ML20197G069

Forwards Sser Based on Containment Sys Branch & Mechanical Engineering Branch Review Re Methodology Used to Calculate Drywell to Wetwell Vacuum Breaker Response to Pool Swell & Chugging Phases of Postulated Loca.Methodology Acceptable

ML20197G069
Person / Time
Site:	Columbia
Issue date:	11/21/1983
From:	Houston R Office of Nuclear Reactor Regulation
To:	Novak T Office of Nuclear Reactor Regulation
References
CON-WNP-0685, CON-WNP-685 NUDOCS 8312020390
Download: ML20197G069 (6)

Text

distri bution central file .

, csb rdg file #

~

rs rdg file

~ wbutler jkudrick feltawila jknight NOV 211983 rbosnak .

yli

!!EMORAi!9t#1 FOR: T. Ilovak, Assistant Director for Licensing, DL FR0't: R. W. Houston, Assistant Director for Peactor Safety, DSI J. Knight, Assistant Director for Components & Structures Engineering, DE

SUBJECT:

!IASHIl!GT0fl PUBLIC PDt!ER SUPPLY SYSTEli, UT!IT 2 lHIP-2: DRYflELL TO WETWELL VACl!Ut1 BREAKER ALVES The Containment Systems Branch and the f:echanical Engineering Branch have completed their review of the liflP-2 applicant's sihaittal that describes the nethodology used to calculate the drywll to vetwil vacuum breaker response to the pool smil and chugging phases of a postulated LOCA and tne design rodi-fications to accomnodate the loads generated during these phases of the LOCA.

Based on our review of the applicant's submittal, as detailed in the enclosure, ta find that the analysis presented is acceptable for use in predicting imact velocities for the vacuum breaker valve body and seat and is acceptable for use in qualifying the vacuum breakers. He also find that the valve nodifications ,

to accon oriate these loads are acceptable. The riodified valves should be in-stalled prior to fuel load.

Odf.v!?Wl 3J R.Wcyqh:qt:1 R. U. Pouston, Assistant Director for Reactor Safety, OSI

- - - ~ ~ . .

k 8312020390 831121 J. Knight, Assistant Director for S ADOCK 05000 '

Components & Structures Engineering, DE

Enclosure:

As st1ted cc: R. Mattson D. Fisenhut R. Vollmer A. Schwencer -

9. Cosnak  ;

R. Auluck f __.o__f _ , _ _ _

"'c" f CS  : . CSB:%/., . CSB:DSI

. RSfQ{Pf . iiEB : DE.. MEG- . iiEB d

" " ~ '

% CSB b,ITg433 ...<. RHoiTitdn Yueh Yyig.Li..tEBn@, 17 JKn i g .1 t.

=>

Ch .

' 4 t ,.

) 11/@G3 11/jT/83.. .... 11/... /S3 . ..

11// I/_83 ~ 11N3/CL a c .na n.v m om n ac .i c:a ._ OFFICI AL RECORD GQPV.

SUPPLEMENT TO THE SAFETY EVALUATION REPORT WASHINGTON PUBLIC POWER SUPPLY SYSTEM, UNIT 2 WNP-2 The WNP-2 containment is equipped with simple, swing check double disc valves to serve as vacuum breakers to equalize the pressure between the drywell and wetwell air space regions so that the reverse direction pres-sure across the diaphragm floor wiLL not exceed the design value. The vacuum relief valves (nine assemblies) are mounted on selected downconers inside the suppression pool air space region.

Following the onset of a loss-of-coolant accident and during the pool swell phase, air will flow from the drywell through the vent pipes and suppression pool into the suppression pool chamber air space. Muring this process, the suppression pool surface will rise and the air space region above it uilt be compressed. This transient pressurization of the wetwell air space may cause the vacuum breaker valves to experience high opening and closing impact velocities. Later on into the transient, at about 20 seconds, the chugging phenomenon takes place. This chugging phenomenon is due to the rapid condensation of the steam at the vent exit. The rapid steam condensation wiLL cause repeated and strong dynamic underpressure conditions

(-10 t o -30 p s ig) in the vent pipe while an overpressure condition of about 20 psig wiLL exist in the suppression chamber. The vacuum breakers will be exposed to this transient pressure differential causing the valves to open repeatly.

Since the chugging phenomenon is repeated every two seconds,on the average, the vacuum breaker may be called upon to function in a cyclic manner and at potentially damaging velocity during these intermittent steam condensation events, i

Failure of a vacuum breaker to close during these times (pool swell or chugging) could result in steam bypass of the pool, thus jeopardizing the integrity of the containment. To estimate the valve disc actua-tion velocities, the Mark II Owners Group developed a vacuum breaker valve dynamic model (NEDE-22178-P(1)) which was submitted for review by the staff.

The Containment Systems Branch (CSB) has completed its review of NEDE-22178-P, which describes the generic methodology to be used in calculating the response of the drywell-to-wetweLL vacuum breaker to certain transients in the Mark II containments (memorandum for T. Novak from T. Speis, dated November 2, 1982.(2)). The proposed approach'was found acceptable.

During a meeting held on June 7, 1983 in Bethesda, Maryland between the staff and the applicant, a presentation was made of the analysis, the tests and the redesign which produced a reduction in the valve impact velocities during pool swell and chugging.

Several changes were made to the valve design which contributed both to a reduction of the impact velocities and to the strength of the valve.

The changes include:

1) Increased mainshaft diameter; 2) added mainshaft outboard and midspan bearings; 3) added special main shaft keys and locking features; 4) increased main shaft / bearing cap clearance; 5) in-creased section properties for arms, columns, linkages and linkage (1) " Mark II Containment Drywell-to-Wetwell Vacuum Breaker Models,"

General Electric Company Report No. NEDE 22178-P, August 1982.

(2) " General Electric Topical Report Review," Mark II Containment Drywell-to-Wetwell Vacuum Breaker Models," November 2, 1982.

I

s'm attachmonts; 6) utilized high strength materials; and 7) rein-forced disc and added damper devices (tuned PSA-3 snubber designed and built to ASME Section III, Class I, Subsection NF requirements)

, which inhibit disc actuation during chugging.

The modified valve, as described above, was then tested. The initial test series was conducted to obtain the valve unique hydrodynamic torque curve as specified by the approved methodology.

Several tests were run in which the two valve discs were locked in selected

, positions using turnbuckles and the torque required to hold them in posi-tion measured as the pressure across the valve was varied. The measured hydrodynamic torque from these tests was used in the dynamic vacuum breaker model (Reference 1) computer code to determine the valve impact velocity.

The next series of tests were run with the valve in its operational con-figuration (i.e., snubber functioning). The valve was cycled several times with pressure sufficient to cause the snubbers to limit the ac-celeration of the disc throughout all but the initial phase of the the opening stroke. Results from these WNP-2 valve unique flow tests were used to calibrate the vacuum breaker valve dynamic models.

j The applicant stated that the dynamic responses of the valves are well predicted by the dynamic vacuum breaker model and that the loading bases used to redesign the valve are, in all cases, fully bounding over the test calibrated dynamic model results.

The predicted pallet impact velocities for the modified valve (using the dynamic model described in Reference 1, time dependent differential pres-sure loading across the vacuum breaker disc derived from the 4TCO test

-e-- - - - -

,y.--.-- - - - - . - - - - - - - _-

.-- .----_m--- - _ ___-- - - - - - - - - -

data and the measured hydrodynamic torque) are opening and closing impact velocities of less than 2 radians /sec for both pool swell and chugging phenomena which are well below the tested values.

The applicant has performed an analysis to verify the modified vacuum breaker valve's structural and pressure integrity. Full scale static cnd dynamic tests of these valves were also performed. As a result of its assessment, the applicant concluded and the staff concurs that under the worst case combined faulted loading conditions, the valve response esmains fully elastic. In addition, post test visual inspection and leakage test showed that valve operability and integrity as a pressure boundary are maintained.

One area of concern during our review was the in-service operability of the tuned PSA-3 snubbers which are utilized as a damper to reduce the opening and closing velocity of the valve disc. The applicant responded to the staff's concern by letter dated November 15, 1983 (3). The applicant stated that the in-service operability of the snubbers is assured by the vacuum breaker valve periodic operability testing as required in Sections 3.5.4.1 and 4.6.4.1 of WNP-2 Technical Specifications. The subject Technical Speci-fications require that the vacuum breaker valve be demonstrated operable at least once per 31 days and within 2 hours2.314815e-5 days <br />5.555556e-4 hours <br />3.306878e-6 weeks <br />7.61e-7 months <br /> after any discharge of steam to the suppression chamber from the safety / relief valves, by cycling each vacuum breaker through at least one complete cycle of full travel. Actuation of any given vacuum, breaker valve disc, via the remotely operated test actua-tion air cylinders, will fully actuate each valve snubber since the valve disc and snubber are close-coupled. The staff agrees that the in-service operability of the snubbers will be adequately demonstrated by performing these required tests.

(3) Letter from G. C. Sorensen to A. Schwencer, " Suppression Chamber -

Drywell Vacuum Breaker Snubber Operability", dated November 15, 1983.

Based on the analyses performed by the WNP-2 applicant which verified the valve's structural and pressure integrity and on the test results which domonstrated the valve operability and functionality, the staff finds that the design of the modified vacuum breaker valves at WNP-2 is acceptable and can accommodate the effects of pool swell and chugging impact loadings fol-lowing onset of a design basis LOCA.