ML20040D651

Nonproprietary Version of OPS-37A36, Verification of CLASIX-3 Computer Program.

ML20040D651
Person / Time
Site:	Grand Gulf
Issue date:	12/31/1981
From:	Haga P, Mariner R, Stumpf H OFFSHORE POWER SYSTEMS (SUBS. OF WESTINGHOUSE ELECTRI
To:
Shared Package
ML19297F290	List: ML20040D647 ML20040D648 ML20040D651
References
NUDOCS 8202020121
Download: ML20040D651 (22)

Text

,i Offshore PowerSystems I

O I '

I .

I I

E I

I VERIFICATION OF THE CLASIX-3

! COMPUTER PROGRAM I

I I

REPORT NO. OPS-37A36 DECEMBER,1981 g$0A$a$!*o$$$oSg

l _

L

I I

I VERIFICATICN OF ' HIE C[ASIX-3 I

bY Dr. II. J. Sttrnpf R. M. Mariner I ^,

/

],1 < >

Approved v - ^ - g P. B. Haga, Dipp'r Power Systens Technology I

I Doctznent No. OPS-37A36 I .

Offshore Power Systens I 8000 Arlington Expressway Jacksonville, FL 32211 I

I - 1ee I

I

il TABLE OF CENTENTS

Page ABSTRACP ii I INTIOIJJCTIQ1 1 II CDMPARIEN WITH CIASIX RESULTS 2 III CDMPARIKN WITH EXTERNAL CALCUIATIQ1S 3 IV SUPPRESSICN IOOL FDDEL VERIFICATIQ4 3 A. Stability 4

) B. Fluid Drainirx3 Fran a Tank 4 C. Simple bhnaneter 5 D. Two Different Size Tanks Connected By a Pipe 5 E. Suppression Pool Response to a bhin Steam I F.

Line Break Suppression Pool Response to a Recirculation 5

Line Break 6 G. Suppression Pool Response to a Wetwell Pressure Transient 7 H. Miscellaneous 01ecks fran CIASIX-3 Canputer Runs 7

. V REFERHJCES 12

,I I

I I

I I i

I

~

l ABSTRPCP i

I The CIASIX-3 conputer program is a generalized version of the CIASIX corputer program for the analysis of reactor plant containment response to y

hydrogen release and deflagraticn. In addition to the capabilities of CIASIX, CASIX-3 has the capability to nodel nore conpartments, flow paths, and passive heat sinks and has the feature of allowing arbitrary selection of flow path locations. New features incorporated into CLASIX-3 are representations of the supprersicn pool of che General Electric Mark III Pressure Suppression Containment System, and spray carryover between cmpartments. The CIASIX-3 canputer program has been verified by canpari-g sons of calculated results to the calculated results of the QASIX program 5 and by conparisons of calculateri results to the results of external calculations. The suppression pool nodel in CASIX-3 was verified by using tiu subroutine to model systcms of various degrees of couplexity subjected to transient boundary conditions and cnaparing the results with available analytical solutions. CIASIX-3 calculated results were also checked to determine if unss was conserved in the suppression pool and if static head in feet was equivalent to the pressure difference between the wetwell and drywell during quiescent parts of the transients. Extensive verification of I the CIASIX-3 program demonstrates that it provides realistic, but conser-vative, predictions of t m peratures and pressures resulting fran a hydrogen deflagraticn. Verificaticn of the suppressicn pool nodel provides a high I degree of confidence that it correctly predicts the dynamic behavior of the suppressicn pool of the General Electric Mark III Pressure Suppression

~

Contairrnent.

I

'l lI ii I

~

I i

I. INTRODUCTIOt1 1 The C[ASIX-3 canputer Irogram is a generalized version of the CIASIX I conputer program, Reference 1, for the analysis of reactor plant containment response to hydrogen release and deflagration. CLASIX-3 l- has all the capabilities of CIASIX except the ability to represent a special flow path for the containment vent option. In addition, CIASIX-3 has the capability to represent more conpartnents, more flow paths, and nore passive heat sinks than CIASIX and allows arbitrary selecticn of flow path locations. New features incorporated in the CLASIX-3 program include representations of the suppression rool of the General Electric Mark III Pressure Suppression Containment g System, described in Reference 2, and spray carryover between 5 conpartments.

The CIASIX conpiter program has been verified by extensive analyses includire conparisons of calculated results to the calculated results of other programs, crmparisons of calculated results to test measured results, conparisons of calculated results to results of external calculations using the CIASIX program methodology, and ntanerous sensitivity studies. The results of these conparisons provide a high I level of confidence in the validity and conservatism of C[ASIX analytical results.

The CLASIX-3 cnnputer program has been verified by conparisons of calculated results to the calculated results of the CIASIX program and by conparisons of calculated results to the results of external calculations. Ccriparisons of CIASIX-3 calculated results to CLASIX calculated results provide confidence in the CLASIX-3 calculations of multi-conpartment pressum and tanperature responses to high enthalpy water and hydrogen addition, burn phenonena, convective and radiant heat transfer to passive heat sinks, ard heat transfer to sprays.

Carparisons of CLASIX-3 calculated results to results of external calculations provide confidence in the CIASIX-3 calculations of spray I

I I

carryover, interconpartmental fan flow, cunpartment voltrne changes  !

dte to water level changes, and all aspects of the suppressicn pool.

I (DMPARITN WITil CLASIX RESULTS II.

CLASIX-3 results were crrnpared to CIASIX results for four cases. '"he j first case, a multi-crrnpartment model with high enthalpy water addition, is one of the cases used for conparison of CLASIX calcu-lated results to 'IMD (Reference 3) calculated results discussed in Appendix A of Reference 1. 'Ihe second case consisted of a one element model with high enthalpy water and hydrogen addition, convective and radiant heat transfer to passive heat sinks, and a hydrogen burn. This is one of the cases usa 3 for cnnparison of I CLASIX calculated results to CIXDCLASS9 (Reference 4) calculated results discussed in Appendix B of Reference 1. The third case consisted of a one elanent nalel with convective and radiant heat transfer to passive heat sinks arri a hydrogen burn. This is one of the cases used for cunparison of CLASIX calculated results to test measured results discussed in Appendix C of Reference 1. The last case consisted of a cne elanent nodel with high enthalpy water and hydrogen additicn and heat transfer to sprays. The model for this case is similar to the nodel for the second case.

I The results of these four conparisons showed excellent agreement between CIASIX-3 calculated values and CLASIX calculated values. In the first case, the multi-campartment nodel, CLASIX-3 calculated values differed slightly frcm CIASIX calculated values but were still conservative relative to 'IMD calculated values. The pressure and tanperature transients for the CIASIX-3, CLASIX and 'IND calculations are given in Figures 1 through 3. The small differences between CLASIX-3 and CLASIX calculated values for this case are attributed to I two phencmena. The first is a small flow oscillaticn which occurs in the CLASIX calculation and does not occur in the CLASIX-3 calcula-ticn. The secrni is that there are differences in the methodology of the analysis necessitated by the change fran the fixed flow path connecticn matrix of CIASIX to the flexible connecticn matrix of

-2~ .

l I

~

I I GASIX-3. In the three single elment emparisons, GASIX-3 and j CIASIX producM essentially identical results.

III. CDMPARISCN WITH EXTERNAL CAICUIATICNS I

CIASIX-3 calculated results were cmpared to the results of external calculations for spray carryover, intercanpartmental f an flow, and conpartment volume charges due to water level changes. (Suppression pool verification is discussed in the following section.) In all these conparisons, CIASIX-3 calculated results were essentially I identical to those generated by external calculations.

these conpariscn results are given in Table 1.

Examples of I IV. SUPPRESSICN POOL K) DEL VERIFICATICN One of the new features incorporated into CIASIX-3 is the ability to simulate the dynamic response of the IMR MARK III Pressure Suppres-sion Pool to pressure transients in the drywell and wetwell. The redel in GASIX-3 is based upon the equations given in Reference 2.

These equations were redified, however, so that reverse flow in the I vents (fran wetwell to drywell) and vent clearing fran wetwell to drywell could be simulated. The pressures in the drywell and wetwell, calculated in CASIX-3, provide the dynamic forces which change the fluid levels in the suppression pool.

It is not practical to verify the suppression pol model by running test cases usire the CIASIX-3 program because of the large a:rount of conputer time that would be required. Instead, the subroutine which models the suppressicn pool was run as an independent program.

Although the feedback effect between wetwell and drywell pressures ard the fluid levels in the suppressicn pool cannot then be simulated

-I as is cbne in CASIX-3, by using the pressure histories as input the required transient runs can be closely duplicatM. The methcds used to verify the suppression pol Itodel are as follows:

I s- .

)

c .

-.s I

s s

x y i) Utilize the subroutine to no3e1 systems of various degrees of cmplexity, subject thun to transient inundary ccnditions and cmpare the resultc with available analytical s51utions.

s I ii) Utilize the subroutine to model the BR MARK III Contairment Suppression Pool, subject the nodel to boundary corditions typical of a main steam line or recirculaticn line break and canpare the dynamic response to available analytical results.

I lii) s Utilize CIASIX-3 results'to determine if russ is conserved in the suppression po51 and if wetw11 to drywell pressure differences durf.ng quiescent parts of the transient are s

consistent with the differences in 43r 13 quit 14vels in the I wetwen ~and drysull, s

~,. ,

s, , -

Experimental results that rea available did not provide sufficient daca concerning the experiment tretus for a cunparison - to be pode with the sqpreession 7.cl ..nel. A' brief' description' of each verification amlysis J and tbe resulte are given in the fo nowing sectior.s. _ ~

A. , stabi g ,,

In arder to check t e stability of t$e suppression poo! rrodel as used in CIASIX--3, tJ.e initial fluic level ard pressure in the wetwell. were set ( e'Iual to those in the drywell and all' ve-I locities were set to zero.

tien ( ~ 10 feet)- resulted A, -stable, anall amplitude oscilla '

' which is ettributed to roundoff

.s errors in the integraticn schdiard the- close couplirg between fluid accelerations and fluid levels. 7 u s B. 4 Flyid Dr'aining fran a Tank s

<The subrout.ine was used to rrodel a tank of water with cne N

outlet., or three outlets at different levels, through which water is' draining fran the tank. 'Ihe liquid level in the tank x .

I 5 g\

g s

% ., h i,

I ,

\

, and the velocity of the fluid at the outlets as a function of 7

' time as calculated by the subroutine were crsnpared with ana-r lytical solutions. The . otznparisons for all cases run wre I excellent. The results for a tank with three outlets are shtun in Figures 4 ar!d 5.

l C. Simple Manczneter Another system nodeled in the subroutine was that of a simple mananeter. Two cases were considered. The first case c:2npared the period of free oscillations of the manoneter fluid as calculated in the program with analytical results. The seccnd I

case crznpared the results of the subroutine with analytical results for the displacement of the liquid level as a function of time and the period of the oscillations for the case in which the mananeter fluid is subjected to an instantaneous pressure difference. Ebr both cases the crznparisons wre excellent. The results for the case of the mananeter fluid subjected to an instantaneous pressure difference are shown in Figure 6.

I D. Two Different Size Tanks Connected by a Pipe I For this ocmparison the subroutine was uced to model two different size tanks connected by a anall pipe. Since the equations formulated in Reference 2 neglect the fluid inertia in the wetwell, a conpariscn between the results fran the sub-routine and analytical results was made for the period of free oscillations of the fluid Vc.en the fluid in the larger tank or smaller tank is neglected. Tae periods of the oscillations conputed in the subroutine differed with the analytical results J by less than 0.4%

l E. Suppression Pool Response to a Main Steam Line Break i In runnirg the subroutine as an independent program it is not feasible to utilize nuss and mergy addition as the driving i lI

s force for the main steam line break transient and crrapare the pressure and fluid level histories frcm Reference 5 with the results of the subroutine. The metled used to run this tran-i sient was to model the suppressicn pool described in Reference 5 and use the wetwell and drywell pressure histories given in the l reference as input to the subroutine. A canpariscn was then made for vent clearing and vent recovery times. The ccriparison is slown in Table 2. The subroutine entry for the top vent l recovery time is rot given since the water level oscillates about the top vents after the drywell fills ard covers and uncovers the top vents several times before finally remaining above the vents. Sane of the difference between the results of Table 2 can be attributed to the following:

f i) Pressure histories prior to 0.1 seconds are not given in Reference 5 and were estimated by extrapolating the pressure curves.

I 11) Detailed geometry of the suppression pool model in I Reference 5 is not given. Data in Reference 6 was utilized to estimate the required parameters.

'I lii) loss coefficients were not given in Reference 5. Data frcm Reference 6 was used to estimate the loss coeffi-cients ard the same value was used for all three rows of vents.

It is felt that in view of the uncertainties listed above tha I ccmpariscn between the results of the suppression pool sub-routine and Reference 5 is quite good.

I

+

I

I F. Suppression Pool Response to a Recirculation Line Break I This transient, run in the same manner as the nnin steam line break transient, utilized the data in Reference 7. The cxn-I parison for vent clearing and vent recovery times is shown in Table 3. No entry is givcn for the top vent recovery time for the reason outlined in Section III-E above. Again, it is felt that in view of the uncertainties listed in Section III-E, the ocznparison is quite cpod.

G. Suppression Pool Response to a Wetwell Pressure Transient I In order to test the logic in the suppression pool subroutine for vent uncoverirg and clearirg fran wetwell to drywell and subsequent refilling and recovery, a transient was run in which the wetwell pressure was increased linearly by approximately 5 psi in 5 seconds, held constant at the elevated pressure for 5 secords ard thcn decreased linearly to the original pressure in 5 seconds. The prescure in the drywell rmained constant throughout the transient. Since no data is available for canparison the primary method of verifying the accuracy of the results is to check that continuity is satisfied and that the

.I vents begin to clear or fill according to the criteria specified in the subroutine. The continuity relations were checked at various times during the transiant and with one exception the 9

relations were satisfied to within a few parts in 10 . The exception occurs during the latter stages of vent clearing or filling where the continuity relaticn between the movirg surface in the drywell, flow in the top vents and drywell fluid velocity belcw the top vent is not satisfied. Since vent clearing or I

filling represents a small partion of the transient and the continuity relations must be re-initialized at the end of vent clearing or filling, the effect of this discrepancy in cne of

the continuity relations en the overall transient will not be l significant. The criteria used to initiate vent clearing or fillirg were satisfied throughout the transient.

~

H. Miscellaneous Checks frcm CIASIX-3 Canputer Runs I A number of CLASIX-3 production runs have been made for a IMR Mark III Pressure Suppressicn Contairrnent Systen. Scme of the I. output grameters fran these runs were checked against hand calculations to gain confidence that the results were correct.

One such check involved a crr:parison of suppression pool static head with the pressure differential betweet the wetwell and drywell cmpartments during quiescent parts of the transient.

The static head, in feet, was converted to a pressure differ-I ential which was then subtracted from the drywell/wetwell pressure differential. In all cases, the difference was less than a tenth of a psi. This check indicates that the links -

between the main program, where the cmpartment pressures are calculated, and the suppression pool rnodel are correct.

Checks were made to ensure that the suppression pool mass was being conserved. Over parts of a transient, a kroan mass of water entered the suppressicn pool. The suppression pool mass was checked before and after this mass addition to ensure that a

.I charge of only the anount added occurred. Also, over an entire transient, the annunt of water and steam added or condensed in

I the suppressicn pool is known. This added mass was ccmpared to the difference in pol mass frcm the beginning of the transient to the end. In all cases, the suppression pool mass is conserved.

I The efforts described above to verify the IER MTGK III (bn-tainment Sappressicx1 Pool nodel in CIASIX-3 provide a high degree of confidence that the nodel correc.tly predicts the dynamic behavior of the suppressicn pool to a variety of I transients. The nodel has correctiv predicted the behavior of systens rargirg in ccmplexity frcm that of a tank fran which water is draining to a BWR mrk III (bntainment Suppression Pool. No significant differences between the nodel results and available analytical solutions have been uncovered.

^

I

TABLE 1 QASIX-3 vs EXTERIAL CAIfUIATIONS 1I TYPICAL CDMPARIS21S I DCrERTAILY CLASIX-3 CAIEUIATED CALCUIATED PARAMETER VALUE VALUE Fan flow rate (cfs) 19.58 19.58 Spray carryover flow rate (lbn/sec) 425.55 425.55 Drywell voltrne (ft ) 2.342 x 10 5 2.343 x 10 5 Wetwell volune (ft3 ) 1.432 x 10 5 1.431 x 10 5 I

I I

I I

I I

I I

I I >

I __ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ - _ _ _ _

I TABLE 2 I CDMPARI9]ti OF VDir CLEARING AND RECDVERY TIMES BETWEDI SUPPRESSIQ1 100L FDDEL AND REFERHICE 5 MAIN STEAM LINE BPEAK VHTr C EARING TIMS VENT RECDVERY TIME

• (SEC.) (SEC.)

W TF NO. bDDEL REF. 5 !ODEL REF. 5 1 ('IOP) 1.04 0.92 -

80.0 2 (MIDDLE) 1.28 1.17 44.0 42.0 3 (B7I'IOM) 1.76 1.52 37.7 34.0 I

I I

I I

I I

I I

I I

I

_ - 1

~

.I TABLE 3 l

l CDMPARIK11 OF VENT CLEARING AND RECDVERY TIMES BETWEDI i SUPPRESSIO1 EOOL bODEL AND REFERDICE 7 RECIRCULATIOJ LINE BREAK I VENT NO.

VENP CTARING TIME FDDEL (sEc.)

REF. 7 VINT RECDVERY TIME FDDEL (src.)

REF. 7 1 (TOP) 1.04 0.94 -

99.0 2 (MIDDLE) 1.29 1.15 33.5 44.0

.I 3 (BOI'IOM) 1.85 1.60 28.3 29.0 I

I I

I I

I I

I I

I I

I I _ _ - _ - _ _ _ _ _ _ _ _ _ - _

.I I V. REFERENCES

1. '*I'he CIASIX Canputer Program for the Analysis of Reactor Plant Contaiment Response to Hydrcgen Release and Deflagration", OPS-07A35 (Proprietary Class 2), October 1981, OPS-36A31 (Proprietary Class 3),

October 1981.

2. The General Electric Mark III Pressure Suppression Cbntainment Systen Analytical Mxlel, W. J. Bilanin, NEED 20533, June 1974.

3. " Ice Condenser Cbntainment Pressure Transient Inalysis Meihxis",

WCAP-8077 (Proprietary Class 2), March 1973, NCAP-8078 (Proprietary Class 3), March 1973.

4. " Zion Probabilistic Safety Study", Mxiule 4, Section 4, 1981 (NRC Docket Nos. 50-295 and 50-304) .

5. Safety Evaluation Report Related to the Preliminary Design of the I GE-SAR-238 Nuclear Islard Standard Design, General Electric (bnpany, Docket Number SIN. 50-447, PB-248 042, Decenber 1975.

6. Vent Clearing Analysis of a Mark III Pressure Suppression Contain-I ment, Rafael Quintana, Nuclear Engineering and Design 55 (1979), p 395-401.

I 7. Mississippi Powr and Light (bnpany Grand Gulf Nuclear Station Units 1 & 2 Final Safety Analysis Report, Section 6.2.

I I

I I

I I

I I

I l

l 8,C

!I i

lI i

!I lI

\E 1

!I 4

I i

!I i

i

!I 1

lE -

!E CLAS1X TMD COMPARISON i 10,000 LBM/SEC BLONDOHN

1205 BTU /LBM

]l NO ICE

. FIGURE 1

!I r

4 il

I

~

a,c I

I ,

I I

I I

I

~

I I

I I

I

~

CLASIX TMD COMPARISON 10,000 LBM/SEC BLONDOWN lI 1205 BTU /LBM NO ICE I FIGURE 2  !

I l

,I

I ,

j a,c I

I I

I I

I I

I I

I I

~

l CLASIX TMD COMPARISON 10,000 LBM/SEC BLONDOWN 1205 BTU /LBM g NO ICE FIGURE 3 I

I I

18 16 -

g14 -

E

$ 12 -

5 t;j 10 - ~

aJ LEGEND g

w B E THEORETICAL RESULTS

' * . SUBROUTINE SUPOOL 36 -

e 6m 4 -

li$

O 2 -

0 80 90 100 O 10 20 30 40 50 60 70 TIME (SECONDS) .

' CHANGE IN HATER LEVEL IN TANK AS A FUNCTION OF TIME AFTER -

START OF TRANSIENT - THREE OUTLETS FIGURE 4

m m m M M M M M M m m m m m m i

LEGEND THEORETICAL RESULTS Ed8 ROUTINE SUPOOL 14 1

12;\

\

E

( 10 -

z@ s,\ 'N O8 -

g s

>. a s s b5 - \ '\

g a. 6 de

\

s '\, BOTTOM VENT i N

>g s N

' l C~ 4 -

t

's 5 I

• l g MIODLE VENT '

s 2 _

[ TOP VENT i \,

0 ' l ' ' I ' ' ' ' ' '

O 10 20 30 40 50 60 70 80 90 100

, TIME MFTER START OF TRANSIENT (SECONDS)

EXIT VELOCITIES AS A FUNCTION OF TIME FROM A TANK HITH THREE OUTLETS -

FIGURE 5

m M M M M M M M M M M LEGEND ANALYTICAL RESULTS

. SUBROUTINE SUPOOL

,2.0 , .

U g . <

5 0

g 1.5 -

z E

E

• n S' ~

U l'O ~~ ~ ~

STATIC EQUILIBRIUM LEVEL

~ ~ ~

5 E

E

v. 5 y 0.5 -

E g =

T = 5.90 SEC.

3 k

5 ' ' ' ' '

0 O 1 2 3 4 5 6 TIME (SECONDS) -

DISPLACEMENT OF WATER LEVEL IN A SIMPLE MANOMETER IN A FUNCTION OF TIME -

-INSTANTANEOUS PRESSURE DIFFERENCE-FIGURE 6