ML20248K841
| ML20248K841 | |
| Person / Time | |
|---|---|
| Issue date: | 06/01/1998 |
| From: | Advisory Committee on Reactor Safeguards |
| To: | |
| References | |
| ACRS-T-3039, NUDOCS 9806100300 | |
| Download: ML20248K841 (152) | |
Text
tep$pp#p$pMWMMQMMW9@6 ka Jp pp M
gdR1d$sa$m$$$$m%0&%f#49N s !ge aae c hM8aggggggmW?msnaGWR 8 T A,s,J9 _4 52 9_2,l8 _fSB S_ _
a 4 b.yggger?;;pman em2Hpssgi 4 wygregggesqw m, casg%oscssy p r e p M eam a a s 1 s
pay 9 1*< io gWl,$ph a? age $)$3% %@q$d
.p w
osm s j; 9
Ap)q&"ap
.g4 N4 (Acns)
RETURN ORIGINAL
@pp gc
$g$g
%QH2e
!gg{ k [g &
sg$
N Oq
.. s e, 4ggf$hs j;k hg(w!
fp n
?3 h
g@f To sawnIrs p;
g kgh gh H em
~ag m Nxs>
ha, ifj 4E L j Jgp%q$esesa$ng$bomA.,,ygg ppany h
he og %a$Q; glqpyfp$$m$g$gQdQp%m
}#
a p
AAg nl 3 L g]g[jjj4 7grd fgi qg Q
aa m
sQ g Q4d dih f]
gpsw spe eWw gd s+,ag$ g$ gp emmemepena pe 45 S g
4agw d
$fj$},ga$gj%I$glg l hh
"/8 4w- #gypogwppusewm$ lit. egg @
,S, _
e,i
"; F 7"5 y
0mf eeg4w%
_et --___
N ag $uangag g g u g _g g geo61ooaoo 9806 1 m
v
'E von scas e?
Mn%v o
3Q % [ M W ??W05 W5$$gm%uendyt[gav w?
O T-3039 PDR
- 1 gu e
yp g:%
sp6%Q*J MW'
%*: jam N
N*
^~ ~
AdMSl-So39 OFFICIAL TRANSCRIPT OF PROCEEDINGS I
O..
i V.
NUCLEAR REGULATORY COMMISSION ADVISORY COMMITTEE'ON REACTOR SAFEGUARDS o
l l
I t-
Title:
. SUBCOMMITTEE ON MATERIALS AND METALLURGY l
j!
TRC4 ACRS Docket No.:-
Rsrar ca rs::;u T': rewa:Ts u
M/S T 2E26 l
415 7 32 l
THMJV C '
Work Order'No.:
ASB-300-313.
.!j!
LOCATIO s
' Rockville, Maryland f
DATE:
Monday, June 1,1998 -
. PAGESi 1 - 149 C
s18 0c,,,ceto:y-R_e:a,r t
g ri
/
10'Ze le0": 50DOIT mi'.73e ANN RILEY & ASSOCIATES, LTD.
'T 1250 I Street, NW, Suite 300 Ac v3,L.OOf
'. Washington,D.C. 20005 7
c.
Y (202) 842-0034
- \\ )
f.
9806100300 980601 PDR ACRS T-3039 PDR a-
t-r-S v
DISCLAIMER i-UNITED STATES NUCLEAR REGULATORY COMMISSION'S l
ADVISORY COMMITTEE ON REACTOR SAFEGUARDS l
JUNE 1, 1998 The contents of this transcript of the proceeding of the United States Nuclear Regulatory Commission Advisory O
- (f Committee on Reactor Safeguards, taken on June 1, 1998, as l
reported herein, is a record of the discussions recorded at the meeting held on the above date.
This transcript had not been reviewed, corrected l
\\
and edited and it may contain inaccuracies.
l 4
l 1
l I
~ (
L____-_
1 1
UNITED STATES OF AMERICA
()
2 NUCLEAR REGULATORY COMMISSION 3
ADVISORY COMMITTEE ON REACTOR SAFEGUARDS 4
5 6
SUBCOMMITTEE ON 7
MATERIALS AND METALLURGY 8
9 10 11 U.S. Nuclear Regulatory Commission 12 11545 Rockville Pike 13 Room T-2B3 14 Rockville, Maryland O
\\,4 15 16 Monday, June 1, 1998 17 18 The Committee met pursuant to notice at 1:31 p.m.
19 20 MEMBERS PRESENT:
l 21 WILLIAM SHACK, ACRS Chairman 22 ROBERT E. UHRIG, ACRS Member i
23 MARIO FONTANA, ACRS Member i
24' THOMAS'KRESS, ACRS Member 25 DANA POWERS, ACRS Member
(
ANN RILEY & ASSOCIATES, LTD.
Court Reporters 1250 I Street, N.W.,
Suite 300 Washington, D.C.
20005 (202) 842-0034
(,
L
2 1
PROCEEDINGS
.G 2
(j
[1:31 p.m.]
3 CHAIRMAN SHACK:
The meeting will now come to 4
order..This is a meeting of the ACRS Subcommittee on 5
Materials and Metallurgy.
I am William Shack, Chairman of 6
the Subcommittee.
7 The ACRS members in attendance, when they get 8
here, are Mario Fontana, Thomas Kress, Dana Powers and Bob 9
Uhrig.
10 We also have'in attendance Mr. Jerry Eisenberg, 11 Director of Nuclear Codes and Standards for the American 12 Society of Mechanical Engineers; Mr. Richard Barnes, 13 Chairman of the ASME Code Section 3 Subgroup on Design; j
14 Professor Asada from the University of Tokyo; Mr. Yamazaki f'
(_)\\
15 from Hokuriku Electric Power Company; Dr. Larry Shao,
'16 Director, Division of Engineering, Office of Nuclear 17 Regulatory Research; and Mr. Richard Wessman, Chief 18 Mechanical Engineering Branch, Office of Nuclear Reactor 19 Regulation.
20 The Subcommittee will discuss the NRC staff i
21 concerns regarding the changes to Class 1, 2 and 3 piping 22 system design requirements contained in the 1994 addenda, 23 Section 3 of the ASME Boiler and Pressure Vessel Code, and 24 the status of resolution of these concerns by the ASME 25 Special Working Group on Seismic Rules.
ANN RILEY & ASSOCIATES, LTD Court Reporters
(
1250 I Street, N.W.,
Suite 300 l
Washington, D.C.
20005 l
(202) 842-0034 i
[
I 3
(
1 The purpose of this Subcommittee meeting is to I\\
2 gather information, analyze relevant issues and facts and Q).
3 formulate proposed positions and actions as appropriate for 4
deliberation by the full Committee.
5 Noel Dudley is the Cognizant ACRS Staff Engineer 6
for this meeting.
7 The rules for participation in today's meeting 8
have been announced as part of the notice of this meeting 9
previously published in the Federal Register on May 12th, 10 1998.
11 A transcript of the meeting is being kept and will 12 be made available as stated in the Federal Register notice.
13 It is requested that the speakers first identify 14 themselves and speak with sufficient clarity and volume so
)
15 that they can be readily heard.
16 We have received no written comments or requests 17 for time to make oral statements from members of the public.
18 We will now proceed with the meeting and I call on 19 Dr. Shao to begin.
20 DR. SHAO:
My name is Larry Shao, I am the 21 Director, Division of Engineering Technology, Office of 22 Research.
The purpose of today's briefing is to present to 23 the Subcommittee our research and NRC position related to 24-ASME revised piping seismic design criteria.
25 The'ASME revised piping seismic design criteria L ()
ANN RILEY & ASSOCIATES, LTD.
Court Reporters L
1250 I Street, N.W.,
Suite 300 l-Washington, D.C.
20005 (202) 842-0034 1
4 1
were issued in 1994.
The staff had major concerns and a l}
2 research project was initiated to review the technical basis 3
for the new criteria.
Now our research has been completed.
4 Today, in addition to the staff's presentation, 5
there will be presentations by our contractor, the peer 6
review group member, ASME and our colleague from Japan.
7 The first presenter is Dr. Nilesh Chokshi.
8 MR. WESSMAN:
Also -- my name is Dick Wessman, 9
Chief of the Mechanical Branch.
With Nilesh Chokshi is 10 Kamal Manoly.
Kamal Manoly is with the Mechanical Branch in 11 Reactor Regulation and, of course, Nilesh Chokshi is with 12 the Office of Research, and these two individuals have been 13 the NRC liaison representatives to the ASME Special Working 14 Group for the last couple of years, and,lof course, are h
15 closet to the relatively recent activities, as well as the 16 older history.
17 And Nilesh will do the majority of the 18 presentation for them, and re-orient us on some of the early 19 history of the issue and then some of the current activities 20 that have been going on between the staff and ASME.
21 MR. CHOKSHI:
Good afternoon.
And I have been 22 introduced now and what I am going to do this afternoon is 23 to give an overview of the issue, the research status, where 24 we are today, and the details of the technical studies we 25 have conducted and findings will be presented by Ken Jaguay ANN RILEY & ASSOCIATES, LTD.
Court Reporters 1250 I Street, N.W.,
Suite 300 Washington, D.C.
20005 (202) 842-0034
l 5
1 of ETEC and Dr. Bob Kennedy.
/~
(
2 Before I start my presentation, I am going to l
's J
\\
3 restate which has been already stated, why we are here.
I I
4 mean what is'the issue?
And the basic issue is that ASME l
5 has revised the piping seismic design criteria and NRC has 6
not accepted it.
There have been some technical concerns.
7' So the focus of the discussion is basically on this issue 8
and we have -- what our studies suggest where the issues are 9
today and what we are doing.
l 1
l 10 In order to give an overall perspective, I would i
11 1 like to briefly start with some background on the piping 12 seismic design criteria, a little bit of history, evolution
)
13 and chronology, how we are -- we wound up where we are 14 today, and briefly describe the new criteria, our concerns,
/N l
()
15 all in a very global sense.
I want to leave the details to 16 the next two presenters.
17' So let me start with the background.
The seismic 18 loads and as indicated here, the pipe break loads, primarily 19 affect the number of supports and direction of supports and l
20 types of supports.
And there has been a big concern that 21 the seismic loads impose -- they are too conservative and 22 require much more supports than you really need them, or
'23 what the experience indicates.
24 The seismic design' practice, particularly for 25 nuclear industries, were developed early in 1970s and during r
[ l' ANN RILEY & ASSOCIATES, LTD.
\\- ' -
Court Reporters 1250 I Street, N.W.,
Suite 300 Washington, D.C.
2000E (102) 842-0034 l
L-- - -- ----- ------
- - - - - - - - - - _ - - - - - - - - - - - - - - - - ------- - _ - -- a
6 1
the course of early '70s and when the designs were being
(
)
2 carried out, it was known that there was significant 3
conservatism.
4 And when I talk about conservatism or margins, 5
there are two parts to this.
One is in determining the 6
loads or response side, given the earthquake ground motion, 7
how you determine what are the loads on the piping. Second 8
is how you are'using the core to size the piping, the 9
capacity or the strength side.
And it's I think important 10 to remember that overall margin is these, of these two 11 factors, what you do in response side, what you do in the 12 capacity side.
And the core is basically addressing the 13 capacity side of the issue.
14 Because of this recognition of this large A(,)
15 conservatism, in early '80s NRC formed a Piping Review 16 Committee and PRC also formed another review committee and a 17 number of activities were taken after the recommendations 18 from these two committees.
Also, a number of changes have R19 occurred in the response side of the calculations.
20 Generally, the analysis procedures, responses, ramping 21 values are in the standard review plans or regulatory 22 guides.
Allowable stresses and things like that are in the 23 code.
And over the years we have made a number of changes 24 on the response sides.
And, in fact, we are in the process 25 of making more changes to reduce conservatism.
O ANN RILEY & ASSOCIATES, LTD.
Court Reporters 1250 I Street, N.W.,
Suite 300 Washington, D.C.
20005 (202) 842-0034
7 1
An example is last year we publishing a new citing (G
2 rule which basically eliminates OBE from the design.
We are 3
currently looking at the approaches for the coupled analysis 4
of building and piping which will reduce a significant 5
conservatism from the response side.
In this viewgraph I 6
have shown a couple of examples like the higher damping 7
values which have been allowed to use -- using the 8
responses.
9 But we are going to talk about today, and what I 10 think when you hear Dr. Kennedy's presentation is on what is 11 the capacity margin and how does the core relate to that 12 core allowable stresses.
13 Some of these changes have already been 14 implemented in some plants where the licensees have gone and 15 removed some excessive snubbers and supports.
So attaching 16 this as a part of the PBRC and PRC recommendations in the 17 early '80s, both EPRI and NRC, mainly EPRI, undertook a 18 large scale testing program of components and NRC, also did 19 some complementary system tests.
These tests were conducted 20 at ANCO and NRC tests were conducted at ETEC.
21 These tests were basically completed in the late 22
'80s and then GE was analyzing the data and had developed i
23 some recommendations for the code changes.
Also during the 24 same time, the Advanced Reactor Corporation was developing 25 the criteria for the advanced light water reactor design.
ANN RILEY & ASSOCIATES, LTD.
d l
Court Reporters 1250 I Street, N.W.,
Suite 300 Washington, D.C. 20005 (202) 842-0034
8 1
-They also had proposed some changes which were -- and p
J 2
Westinghouse had proposed to use those criteria for AP600.
3 But because of the same similar concerns which were for what 4
was being developed by GE, the staff had basically not l
5 accepted Westinghouse proposal, and then Westinghouse had f
6 withdrawn the proposal.
l 7
A lot of this discussion became more focused in 8'
the early '90s when the ASME established a special task 9
group and with -- there were two NRC members also on that 10 group, and there were the discussion of many issues.
11 Basically, in this task group the NRC was in the minority 12 and did not -- and voted negatively to approve the changes.
13 The rest of the special task group had approved the task 14 group report, which formed the basis for some of the
!.(p) changes.
15 16 The rules were published in addenda and I think 17 now its known that we did not accepted.
We informed ASME in 18 December that because of this ongoing water research work, 19 at'present time NRC wasn't able to accept the criteria.
20 Subsequently, there was another letter explaining 21 the basis for the staff's concerns and to give ASME full 22-understanding of what were the concerns.
And then there was l
23 a meeting between NRC and ASME.
Subsequently, there is NRC l
24
-- ASME established a special working grc p, of which Kamal 25 Manaly and myself were appointed NRC observer liaison 1
l ANN RILEY & ASSOCIATES, LTD.
\\-
Court Reporters 1250 I Street, N.W.,
Suite 300 Washington, D.C. 20005 (202) 842-0034
9 1
members and the activity of this work is ongoing and we are s
(
)
2 going to hear from Mr. Barnes about that.
And I will also
\\_/
3 talk about some more interactions we have had with this 4
particular group.
5 With this background and sort of chronology, how 6
we wound up where we are and what has happened.
Let me get 7
back into some more issue oriented discussion.
Again, to 8
repeat, in 1994 ASME published the new criteria.
I am not 9
going to go into any detail.
But, basically, these criteria 10 were -- one of the things they did was to increase allowable 11 stresses significantly higher.
And the basis for this 12 recommendation, or basis for these changes were the tests 13 which were conducted by EPRI and ANCO.
And one of the 14 underlying assumptions in these changes was that the piping r~N.
(,)
15 fails in fatigue and fatigue ratcheting and not collapse 16 mode, and that was one of the major issues in the 17 philosophical and also in implementation of how does the 18 piping fail.
19 Again, I am going to highlight our concerns then 20 in a very sort of -- in a very simplistic way.
And one of 21 the concerns was that the test data did not seem to us to 22 indicate conclusively that piping system collapse will not 23 occur.
You will hear more about the detailed studies and 24 some of the other things we have done since then.
25 There was concern that as the systems become very 1
(
l l
[\\-J ANN RILEY & ASSOCIATES, LTD.
Court Reporters 1250 I Street, N.W.,
Suite 300 Washington, D.C.
20005 l
(202) 842-0034
10 1 :-
flexible,- there are consequences associated with some large lT 2=
deflections which had to be studied further.
V 3.
Another concern was that if allowable stresses are 4'
very high during-the inspection, it would now be acceptable 5
to -- it may not be -- one may not be able to accept even 6'
minor flaws and then there may be -. piping may have to be 7
replaced.
8' One of the issues was that what should be the 9
minimum design margin.and what should be the minimum code 10' margin, and how one-goes about establishing that.
~ 11 Jmd, finally, there was also a question of the 12 matter which the special task group had used to address the 13:
issues of what happens when.the piping, if you have 14-different input other than what was testing in the -- what
' i I
15 was used in the test.
If the systems were different than lu5
, what~was testing, how do you extrapolate, what are the 17' margins for those systems?
And here there is one big issue 18 underlying this so-called frequency effects.
All the tests 11 9 -
were conducted at their resonance level.
And what happens 1
20 Twhen:your' system is not at the resonance.
1 21 So given with this, all this underlying concerns
'22 Land; ongoing dialogues and controversies, we initiated a 23 research program.in 1993, primarily with ETC.
I would like i
24c
~ to a little bit digress.
When we started this program, ETEC 25, was a department of Energy -- as of 1996, it ceased to be 1/
ANN RILEY & ASSOCIATES, LTD.
Court Reporters 1250 I Street, N.W.,
Suite 300
{
Washington, D.C.
20005 i
(202) 842-0034
___-___L-
11 1
the department of Energy -- and after '97 we basically could 2
not -- and Boeing, also in the meantime, Boeing acquired 3
ETEC through Rocketdyne and when we had contracts there were 4
a lot of problems.
So if you see in the report, CASCO 5
Services, this is Ken's personal contract and so just to --
6 you will hear about the report, everything attributed to 1
l 7
ETEC, but the report, in the report it will say CASCO 8
Services.
9 On a more tragic side, our first principal i
10 investigator was Dr. Paul Chan and he passed away under very 11 unfortunate circumstances and then Ken Jaquay took over the
)
12 project.
So if you haven't had to chance to look at the 13 report, you will see the reference to the poles what is the 14 connection between ETEC and CASCO.
V(3 15 Okay.
We started the program at ETEC, basically, 16.
this is the flip side of the concerns.
We wanted to address 17 our concerns, so we established the objectives, and one of 18 the questions was, how do you establish minimum design 19 margin?
And you are going to have a detailed discussion
.20
.from Dr. Kennedy on that one.
21 One of the important items was looking at the test 22 which EPRI and ANCO had conducted.
There were many 23.
questions related to how do you interpret this data, how do 24 you infer, for example, the movements at certain section
'25 from the tail measure data?
How do you make adjustment from l
O ANN RILEY & ASSOCIATES, LTD.
Court Reporters 1250 I Street, N.W.,
Suite 300 Washington, D.C.
20005 (202) 842-0034
12 1
the nominal strength?
What kind of adjustments do you need
[)
2 to make?
How do you evaluate d&ta to arrive at margins?
v 3
I mentioned earlier that was concern with a method 4
used to address the extrapolation issues.
What happens if 5
the system -- how do you extrapolate the insides, the l
6 distaat systems which were not tested, or the components 7
which were not tested.
These studies were conducted by 8
Professor Iwan and Dr. C.T. Wong at Caltech.
i 9
And one of the other important part of the 10 research was to have a peer review group, and there was also 11 a staff working group who sort of looked over ETEC's work, 12 looked at it independently of the types of issues being raised, what the study is telling and make recommendations.
14 And I am going to tell you, I'll put up the names of the p)
(,,
15 group on the next viewgraph.
16.
The peer review group consisted of Professor Iwan 17 of Caltech.
Professor Iwan is a seismic engineering, 18 structural dynamics, mechanical engineer, and I think -- of 19 course, Dr. Kennedy is the architect, engineering, seismic, 20
. PIU4 and many other, I think.
Then Everett Rodabaugh, I 21 think anybody who knows piping, I think, doesn't.
And Larry 22 Shipley is_more of the expert in the design, with architect 23 engineering type of experience.
Ed Wais is Section 3 Code 24:
member and also piping expert.
And, of course, Dr.
25 Wilkowski, materials background plus Section 11.
l
()
ANN RILEY & ASSOCIATES, LTD.
Court Reporters 1250 I Street, N.W.,
Suite 300 Washington, D.C.
20005 (202) 842-0034 l
_ _ _. _ _ _ _ _ _ _ _ _ _ _ _.. _. _ _ - = _ _ _ _
13 1
In the staff working group, it was myself, Kamal
()
2 Manoly and many of the staff experts in the piping.
3 One of the things I think as a project manager, I 4
found this, having these groups, both peer review groups and 5
staff working group, extremely rewarding in many ways.
I 6
think we were able to flush out the issues, discussing that 7
form strateg'r, and there was a lot of guidance given on how 8
to proceed.
I think having this mix of people who had 9
practical experience, knew codes very well, was really sort 10 of -- at least I felt it kept us on a very -- focused on the 11 issues, and that was very important.
12 In particular, with John Fair from the staff 13 working group, with his vast experience in the design and 14 actual knowledge of the codes and things, helped us out in
'7-,
15 many ways, looking over developing models for the Caltech 16' studies to make sure that we were treating real problems 17 when we do parametrics.
18 But I think from this experience I am very much I 19 think proponent of peer review groups.
20 Well, now, I think, as Mr. Shao mentioned, the 21 project is complete, you have received the report, which is 22 basically, it's going to be published as it is.
In fact, I 23 had hoped that it would have been out in blue cover by now.
24 One of the basic conclusions, that as the 1994 25 addenda stands, there are still, there are some lingering --
(
ANN RILEY & ASSOCIATES, LTD.
Court Reporters i
1250 I Street, N.W.,
Suite 300 Washington, D.C.
20005 j
(202) 842-0034 i
I
14 1
some issues remaining and that, you know, we can -- basic
(~s
(
)
2 recommendation is that it is not to accept in the current v
3 form.
That there is insufficient technical basis for some 4
positions.
5 If you look at the Part 4 of the report, there are 6
many issues being discussed.
7 CHAIRMAN SHACK:
That report seemed to focus 8
completely on the EPhI tests.
Are there Japanese tests or 9
other tests that are really relevant that should have been 10 considered in that analysis, you know, when you came to that 11 conclusion?
12 MR. CHOKSHI:
When we looked at few other tests 13 which were available at that time, and Ken can elaborate on 14 that, but there was a test in England, Beany test, and there (Q_)
15 was some test at Liverpool, and I think Japanese test, 16 recent test, probably will have, you know, add a lot to see 17 how'we go forward on this.
18 But at that time we looked at some of the tests 19 available, not all of them.
And we relied also on the 20 investigations we performed.
But I think this new data 21 coming out will probably shed a lot of, you know, further 22 light and maybe help, able to help resolve a number of 23 issues.
24^
I am not going to spend too much time on this 25-because you are going to hear from -- particularly on third l#
ANN RILEY &.'*SSOCIATES, LTD.
Court Repercers 1250 I Street, N.W.,
Suite 300 Washington, D.C.
20005 I
(202) 842-0034 l
15 1
and fourth from Ken and Dr. Kennedy.
And I think that what
()
2 I would like to maybe focus on the last two is to that.
I 3
think this -- one of the important things comi'ig out of this l
4 research is this I would say common understanding of issues.
5 I think everybody -- I feel that everybody knows what the 6
issues are, what are the underlying concerns.
And I think 7
technical agreement, I think, to me, it has been -- now, I 8
think there is a question comes in, how do you use judgment 9
where there are uncertainties, and that's where, how this 10 unfolds remains to be seen.
11 Dr. Kennedy will talk about this framework which 12 is a part of the program, looking at the data and Caltech 13 studies, you know, he has developed, and which has been also la in the report, and which the special working group is also 15 looking at it.
16 One other things I think since the group was 17 formed, we have been in close contact with the special 18 working group and presenting results of our research as 19 things progress.
Ed Wais is a member of the special group 20 so there is a direct pipeline.
Then Bob Kennedy and Gary 21 Wilkowski both have attended meetings, and Ken, Paul Chan, 22 myself and Kamal have.
I think between us we almost 23 attended all of the meetings.
24 You know, I probably took more time than I should 25 have.
ANN RILEY & ASSOCIATES, LTD.
Court Reporters 1250 I Street, N.W.,
Suite 300 Washington, D.C.
20005 (202) 842-0034
16 1
Okay.
So after -- when this work was complete, we
,,)
/
2
\\/
wrote to ASME to let them know what the -- the research 3
findings and basically said this sort of conforms our 4
initial issues and there is an information basis, but that 5
we will -- the special working group has all of the research 6
results and we will continue to work with the consensus 7
process as the further deliberations take place and whatever 8
the recommendations are being developed.
9 In the recent proposed rulemaking in 50.55 (a),
10 which is out for public comment, we had to -- we can't stay 11 silent on it, so we had to take a position, and the position 12 was to, at this time, not endorse that.
13 If there are no questions, Ken Jaguay will first 14 describe the -- basically the evaluation of this data and
()
\\,/
15 the studies and some of the issues in a little bit of 16 detail, plus his findings and recommendations, and then Dr.
17 Kennedy will' discuss more of the proposed framework.
18 CHAIRMAN SHACK:
Any questions or can we let Mr.
19 Jaquay start?
20
[No response.]
21 MR. JAQUAY:
I would rather not be here today.
I 22 would rather Paul Chen was here.
He's a much better speaker 23 and I'll miss my friend.
I 24 The name CASCO -- I really don't have an acronym.
25 It's my wife's company.
We finished this NUREG as an
,(Gy ANN RILEY & ASSOCIATES, LTD.
Court Reporters 1250 I Street, N.W.,
Suite 300 Washington, D.C.
20005 l
(202) 842-0034 l
l
)
17 1
editing service, taking the ETEC report, which was in fairly 2
good shape but it needed some draft corrections and 3
illustrations and things and it was contacted to put out a 4
final document.
So the name CASCO will appear on the NUREG.
L 5
I asked my ETEC management, you know, they should I
l 6
-- I do work at ETEC.
I am an ETEC employee.
If they.would 7
like to have their name on the report, and because they 8
couldn't put a contract in place with the NRC, they though 9
it was awful difficult to explain who they had to -- how i
10 they have a report and they don't have a contract.
But, you 11 know, this is ETEC work reported under my wife's company.
12 CHAIRMAN SHACK:
We won't tell the IGC if you --
13
[ Laughter.)
14
.DR.
SHAO:
IG was sitting here.
15' MR. JAQUAY:
I hope I'm legal.
16
[ Laughter.)
17 DR. POWERS:
Gosh, Larry, you're awfully sensitive 18 to that.
19
[ Laughter.)
20 DR. SHAO:
I know him personally.
21
[ Laughter.)
22 MR. JAQUAY:
My presentation today will go through 23 these items beginning with a focus on the NUREG, and that's 24' where the detail and data is, the information, the arguments i
25 well laid out, it's written well, I feel, as best as Paul
(
ANN RILEY & ASSOCIATES, LTD.
Court Reporters v
1250 I Street, N.W.,
Suite 300 Washington, D.C.
20005 (202) 842-0034 1
.__________-____a
18 1
Chen and myself could express the issues, and makes the case
(~ \\
2
'through the conclusions of.today's presentation.
V 3'
Included in my briefing today will be a background 4
for those'not too familiar with the new rules, a very brief 5
background, hit the high points, and also in fairness hit l
6 the'high points and the technical basis that we'11' offer why 7..
the new rule should be accepted, and that background will be
~
8, valuable, I-think, to some of you,.and'then go through the 9
special studies that were done under the seismic e.nalysis of 10 piping program -- this is the ETEC program -- and conclude 11 the briefing with identification of deficiencies and major 12 issues that came out of the ETEC studies.
13 The overview, called a road map -- I do hope you.
14 had a chance to thumb through briefly the NUREG.
When you
()
15 get'more time to look at it in detail, you will be able to, 16 maybe from this, understand where to me looking for what.
171 The Part 1 section of the report is somewhat 18 historical.
The beginning of the report was to look for 19 issues and try to identify what is important, and a 20 historical perspective of where the rules were coming from 21 with rule backgrounds, other rules being put.out there 22 besides these changed rules, initiatives from the industry 23:
and the NRC on piping structural issue, that status, we'll 24-
.present it, and then an overview of the program test 25 results.
I feel it's like a Reader's Digest version of the ANN RILEY & ASSOCIATES, LTD.
Court-Reporters 1250 I Street, N.W.,
Suite 300 Washington, D.C. 20005 (202) 842-0034
19 1
PFDR volumes, and that's the program, PFDR's, piping fitting
[~)
2 reliability -- I always trip on this -- the pipe -- pipe and 3
fitting dynamic reliability program.
We call it the PFDR.
4 We frankly found very little experimental problems 5
with the testing of the well conducted program for the scope 6
it was.
Our problems lay'elsewhera in the actual conduct of 7
getting the data.
I think ANCO and ETEC, myself, did a very 8
good job of getting data.
9 The new roles and the bases, that is bold because 10 I will be covering that in this presentation in detail.
11 Results of the ETEC review in the beginning -- and I will 12 talk about what we identified in the later studies, funded 13 studies at ETEC of areas that needed studying, and I will 14 present the results of the first margin study that was done
(
15 as an independent study by ETEC of what the margins might 16 be.
17 The appendices contain a lot of data in a very-18 consolidated, very good form, I think, for future
'19 researchers and people looking at this data.
It will save 1
20 them a lot of time.
21 The Part 2 is the special studies begun because of 22 the issues we identified.
We felt we needed to do some 23 additional work, and CALTEC, Dr. Iwan and one of his 24 graduate students, CT Wong, contracted to do the majority of 25
.that work, and I'll give you some results of that.
h ANN RILEY & ASSOCIATES, LTD.
N/-
Court Reporters 1250 I Street, N.W.,
Suite 300 Washington, D.C.
20005 (202) 842-0034 i
20 1
I will also be talking about some of the margins
()
2 that we got with better understanding of behavior coming 3
.from that result, and in that Part 2 are some very detailed 4
structural mechanics type.
Applied mechanics people, 5
dynamicists would appreciate it but probably no one else.
6 But that is an extremely intensive, well-prepared part of 7
the NUREG, and that's where the majority of the response 8
data lies.
9 Then we asked each of the peer review group 10 members that belong to the peer review to, at the initial 11 identification of issues back in part 1 when we came up with 12 the first review, we asked for their comments on that review 13 and then we asked for a commentary, you know, an open page, 14 whatever they wished, to record their thoughts on the new
()
15 rules and the technical basis offered, and we put those in 16 the appendix.
The date generally is a 1997 level.
But we j
i 17 were able, when this funding was available for me to 1
18 continue on as an individual outside of ETEC, to keep Ev 19 Rodebaugh and Bob Kennedy involved a little longer. So they 20 have in that section some more recent thoughts, some very l
21 good thoughts, that take benefit of the efforts, the later 22-efforts.
23 Then part 4 is the heart of the executive summary 24 basis, and it's a very late-thinking summary.
I would say 25 three months ago was the thought here.
And since then, I l
l t
/ h-ANN RILEY & ASSOCIATES, LTD.
\\ssl i
Court Reporters 1250 I Street, N.W.,
Suite 300 Washington, D.C.
20005 (202) 842-0034
21 1
have not felt a need -- that anything needs to be addendumed
()
2 or changed or added.
3 So that is, I feel, the final conclusions of this 4
effort.in summary-form, and the appendices attached are just 5
to help possible future resolutions that come out.
6 So I bett.er -- I'm not used to this -- I should
.7 stand up and that would force me to keep my viewgraphs in 8
line with my talk.
So I am done with the overview of the 9
NUREG and I would like to go now into some of the details.
10 A background, for those not too close to the new 11 rules, in essence, this captures, I feel, the main features 12 of the change, of the piping design rules for seismic.
They 13 have increased the allowable for an equation.
This equation 14 captures the pressure and the moment stresseF, and with 15' these simplified. stress indices for piping components, sets 16 a-limit that.is now higher than it used to be before the 17.
1994 addenda.
It used to be 3 S sub M and 2 S sub Y, the 18
-minimum of those two values.
They've now increased it to 19 fcur and a. half, and at some temperatures for some 20-materials, it may increase the allowable by a factor of two
' 21.
depending on the SY effect.
22 The most significant conditions imposed on the 23 design in order to use these rules are listed here.
The L
.24
. standardization of the analysis procedure, which lays down 1
25 the. exact, to some extent, assumptions used in the analysis O
ANN RILEY & ASSOCIATES, LTD.
Court Reporters 1250 ILStreet, N.W.,
Suite 300 Washington, D.C. 20005 (202) 842-0034
22 1
so' people are getting the same number, and a lot better p-2-
anyway.
3 There's also a B2 limit on weight stress that 4
shouldn't be too much of a burden, and it has its purposes.
5 The design should not be ultra stiff.
These figures are 6
--for those that have been around design analysis as a 7
designer, not a researcher, a designer, you would look at 8
response spectrum curves where, you know, things get stiffer 9
and get soft to the left, stiffer to the right, and they set 10 a limit, and we're going to use that in some other 11 viewgraphs.
RW, R omega -- I call it RW/R omega 12 interchangeably -- has to be greater than
.5.
13 What they're saying in that designer space is that 14 your natural frequency of your piping system has to lay to 15 the left of twice the value of the peak.
And so this is 16 acceptable designs over in here, and you can't have ultra 17 stiff designs.
18 Now, when I show some margins, we're going to work 19 more of a researcher space where off to the right is 20'-
increasing'RW.
This-is a frequency ratio now, if you would, 21 and the effect of frequency ratio on margin generally is to 22 increase the margin, but not always, near residence and drop 23 it_away from residence.
l 24 But in this statement here today, this RW greater 25 than.5 is in these values over here, and we don't need to ANN RILEY & ASSOCIATES, LTD.
Court Reporters 1250 I Street, N.W.,
Suite 300 Washington, D.C. 20005 (202) 842-0034 I
i
23 1
worry.
The designer is not allowed to design in them an RW b
2 less than
.5.
\\J 3
The basis offered for the new rules is on this 4
viewgraph in a very summary form.
The claim is that the 5
testing and analytical studies justified removing reversing 6
dynamic loads from the code collapse criteria.
7 Now, the code collapse criteria is equation 9 with 8
three S sub M and two S sub Y.
These are tied into limit 9
analysis and testing of monotonic collapsed data and things.
10 So people say equation 9, it was always, in most people's 11 mind, a collapsed equation.
12 They justified removing it because of the failure 13 mode not being collapsed but ratchet fatigue, and the 14 statement was the testing, the component test at ANCO (O
,/
15 confirmed that, and under reversing dynamic load, which is 16 seismic, reversing in direction, the things would ratchet 17 fatigue in the failure.
18 I have a viewgraph of those tests, and those are 19 cantilever -- we call them inverse cantilever tests where 20 you have a heavy weight on top of the string of components.
21 It's very rigid on here to put instruments on and 22 essentially make the load a moment load down here on the 23 test article, which is down below, and the sled is then 24 moving underneath it.
25 So those are some illustrations of some of the ANN RILEY & ASSOCIATES, LTD.
\\
Court Reporters 1250 I Street, N.W.,
Suite 300 Washington, D.C.
20005 (202) 842-0034 l
l 24 1
tests if you're not familiar.
2 They-claim that the one test that we felt -- that l
3-a' lot of people felt showed a collapse was just not
'4 representative.
It was not the type of thing you would see 5
in a system.
It was an inverted cantilever, and that's not 6
like a piping system and is not considered representative.
7 They claimed, then, that because the failure mode 8
'is ratchet fatigue, the best way to address it is to' kind of 9
preserve the equation 9.
It works if you preserve equation 10 9 at the higher limit.
We can protect against ratchet 11 fatigue if the limit is high enough.
12 They looked at the testing, came up with the 13 margins,.and found the one that had the smallest margin, 14 which is test 36, and they did analytical extrapolations to
-,G -
V 15-find what would be the worst case for off-resident l
16 conditions, and I think, as was mentioned, the testing was 17 done very, very close to residence, and what that means is 18-that most of the testing was done right up in here, and in 19 here, if_you would, in this space, and the margins generally 20 are higher.
21
-So if you wanted to make a case that you had high
- 22 margins, you've got to do testing down away from residence
-23 or extrapolate analytically, and at that time, they made the 24 decision to extrapolate analytically, and it is difficult to 25 test off residence.
. [v -
Court Reporters ANN RILEY & ASSOCIATES, LTD.
1250 I Street, N.W.,
Suite 300-Washington, D.C.
20005 (202) 842-0034
l 25 1
The extrapolations indicated that the margins on (nj 2
the stiff side of residence were greater than two and
%/
3 greater than four on the flexible side if you adopt certain 4
margin definitions.
I'm not going to argue with that at 5
this point.
Let's go on.
6 The major issues of technical basis were 7
identified, and we went through the whole thing.
The basis 8
was several reports and a lot of test data, and we came up i
9 with three areas that don't resolve all the issues, but the l
10 most important areas to be funded, we felt, for further 11 funded studies, and these are the component test margin j
l 12 extrapolations.
13 We ourselves at ETEC did a lot of work there and l
4 14 we had Ed Wais helping with some of the fatigue damage l
/~~N
(,/
15 models, and at CALTEC, we had the students and Dr. Iwan 16 overseeing some of the reactivity and leading some of it.
j l
17 In the level of acceptable margin, Dr. Kennedy was 18 asked to provide a way of determining what an acceptable 19 level of margin is, and we felt that Section 11 issues ought 20 to be at least looked into.
We hired Bill Wilkowski at 21 Bechtel to look into what the effect is of the new rules on 22 Section ll.
23 These special studies indicated a lot of 24 information that wasn't around at the time the rules were 25 first formulating in people's minds.
We've added, we feel,
[)
ANN RILEY & ASSOCIATES, LTD.
U Court Reporters 1250 I Street, N.W.,
Suite 300 l
Washington, D.C.
20005 (202) 842-0034 l
u__
26 1:
tremendously to.the understanding of dynamic response of pj )
2 piping with these studies.
3 I won't say this first one was anything more than 4'
just reviewing.
We looked at ETEC what the effect on
-5 margins would be if you made a few different assumptions 6
than~what was done by the people presenting the technical 7
. basis, and you can see a list there of some of the things 8
changed.
9' The most important thing changed was the B2 10 indice.
A B2 indice in that criteria reflects the effect of 11 a gross discontinuity or the geometry of a piping component 12
.under moment loading, and a higher indice means that there's l
1
.13 more-of that effect that reduces its collapse strength.
14 It's a limit' load type concept.
i'
'But the failures were occurring not in the l
-15 16-fittings, but they were occurring out in the straight pipe 17 for these teste shown on this viewgraph, and the use of a B2 18-indice when the failure'was in a straight pipe -- now,:the 19 B2 for a straight. pipe is one, and all of these tests shown -
R20 on this viewgraph failed outside the fitting.
We felt there
.21 was no way to justify a higher B2 for that test because the 22 test article was'really not failed, and it's the location of l
23 failure as a B2.indice that ought to be looked at.
- 24 So that was the major difference, and based on 25 just that assumption driving it and some other assumptions ANN RILEY & ASSOCIATES, LTD.
Court Reporters 1250 I Street, N.W.,
Suite 300 Washington, D.C. 20005 l
(202) 842-0034
i 27 1
on strength and others, we were able to show that the
()
-2 margins were not only low, even under one -- and a margin 3
under one means that under an earthquake, it would fail.
l 4
There is no margin.
But the margins with these assumptions 5-were not only low, test 36 was not the minimum, and so they 6
only looked at test.36, and we felt that they should have 7-been looking at other tests.
8 At that time, we called in Dr. Iwan from CALTEC 9-and his students to help us with some of the issues and 10
-other-things.
The point, I guess, made by my margins being 11 very low is that if you hire an expert to give you a number, 12-I could probably -- that expert could probably give you that 13 number.
If you told him what it should be, he can make the 14 assumptions necessary in his method to get that number, and
()
-15 it all depends:on just that issue.
E16 So there's major questions on what are the right, 17 the correct extrapolation techniques, and Dr. Iwan we felt 18 19 CHAIRMAN SHACK:
Can I ask a question?
20:
MR.~JAQUAY:
Yes.
Sure.
-21 CHAIRMAN' SHACK:
The fact that you're getting the 22 failures in the straight pipe location away from where you 23 would expect the stress intensity, is that really a dynamic
'24-thing, that if I did a' genuine dynamic 3-D finite element j
25 analysis, I'would really predict the high stress in that O
l ANN RILEY & ASSOCIATES, LTD.
Court _ Reporters
~
1250 I Street, N.W.,
Suite 300 Washington, D.C.
20005
-(202) 842-0034
28 1
pipe location?
()1 2
MR. JAQUAY:
I don't -- I think if you had a 3
correct lodge displacement, very -- a perfect analysis, you 4
know, it had all the features, you would probably see the 5
failure in the same location that the test occurred in.
6-What is happening is this low cycle fatigue, and this is 7
high plastic level of loading.
8 If you look at the testing done in monotonic tests 9
to try to get limit loads, they had troubles in those tests 10 keeping the failures out of the straights, and they had to 11 stiffen up the straights or put rigids in those so that like 12 the T would buckle and the straight wouldn't buckle.
13 So what's happening is that it's a cyclic failure, 14 a lot of the contribution to that failure is the gross
(~h
()
15 distortion capability, the collapse capability, and what's 16 happening is you're getting what I call plastic hinge kind 17 of going out in that weaker straight pipe before those 18 little peak stresses, you know,. build up very high in the 19 teens, so you have a lot of plasticity in the straight. pipe 20 and not that much plasticity in the T.
So the failure 21 occurs outside the -- so they had a lot of straight pipe 22 tests essentially with different base conditions, and weld 23 tests.
24 MR. BARNES:
Can I ask a question?
l 25 CHAIRMAN SHACK:
Yes.
Introduce yourself.
2 x
s ANN RILEY &-ASSOCIATES, LTD.
'N Court Reporters 1250 I Street, N.W.,
Suite 300 Washington, D.C. 20005 l
(202) 842-0034 lw.
29 1
MR. BARNES:
Richard Barnes, ASME.
/ T 2
Those diagrams that you showed with the -- the V
3 definition of the elements --
4
- MR. JAQUAY
This diagram?
5 MR. BARNES:
Yes, that diagram there, and the 6
crack and so on, are you calling that plastic collapse?
7 Like take test 12 -- is that what you're calling plastic 8
collapse?
1 1
9 MR. JAQUAY:
No.
That is not.
These are ratchet 10 fatigue failure locations.
11 MR. BARNES:
But they're in the straight pipe.
12 MR. JAQUAY:
They're in the straight, the cracking 13 through the wall eventually.
14 MR. BARNES:
And the top one is a crack, right?
()
15 MR. JAQUAY:
Yes.
Yes.
Nine and ten had 16 through-wall cracking in the straights; twelve also failed, 17
'through-wall cracking.
All of these except test 40, for 18 which there's no sketch, failed through ratchet fatigue 19 failure.
20 MR. BARNES:
Okay.
I thought you had said I
21 somewhere in that sentence that it was a plastic type 1
22 collapse.
l 23 MR. JAQUAY:
Yeah.
What happens is plasticity 24 gets introduced into this zone and that weakness that zone, 25 and so the -- if you did a longer test, a lower -- amplitude
' O ANN RILEY & ASSOCIATES,- LTD.
l Court Reporters ss 1250 I Street, N.W., Suite 300 Washington, D.C.
20005 (202) 842-0034 l
u_________________________________
30 1
of loading that took thousands of cycles, you wouldn't get 2
the levels of load that caused this to go plastic, but you 3
would get these little stress risers in the T to fatigue 4
crack in the T, and you would get your cracks in the body of 5
the T.
So, you know, the fact it's low cycle -- it doesn't 6
take many cycles to fail.
You have a lot more plasticity, 7
it goes into this straight, which is weaker than the T under 8
that situation.
9 MR. BARNES:
Thanks.
Thank you.
10 MR. JAQUAY:
There is a motive for reducing the B
)
I 11 indices of these fittings,.because if they're not -- if 12 they're not -- if they're stronger than the straight pipe, 13 maybe there's some place to get some relief there.
So Dr.
14 Kennedy will be presenting some data that might indicate l
g )
15 that's one way to take care of some problems.
16 In the dynamic cyclic response behavior, one of 17 the areas of uncertainty was for non-resident testing, what 18 would the margins be, and we needed some.way to predict 19 off-resident' margins.
20 Dr. Iwan in the civil industry had for years, you 21 know, been looking at hysteredic curves for building, 22
-largely building, which, guess what, they're inverted 23 cantilevers, you know.
So you had something that was very l
24 much like one of.these tests in a lot there's a lot of 25 civil engineering background for things that behave in l
1 4
[T ANN RILEY & ASSOCIATES, LTD.
\\s-Court Reporters 1250 I Street, N.W.,
Suite 300 l
Washington, D.C. 20005 (202) 842-0034
31 1
similar single degree of freedom inverted cantilevers, and
/\\
(
)
2 that's buildings, and his students were very good -- and
-3 himself -- with ways of modeling hysteredic behavior so that 4
you can predict a test with a fair degree of accuracy if 5
it's a simple single degree of freedom cantilever, and this 6
is an example of some of the capability, the hysteredic 7
behavior in a test, and this is a moment versus rotation 8
plot and it goes around in cycles during these tests, has l
9 this very difficult shape to model.
It's not bi-linear, l
I 10 it's not, you know, perfectly flat like a parallelogram.
1 11 It's got the curvature and everything, and you can see, you j
i 12 know, the analytical model that CALTEC was using could 13 capture this type of non-linear behavior.
This is the model 14 down below and this is the test up above, and this does a (3
/
_j/
15 much better job than had been done in earlier work in the 1
l 16 piping area.
(
I 17 What they do is they match the test point with the 18 model and they do it not only in moment rotation response, 19 but also in fatigue strength capability, ratchet fatigue if 20 you would because ratcheting did occur in these, and they 21 take the test data, which is illustrated in these blocks of 22 data, and up this axis is what we call -- when it reaches 23 one, this damage, this fatigue damage means it fails at one.
24 So we have what we call a cumulative fatigue damage.
25 You get more and more fatigue damage each test (h
ANN RILEY & ASSOCIATES, LTD.
g'w'j Court Reporters 1250 I Street, N.W.,
Suite 300 l
Washington, D.C.
20005 l
(202) 842-0034
(.
32 1
run, and finally, at the end -- in the middle in some cases
()
2
-- of the last test when it broke, he tunes his models, his 3
computer models, so he exactly predicts failure when the 4
test failed.
So he has both a hysteredic tuning and a 5
fatigue strength tuning.
6 We have a lot of -- it's good confidence in these 7
models as extrapolation tools a certain distant off the test 8
point, you know.
You can only extrapolate things so far.
9 He would then do one test run and then predict 10 another test run, and he got a lot of confidence that way.
11 By the test runs being slightly different, he could do 12 pretty good predictions of the different test runs.
13 The study was not just one worst minimum margin 14 case, but six different ones.
The studies included, besides O
i,/
15 the basic frequency effect, which I have shown, the s
16 off-resident margins being lower, but also the.effect of P' 17 delta stiffness which is related to, you know, how much 18 gravity contributes to restoring the moments.
'19 In an inverted cantilever, it's a negative P delta 20 stiffness, so that the P delta effect is to resist 21 restoration of the moment back to its original position and
'22 makes things worse, whereas an inverted cantilever, it would 23 be a positive P delta, that the gravity actually helps you 24 from getting off of equilibrium.
25 But the type of loading being narrow band versus
[
ANN RILEY & ASSOCIATES, LTD.
\\-
Court Reporters 1250 I Street, N.W.,
Suite 300
. Washington, D.C.
20005 (202) 842-0034
33 l
1 wide band was looked at, and that's very important.
Up high l [~ l 2
in the reactor sites, in the reactor containment, you have a l
\\_/
3 narrow band type of input; but down low in the ground and 4
close to the base, you have a wider band input.
5 Well, they only tested narrow band, and the effect 6
of the type of loading on response was -- we thought it 7
important -- he looked at that, and several other things, i
i 8
with some of the results shown here just illustrating the 1
9 differences you can get.
1 10 Now, if someone says what is the effect of 11 frequency on margin, the frequency effect is being studied 12 in this viewgraph.
This is the ratio, if you would, the 13 frequency ratio versus the amount of margin.
This is a 14 ratchet fatigue margin.
,m k j) 15 Now, the trend you can see is one way on this 16 particular diagram, but it's totally different on this 17 diagram, and the reasons for the difference is this is a 18 narrow band time history type and this is a wide band.
This 1
19 is maybe what you would expect margin trends to look like 1
20 low down; this is what they look like high up.
And then the 21 difference of how much weight load you have can affect --
22 gee, I'm increasing, you know, I'm decreasing.
And the 23 amount of variation in margin is such a function of so many 24 parameters, it's virtually, I feel, impossible for somebody 25 to write simplified criteria that says as a function of 1
[
ANN RILEY & ASSOCIATES, LTD.
ks/
Court Reporters 1250 I Street, N.W.,
Suite 300 Washington, D.C.
20005 (202) 842-0034
34 1
frequency ratio, I know what margins are on piping.
So then
()
2 we are stuck with a dilemma.
3 We hild a lot of information, we spent a lot of 4
tice getting it, and the information is telling us this is a 5
very complex field and you'll never have enough information 6
to resolve what's going on.
So the more we learn, the less 7
we know is an axiom and it's true in this case.
8 CHAIRMAN SHACK:
So we ought to stop all research, 9
right?
10
[ Laughter.]
11 MR-JAQUAY:
All right.
I think that view graph 12 is done.
Let me move on to this series in the handout.
13 With the knowledge gained, we didn't -- improved 14 margin prediction, and what we ended up doing is in these
(~
15 margin predictions looking at the worst case margin.
The 16 only thing we could hang our hat on is we know that when the 17 RW values get very small, the margins get very small, and we 18 could look at margins in a worst-case fashion with a lot of 19 confidence even though the trends may be difficult.
20 We redid the effort.
In this case, the margins 21 were done based on CALTEC understandings and everything, and 22 we used the test 14 as one test trend.
We then extrapolated 23 all the different tests by the trend of test 14.
2 4 Now, that, you know, obviously is not right.
25 Every margin trend in a different component could be
(
ANN RILEY & ASSOCIATES, LTD.
Court Reporters 1250 I Street, N.W.,
Suite 300 Washington, D.C.
20005 (202) 842-0034
35 1
different.
But we didn't have the margin trends for all the I\\
2 components.
But that and maybe some of the other Q
l 3
assumptions created a bad situation of a lot of scatter in 4
the data.
5 So here we are educated, we understood better 6
margin trends, we had the best data available, and still the 7
scatter was large.
Our data confidence in predicting 8
margins, with the best things we had, was still not there.
9 So Bob Kennedy stepa into the picture at this time 10 and almost regrettably -- obviously, you know, he's helping 11 us, but his message to us was, well, thank you for your 12 effort in showing you really can't do much with that, 13 because it makes the case for a method he was thinking of at 14 that time as maybe offering a solution, and it makes the
)
/~T
(,)
15 case for, you know, we need something that doesn't rely on 16 understanding margin trends that can be all over the place.
17 So his approach will be given to you after this, 18 and I respect it, I think it's a very good one.
I'm 19-referring more to the capacity effort down here.
20 He-also did a study -- I'm pointing to the right 21 thing -- this is --' excuse me.
I'm not used to -- I have 22 been looking at this view graph on the page, and this is the 23 one that'I wish I had in front of you.
24 This is the margin study that had a lot of scatter 25 and this is Kennedy's approach to establishing an acceptable j
1 l
Ow ANN RILEY & ASSOCIATES, LTD.
Court Reporters 1250 I Street, N.W.,
Suite 300 Washington, D.C.
20005 (202) 842-0034 i
36 1
level of margin you will hear about later, and this here is
()
'2
-- I respect and I think the majority, if I don't speak for 3
all of them -- you know, I can't because I didn't get it in 4
writing from one of the individuals -- but the majority of 5
the peer review group members are right solidly behind Dr.
6
_ Kennedy's approach as something to be looked at, and as a 7
frame work for discussions, and should be pursued and looked 8
into further as maybe a way to get us out of this dilemma.
9 It's somewhat of a dilemma that I described with the margin 10 trends going all over the place.
11 There was finally a study done by Bill Wilkowski 12 to make section 11 flow tolerance sensitivity under the new 13 rules better understood, and he wrote a PVP paper on that, 14 and I am not an expert in this area, but the essence was 15 captured in his paper in summary form that I could 16 understand, and it was essentially that if you increase the 17 stress limits to the new allowables and you find any flaws 18 larger than the code acceptable workmanship flaws, then you 19 have to replace ~the piping.
You would not be able to do 20 analysis to make a case for continued operation.
21-DR. SHAO:
May I say something on this.
The code 22 allowable is ten percent of'the thickness.
Anything less 23-than ten percent, you don't have to do anything.
If you're 24 bigger than ten percent of the thickness, you can do a 25 fracture mechanics analysis to show if the flow is ANN RILEY & ASSOCIATES, LTD.
Court Reporters 1250 I Street, N.W.,
Suite 300 Washington, D.C.
20005
~
(202) 842-0034
I 37 1
acceptable or not.
If the flow is arrested, then this flow 2
is acceptable.
3 So essentially what he's saying if you use a 4
^ larger allowable in the code, anything less than ten 1
5 percent, you cannot do a fracture mechanic t est to show that 6
it-will be acceptable. -Anything larger than ten percent, 7
the pipe had to be replaced or repaired.
8 MR. BARNES:
Mr. Chairman,.may I make a comment?
9 I was going to leave this to my talk, but I l
'10 actually -- it's going to take me a couple of minutes.
-11 Richard Barnes, just so that the record is there.
12 The equation that Ken is referring to, when the
'13-new. rules were written, was no longer an equation that
.l 14 discussed or' represented the failure mechanism which we i
-15.
believed was operative, the predominant failure mechanism.
16~
It's a failure -- that equation represents a collapse 17
' failure approach.
18 There was a discussion within the group whether
'19 they should rewrite an' equation to discuss the failure 20-mechanism as they understood it or leave it as it was and do 21' some sort of a manipulation with it as an empirical
'22 equation, because there were so many computer programs 23 associated with piping, they felt to go and change it would l
.24
-be to throw the industry into an upheaval.
That was the 25 sort of discussion that went on at that time.
ANN RILEY & ASSOCIATES, LTD.
Court Reporters 1250 I' Street, N.W.,
Suite 300 Washington, D.C.
20005
. (2 02) 842-0034
38 1
The unfortunate part about it is that everybody
()
2 takes it as being representative of the failure mechanism 3
instead of just as an empirical equation.
I don't think 4
anybody believes that you get 4.5 S sub M actually as an 5
allowable stress.
There is a different mechanism going on 6
there, and the section 11 approach is based on the fact --
7
-the way'I understand it anyway, the section 11-approach is 8
based on the fact that you have this constant stress.
You 9
don't have a constant stress; you have an oscillating 10 stress, it's a reversing stress in the predominant failure 11 mechanism.
12 So the way I see the section 11 problem is you 13 cannot make that logical connection.
What has to be done is 14-section 11 has to re-look at this failure mechanism and come
(
15-up with a new approach'to it and see what its impact is.
16 DR. SHAO:
Yes, but have you guys worked with the I
17 section 11 people?
18 MR. BARNES:
Well, that's one of the issues, 19 Larry, that people are working with.
What we have tried to 20-do is take the '-igher profile or the higher priority issues
'21 first and -- because of the ETEC work that has been coming 22 in.
But that particular issue, really it seems to me that.
23 we cannot draw any conclusions like are being drawn in that 24 area at this point.
25~
DR. SHAO:
Yes, but the thing is, whoever -- the i
ANN RILEY & ASSOCIATES, LTD.
[-
Court Reporters 1250 I Street, N.W.,
Suite 300 Washington, D.C.
20005 (202) 842-0034 L--__________---___
39 1
design people had to work with the inspection people, and
()
2 you don't want to come out with a design criteria the L
3 inspection people cannot live with.
4 MR. BARNES:
Yes, but can I just point out, see 5
that's meant to be a -- that's an empirical equation.
6 DR. SHAO:
Yes.
7 MR. BARNES:
You cannot say that there is a 4.5 S 8
sub M stress there.
You just cannot say that.
9 Now, I actually believe it was a strategic error, 10 what we did, I mean, because I believe people, no matter how 11 we talk about -- I mean, I've heard members of the 12 independent review team who I know should have known better, 13 in my opinion, use it like that, and I was astounded, to be 14 quite frank, and I have '3en saying to some of the people
(~
(,
.15-who are the experts in the area, I said, people are 16 misinterpreting this.
They are not using this as just a
.17 sheer' empirical equation.
It has no meaning.
You cannot 18 vary the variables and get it because it doesn't represent l
19 the failure mechanism.
l l
20 DR. SHAO:
But the thing is, Dr. Wilkowski is a 21.
member of section 11.
22 MR..BARNES:
Yes.
No, I agree with you.
We have.
i 23
'a problem there.
Now, I mean, it's unfortunate that the
'24-people who are really, really in-depth here couldn't have 25 been'here to explain it, but that was certainly my ANN RILEY & ASSOCIATES, LTD.
Court Reporters 1250 I Street, N.W.,
Suite 300 Washington, D.C. 20005 (202) 842-0034
40 1
understanding, and I have -- in fact, I keep on repeating it
(}
2 with these gentlemen to be sure that I'm understanding what 3
they're saying and --
4 DR. SHAO:
But what you're saying is this loading 5
here is the same as constant loading instead of alternating 6
loading.
But the things is --
7 MR. BARNES:
Yes.
Let me just finish.
What I 8
think we should do is, what we have to do is we have to take 9
-- get rid of that equation and put in an equation that 10 represents the predominant failure mechanism.
Now unless, 11 of course, with Dr. Kennedy's work, which people are working 12
.with, we can overcome that.
13
.But, I mean, we do have a problem, in my opinion.
14 that people are misinterpreting that as an exact r
15 representation of the failure mechanism, and it just can't 16 be.
17 DR. SHAO:
Another thing, you know, section 11
)
18 also takes into consideration stresses.
19 MR. BARNES:
Yes.
20 DR. SHAO:
This plus stresses, then you do the 21 calculation.
22 MR. BARNES:
Well, yes, it's an impossible 23 situation'for section 11, that equation.
There's no doubt i
24 about that.
If you take it as a literal stress, it's a very l
I 25 l
l i
ANN RILEY & ASSOCIATES, LTD.
Court Reporters 1250 I Street, N.W.,
Suite 300 Washington, D.C.
20005 (202) 842-0034 lw_---_---
41 1
MR. JAQUAY:
And try to apply it
()
2 MR. BARNES:
Yes.
It's an impossible situation.
3 And I also agree it's an impossible situation for you as a 4
regulator.
5 DR. SHAO:
Yes.
We cannot regulate that.
6 MR. BARNES:
I know.
I know.
I understand.
7' DR. SHAO:
People say NRC --
8 MR. BARNES:
I know.
It's a strategic -- it was a 9
strategic -- in my opinion, this is in my personal opinion.
10 I mean,.I may have people who would grab me by the throat 11 and say, "You're wrong, Barnes", and they would like to do 12 that at times, but the point is, in my opinion, it was a 13 strate;ic error because it's got repercussions that -- and i
14 we do have to address it, but the problem with code work is (O
,/
15 it's done by volunteers, and that's a time -- oftentimes in 16 today's market, _ people are doing it on their own time even.
17 But thanks.
l I
18 MR. JAQUAY:
Okay.
That concludes that section, l
19 just to return you to focus -- we are now through the first l
20 four bullets.
The final bullet is to be discussed next, the
]
I 21 findings.
j 22 I have heard the saying "trying tc ctay on the l
l:
23 same page" as everybody.
It's a little difficult reading u
24 and -- I think I am on this page.
25 Areas of deficiency indicate really a lack of i
ANN RILEY & ASSOCIATES, LTD.
I Court Reporters 1250 I Street, N.W.,
Suite 300 Washington, D.C. 20005 (202) 842-0034
42 1
attention or it's been glossed over.
There's not enough
()
2 given to it -- versus areas of issues.
Issues are things 3
that have some technical meat that we can look at and can 4
look into.
These areas of deficiency, there's like no l
5 attention paid.
6 We felt -- the first bullet, the area of 7
deficiency was the method of extrapolation was not given 8
enough thought.
The variable effect, sensitivity studies 9
should have been done -- how is your extrapolation to this 10 or that and the whole area of time, history, shape, p-delta, 11 stiffness, concentric -- e-centric weights were not really 12 looked at rigorously -- just not enough done here.
13 This was the thing that led us to suggest to Cal i
14 Tech that they should do the studies that would give us
)
15 enough information on the variables to see what margin 16 trends were driven by.
17 Probably the other negative I would say about the 18 technical basis, ratchet fatigue, margin, was it didn't 19 extrapolate considering ratchet interaction with fatigue, i
20 It was strictly a fatigue curve slope and they assume the 21 damage mechanism was fatigue in their extrapolation, not 22 ratchet fatigue, and so they got at a point of ratchet
'23 fatigue failure in the tests and then they extrapolated by a 24 strictly fatigue model.
25 Another area of deficiency -- we get to give me a i
,5 ANN RILEY & ASSOCIATES, LTD.
Court Reporters 1250 I Street, N.W.,
Suite 300 Washington, D.C. 20005 (202) 842-0034
43 1
little break.
I have a videotape I would like to show you.
()
2 It's four minutes -- if I have the time.
3 The failure mode is or isn't x or y has been a big 4
argument and I have come from a high temperature strength 5
background that we can fill things in many ways.
I don't 6
understand why seismic should only be able to fill something 7
one particular way and so they say this is the failure mode.
8 Well, that doesn't have to be a single thing.
You can have 9
more than one failure mode, but they were somehow adamant on 10 saying the collapse was not a failure mode, and they made a 11;
-big case for the ratchet fatigue, so maybe we could roll 12 this video.
13 Do I have to be involved in that?
14 MR. DUDLEY:
I just have to turn the overhead, and
)
15 we'll watch it.
16 CHAIRMAN SHACK:
How much longer do you estimate 17 you are going to need for your presentation?
18 MR. JAQUAY:
Am I running over?
I need -- I can 19 make it as quick as 10 minutes.
Is that okay -- on top of 20 the video?
21 Okay.
This particular test is a. Test 37.
This is 22 low frequency test.
I need to have the sound turned up on 23 it.
24 The words of the test engineer is as important, if 25 not more important than anything anyone is going to learn
[/)
ANN RILEY & ASSOCIATES, LTD.
Court Reporters s-1250 I Street, N.W.,
Suite 300 Washington, D.C.
20005 (202) 842-0034
44 1
from watching this test.
[')
2 Kelly Mertz and ANCO had been testing these
%J 3
components for months, recording information and writing 4
quick look reports, and making his test observations.
Now 5
this particular test is a minute and 15 seconds.
I started 6
right.at the beginning.
It is a very thin walled elbow, 7
Schedule 10.
It is not pressurized.
The weight stress 8
increases as it progressively moves over to the side.
9-Kelly Mettz is in the control room.
He turns the 10 machine off with his right hand.
If there is a problem of 11 facility danger or anything.
This test was destined to go 12~
30 seconds beyond.
He shut this test off early.
That is 13 Kelly there.
14 VIDEO SPEAKER:
This concludes Test 37, run 5.
As
()
15 can be seen, as I come a little closer, we collapsed it.
16 Okay, Bill, there you go -- now the ball is in 17 your park.
18 MR. JAQUAY:
Bill English was the Program Manager 19 of General Electric.
This is Test 39.
20 This test ran the full test duration.
It was run 21 ahead of Test 37.
22 VIDEO SPEAKER:
The first high level earthquake on 23
-the unpressured T.
24 MR. JAQUAY:
This is the first test that we have i
25:
had indications of partial collapse.
This deformation is ANN RILEY & ASSOCIATES, LTD.
Court Reporters j
r 1250 I Street, N.W., Suite 300 J
l Washington, D.C. 20005 L
(202) 842-0034 l
45 1
not located. is not due to the nipple.
It is localized in
/\\
2 the body of the T.
V 3
The area that I am looking at now is slightly 4
depressed.
We walk over to the other side and we have an 5
area that is bulged.
6-The last test I am showing is Test 40. This is not 7
thin-walled.
This is Schedule 40, 8
This one is also stopped prematurely.
9 VIDEO SPEAKER:
Well, Bill, there is your 10 collapse.
Put a nice buckle in it.
I'll show you that in a
'11 close-up in the video in just a minute --
12 MR. JAQUAY:
Too late at night, Bill, to give you 13 a clear, well-lit shots, being it's about 1 o' clock in the 14.
morning, but if you get the idea that it did collapse --
(%
(,).
15 okay, that's it.
16 DR. SHAO:
May I ask a question, Jim?
Did you do j
l'7 an analysis to show what the stress level during collapse?
18 MR. JAQUAY:
Someone may have.
I didn't.
I did 19 not run the stress --
20 DR. SHAO:
3 Sm or 4.5 Sm?
21 MR. JAQUAY:
The numbers could be determined.
I 22-don't know offhand.
23 DR. SHAO:
Another question.
If you have pressure 24 inside, does that make it worse?
l 25 MR. JAQUAY:
Pressure generally helps.
'(' )
ANN RILEY & ASSOCIATES, LTD.
~_)
Court Reporters 1250 I Street, N.W.,
Suite 300 Washington, D.C. 20005 l
(202) 842-0034 l-
46 1.
DR. SHAO:
Help less collapse or more collapse?
(
)
i 2
.MR. JAQUAY:
Less collapse problem.
More 3
resistive to collapse.
4 DR. SHAO:
Combination pressure and seismic 5
together?
6 MR. JAQUAY:
Pressure is good for moment capacity.
7 It's like a balloon.
8 DR. SHAO:
Make it stronger.
9 MR. JAQUAY:
Right.
10 That was a wall buckling on the compressive side.
.11-The pressure would have helped to vent that.
12 DR. SHAO:
Because the pressure c,i wels another 13 compression --
14 MR. JAQUAY:
It's more complex but --
(r),)
15 DR. SHAO:
Something like that.
16 MR. JAQUAY:
It is a stiffening effect and 17 strengthening effect on moment and particularly in elbows.
18 In fact, they would like to give it a negative stress indice 19 to elbows if they could.
They make pressure.
If you have 20 enough, you get relief but --
21 DR. SHAO:
Make it stronger --
22 MR. JAQUAY:
Yes.
But they said we'll just stop 23 it at zero, so most pressure stresses are ignored in elbows.
24 MR. BARNES:
Ken, the three tests that you showed 25 there, were they all unpressurized?
O ANN RILEY & ASSOCIATES, LTD.
Court Reporters 1250 I Street, N.W.,
Suite 300 Washington, D.C. 20005 (202) 842-0034
47 1
MR. JAQUAY:
Two of them -- all three.
Yes, sir, 2
al three were unpressurized.
3 MR. BARNES:
And did that phenomenon occur in any 4
of the pressurized tests?
5 MR. JAQUAY:
These are the only three tests we saw 6
this phenomenon.
7 DR. SHAO:
Okay, thank you.
Should take a look at 8
one with pressure.
9 MR. JAQUAY:
Pressure doesn't necessarily hurt and 10 most likely helps.
11 MR.'WESSMAN:
Excuse me.
Were there pressurized 12 tests actually conducted as part of this group or series?
13 MR. JAQUAY:
Yes.
We had pressurized tests 14 similar to 37 except for pressure.
Identical and it did not 15 have a problem.
16 Now if you have an unpressurized system, loss of 17 pressure, lor:0 of pressure boundary, if you have an 18 unpressurized system, it wouldn't necessarily be a safety 19 concern because there is nothing there, but I have heard 20 there are some cases that during the earthquake it doesn't 21 have fluid.
It's unpressurized, zero pressure, and then 22 some relief valve opens and it has to function, and it is in 23 some cases a thin type piping, and those things, you know, 24 they are not there --
25 DR '. SHAO:
When you collapse like that you may I
t-(d ANN RILEY & ASSOCIATES, LTD.
Court Reporters 1250 I Street, N.W.,
Suite 300 Washington, D.C.
20005 (202) 842-0034
48 1
hit -- it may impel on other pipes or if you have valves l9 2
attached to it --
< v.
3 MR. JAQUAY:
Bulge deformation -- large 4
displacements is a concern -- you know, two of them, one, 5'
valves, things like that.
6 Also system stability -- I consider a code 7
boundary failure -- your prediction of behavior if you don't 8
have strength, if you don't have stability your lineal 9
elastic predictions are going to be way off, if you allow 10 this behavior.
11 Now I promised you ten -- can I have five?
12 CHAIRMAN SHACK:
Take five.
13 MR. JAQUAY:
Okay.
Another area of critical 14 deficiency -- is it the methods, the arguments for margins Osj-15 were based on component data and what we have is systems, t
16 piping system and piping systems are different than 17 components and they can be less -- they can have less margin 18 and they can have more margin contribution in different 19 manners.
20 I feel that we need to look at system behavior 21 when we write rules for piping system and we need to go 22-beyond the component capacity data into some system response 23 data, so that was deficient.
24 Critical issues -- I talked about the B2 values in 25 the testing, being at the failure locations.
Well, that led f-
[
ANN RILEY & ASSOCIATES, LTD.
\\
Court Reporters 1250 I Street, N.W.,
Suite 300 Washington, D.C. 20005 (202) 842-0034 w___________________
49 l
1 us to believe that if you had used the failure location B2s, N.s straight pipe, and you didn't have a fitting next to it, 2
3 that the margins could be very low compared to elbows and 4
other fittings and so we are having trouble correlating our 5
data without some adjustment of the B2 values.
6 The adjustment could be higher B2 values for 7
straights or it might be lower B2 values for fittings, but 8
we need to line up the seismic capacity of those two groups 9
and Bob Kennedy will address that issue too, but my point is 10 that we have inconsistent capacity indicated between the 11 straights versus the ovals.
12 A high temperature issue exists also.
This is one 13 that we don't know it's a problem.
We just don't know it's 14 not -- which I hate to be that way, but I think it is 15 serious enough that we should know this is not a problem.
16 That is, at high temperature carbon steel has a 17 strain aging effect, dynamic strain aging, blue 18 embrittlement.it is called, that may or may not be 19 significant to seismic rates of load, but it has a tendency 20 to change deformation of historetic behavior.
21 It is a potential problem that needs further 22 looking at and only room temperature component tests were 23 run, and since it is strain rate dependent, the strain rates 24 vary in a component.
You can't use simple material data, 25 uniaxial data.
You need gradient data and it almost forces O
ANN RILEY & ASSOCIATES, LTD.
bn
.. Court Reporters 1250 I Street, N.W.,
Suite 300 Washington, D.C. 20005 (202) 842-0034
50 1
you to have to do some high temperature, higher temperature
[
'2 component tests to resolve.
N 3
A critical concern is the test articles didn't 4
have any damage and that needed to be addressed, not 5
necessarily we had the test damaged articles but it needs to 6
be addressed.
7 Then the rules don't seem to exclude some items 8
that may not have the capacity seen in the testing and 9
threaded connections, bi-metallic welds, nonlinear supports, 10 possibly some cast materials, and finally these rules are 11 not just applicable to seismic loading.
They can be put in 1
12 place for any reversing dynamic loading.
13 A seismic event is over pretty quick but one of j
14 the assumptions that we have in this thing is if ratchet (k
15 fatigue cracking is dominating, then the duration of loading 16 is an issue.
If you had something that lasted for an hour, 17 like a chugging event or something -- I don't know reactors 18 but I hear there's BWR chugging events that last hours -- is 19 this failure mode we are studying based on short duration 20 testing really the right failure mode?
21 Less critical issues I won't discuss verbally.
I 22 do feel'that there's other issues besides these.
One of the 23 efforts that ETEC has attempted is to try to filter out some
]
24 of the issues, try to get down to the~ critical issues.
l l
25 We still have some things we can point to and say l
(/)
ANN RILEY & ASSOCIATES, LTD.
N-Court Reporters 1250 I Street, N.W.,
Suite 300 Washington, D.C. 20005 (202) 842-0034
I 51 1
they should be looked at but I don't consider those -- the
()
2 need as bad as the need in the ones I have listed.
3 I am done.
4 DR. SHAO:
I have a question.
As far as 5
metallurgy is concerned, I thought -- it's not fast enough 6
to be considered dynamic as far as metallurgy is concerned.
7 They are not fast enough.
8 MR. JAQUAY:
Are you asking the speed of loading?
9 DR. SIUM):
Yes.
I am talking about metallurgists 10 and the dynamic strength.
-11 MR. JAQUAY:
Right.
The rate that we are talking 12 about is like a second to the negative one -- strain.
13 DR. SHAO:
Yes.
Really not very fast.
14 MR. JAQUAY:
No, it's not impulsive loading.
A)
(,,
15 Well, maybe that needs clarification.
16 When we say dynamic in my presentation today, it's 17 seismic rates loads, not blast impulsive loads.
18 DR. SHAO:
So you are 15 percent -- impulsive 11 9 load, fast dynamic or slow dynamic?
20 MR. JAQUAY:
No, that's slow dynamic, seismic l
21 level.
Gary was using rates -- Gary Wilkowski was using 22 rates, the seismic.
23 CHAIRMAN SHACK:
If there are no other questions, 24 I think we will take a 15 minute break before we resume the 25-subcommittee meeting.
O ANN RILEY & ASSOCIATES, LTD.
\\- I Court Reporters 1250 I Street, N.W.,
Suite 300 Washington, D.C.
20005 (202) 842-0034
j 52 1
[ Recess.]
r~%
( )
2 DR. KENNEDY:
I want to sort of start the i
3 discussion with a discussion of a risk-based approach to i
4 establishing seismic design criteria for piping.
5 The first thing I want to talk about is 6
establishing the required seismic margin for a piping system 7
in nuclear power plants and the purpose of starting with 8
this is trying to put together a framework that might assist
)
9-code committees and the NRC Staff to establish how much 10 conservatism is sufficient conservatism.
11-At least it's been my experience on some other 12 problems if agreement can first be reached as to what is the 13 target degree of conservatism that you are trying to achieve 14 out of a criteria, at least from that stage on, everybody is
!,/
15 trying to aim at the same target.
We may not all hit the 16 same target but we are still trying to aim at the same 17 target and this narrows'the focus of discussions as to what 18 code changes might be desirable versus what code changes 19 might not be desirable.
20 The very first thing you have to do though is 21 decide what is your desired goal.
Well, here, frankly, the j
22 desired goal was a goal that Nilesh Chokshi of the NRC 23 Research and I both felt would be a desirable desired goal, 24 and that is that we establish piping seismic design criteria 25.
sufficiently conservative that we continue to avoid piping
'( ')
ANN RILEY & ASSOCIATES, LTD.
]
Court Reporters 1250 I Street, N.W.,
Suite 300 Washington, D.C.
20005 (202) 842-0034
53 1
becoming a significant contributor to the seismic risk of a 2
nuclear power plant.
3 Our background had been that based on experiences 4
from the review of piping systems that were designed to code 5
versions, and earlier versions of the standard review plan, i
6 seismic PRAs and seismic margin studies have all concluded 7
that piping was not a significant contributor to the seismic 8
risk, and the issue was how much could we liberalize the 9
seismic design criteria and still have it achieve this 10 desired goal of not being a significant contributor.
11 Now there's a number of things that are needed for 12 background information on this.
13 One is that the NRC has used over a number of past 14 years a' thing called the High Confidence Low Probability of 15 Failure, or HCLPF, seismic capacity of a plant as a measure 16 of the seismic margin of the plant, and they have used that 17 then as a surrogate for seismic risk.
If the HCLPF seismic 18 capacity is sufficiently high, the seismic risk will be 19 adequately low.
20 This HCLPF capacity initially was defined as 21 approximately 95 percent confidence of less than 5 percent 22 probability of failure, but based on various studies, 23 particularly as reported on in here, it also can be thought 24 of approximately as a composite or mean 1 percent 25 probability of failure, and I am going to use this b
ANN RILEY & ASSOCIATES, LTD.
V Court Reporters i
1250 I Street, N.W.,
Suite 300 Washington, D.C. 20005 l
(202) 842-0034 l
l i
54 1
approximate definition of the HCLPF to try to keep issues of s ()
2 uncertainty versus randomnecs from crowding in, so whatever 3
I talk here, HCLPF basically corresponds to the earthquake 4
level that we.think we have a 1 percent probability of 5
failure, and seismic margin has just simply been defined by 6-the ratio of the HCLPF capacity to the SSE design ground i
7 motion, so R sub H of one and a half says that you would 8
have to get to one and a half times the SSE before you would j
l 9
estimate there is a one percent probability of failure.
10 Well, in trying to pursue this approach, the first 11 problem area was what is an adequate seismic margin for the 12 whole plant?
13 Basically that has not been explicitly defined to 14 date by the NRC but we probably can make some reasonable O
\\,/
15 estimates of what is an adequate seismic margin by looking 1
16
'at precedents.
We have the precedents that in general 17 existing nuclear power plants have been deemed to be j
)
18 adequately safe and we have every nuclear power plant in the j
19 United States has had to go through a seismic margin study 20 or a seismic PRA study and from those studies we basically 21 conclude that after fairly obvious problems that are 22 discovered in seismic walkdowns have been fixed, our 23 existing nuclear power plants have HCLPF seismic margins of 24 at leant 1.25 typically.
25-We also have at the NRC for the standardized
( (- )
ANN RILEY & ASSOCIATES, LTD.
Court Reporters 1250 I Street, N.W.,
Suite 300 Washington, D.C.
20005 (202) 842-0034
)
55 1
advanced lightwater reactors the design basis earthquake or
()
2 SSE was.3g but they desired a target HCLPF capacity of
.59, 3
which says that for the ALWRs they are looking at a HCLPF 4
seismic margin of at least 1.67, so we can suggest that this 5
is sort of the type of range that we might wish to have for 6
HCLPF seismic margins.
7 Now another thing to look at is if you define the 8
annual frequency of exceeding the STE, and you define your 9
SSE in terms of some ground motion that has a certain annual 10 frequency of excedence, and in current revision criteria I 11 am never sure I know all these numbers -- Reg Guide 1.165, 12 is it? -- that has been defined for future plants as being 13 about 10 to the minus 5 median, which is equivalent to about 14 a 10 to the minus 4 mean annual frequency of exceeding the b) 15 SSE.
q_
16~
If we then say we want some annual probability of 17 unacceptable performance, we can talk in terms of a 18 probability ratio which is simply the ratio of the annual 19 frequency of exceeding the SSE to the annual probability of 20 failure, and you can simply taking existing seismic hazard 21 curves and assuming that fragilities are based on a log 22 normal distribution, you can rigorously with just those two 23 assumptions obtain a ratio between what this HCLPF ratio is 24 and what this probability ratio is.
l 25 For a HCLPF ratio of 1.25, that probability ratio O~
ANN RILEY & ASSOCIATES, LTD.
Court Reporters 1250 I Street, N.W.,
Suite 300 Washington, D.C.
20005 (202) 842-0034
?u________ _ _ _
o -
56 is going to be in this range.
It varies some with the (m) 2 logarithmic standard deviation of your fragility and with 3
the slope of the hazard curve, but for overall practical 4
hazard curve slopes and logarithmic standard deviation 5.
variabilities it is in this ratio, so a HCLPF seismic margin 6
of 1.25 if the SSE was at 10 to the minus 4 leads to f
7 appro:cimately -- well, it leads to this range on probability 8
of failure -- similarly the 1.67.
.9-These also seem to be reasonable numbers that 10 might indicate that this is a reasonable range for the kind 11 of HCLPF seismic margin that we may want plants to have.
12 Now, once we have decided on what-HCLPF seismic 13 margin we want for plants to have, what is an acceptable I
14' probability of failure for a piping segment at this HCLPF
()
15 earthquake level?
This is the level that we expect the 16 plant to have a 1 percent probability of unacceptable 17 performance.
How much of that can be due to a single piping 18 segment?
19 Now, clearly, the probabalistically combined sum 20 of all of the piping failure probabilities has got to be 21 less than 1 percent or else it drives the HCLPF of the plant 22 down.
There are some things that make it a little bit more L23 difficult problem, and that is there are a large number of 24.
potential failure locations.
But in actual fact, when you 25
-look at real piping systems, only a very limited number of O
ANN RILEY & ASSOCIATES, LTD.
Court Reporcers 1250 I Street, N.W.,
Suite 300 Washington, D.C. 20005 (202) 842-0034
57 1
these potential failure locations ever have high stresses to
()
2 start with.
There's only a limited number of locations that 3
get up to code allowables, whatever your code allowables are 4
set at.
5 There's also the situation that failure 6
probabilities at different locations are neither perfectly 7
dependent or perfectly independent.
If they were perfectly 8
-- all perfectly dependent, we could set any one of these up 9
to the 1 percent.
If they were purely independently, any 10' one failure location would effectively only be allowed to 11 have a probability of failure of 1 percent divided by the 12 number of locations.
13 So you need to make some judgments.
These are the 14_
judgement ranges that I considered reasonable _and that I
()
15 used for-the study.
Is it any single piping location?
When 16 we get up to a ground motion level, that is the HCLPF level, 17 or 1.25 to 1.67_ times the design' level, we didn't.want more 18 than about a probability of failure of the piping location 19 to fall outside of this range.
My best estimate, some upper 20_
bounds and' lower. bound ranges to use for looking at what is
-21 an adequate margin.
22 I~am-going to have to make a. couple of L
- 23' definitions.
That'overall seismic margin is made up of both 24~
a capacity margin, which is controlled by the ASME code, and p'
25 we'll call-it therefore a piping code capacity factor of l
i l
O-ANN RILEY & ASSOCIATES, LTD.
l Court Reporters 1250 I Street, N.W.,
Suite 300 Washington, D.C. 20005
_(202) 842-0034
V f~
58 1
safety or margin, and we are going to define that in terms
()
2 of the i percent probability of failure point.
When that 3
capacity is reached, we want to get this factor of safety l
4 such'that if the capacity gets up to that level, we would 5
expect about a 1 percent. chance of failure.
6 And then there's a response margin.
And this is 7
really controlled by the standard review plan, it has 8'
nothing to do with the ASME code, it has to do with the 9
standard review plan.
And over the years, more and more of L
10 the standard review plan has been targeted on the response 11 or demand side to achieving about an 84 percent 12 non-exceedance probability.
But there are some additional 13 conservatism that creep in.
And so what is our factor of 14 safety on the response side relative to the 84 percent O)
(,
15 non-exceedance probability?
16 Now, it turns out that this is an interesting way 17 of looking at the problem because you can write the required 18 code capacity margin in terms of your HCLPF margin that you i
19 are trying to achieve, divided by the margin that you have 20 on the response side beyond the 84 percent non-exceedance 21 probability, times a factor over here.
And this factor is 22 related to the logarithmic standard deviations that you have 23 on-response and capacity and the combination of the two.
24 And if our capacities are defined at the 1 percent 25-non-exceedance probability, the standard normal variant that O
ANN RILEY.& ASSOCIATES, LTD.
Court Reporters 1250 I Street, N.W.,
Suite 300 Washington, D.C. 20005 (202) 842-0034 l
59 1
you put in here is 2.326.
That is simply how many standard 2
deviations 1 percent probability of failure lies away from 3
the mean.
And on the response, since it's at the 84, it's a 4
1.0 because that lies 1.0 standard deviations away, and this 5
factor here is related to what is your acceptable 6
probability of failure of this piping system.
And over a 7
very wide range of practical values, we can estimate this 8
factor here, which is a function of those variabilities.
9 Based on seismic PRAs, other literature, I believe 10 that typical variabilities on capacity, logarithmic standard 11 deviations on piping capacities lie in this range.
Typical 12 variabilities on response lie in this range.
If our target 13 is a 1/10th of 1 percent probability of failure for a piping 14 segment at the HCLPF level, this fac*or term lies in the k
15 range of 1.24 to 1.56.
And that basica21y, if our goal is a 16 probability of failure of 1/10th of 1 percant at the HCLPF 17 level of the plant, this factor can be slightly 18 conservatively, on average at least, approximated as about 19 one and a half.
One and a half is above the midpoint of 20 this estimate, it is not the worst number there.
21 And if you go through the same considerations with 22 these other probabilities of failure, this factor comes out 23 to be anywhere from about 1.1 to 2.
24 Well, a lot of people have argued in the past, 25 well, we have been very, very conservative in our response O
ANN RILEY & ASSOCIATES, LTD.
Court Reporters 1250 I Street, N.W.,
Suite 300 Washington, D.C.
20005 (202) 842-0034
60 1
analyses, therefore, we ought to liberalize the capacity 2
side of the ASME code.
I happen to not agree with that.
I 3
think if you are very conservative on the response analyses, 4
correct those problems, get a rational response or demand 5
analysis, but you can't relax an acceptance code just 6
because you have been conservative in the past on the 7
response side.
8 The goal on the response side has always been of 9
the standard review plan, to make sure this response factor 10 of safety exceeds 1 for the 84 percent non-exceedance 11
. probability response.
In other words, the calculated 12 response is there should be less than a 16 percent chance 13 that the actual response that might occur if that earthquake 14 occurred should be less than about 16 percent.
A(_,)
15 But because of a lot of arbitrary conservatism 16 and simplifications that in the past have been included in i
17 response analyses, in most cases this response factor may 18 exceed 1 by quite a bit.
And this is my estimate of the 19 range.
I think some other people might raise this 4 even 20 higher.
But most of those conservatism have been removed 21 from the standard review plan ever since about 1989.
- Now, 22 our earlier designs out there, they're in, but we now allow i
23 higher damping.
We allow an improved way of doing soil 24 structure interaction analysis, coupled piping analysis, a 25 number of things.
So there is no reason, other than maybe I
ANN RILEY & ASSOCIATES, LTD.
\\-
Court Reporters 1250 I Street, N.W.,
Suite 300 Washington, D.C.
20005 (202) 842-0034
I 61 1
for simplification of analysis, that this response factor
'l 2
significantly exceeds about 1.
And, in fact, we can 3
specifically point to areas where it does not significantly 4
exceed 1.
5 So I don't think we can take a lot of credit on 6
the response side for excess conservatism.
So now I am 7
prepared to go into this equation to try to back calculate 8
how much conservatism, what code margin are we looking for, 9
against the 1 percent probability of failure point.
10 What I have suggested at this stage is that a 11 HCLPF seismic margin, a reasonable level for it from past 12 precedents, is from 1.25 to 1.67.
This response factor, l
13 there's probably some, but we shouldn't take a lot of credit i
14.
for excess conservatism on the response side.
This 15 correction factor, whatever you want to call it, I have 16 suggested ranges from about 1.1 to 2, depending on what 17 probability of failure you are aiming at, since that is sort
)
18 of a judgment call.
That's not the right viewgraph.
19 And if you play around with these numbers to come 20 up with this response -- this code capacity margin, you 21 basically say that the code capacity margin that our ASME 22 code ought to be trying to achieve for a piping system I
23 against the 1 percent probability of failure capacity, 1 j
24 percent -- yeah.
Against a capacity that there is less than 25 or equal to 1 percent probability of failure should lie in l
O ANN RILEY & ASSOCIATES, LTD.
tk-Court Reporters 1250 I Street, N.W.,
Suite 300 Washington, D.C.
20005 (202) 842-0034 l
62 1
the range of 1.4 to 2.7, my best estimate being around 2.
2 I think that aiming at about a factor of 2 has 3
been accepted pretty much by the ASME special working group.
.4 I think it has been accepted pretty much by the NRC staff.
5 It is important that it is not a precise number.
By 6
changing these numbers around, we could get this number 7
anywhere in this range, it's sort of a midpoint of the 8
range.
So it is not a number that should be treated as a 9
hard and fast number that is accurate to three significant 10 figures, or even two.
11
[ Laughter.)
22 DR. KENNEDY:
But it's about the right range for 13 what we should be aiming at, and I don't think there is a 14 lot of controversy at that stage.
O(,/
15 So, next, let's look at the component test data, 16 EPRI component test data to see if this can help us 17 establish a code criteria that will achieve this desired 18 seismic margin.
So at this stage, our desired code margin, 19 and I am again going to define it in terms of a 1 percent 20 probability of failure point, now we can break it up into 21 parts that deal with various parts of our evaluation.
22 It basically can be thought of as being made up of 23 three terms,'a strength margin, which is directly something 24 that the code equations can address; a nonlinear dynamic 25 margin, which, frankly, not be accurately estimated unless O
ANN RILEY & ASSOCIATES, LTD.
\\--
Court Reporters 1250 I Street, N.W.,
Suite 300 Washington, D.C.
20005 (202) 842-0034
63 1
you do a nonlinear analysis of the piping system.
But we
.()
2 can make some judgments on it, but we cannot get this margin t
3 very accurate without doing nonlinear analysis, which people 4
are simply not willing to do ir. current piping code type 5
designs.
6 This is basically -- if you do a linear analysis 7
of a piping system, and you calculate demands, your 8
calculated demands can exceed your ultimate capacities and 9
still have acceptable performance because this is a 10 nonlinear oscillatory demand, and all it means if that you 11 have gone -- that you go into the nonlinear region.
So we 12 will have to estimate this factor.
13 ~
And there is also redundancy margin.
The fact is 14 that when one elbow reaches its moment capacity, that is not O
(,/
15 failure in a piping system because loads can redistribute in i
16 other directions and that depends also on the piping system.
17 And the redundancy margin cannot be accurately estimated
~
18 without at least doing nonlinear study analysis.
19 So our overall capacity margin is a strength 20 margin, a nonlinear margin, a redundancy margin.
And I have 21 cast the equation in this form because what we can use the 22 EPRI test data for is to get the strength margin.
We are 23-going to have to use other means to estimate a nonlinear and 24' a redundancy margin and we are going to want to make some
'25 kind of a generic conservative estimate of that, unless we l
()
ANN RILEY & ASSOCIATES, LTD.
i-x Court Reporters 1250 I Street, N.W.,
Suite 300 Washington, D.C. 20005 1
(202) 842-0034
64 1
want people to have to do nenlinear analysis.
\\
-2 So if our overall capacity margin that we are (d
3.
aiming at is 2,-the strength margin that we have to have our 4
code equation achieve needs to be 2 divided by whatever 5
credit we are going to give to these terms, and that, of 6
course, is one of the places where there is differences of 7'
opinion and controversy and there needs to be some kind of a 8
consensus decision how much credit are you going to give to 9
those terms.
10 The reason to divide it up into this way, one of 3
11 the strong reasons, is a strength margin should not be 12 argumentative.
We have the test data, at least for the 13 tested components we know what the strength margin was, and 14 it is just the ultimate dynamic moment that occurred during (n) 15 the test divided by your code allowable moment, by whatever 16 code equation you ultimately decide on.
17 Well, I looked at the new '94 code provisions, 18 ASME provisions, and the code moment is basically given by 19 this equation.
And everybody has been concentrating on this j
20 4-1/2 here, whether that number should be 4-1/2 or whether j
21 that number should be 3, but there's a lot of other terms in 22 this equation.
This'is a non-controversial term the S sub i
1 23 M, basically, it is a nominal stress term.
The section l
24
' modulus is basically non-controversial.
The effect of l
25 pressure is maybe somewhat controversial, but it is -- on j
i 4
l i
j
(/
ANN RILEY & ASSOCIATES, LTD.
N-Court Reporters i
1250 I Street, N.W.,
Suite 300 Washington, D.C. 20005 (202) 842-0034 i
L
l 65 1
1 all of these tests, this is a negligible term.
This whole
)
2 thing is negligible and is not part of the argument.
3 But B2 is not a negligible term and is part of the l
4 argument.
This number here, divided by B2, I would agree j
3 5
with the comments, that is not -- I mean this is not a real 6
stress.
This is something that says the code S sub M, if l
7 you multiple it by 4-1/2, that's an empirical multiplier, 8
and you divide it by B2, you will calculate a moment that 9
this component can carry.
It is not a real stress.
So it's 10 4-1/2 and the B2 go intimately together.
11 The ultimate dynamic moment is not very 1
-12 controversial.
There may be 5 or 10 percent disagreements i
13 as to what it is.
But we know what acceleration times 14 history was measured for that weight up on the top of every
()
15 one of these components, and we know the moment is equal to 16 that acceleration times the mass, times the lever arm, at 17 every instant time.
Otherwise, we had better decide we 18 don't -- that Newton was wrong and we may as well at that 19 stage go home in earthquake engineering.
i 20 So we know that the tests showed this kind of 21 moment rotation characteristics.
This is one particular 22 test.
We have that for every test.
The ultimate dynamic l
23 moment is the highest moment capacity.
Well, it's basically 24
-- the moment capacity in closing may be different than in 25 opening, and so it's the lesser of these two.
It's the I
\\
ANN-RILEY & ASSOCIATES, LTD.
I
\\- /
Court Reporters 1250 I Street, N.W.,
Suite 300 Washington, D.C.
20005 1
j (202) 842-0034 1
66 1
lesser of'the highest in either direction that occurred
()
2 prior to failure.
3 So we can address -- I mean at this stage, if we 4
base our code equations on not exceeding this ultimate 5
dynamic moment, we can, at that stage, guarantee against all 6
of the possibly failure modes that may have been observed.
7 Collapse -- we can't -- you can't collapse until you reach 8
this moment.
Low cycle fatigue, or low cycle fatigue 9
ratcheting doesn't occur.
None of the tests occur before 10 you reach this moment.
Excessive displacements doesn't 11 occur before you reach this moment.
12 If you take and divide this moment by a reasonable 13 factor of safety, you are basically operating pretty much in 14 the pseudo-linear range,~and so we don't have to worry about
()
15 a lot of issues associated with that.
So that was why I was 16 felt that a. code equation really ought to be based on this 17 ultimate dynamic moment that was observed in the test, 18 because everything else is very controversial, divided by l
19 some factor of safety.
20 Now, these are the tests for which we have good 21 data.
Some are stainless steel, some are carbon steel.
22 This is just the section modulus.
These are the internal 23
. pressures, so you can see which of the tests didn't have any 24 pressure and which tests did.
The B1 and B2 indices at the 25 failure location.
The code moments from that code moment
[ )/
ANN RILEY & ASSOCIATES, LTD.
\\-
Court Reporters 1250 I Street, N.W.,
Suite 300 Washington, D.C.
20005 (202) 842-0034
67 1
equations.
This is that
'94, ASME '94 edition of the code.
Il 2
The ultimate dynamic moments observed in the test and the FS
\\s /
{
3 is just the ratio of ultimate dynamic to code moment.
And l
4 so that is what I call the strength factor of safety.
5 Now, on all of the elbow tests, except for test 6
37, these values are fairly high.
They are variable and i
7 they. range from about 1.6 to 2.7, I believe it is.
The 8
non-elbow tests, though, they are nowhere near as high if 9
you use the B2s, which are the B2s for where the failure 10 occurred, because the failure typically did not occur in the i
11 component, it occurred in the straight section next to the 12 component.
These two it is occurred in the component.
f 13 Now, test 37 has to be addressed and, frankly, 1
14 that is an outlier test and it is outliers for a lot of good
)
15 reasons, but we have to find a way to put a fence around 16 test 37.
If we are going to accept that elbows have these 17 factors of safety, we have got to address test 37 because it 18 didn't have those factors of safety.
But test 37 had a lot j
19 of unique features.
'It was a very low frequency component, 20 but test 30 was also the same low frequency component.
In i
R21 fact, the most important thing I think you have got to keep 22-in mind, test 30 and 37 are identical in every aspect except 23 for internal pressure.
So, clearly, internal pressure had 24 the lack of internal pressure in test 37 clearly had 25 something to do with this low margin.
It wasn't maybe the
[\\-- )
ANN RILEY & ASSOCIATES, LTD.
Court Reporters 1250 I Street, N.W.,
Suite 300 Washington, D.C. 20005 (202) 842-0034 L
68 1
whole cause.
()
2 Test 30 and 37 had by far the largest weight 3
moment that tended to unstabilize you.
They were some of 4
the larger diameter to thickness ratio tests.
They were 5
schedule 10, but there are several other schedule 10s in 6
there.
The clear thing that did happen is,.under cyclic 7
loading, there is greater moment capacity than there is 8
under a monotonic loading.
But test 37 didn't build up that 9
greater capacity because it just ratcheted in one direction.
'10 It never got the real reverse cycles, and I think that had a 11 lot to do with the fact there was an internal pressure in 12 there.
It had a lot to do with the fact that there was a 13 large moment load on the top.
But it didn't get reversed 14 cycling.
And we have got to know a breadth of when that
'O
(,/.
problem might exist if we are going to talk about 15 16 potentially liberalizing the code, there is a need to put a 17 handle around that problem.
18 At-this stage, I am going to ignore test 37.
I l
19 mean any kind of statistical interpretation of this data, I
1 20 test 37 is so different t?om all of the other elbows -- in j
i 21 fact, it lies more.than 5-1/2 standard deviations away from
-22 the mean of all of the other tests and there's no way that
- 23 one test out'of this number of tests should have lay 5-1/2 24 standard deviations away from the mean if this was a 25' homogeneous data base.
It is not a homogeneous data base.
ANN RILEY & ASSOCIATES, LTD.
Court Reporters 1250 I Street, N.W., Suite 300 Washington, D.C.
20005 (202) 842-0034
69 1
Test 37 is unique and we are going to have to solve it
[
}
2 uniquely.
1.
3 Well, another issue, we are trying to estimate a 4
factor of safety associated with a 1 percent probability of f
L 5
failure.
Here's the factors of safety we have in these l
6 tests.
There isn't enough tests to extrapolate -- I mean to 7
literally calculate the 1 percent probability of failure 8
factor of safety.
We have to extrapolate.
Based upon my 9
review of a lot of structural data, a lot of structural 10 strength data does a reasonably good job of fitting a 11 lognormal distribution.
So I am going to initially start l
12 out and assume that this data fits a lognormal distribution.
13 And if we assume it fits a lognormal distribution, we can 14 extrapolate to the 1 percent probability of failure by use
()
15 of this equation.
Obviously, we will have to check whether 16 it really does fit a lognormal distribution.
17 But we also have to check the homogeneity of the 18 data.
We have carbon steel and stainless steel data in 19 these tables, but they seem to be part of a homogeneous data 20 set.
Any kind of hypothesis testing that we might do, we 21 would argue that the carbon steel and the stainless steel 22
. data are both reasonably represented by this S sub M term, 23 that that does a reasonable job of taking care of the 24 differences between carbon steel and stainless steel.
There i
25 are two medians, whether we are talking elbow tests or I
i
- [^h ANN RILEY & ASSOCIATES, LTD.
,s/.
Court Reporters 1250 I Street, N.W.,
Suite 300 Washington, D.C.
20005 l
(202) 842-0034
_.__._m__
70 l
I non-elbow tests, quite close to each other.
/~ T 2
The other thing is the elbow data.
V 3
CHAIRMAN SHACK:
I thought I saw you -- you used 4
20 ksi for all the tests.
5 DR. KENNEDY:
I used tnat S sub M was 20 ksi for 6
all of the. tests.
7 CHAIRMAN SHACK:
So that you didn't really need to 8
distinguish between carbon steel and stainless steel.
9 DR. KENNEDY:
Well, except that 20 kai is 10 basically S sub M for carbon steel and stainless steel that 11 were in the test, but the yield stresses are not the same 12 for those two materials, but the code -- that is the 13 allowable S sub M for those two materials that the tests 14 were done on.
And if yield stress was the important (3) 15
-parameter, or ultimate stress was the important parameter, 16 it should have shown up that when you use S sub M, that 17 there should have been a different in the margin.
18 What'I guess I am saying is it appears that the 19 fact code base is-there, code moment capacity on this term 20 called S sub M, and I'll let one of the code people explain 21 what S sub M is, because.it still confuses me exactly.
22 DR. SHAO:
S sub M for carbon steel and S sub M 23 for stainless steel, they are different.
24 DR. KENNEDY:
No, in this case, these carbon j
25 steels and stainless steels, they --
I I
IN ANN RILEY & ASSOCIATES, LTD.
\\s Court Reporters 1250 I Street, N.W.,
Suite 300 Washington, D.C. 20005
'(202) 842-0034
I 71 1
DR. SHAO:
They have to have same stream, but
()
2 different criteria and terms.
3 DR. KENNEDY:
They have different criteria as to 4
how the code arrives at S sub M.
5 DR. SHAO:
You see, S M is based on -- the code, l
6 for carbon steel, they use two-third of yield.
For 7
stainless steel, they year use 90 percent of the yield 8
because of extreme hardening.
9 DR. KENNEDY:
Yes.
Yes.
So for the same S sub M, 10 the two materials have different yield stresses.
11 DR. SHAO:
Because stainless steel has extreme 12 hardening.
13 DR. KENNEDY:
Right.
But I guess my point is I 14 don't believe that the S sub M is the thing that is the O's,,/
15 problem in the code equation.
But there is clearly a 16 difference between the elbow test and the non-elbow test for 17 those non-elbow test for those non-elbow tests that had B2 l
18 of 1.
Clearly are different from the elbow tests where B2 19 was 3 or greater.
In all thosc elbows, if you go into the 20 ASME code and you get a B2 that corresponds to that elbow, 21 it is always greater than 3.
22 And we have definitely got a different 1 percent 23 probability of failure factor or safety, or median factor of 24 safety.
If we use the ASME B2s, we have clearly got a 25 different factor of safety for elbows versus non-elbows.
()
ANN RILEY & ASSOCIATES, LTD.
Court Reporters j
1250 I Street, N.W.,
Suite 300 Washington, D.C.
20005 I
(202) 842-0034 l
i
_.._________._._________________J
72 l
l 1
And that makes the issue of whether 4-1/2 S sub M is
~2 adequate or not.
I mean these number were for 4'-1/2 S sub M 3
and you might argue maybe 1-1/2 is enough strength factor.of
~
4
-safety, but certainly.75 is not enough strength factor of 5
safety.
l 6
So,' an obvious, and I suggested a possible l'
7 solution to this problem is to change B2.
Use a B2 prime 8
for oscillatory dynamic no-loads and simply say that B2 9
prime is B2 or 2, whichever is greater.
10 In all of the earlier GE work that supported the 11 change'to 4-1/2.S sub M, they always used a B2 prime of 2 or 12 greater, because they used the value associated with the 13
_ component, not with where the~ failure occurred, and where 14 the component was always greater than 2.
So just simply --
15 but the use of 4-1/2 S sub M is very,'very questionable for 16' a straight sectinn of pipe where B2 can be'1, and we don't 17 have tests on the straight section of pipes other than that 18 Ken Jaquay has.said all these component tests that failed in 19 the straight section.were effectively tests on the straight 201 section of pipe.
21 MR. JAQUAY:
No, Bob, there was a straight run 22-test.
23 DR. KENNEDY:
Which one was that?
I can't --
i.
24 MR. JAQUAY:
I believe it was number 15 of the top 25 of'my head.
ANN RILEY-& ASSOCIATES, LTD.
Court Reporters
'1250 I Street, N.W.,
Suite 300 Washington, D.C. 20005 (202) 842-0034
73 1
DR. KENNEDY:
Okay.
So we do have one straight
()
2 section.
You're.right.
3 MR. JAQUAY:
It didn't have the wall thickness --
4 OR. KENNEDY:
And it also had a relatively low 5
strength factor.
6 DR. SHAO:
Can I ask a question?
Okay.
If I 7
understand, you say you keep the stress the same, but you 8
change B2.
9 DR. KENNEDY:
That was what I studied.
10 DR. SHAO:
Okay.
Now, let me ask you --
11 DR. KENNEDY:
I am not necessarily saying that is 12 my end recommendation.
You have to wait till my end slide 13 to see that.
14 DR. SHAO:
Wait a minute.
I have a' question on
)
15 that.
16 DR. KENNEDY:
Okay.
17 DR. SHAO:
If you keep your stress at 4.5, then 18 how do you resolve the inspection problem?
.9 DR. KENNEDY:
Well, I happen to agree with what 20 Dick Barnes indicated.
4-1/2 S sub M is not the stress we 21' have gone to.
I doubt if you could find any location in 22 that elbow that was actually stressed to 4-1/2 S sub M.
23 It's an empirical equation that we get our moment capacity 24 by taking 4-1/2 S sub M and dividing by this arbitrary 25 factor called B2.
That arbitrary factor worked real good Ct ANN RILEY & ASSOCIATES, LTD.
Court Reporters i
1250 I Street, N.W.,
Suite 300
""""i"!'f"64?: 63
)
l l
L
74 1
for monotonic loadings.
It doesn't appear to work so good
/ Y 2
for cyclic loadings.
It's too high.
\\,J 3
So my end conclusion is there's many ways of 4
solving this problem.
We don't have to go to 4-1/2 S sub M, 5
we can go back to 3-1/2 S sub M, but change the B2.
Or we 6
.can-go-to 4-1/2 S sub M and put a lower bound on B2.
7 DR. SHAO:
Okay.
Now, I have another question.
8 DR. KENNEDY:
Because it is an empirical equation.
9 DR. SHAO:
I have another question here.
A lot of 10 plants here, suppose after 20 years, okay, -- when it is 11 new, I have no problem here.
After 20 years, you have some 12 kind of flaw in the pipe.
A certain flaw in a pipe, how 13 does that change your numbers over here?
14 DR. KENNEDY:
Well, I mean the way -- the way I (O,)
15 understand Ken Jaguay's discussion is there are three 16 important open issues that must be resolved.
One is a fence 17 around test 37.
What other kind of situations might behave 18 like test 37 did?
And we may have to do more testing to 19 resolve that.
Two is, What are temperature effects on 20 carbon steel?
And 3 is, What is the effect of various types 21 of flaws or degradation?
There is some limited test data on 22 that third item that I don't believe has had a rigorous 23 study, but a cursory look at it says that as long as we base 24 our capacities on this ultimate dynamic moment, divided by a 25 reasonable factor, it takes a very large flaw to get us an ANN RILEY & ASSOCIATES, LTD.
Court Reporters 1250 I Street, N.W.,
Suite 300 Washington, D.C.
20005 (202) 842-0034 L
2
75 1
ultimate dynamic moment less than that value.
I Y 2
DR. SHAO:
The original thinking is you have V
3 conservative in response, conservative in capacity, is 4
because of these flaw, poscible flaw in the future.
5 DR. KENNEDY:
Sure.
6' DR. SHAO:
So that you don't have to worry about 7
it so much.
8 DR. KENNEDY:
Sure.
But I think we will achieve 9
that if we got to an ultimate dynamic moment divided by some 10 factor.
And there has to be a consensus decision as to what 11 that factor is.
I have my own judgments, I am going to show 12 them if we get to that stage.
13 It has to be consensus of the ASME special working 14 group and the NRC staff.
But I think -- and I think there kq,;
15 has to be further study.
I mean, I don't want anyone going 16 away from this meeting that I believe there has been enough 17' study cn1 this flaw issue.
I think there has to be some 18 further study.
But some very quick looking at it, I'think 19 maybe the flaw issue is not a very big issue.
'20 DR. SHAO:
Depending on the size of the flaw, 21 though.
22 DR. KENNEDY:
Sure.
23' DR. SHAO:
That's why we -- NRC -- I mean, that's 24 why NRC -- anything less than 10 percent, we don't worry 25-about it, but the minute you reach let's say 30 percent, 40
/ D ANN RILEY & ASSOCIATES, LTD.
\\- /
Court Reporters 1250 I Street, N.W.,
Suite 300 Washington, D.C. 20005 g
(202) 842-0034
76 1
percent flaw, there is a different ball game.
m i
1 2
DR. KENNEDY:
All of this test data I am showing G
3 you is on new pipe components, and that issue of how much 4
degradation can you allow and still use these stresses has 5
to be addressed.
6 MR. BARNES:
Mr. Chairman, may I just -- I 7
understand your problem, but, you know, that problem still 8
exists even if you go back with the --
9 DR. SHAO:
If you have a large margin, you don't 10 have to --
11 MR. BARNES:
I'll tell you what the problem is.
12 IF the failure mechanism is a fatigue ratchet, you're basing
-13 all the rules on something.that is not the true failure 14 mechanism, and I think you have'-- we have to -- I think the
-3(_)
15 thrust of my presentation is I believe code rules should
'16 represent the true failure.
17 DR. SHAO:
Oh, really.
Sure.
I have no problem 18 with that.
19 MR. BARNES:
And I believe then you're dealing 20 with the proper foundation to build your rules.
J 21-DR. SHAO:
Okay.
Is it possible that-the failure 22 mechanism would change, too?
23 MR. BARNES:
Well, I think Bob's point he's trying 24 to make is he said he can show through this methodology --
25 DR. SHAO:
No, I don't question the methodology, ANN RILEY & ASSOCIATES, LTD.
Court Reporters 1250 I Street, N.W.,
Suite 300 Washington, D.C.
20005 (202) 842-0034
?
77 1
but the thing is, the methodology covers certain so-called
()
2 inputs.
Whatever you input in, you output out.
If you have 3
no perfect pipe, that's a good answer.
4 MR. BARNES:
Right.
But if we're talking about a I
I 5
flaw that is going to -- and we're talking about it failing 6
in a reversing mode, right?
7 DR. SHAO:
Yes.
You see, we have to worry about l
8 so-called inspection and so-called design, both.
9 MR. BARNES:
Right.
10 DR. SHAO:
It's not only design or only inspect.
i l
11
. What should be the criteria?
12 MR. BARNES:
But we're talking about a flaw and 13 its response under reversing load.
That's what you're 14 really talking about.
s 15 DR. SHAO:
Yes.
16 MR. BARNES:
And what we have to understand is, 17 how does that flaw respond under the reversing load?
We j
18 know how it responds under a sustained load, but the l
l 19 question is, how does it respond under reversing load.
I 20 DR. SHAO:
No.
The flaw changed two things.
The i
21 flaw may change the response also.
The flaw also may change 22 capacity.
23 DR. KENNEDY:
I think the way I would address this 24 issue is for these tests that did not have flaws, we have 25 universally found that we have to get up to this ultimate ANN RILEY & ASSOCIATES, LTD.
Court Reporters 1250 I Street, N.W.,
Suite 300 Washington, D.C.
20005 (202) 842-0034
78 1
dynamic moment before we get failure.
We get up to that
()
2 moment and then we have to go through rotation at that 3
moment.
4 Now, if we hold our capacity below that ultimate 5
dynamic moment, if the flaw does not significantly affect 6
the ultimate dynamic moment but does affect the rotational 7
capacity, we have still guarded against that.
8 If, on the other hand, the flaw reduces the 9
ultimate dynamic moment, the we will have to find a code 10 capacity change that will address it, and I suggest that 11 what you need to do is look at some of these cyclic test 12 data for flawed pipe.
I have looked at a limited number, 13 but not with any care.
I mean, not with the care that needs 14 to be looked at.
15 My general impression is that flaws reduce the 16 rotational capacity but do not significantly reduce the 17 ultimat; dynami; moment capacity.
But that has to be 18 studied, and so it that is an open issue that has to be 19 resolved by consensus between the special working group aad 20 the NRC staff.
21 MR. BARNES:
I think the point that is coming out 22 is the mechanism is different when we don't understand it, 23 but according to what we've seen, maybe it hasn't changed --
24 maybe the end result hasn't changed much, but the mechanism 25 is different.
We're going down with a mind set on collapse,
()
ANN RILEY & ASSOCIATES, LTD.
Court Reporters 1250 I Street, N.W., Suite 300 Washington, D.C.
20005 (202) 842-0034
79 1
a sustained load approach.
()
2 DR. KENNEDY:
Anyway --
3 DR. SHAO:
Sorry to interrupt.
4 DR. KENNEDY:
No, no.
I think it's important.
5 Those are more important than my talk, to have those 6
discussions.
7 If we set this lower bound on-B-2 and say it 8
simply can't get less than two, it turns out that our e
9 strength factors for these non-elbow tests are very similar 10 to the strength factors that I earlier showed for elbow 11 tests, and now we can pass the hypothesis that they're part 12 of the same dataset, we.can combine them.
Combining them,
'13 these are what I estimate from this dataset that we have a 14
-median factor safety of about two, we have a factor safety 115 of about 1.5 against a 1 percent probability of' failure 16 point.
17-Now, so far, I have used the log normal
~
18 distribution without any jurisdiction of the log normal 19 distribution.
My method that I like to use to justify --
20 this:doesn't quite' fit.
I'll put it over like this.
I like 21 to use this cumulative normal paper and plot the logarithm 22 of this' factor of safety on cumulative normal paper.
23 If all of the data fits as a straight line, then 24:
it's log normally distributed.
If I plotted it not the 25' logarithm but-plotted the data, the' raw data, and it fit a ANN RILEY.& ASSOCIATES, LTD.
Court Reporters 1250 I Street, N.W.,
Suite 300 Washington, D.C. 20005 (202) 842-0034
j 80 1
straight line, it would be normally distributed.
If it ffff 2.
doesn't fit a straight line, it differs a little bit from a 3
log normal distribution.
4 This data reasonably fits a log normal i
5 distribution.
If I just use a square, you know, fitting, 6
this is the median and the logarithmic standard deviation l
)
7 and the 1 percent probability of failure point.
8 DR. POWERS:
When you say this reasonably fits log 9
normal distribution, to my mind, that reasonably does not 10 fit.
11 DR. KENNEDY:
It does not fit perfectly.
I agree 12 with you there.
13 DR. POWERS:
What does reasonably mean?
I mean, 14 there are a large number of tests that would tell you l
)
15 whether it fits to some confidence level.
16 DR. KENNEDY:
Sure.
17 DR. POWERS:
Why don't you use those?
18 DR. KENNEDY:
Within my experience, you always i
19.
pass these confidence tests much -- data that fits this 20 poorly passes those tests.
21 DR. POWERS:
It shouldn't.
Not that data.
That l
22-
' data is clearly trended.
23 DR. KENNEDY:
There are a few tests up here that 24 are high, but I've tried the confidence test and the test
-25 that I'm aware of, it does fit'a log normal distribution by
)
ANN RILEY & ASSOCIATES, LTD.
l
~'
Court Reporters 1250 I Street, N.W.,
Suite 300 Washington, D.C.
20005 (202) 842-0034
81 1
the confidence test.
()
2 I think that there is some issue that there are 3;
.some high data points up here.
When I get that kind of a 4
problem,. I would tend to -- if I'm trying to extrapolate in 5
_this direction, I would tend to fit a straight line through 6
these' points as opposed to through all of the points.
7 But.no matter which I do, my extrapolation comes 8f out to something in the neighborhood of 1.5 factor of f
._9 safety.
I'm trying to extrapolate from a limited number of 10 tests down here to the 1 percent point.
Possibly we could 11 find another distribution that it would fit better.
It 12 might fit a truncated log normal better.
But a truncated 13 log normalLhas some problems with its assumption as to 14' what's happening down here, and those assumptions affect you 15-down here.
16 DR. POWERS:
So really, your confidence level that
-17 you want:to. place is not so much on the data, but on this 18
. factor of safety that.you set.
19 DR. KENNEDY:
'Yes.
Yes.
That's what we're really l
.20 wanting, is -- and frankly, to:get this HCLPF capacity,'if 21 we say it's equivalent to a mean factor of safety of 1.5, we 22 really want to estimate the mean factor of safety.
23:
DR. POWERS:
Why the mean?
'24 DR. KENNEDY:
Well, when we've tried -- in 25 fragility curves where we've tried separating our h
ANN RILEY & AS$2CIATES, LTD.
\\s /
Court Reporters l'
-1250 I Street, N.W.,
Suite 300 Washington, D.C.
20005 (202) 842-0034
(-
t
-a__.__.____.____
_m.____.__._
m-___.
-,m_-_-__m._._,-.m..-._-.__-__.--_--_.__..._-_,____..-.---___.--.._m.m._____-__-_..,
82 1
uncertainties from our random variabilities and we've tried
()
2 to have a family of fragility curves, we have found that for 3
typical fragility curves, aiming at a 95 percent confidence 4
of less than 5 percent frequency of failure can be 5
reasonably approximated by the use of the mean 1 percent 6
frequency of failure.
1 7
DR. POWERS:
Okay.
I mean, that's a good reason.
8 DR. KENNEDY:
And that is discussed in NUREG-1407.
9 I could have tried to separate these apart.
I did not try 10 to.
I simply took advantage of that and said that we're 11 going.to aim at a mean 1 percent.
12 So my conclusion is that this component test data, 13 once we put a -- we could go to the 4.5 S sub M if we never i
14 let B-2 get less than two, and if we did that, our 1 percent
()
15 probability of failure strength factor is about 1.5, and 16 therefore, if our goal is to have a capacity margin at the 1 17 percent probability of failure level of about-two, to reach 18 a conclusion that 4.5 S sub M is okay as long as B-2 is not 19 less than two, we would have to be willing to judge that the 20 non-linear factor times the redundancy factor, some 21 conservative estimate of those factors, is at least 1.33, 22 and we must still address elbow component number 37.
23 So let's look briefly at what these non-linear 24 factors area from the component test.
CALTEC did a lot of 25 non-linear studies of these component tests trying to extend
[}
ANN RILEY & ASSOCIATES, LTD.
N-Court Reporters 1250 I Street, N.W., Suite 300 Washington, D.C.
20005 (202) 842-0034
83 1
them into other frequency ratios, these RW ratios.
)
2 All of their test results are reported in terms of 3
this margin over B-2, and I'm trying to talk in terms of a 4
non-linear factor.
But you can show -- I have shown in my 5
written' report that this non-linear factor is just CALTEC's 6
~MR over B-2 times this B-2-over the strength factor that was 7
measured in the test.
8 The two sets that I am going to show you, most of 9
their studies were done on component 14 and component 40, 10 and the plots I'm going to show are for those.
This ratio i
11 here in both cases was close to unity.
I 12 So as a first approximation, for all of the CALTEC l
l 13-plots that I have seen as a first approximation, this 14 non-linear factor is just the MR over B-2 that they have 15 plotted.
But it really is this equation and you should take 16 their data and multiply by this factor, but that factor is i
17 close to unity.
18 So here is two sets of CALTEC plots.
There are i
19 all kinds of them.
Ken Jaguay showed some.
RW, when it's 20' way.over here, it's basically the piping is much stiffer 21-than the center of the power of the input motion.
When 22
.you're way over here, the piping is very flexible compared l
J 23 to the center of the power of the input motion.
J E24 -
One, you're in residence.
This was a very narrow l
t 25 frequency input motion representative of high up in the I) k/
ANN RILEY & ASSOCIATES, LTD.
Court Reporters-1250 I Street, N.W.,
Suite 300 Washington, D.C.
20005 i
(202) 842-0034
l 84 l'
structure.
This:is a broad frequency input motion
'{ )
2 representative of down near the ground.
These are various 3.
assumptions that CALTEC-made, one of them more liberal, one
)
q 4-of them more conservative, but there's all kinds of 5-assumptions.
This is just sort of representative.
i 6
'But what you do see, as RW goes to zero, this 7
non-linear factor goes to unity, and it must.
There are 8
rigorous reasons why it must go to unity over in this 9
region. 'How fast it goes is argumentative.
How far out you 10
.have to go on this RW before you get away from unity is 11 argumentative.
But.over here that non-linear factor is 12 unity, and that's not argumentative.
Well, maybe some
~13 people have argued, but they're wrong.
I mean, there's so 14
.much in the literature on this subject.
)
15 But over here in residence where the tests were 16 done, that factor can be quite large, and over here on this 17 side,-it can be quite large.
So it's highly variable,las 18' Ken said.
It is very dependent on this frequency' ratio, i
19.
it's very. dependent upon the breadth of the content of the j
i 20 input motion.
What we do know'is as we go to zero, it.does 21.
go to one.
At greater than zero, all these CALTEC studies j
.22 show that it typically is in the range of one to eight; and
-23 at.RW of one and higher, it's typically in the range of two 24 to eight.
As you get real high,'it typically drops to about j
25 two.
O ANN RILEY & ASSOCIATES, LTD.
Court Reporters 1250 I Street, N.W.,
Suite 300 Washington, D.C.
20005 (202) 842-0034
85 1
There is no simple relationship that I can see
()
2 from any of this data between RW and this non-linear factor.
3 But even if we could figure out one, even if we could solve 4
that problem, there's also plenty of literature that shows 5
the non-linear factor for a multi-degree of freedom system 6
that has multiple support like pipe does will always be less 7
than the non-linear benefit that you get in a single degree 8
of freedom oscillator like all these components that were 9
tested were.
10 So even if we could figure out what that factor 11 was for the component, it's going to be less for the system.
12 Probably not a lot less, but it will be less.
In other 13 words, this is sort of a -- the component test is an upper 14 bound description of what this benefit is.
()
15 Plus I don't know how, in a multi-degree of 16 freedom system, you'll ever come up with that frequency 17 ratio, or for broad frequency input down near the ground, 18-it's very argumentative what that frequency ratio is.
19 So at some stage, both the NRC staff and ASME are 20
. going to have to reach a consensus as to what level of 21 credit'they're willing to give for this non-linear factor 22-for real piping systems based on the CALTEC studies, 23 non-linear analysis of piping systems, et cetera.
My 24
. judgment is that'it's very unlikely that it's going to be 25.
less.than 1.25.
'l ANN RILEY & ASSOCIATES, LTD.
Court Reporters 1250 I Street, N.W.,
Suite 300 Washington, D.C.
20005 (202) 842-0034
86 1
At the real low RWs, which are really not a I 'l 2
practical region -- I mean, we're not -- we're trying to not U
L 3
design piping out in that region -- it could get less than 4
one, but over most of the region, it's probably going to be 5
greater than two.
6 Now, the redundancy factor.
In a cantilever 7
component test, that's one.
I mean, when that elbow or 8
component is overloaded, there is no other place for the 9
load to go.
If you take a uniform fixed beam -- that's a 10 nice simple one that we can get the redundancy factor very, 11'
'very easy -- that one, it's about 1.33.
The piping systems 12 that I have seen, this redundancy factor is typically in the 13 range of about 1.2 to maybe as high as 1.4, probably in most 14 cases closer to the 1.2 than to the 1.4, but it ranges, and 15 you can have piping things that look just like the component 16 test where that one also will be unity.
17 So the issue becomes that I don't think more 18 studies are the answer.
I think we already have thousands 19 of studies in the literature from which we have to make a 20 consensus judgment, and it's got to be a consensus of a 21 group of people, not any one individual.
22 If you want to take the absolute most conservative 23 view, that factor is one.
It can be one in some cases.
24 Very, very unlikely.
In which case we would need a strength 25 factor of two if our goal is a capacity margin of two.
I ANN RILEY & ASSOCIATES, LTD.
\\--
Court Reporters 1250 I Street, N.W.,
Suite 300 Washington, D.C. 20005 (202) 842-0034
87 1
I think it's more reasonable to use values in this
()
2 range -- this is my personal opinion -- which means our 3
strength factor needs to range from about 1.1 to 1.5.
4 That's my opinion.
I am sure there are people who think 5
that this is too liberal, this 1.33 is too liberal.
I know 6
there are people on the special working group who think the 7
1.8 is too conservative.
You have to look at all those 8
CALTEC test data where the data spread all the way from 1 to 9
8, and maybe there were a few places where it was even 10 higher than 8, and decide from that and from other data how 11 much credit you're going to take for these factors.
12 MR. BARNES:
You could, of course, put exclusions 13 in the code that prevented you getting a 1 and treating it 14 as such.
15 DR. KENNEDY:
You could put exclusions in the code 16 that will prevent you getting a 1.
How can RW be before you 17 get up to a certain level is argumentative.
We know that 18 ultimately, at RW of zero, it goes to 1.
Where does it 19 start getting close to 1 is a little bit dependent upon 20 detailed characteristics of the motion time history.
21 I'm going to go and simplify that code equation.
22 This is not exactly the code equation.
There were some 23 pressure terms that are basically negligible.
But this is 24 the essence.
25 The code capacity is X over B-2, S sub M times the O
ANN RILEY & ASSOCIATES, LTD.
Court Reporters 1250 I Street, N.W.,
Suite 300 Washington, D.C.
20005 (202) 842-0034
i i
l 88 1
section modiolus.
In other words, X by itself is not the
, ()
2 important parameter.
The important parameter is X over B-2.
3 If we're aiming at a capacity margin or two and we're not 4
going to do any adjustment to the current codes B-2 values, 5
our problem is that this X over B-2, if we decide that a 6
strength factor of 1.5 is what we want -- in fact, I show it l
7 on this table -- my judgment is that if our margin should be 8
two, our strength factor should range from 1.1 to 1.5.
In 9
other words, to get a code moment, you find the ultimate 10 dynamic moment, then you divide it by a factor here that 11 could range from 1.1 to 1.5 depending upon -- or if other 12 people's judgment is outside that range, that factor would 13 be different from that.
l 14 If you decide that 1.5 is okay for the strength
, (~h l ()
15 factor, this X over B-2 could be about 4.5 if you don't 16 adjust B-2.
If you decide 1.1 is okay, that could be 6.
In 17 the '94 code, it was 4.5, but for elbows, maybe you could 18 justify even higher than 4.5.
19 But for the non-elbow case, the cases where B-2 is 20 one, this approach and this data suggest that if this is the l
21 aim point, you want to be about 2.25, that's more f.
l 22 conservative than the '89 code.
If this is what you want to l
23' aim at, you could be three.
That is the '89 code.
The '89 l
l 24 code, X is three; the '94 code, X is 4.5.
25 From my studies, the '89 code is too conservative b
ANN RILEY & ASSOCIATES, LTD.
\\-s Court Reporters 1250 I Street, N.W.,
Suite 300 Washington, D.C.
20005 (202) 842-0034
89 1
for those components that have large B-2s, and the '94 code
()
2 is much more, in my opinion, much more justifiable for those 3
cases that had large B-2s.
But the '94 code is too liberal 4
for those cases where B-2 ought to be one.
5 I think that the fact that -- both ASME working 6
groups and the NRC people both have very strong points on
{
7 their side, and that's why there has been so much problem in 8
reaching a consensus.
9 I mean, frankly, I think you could start with 10 either the'old code rules of 3 S sub M and you could 11 liberalize B-2.
For these oscillatory dynamic loads, you 12 could use a B-2 prime that was B-2 divided by some factor, 13 except that it never should be less than one, or you could 14 go with the new code rules of 4.5 S sub M and limit B-2 and
)
15 say that this B-2 prime should be B-2, but it can't fall 16 below a certain limit.
17 Both of these two approaches accomplish exactly 18 the same thing.
To get the closest agreement between elbow 19 tests that had high B-2s and the components that had B-2s of 20 one, this factor should be two to get the closest agreement.
21 So there is some advantage to this factor being two.
22 To get any new revised code provisions to not fall 23 out of the range of the provisions of the '89 code to the 24
'94 code, this factor could be 1.5.
In other words, if this 25 was 1.5, if we had three S sub M, we would be~ permitting a
[)
ANN RILEY & ASSOCIATES, LTD.
b Court Reporters 1250 I Street, N.W.,
Suite 300 Washington, D.C.
20005-(202) 842-0034
I 90 1.
liberalization effectively up to 4.5 S sub M for things that
()
2 had B-2s greater than 1.5.
Here, if we held this to 1.5, we 3
would never be more conservative than the old three S sub M 4
for things that had B-2s of one.
l 5
So there may be some advantages, even rF,.gh you 6
don't get a consistent margin between the twc sets, there 7.
may be some practical advantages to not using a factor as
)
8 big as two.
9 Now, why -- this is not my work, but I think this 10 is important work for you all to consider.
Why are people 11 concerned about whether this number should be three S sub M 12 or 4.5 S sub M?
What is the practical significance?
13 There were 30 comparative piping designs that were 14 high quality new designs.
So there's not unintentional
()
15-conservatism in those designs.
Those designs were made to 16 the '89 code rules and they have been made to the '94 code 17 rules, and Ernie Branch is the one that put this comparison 18 together, and they showed that the average number of
)
19 supports in a piping system was about 15 supports per system 20 by the '80 rules, and by the '94 rules, that could be 21 reduced down to 9.8.
And that's a significant reduction.
22 Now, when I looked at it, it's a little bit of an 23 apples and oranges comparison, though, because there were 24 two changes made.
The '89 rules, they used 3 S sub M, but 25 they also used 2 percent damping.
The new rules were 4.5 S
[
ANN RILEY & ASSOCIATES, LTD.
N Court Reporters 1250 I Street, N.W.,
Suite 300 Washington, D.C.
20005 (202) 842-0034 l
l 1
91
~ 1 sub M but they used 5 percent damping.
(
).
2 I don't think there's much controversy over -- I 3
mean, we're going -- to the best of my knowledge, people i
4 have' agreed on going to the 5 percent damping.
So some of l
l 5
this reduction was not due to going from 3 to 4.5 S sub M.
6 I have tried to make a judgment as to how much is due, and I 7
think about half of it was due to going to 4.5 S sub M and 8
about' half of it was due to liberalizing the damping.
So 9
.this is an overstatement of how much benefit they get by l
10 going to 4.5 S sub M.
'11 Ernie Branch then said, well, what if we limit B-2 12 to two, because that was what I had originally studied.
How 13
-much penalty does that create?
And the penalty is 14 negligible.
The reason why the penalty is negligible is
)
15 it's the'high B-2. things that control the design.
I mean, 16 seldom does the straight section of pipe or the things that 17L have B-2 of one ever control the design.
So putting some 18 lower bound on that makes very little effect.
~
L 19 But that would also be true if you went to stick 20' with the-old 3 S sub M rules and liberalized the B-2 where 21 you had data to justify it.
22 I guess the end part of my presentation is I think qt 23
.every one of these issues is resolvable.
I think some of 24 them may require some new tests.
I just think it's i
25 extremely unfortunate that we have this situation that an i
f' s
l
! t ANN RILEY & ASSOCIATES, LTD.
\\
Court Reporters 1250 I Street, N.W.,
Suite 300 Washington, D.C. 20005 l
(202) 842-0034 L
i
92
-1 agreement has not yet been reached, but it does have to be a
( )
2 consensus agreement between ASME and the NRC staff.
3 I guess my advertisement part -- if ACRS could do anything to force the two groups to reach a consensus, I 4
5 think that would be a good idea, and I think a reasonable 6
consensus could be reached, and it's something in-between 7
the '89 code and the '94 code.
Where in-between, I think 8
there's a lot of judgment.
9 CHAIRMAN SHACK:
One of the other concerns that 10 was raised by a number of people was the fact that the more 11 flexible systems would give you large displacements.
i 12 DR. KENNEDY:
Yes.
13 CHAIRMAN SHACK:
That doesn't seem to be anything 14 that has been addressed by this, fs
~ k -)
15 DR. KENNEDY:
Well, I think it is.
I should have m
16
-said more.
17 If you limit your moment to these ultimate dynamic 18 moments divided by a factor of safety something in the range i
19 of 1.1 to 1.5, typically if you go up as high as 1.5, every j
20 one of these hysteresis loops shows that you're still 21 effectively pseudo-linear.
22 That means that pseudo-linear dynamic analysis 23 will do a reasonable job of predicting displacement so that 24 you'll know whether you're displacements are too large or 25 not, and we're not getting non-linear displacements if we ANN RILEY & ASSOCIATES, LTD.
Court Reporters 1250 I Street, N.W.,
Suite 300 Washington, D.C. 20005 (202) 842-0034 E-________--_____
93 1
limit to that level.
()
2 Now, some of the people who do not like the idea 3
of limiting to something less than ultimate dynamic moment, 4
what their concern is, most real piping systems may have 5
large non-linear factors and we're not giving a lot of 6
credit to that because we're worried about other regions.
{
l 7
My position is if you want to take credit for 8
that, you've got to do non-linear analysis, and I don't 9
think a code that's based on people being allowed to do 10 linear analysis can go -- that it's safe to go beyond that 11 ultimate dynamic moment.
12 DR. SHAO:
You're saying using high limits, you 13 have to --
14 DR. KENNEDY:
If you went beyond the ultimate C\\
(,/
15 dynamic moment divided by about 1.1 --
16 DR. SHAO:
Then you would do non-linear.
17 DR. KENNEDY:
I think you would have to do 18 non-linear analysis.
But if we limit to that level, I think l
i 19 the displacement issue is a non-issue.
20 DR. SHAO:
But that would increase --
21 DR. KENNEDY:
And I think the 4.5 S sub M will 22 work as long as the B-2 is -- I mean, 4.5 S sub M with B-2 23 of one is too high for straight pipe.
I mean, I would be 24 very concerned if we had a code that would allow straight i
25 pipe to be~ designed for 4.5 S sub M.
O-ANN RILEY & ASSOCIATES, LTD.
\\
Court Reporters 1250 I Street, N.W., Suite 300 Washington, D.C.
20005 (202) 842-0034
94 1
DR. SHAO:
But if you did non-linear, of course,
()
2
.your cost of analysis would go way up.
3 DR. KENNEDY:
Well, that's argumentative.
4 DR. SHAO:
Yes.
5 DR. KENNEDY:
I don't think non-linear analysis is 6
very costly anymore.
But I will agree with you, whenever I 7
have suggested to people they consider non-linear analysis 8
9 DR. SHAO:
They don't.want to do it.
10 DR. KENNEDY:
-- they've almost driven me out of 11 the room.
12 MR. BARNES:
Mr. Chairman, part of the expenses is 13 learning the non-linear techniques for the individual f
14 expense.
15 The point, see, I think codes have got to be 16 practical.
)
i 17 DR. KENNEDY:
Yes.
18 MR. BARNES:
And we can't afford to have the pipe 19 wiping out the vessels.
I mean, the pipe is not the god, if 20 you understand; the pipe connects the things that really 21 matter.
So in any design technique we come up with, in any 22
- 1 imitations we put on, we have to take account of 23 deflections and limit them through design methodology.
24 That's the point.
And what we need is engineering tools to 25 predict that, of course, and what I believe has happened is O
ANN RILEY & ASSOCIATES, LTD.
Court Reporters 1250 I Street, N.W.,
Suite 300 Washington, D.C.
20005 (202) 842-0034
l i
l 95 L
1 that the CALTEC experiment, the CALTEC work that was done
()
2 that has led Bob to be able to work with this ultimate l
l 3
dynamic moment, has given us an engineering tool.
t 4
I'll just touch on that when I get a chance.
i 5
CHAIRMAN SHACK:
Okay.
Well, we're running a l
6 little bit late.
7 MR. BARNES:
I know.
I just want to get a chance.
8 CHAIRMAN SHACK:
Perhaps we can catch up in just a I
9 few minutes here.
10 The next presentation is from our -- I have Mr.
11 Yamazaki on the agenda, but actually Professor Asada and Mr.
l 12 Yamazaki will both be making presentations.
I 13 MR. ASADA:
Good afternoon, gentlemen.
It's a 14 great pleasure to introduce our Japanese experts on plastic 15 seismic design criteria for nuclear power piping.
They are 16 here.
We have four who I am pleased to introduce, our 17 project:
Mr. Masahisa and Mr. Higuchi, Mr. Hokuriku and Mr.
18 Ono.
19 Our presentation is composed of three parts.
20 There's the concept of our project, and the third is the 21
. test result.
I will explain the concept and Mr. Masahisa 22 will explain test results.
I 23 Okay.
First, our concepts are very simple.
We 24 put our. idea on the pipes and the components on the -- so i
25 the failure occurs, so if-we can predict the stress behavior l
l f
l fi ANN RILEY & ASSOCIATES, LTD.
V Court Reporters 1250 I Street, N.W.,
Suite 300 Washington, D.C.
20005
.(202) 842-0034
96 1
with local strength, so we can predict the seismic behavior I ).
2 of.the pipe.
Also, we concur -- I'm sorry -- we take into v
3 account the plastic behavior of the pipe.
4 We consider the loading condition for the piping 5
system and the seismic condition, the internal pressure.
IF 6
the seismic load is very high, particularly at a high stress 7
portion, and as a result, so-called ratcheting occurs, this 8
is the history of the strength under the ratcheting, the 9
formation.
10 So if seismic load is very high and also operation 11 load is very high, this curve or this trend comes this way 12 through these -- that will collapse, you know.
13 So this diagram shows schematically two extremes.
14 So almost all cases, the piping system caves under seismic'
()
15 load in-between two extremes.
So this is so-called 16.
ratcheting fatigue, you know.
17 This is composed of two parts, that are fatigue 18 damage and ductility damage, and the designs show the 19 limitation so failure doesn't occur inside the curve and the 20 outside fail curve.
The curve is corresponding here.
21 That's the extremes of that, and the fatigue is located 22 here.
23 So we -- our project yielded this equation on a 24 unique failure mode.
The second part is a component test 25 composed of two parts, the formation control stress test,
/(]/ '
ANN RILEY & ASSOCIATES, LTD.
Court Reporters 1250 I Street, N.W.,
Suite 300 Washington, D.C.
20005 (202) 842-0034
97 1
stress seismic test, and also acceleration control, E
2 so-called dynamic test.
These two tests, component tests 3.
intend at first to verify the element of pressure; second, 4
to identify the different loading, the dynamic loading or 5
so-called strength control condition or acceleration control 6
condition, you know?
7 In this case, dynamic test in between the strength 8
control and the load control condition, you know?
So these 9
are different.
There's seismic loading and dynamic loading.
10 Okay.
The third part is detailed analysis to a 11 very high seismic test -- seismic test, and also to extend 12 the test condition to the large diameter piping, or more 13 complicated piping system.
So we don't make the experiment, 14 but we make an experiment to incorporate our idea.
. (D
(/
15 Finally, we make a coordination based on that 16 numerical experiment, so our -- we are intending to develop 17 a quote applicable for the analysis.
18 So next, Mr. Oyamada will explain this.
19 MR. OYAMADA:
My name is Oyamada, and I would like 20-to explain about the outline of the project.
21 This is the summary of the current result.
We 22 have performed some dynamic test and static test, strength 23 control test, and both static and dynamic test results 24 showed a lot of fatigue failure.
And this loading condition 25 in this test was elastic primary stress was from 5.5 SM to k'--
ANN RILEY & ASSOCIATES, LTD.
Court Reporters 1250 I Street, N.W.,
Suite 300 Washington, D.C. 20005 (202) 842-0034
98 1
12.25 SM, very large volume due to seismic loading and
()
2 internal pressure.
In spite of such severe condition, 3
collapse has never occurred.
We read from this result this 4
strongly supports the validity of ASME present a new 1994 5
addendum.
6 This is the second summary, and both static and 7
dynamic test results show the same elastic plastic and 8
failure behavior.
This similarity of static and dynamic 9
test result will be explained later by Mr. Yamazaki.
10 This summary is about the ongoing effort we are 11 planning now to extrapolate test data by analytical study, 12 which this was explained by Professor Asada, and for 13 example, we analyzed failure behavior under high temperature 14 conditions like that, and then we will establish new O
l V 15 criteria.
16 Our Japanese seismic criteria is slightly 17 different from that of ASME, but anyway, our basic idea is 18 established on the ASME criteria; therefore, your criteria 19 is-very important for us.
Also, we have a plan to do the 20 large piping component test and a piping system test.
1 i
21-DR. SHAO:
What are the differences between 22 Japanese criteria and --
23 MR. OYAMADA:
Oh, yes.
That means, for example, 24 in United States, you have only ten cycle -- but in Japan, 25 we have to consider at least 60 cycles, or like this.
And I
ANN RILEY & ASSOCIATES, LTD.
Court Reporters 1250 I Street, N.W.,
Suite 300 Washington, D.C.
20005 (202) 842-0034 i
99 1
these 60 cycles or like that, we have to -- because of these
()
2 60 cycles, we have to analyze the fatigue allowances.
3 In your ASME code, you have not -- you do not have 4
to calculate the fatigue analysis, but in Japan, we have to 5
do the fatigue analysis.
6 DR. SHAO:
How about your damping values?
7 MR. OYAMADA:
Also about the damping value, in 8
Japan, the damping value itself is very low compared to 9
United States.
10 DR. SHAO:
How low?
11 MR. ONO:
Maximally, it's 2.5 percent maximum.
12 DR. SHAO:
2.5.
13-MR. ONO:
Yes.
14 DR. SHAO:
So you are very conservative, then.
15 MR. OYAMADA:
Oh, yes, very conservative.
We are 16 also trying to increase the damping value, but --
17 DR. SHAO:
I have heard it for the last ten years.
18 You haven't increased.
19 MR. OYAMADA:
Okay.
20 DR. SHAO:
Okay.
Thank you.
21 MR. OYAMADA:
And then this is the proposal, and 22 we. propose to continue technical discussion for ASME new 23 piping sizing in special working groups on the seismic rule.
24 We understand that so many questions just in the i
25
-- in the 1994 addenda, but the -- for example, test 37 ANN RILEY & ASSOCIATES, LTD.
Court Reporters 1250 I Street, N.W.,
Suite 300 Washington, D.C.
20005 (202) 842-0034
100 1
collapsed, showed a collapse, but that was very, very l()
2 special case.
3 For example, that -- for that test, schedule 10, a 4
very thin pipe was used.
But that kind of a thin pipe is 5
never used in the primary system of a -- very important 6
system of a nuclear power plant.
By that, I mean that I 7
believe that the -- not the collapse, but the ratcheting 8
fatigue is important in the seismic design.
And second is 9
that our ongoing effort will be opened step by step after 10 the approval procedure, and that will help you.
11 Then we have two kinds of projects.
The first one 12 is a utility project, and this utility project is jointly 13 funded by ten utilities and three manufacturers in Japan, 4
14 and the purpose of this project is to establish
/^N
(,)
15 elastic-plastic seismic design criteria for nuclear piping.
i 16 The second is the National Project.
This National 17 Project is funded by MITI, our government, and this is 18 performed at NUPEC nuclear power plant, Power Engineering 19 Corporation, and this is to verify new seismic design 20 criteria for nuclear power piping.
21 This is the overall schedule.
This is the UTT 1
22 project and this is the National Project.
We investigated 23 in Phase 1 of the UTT project and we performed some material 24 and components tests in Phase 2 and we are now in the stage l
l 25 over here, and we are analyzing the test results.
1 lgp}
ANN RILEY & ASSOCIATES, LTD.
Court Reporters g
1250 I Street, N.W.,
Suite 300 Washington, D.C.
20005 (202) 842-0034 l
l 101 1
The National Project will start within this year
()
2 and continue until 2005 and this includes large bore 3
component piping and the piping system itself and we are 4
planning to do some standardization or making new rules in l
5 this stage and authorization after NUPEC seismic probing 6
test.
l 7
Then next is the results of components tests and 8
Mr. Yamazaki will explain.
9 MR. YAMAZAKI:
My name is Yamazaki I will 10 explain about the piping test results in Japan.
11 This figure shows an overview of our test program.
12 The study consists of material tests and piping component 13 tests and analytical tests.
14 In our material tests we have done the tensile (h
(,)
15 tests, low cycle ratcheting fatigue tests and cyclic stress 16 stain tests.
17 Under these tests, results are used to make up the 18 components test plan and will be used in the analytical 19 study --
20 CHAIRMAN SHACK:
Excuse me.
The cyclic stress 21 stain just means you determine a cyclic stress stain curve 22 from a small specimen?
23 MR. YAMAZAKI:
How large a stress?
24 CHAIRMAN SHACK:
On a pipe --
25 MR. YAMAZAKI:
No, no, not pipe.
Small specimen b}
ANN RILEY & ASSOCIATES, LTD.
Court Reporters
~
1250 I Street, N.W.,
Suite 300 Washington, D.C.
20005 (202) 842-0034
102 1
test.
(3 2
i)
And we have a piping component test here.
It 3
consists of two parts.
One is a forced displacement test 4
and namely a dynamic test.
We call this test a static 5
cyclic test, and another is load control, the test specimen 6
on a shaker table.
We call this test dynamic test.
7 All those results will be contributed to establish 8
that into the new rules.
Next slide, please.
9 The purpose of my component test is to investigate 10 elastic plastic behavior on failure mode of piping 11 components and the high level seismic loading.
12 This is test condition.
We set up the pipe bend, 13 tee, and straight pipe.
The material may need carbon and we 14 selected stainless steel to compare the materials and n
)
15 diameter is 4B.
The thickness is Schedule 14 and Schedule 16 160.
Schedule 14 is mainly used in Japanese nuclear power 17 plar.ts.
18 On test condition two is we added to that hoop 19 stress to the pipe by internal pressure, that is equivalent 20 to Sm and 0.5Sm nonpressurized.
21 Loading condition is static test and i
22 displacements, and dynamic tests and load controls.
23 Input wave, sine seismic wave so seismic random 24 wave -- in plane -- out of plane.
25 Loading level is 5.5Sm and then 12.25 Sm.
I)
ANN RILEY & ASSOCIATES, LTD.
{
Court Reporters 1250 I Street, N.W.,
Suite 300 Washington, D.C.
20005 (202) 842-0034 1
u
103 1
Before starting the video I will explain our test
()
2 specimen and test facilities.
3 This figure shows that in-pline test facility or 4
bending pipe was -- this is static and this is dynamic 5
testing.
6 The test specimen is not elbow fitting but bending 7
pipe with 228.6 millimeter bending -- namely at 2 DR.
8 In the test, one side is fixed. This is weight 9
about 1300 kilogram.
This is in-plane test so for the 10 dynamic test the weight is guided in this plane, this 11 direction on the shaker table.
12 This is the out-of-plane bending test facilities.
13 In the out-of-plan tests the actuator placed on 14 the wall and forced displacement applies in this.
15 The out-of-plane tests, test weight is here.
This 16 is bending pipe.
Weight is from the top because we want to 17 apply only inertial force for tests.
18 CHAIRMAN SHACK:
Excuse me again.
Why are you 19 using bent pipe rather than elbows?
Is Japanese piping 20 designed all bent pipe?
I 21 MR. YAi%ZAKI :
In Japan probably loop.
We tend to 22 use bending pipe.
23 CHAIRMAN SHACK:
Yes, but how large a diameter can 24 you do that for?
25 MR. OYAMADA:
2 D is a minimum diameter but in
/
~
ANN RILEY & ASSOCIATES, LTD.
Y Court Reporters 1250 I Street, N.W.,
Suite 300 Washington, D.C.
20005 (202) 842-0034
104 1
Japan for almost all the case bending pipe is used.
,G
(
)
2 CHAIRMAN SHACK:
Interesting.
v 3
MR. YAMAZAKI:
This is a table bending test 4
condition, with static and dynamic tests.
5 We have done 11 static tests and 10 dynamic tests, 6
each of number one is a basic test case here and here.
7 On the other test cases are planned in compliance 8
with material, thickness and the internal pressure.
9 Thickness for number one, material carbon.
Thickness is 40 10 and pressure is S by M.
Loading condition is sine wave and 11 the stress level is in this case 12.25 Sm.
Direction is 12 in-plane.
Static condition.
13 The next is strain gauge -- around 40 strain gauge 14 placed on the outer surface of piping components.
/~~
(_)%
'15 Strain gauge layout was determined by analysis.
16 This figure shows the layout of displacement and 17 reaction force instrument.
18 This picture shows the surface of bending pipe, 19 Left picture is outer surface and that picture is inner 20 surface.
We can find the crack status from inner surface 21 because inner surface crack is here, out of surface is 22 here -- inner surface is longer than outer.
23 This table shows the comparison of static tests 24 and dynamic tests on the same stress -- for example static 25 carbon steel stress level.
!O ANN RILEY & ASSOCIATES, LTD.
5/
Court Reporters 1250 I Street, N.W.,
Suite 300 Washington, D.C. 20005 (202) 842-0034
105 1
1 I think it's a phenomenon that is very important j
()
2 for study.
We can use high temperature material test 3
results to estimate dynamic behavior in high temperature.
4 This figure shows the relationship between the 5
action force and displacement from left to right the figure 1
6 indicates all the stages of the loading cycle.
Dashed line 7
is static test results and the solid line is the dynamic 8
test results.
9 For the dynamic test inertial force is applied on 10 test specimen.
A test specimen is deformed by testing.
11 This is strain history data.
We can say that 12 strain behavior of static and dynamic tests are almost the 13 same.
14 This is a comparison between the stress analysis 15 and test results.
The location was in agreement with 16 calculated maximum stress portion.
17 We have not much time so let's start the video.
18
[ Video shown.]
j 19 MR. YAMAZAKI:
I show you the static tests, cyclic l
20 tests of bending pipe, test specimen and dimension.
This is i
l 21 here.
Material is carbon.
22 This is an overview of a test specimen and test 23 facilities.
24 The bottom pipe fixed and actuator placed on the 25 top, total 36 strain gauge -- outer surface of the pipe.
'O ANN RILEY & ASSOCIATES, LTD.
Court Reporters 1250 I Street, N.W.,
Suite 300 Washington, D.C.
20005 (202) 842-0034
106 1
Displacement on the reaction force --
2 T1is cycle is one cycle per minute.
3 This is a close-up of a crack -- maybe you can 4
find it, the water drop near here.
5 In this test internal pressure is zero but we 6
apply slight repressure to the crack.
7 We can find that -- it's a water drop -- 63 cycles 8
in this test case.
9 Next bending pipe and dynamic tests.
Test 10 condition, pipe dimension is same as static cycle tests 11 except dynamic condition.
We added 4.7 hertz and the 12 maximum oscillation is 1800.
13 This is 2800 kilogram weight.
In this test, 14 response to displacement was just plus/minus 33 millimeter, 15 same as static cycle tests.
16 This is the view from over the side.
17 This is the test specimen after the dynamic tests.
18 Crack was found on the side of bending pipe.
Crack 19 configuration was straight along with a curve --
20 Next is dynamic test of bending pipe for stainless 21 steel.
22 In this test through-wall crack was found at 121 23 cycles, longer than that of carbon steel.
24 In this test through.rall crack was found, same 25 portion of the carbon steel pipe but for stainless steel G
ANN RILEY & ASSOCIATES, LTD.
Court Reporters 1250 I Street, N.W.,
Suite 300 Washington, D.C.
20005 (202) 842-0034
107 1
outer diameter was expanded.
l( )
2 This is the crack.
3 In this test displacement adjusts plus/minus 50 l
4 millimeter -- under input cycle.
5 Next is the tee-shaped fitting static cyclic 6
tests.
Piping dimension is same as bending pipe.
In this l
7 test case sine wave was -- sine wave forced displacement is 8
plus/minul 50 millimeter.
This displacement causes around l
l 9
12 Sm.
It's a little sloe.
)
10 Next thing the tee-shape fitting -- test specimen 11 is same as static cyclic test.
I 12 Displacement of top branch pipe is just the same 13 as static cyclic tests -- weights, input cycles -- 1300 14 weight was supported by wire to compensate dead weight.
O)
's,
15 A through-wall crack was found at 157 cycles in a m
16 tee-shaped fitting.
17 Not interesting?
We stop this video and I would 18 say our current conclusions.
.19 In our tests failure mode is fatigue.
Fatigue was 20 ratcheting so from our test results up to around 12 Sm, 21-piping component doesn't collapse but failure by fatigue.
1 22 We have planned to analyze what condition will 23 induce plastic collapse.
We have high tempe-
.c material 1
24 test results, so we have planned to analyze the piping I
25 response by.the force in high temperature.
[
ANN RILEY & ASSOCIATES, LTD.
Court Reporters 1250 I Street, N.W.,
Suite 300 Washington, D.C. 20005 (202) 842-0034 I
108 1
We hope that our ongoing research will contribute
![
2 to new seismic endorsements.
Thank you.
l%
3 CHAIRMAN SHACK:
You haven't completed your l
4 element analysis though to look at the ratcheting analysis l
5 with the Asada criterion?
l 6
MR. YAMAZAKI:
No.
The analysis we are starting l
l 7
this fiscal year so the results will be opened next March i
8 but we the will to open our test results from December ASME 9
meeting.
l 10 CHAIRMAN SHACK:
Any additional questions for j
l 11 Mr. --
l l
12 MR. MANOLY:
Yes, one question.
What kind of 13 index that the 12.25 --
l 14 MR. YAMAZAKI:
I have planned to calculate B2 in 15 this fiscal year.
16 CHAIRMAN SHACK:
How did you get the 12.25?
I 17 mean I assume that was a code pseudo elastic analysis with a 18 B2 of some number, right?
19 MR. ASADA:
That's right.
20 CHAIRMAN SHACK:
What number did you use for B2?
21 MR. ASADA:
Number?
22 CHAIRMAN SHACK:
What value?
23 MR. YAMAZAKI:
No, we corrected -- the strain 24 level is 3 percent.
Yes.
l.
25 MR. ASADA:
a LOCA strain --
ANN RILEY & ASSOCIATES, LTD.
Court Reporters 1250 I Streat, N.W.,
Suite 300 Washington, D.C.
20005 (202) 842-0034 i
109 1
MR. JAQUAY:
It didn't seem t me that if you had a
()
2 collapse mechanism that the test boundary conditions would 3
allow it to happen, nor did it seem you had very high weight 4
loads that might cause p-delta.
5 To say they did not collapse, I wonder if that is 6
misleading some people to say it won't collapse in other 7
test setups.
8 MR. YAMAZAKI:
The purpose of our tests, many for 9
inertial force.
10 MR. ASADA:
Okay.
11 MR. BARNES:
I thought about that point, Ken, and 12
-- Richard Barnes -- and the thought that hit me was that 13 what you would start to see is a dimpling effect or 14 something like that if you started to see collapse and there 15 was no obvious evidence of that.
16 Now, Professor Asada and his colleagues are going 17 to make this available to the special working group and we 18 can discuss this in detail.
Obviously this is preliminary 19 and they speeded it up so they could actually be part of the 20 hearing and provide information, and obviously there is 21 going to be more discussion in September and December, 22 especially December when they bring it out.
23 So I think we can pursue some of those points then 24 more closely.
But I think that's probably the explanation, 25 that you would have started to see the dimpling in if it had
()
ANN RILEY & ASSOCIATES, LTD.
Court Reporters 1250 I Street, N.W.,
Suite 300 Washington, D.C.
20005 (202) 842-0034
l 110 1
a collapse type mechanism.
()
2 DR. OYAMADA:
Excuse me.
Dr. Shack asked how much 3
largeness of a bending pipe can be fabricated.
I can answer 4.
that.as far as a 24-inch pipe can be bended.
1-5 CHAIRMAN SHACK:
Thank you.
6 If there are no more questions, then perhaps we 7
can move on to our lasc presentation of the evening.
8 MR. BARNES:
How many minutes are you prepared to 9
give us so we'll cut it accordingly?
I i
10 CHAIRMAN SHACK:
No, take your time.
This is very 11 interesting and we're certainly willing to --
12 DR. POWERS:
If you do anything less than two 13 hours1.50463e-4 days <br />0.00361 hours <br />2.149471e-5 weeks <br />4.9465e-6 months <br />, you're still within our schedule.
14 CHAIRMAN SHALK:
I suppose I should ask whether
)
)
15 any emergency leak before break situations are occurring and 16 you need five minutes --
17 MR. BARNES:
I would appreciate it.
18 CHAIRMAN SHACK:
I think a couple of us have snuck 19 out and back and forth.
20 MR. BARNES:
I would appreciate one more attempt.
j 21 I have to actually go ahead and book a room for a hotel L
22 because I hadn't confirmed the other one.
23 CHAIRMAN SHACK:
Okay.
24 MR. BARNES:
Can I have a couple of minutes?
25 CHAIRMAN SHACK:
Sure.
Let's take five, then.
1
/)
ANN RILEY & ASSOCIATES, LTD.
-/
Court Reporters 1250 I Street, N.W.,
Suite 300 Washington, D.C.
20005 (202) 842-0034
111 1
[ Recess.]
()
2 MR. EISENBERG:
While Dick is getting his 3
microphone on, I just want to thank the subcommittee for 4
inviting us to speak today.
5 I'm Jerry Eisenberg, director of Nuclear Codes and j
6 Standards at ASME, and this is Richard Barnes, who is 7
chairman of the subgroup on design of the Subcommittee on 8
Nuclear Power.
9 DR. POWERS:
Maybe you could, as a representative 10 of ASME and a representative of a consensus process, explain 11 to me how these thing get to -- when you have some 12 controversy where some significant portion of the technical 13 community is objecting to a candidate's standard, it goes 14 ahead and gets published despite that objection.
rN i
()
15 MR. BARNES:
Can I handle that -- can I handle it 16 in my --
17 DR. POWERS:
Sure, you can.
18 MR. ELSENBERG:
In any event, a lot of the i
19 background history, you've already heard, so we're going to 20 skip over that and go right to Mr. Barnes' presentation.
21 MR. BARNES:
Okay.
Mr. Chairman, gentlemen, what 22 I have attempted to do in the presentation is to take a l
23 philosophical point of view to see where we're going and 24 where we might end up and sort of have a look at the road 25 ahead.
So I don't want to cut out a couple of those; I
/\\
ANN RILEY & ASSOCIATES, LTD.
\\ s/
Court Reporters 1250 I Street, N.W.,
Suite 300 Washington, D.C.
20005 (202) 842-0034
112 1
would like to keep that one back there.
Yes.
()
.2 The first thing is you asked a question of what is 3
the process, and I thought -- I put a note here just to 4
explain to you just the depth that it goes into.
5 We start with a group of experts, develop proposed 6
requirements.
Now, what you find is when you get a couple 7
of experts together, that you've got a variety of opinions 8
and they do fight, you know.
Quite literally, you get quite 9
intense.
And I understand even at this particular moment --
10 I don't attend all the special working groups, but the same i
11 sort of thing is happening.
The people have quite intense 12 discussions.
13 But what we have found is with the -- this special 14 working group reports to the subgroup.
It will bring 15 forward eventually positions on items at the subgroup.
If 16 thir change in code rules -- we're actually going to send it 17 back to the working group on piping.
18 So there's a group of, say, 20 or 30 people there, 19 20 people I suppose, so they will discuss it and they will 20
-voce on it.
It will then come back to the subgroup, that 21 group of 25 people there, they will discuss it and vote on 22 it, and over a period of two or three meetings.
It goes 23 from there to section 3, which is another 25 people of all 24 different background, a balance of backgrounds particularly 25 in section 3, and coming from large companies.
So that if I I)
ANN RILEY & ASSOCIATES, LTD.
\\/
Court Reporters
{
1250 I Street, N.W.,
Suite 300
]
l Washington, D.C.
20005 (202) 842-0034
113 1
am a representative of a large company, I'll take that stuff
()
2 back to my company, and most often, I'll have an expert in 3
there who is on -- I used to work for Ontario Hydro.
I 9
mean, they've got piping experts coming out of their ears because Hydro does a lot of its own design, or did do.
I'm 6
not quite sure where it's going today, but it did do it.
So 7
they would -- and they would give me feedback as a member of 8
section 3 and I would take that sort of information back.
9 It then goes from there to the main committee, 10 which is the actual consensus committee, and that's -- when 11 that one votes, that's what becomes the rule effectively.
12 It goes from there to the Board of Nuclear Codes and 13 Standards, but that is more or less meant to be looking at 14 more on the procedure and things like that rather than the 15 technical expert -- technical input, although you do get 16 technical comments.
17 The point I wanted to make is, you see, in these 18 particular rules that have gone through, effectively it was 19
-- at the consensus committee, it received just one 20 negative, and I'll tell you what we did with these rules.
21 In the subgroup design, we had two days, a Friday and a 22 Saturday in July of that year, where we looked at the first 23 time the rules were proposed to us.
They got all -- it went 24 out for a letter ballot; they came back with the 25 recommendations; and we had a two-day session in July where O
ANN RILEY & ASSOCIATES, LTD.
Court Reporters 1250 I Street, N.W., Suite 300 Washington, D.C. 20005 (202) 842-0034
i 114 1
we sat down and went over this with the full committee.
()
2 They then took all those comments back and we had 3
another'two-day session in September in New York City, and i
4 only after that did things start to go through the section 5
3.
)
6 So there was a huge amount of discussion, a huge 7
amount of input, extensive time given, and the truth is that I don't know if you realize this, but there were Canadian 8
9 regulatory representatives on subgroup design, they voted in 10
. favor of it.
11 We had jurisdiction, like in Canada, ve register 12 the designs.
We had the representative from Ontario who was 13 also on section 3 and on subgroup design, he voted in favor 14 of the rules.
(~N
'( s/
15 Our Japanese colleagues, after the discussion, 16 voted in favor of the rules, although they sent us a letter 17 explaining that there were certain areas with engineering 18 judgment which they didn't believe they could live with.
19 So the way we saw it as a committee, and I can 20 tell you, I can't promise -- you know, Bob Kennedy was 21 there.
He was saying, you can get ASME and NRC now to work 22 together.
I mean, 1 could never promise you that, because 23 these are a bunch of very individually minded people.
I 24 mean, you must see that in the special working group.
They 25 are not in anybody's pockets, they're not each other's
[~
ANN RILEY & ASSOCIATES, LTD.
\\
Court Reporters f
1250 I Street, N.W.,
Suite 300 Washington, D.C.
20005 (202) 842-0034 i
____________z______i___._____
115 1
pockets.
2 Unfortunately, sometimes it gets characterized as 3
NRC versus the rest, and I don't -- I think that's really 4
unfortunate because what I want to see, and in actual fact, i
5 in all truth and honesty, that did happen with the original 6
task group -- it was a very, very unfortunate thing.
And in 7
this task group, we set up a management committee between 8
NRC and ASME in case that sort of thing developed again and 9
we were going to take immediate action if we started to get 10 to that, and Larry Shao and I think Dick may have been at 11 the first meeting we had.
12 We have never had a need to call the meeting 13 together at this point that I have been aware of.
Nobody 14 has given me any feedback.
Because I was going to be hotter 15 than a firecracker on that.
I was not going to let it 16 happen again.
I was keeping a close check on it.
So we had 17 some personality developments in the first crv.p, but the 18 rules just weren't shouted through.
19 Now, I can tell you something else.
At the main 20 committee, there was a little bit of play on it.
I mean, 21 what happened was they used some procedures -- well, the 22 problem was -- I'll just tell you quite frankly -- we had an 23 evening where we had special presentations for section 3, we 24 had special presentations for the main committee.
25 Now, the evening that we were having it for the G
ANN RILEY & ASSOCIATES, LTD.
Court Reporters 1250 I Street, N.W.,
Suite 300 Washington, D.C.
20005 (202) 842-0034
116 1
main committee, it snowed like crazy and several of the NRC
[)
2 people couldn't get there, and I think Mark Hartman ended up LJ l
3 making the presentation, if I remember correctly.
4 Now, the representative on the main committee --
5 well, the person who was sponsored by NRC on the main 6 L committee said that they thought due process wasn't done, 7
that the snow had -- we should put off the vote, and the: 2 8
was a procedural battle, and unfortunately -- nobody won 9
real:y, but the procedural battle meant it went ahead.
You 10 know what I mean?
And that didn't really salve anybody's 11 ego or anybody's feelings of --
12 MR. ELSENBERG:
If I could just try and zero in on 13 the process that you asked about.
In the consensus process, 14 there are a balance of interests, and there may be one O)
(
15 interest or one category of interest that -- or even one 16 individual who may object, and I think that the -- as long 17 as we follow our established procedures and that we consider 18-the negatives and respond to them appropriately, and in this 19 case, we even went an extra -- took the extra effort to make 20 presentations to try to resolve these.
The process has to 21 be served because consensus doesn't necessarily mean 22 unanimity.
23 So there may be a manufacturing sector that may 24 object, there may be a regulatory sector that may object, 25 but in the end, it is according to the procedures how the 1
/'
l k- \\
ANN RILEY & ASSOCIATES, LTD.
/
Court Reporters 1250 I Street, N.W.,
Suite 300 l
Washington, D.C.
20005
(
(202) 842-0034 0
1 i
l 117 1
process is served.
So any one category cannot leverage the
()
2 committee to the extent that it would hold up an action.
3 DR. KRESS:
What does consensus mean?
l 4
MR. ELSENBERG:
Consensus allows for all of the 5
input from the various sectors to be heard, and it doesn't 6
necessarily mean unanimity; it means --
l l
7 DR. KRESS:
Does it mean majority?
8 MR. ELSENBERG:
It doesn't -- well, it l
9 MR. BARNES; Let me take it on.
10 MR. ELSENBERG:
All right.
11 MR. BARNES:
In the main committee, which this set 12 of rules went through, the first vote, one negative vote 13' will hold it up and one negative vote held it up.
It then 14 goes back for consideration by the subcommittees and working O(_j 15 groups if need be, and answers come back up.
And providing 16 there has been no major technical change, it goes back for 17 what is called second consideration, and then if it doesn't 18 get -- anything -- if it gets four negatives, it's defeated.
19 DR. KRESS:
How many members of the committee?
l 20 MR. BARNES:
Thirty.
21 MR. ELSENBERG:
There are 30 members of the 22 committee and --
j 23 DR. KRESS:
If it gets four negatives, it's 24 defeated.
'25 MR. ELSENBERG:
That's right.
So you have to get l
l ANN RILEY & ASSOCIATES, LTD.
\\-
Court Reporters 1250 I Street, N.W.,
Suite 300 Washington, D.C.
20005 (202) 842-0034
118 i
i 1
at least 90 percent of the vote in the end for it to pass.
(
2 MR. BARNES:
And these are people from --
v 3
MR. ELSENBERG:
In fact, that goes even more 4
restrictive than many of the other consensus committees that 5
we have which only require two-thirds.
Even Congress only 1
6 requires two-thirds.
l 7
DR. POWERS:
I mean, I guess what disturbs me or 8
puzzles me a little bit -- maybe it doesn't puzzle me; it's l
9 just a concarn that I'll express here -- is that when the 10 final rule comes out, suppose you did have a manufacturer 11 that had an axe to grind and had raised an objection.
Does 12 the final rule reflect his objection?
I mean, is there 13 something --
14 MR. ELSENBERG:
If there is a public objection,
/O
(_)
15 that is another part of the process.
We have public review 16 of all of our actions.
And this did go out for public 17 review.
We had no comments in the public review.
The only i
18 comments of a negative sort were from the individual who was 19 sponsored by the NRC on the main committee and the board, 20 and that was the only negative that we had, and we went 21 through -- I think we went even further than the procedures 22 require to respond to that, again with presentations and so 23 forth.
24 If there is a public review comment, it has -- all 25 public review comments have to be considered in the same way I
l L
[\\-]
ANN RILEY & ASSOCIATES, LTD.
Court Reporters 1250 I Street, N.W.,
Suite 300 Washington, D.C.
20005 (202) 842-0034
l' 119 1
that you would have to consider them in the Federal Register
()
2
.for a rule, and they have to be responded to.
3 Due process really means the right to appeal an 4
action.
It's not only the consensus process up to it; it's 5
the right to reconsider or appeal an action.
But we didn't 6
get any such public comments on this particular action.
7 DR. SHAO:
You get --
8 MR. BARNES:
No, the NRC representative, or the 9
person sponsored by the NRC, they --
10 DR. SHAO:
Gil Millman, right.
11 MR. BARNES:
Yeah, Gil Millman, he put in his vote
)
12-as a negative, and comments.
On both main committee and 13 board committee, board level.
Yeah.
And were appropriately i
14 answered.
I mean not obviously to his satisfaction.
15' DR. SHAO:
And one thing I would like to say.
I 16 was out of there, okay.
17 MR. BAPNES:
Sure.
18 DR. SHAO:
I don't want to influence my people.
19 But there are seven, eight people who have extensive piping 20 experience working in NRC, both on research -- they are 21 unanimous against it.
This is not say, -- so far all the 22 staff have been unanimous.
Also --
23 MR. BARNES:
Can I -- can I -- see, the trouble 24 with that is -- I mean I did logic when I was in my last 25 year of high school, and the one thing is argument by l
ANN RILEY & ASSOCIATES, LTD.
Court Reporters 1250 I Street, N.W.,
Suite 300 Washington, D.C. 20005 i
(202) 842-0034 l
i
.____]
i 1
120 1
1 numbers'and argument by authority is never a logical OQ 2
argument.
3 DR. SHAO:
That's why I say -- it's very funny.
4 MR. BARNES:
Yeah, but now you see I can point to 5
the AECB who have piping experts themselves.
The government 6
body that I referred to, the jurisdiction, they have piping 7
experts, and they voted for it.
I mean Ontario Hydro has 8
piping experts.
9 You see, now I just -- the trouble with it is, is 10 that I can understand a group -- I mean I believe they had 11 concerns.
In actual fact, I would like to get on with my 12
' presentation and I'll address some of that.
13 DR. SHAO:
Yeah, go ahead.
I 14 MR. BARNES:
Because I actually believe that
)
15 there's -- I think there was a real learning curve here, and i
16 I think there has been real progress made.
But the point i
17 that I just wanted to make was that the first set of rules l
l 18 really had -extensive -- extensive review.
And the one other 19 point I would like to make is it was a fair review.
And how i
20 do I know that?
Because I was the one deciding it, and I 21 said it's fair.
No, that's not the reason.
But the reason 22 it was fair was I had people from both sides mad at me, for 23 different reasons.
And I can tell you that.
So I knew that 24 I was doing the right thing.
I wasn't very -- I wasn't very j
i 25 popular, that was all.
i
[)
ANN RILEY & ASSOCIATES, LTD.
i
\\- I Court Reporters 1250 I Street, N.W.,
Suite 300 l
r Washington, D.C.
20005 (202)' 842-0034
121 1
But, however, that is the sort of -- I hope you
()
2 have got an idea, and I hope -- does that answer your 3
question, Dana, about --
l 4
DR. POWERS:
It confirms my belief that the l
5 consensus process has fallen behind the times in the area of 6
expert elicitation, and this is a process that is guaranteed 7
we will never see expert number 5 in seismic and things like 8
that.
I mean we just never are going to find out about this 9
range of uncertainty with this kind of a process.
And I 10 think that's something these consensus committees are just 11 going to have to get up to date on how you do expert l
12 elicitation.
13 MR. BARNES:
That's a good point.
Because I l
14 actually think that -- I think that point came out of the
()
15 study as well.
16 Okay.
The next.
The next overhead, I just want 17 to -- it briefly shows you -- six, please.
Sequential.
18 These are things that I just want to talk about.
This is 19 just, I want to talk about the perspective, I want to have a 20 brief review of this report, the potential for resolution, 21' the basic conclusions, the road ahead.
Okay.
Next one.
22-Now, we actually -- we do try to get expert 23 elicitation, to some extent.
But the trouble is how much is 24 it and I mean just the depth of it.
But the engineering L
25 expertise is certainly required to develop code work.
Why?
l O
ANN RILEY & ASSOCIATES, LTD.
-/
Court Reporters 1250 I Street, N.W.,
Suite 300 Washington, D.C.
20005 l
(202) 842-0034 L
l
i l
122 1
Because we are developing it for plants, we are developing AQ 2
it'for systems and so.
3 Now, there is no doubt it, the people who are in j
l 4
the code work mainly are engineers.
We have specialty l
5 groups like we have a subgroup fatigue strength, which has 1
6 people who do the research in fatigue, and any questions on 7
fatigue or fatigue curves go back to that committee.
We 8
have a subgroup on materials.
They are sort of service j
i 9
groups.
We have a subgroup -- we have a subcommittee on I
10 non-destructive testing.
11 On the other hand, when you get into experimental j
12 expertise -- experiments, you need experimental expertise to 13 understand and to develop it.
Now, these two, to some 14 extent, are impossible mixes, because engineers see things 15 in a broader scale.
They take information.
They take 16 effectively the basic laws of physics and they say, well, 17 how can we apply them to the benefit?
How can we apply it 18 so that we produce something that is safe?
19 It doesn't allow them most times that sort of 20 thought process to get into the details that the 21 experimentalist, the experimental person works at.
So I see 22 this as a problem area, which I think to some extent makes 23 your point.
That how you elicit the expert, you know, 24 sufficient expertise even is the word.
Because what we did, l
25 you have got to have a team, and we actually had some team.
ANN RILEY & ASSOCIATES, LTD.
Court Reporters 1250 I Street, N.W.,
Suite 300 L
' Washington, D.C.
20005
.(202) 842-0034 p
j 123 f
1 We had people from EPRI, we had people from GE putting input J
l 1
()
2 in it.
This time we had people from ETEC putting it in.
So l
3 it's not that we didn't have some of that elicitation, it's 4
just that we had different people, and we had Dr. Kennedy as 5
well.
I me we didn't actually have it, but meaning as a
{
1 6
group, we had these sort of inputs from people.
j 7
DR. SHAO:
Chairman, may I make a comment?
8 CHAIRMAN SHACK:
Yes.
9 DR. SHAO:
Do you have people from Section 11 10 there or material expert like inspection people?
11 MR. BARNES:
Not on the special working group.
12 DR. SHAO:
I think you should really, in order --
13 eventually, whatever you come out, any time you cannot be --
l 14 it's not practical.
It is going to be a wasted effort 10
(_)
15 again.
16 MR. BARNES:
I'm not sure -- Ken, do you know if 17 we had Section 11 people coming along?
18 MR. JAQUAY:
No, just --
19 MR. BARNES:
I think they came along and put the 20 input in.
That's how we got the issue, wasn't it?
They 21 gave us -- they sent a memorandum around.
But you're right, 22 I agree.
23 DR. SHAO:
It's just a comment.
24 MR. BARNES:
Yes.
25 DR. SHAO:
Because a lot of good work is being j p)
ANN RILEY & ASSOCIATES, LTD.
Court Reporters 1250 I Street, N.W.,
Suite 300 Washington, D.C.
20005 (202) 842-0034
124 1
done, somebody can come up with a different view.
Okay.
()
2 MR. BARNES:
Yes.
Yes.
3 DR. SHAO:
And you should solicit their view.
4 Their view can be important.
If whatever you do, it cannot 5
be inspected later on.
I think this view can be important.
6 Whatever you have done work, maybe you nee.d something else.
7 MR. BARNES:
I understand.
Yeah, you're right.
8 And in actual fact, can I say, we have really 9
tried --
10 DR. SHAO:
Okay.
11 MR. BARNES:
-- to 3_,
as many people as we can.
12 But to be honest with you, the Section 11 people, we took 13 their issue, and I know that they formed the issue and 14 somebody has been working on it, and been working with the
(
15 Section 11 people, right.
I think the person who wrote it 16 was Dick Scott from Ontario Hydro, if I remember correctly, 17 he wrote the letter when the concept of the 4.5 came up.
18 UR. SHAO:
Somebody similar to Wilkowski's 19 background.
It doesn't have to be Wilkowski, but someone 20 similar to him.
21 MR. BARNES:
I agree.
I mean -- that's right.
22 And in actual fact, can I just say this, when the code rules 23 do go forward, maybe one of the steps we should put is we 24 should send it through to the Section 11 people.
25 DR. SHAO:
Yeah.
ANN RILEY & ASSOCIATES, LTD.
Court Reporters 1250 I Street, N.W.,
Suite 300 Washington, D.C.
20005 (202) 842-0034
125 1
MR. BARNES:
As well, before it even goes up to
()
2 the main committee and get comment from there.
l 3
Okay.
Next overhead.
Now, I would just like to 4
say that, No.
1, I think Ken did -- Ken and Dr. Chan, 5
unfortunately, did an excellent job in feeding back to the 6
'special working group.
And Nilesh Chokshi, I mean, obviously, the people were -- you people in research are 8
heavily involved in that particular aspect.
9 There was really good cooperation.
It started off 10 with suspicion, I can tell you that straight, there was no 11 doubt about that.
But in the end I can tell you the reports 12 that I got back were absolutely excellent.
So I feel that 13 this was achieved, and that and the fact that Bob Kennedy is
)
14 working so closely with the people and so on, I just feel
()
15 that has really developed.
16 The special working group, of course, while we 17 have had feedback and we had forerunners of the sort of 18 conclusions that were coming up, obviously, we can't really 19 respond to that report in this short time.
You know what I 20 mean.
We have to go through the process.
21 There were two comments.
The way I -- but I would 22 like to try and make a brief summary of the way I saw it and 23 the discussion I have had with a couple of people-like Andy 24 Branch and Don Landis.
It seemed to me that the ETEC study 25 concluded that the technical basis for the rules was ANN RILEY & ASSOCIATES, LTD.
Court Reporters 1250 I Street, N.W.,
Suite 300 Washington, D.C.
20005 (202) 842-0034
f 126 1
incomplete, and that was the word that was used in the r
l
/^N
(
)
2 report.
And a common conclusion -- there were two sets of I
3 conclusions, really.
If you really look at it, there was l
4 the ETEC conclusion, and then there was the independent 5
review team.
And I think *. hat's a fair characterization of j
6 it.
7 A common conclusion from the independent -- there 8
were several points made by them, but there was one sort of 9
commo1. thread and common sort of major point, that they
)
10 supported the ETEC -- it was that the Kennedy approach 11 provided the required additional basis to make these rules 12 at least defensible.
I think that is a fair 13 characterization of what I read at least.
And when you have 14 got four of the five independent review them, or however
()
15 many there are -- like, for instance, Shipley said, and when 16 I found out today that he was a designer, it made all'the
,17 more sense of what he said, let's take part of the system, 18 let's put some exclusion points out and start to use the 19 roles in a certain -- in a limited area, so we see what the 20 experience is of people.
l 21 MR. JAQUAY:
Can I comment?
22 MR. BARNES:
Yes.
l 23 MR. JAQUAY:
The choice of wording is probably 24 just semantics.
The Kennedy approach is a framework, it is 25 an. approach.
It hasn't the substance of a conclusion yet.
[h ANN RILEY & ASSOCIATES, LTD.
V Court Reporters 1250 I Street, N.W.,
Suite 300 Washington, D.C.
20005 (202) 842-0034
127 F
1 It doesn't form a basis yet until it gets filled in with the 2
numbers and the values.
3 MR. BARNES:
Of course.
I don't agree -- I don't 4
disagree at all.
I agree with you is the way of saying it L
5 in a positive sense.
But the point is that -- but, for l
6 instance, Iwan~-- Dr. Iwan said he would like to see a l
7
. displacement type approach.
Our people feel, that is the-8
. people who write code work, they just feel that that would 9
be such a major change in the way we do business in that 10~
area that it would be impossible for the industry to 11-swallow.
And, therefore, the approach that Bob Kennedy has 12 put forward, which has got the -- effectively ties in.
I 13 mean it's really very clever what you have done, because you 14 have taken -- you have taken the dynamic dimension of this, 15 and the.other failure mechanism, and identified something 16 that can be regarded as a static value, and put it back into
'17
.the static equation to make a comparison with.what the code 18 calculation, which is a static -- I-mean that really, in 19 many. ways, it'i's amazing how that spans the gap and gives' 20.
. consistency.
21 So that's why -- my only, I'll tell you what my l
22 only concern with it is, as I sat here and listened today, l
23
.is I think -- I wonder if people will ever understand that 24 Jthat equation, the way it exists today is--- what is the
'25.
. word?
I have forgotten -- empirical.
Because in its ANN RILEY & ASSOCIAT3S, LTD.
Court Reporters 1250 I Street, N.W.,
Suite 300 Washington, D.C.
20005 (202) 842-0034-
128 1
original form it'wasn't empirical.
('~'j\\
1 2
DR. SHAO:
'Let me ask a question.
If Kennedy -- I l
3 heard what he said, you have to do a nonlinear dynamic 4
analysis.
Can the code live with it?
5-MR. BARNES:
A linear?
6~
DR. SHAO:
Non-linear dynamic analysis.
7 MR. BARNES:
A nonlinear.
Well, he was saying 8
that if you wanted to fill in some of those -- get extra 9
margins.
10 DR. SHAO:
Yeah.
Right.
11 MR. BARNES:
The answer to that question is we are 12 now starting to look at how do we do nonlinear analysis.
13 DR. SHAO:
Nonlinear dynamic analysis.
14 MR. BARNES:
Nonlinear dynamic.
Nonlinear 15.
analysis in the static --
16 DR. SHAO:
No, nonlinear dynamic.
17 MR. BARNES:
I understand.
But I am just pointing 18 out, we are getting into the nonlinear role now in a way.
19 DR. KENNEDY:
This is Bob Kennedy.
Since you are 20 arguing over what I said, maybe.I didn't say it clear.
So 1
21 let me try to say it clear.
If you limit the code to the 22 ultimate -- and I called it ultimate dynamic moment, but 23 maybe we should call it-the ultimate cyclic moment.
If you
]
24 limit the code to that moment, divided by some reasonable 25 factor of safety, which in my judgment is in the l
i i
/(
ANN RILEY & ASSOCIATES, LTD.
Court Reporters 1250 I Street, N.W.,
Suite 300 Washington, D.C.
20005 (202) 842-0034
l 129 1
neighborhood of 1.1 to 1.5, if you limit to that, I believe
(( )
2 you can do linear analysis.
3 I do feel that we are still leaving a lot of 4
conservatism on the table for many, many real piping systems 5-by limiting to the ultimate dynamic moment.
For instance, 6
in displacement controlled tests, you can go beyond that.
7 But to take advantage of going beyond the ultimate dynamic 8
moment, I think you would have to do nonlinear analysis.
9 But up to the ultimate dynamic moment divided by some 10 reasonable factor of safety, the linear analysis is 11 sufficient and'probably leaves, for a lot of piping systems, 12 a lot of margin on the table.
13 MR. BARNES:
Now potential for resolution -- I 14 have heard this statement made many, many times from both
)
15 sides, meaning from the ETEC side and I am including Bob 16 Kennedy in the ETEC side at this point -- since you were i
1*
part of the Independent Review Team.
18 I have heard that from two or three members like 4
19 Ed Weiss.
I think I have heard it reported from Ed 20' Rodabaugh as well -- as vall as our own people believe that i
21' the Kennedy approach has the basis to solve it.
It is 22' actually in your report as well, you know what I mean?
23 So it does seem that we are not far apart.and that i
24 the study -
the special working group is studying Kennedy's
-25 approach and the discussion to date indicates deep interest,
'[ \\
ANN RILEY & ASSOCIATES, LTD.
\\-
Court-Reporters 1250 1-Street, N.W.,
Suite 300 Washington, D.C.
20005 (202)- 842-0034 L__
130 1.
very, very deep interest.
/-
(a) 2 In actual fact, immediately it came up it was very 3
obvious to me that through Iwan's work and your work the 4
concept of the ultimate dynamic moment or ultimate cyclic 5
moment or whatever you want to change it to really seemed to 6
me to actually have an experimental basis, which we had been 7
lacking.
8 I might say the point that you raised earlier, 9
Dana, is very true. We were lacking the sort of skills that 10 Bob Kennedy came in and has started to provide our group 11 because our group sort of came through it based on the work 12 they were doing, but without the probabilistic approach to 13 life that w~s really needed.
14 I agree with you.
One can argue that things can 15 slip through the cracks that way.
It's just the way -- it's 16 true.
17 The thing that I think is very encouraging about 10 Kennedy's approach is that it actually provides what I see 19 as an engineering tool.
It is something that an engineer 20 ca- ' grasp and it's something that it defines a particular 21 concrete aspect and an engineer can generally grasp that and 22 then start to apply that concept to the overall design.
23 Now I go back to that other overhead.
I think one 24
.of the troubles that oftentimes an experimental person will l
25 have is the concept of ceing able to see the whole system
)
ANN RILEY & ASSOCIATES, LTD.
Court Reporters 1250 I Street, N.W.,
Suite 300 Washington, D.C.
20005 (202) 842-0034 L
iL
131 1
approach as distinct from the individual components, uhereas r( )
2 I think the engineer, the engineering designer has a 3
difficulty actually extrapolating out the information that 4
is really important so that it can be used to -- once that 5
is poured out he can then take it and apply it across the 6
system.
7 So I think this provides a very interesting 8
engineering tool which I think people are starting to 9
recognize and in actual fact every time I have heard Bob 10 talk about it, it seems like his own ideas of it are
{
1 11 developing as well -- am I right? -- because your I
12 explanation and the --
13 DR. SHAO:
Let me make a comment.
If you remember 14 we wrote ours many years ago, the reason when you do a
()
15 thermal stress to allow 3 Sm, any general thermal stress 16 allowed has to be less than 3 Sm.
17 The one main reason the code wants to limit 18 thermal stress less than 3 Sm, you can do a linear thermal 19 analysis.
20 MR. BARNES:
Right.
21 DR. SHAO:
If you go beyond 3 Sm, your thermal 22 ana]ysis may not be valid.
That is the reason they remain 3 23 Sm.
Anything that goes beyond 3 Sm a linear analysis may I
24 not be valid.
That is the same reason the thermal stresses.
25 MR. BARNES:
That's right, but it is starting to I
l t
ANN RILEY & ASSOCIAT7S, LTD.
\\~-
Court Reporters i
1250 I Street, N.W., Suite 300 i
L
""""*?!!"e4?: 63
132 1
look at inelastic analysis now and --
(~')\\
2 DR. SHAO:
The code says you can do a general
(
3 linear thermal analysis so long as you are 3 Sm.
You go 4
beyond 3 Sm, you cannot do a linear analysis.
You can do a 5
linear of you go higher.
That is the basis of the code.
I 6
want to remind the group.
7 MR. BARNES:
Now the existing rules will require 8
some adjustment.
There's no doubt about it and in actual 9
fact I understand that the group took a vote on it and they 10 are quite prepared to go along with this B2 equal to 2.
Is 11 that right, Ken?
Didn't the group take a vote on the B2 12 equal to 2 and pass it, at least a verbal vote?
13 MR. MANOLY:
The approach was agreed upon but the i
14 value was not decided.
It was only one vote came.
15 MR. BARNES:
Has that one gone out for later f16 ballot?
17 MR. MANOLY:
That was prior to the last meeting.
18 MR. BARNES:
Okay, yes, that's right.
19 MR. MANOLY:
But only one member responded.
20 MR. BARNES:
Oh, is that right?
21 MR. MANOLY:
That is my understanding, yes.
22 MR. BARNES:
OKay.
It was held in Canada in 23 February and they had to cancel the meeting and then the 24 Chairman, Ernie Branch, had hss travel budget cut from under 25 his legs about the last week and he had to pay $1800 to go
[)
ANN RILEY & ASSOCIATES, LTD.
\\-
Court Reporters 1250 I Street, N.W.,
Suite 300 Washington, D.C.
20005 (202) 842-0034
133 1
to Anaheim so he didn't get along, so we have had two very
)
2 unfortunate meetings.
t%J 3
We have got to do a bit of regrouping.
4 Okay.
Now some basic conclusions.
This is from 5
our point of view, of course.
6 GE tests and the recent work in Japan confirm that
.7 the dominant failure mechanism is fatigue and/or fatigue 8
ratchet, the basis of the existing code rules.
I would just 9
like to comment on that.
10 I agree with Ken.
There are other potential 11 failure mechanism operative, no doubt about it, and nobody 12 has ever said that there aren't, but it seems to me we have 13 been locked into these positions where people say, well, j
14 it's fatigue, fatigue ratchet or it's not.
(D
' _,/
15 Now the people in the code work that I have spoken
\\
15 to and that I talk with about the problem, they say that the 17 test 37 for instance, that that is an unstable design and 18 they would expect and'quite possible would get a collapse 19 mechanism or an incremental collapse.
They don't deny that, 20 and in actual fact they thought by using the
.5 they were 21 keeping out of that range, if I am correct.
22 Now I understand there is disagreement on that.
23 Now in the code we are not against re-examining the rules.
24 That is exactly why the special working group was set up.
25 It wasn't set up just to -- it was because we had l
[
)
ANN RILEY & ASSOCIATES, LTD.
Court Reporters 1250 I Street, N.W.,
Suite 300 Washington, D.C.
20005 (202) 842-0034 1
=
134 1
comments from our Japanese colleagues.
We had comments from
()
2 some of our individual people.
We had comments from the
-3 working group on piping.
It was obvious to me that we 4
needed to set up another group after we'd got the rules 5
through, after it passed through this whole process to 6
investigate these comments.
7 We are not interested in enforcing things down 8
that -- and.not looking at -- otherwise the code and its 9
value decreases, so we are very open to that, and I don't 10 believe we are saying that -- but we are saying t.kat it is a 11 predominant failure mechanism.
The code before '94 didn't 12 address the predominant failure mechanism and in my mind I
13, that is a very serious situation.
j 14 Everyone that I talked to say that, well, the b
\\mst-15 code -- it's safe.
You mean the supports and so one, the 16 dead load -- and it produces a system that is safe, but it 17 seems to me that if you are not using the failure mechanism 18-that is. predominant, that you stand to get something 19 somewhere that nobody has ever thought of that is going to 20 lead to an accident.
That is what I worry about so I
'21 believe the code must address the predominant failure
]
22 mechanism.
I i
l 23 If it doesn't meet the criteria saying you must 24 use the collapse method and therefore there may be a need to i
25' put more restrictions in and make people go to a collapse I
ANN RILEY & ASSOCIATES, LTD.
r Court Reporters j.
1250 I Street, N.W.,
Suite 300 Washington, D.C.
20005 202) 842-0034
- ____--__=____--
135 1
method.
That might we what we have to do, and people are
()
2 quite prepared to do that.
l 3
The potential does exist for unstable designs for 1
4 incremental collapse and maybe there is the potential there 5
in stable designs but nobody at this moment has come up with 6
it and the Test 37 -- it seems to be the opinion of the l
7 study to come through -- is that pressure is the major 8
determining variable.
It may not be the only one, but it s
seems to be the major determining variable.
10 I mean Ken showed three tests that were failed and 11 had that collapse approach, but they are all unpressurized.
12 I have seen the other side of it where the pressure tests 13 were there and it shows no sign of that as far as I can i
14 tell.
()
15 Okay.
The existing rules that address the 16 predominant failure mechanism are technically sound.
It 17 seems to me that -- and why I say that is because the 18 Japanese seem to be starting to show that in their work.
1 19 Even sort of the' report from ETEC I don't think says it is 20 not technically sound, but it is saying there are other 1
21 problems that must be addressed.
22 Would that be a correct characterization, Ken?
23 MR. JAQUAY:
I agree for the majority of the 24 piping system and elbows and fittings it is not a bad set of 25 criteria.
It is --
b ANN RILEY & ASSOCIATES, LTD.
\\/
Court Reporters 1250 I Street, N.W.,
Suite 300 i
l Washington, D.C.
20005 i
(202) 842-0034 1
L__--____________-______
136 1
MR. BARNES:
It is incomplete.
()
2 MR. JAQUAY:
It doesn't cover all cases.
3
'MR. BARNES:
So we need -- and the other way I 4
read it.was -- I guess it's a repeat of this but need to 5
refine based on the Kennedy approach to demonstrate. margin, 6
and we need to get some numbers and we need to get agreement 7
in some of these specific constants.
8 A. cursory. review of the 22 short-term actions 9
listed in the ETEC summary, in that report, in that case 10 there are 14 significant points.
That's Ken's opinion --
11 and Ernie Branch has told me he believes that we can 12 accommodate those and if you would show the next overhead --
13 and I am not going to bore you with going through them j
14 all -- Ernie Branch says that the first 14 points are the
)
15
~ most significant, which is ETEC's.
16 The'next thing he says is adoption of the Kennedy-17
_ approach will accommodate the first six major points and in 18 his opinion it is 99 percent of the issue.
11 9 Now I don't know whether you would agree with
.20
'that, but that is Ernie's opinion.
214 Points 7 and 8 require further discussion.
That i
i 22 is the -- points 7 and 8 are the discussions on temperature, 23-the temperature effect that people have talked about.
It is 24 the opinion of Ernie -- he says that he believes that 25 temperature effects are ceing taken into account, and in i
O' ANN RILEY & ASSOCIATES, LTD.
Court Reporters 1250 I Street, N.W., Suite 300 Washington, D.C.
20005 (202) 842-0034
137 1
actual fact he sent a letter off te our special working
()
2 group, to the subgroup on fatigue strengths for comment, and
{
3 it does seem to be that the initial reply from O'Donnell as 4
the Chair was that it seems like these issues have been 5
taken care of, but there is obviously more work got to be 6
done on it, and nobody is denying that, and we have to 7
investigate it.
8 I mean Bob Kennedy is also saying the same sort of 9
thing, that it needs to be cleared up properly, and the l
10.
other thing, too.
I would agree with Bob and I certainly 4
11 would want to see that when I vote, that Test 37 is properly l
_12 understood.
I believe that it is an outlier and I don't 13 believe it characterizes the way it is, but I do believe it 14 must be explained and I think everybody in the code would
)
15 agree with that that I am aware of, the people I deal with.
16 Points 9 through 14 can be accommodated, but I 17 think you'll have some discussion on.whether it's even 18 necessary.
But there's no doubt about it -- Landers and 19' Branch have said it certainly could be accommodated within 20' the concept of the code.
l 21 But that's where the engineering comes in, because 22 of instead of saying, oh, who's going to worry about that, 23 it doetn't -- you know, it just is not an issue.
But it's 24 an issue from an experimental point of view, no doubt about l
25 that.
But from an engineering point of view, is it?
You E
[
ANN RILEY & ASSOCIATES, LTD.
f
. Court Reporters 1250 I Street,_N.W., Suite 300 Washington, D.C.
20005 (202) 842-0034
l l
138 1
know, and that's the balance.
A) 2 But it's the same -- I think we still can have l
'3 further discussion on these equations.
I think this 4.5 and 4
the concept that it provides, we have to work out something l
5 with section 11.
We've got to work out that this is an 6
empirical equation and they have to be able to do something i
7 else.
Otherwise, they are sort of left stuck with that as a j
8 stress.
It's not very clearly explained.
In fact, it's 9
probably not explained at all really when I think about it, i
I 10 and you as a regulator have a problem with it because you've 11 got to meet the requirements of 3 when you refer back to it.
12 So we do have a problem there that I think needs 13 to be addressed, an issue, but I do believe, first of all, 14 it's a misunderstanding, but then we have to address it.
(O) 15 I just want to reemphasize, as long as I am m,
16 subgroup chairman -- and that's another year or two, and
. 7-then they kick me out I'm open to all input. Every time I li get something -- and you have to testify to this, Mark.
19 Mark came to me one day and he said, I've sent 20 stuff up and nobody has answered-it.
I put it on the 21 agenda, his letter, and I made sure that he got it.
I put l
22 it back on the -- he said, they still never answered it, 23 they still haven't come back with a response or whatever it 24 is.
I put it back on the agenda again.
l l
25 I went and spoke to them about it then, and they O
ANN RILEY & ASSOCIATES, LTD.
Court Reporters 1250 I Street, N.W.,
Suite 300 Washington, D.C.
20005 (202) 842-0034
.l
l 139 l
l 1
said yes, they said they have answered it and they had been
> ()
2 answering it, they thought, through the committee.
3 Obviously, there was a disagreement there, but as far as I i
4 am concerned, all input -- we want all input.
We want to 5
have rules that have integrity.
We want rules that have got l
6 the basis for the integrity, and I don't believe the l
7 majority of the other members disagree with that at all, or i
8 they've got the same attitude is really what I'm trying to 9
say.
10 In actual fact, we have people tod-'r, which Bob 11 can tell you, who say that the rules are just too 12 conservative even with all these things in it, and in fact, 13 they won't vote in favor of the rules because they said, 14' unless you put these -- reduce the conservatism, they would
(_)
15 prefer to go back to the old rules.
16 Now, I don't understand the rationale, and 17 furthermore, I have real difficulty with that because 18 they're not based on the failure mechanism.
19
.Okay.
The next thing that happened is the special 20 working group is going to continue to operate until all 21 relevant input has been fully considered.
This includes the 22 test work undertaken by our Japanese colleague, and in 23 actual fact, I got a fax two nights ago from Dr. Endo, who 24 is a member of the special working group, officially 25 requesting the subgroup to continue to act as a point of eb ANN RILEY & ASSOCIATES, LTD.
'\\ /
Court Reporters 1250 I Street, N.W.,
Suite 300 Washington, D.C.
20005 (202) 842-0034
140 1
contact for the seismic work undertaken in the United States r(gx 2
and Japan.
3 Dr. Endo indicated that the issues raised also be 4
answered by the Japanese tests.
He thought quite a few of 5
those could be.
I don't know.
That's what you're hoping, 6
right, and that will come out.
7 The special working group is at a stage where some 8
of the major issue are almost ready for subgroup -- for the
-9 special working group ballot.
In fact, I've been talking to 10.
Ernie about it.
" i.
The only thing that will stop us from moving ahead 12 on some of these ballots is that this date of December '98, 13 which is the date promised for the start of the reporting on 14 your results, right, and if that's the case, it seems to me
-15 prudent at least to wait two more meetings before we go to 16 ballot just'to be sure that there's no surprises in the 17 issue.
18 I have a final thought which is my own 19 responsibility, and you can certainly kick me out now-20 because my time is spent and it's time tx) go.
I 21 But as a subgroup chair, I believe it's 22 fundamentally important that all stakeholder maintain
-23 active status in the special working group.
I really 24 believe'that for the following reasons:
25 The information will be discussed in detail.
'I ANN RILEY & ASSOCIATES, LTD.
- /
Court Reporters 1250 I Street, N.W.,
Suite 300 Washington, D.C.
20005 (202) 842-0034 t-
141 1
Impacted stakeholder need to participate for the balanced
[J) 2 development of proposals.
w 3
I mean, I can tell you, just the earlier comment, 4
Dana -- I hope you don't mind me calling you Dana --
5 DR. POWERS:
Please. Everybody else does.
6 MR. BARNES:
Is that right?
7 I can tell you that --
8 DR. POWERS:
Actually, other people call me worse 9
things than --
10
[ Laughter.]
11 MR. BARNES:
I have the same problem, especially 12 since I got involved with this seismic stuff.
13
[ Laughter.]
14 MR. BARNES:
But the point that I wanted to make
((_,/
15 was that the major representative that voted negative was a 16 U.S., you know, Regulatory Commission representative.
17 That's fine.
But the way I understood it, there ideas that 18 were considered and many changes came about because of their 19 input at that time.
Now, it wasn't sufficient to satisfy, 20 and that's fine as well, but nevertheless, there is that 21 openness to the ideas, and there is the openness.
Their 22 input or anybody's input is the only way you get a balanced 23 development of the proposals.
24 At the expert level, this is then most efficient 1
25 and effective way to place -- effective place to have input.
1 I T ANN RILEY & ASSOCIATES, LTD.
\\- /
Court Reporters 1250 I Street, N.W.,
Suite 300 Washington, D.C.
20005 (202) 842-0034 l
l l
142 1
There were just technically strong proposals.
See, for
()
2 instance, you don't have to participate and the proposal 3
will go ahead, You will have two or three levels still to 4
put input irt3, although the process is changing.
But you 5
will still have the chance to put that sort of input in.
6 The only trouble is this input is best put in at 7
that stop there where you've got your peers arguing with 8
you, discussing it and developing and refining, and I really 9
think that's a fundamental point that we should not 10 overlook.
This is money very well spent for all 11 stakeholder.
12 The special working group has been very fruitful, 13 in my opinion, very, very fruitful, and I think Larry 14 probably agrees with that from your perspective, precisely 15 because all the stakeholder have participated in one form 16 or another.
17 Finally, I want to express my appreciation to the 18 NRC for undertaking the work.
I really do believe that the 19 work of ETEC took the results and did the thing that Dana 20 was pointing out.
It elicited out a new set of information 21 which I do believe is going to lead to a conclusion, a
22 satisfactory -- a conclusion that represents the reality or 23 at least a conservative dimension of the reality, but it 24 really starts to represent the reality.
25 Finally, I would like to thank you gentlemen for O
ANN RILEY & ASSOCIATES, LTD.
Court Reporters 1250 I Street, N.W.,
Suite 300 Washington, D.C.
20005 (202) 842-0034
143 1-being pat 3ent with me.
Thanks.
()
2 CHAIRMAN SHACK:
I take it -- I mean, this whole 3
thing is in process now, then, and the special working group 4
is active?
Everybody is participating?
The NRC is 5'
supporting the special working group?
6 DR. SHAO:
No.
7 CHAIRMAN SHACK:
No?
8 DR. SHAO:
Well, not supporting them; working with 9
them.
10 CHAIRMAN SHACK:
Working with them.
11 DR. SHAO:
Yes.
12' MR. WESSMAN:
Let me clarify a little bit because 13 there were a couple of recent pieces of correspondence 14 relating to the special working group.,
O(_,/
15' As you know, we've had both Nilesh and Kamal 16 identified as the staff liaisons that have part'icipated with 17 the special working group the last couple of years.
We felt 18 partly from a resource standpoint and partly from an event 19 standpoint that as we reached closure with the ETEC work and
=20-looking at our own resources, we recognized that at various 21-times, we may have as many as three people funded by the 22 staff attending a particular special working group meeting, 23 whether it was Ken Jaguay, Nilesh, Kamal and on occasion 24-other individuals.
We felt we just could not continue that 25 intensity of resources to participate or act as liaisons
- A.
('~}--
ANN RILEY & ASSOCIATES, LTD.
Court Reporters 1250 I Street, N.W.,
Suite 300 Washington, D.C.
20005 (202) 842-0034 I
144 1
Lwith the special working group.
()
2
. Consequently, Brian Sharon advised ASME.that we 3
were going to bac': away from that direct liaison effort but
.4 would continue to stay involved via the regular consensus 5
process, and'that has continued thus far.
6 One.of the meetings went on in Anaheim.
Kamal 7
Minoli was out there and I believe -- is that right? -- John 8
_Putnam..One of the other members of the NRR staff'was out 9
there-and did attend a portion of the special working group 10 tueeting.
11 I fully would expect that type of activity to 12 continue, but we just can't continue this same level of 13 resources.
14 If-I can, let me take a minute and reflect on the
()
15 process because I think you have heard some things about i
16
--and some expressions in the course of the day.
Clearly i
17 that has been a very challenging issue and, on occasion, 18.
contentious issue that involves individuals on the staff, 19 individuals in industry,. individuals in various places.
20 But.you have heard terms-like both sides or push 21
-towards consensus.
Personally, I don't like terms like-22 sides and have to push ourselves.
I believe the consensus 23 process is working, and I think it's working very well and L
24 probably has' worked better from the standpoint of all of the c
I l
25 participants in the last couple.of years than maybe it did l.
[ )-
. ANN RILEY & ASSOCIATES, LTD.
Court Reporters 1250'I Street, N.W.,
Suite 300 i
Washington, D.C.
20005 (202) 842-0034
145 1
five years ago.
(/"%)
2 I think you have to recall that some of the things
\\_)
3
-- as we digressed earlier in Dick Barnes' presentation 1
4 talking about snowstorms and views and this sort of thing, l
l 5
this is five years ago or so.
When the NRC staff felt that 6
they could not support the '94 rules, they did write, but in j
l 7
general, there was a lack of supporting technical bases for 8
certain of the proposed changes, and that was part of the 9
letter that Beckjord and Russell signed out to Arlotto when 10 Arlotto was the chairman of the Board of Nuclear Codes.
11 I think that type of concern was conveyed largely 12 by various members of the staff at various working group, 13 subgroup, you know, this type of communication, but there j
14 was that concern.
,/~h
(,,)
15 That goes back to the time when the ETEC work was 16 being funded and Research probably has spent a million I
I 17 bucks.
I'm exaggerating, I'm not sure of the exact amount, 18 but clearly a lot of money and a lot of effort was invested 19 in getting us from several years ago to the document that we 20 have in front of us right now where, as Ken has expressed, 21 there are still some challenging technical issues yet to be 22 dealt with.
23 I think the consensus process of the special 24 working group and the interchange that has gone on in the 25 last couple of years has been very fruitful, and there is
[\\
ANN RILEY & ASSOCIATES, LTD.
\\--
Court Reporters 1250 I Street, N.W.,
Suite 300 Washington, D.C.
20005 j
(202) 842-0034 u_ __
L 146 1
still technical information emerging.
Whether its' work by
()
2 the Japanese test group or the work by Dr. Kennedy and, of 3
course, the ETEC, but clearly there are still challenges, I
-4 think, that all the participants and~all of those involved 5
in the process face, whether it's deciding on B-2 values, 6
whether it's dealing with things like the inspection, 7
settling on the right S sub M value, collapse versus 8
ratchet, perhaps some Japanese tests that may deal with 9
collapse -- I mean, there's a lot of challenging issues.
10 I think our dialogue that we've had today with all 11 of the involved participants has kind of brought us all to 12 date, brought the subcommittee to date as to where things 13 are and where things are going.
I think, as Dick Barnes 14 alluded, there is work to be done and Japanese testing that
(
15 may not be available until the end of the year.
It wouldn't 16 surprise me that continuing work by the special working 17 group could, you know, take another year or something as the
.18 process evolves, as the ballot work evolves, until final 19 recommendations, final changes, whatever it may be, really 20
-emerges.
21 I think the important thing is that the careful 22 and deliberative process that is done all in this very open 23 forum continues as it has and with full involvement of all 24 of the stakeholder, and I think it will.
I think that's 25 the important message that the subcommittee should carry i
[
\\
ANN RILEY & ASSOCIATES, LTD.
Court Reporters 1250 I Street, N.W.,
Suite 300 Washington, D.C.
20005 (202) 842-0034
147 1
away and try and convey back to the main committee, where
()
2 people are, that there are technical issues, and that things 3
are very much continuing.
4 CHAIRMAN SHACK:
As a practical question, what 5-would be the impact of this rule?
Would people go back and 6
remove piping' supports?
7 DR. SHAO:
You can't -- for existing plants.
For 8
new plants they could just have more flexible piping.
9 CHAIRMAN SHACK:
For a new plant of course you 10 could change the design, right.
I was curious as to what 11 impact it would have in the current fleet of plants.
12 DR. SHAO:
You can remove snubbers in the 13 supports.
'14 MR. BARNES:
That would be one practical -- I
()
15 mean -- yes, I made a mistake.
16 CHAIRMAN SHACK:
If you could find the snubbers 17 and that 18 DR. SHAO:
But the trouble is --
19 CHAIRMAN SHACK:
-- but the problem is that you 20 have to go back and do it.
21 DR. SHAO:
The trouble is they have some kind of 22 support.
They find in inspection some flaws and then the
-23 NRCusays you have to shut down.
They cannot repair with the 24 criteria they have right now.
'25 MR. BARNES:
Yes.
We have to address, we have to
/
ANN RILEY & ASSOCIATES, LTD.
t Court Reporters 1250 I Street, N.W:,
Suite 300 i
Washington, D.C. 20005 (202) 842-0034 j
l 1
l
148 1
clarify that, because I think that would be --
/m (v) 2 DR. SHAO:
But we have no choice.
3 MR. BARNES:
I know.
I understand.
I do 4
understand.
5 DR. SHAO:
Right now we have no chofce.
6 MR. BARNES:
Yes, I agree with you.
The way the 7
words are and given the nature that you have adopted the 8
code, the words, and so on -- I mean I understand.
9 DR. SHAO:
Because we endorse 11 already and 10, 10 the two doesn't gibe.
l 11 MR. BARNES:
No, I know.
I know -- you are right.
12 CRAIRMAN SHACK:
Any additional comments?
13
[No response.]
14 CHAIRMAN SHACK:
Again, I don't foresee any
(-(,,,)
15 immediate action from the committee.
This seems t' me 16 something that is in progress.
That's something we can 17 discuss as we do on report back to the full committee.
18 I thank everybody for their presentations.
They 19 were very interesting.
I thank very much especially our 20 Japanese colleagues who came a long way to make their 21 presentations and I apologize to everybody -- it's been a 22 long day.
23 MR. BARNES:
Good meeting.
l 24 MR. JAQUAY:
Thank you, it was a good meeting.
25 CHAIRMAN SHACK:
Thank you very much.
We are
[s'- f ANN RILEY & ASSOCIATES, LTD.
Court Reporters j
-1250 I Street, N.W.,
Suite 300 l
Washington, D.C.
20005 (202) 842-0034 I
b_._________.____
149 1
adjourned then.
7~
>Q
[Whereupon, at 6:15 p.m.,
the meeting was 2
3 concluded.]
l l
4 l
-5 6
l
.7 l
1 1
8 9
i 10 11 12 1
13 14 1
15 i
16 17 18 19 20 21 I
l 22 l
l-23 i
25-i 3
ANN RILEY & ASSOCIATES, LTD, Court Reporters 1250 I Street, N.W.,
Suite 300
' Washington, D.C.
20005 (202) 842-0034
REPORTER'S CERTIFICATE This is to certify that the attached proceedings 7%
(]
before the United States Nuclear Regulatory Commission in the matter of:
NAME OF PROCEEDING:
SUBCOMMITTEE ON MATERIALS AND METALLURGY (ACRS)
\\
CASE NUMBER:
PLACE OF PROCEEDING:
Rockville, MD were held as herein appears, and that this is the original transcript thereof for the file of the United States Nuclear Regulatory Commission taken by me and thereafter reduced to typewriting by me or under the direction of the court reporting company, and that the transcript is a true and l
accurate record of the foregoing proceedings.
j
(/f\\ \\
b O
Jon Hundley Official Reporter Ann Riley & Associates, Ltd.
l O
_--__--_.-------..-n_
_ -. _ _.