisig i a i i i ri i l ' J 10-3
,R 516 Gr. 78 Low-Carbon Steel -
~
j 2ee c <sse r)
N'" '""'" "**"'
l O
) fa Spectmen FOK-26 .
i j , i Hz sinusoidal ,
~
! - R = 0.2 o E _
, e -
10 4 s&. !
i l - - c j . -
j l'J -
td i g 10'3 - -
E l m
i I -
g g -
, o - -, "I l " 0 J
w ~
U
! U 10 dM e _ U
)
6 : E i .
U k $ 4
- W S 10 o
8 g ~
o - -
l La,, h C 8 La.
- ~
a Room Te.mp. Atr
)) .
D o
R *= C.25
)
~
p 10 6
].
o
/ .-
i .
1 4 ' ' 'I ' 'AI ' '-
10
. I 10 100 l, RPPLIED CYCLIC STRESS INTENSITY, MPaViii I
I 4 FATIGUE CRACK GROWTH HATES FOR A 516 GR. 70 i
- STEEL FOR 288'C (550*F) AIR ENVIRONMENT i
i i
i l
1 k
i i
i
- -. ..----.._ - - .- .~.-----..---- _. . - - - . . . - - - - - - . - . . ~ . - - _ - -
APPLIED CYCLIC Sil(SS INTENSITY, ksaVii. APPLIED CYCLIC STRESS INTENSITY, kstVis.
i 18 tag i 18 tas
. . . .....g .....y ; 3
290 C (Sla rl NG C (558 f t
,g 4 sp . rou-se gg a ,se . rou-as
. . .. .e., :. i ee. .. ....
i
'. ...e
~. 17 = = s ee.s e *
' tr an. .. *
-A. g g - m.eet. q c.e. es.t.e .
q
, t.e se es s.t.e x x 2.
=
x
' a - e.a ,g 4 ;
o a a.a .
g ., o 3 - T-s oce s.se -
f } _ t-s oes s.s / g J - J -
518-' ;-
- [=,
f $10
-8 -
{
x 3 -
. r E 2 3 -
E i e
a a .
s
- 18 4*
5 18 4ya h 5
I N
a :
. u I
u E 18 * - U 18 * -
u t
- 1
' E ' # E l ~
4 4
- 10 ; 18
~
8 :
l 4 -' ' ' 4 ' ' - ' "' '
18 l 18 I 18 'm' I I 18 MPLIED CYCLIC STRESS INTENSITY, MPaK 15 MPLIED CYC.IC STRESS INTENSITY, W e #
l i
FATIGUE CRACK GROWTH RATES OF A 516 GR. 70 l
I i !
_ _ _ _ - _. - -v _ . . - - . . . _ - , . , - - , , , . . . - - . - - , ,. -,--..--m . . - . . . _ . . - . _ _ _ . , _
{
FFPLIE!! CYCLIC SINESS INTENS!TY, hasG. HPPLIED CYCLIC SMSS INTENSITY, ksiG.
l t 18 las i 13 las l
.. . .i . ....rr : i,-s r- --uv"'i'r ta '8 a sts G,. re L -C.,m ss I a sis G,'. 7. L 4 .es ss..i tee C esse rs see c asse rs 3 g,4 _s, . ron-a. g a w. .., o, ., [
. eon-.. q
- .=...i [ . s.
- . s, . ,on-s-i
. er = . ...
. F :
-. . - e. . ; ;; . ir .. ....i i y n-.
[ _j i,-. q
{.
- e.
t-1 . ....
_ 3, 4 q 5 3 1 + o,. ..s.
, ; 5 a y a J -
- a 5 ta '
. g 5 ta-'h
/ j g 5 ** 3 E t - y p'
- ; 13
- d y ,l4 13 5
$ =
u
{ * $
v o 3 4
M M M 2 :a' -
a M ts*h 4 h
E -
E E w
t-
- i. E m
- w I i k l 4
7 13 13 3
i d ' ' " *" ' " d - * " -^ 2d -
tr I la ti'J ta't la 100 ff9 LIED CYCLIC STRLM INTENSITY, NPr5 fffLIED CYCLIC STRESS INTENSITY, t1Pr5 FATIGUE CRACK GRonTl! HATES VS. APPLIED CYCLIC AK FOR L-T AND T-L ORIENTATIONS OF A 516 GR. 70
I i
CONCLUSIONS ,
i
! A 516 Gr. 70 Carbon Steel i
Great Variability, Perhaps Due :
to Microstructure Generally High Environmental Susceptibility Typical Fractographic Features A. Fan-Shaped Features
! B. Brittle-Like Appearance l C. Influence of MnS i
l I
i
/
APPLIED CYCLIC STRESS INTENSITY, km6 G .
APPLIED CYCLIC STRESS INTENSITY, kst G .
I la les
' ' ' ' ' ' l I8 188 -
18' ,,,,m,
.asescc.ct.b'e'.co. ass.i 'i , ,
., ,n gg,,g
- 18 roe c <ssa r> ; . R tes Ce. C L. ~ ~
2ee c csse ri gg -a ate cavec as ,- -
Sp... fop-sa, a-e.2
- , sp e. rop-ss, t e -
18
- e !H .tav..ed.t s,.e. rop-rs, c-a
... tr .H .sa e .ee.i
- -
- N . eau..es.: ; , ~
& ; - s. . roe-s i, n- .r
? n - e.a d s ; -3
$ - 1at )a 18 s g
17 .w. .,no.. ...
J ,
J a. s.e se w.s., . -
E : O J "la-8 :
E E gg -a
[ ,# -
E E : m
- E [ -
3 E 5 5
g g t . # -
a o s -
/ 6 3
e 8 ; t84 5 x .
f gg 4x
(
d
- E M fi /
-: M w j -
u 6 -
.' 6 O gg d #
M //
- a E :
I -
sw !!8' r t./ ,f, ;
- t E
.f '1 -
4 -
f 10 s - gg 4
~
- t :
f gg 4 . . .....I ? . , , , ,,,,1 , ,
l -
gg a 1 Ig igg APPLIED CYCLIC STRESS INTENSITY, t1Pc/ii' I is ggg APPLIED CYCLIC STRESS INTENSITY, t1Paviii FATIGUE CRACK GROWTH RATES FOR A 106 GR. C FOR FREQUENCIES OF 1 Hz AND 17 mHz, AND A LOAD RATIO OF 0.2 AND 0.7
a FATIGUE CRACK GRONTH RATE, mm/ cycle E E E E A A A A a *1**2C* 4
. o 8 G-!n @
P
- 1 =$?: 1 5,.j-e = n m
9 J" "
- .
- : .*- ,n-.' - 3 3 o
n p :
i!: . -
r-p
- . r . a s- n n- -
" I ". .:
m" _ .
ti ;- ES N
m m
I W
. . . . y- _
E
~
- c -
s
= 3 E m
> m 3
g H
w
- o a O_ _: ! : .
c 3 -
- g ",
M % F z O 2
m)
-sl 2
g > ..i,1 . . , ,...I . , , , ,,,,t , , , ,,,,,t_
u O E E E E o M A A A e O
o FATIGUE CRACK GROHTH RATE, in./ cycle t M ar O m
- 64
- Z:
in 2 21 M
d M
O to 2
wi-M *n 2 O H m
- FATIGUE CRACK GROHTH RATE, mm/ cycle o e e
a e
CD -
A A A A a G*?**tC* a m 3 7
o s--in g- $
O M
=
!"15.I-
- ?:
- 1 - .; $' 8" .
G n
Co a
O p : tit **n-
~
e-p
. i. . - .- -
m -
!*t j:: O
~
m mE p , _ ., m- -m N
- k s . . .
E \ 3- 5 m
- E n
. . . . ..\.N -
E
~<_
a _
g -
- j "r
% F 2 .
al -
, ...t , , , ,,,,.s , , , , ,,,,s , , , ,,,,,t-e e e e A A A eIs FATIGUE CRACK GRONTH RATE, nn./ cycle
Ifil4 ! i;I:;'1li,l!. 'l 'i i
_ kE. Nog 56E$J3e1.
I 1
_ I8 s e gg 5 8 y
_ L g
. ~ . . :._- - - - - : 7 - - -[ t . i E a 1 e a R 1 R I
F A 7 A
- P
_ T P P
L
, e p p o H o o S S m 8 e 8C le t si P
L
.e z
.o o o I I
E . a . r ( G E -
D e - 5 r D t F F g 5 . i a O O e 0 C E F C Y , v P P o e - - d F C. Y -
N A C L, C st 9 8 e 3 L V T L , 4 7 o i I I , d M w. C R I I C t g
, a l
L C s e
c.
I R G S 1
- - s og 1 8T S
G O U T R
C L e c e R E
H N E E S
e n S T M S 5 s' .e'e'a Si S
E C I t
e I
N R N N
e . T T A T E
a l
E N
C N S
, s e
S -
K I C
i I
T E F T
,Y N A G ,Yt a t
a .
V T R g ,t ;
. 1 0 k M 8s I I O P a
o e
R G W t V O U T w .
e .
I
.E N E H - - _:: -
M " . ; - - _ - : i_ , - , - 5: i _
1 .-
E C R 3 1
8 i
g gg-N R A 4 4 4 a T A T C E wE.a $5 S8E gJ S e u2 K S G F R O O R W
T A _
H 1
R 0 A
6 T
E G S
.R F -
O C -
R "E N $@ 6o$ 5J : uE .
S g g
i
, IB A T -
5 4 ~
4 E 1
- ~ - . - ~ - - : y' 1 E :._ - - - - ; , i 1 a e a R 3 R 0 L A
? o 3 A P
6 P - e p p a S S s 8 1 P
P e H o e t C 6 i L I L
.o s
. a I G N I E . a . oe ( G E e D
.R D ,
t F F g n O O e e
- S e .r . C 2 C Y
, u P P e C. Y 8 e - - s F . C C ,
S e e 1 L C 8 L ,
st ,e d .9 W L i o
wi I
I e e -
C CI c i. l S S C S B t l L C
- - s t
ai T C .t e r aT T m R E R e E P E n S E h S S
. S L R . e S.
t I
_ I ,
e N N a , e . T I T E
s l E
N e .
N N .
S S . C .
I _
I . T _
3 T A .
,Y 3 ,
Y1 et 9 t t 8 M 8 D i h k es -
- P a t t
C r e N T
6 ,
.n P ; - ;;
W
. - : : - - ; : i_ - . , .
R t e
1 ,
g g
g d 8-5 s wE_N $5 33E gJ ie.>u2 L; !
APPLIED CYCLIC STRESS INTENSITY, kstG.
I 10 100 i i iiii; l
i i iiiiiit i j 3
.A 106 Gr. C Carbon Steel PWR Heter, R - 0.2 2
-2 !? mH sinusoidal 10 33e C -
^
_ --- 2ee C .
93 C
~e .
C 10 E e
l
~
N
- d
/ :
L.I E 10 3
7 / _ J u -
/ -
E u
E :
g /
.f E
5 g
/ 5 x
M l/ x 5 _
10
-s M l/ : 5 s -
s j/ :
M s
~
10
.I
- lI -
I
~
7 10 '
-s i i i i iiil i i i iiitil i gg i 1 10 100 RPPLIED CYCLIC STRESS INTENSITY, MPa W FATIGUE CRACK GROWTH RATES FOR A 106 GR. C STEEL TREND LINES FOR TEMPERATURE DEPENDENCE
- .--_ .-__ -. --__ . ...-- - - - - . - . - . - _ ~ . - - _ .- - - _ . - - - - - -
I 6
1
)
)
i i
1 TEMPER ATURE,* C 400 350 300 250 200 150 100 I I I I i i I ~
10
b i
8 10 7 l
[,, '
\ / 8K = 45 MPa#
2 e
~
\ ~
10* g g \ ,/
2 ~
E
/
! E. N
' O N
-- / AK = 35 MPsv /lii'" a h 10-8 \ / I s -
V E 5 : o s
i, o
- g
- =
i, a U g .
7 10
us 1 u $
s e
c*
, E
- I
- 10 -* 7
- - - - A50s2 -
i
' A 106 Gr. C
~
~. 10 e I -
I I I I I I I 10 ' 20 3.0 1.4 1.6 1.0 2.0 2.2 2.4 2.6 (TEMPER ATURE), = 10 , 8 K
FATIGUE CRACK GROWTH RATES VS. INVERSE TEMPERATURE, i
FOR A 106 GR. C STEEL i
.!i 4
i
)
ge. - - - , , , , , , - - r , 7 . - - .., - - , , . . - - - , .---n-,- --m,--, - - - - - - - . . - - , - - - - - - , , ,--,,~.-g- -.e e.e,-,. .,w-.4 e
CONCLUSIONS A 106 Gr. C For R = 0.2: ;
Relatively Small Environmental Effects Relatively Small Frequency Effects Increasing Temperature Results in Decreasing Growth Rates For R = 0.7 Large Environmental Effect Frequency Effect: Higher Frequencies -
Higher Growth Rates 1
No Orientation Effects Observed For Either R
l! *4!i 4( i!i 2 ! .f 'i]; I ;4 !;jI, l . ! jj ! l i E .d 5yX l ~z @- - jshO I I II B 4
B., a 4
F g . - - , - , ,
- - : :r - - : :T I A R , a 1 n o
R F F
T F I
P
- sp s p sH e 3 5 P L -
I P ,
e e e e E
n 1 ,
I G L I .2 ,e ,e t v C E D
U E , n e a . _
D p p u r s E , s s s e t . C
- - Y C
Y
, S el n = S . C et d e C L C , ,F e nt tee.- LC I R I
.t 2
i nn g S A C10 1
2 2 e
t eg i
g T C S T
e C e s E R
K R s
C 7(
S S E
S S i n N /e
(
5 2 t e .
S G e
- I 5 F el I
N R I N
f 9 3 i T O T E
, n o F
) i E
N W N .
t e i S
I T S I . - i T _
H T
. s i
,Y _
,YI B . e !
i i
i i
g g k R N B
c gs A P iV _
T i E
S d - .
.n i . :; . ,:- - . [ " _ . - . f :;
g g I
g g F B 18 O d -
d 4 s R 'ep8w og5 $gE E J Eyk.
A 3
5 1
C _
F 8
A F _
O E.y vs43A5&e,"{ 3 _
R I 1 0 0 L 3 4 13. f t
'. - - [tCL f{ I[ I
, i
_ O _
L a
- e e 8 2 n i
A C
- 17 s S s sp s ep e e 3
D O ee.se ***e .ee c t c 5
1 r s A R A e
e s
- e. o r i c
- s r P e
P a a g s a A P a t a t L t a s 9" s s e e n t T I E
a t
w s t 8 7 7 m
e r C E
D I D i e t 8 3
,2 e m a c
o o
,s -5 s O c f
n d e e at w t y
c t
S C t
i i
el e " - -
-
- a e S
t I
!1 .a e c C 0 d d . r '
I 2 2 i 15 t O
l 8 .7 -? n 0 -
S
_ F T e l
4 E
S s
s 5
5 0 S ;e S
' I I t N
. e I N
2 T e ' D _
E N !3/ f, l S
I 4
S '
T _
A I T /* ,Y _
N ,Y '
k D P1
' s i
0 a 0 v0 i
I 0V 1
0I
. ni
.E 7
, q 1 lI g
t g g 10 _
4 a~'
- s "E_n vex 345 &e ~ .oi.
. I;! , >, : .
4 ii wE.g 3 M og 5 5 J 3 uxg.
1 g
_ I I
0 s S t i
- 8-
- - - ; , . t _
- ' . ..: - - ' - . : - - a e R 2 R A
_ L R 1 7 e 3 P A
c - e pS pS om 2 3 5, P L
P o H o oo e t C 1 I
P '
e .e a o . . e - (4 C,
- E L F D F I mi a
2 e R R g 5 S A E ' t n S S e e R, C D t u - - od F Y T ' e e l t C. C I C Y
' ts o ot e s SO e 1
a s .
L I
G C L
a d e
W a
t .
C U I
' l t e S t ,. I ST S
E C1 ' e R 8 ei S n E T l S C R o S E s .
R S S
e I N -
A ' n\ Si T C I N
i e t e,
E N _
K T E
D a
e l ,
S I
N t . T .
G S I
' e .
,Y R T ' - .
. 1 -
,YI i 0k O B
ts s
8s W H B
V -
_ T P C i s
H Vi i
.n _
- i
_:- - ' - : : ' - _ ~ ': '
R : - - - 1 A 1 1
9 gg 8 T 8 .- 8 4 4 E .
S wEhwvE55g5@J c *u*
F O
R A
3 5
1 C
F 8
A F
wE g $y 0g5 $w a uxg.
gg O
t 18 IB -
o - f R -
- - - ,- , ' 1 t
- C R e e R 2 R A
?! e 3 P A -
L - e pSpS m o 2 3 5 P A t C 1 -' L P o e H o o c . z o. . o ee I E
T P L
( C D
E I ee a
2 e R R g 5 O ' - 4 T M E ' t n B S e 8 R C -
Y P
D '
t e
s u
e l
- -7 od F C' o '
C L
E C Y '
os st te 4 e
)
a ' I C
- R C L
a d e
M t '
A I al t S' 1 S e t l' T C1 8
"' e at 8T R
U S n E T S R R t
e S E E e '
S s I S ' N S' N
T O I ' one t e
E .
F N N T ' D e' S E a l ' I N ' t ' T 2 S 'e ' ,Y 3 I T ' - '
2
,Y 91 "'1 3 I' 18k 8 4 '8st C M P C V i
s ' n Vii '
i
- . _ " _ ~:~'
', : - - __ :~ - ' : -
1 1
g g m -
8 '
5 4
mgkw v$x $g5 gg c,so^o,=
1 l wE. "gD 53 2 gd $ uxge g l i 1
8 a s g s
d 4 y
- - . ~ - .7 e a R o2 en
, i I L R ! R
- ? P A C = S S s e 3 P P m p p e s L P ,
O .e H e e t c t- i I
.e er ( c L c E e F . E A
I E ,
et n
2 e A R g s o
- s ri D T D , t t
n S B r e n C C a u - - e i Y
I Y , t t e
7 7 e r c. C G C ,
t e l 3 2 e i as . L I
U L ,
n d W t . C I ,
e a E C i t s S S
t e
r t;
e 18T R T i C R E
n E S
R S ,
i
. i S A
C S
I ,
n o
N S e I N
N r si T K T , o *, f- t ei E
N E . .t S N es v G S I
. e i I
T R T . -
i
,Y O ,Y . a i
s W
t n s
y 10k 0m T M P c t H a V
ii-i G R . ' : l - - : - - - : [ _ - - [ : ;
A I 1 g T ' 8 g g E 4 e a
S EO 5e5 5g z g J $sux3e F
O R
A 3
5 1
C F
8 A
F O
R wE. gd 53 2 gd $ ux]e g
g 19 is g A
- r' d 4 1
- - - - ' - - - - - , - , i 1
- 7 T c- n LP e e en ez n A P
E A P
4
. - S S e s a e p p s s P M P '
O s. H e o E
t e
c t i L
I P L r t 2 .e .o r c c E D
A R g s- esi r I
E E ' e a si R D n C
-S -Sa r a a i t
e u Y A C Y
tt s
e 7 P d r c. C L
T C '
a 1
S e i es .
- U L ' a l
d W s .
I C
I ' a a _
- R Ct 'I l t e
sti 1 S E S e r 8T ei R _
T n E R S O
i E e S S g e i F S ' n\ s I I ' ., N si T 2 N t E T ' no o i N 8 E t e S 8 N ' s s I
S ' e i T
- I ' - i
,Y T
C ,Y l '
s i e 'e s g i
1 k 0
s 0s N c t P
e A ' i G i
i
- - - - - : - - - -g
[ :
gg 1 lg 4 9 4 a EO 5e5 5g z gJ [ uxg~e
wE $ D 5gI$J -
o>;
1 g
8 gg g 18 4 4 '
4 1
. ~ . . .: - ~ ~ ..:- -
i I n o" e Sp 3 R f l
e 3 P f e 1 i o e 3 P lP s 7 s .o c 5
L P .
H 1 I
L e e a -
E F I r H R < C
- r. D A E .
s
- z e n S s
- e e D e R. C T C e
e ei u n S t Y I d n e ,9 C Y
mC..
e u a L G C .
s ce s I U L I
w oi d a e C-s .
C E C1 . s d t
.L S. 1 S 8 ., a tg 8T S l e R i
T E C R n E l S R S . e 'e e.
s S
A S nN e I C I . ,i N S. T K N E E
T . oss t e
- e. N N .
e l .
S I
G I S
. - . T R T
. a
,Y O ,Y 01 i
. i l ls -i ak W B ss T M c P tV H e I A
i
, . R I .
R ,.:- - - - - - : 5_ , . l : :;
A 1 1 g g g
T 8 8 g E 4 4
- s S mES5 5 y @z,_z y J E R>;, -
F O
R A
3 5
1 C
F 8
A F
O R wE $ $$rr 5J o>g.
1 1
8 g A ,, 8 '
s
- 4 ' t
~ ~ ..:- - - - , ~ , - I 1
. ~ . . .: -
T n n o e e p S 3 R 3 A P
E F - .e 1 i e e 3 P M IP 7 H .e c tS L P P L
e se e a
.a e H R ; .
e .
/
m
(
RPPLIED CYCLIC STRESS INTENSITY, kst W .
10 100 I .
i i i i i i i iiiig i 3
j i i iiii) ,, 10 R 351-CF8A Cast Statniess Steet PKP Water, R 0.2 _
17 mHr sinusoid al
-2 _
10 I 204 C, 288 C & 338 C
_4 E -
T 10 g K
u s
5 s
J d 3 g Q10 5
E -
E 288 C $
o -
338 C x ,
7 10 vN w
B a
~
O IE 4 -
' 10 204 C
- I T 10
~
10 100 1
10 RPPLIED CYCLIC STRESS INTENSITY, MPa W i
CONCLUSIONS A 351-CF8A Cast Stainless Steel Little Environmental Effect at R = 0.2 Little Temperature Dependence at R = 0.2 i
" Normal" Load Ratio Dependence i Brittle-Appearing Fractography
=
3 Or R E _A~~ O \S ~~AS 1 MeV)
Interaction between irradiation, environment, and stress unclear Proposed remedies:
Impurity control of stainless steel Hydrogen water chemistry with impurity control l
h t Alternative Materials Previous Experience, Laboratory Tests
" Factor of improvement" Impurity Effects l
CERT Tests i
CGR Tests R Ratio Effects Follow-on Work Butter Definitions of Conditions for Crack Growth:
Threshold Effects, R Ratio and Frequency Effects, '
Water Chemistry Initiation Surface Cold Work Chemistry
l Transgranular Stress Corrosion Cracking Based on laboratory testing Potential problem for Type 316NG and probably TP 347 materials Mitigated by Tight control of impurities in reactor coolant
- Residual stress improvement Eliminated by hydrogen water chemistry e
4
- -----w, o- --~e,--r--- . , c4- -,, - e - ---- - - - - 3 .y--y,,,,., --.e--.y - , ,_w-,. _ ---.m, e -~.. - %, -,- - y- . - - - - - - - -_
Weldability of 347 and 316NG: Cooperative Efforts with EPRI, NYPA, and University of Tennessee Weld Overlay Failure Analysis of Hatch Components Residua! Stress Measurements Finite Element Analysis CGR Results MSIP Residual Stress Measurements MgCl2Test Results Future Work Ferritic Steels Irradiated Stainless Steels l
l
Alternative Materials Significant numbers of BWRs have now changed over their piping systems At present all utilities have chosen Type 316NG as the replacement material Fitzpatrick is installing German TP347 in core spray and other non-recirculation system piping Germans are actively promoting TP347; French are trying to interest utilities in cast (CF3) piping Utility concern was probably raised by early difficulties in fabrication with Type 316NG (microcracking in bent components) and questions about weldability and stress ,
corrosion in impurity environments
s -
Type 316NG -- Nuclear Grade Austenitic Stainless Steel in EPRl/GE pipe tests Type 304 SS pipe weldments failed in 100 h; only one failure occurred in 9 Type 316NG SS test specimens (108 weldments) in 5000 h; Type 316NG in Japanese reactors appears to be performing satisfactorily; low carbon materials perform much better in US reactors than conventional materials
" Factor of Improvement" in laboratory tests appears strongly dependent on the particular test In ongoing ANL crack propagation tests Type 316NG is inferior to sensitized Type 304 SS even in high purity water There is now general agreement that Type 316NG is susceptible to environmentally enhanced cracking in '
impurity environments (TGSCC)
To get full benefit from installation of Type 316NG piping it is necessary to maintain a very high water chemistry standard; to get resistance to stress corrosion cracking comparable to that of PWRs it will be necessary to also implement a hydrogen water chemistry program
( _
CERT Test Results for Type 316NG SS Test Sulfate Conductivity, t , o Failure d No. ppm pS/cm h ME$*, Mode m7s,
-1
- L = 4 x 10~ s 169 0.1 0.90 217.4 462 TCSCC
-10 9.74 x 10 916 0.075 0.69 559.4 449 TGSCC
-0
, 2.03 x 10
~
l L = 2 x 10~ s l
! 187 0.00 <0.2 483.3 460 Ductile -
207 0.01 <0.2 497.3 468 Ductile -
228 0.025 0.25 473.8 462 TGSCC
-10 4.71 x 10 i 210 0.025 0.25 532.4 458 TGSCC
-0 0.050 6.33 x 10_y 199 0.47 588.7 449 TGSCC 2.21 x 10 228 0.075 0.64 562.8 456 TGSCC 4.52 x 10 -10
-10 172 0.10 0.90 474.0 461 TGSCC 7.35 x 10 i
i i
i 1
AVERAGE CONDUCTIVITY OF RECIRCULATION LOOP WATER
! FOR OPERATING BOILING WATER REACTORS -
l 15 g 14 ~
- _ WATER _c0NpucTIv!TLI.IMITS AL25*c l 13- l us/cm 12 - '
IIIGII PURITY WATER 0.054 g
CC g_
r
[ BWR OPERATION < 1.0 o-s: , .
g) 9. -
t 'l
- ^
o 8- '
i u 7-
~
i e ~
.o 6- / /
i F 0 ~q i -
5 /, '
z 4- / ,
3- //
/,'
2- '
luf XX A A 7j g 0; , , .
0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1 Average Conductivity (pS/cm)
~
CRACK GROWTH RATE (m s ')
i , , , , , , , , i , , , y x
HIGH PURITY H2 0 7h j d cO Ei n s co o
.fNaNO 3) j, y g co Z.
N x
Na2 B07 4 >
x "%
i 3 3
- >N Qg 1 CD - ;
) Na2 CO 3 h\\hh g DNaCl\\\\hhhh x gh
[ Na3 PO 4 h hh h g Q Na2 S10 3 ykkhkkkkN ZU .
Nk\\\\\\\\\\N y Na2 HP04 O NaOHh\\\\\\\\\\h\\\\\N hQ mm 2
NaAsO2 xh\\\\\\\\k\\k 2O
[- Na2 S0 4 hhkkhhh\\\\\\\\\N l ) Na2 SO 3 Nhhhhhhhhhhhhh\\\\N Q Na2 S kkkh\\\\\\khhhhhhhhhhN Qg s Na2 S'2 0 3 s h\\\\\\\\\\\\\\\\\ $U B &
. CRACK GROWTH RATE (m s)
a a O O i , , , , , , , , i , , i h x
~
, ~. kn m ho O
HIGH PURITY H2 0s
." 33h e- . -
CO u> o 2
'6AW Q
["3" s h%3 ES
!"24hA%W ?!
O
. H3AsO4 \
i"4 A % %' %S i8 f"'s'P O 4
\ mO H AsO g\\ \\\\\\\\
3 3 h i"2 8vi h%'N%'N'%SS $
("sh'%%%'N'%NNM 25 T
E O
$ h
1 Anodic Dissolution Reaction Me = Mez+ + ze-Hydrolysis Reactions Mez+ + H 2O = Me(OH)(z-1)+ + H+,
Me(OH)(z-1)+ + H2 O = Me(OH)2(z-2)+ + H+,
Me(OH)z-2 + + H2O = Me(OH)z-l+ + H +, ,
Me(OH)z-l +'+ H2= Me(OH)7 + H+
Cathodic Reactions 0 2+ 2H 2O + 4 e = 40H-2H+ + 2e = H2 '
SO4 2 + H 2O + 2e = SO 3 - + 20H-NO 3 + H2O + 2e = NO 2 + 20H-
E e impurity Effects on TGSCC 10~9 is ~
e5 E u c
-E 3
9 e
is -
E -
Type 316NG SS ( 289 C )
o k
10-10 i , . , .
0.0 0.2 0.4 0.6 0.8 1.0 1.2 Conductivity, S/cm
L i
I i I itill l I I I I til i I I IIill TYPES 304,316 8 316NG SS (289*C)
O 316 SS (8 ppm 02 + 0.5 ppm Cf-)
10, V 304 SS (0.2 ppm 02+ 0.1 ppm SO2 -) _
Z a 304 SS (0.2 ppm 02)
~
4 O 31GNG (8 ppm 02 + 0.5 ppm Cl-) -
~ 5T e 316NG (0.2 ppm 02 + 0.1 ppm SOk-) (O.3 -
-6 0T
? 10 _
Y ' Ggcc (0.3
$_ }
$\6 c f
j 30 46 '
gee #go :
TO. SAT S pc g
30A e
~7 Q '
10 3\633 ss,TGsc TYPES 316NG 8 316 SS,SA + 650*C/24 h - -
TYPE 304 SS ,SA + 600*C/24 h -
- 6pG SS'TGSc (a = l.0 % , co = l g.m _
go-8 I I I I !!!! I I I I f f!! I I i 11!!!
10-8 10-7 10-6 10-5
- STRAIN RATE (s-l) t
-4y w w y, - - - w, y -e----w- -9g y w *- ---e -e 9-m,-ega.m-,-. - . -- -= - g w-- v 1------, y-7 ----.-----3
-8 10 CONVENTIONAL AND ALTERNATIVE STAINLESS STEELS (289*C) o TYPE 304 SS (HEAT 53319,SA)
O TYPE 316NG SS (HEAT P91576,SA + 650*C/24 h) -
o TYPE 316NG SS (HEAT 08056,SA+ 650*C/24h) _
A TYPE 316LN SS (HEATS 313XX,SA+ 650*C/24h) 7 O.2 ppm 02 + 0.1 ppm SO2 - _
E S = l x 10-6 s -8 g A a
.O O
4 o
-9 10 O O.02 0.04 0.06 0.08 0.10 0.12 0.14 0.16 0.18 0.20 NITROGEN (wt %)
e
J CERT Test Results for Type 347 Stainless Steel (0.2 ppm 02 , 100ppbSO{)
SP ecimen "
Heat Treatment 9, f ' 'f' m i{o rm ' max s
'1
-6 174-1 AW 77.5 27.9 19.4 435 1 x 10 Ductile 174-4 -6 AW + 500*C/24 h 65.5 23.6 18.2 432 1 x 10 Ductile 174-2 136.6 -7 AW 24.6 '
19.0 428 5 x 10 TGSCC 174-5 AW + 500 C/24 h 114.5 20.6 16.6 417 5 x 10 174-8 AW 328.5 23.7 19.9 434 2 x 10-174-3 AW + 500*C/24 h 301.5 21.7 17.8 448 2 x 10-174-9 AW 676.5 24.4 21.6 443 1 x 10-174-6 AW + 500*C/24 h 574.5 20.7 16.4 451 1 x 10- V
52 9
n E
E o.
R o a E8 % o 5 m _ g P ? .
O o% oM m
CD N.
O 5Bo Ed -
e >_-
co o o .
(n O F o w%5 -
8zF0 to O ND LLI I O Q- -Z
>- N O HOO 6
O O O O O O O O O O O O O' O O O O O O O O O O ro N - -
N to e to (o N I I I l l l l (3HS' Aui) lVl1N310d SS 9OE
~
CERT EXPERIMENTS ON TYPE 804%%
IN SIMULATED BWR-QUALITY WATER AT &%%b .
%00-
@ 800- k ,
z < 8
$ 100- t 0- #
MiSfhWMMMifh%
'100'
, a * *
- j g . 9 g 3 g3 I -200' O ***
2 e e h 4 Wa t h h -300= +
0 =f! MULE %%MRRIi'; .
W -500' h^ 4 g -
A g $9)ab '
- 4 o a Wfe e W = WB OfewF M -600' 9
%ggg
-7 0 0 . . . . ., . .. .... . . . . .., . . . ..,
0,1 1 10 100 CONDUCTIVITY at 25'C (pS/cm)
l CRACK GROWTH IN TYPE 304 SS l l UNDER LOW-FREQUENCY CYCLIC LOADING l
l IN SIMULATED BWR-OUALITY WATER AT 289*C
~
I 10'
^ E WATERQi[MJjiRY ae I 200 l 3 a%250, Fl
> 100 @
lE '
e f ,,oa JMMEN
,k ,f 200 R=095 E
- f. l h-300 1 = 8:10'#Hz I'llb l
' b 10 E" b 3** *" l 3: I' a -***
q) I lO ll I, -'88 E l! !' ll gM -700 @ stNsmzAil0N lC l ll ,, 01 1 to 100 EPR.Ckm:
e CONDUCTIVITY at 25*C (pS/cm) lM k a t77 Z1 0 s to i<
'( t. i i 2 It i 11 11 fl 10 ifit ll il " 20 l
O' "
l (1,3,5) (4) (2) (6) (7)
WATER CHEMISTRY CONDITION L._ _..___ _ _._ _ _ ____ _ _.__ _ _ _ _ _ _ _ _ _ _
a
FRACTURE SURFACE OF TYPE 304 SS 1 SPECIMEN No.27 FROM CRACK GROWTH EXPERIMENT !
i SENSITIZATION: LOAD CONDITIONS: WATER CHEMISTRY:
2 u2 EPR=0 C/cm R =0.95, K,=o 28-36 MPa m 0.002-0.2 ppm O2 2 I f = 8 x 16 Hz, TEMP.=289'C 0.1-1.0 ppm SOf l
l= B/2
=l M
? 'l.
1 Q' ?,'(- } .- ./ W#
1,g }d,;<j;{j ' h.
s '. i ,, -
.~L DUCTILE l {..
j '/d .i e' . ,M - '
- _ f { ' l .;,h l __
, g, -
SCC .
(q
- f. a. . . , . ~ . ~ , . :
f '. ' FATIGUE Il N h
~
g.,Q..A?
,: s'j 0b[Im' ' . ' '- -
.. ) .&)'-J.,,b Opmd
! KNAX'UCs H . .
i j CRACK TIP FRACTURE SCC FRACTURE REGION SURFACE MORPHOLOGY I
i i
4 I
l
j l
! FRACTURE SURFACE OF TYPE 304 SS l l
- SPECIMEN No.28 FROM CRACK GROWTH EXPERIMENT '
SENSITIZATION: LOAD CONDITIONS: WATER CHEMISTRY: l j EPR=2 C/cm2 R = 0.9 5, K,= 28-35 MPa m" 0.002-0.2 ppm O2 o !
) f =8 x 162 Hz, TEMP 289'C 0.1-1.0 ppm SO[ l l
,! != B/2 =l
- v- ' -
- q. , c .; _.
1 V- j ';,
. .k'j>' . DUCTILE I:- ' - '-
4
..Q 1 '
' s. .,.
T h,Ek -
k%j,-f l , ji W i - ; i j p ,': SCC-D , i 2 , l
~ '
FATIGUE } _ [ 1:. ~ lb
,60pm
~
1 ' ' 10'O pm' y -A g ,. . CRACK TIP FRACTURE SCC FRACTURE REGION SURFACE MORPHOLOGY l l t l j , i - 1
FRACTURE SURFACE OF TYPE 304 SS i SPECIMEN No. 29 FROM CRACK GROWTH EXPERIMENT SENSITIZATION: LOAD CONDITIONS: 1 WATER CHEMISTRY: 2 u2 l EPR=20 C/cm R =0. 9 5, K,= ,28-33 MPa m 0.002-0.2 ppm Og ) f = 8 x 16 tiz, TEMP.= 289'C 0.1-1.0 ppm SO7 j = B/2
- :. s DUCTILE P' 2 - 9:d,J. 'k l- "'
f, i
;_ t . . t h.@ ? .;'Q.}.*
~. . ,' y SCC' ). j's ',-:'f ..J' 2; . ,; . .
- ] . ':. f> . ; .l.' .*. !'O '!:'.
} l,. -'
FATIGUE
'~
-llp:7N.f.j.
- 300pm
;43, 4 oop.
- c. .%. :
CRACK TIP FRACTURE SCC FRACTURE REGION SURFACE MORPiiOLOGY 4 i l i l v . I
i . l
-3 10
]
CURV _ 2 linlyr NRR o _ 10
-4 '
C e so4
, , a -
k 8 e - 2 -5 / 10 C
/
/
Y 0.2 ppm 07 ; sensitisec at Il50*f/2 h _ ([PR = 15 C/c,m ) - GC < A 0.2 ppm 02 ; "sitirce a t tit 0*r/2 h - I ([PR 10 C/ca ) - GE I .__ O 0.2 99m 02 ; sensitisce .t its0 f/24 a t~ - Gt - O 0.2 vom 0 2; se creir seasitized cr 10_ O - 0 8 c r 0 ; 2seesitized at "s0 r/24 h M Q GC O O GENED D(D HlI ACHI A Nt.
~
G GECR0 < D 8 ;;m 0 A g ,__ 0.04 inlyr a s a, 2
"I'#'d DJ "' d'"9;
,32.,,,, ,(s,13 -
l 3 8 ppm 0 ;2 sensitized at (1292*f /10 min)' 2
+ (932*f/24 h) (CPR 4 C/cn )
l O 8 ppm 0 2; sensitized at (1292 *f/10 min) ~
,7 +
(932*f/24 h) (EPR = 4 C/cm2) l lQ f 0.1 Hr. R
- 0.94 l D 8 ppm 0 ;2 sensittred at (1292*F/10 min)
+
l (842*F/257 h) (EPR = 15 C/cm ) 0 8 p;m 0 ; 2sensitized at (1292*f /10 min) WECEtiT l * (642*f /257 h) (EFA 15 C/ct') f gt;( f
- 0.1 Hr. R = 0.94 l 8 ppm 0 2
; seasttired at 1252*f/14 h UN 2
l (EPR
- 20 C/co ) f 0.008 Hz, R
- 0.95 ;
I X 8 p r 0 ;2sensitized .t 1292 r/i4 h ' I l ((PR = 20 C/cm ) f
- 0.08 Hr.
s . 0.95 s
- 8 1
0 10 20 30 40 50 60 70 - STRESS INTENSITY, K (ksidn.)
l 300 a 316 NG
,, **. . + .,+*
. e a impurity I 200 - '
. . * * * * * * *.* h b high purity U
_ ++* g%a c impurity
+
~
dHWC s- 100 - a l b lcld l e l r e high purity a. f constant load 0 , , , 0 2000 4000 6000 8000 TittE (h)
UNDER LOW-FREQUENCY CYCLIC LOADING IN SIMULATED BWR-QUALITY WATER AT 289 C 10 *- ^ 300-
^
f $ 200- __ WATER CHEMISTRY __ M 2 3 m 100- 6 02, H2 H SO 2 4 E o lv _
' -100- LOADING CONDITION I Lu l F 5 -2 IGSCC
~
<C R = 0.95 l1 LIJ -300-
-2 lI o _40o.
f = 8x10 Hz F 10,g- ' in g } g -500~ K,,x = 29-35 MPa m O . g -800-M -700- ., . , . ., . , , . SENSIT!ZATION__ gg 2 j y _ CONDUCTIVITY at 25 C (pS/cm) EPR, C/cm
- O j< .
304 2 i [ Eli316NG R3 0 j O
" ", ' I
! 1 0 "- . . . i 1,3 2,6 4 5 4
- WATER CHEMISTRY CONDITION
10" _ _ CURV _ 2 1inlyr NRR 0 - C 10-4 'e a _ 3% e
' /
, N6 -
An
/
8 e _ 2 -5 / N 10 c- / 7 C - O ! Y 0.2 ppm 0 2; sensitized at 1150*F/2 h W 2 C
~
(EPR - 15 C/cm ) . GE A 0.2 ppm 0 2; sensitized a t 1150*f/2 h - 2
- - (EPR = 10 C/cm ) - GE
- 0 0.2 ppm 0,; sensitized at tis 0 r/24 a t-
> . GE - Q 0.2 ppm 02 ; 5'" If 5'"5'ti2'd . 10 0 8 ppm 0,; sensitized .t iiS0 r/24 n C - GE (D HITACHI y ~ _ GENED O' ANL O GECR0 att HOrd 8 ppm2 0 3 5'"5'II#'d D# "'Id'"93 g 0.04 in./yr trs at ,32 r/24 = (sR ) - l @ 8 ppm 02; sensitized at (1292*F/10 min)'
- 2
+ (932*F/24 h) (EPR = 4 C/cm ) !
l O 8 ppm 0 2; sensitized at (1292*F/10 min) ~ '
~7 1 l + (932*F/24 h) (EPR = 4 C/cm )
2 lQ f = 0.i Hz, R = 0.94 ; j N 8 ppm 0 ;2 sensitized at (1292*F/10 min) 2 l + (842*F/257 h) (EPR
- 15 C/cm )
@ 8 ppm 0 2; sensitized at (1292*F/10 min) g7 l l * (842*f/257 h) (EPR = 15 C/cm )
f = 0.1 Hr. R = 0.94 Aq l l $ 8 ppm 0 ;2 sensitized at 1292*F/14 h DATA 2 l (EPR 20 C/cm ) f 0.008 Hz, R
- 0.95 l
X 8 pp 0,; sensitized at 1292 r/14 n l (EPR
- 20 C/cm ) f 0.08 Hr.
9 0.95 s 8 l I O 10 20 30 40 50 60 70 STRESS INTENSITY,K (ksiE)
/
. I Crack Propagation Rates in Type 304 SS Specimens (Heat No. 10285) Sensitized to Two Different Levels and Tested in 289'C Water with 8 ppm 0 2 f(Hz) R K (MPa.m ) &(m/s) d*
Specimen 300-C-ll, EPR = 1.4 C/cm
~
0 1 33-34 1.2 x 10_ IG 0 1 36-37 2.9 x 10
-10 0 1 37-38 4.5 x 10 1 x 10-3_3 -10 0.5 31-32 2.6 x 10
-10 2 x 10 -3 0.5 30-31 8.9 x 10
~9 2 x 10
-3 0.5 30-33 3.4 x 10
-10 2 x 10 0.6 32-33 6.6 x 10 0
2 x 10_3 0.7 30-31 3.4x10[y 2 x 10,3 0.7 32-33 5.9 x 10 2 x 10 0.79 -10 31-32 5.5 x 10 2 x 10 -3 0.79 34-36 5.4 x 10
-10
~3 -10 2 x 10 0.8 29-32 7.4 x 10
~3 2 x 10 ~ 0.8 30-31
-10 4.4 x 10
-10 1 x 10 -1 0.94 30-31 3.1 x 10 1 x 10 0.94 -10 31-32 1.9 x 10 -0 2 x 10 ~3 0.95 35-36 1.7 x 10 2 x 10
~
0.95 -10 36-37 1.5 x 10 -10 2 x 10 ~3 0.95 38-39 2.0 x 10 y Specimen 304-C-17. EPR = 1.8 C/cm 0 32-33
-10 1 2.2 x 10 IG 2 x 10 -3
~9
-3 0.5 30-32 2.8 x 10 TG 2 x 10 0.6 28-29 -10 5.6 x 10 1 x 10 ~1 0.94 30
-10 y 2.1 x 10
*IG = intergranular, TG = transgranular.
l t
4 4,30 _ N N M1MN
+
. 01 4 2.15 > .445 --> 4--- --> 4- - .500
+
- .01 + 005 0 0 0 $ 1 0 0 0 0 0 k 8 8 8 8 8 8 I I 8 8 0 0 t i I I 8 0 0 0 3 0 3 3 8 I O 8 8 8 8 8 I
8 I 4 I 0 I 0 8 8 8 8 jy i 8 8 8 0 I I I I 9
-+@
I e I f 8 0 0 t 8 i O 9 8 8 I I 4 I I e I i i I i 4 I I I t i I 9 0 0 8 i e , , 32 I e , e i e i e a i e a e i I t i I I 8 1 >@(l 8 8 If
'DRLL l U 2 PLACESI 8
- rea DRu ' l p --* s50 4- l
' 7 i N -
ANNNNNNN' . l ,, , ,, xx,", ,' .". ^ 88 8 It I II I II $8 8 11 gg
,, . .. +u 48 I it 8 18 8 II il I 11 ]f I l l l e
e s 1 8 8 M ATCH M ARK EACH PECE 1/2 X 20 T AP l SEE NOTE A. DRLt. APO TAP e 2 PL ACES l C I e 8 NOTES: 4 R g A. SERIALEE FD(TURES AND M ATCH M ARK E ACH PECE EXAMPLE: Y-1/Y-1, V2/Y2, , , , Y-16/V-16. SEE NOTE C B. M ATERI AL: 316 ST ANLESS STEEL TO DE SUPPLED BY YENDOR. C. USE COST EFFECTlYE MACHMNO tTTHODS TO MATCH RADUS AT ENDS. ARGONNE NATIONAL LAB INFORMATION : I: CB8 TEST FIXTURE TECmlCAL CONT ACT: Y.F. BURKE l ITY: 16 312-972-5135 TN NUTTER: h: DRAVNO nut @ER: 1~ 1906-1-30 Y. F. Dt.RKE DRAYNO BY:
A - . L- i+ , - - - a-4 8 e I
< 3 0 *
, 25/32 3/16 R i e-5/8 + 4 1 3 /4 ---* 4-- 1 1/4 ---e r_..... U s.....a -- g
> j
- 7/16
<- ----- e n ,---- , __, _ (O'f - -
v TWO HOLES 0.033 DIA X 0.125 DEEP
.250 DIA CREVICE CERT SPECIMEN i
e l i 1 4 4 i i
. - _ . - - , - - - - - .--- ,-.- n ,,,,-, . ..,--.,----.-nr----- , , - , , -,,,,,_n .., --- . , ._,_,..._,,,,,,...,,...,,-,,,,.,----,,n,,--,.,,,.- . , - . . - - .
f 1 } Follow on Work Better Definition of Conditions for Crack Growth in Type l 316NG i
- Threshold effects i
- R ratio Crack tip strain rates 4
l Frequency impurity effects
'1 Comparison with TP347 and CF3 i . :
I Studies on Crack initiation Effect of surface cold work i ! Weld Cold Work Defected specimens j Threshold levels of impurities I for initiation i i ! Transient water chemistries; l during intrusion for long-term l damage i Pipe Tests i l i
Fabrication of Alternative Materials Initial problems with microcracking in 316NG during hot ! bending operations 4 Type 316 and 316NG are more susceptible to weld hot i cracking than conventional Type 304 SS; adequacy of hot j ductility for high restraint welding characteristic of nuclear ! piping systems Prior experience with welding of Type 347 SS leads to substantial industry reluctance to use as a replacement material i Experience has shown weld toughness can be quite different from base material; lew welding processes and novei configurations (Types 347 to 304 SS) need to be examined ' i / i t 4
. _ _ _ - . . _ _ . ~ _ _ _ _ . .
t 2 l l
- -Q
- I -
O '
~
D 2 s $ U 5.0 x 40 x 350 mm x 5 70A-7.5 cm/ min GTA WELDING
] 4 O304 A A L W-C 316 s %3
- C OMggC\N -b E \ O 3 2 COMMERCIAL 304 E
< -T
> I 1
O O 4I 2 O 3 4 5 AUGMENTED STRAIN (%) 4
s 1
- O CRACK SENSITIVE G MARGINALLY CRACK SENSITIVE O CRACK INSENSITIVE O.14 0 O.12 0 PRIMARY AUSTENITIC ' I O
} O SOLIDIFICATION 'O I
] - O O. LOO i
- O C 0
CRACKING 0 gC, I o [ O.080 l O.060 o oo G"o o ig NO CRACKING O O O 1 q O O g Oe O O O g O O.020 _ O ll I.2 1,3 l4 1.5 l.6 l,7 l.8 Creq/Nieq O 1 e w< , c-_, __
e e 37
% 9 m n N e w e e o e m4 4 e D e m 4 m e 9 w C e N N 4 e e m em m
- e. m. e. e. m. e. o. m. . . N. . . m.
w E e. m. e. m. . m = m m mm m a = = = m a O m = = = m = m a e m m e m N e e 4 m e
- N 4 e m em e m e e e F @ 4 m.
@ 4
- e. e. M. e. A. a. M. e. d. e. N.
w e. N. . m. m N m m 4 m e m N 4
- 4. m 4 e
- Om N m
- C C. m o m m o m m m n E m m m m m m m m m m n m m e N O N
- C e a e e N e 4 m m e e e N 4 N w w e 4 e. e. e. N.
4 Q. o. N. e. e. e. N. w e. m. e. m. e. . m o e C. C. C O C C w e mm e e em mem e e e m e m m m C O e N N m N N m N N N N N N Q m 4 m 4 m m e m N m 4 e N e e N w w N 4 4 W C. C. Q. C. Q. Q. Q. C. C. C. O. C. C. C. Q. Q. Q. Q. Q. C. O. 4 COCCOOCQQQCOOOOCCCCCC e e d e a m N m e e e m m D e e d e m e e 6 Q Q O O. Q. Q. O. Q. O. C. Q. Q. Q. C. O. Q. O= O. Q. m. . Q.
- E O. .
e Q Q Q C O Q Q Q Q Q O C O C C Q Q Q Q C C H.
$ m m * * *
- A m C m o N O 4 m ~ O O O O C w $ Q C. C. C. C. O. C. Q . C. C. C. C. O. C.
3 Q. Q. Q. Q. Q. C. C. e O C C O. Q Q Q Q Q Q Q C C C Q. C Q Q C C O c 0 m e m e 4 C 4 N N O e e e N N e = e C e m N 4 Q m m e o m. o. m. m. m. N. O. m. C. =. N. . m. . m. . N.
- m. e. e.
O I N N N N N N N N N N am Q C C O Q Q Q Q Q Q C E t Q w Um e e e a m w e e C e # N 4 e e N e 4 4 C e 4 n e e w a e. e. e. ~. N. e. m
. m. m.
4
- o. N.
m m a
. a. m.
e a O m e e w
. m. e.
w
.Q N
. m.
N
=w E e e e C e oe C C C em e m m m m mm m m m m m m Em e c $m W e o e e e 4 4 m O m O e 9m w C 4 4 C e e C e C 4 w w c w w.
4 4 N. 4
- e. e. e. m.
4
- m. N. N. . w. n. e. m. n. N.
g4
- y e e e e e e mm e e e e 4 e m 4. m m m m m m mm m m m m m m m m m m m m m m em m m o e
.O. 9 !
w c e e m a - N m m m m m =
- c = N N N m = m - m m 4 Q Q Q Q Q Q C. C. o. O. ,
e e O. C. O. Q. . . . . C. C. . Q. . O. Q. C. C. O O O Q O Q Q O Q Q Q Q O O Q Q C C C C Q \ e
- E
% e g l:
n a r wh m M e w w N N N N N N N m e e e C m e C m m e Q uk Q. Q O. C. O. O. O. C. g O. Q. Q. Q. Q. Q. Q. O. O. Q. Q. C. . O O C C Q C O C C C. O C O C O O C C Q Q Q 4 m e m m m e m N e m m 4 e m e e m
- m N N N e m 4 m e c N
- e. N. m. e. m. o. s. e. m. m. m. e. e.
m g e. e. . . N. . .
$ m * = =
- m Q ~ * * ~ ~ = = m * ~
- m ~
- e H
H m e a m a C e N 4 e m m e e m e m 4 N e 4 m 4 4 4 4 4 n e. @. e. e. e. e. e. e. e. e. m. e. m. e. m. O O C O O Q O C Q Q Q Q Q C C C C C C O O m e e m + 4 e N N m
- 4 m m N N mm N m e Q t3 0 C. O.
y . Q. Q. Q. O. O. O. O. O. O. Q. Q. O. C. Q. Q. C. , O O O O O Q Q C C O O O O C O O O O C C O l
,i h 0 0 0 0 0 0 5 J J O J Z z z J J z J z z
- 4 l
= 4
- w w 4 4 4 4 o 4 4 4 4 4 4 m
q Q o O O o o O C C C C A -m m m e m m m m * * *
- 4 mm m m m m m m m m m m m o m m m m m = -
{ l 1 o e m 4 4 e m e o o mN e mem e 4 C 4 N e m e m e a e e N o 4 e C m e g
. m mm e 4 4 m N C O 4 m o e a o oo C0 4 4 i g a w N e e e m N m e mN o a N e m
- e g w e m 4 e e e e 4 m @ m m e N M P > N e 6 E
- e 4 4 4 4 *
- Q 9
- O O O O C w w m w w
't I
I
l l t l l l l l l I
~
i.' fs .! \ /. A;
.'Y 3;5'% K*,t.. . ,.f "f . f.t'%..
l _.? .- : 2+ if .j, j ': '
? *,j i '; y, . :,;}. .*
l 1 l,' .:* .-5 4,% .. ; .-
' ~
l 4 .
.x , r I
~~
./, . \, . \ 't .A ** .! [jh,\ {.'..:-- .
,\
~ " 3 * , , *~ ~ * ;s
' l ': '*w.T i ') %
- o
.e.. x '
' 'f.o. r-l j . 3 ;. . i -
f i .1, Y . U :. ::y .- -J ~x
~. g. g ' .
c Q, l 'i-l, j\ , '.: Ng-
.. + .t,
\.'. _y' ,
f / j. 3 .- . . -- *
~'f
~: ?,
;.;.: "y. ;
p
' ,c ... c- ,
\; , % . .t
. ,,4 ? , (;.' ';j
= . .\ '. \![
s - 'i~ - - . - t \ ;. ,, CRACKS Ar AND NEM WELD FUSION LINE OF TYPE' 34 7 SS WELD.'!ENT INNER SChEACE. 218 X il i l l l ] l l
/
1 e l I ( .I i j l CRACRS AT AND NEAR WF.I.D FUSION l.INE OF TYPE 3'+7 SS WELD.'!CNT 1 INNER SURFACE 433 X l l l 1
Cooperative Efforts EPRI NDE Center Develoment of welding procedures for TP 347 New York Power Authority 347-347 and 347-304 weldments using German and US (EPRI) procedures University of Tennessee Group sponsored program on hot ductility and hot cracking of austenitic stainless steels l
New York Power Authority Weldments Three KWU Weldments using German Narrow Gap Process Two 347-347 weldments One 347-304 with 347 internal cladding Three WSI Weldments One 347-347 weldment using G6rmany Narrow Gap Process One 347-347 weldment using EPRI Consumable Insert Process One 347-304 with 347 internal cladding using EPRI Consumable Insert Process - All six weldments appear to be good Radiography (NYPA) Dye penetrant and UT tests at ANL; metallography being perform 4ed Residual stress measurements on narrow gap weld Fracture toughness tests on narrow gap weld and 347-304 welds Stress corrosion tests on welds and base material
ANALYSIS OF FIELD COMPONENTS WITH OVERLAYS Two 12-in. Pipe-to-Elbow Weldments Two 22-in. Pipe-to-Endcap Overlays a ISI Reported "360' Intermittent" Indications One 12-in. and One 22-in. Weldment Completely Free of Cracking a Localized, Numerous, Short (1-3 cm), Axial and Circumferential Cracks (up to 60% Throughwall) - Heaw Postweld Grinding on Endcaps a Most Cracks Occurred in Forged Components a Cracks Did Not Propagate during Application of Overlay or in-Service Afterwards
1
~ \ ., * *
'\k*'
.g.,fM IM.. jk'
- e. . p g. ..
,_ y +. -s + - ~, ,,,..<i{-
s ph 0.s. . . g s'( ~.? f , ... ' .' ? Y :f $'KM*A [3%,%*, *. *."' j' ,.. ' .e, b.- va
.:+.4.f% #f % g.y , ;g . / c .: 5 .g if.' . -, . , .
O ,
, , :g . n. _
gf ;.. .
*.m t ., m, w % , , f p ;4. .I..~ T~ ; - I,,
v . ', . .[.p, [...y..
'y *
- b,' _[,:f. ,. '![,;.-[ nj b .y*
h j.7
" f [h Q ' . .
l w Q [y. [Y_ (
.y W,
. . _ , . . . 4 : z q. e.,, y. ; ; <
M '
*:.ig. ! d.Q -n *Q W%$M ,y5g ,.. = '. . ;)~. 7 ' - . 4l* .3
.r
,% %. ~ g s h ' f* % Y ..
$.? . f g *- 2 :
sisA n _ ., ': K. ^ (%,.@..r c n
' & .. 2j.c w p g.,q g ?- b ,.f $ ' .'... :
- L ' ' - '^
. ' :. . .;.4 l
e ' f, . '
, d
.')}? ;.W l ; :)W.. . %% , Q' %.%g 3
.g }v b q .C;f. .: .f
. g. f n,.
*l ~
Q, *e .,$
^
- $ F %r .
Wf^f_.%
, QWQ _QP; , . ,, ;l%g...y^f . ', ~k j f,, ?. 4 4"
- j, " ' _. ;, .
l I f yp,i:P. ;
- 4. In Q.l*;h[ * '- '. -k(
.i .4 :
..- ,d " D *
~. '
'S * ' *w.% L
['J f " g' c, . f7.T :..:s' % . g,t w . : c.:% . l >' ,, ' . h:,. ,.k.l, d$ y,4 h. gR
, 3
. . 1M . g. ~9 w . & ' . n - # 9' A ::. .., , J =Q rw, f Tf- hl f ; k Y '{ rl ' . . .,
~ '
,b -f!-s". b', .
.h'h '
. . b .
,1 . .% .) . f . ;.' W.*.. [ .'
e d .M' ...; . '
- s. f "" '
.v ? O f.
$.K.% ~ f $ .l.((lWc g.z
. E .
. - c < :.[ "$.~. ;sD Ndj} * ,.['j ,
. . =h. ?. ,
S s
,f
, i
-7 .* 4 .-. . .
- h .G.,;;'., .f. r **
Q.'*- f '4 '4., ; n .w., (.- -[~E:( e*k, ? e y :. . c. A m dy...a.. Ll:y *
'.neh'a %~ ' y & a,,%.qq; % g,**y %*:.'
,'*f;
' . i .: V 9 : ; ? ',' ,4 ' - y, QMQ .. Q(. .
_ :;-~- *. 4; W js.,&y p g ;,.-w m~ o. ;-, _, . '. ; < g .: ,1 :
'."y 'g
- V a.%. . g . y,. D(,. < :. . ?. y.4 e% .. . . , ,,....W. %%. .'Q d. qs y ; - < "Q
$e . + ; % y- '
.. .W w O. +&_ 1.; J(. .Y::: i... .f?.l
&. J p. ' . . . A i 'J %.
2 v -{y.,. 2,. Q J.%;y. 7.W %
- y s'.. . En? .'
~
*n'*
- f. l,
.[
t g f.3 M
- f + .' ,
_ e.7% 2 .. g. % y. . . '4.,.
>, '?+ . .*. m: 't h..,
' _. i f '
. ,0 Q .y. *f. ..-
k[;[QJ:fQ y '. Q%% .h~ * , >f. )IM O 2 i'
?-7
.. , 9, ; Q
'i,7 [ ( ,4 . M(
. - .", *s' . -
W *
'Q.9Ih.):",Y..S^'=g.%p:'.{'j3-Qb[.;j,..h.
b . . : . f. . ' .' c % ,s
? ".?** * &L .i ;. - ' J ' (: [k - ^ ~ .
1 I'- fI j
' y I : $
k
.'y .d
[. "
- e. ;j AI S.[. ' :l h hd ; *b f, * , ., , 1 {
. ;.:. a
. gW w . . , - ,,8,7 M.C*i..N. . ' 1.2. yg M- k;
,g w.
, 3.g. ....p c. v **.,
;.. ..w _ $.
/ s +
r +.. ." ,*..- ' y ... .
? .,. .- [; $. 4;?yl 'yp:,, .. ?.y~,
A.,' ~~% (e ._c p,:1.. ..r r% {-
?; . ..: h ,.'
s' m.. .g y .; 4 ,s w&,J g 6 ... #
*F.
- . o . ~ .
- ,,.'. ',-, "'{j ,
j'.,,..l ,g- , .c; o i s - ,1 . 'J "N . 4, ;,% .W .' ' , ... '
%..f.%.-
.- . ,,,e s
.m.,.
t 7j . . , ., g
.y M4
. .- . 3 s '" . ;-p - . . !m. {.' . *._ ,
Qh. ..s : M.?. . -~' & -
. " -., x'y -
,. ?: - ;- : a? ,,.4. - e* o.
f . ly' ' . ,~ - . '< _ l *,l . .
s k.~T.
$. $ "~w
' n ,l,'
, .. r- '.*d**$~* , .,' . -f '{ V
~ - :1 V. - f y. f% '
) .-
R f p f . '~ l*l, *mp - e- 'Y~ .~ .
,.R .
.soe
- b. *e; l.Nl ' . .. 'i ,' f ; ~ l *-% j E ? ' ' S .. ~
". l '* ' ,fh
,1
-
- g) .y;'I.- .
,.., , y ...- ; [' -3 1"
- ' y.kg ,, k *
*} y *..'.~
A%
- , , y.
~n.. '
t- .- - p . ' * }
- s. &;i 'f ' ; g ' - _ ' ' '
M '
. ,_ 4
'( . g ' j 9 } ' '
y * ' I
.'J
, , [_ . 1 '# '. . . ' , - .
s
=-
.s a* ' .( ~
[, . .[ ,
. 14
_ _ *M%- ..r.* ~ . ?. 4.**.
, - d % '.
*C
' *~
E-
-t'* ,
...k 9.
,yp/-
.. +
[ .3 - y , y__
.f /
.."< ' g, wj f5 p .
- e..*,
.f k _,(. W *, #
; )3, ,, [ . ; ; macap r- -
f' * ". ~3'
. [. . _- '
.. l '-
f st
- .. . - - ] y , .. .. " . ..
_v .
- ;,, gg.,..,, ., N J . O ,s
., .. A
.,4 ,
', .. 4 .
g, y ,. .. - y
.". ' , * . 9
. * [ $ a '
- g. .. L r,
I
.
- _' il
. , . - J--
' , 'l [c, , ..,, *
.w ~_ '- _. , l '^. l ~
. * ~ ,_ ,.% ' f,
^
. . . Y' ' ' ~
- ,a
[ - ,g9' ' .~' , % , f . -
, p. % . . . ~ . g . -"
/ 8" - 'r c. s-
,.. - . a -
,. , , , , , - 2 .**
/l . . ,, ,
*y , '
T:
* ~
e A f *' . o
}
e 8 #~ 1 egi y ;i - ' , ' . ' ,- . - . .. '. W
. ' ., *- \ . . 6
~
}. .-.t.-.
y -o,,,,m.u f' I n.. g , 4 s.
.4!.
l' -r 4, ) . . . ., . .
- . y i fs. ,.
s s .
.e * : ,
., =r y
* .- ' ' '* 4.-
..,' . k '
l' {* c ap _ -* y an,.
- - - - ~ - -
' '~
~ ~~
*EM
* 'n
,a n. . . ' _
hN pa,w. .. m .
- mw'wreair 4 a. .
y . ' 4 %MyWQW;.
.. n
. .: #*w .'w : . . V:
- r. .g .~ v nn....
.s ;v'
.g- d .
y ., y y;,. 4, . .; .. gr .#.;b, :q'. v..X .
=
$p,. s.v,_ w . ..
..a.-
S.,
. 1 .
- m. . .
I .* * - '- g-[ f, e .g.r s y..p.m.g i mp ..g.g.A
.r
.;:ws W.
g . s 'y>; ,
,., g.,
n e stq.k. ~ .:...'
..rF ., Q *,.
s
'gg . , 3 * -*
._y
.$t' Au%.g ._
- .. z. - '.f
.4 *
.. e .
@ = . . . . x i.eo a-.'.>pe[orf
.I, $ r Y 1. ",- ,.", . %. 6. 4 ..
> . 4+ . e,--e .
_,. . . ' A . * ~. 4 . c . - .
,[-..' .{
.h
. h. ' . a 4
.y s - 3 U-w "g\ . ( ms.,,. y l . .i'.~. g,}( L,N H
- g. ., .
-s. a
., .< s .
.: %. . e w ,
. a 1.
. 3g
(
.s.
,,W } E.. '
.. . x
}.. ..c. g ....,.,s q.. c. : : :: ; 4,. .
.7 7, c
.. . m-. - :: . -
. .v;... w .w..: . . . , . . . . . . . 3 - .R
- <r:'.y. . * ,'., f, F%
.s tg Dp :m, , f a. . -
~
k's x,.*.[;./ $ 5,%l [p. , , ..^ .-. . -5 .v.
'o -. N S u,% . ; ;. . -j; 'q ; 4 7-: w' m yg.}, .
L' ' 6. m.,; ' T,,. t ..t ,,'e.. . . g ~~ *h.. , ,% _- . ;j'lx e -. , j l}
.. ,p
#'. ::: .g._ 1
...5 .x . t*. , -.
3 .e.s 4
.s A, ws.-
n- .
.,c em W%g GQ&Q$hW$k }N
..- 4
, n
?y,A$m e We W k p@ h W '. 4'.'
f ?- [.@4..q. o.R %g&nQqW Mgrl4W
% W.MQ .
m.,, . , > g.
& Q.Qyg%q g,4g .c._l C f Q M.Q Q f. W s.q T
- -: .< gs a. w s_ : ; ..
- m. ,
, . s - .
4
-Y '. '.J. .- j' 'je,eE*f M . -- a $ j e'
.; . . ;g - ..
K[ 7* 4 g.: .,N
- a s . ' a'ls , ...
f n,.~$*, . g.
,A y e f e ./ . p- -
- q. ;.~,W.'",,,,, r- W..,
.; ; '--t,. e 4.
A 4p;; ', ' . y, r3. A9&' s
' ' /.v
%ry.~ .. .%.u;nc.N:
4h . m s .:,; .,.a .J .= _.. ';,b,.4 .
- h w.? 4 , ' _w Q,. .Ca ..
,.i" ; &
ppn -
.y_ ys -
^
T' Q%",M i M Mp"*yWMRsW. 2. - W K.^% a L < ;; ); J
- c. a wy,4.. ..
< gysusv- ..-n%;
c g6 .
. . ,_ _a:c. s. ; s %wv.gw ~;<
.- g, w
,.s~ .: :
. v V O. c:A*
j s .n. g j' L;",* _ _,,-L 7
., :"Qp ,
Q;)
- y. ?
, ~ .e , . \ .. y .-
4, g , s7*<.4;. cz g. y. g :;,.';~_'}i.
; .- ... .. ;n
.. z:. .
t . eW, ,
- 'v
. ., * .. , . : , ~ -
e{ :. .-m;-x
- : . g- - -
- n
- g. .L - -
4 -et -- # 9= _ _~
. _ yg _ ,
4::: ~ s . ..&. . ~ - , , tw L,w -
- m. m i; f . . .. L.'j?:,
, . , , . -((.<. , . .- sJ
.ILi'- '.s:i -
, ;s , :s u 1 y,
. .? ([ (y '. Tr.
..g_
.-( . 4 s' s.
.: .. .-- 1. <r -
.n
' ' ^
,.+
f .;f.,.yf...
- g. : :
.n..
L
,. j{f:.s.//,. -
g x m. *~ ; ; c ; v > ...1 r > . . . L-
. ., / -
Q . . . . . f ..
& g ;
s .. ,
.. a .= -
b . v..-
. , . ** . as- ~ t ,.
g.
. . s -. ,
.. ~ *
.., T..: . .
..( , .-
j3 -
. . . _, ' - . \- * .. , .
_ ._. . ~. ? d _ ~ 9 . 4 .,
- ,e r '
'. . '.,'E .' " ,;-
, ,4 w- Y. ..i ; ,.
$ $ ," .' _.'. . ' .E 3 < . t, i ( .;
.: _ < < '; . ! -!!'. s x, * ' '
,f 3 > *. . . ; -' .'
-. w .y..
~-
. :- : . L l ' ) . ,' ' [' . ,
t^
, , f. [ [ a
;y 4 ..
p
- e. .w. . _ . .
. .. .n.;,t. .. , ;
~
* 'F V
4eN 'Q. ' ._s ~. . - . . . .. 1 9 , f
'p [t* .. A - .m ..
. ~' .y.m. . .., ..
, s-h ., , . < s ,
s
' .-. ' , ; 7; ( ' [.
3 . . - '- 4 s, < 3 Jh. . ,[9 g . g
. m.s , . . . . * :_ e . . :,', . - . .,
F u a. . . . :: .- ..
m a, . ., 2y. . . . e - j
-o
- ~
- ... ;J.,
. .m , -
+ , ,
.:.. . .N.\~'.z: , , )- .
N.((.$.g p e
- . . ': %,g
,. a.- .
. r. ,
.. .f
;Y _ , ,
(4 g ., , ~~
- -1 *
- 4,
+ a. n> > .t 4' -
1 . .. g . 4.s .r. . -
. t J
~
.w: ,.
?
- :.n.
v, . . .
.. i.
... . K; .,.d; , .. ~a,
.* ' W. .,.'r
, , . .4,.' y. ==
i. pf.4 ',./ <
-2
,3, . M g:- w d. . . -., --
?p. .s .t ,. 7.,&. b..,4.;G.dd9Lfm
, . . ~.-
3 . .
- x. .) %515 CRACK TIP OF A'1 AXIAL CRACK IN HATCH-2 B'n'R WELDMENT 2B31-IRC-12BR-C2 1
l l
. e 8
I I I I I t . PIPE SIDE 6 - ELBOW SIDE _
/
A
- - \
N / l
\ i 'E / \
o / %\ 1N / M4
*A \
\
E \ Q-
\
W 2 \
* \ -
\
+ \ .
A s l I I I 'M O I 2 3 4 5 6 DISTANCE FROM WFL ( mm) THROUGHWALL EPR MEASUREMENTS OF HATCH-2 .! OVERLAY WELDMENT 2B31-IRC-12BR-C2 AT i SITES WELL REMOVED FROM ORIGINAL WELD
i 1 INNER SURFACE i PIPE SIDE CRACK ELBOW SIDE i WELD
-OUTER SURFACE
? **
5 4 J MAPPING OF AN AXIAL CRACK IN HATCH-2 BWR WELDMENT 2B31-1RC-12BR-C2 4 1 1 ( 1 l i
l l 4
- I I I I I I I l I
] 0.3 - -
}
l 5 '
,Y\
5 I N 5h h 0.2 / - 5 < I y g 0.1 g 1 0
-4
-0
-6 -5 -3 -2 -1 0 1 2 CIRCUMFERENTIAL LOCATION ON OUTER SURFACE (in.) l Crack Depth vs Circumferential Location.
4
*\
s. Psc s i . Undercut at Weld Fusion Line of the End Cap. e
Measured Residual Stresses on a Hatch 12-in. Elbow-to-Pipe Overlay Weldment 1 Azimuth 0 90 180 270 l ) Inner Surface ! Pipe Side -15 -10 -21 -15 Axial Stress (ksi) ! -37 -27 -66 -45 Hoop Stress (ksi) I Elbow Side -31 -11 -60 Axial Stress (ksi) , -42 -29 -55 Hoop Stress (ksi) Outer Surface l Pipe Side 25 32 . Axial Stress (ksi) 5 10 Hoop Stress (ksi) j Elbow Side 0 Axial Stress (ksi) i
-15 Hoop Stress (ksi) '
l l i i
Measured Residual Stresses on the inner Surface of a Hatch 22-in. Pipe-Endcap Overlay Weldments
- Axlel Stress Hoop Stress i Gage Position (ksi) (ksi)
I l 40cm-pipe -30 -50 { 60cm-cap -22 -43 80 cm-pipe -23 -59 95cm-cap -20 -43 120cm-cap -22 -46 175 cm-cap -24 -82 I l i ._ _ _- _ _ _ .__ _ _ _ _
t . WINIOVE RL AY 10 2.4 mm S2mm i h (0) 25 95mm 7mm HATCH-2 PREP HATCH-l PREP E 118 mm 58mm
$*V/////hl (b) h 95mm
&YY 7mm
, HATCH-2 PREP HATCH-1 PREP L AST PASS 30.2mm 32mm HE AT SIN K WEL_0_ j j
{
~s\
,C, v 9. %, , -
HATCH-2 PREP HATCH-1 PREP
?
/,
C i i i i. p i. p i l i l l
-c-- Mato* *anmur azmum -
cacuwearm. *. eam.sineers - e ._ _ x as-
)tt g. . 225'
-sc !",
- 3'S'
..p i
-4c
= a l gmy e
\ /: -
- -eO - -
j
-7o _ \ -
- u. ,/
/ -
/ #
J - _,oo i lil i i l i l i l i
-60 -40 -20 0 20 40 60 80 AXIAL POSITDI (seen) l l'1' i l' l'l 'l'l -
a e,r u em m umeveur - E g e ce.cres ame _" O Pcsim ,,
?
@@@g @@gg g @35~
3e -40 3 135' 45" l
- ~ '
i.a 3:S* 315' 22S'
~ ~
SHORT PRfP LONG PREP
-so - -
l t l t !I fiff l l l t ! 60 40 20 ( 20 40 60 80 AXtAL PO5 mon (sewn) . l l l l l l
6 I i 4 SHORT WELO LONG WELD ' PREP GEOMETRY PREP GEOEETRY e
- DISTANCE FROM WELD CENTERLINE (mm) 70 60 50 40 30 20 10 1 60203040506070 I I I I I I I I I I I I I I O AXIAL STRESS .
GAGE POSITION X CIRCUMFERENTIAL -20 - 7 STRESS 2
-30 - /
1 . 4 5 3 m 2 i
$1 0 -
W - G . 80
-90 -
e 4 i
-s
, l !' 1 !li
'o I
0 6 I 0 4 0 I 0 2 I ) i s k ( 9 8 7 6 5 4 3 2 0 S O. I l O O O O O O O S 0 E R I IFn.3o h 0 T S 2
- L A
I I X I A _ P N 0 SI O O 4 DI T TS I SO H P n"d CE 6 TG - AA HG i 0 ' O 8 1 ' ,:,I;! i j'.l ;i ;i '
. ; ! !i? i
. i
. I l
I.0 - O 9 0
- 0.9 HATCH STD CCD 0 cb o O.8 oO O CD D E 000000 g 0.7 O OQ O O D 9g Om O CQ3 CD
.5 00dD
@O O to cnod
@ %.3 OCb CO O O O.2 O C O b CEO i O.1
-80 -60 4 20 0 20 40 60 AXIAL STRESS (ksi) '
, - - - - -, , . . _ . ,.-,y . . _ _ . _ _ _ , . - _ ~ . . . . . _ , , - , ,_m . .
Weld overlay 50% Throughwall Crack 2.0 l l 1 l l l l l 50 -1 l ll l I l lil l l i l H
~
CRACK OPENING -
- CRACK PLANE NORMAL -
25 - STRESS - 71.5 -
= e.
E _ d O o 8 - l 1. 0 - 9 ksi y -25 - o -
~
' ~
x 4 -50 - N O. 5 - - R - 6 _ _
-75 -
9 ksi _
' I I ' ' ' ' ' ' I I O -100 I I O O.I O.2 0.3 0.9 0.5 O O.1 0.20.3 0.4 0.5 0.60.7 0.8 OlSTANCE FROM INNER SURFACE (In.) DISTANCE FROM INNER SURFACE (in.)
G l .
8
~ - _ o HN) l
- w I
7
)
i ( n. l ' oE C l l ( e r u s I
- s. A oR F
U s 5 S l e ! L r oRE 1 A M y P N l R I 4 N O i 0 I l N a ' M E k 3 O l N I R l A 5 1 ' 0 F e L E l P S ! 2 C KS 0 NA 3 CE ' T k AR 1 S c l RT ! I a CS 0 D r 1 C 8 l
-~ - ~ -- - - - o l
a 0 5 o 5 0 5 o 5 2 2 o h w g
-*5 -
7
- l u
r o =E gm lR* :; h T
% O 0 .
5 y a 5 l r - _ _ , e - ~ - 0 v o l G '
)
n d N I I ( l W e l N E 4 E P 0 C A O F 1 R K U C S l A 3 R I R C 0 E N 1 i ' N s I k l 5 ! 2 M 1 0 OR 8 F E C l I 1 N 0 AT 8 i S I D
- - ~ _ - ,
o o 3 2 l g1 z g xQ5 _
l i 20 i i g i l 50 ' ' '
- l l l -
l 3 8
- l l l -
10 - 40 -
- i
- 30 -
q N O , i . i - i .
.i .
i i 1 20 t , , c -
.E -10 ./ -
[ 10 - [/ s ~ s g 3-in. OVERLAY ~
.E
~
x -20 - o i i i
~
./
- W' ' '
, N .g - / ~
M~ -30 N - *~ -10 3-in. OVERL AY // - . 4-in. x
/7 N / _
-20 --
OVERL AY '
- gN -
i l l T- i l i
- -30 3 yERLAb '- -
-50 I i i i% -f. I i l 0 0.2 0.4 -40 i 0.6 0.8 1.0 O O. 2 0. 4 0. 6 O.8 1.0 a/t
! o/t l
1 i 4 4 I 9 4
citim., e ~ iii
- ... . . aa wy .
C < m-
.m. .e iir
- d Si .
WELD OVERLAY COMPACT TENSION SPECir1 ENS f I H- -- l
MSIP (Mechanical Stress improvement Process) Developed by O'Donnell and Associates Uses split ring tool somewhat similar to the Pipelock Split ring is installed, shimmed, and tightened to plastically deform the pipe; plastic strains 1-2% One sided application permits use with more complex geometries l No need for cooling on the inner surface -- simplifies scheduling, generally claimed to be cheaper and
, faster than IHSI Monotonic plastic flow with no large tensile loads on the inner surface unlike the reversed plastic flow associated with IHSI l
l 1 l e t
/
l
}
/
/ /
/ /
/ /
/ /
/ /
/ I1/4" /
\' o
\ s 21/4' s "
NC s j/
/ MSIP TOOL
\'
s ._ o \M
\
s s 5a s x \ w " 2
/
/ /
/ /
/ /
h /
/ /
/
/
= 0.6 TO O.7"
/ /
/ /
/ /
l N N N N N N N N N N N N N x x1 . 12" NOM. DIAM. e- y- - ,* c- - ----------y _-,-,.---__9 -e-m -..w,- - ,
12 in. dia MSIP Pipe-to-Pipe Weld Strain-gage residual stress measurements indicate strong compressive stresses on inner surface MgCl2test results at NDE Center consistent with strain-gage measurements with no reported dye penetrant i indications Metallography and dye penetrant tests at ANL aflet sectioning show shallow, short cracks in fillet regions of I counterbore and a longer, deeper crack in the seam weld directly under the tool Morphology of cracks is characteristic of MgCl 2 cracking Current judgment is that MSIP did not cause cracks; local stress risers produced tensile regions that were not rendered compressive by MSIP 9
TOOL POSITION
^
r 2 OUTSIDE (a) r1 r1 r1 JL" >- a r1 s , . a sj e t ) Y2o y 20 y 20 y 20 ymyioy*ygy INSIDE - ~2 OUTSIDE @b@@ r1 A") r1 [ l Y= Y= 3 Y8 UY INSIDE --
-2 (c) i r%
i \'/ w mm W INSIDE @@ @b
-2 DIMENSIONS IN mm l
l l l
l l Maccured Residual Stresses for a 12-in. Pipe Weldment Treated by the Mechanical Stress Improvement Process (0* Azimuth) @cgo Distance from Axial Hoop cition Weld Centerline (in.) Stress (ksi) Stress (ksi) 1 -0.20 -33 -33 2 -0.59 -35 -23 3 -0.99 -26 -13 4 -1.77 -16 -15 I 5 -2.57 -13 -14 6 -3.35 -14 -15 7 -4.14 -17 -13 8 0.20 -31 -36 9 0.59 -44 -33 j besured Residual Stresses for a 12-in. Pipe Weldment Treated by the Mechanical Stress Improvement Process (90* Azimuth) dage Distanc'e from Axial Hoop ition Weld Centerline (in.) Stress (ksi) Stress (ksi) 1 -0.20 -30 -36 2 -0.59 -36 -30 3 -
-1.77 -17 -19 6 -2.57 -17 -21 3 0*.20 -38 -36 1 0.59 -51 -30
Measured Residual Stresses at the Lower Weldment %5 in. from the MSIP Tool (0* Azimuth) Gege Distance from Axial Hoop ~eition Weld Centerline (in.) Stress (ksi) Stress (ksi) 1 - 0.20 -5 -2 2 - 0.59 - 11 4 3 0.20 -7 -10 4 0.59 -5 -9 l Measured Residual Stresses at the Lower Weldment N5 in. from the MSIP Tool (90* Azimuth) i f ge Distance from Axial Hoop
- ition Weld Centerline (in.) ,
Stress (ksi) Stress (ksi) 1 - 0.20 - 12 -16 2 - 0.59 - 27 -13 3 0.20 - 9 -21
- 4 0.59 -
21 -20 0 0
. . - - - - . . . . - . . _ . . , . . . . - - ----..- . , ,, ,, ,,,,,,- - ,- _ .- -. ,,.- -,.. w, ,,,,.,,-n.---,,,.,..,.-,n. ..r-- - - , _ - , ,n, ,,, _ ~ . . ,, ..n,.,
~ a- - - .---,--,,...a , - -
k 1 O 4 O O D to N O g N O_ n O O - (1. -
~
2 +
- O O g m
m + m y - U) O N $ m _ z - M I o o N
- -Q 1 O -u)
O O o N 1 m I y 1 o
< O o
- 8,o i I
< l o o O c) CD N O o e to N - O I
-OOOOOOOOO H.Ld30
)
T O 4 O O c to N O N O O 4 O g _ O. 4ka . . cvi 2 - vv
- o O g m +
m , m w - oy E8 m - l m z Tz oO M O O E m
?
O O C N 1 m I y 1 O 4 O 8 o e in I I O tn O m m N e o e rn N - o
-ddddddddd H.Ld30
I o m - o , T l o
- 4 o m
o N o o o 0 CL - 2 'E
- o ox m m m 4 w - .
90e
$o m o e
Z i F i l T o m 9 N F I o m o N o a o i to I l U o
<r e o 8 v m >i j 1
o 0 9 5 N o o t m N - cy
-000666666 i Hid30 l i
l
, , . - , . , , , . - . - . - - - . - _._-,,n,. -,,_._,-_,,.,__,_,_n_,-- - _ , ,,,_,,,,-,n,-_.,..._-,
REGION OF SMALLER CRACKS A
\= -32.8 =
SEEN AT 34.7,36.9,45.1,47.8
- 25.1 =
fW m
%... \
t I i t 16 i, o ! -9. l-i
=
MSIP PROCESS ~57 = 11 0 ) 1
- DIMENSIONS IN mm l,
I
28 in. dia MSIP Pipe-to-Pipe Weld Weldment received maximum plastic deformation permitted under MSIP specification (2%) Dye penetrant examination before and after sectioning reveal no cracks; metallographic examination confirms only observable dye penetrant indications are due to smearing of the metal by grinding Strain-gage residu:ii measurements indicate strongly compressive stresses on the inner surface; unlike other residual stress improvement techniques MSIP appears as effective on large diameter piping as on small l l 1
)
1 l l 28in MSIp INNER SURFACE STRESSES l 10 O m 0 - 3) m -10 - 5
=
w . s m _m .
- $4 Azrem } *e e g _m _
a e 0
< - a o + 150
-40 -
a 225 e a e,
-50 . , ,
f
-100 0 100 200 g l
l N W CENTER!.fE (mm) l 1 f l l t - _ .
I i l 28in MSIP INNER SURFACE STRESSES 20 a 3 5 E + 8 ' #
, 0 -
v E .
+
+
5 F
-20 -
, 4 E
.. . Amm O
O z e 0
* ~
es + 150 ma
. ,4 a 225 5
-60 . s . . . . .
-100 0 100 200 300 DISTANCE FROM YELD CENTERLfE (mm)
1
....o Future Work Ferritic Steels Work is just getting underway in response to a request from Regulatory Susceptibility to SCC and corrosion fatigue in high-purity oxygenated environments is well-established Impurities can also play a critical role in nonoxygenated environments characteristic of many applications of ferritic steels in nuclear applications 1
l
r Irradiation Induced Stress Corrosion Cracking Early work on cladding clearly establishes susceptibility; currently failures are being observed on components like drywell tubes Radiation has an effect on both material and environment Segregation to boundary under irradiation clearly established at higher temperatures; less direct evidence at LWR temperatures, but empirical results from Lacrosse show low P and S stainless steels perform much better as cladding Radiolysis introduces new cathodic reactions involving the reduction of relatively short-lived radiolytic species (e.g., OH,0 2 , H220) May also increase the rate of reduction reactions k
HOT CELL SHIELDING BACK PRESSURE P REGULATOR MIXED GASES - /, AE AU OCLAVE 2 __ / 11
~ ~
/
~~
I E' COBALT 60 GAMMA SOURCE' SOLUTIONr F4 ~ STORAGE '3 DRAIN T TANK
~
OP =
~~ _ _ _- , SOURCE y SHIELD i HIGH-PRESS. p WATER
--@ PUMP /
SAMPLING' e o SCHEMATIC OF GAMMA IRRADIATION SYSTEM g e i J e 4 W J
1 l ASSESSMENT OF LEAK DETECTION FOR NUCLEAR REACTORS l PRESENTED TO THE ACRS JULY 2,1986 By David Kupperman Materials and Components Technology Division Argonne NationalI2ontory i ~ g
Current Practice Problems Existing Technology Motivation for and Evaluation of Acoustic Leak Detection Recommendations
Current Practice e Regulatory guide 1.45 recommends three methods be l ema oyed for lea < ce ec~ ion l . Suma flow (required)
- e Airborne par-iculate radioac"ivi"y (required) e Condensate flow rate or airborne gaseous racioac"ivi"y
- e Regula"ory guice 1.25 recommends icenti" led anc
~
- unicen"ified sources ae moni~ored seaara"e y to accuracy of ' ga/ min.
i I _ _ _ _ _ _ _
Current Practice e Review of technical specifications of 74 plants indicates the following
. All p ants use at least one of two required systems 4
. Sixty-six use sum 5 pump monitors
. Seventy-one use particulate monitors
. Sump aump moni"oring is "he primary method for detecting lea <s
. . Allowed limits on unicen"ified ea< age are 1 gal / min for al PWRs anc 5 ga/ min for most BWRs i i
CRACK LOCATION r E&KMcTze THERMAL SLEEVE g 1
- r-y+::+_f ' ~ ' ' ' '
% =a p= m \
c:: [IID IGSCC MEASURED CRACK DEPTH
-- ESTIMATED CRACK DEPTH DUANE ARNOLD RECIRCULATION-INLET-N0ZZLE SAFE END CONFIGURATION l
l
Motivation for Research on Leak Detection Methods for Nuclear Reactor Piping Summary of Recommendations for Research in NUREG-1061(v.1), " Report of the U. S. Nuclear Regulatory Commission Piping Review Committee"
" Improved leak detection systems would permit more stringent requirements on unidentified leakage without increasing the occurrence of spurious shutdowns due to relatively benign leakage, and their development should be pursued". l
... Improved leak detection systems would provide additional margin against leak-before-break without increasing the number of false alarms by providing l information about leak location and leak source i discrimination. It is therefore, recommended that improved leak detection systems under development be completed and field tested".
s
Other Active Experimental Programs for Reactors West Germany (KWU) France e g .
i . ). Problems e it is possible that large cracks may produce low leak rates (corrosion fouling or uniform growth of long crack before
- penetration).
e Duane-Arnold safe-end cracking incident suggests reliability ! and sensitivity of existing systems is not adequate for some Cases. ! e In most reactors surveillance periods are too long to detect l 1 gal / min in 1 hour. - l j e Tightening current leakage limits to improve sensitivity may l result in unnecessary shutdowns because of inability to
- identify leak sources.
t l . 4
un Potential Solutions e Imarove the sensitivity and reliability of existing sys" ems. e Install new and more effec"ive ea< detec' ion sys~ ems
. Acoustic Monitors e Moisture Sensitive Ta3e
{ . 4 Improvements in Leak l Detection Initiated by Electric Utilities l e Acoustic Monitors
. Unrepaired welcs t,at , ave crac< indications, welc
- overlays, and valves Endcap weld on 22-in. pipe manifo d (16 sensors)
Manifolc sweepo et (2 sensors on waveguides
- monitoring frequencies above 100 kHz)
Va ves including safety re ief valves using hig, tempera"ure aiezoelectric accelerometers direct y on aiae, monitoring frecuencies ess ",an 50 k-z l
m.__. . -. - . m~.- - _.._m.... . a. ,_ _ _ _ . - m . ._ _ _ . - - - - - i - i l Improvements in Leak Detection Initiated by Electric Utilities
! e koisture Sensitive Tape
. On piaes t,at lave weld overlays or on unre3 aired
- po~errially crac<ed pipes i
Ins ~alled in a" least four reactors j - Insta a" ions concen"ra"ed a" *e" pump risers, main j recirculating line and resicua 1 eat removal piping Leaking valve triggerec kST a arm in one case incicating need for better~ leak characterization
- "o preven" unnecessary squ" cown I
_ _. _ - _ - _ _ _ _ _ _ _ _ _ _ _ __ _ ___ _ _ - _ _ _ _ _ - - _ _ _ _ _ _ _ _ _ _ _ _ - _ . -_i
i i
- Exis;ing Technology t e Current leak detection systems do not provide optimal
! sensitivity, lea < ocation abii"y and f ow ra"e accuracy and are not necessarily sensitive enough to meet regu a"ory goa s. i e Acous"ic moni~oring "echniques can imarove ea< cetec" ion ' capabi'ity but existing systems provide no source discrimi-no" ion and no lea < ra"e information. . i
- O Mois"ure sensi"ive toae is a so in1eren"ly imi"ec in aroviding quantitative informa" ion.
i i l, i
There is an economic incentive to promote leak-before-break to other piping systems, particularly outside containment, where leak detectors may be needed.
-The main steam line, for example, would be a high priority system that is not currently monitored.
-Reliable leak detection systems could lead to acceptance of leak-before-break arguments.
Acoustic leak detection could offer a low cost margin of safety for seam welds in miles of uninspected carbon and stainless steel piping. r m
.---._...._.,,,y . m ,. , _ . .. -- , _-,
Argonne Program In-depth Review of Current Leak Detection Technology ) Employed in Reactors Laboratory and Field Evaluation of Advanced Leak Detection Technology i i l 1 a . a
1 Work has been initiated on validation of the reliability of existing leak detection systems. In depth study to establish actual practices in leak detection and capability of current leak detection systems. Response Time Sensitivity Reliability Establish facility to determine reliability of radiation monitors. x
j . Laboratory Studies at Argonne 4 e Laboratory established to assess adequacy of acoustic l methods to detect, locate and size leaks
- . Studies carried out with field induced cracks 1
! . Acoustic background data acquired from existing reactors
. Acoustic leak detection system installed by utility was
) reproduced by ANL and evaluated under laboratory i conditions 1 e Digital continuous acoustic monitoring system using j advanced signal processing technology developed for field application } e Laboratory facility was also used to evaluate moisture i ! sensitive tape
/
i i 10 0 I I il l l l l l l l l 2 FC #}/
/
/ _
L /
/
J - / -
< /
. 5 'd m iO IGSCC # I,2 8 3 /
/g' OTFC # 2 i $
o [ m@^' / 2 o AfD e
/
oo 1 o O l m af z
< y o,o p - - TFC # 1 _
l 1 1 I i I i I i 1 1 I l O.OOI O.01 0.1 1.0 i LEAK RATE (gol/ min) 4
, sr --6,n-- aJ- , + _n -u_m-- aa m-_- + - . . m 4-_J--_ p - J -+L-A -- ." - - - - ,--.---+AAa-4m a u TRANSDUCER SIGNAL (MICROVOLTS) _
0' 0 - O O O 8 I i i i i t ilj i i i i ilill l 1 i i I ill f~ $ N
- H m p a y' u -
a
. m xz- l mn *a 9d -
8 mE :
,O E .
- um 8rg:
x-co - n 2 O, , s1 -m i-2 a m
.. m\ o z -
r - g 3 Z w
-O -
m -
- - e* a N
i M o - j _ '\ - b . O I 1 1
1 l
- i$
l 3 Os
! -12 0 6
525 F (274 C) i 1100 psi (7.6 MPa) - i
-13 0 IGSCC #I: 0.005 gol/ min (19 cm /3 min) g .
I TFC # 1: 0.010 gol/ min (38 cm 3/ min) dQ s s. mc N. ! 2>-14 0 J, 1 e ,', i:d -150 C a IGSCC #1 , k.$ 0x J l g E cc TFC #1 J H -16 0 - 7 ELECTRO NOISE LEVEL *
~
-17 0 Q
4 O. 200 400 600 800 1000y FREQUENCY (kHz) l
.s V
i i l l
120 i i i i I _ 13 0 50-pm SLIT E ' FLOW = 0.022 gal / min (83 cm3/ min) 2 / T = 342*F (172 C) 140 - - P = 322 psi (2.3 MPa) m FAC-500 TRANSDUCER,S/N 35130 a ,
; El 150 5
a E 160 { INSULATION m 170 _
- w ~
y (N0 INSULATION BACKGROUND g 180 I0 0 200 400 600 800 1000 FREQUENCY (kHz)
i i 1 i h O O cc
-N l I I e8 : 1 e l s l :.
N 2 1 I W D 0
! W
** CE tun.
1 ~. A 2 9 [ 4
*o*
6 enummun s 1 SP'30011 W e i i I
----, --m ,--, . m.- --v- ----..%
m - - - - . - , - ,, - - - - , - 7
i M e i 9 e i ( N q O ro 4 r0 I I d Y
- j. '
1 i q 9 OW N N 1 W~ l I c 1, c
'm s 8 m 8_
s_ O @ N t* D O < O O ' ' O O' L p Ci l l o N l o o o O O o D e N O N_ O_ D (AW) 13A31 lVN91S SW8
s a RMS (v ) 1.0 2.0 3.0 Q .~. . _; -- s-l - Tt%..- l T 4E}4.
..g - '.
_ a__ i 5m i i omo t y E_. m o l m a --0.002 gpm o A" -. d Sc -0.005 I- gpm n . . . . . m '~'~'
-. :' v-v, w 0.006 gpm- m-o -
Ogpm y
+ NOISE -
-,52 .
= fg.:- r .< -
h 0.004 gpm 4-. J. ;2... . - A CD o Ogpm 9-m Di .
- s. .
8
,m- -.
.mn ~: ..
'A. ... . ,_ ...
f e . O
I lh2 ..hdy ,: j, Dw.,; - ' b_ . ,
? =
']- - .
r e e. - E91 Li F p, u, l e '% --
~
i ' l .. I -
, b
. ,c
- 3..
gg, n. i l '"lj j c s v a- _ /, t
.f t'
gi I
/h 'I k;i
. .. o,.'.
l i )' .
, p.
^
s . ! As i L 4k-./-
/
p 1 ACQUIRING AC0USTIC BACKGROUND
- NOISE DATA AT WATTS BAR l
34 ,1 gr gr 5r ga g1 yil i i l l i 30 ACOUSTK BACM6AOUND NOISE :4 pV (LOW) - 16 - 12 - ~
. 1.0 gothnia i 8 -
! 4 - -
t 0 i
-4 - -
0.01
-8 - -
k -12
~
- 16 l-20 -' l'I'l'l'l'l' l' /
< O 2 4 6 8 10 12 14 16 18 20 m
j 20 , y,, , ; , ,,,,,,,,,,,, 16 - 8 20 pV (ME01UM) - 12 - - 7 o 8 - O 4 - O o
;- -4 10 -
5 -8 - - W -12 - - - 5 -16 0.06 E -20 - ' I I ' l ' - I'l'_ j O 2 4 6 8 10 12 14 16 18 20 z o 20 Gi l l ' l l l l ' l ' i ' l I - 40 pV (HIGH) ~ 12 - - 8 - - 4 0 \ ~
-4 -
-8 -
1.0 -
-12 -
-16 -
301 - i N, , IiliIili O 2 4 6 8 10 12 14 16 18 20 OlSTANCE (m) _ _ _ _ - . _ _ . _ - +-
l L 1a ,. l t l ET375 E3 RCOUSTIC UEM ["3 ET375 ! Sit 1ULRTOR 250 se1ong , 3 se dia. M7M VA7Dl
- </////////////////////////////////////////////////
.4 fis////////////////////////s,,,,,,,,,,,,,,,,,,,,,,
tO-in. SCEDULE 60 PIPE l l I I i 1 l
- _ ~ ~ . - _ , . - _ - . - . . - . - . - . . . _ - - - . _ . . - - - . . . - - - , , - , - - . . . _ . . - .- - _ _ .-- .- .--. - .- . - -
i 7,,7 3: LOCATION Of' ACOUSM ee T,e] PATH OtSTANCEtcal N20N MANSOUCERS t D-CL SS C** M e.t.c: LOCATON OF e-----CENTtm c-CL 200 5 (CLI e-CL 225
@^*LE*" es Trel Ti -T2 8000 l I 1 I I I I o ' -
[7$ SOURCE POSITION _ -225cmFROM CENTER _ I g - b +33 SOURCE POSITION a _ +55 cm FROM CENTER _ 2 l $ - - 3 N l
- C +*0 L e s' SOURCE POSITION -
_ +200 cm FROM _ CENTER I I I I I I I I T2 T RELATIVE POSITION OF CROSS-CORRELATION PEAK (AVERAGE OF 9 CORRELOGRAMS) l l e
.--.A -. _ _
O V 2 I
- ~
I
- Es$?BR
$1 d
! osa I
e a i d i . ti-l is, . t of. l I a
stored data as file corp 12.dat 7 "##,z.
- b w **r ~ ^
anotherdataset? I - v . . a o maxinux and Mininux time delag 2.048 ns g l
, , e e i 1
V SPRING WASHER C TRANS0VCER w WAVE GUIDE REFLECTIVE l INSULATION 7g
& w 4g p/ ,
/
1 PIPE STRAPS
h __ . . _ .____,.A _--.m.__a. _.. _ _ . _ _ _ --e_a__ _ . - _ . _ . ___.a2 4 l s + h i i l
- s. I l
- i, l ,
[k'
,9 .
ew %, . .
' *[op i - y,,
- w. .- :: '
*: ': , ll,
- f
.s
, ie '.( - **
_ Y i i,L,n M . L j r uj,:, ',j,' y[ r t s
,%g t
..f. .
%,ML O E' l / ,' , j g DETECTli1G SIMULATED LEAKS AT g e , ,- ,
*f i BRAIDWOOD 1*** . ' *'
. , l
X M e Mas.v.no.; . f .. s.. . - - m.nv s s.re ,' , 7 ' _ _ . & onns..o % *'" &_ f le;yf m p g. . I i mA.. . .i n'
- s' ,, ,
. ,N ..
u ,..-. ..e c . .n I i~ l#2 .ks$*t , , . ,
- l q i. c "$s.y ;
i !
- W fjl
.y
,jrt.rg.
- '. t? ..
1 ,: u s 4 ..- t'g+,.;m . 9. 4, ;. a
.. m ;s- ,
- v. w.::, ,
g
. s,.
. ' i ,l ,
., ee.. .&..:
.8 ,
l ,t
. 1 i
l t i ,' J- - I k la . u ./ , i .. s e,') , I. . ! l .
! t
. 4
.i
> ., 3
..- e
[ . M ( . 1
. ;.,a,: . o ,
r
. .i
- c. .
e. g .i . , .
*1:
-]^ - - t g .
g .... m . ? I.
. .;m. .: e_ .
y' & ;;' ~~~3 g .,
.k
- z. .l
- 1. 4.
,qI.
..a. ..'. .
y % g 3; f.[;, h .,,s l l
. -[,. .
. t.. . <,. . .
. . . .s l .
g . ,,' ' r K.
. e
,, u ..
r ...g I
. .w.~.:
. . . 3 .
;...l.
A
,. n v . g
- . ; i .
(~,, ~ ~~., ~ j
;. -j
,J.
.,.u :. .
.. w.. . . .
s ..t 0 ...
. .. . s . . t ..,
v
+ .. . . g.. m... .
- - c.
t ,n s .
.;. ss-.
ss;...:.
. , , .s ,
s c - c
. :> u s
... s. ,
t '. ;[.t
. .t . . :a,-( .+. . .: . ,.
g
.. .l
..l** ' N. ,....;.
* . . y*;* . .l ',,
p' *.t . A.th.,.hy'.i.. ...
.w s.?.! ,
.g . , , ,;
- c. > 4.
{... , 3 .
= .. . -
1 l l l I l l Leak location capability has been improved by advanced signal processing (cross correlation techniques). Leak characterization has been improved by spectral analysis and pattern recognition techniques. Computer based system tested under field conditions at Braidwood (CECO). Plans initiated to monitor Watts Bar Nuclear Station (Tennessee) with ANUGARD two channel acoustic monitor. e
t .
~
Future Efforts If No Funding The Survey of Existing Leak Detection Systems Will be Completed
- All Experimental Work Will be Terminated No. Technology Transfer .
If Funding Continues Software Development will be Completed Experiments with Large Leaks will be Carried O.ut Field Tests will Continue (CECO, Watts ; Bar) NRC Regulation Guideline for Continuous Acoustic Monitoring Will be Prepared l e O Eh
l O investigations of the Mechanisms of Thermal Aging of Cast Stainless Steels 4 by O. K. Chopra and H. M. Chung Materials and Components Technology Division Argonne National Laboratory Argonne, Illinois 60439 . Presented at ACRS Metal Components Subcommittee Meeting Battelle Columbus, Ohio July 2,1986 l I i 1
b i Current NRC Research Program Seeks To:
- e Characterize the microstructure of long-term in-service l
reactor components and low-temperature laboratory-aged specimens. - e Correlate the microstructural changes with loss in toughness.
. Identify the mechanism of embrittlement.
. Validate the extrapolation of experimentally observed embrittlement to long-term aging at reactor operating temperatures.
e Characterize the loss of fracture toughness in terms of fracture mechanics parameters such as the J ic and J g curves. e Provide understanding of the effects of compositional . and metallurgical variables on the kinetics and degree of embrittlement. . 1
- _ _ _ _ _ _ _ . _ . _ _ _ _ _ _ _ _ _ - _ _ _ - _ - _ - - - _- __ . _ -_ . _ .w
Material, Product Form, and Condition of ASTM A351 Grades of CF-3, CF-6, and CF-8M Cast Stainless Steel No. of Product Heat Material Heais form ^ Treatment Remarks Swiss impact specimens 5 S1 RH Laboratory aged up to 70,000 h of 300,350, and 400*C In-service components 1 PCP PH In-service for ~ 12 y Large heats 6 S2 PH Composition varied to provide different ferrite content and concentrations of carbon and . nitrogen Reactor components 6 P,PC,1 PH Static and centrifugally cast components with different wall thickness Small heats 19 KB PH ' Composition varied to produce different ferrite content and concentrations of chromium, nickel, carbon, and nitrogen
^Si= slob 250 mm thick; S2= slab 76 mm thick; P= centrifugally cast pipe 610-940 mm OD and 38-76 mm woll:
PC= pump cosing; != pump impeller; KB= keel block tapered from 30-100 mm thickneas; PCP= pump cover plate "RH=lobertory hoot treated before aging; PH= production heat treated,
ASTM Specification A-351 for Austenitic Steel Casting
- Chemical Composition (%) Mechanical Properties, min Te nsile Yield Elongation Grade C Mn SI S P Cr Ni Mo (MPa) (MPa) (%)
CF-3 0.03 1.50 2.00 0.04 0.04 17-21 8-12 - 483 207 35 CF-8 0.08 1.50 2.00 0.04 0.04 18-21 8-11 - 483 207 35 CF-8M 0.08 1.50 1.50 0.04 0.04 18-21 9-12 2-3 483 207 30
" Furnished in solution-treated condition.
0 4
g. 30
/ '
o\o GF ANL , o
*s AUSTENITE O O CF- 8 t o o CF- 3 g n e CF- 8M to d 20 +
z O
,-*7
[ o O m Ng6 03 l 1
+ -
_J 4 E 10 n MARTENSITE g.
's FERRITE O
O 10 20 30 40 Creq = Cr + Mo + 1.5 Si + 0.5 Nb SCHAEFFLER DI AGRAM AND FERRITE CONTENT OF CAST DUPLEX STAINLESS STEELS FROM THE AHl_ AND GF Studies-i
30 / EDF FRA *\ o o CF -8 +@ AUSTENITE o CF-3 c a o CF-8M 2 to d 20
+
Z O m --- q Y /I e
, l 0 / I l I @ f
- 4. _ _ _I O
y 10 n MARTENSITE E O E . FERRITE O O 10 20 30 40 Creq = Cr + Mo + 1.5 Si + 0.5 Nb SCHAEFFLER DIAGRAM AND FERRITE CONTENT OF CAST DUPLEX STAIMLESS STEELS FROM THE FRAMATOME' ann EDF STUDIES-1 I
I Time and Temperature for Aging + of the Cast Material for Charpy Impact, Tensile and J-Integral Tests
' Temperature, *C Time, h ~
450 400 350 320 290 100 A .A 300 A A A 1,000 A A A A 3,000 A,8 A,8 A,B A A 10,000 A A,B A,B,C A,B,C A 30,000 A A A,B,C A,B,C' A,B,C
>50,000 A A,B,C A,B,C A,B + Aging time completed for the small experimental heats
' and reactor components = 21,000 h and for large experimental heats = 11,000 h. A = Charpy impact fest at room temperature. B = Charpy impact fest at different temporatures (DBTT) and J-Infogral (1-T specimens) and tensile fests at room temperature and 290*C. C = J-Integral (2-T specimens) tests at room temperature or 290*C.
,..-c,
~
Microstructural Characterization of Low-Temperature Aged Material o Swiss materials (aged at 300,350,400 C for up to 70,000 h) e Fluctuations in Cr content between 15 and 40 at. % observed _ in all specimens by atom probe FIM examination.
. Two other precipitates; G phase (rich in Ni and Si) and Type X (unknown), observed in specimens aged for 70,000 h at 300* or 2.10,000 h at 400 C. Mottled structure of a' phase also seen in these specimens by TEM.
. Size of the G phase obtained by SANS.
e CF-8M steel showed another phase, Type ML, associated with ! G phase.
Microstructural Characterization of Low-Temperature Aged Material Continued o KRB pump cover material (~ 8 yr service at ~ 280 C) e Fluctuations in Cr content observed by atom probe, e a', G, and Type X phases observed by TEM. e Grain boundary carbides are present; may not be associated with aging. 1 . o ANL heats (aged at 350,400,450"C for up to 10,000 h) e a', G, and Type X phases observed by TEM.
. Grain boundary carbides observed in CF-8 and CF-8M grades.
g
- e. ?. s e .s~. .Y y x,. ;K 4..l,*.@... p;: p. i . ~ 7. ..
a.e f *gr.s,..*. <. . .~ , . -
'.
- g * .' ' , '
Ri't 4 e
?. kY$ * /* *, . .,..
C o,vh.,$ .'*g , p % > *1.'>'s..k.
.t.g.w-.sp. .,,;g.,i. . .*-l.
g..v w. .. ....;.' .: . ' : .
, s. . .. .
$y .gp!~_.
U p. . .* l .0.2
, ;,;: , pm
* . l. 's*-a'.*% $%.*.?
o.:s ; l
**W-d-Q dE?;Qq,. F ' ' ' N.
,gp ur.- '
- 7.i;.t , f.
hi *y
'WbW;,, .a I
g4 GCQWN ' $ I. ~. 2
~
._ . .v.-r.R4 .Yi-a. - - .-:W.93 .
Fig. 6. Bright- and dark. field images of precipitates in the ferrite phase of CF-8 stainless steel (heat 280) aged for 66.650 h at 400*C. t
T' i 23 aa;;.s.j:ai;w~x.k n'. unu&x'n!os%e.?G W 4Y
.k. .
R M: f S w c n~. . .t2. : .
..v
;L a.S. "
w.&. .1
'n..
'.,. 2 E,
..,.c
-< .a.
u 45 'n.h;e w . . p s..?..&. y s .;c,
- ..-.u. .n
- n. , 1. V. .g4
%, 5>.w,.,r,Mh
.*.g.n.v...GQ2w~&p.
n
. ~ mu .; w.:. .n..< @..n F. ;5..&.3 ~ : .ngl.
...y.. .
- e. -
V,n 1: .
. .w- . . . - ..g.
,,.,g, :;3w: .m 93.,:}y.: ,k..
';My.y .-a..,3:,rj:i? ,, .
s .,. ..
.m. , . :.c. . 4'c,.e.S. a 4.P
- u. . :n,.< :
.:u&y ..f.,C'yr.. W., , . 4 +
- c .. ..
.:>,.:; *-. . W.= ~~ :5 i:.3.wr:,x s.
n::4: Maan.mic:.s::-..:..,{h)w.sw..m+Y:y ~
. . > , r w 3 6 C ;.uW& @a s a . y.%g.:... m .
7 e i TPr:w#55Naw.:p.uma
+P.w new e p.w&%;.ddAustem.te?
4 w -- N'2'5 w 9 q:?py;;; lO,O t,1m m:;?+m. g gpt
. . . . s e n.,
y 3gy w ~ .44. w..
.~ y3M_ .:cS$
' DI "-$;Mf(Wf""MN': p;r. . h' - 6 Q!YiNfM .Mh 9 di? R.Q
)& ,& %ya A:M'QgkWi%ci;q':4 m ?y, .,s
- s ..Y'*$Weh:2:
s, . 49;. spt.53.x&y.".Q;(.
- p v z, p:.sp y
-:;a. M
'Q.Q r;.9 Q
n.(J Q A r. y>J.A' g;1:7tA%,.s.%
.#.*; .% 4wQ1 kg %
lt? f hr fk k l $
' . Q$
Ff fh2.E%f:',d" .,y ##, h8f.tg2:> 2j&&L#$
&g$gM/p[flNWijM y $; ~~\.5dif$$ti&W(M*"%:i%gh&ppr}
~
b%? - fA$% 5 E:.d.@% V -- ,L. ..fr.5 p .WselFerriteM&C>a. vvar 4 M%FA9,;9,f%.~%~:g3... m .y .A.g .. f.) A;@f;@;g.
> W.f :.$?j.g ..<
e$n . @cc
- f. y2 f c- 44
, p.w ,;c; a.
a x.,,g k.g 9 - 2 :,,
*?.*@M
.e .n yv.de:s 19 F.;gQ; L'pf,Td.9
.'V
.t.
,is...
mt i- . . .: I h m. . Kftigs&X[ l':w; ~. % 1f, e ' t' m .. 4 (: .;%NQ:4W.M e..... ,.ll f.; N.&5N
. g . ,:f. q... jn:k...r.H.
.Q.h,c . ,s
.a ... . k.?h. y',W;
,, ik?f,.2..,f'.lb ,j
%- ' .$ ? .
it -
.g .. s -
+
,i (B) - Fig. 14. Morphologies of a' Observed in the Ferrite Phase of the Reactor Pump - Cover (A) and Heat 60 Cast-Duplex Stainless Steel after Aging at 400*C for 10,000 h (B). M C are denoted by arrows in (23 B).6 precipitates on the phase boundary i. N _ _ J
l l l 1 l 1 l l I 50-l
}
L .. O N Q.. 1 A
- : : : . . : 1 HEAT 280: 400*C, 3,000 HRS
~
5 0-- L .. l O I - X . p .. l 0 I .. l l i : : : : : : : . 1 I HEAT 280: 300*c,' 70,000 HRS l l I i I 6.
50-d IIEAT 280: 300*C, 3,000 IIRS i 50-- "a . L ..
- o '
x l y. ed i -
\ I
-- I 3 i i i
; ; ; ; ; ; . ; --i i
i KRB PUMP COVER PLATEi 200*C, 412 YRS
1.20 g g l l 1 l ' y 1.00 - 5 (A) y 0.80 - - G
- i a
g 060 - -
> i y 0.
- 0 - -
g .. a - S cc O.20 - { i. O. I I l - I I I O 8.00 16.0 0 24.00 32.00 40.00 48 00 5600 , DI AM ET ER (I) I i 1.20 3 I I I i
- i
$ l.00 - -
E te) a ill y 0.80 - -
,l 0
a y 0.60 - - y 0.40 - - w a O y 0.20 - - ) w I I I I O 2000 40.00 60 00 80.00 100.00 - DI AM ET ER ($) Relative Population of Precipitates vs Cuinier Diameter Obtained by Small-angle Neutron Scattering Technique for the C. Fischer Cast-Duplex Stainless Steel IIcat 278 after Aging at 400*C for 1.2 yr (A) and 7.6 yr (B). t .
Fracture Behavior of Charpy Impact Specimens o Room Temperature Tests e Ferrite phase fails by cleavage in the embrittled specimens. Fracture occurs preferentially along the ferrite phase. o Tests at -196 C e Ferrite phase fails by cleavage in the unaged and aged
- . specimens.
The high-carbon CF-8 (all conditions) and CF-8M (high-temperature aged) materials also show a/7 boundary ' separation, i.e., carbides are prese~nt in the unaged CF-8 material. l e Fracture mode for CF-8 and -8M materials aged at 450 C is predominantly groin boundary separation, i.e., carbides precipitate during high-temperature aging.
~
'g' - N Id j' JQ} }
&w r *sa ^
m= hers::.'**2
; f.fK ~~q, ..?T
's. . So pm is.- .. M .r n"EI.. ,. l HEAT 51 y, . . . -
y* y
; ,- ,e
,, s. h 1 -
$j(*
w ,,, . w m y,~ ' -
') %
.' 'M W ,, ' <
l - h\" 50 pm Os ,
' Nb l l f HEAT 60 b
I, t R. w;. D <
, ~ =,- .
w k-
- j. , . l HEAT P4 FRACTURE SURFACES OF CHARPY lMPACT SPECIMENS OF CF-3 (IIEAT 51),
CF-8 (llEAT 60), AND CF-8N (llEAT P4) CAST STEELS AGED FOR 9980 H At i tOO*C AND IESTED AT Room IEMPERATURE.
I 3.y. . - -
,k ~l ,
^*
^
n @.
;}}ggff -
#j f'" y h* '
s A:
#M hv
#Pe:is, c \ .. - ~.,
&, gN- t
. -411 RN 'i
- kW l . ^ Tt4&dG
; p
,o a .. .
~
b_9 '
- ,C '. ...
Y$ ' (
- w. 4 i l
MATING FRACTURE SURFACES OF CHARPY IMPACT SPECIMEN OF CF-8M Cast STEEL (llEAT Pft) AGED FoR 9980 H AT fl00 C AND IESTED AT ROOM ~IEMPERATURE-
.,> l l
v.
't
..e *$ .
5
- 15 pm he -
l l UNAGED d
.k - \. 4 k -
', -l O
5 pm 9980 H AT 400 C
' k '- ,
, ),
YI Y 9980 H AT 450 C FRACTURE SURFACES OF CHARPY IMPACT SPECIMENS OF lINAGED AND AGFD CF-3 (HEAT 51) CAST STEEL IESTED AT -19b'C.
~ -
mqq ~ .. . s ( I ' fO,f5 ' 8;kl f . 'of {
> ; .p.,a .
w. b[&5k d % M } I - - fA ~- h+. s >i .
.g. . ,
f ! UNAGED [ T N ~ f(
% $ W % l $$./. ,*YL 2hyqqq, 5+ -F^
gy
- g4A nw
( ,- , eg .n
^lefbL' z.. , N J ^'~ ~
2 .s.f
;p*qofY ggp
f '
. a ..
p1 , . f ., .A f . [ " K ~ .h " h i $. u
- t v...r U. .id, e ,.. . M. ,e dm j p ,,15 pm 9980 H AT 400 C
['
- p. .
.<, .-- 12? ', ,
l
/M/ 4kh h '..d 9980 H AT 450 C FRACTURE SURFACES OF CHARPY IMPACT SPECIMENS OF UNAGED AND AGED Cf-8 (IIEAT 60) CAST STEEL IESTED AT -196*C.
il I il l' I l-i r N.
\
i t.% gr ., l
. y.
l a $$SW I gp Y!jd! h 1 197FJ&eal4 FRACTURE SURFACE OF CHARPY IMPACT SPECIMENS OF KRB PUMP cover PLATE IESTED AT -196 C. ! l
)
(
Charpy Impact Tests o Room-temperature dharpy tests have been completed for the KRB material and 16 heats that were aged up to 10,000 h at 450,400, and 350 C. o The data indicate that low-temperature embrittlement
- depends on
e the ferrite content e concentrations of C and N, and~ e distribution of ferrite in the duplex steel.
300 -
-Q A A g TEMP ( C)
_ 200 -@ m c 3 5 - 0 450 U # - 3 A o 400
^
3 g o A 350 J 100 _ oi i 80 O g _ 8,J t a w 60 m Z O e w m 40 CF-8 O D OPEN SYMBOLS: HEAT 60,8 = 21 % 0 g
<t -
2 CLOSED SYMBOLS: HEAT PI,8 = 24 % H 20 " 2 3 4 5
- l 300 --E
, 200 , a O N
@Q o .l a' d 100 _
o _ i 80 a - O w 60 6 _ _ p CF-8 o 40 g' - OPEN SYMBOLS : HEAT 59,8 = 13 %
~
CLOSED SYMBOLS: HEAT 61,8 = 13% 20 22 2 3 4 5 ' PARAMETER, P t INFLUENCE OF THERMAL AGING ON THE Roon-TEMPERATURE IMPACT ENERGY oF CF-8 CAST STEELS-
500 _ j [ l _ 400 E 300 -o- 0 o 4 C B - B
}
^
^B j @, i- h -
q e w z 100 - 1-- 80 CF-3 -
@ OPEN SYMBOLS: HEAT 52,8 = 14 % -
g 6O CLOSED SYMBOLS: HEAT P2,8 = 16 % H - 40 :: 1 1 1 2 3 4 5 300 -t: - i g l 0 o g - 200 F O g g ga a j - 00 g
~
cy 100
>- 80 _
e TEMP. ( C) m _ W z 60 O 450 m _ o 400 - A 350 CF-3 40 OPEN SYMBOLS : HEAT 51, B = 18 % CLOSED SYMBOLS: HEAT I,8 = 17 % ~ H 20 --t I I I 2 3 4 5 . PAR AMETER, P If3 FLUENCE OF IllERMAL AGING ON Tile ROOM-IEMPERATURE IMPACT EtlERGY OF CF-3 CAST STEELS-t l
HEAT 51 HEAT 60 ROW 3 ROW 3 t,50 bm 8= 18.0 % l 8= 21.7 % = g -f , , ! - h,!$ hkb i.kk
- h? $
TRANSVERSE SECTION o .h -
'r~ \
~
r.#
~ (\. MW .m dy-g/E fC ' ' .
, ., ?:..:.
%%h. KIT zgk k }f,' t.. k
~' '
., -f l
~~Y'.:W? ' ~ L ' ' .%
[, g "i '
')fb , E.> '
5%.
4%[gM LONGITUDIN AL SECTION fif I s q
}!), '7
*%.7 ~ '~ s f' . ,
/~c :a ga . :. , , e. e
[y.
. ~ .
.. -~t %
,[ ;fggppg~R9
- jI u ,
.]l.,
l).,Y M ,+g_ f
; ^ "b .
Qu ,g- .
- w. MN Nkr$g$~,.N ' _
VERTICAL SECTION
.A = 52 p ri .f. = 63 p rt FEftR ITE IloitPfl0 LOGY lri STATIC CAST STAltiLESS STEEL.
l 9
1 300 H: ; g ; i i TEMR ( C)
- 200 % o "E
o - 4 a O 450 3 2. o 400 - e
,p , a 350 o A.
i 8 100
- g -
g 80 e w o - e - c ^ B o z 60 _ g W -
$ 40 CF-8M ,
f OPEN SYMBOLS: HEAT 65,8 = 23% - CLOSED SYMBOLS : HEAT 64, B = 28 % 1 I I 20 :: 3 4 5 2 , 300-02: g l l 2A 8; - A 200 u - w * [b ,
., g _
l 9 - 3 - o 8 100 A _
>- 80 L9 _ e a _
N 60 - z w - m - D 40 CF-8M a a f OPEN SYMBOLS: HEAT 63,8 = 10 % ~
- ~
- 2 CLOSED SYMBOLS: HEAT P4, B = 10 %
H I I I 20 -t: 2 3 4 5 PARAMETER, P INFLUENCE OF THERMAL AGING ON THE Roon-TEMPERATURE IMPACT ENERGY OF CF-8M CAST STEELS-I 1 t
HEAT P4 I OUTSIDE DIA. HEAT 63 150pm ROW 6 8= 1 1.1 % ' ' l.50 pm 8= 10.0%
.539
'.i-i M. W "9 %2h @W:$
$ :. .) h . *M %[ - } % [ 7 h PW$r$e.
- w. .. . c
. .-:M9
,. :.u:
cMf6@:.~tTshk
.w .'.Mn. n- ..x(>%. , , , ' . . 7.
; f./ .
th :.
. ( , J. W( . N,. . g y:
-.+%n 129 . n.h.. .
* .n . %.
G Y'.'; M. .:s.~.. 7 ~;. ?. ; [. ,f. . s .1
% D ' : ..; &;%.. X..l d,,x. M ?
.n.M ' a .s
%Q h -1 p.1 l .,y:
.W '.'.pj~....f.s..f/*
?.?."!* .
- h. N..4h $}5j.YY*?Y. -[, N..
$[w i. -;.lelEV$@.,
r. s
.9 L @ 2 & .~ h.t h. L :2:M9.% 3A htd':Ni iia.:r$..\@ d TRANSVERSE SECTION 7:- 9 :yr.a..n=n. o
, .c p. ~.m.v ~;.r .m ;< y-,'
% ., 4 y{Jx'k . A.z'. :,.: h,.s y . .~
- ,.y ..:4:d. pik i &;G. m. %.
-::.. c. <.y r.. . - Q.,x?1
. - 0; g..
y 1 %c ,j @ 2 W a '
; sc;a -:.,n
- . .dj .
5'_3,. gf n g, n_
-c
. e pwe ag.w, eg ..... ..y.,s1.,4, z. <
, ,. sg . . ge
'C :1 ' - d f 6,. / , I Y, b ,r~3.M:.g./.9
$,. :.. [r,
........12ri$$fE
. s ::me n , v' e<:. c.. > . .,c.m.m w . s. n
>. :~
.m. r a.....x f
2 i ,-%, s./ ... ,,:. . i; M . = .w. i2 k.,.* Q.
.g .w.; - n w% v..- g.
-Ld $s.w: %? #l5d&A~s; At%g.Q. -:y9Ed.h :yq.(i, &_
W W4y. LO NGITUDIN A L SECTION
-n .
m m rr m-Q m . w ;g^.,1 t.:.ft'E W [~h':y:xr.m~~wcfi&gu f.3 ggfb- ' Y# .-i l..f'3:>A:.cgx+J.*as;.ry]. 3 ,4 Q~ z: {t.2.,?m,d
...s,
.,n.
g.- s> $a g.. -nag y -. M.s y g2tw ul
.&, a.y,m. D.~ t. ..
g S h.- g .w 4 ,. > W, m % y
..x.:.a
. en .
m.g A ; y. ;- 3d y., 5 w%c:.r..w?1 ppccQfp..e&g%{<
. :ipN . L, -
w
-}s- pux..ps:- .
.n u ,;g; 4k_,..g.s.sw phDd@,z- d!~kk. N8,:h 4~p,NNDEMD 4
WSIMTd@;. ifEd5$ ~ O$fi$.j$$$N$ r VERTIC A L SECTION
.f= 182 prt .J,= 81 p n FERRITE Il0RPl10 LOGY IN CF-80 CAST STAINLESS STEEL. !
k
Fracture Toughness and Tensile Tests o J-R curve and tensile tests have been completed (at MEA) on 3 commercial heats aged for 3,000 and 10,000 h at 350 and 400 C. - 0 Fracture toughness results show good agreement with the Charpy data.
DELTR a (In.) 0 .123 .202 .333 .423 .520 7583 , ,, i i i 4B200 AGED CAST STRINLESS STEEL (25'C, L-C, 20% 'SG, 689.5 MPa)
~
,tBIG IRL CODE PIB o ItaGED -
32220 D ,BGED 12228 HRS. e 350*C g _ 4500 -
/ / o RGED 12220 HRS. t 4BB'C d ~
5
.=
/ / s
~
20208 8
; 3220 -
/ / O i o -
/ / o O 0" 0
[ f [cp O o tn - 1222B 1500 - a o 0 c/ o 0 0 B B 2'.5 5.0 .- 7.5 10.0 12.5 15.0 IlEl.TR a (es) DELTR a (In.) 0 .103 .200 .323 .428 O 10223 , i i i RGEll CAST STRINLESS STEEL - 5e220 0229 - (25'C, L-C, 20% SG, 689.5 HPa) 40000 S -
~
6020 - A g 7A d
$ gO/ o -
30000 )
~* O O T
' 4220 -
A O O / o 0 j 8 !
- o -
20000 %
/ o o gooo NnTERIPL CODE P2B 2003 --
o O t;mGED , o cP 7 A AGED 3229 FRS. e 35B'C 18888
)
0 o AGED 10228 FGS. 8 423*C 7 l O($/, , - O I i 0 2.5 5.0 7.5 10.0 12.5 ' DELTH a (ne) CRACK 6ROWTil FRACTURE RESISTANCE BEllAVIOR OF IHERMALLY AGED CAST STAINLESS STEEL AT Roon TEMPERATURE-
Mechanism of Low-temperature Embrittlement o Low-temperature embrit'tlement is caused primarily by a' precipitation. l 0 The a' phase forms by spinodal decomposition. o Aging at 450 C may be outside the miscibility gap for some compositions.
- o Role of G and Type X phases is still unclear.
O Carbide precipitation at a/7 boundaries influences the tough-ness of high-carbon grades aged at temperatures >400 C. O Preliminary results indicate that toughness of the low-temperature aged material may be recovered by annealing at 550 C for 30 min. l l
r~~ y / m
)
. . . _ _ _ . I Jp Z HUCLEATION AND
=0 GROWTH ONLY / JC2 j E
_. r0 _. __ \
't[ \
\
T SPIN 00At. ) T REGION x \ W S P - - - ll I \ a ll, a+a, \ a. i I L t i . i i I f Xcr
. COMPOSITION -
700 . . . . . . . . . 3 650 MULLER AND RA0 AND KU8ASCHEWSKI TILLER 600 - 550 - 500 - - 450- WILLIAMS - M o EXPERIMENTAL - 400 - USING MODIFIED . 3 ROOT 3 RULE M
- s 350- -
W 300- - 250-
] .
r 200- .1 150- ' 100 ' ' ' =
- s. .
t 0 10 20 30 40 50 60 70 80 90 10 0 i At % Cr i SPIN 0DAL IN Ti1E BINARY FE-CR SYSTEM k
I .) l TESTED AT -196*C 0 CF-3 CF-8 CF-BM O O O A UNAGEO
** O E O A 3000 h AT 400*C g 20 a o .A 10000 h AT 400*C o g # g 10000h AT 450'C o Oo o0 a 3 g a .
> 0 o 15 0 5 g oo a
>- O e Et9 5 0 10 a A g g g g O m 0 f o o a s
ag B H a o a 5 0 8 A a0 A
^ d O
! Oo p g ko 4 Of 0 0 0.5 1.0 1.5 8 (%) C (wt. %) CilARPY IMPACT ENERGY OF UNAGED AND AGED CAST STAINLESS STEEL AT -196 C. l
HEAT 51 0-1960 25.00
.....s. a
. ,.#*..... .\
..- t
~~
..- Q N.
f,. .. - - . UNAGED
'f, LCD . f' ~,. '
'\
kn . > f ',
',\ '
'. \
~
9980 h at 450 C
\/ '.,
s
-s,
) .s 's, 9980 h at 400 C " ._
f '..,, g,gg j %[ _ i i i i i i i i i 0,0 2000,0 IIME,usec HEAT 60 0-1960 3 . 25.03 LOD . , 5 kn . '
\ l UNAGED
. Y ;
i 1
\..
\ 9980 h at 450 C
',' j.,
9980 h at 400 C N 'x %, .
. i .....
, s..
g,gg i . .l ._ . _. ... i i i i i i i i i i 0,0 2000,0 IIHE,usec LOAD-IIME CURVES FOR CHARPY V-NOTCH SPECIMENS
1 i Correlations to Predict l Embrittlement at Reactor Temperatures o Existing correlations do not accurately represent the i embrittlement behavior over the temperature range l of 300-450 C.
- o Extrapolation of high-temperature data (i.e., >400 C)
! to reactor temperatures may not be valid for some j grades of steels. l 0 Influence of compositional and metallurgical variables needs to be established. l l l
.. _-]
20 ; i i l GF ANL
. 20.8 9 0 21.5 O O CF-8 N ..
10 e CF- 3 21 9 23 3 N 8 15 ---
- 0 .
m o CF- 8M -
,90,2.00 2
a i .- 0 21.3 22.5. 23 KCU = 52.5-2.19(Cr+Mo+Si)+46/8 (From) 0
* ( i 2h O21.3 22.3 g 10 -
\ \._ -
O
, g \
Cr + Mo + Si= 22.0 - '
- 1 ,.
~ ..
O % " --
< 5 -
N 22 3 21.7 8!23.9
~ -
----- --- 2 2 -
- 't 22.4 23.0 23.6
\\ 23.3 23.9 Q 0 23.4 m %Q,m_ ts 23.6 23.7 24.0 23.8 23.4 24 I I I I o 10 , 20 30 40 50 Fd;RRITE CONTENT, 8(%)
i INFLUENCE OF FERR TE CONTENT ON CHARPY IMPACT ENERGY OF CAST MATERIALS AGED FOR 9980 H AT 1100 C.
\
\
s
\
\
. J
20 ; i ; i i GF' ANL _ g _J- O O CF-8 N g p 10 e CF-3 g 15 -}3-- 0 og a e CF-8M
> c ..
O 3 __ . KCU = 16.7 exp(-l.78 C)+ 0.6 (EDF) O
- O o
>- 10 -
1 - o -- rr n .w 2 0 -- w . I-- o O f5a D k.. $ -
., O O
O 00 , a 3 m_ O I I I I I o 0.5 1.0 1.5 2.0 2.5 8(%) C(wt. %) IflFLUENCE OF FERRITE AND CARBON CONTENT ON CllARPY IMPACT EllERGY OF CAST MATERIALS AGED FOR 9980 H AT 1100 C. , \
i I I Future Work o Complete DBTT, tensile, and J-Integral tests on large expenmental heats that were aged up to 10,000 h. i o Conduct Charpy impact (DBTT), tensile, and J-Integral tests
- on the KRB pump cover material. Correlate the results with data on laboratory-aged material.
! e investigate the influence of N or C content, cast structure,
- and changes in ferrite content on-loss of toughness.
O Continue microstructural characterization of laboratory l aged and KRB material. l i l
ums une mum amm> men umu asmi sus aun amm muu met aus amm smus an uns aus I INTERNATIONAL PIPING INTEGRITY RESEARCH I GROUP PROGRAM l B. F. SAFFELL ! G. WILK0WSKI i 4 ) 1
! ($Battelle l
j
M M M W M S6 WM M M m' W W M M M M M INTERNATIONAL PIPING INTEGRITY RESEARCH GROUP (IPIRG) e INTERNATIONAL PROGRAM MANAGED BY THE NRC FOR THE PERFORMANCE OF COMPLEX PIPING EXPERIMENTS. e PROVIDES A FORUM FOR REACHING INTERNATIONAL CONSENSUS ON NEW PIPE BREAK RULES AND A REPLACEMENT FOR THE DOUBLE-ENDED BREAK CRITER10N. e PERFORMED BY BATTELLE. e COMPOSED OF INDUSTRIAL AND REGULATORY ORGANIZATIONS FROM COUNTRIES WITH NUCLEAR POWER PROGRAMS. e COMMENCE WITH COMMITMENT BY 5 PARTICIPANTS. OBaHelle.
M M M M M M 'M M M M M M M O M M M M M SUBTASK 1.1 ANALYSIS (1) PRETEST DESIGN ANALYSIS e TO ESTIMATE EXPERIMENTAL SIZING PARAMETERS: FREQUENCY AND AMPLITUDE e CONTINGENT: PIPE AND CRACK DIMENSIONS e BEAM ELEMENT FEM ANALYSIS:
- ELASTIC MODAL ANALYSIS - LOCATE CRACK 3
- EMBED P-6 CURVE FROM DP 11 PIPE TEST INTO FEM
- RE-ANALYZE WITH COEFFICIENT REPRESENTATIVE DAMPING.
i (2) POST-EXPERIMENT ANALYSIS e NO STRAIN RATE OR CYCLIC EFFECTS CONSIDERED e USE CURRENTLY ENGINEERING TECHNIQUE
- CALCULATE MAXIMUM LOAD IN ELASTICITY WITH 3-5% DAMPING WITH NO CRACK
- REGARD THE PEAK DYNAMIC LOAD AS STATIC LOAD 1
- USE J-ESTIMATION SCHEME TO PREDICT FAILURE e ASSESS SAFETY MARGIN BASED ON CURRENT METHODS.
OBattelle
- -- - W.. --...----- { D INTERNATIONAL PIPING INTEGRITY RESEARCH GROUP (IPIRG) GENERAL OBJECTIVE DEVELOP, IMPROVE, AND VERIFY ENGINEERING METHODS FOR EVALUATING THE STRUCTURAL INTEGRITY AND PERFORMANCE OF NUCLEAR POWER PLANT PIPING CONTAINING DEFECTS e PLANT OPERATION l e DESIGN CRITERIA , e INSPECTION INTERVALS e ANALYSIS TECHNIQUES. (#Battelle
m W W W m m W W W M M M M M M M M M M SPECIFIC IPIRG OBJECTIVES e DEVELOP AN UNDERSTANDING 0F THE RESPONIE OF HIGH ENERGY. FLAWED PIPING SYSTEMS TO DYNAMIC LOADING. REPRESENTATIVE PIPING SYSTEMS SEISMIC EVENT RELIEF VALVE LOADS s ESTABLISH PIPE FRACTURE AND EXPAND MATERIAL PROPERTY DATA BASE. e VERIFY LEAK-RATE ESTIMATION MODELS. l 3 e COORDINATION OF PROGRAM RESULTS AND REGULATORY ISSUES THROUGH INFORMATION EXCHANGE SEMINARS. l ! OBaHelle
- k
IPIRG PROGRAM TASK 1.0 TASK 3.0 TASK 5.0 l LEAK BEFORE BREAK UNDER FRACTURE OF PIPING INFORMATION EXCHANGE l SIMULATED SEISMIC / DYNAMIC CONTAINING HIGH SEMINARS AND PROGRAM LOADING ENERGY FLUIDS MANAGEMENT TASK 2.0 TASK 4.0 EXPERIMENTAL PIPE FRACTURE RESOLUTION OF UNRESOLVED AND PIPE MATERIAL PROPERTY ISSUES FROM THE NRC DATA BASE DEVELOPMENT DEGRADED PIPING PROGRAM QBattelle
{ h DISCUSSION OF IPIRG BASE PROGRAM TASKS e TASK l.0 LEAK-BEFORE-BREAK VERIFICATION UNDER SIMULATED SEISMIC / DYNAMIC LOADING
- EXPERIMENTAL EFFORTS
- SUBTASK l.1 INERTIAL LOADING
- SUBTASK 1.2 DYNAMIC DISPLACEMENT-CONTROLLED LOADING
- SUBTASK l.3 REPRESENTATIVE PIPING SYSTEM
- ANALYSIS EFFORTS
- MATERIAL CHARACTERIZATION e TASK 2.0 DATA BASE EFFORTS e TASK 4.0 LEAK-RATE ESTIMATION MODELS e TASK 5.0 INFORMATION EXCHANGE SEMINARS AND PROGRAM ADMINISTRATION C4BaHelle
l = = = = = : C = i c = n = a B = ly f = e P s U O R G = H D = C R E T N A A O E = S L U G I T = E R M I N I A Z S D I Y A R = T R E O L E = I T R 0 D C G . N C A E 1 U I R M = T N K K A A H = I S A N S C A E Y T G T R D N S N B / E S L C = M A I
- I P E I S
= I I I R M R Y R P O S E L E F I P A T L E E X N A = = A N B S E A M O K I A e e e T E = A L N = R E T N I { k lL 1 l1ll
W W M M M W W M M M M M M M M M M SUBTASK 1.1 STABILITY OF CRACKED PIPE , UNDER INERTIAL STRESSES SUBTASK OBJECTIVE THE OBJECTIVE OF THIS SUBTASK IS TO DEVELOP AND PERFORM EXPERIMENTS TO ASSESS THE ANALYSIS METHODOLOGIES FOR CIRCUHFERENTIALLY CRACKED PIPE UNDER INERTIAL (TIME-DEPENDENT LOAD-CONTROLLED) STRESSES, l C4Battelle
& S' & & & & M M M g I
Crack Opening Displacement, mm Crack Opening Displacement,mm 20 25 30 0 5 10 15 20 25 0 5 to 15 225 000 1 I a i i i I i i i I 200 - - 200 - 800 g75 _ 800 175 -
- 50 Cydes 50 Cycles I ~
3 S - 600 g _-125 j 50 Cydes 600 ]{r f 125 3 g 100 -
} Y100 - Less than 20 Cycles .$
*l* i
*l5 p 75 -
g 75 - 50 - 50 - 200 200 - 25 - . 25 - - I ! ! ! I ! 0 0 0 0 0 250 500 750 1000 1250 0 250 500 750 1000 Crack Opening Displacement.10 3in. Crack Openmg Displacement.10-lin. DISPLACEMENT CONTROL LOAD CONTROL FATIGUE CRACKING OF CENTER-CRACKED PLATE l CBattelle
TENTATIVE TEST MATRIX FOR EXPERIMENTS ON CIRCUMFERENTIALLY CRACKED PIPE UNDER INERTI AL STRESSES (SUBTASK 1.1) Expenment Bladder Number Material Flaw Type (Yes/No) Remarks 1.1 - 1 Carbon Steel None N/A - Assessment of dynamic analysis damping factors for elastic and then plastic loadings.
- System debugging 1.1 -2 Carbon Steel Through-Wall Crack Yes, - Low toughness material 1.1 -3 Stainless Steel Thrcugh-Wall Crack Yes - High toughness material 1.1 -4 Carbon Steel External Surface Crack Yes - Low toughness material.
- Evaluate compilance changes as surface crack breaks through wall
- Document crack opening area 1.1-5 Stainless Steel External Surface Crack Yes - High toughness material
- Evaluate compliance changes
- Document crack op-ning area 1.1 -6 Carbon Steel internal Surface Crack No - Measure flow rates and thrust loads 1.1-7 Contingent - -
l C4Battelle
W W W W W W M M M M M W W M M M M M M
- Hqh pressure air occumulators .
f r ( f
)
)
't J High flowrote air regulator .
[ k'* %
~
Electricolly isolated fixturing
. ._J C'r cumferentici crack
'~'
Pressure
~
vessel for -
,n i
-;$_Wbler of LWR conditons leakrate U ""~
evoluotons Flowmeter Test pipe with heater tape and insulaton [High flowrote ision servo +e i aantrolled actuator Rsnps and occumulators Q' for high fkNrote o4 supply EXPERIMENTAL SET-UP FOR SUBTASK 1.1 INERTIAL LOADED PIPE EXPERIMENTS C$Battelle j
m a l e e l e H e a B e E P I P O D E K a C A R C m Y L L A I T . W 2 N E R S E S 1 E S F E W K S A M U C R T S T R B I D l C E W l S O R L L O G R F R I T A m P I T A N O C D T E N M R U E M T E C C A A R L M F P S E I P D I P C M P I L E O C V L Y I E C M T C V E R E D E J D B O N O T U M M M M
' llll '
e 6 m a e en IEIB sus as e m e e e a e a g ( h PURPOSE FOR TASK 1.2 EXPERIMENTS
- 1. INVESTIGATE EFFECT OF LOADING RATE ALONE INCREASED CRACK GROWTH RESISTANCE (UNLESS " DYNAMIC STRAIN AGING" DOMINATES)
- 2. INVESTIGATE CYCLIC EFFECT ON FRACTURE RESISTANCE (SOME MATERIALS ARE SEVERELY EFFECTED AND LOAD CARRYING CAPACITY REDUCED OVER MON 0 TONIC LOADED PIPE.)
- 3. INVESTIGATE COMBINED EFFECTS OF LOADING RATE AND CYCLIC UNLOADINGS 1
OBaHelle
M M M M M M M m M M M M M M m m m m a { D Aa,in 0 0.04 0.08 0.12 0.16 0.20 0.24 0.28 0.32 I I 'I I I I I I 2000 -- j
/ A508 0 2 Steel l / 1T CT Specimen - 10,000 l / Room Temperature i
1600 -
/ /
. / ' O 13 cycles l O 75 cycles 8000
/ / Scatterband Solid, Final Aa l / / ~
, 1200 -
// Monotonic .E E 1
/ Loading -
60004 2 I
.E a l/ On 4 8
g0 00 800 - / 4oo9 O O 400 # f - 2000 I I I I I I I I 0 O O 1 2 3 4 5 6 7 8 Crack Extension, Aa,mm A508 CL2 STEEL I J VERSUS Aa FOR A508 CL2 STEEL LOADED CYCLICALLY TO DEVELOP THE R-CURVE OBattelle
M M M M M M M M M M M M M M M M M M M l EFFECT OF CYCLIC PLASTICITY ON LO1D-CARRYING CAPACITY PREDICTIONS OF THROUGH-WALL CRACK PIPEI Initiation Max. Load GE/EPRI NRC.LBB GE/ERPI NRC.LBB Monotonic-loaded 9,882 11,988 14.499 14,337 J-R curve in-kips in-kips in-kips in-kips Cyclic-loaded 7,104 8,011 10,449 12,636 J-R curve in-ktps in-kips in-kips in-kips Ratio of cyclic /monotonic 0.72 0.67 0.72 0.88
- 1. J-R curves from Figure 2-4 were used in this analysis. The pipe geometry and parameters selected for a sensitivity study were:
diameter = 28 inches thickness = 0.94 inches carbon steel linear regression stress-strain curve parameters a = 12.51 n = 5.55 . oo = 57.085 co= 0.0021956 Through-wall crack length / pipe circumference = 0.37 C4Ballelle
== = == == mm - == == == == == == == == == == ==
= l l
! 8 i N E I J B I I 1 - $ 5 - .m 21 1.2 1 6 120 0.562 TYPE 304 SS TWC 550 DYNAMIC 1 AS 1.2-2 6 120 0.562 A106 GR.8 TWC 550 gp 2
^I 1.2-3 6 120 0.562 TYPE 304 SS TWC 550 AT 2
1.2-4 6 120 0.562 A106 GR.8 TWC 550
^ 3 ATf A
1.2-5 6 120 0.562 TYPE 304 SS TWC 550 g 3 1.2-6 6 120 0.562 A106 GR.8 TWC 550 DYNAMIC 3 Ils j, 12) , t 433 , , 4
/ -
x $ 'Jxd <
-v_ .
DETAILED TEST PARAMETERS FOR EACH TENTATIVE PIPE FRACTURE EXPERIMENT IN SUBTASK 1.2 OBa#elle
]
M M M M M M M M M M M M M M M M M M { 6-inch (152 mm) Nominal Diameter ! Test Pipe Wrapped Encirclement Sleeve with i en Heater Tapes and _
,O, ,o,, - Linear Searing insulation
_\ _ _
- - - I . _
l
'" N Pivot Point k
Electrical , i ,.
~
" % Heavy Section I Beam Load Cell # +- Ram 130,000 pound (579 kn) MTS Fatigue Machine
, SCHEMATIC 0F TEST FACILITY CONFIGURATION FOR THE DISPLACEMENT-CONTROLLED PIPE EXPERjMENTS OF SUBTASK 1.2 4 OBallelle j
mus uni uma sum num aus uma use imis amu num aus em num aus sum umu suus sus SUBTASK 1.3 EXPERIMENTS i OBJECTIVE. PROVIDE EXPERIMENTAL DATA TO ASSESS THE MARGIN OF SAFETY IN THE i ANALYSIS PREDICTIONS FOR STABILITY OF A CIRCUMFERENTIAL CRACK IN A REPRE-SENTATIVE PIPING CONFIGURATION UNDER COMBINED INERTIAL AND DISPLACEMENT-CONTROLLED STRESSES RESULTING FROM A SIMULATED SEISMIC EVENT. l I OBattelle l N ]
M M M M M M M M M M M M M M TEST MATRIX FOR SUBTASK 1.3 Experiment Number Meterial Flaw Type ** Comments 13-1 Carbon Steel None Shakedown test to debug system and assess piping stress analysis. Unear elastic and plastic loadings. 13-2 Carbon Steel or SC in Base Metal Low toughness material CF8m having confined ples-ticity at the crack. 13-3 Stainless Steel SC in Base Metal High toughness material Inducing nearly fully plastic condition at the crack. 13-4 Carbon Steel Wold SC in Weld Metal Suscept- Lowest toughness case. able to Dynamic Strain- Evaluate if dynamic Aging strain-aging occurs at seismic rates. i 13-5 Stainless Steel SC in SMAW or SAW Comparable to low Wold toughness carbon steel pipe, but strength mismatch. 13-6 Contingent - -
- 16-inch (406 mm) diameter pipe, Schedule 100. 550 F (288 C), pressurized with subcooled water. Some pipe available from DP3ll program.
** SC = Surface Crack.
l OBallelle u __ _ _ _ _ _ _ _ _ _ _ _ . _ _ _ _ _ _ _ _ _ _ _ _ _ _ . _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ - - _ - _ _ _ _ _ _ - -
i, i' M i i e _ l l e a M B
-~ !*
f M n '5 r
> : (
-Om M
w 0 t - M ' w t s 0 2 o c P M . - " x e F ip p N 0 0 M :
- ON N
1 h c n S o
, it -
r a m - " t e e L c o 6 m) k _ 1 1, x iamca d mrC M >
)
m 54.ed
/, m2 t 4 6( l a _
^ y 0h 4 ct u -
M : s ( hc1 inso P g ~ t e in= = o 6 M N 1 t$ - Z
'5 . -
m # - y a - r 5 m px
, m
/w m
m m m
- , ;ji! ill,i: .:)i I!
I I I I I f High Pressure Air ( Accumulators , f
% 2 fa r
""; t
\ High Flowrate Air
) a Representative
, , gur. .
I Regulato'
- Electrically 9 Precision Servovalve p rw g Check Fisturing I ,,
Pump yagy,
@'. s Computer 't Pressure '
A ' - Controlled Pumps and Vessel Actuator Accumulators I for for High Leakrate Flowrate Evaluations oil Supply I , , _ e,s, ...ters I OBattelle I
uma muu muu num umu muu um muu muu muu muu uma uma uma num num imum lans sum DATA TO BE COLLECTED IN SUBTASK 1,3 EXPERIMENTS APPLIED LOAD ACTUATOR DISPLACEMENT CRACK-M0 urn-0PENING DISPLACEMENTS CRACK GR0wTH (D.C. ELECTRIC POTENTIAL /IIMING WIRES) PIPE IEMPERATURE ACCELERATIONS PIPE STRAINS / STRESSES PIPE ROTATION AT CRACK SECTION PIPE DISPLACEMENTS AT CRITICAL lb,ATIONS INTERNAL PRESSURE FLOW RATES - THRUST LOADS i
~
1 QBaHelle Q j
muu muu mui sum num mun mun suas muu muu nas num um nas ums uma mim E uma { D TASK 1 ANALYTICAL EFFORTS (1) NO ANALYTICAL DEVELOPMENTS e SIMPLE BEAM ELEMENT FEM e J-ESTIMATION SCHEMES. (2) PRETEST DESIGN ANALYSIS e SUCCESSFUL ACHIEVEMENT AND MEANINGFUL DATA IN EXPERIMENTS e PARAMETRIC ANALYSIS FOR OPTIMUM EXPERIMENTAL PARAMETER SETTING
- FREQUENCY AND AMPLITUDE
- HYDRAULIC LOADING SYSTEM.
(3) POST-EXPERIMENT ANALYSIS e USE CURRENT PRACTICE ANALYSIS PROCEDURE e ASSESS SAFETY MARGIN. OBallelle
muu um umm um um muu num ime uma seu um uma um lu m aus seu um muu m SUBTASK 1.2 ANALYSIS (1) NO PRETEST DESIGN ANALYSIS - COMPLETED. (2) POST-EXPERIMENT ANALYSIS e USE J-ESTIMATION ANALYSIS PROCEDURE e ASSESS LOAD CARRYING CAPACITIES INITIATION AND P max DUE T0:
- DYNAMIC STRAIN RATE EFFECTS
- e TESTS 1.2-1
- CYCLIC LOADING EFFECTS s TESTS 1.2-2 THROUGH 1.2-5 I
- COMBINED CASE OF THE TWO
) e TEST 1.2-6 . l C4Battelle N A
ums num uma ammu um amm umm uma muu uma uma umm mum sums mum num um uma mim 1 i SUBTASK 1.3 ANALYSIS (1) PRETEST DESIGN ANALYSIS e FOLLOW THE SAME ANALYSIS PROCEDURE AS IN SUBTASK l.1
- MINIMlZE TORSION AT THE CRACKED PIPE.
(2) POST-EXPERIMENTAL ANALYSIS l e USE CURRENTLY USED TECHNIQUE - J-ESTIMATION
- MON 0 TONIC STATIC J-R
- MON 0 TONIC DYNAMIC STRAIN RATE. J-R e SAFETY MARGINS BASED ON THESE METHODS WILL BE ASSESSED IN COMPARISON
~ WITH EXPERIMENT. 4 1 i 1 l C4Ballelle ( )
uma mum um uma muu num sum num amm muu num umm uma amm uma mum imm ami mum IPIRG MATERIAL CHARACTERIZATION INVESTIGATE EFFECTS OF DYNAMIC LOADING SUCH AS WOULD BE ASSOCIATED WITH A 2 TO 4 HZ EARTHQUAKE SUBTASK 1.1 0 DYNAMIC CHARACTERIZATION OF PIPE MATERIALS THAT WERE CHARACTERIZED QUASISTATICALLY IN DEGRADED PIPING PROGRAM (1) SA 376 TYPE 304 AUSTENITIC STEEL PIPE (DP2-A23) 6-INCH DIAMETER, 0.562 INCH WALL (2) ASME SA-106B CARBON STEEL PIPE (DP2-F30) 6-INCH DIAMETER, 0.562 INCH WALL 0 TEST SPECIMEN TYPES (1) TENSILE (2) CONVENTIONAL FRACTURE TOUGHNESS SPECIMENS (COMPACT OR BEND) (3) NONCONVENTIONAL FRACTURE TOUGHNESS SPECIMENS (FULL-WIDTH-FACE-NOTCH) l 0 TEMPERATURE -- 550 F . I DISPLACEMENT RATES WILL BE THOSE REQUIRED TO ACHIEVE MAXIMUM LOAD IN APPR0XIMATELY 1/8 SECOND (NOTE: THIS IS 5,000 TO 10,000 TIMES FASTER THAN IN QUASISTATIC TESTS) OBaHelle
mum muu num ums num uma em uma um qmm uma sua sum aus uma amm imum suu num SUBTASK 1.3 0 DYNAMIC CHARACTERIZATION OF 4 MATERIALS PREVIOUSLY CHARACTERIZED QUASISTATICALLY (1) SA358 TYPE 304 AUSTENITIC STEEL PIPE (DP2-A8) 16 INCH DIAMETER, 1.0 INCH WALL (2) ASME SA-106B CARBON STEEL PIPE (DP2-F29) 26 INCH DIAMETER, 1.0 INCH WALL (3) SUBMERGED ARC WELD METAL IN TYPE 304 AUSTENITIC STEEL PLATE (DP2-A45W) 1.0 INCH THICKNESS (4) SUBMERGED ARC WELD METAL IN ASME SA-516, GR 70 CARBON STEEL PLATE (DP2-F40W) 1.0 INCH THICKNESS 0 QUASISTATIC AND DYNAMIC CHARACTERIZATION OF ONE ADDITIONAL MATERIAL 0 TEST SPECIMEN TYPES, TEST TEMPERATURE, AND DYNAMIC DISPLACE-MENT RATES WILL BE THE SAME AS THOSE FOR SUBTASK 1.1 OBattelle
- -a
umm immu nas aus sim aus uma sumu can seu aus um num m nas umm unas ami min TASK 2.0 EXPtRIMENTAL PIPE FRACTURE DATA / PIPE MATERIAL PROPERTY DATA BASE DEVELOPMENT SUBTASK 2.1 DATA BASE ON NUCLEAR PIPING MATERIALS SUBTASK 2.2 DATA BASE ON PIPE FRACTURE EXPERIMENTS
~
QBaHelle l
sumi mum uma es imme muu num uma sum num nas seus ums num e uma sum uma num SUBTASK 2.1 OBJECTIVES A. TRANSFER DATA FROM IPIRG PROGRAM TO THE NRC/ MEA MATERIAL PROPERTY DATA BASE IN A FORM CONSISTENT WITH THE PRESENT DATA BASE B. OBTAIN PRESENTLY EXISTING IPIRG MEMBERSHIP DATA FOR INCLUSION WITHIN THE DATA BASE. C. COORDINATE INTERACTION BETWEEN IPIRG MEMBERSHIP AND MEA IN OBTAINING IPIRG MEMBERSHIP DONATED MATERIALS FOR CHARACTERIZATION BY MEA. i D. INSURE THAT MEA TEST RESULTS ARE TRANSMITTED TO MATERIAL DONORS l OBattelle
uma suus mun as muu em num usuu num aus uma ums uma amm ame uma sui aus sum SUBTASK 2.2 OBJECTIVES A. ASSEMBLE RECORD BOOKS (DRB) 0F EACH INDIVIDUAL PIPE EXPERIMENT UNDER THE IPIRG PROGRAM. THE DRB FORMAT l WILL BE CONSISTENT WITH THOSE CURRENTLY PRODUCED IN THE DEGRADED PIPING PROGRAM. 4
- B. ACQUIRE, REVIEW, AND COMPILE PIPE FRACTURE AND RELATED MATERIAL PROPERTY DATA UNDER OTHER PAST AND ONG0ING PROGRAMS.
CBaHelle
SOURCES OF DATA 8 PAST BATTELLE/AEC PROGRAM 0 PAST BATTELLE/EPRI PROGRAM e DTNSRDC/NRC PROGRAM l 8 JAPANESE DATA (HITACHI/NUPEC/JAER) O BRITISH DATA (CEGB) g 8 ITALIAN DATA (ENEA/CISE) 8 WEST GERMAN DATA (MPA/TUV) g 8 FRENCH DATA (FRAMATOME/CEA) I ANY OTHER AVAILABLE DATA I I I I C4Ballelle t
s u a u e s m l e u l l l l e u u a m B _ O a m u _ n _ u _ n s M u O i m R M F A _ S R a m E G O U u S R S P I s G N m u D E I 0 V P _ I _
. L P 4 O m S D
m i K S E R E A N D T U A R m F G E u O D N C O a m I T R N u U E L H O T a S E m u R u u m a u s a m u m u u _ u _ k. m s u s
l m u n s) u s e n u l l m l l e s u a a B 4 m u S L C L E E D a D O m O M u M N N O mu O I I T n T A A M M I T s I S u s T S E E E E T a m T A A R u R
- K 1
K A A E s ~ 4 E L u L s K S F D E A O S T a m B U T E S u S N S E S M A u-u - E N I D N m F E A R D u D E N m E N I A F E N R O A m I T A E U D u L I A V a E V P O R I M E m l E E E E R E ( E l
uma num num uma um imus ums uma uma uma muu es num man sum sum uma uma sum EVALUATION AND REFINEMENT OF LEAK RATE ESTIMATION MODELS LRE MODEL DEVELOPMENT: INTEGRATE EXISTING C0A AND T-H N0DELS TO PREDICT THE LEAKAGE AREA AND LEAKAGE FLOW-RATE FOR A VARIETY OF CRACKED PIPE CONDITIONS LEAK RATE TESTING: PROVIDE DATA NECESSARY TO VERIFY LRE MODEL [ VALUATION OF P6RTICULATE FFECT ON LEAK MATE: DETERMINE SIZE, DISTRIBUTION, AND CONCENTRATION OF PARTICULATES ' IN HPPP SYSTEMS AND ASSESS POTENTIAL FOR PLUGGING EXPERIMENTAL EVALUATION OF PARTICULf.TE PLUGGING: IF JUSTIFIED BASED ON RESULTS OF PREVIOUS TASK, EFFECT OF PLUGGING OR LEAK RATE WILL BE ASSESSED EXPERIMENTALLY. i e i i i OBallelle I k
muu muu unas amu uma umu uma nun uma muu umm aus uma ame usuu uma muu muu ami IPIRG - TASK 5.0 INFORMATION SEMINARS AND PROGRAM ADMINISTRATION e SUBTASK 5.1 - SOLICITATION OF IPIRG MEMBERS
- ORGANIZATIONAL STRUCTURE
- CONTRACTUAL STRUCTURE TWO CONTRACTUAL PATHS NRC/IPIRG MEMBER CONTRACT TERMS INCORPORATED IN BCD/IPIRG MEMBER CONTRACT e SUBTASK 5.2 - PROGRAM ADMINISTRATION
- REVIEW MANAGEMENT ACTIVITIES e SUBTASK 5.3 - INFORMATION EXCHANGE SEMINARS
- C0 ORDINATE WITH IPIRG-TAG MEETING
- TWO-DAY TECHNICAL MEETINGS
- OPEN OR CLOSED MEETINGS TO BE DETERMINED
- BCD STAFF MAKE 3 OR 4 PRESENTATIONS
- EDIT AND DISTRIBUTE PROCEEDINGS WITHIN 60 DAYS. i 1
OBattelle
mum nus ums uma muu uma igue men uma sumi mum mum uma uma um muss imm sums em IPIRG DELIVERABLES SPECIAL l - QUICK-LOOK REPORTS ON EXPERIMENTS (19 EXPERIMENTS)
- DATA RECORD BOOK ENTRIES ON IPIRG EXPERIMENTS (19 DRB) 3
- DATA RECORD BOOK ENTRIES ON DP 11 EXPERIMENTS (62 EXPERIMENTS)
- DATA RECORD BOOK ENTRIES OF OTHER PIPE EXPERIMENTS (35 EXPERIMENTS)
- VIDE 0 TAPES OF SUBTASKS 1.1, 1.2, 1.3, AND EDITED VERSION
- TOPICAL REPORT ON CYCLIC LOAD EFFECTS ON J-R CURVES
- IPIRG MEMBER MATERIAL PROPERTY DATA INCORPORATED INTO MEA COMPUTERIZED DATA BASE.
l OBaHelle
M M M M M M M M M M M M M m m m m m m EXPECTED STATUS OF IPIRG AFTER 3 YEARS TECHNICAL DEVELOPMENTS - a TASK 1.0 .
- EXPECT CRACK STABILITY UNDER INERTIAL LOADS TO BE LESS SEVERE THAN STATIC LOAD-CONTROLLED
- EXPECT DYNAMIC LOADING T0 INCREASE LOAD-CARRYING CAPACITY FOR MOST MATERIALS
- IMPACT OF " DYNAMIC STRAIN-AGING" 0N CRACK GROWTH RESISTANCE EVALUATED
- IMPORTANCE OF CYCLIC LOADING ON CRACK GROWTH RESISTANCE EVALUATED
- MARGIN OF SAFETY ON CURRENT ANALYSIS METHODS DEFINED FOR A REPRESENTATIVE PIPING SYSTEM (MATERIAL SENSITIVE).
e TASK 2.0
- COMPUTERIZED MATERIAL PROPERTY DATA BASE ENHANCED
- INTERNATIONAL PIPE FRACTURE DATA BASE ESTABLISHED.
e TASK 4.0
- VERIFIED METHODOLOGY TO PREDICT C' RACK-0PENING AREA AND LEAK RATES
- EVALUATE PLUGGING EFFECTS.
e TASK 5.0
- INFORMATION SEMINARS ENHANCE UNIFORM TECHNICAL POSITION.
unus mum umm uma muu uma muu umm mum ums sum umm mum uma sum uma imm muu e IPIRG PROGRAM STATUS PROSPECTIVE MEMBERS EXPECTED TO UNDER JOIN
- CONSIDERATION DECLINED UNITED KINGDOM FEDERAL REPUBLIC FINLAND OF GERMANY
; FRANCE ITALY BELGIUM JAPAN SPAIN CANADA ARGENTINA SWEDEN KOREA SWITZERLAND
- TAIWAN EPRI HAS INDICATED THAT THEY WILL JOIN.
OBaHelle N ]
-- -}}