ML20027C120

Technical Evaluation of Fermi-2 Plant-Unique Analysis Rept.

ML20027C120
Person / Time
Site:	Fermi
Issue date:	09/30/1982
From:	BROOKHAVEN NATIONAL LABORATORY
To:	NRC
Shared Package
ML20027C117	List: ML20027C119 ML20027C120
References
CON-NRC-20-82-213 BNL-04261, BNL-04261-01, BNL-4261, BNL-4261-1, NUDOCS 8210120642
Download: ML20027C120 (40)

Text

._ _ _

s 4

a.

s<

8=.

I

, TECHNICAL EVALUATION OF THE FERMI 2

. PLANT-UNIQUE ANALYSIS REPORT

't f

1 i .

Thermal Reactor Safety Division i Department of Nuclear Energy i

Brookhaven National Laboratory Upton, New York '11973

• 9 l September 1982 l

l NRC Contract No. 20-82-213 BNL-04261 ,

~ .

a \

l'8210120642 870923 1PDR ADOCK 05000341 '

'A, PDR_,

- , .. ., __--A

Abstract The objective of this report is to document the post-implementation au'dit of the FERMI 2 plant-unique analysis against the hydrodynamic load acceptance criteria presented in NUREG-0661. A brief description of the audit procedure as well as a summary of the various phases of the audit are provided. In ad-dition, an overview of the results of the audit which highlights the various issues or exceptions to the acceptance criteria identified during the audit is -

included, along with an indication of the status of.each issue. At the pre-sent time, there are still 2 outstanding issues which require supplemental in-fonnation in order for the items to be resolved.

l

=O I -i -

ACKNOWLEDGEMENTS The following individuals participated in the post-implementation audit and contributed substantially to this Report:

J. D. Ranlet, Brookhaven National Laboratory G. Maise, Brookhaven National Laboratory C. Economos, Brookhaven National Laboratory J. Lehner, Brookhaven National Laboratory -

G. Bienkowski, Princeton University i

A. A. Sonin, Massachusetts Institute of Technology The cognizant NRC technical monitor for this program was Mr. Byron Siegel of Operating Reactors Branch No. 2.

s e

f

-ii.

List of Acronyms AC Acceptance Criteria BNL Brookhaven National Laboratory C0 Condensation Oscillation DEC0 Detroit Edison Company DLF - Dynamic Lead Factor FSI Fluid Structure Interaction FSTF Full Scale Test Facility LDR Load Definition Report LOCA Loss-of-Coolant Accident LTP Long Term Program NPS Nominal Pipe Size .

- NUTECH NUTECH Engineering, Inc.

NRC Nuclear Regulatory Commission PUA Plant-Unique Analysis

- PUAAG Plant-Unique Analysis - Applications Guide PUAR Plant-Unique Analysis Report t

I QSTF Quarter Scale Test Facility SER Safety Evaluation Report SRV Safety / Relief Valve STP Short Term Program I

-i i i-

+ c -,

TABLE OF~ CONTENTS Page No. i ABSTRACT- i

, ACKNOWLEDGEMENTS ii LIST OF ACRONYMS iii

1. INTRODUCTION 1

2. OVERVIEW 0F AUDIT PROCEDURE . 2  !

3.

SUMMARY

OF CURSORY AND DETAILED REVIEWS 6

' 4. . SYNOPSIS OF THE POST-IMPLEMENTATION AUDIT 15 1 5. REFERENCES 33

')

l l .

1 .-

1 -~

l r

i

-iv.

i -

t - , ,

1. INTRODUCTION The suppression pool hydrodynamic loads associated with a postulated loss-of-coolant accident (LOCA) were first identified during large-scale tes-ting of an advanced design pressure-suppression containment (Mark III). These additional loads, which had not explicitly been included in the original Mark I contai aent design, result from the dynamic effects of drywell air and steam being rapidly forced into the suppression pool (torus). Because these hydro-dynamic loads had not been considered in the original design of the Mark I containment, a detailed reevaluation of the Mark I containment system was required.

A historical development of the bases for the original Mark I design as well as a summary of the two-part overall program (i.e., Short Term and Long Term Programs) used to resolve these issues can be found in Section 1 of Re-ference 1. Reference 2 describes the staf f's evaluation of the Short Term Program (STP) used to verify that licensed Mark I facilities could continue to operate safely while the Long Term Program (LTP) was being conducted.

The objectives of the LTP were to establish design-basis (conservative) loads that are appropriate for the anticipated life of each Mark I BWR facil-ity (40 years), and to restore the originally intended design-safety margins for each Mark I containment system. The principal thrust of the LTP has been the development of generic methods for the definition of suppression pool hy-drodynamic loadings and the associated structural assessment techniques for the M;rk I configuration. The generic aspects of the Mark I Owners Group LTP f

were completed with the submittal of the " Mark I Containment Program Load De-t finition Report" (Ref. 3) and the " Mark I Containment Program Structural Ac-ceptance Guide" (Ref. 4), as well as supporting reports on the LTP experi-mental and analytical tasks. The Mark I containment LTP Safety Evaluation 4

g , , ,-- --- -

Report (NUREG-0661) presented the NRC staff's review of the generic sup-pression pool hydrodynamic load definition and structural assessment tech-rdques proposed in the reports cited above. It was concluded that the load definition procedures utilized by the Mark I Owners Group, as modified by NRC requirements, provide conservative estimates of these loading conditions and that the structural acceptance criteria are consistent with the requirements of the applicable codes and standards.

The generic analysis techniques are intended to be used to perform a -

plant-unique analysis (PUA) for each Mark I facility to verify compliance with the acceptance criteria (AC) of Appendix A to NUREG-0661. The objective of this study is to perfom a post-implementation audit of the FERMI 2 plant-unique analysis (Reference 5) against the hydrodynamic load criteria in NUREG-0661. A brief description of the audit procedure, as well as a summary of the various phases of the audit, are included in this technical evaluation report.

In addition, the checklists utilized in both the cursory and detailed portions of the audit have also been included for completeness. A chronology of events concerning the FERMI 2 audit and a table summary of items identified during the audit as either exceptions to the AC or as areas where additional infor-mation was required are provided as part of an overview of the audit.

2. OVERVIEW 0F AUDIT PROCEDURE The procedure described here is used for a post-implementation audit of the Mark I plant-unique analysis against the pool dynamic acceptance criteria presented in NUREG-0661 ( Appendix A). The audit is a two-level procedure whereby the acceptability of the methods used by the utility in the PVA will be evaluated. The procedure as shown in Figure 1 consists of both a cursory and detailed review of the pool dynamic methodology and in addition provides for deviations from the acceptance criteria.

The detailed portion of the review concentrates on a set of key loadings and structures in keeping with the premise of perfoming a moderately detailed a

w .

READ PERTINENT SECTIONS OR VOLUMES OF THE PUAR

• PREPARE A LIST OF QUESTIONS FOR CLARIFICATION AND IF REQUIRED A LIST OF PU AC EXCEPTIONS

~

a REVIEW MODIFICATIONS TO SEE IF THEY WILL -

AFFECT THE AC LOAD SPECIFICATIONS ,

a DETERMINE THE AVAILABILITY OF REQUIRED REFERENCE DOCUMENTATION ,

I fl I kl CH v .

~

.)

REVIEW E e PUAR CK g  %%N OF PU AC (%%h%\\3 DETAILED I

S EXCEPTIONS LOAD REVIEW T

p

! U

)

i TER [. -

l FIG'JRE 1. SCHEMATIC OVERVIEW OF IWO - -

~~

LEVEL AuniT PROCEDURE x '

s

. . /

r . .

audit. Each load reviewed -during this phase of the audit will be carefully O scrutinized to verify compliance with the acceptance criteria. The loads selected correspond to the primary l'oading categories associated with a post-

~blated LOCA or SRV actuatiori as identified during the Mark I long term pro-gram, namely, pool swell vertical loads, condensation oscillation and chug-ging loads, vent system impact and submerged structure loads. In addition, a

- 1, load is selected on a plant unique basis to enable the audit to be fitted to the plant under consideration. As a consequence, dead weight loads, seismic loads and pressure and temperature loads thich were previously considered in the original containment design and documented in the plant's FSAR have not been included in this portion of the audit. In the same vein, the internal structures such as the catwalk, monorail, etc. which are not required for the l

7 safe operation of the containment during accident conditions are not consid-ered in detail. However, f ructures are checked during the cursory re-view phase of the procedure to ensure that they were analyzed for the appro-priate loading cer.ditions as specified in NUREG-0661. The cursory overall re-view;of the wetwell structures to ascertain that they have been analyzed for all applica$le LOCA and SRV loads is accomplished by means of a checklist to verify de completeness of the PUA. The checklist used in the FERMI 2 cursory review is included in the following section of the report where the summary of the postdin'plementation audit is presented.

' 2 As shown in Figure' 1, the audit procedure also possesses the capability of dealing with' deviations from the acceptance criteria. The cross-hatched ar-j rows indicate that exceptions to the criteria may arise at three different times during the audit, namely, from the reading of the PUAR, from the cursory checklist or from the detailed load review. It is anticipated, however, that

, ,, trie majority of the exceptions will be found during the reading of the PUAR.

0, -

=

y.

y ,

-p, _ _

w

The AC exceptions or deviations in the plant-unique analysis are carefully re-viewed to determine the acceptability of the method used by the utility to predict the hydrodynamic loads on the wetwell structures. ,

l

.l e

e-

?

i t

i

j. .

{

l l

1 l

3.

SUMMARY

OF CURSORY AND DETAILED REVIEWS The purpose of the cursory review portion of the post-implementation audit is to provide a quick evaluation of the overall completeness of the PUAR under consideration with regard to the NUREG-0661 acceptance criteria. The check-list utilized in the cursory review of the FERMI 2 plant-unique analysis is presented in Figure 2. Check marks are used to indicate that the structures have been analyzed for the loading conditions as specified in NUREG-0661. On the other hand, if compliance with the acceptance criteria is not indicated by the review, a cross is inserted in the appropriate box. This figure is also used to present any plant-unique information useful for an overview of the audit and methods used to satisfy the AC such as any AC approved alternate methods used in the PUAR. The notes in the right-hand margin which accomplish this task are explained after the , checklist in table fashion.

. The detailed load review portion of the post-implementation audit concen-trated on the set of key loadings listed in Figure 3. These loads have been carefully scrutinized and where possible duplicated to verify compliance with the AC. The method used to summarize the results of the detailed portion of the audit is similar to that used for the cursory review with notes being used to provide any additional information required.

In general, various exceptions to the AC or areas where additional infor-mation is required will be identified during both the cursory and detailed re-views. Since these items are not shown in the above figures, if they have been resolved, a complete listing is provided in the following section of the report.

I O

4 S310N  % -N m llX3 H3WODNM00

'EE M M0738 D Ig xdo aW W3iE F 0": ilX3 30 3A08V U E

~81M 'WHON M0738 g g f Fy@ o-ow 3 2f2m 73A37 W31VM 7VW80N 3A08V g gg S3 Nil d HOS10 g gggg 801037330 T D 3:

W (n W30V3H IN3A 'E w

$ (00) S83WOONM00 E E  % cE l- J O

D W30V3H IN3A f) E D d m

0: W F SIN 3A NIVW %) E  % T 3 en W31SAS g gg g E 180ddnS S0801 o m

m

~ll 3 H S S O H O 1 E T'fi) T 3

-n m N01103S OV - N to e e r- m m 4 4 o 1990-03H AN d d ddWdddyg  :--

w =

3 T 3 x

ke m W

w z

w E o

a. 3 2 $ o E 8 Eea o4 F W 5 8 Eax =

w >

et o w M

o $ H o Jw z$ E m m 4 F o l

O 1

w m

m 3

m c

<h<F J

mozF 2 o om4o$

- U- h b$2e $ --.

H Z

2 w a $1 -@W s og< o o w

w H a > J, w 2 m W HwFo_m nJ m

z > 3 wzmZ sJFm 4

M g Z mm m >*F14 knm -<wn F h J DD FoFJ<

l 2

O Z

W O

O m a: 24ooo< m o

o > a. o o >E FF w 'i u. o a. o2 o a 1 .

.s -

r < r -t

> MOO o o < r -t o I OmOO O

-l z > m c- Ez>:n z m -1 o C g o j us

• • c zn -4 *o C

• m r -<mzz 0m ~ u, 0

• Z r -i m Z

o E m -4 u<) zr m u)

Om m mm m r >

c r r or c F _4 yEm m -

O u)OmO rEr yrEr OmO Z z$$o E$$$ 0 h

g*9" s

5*0" E >

r o z

m 5

m

-i r U) o :o o :n - >

O c v> c -f E o o -

O m a H Z

2 c C v> m :n m m (n u>

3 o

M .

N U$UU UU$$ NUREG-0661 P

num- P * .N ~-

Nl*- V a AC SECTION

[; R Q TORUS SHELL

-4 TORUS SUPPORT E E E SYSTEM m u) v ( R MAIN VENTS d o

C E S. RR VENT HEADER O H

C SR RR DOWNCOMERS (DC) x r:1 VENT HEADER u)

, DEFLECTOR S/RV k b DISCHARGE L.NES ABOVE NORMAL m ,

WATER LEVEL z BELOW NORM. WTR. w$~i op$

k k 8 ABOVE DC EXIT oEmT m

ij-i2 k k BELOW $

DOWNCOMER EXIT m

~

$ tu -- -

NOTES l

-w - - - - . - - .

OTHER WETWELL STRUCTURES INTERIOR STRUCTURES AND PIPES p $ 0 a bb z <,

b x

w my J

e m -

m w o m w m wm a 2>

5 2 d A2 p O E W co yw tw [- Oc 3m ow m

w m2 z 'w z < 4H Z oa o Z e zy sE ou m S$ w E E WS dE $

> a m< we 30 mo ow m ms o z wm o om z H z Fw I H om z cc e cc Z 3 zu o" >o <a4 3

@S se LOADS ES S S $- s 8 s o 8 l T-OUENCHER LOADS 5 DISCHARGE LINE CLEARING 2.13.2 d ,

TORUS SHELL PRESSURES 2.13.3 E6 56 '6 JET LOADS ON SUBMERGED STRUCTURES 2.14.3 Id d I AIR BUBBLE DRAG 2.14.4 d d[ I d 6 i THRUST LOADS ON T/Q ARMS 2.13.5 EI '

S/RVDL ENVIRONMENTAL TEMPERATURES 2.13.6 d l

4 FIGURE 2 (CONTINUED)

l I

CONFORMANCE WITH THE ACCEPTANCE CRITERIA IS ALSO REDUIRED OF THE l FOLLOWING:

CRITERIA W

Ez ge m NOT 4$

ww MET MET b

a m

w em Ha F DESCRIPTION SS S% O SUPRESSION POOL TEMPERATURE / 7 I

LIMIT 2.I3.8 SUPRESSION POOL TEMPERATURE /

2 MONITORING, SYSTEM 2.13.9 V L '

DIFFERENTI AL PRESSURE CONTROL o SYSTEM FOR THOSE PLANTS USING A-l 3 DRYWELL-TO-WETWELL PRESSURE / O

DIFFERENCE AS A POOL SWELL MITIGATOR 2.16 4 SRV LOAD ASSESSMENT BY V

/ O IN-PLANT TEST 2.13.9 - I I

1 Frouae 2.(ConTinuen)  !

.. l

NOTES TO FIGURE 2 NUMBER EXPLANATION OF NOTE 1 These loads were examined in depth as part of the detailed review portion of the post-implementation audit procedure.

2 For some structures, the standard procedures described in NUREG-

~

0661 result in unrealistically conservative loads. In these situations, the alternate procedure described in Section 2.8 of Appendix A to NUREG-0661 is used. This procedure enables the Re-gion I froth loads to be based on the high-speed QSTF movies.

3 The post-chug submerged structure load on the ring beam was con-sidered as part of the detailed review portion of the post-im-plementation audit procedure.

4 A concern has arisen as to a possible non-conservatism in the sin-gle downcomer lateral load specifiction (see Item 12 in Section 4). Pending the receipt of a response from the applicant on this issue, this item will be considered an open issue.

5 The T-Quencher loads, in general, were considered as the load to be selected on a plant-unique basis for the detailed review por-tion of the audit. The SRV loads were chosen because of the plant-unique aspects of the FERMI plant T-quencher as compared with the Mark I T-quencher described in the LDR. .

-1 1 -

4

C A series of in-plant SRV tests will be performed to confirm that the computed loadings and predicted structural responses for SRV discharges are conservative.

7 The local suppression pool temperature limit was defined in NUREG-0661 as 200 'F for the generic Mark I T-quencher as de-scribed in Appendix A, .Section 2.13.8. Subsequently, NUREG-0783 provided procedures whereby the limit gould be increased if cer-tain restrictions could be met. Although conformance with the above criteria was indicated in the PUAR, the applicant utilized an unapproved local pool temperature model. Pending final ap-proval of this model or sufficient other justification for the local to bulk tempera.ture differences used in the PUAR (see item

. 8 in next section), the suppression pool temperature limit issue will be considered as an open item.

8 The normal operating drywell-to-wetwell pressure differential is zero for the FERMI 2 plant, thus no pool swell load mitigation l system is utilized.

9 Requirements and/or guidelines are specified in both NUREG-0661 and NUREG-0763 for in-plant tests of SRV discharges. BNL has re-viewed the proposed Fermi SRV in-plant test plan (Reference 6)

! with regard to the hydrodynamic criteria specified in these re-l l ports and find the test plan acceptable.

l

.g - --_ - -- - - , - -

CRITERIA b!

or

$$ MET y a: s LOADS @@ @

1. POOL SWELL A. NET TORUS VERTICAL LOADS 2.3 /

B. AVERAGE SUBMERGED PRESSURE 2.4 / -

C. TORUS AIR SPACE PRESSURE 2.4 - /

D.~CIRCUMFERENTI AL AND LONGITUDINAL TORUS SHELL PRESSURES 2.4 /

2. IMPACT A. VENT HEADER OR VENT 2.6.1 /

HEADER DEFLECTOR B.DOWNCOMER 2.6.2 /

C. MAIN VENT -

2.6.3 /

3. SUBMERGED STRUCTURE A. RING HEADER OR VENT 2.14 / /

HEADER SUPPORT COLUMN

4. CC and CHUGGING TABLES OF 2.11.1 /

RIGID WALL TORUS SHELL PRESSURES 2.12.1 /

5. LOAD SELECTED ON A PLANT-UNIQUE BASIS 2.13 / 2

. FIGURE 3. PUAR D'ETAILED LOAD REVIEW

,o .

NOTES TO FIGURE 3 NUMBER EXPLANATION OF NOTE 1 The post-chug submerged structure load on the ring beam was considered as part of the detailed review portion of the post-implementation audit proce-dure. ,

, 2 The T-Quencher loads, in general, were considered as the load to be selected cn a plant-unique ba-sis for the detailed review portion of the audit.

The SRV loads were chosen because of the plant-unique aspects of the Fermi plant T-quencher as compared with the Mark. I T-quencher described in the LDR. The detailed review of the SRV loads cannot be completed until the open issue associ-ated with the local-to-bulk temperature differ-ences used in the PUAR is resolved (see item 8 in the following section). Upon resolution of this issue the SRV detailed review udll be considered complete.

4

4. SYN 0PSIS OF THE POST-IMPLEMENTATION AUDIT The post-implementation audit as described in this report was used to evaluate the hydrodynamic loading methodologies used for the majar modifica-tions portion of the FERMI 2 plant-unique analysis. During the various phases of the audit, numerous issues were identified as either exceptions to the ac-ceptance criteria or as areas where additional information was required to continue with the review. These issues are listed in Table 1 along with an indication of the type and status of each item. As can. be seen from this table, a significant number of exceptions to the AC were found during the au-dit. In order to resolve these issues, a plant-unique review of additional references and information received from the utility during the audit was per-fo rmed. A chronology of events concerning the FERMI 2 audit is presented in Table 2.

The outstanding issues as noted in the table still require some supple-mental justification in order for the item to be resolved. For completeness, a brief description of each issue identified during the audit is given below.

The numbering system is consistent with the table.

Item 1 - The Acceptance Criteria 2.14.2 section 2b in NUREG-0661 states that l

drag forces on structures with sharp corners (e.g. rectangles and "I" beams) must be computed by considering forces on an equivalent cylinder of diameter Deq = /2'l max, where L max is the maximum transverse dimension. The intent of this criterion is to provide a conservative bound (based on very limited data) that includes non-potential flow effects such as vortex shedding on both the acceleration drag due to hydrodynamic mass and the " standard" drag propor-tional to velocity squared. Since the dominant loads for the Ring Beam and Quencher Beam (the two non-cylindrical structures) are acceleration loads, the

TABLE 1. ISSUES IDENTIFIED DURING  :

POST-IMPLEMENTATION AUDIT-TYPE OF ISSUE .

STATUS OF ISSUE EXCEPTION REQUESTS FOR T0 ADDITIONAL

.' ITEM DESCRIPTION NUREG-0661 AC INFORMATION RESOLVED OPEN 1 PUBLISH 5D ACCELERATION DRAG VOLUMES USED TO DE-TERMINE DRAG ON SHARP CORNERED STRUCTURE. X X 2 P.r.NDOM PHASING OF LOAD- .

k' ING HARMONICS FOR C0 AND CHUGGING. X X 3 DOWNCOMER DYNAMIC LOAD METHODOLOGY FOR UNTIED '

DOWNCOMERS. X X 4 PLANT UNIQUE MULTIPLE DOWNCOMER CHUGGING LAT-ERAL LOAD CORRELATION. X X 5 WATER JET AND AIR BUBBLE DRAG' LOAD METHODOLOGIES FOR FERMI I-0VENCHER DE-SIGN. X X 9

TABLE 1 (CONTINUED) ,-

TYPE OF ISSUE STATUS OF ISSUE EXCEPTION REQUESTS FOR TO ADDITIONAL ITEM DESCRIPTION NUREG-0661 AC INFORMATI0il RESOLVED OPEN 6 USE OF MAXIMUM SOURCE STRENGTH FOR POST-CHUG

^

SUBMERGED STRUCTURE LOADS. X X 7 FSI METHODOLOGY USED FOR C0 AND CHUGGING SUBMERGED STRUCTURE LOADS. X X

.L

? 8 T-QUENCHER LOCAL TO BULK POOL TEMPERATURE DIFFERENCE. X X

'9 SUPPRESSION POOL TEM-PERATURE MONITORING SYSTEM. X X 10 TORUS SHELL PRESSURES PRESENTED IN TABLE 2-2.2-3 X X

]

TABLE l' (CONTINUED) .

TYPE OF~ ISSUE STATUS OF ISSUE- .

EXCEPTION REQUESTS FOR ,

TO ADDITIONAL ITEM DESCRIPTION NUREG-0661 AC INFORMATION RESOLVED OPEN 11 PRESSURE HISTORY PRE-

'- SENTED'IN FIGURE 2-2.2-8. X X 12 SINGLE DOWNCOMER CHUG-GING LATERAL LOAD CONCERN X X 9a i

6

a

• TABLE 2

, CHRONOLOGY OF EVENTS CONCERNING THE FERMI 2 AUDIT

l. March 11,1982 NRC/DEC0/BNL/NUTECH meeting to discuss the FERMI 2 PUAR. An overview of- the PUAR con-tents was presented.

2. May 3,1982 FERMI 2 plant unique analysis report re- i ceived. .

3. June 2,1982 Letter-(Ref. 7) documenting audit procedure sent to NRC. FERMI 2 audit begun.

4. June 16,1982 Letter identifying 8 items as exceptions to the AC or as areas where additional informa-tion is needed was transmitted to NRC.

I

5. June 24, 1982 Conference call between NRC/BNL/DEC0/NUTECH to discuss 8 items described in June 16, 1982 letter. Letter (Ref. 8) with revisions made for clarity and with an additional item-included was resubmitted to NRC on-June 25, 1982.

6. July 15,1982 Draft response (Ref. 9) to 9 items described in June 25, 1982 letter received from DECO.

7. July 22,1982 Preliminary results of BNL review of DECO 4

responses to 9 items phoned into NRC.

8. July 28,1982 Letter (Ref.10) describing 2 additonal items requiring supplemental information .

• sent to NRC.

< 5 h c -,n- --e - .- ,---e., --> -.ng, ..m --en,. . . _ , , -- - - - - , , - , , >-.,,,,,..,ew. .n.- - - . - - , , ,- a---,- .,, , . - - , - ,, --- - - , - - - - - --

TABLE 2 (Continued)

9. August 4, 1982 Conference call between NRC/BNL/DEC0/NUTECH to discuss BNL review of the Reference 9 responses by DECO. Items sent on July 28, 1982 as well as possible additional items were also discussed. ,

10. August 16, 1982 Conference call between NRC/BNL/GE' to dis-cuss r. concerns on itens 5 and 8 of Ref-erence b. A meeting to discuss the local ,

pool temperature model was tentatively set for September 9,1982.

11. August 19, 1982 Attachment to letter (Ref. 6) describing proposed SRV in-plant test plan received from DECO. Draft copy of Fermi technical evaluation report documenting the results of the audit to date was transmitted to NRC.

l

13. August 23, 1982 Conference call between NRC/BNL/DEC0/NUTECH l to discuss the multiple downcomer chugging l

t lateral load (item 4) and the submerged structure drag concern (items 1 and 5).

l

14. August 26, 1982 Draft response (Ref.11) to 2 items de-scribed in July 28,1982 (Ref.10) letter received from DECO.

i '

l

TABLE 2 (Continued)

15. August 27, 1982 Draft response (Ref.12) received from DEC0 providing revised responses to items 1, 4 and 5.

16. September 9,1982 NRC/BNL/GE/ Mark I-Owners Group neeting to

. discuss the local pool temperature model utilized in the Fermi PUAR.

l

17. September 13, 1982 Conference between NRC/BNL/DEC0/NUTECH/GE to discuss single downcomer chugging lat-eral load concern.

1 o

e

' 9

.c + . +- ,.. _

__.-g,w.,. . , , . , . - _ - . . - . . _ . - ., --,

~

issue concerns only the hydrodynamic mass or acceleration volume'and not the drag coefficient in the Fermi Unit 2 plant-specific case.

The PUAR states that " published" acceleration drag volumes listed in Table  ;

1-4.1-1 are used for sharp edged structures.rather than the equivalent cylin- L

- der specified in the acceptance criteria. The detailed response to item 1, in fact, explains that modelling of the actual structures is necessary, and in particular, forces on the web' of. the ring beam are obtained by modelling the .

beam by a circumscribed rectangle. In order to evaluate the implications of '

this modelling, detailed calculations were perfonned on the ring-beam web for-ces for the post-chug loading condition.

A direct application of the PUAR methodology but without inclusion of DLF's to account for the structural dynamics leads to a pressure differential on the web of 4.3 psi. A computation of, the force using the hydrodynamic mass of the equivalent cylinder of the Acceptance Criteria but the real volume for f

the " effective" buoyancy force results in a differential pressure of 5.6 psia, a 30% higher load.

The PUAR accounts for the dynamics of the structure, however, by using an equivalent static load and DLF's based on,a single degree of freedom model i with 2% damping and a natural frequency of 48.5 Hz. This procedure overem-l phasizes the effect of the' 40-50 Hz content of the source term to the extent 1

that this frequency range contributes 87% to the final effective peak pressure differential of 24.1 psi. The direct use of the Figure 2-2.4-3 as an alterna-

. tive representation of the DLF's of the actual multi-degree of freedom ring-beam system, produces a net equivalent static load of only 14.4 psi for the

PUAR acceleration volume and 18.8 psi for the acceleration volume based on the Acceptance Criteria. Since the high-frequency portion of the Post-Chug load w's , ,-e+. - - - , . . , ,,r, . ~ . , , , , , _ , , , , , , , , ,,. , , , , _ , , _ ,

comes largely from the sharp " spikes" within post-chug pressure time' histor-1es, there is substantial conservatism in applying those loads in the fre-quency domain rather than a= an initial value problem for each chug. Indeed,

! the maximum DLF for an initial value problem that one might expect for the

" spike" portion of the chug is only about 2, instead of 25 at resonance of a 4

2% damped single degree of system subject to harmonic loading, or 9 deduced 4

. from Figure 2-2.4-3 of the PUAR. A similar comparison for the quencher sup-

• port beam reveals that substantial conservatism exists in the PUAR methodology for the calculation of equivalent static loads on that submerged structure as

! well.

On the basis of these comparisons we conclude that while the direct use of  ;

f " published" acceleration volumes for sharp edge structures may not in general lead to conservative loads, the PUAR methodology for the application of these

]

loads to the relevant structures, has sufficient conservatism to bound any.hy- ,

' drodynamically produced stresses that could arise in these structures.

Item 2 - The DBA condensation oscillation and the post-chug load definitions 1

j. on the torus 'shell and on submerged structures, accepted in the NUREG-0661, were based on data from a series of blowdowns in the FSTF facility (NEDE- : 24539), subject _ to additional confirmatory tests reported in the General l' Electric Letter Report M1-LR-81-01 of April 1981.

The condensation oscillation load definition as described in NED0-21888 is based on taking tte at' solute sum of I Hertz components of a spectrum from 0 to '

50 Hz. Three. alternative spectra are to be calculated with the one producing I maximum response used for.1oad-definition. The procedure was found acceptable i in the supplement to the SER (NUREG-0661), because the demonstrated high de-gree of conservatism associated with the direct sunmation of the Fourier com-ponents of -the spectrum was. sufficient to compensate for any uncertainties 5- i

.,*--,.r-, ,.-.e .m..v,. . , . . . - -~. , , . . _ & . . - , . . - ~ , . . _ , . . , , , , .,......---.,-.-_--..m.. - - - , . , - , . . - _ ~ .,

concomitant with the data available. The post-chug load definition is based on bounding FSTF chugging data but otherwise follows similar procedures to those used in the C0 load definition.

The Fermi PUAR uses a factor of .65 to multiply the C0 and post-chug loads computed on the basis of the absolute sum of the harmonic components. The justification is based on comparisons of measured and predicted stresses in the FSTF facility using different phasing models (NEDE-24840). The factor .65 is chosen to give 84% non-exceedance probability with a confidence level of 90%. The PUAR does use, however, an additional alternate 4 for the C0 load-ing, based on test M12 from the supplementary tests reported in the letter re-port M1-LR-81-01. The information supplied by Detroit Edison, in response to the request for additional information (Ref. 8) provides additional justifica-tion to show that the conputed loads (using the .65 factor and alternates 1 through 3) bound the neasured stresses at critical points in the FSTF facility by 11 to 69%. The use of alternate 4 on the Fermi plant provides an addition-al conservatism of 11 to 27%.

We have examined this information and conclude that the use of the "phas-ing factor" of .65 coupled with the inclusion of alternete 4 for C0 loading, provides a sufficiently conservative representation of the C0 and post-chug loads to account for possible uncertainties associated with the data base.

Item 3 - The downcomer dynamic load methodology, which va .w epted in the supplement to NUREG-0661, was derived for tied downcomers from the supple-mental FSTF tests (Reference 13). Since the Fermi 2 demicamer pairs are stif-fened at each intersection by a crotch plate and by outer stiffener plates as shown in PUAR Figure 3-2.1-12 and are not tied, additional justification for the use of the methodology was requested (see item 3 of Reference 8). The response to our request for additional information (Reference 9) stated 4

..e ii _ _ _ _ _ _ . _ _ _ _ - - _ _ _ _ _ . _ _ _ _ . _ _ _ _ _ _ _ .__.__.m_. _ _ _ _ _ _ _ _ _ _ _ _ _ __ _ __ _ _J

that a frequency analysis of the Fermi 2 downconers had shown that the pre-dominant fundamental mode of vibration is the sway node, i.e. , both downcomers in a pair simultaneously deflecting in the same direction. As a result, the Fermi 2 downcomers will respond in a manner similar to downcomers which are tied by lateral bracing as was the case in the FSTF tied downcomer pairs.

Based on the additional information provided by the applicant, the use of the downcomer dynamic load methodology is found to be acceptable for the Fermi .

plant-unique analysis. -

Item 4 - The acceptance criteria specified that for multiple downcomer chug-ging the force per downcomer shall be based on an exceedance probability of 10-4 per LOCA. However, a correlation between load magnitude and probabil-ity level derived from a statistical analysis of FSTF data was utilized in the FERMI PUA. During the review of the specification given in the Fermi 2 PUAR, it was found that the values for the load per downcomer listed in T2ble 3-2.2-15 are somewhat less than would be obtained by applying the 10-4 probability of exceedance criteria specified in NUREG-0661. The difference between the Fermi PUAR values for load per downcomer, and the NUREG-0661 specification increases as the number of downcomers in a group decreases, i.e., for 80 downcomers in a group, the PUAR values are only 5% less, for 5 downcomers 20%, and the worst difference is 26% for 2 downcomers. Based on the review of the available information, it was concluded that the equation on which the PUAR values for different downcomer groups are based is not correct.

l Therefore, in order to resolve this issue additional information was requested_

! to justify the currently specified load levels in the PUAR.

In response to these concerns, the applicant delved further into the FSTF chugging data report (NEDE-24539-P) on which the lateral load specifications 4

for Mark I's are based. As stated in the applicant's written response (Ref.12), extrapolation of the FSTF data most applicable to Fermi 2 condi-tions showed that a conservative estimate of the Fermi 2 chug duration is 900 seconds during which 182 synchronized pool chugs can be expected to occur.

(FSTF data showed that only about 33% of all chugs are synchronized pool chugs.) The applicant then concludes the probability that the force per down-comer in a pool chug can be exceeded once per LOCA for Fermi 2 is 1/182 or 5.5 x 10-3, While the probability above is still higher than 10-4 and, in our opin-ion not correctly obtained, we feel that the 182 synchronized chug estimate is correct, and that the lateral loads in Table 3-2.2-15 of the PUAR can be jus-tified based on the FSTF results. Besides showing that only 33% of all chugs are synchronized, FSTF also showed that only 90% of the vents participated, on the average, in a synchronized pool chug and that the time between the first vent and the last vent chugging was never less than 78 milliseconds. There-fore, the term synchronized is a relative one, and assuming all vents in a group to be exactly in phase, when applying the multivent chugging load speci-fication, is a very conservative assumption. Since a triangular pulse of 5 millisecond duration would be a reasonable approximation of the lateral chug-ging load each downcomer experiences, even a one millisecond shift in chugging times between two downcomers would significantly reduce the peak load obtained from a combination of two loadings.

To obtain a quantitative estimate of the conservatism inherent in assum-ing all vents in a group chug exactly in phase, one can consider further the case of a group consisting of 2 downconers. As stated previously, it is for a group of 2 downcomers that the values in the Fermi 2 PUAR are most below the values which would be obtained by following the NUREG-0661 specification for

O the 10-4 probability l evel . The load per downcomer when two downcomers are in a group, is given in Table 3-2.2-17 of the Fermi 2 PUAR as 11.16 kips. For FSTF, the corresponding load would be 11.16/4.276 or 2.61 kips. Extrapolating from Figure 3.9-3 of NUREG-0661 one can find that for 7 downcomers this load corresponds to a probability level of exceeding once per LOCA of roughly 1.2 x 10-2 However, Figure 3.9-3 was obtained using 313 synchronized pool chugs.-

With a simple mathematical manipulation the probability of exceeding a load per downcomer of 2.61 kips corresponding to 182 synchronized pool chugs can be found to be 7.00 x 10-3 i f it is 1.2 x 10-2 for 313 chugs.

FSTF data showed a 907, participation of vents in a synchronized pool chug.

So if one of the vents in the group of 2 chugs, the probability of the second chugging at all is 0.9. FSTF data further showed a time " window" of at least 78 milliseconds during a pool chug. Assuming the prooability distribution of a vent chugging tr be uniform throughout this window, the probability of the second vent chugging at exactly the same time (i.e., during the same millisec-ond interval) is 0.9 x 1/78 = 1.15 x 10-2 This is the probability that the two downcomers chug synchronously - which was asstned as certain to get to the 7.00 x 10-3 probability of exceeding the 2.61 H ps once per LOCA. So the combined probability of synchronous chugging of the two vents and exceeding the 2.61 kips per vent once per LOCA is the product of the two probabilities, i.e., (7.00 x 10-3) x (1.15 x 10-2) or 8.1 x 10-5 which certainly is j

comparable to the 10-4 level of NUREG-0661.

While the above attempt to quantify the conservatism of the "all vents in phase" assumption is a rough one, it does use a conservative time window and in our opinion adequately justifies retaining the chugging lateral load per downcomer values given in the Fermi 2 PUAR.

Item 5 - The applicant uses a T-quencher configuration which differs in sev-eral substantive ways from the Mark I version. Some of the differences in-clude the total hole area (Fermi has 8% less), arm diameter (Fermi uses 20" NPS vs 12" for Mark I), and hole pattern (compare Figure 1-4.2-2 of the PUAR with Figure 1-2 of NEDE-21878-P). These differences could imply that the LDR specification for SRV loads on the torus shell and submerged structures (as amended by the staff's acceptance criteria) may not be applicable for Fermi. ,

The applicant proposes to perfrom in-plant SRV tests to eliminate this un-certainty ia regard to SRV loads on the torus shell. This is consistent with the staff's acceptance criteria (2.13.1 and 2.13.9) and is therefore accept-able. Information obtained from these tests (" bubble pressure" measurements) will also serve to eliminate any uncertainty relative to bubble induced sub-merged structure drag loads (Section 1-4.2.4b of the PUAR). For the water jet loads, however, such confirmation cannnt be obtained from the proposed test program. The applicant takes the position that use of the LDR methodology, with appropriate modifications to account for the Fermi design, is applicable.

However, the modifications were not described in the PUAR (Section 1-4.2.4a).

In response to our request made during the conference call held on August 23, 1982, the applicant has provided a more detailed description of how these mod-ifications were implemented (Reference 12). We have reviewed this information and conclude that they represent a correct application of the LDR methodology.

The procedures proposed by the applicant for defining SRV Water det Loads on submerged structures is therefore considered acceptable. .

Item 6 - The post-chug submerged structure loads, as specified in the accept-ance criteria, are to be canputed on the basis of the two nearest downcomers chugging at the maximum source strength unth phasing between the downcomers that maximizes the local acceleration. On PUAR page 2-2.39, it is stated that the loads were developed using the average source strength. In response to our request for additional information (Reference 8) the above PUAR state-ment was identified as being incorrect since the maximum source strengths were utilized in the Fermi 2 analyses in conformance with the acceptance criteria.

Correction of the PUAR page 2-2.39 will be included in the next PUAR revision.

Item 7 - In response to the request for a detailed discussion of the method used to account for FSI effects on condensation oscillation and chugging sub-merged structure loads, the applicant submitted a techn.ical note by Continuum Dynamics, Inc. entitled " Mark I Methodology for FSI Induced Submerged Struc-ture Fluid Acceleration Drag Loads" (Ref.14). The methodology described in this note is used to compute acceleration fields across a submerged structure anywhere in the torus resulting from FSI, based on knowing the torus boundary acceleration. The method is presented as an alternative to the NRC Acceptance Criteria suggestion of adding the boundary accelerations directly to the local fluid acceleration to account for FSI effects since the latter is deemed too conservative.

I The review of the method outlined in Reference 14 has shown it to be rea-sonable and acceptable. The equations derived for fluid accelerations and l

pressure fields are plausible approximations for the conditions prevai;ing in the suppression pool. Assumed boundary conditions including the driving one at the torus wall are suitable. Overall trends as well as the acceleration fields depicted in the selected results appear reasonable.

The accelerations calculated by the method in Reference 14 are only due to

. FSI and must be added to any local accelerations due to other causes to oitain complete acceleration fields for conputing drag loads.

I

Item 8 - The applicant proposes to use an analytical model to estimate local pool temperatures during SRV transients.. This is not in confornance with the AC as presented in 2.13.8 of NUREG-0661. The intent of this AC was for the ap-

,nlicant to develop a bulk temperature history during these transients using conventional (and staff approved)- methods. The local pool temperature would then be derived using an experimentally determined (from in-plant tests) lo-cal-to-bulk temperature difference. If this were not so, the entire section 2.13.8.2 would be meaningless.

Notwithstanding the above discussion, we need, nevertheless, to examine the viability of the proposed method to serve as an alternative. The appli-cant has provided a description of the model in Reference 15 and during the neeting of September 9,1982. Based on this information, we have come to the following preliminary conclusion.

(a) The " momentum model" component of the analysis will most likely . a an acceptable way to estimate bulk pool response to RHR operation.

(b) The " energy model" component of the analysis is not acceptable at this time because insufficient verification of its adequacy has been provided. The applicant's claim that it has been verified because it reproduces the Monticello in-plant test results is specious. When they state (in Reference 9) that "the model is calibrated to the Mon-ticello test results", we take this to mean that the (numerous) em-pirical constants incorporated by the model were adjusted until the model reproduced what was observed during the Monticello tests. This does not constitute verification.

(c) Verification of the analytical model requires additional in-plant tests. At the September 9th meeting, for example, GE indicated that they were confident that the snpirical constants derived from the l .

Monticello' tests would 'be invariant with quencher submer ence, torus"

^

l - radius, etc. They obviously have to say that since they have no  ;

s information to the contrary. If the model successfully predicts tem-

~

perature responses in a plant with differences in geome'try and RHR I characteristics, that would constitute-its verificatico, or, at

' least, its general applicability.

In .sunmary, we consider this item to be'still an open issue. In our -

i judgement, its resolution requires, at minimum, a more creative use of the Monticello data (coupled with the " momentum model") or additional in-plant tests (add extended blowdowns to the Fermi in-plant SRV test matrix). We would feel most comfortable with the latter option but judge that we can make do with the minimum option.

Item 9 - The description of the Suppression Pool Temperature Monitoring System (SPTMS) that was provided in Section 1-5.2 was judged to be inadequate for the <

purpose of determining that the SPTMS design was in accordance with the re- ,

^

l- quirements of the AC (Section 2.13.3.3). This was because the location of the temperature sensors in the radial and circumferential directions was not_ sup- ,

x .

plied. Only the statement that "... these thermocouples are uliformly dis-f tributed throughout the torus" was provided. The applicants response to NRC D

Question #9 given in Reference 9 supplies this information and pennits us td -

conclude that the Fermi SPTMS is in conformance with the AC.

~

Items 10 and 11 - The intent of these requests for additional information l

l (Reference 10) was to obtain sufficient detail on the Fermi pool swell load i . calculations in order that the detailed portion of the post-implementation .

audit could be accomplished. The information received from the applicant in c

Reference (11) provided the necessary' information to ascertain that the pool '

swell loads were calculated in conformance with the NUREG-0661 acceptance criteria.

l E J e t

- - - e p ,, ,,e--.. , .,. - -- . , -,- +- -e- - -- - , - , , w . -- e-- - - , - . ,- ,-.--n - - , e,

PP

^

'^

~

'o ,

, .s x Item 12 - The applicant 'uses a single vent lateral chugging load for each downcomer bashd on the highest lateral load observed in the FSTF tests. This basis for load magnitude appears to be 'non-consarvative when the number of

~

6teral loadings recorded in FSTF are compared with the number of individual single downceer 1[teral loadings which can be expected' during a postulated P -

LOCl in the Fernii Plant. As indicated in Table 6.2.1-1 of NEDE-24539-P (Full-Scale Test Program Final Report), the approximate total number of downcomer chugs for the eight original tests in TSTF was 1460. .Since of the eight FSTF down:cmers only two, numbers 6 and .8, were instrumented to record i ' lateral loads -(NEDE-24537-P), it is reasonable to assume that approximately

[ 1460 x 2/8 = 365 lateral chugging loads were recorded in FSTF during all of g 'the original eight FSTF tests.

The Fermi Plant on the other hand has 80 downcomers. A conservative es-t'imate of the number of pool chugs occurring with 90% of the downcomers par-w ticipating is 180, resulting in 14,400 downcomer chugs. In addition, many other chugs with only a few domcomers participating can add to this total.

While the above may be a conservative estimate, it does not seem unreasonable to expect on the order of 10,000 downcomer chugs (and therefore lateral loads) to occur during a postulated Fermi LOCA. Therefore, the highest load from the 365 observed FSTF events is very unlikely to bouad the 10,000 expected Fermi events. 0n the basis of the above numbers one could expect the maximum FSTF lateral load to be exceeded about 25 to 30 times during a postulated Fermi LOCA.

A response to this concern of the possibie non-conservatism in the single downcomer lateral load was requested from the applicant. Pending the re-ceipt of that response and its subsequent review, this item will be considered I an open issue.

f I*

f o

5. REFERENCES References cited in this report are available as follows:

Those items marked with one asterisk (*) are available in the NRC Pub!ic Document Room for inspection; they may be copied for a fee.

Material marked with two asterisks (**) is not publicly available because it contains proprietary information; however, a nonproprietary version is available in the NRC Public Document Room for inspection and may be copied for a fee.

Those reference items marked with three asterisks (***) are available for -

purchase from the NRC/GP0 Sales Program, U. S. Nuclear Regulatory Commission, Washington, D. C. 20555, and/or the National Technical Information Service, Springfield, Virginia 22161.

All other material referenced is in the open literature and is available through public technical libraries.

(1) " Safety Evaluation Report, Mark I Long Term Program, Resolution of Generic Technical Activity A-7", NUREG-0661, July 1980.***

(2) " Mark I Containment Short-Term Program Safety Evaluation Report",

NUREG-0408, December 1977.*** ,

(3) General Electric Company, " Mark I Containment Program Load Definition Re-port", General Electric Topical Report NED0-21888, Revision 2, November 1981.*

l (4) Mark I Owners Group, " Mark I Containment Program Structural Acceptance i Criteria ' Plant-Unique Analysis Applications Guide, Task Number 3.1.3",

! General Electric Topical Report NED0-24583, Revision 1. July 1979.*

(5) "Enrico Fermi Atomic Power Plant, Unit 2, Plant Unique Analysis Report",

Volumes 1-5, Detroit Edison Company, DET-04-028-1, Revision 0 (prepared by NUTECH Engineers, Inc.), April 1982.*

l (6) Attachment to Letter from H. Tauber, V. P., Detroit Edison Company to B.

l J. Youngblood, Chief, Licensing Branch No.1, Division of Licensing, NRC, i- August 18, 1982*

(7) Attachment to letter from J. D. Ranlet, Brookhaven National Laboratory, j ' to B. Siegel , NRC,

Subject:

Post-Implementation Pool Dynamic Lcad Audit l Procedure, June 2,1982.*

(8) letter from J. D. Ranlet, BNL, to B. Siegel, NRC,

Subject:

FERMI 2 Plant Unique Analysis Report, Request for Additional Information, July 25, 1982.*

, (9) Attachment to letter from H. Tauber, V. P., Detroit Edison Company to B.

! J. Youngblood, Chief, Licensing Branch No.1, Division of Licensing, NRC, August 2, 1982 (Letter EF2-58,955).*

L

o REFERENCES (Continued)

(10) Letter from J. D. Ranlet, BNL, to B. Siegel, NRC,

Subject:

FERMI 2 Plant Unique Analysis Report, Request for Additional Information, July 28, 1982.*

(11) Attachment to Letter from H. Tauber, V. P. , Detroit Edison Company to L.

L. Kintner, Licensing Branch No. .1, Division of Licensing, NRC (Letter EF2-59,268) . *

(12) Attachment to Letter from H. Tauber, V. P., Detroit Edison Company to L.

L. Kintner, Licensing Branch No.1, Division of Licensing, itRC (Letter --

EF2-59,281) . *

(13) Mark I Containment Program Letter Report: Supplemental Full-Scale Con-densation Test Results and Load Confirmation, M1-LR-81-01-P, April 1981.**

(14) A. J. Bilanin, " Mark I Methodology for FSI Induced Submerged Structere Fluid Acceleration Drag Loads," Continuum Dynamics Tech Note No. 892-15, June 1982. *

(15) " Analytical Model for T-Quencher Water det loads on Submerged Struc-tures," Task 5.14.2/9.4.1, General El ectric Company, NEDE-25090-1-P, Re-vision 1, May 1981.**

d

,r. - . - . v -- .. . , _ . y ,