ML20012F579

Provides Fourth Ltr Rept of Nuclear Safety Research Review Committee.Repts on Two Subcommittees Which Met in Fall of 1989 to Review Certain Programs

ML20012F579
Person / Time
Issue date:	02/28/1990
From:	Todreas N NRC - NUCLEAR SAFETY RESEARCH REVIEW COMMITTEE
To:	Beckjord E NRC OFFICE OF NUCLEAR REGULATORY RESEARCH (RES)
Shared Package
ML20012F577	List: ML20012F576 ML20012F579
References
NACNSRRC, NUDOCS 9004160213
Download: ML20012F579 (11)

Text

O

.!.# *** % '\\

[

userveo states s

NUCLEAR REOULATORY COMMISSION Nucleet Sofety Reneetch Review Committee weehineten, 0.c. seems February 28,1990 Mr. Eric S. Beckjord, Director Office of Nuclear Regulatory Research US Nuclear Regulatory Commission Washington,DC 205$$

Dear Mr. Beekjord:

This is the founh letter repon of the Nuclear Safety Research Review Committee. This letter submits the reports of two subcommittees which met in the fall of 1989 to review certain of your office programs. These reviews are as follows:

• Repon of Containment and Seismic Subcommittee from Program Review of October 19 20, Albuquerque, NM.

• Repon of Human Factors and Reliability Subcommittee from Program Review of November 20 21, Cambridge, MA.

All repons have been reviewed by the full committee. Regarding the Report of the Containment and Seismic Subcommittee,I wish to emphasize my view that, while the LANL shear wall test results raise possible imponant questions, established criteria should be used in deciding the proper role and balance regarding USNRC RES versus utility sponsorship of technically destrable follow up activity.

All Subcommittee members join me in expressing our thanks to you, your staff and the participating contractor personnel for making these reviews constructive and technieslly

. stimulating. Addenda to this letter transmitting the reports of the three additional Subcommittees will be forwarded to you as soon as these repons are completed.

. Sincerely.

[.4 '-

et2b Neil E. Todreas NET:pje xc: All Committee Members J. Taylor, EDO, Inc.

Chairman Carr 9004160213 900312 PDR ADVCM NACNSRRC PDC 1

l j

r Review & Synthesis Associates i

Spencer H. Bush, P.E. e 630 Ceoer / Richland. Washington 99352 January 24, 1990 Professor Neil E. Todreas

{

Massachusetts Institute of Technology i

Nuclear Engineering Department l

Building 24 108 Cambridge, MA 02139 l

Subject:

Review of Containment and Seismic Research Programs October 19 I

20 at Albuquerque f

A meeting of the Containment / Seismic Subcomittee of the Nuclear Safety Research Review Committee met in Albuquerque, New Mexico on October 19 20 to review Containment and Seismic Research Programs. Members present were S. H. Bush, Chairman, J. M. Hendrie. H. S. Isbin and D. Morrison.

The Subcommittee considered presentations by Sandia (SNL), Los Alamos (LANL)

I and USNRC RES.

In addition Dr. H. T. Tang of Electric Power Research l

Institute (EPRI was present for the seismic discussions.

Attachment A

}

is a listing of)those attending.

l i

The Subcommittee was impressed by the combination of experimental results j

and analytic techniques used in resolving both containment and seismic i

issues. Additionally, the peer reviews were excellent, with those on the t

review panels having international reputations in their fields.

l Major experimental and analytic programs in both containment and seismic i

are being completed, some of which will be discussed further. While most l

of these activities are completed or nearing completion, the Subcomittee i

stresses the need for sufficient funds to complete these activities.

Also, it is essential to have a level of continuing funding to maintain the strong relationships with EPRI and various overseas organizations.

This cooperative approach maximizes information transfer at very nominal costs to USNRC RES.

Examples of such payoffs are cited later.

1 l

CONTAINMENT The Subcommittee reviewed the test programs at Sandia and visited the test site to examine the 1/6 scale reinforced concrete model and the failed i

steel containment models. In particular, we were impressed by the quality of research related to the concrete containment.

This research has t

achieved international recognition while providing the USNRC with critical baseline information to va'idate analytical methods so as to be able to i

l estimate margins in other containment designs.

The programs include i

i Telephone: Business - (509) 375 2223 & 375 3749 / Home - (509) 943 0233 l

l I

^

Review & Synthesis Associates

$penter H. $vsh, P E e 630 Cedar / Richiens, Washington 99352 l

2

)

failure testing of the models, testing of containment penetrations including bellows, testing of equipment hatches and drywell heads, and the I

on going separate effects testing program on liner tearing.

The Subcommittee believes a high priority should ba given to completing the liner tearing and bellows programs since failures in these areas following an accident could result in a major rotential release of radioactivity.

The probability of failure of these bellows must be l

established.

Of higher priority is the liner tearing program.

The j

concrete containment failure tests revealed that restrained regions in the i

liner such as near penetrations were potential failure sources.

Particular emphasis should be given to the spacing and size (s) of studs attaching the liner to the concrete containment as they can significantly i

influence liner failure.

Optimization of stud layout could prevent or i

delay failure during severe accidents, j

The Subcommittee listened to presentations on the British Central Electricity Generating Board (CEGB) prestressed concrete 1/10 scale j

experiment.

Because of the preliminary nature of the results we prefer not to comment.

In summary, we found the USNRC RES containment programs of uniformly high quality and directly relevant to future reviews of various containment designs. We compliment those persons at USNRC RES and at SNI. responsible j

for the programs for such quality work.

SEISMit Seismic research covers a

spectrum of programs grouped below.

I Particularly noteworthy is the strong liaison with EPRI where sharing of activities has maximized results while minimizing costs to the participants. This liaison also covers programs in Europe and Asia where j

there is multiple support of programs. We consider continuing support of t

these cooperative programs to be vital in resolving major seismic issues at minimal costs to USNRC RES.

i Programs discussed included:

. Category 1 Structures and Shear Wall Stiffness Reduction Effects:

i

. Resolution of Seismic Piping Issues cited in NUREG 1061;

. Seismic Margins:

. Seismic fragilities;

. Earthquake Experience Data; i

)

Telephone: Business - (509) 375 2223 & 375 3749 / Home - (509) 943 0233

t j

Review & Synthesis Associates Spencer H. Bush. F.E. e 630 Coeat / %hlens.Wuhington 99352 3

. Earth Science Research:

. Appitention of Research to Resolution of Regulatory Issues Several of the programs cited have been completed or are approaching

)

completion.

The next stage will be application of the results in the i

regulatory process.

Examples of such programs include pipe damping, independent support motion, pipe margins under very severe seismic loads, l

application of the seismic design margins approach, partial completion of i

i seismic fragilities and the load combination program.

All of the above i

will directly impact on the regulatory process and result in substantial l

revisions to regulations and other regulatory documents such as Regulatory i

Guides. The Committee wishes to be kept informed of such seismic guidance as it is prepared.

I f

Cateaory I shear Walls Tests at LANL on concrete panels typical of seism 4 integory I structures such as primary auxiliary buildings revealea that structural stiffnesses obtained were less than used in the ACI Codes.

This will require Code i

j changes in the future.

The impact of the lower stiffness is a frequency shift in the floor response spectra which is magnified at higher locations in the buildings.

The regulatory impact is the changed loads that may i

l accompany the frequency shift which could increase or decrease the I

probability of failure of critical components.

The intent is to review t

three sites, a PWR on rock, a BWR on rock, and a PWR on a stiff soil i

column. The Subcommittee agrees with these choices however, analysis of l

a sof ter soil site could lead to a clarification of the behavior of Category I structures on such a site.

Presumably the effect of lower j

stiffness and frequency shift should diminish on such a soft (i.e., a moderately stiff soil). site, but the Subcommittee visualizes possible l

questions arising.

Since the results impact directly on the utilities, we hope someone pursues this issue; if not USNRC RES, then the utilities or EPRI.

Pinino issues The jointly funded RES/EPRI simulated seismic loading of piping has confirmed the large margins against failure. A direct outcome will be a J

relaxation of failure criteria.

The damping results should lead to a relaxation of ASME Section !!! Code design criteria and of regulatory j

documents such as Regulatory Guides and Standard Review Plans (SRP's).

t Generally, the NUREG 1061 seismic piping issues have been resolved, including those related to supports.

I

)

t Telephone: Business -(509) 375 2223 & 375 3749 / Home - (509) 943 0233

i l

8 l

Review & Synthesis Associates Spen::r H. Bush. P.E. e 630 Cedar / Richland, Washington 99352 4

i Seismic Maroins

=== I This methodology developed with EPRI is now being validated on rock and soil sites in connection with a seismic probability approach, it is typical of Code procedures in providing substantial margins against i'

failure. The approach has been applied successfully to various plants.

Seismic Fraci11 ties Emphasis on seismic fragilities has shifted from pressure boundary i

components, where there are ample

margins, to electrical and instrumentation components, which have been found to be much more sensitive to seismic loads. The most critical program deals with relays.

New and ongoing work utilizing EPRI programs in support of the Seismic I

Equipment Qualification Utility Group (SQUG) should be completed to establish how sensitive relays are to seismic loads since they may shut i

down a plant and prevent critical equipment from starting, j

J Earthouake Excerience Data l

The RES and EPRI programs aimed at determining damage to equipment and structures typical of power plants has provided an invaluable data bank directly applicable to older nuclear power plants not designed to current j

seismic criteria.

In essence, these data provide the bases for a

}

probabilistic surrogate to a deterministic seismic design.

The Subcommittee strongly recommends continuing such a data collection.

For i

example, such a review is underway on structures damaged in the recent San Francisco Bay Area earthquake.

j Earth Science Research As noted earlier, most of the engineering and structure related seismic I

l programs are nearing completion.

Most work in those arent should be l

completed in the 1990 92 time frame.

I l

The remaining area where continuing financial support is vital and where I

l there should be a major payoff is a better knowledge of the magnitudes and return intervals of earthquakes in specific regions or sub regions of the l

United States.

An example is a clarification of the region affected by recurring Charleston earthquakes. Early studies had this earthquake with the potential of occurring over a very large region of the Southern U.S..

l Relatively recent paleoseismic research has confirmed that repeated severe u

Telephone: Business - (509) 375 2223 & 375 3749 / Home - (509) 943 0233 l

I l.

1 j

t 1

Review & Synthesis Associates l

i Spenorr H Bush, P.E. e 630 Cedar / Richtend, Washington 99352 5

I 5

earthquakes have occurred in the near vicinity of Charleston over the past l

million or so years. While these quakes cannot be tied to a specific fault structure, they can be limited to a relatively small area.

It's obvious that an extension of this approach can be a powerful tool in l

I l

determining recurrence intervals of severe quakes and specific regions l

that are affected.

We strongly recommend continues funding of these j

trograms.

This has the further benefit of establishing strong ties setween academia, geotechnical consultants and USNRC RES since the J

universities and the consulting groups are the repository of such experience.

Another aspect of the geophysics program is the seismic network which has I

L been operated by USGS with funding from RES. Currently, RES is funding an l

upgrading of the network after which the United States Geological Survey l

(USGS) has agreed to operate it.

Since funding for such activities depends on Congress, there is no assurance that the program will continue.

We feel such a program is vital. If USGS is not funded, then USNRC should l

be prepared to supply funds.

We feel the preferred approach is USGS i

funding with a strong liaison between USGS and USNRC.

Resolution of Reaulatory issues We believe a sufficient level of funding of seismic activities is critical l

to permit activities such as the regulatory resolution of seismic issues, i

support of Individual Plant Examinations for External Events (IPEEE)

]

activities, development of RES in house activities supporting seismic review, and ability to participate in cooperative programs, if funds are not ayajiable, the strong links with EPRI and with other countries could wither with a severe loss in the access to seismic information.

A recent article in ' Mechanical Engineering" argued that the United States is inadequately supporting seismic activities. While this may be somewhat of an exaggeration, there is little doubt that the level of support has been diminishing.

The recent California earthquake illustrates the j

implications of a severe quake. Fortunately the nuclear power plants are in better shape than many other structures.

l l

Telephone: Business - (509) 375 2223 & 375 3749 / Home - (509) 943 0233

,.3 t

t t

l Review & Synthesis Associates

$ pencer H. Bush, P.E. o 630 Cecer / Richtens, weshington 99352 i

l 6

t

{

I l

i t

[

Finally, we compliment those presenters from LANL, SNL and RES for i

providing an excellent overview of the seismic programs.

We feel that major progress has been made in the past decade and many of the seismic l

l issues of the late 70's and early 80's have been resolved, j

Very t.

yours,

{

,/ ffKfY

'Spefcer H. Bush fc4sident i

KEVIEW & SYNTHESIS ASSOCIATES cc/ J. Hendrie i

H. Isbin O. Morrison l

?

i l

t I

l l

l l

I i

i l

J l

I Telephone: Business -(509) 375 2223 & 375 3749 / Home -(509) 943 0233

L j

i I

to:

Prof. Neil Todreas, Chair NSRRC 24108 MIT

(

Cambridge,MA 02139 ft:

T. Sheridan I

f Subcomminee members:

I i

S. Bu rstein N. Todreas f

D. Wilson i

Date: 29 December 39 Report of Human Factors and Reliability Subcommittee, NSRRC l

Meeting on 20,21 November, MIT, Cambridge, M A This report will first give general comments and conclusions, then review our reactions to i

Individual presentations in the order in which they were made.

}

General Comments and Conclusions I

The NSRRC Subcommittee on Human Factors and Reliability represents a diversity of perspectives, some from within the human factors and reliability community, some not.

We agree on the need for a vigorous l{uman Factors Program that is realistically responsive l

to stated needs from users.

As a group we are pleased that the human factors and retlability programs survived budget i

i cuts rather well, and are favorably impressed by the thoughtfulness and energy of the program generally.

We also agree on our anxiety that the Human Factors Program, essentially because it is a new and different entity in hRC culture and the nuclear power community:

4 (1) not spread itself too thin in view of a limited budget. congealing small projects together or just plain abandoning some may be necessary to have critical mass in others:

i (2) explicate implicit assumptions, not only to the Subcommittee (you did that), but also to

{

users:

l (3) ensure that utility engineers have appropriate input across the program:

]

(4) make sure it delivers products which users can affirm as useful; and (5) doesn't promise more than it can produce, since the woods will continues to be full of

)

snipers. To quote one subcommittee member on this last point, "I see great potential i

benefit from the NRC human factors research program, but I am absolutely certain that we i

will not gain the confidence of regulators, users or publics if we make claims for human factors research that we will be unable to deliver. No one will remember budget or schedule constraints.. only the promises made and the promises kept".

We appreciate the considerable efforts put forth by all the presenters as well as by the l

responsible NRC staff to make this an informative and generally excellent meeting.

l I

l i

i

i l

l l

Comments on Individual Presentations Coffman overview for Human Factors Branch it was helpfulto see NRR programs and initiatives in juxtaposition to those of AEOD and NMSS which relate to Human Factors, and then see how the user needs received from each area are translated into i

research on personnel subsystems and performance, human. system interface, organization and management, and reliabillry.

4 As an aside, the prodocinional document from Stello to the Commisioners entitled i

NRC's Human Factors Programs andinitiatives made a good case for human factors activities, was very informative, and raised some additional issues to those presented, i

We appreciate RES e(forts to respond to a diversity of stated user needs in this area, i

i However, the*e may be a problem of how many different directions the Human factors J

Branch can go We warit to monitor whether its penonnel are feeling cut into too many pieces, such that it is hard to achieve closure on anything, f

l There is also the nagging concem, expressed by em subcommittee member, of whether the Human Factors program is searching for issues whose resolution might aid 4

j regulation or resolving regulatory problems already evident.

Persensky overview of Human Factors Section activilles. There was concem I

j expressed that BNL is not the best place to be doing the staffing, shift scheduling and j

other persortnel subsystem work. How does this research differ from what INPO does, j

i and is INPO's experience being brought to bear? Next time we would like to hear from an investigator or two in this stea to make sure that adequate operational experience is being I

J brought to bear, t

We agree that improved measures for training effectiveness need to be found.

l Almost all plants have tmining simulators now. Perhaps they be used more directly, e.g.,

as with the General Physics " performance measurement system" which uses requalineation l

tmining exercises to measure and give detailed feedback on specinc opemtor errors.

i i

Sheron/ Carter /O'Hara on advanced I&C technology. We foundinteresting Sheron's comment that the biggest obstacle to advancing !&C technology is the " utilities not knowing what NRC wants, especially the validation and verification of software".

4 Advanced I&C Seems to us a priority research topic, 1

it seems that the survey, such as was reported by Carter, raises more questions than it answers. We are not surprised that the issues of operator tmst, qual 10 cations of control l

room operator, whether a design guide for advanced I&C looks anphing !!ke that for conventional instruments, and the potential for information overload came out as highest i

2 priority. What about operator underload? It seems that as technology advances, the transient is potentially greater.. from a more exaggerated underbad to a more accentuated overload. Both the commercial aviation and the military are worried about the same phenomenon with the advance of automation. We hope there will be substantial engineering input into this ONRL study as it progresses.

O'Hara's presentation on an advanced control room design review guideline was strong. The danger in this project is that NuReg 0700 merely be polished up without rethinking fundamental differences in new I&C from the old.

l One subcommittee member was concemed whether, because things r.re changing so j

fast in this field, one can suf0ciently anticipate the future.

l Higgins on local control stations. Here is where the human factors principles and economics can be in rather stark con 0let, since the human factors status of many local control stations is not good. In this regard we note that there was considerable confusion felt by the control room design review teams on whether to include local control rooms in r

1 i

i

f o

the TMI.2 mandated resiews. There is need here for NRC to admit to the compromises up I

front, and cenairdy to insist that all that can be done with paint, tape and labels and procedures be done to make cmergency use of these control rooms even credible from a

[

human factors viewpoint. The re are also interesting leaues of functional centralization and l

I

{

communication.

j Barnes on procedure violations and environmental effects. This ts obviously a sensitive area, and one where there was some anxiety expreued by several of the NSRRC i

wbcommittee We agree to the need to distinguish intended " procedure violations" and l

l

" slip" type errors of both ominion and commission in administrative, test and operating j

I procedures. We all had a bit of dimculty understanding the A,B,C category distinction.

One members concem is whether there is wfficient engineering input into the study, and how the remits of weh a study might be used, Anothers is whether procedure violation $' errors are not largely dexndent upon penonal characteristics of operaton and i

plant design. A shared feeling is t se danger in underfunding of such an effort and drawing

[

conclusions based on insumelent evidence.

The presentation on environmental (e.g., vibration, acoustic noit e, temperature.

l I

lliumination) was not convincing, but we appreciate that Bames was wbstituting for the PI l

or NRC monitor. A literature review wrely will reveal that the behavioral and i

physiological effects are already well estabilshed in the literature. The problem, it seems to i

j us, is what exposures to extremes of environment occur in operating a plant, and what risks there are to misoperation and worker safety. We didn't hear that emphuis, i

1 i

Ryan, Haber, Nichols on organizational factors. When research in this area was i

first recommended by the Natl. Research Council Human Factors tepon, NRC seemed i

initially reluctant, but now is taking these recommendations seriously. We liked Ryan's positive introduction. Clearly there are some "name brand" researchers purwing aspects of y

the effort. Clearly also there is a need to "get the problem into PRA space" as Ryan put it.

And it also seems that NRC is doing this rer.earch in sumciently " critical mass", short of i

j which no one would be convinced of much of anything.

For the moment the organization of the research program into " process and t

outcome", " active and passive information" and " indexing" methods seems reasonable, in the laner category one of us is quite happy that NRC is recognizing that subjective

[

judgments, even of numbers, have been playing a major role for r,ome time, and we need i

to understand better what we are doing and try to do it better, it will not be easy to keep l

this research program focussed.

Haber's project - using a Pittsburgh fossil plant and Diablo Canyon as test cases I

seems a good way to focus. The nuclear community will demand that some connection be made to, and some tractable tools be provided for, risk assessment, here is some I

skepticism within the subcommittee that new hypotheses can be discovered and verified in this area with any degree of objectivity and with predictive validity for given plants.

Nichols' presentation and preliminary rewlts on leading organizational indicators,

e.g., that "efrielency and safety are not correlated", also generated subcommittee interest, 2

and for some members more skepticism, t

Ryan overview on tellability assessment. We agree with Ryan that traditionally f

human behavior has been treated peripherally in the PRA and that within the human factors community there has been a lack of dedication to human reliability analysts. These facts seem to stem not from apathy to risk or a regard that human behavior is irrelevant, but more from a skepticism that complexities of human behavior can be reduced to a scalar prediction i

of probability of binary success failure, really the extrinsic statistical reliability (credibility, 6

applicability, generalizability) of the HRA itself. Since WASH 1400 there have been many refinements in HRA, including sensitivity analyses, performance shaping factors, means of handling common mode failures, study of the combined performance of teams of I

=

l l.-

l operaton, and now a beginning encounter with errors of commlulon in addition to those of omiulon.. all of which have ameliorated the resistance to HRA somewhat, lira has

[

not yet accommodated errors in thinking / planning, though there are now serious efrons to l

deal with mental models.

l One subcommittee member feels that PRA's are most useful for those who do f

them, and that their utility to utilities (sorry about that!) depends upon the PRA analysts (e.g., the PLG's) talking to the utilities (or regulaton!). He worried about HRAs making i

the PRAs leu censin, and about how well implielt behavioral auumptions can be explicated. He also wondered whether HRA/PRAs can show where replacing a human with a computer will reduce risk. Finally, will ht. man factors efrons tell us(and saboteun) l too much about how to destroy a plant?

ans !

Buffardi on transfer of non. nuclear data for nuclear HRA, Clearly,1f this hasn't been done before it's good to do. No reason to believe that data won't transfer from non. nuclear to nuclear if tasks are properly qualined.

l t

r Gertaan on M APPS. Bis is a modelling project the subcommittee worried about last j

year, based on promises for a model which combines the efrects of all variables in sight..

l many efTeets orwhich are not now understood independently. If the purpose of the model l

is to describe and organize experimental data for PRA analyst " appreciation" of a given past 1

event (a descriptive model), and the PRA analysts find it useful as such, then we cannot i

i 1

fault the effon. If, on the other hand, the intention is to predict and generalize, then w e l

I l

find ourselves quite skeptical that V&V can be done with anything like acceptable l

standards. This effon has been going on for some time now Perhaps the subcommittee i

needs a closer look and some testimony from users.

(

l Roth on cognitive modelling (CES/ CREATE). This strikes us as bold new l

l approach for modelling what the operator is thinking, quite different from and more in line I

with current cognitive science than MAPPS. We are cager to see if(i can prove itselfin i

HRA usefulness on the test cases chosen.

Ryan on TALENT. nis project is a human factors task. analysis and evaluation of the PRA process itself, which we applaud, since the PRA process is so very costly and can be l

full of human errors and pitfalls. We look forward to the final evaluation of this project.

Samantha/Vesely on risk based criteria for tech, spec, improvements (configuration control), and interface with IAEA.

The configuration control research seems a very good idea. Can the results be put in a l

form which both plants and regulators can find understandable and useful?

With regard to the IAEA cooperation, the Subcommittee has and continues to be supponive.

Comment on External Human Factors Activities Spencer Bush brought to the Committee's attention an example of where pans of RES are panicipating in extemal human factors programs. De example is the PISC.!!! (Progmm for Inspection of Steel Components) concemed with the reliability of technicians performing nondestructive examinations. Variables such as noise, lighting, temperature, clothing,ps! tion of the examiner, duration of the test, etc. are considered. The lab phase is in the UK and an " Industrial" phase at JRC.lSPRA. In this case Engineering is the NRC.RES interface. Such international programs could have significant impact in ASME code examinations. The committee should review this and other " external" activities like it in the future.

I

_ _.