ML20213G552
| ML20213G552 | |
| Person / Time | |
|---|---|
| Site: | Braidwood  |
| Issue date: | 11/07/1986 |
| From: | Atomic Safety and Licensing Board Panel |
| To: | |
| References | |
| CON-#486-1597 OL, NUDOCS 8611180197 | |
| Download: ML20213G552 (98) | |
Text
!
l OR/GINAL O
UNITED STATES NUCLEAR REGULATORY COMMISSION IN THE MATTER OF:
DOCKET NO: 50-456 OL 50-457 OL COMMONWEALTH EDISON COMPANY (Braidwood Station, Units 1 and 2)
LOCATION:
CHICAGO, ILLINOIS PAGES:
16913 - 17008 DATE:
FRIDAY, NOVEMBER 7, 1986
/fA 0\\
i l c ACE-FEDERAL REPORTERS, INC.
l
~
OfficialReporters 444 North Capitol Street Washington, D.C. 20001 (202)347-3700 8611180197 e61107 1
PDR ADOCK 0500 6
ON". m o m G
16913 7-~5 U
1 UNITED STATES OF AMERICA 2
NUCLEAR REGULATORY COMMISSION 3
BEFORE THE ATOMIC SAFETY AND LICENSING BOARD 4
__________________x 5
In the Matter of:
6
(Braidwood Station, Units 1 8
and 2)
__________________x 9
10 Pages 16 913 - 17008 11 United States District Courthouse Courtroom 1743 s
12 219 South Dearborn Street
)'
Chicago, Illinois 60604 s_
13 Friday, November 7, 1986.
14 15 The hearing in the above-entitled matter reconvened 16 at 8:05 A. M.
17 BEFORE:
18 JUDGE HERBERT GROSSMAN, Chairman 19 Atomic Safety and Licensing Board U.
S. Nuclear Regulatory Commission 20 Washington, D.
C.
21 JUDGE RICHARD F.
COLE, Member, Atomic Safety and Licensing Board 22 U.
S. Nuclear Regulatory Commission Washington, D.
C.
23 JUDGE A.
DIXON CALLIHAN, Member, 24 Atomic Safety and Licensing Board l
N U.
S. Nuclear Regulatory Commission
)
25 Washington, D.
C.
i Sonntag Reporting Service, Ltd.
Geneva, 111inois cul34 (312) 232-0262
L i
16914 1
APPEARANCES:
2 On behalf of the Applicant:
3 MICHAEL I.
MILLER, ESQ.
PHILIP P. STEPTOE, III, ESO.
4 Isham, Lincoln & Beale Three First National Plaza 5
Chicago, Illinois 60602 i
6 On behalf of the Nuclear Regulatory Commission Staff:
7 GREGORY ALAN BERRY, ESQ.
8 ELAINE I.
CHAN, ESQ.
U.
S. Nuclear Regulatory Commission 9
7335 Old Georgetown Road Bethesda, Maryland 20014 10 On behalf of the Intervenors:
11 ROBERT GUILD, ESQ.
13 14 15 16 17 18 19 20 21 22 t-l 23 24 25 (s
Sonntag Reporting Service, Ltd.
[
Ueneva, Illinois eu134 l
(312) 232-0262 L.
I i
16915 i
I
)
1 EXHIBIT INDEX Marked Received 2
Applicant's Exhibit No.182 16918 3
Applicant's Exhibit No.183 16959 4
Applicant's Exhibit No. 1 84 16961 5
6 TESTIMONY OF LOUIS OWEN DEL GEORGE l
(Continued)
REDIRECT EXAMINATION s
8 BY MR. STEPTOE:
16916 9
RECROSS EXAMINATION BY MR. GUILD:
16964 10 11 12 i
13 14 15 16 I
17
(
18 19 1
20 21 22 23 24
\\
25 l
l Sonntag Reporting Service, Ltd.
Geneva, 1111nois eu144 (312) 232-0262
16916 (d~'\\
1 JUDGE GROSSMAN:
The hearing is reconvened.
2 This is the 88th day of hearing.
3 Do we have any preliminary matters before we resume 4
with Mr. DelGeorge?
None?
5 MR. GUILD:
Mr. Chairman --
6 JUDGE GROSSMAN:
Mr. Guild?
7 MR. GUILD:
-- I'd like to inquire on the 8
record as to the status of the Joe Hii Pearl Harbor File 9
documents.
10 MR. MILLER:
The documents are in a box in my 11 office.
They are taped up.
Mr. Hii has signed across 12 the tape so that there is positive evidence that no one 13 has disturbed the contents of the box as he taped it up.
14 JUDGE GROSSMAN:
Has he also written " Pearl 15 Harbor" across that tape?
16 (Laughter.)
l l
17 MR. MILLER:
No, sir; just his signature.
l 18 JUDGE GROSSMAN:
Fine.
We'll resume, then, 19 with Mr. DelGeorge.
20 Sir, you are still sworn and under oath.
21 THE WITNESS:
Yes, sir.
22 JUDGE GROSSFAN:
Okay, fine.
I 23 Mr. Steptoe?
I
(
24 REDIRECT EXAMINATION
(~
25 BY MR. STEPTOE:
(
Sonntag Reporting Service, Ltd.
ueneva, Illinois eu134 (312) 232-0262
F 16917 7~
b 1
0 Mr. De1 George, let's go first to Intervenors' Exhibi t 2
187, which is a three-page document entitled " Agreement 3
Rates as a Function of Craft Error Rate, QC Inspector 4
Accuracy and Overinspector Accuracy."
5 A
Mr. Steptoe, I think that may be Intervenors' 188.
6 0
188?
Excuse me.
7 First of all, Mr. DelGeorge, do you agree with 8
Intervenors' use of the term " accuracy" or " accuracy 9
rate" in this table and in these calculations?
10 A
No, sir, I don't.
11 0
Why not?
12 A
I believe the statistic calculated in the way it is 13 shown in this table does not fully account for the 14 effect of the overinspection accuracy and, as such, 15 overstates the agreement rate and thereby misrepresents 16 the accuracy rate for QC Inspectors.
17 0
okay.
18 Why does it overstate -- I'm sorry.
19 You said the agreement rate or the accuracy rate 20 that it overstates?
21 A
It will overstate the agreement rate contained within 22 Intervenors' table.
23 0
Why is that?
24 A
That is due to the fact that the calculation is based on b(
25 the assumption that the overinspector will only detect a j
Sonntag Reporting Service, Ltd.
Geneva, Illinois 6U134
)
(312) 232-0262 t_-
4 %w 16918 s
l 1
percentage of the remaining defectives in the sample i
4 2
presented to him for overinspection.
3 It is our experience, through review of numerous 4
overinspection programs, that overinspectors will not 5
caly detect a percentage of the defects remaining in the 6
sample presented to them; but they will also reject 7
acceptable items contained in that sample.
8 0
And what is the effect of the overinspector rejecting i
9 good work?
10 A
The effect of the overinspector rejecting good work is 11 to reduce the calculated agreement rate that would be 12 presented in this table.
13 MR. STEPTOE:
We just got one document, Judge 14 Grossman.
15 May I ask for two minutes so the witness can see 16 it?
17 JUDGE GROSSMAN:
Sure.
18 MR. STEPTOE:
Thank you, Judge Grossman.
19 I'm handing out a document which we'd like marked I
20 as Applicant's Exhibit 182.
21 Mr. DelGeorge, do you have it before you?
22 THE WITNESS:
Yes.
23 (The document was thereupon marked 24 Apt. cant's Exhibit No.182 for 25 identification as of November 7, 1986.)
l Sonntag Reporting Service, Ltd.
Geneva, 1111nois 6U134 (312) 232-0262
l 16919
()
1 BY MR. STEPTOE:
2 O
Who prepared this document?
3 A
The calculations were performed by myself, and George 4
Orlov prepared the document for me.
5 0
And what does this document illustrate?
6 A
The document illustrates the point I made just 7
previously in my testimony that when one considers the 8
potential of an overinspector rejecting good product in 9
the sample presented to him, that that has the effect of 10 decreasing the agreement rate between the overinspector 11 and the original first-line inspector.
12 To be specific, the sample calculation that is 13 presented in this exhibit is intended to be comparable 14 to the sample calculation offered by Intervenors' 15 attorney in yesterday's proceeding.
16 Specifically, the same assumptions were used with 17 respect to sample size, craft error rate of 20 percent, 18 QC Inspector error rate of 70 percent and overinspection 19 error detection rate of 90 percent.
20 The only change that's been made in the calculation 21 is to take account of the fact that the overinspector 22 will not only identify 5.4 of the 6 remaining defects in 23 the population, as was identified in Intervenors' l
24 calculation; but the overinspector has the potential for
(~^)
c (
25 identifying 8 rejects of the 80 acceptable items i
Sonntag Reporting Service, Ltd.
beneva, 1111nois ov134 (312) 232-0262
16920
/wU 1
presented to him for overinspection.
2 That leads to a total reject -- total number of 3
rejects by the overinspector of 13.4 rather than the 4
value calculated by Intervenors, 5.4.
5 When this is then converted into an agreement rate, 6
that agreement rate, adjusted to account for the 7
conservatism in the overinspection process, is 84.4 8
versus the 94.6 value or 93.7 value -- I think it's the 9
93.7 value -- calculated by Intervenors.
10 So there is a significant difference in the 11 agreement rate, the agreement rates being significantly
~ )
12 lower when the conservatism of overinspection is taken J
13 into account.
14 0
Mr. DelGeorge, what would be the significance of an 15 agreement rate reported of 84.4 percent in, for example, 16 the CSR program or in your analysis of the results of 17 either the PTL overinspection data cr the CSR data?
18 A
Well, agreement rates at that level have been such to 19 elicit further examination by Commonwealth Edison.
20 I think that's been demonstrated both in my 21 testimony as well as the testimony of Mr. Marcus, who 22 reviewed in more detail the PTL data base, where 90 23 percent was established as a threshold for increased 24 investigation.
(v) 25 MR. STEPTOE:
Judge Grossman, I move that Sonntag Reporting Service, Ltd.
ueneva, Illinois eulae (312) 232-0262
16921 rx (v) 1 Applicant's Exhibit 182 be admitted into evidence.
2 MR. GUILD:
Can we reserve on that, Mr.
3 Chai rman?
4 JUDGE GROSSMAN:
Yes.
I have a question, 5
too.
6 Mr. DelGeorge, it seems to me that if you focus not 7
only on the items that were rejected by the QC Inspector 8
but also on those that were accepted by him, then you 9
have to take into account in your calculations the 10 inaccuracies in his acceptance of those items, which 11 you're not doing.
12 THE WITNESS:
No, sir.
If we --
13 JUDGE GROSSMAN:
Only if you -- let me 14 finish.
15 Only if you isolate the rejectable ones have you 16 taken into account all the factors that apply.
17 Is that not correct?
18 THE WITNESS:
Well, even if one were to apply 19 the error detection rate for the overinspector as if it 20 were an accuracy rate for the overinspector and, 21 thereby, apply the 90 percent rate to the population of 22 86 items, as opposed to applying the 90 percent rate to 23 the 80 items and the 90 percent rate to the 6 defective i
24 items, you still wind up with a much lower agreement
(~*g 25 rate.
Sonntag Reporting Service, Ltd.
Geneva, Illinois bu144 (312) 232-0262
4 16922 (G
1 For example, if we apply the overinspector accuracy 2
rate of 90 percent to the total population presented to 3
him of 86 items, he would identify, assuming that he 4
makes errors on a random basis, 8.6 defectives.
5 That 8.6 over the 86 items constitutes a 6
disagreement rate of 10 percent, the inverse being an 7
agreement rate of 90 percent, which itself is lower than 8
the calculated agreement rate calculated in the way 9
Intervenors performed the calculation.
10 JUDGE GROSSMAN:
Well, perhaps I'm missing 11 something, but what I believe is left out of your I~'
12 equation are not the inaccuracies of the overinspector N~-)h 13 with regard to the other 86 items, but the inaccuracies 14 of the original QC Inspector with regard to that.
15 It's only because, in Mr. Guild's calculation, that 16 you take into account only the rejectable items that l
17 you've fully covered the inaccuracies of the QC l
18 Inspector.
I i
19 But if you're considering the entire 100 for your 20 full calculation, you haven't taken his full 21 inaccuracies into account.
22 Am I mistaken in that?
i 23 THE WITNESS:
I don't believe you're mistaken 1
24 in recognizing that distinction.
p j ()
25 But I think Intervenor demonstrates the sensitivity i
Sonntag Reporting Service, Ltd.
Ueneva, 1111nois eu134 (312) 232-0262
16923 p
1 of that difference when he provides to the overinspector 2
the population with def ects removed as well as the 3
population with defects returned after remedial action 4
is taken.
5 There is a very slight difference in the ratios 6
that are ultimately calculated.
I don't expect there to 7
be a very significant difference if one were to take 8
that into account.
9 JUDGE GROSSMAN:
Okay.
Well, now that I've 10 clouded the issue, we'll let Mr. Guild voir dire.
11 (Laughter.)
(m\\
12 Do you wish to now?
\\m /
13 MR. GUILD:
No, sir.
14 If I may, I'd just like to reserve on the document 15 and ponder a little bit and check some math before I say 16 one way or the other.
17 JUDGE GROSSMAN:
Okay, fine.
18 Do you have an objection to that document?
i 19 MR. BERRY:
I'd make the same request as Mr.
20 Guild.
I'd like to study the calculations a little 21 further myself.
22 On the face of it, I don't think I'll have any 23 objection; but I'd just like to be sure in my own mind l
24 how the equation works out.
O 25 JUDGE GROSSMAN:
Well, we'll reserve judgment
(}
i Sonntag Reporting Service, Ltd.
Ueneva, Illinois 6u134
16924
,O,
\\~Y 1
on admissibility of that document now.
2 Chances are we'll admit it as part of the 3
testimony, unless there is shown to be something that is 4
drastically wrong with it and we have some testimony 5
with regard to that.
6 Fine.
Proceed now, and you can use it as though --
7 if you have any further testimony on it, you can use it 8
as though adiaitted.
We'll just make our final 9
determination later on.
10 MR. STEPTOE:
Thank you, Judge Grossman.
11 I do have a few more questions that I perhaps
/~
12 should have asked before in respect to this table.
13 BY MR. STEPTOE:
14 O
Mr. DelGeorge, is it possible that the original QC 15 Inspector could identify as rejectable work that which, 16 in fact, is good work?
17 A
Yes, sir.
l 18 0
Is that fact taken into account in these calculations?
19 A
Not explicitly.
But there is implicit the assumption 20 that the product, the sample, offered.:o the 21 overinspector is, in fact, comprised of all cceptable 22 items.
l 23 So whatever the population, that is a population l
l 24 offered to the overinspector of all acceptable items.
m i
j 25 0
Now, going back to --
Sonntag Reporting Service, Ltd.
ueneva, Illinois du134 (312) 232-0262
g.
,.r.
7
.j 4,
]d 1692(
I s
~
\\
(G'
=-
y y
1 JUDGE GROSSMAN:
LNt?N&'just say:
One thing 2
I'm concerned about is if you don?t fully take into
~ '
3 account the QC Inspector's wrong alls on what he 4
considered' acceptable, you're not taking into account 5
items that would further reduce his accuracy rate,
)
6 possibly in the same proportion that the overinspection 7
accuracy rate has been reduced, and thereby come up with 8
the same figures that Mr. Guild came up with.on his 9
sheet, whatever number that is. '
4 10 Of course, I'm not offering testimony here.
I just' 11 want to make sure that we resolve all the problems
~ )
12 before we close the record.
J 13 MR. STEPTOE:
Thank you, Judge Grossman.
14 BY MR. STEPTOE:
15 0
Mr. DelGeorge, will you address Judge Grossman's 16 observations?
17 A
A calculation I believe can be made to address the 18 observation made by Judge Grossman.
19 It's my belief that, when made, that calculacion 20 will demonstrate the same trend in agreement rates; that 21 is, when overinspector rejection of acceptable product q
l 22 is taken into account, the agreement rate between the 23 overinspector and the reinspector will be deflated even i
j 24 when we take into account the potential for an original 25 inspector rejecting acceptable product.
l Sonntag Reporting Service, Ltd.
Ueneva, Illinois culde (312) 232-0262 t
i 16926 hv 1
I don't have in mind the result of that 2
calculation, but I believe it will demonstrate the same 3
trend.
4 Q
Rettrning to Intervenors' Exhibit 188, Mr. DelGeorge, 5
how many variables are identified which could affect the 6
agreement rates?
7 A
There are three variables identified:
craft error rate, 8
QC Inspector accuracy rate and overinspection accuracy 9
rate.
l 10 0
What variation, if any, would you expect in the 11 overinspector accuracy rate for the CSR data which you 12 used in your testimony measured as a function of the 13 time period -- on the time scale that you used in your 14 testimony?
15 A
I would expect there to be no relation, and let me 16 explain why.
17 The CSR reinspections were conducted over a very 18 short period of time af ter the work in question had been 19 performed.
The over -- the reinspector had no knowledge 20 of the time period at which the original inspection was 21 performed.
22 Therefore, because the CSR reinspection was 4
23 performed over a short period and where the CSR 24 inspection had no direct relationship, in terms of its l
s
)
25 timing, with the timing of the original inspection being l
Sonntag Reporting Service, Ltd.
Ueneva, 1111nois bu144 (312) 232-0262
16927 OV 1
reinspected, it's my belief that there should be no 2
relationship between the two accuracy rates.
3 Beyond that, because of the oversight that was 4
applied to the CSR reinspection program, it's my belief 5
that there was, in fact, very little variability in the 6
accuracy rate for the CSR reinspection.
7 0
In your judgment, is it appropriate to treat the CSR 8
accuracy rate as a constant?
9 A
Yes, sir.
10 0
Okay.
11 That means that two variables remain on this table:
12 the QC Inspector accuracy rate; that is, the Comstock 13 inspector --
14 MR. GUILD:
Excuse me.
Let me interrupt.
15 The last question was unclear.
16 Do we treat overinspector accuracy rate as a 17 constant over time, since "over time" w'as the assumption 18 in the preceding question?
19 Mr. Chairman, I ask that that question be 20 clarified.
It's vague as asked.
21 JUDGE GROSSMAN:
You may ask that question 22 again as voir dire, certainly.
23 MR. STEPTOE:
Well, I don't think it's voir 24 dire, but I'll ask the question.
25 JUDGE GROSSMAN:
Okay, fine.
Sonntag Reporting Service, Ltd.
Ueneva, Illinois bu134 (312) 232-0262
16928 OG 1
BY MR. STEPTOE:
2 Q
Mr. DelGeorge, could you respond to Mr. Guild's comment 3
or request for clarification?
4 A
Yes.
5 I believe it is justifiable to make the assumption 6
that the CSR repeat inspection accuracy rate was a 7
constant as a function of time.
8 0
And the time we' re referring to -- what is the time 9
scale that we're referring to?
10 A
We're talking about the period of inspection -- original 11 inspection reviewed between 1978 and June of 1984, as 12 represented by the CSR repeat inspection interval of the 13 fall of 1985 -- excuse me -- the fall of 1984 through 1
14 the spring of 1985.
15 0
I think that last bit confused me a little bit.
16 Is it -- when you say that it's appropriate to 17 treat the CSR inspector accuracy rate as a col.stant as a 18 function of time, are you referring to the time scale 19 which appears in Attachment 2C DelGeorge-1 in your 20 testimony?
21 (Indicating.)
22 A
Yes, sir.
23 0
And that time scale represents the date of the Comstock 24 QC inspections?
(
25 A
Yes, sir.
Sonntag Reporting Service, Ltd.
Geneva, Illinois 6U144 (312) 232-0262
16929
?N 1
0 All right.
2 Well, then, if the CSR inspector accuracy is 3
assumed to be a constant, we have two variables 4
remaining -- Comstock QC Inspector accuracy rate and 5
craft error rate -- according to this table, 6
Intervenors' Exhibit 188.
7 Which one of those variables or both would you 8
expect to be affected by Comstock QC management 9
harassment and intimidation and production pressure and 10 so forth?
11 A
It's my belief that, given the assumption of production
(
12 pressure by QC management, that the QC Inspector 13 accuracy rate would be affected.
14 0
Would you expect the craft error rate to be affected?
15 A
I know -- I know of no reason why the management 16 performance of QC management should affect production 17 rate -- excuse me -- craft error rate.
18 0
okay.
19 Now, let's turn to Attachments 2C DelGeorge-1 and 20 DelGeorge-2, Revision 2, which are these bar charts 21 which are actached to your testimony.
22 (Indicating.)
23 Mr. DelGeorge, what variation, if any, do you see 24 in the agreement rates reported in these two attachments 25 as a function of time?
\\~ /
Sonntag Reporting Service, Ltd.
Geneva, Illinois 60134 (312) 232-0262
16930 O
1 A
There is very little variation as a function of time.
2 0
So the --
3 MR. GUILD:
Excuse me.
Which exhibits?
4 MR. STEPTOE:
The first and the second 5
charts.
6 The first one is entitled "CSR Inspection Results 7
for LKC, All Populations, All Samples, Welding and 8
Non-Welding."
9 The second one, Attachment 2C DelGeorge-2, Revision 10 2, i s entitled "CSR Inspection Results for LKC, All 11 Populations, All Samples, Welding Only."
12 MR. GUILD:
They are Attachments 1 and 2?
13 MR. STEPTOE:
Yes.
14 JUDGE CALLIH AN:
Mr. Steptoe, these are 15 revisions as of September the 30th,1986?
16 MR. STEPTOE:
That's correct -- October the 17 30th.
18 JUDGE CALLIHAN:
Thank you.
19 BY MR. STEPTOE:
20 0
Now, addressing yourself to the remaining variables set 21 forth in Intervenors' Exhibit 188, what explanation or 22 explanations, if any, are there for the essentially 23 straight-line agreement rates shown on these two 24 attachments?
(
25 A
I think there are two possible explanations for that Sonntag Reporting Service, Ltd.
Geneva, Illinois eU144 (312) 232-0262
16931 A
ID) 1 observation, the first being that both variables remain 2
constant as a function of time or for all practical 3
purposes constant as a function of time, or that there 4
existed an inverse relationship between the variables 5
such that as one increased, the other decreased in an 6
equivalent manner at the same point in time.
7 MR. GUILD:
Mr. Chairman, I'm not clear what 8
variables the witness is addressing.
9 BY MR. STEPTOE:
10 0
Please explain that.
11 A
The two variables that I make reference to are craft f'N 12 error rate and QC Inspector accuracy rate.
13 Q
You regard it as possible that there is an inverse 14 relationship between the Comstock QC Inspector accuracy 15 rate and the craf t error rates?
16 A
I know of no reason why there would be such a 17 relationship as a function of time.
18 0
Well, let's be specific, Mr. DelGeorge.
19 When did Mr. Saklak show up at Braidwood?
20 A
I believe Mr. Saklak took a management position in the 21 QC Department of Comstock at Braidwood in July of 1982.
22 0
Referring to Attachment 2C DelGeorge-1, do you see any 23 variation in the inspection point agreement rate between 24 the period immediately before Mr. Saklak came to 25 Braidwood and the period immediately after he came to l
Sonntag Reporting Service, Ltd.
Ueneva, 111in61s 6ulde (312) 232-0262
16932 O.
(J l
Braidwood?
2 A
No, sir, I see no significant difference.
3 Q
And what are the possible explanations for that?
4 A
In that the composite perf ormance bef ore and af ter that 5
date were identical or that there were -- that there was 6
a relationship that existed that would have had QC 7
performance increasing or decreasing at a rate inversely 8
proportional to the decrease or increase in craft error 9
rate.
10 0
Do you know whether Mr. Saklak had any control or 11 influence over the craft error rate at Comstock?
12 A
Mr. Saklak came f rom the Production Department at 13 Comstock prior to July of '82.
Subsequent to July of 14
'82, I know of no reason why he would have had any 15 effect on the production error rate.
16 0
Well, let me ask you to assume that Mr. Saklak was 17 harassing and intimidating people in the QC inspection 18 organization in such a fashion that their performance 19 suffered.
20 What would have happened to the craft error rate --
21 strike that.
22 Would you expect to see a change in the agreement 23 rate shown in this graph if that were, in fact, the 24 case?
p) 25 A
Well, it's my expectation that if, in fact, Mr. Saklak
(
Sonntag Reporting Service, Ltd.
Geneva, 1111nois culde (312) 232-0262
i
^
16933 (O
,J' 1
at his arrival pressured or harassed OC -- members of 2
the QC organization in a way that would affect the 3
quality of their work, that one would see a decrease in 4
the agreement rates after his arrival.
5 I know of no reason why his performance within the 6
QC Department would have improved craf t performance or, a
7 in effect, decreased craft error rate after his arrival 8
in the QC Department.
9 For that reason, I would expect to see, after Mr.
10 Saklak's arrival, a decrease in the agreement rates as 11 they are reported in this attachment.
12 0
When did Mr. DeWald show up at Braidwood?
13 A
Mr. DeWald returned to Braidwood in a management 14 position in the Comstock QC organization in August of 15 1983.
16 0
Do you see any variation in the inspection point
(
17 agreement rates after Mr. DeWald's arrival as a manager 18 at the Comstock QC organization?
19 A
None of significance.
20 0
Again, what are the possible explanations for the lack 21 of any significant variation in this graph?
22 A
The possibilities I believe are the same as I have 23 previously described; tnat is, the variables either 24 remain constant in time -- that is, both craft error
- ()
25 rate and QC Inspector accuracy rate -- or the craf t Sonntag Reporting Service, Ltd.
Geneva, Illinois 60134 (312) 232-0262
16934 OV 1
error rate changes inversely with the QC Inspector j
2 accuracy rate in a manner that would mask the change in 3
QC Inspector accuracy rate.
4 Q
Do you know of any -- do you know of any reason why Mr.
5 DeWald's assumption of QC management responsibilities in 6
the Comstock organization would have affected the craft 7
error rate?
8 A
No, sir.
9 0
If I asked you the same questions with respect to 10 C De1 George-2, would your answers be the 11 same?
. I' )
12 A
Yes, sir.
D 13 0
Let's move on to Attachment 2C De1 George-3, which is the 14 third attachment.
It's entitled "All Populations, All 15 Samples, Welding Only."
16 This time the data is presented on a weld agreement 17 rate rather than on an inspection point agreement rate; 18 is that correct?
19 A
Yes, sir.
l 20 0
Okay.
21 What variation, if any, is there in the weld 22 agreement rate shown in this attachment, say, for 1983, 23 for the year 1983?
I 24 A
There are some changes in the agreement rates which I
(
25 have not found to manifest any trend.
N-Sonntag Reporting Service, Ltd.
ueneva, 1111nois bu134 (312) 232-0262
1 l
16935 (qJ v
1 Q
Okay.
2 And that's based on your prepared testimony --
3 A
Yes, sir.
4 0
-- is that correct?
5 Again, could you just give me the values of the 6
weld agreement rate -- the approximate values of the 7
weld agreement rate for the four quarters of 1983, which 8
are shown here?
9 A
Those values would be, for the first quarter of 1983, 10 92.36; for the second quarter, 74.91; for the third 11 quarter, 94.10; and for the fourth quarter, 89.81 12 percent, respectively.
13 Q
All right.
14 Now, let's refer back to Intervenors' Exhibit 188.
15 Let's assume that in January of 1983, the first quarter 16 of 1983, before Mr. DeWald shows up, the Comstock l
17 inspector accuracy rate is the highest value shown in 18 Intervenors' table, which is 90 percent.
i 19 All right?
l 20 A
Yes, sir.
j l
21 0
And then let's assume that Mr. DeWald shows up and has I
22 an adverse effect on the QC Inspector accuracy rate.
23 Do you have a figure in mind, based on your review l
l 24 of the record, as to what would be a reasonable figure 25 to assume?
f Sonntag Reporting Service, Ltd.
l Geneva, 111inois bu134 (312) 232-0262
16936
(~h L]
1 A
Based on the testimony provided by Intervenors' expert, 2
Dr. Arvey, I think it's fair to assume a change in 3
performance of about 30 percent.
4 0
And is that -- that assumption is based on your 5
understanding of Intervenors' position in this case; 6
correct?
7 A
Yes, sir.
8 MR. GUILD:
I hope we're not going to quibble 9
at this point whether that's an accurate assumption or 10 not.
11 If the record is to stand as that question N_//}
12 representing the statement of Intervenors' position, it 13 is inaccurate.
It is a mischaracterization of Dr.
14 Arvey's testimony.
15 But so long as it's Mr. DelGeorge's mistaken 16 understanding, I have no objection tc that premise.
17 JUDGE GROSSMAN:
He's not making his own 18 estimate, now.
He's just using that as an assumption.
19 Is that correct, Mr. Steptoe?
20 MR. STEPTOE:
That's right, that's right.
21 BY MR. STEPTOE:
22 0
Is it your understanding, Mr. DelGeorge, that 23 Intervenors claim that in this case harassment, 24 intimidation and production pressure has been powerful
(
25 and pervasive and historic at Braidwood?
)
Sonntag Reporting Service, Ltd.
Geneva, Illinois 60134 (312) 232-0262
l f
l l
16937 g
e t
L) 1 A
Yes, sir.
2 0
And is it your understanding that the Intervenors claim 3
that that production pressure has had significant 4
effects en the QC Inspector accuracy rate?
5 A
Yes, sir.
6 0
Okay.
So let's just take an assumption of a 30 percent 7
decrease.
8 So let's assume that when -- let's assume that when 9
Mr. DeWald shows up in the third quarter of 1983, the 10 accuracy rate goes down to 60 percent.
It's just an 11 assumption; from 90 percent down to 60 percent.
(~ }
12 Now, can you tell me what variations in the craft v
13 error rate, if any, would be necessary in order to come 14 up with these values of 92 percent for the first quarter 15 of 1983; 74 percent -- 75 percent for the second quarter 16 of 1983; 94 percent for the third quarter of 1983, when 17 Mr. DeWald shows up; and approximately 90 percent for 18 the fourth quarter of 19837 19 A
There would have to be significant swings in the craft 20 error rate over those time periods in order to produce 21 these agreement rates.
22 0
Well, let's just take the second quarter of 1983.
23 Again, my assumption is that, since Mr. DeWald isn't 24 there, the Comstock QC Inspectors are performing at
(
25 Intervenors' assumed QC Inspector accuracy rate of 90 v
Sonntag Reporting Service, Ltd.
l ueneva, Illinois culae (312) 232-0262
16938 (D
I
/
1 percent.
2 Where on Intervenors' Exhibit 188 do you find the 3
corresponding craft error rate which will get you to the 4
total agreement rate of 75 percent?
5 A
Intervenors' Exhibit 188 only goes down to a craf t error 6
rate of 30 percent, and it's my belief that the craft 7
error rate necessary to account for a 74 percent 8
agreement rate in that time step would be on the order 9
of 80 percent.
It would be far off the Intervenors' 10 table.
11 0
All right.
f'}
12 Then the next quarter, when Mr. DeWald shows up, we LJ 13 have an agreement rate of 94 percent.
14 MR. GUILD:
Excuse me.
15 Just to be clear, that last answer assumes 60 16 percent inspector accuracy rate?
17 THE WITNESS:
Yes, sir.
18 MR. STEPTOE:
Excuse me.
No.
It assumes --
19 I asked 90 percent inspector accuracy rate for the 20 second quarter.
21 MR. GUILD:
It assumes 90 percent?
22 THE WITNESS:
Then I am confused.
23 MR. STEPTOE:
Okay.
24 BY MR. STEPTOE:
25 0
The second quarter of 1983 --
Sonntag Reporting Service, _ Ltd.
ueneva, 1111nois ou144 (312) 232-0262
16939
['N
\\v) 1 A
Yes, sir.
2 Q
Mr. DeWald is not there yet; is that correct?
3 A
That's correct.
4 0
All right.
5 I asked you to assume that the QC Inspector 6
accuracy rate is 90 percent.
7 A
Yes, sir.
8 0
What craft error rate would be necessary in order to 9
come up with an agreement rate of 75 percent, which is 10 what we've agreed is shown in Attachment 2C De1 George-3?
11 A
If the craf t -- if the QC Inspector accuracy rate is 90
('~'
12 percent and the identified agreement rate is 74 percent,
\\_
13 the craft error rate would be off the scale of the graph 14 and approaching zero.
15 0
okay.
16 Now, let's assume the next quarter Mr. DeWald shows 17 up.
We're going to assume that there's --
18 A
Excuse me.
The craft error rate would be approaching 19 100 percent and would be off the table.
20 MR. GUILD:
I'm sorry.
It was 80 percent or 21 100 percent?
22 TH E WITN ESS :
The previous line of questions 23 I had misinterpreted to be with respect to a 60 percent 24 accuracy rate.
25 MR. GUILD:
Mr. Chairman, may I inquire:
Is Ns._
Sonntag Reporting Service, Ltd.
Ueneva, Illinois eu144 (312) 232-0262
16940 0) tv 1
the witness reading from a document?
2 THE WITNESS:
No.
I'm just looking back and 3
forth between Intervenors' Exhibit 188 and my figure, 4.
5 MR. GUILD:
Did the witness just calculate 6
the 100 percent craft error rate, then?
7 THE WITNESS:
One can, by observation, for a 8
given accuracy rate, by observation of the change -- the 9
rate of change of the calculated agreement rate for 10 various craft error rates, identify how that rate of 11 change will progress as craft error rates increase.
(Vh 12 It's on the basis of a calculation that I'm doing 13 as I'm sitting here that I'm making these estimatos.
14 MR. GUILD:
Fine.
15 MR. STEPTOE:
All right.
16 BY MR. STEPTOE:
17 0
Well, Mr. DelGeorge, can we agree that if you assume a 18 QC Inspector accuracy rate of 90 percent, you can't get 19 an agreement rate of 75 percent using Intervenors' 20 table?
21 A
That's correct.
22 MR. GUILD:
What's the assumed overinspector 23 accuracy rate in this calculation?
24 MR. STEPTOE:
I'm using Table No. 1, so the 25 assumption is the 90 percent accuracy rate as you use Sonntag Reporting Service, Ltd.
ueneva, Illinois eu144 (312) 232-0262
16941 OC 1
that term.
2 MR. GUILD:
Thank you, 90 percent.
3 BY MR. STEPTOE:
4 0
Well, if you assume that the overinspection accuracy 5
rate is 80 percent, Table No. 2, Mr. DelGeorge, do you 6
come to a different answer?
7 A
No, sir.
8 I think by observation one can see that the 9
calculated agreement rate is relatively insignificant --
10 that is, insensitive to the overinspection accuracy rate 11 for any particular calculated agreement rate -- given 12 the manner in which Intervenors calculate agreement 13 rates.
14 Q
Now, let's go back to the next quarter in 1983.
Mr.
15 DeWald shows up, and the agreement rate, as reported in 16 C DelGeorge-3, goes from 75 percent to 94 17 percent.
18 Is that correct?
19 A
Yes, sir.
l 20 Q
All right.
21 Now, let me ask you to assume that the QC Inspector 22 accuracy rate drops to 60 percent because Mr. DeWald is 23 there and he's harassing and intimidating people.
24 All right?
25 A
Yes.
Sonntag Reporting Service, Ltd.
GerAva, 1111Dois 6U144 (312) 232-0262
16942
()
\\ _/
s 1
0 Are you with me?
2 What do you have to assume, using Intervenors' 3
table, about the craft error rate in order to come up 4
with an agreement rate of 94 percent?
5 A
The craft error rate would have to be between 10 and 20 6
percent.
7 0
And that's shown on the table; right?
8 A
Yes, sir.
9 0
And that's shown on the first table, which assumes 90 10 percent accuracy in overinspection; right?
11 A
Yes, sir.
f~
12 0
Now, if we go to the second table, which assumes 80 13 percent accuracy in overinspection, does your answer 14 change?
15 A
It would still be between 10 and 20 percent, and closer 16 to 20 percent.
17 0
All right.
18 So in the second quarter, given the assumptions 19 that we've made, in the second quarter of 1983, we have 20 a craft error rate approaching 100 percent; in the third 21 quarter of 1983, we have a craft error rate between 10 22 and 20 percent.
23 Is that correct?
24 A
Yes, sit.
25 O
Does that seem plausible to you?
l t
Sonntag Reporting Service, Ltd.
Ueneva, Illinois bu144 l
(312) 232-0262 t
16943 (D
\\j 1
A No, sir.
2 0
Why not?
3 A
Based on our continuous observation of craf t performance 4
at Braidwood and our observation of similar performance 5
at all of our sites, we have not identified swings in 6
performance of this type over these short periods of 7
time.
8 0
Now, Mr. DelGeorge, could similar analyses of different 9
time periods other than 1983 be done, using the 10 information supplied in your testimony?
11 A
Yes, sir.
12 0
Well, Mr. DelGeorge, there are variations in the data v
13 shown in Attachment 2C De1 George-3 and also in the PTL 14 data base as well from time period to time period.
15 What do you believe accounts for those variations?
16 A
My review of those variations has indicated that they 17 are generally associated with performance changes of one 18 or a very few individuals who --
19 0
Which individuals, sir?
Of what organization?
20 A
Of the Comstock QC Inspector work force.
21 0
Okay.
I'm sorry.
I shouldn't have cut you off.
22 JUDGE GROSSMAN:
Had you completed your 23 answer?
24 THE WITNESS:
I believe I responded to his
(
)
25 question.
Sonntag Reporting Service, Ltd.
ueneva, 1111nois culae f
(312) 232-0262
I 16944 O
I (J
l JUDGE GROSSMAN:
Okay, fine.
2 MR. STEPTOE:
I need to think of the next 3
question.
4 BY MR. STEPTOE:
5 0
Have you analyzed variations, if any, in the performance 6
of these individuals?
7 A
As I indicate in my testimony, I have attempted such an 8
investigation with greater success with respect to the 9
PTL data base than with respect to the CSR data bas'e.
10 The reason for that is that there is a 11 significantly larger amount of data on a per-inspector
(~'
12 basis in the PTL data base than is contained in the CSR
~'
13 data base.
14 One can make a similar evaluation; but the level of 15 confidence, given the quantity of data, is lower for the 16 CSR data base.
^
17 0
And have those dips been analyzed in testimony that's i
18 been presented to this Board?
19 A
All of the changes in agreement rate for the PTL data 20 base which are of any significance have been discussed 21 in detail in Mr. Marcus' testimony.
22 0
Have you identified any relationship between t
23 Intervenors' claims of harassment and intimidation and 24 these variations in the performance of individual 25 inspectors?
Sonntag Reporting Service, Ltd.
Ueneva, Illinois bu134 (312) 232-0262
16945 f\\
'wl 1
A I have been unable to identify such a relationship.
2 0
Okay.
3 Mr. DelGeorge, what use, if any, did you make of 4
the overall agreement rates which are presented at Page 5
38 of your testimony; that is, the agreement rates of 93 6
percent for PTL and 85 percent for the CSR welding data?
7 A
As I indicated yesterday in my testimony, the use to 8
which I have put these statistics was to provide a basis 9
for comparative assessment of the PTL data base to the 10 CSR data base.
11 I believe I described that in some detail
{v}
12 yesterday.
13 0
What use, if any, did you make of these overall 14 percentages in coming to your conclusions about the 15 effectiveness of Comstock OC Inspector performance?
16 A
I did take account of the agreement rates, but it was 17 not a strong element of the conclusion that I reached.
18 I'll explain that --
19 0
Yes, please do.
20 A
-- if I can, 21 As is demonstrated both by my testimony and by Mr.
22 Marcus' testimony, we have used agreement rates as a 23 screening basis for review of inspector performance.
24 Dased on our experience, based on my experience, 25 the agreement rates that we've observed both in the CSR Sonntag Reporting Service, Ltd.
ueneva, 11:1nois eu1Je (312) 232-0262
16946 OV 1
and the PTL data base are consistent with the 2
performance that we have observed over many years, 3
performance which we have demonstrated to be acceptable.
4 The stronger use that has been made in my testimony 5
of agreement rates has been with respect to their 6
variation in time to determine to what extent any 7
variation exists and, if it exists, whether or not it 8
can be correlated to the alleged histocic, tangible, 9
powerful and pervasive production pressure suggested by 10 the Intervenor.
11 0
What use, if any, did you make of these overall
(~ h 12 agreement rates in coming to conclusions about the L
13 quality of the work?
14 A
I believe I did mention yesterday -- and it is my 15 position -- that these agreement rates are a fairly good 16 indicator of the quality of the work in the plant when 17 one considers the underlying basis for the discrepancies 18 which themselves provide the basis for the number of 19 defects in the calculation.
20 Specifically, if one refers back to the 21 Intervenors' example of yesterday with a sample 22 including 100 items, 20 of which are defective given a 23 craft error rate of 20 percent and a QC Inspector 24 accuracy rate of 70 percent, that quality system would
(
25 have identified 19.4 of the 20 defects in the population Sonntag Reporting Service, Ltd.
GGTmVa, Illinois 6ul34 f
(312) 232-0262
16947
'O 1
of 100 items -- excuse me.
It would have identified 14 2
of the 20 defects in the population.
3 As indicated by the Intervenor -- I believe I'm 4
going to have to correct myself again.
5 With the 70 percent GC Inspector accuracy rate, 14 6
of the 20 defects would have been identified in a 7
population of 100 items, leaving 6 defects in the 8
population of 100 items.
9 The ultimate product quality after QC inspection is 10 94 percent, with there being 6 defects in a population 11 of 100 items.
12 So the composite system of craf t and OC Inspector 13 produced a product at a 94 percent acceptable rate.
14 The agreement rate calculated by use of the 15 Intervenors' algorithms predicts -- or provides an 16 agreement rate of about 94 percent.
So the indicator, 17 agreement rate, is a fairly accurate representation of 18 ultimate product quality.
19 But one can't stop once he's made that 20 determination.
One has to look below the agreement rate 21 and assess the nature of the discrepancies for their 22 relative significance.
23 That has also been done and reported upon in other 24 testimony presented to the Board.
(
25 0
Okay.
Sonntag Reporting Service, Ltd.
Geneva, Illinois 60134 (312) 232-0262
16948 (V
1 And the discrepancies that you're analyzing are the 2
discrepancies that get by the original QC Inspector?
3 A
Yes, sir.
4 O
Let's turn to Mr. Frankel's testimony, Mr. DelGeorge.
5 You were asked a number of questions the other day about 6
possible differences in the items which were reinspected I
7 in the CSR.
8 Do you recall that line of cross examination?
9 A
Yes, sir.
3 10 0
Are you aware of what populations of CSR data -- what 11 CSR data Dr. Frankel analyzed in his testimony?
f}
12 A
It's my belief that he analyzed the random portion of V
13 the CSR data.
14 0
And did he -- did he provide an analysis which, in 15 effect, aggregated all inspection points for that random 16 portion of the CSR sample?
17 A
Yes, sir.
18 0
Okay.
19 And by aggregating those, he took -- did he take 20 inspection points for different work activities and pull 21 them together, consider them together?
22 A
Yes, sir.
J 23 O
And did he take into account -- did he pull together 24 inspection points arising f rom different types of QC 25 inspections?
Sonntag Reporting Service, Ltd.
Geneva, Illinois bu144 (312) 232-0262
_ ~-
16949 A
N~-]
1 A
Yes, sir.
2 0
And lumping those together, he did some calculations on 3
the variation in agreement rates as a function of time 4
for the CSR data that he was working with; correct?
5 A
Yes, sir.
6 0
Okay.
7 In your view, was it appropriate or inappropriate 8
for Dr. Frankel to lump those things together?
9 A
I believe it was appropriate, and the reason --
10 0
Why?
11 A
My reason for that is the following:
l f~'T 12 Again, the Intervenor here has alleged the b
13 existence of production pressure and harassment which 14 has historic, tangible, powerf ul and pervasive ef f ect.
15 They have, in fact, denied that that effect has been 16 isolated or periodic.
17 Because of that fact, it's my view that a fair reading of Intervenors' contention is that that 18 19 pervasive nature of the stimulus will affect all work 20 activities within the QC Department equally.
21 I think this is reinforced by the fact that Mr.
22 Saklak, for a significant period of time, was 23 responsible for all QC inspection activities.
24 Intervenor has made a repeated point of that fact.
\\
25 It is also true that Mr. DeWald was the QC Manager, a
Sonntag Reporting Service, Ltd.
ueneva, 1111nois culse (312) 232-0262
I 16950 0
1 and we have heard repeatedly that he had a pervasive 2
effect over all QC Inspector activities.
That point has 2
been made about the QC organization generally and about 4
the effects of production pressure by Commonwealth 5
Edison as a general matter.
6 For those reasons, I felt it def ensible -- and I 7
believe Dr. Frankel's approach is defensible -- to 8
aggregate all inspection decisions, as he has done, to 9
identify whether, in fact, there are any trends in the 10 decisions -- that is, the adequacy of the decisions --
11 being made by inspectors generally.
f~'}
12 Q
Now, did Dr. Frankel perform any other analysis of v
13 subpopulations other than -- go ahead.
14 A
It is my understanding that he did, yes.
15 0
Okay.
16 And did he perform an analysis of welding 17 inspection points only?
18 A
Yes, sir, he did.
19 0
Okay.
20 In your opinion, was it inappropriate or 21 appropriate for Dr. Frankel to look at welding 22 inspection points only?
23 A
No, sir.
I believe that was an appropriate decision.
24 My judgment is based on the fact that, again, there 25 is a significant volume of welding work that is produced Sonntag Reporting Service, Ltd.
ueneva, 1111nois culae (312) 232-0262
16951
[N V
1 by Comstock and inspected by the QC Inspector work force 2
of Comstock.
3 In fact, the class of 24 inspectors that reported 4
harassment, intimidation and production pressure to the 5
NRC was comprised in large part of inspectors certified 6
in welding.
In fact, a large percentage of those 7
welding inspectors in the class of 24 have data captured 8
within the CSR data base.
9 For all those reasons, I believe that it's 10 justifiable to review the welding work activity 11 separately to determine whether that activity 12 demonstrates an effect of this alleged pervasive
)
a 13 pressure and harassment.
14 0
Let's go back to Dr. Frankel's review of all inspection 15 points.
16 JUDGE GROSSMAN:
Excuse me.
Mr. Steptoe, 17 something is disturbing me about this.
18 My recollection yesterday was, one, the focus of 19 Mr. Guild's questions was primarily not on what Dr.
f 20 Frankel did but on what was left over that Mr. DelGeorge 21 did; and that when Mr. Guild got into Dr. Frankel's 22 work, Mr. DelGeorge was not very familiar with it and 23 declined to answer to a large extent.
24 I'm not saying that we're going to disallow the l (
25 testimony.
I just want to point out now that we're Sonntag Reporting Service, Ltd.
ueneva, Illinois 6U134 (312) 232-0262 i
16952 O
(~-)
1 going somewhat beyond the scope, and we're going to be 2
expanding the cross examination.
3 You may well be covering the ground Mr. Guild might 4
have covered yesterday if there had been more 5
authoritative responses as to what Dr. Frankel had done.
6 So I just want to point out that we are possibly 7
extending the scope of what we're going to be hearing.
8 But, nevertheless, I haven' t heard any objections, and 9
you can go ahead.
10 However, if --
11 MR. GUILD:
Mr. Chairman, I agree with the 12 observation.
I'm eager to hear what Mr. DelGeorge's 13 version today is of Dr. Frankel's testimony.
So I have 14 no objection.
15 JUDGE GROSSMAN:
Now, Mr. DelGeorge may have 16 discussed this with Dr. Frankel yesterday evening and 17 gotten a better understanding of what Dr. Frankel did.
18 But to that extent, perhaps some of those questions 19 really ought to be directed to Dr. Frankel except to the 20 extent that you want to put those as assumptions to the 21 witness and ask for his judgment if these assumptions --
22 if Dr. Frankel had done something.
1 23 What I'm just suggesting is that it's not really 24 fair for Mr. De1 George not to have that good a 25 recollection on cross examination and then to refresh Sonntag Reporting Service, _ Ltd.
ueneva, 1111nois eulse j
(312) 232-0262
16953 4%./
1 his recollection overnight and have a better 2
recollection on redirect.
3 But it really may not make that big a difference, 4
so we'll just let you continue.
5 MR. STEPTOE:
Well, Judge Grossman, I have to 6
disagree with your comments about this.
7 The line of cross examination which this responds 8
to is a line of cross examination which was, "Did you 9
inform Dr. Frankel that there are different size cable 10 pan hangers?
Did you inform him that there were some 11 with DV7's and some with DV22's?
Did you inform him Ou 12 that there were variations in the kinds of work 13 activities?"
The answer was, "I didn't need to inform 14 him.
He knew already."
15 It seems to me a legitimate thing to say, "Was it 16 appropriate for him to lump these together?"
17 JUDGE GROSSMAN :
But my recollection is the 18 further questions were, "Did Dr. Frankel take these into 19 account," and the answers I recollect were uniformly, "I
20 don't know what Dr. Frankel did.
That's something that 21 Dr. Frankel will have to explain."
22 Now you're going further and asking him whether Dr.
23 Frankel -- they're basically the same questions -- took 24 these into account; and the answers are, "Yes, he did,"
4 (v
25 or with a greater knowledge than he had yesterday.
Sonntag Reporting Service, Ltd.
ueneva, 1111nois culae (312) 232-0262
16954
/
\\
\\s-1 I don't dispute that he may have greater knowledge 2
from having discussed this with Dr. Frankel yesterday 3
evening, but we'll then just have to expand the recross.
4 MR. STEPTOE:
But I don't think that's right, 5
Judge Grossman.
The phrase, "Did he take it into 6
account," is where we part company.
7 I think that everybody in this courtroom knows that 8
he aggregated these things.
That was the point of 9
Intervenors' line of cross examination.
I haven't 10 elicited anything more than that --
11 JUDGE GROSSMAN:
Okay, that's fine.
2 12 MR. STEPTOE:
-- from this witness.
I 13 haven't asked him about Dr. Frankel's thought processes.
14 JUDGE GROSSMAN:
All right.
That's fine.
15 Just proceed now.
16 BY MR. STEPTOE:
17 0
Returning to Dr. Frankel's overall calculated -- overall 18 all-inspection-points calculation, he perf orms a 19 calculation for the overall agreement rate before July 20 1, 1982, and after July, 1982; is that correct?
21 A
Yes, sir.
22 0
And he finds there's no statistical difference in the 23 agreement rate -- no statistically significant 24 difference in the agreement rate; is that correct?
(
25 A
Yes, sir.
Sonntag Reporting Service, Ltd.
Geneva, Illinois 60134 (312) 232-0262
l 16955 1
Q And, similarly, he comes to the same conclusion with i
2 respect to the time period bef ore and af ter August, 3
1983; is that correct?
4 A
Yes, sir.
5 0
Do you know what the -- what difference in agreement 6
rates would have been statistically significant, based 7
on Dr. Frankel's testimony?
8 A
Well, it's my belief that the variation would have had 9
to have been on the order of 1 percent or greater, 10 depending on the confidence level at which the test was 11 performed, in order to establish significance.
[}
12 I believe Dr. Frankel is in a position to perform 13 such a calculation.
14 Q
Now, let's go back to Intervenors' Exhibit 188.
Let's I
15 look at the value of the agreement rate.
Let's take a 16 craft error rate of 20 percent and a QC Inspector 17 accuracy rate of 70 percent.
18 The value there is about 94 percent; right?
19 A
Yes, sir.
1 20 MR. GUILD:
On Chart 17 21 MR. STEPTOE:
Chart 1.
22 MR. GUILD:
90 percent accuracy?
23 MR. STEPTOE:
Yes.
24 BY MR. STEPTOE:
I 25 0
Would a difference in the QC Inspector accuracy rate of
,(
Sonntag Reporting Service, Ltd.
Ueneva, Illinois bulae f
(312) 232-0262
n_
i 16956 1
10 percent have given rise to a difference in agreement 2
rates which, using Dr. Frankel's analysis, would have 3
been statistically significant?
4 A
Yes, sir, I believe it would.
5 0
And how do you find that on that table?
6 A
Well, if one refers to any of the columns -- well, if 7
one refers to the column for craf t error rate of 20 8
percent and observes the difference in agreement rate 9
calculated by the Intervenor between increments of QC 10 Inspector accuracy rate of 10 percent, the difference is 11 on the order of 2 percent.
12 0
okay.
\\
13 A
So from that I would conclude that changes on the order 14 of 10 percent in QC Inspector accuracy rate would be 15 detectable through a review of agreement rates even if 16 calculated in the way Intervenors propose.
17 0
All right.
18 Let's turn to Intervenors' Exhibit 190, which is 19 entitled "CSR Reinspection Results for LKC."
20 (Indicating.)
21 Do you have a copy of that bef ore you, Mr.
22 DelGeorge?
23 A
Yes, sir.
24 0
I think it was brought out yesterday that there are no 25 non-welding inspection points in this chart prior to i
Sonntag Reporting Service, Ltd.
ueneva, 1111nois culde (312) 232-0262
l 16957 l
(x_-
1 January of 1981 or prior to the second quarter of 1981.
2 Is that correct?
3 A
Yes, sir.
4 0
What is the explanation for that?
5 A
I believe, as was discussed yesterday, the vast majority J
6 of the work performed at the site through the first j
7 quarter of 1981 was welding work.
8 Given the fact that the samples chosen were in 9
large part random and the fact that the engineering 10 judgment samples were also in part chosen on a random 11 basis, one would expect the distribution within the I
12 sample to be reflective of the distribution of work
\\
13 actually performed.
14 As a result, it is not surprising to me that we see 15 reflected the fact that there is, for all practical 16 purposes, only welding work done prior to January of 17 1981.
18 0
Okay.
19 Is there a sequence of work activities in which 20 various types of items are installed at the plant?
21 A
Yes, sir.
22 0
And Mr. Guild went into that before yesterday, did he 23 not, with respect to the f act that cable pan hangers 24 have to be put up before the cable pans, which have to 25 be put up before the cables?
g 4
Sonntag Reporting Service, Ltd.
ueneva, 1111nois ou134 (312) 232-0262 J
16958 O
1 A
Yes, sir.
2 Q
Do you recall that?
3 With respect to cable pan hangers, are there any 4
non-welding inspection points in this data base?
5 A
No, sir.
6 As is indicated in my testimony and I believe that 7
of Dr. Kaushal, cable pan configuration, which would 8
account for -- cable pan hanger configuration, which 9
would account for the attributes other than welding with 10 respect to cable pan hangers, was determined to be out 11 of the BCAP scope.
12 As a result, it is not reflected in this data base, 13 although those activities would have been proceeding 14 within the period prior to January,1981.
15 Q
Finally, the data reflected in Intervenors' Exhibit 190 16
-- does this reflect random and non-random sample items?
17 (Indicating.)
18 A
Yes, sir.
19 0
Is this the data that Dr. Frankel analyzed in his 20 testimony?
21 A
No, sir.
22 MR. STEPTOE:
If I could have just a moment.
23 JUDGE GROSSMAN :
Sure.
24 MR. STEPTOE:
If I could have a five-minute 25 break, I think that --
Sonntag Reporting Service, Ltd.
ueneva, 1111nois culae (312) 232-0262
16959 1
JUDGE GROSSMAN:
Okay, fine.
2 MR. STEPTOE:
-- I'll have just a few more 3
questions and then I'll be finished.
4 JUDGE GROSSMAN:
All right, fine.
Let's take 5
10 minutes now.
6 (WHEREUPON, a recess was had, after which 7
the proceedings were resumed as follows:)
8 JUDGE GROSSMAN:
Back in session.
9 Mr. Steptoe?
10 MR. STEPTOE:
Thank you, Judge Grossman.
I 11 have only one further thing.
(
12 You asked a question earlier this morning about 13 what would be the effect of taking into account the 14 possibility that the original Oc Inspector might not 15 only miss bad welds that he should be rejecting but also 16 reject good work, and we have performed additional 17 calculations.
Mr. DelGeorge and Mr. Orlov have put j
18 together two more exhibits.
19 BY MR, STEPTOE:
i 20 0
The first one -- Mr. De1 George, would you identify the 1
21 first one, which I'd like marked as Applicant's Exhibit 22 1 83?
23 (The document was thereupon marked 24 Applicant's Exhibit No.183 for 25 identification as of November 7, 1986.)
i Sonntag Reporting Service, Ltd.
ueneva, Illinois bu134 r
(312) 232-0262
1 I
1 16960 1
MR. GUILD:
Which one is that, now?
2 MR. STEPTOE:
It has " Case 2" at the top.
3 A
Case 2, Applicant's Exhibit 183, has been modified f rom 4
the calculation contained in Applicant's Exhibit 182 to 5
take account of the possibility that the original QC 6
Inspector will reject acceptable product within the 7
original sample for inspection.
That's the only change 8
made to the calculation.
9 One derives the agreement rate.
Taking that 10 assumption into account, the calculated agreement rate 11 is approximately 82.2 plus percent, compared to the 12 value of 84.4 percent reported as the agreement rate in 13 Applicant's Exhibit 182, where that assumption with 14 respect to the OC Inspector -- the QC Inspectors' 15 rejection of acceptable work not having been taken into 16 account.
17 BY MR. STEPTOE:
18 0
And what value does that correspond to in Intervenors' 19 Exhibit 188?
20 A
That corresponds to the value of -- I believe it was 21 93.7 in Intervenors' Exhibit 188.
22 0
Now, is there -- you have another exhibit, Mr.
23 De1 George, entitled " Case 3."
24 A
Yes, sir.
25 MR. STEPTOE:
That will be Applicant's l
I Sonntag Reporting Service,__ Ltd.
ueneva, 1111nois culae (312) 232-0262
16961
%)
1 Exhibit 184.
2 (The document was thereupon marked 3
Applicant's Exhibit No.184 for 4
identification as of November 7, 1986.)
5 JUDGE COLE:
We only have one copy of that up 6
here.
i 7
MR. STEPTOE:
We didn't have time to supply 8
copies.
9 JUDGE COLE:
Fine.
We'll work with what we 10 have.
11 A
(Continuing.)
Case 3, shown in Applicant's Exhibit 12 184, goes through a similar calculation, with the only 13 change from Case 2 being the assumption that the 14 overinspector error detection rate changed from 90 15 percent to 80 percent.
16 That has the effect of reducing further the 17 calculated agreement rate f rom 82.2 plus percent to 18 approximately 74.2 percent.
19 That would be a value that corresponds to the value 20 reported in Table 2 of Intervenors' Exhibit 188 of, I 21 believe, 94.4 for an 80 percent overinspection accuracy 22 rate and OC inspection accuracy rate of 70 percent and a 23 craft error rate of 20 percent.
24 BY MR. STEPTOE:
i O
(
)
25 0
Applicant's Exhibits 183 and 184 -- are they accurate, Sonntag Reporting Service, Ltd.
ueneva, Illinois eu134 (312) 232-0262
l 16962 (O
V 1
to the best of your knowledge and belief?
2 A
Yes, sir.
3 MR. STEPTOE:
Judge Grossman, I ask that 4
Applicant's Exhibits 183 and 184 be received into 5
evidence.
6 This concludes my redirect.
7 JUDGE GROSSMAN :
Okay.
8 Are these agreement rates that would have been 9
reflected in the actual agreement rates that were 10 determined as part of the review of the CSR program?
11 THE WITNESS:
Your Honor, if I might respond 12 to that, the purpose of these calculations is to 13 demonstrate that one cannot look to the calculation 14 performed in Intervenors' Exhibit 182 as necessarily i
15 reflective of the agreement rates that one would find in 16 the field where one does not take into account the field l
17 reality of the f act that overinspectors do, in fact, 18 reject acceptable work.
19 I think we --
20 MR. GUILD:
Excuse me.
That's 188, Mr.
21 Chairman.
22 THE WITNESS:
Excuse me; yes.
23 JUDGE GROSSMAN:
I'm sorry.
What did you 24 say?
i 25 MR. GUILD:
188.
Mr. De1 George just Sonntag Reporting Service, Ltd.
Geneva, Illinois bu134 (312) 232-0262
16963 Ak) s 1
misspoke.
2 JUDGE GROSSMAN:
Oh, okay.
3 THE WITNESS:
As we have discussed in 4
previous testimony, there is a conservatism in the 5
overinspection process that finds overinspectors 6
rejecting acceptable work.
7 Intervenors' calculation reported in Intervenors' 8
188 does not take account of that f act.
The 9
calculations we've provided in Applicant's Exhibits 182, 10 183 and 184 demonstrate the effect of taking into 11 account that conservatism.
[v}
12 I believe that while this may represent a limit as 13 to the effect of considering that field reality, it is 14 probably reflective of the kinds of overinspection that 15 we have, in fact, seen both within the CSR reinspection 16 program and the PTL overinspection program.
17 MR. GUILD:
Mr. Chairman, on the question of 18 the offer of these documents, again, if we could have a 19 few moments to review them before we take a position.
20 JUDGE GROSSMAN :
Fine.
I won't rule on these 21 documents, then, until after Mr. Guild has 22 cross-examined the witness.
23 Now, do you wish a few moments now?
24 MR. GUILD:
Yes, if we can to review the 25 cales, that would be helpful, Mr. Chairman.
Sonntag Reporting Service, Ltd.
Geneva, Illinois cul34 f
(312) 232-0262
16964 m
l'w.))
1 JUDGE GROSSMAN:
Sure.
2 (WHEREUPON, a recess was had, after which 3
the proceedings were resumed as follows:)
4 JUDGE GROSSMAN:
Ready, Mr. Guild?
5 MR. GUILD:
Yes, sir, Mr. Chairman.
6 JUDGE GROSSMAN:
Please proceed.
7 RECROSS EXAMINATION 8
BY MR. GUILD:
9 0
Mr. De1 George, you make an interesting concession, sir; 10 that is, that agreement rates -- while we may disagree 11 about what agreement rates do stand for, they are a 12 fairly good measure, all other things equal, of the 13 state of work quality found by the overinspector.
14 A
I did make a statement similar to that, but I also added 15 the caveat that one needs to consider the significance 16 of the discrepancies that are called " defects" for 17 purposes of calculating the agreement rates.
4 18 Q
Sure, and using your dispassionate engineering judgment 19 as the Assistant Vice-President for Nuclear Licensing, 20 you've concluded that, despite the fact that there's 15 21 percent deficient welds in the composite mean agreement 22 rate and that is a meast're of work quality, you've 23 looked at those deficient welds and don' t find that 24 result troubling?
25 A
It is -- it is my view that that result is not l
Sonntag Reporting Service, Ltd.
l Geneva, 1111nois bu1J4
[
(312) 232-0262
16965 1
troubling.
2 The reason is that I do not believe that all of the 3
defects that have been used as a basis for calculating 4
that agreement rate are, in fact, true defects; and 5
there is a large percentage of the defects which are, in l
6 fact, valid defects which are, in my opinion, 7
insignificant.
8 0
And, of course, Sargent & Lundy evaluated those defects, 9
and that's Mr. Kostal's testimony in this proceeding.
10 Now, the agreement rate data is a good measure of 11 work quality except, of course, to the extent that the
/N 12 agreement rate also represents some level of inaccuracy l
13 on the part of the overinspector, in which case if the 14 overinspector missed defects, that agreement rate would 15 understate the rate of defects in the population?
16 A
While that is true, I believe I also testified -- and 17 one can deduce from observation of Intervenors' Exhibit 18 188 -- that those kinds of changes in overinspection 19 accuracy rate are a less sensitive driver of the 20 calculated agreement rate than is the craft error rate 21 and the in-line QC accuracy rate.
22 0
Sure.
I 23 I mean, if it's -- if we have 1 percent more bad 24 welds in the electrical population and 1,000 -- or
{
25 100,000 welds in the electrical population, there would 1
Sonntag Reporting Service, Ltd.
ueWeva, 111inUis 0u134 (312) 232-0262
I 16966 d
1 be 1,000 more bad welds out there, assuming that those 2
were welds not found -- that proportion of welds was not 3
found by the overinspector because of a 1 percent level 4
of inaccuracy?
5 A
I'm not sure that's true, sir.
6 From your table, a 1 percent change in agreement 7
rate would correspond to one-tenth of a percent change 8
in agreement rate if we modify the overinspector's 9
performance by some value.
10 I thought your prior question referred to the 11 performance -- differences in performance of the I~'h 12 overinspector.
% s]
13 0
It does.
14 I'm just simply making the point, Mr. De1 George, 15 that if your agreement rate -- that is, the composite 16 rean saying 15 percent bad welds were found -- if that's 17 wrong and if it's wrong on the order of 1 percent, if 18 instead it's 16 percent bad welds and we're talking 19 about a population of 100 welds, the inference one would 20 draw -- a population of 100,000 welds, the inference one 21 would draw is that 1 percent greater defective welds 22 would represent 1,000 more defective welds out there.
23 A
one might draw that inference, but I think it would be 24 inappropriate to do so --
(Oj 25 0
Yes.
v Sonntag Reporting Service, Ltd.
ueneva, 1111nois bu144 (312) 232-0262
16967 v
1 A
-- without knowing more.
2 0
All right, sir.
3 Well, now, let's see if the agreement rates mean 4
anything else, if they mean anything beyond giving us 5
some measure, at least for the sample, of what the 6
extent of defects are in the population, again assuming 7
all other things being equal about the sampling
]
8 technique and quality of inspection.
9 Let's look at your Exhibit 182, Mr. De1 George.
10 Now, what underlies 182,183 and 184 is an effort I
11 to capture the fact that, in your judgment, accuracy
(
12 rates should be applied on both sides of accurate; that 13 is, to the failure to accurately detect defects, which 14 is what Intervenors' exhibit considered, as well as to 15 the failure of the inspector and overinspector to 16 accurately identify acceptable items; isn't that 17 correct?
18 A
Yes, sir, if by that you mean that the inspector might 19 identify an acceptable item as a rejectable item.
20 Q
Yes; false rejects, in effect.
21 A
Yes, sir.
22 0
And in 183 and 184, you responded to the question of the 23 Chairman by being consistent and calculating both sides 24 of accuracy, both for the overinspector and the original
(
25 inspector; and that produced some changes in the Sonntag Reporting Service, Ltd.
ueneva, Illinois evise f
(312) 232-0262
16968
/N U
1 agreement rate, changes which you characterized as 2
slight or insignificant?
3 A
Yes, sir, with the exception of that shown on -- in Case 4
3, Applicant's Exhibit 184.
5 0
In that one you also changed another variable, though?
6 A
Yes, sir.
7 0
Now, I take it that you agree in principle with Mr.
8 Juran's observations about inspector accuracy.
9 You said so the other day, did you not?
10 A
While I believe that's true, if you would be more 11 specific in identifying the specific position --
f~}
12 0
Yes, Juran, " Quality Control," third edition, Page v
13 12-51.
The title of the chapter is " Inspection Errors."
14 There Mr. Juran says as follows.
It's the second 15 full paragraph -- the second paragraph, really, under 16 the heading " Inspector Errors."
17 I quote:
"Each of these categories has its own 18 unique causes and remedies."
It's referring to a list 19 of types of errors.
.l 20
" Collectively these inspector errors result in a 21 performance of about 80 percent accuracy in finding 22 defects; i.
e.,
inspectors find about 80 percent of the 23 defects actually present in the product and miss the 24 remaining 20 percent."
25 Now, Mr. Juran speaks about accuracy in the sense Sonntag Reporting Service, Ltd.
Geneva, Illinois 6vl34 (312) 232-0262
16969 (3
/
1 A/
s 1
of accuracy in finding defects, does he not?
2 A
In this section of the document, yes, sir.
3 0
And that's where he uses his 80 percent figure:
80 4
percent with respect to finding the defects that are 5
present in a product?
6 A
In this section of the document, that's correct.
7 0
All right, sir.
8 And consistent with that approach, Intervenors 9
displayed defect rates, including an 80 percent --
10 excuse me.
11 Intervenors displayed accuracy rates, including an 12 80 percent accuracy rate as well as a range of accuracy
'~
13 rates above and below accuracy rates in our table, 14 Intervenors' Exhibit --
15 JUDGE GROSSMAN:
188.
16 BY MR. GUILD:
17 0
-- 188.
Thank you, Mr. Chairman.
18 But you stated, in response to Mr. Steptoe, that 19 you thought that that was inappropriate.
20 A
The calculation is a valid calculation, but I don't 21 think it accounts for all of the potential effects on 22 the calculation agreement rate that I know to exist.
23 0
Yes.
l l
24 And so taking that position, you decided that, to 25 be more realistic, you should include the other side of l
Sonntag Reporting Service, Ltd.
Geneva, illinois 6ul34 (312) 232-0262 L
1 i
16970 C(D 1
the coin of accuracy; that is, false rejects?
2 That's what you do in Applicant's 182,183 and 184?
3 A
Yes, sir.
4 O
All right.
Let's see how realistic that is, Mr.
5 DelGeorge.
6 Turn, if you would, in your prefiled testimony to 7
Page 14.
There you speak about the CSR inspections.
8 There you speak about the acceptance rates -- that is, 9
the agreement rates -- found by the CSR inspectors; and 10 you speak about the review of those inspections by the 11 Level III Inspectors -- I'm sorry.
We're talking about O
12 PTL.
v 13 MR. STEPTOE:
Is that on -- oh, I see.
14 MR. GUILD:
Page 14.
15 BY MR. GUILD:
16 O
These are PTL and -- Level III reviews of the PTL 17 results, are they not?
18 A
Yes, sir.
19 0
All right, sir.
20 Now, the finding of the Level III Inspectors of the 21 review of the PTL overinspections was that the rate by 22 which the PTL overinspectors found false rejects -- that 23 is, the rate at which the PTL overinspectors 24 inaccurately identified an acceptabic condition to be 25 rejectable -- was only.72 percent; that is, less than 1 Sonntag Reporting Service, Ltd.
Geneva, 1111MUis 6vl34 (312) 232-0262
16971 m
1 percent.
2 Taking those Level III results into account, you 3
changed your overall PTL agreement rate from 92.56 4
percent up a fraction of 1 percent to 93.28 percent; 5
correct?
6 A
The numbers are correct, sir, but I think implicit in 7
your statement is an error in the understanding of what 8
these values represent.
9 The second of the two values -- that is, the 10 average including overinspection -- that'is, a 11 third-party review -- is representative of a very i
[' N 12 limited portion of the PTL data base.
13 There was, in fact, three-quarters of the PTL data 14 base for which there was no overinspection.
15 0
Yes, sir.
16 These are --
17 A
And --
18 0
Go ahead.
I didn't mean to interrupt.
19 A
So I think the difference between the two numbers is not 20 necessarily reflective of what was, in fact, the reject 21
-- I've forgotten now the reference you made earlier, 22 the term that you used.
23 0
The term is " false rejects."
24 A
Yes.
I 25 I'm not sure that this difference is an accurate i
Sonntag Reporting Service, Ltd.
U6TeVB, Illinois 6U134 (312) 232-0262
16972 fV 1
representation of the true false reject rate.
2 0
Well, if it's an accurate representation of anything.
3 You characterize it as an accurate representation of the 4
accuracy levels of the PTL overinspectors.
5 If it's an accurate representation of that, it's 6
also got to be an accurate representation of what I 7
interpret it to mean; that is, a very small level -- or 8
rate of inaccuracy of the PTL inspectors with respect to 9
identifying false rejects.
10 THE WITNESS:
Sir, I think you misrepresent 11 the position that we've taken.
[O h
12 MR. GUILD:
I'm sure I do.
It's not the same 13 position you take, Mr. DelGeorge, but answer my 14 question, if you would.
15 THE WITNESS:
Could you repeat the question, 16 please?
17 (The question was thereupon read by the 18 Reporter.)
19 A
The problem I have in responding to the question is you 20 refer to an accurate representation of the performance.
1 21 Implicit in the question is a comparison between 22 the two numbers, and I'm having a hard time 23 understanding how to respond to that question.
l 24 DY MR. GUILD:
)
25 0
Well, sir, I think the record is clear.
Your prefiled 3
Sonntag Reporting Service, Ltd.
Ueneva, 2111hols 0U144 (312) 232-0262
16973 m
1 testimony speaks for itself, Mr. DelGeorge.
2 Let's take a look at what happens if we suppose --
3 and I'm asking you to suppose -- a more realistic level 4
of inaccuracy on this side of the coin; that is, the 5
false reject side of the coin; that is, the basis for 6
Applicant's 182,183 and 184.
7 MR. GUILD:
One moment, please.
8 BY MR. GUILD:
9 0
All right, sir.
Let's turn to Applicant's Exhibit 182, 10 then.
11 Using your methodology, let's look at what I would (v}
12 ask you to assume is an accuracy rate that is more 13 realistic -- at least, consistent with my interpretation 14 of your direct testimony, Page 14.
15 A
Sir, in that regard, if I believe your interpretation is 16 incorrect, would you like me to correct you or just make 17 an assumption?
18 0
You've already told me that you believe that, Mr.
19 De1 George.
I'm asking you to assume, sir, some values.
20 Let's see what effects they have.
21 Now, taking your methodology, let's assume a 100 22 percent overinspector accuracy rate; that is, 100 23 percent, of course, applying no errors on either side, 24 either f alse rejects or f ailures to identify rejectable 25 conditions.
Sonntag Reporting Service, Ltd.
Geneva, IltinUis 60134
16974
~h (G
1 Can we agree, sir, there that with 20 percent 2
defective welds and a 70 percent accuracy rate of the 3
original QC Inspector, that we would produce an 4
agreement rate, using Intervenors' two cases -- that is, 5
with the repaired welds replaced into the sample and 6
without replacing the repaired welds respectively -- of 7
94 percent and 93 percent?
8 Again, I would ask you people to check our 9
rounding.
We've done this in the old school of math.
10 THE WITNESS:
I can't, in my head, verify 11 that the 94 percent is correct, and I'd have to take 12 your word at this point for the calculation with 13 replacement.
14 JUDGE GROSSMAN :
Would it help if I gave the 15 witness a pencil and paper?
16 THE WITNESS:
It would help if somebody gave 17 me a calculator.
18 JUDGE GROSSMAN:
Oh, okay.
19 A
I get 93.6, which is different than your 93.
20 BY MR. GUILD:
21 0
I have 94 and 93.
22 94, I take it, is the case where welds are repaired 23 and replaced in the sample.
24 A
Yes.
O}
(
25 Where they' re not replaced in the sample, the Sonntag Reporting Service, Ltd.
I Geneva, Illinois 6ul34 (312) 232-0262
16975 mU 1
population for overinspection would be 94.
2 The number of rejects in that population would be 3
6.
So there would be 88 accepted out of a population of 4
94.
5 I get 93.62 percent.
6 0
All right.
That's close enough.
Let's not argue about 7
it.
8 Now let's take another case; that is, where we 9
reduce by a small increment the level of accuracy of the 10 overinspector.
11 Let's assume a 98 percent overinspection accuracy 12 again on both sides of the coin:
98 percent in
\\_-
13 correctly finding rejects and 98 percent in correctly 14 finding acceptable conditions.
15 Would you agree, Mr. DelGeorge, that, assuming a 98 16 percent overinspector accuracy rate, the agreement rate, 17 again assuming other -- making other assumptions in 18 Applicant's Exhibit 182, the 20 percent craf t error 19 rate, an agreement rate of 92.5 percent repairing and 20 replacing welds in the sample or 91.3 percent if we 21 repair and don't replace those welds in the 22 sample?
23 THE WITNESS:
Again, without validating the 24 numbers, I think the trend is what I would expect to I
25 occur, but it's going to take me some few minutes to
,L l
l Sonntag Reporting Service, Ltd.
Geneva, Illinois 6ul34 (312) 232-0262
i l
16976 OV 1
recalculate this.
2 MR. GUILD:
If you would, I'd appreciate it 3
if you would take a moment to do that, sir.
4 THE WITNESS:
What did you arrive at as a 5
result?
6 MR. GUILD:
One second here.
7 We have a value of 92.5 percent with replacement, 8
91.3 percent without replacement.
9 A
I have comparable numbers, but not identical numbers.
10 BY MR. GUILD:
11 0
What numbers do you come up with, sir?
h 12 A
92.56 and 91.9.
(V 13 0
Fine.
We'll take your numbers.
14 JUDGE GROSSMAN:
Mr. Guild, are you using a 15 calculator, also?
16 MR. GUILD:
We're using our hands.
17 (Laughter.)
18 JUDGE GROSSMAN :
Do you need --
19 MR. GUILD:
It would be a help to have one if 20 we're going to get any further, but I'll accept Mr.
21 DelGeorge's calculation as close enough.
22 BY MR. GUILD:
23 0
One more case, Mr. DelGeorge.
l 24 Again, just to cover the range of possibilities, my 25 assumptions about realistic accuracy levels, let's try Sonntag Reporting Service, Ltd.
ueneva, 1111nois 6U144 (312) 232-0262 i
}
16977
(]N
\\~-
1 the case of a 95 percent overinspector accuracy again on 2
both sides of the coin -- that is, 95 percent accurate 3
in finding defects and 95 percent in recognizing 4
acceptable conditions -- with your methodology and your 5
assumptions f rom Applicant's 182.
6 THE WITNESS:
Are we waiting for me?
7 MR. GUILD:
Yes.
8 THE WITNESS:
Oh.
9 A
I get approximately 90.3 for the case with replacement, 10 and 89.3 for the case without replacement.
11 BY MR. GUILD:
f' 12 0
All right, sir.
We'll take your numbers.
~
13 Let me try one other case and ask you if you can do 14 this.
15 Now, what we did in those cases is we assumed that 16 the accuracy levels were the same both for identifying 17 defects and for identifying acceptable conditions.
18 Let's take one other situation; that is, a 19 situation where the accuracy rates are in the range that 20 Intervenors assume in our exhibit, 90 percent down, but 21 that the -- that's the accuracy in identifying defects 22
-- but that the accuracy in identifying acceptable 23 conditions is much higher -- that is, in the range 24 comparable to the findings of the Level III reviewers of 4
(
25 the PTL data -- again assuming that the Level III l
Sonntag Reporting Service, Ltd.
Geneva, Illinois cul34 (312) 232-0262
16978
~
O l
results are a more realistic depiction of the accuracy 2
on that side of the coin.
3 Let me ask you to run this case:
Let's take, just 4
as you did in Applicant's 182, a 20 percent craf t error 5
rate, a 70 percent QC Inspector accuracy rate but an 6
overinspector accuracy rate of 90 percent for accurately 7
detecting defects -- that's the way we've consistently 8
used that accuracy -- but a 98 percent accuracy rate for 9
correctly identifying acceptable conditions.
10 MR. STEPTOE:
Is there a representation that 11 that 98 percent figure relates to Sargent & Lundy Level j
12 III agreement rates?
13 MR. GUILD:
I'm asking that it be assumed.
14 MR. STEPTOE:
All right.
Well, there was a 15 statement before, and I didn't want that to be a part of 16 the question.
17 BY MR. GUILD:
18 0
Let me tell you what our answers are and see how they 19 coincide with yours.
20 For that case, 93 percent with replacement; 91.9 21 percent without replacement.
93.0 --
22 A
I get 93 percent without replacement, and I have a 23 different value -- excuse me.
Let's see.
24 I have 93 percent with replacement, and without 25 replacement the population size is 84 --
Sonntag Reporting Service, Ltd.
ueneva, 1111nois eu134 f
(312) 232-0262
16979 J
1 Q
How about 86, Mr. De1 George?
2 A
Excuse me; 86.
3 91.86.
4 Q
Close enough.
5 And finally -- this will be the last one -- one 6
last case:
A slightly lower accuracy rate on the down 7
side of correctly identifying acceptable conditions, all 8
other things held equal.
9 That would be a 95 percent accuracy rate for 10 identifying acceptable conditions but again 90 percent 11 accuracy for identifying defects, 20 percent craf t error
(
12 rate and 70 percent original QC Inspector accuracy.
U) 13 Let me give you the answers and see if you come up 14 with the same.
We get a 90.6 with replacement and an 15 89.1 without replacement.
16 MR. STEPTOE:
Can you go over what those 17 assumptions are again?
18 JUDGE GROSSMAN :
I'm sorry.
Mr. Steptoe, 19 even though'you're talking to Mr. Guild, you have to 20 speak up so that --
21 MR. STEPTOE:
I'm sorry.
I asked Mr. Guild i
22 if he could repeat what the assumptions for that last i
23 calculation are' so we can follow.
24 MR. GUILD:
Sure.
(
),.
25 Again, it's 90 percent accuracy for the l
l
\\
Sonntag Reporting Service, Ltd.
Geryva, Illinois 60134 I
(312) 232-0262
l 16980 f
\\
l overinspector in identifying defects but a 95 percent 2
accuracy for identifying acceptable conditions, with the 3
same other assumptions:
70 percent accuracy for 4
original QC, 20 percent craf t error rate.
5 THE WITNESS:
Could you repeat again the 6
values that you got?
7 MR. GUILD:
Yes.
We got 90.6 and 89.1 with 8
and without replacement.
9 A
I get the same.
10 BY MR. GUILD:
11 Q
Now, any way you cut it, Mr. DelGeorge -- let's even f'T 12 take your original way of cutting it in Applicant's 182, d
13 before you made any changes, to apply accuracy rates to 14 both first-line and overinspectors.
15 Before we made any changes to the accuracy rates on 16 either side of the coin, you'd agree, would you not, 17 that the difference between the figure that you've used 18 consistently as a measure of inspector performance --
19 that is, the agreement rate -- and the actual accuracy 20 rate of the Quality Control Inspector is in this case 21 14.4 percent?
22 That is, a 70 percent accuracy rate for the Quality I
23 Control Inspector, even by your calculation in 24 Applicant's 182, produces an 84.4 percent agreement O\\
25 rate?
Sonntag Reporting Service, Ltd.
l ueneva, Illinois cul34 l
(312) 232-0262
16981 A
(a) 1 A
Given the range of assumptions that we've made in these 2
calculations, there is that difference, yes.
3 0
Yes.
4 And given the further assumptions that we made when 5
we asked you to confirm the math, the agreement rates go 6
up to a high of 94 percent where the original Oc 7
accuracy was only 70 percent.
8 That's the case where there's 100 percent accuracy 9
of the overinspector on both sides of the coin; correct?
10 MR. STEPTOE:
I'm sorry.
What are you 11 referring to?
( a' )
12 MR. GUILD:
The first case that we supposed 13
-- that is, a 70 percent GC accuracy rate -- leads to a 14 94 percent agreement rate, assuming a 100 percent 15 overinspector accuracy both in finding defects and in 16 finding acceptable conditions.
17 MR. STEPTOE:
Thank you.
I'm sorry to 18 interrupt.
19 BY MR. GUILD:
20 0
Do you agree with that, Mr. DelGeorge?
21 A
That's what these calculations would suggest.
22 0
All right.
Now let's look at your trending evaluations j
23 that you've explained in response to Mr. Steptoe's 24 redirect examination.
25 Mr. Steptoe asked you to look at a couple of i
Sonntag Reporting Service, Ltd.
I Geneva, 111inofE 6ul34 (312) 232-0262
16982 D.
1 quarters and try to posit some explanations about what 2
was going on here.
He was looking at DelGeorge-3.
3 Could you turn to that, please?
That is welds-only 4
CSR results over time.
5 A
Yes, sir.
6 JUDGE GROSSMAN:
By the way, let me ask you a 7
question first, Mr. Guild.
8 Are you also going to vary the QC Inspectors' rates 9
for both acceptable and rejectable conditions, in 10 putting cases to this witness?
11 MR. GUILD:
One second.
f~}
12 We didn't do that, Mr. Chairman.
I don't know what L/
13 it would do to the results.
14 It's certainly a case that could be tried.
If it's 15 of interest to the Board, we'll make some efforts to 16 calculate that.
17 JUDGE GROSSMAN:
Well, you certainly don't 18 have to do it as part of the evidence.
We have the 19 basic principles in here, and it could be part of the l
20 brief.
21 MR. GUILD:
Yes, okay.
22 JUDGE GROSSMAN:
All I can suggest is that 23 there may be some reason for doing that in that we're 24 still dealing with harassment and intimidation questions
()
25 here and not really questions with regsrd to competence l
l l
l Sonntag Reporting Service, Ltd.
ueneva, Illinois evise l
(312) 232-0262
16983
/'~s 1
of the inspectors.
2 MR. GUILD:
Exactly.
3 JUDGE GROSSMAN:
And, of course, with the 4
harassment and intimidation, you would expect the rates 5
to decline on rejectable and not decline on acceptable.
6 That's only my own observation, and so that might 7
be fertile ground to explore for all the parties.
8 I don't know offhand which way that would cut as 9
far as figures to support you or to support Mr.
10 Steptoe's case.
11 MR. GUILD:
We'll certainly check it, Judge.
12 BY MR. GUILD:
N_-
13 0
Mr. DelGeorge, if you would, do you have your Attachment 14 3 revised?
15 A
Yes, sir.
16 0
Now, the period of time that Mr. Steptoe was directing 17 your attention to was, I take it, the quarters in 1983 18 before and after Mr. DeWald came on the scene.
Mr.
19 DeWald, of course, came on the scene in August.
l 20 Those 1983 quarters show rates that you stated
(
l 21 numerically for Mr. Steptoe as 92 plus percent; 75 22 percent, approximately; 94 percent; and about 90 23 percent.
Correct?
24 A
Yes, sir.
(
25 0
I've rounded, but these numbers aren't of significance I
Sonntag Reporting Service, Ltd.
Geneva, 1111nois bu144 (312) 232-0262
I 16984 O
1 beyond two digits, are they?
2 A
Not as represented here.
3 0
All right, sir.
Mr. Steptoe asked you to surmise what 4
could explain a 74.91 percent agreement rate, which I'll 5
characterize as a 75 percent rate, for that particular 6
quarter.
7 Let's look at that particular quarter.
That's the 8
second quarter of '83.
9 Now, how many welds did we have in our sample for 10 the second quarter of '83, Mr. DelGeorge?
11 A
My recollection is that it was on the order of 300.
12 Q
Less than 500, in any event, isn't it?
13 A
Yes, sir.
14 Q
All right.
I'm just looking at your bar graph here, and 15 it appears to be less than 500.
16 300, approximately?
17 A
Yes, sir.
18 0
All right, sir.
19 Now, do you know how many Comstock Quality Control 20 Inspectors' work was part of the sample of those 300 21 welds?
22 A
Yes, sir.
l 23 Q
How many?
l 24 A
Five.
l
[
25 Q
Five?
l l
l l
Sonntag Reporting Service, Ltd.
l Geneva, 1111nois euise l
(312) 232-0262 L
16985 f'~N I
)
%/
1 A
Yes, sir.
2 O
All right, sir.
3 And what was the lowest agreement rate for those 4
five inspectors?
5 A
Approximately 62 percent.
6 0
What were the agreement rates for the others, since you 7
seem to have them at hand?
8 A
They range from 100 percent to 62 percent, and I believe 9
there were a number of them very close to 90 percent.
10 0
All right, sir, okay.
11 Now, generally you've testified that for a number
/N 12 of periods, the composite mean agreement rate for
(_
13 Comstock shown on a quarterly basis was driven primarily 14 by low agreement rates for one or a handful of 15 inspectors.
16 A
Yes, sir.
17 0
And that's the case in this quarter, isn't it?
18 A
Yes, it is.
19 Q
All right.
20 And, in fact, the small size of the sample and the 21 one inspector who had the very low agreement rate is 22 what mostly explains why there's 75 percent here, and it 23 has nothing to do with the variables that you discussed 24 in response to Mr. Steptoe?
N 25 A
I wouldn't say that it has nothing to do with it, but it Sonntag Reporting Service, Ltd.
uEneva, 1111nols eviae f
(312) 232-0262
16986
[J h
1 certainly does affect the confidence with which someone 2
would make a judgment about this particular interval.
3 0
Well, you didn't tell us anything about that lack of 4
confidence or reduced confidence when you responded to 5
Mr. Steptoe.
6 You postulated a variety of explanations when you 7
knew, Mr. DelGeorge, that you simply didn't have enough 8
data, with any significant confidence, to be able to say 9
anything except that you had one low inspector that 10 quarter.
11 A
No, sir.
I disagree with that.
12 In fact, yesterday, in response to questions f rom 13 Judge Cole, I discussed what would constitute a basis 14 for confidence in drawing conclusions about a specific 15 interval.
16 While I'm not sure we discussed the specific number 17 of welds or attributes that would have been used as a basis for dr' wing comparisons between intervals, the 18 a
19 number that I chose for purposes of analysis was 50.
20 So while I would not argue that, on a statistical 21 basis, there is sufficient data within that time step to 22 draw a defensible statistical inference, I did do what I 23 thought was an adequate engineering evaluation of this 24 time step; and it had enough information within it to do 25 that.
Sonntag Reporting Service, Ltd.
ueneva, Illinois ou144 (312) 232-0262 3
16987
>O
\\~-)
1 0
You just didn't explain that to us in response to Mr.
2 Steptoe's redirect question.
3 A
It's not clear to me that that was the focus of Mr.
4 Steptoe's question.
5 0
I'm sure it wasn't.
6 But you can' t draw any conclusions about even what 7
you characterize as Intervenors' theory of this case --
8 that is, pervasive effects on the group of Comstock 9
Quality Control Inspectors for the second quarter of 10 1983 -- when you only had five inspectors in your sample 11 and one of those inspectors had an agreement rate of 60 12 percent, approximately.
13 A
I disagree, sir, especially because a number of the 14 inspectors in that interval were in the target class of 15 24.
16 0
"The target class of 24"?
17 A
Just for purposes of discussion.
We can call them 18 whatever they want.
I feel a little bad about referring 19 to them as "the complaining inspectors," because I don' t 20 mean to suggest any derogatory context.
21 The 24 --
22 O
You already determined that the 24 -- looking at the 24 23 or looking at the non-24 appeared to have no effect on 24 the results on the basis of agreement rates?
25 Targeting them meant nothing, in your evaluation, Sonntag Reporting Service, Ltd.
ueneva, 1111nois eu144 (312) 232-0262
16988 s
1 did it?
2 A
The fact that they constituted a specific class, when 3
evaluated, identified no trends associated with that 4
class.
5 0
All right, sir.
6 Now, in fact, you only used the gross measures in 7
time of quarters to look for the effects of harassment; 8
you didn't look to see whether or not performance varied 9
on an hourly basis, a weekly basis, a daily basis, a 10 monthly basis, even, with respect to the CSR data?
11 Quarters was the most disaggregated time step that f
')
12 you used, isn't it?
\\s /
13 A
In terms of the evaluation that's represented in this 14 figure, that's correct.
15 But I did review the detailed data that constituted 16 each of the quarters --
17 Q
Yes.
18 A
-- which is presented in what amounts to a daily basis.
19 0
Well, it's presented in a daily basis perhaps, Mr.
l l
20 DelGeorge, because they put a date on it.
21 But if you disaggregate five inspectors' CSR 22 inspections that took place over the second quarter of 23 1983 any further, you get meaningless data, don't you, 24 for purposes of trying to trend?
'O 25 A
I'm not sure I would accept the phrase " meaningless,"
( )
l Sonntag Reporting Service, Ltd.
ueneva, Illinois bulse (312) 232-0262
.. =
16989
(
\\
(
c l
but I think -- so I would say no to that.
2 But it is, I think, important to note that what 3
we're discussing here, as I understand it, is a 4
pervasive stimulus that is supposed to have a pervasive, 5
wide-scale effect.
6 0
Yes, sir.
7 A
And I think that can be detected.
8 0
But your interpretation of the term " pervasive" has some 9
very peculiar meaning here.
10 It simply means that, as you interpret it, it's 11 going to be so gross and so invariable among the 12 affected class of all inspectors that it's going to show 13 up when you aggregate the data on a quarterly basis.
14 That's your assumption, is it not?
15 A
I do assume that it will show up, yes, sir.
16 0
Okay.
l 17 But it can be pervasive in the sense that it l
18 happens a lot of the time, that it happens to a lot of 19 people, and still -- and that it shows up in time steps 20 shorter than a quarter or in groups less aggregated than 21 all inspectors or all welding inspectors or the 24; and i
22 you wouldn't even see it in the CSR data the way you've l
23 displayed it, would you?
24 A
The data has been provided in Dr. Kaushal's testimony in 25 a way that a more refined look is possible.
The data Sonntag Reporting Service, Ltd.
ueneva, 1111nois eulse (312) 232-0262
16990 g
x./
1 has been provided in Mr. Marcus' testimony in a way that 2
a more detailed inquiry is possible.
3 In fact, a part of my evaluation relied on those 4
more detailed data bases.
5 0
Well, I know --
6 A
I did not, in fact, upon review of the detailed 7
information, identify a correlation between specific 8
events and specific inspector performance --
9 Q
Yes.
10 A
-- whether or not that performance would have driven the 11 macroscopic index that's reflected in the attachment 1
12 that we're discussing.
d 13 0
Yes.
14 You've acknowledged in your testimony, have you 15 not, that there are reasons for not disaggregating the 16 data beyond the disaggregation that you have displayed 17 in your testimony?
18 A
Yes, sir.
19 0
You acknowledge, do you not, that the PTL data wasn't 20 processed of a probability sample?
21 A
Yes, sir, but I also discussed what my opinion was with 22 respect to that.
l 23 0
Understood, understood.
24 You don't know whether or not the PTL inspector who
()
25 went out to overinspect picked the 10 percent welds on a l
Sonntag Reporting Service, Ltd.
Geneva, Illinois bulae (312) 232-0262
16991 b
1 hanger that happened to be at the bottom of the hanger 2
because they would require him not to climb to the top 3
of the hanger, do you?
4 A
I don't know that for a fact.
5 But in discussing questions of that sort with the 6
PTL management that oversaw the PTL overinspection 7
program, I don't believe that to be a case that would 8
have found itself in this data base -- at least, not 9
very often.
10 Q
Not very often; correct?
11 A
I don' t know, but that is not the practice, as I 12 understand it from the PTL managers.
13 0
As you understand the practice.
14 You' re aware, aren' t you, that the record reflects 15 that there was no prescribed requirement for how the PTL 16 inspectors picked the welds to be sampled?
17 A
I realize that there's no written procedure, but that I
18 does not necessarily reflect the practice that existed i
19 in the field.
20 0
It's certainly clear that there was not a rigorous 21 random process for selecting the welds or the inspectors 22 or the components to be inspected under PTL7 23 A
I believe I've indicated -- and other testimony 24 indicates -- that there is not a rigorous random i O 25 process.
l l
Sonntag Reporting Service, Ltd.
l ueneva, Illinois bu134 (312) 232-0262
16992 1
But I've also discussed my opinion with respect to 2
the limitations associated with that process.
3 0
Yes, understood.
4 Now, you've referred to the fact that you'd expect 5
the CSR overinspectors' accuracy rate to be constant 6
over time.
7 Do you recall that testimony?
8 A
Yes, sir.
9 0
All right.
10 You don't have any empirical basis for judging 11 whether or not the CSR inspectors' accuracy was constant 12 from inspection to inspection or from one CSR inspector 13 to another CSR inspector, do you?
14 A
I don't have an empirical basis, but I think there's --
15 well, I know there is a good reason that one is not 16 necessary.
17 If one -- if we consider the way in which the CSR 18 reinspection program was implemented, a reinspector did 19 not know that he was going to be reinspecting an item i
20 that was installed in 1978, nor did he know whether it 21 was installed in 1984.
He took it as an item to be 22 reinspected, so his performance --
23 0
You're mishearing my question.
I apologize for 24 interrupting.
I didn' t say "over time," which is the O()
25 element you're now explaining.
Sonntag Reporting Service, Ltd.
ueneva, 1111nois eulse (312) 232-0262 i
16993
(-~
O 1
From inspection to inspection or from one CSR 2
inspector to another, you have no empirical basis for 3
opining one way or the other about variability in the 4
accuracy rates on that basis, do you?
5 A
Well, I think we have some data, and that data is 6
discussed in Mr. Smith's testimony, which I believe 7
reflects a high and -- I'm not sure that it indicates 1
8 whether or not it is uniform over time, but it is a very 9
high performance level of the Daniels inspectors who 10 performed the CSR reinspection.
11 0
All right, sir.
Answer my question, Mr. DelGeorge.
(T 12 You don't know whether there's any variation, do O
13 you?
14 That's a simple question.
15 MR. STEPTOE:
Judge Grossman, I think counsel 16 just interrupted Mr. DelGeorge, and he's also arguing 17 with the witness.
18 MR. GUILD:
I'm trying to --
l 19 MR. STEPTOE:
He asked a question about 1
20 empirical data.
The witness was explaining what 21 empirical data there is, and by was cut off.
{
22 MR. GUILD:
No, sir.
I didn't ask a question i
23 universally about empirical data, Mr. Chairman.
I asked 24 a very precise question, and I've tried three times to (JA) 25 get an answer.
l
%~
Sonntag Reporting Service, Ltd.
i ueneva, Illinois ou134 (312) 232-0262
16994 O) l' 1
I'm trying to short-circuit this, but it's 2
impossible when the wit. ness refuses to be responsive to 3
my question and makes a speech every time.
4 THE WITNESS:
Sir --
5 MR. GUILD:
Excuse me, Mr. DelGeorge.
6 JUDGE GROSSMAN:
It's up for ruling.
7 Well, I don't know that he's trying to not respond 8
or to be evasive, but he hasn't answered directly what 9
the question is.
10 It is appropriate, when he does wander off the 11 track, for counsel to ask that just the question and not f'}
12 something else be answered; and the question is still V
13 before him.
14 MR. GUILD:
Let me try again, Mr. Chairman.
15 I'll see if I can be more precise.
16 THE WITNESS:
Excuse me.
May I make one 17 comment?
18 JUDGE GROSSMAN:
Sure.
19 THE WITNESS:
Just so that I understand my 20 responsibilities here, I thought that I did, in response 21 to Mr. Guild's question, indicate that I did not have --
22 or that I, in fact, did have data; and I was trying to 23 explain the nature of the data that I did have.
24 So in that sense I thought I was being responsive
/~N
(,)
25 to the request of whether or not the data existed, and I Sonntag Reporting Service, Ltd.
Geneva, Illinois eulde (312) 232-0262
16995 th I
wasn't able to complete the description of the data that 2
I think does exist.
3 MR. GUILD:
Let me be precise in the question 4
so we have no argument from counsel or the witness about 5
what question is being asked now.
6 BY MR. GUILD:
7 Q
Do you have an empirical measure of variability in 8
accuracy rates of overinspectors from one overinspector 9
to another, from one inspection by an overinspector to 10 another inspection?
11 Does the data show that variation?
12 Yes or no, sir.
13 A
I believe if one looks to the overinspection activities 14 of Commonwealth Edison as a part of the BCAP CSR 15 overview, that there is data therein that provides such 16 an index.
17 0
And what does that data show, Mr. De1 George?
18 I want a precise measure of variation, if indeed 19 your answer is yes to my question that there is an 20 empirical measure of variation from one inspector to 21 another and from one inspection to another.
22 A
I'm not aware of an index that says, "There is this 23 calculated variance between inspectors or this 24 calculated variance for an inspector between 25 inspections."
Sonntag Reporting Service, Ltd.
Ueneva, Illinois eu134 (312) 232-0262
16996
/x 1
What I do know exists is a data base of 2
overinspection of those CSR inspectors which takes into 3
account their work over time and over different 4
populations of work, from which such a calculation can 5
be made.
6 It's my view that the results of the CECO QA 7
oversight of Daniels identified a uniform and very high 8
agreement rate, which leads me to believe that the 9
variance, to the extent one exists, is very small.
10 0
And that's Mr. Smith's testimony.
11 The record will reflect whatever the limits are on 12 that information in Mr. Smith's testimony; is that 13 correct?
14 A
Well, I won't say with certainty -- my understanding is 15 that Mr. Smith discussed the oversight program and that 16 the data associated with that program is in the BCAP 17 final report, and I thought that was a matter that was 18 at least addressed in this proceeding.
19 I may be wrong.
20 0
You mentioned that Dr. Frankel, as you understand it, 21 considered only the so-called random samples in CSR; 22 that is, the samples that were selected using the random 23 number table; correct?
24 A
Yes, sir.
OV)
(
25 0
All right, sir.
Sonntag Reporting Service, Ltd.
Geneva, 1111nois 0u134 (312) 232-0262
16997 mU 1
And that that -- that your attachments, your bar 2
graphs by contrast, aggregate both the so-called random, 3
as that term is used in your testimony, and the 4
non-random samples and results?
5 A
Yes, sir.
6 0
All right, sir.
7 Now, do we have data to show what the various 8
constituents are of the samples over time that Mr. --
9 that Dr. Frankel used comparable to your attachments; 10 that is, data showing what proportion of weld 11 inspections were in the samples, say, before 1982, what 12 proportion of welds were in the sample af ter 1982, et 13 cetera?
14 A
I'm not sure whether that information has been provided 15 or not.
It is not included as a part of my testimony.
16 0
No, it's certainly not, nor is it included as a part of 17 Dr. Frankel's testimony.
18 My question to you, Mr. DelGeorge, is:
Does it 19 exist?
20 A
The data base upon which Dr. Frankel performed his 21 calculations exists, and I believe it would reflect l
22 information of the type that you describe.
23 MR. GUILD:
Mr. Chairman, then I would ask 24 that Applicant be required to produce the data 25 reflecting what the samples were that Dr. Frankel used Sonntag Reporting Service, Ltd.
ueneva, Illinois eulae (312) 232-0262 i
i 16998
/m, 1
that were referred to by Mr. DelGeorge in redirect in 2
response to a question on redirect examination; that is, 3
the so-called random data in a format that is comparable 4
to the DelGeorge attachments, please.
5 JUDGE GROSSMAN:
If that was the subject of 6
your redirect, Mr. Steptoe, I think you ought to supply 7
the information.
8 MR. STEPTOE:
Well, Judge Grossman, I think 9
we have provided that information a long time ago.
10 But with respect to the format that Intervenor now 11 wants, which is bar graphs, we can generate those.
f~ }
12 We've given him the entire data base and the print-out V'
13 and the floppy disks, and we can generate some bar 14 graphs for him if that's what he wants.
15 MR. GUILD:
What I want --
16 JUDGE GROSSMAN:
I'm not asking you to 17 generate any bar graphs.
18 Is that what your request is?
19 MR. GUILD:
Here is the point, Mr. Chairman:
20 If now we learn that some evidentiary reliance is being 21 placed by Applicant on the fact that the variations in 22 the data sources that Mr. De1 George relies on may not be 23 a flaw in Dr. Frankel's testimony, then I'm asking for 24 that data, that evidence.
O) 25 Now, the format, to be meaningful, I suggest needs
(
Sonntag Reporting Service, Ltd.
I beneva, Illinois culas (312) 232-0262
16999
(~V 1
to be in the same format that Mr. DelGeorge used.
For 2
Mr. Steptoe to tell me that I have the files available 3
to me for every CSR data point and I could have dug that 4
information out is somewhat disingenuous.
It's probably 5
true that that information has always been kicking 6
around.
7 The fact of the matter is I now hear that Dr.
8 Frankel picked out the random CSR sample items and that 9
data exists reflecting for over time -- which is what 10 Dr. Frankel addresses, the trending over time -- the 11 constituents of the samples that he looked at.
[ T 12 Now, the format could be in a bar graph, which is
' O 13 what Mr. DelGeorge uses, or it could be in the raw data 14 that would be used to make a bar graph.
15 But I need to know, for example, for each quarter 16 that Dr. Frankel looks at, how many weld inspections 17 there were, how many there were for the 24 QC 18 Inspectors, how many non-welding inspection points there 19 were, the data that Dr. Frankel looked at.
20 JUDGE GROSSMAN:
I wasn't suggesting that Mr.
21 Steptoe merely has to dump all the raw information on 22 you or the raw data.
23 What I was suggesting is that whatever was compiled 24 already by Mr. Steptoe and Dr. Frankel is what ought to 25 be turned over and that we're not putting a burden on Sonntag Reporting Service, Ltd.
Ueneva, 1111nois bu134 f
(312) 232-0262
17000
/'
1 Applicant to construct bar graphs if they weren't in the 2
form of bar graphs.
3 But if the information was aggregcted in some form 4
that Dr. Frankel made use of and that the witness here, 5
Mr. DelGeorge, has referred to, that is what ought to be 6
turned over to Intervenor.
7 MR. STEPTOE:
Judge Grossman, we can give 8
Intervenors the print-out of the data that Dr. Frankel 9
analyzed.
10 However, I do want to take exception to Mr. Guild's 11 repeated comments that he learns for the first time now 12 that Dr. Frankel did something.
13 This testimony was filed back in August; and in the 14 prefiled testimony of Mr. DelGeorge, for example, Page 15 29, a question that Mr. Guild himself read from earlier, 16 the question says, "Dr.
Frankel testifies that, in i
l 17 perf orming his analysis, he only used the items from the l
18 CSR sample which were selected using random sampling."
19 That same statement appears, of course, in Dr.
20 Frankel's prefiled testimony.
So it's just simply not 21 true that we've sprung something here on Mr. Guild.
Mr.
22 Guild's recognition that Dr. Frankel is just using the 23 random portion of the data comes pretty late, it seems 24 to me.
25 JUDGE GROSSMAN:
I don't believe that what l
Sonntag Reporting Service, Ltd.
ueneva, Illinois 60134 (312) 232-0262
17001 Chb 1
Mr. Guild is saying is that he didn't know that Dr.
2 Frankel used the random sample portion.
3 What I think he's saying is that the witness now 4
categorized what Dr. Frankel used those random sample 5
portions for and that, if that data was aggregated for 6
that purpose, then that's what ought to be turned over 7
to Mr. Guild.
8 I don't hear you saying you object to doing that.
9 So we have a long weekend, and I hope you would do 30 that sometime in advance of Wednesday.
11 MR. STEPTOE:
We will give Mr. Guild the data 12 that Dr. Frankel used in performing his calculation in 13 the form that he was given it.
14 JUDGE GROSSMAN :
That's fine.
15 MR. GUILD:
Mr. Chairman, let's be clear.
36 Let's try to lower the tone of this acrimony between 17 Applicant and Intervenor.
18 The new information is the redirect examination of 19 this witness, and that is the distinction that somehow variability and the sample constituents are not a flaw 20 l
21 in Dr. Frankel's testimony because he didn't look et the 22 same thing Mr. DelGeorge did.
t 23 Now, let's lay that aside.
l 24 MR. STEPTOE:
I don't think that's what Mr.
l (N}
25 DelGeorge testified to.
l Sonntag Reporting Service, Ltd.
Geneva, Illinois bu144 (312) 232-0262
17002 0
1 MR. GUILD:
That's fine, Mr. Steptoe.
If 2
you'd just hold your tongue a moment, sir, perhaps 3
you'll have your chance to put your three cents' worth 4
in.
5 The fact of the matter is that what's new, what has 6
been sprung on Intervenor, is the redirect examination 7
that tries to draw this distinction.
8 Now, I would like to have a reasonable opportunity 9
to probe the basis f or this distinction, because I don' t 10 believe that the distinction is meaningful and I think 11 it misrepresents the record.
fh 12 Now, I think what will be shown is when they show O
13 for over time, using the same format as Mr. DelGeorge's i
14 testimony, for Dr. Frankel's random sample, that they'll 15 be predominantly welding inspections in the early part 16 of the period, pre '82; and there will be predominantly 17 non-welding in the later portions because that's when 18 the work was done.
1 19 That was not what Mr. De1 George suggested when he 20 tried to draw this distinction.
21 If all I'm going to get, Mr. Chai rman, from Mr.
22 Steptoe is simply a list of data points that aren't 23 rooted in time -- in other words, "Here is the pile of 24 weld inspection results.
Here is the pile of non-weld
()
25 inspection results" -- it's meaningless.
Sonntag Reporting Service, Ltd.
Geneva, Illinois bu134 (312) 232-0262
i 17003 (3
V 1
So I hear Mr. Steptoe being very craf ty in giving 2
me an answer that appears to suggest that what I'm going 3
to get is meaningless.
I just don't want to go through 4
a meaningless exercise with Applicant.
5 JUDGE GROSSMAN:
Well, we're not going to 6
make any judgment in advance as to whether it's 7
meaningless or not.
8 I'm waiting until Mr. Steptoe confers with Dr.
9 Frankel.
10 Are you finished on that?
11 Because I don't want to say things that slip by Mr.
12 Steptoe.
13 I would assume that these questions will be put to 14 Dr. Frankel.
If, in fact, he doesn't substantiate what 15 Mr. DelGeorge testified to, we would expect that you 16 won't, Mr. Steptoe, refer to Mr. DelGeorge's testimony 17 as supporting any item that Dr. Frankel himself can't 18 support.
19 Now, with that in mind, I don't think we'll have 20 any problem.
Whatever Dr. Frankel can support and has 21 an evidentiary basis for is fine.
Whatever he doesn't 22 as far as Mr. DelGeorge testifying to those facts, they 23 really shouldn't stand as evidence.
24 I don't see you having any disagreement with that, 25 do you, Mr. Steptoe?
l l
l Sonntag Reporting Service, Ltd.
Geneva, Illinois 60134 I
(312) 232-0262
(
17004 V
1 MR. STEPTOE:
I don't have any disagreement.
2 There is a relationship between the two pieces of 3
testimony, and Mr. Frankel makes it expressly clear in 4
his prefiled testimony what he's relying on Mr.
5 DelGeorge for and what he's providing and vice versa.
6 If Mr. De1 George has made a statement within the 7
scope of Dr. Frankel's responsibility that Dr. Frankel 8
can't support, then I'm not going to refer to that in 9
the findings.
10 JUDGE GROSSMAN:
Fine.
I don't see that we 11 have any problem, then.
/~ \\
12 So we'll just expect that all the underlying data, U
j 13 in whatever form Dr. Frankel used, will be turned over 14 to Mr. Guild; and it will be set to go next week when 15 Dr. Frankel appears on Thursday.
16 Mr. Guild, do you still have more?
17 MR. GUILD:
I do have more, and it appears to 18 be 11:00 o' clock.
I would suggest recessing.
19 JUDGE GROSSMAN:
So you do wish to have Mr.
20 DelGeorge back on Wednesday?
21 MR. GUILD:
Yes, I do.
22 JUDGE GROSSMAN:
And we'll hope that you and 23 Mr. Steptoe can arrive at some understanding of how long 24 it might take.
O 25 If you tell him that you have extensive testimony, t
J Sonntag Reporting Service, Ltd.
Geneva, Illinois 6U134 (312) 232-0262
17005
\\
l then we won't have any back-up witness for Wednesday and 2
we'll be prepared to go with Dr. Frankel on Thursday.
3 We'll go with Dr. Frankel first thing on Thursday 4
anyway, I expect.
5 MR. MILLER:
Your Honor, I have been in 6
communication with Mr. McGregor's lawyer, as I said.
He 7
has his very own schedule, which he asks be 8
accommodated.
9 If he receives the appropriate representations from 10 the Office of Inspection and Audit or whatever the 11 initials OIA stand for, he said that he is prepared to 12 go forward with Mr. McGregor on Wednesday afternoon.
13 I will try to confirm that during what's left of 14 today and communicate with Mr.' Guild and counsel for the l'
15 Staff and the Board.
16 It would be my suggestion that if, in fact, Mr.
17 McGregor and his attorney are available only at a set 18 time and that set time happens to be next Wednesday 19 afternoon, that we try and squeeze them in, unless Mr.
20 Guild -- there's going to be -- Mr. DelGeorge is 21
.available.
He's here.
I would just like to accommodate 22 Mr. McGregor and his attorney if we could.
23 MR. GUILD:
I think that's wise, Mr.
24 Chairman.
I have no objection to that.
25 I would suggest, so that we don't have any gaps --
Sonntag Reporting Service, Ltd.
Geneva, 1111nois culde (312) 232-0262
l 1
4 1700'6 f3 t
i V
1 or so Intervenors is not blamed f or that, in any event 2
-- that they have Dr. Frankel ready to go, too, in the 3
event that Mr. McGregor is finished or can't make it.
i 4
JUDGE GROSSMAN:
Dr. Frankel can't make it on 5
Wednesday.
6 MR. GUILD:
Okay, fine.
7 JUDGE GROSSMAN:
I think we're going to fill 8
up Wednesday.
The only problem I foresee is that we i
9 won't finish with Mr. DelGeorge before we have Dr.
10 Frankel.
11 But if that doesn't pose a problem for you, Mr.
12 Guild --
G 13 MR. GUILD:
It doesn't.
14 JUDGE GROSSMAN:
-- it certainly doesn't for 15 us.
16 Now, let me again urge on the record that Staff 17 complete whatever is necessary in order to get Mr.
38 McGregor on here on Wednesday, to begin with, and 19 possibly on Friday if we don' t complete him on Wednesday 20 or even get to him.
21 Again, let me say that I can understand his 22 attorney not wishing to take any chances until'he has 23 something in writing with regard to that second 24 investigation.
25 So I do want to put the ball in Staff's court, but Sonntag Reporting Service, Ltd.
ueneva, 1111nois ou14e (312) 232-0262
4 s
17007 h
i 1
really not in your court, Mr. Berry.
I don' t think that 2
you have any control over that, except to convey what 3
we're saying here to those who can complete that second 4
investigation.
5 If you wish to comment, certainly.
t, 6
MR. BERRY:
Thank you f or the opportunity, 7
Mr. Chairman.
8 Just so we're clear, I understand from Mr. Miller 9
that Mr. McGregor's attorney has indicated that Mr.
10 McGregor, if he receives requisite' assurance, will be j
11 available to testify next week.
fN 12 What I would just ask f rom the Board by way of As 13 clarification is:
Is the Board suggesting or ordering I
14 or ruling or holding that the Board itself will not 15 compel Mr. McGregor to resume the stand until this l
' 16 investigation has been completed?
l 17 JUDGE GROSSMAN:
Absolutely.
We don't intend 18 to put him in any kind of unnecessary predicament here; l
19 and I think, from what I've heard, that it is t
J 20 unnecessary because it appears as though all that is i.
21 required -- a'nd that I believe is represented to his 1
22 attorney -- are the formalities necessary to close that 23 investigation.
^
24 If there's something more substantive, that has not 25 been told to us or to Mr. McGregor's attorney.
So in
(
j s
l Sonntag Reporting Service, Ltd.
t j
ueneva, Illinois bu144 (312) 232-0262
l 17008 (N,)
I the absence of the formalities, I think that we 2
shouldn't put him in that kind of predicament.
3 If f ormalities are all that's required, that's an 4
unnecessary predicament.
5 MR. BERRY:
I understand, Mr. Chairman, and I 6
certainly will relay this back to the responsible people 7
in the Staff.
8 JUDGE GROSSMAN:
Thank you.
9 Well, then -- Mr. Steptoe?
10 MR. STEPTOE:
Judge Grossman, I do have a 11 motion pending f or Exhibits 182,183 and 184.
/~'N 12 Can they be admitted now or shall they be reserved NY 13 until next week?
14 MR. GUILD:
I ask that they be reserved, Mr.
15 Chai rm ari.
16 JUDGE GROSSMAN:
Okay, fine.
So we'll 17 reserve that also until Mr. DelGeorge resumes the stand.
18 So we're adjourned now until 2:00 o' clock on 19 Wednesday, and we'll expect that the witnesses will be 20 here as agreed to by counsel.
Thank you.
21 (WHEPEUPON, at the hour of 11:05 A.
M.,
22 the hearing of the above-entitled matter 23 was continued to the 12th day of 24 November, 1986, at the hour of 2:00 f
25 o' clock P.
M.)
w Sonntag Reporting Service, Ltd.
Ueneva, Illinois culde (312) 232-0262
NO PAGE NUMBER CERTIFICATE OF OFFICIAL REPORTER S
This is to certify that the attached proceedings before the UNITED STATES NUCLEAR REGULATORY COMMISSION in the matter of:
NAME OF PROCEEDING:
BRAIDWOOD STATION UNITS 1 42 COMMONWEALTil EDISON (IIEARING)
DOCKET NO.:
50-456/457/0L PLACE:
CHICAGO, ILLINOIS g~s
(
DATE:
FRIDAY, NOVEMBER 7, 1986 were held as herein appears, and that this is the original transcript thereof for the file of the United States Nuclecr Regulatory Commission.
~(siat) A r).J G
',/
(TYPED)
Official Reporter Reporter's Affiliation 8
..