ML20128P108
| ML20128P108 | |
| Person / Time | |
|---|---|
| Issue date: | 10/04/1984 |
| From: | Ryder C NRC |
| To: | |
| Shared Package | |
| ML20127A894 | List:
|
| References | |
| FOIA-85-110 NUDOCS 8507130203 | |
| Download: ML20128P108 (56) | |
Text
_ - _ _ _ _ _ - - ___-
.4 o
I
~,
l CODE COMPARISON This report is a sumary of "Results and Code Predictions for A8 COVE Aerosol Code Validation
- Test ABS," HEDI-TMI 83-16.
R. K. Hilliard, Hanford Engineering Development Laboratory The summary was prepared by Christopher Ryder, U.S. NRC NOTE: The information in this report is applied. technology.
It is not to be published or disseminated without the written permission of the U.S. NRC.
l e
8507130203 850415 PDR FOIA ALVAREZ85-110 PDR
~
L
OBJECTIVE This report describes the first confirmatory test.in the ABC0VE program, test AB5, and compares the computer code predictions with the experimental measure-ments. The objective of test ABS is to collect experimental data on aerosol behavior for une in validating aerosol behavior computer codes. The expericent is of moderate duration with a strong aerosol source generated by a sodium spray in an air atmosphere.
RELATIONSHIP TO SOURCE TERM RESEARCH The Aerosol Behavior Code Validation and Evaluation (ABC0VE) program has been developed in accordance with the LMFBR Safety Program Plan. The ABC0VE Program is a cooperative effort between the U.S. Department of Energy, the U.S. Nuclear Regulatory Commission, and their contractors currently involved in nuclear aer-osol code development, testing or application. A series of large-scale con-firmatory tests are to be performed in the Containment Systems Test Facility (CSTF) vessel covering a range of aerosol source rel' ease rates, source duration times, and aerosol composition. When experiments cannot be performed under the full range of postulated accident conditions the experiments must demonstrate that all of the significant aerosol mechanisms have been properly modeled and that the assumptions used in the modeling are valid.
EXPERIMENTAL FACILITY AND INSTRUMENTATION The test was performed in the Containment Systems Test Facility (CSTF). The CSTF is a model containment vessel which is located in a ventilated concrete building. Associated eouipment includes a sodium supply system, instrumentation t/ stem, control room and data acquisition system, data reduction and analysis system, chemical laboratory rooms, utility services, maintenance shop, and of-fices.
Containment Vessel. The CSTF containment vessel is a 850-m3 (30,000-ft )
3 carbon steel vessel with a design pressure of 0.517 MPa gauge (75 psig). All interior surfaces are coated with a modified phenolic paint, and exterior surfaces are covered with a 25.4-mm layer of fiberglass insulation with an outer aluminum vapor barrier.
Sodium Spray System. Commercial grade sodium is melted in a portable clam shell heater ano charged into the sodium supply tank. The supply tank is suspended from a lead cell so that the combined weight of tank and sodium can be measured.
Two valves and a magnetic flowmeter are located in the sodium line. Two sodium spray nozzles are installed in the containment vessel 4.2 m above the catch pan.
The nozzles are hollow cone types oriented to spray in the upward direction.
. j Aerosol Characterization. The suspended mass concentration, the particle size distribution, ano the chemical composition is measured periodically by direct sampling at various times and locaticns during the tests.
In addition, some information en shape and size is obtained by electron microscopy.
The mass concentration of suspended particles is measured as a function of time by periodically passing a measured quantity of gas through small filters located directly in the containment atmosphere and subsequently analyzing the material collected on the filter for total mass and for Na. Two types of samp-ling techniques are used:
in-vessel fiter clusters and through-the-wall samplers.
The aerodynamic size distribution is determined by sampling with cascade im-pactors inserted through the wall. Two types of cascade impactors are used:
Andersen Mark III 8-stage and Sierra model 225 6-stage. Previous tests have shown that these instruments give good agreement when properly calibrated.
Glass fiber collection surfaces provided by the manufacturers are used.
Cha.nical identification of the aerosol is determined at various times during each test by collecting aerosol on a membrane filter paper at a wall station and analyzing for various chemical species by x-ray diffraction and wet chemistry. The sample is protected from ambient atmosphere to minimize chemical changes that might occur after the sample is taken.
The instantaneous deposition rate of particles is measured by exposing cou-pons in a horizontal orientation for brief periods. The top surface of the coupon is washed and the rinse water analyzed for sodium. The deposition rate is calculated as a total mass flux of particles. No information is obtained on settling as a function of particle size by this technique. The " deposition velocity" was calculated by dividing the flux by the airborne concentration.
Temperature Measurement. All temperatures are measured by calibrated Chromel-Alumel tnermocouples with stainless steel sheaths. Readout is in parallel on strip chart recorders, magnetic tape, and paper tape.
Pressure Measurement. The absolute pressure and the gauge pressure in the con-tainment vessel is measured by a pressure transducer and a Heise gauge. The differential pressure between the cover gas in the sodium supply tank and the containment atmosphere is measured by a differential pressure transducer.
Gas Analysis System. The composition of the containment gas is measured con-tinously at five locations by pulling samples through tubing to on-line analyzers located ex-vessel. Filters at the tube inlet prevent aerosol from entering the analyzers. A few grab samplos are taken for subsequent analysis by mass spec-tremetry.
O i :
Data Acouisition System. Many of the key experimental measurements are made
~
manually and periacically, such as the filter samples, the cascade impactor-samples, the electron microschop samples, and the chemical analysis. The data associated with these manual samples are logged by technicians onto data sheets or recorded in notebooks, i
The on-line instrumentation includes thermocouples, pressure transducers, a sodium flowrate mater, a sodium supply tank load cell, and gas analyzers for 0 '
2 i
H, and water vapor. The output of these sensors is recorded in parallel on 1
strip chart recorders and on a 120-channel digital data acquisition system.
l For test ABS,108 channels are recorded on magnitic tape every 9.0 seconds initially, with decreasing frequency at later times. For times greater than 90 minutes, the measured parameters change more slowly.
Chemical Analyses. Filter papers from cascade impactors, through-the-wall aerosol concentration samplers, and in-vessel fiter clusters are analyzed for 2
sodium by either acid titration or emission spectrometry. Corrections are made to account for background sodium in the filter paper and demineralized water.
Approximately half of the fiter papers are weighed before and after exposure to determine the total mass of aerosol on the paper. Chemical forms of aerosol are detemined by performing a combination of chemical techniques, including x-ray-diffraction, chromatography, and wet chemistry for metallic Na, Na 0 and CO content.
22 2
RESULTS particle Size Measurement. The data from the cascade impactor, the electron j
microscope photographs, and the deposition coupons are compared. The output of the cascade impactor measurements is the AMMD. The settling mean diameter, d,
1 is the output of the deposition coupon and mass balance computations.
Inord$r i
to compare the three methods, the cascade impactor data are converted to settling mean diameters.
If the particle size distribution is log-normal, the aerodynamic settling mean diameter can be calculated frcm the cascade impactor data; j
i where: d aerodynamic settling
=
2 5
AMMD exp(In 0) mean diameter d
=
s 9
0 geometric standard
=
9 deviation The ratios of d measured by the various methods averaged near unity for the i
entire test perlod, hence, the methods are in general agreement. However, a l
significant discrepancy is noted between the cascade impactor measurement and j
the mass balance method for the time period bracketing the source cutoff; the impactor method showed that the particle size increased to very large values i+;,
imediately after source cutoff; the size calculated by mass balance on the i
containment atmosphere showed that the particle size decreased steadily after source cutoff. One explanation is that the impactor data for the time immedi-ately after source cutoff are based on one sample (T316) in which 53", of the
{
aerosol mass is greater than the first stage cutoff diameter. Thus, no infor-mation is available concerning the distribution of particles larger than 20 um.
i
-, - - - _ _ ~ - -, _, _..
...____m.,-.-
i
. The assumption of log-normality may be inapplicable on this sample. The depo-sition coupon data agree well with the other methods except for the samples taken at longer times when the suspended concentration is low. Small quantities of F
resuspended aerosol might have contaminated the coupons and biased the measure-ment toward larger sizes.
Eleven individual code cases are run both prior to the test and after the test.
The post-test runs are done " blind" in that no experimental data on aerosol be-havior is made available; only information on the thermal conditions and aerosol mass is made available.
Pretest predictions are based on a test plan and chosen inputs. After the ABS test is completed, data are transmitted to the participants.
Most of the code cases used identical or very similar numerical values. The exception is the source size for the MAEROS code. The MAEROS uses discretized distributions and, for the post-test case, all of the particles are placed in the bin whose diameter limits ranged from 0.1809 um to 0.2599 um. The use of the one size bin simplified the setup of the code case. The difference in source size do not have a great impact on predicted aerosol behavior. Because of differences in modelling, the same input parameters for different codes may not infer the same calculations.
Some input parameters are as follows:
CHI is a dynamic shape factor that allows the particle drag to be related to Stokes' law for spheres. CHI is a denominator factor in Stokes' law.
Because non-spherical agglomerates settle more slowly than sperical ag-glomerates, CHI is equal to or larger than unity.
GAMMA is a factor which relates the effective collision radius of a par-ticle to the actual particle radius. Because non-spherical agglomerates are able to collide more effectively than spherical agglomerates, GAMMA is equal to or larger than unity.
ALPHA is a density modification factor used to account for the reduced set-tling velocity of agglomerates compared to solid spheres. Generally, it is a numerator factor in Stokes' law and its value is less than or equal to unity.
EPSILON is a gravitational collision efficiency used in the HAA codes.
It relates to the fraction of particles in a swept volume that is captured by a falling particle.
KLYACHK0 is a parameter that allows deviations from Stokes' drag to occur at high Reynolds numbers to be taken into account. This factor becomes important for particles larger than 100 um.
E i
. The valuse of most parameters are selected by comparing code predictions with earlier large-scale sodium fire aerosol tests in the CSTF. The selected vari-ables cover a significant numerical rangg. Diffusional plating boundary layer thickness is assigned values from 1 x 10 m to 1.5 x 10-m, a variation of three orders of magnitude. The small value for delta used in the HAA-3 code re-sults from the assumption that all plateout is caused by diffusion; delta is assignedanempiricalvaluetomatcggreviousexperimgngs. The wall temperature gradient values range from 4.7 x 10 K/m to 1.6 x 10 K/m, almost three orders of magnitude. The assigned values of the ratio of gas to particle themal con-ductivity varies from 0.001 to 0.11, two orders of magnitude.
COMPARISON WITH THEORY AND CODES Some predicted parameters of the aerosol behavior are compared with each other 2
and with the experimenal data.
Suspended Mass Concentration. The concentration predicted by log-normal codes decreases more rapidly than the data after the end of the source period. The discrete codes give good agreement with experiment after the end of source period. All codes agreed with the data within a factor of 2 for the high con-centration period.
Aerodynamic Mass Median Diameter (AMMD). The log-normal codes generally over-predict the AMMD.
The discrete codes underpredict the AMMD. For the first few hundred seconds, all codes underpredict the AMMD, which suggests that the ccde input values for source particle size are too small.
For the MAEROS code, the AMMD is not reported. The values for MAEROS are cal-culated by plotting the size distribution on log-probability paper. The mass median diameter (MMD) obtained from this plot is converted to AMMD; p
1/2 AMMD MMD where: p material density of aerosol
=
=
particle dynamic shape factor x =
The data obtained from measurements using cascade impactors. Since the impactor measurements are not made precisely at the computer code times, the measured values are extrapolated experimental AMMD as a function of time.
A factor of 1.5 is used for evaluating particle size parameters rather than a factor of two as is done for other parameters. Particle sizes and standard deviations do not vary over as wide a range as other parameters and, for this reason, the error band is assigned a value of 1.5.
l j
Geometric Standard Deviation (0_). Compared with cascade impactor data, the log-normal codes underprecict 0 The discrete codes are generally better, but overpredict 0 duringthes00r.ce release period and underpredict 0 after the source ended.9 Values for 0 are not reported for the MAEROS and C@NTAIN codes; the values for MAEROS an8 CONTAIN are calculated by plotting the re-ported size distributions on log-probability paper and using the equation be-low to detennine an approximate value for O ;
g size O
=
g 50% size Although this method is not rigorous for size distributions that are not log.
normal, it is believed to be reasonable.
The experimental measurements are not made at precisely the times reported for the computer codes, the measured values for 0 are plotted and the data are taken fron this curve for the desired times. 9 The number of times that individual code cases preoict the experimental value within a factor of 1.5 is determined. The comparisons may not be relevant for the discrete codes because the distribution is not usually log-normal and a well-defined O may not exist.
g Aerodynamic Settling Mean Diameter (d_).
The aerodynamic settling mean diameter, d
is definea as the diameter of a unit density spherical particle which has a $e,ttling velocity equal to the sedimentation velocity for the whole aerosol.
3 The experimental data are calculated from the rate of change of suspended mass concentration, a knowledge of the source release rate, and Stokes' law for set-tling of unit density spheres. Wall plating is insignificant compared with sed-imentation.
All of the codes are generally able to predict the d better than the AMMO.
ThecodeswhichunderpredicttheAMMOalsooverperdi8tthe0 and vice versa.
Since the settling mean diameter is proportional to both AMM6 and 0 the er-rors in predicting AMMD and 0 tend to compensate and result in a refs,onably good prediction of d The 9 umber of times that individual code cases pre-dictedtheexperimenl$1valuewithafactorof1.5isdetermined.
Leaked Mass. An estimate of leaked mass is not provided by the MAEROS code; the experimental values from which its curve is drawn are calculated assuming the aerosol leak is at a constant rate of 1% per day of the suspended mass.
e
.__,_____,___,m._,______,_
__..__mm.,,___._.,
,. -,, - _ _ _ _ _ ~.. _ - _ _ _., - _..,, _. - _. _., _
. The codes are generally more accurate in predicting leaked mass than they are in predicting suspended mass concentration at discrete times. The number of times that individual code cases predict within a factor of two the experi-mental value is detennined..
Settled Mass. The settled mass is measured experimentally only at the end of the experiment. All of the codes accurately predict the settled mass with nine of the 11 cases predicting within 10g of the test result. Since all the codes show that settling is complete by 10 seconds, the end-of-the test result is 4
reported for times 10 and later.
The number of times that individual code cases predict the experimental value witnin +15% is determined.
Plated Mass. The plated mass is measured at the end of the test. The code pre-dictions range from 1.7% to 1000% of the measured value. Four of the code cases predict the experimental value within a factor of two. The code pre-dictions for plated mass have the most variability in accuracy than for any other parameter except suspended mass concentration.
SUMMARY
The codes in the ABCOVE program which use the log-normal assumption are HAA-3, HAA 4, and HAARM-3. The codes which use the discrete particle size groups are QUICK, MSPEC, MAEROS and CONTAIN. Each type of code has its advantages and disadvantages. Discrete codes are usually more accurate than log-normal codes, but log-normal codes are usually more efficient. No comparison of efficiencies is attempted for test A85.
Test AB5 uses a high sodium spray rate for a sufficiently long time so that high aerosol concentrations are achieved. Agglomeration is important. Previous studies carried out by the ABCOVE participant from Battelle Columbus indicate that the log-normal assumption causes an overestimate of agglomeration. The high aerosol source rate used in test ABS is chosen to explore this apparent limitation of log-normal codes.
A comparison between any two codes is difficult without considering differences in input parameters. Comparing average values of parameters is inadequate, but this is done for suspended concentration and particle size parameters.
The following tendencies for suspended concentration are evident:
Discrete codes tend to predict close agreement or low during the o
source release period compared to data. The concentration decreases at a slightly lower rate than data after source termination.
Log-normal codes tended to precict high during the source period and o
to give a more rapid concentration decrease after the source termina-tion compared with data.
9
o The log-normal codes tend to overpredict the AMMD at all times and underpredict the 0 for all except the latter stages of the source release period,9 compared with data.
o The discrete codes tend to underpredict the AMMD at all times and overpredict O during the source period, compared with data, g
Both discrete codes and log-normal codes predict the aerodynamic set-o tling mean diameter reasonabley well during the source period compared with data. The discrete codes predict best after the source is terminated.
The discrete codes appear to predict aerosol behavior both during the source period and after source termination better than log-normal codes. The discrete codes compare more favorably with the data if plateout had not been overpredic-ted by two cases involving discrete codes. The overpredicted plateout causes an underpredicted concentration during the source period and a slow decrease in concentration after source termination. While the more rapid concentration de-cay predicted by the log-normal codes does not significantly affect the leaked mass calculation (at constant leak rate), it could be important for accident cases where a containment fail after source termination. For example, at 30,000 sec. (8.3 hr.), the mean of the log-normal codes underpredict the con-centration by a factor of 22. However, the concentration at 30,000 sec. is only 0.0004 that of the maximum value during the source period.
Brownian diffusion and thermophoresis is used to model plateout. For the HAA-3 codes, thermophoresis is not included as a deposition mechanism; diffusion is the only plateout mechanism for these two codes. Predicted values of plated mass vary by severaj orders of magnitude. Predigtedvaluesofplatedmass vary from 3.17 x 10 g (QUICK, ORNL) to 1.9 x 10 g (MAPEC, BCL). Part of the discrepancy between the two extremes is due to differences in input values for the thermal gradient. The input values for peak gradients vary by a factor of 10 between the two codes. This difference in gradients is expected to cause as much as an order of magnitude difference, but not an observed difference of a factor of 600.
It is apparent that other differences must also exist. Of in-terest is the difference in predic predicts a plated mass of 2.2 x 10} ions when the same inputs are used; HAARM-3 g, QUICK and MSPEC predicts plated masses, 8 times larger.
Some deviations from the test plan occur and require blind post-test pre-dictions to be made. A significant change from the test plan is the aerosol source rate, which is 0.770 of the test plan value. Other deviations include:
the sodium fraction in the aerosol is 0.574 rather than 0.588; the material density is 2.50 rather than 2.72; and the containment temperature and pressure are slightly lower than planned. The total mass of aerosol released is 0.746 of the test plan value.
4
.g.
Some of the variation arises because the codes do not predict for the entire 4 x 10" sgcond period; thus the number of codes being averaged change with time (after10 sec.). The post-test predictions for suspended concentration, leaked mass, settled mass and plated mass averaged 70% of the pretest values. Pretest predictions might seem to be unnecessary as long as blind post-test predictions are needed. The merit of performing the analyses prior to experiment is that no beneficial knowledge of test results are available when the predictions are made.
FUTURE PLANS The three experiments have been completed. Test AB5 is described in this summa ry. Test AB6 is done on an aerosol system made of two chemical species.
Test AB7 is done to address issues arising from test AB6.
Several chemical analyses have yet to be completed. The results will be reported.
e
%S
Enclosure b
INDEX Experiment (Facility and Instrumentation)
Pace Containment Vessel Arrangement 1
Containment Vessel Properties 2
Sodium Spray System Characteristics 3
Test Conditions 4
Sodium Mass Balance 5
2 Containment Temperature and Pressure, Plot 6
j Source Aerosol Particle Size, Comparison of Methods 7
i Containment Aerosol Particle Size, Comparison of Methods 8
i Aerosol Particle Distribution, Plot 9
Suspended Mass Concentration in Containment, Plot 10 Mass Spectrometric Analyses 11 Experimental fieasurements and Accuracy 12 Comaarison (Data from Experiments and Predictions from Codes)
List of Participants 13 Code Cases for Test ABS 14 Pretest :nput Parameters 15 Pretest Estimates of Temperature and Pressure - Plot 16 Test Ccnditions for Post-Test Code Predictions 17 Code Input Parameters 18 Code Predictions for Suspended Itass Concentrations - Plot 19 Code Predictions for Suspended Mass Concentrations - Plot 20 Correct Predictions of Suspended Mass Concentrations 21 Code Predictions for Aerodynamic Mass Median Diameter, Plot 22 Correct Predictions of Aerodynamic Mass Median Diameter 23 Cade Predictions for Geometric Standard Deviations - Plot 24 Correct Predictions of Geometric Standard Deviations 25 Code Predictions for Aerodynar.ic Settling Mean Diameter - Plot 26 Correct Predictions of Aerodynamic Settling Mean Diameter 27 Code Predictions for Leaked Mass 28 Correct Predictions of Leaked Mass 29 Leaked Mass Predictions 30 Code Predictions for Settled Mass 31 i
Correct Predictions of Settled Mass 32 Code Predictions of Plated Mass 33 Correct Predictions of Plated Mass 34 Code Predictions of Overall Removal Rate - Plot 35 Correct Predictions of Overall Removal Rate 36 Comparison of Pretest and Post-Test Code Predictions 37 Comparison of Log-Normal and Discrete Code Predictions 38 Comparison of Log-Normal and Discrete Code Predictions 39
2 Na FLOWMETER ARGON O
EQUIP SUPPORT BEAMS gOXYGEN (24 NOZZLES)
INTERNAL AEROSOL O
O F
SAMPLERS (TYP. OF 6)
SODIUM THRU THE WALL SUPPLY SAMPLERS (TYP. OF 4)
TANK C
d j
WINDOW (TYP. OF 3)
I3 Y
O ELEV
///,
'////
/
a
/
/
/
/
. g,,i,
Na SPRAY NCZZLES (2) 9
(-) 4.36 m ELEV H
WINOOW (.) 5.52 m ELEV 1.1 m N
MOVIE CAMERA GAS SAMPLE (TYP. OF 5)
O g
/
AND MIRROR CATCH PAN
(-) 8.66 m ELEV
(-)9.51 m ELEV m eum Schematic Elevation View of the CSTF Containment Vessel Arrange-ment for Test A85.
i Page 1 (experiment) ne l
1 l
CSTF CONTAINMENT VESSEL PROPERTIES l
General Cooe ASME Section VIII, 1962 Material Carbon Steel, SA 212-8 Interior paint (phenolic) 0.51 mm Exterior thermal insulation Fiberglass, 25 mm thick, k = 0.0467 W/m*C @ 100*C Design pressure 0.517 MPa at 160*C (75 psig at 320*F)
Nominal leak rate 1.0% per day Dimensions Diameter (ID) 7.62 m Overall height 20.3 m Cylinder height 16.5 m i
}
Enclosed volume 852 m3
[.
Weight, kg (lb)
Top head 8.753 Bottom head 8,753 Cylinder 69,390 Penetrations and doubler plates 10,295 Catch pan 500 Internal components 5,580 Total Weight 103,260 Surface Areas for Heat Transfer, m2 Top head 63.0 Bottom head 63.0 Cylinder 394 Total area for heat transfer to environs W
Internal components 232 Surface Areas for Aerosol Settling, m2 Bottom head 36.7 Catch pan 11.1 Personnel deck 4.2 Internal components 36.2 Total ITT Surface Areas for Aerosol Plating, m2 Vessel shell 520 Internal components 232 Total 7!If i
Thickness for Heat Transfer, mm (Average lumpeo values)*
" Average Thickness =
Top head 18.1 Bottom heaa 18.1 Cylinoer 22.9 Weight Internal components 3.4
.rea) (density of steel)
Page 2 (experiment) s
SODIUM SPRAY SYSTEM CHARACTERISTICS a
humoer of nozzles 2
i hazzle orientation upward Sodium temperature 563*C Pret:sure drop across nozzle 200 kPa (29 psi)
Sodiam spray rate 265 g/s hozzle distance above catch pan 4.2 m hozzle orifice diameter 5.51 mm i
Spray angle 72*
e i
l l
l t
~
l i
l l
l
\\
(
l l
l Page 3 (experiment) i
SUMMARY
OF TEST CONDITIONS FOR TEST ABS i
Test Condition Value Initial Containment Atmosphere Oxygen concentration 23.3 + 0.2 vol %
Temperature, aman
- 29. l *C Pressure 0.122 kPa (17.7 psia)
Dew point -
16 + 2*C Nominal leak rate 1%~per day at 10 psig Sodium Spray Sodium spray rate 256 + 15 g/s Spray start time 13 s Spray stop time 885 s Total Na sprayed 223 + 11 kg Sodium temperature 563*C Spray crop size, MMO 1030 + 50 um Spray size geom. std. dev.
1.'T 0xyoen Concentration Initial 02 concentration 23.3 + 0.2 vol %
Final 02 concentration 19.430.2vol%
0xygen injection start t = 1 minute Oxygen injection stop Total 02 injected t=14minuteg 47.6 std. m Containment Conditions Durino Test Maximum average atmosphere temperature 279'C Maximum average temperature of steel vessel 93.5'C Maximum pressure (absolute) 213.9 kPa (31.0 psia)
Final dew point
-1.5'C(29'F)
Aerosol Generation Generation rate, g/s as aerosol 445 Mass ratio, total to Ng 1.74 Material density, g/W 2.50 Initial suspended concentration 0
Page 4 (experiment)
I
~
l
-l l
SODIUM MASS BALANCE
,2 kg Na kg Na/m Delivered to Containment (a) 222.8 RecoveredfromContainment(b)
Bottom Head 36.7 117.2 3.19 Catch Pan 11.2 42.8 3.81
~
Personnel Platform 4.2 15.4 3.68 All other Surfaces (C) 67.4 Total CV Washes 242.8 Samples 2.2 Total Accounted For 245.0 Difference 22.2 gain (a) Measured by load cell on Na supply tank.
(b)
(c) Measured by water wash volume and Na concentration.
Vertical walls, top dome and internal components.
Page 5 (experiment)
.w4,m a
O%
(i 400 500 y,
..,,ini
'isini sein TEST AB-5 350
- 1) AVERAGE ATMOSPHERE TEMP. - 450
- 2) AVERAGE STEEL TEMP.
300
- 3) CV PRESSURE m
3_
400 I
m 1
n.
y o 250 350 M
E ad O
m m
2 ui g
5M
- M S l
4 0
i i
tu 3
m u
e m
m m
n.
150 O
i 2
2 m
250 Q-g m
F 100 3
i C
200
)
~
2 50 -
150 2
3
I 0
-l 100 l
101 102 103 104 105 106 TIME, SECONDS HEOL 8306 006 5 Containment Temperature and Pressure as a function of Time.
1 i
e l
i" l
t
~
SUMMARY
OF INFORMATION ON AEROSOL SOURCE PARTICLE SIZE MMO e
Methoo (um) g
+0.55
~
1.
Electron microscope sample 0.63 -0.27 2.25 + 0.25 at 1140 s.
Optical sizing.
2.
Cascade impactor sample 1.6 + 0.07 1.9 + 0.2 taken at 63 s.
~
3.
Test plan assumed, based 0.5 1.5 on literature.
Page 7 (experiment)
~~
COMPARISON OF AEROSOL SIZE IN CONTAINMENT ATMOSPHER AS MEASURED BY VARIOUS METH005
- de, Settling Mean Dias. (um)
Ratto of ds Time cascace Dec.
Mass -
covoon Mass Sal.
Mass. 547 measured (s) locactorial Covoonib) 8alance(c)
To too.
To 1=o.
To couoan 3.00(2]I8I 37.1 25.0 0.67 4.00(2) 24.3 60.5 5.00(2) 19.4 2.49 49.6 8.85(2) 18.4 2.55 49.6 1.00(3) 66.2 2.55
.46.1 1.26(3) 129.
0.70 33.8 2.08(3) 33.8 20.9 'I 0.26 I
20.7 0.62 0.61 0.99 3.18 (3) 23.8 18.9 14.5 0.79 0.61 0.77 5.30(3) 14.6 16.3 'I I
11.3 1.12 0.77 0.69 7.50 (3) 9.84 9.04 1.07 (4) 8.32 10.2 7.82 1.22 0.94 0.77 0.92 1.62 (4) 6.98 10.9 'I I
6.45 1.56 0.92 0.59 3.50 (4) 5.07 8.5 'I I
4.39 1.68 0.87 0.52 5.00 (4) 4.65 3.77 0.81 1.00(5) 3.65 2.69
- 0. 74 2.05(5) 2.88 1.98 Mean 0.69 1.17 1.07 0.72 (ajQalculated by Eevetion (11) using Ff9ures 21 and 22 (blFrom Taele 18.
(c)From Table 19.
(e) Average of multiple measurements.(a)Numeers in parenthesis are expone I
5, l
4 1
e e
e Page 8 (experimnet) i n~
i 10.0
.I l
~
TEST AB-6 ELECTRON MICROSCOP5 SAMPLE EM 1 TIME = 1140 SECONDS 1.0
~
f 2
C O
g O
1 8
5 O
Q O
"!o i:
E<
n.
o 0.1 O
O
~'
CMD = 0.17 0.02 pm og = 2.25 0.25 0.01 2
5 10 20 30 40 50 80 70 80 90 96 98 PERCENT OF PARTICLES (BY NUMBER)
LESS THAN STATED SIZE not amma.
Log-Probability Plot of Primary Particle Size Distribution.
Page 9 (experiment)
- F
102
,,)
,..N.O..R.E.M.O.V..A.L.
TEST A86
~
MEAN OF THROUGH THE WALL
~
102 MEASUREMENTS. WITH STD. ERROR
~
Y I MEAN OFIN VE5SEL FILTER
~
, f" I CLUSTER SAMPLES ^
o' l
101.-
j l
J t
i
,0,
~2 h
f 4
4 N.*
U 101 -I 5
i a
u e
o
=\\
w A't FAN ON FOR 24 MIN' E
k FILTERED RECIRC.
- 3. wa7 g-3 is t
s a
N Id k
~
E e
5 5
s M
.I 35 I
101 102 103 104 108 108 TIME (SECONOS)
% maan Suspended Mass Concentration in the Containment Atmosphere.
Page 10 (experiment)
'E N C -
T.
C.
C.
C.
C.
W1 m.
~J w
- V N v v C'
C @ C - O C e
C g
=
% C -
E N.
C.
N.
C.
C.
C.
W1 m
C
@ C @ C - O C t
w @
- V N y
v v
e C
O.
a e
e E
@ C -
w
- m. C.
C.
C.
C.
ui.
J 4
J
@ C m C - C C
- V N v v 2
-+
v1 Ce a3 w (2
g b
E N m -
A
>=
w C.
W1 O.
- m. C.
W1 6m N.
C 7
@ C @ C C C C QJ
- V N v v m>-
C e
w4 h
E m
=
> >=
u aZ
@ C 4W c.N Z
m -
E
@ e -
M.
C.
O.
e.
C.
N u)
==
a e
U4 e
e m C m C C C C
. - - > =
- v N v v 2Z m
+
>- C wu m
3 5 e=
h E
N N -
uw 6
N.
O.
- m. C.
- m. C.
u1 w
C 4
C C G3 C C' C C A
C N v N v v o
e m
g E
z
-C N
=> m E ui C -
m.
C.
C.
C.
C.
ui.
G3 C C C - C C
- V G3 v v
+
C E
e m -
-a N.
O.
- m. C.
e.
C.
e.
e.
N O @ C C C C w@
N y N v v e
e-
.m m ~
~C
=C
~
v C
C z Z v < uu Page 11 (experiment)
L_'
EXPERIMENTAL MEASUREMENTS AND ACCURACY he. of No. of Standard Measurement locations Times
_frror Method Suspended aerosol mass concentration 6
1
- 255 In-vessel filter clusters Suspended aerosol mass concentration 4
52
- 151 Through-the-mail samplers l
Aerosol particle site and 's Aerosol particle site (actuall 3
16
- 20E Cascade impactor l
,2 N/A flectron microscopy, slalag Aerosol particle shape y
1 2
N/A flectron microscopy Aerosol chemical compositten 1
5 N/A Aerosol lastantaneous deposition rate Various; Chemistry Lab.
1 9
+ 205 Through-the-wall coupons N
Integral settled mass / unit area 23 1
- 101 Fall-out pans lategral plated Na on vessel walls per unit area m
2 I
- 205 Vessel well smears Aerosol settled / unit area during Na spray period 1
2
- 251 E
,Na mass deposited in catch pan Special samplers s
1 1
- 10E Wash and analyse for Na y
total settled Na mass 1
I
- 10E Wash vessel floor lotal Na wall platenut i
1
- 301 8
lemperature of containment atmosphere Wash vessel walls 28 (a)
- 21 Thermocouples temperature of vessel surf ace 18 la)
- 25 Thermocouples temperature of Na sprayed 2
(a)
- 21 Therwoccuples Containment pressure I
la)
+ 15 pressure Transducer Containment 0 concentration 2
5 (a)
- 21 On-line 02 *"*IY
l Containment H c ncentrations 2
5 (a)
- 205 On-line H Containment moisture concentration 2 '"* U
2 la)
- 308 Convection velocity On-line humidity analyser 1
5 44.known Anemometer Sodium Spray mass flow rate I
la)
- 105 Magnetic flommeter and load cell Overall Na mass balance N/A 1
- 105 Weighing, Washing, volume, chen. analysis, calculation (a) Continuous
.i l
LIST OF PARTICIPANTS FOR ABCOVE TEST ABS Particicant Affiliation Aodress Emil Gluekler General Electric Company P.O. Box 5020 310 DeGuione Drive Sunnyvale, CA 94086 R. K. Hilliard Hanford Engineering P.O. Box 1970 Development Laboratory Richland, WA 99352 Safety Systems Development Hans Jordan Battelle Columbus 505 King Avenue Laboratories Columbus, OH 43201 T. S. Kress Oak Ridge National P.O. Box X Laboratory Oak Rioge, TN 37830 K. K. Murata Sandia National P.O. Box 5800 Laboratories Albuquerque, NM 87115 J. M. Otter Rockwell Internat'ional 8900 De Soto Avenue Energy Systems Group Canoga Park, CA 91304 M. G. Piepho Hanford Engineering P.O. Box 1970 Development Laboratory Ricnland, WA 99352 Containment Systems Analysis Page 13 (comparison)
CODE CASES FOR TEST A85 Coce Case
-No.
Code User
'~
l HAA-38 GE 2
HAA-3C HEDL/SSD 3
HAA-4 ROCXWELL/ESG 4
HAARM-3 HEDL/SSD 5
HAARM-3 BCL 6
HAARM-3 ORNL 7
QUICK BCL 8
QUICK ORNL 9
MSPEC BCL 10 MAEROS HEDL/CSA 11 CONTAIN SNL i
4 Page 14 (comparison)
PRETEST INPUT PARAMETERS TRANSMITTED TO CODE USERS BY THE TEST PERFORMER C
Parameter Value 3
Source rate, g/s em 6.8 x 10~7 Source 301 radius, um 0.25 Source sigma, e 1.5 g
Initial aerosol concentration 0
Source cutoff time, sec. 900 Maximum time, days 5
Leakage rate, %/ day 1.0(constant) 2 5
Settling area, cm 8.8 x 10 2
6 Plating area, em 7.5 x 10 3
8 volume, em 8.5 x 10 3
Density of aerosol, g/cm 2.72 Temoerature of atmosphere Figure 26 Temperature of CV walls Figure 26 Pressure Figure 26 4
\\
Page 15 (comparison)
I
e (etsd) d
.E m
a a
a e
s"A
.E i
=-
I i
t 2
l I
l E
a I
I r
e
-5
=
- n.
/
s, n
I
-2 e
n p.
a-j
/
E 5
/
o 0
/
z e
7 e
i-
% M i.
.\\
's E
e 1
5 4
\\
3 b
E
\\
$s I
O
\\
m E
A P
's w
C g
g s
s 1
x ss 1
a 3
\\
t s
m N
I
,1 R
R R
R H
E 8-10 11 Page 16 (comparison)
TEST CONDITIONS TRANSMITTED TO CODE USERS FOR USE IN MAKING BLIND POST-TEST PREDICTIONS Post-Test Pretest Measured Parameter Test Plan or Estimateo Pressure Drop Across Spray Nozzle (psi) 40 29 Sodium Spray Sodium Spray Rate (g/s) 340 256 SprayStartTime(s) 0 13 Spray Stop Time (s) 900 885 Total Na Sprayed (kg) 306 223 Sodium Temperature (*C) 560 563 SprayDropSize,MMO(um) 980 1030 Spray Size Std. Dev. (a )
1.4 1.4 g
Aerosol Generation Mass Ratio, Total to Na 1.70 1.74 Generation Rate (g/s) 578 445 3
Material Density (g/cm )
2.72 2.50 Initial Time (s) 0 13 Cutoff Time (s) 900 885 Initial Conc.
0 0
- i Source Mass Median Radius (um) 0.25 0.25 Source Sigma 1.5 1.5 Page 17 (comparison)
,_,w,
CODE INPUT PARAMETERS RELATED TO THE AEROSOL SOURCE Material Densig;J r
Source Size Source Rate Code User (g/cm dse (pm) og (No./s cm3)
(q/s m3)
HAA-38 GE 2.72 (2.5)(8) 0.5 (0.5) 1.5 (1.5) 8.0 E6 (6.69 E6)(b) u llAA-3C HEDL/550 2.72 (2.5) 0.5 (0.5) 1.5 (1.5) 8.0 E6 (6.69 E6)
IIAA-4 RI/ESG 2.72 (2.61) 0.5 (0.5) 1.5 (1.5) 8.0 E6 (6.69 E6)
E ilAARM-3 ' HEDL/SSD 2.72 (2.5) 0.5 (0.5) 1.5 (1.5) 8.0 E6 (6.69 E6)
IIAARM-3 BCL 2.72 (2.5) 0.5 (0.5) 1.5 (1.5) 8.0 E6 (6.16 E6)
L 11AARM-3 ORNL 2.72 (2.72) 0.5 (0.5) 1.5 (1.5) 8.0 E6 (6.16 E6)
QUICK BCL 2.72 (2.5) 0.5 (0.5) 1.5 (1.5) 0.68 (0.52)
QUICK ORNL 2.72 (2.72) 0.5 (0.5) 1.5 (1.5) 0.68 (0.52)
MSPEC BCL 2.72 (2.5) 0.5 (0.5)
- l.5 (1.5) 0.68 (0.52)
MAEROS HEDL/CSA 2.72 (2.5) s1.5 (0.21)
IIIC) (1.0) 0.68 (0.52)
CONTAIN SNL 2.72 (2.5) 0.5 (0,5) 1.5 (1.5) 0.68 (0.52)
(a)Lx-parenthesis nimiber (b)8.0 [6 means 8.0 x 10g 'are pretest; mediers within parenthesis are post-test.
(c)A histogram distribution was used.
1 e
.e t
3 la i
M
[g k
~~
- t. AB-5 TEST RESULT I.
- 2. HAA-38 / GE
=
E
- 3. HAA-3C / HEDL
- 4. HAA-4 / RI I
r
- 5. HAARM-3 / HEDL y
'\\
- 6. HAARM-3 / BCL i
18 ; j
- 7. HAARM-3 / ORNL M
i
- 8. QUICK / BCL i
m
- 9. QUICX / ORNL 4
O
- 10. MSPEC / BCL
}
I 1i. MAEROS / HEDL
.]
s Ig
- 12. CONTAIN / SNL 3
m t
4 (k:
2 8
e IS #
8h j
a I,
4 la 2
o 5
7 6
12 o
n 19 i
'N i
e
\\~
z 4
a 5
~
d N
b lg c
e 3
i l
.\\
3 II -4E s
11
.g gg 16 18 IB '
IB '
2 4
TN, SE0005 Plot of Code Predictions of Suspended Mass Concentrations for the Entire Test Period.
Page 19 (comparison)
3la L.
f-u r
E b
d u
5 m
fa 4
7 N'.a:
a i
>3 m
2'
\\
~
{
4 3
1.\\
I.
s g
~
li 12
\\
e s
s
- i2 o
2
{ \\g' s s u
1 6,8 %
g m-1 y
gg /
a e
Q tu
- 1. AB-5 TEST RESULT
- 7. HAARM-3 / ORNL Sw
- 2. HAA-38 / GE
- 8. M cK / BCL f
c.g
- 3. HAA-3C / HEDL
- 9. M CK / ORNL
- 4. HAA-4 / RI
- 10. MSPEC / BCL 6
- 5. HAARM-3 / HEDL i1. HAERCS / HEDL
- 6. HAARM-3 / BCL
- 12. CONTAIN / SNL i
i i
i i
gg 188 288 388 48 515 m
788 m
a tm TM, M l
l Plot of Code Predictions of Suspended Mass Concentrations During Source Period.
l Page 20 (comparison)
i CODE CASES WITH CORRECT PREDICTIONS FOR SUSPENDED MASS CONCENTRATION Time Code Case (a,d)
(sec) 1 2
3 4
6 o
/
8 9
10 il ICI 100 X
X X
X X
X X
X X
X X
300 X
X X
X X
X X
X X
500 X
X X
X X
X X
X X
X M5 X
X X
X X
X X
X X
1(3)IDI X
X X
X X
2(3)
X X
5(3)
X X
X X
1(4)
X X
X X
X X
3(4)
X X
X 1(5)
X X
X X
4(5)
X TOTAL CORRECT 6
5 5
3 3
3 7
7 6
9 9
(a)See the curve / code identification below.
(b) Number in parenthesis is exponent of 10.
(c)An X indicates that code predicted within a factor of 2 for the indicated time.
(d)A dash indicates that no data were submitted by the code user.
CURVE / CODE IDENTIFICATION Code Code Code Number Code Number Code Number Code 1
HAA-3B 5
HAARM-3 9
MSPEC 2
HAA-3C 6
HAARM-3 10 MAEROS 3
HAA-4 7
GUICK 11 CONTAIN 4
HAARM-3 8
QUICK Page 21 (comparison)
1 gg s
L
- 1. AB-5 TEST RESULT j
l.
- 2. HAA-38 / GE j
g
- 3. HAA-3C / }EDL w
- 4. HAA-4 / RI i
i C
- 5. HAARM-3 / HEDL E
- 6. HAARM-3 / BCL i
7
- 7. HAARM-3 / ORNL i
E
- 8. QUIOC / BCL
- 9. QUICK / ORNL 2
- 10. MSPEC / BCL c
[
[
[
t1. MAEROS / tEDL 8
g 4
- 12. CONTAIN / SNL m
O 1
z 18 I
l
,1
]
~
3
[
12
]
c I
5 12 l
e W
11 E
[5
}
1 I
(/
O Il 6
k l s' f
/ 18 o
O E
5 2h 4
3 1...
t 8
4 3
8 is 18 18 18 18 IN,SECQOS Plot of Code Predictions of Aerodynamic Mass Median Diameter.
1 Page 22 (comparison)
CODE CASES WITH CORRECT PREDICTIONS FOR AERODYNAMIC MASS MEDIAN DIAMETER L'
Time Code Case (a,d)
(sec) 1 2
3 4
5 6
7 8
9 10 11 100 300 X(c)
X X
500 X
X X
X
~
M5 X
X X
1(3)ID)
X X
2(3)
X X
X X
i 5(3)
X X
X X
X 1(4)
X X
X X
X X
3(4)
X X
X X
X X
X i(5) x x
x x
x x
4(5)
X x
x x
x x
TOTAL CORRECT 0
0 0
5 6
1 5
7 3
9 9
(a)See the curve / code identification below.
(b) Number in parenthesis is exponent of 10.
(c)An X. indicates that code predicted within a factor of 1.5 for the indicated time.
(d)A dash indicates that no data were submitted by the code user.
CURVE / CODE IDENTIFICATION Code Code Code Number Code Number _
Code Number Code 1
HAA-3B 5
HAARM-3 9
MSPEC i
2 HAA-3C 6
HAARM-3 10 MAEROS 3
HAA-4 7
QUICK 11 CONTAIN 4
HAARM-3 8
QUICK Page 23 (comparison)
7
- i altl[
' i i 6tlll
' i i s t lll i a 11ll
-t. AB-5 TEST RESULT
- 2. HAA-38 / GE 6
12
- 3. HAA-3C / FEDL
- 4. HAA-4 / RI
- 5. HAAR?i-3 / HEDL
- 6. HAARft3 / BCL 5
- 7. HAAR?f-3 / ORNL
~
8
- 8. QUICK / E
- 9. QUICK / ORM.
- 18. MSPEC / BCL b
4 1
H 0
- 11. MAEROS / PEDL x
j$
$2. EM / E
/
I 1
3 4
7s 18 3
1 2
5 6
7' 2,4 11 8,8
'N 3
1
,, i t ill
.,,iim
..iiim r ill 2
3 4
3 13 18 18 18 18' TM, SB3DS
+
Plot of Code Predictions of Geometric Standard Deviation.
1 Page 24 (comparison)
=
i e
CODE CASE.", WITH CORRECT PREDICTIONS FOR GEOMETRIC STANDARD DEVIATION Time Code Case (a,d)
(sec) 1 2
3 4
5 6
7 8
9 10 il L
s-100 300 X
X X
X X
X X
X ICI 5 00 X
X X
X X
X X
X X
885 X
X X
X X
X X
X X
X 4
1(3)ID)
X X
X X
X X
X X
X X
2(3)
X X
X X
X X
5(3)
X X
X X
X X
1(4)
X X
X X
X X
X 3(4)
X X
X X
X X
1(5)
X X
X X
X X
4(5)
X X
X X
X X
TOTAL CORRECT 3
3 4
4 10 3 10 10 10 10 8
(a)See the curve / code identification below.
(b) Number in parenthesis is exponent of 10.
(c)An X indicates that code predicted within a factor of 1.5 for the indicated time.
(d)A dish indicates that no data were submitted by the code user.
CURVE / CODE IDENTIFICATION l
Code Code Code Number Code Number Code Number Code 1
HAA-3B 5
HAARM-3 9
MSPEC i
2 HAA-3C 6
HAARM-3 10 MAEROS 3
HAA-4 7
QUICK 11 CONTAIN 4
HAARM-3 8
QUICK Page 25 (comparison)
e a e as
- l. A8-5 TEST RESULT
- 2. HAA-33 / GE w
- 3. HAA-3C / HEDL
.v t-W
- 4. HAA-4 / RI 8
(
- 5. HAARM-3 / HEDL
- 6. HAARM-3 / BCL a:
- 7. QUICK / SCL o
9
- 8. MSPEC / SCL y
1
- 9. CONTAIN / SNL s
d 7
w E
k 4
H o
2 IO g
5
]
.I.
3 e
M 8
5
[
1 N
.I I
\\
i o
4 g
4 1
2 o
ew 4
I 3
2 3
.i 2
3 4
3 18 tg 18 18 18 '
TLY, SECNS Plot of Code Predictions of Aerodynamic Settling Mean Diameter.
Page 26 (comparison)
CODE CASES WITH CORRECT PREDICTIONS FOR AERODYNAMIC SETTLING MEAN DIAMETER Time Code Case (a,e)
(sec) 1 2
3 4
5 6
7 8
9 10 11 100 ICI 300 X
X 500 X
X X
X
-X X
X 885 x.x x
x x
x-x 1(3)(b) x x
x x
x 2(3)
X X
X X
X X
X X
5(3) x x
x x
x x
x 1(4) x x
x x
x 3(4)
-- -- X X
X 1(5) x x
4(5) x x
TOTAL CORRECT 5
5 5
5 4
(d) 5 (d) 9 (d) 10 (a)See the curve / code identification below.
(b) Number in parenthesis is exponent of 10.
(c)An X indicates that code predicted within a factor of 1.5 for the indicated time.
(d)Not reported.
(e)A dash indicates that no data were reported at that time.
CURVE / CODE IDENTIFICATION Code Code Code Number Code Number Code Number Code 1
HAA-3B 5
HAARM-3 9
MSPEC 2
HAA-3C 6
HAARM-3 10 MAER05
-3 HAA-4 7
QUICK 11 CONTAIN 4
HAARM-3 8
QUICK Page 27 (comparison)
40
'a 6illjj
' + 41 lll
' + i : 11ll
' + iillj 3fL l
L 5
J l
d
- 1. AB-5 TEST RESULT a
- 2. HAA-3B / GE
- 3. HAA-3C / HEDL
- 4. HAA-4 / RI U
- 5. HAARM-3 / FEDL
~
j 3.
I
- 6. HAARM-3 / BCL
- 7. HAAR&3 / ORtt
- 8. QLrICK / BCL
- 9. QUICK / ORNL
- 10. MSPEC / BCL E
3.
I1. CONTAIN / SNL I
o 3
w E
15 7'
W J
2[!s f
'll 4 18.
9 8
's,
5 18 g
,,,itill
,iitill
.. i i11ll
. i i iIll 2
3 4
5 I
gg Il 18 18 gg lW, SBMS Plot of Code Predictions of Leaked Mass.
Page 26 (comparison)
CODE CASES WITH CORRECT PREDICTIONS FOR LEAKED MASS Time Code Case (a)
(sec) 1 2
3 4
5 6
7 8
9 10 li 100 X(c)X X
X X
X X
X X
X 300 X
X X
X X
X X
X X
X 500 X
X X
X X
X X
X X
885 X
X X
X X
X X
X X
1(3)(b)
X X
X X
X X
X X
2(3)
X X
X X
X X
X 5(3)
X X
X X
X X
X 1(4)
X X
X X
X X
X 3(4)
X X
X X
X X
X 1(5)
X X
X.
X X
X X
4(5)
X X
X X
X X
X TOTAL CORRECT 11 11 11 4 5
11 11 11 2 (d) 11 (a) See the curve / code identification below.
(b) Number in parenthesis is exponent of 10.
(c)An X indicates that code predicted within a factor of 2 for the indicated time.
(d)Not reported.
CURVE / CODE IDENTIFICATION 4
Code Code Code Number Code Number Code Number Code 1
HAA-3B 5
HAARM-3 9
MSPEC 2
HAA-3C 6
HAARM-3 10 MAEROS 3
HAA-4 7
QUICK 11 CONTAIN 4
HAARM-3 8
QUICK Page 29 (comparison)
5
SUMMARY
OF LEAKED MASS PREDICTIONS Code User Ratio: Code to Test HAA-3B GE 1.05 HAA-3C HEDL/SSD 1.17 HAA-4 RI/ESG 0.84 HAARM-3 HEDL/SSD 2.26 HAARM-3 BCL 0.43 HAARM-3 ORNL 1.07 t
QUICK BCL 0.51 QUICK ORNL 0.56 MSPEC BCL 0.45 MAEROS HEDL/CSA
'CONTAIN SNL 0.98 AVERAGE 0.937
~.
k Page 30 (comparison)
a i t illl intilj
- i 111ll
- i n it(
a
- 1. AB-5 TEST RESULT
- 2. HAA-38 / GE
- 3. HAA-3C / HEDL l
g
- 4. HAA-4 / RI j
[
- 5. HAARM-3 / HEDL
- 6. HAARM-3 / BCL
- 7. HAARM-3 / ORNL
- 8. QUICK / BCL t
- 9. CUICX / ORNL mz 53
- 10. MSPEC / BCL
[
I1. MAEROS / HEDL e
- 12. CONTAIN / SNL o
a Q
J l
I n'
49
-7 2
a 4'
n s
s s
.y
/
0-1 s
s 4
8 3g IV 3'
J
'T r
6' I
0 318.
a s
3 8
1 3.
3 18
{
la ti l
.,,itill
,, irtill
,,iriifl t
3 4
5 s
jg 13 18 13
- g l
l E, E Plot of Code Precictions of Settled Mass.
l l
Page 31 (comparison) l
i CODE CASES WITH CORRECT PREDICTIONS FOR SETTLED MASS Time Code Case (a)
(sec) 1 2
3 4
5 6
7 8
9 10 11 1(4)(b,c)
XId)X X
X X
X X
X
~ (4)
X X
X" X
X X
X X
X l
3 1(5)
X X
X X
X X
X X
X J'
4(5)
X X
X X
X X
X X
X TOTAL CORRECT 4
4 4
4 4
4 0
4 0
4 4
(a) See the curve / code identification below.
(b) Number in parenthesis is exponent of 10..
(c) Experimental result not available at t < 10.
4 (d)An X indicates that code predicted within *l5% for the indicated time.
~
J e
t CURVE / CODE IDENTIFICATION Code Code Code Number Code Number Code Number Code 1
HAA-3B 5
HAARM-3 9
MSPEC 2
HAA-3C 6
HAARM-3 10 MAEROS 4
3 HAA-4 7
QUICK 11 CONTAIN 4
HAARM-3 8
QUICK Page 32 -(comparison)
- I
I lg 18 5
8 I
II 4
5 12 6
1 s
s x
m'3
\\
\\
\\
/
\\
\\
4
\\
-lg y
l
/
/
2' 3
g Il i
j e
7 3
E 18 i
1 I
9 4
o i
tue J
t A 18 i
i
[
- 1. AB-5 TEST RESULT
- 7. HAAR!+-3 / ORNL j
- 2. HAA-38 / GE
- 8. GJICK / SCL i
- 3. HAA-3C / HEDL
- 9. GJICK / ORNL
- 4. HAA-4 / RI
- 10. MSPEC / BCL IO l.i
- 5. HAAR!t-3 / lEDL
- 11. MAEROS / HEDL
[
- 6. HAARM-3 / BCL
- 12. CONTAIN / SNL j
i j
.l 2
3 4
8 8
33 13 18 18 18 DE,SECQOS Plot of Code Predictions of Plated Mass.
I i
l Page 33 (comparison)
I L
CODE CASES WITH CCRRECT PREDICTIONS FOR PLATED MASSLD)
Predicted Plated Code Mass Within Factor of 2 Case (a)
YES NO 1
X 2
X 3
X 4
X S
X 6
X 7
X 8
X 9
X 10
.X 11 X
(a) See the curve / code identification below.
(b)Forend-of-testconditions.
CURVE / CODE IDENTIFICATION Code Code Code Number Code Number Code Number _
Code 1
HAA-3B 5
HAARM-3 9
MSPEC 2
HAA-3C 6
HAARM-3 10 MAEROS 3
HAA.4 7
CUICK 11 CONTAIN 4
HAARM-3 8
QUICK Page 34 (comparison)
I
18
- 1. AB-5 TEST RESULT 7
6
- 2. HAA-3C / HEDL i!
- 3. HAA-4 / RI I
- 4. HAARM / HEDL l
- 5. HAARM-3 / BCL
- 6. HAARM-3 / CRNL I
[3
- 7. QUICK / BCL l
9 IO.2 i
- 8. QUICK / ORNL o
8 1,
g
- 9. HSPEC / BCL
- 10. CONTAIN / SNL 1
b
[
ki 8
i 5
6 ew i
g il -3 N
i ss W
ht 2
<E i
2
{\\
i S
I w
'N 4
i E
6 N
ll :
7 j
s d
s
.g.
}9 5
4 2
10 4
8 gg
.x..
2 3
4 5
8 1g '
is Ig Ig gg TDE, SECQOS Plot of Code Predictions of Overall Removal Rate.
Page 35 (comparison)
CODE CASES WITH CORRECT PREDICTIONS FOR REMOVAL RATE Time Code Case (a,e)
(sec) 1 2
3 4
5 6
7 8
9 10 11 100 i
300 X
500 X(c)X X
X X
X X
885 X
X X
X X
X X
1(3)(b)
X X
X X
X X
2(3)
X X
X X
X X
X f
5(3)
X X
X X
X X
X X
1(4)
X X
X X
X 3(4)
X X
X X
X X
1(h)
X X
X-4(5)
X X.
X X
TOTAL CORRECT (d)5 4
4 9
1 6
8 8
(d) 9 (a See the curve / code identification below.
(b) Number in parenthesis is exponent of 10.
(c)An X indicates that code predicted within a factor of 2 for the indicated time.
(d) Mot reported.
(e)A dash indicates that data on removal rate were not reported at this time.
CURVE / CODE IDENTIFICATION Code Code Code Number Code Number _
Code Number Code 1
HAA-3B 5
HAARM-3 9
MSPEC 2
HAA-3C 6
HAARM-3 10 MAEROS l.
3 HAA-4 7
QUICK 11 CCNTAIN 4
HAARM-3 8
QUICK Page 36 (comparison)
l 1
COMPARISON OF PRETEST AND BLIND POST-TEST PREDICTIONS Geometric Mean of All Codes, Ratio of Post-Test / Pretest Time susp.
M M
M (s)
Conc.
APet0
- g ASMO L
S P
R 1(2) 0.67 0.77 0.81 1.16 0.58 0.49 0.44 0.96 3(2) 0.76 1.00 1.00 2.08 0.70 0.69 0.62 0.96 5(2) 0.75 0.93 1.02 1.45 0.72 0.84 0.56 0.99 8.85(2) 0.73 0.93 1.09 1.43 0.70 0.80 0.64 1.06 1(3) 0.64 0.87 1.02 1.19 0.72 0.86 0.62 1.89 2(3) 0.68 1.11 1.03 0.92 0.69 0.55 0.61 0.70 5(3) 0.70 1.11 1.03 0.97 0.68 0.77 0.61 1.33 1(4) 0.54 1.24 1.00 0.89 0.68 0.77 0.61 1.58 3(4) 0.75 0.93 1.05 1.07 0.68 0.77 0.61 0.95 1(5) 0.80 0.94 1.03 1.00 0.68 0.77 0.61 0.95 4(5) 1.41 0.89 1.06 1.60 0.68 0.77 0.61 1.28 Page 37 (comparison) ce
[,... -
I COMPARISON OF SUSPENDED CONCENTRATIONS PREDICTED l
BY LOG-NORMAL AND DISCRETE CODES Time Geometric Mean Concentration (g/m3)
Measurea (s)
Log-Normal Codes Discrete Codes Concentration (g/cm3) 100 4.4(1) 4.4(1) 3.7 (1) 300 1.1(2) 8.7 (1) 1.4 (2) i 500 1.4 (2) 8.1(1) 1.1 (2)
{
885 1.4(2) 8.2 (1) 1.1 (2) 1000 6.0(1) 3.8(1) 6.5(1) 2000 9.0(-1) 3.9 (0) 6.8(0) 5000 2.5(-1) 9.3 (-1) 1.2(0) 10000 2.7(-2) 3.5(-1) 3.8(-1) 30000 2.1(-3).
7.9(-2)-
4.7 (-2)
Page 36 (comparison)
>r
R i
e COMPARISON OF PARTICLE SIZE PARAMETER PREDICTED BY LOG-NORMAL AND DISCRETE CODES i
i Geometric Mean of.
Geometric Mean Geometric Mean Aerodynamic Settling Mean Time of AMMD (pm) og Olaneter (wn)
(sec)
Log-N Discrete Test Log-N Discrete Test Log-N Discrete Test I
jg 100 1.10 1.64 -
4.0 1.58 1.57 2.9 1.34 2.02 h;
300 9.16 3.97 7.2 2.66 4.00 3.55 20.6 42.7 26.
500 13.0 3.41 5.6 3.'26 4.42 3.05 42.0 45.0 50.
,,l 885 14.2 3.46 5.5 3.36 4.43 3.00 45.9 47.1 50.
Ih 1000 30.5 5.97 13.0 2.78 4.28 3.50 72.8 52.6 47.
j[
2000 17.2 5.67 9.4 1.85 2.96 3.3 23.9 21.7 20.
EL 5000 11.4 4.65 6.2 1.54 2.29 2.55 14.0 11.9 11.
10000 8.81 3.99 5.0 1.40 1.99 2.2 10.4 7.0 7.8 30000 5.47 3.03 3.4 1.27 1.42 1.97 6.8 4.4 4.7 me
/
Report 11 A Comparison of Aerosol Behavior Codes J. A. Gieseke Battelle Columbus Laboratories t
e t
+
0 e
6
I O
A COMPARISON OF AEROSOL BEHAVIOR CODES by J. A. Gieseke Battelle Columbus Laboratories There are a number of mechanistic aerosol behavior codes developed to evaluate the transport, deposition, and leakage of fission products in reactor containments. Most of the. codes have evolved from models designed for analyses of aerosols in LMFBR systems and therefore include to varying degrees, mechanisms added to extend their applicability to LWR conditions. Among the codes available for containment analyses are several which were developed for fission product behavior in primary systems. These are also applicable to containment conditions through their ability to analyze aerosol behavior in general and in some cases have been used for containment analyses. The CORRAL code, even though historically important, has not been included in the comparison since it is largely empirical.
The major technical features or aerosol behavior mechanisms included in all of the codes are listed in the attached table where the inclusion or exclusion of each mechanism in each of the various codes is noted. A brief discussion of the major features and differences among the codes is provided to supplement the tabular information.
There are two major classes of codes, those developed for the containment and those developed for the primary system (RAFT, RETAIN.
TRAP-MELT 2). The RETAIN code has been extended for use in the contain-ment and TRAP-MELT C is an adaptation of the TRAP-MELT 2 code for contain-ment conditions. Primary system codes include fission product vapor interactions with aerosols and surfaces and do not consider water conden-sation, although the TRAP-MELT C code is an exception in this regard.
The containment codes are focused on aerosol behavior under containment conditions with steam condensation on aerosols and surfaces. The MATADOR code, which is a fairly mechanistic code for risk assessment and is intended to be a replacement for CORRAL, treats noble gases and fission product vapor (assumed to be 1 ) in addition to aerosols.
2 A significant difference in representation of the aerosol size distribution exists among the codes with the ABC-3C. HAA-4A, HAARM-3, MATADOR, and RETAIN codes assuming a log-normal distribution. The log-normal distribution has been shown, by comparison with experiments,
2 to overestimate aerosol depletion rates, and the codes using a discretized or sectionalized size range show much better agreement with experiments and are therefore much preferred.
Related to the size representation is the question of aerosol composition as a function of size. Only the QUICXM and MAEROS codes compute composition by size range, with other codes following only water /
solid aerosol ratios with size (NAUA-4. TRAP-MELT C) or assuming equal composition for all particle sizes.
It should be noted that the QUICXM code is the only code that tracks aerosol material density and shape factors as a function of size (composition). The theoretical basis for size dependent composition is quite appealing and significant effects on airborne materials are computed for some circumstances; however, there is not yet a definitive experimental basis for supporting the code predictions. Experimental work in this area is progressing and the second multicomponent ABC0VE experiment was matched fairly well in aerosol composition by the QUICXM code. However, the use of codes treating size dependent composition should proceed cautiously until considerably more experimental data are available.
There are some situations in the progression of accidents where homogeneous nucleation of aerosols may be important.
The first situation would involve rapid steam injection into the containment with low aerosol concentrations such that the gas becomes supersaturated with water vapor and homogeneous nucleation would occur in parallel with condensation on pre-existing particles.
The NAUA-4 Mod code permits homogeneous nucleation of water to take place under these conditions. A second situation could exist for BWR cases in which decay heating in the containment is considered, the drywell volume may reach temperatures at which many fission products are vapors and the timing is such that previously existing aerosol species have been swept from the volume. As fission product vapors move from the drywell volume or as the drywell temperature drops, homogeneous nucleation is expected to occur. Similar situations can exist within the primary system. Another situation can exist where effects of hydrogen burning may be such to produce vaporized fission products while there is a low non-volatile aerosol component airborne. Again homogeneous nucleation may be important as cooling occurs.
Only the RAFT code includes homogeneous nucleation. However, it is unlikely that truly homogeneous nucleation can occur in the high fon concentrations expected in a radiation environment so the issue and inclusion of homoge-neous nucleation is subject to further evaluation, k
l o
3 In general, the aerosol behavior codes using discretized size i
distributions have been found to predict experimental results with good
{
accuracy (to within a factor of 3 for airborne mass concentrations) in j
blind predictions. A sumary of predictions with some of the codes compared with ABCOVE experiment AB-5 are shown in the attached figure.
The agreement between experiment and predictions for codes using a discretized size distribution is quite encouraging and suggests that j
major processes are being modeled successfully.
t j8 I
l 1
k i
I i
l I
i t
I n
a t.
)
j
[
~
no.
n. $_.
o =
o o
.e c
.e a
- ggy 3K 3
c n
- a
.o.
.c ecc
=
c 3
.a 1
3 _
=
a****
2 a
.a 3
I n.
- oa 3
-e
-2 c.
~
-m
-n
-1 2
1.g g**
, a n
o g
~
a.
3 8 3' t
.R
--+
a ggo-m.a o
.a.
TEc 23 a
g 8
o
.-at.
=
gg.,
. =
MC.3c o.,
)
44 23=g
==
. =
==
MnOSIM (m) 1.1. gt
.,. 4
=
=
AEa050L (France) g 2 3.~.
2o vs CONTAIR/MEROS
=
= e
~
. =
n >.
o 7
mm NAA-44 (AI) b
. =
4:
-e o
. =
. =
. =
. =
MAAAM 3 (KL) n O
4.
o e
a.
- =
MTA004 (KL) m m
3-2 M,UA-4 M O
=
=
. =
m o
a n
x m.
. =
=
PAADISEE0 3 (FAG) 3 c
c..
2 g o..-
~
, E ; **.1 u.
. =
. =
qugcg (gt) 3
- a. 2. 4 3 3
-=
un
..g,g-
=
- ~
QutCKM (KL)
.4s
- o 2...,
E't$&*
=
=
=
~.
=
~ =
- t**-
yyy{w) 2aoE'. 3 o
.,. 2..,
3
=
a.
. = - =
. =
=
=
i yyagn gggoa}
. o.a.
1 1
o2g I
." 2.
=
. =
. - - =
TAAP-MELT 2 (KL)
.-2E*
=
=
e.
. =.. = = = -
inAP-E LT C ( KL) 4
=
e
. ~..
. >.. =. =..
=
ggg ggg}
o_
3
==
O g
5 F
W I 5
yy a y g,
l 4
2 6
\\
8 18 7 ti
\\ fxg '
9 g n s-n s
7 9
8 9
18 5,7
-- g 4
8 5
8 -l!
H 6
g 6
w 3
4 Q.
X 4
o Il 5
i r
W l
l 8
u S ;g -1 TEST A8-5 F
m
- 1) HAA-38 GE
- 7) CUICX BCL
- 2) HAA-3C HEDL / SSD
- 8) CUICX CRNL ll j
- 3) HAA-4 RI / ESG
- 9) MSPEC BCL l
- 4) HAARM-3 HEDL / SSD
- 19) MAEROS HEDL / CSA
- 5) HAARM-3 BCL
- 11) CCNTAIN $NL
- 6) HAARM-3 CRNL gi 2
3 I
8 to 18 18 '
19 13 EE, EES Ratios of Code Prediction to Ex:eriment for Suscended.%ss Concentration -- Test A85.
- -. _ _. - _ _ _ _ _ - _ -