ML19261C928

Summary of 781111 Meeting W/Nrc QA Review Team Re Plans to Review QA Practices

ML19261C928
Person / Time
Issue date:	04/12/1979
From:	Bennett G NRC OFFICE OF NUCLEAR REGULATORY RESEARCH (RES)
To:	NRC OFFICE OF NUCLEAR REGULATORY RESEARCH (RES)
References
NUDOCS 7904190311
Download: ML19261C928 (80)

Text

{{#Wiki_filter:..... _.. _ - -.. $fA l f pd UNITED STATES 3V .. 't NUCLEAR REGULATORY COMMisslON j WASHINGTON, D. C. 20555 ++,*** APR 12 1979 l Those on Attached List Gentlemen: I

Subject:

Summary of Meeting of QA Review Team, November 11, 1978 l t On November 11, 1978, the NRC QA Review Team held a coordinatic. f meeting prior to making the first of the field trips. E.W osure 1 l is the list of attendees. Two representatives from DOE /HQ attended this meeting. Below is a summary of this meeting. Gary Bennett provided the team with copies of background material on f the plan to review QA practices. Action items were discussed. (see Encl. 2) I M. E. Langston said OES may be the lead group in DOE for QA. He 4 made several coments and offered suggestions for consideration by l the review team: NRC/RES should have a policy which says that there will be ( i QA on the research programs a contact should be identified for QA guidance, otherwise 2 the NRC programatic staff can provide the technical guidance 1 l within DOE the field offices are responsible for QA I there is a nuclear QA policy (exclusive of weapons and naval) which states that it is a goal to work toward national QA standards (He agreed to send the team a copy of this document) I - new QA standards should not be used if they are not cost-effective ultimately it is up to the laboratory to carry out the QA the DOE approach is to put QA in at the project level i a good QA document " trail" should be established The NRC/DO: interface on the review of QA on NRC/RES-sponsored programs was discussed. DOE /HQ personnel will be invited to attend all meetings with the DOE field offices and laboratories. The NRC Review Team plans to meet with the cognizant fit.ld office staff before meeting with the laboratory staff. Langston said DOE conducts an annual QA review of celected laboratories and he thought NRC could participate in these reviews. J. W. Gilray suggested starting the QA with NRC/RES. 4 790nsc3// m,

.s .a i t APR ; 2 73, 9 Those on Attached List. I F. J. Liederbach said NRR uses a graded approach to 10 CFR 50 Appendix B. Plans for the forthcoming trip to Sandia and LASL were discussed. Sinc r / G ry L. nne ief Research Se ort Branch i Division of Reactor Safety Research 4 i

Enclosure:

as stated I i 4 4 I I I ~ f f i s t .p,

c a c.. 'J' d Addressees for Letter dated ~ Saul Levine T. E. Murley C. N. Kelber B. Snyder, OPE J. Milhoan, SD J. W. Gilray, NRR W. P. Haass, NRR F. J. Liederbach, NRR W. R. Rutherford, IE F. Jablonski, IE-III M. E. Langston, DOE /HQ T. McSpadden, DOE /HQ R. Kuhnapfel, DOE /HQ r G. P. Dix, DOE /HQ l PDR (2) t i i l ~~ l 5 ~ ~ p .....-- - ~. - .m

aa 5 ~~' ENCLOSURE 1 1 ~~ LIST OF ATTENDEES MEETING OF QA REVIEW TEAM November 11, 1978 I Name Organization Gary L. Bennett NRC/RES 9 R. Feit NRC/RES I J. W. Gilray NRC/NRR W. P. Haass NRC/NRR M. E. Langston DOE HQ f F. J. Liederbach NRC/NRR T. McSpadden DOE / DES W. R. Rutherford NRC/IE I 1 I t i .m.

..u a ENCLOSURE 2 j STATUS OF ACTION ITEMS i At the October 6,1978 meeting of the QA review team, the following action items were assigned. The status is reported below each one. ACTION ITEM #1: Bennett to write a justification for the programs selected, f STATUS: A draft justification was distributed (see Enclosure 3). This i action ~ item is completed. l j ACTION ITEM #2: Bennett to set up a meeting with RES program managers during the week of October 23, 1978. The plan for revi aing RES programs for QA adequacy is to be reissued to the program managers prior to these meetings. i STATUS: Meetings were set up. Action item completed, j ACTION ITEM #3: Bennett to discuss QA review with RES/GRSR. i ? STATUS: Bennett discussed the QA review with L. C. Shao who said the i GRSR programs were too new for this effort but GRSR would be willing to cooperate and believes in QA. ACTION IT5M #4: NRR and IE to review briefing book and laboratory QA manuals before October 23, 1978. f STATUS: Completed. MISCELLANEOUS ACTIONS A separate action item was for Bennett to provide document of a sample program from research request through implementation. This is given in l Bennett was asked to provide more information on the process of code assessment. This is given in Enclosure 5. l i .m

h. aumind

.a i i DRAFT GLBennett 11/7/78 JUSTIFICATION FOR SELECTION OF PESEARCH PROGRAMS FOR QA REVIEW I i Loss of Fluid Test (LOFT) LOFT provides an information base to be used in testing the predictive capability of computer codes and to assist in the understanding of the l i more complex nuclear tests. LOFT is probably the world's only nuclear thermal hydraulic test facility for the study of postulated loss-of-coolant h accidents -- as such it has the unique feature of being able to simulate the integrated effects of nuclear heat, primary pump performance, metal-i f to-water heat transfer and steam generator heat transfer. Finally, LOFT is NRC's most expensive single budget item; therefore, the quality of i f the data must be good in order to make the program cost effective. Power Burst Facility (PBF) The Power Burst Facility is being used to produce experimental conditions i in which fuel rods can be tested to failure in a nuclear environment. These tests will permit detailed examination of the mechanisms of failure and of possible propagation of failure from one rod to another. PBF is thus the i principal facility for the checkout of the transient fuel behavior code FRAP-T which is expected to be a mainstay fuel code in the licensing process. i Semiscale Semiscale is a one-dimensional, nonnuclear representation of the thermal j hydraulic aspects of a pressurized water reactor facility. The primary l coolant system is designed with the same system elevations and the same i I

_ ____ L - f e L ratio of total volume to core power as exists in LOFT. Thus, Semiscale f provides important experimental information on scaling thermal hydraulic g phenomena. From these tests, NRC will gain confidence in scaling its analytical tools to full size reactors. Moreover, because of its vers-l atility, the Semiscale results have greatly assisted in the development and testing of computer codes. Furthermore, since Semiscale can be used more often than the more complex L;rT facility, it is possible to study thermal hydraulic phenomena in more detail with Semiscale. In general, Semiscale is closely coupled to LOFT. f Qualification Testing Evaluation (QTE) Program The qualification testing evaluation program is designed to obtain data to further improve the technical basis for current standards and regulatory I guides for Class IE safety-related equipment. This program is very closely related to generic issue A-24 on equipment qualification. Because of its direct relevance to day-to-day licensing issues, and the occasional use of " demon-i stration" hardware, this program is a big part of the NRC operational safety research effort. Fire Protection Research Program The fire protection research program is aimed at obtaining data in support of current regulatory guides and standards for fire protection and control in light water reactor (LWR) power plants. Considerable interest has center 6d on this program, largely because of the Browns Ferry fire and the

...--.--.---..J -a l g requirement to upgrade fire protection measures in nuclear plants. This k program also has occasion to test " demonstration" hardware or designs. t This program is also a big part of the NRC operational safety research ll effort. b i l PWR Blowdown Heat Transfer (BDHT) Research conducted in the nonnuclear Thermal Hydraulic Test Facility (THTF) i { will provide data to determine hydrodynamics behavior, time to critical heat l flux, and transient heat transfer rates during coolant depressurization 1 l (" blowdown"), as influenced by variations in power, system pressure, i coolant flow, and break location. The critical item is determining the time of critical heat flux because this helps determine how long the l fuel rod cladding will remain intact. Also, until the advent of Semiscale Mod 3, this was the only NRC PWR blowdown test facility with a full length (12-ft) core. As such, THTF provided key heat transfer data under more realistic blowdown conditions, thus helping checkout computer i i codes. Heavy Section Steel Technology (HSST) Program The HSST program fccuses on the integrity of the reactor pressure vessel, one of the principal safety design features in a nuclear power plant. The HSST program helps provide a more confident basis for criteria and analytical procedures for the design, fabrication and operation of the componept (vessel, piping, etc.) of the pr imary system pressure boundary of LWRs. This programhelps determine the probability of failure and mechanisms for reducing that probability. The HSST program often tests i ..m.

i ( I material and assembly techniques typical of those used in nuclear power 6 plants. 1 Annular Core Research Reactor (ACRR) This research program provides experimental information on the interaction of nuclear fuel with sodium coolant and structural and coolant materials. Specifically, the reactor provides information on the behavior of LMFBR-i type fuel rods in reactivity excursion accidents. Thus, ACRR gives NRC an independent capability to assess the safety margin in LMFBR-type fuel rods. Transient Reactor Analysis Code (TRAC) TRAC is an advanced, best estimate computer program for analyses of postulated accidents in LWRs. It features a nonhomegeneous multidimensional fluid dynamics treatment of the basic physical phenomena that occur under accident conditions. As such, TRAC will be both the future NRC LWR safety analysis code and the " benchmark" for other codes. RELAP The RELAP series of-codes represent the existing basis for licensing calculations. This series of codes is also used in the analysis of LOFT Semiscale and THTF results. l

EMCLOSURF 4 SAMPLE OF RESEARCH IMPLEMENTATION DOCUMENTATION

9 [O UNITED STATES / / 3^ ^ NUCLEAR REGULATORY COMMISSION 4 WASHINGTON, D. C. 20555 JUN 11978 [.. ,l q ly;l' Q.Y $ N._f-i MEMORANDUM FOR: Saul Levine, Director iN h M. Office of Nuclear Regulatory Research FROM: Edson G. Case, Acting Director M Office of Nuclear Reactor Regulation 'lli 00

SUBJECT:

REQUEST FOR ACCELERATION 0F VERIFICATION TESTING OF VERTICAL AND HORIZONTAL CABLE TRAY FIRE PROTECTION SYSTEMS (RR-NRR-78-13) The NRC fire protection guidelines in Branch Technical Position ASB 9.5-1 state that in plant areas where potential fire damage may jeopardize safe plant shutdown, the primary means of fire protection should consist of fire barriers and fixed automatic fire detection and suppression systems. The purpose of this memorandum is to propose fire suppression tests to verify that these NRC fire protection guidelines provide adequate fire protection measures for grouped cable trays. The tests are intended to demonstrate sprinkler and gas floodirg fire suppression systems effective-ness on cable tray configurations representative of installations in operating plant or plants currently under construction. A memo from L. S. Rubenstein to L. S. Tong dated February 10, 1978, "NRR Comments on RES Branch Plan for Fire Protection Research," stated that NRR considered it important and of a general applicability to have cer-tain fire suppression systems perfomance verified on an accelerated schedule via full scale fire tests involving horizontal and vertical cable trays. On April 28, 1978, NRR, SD, and IE representatives on the Fire Protection Research Review Group concluded the fire protection research program could be beneficially expedited by conducting proof tests on sprinkler systems this fiscal year instead of in FY 1979 as currently scheduled. Attached for your use is a description of suggested fire test configura-tions that represents selected generic nuclear plant fire protection concerns relevant to selected cable tray systems. Both the horizontal and vertical cable tray arrangements have been reviewed by DSS and 00R for their generic applicability. Completion of the tests described in the attachment will provide useful guide information on effectiveness of extinguishing systems for specific problem areas which is the stated objective of RES Branch Plan program element number 8 that covers testing of fire extinguishing systems.

Saul Levine ggg 1g7g The fire test progi 2m described will verify certain proposed or existing fire protection systems, consistent with NRC guidelines, using full scale vertical and horizontal cable tray configurations. Fire protection systems that need to be investigated and verified include: various fire resistant coverings, e.g., kaowool; smoke detectors; sprinklers; halon; and CO - 2 We understand that RES plans to conduct the sprinkler fire suppression test during July and August of 1978 and is reprogramming funds in the amount of $100K to allow the tests to proceed. We also understand that RES is taking steps to provide Sandia with additional equipment funds to enable the laborator; to perform a timely series of full-scale cable fire gas suppres-sion tests in early FY-1978. We request that discussions be held between NRR and RES personnel and the RES contractor to agree on specific test procedures and to arrange for a revies of the test facility arrangement prior to testing. The NRR contact will be G. Harrison, Extension 27763. ', j, ; y b ,u. ~ - Edson G. Case, Acting Director Office of Nuclear Reactor Regulation cc: V. Stello D. Eisenhut L. Rubenstein W. Butler R. Ferguson R. Mattscn .. Tedesco V. Benaroya P. Matthews W. Paulson V. Panciera G. Imbro A. Ungaro C. Long G. Harrison G. Bennett R. Feit

ENCLOSURE Cable Tray Fire Suocression Tests The purpose of perfot - ng demonstration fire tests on selected fire suppression systems is to verify that certain NRC fire protection guide-( lines provide adequate fire protection mersures for grouped cable trays. A. Vertical Fire Sucoression Demonstration Figure 1 shows a plan view of the vertical cable tray arrangement that represents certain plant conditions of common interest to D0R and DSS. Test parameters pertinent to the vertical tray tests include the following: 1. the cable tray width should be 18 inches; 2. no cable tie-downs should be used except that the cables will be secured at the top of the tray; 3. the cable tray height should be 12 feet; 4. cable that qualifies to IEEE 383 should be tested before non-383 type cable; 5.

• de cable trays should be filled to the top of the side rails (40% fill) with the identical cable size mix used in the July,1977, Sandia fire test;

-2 6. two types of test fires shculd be ccnducted: (a) one will be a prepane burner wnich peduces 70,000 Stu per hour and be located to expose a cable tray proper; (b) the other will..:cdel a two-gallon spill of fla=able liquid be' ween redundant cable divisions (only one fire source would be tested during any single fire test); As a prerequisite for these vertical cable tray tests, the igni-tien source will be parametrically varied to establish a standard fire as was done in the hori:entai tray tests. This technique is in agreement with the overall program objectives for develepcent of censistent and reproducible test methcdology. Tne standard fire source will be one that censistently produces a sustainable fire in a given cable test configuratien with the minicum igni-tien source. 7. the sprinkler heads anc detectors shculd have their location based on standard engineering practice and will be deter =ined later by NRR; response of the scrinkler system and f u Iccaticn would be chosen to be recresentative of accected perfor :ance and plant ccnfigurations; 3. the vertical cable tray test set up will use a cerner arrange ent rather than a reca; :nis arrangerant would create the radiatier effects nat walls and a ceiling have en a # ire and is similar

. to the method of full scale fire testing of wall and ceiling finishes conducted by FM, et. al; 9. the fire protection scheme to be tested would include a sprinkler system, fire resistant barrier (kaowool) on both divisions of cable trays, and a smoke detection system. B. Horizontal Tray Fire Suporest ion Test Figures 2 and 3 illustrate the horizontal fire tests representing two divisions of cable trays at. or near the ceiling in a corridor situation. Test parareters pertinent to the horizontal tray tests include the following: 1. The corridor should be approximately 20 to 25 feet in length with a width of 8 feet and a ceiling height of 12 feet; 2. One end should be closed off with closure capability at the other end; 3. One fire source would represent an exposure fire and would be two-gallons of spilled flammable liquid located between the cable tray divisions; 4. A fire in the bottom tray of one tray stack should also be considered; 5. Two different protection schemes are envisioned to be tested; one scheme consists of sprinklers in combination with fire resis-tant barriers and the other scheme consists of either a halon or a CO2 system in combination with fire resistant barriers. Both

W 0 l / ^ \\ / Y / 3 / / v / VDtv I Osv2 ,g / loY a /+-7 M oivi & oiv 2 T i / {f l yl $rf-fin. lcedLr i owa +- I8 '--> /k Yr7lTcAL Co<m<%f -ie o et i J

N D<EMm f c.gensd X W@ Fi a Re s fs l On I 0; a i __p n-eswier f 96l2' ~ __s c=, W--R << red D Sp ddNler p r' f i.t k p?h Loccos 2' @-Acke Di&%- n" xv f? aa 2 b<tdler Sir $e r% L<br< C 9 zrc K, ry,~a ca >a s coaik 4 W@ ,,m a resiin t van wa ^ Ba <riz y .-2 __ _ a r ,..S G-A hktfrr.re$ Erlod se 1 gi4 I C4 96 y, J f'" V _ / $fCbAft?If fla (t 5 ~~ ,9 CA, f~ (hg /M P <4t.1 e,u-C sa Eawe F?a protection schemes will incorporate smoke detectors. The sprinkler tests should be perfomed first. In summary, the fire test program envisioned is one of verifying certain proposed or existing fire protection systems, consistent with NRC guidelines, using full scale vertical and horizontal cable tray configurations. Fire protection systems that need to be inves-tigated and verified include: varicus fire resistant coverings, e.g., kaowool, smoke detectors; sprinklers; halon; and CO. The 2 test facility for the horizontal tray tests has to accomodate the evaluation of gas systems effectiveness. Test fire locations for the vertical cable trays are indicated on Figure 1, and the horizonal cable tray fire locations are shown in Figures 2 and 3. The cable trays should be loaded to 40% full with mixed cables placed in a normal fashion. The initial halon gas concentration should be 10%, whereas, the CO2 concentration should be 50%. The sprinkler / water 2 spray discharge density should be approximately 0.30 gpm/ft. Other test parameters and details will need to be developed in greater detail in conjunction with the testing laboratory.

Draft 6/2/78 Vertical Cable Tray Procram

Background

In a draft letter to RE51, NRR has outlined a series of fire tests involving an arrangement of vertical cable trays. This tray configuration is to be subject to a series of exposure fires in order to demonstrate the performance of water suppression systems in putting out fires before involvement of redundant safety divisions. The physical configuration of cable trays for the test is intended to model a " worst case" situation typical of an actual nuclear power plant configuration. While this vertical tray testing appears to be a relatively straightforward experimental program, there are reasons to be wary in the undertaking of such an experimental effort. These reasons are (1) a previous history of non-reproducibility in similar configura-tions, (2) the " demonstration" aspect which suggest that the tests may be the basis of regulatory policy and, therefore, subject to intense industry scrutiny, and (3) the possible need to meet new QA requirements imposed on RES research programs. These three issues will be settled most satisfactorily from an experimental standpoint by a carefully planned program of single tray tests followed by a few tests of subunit tray configuration (two trays or more) conducted without suppression but with passive protection of redundant systems and tests conducted without either protection system. The " demonstration tests follow this ^ preliminary testing program and represent system response to a known threat.

_2-The purpose of the single tray tests is to evaluate the configuration of cables within the trays necessary to provide a situation in which a fire will, if unchecked, envelop the entire tray. The configuration sought must be reasonably typical of power plant experience in order to assure acceptance of the test p r og r am ; however, this may limit severity of the fire experienced at envelopment of the entire tray. Assuring complete development or the " worst fire" in a tray provides the most credible threat maximum test of the protection systems, and therefore will add to the overall credibility of the entire test program. The subunit tests are intended to provide information on temperatures and heat fluxes experienced in adjacent trays which can "see" the fire. Of particular interest is any area where plume impingement from the burning tray may provide the mechanis= for lighting an additional tray in the same or redundant safety division. These tests will, therefore, provide basic information on the effect and need for barrier / spacing on fire propagation. Since the physical configuration selected by NRR for the demonstration tests is one in which several redundant systems are arrayed in close proximity it is presumed that the issue of propagation of a fire between safety divisions is also of interest in this de=onstration program. Thus, it would appear that one experiment should be devoted to the evaluation of propagation between division without the programmed intervention of a suppression system or other active or passive fire safety devices. In this way it will be possible to evaluate the actual need for other techniques for controlling ':he propagation hazard in this situation. Propagation will ? ovide a strong supporting case for suppression and/or other measures, and, in addition, will allow measurement of the time between fire inception and the time that propagation occurs during which control must be established. Completing the test in this manner would permit the evaluation of a full spectrum of remedies for use in this, and similar, cases, which would become part of the experience base to be used by NRR in judging the minimum protection needed in similar situations. Thus the testing program Sandia would recommend is comparable to, and modeled on, that used successfully in the horizontal tray testing program. The test sequence would be approximately as follows: 1. Single Tray Tests ( Full) -- A series of eight to twelve tests to evaluate reliable methods to arrange cables in a vertical tray which will assure total fire involvement (top to bottom - s'ide to side) and also investigate the minimum ignition source necessary to achieve full involvement. Some detailed configurations mignt be: a. Cables (383 qualified) parallel and evenly spaced (perhaps weighted or in locating fixtures at top and bottom (not realistic but reproducible) b. Cables (383 qualified) bundled in 3's or 4's - bundles parallel and spaced evenly top to botton - bundles tied to trays, c. Cables (383 qualified) placed in full (or nearly full) layers with first layer tied to tray rungs and with subsequent layers spaced from adjacen' layers to allow flame passages over the length ( of the tray. d. Configurations as above tested with propane burner turned on for various lengths of time to establish minimum time necessary for full involvement. Evaluation of optimum configuration found above with e. non-383 cable to detect time-temperature difference ~ in phenomenology. A number of these tests would be completed at Sandia Labs in order to establish the standard donor fire for the testing at UL and to support future test efforts in vertical tray configuration at Sandia's facility. Instru-mentation would be as shown in sketch fl. It is proposed that one test of the single tray configuration be conducted at UL to calibrate data and phenomenology between the two laboratories. 2. Subunit Tray Tests -- A series of tests to evaluate tempera-ture and heat flux in the donor tray and in its vicinity on receptor trays or mock-up trays arranged in typical vertical tray arrays. Heat flux gages and thermocouple

__ arrays will be used to characterize the convective and radiative heat transfer environment (see sketch 42). These tests would consist of: Single vertical donor tray ( filled with 383 qualified a. cable as determined in earlier tests) with Marinite boards simulating other trays adjacent to both tray faces and spaced 10-1/2" apart. Single donor tray as above with simulated trays 3E2 3' b. away. (This test may be omitted if temperatures at 10-1/2" are low enough to suggest that no combus-tion would occur. c. Repeat of "a" with non-383 cable. d. Repeat ob "b" with non-383 cable Tests "a" and/or "c" would be repeated at UL to assure the results are relatable to those obtained at Sandia. 3. Full Scale Tray Tests - A series of tests will be run at UL to model the fire situation in a realistic configuration (see sketch 43). a. Trays will be installed as shown in the NRR sketch with sprinklers in place but not operative, no barriers, and smoke detector (s) operative. Cable fill is 383 rated material instrumented as shown (sketch $4) in trays 1, 2, and 3. The test will be initiated by propane burners. Indication of sprinkler and detector operation will be noted, as

6 . Test to will propagation from one tray to another. system be terminated if propagation to a redundant occurs. in place. Same as shown above except barrier b. initiated by o il spill. Same as "a" except c. Same as "a"- c" except with sprinkler operative. d. h ique The progression of tests would be selected by a tec n hich similar to that shown schematically in sketch 45 w It is expected should minimize the total number of tests. 6 or fewer tests will be required to provide complete that d configu-information on both the vulnerability of the tested in this ration and the phenomena which govern fire sprea configuration.

4 L /[ l b c e # l 3 SCD j O h I r 4 i G O 3r j li h i 3' s 2 O DD t1 O I b i h [ i 3 (( /E' i j w '7-Y I b B TC $ hl yn k>,.9 In S( SeN/1 #/ S kha men labbn (be) y

9 e hgrrwols f4 aL Gp D = x - -x n,. O t ( 0 t I fh EAUX $hY I b I Sense r-8 d7IN I ~ fo% +" (c'/Ob'-* E' O-<x o 9 48" l l o A o, 4 3 t." /D E l f O o I zf I

[

O-x x x, t!F

• i ll l

i ? hY J

? 9 9 ,Cacl Tia s.El do,,w g f 'Ay 5 a M M [My A a s g l x a A i 7' r ~ A a Y I

9 e b nok: cally heiy/at e /4 ' m khaw af }C p.r ac, n- '4 -Tr. &je i% _ hitf2/A1/rd" [Nybber. ?Gai. 2. Pt/2177~totl 'e h f [ ~+4tta.~, // / / / V /I / J / / / / /_^7 / f / / / / / / / 'I \\ Wl.1 / i h b i org . =.a L n,' / M Y 2 l&'TWQ Cpb f 2 / t 13"! ,[C^ ~li.c~-$n.lcc8a.c / 3' y wi l e ;g y sPr inkler o ner /j, receF<er N 1, le-7 - h & aI, a u a p..,.g, ~ hsor-

1. Clb: C e

% %L % C w 7st O fic oft { a d 8 as g y c k,

• 3 S t/ Sale E t layd w-a

9 e \\ '}YV 1 N Wh e , b c- -g.s ],1 s dj ( 'T 4 wi 4 45 -t Q 9 !.5 y2' -~ .i .s e 7 -f 4 T ) w Y Ij r i s 7 u 't c b f = g E tt 4 d a j 57m x+g a 4 0 Y 4 0 s m *g J ' s. 3 .u 2 c = e 's.p,

• ,o y = %

e o?? u .m s e 5 i s nwe a c p-( tu -.0 P. p m ;_ p ,4 x u d4 n _d I k, y oc 1 1 i ~ l l 3 I 4 -m 'e x Sk, 4 5 s i f,, Y k,-u$.'J[- $ g'.Y 1

r

~' b g u% ' f n % +v. 3. c em es 24 A c 9 i e - E i 1a tt -gI t% h - o N A g n.8 s. U, h e a-- M'@ +" g = 3 {. 8 f hA ? f 't N 2 J4 u ^ s C %S L FX., 4 ~ ij m d ::1 g, o =g a s. A 2 ~ A.h O P * +h - w 4' I h7

• R i h

5., Y ,a g N -6> - -~ g

9 e 3 Y.c ~7 3 o.-A ' " c.t. u w g _.- N ~-- m s ___ s YZ M [ ='~ N M 4-wo "s -

r. 4 CL.

"r% o 4 w do

1. 1

\\ A L n o,, b f s . g( c - c w e 0 3 v = .D d 'l a s Q* C I I k J 0 \\ M [- o m c kLc c Q k-v A v' e_ g9 a in m +4 9 s o kg g E C h a b $ _C 4 C .Y 4 4 f C -O .a -i-- o o o C R A e s .- a V l h i 4 a k 0 W g e S S I N i N i L \\ .2 V =_ LM $, ~_ p*h. s* m o l n 4 % %,, p *C.L U A E. ~ 0 i. s o c 4 -% e

• i' 1

0 h n p +, \\ t d-N JC ( l

• 5' o

.Y N I t. A 1 A I .I a y o 'l

1. j S'

\\ t I 3 l t ( 40 } a e t a L s y y 'g - - M

• k 'y~ '

M i N$ $ m c . s s v et o -g-; ; W n a. .o

n.a a ;

• m.t i

-1 a t-h 6a3i "3 U J". l l

.~, v..- 5p,,nkler ^le+ 7,cc,/ 5for Q E Gr,,,,e// A,,kce/- Ac., e 4, - e } "/'r.y.) sc % tac la,cc. "% cia. c e l, ..- Gi.,,cll Q,, <t u ; J ..) (en Onngsn m Sb a-GE Gr >n net / D.< (c../G/cle S.,va// fjr.,fr..,,e s Fe.,.../ 5.,< Gr. .1 :a - e G (i, ew ha., ak.gc) 1 rer f ri Sea, __ e 1 1 t i I e o 9

.r k*/l,yy-Underwriters Laboratorics, Inc. 333 Pfingsten Road Northbrook, IL 60062 (312) 272-8300 STATEMENT OF WCPM Phase 1 FIRE PERFORM VERTICAL CABLE TRAY TESTS WHICH ARE TO BE DES A A. THIS S1ATEMENT OF WORK single tray configuration test to calibrate data and (a) One phenomenology between Underwriters Laboratories and Sandia Laboratories. See sketch #1. Two single tray tests to evaluate remperature and heat (b) flux in the donor tray and in its vicinity on receptor mock-up trays arranged in typical vertical tray arrays. Heat flux gages and thermocouple arrays will be used to transfer characterize che convective and radiative heat environment. See sketch 42. One tray shall be filled first of these with IEEE-383 qualified cable for the test The second test shall consist of a repe,a tests. using non-383 cable. Full scale tray tests -- A series of tests will be run to (c) model the fire situation in a realistic configuration. See sketch #3. For the first of these tests, trays will be in place, barriers in place, and smoke detectors Cable fill is to be with non-383 qualified operative. simulated cable. "ethod'of ignition will be with a fuel spill using Heptane.

Indication of sprinkler an.d detector operation will be noted, as will propagation from one tray to another. Test to be terminated if propagation to a redundant sys-tem occuro. The progression of these full scale tray tests would be selected by a technique similar to that shown schematically in sketch 45 to minimize the total number of tests. It is expected that nine or fewer tests will be required to provide information on both the vulnerability of the tested configuration and the phenomena which governs fire spread in this configuration. Four tests are the minimum required and this phase (?' ase

1) will cover these four tests.

The top of each cable tray shall be at or in close proximity to the room ceiling. Both photoelectric and ionization detecters to be located 15 feet from each side wall. Sprinkler height and spacing to be deter-mined by current design practices. The fusible links on each sprinkler head shall be monitored so that a time-to-open for each head may be recorded. Sprinkler heads from three different manufacturers shall be positioned at each proposed sprinkler location with only one to be actuated after any two open. Sprinkers will be set at -165'F with 30 psig available at the actuated head. For the purposes of these tests barriers will be considered to be 2-inch cohamic wool blankets wrapped around each cable tray. Fire initiation by propane will use standard IEEE-383 ribbon burners. Fire initiated by fuel spill will use Heptane as fuel contained in a sheet metal pan

4* _3-at the base of the cable trays. Reports -- Furnish a report including still phc:cgraphs before and after each test. Colored movies should be ta.len during the test and one copy of these films provided with the report. Closed circuit television should be recorded only as a backup for movie camera failure. Two copies of the report should be provided to Sandia Labora-tories to cover the first full scale test. A final report should be provided after completion of all full scale testing. This testing may be completed with phase 1 in which case only four full scale tests are needed er with succeeding phases in the event as many as nine full scale tests are needed. Services to be provided by Sandia Infrared ".termovision will be taken of selected full scale tests by Sandia personnel with Sandia ecuipment. Sandia Labora-tories agrees to ptovide the necessary electrical cable and barrier materici to Underwriters Laboratories. C e

6 s e

• Ne n

i i i ~ [. i i i l lr j! aaa 6 4;;. ji i i i i i l i 4) i I !l i i si e i h' i a, I .{ l 1 l 1 t I' l 3 t ~- ! l, N I I II .t l - ---{.- a o !3 0 .s 9 d } 1 t I 2 l l 1 ii I h h h h l l /2* l I O ---3: gag t (,, 6 t_. y . set Ho f d' We l h (( e, /r.<y Ln:aWa men & s.J fOc<'ar v .6 i &p,ien to n w

e "I l i~/ 11 prualesoll i -. = = = =. ..9 .I l l l l 1, I; e. i j f VI o n _x .,_ e,.,.-, a [ I O ' / f r l C

• • t O

0 T.1,. .VY } l r.e at :. I. i g /.. l / 4 g. I r - gg --9 g /0/, -

• t h

=- < / t w v g I,. l i 48 [. i am 4. 6 3e" I+ ) l. sf 4 -- O, O, i f I 24' l s. C -- O i f (-....-._.. l: 1 Y _ I I w, K1M TC.h 'd ywo: r I vnti h a u e, ~, d I v s S

q o O /. [8 4 ? a af -[ I l 'i i i ! l. s' cl e h _s - l,' ' & / g..- $f A TNgl f q jl 'i p{_ 3 ;. 9 4 /'(d : /) g i ,-.g---- 7 ( 5 l 7s es =8 e j,* -W r- = f a g hS !h I : I !' l i I i k i NOTE,* C E El. IN(o" ltl Yf 3 ll2 I/) IE' Gr M 20/71 I s l l i b 4p O -FQ~ A e' rm ni / '-) j (., H,s T I gi M, lLv i \\ m. 1 'v \\i j , si r~ l' u g e \\(, gj j g c g,,j

8 4 kY &fff $ l h)f ano e I{ 'I EacA TFa w,Th /aur l .l /ky 7~c 's l 1 ! I O ! De*1 Dap3 i l I I I i< r x i ( ia l I i i i i l 7 t j i i ~ l l i i. L l4' l x e i l-

I e 1 + i ._q. l _ L. 5 e s x { L* (M' s.4 \\ Q; 'N, 4'. ves - x q w. t e O s 5, k

N7 8 e S N

9 >) % w,.s t e; s e = s* ~ g t t o .n en % (- ,h

• E,

k lM e g. s el y'st, s i e, i e w-4- ',t. e i q e i e i e i i l j i i i A C A r o o i, o-R d e m I w y' w l l h 5 i s ,f i w i a. y -t (1 N ') Y i e had ( E d( A 9 F. %. w ~3 < y e o n }q x% F 9 % s< n-o + he v 'i \\s i A b 9t ~ % 't 's Oy y s'

• c

,s _.y q g C. e-q g i 1 1 N .'7 j ,.g g F o! 'N O b D,' ,O ~ i o, c

• 7 c 3, -

,p I o i u, N ,,...I N at L. A vs y S. p / 'O i, >'1 %) [ t g S' i y q w M 2,,; v / e 3 < 4., m. _ y3 s s .x / g g, l 's ! t g kt g , R l/: e1

e e

f / 6 e T-(. . c.e 4 ; 3-v 0 .= .' R D! %w 's e w N / O t i 67 i-r _, %g / + e,j s l / s. i b +- i / 7 a \\ [ R C m O a t q j ;_ e +.., s f 9 i 7 e

• .o

., L' ( > r e s i 4 s O N iC q .( < e c / it e c N 5 s s O

g .-&12 ~ p JUE. 3M Mr. H. E. Roser. Manager Albuquerque Operations Office U. S. Department of Energy P.O. Box 5400 Albuquerque, New Mexico 0115

Dear Mr. Roser:

FY 1978 NUCLEAR REGULAT MY RESEARCH ORDER NO. 60-78-208 FOR SANDIA LA80RATORIES Please authorize Sandia Laboratories to execute the program described in L, the enclosed NRC Order.. \\\\v If this meets with your approval, it is requested that acceptance be i indicated on the w1osed forn and that the original be returned to this office and a signed copy sent to the NRC Controller. .\\. - Sincerely. ow sw as ,i T. E. Murley Thomas E. Murley Director Division of Reactor Safety Research Distribution

Enclosures:

subj

1. MRC Order circ

2. Program Brief chron Riggs:rdg cc w/ enclosures:

J. Lincoln, CON M. Sparks. Sandia R. Shumway, CON -7"A. Snyder. Sandia E. Case, NRR R. W. Barber. DOE /MSC T._Murley, RES H. G. Fish. DOE /ASEY:0PC C. Johnson, RES R. Feit, RES f R. Hoskins, RES M. Hayes, RES i e-09 '~ /, y,_A .b:A&PO N RSR: Wad R:WbS RSRiWh arrie = > RES:A&PC WDn Ric#s/ mig show/H< : 1s Feit N [ BdM John" son / kong k/1W /'[/14/78 M/lI/78' 6//5/I8" 6AN78 6/ /78 6/478 g.,,, { NRC Pout 318 (9 76) NICM 0240 Tr u. s. oovsamassser ramvme orrien sere -emeene

NRC *ORu 173 u S. NucLt AR RE GUL ATOR Y COMMISSION ~ ORDE A NUMBE R l (2 78) 60-78-28 l ^ STANDARD ORDER FOR DOE WORK JUL 3 1978 I ISSUED TO (DOE Office) ISSUED BY: (NRC Of fece) ACCOUNTING CITATION '^ ' " " ' Albucuera,ue Operations Office Division of Reactor Safety ""b"020.608 3 Research PERFORMING ORG ANIZATION AND LOCATION 60191001 FIN NUM BE R Sandia Laboratori.;s A1010-8 WORK PERIOD THIS ORDER FIN TITLE FIXED O ESTIM ATED O FROM. TO: Fire Protection Research 07-01-78 09-30-78 OBLIGATION AVAILABILITY PROVIDED BY: A. THIS ORDER S ]QQ,000 B. TOTAL OF ORDERS PL ACE D PRsOR TO THIS DATE WITH THE PERFORMING ORG ANIZ ATION UNDER THE SAME "APPROPRI ATION SYMBOL" AND THE FIRST FOUR DIGITS OF THE S ]2.911.000 S&E NUMBER" CITED ABOVE C. TOTAL ORDERS TO DATE (TOTAL A & B' S 13,011,000 0 AMOUNT INCLUDE D IN "C" APPLICABLE TO THE " FIN NUMBE R" CITED IN THIS ORDE R. S 400,000 FINANCI AL FLEXaBIOTY. O FUNDS WILL NOT BE REPROGR AMMED BETWEEN FINS. LINE D CONSTITUTES A LIMITATION ON OBLIGATIONS AUTHORIZED. g FUNDS MAY BE REPROGRAMMED NOT TD EXCEED 210% OF FIN LEVEL UP TO $50K. LINE C CONSTITUTES A LIMITATION ON OBLIGATIONS AUTHDRIZED. ' STAND ARD TERMS AND CONDITIONS PROVIDED DOE ARE CONSIDERED PART OF THis ORDER UNLESS OTH ERW:SE NOTED. ATTACH ME NTS. THE FOLLOWING ATTACHMENTS ARE HEREBY SE CURITY: MADE A PART OF THIS ORDER: O WORK ON THIS ORDER l$ NOT CLASSIFIED. 3 STATEMENT OF WOR K O WORK ON THIS ORDER INVOLVES CLAS$1FIED C ADDITION AL TF.RMS AND CONDITIONS INFORMATION. NRC FORM 187 IS ATTACHED. O OTHER R EM A R KS: 1sSUIM ApTHORITY A f,/ ACCEPTING ORGANIZATION SIGN ATU RE '/

• [

SbGNATURE Thomds. Murley irector Ta TL E V TITLE Division Of Reactor Safety Research WRC FORn4173 82 785 ..s we

FY 1978 PROGRAM BRIEF PROGRAM: WATER TITLE: TASK: CONT: FIN NO: A1010 FIRE PROTECTION RESEARCH CONTRACTOR: SANDIA SITE: ALBUQUERQUE STATE: NEW MEXICO NRC TECHNICAL MONITOR: RONALD FEIT PRINCIPAL INVESTIGATOR: LE0 KLAMERUS OBJECTIVE (S): TO PROVIDE DATA NEEDED TO CONFIRM OR MODIFY NRC FIRE PROTECTION REQUIRE-MENTS CONCERNING: (1) VULNERABILITY OF NUCLEAR POWER PLANT TO FIRE (2) CONTROL OF A NUCLEAR POWER PLANT FIRE (3) MITIGATION OF THE EFFECTS OF A FIRE OR NUCLEAR POWER PLANT SAFETY SYSTEMS, BUDGET ACTIVITY: 60191001 FY 1978 ADDITIONAL SCOPE (7/1/78 - 9/30/78): OBLIG - PREV OBLIG $ 300K THIS OBLIG 100K TOTAL OBLIG 400K 1. TO CONDUCT PROOF TESTS OF CABLE SYSTEMS. THIS INCLUDES FIRE BARRIERS AND FIXED AUTOMATIC FIRE DETECTION AND SUPPRESSION SYSTEMS AS WILL BE INSTALLED AND WED WITH TYPICAL VERTICAL CABLE TRAY CONFIGURATIONS. 2. TO CONDUCT SEPARATE EFFECTS TESTS ON VERTICAL CABLE TRAYS TO DETERMINE THE EFFECTIVENESS OF: A) CABLE TRAY SEPARATION B) FIRE BARRIERS AS USED ON AND BETWEEN CABLE TRAYS C) AUTOMATIC SPRINKLER SYSTEMS. e

Department of Energy Albuquerque Operations Office P.O. Box 5400 3050 Albuquerque, New Mexico 87115 JUL 171978 T. E. Murley, Director, Division of Reactor Safety Research, NRC NRC STANDARD ORDER NO. 60-78-208 Enclosed is a signed copy of NRC Form 173 acknowledging our acceptance of NRC Order 60-78-208. k D. K. Now Director LSP:MCB(608) Special Programs Division

Enclosure:

As stated cc w/ signed cy: L. W. Barry, Controller, NRC so'. Ill, 4 f' 4 ,e e r

1. 'Ye; a

f nV 4

e NRC *ORw 173 u s. nucle AR RE GULATOR Y COUutSSION OROE R NUM8E R ' (27P' 60-78-2G STANDARD ORDER FOR DOE WORK ^ JUL 3 1978 ISSUE D 70 IDOE Of ficel 8SSUE D BY. (N AC Off.ce) ACCOUNTING CIT ATION AppR OPRI ATION SvMBOL Albucueraue Operations Office Division of Reactor Safety Research 31X0200.608 m, y o,,,, PERFORMING ORG ANIZATION AND LOCATION 60191001 Sandia i.aboratories FIN NUMBE R A1010-8 WOR K PE RIOD. THIS ORDE R FIN TITLE fixed O lESTIM ATED O FROM. TO-Fire Protection Research 07-01-78 09-30-78 OBLIGATION AVAILABILITY PROVIDED BY: A THIS ORDER S 100,000 6. TOT AL OF OaCERS PL ACE D PRsOR TO THIS D ATE WsTH THE PERFORM 4NG ORG ANIZ ATION UNDER THE S AME ' APPROPRI ATsON SYVBOL" AND THE FIRST FOUR CiGITS OF THE S T &E NuvSER ~ CITE D ABOVE 12,911,000 C. TOTAL ORDERS TO DATE ITOTAL A & B1 S 13,011,000 0 AVOUNT INCLUDE D ;N "C** APPLICABLE TO THE " FIN NUMBE R" CITED IN THIS OROE R S 400,000 FINANCI AL FLExlBILITY. O FUNDS WtLL NOT BE REPRCORAMvED BETWEEN FINS. LINE D CONSTITUTES A LIMITATION ON OBLI AU THO RIZ E D. S FUNDS MAY BE REPROGR AYMED NOT TO EXCEED 21C% CF FIN LEVEL UP TO S50K ON CBLIGATiCNS AUTnORi2EC LINE C CONSTITUTES A LIM 6TATION STAND ARD TERMS aND CONDITICNS PRDv DED DOE ARE CONSIDERED PART OF THIS ORDER UNLESS OTHERW:SE NOTED. ATTA CH VE NTS. THE FOL OMNG ATTACHVENTS ARE HEREBY SE CURL TY: MACE A PART OF THIS ORDER: O WORK ON TH15 ORCER l$ NOT CLASSIFIED. 3 STATEMENT OF WOR K C ADDITICNAL TERYS AND CONDITIONS O WORK ON THIS ORDER INVOLVES CLASSIFIED INFORMATICN. NRC FORM 187 IS ATTACHED. O OTHER R E M A R KS: ISSUING AUTHORITY f_ f,/ ACCEPTING ORGANIZATION Thom]as[.Murlev/directort~/*v 7g y ;, W e A i / sicNATuRE sec e '~ mM-Ta te E V TsTLE ision of Reactor Safety Research ). K. Nowlin, Director, Special Prograr_s Div. F...#Ru 373 e2. Tai ^" g 17.10ZS. _ u_

/ ( 3 o Mr. H. E. Roser, Manager Albuquerque Operaticns Office U. S. Department of Energy N0y 15 tn7 P.O. Box 5400 Albuquerque, New Mexico 87115

Dear Mr. Roser:

FY 1978 NDCLEAR REGULATORY RESEARCH ORDER NO. 60-78-026 FOR SANDIA LASORATORIES Please authorize Sandia Laboratories to execute the program described in the enclosed HRC Order. If this sneets with your approval, it is requested that acceptance be indicated on the enclosed form and that the original be returned to this office and a signed copy sent to the NRC Controller. ' /,. Sincerely, onginaf signed tg T. L Murfey Thomas E. Murley, Director Division of Reactor Safety Research

Enclosure:

DISTR: NRC Order J. Lincoln, CON R. Shumway, CON cc w/ enclosure: C. Beckwith, CON M. Sparks, Sandia E. Case, NRR A. Snyder, Sandia S. Fabic, RES m. R. Barber, DOE /NCS W. Johnston, RES UA3 H. G. Fish, DOE /AES:0PC G. Bennett, RES L. S. Tong, RES 9 T. Murley, RES (2) M. Hayes, RES Subj Circ Chron Sokolow:rdg / y h tn / / b......- I e5:n ec_ Je: ' +9 i Ha Fabic [yf. JEsla_y5.;Sa_ gsR_ masz. - 3f_ DIN Johnston nett/ Tong Murley _..gk/77 Y 7/77 10/ \\ 10/20/77 10/ c/77 /0 / p /77 ll /lh77

, 173 v s ave t an an s vi.at o = * ::== ssio= omst a avw ss a 60-78-026 DATE STANDARD ORDER FOR ERDA WORK NOV 15 577 issse : te ia= =... on... rssst: a T. c a.- c A :ovaTm :: Tan a e.ac....rio s w.o w Division of React:r Albucuerque 0:erations Office 535,;v .eco.-w 31X0200.603 Ps aa : n w.= = : :n a A T a..se e.es w ee....

• • • • w"***

See Attacked em =vusta Sandia Laboratories ga. 3...,w s Pth T1TLE WWO a t P'E Rio:. TMrs OR D( a paow To See Attached 10/1/77 9/30/78 es-06LIGAT1CN AV AILABILITY PROVIDED BY: A. rwisca:ta s 2,181.CC-J s T:T4. :' ca:Ea5 e a:E; perca T: Ns ca Ti nit = T= E PE a s ca w,%;, Oc% Tea:T:a us:ta !=t sawi a"a:>a.aT40% svus: A8:. Tat Sims? 5 0va :ois :3 mt saa %,;weem :iTE: as:ve 5.811.000 1

. to AL oa:Ea s i: : ATE (TSW a & SI g

7.991.000 o w:vs? m: v:t: m r ara. :as.t : Tat -sis wusta :: 10 8 mis emota - See ' M ed 'n,s,,o.,nr sesr,u.a. e in are., e canar.ws a.cor.cee ta se seeme ro* re wef.c s' ra soa* = =e am 's ~~tr-atuAncs Sandia may adjust funding prior to fiscal year end without RSR approval up to +10" not to exceed $50K for any FIN No. pr0vided that the dollar total of all RSR orders for Sandia frca the six digit S&R category (601910) is not exceeced. Requests for approval of changes outside these limits require adequate justification and RSR approval. ACOEPTING CGCANilATION Iss;.JtNG AUTMCRITy l / ,,f' / b!s OROER is AOCEPTABLE T w, , M'.Luu C\\ / if si %ATvat ^ Th~ 2 e E. Puffev. Di" ecto # - OCO I' 0 "' T TLE g c,o. TiTb1

3. 4. Newli:. Dire:::::

Dt ST RI B U*1C N. g eein p.n.3 3 - i,, m i

=c%.fms:ss.:t.8==

ggy, g ;g77 e,1, O Orts t cs Twt : winct;ga,%a: N.1C s c a w 173 0 2 75: . - -. -. - agf - e.-. i eowes i <.t a unnntion i W.n e, a+ + / *a - 1 14 ' '. 4 "'

,Q 3N 3 T g?\\ s -m b f Sandia Laboratories FY 1978 B&R No. Fin No. Title Obligation 601910012 A 1010-8 Fire Protection Research 300K 601910012 A 1051-8 Quality Testing Evaluation 675K m 601910012 A 1194-8 iluman Engineering Consultation 10K 60191003 A 1205-8 Statistical Analysis 135K 60191003 A 1207-8 Ulli RELAP Hodel Development 135K 601910041 A 1019-8 Holten Core Interactions 385K 601910041 A 1030-8 Steam Explosion Phenomena 450K 60191J041 A 1206-8 Resident Engineer - Karlsruhe 91K TOTAL 2,181K

5565 B Depa.tment of Energy Albuquerque Operations Office P.O. Box 5400 Albuquerque, New Mexico 87115 NC'l ' a 1977 T. E. Murley, Director, Division of Reactor Safety Research, NRC NRC STANDARD ORDER NO. 60-78-026 Enclosed is a signed copy of NRC Form 173 acknowledging our acceptance of NRC Order No. 60-78-026. U mw D. K. Nowlin, Director g LSP:GWJ(1014) Special Prograss Division 4'

Enclosure:

As Stated cc w/ signed cy: R. J. Friedman, Controller, NRC m 7.~. s: m .rj; h: &, ? &r Q! ' V >, C@ . ?.~f.._, q \\ s a us l I)w 1 h$ Ea:. Y . hsd

FY 1978 PROGRAM BRIEF PROGRAM: RATER TITLE: TASK: CONT: FIN NO: A1010 FIRE PROTECTION RESEARCH CONTRACTOR: SANDIA ~ SITE: ALBUQUERQUE STATE: NEW MEXICO K8C TECHNICAL MONITOR: RONALD FEIT PRINCIPAL INVESTIGATOR: LEO KLAMERUS OBJECTIVE (S): TO PROVIDE DATA NEEDED TO CONFIRM OR MODIFY NRC FIRE PROTECTION REQUIRE-MENTS CONCERNING: (1) VULNERABILITY OF NUCLEAR POWER PLANT TO FIRE (2) CONTROL OF A NUCLEAR POWER PLANT FIRE (3) MITIGATION OF THE EFFECTS OF A FIRE OR NUCLEAR POWER PLANT SAFETY SYSTEMS. BUDGET ACTIVITY: 60191001 FY 1978 SCOPE (7/1/78 - 9/30/78): OBLIG - $100K 1. TO CONDUCT PROOF TESTS OF CABLE SYSTEMS INCLUDING FIRE BARRIERS AND FIXED AUTOMATIC FIRE DETECTION AND SUPPRESSION SYSTEMS AS WILL BE INSTALLED AND WED WITH TYPICAL VERTICAL CABLE TRAY CONFIGURATIONS. 2. TO CONDUCT SEPARATE EFFECTS TESTS ON VERTICAL CABLE TRAY'S TO DITERMINE THE EFFECTIVENESS OF: A) CABLE TRAY SEPARATION B) FIRE BARRIERS AS USED ON AND BETWEEN CABLE TRAYS C) AUTOMATIC SPRINKLER SYSTEMS. MM

e loaeo,4 u C j UNITED STATES y'w NUCLEAR REGULATORY COMMisslON j j f)j WASHINGTON, D. C. 20555 's October 26, 1978 NOTE T0: L. S. Tong L. C. Shao C. N. Kelber S. Fabic W. Johnston / G. L. Bennett 1 R. Curtis BURSSELS PAPER Enclosed for your information is a final version of the paper and slides that I presented at the Brussels ANS/ ENS Conference. hhA Tom Murley

Enclosure:

As stated

VERIFICATION OF REACTOR SAFETY CODES BY THOMAS E. MURLEY DIRECTOR, DIVISION OF REACTOR SAFETY RESEARCH U.S. NUCLEAR REGULATORY COMMISSION WASHINGTON, DC 20555 FOR PRESENTATION AT ENS /ANS MEETING ON NUCLEAR POWER REACTOR SAFETY BRUSSELS, BELGIUM OCTOBER 1978

ABSTRACT iERIFICATIONOFREACIORSAFETYCCDES by Themas E. Murley U.S. Nuclear Regulatory Cenmission Washington, D.C. 20555 The safety evaluation of nuclear power plants requires the investigation of a wide range of potential accidents that could be postulated to occur. Many of these accidents deal with phenomena that are outside the range of normal engineering experience. Because of the expense and difficulty of full scale tests covering the cceplete range of accident conditions, it is necessary to rely on complex cceputer codes to assess these accidents. The central role that computer codes play in safety analyses requires that the codes be verified, or tested, by comparing the code predictions with a wide range of experimental data chosen to span the physical phenomena expected under potential accident conditiens. This paper discusses the plans of the Nuclear Regulatory Ccamission for verifying the reactor safety codes beir.g developed by NRC to assess the safety of light water reactors and fast breeder reactors. O e

VERIFICATICM OF REACTCR SAFETY CODES Themas E. Murley I. INTRODUCTICN Over the years the Nuclear Regulatory Commission has evolved the safety philosophy of defense-in-depth. The first line of safety defense is to make sure the plants are designed and built c0rrectly in the first place -- that is, t0 be sure the design, the materials, the fabrication metheds, the construction practices and the testing and operatien are of very high quality. The seccnd line of safety defense is to provide protective systems to shut down the reacter plant in a safe condition in the event of equipment failures or breakdowns which can happen from time to time. To provide a third line of safety defense, the NRJ staff postulates that sericus accidents happen in spite of their very lcw probability, and then requires engineered rafety features to mitigate the consequences of even these icw probability accidents. Thus, the safety evaluation of nuclear power plants requires the investigation of a wide range of potential accidents that could be postulated to occur. P.any of these acciden s deal with phencmena that are outside the range of normal engineering experience. For example, one of the severe accidents postulated by the NRC staff for a Pressurized Water Reacter (PWR) is the instantanecus dcuble-ended rupture of the largest inlet cooling pipe leading to the

. reactor vessel. The staff requires that the plant design include a number of features, including emergency core cooling systems, to make sure that the consequences of such a loss-of-coolant accident (LOCA) are within acceptabls limits. Similarly, one of the accidents postulated for a Fast Breeder Reactor (FBR) is a loss-of-coolant flow to the core coupled with a simultaneous failure of the reactor shutdown system. The staff requires suitable engineered safety features in the plant design to make sure that the consequences of this accident, tenned a core disruptive accident (CDA), are within acceptable limits. In order to assure that the plant safety features are adequate to mitigate these postulated accidents, the NRC sponsors a broad program of confirmatory safety research. Because of the expense and difficulty of CJnducting full scale tests Covering the complete range of accident conditions that could be postulated, it is necessary to use sophisticated computer codes to evaluate these accidents. The central role that computer codes play in safety analyses requires that the codes be verified, or tested, by comparing the code predictions with a wide range of experimental data chosen to span the physical phenomena expected under accident conditions. The need to verify the quality and reliability of safety computer codes arises both from our desire to understand the safety margins in nuclear plants and from our desire to ensure public acceptance of the basis for making safety judgments.

. In this paper,the term code verification is synonymous with code assessment: to determine how well a code is capable of simulating the physical processes observed and measured in experiments; to assess the code's applicaatlity to analysis of accidents in full scale plants; and to assess the uncertainty associated with the code's prediction of important safety parameters. Further information on NRC's code development programs can be found in References 1-4, and on our code verification plans in References 5-7. II. THE DEVELOPMENT, TESTING AND APPLICATION OF SAFETY CODES NRC's plan for the development of advanced safety codes follows three phases as illustrated in Figure 1. The first phase, com-prising the bulk of the effort, is to develop the overall framework and the detailed physical models embodied in the code. Durir.g the developmental phase, before the code is publicly released, many comparisons are made with the available test data, and models in the code are changed when necessary. Various sensitivity studies are performed to identify important parameters and in some instances to point out where additional research data may be needed. The emphasis during developmental checkout is on (a) cumerics, (b) models of physical processes described by the constitutive equations, (c) system component behavior, and (d) integral system behavior. Once the code developers believe they have a satisfactory

. version of the code, it is frozen and all cases are rerun prior to documentation and public release of the code. The developers will then usually begin working on a newer version of the code which em-bodies more sophisticated models, improvemencs in numerics, improve-ments in user convenience, or inclusion of models applicable to different reactor systems (such as BWR vs. pWR for example). The second phase of the code development process is an assessment of the code by an indt9endent team. We believe that this ',tep will help ensure an objective evaluation of the accuracy of tie code, of the completeness of the code manual, and of the user convenience of the code. In this independent assessment phase, en.phasis is placed on blind predictions of tests conducted in new test facilities and, also, significantly changed test conditions or system configur-ations in older test facilities. The purpose here is to exercise the code over a wider range of test conditions than was used in the developaent phase and thereby test the true predictive capability of the code. During this activity the results will be communicated to the code developers in order that follow-on versions of the code can correct any weaknesses found. The output of the independent assessment phase will be a code assessment report describing all of the comparisons of code predictions with test data, evaluation of code error, and extrapolation of that error to large scale plants.

. When the NRC staff is satisfied that a version of the code has been adequately tested against experimental data and that the code meets NRC acceptance criteria, the code will be used in the application phase to aid the staff in assessing margins in the safety design of nuclear plants. III. EVALUATION OF UNCERTAINTIES One of the primary aims in developing advanced safety codes is to permit NRC to evaluate regulatory safety margins. To illustrate how this may be done, let us consider a comparison o ' a licensing evaluation model (EM) calculation with an advanced best estimate (BE) calculation of the peak cladding temperature (PCT) reached in a loss-of-coolant accident in an LWR. In order to make this com-parison, one must know the uncertainty bnnds on the best estimate calculation. There are three sources of uncertainties that must be accounted for: 1. Code Model Errors (CODE ERROR) physical models may be wrong or incomplete numerical solution errors stochastic phenomena not represented etc. 2. Code Data or Built-In Coefficient Uncertainties (DATA) fuel thermal conductivity fuel-clad gap conductance decay heat

• heat transfer momentum exchange etc.

3. Reactor Plant Condition Uncertainties (REACTOR) availability of off-site power core conditions at time of accident functionability of safety systems

• etc.

One can begin by estimating the contribution of data uncertainties to the overall best estimate calculation uncertainty. This is done by assuming the code models are correct and considering fixed re-actor conditions. Prior code sensitivity studies can tell us which parameters significantly influence the calculated peak clad temper-ature, while plots of basic test data on physical phenomena can tell us the uncertainty range of interest for each important parameter. One then carries out a large number of computer calculations in which the selected parameters are varied over the established range of un-certainty, using a prescribed sampling procedure. The result is a Response Surface -- a complex algebraic expression defining the effect of each parameter, xj, on the peak clad temperature. Using tae Response Surface in conjunction with Monte Carlo sampling from the individual probability distributions, p(xj) for each selected parameter, one finally obtains a probability distribution function for the peak clad timperature, P (T) vs. T, which represents the un-d certainty in the calculation due to code data uncertainties. Calcu-lations of this type are already being carried out on a trial basis using older, less sophisticated codes. The evaluation of uncertainty from code errors is less straight-forward, and we have not yet worked out the complete formalism. During the independent assessment phase, a large number of com-parisons between calculations and measurements will be made. If one plots the measured vs. calculated values of selected parameters from all comparisons, the result will be a series of scatter plots illustrated schematically in Figure 2. If the calculated values agree perfectly with the measurements, all of the points would fall on a 45 line. There will be a scatter of points, of course, and this scatter will be due to measurement uncertainty and code data uncertainties as well as code error. The next step is to strip out the measurement uncertainty and the code data uncertainties to get a statistical assessment of code error. The goal here is to obtain a probability distribution function of peak clad temperature, P (T) c vs. T, which represents the uncertainty in the calculation solely due to code errors. In the next step, the code error probability distribution function P (T) is convoluted with the code data uncertainty probability dis-c tribution function P (T). The resulting probability distribution d function Pcd(T) represents the uncertainty in calculated peak clad temperature due both to code errors and to data uncertainties. In assessing the safety margins for an actual reactor plant, one must also account for uncertainties in the reactor conditions before and during the accident. This uncertainty can be obtained by sampling from probability distributions representing the possible combination of reactor conditions, and, for each set of conditions, making best estimate calculation of the peak clad temperature. We know, however, that each calculated best estimated PCT has its own proba-bility distribution due to code errors and data uncertainties. The resulting combination of code errors, code data uncertainties and re-actor condition uncertainties is a probability surface represented by Figure 3. Having obtained the PCT probability surface, one can obtain the probability of exceeding a given temperature T by integrating the volume under the surface intersected by the plane T = constant, and dividing by the total volume under the surface. The final result can be displayed as a probability distribution function as shown in Figure 4. An assessment of the margin of safety can be made by comparing the best estimate of peak clad temperature (BE) with the licensing evaluation model limit (EM) and calculating the probability of exceeding the licensing limit. IV. CURRENT STATUS OF ADVANCE 0 SAFETY CODES The foregoing discussion has outlined NRC's plan for developing, testing and applying advanced safety codes. Although we have several safety codes under development, the principal advanced LWR

9_ system code is TRAC (Reference 3). An initial version has been developed and checked out, and an independent assessment is beginning, with the emphasis being on blind calculations of Semiscale-MOD 3, LOFT Nuclear Tests, PXL-Core 2, and LOBI. The final detailed un-certainty analysis and assessment of margins will not be concluded until all test data have been obtained from the planned LOCA/ECCS test program and the final version of TRAC has undergone independent assessment. This process will clearly take several years to complete. With regard to advanced reactor safety codes, Figure 5 shows the codes NRC expects to use to analyze core disruptive accidents in fast breeder reactors. The principal CDA analysis code is SIMMER (Reference 4), and an initial version has been released and is under-going checkout testing. Independent assessment of the advanced re-actor safety codes has begun to a limited extent, but it will be many' years before there is a data base adequate to permit assessment of regulatory margins in the same detail as for LWR's. REFERENCES 1. Fabic, S., " Review of Existing Codes for Loss-of-Coolant Accident Analysis," Advances in Nuclear Science and Technolocv, Vol.10, (Plenum Publisning Corp., 1977) 2. Fabic, S., " Computer Codes in Water Reactor Safety: Problems in Modeling of Loss of Coolant Accident," Paper No. C201/77, Conference on Heat and Fluid Flcw in Water Reactor Safety, Septemcer 13-15, 1977, Manchester, UK., Proceedings Puolisnec ey Institution of Mechanical Engineers. 3. Jackson, J. F., et. al., " TRAC-Pl: An Advanced Best Estimate Computer Program for PWR LOCA Analysis; I. Methods, Models and User Information, and Programming Details," NUREG/CR-0063, LA-7279-MS, Vol.1, June 1978. 4. Bell, C. R., et. al., " SIMMER-I: An Sn, Implicit, Multifield, Multi-component, Eulerian, Recriticality Code for LMFBR Disrupted Core Analysis," LA-NUREG-6467-MS (January 1977). 5.

Fabic, S., " Accident Analysis," Chapter 6.6 in Handbook of Multichase Systems, edited by G. Hetsroni, Hemisphere Publishing Co., (to be published in 1979).

6. Basdekas, D., Silberberg, M., Curtis, R., and Kelber, C., "A Qualifi-cation Testing Program Plan for SIMMER," USNRC Publication NUREG-0457 (July 1978). 7. Bell, C.R., et. al., " SIMMER-II Analysis of LMFBR Post-Disassembly Expansion, these proceedings. O

5 G PHASES OF CODE DEVELOPMENT, TESTING AND APPLICATION AND ASSESSMENT ANALYSIS CHECKOUT e FIGilRE 1

CODE TESTING Example: Measured PCT vs. Calculated PCT in SEMISCALE, LOBI, PKL and LOFT bc. INDICATES + ++ SYSTEMATIC c: + + ERROR + + y + + CALCULATED PCT / N / 6 / w + g a f', + ->, /, y 9 + CC / o /+ + +,/ + /, - f /,s', CALCULATED PCT I l i CODE i P(T) l ERROR l l T

PEAK CLAD TEMPERATURE PROBABILITY SURFACE d COMBINED PROBABILITY 's N \\ ' y PROBABILITY OF PLANT CONDITION ,/ '- -f \\' h PCT PROBABILITY DUE TO / ,/ f CODE UNCERTAINTY PLUS j j ,p CODE ERROR e L& l@ /h

• f* #

M*

1. l#, s* j s 9 #

CURRENT EM LIMIT (22000Fi / d FIGURE 3

• /e

ESTIMATION 01: MARGIN e REACTOR CONDITION UNCERTAINTIES e CODE MODEL UNCERTAINTIES e CODE DATA UNCERTAINTIES I I I l l P(T) l l I l l l l l 1 I I BE EM T FIGURE 11

CORE DISRUPTIVE ACCIDENT ANALYSIS LOSS OF CODES OVERPOWER COOLING SAS-EPIC SAS-LAFM REACTOR TRANSIENT BENIGN TE R MIN ATION NUCLEAR SIMMER DISASSEMB LY 6'M PRIMARY SYSTEM DAMAGE RE O Y AEROSOL BEHAVIOR HAARM CONTAINMENT INTEG RITY CONTAINMENT CODE fir-URE 5

PHASES OF CODE DEVELOPMENT, TESTING AND APPLICATION e M ^ DEVELOPMENT INDEPENDENT T A AND ASSESSMENT CHECKOUT ANALYSIS i 9

RESEARCH PLAN FOR DEVELOPING LWR SAFETY ANALYSIS METHODS Separate Effects Tests Integral System Tests e Biowdown Heat Transfer e LOFT e ECC Bypass e Semiscale o Reflood Heat Transfer e Pump Characteristics e Core 3D Effects Component Codes Tests Tested ure ECCS System Code Development ECCS Code Iterations A Code Fuel Codes Development Fuel Behavior Tests

• PBF

k EXPERIMENTS FOR ASSESSING LWR SAFETY CODES SEPARATE EFFECTS TESTS FUEL BEHAVIOR TESTS e Blowdown Heat Transfer e Multirod Burst Tests e Reflood Heat Transfer e Basic Materials Tests e ECC Bypass e PBF Tests e Pump Characteristics e NRU Tests e Core 3-D Effects o ESSOR Tests e BWR CCFL Tests e Upper Plenum Tests INTEGRAL SYSTEM TESTS e LOFT e Semiscale ePKL e LOBI e CCTF O

u.. - _ I 5 I 4 E 3 11 ~

a.;

+ /. g N 2 337 = / / B 5 / 25 d.4 / 3$ g5 ~ a }j: j / S E

j 4

/ / j; , '[l j 82- / / s ;]t / / = ,a. / / I s3 5 s / e / / s' ./ 3 / /# - t i e t E "7 3 E M s s s

• 3 / 4 288*4 M onet to sop oeg l

l l l f / /

T .i TEMERATURE(C) 1500 1400 1300 1200 1100 1000 900 t.... c, r s.. =__ s-. .r.. 3 '----S.x a;.- %..- N. ~~ i;t = Wars =__.-m.. > -- = --r -+3 2-s # ; s.+... lEl x _- Z =-V iri -w - ' -- 4,---- = u-3.,---.- B AKER-JUST _# : := W f:-WSi ._Il i o.. --y.. j.... =.. =... ----:---=--:...._... n \\g:. v +. t--..~ - -- - -- .--:------=. E - 10,g _.--. g. y E. .ss_ a j.v T- [ 3O N w r... En. WPI..NN \\ ~..~ g ,g g. w g, _ s E4L \\ g u.. C3 g t_. X .6 g i - A \\; g

==f= = ~; h =- i:.- 4-o .%g. it =-- .i y.;

i.

--I.. - t -

..u . yM.x... :.. _ - - -.. w - +- =. - .C', 2 .-.-.- =.. z._... _. :... - _.;... --_-*--=:=__.- 10,,7 W s s Q W sg. .,-a ~ wx y. -m . x.v., g_....... - = .. - - - - L .f_ o -.- a.--~ =. -f-N-:.R : -- ' v w ._-.w--- =- ORNL o w =-. -m - rs". E HEIDRICK ~ - m oi ,..== -=.m.-. _ -1_= - =_.a. - -m - =\\4_5e (CANADIAN DATA - -===- w- _ - LEISTIX0W - - --- Z' :

=. E (GERMAN CATA) T "-

- Osii - --iTW=h x --E-~= =a -E =: E : -@ ' - M XAWASAXI 4 (JAPANESE DATA) = 10-8 s- - .-.n. r =.- y, ?' g; ., v-5---:_r_-+ , --jE. .u=L M. O -: , _ M-.-: '::. -d h._ _-? - : - E- ' '. , -- :..... p -... w. 4: : - - --'^t__ _Q-2. 0.60 0.65 0.70 0.75 0.20 0.85 1000/T (X-1) ARRHENIUS PLOT TOTat OXYGEN RATE CCNSTANT I I FOR STEM-ZIRCiLOY 4 REACTION \\ /

PWR ECCS PERFORMANCE 2500 - NRC EVALUATION MODEL / (-2200 F) g-2000 - 1 IE 3 ~_s IJ 1500 - 's% x y ,/ 2 m / REALISTIC ESTIMATE 's j" \\ ./ (-1500 F) 's V 's g 1000 - o E h" S00 '- 0 I I I I I I I I I O 20 40 60 80 100 120 140 160 180 200 TIME AFTER BREAK (seconds) S

SENSITIVITY STUDIES f NOT AN IMPORTANT f f,/, PARAMETER / TEMP i g l,' \\/ f i TIME s ~/ ,, 7 e' IMPORTANT PARAMETER lt I'/ . TEMP ts /t /~'N t l l, l' TIME

UNCERTAINTY STUDY FOR CALCULATED PEAK CLAD TEMPERATURE n \\ RESPONSE SURFACE .x, m p(X ) 2 F (X, X,.., X,, ) y 2 P iecT) p(X,,) ( X,, PCT COPE DATA UNCERTAINTIES

CODE TESTING Example: Measured PCT vs. Calculated PCT in SEMISCALE, LOBI, PKL and LOFT b c. g INDICATES + c: + + SYSTEMATIC ] + + ERROR + + + + CALCULATED PCT / / i._ / N / + '+ + / 9 cc / g, 3 /+ + h +,/ + / f /,/, CALCULATED PCT I i l CODE P(T) l ERROR l l t T ,. a

M PEAK CLAD TEMPERATURE PROBABILITY SURFACE a COMBINED PROBABILITY - s T _f. [ N, \\ y PnOBABILITY OF PLANT CONDITION -,s / y / \\ 'ih[ PCT PROBABILITY DUE TO / y[ i / CODE UNCERTAINTY PLUS f / jgo cOoeeanOn p s + Nm 49 ,4s9 AiP

1. j#

s eo o ol# s* j d m f#' ceeeeNT _ L,.,T _ 4 g % ::3 fo, p

k ESTIMATION OF MARGIN e REACTOR CONDITION UNCERTAINTIES e CODE MODEL UNCERTAINTIES

• CODE DATA UNCERTAINTIES 1

I I I I P(T) l I I I I I I i I I i BE EM T

VdHEN TO STOP7 I I I I I I I l P(T) I i PROBABLY OK TO STOP l I I I I I I BE EM T l 1 I 1 l' I l j i LICENSING ACTION INDICATED p,,, I I I I I I I I BE EM T I l l l l PIT) I PROBABLY NEED MORE WORK I I I I ^ DE EM T

CORE DISRUPTIVE ACCIDENT ANALYSIS >VERPOWER LOSS OF CODES COO LING r SAS-EPIC SAS-LAFM REACTOR TRANSIENT .sW4 ~ I~ BENIGN TERMINATION NUCLEAR SIMMER DISASSEMB LY PRIMARY SYSTEM DAMAGE 6iMMER RE C AEROSOL BEHAVIOR HAARM CONTAINMENT INTEG RITY CONTAINMENT CODE O ~ue e.}}