ML18029A183

Forwards Response to 840726 Request for Addl Info Re Mark I Containment long-term Program plant-unique Analysis Rept

ML18029A183
Person / Time
Site:	Browns Ferry
Issue date:	10/11/1984
From:	Mills L TENNESSEE VALLEY AUTHORITY
To:	Harold Denton Office of Nuclear Reactor Regulation
References
NUDOCS 8410240289
Download: ML18029A183 (102)

Text

REGULATORY l rORMATION DISTRIBUTION SYS M (RIDS)

ACCESSION NBR;8410240289 DOCiDATE! 84/10/11 NOTARIZED:

YES FACIL;50-259 Browns Ferry Nuclear Power Stationi Unit ii. Tennessee 50 260 Browns Ferry Nuclear Power Stationi Unit 2i Tennessee

'50 296 Browns Ferry Nuclear Power Stationi Unit 3<

Tennessee AUTH'AME AUTHOR AFFILIATION MILLsiL,M, Tennessee Valley Authority RECIP ~ NAME'ECIPIENT AFFILIATION DENTONgH ~

RE

'ffice of Nuclear Reactor Regulationp Director

SUBJECT:

Forwards response to 840726 request for addi info =re Mark I containm nt ion ter program plant-unique analysi-s r ept-,

DISTRIBUTION CO:

A02 'PIES RECEIVED!LTR f ENCL 3 SIZE:

TITLE:

OR Submi tal:

USI A"7 Mark I Containment NOTES:NMSS/FCAF 1cy ~ icy NMSS/FCAF/PM'Le06/26/73 NMSS/FCAF icy'cy NMSS/FCAF/PM'L:06/28/74 NMSS/FCAF,icy, 1cy NMSS/FCAF/PM'L:07/02/76 DOCKET tt 05000259 05000260 05000296 05000259 05000260 05000296 RECIPIENT ID CODE/NAME NRR ORB2 BC 01 INTERNALe ADM/LFMB ELD/HDS4 13 NRR. SIEGELIB NRR/DE/MTEB NRR/DL/TAPMG Q4 COPIES LTTR ENCL Q

1 0

1 0

5 5

1 1

1 1

1 1

REC IPIENT ID CODE/NAME EDO NRR DIR NRR/DE/MEB NRR/DL/ORAB 10 NRR/DS I/CSB 11 RGN2 COPIES LTTR ENCL-1 0

1 0

4 4

1 1

1 1

1 1

EXTERNAL; ACRS NRC PDR NTIs NDTEs:

12 10 3

02.

1 1

1 1

LPDR NSIC 03 05 1

1 1

'TOTAL NUMBER OF COPIES REQUIRED:'TTR 39 ENCL 24

n

~

~

~

~

~

I l(

TENNESSEE VALLEYAUTHORITY CHATTANOOGA. TENNESSEE 87401 400 Chestnut Street Tower II October 11, 1984

~!

Mr. Harold R. Denton, Director Office of Nuclear Reactor Regulation U.S. Nuclear Regulatory Commission Washington, D.C. 20555

Dear Mr. Denton:

In the Matter of the Tennessee Valley Authority Docket Nos. 50-259 50-260 50-296 By letter from D. B. Vassallo to H. G. Parris dated July 26,

1984, we received a request for additional information regarding the Mark I Containment Long-Term Program Plant Unique Analysis Report (PUAR) for the Br owns Fevry Nuclear Plant.

In response to your request, meetings were conducted at TVA's offices in Knoxville, Tennessee, on September 5 and 13, 1984.

Enclosed are the responses to the request for additional information.

The enclosure reflects discussions that weve held with your staff after the.

meetings.

Based on discussions with your staff, we believe that the responses provided here are acceptable for resolution of your staff's concerns.

If you have any questions, please get in touch with us through the Browns Ferry Project Manager.

Very truly -yours, TENNESSEE VALLEY AUTHORITY L. M. Mills, Ma ager Nuclear Licensing Subscribe sworn to fo e me 'this da o

1984.

otary Public My Commission Expires Enclosure cc:

See page 2

84i0240289 84iOii PDR ADGCK 05000259 P

PDR An Equal Opportunity Employer

c~q

'V Vj I

Cg (J

Mr. Harold R. Denton N

October 11, 1984 co (Enclosure):

U.S. Nuclear Regulatory Commission Region II ATTN:

James P. O'Reilly, Regional Administrator 101 Marietta Street, NW, Suite 2900 Atlanta, Georgia 30323 Mr. R. J. Clark Browns Ferry Project Manager U.S. Nuclear Regulatory Commission 7920 Norfolk Avenue

Bethesda, Maryland 20814

~

~a 1

ENCLOSURE

RESPONSE

TO D. B. VASSALLO'S LETTER TO H. G.

PARRIS DATED JULY 26, 1984 MARK I CONTAINMENT LONG-TERM PROGRAM PLANT UNIQUE ANALYSIS REPORT LOADS EVALUATION BROMNS FERRY NUCLEAR PLANT Attachment 1 - Responses to Franklin Research Center Questions Attachment 2 - Responses to Brookhayen National Laboratory Questions

~ WP

~ '

ATTACHMENT I BFN PUAR FINAL T.VA RESPONSES IO NRC'ND'RANKLIN RESEARCH CENTER QUESTIONS.

Genera 1 Response to PUAR Ques t i ons BFN LTP analysis and design activity has proceeded on a

schedu l.e necessary to suppor t i'nsta 1 la t ion of a l,l mod i ficat ions dur iing the Cycle 4 and 5 refuel ing outages of each unit, as required by NRC.

The

..first BFN Cycle 4

refueling outage began in Apr i 1

1981, and most of.

the major mod i f icat:ion designs were complete by May 1981.

Remaining mod i f ica t ion designs, pr imari ly for torus at tached pip ing external supports were complete in, time to support installation during the Cycle 5 refueling outages.

In order to satisfy schedule commitments, it was necessary to make interpretations of LDR 'and NUREG 0661 requirements based upon the 'best available information ~at the time of analysis Most of the interpre.tat'ions ivere originally established in 1979 and early 1980.

A continuing effort to remove excessive conservatism from load definiitions and analysis method.

was made, particularly when that conservatism would result in unnecessary, impractical modifications.

When la ter analysis me prev'ious in for reanaly

.unconservat For example resulted fr oscillation in forma t iion on load de fin i t ions and assoc ia ted thods became available, it~ was compared to the ter pretations.

The later information was used s is and assoc ia ted des ign work if a s igni fican t ism in the previous interpretation was indicated.

the fina1 downcomer t iebar/v-brac ing mod i f ica t iion orn November 1981 changes in the DBA condensa't ion

'a tera 1

load de fin ii t ion.

Sometime:s, la ter information was used to remove excessive conserva t ism in rema in ing analysis and des ign work.

For

example, the:l.

1 SRSS load combination technique was permit ted for.torus at tached piip ing ana)lysis af ter NRC's final po:sit ion on this subject was def ined in April 1983 by PUAR Reference 58.

An absolute sutrsm<t ion combination technique was required pr ior to t,hat t irne.

Finally, when the later information showed the previous load definitions and analysis methods to be adequately (but not excessively) conservative, the original interpretations'ere retained.

In these siituations reanalysis utiliz'ing th'

'later information would have been unnecessary and costly,

and, in some

cases, would have resulted in delays in th' installation of modifiic'ations.

Many of the E>UAR questions der ive from:si t uat original inte rpretations stated in PUAR Sect i for analys is.

Justi f iicat ion for these interp provided in E~UAR Section 4,

Sec ti,on 5, and Ap Addi t ional technical

.just ifica t ion follows in to specific questions on these topics.

Other simply request additional information, whiich the respon,ses.

ions where the on 4 wer e

'used'etationswas pendix C.

the responses'UARques'tions is provide'd in GR-1 PUAR.04

I t is TVA's posi t ion that the BFN PUAR (Revis ion 1')

and our review quest ion responses demonstrate compliance with the intent of the Mark '1 Conta inment Long-Term Program and NUREG 0661

( i.e.,

to upgrade the containment system safety margins, for 'a 1 1 pos tu la ted hydrodynamic load i ng. cond i t ions, to those intended by the or iginal design speci f ica t ions).

On this

basis, we feel that all indicated safety concerns are fully and sa t is factor i ly addressed, and we respec t fu 1 ly r eques t a

favorab le final evaluat ion for the BFN LTP..

GR-2 PUAR.04

tTEM Provide a more deta i led'escr ip t ion of the vent,system ana lys is regard ing downc'orner 1'a tera 1 'loads (Sec t ioii 4.4.5 (5))

0

RESPONSE

As descr.ibed in tihe LDR, the condensation osci llation lateral load

'is simulated foreach downcomer pair by adding, a di fferent ia 1 pressure for one downcomer to t'e intern'al'ressure, tha t occurs, in both downcomer.,

thereby produ'cfog higher load in one downcomer than in t'e other,.

Thus, fr om Figure'.4'.3.4 of'he LDR,,

a darkened downcomer indicates that: thIe differential. and internal pressures are work ing s iniu 1 taneI~us ly, wherea.

the o ther downcome r only exper iences the in terna.l pressure.

A 45o beam model, Figure 6-2,,'ra0 uised to 'analyze both IBA CO and DBA CO.

Based on the primaryowncomer swing frequency extracted from a modal analysis of the system, sinusoidial

'orc ing functions were appl i ed 'o'he'owncomer pair s consider ing the load cases def iine'd by Figure 4.4 3-3 in the LDR.,Since the primary swing rood'e foII the BFN System occurs in the 8

Hz, range the 1st,

2nd, an8

!hard harmonics were addressed by sinusoidal functions in the 4, 8,

and 12 Hz

ranges, respectively.

An added c'on'servatism in the BFN analysis was the hppl ication of the first harmonic forces to the coincident

& Hz swing frequency.

The response of <<he system to this-single i'requency'oad enveloped responses from the sum of 'the three harmonics defined by the LDR.

Also, the first harmonic force.,ampl i tudes were applied with 16 and 24 Hz sinusoidal functions to verify that higher frequency responses do not in>pact tYie total CO respon'se.

In fact, 30 individual si~usoidal functions were applied for each load case to account, for poten t ia 1 harmonics a't the 1/2, 1, l-l/2.,

2, andI 2-1/2 ha rmon ics of the s ix d i sc ree t p r ima ry swing mode frequencies in the 8

to 9

Hz range.

The BFN vent system was analyze'd'foir the fou'r itn it ia1 d i fferen t ia I pressure ca,sea spec'i fi ed by a May 1981 dra'f t of LDR Section 4.4.3.

(See PUAR Figute 6-12.) Subsequently, four add'i t iona 1 mi r ror image ca'ses

>vet e in'eluded in the fical CO latera,!

load definit ion.

Evaluation of both the ir. it ia1 f'our cases and theiir four mirror images demonstrated FRC'-l PUAR.04

t tha t Load Case 1 is control.l i.ng and addi t ional r igorous analysis of the mirror images was unnecessary.

(See response to BNL Item 11 for further discuss ion. )

The DBA CO downcomer la tera 1

load ef fec ts were combined with DBA CQ fluid drag loads and other loads in the controlling load comb ina t ions and eva lua ted to the govern ing stress levels.

The chugging la tera 1

loads were calcu la ted in accordance with the LDR and NUREG 0661, us ing fr equenc ies from the modal analyses per formed'. on the 45 and 180 vent system beam models.

These loads were appl ied to s'ingle downcomer s chuggingexclus,ively and to mu 1 t ip le downcomers chugging synchronously.

The 45o beam model (PUAR Figure 6-2) was used for analysis of single downcomer chugging lateral'.

loads, and'he 180 beam model (PUAR Figure.6-3) was used for analysis of downcomer synchronous chugging la terai loads.

The r esu 1 t ing e ffec ts were then comb ined wi th other

loads, including chugging fluid drag loads on the tiebars and v-bracing for stress, and fatigue evaluation of the entire vent system.

Stresses were determined by applying appropriate intensi-fication factors at intersections and by direct application of loads to finite. element models shown on PUAR Figures 6-5, 6-7, and 6-11.

~ 1

'I

.I I

~ ~

k ~

FRC 1-2 PUAR.04

ITEM.. 2 I Provide. the Physical detafi ls of 'the sei'smic lugs. that restrain

.the. torus against hor izon tal seismic rrot ion yet

'a 1

1 ow therma 1

g row th;

RESPONSE

The er ec t ion drawing for. the seismic 'lu'gs's PDM-'E12.

,Fabrica t ion de ta i ls. foi the cohyon@n ts are shown on PIM drawing 41.

(Copies of the"draw~inps are avai lable for

.rev iew." )

FRC 2-jl'URR 04

ITEM 3:

Indicate how the ring girders were analyzed for loads from attached internal structures.

Any, dynamic load factors that may. have been used in the analysis must be provided and justified.

RESPONSE

The effects of the large'r systems on the ring girder and other portions of the torus were considered as follows:

header pool swell impact.were corisidered directly, as described in Sec'tion 5.4;2.7 of. the PUAR.

In addition, the vent system masses were included in the dynamic

'22-1/2o torus model, so tha.t the mass times acceleration

(rigid body) inertial effects were, developed for al,l dynamic loads.

'2.

ECCS Header:

The torus cradle stresses due to the react loads at the ECCS h'seder supports were added to the stress intensities i'n tha't region for the torus model, without exceeding,al lowables.

The mass of the ECCS was also included in the dynamic torus model.

Thus, the rigid 'body por.tion of the ECCS header support reactions were conservatively included twice.

3.

HPCI and RHR:

The masses of these systems were also included in the dynami'c

- torus model.

The other

'systems were judged not to a'f feet overal:1 torus

behavior, but to produce only local'ized.'f fects.

Suppor't connections to the ring. gird'er were heavi ly reinforced.

The line-of-action.of p'ipe bracing members was applied near the base of the ring.girder 'to el'iminate any signi fica'nt overturning tendency.

For example,

.see PUAR Appendix G

Plates 18 and 21.

Addi t.iona 1 ly, 'the support system for S/RV discharge lines and quenchers includes a 15-inch x 15-,inch box-beam which forms:a:continuous ring, inside the torus and

..preven,ts any possibi.'lity.of overturning

.each ring, girder in the region 'of,attachment.

The main suppor,t members for the

,catwa'lk perform a similar function at, each ring girder.

'See PUAR Plates 12, 13, 26, and 27.

Local accelerati'on response spectra for.each dynamic load were defined at, each ring girder attachment point.,'he attached piping systems and s'tructures were analyzed for these input.spectra and assoc.iated.

displacements.

FRC 3-1 PUARe04

Reactions were calculated from the piping analyses per Sections 7 and 8 of the PUAR',

and fior the catwalk per Section 9.

The local reinEorcement was designed for these r eact ion<<.

The lc>cal ized stresses transmi t ted to the r ing

'irder were. limited to 3 ksi.

Whlen combined with the general ring girder stresses, rio allowables were exceeded.

FRC 3-l?

PUAR. 0 4

Wi th respect to the 22-1/2o torus model ment ioned in Sect ion 5.4.1.1 of the PUAR (5),

the boundary condi t ions ar e based on the assump.t ion that al 1 loads are appl i ed equal ly to each of the 16 segments.

However, the saf'ety-relief valve and chugging loads are asynmetrical.

Justify the use of a 22-1/2 model to evaluate the torus for S/RV and chugging instead of the 180o model required by the criteria (1).

RESPONSE

The following, po.ints perta:in.:

1.

Synxnetric loading, for S/RV and chugging, is certainly bounding for cradle.

loads just from the point of view of the magnitude of the ne.t applied load.

2.

Shell'esponses are primarily a localized phenomenon.

They can be affec.ted by ring girder oval.ling, but this too is related to the, ne,t

load, and so would be more severe for syranetric loads.

3'.

The BFN: t.orus. support system, inhibits asymmetric response.

The development of asyrrmetric modes would t'equire 1'ongitudinal mot,ion, of the torus, which is prevented by the, se.ismi,c. lugs.

It would also require radial movement of the, ring girders which is prevented by the torus snubbers (PUAR Plate 1'),.

Friction at the cradle support pads, would, also, inhibit any tendency for asymme tr-ic response.

I f s igni ficant asymmetr ic response could develop, i t would have been present in the single-and mul't I-valve AS/RV tests, None. was egident, aqd analysis results based on. synmetrl'c.

1'oading boundary cond,itions were shown to be conservative (PUAR Appendix 'C).

Finally, if is important to recognize that Section' of the PUAAG (Refer nce (I) of the questions) is not a criteria.

It is

-a guideline for analysis methods.

FRC 4-1 PUAR. 04

ITEiE 5:

Figure 5-6 in the PtJAR (5),

whiichi depicts the 180o model of. the toius, shows only the lowe.r half of the torus shell.

lndica te whether

.the model inc Iludes the torus supp5)r ts.

II

RESPONSE

Figure '5 -6.of the PUAR depicts

.only the lowe;r. ha'1 f of the',

180o model for clarity.

The actual model includes the upper and lower halves.

Al,jl supports are included (i.e.,

the torus snubber s, seismIic

lugs, and, the suppor t'ad-t iedown sys tern).

FRC 5-1 P.VAR., 04 0

Since NRC Regula'tory 'Guide 'I.'6'I "(I4) deal's with dampi'ng values fo'r the.seismic design..of, structures, expla in how this. Regula'tory Gu.ide val'idates th'e use of 4 percent damping for the '0.'0 AP pool'well analysis of 'the torus:(S'ection

'5..4.2'.7 (5) ).

RESPONSE

The use of 'NRC Regu'I'a tory. 'Gu'i de I-.'61 damp i'ng was accep ted for ana:lys;i's, by Section 4.4.2 of NUREG 066;l.,

and the 0.'0 hP pool swel'I "case

'was designa'ted as a 'Servi'ce Level D

condi't ion by Sec:t'ion '4.3.3.1 of 'NUREG '0661.

Regulatory Guide '1.61 spec'i'f ies 4 percent damping.for 'welde'd steel

'structures, under.'SE;loadi'ng" which: is'ormal'ly associat'ed

.with Servi'ce Leve'ls C..and D condi't.ions.

'The:torus and vent

system'are

'welded steel s'tructures.

'Two perce'nt damp:ing was "conserva't.'ively used for BFN vent system '0.0. A'P 'pool -swe'I'I 'aInEa'I'ys'is.

'Fo'ur percent

.damping was used 'for the 'torus,.

Th'is "a's"sump't'ion i'n comb:i;n'a't'ion wi'th the overa I:In'a'I'ys is me'thod <produ'eed

'a r ea'sonab'ly conser vat ive dynami'c re'sponse

'.p'red'ic't i'o'n;:BNL I'tems 'I "th'rough 4 provide addi,t:i'ona:I informa.t'ion on 't'e BFN poo'I 'sw'e'll analysis

- me,thod.

Final I'y,

'i t 'is noteNoIrIthy t'h'at two per'cen't

'damp'i'ng was con'ser vat ively assumed:

fo'r.al I S'erv'i'ce Le'yel 'C I'oad

, comb:ination 'toru's a'nd. v'en't 'syem analyses, to 'reduce the

~

numb'er of c'ases:fo'r ana'lys'is;.

Fou> pere'ent damp i'ng, 'is cons'i'dere'd just i"fi'able for the.Service Lev'el

'C "a'nd D-load

~ 'combinati'ons on th'e basi's -'of Regulatory Gu 1'de I.61..

FRC 6-1 PUAR.04

Ni th respect to Section. 5.4.2.11 of the PUAR (5), provide the technical basis and just i fi'ca't ion for consider ing the forcing functions from 0

t~) 30 Hz ihstehd 'of the ful 1 0 to 50 Hz for post-,chug analysis of the torus.

RESPONSE

The intent of the discussion in. Section 5.4.2.11 of the PUAR was to emphas ize the fo1 lowing ma jor poin ts:

1.

There are signif'icant conservatisms in the BFN post-chug analysis method which offset 'th'e effects of not considering the harmonics in the 80't6 50 Hz range.

,2.

The 30 to

50. Hz harmonics dere kohsidered in th'e drag load analyses.

(See Sectidn D.1.2.4..1 of the PUAR.)

3., Pre-chug generally controls oliver post-chug for torus analysis.

The.

BFN pre-.chug analysis was performed in complete accordance with NUREG 0661 and the IDR, including additional conser'vatilsmk inherent in the method.

4.

Any remaining concern with 'te 'response of high frequency modes for torus attachments is offset by the high frequency content in the pool swell analySes.

Finally,

'a strong empirical indication (not a 'rigorous analytical proof) of the conservatism of the BIN chugging analyses (both pre-and post-chug) is seen in the attached table.

BFN shell* surface streSs

'an'd Support reactfor>

forces are presented, factored 'as nearly as possible to an FSTF-equivalent

basis, and Compared to 'measured and

~

calculated NEP values for the O'STF.

The conservat'ism of the BFN responses to post-chug 'an'alpsls results is due to the absolute sunmation of all 30 harmonics in the 0 to 30 Hz range.

(Note that the dominant BFN torus 'modes are in the 0 to 30 Hz range.)

PUAR Reference 20 recorrmends absolute suranation of 5 harmonics plus SRSS of remaining harmonics to achieve an 84 percent NEP.

FRCl 7-1 PUAR. 0 4

TRBLE FRC-7-1 COMPARISON OF FSTF AND BRONNS FERRY CHUQQINQ RESPONSES

RESPONSE

=BFN RESPONSES BFN FACTORED TO CALCULATED EQUIVALENT CHUG,.PER

.RESULTS

-FSTF REF 2P PRE-

'POST-PRE-POST-TABLE 5-1 50%

NEP FSTFs PER REF.2P TABLE 4-1 84%

NEP FSTFs

'PER REF.2P TABLE.4-1 OUE TO POST-CHUG ONLY BDC SURFACE STRESS INTENSITY 1.10,1,.12:1.31 1.33 (KSI) 0.90 0.88 0.98 INSIDE REACTION 1'02 100

-'57 (KIPS)

.56

-,31.2 17.0 19.,2 OUTSIDE REACTION 113 110 (KIPS) 62 32.3 17.7 20.2 (1)

FSTF EQUIVALENT SHELL STRESS INTENSITY = 8 BFN (R/T) 1.19 8 BFN (R/T)FSTF NHEREi R = MINOR RADIUS OF THE TORUS'

= SHELL THICKNESS'.AND 8 = STRESS INTENSITY (2)

FSTF EQU IV ~

SUPPORT REACTIONS (BFN REACTION)

FSTF POOL AREA PER COLUMN PAIR

= P.562 (BFN REACTION)

t,>g-ngr~ere~g,qg-<<",

o

-wr e-v r~

asser,

~

~

~

ITEM.82 Items

.2 and

.3 in Section 5.4.2.11 oi the PUAR:(5) suggest that the pre-chug load bounds the post-chug load in the analysis of the torus;

however, I tens S,in Sec t iion '5.4.2.11 indicates.

a higher. sur face stress fcir post-chuga Expla in this apparent inconsistency and indiicate whether pre-chug:

or, post-.clhug was considered in the control ling load combinations.for the.torus.

RESPONSE

The'iscussion of It'em 5 in Section Se4.2.11 of the PUA'R

.demonstrates-the inherent conservatisn>

ctf the BFN post-chug a~nal si's, relative to the FST'F data, The discus ion of

,I tems 2 and 3 of'.4.2.11 show that the LDR prescr ibed pre-chug analysis method bounds the actual measured FSTF chugging responses due-to both the pre-and post-chug phases.

'This is not to say tha't ipr>-chug 'wi'll always, bound post-chug.

It only sa'ys that consideration of the pre-chug phase alone, is sufficient to. demonstrate the conServat ism of

.torus results

-based on the.LDR method as compared to actual measured FSTF -results for the combined pre-and*

post-ch'ug'hases.

Al,so see the response itoi FRCiItem 7, including

-Table FRC-7-le For al 1 load comb inat ions involiviingi chu'gging, the max iniium

,s tress -due to e tither preor post'hug was used

( i. e.,

an envelope of pre-and post-chug,'re'sponses).

FRC 8-1, PUi&.0.4

ITEM 9':

Wi th respect to the fatigue analysis of the torus presented in Sec t ion 5.4.6 of the PUAR (5),

spec i fy the elas t ic i ty methods used to calculate stress intensification factors at the penet'rations.

RESPONSE

The stress

,intensification factors presented in Section 5.4.6 of the PUAR were calculated using formulas presented in the following text:

Formulas for Stress and Strain, 5th edition, by R. J.

Roark and W.

C.

Young, McGraw-Hill.

The insert,pad to shell junction intensifi'cation factor was based on Case 14, page 598.

For the insert pad to nozzle

junction, Case 5,

page 593 was used.

FRC 9-1 PUAR.04

ITEM 10:

Provide and just i fy the bound ing technique used to det'erkike

'he contro'll'ing load cases presented in the PUAR (5) i'n 'the fo 1 lowing sec t ions:

5.5.1, page 5-21 6.3.2 (and Table 6-5),

page 6-6 6.4.2 (and Table 6-7),

page 6-7 6.5.2 (and Table 6-9),

page 6-8 6;7.2 (and Table 6'-12),

page 6>>12 6.8.2 (and Tables 6-15 and 6-16),. page 6-14 6..9.2

'(and Table,s 6-17, 6-18, and 6-19),

page 6-.15 7.3.1 (and Table ?-I),

page 7-7 7.4.1 (and Table,s 7-2 to 7-4),

page

'7-12 8.2'.22 (and Table 8-2),

page 8-3 9'.1, page 9-2 9.2, page 9-2 RESPONSE.':

PUAR Section 5.5el No bounding techniques were used for determining controlling loa'd cases for the torus analy'si's.i All 1'oad 0:ombination.

were analyzed, including the calculation of stress intensit ies and reaction loads for the cradle and torus snubber.'.

The referenced section was stating which of the comb ina t ions produced the highest,st ress in tens i t ies.

PUAR Sectiooe 12.2.2 theo~9 0 0.9.2 PUAR Tab les 6-5, 6-.7, 6-9, 6-12 2

6" 13, 6-17, 6-18, and 6-19 give the event,,

combinati<>n, and syrvice level of the v'arious locations'f inspI.etio'n.',

gatile', 3-1, of the PUAR shows th.is information in a diffkrknt form.

The bounding'echnique for the table,s in PUAR Section 6 was such that Service Level C event combinations would be qual i fied i)sling'ervice Level B allowables when possible.

When this was not

possible, actual service level combinations coincided With the assigned service level str'ess Ijimits in PUAR Table '3 'l.

A logic desc'ription of the bounding justification for each table follows:

FRC 10-1 PURR.'04

PUAR Table 6-5 Lo ic 0

0 0

Load Comb ina t ion (LC),

15 to,,Service Level (SL')

B allowab,les envelops LC 1

th.rough LC 14.

LC 27 to SL B al:lowables envelops LC 17, LC 20, L'C 21, LC 23, and LC26.

LC 25 to SL B al lowables envelops LC 16, LC 18, LC 19, LC 22, and LC24.

o IBA is not indicated in LC 15 because:

('1)

SBA chugging.

is, no less severe th'an IBA chugging.

(2)

IBA'O is enveloped by. DBA CO in LC 27.

PUAR Table 6-7 Lo ic o

LC 18 to SL B allowables envelops LC 16.

0 0

For. the vacuum breaker to, main vent cap inter-

section, dynamic loading due to, pool swell vent response far exceeds any chugging, CO, or S/RV effect.

Ther,efore, evaluati,on of LC 1

through LC 15 plus LC 17',

LCs 20 through 23, and LCs 26 through 27 i's not necessary.

LC 18 does not envelop

.LC 19, LC 24, and LC 25.

However, 'the load contribution from-SSE versus OBE and the S/RV contribution are small and have been.

neglected consider.i'ng the increased allowabl:e for SL C r.e 1 a t i v e 'to SL B.

Note that the pool swell, load considers pool swell vent response and direct impact, of the swell on the vacuum breaker shell.

(The vacuum. breaker shell is actually.partially shielded by the vacuum breaker access plat.form.)

PUAR 'Table 6-9 Lo ic 0

This logic is similar to PUAR Table 65 except LC 27 could not be,sa tis'f'ied for SL. B pr imary plus secondary stresses (P

+ Q~3 Smc).

Ther efor e, LC 21 was evaluated for pr lunar'y plus, secondary instead.

(LC 21 is the same as LC 27 wi th no S/RV.

PUAR Table 6-9 incorrectly indicates LC 27 instead o'f LC 21.

This correction will be: made in the PUAR revision 2.)

FRC 10-2 PUAR. 04

PUAR Table 6-12 Lo~fc o

Again this logic is simi lar to PUAR Table 6-5 except LC 27 would not meet SL B al lowables.

Therefore, LC 27 was evaluated to SL C al lowables and LC 21 was evaluated to Sl. 9 al lowables for both pr imary

'and,'pr irnary plus secondary, stresses.

(PUAR Table 6-12 incor r ectly indicates LC 27 instead of LC 2'1 f'r the Serv ice Level 8 comb ina t ions.

'Th is correc t ion wi 1 1 be made.

in PUAR revis ion 2. )

PUAR Tables 6-15 and 6-16 L~oic o

Again this logic is simi lair to PUAR Table 6 5.

The noted loads are the wors t poss ib le comb ina t ion o'

any accident condit ion, including thermal effects, and the buck 1 i ng evaluation is per formed on

ttI, is'as iIs.

'PUAR Tables 6-.17~ 6-18~and 6-19

~Lo i c o

Again, LC 15,, IC 25, and IC 27 are the worst case'oad combinations for the systemThe indicated s ti esses ar e maximum sur fa.ce inc Iluding secondary effects wi th s igni ficant margin against 1.5 SnIc.'heref'ore, the primary plus secondary stress range evaluation i. aut'omatically assured.

PUAR Sectilon.7.3.1 Section 7,.3.1 provides a general d'escription of the

'bound'ing technique that was used to determine controlling load cases for S/RV piping in the drywell at Browns Ferriy.

Specifically, the controlling load cases for drywell S'/RV piping were determined by the following process:

(1)

Survey'll defined normal,

seismic, and LOCA ;oad de f in it ions to determine which oi'hese have sign~i f~ichn t e f fec t,on drywe 1 1 S/RV piip ing.

No te that separate modela Cf 'th'e /itin'g'ystems w'erie deve'loped to analyze the drywe1 1 and we twel 1 por t ioins of the system.,

The we twe 1 1 models were devel, oped in s igni fi can t de ta i 1 to s tudy torus hydrodynamic phenomenon closely.

These models extended a

s igni fican t dis tance into ma in vent to account for at ten at ion.

it was found that the S/RV piping in the main vent is isola ted from most of the hydrodynamic ef fee ts of S/RV di: charge or

.IAX'A exci tat io'n 'of'he suppression pool.

(An exception to this is the containment vent response FRC 10" 3 PUAR.04

I~

induced by DBA LOCA,pool swell. ),

The drywel 1 piping is su'fficiently removed.

from the suppression pool to discount ef fec ts 'from wa te'r clear ing trans fents in wetwel 1 port ions of the S/RV 1 ines.

V The load sources that were determined to have a

s igni f icant effect on drywell S/RV piping are:

a.

Deadweight b.

Seismic

- OBE and SSE c.

Al 1 S/RV blowdowns d.

Pool swel I v.ent response e.

Therma.l expans'ion f.

Pressure (2)

Per:form an i'nspec t i'on of Table '5-2 in the PUAAG (PUAR Re fer ence 13')

cons ider i ng the resu1 tant S/RV load sources noted in step 1 above.

A sur'rxnary of. the 'fi'nd'ings with! respect to PUAR Tab,le '7-1 f'o 1 1ows.

Case 1:,Sat is,f~i'es L'C l.

Case

-2:

Sat is fies LC 3 wh'ich envelops LC 2.,

Case 3:

Satisfies,LC 15 whi'ch 'envelops L'C'.4 through LC 14 except a's indi'cated by 'note

.6 on'PUAR Table 7-1..

Case 4:

Sails f i'es LC27 which 'envelops LC 16,;through L'C 26.

PUAR Section.

7,'.,4.;1 Section

?.4.1 pro'vides a general discussion o'f the bounding technique for wetwel1 S/RV piping.

Al I load sources, are treated.

A suianaiy of the envelopirig logic foi PUAR Tables, 7-2, 7-3, arid 7-'4 with resp'ect to Tab'le 5.-2 of the PUAAG (PUAR Reference I i) is 1 is ted below.

PUAR Table..7-2 Case l,and Case. 2:

Sat is fies LC l.

Case 3 arid Cise 4:

Satisfies LC 3 w'hich envel'ops LC 2.

FRC 10-4 PUAR.04

PUAR" Ta b I e 7-3 Case I ahd Case 2

'Sati'sf i-e 'C: I'I, r.ow I'I,, which envelops LC' through LC '10 for row 11.

Ca se 3 and Ca se 4.",

Sa ti,s fies LC 15, row ll,,which.'nvelops

'LC 12 through LC 14 for-row 1-1.

Case

'5. and Ca'se 6,:

Sa'tis f iies LC 15, row'0, whi'ch

'nvelops LC 4

tl~ir'atugh LC 14 for rlaw

,10.

PUAR"Table T.-I Case, I:

Sat'is fles

.LC 27 which, envelops LC 17, LC 20,,

LC.2,1, LC 23.,

and LC 26.

(Note thai'CO,'nld chuggir<g do not occur si'inul taneously and

.chuggirig, Is a'ddressed

@pre conservat ively in

'PUM Table 7'-3 for. SBA/IBA events e )

Case 2!

Sa'ti: fies LC 25 whic'h 'envelop's LC 16, LC 18, LC,;19, LC'2,. and LC 24.

.Case':

Satisfies LC 16 for ',the Ct.0 AP case.

(Due to'he love Iprobab I:1 i ty of,occurrence,,

S/RV blowdown.

and earthquake

'are not, assumed to be conc'urrent'ith'th'.

0.0'.ZiP pool swell.

This is in accord'ance wi,th. the:PUAAG.)

Q PUAR Sect.ion:Se2.2.2 The.boundary:yt tec'hnique used to. reduce,.the number of 'lokd

'case combina,tions shown:, in PUA'AG; Tabl'e

.'5-'2.

itPUAR Reference,

13) to those

'shown. in PUAR'abile. 8-'2 was based on using the most-conserva'tive combination of 1Oad cases associated, with, each of the serv-i'ce levels and ASME Sec'tion II'I,,NC-3600

'PUAR Reference

68) equation 9 stress limits.

(Note that fo,r different.local. combin'ations,',

thetii.ss limits could

1.2 S,.1.8 'S, or 2.4 S,

carr,esponding to Set v'i'ce Levela B,

C, and D, respec:tively.,)

The following are two examples o~f how this bounding techniiquae w'as applied:

When two ser,ies of -load combinatiio'ns li's:ted in the PUAAG were, the same except tha't

'one Included OBE a~nd the other, SSE and both sets of'oad combinations had the same s,tres,s.limits, the OBE,and SSE loa'd cases were envel'oped.

In this way the two sets of lo@d

'ombinations could thus be reduced to one.

",FRC 10- 5 PUAR. 04'

(2)

Another example would 'be when one set of load comb inat ions cons is ted 'of a 1 1

the load cases found in

,another set of load combinations plus at l,east one more load case.

If both sets of load'ombinations had to meet the same stress limits then only the combination with the grea'ter number of load cases was evaluated.

The controlling load combinations for each seivice level equation 9 stress li'mi:t were 'found in.th'is way.

In addi-tion,

,.NC-3600 equation 1'0 'or ll was satis.fied for each of the cont'rol 1 i'ng load combinat'ion's.

PUAR Section 9.1 The.new catwalk finite element model was analyzed for all, applicable

.load events.

The results showed that the largest

stresses,

'by far, were due to pool swell impact and drag.

Therefore, the most severe condition invol.ving 'th,is load was

limit;ing.

For.Load Combina'ti'on 25, the effects of pool swe.ll impact-drag, p'ool swel"1'nd vent header mot,ions, S/RV moti'ons, deadwe.ight.,

and 'SSE were add'ed absolute,ly.

PUAR Sect'ion 9.2 In.the same, manner as

.the catwalk, the vacuum breaker valve pla:tform is most severely affected by pool swell impact-drag loads,,

and the 'limi'ting 'load combination was determined

'in

.the same way.

FRC 10-6 PUAR.04

ITEM 11:

Pt'ovide "the stress resul ts from the analys,is of the torus shell and supports.

. I

RESPONSE

Stress intensit ies wiere calcu~la'ted by postprocessor computer codes for ail load comt) inatio'ns'nd fo'r al II elements in the finite element model.

The results were screened to locate predic ted overs tresses

Fo1 lowinj;. mod i f 'i ca t ions, al 1 stresses were below allowable's.

'he most highly stressed torus support locations are in 'th'e

cradle, adjacent to the scab plateis described in Section 5.2.4.3 of the PUAR.

The. addi t ion of the',scab pla tes

'educed the laical cradle stresses.

for Load Combination 14 from 24.4 ksi to 19.7 ksi.

Rielative to the Service Level p

allowable of 21.6 ks i (see Sei;.tijou 5,.3.2 oi'he PUAR),'he' tress factor was reduced'.froIn lf. 13,to.0.91.

The maxImuIn stress for any Service Level C or D load case was in the'ame region of the craclle and'z,, clar:i fy whether, the seismic and', thermal'esponse o f the drywe.l 1 was. cons'idered (8'ei". t lions; 6,.2. 1;.2.7.and

.6.2.i;.'.9 (S.)),

RESPONSE

Thermal'rowth o'f the-conta inment shell'was considered jn the anailys is of'he'a in vent/dry'we 1 1 in tersec,t ion.

Therrre 1 displacenIients were. ca'lcultated for 'the,drywall based uiIok maximum air.tI.mj>era tu'rIes: occurr ing,clur ing the KIBA, IBA, and

,.SBA events..

These displacement

s. >vere input ht the nodes re'presenting

'the drywel 1/main vent."intersect ion.

The'.t?Iierma'1'.

analysis was

'thin completed'onsider ing expansion':and restra int;of 'free end displacement cIf'he vent.system..

~

The, BFN LTP.seismic ana'lysis was based upon the'ethods.

~.employed in, the or iginal-plan t des ign.

'Se ismic.

r esponse.

o'f the'rywel 1/main, vent intersection wa,s analyzed using equivalent s'ta t ic loads de term'ineA from appropr iate accelera t ioix 'levels of. the ven't 'y't'm.

This is.cons i' tent w.i th the gene'ra 1 gu,ideIIines of-NUEtEG 0661, Section 4.,4.1 as.

wel,l as the

-P1IAAG (Pl'JAR Reference 13).

FRC 12 '-'1 PUAR,04 0

ITEM I 3:

Prov ide a

sunma ry of the ana 1'ys i's of the

.vacuum breaker valves;,

i nd ica te whe ther they are cons idered CI ass 2'omponents as required by, 'the cri ter ia (1).

RESPONSE

Appar ent ly this request we twe 11 vacuum b r eak.er s chugging events.

Since I Containment Long-Term sent to the NRC.

relates to analysis of the drywel1/

for cyclic loads occuring during this concern is not part of the Mark

program, a separa te response will be Ii'h is request rela tes to th'e new 10-inch S/RV vacuum

breakers, these valves have been analyzed and subsequently mod ified. to sat is fy ASME Section III Class 2 stress limits for all postulated condit ions including opening impacts.

The mod i fied S/RV vacuum breakers are shown on TVA drawing 47W401-9.

ADDENDUM

'I Addit ional information on quali fication test ing of the mod i fied S/RV vacuum breakers for opening impact loads was

requested during the September 13, 1984 meeting with NRC and FRC representa t ives.

That in forma t ion follows:

Pr eliminary forcing functions for design. of opening impact modifications were based on conservative predictions extrapola ted from Mont icello test results.

Mod i fication designs were made.

and preliminary tests for short-term, adequacy were conducted on that basis,.

The final forcing function for the S/RV line E vacuum breaker was determined from discharge line pressure measurements taken dur ing the Apr i 1 1983 BFN S/RV tests (PUAR Reference 41).

This forcing function was analytically extrapolated for all'FN S/RV lines and all long-term program load conditions.

Then a,prototype vacuum breaker was tested at Wyle Labora tory, Huntsvi lie, Alabama, to demonst ra te operab i 1 i ty for al 1 fore ing func t ions and the full 40-year plant life.

FRC 13-1

.PUAR.04 tMW0

~

~

I

~

~ t ff I

tV' 0

~

h t

'%W

~

'eA

ITEM 14:

The PUAR.(5) indicates that t'e chloulated stress>>values at the following locations are very close to.the respective a I 1 owab I e s:

0 o

dowi>comer/vent header inter sec t ion (Sect ion 6.5.4.1) o downcomer/t iebar in tersec t~ io~n (Sec t io'n 6.7.4.1)

Indica te. conservat isms in the a~nakys is to show tha't the. e

'calculated values would not be'xceeded if a diffe'rent analy t iica 1 approach were to be

~used>>

RES PON'iE:

Although the, stress values at the intersections mentioned above were close to the respect~ive all'ow'able stresses, this should not represe nt a significant concern.

Design modifications were made such that the stresses resulting from the new configurations wer~e

~just below the acceptable values.,

This would normal ly be ant iic ipa ted.

There are conservatisms which could be removed to obtain a

greater di fference in the allowable and actual calculated st-ress values for the above i'nt'er.,'sections.

For example::

(])

Absolute surmation was used in the. combination of loads.

(2)

The downcorner /t iebar was c~onser vat ively analyzed. using Service. Level B'llowables for Service Level C loads.

(SSE seismic loads were in'eluded in'he actual loading in place, of the -prescribed OBE seismic I"oad.)

(3)

There, are other.

conservati'sms associated with the DBA CO lateral load analysis method as described in 'the response to PRC I tern PRC 14-1 PUAR.04

ITEM 1'5:

Stress intensi f ication factors 'for the mi t'e'r bends in the vent system are not found in Table 6-17 as s ta ted in" Sect ion

.6.9'.

1 of the PUAR (5).

'Pr ovide these factors.

RES P.ONSE:

Main vent mitre bend SIF

- '3.85 Vent.header m'i:tr e bend, 'SIF 8.2 Downcomer mitre bend SIF

.- 3:.82 (An appropr'i'ate revision,wiI'I made to the PUAR:)

FRC 15-1 PUAR.04

ITEM 1 6:

Regard 1 ng the torus bel lows ana'lys is in Sec t ion 6. 10. l. 1 of the PlJAII (5'),

pr<>v i de the method and technica 1'as is 'for calcu lat inj,, the spr;-ing v'alues,tha t represent the bel I'ops flex ibi lii ty in the con~uter. models'f the vent system

'F.igur.es 6-2 and 6-:3 (5)).

RESPON6E:.i The spr ing values that:repr'esent the bellows flexibility were calculated using the "Standards of the Expansion Joint

- Manufacturers Associiation, Inc.,"

(PUAR Reference 24).

Appropr ligate data. for the BFN bellows was input including convolu t ion dep th, thickness, number of'onvolu t ions, and modu lus of elas t ic i ty.

FRC 16-1 PUAR. 0 4 0

~

~

ITEM I 7:

Provide and justify the a'pproach for the fa'tigue evaluation

.o f the be 1 1ows men t i oned in: Sec t i on 6. 10;.3 o f the PUAR (5 ).

~ '1 The fatigue evaluation was carr ied out using "Standards of the Expansion Joint Manufacturers Association, Inc."

(PUAR Reference 24) and the Mark:I. Containment

'Pro ram Au ented Class 2/3 Fati ue Evaluation Method and.Resul ts or T ical Torus Attached and S V,i sn S stems, PUAR e erence 21 Deflect.ions for the torus and bellows were obtained for each load event using computer analysis resul'ts and hand calcu-

,la t ions.

These de flee t ions 'resu I ted in bel lows stresses which were then combined in ac'cordance with the fatigue eva 1 u a t ion me thod no t e d abov'e.

FRC 17-1 PUAR.'0 4 El(h ( Phh'\\

~

( lff E ~ i

'lg I (

I' (h

i ~

(( h((

I)( i ~

) ~

P

~

"1

ITEM. 18:

Accord iing'to Sec't i ori 7,.3.3 n I of the PUAR',(5) the sa f e ty-

,rel i ef val'v.e I i i>e p'enetrat I'oti,o.f 'th&, re i'n "v,ent was modeled us,ing cy 1 1 ndr'le~el shell'1 exib i 1'i 'ty 'harac ter,i s t ics.

Indicate, the ice thod for'e'termi'ni'ing 'h'es'e character is t ics. I

RESPONSE

For'he S/RV-'ine penetratfion of 'the main 'vent, a six degree of,fr eedom "suppor t" 'was: modele'd.

For three degrees of freed'on> (pipe torsion arid th'e two transla't iona'I she'ar'dir'ection's) ful ji fi'xity was ass'umisd..For the ci r cumferent ia1 and ion'gi tudi'nal bending directions,'i j laard's methods (PUAR Reference

64) were utili ~ed to de'ter~mine rotat,ional spr iing rates.

F'r the main v'ent'adial direction, a transla tional

.spi ing r'ate was'etermi'ned.

pdr th4 R. J.

Roark text, IFormuias for Stress,and Strain.

.FRC 18-1 PUAR.04

ITEM I 9:

Prov ide the technical bas'is: for ob ta in ing. the s tress intensi'f'ication. fact'ors u'seda in the anal'ys'is "of the saf'ety-re'1 1'.ef val.ve dischar.ge pip'ing sys,tern

('S'ect.ions 7.'3.3.1 and 7.4.3.1 (5)).

RES PONSE':

In genera:1 the ND-36'V3. 2 ( b) -1 )

I n tens i.fi ca t ion p ip i'ng, ana lys'is 1'97V ASME'od'e S'ect,ion I I'I'(Figure is. the technical basis for "the stress factors. uti;lized i'n the S/RY'ischarge wi th the fol'lowing excep t ions:

Bas.I s:

~ ~

'1 Weld-0-Let Bonny Forge s,tress

'in.tensi'f ication factors and'. stress.

i'ndices for we 1 d'-o -1 e t s..

Sweep-0=Le t Quencher Near Collar'upport Stress intensi fication.factors and stress indices for. the Bonny Forge sweep-o-1 e ts.

Stress. intensification factor based, on eff'ective secti'on of quencher.

Assume's hole zone of quencher provides no structural contribution.

i =,Section. modulus of 12 inch Sch 80 Pi e

Ef fective section modulus of quencher I

FRC 19-1 PUAR.04

• ~

~

r r

I err

~re r r ~

~

• , r ss+v sr

~ '

. ~ rr ~ ~

rt

~

~

iTEM 20 Provide, the stress,resul't s

from the 'wetwel 1 and drywell 1 sa fety-re,.l i ef'.val,,v,e.dischar ge p.ip ing analys'is

'(Sect ions 7,'.3.4..l. ari'd. '7.'4<..4'..1;,(,5)-).

RESPONSE

I Tables., FRC-.20-,1 through,:FRC-20-4 provide a

suranary. of the

'aaximum equ~iti'on, 9 st'resses from'he S/R'V discha'rge piping

-analys'is.

FRC 20 1

PUAR;. 0 4

TRBLE-FRC-20-i ORYHELL LOAO CQMBINATIGNS, MAXIMUM STRESS LOAO CASE '( 1')

I NOOE

'STRESS'KSIr)

'STRESS RATIO

-LINE C '(2)

CASE 1

CASE,2 CASE 3

. CASE 4 59 59

'47

.17,.

8'0.5

.22. 1

.20.6 0.991 0.760 0.61'3 0.573 LINE E (2)

CASE 1

CASE '2 CASE 3'CASE 4

157

15'57

157

,1'6,.2 23'.'8

,23,. 8

,,23.. 9

0.. 898

,0.883 0.'662 0.664

,1..LOAO CASE.PER TASL'E'-1 OF.PUAR.

,2.. LINE C IS A REPRESENTATIVE SHORT.LINE.

LINE E, IS A REPRESENTATIVE, LONG L'INE.

'CALCULATED STRESS ALI OHABLE STRESS

T,Rf3t.E, FF/C;-l20-.2 NOC SERVjiCE II EVEL B Atlo C.LOAD COMBINRT,IOhlS MAXIt tut<<

.;TIRESSES.

. l~E.'rWELL.EVRL.URT:ION 4

LOAD CASIE'(1)

NODE STRESS (KSI

) 'TRESS'(4' fkRl IOl LINE (3)

CASE 1

,CASE. 2 CASE',

CASE 4 10,.3

~

0.573 ENIVELOPED BY',CASE 10.9 0.404'ENIVELOPED BY CRSE '2, (2)

'LINE H (3),

CASE 1

CASE 2 CASE':3 CASE 4

'8 13'.3 ENVEIL".OPED BY -CRSE 4 (

0 '758 1:4. 1 0.'522.

ENVEIL'Oi'ED BY CASE 2 (2) 0 1'.

LOAD 'CASES PER TABLE '7'2 OFPURR',.

2.

THE,ENIVELOPING 'OCCURS BECRUSE R WORST'.CRSE

'NOC BLOWDOWN (SCREENED BENI WEEN'IRST -AND 'SECONDACTUATION)

IS'.USE(I)

NOR'HE STRESS ANAL.YSIS.,

3'.

L'INE L TYPICPIL SHORT 'LINE'.

LINE Hl -TYPICPIL LONG LINE.

4; STRESS RATIO CALC'ULATED STRESS ALLOWABLE STRESS

TRBLE:FRC-.20=.3 SBA/'IBA -- SERVICE 'LEVELS;C.AND 0 LOAD.COMBINATIGNS

,MAXIMUM STRESSES "HETNEL'L EVALUATION LOAD CASE.('1)

'NODE STRESS

',(KSI')

STRESS'RATIO CASE

.1 CASE 2 12

.2 21.'6 24.9 0,. 887 0.922

LINE,

.L CASE.3 CASE 4,

ENVELOPED BY CASE 5

,ENVELOPED'Y 'CASE 6 CASE 5

'CASE 6 12 2

.21.8 26.8 0.674 0.745 L'INE CASE 1

'CASE 2

.CASE.3 CASE 4

'CASE 5

CASE 6 8

10 26.5 26.0 ENVELOPED BY CASE '5 ENVELOPED 'BY. CASE 6 j

31.7 32'..2 0.982 0.963 0.881 0.894

1. 'LOAD CASES 'PER TABLE 7-3 OF, PUAR.

2.

SBA AND'RA BL'OWDOWNS SCREENED TO PROVIDE WORST CASE FOR'NALYSIS; CALCULATED'TRESS

',ALLOWABL'E STRESS

'I PiBL)=- FRC--20.->>

OBPi

- SERVICE.:LEVEL, D..LOAD-COMBINATIONS MAXIMl)M STRESSES

'NETNELL',E'.VALUA'T IONI NE L

'"LORD CASE Ei'i

,CASE 1

CASE 2'ASE.:3 NODE 235

'L STRES!~.

IIKSI,)

22.4 33 4 31.7 STRESS E>>,

'RATIO 0 '692 0'.

927'.880 LINE CASE 1

CASE 2 C,RSE 3 28.1

'28'.,'3 27.4,

.0.'7'81 0

.'787'.760 0

1.

LOAD, CASIES 'PER TABL'E 7-'4. OF

'PURR'ALCULATED

STRESS,

.AL'L'CIHABLE STRESS

ITEM 21:

Prov ide and just i fy the. al 1'ow'able safe ty-rel i e f valve nozzle loads which were.referred t'o in: Section 7.3.4.2 of the PUAR (5) ~

RESPONSE

As surrmarized'n the PUAR, relief valve nozzle loads calculated in. the d.rywell S/RV. piping ana'lysis were compared to a set of allowable nozzle loads used in the original S/RV analysis performed by Teledyne Engineering, Services and documented by PUAR Reference 45..

These allowable loads were provided by the reli.ef valve. vendor, Target Rock, and are incorporated'n the design report for. this component.

The alldwables and worst case calculated loads, are:

Va 1 v,e

~Flan e

Inlet Ou t'1 e.t.

Worst Load*

3 209 6,1 IN=,-LB.

287;,5'68:

IN.=LB Allowable Resultant Bending Moment from D~naml o Loads 400,000 IN-LB 300,000: IN<

~

Va 1 ve Flange Cond i t ion Wors t Load

(,IN-LB)'l 1 owah 1 e

( IN-L6)

In let Ou t le t Inlet Outlet Inlet Outlet S

'S S

+ D (B).

S,+ D'(B)

S

+ D (CD)>>

S

+ D (C,,D)>>

296843 306205 390466 353947 637582 515009 437321 372971 874642',

745942 1375829 1096778

<<Direct add.i't i,on of dynamic load components.

As can be seen, all loads are acceptable.

Condi,t ion notes:

D (B)

(C)

(D)

Static Dynamic Service Level 'B Service Level C

Service level. D FRC 21-2 PUAE.04

ITEM 2 2:.

Wi th respec t to Sec t ion 7.4.3'..2.

1 of the PUAR (5), prov ide and just i fy all dynamic ampli.f ication. factors used in 'the calculation of'afety-relief valve discharge-induced fluid drag., forces on the sa fe ty-r el'i e f valve sys tern.

RESPONSE

Safety,-relief valve (S/RV) discharge-induced fluid drag forces were applied pseudo-sta ti.cally to the S/RV system.

The TQFORBF computer cod'e was used with S'/RV line input properties that would produce the highest force amplitudes possible from any line for any S/RV discharge case.

Maximum amplitude force-time histories were thus determined.

From these time histories, the peak amplitudes of the distr'ibuted forces were taken for equivalent static application to the system.

The equivalent pseudo-s ta t ic force dis tr ibut ion was determined by conservatively assuming the distribution of peak forces to act as a perfectly steady-state sinusoidal fore ing func t ion.

A modal analys is. of the S/RV sys tern indicated that there were no potentially responsive modes of the system within the broadened; 4.2-14.7 Hz ran'ge (see PUAR Section 5.5.2) of possible S/RV forcing frequencies.

Therefore, the lowest potentially responsive natural frequency of the system was assumed to be driven by the highest possible broadened forcing frequency of each load case considered.

This pulls the two frequenc ies

( i. e.,

the forcing frequency and the system frequency) as close together as they can ever -possibly be, thereby conserva-tively maximizing the dynamic load factor,(DLF).

The DLF was computed for each S/RV load case considered using the following expression for a harmonic forcing function (see PUAR Equation D.1.2-24):

1 (i-Qglg)'g

+

4 iInQ/~) g where damping ratio

(-2% was use

)

Q= forcing function frequency system natural frequency FRC 22-1 PUARe04 e ~

r ~: 'e'"

~ ~ere v" '+r'n" ~ ma'"'""'

e

• r

~

The load cases considered and cdrresponding DLFs we'r'e

'as'ollows:

Ca s'e No.

Descr:iot ion NOC,'DBA-1st Actuation NOC,DBA-2nd Actuation SE1A,IBA-1st Actuation SE1A, IBA-'2nd Ac tua t i on (H.z )

8.34 10.2~J 12.3 1!

14.6~J 18.25 18.25 18.25 18.25 DLF'..26

1. 4i6 1'. 8 4 2.83 Finally BFN S/R'V test results showed significant conserva,tism

~ of S/RV discharge l.in,e and suppo'rt" st'resses relative to analytically predicted values.'ee PUAR 'Se'ctions C.'7.1 and C.7.2.

0 FRC 22-2 PUAR.04

ITEM 23:

,Wi th respec t to,Sec t io'n 8.2.2.3 o f the PUAR (5)', prov ide and j us t'i fy the r.eason's for not cons'1'der ing the contr ibu t ions of h igher modes above 2'0 Hz for 'seismic analysis

'of torus attached piping systems.

'RESPONSE:

Seismic.analysis of to'rus-attached pip"ing systems was performed using the origi'nal analysis methodology as permi.t ted by. Section 4.4. 1 of NUREG 0661.

Original seismic piping analysis methodology of BFN documented in '-FSAR

.Appendix C.3.'2.l..a, includes use oi'0 Hz as the "cut-off" frequency.

FRC 23-1 PUAR.04

IUI th respect to.Section 8.2,.5.2 of the PUAR- (5), provide justification for considering'taAch lines having peak spectral accelerations below 5.0 g at the point of attachment to the pr ocess 1 in'e to 'be quali f ied without fu r ther eva lua.t.ion.

i

RESPONSE

TVA's cri ter ia for excluding br~'t'ncih lines from addi t ional analysis may have been mis interpreted.

The exclusion limit is.not the acceleration input to the branch line from the process 1 inc.

I t is the ampli fl'ed mot ion of the process line, i.e, the exc1usion limit is based on the dynamic reponse spectra for the branch line.

The 5-g limit was or iginally selected based on TYA'0 experience ivith seismic quali.f ication of small lines with typical BFN con figu ra t i ons.

Exper ience with BFN LTP analysis of branch liines which exceeded the 5-g limit prov ide d fu rther ver,i f i ca t ion of: the accep tab i 1 i ty of 'th~is imi t for BFN branch lines.

ADDENDUM 0

Dur ing the Septeimb er 13, 1984 meeting, FRC r epresenta t iv'es'ndicated some concern with this response and during a'elecon on September 24, 1984,, addit ional information iwa's requested.

Tha t in forma t ion fol lows:

All branch lines which connect to 'thie tot'us'ttached piping process lines were evaluated for dynamic response of the'ranch

line, thermal and dynan1ic dispiac~ment of the a t tached pr oc ess 1 ii ne, and su. ta ined loads (deadwe igh t and pressure)

Re ferr ing to paragraph NC3650 of the 1977 ASIDE Section II I Code (PUAR Reference 23),

branch line dynamir'.

response s tresses plus susta ined load s tresses are included in code. equation 9 whereas bran'ch link s'tresses clue to process line displacements and susta ined loads are included in equa t ion 11.

The 5-g limit for branch line dynamic response analysis was spec i f i ed on the basis of exper ience's descr ibed above.

The adequacy of thai t 1imi t andi the fac t tha t equa t ion 11 FRC 24" 1 PUAR.04

s t'resses are typ ical ly more cr i t ical for BFN branch lines than equa t ion 9

s t'resses is shown by Tab le FRC-24-1.

Tha t tabula t ion giv'es resu 1 ts for ten BFN Un i t 3 branch 1 ines which had response spectra exceeding.

the 5-g 1imi t.

Stresses are presented as a percentage of the code equation 9 and 11 a-l.lowables.

The peak of response spectra accelera.tions-and branch line identi;fiers are also given.

Recognizing that the equation ll stresses.

were more critical, all branch lines were analyzed for those condit ions.

It was not necessary or cost effective to r igorously analyze all branch l,ines for equat ion 9

s,tresses--hence the 5-g limit.

Finally, the small compact valves which are located in BFN

'branch lines are structurally adequate for accelerations much greater than 5-g's.

Therefore, the 5-g 1imi t is also appropriate from a component operabi'll-ty standpoint.

FRC 24-2 PUAR.04

TAELE FRC:-.2 i-1 TYPICAL HRPSCH LINE ARQ YSIS K~TS Process Li,'ne Penetration X212 Process 'Line Node Point

.37 30 40 46 107

~F~i~~~l ovia~)e Str~eq Ettuat ton 11 Eguation

'9

~ 3

~3

.9.

.1

.1

..1

.1

.1

.4

.3, Peak Spec tra1 Acceieration,a'

'9.5 6.1

'9.3 6,3 9.9 X214 X223B X231 55 E30 55

- 75X 75K

~2

,.6 n 2

,;5

~ 4.I 9.5 10.8 19 a9

. 20'. I

20. 1 FHC 24-3 HJAR.'04

ITEi>I 25:

Wi th respec t to Sect ion 8.2.5.5 of the PUAR (5), provide just i;fication for considering the valves wi.th accelera t ions less than 3-g horizontal and,2-g vertical and -having no operator supports to be quali'fied without further eva lua t,ion.

RESPONSE

The 3-. g horizonta'1 and 2-,g vertica'1 accelera ti'on limits on valve acceler.a tions.are justi'fled by our experience with

,seismic qua.li f.ication of similar.'va.ives on four TVA nuclear plants (Browns Ferry.,

Sequoyah, Wa,tts iBar, and ~Bellefonte).

ln addition,

.none of.the:numerous valves which were evaluated for the BFN LTP,,had any problem with sa tis'fying

the requiremen~ts of PUAR Section 4..3..3 with applied accelera,tions, in excess of the 3-g/2-g limits.

FRC.25-1 PUAR. 04

ITEM 26:

,Prov ide a schedu le for the corrlp 1 4 t ion o f p i joe suppor't

mod i f ications. for Units

'2 and 3.

,RESPONSE:

The BFN Unit

'1 and Unit 3 Ipipe sup'port modifications are comp le te, and al 1 Un i t 2 in terna i pipe support mod i f ica t ions a re comp le t,e.

The.date for completion of Unit

-2 external pipe supper t'odif ications is currently (September'984) under negotia,tion with NRC.

An integrated modif ication. schedule was submit ted iin Augu,st 1984'or NRC review and approval, indicating completio'n of Unit 2 -external pipe support

=

mod i f i ca t iIon s du r ing the Cycle 6 re fue 1 i ng outage.

FRC 26-1 PUA]R. 0 4

ATTACHMENT 2 BFN PUAR FINAL TVA RESPONSES TO NRC AND BROOKHAVEN NATIONAL 'LABORATORY. QUESTIONS tl'r r

~rrpl

~

il~

il~

ll

General R~es ense, to PURR nest l'ons.

'.BFN L'TP anal;ys.i,s and des.ign activi.ty has proceeded on a

sc'hedul:e necessary to support,instal::lation of all

.modif'i,ca.tions dur.:1'ng, the, cyc'le

'4 and 5 refueling ou,tages o'f each uni't,.as requ'ired by.NRC.

The f:i;rst BFN cycle 4

refuel,ing outage began in 'Apr'itl 1981,.a'nd'ost of the major modi:f'i'cati'on. desi'gns were compl'ete

'by" May 1981.

Remaining modification designs,. pr.imarily.i'or torus attached pi.ping ext'ernal supports.,

were complete in t'ime.t'o support i.ns.tallati.on duriqg,'the cycle 5 r.efuel:ing outages.

,In order to, sa't,isfy schedule comnitments, it was necessary to make in'terpr.etations of, LDR and NUREG 0661.requirements based.

upon. the best available,.information at the time of analysis.

Most of'he, interpretations, were originally es.tablished in 1979 and early 1980.,

A continuing effort to remove excessive conservati'sm,from load defin'i'tions and analysis "methods was made, parti,cularly, when that conservat.ism, would: result i,n unnecessary, impractical modifica.tions.

,When,'later information. on, load definitions and associated analys.is: methods.

became avai.lable, it was

compared, to the lprevious interpretations-..

The I.ater information was used for reanalys.is,.and,.associ.ated des,ign work i.f a si,gnifi,cant

'unconservat.ism, in the prev;ious, interpretation was i'ndicated.

'For example,,

the f:inal downcomer ti.ebar.'/V,-brac.ing

,modification resulted from; November 1'981 changes in t'e DBA condensation osci'llation lateral l.oad defin'ition.

Somet imes, later i pf orma't i on was used. to r emove excess i ve conservatism, in remaining.analysis, and'esign work.

For

example, the 1.1 SRSS load combination technique was permitted for torus attached piping analysis after NRC' final position on 'this subject

'was defined in Apri'1 '1983 by PUAR 'Reference 58.

An absolute surrrnation combination technique was required prior to that time..

Fi.nally, when 'the later information showed the previous load defini:tions and analysis methods to be adequate'ly, (but not excess'ively) conservative, the original interpretations were retained.

In these situations, reanalysis util.izing the later information would have been unnecessary and costly,.

and, in. some cases, would have resulted in delays in the installation of modifications.

Many of the PUAR quest,ions derive from situations where the original interpretations stated in pUAR Section 4 were used for analysis.

Justification for these interpretations was provided in. PUAR Section. 4, Section 5,

and Appendix C.

Additional technical 'justification,'ollows in the responses to.,specifi.c.questions.

on 'these topics.

Other PUAR questions simply request add'itional i.nformation, which's prov.ided in the responses.

GR-1 P,UAR. 00

It is TVA's posi't ion that the,BFN PVAR (Revision 1) and 'ou'r review question responses demor'~st'ra'te compl lance wi th thee intent of the Mark I Containmedt Long-Term Program and NUREG 0661 (i.,e

, to upgrade

'the contai'nm'ent system safety.

margins,,

for all,'postul'cited hydrodynamic loading conditi'on's,'o those intended by the original design spec,ifications).

On-this basis, we feel that all indicated safety concerns are fully and satisfactorily addr!es'sed,

'an'd we respectfully request a favorable final evalilatioh for the BFN LTP.

i L

GR-2 PVAR. 00

According to Section 4.2.5 of the

PUAR, BFN used the loads

. defined by the PULD and the LDR Sect:ion 4.3.2 for pressure 1'oads on the torus..

However, BFN applied a much smaller

. margin on, the LDR.load 't'han stipulated in NUREG 0661, page A-6.

The BFN margin, of '6.5 percent on the LDR upload is justif-ied in the BFN PUAR on the basis that the 15 pe'rcent margin r'econmended on page 39 of NUREG 0661 i"s unnecessary because the EPRI 1/12-seal'e model

'had the BFN geometry, and that the Acceptance Criteria (AC) margin of 21.5 percent should therefore be reduced by 15 per'cent to.yi'eld 6.5 percent.

This does not meet the intent of the AC.

The 15 perce'nt margin of NUREG 0661 was,i'mposed for'everal reasons (see pp.

36.-3'8 of NUREG 0661).,

the geometry being only one of the concerns.

Consequently, a full justification for the reduction of the margin from 21.5 percent to 6.5 percent is

needed, or the ability of the torus to withstand

'a 15 percent load increase mist be dern'on'strat'ed.

RESPONSE:,

'I The uncer'tainty margins us'ed for BFN pool swell load dei'ini,tion and the BFN spool swell.analysis procedure ensured conservati've structural response predictions.

Some justifi-cation for,this fact is given, in 'Section 4.2.5.2'of the PUAR.

Addi.tional justification follows:

-1.

,Uncertainties regarding.the.2D/3D test model results were minimized because the 1/4 scale 2D and 1/12 scale 3D models for the generic Mark I LTP tests were prototypical of BFN.,geometry.

2.

Significant, conservatism was added to the BFN pool swell load definition because fluid,compressibility effects in the vent system were not considered in the 1/4 scale plant unique tests.

This conse'rvatism is recogni'zed and quantified in Section 2.4 of Supplement 1

'to NUREG 0661..

3, BFN plant unique 1/4 scale tests for normal operating conditions were conducted at min'imum dP and maximum downcomer submergence, thereby ensuring upper-bound p'ool swell load predictions.

4.

BFN plant unique I/4 scale tests for 0.0 hP condi tions were. conducted at 0.0 ALP and maximum downcomer submergence, thus ensuring upper-bound pool swell load predictions.

BNL, l-l PUAR. 00

S.

The

'BFN v pool swel Sections vent syst 5.4.2.7 o

procedure incl'uded system an for torus a's s ump t:l o comb i'na t i ent system and toruS analysis procedures for 1

loai1ing Included 'slgnifican't

'conservati.sms.'.4.5 akid 6.'ll of'he IPUAR'surrm'arfze the BFht em analysis procedur,e.'ections 4.4.4 and f the PVAR sunmarize the Bl N torus a'nalysis S,igni'f.icant-analgtilcal r.onservatisms the tvio percent da'mginlg Iaskumption for vent'lysis, the 80 percent waI ter 'mahs'assumption

analysis, and the twd percent torus damping re for all operating'AP

pbol,

's'we11 load ons; 6.

The BFN uncertainty margin,fair poiol swel 1 loads (S.S

.percent) was,conservat i-vely-appl i ed to the pr edicted'orus response includling vent system input effects,.

wher eas the download, and upload margi ns i,n NEC' acceptance cr i ter ia (Appendi x A of hIUREG. 0661') are appl icable for torus hydrodynamic pressure loads '.only.

V.

The BFN uncertainty margin (t).5 pIercent)

exceeds, NRG's recommended download margin (5.~4.gedce'nt).,I,t also exceeds one,standai d deviation of ttie 'BFN I/4'cale

.resul ts for operating bP condi tions.

Those standard

" deviations were appr'oximately 3.6 percen't and.4,.0.

percent for upload,and download respectively.

8.,

BFN operati.ng AP,pool swel 1.dynamic resp<inses were conservatively combined wi th dynamic respons'es

.fram other load sources

.by, the: methods described i.n. Section 4.4.2 of the PUAR.

Additional assurance.

regarding any remaining upload concern is provided by the i'act that the'~BFN torus tiedown desigri lis not cont.rollied, by a pool, swel;1 load combination.

This would remain true even if an additional 15 percent upload ma'rgin were added 'for pool, swe.l,l loads.

BNL 1,-2 PUAR. 00

ITEM:

What margin was appl ied on the LDR download?

Is the downl;oad speci fication consis~tent wi th Section 2.3 of the the Acceptance 'Criteria?

RESPONSE':

A 6.'5 percent uncertainty margi:n, was applied'or 'both down-

,l,oad: and upload

.as, descr.ibed in, the response to,'BNL item l.

The speci.fica'tion Iin Section 2.3 of'RC's acceptance criteria requires.

a download. margin of 5.4 percent, based upon a peak download of '2700 pounds for BFN 'I/'4 seal'e oper at'i ng"hP,tes ts.

BNL 2-1 PUAR. 00

ITEM 3:

For what structures would the losid exceed acceptable levels i f the tor us pre. sure loads were made c&ndis'teh't wi th NUREG 0661?

By how much, and for what load combinations?

RESPONSE

To make the BFN torus pressure

'loads consistent with NUREG 0661 an additional 15 percent margin would 'be added to,th'e upload phase--if the other conservatisms

-in the BFN pool swell load definition were, disregarded.

The download margin would be reduced by 1 percent.

Assuming that the

same, conservative analytical procedure was applied, maximum stresses in the download phase would decrease slightly and maximum stresses in the uploaId phhs& Would'increase by less than 15 percent.,

(An increase of 5 to 10 percent is estimated.)

This level of potential stress increase could readily be compen. ated by rem'oval

'ofom6, of thI'onservatism in t,he analytical procedure and load combination technique.

Therefore, realistically, there" is no potential to overstress a

BFN stru0tatre by changing the torus pressure load" defi,nition to c'omply with NUHEG 0661 generic margins.

iBNL 3-1 PUAR. 00

ITEM 4:

Was the vent header impact load def ini tion of;pages 6-17 of the PUAR i.n. accordance wi.th Secti-on. 2.10.,1 of NUREG 066,1?

If.not, explain the differences and provide estimates showing that sufficient margin exists to acconmodate the NUREG load.

RESPONSE

The vent header impact 1'oad definition was not in accordance with-Section 2.10.1 of'he NUREG 0661 acceptance criteria.

S'ection 2.1:O.l addresses loads on a vent header deflector.

BFN does. not. have a vent header deflector;

however, the BFN vent headers were reinforced near the center of each non-vent bay as, a result of pool swell impact loading anal'ysis as described in Section 6.11;3 of the PUAR.

A typical, BFN header reinforcement installation is shown by PUAR Plates 7 and 8.

The BFN vent system pool swel.l impact load analys.is (PUAR Reference 17) and. header reinforcement. modi,fication design were: performed i.n 1979, pr.ior to the release of NUREG 0661.

The longitudinal velocity, and impact timing:profiles were based upon EPRI 1/12 scale, split orifice tests for operating 8P and 0.0. 4P pool swell conditions.

NUREG 0661 specified the use of a single "conservative."

profile for impact velocity and timing for all conditions.

However, a comparison of-the resulting peak. impact pressures on the BFN vent header showed that the existing analytical values were more conservative i'or the entire non-vent bay.

This was particularly true in the critical region where the BFN reinforcement, modification is located.

Therefore, within the non-vent bay i,t was conclude4 that the existing analysis results were conservative relative, to the revised load def,ini tion from NUREG 0661.

Within the vent bay the peak impact pressures would be somewhat higher with the revised load definition.

However, conservative estimates of the increased vent system stresses in this region showed all stresses to be less than 24 percent of allowables for the operatinghP case and 37 percent for the 0.0 AP. case.

Further consideration of this information leads to the basic conclusi.on that the 1979 analysis was appropriately conservative and sufficiently accurate to address all structural concerns of the BFN vent system for pool swell impact and drag loads.

Additional analysis was not necessary.

BNL 4-1 PUAR. 00

-ITEM 5:

Ner e the IOCA jet and bubble dr'ag loads for'FNvaluated i n accordance w'I th the LDR and NUREG 066'I?

RESPONSE

Yes, LOCA jet and bubble drag loads for,,BFN were evaluated in accordance with the LDR and, NUREG 0661 (See

PUAR, Appendix D, Sections 'D.l.1.2 and D.l.l.l, respectively; for di scus's i ons )'

'BNL S,-l PUAR. 00

For analyzing structures, af'fected by CO loads, the LDR and NUREG. 0661 prescribe absolute sumnation of the CO load harmonics at li-Hz intervals.

from.

1 to 50 Hz.

BFN used an alternate approach where:

(i) forcing. frequencies above 31'z were neglected, and (ii) four parti,cular load harmonics (the ones at 4-5, 5-6, 10-11, and 15-16 Hz)'ere added absolutely and added.

to the SRSS of'he remaini,ng 26.

Justify the neglect of forcing frequencies above 31'z for

('a) torus shell loads, and (b) submerged structure drag loads.

(Arguments about smal;I torus. response do not apply fordrag loads.)

Why were CO drag l.oads (page 4-4) analyzed i'or 1-31 Hz only:,. but. post-chug: drag loads (page 4-5) for 1-50 Hz?

RESPONSE

0 When the. DBA. CO load def:i:ni:tions were provided by, the

LDR, it soon

became, ap'parent that there were'ignificant inherent conservatisms,,

not the least of which was the lack of any information about the phasing relationships between the Fourier harmonics.

Clearly, conservatisms could have been maximized by applying all 50, CO harmonics using an absolute sunmation rule, and whi'le some might infer this approach from the LDR and NUREG 0661, it was not prescribed.

Various experts identified. specific conservatisms and recorrmended approaches that would allow more realistic accounting for the potential DBA CO event.

Some of the key findings by these experts f'ollow:

From PUAR Reference 19 (or equivalently.:

GE/NEDE-24840).,

Section 3.4:

(1)

The 5.5 Hz harmonic was the dominant content of the loading.

(2) "... a.ll harmonics appear to be randomly phased relative to the dominant harmonic at 5.5 Hz."

(3)

"... investigators have seen a tendency for a fixed phase relationship between the dominant harmonic and one or two mul tiples of the dominant (e.g.,

5.5, 11, and 16.5 Hz) from examination of data from individual pressure transducer records for the FSTF test, but showed random phasing for all other harmonics."

I BNL 6-1 PUAR. 00

~

s r r

~ -, ~

~

~

rf

~ r rrl rri h <<Ml r

~'

r9'irrr lh w ',

5 rh r ~wlehhh, M 'rtr hht

~TIE Ehhhr hh J(r'.rulrr'r 'h lrrhlh EMIrr 'l."s

~ ~

~

l9

(4)

EVhf le 1 t was admi tted that there appeared to be some relative per iodici ty of t'e harmonic ampl

l. tudes that would preclude the appropriateness of pure.

SRSS combination o1i'he harmonic amplitudes, it was stated that ",;.. it is highly improbable for miore than about three harmonics to be'orst case phased at any one time...."

(5) "...

a rule which requires about three harmonics to be absolute combined INi th all additional harmonics SRSS combined is consistent with the as. umptfon of steady-state periodic amplitudes arid randem phasi ng."

-(6)

The LDR amplitudes are-defined with significant conservatisms as can be seen from Figure 3-11 (especially in the frequency range from 40 to 50 Hz where most LDR amplitudes ar,e much more than 100

,percent greater than thie average FSTF amplitudes).

(7)

A'iso from Figure 3-11, i t can be seen t',hat actual FSTF ampl f tudes seem to show that CO has relati,vely 1 lttie' requency content above 30 Hz.

From PUAR Reference 42, Section 4:

(8)

Conclusion

'No.

2 states thats "For structures wi*th frequency content similar to the FSTF or, Oyster Creek torus encl supports, only the harrnonfe responses belowi 30 Hz need to be computed and in'eluded."

Based, on the f'indings 1 isted and our own best technical judgment,,

T'VA feel,s that neglect of the fair cf ng, frequenci e's above 30 Hz, is justif;ied for".

(a) torus.,shel 1

loacls (b) submiergecl structure drag loadS An explanation of our reasoning followst As stated f n findi;ng (7), there is 1 i ttie frequency content of the CO lioacling. above 30 Hz; this.is the main reason for neglect i ng 1 t.,

Addi tional ly, for the BFN 'o'rus (even af ter substantial modf fications that resul ted'irectly fr'om CO loads analysis),

t',he responsive structural modes occur at frequenei es below,30 Hz.

Shapes 'f the'igher fr'equency modes of the torus will niot part.iei'pa'te'igni ficantly wi th the shape of the CO pressure distr ibution.

This argument al so appl ies for t'e post-chug loacli ng on the.torus since i

t'as the same distribution -shape as iCO wi th only the harmonic ampl i tude cioef fiei ents bei ng di ffec ent.

BNL 6-2 PUAR. 00

For harmonic for cing components wi th.frequencies more than 1.5. t.imes-the natural struc:tural frequencies, the dynamic load factors (DLFs) become less than 1.0.

For harmon.ic forces with frequencies greater than twi ce the natural structural frequencies, the DLFs are small and asymptotically approach zero.

In this range,,the forcing components would have negligible effect on the structure.

This is. the case wi.th the torus f'r high frequency

.harmonics.

Further, empirical evidence of the adequacy of the BFN analytical approach for the DBA CO loading on the torus is provided in the responses to BNL item 7 (see Table BNL-7-1)..

Similar evidence is provided for the chugging.

loading on. the. torus in the response to FRC item 7.(see Tabl e FRC-'7-l,).

Submerged s.tructures are a different matter, however.

Init'ially, most submerged structures were primar'ily responsive in the lower frequency ranges and were dominated by the DBA CO harmonics, In.order to avoi.d highly. amplif,ied responses due to CO, pre-.chug, and S/RV load definjtions, virtually every.submerged structure required substantial'edi'fication to stiffen and strenghten it.

These modifications took some.of the submerged structures into a responsive range with.post-.chug fluid,drag.

While it was, impractical to stiffen submerged structures (most of them 1nternal portions of large piping systems) above the, post-chug forcing frequencies and stil'1 maintain v1able designs for thermal loa'ds, it was possible, after-many iterations, to obtain 'designs that had sufficient strength to,meet allowables for all load.combinations.

This left BFN with

~ st:iff submerged structures that were well within a range of post-chug

drag, thus, necessitating use of. the full 0 to, 50 Hz range of the post-chug,.defini-tion prescr.ibed in the LDR.

FUAR Pla.tes II,:I'8, 1'9, 20, and

'21 show some of the stiff'ened submerged structures.

I 4

~iii e

,I BNL 6-3 PUAR. 00

~f)'

~

~ t...Q, r ~ g r aV Jr' r

~Ilail

.;p prV

+,I ~'~~ tt 8, tYr/ std;,,'

C ~, t.

~ s ', 7, '

~

~

The. appr oach of GE/NEDE 24840 - which i <<

i tsIel f a depar ture from the LDR - call's for taking the sum of the four harmonics which produce the highest structural

response, and adding them to the SRSS of the remaining harmonics.

Were the forcing functions at 4-5, 5-6, 10-11, and 15-16 Hz, the ones which produced the highest structural response for both tor us shell and all draj loads'?

In the work done by SMA (References 19 and

.42 of the BPN PUAR), the absolutI.

surmnation of the four 'highest Iharmonics had nothing to do with phase relationsh,ips, but was an artifice used to arrive at an 84 ipercent nonexceedance pr'obability (NEP).

Bas:ed on the discussl,on in the P~UAR, BFN's proIcedure does not guarantee an NEP of 84 ipercent.

Justify BFN's departure from the recommended procedure and/or demons'trate structural margins which would adequately cover incr easIes in tlute CO loads.

Was Alternate 4 of the CO basIel inc r igid wal 1 pressure spectrum applied to BPN!'ESPONSE:

While the approach recommended in PUAR Reference 19 (or GE/NEDE-24840) i s a diepar ture from thIe LDR, TVA bel 1 eves i.t is well justified by the thorough,studies and findings of many experts.

Those fii'ndings clearly stress the extreme conservatism, hence inappropriateness,,

of absolute sutrrnation of all response harmonics As the approach is specified, it requires the identification of the three or four highest response harmonics of a

structure to be combined absolutely with the SRSS of those remaining.

ThIis identification, however, ls an impracticable task when one consiiders the number of. structures to be analyzed and the response quantities of interest for each (e.g.,

dispiiacemenit, acceleration, force, stress, stress intensities).

Also, while this approach may guarantee 50 or 84 percent NEP for the CO loading alone, there can be no such claim for the controlling dksign ibad combinations involving CO since the pol',nts of maximum respo'nses for CO load combinations are likely to be different than for CO alone.

Therefore, TVA chose to vary slightly from the approach suggested in Section 6 of PUAR Reference 19.

The approach used was justifiable encl practical for timely, cost efficient implementation.

I t cannot'd g'ua'rar'>te'ed fcir both the tor us shel 1

and al 1 submerged structures that forcing functions at 4>>5, 5-6, 10-11, and 15-16 Hz were t'e ones producing the highest structural responses, although for some they may be--

especially for t'e torus shell since its primary response BNL 7-1

.PUAR. 00

occurs around

.these frequenci es.

However, f t i s impor tant to note that the suggestion, to use the four highest response harmonics was an art-iflce to obtain an 84 percent NEP of an artificial load definiti~on;-one which Dr.

A'I'an Bilanin has stated is 33 percent conservative just due to the presence of the bulkheads on the FSTF (see Reference

'BNL-'7.,l, Section 1,

Equa t-i on 1..7,),.

There are the, addi tional conservati'sms of the LDR prescribed ampl i tudes, as already addressed by 'f inding. 6 1 isted in our response to BNL Item 6.

And, speci fically concerning torus

.shell responses,,

TVA has the conser vat f sm of'aving appl t ed a

maximum envelope of the three LDR'lternatives speci fied for harmonics between,4-16 Hz rather.

than select'ing the one alternative producing maximum response.

To pursue the issue of why TVA chose forcing functions at 4-5, 5-6, 1'0-11, and 15-.16 Hz,. the following arguments are offered:

(1)

As contended, it is a practical impossibility to identify the three or, four-highest CO response harmonics for, all structures and,response quantities.

Therefore, TVA sought

a. practical al'terna'tive that would maintain some conservatism over a pure SRSS combinat,ion of the CO harmonics.

Because of the reported indications that there may be s'ome fixed phase relationship be, tween the dominant harmonic at 5.5 Hz and its first few mul"tiples, we decided to use the 5-6, 10-1'1, and 15'-16 Hz LDR harmonics.

The 4-5 Hz.harmonic was added to. t'his li'st since it was the next largest amplitude harmonic fn the LDR definition.

I;t should be noted that three of these (4-5, 5-.6, and 10-..11 Hz) are the 'highest of al'1 the amplitudes provided in the LDR.

(2)

Comparing the LDR prescribed'BA CO and post-chug amplitudes, it can be seen that for s'truc$ ures 'having primary response modes bel'ow about 20 Hz, CO load combinations woul'd be expected'o control for design since the CO amplitudes: below 20 Hz're generally larger than the post-chug amplitudes.

Structures having primary response modes, above 20'z would'ost likely be controll'.ed for design by post-chug load combfnat:i'ons since post-chug amplitudes are hi'gher in this range.

Therefore, by picking thi'ee of'he hi.ghest

'CO amplitudes for absolute suranation, the three potentially most damaging load components are assured of receiving conservative combination in the response predictions for structures likely to be cont'rolled in their design by CO load, combinations.

BNL 7-2 PUAR. 00

I I

JI I

~

~

1

~

4 I

q,j) l (3)

(4)

PUAR -Ref'er ence 19 reconwnen~'fs a

pr ecedure for obta ini'ng" 50 percent or-84 percent NEPs and -shows that this is achi'e'ved when t,he three or f'our. hi.ghest res'ponse harmonics are combined absbluteiiy.

This is based on compar'isons with, predicted response values a't these probabilities as taken from constructed CDF curves for both'he FSTF and Oyster Creek. torus.

As can be seen from; the CDF curve's of Figrlireis 4-6 through 4-10 from Reference 19 all hav'e r'elatively small variance as'ndicr<ted by their steep slope.

It can also be, seen th'at the response values pr edi.cted'y total absolute sum of all 'harmonics is wel,l abov'ei the respons'e value assocliated wiith 1~00'ercent NEP.

Therefore, while even a, totrLI SRSS combination of,', all heirmonics would be only

~sl i~ht~l unconservative relative to the CDF 50 percer>t and 84 per cent

.NEP" response

values, a 'total 'absolute combination would'e grossly overconse'r vitive*

IUhi'le. there fIs no reason to suspect that a total SRSS combirration c>f,all harmonic responses would prodluce. a

.resporise valise as low even as that associated. with a 0 percent NEP; it 1<< interesting to observe from PUAR Reference, 19, Figures 4-6, 4-7, 8, 4-9, and 4.-10, that the percentage differences between the 0 and 50 per;cent NHP response values

are, 11.3, 15.6 16.3, 24.i0,;and 18.,2. percent, respectiveily.

Th,erefore, even a -piLrre SRSS combination rule wouldl result in at least a

0 percent NEE',esponse

value, meaning the potential unconservatism could be no more than the above pericer<tage

-di fferences between

.0 an'd 50 percer>t NEP vilueii.

Further, since TVA's approach is mor' con'servat'ive than pure SRSS,,our predicted response values would.be sti'll less of.a percentige difference.

All of'these arguments mean that any, slight uncon-servatism there may be in TVA'.s approach is much more than -offset. by.the:inherent conservatism in the FSTF based'oad defini"t,ions.

.Inherent in those definitions, as iil,r.eady, meintioned,

'i,s at least.

a 33 percent conservat'ism acco'rding to Dr. Alan Bila'nin.

So,.while our approach doesot" ri'gorously assure

-84

'per ceait NEP -alf the conserva't i'e LDR load def ini t ions

'per se,

-we ':feel very confident.that thi',s

leVel, or, more woul,d be achieved: if more r'ealidti'c load definition eind anaiyssis techniques were,possib)'e'.

A strong indication of the. conservati'sm. of the. BFN BBA', CO'dna',lysis is sekn

'in 'the, attach'ed=,Table BNL-7-1.

iBFN,stresses and reaction, forces are. presented, factored as nearly as possible to an FSTF-equi. valent basfs, and compared to measured and caiciilated NEP vjaliies for the FSTF.

BNL 7-3 PUAR.00

Looking at Table BNL-7-1, the factored BFN cradle support pad reaction forces are seen to be conservative with respect to the 84 percent NEP values (this would indicate similar conservatism of the stresses in the critical cradle regions which required extensive modifications).

Also, the BFN BDC membrane stress intensity (f'actored

-to an FSTF-equivalent basis) is almost exactly the same as the 84 percent NEP value for the FSTF.

While this degree of closeness may be coincidental, it does provide additional evidence that there is no large deficiency in stress intens'ity predictions in the BFN analysis.

Finally, there are two additional conservatisms worth noting about the BFN analyti'cal approach.

Firs.t, 2

percent damping was used for the'BA plus S/RV 'load combination analyses of the torus and submerged s tructures even though higher damping is justifiable because of the higher service level allowables.

Second, conservative load combination techniques were applied (see PUAR Section 4.4.2).

Concerning the final question about Alternate 4 of the CO baseline rigid wall pressure

spectrum, we do not know to what this refers.

Additional'eference:

BNL-7.1 Structural Mechanics Associates, "A Statistical Basis for Load Factors Appropriate for Use with CO Harmonic Response Combination Design Rules," Report No.

SMA 12101.04-R003D, March 1982.

Addendum In the September 5,

1984 meeting, BNL expressed remaining concerns over. TVA's use of the absolute sum of the four highest DBA.CO harmonic amplitudes, rather than the four harmonics. causing the greatest response.

BNL also asked:

"How much greater could the DBA CO load be without exceeding allowables"7 (paraphrased)

Oui response considers the torus separately from the tiedown system for reasons which are explained below.

The most highly str'essed regions of the torus, relative to allowables, are in the cradle adjacent to the scab plates described in PUAR Section 5.2.4.3.

(Also see the response to FRC 'Item.11.)

The controlling load combination is number

, 14 which'oes not include DBA CO.

There are large margins BNL 7-4 PUAR.00

for all load! combinati'ons involving DBA'O.

If one assume's that cradle" stresses are di,rectly proportional to the net compressive

loads, the DBA CO rehctio'ns'ould be 2.5 times the values computed by the TVA, analys'is'fthout exceeding allowables.

The tiedown system,

.on the other

hand, is controlled by lolad combination number 2'7 which incltzdes DBA CO..

Tiedown stresses are directly propor tionhl 'to'est 'uplift loads.

From computer calculated responses to the '0 to 30 Hz unit amplitude harmonics and hand calculations, the increase in reactions by taking the four highest responses rather thah,'he responsIes to the four highes.t a!mpl'itudes',

has been quantified.

The conservatism of TVA's method of enveloping the three alternate sets of amplitudes has also been quan-tified.

These calculations show that the DBA CO reactions presented in Table BIIL-7-1 would be 9 percent higher i"f th0 four highest harmonic response's had b'eel u!sed,'while retaining the conservatisrn of enveloping the a]!ternate amplitudes.

If the i,ndividual altern'ate amplitude sets

'are

used, the reactions would be 5.4 percent higher than those'n the table.

With the 5.4 per'cent

increase, the tiedown system stresses do not exceed allowables.

It is intpo'rtant to reemphasize that the SIM method of Reference 19 pro>'rides large margins'or support reactions.

It is

'alamo 'no'tee'd'or'th'y that the limiting load combi'nation (number,27) includes the highly unlikely simultaneous:

occurrence'of the'maximum responses due to the safe shutdown earthquhk6, DBA CO, and a single valve S/RV actuation.

TVA combir>edi t'geese three dynatnic events absolutely.

If,. 1'or example, a I.l SRSS combination of the three dynamic loads had been

used, the uplift loads would have been barely great enough to *vercome the deadweight.

BNL 7-5i PUAR.OO

TRBLE BNL-7-i COMPARISON OF FSTF AND BRONNS FERRY DBA C.O.

RESPONSES BFN RESPONSES MEASURED BFN FACTORED TO FSTFi PER CALCULATED EQUIVALENT REF.I9

RESPONSE

QUANTITY RESULTS FSTF TABLE 7-2 50%

NEP FSTFi PER REF. I9 TABLE 6-2 84/

NEP FSTFr PER REF.19 TABLE 6-2 BDC.MEMBRANE STRESS INTENSITY (KSI) 1.SB 2.77'"

2.6 2.47 2.79 INSIDE REACTION (KIPS)

OUTSIDE REACTION (KIPS),

306 333 202~

220~

93 110 122 140 140 159 I (R/T)FSTF] I 1

BFN L (R/T) BFN J LPLANT UNIQUE PRESSURE FACTOR=0.85 BFN HHEREz R = MINOR RADIUS OF THE TORUS' SHELL THICKNESS'ND O' STRESS INTENSITY (2)

FSTF EQUIVALENT SUPPORT REACTIONS =

(BFN REACTION)

BFN POOL AREA PER CRADLE 1

~LANT UNIQUE PRESSURE FACTOR

ITEM 8:

Were pne-chug loads applied to O'FN according to the LDR and NUREG 0661 speci'fications regarding amplitude, circumferential and vertical distribu'tion and cycle durat'ion?

If not, provide quantitative 'justification,

RESPONSE

l l'

Th'e pre.-chug loads were applied in coinpl etc 'accordance with the LDR and NIJREG 06(iI, including don'siderations

'of

.amplitude, circumferential and vert, ical di'stribution,, and cycle duratio&.

With regard to the latter, the respohses of 71 shell mcides due to siz Independen1).

harmonic (implying in'finite dura't,ion) forcing funct.ionls, fiine-'u'ned to the structural:frequencies, were enveloped for all points in the'odel.

No credit was taken for thei finite duration. of the'ctual even't.

~ 4 ~

~i BNL 8-1 PVAR,.00

For post.-chug

loads, were the harmonic forci,ng functions used in the 1-30 Hz range the ones speci, fied in the

LDR, and were they applied in the manner prescribed

.in the LDR?

if not,. justif'y departures.

RESPONSE

h For both.the pos:t-chug, pressure loads applied to the torus and the drag loads; appl'1;ed to internal'tructures.,

the harmoni'c forcing funct.ions, used in the 1-,30,Hz range were those specif.ied in, the LDR,, and, they were appl:ied in the

manner, prescri,bed by the LDR.

Appendix D,. Section D.1,.2.4.2 o,f the BFN PUAR expl'ai;ns. how the, LDR prescribed, method was applied. for,submerged. structures by -consider,ing

.the closes't

'owncomer l.oad sources.

together. with worst-case phasing be. tween the sources.

BNL 9-1 PUAR. 00

ITEM 10.:.

The finite element model. of Figure 6-7 shows amputated downcomers.

How were thy CO and CH loads applied to these amputated downcomers?

RESPONSE

The finite element model in Figure 6-.7 was used to examine more clos'ely t,'he vent downcomer/head'er intersection.

Loads for the dii ffereni: combir>ation even'ts'hich i'nclud'ed condensatiion osci Il'ation. and chugging iwe'ire extracted:

from the 45.o vent system beam model ',(PUAR Ffg'use 6-2) at'he node.

representing the downcomer/vent header. shel.l intersection.

~

These loads were then:input into the truncated model't. the end of the downcome,r.

(The downcomer end's comprised of

,rigid beams connected by a, node in the center.)

The appro-priate stresses were then extracte'd

.and compared to the stress allowables.

BNL 10-1 PUAR 00

ITEM 11:

Were the CO I'oads appli'ed to the downcomers in accordance wi'th the L'DRand NUREG 0661?

Were the eight load cases of Secti,on 4.4.3'.2 of the LDR'nalyzed for all relevant vent system parts, (including main vent/vent 'header intersection, drywell/main vent interaction, downcomer/vent header I'ntersection,,

etc.)?

The:PUAR expl'I'cftly mentions consi'deri'ng different load'ases onl'y f'r the downcomer/

tiebar Intersection,

'and 'fn that case refers only to four load'ases rather than the eight of the LDR (Section 6.V 1.2 1).

Why is 2

5 percent damping justified for BFN f'r CO lateral load analysis?

Note that Table 6-10 shows

.no.margin for the downcomer/vent header intersection in Load Combination 27 which involves CO.

RESPONSE':

The CO downcomer.

loads Were applied in a manner consistent wi.th the intent o'f the LDR and NUREG 0661.

The load from the differential pressure for one 'downcomer was added to

" the internal pressure, that occurs simultaneously in all downcomers, thereby producing a higher load in one downcomer fn 'each paii.

Thus, from Figure 4.4.3.4'f th'e

LDR, a

darkened downcomer indicated that the di.fferential and internal pressures were working together simultaneously, whereas

'the other downc'orner in the pa'i,r experfenced only the internal pressure.

Based on the primary downcomer swing frequency "extracted from a modal analysis of the system, s'Inusoidal forcing, functions were applied to downcomer pairs defined by Figure 4.4.3-3 in the LDR.

'Since the primary swing mode f'r t'e BFN system occurs at approximately 8 Hz, the 1st,

2nd, and 3rd harmonics were applied In the 4, 8,- and.l2 Hz ranges, respectively.

Another aspect of. the BFN analysis was the application of the first harmonic forces to the coincident.8-Hz swing frequency.

Response

of the system to this single frequency load envelops the sum of the three harmonfcs defined by the LDR.

This load was subsequently

.applied in the stress evaluation.

Also, the first harmonic force amplitudes were applied with 16,and 24'z sinusoidal functions to verIfy that higher-frequency responses do not impact the, total CO response.

In actuality, 30 individual sinusoidal functions were applied for each load case to account for potential response at the 1/2, I, l-l/2, '2, and 2-1/2 harmonfcs of the six discreet primary swing mode frequencies in the 8-9 Hz range.

Note that this load was in addition to the vent system CO loads which were applied In a separate analysis.

BNL 11-.1 PUAR. 00

Four load'ases were ini tial I'y analyzed'or downcomer CO literal loads as indicated by PUAR Figure 6-12.

From inspection, of the first, four load

~+ases'an'd resulting

stresses, it visas evident that in each

instance, the.worst effect'n a

daiwnc'orner would occur on one that was loca'ted on the inboarci side of the vent hea'de'r.

Furthermore, the highest loadeci'owncomer (unreinfoi'ced

'She'1 I) i esulted, from the applicati,on of L'oad Case, I (differential, pressure applied to all, inboard downcomers).

Sine!e 'the sec'ond four.load cases defined in revision 2', of the LDRiarie mirror images of the f,irst four cases and greater response resul,tied from Case I, it was resolved that'he worst'oading had aire,ady been analyzed.

Therefore, no further analysi,s was performed.

All'ri-tical 'locations of the ven DBA CO lateral Ioaci combinations.,

s,tress margiin rela'tive to Service CO, combination, in Table 6-10 of t the primary plus secondary stress

,level should be evaluated with co conservatism in the load comb'inat Level C combination) and the.fact Case I is the worst of the eight configurations.

t systEm were evaluated

'fbr As noted in. item ll, the Leviel. B.allowables for the he PUAR is-close to 1.0 for catiegory.

This stress

'sidieration of the, ion (event 21 is a Service that clowncomer lateral Load potential load Prel iminary anal'ysis of tt>e BFN c.ontat,nment vent system for the DBA CO Iateril load def-initidn 'indida't'ed'

'su'rface stress level in the vent header shell.near the down'comer that approached the yield stress value for SA-516.Grade VO steel.

Under this'ituatio'n the tot'al stress level for l,oad

.combination No.

21 would not me'et the allowable for primary plus secondary stress range.

In an effo'rt to avoi.d additional modification

( I.e., downcorner/vent header reinforcement.gu,ssets),

the 2 perceinti reconmended damping rat'io was investigated as a source of excessive conservatism.

Per Regulatory Guide 1.61,,

a 3 percent clamping value 'is r econmended for arialysi,s c>f large diamet,.er, piiping systems for the safe shutdown earthquake.

Furthermoire, the cal'culated damping value 'resulting from the snap putll test of a tied downcomer arrangiement (seE. Figure,4-5 in Reference 5 to supplement I of I%UREG Oi861) was found to be approximately 2'.'2 percent for a 50 percent yield; Also, the damping; versus strain curve iridicates ia rapid IIIIcrease in damping above the 50 percent yield'evel.

Based oui these findiings, i t iias concluded that a damping 'value greater than 2 percent but less tha'n 3 per cient is appropr tate for the Browns Ferry c'o'nf iguration.

~ The 2.5 percent value was selected as the midpoint of the',2 percent:range and utilized for the DBA COi downcomer lateral analysis.

Resul tinIg surface stre: ses. in the vent iheader shell at the down'comer intersection're 20 to 24 ksi (f'r DBA CO,.loadlipg,:alone) as compcired to -a yield stress value of,'32.6 ksi f'r SA-516'rade VO steel at 400oF BHL 11-2 PUAR. 00

Therefore, the 2.5 percent; damping value is justi.f led based on:

(1.)

The s'tructural response of the Browns Fo"ry downcomer/

vent header con'f igurat ion.

(2)

The projected, resul"-ts of the snap test for h'igher ini ti,al str ess l,evel,s.

4 (3)

The damping criteria, del ineated in Regulatory Guide

1. 61.

,)I

~ jt; 0%

~ Yi

~ l

'a I'

~

I

(

BNL 11-3 PUAR. 00

ITEM I 2'.

1Veie.the, chugging loaiis appl'f,.ed to th'e.d2>wncomers:in accordanc'e with the LI)R. arid NUREG 0661?

.We!r.e tlie.mu.l triv.ent'chugging lo'ads. accounted for, on1 l v'en't, system,part's in

~acco'r dance. wi.th;the LI)R 'and NUREG 0(i61',?

Note that accord ing to Table 6'-'10, IIoad, Comb inat ion:15, which involves CH,'uis'..i ela t iv,ely,t i,t,t le, margrin.I RESPONSE':

Yes, the chuggiing loads. were app'll'ed,;to, the vent system i'ri

'accordance, with,'the LI)R.and,,lIUREG 0661.

LCCA,chuggin'g.load's, included post-rchug

drag, chugging late'ral, acoustic yerit system pre,ssure ciscill'ation,'and gross y'ent syst'm pre.~su.re'sci 1 lat iaIn, which:were,aplpl ied tIo the bIeam models shoiivn in'PUAR'igur'es

'.6-2 and. 6-3.

'.Responsea Were determined from

-those models and- 'loca'.I stresses w4r6 ctalt.'u/at'ed'.for the cr i t'ica*l 'loca t ii'oats..

4 I,

212 I

I BNL 12-.1

~PUAR;, 0 0

ITEM 13:

What hydrodynamic load definit ion was used for the vent pipe drain referred to on page 6-10 and shown on Figure 6-8?

RES PONSE:

From a Stardyne model of the vent drain and support

system, the fundamental natural frequency of the system was found to be 35.1 Hz.

Clearly, high ampli tude harmonics.. of post-chug around this frequency would lead to a strong expectation that a load combination involving post-chug woul'd be

'ontrolling; although, the possibi lity of a combination with pool swell was also considered.

From investigations of all potentially controlling design load combinations (including associated service level allowables) it was determined that combination ll (see Figure 4..3-1 of NUREG 0661) was controlling.

Specifically, the design case determined to be controlling was the SBA combination of S/RV plus post-chug fluid drag loads under service.

level A allowables.

The dynamic loads were very conservat.ively accounted for by the "Equivalent Static Load Method" explained in Appendix D, Section D;1.2.3 of the BFN PUAR report.

All 50 post-chug bubble source amplit'udes and. FSI accelleration coefficients were summed and used as multipliers of the unit forces (For)BUB and (For)FSI described in Equations D.1.2-6 and D.1.2-8, respectivel'y.

To these a resonant DLF = 25 (assuming 2 percent dampingI) was conservatively applied.

The S/RV'oad, contributions were applied with a harmonic DLF = 1.2 based.

on the ra.tio of maximum S/RV bubble frequency-to-system frequency, of 1'4.7/35.1 (again assuming 2 percent damping).

Addendum In the September 5,

1984 meeting,,

BNL's consultant, Professor Son'in of MIT, asked if the potential for chugging

through, the.

vent drain pipe and the effects of the resulting lateral loads had been considered.

TVA responded that the LDR did not include a method for defining such. loads, and the effects could therefore not be evaluated.

Professor Sonin then asked to be provided the properties and dimensions of the drain pipe and its support and the drag loads which had been applied to them.

BNL 13-1 PUAR.00

A set of five pages of calcula tions os the ef fec ts of S'/RV and pos t-chug drag loads, were subse'quent ly transmi t ted to Professor SonIi'n through l&C.

The calcula t ions show the post-chug draj~ loads, part;icular ly, were, d'ef ined in an extremely,, conservat ive manner wh'ich should'ompensate f'r'he lick of a directly applied ~la'teial. chugging load.

BNL,, 13-2 PUAR. 00

ITEM I4.$

Combining individual S/RV shell pressures by SRSS to obtain mul tiple valve shell pressures is an exception to the AC.

Jus t i fy thi s,procedur e for BFN.

RESPONSE$

The use of SRSS to obtain multiple valve she'1'1'ressures for analysis of S/RV discharges is justified by the BFN plant unique S/RV tests (PUAR Reference 41) and the correlation of analysis and test results (PUAR Appendix C).

Section C.3.2 of the PUAR specifically addresses this issue.

The measured peak shell pressures during multiple valve tests were approximately 4'5 percent of the analysis values for single.valve tests and 54 percent of the analysis values for multiple valve tests.

Thus the multiple valve test pressures are correlated by a 1.2 SRSS of single valve test pressures, but the overall BFN analysis and design approach was clearly conservative,.

It is also notewort'hy that considerable care was taken in the BFN test to ensure simultaneous actuation of three S/RVs with adjacent discharge locations in the torus.

This represents a "worst loca'tion" in the torus.

Referring to PUAR Figure 7-3, S/RVs D, E, and M were actuated simultaneously

'i'or mul.tipl e valve tests.

S/RV E was actuated for single valve tests.

Excellent repeatability was demonstrated for both test series (five single valve tests and four multiple valve tests.)

BNL 14-1 PUAR.00

ITEM I5:

Clarify the statement that the torus was analyzed quasi-statica'lly for S'/RV hydrodynamic shell pres. ures.

Where does g(t), i.e.,

the wave form of the pressure

history, in the expression on page 5-1'3 of the PUAR come from?

Are pressure,s applied statically as stated on page 5-12 or is there a time variation as implied by the expression on page 5-13?

RESPONSE

The wave form,,of t'e pressure history, g(t) was generated by the QBUBS02 computer code.

The pressures were applied statically to the torus shell to determine torus stresses deflecti ons, and support loads.

Subsystems,,

such as attached piping, weve analyzed dynamically for, the acceleration resp'onse resulting from the assauned shell mo'tion (see page

'5-14 of the PUAR).,

BNL 15-1 PUAR. 00

~,

~

ITEM 16,:.

't t

Provide the fol lowi'ng addi tional information regarding the in-plant. S/RV: tests conducted at. BFN and. the S/RV design loads extrapolated from the tests:

t 1.0 Descripti'on.of the tes"ted: Quencher Device-1.1 Drawings showing deta.ils of the quencher geometry-plan, elevat.ion,,

arm length, arm diameter, hole arrangement, spaci'ng,

size, etc.,

1.2'ocation of quencher device relative to suppression pool boundaries and suppression.

pool surface.

2;,0 1.3 Any difference between the tested quencher configu-rat-ion and'. the Monti.cello version. (as described in GE/NEDE-24542-P) highlighted and quantified.

A descript'ion. of the loads observed during testing-2.1 Peak overpressure

('POP) and underpin-essure (PUP)

.recorded on the torus shell during each relevant S/RY, actua,t i:on.

t 2.2 A tmeas.ure of the frequency, content of each pressure

- signature.,

3..0 A description of the test conditions 3.1 Geometry of the tested SRVDL ('diameter,

length,

'free volume, and routing below pool, surface).

3.2 Geometry of any SRVDLs in. the plant that differ significantly from the tested SRVDL.

-4. 0 3.3.

S/RV steam flow rate (MS,), pool'emperature (TPL),

pipe temperature (TP), wate'r leg

~ length (LW) and.

pressure di:fferential (hP),. if any, for each test.

3.4 Minimum AP permitted by NRC Technical "Specification and corresponding LW for all 'SRYDLs.

A description of the design conditions for each load case used for design-4.1 Geometry of all SRVDLs involved and thei r azimuthal location in the torus.

4.2 TP, TPL, MS, hP, and LW for al,l SRVDLS involved.

'BNL 1 6" 1 PUAR. 00

5.0 A description iof the design loads for each load case'-

5.1 5.2, 5.3 5.4 hformali:zed pressure signature.

Single valve POP/PUP values.

Spatial atteriuiati'on of the POP/PUP values (i f this di ffera from, the LDR methodology,,

suf f1'ci'ient addi t ional -torus shell pressure data must be supp 1 i end to just i fy,such dev ia t ion),i Frequency range considered.

RESPONSE

1.0 Description of thie tested quencher device The plan view of all BFN quenchiers ln units

.1

<'md 2

is shown on TVA drawing 47W401-7.

For unit 3 thk plan view is shown on drawing 47W401-3.

The tested quenchers were in unlit 2 at azirnuths 78o-45'D),

101o 15'E)

~

and 123o,45'M)

S/RV E. w actuated. f'r single valve, tests and all three (0,

E,,

and M) were actuated fior mu 1 t iple valise tests.,

These plan vi,ews corriespond to PUAR Figur'e 3

~

BFN quencher arm details are shown on TVA drawing 47W401-5.

Al 1 BFN quencher s are ident ical in dies,ig n.

Copies of all refer enced drawings are available for r ev i ew.

Oi 1'. 2 1.3 The BFN quencher device loca tilons are shown on TVA drawings 47W401-3, 47W401-5, and 47W401-.7.

Hach quen'cheer center li. ne is at elevat ion 526.-5 wh ich is 50 feet above the bot tom of the torus shel.l.

This yields a submergence of appi~oximately 10.0 feet.

Typical BFN quencher installations are shown on PUAR p la tes 10,,

11, and 13.

The BFN quencher device u t i Aires the previously.

ex i s t i ng 10-'i nch ramshead as i nd i ca ted on drawi'ng'7W401-5, while the Mont icel lo version has a

12-inch ramshead.

The BFN device has a

10-inlch2-inch reiducer between the ramshead and quencher arm whi 1'e the Monticel lo gersion does not..BFNnd'onticello quencher arm designs are identical except for minor var i,'at iohs in support type and BNL 16-2 PUAR.00'L

~

~

~

location.

The BFN weld cap hole pattern matches the pattern

.shown i',Figure 1-2 of NEDE-24542-P; it does not match

.the pattern in. Figure 1-3 of that report.

2.0

A descripti'on of the loads observed during. testing 2.1 Torus shell pressures resulting from the S/RV test are discussed in Appendix C, Paragraph C.5..2 of the PUAR.

Table C-2 of the=PUAR presents a 'compari,son of the analyt'i~cally predi'cted pressures versus the average o'f the peak.pressures from =e'ach test.

Thi's information 'is from the TES: Report No.

5172 (PUAR Reference 41).

Pages 1 through 54 of Volume III of the TES report show the pressure traces of the t'orus shell for each test.

'A suranary of the maximum and minimum.press'ures (POP and PUP) reco'rded during each test are shown, in Tabl:es BNL-1'6-1 and BNL-16-2.

2.2 The.pressure traces for all locati,ons and each test are f'ound i'n the TES Report (PUAR'Reference 41).

All pressure traces are similar i'n shape ahd primary f:reque'ncy; The 'primary frequency 'of the

,pressure traces ranges f.rom 5.'5 to 6.5 Hz.

Typical pressure,'traces are shown i,n F'igure BNL-1'6-'1'.

3.0 A 'description of the test conditions 3.1 R 3.2 The geometry of all SRVDI s.

i.s shown on the TVA 47W401 drawing ser'ies.

All di scharge 1 ines are IO" SCH 40 i,n the drywel1 and 10" SCH 80 or 1'.0" SCH 60 in the. wetwell.

The routi;ng for all lines below the pool is the same.

The 'rout'ing in the torus above the

.pool can be grouped into two categories--

lon'g lines and short lines.

SRVDL, E, a long line, was chosen for the single valve 'tests since the long line 'should rept esent the

'orst case for S/RV blowdown.

S/RV's D, E,

and M.were actuated simultaneously for the multipl'e valve tests.

The initial gas volume foi all lines is shown in Tabl.e BNL-16-3.

Also, see 1-.1 above.

BNL 1 6-3 PUAR. 00

~

~

~

3.3,

-Test Conditions Line.D

.Line S

Line M MS;, lb/sec TP.,

oF LWid F'T QP; psid

,26 8 78-&1 238'-355 7,.'0 1 e 2.-1'.33 268 78-81 220-3V 9 7.0 1.2-:1.33 268 V8-81 246-361 V.'0

l. 2-le 33

+lnit.ial aIad final-.temp'eratures of'rywell 1:pipe'gauge.

3.4'he minimum&F3 permit ted by technical speci f.ications is 1.10 psid.,

The corresponding water,leg leng th. is approximately V.S fe.et mea'sui ed from the'quencher centerline elevation.

4.0

'A description of the design condit ions for each load case used.

f.or -design 4'..1

-Figure 7.3:.o,f the PUAR sh,ows a plan view of,thy S/RV di'scharge in the torus.

This information's well as other information r'egarding 'geometry is avai'lab'le from TVA'rawing series 47W401.

Ai-s*,

'ee le,l above.

4'.2 The following. parameters wer e extracted from selected RVRlZ; RVFOR input.

Case Al..l '(NOC)

SRVDL E,1'P oF

'rPL oF 115.

'.5 115 VS M~Slb/.sec AP~sid LW.ft 303 l. 2>>

2.,13 308 1.2*

V.13 Cess.C3.3 (ISA. with Steam in D~W Second.Actnntion)

SRYDL E

L.

TP toF'PL oF,MS 1b/sec dP s I d 350 90.

308

l. 0+

350 90

.308

1. 0+

LW i'It

40. T2 31.,56

• ,These 'values were.used in RVRIZ, a value of zero wa's used for RVFORe 5..'0 A'escription of the design loads for each iioad ca'se.

1 Normal.i'zed. pressure s ignature We in terpre t the term "normal ized pre.ssure s ignature" to mean th6 s~ar ia'tion of the shell

pressu~i es wi th times Thieref'ore, i t-is the sake

.the var iable g( t} dtbf ined

.in', S'ec t ion 5.,4.2.8,of

'NLi 16-4 PUAR.00

5.2 The largest magn.i'tude POP and PUP values generated by QBUBS02 were -.appl'led for.each SRVDL in the torus.

Per 'the 'LDR, first actuation pressures were conservatively

assumed, to.be possible for second actuation (reflood) conditions.

The single valve values are as follows:

Pressure

( si

)

Event NOC or DBA SBA or,IBA POP

16. 2 1'9. 7 11.9

16. 6 5 e 3 The spatial a t.t en ua t i on functions used we r e as defined

'i n.the LDR and ge n e r a t ed by QBUBSO 2.

The only de v i a t i on 'from the LDR i n t h i s r ega r d wa s the use of SRSS 'for combining the ef fects of mul tiple valve actuations.

The SRSS issu'e fs addressed by the respons'e:to:BNL Item 14.

5.4 The f,requency.range.used, for design, as provi'ded 'by Section.5..5.'2 of the PUAR, is as follows:

Fre uenc (Hz)

.Even.t NOC or DBA SBA or IBA Min imum

4. 16

5. 58 Max imum 10.'29

14. 69 BNL 1 6 "5 PUAR. 00

8.0 6.0 P

4;0 2.0.,

1 VI.

-2.0 $

-0.0 $

0I0L I

(

I V

'R,A.A. A

/

X./: V V

t$.V0.

I '

.0.2

~

'.0.4 0;6, 0;.8,'1.0 1".2-1'.4'.1.6 1.8

.2.0 P-6.0

'5.0 4.0 3,0 $

2.0 f

/l

/1 VIV

~

n 2 ~ M 2.0

-3.0

-4;0 5.00;0 0ec 0.4 0.6 0.8:1.0 TIHE SECONDS V

.1..4

'1.6 1.8I 2.0 X100 t=st"Uxi BN" -ib-,i

'YPICAL PRESSURE

'TRACES'I-UR'BFN, lORuS SHELL'

TRBLE'NL-.iS'-i

'MAXIMUM AND MINIMUM PRESSURES (PS:I )

SINQL'FVALVE. ACTUATION. TESTS S2 S3 S4 S5 2

4.

7 8

C7 S

1'0 12 13 5.648

-4'. 950, 5,.'870.

-5'.203 5.393

-4.825'.593

'3.950,

3. 105

. -2'.354:

5. 195'

-3'.554'1'.757,

. -1'.1'1'3

'.:2.'539 1'.716 2.226 1:.648

'6..158.

'4'.755 4 887

-3.655 6.417'5.5S6 8.743.

-5.898:

6..455'5.386,:

5.285

-4:.47.7';;07,7,'4'097

5.7,46.

-3.9SO; 1'.'670'1.482 2'.737"

1'.'988'.369

- 1'.859 6'.'996

-5.397,'.615

-3'. 907:

6.71'4

-5.538 7 233

-5.827

.6'.S7S

-5.'45.1, 6'.'004

-4.650'.'297,

-4:.'258 5,.'501

.3.S69:

1'.878

-,1.341 2'. 821;

-1.961'2.580.

-2'.036 8.,757

-5.249 5'-517

-3'32 8'.6S5

-5.'545 7:. 1'12

-5.849 6;550

-5.429'.

709

-4..'509'..620

"-3'.929',.814

-4.. 1'12 1'.810

-1.354 2.751

-2.008 3.057 1'.838

'6,S1S.

-,5.333 5,496

.-3.858

'5.971-

-5.293 6.232

-5.52S 5.800

-4.949 5.183

-4. 195 4:. 151,

-3.,467, 5.. 1.13

-3.915 1.650

. 1'.341

~

2'512

.1.900 2'.764.

= 1.634 7.046

.5' 383 5.461,

-3.970

TRBLE -BNL-16-2 MAXIMUM ANL) MINIMUM F'RP'SSURES.

(PS I )

MUII TIPLE VALVE ACTUATION TESTS MI2 I%3 M4 7

8 Cl S

10 12 13 14 I).685

-5.997 6 211

'-6.310 ii.787

-6. 150 CJ.204

-.6 016 6.471

'; -.4.922

'5.828

-4.623 1.348

-1.831 2 ~ 151

.2~ 1S2 11.98

-7.482 7'.616

-6 108 5.314.

-5,.300 6.029

.6.062 7.283

".438 8.857

-6.434 10.01

-6.485 6.310

-4.613 6;7S5

-5. 174 1.764 1;777 2.669

.2.185 1 1.00

-7.639 8.328

-5.876 6.140

.5.356 S.,i123

-;6.,333 9.,135

--6.,771 9.,956

-6.,688 1 0.,79

-6.,760 8.,523

-5.,659 8.,224

-5.,331 2.,642

-'1.,837 3;,288

-2.,383 1 0-.,36,

-7.,162 7.;S19

-6.,524 6.,343

-'5 '097 7.005

-6.35S

8.'617,

-6.786 10.00

.6.66 1'0.66

-6.491.

6.216-

-5.686 8.007

-5.590 2.903

. 1.50S 2.914

-2.240 1 1.16

-7,.1S6

'7.264

-5.S53 5.076

-4.57S

TRBL'E:BNL-i:6-3 INITIAL GAS VOLUME (FT~ )

'A1. 1 'NORMAL OPERATING CONDITION BLOWDOWN 73.80 75.37 68.'16 LONG LINES 58.31

'53.33

'55 07 52.'1'4 54;82

'51.'00 52'.'89 55;60 SHORT LINES 54.'82 54.93

ITEM I7:

Elaborate o'n the, methodology used'o a'ccount for fluid-s tr ucture inteicact ion ef fects during CO hand c'hugging, drag loads in ElFN, ivhich is mentioned in1Se1ct)on 4.4,.9 of the P,UAR.

RESPONSE

,For an elaboration on the methodology used to a'ccount for fluid-str uctur e interact ion ef fects during CO and chugging drag loads, see PUAR Section 4.4.9 tpage 4>>24R1),

wi th, revi sion 1 change, and PUAR Sect i,on D.1.2.1;3, where tlie methodology is discussed in detail.,

1BNL 17-1 PUAR. 00

ITEM.18"..

What i's the verti'cal loca't'ion: of the suppression pool temperature sensors in rel'ati'on to the S/RV T'/Quencher cen'ter l,inc?

RESPONSE

The sensors are loca'ted approximately 20 i'n'ches above

the, T-Quencher'ent'erli'ne and:at mid-b'ay.

(See PUAR Figure 10-7.)

BNL 18-1 PUAR. 00

ITEM 19

~

'Were, there, any'xceptions',to the AC -for the hydr'odynamic'i loads..appl.ied

'for analysis of the Torus Attached Piping?

I'f so,. elaborate;.

RESPONSE

Other than the general.

interpr'etations elaboi ated In.Section.

4.2 of'he E)FN Pl'JAR,, there,were;no speci'f I c exceptions to

the NUREG.06i61

A'C for the,'.,hydrodynamic loads'pplied'or anal'ysi's of the,torus:attached piping.,

BNIi 19-.1.

PUAR. 00

ITEM 20:

In the calculation of va'r ious drag loads for BFN, the computer codes

LOCAFOR, CONDFOR,

TQFORBF, and TQFOR03 were used.

Do the algorithms o'f these codes follow approved AC procedures?

State any exceptions and justify them.

RESPONSE

The GE computer codes I'OCAFOR,

CONDFOR, TQFORBP, and TQPOR03 were used in the calculation of various drag loads for BFN.

These codes were "put up on Control Data Corporation computers around the country for access by the different AEs performing Mark I plant unique long-tecum program evaluations.

These codes were develo'p'ed, documen'ted, and verIfied by consultants u'nder contra'ct with GE, not by the'Es performing the Mark I anlayses.

The codes are proprietary t'o GE and

- Were only provided as "black boxes" with instructions on their use (including descri'pti on of requi red input data)

provid'ed in the form o'f Application Guides.

The AEs (includ'ing TYA) therefore.

do not have the direct access to t'e specific algor'ithms

'o'f t'hese codes whic'h. would be necessary to a'newer youi questi'on definitively.

It is TYA's understand'ing>

however, that the codes

LOCAFOR, CONDFOR, and TEE@FOR, used for evalu'ation of pool sfiiell, CO and chugging, and S/RV drag loads, r'espectively, follow approved NUREG 0661 AC procedures.

That me'ans that TVA has defined only S/RV drag loads with codes not thought to specifically follow all NRC-approved AC proce'dures.

The only significant differences between the appr'oved code TEE/FOR (not used by TVA) and the TQFORBF and TQPOR03

codes,

,to TVA's knowledged are

'as follows:

T~FORRF This code is different in that two bubble pressure

factors, BFAC(l) and BFAC(2), were incorporated to be used as mul tipliers of the n'egative ahd posi tive bubbl'e pr essures; respect ively.

These were 4mpl rically d'eveloped factors used to obtain more realistic comparisons of code predictions to Monticello test results (see Appendix B of GE Appl ication Guide 5, Revision 3 - a later revision of PUAR Reference 63).

~T FOR03 This code is dlffe'rent In the bubble dynaMics portion o! the

.code which uses QBUBS03 instead of QBUBS02.

The resul t i s that far moi e real istic, load predictions are obtained from this code due BNL 20-1 PUAR. 00

~

~

I

~ %

to the at tenuat,ion tn bublble energy as i t rises tIo the. surface of the torus pool.

Particul'arly fort structures, located high.in 1'.he.pool, this code predicts'rag loads, that lark sighific'antly attenuated in amplitude with 'time.

Whil,e bo'th of these

-codes are felt to be more real istic than the; extremely conservatiive TEEQFOR code, it should be noted that the least conservat'ive

code, TQFOR03, was, only'sed fear the.down'comer.S/RV drag load prediciti*ns.

That was because the, downcomers az'e,-located, very. high in ithe pool where,'realisticall'y, the S/RV bubbles have attenuated substantially'rom

'their exit str'ength.

Loads pr ediicted for,this structure's i,ng even the 'TQFORBP codle (which is slightly 1 ess conser,-

vative'han.'PEEQPOR) w'ere found to be un'rehlistically high),

For'l"1subrrierged s'tructures otheir thorn. the downcomers, the still very conservative Tg!PORBF code was used to,obtain peak

'S/RV drag force amiIil i tudes.

Thi's ciode iwas used for the

.other s'ubmer'ged s'tructures because.

i t is cheeIper.to run thein TQFOR03 and the degree of overconservat'1'sm in TQFORBF, versus TQFOR03: is not 'too significant for structures in lower'levations of'he pool'.-

Other than the downcomers, most BFN

,submerged structures ar'e in the lower pool elevat'ions.

The use of TQFORBF and TQFOR03 was

',believed to be well justi'fled. by 'good engineer in'g...judgment and especially by the fact that TVA plannied S/RV tests which were. expected to

,support tliat.judgment'.

Additionally,'he S/RV,drag, (exc'ept forth'e downcomers) were.co'nserva'ti'vely

'defir'>ed assuming worst-case peak

'.load amplitudes, app11ied as steadier-state harmonics at worst.-case, frequencies..

As expected,,

the S/RV tests provided conclus'i,ve evidence of the adequacy~

of the, analytical'appraoch for S/RV fluid dirag loads.

~ Ajipendix C of the.PUAR desctibes'lhe correlation of analytical and test:. dat'a.

For examipl<!, the analytical approach for the downcorners (using, TQFOR03 loads)

.was showvi to overprIedict stre,sse's by a fac'toi of four relative to single anil niiilti~rle S/R11 test results.

i

~ ~'~'i

-'ir PJ C

BNL 20-2 PUAR.00,,

ITEM 21:

Are there any di fferences between Browns Ferry Un i ts 1,

2, and 3 which were s igni f.icant enough to warrant separa te analyses for any unit? If'o, state the di f ferences and the analyses used.

RESPONSE

Tot ns The Units 1 and 2 tori are virtually identical.

Unit 3

used lighter construction in the following areas:

Location Uni,ts 1 and 2

Untt 3

Ring gir'der inside flange l-l/2" x 12" 1" x 10" Ri'ng g'i rder web 1-1/4" x 12" 3/4" x 12" Cradle edge p'lates 1-1/4" x 12" 1" x 12" All dynamic torus anal'yses and the definition of modifi-cati'ons were 'ba'sed on the Unit 3 properties.

Modification studies showed that stif'I'ening the ring girder-cradle system always improved performance.

Hence,,

the definition of modifi'cation for Un.its 1,'nd' fr'om the Unit 3 ana.lysis results is conserv'ative; Vent S stem The vent systems for al'1'. BFN units 'are virtually identical.

Tor us Attached Pi in Ex'ternal BFN torus a't tached piping conf igura tions are d i fferen t for each BFN un it.

There fore, g'enera 1 ly a separate pipirig analysis was required for every piping system on each unit.

Internal torus attached piping Configurations are virtually idehtical i'rom unit to unit but th'ey are included in the external piping analytical models.

S/RV Dischar'e Pi in Two basic S/RV piping configurations are used in each BFN torus as shown by PUAR Figures 7-8 and 7-9.

The arrangement of all 13 lines in eacli BFN torus is shown in plan view by PUAR Figure 7-3.

BFN S/RV drywell piping configurations v'ary from line to line and and in some cases unit to unit.

Therefore, various analytical models were used for the S/RV piping in the drywells.-

Nonsafet Related Interrial Structures The nonsa fety related Iriternal stiuctures are virtually identical for all BFN units.

BNL 21-1 PUAR. 0 0

ITEM 22

'Th i'; i',an adidi t,"ion'al'I tern to. re'spond:

t'o a;verb'al,inqui ry

'rom'BNL't

.the Sep'temper'"5, 1984 meeting;,BNL asked how close'o impact with,the ma in 'ven't'ellows does, the pool swell come.

V RESPONSE::

Calculations; ba'sed on the;conse'rvatiye LDR methodology>.

predict. ihat-;the pool swel'1. profile woul'd'omeithin" 1.1 inches. of the bellows.

If the pool'well were to slightly impact the bellows, the velocity would, be, very

small, approaichilng,,zero at inc ipient Impaot

,Examinatiion of the, construction 'of'.the.'bellows leads'o the asse'ssment

.that it hiss ver', good 'local-'mpact res is tance.

i

'Any poten'tial for damage or leakage would r'e!qui're 'Impact over'a 'large, area,

which, in turn, would'equire the pool to,r ise at'" leis t, a foot,air,more ab'ove the;bcit tom of the.

.b e I'1 ows TVA, concludes that ther,e is no digni,ficant'.'saf'ety

'Co'ncern'or pool., swiell jimp'act. on the vent'ystem.bellows's

~

~ 5

~ ~

~ t r

~ 'I BNL'2'-I'UAR.,OO'

'HAX 'POOL SHELL PROFIL'E 4 +46' EL 54m -SO~E 3 o87' m

t, io I

I EL 537~-0 7+75'c

~S F'R PoiNT B oN 'vENTI YB +

5,07'S

~ YHAX POOL AT )L/R ~ 6.5<

YHAX

~Bol OOL HISSES 4Y i09'A ASOUT ii AT THE CLDSEST POINT>

FICURE BNL--ZZ=I TVPICAL, HID VENT BAY CROSS-SECTION

"r4 ~

~ Q

~

~

0 l

>>I I

.1 4

)

S 7

~,'A l

l I'

~

i>"

',f~

1 4

~4laJ ~ 1