ML20125C347
| ML20125C347 | |
| Person / Time | |
|---|---|
| Site: | North Anna  |
| Issue date: | 05/21/1985 |
| From: | Levine H, Sandoval R, Weber J SANDIA NATIONAL LABORATORIES |
| To: | |
| References | |
| OLA-1-A-003, OLA-1-A-3, SAND82-2365, TTC-0398, TTC-398, NUDOCS 8506120017 | |
| Download: ML20125C347 (113) | |
Text
. -
Fu s
,s 3
RANDIA REPORT SAND 82-2365
- TTC-0398
- Specified External Distribution Only' Printed June 1983 I
An Assessment of the Safety of o.cxtreo USNRC Spent Fuel Transportation in Urban Environs
'85 JUN -7 A10 :49 LFFCE Cr EEC3:~~
0 Z..
i p. E '
t R. P. Sandoval, J. P. Weber, H. S. Levine, A. D. Romig, J. D. Johnson, R. E. Luna, G. J. Newton, B. A. Wong, R. W. Marshall, Jr., J. L. Alvarez, F. Gelbard i-Prepared by j
$andia Nat6cnal Laboratories Albuquerque. New Mealco 87185 and uvermore, California 94550
])
for the Umted States Department of Energy under Contract DE ACO4 760POO789
- Only those recipiente enternal to SNL se listed under "Dietribution
are authorised to receive copies of thee report. They are not autho-rised to further dieseminate the information without permiselon from e
the originator.
NUCLEAR REGULATORY COMMl31104 z z %'- c L A - l
- #4 * / Official Exh. Ko-Becket Me le the antist of_
Ab-M Staff l'INTiriL*.
Applicant
--RICElVED _
latervanor REJECTCD-Cent's Off'r I"
Centractor DATE Other
-Witness ReportdrheY YW
~
/
U 8506120017 850521 5,2,oo m. 2, PDR ADOCK 05000338 G
l.
. h e
r Issued by Sandia National Laboratories, operated for the Cnited Statae Department of Energy by Sendia Corporation.
.{
NOTICE: This report wee prepared as en account of work sponsored by an agency of the United States Government. Neither the United States Govern-ment nor any agency thereof, not any of their employees, nor any of their contractors, subcontractors, or their employees, makes any warranty, espress or kmphed, or assumes any legal liability or responsibihty for the accuracy, compietenees, or usefulnese et any information, appetetus, product, of pro-ceae disclosed, or represente that its use would not infnnge privately owned rights. Reference herein to any specific commercial product, process, or service trade name, trademark, manufacturer, or otherwise, does not necessari constitute or imply its endorsement, recommendation, or favonna by the sted States Government, any agency thereof or any of their contractors et subcontractors. The views and opinione espreesea hereta do not neceneanly state or reflect those of the United States Government, any agency thereof or any of their contractors or subcontractors.
-s a-r w
~
y n
,r-
f 9
3 a-
' M As Assessnest of the Safety of Op+at Fzel i
Trasspcrtatics;is~yrtas E=viress 3AIC42-2365, TTC 0394
- s V
.A. P. Sandoval J.,-P. ideber. N. 5. Lewise. A. D. R:sig. J. D. Jchssce. 2.
E. t.ana, sandia untional Laterateries.1 Altesserque. New Merico 87135 4
C. J. N4 tes. 3. A. idesg. Lcvelace Islalatics Tericolcgy Eesearch Irstitute.2 P. O. Ecx 5890. Albsqsergie. New Mexico 37135 i
R. W. Marshall, Jr., J. L. Alvarer. Idaho Naticsal Est teeri:g Laboratories.3 EOM. Idaho. I
.. P. O. Ecz 1629. Idaho Falls. ID $3415 Fred Celbard Cc=ssitant. 561 Scuth Clesdeliff s31. La Matra. Califersia.
10531 1
'T g4 s -
\\.
p
. 'i.,
.1 4
s s,
s k
(
g j'
c 1.
Under US Department cf E:ergy Centract OE-ACC1. ISDcclti 2.
Under US Department cf Eseru Ccstract $E-ACC4-76ETO1013 s.
3.
Under US Cepartmest of Eme gy Ccatract OS-AC07 751D01573 me -
a
'T b
w..
--b
t 2 ABSTRACT The results of a program to provide an experimental data base for estimating the radiological consequences from a hypothetical sabotage attack on a light water reactor spent fuci shipping cask in a densely populated area are presented. The results of subscale and full-scale experiments in conjunction with an analytical modeling study are described. The experimental data were used as input to a reactor safety consequence model to predict radiological health consequences resulting from a hypothetical sabotage attack on a spent fuel shipping cask in the Manhattan borough of New York City.
The results of these calculations are presented in this report.
G
t
?
, ACKNOWLEDGEMENTS The authors would like to express their appreciation and thanks to the following people for their technical support and assistance in this study:
D. J. Alpert, Org 9415, SNL R. E. Bohannon, Org 1533, SNL D. M. Ericson, Jr., Org 9414, SNL B. J. Joseph, Org 9783, SNL W. B. Leisher, Org 9782, SNL F. H. Mathews, Org 1533, SNL N. G. Perdue, Org 1533, SNL R. T. Reese, Org 9782, SNL D. R. Stenberg, Org 9783. SNL M. M. Sturm Org 1822, SNL M. G. Vigil, Org 9782, SNL L. Isaacson, EC&G Geo-Centers. Inc.
R. C. Green, EG&G Idaho, Inc.
B. B. Kaiser, EG&G Idaho, Inc.
V. J. Novick, EC&G Idaho, Inc.
Y. S. Chen6, Lovelace Inhalation Toxicology Research Institute B. B. Boecker, Lovelace Inhalation Toxicology Research Institute G. M. Kanapilly*, Lovelace Inhalation Toxicology Research Institute R. O. McCellan, Lovelace Inhalation Toxicology Research Instituto R. C. Yeh, Lovelace Inhalation Toxicology Research Institute
- deceased
t
)
-111-AN ASSESSMENT OF THE SAFETY OF SPENT FUEL TRANSPORTATION IN URBAN ENVIRONS TABLE OF CONTENTS P. ale.
1 1.
Executive Summary.
5 2.
Introduction 7
3.
Pro 5eam Scope............
10 4.
Summary of Resdits 4.1 High Energy Device Evaluation Tests 10 4.2 Development of Measurement Techniques and Assessment of I~
10 Experimental Precision:
Subscale Tests 4.3 Quantification and Characterization of Material Released From a Full-Size Reference Sabotage Incident:
36 Full. Size Test 4.4 Development of a Correlation Between Depleted UO2 and Spent Fuel.
52 4.5 Analyses of Fuel and Cask Breakup and Aerosol Production.
56 4.6 Analysis of Radiological Health Effects 80 5.
Conclusions 89 5.1 Measured Source Term Release for a Full-Size One PWR Fuel Assembly Reference Event.
89 5.2 Calculated Source Term Release for a Three PWR Fuel Assembly Truck Cask Sabotage Event 89 l
5.3 Calculated Health Effects 89 5.3.1 One-PWR Fuel Assembly Truck Cask Reference Event 89 5.3.2 Three-PWR Fuel Assembly Truck Cask Sabotage Event 89 l
5.4 Degree of Precision and Accuracy in Measured Parameters i
and Calculated Results 94 1
l t
O
-lv-Page.
5.5 Comparison of This Study's Results With Other Study's Hesults.
95 6.
References
'98 Volume _II, (Classified CNSI) 7.
App.'ndices 1.1 APPENDIX A-HED Evaluation Tests:
Test Data 100 1.7 APPENDIX 8--Scaling Analyses:
Extrapolation to Higher Ener5y Attack Devices.
115 7.3 APPENDIX C-Subscale Tests / Full Scale Test: Test Data 120 _
volume III, (Unclassified) 7.4 APPEND 1X D-Correlation Experiments:
Test Data.
130 1.5 APPEND 1X E-Analytical Methods 227 1.6 APPENDIX F-Radiological Health Consequence Modeling: CRAC.
237 l
I l
t
)
-v-List of Figures Figure Page 4.7.1 Schematic of Confinement Chamber and 1/4-Scale Cask Used in Tests 3 Through 7 15 4.2.2 Schematic of the Experimental Setup for " Dry" Test 6 Prior to Detonation of HED.
17 4.2.3 Schematic of Test 6 (Dry) Tmmediately After Detonation.
18 4.2.4 Time History of UO2 Aerosol Mass Concentration Based on Sequential Filter Samples..
70 4.2.5 Time History of Mass Median Aerodynamic Diameter and
,ETI Geometric Standard Deviation for Test 6 (Dry).....
4.2.6 Schematic of Fuel Pin Damage and Damage Path Caused by HED Action for 1/4-Scale Test 6 (Dry) 22 4.2.7 Photograph Showing Cask Damage for 1/4 Scale Cask Test 6 (Dry).
73 4.2.8 Schematic of Test 1 (Wet) Immediately After Detonation.
25 4.2.9 Time History of Aerosol Mass Concentration Within Chamber Based Upon Sequential Filter Samples and Front Surface Reentrant Filter Samples for Test 7 26 4.2.10 Time History of U and 2r Mass Concentration Based Upon Rxtractive Filter Samples.
27 4.2.11 Time History of Aerosol Site and Geometric Standard Deviation for Test 7 (Wet).
28 4.2.12 TEM Photomicrograph of Aerosols Collected With Electrostatic Precipitator (ESP) Sampler for Test 7 (Wet) 29 4.2.13 TEM' Photomicrograph of Aerosols Collected with Electrostatic Precipitator (ESP) Sampleer for-Test 6 (Dry).
30 4.2.14 UO3 and Total Mass Distribution Based Upon Steved Debris Released From Cask Into Chamber for Tests 6 and 7.
32 4.2.15 Total Aerosol Mass Concentration Based Upon Continuous Flow Condensation Nuclel Counters of Tests 6 (Dry) and 7-(Wet) 33 f
8
I
-vi-List of Figures (continued)
E111re Page
~
4.2.16 Schematic of Fuel Pin Damage and Damage Path caused by Action of HED for 1/4-Scale Test 1 (Wet).
34 4.2.17 Photograph Showing Cask Damage for 1/4-Scale cask Test 7 (Wet).
35 4.3.1 Schematic of the Full-Scale Reference Test Configuration Showing Finned Cask Inside Pressure Chamber.
37 4.3.2 Schematic of the Full. Scale Test Configuration Tmmediatol)
After Detonation Showing Damage and Net Mass Loss 41 Fuel ' Assembly Showing 4.3.1 Photograph of 15 x 15 Depleted UO2 Damage Caused by HED in Full-Scale Reference Test 44
?-
Aerosol Mass Within Pressure Chamher
~
4.3.4 Time History of UO2 Based Upon Sequential Filter Samples for Full-Scale 45 Reference Test.
4.3.5 Total Mass and UO2 Mass Size Distribution for Steved Debris of Full Scale Test..
46 4.3.6 Aerosol Size (MMAD) and Geometric Standard Deviation as a Function of Time for Full-Scale Test.
47 4.3.7 Scanning Electron Micrograph of a Typical Rotating Plate Sample Taken 400 msec Af ter Detonation of the Full-Scale Event. Magnification is 200,000.
48 4.3.8 Scanning Electron Micrograph of Time Integrated (10 minutes)
Aluminum Planchet Sample Showing a 70.m (real diameter)
UO2 Particle.
50 4.3.9 Comparison of Total Mass Concentration (mg/ ) as a Function of Time for Subscale Tests No. 6 and 7 in the Full-Scale Test.
51 4.5.1 UO2 and ZrO2 Aerosol Concentrations as a Function of Time for Test No. 7 (Wet) 57 4.5.2 FeO and Pbo Acrosol Concentrations as a Function of Time After Detonation for Test No. 7 (Wet).
Analyses Are by XRF.
58 4.5.3 002 and ZrO2 Aerosol Concentrations as a Function of Time for Full-Scale Test (Dry)................
59 4.5.4 Pbo and FeO Aerosol Concentrations as a Function of Time After Detonation for Full-Scale Test (Dry).
60
i
-vil-List of Figures (continued)
Page Elgute 4.5.5 Aerosol Fraction for 2rO2 and UO2 as a Function of 61 Time After Detonation for the Full-Scale Test.
4.5.6 Aerosol Fraction of ZrO2, PbO, and FeO Components as a Function of Time for Full-Scale Test.
62 4.$.1 SKM Micrograph of Material Depos led on the Valve Door 64 During the Full. Scale Cask Test 4.5.8 X ray Spectra Showing the Elements Preseht in Figure 4.5.7.
65 4.5.9 Macrophotograph (1.5x) of the Stainless Steel Slug Taken From Cask After the Full Scale Test Event.
66-4.5.10 Micrograph of the Stalnicos Stcol Slug in Cross 67 Section (500x).
4.5.11 Macrophotograph of a Piece of Zircaloy Cladding After the 68 Full-Scale Test Event (1.5x).
4.5.12 Micrograph of Zircaoy cladding in Cross Section (100x).
69 4.5.13 Microscopic Cross Section of UO2 Fuel Pellet (100x) 70 4.5.14 Plot of Measured Aerosol Mass Concentrations Obtained From Sequenced Cascade Impactors for Full-Scale Reference Test.
73 4.5.15 Plot of Total Suspended Aerosol Mass Concentration as a Function of Time.................
78 4.5.16 Plot Showing Comparison of Calculated Mass Concentration From Model With Measured Values After 52.5 Seconds Postdetonation.
79 I
i
~
-a
t s
i l
-vili-List of Tables Table Page 4.1.1 Construction Material Data for Truck Spent Fuel 11 Shipping Casks 4.1.7 High Energy Devices Surveyed.
12 4.2.1 Summary of Scaled Cask-High Energy Device Tests 13 4.2.2 Sampling Instrument Used in 1/4-Scale Chamber Tests 16
.a 4.3.1 Sampling Instruments Used in Full-Scale Cask Test 42 4.3.2 Summary of Results of Full-Scale Test 4.4.1 Filter Data From Battelle Columbus Laboratory's Studies on Spent Fuel Release Fraction.
54 j-4.5.3 Surface Area for Aerosol Deposition.
14 4.6.1 Measured Release Fractions for a 1 PWR Fuel Assembly and Calculated Release Fractions for a 3 PWR Fuel Assembly Track Cask.
82 4.6.2 Spent Fuel Cask Radionuclide Inventory Used in This Study 84 4.6.3 Population Distribution Used for This Analysis.
86 4.6.4 CRAC Computed Health Consequences for This F.xperimental Study.
87 4.6.5 Peak Thyroid and Bone Marrow Doses as a Function of Distance From Release Point.
88 5.1 Summary of Release Parameters for a 1 PWR Assembly Cask Event.
90 5.2 Summary of Release Parameters for a 3 PWR Assembly Cask Event.
91 5.3.1 Health Consequences for a One PWR Assembly Truck Cask Sabotage Event.
92 5.3.2 Health Consequences for a Three PWR Assembly Truck Cask Sabotage Event.
93 5.5.1 Comparison of Extrapolated Test Results With Urban Study Results.
96 5.5.2 Comparison of CRAC Computed Health Effects With Urban Study Results 97 i
t t
1.
EXECUTIVE
SUMMARY
In 1978 a study of radiological impacts from transport of radioactive l
material through urban areas, the 1978 Urban Study, indicated very severe consequences from a suc'cessful malevolent act on spent fuel shipments. On the basis of that analysis the NRC instituted stringent physical security require ments2 for spent fuel transport which were designed to prevent sabotage events in urban areas. A subsequent version of Reference 1, the 1980 Urban Study,3 reduced the postulated release quantity by a factor of 14 and thus showed reduced numbers of early fatalities, morbidities and latent cancer fatalities. As a result of the second report, the NRC reduced the stringency of the physical security measures, but they remain a serious restriction on the shipment of spent fuel and have resulted in increased shipping costs.
Since no relevant experimental data base was as411able for use in the Urban Studies, source term estimates were based upon assumed physical and chemical characteristics and estimated quantitles of the released fuel.
Consequently, there was a high degree of uncertainty in the estimated _ source ter_ms and radiological consequences.
A need existed to provide experimental data char'acterizing the quantity, physical, and chemical form of fuel released from hypothetical attacks on spent fuel shipping casks.
This report describes the results of a program conducted at Sandia National Laboratories (SNL) to provide the experimental data base for estimat-ing the radiological consequences from a hypothetical sabotage attack on a spent fuel shipping cask. The primary objectives of the program were limited to (1) evaluating the effectiveness of selected high energy devices in breach-ing full-size spent fuel casks, (2) quantifying and characterizing relevant aerosol properties of the released fuel, and (3) using the resulting experi-mental data to evaluate the radiological health consequences resulting from a hypothetical sabotage attack on a spent fuel shipping cask in a densely populated area.
Subscale and full-scale experiments in conjunction with an analytical modeling study were performed to meet the programmatic objectives. The program was divided inte the following tasks:
i Perform subscale and full-scale tests to evaluate the capability of available high energy devices (HEDs) to breach generic spent fuel truck casks and disperse cask contents.
According to the results of this evaluation, a reference case HED was selected for further evaluation and full-scale testing.
Perform subscale tests on casks filled with surrogate fuel subjected to scaled versions of the reference HED to establish measurement techniques and to determine a preliminary release fraction for surrogate matreial subjected to HED environments.
j
~
t t,
i Perform subscale tests on surrogate and actual irradiated fuel pellets subjected to scaled HED attacks in order to quantify and characterize radioactive particle production for various high energy environments. A correlation.between aerosol and fine partjele_ parameters _for spent.fue I f*"8' Of shock impact,_lo_adings _was_ developed._
and depleted UO2 f0f
- Conduct a full-scale source-term characterization test using the reference HED and a generic truck cask containing unieradiated UO2 fuel to characterize the released fuel (quantity, particle' size, composition, etc).
Develop an understanding of the cask and fuel material response to high energy / explosive environments and to develop a basis for predicting the response of similar casks to larger HEDs.
Use the experimental data obtained to evaluate the health consequences resulting from a hypothetical sabotage attack on a spent nuclear fuel shipping cask in a densely populated area.
,7-HED Evaluation An extensive survey of available high energy devices (HEDs) was performed to select these that might be capable of breaching a full-size spent fuel truck cask.
From the many different types of attack devices considered in the survey, four general types of HEDs were selected for testing and further evaluation. These devices were those discussed in the 19181 and 19803 Urban Studies:
1.
Conical-shaped charges.
2.
Contact breaching charges.
3.
Platter charges.
i 4.
Pyrotechnic torches.
Tests subjecting both simulated and actual spent fuel truck casks to the four types of HEDs were performed to provide data for final selection of a ref-erence HED which showed the greatest potential for penetrating a full-size cask and dispersing its contents.
An HED was selected from the four types tested and was used as the reference attack device for the full-scale source term characterization test from which the needed source term data base was obtained. Volume II of this report elaborates on the selection of the HED; more detail cannot be provided here because of national security limitations.
Subscale Tests Five subscale tests of 1/4-scale casks containing
~
full-size fuel pins made up of unieradiatedN tre1Tatirn targets for 2
scaled versions of the selected reference HED were conducted. These tests provided initial experimental data characterizing the fuel material released from a cask subjected to a sabotage incident. The results of these experi-ments indicated that approximately 48.6 1 5 g of UO2 fuel mass was released from the 1/4-scale cask as a result of the attack. Approximately 1.6 percent 1
(0.78 g) of the total released UO2 mass was in the respicable size range (ie, less than 10 micrometers aerodynamic diameter).
A calculation of the i
4 l
~..
t t,
l fraction of released airborne respirable aerosol for a full-size event assuming a three assembly pressurized water reactor (PWR) cask (1.4 t cf heavy metal (tHM) inventory) of the type used in the Urban Study 1.3 was made based on the measured 1/4-scale release parameters. This calculation assumed
)
the longest path of interaction through the cask together with rupture of, and p subsequent release through, the cask's walls. The results of these extrapola.
g act tions of the scaled tests indicated that approximately 0.0023 percent (32 g)
[
of the total solid fuel inventory could be released from a full-scale sabotage event as respicable radioactive materials.
Full-Scale Test A full-scale test subjected a 25.45 t generic truck cask containing a single PWR-like unieradiated depleted UOg fuel assembly to the reference full-scale HED.
A total UO2 fuel mass of 2.548 kg was released from the cask as a result of the explosive attack. Approximately 0.0006 per-cent (3 g) of the total unieradiated fuel inventory,(0.5 t) was released as a respirable radioactive aerosol. These full-scale test data were used to calculate the quantity of radioactive material that could be released as a
.E result of an esplosive attack on a three PWR fuel assembly generic truck cask.
These calculations indicated that approximately 0.0005 percent (6 g) of the total unieradiated fuel inventory (1.4 t) could be released as a respir-abic radioactive aerosol as a result of an explosive attack on a 3 PWR fuel assembly truck cask.
Co r rel a t i_qn. T.e s t s._ Effects of the high energy environments created by a variety of HEDs on breakup and particulation of spent conumercial nuclear reactor fuel and its surrogate, depleted uranium dioxide (d-UO ), were 2
evaluated in a series of single pellet tests. Tests conducted on single irradiated fuel pellets and single depleted UO2 pellets enabled measurement of the radioactive aerosols typical of the high energy environments.
l l
Correlation functions were obtained from both filter and sieve data relating particle size distelbutions for fracture, breakup, and aerosolization of depleted UO2 fuel to that of irradiated fuel. Using filter data, a spent fuel to depleted UO2 Particle mass production ratio of 0.53 was obtained.
j
~
This would result in a smaller respicable aerosolized release than obtained in gge' the scaled and full-scale cask tests with depleted uranium surrogate fuel
- 'l l
rods. An extrapolation of wet sieve data into the respirable range (a doubt-ful procedure) resulted in a correlation ratio of 5.6.
This value would increase the released quantity of aerosolozed fuel from the scaled and full-scale tests. Similar data are available from an NRC-sponsored test program at Battelle Laboratories in Columbus, Ohio.4 In their experiments a ratio of j
spent fuel to depleted uranium respirable release of 0.7 was obtained.
l In considering which correlation ratio is the most appropriate, it at, pears that a value less than one is the most probable. This implies that the aerosolized respirable release from the reference base incident (one PWR fuel assembly cask) would be less than 3 g of irradiated fuel, and the aerosolized respicable release from a 3 PWR fuel assembly cask of the type used in the Urban Studyl would be less than 6 g. __However, for conservatism _in_thJ l
health risk assessment the correlation value of 5.6 was used; this leads to an l
l
f f upper limit release value of 17 g (3.4 x 10-3 percent) of aerosolized respicable irradiated fuel for a one-PWR assembly cask and 34 g (2.4 x 10-3 percent) for the maximum respirable aerosol release from a three-PWR assembly shipping cask.
Health Effects Evaluation The reacter safety study consequence model, CRAC,5 was used in the Urban Study to estimate human health consequences from an attack using the reference HED on a 3 PWR fuel assembly truck cask.
The basic scenario as defined in the Urban Study was (1) the attack occurred in the borough of Manhattan in New York City, (2) the attack occurred on a weekday, midafternoon, (3) the spent fuel inventory was typical of PWR assemblies,after_i_5O days cooling at the reactor, (4) all consequence esti-mates were made without any evacuation to avoid early exposure.
For this scenario the Urban Studyl estimated the health consequences to be 4/60 (mean/pcak) early fatalities, 160/1600 (mean/ peak) early morbidities and 350/1300 (mean/ peak) early latent cancer fatalities. Using the same CRAC model and assumptions and this study's experimentally determined release fraction (3.4 x 10-3 percent) for the same attack mode on a si_ng_1e,PWR fuel 37 assembly truck cask (0.5 tHM, 150 days cooled), values of health consequences were found to be 0/0 (mean/ peak) early fatalities, 0/0 (mean/ peak) early morbidities, 0.3/1.3 (mean/ peak) early latent cancer fatalities, and 2/7 total latent cancer fatalities.
Extrapolating the experimentally determined release fractions for a single PWR fuel assembly cask to the Urban Study 1 3 PWR fuel assembly cask scenario, an estimate of the health consequences of 0/0 (mean/
peak) carly fatalities, 0/0 (mean/ peak) early morbidities, 1/3 (mean/ peak) early latent cancer fatalities, and 4/14 total latent cancer fatalities were obtained. These newly calculated latent cancer fatalities are smaller by a factor of 350/433 (mean/ peak early latent cancer fatalities) than the original Urban Study predictions upon which the NRC interim regulations for US transport of spent fuel were based.
Overall Program Result The data from this program, together with that from the NRC sponsored BCL program, indicate that the Urban Studies 1,3 4
greatly overestimated the impact of malevolent acts directed at spent fuel casks in urban environs. From that standpoint this work could be the basis of additional regulatory revisions of the NRC physical protection requirements.
In a larger sense this work can also be the basis of more credible " worst case" analyses since it defines the actual result of an event which is well beyond any expectation of cask failures in accident environments. Thus this experimental program has provided significant new information on the behavior of spent fuel and su: rogate materials under severe shock and thermal environ-ments which can be the basis of a better understanding of spent fuel transport risks and safety analyses.
l 1
i f
f I l
2.
INTRODUCTION i
l Public attention has been focused on the environmental impact that could result from the sabctage of spent nuclear fuel shipments in densely populated
[
urban areas.
Previous studies have been sponsored to assess the potential of several generic HEDs to breach large spent fuel shipping casks and to evaluate the radiological hazards which could result from the transportation of radio-active material (RAM) in urban areas for various types of environments, including those caused by sabotage.le3 The most recent version 3 of the l predicted approximately 100 total latent cancer fatalities could 1978 study
]
occur from the successful sabotage of spent furi shipping systems. Since no experimental data were available for these studies, certain bounding condi-tions and specific chemical and physical characteristics of the released material were assumed. Since the analysis became the basis of a potentially costly Nuclear Regulatory Commission (NRC) regulation, the need to verify the assumptions in the snalysis received high priority in the Department of EnergyL (DOE).
This report describes the results of a program conducted at Sandia l
National Laboratories (SNL) to establish this needed data base. Subscale and '~~~~~
7 full-scale experiments were performed in conjunction with an analytical modeling effort to develop a data base characteri:Ing the release of radio-i active material from a spent fuel shipping cask subjected to a hypothetical sabotage event.
The experimental data base was used to develop improved estimates of the radiological health consequences resulting from the sabotage j
of spent fuel transports in urban regions.
The origin of the program can be traced back to 1975 when the US Department of Energy (DOE) sponsored a study at SNL to determine the ability l
of several generic HEDs to disrupt a large truck spent fuel shipping con-tainer. The results of this study indicated that it was indeed possible for l
certain HEDs to breach a large spent fuel cask.
l In 1977, the NRC published a final environmental impact statement on the transportation of radioactive materials by air and other modes (NUREC 0170)8 l
which concluded that spent fuel shipments do not constitute a threat to the l
pubile health and safety. This same study stated that the risk of sabotage of
[
radioactive materials transports is sufficiently small to constitute no adverse major impact to the environment or to public health.
O However, in 1978, the NRC published a draft environmental assessment of the transport of radionuclides in urban environments.1 The so-called " Urban Study" evaluated the radiological hazards resulting from the transportation of radioactive material in urban areas for various types of environments includ-ing those caused intentionally. The first draft version of this study, SAND 7 7-192 71 predicted that several hundred latent fatalities could occur from the successful sabotage of spent fuel shipments subjected to certain 3
modes of attack. A second version of the urban study, NUREC/CR-0743 reduced the latent fatalities to fewer than 100 based upon a reevaluation of released quantitles of radioactive material.
t
. In 1919, the NRC reacted to the initial urban study by requiring physical 2 for spent fuel shipments in the United States pending protection measures the availability of credible experimental data supporting or disproving these predictions. These safeguards include (1) the ecuting of truck and rail shipments of spent fuel to avoid densely populated areas, where possible, (2) requiring armed escort (s) for shipments traversing heavily populated areas, (3) requiring that the transport vehicle be equipped with NRC approved immobilization devices, (4) requiring that the shipments be accompanied by at least two drivers (escorts) and (5) requiring that the shipment be under constant surveillance at all stops.
9 in 1979, the Comptroller General of the United States published a study of federal actions needed to improve safety aad security of nuclear material transportation. This study recommended that the NRC and DOE develop experi-mental data bases characterizing the quantity of material that could be released from the sabotage of spent fuel casks.
Y In response to these data requirements and requirements to support DOE fuel programs, the DOE initiated a study at SNL to evaluate the effects of intentional acts on spent fuel shipping systems and to determine experi-mentally the quantity, size, and chemical form of any released material. This source term data base was to be used in existing health consequence models to assess the safety and security of spent fuel transportation for these types of environments. This report describes the details and results of the DOE sponsored program.
i e
I
i e
1 3.
PROGRAM SCOPE The program scope was to conduct subscale and full-scale tests in conjunction with an analytical modeling study to fulfill the following objectives:
1.
To evaluate the effectiveness of selected high explosives and high energy devices (HEDs) to breach a generic spent fuel cask and disperse cask contents.
2.
To quantify the fraction of cask contents released as a result of the high explosive or high energy attacks on a generic cask. Mass fraction and, in cases where applicable, activity fraction were to be measured.
~
3.
To characterize the physleal and chemical properties relevant to human health risk estimates of the released radioactive material. Particle size, morphology, mass, mass concentration and elemental composition were E~
some properties of the released material that were to be measured.
4.
To develop an understanding of the material response (of both cask and fuel) to the high explosive /high energy attacks in order to develop a basis for predicting the response of similar generic casks to larger quantities of high explosives or HEDs of similar design.
5.
To use this experimental and analytical data base to evaluate the health consequences resulting from a successful sabotage attack on a spent nuclear fuel shipping cask in a densely populated area.
6.
To provide data to regulatory agencies for setting standards governing the shipment of civilian reactor spent fuel in the US.
l Early in the program, it was realized that it would not be feasible from cost and safety standpoints to perform full-scale tests involving spent fuel l
in.the atmosphere.
It was also realized that in order to achieve the program matic objectives in a cost efficient manner, a series of scaled tests using irradiated and unieradiated fuel should be performed.
It also became clear that to maximize the precision of mass accountability and to minimize the mass loss from the event, the subscale and full-scale tests should be performed in a pressure vessel (aerosol chamber). The chamber offered the advantages of confining the released material and permitting all of the dispersed material to be recovered for mass balance and particle size measurements.
Another area of concern was the need to select a worst case attack device
(
(ie, a device capable of causing the greatest amount of damage to the cask and l
maximum release of fuel while being relatively available and requiring minimal technical expertise) for the full-scale reference event.
L
a i In order to achieve the programmatic objectives and to address these areas of concern, the program was divided into the following tasks:
Task 1: To evaluate and characterize the effectiveness of several types of HEDs to breach generic truck casks and aerosolize and disperse spent fuel elements. Also important in this evaluation was the scaling of cask and spent fuel response parameters. This task consisted of full-scale tests using simulated cask walls and/or full-scale generic truck casks as targets for four types of HEDs described in the Urban l
Study :
(1) conical-shaped charges (CSC), (2) contact-breaching charges (CBC), (3) platter charges, and (4) pyrotechnic torches.
St rogate fuel elements (depleted UO2 fuel pins) were placed between the simulated cask walls in some of these tests. The results of this experimental evaluation were used to select a worst case attack device (HED) for the full-scale ~ test (Task 4).
Task 2:
To develop measurement techniques and establish maximum achievable
,E~
measurement sensitivities and precisions, subscale tests were l
performed on spent fuel casks filled with depleted UO2 fuel pellets i
(zircaloy cladding) subjected to a scaled version of the reference
{
base HED selected in Task 1.
These tests were conducted in a pressure chamber to confine the released material and permit recovery of all of the dispersed material for mass balance and particle size measurements.
Task 3:
To develop a correlation between selected radionuclide particulate size distributions for spent fuel subjected to scaled explosive attacks and that for depleted UO2 fuel, single pellet tests were performed. This task was completed by EG&C/1daho National Engineering Laboratories (INEL) and involved subjecting single H. B. Robinson 2 spent fuel pellets and depleted UO2 Pellets to a scaled version of the reference base HED selected in Task 1.
The correlation function developed was used to quantify released j
radionuclides and their fractionation for comparison with the i
full-scale reference test results (Task 4).
l l
Task 4: To quantify and characterize the release of unieradiated fuel in a I
full-scale event, a generic truck cask was subjected to the reference base HED selected in Task 1.
The full-scale test provided a source j
term data base for the reference base-sabotage event which in j
conjunction with the results of the subscale experiments performed in Task 2 were used as primary input and data base to the Consequence Reactor Safety Model (CRAC)5 for estimating the radiological consequences.
Task 5: To develop an understanding of the cask and spent fuel material response to selected HEDs End to model the fuel breakup for various types of energy loadings, an analysis was performed to determine thermal and mechanical response of the unieradiated fuel and cask material to the reference HED.
A method was developed for scaling
. - ~
3
.g_
subscale source ters parameters (released mass fractions, size fractions, etc) to the full-scale event through detailed consideration of HED characteristics, material properties and geometry.
Task 6:
To complete the study, the evaluation of the radiological health effects using the experimental data derived in Tasks 1 through 5 was completed using the Consequence Reactor Safety Model (CRAC) used in the Urban Study. The expected health and economic consequences were calculated assuming a release in the Manhattan borough of New York City.
This multidisciplinary research effort was divided among the following laboratories or contractors:
(a) Tasks 1, 2, 4 and 6 were performed at Sandia National Laboratories, Albuquerque, New Mexico.
Lovelace Inhalation Toxicology Research ;_
Institute, Albuquerque, New Mexico supported the aerosol measuremest activities of Tasks 1, 2, 3 and 4.
(b) The experimental work of Task 3 was performed at EG&C Idaho National Engineering Laboratory, Idaho Falls Idaho.
(c)
Task 5 was performed by SNL personnel and by Professor F. M. Gelbard while at the Massachussetts Institute of Technology.
The analytical and experimental results of the work performed in Tasks 1-6 are summarized in Section 4 Summary of Results.
D i
L
-10 4.
SUMMARY
OF RESULTS 4.1 HED Evaluation Tests An extensive survey of attack devices which could be available to a saboteur and which showed a potential for breaching a nuclear spent fuel shipping container was performed.
The types of shipping containers with which this study was concerned limited the modes of attack to high energy and/or high explosive devices.
Because of the heavy radiation shields and container damage resistance requirements, spent fuel shipping casks are very large and massive and may weigh from 22 to 88 t or more depending on the type of cask.
Wall construction of the e casks include combinations of stainless steel, lead, depleted uranium, water, wet cement, and resin. Table 4.1.1 provides wall material data for five spent fuel truck casks considered in this study.
A previous study at SNL of several high explosive devices has shown that the variety of attack devices capable of penetrating and/or breaching a full-size truck cask is limited.
A survey of the attack devices shown in Table 4.1.2 was performed in this study for the purpose of selecting several candidates for more detailed evaluation and subsequent testing.
From Table 4.1.2, the devices were selected based on their availability to the perpetrator and their potential to becach truck casks of the type listed in Table 4.1.1.
The details of the tests performed and the HED device selected is contained in Appendix A of Volume II of this report.
4.2 Development of Measurement Techniques and Assessment of Experimental Precision:
Subscale Tests Seven tests using 1/4-scale casks and/or depleted UO2 fuel pins were conducted in support of the second task to develop source acrosol measurement techniques and to provide a data base for assessing measurement precisions and experimental feasibilities in anticipation of the full-scale cask /fuct test.
The results of two preliminary tests * (Tests 1 and 2) subjecting simulated cask walls and depleted UO2 fuel pins to explosive charges in an open environment indicated that it was not feasible to make accurate mass balances of uranium fuel released in an unconfined area.
For this reason, and also in order to obtain time histories of source aerosol parameters, subsequent scaled and full-scale tests were conducted in containment.
The containment pressure vessel offered the advantages of confining the aerosol and permitting all of the disrupted and/or dispersed fuel to be recovered for mass, elemental and particle size analyses. Table 4.2.1 lists the five confined scaled eask/UO 2
fuel tests (3 through 1) conducted together with the purpose and results of each test.
- Appendix C:
Subscale/ Full-scale Tests: Test Data
TABLE 4.1.1 Cowetruction hterial Data f or Truck Spent Fuel Shipping Casks NFS-4 NLI 1/2 TN8 1N-9 CASK (mm)
(mm)
(mm)
(mm) h terial Layers - Sidel 7.9 321 SS 1.27 SS 6
304 SS 6
304 SS (Thickness) 168.4 Lead 69.9 Dep. Uran.
5 Copper Plates 5
~
Plates 150-Lead 150 Lead 185 31.8 321 SS 54 Lead 10 Wet cement 10 Cement 114.3 Bordted 22.2 SS 20 Carbon steel 20 ca rtnn water Steel 4.2 321 SS 127 Water 150 Borated resin 150 Borated Resin l
H 6.4 SS 200 Copper fins 200 Copper Y
Fins hterial Layers - Bottum 203.2 SS 31.8 SS (Inner 7.6 304 SS 7.6 304 SS (Thickness) cavity) 88.9 SS 226.1 Lead 226 Lead 200.7 Dep. Uran.
16 304 SS 16 304 SS 80.8 SS h terial Layers - Top 190.5 SS 48.3 SS 38 304 SS 38 304 SS (Thickness) 193 Dep. Uran.
208 Lead 208 Lead 80.8 SS 25.4 304 SS 25.4 SS 72.4 SS (outer closure)
I Dimensions given in inches. hterial layers are listed in order f rom inside of cask.
I References Directory of Certificates of Cimap11ance for Radioactive Mate %9 s Packages, US Nuclear 1
j Regulatory Commission, NUMEG-0383, December 1981 I
TABLE 4.1.2 liigh Energy Devices Surveye<l L
TiiERMAL MECliANICAL (IDU VEIDCIfY) 14ECilANICAL (IIIGli VEIDCITY)
ELECTRICAL CHEMICAL Pyrotechnic Torch Demolition Saw Explosive Air Blast Arc.
Acid Gas Torch Grinder Explosive Shaped Charge Electron Beam Exothermic The rinite Iligh Speed Drills.
Explosive Flyer Plate Laser Reagents Burn Bar liigh-Velocity Projectiles Rod. Penetrators U
D: plosive Contact Charge i
High-Velocity Particles 4
s 9
A l*
i 9
i TABLE 4.2.1 Summary of Scaled Cask-Explosive Device Tests No.
Date Target
__ Test T128..__
Purpose and Results 3
.1/80 Thick steel target Chamber Feasibility of chamber test.
(1/4-scale HED)
Evaluation of aerosol sampling instruments and isolation valve.
Multicomponent aerosol observed requiring several aerosol instruments for measurement of mass and size.
4 2/80 Thick steel target Chamber (Same as Test No.3) _l~
(1/4 scale HED)
S 3/80 Zirealoy clad Chamber (Same as Test No.3) Verifica.
depicted UO2 (1/4-scale HED) tion of uranium fluorometric fuel pins analyses.
6 6/80 1/4-scale cask Chamber Determined respicable UO2 containing short sec-(1/4-scale HED) release fraction based upon tions of a 5 x 5 removed fuel mass (0.0036) array of surrogate from dry 1/4-scale cask using fuel pins. No water 1/4-scale EID.
coolant was used in this test.
1 3/81 1/4-scale cask Ch amber Determined respicable UO2 containing short see-(1/4. scale HED) release fraction based upon i
tions of a 5 x 5 removed fuel mass (0.00014) array of surrogate from a wet 1/4-scale cask fuel pins. Water using a 1/4-scale KID.
was used in fuel j
cavity and outer l
water jacket.
l l
, Tests 3, 4, and 5 were conducted to evaluate acrosol measurement and analytical techniques for explosive environments and to evaluate the overall experimental system.
Five 3 in. thick steel blocks were used as targets for a scaled version of the reference HED in Tests 3 and 4, and three 30 cm long sections of depleted UO2 fuel pins were the targets for Test 5.
Test 6 used a 1/4-scale dry cask containing a 5 x 5 array of 0.9 m long full-scale UO2 fuel pins and Test 7 used a 1/4-scale cask containing 3.96 x 10-3 3 or m
water coolant and a 5 x 5 array of 0.9 m long full-scale UOp fuel pins in the fuel cavity. The results of Tests 6 and 7 were used to predict release parameters for a full-scale event and to provide an understanding of the achievabic precision measures and experimental resolution of measured release parameters.
Figure 4.2.1 is a schematic of the steel confinement' chamber which was used in Test 3 through 1.
The steel cylindrical chamber.was a 28.9 m3 (net volume) pressure vessel scaled at one end and having an air-tight door at the other end.
The HED was mounted externally to the chamber and fired into the chambe r t hrough a 6.4-mm diameter poet in a flanged assembly mounted. externally i-to the chamber.
An explosively actuated isolation slide valve was used between the HED and the chamber to prevent release of gases and dispersed fines f rom t he chamber.
Five sampling ports penetrated the chamber in various locations. The ports were closed by remotely operated pneumatic 2.5-cm diameter valves. The valves were kept closed until af ter detonation to prevent damage to the aerosol sampling equipment by the shock wave.
Since no single aerosol ~1nstru-ment can size particles over the size range of interest (from 0.01 pm to about 2 mm diameter), a battery of instruments listed in Table 4.2.2 was selected to measure the size and mass of particles collected.
Aerosol size parameters as a function of time were determined from cascade impactor samples obtained at various time intervals after HED detonation.
Similarly, filter samples provided a time history of the change in mass concentration.
Changes in morphology were shown by sequential electrostatic precipitator (ESP) samples and changes in number concentration were shown by continuous recording condensation nuclei counters.
After each test, all debris were collected from surfaces inside the test chamber and separated into material not containing uranium and materials suspected of containing uranium. All uranium containing 4
l material was sieved, the mass determined, and uranium fluorometric and wave length dispersive x-ray fluorescence analyses performed. Transmission (TEM) and scanning (SEM) electron microscopy and energy dispersive x-ray analysis (EDKA) were used to determine the elemental composition and morphology of particles co11ceted by ESPs.
l The acrosol-sampling procedure was designed to provide a time history of selected acrosol parameters (such as concentration, particle size distribu.
I tion, and morphology) in the chamber.
From the aerosol time history, l
calculations could be performed to determine initial release parameters.
Figure 4.2.2 is a schematic of the experimental setup for Test 6 prior to detonation of the HED.
A total UO2 fuel mass of 15.3474 kg was measured before the event. Figure 4.2.3 is a reconstruction of the test configuration l
l
i
- h i
L
.I I
e ENE_2GY OEvCE AERC50L"M (4is 71 J
vi a
- 1.,i 9
7 AERCSCL w.>G a
v f 'S~EE.
\\(p m aTEin e
pag +'
/,\\ k M *M.hG
/hp dC M
'N
/
%+
Su4RCGA*E F.E. 89d5
~" "
(SL% ***!)
k.p
=. _ _ -
m a
Nl
._e-_
a =m - y my r
W (324 m I
l!
Q l!
< A m w.sc b-s\\
k!
i!
//
m woue mort m
- .1.!
- 1 i
Fir re 4.2.1 5:%e.=:i ef c:.f i,~_ m-: e..g. eg if4 3_ ; 3,g
,,3 g.
Tes s 3 S_.72-3 7.
. l TABLE 4.2.2 Sampling Instruments Used in 1/4-Scale Chamber Tests Instrument and Applicable Size Ra n_Le Purpose of Samples Analytical Model LMJ Cascade im-Aerodynamic size, geometric standard Gravimetric (Cahn micro-pactors 0.5 -
deviation of total aerosol mass and balance) and fluorometry 12pm UO2 mass.
for determination of ura-nium.
Point to-plane Particle morphology, count distribu-Transmission and scanning electrostatic tion and elemental distribution electron microscopy and precipitator energy dispersive x-ray
~
TEM & SEM analysis 0.0 L - 12 p m Filter 37 mm, Provide a time history of total aero-Gravimetric for total sequential, sol and uranium mass after EID deto-mass and uranium by 0.01 12 m nation. Samples for surface area fluorometry, BET nitro-measurements.
gen absorption for surface area measurements.
Filter 37 mm, Sequential filter samples from fil-Gravimetric for total front surface ters inside chamber to compare with mass of aerosol and ura-reentrant Ell-filters obtained by extractive tech-nium by fluorometric ter (FSRP) niques to address aerosol line techniques.
0.01 - 12 m losses.
l bCondensation Total count of aerosol particles vs.
Optical light scattering l
nuclei counter time after detonation.
Instrument.
(CNC) 0.001 -
12 pm i
i
- Klectrical Size distribution parameters for Electrical mobility.
aerosol analy-particles between 0.01 and 1.0 pm.
zer (EAA) Dia 1.O p m l
Steves (38 -
Provide size distribution data on Mechanical sieving fol-l 2000 p m) larger particles of surrogate spent lowed by weighing for to-fuel.
tal mass and fluorometry for uranium determination.
aLovelace Multijet.
bData from these instruments are not included in this report.
J l
I l QUARTER SCALE TEST (DRY)
SURROGATE FUEL / CASK H E D ----
A
-84 g Cu I
E E E I
D E
a a u E
I I
I I
I E
II II 18,890 g fuel pins uu 103.6 kg STEEL (
l BEFORE EVENT CASK wt. = 316.04 kg 18,890 g fuel pins -
15,347.4 g UO + 3,548.8 g Zr-4 2
l Figura 4.2.2. Schematic of the Experimental Setup for " Dry" Test 6 Prior to Detonation of HED
. QUARTER SCALE TEST (DRY)
AFTER EVENT Y.
I I.
I l
y MASS OF AEROSOL (ts o.2-3.0 mm)0.78 g UO
'"I i-n 1
p CASK CEPOSIT@ 123.1 g uo2 CHAMBER WAt L DEPOSitiOr4 4 7.0 g U0 l 18,680 g fuel pens 2
r 2
2 103.6 kg STEEL (
TOTAL SWEPT MASS (feel pas) 216.0 g l
SWEPT MASS UO 379 5 9 2
SWEPT MASS OF Zr-4 36.5g l
MASS BALANCE 95.6 %
Figure 4. 2. 3.
Schematic of Test 6 (Dry) Icnediately Af ter Detonation l
, immediately after detonation. A total UO2 fuel mass of 15.1679 kg remained in the cask after the event which indicated that approximately 179.5 g of fuct was removed from the fuel assembly as a result of the HED action.
UO2 One hundred and twenty three grams of UO2 were deposited in the cask, 47.8 g of UO2 were deposited and recovered from the chamber walls and floor and 0.78 g of UO2 was released as an airborne aerosol. Not accounted for in this test was 7.8 g of UO2, approximately 4.3 percent of the total UO2 fuel mass.
The unaccounted for UO2 mass was believed to be of particle sizes greater than 30 ym which were deposited, but not collected, on surfaces inside the test chamber.
Figure 4.2.4 shows a time history of the UO2 aerosol mass within the chamber based on sequential filter samples. The exponential decay of UO2 acrosol mass indicates that no unusual phenomena were occurring in the chamber and that the first filter sample taken from 0.2 min to 3 min after detonation is consistent with all subsequent filter samples.
6 maximum UO2 aerosol mass concentration of 0.27 mg/l was detected at 12 seconds postdetonation.
Using the containment chamber volume of 2886 1 and assuming a uniform spatial _
concentration, a total released UO2 aerosol mass of 0.78 g was calculated.
Assuming that 100 percent of the measured UO2 acrosol mass is in the respicable size range, a total UO2'respirabic mass of 0.18 t 0.005 g was released from the cask as a result of the event.
Sequential cascade impactor samplers were also taken to give a time history of aerosol size distribution parameters.
Figure 4.2.5 shows the mass median aerodynamic diameter (MMAD) and geometric standard deviation as a function of time for Test 6.
Assuming a single mode lognormal distribution, mass median aerodynamic diameter (MMAD) reached a maximum of 3.5 pm af ter appenximately 12 min.
The size then stabilized at 2.0 pm af ter 30 min.
Fluorometric determination of uranium dioxide concentrations indicated concentrations ranging from 7 percent of the total aerosol mass at early times to about 2 percent 10 min after detonation.
Similar UO2 aerosol percentages were also observed in EDKA studies of ESP collected samples with spectra attributabic to Pb, Cu, A1, Fe, and U.
Figure 4.2.6 summarizes the fuel pin damage caused by the action of the HED.
Ten fuel pins sustained a net mass loss of 216 g (zircalloy and UO )
2 and the remaining fifteen fuel pins sustained no detectable mass loss.
Twelve fuel pins sustained some degree of cladding failure ranging from small cracks to complete shearing of the cladding and five of these fuel pins were completely sheared. The average fuci pin length completely removed by the action of the HED was 21 mm.
The average diameter of the hole in the 5 x 5 fuel pin array was approximately 19.8 mm.
The average diameter of the entrance hole in the cask steel wall was 12.7 mm.
The average entrance hole diameter in the cask lead wall was 50 mm.
The 27.9-cm-diameter cylindrical eask was completely penertated by the action of the HED.
Figure 4.2.7 is a photograph of the cask cross section showing relative hole sizes in the steel / lead cask wall. Further details of Test 6 are reported in Appendix C-Subscale/ Full-scale Test Data.
^
l
7 M
s
/
g 4
+
' /
W-1.0
~
3 UO CONCENTRATION (mg/L) 2
~
FOR TEST 46 (6/80)
E w
0.1
- 0.05-O
+
\\
p Z
9 0.01 -
r
+
( 0.005-
+
Z s
+.
m O 0.001 -
Z Q0.0005-O
[
'?
l 0.6001 -
0.00005-0.00001 0
20 40 SO SO 100 120 140 180 100 200 220 240 200 ELAPSED TIME (m!n) AFTER DETONATION e
~
I Figure 4.2.4 Timo History of UO2 Aerosol Mass Concentration Based on l
Sequential Filter Samples.
. TIME HISTORY OF MASS MEDIAN AERODYNAMIC DIAMETER AND GEOMETRIC STANDARD DEVIATION.. TEST #6. 6/80 4-E 3
2- +. -*-
+
+
+
+
+
3 2
0 69 4
2-
+
. i i i.......,, i i...
i i i s
0 30 80 90 120 150 180 210 240 TIME AFTER DETONATION (min.)
Figure 4. 2. 5 Time History of Mass Median Aerodynamic Diameter and Geometric Standard Deviation for Test 6.
)
~
4 COMPLETELY SHEARED (5)
MASS LOSS (10)
NO MASS LOSS (15)
CLADDING FAILURE (12) l l
Figure 4. 2.6 Schematic of Fuel Pin Damage and Damage Path Caused by Action of HED for 1/4-Scale Dry Test 6.
i
l I
i 1 !
i p
I,..
l_
M, 2 5. 4 ~
"~
,(,,
T-3. ','
l
'? ~
n.
i
,5,
s.+--
9 l
i j
b D
l d
..qe I
I~"&
. Els
--. i %,
fg i
)
\\ !' & [ k.
i
- 4*
-f l
<33 a.
l
~
- ?
3
}
~,. 'R.
I
, ' - - A
. 7?$-
i.
r g
I '
,,1 y
- )
\\
l']
,c: "
"_' r I
a.
l
^.
s,
~-
I g.
.<t..
Figure 4.2.7 Photograph Showing Cask Damage for 1/4-Scale Cask Test 6 (Dry) 1 l
l
. Some light water reactor shipping casks have been designed so that the outer shell and fuel cavities may be filled with water (see Table 4.1.1).
This configuration was investigated to determine the impact on aerosol production from a " wet" shipping cask subjected to high explosive attacks.
Figure 4.2.8 is a schematic for Test 7 showing the cask immediately after detonation. The cask fuel cavit and cask outer shell contained 3.96 X 10-3 m3 and 16.6 X 10-4 m of water respectively. A total respirable uranium dioxide mass of 19.6 + 2.0 mg was measured at 12 seconds af ter detonation. This mass of released aerosol from the " wet" scaled Test 7 was approximately a factor of 40 less than that for the " dry" Test 6.
A similar reduction was also reflected in the total removed fuel mass of 143 g from the " wet" fuel assembly versus 216 g in the case of the " dry" test.
" Removed" in this context is defined to be that mass which is displaced from the fuel assembly but not necessarily from the cask cavity. Apparently, the presence of water in the cask water Jacket and cavity resulted in scrubbing of the suspended particles and reduction in the total aerosol release. Further details of the scrubbing process is described in Appendix C:
Subscale/ Full-scale.
Figure 4.2.9 shows a time history of the aerosol mass concentration within 3-the chamber based upon sequential filter samples extracted from the chamber and also shows front surface reentrant filter measurements. The exponential decay of the aerosol mass in Figure 4.2.9 indicates that the first filter samples (used for respirable release calculations) is consistent with all subsequent filter samples obtained. This figure also indicates that the front surface reentrant filter saw essentially the same concentration as did extractive techniques using externally counted filters and that aerosol transport line losses in the extractive filter samples did not account for significant losses.
Figure 4.2.10 shows the time history of the uranium and zirconium concentration versus time. These samples were based on uranium fluorometric analyses of extractive filters, uranium x-ray fluorescence analyses of alternate front surface reentrant filters and zirconium x-ray fluorescence analyses of front surface reentrant filters. These data indicate that no significant line losses occured in the extractive sampling system and that the fraction of aerosolized zirconium was proportional to the fraction of zirconium present in the fuel pins before testing.
Figure 4.2.11 shows the
)
time history of the aerosol size for the subscale wet test. Whereas in the case of the dry test, the MMAD stabilizied around 2 gm, in the case of the wet test, the MMAD started out at about 1 gm at t = 0 and stabilized near 0.8 gm after 30 minutes post detonation. The geometric standard deviation (a )
g also remained fairly constant at around 2.5.
The morphology of aerosol particles from the wet test is dramatically different than aerosols obtained in the dry test.
Figure 4.2.12 is a TEM photomicrograph of aerosols obtained with the point-to-plane ESP for Test 7.
Figure 4.2.13 is a TEM photomicrograph of aerosols obtained with the point-to-plane ESP for the " dry" Test 6.
In contrast to the aerosols from the
" dry" test wherein the aerosol size distribution was dominated by the ultrafine chain agglomerate aerosols, the aerosols from the " wet" test were larger particles consisting mainly of collapsed ultrafine chain agglomerate filaments. Apparently, the large amount of vaporized water results in supersaturation within the chamber and subsequent scavenging of the' chain agglomerate aerosols.
Subsequent evaporation of the droplet collapses the
j I
-2s-OUARTER SCALE TEST (WET)
AFTER EVENT I I i
aj ga 1
m u
i M ASS OF AEROSOL (t: 0.5-3.0 min):19.6 mg UO2 II 11 f, I.
l 17.899 g f uel pins iF t_
in q r Y
/
103.6 kg STEEL g TOTAL SWEPT MASS (fuel pms) 142.5 g 115.1 g SWEPT MASS UO2 SWEPT MASS OF Zr-4 27.3 g Figu re 4. 2. 8.
Schematic of Test 7 (Wet Test) Immediately After Detonation
1.0 AEROSOL MASS CONCENTRATION v.s. TIME "5~
(CORRECTED) TEST 7, (3/81)
Of4
+ -e--e +
+ EXTERNAL FILTERS
~
+ FSR FILTERS 3
'r A
0.1 -
+
+
E E
+
O 0.05 -
+
H E
+
b 3
W e-02
+
+
0 0.01-O en E 0.005 -
W 4
0.001 --,
0 20 60 100 140 180 220 TIME POST DETONATION (min)
Figure 4.2.9. Time History of Aerosol Mass Concentration Within Chactber Based Upon Sequential Filter Sanples and Front Surface l
i Reentrant Filter Samples for Test 7.
. 10.0 CONCENTRATION OF U & Zr v.s. TIME (CORRECTED) 8- -*r TEST 7, (3/81)
- U FROM FILTERS
-e-4 U FROM FSAF
- e
+ Zr BY XRF FROM FSRF 1.0 -
++
3
^
1 0.50-
+
g
+
!E
+
o
+
8 0.10-1 l
+
u 0.05-4 0.0 i -
0 2b 60 100 150 130 2h0 TIME POST DETONATION (min)
Figure 4.2.10.
Time History of U and Zr Mass Concentration Based Upon Extractive Filter and front surface reentrant filter (FSRF) Samples. X-ray Fluorescence analyses were performed on FSRF samples.
1
)
SANDIA TTC SSEP TEST #7,3/5/81 8-AEROSOL SIZE DISTRIBUTION 6-e 8-
~r 4
e e
e-
+
2-
+-
e.
1.0 - e E
3o 0.8-
+
22
+
+-
)
+
0.6 --
j
+
0.4 0
30
, 60 90 12C 150 180 210 240 TIME POST-DETONATION (min)
Figure 4.2.11 Time History of Aerosol Site (MMAD) and Geometric Standard Deviation (ag) for Wet Test 7.
1 ':
.,.c_
ear 1.0ym.
,l
..i l
I i
i 4
, e
'[
Figure 4.2.12 TEM Photomicrograph of Aerosols Collected With Electrostatic Precipitator (ESP) Sampler for Wet Test 7.
Magnification is 40,000x.
[
I
.....ll.!:Wf - y,:
Y l
- '} }
- ll-- f f'
k e s.
m ya -<y w
y, %.. -e'
..(.
n w. :
.... g yA.-g);:. ;. -
.y.
3,; x q. j {3; 4.:. g..
- ....;m,-
. +
+
m.4[h,7,g, A-
'/ l [.:. }[.
g -
y
. j.+
. :. [..
- i., :
L 1_,)..>..[. _ c'.. g
- g. w'. - (
- ~ W '.; }.
,.4..,.
.4 g ;.4,.,; y
- +
e
+
w;'.
-l
. ).,-
.b,.
g '--'...
'.y p,. ;
- s...... e.. 4.........
7 r
y 4.. - -.,
- c. -
.r ;;
yw.....;.r.
- %y.. <
-"4
,_ g.
1 - '
. A c,.,' {%.yy,a.'
..,.y u,
.., ;.',:u
- s.
,g 9
s
<gs.A ;,... ;..... Y 4 'Q ~ s w
-,t
.e.
v.
. r. ;,-..
. o
, c. m. p..
p 4
t, ;,
<G,.*:yl+.;:5',".?,.'*
~
\\. ~ -. - :.
z.
.'..,.s......:
4
.a 4.,
.m.
s g
4
.. ~.
s..... -
4.-
,s Q,~
... b 6 t,
- l. j.'.[;j g ?_;.[c.+,.. -[ :. }., '
q.l-V u ;
- - Q ~'.. -;
.% - L 5,.f:
+ 3 Q_.f 1.
V. s
.t,- - gy s
...., g, a
,;;.. h. i c'
~ y -: '
- <g 4.
' y..
- i. [ g y )_!c("';; y 8
- p.., ;_ _ 3, 7..,
'q
. \\1.1 7.g.*l4[..^.:,[ j (jy. M} j. fy;
._.3 n ;:
..->'rz y -.,y ' _
4_
_ sq lp
,
- & -) 3-. 3.'
..e :.'
l
- - \\.i,...
.. e w.=-
w
- [, : :[,
- 7. _. :[.: :
i l
k'--
, _..,:.. f k l[ ;., ;
' _ '. : :. ". '. R ' l ~ ~ [..... :;, _,
T
. l.' -
.,l
' _j
.g
.. j...
g.:. h./.
k _.-;,,.:..- l;' s -...g ;
y * - '..;. _
\\..., --
.:,,l.-' ;...r *..;;..#...
.., e
_:} g %.f.....g 3
?., 4-
. ;. ]..,...
, '.. g! -
3
" \\...,
-.. > _ _l h._ ~.'. :. f ),y' ' - Q N,f '. ' -..
'l
- [.,;
y.'[.
g.% 4
. k [s.
h,.
- " ph-
~~^
r' '-
Ko,Dhl-4
- . f N S'I'
- l,_,'
- j.-
(.-
pf I'.
. -Q D '
'.?
l U i-
.?,.
W-z:. }..
9
....~-.y...-
.,_ ; 4.,. '... +,
p..
y-4.t,.
y...
3.--
e.
- .y -
41Jh.
.3
- z.
. ? y:
......y
,.a-r..,s ys p,,
f,.v.
n.
l.f.k
^
f fl
/
l
- ;- - *W 1
- *y, '
'(;
y-(e..y _ ~. %,, l
' W
..;.. = :.. -
,l
- j. :'...
-y _,; -..'
..=g..; _
- w
}
7 g....4 E
_,. m pl'.f:'j-
... ; y,.,,.a.
~
Q: ' y p.
.1". 4l. K
. 52 gp[f * ' -'l'.,.'..-
.e.
.s.,
A k. h.~ s, % '. h. I..
s.i.l*-
. ?" -
T... f 19 f ;) ?
d:
{
s.' i [.<1;..
-'; :..:M..'l Y,. ? '-.
? A..%
- ' ? h.h f,._'. ; ' h ?
'.3 T
,..A 1
- 4 n.
y'
, % L.._
,. c. pl, ' *,_
f.f^l
.?
f';. -, :. e
.[, "**'; ' ' * ',l ;. ' ; '..
..)
...T l; ".1
.., 4.,.. 'g
- . i ;
- .%.* :..y_W.* - ~
,.T
^
- '*. ;*.= ';...,
W
\\ [....\\.ll:? l ' I N f =_. ;.'.:
h
. f ~ ~ [.',0 l?;_l ~..g, g f!$:fh.
.j. '
..i 3
9 kN$$ kl.I.k.
l f,lf.f ;* ' I ]? ?. ':'
$4 f. *f 5
!-l.N.?
5 5
5.0 pm Figu re
- 4. 2. i 3.
2 04 Photomi c ro f r ap" M hrosols Collected Wit h Elect ro s t a t ic Pre : tp it e.o r (ESP) Sampler for Drv ^:'e s t 6.
I
, filamentous structure of the chain agglomerate aerosol. The resultant size distribution is shifted from the 2 pm aerodynamic diameter typical of the chain agglomerate structure, as seen in the highly concentrated aerosol from the " dry" test to a smaller size distribution typical of particle clumps of mean diameter of about 0.8 pm in the case of the " wet" test. A large amount of debris varying in size from several micrometers to centimeter-size chunks of metal and fuel pins remained on the floor of the chamber following both types of tests. These samples were passed through a series of sieves ranging in size f rom 2,000 pa diameter to 38 nm diameter. The results of the sieve analysis are shown in Figure 4.2.14, where sieve diameter versus cumulative mass percent for both " dry" (Test 6) and " wet" (Test 1) are displayed.
Both total mass and UO2 mass distribution are shown for Tests 6 and 7 These data indicate that the total debris were bimodal in stze distribution, probably as a result of two different mechanisms of formation: (1) a mechanical fracture of the UO2 fuel target; and-(2) vaporization /mciting of some of the components (lead and steel) in the 1/4.tcale cask.
Figure 4.2.15 shows the results of aerosol mass concentration calculated }-
from the continuous flow condensation nucici counters.
These data reflect the fact that the " dry" Test 6 produced 40 times as much acrosolized mass as did the " wet" Test 7.
Additionally, the wet test shows a " catastrophic decay" of aerosol mass concentration after a time period of ~200 min postdetonation.
This phenomenon has also been reported by investigators in the Netherlands.10 Figure 4.2.16 summarizes the fuel pin damage for Test 7.
Eight fuel pins sustained a net mass loss of 142.5 g (UO2 and zircaloy). The remaining 16 fuel pins sustained no detectable mass loss.
Nine fuel pins showed some degree of cladding f ailure and eight of these fuel pins (the same fuel pins which sustained a net mass loss) were completely sheared by the action of the HED.
The average fuel pin length which was completely removed by the action of the HED was 20 mm.
The average diameter of the entrance hole in the cask steel wall was 12.5 mm.
The outer 0.76 mm thick steel shell was peeled back from the cask as shown in Figure 4.2.17 and was caused by hydrodynamic forces from the HED transmitted through the water in the outer jacket. The average entrance hole diameter of the cask lead shell was 48 mm.
The 27.9 cm diameter cylindrical cask was completely penetrated by the HED.
Summary _of Subscale Test.
Measurements conducted in the subscale test program are summarized in the following statements:
1.
A measured released respirable quantity (maximum) of 0.18 g +,0.05 g of UO2 was determined for a 1/4. scale dry cask Test 6.
A linear extra polation f rom the 1/4-scale event to a full-scale event using the selected reference HED and a full-scale truck cask containing three PWR assemblies yields a respicable acrosolized release fraction of 0.0023 percent of the total heavy metal inventory.
m
e 3000 ls's' SIEVED C ASE DEGRIS. TEST *S. 6/80 total mese
}
,4
.,e
}
f, e' s00-g 330-
,e l
, /,?
UOs 5
/
/
/[
100-
~
- r
.=
30 se 90 70 50 30 10 3
CUMULATIVE PER CENT OVERSIZE SIEVEO CH AMBER DEBRIS TEST + 7 Ile-SC ALE W ATER FILLED C ASK 2000 1000 -
TOTAL MASS %
500 -
p SuD a 2 42 um
[
ogte0 e=
7 100 -
/
N 3g-1g/
UO DISTRIBUTION 2
3e 5
s
/
.y to -
p
/
5.0 -
l
/
o 0
9 9
f 10 99.99 98 90 70 $0 30 to 2 01001 CUuul ATIVE PERCENT OVERSIZE Figure 4.2.14. UO2 and Total Mass Distribution Based Upon Sieved Debris Released From Cask Into Pressure Chamber for Tests 6 and 7
. 4 10.0 5.0 _
MASS CONCENTRATlON v.s. TIME L
i A
A md 1.0 -
Ag CD
[
0.5 Jb A
z A
^
O 4,
b A
e (I
01-F-
e a
2 0.05 -
e g
o z
O ADRY o
u) 001'-
y)
WET 4
0.005 -
7 i
0.001 ---
O 100 200 300 TIME POST DETONATION (min.)
1 l
!t i
Figure 4.2.15. Total Aerosol Mass Concentration Based Upon Continous Flow r
Condensation Nuclei Counters of Tests 6 (Dry) and 7 (Wet).
l
.l r
8@@@h
$>4 COMPLETELY SHEARED (8)
NO MASS LOSS (16)
MASS LOSS (8)
CLADDING FAILURE (9) i Figure 4.2.16. Schematic of Puel Pin Damaga and Damage Path Caused by Action of HED for 1/4-Scale Wet Test 7.
l
[ !
l E-
'l, o
i J
=
I
$f v$$5N
.A 1%d.
d
)
,7 1
Y 3
i I
%% 9;
.o i
k F
g.
fQ' Q.
.NT,
F17ure 4.2.l'.
P"otograp-5".aw;.g Ms< D ria g e for
- 1. 4-S c a le Cask Test
,det}.
. Law of ex;t '.?;e pr,ducei by MED.
. 2.
A measured maximum released quantity of 19.6 t 2.0 mg of UO2 was determined for the 1/4 scale wet cask Test 7.
This aerosol mass was approximately 40 times less than that for the dry cask Test 6.
Based on these results and the intent to select the more conservative test case, the dry cask configuration was chosen for the reference base full scale test.
3.
Testing and aerosol sampling techniques which provided aerosol samples as a function of time following an explosive attack on a target in a contain-ment structure were evaluated and developed for the full-scale reference base test.
4.3 Quantification And Characterization of Material Released From A Full-Size Reference Sabotage Incident:
Full Scale Test A full-scale test subjecting a 25.45 MT generic-truck cask containing a section of a single surrogate PWR spent fuel assembly to an attack by a T-reference full-scale high explosive device (HED) was performed to extend the
~
results of the subscale experiments. The surrogate fuel assembly consisted of a 15 x 15 array, 1.2 m long zircaloy tubing filled with depleted UO2 fuel pellets (9.33 mm diameter by 15.2 mm length). The dimensions and mass of the U02 pellets were similar to those of fresh reactor fuel pellets. The stain-less steel / lead cask wall consisted of a 2.54 cm thick stainless steel outer shell, 21.3 cm thick lead middle shell and a 1.9 cm thick stainless steel inner shell. The cavity dimensions were 38 cm in diameter and 356 cm in length. The shipping cask was placed inside a 3.1 m 1.D. x 0.02 m thick wall x 6.1 m long cylindrical chamber for aerosol containment. The HED was mounted and detonated externally to the chamber.
An explosively actuated sliding isolation valve placed between the HED and chamber port was designed to close milliseconds after detonation in order to prevent the release of the source aerosol and fragments to the surrounding area.
Eleven sampling ports penetrated the chamber at various locations.
These 2.5 cm I.D. sampling ports were closed before and during detonation by remote controlled pneumatically actuated ball valves and were opened shortly after detonation in order to allow sampling of the acrosol.
Figure 4.3.1 is a schematic diagram of the test setup for the full-scale test and shows the spent fuel shipping cask inside the acrosol containment chamber. Fuel pins weight'g 258.048 kg were n
configured in a 15 x 15 array inside the cask cavity.
A 61-cm cubical high-pressure fiberglass filter was connected to the chamber via a 15-cm I.D. Port to vent the high-pressure gases from the chamber to the atmosphere. This filter was designed to handle the high-pressure shock 3
wave and a gas flow rate of 31 m / min. The filter had a flitration effi-ciency of 99.99 percent for 0.3 pm monodisperse aerosols. After the test, the filter was washed in an HNO3 acid solution and analyzed for uranium using uranium fluorometry.
Based on experience with subscale tests, the battery of instruments listed in Table 4.3.1 were selected to sample and characterize the released particles
. EXPLOSIVELY ACTUATED ISOLATION VALVE (C.gV T 0.
01.
Ej STEEL TEST CHAMBER (3.10 m ID X 3.56 m LONG)
/ X 2.22 cm WALL X 6.10 m LONG)
\\
/
4 c,
^
__p.6...................................................~ 3_
ASSEMBLF Ilill lllllllllllllllllllllllilllllllll111111111llll1lllllllly
_jl r
N llllll1lllllll!Illllllllllll!!llllIll!!IllllllllllllllllII'!Il,dk, i
b
(
i e
i u;
m T
T YT T
T/
4,
j
_m ft'
'N !.' yj w
%.*:* -'97 m.f,y STEEL STOPPING BLOCKS BEFORE EVENT 258.048 kg FUEL PINS: 201.053 kg UO2 + 56.995 kg Zr-4 Figure 4.3.1.
Schematic of the Full-Scale Reference Test Configuration Before Detonation Showing Finned Cask Inside Pressure Chactbe r.
1
. TABLE 4.3.1 Sampling Instruments Used in Full-Size Cask Test Instrument and Applicable Size Range Purpose of Samples Analytical Method Cascade impactor Aerodynamic size and geometric Gravimetry (Cahn micro-0.5-12 um standard deviation, rotal aerosol balance) and fluorometry mass and UO2 mass, for determination of uranium.
Point-to plane Particle morphology, count distri,-
Transmission and scanning electrostatic button and elemental distribution.
electron microscopy and,
precipitator energy dispersive x-ray E~
0.01-12 um analysis.
Filter, 47 mm, Provide a time history of total Gravimetry for total i
sequential, aerosol and uranium mass after mass and uranium by 0.01-12 um HED detonation. Samples for fluorometry. Nitrogen surface area measurements.
absorption for surface area measurements.
Filter, 37 mm, Sequential filter samples from Gravimetry for total front surface filters inserted inside chamber mass of aerosol and reentrant filter to compare with filters obtained uranium by fluorometric (FSRP) by extractive techniques to techniques.
0.01-12 gm address aerosol line losses.
Screen type dif-Total count of aerosol particles Optical light-scattering fusion battery vs. time af ter detonation.
instrument.
with condensa-tion nuclei counter (CNC) 0.005-12 gm i
-3 9-TABLE 4. 3 1 ( Continued)
Sampling Instruments Used in Full-3ize Cask Test Instrument and Applicable Site Ra nge Purpose of Samples Analytical Method Electrical aero-Size distribution parameters for Electrical mobility.
sol snalyzer particles between 0.01 am and 3
(EAA) < l.0 um 1.0 am.
Sievus ( 38 -
Provide size distribution data on Mechanical steving fol-2000 um) larger particles of surrogate lowed by weighing for spent fuel.-
total mass and fluorom-etry for uranium deter-2 mination.
Integrated time Provide size, morphology, and ele-Scanning electron mi:ro-I deposition alu-mental composition data of parti-scopy (SEM) and energy minum planchets cles diffusively and explosively dispersive x-ray (ED3)
(0.01 am to deposited on chamber surf aces.
analyses (Z > 11 of
- 00 m)
. deposited pa rticle s).
I Rotating Plate Provide Size, mo rphology, and ele-Scanning electron micto-3amplers ( RPS ) :
mental composition data on explo-scopy (SE4) and energy ilscrate time sively deposited particles.
dispersive x-ray (EDS)
(0.01 am to analyses (Z > 11 of 4
100.m) ieposited pa rticles).
l
O 1
l 4 and debris. The sampling procedure was designed to provide measurement of high velocity particles as well as the lower velocity aerosols within the chambe r.
From these data, calculations could be performed to determine the initial release parameters, such as initial fuel mass aerosol concentration and released fuel mass. Aerosol size parameters as a function of time were l
determined from cascade impactor samples obtained at sequential time intervals after HED detonation. Similarly, filter samples provided a time history of the change in elemental mass concentration and total mass. Changes in particle morphology and elemental composition as a function of time were determined using sequential point-to-plane electrostatic precipitator (ESP)
}
samples. Scanning electron microscopic (SEM) and energy dispersive spectroscopic (EDS) analyses of the ESP samples provided information on j
particle size, morphology, and elemental composition. The electrical j
analyzer provided information on size distributions for size modes j
than 1 um aerodynamic diameter.
l i
l A Sandia Laboratory developed time-resolving rotating-plate aerosol 37 sampler was used to collect the high velocity particles for examination by SEM and EDS. These samplers consisted of a collector plate which rotated beneath a cover plate containing two 3 mm x 4 mm rectangular sampling ports. Three millimeter diameter copper grid discs were placed 6 mm apart on an inside circular path and 12 mm apart on an outside circular path of the rotating i
collector plate. This grid spacing allowed sampling at 200 ms intervals. The sampling grids were removed af ter the test for examination using SEM and EDS 3
j analyses. The problem of diffusively deposited particles covering and obscuring the early-time explosively-driven particle deposits was solved by pressurizing the sampler housing with air so that the gas flow exiting the i
sampling ports permitted entry only of early-time, high velocity particles and prevented entry of lower velocity aerosols.
In addition to these surface deposition samplers, time-ictegrated deposition aluminum planchets were placed at various locations in the test chamber to sample particles deposited over 4
long periods of time. - These aluminum planchets provided information on the size, morphology, and elemental composition of dif fusively and explosively j
deposited particles.
i After the test, all debris were collected from surfaces inside the test chamber and separated into nonuranium material and material suspected of containing uranium. All suspected uranium debris was sieved, the total mass determined, and uranium fluorometric analyses performed. The intact UO2 j
fuel and zircalloy-cladding mass were measured before and af ter the test and j
the quantity of UO2 and zircalloy mass removed'as a result of the HED attack vas determined.
l 7
Figure 4.3.2 is a reconstruction of the test configuration immediately af ter detonation. A total UO2 fuel mass of 195.593 kg remained in the fuel l
pin assembly in the cask af ter the event which indicated that 5.460 kg of UO2 fuel was removed from the fuel assembly as a result of the event. Table 4.3.2 summarizes the measured values and results of the full-scale test.
Af ter detonation, 198.504 kg of the UO2 fuel remained in the cask; 2.549 kg 1
. "'s* M EAsu IEl (C AvlTY.1s la STEEL TEST CHA40SER to so X s in WALL
/(t2 A140 In LOe00 I
a \\
p AEROSOLIZED UO2 P
'**FbEL AssatsSty M A SS = 2.93 g F
m h-aL 9
'IIllIllllllIIIC, e1illlllllllllllllllll'lllllll!!il J
's
$,11!lillllll!C 7... !llilllillllihililpillHipili_
_i n
( M
~'J J--
1
\\ )
I L
f te;
.i.
T T
T T
Y7 p
fr
% :.=?.' y,
l
~
,2.* : *:.'
- y m g1 STEEL stonemo stocas AFTER EVENT l
REMOVED UO MASS. 5.460 kg REMOVED Zr-4 MASS : 2.063 k g 2
Figu re 4. 3. 2.
Sche:natic of the Full-Scale Test Configuration Immediately After Detonation Showing Damage and Net Mass
!oss.
. TABLE 4.3.2 Summary of Results of Full-Scale Test No.
PARAMETER VALUE 1
UO2 fuel mass before event (kg) 201.053 2
Zr-4 mass before event (kg) 56.995 3
002 fuel mass remaining in cask after event (kg) 198.504
}
4 UO2 fuel mass removed from assembly (kg) 5.460 5
U02 fuel mass released from caska (kg) 2.549 6
Maximum UO2 fuel mass fracturedb (kg) 20.820 7
No. of fuel rods before event 223 8
No. of fuel rods with mass loss 111 3
Maximum length of UO2 fuel fractured (mm) 275 10 Maximum length of fuel rod removedC (mm) 76 11 Average entrance hole diameter of cask (mm) 152.5 12 Average exit hole diameter of caskd (mm)
None 13 UO2 fuel mass released as aerosol (g) 2.94 14 UO2 fuel mass released as respirable aerosol (g) 2.94 3All sizes.
bNot necessarily released or removed.
C As a re sult of t.".e event.
dHED did not completely penetrate cask.
0 1 of UO2 was released f rom the cask as a result of the event.
Approximately l
0.115 percent (2.94 g) of the released UO2 fuel was airborne aerosol and all of the airborne UO2 aerosol was assumed to be respirable. Fifty percent (111) of the 223 fuel rods sustained some degree of mass loss (damage).
Figure 4.3.3 shows the damage to the simulated fuel bundle. Approximately 10.3 percent (20.820 kg) of the pretest UO2 fuel mass was fractured. The maximum missing (removed as a result of the energy loading) fuel pin length was 76 mm.
The entrance hole in the 2.54 cm thick stainless steel skin was i
approximately 15.25 cm in diamater (average). The opposite cask wall was not l
completely penetrated but the 1.9 cm thick inner stainless-steel shell and l
14.29 cm of the 21.27 cm thick lead shell was penetrated. The outer 2.54 cm thick stainless-steel skin was not breached.
Figure 4.3.4 shows a
~
time history of the UO2 aerosol mass within the chamber based on sequential filter samples. A maximum 002 aerosol mass concentration of 23.8 gg/l was detected at 12 seconds postdetonation. Using the containment chamber volume i
of 42.29 m3 and a measured uniform spatial concentration of 23.8 gg/l, a total released UO2 aerosol mass of 1.01 g was calculated. Another 1.93 g of 3_
UO2 was collected by the pressure release fiberglass filter assembly.
Assuming that 100 percent of the measured UO2 aerosol mass is in the respir-able size range, a total respirable Uo2 mass of 2.94 + 0.30 g was released from the cask as a result of the event.
1 The debris swept and vacuumed from the chamber walls and cask surfaces were collected, sieved and analyzed using uranium fluorometry to determine the i
size distribution of the larger U02 Particles.
Figure 4.3.5 shows the a
cumulative UO Particle size distribution of the sieved debris.
Extrapo-2 1ation of the size distribution to the respirable regime suggests that less
{
than 10-3 percent of the chamber and cask surface deposited 002 debris was I
smaller than 10 gm aerodynamic diameter. Since the collected surface debris l
contained = 540 g of Uo2, the inference is that less than 5 mg of UO2 particles deposited on the chamber surfaces was smaller than 10 gm aerodynamic diameter.
Figure 4.3.6 shows the aerosol size parameters versus time for the full-scale test. Whereas in the subscale tests, aerosol size distribution MMAD was approximately 2 gm for the " dry" Test No. 6 and slightly less than 1 m for the " wet" Test No. 7, in the case of the full-scale test, the MMAD was 3 gm immediately af ter detonation and stabilized near 2.5 pm, 30 minutes after detonation. The geometric standard deviation (09) also was larger for the full-scale test than that for the " dry" Test No. 6 (2.6 versus 2.2).
The larger MMAD in the case of the full-acale test is attributed to the fact that the relative ratio of the chamber volume for the full-scale test to that for the subscale test was smaller in proportion to the explosive charge mass ratio and, therefore, a much higher concentration of aerosol from all sources l
re sulted f rom the explosive attack in the. full-scale test. A higher aerosol l
concentration would lead to more particle interactions, agglomeration and, I
hence, larger diameters.
Figure 4.3.7 shows an SEM photomicrograph of a typical rotating plate sampler (RPS) sample taken 400 millisec after detonation. High velocity I
particles were collected on this sample. No uranium particles and no uranium d.
Y.*g g',.'..;. g..
.. ? <,,, y.-
4.
. /
s,,.7,
. ;- s..,,,,.,.; y. - 4
. u,c.
m s g., -- -
3
. n.r,,...-;w m,v..,,..., - -
3,
.g s,
,T.
= _
y e,-
e s
u,,
v...
-s.-
. ~-
x f '- -
a.
4 g
s g, :. i.,
c
. a ). : -
- t. ~,..
- p. 4..
s 9...., <:... -( y, e. ',
,-.c-
.4 s?.
o.%'
sn,.c
- 3 a
+,
^. -.'
w
,e-
..- ! 4. t ' _ c, s 3
i:
a
',. '.d,., A"e.-
r e.
.rs
,b L.
.g s.
-.s,t - - - r
.n g
,t.
, s. y'.m.
.s
_7:
. s,ky
..r... -..:
s.,
,...t f... -.. s,f..
wo. _., - -
,r a.
'g
.l-a'
.g.
..e....
1
+y,.
a -.
q 4;, ;
+
f..
.\\
. 3..
<,-)g. - 3.,.,. ( -
~ ' -s f, q -,
}..,
w.
~ q.y w
< s..
L.,
4;
,\\
l,,
y a
.gg
- d., ' *
.j~,.
- 9 s'
=,
v.u',, ;
s; y - -,9 4
,,4,e,,-
<y
. s y
2
.5
+.
.. :,. f 1.,
..4+-.
,,.-4 s
s v.
n y. '. '...
- lts
,+
r ** ;
j r.
'< ~
v.,
g.'
,.),,. ' y.
.- He ig,,,..O 4 ' - p-. -
.a 1..
- i? P:
~
E,
<[,,
l 1.,., - 4
)l, ~..
',.t N.
h * '_
.,e a
w'
'h l,,
1
.'*[
i
, p
., L.. ) I,
.I
. *[kd '" '<'
[
k.,
.[,p. '
[
s
'1
,- [ r E
x. g(,,, '..
- N
,b
+.-
t A, *. -,;.,...,.
e, o..
,'s
, g
~. p.,
-~
- '.,. -.p
.s i
r 4,
u., ;, gr, 2 n,,
".: l. < ),,
J'
-- ~
_-. m.
,,* '; -l f,. '
p
_ dum
..I
[,
k g,
., T 4.?,..,g
< k t.
.,e.
,* \\
n,., ; s w
,.,6.,>
n. . %
3...
s, 83(
J w
,~,..'.ar*,
', " 4
.-, ~. -
..,.." "kf,"'. ' _*, +
Y
},.. &>h
. '/. * ?z.-,.'
6 w.
+;
a.,.
-., w
, s 4 his,
.y
.s' *
.. ' *..,, e <
+.t.,
, I. ' ** (
+,#
+.E \\., -
w
~-
y -
e u
.y" sr
/ M3..:
.,f'[
- f-j..,'
,t s
- 'h.',.-.
y 4,
-#i.. '.= ' _, ',"
' #, ', ;I
', o '[ y
(-] ^ '}
..' ', f ;-l j'
+
.yy
. ;e
. f E, ; .., ', -
- 3*.h ' g-[7
.. b.
~
M u
y I',,4 a.y i7 J Ml.'.].'} ( p.;(-Q( i ~ ~. *- j -;[
a.
p
- * [ @r ],
J
[ ~, ", ' -
?,,7 g.
4,
,, 4
' i 2 b
',.k'.Q 8 $f.a
+.
h'..,
.q. C$, ;*-
vi ;
j[.
...[ &f,
- e:,',,;. u :
t :.
~w~,,
p,,
w R
~
.'. 1 e:s.
~
- v. y.'
r, y.
,3:
s I
^,, t,
(
- ' > Q
%. t*;.
y.
...'.5,
.R..~,','t.'**
e,
,..7
, ;.,,.,.*,-.-',. :..[a. : > na g.
a<< v c_
- r.
C
.a
- 9,,,
.3 y;
?' ~ $ 5.&.
a,.
_s
- y 1
- Y., s,'S, e
..e-i
.n..,
y
- r.,-,
. '.,. e,
- A e l'. yf <
.,~.aq-, ' ; -- )....e..
....r
~ 1.N.. 'y,' 's > (*
, y1
,4 I <s, ;.,: ~
&,' s t
>f' o,ss,..'3.:.
,.. f%
4 1g s.
+.-
V
,l,
,'4.r,.,.'<,e
.-.r' n
y
,y.r.
aspefv -
.'.a,'y-t p
.,...y.'
,,j,,..,.3, m m.f; ' g., y,.
M U -
~ M **
R. t, g 3.g c
- e. :
M;s,.y,
~
&fF
-~..
-4 4
v a
v-
~
1 I
Figu re 4,3,3 Photograph M 4 e 2 Fuel Assembly Showi.sg Damage Caused g
e e rence N st.
I
~ - - - - ~ -
o 20 3
I UO CONCENTRATION (CORRECTED) 2 E
FULL SCALE TEST +8,1/10/81 zop 5-T z
W Oz Oo t-2 0
$ 0.5-Tw<
O" a
i i
e i
i e
i i
i i
0 20 40 60 80 100 120 140 160 180 200 020 TIME POST DETONATION (min) i l
l Figure 4. 3.4.
Time History of UO2 Aerosol Mass Within Pressure Chamber Based Upon Sequential Filter Samples for Full-Scale Reference Test.
. M i
i i
e i
i i
i TEST a FtLL SCALE 2000 CHAMBER DEBRIS TOTAL MASS a 1000 UO
~
2 o
o o
500
-r x
MMSD = 210 pm W<
o O = 2.5 g
l 5
100 o
a g
50 i
I to l
5 1
I I
I I
I I
i 0.01 0.1 1.0 to 30 50 70 90 99 99.99 l-CUMULATIVE PER CENT (ACERSIZE l
Figure 4.3.5.
Total Mass and 002 Mass Size Distribution for Sieved Debris of Bill-Scale Test.
l
a
, AEROSOL SIZE PARAMETERS v.s. TIME: FULL SCALE TEST 2.6 e
_e s
E e,e ty2.2
~
~
_r e
1.8 e
e 0
1.4 3.4
~
E 3 3.0
- ee o
- 2.6 -
i s
3 e
2.2 E o
B l
1.8 O
i i
l O
40 80 120 160 200 TIME AFTER DETONATION (min)
O FAULTY DATA POINT (LEAK IN CASCADE IMPACTOR) l Figure 4.3.6.
Aerosol Size (MMAD) and Geometric Standard Deviation (o ) as a Function of Time for Full-Scale Test.
g l
1 1
1 l
l ? l:.?'
-]. jj.'..fQ [p.'y
. ':'i
- ' }- Q r.
'... [ g j :9 y.;.-.(
yl.
- r. Q: '
- 1. [. + :. h' s[" R. c '.,_
. ag. ;. ?
..c '.] 3 ~:;
,. --. W,'
-o..[ q _.. g,..f_,
y.
=,,,
\\
. 1. r z.
s j f[
" 3.m [ -
~.
[.
. %t.
ey
. s
, -z; 3,.:.....
. ~..
- .
- t. -
N
. g,c..,
Ty,
.<, -... s.,
- ~.P. '.,
?.
r g ) (..g.y. ~.
';5.. s - -
6 '.t.
~
..m
<s s
J., -
y,
... i.;w
<4
~
..,e-
.- g.
'?
j;..g; i
f
,.. 6 A,
y,g y
+':,y, t....
.s s
- .
- . ~.;. -
s._
as p,y r.,
..g..
-f.
't
~. _
s,-
7._
+s
.5
- j., --
},,,,y;.
~,..;
-.}?-,
j s
'. T]
'. ' Q,,
.Q,:
%f
+
.q a;
.y
..yp
.y:
.4... - - - <,
A,. _.
... ~
o.
- g..
r.:y, %y,
.,s,_
s.( e '.s,-
. ~,.
r
..b
.{y '
- 4
~ {) ; ', ' ', f, i,.7
.. g l..
- MT*' _,.. y.,.
.; - ' ; [,_Q:::
+
c n -
E
,y y.
7,'
y.
~ % _.;. ' u. '..
. w
.. 3 nl-l *
- .. a...;.-
q_
. %.v.~.it n,?.~.'
. 7%
4.-
-, Q,. &_ - ((,_;p,,
f-y
.y..,_
~;;,a }
- y. J...
~ww,,.,,.;. !.. r.
1 ^ ".:.
.y
.y
... j..- -
4
' '.g-
? ;
l 1,.
.l., 4 -
o r 1
'?
w
-n A
'Y.
m8N...
g,
+,
- ',.'.'.,1 g-
, rl m
'+'4 3.
1,
%...h ;E.;Y h.
g4
.?
g:.
Aw, t
~
k.
66
.h'..*% ' ~ ;,,, l
,[/
M
..'.'.y-U e.g.
- ', '. ":-. $ -l.'-l' ' ;< - '.
~~
e
..?
j
'l N Y)f. f ? ',
,j Y,.h*.,' -.f
' " '._{'.+t. gl\\ '., \\,,,
. -<-s
' f.
- f.o 3..}g
/g?..
a 1
T ' ', }
- .l.?
. g ]...
,,'..6*
- 4.,
,*1.
L' a'
i ff t
' ' e
,l c
. ' ", j
%, '. [, p
_ cl'og ' M....' [
3
.r
+.
e,,
-+'.:
,.,.:, r %
l 4;
,}-
r.
. - 4 '
'a y
?
s
.e e,
4.#.',
. Y.4 %
j.
'.2 -.
-,' - 3,y 1
.", hj i
- ' w
,es
>; 3;,
. A...
,p.
l; _ - =.
- g.,,..
v
..y ~-:.y
,._/ ;,
. j.,,,, * ' #
, :. ' ' _ _ ', ',/ ',f
? 6 Q-
._9
,g-gy
. :q :
y
-s-u.
".w.
h ',
_ g.
,.'.[g,h 3.
4
't
. ' k, 9 %%&,,. c.*..
,; e \\;;,.
s 3,,
".5,,
j
)Y.... '. ~'_
'.h-'..
.3.-
.. f %' ai
?
.5
i'
.,..
- p'.
g t
..,, py 4.,
- e 8
- h. '
,.S, 4
...,.4.
,1
" -+
(,
I ,,;;
.y'-
~
3
.. i (g 4 i
". 4,
t, i
'[' *:
'd-M,,
, ' qf _
f.. %
- L 't %,.
W 4 E.._ M i
f.
-f,.,.
'{
T, p
..l ' *t
.f
.p. E...
. 4
-/
c-
!.'.' r.i- -'e f-~.f*S>.-
M i
i s-
,...j
,N-y.
g.
-...~
4,.
is.
s-**
I Figu re 4.3.7.
Scanning Electron Micrograph of a "'ypical Rotating Plate Sample Taken 400 msec Af ter Detonation of the Full-scale Event.
Magnif ication is 200,000.
f
=
=
l l
) coating was detected on this sample or any of the other RPS samples analyzed.
However, numerous spherical globules of stainless steel (Fe, Cr, Ni) ranging in size f rom submicrometer to tens of micrometers were found on the copper substrates. Additionally, networks of submicrometer lead fibers and wisps of lead were detected on the surfaces of the stainless steel globules and the copper substrate surfaces. Figure 4.3.8 shows an SEM photomicrograph of a time-integrated aluminum planchet sample of both high and low velocity mate-rial ejected during and in the time period following detonation. Fractured particles of uranium ranging in size from 70 to hundreds of micrometers real diameter were detected using EDS.
The morphology of the uranium particles was irregular and was indicative of a mechanical fracturing mechanism. The aluminum substrate was covered with stainless steel globules and both the stainless steel globules and uranium particles were covered with a network of ultrafine submicrometer lead fibers. This lead coating had the appearance of having been formed from a vapor state. The stainless steel globules had the appearance of having been deposited in a molten state.
The analyses of the RPS and time-integrated aluminum planchet samplers I
cuggested the following:
(1) A lead vaporization / condensation aerosol forma-tion mechanism, (2) a stainless steel melting / solidification aerosol formation mechanism, and (3) a UO2 nonmelting mechanical fracture aerosol formation mechanism.
The conclusion from these analysis was that the maximum temperature seen by the fuel mass was greater than 1744*C but less than 1850 *C.
Estimating the precision of measured parameters and using standard error propagation technigpes, the uncertainty of the measured respirable UO 2 mass released from the cask was calculated. Uncertainties of the measured para-meters, such as mass concentration, chamber volume, sampling flow rates and mass fraction of uranium determined by fluorometric spectroscopy were estimated and propagated to determine the uncertainty of the derived U02 respirable release mass. The estimated error of the released U02 aerosol mass based on this analysis was +10 percent.
Figure 4.3.9 compares condensation nuclei counter (CNC) data where mass has been calculated for both the two subscale Tests No. 6 and 7 and the full-scale tests.
In all cases for the dry tests, mass concentration starts out at about 7 mg/l and decays in time through several orders of magnitude.
As previously mentioned in the case of the subscale wet Test No. 7, the concen-tration of UO2 aerosol was a factor of 40 less than that for the dry tests (subscale and full;-scale) and a catastrophic decay of mass concentration was observed at about 200 minutes postdetonation. This indicates that the presence of water in the cask jacket and cavity would reduce the consequences of a malevolent act on spent fuel shipping casks.
The results of the full-scale HED attack on a full-scale generic shipping cask may be summarized as follows:
A total respirable Uo2 mass of 2.94 + 0.30 g was released as a result of a successful sabotage attack event on a single PWR fuel assembly truck cask containing 0.5 metric tonnes of heavy metal charged to a light water reactor.
This released quantity corresponds to a release fraction of 6.0 X 10-6 of the total fuel inventory in the cask.
l T Q
t y
s.
y >-
3%
p-
. ', (
r 1
=
-}
\\
- ' e i 1
n.
r
~
4 T4 *-%
g
^ ".
c y
".se e t.
-e.
%J' M '
M l
l Figure 4. 3. 8.
Scanning Electron Micrograph of Time-Integrated (10 minutes) Aluminum Planchet Sample Showing a 70 sm (real diameter) UO2 Pa rt icle. Magnification is 20,000 X.
s
. 10.0 s.0 -;
MASS CONCENTRATION v.s. TIME b
A
-d 1.0 - +Ag G
g,%
[
0.5 JL g
Y A
+
z A
O
- e
+
A' F
A
+
+
4g 0.1 -
e
+ FULL SCALE H
4
[
0.05 -
o O
A z
O ADRY O
u) 001-WET y) 0.005 -
e l
0.001-0 100 200 300 TIME POST DETONATION (min.)
l Figure 4.3.9.
Comparison of Mass Concentration (mg//) as a Function of Time for Subscale Tests No. 6 (Dry) and 7 (Wet) and the Full-Scale Test.
f
. 4.4 Development of a Correlation Between Depleted UO2 and Spent Fuel In order to obtain a correlation function for the aerosolization of spent fuel versus depleted uranium dioxide, it is useful to consider data produced during this program and that generated during similar experiments at BCL.4 Since no aerosol measurements were attempted for a full-scsle reference base attack on spent fuel, data from subscale attacks on both spent and depleted uranium dioxide must be used.
Subscale Tests Performed by EC&C/INEL Tests performed by personnel at EG&C/INEL in support of SNL's research effort involved single fuel pellets of both depleted uranium dioxide and H.
B.
Robinson Unit-2 spent fuel subjected to subscale attack by a scaled version of the reference base attack device (see Appendix D for details).
Two types of measurements were made that can be used in calculating a spent fuel to depleted uranium dioxide aerosol production correlation. The first measurement type was obtained from filters which collected aerosols generated
,E-following exposure of both spent and depleted uranium dioxide fuel pins to HED attacks. Because of their refractory nature the radionuclides 154-Eu and 144 Ce were used as a tracer for uranium in the spent fuel pellets; amounts of these two fission products were determined by gamma spectrascopy.
For the case of depleted uranium dioxide the total mass of aerosolized materials collected on the filter was 17 mgll (gravimetrically determined). Assuming that the total mass consists of combustion products, fuel pin cladding and pellet material, an estimate of the relative masses represented by the three main components was made.
It was estimated that 25 percent of the total collected aerosol mass was combustion products. The pellets consisted of 89 percent by weight of UO2 for experiments with either spent fuel or depicted uranium dioxide. Therefore a 17 mg aerosol sample implied that 11.35 mg was uranium dioxide. The measured mass of aerosolized spent fuel was found to be 5.95 t 1.67 mg.
This analysis suggests the correlation ratio is approximately equal to 0.53:1.
A second series of measurements used for the calculating the spent fuel versus depleted uranium dioxide correlation was based on sieving of debris resultlag from identical experiments involving spent fuel and depleted uranium dioxide. Details are included in Appendix D in this report.
Steves used for classiflying particles ranging in size from 5 pm diameter to 212 pm diameter were used to separate masses of spent fuel and depleted uranium dioxide. The mass remaining on each sieve were then used to calculate a ratio of spent fuel to depleted uranium dioxide for each sieve size fraction. A regression line was fitted to the data and a ratic axtrapolated for the respirable size range
(<3 um actual particle diameter). For particles 3 pm and smaller, the ratio of spent fuel to depleted uranium dioxide mass was 5.6:1.
Subscale_Te_s.ts Performed at Battelle Columbus _La_boratories Under the aegis of the U.S. Nuclear Regulatory Commission (NRC), a series of subscale tests on spent fuel was performed by personnel at Battelle Columbus Laboratories (BCL).4
o
. These tests used a short five-pin array of H. B. Robinson Unit-2 spent fuel which was subjected to attack by a subscale precision version of an HED device. Results of selected early time aerosol measurements by filters are summarized in Tabic 4.4.1.
These data indicate that the uranium aerosol concentration from the spent fuel ranged from 50 to 500 g per liter for the time period 0 3 minutes post-attack. The mean of these measurements was 118.2 pg per liter with a standard deviation of t130 pg per liter.
If we eliminate the extremes of the measured aerosol concentration the resultant mean and standard deviation are 70.4 1 31.4 gg per liter, respectively.
Eliminating the two highest measurements is justified for the following reasons:
1.
During " Hot Shot 8" one filter cass tte measured a concentration of 447 and All pg per liter of uranium dioxide while another filter cassette operated simultaneously with the 447-411 measurements indicated a UO2 concentration of 43 to 130 pg per liter.
2.
A statistical test (" standard box plot"), which assumes a Caussian T-distelbution, was applied to the data to evaluate the validity of the -
outlyers. This statistical test indicated that the values of 447 and All pg/l have less than 2 chances out of 1,000,000 probability of belonging in the assumed Caussian distribution.
Based on scaling theories developed at BCL for extrapolating to the 3 fuel assem'ly PWR truck cask incident, an aerosolized uranium mass of 4.24 1 4.66 g b
was calculated for the case using all of the filter data from BCL.
If the aerosolized mass is calculated based on filter data where the extremos of the measurements are not used, then 2.53 1 1.12 g of aerosolized uranium would be released as a result of the reference base attack on the 3 PWR fuel assembly l
truck cask.
The data from the subscale and full-scale experiments performed in this study suggest a release of 6 grams from a reference base attack on a 3 PWR fuel assembly truck cask.
From these mean values a correlation ratio (spent fuel:D-UO ) of 0.71:1 and 0.42:1 is calculated for all filter data 2
and for the truncated versions, respectively.
BCL, in the analysis of their data, utilized the maximum measured release instead of a mean value. This led to a correlation ratio 5 times the largest value quoted above or about 3.0.
An.a_1 y s i s There are potential errors implicit in all of these correlation ratios.
While estimates for the BCL results are based on actual spent fuel acrosolized mass, there were no comparable results for depicted uranium dioxide in the same test set-up.
Given this uncertainty the most definitive statement that can be made is that the value is likely to be on the order of unity.
l Calculations based on EG&C/INEL data for filter measurements were suspect because the mass of depleted uranium dioxide was not determined analytically but was calculated from the total mass collected by sempling filters. Unknown I
amounts of test fixture support materials could lead to an overestimate of
)
\\
1 l Table 4.4.1 Filter Data From Battelle Columbus Laboratories Studies 4 on Spent Fuel Release Fraction Time Concentration Experiment Filter No.
Interval (Min) pjujL._ _ _
r hst shot
- 1.11 3
1.0-2.0 93.8 1.12 2
0.5-1.0 60.2 1.12 3
1.0-2.0 79.8 T-Hot shott 2.11 3
1.0-2.0 50.6 2.12 2
0.5-1.0 95.5 Hot shot
- 7.11 1
0
-0.5 54.5 2
0.5-1.0 25.4 3
1.0-2.0 27.8 7.12 1
0
-0.5 67.7 3
- 1. 0- 2.0 84.9 Hot shot
- 8.11 1
0
-0.5 447 2
0.5-1.0 411 8.12 1
0
-0.5 107.7 2
0.5-1.0 129.9 3-1.0-2.0 43.1
= --
- Spent Fuel Experiment
. l l
acrosolized depleted uranium dioxide. Since it is unlikely that the estimates of mass allocation are an order of magnitude off, the available data suggests the correlation is of order unity.
The correlation based on wet sieving of debris from both spent fuel and depleted uranium dioxide was based on extrapolating sieve results down to the respicabic regime.
In order for this process to be valid, a single mode particle size distribution was assumed to extend into the respirable regime.
Since the ultrafine aerosol size distributions in the scaled and full scale tests were determined to be multimodal as were larger diameter sieve size distributions, the assumptions of a single mode representing both wet sieve fractions and respirabic size fractions may not be valid. The value of 5.6 obtained for the ratio was greatly impacted by the 1 - 20 pm sieve fraction which exhibited a ratio of 125:1, spent fuel to depleted uranium dioxide.
This extremely high ratio has a significant effect on the calculated regression line that was fitted to the data.
Since.this size fraction is very similar to the original grain size of the uranium dioxide prior to pellet fabrication it may represent an anomolus peak resulting from a breakdown in the sintered matrix of the fuel produced by grain swelling in the reactor environment.
If this hypothesis were true, then the ratio of spent fuel to depicted uranium respirabic acrosol would be of order unity and not 5.6.
Summart of correlation Data The following statements summarize the calculation of a correlation function:
The NRC/BCL studies yielded a maximum UOg release value of 500 g per liter which resulted in a spent fuel to depleted uranium dioxide correlation ratio of 3.
Using the original data from the BCL studies and using standard statistical tests an estimate of 0.42 was obtained.
Using filter data from EG&G/INEL that was collected in support of this DOE program, a value of 0.53 was obtained.
Based on wet sieve data from EC&G/INEL on material released as a result of subscale attack on both spent-fuel and depleted uranium dioxide, a correlation ratio of 5.6 was obtained.
In considering which correlation ratio is appropriate for use in risk analysis and calculation of radiological impacts, it would seem that a value of unity is most appropelate. This implies that the acrosolized release from i
the reference HED used on a 3 PWR fuel %ssembly truck cask would yield l
approximately 6 g of spent fuel as a respirable aerosol.
However, for conservatism in the health risk assessment, a maximum value of 5.6 will be
l l
l
O D
a
, 4.5 Analyses of Fuel and Cask Breakup and Aerosol Production In order to develop an understanding of mechanisms of aerosol production, and the mechanisms causing breakup of fuel and cask, elemental analyses of aerosol samples, using x-ray fluorescence, standard metallurgical analytical techniques, and aerosol dynamics modeling were performed.
4.5.1 Elemental Analyses and Quantification of Aerosol Samples by Wavelength of Dispersed X-Ray Fluorescence (XRF) Spectroscopy.
Aerosol samples collected on filters during a 1/4-scale wet and a full scale dry test were analyzed for U, Zr, Pb, and Fe by the use of a new x-ray fluorescence (XRF) thin film technique. These analyses were made to characterize the composition and time-dependent behavior of the aerosol produced by a reference HED attack on a fully loaded spent fuel shipping cask. I-Figures 4.5.1 and 4.5.7 display the time-dependent behavior of the four acrosol components in the 1/4-scale wet Test No.
7.
The corresponding results for the full-scale dry test are presented in Figures 4.5.3 and 4.5.4.
The UO2 acrosol concentration decreased in both tests from roughly 5 to 0.1 pg/g over a period of 150 min after detonation.
The results for UO 2 concentration correlated well with those obtained by U fluorometry.
The 2r02 aerosol concentration varied in roughly the same concentration range as that of UO2 whereas those for Pbo and Fe0 were roughly one to two orders of magnitude greater with the Pbo > Feo in the full-scale test and FeO > Pbo in l
Test 7 (wet test).
i l
The essential behavior of the full-scale dry test aerosol is illustrated in Figures 4.5.5 and 4.5.6 where the aerosol fraction of the components are 1
presented as a function of time.
The aerosol composition at any particular time was defined by the ratios of the various oxide component concentrations to the total aerosol concentration. The aerosol fractions of lead oxide, iron oxide, and zirconium oxide in the full-scale test were found to be 0.64, 0.20, and 0.003, respectively, independent of time. The aerosol fraction of uranium oxide,
~
however, was found to decrease exponentially with time from about 0.005 initially to 0.0003 after 180 min. The aerosol fraction (AF) by (XRF) for UO2, was found by multiple regression to vary as AF
= 0.0040 exp [.012t]
UO 2
where t is in minutes so that initially (t = 0) it is 0.4 percent of the total acrosol.
. 6.0 4.0 e UO2 - XRF E UO2 - U FLUORESCENCE A ZrO2 - XRF
- 8*
A 1
32.0 z
9 s-zw y1.0
^
O O 0.8 UOg z
O I
ZrO2 O
j 0.4 O
v2 w<
0.2 -
Zrog 0.1 -
0.0 I
'\\
O 40 80 120 160 200 240 280 MEAN TIME AFTER DETONATION (min)
(
Figu re 4. 5.}.
UO2 and ZrO2 Aerosol Concentrations as a Function of Time for Wet Test No. 7.
Comparison of XRP and uranium g
fluorescence analyses.
\\
\\
. i i
i i
i i
100.0 4
80.0 60.0 A FeO - XRF
~
e Pbo - XRF 40.0 3
]g 29 7 20.0 m>
e e zwo 8
o g 10.0 i
g o 8.0 g
8 6.0 4
.J O
en Oz 4.0 w
l 4
FeO _
e PbO 2.0 O
40 80 120 160 200 240 280 MEAN TIME AFTER DETONATION (min) 3
?>
Figure 4. 5.2.
FeO and PbO Aerosol Concentrations as a Function of Time Af ter Detonation for Wet Test No.
7.
Analyses are l
h XU.
-5 9-i 30.0 3
i i
i i
6 E
e U FLUORESCENCE A XRF - UO2 3 XRF - ZrO2
- e 6.0 C
l a
~ 3.0 z
9-
<Ewz e
o y 1.0 g
0 A
ZrO2 a
8 0.6 OE y
A a
A 0.3 UO g
0.1 0.06 O
30 60 90 120 150 180 210 MEAN TIME AFTER DETONATION (min)
Figure 4. 5. 3.
UO2 and ZrO2 Aerosol Concentration as a Function cf Time for Full-Scale Dry Test.
Comparison of XRF with uranium fluorescence analysis technique.
. 1000 g
i i
g 3
3 e
e 600 e Pbo - XRF
- E FeO - XRF 300 E
9
~
r 3
~
z9 E
Pbo a:
6 100 g
O o
.J
@ 60 E
O g
y FeO 30 E
i I
I I
I I
I 10 O
30 60 90 120 150 180 210 MEAN TIME AFTER DETONATION (min)
Figure 4. 5.4.
Pbo and FeO Aerosol Concentrations as a Function of Time Af ter Detonation for Full-Scale Dry Test i
~ _.
d
+
i '
I i
1 1
1 0.004 e
T 0.003 ZrO2 Q
e O<m u.
f 0.002 8
=
w<
e e
0.001 o
e UO 2 l
0.000 I
I I
I O
40 80 120 160 200 MEAN TIME AFTER DETONATION (min) i I
l l
l l
Figure 4.5.5.
Aerosol Fraction for ZrO2 and UO2 as a Function i
of Time Af ter Detonation for the Full-Scale Dry Test (No. 8)
4 l
i i
i i
4.0
~
e e
e 3.0 e
e _ 103 ZrO2
-p z9 e
G E
O 2.0
~
o b<
~
e-PbO + FeO 04 O
40 80 120 160 200 MEAN TNE AFTER DETONATION (min) l Figure 4.5.6.
Aerosol Fraction of Zr0, and Pb0 plus Fe0 Components as a 2
Function of Time for Full-Scale Test
e '
These results for the dry full-scale test describe an aerosol with j.
essentially two components in terms of timo dependent behavior. One component is formed of Zr0, and the other is Pb0 plus Feo which all fall out at the 2
l same rate as the total aerosol. This unique result suggests condensation and/or coagulat ion of these components af ter formation by vaporization and/or melting of metal components (Pb and Fe) in the shipping cask. The UO2 aerosol, however.. falls out exponentially more rapidly than the total I
aerosol.
It appears to be a completely different component of larger particles (than the lead and iron oxides) and could have been formed by fragmentation only of the UO2 fuel.
Comparabic results for the acrosol fraction in the 1/4 ccale wet test l
(Test No. 7) could not be obtained because of moisture in the weighed filters.
4.5.2 Metallurgical Analyses Metallurgical analyses of the materials recovered from the full-scale test were performed to determine the maximum temperature, Tmax, attained in the l
region where the blast front, generated by.the HED, impinges on the fuel
]I I
rods. A knowledge of the temperature of this event is important, since it l
would be useful in estimating the amount and species of fission product j
radionuclide contaminated aerosols that would be released during an attack.
Thermocouples were not a viable way to determine the temperature.
Thermocouples have too large a thermal mass, and therefore would respond too slowly.
Alternatively, it is possible to make some reasonabic estimate of temperature by examination of the phase transformations (microstructures)
Induced during the explosive event.
Several components from the full-scale cask test have been examined with a variety of metallurgical techniques, including optical metallography, scanning I
clectron microscopy (SEM), and x ray diffraction.
Other electron optical techniques, including transmission electron microscopy (TEM) and electron microprobe analysis were also used. However, in this case, these more compicx techniques yielded no additional information eclative to the SEM.
Since the objective of this analysis was to determine the marinum temperature attained at the fuel rods during the test, several samples were used.
Samples analyzed included (1) the explosive actuaied valve door, (2) millipore filters, (3) aluminum SEM stubs, (4) Cu discs which collected (a) explosively driven particles only and (b) explosively driven plus diffusion l.
particles, and (5) a disc sampler-similar to (4) but containing 3-mm TEM grids..These samples also included (6) a stainless steel clug from the cask, (7) zirealoy cladding, and (8) a UO2 fuel pellet, all from the site of blast impingement.
Figures 4.5.7 through 4.5.14 show the results of this analysis.
Each figure is fully described by its caption. Briefly, the SEN analysis of the samples showed evidence, as shown in Figures 4.5.7, and 4.5.8 of vaporized Icad and molten stainless steel.
I l
lF
l i
l I
i l
i 1
I t
t
]
I l
i r
l 4
9 l
7 I
y l
k i
s l
l i
~
I i
A
'g 1
l
.00fjk
- E itu w
I Figure 4. 5. 7.
SEM Micrograph of Material Deposited on the Valve Door During the Full-Scale Cask Test. Spherical particles of droplets of stainless steel, indicating the stain-less steel f rom the cask was melted. The wispy material is lead, probably condensed from a vapor.
-..~.-.--
i
i l
1 i
i 1
l 1 i
)
l 1
4 I
i 1
,i I
l i
l j
I 6
i t
I l
i-I t
i l
l i
i t
l t
l l
t
. c 4 * '
i V
J W
s cf e
?
e e'
f Figure 4. 5. 8.
X-Ray Spectra Showing the Elements Present in Figure 4.5.7.
Fe, Cr, and Ni are f rom the stainless steel, Pb is simply Pb, and the Si is probably due to the presence of a small amount of sand. The Pb and stainless steel are probably f rom the cask.
i l
l
. i sY l
i I
l l
i m
,t
,f &
- M.Ed._ i, J
. 3v y s:
c
- r;,
.v.
7
..4-e r4 i
1 I
I i
i J
1 1
i l
)
i Figure 4.5.9.
Macrophotograph (1.5x) of the Stainless Steel Slug Taken From the Cask Af ter the Full-Scale Test Event.
{
l
}
--- -,. -,-- ~
-r
,-r----
,. -~--
I l l l
l i
l i
I i
w; I
~, n. <. - - v ' f.. so-r.
e ' e..
.s y
M.
].3 " ' . * *as
, p.,.,3q,;,,
'. T f.'. +)
M l
a.
g 4,
- A
<.,..f,-
~
ry.
. 74 ~ M',, ?,, ** -
m..
h
.'6-W
- /
../F
,.4 g
e.
...... ",.,. -....;,=
s-
-g... ee o. r,M -
.a :.
a
=..,
4 8
p' U;#1 ' '.
.
- AL4 ~ '
- - t
,6 >., f
.s Mh ~
.x..
. 9-s.., 2 y
n-M Figure 4. 5.10.
Micrograph of the Stainless Steel Slug in Cross Section (500x). Sigt,s of surf ace melting are present. Exterior of slug is co'sted with Pb.
-6 8-
-u
.,x,
e,...,..-
v. w -...
f,'. 'j?;r giy 'Ty'er *Chf,", _
%ga-Q@h.o;?qp.
a uc Eu Ns.6.1?
3 rn 9.e.
r.*5 93f.. 7 :c S "A.e:
c
- ~wts :
w.n,4* :> n,.:h.'.,n
.,.e y
- n u n.,i x
~
..3 s.
a+
. -ff "-*;M Ed'.,IYD prK4 b.T E K 6.gN i
tt
Qy.:
,... f %g ~ ".yp ? '
jyy:
y t
ei w.,,;n Tf.y Qy*;'*?
t Qm, as,+4
,, n y
'k%kQ'.E n.;*;
' blhh
- Wl2h
+u.
9 e
?#v%
g/h-d?M.w3,~,n o$W-e~,,
1;l.fy v.. % f',C2.j c.p>;,.;.N.,,,?<
Q' 9 ;{.. r,,;j,u:: -i re '.. ',,.,'n,'q'
.,.. w r
- r,,
.b s
+
s
- ',.4
, 1.s.
. ; ;,,,, u,:,9, 3. 7;;}f
. ~,
,c.--
- *.' ;% % ':.;<.,'eM *t=5..
..,., '. '. ' '. :. * * '^
-s
- N*)Qr.? "6
. z L::2 ? 2d.L.. R
.: =,
3.a.
_i
['
~~ i
+ ?xl 0;N f,h.
[_-
' '*}
'J l
s
'.. ~.
1 n.
.1 l
l l
Figure 4.5.11.
Macrophotograph of a Piece of Zircaloy Cladding Taken After Full-Scale Test Event (1.5x) 1 m.
i P
I l
-6 9-1 i
i l
,L f
I I
i i
I J
t
- r
, e 4.
-r i
1 y;.
,.., a
.,;,,,. 4
- 7f..;
i
. d'
- ~.L.
~,...%' '
- 1:....
s
~: (,;--.
7.v,..
0*
n i
~ r..y t
l
.. - 4 w_,q;,3 n.p i
(
- s;..:g??m,,.-s4 i
. +
". -, v. o2...,
l I
..,,. ~,
~ !a,. rl',fj
.,,..s,c,
'- y,_ :)l.,.u,9
~..n
,.. - - ~,. -
x,,'.s,~ _ 'c f
^ o.,-.D..
,,. f. :,t.r
.4
'... f.. Y
.3
-i
- c. m..
- 4 L
i i
i l-Figure 4. 5.12.
Micrograph of Zircaloy Cladding in Cross Section (100x).
No signs of raelting or ablation.
. I
~b a.
0
's f
"b l*
~
S',$ #* \\ %
y'.?
/
Tf
~
k, M
e
~
g..,-.
WW F
f l
~
j' S' a 4 s
',,1-A y.
4
+'.
,j
&gU.. 9k,,-l' A.-
i s,
~
fl W
__9 t
[
l P
Figure 4.5.13.
Microscopic Cross Section of UO2 Fuel Pellet (10 0x).
Basic microstructure has been unaltered by the explosive l
event.
l l
I
-71 The OIns of stainless steel, probably from the cask itself, is shown macroscopically in Figure 4.5.9.
Meta 11ographic examination of the slug in cross section shows signs of melting and ablation, as shown in Figure 4.5.10.
The sample is also coated with lead, which was probably deposited after condensation from a vapor state. This evidence points to a temperature in excess of the molting point of stainless steel, 1400*C, and the vaporization point of Icad, 1744*C.
The sample of zircaloy cladding is shown macroscopically in Figure 4.5.11 and in microscopic cross section in Figure 4.5.12.
Although the cladding material is coated with vapor deposited Icad, there are no cigns of melting or ablation of the zircaloy cladding itself.
Hence it may be assumed that the melting Lem stature of zircaloy, 1850*C, was not attained.
It has been e
suggected that since it is highly pyrophoric, it may have oxidized during the event and all signs of melting removed.
X-ray diffraction studies of fines recovered from the tect showed the pecuence of Zr02., However, it is contended that physical signs of molting would have been retained in the material despite oxidation.
Such evidence could include subsurface hot 3
plastic flow or pockets of trapped oxide (Zr0 ).
2 The UO2 fuct pellet was examined by x ray diffraction and optical metallography. The x-ray dif fractometer data showed only U02; nc other forms of uranium oxide were present. There was no evidence of the formation of a glassy phase. No signs of melting or ablation were apparent in the metallographic samples, as shown in Figure 4.5.13.
Depending on the reference, the melting point of UO2 is listed to be between 2500*C.and 3000*C.
It has been suggested that it may be possible to infer the thermal history of the fuel pellet accurately by detecting the presence of uranium oxides other than UO2 Such inferences would be tentative and of dubious value.
Oxidation reactions are dependent on many parameters other than temperature including (1) total pressure, (2) identity of all gaseous species present (CO2, CO, HxCy, etc), and (3) the partial preocures of each species listed in (2).
It is not possible to accurately define these parameters.
In addition, all cimple thermodynamic calculations intended to predict phase stability apply only to the state of equilibrium. The assumptions of equil.
ibrium between the cask solid materials and its components and atmospheric and blast gasec are probably not valid.
In summary, the evidence derived from examination of these components from the full-ccale test indicate the maximum temperature attained was most likely between the vaporization point of lead, 1744*C, and the melting point of zirealoy, 1850*C.
No particles of UO2 indicating melting or vaporization were found.
All UO2 Particles examined indicated particles produced by mechanical fracturing mechanisms.
From these results, it is predicted that a full.ccale attack on spent fuel would result in similar conditions of temperature and acrosol formation.
Consequently, from these temperature ranges inferred, certain volatile fission product radionuclides such as
i i
137Cs, 106Ru, and 125Sb could be vaporized and could be expected to have an enhanced release as compared to the fuel matrix.
Furthermore, it is i
expected that these vaporized fission product radionuclides would then condense onto available surfaces and would be associated with smaller (respirable) particle sizes. None of the more biologically significant radionuclideo cuch as the actinides (239Pu, 238Pu, 242-244Cm and 241Am) or beta gamma emitters such as 90 144 37, or Ce would be vaporized and thereby result in an enhanced release compared to the fuel mass.
Although l
certain volatile fission-product radionuclides (137Cs and 106Ru) would exhibit an enhanced release, all of these radionuclides are less biologically significant than the actinides and the resultant calculated dose increase does j
not affect the overall risk estimate.
4.5.3 Modeling the Dynamics of Aerosol Production l
A mathematical model of aerosol dynamics is essential for understanding the processes which govern the particle size distribution produced from an explosive environment. The model serves not only to elucidate the dynamics of,
meacured particle size distributions but also to predict the effects of condi-E-i tions not encountered in the course of the experimental program. Since it i
wocid be impractical and cost prohibitive to arrange an exhaustive set of i
experiments to encompass all possible scenarios, a model has great utility in i
i wupplementing existing data in regions not explored experimentally. Thus the major objective of the model presented is to elucidate and quantify the mechanisms contributing to the dynamics of the particle size distribution of a confined aerosol generated by an explosion.
I Since quantifying the confinement effects are crucial to extrapolating the data to unconfined conditions, it is important to determine the mechanisms for aerosol removal in the chamber.
If surface diffusive der. ItIon ic import. ant.
then it should be noted that this mechanism will not exist in an open atmo-sphere. However, gravitational settling would be present in confined and unconfined conditions. Thus, it is important for the model to determine which process resulted in aerosol removal in the confined experiments.
Figure 4.5.14 shows the aerosol mass concentration from each cascade impactor stage for sample times taken at 5-30 sec., 1-1.5 min and 13-13.5 min. postdetonation for the full-scale reference test (No. 8).
Note that the mass concentration over an aerodynamic particle diameter range is drawn as a horizontal line.
Thus for particles in the diameter range 0.15-1.73 km, approximately 2.0, 0.5, and 0.085 g/m3 were measured at the three compling times, respectively. From Figure 4.5.14 we note that more than half of the aerosol mass was removed within about the first minute of sampling.
l In this work, three models are considered to elucidate the observed initial rapid aerosol mass loss.
It is shown that diffusive wall deposition and gravitational settling alone, do not explain the data.
Finally a coupled coagulation settling model is shown to provide some insight to the data.
From these calculations, a lower limit estimate of the median aerodynamic diameter is made for confined aerosol within the first minute af ter the explosion.
. 1 l
l 2
TIME INTERVAL (MIN SEC)
~
0:15 - 0:30 E 1.6 E
--- 1:00 - 1:30
+
3
......... 13:03 - 13:30 zo 1.2 P4 Zw 0.8 o
r7 8
i o
J i
I L,
m 0.4 m
I 4
I "s_
2
....L- - _
_ _ _ _ _ _ _ _ _ _ _ _ 3.. ;
g 10-2 jo-1 100 101 102 l
PARTICLE DIAMETER (pm)
Figure 4.5.14 Plot of Measured Aerosol Mass Concentrations Obtained From l
Sequenced Cascade Impactors for Full-Scale Reference Test.
l The impactor measurements were taken at 15 to 30 seconds,1 to 1.5 minutes, and 13 to 13.5 minutes postdetonation.
" 4.5.3.1 Diffusive Wall Deposition For calculating aerosol dif fusive deposition losses, the boundary 1ryer thickness must be specified. Since this parameter is not easily m:asured, it may be calculated form the measured aerosol mass loss. This calculated thickness can then be compared to other experimentally determined values to determine if the calculated boundary layer thickness is reasonable.
A physically realistic calculated boundary layer thickness would support diffusive deposition as a mechanism for aerosol mass loss.
Under quiescent conditions, Harrison (1979)12 has experimentally dstermined tnat the boundary layer thickness a, may be approximated by
~
a = 12(D/2)
(4.5.3.1) where D, tne particle diameter and a are in microns.
For comparison, the esrosol diffusive deposition loss may be modelled as f f = f fh) Q (4.5.3.2) where t is time, Q is the aerosol mass concentration, D is the particle diffusivity and A/V is the surface area-to-volume ratio of the chamber.
Integrating Eq (4. 5.J.2), we obtain a = Dt(A/V)ln(Q /Q)
(4.5.3.3) where Qo is the initial mass concentration. The chamber volume is 42.6 m3 and f or a particle diameter of 1.0 gm, D = 2.77 x 10-7 2
cm /s (Friedlander (1977) p. 33).13 The surface area for deposition is given in Table 4.5.'3.
TABLE 4.5.3 Surface Area for Aerosol Deposition Pa rt Area Chamber Walls 74.4 m Cask Body 12,7 m Cask Fins 32.7 m Total 120.0 m 3
By substituting Qo = 6.442 g/m, Q = 2.712 g/m3 and t = 52.5 s into Eq. (4.5.3.3), we obtain a boundary layer thickness of 3.6 x 10-3 gm.
s 1,
I However, from Eq. (4.5.3.2), c = 39 gm.
Note that the two results differ by 4 orders of magnitude. Even if D and ( A/V) are each inaccurate by an order of i
nagnitade, this would still not explain the extremely small boundary layer thickness needed to explain the observed rapid aerosol mass loss. Further-mo re, it is doubtful that the turbulence remaining 15-30 seconds af ter the blagt would reduce the boundary layer thickness by 4 orders of magnitude to 35 A.
Thus the data do not support dif fusive deposition as a significant mechanism for aerosol removal.
4.5.3.2 Gravitational Settling 4
over the initial time me..surement intervals, the largest aerosol mass loss was for particles' in the diameter range 0.7 5-1. 23 um.
For a particle, 1.0 am in diameter, the settling velocity is 3.52 x 10-3 cm/s (Friedlander i
- p. 33).13 Thus after 52.5 seconds, a spherical particle would have fallen only 0.185 cm.
Clearly, since this distance is much smaller than the chamber dimensions, the data do not support gravitational settling alone as a signif-icant mechanism for aerosol mass removal. Note that if the nonspherical shape 3* the particle is accounted for, the settling velocity would be even smaller than that given above, 4.5.3.3 Coupled Coagulation and Settling t
A.
Preliminary Assessment Although coa 7ulation alone does not remove aarosol mass and settling
'us been previously shown not to be a significant removal mechanism, the combined effects of particle groath by coagulstion and gravitational settling any explain the data.
One can quickly assess the feasibility of this process bf 3Actsfying two criteria.
First, the time scale for coagulation must be less than 52 seconds to enable particles to grow over the time scale of inte re st.
.Second, the particle size needed to attain a settling velocity large enough to account for the data should be physically realistic. We can evaluate both criteria as follous.
The coagulation coefficient p, for particle sizes of interest is in cm /s (Friedlander p. 193).13 The character-3 the range 10~9 to 10-8 istic time scale for coagulation is approximately 7 = 1/pN, where N is the serosol number concentration. To estimate N, consider the particles in the diameter range 0. 7 5-1. 2 3 um, to be spherical and 1.0 gm in diameter, with a 3
density or 1.0 g/cm. Thua since the initial mass concentration is 3
1.986 g/m, the number concentration is approximately 6Q-S
-3 N=
= 3.8 x 10 cm (4.5.3.4) 3 rpD and l
- 7. = 1/pN = 2.6-26 sec.
(4.5.3.5) r
./
. m since 7 <52 seconds, coagulation is significant over the first sampling period and the first criterion is satisfied.
?or gravitational settling loss, Q may be modelled by f=-p(
Q (4.5.3.6) where v is the settling velocity. Since the particles are much larger than the moon free path of air, Stokes law may be used to obtain 2
y= MEE-(4.5.3.1) 18 4 where p is the particle density, g is the gravitational constant and 4 is the viscosity of air.
The material densit is approximately 8 g/cm3 and based ontheworkofClesekeandReed(1977){4 one would expect an effccLive 3
particle density of at least 1.6 g/cm.
Substituting Eq. (4.5.3.1) into Eq. (4.5.3.6) and integrating, results in I
18 4 f)In(Q/Q)
D (4.5.3.8)
=
p The projected horizontal surface area is approximately 18.9 m2 and thus using a particle material density range of 1.6-8.0 g/cm3, the particle geometric diameter needed to explain the rapid mass loss is 12-28 pm which is equivalent to an aerodynamic diameter of 35 pm.
However, since Q hardly varied over the sampling time of 52.5 to 105 sec, a factor of 2 in time would reduce the aerodynamic diameter from 35 to 25 pm.
Since 25-35 pm is a i
reasonable particle size range, the second criterion is satisfied.
B.
Coagulation _ Settling Model Description The model used in this work solves the baste differential equation governing coagulation and settling as discussed by Celbard and Seinfeld (1980).15 The basic assumptions are:
I l
1.
The particles are uniformly distributed in space, which
}
is supported by the relatively constant mass density l
measured at several sampling ports in the full.ceale test.
t.
l 2.
Coagulation is due to Brownian motion. turbulence and i
differential gravitational settling.
l 3.
Particles are removed from the atmosphere by settling.
4.
Particles can be uniquely characterized by their mass.
The fourth assumption is best suited for spherical liquid droplets.
j However, due to coagulation of solid particles in these,cuperiments, we are j
usually dealing with irregular shaped agglomerates. Unfortunately the theory i
l
g
.. for leregular shaped part icles is not well developed and one generally resorts to introducing correction factors to the equations governing spherical particle dynamics (Cleseke et al, 1918).14 By using the above assumptions one can develop a set of mass balances for the acrosol mass within a set of particle size ranges. The net rate of aerosol mass accumulating in each size range is equal to the aerosol mass forming in that size range by coagulation minus that leaving the size range by coagulation and settling. The expressions for the rates of coagulation and settling are rigorously derived for spherical particles and corrected for nonspherical shapes.
C.
Results Figure 4.5.15 chows the total suspended acrosol mass concentration as a function of time.
Note that the data, as shown by the solid line, indicates a very rapid inillal mass loss. To account for the nonspherical nature of the particles. the agglomeration shape factor y, was varied in the calculations t For ideal spherical particles y i 1, and a larger value of y enhances the
~
coagulation process. Unfortunately, y is not a readily determined parameter and is usually adjusted to fit the data.
Due to the formation of void space in the fluffy agglomerates, the effective density should be less than the 3
material density of 8.0 g/cm. Since the work of Gieseke and Reed (19/1)16 indicate a reduction by at most a factor of 5, the effective 3
density was varied from 1.6 to 8.0 g/cm,
In the absence of coagulation, it was shown that gravitational settling could not explain the rapid drop in aerosol mass concentration.
From Figure 4.5.15 it is noted that by increasing y, which enhances coagulation, thus more quickly forming large particles, rapid drop in aerosol mass concen-tration with time can be obtained.
Furthermore, it is seen from Figure 4.5.15 that a value of y = 10 can result in a rapid aerosol loss by coagulation and settling which could not be achieved by any of the previously discussed models. Although it appears that the initial processes may be modelled by coagulation and settling. Figure 4.5.16 shows that the calculated aerosol distribution does not agree with the measured distribution after 52.5 seconds of sampilng. The calculations predict more aerosol in larger particle sizes than was actually mea :ured.
The calculated results are not surprising since as previously shown, an aerodynamic particle diameter of 25 35 pm is needed to justify settling as the removal mechanism.
Furthermore, note from Figure 4.5.16, if the dynamic shape f actor X is 2, thus halving the settling velocity, even more mass remains aleborne for larger particle sizes.
(A value of X = 1, which was used in Figure 4.5.15 neglects reduced settling velocities due to nonspherical particles.) Thus for coagulation and settling to explain the rapid aerosol mass loss larger particles should have been present than were mescured from the cascade impactors.
Even if one only considers settling alone, particle diameters in the chamber would have to be 25-35 g m, which is much larger than reported by the cascado impactors. Thus one is lead to the I
hypothesis that the impactor data are artificially skewing the data toward I
smaller particle sizes. This type of err or is well known and can of ten be L
. 1 8
i i
57 E
s
,,O 6
f.!
g N
y g
g
-e 7=1.0 9 = 1.6 g/cc
\\
N5
\\
\\
s l-
~~
~ -
- z
\\
\\
s W
N o
s s
s z
s s
\\
u 4 g
s o
N s
s i
\\
s e
N e
g s
$3 g
\\
s s
\\
\\
s N
N g=5 p = 8.0 g/cc N
7 y
k s
o O 2 N
\\
N
~%
g W
N N
4 N
^
"f = 5 p-1.6 g/cc 7
s 1
s
\\'
4 -
DATA s
w
~~'
7= 10 p = 8.0 g/cc g~~"-------- wy= 10 p = 1.6 g/cc 0
t i
e i
O 200 400 600 700 800 SAMPLING TIME (seconds) l Figure 4.5.15.
Plot of Total Suspended Aerosol Mass Concentration as a Function of Time.
Solid line indicates neasured total aerosol mass based upon sequential filters. Dotted lines are for different agglomeration shape factors, y.
a
.. 1 EXPERIMENTAL DATA
........... y - 10, p = 1.6, X = 1.0
^n
y = 10, p = 1.6, X = 2.0 E O.8 2
8 z
r-i 9
_ _r, l L,
y 0.6 li i
" * ! L _l L,
g l
i::
L, z
W i
i S 0.4 i
8 o
-.I 7
i I
g i
h....; (7 m
4 0.2
- ..**: I lE
- I
- ..J n
i O
I
!I I
10-2 jo-1 joO 10 10 1
2 PARTICLE DIAMETER (pm)
Figure 4. 5.16.
Plot Showing Comparison of Calculated Mass Concentration From Model With Measured Values After 52.5 Seconds Postdetonation
d e attributed to particle bounce (Cheng and Yeh, 1979)l7 and to agglomerate breakage when passing through the impactor. Also, the extractive sampling line geometry (multiple bends in the sample lines) tends to provide an inherent maximum particle sampling diameter which precludes larger particles
(>10 pm) from reaching the sampling instruments (C. J. Newton, private communication, 1983).
Note that these errors should not affect the total mass concentration measurements which do agree with the filter measurements.
4.5.3.4 conclusions Three models were used to explain the observed rapid aerosol mass loss.
First, diffusive deposition would require an unrealistically small diffusion boundary layer thickness to support this mechanism as an explana-tion Second, gravitational settling alone of the measured particle sizes was too slow to explain the rapid mass loss.
Finally, a combined coagulation-settling model, in which particles grew by coagulation and settled out could explain the rapid aerosol mass loss.
This model, however, indicates larger particles should have formed than those reported from the cascade impactors.
Since an aerodynamic particle diameter of 25-35 pm is needed to obtain a settling velocity which would explain the data, it is hypothesized that the impactor size distribution data is in error due to overloading of the impactor stages and that the actural aerodynamic diameters were larger than measured.
This hypothesis is supported by the observed rapid loss in total aerosol mass concentration which is reported from both the cascade impactors and the total filters.
Using the calculated results shown in Figure 4.5.16, the median aerodynamic diameter was larger than 2.0 pm.
Since neglecting coagulation would indicate that particles formed at 15-30 seconds postdetonation were 25-35 pm in diameter, 2.0 m is considered to be a lower limit of the median aerodynamic diameter at about 1 minute postdetonation. Note that this airborne particle size underestimation does not affect the uranium mass concentration measurements upon which the health consequence estimates in this study are based. The fact that coagulation and gravitational settling plays a dominant role in the rapid loss in total aerosol mass concentration in these confined volume tests becomes a secondary mechanism in an atmospheric free-volume explosive event in which the important consideration is the early time (t = 0) primary particle concentration upon which this study's release j
fractions and health consequence estimates are based.
4.6 Analysis of Radiological Health Effects l
The health consequences due to radioactive release from spent fuel casks subjected to sabotage were estimated using the consequence reactor safety 5
model CRAC employed in the Urban Study.1'3 Although CRAC was not developed specifically for transportation accident environments, CRAC may be used to provide useful consequence estimates if interpreted with appropriate regard for the modeling assumptions. The Urban Study 1,3 has reported that CRAC estimates are generally in good agreement with those of other calculational techniques. However, CRAC was selected for this study primarily because the NHC interim regulations regarding civillaa reactor spent fuel transportation in the US are based upon the CRAC estimates as reported in the Urban Study.1 I
L
}
O e
81-The consequence model uses calculated airborne and ground radinnuclide concentrations to estimate the public's exposure to (1) external radiation from airborne radionuclides in the cloud and radionuclides deposited from the cloud onto the ground and (2) internal radiation from radionuclides inhaled directly from the passing cloud, inhaled resuspended radionuclides and ingested contaminated food and milk.
Radiation exposure from sources external to the body is calculated for time periods over which individuals are exposed to those sources and the exposure from sources internal to the body is calcu-lated over the remaining life of the exposed individual.
Based on the calculated radiation exposure to individuals downwind, the consequence model estimates the number of public health effects that would result from a radioactive release.
Early injuries, early fatalities, latent cancer fatalities, and thyroid and genetic effects are computed. Early fatal-itics are defined to be those fatalities that occur within 1 year of the ini-tial exposure. They are estimated on t he basis of exposure to the bone mar-row, lung, and gastrointestinal tract. Bone marrow damage is the dominant cont ributor to early f atalitics.
Early fatalities are calculated assuming supportive medical treatment of the exposed individual.
Early injuries (ie.,
morbidities) are defined as nonfatal, noncaccinogenic illnesses that appear _
within one year of the exposure and require medical attention or hospital I
treatment. Latent cancer fatalities occur over any time subsequent to the exposure as a result of the initial exposure (carly cancer fatalities) and of any long term chronic exposure to low-levels of radioactive contamination (chronic cancer fatalities).
The expected health consequences were calculated using the measured release fractions shown in Table 4.6.1 as input to CRAC.
The solid particu late release fractions of 3.4 x 10- 5 and 2.4 x IO' S of the solid radio nuclide inventory for 1 and 3 PWR fuel assembly casks, respectively, are based on t he full-scale test data.
The nobic gas release of 0.5 and 0.34 for 1 and 3 PWR fuel assembly casks are based upon the mass loss sustained by 50 percent of the 223 fuel rods used in the full scale test (Section 4.3) and are extrap-olated to a 3 PWR fuel assembly sabotage scenario. Thirty four percent (222 fuel rods) of the total fuel rod inventory of a 3 PWR fuel assembly truck cask are estimated to sustain some kind of cladding failure and that each failed rod would release all gases contained in the plenum. This assumes that only two out of three fuel assemblies could sustain maximum damage if the longest path of interaction for each assembly is assumed.
Other input data required for CRAC include (1) site-related data such as meteorology and population distributions, (2) radionuclide inventories and release parameters, and cool ing periods analyzed exterior to the CRAC code, and (3) emergency response scenarios.
The spent fuel radionuclide inventory used for this analysis has been generated using the fuel burning code ORIGEN18 assuming light water reactor fuel with 33,000 mwd /t of heavy metal burnup at 40 kw/kg power density and 150 days cooling.
A 1 PWR fuel assembly shipment and a 3 PWR fuel assembly truck cask shipment were used for this analysis. The 1 and 3 PW2 fuel. assembly casks were assumed to contain 0.5 t and 1.4 t of heavy metal charged to the reactor, respectively. The 1 PWR fuel assembly radionuclide inventory used in
~
this study is shown in Table 4.6.2.
The 3 PWR fuel assembly inventory may be obtained by multiplying the Curie inventory of Table 4.6.2 by 3.
i-
+
j TABLE 4.6.1 Measured Relesse Fractions for a 1 PWR Fuel Assembly and Calculated Release Fractions for a 3 PWR Fuel Assembly Truck Cask.
b One" Three Lnventory.
As_s embly.
A s_s e_mb lj e_s, d
Noble cases (Xe, Ke )
0.5 0.34 Solid RadionuclidesC 3.4 x 10- 5 2.4 x 10-5
- Based upon an inventory of 0.5 t of heavy metal b Based upon an inventory of 1.4 t of heavy metal
_r.
I C Solid radionuclides are airborne particles having smaller than a 10 m aerodynamic diameter.
d Gas release based upon cladding failure in 50 percent (111 fuel rods) of the total number of fuel rods (223) used in the full-scale reference test (see Section 4.3).
i I
e i
. The detailed population distribution employed in this model is equivalent to the Manhattan borough of New York City. The detailed distribution is shown in Table 4.6.3.
The detailed population distribution accounts for the fact that there is no population in the researched area by setting the population equal to zero in certain segments. The total population used in CRAC closely approximates the actual population within 800 km of the assumed release point.
One hundred sequences of New York City weather conditions representative of weather near the release point were used in these calculations. The estimated time of release was midafternoon and a street intersection was the assumed release point. A thermal source in CRAC was used to account f.,e the effects of high explosives lofting the material and thus reducing the close in ground level concentrations. All the consequence estimates have been made with the population in place. No attempt was made bo model or account for evacuation to avoid early exposure because evacuation may not be possible in all instances.
Table 4.6.4 shows the results of the CRAC calculation for experimentally determined releases of a single and three assembly cask sabotage event.
Because the source terms used never produced the threshold dosage for early fatalities and morbiditics, the number of early fatalities and morbidilles predicted are zero. The total latent cancer fatalities are a result of initial exposure, particle resuspension and long term exposure to contaminated ground.
For the reference base event of a sabotage attack on a one assembly truck mounted cask in downtown New York City, total latent cancer fatalities of 2/1 (mean/ peak) are predicted.
For a three assembly cask sabotage event in downtown New York City, total latent cancer fatalities of 4/14 (mean/pcak) are predicted.
Peak thyroid and bone marrow dose in rems was also calculated as a function of distance from the release point. Table 4.6.5 lists peak thycotG and bone marrow dose for a one assembly and three assembly truck cask release.
At a distance of 30 m from a three assembly cask release, the peak bone marrow dose was calculated to be 900 meem and the peak thyroid dose was calculated to be 455 meems.
At 1.61 kilometers from the release point, the peak bone marrow dose was calculated to be 22.3 meems and the peak thyroid dose was calculated to be 15.8 meems for a three assembly cask. The Protective Action Guide 19 (PAC) threshold for these distances is 1 rem.
The peak bone marrow and thyroid dose for distances of 30 m or more from the release point for a single and three assembly cask are significantly less than PAC threshold.
In summary, the early latent cancer fatalities calculated using the experimentally determined releases from a three assembly truck cask are smaller by a factor of 350/433 (mean/ peak) than the original 1978 Urban l
St udy predictions upon which the NRC interim regulations for U.S.
transport of spent fuel were based.
q r
.a
. Table 4.6.2 Spent Fuel Cask Radionuclide Inventory Used in This Study 0.5 MTHM Charged to Reactor (1 Assembly) 33,000 PMd/MTHM Burnup at 40 kW/kg 150 Days Cooling Radio 2clide*
Curies l
Co-58 1.09 x 10
Sr-89 7.48 x 10' Sr-90 4.01 x 10' 0
Y-90 4.01 x 10 Y-91 L.14 x 10 i
Zr-95 1.81 x 10 Nb-95 3.37 x 10' i
Rn-103 5.79 x 10 j
Ru-106 1.95 x 10 Te-127 2.94 x 10 Te-127m 3.00 x 10 Te-129 1.23 x 10 Te-129m 1.93 x 10 i
Cs-L34 1.20 x 10 1
Cs-L36 1.10 x 10 Cs-137 5.32 x 10 2~
Ba-140 2.79 x 10 i
~
- Radionuclides with significant health effects based upon
- Reactor Safety' Study, NUREG 75-014 (WASH 1400) i
)
- 4 q s
+
-e r-n
,.n, e
--c---,
\\
i i.
' TABLE 4.6.2 (cont'd)
Radionuclide*
Curies 1
2 s
Ta 140 3.21 x 10 Ce 141 3.71 x 10 Ce 144 4.57 x 10 Pr-14 3 4.23 x 10 Nd 147 2.99 x 10 l
Np.239 1.02 x 10 Pu-238 1.48'x 10 2
Pu-239 1.55 x 10 1
2 Pu-240 2.29 x 10 Pu 241 5.18 x 10 Am-241 6.18 x 10 cm 242 8.13 x 10 Cm-244 1.46 x 10
- Radionucildec with significant health effects based upon Reactor Safety Study NUREC15-014 (WASH 1400) i i
4 l
l l
l 9
p
,----w-
~m---
L a
w JuL-
.,4--
J
.s
+*k lt Jx LJ-A 4mm c
d 1 i TABLE 4.6.3 Population. Distribution Used for This Analysis I
i i
l Radius From Releaae Population Density (km)
(people /km )
t i
10 39,000 j
10-16 16,000 -
j 16-48
~
3,900 48-88 380
-r 88 38 i
4 l
4 J
f I-l J
il 4
I e
+
i o
P f
e p
~~u
+~g
-e q.
~
w, e~+.
eo a-
- m w
i TABLE 4.6.4 CRAC Computed Health Consequences for This Experimental Study RELEASE EARLY EARLY EARLY TOTAL FRACTLON FATALITIES MORBIDLTIES LCF LCF mean/pcak mean/ peak mean/ peak mean/ peak One Assembly Cask
- O/O 0/0 0.3/1.3 2/7 3
Three Assembly Cask 0/0 0/0 1/3 4/14 a.
50% of the noble gases and 3.4 x 10' % of the total solid. inventory.
b.
34% of the nobic gasco and 2.4 x 10~ % of the total solid inventory.
c.
Early fatalities occur within one year after exposure to the radioactivo material.
d.
Early morbidities are illnesses appearing within one year after exposure.
e.
Early latent cancer fatalitics are a result of the initial exposure only and can occur over any time cubscquent to the initial exposure.
f.
Total latent cancer fatalitics are the sum of early and long-term exposure cancer f atalities and can occur at any time subsequent to exposure.
~ -.
i
' TABIE 4.6.5 4
Peak Thyroid and Bone Marrow Dose as a Function of Distance from Release Point Peak Peak a
b b
Distance.
Bone Marrow Dose Thyroid Dose (m)
(mcem)
(mrem)
One Assembly tree Assembly One Assemoly tree Assemoly Cask Cask Cask Cask 30 424.0 900.0 215.0 455.0 1400 12.2 25.8 8.6 18.1 1600 10.5 22.3 7.4 15.8 2000 7.3 15.4 5.2 11.1 4
I a
j a.
Distance from release point.
b.
External ground exposure for 24 hours2.777778e-4 days <br />0.00667 hours <br />3.968254e-5 weeks <br />9.132e-6 months <br />.
i d
4 i
d 7.-
r
-m_.
,.,--,,~-,-,,,m.
,y
-wr
,---..w-
,---,-_--,..cw--,-,,
,%-y--.y,,,,
e-w.v.
.... 5.
CCNCLUSIONS 5.1 Measured Source Term Release for a Full-Scale 1 PWR Fuel Assembly Truck Cask Reference Event Table 5.1 sunmarizes the measured source term release parameters for a single PWR fuel assembly truck cask (containing 0.5 metric tonnes of heavy metal charged to a light water reactor) subjected to a simulated sabotage event.
A released quantity of 17 g of respicable radioactive particulates was determined based upon full scale and subscale test data. A solid respicable particulate release fraction of 3.4 x 10-5 of the solid radionuclide inventory and a release fraction of 0.5 of the noble gas inventory was determined.for the reference base event.
5.2 Calculated Source Term Release for a 3 PWR Fuel Assembly Truck Cask Sabotage Event Tabic 5.2 summarizes the calculated source term release parameters for aI three PWR fuel assembly truck cask (containing 1.4 metric tonnes of heavy metal charged to a light water reactor) subjected to a simulated sabotage event.
A release of 34 g of respirabic radioactive particulates was calcu-lated based upon full-scale and subscale test data.
A respicable particulate release fraction of 2.4 x 10-5 of the total solad radionuclide inventory and a release fraction of 0.34 of the nobic gas inventory was calculated for a three element truck cask sabotage event.
5.3 Calculated Health Effects 5.3.1 1 PWR Fuel Assembly Truck Cask Referenced Event Table 5.3.1 summarizes the health consequences due to a radioactive release from a single PWR assembly (0.5 t) spent fuel cask subjected to a simulated sabotage event. The reactor safety model CRAC employed in the Urban Study was used to calculate the health cor. sequences. The experimentally determined release parameters summarized in Table 5.1 were used as input to the CRAC estimates.
Because the source terms used never produced the threshold dosage for early fatalities and morbidites, the number of early fatalities and morbidities predicted are zero.
The total latent cancer fatalities for a hypothetical attack on a one PWR assembly truck cask in downtown New York City are predicted to be 2/1 (mean/ peak).
5.3.2 3 PWR Fuel Assembly Truck Cask Sabotage Event Table 5.3.2 summarizes the health consequences due to a radicactive release from a three PWR assembly (1.4 t) spent fuel cask subjected to a simulated sabotage event. The calculated release parameters summarized in Table 5.2.1 were used as input to the reactor safety model CRAC to compute the health effects.
The consequences were calculated for a simulated release in downtown New York City at midafternoon. The number of early fatalities and morbiditics are predicted to be zero. Total latent cancer fatalities of 4/14 (mean/pcak) are predicted.
. TABLE 5.1 Summary of Release Parameters for a 1 PWR Assembly Cask Event PARAMETER VALUE Total Fractured Fuel Mass (kg) 20.820 Total Spent Fuel Mass Removed From Assembly (kg) 5.460 Total Spent Fuel Released From Caska (kg) 2.549 Total Spent Fuel Released as Respirable (g) 17 Fraction of Fuel Inventory Released
- 5.1 x 10-3 Fraction of Fuel Inventory Released as Respicable 3.4 x 10-5 Fraction of Noble Gas Inventory Released 0.5 aAll sizes i
I l
1 I
i j
.... TABLE 5.2 Summary of Release Parameters for a 3 PWR Assembly Cask Event i
i PARAMETER VALUE Total Fractured Fuel Mass (kg) 41.612 Total Spent Fuel Mass Removed From Assembly (kg) 10.918 Total Spent Fuel Released From Caska (k )
5.098 Total Spent Fuel Released as Respicable (g) 34 I
Fraction of Fuel Inventory Releaseda 3.6 x 10-3 Fraction of Fuel Inventory Released as Respirable
- 2. 4 x 10- 5 Fraction of Noble Cas Inventory Released 0.34 aAll sizes i
a f
k
-l TARLE 5.3.1 Health Consequences for a 1 PWR Assembly Truck Cask Sabotage Event a
b C
Release Early Early Early Total'I Fraction Fatalities Morbidities LCF LCF (mean/ peak)
(mean/ peak)
(mean/ peak)
(mean/ peak) 0.50 Noble Cases 0/0 0/0 0.3/1.3 2/7 3.4 x 10-5 Solid Inventory v
aEarly fatalities occur within 1 year after exposure to the radioactive material.
bEarly morbidities are illnesses appearing within 1 year after exposure.
CEarly latent cancer fatalities are a result of the initial exposure only and can occur at any time cubsequent to the initial exposure.
dTotal latent cancer fatalities are the sum of early and long-term-exposure cancer fatalities and can occur at any time subsequent to exposure.
5
s
. ~.
TABLE 5.3.2 Summary of Computed Health Consequences for a Three PWR Assembly Truck Cask Sabotage Event D
C Reicase Early&
Early Early Totald Fraction Fatalities Morbidities LCF LCF (mean/ peak)
(mean/ peak)
(mean/ peak)
(mean/ peak) 0.34 Noble Gasec 0/0 0/0 1/3 ta'14 2.4 x 10-5 Solid Inventory
~
aRarly fatalities occur within 1 year after exposure to the radioactive
- material, bEarly morbidities are illnesses appearing within 1 year after exposure.
cEarly latent cancer fatalities are a result of the initial exposure only and can occur at any time subsequent to the initial exposure.
4 dTotal latent cancer f atalities are the sum of early and long-term-exposure cancer fatalities and can occur at any time subsequent to exposure.
i 4
i
O
. t 5.4 Degree of Precision And Accuracy of Measured Parameters and Calculated Results The accuracy of the calculated health consequences is dependent upon at least three parameters:
(1) the accuracy of the CRAC model to predict health effects, (2) the accuracy of the measured full-scale and subscale released UO2 mass, and (3) the accuracy of the correlation ratio between spent fuel and surrogate fuel. The accuracy of the measured release mass is in turn a function of the uncertainty in the measurement of the uranium mass by uranium fluorometry, the flow rate of the samplers, the sampling period, the pressure chamber volume and the ability to fit the data and extrapolate to t = 0.
The uncertainty of the measured parameters such as flow rate, uranium mass determination by fluorometry, pressure chamber volume, etc, has been estimated and using standard error propagation methods, the uncertainty of the der.ved UO2 released te irable mass has been determined. The estimated most probable ceror the released UO2 respicable mass 's 17 percent.
r l
The accur,
f the measured correlation ratio betwcon spent fuel and l
Uo2 fuel can be estimated using the results of two independent studies. The 37 NRC sponsored study at BCL indicates a value of 1 1 0.75 for the correlation spent fuel: D - UO2 ratio. The EG&G/INEL experiments indicate an average value of 3.06 i 2.54 with outlying values of 0.53 and 5.6.
The most probable error for both the BCL and EG&C/INEL experiments is 183 percent, maximum.
For conservatism in the health risk assessment, the maximum value of 5.6 was used for the correlation ratio.
The accuracy of the CRAC model's health consequence predictions was also considered in this study. Because of the large uncertaintics in weattering and decontamination assumptions of CRAC and the assumptions used in the dispersion and deposition submodel of CRAC, it is difficult to quantify the uncertainty or accuracy of the health consequence predictions.
- However, because the assumptions and statistical approximations used in CRAC and its submodels are conservative and assume conservative values for most of the model's parameters, it is expected that the maximum total latent cancer fatalltly values of 7 and 14 for a one and three PWR assembly fuel cask, respec-tively, are conservative and represent the largest values possible for the type of simulated sabotage events considered in this study. This together with the fact that a maximum value of 5.6 was used for the correlation ratio provides the basis for a conservative upper bound estimate of the health consequences that could result from a sabotage event in downtown New York City.
l l
l l
I l
5.5 Comparison of This Study's Results with Other Study's Results Table 5.5.1summarizesandcomgarestheresultsofthisexperimentalstudy for a simulated sabotage event. This with those of the 1980 Urban Study simulated sabotage event assumes an attack on a 150 day cooled three PWR assembly fuel cask containing 1.4 metric tonnes of heavy metal fuel charged to the reactor. The reference HED le the assumed attack device in both cases.
The 1918 and 1980 Urban Studies were based upon engineering judgement and limited data available at the time of the study. No comparisons were made with the 1978 Urban Study release parameters because the release fractions reported in that study were based upon a range of attack devices and not exclusively on the, reference device used in this experimental study and the 1980 Urban Study. The predicted respitable fuel mass released from a three PWR fuel assembly cask as a result of the simulated Attack.is 29 times larger for the 1980 Urban Study than for this study.
5 i
Table 5.5.2 compares the CRAC model computed health cnnsequences based upon this study's experimental data with the 1978 Urban Study's predictions.
Early latent cancer fatalities are 350/433 (mean/ peak) times hl5her based upon l analyses than that of this study.
i the 1978 Urban Study f
s l
i l,
l i
i I
i O
t I
1
.a.
TABLE 5.5.1.
COMPARISON OF EXTRAPOLATED TEST RESULTS WITH 1980 URBAN STUDY" RESULTS THIS STUDYb URBAN STUDYD TOTAL FRACTURED FUEL MASS (g) 41,612 140,000 TOTAL REMOVED FUEL MASS (g) 10,918 TOTAL FUEL MASS RELEASED FROM CASK (ALL SIZES) (g) 5,048 14,000 1
TOTAL FUEL MASS RELEASED AS RESPIRABLE (g) 34 980 FRACTION OF FUEL INVENTORY RET.HASED (ALL SIZES) 3.6 x 10-3 1 x 10-2 FRACTION OF SOLID FUEL INVENTORY REI. EASED AS RESPIRABLE 2.4 x 10-5 7 x 10-4 FRACTION OF GAS FUEL INVENTORY RELEASED FROM CASK 0.34 0.1 i
a NUREC/CR0743 b THREE PWR FUEL ASSEMBLY CASK I
i
-... TABLE 5.5.2.
COMPARISON OF CRAL COMPUTED HEALTH RFFECTS WITH THE 1918 LlHBAN STUDYa REStil.TS THIS URBANA STUDY STUDY _,
EARI Y FATAL.lTIES (MEAN/ PEAK) 0/0 4/60 EARLY MORHIDLTIES (MEAN/ PEAK) 0/0 160/1600 EARLY LCFc (MKAN/PKAK) 1/3 350/1300 TOTAL LCFC (MKAh PKAK) 4/14 b
a.
1918 tlRRAN STUDY SAND /1 1927 b.
NOT REPORTED IN 1978 URRAN STUDY c.
I.ATENT CANCKR FATALITIES l
l l
l L
d
. 7.
REFERENCE 3 1.
A.
R.
Ducharme, JR., et al, Transport of Radionuclides in Urban Environs: Working Draft Assessment, US Nuclear Regulatory Commission Interim Report SAND 77-1927, Sandia National Laboratories, May 1978.
2.
Code of Federal Regulations, Title 109 Part 73.37, " Requirements for Physical Protection of Irradiated Reactor Fuel in Transit," January 1, 1982.
3.
N.
C.
Finley, et al, Transportation of Radionuclides in Urban Environs:
Draft Environmental Assessment, NRC Research and Development Report, NUREG/CR-0743, Sandia National Laboratories, Albuquerque, NM, July 1980.
4.
E. W.
Schmidt, M.
A. Walters, B. D. Trott, J. A.
Gieseke, Final Report on Shipping Cask Sabotage Source Term Investigation, US Nucle ar Regulatory 1-Commission, NUREG/CR-2472, October 1982.
~
S.
I.
B. Wall, et al, overview of the Reactor Safety Consequence Model, NUREG-0340, US Nuclear Regulatory Commission, Washington, DC, October 1977.
6.
E. L. Wilmot, Transportation Accident Scenarios for Commercial Spent Fuel, SAND 80-2124, Sandia National Laboratocies, Febolary 1981.
7.
R.
P. Sandoval, J. P. Weber and G. J. Newton, " Safety Assessment of Spent Fuel Transportation in Extreme Environments," Waste Management '81 ANS Topical Meeting, Feb rua ry, 1981.
8.
Final Environmental Statement on the Transportation of Radioactive Material by Air and Other Modes, Vols I snd II, NUREG-0170, Office of Standards Development, US Nuclear Regulatory Commission, Washington, DC, December 1977.
9.
Federal Actions Are Needed to Improve Safety and Security of Nuclear Materials Transportation, Comptroller General of the United States, Report to Congress, EMD-79-18, May 7, 1979.
10.
J.
F. VanderVate, Investigations Into the Dynamics of Aerosols in Enclosures as Used for Air Pollution Studies, ECN-86, Netherlands Energy Research Foundation, Petten, Netherlands, 1980.
11.
Personal Communication f rom Dr. J. L. Alvarez, EG&G/INEL, September 17, 1982.
12.
A. W. Harrison, " Quiescent Boundary Layer Thickness in Aerosol Enclosures Under Convective Stirring Conditions," J. Colloid Interface Science 69) 563.
a si e 13.
S.
K. Friedlander, Smoke, Dust, and Haze, John Wiley (1977).
14 J. A. Giesehe, K. W. Lee, and L. D.
Reed, "HAARM-3 User's Manual,"
BMI-NUREG-1991, 1978.
15.
F - Gelbard and J.
H. Seinfeld, " Simulation of Multicomponent Aerosol Dynamics," J. Colloid Interface Sci., 78, 485, 1980.
16.
J. A. Gieseke and L. D. Reed, " Aerodynamic and Thermophoretic Behavior of Coagulated Sodium Oxide Aerosols," Airborne Radioactivity, ANS Winter noeting, California (1977).
L7 Y. S. Cheng and H. C. Yeh, " Particle Bounce In Cascade Impactors,"
Environmental Science and Technology 1. lj,(1979).
18.
D.
E. Bennett, SANDIA-ORIGEN User's Manual, NUREG/CR-0987, SAND 79-0299, s
Sandia National Laboratories, Albuquerque, NM, October 1979.
19.
Manual of Protective Action Guides and Protective Actions for Nuclear Inc idents, EPA-520ll-75-001, September 1975, US Environmental Protection Agency.
O
-s, o
SPECIFIED EXTERNAL DISTRIBUTION:
Battelle Columbus Laboratories Inhalation Toxicology Research Institute 505 King Avenue P. O. Box 5890 Columbus, OH 43201 Kirtland Air Force Base Kenneth D. Kok Albuquerque, NM 87115 Richard J. Burian R. McClellan B. Dale Trott C. J. Newton (10)
M. Audeen Walters
- 3. C. Yeh B. B. Boecker B. A. Wong Y. S. Cheng Oak Ridge National Laboratories EC&G Idaho National Engineering Labs.
P. O. Box X P. O. Box 1625 Oak Ridge TN 37830 I aho Falls ID 83415 A. P. Malenauskas (5)
R. W. Marshall, Jr. (10)
C. W. Parker J. Sanders J. L. Alvargz I
Office of Nuclear Regulatory Com.
U.
S. Department of Energy Washington, DC 20555 Div. LWR Fuel Cycle Projects W. R. Lahs (10)
NE 43 C. Sawyer (10)
Washington, DC 20545 T. Sherr A.
Katz R. Nelsen F. Falci (10)
R.
F. Burnett C. Ortel T. Margulies R. B. Chitwood (10)
U. S. Department of Energy U. S. Department of Transportation Albuquerque Operations Office NE 341 Albuquerque, NM 87115 Mail Stop B-107 K.
A. Carlson Washington, DC 20585 R. Y. Lowrey Thomas D. Anderson Sandia National Laboratories Albuquerque, NM 87185 1822 H. Levine 9761 J. Jackson 1832 A. Romig 9780 R. M. Jefferson 1533 F. Matthews 9781 R. E. Luna 1533 J. Weber 9781 R. P. Sandoval (30) 9268 D. Tomasko 9782 R. B. Pope 9413 J. D. Johnson 9783 G. C. Allen 9413 N. C. Finley 9783 E. L. Wilmot 9414 D. M. Ericson 9784 B. D. 7.ak 9415 D..J.
Alpert 9784 H. Church Fred Colbard E. 1. duPont de Nemours and Co.
l 861 South Glencliff Savannah River Plant Atomic f
Apt. 31 Energy Division La Habra, CA 90631 Bldg. 113-28 A Aiken, SC 92808 R. Baxter l
l l
L
\\
e,,-
E. 1. duPont de Nemours and Co C.
I. W. Llewelyn Savannah River Laboratory U. K. Atomic Energy Authority Rids. 173. IIA 11 Charles Street Aiken, SC 92808 London, SW14 42P R. Moyer Dr. Klaus A. Schneider W. R. Merritt Transnucklear GmbH 810 Belfast Rd.
Postfach 110030 Ottawa, Ontario KlA OR2 D-6450 Hanau 11 I. H. Fraser Nello Bonicelli 810 Belfast Rd Chief. Test Reactors Branch Ottawa, Ontario KIA OR2 U. S. Department of Energy 550 Second St.
Idaho Falls, ID 83401 James C. Tosces Marty Semonin Commonwealth Edison Illinois Department of Nuclear Safety P. O. Box 16/
1035 Outer Drive Chicago, IL 60690 Springfield, IL 62704 I
,r Gerald Womack H. F. Mac Donald Central Electricty Generating Board Central Electricity Generating Board Barnwood Gloucester, England Cloucester, England GI,139PB Dr. Florentine Lange D. R. Poulter Gesellschaft fur Reaktor sicherheit UKAEA Safety And Reliability Directorate Schweetncesasse Wigshae Lane, Culcheth D 5000 Koln Cheshire, England FRG i
l t
a-,
- +
- _
u _..+-_,-
-_a.e4 m
m 3
AA__
_a
..-4.-.aae->A..
a
.i~.-,4 a
a a
a 4
a j
e. f-6 y
i i
3 n.
w, P c a e, 0,3 e ca.
Nm w.' 2 by 4
l
+
+e
==^-6whw-
.hh*
ewe.-
he -
-*eh whm
- =ie.p-i 4
)
t 1
f
,f
._. -..........~.,. -. -,
A J
_ =.
_s-.
-..-... ~ _ - -.-
4
=w+mee.
--.ew.+a e.--
m._.
-e-hN
.hm-->h-.em
.-y.
+ --.
-mwhi nu h w-4
.,g,.
,w..
t
~___- -... _
~"
. - -. -......~~
--.~
l j
d i
h Sandia National Labo
..