ML20132B280

Evaluation of CLASIX-3 Conservatisms & Quarter-Scale Tests

ML20132B280
Person / Time
Site:	River Bend
Issue date:	08/31/1985
From:	GULF STATES UTILITIES CO.
To:
Shared Package
ML20132B272	List: ML20132B280 RBG-22010, Forwards Evaluation of CLASIX-3 Conservatisms & Quarter-Scale Tests
References
NUDOCS 8509260147
Download: ML20132B280 (41)

Text

..

A A EVALUATION OF-CLASIX-3 CONSERVATISMS AND QUARTER SCALE TESTS i

River Berd Station - Unit 1 August, 1985 I

i-t i

1

}

f i

! 5 8509260147 850904 PDR .ADOCK 05000458 PDR-

' P l

'l A v

1.0 Introduction Following the 'IMI-2 accident, the NRC staff expressed concern that the

. ice condenser and Mark III containment designs might be vulnerable to i overpressures produced by hydrogen deflagrations. Consequently the CLASIX cmputer code was developed to provide a conservative estimate of the hydrogen deflagration pressure. The BhR version of this code, CLASIX-3, contained the same conservatisms as the original CLASIX code.

%e fact that the CLASIX-3 code was designed to provide a conservative prediction of deflagration pressures in turn resulted in a conservative prediction of taperatures resulting frm deflagrations. Recent testing performed by the Hydrogen Control Owners Group (HOOG) at the Quarter Scale Test Facility (OSTF) shows that the pressures and tmperatures predicted by CLASIX-3 are overconservative.

%e RBS survivability analysis ccmpleted to date, based on the CLASIX-3 analysis, indicates that all equipnent located in the lower intermediate volume just above the HCU floor will survive the thermal enviroment present in this region. We survivability analysis capleted for hydrogen igniters and igniter power supply cable in the wetwell region indicates that survivability of these cmponents is not assured for the "i entire 75% LWR transient thermal environment as predicted by CLASIX-3.

We thermal environments predicted by CIASIX-3 are excessively conservative for assessing the ability of equipnent in the River Bend Station to survive hydrogen cmbustion. As indicated above, a nunber of conservative assumptions have been incorporated in the CLASIX-3 analysis to assure conservative predictions of contaiment pressure. In addition, based on 1/4 scale testing, the serial deflagrations predicted

by CLASIX-3 constitute a significantly more severe thermal environment than the actual thermal environment which would be expected to occur in a Mark III containment.

The following report provides an evaluation of the conservatisms present in CLASIX-3, a discussion of the cmbustion phenmena observed in 1/4 scale testing and an evaluation of the ability of equipmnt to survive the thermal enviroment expected to occur in a Mark III containment.

1

2 a, i

2.0 Evaluation of Conservatism in CIASIX-3 Assumptions We CLASIX-3 code assumes that each cmpartment modeled by the code is instantaneously cmpletely mixed. Thus embustion in a cmpartment cannot occur until sufficient hydrogen has been injectal into the~

cmpartment to bring the hydrogen concentration throughout the entire volume up to the hydrogen concentration which has been specified as the concentration rcquired to support cmbustion. The CIASIX-3 code allows cmbustion in a volume, e.g. the wetwell, when the average concentration reaches 8%. Upon ignition, the volume is swept out by a flame front assumed to travel at 6 ft/sec. This burn is assumed to go to 85%

cmpletion. %e CIASIX-3 code accounts for heat losses to walls and other surfaces. ' he heat losses during burning are directly affected by flame speed since the flame speed will determine the time to lose heat during flame propagation. he degree of conservatism resulting fr m this cmbination of assumptions can be quantified by assessing the deflagration pressures and tmperatures expected to occur in a Mark III containment.

Two general types of release histories have been injected into the 1/4 scale test facility to date. One (case c',150 gpn reflood) begins with a quickly increasing hydrogen flow rate which should result in a large vertical hydrogen concentration gradient in the wetwell. W e other (case B, 5000 gpn reflood) injects hydrogen for a relatively long period at a low rate before a large spike in hydrogen flow is introduced. Wis history, at least prior to the spike, should be representative of the minimum vertical hydrogen correntration in the wetwell. A total of 21 scoping tests have been performed with such histories and in no case did the initial lightoff deflagration (only deflagration observed in any test) result in pressures or tmperatures approaching those calculated for the full scale plant using CIASIX-3 for the same (scaled up) release histories. Figures 2-1 and 2-2 provide a cmparison of pressures predicted by CIASIX-3 and 1/4 scale pressures respectively. Figures 2-3 and 2-4 provide a cmparison of taperatures calculated by CIASIX-3 with the 1/4 scale tmperatures. Based on the above, the cmbination of assumptions used in the CIASIX-3 codo yield pressures and tmperatures well above those which actually are expected to occur in Mark III containment due to deflagrations.

Frm the above discussion, it can be concluded that CIASIX-3 severely over-predicts both the expected full scale tmperatures and pressures.

The following discussions identify key CIASIX-3 assumptions which may produce this over conservatism. In the CIASIX-3 code, cmbustion is assumed to be initiated in a volume when the hydrogen concentration by volume reaches 8%. This represents an upper bound on the hydrogen concentration at which deflagrations would be initiated by igniters. A large nunber of tests including the recently capleted Nevada Test Site Tests d monstrate that mixtures with hydrogen concentrations as low as 5.8% can be ignited. It is cmpletely reasonable, based upon tests cmpleted by Acurex for EPRI, tests empleted by Fenwal Laboratories for Westinghouse, and tests empleted by Whitoshell Iaboratories for EPRI, l

{

1 1 r, l l

to conclude that mixtures with volumetric hydrogen concentrations of 6%

will be reliably ignited in the Mark III containment.

Another CLASIX-3 assumption which may result in over-prediction of tmperatures and pressures is conbustion empleteness. hhen mixtures with lower volumetric hydrogen concentration are ignited, less of the hydrogen present in the mixture is burned. For example, cmbustion of mixtures with a hydrogen concentration of 6% by volume will consunn only about 65% of the hydrogen present in the mixture even when very high levels of turbulence are present. The carbination of initiating carbustion at lowr hydrogen concentrations than assumed in the analysis of wetwell hydrogen cmbustion and burning less of the hydrogen would result in a considerable reduction in peak tarperature for wetwell burns. Since the peak tarperature has a significant effect on radiant heat transfer in equipment survivability analysis, use of higher hydrogen concentrations to initiate carbustion and higher burnup fractions results in considerable conservatism in thermal environment definition.

The burn duration assumed in CLASIX-3 will have a direct affect on peak pressures and tanperatures. S e rate at which energy is added to a volume by hydrogen cmbustion in the CIASIX-3 carputer code is controlled by a burn duration time input for each volume treated by the code. The burn durations used to date in GSU's CIASIX-3 analysis are based on an average flame propagation speed of 6 feet per second. This is a conservative basis for defining canbustion duration for the River Bend Station. Flame speeds for canbustion propagation decrease significantly when carbustion is initiated at lower volumetric hydrogen concentrations such as 6% hydrogen concentration mixtures. In addition, flame speed is related to the turbulence levels present in the containment. Since the River Bend Station does not utilize containment sprays to provide bulk containment heat raroval, but rather uses safety grade containment unit coolers, the relative turbulence levels in the River Bend Station containment should be significantly lower than the turbulence levels present in other Mark III containment plants. Imer flame speeds would result in greater burn durations. This would result in more time for pressure equalization, more uniform mixing of the containment air spaces and reduced tmperatures due to dilution by the entire containment volume. In addition, a longer burn duration will result in a lower heat addition rate to the containment which will allow more time for heat removal by the RBS containment heat sinks and unit coolers.

The methodology used to calculate heat transfer frm a carpartment atmosphere to cmpartrient heat sinks is extrmely conservative. At the NBC staff's suggestion, heat transfer correlations uso'l to calculate envirorrnental conditions following design basis accidents have been used to calculate heat transfer to containment passive heat sinks. This methodology is described in detail in NUREG 0588. 'Ihe conservative character of these heat transfer correlations is intended to provide adequate margins in defining the thermal environments produced after a l

'l s _

design basis accident. 'Ihese - conservatisms are not awwg.iate for definition of thermal envirorments following degraded core accidents since the same levels ~ of margins are not warranted for less probable l recoverable degraded core accidents.-

'Ihe above CLASIX-3 conservatisms when coupled with the conservative modeling of the River Bend containnent heat sinks and conservative

, modeling of the containment unit coolers result in a thermal envirornent i significantly more severe than that which would be expected at full scale.

l f

j I

l l

I i

J 4

d

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

Am..

h FIGUPI 2-1 GSU/ RIVER WETWELL BEND 5 VOL. AUG.85 PRESSURE 35 -

30 -

25 -

W m- ) )LLULJL,dtAdduMJJijt JUUJJJL d U t/

15 --

f 10 -

5-0 4 8 12 16 20 24 28 TlWE (1000 SECONOS) 4 i

, - - , - - ~

r l 36.6N i

/\

/%

4 H

0)

G.

m 3 /(

,4

! G. /

14.0 #

! GN:43:19.23 TIME (SEC ) -> Mi:13:19.23 EPRIf1RST SCTIE TEST RR CASE B REl. EASE, TEST E. S4 4#

TESTDATE: 18-Jl&35 START TIE: 14:16:23.8 pin MINI M 16.797 PSIA PiH MAXI M 21.814 PSIA rigure 2-2 Prersure Profile - Test S.08 e

s FIGURE ~2-3 GSU/ RIVERWETWELL BEND 5 VOL. AUG.85 TEMPERATURE 2.4 2.2 -

2-1.8 -

1 g 1.6 -

+

Wj 1.4 -

$C

• I 1.2 -

o Efi 1-I o.a - ,

0.6 -

0.4 -

0.2 - '%kbskkOb #

.t -

0 , , , , , , , , , , , , ,

O 4 8 12 16 20 24 28 TlWE (1000 SECOHOS) 1 h

1

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

i I 426.7 A

/%

O O  :

W Q

v N

(D w4 n.

.- b- -

I..

_..m

.J M h- ..

-17.8 412:34:9.23 TIME (SEC ) -3> 413:44:5.23 l

EPRNIRST SOFM TEST RR CASE B RELEASE, TEST NO. S-46 TEST MTE: 1FMHl5 START TIE: 14:16:23.8 T187 MINIM 38.5 EGC T187 WIM 327.4 EGC Figure 2-4 Temperature Profile for area adjacent to aten, tunnel and under !!CU rioor - Test 5.08

t 3.0 COMBUSTION PIBGENT OBSERVED IN 1/4 SCALE TESTING ,

Testing conpleted in 1983 in a 1/20 scale model of a Mark III contalment plant (reference 1) indicated that for full scale hydrogen release rates above approximately 0.4 lin/sec, steady diffusion flames would be produced on the suppression pool surface. %e 0.4 lbn/sec hydrogen flow rate was defined as the threshold above which steady

, diffusion flames would be e;q w ted. It was asstzned that the repeated deflagrations predicted by the CIASIX-3 cmputer code would provide a conservative representation of the cmbustion phenmena below the diffusion flame threshold.

HOOG ccmnitted to ccmplete additional testing in a 1/4 scale simulation '

of a Mark III containment in order to define the thennal envircernent produced by steady diffusion flames. Based upon discussions between HOOG and the NIC staff, !!OOG also ccanitted to cmplete testing in the 1/4 scale test facility which would allow IICOG to demonstrate the degree of conservatism present in the CLASIX-3 cmputer analyses.

HOOG has now ccupleted the scoping test portion of the 1/4 scale test progrm. H00G's initial evaluation of the scoping test program results as they relate to evaluation of parameters which could affect the thermal environment produced by diffusive hydrogen cmbustion are contained in reference 10. The scoping tests have provided considerable information on the cmbustion phenmena which is expected in full scale Mark III containments. %e cmbustion phencmena observal were not discussed in detail in reference 2.

Three scoping tests were ccmpleted to evaluate the threshold for existence of steady diffusion flames. We first two were intended to evaluate the threshold for steady diffusion flames under conditions which would be representative of degraded core accident conditions. We third test was intended to provide a ecmparison with the threshold <

testing ccrpleted in the 1/20 scale test facility.

Test S.08 was ccmpleted to evaluate the threshold for existence of steady diffusion flames under degraded core accident egnditions when a single safety relief valve is stuck open. The 330 simulated safety l relief valve was assumed to be stuck open for this test. Since all +

scoping tests were ccmpleted using a Mark III plant gecnetry which has a l larger core than River Bend, eight simulated ADS safety relief valves are open for the tests. In order to simulate ptential degraded core accident conditions, a hydrogen release history corresponding to hydrogen production following recovery of an ECC syst s with ficw capacity of 5000 GPM was injected. Reference 3 discusses the hydrogen release histories used in the accping test program. Following the EOCS reflood hydrogen release history,. the hydrogen injection rate is dropped to 0.21 lbn/sec (all hydrogen injection rates are full scale equivalent.

values) and held for approximately one minute. W e flow rate was then reducrx1 to 0.14 lbn/sec for another minuto. %e flow was tien decreased to 0.07 lbn/sec and held constant at this value for 45 minutes to define 3_

I

a the dcminant. embustion phenmenon present at very low hydrogen injection rates.

%e hydrogen release history used for test S.08 is shown in figure 3-1.

Wis release history presents actual 1/4 scale hydrogen injection flow

. rates. An initial deflagration established a steady diffusion flame on the suppression pool surface coincident with the rapid hydrogen release associated with reflooding the reactor vessel with an Irc systs. Since the release history used in this test has the highest hydrogen injection rate, this will produce the maximum hydrogen gradient in the wetwell ccmpared to other release histories. Infrared video cameras in the test facility showed horizontal propagation of the flamefront with the apparent- point of initiation under the steam tunnel. We pressure rise produced by the deflagration was very small indicating that the total amount of hydrogen consumed in the deflagration was not appreciable.

mis indicates the conservative nature of CLASIX-3 analysis which prcdicts an approximately 9 psi initial deflagration pressure rise for the same release history. We pressure history for a pressure transducer located in the wetwell is shown in figure 3-2. In addition, the hydrogen concentration which is measured continuously in the wetwell by sensor 11190 shows that the hydrogen concentration measured by 11190 at the time of the initial lightoff deflagration was approximately 4%. We hydrogen concentration measurment from instrument 11190 is shown in figure 3-3. In addition, the hydrogen concentration measured by 11001 is approximately 4% and the concentration measured by !!002 is approximately 5% at the time of the initial lightoff deflagration. %ese measurments indicate that the global hydrogen concentration measured in the wetwell and at higher elevations at lightoff is in the range' of 4-5%. These measurements agree with the 4.7% hydrogen concentration calculated by asstming that all hydrogen released prior to lightoff is uniformly mixed throughout the contairment, rollowing the initial lightoff deflagration, hydrogen burned as a steady diffusion flame on the suppression pool surface until the hydrogen injection rate dropped below 0.14 lbn/sec full scale equivalent. At about thirty minutes into the transignt, a weak and intemnittent diffusion flame appeared in the 315 chimney for approxinntely 10 minutes. This diffusion flame appeared with no visibly propagating deflagration. W e flames appeared to originate under the steam tunnel.

A thermocouple trace for instrument T-187 which is inmediately adjacent to the steam tunnel and under the 11CU floor is shown in figure 3-4.

This figure demonstrates the weak character of the diffusion flame in emparison to the diffusion flame present during the earlier high hydrogen release.

Test S.10 was empleted to evaluate the effects on diffusion flame threshold of assuming that the stuck open relief valve was actually an ADS valve. For this test, only 8 simulated safety relief valves are open. 21s is similar to the River Bend case in which there would be 7 ADS valves plus one stuck open relief valve open. The same hydrogen release history is used for this test as the release history used for

a test S.08. The hydrogen release history injected into the facility is shcun in figure 3-5.

In test S.10, a steady diffusion flame is established on the suppression pool surface before the rapid hydrogen injection associated with DCC recovery occurs. The initial deflagration which ignites the diffusion flame again appears to originate under the steam tunnel. As with the initial deflagration which occurs for test S.08, a relatively small pressure rise is produced by this deflagration. Figure 3-6 shows the pressure history for instrument P-100. The initial deflagration occurs when the global hydrogen concentration is slightly less than 4% by volume. Figure 3-7 shows the hydrogen concentration measu: x1 in the upper region of the wetwll by instrument H-190 for test S.10.

After the hydrogen injection decreases below 0.14 lbn/sec full scale equivalent for test S.10, the diffusion flames on the suppression pool surface extinguish and do not reappear. tb additional ccxtbustion is visible on the videotapes frm the infrared television cameras in the wetuell. As can be observed in figure 3-7, the hydrogen concentration reaches virtually a steady state value during the transient. This conclusion is reinforced by the continuous hydrogen concentration measurments frczn instrument H-410 which is located imnediately below the top of the 45 chimney. Figure 3-8 shows this hydrogen concentration masurment as a function of time. Since hydrogen is being injected thrcughout the test, it is apparent that scxne type of carbustion must be consuming hydrogen. tornoccuple data in the 45 chimney indicates that scxne type of weak, localized embustion is occurring in this chimney. Figures 3-9 and 3-10 show tmperature traces for themoccuplesgT-309 and T-410. These thennoccuples are in the upper regions of the 45 chimng.

%e third test completed to investigate the threshold for establishing steady diffusion flames was test S.04. This test was intended to replicate as closely as possible the threshold tests empleted in the 1/20 scale test program. Eight simulated SRV spargers were used in this test. A steady diffusion flame was established as early as possible with an initially high hydrogen flow rate. his prevented accumulation of a- significant background hydrogen concentration which 1100G believes contributes to a lower threshold for establishing diffusion flames. The hydrogen flow rate was then stepped down to 0.281hn/sec and held at a constant value. We hydrogen flow rate was then reduced to 0.21 lbn/sec and following that to 0.14 lbn/sec. The diffusi,n flames became intermittent when the flow rate was reduced frcxn 0.21 lbn/sec to 0.14 lbn/sec. The diffusion flanes did not extinguish until tim flow rate was lowered to 0.07 lbn/sec. The apparent lowering of the flow required for initiation of intermittence, (threshold) at 1/4 scale is partially attributed to the improved modeling of the sparger devices which have vertical slits simulating the flow frcun each column of sparger holes as opposed to the 1/20 scale spargers which contained only 4 holes in each side of each arm. In addition, the improved overall modeling of the phenmena at 1/4 scale, i.e. fully turbulent flow vs. scxnewhat laminar flow off the pool at 1/20 scale, is also a contributing factor.

.: e i

The HCOG cmpleted two tests to identify the limiting thennal envirorment p.h by hydrogen cmbustion. Test S.11 was empleted with a hydrogen release history corresponding to a 150 GPM reflood of the vessel and a sustained hydrogen release of 0.14 lbn/sec full scale 1 equivalent hydrogen release until the total hydrogen produced equaled the amount produced by , oxidizing 75% of the active core cladding inventory. We hydrogen release history for this test is shown in figure 3-11. We purpose of this test was to identify a limiting thermal envirorment produced by diffusion flames when total hydrogen production reached a 75% metal water reaction (75% WR) . The hydrogen was released through the 8 simulatcd ADS safety relief valves aid a single safety relief valve assumed to be stuck open in the 45 chimney.

4 The stuck open relief valve was postulated to occur in the 45 chimney in order to create the most severe diffusion flame environment. Since test S.11 was intended to define a limiting thermal envirorment for plants with containment sprays, the sprays were actuated when the n average contaiment air space tmperature reached 185 in accordance with the existing primaq cmtrinment mergency procedure cuideline.

During test S.ll, the initial deflagration which establishes the diffusion flame on 'ne suppression pcol surface occurs during the hydrogen production spike associated with initial injcction of 150 GPM into the reactor cressure vessel. A steap diffusion flama exists on the suppression ;xcl surface in the 45 chimney throughout the transient. Diffusion flames exist on the pool surface above each spargerdevgcedaringthehydrogenproductionspikeand reappear twice in the 315 chimney during the sustained hydrogen production portion of the transient, Firmm ?-12 shows the ternperature profile in the 45 chimney above the HCU floor as measured by p rmocouple T-204. Figure 3-13 shows the tarperature profile in the 315 chimney above the HCU i

floor as measured by thermocouple T-287. These two thermocouples represent limiting thermal environments in these two chimneys at the HCU floor.

Test S.09 was empleted to evaluate the limiting thermal environment -

which could be produced by localized cmbustion for accidents involving total hydrogen production equivalent to 75% MNR. The same hydrogen release history used in test S.11 was used in test S.09 except that the sustained hydrogen injection following the 150 GPM reflood hydrogen was reduced to 0.07 lbn/sec. Figure 3-14 shows the hydrogen release history used in test S.09. As in test S.11, the contalment sp;ays were actuated when the average contaiment tmperature reached 185'T. We hydrogen was released through the eight ADS valves which correspond to the scoping test ADS locations. W e stuck open relief valve was assumed to be an ADS valve for this scenario since the only test in which localized cmbustion had been cbserved was test S.10 which used only the 8 ADS spargers.

During test S.09 the initial deflagration which establishes the 1 diffusion flame on the suppression pool surface occurs during the hydrogen production spike asscciated with the 150 GPM reflood. A steady

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

o i

i diffusion flame exists in the 45 chimney throughout the period of hydrogen injection into the facility. Figure 3-15 shows the tauperaturg measured by thermocouple T-209 which is above the HCU floor in the 45 '

chimney. -No evidence of the localized cmbustion which occurred in test S.10wasobgervedduringthistest. 'Ihermocouples in the upper regions of the 45 chimney which provided evidence of localized cmbustion in test S.10 seem to indicate only the presence of diffusion flames on the suppression pool surface. Figures 3-16 and 3-17 show the tmperaturg response for thermocouples located in the upper regions of the 45 chimneys. A comparison of these tmperature plots with figures 3-9 and

, 3-10 verifies that the same phenmenon present in test S.10 is not occurring in test S.09. [

i The testing cmpleted to date has danonstrated that in a Mark III i contairinent, hydrogen cmbustion will be initiated before bulk average ,

l wetwell hydrogen concentration reaches 6%. The testing has shown that l for very low hydrogen generation rates, it is still possible to maintain intermittent diffusion flames on the suppression pool surface. All 4 deflagrations observed to date in the facility are very weak, and in many cases virtually imperceptible. Bulk average hydrogen concentration i throughout the test facility never exceeds 6% by volume.

i I

/

f i

l J

b t

1

,r_,. _ . . _ _ . - , _ - . - , _ - - . _ _ _ - , - , , _ , . . _ - .

l l

l 1 4,806 h

n Z

I N

s v

,4 O

(D I

b i 4.000 g b I

0N:43:19.23 TIME (SEC ) 4 # i:13:19.23 0

EPRIflRST SCTIE TEST FOR CASE B ELEE, TEST W. S-g 42-TEST NTE: 18- M SIM IN: 14:16:23.8 MBli W I M 6.000IM mi MIM 3.844IM n eu n, 3 i .

Ilydroren Pn1 ens" !!!ntory - Tc.-t. ",04 i

l 36.m i

/\

n

4 H

U)

Q.

(  :

+ -

J'

( * ~

x '

1 7 14.N6 W:63:19.23 TIME (SEC ) -> N1:13:19.23

" EPR!flRST SQFING TEST FR CA6E I IEWdE,iEST NO. S4 41**

TESTDATE

IFMH6 STARTTM: 14:16:23.8 P1N MINI M 16.797 PSIA PIN WIM 21.814 PSIA v i r o, ,, , . :. ,

l'att.r un e. '" o t si ( ur t i e i f ) . >i . e. , ;,, c f

a l

l I i 14.m

/\

a .

X

.J O

v

._ m

/ -

-W%  :-w :p

/

I i

! -2.50 ti2:M:5.23 TIME (SEC ) --> 613: # : 5 .23 EMI HET SWIM RST RR CASE I EEE,HST W. H8 *"

TEST MK: 18-MHE STMT TM: 14:16:23.8 H1N MINIllM 4.065 RX HIM WIM 4.979 R X Firurn .1 - 1 flydroren Conrefitrnt inn in 49' Chiratiny in wetwnll - Tent,/1,04

I i

)

I 426.7

/\

i n

O O l W l 0

N I

(D y 1,

.. - -. L. I..

.u . J N - .-

-17.8 412:34:5.23 TIME (SEC ) --> 613:44:5.23 DRI f1RST SCOPING TEST RR CASE i RELDGE, TEST NO. S 48 TEST W E: 19-MM SIMI TN: 14:16:23.8 T187 MINIM 38.5KGC T187 W I M 327.4NGC

,,,...,..o.,..

Tg,r p t,rn f liro f rO f f I t' f 0 f' U I fl fl'!) d C III t U nt r,nn f,unno t nn't un<for 11C0 f!Oor - ie.1t .9.04 '

l l

t 8

! 4.900 i

/\

n Z

I

! N i e v

i si  ;

O (D

I lL- , _

6.000 000:43:23.3 TIME (SEC ) --> 601:13:23.3 EPRI-SCOPING TEST, CAGE T RE1. EASE, All6 OM.Y, TEST NO. S-il TUT NTE: 2NilHE STARTTM: 13:47:24.6 MiMINDtM 4.004IM Mi WIMM 3,833IM F l vin re* l r, ,

flydrnce*n It<*lanno flint ory - Tt*"t t1.10 I,

I l 30.H6

/\  ;

n ,

4 H

U)

E.

Q = .. -

x  :

\ a. /

14.W W:43:23.28 TIME (SEC ) -> #1:13:23.28

-s oo EN!-SQFIE TEST, CASE 'l' RElDSE, A06 M.Y, TEST W. S-il TEST M TE: 21-AAHE START T!E: 13:47:24.6 PIMMINIM 16,875 PSIA PIN WIM 21.861 PSIA l'icore "f.,-

l'i nnt.ur n Tresco (wntwoJl) - Tnt.t '.;.10

. t 1

i i

1 14.000 l

1 l

1

/\

l n

i N i .J

O W

M wk _

! $ J@

,4 N

I 2.000 .

412:31:N.8 O

TIME (SEC ) -> 413:41:N.3 Hec

! RI40!PN TUT, CNE T EIAGE, ADS KY, EST W. S-il

TUT ME

ti-JtHE STARTTM: 13:47:24.6 l

Hi# MIDtM 6.5 EX Hi# WIM 4.979 E X i Figure 3-7.

l Ilydrogen Concentration in 11r;" Chimney in wetwoll - Test S.10 l

j i 14.000 h

^

X '

_J O

v

, t% _m __

r=  : =- - -< >-

, ,4

+

I

-2.006 TIME (SEC ) ->

012:31:03.28 013:41:03.28 EPRI-SC(PIE TEST, CASE T EEASE, ADS QU. TEST 2. S-il TESTDATE: 21-MHE START TN: 13:47:24.6 H414 MINIM 4.053 WLX H414 N I M 4.754 WLX Figurn 3.n, flydroger. Enncentration below top of 45" Chinnny

- Tent S.10

i l

I 8 426.7

/\

l 4

O O

O W

A v

l @

G L r0 L , 1. i 1 m 6 .

_ _f 8

_ l dd l b L .

-17.8 412:31:43.28 TIME (SEC ) 4 613:41:63.28 o 9zco EPRI400 PING TEST, CASE T RELEASE, ADS OM.Y, TEST NO. S-il TESTDATE: 21-JGHl5 STARTTM: 13:47:24.6 T349MINIMM 42.4 KGC T349MAXIMM 243.9 EGC Figure 3-9.

Temperature Profile in upper region of the 45 Chimney - Test S.10 l

l

- r __

I 426.7 A

b O

hl A

v l O

$ b Y A & L r

ft

~

2 L,.

-17.8

, 612:31:43.28 o

TIME (SEC ) -> 413:41:43.28 t

9200

EPRI-SCTING TEST, CASE 'B' REl. EASE, ADS OM.Y, TEST NO. S-14 TESTDATE

21-MHl5 STMT TIE: 13:47:24.6 T416 MINIM 42.1 E GC T416 MAXIM 334.8 ECC Figure 3-10.

Tenperature Profile in upper region of 45 Chimney - Test S.10

a 1

i 2.M b .

! ^

Z J LI I

ig g

s v

• i G r (D

I ,

J  :

6.m c 666:43:22.0 TIME (SEC ) 4 662: # :42. 0

~o V**

EPRI-S-11 RELEASE TO 75 PERCENT WITH TAIL JUST ABOVE TWESE3 TEST MTE: 9-Jll-95 START TI!E: 15:47:24.1 M861HINIM 6.066 IM 1961 MIM 1.824IM l

Figure 3-11.

Hydrogen Release - Test S.11 l

___j

l 426.7 h

c 0

i1 l l Id A

j' l N '

.4 'il1 # Wl it. t$ llt Jlld l- 7l y Vjh fII'T IlTMi }g ,_.

-17.8 M5:19:42.E TIME (SEC ) --> M7:16:22.E EPRI-S-il REl.EME TO 75 PERCENT WITH TAIL JUST ABOVE THESH1D

. TESTDATE: 9-AL-05 STARTTIlf: 15:47:24.1 l T2M MINIltM 45.5 KGC T264 MAXI!!M 283.5 EGC Figure 3-12.

Temperature Profile *in 45 Chir:ney above the HCU floor - Test S.11

l 426.7

/\

n O .

O Id O

i

, N l

(D N i .l. .1L

" r} _ ::_OW_

u.1 .

-' A $56 :_ . - -

-17.8 H5:19:42.E TIME (SEC ) 4 M7:16:22.0 EPliI-S-il RELEASE TO 75 PERCENT llITH TAIL JUST AB0VE T}RESRD TESTDATE: 9-Jll-05 START TIE: 15:47:24.1 T287 MINIM 43.7DEGC T287 MAXIM 2 H 6 KCC Figure 3-13.

Temperature Profile for 315 Chimney above the HCU floor - Test S.11 l

l 2.4#

/\

n Z

I N

2 v t 1

,4 G

(D I

l t.He A l

666:63:23.M TIME (SEC ) -> H2:58:23;#g, EPRI-S-49RELEASETO75PERCENTWITHTAILJUSTBELOWT TESTDATE: 15-Jll-85 START TIE: 16:56:24.3  !

l M891 MINIM 9.000IM M801MAXIMUN 1.826IM Figure 3-14.

Hydrogen Release - Test S.09

l l

l l

1 426.7 A

n O

O hl O

l 0

O N

Y 3

R DA ubl&L \

N h\L&b

-17.8 616:44:43.N TIME (SEC ) 4 613:39:43.#

EPRI-S-69 RELEASE TD 75 PERCENT WITH TAIL JUST BELOW TIESHU TESTDATE: 15-JtL 4 START TIE: 16:56:24.3 T269 MINIM 44.6DEGC T269 MAXIM 192.3DEGC Figure 3-15.

Temperature Profile for 45 Chimney above the HCU floor - Test S.09

I t

l<

l 426.7 -

s ll A  :

~i a

o ll.

dl 0

LLI Ci O J rq m .

f-y y

MMMMW3' """brIAM ,

-17.8 919:44:43.# TIME (SEC ) 4 913:39:43 # -

~

l EPRI-S-49RELEASETO75PERCENTWITHTAILJUSTBELOWTIRESHDL l I

TESTDATE: 15-Jll-85 START TIE: 16:56:24.3 .

i T369 MINIM 45.0DEGC T309 MAXIM 214.7DEGC Figure 3-16. l Temperature Profile for upper regions of 45' Chimney - Test S.09 l

- j

1 l

l 426.7

/\

n O

O 111 Q j v

N g -_

w l 4 x e , ,.

l- (j'.';A ?ft'#'+::,P#A6; *r + (

-17.8 616:44:43.66 TIME (SEC ) 9 613:39:43.60 EPRI-S-69RELEASETO75PERCENTWITHTAILJUSTBELOWTIRESHOLD TESTDATE: 15-Jll-85 START TIE: 16:56:24.3 T416NINIM 45.1 EGC T416 MAXIM 221.0 KGC Figure 3-17 Tenperature Profile for upper regions of 45 Chimney - Test S.00

4.0 Assessment of Equignent Survivability Gulf States Utilities has provided a prelhninary assessment of equipnent's capability to survive deflagrations. As noted in section 2.0, a number of conservative assumptions were included in the CLASIX-3 analysis which was used to define thennal environments for assessing equignent survivability. The analysis discussed in reference 4 showed that the calculated peak equipnent surface tanperature or the tanperature of the critical ccrponent for the igniters and cable located in the wetwell exceeded the equignent qualification tanperature. As noted in section 3.0 of this report, additional testing in the H00G's 1/4 scale test facility has irrlicated that the CIASIX-3 predictions of thermal environments is excessively conservative. Gulf States Utilities has canpleted additional analyses to assess equipnent's ability to survive canbustion phenonenal observed in the 1/4 scale testing.

2 e 1/4 scale facility has been designed to allow simulation of each Mark III containment's plant unique geanetry. The scoping testing phase of the 1/4 scale test program has been coupleted using the plant unique geanetry for another Mark III containment. This geanetry also provides a reasonable basis for assessing equipnent survivability at the River Bend Station. The principle geanetric difference between the scoping test geanetry and 'Ihe River Bend Station is the significant flow restriction present at the refueling floor in the River Bend Station design. This flow restriction will not affect conditions in the wetwell or near the HCU floor. Ibsts canpleted in the 1/20 scale test facility denonstrated that the extensive restriction to flow at the refueling floor elevation will not alter the character of canbustion.

A HEATING-6 rnodel of the hydrogen igniter has been developed to provide verification of the modeling documented in reference 4. This model was used along with the thermal environment measured in the 1/4 scale facility to calculate the igniter's tanperature response. Data fran test S.12.2 was used for this analysis. This test involves total hydrogen release corresponding to 75% MWR. Although the containment sprays were operational for this test, this test is believed to provide the best basis using currently available 1/4 scale test data for assessing equipmnt's ability to survive accidents where total hydrogen generation equals 75% MNR. The use of sprays in this test will not affect the applicability of test results to River Bend since the sprays were not activated until the very end of release historf A'. Since sprays are not activated until late in the transient, the prior thermal environment, which poses the greatest threat to equipnent survivability, is applicable to RBS.

Tanperature data fran thernoccuple T-202 in the 1/4 scale facility was used to evaluate the thennal response of the igniter. Thernoccuple T-302 represents a limiting diffusion flame thermal envirorrnent in the 45 chimney for test S.12.2. Since the safety religf valve spargeg which is assumed to be stuck open is placed in the 45 chimney, the 45 chimney should represent the most limiting thermal environment.

, . i l

l l

The igniter assembly surface tempen ture and the tcsuperature of the igniter transformer are plotted in figure 4-1. This figure also shows 4'

the tcraperature trace which has been hand digitized fran 1/4 scale test data. Figure 4-2 shows the tanperature data measured by thermocouple T-202 during test S.12.2. Free convection has been used in this calculation along with radiation fran the canbustion products plume. As indicated in figure 4-1, the igniter's response ranains well below the equipnent qualification tanperature.

The response of a hydrogen igniter power supply cable to the 1/4 scale thermal envirorment has also been calculated. The conduit through which the hydrogen igniter cable is routed is included in the HFATING6 model.

The taaperature of the conduit, the cable insulator, and the conductor are shown in figure 4-3. Figure 4-4 shows the 1/4 scale tanperature

, plot for thermocouple T-200 which has been used for analysis of the cable thermal response. This thennoccuple was selected for evaluating the response of the igniter power cable in order to provide diversity in

! the thermal environments used to assess equipnent survivability. As with the analysis for the igniter, free convection to the boundary of the conduit has been used. Because the analysis has been canpleted in cylindrical coordinates, the radiation has been applied uniformly to the entire circumference of the cable. This represents a significant conservatism in the analysis. Even including this conservatism, figure 4-3 indicates that the cable's ability to survive is not jeopardized.

The response of a pressure transmitter to the diffunion flame thermal environment has also been calculated. Data fran thermocouple T-200 was used for assessing the ability of the pressure. transmitter to survive hydrogen canbustion. This thermocouple was used to assess survivability ,

of the pressure transmitter because several HOOG mernber plants have instrument racks containing pressure transmitters located near the steam tunnel. This location corresponds to the location of thermoccuple T-200 in the 1/4 scale facility. Figure 4-5 shows the surface tenperature for the pressure transmitter as a function of time along with the hand digitized tanperature data frcm thermocouple T-200. As with the analyses of the igniter and the igniter power cable, free convection and radiation have been applied to the boundaries for the canponent. Figure 4-5 denonstrates that the pressure transmitter has considerable margin in its ability to survive hy%eu cc nbustion.

1

.=

Figure 4-1.

HYDROGEN IGNITER ASSEMBLY RESPONSE TO TEST DATA 550 1

'\

500 - \

\

k a_ m L .

C 400 - \

g

, w 350 -

W 3

3 0 0 _-

g . \ _

250 - ,[ \

200 - 4

\ -

~

s 150 - N-..

o --

100 , , , , , , [ , , , , , , , , , , , ,

O O.2 0.4 0.5 0.5 1 1.2 1.4 1.8 1.8 2 (Thous an ds)

TIME (SECONDS)

BOX SURFACE + TRN4S FORf.1ER 0 OSTF S12.2 T202

i

315.6

/\

ll

^

11 E

O g O sis W l'll O

Il N

@ ll f I , , h la u l R.1 2 G al F0

\

%pyr ri -  : iv ' ' - ' -

K+ h

~

37.8 l 2ti.N TIME (SEC ) -> 7701.66 EPRI-S-12.2SECONDREPEATTESTWITHGRATINGANDR.0CKAGES TESTDATE: 1-AUG-85 START TIE: 13:45:24,9 T262 MINIM 44.2DEGC T262 MAXIM 274.3DEGC Figure 4-2.

Temperature Profile for 45" Chimney above the HCU floor - Test S.12.2 l

l

Figure 4-3.

RBS CONDUITED IGNITER CABLE EXPOSED TO QSTF TEST S12.2 500 MO 400 -

,g b 350 - a a

w

$ 300 -

3 .

NQ

~

$ 250 -

1 W

200 -

150 - ,

IE 100 , , , , , , , i , ,

0.8 1 1.2 1.4 1.8 1.0 (Thous an d s)

TIME (SECONDS)

D CONDUCTOR + INSULATOR C CONDUlT 4 T2OO M

...s i

i 315.6 l

^ \\\ l l ikbl U l 8 a"MhA I B FW1 l 9 FI F 11 N

II I I %J I I I l p I I I I man f

n l l l 37,g ______ __..~ ,

l l l 291. 9 TIME (SEC ) 4 1241. 9 EPRI-S-12.2 SECOND REPEAT TEST lilTH CRATING AND BL000lCES TESTDATE: 1-AUG-05 START TIE: 13:45:24,9 T2M MINI M 43.8 KGC T2M MAXI M 275.3 EGC Figure 4-4.

Temperature Profile for 45" Chimney above the HCU floor - Test S.12.2 i I

. _ . . _ . _ _ _ _ - , a _ _ _ _ a _a4 6 m . ._.__ ,. _ . _ _

h I

_w e <-.

_ O.

o W -

,9

_4 -

ri,

[1 4 k m (f) "4 O '8 _N t

" np.

i e Zs m ui  !

Ib E"

$ - - 5 8.

E l- 0y <. _ sN Ov M _$

T@  ?

D$ -

(f) " _9 (O 62 Ld y e

Q_

a

_N .

o n.

0 l 1 i i i i i O O O O O O O O O O O IJJ O tG O 10 O u) O g) 4 't r) F1 r1 ti e- r (a eaa) aanivaannal l

I l

'e .. .

1 g .. g 5.0 Conclusions 4 '1he CLASIX-3 code has been used to. date for purposes of assessing eglipnent's ability to survive hydrogen cmbustion in the form of deflagrations. Very conservative assunptions have been used in i cmpleting the CLASIX-3 analysis. This has resulted in very i conservative predictions of both the peak pressure produced by hydrogen ocarbustion in the Mark III containnent, and the tmperature environment to which equiptent would be exposed during hydrogen generation events.

Although test S.08 did not involve total hydrogen injection equivalent l to 75% WR, the ECCS reflood transient and the sustained, low hydrogen

! production of 0.07 lbn/sec are directly emparable to the hydrogen j release rates used in the CLASIX-3 analysis. The cmparison provided in section two clearly dmonstrates that the CIASIX-3 thernal profiles are extremely conservative in cmparison with the tmperatures measured during 1/4 scale testing. In addition, as discussed previously, the .

car

parison of 1/4 scale pressures with pressures predicted by CIASIX-3 l reinforces the conclusion that CLASIX-3 is overly conservative. -

Several key assunptions used in the CLASIX-3 analysis appear to be excessively conservative. The 1/4 scale tests have dernonstrated that

cmbustion will be initiated well before global hydrogen concentration reaches 5%. In fact, the tests have dmonstrated that the hydrogen concentration does not exceed 6% for all tests cmpleted to date. The i testing to date has dmonstrated that steady diffusion flanes can exist on the suppression pool surface for hydrogen flow rates of 0.14 lbn/sec for all sinulations of degraded core accident hydrogen production.

When hydrogen flow rates are below the threshold for steady diffusion flames, repeated deflagrations are not observed. 'Ihe only deflagration resabling the deflagraticns predicted by CLASIX-3 is the initial lightoff deflagration. When diffusion flames are reignited after extinguishing thernselves, the deflagration which reestablishes the diffusion flame is virtually inperceptible on the infrared camera videotapes. This further ephasizes the conservatism of the CIASIX-3 predictions of thermal enviroments associated with hydrogen cabustion in the Mark III containment.

Gulf States Utilities is a mertber of the HCOG. HCOG has a long term program of analysis and testirg in progress to assure resolution of issues associated with degraded core hydrogen control. This program will result in ccmplete definitica of the thermal environments produced by hydrogen ombusticn including definition of thermal envirorsnents produced by steady diffusion flames, and by deflagrations. For the purpose of initial plant licensing, Gulf States Utilities has ruhnitted plant specific analyses of contairment response and equipment survivability. These analyses have demonstrated that the contaiment structure will survive the peak pressure produced by hydrogen cmbustion without failure. The analyses have also dm onstrated that with the extremely conservative thermal environments predicted by the CIASIX-3

ca puter program, selected caponents may reach or exceed their

t*'s-  !

)

equipnent qualification taperature. Based on the information available fran the 1/4 ' scale test program and additional analyses of equipnent response to diffusion flame thermal environments defined frun 1/4 scale data, Gulf States Utilities concludes that the long term program of analysis and testing currently in progress through HOOG will result in demonstration of ~equipnent survivability. Accordingly, Gulf States Utilities considers it evident that sufficient information has been presented to warrant licensing of the River Bend Station with a license ccndition to ccrplete the HCOG generic program of analysis and testing.

r t JJ ' 4 References

1. 1/20' Scale Final Test Report subnitted to the NRC staff by letter IKE-014 dated February 9,1984

2. Report on 1/4 Scale Scoping Test Results sulmitted to the NRC staff by letter IKN-053 dated August 1,1985

3. Final Scoping Test Hydrogen Release Histories subnitted to the NRC staff by letter IKN-031 dated March 13, 1985  :

4. Preliminary Equipnent Survivability Report Supplement Three subnitted to the NBC staff by letter RBG-21,912 dated August 22, 1985