ML17319B198

Hydrogen Mixing & Distribution in Containment Atmospheres.

ML17319B198
Person / Time
Site:	Cook
Issue date:	12/31/1981
From:	Hudson F, Shiu K, Wilder J ASSOCIATED ELECTRIC COOPERATIVE, INC., DUKE POWER CO., TENNESSEE VALLEY AUTHORITY
To:
Shared Package
ML17319B196	List: ML17319B195 ML17319B197 ML17319B198 ML17319B199 ML17334A364 ML20041B520
References
NUDOCS 8202240182
Download: ML17319B198 (30)

Text

HYDROGEN MIXING AND DISTRIBUTION IN CONTAINMENT ATMOSPHERES PRELIMINARY PROJECT REPORT DECEMBER 1981 Prepared by:

J. J. Milder, Tennessee Valley Authority F. G. Hudson, Duke Power Company K. K. Shiu, American Electric Power

'7 Pro)cot Conducted by:

Hani'ord Engineering Development Laboratory Prospect Sponsors:

American Electric Power Service Corporation Duke Power Company Electric Power Research Institute Tennessee Valley Authority 8202240182 820217 PDR ADQCK"050003i5 P PDR re@ y qy ~pqM >

TABLE OF CONTENTS

1.0 INTRODUCTION

2.0 TEST FACILITY 3.0 TEST MATRIX 4.0 TEST RESULTS 5.0 PRELIMINARY CONCLUSIONS

l 0

1.0 INTRODUCTION

The oh)ective of this program was to measure hydrogen mixing and

, distribution in a simplified scaled light water reactor subcompartment. The test matrix for this program was designed to characterize hydrogen distribution in a typical lower compartment region of an ice condenser 'containment under two different release conditions. The test vessel used for these tests was modified to geometrically simulate the lower compartment volume and the associated flow paths into and out of that region. Hydrogen and steam release rates used in these tests were scaled to model a small break loss of coolant event based on analytical studies (MARCH computer code analysis of Sequoyah Nuclear Plant) performed for the NRC by Battelle Columbus.

The test results presented in section 4.0 are based on preliminary results provided by G. R. Bloom and L. D. Muhlestein of the Hanford Engienering and Development Laboratory. This report has been prepared by the utilities.

2.0 TEST FACILITY The Hanford Engineering Development Laboratory Containment Systems Test Facility (CSTF) houses a carbon steel test vessel 25 feet in dimeter, 67 feet in overall height, which encloses a volume of 30,000 ft . The vessel has standard elliptical disked heads at'he top and bottom and models a scaled down LMR containment building with minimal interior structures.

Hydrogen mixing in the upper compartment of an ice condenser containment is expected to be very good. Upper containment sprays will induce turbulent'mixing of the gas and the upper compartment geometry should prevent stratification or stagnant regions in'that large open area. This program, therefore, is focused on the lower compartment where hydrogen from a degraded core event'would be expected to be released.

For the purposes of these tests, geometric similarity was retained the test compartment and the lower compartment of an ice 'etween condenser containment. Dimensions of the compartment constructed in the CSTF vessel are shown in Figure 8. By comparison, the ice condenser crane wall diameter is 83 feet and the CSTF outside diameter is 25 feet. The height to diameter ratio of the test compartment is equal to that for the plant. The linear scaling factor between the test compar tment and the plant compartment is 0.3.

The high velocity Jet nozzle shown on Figure 8 was located in one corner of the test compartment directed horizontally 60 degrees away from the center of the vessel at a height of approximately 5 feet from the floor. This location was chosen because the ma)ority of the reactor coolant system piping in an ice condenser plant is located in the bottom half of the lower compartment. The corner location was

3 chosen to be conservative. The second hydrogen release location was positioned vertically upward at approximately the 140 degree position on Figure 8 to be geometrically similar to the location of an ice condenser plant's pressure relief tank rupture disk.

3 ' TEST MATRIX 3~ 1 SCALING As mentioned above, this test program was geometrically modeled to simulate hydrogen releases in the lower compartment of an ice condenser containment. The test parameters: hydrogen and steam flow rates, forced air flow into and out of the compartment, and release locations and sizes, also had to be modeled so that the results of these tests could be integrated into the full scale degraded core accident analysis.

Modeling is said to be ideal if all dimensionless groups characterizing the event (including geometric scale factors) are identical in the model and the prototype. However, in complex models it is often difficult or impossible to select model test parameters so that all dimensionless groups are identical to those of the prototype. For this reason non-ideal modeling is often used. In such non-ideal models, similitude for one or more of the dimensionless groups is compromised so that experiments can be carried out to evaluate the dominant controlling phenomena. When non-ideal models are used, it is obviously important that the dominant parameters be identified, and that test conditions be selected to yield parameters which are essentially the same in the prototype.

HEDGE, with the help of their consultant, Dr. Arlin Postma, examined the dominant mixing processes and identified the following four dominant mixing processes in the ice condenser annular volume for the release conditions under consideration;

1. High speed horizontal get mixing

2. High speed vertical get mixing

3. Fan induced recirculating air flow and Natural convection flows along surfaces It was determined that in modeling these mixing processes that the following criteria would be met.

1. Preserve the similarity of the densimetric Froude numbers. The densimetric Froude number is the ratio of momentum forces to buoyancy forces.

2. Preserve geometric similarity.

3. Preserve scaled relative times which is required because of competing mixing processes.

4. Preserve density ratios of Jet to ambient atmosphere.

One other important factor in modeling these tests was the substitution of helium for hydrogen. The reason for a substitute gas was that safety requirements at Hanford would not allow hydrogen combustion at this facility. Helium was selected because it has nearly the same buoyancy in air or nitrogen as hydrogen and exhibits similar mixing characteristics. The thermal conductivity of helium is near that of hydrogen and the same thermal-conductivity type concentration monitoring equipment was used for both gases, thus, eliminating the need to reinstrument the test compartment for the one hydrogen test. The hydrogen test was conducted in an oxygen reduced atmosphere to confirm test results using helium as a simulant.

3.2 Mixin Tests The mixing test matrix is given in Table Z. Hydrogen or helium concentrations, gas velocities and temperature profiles were measured during simulation of the two hydrogen release scenarios. Mixing and distribution were determined from the concentration, velocity, and temperature profiles.

The four preliminary tests, HM-Pl through HM-P4, were designed to determine the separate and combined effects of natural and forced convection air recirculation. Natural convection air velocities were measured at near ambient and at an elevated containment gas temperature of 150 0 F.

Forced air recirculation and gas velocity in the compartment were measured in both ambient and elevated temperature cases.

Mixing tests HM-1 through HM-5 were all simulations of a steam-hydrogen release fr om a two inch pipe break Tests HM-1 and HM-2 were cases without the air recirculation normally

.provided during an accident scenario for an ice condenser lower containment region. Tests HM-3 and HM-4 were conducted with the modeled air recirculation flows. Tests HM-1 and HM-3 were conducted with helium-steam release rates arbitrarily set at one half the reference release rate to determine the effect of reduced jet mixing in the test compartment volume. Mixing test HM-5 was identical to test

k

'I 0 ~

HM-4 with the difference that hydrogen was used as the source gas and nitrogen was used as the containment gas.

This test was conducted= to determine whether helium was a valid simulant for hydrogen in these tests.

The final two tests HM-6 and HM-7 simulated a vertical hydrogen steam jet release directed upward from a location geometrically similar to that of a pressure relief tank (PRT). The helium-steam source rates were identical to those of tests HM-3 and HM-4 respectively. These flow rates taken from an S D type event were considered representative of the releases from a 2-inch pipe br eak as well as a rupture of hte PRT rupture disk.

4.0 TEST RESULTS Tests HM-P1 and HM-P3 demonstrated the difference in natural convection create) by ambient (85 F) and elevated gas temperatures ( 150 F) in the lower compartment. When comparing the effect of natural convection forces in the ambient air test HM-P1 with test HM-P3 the local average air velocity in the lower and upper volume increased by a factor of four when the compartment was heated to 150 0 F. Tests HM-P2 and HM-P4 were conducted with ambient and elevated gas temperatures respectively with the added parameter of scaled forced air recir culation. In these forced air recirculation tests, ambient air case HM-P2 and elevated air case HM-P4, natural convection effects were the most significant in the bottom of. the test compartment. The local air velocities in these two tests were of the same order of magnitude with the elevated temperature case being slightly higher.

Hanford experienced failure of several of their gas velocity probes in the middle region of the test compartment on test HM-P3. They could not repair them in time to keep the tests on schedule so they elected to proceed with the test series.

Hanford felt that sufficient gas velocity information was available from the remaining probes to describe the gas mixing process.

The helium and hydrogen concentrations in these tests were measured by 12 thermal conductivity analyzers. Ten of the analyzers were located in the lower compartment and two in the upper compartment. The analyzers required a constant gas flow rate of 150 cc per minute. The'gas sample was cooled, the condensate separated, and the gas dried prior to being infected into the analyzers. The concentrations presented in figures 1 through 7 have not been corrected for steam and are therefore conservative.

There is a modeling time scale of approximately 3.33 to 1 from the test to the plant case. Test helium transients will happen three times faster in the test compartment; therefore, one should multiply the time scale of Figures 1 through 7 by 3.33 to predict transients in the plant containment.

Helium concentration transients for ten test compartment points are presented in Figures 1 through 7 for tests HM-1 through HM-7 respectively. During the helium-steam release in tests HM-1 and HM-2, the maximum concentration difference between all measurement points in the test compartment was 2 vol $ He.

Following the helium-steam release, the test compartment, unlike the plant, developed a vaccum as the steam in the compartment condensed. This reverse migration coupled with the lack of a mixing mechanism from either the fans or the.get itself created a concentration difference of 7 vol $ He for test. HM-2. This was the largest difference observed in any test. Hanford did not believe that this portion of the test was in any way prototypic of the plant. The maximum concentration difference occurred approximately two minutes after the get release was stopped.

Mixing tests HM-3 and HM- 4 both exhibited good distribution of the helium during and after the source was terminated. The maximum difference in the test compartment at any time was less than 2 percent for both tests.

The hydrogen concentration transients for test HM-5 is presented in Figure 5. The hydrogen release rate was 0.66 lb/min or about 27$ lower than the modeled release rate of 0.9 lb/min due to problems with, the hydrogen flowmeter. The low hydrogen flow rate is not considered serious since the hydrogen concentration data can be normalized for comparison purposes and the jet mixing Froude number for test HM-5 was near that calculated for test HM-The hydrogen concentration transient in Figure 5 is identical to the helium concentration transient reported for test HM-4 concentration data is normalized. The data was normalized by if the dividing the concentration at any time by the peak-concentration for the test. Comparison of the hydrogen and helium concentration transients demonstrates that the curves are identical in every detail. The agreement between the test results of tests HM-4 and HM-5 demonstrates that helium is a valid simulant for hydrogen.

During test HM-6 the helium flow rate was too high by a factor of two during the ninth minute of a 10 minute release.

Extrapolation of the helium concentration transient prior to the inadvet tant high helium release rats leads to a prediction of a peak helium concentration of 6$ .

Comparing vertical and horizontal helium-steam release tests (HM-6 to HM-3 and HM-7 to HM-0) to each other shows relatively close agreement. The peak helium concentration of test HM-3 was 5.5 volume percent helium and the corrected peak for HM-6 was 6 volume percent.

The maximum observed helium concentration difference between points in the test compartment during the source release was 1.9 volume percent helium during test HM-6 and was 1.3 volume percent helium during test HM-3.

Tests HM-0 and HM-7 also had similar helium concentration transients. The peak helium concentration was 10.8 volume percent. for both tests. The maximum helium concentration difference between points was 2.8 volume percent helium in tests HM-7 and 1.5 volume percent during test HM-3. The conclusion from these test results is that Jet orientation plays an observable but minor role in the gas mixing process and the resulting gas concentration transients.

5.0 PRELIMINARY CONCLUSIONS From the preliminary test results we have drawn the following conclusions concerning hydrogen mixing:

1. Helium is a valid simulant for hydrogen in these hydrogen mixing and distribution tests.

2. The test compartment volume is well mixed with less than 1 volume 5 concentration difference between points within 20 minutes after stopping the hydrogen-steam or helium-steam source for all cases.

3. In all cases with forced air recirculation, the concentration difference between points was less than volume per cent 1

within 5 minutes after stopping the source gas.

The air recirculation fans minimize both the peak helium concentration and the maximum helium concentration difference between points in the test compartment.

5. Test compartment volume gas mixing is not strongly dependent on source get release orientation.

TABLE 1 H MIXING TEST MATRIX Recir c. Orientation He or Test Cont. Flow Source Source H

Flow Rale blow Initial l;ontaznment No. Gas (CFH) Gas Gas Jet (lb/min) (ib/min) Gas Tem ( F) fM-Pl Air 0 85 fM-P2 Air 3700 85 fM-P3 Air 0 150 fM-P4 . Air 3700 150 fM-1 Air 0 He-Steam Horizontal 0.9 27 150 fM-2 Air 0 He-Steam Horizontal 1.8 150 Air 3700 He-Steam Horizontal 0.9 27 150 fM-4 Air 3700 He-Steam Horizontal 1.8 54 150 IM-5 N2 3700 H2-Steam Horizontal 0. 66 150 Air 3700 He-Steam Vertical 0.9 27 150 fM-7 Air 3700 ffe-Steam Vertical 1.8 150 E71352. 29

~ i ~ i @4' eI ~ s 4I 4 4 4 W ~ s. 4 LI $ ~ 'J s s ~

-., I .l

': I i i

~

-z-

bJ.

C+

Figure 1. Test llH-1 llel ium Concentration Transients.

sa '03 'l$

00l0 9V s>>asss

~

os aa ss ~ Is x s.ical U1$ V Xu$ 1oi J>ARss o> X ol Ho)i

~ ~ I t

I i 4

~ >t I . 3 I

t l

tI a-

\

Figure 2. Test lit<-2 llel ium Concentration Transients.

Itt ttt ii tt '00 li3SS3 II 13 I JA3II OOOO 9P

~

S3IIDNI Ol X

~

I ~ IIJIII 3fu OX OI X Of R'>l

Figure 3. Test IIH-3 Ilelium Concentration Transients.

>4 ~ Ili4ee '03 UlSS3 00l0 9V S3IIDnl OI X l ~

V 1JJiALN ll3NI 3IIL Ol. 01 X Ol

a

~ ~ ~

~

I

~ ~ ~

g l I

~

'I l I

I I li I~

I'

.':Ill!

~ ~

Ii s

I

~ I Figure 4. Test lite-4 Ilejium Concentration Transients.

00LO 9V ~ vs a as sova 03 U3ss3 ss 13INI3)s S3IS3UIOI X I ~ IIDIII3IIL OL Ol X Ol M H

..:.la.

~ I Y.DRO

]jj I:

~ ~

~ ~

.I ~ I I I '

t/

~

~

/ I

~ ~ ~

) I I I

~

~ ~ I

. ~

j l I ~:

~

s~ >

~ I I y S g ~ ~ ~

FigUre 5. Test flH-5 llydrogen Concentration Transients F'i%I'I Nlll O> u3SS3 V l3JJA3)I OOOO 9l ~

SBHDNl 0> X C ~ HONE 3HL OA Ol X Ol

~ ~ -~ ~ T

-I- ~ ~

I C

f

~ I I'

- ~ ~ T f f+ I:.

~

~

I

~

T

~

IIELI t

~

.8

~ --t----

!h l

f I ~

~ ~ I

~ ' I f  !

f I ~ '.

• .;r"rfi ):,l! I:..

I I I  ;.!! I .!

~

f ~ ~

I Ii, (

.ME.'

I ~

I

~

I

~

~ ~ I I

< I: I I Figure 6.. Test HH-6 llelium Concentration Transients

>l1 <II II<IT 05 <<ASST h< 13JJA))I AS% ~

00l0 9L ~

Su<><<<0< X f ~ 1<3HI 314L OL Ol X f<l W H

I P

e O' LkW,~( ik p j;

~ 4 r

I ~

I ~

',T I

! j l I (

HEL j M 'CONCE 3R l ~

~

i

.i.-2-,-'

~

I I I

~  ! '

~ ~

Kt bJ

~

Q.

~

I---

I

~ 't r

~ ~

f

~ ~

- ~ ~ i-I- 0

~ ~ ~

I-I I f ~ I I ~ ~ I ~

j I I

' I I

Figure 7. Vest IIII-7 Ilel ium Concentration I vansients eS0 is u~~ 'OP 83S$ 3 V 133 JA3>l OOOO 9P S3HO!!I 0! X! ~ IIQN! /HE QJ, PI X PI

TEST COMPARTMENT SCllEMATlC CSTF VESSEL ANNULAR TEST

1 ~k~

I'. -)+ Wp,O 1 6 ft RAO.

I 346~

4 LOWE IA'ARRIEh OCK I'k Alh l1 g V RID~

F-4 It iCSTF VESSEL I>> jf WALL ANNULAfl TEST

,,'~; OII:ISIA w~TD It ~

gt 1~

16e 16 COI1BARTMENl'-

REMOVABlE RAOIAL WALLS A one OECK ~ I' 3 It

'ENSOR LOCATION STEEL BARRIEfl WWOW1111 Figure 8

Attachment No. 7 to AEP:NRC:00500G Donald C. Cook Nuclear Plant Unit Nos. l and 2 Additional Enformation on Hydrogen Mitigation and Control Core Recovery Studies

A preliminary draft report on the analysis of hydrogen production during the core recovery phase of the S2D event has been received from Westinghouse and is undergoing inhouse review. The results of the analysis will be transmitted to the Commission subsequent. to receipt and review of the final report.