ML20127E243
| ML20127E243 | |
| Person / Time | |
|---|---|
| Issue date: | 06/18/1985 |
| From: | Advisory Committee on Reactor Safeguards |
| To: | |
| References | |
| ACRS-T-1419, NUDOCS 8506240490 | |
| Download: ML20127E243 (163) | |
Text
0R\G#RL aesr-d UNITED STATES OF AMERICA
,e NUCLEAR REGULATORY COMMISSICN
%.j' In the matter of:
ADVISORY COMMITTEE ON REACTOR SAFEGUARDS Subcommittee on Maintenance Practices and Procedures Docket No.
)
Location: Washington, D. C.
Date: Tuesday, June 18, 1985 Pages: 1 - 123 3 4 k '$
M0" 3090VB "f 0m AC35 C" fine ANN RILEY & ASSOCIATES Court Reporters p.,)
i 1625 I St., N.W.
Suite 921 dgo Washington, D.C. 20006 og t
(202) 293-3950 g62 0 050618 T-1419 PDR
I .
1
\
V 1 UNITED STATES OF AMERICA 2 NUCLEAR REGULATORY COMMISSION 8
4 - - -
5 6 ADVISORY COMMITTEE ON REACTOR SAFEGUARDS 7 SU0 COMMITTEE ON MAINTENANCE PRACTICES AND PROCEDURES 8
9 - - -
10 Room 1046 11 1717 H Street, N . W e*
12 Washington, D.C.
19 Tuesday, June 18, 1985 14 The Subcommittee on Maintenance Practices and 15 Procedures of the Advisory Committee on Reactor Safeguards 16 met, pursuant to notice, at 8:30 a.m., Glenn A. Reed, Chairman 17 of the Subcommittee, presiding.
18 PRESENT:
19 G.A. REED, Chairman 20 W. KERR, Member 21 C. WYLIE, Member 22 D.A. WARD, Member 28 n
24 25
1 ALSO PRESENT:
2 H. ALDERMAN, Cognizant Staff Member l l
d 8 G. CWALINA l
4 H. BOOHER l 5 J. JANKOVICH t
6 P. MCLAUGHL1N 7 T. LEE 8 J. KOONTZ 9
10 11 .
13 14 15 16 17 18 19 20 21 22 23 24 25
I i
V 1 P ROC EED i NGS 2 MR. REED: I guess we are ready to get going. The 8 meeting will now come to order.
4 This is an open meeting of the ACRS Subcommittee on 5 Maintenance Practices and Procedures.
6 My name is Glenn Reed, and I am Chairman of the 7 Maintenance Practice and Procedures Subcommittee. The other 8 members present today are Dr. Kerr, David Ward, Charlie Wylle, 9 and the purpose of this meeting is to review the NRC 10 Maintenance and Surveillance Program Plan, MSPP, Phase I, 11 Herman Aldorman, on my right, is the assigned ACRS 3
\/ 12 Staff member for this meeting.
13 A transcript of the meeting is being kept, and it is 14 requested that each speaker first identify himself or herself 15 and speak with sufficient clarity and volume so that he or she 16 can be readily heard.
17 We have received no written comments from members of 18 the public. We have received no requests for time to make 19 oral statements from members of the public.
20 1 am new on this Maintenance Subcommittee as the 21 chairman, and I came from industry to the ACRS so I have some 22 feelings with respect to maintenance. So my comments will
) 29 likely reflect a number of years in the workplace, 24 Before we gut going, I would like to know if any 25 other members of the Subcommittee have any comments they wish
- f. l 4
q
~- 1 to make at this time.
2 MR. WARD: My comments will also reflect many years 8 in the workplace. I'm sorry. I couldn't resist.
4 MR. KERR I have never worked in my life.
5 [ Laughter]
6 MR. REED: Charlie, do you have anything?
7 MR. WYLIE: Not after that.
O MR. REED: Well, let's hope with all this background 9 that we have, that we will be contributory. We are accused 10 sometimes of not being.
11 First I would like to call on Dr. Bocher, who I-s G
j 12 going to discuss from the Staff's viewpoint the background and 19 status of the Maintenance and Surveillance program plan.
14 CSlide3 15 MR. BOOHER: I think most of you met me about a year 16 ago. We came down to give a kind of preliminary review of the 17 Maintenance and Surveillance program plan.
18 I am Hal Bocher. I work for Bill Russell, who lo a 19 key instigator of this whole program area, and Bill has kind 20 of dropped off with minor things like Three Mlle island and 21 things like that and is letting um carry the maintenance load.
22 CSlidel 1 think before I get into this background, I think
{}
'V 29 24 one of the most important things that has happened is we now 25 have a formal Maintenance and Surveillance Section formed in
{
5
/
1 the Licensing Qualifications Branch, and I think i brought 2 down just about the whole section to meet you people. So if I 3 could, I would just go through.
4 As Mr. Reed emphasizes, qualifications of personnel 5 are very important. We tried to get the best people we 6 possibly could to form this section, so if you would just 7 allow me to introduce them and give a little of their 8 background.
9 The section leader is Gregg Cwallna. He was 10 selected in January. Gregg has a nuclear engineering degree 11 from the University of Virginia. He worked for Bechtel Power 12 Corporation as a nuclear engineer before he came to NRC. He 13 has been at NRC since 1977, and he has had a wide experience 14 in the nuclear area. He has worked for the Environmental 15 Evaluation Branch, he la program Manager for Dresden 2 in the 16 Systematic Evaluation program Branch.
17 Most recently when we acquired him he was with IE 18 Headquarters, and he was doing important to safety inspections 19 and particularly looking into the maintenance area at that i 20 time. He also was the IE coordinator, and I think he was down 21 here talking last year from an IE point of view when we were 22 developing the Maintenance and Surveillance program plan.
23 The next person, who has been with us from the 24 start, is Dr. John Jankovich. John received his master of 25 science degree in 1963 in Hungary, and there they emphasize
T 6
O)
(
\/ 1 such things as maintenance in the engineering courses. He was 2 specializing in manufacturing and maintenance technology.
8 in 1970 when John was in the United States he got 4 his doctorate in mechanical engineering at the University of 5 Kentucky, and he has worked in Human Factors for the Navy, has 6 worked on automotive safety systems for the Department of 7 Transportation. He has been with the NRC since 1932 working on 8 simulators in maintenance. He has been on a number of site 9 evaluations and INDO plant evaluations, so he can give some 10 personal experiences there.
11 He is the program manager for the Japanese study 12 which has been recently completed, what we have done so far, 18 and I would say he was a key individual in developing the 14 Maintenance and Surveillance program plan.
15 Jennifer Koontz has also been with us from the 16 start in developing this program plan, Jennifer has a 17 master's degree in engineering psychology. Along with me, we 10 are the psychologist arm of this section. She has worked in 19 the Bureau of Standards. She spent three years with NUS 20 Corporation. She was at NRC both before and after NUS, so she 21 has had a lot of experience in specializing in human factors 22 engineering, personnel training qualifications.
23 She has been deeply involved in the Human Factors
(
24 study of nuclear plant operations and maintenance. She is our 25 primary in-house specialist on maintenance personnel
7
\~ / 1 selection, training and qualifications.
2 Tommy Lee, another acquisition from IE. He is our 3 resident equipment qualification specialist. He was with the 4 Division of Quality Assurance and Inspection programs. He was 5 equipment qualifications engineer in NRR before that. Tommy 6 has a master's degree in electrical engineering. He spent 7 quite a few years at TVA before he came to NRC. There he was 8 a design engineer, a pre-operational test engineer, and he was 9 a maintenance support engineer, quite a wide background, and 10 much needed. He is also a key participant in our IEEE 11 Standards Working Groups. Most recently there has been e new g 12 one formed, and we will hear about that in a little more IS detail later, but Tommy has had key inputs into that group.
14 Drew persinko is not here because he is out today 15 working on wrong unit, wrong train site visits, so we are 16 participating, our section, in all these visits which are to 17 look for short-term solutions right now. Drew has a master's 10 degree from MIT in nuclear engineering.
19 Bernie Grenier --
I don't know if Bernie is here 20 yet. Bernie is acting as our program manager on our technical 21 assistance contracts and maintenance. He has helped us with 22 our overall program and management coordination. He has a g 23 master's degree in business administration. He also has 17 b('w 24 years of emperience with the Ccemission going back to the 25 beginning. He has a lot of emperience in budget contracting
8
(
\ 1 and management information systems, and since we have so many 2 interfaces with so many other programs, we really want to keep 3 this as an integrated project, and Bernie has been very
- 4 helpful with that.
5 poter McLaughlin, who does not work in the section 6 but has been a major contributor to our management 7 organization problems, you will hear him today. He will 8 present the results of our maintenance Indicator study.
9 And finally, I don't know if he got down here --
10 it's his second day of work -- our most recent acquisition, is 11 pinto Kutename, pinte is from Helsinki, Finland. He i s on 12 loan to us for one year. He is from the Finnish Center for
)
13 Radiation Nuclear Safety, where he was a section head there, 14 and he expressed an interest in maintenance, so we have him 15 working with us and we asked him to come down today. I don't 16 know if he has been able to find his way down. He probably 17 got lost. He has had a hard time.
19 MR. WARD: It is probably a bad way to start out a 19 new employee.
20 MR. REED: le that the rundown on your staff? I'm 21 very pleased to hear that. You remind me of something. Yes, 1 22 am a great exponent of selection, and particularly in the 29 maintenance area for people in the workplace, and you have 3
24 heard me remark on one of the problems, called all-thumbs 25 mechanics with zero mechanical comprehenelon. I am interested
f l
l
[~'N i \ )
\'
1 in your people, and it sounds very good.
i 2 One thing we ought to keep in mind is we are going 8 to have to adhere to schedule today because we have a very
! 4 important session this afternoon, which is a camb i ned sess ion ,
5 and we will be addressing such things as' selection, and i 6 think it would be --
7 MR. BOOHERr That should be an interesting session, i
l 8 In fact, we will have sor.e peop lo here to be sure and hear 1
9 what the Industry is doing in that area.
10 CSilde3 i
! 11 Since John will be giving the details of the program y j 12 plan, 1 j
,ust wanted to go over quickly the objectives to kind 19 of uet the stage. The thing we have been saying all along is 14 it is not our intent to address specific technical problems on 15 how to maintain cpecific components at this point. Our 16 objective is to identify factors that contribute to
[
, 17 programmatic effects and aspects of the overall maintenance l
l
< 10 effectiveness.
19 Of course, through our site visits and other l
20 offerts, we do enpact to identify practices which do reduce 21 human error rate in performance and maintenance.
22 lJe have done a lot of talking about comparison with
(
29 the Japanese, but we know that there are excellent plants and 24 maintenance management programs in the United Gtates, ard that 25 le the thing that we have been given the task to do in this L
10 1 first phase, is to ferret those out, make sure we know of 2 them, and then make sure the industry is aware of what can be a done.
4 We also want to look at we are very concerned about 5 improving the effectiveness of maintenance programs for this 6 operability of safety systems. We think the trip rates are too 7 high in so many areas, and there have definitely been, from 8 our studies of 1988 and 1984 data, too many challenges to the 9 safety systems because of this.
10 MM. KERR: How will you know when you are getting 11 the correct number, John?
f
' 12 MR. BOOHER: I don't think we are going to be able 18 to say what the correct number is right now. Right now at 14 this point we have -- it's just that the tall is so big. John 15 has the distribution of people who are out there having 12 or 16 13 trips per year, whereas in Japan it's like .6 per year, and 17 Oermany is about one. We have plants with zero and one 10 consistently.
19 So the important thing is to try and bring those 20 talls in.
21 MR. hEHR What is your goal?
22 MM. BOOHER: Our goal right new is to find out how 23 bad the problem is because in phase I we can't say even for 24 sure that those ones out there having the trips are 25 necessarily having that big a problem. We are suspicoud, i i
$ 11
(
1 but we haven't been able to prove that yet.
2 Also we want to reduce, if possible, the 3 radiological exposure to maintenance personnel, primarily 4 because maintenance personnel are the ones getting all the 5 dosage. Over 75 percent of it occurs to them. So I think we 6 do have a responsibility in reducing exposure to those people 7 if there is any way possible to do that.
8 Now, there havu been arguments that perhaps a better g maintenance program might increase the dosage rate to some 10 people.
11 MR. KEHR: Are any of them getting more than the g j 12 maximum permissible exposure?
19 MR. BOOHER : I can't state that. I'm sure some are, 14 but i don't have data on that.
15 Gregg, do you have any data on that?
16 MR. CWALINAr I believe some of the plants --
17 separate Individuals are getting more than the maximum 19 allowable exposure. That happens occasionally. Some minor tg type accidents. They stay in there too long. But on the 20 whole, the plants will just run in enough people so they don't 21 burn out anybody.
22 MM. KEHR And what is your goal there? Meduce it 29 to what?
24 MR. NOOHEM I Again, in this phase we are not setting 25 a goal to reduce to anything. Our overall goal le to reduce
F 12 O
xl 1 it if possible, but phase i has to determine how bad things 2 are and whether or not their goal can be set. So these are, 3 again, motherhood statements to a degree. There is a lot of 4 concern that, you know, if 75 percent of the personnel are 5 getting this in maintenance, has there been appropriate 6 attention to this?
7 John, did you have a comment?
8 MR. JANKOVICH: John Jankovich of the Staff.
9 A recent publication in Nuclear News indicated, for 10 example, that in the Midwest, the north Midwest has excellent 11 maintenance program, good availability, and still very low
() 12 radiation exposure of the workers, so the two things, good 13 maintenance and the low radiation exposure, are compatible.
14 It depends on the individual program management at the site.
15 CS11de3 16 MR. BOOHER I won't go through all of these, but 17 this just gives ynu a' history -- I think in your handout you 10 have it -- of where this program plan has been. it really 10 started with a couple of workshops in which we developed some 1
20 of the things we brought down a year ago to talk about, mostly i
Cl emphasizing all the problems, things like number of abnormal l
l 22 occurrences percentage-wise s**med to be increasing, faster i
29 rata than what the total abnormal occurrences are.
{
i 24 We briefed the ACRS on the maintenance and program 25 plan and also what we were planning to do with the e
[
18
(~N
\
1 Japanese. We briefed all the standards groups --
2 MR. REEDS Dr. Bocher, I notice you have mentioned 8 now twice the Japanese study, and I know we have gotten a lot 4 of literature and I have some documents with respect to the 5 Japanese interfacing. We are not going to be able to discuss 6 that in open session, and we are not discussing it here todays 7 is that correct?
8 MR. BOOHER: That is correct.
9 MR. REED: Thank you.
10 MM. BOOHER : There is a publication, a NUREO which Il le being published right now, which is available to the 12 public, is that right, John? But there is one part that is 19 confidential.
14 MR. J ANKOV I CH : These results of the Japanese study 15 --
this is John Jankovich of the Staff. The results will be 16 published in two volumes. Volume 1 is NUMEG/CR-3999, which is 17 available to the public. That is the general summary of the 10 findings, and it discussee the Japanese operating experience 19 and management practices. Volume 2 is NUME0/CR*3GODO. That 20 containe proprietary information where we compare item by item 21 the maintenance practices of certain selected Japanese nuclear 22 plante to their sister plants in the U.S. That is a table
() 23 format of maintenance schedules and practices, where they are 24 compared item by item.
25 Both of these are in printing at the moment.
14 1
0}
V 1 MR. REED: A folIow-up questlon. Wiii there at some 2 time be a discussion between ACRS members and the Staff with S rspect to perhaps the final conclusions on the Japanese review 4 and what might be considered for adoption in the U.S., if 5 anything?
6 MR. BOOHER THere ia nothing pianned but i see no 6
7 reason why something like that couldn't be set up. The one 8 thing is that the Industry itself is going this month through l L
9 INp0 coordinatlon to Japan to do their own survey and 10 evaluatlon of the Japanese program, and we were told by 11 Mr. Denton to pretty much stop untii that comes back and we 12 compare notes. I think probably at that time it would be 13 appropelate for both us and the i ndu s t r y to talk with ACRS.
14 MR. REED: Well, I think that's a good idea, to go 15 slow and let the INpo NUMHRC or whatever it is material come 16 in. The socia 1 dI(ferences and people dI(ferences and how 17 people are led up through their careers in Japan make it very 10 complex and diifIcutt to make a trun cceparison.
19 MR. BOOHERi There ie no queetion. And aiso, i 20 think we see it as such a difforent approach to maintenance.
l I l 21 i can certainly tell you right now their approach is one of l 22 preventive maintenance as opposed to oure of more surveillance l
l l x 28 to ensure operabIIity of eyetems. They are completely
) r i
24 different approaches. We both are trying to learn frem nach !
I I
25 other, but it's unlikely we can up and adopt very many things
[
l
- 15 1 directly, and I think that probably is one of the more needed I
2 studies, which is to compara the trade-off between 8 surveillance testing and preventive maintenance because 1 4 don't think you want to go either way completely. I think I
l 5 both countries are finding that is a problem.
l 6 MR. WYLIEt Let me ask a question. la the INp0
{
! 7 NUMARC report going to be published or made available?
1 l 8 MR. 900HER: I don't know what they have in mind for l
l 9 a report. All I know is they are just going over to do their l
l 10 own survey, 11 MR. WYLIE: What about --
'l2 MR. BOOHER: See, the INp0 shouldn't be just for the 19 utilities that participate. I should think it should be made 14 public.
15 MM. WYLIE: IJell, I guess -- do you have access to l 16 it or have you got any briefing from them?
17 MR. BOOHER: Well, we have no access to anything to yet.
19 MR. WYLIE: Do you have any agreement with them to 20 share that information?
21 MR, BOOHERi I don't think so. Oregg, do you know?
22 MR. CunLINA I don't k n esw . I think it may be 23 worked out through the current HMC INDO agreement that we 24 have.
I 25 MR. BOOHER: We haven't addressed that one 1
16 1 specifically. The one we are really working on right now is to 2 get the data that INp0 and NUMARC are collecting on plants, on 3 their' maintenance indicators, which we will get into in just a 4 little bit.
5 MR. REED: You mentioned sister plants, and it was a 6 little disappointing to me to see the way sister plants were 7 selected between Japanese and U.S. plants because we should e keep in mind there are a few -- I would say probably a very 9 few -- plants in the United States that rival or do better 10 than all of the Japanese performance, and those did not appear il in any of the sister trade-offs.
12 MR, BOOHER : Yes. John, would you like to speak to 18 that as far as our selection?
14 MR. JANKOVICH The plants were selected on a number 15 of criteria. When they were, built, each design, BWR and PWR, 16 and the technical specification, these were major categories 17 by which we rcreened plants for incluelon into the study, We 10 compared one Japaneen plant versue two American plants. We at 10 that time couldn't find better criteria, and we certainly 20 didn't want to compare them on the basis of operstitig 21 emperience becauen that was a dependent variable what we 22 wanted to emplore, The others, the selection criteria served 1
29 as the independent variables,
(
24 HR. REEDt You sort of lost me on that last one.
25 You eald you certainly couldn't compara them on the baele of l
p 17 ss 1 operating experiences for instance, take the best Japanese and 2 the best U.S. Why wouldn't you do that? After all, you are 8 looking for indicators. You are looking for what is wrong or 4 how to improve. Why wouldn't you take the best Japanese and i 5 best U.S. and then maybe the worst U.S. and the worst 1
6 Japanese?
7 MR. BOOHER: That might be a good follow-up study, O but I think the main thing was to compare comparable 9 equipments and then look at the practices that are being 10 used. There is a lot more to be done.
11 MR. REED: Keep in mind, now, equipment is not The 12 leeuw, necessarily, or that shouldn't be wooded out. The 19 issue is humans in the workplace, isn't it, to find out why 14 Japanese performance is better than ours? Its administration, l
15 its emphasis on surveillance, its emphasis on overhaul, 16 maintenance and frequencies and all those things. It's the 17 whole human aspect, o t necessarily the equipment n'
IG comparativeness, 19 MR. JANKOVICHi i completely agree with you, and 20 that is what we wanted to find out. But had we selected the 21 plants with different equipment, then that would have been 22 even more variables in the comparison than at the present l
20 time.
1 ,
24 MR. BOOHER Our problem is that if the differences 25 came out, it might be argued that it was because of different
Ii 18 Os (m l 1 equipment, not because of human differences. So I think we 2 have found enough differences even so, even with --
8 MR. REED- Well, I think that the equipment thing is 4 more factual and finite and can be subtracted, but the human 5 thing is complex, and, quite frankly, I would have taken 6 comparative performance plants for the evaluation and see if 7 there isn't threads of human administrative aspects that run 8 through the comparative plants with good performance and bad 9 performance.
10 MR. BOOHER: I think that's a good suggestion. If 11 we get to do some more, we will try to do that. We will to 12 looking at, I think, some good performers in our own survey (O}
13 here.
14 I think we had better move along so we don't lose 15 time.
16 I have already introduced the people from the 17 Maintenance Section, and all these things will be covered.
10 Basically I just want to say that we have made sone pret ty 10 good progrees, particularly in getting our people experienced 20 and out in the field, and we are trying to approach this free 21 a very practical point of view and not sit back in the Ivory 22 tower and come up with things that people cannot live with.
29 John, I think we ought to move on to Dr. Jankovich, 24 who will give us the overview of the Maintenance and 25 Surveillance program plan itself.
m
(
1 [SIide]
2 MR. JANKOVICH: The objective of my presentation is 3 to give you an overview of the Maintenance and Surveillance 4 program plan. This is the guideline which we will follow in 5 the course of our action. We sent to the Subcommittee already 6 a copy of this program plan. I understand that it's a big, 7 hefty document of 30 pages single spaced. You don't have time 8 to read all the details in it, so I would like to highlight 9 the important aspects of it and show you where we are and how 10 we are to go about the specific action items in that program 11 plan.
O) v 12 CSilde3 13 So we start out as usual with the definition. We 14 have been talking about maintenance already this morning, and 15 we want to clarify that in our definition of maintenance, we 16 include not just the conventional term of preventive and 17 corrective maintenance activities which go on daily in the to plants: we want to include in the maintenance program at the 19 moment the surve i l lance ac t ivi t ies becau:se we think they are 20 closely related to maintenance, 21 Of course, we want to include interactions with 22 operatione and poet-maintenance testing, and administration
(N 23 and management of the maintenance program, personnel aelaction 24 and qualifications, procedures, documentation, even material l
I 25 control and spare parte, i_ _ _________ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
F 20
{
\/ 1 As you see by this list, we want to look at 2 maintenance in overall aspects of it as a complete, systematic 3 approach to managing maintenance at the site. We don't want to 4 select certain aspects of maintenance.
5 CSildel 6 Maybe the pictorial representation of these items 7 gives you a better idea --
8 MR, KERR Mr. Jankovich, I am convinced that 9 maintenance is extremely important. The program you are 10 describing strikes me as being the sort of program that the 11 Industry must be carrying out or ought to be carrying our in
'12 order to better organize their maintenance.
13 How, what is NRC going to do with this information?
14 la the ultimate objective that you will set up some new rulee 15 which specify maintenance, or are you trying to understand it to so that you can observe it? What are you doing to do with 17 thee information that you collect and analyze and write to reports on?
10 MR. JANKOVICH: The program plan gives clear answere 20 to it. We have been til scuss ing the subject ourselves, too.
21 We reallee completely that it le boet to handle maintenance by 22 the industry itselfa therefore, our program plan is aimed that 28 in the first year we will just look at the present statue of 24 maintenance in the industry, 25 1 could use an analogy, for emample. To take the
21
\- 1 temperature of a patient, you use a thermometer and measure if 2 somebody is healthy or not. That is what we are doing at the 3 moment, measuring the temperature of the patient, doing this 4 survey, overall investigation on how maintenance is being 5 done la this adequate, does it affect safety systems, do we 6 find any degradation of safety systems due to improper 7 maintenance?
8 Once we Identify those problems, then in phase 11 --
9 that is in the second and third year of this activity -- wo 10 can define the specific actions for NRC to take. At the 11 moment we don't have any prescription for what to do or what n( f 12 rules to come out with because we don't know the status. We 13 don't know the temperature of the patient.
14 By the way, we also have to realize that we don't 15 have a handy thermometer, means or standards to measure that 16 temperature, so that's why we are working on tools to do this 17 survey, to establish the present status, 10 MR. KEftR From what you told me earlier, you don't 19 oven know what the temperature of the patlant should be, so 20 that you are going to have to decide that, as well.
21 MM, JANKOVICH Yea.
22 MR. KEMit t And how are you going to decide that?
23 MR. JANKOVICH: Well, the tools we are going to s
24 develop will help us. These tools are site surveys, the 25 Indicatore, activittee of the standard groups, and all through
22 t
h~#
1 these activities, interaction with the industry groups 2 involved in maintenance.
8 MR. KERR: Well, it appears t o me that that will 4 tell you what is going on. It's hard for me to see how that 5 will tell you what should be happening.
6 MR. JANKOVICH: We don't do this in isolation. We 7 are discussing,the subject with, of course, the Subcommittee, 8 and we are comparing our situation with other countries, and 9 the conclusion of our first year activities should give us a 10 good indication of where we are. Don't forget, one of the 11 objectives of this work is to identify good pract ices of -
("N,
( j 12 maintenance, and once we have those, that could serve as sort 18 of a desirable goal 14 MR. KERR: Thank you, i
15 MR. WARD: Could I follow further on this? As far 16 as -- I mean there seem to be some implicit goals that you 17 think a plant's maint'enance program should achieve, and it 18 sort of sounds now as if you are talking about the performance 19 of Japanese plants with respect to how they are maintained as 20 a possible goal But since there is such a range of 21 performances among U.S. plants -- as I mean as Mr. Reed and 22 others have pointed out, some U.S. plants perform better than
~~
28 most Japanese plants in several respects, but in particular,
/
\~/
24 apparently, with regard to maintenance.
25 Why aren't the emphases of comparisons on --
just on
23
(_s\
1 the domestic plants where it would be easier to do and perhaps 2 more useful information?
3 MR. BOOHER: Could I answer that? The emphasis is 4 not on the Japanese plants for this phase. Emphasis is on 5 domestic plants. The whole phase 1 is to look only at 6 domestic plants, and we curtailed any further work on looking 7 at foreign experience until phase 11, if there is a phase ll.
8 MR. WARD: Okay. I guess I misunderstood, 9 apparently. Thank you.
10 MR. BOOHER: The problem was that we had looked at 11 the foreign experience, particularly Japanese plants, bec~ause I 12 of the trip rate differences, and when the people from NRC 13 went over, including one of the commissioners, to look at the 14 Japanese, the Japanese claimed their low trip rates were due 15 specifically to their preventive maintenance program.
16 Therefore, that urged us to look into this problem deeper.
17 CSlidel 18 MR. JANKOVICH: Let's look at how we are going to l 19 accomplish our objectives. As I indicated, this definition of l
i 20 maintenance is broad in scope. We include surveillance and I
i l 21 overall management of maintenance programs, and we want to l
22 focus on technical issues, which I will address in A. minute on
/~5 23 the next few slides, not just on administrative aspects.
l i
k_
24 We used a phased approach. That means we don't want 25 to run into and accomplish everything in the first few l
l
24
.(
\
h 1
'/ 1 months. As 1 indicated, in Phase I, which is about one year, 2 we want to accomplish this survey, and in Phase 11, then, we 3 will define the specific problems and determine what impact 4 maintenance has on safety issues and what actions NRC should 5 take.
6 For example, the products, what we envision for 7 Phase 11 is the definition of NRC concerns, and maybe an 8 outline of NRC actions. We hope to have a good handle on 9 human error aspects of maintenance. We want to see if there 10 are indicators which are universally accepted by both the 11 industry and the NRC, and those indicators are validated.~ That
) 12 means that they are a valid tool of showing what they are la supposed to show us. And we want to have some good practices 14 all ready for recommendation in Phase II And we want to see 15 if industry self-activities are sufficient.
16 Once we have those tools -- you know, the 17 thermometer --
then we can follow the activities and see their 18 effectiveness. And there are other aspects that we want to 19 accomplish with the Maintenance Program Plan. We all know 20 that there are a number of different programs going on in NRC
(
21 -- for example, the aging program, reliability program, the l 22 ALARA program. They all have some aspects which are related to
~') 23 maintenance. And NRC doesn't want to come out with fragmented i
[O 24 attitudes towards industry, that you do the maintenance 25 here, someth.ng in the radiation area, another thing in the
x 25 1 reliability area.
2 This program, the Maintenance and Surveillance S Program, is designed to pull together the results from the 4 various NRC programs and show one integrated approach to the 5 industry and one set of requirements, if there will be any.
6 And, of course, we want to coordinate with the industry groups 7 as we have discussed before.
8 CSlide]
9 I want to indicate to you what are the specific 10 technical issues which warrant maintenance to be treated in a
~
11 program like this, and hopefully these problems and technicar
(%
( )
12 issues, as they are called here, would be improved on. For 19 example, we all know that there is a lot of human errors in 14 the performance of maintenance activities. For example, 15 two-thirds of the LERs are due to human errors, 60 percent of 16 the LERs are maintenance related; and in our previous 17 presentations, we provided a number of specific examples, i
l 18 plant-by plant examples. I don't want to repeat them.
19 So that is one of the technical i s s u e s .. Others i
20 are: What are the indicators? How can we?
21 MR. REED: Back to item 1, human error. In your 22 Phase I, you are going to be following up on auditing, 28 monitoring, reviewing and looking for, let's say, the root l
24 causes in human error; is that correct?
25 MR. JANKOVICH: Correct. Not just in one program,
l w 26 1 but in a number of programs.
2 MR. REED: Do you have personally any opinion of S what might be a large problem contributing to human error?
4 For instance, would you say it's lack of training? Would you 5 say it's drugs in the workplace? Would you say it's low 6 mechanical comprehension? Or what would you have as an 7 opinion? Do you have a personal opinion?
8 MR. BOOHER: Could I answer that? My personal 9 opinion is based upon industry studies just completed, and 10 they actually tried to do root causes based on the INPO data
~
~
11 in our LERs, and they came out with something like procedures,
. -s
( 12 errors in procedures being the root cause of most of the 13 greatest p sportion of human error in maintenance. Something 14 like 12 percent was due to training. I don't know some of the 15 others. I know C.O. Woody had pointed out in a conference just 16 last week that it seemed, surprisingly, it was in the 1
17 procedures a r.s a .
18 MR. JANKOVICH: Human errors originate from more
! 19 than one source. Certainly training and qualifications is one 20 of them. Therefore, we have a specific project to address 21 that. We want to see if procedur(s are an important cause of 22 human error s therefore, we have another project, the wrong 23 train / wrong unit project, here in this program which looks at 24 that.
25 There could be organizational causes, too, and our
s g'
(
27 1 survey will be designed to explore that aspect too.
1 2 MR. REED: I recall in the last few months at one 3 utility there have been one death and two or three near 4 fatalities, and I am trying to relate -- this thing with 5 procedures bothers me. It seems to me that if a person is 6 operating a crane with a freight cable or a hook that is 7 improperly rigged, that written procedures are not going to O take care of those things. There has to be a certain aptitude 9 with respect to the people that are involved in and around 10 equipment.
11 So I hope that you will seriously look at this ~ issue A
k,,) 12 of natural ability.
18 MR. JANKOVICH: I will tell you an interesting 14 example which comes from our first site visit on the wrong 15 train / wrong unit problem. We went to Dresden where they had 16 this problem. The maintenance mechanic went to the wrong 17 weld. Why? He was a new trainee. This was his first 18 assignment. He used procedures which had been used eight 19 times previously in the entire 14 years of operational 20 experience of the plant, and the unwritten procedure was that 21 a person who has already done that function would discuss the 22 situation with the new person who goes to the job.
23 So this did take place. An experienced mechanic told
[j\
24 the new mechanic to go out and do it, and he still ended up at 25 the wrong unit. There were color codings and it still
m fg 28 1 happened. So it is hard to really define what was the root 2 cause in this case. It was human error, of course, because he 8 went to the wrong units but why? He was trained, he was told, 4 but it still happened. So this kind of specific example, here 5 the obvious cause is a little refinement of the procedure is 6 important.
7 MR. WARD: What was the root cause of that? Was the 8 procedure not clear? Was the training improper? Was the 9 mechanic not able to comprehend the work? Or what was the 10 root cause there?
~
11 MR. JANKOVICH: The conclusion of the team was that k ,) 12 there were two basic root causes. One is the procedure --
18 MR. WARD: Uh-oh. When there are two, that often 14 means that neither of them are the root.
15 MR. JANKOVICH: Well, this is the conclusion as of 16 the moment. This was about three weeks ago. They said the 17 procedure was a modified procedure. They used it elsewhere, 18 and when they went to this par *icular weld eight times in 15 l
19 years, then they modified it. So that was not clear, not a 20 valve-specific procedure.
21 And the other thing is this person was the first 22 time on his own alone. He should not have been given this l
28 critical function. Of ch;urse, this could be a management i s 24 problem, but at the same time, $he trip report identified it w
25 as sort of a training problem.
r
, 29
1 MR. WARD: This was the Dresden plant. Glenn, we 2 will have to find out if Dresden uses the eel natural ability 3 selection.
4 MR. REED: I'm pretty sure they do not --
5 MR. CWALINA: I would like to add to that if 1 6 could. What happened on that is the person who was trained 7 had been trained for about six months, and at that time all 8 the Dresden training was on Unit 2. He had been trained solely 9 on Unit 2, and his first job assignment was to work on Unit 10 S. So he had a mindset to work on Unit 2, and he went in
~
11 and worked on Unit 2 instead of Unit 3 and caused the trip.
12 MR. REED: I didn't want to bring this up, but i 13 have been goaded into it by what has been said. I'm thinking 14 of another case of wrong train / wrong unit, and actually, there 15 were eight mistakes in series. I am thinking of the analysis 16 that was done to try to ferret that out, and I'm thinking of 17 the fact that the person had 25 percentile mechanical 18 comprehension, which was a marginal pass score, a low 19 margin. Probably he should not have been passed.
20 Wrong train / wrong unit is a very serious maintenance 7
21 issue, a maintenance error problem issue. It has many facets, 22 and they are identification, they are procedures, they are
[ h 23 training. In my opinion, the one that is most often forgotten 24 is natural ability. So that's a tough one, and it's one that 25 you people should certainly be addressing.
/ SO (NU) '
, I think one of the first things that we could do and 2 get out of the way and over the hump on is wrong train / wrong 8 unit, in most cases. We do have these two-unit plants that are 4 pretty much integrated.
5 MR. WARD: Could I go back and ask Hal Booher a 6 question? Hal, you referred to an industry study which 7 attempted to establish root causes for some sample of the LER 8 reported maintenance errors. What study was that? Could you 9 be more specific?
10 MR. BOOHER: I don't know the exact title.
11 Mr. Woody presented this at the ANS conference in Boston last D
1 i
\ss/ 12 week, and I don't know if they have really published anything 18 on this, but it is part of the NUMARC initiatives. They had 14 several One was to develop maintenance indicators. Another 15 was to do their own surv'ey of root causes to see whether or ty i-16 not our Maintenance and Surveillance Program Plan was tackling 17 the right things.
18 They came upavith essentially the same numbers that 19 we had in terms of that there is a maintenance problem in the 20 industry. They could not find any programmatic cause, 21 however. They felt it was spurious and probably plant 22 specific. But in doing so, I think they analyzed something
[ \ 23 like --
I forget the actual total numbers. I would have to 24 talk to Mr. Woody. If you would like that information, t 25 could get more specifics on it.
F~
-s gj 1 MR. WARD: Well, this was a paper at a conference, 2 then?
8 MR. BOOHER: it was a paper at a conference but it 4 also has been reported within industry groups themselves, 5 their working groups. And basically they are saying that our 6 numbers and their numbers aren't that far offs it's how we go 7 about tackling the problem, is the issue.
8 MR. REED: Just one last question on this issue of 9 wrong train / wrong unit. What are'your plans? It comes under 10 your number one human error problem. Do you have any specific 11 plans in phase I for trying to improve the situation on wrong O
ss! 12 train / wrong unit? We get notifications, it seems like, once a 18 month with respect to this aspect.
14 MR. JANKOVICH: Our specific plan at the moment is 15 that we study these reportings and summarize the root 16 causes. That will give us son.e bas i s to come up with 17 recommendations. Right now I would say we are at the level of 18 hearsay only: this is a wrong train / wrong unit event, that is 4
19 a wrong train / wrong unit event. There has got to be a common 20 basis to see what are the root causes, and if we can identify 21 one, two or three root causes, then we can recommend some 22 specific action.
4
( j 23 MR. KERR: I must say that I have a little 24 skepticism about this root cause because I have seen a good 25 many examples of people who have been trained on a unit who l
m 32 p_
/
k also made mistakes, and it seems t o me it's easier to pick 1
2 this out and say, well, he was trained on another unit, that 3 must be the reason he made this mistake.
4 people make mistakes even if they have been trained 5 on the same unit. There are going to be human errors, and it 6 seems to me -- and you must certainly have reached this 7 conclusion -- that we are not ever going to reach that state 8 of perfection where there won't be any human errors. We can 9 perhaps decrease them some, but our systems have to be 10 designed to be human error tolerant to some extent, and they 11 are.
(O\) 12 So I think of electronics repair. Indeed, in 13 training people, all the training centers that I have seen, 14 they start out, at least, training people on this is a valve, 15 and you don't have to go to the plant, it seems to me, to 16 learn that this is a valve. So that may be a contributor.
17 But if that turns out to be a root cause in a lot of 18 situations, I would be skeptical of it.
19 Maybe you aren't, or maybe you are equally 20 skeptical. I don't know.
21 MR. JANKOJICH: We realize that we cannot eliminate 22 human error. The objective here is identify the role and the
[\ 23 significance of human error and then see what we can do about 24 reducing them by compensatory measures.
25 MR. CWALINA: If I may, I would like to add to what t
sa t
1 John was talking about. Right now the Section is 2 participating in the Wrong Unit / Wrong Train study. That is 8 Joint with us, AEOD and the TSOB Branch in NRR, and currently 4 one of our members is out on one of the inspections right now, 5 and he would participate in the Dresden inspection. So they 6 are going to be going to about ten or eleven plants that have 7 experienced wrong unit / wrong train events and try to determine 8 the causes or contributing factors to those events, and we 9 intend to report on those in the final report to the AEOD --
10 or to EDO, rather, in April of '86.
~
11 So we are looking at these closely, and we agree
(^^
\
12 that there are instances of human error that you just can't 13 avoid, and we are trying to identify the contributing factor.
14 in the Dresden case we believe the training on just one unit 15 was a contributing factdr, but it certainly wasn't the root 16 cause. The floor around the valve was painted blue. The 17 floor around the Unit 2 valve was painted yellow. The fact is 18 the guy made a mistake, and I agree with you those things l 19 happen.
l 20 MR. JANKOUICH: We are on item number S, the role of i
21 preventive maintenance in aging. We have to face the fact 22 that there is an aging power plant population there, and aging 23 factors will manifest somehow themselves in reduced 24 performance s but how do we compensate for that? What method l
i 25 of detection should be applied and what type of maintenance
34 7 ,,
should we use once there are noticeable effects? Should it be I
2 overall or should there be replacement? So that is the thrust 3 of this program in the aging effects.
4 We all realize that item 4, management and 5 orgnalzation of maintenance is very important, crucial for 6 good performance, so we want to see what are the problems 7 there and what NRC should do, if anything.
8 MR. WYLIE: You are coordinating that with the 9 research investigation into that subject?
10 MR. JANKOVICH: They did plan a program along those 11 lines, exactly. The management aspects of maintenance. 1Bu t
) 12 since they have a budget cut, they don't have anything going 13 on at the moment. We have to compensate for that fact here in i
14 this program. The way we want to do it is incorporate review 15 of management and organization of maintenance in our plant 16 surveys. The plant surveys will be conducted on a systematic 17 basis by a strict protocol We will descuss that a little bit 18 later in a separate presentation. And there will be questions 19 directly addressing these aspects of maintenance.
20 MR. KERR: May I go back to number 3 and make sut*e 21 or try to understand it better than i do? What is meant by 22 the role of praventive maintenance in countering aging
[~'\ 23 effects? If aging effec 2s means, for example, waarout, then 24 it seems to me there has been a tremendous amount of study 25 already in the military of mean life and it can be used as a
r-
,e 35 1 guide to replacement, for example.
2 So it isn't that that you are studying, apparently, 3 because that has been pretty,well studied. What is it that you 4 are trying to find out about~>
5 MR. JANKOVICH: We are coupling our work with the 6 work of the aging program in 9esearch. They see the effects of 7 aging on specific components just as you described. We want 8 to see how it is implemented at the present time in the 9 industry, and then if there is any role for NRC in doing 10 something in countering agir2 It is still not clear.
~
11 Let's see. The aging program already has two O
( ,) 12 reports. One addresses motor-operated valves. The other 18 addresses, I think, trip break. Both of these reports 14 recommend certain maintenance practices. As I told you 15 earlier.this morning, we don't want duplicate efforts going on 16 at NRC, so the maintenance program takes these recommendations 17 from the Research program and presents one uniform standard to 18 the industry.
19 Item 5. We don't have any specific criteria or 20 standards for acceptance or rejection of any maintenance 21 program. We don't want to be proscriptive anyhow, but we want 22 to encourage some uniformity or some sort of measure. That is T 28 what we mean by the lack of criteria and standards. And we j
24 identified another technical issue. That is the lack of 25 interface between operations and maintenance, disconnecting
36 1 systems for maintenance and returning them to operation. We 2 see that that is a source of error and events which we could a maybe eliminate.
4 MR. KERR: There is one item, it seems t a me , that 5 one might want to consider adding to that list, if 1 6 understand. Currently, I believe, a significant number of 7 trips and challenges to safety systems occur during 8 surveillance testing, it's mandated by technical 9 specifications to occur when the plant is operating. Now, the 10 schedule for those surveillance tests is somewhat arbitrary.
11 It is supposed to be calculated based on reliability -
( 12 considerations, but to a considerable extent, one could argue 18 that those are artificial.
14 It seems t o me that one ought to look at the 15 possibility of readjusting testing schedules. I don't want to 16 harp on the Japanese, but it is my impression, at least, that 17 they do not do testing when the plant is in operation, if one 18 did not do testing when the plant is in operation, I have an 19 Idea that there would be a significant decrease in challenges 20 to safety systems, and particularly the number of unscheduled 21 trips.
22 Are you looking at this sort of thing?
23 MR. JANKOVICH Yes, we do. And you are aware,
\
24 probably, that there is a technical specifications group 25 working at the moment on revising the guidelines for technical
F fT 87 1 specifications, and we are cooperating and working together 2 with them.
3 We didn't identify that problem as a technical 4 issue, but it is addressed in the program plan several times 5 as a possible trade-off, if there is a good preventive 6 maintenance program, then we can modify the technical 7 specifications requiring surveillance testing, b'at that is a 8 project which should get maybe more emphasis. But we think we 9 will accomplish that in phase 11. We are too early in the 10 project at the moment.
11 MR. REED: I think what Dr. Kerr is saying in the 12 focus of his comments relates to who is studying, if anyone, 18 and under whose umbrella is the issue of frequency of 14 surveillance testing in the operating mode. Now, I don't 15 think the tech spec group is addressing that issue because 16 what they are addressing is format and overhauling and 17 separation of tech specs, but I don't think they are 18 addressing the issue of frequency of testing.
19 Now, is it under your umbrella to look at,this 20 frequency and to perhaps come out somewhere along the linq in 21 Phase 11 or something, recommending a lesser frequency?
22 MR. J AN.".0V I C H : We don't do it in phase 1, but we p 28 are thinking about it and we are hoping to include it in phase 24 11 25 MR. KERR: Well, why don't you do it in phase l?
F 38 1 MR. BOOHER: Could I interject? I understand there 2 are two tech spec groups. There is another tech spec group 3 which was headed under Don Beckin for tech spec improvement, 4 and I believe that group is looking at other aspects of this 5 MR. KERR: I am not talking about tech specs I'm 6 talking about testing.
7 MR. BOOHER: Well, they are including testing in 8 this.
g MR. KERR: I'm talking about surveillance and 10 maintenance, which you guys are talking about, and I'm talking
~
11 about human error. Now, it is possible that there could tne 12 some human error within the NRC Staff. The human error may be 18 that you are requiring survelliance testing too frequently or 14 not frequently enough, and it seems to me it's unfortunate not 15 to include a considerati*on of this.
16 MR. BOOHER : I would just like to support you on 17 that. I believe, quite frankly, when we set up these technical 18 lasues, they are the same technical issues we had almost a 19 year and a half ago, and I would have to concur with you that 20 we probably should have had a seventh technical issue, which 21 is this very issue, and I think that what we are finding 22 with the difference between the Japanese and the U.S. is that 23 the whole approach is so different that in order to conduct 24 any kind of tradeoff between too much testing and not enough, 25 and also where does preventive maintnenance allow this to
39 O
'- 1 change, it should be looked at under this program.
2 MR. WARD: Of course, when you look at that, I think 3 it's obvious that if you back off on on-line surveillance 4 testing, it's going to help plant performance, but the down 5 side is whether it affects the reliability of a safety system, 6 so if you are going to look at this, you have to bring in, you 7 know, risk analysis people or someone who will assess the 8 possible down side.
9 I mean the Japanese have part of their good 10 performance with regard to low spurious scrams as attributed 11 to the fact that they don't test the scram circuits on i The .
12 That's my understanding, anyway. On the other hand, I don't 18 know whether or not the Japanese have a better or a poorer ,
l 14 performance as far as reliability of the scram system as a 15 result of that. You know, I don't think anybody knows, but l
'16 that has to be weighed and made a part of this thing.
17 MR. REED: Let me try to pin this down again because 18 I think i got a statement that under the umbrella of the 19 maintenance activities of the Staff, you are going to look at 20 frequency or you hope to look at the frequency and whether or not the frequency is a good idea. Don't get me wrong. I have 21 22 always felt that a lot of trips in the first ten years or 23 first two decades of the reactor business are good, and if 24 they come from human error in surveillance testing, fine, they 25 shake the system.
F.
40 <
N 1 Some people say they challenge the systems and that 2 is bad, but there are sometimes things to be learned from 8 testing where human error occurs and this, that and the 1
i 4 other, but as we get older in the business, the things to be 5 learned -- that is, the benefits from frequent surveillance 6 testing -- should shrink into being a negative benefit, and 7 the equipment and the circuits are all fine and we are just 8 introducing human errors and challenges. f 9 So is it, in fact, going to be your pursuit, this i 10 issue of frequency of testing? I really don't believe you are 11 right when you say it's under tech spec special program -
~%
12 because I am on that committee.
18 MR. BOOHER: No, I didn't mean --
that may be 14 the solution, but they have been starting to look at this 15 issue. The way this program started out is we were 16 maintenance only, but it has grown, and NUMARC has said that 17 our program plan ~~ we have changed it to be called 18 Maintenance and Surveillance, not just maintenance, by which 19 we mean to infer that we do Intend to cover the testing and 20 surveillance aspects in our scope.
21 l'm not saying in the first year we are going to i
22 have the solutions, but we have to look at it this broadly, 29 and we have discussed this with NUMARC, who claim that we now O 24 cover something like 75 percent of all the activities at the 25 plant within this program. That doesn't mean we are going to
- - . . - - - - _ - -- - .-_.- .-.~ --___ . . . _-- - -
I 41 7s
('
1 come out with regulations for everything, but we are using a 2 broad enough scope to find out what is the real cause, and if a the real number one cause is too many trips, that might be the 4 very first thing we need to work on and to correct.
5 MR. JANKOVICH. Mr. Ward, I would like to add to 6 this answer. We tried to initiate a pilot study. We realized 7 that we have some facts already that speak much better than 8 any paperwork, so we went to LaSalle in October and presented 9 them with our maintenance approach and explained it clearly to 10 them that there is a possibility to do some tradeoff between 11 preventive maintenance and change in surveillance t es t i ng".
12 The reason that LaSalle gave was it was one of those 18 plants which was compared to Fukishima, an identical Japanese 14 plant where they have noch bet ter operating experience and 15 less surveillance testing. That was in October.
16 in early November, Mr. Denton met with the president 17 and Vice president of Commonwealth Edison on a separate l
18 occasion. Mr. Denton presented the same idea to top 19 management there, and it was discussed, and by January, the 20 whole subject just died away.
21 So we tried to initiate a pilot study to make this f
[ 22 tradeoff, and nothing happened. That was even before we i
i
/) 28 started the phase 1. So the issue is not dead. We are still N_)
24 trying to accomplish that.
25 MR. WYLIE: Did you get a clear understanding, L
42
,m s- 1 Glenn, to your question? I'm not sure i did.
2 MR. REED: Well, I think i got an understanding that a the word " surveillance" is in the plan and that surveillance 4 refers to the frequency, and you are going to be addressing 5 that lasue.
6 MR. BOOHER: That is correct.
7 MR. WYLIE: Welle let me see if I can state it.
8 Will the plan include frequency of testing and analysis or 9 assessment of the effects on safety? Is that the question?
10 is that going to be a part of the plan?
11 MR. BOOHER: That will be a part of the plan. -1
() 12 cannot say how specifically that will be addressed, whether in phase i or Phase 11, because Phase I is the more broad aspect 13 14 to look at all the problems in the plant that might be 15 affecting safety.
16 MR. WYLIEt Thank you.
17 CSlide]
18 MR. JANKOVICH: We are at the milestone chart. This 19 shows how we schedule our program, phase I will take 20 approximately a year, and then we have some time for the EDO 21 to review our progress and then approve phase 11, which would 22 last two years, in tne meantime, around half-time, we include
-~ 28 another EDO review, and at the end, of course, there is V 24 another EDO review, 25 We are now in the first quarter of the first year at
I 43
\d )
1 the very beginning, so there is no slippage or delay at the 2 moment.
S Please notice the lower half of the chart where we 4 Indicate continuous interaction with the industry groups, with 5 NUMARC, INpO, EpRl and the standard activity. That is very 6 important. We want to contribute to their activities and we 7 want to benefit from the results of their activities.
8 CSildal g MR. WYLIE: You mentioned your agreement with 10 NUMARC and the exchange of information. What is that 11 agreement? -
( 12 MR. JANKOVICH: I couldn't hear you quite clearly.
13 MR. CWALINA: There was an agreement made a couple 14 years ago with NUMARC and INpO, or rather specifically with 15 INPO about exchange of leformation, allowing us to look at 16 their plant evaluations and that type of thing, where we go 17 down to Atlanta and look at them at their offices. What wo 18 are trying to do in our plan is to find out if we can expand Ig that agreement so that it covers the maintenance activities.
20 The feelings of the NUMARC group are that it is 21 sufficient, and we are working on that with INp0 right now.
22 MR. WYLIE But so far you don't have that 28 agreement.
Os We are 24 MR. CWALINA: We assume that it's in effect.
25 trying to verify that it is.
r 44
(%
k, s 1 MR. REED: Charley, I don't think your mike is 2 working.
3 MR, JANKOVICH: These are the nine specific action 4 items, specific projects that we want t'o accomplish in phase 5 1. Some of them ar e sufficiently important that they warrant 6 a separate presentation, so three of them, the survey, the 7 second one, the Indicators, and the qualifications and 8 training, which is number 7, will be presented to you after my 9 presentation separately and expanded.
10 Here I want to put them into their proper place, and 11 then I want to give an overview to you and see what is -
12 involved in each of these specific projects, (O) 13 CSildel 14 So I will talk about number 2, which is the 15 maintenance indicators, and introduce that to you. Of course, 16 the objective is to establish some thermometer, as wo 17 discussed earlier this morning. Why? We want to see what is 18 the status of the maintenance at present. Interestingly, 19 HUMARC initiated a set of indicators last December, We 20 reviewed it in January, and that was a set of 11 indicators, 21 in February 1985, the NUMARC finalized their indicators, and 22 now we have ten. We think they are sufficiently good at the 23 moment that we could try them out, so in May, our EDO wrote to 24 NUMARC saying go ahead and try it, and at the moment we are 25 waiting, expecting the first set of data tapes to see how
F 45 fs\
\~ 1 those indicators do what they expect them to do.
2 You are aware also that earlier last year we 8 completed a smaller study. That was a pilot study on 4 indicatcrs where we demonstrated that, yes, indicators can be 5 established with relate plant safety to the performance of 6 maintenance programs. You will hear the results of this pilot 7 study later in detall.
8 We also want to use the Indicators to verify the 9 progress of what industry is making. Of course, it will need 10 validation. Hopefully, when we look at the data and analyze 11 It, we can accomplish that, and there is no decision at the
,Oi 12 moment on whether NRC should go ahead and develop a second set
( /
18 of Indicators or not 14 We will see the results of the NUMARC activity in 15 that area and the we can make a decision. I want you to be 6
16 aware of this, but we still are in a position to pursue our 17 pilot study further a'nd go along those lines and do that kind te of analysis.
19 CSilde3 20 in the meantime, we are plotting some convenient 21 data, and here, for example -- 1 don't want to dazzle you with 22 colors here, but it's interesting. Here we plotted the 28 abnormal occurrences which were reported to Congress. We are
\
24 required by law to do this. The total number of abnormal 25 occurrencos is indicated by the red line here, and regression
l 46
! \- 1 activation is the straight red line.
2 There is a tendency of increasing numbers --
3 MR. KERR It's interesting that if you eliminate 4 the last year, you could very readily draw a flat line, and if 5 it's total number of abnormal occurrences --
6 MR. JANKOVICH: Right, and we have to realize that 7 there are more plants each year as we go ahead. So I think it 8 would be better to normalize this data per plant or per every 9 ten plants, and we intend to do that. This is just the very 10 preliminary raw results, and I show it to you just for the 11 sake of discussion for this very reason, so we could see -i f 12 it's worthwhile normalizing, pursuing it further or not.
(D]
18 MR. WARD: John, could I ask you a question? As far 14 as normalizing, in particular the first year, do plants tend i 15 to have more occurrences of this sort in their first year of 16 operation?
18 MR. WARD: So I think a lot of plants came on line 19 in '84, didn't they?
20 MM. JANKOVICH: Right. Here we don't have control 21 over that consideration. On the next graph, I will show you l
l l 22 where we eliminated plants in the first three years after l 28 getting their operating license. The interesting trend here i
i l 24 la the blue line, the zig-zagged line is absolute number of 1
25 maintenance-related occurrences. The straight line is the
47
( ,/ 1 regression line computed for those data. The slope is twice 2 as much as for the total number of events, as for the straight S red line.
4 MR. KEHHr But it seems to me with this small number 5 of abnormal occurrences, the statistical fluctuation, if these 6 things were random, would be not much more than what you are 7 seeing here.
8 MR. JANKOVICH: Correct.
9 MR. KERRi i don't see that you can draw any 10 conclusions at all from these data. I mean they are 11 Interesting and I think you ought to look at them, but having 12 looked at them, I would film this away somewhere and use it in
}
13 another report.
14 Claughter3 15 MR. 900HERt We hope that's the last presentation of 16 that chart.
17 CSildeJ 18 MR. JANKOUICH: Here we have the number of trips per 19 plant versus the number of reactors. For example, let's look 20 here at number S. Three reactors had -- well, let's look at 21 the next one. Four reactors at the beginning of the blue line 22 had zero trips in 1983. That's what it means. And we have 23 calculated the means that is, the blue mean for 1983, the red 24 mean for 1984. This is supposed to be the regular, normal 25 distribution, the bell-shaped curve.
48 rN r
\s , 1 We see the standard distribution, plus and minus the 2 mean for both of these years. This tell us two little things, l l
3 If we want to use this kind of representation as an indicator, 4 that there is a large number of plants here with many trips 5 per year, and also, there are excellent performance here with 6 low number of trips. l 7 MR. WARD: Is that ordinate labeled right? I mean 8 zero reactors --
9 MR. JANKOVICH: No, zero trips. The number of trips 10 is the X, 11 MR. WARD: Yes, but the ordinate is the number of
'12 reactors.
1 13 MR. JANKOVICH: Right. Four reactors with zero l 14 trips. We are talking about --
l 15 MR. WARD: Wha,t I'm puzzled about is the points you 16 have for zero reactors.
17 MR. REED: There is no point there, is there?
18 MR. WARD Well, look at the other end. You have f i
19 got some points down there. What do those mean?
20 MM. JANKOVICHs That means that no reactors had 14 i
21 trips.
i l 22 MR. WAHD: And no reactors had 100 trips, either.
23 MR. JANKOVICHi No, that's right. We end up at 15, g-'g i h 24 That was the highest number.
i 25 MR. WARD I don't understand that plot, f
m l
49
\ 1 MR. JANKOVICH: Well, for example, look at 13. One 2 reactor had 13 trips. That is just how it happened. No 8 reactor had 14 trips. This could be another indicator if we 4 chose to do so.
5 l also want to make you realize that here we did not 6 normalize either. We could normalize, for example, each 1000 7 hours8.101852e-5 days <br />0.00194 hours <br />1.157407e-5 weeks <br />2.6635e-6 months <br /> of criticality or something.
l l 8 '9 KERR: Do you know whether the same group is 9 beyond the nine trips in each of the two years?
I 1
10 MR. JANKOVICH: Yes. There are, for example, five 11 repeaters here in the outside range. -
12 CSilde3 13 Action item No. 3, monitor industry activities. This 1
j 14 is a very important aspect of our work, and we have discussed
! 15 it earlier this morning. Why we do it? First of all, wo 16 don't want to act in a vacuum. We want to cooperate with our 17 counterpart in the industry. We want to summarize their l 18 findings, see if they are applicable to the NRC mlssion, and i
19 then document any results which will come out of their 20 activities.
21 There are a number of industry groups who we are 22 working with, NUMARC le maybe the most important. Starting 23 last June, we informed them about our program plan 24 development, in October we gave them another copy of our j 25 draft. In January and February we briefed them, we discussed i
50 f~N ws 1 their indicator work with them, we commented on the 2 indicators.
8 Another industry group is EpRI They do a number of 4 maintenance-related projects. We attended their seminars, we i
5 commented on their programs, and they brief us regularly, 6 about twice a year, on what the results they are getting.
7 MR. REED: John, I would like to remind you you have 8 got about 10 or 15 minutes to go, so sort of look at what --
9 MR. JANKOUICH: Fine. Well, let me just point out 10 one more interesting thing here.
11 LSlide3 -
O l 12 Last week we got a report from MIT. That was a
(
%.)
18 joint study, the last one here, the EPRI MIT study. They 14 looked at the operating results from German reactors. They 15 attempted to do the same thing that we did with the Japanese 16 reactors except they limited themselves to the operating I 17 experience, the factor of availability and the number of 18 trips. They didnt go any further, like trying to reduce the 19 differences to maintenance or looking at management practices, 20 as we did in our Japanese study.
21 They are asliing for funding. That's why they sent 22 us the report.
23 [Silde]
l ~
24 Standard groups. That is very important. We want 25 spontaneous industrial activity in the maintenance area.
51 1 MR. KERR: Any chance that we could get a copy of 2 that EpRI MIT report?
3 MR. JANKOVICH: Yes.
4 MR. WARD: You didn't tell us what conclusions were 5 reached in that report.
6 MR. JANKOVICH: I can tell you they looked at 7 availability for 1982, 1983, and 1984. The German reactors 8 are in the 70 and 80 percent range. We are in the 50 and 60 9 percent range. We plotted the findings in our Japanese study 10 on the same graph because we studied practically the same 11 years. -
I 12 The Japanese started lower than we are, and they are Y'
18 approaching the German results.
14 MR. REED: You know, when you get into German, 15 Japanese and United States, there is one thing you need in 16 indicators that you are going to have to sort out, and I would 17 like to see it someday put on a common, appropriate basis, and 18 that is the difference between availability and capacity 19 factors.
20 I believe the Germans use pretty exclusively 21 availability as their indicator of most importance. I believe i
22 the Japanese use the capacity factor, and unfortunately, we 7-wg 23 tend to use capacity factor. Really, I think the indicator of U 24 most significance is availability, so that is one of the 25 things you have got to sort out when you try to use these
52 b
\ 1 numbers.
2 MR. JANKOVICH: The MIT study looks at both of 8 those, availability and capacity, and we don't intend to do 4 anything further with it at the moment, 5 Standard groups. Let's see. This is the wrong 6 slide.
7 CSlide]
8 If there are industry standards, NRC could endorse j 9 those standards with Regulatory Guides. That would be an 10 ideal approach. Therefore, we want to participate in the 11 standards activities, either by encouraging the development of 12 standards and/or providing our own contribution.
\_/
18 There are three major standardization grouos where 14 maintenance-related activities go on: ANS and the group, S.9 15 naintenance group last year, and three events happened. In 16 September the committee was formed. In March '85, they 17 circulated an outiine of a standard, which was just a very 18 bare skeleton, and NUMARC indicated in a presentation to the 19 Commission in March that they are working on a draft standard 20 which will be ready in the fall, and it will be presented to 21 S.9.
22 So there is a standstill at the moment in the ANS.
28 MR. REED: Is there a member on the ANS 9.9 from the L
24 Staff?
25 MR. JANKOUICH: Dr. Booher.
53 1 ASME. Our division director got a letter in April 2 from the Operations and Maintenance Committee Chairman that 3 they don't want to do any preventive maintenance 4 standard. Their reasoning standard is that the subject is too 5 broad. However, we see a lot of other activities going on in 6 the Operations and Maintenance Committee.
7 For example, ASME is providing seminars for the 8 utilities on how to use the ASME standards. There are six 9 standards which are designated as.OMM 1, 2, 3, 4, 5, 6. Those 10 are addressing specific components: snubbers, the flow
~
11 circuit, the cooling systems, and they are practically test O\
( ,) 12 procedures.
18 They have a number of other standards under 14 development, specifically six other standards an the 15 development stage now by the members. Eleven other standards 16 are being developed. Some of them are revisions of earlier l 17 oncs, and nine other standards are in the long-range 18 planning. They are looking at the membership in questioning l
19 them, if there is a need for standards in that area.
l 20 But we must realize that these are 21 component-specific test procedures.
, 22 There is a lot of activity going on in the IEEE t
28 Working Group S.3. That group was formed last fall Their 24 objective is to develop good practices in their respective 25 areas for maintenance, and we are invited to participate, and
I~
54 k) m 1 Dr. Booher and Tommy Lee are members there. Right now they 2 are divining the scope of their work, what systems they should 3 be addressing, and this group is very enthusiastic and we can 4 expect a lot from them.
5 [ Slide]
6 Program integration. I can go through this quickly 7 because we already discussed it. NRC wants to come out with a O unified posture on maintenance, therefore, we want to 9 integrate and combine all maintenance-related outputs from 10 this agency into one through this program.
11 [ Slide] -
12 In the next picture, I will show you which are these
[\u))
13 programs. This is a list. This is in the Maintenance Program 14 Plan, too. We are getting their report. We are coordinating 15 our milestone schedule with their milestone schedule, and we 16 will review their reports, abstract the maintenance section 17 from it, and provide it to the EDO at the end of our Phase i 18 [ Slide]
{
19 We successfully concluded the next action item l
20 already, study of the Japanese maintenance programs. These i
! 21 are the two reports which I mentioned at the beginning of this l
c 22 meeting.
(g 23 I want to emphasize here that we compared the
\_ 24 maintenance practices, but they are not completely comparable 25 because the Japanese do a lot of preventive maintenance, but
E
,, 55
'\
'# 1 they also have some surveillance practices, so in our 2 comparative tables, we compare preventive maintenance with a some surveillance testing at the U.S., the Japanese preventive 4 maintenance against U.S. surveillance.
5 If we want to do anything more in this area, the 6 Japanese surveillance must be further explored.
7 MR. KERR: What would you say is the most important 8 one or two things that you learned from the Japanese study?
9 MR. JANKOVICH: Good results of their operating 10 experience. It's not that much due to the cultural 11 differences. It is due to their emphasis on preventive -
0)
( 12 maintenance, so much so that they think that they are 18 overdoing their preventive maintenance, and they want to share 14 the American operating experience and reduce their preventive 15 maintenance. That means in plain English they want to learn 16 from our faults.
17 MR. REED: 'I think we have touched this just a l 18 little bit, Dr. Booher, but somewhere down the trail a year 19 from now when you start to publish conclusions with respect to 20 the Japanese comparison, I think we should hold a closed 21 meeting prior to any final words that you put out.
22 MR. BOOHER: I see no problem with that.
p) s_
28 [Siide3 24 MR. JANKOVICH: Item 8, in-service inspections.
25 This is an interesting study, sort of small scale,
c 56 t
-- 1 but we still benefit from a larger program. You have to 2 realize Research is sponsoring a study of the ulstrasonic 8 testing devices where they see operating performance, how well 4 they do in a laboratory setting. They have different pipe 5 cracks set up in the lab, and they have a group of experienced 6 instrument operators coming in and performing inspections and 7 seeing how well they diagnose it.
8 So we are piggybacking to this project, identifying 9 the human elements, the performance-shaping factors which may 10 influence the results of their findings. For example, we look 11 at the operating experience of the individuals, differen6es in O)
(
%d 12 the human factors aspects of the various instruments, see 18 the man-machine interface, if there are factors which 14 influence the findings.
15 Of course, we look at fatigue, what could be the 16 human errors which result from the various instruments. We 17 use a number of accepted human factors test techniques to come 18 to some conclusions. For example, we used the critical 1
( 19 incident matter. This is the matter that they use in airline 20 accidents, if there is a mistake, for example. The operator i
! 21 didn't find a crack. Then he will be questioned: why didn't i
j 22 you find it, what were the contributing factors? Just like 28 in an airline accident they interview the pilot if he is 24 available, what were the contributing factors to the accident.
25 We will have a questionnaire looking at all the
i l
I l
1 v
1 backgrounds, and we will measure performance. So we will 2 benefit greatly from the technical program, the mechanical S evaluation of the instruments by using the human factors 4 methods.
5 And the last action in the program is the wrong 6 unit / wrong train. We have discussed it. As you see, all 7 through the years up to January '86, we are participating in 8 site visits trying to come up with some categories of human 9 error.
10 MR. REED: Is that it, John?
11 MR. JANKOVICH: Yes.
b
(_sl 12 MR. REED: Any more questions from the ACRS 18 members? All right, we thank you very much, and for keeping 14 on schedule, too.
15 And now we will take ten minutes and be back at 16 10:15.
17 [ Recess.3 18 MR. REED: Now, as we reconvene, the next 19 presentation is Mr. Cwalina on the survey project.
20 MR. CWALINA: My name is Gregg Cwalina. I am here to 21 talk about the survey. John touched briefly on the survey 22 project that we have. This is the biggest part of Phase i of
() 28 the Maintenance and Surveillance program plan, so we wanted to 24 go into this in a little bit of detail 25 We have issued a Technical Assistance Contract at
58 1 pNL to aid us in accomplishing this survey, and let me briefly 2 tell you the parts of it. What we have is we have broken it 3 down into about six tasks: data collection and assessment, 4 looking at the Salem pM program, developing a questionnaire 5 and a protocol, performing the site surveys, and then the 6 final part is just the summary report.
7 [ Slide]
8 This will just show you a little bit of the --
9 MR. KERR: Are you going to give us some idea of 10 what it is? This says a survey. What are you trying to find
~
11 out?
12 MR. CWALINA: In the survey what we are trying to la find out is what are the current programs and practices that 14 exist in the nuclear power industry today in maintenance.
15 What exactly is out t h e r.m ? This is our baseline data. We are 16 trying to find out what do they have out there in the 17 industry, and then in phase 11 we are going to be looking at is specific problem areas and trying to find out where they can l
19 be improved.
20 MR. KERR: What if what they have out there is much 21 too broad? What part of what they have out there are you 22 looking for?
i
/ 23 MR. CWALINA: We are looking at the practices in the
! 24 maintenance program, what percentage is corrective maintenance l
25 and what percentage is preventive maintenance. How do they
6
/~' 59 I
's 1 plan their preventive and corrective maintenance? Where do 2 the maintenance and operations interface? Where is the 3 interface with rad protection? How much of it is done by 4 contractors, how much of it is done by utility personnel?
5 MR. KERR: Okay. That helps.
6 MR. CWALINA: It's a big, broad look at it, but 7 there are specific elements that we are going to be looking at 8 and that is going to be coming into developing the protocol 9 and the questionnaire, which I will get into in a little bit.
10 MR. WARD: Your title doesn't say anything about 11 surveillance. in just scanning through the slides, I'm not
)
/ 12 sure surveillance is included. That is part of the program 13 plan for some reason? Is that not part of this survey?
14 MR. CWALINA: It's part of the survey. It's kind of 15 an oversight, just not putting it in there. A lot of the 16 slides were prepared a long time ago when it was just a 17 maintenance plan and just wasn't carried over.
18 MR. WARD: But specifically you are looking at the 19 surveillance --
20 MR. CWALINA: We are looking at surveillance 21 practices too, part of the surveillance practices are going 22 to be how they determine their corrective, preventive 23 maintenance based on their surveillances. So bear in mind 24 that is in there even though it may not come across per se in 25 the slides.
60 0 1 This is just a little bit of a matrix layout showing 2 you just the interactions of each of the tasks.
S [ Slide]
4 As you can see, all the tasks are kind of --
they 5 kind of feed into each other, the questionnaires and surveys 6 and protocol development. They all blend in to the whole 7 survey project.
8 [ Slide]
9 MR. KERR: What should I learn from this slide?
10 MR. CWALINA: This slide will tell you that 11 everything is blended in together. The survey project h rs (v 12 five or six different tasks, but each one feeds the other to 18 some extent. Just trying to show you the interface, the 14 interactions between each part or each task of the survey.
15 MR. KERR: As I look at this, I can see that they 16 all have interactions, I just didn't know whether I was 17 missing some important message or not.
18 MR. CWALINA: The important message is that each 19 piece is not being done independently. We are not on five i
20 different straight line paths to a conclusion.
21 MR. KERR: Okay. Let's take, for example: Visit 22 Salem and assess preventive maintenance. What do you
[ 28 acceeplish by visiting Salem that you couldn't accomplish by V
j 24 reading some reports? Do you want to see what the people 25 thern look like?
61 1 MR. CWALINA: No. Salem is an interesting case 2 becuse they are in a preventive maintenance program that's 8 dictated by the NRC. Because of the Salem ATWS event, we told 4 them that they have to develop a comprehensive preventive 5 maintenance program. They have done what looks like a pretty 6 good job of developing a big maintenance program on all their 7 safety-related equipment.
8 When we were there about a year ago --
the 9 maintenance group and I were there as an ISE representative --
10 they liked it so much that they have expanded it to balance of
~
11 plant equipment.
O
( ,) 12 By reading the reports, you can't get a real flavor la for what they do at the plant, how they plan it, how they go 14 about performing their preventive maintenance. When you go 15 out there and actually see them in action and talk to the 16 people, that's when you really learn something. You learn a 17 lot more than in a report by talking to the people at the 18 site. That's why we are going on these site visits.
19 MR. KERR: What sorts of things, for example, do you 20 learn that you wouldn't learn from a report?
21 MR. CWALINA: For instance, you don't get to see the 22 paper flow, the actual paper flow. You don't get to see how 28 the people go about actually planning the maintenance, who
\
24 works with whcm. It could be written down in the report, but 25 in a report they will write down: Well, I tr,terfaced with
(O 1 this guy, but do they really interface? Do they talk to each 2 other? Does one guy send the other guy a report? Things that u
8 are actually written in the documents don't necessarily happen 4 in the plants. I'm sure you are well awarn of that. And a 5 lot of the plants have a lot more going on than what is 6 exactly written in the documents.
7 CSlide]
8 MR. BOOHER: I want to add just one thing. We are 9 on a very tight schedule. In order to meet our final report, 10 which should be comstated more like December in order for us
~
11 to come up with a report to the EDO in March, and this is more f\
(j 12 of a kind of breakdown of the tasks and their contractual 18 agreement with the contractor to make sure there is no 14 slippage.
15 MR. CWALINA: The first task on the project is just 16 data collection and assessment, What we are looking at is 17 background information, any kind of a maintenance program 18 characterstic that we can identify, and some measure of the 19 maintenance program effectiveness. That will come in the 20 indicators work in just looking at some of the stuff John 21 talked about, normal occurrences, trip rates, number of 22 LERs. Those type of things will give us some idea of the
~') 28 maintenance program effectiveness.
[V 24 MR. REED: I just want to throw out a personal 25 prejudice. When you are looking at these indicators, in my
i l
i 68 1 opinion the most important maintenanco performance indicator r
2 is forced outage rate, and that's what I hope I will hear a j 8 lot about.
l 4
4 MR. CWALINA: That is one of the NUMARC maintenance 1 5 performance indicators. That's also one of the ones that we ,
6 identified.
7 Data analysis. We are going to be trying to look at 8 good features of shortcomings, deficiencies in the maintenance 9 programs. What exactly those are we haven't determined yet.
10 We are hoping to see something stand out. I'm not sure that 1
' ~
11 something will. And a comparative analysis, program I k 12 differences versus results on a plant-to-plant basis. Why is I
18 Plant A a good actor whereas Plant B doesn't seem to be such a 14 good actor? What are the differences? Can we attribute that i
15 to some of their maintenance characteristics or their >
16 maintenance program?
4 17 MR. REED: Now, this data collection and assessment 1
18 task parallels,'doesn't it, quite closely a lot of the stuff L
19 that INPO does, right?
20 MR. CWALINA: Right, it does.
21 MR. WARD: And are you working closely with them in j 22 their in their assessments?
g 28 MR. CWHLINA: We are working with INPO. We are 1
f j 24 trying to reach an agreement. We should have the agreement to I
25 be able to go down to the INPO offices and look at their
64
['N 1 Individual plant evaluation reports. That's part of it.
2 MR. BOOHER: The one thing that we have been tasked 3 with in this phase 1, though, is to have an independent 4 assessment for NRC, so although we are working together and 5 cooperating, we do need to come up with the Staff's position, 6 and we hope that that should be fairly close to INpO's but we 7 do have to maintain a certain level of independence.
O MR. JANKOVICH: I would like to add to that 9 response, too, that the INp0 plant evaluations consist of a 10 team of 15 people going there to the site, and two of those 11 people look at maintenance. They identify specific problems O
\ms 12 and make recommendations to improve them, and that is the 18 extent of their survey. We don't have any indication -- NRC 14 in this case -- we don't know about good practices. We don't 15 know, for example, how maintenance is done at that site in 16 generalt what is the organization, is there an overall 17 management commitment to good maintenance? So we are doing 18 more in our survey than INpO, 19 lNp0 has a specific function: that is, to take care 20 of problems, and that their team does very effectively, but 21 that doesn't provide the nation, NRC with an overall status of 22 maintenance, q f 23 MR. REED: What I am thinking, though, is maybe 24 their teams only have two that focus on maintenance, but they 25 are going from plant to plant to plant and they cover all the
65 l
k I plants, I guess, in a certain period, and so that takes really l 2 a lot of manpower over a long period of time. And they 8 certainly must be able to integrate this information into best 4 practices and best indicators.
5 You could only do, what, one or two plants on a more 6 thorough basis?
7 MR. JANKOVICH: We plan to visit six sites this 8 year, Still, we have to see INpO's output on good 9 practices. That treasure of knowledge may exist in the mind of 10 the plant evaluators. They have a whole department to 11 accomplish that. But we don't have written documentation.
\_- 12 MR. REED: Your effort is to make sure, at least as 18 an independent look, whether you agree with what they are 14 doing or disagree, and that's it.
15 MR. JANKOVICH: Correct.
16 MR. CWALINA: And the data collection assessment, wo 17 are also looking at l&E Inspection reports, PAT team reports, 18 construction appralsal team reports, these type of things. So 19 we are trying to get as broad a base as possible in these 20 items. For instance, our site surveys, part of the selection 21 was based on have they had a recent PAT team inspection. If 22 they have, then we are going to draw on that knowledge instead g 28 of going ahead and duplicating it.
s 24 The Salem preventive maintenance program has a 25 couple of purposes involvedi identify the structure and i
r .
66 x/ 1 characteristics of the preventive maintenance program, at 2 least as dictated by the NRC, as implemented by the licensee 3 in compliance with NRC is probably a better way to put its to 4 identify the safety benefits and trying to do a cost-benefit 5 analysis. Look at the implementation costs of this program.
6 That will be done reviewing all the literature involved plus 7 going on site for an actual up-close look for about a week 8 inspection on site.
9 CSlide]
10 in order to expand the data base, we are going to be
~
11 developing a questionnaire for use by resident regional 12 inspectors. Since we realize that we are going out and we are 13 only looking at six sites, the amount of information we are 14 going to get from those six sites is going to be limited, so 15 we are trying to expand it to be able to cover all ninety 16 sites, by developing a questionnaire that can be answered by 17 regional resident inspectors to broaden our da t a base as n.uch 18 as possible to identify programs and characteristics of 19 maintenance programs in the industry.
20 We have briefed all the regions on this. I have 21 just come back from briefing four of the regions, and we are 22 going to get some more input from them. We are going to be 23 developing the questionnaire, sending it through them for 24 review and comment before we come up with a final 25 questionnaire.
p 67 U 1 The regions are heavily involved in this, in both 2 this and our plan and are participating in the site visits.
8 MR. REED: I think that's a good idea. I would be l
4 interested, I would personally, in seeing the questionnaire l 5 when you have it developed. Is that going to be made 6 available to the members?
7 MR. CWALINA: Sure. We can make that available.
8 MR. REED: 1 might point out that i feel the 9 resident inspectors, since they live at the sites, if they 10 weren't before they got there, they will get very familiar 11 with the workplace and the pros and cons and the climate and 12 tone and all these things, and I think they can provide 13 valuable input to anything such as a physical activity, mostly 14 physical, like maintenance.
15 MR. CWALINA: Yes, that was our feeling, too, and 16 that has come across just in the briefings I have had with 17 them. They have been very interested in the whole maintenance 18 area and very cooperative in this issue.
19 MR. BOOHER: Again, when we say maintenance, we mean 20 maintenance and surveillance for these purposes.
21 [ Slide 3 22 MR. CWALINA: Maintenance review protocol is also a 23 tool that we are developing for use i n our site surveys. The (O}
24 intent of this is to develop a good protocol so when we go out 25 to the six sites, we look at the same things in each site no
-n- -----.-w -,--,wr - - . - , --- - - - - , ,- - , ,
--w-- - - , - - - , , , , - - ,-n- x- - , - - - - --
68 i
( 1 we have the same data base for the six sites, plus this will 2 provide a backup. When we get the questionnaires from the 8 regional offices, we will be able to look at the 4 questionnaires and determine how they answered them in terms 5 of what we saw there and get a flavor for how the residents 6 are answering it, and it will help us in evaluating the other 7 sites.
8 Again, we have been working with the regions and 9 other offices in the NRC to develop this protocol. We have 10 gotten responses, I believe, from just about all of them now
~
11 in developing a protocol, and we will be developing it and
.O
( ,/ 12 sending it to them for comment and review before we go on site 13 visits.
14 We are going to be using it on the pilot site visit, 15 which is actually the first site visit that we are planning at 16 the end of July, and after that site visit, we intend to 17 refine it as much as possible and then use it steadily 18 through the next five site visits. Again, mainly we want to 19 make sure we cover the same things in all site visits so we 20 have a common data base when we make our conclusions and 21 recommendations.
22 MR. REED: These site visits, you will be sending a 29 team. How many people are you going to have on the temm?
x 24 MR. CWALINA: Right now we are having someone from 25 our section being the team leader, We will be having two
O contractors from PNL on the site survey. We have asked for a 1
2 regional and an l&E headquarters representative, so we could a have as many as five. We have not confirmed that. On the i 4 first site visit, I believe Region 3 has already agreed to 5 send a representative, and we are working with 1&E 6 headquarters because we would like to get a representative 7 from ISE headquarters also.
O MR. REED: pHL. Who la PNL7 9 MR. CWALINA: Pacific Northwest --
10 MR. REED: Oh, the Battelle group, who have been 11 actively involved in the Japanese work. All right, thank you.
\ 12 CSilde3 13 MR. CWALINA: The site survey. As I mentioned, we 14 are planning on looking at sites. Site selection. We have 15 already come up with a preliminary set of six sites. They 16 haven't been finalized. I believe two of them have already 17 agreed to the site visits. We are tryeng to work with the 18 licensees to make sure they can support us so we can get the 19 information we need.
20 The base of the site visits. We are trying to get a 21 pretty broad cross-section, so right now we have got two 22 Westinghouse, two GE, a B&W and a CE selected. We are trying 23 to get a range in age, a range in reactor size, a big utility, 24 a small utility, trying to get as broad a base as we can on 25 these six site surveys, and that was the basis for the six
n 70 l 1 that we have selected.
2 MR. REED: Are you indicating that you are having 3 some resistance from the licensee with respect to entertaining 4 this team group?
5 MR. CWALINA: No, we haven't had any at all. I'm 6 just saying that there may be reasons. For instance, they are 7 pretty busy during refueling outages and that type of thing, e so if there is scene k ind of a problem where they can't support 9 our group, then we will back off and go to another site, but 10 so f ar we have met no resistance. The two we have contacted
~ ~
11 have already agreed. We have contacted a third and we just 12 haven't heard back from them yet.
18 MR. REED: Well, let me point out that there are 14 some different kinds of people in the maintenance arena than 15 you normally find in headquarters, be it in company 16 headquarters or NRC headquarters. They are people who work 17 with their hands and their physical extremities and so on, and 18 they don't like a lot of paper and they are high on skill if 19 they are good maintenance people at all. I wouldn't be 20 surprised if they would be quite apprehensive of a white shirt 21 group coming in to review maintenance and worry about more 22 " call me stupid" paper, as I heard a maintenance man very 28 unhappily refer to some of the things that are coming down, as 24 " call me stupid."
25 MR. CWALINA: Well, I went on five site surveys on
1 the important to safety aspect last year, and I interviewed 2 the maintenance people. Most of the time they are very S cooperative. They didn't seem to have that attitude. If they 4 knew that you were coming out there for information gathering, 5 to learn on a survey, and you are not primarily interested in 6 enforcement actions and pointing the finger of blame at 7 someone, they are a lot more cooperative.
O We also had taht experience a couple of weeks ago in 9 the wrong unit / wrong train survey at Dresden. They said the 10 people were very cooperative and helpful. Once they got over 11 the initial hump of letting them know what they were there k 12 for, they really opened up and told them what they really 18 thought the problems were and the causes.
14 [ Slide 3 15 The last issue is just a summary report, of which my 16 slide seems to be missing, but again, it's just going to be 17 putting all the tasks together into a final report which will 18 be included in the AEOD, or rather the report to the EDO in 19 April, generally trying to characterize the current status of 20 maintenance, identify any generic issue and identify prompt 21 action required f there are any. I don't really expect to 22 come across any just based on my experience in the last year
() 23 and previous experience.
24 And then try to group plants by characteristics and 25 performance, see if we can identify what are the good
I l
72
's' 1 practices of programs that are out there, and if they ome 2 across in the maintenance indicators, what do they have in a common that makes them good or what do the bad actors have in 4 common that makes them bad.
5 MR. REED: You recognize that your summary report is 6 going to be quite a challenge and, if done well, could be very 7 contributory, if it is not done carefully and well, it could 8 just cause a lot of dissension. It's a real challenge to come 9 out with a good report in this area where so many people are 10 involved and workmen who can embarrass engineers are involved
~
11 and all kinds of things.
12 MR. CWALlHA: Yes, I think we recognize that.
13 MR. KERR I don't know whether this is a generic 14 issue, and maybe it comes under some other group, but if some 15 of these practices are embedded in tech specs, they need to be 16 changed. One of the impediments to making that change is, i 17 think, if I understand the current situation correctly, the 18 fact that public hearings are necessary if tech spec changes 19 occur. There isn't a thing that you or your group can do about 20 that except make recommendations that somehow these changes 21 be ensler to accomplish, and somebody within the organization, 22 I think, needs to do something about that.
23 MR. CWALINA Yes, that's right. I agree. We A
24 can make recommendations, but someone has to implement it.
25 And we are also limited to the fact that our recommendations
E l
l 73 1 are going to have to be based on plant safety. We can't tal l
l 2 about availability --
! 8 MR. KERR There has been a recent situation which 1 4 personally think affects plant safety, and that is diesel 5 testing, in which the Staff would like to see licensees change 6 their tech specs, and they, I am told, are encoutering some 7 reluctance on the part of licensees, and I would bet that one 8 of the big reluctances is just that idea of having to open l
9 things up to a set of challenges. That is a situation which, l
l
! 10 at least in my view, certainly affects plant safety.
11 MR. REED: Okay. Is that it?
12 MR. CWALINAr That's it.
- 13 MR. REED: All right. We are right on schedule, and i 14 we will listen to the Maintenance Indicator Study by l
15 Mr. McLaughlin, l
l 16 CSilde.]
17 MR. McLAUGHLIN: Good morning. I am poter 18 McLaughlin, and the portion of the maintenance and 19 surveillance program that I would like to talk to you about 20 this morning is that dealing with the maintenance indicators l
21 activities.
22 The effort that is planned for the future is 23 somewhat uncertain at the moment, so the majority of my 24 remarks are going to concern themselves with the report that 25 our technical assistance contractor, Battelle, is -- we're l
~
l r
a
74 1 about ready to publish as a NUPE0/CR that will be presumably 2 vallable in July.
(
3 CSilde.3 4 The effort in indicators --
5 MR. KERR I'm sorry. What is it that is in 6 doubt? You said something was in doubt.
7 MR. McLAUGHLIN: The NRC, as John dankovich 9 Indicated in his earlier comments, we are in a " walt and see" 9 attitude at the moment as to the accuracy, the viability of 10 the NUMARC initiative in indicators, and how that program 11 progresses before we continue any redundant effort in that 12 area.
13 MR. KERR1 Okay, thank you, s
14 MR. McLAUGHLIN: The Indicators work started 15 initially as a research project. The effort, and their report 16 on their effort, was concluded in July. The NRR Interest 17 started in July of that year and started to do an evaluation 19 of the management organization and its relationship to plant 19 safety. For those of you -- and I think most of you were in 20 attendance when we presented our report on the NUMEG/CR-8737, 21 Dr. Thurber from Battelle and myself came down and briefed the 22 Human Factors Subccensittee at that tilme and discussed the 23 findings that had come from that effort.
24 At the conclusion of that effort in November of 1984 s/ 25 and with the increased interest in maintenance and its
75 f
t contribution to plant safety, we shifted our emphasis in the
()
l 2 program into maintenance and started to refine the model that 8 had been developed for the management and organization 4 Indicators effort, and started to look at something that we 5 saw as a necessary part of this effort that would be to 6 establish the temporal eelationship between what we're going 7 to call intermediate Indicators and final indicators.
l e in other words, we didn't things happened in a plant g concurrently, and since one of the objectives of the program 10 was to have some predictive capability, we wanted to know how 11 a program was progressing, how it was performing before 12 degradation of plant safety could possibly occur.
I 19 So one of the things that you will see in this 14 upcoming work in the report that will bee published will be a 15 term that is probably new to you, which is an intermediate 16 Indicator as opposed to a final indicator, 17 The appedach that we have taken in this effort deals I
le primarily with identifying and coding of existing plant 19 maintenance program characteristics.
20 C5ilde.J l 21 We are going to call those resources in this effort, 22 and the emphasis should be on the fact that these are 23 existing. These were already pre-available data to us that 24 had been used in the earlier management and organization O. 25 elfort.
L
70 t We had the characteristics of plant maintenance
[~'T 2 programs which we broke down into three subcategories of U
8 administration, which included the organizational structures, 4 policies and procedures about how they conducted their 5 business a the human subcategory which dealt with the 6 experience, qualifications and training of the Individuals and 7 staff that were involved in the maintenance programs and 0 finally, the technical subcategory which dealt with the tools, 9 equipment, the parts availability, those types of issues, that to would be available to the maintenance program.
11 We then looked at performance of the maintenance 12 program which we are calling maintenance intermediate safety 13 indicators, and we established parameters of that, and p_
14 hopefully, if our assuFption is correct, what we would see is 15 an indication of relationship between the performance of the 10 maintenance program and the resources assigned to that 17 program, and the overall safety of the plant, 10 CGilde.1 19 MH. HEED I am sitting here musing. You speak of 20 indicators and things that you're following, and i guess if 21 you're following the paper, the same paper that I see, we have 22 seen some serious problems at TUH with respect to injuries and 23 shutdowns and all units down, and these kind of thinge.
24 Are you now collecting indicators -- or have you 25 been plotting Indicators with respect to TV47
l 77 1 MM. McLAUGHLIN: Not specifically, no. The effort
/\ 2 that I am reporting on today, and our effort to date, has 3 concentrated on available data of the 1981 vintage that has l 4 been used in the earlier study for management and 5 organization, and there has been no plant-specific collection i 6 of data, nor use of new data from later timeframes than 7 1991-02.
l 6 MM, MEED: So you have picked a timeframe, and I l
l 9 therefore, you start your activity of following indicators and 10 trying to see if there's something wrong.
l 11 It seems to me that where we have all this i
l l 12 preliminary notices and abnormal occurrence reports and all 18 the kinds of reports, that sometimes it almost baccees ==
t 14 well, everyone is interested in what's wrong at TVA. It l
15 seemed to me that you, in the interest of maintenance, would 1
I l
16 have been following these indicators with respect to TVA.
17 MM, DOOHEM: Could I make a comment? One of the 18 important things of this whole program at the EDO's request l
19 and the Interaction with NUMAMC la that we are not going out l l 20 specifically collecting data on indicators on plants at this l
l 21 time. We don't want to come up with a reactive approach, We t i l
22 are supposedly to get all the indicator data through INPO and 28 our agreement, and we will be looking at all the plants, i 24 including TVA, as soon as we can get that data.
I
) 25 HM, MEED: You're saying okay, then, you don't want m._- _ __._._ ____ _ - , _ . . . _ _
78 1 too many. cooks at this time s you want to let them come in with
[ 2 it, and then you're going to review and see how it fits s
8 together.
4 MR. BOOHER: Right.
5 MR. McLAUGHLIN: Okay. The approach that i 6 described could be graphically modeled in this fashion. The 7 primary interest ,of the study was to determine the 8 relationship through the lower chain here Cindicating]. In 9 other words, the impact of maintenance resources which are 10 given to a maintenance program, the performance of that 11 program and then the ultimate effect of that program's 12 effectiveness on the overall operation of the plant from a 13 safety perspective.
14 The second areg of interest was the direct or 15 unchartered relationship between the maintenance program, how 16 it affects other programs within the plant -- for instance, 17 the relationsh.ip.between maintenance and operations may have 18 some secondary effect on the overall operation of the plant, 19 and we wanted to see how that fit into the program.
20 [ Slide.]
21 The maintenance resources that were used, or 22 characteristics of the maintenance program that were used in 28 the study were these. They come from a fairly extensive list 24 of ISS that we identified in the human, administrative and 25 technical area that I described earlier. Unfortunately, the L.
79 1 number that was available already coded and in the system in 2 machine-readable form that could be used in the study without S relying on additional information or new information, as 4 Mr. Reed suggested earlier, are these eleven.
5 MR. REED: You've got to beware of these maintenance 6 resources. Staff size, I automatically think well, big is 7 better. Some people think it. But that's not true. And 8 I might just point out that in my experience, that I know of, 9 the smallest staff size with the least number of ranks, which 10 i think you mean different job categories, has probably got 11 the lowest forced outage rate over the last 10 years in the 12 United States of America.
13 You've got to be very careful in evaluating what is 7S 14 a resource and what is a debit.
15 MR. McLAUGHLIN: Did you get our preliminary copy of 16 the report, Mr. Reed?
17 MR. REED: No, I didn't.
18 MR. McLAUGHLIN: That's exactly the result that was 19 found, and I will discuss that once we get down to the last 20 slide. That's one of the features of the maintenance program 21 which appears to have a significant effect statistically on 22 the operation of the program itself and on the overall plant 28 safety -- is the smaller the size of the maintenance operation 24 the better the performance. And consequently, the plant O
(_j/ 25 safety.
80 1 CSlide.]
2 The maintenance program performance characteristics f) v 8 that were, again, available to us from existing data, and 4 unfortunately this does not include the forced outage, forced 5 reduction, the trip rate statistic that all of us would like 6 to have in here, because when the original study was done by 7 the Battelle people we did not code those trips to indicate 8 those that were specifically related to maintenance or 9 surveillance causes.
10 Hopefully, if the effort that is proposed by NUMARC 11 and will be implemented by INp0 is successful, we will have 12 that type of information and be able to have a better handle
-sg 18 on that 14 The ranking here, or the parameters here used in the 15 modsl were primarily just breaking out the maintenance-related 16 LER's, the numbers, the maintenance SALp scores, the 17 surveillance SALp scores for plants, and then a separation of 18 teh 766 file non-compliance data into those rated 1, 2 or S 19 and those rated 4, 5 or 6 is what was done. Not a very l
20 rigorous attempt to quantify the data, but that's what we have i
21 used at the time.
22 MR. WARD: Do you know if the NUMARC study, in l, -
l 28 addition to looking at forced outages and that sort of thing, i
1 24 is looking at perhaps industrial safety performance for the l f% 25 maintenance department as an indicator?
i
81 1 MR. McLAUGHLIN: I believe they are. Greg, do you w 2 have that slide of the NUMARC indicators? I didn't bring one 8 with me, but we have one that would give you what they have 4 proposed to use. And we have looked at this internally, and 5 it looks like a fairly comprehensive list. It will give us 6 some real time data.
7 CSilde.3 8 There is a lost time accident rate for maintenance 9 personnel 10 MR. NARD: Yes, okay.
11 MR. McLAUGHLIN: They seem to have done a really 12 good job in trying to put together characteristics of 18 parameters that could be monitored, and to give you a feeling (N
(,,, 14 for how things are going.
15 MR. BOOHER: At last count, INPO had data from all 16 except, I believe, two plants on these indicators, for the 17 first go-around.
18 MR. REED: I think those are very good-looking 19 indicators there.
20 MR. WARD: Are they going to share that with you? I 21 presume it's not going to be public information. ;
22 MR. McLAUGHLIN: No. It will be shared under the 28 same system tnat we presently have for the plant-specific 24 NPRDS data, which is considered proprietary. And 1 25 understand, I guess, that the linkago will be through a
I l
82 l 1 computer and it will be a " read only" at our end, to get the j l
2 data. We will have no manipulative power over the results.
8 MR. WARD: Then you have to treat it as l
4 proprietary? Or is it somehow made -- l 5 MR. BOOHER: That's correct. We would have to treat that as proprietary data, i 6
7 MR. WARD: All right. Thank you.
8 MR. McLAUGHLIN: I can make a copy of this available 9 to you if you'd like.
10 MR. REED: I think the members would like a copy.
11 MR. KERR: I thought I had seen it in one of the 12 papers.
18 MR. McLAUGHLIN: Going back to the model that we had 14 used in the statistical sense, --
15 CSlide.3 16 -- in the maintenance indicators area, these are the 17 parameters that were used to indicate a final safe --
final 18 indicators of plant safety.
i 19 CSlide.]
20 MR. KERR: It seems to me there's an interesting 21 contrast here between --
if I understand the intent of the 22 indicators -- between this and the results of risk analyses.
28 The risk analyses all seem to indicate that one has a 24 principal contributor only if one gets core melt. There's
(, 25 not, at least to me, an obvious connection between these
88 1 indicators and core melt. Now, maybe one can establish one,
'\ 2 and if so, it will sure eliminate a lot of need for PRA's.
8 Sometime in your study it seems to me you need to 4 give some thought to whether one can correlate these safety 5 indicators with what one calculates with PRA's. I don't have 6 a good suggestion, but it seems to me one cannot ignore the 7 results of PRA's in this study.
O MR. McLAUGHLIN: Mr. Kerr, one of the things that we 9 had attempted to do here -- again, to get back to the 10 predictive nature of what we were trying to show, and again 11 trying to do it statistically --
we had identified a core melt 12 situation or incident with a massive radiological release as 18 the pen ultimate indicator of plant safety.
i' 14 Unfortunately, from a statistical sense, or 15 fortunately for us, not fortunately for the statistics, there 16 has not been a large number of those. A n'd consequently, it 17 doesn't lend itself very well to statistical analysis.
18 And what we tried to do here was to show something 19 that could be conceived, say, from the ALARA approach that 20 could be construed as a possible safety factor. And we had 21 numbers to work with.
22 MR. KERR: I understand the lamp post syndrome very 28 well.
24 MR. WARD: That's exactly what it is, I think.
A is, however, that what one is,
() 25 MR. KERR: My point 1
84 1 think, trying to do, all of us, is to avoid those things that
[d \ 2 contribute to risk. And if these really aren't significant 8 risk contributors, then we ought to be looking for something 4 else.
5 At this point, 1 don't know whether they are or 6 not. All I see is that we're l'ook i ng at some things which a 7 normal PRA, which is supposed to predict reactor risk, doesn't 8 aven consider.
9 MR. WARD: I would go further to say I think it is 10 dead wrong to do it this way. Would you put up that box? The 11 second slide you had, where you had the three boxes. And i 12 guess I didn't understand that your final safety indicators 18 box meant these things. The final radiological release,
(~)N t
14 occupational exposure.
15 Well, if by radiological release apparently what you 16 mean are the routine releases which you just showed. And you 17 know that is not the safety problem for nuclear power plants.
18 If that was all we had to be concerned about in nuclear power 19 plants, none of us would be here.
20 So I think to use that as a final safety indicator 21 is just wrong. I don't think it's going to tell you what you 22 want to know.
28 [ Slide.]
24 MR. BOOHER: This is not the ultimate solution.
s 25 Again, this was a fairly quick study done with data we could
85 1 get to see whether or not we should even have any hope of --
2 even if we could have data -- of showing any kind of !
l 8 relationships on an intermediate level 4 The other problem is much deeper and is going to 5 require a lot more research to resolve.
6 MR. REED: I'm not sure that I'm following what's 7 going on here but I am certainly disturbed about what i see as 8 final safety indicators and how that affects this maintenance 9 surveillance plan, and how it affects maintenance performance.
10 Certainly, these final sefety indicators are a great 11 deal related to the reactor concept, BWR versus PWR, and a 12 great deal related to design aspects or design problems such I
18 as steam generators which are, let's say, in some cases
\_/ 14 mis-designed or have to be changed out.
15 Well, how do the final safety indicators impact this 16 work that you're doing?
17 MR. McLAUGHLIN: There is a normalization factor 18 which had been done on this data so that the differences, the 19 chemical plant differences between BWR's and PWR's, for 20 instance, was accounted for -- the age of various plants was 21 accounted for, size was accounted for. Those factors were 22 taken into account during the regression analysis that was 28 done.
l 24 MR. REED: Dave, are you happy with what you were r
25 questioning? I'm unhappy that these final safety indicators
86 1 are even involved.
s 2 MR. REED: Yes. You see -- I mean, I think those
{Jb 3 might be another class of intermediate indicators, or some 4 sort of partial indicator. But your whole program there, 1 S mean the maintenance program, isn't directed toward minimizing 6 those things up there that you've called final safety 7 indicators. The ma is,tenance program is directed towards 8 minimizing the likelihood of a serious core over-heating 9 accident, a co*e melt accident.
10 MR. BOOHER: Excuse me. That's true. Again, I want 11 to emphasize that this was a short study. One of the big 12 problems is that because there is so much lack of 13 standardization and so many factors, is it even possible to do r
(g
%_/
)
14 through some kind of statistical analysis to show any kind of 15 relationship based on data that we could get.
16 The purpose of this is to show that -- and probably 17 the word " final" should"not be used there -- but the purpose 18 was to show that in spite of all these differences, if you use 19 just this model, we were able to come up with some 20 relationships that rose above all these unstandard practices.
21 MR. REED: Well, I hope what you're saying, 22 Dr. Bocher, is that you are not really going to use these as i
23 the final safety indicators, the overall evaluation.
i 24 MR. BOOHER: Not all I
25 MR. REED: I can see man rem exposure as important l
l l
l
.v y-- . - _ , - _ . - - - - -
87 1 as an indicator on a par with many other indicators.
'~5 2 MR. BOOHER: The indicators are going to be the (a I 3 industry indicators that we have already established. We also 4 have an NRC task force that looked at the industry indicators 5 and came up with the same indicators that you saw on the 6 earlier list. There are two additional indicators that the 7 industry is trying to develop which the NRC recommended, which S they will have not collected data on as yet.
9 MR. KERR: But the thing that caused me to raise the 10 original question is that these are supposed to be indicators, 11 I think, of safety. Now, it appears to me that you're going 12 to demonstrate that they are indicators of safety by showing 18 that there is a correlation between these and normal
(~%
i%- ) 14 radiological effluents. That is going.to be your method of 15 validating that they are safety indicators.
16 MR. BOOHER: Well, if you look at the NUMARC 17 indicators I think that's just one of many.
18 MR. KERR: Well, I have to judge what you're doing 19 on the basis of the presentation made here today, and it was 20 my impression that you are going to try to demonstrate that 21 these are valid indicators by running a correlation. If I'm 22 missing something, tell me.
28 MR. BOOHER: Not so much that these are the 24 indicators, but that the approach is worth looking into a 25 little more deeply, f) w.
88 1 MR. KERR: Well, you're going to demonstrate then, g-~ 2 that the approach is worth looking into more deeply by doing a 8 correlation. If the correlation is good, presumably the 4 approach is worth looking into more deeply.
5 MR. BOOHER: That is correct. And the kind of 6 things you brought up are the kind of things that have to be 7 looked at more in order to come up with any kind of final 8 safty indicators above and beyond the one on radiological 9 releases.
10 MR. KERR: I will remind you of something you have 11 probably seen in other contexts. .But I remember a friend of 12 mine wrote a book on systems analysis a number of years ago,
~
18 and in the opening chapter of the book he pointed out some of
'O, 14 the pitfalls in using correlations. And he noted that during 15 a 30-year period in Sweden I think it was there was a very 16 high correlation between historic population and the birth 17 rate. And I just leave that as one of the possible pitfalls 18 of using correlations.
19 MR. BOOHER: I think we would concur with that.
20 MR. WARD: Well, I guess the problem is, you know, l 21 it may be that what you're doing is useful, or certainly, I i-22 think those radiological releases are important. But you are l
28 couching it in terms of -- well, I mean you are looking at it 24 and presenting it and maybe you're going to act as if you did r
[ 25 have a handle on the true final safety indicators. And l
\
l %s .
l l
l
89 1 don't think you can do that.
[~'T 2 You know, it's impossible to take -- I mean, if I 3 just start out looking at those three blocks with their titles 4 on them, I agree with you that that would be an ideal way to 5 approach the problem. I mean, an academic way to look at it.
6 But it turns out you don't really have anything to put in the 7 last box. So I think you have to abandon the approach or make 8 some sort of attempt at correlating the things you've got in 9 the final bow or something else with what you would really 4
10 like to have in the final box.
11 MR. McLAUGHLIN: Well, I think that, in effect, is 12 where we are. What we started out to do was to develop the la model, which I think we have a model We tried to fill it t
14 with whatever existing data there was available to us, and we 15 did that within four months. And what we have are 16 inconclusive results due to a lack of substantive data and a 17 lack of good, final indicators of plant safety. Hopefully, 18 the model is in existence and the model still works.
19 If, in fact, the NUMARC indicators provide us better
'O data that is more reasonable to indicate a condition of plant 21 safety, then we will have that information available to us, 22 the data available to us, and we'll be able to exercise this 23 model using those different indicators.
24 MR. KERR: When you talk about a model, O)
( 25 Mr. McLaughlin, to what model do you refer?
I~
I 90 i l
1 MR. McLAUGHLIN: The model in our approach. This l
es 2 would be the model representation of what we have done.
U S MR. KERR: That is, that one does a correlation 4 between the indicators and the things in the final box, and if 5 it's a good correlation one concludes that the model works?
6 is that it?
7 MR. McLAUGHLIN: Well, there are other aspects. I 8 mean, we are not here to prove that there is irrefutable 9 evidence that says these are the final indicators and this is to our approach. There's three other -- the relationship of this 11 to this that we're looking at I mean, there are other things 1
12 to look at. i 18 MR. KERR: Please. I'm trying to understand. I'm l
I j
O) t
\_/ 14 not trying to be critical And I didn't use the term 15 irrefutable. I was trying to say, in other words, what i 16 thought you had said. And if it doesn't say that, correct me.
17 it was my impression that you have a model that 18 involves some computer calculation in which you run a 19 correlation between these indicators and the things in that 20 box. And I thought you said you had demonstrated that the 21 model worked. Did I misunderstand you?
MR. McLAUGHLIN: No. I would say that's a fair 22 representation of what I said. Yes.
28 MR. KERR: Okay. Now, my impression was that you 24 f5) 25 demonstrated that the model worked by showing that one got a
91 1 good correlation. Again, if I am misinterpreting you, that's 2 what I want you to tell me.
v 3 MR. McLAUGHLIN: Well, in most cases we did not get 4 a good correlation. As a matter of fact, the results seemed 5 to run contrary to intuitive logic that would say some things 6 affect maintenance and overall safety of the plant which 7 proved inconclusive or not obvious through the use of the 8 model 9 MR. KERR: Then when you say that the model worked, 10 you mean that it will run on the computer? Is that what you 11 mean?
12 MR. McLAUGHLIN: Well, I guess my interpretation of 18 that would be that we think we have something that will work O
w 14 given the right data, the computer coding and things that were 15 done or in fact in place, we've had experience with it.
16 MR. KERR: I'm trying to understand what you mean by 17 saying it will work.
18 MR. McLAUGHLIN: It will run on the computer, is 19 that what you said?
20 MR. KERR: No. I want to know what you mean by 21
- saying that your model will work. You tell me that that 22 doesn't mean that you've got a good correlation between the 23 Indicators and the things in the box. What does it mean to 24 say that the model worked?
O 25 MR. McLAUGHLIN: It means that we have developed an l
()
e --w , , - - , - - - . -
n- - -
92 1 approach which we think is viable and feasible, that the data 2 that we had available to us we used, we coded, we ran it on (A}
a the computer and we got some results.
4 MR. KERR: But you don't know whether the results 5 have any significance or not at this point?
6 MR. McLAUGHLIN: To safety?
7 MR. KERR: Well, to anything. That's meant to be a 8 question, not a statement.
9 MR. JANKOVICH: When we say that the model works we 10 mean that we found correlations at the moment at the 5 and 10 11 percent level, significance level, between the factors which 12 we considered. .
13 MR. KERR So you judged that the model works on the 14 basis of the correlations you got which you considered to 15 demonstrate a good correlation.
16 MR. JANKOVICH: Yes.
17 MR. KERR: Okay, because i thought when I asked 18 Mr. McLaughlin that question he didn't think that you had 19 gotten a good correlation.
20 l'm not trying to question it, I'm just trying to 21 understand how you --
22 MR. BOOHER: In a lot of areas we did not get good 23 correlations where we thought intuitively that we would get 24 good currelations, I think is what he was saying. In other O
\)
s 25 areas, there were correlations; perhaps some that were not
98 1 expected, but there were. In spite of the fact that there are
[~D 2 so many differences among plants and the data and situations, 8 when we took out this information we still were able to get 4 correlations above and beyond those factors such as age and 5 things such as that.
6 MR. KERR: Well now, when you get these correlations 7 that are contrary to intuition, are you convinced that your 8 intuition is wrong, or that maybe the model is not so good?
9 MR. JANKOVICH: If I read the first factor, for 10 example, from the draft report, that will clarify many of 11 these questions.
12 The first final safety indicator as it was defined 18 in the report is man rem exposures for 1982. What are the 14 significant maintenance-related factors which may contributed 15 to that? Number one, these are the elements in the middle of 16 the model Maintenance manager rank 3 significant at the 5 17 percent level. The higher the rank of the maintenance 18 manager, the better; the less man rem exposure we get.
19 Number 2, maintenance-related factor is the number 20 of ranks between the manager and the individuals. Significant 21 at the 5 percent level; that is, the fewer the ranks, the 22 smaller the chain of management, the better it is.
28 The third one. Combined maintenance and l&C 24 operations. l&C combined with the maintenance department.
W 25 Significant at the 5 percent level 3 if they are combined, we
E 94 1 find less man rem exposure.
['V } 2 Fourth factor is: is the maintenance department 3 combined with oprations. That's significant at the 5 percent 4 level, if they are combined, we get less man rem exposure.
5 Last factor is the size of the maintenance staff.
6 Significant at the 5 percent level That is, the smaller the 7 maintenance staff, the less man rem exposure per individual 8 So this is the kind of intermediate --
9 MR. KERR: Walt a minute. You mean there's a 5 10 percent correlation between the smallness of the maintenance 11 staff and the man rem euposure?
12 MR. JANKOVICH: Yes. That means if there is a large 18 maintenance department with several hundred people, it is more
(~%)
(J 14 likely that you will get higher exposures than in a tight, 15 smaller organization.
16 MR. KERR: And you consider 5 percent significance a 17 fairly good indicator?
18 MR. JANKOVICH: That's the best one we have at the 19 moment.
20 MR. REED: Pete, is that the end of your 21 presentation?
22 MR. McLAUGHLIN: Well. -- I don't want to leave in 28 flames. I had one more slide.
24 MR. WARD: Well, I want to say something about this.
i
( 25 MR. McLAUGHLIN: Actually, I think it would probably l
6 95 1 clear up some of Mr. Kerr's concerns if we talked about it.
l O i 2 But go ahead.
V 3 MR. WARD: Well, I think this is a useful model 1 4 mean, this is what you're really trying to do. But for the 5 third box, what you really want in there is, you know, 6 accidental radiological releases. There's now way to get that 7 from experience we're doing such a good job we don't have a 8 statistically significant number of those. There's no way to 9 get those at the present time from risk analysis because risk 10 analysis is too coarse of an art to be able to look at this 11 sort of thing.
12 So I think it's useful to keep this model in fr,ont 13 of you, but in order to use it in any way I think what you're
(^w 14 going to have to do is just say look, folks, at the present 15 time there's no way to really use the third box in this 16 model What we're going to do in the meantime is by expert 17 opinion, judgment, faith, voting, we're going to take certain 18 of those intermediate indicators and say we think that these 19 are the strongest safety indicators we have and we can't show 20 that they correlate with real final safety indicators, but on 21 the basis of voting and expert opinion this is what we're 22 going to use. And then you get on with the important business 23 of deciding what correlations you have between the first and 24 second boxes.
['N l ss 25 And if you feel strongly enough about the validity l
96 1 of your voting and expert opinion procedure, then that gives 2 you a basis for insisting that people adjust the resources in 3 the expectation that they will get better numbers in the I 4 second box.
5 1 mean I think it's a good approach, but you have to 6 keep the fact that you don't really have anything in the third 7 box transparent in front of you.
O MR. JANKOVICH: May I add something to this 9 comment? That is very true, and we have to look at our study 10 in the proper perspective. One is that it was the first 11 preliminary study. Nobody ever before even talked about 12 indicators. We were in complete darkness, and from the 18 darkness we picked these factors, and we found something.
v 14 And the other thing is we tried to put it on some 15 statistical basis. We also have to see the NUMARC proposed 16 set of indicators and they haven't been subjected to any kind 17 of statistical validation. That will have to come yet.
18 And that's where we are at the moment; at the early 19 phase of this indicator work, t 20 MR. REED: Well, in order to keep on schedule --
21 MR. KERR: I have just a short question. in some 22 contrast to what Mr. Ward is saying, I am not sure that these 23 are necessarily bad indicators, but I think if they are to be 24 good indicators, one needs to show some relationship between i 25 these and reactor risk.
i 1
97 1 Now, I have an idea that one could at least make a
[)
v 2 plausibility argument that these are indicators of human 3 error because if any of these things goes up it probably has 4 to do with human error in the operation and maintenance of 5 the plant. And it may well be that we're at a stage in pRA 6 and in plant operation that this is really the biggest 7 contributor to risks that the equipment is in pretty good 8 shape.
9 So I wouldn't completely abandon these, but I do 10 think that one needs to establish, at least on a plausibility 11 basis, that there is a relationship between these things and 12 the risk from core melt. And I don't see any effort to do 13 that, and it seems to me if you're going to continue to use
/^\
14 them you need to give some thought to that possibility.
15 MR. REED: Well, I think, Dr. Booher, you certainly 16 have discovered that the ACRS members have a great deal of 17 concern about your final safety indicators and the weighing of 18 those in final decisions in reports. I agree with essentially 19 what Dave Ward said. I really think that the only one of j 20 those final safety indicators, man rem exposure, may possibly
{
21 relate. Take tritium as an example, it can be related to 22 cladding, stainless versus zirc alloy, which now more or less J
23 everything is zirc alloy. It can be related to how you
( 24 recycle water, whether you throw it away or whether you plow
) 25 it back into your fuel tanks and your other tanks.
l l
98 1 I think they are really not very good indicators
/ \
2 unless you have a major accident, and that's not what we're V)
/
3 here to do with respect to maintenance. So you ought to have 4 a review of your final safety indicators, I think, and that is 5 the consensus of our judgment.
6 MR. GOOHER: I appreciate your comments, and I hope 7 you will have those in your letter. As far as the problem of S doing more with risk analysis, I don't believe we have 9 considered that in any depth, and I would appreciate any 10 guidance from ACRS on that 11 MR. REED: Okay. Thanks, Pete.
12 Next on the schedule is Ms. Koontz on the personnel f s, 13 qualifications issue. Now, that sounds like an interesting i i V 14 subject.
15 MS. KUNTZ: I am Jennifer Koontz with the Licensee 16 Qaullfications Branch, and I would like to talk about some 17 problems that we have been able to identify in the area of 18 maintenance personnel qualifications. I would like to add at 19 the outset here that the emphasis, as you know, since TMl as 20 far as personnel goes, has been on operating personnel and 21 members of the operating shift crew. So the NRC has not, 22 until now, been involved in any extensive investigations of 28 the issue of maintenance personnel qualification, selection, 24 training, what have you.
s 25 There has not been any regulatory emphasis on
99 1 maintenance personnel qualification, selection and training; 2 in fact, there hasn't been a regulatory emphasis on personnel
[/}
types of issues at all in the past. And so we are approaching 8
4 an area here that is relatively uninvestigated from an NRC 5 standpoint.
6 It merits investigation because of the problems that 7 we know about with human errors occuring during maintenance, 8 and with maintenance personnel being a group of people that 9 contribute significantly to not only maintenance errors but 10 operating errors as well And some of the problems that --
11 MR. KERR: Excuse me. I must confess to some 12 puzzlement at your statement. I have attended a number of 13 ACRS meetings in which we were involved in looking at license l T U 14 applications, and I have seen long and extended lists of 15 maintenance people and training programs for maintenance 16 people and selection processes for maintenance people. So 17 there must be somebody in the NRC that has been looking at 18 this.
19 MS. KOONTZ: It has been a fragmentary effort at 20 best, and as you know, during a licensing review the initial 21 staff of a facility, including the maintenance staff, is 22 reviewed as far as the assurance that the licensee or the 23 applicant has taken measures to provide adequately qualified 24 staffing and personnel for startup and operation of the plant.
25 1 guess what I mean to say is there has been no
F 100 1 research emphasis, there has been no systematic approach to a 2 determination of what should be the job-relevant types of a qualification and training specifically for maintenance 4 people.
G*
5 MR. JANKOVICH: When it comes to qualifications and 6 training of maintenance people, most of the commitments from 7 the utilities refer to standard ANS 3.1 Yes, it defines that 8 maintenance people must be properly qualified, and that's 9 about all it contains.
10 And our concern here in this area is the inadequacy 11 of the definition and the variability. Of course, we know 12 that unions are involved. They say they have qualified 18 people, but their criteria vary from state to state, too.
( 14 MS. KOONTZ: To start with, part of the problem is 15 that there is a lack of standard maintenance personnel job 16 classifications, and that is, at different plants maintenance 17 personnel are called different things. At one plant, you may
( 18 have personnel grouped under electrical, mechanical and 1&C j 19 department; at other plants you may have additional breakdowns l
20 of people into laborers, pipefitters, welders, et cetera. And 21 all these various and sundry titles have not been consistently i 22 used by the utilities, have not been consistently addressed by l 28 the NRC or by industry standards to date.
24 The most common classification that is currently 25 used for maintenance workers happens to be the electrical, the
101 1 mechanical and the l&C breakdown, and that is probably for 2 several reasons. But one reason is because that is the 3 classification that INPO uses in their, job task analysis data 4 base and in their evaluation of training programs for purposes 5 of accreditation.
6 So part of the problem has been that the population 7 essentially is not well defined. You can talk about 8 maintenance people but you are talking about everything from 9 the people who apply paints and coatings to surfaces to people 10 who at other facilities may be involved in fuel manipulations 11 and quite critical operations in t'h e day-to-day running of the 12 plants valves, pipes, electrical, ISC jobs and so forth.
~ 13 MR. REED: I can understand your listing, I think,
\'
14 of lack of standard maintenance job classifications as a 15 problem. But I think it's a problem that you can't face up to 16 and you can't do much about for a long period of time. It 17 seems t o me it's a problem that ought not to be one of focus 18 because as was just mentioned, you have union aspects, 19 management aspects and all kinds of aspects that impact upon 20 the number of differentiations or ranks that you have in the 21 maintenance group, and their job-allowed activity and all 22 these things.
23 1 really don't think that it's a major problem. I l 24 think that the fact that there isn't any standard s
25 classifications --
and I doubt that there ever will be --
is l
r 102 1 something that we should look --
we should look at other
/'} 2 things. Look at the ether things and focus on the other V
8 things to cure maintenance problems if there are problems.
4 MS. KDONTZ: That's true. This is not necessarily a 5 behavioral or a causative problem in terms of human errors, 6 but it's a problem from a regulatory perspective, or even from 7 an industry perspective, that tries to regulate itself, if you 8 will, through standards and standardization. And that's where 9 this problem comes into play. And that is that anything that 10 the industry tries to do, such as accreditation, is going to 11 run into implementation problems, if you will, because of the 12 various types of personnel that are involved, the way the 18 unions deal in job classifications, job titles and various O 14 levels of qualification within a job title, and then the way 15 that the NRC deals with various groups of people as well, 16 to i merely present that as being an implementation 17 type problem, when you think in terms of remedies or solutions 18 that may be applied to some of the maintenance personnel 19 issues.
20 The next one is probably 4 little more obvious s l
21 education and experience and training for maintenance 22 personnel vary widely. And this is for several reasons, 23 again. This is because from plant to plant, there is no 24 standard criterion other than occasional adherence to a
\ ,/ 25 regulatory guide which endorses a standard, an old standard, L
103 1 ANSI N18.1 1971, which addresses maintenance personnel 2 qualifications in a very cursory fashion.
3 Not only do standards and their application differ 4 from plant to plant, but the plant-specific training programs 5 vary widely as well. And this is something obviously that 6 INPO is trying to grapple with through their accreditation 7 program.
8 part of the problem inherent, however, is that the 9 training programs as implemented today in the nuclear industry 10 are plant-specific and as such, they often do not cover --
or 11 they often are not participated in by a large proportion of 12 plant maintenance personnel who are hired under contract or as 7, 18 temporary workers during an outage, for example.
14 So the problem of education, experience and training 15 for maintenance personnel is one, again, from an 16 implementation standpoint that presents a difficulty in terms 17 of implementing standards criteria or possibly any kind of 18 regulatory guidance because the training programs vary 19 widely.
20 The apprentice programs, which the craft personnel 21 receive, are generally not nuclear programs at all An 22 electrician comes up through an electrical craft apprentice 23 program, which is not a nuclear training program, it's not a 24 nuclear experience program. There is no nationally-sanctioned I craft apprenticeship training program that is g ,/ 25 I
E
, r 104 1 nuclear-specific.
2 MR. REED: Do you think there should be 8 nuclear-specific training for craft work? I be l l eI/e we had a 4 report here some months ago by a number of people -- I wish I 5 could focus on who they were -- where somebody said that 80 6 percent of some of these jobs, such as becomi rig a p i pef i t t er 7 or a welder or so on -- really, it's the same whether it's e fossile or nuclear, that there isn't much difference with 9 respect to nuclear when you get down to craft work.
10 MS. KOONTZ: That is a popular argument and it's 11 true that the actual hands-on manipulation of the tools and 12 the equipment may be similar from a fossile to a nuclear 18 station, but I think that the nuclear-specific training comes O 14 in in terms of the emphasis on systems, the emphasis on 15 importance to safety and safeguards that the nuclear 16 maintenance workers, I think, need to be aware of, and need to 17 be aware of the implications of their work and the interface 19 of their work with plant operations. And that is certainly 19 different at a nuclear plant than it is at a fossile plant, 20 from a safety standpoint.
21 MR. REED: I agree with you on that. And that maybe 22 is where the 20 percent comes in. Certainly, -- and this 28 doesn't have to do with a craft skill -- there certainly is a 24 difference in the honesty that the person should-displays the 25 integrity that he should display, the recognition of the i
105 1 safety consequences that he should be trained in and display.
("N 2 These are the different things, and I think they
\vI 3 relate a little bit more closely to personality, and we'll get 4 to that later.
5 MS. KOONTZ: Well, there are a lot of different 6 kinds of knowledges, skills and abilities involved. I don't 7 know how I would classify those couple of things that you just 8 mentioned, but they might come under interpersonal 1
9 communications or communicative skill or ability, somehow.
10 And certainly, these types of things are related to certain
!?
I 11 types of tasks that the workers must perform.
l 12 So the point is, I guess, if the problem is one of 13 variability, as I'm trying to point out here, then maybe we G
s- 14 need to take a systematic look in order to determine what are 15 the requisite kinds of knowledges, skills, abilities including 16 communications, interpersonal skills, dexterity and special s
17 abilities. But they must be shown to be job-relevant in order 18 for, number one, the industry and the workers to accept them 19 and in order for them to be legally defensible, and in order 20 for a regulatory strategy, if any could be developed to apply 21 standardization.
22 MR. KERR As I listened to you and Mr. Reed, the 23 impression I get is that nuclear welders must be better 24 welders than non-nuclear welders, and that is the principal y ) 25 difference.
106 1 MS. KOONTZ: Well, those are a special class 2 because, of course, they are certified to their job. But 1 3 think that a generalization can be made there.
4 MR. KERR: Well, that means to me then that the 5 nuclear welders don't have to know a lot about radiation or 6 nuclear. power, that they just need to be better welders. Is 7 that the sense that I get from this?
8 MS. KOONTZ: I think it could be shown from an 9 analysis of the job tasks that the nuclear welder must do that 10 there would be certain components of nuclear systems 11 knowledge, I will call it, or nuclear engineering type of 12 knowledge, that the nuclear 'valder must apply to do his job 13 safely and efficiently that the welder in a fossile plant l
'"# 14 might not need to know. And likewise, the welder in a-fossile 15 plant may certainly need to know certain things to do his job 16 safely and efficiently that the welder in a nuclear plant 17 wouldn't. Specialized types of things.
18 MR. KERR: You're telling me you think one could 19 show that, but at this point, you haven't.
20 MS. KOONTZ: We haven't. It can be done; it's a-21 matter of establishing what we call criterion reference 22 validity which, as you may be aware, has been done. We have 28 made a lot of progress in it for members of the operating 24 shift crew by using job task analysis data. The same thing 25 could be done for maintenance personnel by using job task
107 1 analysis data that currently exists.
[ 2 MR. BOOHER: Could I make a comment? When i 3 .sttended the INp0 maintenance superintendents
- workshop there 4 was a consensus that if anyone had to work in a nuclear plant 5 or a fossile plant doing the same job, there's no way they 6 wouldn't select the fossile plant. The level of frustration 7 is maybe another characteristic that needs to be placed in 8 here in terms of the work environnment. And for the same pay, 9 it's a much easier job, according to them, to work in the 10 fossile plant on the same equipment.
11 MR. KERR: That's another way of saying that the 12 welders in the nuclear plant must be much better welders they 18 must be more careful and do a better job.
V 14 MR. BOOHER: Well, if you give them a performance 15 test you may not be able to show up a difference, but in the 16 work environment over the long term, it would certainly show.
17 MR. KERR: And so I'm trying to look for this 18 nuclear training that the welder should have and all 1*ve l 19 heard so far is that the quality of his work must be better.
20 Must be demonstrably better.
21 MS. KOONTZ: I think the answer to your question, 22 though, lies in a very large database that INp0 has created
! 28 for the jobs of electrical, mechanical and l&C maintenance f
24 personnel, and I think you would have to do a systematic
[
( 25 analysis to determine what kinds of knowledges, skills and l
r
108 1 abilities the nuclear worker must have that the fossile worker
/~N . 2 must not have. But that's not really the point.
8 MR. KERR: Systematic analyses are fairly good if 4 you know waht you're looking for. But if you don't know what 5 you're looking for, you can do systematic analyses that will 6 turn out to be very unsystematic.
7 MS, KOONTZ: You're telling me that I need to look 8 for the things that the nuclear worker must know or have that 9 the non-nuclear worker need not.
10 MR. KERR: I'm saying it seems to me it's worth-11 giving some thought to this before you start doing systematic 12 analyses; otherwise, the systematic analysis may not tell you 18 very much.
O)
+
\- /
_ 14 MR. BOOHER: I don't think it wo*Jid be very 15 systematic if we didn't do what you said in the beginning.
16 MS. KOONTZ: Well, another part of the problem here 17 is the fact that the knowledges, skills and abilities of 18 maintenance personnel are not well documented at the 19 practicing level, and that is, plants do not currently 20 undertake a systematic analysis or evaluation -- at least all 21 plants do not -- of the knowledges, skills and abilities of 22 their incoming maintenance personnel 28 So while they may have a criterion for the hiring of 24 a person with a number of years of experience which brings 25 them up to the journey level of qualification within their job i
1
109 1 classification, there is no way at present to systematically
[~N 2 identify the knowledges, skills and abilities of the job s
b 3 incumbents in the nuclear industry, and the problem is 4 exacerbated by the fact that up to two-thirds of the 5 maintenance work force are temporary workers or contract 6 persennel, and there's no way that we currently know of of 7 identifying, much less documenting, the knowledges, skills and 8 abilities in terms of education, experience and training for 9 that large proportion of the work force.
10 They are not covered necessarily by unions, they are 11 not covered by our current regulatory strategy, which is 12 meager in this case anyway, and a lot of times, they are not IS even covered by the plant standards themselves, by the plant t
14 selection criteria that they apply to their inhouse personnel.
15 MR. KERR: What do you know about the relative 16 skills of temporary workers as compared to the permanent? Do 17 they tend to be l o w - s k i l' l e d people, or is there a distinction?
18 MS. KOONTZ: We don't know.
19 MR. WARD: Earlier there was an observation that 20 plants with small maintenance departments, small numbers of 21 personnel in the maintenance department, seemed to have a 22 better maintenance performance record. I mean, can I conclude 28 from that that if the maintenance department is small that 24 they use a lot of extra personnel, contract personnel? And if r"
()/ 25 that's true, then I might conclude that the contract personnel s
110 1 must do a better job -- well, you know, that's a little f"% 2 simpleminded, but I have been driven to that conclusion.
U 8 I wonder about that.
4 MR. REED: David Ward, you are beginning to conclude 5 very well I think there are some mixed-up statements around 6 here. I would like to go back to one where it was said that 7 maintenance personnel rather prefer to work in fossile plants 8 than in nuclear plants. And let me point out what I really 9 think the reason for that might be.
10 The reason for that might be that if you work in a 11 nuclear plant, you are held to performance standards, and 12 peopie don't like to be inspected and inspected for the numbe-18 of slag inclusions in their welds. And that's the reason they
/
t
\m- 14 don't like to work in nuclear plants, is because they get 15 inspected several times over and they can throw a weld in at a 16 fossile plant and get away with it, even if the slag inclusion 17 is there.
18 So really, if the fossile plants' standards were the 19 s a:ne for a weld as they are at a nuclear plant, you would find 20 that they prefer to work at a nuclear plant. So there are 21 some subtle things here that go beneath the surface.
22 Quite frankly, I thought that what David Ward just 23 concluded was under the surface and you have to look for these 24 subtleties.
(a 25 MR. JANKOVICH: it is not that funny, actually. If
111 1 I may go back to our Japanese study, some of the Japanese
/"'s 2 utilities have only 50 staff people in maintenance at the l 6
%. J S site. The rest of the maintenance work is done by 4 contractors, and the contractor people are highly qualified 5 with double quality control checks.
6 MR. REED: Yes, but we're talking about the USA when 7 we talk about the preference.
8 MR. JANKOVICH: So this argument can be carried out 9 and can be valid.
10 MS. KOONTZ: I think it would be pretty hard to draw 11 a firm conclusion about the relationship between a staff size 12 and plant performance because of a lot of mitigating factors, 13 including the way in which you count and measure your worker b
> 4
\wd 14 exposure, for example, and whether you include your contractor 15 personnel in that count or whether they are reported 16 separately, and so forth.
17 MR. REED: (Jell, I want to point out that in i-18 U.S. plants that I'm familiar with where they are small staff 19 size for maintenance -- and I know two or three of them --
20 they do use a lot of contract personnel, and these contract 21 personnel are of very, very high quality. And if they are 22 reptitiously cycled into this paint from a curtain contractor 28 who is quite normally selected, these people are used to the j 24 kind of inspections and standards that they must meet to put a
) 25 weld in, in a nuclear plant, u
112 1 Since we're running slow on time, let's hit that
h 2 selection bullet for a minute.
{J 8 MS. KOONTZ: Okay, Well, the problem is fairly 4 straightforward, and as I mentioned before, selection and 5 screening criteria have not been regulated in the past for 6 maintenance personnel, nor for operations personnel NRC just 7 has not been in that business.
8 From an industry standpoint, there are, we know, 9 utilities out there who apply very stringent selection and 10 screening criteria as a matter of practice for all their 11 personnel But again, the variabi'lity comes in between plant 12 personnel and contractor personnel, and a variability also 13 comes in from plant to plant. Because we also know that there O
-' 14 are plants out there who do not apply selection and screening 15 criteria to the hiring of their maintenance work force.
16 MR. REED: Will you attend this afternoon's session 17 on selection of personnel?
IS MS. KOONTZ: Yes.
19 MR. REED: Very good. I'm glad you'll be there.
20 MS. KOONTZ: So again, the problem is one of 21 variability; the problem is an area that has not been 22 addressed through regulatory strategy in the past, And in 28 fact, it's one that is kind of falling between the cracks, 24 even in the industry improvement programs today. Because as
(_ 25 you know, INPO is moving to accredited training programs for
113 1 electrical, mechanical and l&C types maintenance personnel
(~' 2 But INPO's accreditation does not address personnel selection
(
3 and screening. It addresses only in-plant training programs, 4 and it will look to see if the utility has a selection and 5 screening program that they use to feed people into their 6 training, just in order to insure that it's carried throughout 7 the training program. But they will not evaluate whatever 8 that selection and screening practice might be. And they will 9 not evaluate if there is none.
J 10 MR. REED: Do you think that any person can be 11 trained to do anything? And as an example, do you think a 12 person with two percentile mechanical comprehension can be 18 trained, no matter how widely the training program varies, can
\- 14 be trained to be a good maintenance worker?
15 MS. KOONTZ: Well, that's an empirical question, 16 that's for sure.
17 MR. REED: l'Il ask you again at 5:00 o' clock this 18 afternoon.
19 MS. KOONTZ: Well, I think that we need to determine 20 the answers to such questions by seeking to find out what are l
21 the job-relevant knowledges, skills and abilities for 22 maintenance personnel And i emphasize that we need to find 28 out the job-relevant knowledges, skills and abilities in order 24 that we would subsequently have a legally defensible selection
( 25 and screening standard.
l l
< 114 1 As you know, other industries, and airlines in f' 'h 2 particular, have come under a lot of scrutiny and have faced I (v) 3 legal recourse because of the need to defend on a criterion !
4 reference validity basis any selection and screening practices 1
5 that they adopt and that they impose via regulation.
6 So we want to be very careful to adopt a valid 7 strategy.
8 MR. REED: When you say "we" and you talk strategy, 9 are you aware of the POSS, the eel pOSS tests?
10 MS. KOONTZ: I am aware that there are some tests 11 that are used in the industry. I am,not familiar with the eel 12 methodology, nor with the methodologies used by General 13 Physics and some of the other contractors in theindustry, to n
A- 14 administer what they use as selection and screening 15 instruments.
16 MR. REED: Well, the eel tests are validated tests, 17 the POSS tests.
18 MS. KOONTZ: Well, there are different types of 19 validity, and they may be validated in a number of ways. They 20 may be validated against a population sample, they may be 21 validated against some type of performance measure. But i 22 would like to assert that since the job task analytic data has 28 only recently come available through INPO's extensive effort, 24 I think it's necessary to validate using such data, any kind
) 25 of selection and screening mechanisms, that the industry would w./
115 l 1 hope to promote or use or adopt as a standard practice.
/~% 2 MR. KERR: The next question coming up shortly is:
v' S do you believe in inherent mechanical ability. But I'm not 4 going to ask it.
5 MS. KOONTZ: That's a dangerous question to ask a 6 psychologist, you know.
7 Well, just very quickly to sum up what I think the 8 need is in this area, --
and I hope we do get into this this 9 afternoon a little bit --
I think that we do need to determine 10 the knowledges, skills and abilities required for maintenance 11 job tasks. And again, this can be done using job task 12 analytic data and other means of systematic analysis.
18 We need to then identify the relevant sources of the
\~- 14 knowledges, skills and abilities in terms of education, 15 training and apprenticeship programs and other sources of 16 abilities.
17 MR. KERR: Where do genes come into this?
l 18 MS. KOONTZ: That's why I said "and other sources" 19 of abilities which can be measured. And then we need to 20 conduct an analysis of the applicable industry practices and 21 guidelines and whatever eel is doing and so forth, in order 22 to, again, come up with a legally defensible strategy for 23 ensuring not only the right qualifications but the standard 24 qualifications, if that's the way the industry feels they need 25 to go, f
- - - . - - - - . . , , . - - - - . . . - . , _ . ,,_-n-_ . , . , ,-,--.--.n. . , , r - - - - , - ,
116 1 MR. WYLIE: Does this represent the objective of
-(~ 2 your work?
S MS. KOONTZ: Thees would be the objectives if 1 4 approved, that a project under phase I of the maintenance and 5 surveillance program would attempt to meet. Yes.
6 As you know, we have done similar analyses for 7 operating personnel in the past couple of years which have 8 been successful and have provided input into not only 9 standards development, but licensed operator examination 10 improvements as well And I would hope that similar strides 11 could be made toward promoting the appropriate knowledges, i 12 skills and abilities for maintenance personnel and promoting
~
18 standardization of selection, screening, education, experience
\s 14 and training in the industry so that that maintenance work 15 force is the best that it can be.
16 MR. WYLIE: I would assume that the end product of 17 it is basically a survey of roporting of these needs, and then l
18 a recommendation?
19 MS. KOONTZ The end product would be a 20 recommendation for, I think, how far the industry currently is 21 from where it should be, in terms of ensuring the right kinds 22 of education, experience and training for maintenance 28 personnel.
24 So whether that recommendation would be for 25 increased use of standards, increased use of plant-specific
I 1
117 i 1
1 selection mechanisms or increased regulatory emphasis, you )
2 know, I can't say. But the input into any of those would be O1 3 the same, and that would be identified job-relevant 4 knowledges, skills and abilities and where they should 5 appropriately come from for different types of maintenance 6 workers.
7 MR. WYLIE: Well, are you going to limit your 8 activity to knowledges, skills and abilities?
9 MS. KOONTZ: Knowledges, skills and abilities, those 10 are the buzz words for the "right stuff" to do the job.
11 MR. WYLIE: And that includes natural ability.
12 MS. KOONTZ: Yes, sir.
13 MR. REED: Okay. Is that the end of your
- (
14 presentation?
15 MS. KOONTZ: Yes.
16 MR. WARD: I have a question. Jennifer, a staff 17 member just handed me a couple of reports here from 1982 from 18 Bore NL where some of this work seems to have been done, but 19 Job analysis of maintenance mechanic position for the nuclear
, 20 power plant maintenance personnel reliability model -- is that 21 useful?
22 MS. KOONTZ: Yes, there are four reports in that 28 series which involved a limited job analysis which was done 24 for development of the MAppS models that is, the maintenance
) 25 personnel performance simulation model it's useful it is
118 1 extremely limited because the data was collected on a limited 2 sample of the maintenance work force population and because N_/
8 the data, as we have it in a couple of reports that resulted 4 from the project, are very, very synthesized.
5 And that is, that the task listing that they include 6 in those documents have been very, very boiled down. So you l
7 don't have a comprehensive set of what the job actually 8 consists of. It was somebody's Idea of a boiled-down set of 9 what the jobs actually consist of. It could be used for a 10 study of the type that's needed, but the INPO database is 11 much, much more comprehensive and would be a logical starting 12 point. This data would serve as a point of reliability check ,
18 and verification of one database against the other.
(I g
We can conclude that the study you 14 MR. KERR:
15 propose would be a boiled-up version rather than boiled down?
16 MS. KDONTZ: pardon me?
17 [ Laughter.]
18 MR. KERR: I don't think it should be repeated.
19 MR. REED: Let me add to and say it differently than 20 what Dr. Kerr is saying. I am concerned, when you use the 21 word " limited" and you talk about more work and more research 22 and more studies and redo the POSS tests because they aren't 28 validated against the latest INPO task analysis and so on and 24 so forth, I believe that many, many years ago Admiral Rickover O 25 said something about building a submarine. He said, Well, if A . . . _
119 1 we continue to do research we will never build a submarine; fx 2 there can always be research and this, that and the other.
3 And I think this may apply here.
4 1 think we should beware that you can research 5 things to death; in the meantime the core melts down, or the 6 maintenance work doesn't get done. It seems t o me that there 7 are cutoff points that may be limited analyses, but are 99 8 percent on target with respect to the issue.
9 MS. KOONTZ: Yes. I wouldn't propose that we redo 10 any of the analyses that have already been done, but rather, 11 use those as a starting point for identification of the 12 appropriate categories of knowledges, skills and abilities.
13 And that's what we would do.
O%_/ 14 MR. KERR: I think it was Admiral Rickover also who 15 said, " Damn the torpedoes, full speed ahead."
16 ELaughter.]
~
17 MR. REED: Okay. Thanks very much, Ms. Koontz.
18 Now, we should revie a bit what we want to have 19 brought in to the full committee meeting, among ourselves.
20 MR. WARD: Could I make one comment before we 21 finish, to Mr. McLaughlin on the final intermediate safety 22 indicators.
28 i didn't mean to be excessively negative about that 24 model I mean, I think that's a good way to look at it, What 25 1 might suggest is that if you develop some sort of a
[Vh
[
120 1 correlation between what you call the final indicators, which
[s 2 really aren't in the intermediate indicators or between that 8 and the resources, that that might be one input into the 4 judgments you make about what intermediate indicators to use.
5 You see what I mean? But I would regard that only 6 as one input. I mean. I think you're going to have to have 7 something practical to work at; you're going to have to decide 8 on the basis of judgment which of those intermediate 9 indicators is of most value, j
. 10 And one input to that might be how they correlated 11 with your routine release indicators. But I would say that is 4
12 only one input.
i 18 MR. REED: 1 might point out to the committee N 14 members that we got scheduled for one and one-half hours for 15 this subcommittee report at the SOSrd meeting in July. Now, 16 the question is one and one-half hours seems like a lot of 17 time with respect to the maintenance / surveillance program 18 plan.
19 What should we do? Should we have a number of 20 people come in from the staff?
21 MR. KERR The first thing we should do is cut that 22 down to a half an hour.
28 MR. REED: You would vote for a half hour. I think l 24 it's long. Quite frankly, there was a statement made -- and i 25 think this is representative of the maintenance / surveillance
121 1 program plan -- that there's a bit of a " wait and see."
2 There's waiting for NUMARC, there's waiting for INPO, there's
\.
8 a lot of study and review and work activity going on. But as 4 far as major milestones, there aren't that many.
5 MR. KERR: I would suggest really a brief 6 presentation by Mr. Booher. We have heard what they're 7 doing. It seems to me they're headed in a reasonable 8 direction and we can report this to the committee.
9 MR. WYLIE: I would think that the way to approach 10 it is to have Dr. Booher make an overall presentation on the 11 program, and it seems t o me that could be effectively done 12 with all the slides that have been presented today.
18 MR. KERR: You mean with an astute selection of O
( 14 them, not all of them.
15 MR. WYLIE: Oh yes, sure.
16 MR. REED: Well, I vote for that, so that's three to 17 one, Dave, unless you want to vote for it, too.
18 MR. WARD: I think you might need a little more than 19 half an hour.
20 MR. WYLIE: Well, I would think you may, once they 21 get to asking questions.
2 2. MR. REED: Well, I think i see it that way.
28 Dr. Booher, why don't we think about 45 minutes at the July 24 meeting.
25 MR. BOOHER: What's the date of that meeting?
122 1 MR. ALDERMAN: The actual date hasn't been selected O 2 yet.
8 MR. WARD: Well, you know it's going to be Thursday 4 or Friday. It will be either the lith or 12th of July.
5 MR. BOOHER: Okay, I can be available then.
6 MR. REED: Now as far as the subjects, it seems to 7 me the full committee would want to hear the plan pretty much, 8 and I guess there are the phases, the plans, the major 9 breakdowns, the indicators. I think the full committee would to be quite interested in what you can say on indicators at this 11 point in time. I believe that the qualification issue will be 12 interesting simply because of this afternoon's meeting that's fg 18 going to take place and this issue of selection versus
(
14 training is going to come up hot, I think.
15 MR. BOOHER: la there any proposal to have the 16 industry brief you on their indicators and other work in the 17 near future?
18 MR. REED: Not that we know of.
19 MR. BOOHER: Because as you know, we are, as you 20 say, in a " wait and see" i n a lot of these areas with the 21 industry.
22 MR. WYLIE: That might be worthwhile.
28 MR. REED: Well, it's not something that we're going
! 24 to get into at the 808rd meeting, is it?
s 25 MR. KERR: I think in your presentation you simply
..-- - _ _ _ , _ _ . __ . .,._--.-_-..,___m_.. -
..- m.-, ,
J 123 1 present them and say that at least on a preliminary basis they
[ 2 seem reasonable or unreasonable, as you think.
V) 8 MR. REED: Herman, have you got anything?
4 MR. ALDERMAN: No.
5 MR. RFED: 45 minutes total for the maintenance 6 issue. I won't have much to say.
7 All right. I wish to thank the staff very much, and 8 a word of advice. I hate to do this, close with a word of 9 advice, but maintenance is a people issue, and I don't know 10 how it is that we sometimes get maintenance workers to say 11 that they are being treated as " c a'i l me stupid" people. We 12 must think about that.
- 18 [Whereupon, at 12:06 p.m., the meeting was 14 adjourned.]
15 16 17 1
18 19 20 21 22 23 24 25 I
4
, - - . - -...v, m--e. - , , _ , - , . _ . _ _ _ . _ _ _ . . _ _ _ , _ _ _ _ _ _ _ . - _ _ _ _ . _ , _ , _ _ _ _ _ _ _ _ _ _ _ , _-_-.,,y.....,,,,-y. , . _ _ - _ _
l O 1 CERTIFICATE OF OFFICIAL REPORTER 2
3 4
5 This is to certify that the attached proceedings 6 before the United States Nuclear Regulatory Commission in the 7 matter of: Advisory Committee on Reactor Safeguards e
9 Name of Proceeding: Subcommittee on Maintenance Practices and Procedures to 11 Occket No.
O 12 Place: Washington, D. C.
la Cate: Tuesday, June 18, 1985 14 15 were held as herein appears and that this is the original 16 transcript thereof for the file.of the United States Nuclear 17 Regulatory Commission.
18 /
(Signature) ,
,g ,
(TypedName'6fRegrter) ,/'u S zaqple B . Tedng 20 21 22 23 Ann Riley 4 Associates. Ltd.
24 25
/YWbi ,;\
E i
l 4
+
l NUCLEAR. REGULATORY COMMISSION
!. - MAINTENANCE AND SURVEILLANCE PROGRAM 4
O PROGRAM MANAGER DR. HAROLD R. B00HER, CHIEF LICENSEE QUALIFICATIONS BRANCH DIVISION OF HUMAN FACTORS SAFETY /NRR GREGORY C. CWALINA, SECTION LEADER MAINTENANCE / SURVEILLANCE SECTION O
r 4
NRC SPECIFIC OBJECTIVES DETERMINE EFFECTIVENESS OF CURRENT MAINTENANCE PROGRAMS t IDENTIFY PRACTICES WHICH REDUCE HUMAN ERROR RATE IN PERFORMANCE OF MAINTENANCE O IMPROVE EFFECTIVENESS OF MAINTENANCE PROGRAMS IN ASSURING OPERABILITY OF SAFETY SYSTEMS REDUCE UNNECESSARY AND UNANTICIPATED RADIOLOGICAL EXPOSURE TO MAINTENANCE PERSONNEL DETERMINE REGULATORY APPROACH TO ASSURE EFFECTIVE i MAINTENANCE PERFORMANCE i
j
MAINTENANCE PROGRAM DEVELOPMENT NOV 83 NRC MAINTENANCE WORKSHOP f MAINTENANCE INDICATOR PILOT STUDY INITIATED JAN 84 COMMISSION POLICY AND PLANNING GUIDANCE MAY 84 US/ JAPANESE STUDY PART I COMPLETED ACRS BRIEFED (MAINTENANCE SUBCOMMITTEE AND FULL COMMITTEE JUN 84 DRAFT PLAN TO NUMARC ASME BRIEFED PUC BRIEFED JUL 84 IEEE BRIEFED
. SEP 84 ANSI BRIEFED OCT 84 UPDATED PLAN TO NUMARC ACRS - MAINTENANCE SUBCOMMITTEE - JAPANESE STUDY BRIEFING DEC 84 MAINTENANCE INDICATOR PILOT STUDY COMPLETED JAN 85 PLAN PHASE I APPROVED NUMARC PROPOSED INDICATORS NRC INDICATOR TASK FORCE FORMED MAR 85 REVISED MSPP (NUMARC COMMENTS)
APR 85 SUBMITTED TO COMMISSION (SECY-85-129)
MAY 85 SURVEY PROJECT INITIATED O JUN 85 COMPLETED REGIONAL BRIEFING / COORDINATION ACRS - MAINTENANCE SUBCOMMITTEE
O PROGRESS f 3
t MAINTENANCE SZCTION FORMED 4
TASK FORCE REVIEW NUMARC DRAFT INDICATORS 4
PARTICIPATION NEW STANDARDS EFFORTS REVIEW JAPANESE MAINTENANCE /0PERATIONS EXPERIENCE i
PARTICIPATION IE INSPECTION j
Q FEEDBACK NRC REGIONAL / RESIDENT INSPECTORS l
5 i
i lO l
l P
! A+3 ;
i i
, i i :
i t L@
l l
l i
b t
l-i' !
I' ;
1 1 l
+
b !
L
! MAINTENANCE AND SURVEILLANCE PROGRAM PLAN I
i
! JUNE 18, 1985 1-i i
i i, JOHN P. JLNK0VICH-i MSS /LQB/DHFS/NRR X24811 i
l t
i t
i 1
1 4
i 1
1
- - - - - - ~ _ . _ - . . _ . _ _ _ _ . _ _
O ,
MSPP SCOPE
)
i
- ALL ASPECTS REQUIRED TO CARRY OUT A SYSTEMATIC MAINTENANCE
. PROGRAM i
SURVEILLANCE AND TEST ACTIVITIES I
EQUIPMENT REMOVAL FROM/ RETURN TO SERVICE
- O POST-MAINTENANCE TESTING l - MAINTENANCE MANAGEMENT / ADMIN, CONTROL
- PERSONNEL SELECTION, QUALIFICATIONS, TRAINING i
PROCEDURES DOCUMENTATION l
! THOSE COMPONENTS WHICH AFFECT PERFORMANCE OF SAFETY SYSTEMS i
l O
O ,
MAINTENANCE MAINTENANCE ENGINEERING MAINTENANCE MANAGEMENT MAINTENANCE EQUIPMENT, SELECTION AND TOOLS AND FACILITIES TRAINING ALARA SPARE PARTS PERSONNEL MAINTENANCE PLANT / EQUIPMENT DESIGN CERTIFICATION
\
CORRECTIVE PREVENTIVE TECHNICAL QUALITY DATA CONTROL
\qPRODUC CONFIGURATION PLANNING AND MANAGEMENT SCHEDULING
'10R
- K MAlflTENANCE mal'1TENN!CE C0'lTP.ol AUT!:0RICATIO';
PROCEDURES FIGURE 1: A Pictorial Representation of Maintenance
STRATEGY ,
BROAD SCOPE .-
FOCUS ON TECHNICAL ISSUES USE PHASED APPROACH INTEGRATE STAFF ACTIVITIES COORDINATE INDUSTRY INITIATIVES
M ,
O '
TECHNICAL ISSUES l 1. HUMAN ERROR
- 2. INDICATORS OR MEASURES OF MAINTENANCE EFFECTIVENESS -- IS i INP0 EFFECTIVE 7
- 3. ROLE OF PREVENTIVE MAINTENANCE IN COUNTERING AGING EFFECTS f
i 4, MANAGEMENT AND ORGANIZATION IMPACTS ON MAINTENANCE -
- EFFECTIVENESS
! s
) 1 4 5. MAINTENANCE PROGRAM CRITERIA AND STANDARDS i
i 6, INTERFACE WITH OPERATIONS t
l I
l lO I
4
PRMRN4 PHASES O PROGRAM ) -
INITIATION / A YEAR 1 YEAR 2 . YEAR 3
. l QUARTER 1 2 3 4 1 2 3 4 1 2 3 4 Phase I Survey anti I
Evaluation A A l l EDO Review A A !
Phase II Identification '
I of Problems and Impacts lA A A-
i e i 2
l I I
~
Industry Efforts l l l i 1-NUMARC A A' .A A STANDARDS GROUPS n
e 9
6 i .
O
a s .- 2-__ _ - ,a_._4- a__., -a..,..- e m* .-- A. - - _. _- , a~=
4 .
0 .
(
PHASE 1 PROJECTS
- 1. SURVEY OF CURRENT PRACTICES
- 2. MAINTENANCE PERFORMANCE INDICATORS
- 3. MONITOR INDUSTRY ACTIVITIES
- 4. PARTICIPATE IN STANDARDS GROUP
- 5. PROGRAM INTEGRATION (NRC AND INDUSTRY)
- 6. ANALYSIS OF JAPANESE /U.S. MAINTENANCE PROGRAMS
- 7. MAINTENANCE PERSONNEL QUALIFICATIONS
- 8. H.F. IN IN-SERVICE INSPECTION
- 9. HUMAN ERROR IN EVENTS INVOLVING WRONG UNIT OR WRONG TRAIN (GENERIC ISSUE 102)
O
. . --. . .- _ __ _-. . _ . . . - . = . -. _ . _ _ - . . _ - .
l O . ,
, 2. MAINTENANCE PERFORMANCE INDICATORS
- SCOPE:
- MONITOR INDUSTRY INDICATORS
- DEVELOP NRC INDICATORS IF WARRANTED MILESTONES:
PILOT STUDY, PNL, JAN - APR 1984 END0RSEMENT OF NUMARC INDICATORS, MAY 1985
( })
- VERIFICATION OF INDUSTRY DATA DECISION ON NEED FOR NRC INDICATORS O
O ABNOR'iAL OCCURREllCES AINAL !!UCLEAR 00HER PLANTS l.
1975 - 1934
, (SOURCE: fluREG-0090) 10- TOTAL ,
9-8 +
- 0 .
M7-W E
86-a 4
A E5- + +
!2 .b's *
/ s. . 'MIllTEIRflCE 4- A----As p M'y 'e RELATED u.
o . \. ./ * '"
,5
'\ / N. /
jp 3 -
P
\-
- c _ _ /
/
\.y. ',
. */,.
/.
2- g, 8--- 7 , -.-.-.e $-- - d F CIENCY s . - s
/'s
~~' /
/
~'. ' '
- i. s[-.g
- s. /
i % ---A, s .
/
-4~,,,~~
s%
0 %. h --- '*--- I I I I I I I a i r i 75 76 77 78 79 80 81 82 83 84 YEAR e
l.
.i j sde4
- SD'84 1, sg, s YEARLY TRIP RITES i 50'83 ' i sD'83 t
10 - l' l' ' ' '
@ r j 9 -
l I I i l is el $ '
/
! 8 _ l i / \ ~3i wt 5W I i
's E l l j El .
7
/ b- g]
l 4 ,'
I t
g E 6 -
3
/
lN/D I
! 6
\ 'I I
,I984 u \
l \
l
$' S - -
t R$
- i P~~% jl983 4 - -
l i
'k \* /
\
/ i l
\
m l t #
- i ss i l 5 I t 3 -
8 i 'd *
~
l 2 -
1 l
l I
1 l
I l
\,'
kN b
/ 8 l
j .
I l
i l
's /
/
l g I Ks 8/
0 ,
, , y i
e i i :i s
V,.MN./ /
O i I I I 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 d
NU'mF.R OF TRIPS -
4 1
() ,
l
- 3. MONITOR INDUSTRY ACTIVITIES SCOPE:
SUMMARIZE FINDINGS ASSESS APPLICABILITY DOCUMENT ACHIEVEMENTS MILESTONES:
C0 ORDINATION WITH NUMARC JUNE, OCTOBER 1984
[)
JAN, FEBRUARY 1985 ATTEND EPRI SEMINARS, MARCH, APRIL 1985 1
REVIEW EPRI/MIT REPORT, JUNE 1985 O
. .--.-*.,,---e-, , ,,.m._- - - - ~ , - - , ., ._--.m-.-, , , . - , ,m.,~,. , + _ _ . _ .
- 4. PARTICIPATE IN STANDARD GROUPS SCOPE:
ENC 0VRASE INDUSTRY INITIATIVES PROVIDE NRC CONTRIBUTIONS MILESTONES:
i -
ANS 3,9 SEPTEMBER 1984 - SUBCOMMITTEE FORMED MARCH 1985 -
OUTLINE CIRCULATED
[}
FALL 1985 -
NUMARC DRAFT DUE ASME OPERATIONS AND MAINTENANCE COMMITTEE APRIL 1985 -
NO PM STANDARD SPRING 1985 -
SEMINARS COMPONENT STANDARDS DEVELOPMENT CONTINUES IEEE W.G. 3.3 -
" PRACTICES" OCTOBER 1984 -
W.G. FORMED SPRING 1985 -
SCOPE DEFINITION
[)
t
O ,
- 5. PROGRAM INTEGRATION SCOPE:
IDENTIFY RELATED PROGRAMS INTERPRET CONTENTS AND SCHEDULES PREVENT OVERLAP PRESENT UNIFIED POSTURE MILESTONES:
O esosaeSS ace 0aTS. FINAL ase0aTS Aae seING REVIEWED l
f f
O f
O
\> ,
l NRC PROGRAMS RELATED TO MAINTENANCE RESPONSIBLE ORGANIZATION QUALITY ASSURANCE PROGRAM, R.G. 1.33 IE SYSTEMS IMPORTANT TO SAFETY IE SAFETY IMPLICATIONS OF CONTROL SYSTEMS (USI A-47) NRR COMPREHENSIVE REEVALUATION OF STANDARD 4 TECHNICAL SPECIFICATIONS RES l SUP.VEILLANCE AND TEST REQUIREMENTS
(]) (ECCS OUTAGE CRITERIA) RES ,
NUCLEAR PLANT AGING RESEARCH RES EFFECTIVENESS OF INDUSTRY ALARA PROGRAMS NRR EQUIPMENT QUALIFICATION - R.G. 1.89 NRR l RELIABILITY RESEARCH RES IMPROVING QUALITY IE TRAINING RULE - SECTION 306 WASTE ACT NRR l
l
- O
()
- 6. ANALYSIS OF JAPANESE /U.S. MAINTENANCE PROGRAMS SCOPE:
COMPARE OPERATING EXPERIENCE COMPARE MAINTENANCE REQUIREMENTS ANALYZE ORGANIZATION AND MANAGEMENT MILESTONES:
APRIL 1984 -
PROGRAM INITIATION MAY 1985 -
PROJECT COMPLETED
[)))
JULY 1985 -
NUREG/CR-3883 AND 3883P PUBLISHED j.
l t
I l
O F
- 8. HUMAN FACTORS OF IN-SERVICE INSPECTIONS SCOPE:
IDENTIFY HUMAN ERROR POTENTIAL MILESTONES:
4 MAY 1985 -
PROJECT INITIATION JULY 1985 -
DATA COLLECTION DECEMBER 1985 -
FINAL REPORT t j .
- O i
i I
i O
() .
- 9. HUMAN ERROR IN WRONG UNIT / WRONG TRAIN EVENTS SCOPE:
IDENTIFY PROBLEMS ANALYZE ROOT CAUSES DEFINE ACTIONS FOR RESOLUTIONS MILESTONES:
JANUARY 1984 -
AE0D REPORT MAY-JULY 1984 -
AEOD/NRR COORDINATION
({])
JUNE-DECEMBER 1985 -
SITE VISITS JANUARY 1986 -
FINAL REPORT i
i l
lO
[haA E SURVEY OF CURRENT MAINTENANCE PRACTICES r
TASK 1 -
DATA COLLECTION AND ASSESSMENT
- TASK 2 -
SALEM PREVENTIVE MAINTENANCE PROGRAM TASK 3 -
QUESTIONNAIRE
- TASK 4 -
MAINTENANCE REVIEW PROTOCOL L
TASK 5 -
SITE SURVEYS -
TASK 6 -
SUMMARY
REPORT 2
I O
SUMMARY
APPROACH AND SCHEDULE
- TASK 3 Develep l&E j MAY 15_1905 ouestionnaire -
Administer Analgre Data Prepare N ARCH 15.1906 kstiermire Report N
i TASK 1 Este Data Perform UPdate Data Ero )
l Base Cemparative -
Base >
Structure Analysis t TASK 2 l Este Maint identify Data Colket Data Arrany Visit Salem TASK 6f j Prepare Prepar e Taxonomy -
Heeds en salem PM Salem Site - and Assess" -
Report Aneg/CR -
f Visit PM TASK 4 Develop Haint Conduct Pilot Compkte Review - Site Visit -
Proteeet Protel - Protocol START TASK 5 -
Determine Arrange Site CenJuct Site Corr.plete Site l Plants to Visit - Visits Visits Visit Reporis l
4 i
a, .._. AL----
DATA COLLECTION ASSESSMENT
[) ,
DATA SELECTION l BACKGROUND INFORMATION
. MAINTENANCE PROGRAM CHARACTERISTICS MAINTENANCE PROGRAM EFFECTIVENESS
, DATA CATEGORIZATION DATA ANALYSIS GOOD FEATURES
[)
SHORTCOMINGS / DEFICIENCIES COMPARATIVE ANALYSIS PROGRAM DIFFERENCES AND APPARENT RESULTS PRELIMINARY CORRELATION OF CHARACTERISTICS AND RESULTS 1
O 4
. - , - ,- y __.. ._ . . , , . - , _ , . , , _ . , _ _ _ , _ - - - , , - . . - - - _ , - - - .
4
~
i
! SALEM PREVENTIVE MAINTENANCE PROGRAM
!O F
1 I IDENTIFY STRUCTURE / CHARACTERISTICS OF PM PROGRAM i
. IDENTIFY SAFETY BENEFITS i
4 IDENTIFY IMPLEMENTATION COSTS i-1 l
~
- i. O :
- i i
4 4
i i
I 4
i 4
O t
1
- - - - - - - - - - .,mw.e..-...,_,, m,,,,,,,.,.__w.w-, ,. _ _ _ .,my, , , . _ _ . , , , . _ __
QUESTIONNAIRE OBTAIN IE/ REGIONAL INPUT EXPAND MAINTENANCE DATA BASE (TASK 1)
~
O
s O MAINTENANCE REVIEW PROTOCOL r
-l DEVELOP PROTOCOL FOR USE ON SITE VISITS i
9 I OBTAIN NRC 0FFICE INPUTS i
i i USE DURING INITIAL SITE VISIT >
i l
i
! REFINE FOR OTHER SITE VISITS l
~
!O i 1
I i
l l lO l l i
O Site SuavEv
- r SITE SELECTION REGIONAL INPUT i
CROSS SECTION OF PLANTS J
SURVEY REPORT
. O l
O
-. . . - . - - - - _ . -.. .= _ _ -. - - - . . -.. --. _ . . - .
SUMMARY
REPORT l
O .
r b
- CHARACTERIZE CURRENT STATUS OF MAINTENANCE I
IDENTIFY GENERIC ISSUES REQUIRING PROMPT ACTION I
! GROUP PLANTS BY CHARACTERISTICS / PERFORMANCE O
i i
+
- i I
i .
iO I
)
n
, - - - - - - - - . . , _ , - -. . _ _ _ _ , - . --- - . _.en,, -
, -- _,.------n
,w--,----,,w.
/XC Aug/d n
NUMARC MAINTENANCE PERFORMANCE INDICATORS '
I. DATA AVAILABLE THROUGH EXISTING SOURCES
- 1. UNIT FORCED OUTAGE RATE
- 2. UNIT EQUIVALENT AVAILABILITY .
- 3. NO. OF UNPLANNED AUTOMATIC SCRAMS DUE TO MAINTENANCE II. DATA TO BE OBTAINED FROM UTILITIES
~
- 1. TOTAL RADIATION RATE PER UNIT DUE TO MAINTENANCE
'O 2.
3.
OVERTIME WORKED BY MAINTENANCE PERSONNEL LOST TIME ACCIDENT RATE NO. FOR MAINTENANCE PERSONNEL
- 4. AMOUNT OF OUTSTANDING NON-0UTAGE CM WORK
- 5. RATIO 0F HIGHEST PRIORITY NON-0UTAGE CM WORK REQUESTS
- 6. PM ITEMS OVERDUE
- 2. SAFETY SYSTEM AVAILABILITY l
l O
1 l
I
NRR STAFF PRESENTATION TO THE O ACRS .
SUBJECT:
(6AIN"M TNbICA7 eft.S DATE: Jgwg )3 y Sf PRESENTER: hg4 N , f(t Sqd Lf A PRESENTER'S TITLE / BRANCH /DIV: $/1,MA%4 ANA$F k(ME &%4cArtant $2ANcn PRESENTER'S NRC TEL. NO.:
SUBCOMMITTEE: MhlM4NAMT O
1
- ~~~~ -~ ~ ~~ ~
BACKGROUND 3
(U ,
4 NUREG/CR-3215 - ORGANIZATIONAL ANALYSIS AND SAFETY FOR UTILITIES WITH NUCLEAR POWER PLANTS (JULY 1983)
NUREG/CR-3737 - AN INITIAL EMPIRICAL ANALYSIS OF NUCLEAR POWER PLANT ORGANIZATION AND ITS EFFECT ON SAFETY PERFORMANCE (NOVEMBER 1984) s
()
- FINAL DRAFT NUREG/CR-4284 - AN EMPIRICAL ANALYSIS OF SELECTED NUCLEAR POWER PLANT MAINTENANCE FACTORS AND PLANT SAFETY (ESTIMATED JULY 1985)
()
i
APPROACH IDENTIFY AND CODE EXISTING CHARACTERISTICS OF PLANT MAINTENANCE PROGRAMS IDENTIFY AND CODE MAINTENANCE INTERMEDIATE SAFETY INDICATORS i
r IDENTIFY AND CODE FINAL INDICATORS OF PLANT SAFETY t
>O
- EMPIRICALLY TEST THE RELATIONSHIP BETWEEN THESE VARIABLES i
I
! i t
l I
rO :
l 1
l-
--,.___-,,...m...-.- - - - . --__,,,_r_.-_____._,,,,..,__ _.-_ _ _ ___. - _ _ . - . - _ - _ . _ . . . - - _ _ _ - , . _ - - - . -
MAINTENANCE RESOURCES
- 1. STAFF SIZE
- 2. COMBINED MAINTENANCE /IaC
- 3. COMBINED MAINTENANCE /0Ps
- 4. COMBINED MECHANICAL / ELECTRICAL
- 5. NUMBER OF RANKS
- 6. MAINTENANCE MANAGER RANK
- 7. ENGINEERING SUPPORT
~
- 8. EXPERIENCE O 9. NAVY EXPERIENCE
- 10. NUCLEAR EXPERIENCE
- 11. R0/SR0 LICENSE O
. - - . - . . . - _ - _ _ . - - - - - - = . . - . - - . . . - - - - - _ - _ . . _ - . _ _ _ _ _
e MAINTENANCE INTERMEDIATE SAFETY INDICATORS J
l
[
- 1. MAINTENANCE LERs
- 2. MAINTENANCE SALP RATING f
i
- 3. SURVEILLANCE SALP RATING 1
i 1
- 4. SERIOUS MAINTENANCE NONCOMPLIANCES l
. 5. LESS SERIOUS MAINTENANCE NONCOMPLIANCES 1
l
. 9 I
i O ;
i f
i 4
i FINAL SAFETY INDICATORS .
l 1
i i
- 1. MAN / REM EXPOSURES - 1982 I
l
- 2, AIRBORNE EFFLUENTS - 1981 (CURIES !
1 I
l r 4
l 3, 1-131 AND PARTICULATES - 1981 (CURIES)
- 4. LIQUID EFFLUENTS - 1981 (CURIES) ,
t
' i 1
l 1
!v 5. . TRITIUM EFFLUENTS '1981 (CURIES) i i !
s 4
- l. i i t I
. i i [
t
> j i'
1 ,
!O 1 1 i
i I
. ~ _ _ , - . . . _ . . . _ . . - . . . . . _ _ . .
t
[)b RESULTS e
WEAK RELATIONSHIP BETWEEN MAINTENANCE INTERMEDIATE SAFETY INDICATORSAND THE FINAL SAFETY INDICATORS DUE PRIMARILY TO INCOMPLETE DATA SETS AND LACK OF TIME SEQUENCE INFORMATION RELATIONSHIPS OF STATISTICAL SIGNIFICANCE BETWEEN MAINTENANCE RESOURCES AND MAINTENANCE INTERMEDIATE
, SAFETY INDICATORS SHOWED AB0VE-AVERAGE PROGRAMS TENDED TO
' BE SMALLER IN STAFF SIZE, COMBINED MAINTENANCE AND c
OPERATIONS, SEPARATED MECHANICAL, ELECTRICAL AND 18C, AND HAD FEWER SUPERVISORY RANKS.
RELATIONSHIPS OF STATISTICAL SIGNIFICANCE BETWEEN MAINTENANCE RESOURCES AND THE FINAL SAFETY INDICATORS SHOWED BETTER PERFORMING PLANTS TO HAVE SEPARATE MAINTENANCE AND OPERATIONS DEPARTMENTS, SMALLER STAFFS, AND FEWER SUPERVISORY RANKS.
+
\
O
\
4
$W5p O
MAINTENANCE PERSONNEL QUALIFICATIONS PROBLEMS LACK 0F STANDARD MAINTENANCE JOB CLASSIFICATIONS EDUCATION, EXPERIENCE, AND TRAINING FOR MAINTENANCE PERSONNEL VARY WIDELY
~
DOCUMENTING KNOWLEDGE, SKILLS, AND ABILITIES OF MAINTENANCE PERSONNEL DIFFICULT SELECTION AND SCREENING CRITERIA VARY LIMITED NUCLEAR TRAINING FOR CRAFT PERSONNEL HIGH PROP 0RTION OF TEMPDPC'Y tIRSONNEL (2/3) 2
, + - - - - , . , - - - - - - , - - -
i O .
MAINTENANCE PERSONNEL QUALIFICATIONS NEED TO DETERMINE THE KSAs REQUIRED FOR MAINTENANCE JOB TASKS TO IDENTIFY THE RELEVANT SOURCES OF THE KSAs IN TERMS OF EDUCATION, TRAINING, AND APPRENTICESHIP PROGRAMS TO CONDUCT AN ANALYSIS OF APPLICABLE INDUSTRY GUIDELINES AND STANDARDS AGAINST JOB RELEVANT KSAs AND SOURCES 1
O
.. . _ _