ML20135E649
| ML20135E649 | |
| Person / Time | |
|---|---|
| Issue date: | 12/03/1996 |
| From: | Advisory Committee on Reactor Safeguards |
| To: | |
| References | |
| ACRS-T-2084, NUDOCS 9612110391 | |
| Download: ML20135E649 (246) | |
Text
- _ . ._
Officict Trcnscript cf Prscccdings O NUCLEAR REGULATORY COMMISSION l ACRS F-2o4
Title:
Advisory Committee on Reactor Safeguards Human Factors Subcommittee l TRO4 (ACRS)
REWRN ORIGINAL I Docket Number: (not applicable) M B3kHIra l M/S T-2E26 !
415-7130 THANKSI \
Location: Rockville, Maryland O Date: Tuesday, December 3,1996 i
Work Order No.: NRC-922 Pages 1-210
- g 21 g 1 961203
! T-2004 PDR O7 Oj\ld " I}, k -
NEAL R. GROSS AND CO., INC.
Court Reporters and Transcribers l 1323 Rhode Island Avenue, N.W. ()I W shington, D.C. 20005
()~ 1100D (202> 234-4433
~
n q AqlS v c... ce so 3y - Re~e,1 .,
.~
'0 ":^e _ITe 0 : e ,o:---
. mee
DI8 CLAIMER PUBLIC NOTICE BY THE UNITED STATES NUCLEAR REGULATORY COMMISSION'S ADVISORY COMMITTEE ON REACTOR SAFEGUARDS DECEMBER 3, 1996 J
j The contents of this transcript of the proceedings of the United States Nuclear Regulatory i
Commission's Advisory Committee on Reactor Safeguards on
- DECEMBER 3, 1996, as reported herein, is a record of the i
! discussions recorded at the meeting held on the above date.
I j This transcript has not been reviewed, corrected i
l and edited and it may contain inaccuracies. l a
i j
i i
!O 4
h a
1 1 l
1 UNITED STATES OF AMERICA I
,.- 2 NUCLEAR REGULATORY COMMISSION u 3 +++++
l 4 MEETING l 1
l 5 ADVISORY COMMITTEE ON REACTOR SAFEGUARDS 6 (ACRS) ,
i 7 HUMAN FACTORS SUBCOMMITTEE 8 +++++
9 TUESDAY, DECEMBER 3, 1996 10 +++++
11 ROCKVILLE, MARYLAND i l
12 +++++
I 13 The Subcommittee met at the Nuclear Regulatory !
f~)
/ 14 Commission, Two White Flint North, Room T-2B3, 11545 15 Rockville Pike at 8:30 a.m., George E. Apostolakis, 16 Chairman, presiding.
17 18 COMMITTEE MEMBERS:
19 GEORGE E. APOSTOLAKIS Chairman 20 THOVAS S. KRESS Member 21 DON W. MILLER Member 22 ROBERT L. SEALE Member l
23 l
24
, 25 NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N W.
(202) 234-4433 WASHINGTON D C. 20005-3701 (202) 234-4433
2 1 ACRS STAFF PRESENT:
, 2 Amarjit Singh t t 3 Noel Dudley 4 John T. Larkins 5
6 ACRS CONSULTANTS PRESENT:
7 James J. Carroll 8
9 ALSO PRESENT:
10 Frank Coffman 11 Frank Collins 12 Fred Combs 13 Dick Eckerode
~
'( )) 14 Juel Kramer 15 Jay Persensky 16 Josephine Piccone 17 Jack Rosenthal 18 Ernie Rossi 19 Dennis Serig 20 Cecil Thomas 21 Eugene A. Trager j i
22 23 24 O.
25 NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N W.
(202) 234-4433 WASHINGTON. D C. 20005-3701 (202) 234 4433
3 1 I-N-D-E-X 2 Aaenda Item Pace
\')
(
3 Introduction by the Subcommittee Chairman 4 4 Answers to ACRS Questions 6 5 (Continued) 132 6 Office of Analysis and Evaluation of 7 Operational Data 76 8 Office of Nuclear Material Safety and Safeguards 132 9 Discussion 10 11 12 13 0
14 15 1
16 j 17 18 19 20 ;
i 21 22 23 i
24 )
n 25
()
NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISL AND AVE , N W.
(202) 234-4433 WASHINGTON. D.C. 20005-3701 (202) 234-4433
i 4
1 P-R-O-C-E-E-D-I-N-G-S 2 (8:36 a.m.)
7-( i 3 CHAIRMAN APOSTOLAKIS: The meeting will now I
4 come to order. This is a meeting of the ACRS Subcommittee j l
4 5 on Human Factors. I am George Apostolakis, Chairman of l
6 the Subcommittee.
7 The ACRS members in attendance are: Tom Kress 8 and Robert Seale. We also have in attendance as a 9 consultant Mr. James Carroll.
10 The purpose of this meeting is to hold l 11 discussions with representatives of the NRC staff I 12 concerning questions raised during the September 20, 1996 13 Human Factors Subcommittee meeting and the activities of
/%
- 14 the Office for Analysis and Evaluation of Operational Data 15 and the Office of Nuclear Material Safety and Safeguards 16 in the human factors area.
17 The Subcommittee will gather information, 18 analyze relevant issues and facts, and formulate proposed 19 positions and actions as appropriate for deliberation by 20 the full Committee.
21 Amarjit Singh is the cognizant ACRS staff 22 engineer for this meeting.
23 The rules for participation in today's meeting 24 have been announced as part of the notice of this meeting r
( 25 previously published in the Federal Register on November NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N W.
(202) 234-4433 WASHINGTON, D C. 20005-3701 (202) 234-4433
5 1 19, 1996.
e, 2 A transcript of the meeting is being kept and L) 3 will be made available as stated in the Federal Register 4 notice. It is requested that the speakers first identify 5 themselves and speak with sufficient clarity and volume so 6 that they can be readily heard.
7 We have received no written comments or 8 requests for time to make oral statements from members of 9 the public.
10 Based on the staff presentations at the 11 September 20, 1996 Human Factors Subcommittee meeting and 12 at today's meeting, the Subcommittee should determine what 13 issues should be brought forward to the December 5th ACRS
'- 14 meeting. The full Committee is expected to review and 15 comment on these issues.
16 Comments on the Human Factors Research Program 17 need to be provided prior to January 1st, 1997 to support 18 the ACRS letter to Congress. Time has been scheduled at 19 the end of the meeting for discussing our thoughts and for 20 providing guidance to the staff on information to be 21 included in its full Committee presentation.
22 Dr. Seale will chair those portions of the 23 meeting related to work done by INEL due to concerns 24 related to conflict of interest.
25 We will now proceed with the meeting. And I NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE. N W.
(202) 234 4433 WASHINGTON, D C. 20005-3701 (202) 234-4433
6 1 call upon Dr. Thomas, Chief of the Human Factors j
,_ 2 Assessment Branch, NRR, to begin.
4
) 1 3 MR. THOMAS: Thank you and good morning, )
4 gentlemen.
5 ANSWERS TO ACRS OUESTIONS 6 MR. THOMAS: This morning we have proposed a 7 panel format to address the questions that you posed to us 8 as a result of the last Subcommittee meeting. If it's 1 l
9 okay with the Subcommittee, we would just go through the I 10 questions one by one and pose the answers.
11 At the table today are Frank Coffman of l 12 Research, Jay Persensky of the Office of Research, Dick 13 Eckerode from NRR with me, and from AEOD Jack Rosenthal. 1 O
\~) 14 And we have one other person who will join us later from i 15 the Operator Licensing Branch in NRR to address the answer 16 to Question 16 concerning simulator fidelity.
17 Does anyone at the table have anything they 18 would like to open with? If not, I'll proceed with 19 Question 1. Would you like for me to read the question 20 and then --
21 CHAIRMAN APOSTOLAKIS: I think it would be 22 useful, yes.
23 MR. THOMAS: Okay. Question 1 is: What are 24 the staff plans for developing a Human Performance Program O
() 25 plan activities road map, which would be useful for NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N W.
(202) 234-4433 WASHINGTON, D C. 20005-3701 (202) 234-4433
7 1 allocating resources to, scheduling, and understanding the
- 2 relationship between the activities?
i
\' 3 The simple answer is that at this time we only 4 have plans to maintain the Human Performance Program plan 5 in essentially its current form for the following reasons.
6 The Human Performance Program plan in its 7 present form was what was envisioned by the Human Factors 8 Coordinating Committee charter. The program plan is 9 adequately serving the staff's needs, as evidenced by a 10 report from each of the offices that use the plan.
11 When we developed Revision 1 to the plan, that 12 -- was the latest revision that was issued in July -- we 13 attempted to do a number of the things, certainly not all 14 of them, but a number of things, that were suggested to be 15 in a road map.
16 For example, we did try to establish a 17 relationship among the activities that was missing from 18 the first, the initial plan. We did include references to 19 related activities and attempted to address how they were 20 related. We also included schedules for accomplishing 21 each of the activities and intermediate milestones. And 22 to the extent that the schedules that depended on the l
23 outcome of other activities, we attempted to identify ;
1 24 them. I
() 25 We are committed to maintaining the program NEAL R. GROSS l COURT REPORTERS AND TRANSCRIBERS j 1323 RHODE ISLAND AVE., N W. l (202) 234-4433 WASHINGTON. D C. 20005-3701 (202) 234-4433 J
, 8 1 plan. And we expect that we will maintain it and wc will s 2 refine it with continued usage and experience. But we at )
(#l
'- 3 this time can't really justify using our limited 4 resources, diverting our limited resources from other -
t 5 work, to developing a more detailed road map.
6 l'd be happy to elaborate on any of those l 7 points if the Committee would like. ;
8 MR. CARROLL: Is there some routine interval 9 at which there is going to be a review of the plan?
10 MR. THOMAS: Yes, there is. As I reported at 11 the last -- I'll try to recall it from memory. At the 12 last ACRS Subcommittee meeting, I reported that the Human 13 Factors Coordinating Committee had been abolished because r'N s 14 it was redundant to the normal coordinating committee that 15 each of the offices hold every quarter. The members of 16 the coordinating committee were the same. But we did 17 commit to the Deputy EDO and received his approval to 18 continue those meetings and to maintain the plan.
19 The way that is to be accomplished is 20 consistent with the charter for the coordinating 21 committee. Every six months the branch chiefs of the 22 cognizant offices would meet and review the program plan 23 and consider the need to update it.
24 On alterr'te quarters, the meeting would be at
/"N
(_,) 25 the division level. And the division directors who signed NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N W.
(202) 234-4433 WASHINGTON, D C. 20005-3701 (202) 234-4433
~
~1 1
1 out Rev. 1 to the program plan, we would report to them.
,_ 2 And then they would approve revisions to the plan.
~
'_] 3 So we have committed to Mr. Milhoan, who has 4 agreed that this is the proper way to proceed. And it 5 really does exactly what the coordinating committee would 6 have done. It was just redundant.
1 i
MR. CARROLL:
7 What other activities does this 8 coordinating committee concern itself with?
9 MR. THOMAS: The coordinating committee has 10 existed for a number of years, primarily to discuss items 11 of mutual interest. One of the, perhaps the major, item 12 was research, user needs and the status of them. We did j l
13 attempt to discuss things like databases and so on, where p ]
k- 14 there was potential for duplication and overlap.
15 We think the program plan has gone long ways i 16 to document and codify the results of those discussions. ,
1 l
17 And now with the commitment to the EDO, we will certainly 18 do it on a quarterly basis.
19 CHAIRMAN APOSTOLAKIS: Question Number 2.
20 MR. THOMAS: Frank?
21 MR. COFFMAN: We split up the questions. So 22 we'll be bouncing back and forth.
23 Question Number 2 is: The activities ;
1 24 delineated in the Human Performance Program plan appear to O)
(, 25 be focused on reducing the assumed risk worth of human NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N W.
(202) 234-4433 WASHINGTON, D C. 20005-3701 (202) 234-4433 i
I lo l
1 actions used in PRAs. What is the risk worth of human j l
,s 2 actions? Why does the staff believe the risk worth is too N
3 high and should be reduced?
l 4 As I understand the concept of risk worth, the !
i 5 Human Performance Program plan was not really implemented 6 based upon the concept of risk worth but, in contrast, was 7 more built upon the feedback from experience and the 8 expertise it accrued from studying human performance over 9 time. ;
i l
10 There are several studies that allude to risk l
11 worth or address risk worth of human actions. There's an 12 old one where there was a Brookhaven sensitivity study 13 using the LaSalle plant and some other plants where they i
\
14 just looked at the changed core damage frequency as an 15 effect from changing the values for the human error )
i 16 probabilities rather uniformly to show its sensitivity. l 17 But more recently the --
18 MR. CARROLL: That was the study that Tom l
l 19 Morley used to --
20 MR. COFFMAN: Yes, sir.
21 MR. CARROLL: Yes.
22 MR. COFFMAN: More recently, though, the IPE l
l 23 draft report, NUREG-1560, lists some BWR and PWR scenarios 24 where human actions are important and lists them by r~N
!s s) 25 percentage of the plants at which they are important.
- NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 (202) 234-4433
11 1 It concludes in the final conclusions of the 7
2 draft report that human actions are important contributors 3 to core damage frequency in the IPEs. And correct 4 operator actions often significantly reduce core damage 5 frequencies. However, there is considerable uncertainty 6 associated with determining human error probabilities for 7 operator actions.
8 As a result, improved modeling of human 9 actions would significantly improve the understanding of 10 the risk associated with the operating of nuclear power 11 plants, which their conclusion I think is basically that 12 human actions add and subtract risk.
13 But the concept of risk worth requires that s
(
k-l' 14 human actions be converted into some risk measure. And 15 the methods by which human actions are converted into risk 16 measure has with it large sources of uncertainty -- I'm 17 sorry -- is a source itself of uncertainty in determining 18 the risk worth, particularly in the completeness of 19 modeling the errors of commission and cognition.
l 20 So it may be a little premature to judge the 21 risk worth of human actions prior to completing method 22 to convert the human actions into risk measures, out in 23 the meantime the Human Performance Program plan was really 24 built upon experience and the experience accrued over
(,3) 25 time, a study of human error.
NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N W (202) 234-4433 WASHINGTON, D C. 20005-3701 (202) 234 4 433
12 1 MR. ROSENTHAL: Let me just -- Jack Rosenthal. !
, 2 Just a few AEOD activities. In the accident sequence C> 3 precursor field, we're moving to a more sophisticated 4 ASP-based performance shaping factor. Human factor model, 5 the historical model, is quite crude.
6 We do have plans to, I mean, it's our goal to, 7 be able to go from operating experience that we observe to j 8 quantitative HRA numbers. And that's Items 225, 226, 227 9 in the plan.
10 So we definitely want as a goal to be able to 11 quantify human performance or HRA based on operating 12 experience, but that's not to say that in the questions 13 the risks were too high and should be reduced. That's not
()
l
(_/ 14 the current goal. The goal is to just understand what we 15 have.
16 CHAIRMAN APOSTOLAKIS: Now, risk worth does i
17 not necessarily mean quantitative numbers. You could look 16 at operating experience and using your own judgment and 19 that of a lot of other people decide that certain things 20 are happening that require study. Plus, again, even 21 without numbers or definitive numbers, you can look at l 22 some of the recovery actions, for instance, at PRA and 23 say, "Yes. This particular action is important. I don't
, 24 care what the number is, but it is important."
( ) 25 Has any of that thinking taken place when the NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N W.
(202) 234-4433 WASHINGTON, D C. 20005-3701 (202) 234-4433
13 1 plan was put together?
/
g 2 MR. COFFMAN: I think that, to my knowledge,
(
' ') 3 it was generally. But using the term " risk" more broadly, l 4 not in the focused use of the term, as it's at the NRC, 5 but using it broadly, yes.
6 I mean, certainly people are aware, the people 7 who work with human factors are aware, of the frequency of 8 occurrence of human contributors, human performance j 9 contributors, in the LERs.
10 There was an AEOD study Jack may want to 11 address about 16 events that addressed human performance.
12 We're continually looking at the databases for indications l 13 of relative contribution from different aspects of human j O.
k-) 14 performance procedures, the relative frequency of )
15 procedure problems, the relative frequency of i
16 communication problems. So there are other indicators in 17 the broader sense of the word " risk."
18 MR. ROSENTHAL: Yes. Item 228, " Routinely 19 analyze and use information in the human performance ,
20 databases." I think that's the point that you brought up. l l
21 Please note that they delete offices all, and 22 I think all of us are looking at the contribution of 23 procedures, of training, of management organization, 24 trying to figure out what to do with it. So all the i n
(,/ 25 groups are qualitatively looking at the data and trying to NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N W.
(202) 234-4433 WASHINGTON, D C. 20005-3701 (202) 234-4433
14 !
1 decide where to go. And that's a continuing activity I
g- 2 called out. I
,]
3 CHAIRMAN APOSTOLAKIS: That's 228?
4 MR. ROSENTHAL: Yes, sir. ,
l 5 CHAIP. MAN APOSTOLAKIS: So it says AEOD plans 6 to start generating new studies; right?
7 MR. ROSENTHAL: You'll hear more about that at 8 10:30, but the lead office is really all. l l
9 CHAIRMAN APOSTOLAKIS: So it says RES l
10 routinely analyzes SCSS and HFIS when initiating a 11 research project. What does that mean? That's Page 21. j i
l 12 MR. COFFMAN: Yes. That means that -- well, 13 there are two aspects o the answer. Let me give you the i l'~ ,
\--)' 14 general discussion first. And that is that routinely in 15 determining the need for a research project, we will, 16 either ourselves or ask a contractor, do sorts and i
i 17 searches through sequence coding and search system or 1
18 NRR's human factors information system.
19 An example might be the recent work on 20 communications was as a result of having determined that i
21 35 percent of the LERs involve communications contributors j
22 and, therefore, that it felt like that was an important 23 aspect to be addressed, see if there would be anything 24 that could ameliorate the problem. So it's just done as a r
~
I 25 routine. And I could pick other examples.
NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N W.
(202) 234-4433 WASHINGTON, D C. 20005-3701 (202) 234-4433
15 l
1 Do you want to cover some of this? I g- 2 MR. THOMAS: From an NRR point of view, I 3 would just show you some of the results of what we i
, 4 routinely do. We use the human factors information !
5 system, which is a database that is built upon review and 6 coding of inspection reports and LERs for human 7 performance information. I addressed this at the last 8 Subcommittee meeting.
9 We use the data on a plant basis to support i
10 insights for the senior management meeting screening !
11 process, screening meetings. Also, programmatically we 12 integrate the data and look at a breakdown of half a dozen i 1
13 areas that are contributing to human performance problems
(~N
\> 14 to make sure that we're focusing on the right areas.
15 (Slide) 16 MR. THOMAS: This slide is an example of LER 17 data over the past year or so that shows where the 18 problems really are. And you can see that most of the 4
19 events involving human performance have procedures as a 20 contributing factor, a strong contributing factor.
21 Communications and management and planning are other 22 strong contributing factors.
23 We have a similar slide for inspection reports 24 that really validates the slide, what we saw for LERs,
() 25 although the numbers may be slightly different.
NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N W.
(202) 234-4433 WASHINGTON, D C. 20005-3701 (202) 234-4433
16 1 This shows that we need to continue to update
,s 2 our inspection guidance and procedures to focus on 3 procedures and management and organization. And I think 4 from the other offices' point of view, this does direct .
l 5 our discussions and our focus and direction.
6 MR. CARROLL: I've had the impression over the 7 years that everybody thinks we ought to do something about 8 LERs so you get a better picture of really how human 9 performance entered into the event. Are there still plans 10 to improve the LER reporting system? l 11 I think it's very spotty. I think some 12 utilities probably provide pretty good human factors 13 information and others, every problem is a mechanical j
('~h <
\-) 14 problem. l 15 MR. THOMAS: We do the best we can. To some j I
16 extent, you're right, Jay. It's like trying to squeeze 17 blood out of a turnip.
i l
18 We do the very best we can. We have 19 contracts, NRR. We have contracts with Oak Ridge to do 20 the LER coding. And we do our own inspection report 21 coding. We keep trying. And I'll speak for NRR and ask 22 Jack especially to address it from AEOD's point of view.
23 Ne continue to support that. We'll do everything we can.
24 I would just mention from an international
/~
'( ,T/ 25 point of view we are a member of the NEA Principal Working NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N W (202) 234-4433 WASHINGTON, D C. 20005-3701 (202) 234-4433
17 1 Group Number 1 expanded task force in Paris. And we have
,_s 2 just this past year taken a step that we believe would be
( I
\~ 3 successful for incorporating some fundamental amount of 4 human performance information into reports, IRS reports.
5 And hopefully that will fly, and hopefully that will serve 6 as an impetus for some work in a sir.ilnr thing here.
7 Jack, do you want to talk about where you're 8 going on that?
l 9 MR. ROSENTHAL: Yes. We --
10 MR. CARROLL: I guess the thing that troubles 11 me is that you don't seem to be making a lot of progress 12 --
13 MR. ROSENTHAL: We're not -- )
,2 \
k_ 14 MR. CARROLL: -- changing the LER system to 15 really include this service, though.
16 MR. ROSENTHAL: Right. The progression would 17 be as follows. We built a human performance-based 18 database based on the AITs and IITs and er.tracted 19 information for the corresponding LER. You'll hear more 20 about that at 10:30.
21 Then you have to study that database. And 22 then you have to conclude that: Is there or isn't there a 23 problem in the sense that: Is the agency getting enough 24 information for its needs?
r~
(s) 25 MR. CARROLL: Yes because the AITs and IITs NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE IS'RJD AVE , N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 (202) 234-4433
I 18 )
1 1
1 are things that went south. LERs, many of them, are 7_s 2 things that had the potential for a serious problem. And
(' ') 3 something prevented it from happening. But I don't think 4 you're really capturing what's going on in those kinds of I
5 events with the present system.
I 6 MR. ROSENTHAL: Yes. )
I 7 MR. ROSSI: This is Ernie Rossi. I'm the !
I 8 Division Director of the Safety Programs Division in AEOD.
l 9 We have had for some period of time an effort l i
10 in AEOD to revise NUREG-1022 and take a look at the 11 reporting rules themselves. And, as you I'm sure are )
l 12 aware, we currently have an effort underway to collect i 13 reliability and availability data in a very consistent
/~%
14 way.
15 So at some point in time, we're going to try j 16 to integrate all the things that are going on into looking 17 at either NUREG-1022, which, of course, is the guidance 18 for the LER, or the LER rules themselves.
19 And when we do do that, we'll include all the 20 things that are going on in the agency. That would be 21 changes as a result of getting reliability, availability 22 data, which we expect to eventually get. That may mean 23 that we need different or less kinds of things reported in 24 the LERs.
(~%
( ,) 25 And at that point in time, we'll be looking at NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N W.
(202) 234-4433 WASHINGTON, D.C, 20005-3701 (202) 234-4433
l 19 1 what, if anything, needs to be done with respect to g~-
2 getting better reports on performance. l
) 3 MR. CARROLL: I would certainly encourage 4 moving anead with that because, as with anything, data is l l
5 what you make decisions on. I 6 MR. ROSENTHAL: Item 225 is revise the 7 NUREG-1022 guidance. And that will involve a backfit in 8 all likelihood or a change from state of practice. So we !
l 9 have to analyze the data that we have and then draw a
- 10 conclusion. Do we ne or do we not need more 11 information? And let's not prejudge it.
12 But the progression would be analyze what we i
13 have, then attempt to revise NUREG-1022 based on some i OI
's 14 factual determination as Item 225. And beyond that would I
15 be to revise the rule itself.
16 MR. CARROLL: I would point out, Jack, under i 1 17 " Schedule" in 225, TBD. I 18 MR. ROSENTHAL: That's right. It depends on I i
19 the database. i 20 MR. CARROLL: I would feel much more 21 comfortable if I saw a date in there.
22 MR. ROSSI: Well, a number of the things 23 depend on other things getting completed, but the big 24 driver is resource allocation as to where we best use our q
q,) 25 resources at any given time. So that's part of the reason NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N W.
(202) 234-4433 WASHINGTON, D C. 20005-3701 (202) 234-4433
20 1 why the dates at this point in time are not there.
2 MR. CARROLL: And the priority is medium?
7-(~/ 3 MR. ROSSI: Right.
4 MR. CARROLL: I guess I would suggest that 5 maybe the committee might want to say something about this 6 in a letter because if you don't have the data, you've got 7 nothing. You're just speculating. Of course, you've got 8 some data, but the LERs are a real source of information 9 that isn't really being done. Okay.
10 CHAIRMAN APOSTOLAKIS: Let me interrupt this.
11 Can we all agree or would you please give us a definition 12 of human factors? When we say " human factors," what do we 13 mean?
7 (b) 14 MR. COFFMAN: I can give you a start on it. I 15 guess everybody has their working definition. But it's 16 the technology of the use, the safe and effective use, of 17 machines by personnel, I mean, just generally. Is that 18 what you're -- .
l 19 CHAIRMAN APOSTOLAKIS: No. I'm not testing 20 you. I'm just trying to --
21 MR. COFFMAN: No. I meant is that the level 22 which you were looking for? l 23 CHAIRMAN APOSTOLAKIS: Specifically, when we l
24 say " human factors," do we include in that human error?
r
/s n MR. COFFMAN:
V 25 Yes. ;
NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N W.
(202) 234-4433 WASHINGTON, D C, 20005-3701 (202) 234-4433 l
21 We do?
1 CHAIRMAN APOSTOLAKIS:
7- 2 MR. COFFMAN: Yes.
i
(
3 CHAIRMAN APOSTOLAKIS: So anything that 4 involves humans, in other words, falls under human 5 factors?
6 MR. COFFMAN: Yes.
7 MEMBER MILLER: Can you define human error, 8 then?
9 MR. COFFMAN: I'm scratching my head. I'm 10 sure that lots of folks can give you alternative 11 definitions that may be variations on the theme, but the 12 purpose of the system is that a system has an intended 13 function. And it really is not a system until you get the
\_/ 14 machine and the operator and maintainers together. I 15 mean, then it becomes a system.
16 When the functioning of the operator or the 17 humans within that system goes against the basic intended 18 function or not the basic, but intended, functions, then
< 19 you've got a deviation from intention. And I would call 20 that an error.
21 CHAIRMAN APOSTOLAKIS: Now, of course, as you 22 probably know, that implies that the poor human was 23 responsible for the error. And sometimes that's not the 24 case. So people now are using words like " unsafe act."
( 25 In other words, the human committed an unsafe act without NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N W.
(202) 2344433 WASHINGTON, D C. 20005-3701 (202) 234-4433
22 1 necessarily doing anything wrong. I mean, if the
, 2 indications were wrong, for instance, the human acted N
3 correctly.
4 So people are beginning to see these things 5 now because we used to blame the operators for anything 6 that went wrong. Then we realized that it's not the poor 7 guy's fault many times, that the context really leads them 8 to make these mistakes. But basically that's a correct 9 description.
10 Okay. So human factors is a broad, a way of 11 using it in a broad, sense. And that includes everything 12 from the design of control room all the way to human error 13 or unsafe acts.
V 14 MEMBER SEALE: It includes cognitive errors as l 1
15 well as functional errors.
16 CHAIRMAN APOSTOLAKIS: The whole works, the 17 whole works.
18 MEMBER SEALE: Yes.
19 CHAIRMAN APOSTOLAKIS: Okay. We can go back 20 to the questions, then. I guess Number 3. Did we cover i 21 that? No.
22 MR. THOMAS: No. Numbe.- 3 is: How does the 23 staff set priorities for the Human Performance Program 24 plan activities? And what does the priority ranking mean?
!Q 1 Q 25 We addressed priorities in the initial Human NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N W.
(202) 234-4433 WASHINGTON, D C. 20005-3701 (202) 234-4433 i
23 1 Performance Program plan. And I'll read you the couple of
,, s 2 lines that deal with that. "Each program activity was
( )
V 3 assigned a priority, high, medium, or low, by the 4 coordinating committee. The priorities represent the 5 coordinating committee's qualitative assessment of the 6 activity's relative importance, taking into consideration 7 their perceived safety significance, immediacy, management 8 direction, and regulatory need."
9 And they really are no more than that. They 10 were, in effect, a collaborative vote of the coordinating 11 committee members based on their own perception of those 12 factors. And the idea was to quite honestly, separate the 13 activities, to distinguish one from another in relative
\_) 14 importance, as opposed to try to align them with, for 15 example, NRR's priority-ranking scheme.
16 MEMBER KRESS: Does the priority ranking, 17 then, get fed back into the amount of resources put into 18 each of these item and the amount of attention it gets?
19 MR. THOMAS: Ideally, yes. I can't say in 20 every case it does, but that was the intent.
21 MEMBER KRESS: So that's what the rankings 22 mean? It's how you're going to allocate your resources 23 and scheduling and --
24 MR. THOMAS: That was the intent, f~%
i ) 25 MEMBER SEALE: One of the terminology tricks,
\_/
NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N W.
(202) 234-4433 WASHINGTON, D C. 20005-3701 (202) 234-4433
24 1 if you will, or ideas that seems to be kind of pervasive
,- 2 is the concept of ownership, whether it's the ownership of 3 a system within the plant by some plant engineer, perhaps 4 more recently the concept of ownership of the design basis 5 for a plant and so on.
6 When you established a priority for a 7 particular issue, did you then try to assign ownership for 8 that issue to someone within your organization? In other 9 words, is there an advocate for the high priority issues 10 or the medium priority issues or whatever that you have 11 identified within your organization so that there is 12 supposedly someone who is pushing it?
13 MR. THOMAS: Yes. The program plan identifies A
i 1
\~/ 14 the responsible office. And the responsible office and 15 someone within that suborganization or individual within 16 that office would have responsibility for advocating and 17 following up on it.
18 MEMBER SEALE: Okay.
19 MR. CARROLL: You limited this to high and 20 medium. How about low?
21 MEMBER SEALE: Yes.
22 MR. THOMAS: There really weren't. I guess 23 there was -- I think there were, what, two lows?
24 MR. COFFMAN: I think two.
m
\ '
25 MR. THOMAS: The lows are NRR's. I'm assuming NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N W.
(202) 234-4433 WASHINGTON. D.C. 20005-3701 (202) 234-4433
25 1 responsibility for them in their relative ranking.
l 2 MR. CARROLL: Does low mean the same thing as
~'
3 it does with generic issues, --
4 MR. THOMAS: No.
5 MR. CARROLL: -- the equivalent of drop?
6 MR. THOMAS: No.
7 MR. CARROLL: Okay. I noticed one of those 8 lows is one of my favorite subjects, 1.3.2, method to l
l 9 quantify organizational performance factors.
10 MR. THOMAS: Frank?
l 11 MR. COFFMAN: We come to this in Question 8. I 12 MR. CARROLL: Yes, I know you do. I just -- !
l 13 i MEMBER SEALE: He's doing a little bit of l
/'N l 14 baiting. i 15 CHAIRMAN APOSTOLAKIS: I understand how the 16 coordinating committee used its judgment, and I think 17 that's inevitable. But, as you know, a big thing 18 happening now in this agency is a move towards 19 risk-informed regulation. So I wonder whether this would 20 be a place also where these prioritization would be 21 risk-informed. And what would that mean in this context?
22 Risk-informed regulation is not only for one 23 part of the agency.
24 MR. THOMAS: I would like to say I wish one of
(-
\j 25 the considerations was risk. Maybe to some degree it was.
NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.
l (202) 234-4433 WASHINGTON. D C. 20005 3701 (202) 234-4433 I
l
26 1 It wasn't enough that we could call it out as a 73 2 consideration and be honest about it.
)
\~/ 3 We're not there yet, I think as Jack described 4 earlier. I wish we were. If we were closer and could do 5 more, we would. Then it would be a factor.
6 CHAIRMAN APOSTOLAKIS: Now, you said at the 7 beginning that the committee does not exist anymore; 8 right?
9 MR. THOMAS: Correct. As an ad hoc committee, 10 it doesn't. However, the members -- effectively, it still 11 does.
12 CHAIRMAN APOSTOLAKIS: The members were not 13 abolished, evaporated? The members still exist?
p#
k> 14 MR. THOMAS: Not yet anyway.
15 CHAIRMAN APOSTOLAKIS: Okay.
16 MR. THOMAS: In fact, you're looking at three 17 of them right now.
18 CHAIRMAN APOSTOLAKIS: What is it that's 19 stopping you or the agency from revisiting this plan and 20 bringing it up to date and respond to some of our !
I i
21 questions and making it more rick-informed? I mean, what i 22 is --
23 MR. THOMAS: Nothing is in the way. And as a l
24 matter of fact, as I perhaps didn't make it clear enough,
() 25 every six months we get together. And we consider our own NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N W.
(202) 234-4433 WASHINGTON, D C. 20005-3701 (202) 234-4433
27 1 office's needs. We consider the needs that others bring i
2 to us, including the committee. And we'll modify it or t r l 3 recommend its modification and reissuance on a six-month 1
4 basis.
l l
5 CHAIRMAN APOSTOLAKIS: I think we have a real j i
6 problem with not just -- when I say " problem," not with !
7 you, as a community, with this risk-informed business, the ,
i 1
8 relevance to risk. And it comes back to the second 9 question as well. It should address the question of risk 10 worth of human actions.
11 It can work both ways. You can miss something 12 that's important to risk or you can exaggerate the 13 importance of something simply because you don't have any
( )
b 14 way of evaluating its risk worth.
1 j
l 15 And I think that's really a pressing need l l
16 right now. And I'll come back to this speech by Mr. Pate, 1
l 17 Dr. Pate. '
18 MEMBER SEALE: Admiral Pate.
l 19 CHAIRMAN APOSTOLAKIS: He has a Ph.D. from i
20 MIT. So that takes precedence over -- l 21 MEMBER SEALE: Okay. Okay.
22 (Laughter.)
23 CHAIRMAN APOSTOLAKIS: So that speech where he 24 lists, I believe, eight actual incidents and created quite n
Q 25 a stir, as you know, -- people said, "Oh, gee," this and NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N W.
l (202) 234-4433 WASHINGTON. D C. 20005-3701 (202) 234 4433
l 28 l
1 that -- I mean, if you sit back and look at them, it's not j
- 2 obvious to me what the risk significance of all of these t l 3 things was and whether they were significant at all.
4 I think the significance of the speech is that 1
5 such a high-level person attempted to talk about these 6 things and gave specific incidents. So he brought the l
7 attention of very senior people to that issue. But l 1
8 lacking a framework, I'm not sure that all of these really l 9 were significant from the risk perspective.
l 10 So it can work either way. You can think that 11 something is not important when it is or you can think of )
i 12 something as being very important when, in fact, it may 13 not be.
O (l 14 So I think that's what's behind this 15 risk-informed move to try to place things in perspective. I 16 And I fully appreciate the problem that Frank raised 17 earlier that for some of these things, we don't know how 18 to quantify them.
19 But I do believe we have a framework. At 20 least we have a framework where we can make some informed 21 judgments. And I really think that's very key here. It's 22 very important. It's very important, I think, to try to 23 develop some criteria that addressed risk. And in your 24 case, I mean the human factors case, a lot of them would
/"\
( ) 25 have to be qualitative simply because we don't have the NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 (202) 234-4433
1 29 1 numbers.
2 But then, at least if we saw that list of
/-ws
\ )
3 criteria, we would be able to say -- and not just data; I 4 don't mean just data -- "Yes, this is important," "Yes, 5 this is not important." And I think the community, in l
6 fact, needs that.
7 I was in Berlin this past weekend at a human 8 factors workshop, I guess. People presented their work.
9 And at the end, you really didn't know what was important I
10 and what wasn't because there was no attempt by anybody to 1 I
11 prioritize. Maybe that notion is foreign to that 12 community.
13 MR. CARROLL: Prioritized based on?
7",
'V) 14 CHAIRMAN APOSTOLAKIS: Based on something. I j 15 don't know what that something is, but on something. So, I i
16 you know, somebody talked about leadership. What does 17 leadership mean and at what level? Is leadership the 18 same? Does it have the same attributes when you talk 19 about the vice president and when you talk about the head 20 of a small group that fixes things when they break?
21 So leadership was the subject for 20 minutes.
22 Then we moved on to something else: man-machine 23 interface. Okay. But at the end of the two and a half 24 days, I myself didn't know what was important and what 25 wasn't.
NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N W.
(202) 234-4433 WASHINGTON. D C. 20005-3701 (202) 234-4433 i
30 1 1 So it's not really that everybody is doing it g~s 2 and you guys are not doing it. I mean, it's really a ,
(' ')
3 problem I think in this area not having criteria by which 4 we can prioritize.
5 Now, in our case, being a regulatory agency 6 and for protection of public health and safety being our 7 goal, the criteria have to be safety-related; right? And 8 since safety now is measured in terms of risk, I mean, the 9 conclusion is inevitable.
10 So I would really like to see something along 11 these lines going beyond just the judgment of the 12 Committee or maybe it will be the judgment of the 13 Committee but now you will have a well-defined list of
(~'y
's /
14 criteria to guide the Committee to make those judgments 15 because otherwise I don't know.
16 I mean, I read Dr. Pate's speech, and I say, 17 " Gee, that's terrible." And then I say, "Well, why is it 18 terrible?" I don't know that's terrible.
l 19 MR. CARROLL: I agree with you, George. I had 20 exactly the same reaction to Admiral Pate's points. Those 21 items were a mixed bag.
22 CHAIRMAN APOSTOLAKIS: Yes, that's right.
23 MEMBER SEALE: I think his main focus was the i
24 question of the management culture within the utility and i
(~%s,) 25 the apparent disinterest in what happened in the routine j NEAL R. GROSS I COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE,, N W.
l 0 (202) 234-4433 WASHINGTON. D.C. 20005-3701 (202) 234-4433
31 1 operation of the plant.
2 CHAIRMAN APOSTOLAKIS: I fully agree.
t 1
~~'
3 MR. CARROLL: The way he got people's 4 attention.
5 MEMBER SEALE: Yes, right.
6 CHAIRMAN APOSTOLAKIS: And I think he did a 7 good job.
8 MEMBER SEALE: But I have a. question. Earlier 9 we heard discussion on mention of the fact that the whole 10 question of the requirements for LER information and so on 11 were being examined with the idea of trying to make the 12 LERs more useful and so on. Included in that, is there 13 consideration of the things that would be perhaps helpful
(~h t I L 14 in trying to get a better handle on the risk that's 15 actually involved in the events or from the events that 16 are embedded in the LER? I mean, is there a risk measure 17 or --
18 MR. ROSSI: Well, certainly when we look at 19 the things that are required of licensees to be reported 20 to the agency, risk, at least in a qualitative sense, is a 21 part of it. I mean, when we look at perhaps safety system 22 failures need to be reported to the agency, we would look 23 at the risk of those safety systems.
24 I don't know that we have developed the I
25 U mechanism again for the human factors area to the degree NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS
'l323 RHODE ISLAND AVE., N W.
(262) 234-4433 WASHINGTON, D C. 20005-3701 (202) 234-4433
32 1 where we can look at that, but I think that Cecil will put
_s 2 that chart up there again. And I think you were probably t
3 going to say some words about that being some indication l
I 4 of the things that we're seeing in LERs that are important I 5 in the human factors area or where human factors has 6 contributed to it. ;
7 MEMBER SEALE: My thought was more --
8 MR. CARROLL: Maybe.
9 MEMBER SEALE: Yes. Maybe. My thought was 10 more along the idea that so often it seems like we get 85 l 11 percent of the bundle. And then we're stuck with 12 inferences of what it takes to complete the picture in l 13 order to really make a risk evaluation or an assessment.
1
(~h kl m 14 When you're looking at your new ensemble of 15 what constitutes completeness in an LER, are you trying to l l
l 16 be sure you get everything in so you don't wind up with an 17 85 percent bundle? i 18 MR. ROSSI: Well, I'd be a little cautious on 19 that because one of the things that you have to be careful j
l 20 of is that you don't want to ask for ten percent more than l
21 of things that don't have a high importance to the agency 22 because there's some trade-off of costs and benefits.
23 MEMBER SEALE: Sure. That's true.
24 MR. ROSSI: So I think the answer to your A
(%- ) 25 question is a somewhat qualified --
NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N W.
(202) 234 4433 WASHINGTON, D C. 20005-3701 (202) 234-4433
33 1 MEMBER SEALE: We hope so.
, 2 MR. ROSSI: -- yes, we will try to get the
(" )
3 important things, however we define those, reported to the 4 ejency. The ones that are not important, we would try not 5 to ask for those because there's not a cost benefit in 6 doing so.
7 MEMBER SEALE: I appreciate that problem.
8 MR. CARROLL: Well, Bob, of course, the LERs, 9 among other things, are screened. And that's input into 10 the accident sequence precursor program.
11 MEMBER SEALE: Sure, yes.
12 MR. CARROLL: And when something triggers, you 13 don't just have more than the LER. You review --
\
(V 14 MR. ROSENTHAL: We may have an AIT or an IIT 15 or a special inspection or something that a special human 16 factor --
17 MR. CARROLL: Or you may go to a utility and 18 get some more information.
19 MR. ROSENTHAL: That is a good. We have --
20 CHAIRMAN APOSTOLAKIS: Wait, wait. Yes? The 21 mike.
22 MR. ROSENTHAL: I'm sorry. And we routinely 23 do that on a daily basis with NRR and AEOD both 24 participating in morning calls and decide when it's O
(j 25 appropriate to investigate an event more. And that gets NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N W.
(202) 2344433 WASHINGTON. D C. 20005-3701 (202) 234-4433
~
34 1 coded in.
f~ 2 The current LER rule does require reporting of
( )
'~'
3 human factor information. It's like one sentence long. l l
l 4 It is there. I 5 MR. CARROLL: And what people typically do is 6 say, "The operator screwed up. And we have instituted 7 some training to fulfill this." That may not even come 8 close to what the problem is.
9 MR. ROSENTHAL: Right, right. !
10 MR. THOMAS: But, if I may chime in, we do not 11 just limit our look, at least in NRR, and I know Jack 12 doesn't either to LERs. We have done an -- the agency as 13 a whole has done a number of 40-500 inspections, root
(
. x- 14 cause, licensee self-assessment. And we do go down 15 several layers beyond that and look to determine really 16 what the root cause is. There's also a great deal of 17 information from other inspection reports that are rich
, 18 relative to LERs.
19 The point I wanted to make, Mr. Chairman, when 20 I put this up is absent the risk-informed information. In 21 NRR, where we're really practitioners of what comes out of 22 Research and information from AEOD by doing inspections 23 and reviews and so on, we're guided programmatically by 24 what we have here, the results of our information system.
3 s_j 25 I know you can slice the pie as many ways as NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHoDE ISLAND AVE., N W.
(202) 234-4433 WASHINGTON, D C. 20005-3701 (202) 234-4433
35 1 there are people. These are basically the six areas that
,x 2 we look in in human performance. And the program plan, as
/ 1 3 far as my vote on priority, was guided by this.
4 For example, procedures are very important.
5 Training is very important. And management is very 6 important, communications also. And you'll find that most 7 of the elements in bere in those areas do reflect the i
1 8 higher priorities.
9 I understand, for example, that it's hard to 10 quantify or even to talk about the risk-effective 11 management because that's all pervasive; whereas, 12 procedures, you can be a little more specific about it.
13 But I just wanted to let you know that, at
( } l
\/ 14 least as far as NRR's vote goes, there was some rhyme or 15 reason as a surrogate for risk-informed numbers. And I l 16 think the other offices are guided by this as well.
17 CHAIRMAN APOSTOLAKIS: And I never disputed 18 that. All I'm saying is that the next step now would be 19 to go to this risk-informed framework. And that I think 20 will build upon this.
21 MR. THOMAS: We wish we could.
22 CHAIRMAN APOSTOLAKIS: I'm not talking about 23 an entirely different approach. It's just that things 24 evolve in time, and this I'm sure has served its purpose.
m
.) 25 In fact, what I see there makes perfect sense to me. I NEAL R. GROSS COURT REPORTERS AND 1RANSCRIBERS 1323 RHODE ISLAND AVE., N W.
(202) 234-4433 WASHINGTON, D C. 20005-3701 (202) 234-4433
36 1 mean, things evolve.
s 2 MR. THOMAS: Sure.
I b 3 CHAIRMAN APOSTOLAKIS: We have been doing 4 regulation a certain way now for 30 years. And now we are 5 about to release, as you probably know, a number of reg 6 guides and new sections to the SRP. So things change; 7 right?
8 And I think we could -- incidentally, I just 9 remembered -- maybe that belongs to the discussion, but 10 since I remembered it, I'll mention it.
11 Professor Reason gave a very good talk at that 12 workshop where he argued that perhaps we are overdoing it.
13 They also started realizing they need some criteria for
/~'s
\ l s/ 14 limiting research. He claimed that people have begun wide 15 with this sociotechnical system and studying management 16 and organizations and talking about societal issues and he 17 says well, maybe we should limit ourselves to what we can 18 do something about and we know we cannot change society at 19 large, so why talk about it ad infinitum?
20 Now a lot of people felt that that was the 21 wrong approach, for example, deregulation is a big issue 22 which will affect human factors and so on, but what was 23 interesting to me was that one of the leaders in this 24 field of human error felt that the researcher had gone too O
(_,) 25 far in certain areas, simply because people identified l NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE, N W.
(202) 234 4433 WASH:NGTON. D.C. 20005-3701 (202) 2344433 l
37 1 problems they could do nothing about.
,s 2 If society is in a certain way, society is in
/ i !
i
\ ,)
3 a certain way and we're not going to change it. So I l 4 think we do need criteria to focus the research and now 1
5 again, not everybody agrees with what he said, by the way, 6 but that's the nature of the thing. I don't think very 7 many people agreed on anything.
8 It was very interesting that the leader in the 9 field felt that we had gone too far.
10 MR. ROSENTHAL: If you accept that the ASP .
l 11 events that score high are risk important or AITs and IITs 1 I
1 12 which don't necessarily always score high in ASP space are ;
1 13 the most important events. At least to our goal to take k- 14 those events or those ASP cases and say what were the 15 contributing factors to those and as it turns out, in all 16 the high scorers, human performance is important, the I
17 Davis-Besse, the most recent Vogtle event, etcetera. So 18 you go down a step farther and you say okay, what affected 19 those events?
20 One of the blocks is management and l
21 organization. And not just a block like that, but you say l
22 what aspect, what specific aspects of management and l
l 23 organization affected the course of that risk important 24 event and that's where we would intend to work.
( ,s) 25 To that extent, that's not all of society.
f l
NEAL R. GROSS l COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N W.
(
(202) 234 4433 WASHINGTON. D C. 20005-3701 (202) 234-4433 l
38 1 CHAIRMAN APOSTOLAKIS: No, I'm not saying 2 ; hat.
!' )
'/ 3 MR. ROSENTHAL: It's those things that drove 4 real important events.
5 CHAIRMAN APORTOLAKIS: Yes.
6 MR. ROSENTHAL: At least that's the direction 7 that we're going. Now are you suggesting that we get 8 quantitative or at least be able to draw that sort of road 9 map? ,
10 CHAIRMAN APOSTOLAKIS: I think 3 went beyond 11 what I was trying to say. I was just mentioning what 12 Reason said, but we will come to the details of what you 13 just gave. An example he gave was this fellow who G 14 committed suicide after Chernobyl and he left a note and 15 he said Chernobyl was a natural consequence of our Soviet 16 system and Reason said, "how does that help me? I can't 17 do anything about the Soviet system." That's too much.
18 But the other things you mentioned, I think, are very 19 reasonable.
20 MR. ROSENTHAL: I'll give a specific concept 21 and that is if management has scheduled too much work to 22 be done on a shift during a change in mode, that is j 23 something that this Agency can address.
24 CHAIRMAN APOSTOLAKIS: Exactly. Exactly.
'O Q 25 Exactly.
i NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N W.
(202) 234 4433 WASHINGTON. D C. 20005-3701 (202) 234-4433
39 1 MR. CARROLL: Having said that, Jack, how come
- 2 item 1.3.2 is low priority?
%'/ 3 CHAIRMAN APOSTOLAKIS: 1.3.2?
4 MR. CARROLL: Develop methods to quantify 5 organizational performance process.
6 MR. COFFMAN: Let me, yes, we're coming to 7 Question 8 again. We may as well address it. I would 8 like to address Dr. Reason's comment because it comes to a 9 very practical question for us in research and I think 10 part of becoming a professional in an area is knowing when 11 to stop and that's when doing a root cause analysis you 12 can go too far. You can go into those areas where you 13 can't reach, you can't change personalities, you can't kI 14 change culture, at least very readily. So in the 15 development of the human performance investigation process j 16 guidance, we were guided by the principle that the root 17 cause we stopped in determining the root cause at the s
18 point where it could be fixed. Just a very practical 19 limitation, but there is that consideration that was done 20 in research and I don't think we're too guilty at the NRC 21 of going too far into some of these vaguer issues.
22 MR. CARROLL: Particularly in an area --
23 CHAIRMAN APOSTOLAKIS: That's Question 8.
24 (Laughter.)
p) y, 25 MEMBER SEALE: Can we go on?
NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISMND AW. N W (202) 234-4433 WASHINGTON, D C. 20005-3701 (202) 234-4433
40 1 CHAIRMAli APOSTOLAKIS: Number 4?
l 73 2 MR. THOMAS: Number 4, how does the staff l (s ,
'~'
3 decide that an independent program element is required and l l
4 why has the staff decided that data gathering should be 5 separated from developing guidance and that the two 6 activities should have different priorities.
7 I'll answer the first one, give an example of 8 the second one and I think Jack has another example of the 9 second one.
10 It was just judgment on what should constitute 11 a program element. It's just a judgment, a manageable, 12 segregatable block of activities and in some cases it i
13 corresponded to a user need, other cases it was a program I
[h
\
\- 14 of activities such as oversight of INPO accreditation 15 process. There is really nothing magic and no established 16 criteria, it was just whatever was a block that could 17 stand on its own and be described and assigned a 18 responsible organization and have a meaningful schedule 19 attributed to it.
20 The second part, I'll generalize the question 21 as I understand it. There are some apparent 22 inconsistencies and priorities for apparently related 23 activities. The one example that I will address involving 24 NRR and research are, and I think these were raised during (3
( ,)
s 25 the last subcommittee meeting, any apparent inconsistency NEAL R. GROSS i
COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D C. 20005-3701 (202) 234-4433 l
41 1 between items 1.2.2 and 1.2.6. 1.2.2 says, has to do with s 2 conducting reviews of control room modifications and 1.2.6
\ l i s~/
3 has to do with developing guidance for hybrid control 4 rooms. The first was given a medium priority and the 5 second a high priority.
6 The reasons are that conducting reviews of 7 control room modifications is a task that the Human 8 Factors Assessment Branch and NRR has in conjunction with 9 the INC branch. When utilities decide to do a digital 10 retrofit of an analog system for whatever reason, that it 11 requires prior NRC approval before they can implement it.
12 If there are human factors considerations, my branch 13 becomes involved and reviews those considerations, namely t
r'N 4
\~/ 14 things like man-machine interface and any time for 15 operator action and so on.
16 By definition, this is a priority 2 item in 17 the NRR prioritization scheme. And so there aren't, first 18 of all, there aren't very many. Secondly, it's not the 19 highest priority and the priority that the Committee gave l
20 it is consistent.
21 Contrast that with 1.2.6, development of 22 guidance for hybrid control rooms. That's a research tack 23 that -- whose purpose is to really see if there's an 24 underlying problem of having both hybrid and analog
() 25 systems and in converting from one to the other and if so, f NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N W.
(202) 234 4433 WASHINGTON, D C. 20005-3701 (202) 234-4433 i
I
42 1 to develop appropriate regulatory guidance. Because of s 2 the potential safety significance of that question, it's
[G \
3 important to find out the answer as soon as possible 4 because there's a long lead time to developing the 5 regulatory guidance.
6 So we assigned that high priority just to find 7 out the answer and if the answer is no, it's not a high 8 safety significance, we would probably revise the priority 9 downward. Until we know that answer, it's important to ,
10 find out and hence the high. So that's the reason for the 11 apparent disconnect of those two items.
12 I think Jack, you had an example also?
13 MR. ROSENTHAL: Dana Powers raised the issue O
k >I 14 of 2.2.6 and 2.2.7 where 2.2.6 is medium and 2.2.7 is 15 high. As I sit here and read it with a few months from 16 when it was written, I can see why it would be confusing.
17 2.2.6 is determine information needs for HRA 18 and 2.2.7 is assess HRA models to assure human performance 19 data bases contain information useful for HRA models.
20 It was simply a sequence. It will just take 21 me a moment to get through it and that is let's take the 22 data that we now have, put it into data bases as a means 23 that support HRA and attempt to do HRA using operating 24 experience data, priority high.
I s 25 Should it turn out that we believe that you NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N W.
(202) 234-4433 WASHINGTON, D C. 20005-3701 (202) 234-4433
43 1 can support the current HRA models with the data that you 2 have, then we're going to have to revise the information 7~i r
\/
3 that we collect, support the HRA models or change the HRA 4 models to be consistent with the level of information that 5 you have. And that was ranked medium because we just had 6 to do the first task, assess what we had before we went 7 out and started to change the world. So that was the high 8 and medium.
1 9 CHAIRMAN APOSTOLAKIS: Now there is also a 10 note attached to 4. Are you going to address that?
11 MEMBER SEALE: Actually, it has to do with 3. l l
12 MR. THOMAS: We interpreted the note to relate !
13 more to number 1.
p-k..) 14 CHAIRMAN APOSTOLAKIS: Yes.
15 MR. THOMAS: On models, human performance 16 models and I hoped I addressed it when I answered one, 17 that we really only planned to at this time maintain a 18 program plan and essentially in its present form.
19 CHAIRMAN APOSTOLAKIS: Okay, let's move on to 20 5 then.
21 MR. THOMAS: Number 5, what does the staff 22 mean by effective and adequate as used in the objectives 23 and goals of the program plan? How does the staff know 24 what must be done and when the goal or objective is
( 25 achieved?
NEAL R. GROSS 4
COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N W.
(202) 2344433 WASHINGTON, D C. 20005-3701 (202) 234-4433
44 1 By use of the terms " effective" and " adequate" f3 2 they have no special meaning an/ as a matter of fact, they b 3 could even be eliminated.
4 How do we know when the goals and objectives 5 have been met? When all the elements under the goals and ,
I 6 objectives have been accomplished. But recall, it's a I 1
7 living document. There will probably be elements added ;
8 and deleted with time, so it's -- I don't envision a goal 9 or objective ever being met.
, 10 Number 6, should the staff be pushing 11 licensees towards the state of the art in human factors 12 and human reliability, rather than approve an adequate 13 state? I think the answer to that is you have to look at l' )
\- 14 it in two parts: current (Oerating plants and future 15 advanced reactor plants. For the current operating 16 plants, there's certainly no requirement to be at the 17 state of the art. We think it's a good idea, both for the 18 licensees and for the staff to be aware of the state of 19 the art and human performance and human factors and 20 reliability and risk and so on. Just so they know. Just 21 so they know what's -- the licensees know what's available 22 and so we can be in a position to evaluate it and provide 23 guidance. Because the answer is different for advance 24 reactors. Actually, the post-TMI requirement for control
,e
( ,) 25 rooms in 10 CFR 50.34 (f) which was for advanced reactors NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N W.
(202) 234-4433 WASHINGTON D.C. 20005-3701 (202) 234-4433
45 )
i 1 was referenced in part 52, does require a state of the art 2 human control room, a control room that quote reflects
(,l
%J state of the art human factors principles. So it is 3
4 anticipated that for advanced reactor designs, that they i
5 do reflect the then state of the art principles. That's 6 one of the reasons that our approach to reviewing advanced 7 reactors is as it is. That is, we have a program review 8 model. It's a process as opposed to specific criterion 9 requirements that would lock in technology because the 10 design certifications are granted for a certain length of 11 time, quite a few years actually and this is a rapidly 12 moving technology and we don't want to have these designs So we go with the method or process. I 13 locked in. !
(~'s
- 14 CHAIRMAN APOSTOLAKIS: Now how does the staff 15 keep up to date with state of the art?
16 MR. THOMAS: Many ways. We interact on a 1
17 daily basis with, among offices, let's start at that i
18 level. We are out in the field, working with licensees.
19 We're members of committees. We attend meetings, l 20 workshops, symposia that licensees also attend, that's put 21 on by the industry. We're members of working groups.
22 MR. CARROLL: Was anyone from the NRC at the 23 Berlin meeting?
24 MR. THOMAS: We do our best, given the money
( ) 25 we have. We are a member of NEA. We do a lot of work NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N W.
(202) 234-4433 WASHINGTON. D.C. 20005-3701 (202) 234-4433
46 1 with IAEI. We have limited resources. We do the best we
, 2 can. We're on a number of human factors committees,
. 4
~
3 including IEEE and ANS committees. We share the 4 information among offices as best we can. That's another l 1
5 thing we accomplish in our quarterly meetings is to 6 exchange information that one office may gather that the l
7 other ones are not aware of. l 8 We subscribe to periodicals. We review i
9 literature.
10 CHAIRMAN APOSTOLAKIS: Okay, number 7.
11 MR. PERSENSKY: Can I just add to that? i 12 CHAIRMAN APOSTOLAKIS: Sure..
13 MR. PERSENSKY: You keep asking for road map l 1
) 14 of the plan and that's the intent of Section 3 of the I 15 plan. Section 3 of the plan talks about how we keep up i 16 with the state of the art, how we interact with others.
4 17 Cecil sort of focused on the nuclear, but we also spend a 18 lot of time interfacing with people in other agencies like 19 the FAA, Maritime Administration, because they have 20 similar problems. So we try to keep up to date through 21 that process and that's why element 3 of the plan really 22 is there. It's not only coordinations, but how do we keep 23 ourselves smart in human factors.
24 MR. THOMAS: Number 7 is NMSS's and I
/^T 25 understand they were going to address that as part of
()
NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N W.
(202) 234-4433 WASHINGTON, D C. 20005-3701 (202) 234-4433
47 1 their presentation this afternoon. Is that correct, Jit?
2 MR. SINGH: That's correct. This afternoon at 3 2:30.
4 CHAIRMAN APOSTOLAKIS: Okay, 8.
5 MR. COFFMAN: The question is how does the 1
I 6 staff plan to respond to the ACRS advice concerning 7 developing metrics for organizations and managements that 8 correlate with risk or performance? :
l l
9 The staff has interacted with the ACRS l i
10 throughout the years on this topic in '89, '90, '91 and 11 the last time in '93 in any detail. And then the ACRS l 1
12 wrote a letter to Chairman Selin on the review of !
13 organizational factors research program. That one was
~ ;
(~')'\
K- 14 dated April '93 and the committee indicated that it had i 15 been unable to arrive at a consensus concerning the 16 research. And the staff responded, I think, identifying 17 with the ACRS on this that the need to conduct regulatory 18 research on organizational factors, even though the 19 organizational factors appeared to influence safe 20 operation, articulating specific researchable issues that 21 study systematically the complex subject of organizational 22 factors and then quantify it is a difficult task. So l 27 where we are at this point in the research is we're at a l
24 juncture where the staff feels like or is considering
()~.
25 reformulating the research question from the generally NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N W.
(202) 234-4433 WASHINGTON. D C. 20005-3701 (202) 234 4433
48 1 worded issue of folding organizational factors into PRAs 2 into more specifically focused issues and worded issues 7_s
' k ')
3 and we're considering including a workshop of experts to 4 aid in that effort.
5 I might go on to say that the issue is rated 6 low because at the time the likelihood of success of doing 7 research in this area was considered low. It was 8 difficult because the plan was to validate the work that 9 had been done to date by gaining access to the senior 10 management process, senior management meeting process, but 11 that remained opaque and it did not look like it was going 12 to open up. Well, experience has -- basichlly we have 13 been overtaken by events. Experience has refocused.
- k. / 14 They're underscored. The potential of the qualitative 15 risk importance, risk worth of organization factors, now 16 the senior management meeting process is being examined 17 openly and there's SECY papers been written on it. There 18 are procedures being developed and so we are thinking what 19 further work can be carried on in that area and so we've 20 had some interoffice brainstorming sessions and trying to 21 formulate further plans.
22 MR. CARROLL: What are these SECY papers?
23 MR. COFFMAN: I don't have the numbers.
24 MR. CARROLL: Can you get them to Jit?
25 MR. COFFMAN: Yes. It may be singular. It NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N W.
(202) 234-4433 WASHINGTON, D C. 20005-3701 (202) 234-4433
49 1 may just be --
s 2 MR. SINGH: Just give to me.
I N)\ 3 MR. COFFMAN: Okay. I need to -- this is a 4 big subject and there are others here who could address 5 some other aspects if I've forgotten an important part.
6 MR. CARROLL: One of the things that troubled 7 Dave Ward when he was on the committee and I sort of !
8 picked up on it is that it just seemed to us that there 1
9 are people who make a professional career out of 10 understanding the dynamics of organization, the academic 11 community. And somehow or another we haven't figured out 12 a way to tap that knowledge. I'm an expert on 13 organizational factors. I was part of an organization for h ,
("~l k - 14 many, many years and I understand all this stuff, and I'm l 15 sure that's the attitude of a lot of people, both in j l
16 industry and the agency. We're smart engineers, but 17 there's a difference between that perspective and those 18 that have made a real study of the subject. I admit I've 19 met a bunch of those people and I wouldn't give them a 20 nickel to do research for me because I just think they 21 talk a lot. There's got to be something to this. There's 22 got to be people out there that could help.
23 George, do you have a thought on that?
24 CHAIRMAN APOSTOLAKIS: I have many thoughts.
O
(_j 25 MEMBER SEALE: I'm reminded that everybody has NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N W.
(202) 234-4433 WASHINGTON. D C. 20005-3701 (202) 234-4433
50 1 an opinion on education because we all have one, so there l l l I
gS 2 are a lot of experts on education out there too. Maybe '
b 3 that's part of our problem there.
l 4 MR. CARROLL: Yes, but the people that are )
i 5 professional experts on education, there's a lot of them l
6 that don't know anything.
7 MEMBER SEALE: Oh yeah.
l 8 CHAIRMAN APOSTOLAKIS: I think the same thing l
9 happens in cognitive psychology, Jay. I think the problem 10 with these people who study cognitive issues or 1
11 organizational issues from the theoretical point of view 12 and they tend to be academics is that their only interest 13 is to write papers for their peers, because that's how i
\- 14 they advance and well, I'm sorry, but that's the case.
15 And they are not really interested in solving anybody's 16 problem.
17 There's a conflict, I think, between engineers 18 and these types of people and a big mistake is to make l
l 19 these people the leaders of a group that is supposed to 20 solve our problem. I think they should be important j 21 members of the team, but never the leaders.
22 We simply have a very hard time in this 1
l 23 industry understanding that, that we don't take a 24 psychologist and make him the leader of a team that's A
, k_) 25 going to solve problems and you don't take an NEAL R. GR(lSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISMND AW., N W (202) 234-4433 WASHINGTON. D C. 20005-3701 (202) 234 4433
l l
51 1 organizational theorist and make him a leader of a team l
_s 2 that will solve nuclear problems. They should be i ,
3 advisors. They should be members of the team, but not the 4 leaders. They don't understand the culture of the 5 industry. They don't understand what it means to solve a 6 problem. You try to present some sort of a solution and 7 they give you 25 reasons why this is not perfect and so 8 on, so there have been a lot of problems. In fact, I just 9 remembered, you sent me your ideas on organizational 10 issues after you left the committee. It was very 11 interesting to see how many times this Agency has started 12 on that issue and then stopped and how many times they 13 have said we are thinking about it and again nothing came l
/~'i i
\
k/ 14 out of it. So --
15 MR. CARROLL: I really don't like the answer, l
16 well, it's a difficult problem. Humankind would still be '
l 17 hunters and gatherers trying to live at a subsistence l
18 level if we said it's a difficult problem.
19 You've got to bite the bullet. You got to try 20 at least.
21 MR. COFFMAN: I might say that certainly 22 clinical psychologists would not be the ones to address 23 the issue, but industrial psychologists have contributed 24 to what progress we have made in this area.
,~.
( 25 One of the things that we try to do, one of NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N W.
(202) 234-4433 WASHINGTON. D C. 20005-3701 (202) 234-4433
52 1 our objectives in research is to develop a technical basis 2 which includes a repeatable method per quantification or a p_
l
'~'1 3 repeatable method for measuring which was the objective of 4 the initial user need that came over on this area.
5 And it certainly -- we can find ways to 6 quantify the effects, the organizational effectiveness, 7 but to get those repeatable across different evaluators i 1
8 was a difficult aspect. And so there was progress made. i 9 CHAIRMAN APOSTOLAKIS: So you have had studies 10 then when you gave the same model to different people and 11 you found that it was not repeatable? And you have done 12 that with SLIM MOD? You have done it with all sorts of i 13 human error models that this Agency has developed and you i l0
\/ 14 found that those were repeatable?
15 I don't understand what repeatable means in j 16 this context because I submit that they will never be 17 repeatable.
18 This is not Newton's law here.
19 MR. COFFMAN: We did use SLIM MOD that was one 20 of the methods that was used. Whether we tested it fully 21 or just could look at the process and determine that it 22 was not repeatable.
23 CHAIRMAN APOSTOLAKIS: I mean there is a NUREG ,
24 that describes SLIM MOD in the context of human error, so 25 has that been submitted to that test? In other words, if NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N W.
(202) 234-4433 WASHINGTON. D.C. 20005-3701 (202) 234-4433
53 1 I give two errors to two different people and I ask them
,-s 2 to use SLIM MOD, they will come up with the same numbers?
b 3 Surely, you don't mean that and yet SLIM MOD 4 is a method that a lot of people think helps you structure 5 your judgment and get some answers.
6 MR. COFFMAN: The convergence criteria for 1
7 repeatable on human performance is more in the order of 80 i
8 percent or something like that. Now I don't know if we --
9 Joel, you can aid my memory on this.
10 Did we subject the Brookhaven SLIM MOD 1
11 approach to an independent check or was it just from staff l i
12 review that we determined that --
13 MR. KRAMER: We did not have an external O
- 14 review, but we applied that method twice at two plants, 15 Diablo Canyon and Limerick and using that process and that 16 model we could distinguish the difference between what 17 were two good plants at that point in time. Both were 18 good and you could pick up the difference, but that's with 19 respect to organizational factors and not the human error 20 question.
21 CHAIRMAN APOSTOLAKIS: That's very different 22 from demanding repeatability. I don't think that 23 repeatability is a requireraent when it come to the soft 24 sciences. There's just no way you can repeat things. We p
( ,) 25 cannot conduct experiments to show that you get the same NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N W.
(202) 234 4433 WASHINGTON, D C. 20005-3701 (202) 234-4433
54 1 answers under the same conditions and there's this 7- 2 infamous benchmark exercise from ISPRA where the 12 teams
'~) 3 used different methods and their results are scattered all 4 over the place. I mean that's the nature of the beast.
5 What you are trying to do is structure your thought 6 process and hope for the best and hope that you will 7 convince your peers that you have found something that is 8 worthwhile, but I don't think that we can impose the 9 requirement of repeatability the same way that physics 10 does or chemistry where unless you come within 11 experimental error your proposed model is garbage.
12 MR. KRAMER: George, I would agree with what 13 you've said, but I think there's another aspect from a
\s)
(
14 technical basis standpoint aside from reliability or 15 repeatability and that's the validity issue. And 16 establishing valid indicators for organizational 17 effectiveness is an issue.
18 CHAIRMAN APOSTOLAKIS: But I don't see that 19 any different for human error or for anything else in this 20 arena. I think the issues are there. I mean how do you 21 know that they have a good operators at a certain plant?
22 I mean you're trying to develop certain indicators to tell 23 you, but you can never really be 100 percent sure. In 24 fact, I'm told that your own inspectors when they walk O
(_/ 25 into a plant they can tell in 5 minutes whether this is a NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N W.
(202) 234-4433 WASHINGTON, D C. 20005-3701 (202) 234-4433
55 l
1 well run plant or not. So there must be some indicators. I 1
_ 2 It's just that we don't, in fact, I hear from a lot of my 3 colleagues who have done PRAs in the United States and 4 they have gone to Switzerland, every single one of them 5 comes back and they're so impressed by the Swiss plants. I 6 They say the moment you walk in there, you know this is a 7 well run organization and you ask the guy why and he 8 doesn't know. He doesn't know why he has that feeling.
1 9 MR. COFFMAN: We have asked, at least not -- l 10 well, maybe some of the staff have asked foreign 11 inspectors, but we've asked certainly U.S. inspectors and l 12 there is a range of indications and they come at a range i
13 of the -- they come at different points in the process at )
~g (x/ 8 14 which a plant is either becoming a poorer performer or 15 becoming a better performer. But to get -- we can't 16 regulate. We need this criteria of repeatability for the ;
17 purpose of establishing the validity of the process or 18 method developed to be able to regulate. I mean well 19 can't regulate --
20 MR. CARROLL: How are you currently 21 regulating? It's somebody's subjective judgment.
22 Now wouldn't it be better to at least t: Y to 23 quantify to the extent you can these attributes uat make 24 for a good organization?
A
( ,/ 25 MR. COFFMAN: Yes, and I think that effort, I NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N W (202) 234-4433 WASHINGTON. D C. 20005-3701 (202) 234-4433
56 1 need to get you that SECY paper because I think there's
_ 2 the effort to make this subjective judgment systematic.
U 3 The research was to go beyond what the Agency is 4 practicing to try and improve on that process.
5 CHAIRMAN APOSTOLAKIS: None of these sciences 6 is repeatable. I don't think economics is repeatable.
7 People try to develop models to provide insights and maybe 8 some good indicators and that's the best you can hope for, 9 I think. And the question is if we formed another 10 coordinating committee now, with five other people, nct 11 you and we gave them the same issues and asked them to 12 rank, how repeatable would the results be?
13 MEMBER KRESS: George, let me ask you a
/^\
(--) 14 question. ;
l 15 CHAIRMAN APOSTOLAKIS: Yes.
16 MEMBER KRESS: I conduct a PRA on a plant.
17 Where would I find organizational factors and human 18 factors showing up? I'll find them showing up explicitly 19 in human errors of operator actions that are required by 4 20 the procedures, whether or not they follow these or not in 21 fault trees and in event trees and I'll find them showing 22 up implicitly in reliabilities of equipment. So I 23 maintain that organizational factors and human factors are 24 already built in the PRAs. We just don't know how to fh)
(, 25 quantify them when we put in our failure rates. And the NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N W.
(202) 234-4433 WASHINGTON D C. 20005-3701 (202) 234-4433
57 1 question I'm leading up to to you is if one had an
,s 2 appropriate LER process or one quantified on a plant .
t T I N 'l 3 specific basis, things like reliabilities, fail".re rates j I
4 and got plant specific values for those, doesn't that ;
l 5 automatically include human factors and organizational !
6 factors for that plant in your PRA and reflect itself in 7 the risk calculation that one makes?
8 CHAIRMAN APOSTOLAKIS: No. What you said is 9 only partly true. For example, for the failure rates, 10 certainly the data that you get reflect organizational )
i 11 practices. You may get a higher failure rate in a plant 12 where maintenance is not very good, for example. And that 13 certainly is built in.
( 't A/ 14 There are many other areas where you don't 15 really see it. For example, in many cases we use generic 16 error rates. So what does that mean now?
17 MR. CARROLL: You don't have enough data.
18 CHAIRMAN APOSTOLAKIS: You don't have enough 19 data. Second, there may be correlations that are 20 introduced by organizational practices that correlate a 21 number of failure rates, a number of events and you don't 22 see that in the data base either because these are usually 23 rare events. Or even if you see it --
l l
24 MEMBER KRESS: What you're saying is there's
(~N i () 25 not enough plant specific data to reflect --
NEAL R. GROSS l COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N W (202) 234-4433 WASHINGTON. D C. 20005-3701 (202) 234-4433 l
l
58 1 CHAIRMAN APOSTOLAKIS: But also you need the 73 2 model because you may look at it, but you don't see it.
A 3 MEMBER KRESS: You don't need a model if j 4 you've got empirical data.
5 CHAIRMAN APOSTOLAKIS: You see data by itself 6 doesn't mean anything unless you have supplemental model.
7 MEMBER KRESS: You can quantify the risk level l
8 of that plant. l l
9 CHAIRMAN APOSTOLAKIS: NO, no, no, no. Let's 10 say that the two distributions are correlated, okay? You 11 have to have an overwhelming amount of data to say gee, 12 they're correlated.
13 MEMBER KRESS: You have to have a lot of ym
- \--) 14 empirical data.
15 CHAIRMAN APOSTOLAKIS: But if you just see one l
16 or two here and there, your mind will never go to that 17 unless you have a model that says look for this and in my l l
18 mind this is the most critical thing. I mean our plants
)
19 are so redundant that unless you worry about dependencies, 20 everything else is really minor. You may create new 21 initiating events. In fact, there was an incident where l 22 there was a path and we lost several thousands of gallons
~
23 of water again and that was because people ran the wrong 24 valves at the same time. Now why should you do that? Is r~
( ,}/ 25 there anything in the book that says you shouldn't be NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N W.
(202) 234 4433 WASHINGTON, D.C. 20005-3701 (202) 234-4433
59 1 doing that? See, these are the kinds of questions -- now, es 2 you can treat that as another initiator or you can read
\ t
%./
3 more into it as some sort of bad practice.
4 Organizational practice is partly in the data, 5 but there are also other things that are not in the data 6 and unless you have some sort of a model, you will never 7 catch them. Okay, so that's why it's a hard problem, 8 because you can say everything is in the data, but again I 9 think it's the dependencies of the -- I mean look at 10 common cause failures. In the original studies in the 11 late '60s and early '70s, we said failure rate is 12 everything. You talk about failures and the failure rates 13 until people started realizing that the failure rates are r~s
( I k/ 14 squared, then cubed because of their redundance and you )
1 15 say well, wait a minute now. Some of these failures don't 16 happen that way. You have common causes. And it took us 17 20 years to come up with something that is not repeatable 18 yet by the way. But at least people are beginning to say 19 well, this makes sense.
20 In my opinion, in risk assessments, unless you 21 are talking about dependencies, you're not talking.about a 22 serious issue because the plants are so redundant. The 23 dependence is on correlations. Even if you multiple the 24 failare rate by 10, it doesn't matter because you square A
k ,), 25 it or you take the cube power at some point. So it NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N W.
(202) 234-4433 WASHINGTON. D C. 20005-3701 (202) 234-4433
60 1 doesn't matter. You raise it by epsilon, but if you say
,-s 2 no, this affects five things at the same time. That's a
! 1
%~/ different story now.
3 I think that's what Jack is worried 4 about all the time, right, dependence. It's common cause 5 failures, commonalities and that's the stuff that's 6 missing.
7 MEMBER KRESS: Thank you.
8 MR. CARROLL: And bad management is a common 9 cause.
10 CHAIRMAN APOSTOLAKIS: Exactly,it's the 11 biggest common cause.
12 MEMBER MILLER: Let me pick up on what Jay 13 said. I think organizational factors -- and I've not made
,G
%s 14 a study on all of this -- but have to start right at the 15 top.
16 MR. CARROLL: Exactly.
17 MEMBER MILLER: For example, let's forget 18 about PRA and nuclear power plants. If you want to invest 19 in a company, for example, the first thing I look at is 20 the mentality of the management. I don't care about the 21 product. You can go into a plant, for example, and 22 interview the plant supervisor. You can probably tell if 23 the plant is run well or not just by interviewing one guy. 1 24 But the style of good management is hard to quantify. !
I
/(-)s 25 CHAIRMAN APOSTOLAKIS: That's right.
I NEAL R. GROSS I COURT REPORTERS AND TRANSCRIBERS !
".323 RHODE ISLAND AVE., N W. l (202) 234 4433 WASHINGTON, D C. 20005-3701 (202) 234-4433
61 l
l 1 MEMBER MILLER: You can take styles all over l 1
_ 2 the map and they all often work. You can take styles all I 3 over the map and they cannot work too. It's kind of built j i
i 4 in. You got to be born with it in my opinion. l l
5 CHAIRMAN APOSTOLAKIS: I mean sometimes hands 1 6 on management is the best. Other times, detached 7 management is the best. You never know.
8 MEMBER MILLER: I don't think you're ever 9 going to quantify organizational factors very well, 10 MR. CARROLL: No, but I think you could 11 understand them a lot better in nuclear power plants than 12 we do right now.
13 CHAIRMAN APOSTOLAKIS: I think I'm in the same
(
(3 ~) 14 league as dynamic human errors. We will never reproduce 15 numbers that you say this is it, but in the process of 16 getting there, you are learning so much.
17 By the way, speaking of culture, I don't want 18 you to feel you are falling behind because a lot of the 19 discussion that is supposed to take place at 2:30 is 20 taking place now. So we're not really falling that far 21 behind. Plus it's very easy for me to say skip the next 22 five questions, right?
23 (Laughter.)
24 It's the Chairman's prerogative. We had an f~'s is ,) 25 interesting presentation the other day. It was a video NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N W.
(202) 234-4433 WASHINGTON, D C. 20005-3701 (202) 234-4433
62 1 conference, Southern California Edison where they talked ,
I 7- 2 to us about their risk monitor. Now the thing that really
('~') 3 impressed me was one thing. Tom Hooks said we have those l 1
1 l
4 monitors everywhere and it has created a tremendous change 5 in the culture of the plant. He said people who felt sort 6 of isolated before, you know, a guy who paints a tank, for 7 instance, just this job, he does it, he doesn't really 8 care. Now he looks up and looks at the monitor and sees 9 how the core damage frequency is affected by the fact that 10 the tank he's painting is out of service. This is a 11 tremendous change in culture.
12 Now here is something out of the blue that has 13 had a tremendous impact and people have been talking about t'3 sl 14 culture all these years and everybody kept saying well, 15 gee, how do we change that? Culture is culture. Very 16 simple thing. Now they are acting more like a team 17 although I wouldn't really use that word, but really there 18 is a big difference. Here you have a guy hat he works at 19 the nuclear plant, would have been whatever, a rice 20 producing plant. All he does is paints tanks. And then l 21 all of a sudden he sees that what he is doing is really
! 22 important and the work he has done, that curve over there 23 starts going down.
24 MEMBER SEALE: George, rice grows in fields, 10
() 25 but --
l NEAL R. GROSS COURT REPORTERS AND TaNSCRIBERS 1323 RHODE ISLAND AVE., N W.
(202) 234-4433 WASHINGTON, D C. 20005-3701 (202) 23( J33 l
63 i
1 CHAIRMAN APOSTOLAKIS: But they paint tanks g~g 2 too, i $
3 MEMBER SEALE: But I think the interesting 4 aspect to this is you referred to it as a team attitude, l
1 5 but in a real sense, the individual reaction is more one '
l 6 of ownership.
7 CHAIRMAN APOSTOLAKIS: Exactly.
1 8 MEMBER SEALE: And it's fascinating that 9 ownership is a strong element in team identification and l l
10 it's those kinds of interrelationships, I think, that we 11 need to try to understand better. It should be no 12 surprise to us that providing a person with information 13 about how their particular responsibility affects plant O
k- 14 safety status is an effective way of involving that person 15 emotionally and intellectually in the operation of the 16 plant.
17 MEMBER MILLER: Let me go back to that risk 18 monitor. I agree with Bob. I spent half a day in that l 19 plant last summer.
20 CHAIRMAN APOSTOLAKIS: Which plant?
21 MEMBER MILLER: San Onofre.
22 CHAIRMAN APOSTOLAKIS: Okay.
23 MEMBER MILLER: And I saw the risk monitor 24 before we saw it here. And I agree the risk monitor is a
( 25 key factor in that, but I got to talk to all the managers, NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N W (202) 234-4433 WASHINGTON. D.C. 20005-3701 (202) 234 4433
64 1 except Harold Ray, I missed him. I could see right away, 7
~~ 2 just talking to those people they created the environment 3 so the risk monitor is a powerful tool for their style.
4 So it's all put together into one -- if you put a risk 5 monitor in another plant where the manager has a different 6 attitude, it may not work.
7 CHAIRMAN APOSTOLAKIS: That's true.
8 MR. CARROLL: That's exactly right. You could 9 have a plant just exactly like San Onofre 2 and 3 in terms 10 of physical design and put a management in charge of it --
11 MEMBER MILLER: And give him a risk mc 'itor 12 and you would not have the same culture.
13 MR. CARROLL: A risk monitor and if the Vice
( )
- > 14 President looks at the thing and says ch, that's a big 15 joke.
16 CHAIRMAN APOSTOLAKIS: Maybe he wouldn't even 17 give them the risk monitor though if he feels that way, so 18 I'm not sure you can separate the two. But I'm curious, 19 how long were you there, Don?
20 MEMBER MILLER: About a half a day.
21 CHAIRMAN AFOSTOLAKIS: You looked at the 22 monitor?
23 MEMBER MILLER: Oh yeah.
- 24 CHAIRMAN APOSTOLAKIS
- Did you feel part of
( ,m
'(_,/ ) 25 the team at the end?
NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE island AVE., N W.
(202) 234-4433 WASHINGTON. D C. 20005-3701 (202) 234-4433
l 65 1 (Laughter.)
i f3 2 MEMBER MILLER: I was amazed at what they were
('~# ) ,
3 doing with it,yes. I had read the paper before I got )
4 there, but reading the paper doesn't have much impact 1
l 5 until you see, you actually go around the plant, exactly J 6 what you saw in the video, people looking at it and 7 talking about it and talking about what it's doing for 8 them.
9 CHAIRMAN APOSTOLAKIS: By the way, ownership 10 is really a big deal. Here is a way to --
11 MEMBER MILLER: Management provides the l 1
i 12 atmosphere for ownership and then the management takes j 13 advantage of tools to make that happen.
I k/ 14 MR. CARROLL: But I guess I have to put in a 15 word of caution about something like a risk monitor.
16 Unless you do it right, it can mislead you. It didn't 17 involve a risk monitor, I know of a couple of instances 18 where people have taken their PRA and made a decision of 19 equipment to take out of service, misinterpreting or I
20 misusing the PRA, doing something very risky and it should 21 have been intuitively obvious to them, but they said oh, 22 the PRA says this is all right. That is scary.
23 MEMBER MILLER: You can't get too dependent on 24 the data. It's counter intuitive then. 1 l
s ;
y_) 25 CHAIRMAN APOSTOLAKIS: Question 9. l NEAL R. GROSS ,
COURT REPORTERS AND TRANSCR:BERS {
1323 RHODE ISLAND AVE., N W. I (202) 234-4433 WASH!NGTON D C. 20005-3701 (202) 234-4433
66 1
1 MR. THOMAS: What are the technical bases for ;
2 defining the staffing levels inside and outside of the
(~N
( ') 3 main control room and for communication procedures?
4 The technical bases for the staffing level 5 requirements that are found in 10 CFR 50.54 are described 6 in the Federal Recister notice that promulgated the l l
7 requirements back in '83. The citation is Volume 48, No.
8 133, July 11, 1983.
l 9 MR. CARROLL: That's the legal basis. What's !
10 the technical basis? j I
11 MR. THOMAS: The technical basis, yeah, I just j l
12 in case somebody wanted to look, the technical bases had I l
~
13 to do with plant operating modes, activities that gross, l
\_/ 14 broad activities that must be performed, operator i 15 qualifications, that is SRO versus RO and allowing a 16 certain amount for overlap and contingencies. I would 17 emphasize that these were minimum staffing levels. We've 18 always maintained that the requirement to determine .
I 19 adequate staffing levels was a responsibility of the i 20 utility and these were just the minimum ones for control 21 room or on-site or at the controls at that time.
22 MR. CARROLL: Do you have a sense as to how 23 many fraction of utilities use the minimum requirements as 24 opposed to --
()
("% 25 MR. THOMAS: Yes, almost all of them. We've NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W (202) 2344433 WASHINGTON, D C. 20005-3701 (202) 234-4433
67 1 done a number of studies and taken a number of looks.
s 2 Almost all of them staff above the minimum. In fact,
! \
V 3 their tech specs are typically above the minimum.
4 MEMBER SEALE: Do you have a basis for 5 deciding how many is too many?
6 MR. THOMAS: No. We know that too many can --
7 you can have too many, but we don't have criteria for --
8 MEMBER SEALE: I think INPO pays very close 9 attention to the question of too many people in the 10 control room.
11 MR. THOMAS: Yes. The second part of the 12 question dealt with communicating the basis for our 13 communications procedures. These are communications, (nx-) 14 review guidance that we use in our inspection procedure to 15 evaluate licensees' corrective action programs that 16 involve communication related problems.
17 MR. CARROLL: Are you talking about 18 communications in the context of control room 19 communications?
20 MR. THOMAS: All kinds, including 21 communications of management expectations, person to 22 person communications in the control room, any place that 23 the transfer of information is important.
24 What we did on -- it's mainly person to person 7-( ,) 25 communications in the control room, but it does include NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N W.
(202) 234-4433 WASHINGTON, D C. 20005-3701 (202) 234-4433
68 1 others and our inspection guidance, for example, is 7- 2 designed to follow your nose and if you find it, there are
('~) 3 communications problems as a result of inadequate l
1 4 communication of management expectation and we have j i
5 procedures that lead you in that direction too. !
6 But our procedures were based on a model of --
7 a model of communication processes, input from a subject 8 matter expert panel and a lot of data that was gleaned 9 from a systematic review of our LERs and inspection 10 reports that were concerned with communication errors.
11 By the way, that inspection NUREG that's 12 referenced in our program plan has been developed and it's 13 out with the regions right now for trial use and comment. ,
f~) l
- 14 MR. THOMAS: Okay. Number 10 is research. )
15 MR. PERSENSKY: The question is where are the 16 deficiencies or holes in NUREG 0700. In fact, we term 17 them in I guess it's items 1.2.8 and 1.2.9 gaps. These 18 gaps were identified in NUREG/CR-5908, which was the 19 predecessor to NUREG 0700 in terms of the advanced control 20 systems. This is a list of them. There are 14 of them.
21 The priorities were set by the contractor at the time, 22 based on information that they had regarding how soon some 23 of these things might be happening in the nuclear setting.
24 The single starred items are things that we (O)
(2,i 25 are already addressing based on user needs. The double NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS l
1323 RHODE ISLAND AVE., N W.
(202) 234 4433 WASHINGTON, D C. 20005-3701 (202) 234-4433
l 69 1 starred items are those that are being addressed in part
,ss 2 under the bottom one there, integration of advanced
'~] 3 systems and conventional CR, control rooms, which really l
4 we usually term it as the hybrid project. It's talking 5 about putting two types of systems together.
6 In starting that project, the first element 7 was to determine what issues might be associated with the 8 hybridization. So some of those things came up again. So 9 they are being addressed under the hybrid project.
10 MEMBER MILLER: Under that second priority, I 11 am intrigued by test and maintenance of digital systems.
12 I'm intrigued, why does that come into this? Why is that 13 in NUREG 0700 or why is it a gap in NUREG 0700?
\- 14 MR. PERSENSKY: It's a gap because we do not 15 have any guidance in NUREG 0700 for it, because we didn't 16 feel there was adequate basis for developing criteria to 17 put in as a standard.
18 MEMBER MILLER: Is this where it belongs? Is 19 that a human factors issue?
20 MR. PERSENSKY: The human factors elements of 21 it are, yes.
22 MEMBER MILLER: Oh, that's what you mean, 23 okay.
24 MR. PERSENSKY: Yes. It's the human factors
(~'s
(_) 25 elements of the problem. Part of it gets into questions NEAL R. GROSS COURT REPORTERS A,l9 TRANSCRIBF.RS 1323 RHODE ISLAND .'VE., N W.
(202) 2344433 WASHINGTON, D C. 20 0 5-3701 (202) 234-4433
70 1 of maintainability, which may be a design issue as opposed 2 to --
[..s\
MR. CARROLL:
~
3 Now 0700 was the basic document 4 used in the program design review?
5 MR. PERSENSKY: The original 0700 was used for 6 the DCR, detailed design control room, control room design 7 review guide, yes.
8 MR. CARROLL: What we're talking about bere is
! 9 additions to it to enable or to provide guidance on how to 10 look at modifications to existing control rooms?
11 MR. PERSENSKY: Modifications to existing 12 control rooms as well as advanced reactors and advanced l
l 13 control rooms, if there are any.
14 The original 0700 was developed back in the l
l 15 early 1980s, right after TMI. It was basically analog l
16 controls, because it was analog control rooms at the time.
17 0700 Rev. 1 which was just issued this summer 18 included something, a good deal of the digital aspects 19 because of the need for review for the advanced control, l 20 advanced reactors, as well as some changes that are coming i
t l 21 into control rooms.
I 22 In doing that effort, developing this Rev. 1, 23 we identified some areas, which are these gaps here, that 24 we didn't feel there was sufficient guidance or sufficient
( 25 technical basis to develop guidance. So that's why we NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N W.
(202) 234-4433 WASHINGTON, D C. 20005-3701 (202) 234-4433
71 1 consider these gaps.
, _s 2 MEMBER MILLER: So 0700, this version or it's e ;
I /
3 a version like it, was used? I wasn't on the committee 4 when we did the evolutionary plants, you know, the ABWR 5 and the System 80. But 0700 was used as part of that 6 review or elements of the 0700?
7 MR. PERSENSKY: Elements of 0700 Rev. 1 8 NUREG/CR 5908 was used. NUREG 0711 was used from a 9 process standpoint. This was being developed at the same 10 time we were doing some of the reviews. In fact, the 13 people that were developing it were also helping 12 performing the reviews for NRR.
13 MR. THOMAS: They came up in parallel, Don.
./ ~ N
' k m- 14 The reviews were done as part of the upgrade. j 3 MEMBER MILLER: I need to go back and look at l 16 those. In the last several months, I've been in two power l 17 plants who in their control rooms are dramatically 18 different. I don't think either one of them are advanced ,
19 reactors.
\
l 20 The plant in Japan, where they build our ABWR, i i
1 21 but they did not use our control room. They did their 22 own. Then I was in the N-4 --
l 24 MEMBER MILLER: ABWR. They did not use the G
/
25 --
NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., NW.
(202) 234-4433 WASHINGTON O C. 20005-3701 (202) 2344433
72 1 MR. CARROLL: KK67.
, 2 MEMBER MILLER: Yes. I don't know the
, \
V 3 difference because I don't know what the ABWR was. But 4 it's substantially redesigned by Kepco.
5 I was also in the N-4 reactor in France.
6 There's dramatic difference in the control room philosophy 7 there. In each, they told me very pointedly, the 8 operators had total involvement in those two plants, while 9 the N-4 reactor had not yet built a simulator. Of course 10 the Kepco built a simulator two years before they built a 11 plant. I am wondering how that all fits into this. Would 12 0700 agree with what they have done in those two plants?
13 MR. PERSENSKY: Well, in fact, 0700 was sort
/^N, V 14 of peer reviewed in both of those instances. We are going 15 to do further work in terms of future revisions of it with 16 them. You have got to remember, these are review 17 guidelines as opposed to design guidelines. So the basic 18 process is adopted. It's not meant to be this is the way 19 all control rooms should be. It's just these other 20 criteria that make it acceptable from a human factors 21 standpoint.
l 22 MEMBER MILLER: So staff from NRC have been to l
l l 23 those two plants then?
24 MR. PERSENSKY: Yes.
%j 25 MEMBER MILLER: I would like to talk to them.
f NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS l 1323 RHODE ISLAND AVE , N W.
(202) 234-4433 WASHINGTON. D C. 20005-3701 (202) 234-4433 l
73 1 I am just curious how they felt about them. You know, I 7
- 2 come in there and spend three hours and see dramatically
\'~ ']
3 different plants.
4 MR. CARROLL: Dick was part of the team that 5 went to KK67.
6 MEMBER MILLER: How different is that from the 7 GE control room, which I haven't looked at?
8 MR. ECKENRODE: GE does not have a control 9 room. l l
10 MEMBER MILLER: Oh, they didn't have it?
l 11 MR. ECKENRODE: No.
12 MEMBER MILLER: That wasn't part of the 13 design?
k- -
14 MR. CARROLL: Now they have a DAC.
15 MR. ECKENRODE: They have a design process by 16 which they have to do their control room. They deferred 17 most of their information to Japan to Mitsubishi and --
18 MR. CARROLL: Hitachi and --
19 MR. ECKENRODE: Hitachi and Toshiba.
20 MEMBER MILLER: So when Taiwan builds theirs, 21 they can basically do what they want to do?
22 MR. ECKENRODE: That's correct.
23 MEMBER MILLER: They can go back to Kepco and 24 just take Kepco's design?
/"
( ,N) 25 MR. ECKENRODE: That's correct.
NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON. D.C. 20005-3701 (202) 234 4433
74 1 MR. CARROLL: Which is a very well thought out I
fS 2 design. I N~) j 3 MEMBER MILLER: Oh I really liked that one, l 4 just my sense.
5 MR. CARROLL: They did a good job of giving 6 the operator really meaningful tools.
7 MEMBER MILLER: I could sit in the middle of i I
8 the room and I couldn't read Japanese obviously, and I ,
l 4
9 could tell what was going on at the plant.
10 MR. ECKENRODE: Yes. They basically took two .
l l
11 individual designs, a Hitachi design and a Toshiba design, l I
12 picked the best features of both, and integrated them into 13 a control room. Something we could never do here in the
(,_) I
\/ 14 United States.
15 MR. CARROLL: As you pointed out, there was a 16 lot of input on that from Kepco.
17 MR. ECKENRODE: Oh yes.
18 MEMBER MILLER: Oh yes. They took a lot of 19 pride, this is Kepco's design.
20 MR. CARROLL: And interestingly enough, the 21 hardware for all of that is on one unit it's Hitachi and 22 on the other unit it's Toshiba, but it's --
I 23 MR. ECKENRODE: The same design.
l 24 MR. CARROLL: Transparent to the operator, f
(,) 25 MEMBER MILLER: I didn't know that. It's NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W. l (202) 234 4433 WASHINGTON, D.C. 20005 3701 (202) 234-4433 l
75 1 transparent to me.
-s 2 CHAIRMAN APOSTOLAKIS: I think it's time for a i
)
/
't'~
3 break. Now let me make a couple of suggestions here.
4 There are, I don't know, six or seven questions to 5 discuss. We have an hour and a half for lunch. What we 6 could do is break now, come back at 10:35, start the AEOD 7 discussion according to schedule. I don't know, do you 8 need an hour and a half, Jack?
9 MR. ROSENTHAL: We can cut it back.
10 CHAIRMAN APOSTOLAKIS: Cut it back a little 11 bit. Then maybe go to lunch, but take only an hour for 12 lunch, and come back and continue the discussion of these 13 questions. Is that okay, Amarjit? That doesn't create
/N I
f L) 14 any problems?
15 MR. SINGH: Not for me, as far as the staff is 16 concerned.
17 CHAIRMAN APOSTOLAKIS: How about you?
18 MR. CARROLL: Where did the hour and a half 19 come from?
20 CHAIRMAN APOSTOLAKIS: I don't know.
21 MR. ROSENTHAL: We just have our 45 minute 22 presentation and 45 minutes for questions, so it depends 23 on you.
24 CHAIRMAN APOSTOLAKIS: Because I would like to (D 25 finish the whole thing by 3:00, if possible. But I think
()
NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N W.
(202) 234-4433 WASHINGTON. D.C. 20005-3701 (202) 234-4433
76 1 we've already done a lot of the discussion.
,s 2 Okay, so we'll be back at 10:35. By the way, e s 3 is Mr. Mayo here?
4 MEMBER MILLER: Charlie Mayo?
5 CHAIRMAN APOSTOLAKIS: Yes.
6 MEMBER MILLER: We traded voice mail, E-mails.
7 He did not feel it's worth his time to come up. This is 8 on Thursday. Why would he be here today?
9 CHAIRMAN APOSTOLAKIS: Today.
10 MR. SINGH: He's supposed to be here today.
11 MEMBER MILLER: Oh, he is?
12 MR. SINGH: Yes. He had talked to us 13 yesterday. He said he's supposed to --
tm
- 14 MEMBER MILLER: He E-mailed me yesterday and 15 he said he would not be here. I guessed maybe he meant 16 Thursday.
17 MR, SINGH: He meant Thursday.
18 (Whereupon, the foregoing matter went off the 19 record at 10:25 a.m. and went back on the record at 10:41 20 a.m.)
21 CHAIRMAN APOSTOLAKIS: So we will now continue 22 with the AEOD activities. Who is the presenter?
23 MR. TRAGER: Gene Trager.
24 CHAIRMAN APOSTOLAKIS: Okay, fine.
r}
( j 25 MR. TRAGER: I am with the office of AEOD and
)
NEAL R. GROSS 1 COURT REPORTERS AND TRANSCRIBERS l 1323 RHODE ISLAND AVE., N W.
(202) 234-4433 WASHINGTON, D C. 20005-3701 (202) 234-4433
77 1 Safety Programs Division, have been about 15 years. Much
,- 2 of that time has been spent in studies of human
( )
~
3 performance during operating events.
4 In the agenda, I know we're listed as 5 presenting during this period the activities and products.
6 CHAIRMAN APOSTOLAKIS: I'm sorry. I'm sorry.
7 I think Dr. Seale has to chair this.
8 ACTING CHAIRMAN SEALE: Okay, fine, whatever.
9 MEMBER APOSTOLAKIS: Sorry to interrupt.
10 ACTING CHAIRMAN SEALE: Continue.
11 MR. TRAGER: On the agenda, it lists AEOD 12 activities and products. I think I'm going to touch upon 13 that briefly toward the end, but the activities and the
, ,\
-- 14 ongoing products being produced are pretty much covered in 15 the human performance program plan.
16 What we are trying to cover during this time 17 period is a human performance event data base that we are 18 working on, to briefly discuss the effort so far and our 19 plans for the future.
20 A brief history of the data base goes 21 something like this. Although we have performed a number 22 of studies of human performance going back to the mid 23 1980s or the early 1980s, in approximately 1990, we began 24 studies with multi-disciplinary teams on-site studies with
^
iQ i
q,j 25 the assistance of Idaho National Engineering Laboratory NEAL R. GROSS COURT REPORTERS AND TRANSCR'BERS 1323 RHODE ISLAND AVE., N W.
(202) 234-4433 WASHINGTON, D C. 20005-3701 (202) 234-4433
78 1 and the human factors aspects of the events.
7-w 2 All the studies were performed using a
\)
3 protocol, so we asked the same questions during each of 4 the event investigations. We completed 20 of these over 5 the time period from 1990 to 1993.
6 In 1992, the decision was made to start or 7 develop a human performance event data base. We solicited 8 or asked NRR and RES to assist us in developing the 9 structure of that data base. The structure is based on a 10 human performance investigation process or protocol that 11 have been developed by the Office of Research. The NRR 12 office working group completed the data base specification 13 in 1993. AEOD authorized INEL to go ahead and construct
- 14 the data base.
15 The work has proceeded since that time for a 16 delay, a brief delay in 1995 because of budget cuts we 17 experienced. At the beginning of 1996, we resumed and 18 expanded and scope of the project.
19 ACTING CHAIRMAN SEALE: Could I ask the multi-20 disciplinary team studies that you referred to that took 21 place beginning in 1990, these were in response to 22 specific events?
23 MR. TRAGER: Yes. They were. They were 24 events -- our intention was to collect information in a
(^g
(_j/ 25 data base as an aide in determining what information NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N W.
(202) 234-4433 WASHINGTON, D C. 20005-3701 (202) 234-4433
79 1 should be reported in general on an event. We performed gs 2 these studies at sites, sometimes giving assistance to V 3 augmented inspection teams, inspections, and sometimes 4 study of events just because they appeared to have 5 interesting human performance during the event.
6 ACTING CHAIRMAN SEALE: Have you correlated or 7 attempted to correlate any of those results perhaps with 8 respect to things like SALP scores or something of that 9 sort?
10 MR. TRAGER: We have not done that yet. j 11 Actually, we're at the point now where we just have the 12 raw material. We have the data base at this point. We're
_ 13 at a point where we can release it to NRR and Research and j 14 solicit their comments and advice, try and resolve any 15 problems they have, and go forward.
16 We have studies planned for the near future 17 though, of the data and the data base.
18 ACTING CHAIRMAN SEALE: Maybe I am 19 anticipating what you are going to tell us a little later, 20 but will those studies try to ferret out specifically 21 things that might be a result of management -- I won't say 22 mal, but mispractice, perhaps is a better way to 23 characterize it, as opposed to operator.
24 MR. TRAGER: That is one of the areas that's (Q_) 25 covered in the data base.
NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODF. ISLAND AVE , N W.
(202) 234-4433 WASHINGTON, D C. 20005-3701 (202) 234-4433
80 1 ACTING CHAIRMAN SEALE: So you are not just 2 going to pick on the operators then?
3 MR. TRAGER: Absolutely not.
4 MR. CARROLL: There is a good report on those 5 studies.
6 MR. ROSENTHAL: Let me chime in that that's 7 NUREG-1275, volume 8, reports those studies.
8 MR. CARROLL: You might want to send Jit 9 several copies of that to distribute to the members.
10 ACTING CHAIRMAN SEALE: What was that one?
11 MR. SINGH: 1275.
1 12 MR. ROSENTHAL: Volume 8.
13 ACTING CHAIRMAN SEALE: Eight?
(~
kJ 14 MR. ROSENTHAL: Yes. And the HPIP process 15 itself is a MORT-like process, in which you start with an 16 unsafe -- you are familiar so that the operator had, to 17 use Dr. Poslaka's phrase, an unsafe act. Then you ask is 18 this due to training and inadequate procedure. You go 19 progressively lower.
20 ACTING CHAIRMAN SEALE: Yes. I met MORT once 21 before.
22 MR. ROSENTHAL: You've been MORT'd, but it's 23 not just in the problem of the operator.
24 MR. TRAGER: Information currently in the data O
(,) 25 base, this is what has currently been incorporated. As NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND A' E., NW (202) 234-4433 WASHINGTON, D.C. 20005-3701 (202) 234-4433
81 1 you can see, we have a number of AITs from the 1990 2 through 1996 time period. Human performance studies that
\'] 3 were performed by AEOD from the years 1990 through 1993, 4 we have started picking up information that NRR developed 5 in their team inspections. We also have licensee event 6 reports for these events. So this, I guess you could call 7 it double counting.
8 Then the next question is, well how much do 9 you double count. I made up a -- this is the same 10 transparency, but it's got some totals. The printer 11 wasn't working this morning, so the last column is a 12 little -- but as you can see, that's actually the number 13 of events we covered. It wasn't because the events p) i
\- 14 fluctuated this much, I think, or that there were more or 15 less problems in any given year. It's just the events 16 that were studied.
17 But there is some double counting, as I said.
18 On some occasions, we were asked to assist in an AIT, so 19 we would conduct a study as part of that team effort, turn 20 it over to the AIT team leader, who would decide what 21 portions of the study that they would use in their report.
3 22 So for some events, for example, it might be an AIT 23 report, an HPS report, and an LER. But this is -- we're 24 going to start to process this data and tell you what C) q 25 differences there are in different types of events reports NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 (202) 234-4433
82 1 and where information might be improved.
,_s 2 ACTING CFAIRMAN SEALE: Well, that's double
(#)
'~
3 counting within columns. I assume that's what you mean.
4 Between columns.
5 MR. TRAGER: Right.
6 ACTING CHAIRMAN SEALE: But from 1990 through l
7 1993, they are essentially double digit LERs. After 1994, 8 1994 and afterwards, they are single digit. That's what 9 you looked at rather than what was submitted from the 10 licensees, I assume.
11 MR. TRAGER: Well, for example --
l 12 MR. ROSENTHAL: In 1994, there were three l 13 AITs. Those three corresponding LERs were --
,r-)
kJ 14 ACTING CHAIRMAN SEALE: I understand.
l 15 MR. ROSENTHAL: And the three AITs were put in 16 the data base. If there had been more AITs, there would 17 have been more LERs. The purpose here is to compare the 18 richness of the reported human performance information and 19 AITs, special team inspections by NRR, to what the 20 licensees are telling us in their LERs.
21 ACTING CHAIRMAN SEALE: Yes. I understand 22 that. I think I understand that. Okay. I got you. I 23 think I got you now.
24 MR. CARROLL: I am curious why, for example,
( ,/
25 in 1995, you had two AITs and only one LER.
NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N W.
(202) 234-4433 WASHINGTON. D C. 20005-3701 (202) 234-4433
83 1 MR. TRAGER: My guess, and this is just a 7- 2 guess, I would have to double check, but I think an LER Y 3 may not have been written.
4 MR. CARROLL: Well that's probably -- why not?
5 MR. TRAGER: It should be, but I am thinking i
6 either it wasn't written or we couldn't find it. Our goal 7 is to have one to one correspondence between events and 8 LERs.
9 MR. CARROLL: The same is true of 1996, five 10 and four.
11 ACTING CHAIRMAN SEALE: Yes.
12 MR. CARROLL: But some of my earlier 13 statements about how poorly the LERs captured the human
\' 14 performance aspect of events is based on looking at the 15 1990 through 1993 data.
16 The other thing that jumped out at me were 17 some of the HPS studies that were done, it seemed to me 18 clearly indicated that an AIT should have been done and it 19 wasn't. That was two things that jumped out at me when I 20 was involved in this study.
21 MR. ROSENTHAL: When we were presenting the 22 HPS studies, you encouraged us to attempt to incorporate 23 that kind of work into the normal agency's processes. In 24 this very room I think you made those comments.
i
/~'s
(_) 25 MR. THOMAS: No. Probably not. It was NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N W.
(202) 2344433 WASHINGTON. D C. 20005-3701 (202) 234 4433
i 84 1 probably over in Phillip's Building, but okay.
1
,S, 2 MR. ROSENTHAL: Since then, NRR is conducting
, Y ,]
3 special team inspections of selected events which are very 4 similar to the HPS studies that we did.
5 I also get to participate as the AEOD 6 representative on a marathon phone calls about whether we 7 should have an AIT and to ensure that as part of the AIT 8 charter, we do include a request for an investigation of 9 the human aspects, so we have done that.
10 MR. TRAGER: And I think this is -- where the 11 data base is now, there will be more information I'm sure 12 in 1996. It hasn't found its way in there, but I think we
, 13 have the years up through 1995 pretty well covered.
/~Ni i
l 14 The information in the data base itself --
15 MEMBER APOSTOLAKIS: Can we get a copy of this 16 viewgraph? I don't think we have it.
1 17 ACTING CHAIRMAN SEALE: Just add the last 18 column.
19 MEMBER APOSTOLAKIS: You mean I am supposed to
- 20 do that?
l l 21 MR. SINGH: George, I'll get you a copy.
22 MR. CARROLL: Professors.
23 MR. TRAGER: As I said earlier, the data base 24 structure was based on a human performance investigation o
(s) 25 process or protocol. These are the questions that are NEAL R. GROSS l COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N W, l
I (202) 234 4433 WASHtNGTON. D C. 20005-3701 (202) 234-4433
85 1 asked in sort of an MORT process.
l
-s 2 As you can see, we asked these questions. We i
/ \
\q,] l 3 asked questions about these things consistently. It just j 4 a lot of things we take for granted. You know, we find j 5 out things in the event or in the study that we would not 6 have otherwise found out, some surprising things. These 7 are the types of information.
8 MEMBER APOSTOLAKIS: Yes, but this is really 9 where you need a model I think because I mean if you look 10 at these bullets, they certainly make sense, but the 11 question is, is there anything that's left out? When we 12 say management and organization, what do we mean? For 13 example, in several of incidents that I have looked at,
, 's l 14 organizational learning seems to be a weakness in the 15 sense that a similar incident happens somewhere else or 16 maybe in the same utility, but the lesson was not learned 17 and was repeated. Now that goes a couple of levels below 18 just the heading of management and organization.
19 So this is where you would benefit, I think, 20 from some sort of a structured approach. That was the 21 idea behind these figures that we attached in our 22 comments.
23 ACTING CHAIRMAN SEALE: But isn't the MORT 24 approach a structured approach?
) 25 MEMBER APOSTOLAKIS: A lot of them are NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 (202) 234-4433
86 1 structured approaches, but regarding human error, there j
,_s 2 are some models that people have developed the last 10
(' )
3 years that would help make this -- I mean what does work ,
I 4 environment mean? It can mean a lot of things. What does I l
5 procedures, is it that the substance is wrong or they are 6 written in a bad way so the operators make mistakes? Why 7 is fatigue and stress work load at the same level as 8 management? I mean these are the kinds of questions. I 9 mean obviously this is an empirical list.
l 10 MR. TRAGER: These are just things that -- l 11 MEMBER APOSTOLAKIS: That have popped up in 12 the past.
13 MR. TRAGER: This is just a summary. j n\~/ '
14 MEMBER APOSTOLAKIS: Yes.
1 15 MR. TRAGER: But actually the coding sheet l 16 that they use is a lot more detailed. The people that do 17 the coding do it -- you know, they have some training in 18 the use of this HPIP process. So a lot of the things that 19 maybe will come out a little bit -- part of the problem I 20 think is in coding any event is that the -- we can only --
21 the information that goes in the data base is what's 22 available in the report. If it's not in the report -- the 23 licensee event reports many times have the bare minimum 21 information, which can be a problem.
,~,
( ) 25 MEMBER APOSTOLAKIS: Which brings us back to NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N W.
(202) 234-4433 WASHINGTON. D C. 20005-3701 (202) 234-4433
87 1 Mr. Carroll's point earlier that maybe we should tell them
- 2 what to put in the report. Right? At some point, we have 3 to go back and tell them this is not enough information.
4 MR. TRAGER: The inter-office working group 5 that worked on the data base, I guess that was a proposal 6 back in December of 1993, was to perhaps issue additional 7 guidance on NUREG 1022 on what goes into a data base.
8 That's something that may someday be done.
9 MR. CARROLL: The problem, Jack probably can 10 talk about it better than I can, is the problem of 11 backfit. Just trying to make the case that can be -- can 12 and should be done. Providing additional --
13 MR. ROSENTHAL: Okay. 50.73 and 50.72, but
(_- 14 50.73 does require you, it's only a sentence or two long, )
15 to report human performance information.
16 NUREG 1022 is a whole NUREG of guidance in 17 preparing LERs. Changing NUREG 1022 to require the 18 licensees to report more information than they have in the 19 past would be a backfit. In order to justify that 20 backfit, you have to have a reason, a use, motivation for 21 data.
22 We thought that by constructing this data 23 base, bring some studies from it, we could show both the 24 value of human performance information and that it wasn't m
) 25 coming in an LERs, but that things like AITs and the NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N W (202) 234-4433 WASHINGTON. D C. 200C5-3701 (202) 234-4433
88 1 special team inspections from NRR were richer sources.
- . , 2 That would then provide a basis to go forward. We just
( \
3 need to go through those steps one at a time. l 4 MR. CARROLL: We've been working this for 5 three years now. Is that right? If I come back three 6 years from now, what can I expect to see?
7 MR. TRAGER: Hopefully in a couple of months 8 we'll know mort. We have the data base available now. I 9 think it's pretty good form for it to be able to say some 10 things about the events and information on the events.
11 The other type of information that's included 12 in each record, each record consists of 54 records. The 13 information fields you saw earlier, but it also includes :
-- 14 other information on the types of personnel that were 15 involved, their training, their shifts, rotation, see if 16 we can say anything about that.
17 MR. CARROLL: That viewgraph seems to have a 18 blame the operator mindset to it.
19 MEMBER APOSTOI AKIS: Have what?
20 MR. CARROLL: Blame the operator.
21 MEMBER APOSTOLAKIS: Oh yes. In fact, you 22 know what I was doing in my mind? I took this figure from 23 Reason's book. I was trying to see how these things fit.
24 That's on page eight of the handout. Figure 79.
,' ~x k ,) 25 MR. SINGH: Gene.
NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N W.
(202) 234-4433 WASHINGTON, D C. 20005-3701 (202) 234-4433
89 1 MEMBER APOSTOLAKIS: On page eight, l 1
'- 1
-~3 2 handwritten eight There are two figures there. The top
\) 3 one deals with a classification of unsafe acts themselves. l i
4 It would be interesting to take some of these events and 5 see where they fit in here. In fact, there is a NUREG ll 6 that shows that most of the major incidents in nuclear 7 power were violations at the very bottom, which I found I 8 extremely interesting. They didn't call it a violation, 9 they called it circumvention.
10 If we go down though, I think you know, look 11 at the sequence here. You have the unsafe act, then you i
12 have the immediate psychological precursors. Then you j
p 13 have line management deficiencies. l t \
s 14 Now, Mr. Carroll just said that these seem to 15 be centered around a box, psychological precursors of l 16 unsafe acts. In other words, the operator's fault. It's 17 not always the case.
18 So you know, these are very simple ideas here.
l 19 These are -- we're not talking about Nobel Prize type, i 20 It's very easy for me to remember a simple figure like 21 that.
22 MR. TRAGER: Sure.
23 MEMBER APOSTOLAKIS: Than to remember, you 24 know, 12 bullets that really don't seem to have any
'% / 25 logical connection. The question now is, does this NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N W.
(202) 234 4433 WASHINGTON. D C. 20005-3701 (202) 234-4433
90 j l
1 evidence support what Reason says here? Do we learn l l
1 7- 2 anything from it? Do we spend more time on line i
> l
\_)i 3 management deficiencies, or is it really the psychological l
l l
l 4 precursors that really matter here? I don't know, what is i 5 the message? j 6 MR. TRAGER: Well, we have some viewgraphs of 7 the simple things we have done with some of the data that 8 will I think show the value --
9 MEMBER APOSTOLAKIS: Okay.
10 ACTING CHAIRMAN SEALE: A specific question l
11 though with regard to whether or not this particular 12 process is, if you will, predisposed to blame the l 13 operator, when you say numbers, titles, and types of I l') l
~/ 14 personnel, how far up the organizational ladder do you 15 crawl?
l 16 MR. TRAGER: Anybody directly involved in the l
17 event. That will probably go as high as maybe a shift 18 engineer or an operator, or there's evidence in the report 19 that it was a higher level management person.
20 ACTING CHAIRMAN SEALE: So occasionally you l
! 21 might include the shift supervisor?
l 22 MR. TRAGER: Yes.
23 ACTING CHAIRMAN SEALE: Yes, I agree with you, 24 George.
/~
(_)T 25 MR. TRAGER: We're at the point now where we NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 (202) 2344433
91 1 just have raw material to work with.
fs 2 ACTING CHAIRMAN SEALE: I understand. I k')
3 understand, but --
l 4 MR. ROSENTHAL: We used -- I know that Gene 5 read Dr. Reason's book, and then he gave it to me to read.
1 l
6 Then we did use Dr. Reason as a consultant to us in NUREG-l l
7 1275, volume 8. I think, yet we hear the message. We're l l
8 going to have to go to some structure. l 9 MEMBER APOSTOLAKIS: It's just that it makes 10 it so much easier to look at simple models like that.
11 We're not talking about a revolution. In another context 12 we keep referring to a revolution with a small R. That's 13 the risk-informed performance-based regulatory effort. So (o)
\_/ 14 maybe we can have something like that here. It's probably i 15 an evolution here, not even revolution. But going to 16 something as simple as that. l 17 For example, I don't think Rasmussen's figure 18 is simple, on page seven, even though it contains probably 19 the same amount of information. But it's not something i
20 that's easy to remember. So I think mental models are 21 very important. You guys are experts and know this 22 anyway. So please.
23 MR. TRAGER: So that's the type of information 24 in the data base. Briefly, we had a process for getting
(_,) 25 the information into the data base, for developing the NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 (202) 234-4433
i 92 l I
1 data base, and for coding the events. We're having a l
- 2 report issued toward the end of this month on this process 3 also, and on the data base.
4 MEMBER APOSTOLAKIS: I would like to get a 5 copy of that, by the way.
6 ACTING CHAIRMAN SEALE: Who are your 7 independent reviewers? Have you identified those yet?
8 MR. TRAGER: The independent reviewers 9 actually as part of the process, it would be two people 10 who would code each event.
11 ACTING CHAIRMAN SEALE: Okay. Okay. You are 12 talking about people who are embedded in the --
13 MR. TRAGER: As part of the process, coding
(.
14 process.
15 ACTING CHAIRMAN SEALE: I beg your pardon.
16 Okay.
17 MR. CARROLL: How about the third party?
18 MR. TRAGER: The third party independent 19 review is you have two people doing the coding 20 independently, manual input of data. They get together, 21 they do the joint consensus on what should be in the 22 record. Then you have a review by a person, say an 23 operator, a person with operator experience. He'll look
. 24 at it and do a reality check on the content of the record.
() 25 MR. CARROLL: All of these people are from NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.
(202) 2344433 WASHINGTON, D.C. 20005-3701 (202) 234-4433
93 1 INEL?
ry 2 MR. TRAGER: Yes. They are.
(',)
3 To test out the data bases, it's still very 4 early. We developed just a couple simple questions, you 5 know, the kinds of things you might ask. These are some 6 of the sample questions we tried to get answers to. I'll 7 show you some of the results we got.
8 The first is simply a count of events -- and 9 this is a sample, let me stop and say that. This is not -
10 - this does not include the 160 records, but 40 AITs and 11 the 20 human performance studies.
12 As you can see from the distribution, events 13 happen all times of the day and more frequently during the w- 14 day as you expect, more opportunities for error, during 15 day shift.
16 ACTING CHAIRMAN SEALE: But those 4:00 in the 17 morning ones are the doozies.
18 MR. CARROLL: They are the bad ones, but --
19 MEMBER APOSTOLAKIS: Between at 8:00, you had 20 10?
l l
l 21 MR. CARROLL: Between 8:00 and 10:00.
j 22 MR. TRAGER: Yes. Between 8:00 and 10:00.
23 MEMBER APOSTOLAKIS: Why is that so?
24 MR. CARROLL: Because they sent people off O
( ,) 25 doing maintenance.
NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N W.
(202) 234 4433 WASHINGTON. D C. 20005-3701 (202) 234-4433
94 1 MR. TRAGER: They haven't had coffee yet.
em 2 MEMBER APOSTOLAKIS: The number of (s/ 3 opportunities.
4 ACTING CHAIRMAN SEALE: The number of eyeballs 5 went up.
6 MR. TRAGER: And this is the distribution of 7 different contributing factors for AITs and for human l l
8 performance studies.
9 MR. CARROLL: Now why is this a different 10 categorization than what Cecil showed us that NRR is 11 doing? Do you have a different notion of what's important l
12 to human events?
l 13 MR. THOMAS: Yes. The categories are l (~'\
\- 14 different. I notice one of theirs is human factors. In )
! 15 ours, we consider all of them human factors.
16 Ours consider all LERs and all inspection 17 reports, so that there's much more data. The periods of 18 time are different so that might make -- that will account 19 for something. But actually the significant categories 20 are the same. Procedures are the biggest in both NRR's 21 and AEOD's. M&O is big. Communications is big, and l 22 training is big. So probably the major contributors 23 aren't significantly different.
24 MR. CARROLL: And these overlap. In other l')
'\ _,/ 25 words, procedures is 95 percent. The bases involve NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N W.
(202) 2344433 WASHINGTON, D.C. 20005-3701 (202) 234-4433
95 1 procedures, but the same case can also involve -- it
- 2 doesn't add up to 100 percent.
J 3 MR. TRAGER: This too is sorted by what was 4 found during AITs. So unfortunately, there's two 5 different types of percentages displayed in this. But you 6 see it increases. The work environment is a shorthand.
7 The human factors is supposed to stand for human system 8 interface, human factors engineering. That label covers a 9 lot of different things.
10 MEMBER APOSTOLAKIS: In fact, all the 11 categories are human factors, aren't they?
12 MEMBER MILLER: Could you give me a few 13 examples of what that might be?
/^T k- 14 Now we have a different categorization than we 15 had before. I agree. What does that mean?
16 MR. CARROLL: The question is what does human 17 factors mean?
18 MEMBER MILLER: In this context.
19 MEMBER APOSTOLAKIS: And how is it different, 20 for example, from individual personnel factors?
21 MR. TRAGER: Well, something like the layout 22 of a control panel, for example, was a human system l
l 23 interface.
24 MR. ROSENTHAL: The tigon tubing that you
() 25 couldn't read on the draindown event.
NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N W.
(202) 234-4433 WASHINGTON, O C. 20005-3701 (202) 234-4433
96 1 MR. TRAGER: Right. Yes.
(_,
2 MEMBER MILLER: Okay.
\#
3 MEMBER APOSTOLAKIS: Wouldn't that be also a l
4 control room system design?
5 MR. TRAGER: The other thing I guess I ought 6 to mention here is when the events are coded from reports, j i
7 we gave INEL the instruction that do"'t read anything into 8 the event. In other words, only include precisely what is 9 stated in the report. ;
10 MEMBER MILLER: I guess I am confused what 11 that means.
12 MR. TRAGER: Well, if for example, if in an i 13 LER, for example, the root cause is found to be one j
,m.
(V )
14 specific thing, say an operator made a mistake, then that l
j 15 LER would just have in this data base, would have one ;
16 count. That would be the --
17 MR. CARROLL: Individual personnel.
18 MR. TRAGER: Factors.
19 MR. ROSSI: But if you had an inspection 20 report that came to a different conclusion, like an AIT 21 that had a different route cause or another route cause, 22 that would also go under.
i l 23 MR. TRAGER: Yes it would. It might say well, 24 we're not sure. The guy made a mistake, but it was cold, g/ s 25 the lighting was bad. These are two environment factors.
NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N W.
(202) 234-4433 WASHINGTON D.C. 20005-3701 (202) 234 4433 I
97 1 We're not sure whether he was adequately trained. We 2 don't know whether he had a mindset problem, which might 7-t )
'~#
3 be some other personal factor. There's a lot of things 4 that could possibly be involved.
5 Of course the more the human performance 6 studies that we did, having a protocol as we did, asking a 7 lot of questions, and interviews with the operators and 8 people who are involved, we found out a lot more. So 9 that's why I think you tend to see the higher percentages 10 for those.
11 The AITs have a charter of things that have to 12 be covered by the team. A lot of times the charter items 13 have to be addressed and nothing else.
(g
-. 14 In more recent years, the charters of AITs 15 should include a question about the human factors aspects, 16 the human performance during the event. But that was not 17 true years ago.
18 ACTING CHAIRMAN SEALE: If you look at those 19 bars, it appears that there's a factor of two roughly 20 biased between the HPS set and the AIT set on everything 21 except management and organization and system design and 22 configuration.
23 MEMBER MILLER: And procedures.
24 ACTING CHAIRMAN SEALE: No. No. I'm saying 25 that in every case but those two, the HSP spreads the i
NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS l
1323 RHODE ISLAND AVE., N W.
l (202) 234-4433 WASHINGTON D C. 20005-3701 (202) 234-4433 l
I.
l 98 1 blame more broadly. And in the case of management and
- 2 organization, and system and design and configuration, the 7-s
'~
3 AITs did come up with larger percentages. That may be 4 part of this predisposition you talked about in terms of 5 the way the AITs were set up.
6 MR. TRAGER: The AITs are usually -- they have 7 a charter.
1 8 ACTING CHAIRMAN SEALE: That's what I mean.
9 MR. TRAGER: If you go outside of the charter, i
10 that may not be time well spent. There's a lot to cover I 11 in an AIT.
12 ACTING CHAIRMAN SEALE: Well, that's an 13 interesting question because here you have got two
'r3 l
'\/ 14 different ways of slicing the same loaf of bread. As I l i
15 say, the appearance is that the HPS approach looks to be l 16 wanting to spread the blame around in the sense that 17 everybody is involved, almost everyone is involved in at 18 least half of the events.
19 MR. TRAGER: One thing that does not come out 20 here that's -- well it's not obvious, but the events that 21 we did human performance studies for, they were events we 22 knew in advance. We're kind of complexed that they did i
l l 23 have human performance problems. So it's not really fair s
24 to compare the two.
C'
(,,,) 25 ACTING CHAIRMAN SEALE: Okay.
NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE , N W.
(202) 234-4433 WASHINGTON, D C. 20005-3701 (202) 234-4433
99 1 MR. TRAGER: AITs may have involved equipment j l
,3 2 failures also. So really it's not fair to compare. I i )
%./
3 ACTING CHAIRMAN SEALE: Well, it's important ,
I 4 that you be prepared to tell spectators like me that 5 that's part of the problem that's in that.
6 MR. TRAGER: The human performance studies, we 7 went and studied events where it appeared on the surface j 8 there were no problems at all. It wasn't until we started 9 asking questions that it turned out that the performance 10 wasn't as good as it appeared to be on the surface, that 11 there were some problems.
12 MR. CARROLL: And if I recall correctly, there 13 were at least two of the events out of the what, 20 some ,
(-
'- 14 odd -- i 15 MR. TRAGER: Twenty.
16 MR. CARROLL: Where it was clear to me at 17 least that there should have been an AIT and there wasn't.
18 ACTING CHAIRMAN SEALE: Yes.
19 MR. CARROLL: After people started looking, 20 looking under the bed.
21 MEMBER APOSTOLAKIS: I'm curious about the 22 procedures. I read here in Reason's model that there are 23 two ways you can make a mistake. Misapplying a good l
l 24 procedure or applying a bad procedure. Which one seems to x
(_) 25 dominate here? Obviously the causes are different. I NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 (202) 234-4433
l 100 1 mean if the procedure is bad, then you can't blame the
- 2 poor operator. But if he is misapplying a good procedure, 3 then you can blame him or her.
4 So when we have there a number of procedure 5 related issues, which one is which?
I 6 MEMBER MILLER: Which is a dominant issue? l l
7 MEMBER APOSTOLAKIS: Yes.
l 8 MR. TRAGER: We'll have to -- there are 9 certain things we can tell from this data, but you really 10 have to get into the details of the data base to say more 11 than superficial things.
12 MEMBER APOSTOLAKIS: And you can do that?
13 MR. TRAGER: We can.
( ~' ,
\-s'\ 14 MEMBER APOSTOLAKIS: The details are there? !
15 So all you need is a model like this one to do that. )
16 MR. TRAGER: All we need is a model.
17 ACTING CHAIRMAN SEALE: And it seems to me 18 that's a good example --
l 19 MR. CARROLL: Go ahead.
20 ACTING CHAIRMAN SEALE: That's a good example 21 of where you really need from the very beginning to make 22 that distinction, because I mean that's pretty 23 fundamental.
24 MR. ROSENTHAL: It's in there. We are
/ \
t, ,) 25 actually planning to put out the data base, make it NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N W.
(202) 234-4433 WASHINGTON. D C. 20005-3701 (202) 234-4433
101 j 1 available throughout the agency and to others by the end l
2 of the next couple of months. Then as Gene said, this is t7-s) 3 just to give you an idea of what kind of information is in 4 the data base. This word superficial is I think well 5 takan. The plan would be to then slice and dice this 6 information in far finer detail, with the first report 7 coming out this spring and then we would take on not just 8 calling it M&O, but just what are we really talking about.
9 Not just procedures, but what aspects of procedures.
10 MR. CARROLL: My intuition, George, tells me 11 that it's probably bad procedures.
12 MEMBER MILLER: Yes. I would say.
13 MR. CARROLL: Because I've never been involved l O
\s l 14 in an incident where procedures were a factor where you 15 couldn't figure out some way to make them better. l 16 MEMBER APOSTOLAKIS: Now bad procedures in 17 terms of content or presentation?
18 MR. CARROLL: Both.
19 MR. THOMAS: To share with you NRR's results, 20 we've as I mentioned earlier, we analyze our human factors 21 information system data in support of the senior 22 management meeting pre-briefs. We do break ours down into 23 two parts. It turns out they are approximately equal as l 24 far as contributors go. The procedures themselves r~N tv) 25 contribute to about half of the events, and misapplication l
NEAL R. GROSS
, COURT REPORTERS AND TRANSCRIBERS I 1323 RHODE ISLAND AVE., N W.
l (202) 234-4433 WASHINGTON. D C. 20005-3701 (202) 234-4433 l
102 1 of the procedures contribute to about half.
2 I guess at first blush you would think maybe i,,,)
3 it would be misapplication, but when you look at 4 procedures, the procedures we look at include not only 5 EOPS, but also normal operating procedures, abnormal 6 procedures and emergency operating procedures. The 7 operating procedures and abnormal operating procedures are 8 not developed with the rigor that emergency operating 9 procedures are. You find many contributors in those two 10 categories.
11 So when you look at the final results, it's 12 approximately equal between misapplication of procedures 13 in some way and some problem with the procedures.
ry k- 14 MEMBER MILLER: That surprises me, actually.
15 MEMBER APOSTOLAKIS: It does?
16 MEMBER MILLER: I'd agree with Jay, based on 17 my experience. It would be poorly written procedures that i
18 would dominate over misapplication.
19 MEMBER APOSTOLAKIS: On the other hand, the 20 experience from aviation is that shortcuts really, it's 21 not the fault of the procedures that pilots don't follow 22 the procedures and checklists and all that. So many times 23 they find it very boring and they just glance at it and 24 they say it's fine.
(3 l
y ) 25 MR. TRAGER: That's Jim Reason's violation, as NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHoDE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON. D.C. 20005-3701 (202) 234-4433
103 1 you know a better way. It's not --
,- 2 MEMBER APOSTOLAKIS: That's a violation, yes.
( )
'~'
3 But since you don't have violation here, probably it's 4 hidden there in the procedures.
5 MEMBER MILLER: So you might have operatorn 6 that are doing like pilots aze. doing.
7 MEMBER APOSTOLAKIS: Yes, which brings me to 8 another question.
9 MEMBER MILLER: That should be a third i 10 category.
11 MEMBER APOSTOLAKIS: The issue of formal 12 versus informal culture. There is a growing realization 13 that what we have on paper and what actually is happening 13
-- ) 14 at the plants are not always the same. The pilots again, i 15 the studies in aviation have confirmed that. That they I l
l 16 don't really fcllow every single checklist and every item. 1 17 They don't confirm and so on because it takes too much ,
i 18 time and they don't think it's worth it. l l
19 Is there any effort within our agency to try 20 to see whether there's a significant difference between 21 what we think is happening and what is actually happening 22 there? Do the operators take shortcuts there? Do we know 23 about them?
24 MR. SINGH: George, I think most of the time
()
in 25 procedures might be good, but there's inadequate training.
NEAd. R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N W.
(202) 234-4433 WASHINGTON, D C. 20005-3701 (202) 234-4433
1 104' Sometimes the operators have not been tr i a ned on those
! O 3 2
particular procedures.
}
MEMBER APOSTOLAKIS:
Well, that's another 4 issue.
I'm talking about intentional. i 5 I mean that's 6
related to this violation business whichcalled we have circumvention now.
How many times do they circumvent 7 things?
8 ACTING CHAIRMAN SEALE: Well, let me ask the 9
question in a little bit different way .
You go through 10 and you look at an LER and you try to understand what 11 happened in the case of that specific event.
12 Is there any information that would tell you how many times that 13 particular procedure had been invoked previous t 14 o o that ccasion successfully?
15 You know that's really the culture you are 16 t alking about, because what has happened is that the 17 o perators have learned how to do it 18 procedure.
right in spite of the 19 i
It's only this neophyte maybe who doesn't h ave the right senior operator looking over his shoulder or 20 w hatever, the guy that does it by the book that runs 21 h imself into the ground.
22 MEMBER MILLER:
23 And the other is there's certain procedures I'm certain are very high vi sibility on 24 the training simulator.
25 ACTING CHAIRMAN SEALE: Sure.
NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N W.
1 (202) 234-4433 WASHINGTON, D C. 20005 3701 (202) 234-4433
--~ _
104 1 Sometimes the operators have not been trained on those 2 particular procedures.
3 MEMBER APOSTOLAKIS: Well, that's another 4 issue. I'm talking about intentional. I mean that's 5 related to this violation business which we have called l
6 circumvention now. How many times do they circumvent 7 things? i i
8 ACTING CHAIRMAN SEALE: Well, let me ask the 9 question in a little bit different way. You go through 10 and you look at an LER and you try to understand what I l
11 happened in the case of that specific event. Is there any 1
12 information that would tell you how many times that 13 particular procedure had been invoked previous to that k-) 14 occasion successfully?
15 You know that's really the culture you are 16 talking about, because what has happened is that the 17 operators have learned how to do it right in spite of the l 18 procedure. It's only this neophyte maybe who doesn't have 19 the right senior operator looking over his shoulder or 20 whatever, the guy that does it by the book that runs 21 himself into the ground.
22 MEMBER MILLER: And the other is there's 23 certain procedures I'm certain are very high visibility on 24 the training simulator.
O)
(_, 25 ACTING CHAIRMAN SEALE: Sure.
1
\
NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N W.
(202) 234-4433 WASHINGTON, D C. 20005-3701 (202) 234-4433 l
105 1 MEMBER MILLER: Certain procedures probably l
2 rarely get a glance by the simulator work in training.
3 3 ACTING CHAIRMAN SEALE: So what you are really 4 talking about is the internal process within the utility i 1
5 to identify the procedures that don't really work, and do 6 something about them rather than working around them and 7 ignoring them. You don't have that in your LERs or any 8 other process.
l 9 MR. CARROLL: Is that a management, an 10 organi::ational f ailure?
11 ACTING CHAIRMAN SEALE: Sure. Sure it is.
12 MEMBER MILLER: Senior shift supervisor should 13 recognize that.
/~N ws 14 MR. THOMAS: But again, don't focus just on 15 the operators, because when I talked about the equivalency 16 or roughly equivalency in carrying out procedures versus 17 the procedures, I am including all procedures, maintenance 4
18 procedures and so on, outside the control room.
19 ACTING CHAIRMAN SEALE: I appreciate that.
20 Sure. Operator with a little O.
21 MEMBER APOSTOLAKIS: But even there you can 22 have certain --
23 MR. ROSENTHAL: The only way you are going to 24 know that is to leave Washington, and either in the form
()
( ,j 25 of an AIT or an HPS, or a special team inspection, is to NEAL R. GROSS COURT REPORTERS AND TRANSCRIDERS 1323 RHODE ISt.AND AVE., N V.'.
(202) 234-4433 WASHINGTON. D C. 20005-3701 (202) 234-4433
I 106 1 go there and interview as quickly as possible real people.
2 MEMBER APOSTOLAKIS: And it's not easy.
(h i
'~'
/
! 3 MR. ROSENTHAL: And those are the events which 4 tend to be the more safety significant events, by the way, 5 that we're trying to focus on.
l 6 ACTING CHAIRMAN SEALE: But that's the sort of l
7 thing you might try to get the resident inspectors to be a 8 little bit more attuned to. That is, to ask those kinds 9 of out of the box questions after an LER comes in.
10 MR. ROSENTHAL: Well, at the time, we have, 11 it's been slowly, but it's happening. RES has presented 12 the human performance investigation procedure process at 13 the regions, and we're slowly getting the regions to use C
( 14 these techniques.
15 MEMBER APOSTOLAKIS: Maybe another way of 16 approaching this is to divorce it from an individual 17 incident, because people may be reluctant to talk about 18 it, and do a more generic study where you do interview 19 operators, people who know, and ask them these specific 20 questions that you think about in advance. Because it 21 seems to me that this is a very important subject.
22 I mean if there is a significant difference 23 between what we think is happening and what is actually l
l 24 happening, we ought to know about it. It seems that j
pj s 25 people who have looked, like the French for example, have l
NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS l 1323 RHODE ISt.AND AVE.. N W l (202) 234-4433 WASHINGTON D C. 20005-3701 (202) 234-4433 l
1 107 1 concluded yes, that there is a difference. They don't
,s 2 know whether it's significant yet, but they can not i \
x
/
3 conclude that it's insignificant either.
4 So maybe there are ways of approaching it, 5 Jack, so that people will be freer to talk about it, you l
6 know, without -- because that's certainly a subject that I
7 is not usually discussed.
8 ACTING CHAIRMAN SEALE: Sure.
l I
9 MR. CARROLL: Don't forget, George, that the 10 resident inspectors are coming in on cff shifts. They are 11 better trained to look at that kind of thing.
12 MEMBER APOSTOLAKIS: Yes. That's one way. I 13 mean maybe a workshop, for example, or some other forum, g
k-) 14 where people will not feel inhibited.
15 MR. CARROLL: Before we leave this viewgraph, 16 I would strongly suggest that the category human factors 17 be better defined, for any confusion.
18 MR. TRAGER: Yes. It's in an earlier 19 viewgraph, it should be here, but to get it on the --
20 there was a shortcut. But it's human system interface, 21 human factors engineering. That covers a lot of ground.
22 MEMBER MILLER: So it's basically mismatched l
23 between the system and the human. Is that what it is?
24 MR. TRAGER: Yes, and other human factors, C\
s q ,) 25 engineering problems.
NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N W-(202) 234 4433 WASHINGTON, D C. 20005-3701 (202) 2344433
108 1 MEMBER APOSTOLAKIS: Now at the last
,g 2 subcommittee meeting, the question of the accuracy of the
()
3 information that is displayed in the control room was 4 raised. Where is that here? When we say man-machine 5 interaction, what do we mean? Is there such a thing 6 there? Supervision, command and control, perhaps? No, it 7 can't be there.
8 MR. TRAGER: That would be a human factors, I
9 and a display, for example. If it wasn't clear, say you l
10 have a bunch of enunciators that go off following a 11 reactor trip and you don't know what caused it.
12 MEMBER APOSTOLAKIS: But why would that be a 13 human factors problem? Maybe it's a problem with the
(~h kY 14 engineers who designed the system.
1 15 MR. CARROLL: Well that's included in what I
16 they mean by human factors. i 1
17 MEMBER APOSTOLAKIS: It should not. In my l
18 opinion it should not be included. These are different !
l 19 humans. These are not the humans in the plant.
- 20 MR. TRAGER
- It's a human factors problem.
21 For example, if the operator didn't know what step to take 22 next --
23 MEMBER APOSTOLAKIS: But he didn't know why?
24 Because he didn't have the appropriate information. Not O
' () 25 because he himself didn't know. That's what I'm saying.
NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N W.
(202) 234-4433 WASHINGTON, D C. 20005-3701 (202) 234-4433
l l 109 l
l 1 That it was somebody else's fault.
2 MR. CARROLL: Isn't that in the category of i (7'~x) l 3 human factors as you were describing it?
4 MEMBER MILLER: No. It shouldn't be.
5 MR. ROSENTHAL: Dr. Apostolakis, I read the 1
6 transcript last night, Ivan catton's questions.
7 MEMBER APOSTOLAKIS: Yes. That he raised in 8 the context of simulators, yes.
9 MR. ROSENTHAL: Things like a draindown event 10 with bad tigon tubing and given the wrong level to the 11 operator, so he has the wrong information. It's an 12 instrument problem, it would have been in this man-machine 13 human factor interface.
O) i N' 14 MEMBER APOSTOLAKIS: Where is that?
15 MR. ROSENTHAL: Well, what we've labeled human ;
16 factor in which we have already said we will relabel.
17 Now Ivan was interested in, as I understand 18 it, in a different sort of thing. That is, that because 19 you don't' understand the thermal hydraulic phenomenology, 20 then your simulator is wrong and the information given to 21 the operator is wrong.
22 I was trying to -- the best example I can l 23 think of of his type concern would be let's say a boiler 24 ATWS where you expect because he tripped a reactor, the
() The
(_,/ 25 recirc pumps, to fall into an area of instability.
NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N W.
(202) 234-4433 WASHINGTON, D C. 20005-3701 (202) 234-4433
110 1 cold water level would be oscillating. All of the
,- 2 operators training would have been on a simulator. It
\ /
3 would show the constant water level. Here the water level 4 is showing up and down. I think that that was more like 5 Dr. Catton's concern.
6 MEMBER MILLER: That was a level?
7 MR. ROSENTHAL: In this case, it would be j 8 water level, where you have trained the man, in fact, you 9 tell him to reduce the cold water level, and then the most j i
10 recent guidance to raise it. But you have told him to l 11 take an action based on an instrument which he thinks in 12 his training will be relatively stable. He can see where l l
13 he's at, where he's supposed to go. Because of the !
r~% i I )
k/ 14 underlying TH phenomenology, in fact, that level is j l
15 oscillating around. That's clearly not in here. We don't 16 have examples like that from operating experience.
17 I told Mr. Carroll that on the break, if you 18 think about the sparsity of data, one event that we did 19 investigate was a stuck open pressurizer safety vaive at 20 Fort Calhoun. This is the best integral test, post-TMI, 21 that I know of, in which the operators used their 22 emergency operating procedures. They used RVLS. They 23 used sub-cooling. They used their SPDS. They did it 24 right. But if we only have one like that, one event of A)
( 25 that severity, where you could see them perform.
NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N W.
(202) 234-4433 WASHINGTON D C. 20005-3701 (202) 234-4433
111 1 ACTING CHAIRMAN SEALE: I think the other 2 concern that Dr. Catton had was just the general queaion
(, ,)
3 of do the operators know when the simulator ceases to 4 track the plant. So you just don't use the simulator any 5 more. You go to other say emergency procedure kinds of 6 diagnostics, without believing the simulator.
7 MR. THOMAS: I'd like to defer that question 8 number 16 was question that the subcommittee raised.
9 Frank Collins of the Operator Licensing Branch is prepared 10 to talk to that this afternoon.
11 MEMBER APOSTOLAKIS: Now at the beginning of 4
12 our meeting today, I asked you for a definition of human 13 factors. Now these human factors here is not the same as b 14 the human factors of two hours ago.
15 MR. CARROLL: No. Why don't you make 16 everybody clear on it. Why don't you tell us how you're 17 going to relabel.
l 18 MEMBER APOSTOLAKIS: I think they should just 19 throw it away and use a model, where the definitions are 20 clear, the relationships are clear. I don't think it's a 21 matter of just renaming. I mean we have things, 22 management, and then work environment, and then 23 communications. I do not see any logic in it. It's just 24 things that people thought were important, and they are.
f 25 I mean they are important, but whare is the logic.
NEAL R. Gi!OSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 (202) 234-4433
112 1 It is a more fundamental problem than just s 2 relabeling something.
(s )
i
'~
3 MR. CARROLL: No. But I think there is some 4 confusion.
5 MEMBER APOSTOLAKIS: In more places than one.
6 MR. TRAGER: The whole reason for this data 7 base, if I'm thinking back correctly on this, was that 1
8 everybody agreed there's not enough human performance l 9 information in reports, in LERs, in any kind of report.
10 But you can't get two experts to agree on what should be 1
i 11 in it.
12 So a goal I think of developing this data base 13 was to get a consensus on the kinds of information that's i
l \
r^s
\--) 14 important to know about human performance during events.
l 15 That's what the attempt is.
l 16 MEMBER APOSTOLAKIS: And what I'm saying is 17 that the next natural step here would be now to look at 18 some of the models that people have proposed. The 19 disagreements now are not great any more. Then see how l
20 you can merge the two. Tat's all I'm saying.
I 21 MR. ROSENTHAL: Yes, and related to risk.
l 22 MEMBER APOSTOLAKIS: Yes. You expect me to 23 say no?
24 (Laughter.)
25 MR. CARROLL: And you'll have all that done NEAL R. GROSS ,
COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N W.
(202) 234-4433 WASHINGTON, D C. 20005-3701 (202) 234-4433
I 113 1 before the next subcommittee meeting.
-- 2 So how are you going to or how would you L) 3 propose to change the presently labeled human factors?
4 MR. TRAGER: Human factors could be human 5 system interface / human factors engineering.
1 l
6 MR. ROSSI: I think this is something we'll !
l 7 have to go back and think about and look at. I don't know l 8 that we can tell you exactly how to change it here or l
9 break it down.
10 MEMBER APOSTOLAKIS: I agree.
11 MR. ROSSI: I would like to say a couple of 12 other things about this whole area in the data base. That 13 is, that the ]ast several years, as I'm sure you know, I
\ ') 14 we've been working on the reliability, availability data 15 rule or as an alternative just to collect the information 16 voluntarily. That's a much much more simple kind of thing 17 to do because it's equipment. What you need is the 18 failures of the equipment. Then you need a denominator.
19 We have had a lot of difficulty in just 20 agreeing to definitions on when do you consider it a 21 failure and how do we get the denominator. Here this is 22 much more complicated. We just have not gotten that far 23 yet. So I think we're going to have to look carefully at 24 all this. It's going to evolve, I think, as you have A)
(_ 25 pointed out.
NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 (202) 234-4433
114 i
- 1 Many of the questions you are asking I think '
i l gw 2 have to do with what's the denominator. Somebody used a
~
3 procedure and they make an error, other people must have 4 used it in the past for many many times. But how do we 5 get any kind of a measure at all how many times. These 6 things are much much more difficult than dealing with the I
7 equipment. We have had lots of problems just defining i
8 things, get reliability data of the equipment. l 9 MR. CARROLL: What is tue current status of 10 the reliability?
11 MR. ROSSI: I'll give you one sentence, okay?
12 We're working with the industry to try to get the 13 information voluntarily. We believe there's a high
,_ l f
\' 14 likelihood that that will happen. We're evaluating what i 15 we're getting from the industry. Once we're finished with 16 that, we'll decide whether voluntary approach will work.
17 It will probably be, I believe the schedule was something 18 like four or five months to evaluate the industry 19 approach, see whether that's going to work.
20 MR. CARROLL: Industry has come back and 21 proposed --
22 MR. ROSSI: Yes. They have, right.
23 MR. CARROLL: Good. I'd glad to hear that.
24 MEMBER APOSTOLAKIS: But I would like to point
()
(, 25 out though that it's not just an issue of the denominator.
NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 (202) 234 4433
115 1 I think that's an important issue, but that's not the only l 2 issue. I think another issue here is really a taxonomy of i,-s) i
(_/ 3 --
4 MR. ROSSI: Well, it's defining how many 5 numerators are there that you want denominators for.
6 That's a part of the problem too, the definition.
7 (Laughter.)
8 ACTING CHAIRMAN SEALE: And what is there 9 paired up denominators.
10 MEMBER APOSTOLAKIS: Okay.
11 ACTING CHAIRMAN SEALE: Okay. Moving on.
12 MR. TRAGER: This is a blow up of the same 13 thing, the same graph.
I
\/ 14 ACTING CHAIRMAN SEALE: Okay. Now that's a 15 column of numbers that needs a denominator.
16 MR. TRAGER: This is just, table 5 is the 17 numbers of AITs that occur during different modes. Mode 18 isn't completely correct. It's actually the sequence 19 coding and search system reactor status, which is not 20 exactly the same as the five modes. But you can see when 21 the events occurred during operations or shutdown.
22 ACTING CHAIRMAN SEALE: Now let me see if I 23 understand that. What you are saying is that 47 percent -
24 - 43 percent of all of those -- of the events that r
(_,)N 25 occurred that constituted an AIT entry, occurred at full NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHC .SLAND AVE.. N W.
(202) 234 4433 WASHINGTON, D.C. 20005-3701 (202) 2M-4433
116 1 power.
p- 2 MR. ROSENTHAL: Started.
3 ACTING CHAIRMAN SEALE: What?
4 MR. ROSENTHAL: Started.
5 ACTING CHAIRMAN SEALE: Started at full power.
6 That's what I mean.
7 MR. TRAGER: We have two labels here for power 8 operations. One is F and the other is E. So it's the sum 9 of those two.
10 ACTING CHAIRMAN SEALE: Yes. The load change, 11 so it's E and F.
I 12 MR. TRAGER: E and F.
13 ACTING CHAIRMAN SEALE: Yes.
O
\~I 14 MR. TRAGER: Forty eight percent, roughly 15 half.
16 ACTING CHAIRMAN SEALE: I think that Dr.
17 Powers, if he were here, would --
18 MEMBER MILLER: Yes. I was going to say.
19 ACTING CHAIRMAN SEALE: Would remind you that 20 that's part of the reason that he has concern about low 21 power or shutdown risk. This clearly demonstrates that 22 it's a significant part of the pie.
23 MEMBER MILLER: Yes. Clearly if I add up the 24 percents, more than half.
) 25 MEMBER APOSTOLAKIS: Yes. But again we don't NEAL R. GROSS I COURT REPORTERS AND TRANSCRIBERS l 1323 RHODE ISLAND AVE., N W.
(202) 234-4433 WASHINGTON, D C. 20005-3701 (202) 234-4433 l l
117 1 know how important these were.
2 MEMBER MILLER: I understand that, but inst
!7 s)
~
3 these events.
4 MR. TRAGER: Well, these are AITs, so they 5 were pretty important. ,
i ACTING CHAIRMAN SEALE:
6 That's the other part 7 of it. How much, you know. And if that's the case, then l
l 8 the non-power operations jump way up. You know, it's kind I l
9 of like airlines statistics. There are three numbers.
10 Takeoffs, landings, and per hour in flight. Until you j l
11 denominate -- until you put the denominator in, they are !
12 just numbers.
13 MEMBER APOSTOLAKIS: Takeoff and landing is --
Q
~) 14 MEMBER MILLER: Anybody can fly a plane when 15 it's level, but takeoff and landing is the hard part. j l
16 ACTING CHAIRMAN SEALE: Almost anybody, yes. ,
17 MEMBER APOSTOLAKIS: Maybe just a few people.
18 MR. CARROLL: I have a new one on my computer.
19 I am having all sorts of trouble flying it. I keep trying 20 to land on my aircraft carrier. I'm running into the 21 stern and the island.
22 MEMBER MILLER: Is that what you do when you 23 quit the ACRS? You play computer games?
24 ACTING CHAIRMAN SEALE: Well let's get back to fx ts,) 25 the subject. But seriously, the denominator is an NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N W.
(202) 234-4433 WASHINGTON, D C. 20005-3701 (202) 234-4433
118 1 important concern here.
f3 2 MR. TRAGER: It's kind of amazing, a quarter
/ )
\
'"'/
3 of the events happen during refueling.
4 ACTING CHAIRMAN SEALE: Yes, but only about 10 5 percent of the total time is spent during refueling.
6 MR. TRAGER: And these are AIT, so something 7 significant did happen.
1 8 MEMBER MILLER: You have 80 percent 9 availability factors, which are almost that. Eighty 10 percent of the time you are in power, 20 percent of the l
11 time you are not. Right?
12 ACTING CHAIRMAN SEALE: And about half of that 13 you are actually in the process of refueling.
(~\
- I 14 MEMBER APOSTOLAKIS: So I think we have 15 similar questions here.
16 ACTING CHAIRMAN SEALE: Yes.
17 MEMBER APOSTOLAKIS: Look at management and 18 organization now. Wow, 90 percent.
19 MR. CARROLL. It's low priority.
20 MEMBER MILLER: It would be interesting to see 21 what all falls into that category.
22 ACTING CHAIRMAN SEALE: What's percent PO?
23 MR. TRAGER: That's at power operation, events 24 at power operation. So that would be the combination of es
( ,s) 25 the two steady state and load change.
NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N W.
(202) 234-4433 WASHINGTON. D.C. 20005-3701 (202) 234-4433
119 1 ACTING CHAIRMAN SEALE: Okay. I beg your p 2 pardon.
N.) MEMBER APOSTOLAKIS:
3 Sc the dark line is cold 4 shutdown.
5 MR. TRAGER: This just shows the relative 6 proportions for which contributors were found to be.
7 Management and organization keeps popping up as being a 8 very high percentage.
9 MEMBER APOSTOLAKIS: This reminds me. There 10 was an interesting paper at the last PSA 96 meeting, where 11 some people from the Institute of Nuclear Energy Research 12 in Taiwan interviewed managers and technicians to see what
_ 13 their attitude was during outage. Interestingly enough, i\ ') 14 management thought that management was not important, and 15 all that really mattered was technical knowledge and being 16 able to do your job. The technicians felt that management 17 was extremely important in things such as ownership and so 18 on, was very important, which were completely dismissed by 19 management. They felt things like ownership and so on 20 were nothing.
21 So that's a very interesting study, in my 22 opinion. Maybe it explains a lot of this.
23 MR. CARROLL: It also depends on the 24 organization.
("
(,) 25 MEMBER APOSTOLAKIS: Yes. Well this is a NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 (202) 234-4433
120 1 specific organization.
2 ACTING CHAIRMAN SEALE: And the culture in
/-'s 3 which it is embedded.
4 MEMBER APOSTOLAKIS: Yes.
5 ACTING CHAIRMAN SEALE: i find it interesting l
i 6 that fatigue, stress, and work load only occurs during I
7 shutdown events and work environment is only important 8 during power operations.
i 9 MEMBER MILLER: Yes.
10 ACTING CHAIRMAN SEALE: That may be a taxonomy ;
i 11 problem.
i
! 12 MEMBER APOSTOLAKIS: And system design and i !
13 configuration is less important during shutdown?
r d
i
's / 14 MR. TRAGER: Well, it was just found by the -- I i
15 according to this data base, that was what --
16 MEMBER APOSTOLAKIS: I thought that was a 17 major problem with shutdown, wasn't it? Things are out of 18 service -- l 19 MR. ROSENTHAL: Configuration management is a i
20 shutdown issue.
21 MEMBER APOSTOLAKIS: Yes, but what does it l 22 show here though? It shows that what is it, that you have 23 more events occurring during power operation that have to 24 do with configuration.
( 25 MEMBER MILLER: That certainly goes against NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N W.
(202) 234-4433 WASHINGTON, D C. 20005-3701 (202) 234-4433
121 1
1 intuition.
j 2 MEMBER APOSTOLAKIS: Is it because you are 7-3 more careful when you shutdown? Because that's the main 4 reason why people want to see more detailed risk 5 assessments for shutdown operations.
)
l 6 MR. TRAGER: When we start doing things with 7 this data, I think we might be able to make some more 8 interesting observations about at least what's in the data -
l l
9 base anyway.
10 MEMBER APOSTOLAKIS: That's kind of strange, 11 isn't it?
12 ACTING CHAIRMAN SEALE: Yes. Kind of strange.
13 MEMBER MILLER: Counter intuitive. l
,.-~\ \
\ -)
14 MR. TRAGER: Things we're planning to do in 15 the future with the data base is first distribute it to l 16 NRR and RES. We're going to be able to do that very soon.
17 MEMBER MILLER: What are they going to do with l 18 it?
19 MR. TRAGER: We hope to get some feedback on -
20 -
21 MEMBER MILLER: Is feedback going to be 22 voluntary or is there going to be some sort of way of 23 forced feedback, so to speak?
24 MR. TRAGER: I don't know that. But I think
) 25 they would be only too happy to comment on it. We had an NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N W.
(202) 234-4433 WASHINGTON, D C 20005-3701 (202) 234-4433
122 1 inter-office working group to develop the data base. It i
2 was a pretty significant effort that went on involving a 7-( ')
3 group of people for a period of time. So I think there is 4 an interest in this. There's a lot of work and effort 5 that has gone into this.
l l
6 We thought we understood what the group 7 thought would be the data base that was the best data base 1
8 for human performance information. Whether it's turned 9 out that way, whether improvements are needed, we can find 10 that out.
l l
11 We are going to plan to continue to update the l l
12 data base with information from AIT reports, performed j 13 studies using the data, and add data from other source, I r~s 1 I
)
\~/ 14 which we haven't tried entering yet. The regional 15 inspections, when they were done in accordance with the 16 human performance investigation protocol, certain accident 17 sequence precursor events, foreign event reports that we 18 can get the reports released.
19 Then we are going to start looking at risk 20 significance of the events, and try to develop information 21 for use in PRAs and human reliability assessments.
22 MEMBER APOSTOLAKIS: Now that's interesting 23 now. That's very interesting. You are writing this as if 24 -- are you aware of the ATHENA project?
\
3 Technical Training Center.
4 MEMBER APOSTOLAKIS: When is this? When 5 again?
6 MR. ROSSI: It's in the fall of 1997. I think 7 it's probably scheduled for October, as they have 8 indicated up there. That specialists workshop will be on 9 the subject of the events involving human performance.
10 One of the reasons that the United States 11 offered to have it at Chattanooga next year is that 12 Chattanooga of course has a number of plant simulators, so 13 the simulators can be used in walking through some of the
(~T
\
'- '/ 14 events that have important factors related to it.
15 So we'll keep you informed of development of 16 the agenda for the meeting. Many of you will be d
17 interested. It also ties in with your question about how 18 we keep up to date within the agency, how people keep up 19 to date with state of the art. These sorts of workshops 20 which Cecil mentioned, go on frequently. Obviously we 21 don't send NRC people to each and every meeting that 22 occurs in the world, but we go to a large number of them 23 and we try to select them. That keeps people in all of 24 the offices pretty well up to date.
Q).
(_, 25 MEMBER APOSTOLAKIS: Now this particular NEAL R.. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N W.
(202) 234-4433 WASHINGTON, D C 20005-3701 (202) 234-4433
1 125 1 meeting will not address itself to the models themselves.
,_ 2 Just the events?
,3 L) 3 MR. ROSSI: Well, it's hard to say exactly 4 what may be covered in the meeting because we'll be 5 developing an agenda and inputs into it, but that's likely l 6 to be the case. But it may very well cover that too. I 7 don't know.
8 MEMBER MILLER: That will have international 9 participation, I assume?
10 MR. ROSSI: Yes. It's an international l
l 11 nuclear energy agency, which is the --
12 MR. THOMAS: It's the OECD countries.
13 MEMBER APOSTOLAKIS: But they will be mainly I A r
\/ ' 14 engineers and regulators?
15 MR. ROSSI: That's correct. I 1
1 16 MEMBER MILLER: So it will be European and 17 United States basically?
18 MR. ROSSI: Well there are many other 19 countries in there now too.
20 MR. THOMAS: Yes. Japan is involved. Korea 21 just recently joined.
22 MR. ROSSI: It's pretty worldwide. j l
23 MEMBER APOSTOLAKIS: I think worldwide is 24 important. But also who is invited is important. I mean
(~h 25 if you invite people who are doing this kind of thing,
(_f i
NEAL R. GROSS I COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N W.
(202) 234-4433 WASHINGTON, D C. 20005-3701 (202) 234-4433
126 1 then you get the false sense again that this is the state 2 of the art.
/,,T
~'
3 MEMBER MILLER: Are the dates for this 4 established?
5 MR. ROSSI: I'm not sure that the definite 6 dates are established yet.
7 MEMBER APOSTOLAKIS: If we know the date, we 8 can show up?
9 MR. ROSSI: Oh sure.
10 MR. CARROLL: They'll schedule it for ACRS.
11 ACTING CHAIRMAN SEALE: Take care of that 12 problem.
13 There's a common problem in how you examine or
(--) 14 the ability to examine some of these data in detail. I 15 know in -- there is a desire, at least in certain kinds of ;
1 16 data, particularly the -- and now I'm thinking about the 17 failure data that INPO collects in their NPRDS system, for 18 anonymity of exactly where it occurred.
19 Yet it seems to me that ultimately as the data 20 base gets large enough and so forth, that there should be 21 a bifurcation in some of the statistics reflecting those 22 plants which do a good job, and those plants which don't 23 do a good job. Whether you measure that with SALP scores 24 or whatever else there may be.
{^j%
i 25 As you look at these data and try to carry out NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N W.
(202) 234-4433 WASHINGTON, D C. 20005-3701 (202) 234-4433
127 1 these various evaluations and so forth, are you going to
,-s 2 try to begin to see whether you can make any judgements on N~j 3 the quality of let's say on the output in terms of l
4 reliabilities and so on between good performers and bad 5 performers, or are you still going to look at just the ;
I 6 collective -- l 7 MR. ROSSI: No. I'm sure that that will be 8 done. As a matter of fact, as Cecil indicated to you 9 earlier, he provides information to the discussions that j 10 take place with respect to the senior management meetings. I 11 That gets used to the degree that it's useful at this 12 point in time. I'm sure as time goes on and we learn more 13 about how to do all these things, that that will be done.
r~s l
k/ ) 14 Needless to say, NRC data bases depend on LERs and 15 inspection reports, are not proprietary.
16 ACTING CHAIRMAN SEALE: I understand that, 17 yes.
18 MR. TRAGER: This concludes this presentation 19 on this data base. I have got one more transparency I'll 20 put up in a second. But I guess trying to summarize this, 21 is human performance event data base. The structure was a 22 joint -- was a cooperative effort of three offices. We 23 have had INEL construct a data base. We have entered a 24 number of records. We are beginning to work to analyze (O ,/ 25 the data. We should have some results shortly.
NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N W.
(202) 234-4433 WASHINGTON, D C. 20005-3701 (202) 234-4433
128 1 Continuing. There's one other. This briefly 7-~ 2 covers the stated topic of the presentation, which is some i )
3 of the recent products of AEOD. The first listed is an 4 Oconee Electrical System Design and Operation study that's 5 going to be produced. We expect it will be published in 6 March of 1997.
7 We just recently finished an assessment of 8 spent fuel pool cooling. A review was performed of --
9 MR. CARROLL: Does this deal with the 10 Susquehanna situation and the similar problems exist on 11 PWRS?
12 MR. ROSENTHAL: Yes. We're sort of getting I
13 off our topic, but the chairman and the EDO asked us to do
+/~'s, l
'w/ 14 an independent evaluation of a spent fuel pool -- spent 15 fuel cooling problems. We wrote a separate report and did 16 brief the Commission.
17 MEMBER MILLER: That's the one we saw. We had 18 --
19 MR. ROSSI: Yes. I think at some point I i 20 think we are going to be here to brief you on that study.
21 We did brief the Commission on it already in the last two 22 or three weeks.
23 MEMBER MILLER: That's the one where there's 24 the instrumentation question and things like that?
(q) ,
25 MR. ROSSI: Right. Now some of these studies I NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE.. N W.
(202) 234-4433 WASHINGTON, D C. 20005-3701 (202) 234-4433
129 1 up here are far broader than just human factors issues.
l
,s 2 Probably the one that is mostly devoted to human factors .
! \ l kJ issues is the operating events within appropriate bypass l
i 3
4 or defeat of engineering safety features. That's 5 primarily zeroed in on human factors things. !
6 Jack taay want to say something about other 7 AEOD studies over the years that are more directed at 8 human factors.
9 MR. ROSENTHAL: And of course we did look at i
10 the NTSB work. Gene is being very modest because he is 11 the one who did the agency's first wrong unit, wrong train 12 study, and participated in the NUREG 1275 volume 8, which 13 is the summary of the 20 events, and actually went out on c
k- 14 many of those events himself. So there's several 15 publications that we have done over a long period of time.
16 MEMBER MILLER: Is the Oconee one the one 17 situation they are now in right at this time?
18 MR. ROSENTHAL: Yes.
19 MEMBER MILLER: The one with the -- I can't 20 remember all the details.
21 MR. CARROLL: They don't have diesels.
22 MEMBER MILLER: Charlie Wylie should be back 23 for this.
24 MR. CARROLL: I don't' know that he wants to
( 25 come.
NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N W.
(202) 234 4433 WASHINGTON. D.C. 20005-3701 (202) 234-4433
130 1 MR. TRAGER: There's a typo in there, by the c~ 2 way. The assessment of spent fuel pool cooling, that's a V) t 3 special study 96-02.
4 MEMBER APOSTOLAKIS: We noticed that.
5 MR. TRAGER: These are just recent products, 6 examples of recent products. So if there are no other l
7 questions, I think we beat an hour and a half by five i l
l 8 minutes.
9 ACTING CHAIRMAN SEALE: Very good. Are there 10 any questions from any of the members further? Any 11 comment from the rest of the staff? l 12 MR. ROSSI: Well, I would like to say that as 13 we start trying to use this data base, we'll be looking l
, ,\
I v' 14 for information that we can use out of the data base as it i 15 exists now. We will also be looking at possible changes 16 that need to be made to the data base based on comments 17 that you have made, and based on our own --
18 ACTING CHAIRMAN SEALE: Certainly we'11 be 19 willing to --
20 MR. ROSSI: To try to use it in a more 21 practical way. All the offices will be looking at it and 22 working on that together.
23 ACTING CHAIRMAN SEALE: Certainly we'11 be 24 willing to hear anything you might have that would be of Q 25 interest to us.
NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N W.
(202) 234-4433 WASHINGTON, D C. 20005-3701 (202) 234-4433
131 1 One other question. Is there -- I notice 2 there's one person from the public here. Would you like
[,'i 3 to make any comment?
4 AUDIENCE PARTICIPANT: No.
5 ACTING CHAIRMAN SEALE: Okay. Well, I guess 6 it's about time for chow, right?
7 MR. CARROLL: I would just like to reinforce 8 George's point though. Data is wonderful, but you have 9 got to fit it to a model if it's really going to be 10 something you can fiscally apply.
11 ACTING CHAIRMAN SEALE: Good point. Okay, 12 George.
13 Okay. We'll recess for an hour. We'll be (D
(_) 14 back at 1:00 to resume our discussions on the questions. i 15 (Whereupon, the foregoing matter went off the 16 record at 12:00 p.m.)
17 18 l
l 19 l
l 20 i
21 22 23 24 Lj 25 NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N W.
(202) 234-4433 WASHINGTON. D C. 20005-3701 (202) 234-4433
132 1 A-F-T-E-R-N-O-O-N S-E-S-S-I-O-N
, .s 2 (1:10 p.m.)
/ i 3 CHAIRMAN APOSTOLAKIS: Okay, we're back in 4 session, and we'll continue with the questions. Where are 5 we -- question 11?
6 DR. THOMAS: Mr. Chairman?
7 CHAIRMAN APOSTOLAKIS: Yes? !
l 8 DR. THOMAS: With your permission, we would l I
9 like to move -- we are ready to resume with 11. We would 10 like to move to 16 because we have a person from Operator 11 Licensing Branch who's been here this morning, and has 12 enjoyed enough of a vacation.
13 CHAIRMAN APOSTOLAKIS: No problem.
(~h ks 14 DR. THOMAS: Thank you.
15 CHAIRMAN APOSTOLAKIS: So we'll do 16 first 16 and then go back?
17 DR. THOMAS: Yes. l l 18 CHAIRMAN APOSTOLAKIS: Okay. Sure.
, 19 DR. THOMAS: I'd like to introduce Frank 1
20 Collins of the Operator Licensing Branch of NRR, who will i 21 address 16.
l l 22 DR. COLLINS: Good afternoon. As Cecil said, l
23 I'm Frank Collins; I'm an Operator Examiner with the 24 operator Licensing Branch. I spend almost all of my time (O ,/ 25 on the subjects of simulators and simulator certification.
NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N W.
(202) 234-4433 WASHINGTON D C. 20005-3701 (202) 234-4433 i
133 1 The question 16 is in three parts. It's: how g ~. 2 does the staff assure simulator fidelity; how important is V 3 good fidelity to emergency operator procedure training; 4 and what does the staff expect an operator to do if 5 unexpected plant behavior occurs during a severe accident.
6 To address the first part I'd like to, if I 7 may, refer to this visual aid that I brought with me to 8 put fidelity in the context of the development of the 9 simulator and certification of a simulator and maintenance 10 of it in a training and examination process.
11 The simulator fidelity which we see here on 12 the flow chart is not really defined in the standard. If 13 you go to the standard ANSI 3.5, fidelity is only defined l 1
\/ 14 in terms of physical fidelity or physical relationship to )
15 the panels.
16 The fidelity as I interpret the question here, 17 as the replication of the plant's characteristics and -- )
i 18 plant's characteristics and the plant's responses, is not 19 defined per se. And it's not a tangible item that you can 20 go to in any one spot and say there, we have fidelity now.
21 It's actually an outcrop of this whole process that starts 22 over here with the software design verification and l
23 validation processes.
24 And this process in initial simulator
(~T
's ,) 25 development probably takes somewhere on the order of six NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N W.
(202) 234-4433 WASHINGTON, D C. 20005-3701 (202) 234-4433
134 1 to eight months of design and verification; another six g-) 2 months of validation; and not unlikely to have another six V 3 months beyond that, of factory acceptance test by the 4 facility licensee before they'll accept a simulator from 5 the vendor.
6 And the aggregate of all that testing and all 7 that software documentation, is fidelity. That is what 8 the licensee certifies to here, on their NRC Form 474.
9 Saying that they have a facility in place that meets the 10 requirements of 10 CFR 55.45, for us to be able to come in 11 and do operator licensing examinations, and sample across 12 the 13 broad sampling areas that are specified there.
13 And they specify that they have built a
(~'i
\- 14 simulation facility and are maintaining it to an ANSI 15 standard. Typically, it's ANSI 3.5 and typically there 16 are very few exceptions taken.
17 MEMBER MILLER: It's my understanding that 18 many plants involve their operations staff directly in the 19 V&V process, is that true?
20 DR. COLLINS: Yes sir. Almost all plants 21 involve the operations staff in the V&V process. In fact, 22 experience has shown that the ones that didn't get them in 23 early, struggle.
24 MEMBER MILLER: Right.
()
(, 25 DR. COLLINS: To get to that final fidelity NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N W.
(202) 2344433 WASHINGTON, D C. 20005-3701 (202) 234 4433
135 1 where you can say, I'm carrying a minimum number of fs 2 ongoing problems. You really have to get them in there
/ 4 L.]
3 early.
4 MEMBER MILLER: So each different plant 5 personnel may have responsibility for one set of systems 6 or something like that, I believe?
7 DR. COLLINS: Typically, they're staffed with 8 three or four software engineers, one being a --
9 MEMBER MILLER: That's the vendor?
10 DR. COLLINS: I'm sorry?
11 MEMBER MILLER: The vendor is?
12 DR. COLLINS: The vendor typically is staffed 13 with, anywhere from a dozen on a smaller simulator to a
\/ 14 larger simulator being up to two dozen; engineers of 15 unique disciplines. You'll have core engineers, you'll 16 have thermal hydraulics people, you'll have electrical 17 people. And --
18 MEMBER MILLER: No, I'm talking about the 19 plant personal that get involved in the V&V --
20 DR. COLLINS: Oh, the plant personnel that 21 come in are usually operators.
22 MEMBER MILLER: Those are primarily operators?
23 DR. COLLINS: Primarily they're operators and 24 instructors.
) 25 MEMBER SEALE: Yes, the instructors are a big NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N W.
(202) 234-4433 WASHINGTON, D C. 20005-3701 (202) 234-4433
l 136 1 part of that whole process, too.
,s
, 2 DR. COLLINS: They are integral to the j I )
3 process.
4 MEMBER SEALE: Yes. Yes.
i 5 DR. COLLINS: And they're valuable in the fact j i
6 that the vendor's personnel that you were discussing, I l
7 generally don't have the integrated plant view. And what 8 the instructors and what the operators bring is the 1
1 9 integrated plant view that can tell one particular 10 discipline, this is fine what you're doing here in 11 preliminary design but you have to factor in this also, i
12 and we have to establish handshake with another system.
13 MEMBER SEALE: Now -- well, I'm sorry. Go
(
/~s
)
\_/ 14 ahead. i l
15 DR. COLLINS: I'm sorry. What we're '
16 certifying here on Form 474 is that they're meeting the 1
17 ANSI standard and typically there are very few exceptions 18 taken to that. We endorse the ANSI standard and Reg Guide 19 1.149, and currently we are endorsing both 1985 and the 20 1993 versions.
21 We answered in 1987, industry's questions 22 regarding how that implementation would be a NUREG-1262, a 23 series of questions and answers that accompanied the 24 implementation of Part 55 rule change in 1987. And many
() 25 of the operational practices of certification are really l NEAL R. GROSS l COURT REPORTERS AND TRANSCRIBERS l 1323 RHODE ISLAND AVE., N W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 (202) 234-4433 l
1
137 1 defined in the NUREG-1262.
fs 2 But what the Reg Guide and the NUREG !
1 3 specified, was that even after you've established this 1
)
4 initial -- pedigree, if you will -- for your simulator, !
1 5 that they will undertake a program of periodic testing.
l 6 And it involves both annual testing on a 7 macroscopic scale, and quadrenially-spaced testing on a i 1
8 microscopic scale, to make sure that you can go through 9 that whole software package, and to make sure that as you 10 incorporate plant modifications which are constantly 11 coming in from the plant, being fed back up into the 12 design and scope process and then retrofit, or as you i 13 incorporate changes that operators see and say, that's not I 14 just like the plant. The plant actually hits up faster, 15 cools down slower, or what have you.
16 As these things are done, this periodic 17 testing program is to make sure that you still maintain 18 that validity, and you can take the test results that you 19 had from previous runs that were validated, compare them 20 with what you have now.
21 And that's an ongoing test; those records are 22 maintained at the site and can be periodically reviewed.
23 That's what establishes simulator fidelity, and we assume 24 simulator fidelity when we go to do our operator
-s
) 25 licensing.
NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N W.
WASHINGTON, D C. 20005-3701 (202) 234-4433 l (202) 234-4433
138 1 MEMBER SEALE: I think the question came up in
,s 2 the context of how far away from the nominal performance
(
~'
)
3 of the plant can you trust the simulator to give you a 4 reasonable simulation of the plant performance? And how i
5 do you know when you have to go to more fundamental i 6 diagnostics in order to really understand the status in 7 the plant; when do you flush the simulator, essentially, 8 in a severe accident scenario?
9 DR. COLLINS: The simulator will be bounded by I i
10 a set of criteria that are established during preliminary 11 design that typically will take you up to the point of 12 physical failure of components, a loss of coolable 13 geometry in the core -- physical failure again, of
(%
14 containment, of turbine, things like that --
15 MEMBER SEALE: Sure, sure. l 16 DR. COLLINS: And up until that point, there's 17 a great deal of confidence because the initial comparisons 18 were done against best estimate analysis that included all 19 of the analyzed transients from the safety analysis 20 report, and a series of transients sort of specified in 21 the standard.
22 The operator has reasonable faith in it as 23 long as he continues to see the expected results that 24 don't defy laws of physics. And my experience has been p)
(, 25 that as long as you're not defying the laws of physics NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N W.
(202) 234-4433 WASHINGTON, D C. 20005-3701 (202) 234-4433
139 1 with the model and you can still operate within the ex 2 boundaries of your procedures.
U 3 The model has proven to be fairly trustworthy, 4 if you take the data off and go back to compare it to what 5 you had benchmarked earlier. There are very few cases 6 where the model does start to fall apart, and only when 7 you get out to those extreme limits out where you're i I
8 boiling in a suppression pool or you're on the edge of 9 losing poolable geometry, something like that.
1 10 MEMBER SEALE: Well, the earlier example that 1 11 was mentioned in our discussions this morning had to do 12 with the possibility of a chugging kind of process as a 13 part of an ATWS event, in a boiler where you had some cold ,
(s ,
14 water giving you condensation and then flashing and then 15 some more condensation. l 16 DR. COLLINS: The newer, advanced models 17 should be able to handle that without great difficulty.
18 Earlier models were basically 1- or 2-node thermal 19 hydraulic models, and they depended on a mass and energy 20 balance equation.
21 MEMBER SEALE: Yes.
22 DR. COLLINS: If you look at what's being done 23 in the current models -- and I say current, I'm going back 24 from around 1985, starting into what they called the
. 25 advanced models; they've got various trademark names to l
t NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N W (202) 234-4433 WASHINGTON, D.C. 20005-3701 (202) 234-4433
140 1 them -- but they involved, first of all, not having that
,s 2 strict coupling, a more nodal relationship where the i
s
)
'~'
3 process was being resolved in many more nodes throughout 4 the vessel, throughout the primary loop, throughout the 5 boiler.
6 And we didn't see that on the earlier ones 7 primarily as a result of computing power. The computing 8 power question has been obviated now, so we're at the 9 point now where the advanced models have taken into 10 account such things as the difference in densities between 11 the fluids, 2-phase flows at each node, and the pressure 12 differentials and flow differentials and reversals that 13 can be caused between nodes, and which would give you the t
(n%-) 14 chuggina effects.
15 MEMBER MILLER: What it won't do is, it won't 16 give you the spatial effects say, in a BWR under an 17 instability situation, I don't believe.
18 DR. COLLINS: In what regard?
19 MEMBER MILLER: Well, like take the LaSalle 20 event. We had instability in the core in a spatial sense.
21 DR. COLLINS: Some of the newer models are l
22 giving you that, yes.
l 23 MEMBER MILLER: I used the Perry one quite 24 often. We tried -- we'd like to do that and we can't do f%
q,) 25 it with Perry.
NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS i 1323 RHODE ISLAND AVE., N W.
(202) 234-4433 WASHINGTON, D C. 20005-3701 (202) 234-4433 i
f 141 1 DR. COLLINS: I don't believe the Perry one 7_
2 has the updated model in it.
(\') 3 MEMBER MILLER: No, it's probably a couple of 4 years old.
5 DR. COLLINS: Perry model was one of the 6 earlier generation -- I think we built Perry when I was 7 with GE back in 1980.
8 MEMBER MILLER: No, they were updated since 9 then.
10 DR. COLLINS: Since then?
11 MEMBER MILLER: I guess they're -- the one 12 we're using now probably was updated three years ago.
13 DR. COLLINS: Oh, okay.
,n i \
'w /
14 MEMBER MILLER: But we want to do some spatial 15 effects and we can't do it.
16 DR. COLLINS: I don't think you can get a 17 broad scale -- you can get power chugging but it will be 18 uniform throughout.
19 MEMBER MILLER: Right.
20 DR. COLLINS: In the second-generation models 21 -- Perry probably is carrying a RETAC model and in that 22 generation model --
23 MEMBER MILLER: You can't get the neutron flex 24 distribution spatially.
,m ,
(s- ) 25 DR. COLLINS: No. No, neutronics -- !
NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N W.
(202) 234-4433 WASHINGTON, D C. 20005-3701 (202) 234-4433
142 1 MEMBER MILLER: I have a graduate student like 7_
2 to do that --
( $
3 DR. COLLINS: It doesn't have that level of 4 definition.
5 MEMBER MILLER: Right. But the one they have 6 now is far different than the one they.had way back in the 7 '80s --
8 DR. COLLINS: Oh, yes. It's an order of 9 magnitude difference. And now they're going to -- we're 10 actually --
11 MEMBER MILLER: Maybe two orders of magnitude 12 difference.
13 DR. COLLINS: Excuse me?
V O) 14 MEMBER MILLER: Maybe two orders of magnitude 15 difference.
16 DR. COLLINS: Yes, it's greatly different.
17 And now you're getting to the point where you actually can 18 run design codes in real time. You're getting to the 19 point where RELAP is being fielded in leal time now, and 20 with a full 5-equation set for thermal hydraulics.
21 Neutronics is still done in a multi-node 22 simulation, but a lot of what you see is not calculated in l
l 23 on each individual detector. Each individual LPRM is not 24 a separate calculation in that core model. They have an
/~
Q} 25 array of nodes and then what the reading would be l
NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N W.
l (202) 234-4433 WASHINGTON. D C. 20005-3701 (202) 234-4433
143 1 representative for that LPRM is calculated based on a l l
,s 2 closest node. That's not being actually calculated at
/ \
3 each LPRM location.
4 DR. CARROLL: Back to your ATWS in a boiler.
5 I was up at Pilgrim, oh, it's been several years now, but 6 they demonstrated an ATWS on their simulator and as I 7 recall, it did get oscillation.
i 8 DR. COLLINS: Oh, they'll get into power and 9 pressure oscillations -- almost all of them will do that -
10 - but will you will be able to see it at different parts 11 of the core and different parts of the vessel?
12 DR. CARROLL: I understand, I understand. l 13 DR. COLLINS: There you don't have that level (3
\- l 14 of definition in some of the older models. Maybe some of 15 -- maybe RETAC models that are going on real-time now will 16 be able to do that, but that's a tremendous amount of l
17 computer power.
18 MEMBER MILLER: Oh, yes. If you wanted to I i
19 simulate all of the LPRMs, you're talking about 170- some I 20 of them.
21 DR. COLLINS: Yes, if you take a look at a 22 generation simulator that you're talking about with Perry, 23 they basically ran two Gould 8780 processors and used one l
- 24 exclusively for the core model.
) 25 MEMBER MILLER: That's the old one.
NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N W.
l (202) 234-4433 WASHINGTON, D C 20005-3701 (202) 234-4433 1
1 144 l
1 DR. COLLINS: Yes. One exclusively for the l
,y 2 core model. Now we're getting to the point where they're
('~' )
3 getting down on a single Pentium processor. They're 1
4 running multiple Pentium processors to run advanced codes.
5 MEMBER MILLER: But see, the ATWS to dett;t 6 the instability in the BWR now uses the LPRMs. I don't 7 believe -- at least the Perry does not simulate that.
8 DR. COLLINS: Not at that level of rigor. The 9 second part of the question, how important is the fidelity 10 to emergency operating procedure training? We see that in 11 operator licensing in terms of our exams.
12 We built our exams in the emergency operating 13 procedure realm, where we take our scenarios down through
/ .,N ,
e
\ i 14 the emergency operating procedures ensuring that we have a i
15 success path that you can get from the initiation point to l 16 a point where the operator says yes, I am in control --
17 I'm either recoverable to some extent or I'm stable. .
18 MEMBER SEALE: But once you get to the point 19 where you have equipment failures breakdown, loss of 20 containment or that kind of thing, then you don't rely on l 21 the simulator any more to --
22 DR. COLLINS: No, you can't. By the ANSI 23 standard, what will happen then is the simulator will 24 inform the instructors, and most of them will go into a
(-
(_,/ 25 freeze mode of operation. They'll cease displaying so NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 (202) 234-4433
145 1 that they don't present a negative training scenario at l
,e 2 that point.
( )
'~
3 MEMBER SEALE: Very good.
4 DR. COLLINS: And inform the instructors that 5 we are now out of real-time or out of the validated realm. l 6 MEMBER MILLER: Is the ANSI standard in the 7 process of being updated?
8 DR. COLLINS: Yes, it is. In 1998 we're 9 hoping to have that revision out. We just about have it -
10 -
11 MEMBER MILLER: That's ANS, right?
12 DR. COLLINS: Beg your pardon?
13 MEMBER MILLER: That's an ANS --
! )
k/ 14 DR. COLLINS: Yes, it is. This is 3.5 and 15 we've just about got it now to a consensus draft. I'm on 16 that writing group. A consensus draft just went out in l 17 the mail this week, as a matter of fact, and we're hoping 18 to get the people together later this month and see if we ,
l 19 are at the level of consensus that we think we are. That 20 may be a surprise.
21 And if so, we'll be ready to move it up to 3.0 j 22 for consideration, but the major changes that we're 23 looking at in the next revision are to provide a software 24 verification validation and testing structure that will --
p'
! 25 MEMBER MILLER: Will that follow the IEEE V.
NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISMND AW., N W (202) 234 4433 WASHINGTON, D.C. 20005-3701 (202) 234-4433
7 146 1 standards?
2 DR. COLLINS: That will more closely conform
(,_h i
/
3 to the IEEE standards and at the same time will allow the 4 facility licensees to do more effective testing in some of 5 the testing they're doing now.
6 MEMBER MILLER: Will that specify then, 7 specifically that operators are part of the V&V?
8 DR. COLLINS: I'm sorry, sir?
9 MEMBER MILLER: Will that specifically specify 10 operation staff to be part of the V&V process, then?
11 DR. COLLINS: No, it doesn't specifically say 12 that operators and the staff will be part of the V&V 13 process. It does say that the software development and
\_/ 14 implementation will be part of their -- shall be part of a 15 structured software development process. I'm trying to 16 recall the exact words. We said it without referring 17 explicitly to the IEEE standard because ANSI standards 18 don't like to invoke other standards; they like to be 19 free-standing as much as possible.
20 But the problems that you have with simulator 21 fidelity today are over here, when changes have to occur 22 and models become updated -- and the new cores go in, for 23 example. The new core goes in, it requires a whole new 24 validation, it requires a whole new stamp of pedigree to y~.,
25 ensure that you still are good.
(x -)
NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N W.
(202) 234-4433 WASHINGTON. D.C. 20005-3701 (202) 234-4433
147 1 And this is being done in the heat of battle, 2 if you will, with all the pressures that are put upon the
[,. \
\- 3 training staff and the simulators. And there is a real 4 challenge for them in being able to get the appropriate 5 testing done and go through that design and verification 6 process in a structured manner.
7 MEMBER MILLER: So with all the plants 8 upgrading their power that give simulator people 9 headaches?
10 DR. COLLINS: It can be, because when a plant 11 upgrades in power, what you can find is that you've got 12 whole new baseline data, and that would show up over here 13 in plant mods.
,/
(_ / 14 And the plant mods then would, if you take 15 that back to scope and design and say, where do I stand 16 now? I now am at a position where if I run my heat 17 balance test for the ANSI standard annually, I'll be 18 outside the two-and-a-half percent or two percent on the l
i 19 1985 standard, one percent on the 1993 standard.
20 I'll be outside of my allowable tolerance on 21 critical parameters, and I will subsequently fail the 22 test. So they have to go back and revalidate that again.
23 It's a significant effort. It takes most !
24 facility licensees three to four full-time staff (n) v 25 equivalent to maintain a similar.
NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N W.
(202) 234-4433 WASHINGTON, D C. 20005-3701 (202) 234-4433
148 1 MEMBER MILLER: What fraction of the 2 operator's exam is on the simulator?
I7-~\
\s^/
3 DR. COLLINS: The operator's exam is in two 4 parts. There's a written examination and there's a -- you 5 can say it's 50/50. There's an operating examination.
6 There's some plant walk-through or some of the job 7 performance measure are done, but all of the scenarios are 8 done on the simulator. The operating examination is built 9 around the simulator.
10 CHAIRMAN APOSTOLAKIS: Have we discussed 16 11 enough?
12 DR. COLLINS: I think that we've discussed all 13 but the last part, and the last part is what does the
,y
\~-} 14- staff expect an operator to do if unexpected plant 15 behavior occurs during a severe accident? I think it's 16 not a matter of if, it's a matter of when.
17 By definition, I think what we'll see is we'll 18 see the operators becoming more cognitive in their 19 activities; they'll go into a much more active 20 communications role with the technical support center.
21 CHAIRMAN APOSTOLAKIS: But you do expect them 22 to do the right thing?
23 DR. COLLINS: By all means. They still have 24 the responsibility under their license to protect the
) 25 health and cafety of the public, and their conditions have NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 (202) 234 4433
l 149 l
1 changed in the plant, but they still have that same i 2 responsibility.
l O
! 3 CHAIRMAN APOSTOLAKIS: Okay. Thank you very 4 much.
5 DR. COLLINS: Thank you.
6 CHAIRMAN APOSTOLAKIS: Now, we have a choice j 7 here. Actually, it's not much of a choice. I think we 8 really ought to hear from the NMSS staff and then come 9 back again, here in the discussion pe';iod. j l
10 (Off record discussion.) l l
I 11 Well then, I declare a break until 1:30.
12 (Whereupon, the foregoing matter went off the i ,_ 13 record at 1:30 p.m. and went back on the 1 (
\- 14 record at 1:33 p.m.)
15 DR. COMBS: Good afternoon. I'm Fred Combs, l
16 I'm Deputy Director of the Division of Industrial and 17 Medical Nuclear Safety. With me is Dr. Josephine Piccone )
18 who's Director of the Operations Branch, and also Dr.
19 Dennis Serig who works for Dr. Piccone in the Operations 20 Branch.
21 We're here to respond to a question you raised l 22 concerning the status of the NMSS program in human I
i 23 factors. Dr. Serig will make the presentation.
24 DR. SERIG: On your first slide -- I'm not x_) 25 going to do overheads -- but on your first slide you can NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N W.
(202) 2344433 WASHINGTON, D C. 20005-3701 (202) 234-4433 l
t
150 1 see that the program was initiated in May of 1988, at the
- f. s 2 direction, basically, of Hugh Thompson, now the Deputy I \
~
3 Executive Director for Operations.
4 It had limited resources at that time. It has 5 limited resources at this time. But you have to 6 understand that all of NMSS has limited resources in 7 comparison to the resources Lhat we direct toward reactor 8 issues in this agency. At that time there was one human 9 factors analyst and there still is one human factors 10 analyst.
11 DR. CARROLL: And that's you.
l 12 DR. SERIG: That's me.
I 13 DR. CARROLL: Why don't you tell us a little
'i r"'
I' 14 bit about yourself? How did you get to be a human factors I l
15 analyst?
( 16 DR. SERIG: All right, fine. I'm a Ph.D.
! 17 Psychologist, from Penn State University. I left --
18 DR. CARROLL: Oh, oh. We knew somebody that 19 went to Penn State, didn't we?
20 DR. SERIG: Yes, you did. I spent some time l
l 21 working in human factors for the Army at Aberdeen Proving l
l 22 Ground in the human engineering lab -- what was the human 23 engineering lab at that time. Soon after Three Mile 24 Island the agency determined that human factors was
/^)\
(_ 25 something of interest to them here, and I was one of the NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N W.
(202) 234 4 433 WASHINGTON, D.C. 20005-3701 (202) 234-4433
I 151 1 group of analysts that was hired at that time to look at
,. 2 control and design reviews and safety parameter display
(, s) 3 systems.
4 I did that for about s.ix years. I spent a 5 year-and-a-half in the Office of Research working on some 6 projects that were associated with reactors. At that time 4
7 an opportunity came to work at NMSS and I was selected to 8 do so. So I am one of the limited resources in NMSS.
9 MEMBER MILLER: How many personnel in NMSS?
10 DR. COMBS: There are approximately 300 11 personnel in NMSS.
12 DR. SERIG: So it's about 10:1 -- 9:1, 10:1 13 personnel resources, reactors to materials. At the same
/~~~%
\V) 14 time NMSS has limited resources, it also has very broad 15 responsibilities. We have between 6- and 7,000 licensees, 16 and those licensees extend over a broad gamut of
- 17 activities which I'll review with you in just a second.
18 ,That meant that, just like everybody else in 19 the agency, we had to prioritize our activities, but it 20 really meant that we had to be much more strict in 21 prioritizing our activities than some of the other offices 22 have been.
23 We had to determine very specifically, what 24 safety problems warranted attention and what attention we p
y) 25 could afford to devote to those. We did that early on by NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N W (202) 234-4433 WASHINGTON, D.C. 20005-3701 (202) 234 4433
152 1 a committee meeting, a group sitting around the table and
,.s 2 looking at all the kinds of things that NMSS regulates and i i t
/
3 the kinds of things that the Commission at that time and 4 the office were concerned with, and selecting from those, 5 some things that we thought we could do something about.
6 Notice that on the first slide also, there's a 7 need to leverage resources where possible. We try to use
. 8 other folk's resources to do the work that we see that 9 needs to be done to the extent we can. That includes the 10 resources of NRR on occasion, RES, and some outside
)
11 organizations such as the Association for the Advancement 12 of Medical Instrumentation -- that I'll speak about in a 13 little while. l
,y kls 14 Basically, we tried to find out what things 4
15 needed to be done, who's doing them or could do them, and !
16 then we try to benefit from the work of others to the 17 extent we can. If we cannot, then we do some things on l 18 our own.
19 The next slide shows the areas of 20 responsibility, or some of the areas of responsibility in 21 NMSS. Medical use -- I'm not sure how familiar you are 22 with our licensees. Medical use involves nuclear medicine 23 which is the ingestion, injection, or inhalation of a 24 radiopharmaceutical, and then usually some kind of towning (O _j 25 or imaging process as a diagnostic; but it might also be NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N.W.
(202) 234-4433 WASHINGTON, D C. 20005-3701 (202) 234-4433
153 1 therapy for something like cancer of the thyroid, or for
- 2 now, palliation of bone cancer.
/
\'~') 3 Then there's radiation oncology which involves 4 primarily sealed sources, either by which a patient is 5 irradiated from a distance -- teletherapy -- usually using 6 cobalt-60, or something called brachytherapy where sources 7 are installed in a patient's body and left for a specific 8 period of time and then removed; or in some cases, because 9 of rapid decay and small energies left in place. l 10 We also regulate industrial use and non-11 destructive testing; industrial radiography.
12 Sterilization, primarily of medical products, in large 13 irradiators having millions of curies of material; O- 14 engages, where people are looking at moisture density and l 15 things like that, on parking lots and compacted soil sites 16 and general construction sites. I 17 Academic --
18 CRAIRMAN APOSTOLAKIS: Excuse me.
19 DR. SERIG: Yes?
20 CHAIRMAN APOSTOLAKIS: I just looked at your 21 whole presentation and it doesn't seem to me you're 22 addressing the question. Why isn't NMSS as involved with j 23 human performance efforts as the other offices? I thought 24 that meant in the plan. In developing the human factors
() 25 plan. Unless I misunderstand the question. Because you NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N W.
(202) 234-4433 WASHINGTON, D C. 20005-3701 (202) 234-4433
1 1
154 l
1 do list here the activities of NMSS in the human factors
' (q t
2 area --
RJ 3 DR. SERIG: Right.
4 CHAIRMAN APOSTOLAKIS: -- and you're doing a 5 pretty good job of it, but I thought the origin of 6 question number 7 --
l 7 DR. SERIG: Is that we have very few bullets 8 in the human performance program.
9 CHAIRMAN APOSTOLAKIS: And you are not -- you 10 don't seem to be an actor in formulating the plan. I 11 mean, the gentlemen we had here this morning did not l
l 12 include an NMSS representative. Am I correct?
13 MEMBER SEALE: That's right.
i t
V 14 CHAIRMAN APOSTOLAKIS: And that was really the 15 origin of question number 7. j 16 DR. COMBS: Yes.
l 17 DR. THOMAS: In defense of NMSS, Dennis did 18 participate as a member throughout the development of the 1
19 human performance program plan, but I think the question
! 20 really goes to the number of bullets that they have, 1
l 21 because NMSS didn't participate completely with the plan l
1 22 development.
23 CHAIRMAN APOSTOLAKIS: I mean, you are 24 describing human factors activities in NMSS. If we can n/ 25 comment on this -- if it's any of our business, actually.
NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N W.
(
l (202) 234 4433 WASHINGTON. D C. 20005-3701 (202) 234-4433
155 1 I don't know; probably not.
fs 2 DR. COMBS: It's just that what we're doing
('~' )
3 doesn't appear to look the same as what the other offices 4 are doing. And I felt that what we would do here was to 5 establish the basic reasons for the nature of our 6 involvement and the fact that because their plan is I
7 developed differently and because we all do human factors, l
8 there is a bit of interrelation. l 9 However, we're being a bit more specific in 10 ours. We've agreed with the human factors performance i i
11 plan and we do have data in there. We do have 12 circumstances, however, which for some of our activities, 13 we haven't identified a target date due to assessments 14 being done for strategic assessments in other initiatives.
15 CHAIRMAN APOSTOLAKIS: But if we look at what 16 you have here, I mean, you do have a human factors program 17 in NMSS, but it doesn't seem to either receive information 18 or give information to the rest of the agency in terms of 19 experiences, insights, models. I mean, AEOD is going to 20 do all this work, evaluating data immediately from 21 reactors --
22 DR. SERIG: Well, I certainly can address that 23 historically, and give you some projection into the 24 future. As I said, I came out of NRR's program and much q
k ,/ 25 of what we have done has been based on the activities that NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N W.
, (202) 234 4433 WASHINGTON. D C. 20005-3701 (202) 234-4433
i 156 l 1 were started in reactors. We've translated those,
,3 2 however, into programs that were suitable for our r i 3 licensees.
4 So you'll see in the slides here that we've 5 done things called human factors evaluations. Those are l 6 very much based on detailed control room design reviews 7 that we required of our reactor licensees, but they also 8 integrate the question of procedures, training, and 9 management, that were sort of separate issues when they 10 were addressed in NRR.
11 We've used the resources of Research -- we've 12 interacted with Research extensively to get those human 13 factors evaluations done. You'll find on each of the l g3 k 14 reports that was completed, an author from the Office of l 15 Research and an author from NMSS, as well as the t
l 16 contractors because of the extent of the involvement of l
l 17 the two offices in development of the product.
l 18 We have worked AEOD involved in something l 19 called the nuclear material events database. We've t
l 20 described to them what kinds of information we would like i
1 21 to receive from that, and we've provided them some 22 feedback on where we think they are, being able to serve 23 us now.
24 They're right now working with research on a (j) 25 project to locate task network modeling as a tool for NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 (202) 234-4433
157 1 assessing a number of things. One of them is the risk 7.-
2 associated with material systems. But we're also trying 3 to integrate a look at human events with a look at 4 hardware and software events.
5 CHAIRMAN APOSTOLAKIS: So it seems to me that 6 you gave us a very strong argument in the last couple of 7 minutes as to why you should be an integral part of the 8 human factors plan. You have both benefitted yourselves 9 from activities in AEOD and NRR, and presumably your 10 experiences will be of benefit to them. So I don't see 11 why you're acting sort of separately.
12 DR. SERIG: I'm not saying we are.
13 CHAIRMAN APOSTOLAKIS: They're not part of the em
( )
k/ 14 plan.
1 15 DR. SERIG: What?
16 CHAIRMAN APOSTOLAKIS: They're not part of the 17 plan. The plan is developed independently of these 18 gentlemen.
19 DR. THOMAS: No, it's not.
20 MEMBER MILLER: That's the whole point.
21 CHAIRMAN APOSTOLAKIS: What is the point?
22 MEMBER MILLER: I thought their point first of 23 all, was to give a perspective of what they're responsible 24 for, I think which gives us a little different view than
(*)
( 25 just reactors. Then the point was, here's how they're NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N W.
(202) 234-4433 WASHINGTON, D C. 20005-3701 (202) 234-4433
i 158 1 involved even though they have only one person basically,
!,y 2 being involved. That was the point they're going to make G'i e
i 3 --
4 CHAIRMAN APOSTOLAKIS: No. As I remember --
5 DR. SERIG: -- get involved in development l
l 6 plan --
l 7 MEMBER MILLER: I wasn't here September 20th.
8 CHAIRMAN APOSTOLAKIS: September 20th the 9 question arose because NMSS was not here. And the i
10 question maybe is not phrased very well here, but the 11 question was, why not? I mean, here the agency is 12 developing a human factors plan, and NMSS is not present.
13 That was the thrust of the question. I don't care about l
\- 14 the actual wording here. So now we're hearing that NMSS 15 has its own program.
l'i DR. SERIG: NMSS has a program which is to the 17 extent that's reasonable, been integrated with the other 18 offices.
19 MEMBER MILLER: But their scope is so much 20 different than the reactor scope --
l 21 MEMBER SEALE: Yes, could we back off just a 22 little bit? Maybe it might help us. Your primary vehicle 23 for -- in a very real sense your customer is the agreement 24 state rather than the licensee?
p
'N ,) 25 DR. SERIG: We have about 6- to 7,000 of our NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N W
! (202) 234 4433 WASH!NGTON, D C. 20005-3701 (202) 234 4433 l
159 1 own licensees. The agreement states have perhaps, double fs 2 that number; two to three times that number.
( )
w' 3 MEMBER SEALE: Yes, okay. But in any event, 4 the operational enforcement generally -- in many cases, 5 anyway -- is the agreement state's responsibility, isn't t
6 it?
7 DR. SERIG: No.
8 DR. COMBS: Only in those states that are 9 agreement states where we will allow them --
10 MEMBER SEALE: Well, I mean in the agreement 11 states.
12 DR. COMBS: Yes, where we ceded our authority.
,__s 13 And that's an entirely different relationship. We e s
'~
14 regulate the same materials; in fact, they regulate more I
15 than we do. i 16 MEMBER SEALE: Yes. Actually, one of the 17 reasons that Dr. Powers, who's not here, raised the 18 question. It had to do with the fact that there seems to 19 be a, let's say, disproportionately large number of events 20 of overexposures that involved medical applications rather 21 than reactor activities.
22 And I think if you look at it, in many cases, 23 those overexposures are the result of not radio isotope l I
24 applications but accelerator applications, where people b)
(m 25 are using --
NEAL R. GROSS l l
COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N W l (202) 234-4433 WASHINGTON, D C. 20005-3701 (202) 234 4433 l
j
160 1 DR. CARROLL: They're not concerned with those
,- 2 --
i
\'
3 MEMBER SEALE: I know that, but the things 4 that get the noise about overexposures of people in many 5 cases, involve machines that aren't even in your 6 bailiwick.
7 DR. SERIG: If you read the popular press that 8 is undoubtedly true, because there are many more linear 9 accelerators than there are cobalt teletherapy machines or 10 brachytherapy machines.
11 MEMBER SEALE: Right, that's right. And many 12 of the state agencies who have the state responsibility 13 for the agreement state, have a broader responsibility
/O
- 14 than the NRC's responsibility. I 15 DR. SERIG: Many do.
16 MEMBER SEALE: They do in fact, have 17 responsibility for the regulation of radiation-producing 18 machines. So it kind of gets lost in the translation --
19 DR. SERIG: But if you were to look at our 20 misadministration reports which we have some -- let's pick 21 a number -- 30 a year since 1992 --
22 MEMBER SEALE: Gkay.
23 DR. SERIG: Those are involving regulated 24 materials, and if you were to break those --
p)
( 25 MEMBER SEALE: Oh, sure.
NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N W (202) 2344433 WASHINGTON, D C. 20005-3701 (202) 234-4433
161 1 DR. SERIG: If you were to break those down
,_ 2 you would find cobalt teletherapy accounting for four or
( )
3 five, six a year; you would find remote afterloading 4 brachytherapy accounting for something like the same 5 number; you'd find manual brachytherapy -- which is akin 6 to, but a lot different than remote afterloading 7 brachytherapy -- accounting for maybe half of the 8 misadministrations; and the remainder being nuclear l l
9 medicine, again, injected, ingested, inhaled products.
10 One of the projects that we were involved l 1
l 11 with, with the Office of Research, did in fact cross the i
12 line and look at linear accelerators -- our te]etherapy ;
13 project -- because we realized that one possible source of
("h
(/ 14 error was that people walked back and forth between the 15 two types of machines in hospital settings, and that they 16 were different enough so that you'd have negative transfer 17 of learning effects.
18 And so the teletherapy human factors 19 evaluation did in fact, look at linear accelerators, and 20 there were so many more linear accelerators than cobalt 21 machines. That is sort of biased that way in the report.
22 MEMBER SEALE: Yes, yes.
23 DR. SERIG: So that probably is of more 24 benefit to our agreement states than it is to us, in that 25 it has that broader scope.
NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N W.
(202) 234-4433 WASHINGTON, D C. 20005-3701 (202) 234-4433
1 162 1 CHAIRhWJ APOSTOLAKIS: So let me come back to i
, 2 the real issue. The opening statement of the human
'.]
3 performance program plan is, mission to ensure safety of l
4 nuclear facilities through a vigorous and comprehensive 5 human performance program. And the facilities that you l 6 regulate -- nuclear facilities in this sense -- so you are l
l 7 part of this?
8 DR. SERIG: Yes.
9 CHAIRMAN APOSTOLAKIS: Okay. Now, if I go to 10 your slide 6, human factors in NMSS; initiate human 11 factors evaluation of teletherapy, initiate human factors l
12 evaluation of remote afterloading brachytherapy. So if I i
13 go to the human performance program plan, somewhere here I l 1
( 14 will find these items with a high, medium, or low .
15 priority?
16 DR. SERIG: No, you will --
17 DR. CARROLL: Yes, on page 17. )
1 18 CHAIRMAN APOSTOLAKIS: On page 17.
19 DR. SERIG: You will find that they were 20 completed.
i 21 CHAIRMAN APOSTOLAKIS: I will find that they j 22 were completed, but these were random choices on my part.
23 DR. SERIG: And that we were going to use the 24 information for further assessment as to how to address l
/ \
h 25 the problems that --
NEAL R. GROSS I COURT REPORTERS AND TRANSCRIBERS I 1323 RHODE ISLAND AVE., N W.
(202) 2344433 WASHINGTON D C. 20005-3701 (202) 234-4433 l
t
163 !
1 CHAIRMAN APOSTOLAKIS: Okay, so all of the
, 2 things that you have here are part of the plan? But all ,
NJ 3 doesn't mean 100 percent, but most of them? Is that -
4 correct?
5 DR. SERIG: Well, on page 6 this is history.
6 CHAIRMAN APOSTOLAKIS: All right. Page 7 -
7 then? Achievements. So that must be history, too?
8 DR. SERIG: I think if you turn to page 8 9 you'll find our ongoing activities -- j 10 CHAIRMAN APOSTOLAKIS: Okay. So these are in 11 the performance plan and they have a categorization as 12 high, medium, or low?
13 DR. CARROLL: No. TBD.
OT #
'w/ 14 CHAIRMAN APOSTOLAKIS: But they are mentioned?
I 15 DR. CARROLL: Go to page 17. To TBD on 219. l I
16 CHAIRMAN APOSTOLAKIS: Seventeen. I i
1 17 DR. CARROLL: 2.1.8.
18 CHAIRMAN APOSTOLAKIS: Yes? TBD, TBD, TBD, 19 yes. Radiography. So I guess I don't see what the 20 question was all about. Maybe it was just a complaint !
l 21 that you were not here on the 20th. j 22 DR. COMBS: That's what it sounds -- well, in i l
23 part.
24 DR. SERIG: I think it may be a question of i 25 wait.
NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N W.
(202) 234-4433 WASHINGTON. D.C. 20005-3701 (202) 234-4433
164 1 CHAIRMAN APOSTOLAKIS: That's not what the l
,g
, 2 question -- no. l i 1 V 3 DR. CARROLL: What Dana was getting at is, if 4 you look at any number of places you find -- for example, 5 the report to Congress -- you find in terms of significant i
6 events, they're all in your bailiwick. In fact, the last l
7 several years there haven't been any --
8 DR. COMBS: Reactor events.
9 DR. CARROLL: -- reactor events; or very few, 10 at least. And I think what Dana was saying was, a lot of 11 these events are human factors issues.
12 DR. SERIG: That's correct.
13 DR. CARROLL: So why aren't you guys spending l
\
/~'st
\/ 14 a lot more money trying to deal with these problems than '
l 15 it appears you are?
16 DR. SERIG: Well, I think you caught us in the 17 middle of something where number one, we had a lot of work
(
j 18 looking at the medical misadministrations. We've done, 19 with the Office of Research, two very significant 20 projects. We've also got databases on the medical 21 misadministrations that have provided us some insights.
22 However, we're at the point in the Commission's decision 23 cycle where they're not sure that they're going to 24 continue looking at medical issues, and so we're holding a (O_) 25 couple of things in abeyance -- that's TBD.
NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISMND AW., N W (202) 234-4433 WASHINGTON D C. 20005-3701 (202) 234-4433
165 I DR. CARROLL: I thought maybe you transferred
, 2 to the --
k'~' /
3 MEMBER MILLER: Statra.
4 DR. CARROLL: The FDA.
5 DR. SERIG The exact disposition of what 6 we're going to do with the medical activities is unknown 7 at this time.
8 DR. CARROLL: Who presently regulates, for 9 example, lin. acc.s? The states?
10 DR. SERIG: FDA has approval of devices but 11 they don't have the broad spectrum that we have, and we 12 don't have very much.
13 CHAIRMAN APOSTOLAKIS: Dr. Sealc just pointed O
\~2 14 out to me that on page 30 of the plan, 3.1.7, a project is 15 to develop an NMSS human factors program. Is that what 16 you're referring to when you say everything is on hold?
17 Priorities, medium.
18 DR. COMBS: Yes.
19 DR. SERIG: Yes.
20 CHAIRMAN APOSTOLAKIS: And that, what you are 21 describing in the sljdes --
22 DR. SERIG: And that program -- I'm not sure 23 of the wording you have there -- but the intent is to 24 include 1: that program development, what we've learned
(~N
( ,) 25 from using our databases, what we've learned from the NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERE 1323 RHODE IFLAND AVE., N W (202) 234-4433 WASHINGTOb', D C. 20005-3701 (202) 234-4433
166 1 human factors evaluations, and how to integrate those 7, 2 findings into activities that will affect public health 3 and safety -- improve public health and safety.
4 CHAIRMAN APOSTOLAKIS: Well, it's only one 5 sentence; I can read it to you. Description. Develop an 6 NMSS human factors program plan designed to integrate as 7 appropriate, the goals and objectives of the human 8 performance plan into the process of regulating materials 9 licensees.
10 DR. SERIG: Right.
11 CHAIRMAN APOSTOLAKIS: So you will integrate I
12 the goals and objectives of the human performance plan --
13 DR. SERIG: And in fact, some of them already
(~h
\ #
w/ 14 are. There are some activities in there that are intended 15 to improve the knowledge of other individuals within the i l
i l
16 NRC who rub elbows with human errors in their day-to-day l
17 life -- inspectors, for instance -- but we've already done l
18 some of that.
19 And what we will simply be doing is l
20 formalizing some of the things that have already been l
21 happening --
l 22 CHAIRMAN APOSTOLAKIS: So do you feel that you 23 are a part of the human performance plan and that you have 24 had input to it?
( ,
, 25 DR. SERIG: Yes.
NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N W.
(202) 234-4433 WASHINGTON, D C. 20005-3701 (202) 234-4433
167 1 CHAIRMAN APOSTOLAKIS: Commensurate to your g3 2 importance?
k' ' 3 DR. SERIG: Yes.
4 DR. COMBS: Of course not.
5 DR. CARROLL: If our criteria for importance 6 is people killed, they are much more important that the 7 reactor guy.
8 DR. SERIG: Commensurate with our resources, 9 but perhaps not with our --
10 CHAIRMAN APOSTOLAKIS: Well, but the resources 11 I think, is something that the plan does, right? Besides 12 on the allocation, by assigning these medium, high, and 13 low?
U 14 DR. THOMAS: Yes. But I would say present j 15 resources. As former chairman of the coordinating 16 committee I repeat myself. Dennis participated in, I 17 think, almost every single meeting of the coordinating 18 committee and the development of the plan, and I would say 19 it was commensurate with the resources that they had at 20 the time.
21 The resource implications of the plan are more 22 in the future -- are in the future at least as much as the i
l 23 present. Priorities, resources, schedules, and so on.
24 CHAIRMAN APOSTOLAKIS: And you're saying that O
Q 25 the commission is not sure whether they want to continue l
l NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N W.
(202) 234-4433 WASHINGTON, D C. 20005-3701 (202) 234-4433 l
168 1 regulating this --
2 DR. SERIG: That's correct.
3 CHAIRMAN APOSTOLAKIS: So as far as I'm 4 concerned, you answered question 7. Any members have any 5 questions? The problem seemed to be that you were not 6 here on the 20th of September, now you're here.
7 DR. CARROLL: I don't think that's it.
8 CHAIRMAN APOSTOLAKIS: Plus they don't get the 9 -- I think that was the problem.
10 MEMBER SEALE: I think Dana's concern was the 11 high profile that misadventures involved in medical 12 applications, be they radioisotope or radiation-producing 13 machine in origin, had -- and the fact that that seemed to
\/ 14 be relatively absent from what we had been hearing --
15 DR. SERIG: I think it has in fact, been the 16 major activity of NMSS's human factors program. We've 17 done a large number of things, and if the plan does not 18 give you that impression, you know, I understand that 19 fully. But there is a great deal of -- has been a great 20 deal of work in the past and -- I think we skipped over a 21 slide that's of some importance here.
22 MEMBER MILLER: Which one is that?
23 DR. SERIG: There's one that sort of tells you 24 where our starting point was, and our licensees -- page 5, (3
1
,/ 25 the last bullet -- really had no appreciation of human NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N W.
(202) 234-4433 WASHINGTON. D.C. 20005-3701 (202) 234-4433
169 1 error or human factors -- I mean, we think of human error p 2 as a 2-word phrase; they always added a third word, and O 3 some continue to do -- random human error. As in, I can't l
i 4 do anything about it. i 5 Medical community in particular, was prone to 6 do that. We've seen over the last eight years, something i 7 of a turnaround. I just participated in a conference in 8 October that was sponsored jointly by the American Medical Association, the Joint Commission on Accreditation of l 9 l l
10 Hospitals, and a number of other notables.
11 And they were specifically looking at error in 12 health care. They were bringing in resources from 13 outside: people who had experience with systems and 0 14 addressing human error; folks from the aviation industry, 15 folks from some of the chemical processing industries.
16 And I was among one of the people selected to make a 17 presentation.
18 And I described the efforts that our folks in i 19 Research did, and indicated the value that we thought ;
20 those had given us to date, and would in the future. And 21 one of the announcements at that meeting was the formation ),
l 22 of something called the National Patient Safety -
23 Foundation, which we'll carry on. We're looking 24 specifically at human error in medicine now. !
]
l b 25 So I think, we started from well back of where j 1
NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N W.
(202) 234-4433 WASHINGTON. D C. 20005 3701 (202) 234-4433
1 170 1 the reactor community was in terms of knowledge of human
,, 2 error and its causation and how to address it. At least '
(~) 3 we had EPRI working on human error and human factor as l
4 well before Three Mile Island. We had nothing in the '
i 5 medical community, and they've really come a long way.
l 6 DR. CARROLL: Let me ask this question. Let's
]
i 7 pick the last three years. Of the bucks Research is i 8 spending on human factors issues, what percentage goes 9 towards -- i 10 DR. SERIG: Jay, can you help me?
11 DR. CARROLL: -- and radiation uses?
12 DR. SERIG: We published in '95. !
13 DR. PERSENSKY: I think probably -- well, I
(~)
U 14 know in the last two years we've spent no money on NMSS 15 directly.
16 DR. CARROLL: Spent what?
17 DR. PERSENSKY: No -- zero. The work that we 18 did -- as far as user needs were concerned. We have done 19 some work in the modeling -- that I'll talk about a little 20 bit later -- that relates to use of the data that we 21 collected for NMSS. The two studies that Dennis described 22 were completed about two years ago and I think they went 23 on for about a 3-year period and we probably spent about 24 ten percent of our budget on those efforts. l
/ \
Q}
25 DR. CARROLL: What's the ratio of the NEAL R. GROSS l COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N W.
(202) 234-4433 WASHINGTON, D C. 20005-3701 (202) 234-4433 l
)
i 171 l
1 fatalities that -- these activities as opposed to
-- 2 reactors?
3 DR. SERIG: Well, I don't think reactors have 4 had any fatalities, and we would probably claim one. We 5 would probably claim one over the last four years.
6 CHAIRMAN APOSTOLAKIS: So the ratio is 4 I
7 infinite.
8 MEMBER MILLER: Over the last three years?
9 DR. SERIG: Four years. ;
10 MEMBER MILLER: Now, the largest one was a 11 Canadian medical --
12 DR. CARROLL: That was a lin. acc.
13 MEMBER MILLER: That was a lin. acc., that's
)
\/ 14 right; that was a programming -- that was a --
15 DR. SERIG: They're at 25, but those cases 16 were in the United States.
17 MEMBER MILLER: Yes, but if any of them were 18 in the United States you know those Canadians --
19 DR. CARROLL: Help me out on that one. I have 20 read Nancy Levinson's paper and I just can't believe that 21 she wasn't overstating the incompetence of the medical 22 community in that paper. Have you read her paper?
23 DR. SERIG: No, I haven't, but I understand 24 that there was some programming problems, not just --
( 25 DR. CARROLL: Yes, but --
NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 (202) 234-4433
l 172 l i
1 DR. SERIG: -- problems.
_ 2 DR. CARROLL: They were zapping people with
'd 3 that machine and they were getting obvious burns and so 4 forth, and according to her story, nobody could see a l
5 problem. They just didn't correlate the fact that they i i
6 were injuring people.
7 DR. SERIG: In some cases I think -- and this 8 is from a non-medical person's appreciation -- but I've 9 visited some hospitals and maybe Josie can help us because 10 she does have a medical background -- the symptoms you 11 get from excess radiation are masked by the disease ;
12 symptoms. I mean, you simply don't know until you stack 13 up a lot of injuries, that you've got a problem.
ry
( )
s/ 14 MEMBER MILLER: Well, there's one situation in 15 my home town some years ago, where a radiation physicist 16 made a mistake on a calculation. A lot of people 17 shortened their lives, I guess the way to put it.
18 DR. SERIG: Riverside Hospital?
19 MEMBER MILLER: Riverside, right. And -- that 20 a human factors there, I don't know.
21 DR. SERIG: Yes.
22 MEMBER MILLER: I'd say it was, yes.
23 DR. SERIG: les.
24 CHAIRMAN APOSTOLAKIS: I'm at a loss here.
O 25 What is the issue we are discussing?
( ,/
NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N W.
(202) 2344433 WASHINGTON. D.C 20005-3701 (202) 2344433
173 1 DR. CARROLL: I think we've exhausted the 2 subject.
I-(_/
3 CHAIRMAN APOSTOLAKIS: Okay, we've exhausted 4 the issue I don't understand. You are part of the human 5 performance plan?
l 6 DR. SERIG: Yes.
7 CHAIRMAN APOSTOLAKIS: Okay.
l 8 DR. SERIG: And we have an active program and l l
9 will continue -- l 1
10 CHAIRMAN APOSTOLAKIS: You have an active 1
11 program so the problem was that on the 20th of September i l
1 12 you were not here? Plus Dana and others felt that they 13 were not getting sufficient resources here, is that 1 t )
kl 14 correct?
15 MEMBER MILLER: Well, I think that's still I l
1 16 probably the situation.
17 CHAIRMAN APOSTOLAKIS: Yes. )
l 18 MEMBER SEALE: I think the next question is, l
19 the shoe hasn't dropped on what the agency is going to do 20 in this area, and when it does, it may very well be that i
j 21 this question of appropriate resources for human factors l
22 should be raised again.
23 CHAIRMAN APOSTOLAKIS: Yes.
24 DR. COMBS: We'll decide what the mix of tools
,Q
(/ 25 are to solve the problem --
NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N W.
(202) 234-4433 WASHINGTON, D C. 20005-3701 (202) 234-4433
174 1 CHAIRMAN APOSTOLAKIS: And you will be there, n 2 next time?
e s V 3 DR. COMBS: Yes.
4 MEMBER SEALE: And if we can help you, let us 5 know.
6 DR. COMBS: Thank you.
7 DR. CARROLL: What's the extreme scenario?
8 That you get rid of all the --
9 MEMBER SEALE: Yes, that's one of the options 10 --
11 CHAIRMAN APOSTOLAKIS: Excuse me. What is 12 extreme scenario?
13 DR. COMBS: One extreme scenario is that we t')
(
14 would be completely out of medical regulation.
15 DR. CARROLL: But still --
16 DR. COMBS: Still in industrial, right.
17 CHAIRMAN APOSTOLAKIS: Who will take over the 18 medical? We don't know?
19 DR. COMBS: We don't know.
20 CHAIRMAN APOSTOLAKIS: Okay.
21 DR. SERIG: That takes --
22 CHAIRMAN APOSTOLAKIS: And the other extreme?
23 DR. SERIG: -- an act of Congress to make 24 decisions like that.
/"\
V 25 MEMBER SEALE: It stays essentially as it is.
NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N W.
(202) 234-4433 WASHINGTON, D C. 20005-3701 (202) 234-4433
175 1 CHAIRMAN APOSTOLAKIS: Or you get more
,- 2 resources to?
( ,)
3 DR. COMBS: Right, to do it. Or we do more of 4 what we've been doing.
5 CHAIRMAN APOSTOLAKIS: Okay.
6 MEMBER MILLER: I only have anecdotal 7 information. But where I come from, a major university 8 with comprehensive radiation use -- at Ohio State you 9 probably know as well --
10 CHAIRMAN APOSTOLAKIS: Right.
11 MEMBER MILLER: Anecdotally, it's been known 12 for me, your group has had far less funding than the 13 reactor group has for years and years and years, for your O) t
\- 14 responsibilities. That's just anecdotal. Applying to j 15 colleagues you take care of our radiation and protection l l
16 program at Ohio State, which is, I say, the medical area l
l l 17 is very complicated there. So you're not telling me i 18 anything I haven't heard before anecdotally.
l 19 CHAIRMAN APOSTOLAKIS: So why are you smiling?
20 MEMBER MILLER: Why? Because I'm telling him j 21 what he already knows.
22 CHAIRMAN APOSTOLAKIS: Exactly. Well, it 23 seems to me I should thank you very much for coming here.
24 MEMBER MILLER: I appreciate the information l
/N
(_) 25 and always anxious to learn about other uses of radiation.
NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N W.
(202) 234-4433 WASHINGTON, D C. 20005-3701 (202) 234-4433
176 1 CHAIRMAN APOSTOLAKIS: Thank you, sir.
p._
2 DR. COMBS: Thank you.
t i
\~
3 CHAIRMAN APOSTOLAKIS: This was very 4 informative. And we're back to discussion now. Which 5 question are we going to discuss?
6 MEMBER SEALE: Eleven. Yes, 11 is the next 7 one on the list.
8 CHAIRMAN APOSTOLAKIS: So we have 11, 12, 13, 9 14, and 15? Can we do all this in an hour, you think?
, 10 MEMBER SEALE: Oh, yes.
11 DR. CARROLL: As long as the chairman 12 restrains member Apostolakis from asking too many j 13 questions.
/^)\
(~ 14 CHAIRMAN APOSTOLAKIS: I don't ask questions.
15 All them are coming from my left.
16 DR. COFFMAN: Question 11 is: how are 17 standards adopted by the staff formulated, and how does 18 the staff assure that the standards are necessary and 19 sufficient to meet the regulatory needs? Yes sir, I took 20 it as a loaded question.
21 But I think the basic answer is that consensus 22 standards are based -- I think the heart of the issue is 23 consensus standard. And consensus standards are based 24 upon experiential evidence and are used in lieu of CT (j 25 resources to develop experimental bases.
NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS i 1323 RHODE ISLAND AVE , N W. l WASHINGTON, D C. 20005-3701 l (202) 234 4433 (202) 234-4433
177 1 There's just no guarantee of sufficiency when f.s 2 using consensus standards, and so the scope of the
() 3 standard has to necessarily not exceed the scope of the 4 experiential basis.
1 5 And the only other point I can make is that 6 the staff is encouraged primarily for resource efficiency, l
7 to rely upon consensus standards as much as is defensible.
8 MEMBER SEALE: I understand your dilemma and j 9 there's an aspect of the reliance on standards, though, 10 that I think is interesting. Recently, the committee has 11 been looking at standards in a couple of applications.
12 One has to do with materials, where consensus 13 standards are used on things like tube ruptures and that
(~
k-)b 14 sort of thing. And then another is on the I&C plan where 15 the standard basically is a process, rather than a 16 product-oriented standard. ;
17 And so the discussion is with respect to the 18 process that's used to develop hardware and software to 19 meet the I&C needs, rather than the performance of the 20 product.
21 MEMBER MILLER: Well, there's still debate on 22 that one.
23 MEMBER SEALE: Yes, I understand, but that's -
24 -
/~'N
(_,) 25 MEMBER MILLER: That's a consensus standard, NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 (202) 234-4433
.i 178 l
1 first of all, those were not developed in the nuclear !
,c. 2 industry.
( 4
' %J MEMBER SEALE:
3 I appreciate that. I'm talking 4 about standards in. general. Now, in the human factors i
5 area we've said we're going to use a similar approach and l
6 that approach is going to be a consensus approach.
l 1
7 The problem I have is, in the case of pipe !
8 rupture, I know what a failure is. Are we sure we know 9 what constitutes failure in the human factors arena?
10 DR. COFFMAN: Only when it's defined in 11 specific context, and sometimes the context gets so 12 detailed that you just can't generalize.
13 MEMBER SEALE: That's right. And it bothers rs l
14 me that, you know, in the standards that are developed --
15 you say they're consensus standards -- but in some of 16 those standards I can't find what specifically is the 17 on/off switch for failure. There's a discussion of the 18 standard way of doing things, but how do I know when I've 19 failed?
20 DR. COFFMAN: The standards take on the nature 21 of more like recommended practices. l 22 MEMBER SEALE: Precisely, precisely. So it's 23 not exactly the same thing as a consensus standard based 24 on what's happened to all the pressure vessel performance fq 25 data since the Robert E. Lee?
NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N W.
(202) 234-4433 WASHINGTON. D C. 20005-3701 (202) 234-4433
179 1 DR. COFFMAN: No, because we don't have the g3 2 detailed measures.
i )
</
3 MEMBER MILLER: Well, all standards group --
4 they have both standards and recommended practices, so --
5 what standards -- who's the mainstay of standards in human I
6 factors?
7 DR. COFFMAN: Well the responsibility -- RES 1
8 has the responsibility for --
9 MEMBER MILLER: No, what professional 10 organization? Is it IEEE?
11 DR. COFFMAN: IEEE, ANS --
i 12 DR. PERSENSKY: ANS has some standards --
13 depends on what it is. ANS says the standard is 3.5 -- l t i
\~/ 14 from the standpoint of 3.5 fo_ simulators, 3.4 is for the 15 medical, 3.1 is training and qualifications. IEEE has a 16 set of standards designed more towards the measurement of 17 performance and the interface. ANS doesn't have an l
18 interface standard, for instance. ,
19 MEMBER MILLER: But they're all standards 20 developed specifically for the nuclear -- !
l 21 DR. PERSENSKY: Those particular ones are for l i
22 the nuclear or for the power. IEEE is not necessarily l 0
23 only for nuclear -- !
24 DR. SERIG: There is for the medical 25 community, a medical device standard that the Association 4
NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N W (202) 234-4433 WASHINGTON, D C. 20005-3701 (202) 234-4433
180 1 for Advancement of Medical Instrumentation is responsible 2 for. It's a 5-year program process and it is currently an 73
(
)
3 ANSI standard.
4 MEMBER MILLER: What about human factors, 5 though?
6 DR. SERIG: Human factors? It is human 7 factors for medical devices, human factors -- engineering 8 guidelines and preferred practices for medical devices.
9 MEMBER MILLER: So -- but I guess the question 10 is, all the standards that we depend on in the nuclear 11 business are specifically designed for nuclear occasions, 12 or do we use generic standards or human factors for just 13 anything?
14 DR. PERSENSKY: Well, in part, it may be that 15 Dr. Seale's talking about, where we gave a presentation at 16 the last meeting regarding our adoption of standards for 17 instance, from the military. The military --
l 18 MEMBER MILLER: Okay. Mil spec.
19 DR. PERSENSKY: Mil specs that we have adopted l 20 or adapted to the nuclear community. For instance, 0700; 21 0700 is based primarily on Mil Spec 1472 and other ones l i
l l 22 since then. So they are design standards and purchasing l l
23 specifications in a sense, that we have adapted into our 24 guidelines, which must be more good practices perhaps, O
( ,)\ 25 than a yes or no.
NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N W l (202) 2344433 WASHINGTON. D C. 20005 3701 (202) 234-4433
l 181 !
1 DR. CARROLL: This necessary and sufficient ,
i I
i 2 issue comes from Dana Powers, and last I talked to him l i /-s i N.~)T 3 about it, two or three months ago I guess, this was a 4 question that DOE was asking itself regarding the adoption 5 of standards for DOE facilities. Sounds like something 6 some clever lawyer came up with. It's a tough question to j 7 answer. l 8 CHAIRMAN APOSTOLAKIS: I think it's way 9 premature to ask that question in the human factors area, 10 is it not? I mean, especially when it comes to human I
11 error. And we don't even have models to understand what's 12 going on. Now to start developing standards.
13 DR. CARROLL: Well, it came up with Dana in O i
%# 14 connection with tae, principally the IEEE standards in the 15 digital I&C.
16 CHAIRMAN APOSTOLAKIS: Yes, that's an entirely 17 different story. You see, the difference is that they 18 don't have model -- but they do have standards.
19 MEMBER MILLER: That's close.
20 CHAIRMAN APOSTOLAKIS: Shall we move on to 127 21 Or if somebody has a burning question regarding 11?
22 DR. COFFMAN: Well, I think there was a second 23 part of the question dealing with the focus on process 24 versus product.
,a
!j 25 CHAIRMAN APOSTOLAKIS: On 11?
NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N W.
(202) 234-4433 WASHINGTON D C. 20005-3701 (202) 2344433 i
182 1 DR. COFFMAN: Well, in the discussion --
-s 2 CHAIRMAN APOSTOLAKIS: In the I&C there was.
t )
3 I mean, if you want to discuss that, raise that issue 4 here, that's fine with me.
5 DR. COFFMAN: No, I took it that Dr. Seale l
6 were -- l 7 MEMBER SEALE: No, I just -- I was trying to 8 draw the distinction between a material standard where it ;
9 was more a performance standard; that is, it failed or it 10 didn't fail kind of thing. Whereas, the other standard 11 was strictly a process-oriented standard.
12 DR. COFFMAN: Okay. l 13 MEMBER SEALE: But they're both standards. !
{
\_- 14 DR. CARROLL: Before I can understand the 15 answer somebody's going to have to explain to me what's 16 wrong with the last sentence of 12?
17 DR. PERSENSKY: It should be a "do not exist".
18 DR. CARROLL: Oh, do.
19 DR. PERSENSKY: That's the way I read it, at 20 least. The staff schedule in Item 1.2.11 of HPPP, develop 21 guidance for computerized job performance aids is to be 22 completed as technology is developed.
l 23 What standards does the staff have for such 24 aids that would foster the development of such technology?
() 25 If standards do not exist, what are the staff plans for i NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N W.
(202) 234-4433 WASHINGTON, D C. 20005-3701 (202) 234-4433 l
i I
183 1 developing such standards? !
f3 2 To start off with, it's not the role of the
(
)
3 HPPP or the NRC, to foster development of technology.
l l
4 Just to set that straight. Our role is to review and l 5 evaluate industry technology and perform confirmatory 6 research necessary to support the review and evaluation.
l 7 Now that being said, what we are doing i
8 essentially, is following what the industry is doing; try '
9 to keep up-to-date. We have another -- we've had a number 10 of efforts to keep in touch with the state-of-the-art:
11 SERIAC has done literature search; we've held a workshop 1
12 with the industry; and we've also had an SBIR project on l
13 sort of describing what the existing standards are, r~n
( '
)
\~/ 14 DR. CARROLL: What is SERIAC? l 15 DR. PERSENSKY: SERIAC is the Crew Systems 16 Ergonomics Research Information Center.
17 DR. CARROLL: You thought that was the 18 information --
19 DR. PERSENSKY: Sorry about that. It's part 20 of the Departmant of Defense information centers, and it's 21 run out of Wright-Patterson Air Force Base. And it is 22 primarily the human factors information center for the 23 Department of Defense.
24 We've also done a few evaluations. There was
{')N s_ 25 a NUREG/CR-6398 which was an attempt to evaluate NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N W.
(202) 234-4433 WASHINGTON, D C. 20005-3701 (202) 234-4433
184 1 computerized procedures versus hardcopy procedures that we g- 2 did at the North Carolina State University. And we also N.]I, 3 used that as a basis for some of our modeling studies.
4 The future plans. NRR has plans to come up 5 with a supplement for NUREG 0711 which is the review 6 process that will incorporate more information on 7 computerized procedures because we're seeing the 8 possibility of more of them coming into play.
9 In the study I mentioned before, the hybrid 10 study, it came up that computerized procedures, since 11 they're probably one of the more likely things to occur in 12 a hybrid control room, that is one of the areas that 13 they're going to be studying further.
7_
(' )
'/ 14 The Halden reactor project which we fund has 15 developed their own computerized procedure program called i 16 CAMPA; and that's an ongoing effort to evaluate that and 17 to use it. There's been some interest in, actually some 18 of our vendors, in incorporating that.
19 So the plan right now is just to keep up-to-20 date and to continue, particularly through the hybrid 21 study, some of the work we've been doing in terms of 22 looking to what standards exist on the outside.
23 Any questions?
24 CHAIRMAN APOSTOLAKIS: Thirteen.
r-k_N) 25 DR. PERSENSKY: Thirteen?
NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N W.
(202) 234-4433 WASHINGTON, D C. 20005-3701 (202) 234-4433
185 l
1 CHAIRMAN APOSTOLAKIS: I suppose.
s 2 DR. PERSENSKY: Pardon?
\ i
)
3 CHAIRMAN APOSTOLAKIS: Isn't 13 next?
4 DR. PERSENSKY: Yes, 13. What is the staff 5 approach to develop a performance-based fitness for duty l 6 criteria? I believe the plan indicated that we were going l l
7 to maintain awareness of what's going on. Again, this is 8 one of these efforts where the practice is -- in other 1
9 industries, there's a lot more being done in 10 transportation, I guess some of the more well-known --
11 UNIDENTIFIED: National Football League.
12 DR. PERSENSKY: National Football -- but we're 13 trying to track research that's going on in the
\~) 14 government. Department of Transportation has a number of 15 projects going on in this area. Various industries have 16 incorporated it. United Postal Service, or something like l
17 that, has done this for their drivers, using various types l
l 18 of behavioral measures rather than urinalysis or blood 19 tests.
20 There are a number of vendors that are coming 21 up with methods, but probably the most telling thing is 22 NUREG/CR-6470 which is titled, " Fitness for Duty in a 23 Nuclear Industry and Update of Technical Issues 1996". In 24 fact, did a review of current behavioral measures.
,a
(,j) 25 They identified the fact that these tests do NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N W.
(202) 234-44',3 WASHINGTON, D C. 20005 3701 (202) 234-4433
186 1 help determine if there is impairment from a very broad
,-s 2 range of agents. One of the things that we're talking V 3 about here of course is, to perhaps include fatigue as one l 4 of the fitness for duty issues as opposed to just drugs, 5 because this would pick that up whereas urinalysis would 6 not, necessarily.
l 7 It does detect impairment but not the cause of :
l I
8 the impairment, and of course, these tests are not 9 predictive. They can only tell you what the situation is 10 at one particular point in time; whether it's when they 11 walk into the gate or if it was done on some for-cause 1
12 basis later on in the shift.
13 DR. CARROLL: It may become very important now l p
k-s 14 that -- as long as you have a note from your doctor in 15 California, you can smoke pot on the job.
16 MEMBER MILLER: And the qualification of the 17 doctor is not too well specified. j 18 DR. CARROLL: Not at all. Boy, people are 19 dumb what they voted for.
1 20 MEMBER SEALE: Yes, that's right. Just --
21 DR. PERSENSKY: Sure.
22 MEMBER SEALE: There are some things that have 23 gone on in the industry, though. I remember hearing about 24 a program that ANO1 had, that involved some test
() 25 activities, and they were able to do things because they NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISMND AW., N W (202) 234-4433 WASHINGTON. D C. 20005-3701 (202) 234-4433
187 1 didn't have a strong union.
2 And, you know, not that this is an excuse to h
(a 3 go around and beat up on labor unions, but it does suggest 4 that there are a lot of external influences on this kind 5 of problem, maybe more than we might originally think of, 6 that need to be considered in coming up with appropriate 7 approaches.
8 So I guess I'm just saying that the problem is 9 maybe even more compounded with difficulty than one might !
10 first presume, if you think about it.
1 11 DR. PERSENSKY: Yes, I think we're aware of 12 some of these tests that have been going on, and we do 13 know that this whole area of fitness has been a big
(.s 14 question, but I think one that will be with us for some 15 time.
16 MEMBER SEALE: But the other problem that i
17 bothers me is, the real problem from a safety point of !
l l
) 18 view is the impairment -- ;
l l
19 DR. PERSENSKY: Correct.
l 20 MEMBER SEALE: -- not what caused it. Now, l
21 you may not want to keep the guy around if he's abusing j 22 something, but the real question is --
23 DR. PERSENSKY: Can I do my job right now?
24 MEMBER SEALE: That's right. And if that's 25 the case, then the case for the fitness for duty test if NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N W.
(202) 234-4433 WASHINGTON. D C. 20005-3701 (202) 234-4433
1 188 1 you will, is really a much more general question than just 2 abuse of any kind of narcotic or anything like that. You
(-
s) 3 know, it strikes me that that's a very important 4 distinction to make. j 1
5 DR. PERSENSKY: Well, I think that's why a 6 behavioral test may be a more appropriate test --
l 7 MEMBER SEALE: Exactly, exactly. !
8 DR. PERSENSKY: And perhaps even more j 9 acceptable to the unions or -- l 10 MEMBER SEALE: Well, that's my other point, 1 l
11 because that's the real safety issue and now not a 12 question of what particular chemical culture you happen to 13 belong to. It strikes me that that's a legitimate basis 14 on which to enforce a testing program.
l 15 DR. CARROLL: Yes, but I'm afraid that might 16 have been the philosophy some years ago, but the 17 government and private companies made such a thing out of 18 a drug-free workplace -- that I'm not sure -- the drug-19 free workplace has become so ingrained I don't think it's 20 reversible.
21 MEMBER SEALE: I'm not saying you necessarily 22 have to reverse it, but I don't think you have to --
23 DR. PERSENSKY: I think there may be 24 approaches as well, in terms of, if you use the behavioral q,/ 25 test as the initial screen, and if there's reason to l
NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N W.
(202) 234-4433 WASHINGTON. D C. 20005-3701 (202) 234 4433
189 1 believe that you want to go further with it, then to go eS 2 into the more intrusive testing.
A 3 DR. THOMAS: From NRR's point of view, one of 4 the aspects that we're involved here is the one of working 5 hours5.787037e-5 days <br />0.00139 hours <br />8.267196e-6 weeks <br />1.9025e-6 months <br /> and fatigue. Fatigue is mentioned explicitly in 6 Part 26, and several years ago we considered whether or 7 not we could appropriately amend the Commission's policy 8 statement on working hours because it had been out of date 9 for some time, based on more current research.
10 And we looked at the different applications or 11 different ways, and really, working hours are a way to 12 attempt to control fatigue, but it's a very poor way.
13 This would be a more direct measure of it. But we 7_
14 couldn't even make a case, although there had been a few 15 documented cases where fatigue had entered into events as 16 causal factors.
17 We couldn't make a case to even recommend to 18 the Commission to modify their policy statement on working 19 hours2.199074e-4 days <br />0.00528 hours <br />3.141534e-5 weeks <br />7.2295e-6 months <br />, let alone to go forward with any kind of 20 requirement for a fitness for duty requirement to measure 21 fatigue.
22 DR. PERSENSKY: Fourteen? Question 14: What 23 is the staff approach to evaluating task network model 24 espoused by the Department of Defense, and how will the O
(sl 25 staff decide if the model is applicable and useful for NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N W.
(202) 234-4433 WASHINGTON. D C 20005-3701 (202) 234-4433
190 l
1 regulatory needs? l i
l 7s 2 The approach that we're taking -- l k ,}
s l
3 DR. CARROLL: You've got to tell me what task ;
4 network model is first.
i 5 DR. PERSENSKY: Okay. A task network model is !
6 really a shell; it's not a model in the same sense that 7 we've been talking earlier. It's a way of describing 8 behavior in terms of actual tasks and functions and the 9 steps associated with it. Because you can go through 10 multiple steps you get a network out of it.
11 We started this process or this evaluation, 12 look at this effort, some years ago; actually through some 13 small business innovative research grants that had been
(
\
14 given in terms of how we might improve the data that we 15 can use from a research perspective as opposed from a 16 regulatory perspective. This work has been done almost 17 all entirely on research initiative rather than user need.
18 We looked at various promising techniques in 19 terms of various things that had been proposed to us 20 through the small business innovative research program.
21 And at that point selected a candidate -- in this case, 22 task network modeling -- there have been some feasibility 23 studies.
l t
24 We're about at this point here in the study, 25 where we've done some targeted validation studies in terms i
NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISMND AVE., N W (202) 234-4433 WASHINGTON, D C. 20005-3701 (202) 234-4433 l
191 1 of trying to do a shadow study of some laboratory effort
,- 2 that we were doing that was more data-rich. We're also
(\ ') 3 looking now at refining enhancing the technology of the 4 model itself.
5 The next kind of things we want to do is to I 6 link it more to the safety assessment and risk models, i
7 perhaps. The status -- as you can see the blue ones at 8 the top -- that you can't see -- are generally the data j 9 elements of the work. That was a procedure study where l 10 they compared hard copy procedures to computerized 11 procedures.
12 We did a modeling study that shtdowed that by 13 using the paper procedures data as a baseline in trying to
()
k/
m 14 predict using the model -- to predict what would happen if j l
i 15 we went to computerized procedures. That one came out so-16 so. There are NUREG/CRs on this.
17 We've also done some data collection in the 18 area of staffing at BNL which we built the model around 19 that. That's currently in a letter report. We're 20 awaiting the report from Halden on the Loviisa staffing 21 study which was one of the parts of the advanced reactor 22 staffing project.
23 We've also been looking at some utility event 24 data and trying to model that for some smaller human
() 25 factors models. We've also -- down here, the NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N W.
(202) 234-4433 WASH:NGTON. D C. 20005-3701 (202) 234 4433
192 1 brachytherapy -- that's where we've done some work for s 2 NMSS on taking a task network -- or, the task analysis
( \
V 3 that was done for the NMSS work and trying to convert that .
I l
4 into a model of -- human operator model from that data. !
5 The work has been done almost entirely by one 6 contractor, but through peer review of others that look at 7 the work. Currently, the project we have going on is done 8 through the Army Research Laboratory because they have the 9 most experience in this modeling technology. ;
10 They have developed a number of models for i
11 their use in purchasing equipment and testing equipment 12 for the Army. And through that task order we're now doing 13 work on the medical systems, alarm reduction study, and t
\~- 14 also the integration of modeling with both PRA models, or 15 risk analysis models, and models of the plant, so we get a 16 closed loop system.
17 So that's kind of where we are now. We would i 18 hope that if all this validation effort where we compare 19 data rich environments with the model do prove to be 20 valid, that we may be able to get more involved with use 21 of the model in a regulatory environment, rather than just 22 the research environment.
23 Thank you.
24 DR. COFFMAN: Fifteen is, how does the staff
) 25 decide on the allocation of resources between human l
NEAL R. GROSS i COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N W.
(202) 234-4433 WASHINGTON, D C. 20005-3701 (202) 234 4433 l l
l
193 1 factors research and other research activity such as
- 2 thermal hydraulic modelm
\
['~'w/
3 DR. CARROLL: I wonder who asked that 4 question?
5 CHAIRMAN APOSTOLAKIS: I think it was Ivan.
6 DR. COFFMAN: I'd like to address it in two 7 parts. Just speaking generally about the process, just 8 very briefly to summarize the process.
9 Process flows up generally, from the branches 10 to the divisions to the office where the final allocation l l
11 is made. And judgments are made based upon some common 12 elements at each one of those levels; generally things 13 like the directness to which it ties to safe operations i
k/ 14 and maintenance.
15 So its coupling to safety is generally one of 16 the first considerations. The relationship to the formal 17 user need, whether it's an existing user need or an 18 anticipated user need from one of the regulatory offices, 19 comes into play in determining the relative weights of 20 budgets given among the different topical areas that are 21 addressed in research.
22 And then the very subjective, the degrees to 23 which, you know, experience, previous experimental l
24 results, analysis, maybe insights, determine how it q,,) 25 relates to regulatory issues. Nothing -- I mean, that's NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE , N W.
(202) 234-4433 WASHINGTON. D C. 20005-3701 (202) 234 4433
194 l 1 just a general characterization, probably what you already g) 2 knew. And that's from the branch chief perspective on the 3 thing.
4 What we did last year was a little more j 5 specific, but we're not going to follow the process this l 6 year though, because we're in the middle of reviewing the 7 direction-setting issues and strategic planning, so we're 8 not going through, at least from the branch level, this ;
9 formal process.
10 But last year we went through formal process 11 of answering five questions on each of the topical areas, 12 and then the judgments were made based upon the answers to 13 these questions. And the safety significance was just i \
l 'd 14 rated in three categories: high, medium, or low. And 15 here again, it was the tie to safe operations.
i 16 And then regulatory significance was not in 17 contrast with safety significance but was supplemental to l
l 18 safety significance in that what have the results to date, 19 what have been the results to date? And does this work l
20 help maintain expertise that would be necessary over the 21 long run?
l 22 So there's more organizational -- well, I mean 23 that in the sense of the more agency-related concerns were 24 addressed in our justification in each one of the topics o
(>\ 25 as we addressed these questions.
NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N W.
(202) 2344433 WASHINGTON, D C. 20005-3701 (202) 234-4433
195 1 Likelihood of success and timely closure. You
,- m 2 know, has similar work on other projects been successful?
[G 9 3 Are we trying to recover sunk costs? Those types of 4 considerations were brought into play in addressing that 5 question. Incremental benefit versus incremental cost is 6 always a consideration.
I 7 And then finally, there was the consideration, l
8 or the question addressing whether there was the 9 opportunity for leverage working in cooperative agreements 10 with foreign research institutes or standards bodies. But 11 was there any opportunity to leverage funds?
12 DR. CARROLL: That one bothers me. That's 13 different than the other four.
(/,-~.l
% 14 DR. COFFMAN: Yes sir.
15 DR. CARROLL: That's saying, if I can get it )
l 16 done for cheap maybe I should do it. I think the real 17 question is, should I do it?
18 DR. COFFMAN: I think that's the first 19 question --
20 CHAIRMAN APOSTOLAKIS: It changes number 4.
21 Because if you can leverage it, the ratio becomes --
22 DR. CARROLL: Yes, but I don't think whether 23 you can or you can't should --
24 CHAIRMAN APOSTOLAKIS: And I have the same p).,
( 25 problem with the third ene.
NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N W.
(202) 234-4433 WASHINGTON, D C. 20005-3701 (202) 234-4433
i 196 1 DR. COFFMAN: But if we --
7, 2 DR. CARROLL: We'd still be living in caves.
(
3 CHAIRMAN APOSTOLAKIS: That's right.
4 DR. CARROLL: With our ancestors. And 5 somebody said, let's invent fire.
1 1
6 MEMBER SEALE: You know, regulatory l
7 significance at this state and the way you defined it is ;
1 8 an act of faith. I haven't seen an awful lot of 9 regulations change yet because of our so-called risk 10 informed regulations. We believe they will; we think they 11 will.
12 We've seen some useful and helpful initiatives 13 among the utilities themselves on things like in-service (G
s/
4 14 inspection and other things like that, where they're j 15 presumably going to cone forward with the recommendations.
1 16 But I haven't seen any regulatory significance yet.
17 DR. COFFMAN: Well, it's not always just 18 regulations, but it can be to facilitate the regulatory 19 process, perhaps creating a computer-aided software 20 engineering tool, or --
21 MEMBER SEALE: But ultimately, the idea that 22 you might change some regulatory requirements was also a 23 valid potential there.
24 DR. COFFMAN: Yes, but --
()
(j 25 DR. CARROLL: Anyway, you're not using any of NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N W.
(202) 234-4433 WASHINGTON, D C. 20005-3701 (202) 234-4433
i 197 1 these this year.
,x 2 DR. COFFMAN: The best I could do is tell you e i
\ / i 3 what we did and the budget allocations are always a 4 subjective -- it boiled down to a subjective judgment.
5 There's an attempt though, to be systematic in building up 6 the information on the topic to be decided upon, or to 7 have its budget allocated in a way that I think seems -- I l
8 it's systematic. And these are kind of common to the 9 process that we go through every year.
10 CHAIRMAN APOSTOLAKIS: So according to this 11 now, is it worthwhile for this agency to launch into the ;
12 new thermal hydraulic code development, or to develop more 13 -- to spend more resources on human error?
/ \
/ 14 MEMBER KRESS: Are you asking us or them?
15 CHAIRMAN APOSTOLAKIS: It's a rhetorical 16 question. ,
1 1
17 MEMBER KRESS: We've already established a ;
i 18 position on that in a nice letter on the subject.
1 19 CHAIRMAN APOSTOLAKIS: But that letter I 20 addressed a different question. It did not address the 21 relative merits of things. It just said it would be a 22 good idea to have a new set of thermal hydraulic codes.
23 But the question of whether it's a better idea than doing 24 something else was never raised.
(\
( ) 25 DR. CARROLL: Well, I wouldn't trust this NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE , N W.
(202) 234-4433 WASHINGTON, D C 20005-3701 (202) 234-4433 l
198 1 group because we don't have up there an advocate or anyone q 2 --
3 CHAIRMAN APOSTOLAKIS: That's true.
4 DR. CARROLL: In my experience in allocating 5 resources, oftentimes, he who speaks loudest -- persuasive 6 but not necessarily the merits of his case.
7 CHAIRMAN APOSTOLAKIS: Well, I mean, you can 8 also make the argument that if you look at major incidents 9 over the last 15, 20 years, thermal hydraulic ability had 10 nothing to do with these things. The work that these guys 11 are doing or the subject matter these guys are studying 12 was always there. All the major nuclear accidents have 13 involved humans.
A 14 DR. CARROLL: Or parts of organizations.
15 CHAIRMAN APOSTOLAKIS: Never codes.
16 MEMBER MILLER: Codes are developed by humans, 17 though.
18 CHAIRMAN APOSTOLAKIS: Ah, well I think 19 there's a lot to be said about inertia, and engineers l
20 liking engineering problems.
21 MEMBER MILLER: Codes are fun.
22 CHAIRMAN APOSTOLAKIS: They're never going to 23 spend millions of dollars on human error, but soft 24 science. Anything else?
25 DR. CARROLL: And besides, us engineers know NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N W !
(202) 234-4433 WASHINGTON, D C. 20005-3701 (202) 234-4433 !
l 1
199 1 all there is to know, i l
2 CHAIRMAN APOSTOLAKIS: That's right.
(3 U 3 MEMBER SEALE: Well, are we going to have i
1 4 anything else -- are we going to hear anything else from 5 these people? l 6 CHAIRMAN APOSTOLAKIS: Well, is there a time 7 the full Committee meeting -- first of all, do we have to j 8 write a letter this time? We do? Why? Has anybody asked 9 us?
10 DR. SINGH: I think the SRM.
11 CHAIRMAN APOSTOLAKIS: Oh , there is an SRM?
12 DR. SINGH: I'll find out.
13 CHAIRMAN APOSTOLAKIS: I'm not sure there's an l l 14 SRM.
15 DR. SINGH: You don't think so? I thought --
16 CHAIRMAN APOSTOLAKIS: Can Noel come here for 17 a second? He probably knows. So we have -- oh, no? Are i
l 18 we responding to an SRM here, or we're writing a letter 19 because we like to write letters?
20 DR. DUDLEY: No, the reason for writing the 21 letter is that the last time the Committee reviewed and j
l 22 commented on the human factors program I believe, was 23 three years ago.
24 CHAIRMAN APOSTOLAKIS: I'd like to get a copy
(-
V 25 of that letter. It's in the package you gave me? Okay.
NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N W.
(202) 234-4433 WASHINGTON. D C. 20005-3701 (202) 234-4433
200 1 DR. CARROLL: But we did at the time -- was it
,s 2 February this year -- version 1 of this was presented to (x
3 us? We said we'd probably comment when it was --
4 DR. DUDLEY: The other factor -- that's 5 correct. There was a commitment there that additional 6 subcommittee meetings would be held so the Committee would 7 have more information on which to base recommendations.
8 CHAIRMAN APOSTOLAKIS: And this letter will be 9 addressed to the chairman?
10 DR. DUDLEY: That's correct. The second 11 reason for writing a letter is that if there any comments 12 or views in terms of the research program, it would help 13 if they were made before January in order to incorporate
/^h (s" )
14 your recommendations in the letter to Congress.
15 CHAIRMAN APOSTOLAKIS: All right. So we have 16 to write a letter by Saturday? ;
17 DR. DUDLEY: That's correct.
18 CHAIRMAN APOSTOLAKIS: And there is time l l
19 scheduled during the full committee meeting for these 20 gentlemen to come back?
21 DR. DUDLEY: That's correct. One thing that 22 needs to be done is to provide them some type of guidance 23 of what you would lika to hear during that meeting.
24 CHAIRMAN APOSTOLAKIS: Well, it seems to me O
(_j 25 that we can select maybe one or two or three or groups of NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N W.
(202) 234-4433 WASHINGTON. D C. 20005-3701 (202) 234-4433
201 1 these questions and have you address them at the full
- 2 committee meeting, especially since Dr. Powers is not here
~
3 today.
4 The issue of models and the appropriate use of 5 models by this agency -- human error models -- I think 6 certainly deserves attention, and I will be right up front 7 with you. I will not -- I will make sure that whatever ,
i 8 letter we write will say something about it. I 9 I don't like the fact that everybody else in 10 the world seems to be using Rasmussen's and Reason's and 11 those guy's models, except the Americans. I just don't; 12 I'm sorry. And now the PRA branch has finally started 13 using them in developing the human reliability model, but O #
\~/ 14 the human factors activities seem to rely on traditional l
15 judgment. So that's something that you may want to j I
16 address whenever this -- Wednesday or Thursday, Friday?
17 DR. DUDLEY: Thursday.
18 CHAIRMAN APOSTOLAKIS: Thursday. Now, I'm 19 sure Powers would like to know more about some of the 20 later questions here. I'm not sure it's worth addressing 21 each one again.
22 MEMBER MILLER: No, we don't have time. ]
23 CHAIRMAN APOSTOLAKIS: So maybe what you can 24 do is, group them as you see fit and address these 73 25 questions. Because these are really the concerns of the
()
NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N W.
(202) 234-4433 WASHINGTON, D C. 20005-3701 (202) 234-4433 j
I 202 1 committee. I don't think we --
2 MEMBER MILLER: Can we highlight any ourselves t' m\
3 now, based on what we hear today?
4 CHAIRMAN APOSTOLAKIS: I told you which one is 5 mine -- one, plus the note. And two, and the rest I leave 6 it up to you. Okay, now --
7 DR. CARROLL: Seven and eleven.
8 CHAIRMAN APOSTOLAKIS: Seven and eleven, yes.
9 MEMBER SEALE: Well, yes, but -- I don't think 10 we need to beat on seven too much.
11 CHAIRMAN APOSTOLAKIS: Yes, I mean, we're 12 satisfied now. As long as you show up. l 13 DR. CARROLL: Why doesn't George summarize j q <
\
l
's/ 14 what we learned and say, does that answer your question?
15 DR. SERIG: I'11 show up anyway. )
16 CHAIRMAN APOSTOLAKIS: No, but they will be l
17 here anyway? Yes.
18 DR. THOMAS: Mr. Chairman?
19 CHAIRMAN APOSTOLAKIS: Yes?
20 DR. THOMAS: Might I offer that the last time I
21 I addressed the full committee was in a capacity of 22 chairman of the coordinating committee. The coordinating 23 committee has now been abolished. If you like, I would be 24 happy to start off with a five to ten minute, how we got O
'Q 25 from where we were then to where we are now.
NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N W.
(202) 234-4433 WASHINGTON, D C. 20005-3701 (202) 234-4433
203 1 CHAIRMAN APOSTOLAKIS: Sure. So all five of 2 you will be here on Thursday? Is that the plan? Okay.
!7- T U 3 DR. THOMAS: We would certainly have 4 representatives from each of the offices.
5 CHAIRMAN APOUTOLAKIS: I don't know about this 6 effective and adequate and all that. And fitness for ;
7 duty. Allocation -- come on now. This is such a high-l 8 level decision. The simulator fidelity though, perhaps --
9 MEMBER KRESS: I think that's a reasonable 10 one to have on the list. That one was particularly --
11 CHAIRMAN APOSTOLAKIS: And also number 3, how
]
12 priorities are set.
13 MEMBER KRESS: Well, that's so easily
[ ) i
\/ 14 answered, though.
15 CHAIRMAN APOSTOLAKIS: I think that one --
16 well, I mean, yes, fine. They can answer in 30 seconds.
17 I think we've beaten eight to death. We've talked about 18 it and talked about it and always say it's important so, 19 I'm not even sure it's worth raising it on Thursday.
20 DR. CARROLL: Just put it --
l 21 CHAIRMAN APOSTOLAKIS: Speak louder, please.
l 22 MEMBER KRESS: He said as long as you put it 23 in the letter.
24 DR. CARROLL: A strong recommendation.
tO
( ,) 25 CRAIRMAN APOSTOLAKIS: I don't know about NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N W.
(202) 234-4433 WASHINGTON, D.C. 20005-3701 (202) 234 4433
j 204 l 1 that.
l g-s 2 DR. CARROLL: Staff needs to get serious about l Na' organizational matters.
l 3 l
l 4 CHAIRMAN APOSTOLAKIS: Well, there's also this 5 evaluation being done by one of the management consulting 6 firms now, right? Is that part of this?
7 DR. ROSENTHAL: Not really. The, study by 8 Anderson Consulting is really a study of the NRC, and how 9 it goes about setting priorities; what information do we 10 have and then what was the decision process and the matrix 11 that the senior managers used for picking the problem --
12 have we missed some of, you know, type-1, type-2 errors?
13 DR. CARROLL: Arthur Anderson?
[_ )
s' 14 DR. ROSENTHAL: Yes.
15 CHAIRMAN APOSTOLAKIS: By the way, the staff 16 is doing something on eight. Don't you guys have a 17 contract right now, with JBF Associates?
18 DR. COFFMAN: Yes, there is a contract.
19 CHAIRMAN APOSTOLAKIS: So the metrics are 20 being developed, and they correlate with risk performance, 21 so I don't even know why you guys didn't say that today.
22 I mean, the question is, how does the staff plan to 23 respond to the ACR's advice concerning developing metrics 24 for organizations that correlate with risk or performance?
()N
(_ 25 Isn't that what the project's doing?
NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N W.
l l
(202) 234-4433 WASHINGTON. D.C. 20005-3701 (202) 234-4433
I 205 l
1 DR. COFFMAN: That's the objective of the ,
I g 2 project.
(" ) I 3 CHAIRMAN APOSTOLAKIS: Okay. But you didn't 4 say that.
5 DR. COFFMAN: Well --
l 6 MEMBER SEALE: You don't sound particularly 7 enthusiastic about the results.
l 8 CHAIRMAN APOSTOLAKIS: Well, this is not for 9 this committee to decide that right now.
10 MEMBER SEALE: I'm talking about him.
11 MEMBER KRESS: That's not --
12 MEMBER SEALE: I understand.
13 CHAIRMAN APOSTOLAKIS: That's not appropriate.
en 14 Is there anything else that we would like --
15 MEMBER MILLER: Is there anything under number 16 10 that's worth talking about? Goals and deficiencies?
17 CHAIRMAN APOSTOLAKIS: Frankly, I don't think 18 so, unless --
19 MEMBER MILLER: I'm kind of asking for the 20 members who are not here.
21 CHAIRMAN APOSTOLAKIS: I don't even know who 22 raised that.
23 MEMBER KRESS: I don't either, but --
l l 24 CHAIRMAN APOSTOLAKIS: Well, I think you
.C'i Q 25 should be prepared to answer anyone, because God knows NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N W (202) 234-4433 WASHINGTON, D C 20005-3701 (202) 234-4433
206 1 what our colleagues will come up with, but in terms of 7
2 prepare the comments, I don't think we should go over each NI 3 one.
4 MEMBER SEALE: Comments from the far left or 5 the far right are pretty divergent.
6 CHAIRMAN APOSTOLAKIS: Yes.
7 DR. CARROLL: I think maybe Jack's group's 8 presentation on the -- studies that --
9 DR. ROSENTHAL: Just a few minutes?
10 DR. CARROLL: -- might be useful, because I 11 don't know that a lot of people are aware that --
12 CHAIRMAN APOSTOLAKIS: You mean AEOD?
13 DR. CARROLL: Yes.
C'\
b 14 CHAIRMAN APOSTOLAKIS: Yes, oh, definitely. l l
15 DR. CARROLL: Or better yet, bring copies and 16 hand them out.
17 CHAIRMAN APOSTOLAKIS: So people can read them-18 and have comments for the letter.
19 DR. SINGH: Jack, why don't you give me a copy 20 and I'll get the copies done?
21 DR. ROSENTHAL: Of the actual reports? You 22 want the ones --
23 CHAIRMAN APOSTOLAKIS: I want the report, but 24 not necessarily on Thursday. I'd like to look at the O
25 reports, yes.
NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N W.
(202) 234-4433 WASHINGTON D C. 20005-3701 (202) 2344433
207 1 DR. CARROLL: It isn't that big.
2 CHAIRMAN APOSTOLAKIS: But I think that more
,- 3t
^
3 from the mental --
4 DR. ROSENTHAL: I'd be glad to do it. A 5 couple of inches of paper. Because --
6 CHAIRMAN APOSTOLAKIS: I think the fundamental 7 question -- that I think we owe it to you to tell you 8 right now -- the fundamental question still is whether 9 this is a plan. Okay?
10 And not a compendium of various activities 11 that people like to -- just happen to like. So is this 12 really a plan or do we need to think about developing a 13 plan? Any other comments or questions from people around kj 14 the table? Members?
15 MEMBER MILLER: I would comment on your 16 question. I was not here the 20th of September, 17 unfortunately, but what I've seen today and what I saw 18 last -- February it was?
19 CHAIRMAN APOSTOLAKIS: Yes.
20 MEMBER MILLER: I don't see a plan yet. Maybe 21 a plan was --
22 CHAIRMAN APOSTOLAKIS: I think a lot of --
23 MEMBER MILLER: Maybe that plan was clear in 24 September and I missed it.
(%) 25 CHAIRMAN APOSTOLAKIS: No. That's why we're NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N W.
(202) 234-4433 WASHINGTON, D C. 20005-3701 (202) 234-4433
208 1
l 1 havii this --
s 2 MEMBER MILLER: Even the questions don't lead I-s4
'w/ me to believe that there's a plan here.
3 4 CHAIRMAN APOSTOLAKIS: In my mind the 5 fundamental question is, where does this agency want to be 6 five years from now, ten years from now, with respect to 7 human factors? What tools do we have now, what do we need 8 to do to get there? And that fundamental question is not !
9 answered by this plan.
10 MEMBER SEALE: Yes, it is.
11 MEMBER MILLER: Where?
l 12 MEMBER SEALE: It's wherever the state-of-the-13 art in other industries leads us.
C\
k-) 14 CHAIRMAN APOSTOLAKIS: Well, that's one 15 answer.
16 DR. THOMAS: Could I quickly go over and 17 highlight the ones that you do want to hear, just to make 18 sure we're complete?
19 CRAIRMAN APOSTOLAKIS: Sure.
20 DR. THOMAS: I have one, two, three --
21 CHAIRMAN APOSTOLAKIS: The question -- the 22 fundamental question of maybe it's very clear is, is this l
l 23 a plan? Where do we want to be five years from now, ten 24 years from now? And what do we need to do to get there?
(3
(_,) 25 Now, if that takes a lot of new thinking on your part, I NEAL R. GROSS I COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N W (202) 234-4433 WASHINGTON, D C. 20005-3701 (202) 234-4433
209 1 just say, you know. we have already done it.
,r 2 MEMBER MILLER. %11, if the plan is to follow
\q v) 3 the industry and the state-of-the-art, then maybe that 4 should be stated.
5 CHAIRMAN APOSTOLAKIS: We don't know.
6 MEMBER MILLER: If that's their plan.
7 CHAIRMAN APOSTOLAKIS: And whose state-of-the-8 art? We're getting into details now. The second is, the 9 road map. I mean, that's related. Question 1 plus the l 1
10 note, as you correctly said, the note really goes with one 11 -- the three figures. l l
12 Then I thought we agreed that we would address 13 2 and 3 as one, which is really setting priorities, risk
/ \
\ ) \
's 14 worth and all that. I thought we were going to discuss 11 i l
1 15 --
l l
i 16 DR. SINGH: Seven, also.
17 CHAIRMAN APOSTOLAKIS: Seven, but not with the 18 same extent as today.
19 DR. CARROLL: Well, just give Dana a chance to 20 make sure the question is understood and --
21 CHAIRMAN APOSTOLAKIS: But again, not the nine 22 slides that you made.
23 MEMBER MILLER: No, just make certain that 24 somebody's here.
25 CHAIRMAN APOSTOLAKIS: Eleven, and 16. That NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N W.
(202) 234-4433 WASHINGTON. D C. 20005-3701 (202) 234-4433
i 210 i 1 should keep us busy. It's only an hour I believe, isn't
- 2 it?
3 DR. DUDLEY: An hour-and-a-half.
4 CHAIRMAN APOSTOLAKIS: An hour-and-a-half. So 5 any questions from anybody?
6 DR. CARROLL: Why are we ending five minutes 7 early?
8 CHAIRMAN APOSTOLAKIS: Because we were very l 9 efficient. Thank you very much. l l
10 (Whereupon, the Human Factors Subcommittee 11 meeting was concluded at 2:58 p.m.) I 12 13 14 15 16 1
17 l
18 l 19 20 21 22 l 23 i
24 m
25 l NEAL R. GROSS COURT REPORTERS AND TRANSCRIBERS 1323 RHODE ISLAND AVE., N W.
(202) 234-4433 WASHINGTON, D C. 20005-3701 (202) 234-4433 l
O CERTIFIC&TE This is to certify that the attached proceedings before the United States Nuclear Regulatory Commission in the matter of:
Name of Proceeding: ACRS SUBCOMMITTEE ON HUMAN FACTORS Docket Number: N/A Place of Proceeding: ROCKVIILE, MARYLAND l were held as herein appears, and that this is the original transcript thereof for the file of the United States Nuclear Regulatory Commission taken by me and, thereafter reduced to O 4 typewriting by me or under the direction of the court i reporting company, and that the transcript is a true and accurate record of the foregoing proceedings.
)
&& P>um CORBETT RINER' official Reporter Neal R. Gross and Co., Inc.
O
I J
O INTRODUCTORY STATEMENT BY THE CHAIRMAN OF THE l
i HUMAN FACTORS SUBCOMMITTEE i 11545 ROCKVILLE PIKE, ROOM T-2B3 l ROCKVILLE, MARYLAND '
DECEMBER 3, 1996 The meeting will now come to order. This is a meeting of the ACRS Subcommittee on Human Factors.
I am George Apostolakis, Chairman of the Subcommittee. .
The ACRS Members in attendance are:
Robert Seale and Don Miller. We also have in attendance Mr.
Charles Mayo, Chairman of the NSRRC subcommittee on advanced I&C and human factors.
ACRS Consultant in attendance is James Carroll.
The purpose of this meeting is to hold discussions with I representatives of the NRC staff concerning questions raised during the September 20, 1996. Human Factors Subcommittee meeting, and the o activities of the Office for Analysis and Evaluation of Operational Data and the Office of Nuclear Material Safety and Safeguards in the human factors area. The Subcommittee will gather information, analyze relevant issues and f acts, and formulate proposed positions and actions as appropriate, for deliberation by the full Committee.
Amarjit "Jit" Singh is the Cognizant ACRS Staff Engineer for this meeting.
The rules for participation in today's meeting have been announced as part of the_ notice of this meeting previously published in the Federal Register on November 19, 1996.
A transcript of the meeting is being kept and will be made available as stated in the Federal Register Notice. It is requested that the speakers first identify themselves and speak with sufficient clarity and volume so that they can be readily, heard.
We have received no written comments or requests for time to make oral statements from members of the public.
(Chairman's Comments follow:)
1 Based on the staff presentations at the September 20, 1996 O Human Factors Subcommittee meeting, and at today's meeting, the Subcommi t tee should determine wha t issues should be
, - . _ . . . , _ _ _ _ ~ _ _ _ _ _ _ _ _
brought forward to the December 5 ACRS meeting. The full
(~) Committee is expected to review and comment on these issues.
G Comments on the Human Factors Research Program need to be l
l provided prior to January 1, 1997, to support the ACRS letter {
co Congress. Time has been scheduled at the end of the meeting for discussing our thoughts and for providing guidance to the staff on information to be included in its full Committee presentation.
Dr. Seale will chair those portions of the meeting related to work done by INEL and to Organizational Factors, due to concerns related to conflict-of-interest.
We will proceed with the meeting and I call upon Dr. Cecil Thomas Chief of the Human Factors Assessment Branch, NRR, to begin.
l l
l t
l I
4 O ,
l
O O O ,e ;
I f
W
%<g%
- 4 6 O o n
i J/
- 4***
HUMAN. PERFORMANCE EVENT DATABASE AEOD Staff Presentation to the ACRS Subcommittee on Human Factors December 3,1996 Eugene A. Trager 415-6350 '
Reactor Analysis Branch Safety Programs Division Of Fice for Analysis and Evaluation of Operational Data 1
I -
l O O O ,
i DEVELOPMENT OF THE HUMAN FERFORMANCE EVENT DATABASE (HPED) it e 1980's: AEOD began studies of human performance during events
- e 1990
- Began multidisciplinary team studies of human performance i
INEL provided technical assistance in studying the human factors On-site studies performed in accordance with a protocol Twenty event studies completed from 1990 to 1993 e 1992: Human performance database initiated I
Working group representatives from AEOD, NRR, and RES e 1993: Specified database content and authorized INEL to construct database :
- 1995: Delayed work because of budget cuts ;
- 1996: Resumed work and expanded scope 2 l i
__ _ _ _ _ . _ . . _ _ . _ .. _ _ _ _ _ _ _ _ _ _ __. _ _ _ . .. _..._ _ _ _.. .____ _ _ _ _ ~_ .__ ..
ORIGIN OF CODED INFORMATION IN HPED AITs/IITs HPS STis LERs Total 1990 9/1 7 -
14 31 1991 15/1 6 -
16 38 1992 11 3 -
11 25 1993 17 4 2 20 43 1994 3 0 1 3 7 1995 2 0 4 1 7 1996 5 0 0 4 9 62/2 20 7 69 160 AIT - Augmented Inspection Team report HPS - Human Performance Study report IIT - Incident Investigation Team report STI - Special Team inspection report 3
._ _. _ _ _ _ _ ._.__ ____. .i
O O O .
HUMAN PERFORMANCE EVENT INFORMATION i
Contains information on the event and on human performance factors / issues that influenced performance by crews and individuals e Human Factors Engineering / Human-System Interface e Work Environment e System Design aiid Configuration e Procedures ,
e Training e Communication and Coordinatica e Supervision e Management and Organization e Fatigue / Stress / Workload e Individual Personnel issues (e.g., situation awareness) 4
HUMAN PERFORMANCE EVENT INFORMATION (CONT.)
Database Structure has 54 Fields to store human performance and other general event information, including,
- Event Summary and Classification e Time and Date e Operating Mode and Power Level l
e Preceding or Subsequent Subevents e Numbers, Titles, and Types of Personnel e Whether Licensed and Number of Years e Personnel Training i e Shift Rotation and Duration e Activity Type and Time on Shift l
5 1
O O O.
INEL HUMAN PERFORMANCE DATA ENTRY PROCESS I
e Data Sources
- Coding Sheets e Coding " Boundaries"
- Coding Teams e Coding Process Independent Data Coding Joint Data Entry Procedure Third Party Independent Review 6
O O O .
DATABASE QUERIES DEVELOPED e What time of day did the events occur?
- Create a matrix of the number of times that the contributing factors were coded for AIT and HPS events.
i
- Are there differences in the kinds of contributing factors that were identified in different reactor modes i of operation?
e For AIT events that occurred during power operations, what was the relative frequency of the contributing factors?
7
o a, o i
i Table 1 D.._i_st_nb__ut.io,ns of event times per t_wo.__ hour peno_d_for _a sample _of _ev.e_n_.t_s_and AITs 4 i
~ ~ ~ ' ' ~ ~ ~ ~ ~ '
2-Hour Time Period l i
(2_4-hr cloc._k) i_ _
! i~
~ ~
f4 -f~
>0200 i
>0400
...2
-_. 3_ _
2
-. __ _ _.__-4 I
. _ _. J :
t l
}________.. _p_._. i
~~
~ ~~ _u___.. f l
~
>0800 10 5
. _a _ _ . . _ o ! i, ,
>1000 6 6 J l i -
i !
>1200 ~
4 3 1 i i~~ ~ ~ ~I i !
N 06 6 4 !~ i~
44 2
>1600 4
>1800 3 2 ~i~
i
__4'_______ f _ ._ __
f
>2200 6 3
. - - _ _ + . .. ._
_ - . _ _.i
_t J1.._ Z 111 _. _2__ _ _.__f E i~ ' t _ _ -t Distribution of Event Times Per 2-Hour Time Period 10 -
9 !
8 7 i
$ 6 -
o 5 l I
j 4 _ , _
g 3 _
i' z
1 '
F- t
- f 0- I ' ' ' ' ' + - - ^
-*- -~+- + H I
>0000 >0200 >0400 >060 >0000 >1000 >1200 >1400 >1600 >1800 >2000 >2200 f.ms_ l 2-Hr Time Penod l
. . Table 2 Tab!e 2, . Contributing Factors for a S_arnple of AIT and HPS Reports
~ ~~
~~
.]_.'_i - ]
~
Con}tnbutin[ Factors
. ...._ .. .. AIT (of 40 Total) HDS (of 20 Total)J % of AITs % of HPSs i Ma.nagement_& Organization 33 .13 _82.5 _65 Procedures _ _ _ .._ _ ,_26_ __ ___ _ 1_9_ 65 95 l System Design _&_C_onfiguration 23 __ _ __ 9_ _ .
,57.5 _45 Supervision / Command & Control .17._.,_ _ . _ 16__ _ _ _ 42.5 80 Human Factors 16 14 40 170
, Training 16 16 40 80 Individual (Per_ son.n. el) Fa_cto_rs ._ _ :15'13
_..__ ' ,37.5 65 j Fatigue / Stress / Workload 4 9 :10 '45 Work Environment 3 2 7.5 11 0
_ . _ _ _ _ _ _ _ . . _ . I Contributing Factors for AIT and HPS Event Studies i
Work Environment h k i
d Fatigue /StressN/orkload 6 Y l t
Communcat ons
? l l Indmdual(Personnel)
Factors y f l Training 4 r I
l j:: Human Factors l
f u y l l~
SupervissorVCommand &
I Control System Design &
Conriguraten l h Procedures l
Management &
Organization l
~E % of HPSs j o 10 20 30 40 50 60 70 80 90 100 l E % of AITs Percent (%) a Contributor
e5 '
Table 5. Contnbuting Factors in Different Modes o_f_ Operation for a Sample of AITs (40) ev,e_n. ts_ _ _ - .
3
~ ~ ~ ' ~ ~
~ ~ ~ ~
~ ~ ~ ~
Reactor Status Code
~ " ~
'i ~ ~ ~ '~
~ ~ ~
All AITs ~~~ ~
E] power operations steady-state . , _ _1'3__ _ _ _
5 _ ]~ ~ 18~ ~ ~ ~ i ' -f,
__._. 6 H refueling _ ._ . _ _ _ _ _ 4 10 [
{
l cold shutdown 4 2 6 i i F}~Mr operations $ioad change [_[)) ~1 [ ~~l. [ _[2] ] [ ] ~~
~ ~
1 C_ r_outine start-up _ _
0 2 2 K hot shutdown 1 0 ~ ~~
1 l ho_tjtandby___________
1 26 0
14 ---
40 1- ~ ~ ~ .._]_] -[_ ]_. {_] ;
,. -rw.-- -...u.-------.---- . _,._,,_ .. ,__ _._m__.
_ _ _ _ , , _ . _ . ~ ~ . - - '~' ~ '
~ ~**~ ' ' ' ~~* * *
-.- _'~ ~ '
- _ ___ _ ~~""_'-__~"'~'
. _ , __ 7.__. . . . _ _ . _. r
- ., ~ - - . - -- . - ---- - - - - - - - - -- - - - - . . - - - - - - - - - . - - - . . . - -
__ -..- . -~. ..- . . -
h '
- au-h- =.i.-
.mai .,_4. =
_. . ... . ~_ _---._ - ._ _ ..- _ _ _ , ,._. . . _ _ _ _
_. .. _ . _._.~._ _._ . _ _ _ _ _. _ . _ _ . . _ _ _ - .
%N -hsm--.rwe g.-.m%...e.m..e.-we.-
_-h m. .i.,.m.-As.i.e ps-
., em.m.-M>
e.-?v- -+4.wa.m.-a.m.*a.m 'd-.-4m+eam.-n.-.-4 .. e+-mea
.---wio%%,,._---., -
--h.-
__ ___ ____ . _ __ .____.. ---- - - - - -- - - - - - - - - - - - ---E - ------ -
f
}---
Page 1
-_.-.n.-_m
- M-er-M se e J----- --..M= -"NemW Gh. A A M .W h aAh he m m.4w M M M. 5 M an am l*
I O 5 I
s.
s7 h
I l
f I e 8 E i
5 g i a n g I E 1
1 m
. ?ySf e N
~
p 3
a ac '
gi E t e
iM <
3 O
I f
- o 34 r~
w J< .,
Z
> '?
$$ 8
- is
- I
! I
.I J
1 i l lO - - ' - - - - - - . . - ~ , _ - _ _ _ , _ _ _ _
-*"'*Nw--*- _
J
}
O O Th.de 5d O.
Table 5d % of a Sample. of AITs (40) for which certa _in facto _r_s were fo__un.d_ t_o be c___ontnbutors (Sorted by %PO Events),
7- ,
~
~
Contnbuting Factors I' I Power Operations Cold SD'and Refuel I
~
^
_ , ~(PO) Fraction '~~(CS&R)'F action ~ I % PO Events ,
% CS&R Events Management & Organization '
15/20 -~~ ~
14/16 ! 75% : 88 %
System D'esign & C_o_nfigu_ ratio _n _ . . 15_/20 _ - _ _ -
.~6/16~ __
_4 I 75 % ,I 38%
~ ~ ~ ~ ~
~
ind'ividuaf(fe'~rsonnel) Factors ~~~8 /20 ~ ~]4/16 l~ [ _ [ 40% __ { 25%
Supervision /Co_m_ mand & Control 8/20 8/16 40%_ l 50 %
Training _ 8/20 4/
. _ _ __ _ 1_6_ _ _ _0%_. 4 25 %
3/20 8/16
.f _ .
50 %
Communications (__ __ ___15%_ _
.L Work Environment 3/20 0/16 I_ 15% ! 0%
7~ !'
~
Fatigue / Stress / Workload 0/20 2/16 ~ 0% ~ 13% i f
Percent AIT Events that Resulted from Certain Contributing Factors b
l Fatgue/ Stress / Workload ME Work Ermronrnent mm Cornmunicatens N
= I ,
f Training M m !
a, SupennservComrnand & Control M j indmdual(Personnel) Factors M m ;
f Human Factors M ME Procedures System Design & Conriguraten Mm Management & Organizaten p . _ . _
' ' ,E % CS&R Events i
0% 10% 20 % 30 % 40 % 50% 60% 70 % 80% 90% !
5 % PO Events
- _~ >
p I
L t
O O O-PLANS TO DEVELOP THE HPED e Distribute Draft HPED to NRR and RES e Update the database with event information from AIT and other reports '
- Perform studies using HPED data e Continue to test and develop the database and consider including data from other sources, for example, Regional inspections performed in accordance with the Human '
Performance Investigation Protocol (HPIP)
Accident Sequence Precursor (ASP) events with high CCDPs, and Foreign event reports (that have been cleared for release) e Examine risk significance of events and develop information for use in studies of human reliability / risk assessment e Distribute the database at the NEA PWG-1 Specialists Workshop t October 1997 in Chattanooga, Tennessee.
- Estimate costs to maintain and access database 8
o o o:
RECENT AE00 STUDIES INVOLVING HUMAN FACTORS ISSUES
. Oconee Electrical System Design and Operation (ongoing, projected 3/97)
. Assessment of Spent Fuel Cooling (S96-01)
. Review of the National Transprtation Safety Board's Safety Study NTSB/SS-94/01, A Review of F'ightcrew-Involved. Major Accidents of U.S. Carriers,1978 Through 1990 (T95-03)
. Reactor Coolant System Blowdown at Wolf Creek on September 17, 1994 (S95-01)
. Operating Events With Inappropriate Bypass or Defeat of Engineering Safety Features (E95-01)
1 i
s Scope & ANSI 3.5 Part 55 ES-301,-3 ES-604 Standard 10 CFR 55 -
Comments RG 1 149 Scenano Attchmt 3
[ Design Summary Checklist 8.1.c l
l
- ES-604-1 Verification Scenario
' =
NRC NUREG Checklist o Form 474 1262 Simulator Validation Acceptance 4 Periodic :
NRC Testing h NUREG 1258 Simulator = '
Evaluation Simulator Trainina Fidelity m
J. .esting. & E,xam :
i Fee &ack u Plant SAT Ns ._ hjnynt y
Difficulty r grams ,
. Exam .
Training 1
=
- . g'
- O BRIEFING ON
HUMAN FACTORS 4
- o
- +
i i
1 1
i I
Dennis I. Serig, Ph.D.
! Senior Human Factors Analyst NMSS/IMNS/IMOB
- O December 3. 1996 1 I
O HUMAN FACTORS IN NMSS o Program initiated in May 1988 o Limited resources O o aroad responsidiiities o Need to prioritize activities o Need to leverage resources where possible O December 3. 1996 2
)
O HUMAN FACTORS IN NMSS i
o Areas of responsibility o Medical use o Industrial use o Academic use ,
O ,
o Fuel cycle o Waste management !
o Transportation o Safeguards O December 3. 1996 3 i
lO HUMAN FACTORS IN NMSS o Priorities o Teletherapy misadministrations o Remote afterloading brachytherapy misadministrations O o Nuclear pharmacies and nuclear medicine misadministrations o Industrial radiography personnel l overexposures o Safeguards personnel training I
i O December 3. 1996 4 4
l .
L O HUMAN FACTORS IN NMSS o Starting point o Little or no previous human factors work associated with systems used by NMSS licensees o Systems used by NMSS licensees not ,
adequately modeled in terms of functions nd tasks O
o No immediately useful base of data from which to begin the analyses necessary to :
identify human factors problems in i systems used by NMSS licensees o Lack of knowledge and appreciation of the potential benefits of human factors by NMSS licensees o December 3. 1996 5
O HUMAN FACTORS IN NMSS o Approach l o Participate in the development and ongoing i revision of Humanfactors engineering guidelines and preferred practices for the
\
design of medical devices o Initiate human factors evaluation of teletherapy o Initiate human factors evaluation of remote afterloading brachytherapy O o Encourage a large nuclear pharmacy chain to initiate its own human factors program
~
o Initiace human factors evaluation of industrial radiography o Address need for performance based training of safeguards personnel o Participate in day-to-day review of events and follow-up for events that appear to involve i
significant human factors problems December 3. 1996 6
I i
f 1
O HUMAN FACTORS IN NMSS ;
o Products / achievements )
1 i o AAMI/ ANSI standard on Human factors
- engineering guidelines andpreferredpractices for the design of medical devices o Reduction in misadministrations attributed to nuclear pharmacies o Flexibility in licensees training plans to assure
, security personnel have necessary knowledge, skills, and abilities
- O i o Human factors evaluations of teletherapy and j remote afterloading brachytherapy
! o Database of nuclear medicine ,
- misadministrations (1989-1990) j o Database of all misadministrations (1992-present) o Findings incorporated into books, inspection reports, and newsletters o Training of inspectors O December 3. 1996 7
_=
I 4
!O HUMAN FACTORS IN NMSS ,
l o Ongoing activities )
o Work with AAMI on revision of Human factors engineering guidelines and preferredpracticesfor the design of medical devices o Work with RES on follow-on activities related to human factors evaluations o Briefm' gs to the medical community, ACMUI, other regulators, and the human O factors community on the method and findings of the human factors evaluations of teletherapy and remote brachytherapy o Briefings to the medical community, other regulators, and the human factors community on the use of event databases for identifying human factors problems and on the results of that use o Daily review of events o Participation in event follow-up O December 3, 1996 8
t
'O HUMAN FACTORS IN NMSS o Conclusion o NMSS has an active ongoing human factors program
- o That program began from a starting point
- well behind the starting point for the human factors programs associated with
! reactors i
o The program's research and data based PProach has succeeded in identifying, O prioritizing, and evaluating alternatives for resolving the human factors problems experienced by NMSS licensees o Pending Commission direction on the future of the medical program, NMSS's human factors program is in a position to assist in revision of 10 CFR 35 o There is evidence that NMSS's efforts have resulted in an appreciation for the need of human factors in medicine O December 3. 1996 9
Q14. Staff approach to evaluating human performance modeling Select a candidate Evaluatefeasibility Evaluate promising 4 technique (s) oftechnique(s) technique (s)
' I Y
Develop /refinelinks 3 f between models and safety assessment Conduct tarsted Refine / enhance "**##
validation studies in modeling a nuclear setting technology Validate integrated humanperformance andsafety assessment Extend models to models develop regulatory I guidance Use integrated models to evaluate risk ofhuman performanceissues o
O O O
j Q.14 Status of Modeling Efforts S
Evaluate promising Select a candidate Evaluatefeasibility technique (s)
M technique (s) oftechnique(s)
MicroSaint MA&D, BNL, HRP, 1 NUREG/CR-5439 Charles River Charles River Assoc.
NUREG/CR-6465 Dynamic Howgraph ASCA 1 f 1 Y Developfrefinelinks i 1 V between models and
{ Refinelenhance safety assessment
- Conduct targeted modeling methods validation studies in technology V 4
a nuclear setting SBIR(MMI, Plan ) Validate integrated
{ ARL(Med.,Hdwr, ,
hemanperformance
' Alarm, HRA)
NUREG/CR-6159, Ltr.Rpt. andsafety assessment Ltr. Rpt. (Phase 1) modelf Ltr. Rpt. (Phase 1) y Useintegrated models to Extend models to evaluate risk ofhuman develop regulatory performanceissues
. guidance O O O
w xw a a ,a s _ e .k m .== = n w.a m.w-o, m.e mnon m, :.a g, . - - - _ --...--.n. . - - - . .
s . . , _ , , , _ _ . . _ _ _ ..
s
~
j .
t i
1 s
l l'
4 EEE . , . <
alf[i;M W
' isinfendj" l 43!islasadeta @ bm*cN j"q.i' 1
%sNRie -
s.
ek l.
l if i , ,
1 i . . . . . . .
j .
,J881Rilj l.m...
Procedure. e';i '
l l
10 b. -
l N
! l
( i r i
-ARL1 c
JTAEOMer? *
{ . % Medical j
a mem m i
f y W iWi J
I j e '
1
) .__ -
, - s.,
[ j United States Nuclear Regulatory Commission Q13. Staff approach to developing k
performance based fitness-for-duty criteria Maintain cognizance of state-of-the -practice
- Track government research Follow industry / commercial experience Contact vendors NUREG/CR-6470 Fitness for Duty in the Nuclear Industry:
Update of the TechnicalIssues 1996 Tests identify impairment from broad range of agents Detect impairment, but not cause Not predictive O O O
.p-m ,
( ) United States Nuclear Regulatory Commission l Q12. Standards for Computerized Job L Performance Aids HPPP ltem 1.2.11
. Fostering the development of technology is not the role of the HPPP or the NRC.
! The NRC's role is to review and evaluate industry technology and perform confirmatory research as necessary to support the review and evaluation.
G 9 9 u___ _ _ _ _ -_ _ _ __ -- _ _ _ _
] -:
s p * "*%,,,
, ) United States Nuclear Regulatory Commission Q10. What are the GAPS in NUREG-0700, Rev.1 High Priority Secondary Priority 0 Alarm systems
- O Knowledge-based O Graphical presentation Systems O Computer-based O Flat panel displays procedures ** O Display hardware O Automation interface O Soft controls **
methods O Workstation integration O interface management
- O CR layout and O Large screen displays environment O Integration of advanced O Test & maintenance of systems in conventional digital systems **
CRs*
- User need active research project ** Incorporated in Integration project O O O
-- _ - _____ -- _