ML20202B652

Sanitized Version of Rev 0 to Mark I Wetwell to Drywell Vacuum Breaker Load Methodology

ML20202B652
Person / Time
Site:	Pilgrim
Issue date:	02/29/1984
From:	Bilanin A CONTINUUM DYNAMICS, INC.
To:
Shared Package
ML20202A858	List: BECO-86-041, Forwards Sanitized Rev 0 to Mark I Wetwell to Drywell Vacuum Breaker Load Methodology & Rev 0 to Mark I Wetwell to Drywell Differential Pressure Load & Vacuum Breaker Response for Pilgrim Station Unit 1, Per 860123 Request ML20202B652 ML20202B679
References
84-3-01, 84-3-1, 84-3-R, 84-3-R00, NUDOCS 8604110257
Download: ML20202B652 (68)

Text

-. -

• Sekenoce- \

4 C.D.I. REPORT NO. 84-3 MARK I WETWELL TO DRYWELL

, VACUUM BREAKER LOAD l METHODOLOGY i

l Revision 0 Prepared by CONTINUUM DYNAMICS, INC.

P.O. BOX 3073

, PRINCETON, NEW JERSEY 08540 l

l Prepared under Purchase Order No. 205-YF495 for GENERAL ELECTRIC COMPANY 175 CURTNER AVENUE SAN JOSE, CALIFORNIA 95125 Approved by bF AMed Alan J. ilanin February 1984 8604110257 860407 PDR ADOCK 05000293 p PM

i

> e 6

DISCLAIMER OF RESPONSIBILITY Neither the General Electric Company nor any of the contributors to this document makes any warranty or representation (express or implied) with

- respect to the accuracy, completeness, or usefulness of the information contained in this document or that the use of such information may not infringe privately owned rights; nor do they assume any responsibility for liability or damage of any kind which may result from the use of any of the information contained in this document.

PROPRIETARY INFORMATION NOTICE This document contains proprietary informatien of General Electric L Company and is submitted in confidence solely for the purpose or purposes stated in the transmittal letter. No other use, direct or indirect, of  ;

! the document or the information it contains is authorized. Furnishing j j this document does not convey any license, express or implied, to use any l l

patented invention or any proprietary information of General Electric l ,

disclosed herein, or any right to publish or make copies of the document l without prior written permission of General Electric.

n-I l

1 i

i i

1 _ . , - . . - _ _ - - _ . , _ -..,_ -_ , -, _ -

SUMMARY

The following report documents the methodology used to determine Mark I wetwell to drywell vacuum breaker (VB) actuation velocities during the chugging phase of a postulated loss of coolant accident (LOCA). Differential pressure loading across the vacuum breaker exists during chugging as a consequence of the rapid condensation occurring at the downconer exits.

Sufficient data has been obtained during the full scale test facility (FSTF) test series to define this differential pressure load conservatively. The FSTF test series, however, because of relatively small drywell size, produced differential pressure loads across the VB far in excess of what would be anticipated in Mark I plants. This differential pressure loading is adjusted for plant-unique drywell volumes with a vent dynamic model which has been validated with independent FSTF test data. The resulting plant-unique differential pressure loading is then used to drive a wetwell to drywell valve dynamic model which calculates the disc actuation velocities on a plant-unique basis. The valve dynamic model has also been validated against full scale test data and conservatively predicts actuation velocities. Valve disc actuation velocities are then used to confirm the adequacy of the installed design.

b d

!F i

l 1

TABLE OF CONTENTS Section h

^

i Summary 1-1 1 INTRODUCTION FSTF TEST PROGRAM 2-1 2

4 THE VENT DYNAMIC MODEL 3-1 3

3.1 Mark I - FSTF 3-1 3.2 Model Description 3-1 3.2.1 Dynamics of the Steam Vent System 3-3 li 3-8

' 3.2.2 Dynamics at the Sisam Water Interface (Vent Exit)

' 3.2.3 Dynamics of the Pool and Wetwell Airspace 3-9 3.3 Model Validation 3-14 3.3.1 Determination of the Condensation Source 3-14 3 Velocity and Prediction and Comparison of Ring Header Pressures 3.3.2 Effect of Vacuum Breaker Actuation on 3-24 Measured Vent Pressures I

4 PLANT-UNIQUE DIFFERENTIAL PRESSURE LOADS 4-1 I 4.1 Data Set Selection for Plant-Unique Loads 4-1 4.2 Internal Vacuum Breaker Forcing Function 4-1 l

i 4.3 External Vacuum Breaker Forcing Function 4-3 l-

!- 5 VACUUM BREAKER VALVE DYNAMIC IODEL 5-1 E 6 CONCLUSIONS 6-1 7 REFERENCES 7-1 APPENDIX A-1 ii

l

. 1 I

l l

LIST OF ILLUSTRATIONS l Fiaure Page,,

2-1 A Schematic of the FSTF torus 2-2 2-2 Ring header pressure during FSTF run M1 2-4 2-3 Vacuum breaker disc angular displacement time history 2-5 during run M1 3-1 Steps in determining plant-unique vacuum breaker 3-4 forcing functions 3-2 Schematic model of the vent system 3-5 3-3 Details of the deflector plate mode 3-7 3-4 Details of the steam water interface 3-10

.; 3-5 Details of the pool dynamic model around each downcomer 3-11 3-6 Measured drywell pressure fluctuations during 65.9-105.9 3-15 seconds of FSTF run M1,

a. First ten seconds

[

3-6b Second ten seconds 3-16 3-6c Third ten seconds 3-17 3-6d Fourth ten seconds 3-18 3-7 Predicted condensation source velocity at the vent exit 3-19 for M1. a. First ten seconds

! 3-7b Second ten seconds 3-20 i

f 3-7c Third ten seconds 3-21 3-7d Fourth sin seconds 3-22 3-8 Predicted (solid curve) and measured (dashed curve) 3-25 pressure time histories at transducer P5901 for M1

a. First ten seconds 3-8b Second ten seconds 3-26 3-8e Third ten ser.mds 3-27 3-8d Fourth ten seconds 3-28 3-9 Predicted (solid curve) and measured (dashed curve) PSD's 3-29 for M1 111

-. .,_---n

-. , , . , , - - - , - - . . , - - , - - _ , , , - . . . - _ . , . . , - - , , - _,n., , - -.

~.

LIST OF ILLUSTRATIONS (Cont'd) l Figure Page, 4-1 ras predicted pressure at the main vent / header junction 4-5 as a function of drywell volume / vent area 4-2 Pressure time history prediction at the usin vent header 4-7 junction for the conservative Hatch 2 drywell volume / vent area of $32.87 f t (Group 2). The conditions of run M1 as described in Fig. 3-6 apply. The hydrostatic head has not been added to this figure. a. First ten seconds 4-2b Second ten seconds 4-8 4-2c Third ten seconds 4-9 4-2d Fourth ten seconds 4-10 i

4-3 Schematic of the external pipe end loss mechanism 4-12 4-4 Pressure time history predictions across the external 4-14 vacuum breaker for Dresden, using a nonlinear and flow damping mechanism. The conditions of run M1 as described e in Fig. 3-6 apply with the hydrostatic head not added to the time history. a. First ten seconds 4-4b Second ten seconds 4-15 4-4c Third ten seconds 4-16 4-4d Fourth ten seconds 4-17 A-1 Schematic of a vacuum breaker valve A-2 i

t iv

e O

, ~.

LIST OF TABLES Table h i

2-1 FSTF Test Matrix 2-3 2-2 Su.anary of Chugging Data Base 2-6 3-1 Plant-Unique Characteristics 3-2 i 3-2 Basic FSTF System Parameters 3-23 i

4-1 Vacuum Breaker Actuation During Chugging 4-2 4-2 Mark I Drywell Volume / Total Vent Area 4-4 4-3 Drywell Volume Ef fect 4-6 A-1 Vacuum Breaker Characteristics A-5

- !i J

]

, is

'jI

^ !'i.

ij.

}l!l

/

! f

1' l'

I l

I I

l 4

l l

l 0

~.

1. INTRODUCTION The FSTF was constructed as part of the Mark I long term containment program to evaluate containment loads during a LOCA. This facility was a 1/16 sector of a Mark I containment torus and contained a ring header and eight downconers. The facility included the installation of a GPE wetwell to drywell vacuum breaker on the ring header. The function of this vacuum breaker is to mitigate the differential pressure between the torus wetwell airspace and ring header /drywell.

The first test, in a series of 13 tests, was a simulation of a small break accident (SBA). The wetwell to drywell vacuum breaker, which was instrumented h with a dise position indicator, cycled during this test. Pos t-tes t data I analysis revealed that chugging occurred for approximately 300 seconds and d that the vacuum breaker cycled repeatedly and sustained some damage to the dise hinge and gasket. These results initiated the Mark I vacuum breaker loads program documented in this report.

4 W This loads program has shown that the actuation velocities sustained in Ip, the FSTF test program are not protypical and result from conservative sizing of the drywell volume as will be discussed below.

This report is organized as follows: In Section 2, the FSTF test program is briefly reviewed and the chugging data base, used for the differential pressure load definition, is discussed. Mark I containment geometries are compared to the FSTF facility in Section 3, and a vent dynamic model is described to adjust FSTF data for application to existing Mark I containments. Described in Section 4 is the calculation of plant-unique )

differential pressure loading from FSTF data for internal and external vacuum breakers. In Section 5, the description and validation of an analytical vacuum breaker dynamic model is summarized. Lastly, in Section 6, conclusions are offered.

l 1 1-1 l

i

1

2. FSTF TEST PROGRAM 1

The FSTF test facility was constructed to determine unsteady loads in Mark I containments during a postulated LOCA. The facility was a 1/16 sector ,

l of a Mark I torus and contained eight downconers extending from a ring header

into the containment pool. On the ring header was a GPE 18" internal vacuum breaker. A schematic of the FSTF facility showing vacuum breaker location is shown in Figure 2-1. The primsry function of these vacuus breakers was to linit possible wetwell presure differential above drywell pressure.

1

, The FSTF test program was undertaken between the spring of 1978 and the

,l summer of 1980 and was comprised of 13 tests or blowdowns. Each blowdown was j a simulation of a postulated LOCA. The conditions for these tests are

summarized in Table 2-1. During the first test of the program, test M1, which l

simulated an SBA, an unsteady condensation phenomenon known as chugging was

' j" observed to occur (Ref. 1). The measured pressure in the ring header is shown in Figure 2-2. Note the oscillations which are superimposed on the containment pressure. During the time interval shown here, a vacuum breaker iJ disc position indicator was functional and the angular displacement time ll history recorded is shown in Figure 2-3.

I

! Subsequent analysis of all the test data indicates that only SRA tests produce chugging with sufficient magnitude of differential pressure to result AI in vacuum breaker disc actuation. On Table 2-2 are summarized those runs and time intervals over which chugging was observed to occur. Note that over 1000 i seconds of chugging data were recorded in which there were nearly 400 distinct chug events which actuated the vacuum breaker 179 times.

4 i

i The most severe vacuum breaker actuation was surmised to occur during test M1 between 65 and 105 seconds into the blowdown. During this time, the j minimum differential acting across the vacuum breaker disc was 1.82 psid.

This was the most severe differential pressure loading seen during the FSTF 1

program.

1 2-1

~

l l

1 1

l

} STEAM STEAM FROM FROM DRYWELL DRYWELL e Vacuum snaAmen

}' /

%F r T

(6[

m m m m

./

rr 1 7 ,vs.,

Insnoen L L

- - - J J

__ - _ _  %\

> \ oow=cousna i

LOCATION OF FSTF VACUUu enEActER I

I I

Figure 2-1. A schematic of the FSTF torus i

]

l I 2-2

i TABLE 2-1 FSTF Test Matrix BREAK WEWELL NOMINAL INITIAL CONDITIONS Test

• Date Number Performed Size Type Submergence Temperature Pressure First M1 5/5/78 Small Steam 3 f t 4 in 70'F 0 psig Series M2 5/12/78 Medium l M3 5/25/78 Small Liquid p l !, M4 6/17/78 Steam u 5 psig i M5 6/26/78 o 120*F G puis s

f M6 7/6/78 1 f t 6 in j M9 7/11/78 4 ft 6 in 70*F L

M10** 7/27/78 o 3 f t 4 in M7 8/10/78 Large o 1

M8 8/22/78 Liquid u I

i il

, Second SDA 4/8/80 Small Steam 3 f t 4 in 70*F 0 psig i Series Large M11B 8/22/80 Liquid q M12 8/27/80 ( } U 95'F e i

J l

• Shown in order of performance

• • Air sensitivity test performed with vacuum breaker replaced with i rupture discs 2-3

l

{

i as M1 , , , ,

2 27. -

M l o

i 24- _

i

- j

CD

~

m O_ 21. ~

i i

18 . ~

i 15 ~

l 1 ' ' ' '

l 0 10 20 30 40 50 TIME CSEC) i j Figure 2-2. Ring header pressure during FSTF run M1

,l  !

g n

i 0 r W 5 u d

_ p y r

o t

s i

h M e

. . 8 0 m i

4 t t

n e  ; e m

e c

a l

p

. . 8 s 0

3 )C i E d

" S C

r a

l u

E g n

Ma I e 1'

g

. . a 0T 2 i c

d w r e

k a

M b e

r a1 t

a 0 uM

_ 1 u

_ cn

_ au Vr l

. 3 1 2 M . _ - . . . e 6 3 , 1B r 7 8 5 2 9 u 2 M M 1 1 1 - g i

F n8s J

4 g 5 e~a:o u3o ga auo E 5 S Yw

il1 1i!

l!  ;

i-;' ) }il;4j i i! !.

TABLE 2-2 l

Summary of Chugging Data Base Test Number M1 M4 M9 M10 SDA Total i

Approximate chugging period, seconds30-330 26-119 25-305 20-120 20-300 250-305  ;

i

. Seconds of chugging data recorded 300 90 280 155 280 1105 Determined ring-header chus events 89 42 101 59 101 392

Number of vacuum breaker actuations 43 26 37 38 35 179 1

l i.

i l

l l

9 2-6 ,

i

i i ..

i

,1

3. THE VENT DYNAMIC M) DEL 1

l' 3.1 Mark I - FSTF i l

! The sizing of the FSTF facility was protypical of Monticello (Ref. 1). l The FSTF torus represented a 22-1/2' segment (1/16) of the Monticello torus and the FSTF drywell volume represented 1/16 of the Monticello volume. For l

j reasons of adding conservatism to the torus hydrodynamic loads, however, FSTF l was constructed with 8 of Monticello's 96 downconers, or 1/12 of the total.

l Therefore, FSTF's drywell volume / total vent exit area was approximately 12/16 i; . or 3/4 of Monticello's. On Table 3-1 is tabulated plant-unique vent system characteristics. Note that the FSTF facility has the smallest drywell volume to total vent exit area ratio of all domestic Mark I plants and that this I' ratio for FSTF is over 2.5 times smaller than that of Duane Arnold. It is s

!l important to keep in mind this ratio, for it will be shown to strongly control 3

l' the magnitude of the vent system pressure oscillations during chugging and, hence, strongly influence vacuum breaker actuation.

i Since plant-unique details may affect the actuation of the vacuum breaker, a need exists to quantify the ring header /wetwell pressure fluctuations to be anticipated in the Mark I plants. It is important to identify a conservative l yet realistic vacuum bresker forcing function. This section details a 6 methodology by which quantitative vacuum breaker forcing functions for plant-I' unique application asy be extracted from FSTF data.

h i

e 3.2 Model Description i

l The analytic approach taken is analogous to that used (Ref. 2) to

! determine torus rigid wall loads from measured FSTF wet shell pressures.

Here, however, instead of determining an unsteady vent emit pressure which l

reproduces measured pressure and acceleration data on the torus, a model is l developed that allows the extraction from the data of the unsteady 1

i condensation rate at the vent exit. The key assumption then made is that this condensation rate is a facility independent quantity. This assumption is

! 3-1

i l

TABLE 3-1 i

1 4 Plant-Unique Characteristics Drywell Header j Volume / Downconer Area / Vent Area /

4 Vent Area / Downcomer Downcomer Main Vent Header Downconer Plant Area (ft) Pool Area Area Area Length (ft) Length (ft) Length (ft) 5 l i

j 1 FSTF 292.78 .045 1.47 .99 37.3 15.0 10.8 Monticello 413.62 .035 .98 .99 25.0 19.5 10.8 l Pilgrim 455.91 .033 .94 .95 25.3 20.3 11.6 i Browns Ferry 490.70 .027 .98 .99 31.1 22.2 11.5 Quad Cities 491.40 .030 .97 .95 27.0 21.7 11.4

[ Match 2 532.87 .026 1.06 1.02 29.7 21.3 11.5 t

{

Cooper 542.25 .027 .91 .91 27.4 20.2 11.4 l Peach Bottoe 549.75 .027 1.00 .98 31.2 22.2 10.3 -

Hatch 1 586.42 .026 1.06 .91 30.0 21.3 11.2 l

i Brunevick 591.03 .032 .83 .82 25.2 21.7 10.7 1 t i Nepe Creek 642.24 .022 .96 1.0! 32.6 22.4 10.5 i .023 .94 .94 32.5 22.4 10.2 Fermi 658.06 l

, Duane Arnold 755.11 .018 1.07 .98 27.3 19 .6 8.5 i 1

! Dresden 490.45 .030 .97 .95 28.06 21.68 11.43 i

4 Fitspetrick 476.17 .029 .98 .99 30.08 21.48 11.99 i t

l Millstone 458.11 . 033 .94 .95 25.14 20.29 9.37 Nine Mile Pt 434.59 .037 .96 .95 35.28 19.54 10.13 f

) Oyster Creek 487.18 .039 .91 .92 27.93 15.99 10.51 4

Yermone Yk 411. % .035 .98 .99 24.48 19.49 12.29 i

i 7

a  ; .

I

k i

supported by the observation that the condensation rate is fixed by local

! conditions at the vent exit; i.e., steam mass flow rate, noncondensibles and i thermodynamic conditions, and that these local conditions vary slightly I

between plants.

i i

' The steps in the approach to obtain plant-unique vacuum breaker forcing functions are shown in Figure 3-1. Since unsteady pressure data was taken at i several locations along the vent and in the drywell, existing data may be used

! to validate the model (see Step 3 in Figure 3-1).

t

{

1i 3.2.1 Dynamics of the Steam Vent System t

A schematic model of the vent system is shown in Figure 3-2, depicted by 12 dynamic components. Each component except for the jet deflector (component 9) and possibly the drywell (component 10) are i

assumed to be governed by one-dimensional linear acoustics theory; 4

consequently, the unsteady pressure field satisfies the equation i

i 2 2 Ku

}

I [ - k Iat[

3x a aL 2

=0 (3-1) l

! where i

j a -

acoustic speed in steam K -

loss factor j Ia,

- mean steam velocity I L -

length of the component I

i i At the junction of one component with snother, pressure continuity and asss conservation are enforced. Steam velocity is taken as sero at all

! closed ends of components: to the right of component 4 (a reflecting i

! plane); at the top of the drywell, component 10, when the drywell is a

modeled as an acoustic leg; and on either side of an external vacuum l breaker (the vacuum breaker is assumed to rensin closed), when one exists i between components 11 and 12.

i i

3-3 1

\

l l

t

! STEP 2

Develop a dynamic model of the i vent system, steam water inter-i 1 face and pool slosh with the

! condensation rate at the int.er-i face unknown, i 1

l 1 4

Use measured drywell pressure to 2

determine the condensation rate.

.i,

!: With the condensation rate il 3 determined, predict unsteady j pressures at other vent locations i to validate the model. ,

l O

1[i

Use the condensation source at l the vent exit to drive dynamic 4 models of Mark I plants to determine unique vacuum 1 breaker forcing functions.

{

l jl Figure 3-1. Steps in determining plant unique vacuan breaker i, forcing functions i

i 1

1 I

i l

3-4 i

i

i 10 - Drywell External Vacuum ---

Breaker Piping 9 m Jet Deflector Plate I

\ 8 ~

Main Vent EL Header s ..... .........

7 Wetwell  ! 6 5 ' 4

'L Airspace -- --

J-12 3 2 1 M ~--  % ~ n l!i <l

/

Downcomers Figure 3-2. Schematic model of the vent system I

l 3-5

1 This type of modeling is known as acoustic circuit analysis and is of ten used to determine natural frequencies of complex piping systems.

1 Figure 3-2 represents one-half of the FSTF vent system. Thus, the volume i of the drywell (component 10) is one-half of the total dryvell volume in FSTF, and the cross-sectional areas of elements 7, 8 and 9 are one-half of l the plant main vent area. Along the ring header, three downconer pairs are provided, although in a plant-unique analysis one or more of the downconer pair cross-sectional areas may be set to zero. The predicted condensation rate is assumed to be the same at the exit of all downconers since sufficient data does not exist to determine downconer-unique condensation sources.

The drywell is modeled in two different ways to account for the different geometry between the FSTF drywell and one in a plant. In FSTF, b the drywell has a high aspect ratio tank (length / diameter) and is more

$ sppropriately treated by one-dimensional linear acoustics, whereas Mark I i drywells are of aspect ratio near unity, so it is more appropriate to treat the steam properties in the drywell as being unif orm. Thus, the pressure in the drywell is related to the steam velocity exiting the drywell by the equations p

Il h =hp,au (3-2) j where u - unsteady component of steam velocity !nto the drywell p, - mean steam density A - flow area of the deflector plate V - volume of the drywell it may be shown that both models for the drywell are equivalent for sufficiently low frequencies.

Component 9 is a model of the jet deflector plate positioned at the l l

end of the main vent, shielding its opening into the drywell. Referring to Figure 3-3, the plate typically has a radius R p greater than the

\

3-6 l l

1 l

~. -_. _ ._ _ _ _ __ __ _ _ . _ __

l l

l l

Drywell Floor i < 2R p  ;

Jet Deflector i Plate j, f

i. o s s

1. \ h l j U dp 4 U

l U dp l

. Umv  !

Main Vent

( A J l

< 2Rm  ?  !

! 1 i l l Figure 3-3. Details of the deflector plate model 3-7 i

L__-__--_....-.-.-.....-...-._.------.-._.--_..__.

radius R, of the main vent, and is positioned a distance h above the end of the vent. Mass continuity requires the velocity leaving the deflector plate udp to be related to the velocity leaving the main vent u,y by the relationship R

~

"dp " 2Rp h "av neglecting the small density change across the deflector plate.

j Fluid momentum gives the relationship for pressure:

l j du (34) pdp " p,y-P s d p n(Rp /R,)

where pdp is the pressure entering the deflector plate from the drywell l

side and p,y is the pressure on the main vent side of the deflector plate.

other details of the model are that the acoustic speed in steam has l been assumed to be a= 1500 f t/sec, vent system f rictional losses are assumed to be distributed along the downco'mers only, and the mean steam velocity, from a consideration of FSTF blowdown flow rates, is taken to be u,= 75 ft/sec .

i j 3.2.2 Dynamics at the Steam Water Interf ace (Vent Exit) i o

l The boundary conditions on pressure, density and fluid velocity at

! the steam water interface at the ends of the downconers are jump conditions obtained using momentum and mass conservation. Conservation of i momentum requires thats l

l 3-8

2 P, + 9,U,

• P,+ p,U,2 (3-5) where ( ), are steam conditions and ( ), are water conditions, in a coordinate system fixed on the interface. Mass conservation requires:

s = p,U, = p,U, . (3-6)

In an inertial coordinate system, l

l dh, "o " s + dt (3-7) l l. dh "w " Uw + dt where dhy /dt is the velocity of the steam water interface. Figure 3-4 schematically details the interface.

i i  !

l 3.2.3 Dynamics of the Pool and Wetwell Airspace I

i The wetwell pool per downconer is shown schematically in Figure 3-5 One-dimensional asse conservation in the pool givest a

j h (h Agp) + h (h,A,) = -u,A, (3-8) 1 1

1 3-9

.t i

l o

4 l

4 Steam side

-u s

4, Steam water d_ _ interface i,

-u U W l

dh water side W

dt l

l 9

i If Figure 3-4. Details of the steam water interface, i

i I

'l I

l i

3-10

1 l

l j Wetwell Airspace b /

eA  ;

p

-w _ _ Pool a f v li j t H

A,

• h t w o

nen 4 ,

u, hW

'. 4 X

i i

-1_____-

4 u=0 t

[

d 4

l

! Figure 3-5. Details of the pool dynamic tr.odel around each downcomer.

I 3-11

l 1

where hg - ef fective depth of the pool h, - effective height of the water column in the downconer-Ap -

pool surface area per downconer

A, - downconer area condensation source velocity at the end of I ue -

the downconer with the pool surface velocity defined by:

dh ug = dt *

~'

ll it An energy balance in the pool analogous to that used in Ref. 3 yields:

)

A A dh

[ h (h,u ) + [ h (h u gg ) + W ) [A uhu, dt 1

• (3-10)

A A j

=

g[ u,p, g/l upgg - gu ,h,A, guggp hA w w 1:

'!; where p, - pressure on the water side of the steam water ,

i interface pt - wetwell airspace pressure i

! 4 - pool damping factor j

Equation (3-10) is nonlinear and is linearized about a mean (denoted by overhars) to obtain linear equations for fluctuating quantities.

Assuming a mean plus a perturbation for all velocities, pressure and lengths, Eqs. (3-8), (3-9) and (3-10) yield 3-12

- l l

I du h,pw de + (l+")'s"s"w " P g P, 83,(h-h,) t l

l dh, dt " "w ~ "c (3-11) dh*

dt

= - u ww A /A p

Pg" tt t where I5, is the ef fective mean depth of the water below the downcomer exit and pg is the unsteady pressure in the wetwell airspace. This pressure fluctuation crises from oscillations of the pool surface elevation which isentropically compress the air in the airspace.

The model is completed by coupling the pool dynamic model to the i interface jump conditions; these equations are then coupled to the steam h' vent dynamic equations. The interface jump conditions are linearized to give:

Im P, " P, - h,E,(u, u,) - ( p,-p,)u,u (3-12) h

!h 1

up I

_s s (3,33) u, = u (p,/p, - 1) + u ,

p,a There exists one more dependent variable than there are linear equations. The unsteady condensation source velocity u e on the water side is taken as the unknown. The above system of equations is Fourier transformed in time and the resulting system of linear equations is solved exactly.

To determine the condensation source velocity, FSTF vent pressure time history at one location is Fourier decomposed. A unit condensation source is placed at the vent exit, frequency component by frequency component, and the pressure is predicted at the location where the vent 3-13

0 The ratio of the predicted pressure to pressure time history is given. In as the measured value is used to ratio the unit source at the vent exit.

this manner, the condensation source is reconstructed frequency component ,

by f requency component, then Fourier inverse transformed to give the time l history of the unsteady condensation velocity.

• l 3.3 Model Validation E The solution to the equations determines the unsteady condensation source velocity ue (t) from measured pressure time histories in the steam vent system and is based on a linear dynamic model of the wetwell airspace, pool, steam vent system. While the model is steam water interface and straightforward conceptually, the many elements which comprise the model each contain some approximation. Therefore, it is prudent to check the model against data to assess its accuracy.

3.3.1 Determination of the Condensation Source Velocity and Prediction and Comparison of Ring Header Pressures .

I In FSTF, pressure transducers were located at several positions Therefore, these along the steam vent system including the drywell.

Since the independent measurements provide a means of checking the model.

l drywell is the component which is the greatest distance from the vent exit where condensation forcing is taking place, a severe test of the model would be to use a pressure time history in the drywell, determine the condensation source velocity at the vent exit, then use this source to Figure 3-6 predict the unsteady pressure elsewhere in the vent system.

shows the measured drywell pressure fluctuations during run M1 (65 to 105 seconds), while Figure 3-7 shows the corresponding predicted unsteady condensstion source velocity at the vent exit. Note that this velocity is a water velocity. The FSTF parameters used to obtain this result are listed in Table 3-2.

3-14

g. ,

8

A 8

8.

p fgy Z

VM bj[ b#hi y ~8 4

G Ef

3 I k

11-8 i'

! 8 ,,

! f -

i

, 8,

'O,ge 3.00 2.00 S.00 4.00 5".00 8".00 I.M NGO O'.W l'O.R TIME IN SECONOS l Figure 3-6._. Measured drywell pressure fluctuations during 65.9-105.9 seconds of FSTF run

M1.

i a. First ten seconds.

i

z a -L- - - p -

j B

..f 8

.f B.

B.

4 (,

u 8

. .a-i

', C

l 1 2

Bo*U

=

=CW F)

Z

~

B.W

. .*E "ame b

i

,1 B -

.f 8 l 8 U

G

~

B =

. .4- e G

AJ T

B* 8 U

• e"e G sn.

o 8 i ja

^

00 i- 00.g. 006-00 Jg 00*2 00*1- 00's-00't,g.}Y;9g J N1 30011 W 6 7 no w

la.

3-16 l

l

4 e

B

..R B

.i B.

4>

B A

.u B,

h, '

i, ai n m

"- C 2

x Be u

• CW m

2 B.w

ll! *

• E~E

'D iji B .

d .E e

o U

G B = l "N g u

'tl)

N B. *r4 i

• E h .

'i '

B. M E e 5 m 's m*z

""! .G'I S d NI bbO1I"1

$14 l

l 1

3-17

ll l g

o

c. -

go -

o n.

e t o

s o s em a

n.

e s

o c.

7 3

o n.

e s

S

.- D w oN c.OC

sE s

S N

I

=

o n.E

tM s

_ I T

o

= c. s s d

=.

=

s n o

k c

e o s c.

as n e

t h

- t o r

- c. u

3 1

o F

d o 6

c. 3

. o

'i 8 A's e 8a E o 8 :. . *7" 8i 8J r s 8d d,_gm.' *~ wo3> .a s8.E E u g

i F

ig

E[.

R*

~

l h

38 i

C+

4 O

N$..

g

{ 7" i R

+

l a.no [.no /.no d.m ' ' ' '

TIEINSUONOS Figure 3-7. Predicted condensation source velocity at the vent exit for M1 l . a. First teh seconds, i

+m< _ _ a- 2 -,,

e 8

.;g 8

.i 8

.f B

A-C

- E o

S.u

• CW en x

B.w

/

• g E.

l B.

,m W

, t3

'g U

G B "

sEu "t3

s. E U

.>= " G t/2 N

1. 8 A o

e 00*5- g .j. M *E"w G i ac t m 'a ca .jin afdla uf'.g.

utso g

• i.

m p

g u

lag 3-20

• . j l

l'

)

8

.R B

i 8

f f

L B.

m C

2 Be

. .gS m

E i

8.w l

gE N

5 .

T i 8  ;

1 i u, i 8-

= l u g [

a T '

B .b "d fj U

N 8

8 m a ,

• 80 #8 g$'O X)3fS id Nf 5130 A e 3-21

_ _ _~ _-.. .. --_.__.-_ .._____. _. _ ______

l f

5

.g i

\

B i l i

8

. .g 5

. .g B

N fl m a

v. . z o

Sc

• U R o gw e

z h

B*W N '

I 1

r K5

-l l n- B N "N .

E B

.. . h . .g

=

E E

• g.g.

go i. os.i. go .g. go.s E 1 80 1 as'o co o- os a- e

( -01X)32S/13 NI A113972A w D

00 M

. fa.

3-22

TABLE 3-2 Basic FSTF Systen Parameters a speed of sound in steam 1500.0 ft/sec 3, water density 62.4 lba/ft 3 j 3, steam density 0.073 lbs/ft3 u, mean steam velocity 75.0 ft/sec nu characteristic water expulsion 17.5 ft/sec vent velocity h, effective mean depth of vent water 2.0 ft 1

columm ,

i f

,I A,/A p vent / pool area ratio 0.045 2

Ypg /(H-h ) pool pressure coefficient 3.82 lbf/ft/in a

V/A, drywell volume / main vent area ratio 292.78 ft t

K downconer loss factor 4.8 h standoff distance of jet deflector 1.5 ft plate 2

A dp jet deflector plate area 19.0 ft 2

,[ A dt drywell area 61.36 ft 2

A, main vent area 11.95 ft 2

A h

header area 17.72 ft 2

Ad downcomer area 3.01 ft

L dt drywell tank length 50.0 ft L, main vent length 37.32 ft L4 component 4 header length 2.0 ft L3 component 5 header length 5.0 ft L6 component 6 header length 8.0 ft Ng number of component I downconers 2 N2 number of component 2 downconers 2 N3 number of component 3 downconers 0 3-23

)

. I With the condensation source determined. UNCOM1 was rerun and predictions of ring header pressure at transducer P5901 were made and are shown plotted against measured pressure time histories in Figure 3-8.

Figure 3-9 shows the corresponding PSD comparison. As can be seen, the comparisons are most favorable, supporting that the model developed contains the essential physics required to predict pressure oscillations in Mark I steam vent systems.

3.3.2 Effect of Vacuum Breaker Actuation on Measured Vent Pressures l

Vacuum breaker actuation during a blowdown will modify the pressures measured in the vent system. This effect could possibly make the measured pressures and, hence, condensation source velocity nonconservative. The i l- possibility then would be that plant-unique analysis to determine vacuum

'(o 1

breaker forcing functions might be nonconservatjve.

L An estimate of the effect of mass flow through the vacuum breaker in FSTF was made by first determining the pressure difference between the ring header at the vacuum breaker location and wetwell airspace. This i t pressure difference was then used to drive the nonlinear vacuum breaker 1 1 valve model described in the Appendix. For this test, an idealization of l g

$ a GPE 18" valve with a 0.2 psi equivalent magnetic catch was used. The

angle of opening (as a function of time during the 40 seconds of forcing)

I was translated into the effective area available for flow passage, 2

and Ap =

1/2 p,u was used to compute the velocity through the valve. From this, the total net mass flow through the valve amounted to 3

80 f t passing from the wetwell airspace- to the header during the 40 seconds of condensation source simulation. This volume is less than 2 percent of the total vent drywell volume in the FSTF; and the effect of vacuum breaker valve actuation on measured vent system pressure is judged to be small.

3-24

Y 5

i 0' ;f 8

N-

2. 2- ,,jp,,3fg,3 2=

T.o. 2= 5= d= d= i=

4

" * " " -" H*n:':d !"oe 1h#90Y r M1 a. First ten se onds.

1

' : 1 1 1 4

d 4 _

4 I

l 4

l r ..

4 m g, lsi, g

g h.i, 7:

m y

= . . s m A d n

o

- c

- e

.s n 4 e t

d n

. o c

i i e S.

b 8

. 3 8j 8i 8i 84 e w 84 8 r u

g i

F l

il

I B

"i B

.g 8

. .g

't B.

=~ C

-~_____.--

. r 8

1 -

. .g 1 . . -

-- - E i

f =- --

- so 1 -

d$

!!' 4 %-- a.w l7 pt

. g.x. .

t r l.

g

u. _

_

.g .

0

% c 0

'I sJ I 8 m

• k 1 <

.;;; w C

1 ' B E I . r ,,,,i A

' " '8

1. 2#

co't ,G,I S J NIoni- ocg J on.j_

2001IldWtf go.j_ . .f.

• w l :3 to w

$4e 3-27 i

L-- - - - - - . _ _ _ _ _ _ _ _ - _ - - - - - - - - _ _ _ _ _ _ _ _ _ _ _ _ _ - _ - ~ - - - - - - - - _ _ _ _ . _ _ _

i 1

l 8 *

.,g

=

.1 8

j < ,g

.c ~,.

B

.g

.~.,,

,g

._. __ g

~ - - .

. .g

p -

n

.: m

- c z

w Bo u a

• w

.dP . , M s

~

- B.w dn a

.g~z d:l1 is j!. E .

'E E

. S --- g o

-- a  :

"R c e

U e

e

--- .c u

B 8

. . . n i co's oo's an's oo'o co i. co i- ooi- co i. no s E .

u

• G *I *S *J NI 300117dWW o '

to 1

.e4 B&e I l

i 3-28

i e ,,

c4 N

i i

i E.

l w

w.

e I

o I w U%*'k. e i i

~

ao I

o m

i

'4 co i o

! M i

i s  ;

8 i

o. I j o.m

.no e.se s.se

_ __ t _._ - 'L e.se s.on s.no dl v.se I.k m__. ..

e.no

e.no _ _ -

so.so si.se n't .m ~s.m i FRE00ENCY IN HZ ab.m sie Figure 3-9.M1. Predicted (solid curve) arid ineasured (dashed curve) PSD's at transducerorP590 1

l I

l 5

3 . .

4. PLANT-UNIQUE DIFFERENTIAL PRESSURE LOADS 1

4.1 Data Set Selection for Plant Unique Loads Three runs in the FSTF were noted to have significant chugging: runs M1, i M4 and M9. In order to assure that the most conservative data set was used to extract an unsteady condensation source velocity function, data from these runs were used to drive a vacuum breaker valve model (see the Appendix for I details of this model), using the pressure difference between P5901 in the header and P3009 in the wetwell airspace, with the hydrostatic head determined from FSTF run conditions. The data for runs where chugging was thought to be j most severe were provided for runs M1 (30-325 seconds), M4 (26-110 seconds) 1 ; and M9 (25-100 seconds). Pressure data (with the two signals subtracted, ai ll linear trends removed every five seconds and hydrostatic head added) were i

p obtained for the above time intervals. These data were then used to drive an 18" valve dynamic model for the internal GPE valve with and without a 0.2 psi equivalent magnetic catch and the A&M valve. The results of this study, showing the maximum impacts of the valve disc on the body and the seat of the

'I valve, are shown in Table 4-1.

i Il It may be seen that only during the time increment between 65.9-131.8 l seconds of run M1 is the body of the valve predicted to be struck by the valve  ;

Ii disc, even though comparable closing impacts occur in 30-65.9 seconds of run I

! M1 and 26-64.5 seconds of run M4 The maximum closing impact velocities

^

l during run M1 occurred at 75.4 seconds, while the maximum body impact occurred i

at 92.8 seconds. For these reasons, the time interval 65.9-105.9 seconds was chosen as the data set which would be bounding for all FSTF data. The l

condensation source velocity for this time interval has been shown previously in Figure 3-6.

1 i

4.2 Internal Vacuum Breaker Forcing Function l

! Analysis with the vent dynamic model have shown that for a given

! condensation source at the vent exits, ring header pressure fluctuations are 4-1

I 4

i <

! Table 4-1 Vacuum Breaker Actuation During Chugging i, i MAKIMUM CLOSING MAXIMUM OPENING

)

l ANGULAR VELOCITY ANGULAR VELOCITY l

1 (rad /sec) (rad /sec) j GPE GPE GPE GPE RUN TIME (sec) (no msg) (msg) A&M (no mag) (mag) A&M ,

i l

i i j

N M1 30.0- 65.9 23.59 20.41 4.98 no impact - - -

i 65.9-131.8 23.50 21.02 7.16 16.80 17.46 3.95

131.8-197.8 11.29 12.37 3.87 -

- no impact 197.8-263.8 4.00 4.18 0.56 - no impact 263.8-325.0 no opening -

- no impact 4

i M6 26.0- 64.5 27.49 22.51 5.39 no impact 64.5-110.0 14.19 16.22 4.42 --

no impact M9 25.0- 64.5 17,43 15.41 3.74 no impact 64 .5-100.0 19.21 11.95 3.93 no impact l

1 All results use actual FSTF pressures not predicted pressures. Valve characteristics are supplied in the Appendix.

8

._ , . . _ . ~ . ~ _

i most significantly influenced by the parameter drywell volume / vent area, whose "

values for Mark 1 plants with internal vacuum breakers are shown in Table 4-

2. Figure 4-1 shows the variation of ras pressure for several plants as a volume / vent area. A sensitivity study of other function of drywell characteristics of the plants (such as main vent length, for example) results as in small variations in predicted pressure levels. The drywell volume effect on the minimum differential pressure can also be illustrated by comparing the E

computed minimum pressure with that value obtained by reducing the minimum These results are shown pressure observed in FSTF by the drywell volume ratio.

the plants having internally mounted on Table 4-3. As seen in Table 4-2, vacuum breakers fall into three convenient groups (even though the drywell volume / vent area in Duane Arnold is large, results vary only slightly from those of Hope Creek and E. Fermi).

1 Internal vacuum breaker loadings were computed for Monticello, Hatch 2 and Since each of these plants had I Hope Creek for Groups 1, 2 and 3, respectively. '

the smallest drywell volume / vent area in their group, these loading functions will be conservative for the other plants in their respective group. Figure 4-2 gives the predicted internal vacuum breaker forcing function for Hatch 2 in t

l Group 2. The condensation source velocity used to drive the Hatch 2 dynamic a

L model was shown in Figure 3-6. Vacuum breaker forcing functions in graphical j and tabular form for Groups 1-3 have been computed. '

m 4.3 External Vacuum Breaker Forcing Function j

! Six of the Mark I plants have external vacuum breakers, using external A sensitivity study piping to connect the wetwell airspace with the main vent.

l I has shown that the lengths of the external vacuum breaker piping (especially on the main vent side) tend to strongly control the pressure across the external i

l vacuum breaker. The external piping lengths are such that predicted undamped Since the unsteady i

natural frequencies are in the 9-13 Hz frequency range.

condensation source velocity has energy in this f requency range, damping must be conservatively introduced to keep pressure levels realistic. A conservative estimate of the damping is introduced here which yields a prediction of

< conservative pressure forcing functions for the external lines.

4-3

Q TABLE 4-2 Mark I Drywell Volume / Total Vent Area Group Plant Value (ft)

FSTF 292.78

, Monticello 413.62 1 Pilgrim

' 455.91 Browns Ferry 490.70

, Quad Cities 491.40 o

lj; i.

h I rHatch 2 532.87 Cooper 542.25 2 ( Peach Bottom 549.75 Hatch 1 586.42 p $ Brunswick 591.03

\,

a d

w l:. Hope Creek I and 2 642.24 I

i 3 Enrico Fermi 2 i

658.06 I

Duane Arnold 775.11 l

4-4 i

0.8" A

e O m a 0.7 "

h g w o O N b o x a:

z 0.6- -

5  %

F w I

h

$ 0 z

O t

i Prms. Psi 0.5"

)

O.4- -

.i i

.l >

0. 3 " '

~ li; lIli

! j::

c- 0.2- -

0.1 "

0 0 500 400 doo 800 l Drywell volume / vent area, ft Figure 4-1. rms predicted pressure at the main vent / header junction as a function of drywell volume / vent i area.

l l

4-5

- 1

. l TABLE 4-3 Dryvell Volume Effect l Pressure Drywell Volume / A pain (psid) A p (psid)

Signal Vent Area V(ft)

FSTF 292.78 3.32 3.32 r

Group 1 413.62 2.30 2.35

![

i Group 2 532.87 2.07 1.82

. lj

. il - Group 3 642.24 2.05 1.51 l V FSTF

/ vent exit area lt[ AP " PFSTF Vp g / vent exit area I!

ic iI h

i a

i 4-6 i

l

n o

h"*

5 l*

a 5"

c.

[s 1**

o

• n l e

E d*

0 0 E ** _

h. " "

N.

5 o

c "

o $

f "r

" s

" d n

[.

m 5$

$E

"[

• s n

I n

• e

, / t 5

N" I!

EE NN 4 5 Ia MNe E"

g

s. *O i

Ed r E s s Ni s c.

'e

- s e

I e 0

0 =id n!

5 2 8 -

0 '

{

_?  ! 8i

= "

a

!, ,  ! L i:  !

1 t

a ,

g 1 i

g a

s s

i i-0 _

F h -

s.E y

lp  % $

-l-f .

Y ~ .

a

.a 8

L -

a h

i &

~

s 9 ,

' " " r ge, ,,=;i; n1 ,

='i . o . e g i- .4 ....e  ;.

60 we IAe l

4-8

i

\

l l

1 8 -

..g E

E I

. \

' \

8 l

. .a=

m ._

t E

. .g l

! h n E l

.w .

% se u 1

.Ew ll  % .

E u.s

.c.z.

I,.t = e-i ,

B .

n . m

• N e E

o B

d g

_M e i

M f 8 7*

e  :  :  :  : .

:  : co s E co s co o co t- co a- co s- co t- e

• n oo s co a 9

• 0*!*S*J NI 300117dWW :s 60 r4 la.

4 4-9

5

..g L .

B

. .g B.

.g g

. .g

. en

, 1. o z

Bo

• U o gw m

z d "

E5 l '

[ .

. B. .=

a

!, E g U

e e

. C B

"d s2 N

i g 4 i

% oo;s oc;a ao as t oo'i- co'i- ao'si- co'i- 00*sg *

• G'I }o;oS*J NI 3001!1dWW e e

ord ILe 4-10

. _ . ~ . _ - . _ - _ . - _ . - - __ _ - - . _ _ _ _ _ . . _ - _ _ - . _ _ _ _ _ - _ _ ,

5 i

.J I

The generic structure of FSTF is maintained by considering components 11 and 12 in Figure 3-2: piping connects the wetwell airspace to one side of the external vacuum breaker and connects the other side of the vacuum breaker to a position along the main vent close to the drywell. Since the lengths of the f

external piping are extremely important, each Mark I plant with external vacuum

! breakers must be considered on a plant-unique basis.

\

1 A damping mechanism is suggested by considering the flow of the fluid out of the external pipe into the stressing main vent flow, as shown in Figure 4-

3. Any fluid discharged into the main vent will be swept away by the mean 4

! steam velocity; consequently, kinetic energy losses occur at the external piping main vent junction. The dissipative force per unit area on the fluid in the external pipe is estimated to be h@

F= 3,u,lu,l (4 1) a

)

! This loss is introduced as a pressure drop at the junction of the external

, vacuum breaker piping with the main vent so that l'

ih 1  !).

OP = P(L) - Py =F (4-2)

Y,  !

p

where p(L) and p, are the unsteady pressures at the main vent on the piping i side and asin vent side of the junction, respectively. Note that this damping

) sechanism is nonlinear since the loss is proportional to the unsteady velocity squared. The solution for the unsteady pressure at the vacuum breaker p(0)

then involves the solution of a nonlinear algebraic equation, for which the pressure levels are dependent upon the loss factor K* . References 3 and 4 1 e

offer similar analyses of this probles and suggest that the value of K may I be of the order of 0.7 . An experimental program was undertaken to quantify l j K* and is described in Reference 5. This program confirmed that K* = 0.7

) l i results in underestination of losses and hence high predictions of pressure l s

1evels in the external lines.

l ]

! I I I 4-11  !

I

_.. _ .. _ _ _ _ _ _ _ .- ._-_ _ . _ _- _ ___ __ _ _ . _ ~ . _ _ . _ _ _ . . . _ . . . _ . _ . _ _ . _ . . . - -

" K 3

! moln vent

)

i 4 L 7

ll- ) -+ x u, Vacuum I - r-I g Breaker "P(0)

I {p(L): i op 1 l i V

_ L ..J e

l k:

tit-h y

D l-j" u, I

f2 Figure 4-3. Schematic of the external pipe end loss mechanism.

l l

4-12

I Using UNCOM1 and including the nonlinear damping abcVe yields the pressure prediction sho m in Figure 4-4 for Dresden. It may be seen that the pressure levels are of the order of those predicted for the Hatch 2 plant in Figure 4-2.

O 5

.i d.

!l!

i?

fI lio

+l ! "

r.

l i

1 1

4-13

_. _ _ _ _ _ . _ . _ _ . . _ _ _ _ _ _ . _ _ _ . _ _ . _ . . _ . _,. .__.___ - . . _ , _ . _ , _ . - . _ . . . _ _ , . . _ - _ _ _ . . _ ~ . _ _ . . - . .

=g. _

- _ - = .- =

J" 8

8 I

"8

z. e a'8 E+

j "s'

8

! i" 1

i 8 id.co

/.co /.co /.co =

,s. g .co i E.so d.oo d.co /s.oo Figure 4-4. Pressure time history predictions across the external vacuum breaker for Dresden, using a nonlinear end flow damping mechanism. The conditions of run M1 as described in Figure 3-6 apply with the hydrostatic head not added to the time history. a. First ten seconds. ,

B.

..E B

.f 8

.f

. . e-

~

. [%N 8 2

I m

! - e

.h

! z Bo *U cw m 2

ii

- 3.w R .

.n z v s. .

. .g o i t g u

e

=

E.

2 g u

o a 8 o

e us g -

s .

00*n co E co'r 4 o0 : oo'o co i. no.g. no g. no.g. g .g.E ,

,G,I*S*d NI 300117dW8 9 o

60

.e4 lae 4-15

B.

"R E.

"2 8

ng B

w --

w

• . .:N

?

^-

g

- "d e

m E

so d[w

% w i zs r

7

• i n B. -

O 8 8

u 8 *

$ c e

_w B. u G f a s e

oo s oo's oo;z a co t oo'o co*t- co a- co's- co t'-

00 s d 4 l e

' *G*I*S*d NI 300117dWW w E

lLe 4-16

l 1

E i

.i  !

8  !

a g I 1

8

. .g B

. .g

- ,. Y N E i

i

.g b M

- r: e p- z

.s Bou RWm Z

e g q: & W 1:1 m '

K3 i l

'l h i

' E .

- a m,; e ,

f 'ts ;

M b I l

-l

.C i

00*n 00*E 00*2

[--

00*t 00*C 00*t-a r*S*d NI 300H W B 00*E-00 *E- 00*n- 00*S R

= k

=

60 w

LL*

4-17 l

' 1

5. VACUUM BREAKER VALVE DYNAMIC HDDEL The valve dynamic model used to predict plant-unique vacuum breaker disc actuation velocity is documented in Reference 6. This model has been validated against full scale test data and has been shown to result in conservative predictions of vacuum breaker disc actuation velocity. The valve model requires as input valve unique mass and geometric characteristics as well as an unsteady plant-unique differential pressure forcing function across the valve disc. The dynamic model then predicts vacuum breaker disc actuation velocity which is used to assess adequacy of the installed design.

i i

r i l

I I

t l

l i

5-1 1

--- ,c.-- --- , - - - . - . . . - , , , , - - , . _ . . , , - . - . . . -,,.w.. -

, - - , , . . - - . , - - , . - . , - , , - , , - . ~ , , - . - - - . . , . .

a

)

6. CONCLUSIONS The Mark I vetwell to drywell vacuum breaker load definition is based on full scale test data corrected to plant-unique geometries using a validated analytic model. The approach taken has been to:

o Identify the FSTF test data which will give the most severe differential pressure loading across a vacuum breaker disc o Develop and validate a vent system dynamic model to correct this data for plant-unique geometry 1

i o Develop and validate a vacuum breaker dynamic model to predict l vacuum breaker disc actuation velocity o Predict plant-unique vacuum breaker disc actuation velocity using the plant-unique differential pressure load definition and the valve dynamic model o Use this actuation velocity to evaluate the adequacy of the installed design.

)

6-1

a

7. REFERENCES

1. " Full Scale Test Program Final Report," NEDE-24539-P Class III, April 1979.

2. " Analysis of Full Scale Test Facility for Condensation Oscillation i Loading," NEDE-24645-P Class III, July 1979.

3. Ingard, U. and Ising, H., " Acoustic Nonlinearity of an Orifice," J.

Acoust. Soc. of Amer. , Vol. 42, No.1, pp. 6-17,1967.

4. " Liquid Propellant Rocket Ccabustion Instability" (D. Harrje, ed.), NASA SP-194,1972, 399-405.

5. " Mark I Experimental Determination of External Lines Losses for Definition of External Vacuum Breaker Loads," Continuum Dynamics, Inc. Report No. 81-

g 2, July 1981.

> h; j !! 6. Sullivan, J.M., " Mark I Vacuum Breaker Improved Valve Dynamic Model,"

j!.  ! C.D.I. Tech. Note 82-31, August 1982.

jU l iii

.m I

o S

I i

I 7-1 i

e l

APPENDIX Vacuum Breaker Dynamic Model Used to Select FSTF Chugging Data for Differential Pressure Load Definition A vacuum breaker valve, shown schematically in Figure A-1, typically consists of a circular valve disc kept in a closed position by gravity and the overpressure present in the main vent or header during a postulated LOCA. The disc is attached to an arm that freely pivots about a shaft above it. In some i

valves, a counter-weight is also attached to the arm. The actuation of the

valve obeys the following dynamic equation:

• T +T gravity gravity where I

gravity " -b Gmg sin 0

't T

pressure"~bD AA cos((0-Og ) /(0,,,-Og ))

i where I = moment of inertia of the valve system about the pin La = moment arm from the shaft pin to the center of gravity of the entire valve system Lp = moment arm from the centerline of the shaft to the center of the disc 1

m = total mass of the entire system l

A-1 l

l i

D i

I i =

n m

E

~

/s G

((+)L /

a s C.G.

j 'G j.

li 4 Wetwell side Main vent or 9 9l header side I

k a i l

E

"; I l '

l i.

l l

l-Figure A-1. Schematic of a vacuum breaker valve.

4 A-2 l

4 g = gravitational constant A = area of the circular disc AP = applied differential pressure as a function of time, used to drive the valve model (positive P tends to close the valve) 0 = angle measured from the vertical, giving the location of the system center of gravity as a function of time Q = rest angle of the system center of gravity O = maximum valve opening angle

, , max

! The characteristics of the physical valve under consideration dictate all i

of the system characteristics; while the predicted (or measured) pressure forcing function AP(t) ,

af ter accounting for hydrostatic head effects, y

completes the inf ormation needed to provide a prediction of O(t) .

O is greater than Og , the valve has opened. Impacts upon When 0 j closing or upon full-opening are handled by assuming coefficients of The valve model uses a modified predictor-corrector scheme to restitution.

solve for 0 as a function of time. The only additional feature in the case of simulating the GPE valve is to add an additional torsion ters of the form T

magnet

= -L AAP mag {0mag - (0-0 G)}/0 mag ,

7 0-0G<0mag l

=0 0-Og > 0,,

l

(

A-3

b where 1

mag = equivalent positive pressure due to the presence of a magnetic catch O = effective angle over which the magnetic force is felt by the mag disc.

i The properties used in the computations discussed in the asin text are shown in Table A-1.

i llel I

o1

..i.

g.l a

A-4

=-.

Q' a Table A-1 Vacuum Breaker Characteristics Am 18" CPE 18" A m 18" INTERNAL INTERNAL EXTERNAL I, system moment of inertia (Ib-in-s 2 ) -

54.57 20.*38 23.23 Q , system annent are (in) 1.60 10.71 4.20 E L D, disc moment are (in) 11.38 11.47 11.38 m, systen esas (Ib) 154.8 50.9 124.6 A, disc area (in 2) 261.59 375.82 261.59

! Og , rest angle (red) 1.27 0 0.49 0 ,,, maximum opening angle (rad) 2.06 1.32 1.27 Seat Coefficient of Restitution 0.6 0.6 0.6

~

i