ML20135F949
| ML20135F949 | |
| Person / Time | |
|---|---|
| Site: | Calvert Cliffs  |
| Issue date: | 02/28/1997 |
| From: | Allen M, Blanchat T, Pilch M SANDIA NATIONAL LABORATORIES |
| To: | NRC OFFICE OF NUCLEAR REGULATORY RESEARCH (RES) |
| References | |
| CON-FIN-W-6162 NUREG-CR-6469, SAND96-2289, NUDOCS 9703170250 | |
| Download: ML20135F949 (195) | |
Text
_-
NUREG/CR-6469 SAND 96-2289 Experiments to Investigate Direct Containment Heating Phenomena with Scalec. Moc els of the Calvert Cliffs Nuclear Power Plant T. K. Blanchat, M. M. Pilch, M. D. Allen Sandia National Laboratories Prepared for U.S. Nuclear Regulatory Commission 7 09 jn188n?d88)*;7 iiiialsnais!siti!ssili.n!!i g
~
s AVAILABluTY NOTICE Availability of Reference Materials Cited in NRC Pubhcations Most documents cited in NRC publications will be available from one of thw following sources:
1.
The NRC Public Document Room 2120 L Street, NW., Lower Level, Washington, DC 20555-0001 2,
The Superintendent of Docurnents U.S. Government Pnnting Office, P. O. Box 37082. Washington, DC 20402-9328 3.
The National Technical information Service, Springfield, VA 22161-0002 Although the Rsting that fonows represents the majority of documents cited in NRC publications, it is not in-tended to be exhaustive.
Referenced documents available for inspection and copying for a fee from the NRC Public Document Room include NRC correspondence and internal NRC rnemoranda; NRC bulletins, circulars, information notices, in-spection and investigation notices; licensee event reports; vendor reports and correspondence; Commission papers; and applicant and licensee documents and correspondence, The following documents in the NUREG series are ivailable for purchase from the Government Printing Office:
formal NRC staff and contractor reports, NRC-sponsored conference proceedings, international agreement reports, grantee reports, and NRC booklets and brochures. Also available are regulatory guides, NRC regula-tions in the Code of Federal Regulations, and NucIcar Regulatory Cnmmission Issuances.
Documents available from the National Technicallnformation Service include NUREG-series reports and tech-nical reports prepared by other Federal agencies and reports prepared by the Atomic Energy Commission, forerunner agency to the Nuclear Regulatory Commission.
Documents available from public and special technicallibraries irolude all open literature items, such as books, journal articles, and transactions. Federal Register notices, Federal and State legislation. and congressional reports can usually be obtained from these libraries.
Documents such as theses, dissertations, foreign reports and translations, and non-NRC conference pro-ceedings are available for purchase from the organization sponsoring the publication cited.
Single copies of NRC draft reports are available free, to the extent of supply, upon written request to the Office of Administration, Distribution and Mall Services Section, U.S. Nuclear Regulatory Commission Washington, DC 20555-0001, t
Copies of industry codes and standards used in a substantive manner in the NRC regulatory process are main-tained at the NRC Ubrary, Two White Rint North,11545 Rockville Pike, Rockville. MD 20852-2738, for use by the public, Codes and standards are usualty copyrighted and may be purchased from the originating organiza-tion or, if they are American National Standards. from the American National Standards institute,1430 Broad-way, New York, NY 10018-3308.
DISCLAIMER NOTICE This report was prepared as an account of work sponsored by an agency of the United States Government.
Neither the United States Government nor any agency thereof, nor any of their employees, makes any warranty, expressed or implied, or assumes any legal liability or responsibility for any third party's use, or the results of such use, of any information, apparatus, product, or process disclosed in this report, or represents that its use by such third party would not infringe privately owned rights.
i
A NUREG/CR-6469 SAND 96-2289 i
Experiments to Investigate Direct Containment Heating Phenomena with Scaled Models of the Calvert Cliffs Nuclear Power Plant Manuscript Completed: August 1996 Date Published: February 1997 Prepared by T. K. Blanchat, M. M. Pilch, M. D. Allen Sandia National Laboratories Albuquerque, NM 87185 R. Lee, NRC Project Manager Prepared for Division of Systems Technology Office of. Nuclear Regulatory Research U.S. Nuclear Regulatory Commission Washington, DC 20555-0001 NRC Job Code W6162
ABSTRACT The Surtsey Test Facility at Sandia National Laboratories (SNL) is used to perform scaled experiments for the Nuclear Regulatory Commission (NRC) that simulate High Pressure Melt i
Ejection (HPME) accidents in a nuclear power plant (NPP). These experiments are designed to investigate the effects of direct containment heating (DCH) phenomena on the containment load.
in previous experiments, high-temperature, chemically reactive (thermitic) melt was ejected by high-pressure steam into a scale model of either the Zion or Surry NPP. The results from the Zion and Surry experiments can be extrapolated to other Westinghouse plants; however predicted containment loads cannot be generalized to certain types of Combustion Engineering (CE) plants.
In most Westinghouse plants, there is (1) an intermediate compartment that is large compared to the reactor cavity but small compared to the main containment volume, and (2) there is no significant line-of-sight pathway for debris transport from the cavity to the main containment volume.
Containment compartmentalization is the dominant mitigating feature. These two conditions are not satisfied for five CE plants: Calvert Cliffs 1 and 2, Millstone 2, Arkansas Nuclear One Unit 2, and Palisades. In particular, although these plants have an intermediate subcompartment, that is, the steam generator compartment, there is no flow path from the cavity to that compartment. The dispersal of melt from the cavity is predominately to the dome through the annular gap around the RPV. This circumvents the main mitigation associated with containment compartmentalization that exists in Zion and most other Westinghouse PWRs.
In all of the DCH integrai effects testing conducted to-date, single phase superheated steam was used to drive the molten core simulant. This is based on the assumption that no water remains in the lower plenum at vessel breach because all the water has been vaporized by the time the vessel fails. Ilowever, if saturated water is present, then the driving fluid for molten core materials will be a two-phase water mixture at the time of vessel breach. The discharge of the water as film and droplets with the entrained debris will provide a potential heat sink since some of the thennal and chemical energy in the debris would be used to vaporize the water and may quench part of the melt.
The increased amount of steam may also increase the hydrogen production from debris / steam oxidation. However, the potential for enhancing the DCH load due to hydrogen burning may be reduced due to an increased steam fraction that may inert the atmosphere and suppress a hydrogen burn.
Calvert Cliff-like plant geometries and the impact of codispersed water were addressed as part of the overall DCil issue resolution. Integral effects tests were performed with a 1/10* scale model of the Calvert Cliffs NPP inside the Surtsey test vessel. The experiments investigated the effects of codispersal of water, steam, and molten core simulant materials on DCH loads under prototypic accident conditions and plant configurations. The results indicated that large amounts of coejected water reduced the DCH load by a small amount. Large amounts of debris were dispersed from the cavity to the upper dome (via the annular gap).
iii NUREG/CR-6469
CONTENTS Page Abstract.............
.......................................................................iii E x e cu ti ve S umm ary............................................................................................................. xv Ackno wledgments............................................................................................... xvii No m e nc l at ure.............................................................................................
1.0 I ntrod uction..............................................................................................................I 2.0 Experiment Description......................................................................................3 2.1 Geometry and Initial Conditions: The Design Basis.
..........................................4 2.1.1 Facility Geometry..................................................
.....................4 2.1.2 Melt Mass and Composition..................................................................... 6 2.1.3 Driving Pressure and Fluid.................................................................... 7 2.1.4 Test Setup and Initial Conditions............................................................... 7 2.2 Measurements and Instrumentation..................
.....8 2.2.1 Pressure Measurem ents...............................................................................
2.2.2 Temperature Measurements...........
...........9 2.2.3 Gas Compo siti o n.................................................................................. 10 2.2.4 Posttest Debris Recovery.................................................................... 10 2.2.5 Cameras.........................................................................................................11 2.2.6 A dditional Measurements........................................................................... 1 1 3.0 Ex perimental Results................................................................................................
3.1 B l o wdo wn Hist o ry........................,................................................................ 2 3.2 Vessel Pressure.................................................................................... 2 6 3.3 C avi ty Pre s s ure.......................................................................................,
I 3.4 Vessel Gas Temperature s..................................................................................... 2 i
3.5 Video Results and Interpretation.................................................................... 28 3.6 D ebris Recovery Stunmary.................................................................................. 3{
3.7 Gas Composition Measurements............................................................................. 31 1
i 4.0 An al y s es...........................................................
....................................................109 l
1
)
CONTENTS (concluded)
Page 4.1 Debris Dispersal Prior to Water Delivery.............................................................109 4.1.1 Initial Displacement of Melt..................................................................... 109 4.1.2 Entrainment Prior to Water Delivery............................................................. I i 1 4.2 Pre-HPME Vessel Behavion.......................................................................... 1 12 4.3 Thermite Reaction and HPME Interval............................................................ 117 4.4 Accumulator Depressurization....................................................................... 1 19 4.4.1 Introduction..............................................
.....................................119 4.4.2 Accumulator Depressurization During and After Ejection of a Nonflashing Liquid.................................................................................. I 19 4.4.3 Steam Only Blowdown................................
.. 126 4.4.4 Ejection of a Flashing Liquid Followed by Steam Blowdown..................127 4.5 Coherence of Debris Dispersal and Blowdown................................................129 4.6 Flow of Dispersed Material Into the Subcompartments.....................................132 4.7 Validation of the TCE Model in Open Geometry Experiments............................. I34 4.8 Impact of Coejected Water on DCH Loads in Calvert Clifts Geometry.............138 4.9 Consistency in Pressure, Temperature, and Moles.............................................140 5.0 Conclusions...................................................................................................................169 6.0 References...........................................................................................................171 NUREG/CR-6469 vi
LIST OF FIGURES i
Figure P_ age 1
i i
1.'
The Surtsey vessel, high-pressure melt ejection system, subcompartment structures, i
and thermocouple arrays used in the 1/10* scale CE DCH experiments................................19 2.
An isometric view of the Calvert Cliffs subcompartment structures and RPV model.......... 20 3.
The RPV model (with melt generator) and cavity used in the 1/10* scale CE DCH i
t experiments............................................................................................................................21 i
l 4.
Plan view of structures at Level 4.......................................................................................... 22 5.
Plan view of structures at Level 5........................................................................................... 23 6.
Plan view of structures at Level 6.................................................................................... 24 7.
Cavity configuration for delivery of melt in the CE experiments....................................... 25 8.
Blowdown history of the CES-1 experiment........................................................................... 50 9.
Blowdown history of the CES-2 experiment........................................................................ 51 10.
Blowdown history of the CES-3 experiment........................................................................ 52 11.
Blowdown history of the CE-1 experiment.......................................................................... 5?
12.
Blowdown history of the CE-2 experiment........................................................................ 54 13.
Blowdown history o f the CE-3 experiment......................................................................... 5 5 14.
Blowdown history of the CE-4 experiment......................................................................... 5 6 15.
Accumulator pressure in the CE DCH experiments...................................................
. 57 16.
Accumulator gas temperatures in the CE DCH experiments............................................ 58 17.
Vessel pressure in the CES-1 experiment from -60 to 600 s................................................ 59 18.
Vessel pressure in the CES-1 experiment from -20 to 30 s................................................ 60 19.
Vessel pressure in the CES-1 experiment from 0 to 10 s.................................................... 61 20.
Vessel pressure in the CES-2 experiment from -60 to 600 s...................................... 62 vii NUREG/CR-6469
LIST OF FIGURES (continued)
Figure Pm 1
21.
Vessel pressure in the CES-2 experiment from -20 to 30 s.................................................... 63 22.
Vessel pressure in the CES-2 experiment from 0 to 10 s..................................................... 64 l
23.
Vessel pressure in the CES-3 experiment from -60 to 600 s.............................................. 65 24.
Vessei pressure in the CES-3 experiment from -20 to 30 s................................................... 66 25.
Vessel pressure in the CES-3 experiment from 0 to 10 s........................................................ 67 l
26.
Vessel pressur;: in the CE-1 experiment from -60 to 600 s............................................... 68 27.
Vessel pressure in the CE-1 experiment from -30 to 30 s............................................ 69 28.
Vessel pressure in the CE-1 experiment from 0 to 10 s........................................................... 70 29.
Vessel pressure in the CE-2 experiment from -60 to 600 s................................................... 71 30.
Vessel pressure in the CE-2 experiment from -50 to 30 s..................................................... 72 31.
Vessel pressure in the CE-2 experiment from 0 to 10 s.......................................................... 73 32.
Vessel pressure in the CE-3 experiment from -60 to 600 s........................
.................74 33.
Vessel pressure in the CE-3 experiment from -30 to 30 s....................................................... 75 34.-
Vessel pressure in the CE-3 experiment from 0 to 10 s........................................................... 76 35.
Vessel pressure in the CE-4 experiment from -60 to 600 s..................................................... 77 36.
Vessel pressure in the CE-4 experiment from -30 to 30 s..................................................... 78 37.
Vessel pressure in the CE-4 experiment from 0 to 10 s........................................................... 79 38.
Vessel pressure in the CE DCH experiments........................................................................... 80 39.
Vessel pressure during the HPME in the CE DCH experiments............................................. 81 40.
Cavity pressure and vessel pressure in the CES-1 experiment............................................. 82 41.
Cavity pressure and vessel pressure in the CES-2 experiment..
.......83 4
NUREG/CR-6469 viii
l LIST OF FIGURES (continued) i.
Figure P, age j
l
- 42.
Cavity pressure and vessel pressure in the CES-3 experiment.............................................. 84 43.
Cavity pressure and vessel pressure in the CE-1 experiment.................................................. 85
,1 44.
Cavity pressure and vessel pressure in the CE-2 experiment............................................. 86 4
45.
Cavity pressure and vessel pressure in the CE-3 experiment.................................................. 87 I
46.
Cavity pressure and vessel pressure in the CE-4 experiment.................................................. 88 i
V i
47.
Cavity pressure in the CE DCH experiments....................................................................... 89
+
i
- 48.
Average gas temperatures in the CES-1 experiment............................................................ 90 t,
49.
Average gas temperatures in the CES-2 experiment......................................................... 91
[
50.
Average gas temperatures in the CES-3 experiment............................................................. 92 I
j 51.
Average gas temperatures in the CE-1 experiment.................................................................. 93 i
4 52.
Average gas temperatures in the CE-2 experiment.............................................................. 94 53.
Average gas temperatures in the CE-3 experiment............................................................... 95 l
54.
Average gas temperatures in the CE-4 experiment................................................................. 96 1
i i
55.
Comparison of the dome average gas temperatures in the CE experiments during the thermite reaction interval........................................................................................................ 9 7 4
56.
Comparison of the dome average gas temperatures in the CE experiments l
d urin g the H PM E........................................................................................................
4 d
57.
Comparison of the vessel mole-average gas temperatures in the CE experiments during the thermite reaction interval.......................................................................................................... 99 58.
Comparison of the vessel mole-average gas temperatures in the CE experiments d urin g the HPM E...........................................................................................................
1 s
59.
Vessel gas pressure, average temperature, and moles in the CES-1 experiment.................101 60.
Vessel gas pressure, average temperature, and moles in the CES-2 experiment..................102 ix NUREG/CR-6469 J
LIST OF FIGURES (continued)
Figure Py 61.
Vessel gas pressure, average temperature, and moles in the CES-3 experiment.................103 62.
Vessel gas pressure, average temperature, and moles in the CE-1 experiment...................104 63.
Vessel gas pressure, average temperature, and moles in the CE-2 experiment.................. 105 64.
Vessel gas pressure, average temperature, and moles in the CE-3 experiment..................106 65.
Vessel gas pressure, average temperature, and moles in the CE-4 experiment.............107 66.
Sieve analysis of debris recovered from the operating deck............................................108 67.
Conceptual layout of the system......................................................................................... 151 68.
Details of the melt generator / cavity layout......................................................................... 152 69.
Hydrogen combustion in Zion geometry SN11IET tests..............................................153 70.
Hydrogen combustion in steam-driven tests................................................................. 154 71.
Hydrogen combustion in saturated water-driven tests.......................................................155 72.
Vessel pressurization during thermite bum in cavity and prior to HPME event.................156
- 73. - Pre-HPME dome temperatures in the CE-1 experiment....................................................157 74.
Dependence of melt retention on the delay between thermite ignition and blowdown........158 75.
Heat losses from thermite on the cavity floor prior to blowdown................................... 159 i
76.
Predicted temperature history of thermite in the cavity prior to blowdown..................... 160 l
t 77.
Conceptual layout o f the system...................................................................................... I 61 78.
Comparison of the predicted depressur,zatian history for CES-1 with experimental data.......................................................................................................................162 79.
Comparison of the predicted depn.ssurization history for CES-2 with experimental data.......................................,.......................................................................163 NUREG/CR-6469 x
LIST OF FIGURES (concluded) i Figure fage j
i 80.
Accumulator depressurization in the CES-3, CE-1, and CE-2 experiments.........................164 i
81.
Procedure for estimating the coherence interval.................................................................165 l
j-82.
Coherence of dispersed debris and blowdown during cavity dispersal................................166 i
4 83.
Validation of the TCE model for Calvert Cliffs geometry and other open geometry tests..167 l
r i
84.
Validation of the TCE model in "open geometry" experiments..,....................................168 I
/
i l
l LIST OF TABLES Table Page 1.
Geometric comparisons for the CE DCH experiments..............................................13
\\
2.
M elt compo siti o n....................................................................................................... 1 6 3.
Material properties of the melt............................................................. 16 4.
Initial conditions for the CES and CE experiments.............................................17 5.
CES and CE experiments instrumentation summary.........................................18 6.
Debris mass balance in kg for the CES and CE experiments..................................... 40 7.
Mass balance for the CES and CE experiments................................................... 41 8.
Gas concentrations measured in the CES-1 experiment....................................
... 42 9.
Gas concentrations measured in the CES-2 experiment.............................................. 43 10.
Gas concentrations measured in the CES-3 experiment................................... 44 11.
Gas concentrations measured in the CE-1 experiment.................................
.45 12.
Gas concentrations measured in tne CE-2 experiment................................................ 46 13.
Gas concentrations measured in the CE-3 experiment....................................................... 47 14.
Gas concentrations measured in the CE-4 experiment................................................. 48 15.
Results from the CES and CE DCH experiments......................................................... 49 16.
Debris dispersal prior to water delivery............................................................................142 17.
Hydrogen combustion and analyses in CES/CE tests...................................................143 18.
Material properties for thermite cooling..............................................................................145 19.
Categories for accumulator blowdown histories.................................................. 145 20.
Input parameters for the CES-1 experiment...
......... 145 l
l 21.
Input parameters for the CES-2 experiment......................................
............... 14 6 NUREG/CR-6469 xii
i i
i LIST OF TABLES (concluded) i Table Pg 22.
Initial conditions and computations for CES-3......................................................................146 i
23.
Experiment data on coherence ratio for Calvert Cliffs cavity geometry...............................147 i
24.
Key parameters characterizing the coherence correlation................................................147 l
25.
Applicability of the Calvert Cliffs database to reactor applications......................................147 i
a 26.
Coherent steam and water during dispersal in the SNUCES and SNUCE tests.................148
- 27. Assessment of debris flow through nozzle cutouts..............................................................149 j
28.
Loads mitigation in open geometry experiments...............................................................149 29.
Experiment insights on coej ected water.............................................................................. 149 5
30.
Test o f P, T, N consistency.......................................................................................... 15 0 l
l xiii NUREG/CR-6469
i EXECUTIVE
SUMMARY
A series of seven experiments were performed to investigate DCH phenomena in a 1/10*
scale model of the Calvert Cliffs nuclear power plant. The Calvert Cliffs plant is typical of Combustion Engineering (CE) plants with a Bechtel annular cavity, a design that represents 5 out of the 15 CE plants in the United States. These types of cavities do not have instrument tunnels like Westinghouse plants, as do the Zion and Surry plants studied previously; in plants with Bechtel annular cavity designs the only debris dispersal pathway to the dome region is through the annular gap between the RPV wall and the biological shield wall. In these types of plants, mitigation of containment loads due to trapping in the subcompartments that was important in resolving the DCH issue for Westinghouse designs is absent, i.e., debris can be transported directly to the dome.
Understanding the impact on DCH loads of debris transport to the dome was a primary goal of these experiments.
Plant analyses of Zion, Surry, and Calvert Cliffs with SCDAP/RELAP5 showed that if the RPV dries out, the surge line or hot leg will fail and the RCS will rapidly depressurize. Hence the most likely scenario that can lead to a HPME is an accident in which the operator intervenes and refloods the RPV but is unable to prevent vessel failure. Previous DCH experiments were performed with steam as the driving fluid. However, since water overlying the molten corium is likely, the CE experiments were designed to investigate the impact of coejected water on DCH loads. The results indicated that large amounts of coejected saturated water reduce the DCH load by 15 to 20% in the experiments.
The methodology chosen to simulate a coejected water accident scenario involved reacting the thermite directly in the cavity and, after waiting for the reaction to proceed to completion, introducing high-pressure water or steam into the cavity through a hole in the bottom head of the RPV. Results of all seven Calvert Cliffs experiments indicated that 58% of the total debris recovered posttest was transported to the upper dome. In the Zion and Surry tests without the annular gap modeled, only 7% to 10% of the total debris recovered was found in the upper dome. In addition, these tests indicated that a Calvert Cliffs cavity has a substantially smaller coherence ratio than in the Zion and Surry configurations. Measurements indicated that debris was dispersed from the cavity in less than 0.1 second, whereas the blowdown of the scaled RCS volume was several seconds.
Significant amounts of hydrogen, preexisting in the Surtsey atmosphere and also produced by the thermite reaction with condensate water, burned in the reactive atmosphere tests prior to the HPME event and moderately pressurized the vessel (~0.1 MPa).
Hydrogen combustion prior to the HPME event is an artifact of the experimental method and, therefore, is not prototypic of a NPP accident. Experiment data suggest that potential hydrogen combustion during the HPME event did not contribute to loads, which may simply mean there was not enough hydrogen remaining to make a significant impact on the loads. These uncertainties in hydrogen combustion and timing in the experiments preclude taking credit for potential mitigation of hydrogen combustion resulting from coejected water in NPP analyses.
s xv NUREG/CR-6469
ACKNOWLEDGMENTS The authors express their gratitude to Michael Oliver and Tom Thornhill, who were the electronics and instrumentation engineers for these experiments, and to Tim Covert, Frank Wilkins, and Gilbert Jefferson, who were the mechanical technicians.- Three summer students participated in the performance of the test: Eric Antrim, Reginald Tillman, and Jose Walters. All operations at the Surtsey test site were performed under the capable management of the site resident mechanical engineer, Robert Nichols. T.Y. Chu and Richard Griffith reviewed the report, providing numerous helpful comments. Mary Lou Garcia typed, compiled, and edited the manuscript.
Additional guidance for the experiment programs was provided by a six-member DCH Experiment Technical Review Group (TRG). They included R.E. Henry (FAI), M. Ishii (Purdue),
F.J. Moody (GE), S. Levy (Levy and Associates), M. Corradini (U. of Wisconsin), and R. Schneider (ABB Combustion Engineering). The DCH working group for the experimental program consisted of representatives of the sponsor (NRC), universities and industry (TRG), and the national laboratories (SNL), and met periodically to discuss new results and decide future directions.
This work was sponsored by the Accident Evaluation Branch of the Office of Research of the U.S. Nuclear Regulatory Commission.
Input and valuaNe guidance were supplied by C.G. Tinkler and R.Y. Lee.
I i
xvii NUREG/CR-6469
NOMENCLATURE A,a area of annulus
=
cavity floor area Ar
=
A floor area
=
flow area in gap A,
=
hole area An
=
A.
manway area
=
ANL Argonne National Laboratory
=
A, flow area ofnozzle cutouts
=
projected area ofnozzles on gap A,%
=
cavity wall area A,
=
c4 drag coeflicient a
=
c,,a specific heat ofdebris
=
c.
specific heat of atmosphere
=
Ca discharge coefficient
=
hole discharge coefficient Ca.n
=
CC Calvert Cliffs
=
CE Combustion Engineering
=
CES combustion engineering scoping
=
centimeter cm
=
specific heat C,
=
Ca, constant in coherence correlation
=
d drop diameter q
=
dt time derivative
=
dV, change in accumulator water volume
=
D drop diameter
=
DCH direct containment heating
=
D, hole diameter
=
initial hole diameter Dn
=
statistical bias e.,
=
w Ew total energy of corium or thermite melt
=
FAI Fauske & Associates,Inc.
=
=
f'wc initial mole fraction of noncondensibles
=
f,a, fraction of blowdown steam coherent with dispersal
=
fa,,
dispersed from cavity
=
f transported outside subcompartment
=
am f,,,
ejected into cavity
=
f, fraction ofdispersed flow
=
f, subcompartment vohune fraction (0.43) or dome volume fraction (0.57)
=
f.,
fraction of dispersed melt going out manway
=
fs, noncondensible gas fraction
=
f, fraction of melt flow going out nozzle cutouts
=
xix NUREG/CR-6469
~. -.
NOMENCLATURE (continued) f fraction of melt flow carried by gas out the nozzle cutouts
=
f.
fraction of melt flow going out nozzle cutouts due to splashing
=
f-recovery fraction
=
FSAR Final Safety Analysis Report
=
f.,
transported to dome
=
f, fraction ofliquid water vaporized
=
F force
=
g acceleration due to gravity
=
G mass flux
=
Gu Bemouli mass flux
=
experiment mass flux G.,
=
'GGS gas grab sample
=
Ga homogeneous equilibrium mass flux
=
'h distance between orifice and liquid surface
=
hmo3 heat transfer coefficient through alumina
=
h, concrete heat transfer coefficient
=
h, depth ofdepression
=
ha,,
debris / gas heat transfer coefficient
=
ha.,
downward heat transfer coefficient
=
hr, heat of fusion
=
he,mo3 heat of fusion for alumina
=
h,,
iron phase heat transfer coefficient
=
h, enthalpy of saturated steam at containment pressure
=
ht initial enthalpy of saturated water
=
h, radiation heat transfer coefficient
=
h,
' =
upward heat transfer coefficient H
dome or containment height
=
idgh pressure melt ejection HPME
=
IDCOR Industry Degraded Core Rulemaking Program
=
Integral Effects Tests IETs
=
kg kilogram
=
kN force, kilonewtons
=
Kmo3 thermal conductivity of alumina
=
K, thermal conductivity ofconcrete
=-
Kr.
thermal conductivity ofiron
=
Ku Kutateladze number
=
characteristic turning length L
=
LHS lefthand side
=
LOCA loss of coolant accident
=
meter m
=
mass of atmosphere m,
=
mass ofdebris ma
=
m, ge mus
=
_m i
a g
1 NOMENCLATURE (continued) j
)
, m*,
~
initial gas mass
=
initial water mass m*,
=
megapascal MPa
=
l mw, molecular weight of gas
=
sj mass flow rate
=
A' initial gas flow rate
=
initial water flow rate K
=.
4 M
=
mus thermite mass M,
i
=
4 mass dispersed from cavity M
=
M, ge mus
=
RCS gas mass at end of entrainment interval M,,,
=
l MJ megajoule -
=
l M'a initial thermite charge.
=
initial gas mass
~
M',
=
mass frozen on roof j
Ma
=
metric tons mtonnes
=
mass ofcold water M,w/
=
effective molecular weight Mw,,
=
i MW, gas molecular weight
=
molecular weight of water MWmo
=
A jet momentum flow
=
h,'
initial gas mass flow rate
=
3 h,
water flow rate
=
gas moles N
=
- Nm initial number of gas moles of hydrogen
=
N*,
initial number of gas moles of species i
=
initial number of gas moles of nitrogen Nm
=
l N
initial number of gas moles ofoxygen
=
02 i
N' a the total pretest moles of gas
=
coherent moles Na
=
moles ofhydrogen Nm
=
hydrogen entrainedintojet Nm
=
subcompartment or dome gas moles N
=
i initial moles in accumulator N.
=
total moles initialin dome i
N*m,
=
nuclear power plant i
=
Nuclear Regulatory Commission NRC.
=
moles of hydrogen in the Surtsey vessel at time t N'm
=
moles of hydrogen bumed in the Surtsey vessel at time t N'm w
=
xxi NUREG/CR-6469
. _.._._ ~ _ _ _ _..
l
{
NOMENCLATURE (continued)
L N'mm moles of hydrogen produced in the Surtsey vessel at time t
.=
. N'i moles of gas species i in the Surtsey vessel at time t
=
N'm moles of oxygen in the Surtsey vessel at time t
=
P
. vessel pressure
=
P'-
=
normalized pressure 1
P+-
normalized pressure transient term
=
P, ambient pressure
=
P, critical pressure
=
P, pressure at end of entrainment interval
=
l P,3cs RCS pressure at end of entrainment interval
=-
P.
minimum pressure
=
P,3 minimum pressure
=
P*
initial pressure
=
Pacs initial RCS pressure
=
psi pounds per square inch
=
psia pounds per square inch, absolute
=
+
l P,
experimental depressurization rate
=
l PWR pressurized water reactor
=
1 R
universal gas constant
=
1 RCB reactor containment building
=
RCP reactor coolant pump
=
=
RHS right hand side
=
R, coherence ratio
=
. R, universal gas constant
=
le,y heat transfer rate
=
=
second s
=
SASM Severe Accident Scaling Methodology
=
=
Sh Schmidt number, ratio of the pneumatic viscosity to the mass diffusivity
=
SMMD sieve mass median diameter
=
1 SNL Sandia National Laboratories
=
time t
=
l t'
normalized time
=
l
' t, entrainment interval
=
reference time t,,,
=
T-temperature
=
TCE two-cell equilibrium l
=
T, final temperature
=
T, gas temperature
=
T subcompartment or dome average gas temperature
=
i NUREG/CR-6469 xxii L
1 i
NOMENCLATURE (continued) debris temperature T.,,
=
Three Mile Island - 2 l
TMI-2 melting temperature ofoxide T
=
time of flight TOF
=
initial temperature P
=
initial atmosphere temperature P,
=
initial debris temperature P
=
4 initial gas temperature l
P,
=
initial RCS temperature T acs
=
Technical Review Group TRG
=
initial wall temperature T.,o
=
v4 debris velocity
=
volume V
=
V',
orifice gas velocity
=
cavity volume V,
=
cavity volume V.
=
volume of gas V,
=
fmal gas volume V,f
=
i subcompartment or dome gas volume V
=
initial volume V*
=
initialvolume ofgas V*,
=
initial volume ofgas in the nozzle V*,
=
i initial water volume V.
=
Vacs RCS volume
=
freeboard gas volume V,
=
water volume V.
=
fmal water volume V.,
=
X'i initial (background) mole fraction of species i at time t=0 in the
=
containment vessel fraction of blowdown water that flashes to vapor X,
=
hydrogen concentration Xm
=
downward flammability limit Xm(dwn)
=
upward flammability limit Xm(up)
=
oxygen concentration Xo2
=
X',
mole fraction of species i at time t
=
X'm mole fraction of nitrogen at time t
=
Greek a,
thermal diffusivity ofconcrete
=
l Ae4 specific energy ofdebris
=
hydrogen combustion energy Ae
=
m l
xxiii NUREG/CR-6469
il NOMENCLATURE (concluded) specific heat of reaction Ae.
=
specific thermal energy Ae
=
w specific combined energy Ae
=
m AEi energy contributor to DCH
=
change in gas mass AM,
=
change in pressure AP
=
APw measured pressure rise
=
g predicted pressure rise AP
=
steam / hydrogen binary diffusion coefficient 3
=
emissivity c
=
p melt density
=
water density
)
p,
=
p,,,,
gas density in cavity
=
py (gas) steam density j
=
pu liquid (water) density
=
p*
initialdensity
=
p, gas density-
=
pa drop density
=
liquid density pt
=
- p, orifice density
=
blowdown time constant t,
=
entrainment time t,
=
2 variance o,,,
=
" burp" time constant t
=
T.
time constant to move liquid away from impingement region
=
g dispersal time constant i
t
=
. measured efficiency
=
4 mass fraction frozen on roof
-l
=
kinetic efficiency th
=
t, trapping time constant :
=
heat transfer time constant Tur
=
y debris / atmosphere heat capacity ratio
=
stoichiometric coefficient v-
=
p melt viscosity
=
isentropic exponent y
=
scaling group ni
=
scaling group x2
=
Stefan Boltzman constant, surface tensicn a
=
S melt layer thickness
=
NUREG/CR-6469 xxiv
-. ~. -. - -.. - -. _ -.-
i
1.0 INTRODUCTION
')
In a core melt accident, if the reactor pressure vessel (RPV) fails while the reactor coolant system is'at high pressure, the expulsion of molten core debris may pressurize the reactor containment building (RCB) beyond its failure pressure. A failure in the bottom head of the RPV, followed by melt expulsion and blowdown of the reactor coolant system (RCS), will j
entrain molten core debris in the high-velocity steam / water mixture. This chain of events is l
called a high-pressure melt ejection (HPME). Three mechanisms may cause a rapid increase in 1
pressure and temperature in the reactor containment: (1) efficient debris-to-gas heat transfer, (2) exothermic metal / oxygen reactions, and (3) hydrogen combustion. These processes that lead l.
to increased loads on the containment building are collectively referred to as direct containment heating (DCH).
l DCH experiments have been previously conducted at Sandia National Laboratories (SNL), Argonne National Laboratory (ANL), and Fauske and Associates (FAI). These early DCH experiments were reviewed as part of an NRC-sponsored effort known as the Severe Accident Scaling Methodology (SASM) Program (Zuber et al.1991). As a result of SASM recommendations, the NRC-sponsored experiment programs were redirected towards performing counterpart experiments at two different physical scales: 1/10* linear scale at SNL and 1/40*
linear scale at ANL. These counterpart experiments included geometrically scaled simulations of the Zion or Surry nuclear power plant (NPP) structures and had the initial conditions closely tied to postulated accident scenarios. These experiments, called the Integral Effects Tests (IETs),
were designed to provide integral effects data on HPME/DCH phenomena from large-scale, prototypic experiments. The primary measurements include pressures, temperatures, and gas l
concentrations.
The initial integral effects tests were conducted by SNL at the Surtsey Facility using 1/10* linear scale models of the Zion NPP structures; these tests are designated as IET-1, IET-IR, IET-3, IET-4, IET-5, IET-6, IET-7, IET-8A, and IET-8B (Allen et al.1994). These experiments used models of the Zion structures, including the bottom head of the RPV, biological shield wall, reactor cavity, instrument tunnel, containment basement floor, seal table room, refueling canal, steam generators, reactor coolant pumps (RCPs), and operating deck.
Four additional integral effects tests were performed by SNL under even more prototypic conditions with scale models of the Surry NPP (Blanchat et al.1994). The experiments were conducted at 1/6* linear scale (IET-9, IET-10, and IET-11) at the Containment Technology Test l
Facility (CTTF) and at 1/10* linear scale (IET-12) at the Surtsey Facility. Hydrogen combustion
- was examined under more prototypic atmospheric conditions, i.e. air / steam / hydrogen atmospheres likely to occur in an accident scenario.
While it appears possible to extrapolate the results obtained thus far to other Westinghouse plants, predicted containment loads cannot be generalized to certain Combustion Engineering (CE) l plants. In most Westinghouse plants, there is (1) an intermediate compartment that is large compared to the reactor cavity but small compared to the main containment volume, and (2) there is no significant line-of-sight pathway for debris transport from the cavity to the main containment 1
Introduction volume. Containment compartmentalization is the dominant mitigating feature. These two conditions are not satisfied for some CE plants, specifically, Calvert ClitTs 1 and 2, Millstone 2, l
Arkansas Nuclear One Unit 2, and Palisades. In particular, although the Calvert Cliffs-like plants have an intennediate subcompartment, that is, the steam generator compartment, there is no flow path from the cavity to that compartment. The dispersal of melt from the cavity is predominately to the dome through the annular gap around the RPV. This circumvents the main mitigation associated with containment compartmentalization that exists in Zion and most other PWRs.
In all of the DCH integral effects testing (e.g., SNL/ANL IETs for Zion, SNL IETs for Surry) conducted to-date, single-phase superheated steam was used to drive the molten core simulant. This was based on the assumption that no water remained in the lower plenum at vessel breach because all of the water had been vaporized by the time the vessel failed. However, analyses of core melt progression indicated that saturated water will still be present in the lower head at the time oflower head failure. In addition, operator intervention accidents (e.g., TMI-2) are likely to have large quantities of subcooled water in the vessel at the time of vessel failure.
The coherent water / steam and debris entrained from the cavity would provide a potential heat sink since some of the thermal and chemical energy in the debris would be used to vaporize the liquid water and quench part of the melt. The increased amount of steam may also increase the hydrogen production from debris / steam interactions. However, the potential for enhancing the DCH load due to hydrogen burning may be reduced due to the increased steam fraction which could inert the atmosphere and suppress a hydrogen burn.
The potential effects upon DCH loads caused by the coejection of liquid water with molten core materials could be significant. Therefore, the NRC requested that SNL design and test an apparatus that can be used to conduct high pressure melt ejection experiments with coejection of water and molten core simulants. The technical guidance for the initial conditions of the coejected water experiments was provided by the Accident Evaluation Branch of the NRC and a six member Technical Review Group (TRG). The TRG included R.E. Henry (FAI), M.
Ishii (Purdue), F.J. Moody (GE), S. Levy (Levy and Associates), M. Corradini (U. of Wisconsin), and R. Schneider (ABB Combustion Engineering).
Many scoping experiments were performed to demonstrate the feasibility of driving a corium simulant out of a RPV lower head model into a scaled cavity using high-pressure water.
Most of the scoping tests yielded unsatisfactory results, either due to early water interaction with unreacted thermite that caused an incomplete thermite reaction or due to large steam vaporization pressures inside the melt generator when the high-pressure saturated water contacted the molten thermite. As a result, the methodology chosen to conduct the coejected water tests involved reacting the thermite directly in the cavity and, after waiting for the reaction to proceed to completeness, introducing high-pressure water or steam into the cavity through a 4-cm hole in the bottom head of the RPV. The following sections describe the design basis for the tests and give the test description and results of seven DCH experiments conducted in the Surtsey test vessel that used a 1/10* scale model of the Calvert Cliffs NPP.
- 2.0 EXPERIMENT DESCRIPTION The Calvert Cliffs-like design was based on three principles: 1) geometrically scale the key parameters as close as possible,2) maintain design flexibility, and 3) perform cost effective and efficient modifications to existing 1/10* scale Surry NPP structures. Certain key design parameters were identified, mostly flow areas, obstruction areas, and flight paths.
Figure 1 shows the 1/10* scale Calvert Cliffs structures installed in the Surtsey vessel.
The main structure modifications included building a refueling canal, missile shield, and operating deck. The existing Surry operating deck was removed. The new operating deck was located at the top of the existing Surry crane wall. Appropriate scaled openings were placed in the operating deck to simulate the reactor coolant pump (RCP) and the steam generator (SG) vent paths. Most of the basement crane wall openings were sealed to obtain the scaled flow area. A 1/10* scale missile shield was designed to sit above the top of the refueling canal wall.
Figure 2 shows an isometric view of the Calvert Cliffs subcompartment structures and the RPV model. Since the dimensions for the Calvert Cliffs RPV were very similar to the Surry RPV, the existing RPV model design was used. A robust cavity design was necessary because large cavity pressures could not be ruled out during the HPME. It was necessary to build a new cavity to meet the design requirements. The cavity modifications include sealing the pathway to the in-core instrument tunnel, raising the floor, decreasing the cavity diameter below the nozzles, and cutting holes for the access hatch and primary loop piping pathways.
Figure 3 shows the RPV model and cavity that was used in these experiments, which are referred to as the Combustion Engineering Scoping (CES) tests and the Combustion Engineering (CE) tests. Although these experiments are referred to as the CES and CE tests, only 5 of the 15 CE NPPs in the United States are Calvert Cliffs-like with a Bechtel annular cavity design in which the only flow path out of the cavity is through the annular gap between the RPV and the biological shield wall. The crucible (or melt generator) was not used to react the thermite as it was in the earlier IET tests; the thermite was reacted on the cavity floor in all of the CE experiments.
Figures 4, 5, and 6 show plan views of the subcompartment structures in the Surtsey vessel near the Surtsey port levels 4, 5, and 6, respectively. Level 4 shows a view at the basement level, level 5 shows the view at the refueling canal, and level 6 gives the view from above the operating deck.
The melt / water delivery setup for the coejected water tests is shown in Figure 7. A 12.7 cm diameter tube filled with iron oxide / aluminum thermite (33.2 kilogram (kg)) with a small amount of alumina diluent was placed on the floor of the cavity. The tube was about 142 cm long. In the three CES experiments, which utilized a cold, nitrogen-inerted Surtsey atmosphere, the bag was formed from polyethylene material that was heat sealed at the seams. In the four CE experiments, which utilized a prototypic air / steam / hydrogen atmosphere, the bag was formed using Teflon material that was sealed at the seams using a chemical etching process along with a polysulfide adhesive. A prototype bag was tested at 373 K and was determined to 3
Experiment Description be water-tight. It was estimated that a 2.5-cm deep pool of melt formed in the cavity after the thermite reaction was complete. A 2.5-cm high concrete plug was attached to the cavity floor below the exit hole to prevent jetting of water directly into the melt pool. A thin (2-cm thick) concrete plug was formed in the cavity access hatch to prevent melt from flowing out of the cavity during the reaction process. A flow nozzle attached to the lower head of the RPV model was used as a transition piece from the 10 cm pipe to the 5 cm schedule 40 pipe. The flow nozzle had either a 5.25-cm diameter exit hole or a 4-cm diameter exit hole. The nozzle ensured that as the water rushed down the pipe that the. compression of the gas in the pipe would sweep melt away from the exit hole prior to water ejection. The nozzle also reduced the gas volume in the crucible and minimized the amount of melt that would be entrained into the annular gap and out of the cavity by the gas jet. A 0.6 cm steel tube penetrated the cavity access hatch plug and was used to drain condensate water out of the cavity during the vessel and structure heatup.
2.1 Geometry and Initial Conditions: The Design Basis The goal was to perform integral effects tests in geometrically scaled Calvert Cliffs-like structures with initial conditions generally selected to be well within the expected range of full-scale plant behavior. The geometry and initial conditions selected for the CE DCH experiments were guided by the pump seal LOCA sequence initiated by a station blackout. The Calvert Cliffs NPP was chosen as a representative CE plant with a Bechtel annular cavity design to study the 3
effect of key structures on DCH loads (Pilch 1994a). The decision was based on (1) the expected similarity of loads for all CE plants with Bechtel annular cavity designs, and (2) the Calvert Cliffs ~IDCOR Type F narrow gap cavity design would maximize debris transport to the dome, which in turn should maximize potential DCH loads.
2.1.1 Facility Geometry The cavity and RPV holddown were redesigned to meet the following requirements:
- 1) cavity design pressure of 6.9 megapascals (MPa) with a safety factor of 2 to yield and 2) RPV holddown tabs designed to 6.9 MPa with a safety factor of 4 to yield. The existing Surry cavity steel cylinder (1 cm thick) could not withstand the required design pressure. - Various means of strengthening the cavity were reviewed (internal and external steel bands and inserts, steel rope 1
or braid, etc.); however, it was decided to construct a new cavity and insert it into the existing cavity. The decision was based on cost, schedule, and confidence of the design analysis.
The new cavity was constructed from rolled 2.5-cm thick ASTM-A36 steel plate. The steel cylinder was welded to a 7.6-cm thick base plate that was then welded to the existing cavity floor. All welds in the cavity and RPV holddown assembly (Figure 3) were NDE tested by dye penetrant to ensure weld integrity in accordance with the methods suggested by the ASME Boiler and Pressure Vessel Code. The upper part of the steel shell (at and above the nozzle penetrations) was constructed using rolled 3.8-cm thick A36 steel plate. Six RPV holddown tabs (5 cm thick by 15 cm wide) were welded to the top of the cylinder. Finite element analysis (FEA) was performed on the cavity using a boundary condition of 6.9 MPa on all internal surfaces with the exception of a 27.6 MPa pressure load on the underside of the tabs. The FEA showed a Von Mises stress of 117 MPa in the center of the cylinder (a small part of the nozzle NUREG/CR-6469 4
Experiment Description cutout reached a stress of 152 MPa; however, the analysis did not take into account the stress reduction that would occur after the piping representing the flow area through the biological shield was welded to the cutouts). The maximum stress on the tabs was about 58.6 MPa. This class of steel has a yield strength of 248 MPa and an ultimate tensile strength of 476 MPa.
Therefore, based on the FEA, the steel shell had a safety factor of 2.1 to yield and the tabs had a safety factor of 4.2 to yield. Hand calculations were performed that gave good agreement with the FEA.
Finite element analysis was also performed on the existing RPV shell and holddown using a boundary condition of 6.9 MPa on all external surfaces. The FEA determined a Von Mises stress of 110 MPa in the center of the RPV cylinder (with small hotspots below and between the nozzles reaching a stress of 145 MPa). The maximum stress on the holddown blocks was small, only about 34.5 MPa. Therefore, based on the FEA, the existing RPV steel shell had a safety factor of 1.7 to yield. This was deemed sufficiently close to the design requirement to not incur additional cost redesigning the RPV model. Hand calculations were also performed on the RPV model that gave good agreement with the FEA.
The only modification to the RPV assembly was the addition of twelve 1.9 cm bolts (two bolts per tab and associated holddown block) that were used to snug the RPV holddown blocks against the bottom of the cavity tabs and remove any clearance that could contribute to dynamic loading.
The length and width of the Calvert Cliffs missile shield are 5.4 m and 7.6 m, respectively. The shield sits about 1.5 m above the top of the refueling canal wall, with a 0.9 m gap between the edge of the missile shield and the steam generator room wall. Therefore, the gap is a flowpath along all four sides of the shield (though probably small along the SG room sides). The 1/10* scale missile shield design incorporated these features.
The 1/10* scale missile shield and tie-down was designed to be very robust. The missile shield was constructed using a 10 cm steel channel framework with #5 bar welded to the channel on 15 cm centers. The minimum concrete compressive strength within the frame was 20.7 MPa.
~
For the tie-down, four threaded steel rods (1.9-cm thick) were welded to the cavity steel. The threaded 1.9 cm rods anchored the four comers of the missile shield to the cavity. It was estimated that there would be about 2.2 kN applied to the missile shield by debris flowing out of the cavity. The safety factor to yield was over 100 (assuming only one rod in pure tension).
The RPV lower head hole diameter plays a key role in determining the rate of RCS blowdown, which in tum controls the rate and magnitude of melt dispersal from the cavity. A scaled hole of ~4 cm was chosen for the CE DCH experiments to allow comparison with the 1/10* scale Zion and Surry DCH experiments. The scenario considered for the Zion and Surry experiments was a penetration-type failure of the lower head. Such a failure could occur by the ejection of an in-core instrument guide from the lower heed or by melt flow into the guide tube causing the tube to rupture outside the lower head. The initial size of such a failure is
~0.025 meter (m), but melt flow through the hole will cause it to ablate to a much larger size. A final hole size of ~0.4 m was computed with an ablation model (Pilch 1994c). The calculation was carried out using the melt mass (scaled to Surry, i.e., 43 metric tons (mtonnes)) and 5
Experiment Description composition specified in the SASM document (Zuber et al.1991). Note that the Calvert Cliffs lower head contains no penetrations.
Table I lists geometric comparisons for the CE DCH experiments. The table is based on the Calvert Cliffs (CC) subcompartment structures that were shown in FSAR drawings. The key design parameters (flow areas, obstruction areas, and flight paths) are shaded in the appropriate i
table row. The key parameters have been closely scaled; this is shown by comparing the Scaled o
column with the Surtsey column.
2.1.2 Melt Mass and Composition The experiments employ iron oxide / aluminum thermite as a high temperature, chemically reactive simulant for corium. Geometric scaling of th: melt mass for the. experiment is not strictly applicable because of material property differences between corium and thermite. The amount of thermite used in the experiments was selected so that the experiments would have the same potential for pressurization as the reactor application.
The mass of thermite chosen for the CE DCH experiments was based on the Calvert Cliffs upper bound melt mass distributions for the splinter scenario V, where the total melt mass Was I
63.7 mtonnes (Pilch et al.1995). Scenario V represents a core melt accident wheie operator actions are assumed to repressurize the RCS to 16 megapascal (MPa). The RPV is refiled with water to the hot leg nozzles (80-100 mtonnes) and the steam remaining in the RCS is at saturation.
In the Zion and Surry DCH experiments, small amounts of chromium were added to the iron oxide / aluminum thermite to cool the molten thermite to temperatures more prototypic of corium and to make the oxidation potential more prototypic of corium. The melt' composition for Scenario V is expected to be largely oxidic; hence, the chromium was replaced with an appropriate amount of alumina in the CE experiments to reduce the metallic component The amount of alumina that was added was based on maintaining the same ratio of alumiaum to iron oxide and the same heat capacity of the chromium-doped melts. Table 2 presents constituent mass, mole, and volume fractions of the melt products for an upper bound Calvert Cliffs corium from Scenario V and an alumina doped thermite. Thermophysical properties of the Calvert Cliffs NPP melt and a 1/10* scale Surtsey CE test melt using an alumina doped thermite were j
determined using the TCE model (Pilch 1991) and are given in Table 3. The 63.7 mtonnes of corium has a total combined thermal and chemical energy of 95 x 10' megajoule (MJ). This yields a required melt energy of 95 MJ for the 1/10* scale CE DCH tests. Therefore, based on the combined thermal and chemical specific energy of the thermitic melt, the required melt mass i
for the CE experiments was 33.2 kg.
Three scoping tests were performed with the oxidic thermite melt simulant (Blanchat 3
1995). The purpose of the scoping tests was to: 1) ensure ignition of the new thermite mixture in which the chromium was replaced with alumina,2) measure the time from ignition to melt plug failure (using a 1/10* scale crucible and 30 kg of oxidic thermite), and 3) measure the melt NUREG/CR-6469 6
Experiment Description
[
temperature. The thermite was ignited using a pyrofuse. Melt temperatures of 2200 K to 2500 K l
were measured. The average burn time was 6.0 seconds (s).
Two scoping test were performed using a 33 kg tube of oxidic thermite laying on the concrete floor of a test cavity (dimensions similar to the scaled Calvert Cliffs cavity). The l
l purpose of the tests was to determine the minimum delay time after ignition of the thermite prior to commencing the HPME (to ensure reaction completeness). The concrete floor was dry in the first experiment. The' thermite appeared to be fully reacted after 12 s. A dark slag appeared to form on the surface of the melt pool by 30 s. A 0.6 cm deep pool of water was placed on the l
floor in the second test. Again, the thermite appeared to be fully reacted by 12 seconds.
However, the burn seemed more intense (due to the production of hydrogen from the iron oxidation reaction with water) and slag was not seen until about 50 s after ignition.
2.1.3 Driving Pressure and Fluid One of the main objectives for the CE DCH experiments was to investigate the effects of codispersal of water, steam, and molten core materials on DCH loac. Tests that use only small amounts of lower plenum water may not significantly capture the potential DCH load reduction mechanism. Therefore, the amount of water and melt that was used was scaled according to the Scenario V amounts of saturated water (80-100 mtonnes) and the corresponding core melt mass (63.7 mtonnes). A driving pressure of 8 MPa was chosen (versus 16 MPa) because of:
- 1) recommendations of the Technical Review Group,2) to allow comparison with the 1/10* scale Zion DCH experiments which were performed at 7 MPa, and 3) the cost to modify the system to operate at higher pressures. Note that there is a related scenario (splinter scenario VI) in which water exists only in the lower plenum (<20-30 mtonnes); the RCS gas is superheated to ~1000 K which yields a driving pressure of about 8 MPa. One test used a driving pressure of about 4 MPa to determine the effectiveness of depressurization procedures that may be used in accident mitigation strategies.
2.1.4 Test Setup and Initial Conditions In CES-1, CES-2, and CES-3, following leak checks of the Surtsey vessel (the vessel leakage rate was typically on the order of 690 Pa/hr, based 12-hour leak checks at 0.2 MPa), the vessel was inerted with nitrogen gas using a feed and bleed procedure and was then pressurized with nitrogen to about 0.2 MPa. The oxygen concentration in the vessel at the beginning of the tests was about 0.2 mole percent (mole %).
In CE-1, CE-2, CE-3, and CE-4, following leak checks of the Surtsey vessel using bottled air, the vessel was vented to about 0.09 MPa at a temperature of about 285 K. At test conditions, this amount of air in Surtsey would provide the same amount of air as in the Calvert Cliffs NPP at operating conditions (0.1 MPa,311 K). The vessel was then heated and filled with steam until the vessel pressure reached about 0.22 MPa. The average gas temperature inside the Surtsey vessel was about 377 K. at the end of the heatup. Water condensed on the vessel walls and the 7
Experiment Description Calvert Cliffs structures during the 5-7 hour heatup and the condensate water was manually drained from the cavity (~17 kg) and vessel floor (1700 kg) during the steaming process.
In CE-1, the air and steam concentrations in the vessel at the beginning of the test were approximately 58 mole % and 42 mole %, respectively, based on partial pressures. A small amount of preexisting hydrogen gas was placed inside the vessel in the CE-2, CE-3, and CE-4 experiments, yielding air, steam, and hydrogen concentrations of about 55 mole %,41 mole %,
and 4 mole %, respectively, based on partial pressures and the measured addition of hydrogen.
The accumulator driving fluid varied. In CES-1, cold water (100 kg) was placed in the 1/10* scale accumulator (empty volume = 0.2544 m ) and was driven using nitrogen gas. In 3
CES-2 and CE-3, saturated steam was used as the driving fluid. In CES-3, CE-1, CE-2, and CE-4,100 kg of water was placed inside the accumulator. Heaters on the accumulator and 10 cm piping to the burst diaphragms were energized. After about seven hours, the accumulator wall, water, and steam temperatures had equilibrated at 571 K (532 K for CE-4). This produced saturated water and a saturation steam pressure of about. 8.4 MPa (4.3 MPa for CE-4) inside the accumulator. Note that the RPV exit hole diameter changed from 5.25 cm to 4.0 cm after the CES-2 experiment.
A pyrofuse embedded in the tube was used to ignite the thermite. In CES-1, CES-2, and CES-3, thirteen seconds after the thermite ignition, the burst diaphragms that contained the water and/or steam in the accumulator were failed to initiate the high pressure melt ejection process. In CE-1 and CE-2, controller problems delayed burst diaphragm failure until 29 s and 45 s had elapsed, respectively. A delay time of 29 s was chosen for the CE-3 and CE-4 experiments to allow replicate experiments using prototypic atmospheres. The time of burst diaphragm failure was used to set the zero time for the HPME. The initial conditions for the CE DCH experiments are summarized in Table 4.
2.2 Measurements and Instrumentation The most significant variables to be measured in the CE DCH experiments were: (1) the increase in pressure and temperature in the Surtsey vessel, (2) the cavity pressure, (3) the accumulator pressure, (4) the number of gram-moles (g moles) of driving water / steam, (5) the number of gwies of hydrogen generated by the reaction of metallic debris with steam and water, (6) the number of g moles of hydrogen burned, (7) the mass and location of debris recovered from the Surtsey vessel, and (8) the debris particle size.
In addition, strain measurements on key components in or near the ejection path, flow velocity of the ejected accumulator water, accumulator water level, and visual recordings of the event were made. The instrumentation and techniques used to make these measurements are summarized in Table 5 and described in the sections below.
2.2.1 Pressure Measurements i
l Pressure transducers with ranges of 0-0.69 MPa were used to measure the pressure in the i
upper dome of the Surtsey vessel. Pressure transducers with ranges of 0-20.7 MPa were used to 1
l
Experiment Description measure the gas pressure in the steun accumulator. Two strain gauge-type pressure transducers with ranges of 0-3.5 MPa and 0-34.5 MPa were used to measure the gas pressure in the scaled reactor cavity. These gages were located in tapped holes at two locations in the cavity wall, near the cavity floor, and in the annular gapjust before the cavity exit (Figure 3). Two quartz pressure transducers with ranges of 0-34.5 MPa and 0-104.4 MPa (2 ms response time) were used to measure dynamic gas pressures in the cavity. Air and hydrogen manifold pressures (and temperatures) were recorded. The number of gas moles added to the Surtsey vessel was calculated using the number of standard 44 liter compressed gas cylinders installed on a j
manifold, the cylinder volume (0.044 m'), the manifold initial gessure and temperature, and the manifold final pressure and temperature. The noncondensible gas and steam fractions at the start of each experiment were then calculated using the initial moles of air and hydrogen and the measured pressure and average gas temperature data at time t = 0 minutes, along with P, V, T ideal-gas law relationships.
The specified accuracy from the manufacturer for the pressure transducers is less than
- 0.50 percent at full-scale output. These instruments are routinely recalibrated at SNL against instruments traceable to the National Institute of Standards and Technology, and accuracies are j
always within the manufacturer's specifications. The data acquisition system recorded data from the strain-gage pressure transducers at a rate of 1400 data points per second per channel from thermite ignition to about 120 seconds following the HPME. Data from the quartz gages was recorded at a rate of 20,000 data points per second per channel (50 microsecond resolution) from thermite ignition to about 5 seconds following the HPME.
2.2.2 Temperature Measurements The bulk gas temperature above the operating deck in the Surtsey vessel will be measured with thermocouple rakes. Figures 1 and 2 show the 20 thermocouple locations for the bulk gas measurements. There were three vertical thermocouple rakes installed in the vessel; the rakes were located ~0.76 m from the vessel wall with equally-spaced thermocouples (0.61 m spacing).
Arrays A, B, and C were installed on the operating deck. Two thermocouple rakes (array SCA and array SCB) were installed in the basement below the operating deck. Figures 2 and 3 show the locations of these arrays. The three type-K thermocouples on these arrays are also equally-spaced (0.91 m). All type-K thermocouples used to measure vessel gas temperature were made of 0.127-millimeter (mm) wire with a thin Teflon sheathing. The time constant for these thermocouples is ~0.1 s. The temperature range is 273 K to 1523 K. The maximum error using the manufacturer's calibration is
- 9.4 K at 1523 K.
Type-K thermocouples were installed in the Surtsey vessel steel walls and also in the concrete subcompartment structures. In some tests, four high-temperature tungsten rhenium type-C thermocouples, comprised of 0.38-nun diameter wire with a 1.6-mm diameter stainless steel sheath, were installed in the cavity and annular gap. These thermocouples measured the temperature of the debris / gas as it exited the cavity and entered the subcompartment structures.
The temperature range for the thermocouples are 273 K to 2593 K. The maximum error using the manufacturer's calibration is 25.9 K with a 0.9-s time constant.
l Expcriment Description The temperature of the driving steam / water in the steam accumulator tank and connecting 10 cm piping was measured using nine type-K thermocouples. Measurements from these thermocouples were important because the measured temperature and pressure in the accumulator tank was used to calculate the number of g moles of steam or nitrogen driving gas.
The temperature of the steam accumulator shell was measured using three type-K thermocouples that were placed in the top and bottom hemispheres and in the vertical cylindrical wall. These thermocouples monitored and controlled the electric heaters on the accumulator shell, which heated the accumulator steel and water to the desired temperature. The 10 cm piping, from the bottom of the accumulator to the rupture disk holder, had similar heaters, controllers, and instrumentation.
An optical pyrometer was used to measure the temperature of the debris as it exits the cavity and also the timing of the debris entrainment out of the cavity. The pyrometer (type i lx30, Ircon Inc., Niles, IL) was located in the basement and was focused (through a window in the refueling canal) just above the cavity exit. The optical pyrometer had a response time of 1.5 ms to 95 percent ofits full range. A mid-to-high range controller was installed on the 1lx30 pyrometer. The controller can measure temperatures between 1873 K and 2773 K with a specified accuracy of 1 percent of the full-scale temperature. In a transient event such as a HPME experiment, the accuracy of the pyrometer measurement is expected to be no better than
- 25 K.
2.2.3 Gas Composition Twenty pre-evacuated 500-cm' gas grab sample (GGS) bottles were used to collect samples from the vessel at several locations and times. Five GGS stations were mounted on Surtsey. One station was located on the Surtsey top head and sampled gas high in the dome
'(5.5 m above the floor). Two stations were located circumferentially about Surtsey and sampled gas at a height of 3.05 m above the floor (through the level 6 ports). One station was located at the level 4 port and sampled basement gas at 0.61 m above the floor. One station sampled gas in the refueling canal (only at 15 s and 30 s into the transient). Each station contained four GGS bottles and the sample times were : 1) background, 2) 15 s, 3) 30 s, and 4) 2 min. after the HPME. In addition, two samples were taken at thirty minutes after the HPME after the vessel was mixed using internal mixing fans (in some experiments mixing fans were turned on earlier; in two experiments the mixing fans were not tumed on).' With the exception of the 15 s GGS, each sample line was purged for at least 30 seconds inunediately prior to sampling. All of the gas samples were analyzed using gas mass spectroscopy by the Gas Chromatography and Mass Spectrometry Laboratory at Sandia.
2.2.4 Posttest Debris Recovery The total debris mass dispersed into the Surtsey vessel and the debris mass in specific locations was detennined by a very careful posttest debris recovery procedure. The following measurements were made: (1) all cavity surfaces (including the annular gap between the cavity wall and the RPV model), (2) all surfaces on the refueling canal, (3) on the operating deck, (4) on the vessel wall, dome surface, and structures above the operating deck, (5) all surfaces inside the NUREG/CR-6469 10
Enperiment Description basement, and (6) in the vertical annulus between the vessel wall and the crane wall and on the l
Surtsey floor. A posttest sieve analysis of the debris that was recovered outside of the subcompartment structures was perfomied for each test. A standard set of 35 sieves was used (U. S. series 9.5 mm to 3R micrometer).
1 l
2.2.5 Cameras Two high resolution 1.3 cm CCD color cameras were used in the CE DCH experiments.
One camera was mounted outside the dome penetration, viewing the DCH event from above through a 2.5 cm thick tempered glass window. One camera viewed the DCH event through a tempered glass window mounted on a level 6 port, with a view that looked across the operating deck at the missile shield and cavity exit.
l 2.2.6 Additional Measurements Breakwires were placed across the annular gap exit and at the refueling canal openings.
The breakwires were intended to give timing information on entry of debris out of the cavity and into the Surtsey dome region. The breakwire failure time, in conjunction with measured distances, should yield debris velocity information.
In two tests, four photodiodes were used to measure the timing of the debris entrainment in the annular gap and out of the cavity. The photodiodes (equally spaced ~0.4 m apart) were mounted in the cavity wall along a vertical axis.
Two 0.6-cm square strain gages with ranges of-2 to 2% strain were attached to the cavity steel liner during construction. Identical strain gages were attached to the missile shield holddown bolts.
Accumulator water level and water velocity measurements through the 10 cm blowdown piping were attempted. Calculations indicate that 70 kg of water would be expelled from the accumulator in about 0.4 to 0.7 s, followed by a ' typical' accumulator blowdown of the remaining steam in about 3-4 s. The water expulsion is very quick, with a transient peak water velocity on the order of 25 to 40 meter /second (m/s) in the 10 cm pipe; therefore, these instruments must have a fast response time (<50 ms).
A differential pressure transducer with a range from 0-26 m (0-0.1 MPa differential) of water (that can operate in a 13.8 MPa saturated steam system) was used to measure accumulator water level. The response time for this instrument was 2 ms. The transducer was mounted outside the Surtsey vessel. Fluid filled instrument lines were used to connect the instrument to the accumulator to minimize line effects on the response time of the measurement.
l A pitot-static tube for the 10 cm pipe was designed by SNL and was used with a 0-6.9 MPa differential pressure transducer that can operate in 13.8 MPa saturated steam conditions. The response time for this instrument is 88 ms. The transducer was mounted outside l
. _.. _ = _
l Experiment Description the Surtsey vessel. Fluid filled instnunent lines were used to connect the instrument to the accumulator to minimize line effects.
t, NUREG/CR-6469 12 Y
j Experiment Description f
Table 1. Geometric comparisons for the CE DCH experiments
)
Geometric Parameters Calvert Scaled Surtsey i
Cliffs Scale Plant 1/10*
1/10*
Reactor Coolant System 1RCS, Volume (m')w x ; M;< m o W %; 414.MW W03198M? #02544 Melt Mass (kg) 59.3 x 10' n/a 30.0 Melt Volume (m )
7.27 n/a 7.73 x 10
RPV Lower Head I.D. (m) 4.368 0.4368 0.3969 RPV Shell Mid-Vessel O.D. (m) 4.822 0.4822 0.5461 l
RPV Seal Ledge O.D. (m) 5.588 0.5588 0.6223 j
RPV Lower Head Volume (m')
21.830 0.0218 0.0276 Melt / Lower Head Volume Ratio 0.3330 0.3330 0.2801 j
r Length Rom-Mat =Plugte CisvityFaboi $n)?%rg/ @4%$50RR %F @M!Tust l
Length from Melt Plug to Nozzle Centerline (m) 8.300 0.8300 0.7833 j
Length of RPV Shell and Bottom Head (m) 10.377 1.0377 1.2186 l
RPV Nozzle Average O.D. (m) 1.434 0.1434 0.1524 j
ghgMpWay "$gggpgl g@g%f ypg Q' jig pgpg
]
Annular Gap / Cavity l
Effective Annular Gap below Nozzles (m) 0.493 0.0493 0.0443
)
Effective Annular Gap at Nozzles (m) 1.119 0.1119 0.0995 Effective Annular Gap at Cavity Exit (m) 0.787 0.0787 0.0847
@FlowAma.helbwC M$$)&MPs%9sFMVWGM sn WRC _ _ MM4 2
Area Blocked by Nozzles (nozzle centerline) (m )
5.798 0.0580 0.0910 Cavity Diameter below Nozzles (m) 5.808 0.5808 0.6349 l
Cavity Diameter at Exit (m) 7.163 0.7163 0.8191 Cavity Empty Volume (m')
326.0 0.3260 0.4694 l
Cavity Free Volume (m')
145.4 0.1454 0.2169 Cavity Floor Area (m )
26.49 0.2649 0.3167 Cavity Height (m) 10.833 1.0833 1.1826
)
Cavity Access Hatch Width (square)(m) 0.762 0.0762 0.0762 l
liCavity? Access. Hatch FloWAiek (of)P75 mms M ?O48E 5%.QU:HIW/ E 0.00554P Loop Piping and Cavity Cutouts (Hot 2X, Cold 4X)
Hot Leg O.D. (m) 1.276 0.1276 n/a Cold Leg O.D. (m) 0.908 0.0908 n/a j
Hot Leg Cutout Diameter (m) 1.727 0.1727 0.1365 Cold Leg Cutout Diameter (m) 1.372 0.1372 0.1365
]
Hot Leg Area (m )
1.279 0.0128 n/a 2
l 13 NUREG/CR-6469
Experiment Description Table 1. Geometric comparisons for the CE DCH experiments
]
Geometric Parameters Calvert Scaled Surtsey Cliffs
~ Cold Leg Area (m )
0.648 0.0065 n/a 2
Total Leg Area (m')
5.149 0.0515 n/a i
IIot Leg Cutout Area (m')
2.343 0.0234 0.0146 2
Cold Leg Cutout Area (m )
2.272 0.0227 0.0146 Total Cutout Area (m )
13.775 0.1377 0.0878 iTotal Bypaqs Plow Area (Odan Leg)(m*p% > c M8.626fY W 0.0863 W, d0.0878w Refueling Canal Length (m) 21.361 2.1361 2.1336 Major Height (m) 11.418 1.1418 1.0160
- Minor Heighete)1inbove biologioni sblold)wwFe 27620w *e0J6200m N O.71t2 A 4Widthim)#MRR@nuhs%MsWMWw mV7.632m. 2 0.7632 % $ 0.81 h Volume (m')
1554.9 1.5549 1.8521 Missile Shield Length (m) 5.398 0.5398 0.5398 Width (m) 7.632 0.7632 0.8726 4 Area (mb W r % -* % W W N # @ n9M". :P 4 41;195 2 3 8 0.41_199 :- 20.47616R FHeight-AboveCavityExit(m)msfH Ama w d$7.925au d o.7925.H A 0.83825 -
Refueling Canal Openings (2X)
Effective Length (m) 7.982 0.7982 0.7965 Width (m) 7.632 0.7632 0.8179 2
Flow Area (m )
60.918 0.6092 0.6515 GTotal Flow Areai(m")N 1mW ';# #
v 121.84?: W1:2184P A1.3032*r s
RCP Vent Openings (4X)
Length (m) 3.759 0.3759 n/a Width (m) 2.553 0.2554 n/a 2
Flow Area (m )
9.599 0.0960 n/a ETotal Flow Area (m"); - %s W -
~.
m138;341 a n 0.3840:,
w 0.=4032 %
Steam Generator Openings (2X)
Length (m) 7.693 0.7693 n/a Width (m) 5.855 0.5855 n/a Occluded Area (m')
29.13 0.2913 n/a 2
Flow Area (m )
15.92 0.1592 n/a 4TotalFlow. Area (m")s ::a N 01 231.84?
i0.3184 ^
0.3084J Basement / Operating Deck Openings Containment Cross-sectional Area (m )
1233.1 12.3312 10.5071 2
Operating Deck Area (without openings) (m )
676.8 6.7680 6.3150 NTotal, Operating Deck 9==d== Ares (m%
- . 1-J 70.23 n :
-2 70236v y0.71169 0
4 1(S/G ind RCP RespovalID$ch Onanisaa)b ' ' **
il?
- l.
MO*F*
~ "'
gasem.ent/Oranewalfi ogM(m3Mighgh gg209M: ggA090lig gSM664g NUREG/CR-6469 14
Experiment Description l
Table 1. Geometric comparisons for the CE DCH experiments Geometric Parameters Calvert Scaled Surtsey Cliffs Containment Building l
Length (m) 55.37 5.5372 5.4864 Diameter (m) 39.62 3.962 3.6576 Aspect Ratio (L/D) 1.397 1.3974 1.5 Length from Basement Floor to Operating Deck (m) 17.98 1.7983 2.2352 Length from Cavity Exit to Dome (m) 44.65 4.4653 3.4925 Empty Volume (m')
62241.5 62.2415 56.710 Structure Volume (m')
5607.8 5.608 7.260 4TotalFriestmanlWhene(st')EMi" :n
,::C J 56633.7 -
- 56.634)
- 49.450 ~
+FreebondVelonwebow0ADeck (m') We?e~
- 34458.2 5
% 34.458w 127.927- <
l 2 Freeboard, Volume hdoW Opcseting Dock (m') Anix.m.
J22175.S u 4 22.175; 121.523.
9:FreeboardN6haneRatioW#"2 XnMite % f% ;.s M 0.6084 e rg0.6084:
.,;0.565.:
9A60'vti'OPST)pclcf5qsdj sshis?WO$$f)Mid9 U$ YW # Je sgl': w.5539 i g d.5539; 3Freeboa#74@ywokpegnMfggaee4me cf.298i QpsMig 1
enaseapsm 15 NUREG/CR-6469
..~.
l Experiment Description Table 2. Melt composition Mass Fraction Mole Fraction Volume Fraction Constituent Corium '
Thermite 1 Corium Thermite Corium Thermite UO, 0.8477 0.0000 0.7056 0.0000 0.7503 0.0000 ZrO 0.1272 0.0000 0.2323 0.0000 0.2149 0.0000 2
l Zr 0.0251 0.0000 0.0620 0.0000 0.0348 0.0000 Fe' O.0000 0.5315 0.0000 0.6559 0.0000 0.3172 l
Cr 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 l
Ni 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 l
Al 0 0.0000 0.4543 0.0000 0.3078 0.0000 0.6527 2 3 l
Al 0.0000 0.0142 0.0000 0.0363 0.0000 0.0300 8
Corium mass fractions are based on the Calvert Cliffs scenario V upper bound limits, shown in NUREG/CR-6338.
1 Alumina-based oxidic thermite.
Steel specie mass fractions of 0.72 for Fe, 0.18 for Cr, and 0.09 for Ni were used if steel was present.
Table 3. Material properties of the melt Property / Parameter Corium Thermite Scenario V w/ alumina Mass (kg) 63700 33.15 Volume (m')
7.9177 0.0085 Moles (g. moles) 283436 479.7 Mw,y(kg/ mole) 0.2247 0.0691 Ae,,,,a (MJ/ mole) 0.0371 0.0157 Ae,,,o (MJ/ mole) 0.2980 0.1820 s
Ae,sio,a (MJ/ mole) 0.3350 0.1980 4
co Ew (MJ) 95.0x10' 95.0 v (moles H / moles melt) 0.1241 0.7104 2
C,(J/ mole /K) 119.1 82.8 C,(J/kg/K) 525.7 1198.5 K (W/m/K) 5.0 19.6 p (kg/m')
8045 3879 p (mole /m')
3.58 x 10' 5.64 x 10 4
(Pa s) 0.0151 0.0074 T.,,,,ia, (K) 2450 2200 T.a(K) 2800 2500 l
l NUREG/CR-6469 16
Table 4. Initial conditions for the CES and CE experiments CES-1 CES-2 CES-3 CE-1 CE-2 CE-3 CE-4 Thermite composition (kg) iron oxide 24.09 aluminum 7.49 alumina 1.64 Mass of the thermite charge (kg) 33.22 Accumulator pressure (MPa) 8.274 8.523 8.329 8.634 8.340 8.030 4330 Accumulator gas or steam temperature (K) 311 607 571 573 571 622 532 Accumulator volume (m')
0.2544 Moles of accumulator gas or steam (g moles) 494 538 301 310 301 477 156 at t = 0 s (N )
2 Moles of accumulator water (g moles) at t = 0 s 5556 0
5255 5245 5255 0
5400 Time between thermite ignition and HPME (s) 13 13 13 29 45 29 29 Exit hole diameter (m) 0.0525 0.0400 Exit hole area (m )
0.0022 0.0013 U
Annular gap area (m )
0.0825 2
Vessel pressure prior to ignition (MPa) 0.2024 0.2027 0.2021 0.2103 0.2195 0.2360 0.2355 Vessel temperature (K) 298 316 305 378-376 384 380 Vessel gas moles (g moles) 4091 3851 3986 3346 3506 3700 3730 Gas composition Dry Wet Dry Wet Dry Wet Dry Wet in the Surtsey vessel (mole %)
Steam 0.0 0.0 0.0 0.0 41.9 0.0 41.0 0.0 43.6 0.0 44.2 N
99.3 99.6 99.1 78.1 45.4 73.2 43.2 72.9 41.1 72.5 40.4 2
0 0.2 0.4 0.2 20.9 12.2 19.6 11.6 19.6 11.0 19.4 10.9 7
2 H
0.0 0.1 0.4 0.0 0.0 6.2 3.7 6.7 3.8 7.2 4.0 3
[
2
[
Other 0.50 0.0 0.3 1.1 0.6 0.9 0.5 0.9 0.5 0.9 0.5 Freeboard volume inside subcompartment 21.5 P.
Q structures (m')'
y Q
Freeboard volume in upper dome (m')
27.9 y,
k 3.
Total freeboard volume (m')
49.4 8
'Ihe Surtsey empty volume is 56.71 m'.
Experiment Description Table 5. CES and CE experiments instrumentation summary Description Range Manufacturer and Model #
Response Time Comments Vessel Press 0100 psia Kulite/ Strain Gage BM-1100 122 usec Vessel Press 0-100 psia Kulite/ Strain Gage BM l100 122 psec Vessel Press 0-300 psia Precise Sensor / Strain Gage 555 182 psec Vessel Press 0-300 psia Precise Sensor / Strain Gage 555 182 psec Accumulator Press 0 3000 psig Trans Metrics / Strain Gage P53HT 57.2 psec Accumulator Press 0-5000 psig Trans Metrics / Strain Gage P531{T 57.2 psec Burst Diaphragm Press 0-5000 psig Trans Metrics / Strain Gage P53HT 57.2 psec Burst Diaphragm Press 0-5000 psig Trans Metrics / Strain Gage P53HT 57.2 psec Cavity Press 0-500 psia Trans Metrics / Strain Gage P52HT 112 psec Below RPV Cavity Press 0-500 psia Trans Metrics / Strain Gage P5311F i12 psec Near Exit Cavity Press 0-10000 psig Trans Metrics / Strain Gage P53HT 18 psec Below RPV Cavity Press 0-30000 psig Kistler/ Quartz 6230 i see Below RPV Cavity Press 0-30000 psig Vfstler/ Quartz 6230 1 psec Below RPV Cavity Pyrometer 1873-2773 K Ignitor Trigger Ignitor Battery Accumulator TC Type K TC Watlow Gordon 4 see Gas High Accumulator TC Type K TC Watlow Gordon 4 see Gas / Water Mid Accumulator TC Type K TC Watlow Gordon 4 sec Water Low TC Array Al Type K TC On Site 150 msec Top TC Array A2 Type K TC On Site 150 msec TC Array A3 Type K TC On Site 150 msec TC Array A4 Type K TC On Site 150 msec Bottom TC Array SA!
Type K TC On Site 150 msec Top (Below OP Deck)
TC Array SA2 Type K TC On Site 150 msec TC Array SA3 Type K TC On Site 150 msec Bottom TC Array B1 Type K TC On Site 150 msec Top TC Array B2 Type K TC On Site 150 msec TC Array B3 Type K TC On Site 150 msec TC AITay B4 Type K TC On Site 150 msec Bottom i
TC Array SBl Type K TC On Site 150 msec Top (Below OP Deck)
TC Array SB2 Type K TC On Site 150 msec TC Array SB3 Type K TC On Site 150 msec Bottom TC Array Cl Type K TC On Site 150 msec Top TC Array C2 Type K TC On Site 150 msec TC Array C3 Type K TC On Site 150 msec TC Array C4 Type K TC On Site 150 msec TC Array C5 Type K TC On Site 150 msec TC Array C6 Type K TC On Site 150 msec Bottom Gas Grab Various Levels 20 Samples Various Times Cavity TC Array Type C TC Watlow Gordon 500 msec 2 top. 2 bot _
Cavity Photodiode Array 4 diodes Motorola top, mid, bot _
Missile Shield Strain i 2% strain Micromeasurements Holddown Bolts Break wires Annular Gap and On Site Refueling Canal exit 4-inch Pipe DP 0-1000 psid Sensotec TJE/7564-01 89 psec Flow Rate Accumulator DP i 15 psid Sensotec A-5/5466-01 2 ms Water Level NUREG/CR-6469 18
l Experiment Description l
I 1
l DOME i
]
i f
ID b
ARRAY A ARRAY B l
i REFUELING l
CANAL BURST j
s.
4 CRUCIBLE CAVITY ACCESS l
HATCH s
i l
4 l
3 i
SURTSEY l
l Figure 1. The Surtsey vessel, high pressure melt ejection system, subcompartment structures, and thermocouple arrays used in the 1/10* scale CE DCH experiments, j
19 NUREG/CR-6469 1
l
Experiment Description MISSI.E ACCUMULATOR
/
l ARRAYB I
1 i
ARRAY A
~ ' '
~~'
ARRAY C l
'}
4 IA I RAGM
~
DE 1
I i
SUBCOMPARTMENT IfidY A i
i
~
CRUCIBLE CAVITY Figure 2. An isometric view of the Calvert Cliffs subcompartment structures j
and RPV model.
j NUREG/CR-6469 20 i
i i
l Experiment Description l
MISSILE SillELD TO ACCUMULATOR i
REFUELING CANAL BURST DIAPIIRAGM l-I CAVITY llOLDDOWN TAB RPV IlOLDDOWN BLOCK i
i ANNULER GAP RPV MODEL
{
p CAVITY PRESSURE GAGE i
4
. CAVITY ACCESSIIATCIICUTOUT l
l f
i l
l i
l Figure 3. The RPV model(with melt generator) and cavity used in the 1/10* scale CE DCH experiments.
1 j
21 NUREG/CR-6469 s
Experiment Description SG PEDESTAL ARRAY A i
(N U$
SURTSEY ACCUMULATOR l
LEGS CAVITY i
ARRAY C BASEMENT po SG PEDFSTAL Figure 4. Plan view of structures at Level 4.
i Experiment Description l
PIPING ACCUMULATOR LEGS i
l i
REFUELING CANAL t
i I
i l
I ARRAY C CRUCIBLE ARRAY 11 i
1 i
Figure 5. I';an view of structures at Level 5.
1 23 NUREG/CR-6469 y
7
Experiment Description l
LOOP PIPING SiiOWN FOR CLARITY ARRAY A ACCUMULATOR OPERATING DECK ARRAY C j
SURTSEY OPERATING DECK MISSILE ARRAY B CUTOUT S11tELD Figure 6. Plan view of structures at Level 6.
4 i
Experiment Description l
missile shield refueling canal to accumulator i
burst diaphragms I
arv i
l l
l annular gap i
i i
i i
l access hatch (with concrete plug) r thermite flow nozzle Figure 7. Cavity configuration for delivery of melt in the CE experiments, j
25 NUREG/CR-6469 j
3.0 EXPERIMENTAL RESULTS 3.1 Blowdown History
~
Figures 8 through 14 give the blowdown history for the seven CE experiments. These figures show the relationship between burst diaphragm pressure and accumulator pressure, and the resultant cavity pressure as the melt is entrained out of the ca'vity during the HPME. Note
)
' that the ratio of the cavity pressurization time to the accumulator blowdown time is very small.-
At t = -0.2 s, the signal to fail the burst diaphragms was sent. At t = 0 s, the burst diaphragms failed. Note that the longer blowdown times in CES-3, CE-1, CE-2, CE-3, and CE-4 were caused by the change in the flow nozzle diameter (from 5.25 cm to 4.0 cm). Also note the I
increased rate of pressure decay in the accumulator pressure data in the CES-1 test (at 0.8 s), in the CES-3, CE-1, and CE-2 tests (at 2.4 s), and in the CE-4 test (at 3.2 s). This was due to the transition between water ejection and gas ejection.
l l
l Figure 9 shows that the accumulator pressure did not track the blowdown very well in the
)
l CES-2 experiment. Typically, accumulator pressure tracks burst diaphragm pressure after l
' equilibration. The gages were affected by heat after melt landed on them. Figure 15 compares l
the accumulator pressure during the blowdown for the seven. experiments. For the CES-2 l
experiment, the burst diaphragm pressure was assumed to give accumulator pressure after l
equilibration. Pinhole leaks (due to melt impacts) were found posttest in the lines connecting the
[
pressure transducers to the burst diaphragms in the CES-3 experiment. These leaks caused the l
lower than expected measured burst diaphragm pressure after the HPME (see Figure 10).
Accumulator gas temperature is shown in Figure 16. The gas temperature was used in conjunction with the accumulator pressure to determine the amount of moles of driving gas at the start of the blowdown. The steam temperature closely tracked saturation temperature during the blowdowns that involved saturated water or saturated steam.
l 3.2 Vessel Pressure L
Three figures plot vessel pressure at different time scales for each experiment (Figures 17 through 37). The first pressure plot shows the big picture, giving vessel pressure 60 seconds prior to melt ejection (t = 0 s) to 600 seconds after the HPME. The thermite ignition time is also shown. The next figure examines the two pressure increases that occurred in each test: the first increase was due to heating of the gas in the vessel during the thermite reaction interval and the second increase was due to the HPME. The third figure of each set examines in detail the vessel l'
pressure increase during the HPME.
. The vessel pressurized immediately after thermite ignition.
In the tests with the nonreactive atmmpheres, the vessel pressurized slightly (0.02 MPa).
In CE-1, a reactive atmosphere without preexisting hydrogen, the vessel pressurize increase was about 0.06 MPa. In the CE-2, CE-3, and CE-4 tests, which contained a reactive (air and steam) atmosphere with l
preexisting hydrogen, the vessel pressurize increase was about 0.12 MPa. The cause for the initial pressure rise was apparent from the camera views inside the Surtsey vessel. When the NUREG/CR-6469 26
Experimental Results vessel was inerted with nitrogen, a thick black aerosol cloud rose out of the cavity annular gap, engulfed the refueling canal and missile shield, and filled the Surtsey vessel in about six seconds.
When the vessel atmosphere was reactive, a gray / white (possibly steam) cloud and flames rose out of the cavity annular gap. The fire was very intense; the flames erupted from the refueling canal and rose at least one meter above the operating deck. Some molten particles were ejected out of the cavity and onto the operating deck. The cloud was much lighter in color than in the nonreactive atmosphere tests. Note that interactions of the melt with residual condensate water in the cavity could form steam and hydrogen, which could have bumed as it was pushed out of the cavity.
Figures 38 and 39 compare vessel pressure at different time scales for all of the CE experiments. Direct comparisons of vessel pressures between the CES experiments, performed in nitrogen atmospheres, and the CE experiments, performed in air / steam atmospheres (sometimes with hydrogen), are misleading due to differences in the specific heats of nitrogen compared to air / steam. These differences will be accounted for in the Analysis section. Figure 39 shows that during the HPME, the vessel peak pressure increase ranged from 0.2 MPa to 0.3 MPa and was reached at about t = 3 s; however, between 75% and 90% of the vessel pressure increase occurred by about t = 0.1 s (closely tracking the cavity pressure transient). The side camera usually captured the HPME in only one frame; this yields a debris entrainment interval out of the cavity in the range of 33 ms. In that one frame, the molten debris rose upward in the refueling canal; some of the debris impacted the bottom of the missile shield and was deflected to the side and out of the refueling canal. A few tenths of a second later, molten debris would fall downward in front of the side camera window (over a one second interval).
3.3 Cavity Pressure Figures 40 through 46 compare cavity pressure with vessel pressure for the CE experiments. The cavity peak pressure usually occurred at about t = 0.06 s, followed by an equilibration with the vessel pressure by t = 0.1-0.2 s. The debris entrainment interval is defined 4
as that period of time when the cavity pressure exceeds the vessel pressure. The cavity pressurization was due to the acceleration of debris out of the cavity and corresponded closely to the ejection measured by the pyrometer, which was mounted on the outside of the refueling canal, looking through a quartz window directly across the cavity exit. In a few tests, the 4
pyrometer trace and the cavity pressure trace both showed an entrainment interval of about 0.1 to 0.2 s (confirming the video interpretation). However, the pyrometer time was normally unusable due to the intense aerosol generation during the thermite reaction interval. Again, note that the cavity pressure equilibrated with the vessel pressure immediately after the debris entrainment interval, even though the accumulator blowdown has just started (with high pressure and large amounts of water remaining in the accumulator). The cavity does not pressurize during the remaining blowdown because of the large area ratio (>38) between the annular gap and the flow nozzle exit.
Figure 47 compares the cavity pressure in the seven CE DCH experiments. Peak pressure in the inerted tests (with thermite reaction intervals of about 12 s) ranged from 2.8 MPa to 3.3 27 NUREG/CR-6469
- Experimental Results MPa. ~ Peak pressures in the reactive atmosphere tests (with thermite reaction intervals of 29 s to 44 s) ranged from 0.4 MPa to 1.7 MPa.
3.4 Vessel Gas Temperatures Figures 48 through 54 show the vessel average gas temperatures determined from the thermocouple arrays located above and below the operating deck in the CE experiments.
Standard linear averaging was used for the dome and subcompartment gas temperatures. A mole-average bulk gas temperature is derived below.
i E P V, To E No To R To Vra 1
<T>m g3,gy j
g 73 g[
E No PV V
R To To To
)
where f
= subcompartment volume fraction (0.43) or dome volume fraction (0.57) i
-N
= subcompartment or dome gas moles i
P~
= vesselpressure R
= universal gas constant T
= subcompartment or dome average gas temperature i
Vr, = freeboard gas volume V
= subcompartment or dome gas volume.
i
.I Note that the region below the operating deck (43% of the total freeboard volume) stays relatively cool both during the thermite reaction period and also during the HPME. -This was typical behavior for all of the experiments. Figures 55 and 56 compare the calculated dome average gas temperatures in the Surtsey vessel in the CE experiments during the thermite reaction interval and during the HPME, respectively. Figures 57 and 58 compare the calculated mole-average gas temperatures in the Surtsey vessel in the CE experiments during the thermite reaction interval and during the HPME, respectively. Figures 59 through 65 show the calculated number of moles of gas in the Surtsey vessel (using the ideal gas law with pressure data and mole-average gas temperatures) for each experiment.
3.5 Video Results and Interpretation Two CCD cameras were used to view the inside of the Surtsey vessel. One camera looked down from the window port in the Surtsey upper head and one camera looked across the operating deck directly at the missile shield through a level 6 window port.
In CES-1, the thermite was ignited at t = -13 s.
Almost immediately, a thick black
'l aerosol cloud rose out of the cavity annular gap, engulfed the refueling canal and missile shield, and filled the Surtsey vessel (by t = -6 s). The camera view was obscured for the next six l
_____m Experimental Results seconds. At t =i 0 s, the burst diaphragms failed, and a flash of orange light was seen in both views. The upper head video showed molten particles impacting the top window. Debris
[
ejection out of the cavity was captured by the level 6 port camera in only one frame; this yields a cavity entrainment interval on the order of 33 ms. In that one frame, the molten debris was seen rising upward in the refueling canal; some of the debris impacted the bottom of the missile shield j
- and was deflected to the. side and out of the refueling canal. A few tenths of a second later, i
molten debris was seen falling downward in front of the level 6 camera window (over a one second interval).
In CES-2, the thermite was ignited at t = -13 s. The top view showed that a thick black aerosol cloud rose out of the cavity annular gap (t = -8 s), engulfed the refueling canal and missile shield, and filled the Surtsey vessel (by t = -3 s). The top camera view was obscured for the next three seconds. The side view showed that the aerosol cloud filled the refueling canal by t = -7 s. At t = 0 s, the burst diaphragms were failed, and a flash of orange light was seen in the top (upper head) view. The upper head video showed molten particles impacting the top window over an interval of about 0.15 s to 0.30 s. The side camera failed due to a circuit trip during the HPME (caused by melt impacting and breaking the internal lights), and no entrainment of debris from the cavity was seen in that view.
In CES-3, the thermite was ignited at t = -13 s. The top view showed that a thick black aerosol cloud rose out of the cavity annular gap (t = -8 s), engulfed the refueling canal and missile shield, and filled the Surtsey vessel (by t = -4 s). The top camera view was obscured for the next four seconds. The side view showed that the aerosol cloud filled the refueling canal by t = -8 s. At t = 0 s, the burst diaphragms were failed, and a flash of orange light was seen in the top (upper head) view. The upper head video showed molten particles violently impacting the top window over an interval of about 0.17 s. Slower moving molten particles were seen either slowly rising or dripping from the upper head for an additional 0.5 s. The side camera captured the HPME in only one frame (0.033 s). The side view was dark for the next four frames, then falling molten drops were seen for about 0.25 s, followed by 0.75 s of drifling molten drops.
In CE-1, the top view showed light reflecting off of the wet concrete surfaces. Water was seen dripping from the dome onto the missile shield. Note that in the CES tests, the surfaces were dry. The thermite was ignited at t = -29 s. The top view showed that smoke and flames started to rise out of the cavity annular gap at t = -25 s. An aerosol or steam cloud obscured the flame at t = -24 s. The cloud was much lighter in color than in the other tests, gray / white versus dark black, which may be indicative of steam formation. At t = -22 s, the cloud reached the top window and obscured the top view. At t = 0 s, glowing orange molten particles impacted the top window, The molten particles violently impacted the top window over an interval of about 0.25 s. Slower moving molten particles were seen either slowly rising or dripping from the upper head for an' additional 0.25 s. A combustion flame could not be seen through the aerosol cloud.-
The side view showed that a smoke or steam cloud and fire impacted the bottom of the missile shield at t = -25 s. At t = -24 s, the whitish-colored cloud filled the refueling canal and obscured the flames that exited the cavity. The view was still mostly obscured by the cloud at t = -22 s; however, flames were seen licking the missile shield. At t = -21 s, an intense fire ball erupted from the refueling canal. Molten particles were ejected out of the cavity and fell onto the 29 NUREG/CR-6469
Experimental Results operating deck. At t = -19 s, either the flames had stopped or the aerosol or steam cloud completely obscured the view, at t = 0 s, when the burst diaphragms were failed, the side camera did not capture the HPME (note that in the previous three tests, the HPME was seen in only one frame). The side view showed molten drops falling for about 0.25 s, followed by 0.75 s of drifting molten drops.
In CE-2, the thermite was ignited at t = -45 s. The top view showed an aerosol or steam cloud, flames, and molten debris violently rising out of the cavity annular gap at t = -44 s. The glowing-orange cloud reached the top of the vessel about three seconds later and obscured the camera view. However, unlike the previous tests where the view immediately turned black, the cloud glowed orange for an additional three seconds, as ifit was backlit by a flame source. At t = 0 s, glowing orange molten particles impacted the top window. The amount of molten particles seen in the top view seemed substantially less than that seen in previous tests. The side view showed that an intense fire and a small burst of molten debris erupted from the refueling canal at t = -43 s. The flames were seen for about three seconds, until an aerosol or steam cloud obscured the view. At t = 0 s, when the burst diaphragms were failed, the side camera did not capture the HPME (as in CE-1) The side view showed molten drops falling for about 0.5 s.
Again, the amount of falling molten particles seen in the side view seemed substantially less than that seen in previous tests.
In CE-3, the thermite was ignited at t = -29 s. The top view showed an aerosot or steam cloud, flames, and molten debris violently rising out of the cavity annular gap at t = -26 s. The glowing-orange cloud reached the top of the vessel about two seconds later and obscured the camera view. The cloud then pulsed and glowed orange for an additional six seconds, as ifit was backlit by a flame source (similar to the CE-2 test, except much stronger and about three seconds longer). At t = 0 s, glowing orange molten particles impacted the top window. The particle stream lasted about 0.5 s, but again, as seen during the thermite reaction interval, the view pulsed orange (as if from a flame plume) for an additional two seconds. This glowing view immediately after the HPME was not seen in the CE-2 water-driven, melt ejection test. The side view showed that an intense fire and a small burst of molten debris erupted from the refueling canal at t = -26 s. The flames were seen for about three seconds, until an aerosol or steam cloud obscured the view. The view remained black until t = 0 s. At t = 0 s, when the burst diaphragms were failed, the side camera did not capture the HPME (as in CE-1 and CE-2). The side view showed molten drops 411ing for about I s; no orange glows were seen.
In CE-m the thermite was ignited at t = -29 s. The top view showed an aerosol or steam cloud, flames, and molten debris violently rising out of the cavity annular gap at t = -26 s. The glowing-orange cloud reached the top of the vessel about two seconds later and obscured the camera view. The cloud then pulsed and glowed orange for an additional four to five seconds, as ifit was backlit by a flame source (similar to the CE-2 and CE 3 tests). At t = 0 s, glowing orange molten particles impacted the top window. Particles could be seen in the top view for about I s. The top view did not pulse orange after the HPME (as was seen only in the CE-3 test).
The side view showed that an intense fire and a small burst of molten debris erupted from the refueling canal at t = -26 s. The flames were seen for about three seconds, until an aerosol or steam cloud obscured the view. The view remained black until t = 0 s. At t = 0 s, when the burst 4
i 4
Experimental Results
}
diaphragms were failed, the side camera did not capture the HPME (as in CE-1, CE-2, and CE-
- 3). The side view showed molten drops falling for about I s; no orange glows were seen.
3.6 Debris Recovery Summary Debris in the Surtsey vessel was recovered from six locations: (1) all cavity surfaces (including the annular gap between the cavity wall and the RPV model), (2) all surfaces on the refueling canal, (3) on the operating deck, (4) on the vessel wall, dome surface, and structures i
above the operating deck, (5) all surfaces inside the basement, and (6) in the vertical annulus between the Surtsey vessel wall and the crane wall and on the Surtsey floor. A posttest sieve analysis of the debris that was recovered from the operating deck floor was performed for each
[
test. A standard set of 35 sieves was used (U. S. series 9.5 mm to 38 mm). Table 6 gives the debris recovery summary which lists the locations of all debris recovered in the CE DCH experiments. Table 6 and Figure 66 also show the posttest sieve analysis results. The particle size analysis discounted all debris with sizes >9.4 mm and <0.038 mm. The particle size sieve mass median diameter (SMMD) for all of the tests was on the order of 0.3 mm with a near lognormal distribution. Note that the SMMD was ~0.6 mm in the CE-4 experiment; the slightly 1
larger SMMD was probably due to the lower driving pressure.
)
l Table 7 gives the mass balance for the CE experiments. A recovery fraction greater than one indicates that the total mass available for dispersal into the Surtsey vessel was greater than i
the initial thermite charge due to ablation of concrete in the cavity, contaminants (breakwires, i
thermocouples, etc.), and oxidation of metallic debris. Table 7 also gives transport fractions based on the mass balance. The definitions for computing the transport fractions from the mass balance are also shown. The transport fractions depend on the mass recovered from the locations specified. Debris from the floor area could not be collected following the CE-2, CE-3, and CE-4 experiments because the CE structures were not removed between tests. The amount of debris in the Surtsey vessel (dome, walls, and floor) for those experiments was estimated using an average total recovered mass of 38.55 i 0.16 kg, based on the four previous experiments. This was done to allow calculation of the mass balance and the transport fractions.
Some of the debris that is transported to the dome area falls back into the refueling canal.
All of the debris found in the refueling canal was included in the debris found outside subcompartment structures because: (1) the missile shield does not appear to be very effective in trapping debris, and (2) the debris in the refueling canal directly heats the gas in the region above the operating deck.
3.7 Gas Composition Measurements Gas grab samples used to measure the vessel atmospheric composition were taken at a dome penetration, at two level 6 port (operating deck) penetrations, and also in the basement and in the refueling canal (using extension lines) in all experiments with the exception of CES-1 (operating deck samples only). All samples were taken following a 30 s line purge (with the exception of a 15 s purge for the 15 s gas samples). The times of the samples were set for 31 NUREG/CR-6469
Experimental Results background,15 s,30 s,2 min., and 30 min. Mixing fans attached to the underside of the Surtsey 1
vessel upper head were operated prior to taking the background samples. The 30 min. samples
_ were also well-mixed with the exception of CES-1 and CE-1. The mixing fans were energized earlier in a few experiments. The gas concentrations measured in the CES and CE experiments are given in Tables 8 through 14.
The gas grab samples were taken from an atmosphere containing a mixture of steam and noncondensible gases. If the sample bottles were cold, it was determined experimentally that they would pressurize to vessel pressure with only noncondensible gas. Gas mass spectroscopy was l
performed on gases from sample bottles at room temperature. Since the steam in the bottles
- condensed prior to analysis, the measurements are only of noncondensible gases; thus, the mole
]
percent of the individual gas species determined for each bottle must be adjusted by a noncondensible gas fraction (fuc). The pretest or background noncondensible gas fraction could be calculated based on the gas and steam additions to the Surtsey vessel during the charging process.
1 For example, in CE-4, venting air from the vessel after the leak check (to 0.092 MPa at 288 K) prior to the steam addition placed 1931 g moles of air inside the Surtsey vessel. After steam was added to Surtsey to adjust the atmosphere conditions to 0.234 MPa and 378 K, then about 149 g moles of hydrogen gas was added. Immediately before the thermite ignition, the total moles of noncondensible gas (air and hydrogen) and steam was 3730 g moles. Therefore (assuming no leakage), the background noncondensible gas fraction was 0.558.
This method yields i
background wet-basis gas concentrations inside the Surtsey vessel of 44.2 mole % steam,51.8 mole % air (40.4 mole % nitrogen,10.9 mole % oxygen), and 4.0 mole % hydrogen. On a dry j
basis, the concentrations were 92.8 mole % air and 7.2 mole % hydrogen.
There were usually small differences between the concentrations calculated with the j
method described above and the values determined from mass spectroscopy analyses.' For example, Table 14 shows the results of dry-basis gas mass spectroscopy analyses performed by
'SNL after the CE-4 experiment. Multiplying the concentrations of the background gas grab sample measurement (listed in Table 14) by the background noncondensible gas fraction yields wet-basis gas concentrations. The background wet-basis gas concentrations inside the Surtsey j
vessel determined from mass spectroscopy were 44.8 mole % steam,40.6 mole % nitrogen,10.3 mole % oxygen, and 3.7 mole % hydrogen.
The need to estimate the posttest noncondensible gas fraction introduces uncertainty in the calculated amounts of posttest hydrogen. The nitrogen ratio method described below does not require an estimate of the posttest noncondensible fraction (Blanchat et al.,1994). It does, however, require the pretest noncondensible fraction. The data and assumptions required for the nitrogen-ratio method are listed below:
1.
The initial noncondensible fraction, f'sc, must be known.
2.
The total pretest moles of gas, Nb, including steam and noncondensible gases, must be j
known. -
1 3.
The measured ratios of the pretest and posttest noncondensible gases must be known.
Experimental Results 1
4.
It must be assumed that nitrogen is neither produced nor consumed by chemical l
. reactions.
5.
It must be assumed that leakage between the time for which the pretest numbers apply and the time of the posttest samples does not change the ratios of the noncondensible i
fractions.
Let X'i be the initial (background) mole fraction of species.i at time t = 0 in the
' containment vessel and let Nb be the initial number of steam and noncondensible gas moles in I
the vessel. The initial number of gas moles for all species is i
l Ni = Xf Nfos.
(3.2)
Let X'i e the mole fraction of species i at time t. For the various posttest times, the number of l
b moles of nitrogen is assumed to be unchanged, and the numbers of moles of the other gases are therefore given by (3.3)
NI= NL,Xu, It is not necessary to know the posttest noncondensible fraction; only the ratio of the posttest gas species mole fraction is needed. Furthermore, provided all noncondensible gases leak in the same proportion, a correction for posttest leakage is not needed.
- Given the pretest moles of O and H from the noncondensible fractian method and 2
2
- posttest moles of 0 and H from the nitrogen-ratio method, the moles of H burned and the 2
2 2
moles of H2 Produced can be computed from N',. 6, s = 2 (No, - No)
(3.4) n N',.rnsaas = Nk, - Nk, + Nh,. s.,,.a.
(3.5) n Table 15 gives the results for the amounts of hydrogen produced and bumed for all of the experiments based on the nitrogen-ratio method and using the 30 min. gas grab sample data. The nitrogen-ratio hydrogen combustion results assume metal / steam reactions only; i.e. it assumes that none of the oxygen decrease was due to direct metal / oxygen reaction. The gas sampling procedures and results for each experiment are described below.
CES-1 Gas grab samples to measure the vessel atmospheric composition were taken at a level 6
_ port connection following a 30-s purge; the times of the samples were background,2 minutes, and 30 minutes. Table 8 shows that the average background oxygen concentration was 0.2 mole
%. The 2-min. and 30-min. samples measured 0.1 mole %. This small change was either due to 33 NUREG/CR-6469
Experimental Results hydrogen combustion or debris oxidation and will be considered negligible.
The 2-min.
hydrogen concentration was 2.9 mole %. The hydrogen concentration measured at 30-min. was 2.3 mole %. The gases in the vessel were probably not completely well-mixed at 2-min.;
therefore, the hydrogen production amounts will be based on the 30 min. data. The initial vessel gas moles was 4091. An additional 494 g moles of titrogen gas was added to the vessel during the accumulator blowdown. The total amount of gas moles in the vessel at 30-min, was 4585 g moles. Therefore, a maximum of 105 g moles of hydrogen were produced (4585 g moles x 0.023).
i The 105 g moles of produced hydrogen is probably over estimated. Note that mixing fans were not turned on prior to obtaining the 30 min gas samples. Later experiments showed that gross stratification across the operating deck can exist. For example, in CES-3, hydrogen concentrations at the operating deck were 5.2 mole % and 3.3 mole % in the basement. The hydrogen concentration above the operating deck reduced to 3.7 mole % aner the fans were turned on. Simple mole-averaging of the above and below deck concentrations yields a value of 4.2 mole %, close to the well-mixed measured value. The above deck hydrogen concentration in CES-3 decreased by a factor of 0.71 af er the mixing fans were turned on. Assuming that the i
same reduction would have occurred in CES-1 if the mixing fans were turned on, yields a well-mixed hydrogen concentration of 1.6 mole % and a hydrogen production of 75 g moles.
CES-2 Gas grab samples to measure the vessel atmospheric composition were taken at a dome penetration, at two level 6 port (operating deck) penetrations, and also in the basement and in the refueling canal (using extension lines). All samples were taken following a 30-s line purge (with the exception of a 15-s purge for the 15-s gas samples). The times of the samples were background,15 s,30 s,2 min., and 30 min. There are five gas grab sample stations with four sample bottles at each station. During the HPME, melt impacted a dome light and caused a GFCI to trip; the electrical fault tripped all breakers attached to that circuit. Unfortunately, gas grab stations 1 and 4 lost electrical power. The result was that the 15 s, 30 s, and 2 min.
refueling canal samples and also the 2 min. operating deck samples were not taken.
Table 9 shows the results of gas mass spectroscopy analyses performed by SNL. The average background oxygen concentration was 0.4 mole %. The 15 s,30 s, and 2 min. average hydrogen concentrations were 5.6 mole %,5.7 mole %, and 5.1 mole %, respectively. All of these concentrations represent volumes above the operating deck. Thirty minutes after the HPME, the mixing fans were tumed on for two minutes, the samples lines were purged for 30 s, and then two samples were taken at the level 6 port. The average hydrogen concentration in the Surtsey vessel was 3.65 mole %. The decrease was probably due to the forced mixing of the basement volume (with suspected lower hydrogen concentrations) with the dome volume. This conclusion is supported by the temperature data which indicates that a strong stratification immediately occurred and remained in place through the 30-minute sample time.
Experimental Results The hydrogen production amounts are based on the 30-min. gas grab sample data. The initial vessel gas moles was 3851. The steam moles from the accumulator blowdown had condensed by thirty rninutes. It is assumed that the number of moles in the vessel at thirty minutes is equal to the initial amount (note that at t = 30 min., the ideal gas law yields 3711 g moles, based on a pressure of 0.20 MPa and a volume-average gas temperature of 324 K).
Therefore,141 g. moles of hydrogen were produced in CES 2 (3851 g moles x 0.0365).
CES-3 Gas grab samples to measure the vessel atmospheric composition were taken at a dome penetration, at two level 6 port (operating deck) penetrations, and also in the basement and in the refueling canal (using extension lines). All samples were taken following a 30 s line purge (with the exception of a 15 s purge for the 15 s gas samples). The times of the samples were background,15 s,30 s,2 min., and 30 min. Mixing fans attached to the underside of the Surtsey vessel upper head are operated prior to taking the background and the 30 minute samples.
Table 10 shows the results of gas mass spectroscopy analyses performed by SNL. The average background oxygen concentration was 0.2 mole %. The 15 s,30 s, and 2 min. average hydrogen concentrations in the dome and operating deck regions were 5.8 mole %,5.3 mole %,
and 5.2 mole %, respectively. All of these concentrations represent volumes above the operating deck. Twenty five minutes after the HPME, the mixing fans were turned on for about four minutes, the samples lines were purged for 30 s, and then two samples were taken at the level 6 port. The average hydrogen concentration in the Surtsey vessel was 3.65 mole %. The decrease was due to the forced mixing of the basement volume (with measured lower hydrogen concentrations at 15 s,30 s, and 2 min. of 2.9 mole %,2.8 mole %, and 3.3 mole %, respectively) with the dome vohune.
The hydrogen production amounts are based on the 30-min. gas grab sample data. The initial vessel gas moles was 3986. The steam moles from the accumulator blowdown had condensed by thirty minutes. It is assumed that the number of moles in the vessel at thirty minutes is equal to the initial amount (note that at t = 30 min., the ideal gas law yields 4120 g moles, based on a pressure of 0.22 MPa and a volume-average gas temperature of 318 K).
Therefore,145 g moles of hydrogen were produced in CES-3 (3986 g moles x 0.0365).
CE-1 Gas grab samples used to measure the vessel atmospheric composition were taken at a dome penetration, at two level 6 port (operating deck) penetrations, and also in the basement and in the refueling canal (using extension lines). All samples were taken following a 30-s line purge (with the exception of a 15-s purge for the 15-s gas samples). The times of the samples were set for background,15 s,30 s,2 min., and 30 min. However, there was a 104-s delay in the start of the gas grab sample sequence; consequently the sample times were 119 s,134 s,3.7 min., and 30 min. Mixing fans attached to the underside of the Surtsey vessel upper head were only operated prior to taking the background samples.
- Experimtntal Results The pretest or background noncondensible gas fraction could be calculated based on the gas and steam additions to the Surtsey vesel during the charging process. Venting the vessel (to 0.0938 MPa at 290 K) prior to the steam insertion placed 1945 g moles of noncondensible gas inside the Surtsey vessel.
Immediately before the thermite ignition, the total moles of noncondensible gas and steam was 3346 g moles. Therefore (assuming no leakage), the background noncondensible gas fraction was 0.5813.
Table 11 shows the results of gas mass spectroscopy analyses performed by SNL.
Multiplying the concentrations of the background gas grab sample measurement by the background noncondensible gas fraction yields the wet-basis gas concentrations.
The background wet-basis gas concentrations inside the Surtsey vessel were 41.9 mole % steam,45.5 mole % nitrogen,12.0 mole % oxygen, and 0.0 mole % hydrogen.
Thirty minutes after the HPME, the samples lines were purged for 30 s, and then two samples were taken at the level 6 port. The mixing fans were not turned on (inadvertently) prior to the taking of the 30-min. gas grab samples. However, it appears that some mixing did occur between the 3.7 min. sample time and the 30 min. sample time, evident by the decrease in hydrogen and the increase in oxygen concentrations measured above the operating deck. Mole-averaging of the oxygen and hydrogen concentrations at 3.7 minutes gives results very close to the measured values at 30 minutes. The mixing may have been enhanced by the hot structures (at 373 K) as opposed to the cold structures in the CES experiments which promoted thermal stratification. The average gas concentrations (dry-basis) in the Surtsey vessel were 0.0 mole %
steam,77.9 mole % nitrogen,' 17.2 mole % oxygen, and 3.1 mole % hydrogen. The posttest wet-basis gas concentrations were 49.2 mole % steam,39.6 mole % nitrogen,8.7 male % oxygen, and 1.6 mole % hydrogen. The posttest moles of O and H (and other noncondensibles) along
]
2 2
with the posttest steam fraction were computed using the nitrogen-ratio method. The hydrogen production amounts are based on the 30-min. gas grab sample data. Hydrogen moles increased by 60 and oxygen moles decreased by 65. Therefore,191 g moles of hydrogen were produced and 130 g moles of hydrogen were burned in CE-1.
CE-2 Gas grab samples used to measure the vessel atmospheric composition were taken at a i
dome penetration, at two level 6 port (operating deck) penetrations, and also in the basement and in the refueling canal (using extension lines). All samples were taken following a 30-s line purge. The times of the samples were set for background,15 s,30 s,2 min., and 30 min.
However, a problem concerning the burst diaphragm failure time also affected the posttest gas grab sample times. The operator had to manually obtain the gas grab samples; consequently the sample times were 120 s,211 s,5.7 min., and 30 min. Mixing fans attached to the underside of the Surtsey vessel upper head were operated prior to taking the background samples and all of the posttest samples.
- =
Experimental Results The pretest or background noncondensible gas fraction could be calculated based on the gas and steam additions to the Surtsey vessel during the charging process. Venting the vessel (to 0.0896 MPa at 278 K) prior to the steam insertion placed 1939 g moles of air inside the Surtsey vessel. About 129 g moles of hydrogen gas was added. Immediately before the thermite ignition, the total moles of noncondensible gas (air and hydrogen) and steam was 3506 g moles.
Therefore (assuming no leakage), the background noncondensible gas fraction was 0.59.
Table 12 shows the results of gas mass spectroscopy analyses performed by SNL.
Multiplying the concentrations of the background gas grab semple measurement by the background noncondensible gas fraction yields the wet-basis gas concentrations.
The background wet-basis gas concentrations inside the Surtsey vessel were 41.0 mole % steam,43.2 mole % nitrogen,11.6 mole % oxygen, and 3.7 mole % hydrogen.
The posttest moles of 0 and H (and other noncondensibles) along with the posttest 2
2 steam fraction was computed using the nitrogen-ratic, method. Thirty minutes after the HPME, the samples lines were purged for 1 min., and then two samples were taken at the level 6 port.
i The mixing fans were turned on prior to the taking of the 30-min. gas grab samples. The average gas concentrations (dry-basis) in the Surtsey vessel were 0.0 mole % steam, 76.3 mole %
nitrogen,15.4 mole % oxygen, and 5.9 mole % hydrogen.
The posttest wet-basis gas concentrations were 47.3 mole % steam,40.3 mole % nitrogen,8.1 mole % oxygen, and 3.1 mole % hydrogen.
Table 15 gives the results for the amounts of hydrogen produced and burned. The hydrogen production amounts are based on the 30-min. gas grab sample data. Hydrogen moles
]
decreased by 12 and oxygen moles decreased by 102. Therefore,191 g moles of hydrogen were
)
produced and 202 g moles of hydrogen were burned in CE-2.
CE-3 Gas grab samples used to measure the vessel atmospheric composition were taken at a dome penetration, at two level 6 port (operating deck) penetrations, and also in the basement and in the refueling canal (using extension lines). The times of the samples were background,15 s, 30 s,2 min., and 30 min. All samples (except the 15 s) were taken following a 30 s line purge.
Mixing fans attached to the underside of the Surtsey vessel upper head were operated prior to taking the background samples and the 2-min. and the 30-min. posttest samples.
The pretest or background noncondensible gas fraction could be calculated based on the gas and steam additions to the Surtsey vessel during the charging process. Venting the vessel (to 0.091 MPa at 284 K) prior to the steam addition placed 1934 g moles of air inside the Surtsey vessel.
About 139 g moles of hydrogen gas was added. Immediately before the thermite ignition, the total moles of noncondensible gas (air and hydrogen) and steam was 3700 g moles. Therefore (assuming no leakage), the background noncondensible gas fraction was 0.56.
Experimental Results Table 13 shows the results of gas mass spectroscopy analyses performed by SNL.
Multiplying the concentrations of the background gas grab sample measurement by the background noncondensible gas fraction yields the wet-basis gas concentrations.
The background wet-basis gas concentrations inside the Surtscy vessel were 43.6 mole % steam,41.1 mole % nitrogen,11.0 mole % oxygen, and 3.8 mole % hydrogen.
The posttest moles of 0 and H (and other noncondensibles) along with the posttest 2
2 steam fraction were computed using the nitrogen-ratio method. Thirty minutes after the HPME, the samples lines were purged for 1 min., and then two samples were taken at the level 6 port.
The mixing fans were turned on prior to the taking of the 30-min. gas grab samples. The average gas concentrations (dry-basis) in the Surtsey vessel were 0.0 mole % steam, 80.2 mole %
nitrogen,13.0 mole % oxygen, and 4.5 mole % hydrogen.
The posttest wet-basis gas concentrations were 48.9 mole % steam,41.0 mole % nitrogen, 6.6 mole % oxygen, and 2.3 mole % hydrogen. Hydrogen moles decreased by 54 and oxygen moles decreased by 161.
Therefore,269 g moles of hydrogen were produced and 323 g moles of hydrogen were burned in CE-3.
CE-4 Gas grab samples used to measure the vessel atmospheric composition were taken at a dome penetration, at two level 6 port (operating deck) penetrations, and also in the basement and in the refueling canal (using extension lines). The times of the samples were background,15 s, 30 s,2 min., and 30 min. All samples (except the 15 s) were taken following a 30-s line purge.
Mixing fans attached to the underside of the Surtsey vessel upper head were operated prior to taking the background samples and the 30-min. posttest samples.
The pretest or background noncondensible gas fraction could be calculated based on the gas and steam additions to the Surtsey vessel during the charging process. Venting the vessel (to 0.092 MPa at 288 K) prior to the steam addition placed 1931 g. moles of air inside the Surtsey vessel.
About 149 g moles of hydrogen gas was added. Immediately before the thermite ignition, the total moles of noncondensible gas (air and hydrogen) and steam was 3730 g. moles. Therefore (assuming no leakage), the background noncondensible gas fraction was 0.558. This method yields background wet-basis gas concentrations inside the Surtsey vessel of 44.2 mole % steam, 51.8 mole % air (40.4 mole % nitrogen,10.9 mole % oxygen), and 4.0 mole % hydrogen. On a dry basis, the concentrations were 92.8 mole % air and 7.2 mole % hydrogen.
Table 14 shows the results of dry-basis gas mass spectroscopy analyses performed by SNL. Multiplying the concentrations of the background gas grab sample measurement by the background noncondensible gas fraction yields wet-basis gas concentrations. The background wet-basis gas concentrations inside the Surtsey vessel determined from mass spectroscopy were 44.8 mole % steam,40.6 mole % nitrogen,10.3 mole % oxygen, and 3.7 mole % hydrogen.
The posttest moles of 0 and H along with the posttest steam fraction were computed 2
2 using the nitrogen ratio method. Thirty minutes after the HPME, the samples lines were purged NUREG/CR-6469 38
Experimental Results for one minute, and then two samples were taken at the level 6 port. The mixing fans were tumed on prior to taking the 30-min. gas grab samples. The average gas concentrations (dry-basis) in the Surtsey vessel were 0.0 mole % steam,77.8 mole % nitrogen,14.3 mole % oxygen,
- and 5.6 mole % hydrogen. The posttest wet-basis gas concentrations were 51.8 mole % steam, 37.5 mole % nitrogen,6.9 mole % oxygen, and 2.7 mole % hydrogen.
Table 15 gives the results for the amounts of hydrogen produced and burned based on the nitrogen-ratio method. The hydrogen production amounts are based on the differences between the background concentration data (from gas addition measurements) and the 30-min. gas grab sample data. Hydrogen moles decreased by 41 and oxygen moles decreased by 128. Therefore, 215 g moles of hydrogen were produced and 256 g moles of hydrogen were burned in CE-4.
Table 6. Debris mass balance in kg for the CES and CE experivients
[
Location CES-1 CES-2 CES-3 CE-1 CE-2 CE-3 CE-4 8.l Q
Basement 6.990 -
7.100 7.005 6.870 6.215 5.215 4.020
- n Missile shield (lower combined 2.070 1.050 0.840 1.760 1.910 1.640 E
T surface) with refueling g
canal
{
Refueling canal '
7.515 3.195 3.000 1.860 2.900 4.575 1.840
=
Operating deck (including 6.925 4.050 3.610 4.210 4.135 4.885 3370 top of missile shield) '
Surtsey vessel (dome, 12.025 16.285 18320 18.585 not recovered not recovered not recovered walls, and floor) '
(I5.71)2 (8.230)2 (11.105)2 Cavity (walls, floor, 5.070 5.670 5380 6.070 7.830 13.735 16.575 annular gap, and RPV walls)
$i Crucible N/A Accumulator (top) '
O.220 combined 0.140 0.145 combined with combined with combined with with refueling operating deck operating deck operating deck canal i
Total Recovered 38.745 38.370 38.505 38.580 38.550 38.550 38.550 (22.840)2 (30320)2 (27.445)2 j
Particle size SMMD (mm) 0.286 0.289 0335 0.274 0306 0.272 0.568 6
Debris considered as outside structures for the mass balance.
2 The amount of debris recovered from the Surtsey vessel was calculated by subtracting the actual amount of debris that was recovered in CE-2, CE-3, and CE-4 from the total recovered debris average (38.55010.156 kg, based on CES-1, CES-2, CES-3, and CE-1).
l
Table 7. Mass balance for the CES and CE experiments MASS BALANCE (kg)
CES-1 CES-2 CES-3 CE-1 CE-2 l CE-3 CE-4 Initial thermite charge, M/(a) 33.22 33.22 33.22 33.22 33.22 33.22 33.22 f
Crucible (b)
N/A Cavity (c) 5.070 5.670 5.380 6.070 7.830 13.735 16.575 Inside structures (d)(d = f-e- c) 6.990 9.170 7.920 7.710 7.975 7.125 5.660 Outside structures (e) '
26.685 23.530 25.070 24.800 22.745 '
17.690 '
16.315 '
Total Recovered (f) 38.745 38.370 38.505 38.580 38.550*
38.550' 38.550' Recovery fraction, f__, = f7a 1.17 1.16 1.16 1.16 1.16 1.16 1.16 THERMITE TRANSPORT FRACTIONS Ejected into cavity, f% = 1 - b/a (only if < 1) 1.0 1.0 1.0 1.0 1.0 1.0 1.0 Dispersed from cavity, fu = (de)/(c+d+e) 0.869 0.852 0.860 0.843 0.797 0.644 0.570 g
Transported outside subcompartment, 0.792 0.720 0.760 0.763 0.740 0.713 0.742 f
= e/(d+c)
Tran., ported to dome, f,,,, = f,
- f
- fw 0.688 0.613 0.654 0.653 0.590 0.459 0.423
'Ihermite transported to dome, M = M/
- f.,,,
22.86 20.38 21.73 21.37 19.60 15.25 14.05
'Ihe amount of debris found outside structures assumes that the average total recovered debris mass is 38.550 kg. See Table 6 for details.
,Y b.
B Q
e.
n Y
W
?
e if i
l
. _ ~
Experimental Results Table 8. Gas concentration measured in the CES-1 experiment Location Label Start Time Species
- Duration (mole %)
N O
H CO' CO 2
2 2
2 1
G1-B
-2 m -+ 10 s 99.1 0.2 0.0 0.0 (Level 6)
G2-B
-2 m -+ 10 s 99.4 0.2 0.0 0.1 G3-2m 2 m -+ 10 s 96.4 0.1 2.9 t1 G4-2m 2 m --> 10 s 96.2 0.1 2.9 0.'
2 G5-30m 30 m
- 10 s 97.2 0.1 2.2 0.1 G6-30m 30 m -+ 10 s 97.0 0.1 2.4 0.1 Background Mean 99.3 0.20 0.00 0.05 Background Standard Deviation
- 0.2 0.00
- 0.0
- 0.07 Posttest Mean (2 m) 97.1 0.10 2.9 0.10 Posttest Standard Deviation
- 0.1 0.00
- 0.0
+0.00 Posttest Mean (30 m) 96.3 0.10 2.3 0.10 Posttest Standard Deviation
- 0.14
- 0.00 0.1 0.00 CO analyses were not performed due to equipment failure.
2 Mixing fans were not tumed on prior to taking the 30-min. samples.
j.
Experimental Resdits Table 9. Gas concentrations measured in the CES-2 experiment Time Label Location Species (mole %)
N2 O2 H2 CO2 CO Argon background 1-box 5 2 dome 2-boxl ops deck 99.5 0.5 0.1 0.0 0.0 0.0 3-boxl ops deck 99.6 0.5 0.1 0.0 0.0 0.0 4-box 4 basement 99.8 0.2 0.1 0.0 0.0 0.0 15 s 5-box 5 dome 94.0 0.1 5.9 0.1 0.5 0.0 6-box 2 ops deck 93.7 0.4 5.9 0.1 0.4 0.0 7-box 2 ops deck 93.6 0.2 6.0 0.1 0.5 0.1 8-box 4 I basement 9-box 2 refuel canal 95.0 0.4 4.5 0.1 0.1 0.0 30 s 10-box 5 dome 94.2 0.1 5.7 0.1 0.5 0.0 11-box 3 ops deck 2
ops deck 93.8 0.5 5.7 0.1 0.5 0.0 12-box 3 basement 13-box 4 refuel canal 95.8 0.2 4.0 0.1 0.3 0.0 1
14-box 2 2 min 15-box 5 dome 94.7 0.1 5.1 0.1 0.4 0.0 16-boxl ops deck 1
ops deck 17-boxl basement 1
18-box 4 1
30 min 3 19-box 3 ops deck 96.0 0.3 4.0 0.1 0.3 0.0 20-box 3 i ops deck 96.7 0.2 3.3 0.1 0.2 0.0 background mean i std. dev.
99.610.2 0.410.2 0.110.0 0.010.0 0.010.0 0.010.0 15 s mean i std. dev.
94.li0.6 0.3i0.2 5.610.7 0.110.0 0.4i0.2 0.010.1 30 a mean i std. dev.
94.6il.1 0.3i0.0 5.70.0 0.110.0 0.410.1 0.010.0 j
2 niin mean i std. dev.
94.71n/a 0.lin/a 5.lin/a 0.lin/a 0.4tn/a 0.0in/a 30 min mean i std. dev.
96.410.5 0.310.1 3.710.5 0.110.0 0.310.1 0.0i0.0
' Sample stations lost electrical power.
2 Bottles leaked.
' Mixing fans were turned on prior to taking the 30-min. samples.
1 43 NUREG/CR-6469
Experimental Results Table 10. Gas concentrations measured in the CES-3 experiment Species (mole %)
Time Label Location N2 O2 H2 CO2 CO Argon background 1-box 5 dome 99.4 0.0 0.2 0.1 0.2 0.0 2-boxi ops deck 99.5 0.2 0.1 0.2 0.4 0.0 3-boxl ops deck 98.6 0.6 0.4 0.1 0.2 0.0 4-box 4 basement 98.8 0.1 0.7 0.1 0.0 0.0 15 s 5 box 5 dome 93.5 0.1 6.0 0.1 0.4 0.0 6-box 2 ops deck 93.8 0.1 5.6 0.1 0.4 0.1 7 box 2 ops deck 93.7 0.0 5.7 0.1 0.3 0.1 8-box 4 basement 96.5 0.2 2.9 0.1 0.0 0.0 9-box 2 refuel canal 96.5 0.0 4.4 0.0 0.3 0.1 30s 10-box 5 dome 94.2 0.0 5.7 0.1 0.4 0.0 11-box 3 ops deck 94.0 0.7 5.0 0.1 0.2 0.1 12-box 3 ops deck 94.7 0.0 5.2 0.1 0.2 0.0 13-box 4 basement 96.6 0.5 2.8 0.2 0.1 0.0 14-box 2 refuel canal 95.5 0.1 4.0 0.1 0.3 0.0 2 min 15-box 5 dome 94.1 0.0 5.4 0.2 0.3 0.0 16-boxl ops deck 94.3 0.1 5.1 0.1 0.4 0.1 17-boxl ops deck 94.0 0.0 5.0 0.3 0.3 0.0 18-box 4 basement 96.6 0.1 3.3 0.0 0.2 0.0 30 min 3 19-box 3 ops deck 95.9 0.0 3.6 0.1 0.2 0.1 20-box 3 ops deck 95.9 0.0 3.7 0.0 0.1 0.0 background mean std. dev. I 99.110.4 0.210.3 0.410.3 0.110.1 0.210.2 0.00.0 15 s mcan std. dev. 2 93.710.2 0.120.1 5.810.2 0.Ii0.0 0.410.1 0.1i0.1 30 s mean i std. dev. 2 94.310.4 0.210.4 5.310.4 0.110.0 0.310.1 0.010.1 2 min mean i std. dev. 2 94,1i0.2 0.010.1 5.210.2 0.210.1 0.310.1 0.0i0.1 30 min mean i std. dev. I 95.910.0 0.010.0 3.710.1 0.li0.1 0.210.1 0.li0.I 1 Averaged over all samples.
2 Average of dome and operating deck samples.
3 The mixing fans were turned on prior to taking the 30-min. samples.
Experimental Results Table 11. Gas concentrations measured in the CE-1 experiment Time Label Location Species (mole %)
N2 O2 H2 CO2 CO Argon background 1-box 5 dome 78.3 20.6 0.0 0.0 0.0 0.9 2-boxl ops deck 78.3 20.7 0.0 0.0 0.0 0.9 3-boxl ops deck 78.0 20.6 0.2 0.0 0.0 0.9 4-box 4 basement 78.2 20.7 0.0 0.0 0.0 1.0 119 s 5-box 5 dome 78.1 16.9 3.3 0.7 0.4 0.9 6-box 2 ops deck 77.7 16.8 3.5 0.8 0.2 0.9 7-box 2 ops deck 77.8 16.7 3.6 0.9 0.3 1.0 8-box 4 basement 78.1 18.6 2.0 0.4 0.1 0.9 9-box 2 refuel canal 77.9 17.8 2.6 0.6 0.0 1.0 134 s 10-box 5 dome 78.9 17.1 2.1 0.9 0.2 1.0 11-box 3 ops deck 77.6 17.3 3.0 0.8 0.2 1.0 12-box 3 ops deck 78.2 17.0 3.0 0.8 0.3 1.0 13-box 4 basement 77.8 19.0 1.8 0.4 0.1 1.0 14-box 2 refuel canal 77.9 17.9 2.5 0.6 0.1 1.0 3.7 min.
15-box 5 dome 77.8 16.9 3.2 0.8 0.3 1.0 16-boxl ops deck 77.9 16.7 3.5 0.8 0.4 1.0 17-boxl ops deck 77.6 16.8 3.7 0.8 0.3 1.0 18-box 4 basement 78.1 18.0 2.3 0.5 0.2 1.0 30 min.3 19-box 3 ops deck 77.9 17.2 3.0 0.7 0.2 1.0 20-box 3 ops deck 77.8 17.2 3.1 0.7 0.2 0.9 Dry-Basis background mean i std. dev.1 78.210.1 20.710.1 0.010.1 0.010.0 0.010.0 0.910.1 119 s mean i std. dev. 2 77.9i0.2 16.810.1 3.5i0.2 0.8i0.1 0.3i0.1 0.910.1 134 s mean i std. dev. 2 78.2 0.7 17.110.2 2.710.5 0.8i0.1 0.210.1 1.010.1 3.7 min. mean i std. dev. 2 77.8i0.2 16.810.1 3.5i0.3 0.810.0 0.3i0.1 1.0i0.0 30 min. mean i std. dev. I 77.910.1 17.210.0 3.lio.1 0.710.0 0.210.0 1.010.1 Wet-Basis j
background mean 1 45.5 12.0 0.0 0.7 30 min. mean 1 39.6 8.7 1.6 1.0 1 Averaged over all samples.
2 Average of dome and operating deck samples.
3 The mixing fans were inadvenently not turned on prior to taking the 30 min. samples.
l Experimental Results l
l Table 12. Gas concentrations measured in the CE-2 experiment Time Label Location Species (mole %)
l N2 O2 H2 CO2 CO Argon background 1-box 5 dome 73.5 18.8 6.3 0.2 0.0 1.0 2-boxl ops deck 73.5 18.9 6.2 0.1 0.0 1.0 3-boxl ops deck 73.7 19.0 6.2 0.1 0.0 1.0 4-box 4 basement 73.4 18.8 6.2 0.0 0.0 1.0 120 s4 5-box 5 dome 76.4 14.2 5.8 1.6 0.0 1.0 6-box 2 ops deck 77.4 14.3 5.9 1.5 0.1 0.9 7-box 2 ops deck 77.0 14.3 5.9 1.4 0.0 0.9 8-box 4 basement 76.1 16.1 5.7 1.0 0.0 1.0 9-box 2 refuel canal 77.0 14.8 5.5 1.3 0.0 0.9 21Is4 10-box 5 dome 76.1 15.4 5.6 1.2 0.0 1.0 11-box 3 ops deck 77.4 15.3 5.2 1.1 0.0 0.9 12-box 3 ops deck 76.4 15.3 6.1 1.1 0.0 0.8 13-box 4 basement 76.1 16.2 4.8 1.0 0.0 1.0 14-box 2 refuel canal 75.8 15.3 6.4 1.0 0.0 0.9 i
5.7 min.4 15-box 5 dome 75.5 15.3 5.8 1.3 0.1 1.0 16-boxl ops deck
)
17-boxl ops deck 76.3 15.6 5.8 1.2 0.0 1.0 18-box 4 basement 75.9 15.6 5.8 1.2 0.0 1.0 30 min.4 19-box 3 ops deck 76.2 15.3 5.8 1.0 0.0 0.9 20-box 3 ops deck 76.2 15.4 6.0 1.0 0.0 1.0 Dry-Basis background mean i std. dev. I 73.610.1 18.910.1 6.2i0.1 0.li0.1 0.010.0 1.010.0 1
120 s mean i std. dev. 2 77.110.3 14.3i0.1 5.910.I 1.510.1 0.010.1 0.910.1 211 s mean i std. dev. 2 76.610.7 15.310.1 5.6i0.5 1.I10.1 0.0i0.0 0.910.1 5.7 min. mean i std. dev. 2,5 75.910.6 15.510.2 5.810.0 1.310.1 0.110.I 1.010.0 30 min. mean i std. dev.1 76.210.0 15.410.1 5.9i0.I 1.0i0.0 0.0i0.0 1.010.1 Wet-Basis background mean 1 43.2 11.6 3.7 0.5 30 min. mean 1 40.3 8.1 3.1 1.4 1 Averaged over all samples.
2 Average of dome and operating deck samples.
3 Sample 16-boxl leaked.
4 Mixing fans were operated prior to taking all posttest samples.
=
Experimcntal Results Table 13. Gas concentrations measured in the CE-3 experiment i
Time Label Location Species (mole %)
N2 O2 H2 CO2 CO Argon background 1-box 5 dome 73.6 18.7 6.7 0.1 0.0 1.0 2-boxl ops deck 73.4 18.7 6.7 0.1 0.0 0.9 3-boxl ops deck 73.4 18.7 6.6 0.1 0.0 1.0 4-box 4 basement 73.4 18.7 6.6 0.1 0.0 0.9 15 s 5 box 54 dome 6-box 2 ops deck 89.2 6.1 0.9 2.6 0.1 1.2 7 box 2 ops deck 89.2 6.2 0.9 2.5 0.1 1.1 8-box 4 basement 77.2 5.1 5.1 0.6 0.1 1.0 9-box 2 refuel canal 77.7 13.2 7.0 1.1 0.2 1.0 30 s 10-box 54 dome 11-box 3 ops deck 88.3 6.4 1.3 2.6 0.1 1.1 12-box 3 ops deck 87.5 6.3 1.5 2.5 0.1 1.1 13-box 4 basement 78.2 16.4 3.7 0.6 0.0 1.0 14-box 2 refuel canal 80.5 12.1 5.0 1.3 0.1 1.0 2 min.
15-box 5 dome 80.8 12.6 4.3 1.2 0.1 1.1 16-box 14 ops deck 17-boxl ops deck 80.9 12.4 4.3 1.3 0.2 1.1 18-box 4 basement 79.5 13.8 4.4 1.0 0.1 1.0 30 min.
19-box 3 ops deck 80.1 13.0 4.5 1.1 0.1 1.1 20-box 3 ops deck 80.2 13.0 4.5 1.1 0.1 1.0 Dry-Basis background mean i std. dev. I 73.510.I 18.710.0 6.710.1 0.110.0 0.0i0.0 1.0i0.0 15 s mean i std. dev. 2 89.210.0 6.210.1 0.910.0 2.6i0.0 0.li0.0 1.210.I 30 s mean i std. dev. 2 87.910.6 6.410.1 1.410.1 2.610.1 0.110.0 1.110.0 2 min, mean i std. dev. J 80.410.8 12.910.8 4.310.1 1.210.2 0.110.1 1.110.1 30 min. mean i std. dev I 80.210.1 13.0f0.0 4.510.0 1.li0.0 0.I10.0 1.1i0.I Wet-Basis background mean 1 41.1 11.0 3.8 0.5 30 min. mean 1 41.0 6.6 2.3 1.2 1 Averaged over all samples.
2 Average of dome and operating deck samples.
3 Average of all samples (mixing fans were turned on at I min.)
4 Samples 5-box 5 and 10-box 5 did not pull, sample 16-boxl leaked.
1 Experimental Results Table 14. Gas concentrations measured in the CE-4 experiment Time Label Location Species (mole %)
N2 O2 H2 CO2 CO Argon background 1-box 5 dome 72.7 18.6 6.9 0.0 0.0 0.8 2-boxl ops deck 72.9 18.5 6.8 0.7 0.0 0.8 3 boxl ops deck 72.0 18.3 6.8 0.3 0.0 0.8 4-box 4 basement 73.1 18.6 6.9 0.1 0.0 0.8 15 s 5-box 5 dome 82.7 9.7 3.4 2.2 0.2 0.9 6-box 2 ops deck 82.8 10.0 4.0 2.2 0.1 1.0 7-box 2 ops deck 81.0 10.2 3.9 2.0 0.1 1.3 8-box 4 basement 75.0 16.4 5.9 0.6 0.0 0.8 9-box 2 refuel canal 81.0 11.0 4.8 2.0 0.2 1.0 30 s 10-box 5 dome 83.1 9.8 3.9 2.0 0.2 0.9 11-box 3 ops deck 82.3 10.8 3.8 1.9 0.1 1.1 12-box 3 ops deck 81.2 10.9 4.4 1.9 0.1 1.0 13-box 4 basement 75.9 16.8 4.4 0.5 0.0 0.9 14-box 2 refuel canal 80.0 12.3 5.0 1.7 0.1 1.0 2 min.
15-box 5 dome 81.1 11.2 4.2 1.8 0.1 0.9 16-box 14 ops deck 17-box 14 ops deck 18-box 4 basement 75.7 15.8 6.2 0.8 0.0 0.9 30 min.
19-box 3 ops deck 77.7 14.2 5.5 1.1 0.1 1.1 20-box 3 ops deck 77.8 14.3 5.6 1.2 0.0 1.0 Dry-Basis background mean i std. dev. I 72.710.5 18.510.1 6.7i0.1 0.3i0.3 0.010.0 0.810.0 15 s mean i std. dev. 2 82.2il.0 10.010 3 3.810.3 2.1 0.1 0.110.
1,1i0.2 30 s mean i std. dev. 2 82.2il.0 10.510.6 4.0i0.3 1.9i0.1 0.110.1 1.010.1 2 min. mean i std. dev. 3 31.1 11.2 4.2 1.8 0.1 0.9 30 min, mean i std. dev.1 77.810.1 14.310.1 5.610.1 1.210.1 0.110.I 1.110.1 Wet-Basis background mean 1 40.5 10.9 4.0 0.5 30 min. mean 1 37.5 6.9 2.7 1.2 1 Averaged over all samples (mixing fans were turned on prior to background and 30 min. samples).
2 Average of dome and operating deck samples.
3 Dome sample.
4 Melt impacted internal lights and tripped an electrical circuit which prevented mixing fans from j
starting prior to t = 2 min. and also prevented the 2-min. ops deck samples (16-boxl and 17-box 1) from j
pulling.
i l
NUREG/CR-6469 48 1
Table 15. Results from the CES and CE DCH experiments CES-1 CES-2 CES-3 CE-1 CE-2 CE-3 CE-4 Inisimi vessel rin._.m (MPa) 0.2024 0.2027 0.2021 02103 0.2195 0.2360 0.2353 Initial vessel temperature (K) 298 316 305 378 376 384 380
~6 s aft:r thermite ignition AP (MPa) 0.015 0.014 0.014 0.059 0.121 0.137 0.125 Dome gas tempo.h-w mcrease (K) 16 15 15 129 224 257 257 Su' wrur-huent gas temperature merease (K) 0 1
1 8
13 16 12 u
Vessel average gas temperature merease (K) 9 7
8 67 109 122 120 Mass thermite houspuited to dome (kg) 1o.92 2038 21.73 2137 19.60 153 14.1 hauite houarnt fre (to dome) 0.69 0.61 0.65 0.65 0.59 0.46 0.42 Vessel rswc prior to HPME(MPa) 0.2158 0.2156 0.2157 0.2558 0.2619 0.2999 0.2 % 8 Vesseltempa. hue prior to HPME (K) 307 324 314 440 456 488 479
~1 s after HPME AP (MPa) 0.234 0316 0.293 0.242 0.208 0.253 0.217 Dome gas temperature mcrease (K) 404 503 425 152 35 228 202 Subh- ;-Luent gas temperature mcrease (K) 38 68 68 94 54 126 77 Vessel average gas temperature increase (K) 175 221 201 121 48 169 I16 8
Cavity AP (MPa) 3.08 3.28 2.78 1.74 0.41 0.72 0.69 Initial gas composition Dry Wet Dry Wet Dry Wet Dry Wet in the Surtsey vessel Steam 0.00 0.0 0.0 0.0 41.9 0.0 41.0 0.0 43.6 0.0 44 2 (mole %)
N 993 99.6 99.1 78.1 45.4 73.2 43.2 72.9 41.1 72.5 40.4 2
O 0.2 0.4 0.2 20.9 12.2 19.6 11.6 19.6 11.0 19.4 10.9 2
H 03 0.1 0.4 0.0 0.0 6.2 3.7 6.7 3.8 7.2 4.0 2
Posttest gas composition Dry Wet Dry Wet Dry Wet Dry Wet in the Surtsey vessel Steam 0.0 0.0 0.0 0.0 493 0.0 47.2 0.0 48.9 0.0 51.2 i
(mole %)
N
%3
%.4 95.9 77.9 39.5 76.2 403 80.2 41.0 77.8 37.5 j
2 O
0.1 03 0.0 17.2 8.7 15.4 8.1 13.0 6.6 14 3 6.9 2
H 23 3.7 3.7 3.1 1.6 5.9 3.1 4.5 23 5.6 2.7 y
2 Moles of preexisting combustible gas (g-moles) 0 0
0 0
129 139 149 8j-g g
Moles of combustible gas proA_W (g-moles) 75 141 145 203 191 269 215 Moles of combustible gas burned (g-moles) 0 0
0 144 202 323 256 a
a8 Net difTerence between production and combustion 0
0 0
60
-12
-54
-41 E
Y Fraction of available combustible gas that bumed 0
0 0
0.71 0.63 0.79 0.70
- xs 3
(Nd(N,,, + N,))
5?
Experimental Results 10
/
--*- accumulator
-e-burst diaphragm
+ cavity 9_
--5 8-7-
1 g
a.
6-2 i
g a
e 5-n.
0
- s 3g 4-i 3-2-
f 1-j
^m ts-s--s--s- + u
.h - O
,4 s
s n
-1 0
1 2
3 Time (s) 4 4
l 1
Figure 8. Blowdown history of the CES-1 experiment.
i i
4 NUREG/CR-6469 50
Experimental Results 10 l
-*- accumulator
-*-- accumulator 4 burstdiaphragm 9-
-v-cavity 8-7-
E a.
6-E E
e 5-c.
o5
~
4-3-
2-l 1-L
._______J 0
i i
-1 0
1 2
3 Time (s)
Figure 9. Blowdown history of the CES-2 experiment.
Experimental Results i
10
--+-- accumulator
-e-- burst diaphragm
+ cavity g_
8-7_
i l
g n.
6-1 2
=
e 5-c.
2 o
~o E
4-4 i
3-l
)
2-i 4
i 1-Wa c r-0 i
i
-1 0
1 2
3 4
5 Time (s)
Figure 10. Blowdown history of the CES-3 experiment.
l Experimental Results 10
-+-- accumulator
-*-- burst diaphragm 1
+ cavity 9-8-
4 4
7-o.
6-
- s h
5-c.
4-0 3-i 2-
\\
1-
.I O
i
-1 0
1 2
3 4
5 Time (s)
Figure 11. Blowdown history of the CE-1 experiment.
Experimental Results 10
-*- accumulator
-*- burst diaphragm n cavity l
9-i i
8-l 7-l k
6-E 2
5-c.
e
_5 4-3-
2-1-
---n_w 0
-1 0
1 2
3 4
5 Time (s)
Figure 12. Blowdown history of the CE-2 experiment.
l NUREG/CR-6469 54
Experimental Results l
l 10
-+-- accumulator
-*- burst diaphragm cavity 8-l 7-n.
6-3, e
e g
4-3-
2-1-
- --a a N -6
- w -6_6 6" 6. # 6_
0 i
-1 0
1 2
3 4
5 Time (s) l Figure 13. Blowdown history of the CE-3 experiment.
l 55 NUREG/CR-6469
Experimental Results 5
-*- accumulator
-e-burst diaphragm
.1 cavity 4_
[3-E I
a e
n.
e5)2--
n n
g J
1
- + 6-- 6-66--6c6-6 -44 -- 6 _%% h.
e6.-
0 i
i i
-1 0
1 2
3 4
5 6
Time (s)
Figure 14. Blowdown history of the CE-4 experiment.
4 NUREG/CR-6469 56
Experimental Results 4
'i....i
...i....i...
i g
--o-CES-1
-o-CES-2 4 CES-3
-*- CE-1 8-
+ CE-2
-* CE-3
-+ CE-4 7-6-
)
?c.
\\
2 u
~5-2 g
4
~
k 4_
4
~
.%,No g o
3-
\\$
2-k j_
_: g
"IU c
0 i
0 1
2 3
4 5
Time (s)
Figure 15. Accumulator pressure in the CE DCH experiments.
Experimental Rtsults 650
-*- CES-1
-*- CES-2
-*- CES-3
-v-CE-1 600 -
+ CE 2
(
- CE-3
-mm m,
-+ CE-4 550 -
+
1
-W
+
+
500 -
+
525 ll 450 -
e
~g
- -ee - -
_?}-
400 -
350 -
300 -
2 2 --:
250 i
i i
i i-i
-1 0
1 2
3 4
5 6
Time (s)
Figure 16. Accumulator gas temperatures in the CE DCH experiments.
Experimental Results l
l 0.50 0.45 -
0.40 -
)
l l
S.
E 0.35 -
l 2
peak pressure increase due to HPME = 0.234 MPa
=
i n.
22 0 i
.30 -.
1 0.25 -
1 i
l l
peak pressure l[1Nease prior to HPUb = b.015 Mka
]"' "
0.20 -
t, = -13 s 0.15
-60 0
60 120 180 240 300 360 420 480 540 600 Time (s)
I l
l Figure 17. Vessel pressure in the CES-1 experiment from -66 to 600 s.
1 59 NUREG/CR-6469
1 Experimental Results l
O.50 peak pressure increase due to HPME = 0.234 MPa 0.45 -
l 0.40 -
7 k
- 0.35 -
t 1
0.30 -
t, = -13 s 0.25 -
.......... y 0.20 -
peak pressure increase prior to HPME = 0.015 MPa 0.15
-20
-10 0
10 20 30 Time (s)
Figure 18. Vessel pressure in the CES-1 experiment from -20 to 30 s.
1 1
a.
2 k'.
A Experimental Results 0.50 0.45 -
0.40 -
g E
- 0.35 -
23i n.
0 peak pressure increase due to HPME = 0.234 MPa 1
30 -
0.25 -
0.20 -
0.15 i
i i
i i
i i
i 0
1 2
3 4
5 6
7 8
9 10 Time (s)
Figure 19. Vessel pressure in the CES-1 experiment from 0 to 10 s.
Experimental Results 0.55 a
0.50 -
0.45 -
. g 0.40 -
n.
E e3 peak pressure increase due to HPME = 0.316 MPa '
O.35 -
n.
22 O
E< 0.30 -
0.25 -
v..
Peak pressure increase prior to HPME = 0.014 MPa 0.20 -
t
= -13 s p
0.15 i
-60 0
60 120 180 240 300 360 420 480 540 600 Time (s)
Figure 20. Vessel pressure in the CES-2 experiment from -60 to 600 s.
Experimental Results
/
0.55 0.50 -
0.45 -
peak pressure i 1 crease d to HPME = 0.316 MPa g 0.40 -
e3 0.35 -
n.
2 s
4 0.30 -
Q = -13 s 0.25 -
....... y 0.20 -
peak pressure increase prior to HPME = 0.014 MPa 0.15
-20
-10 0
10 20 30 Time (s)
Figure 21. Vessel pressure in the CES-2 experiment from -20 to 30 s.
Experimental Results 0.55 1
0.50 -
l 0.45 -
\\
1 g 0.40 -
E
~
2 3
0.35 -
n.
32 peak pressure increase due to HPME = 0.316 MPa 4 0.30 -
0.25 -
0.20 -
0.15 i
.i i
0 1
2 3
4 5
6 7
8 9
10 Time (s)
Figure 22. Vessel pressure in the CES-2 experiment from 0 to 10 s.
Experim:ntal Results 1
0.55 0.50 -
l 0.45 -
g 0.40 -
4 a.
I E
w t
peak pressure increase due to HPME = 0.293 MPa
=
0.35 -
a.
1 4 0.30 -
0.25 -
- ~.............
/..
.. Peak pressure increase prior to HPME = 0.014 MPa 0.20 -
t,= -13 s 0.15
-60 0
60 120 180 240 300 360 420 480 540 600 Time (s)
Figure 23. Vessel pressure in the CES-3 experiment from -60 to 600 s.
=
Experimental Results 0.55 peak pressure increase due to HPME = 0.293 MPa 0.50 -
0.45 -
9 0.40 -
n.
E e3 0.35 -
i]
4 0.30 -
i t
Q = -13 s 0.25 -
.... y 0.20 -
peak pressure increase prior to HPME = 0.014 MPa -
0.15 i
-20
-10 0
10 20 30 Time (s) l 1
i Figure 24. Vessel pressure in the CES-3 experiment from -20 to 30 s.
j l
4 NUREG/CR-6469 66
Experimental Results i
i 1
0.55 1
1 f
j 0.50 -
0.45 -
l l
t 9 0.40 -
n.3 2s 0.35 -
n.
eI peak pressure increase due to HPME = 0.293 MPa 4 0.30 -
0.25 -
~~
0.20 -
0.15 i
i i
i i
i i
i i
i 0
1 2
3 4
5 6
7 8
9 10 Time (s)
Figure 25. Vessel pressure in the CES-3 experiment from 0 to 10 s.
i l
Experimental Results 1
0.55 0.50 -
0.45 -
g 0.40 -
1 E.,
e=
peak pressure increase due to HPME = 0.242 MPa 0.35 -
n.
2a 4 0.30 -
0.25 -
peak pressure increase prior to HPME = 0.059 MPa i
\\.....
0.20 -
t,
= -29 s 0.15
-60 0
60 120 180 240 300 360 420 480 540 600 Time (s)
Figure 26. Vessel pressure in the CE-1 experiment from -60 to 600 s.
Experimental Results 0.55 O.50 -
(
0.45 -
g 0.40 t
E3 peak pressure increase due to HPME = 0.242 MPa 4
4 0.30 -
..............m..
0.25 -
peak pressure increase prior to HPME = 0.059 MPa 0.20 -
4 t,, = -29 s 0.15
-30
-20
-10 0
10 20 30 Time (s)
J Figure 27. Vessel pressure in the CE-1 experiment from -30 to 30 s.
4 69 NUREG/CR-6469
Experimental Results 0.55 O.50 -
)
1 0.45 -
g 0.40 -
n.E l
23 0.35 -
2 j
peak pressure increase due to HPME = 0.242 MPa
.I4 0.30 -
'M-0.25 -
0.20 -
0.15 i
0 1
2 3
4 5
6 7
8 9
10 Time (s)
Figure 28. Vessel pressure la the CE-1 experiment from 0 to 10 s.
Experimental Results O.50 i
0.45 -
0.40 -
4 1
4 peak pressure increase E
due to HPME = 0.208 MPa M
- 0.35 -
2 3i n.
3 i
2 0.30 -
0.25 -
peak pressure increase prior to HPME = 0.121 MPa
-d s.
j 0.20 -
Q = -45 s 0.15 i
i i
i i
-60 0
60 120 180 240 300 360 420 480 540 600 Time (s)
Figure 29. Vessel pressure in the CE-2 experiment from -60 to 600 s.
s 71 NUREG/CR-6469
=
Experimental Results 0.50 0.45 -
0.40 -
l i
k i
E- 0.35 -
E l
peak pressure increase due to HPME = 0.208 MPa 0.30 -
l peak pressure increase 0.25 -
prior to HPME = 0.121 MPa 1
1
)
1 0.20 -
t, = -45 s 0.15
-50
-40
-30
-20
-10 0
10 20 30 1
Time (s)
Figure 30. Vessel pressure in the CE-2 esperiment from 50 to 30 s.
Experim ntal Results 0.50 0.45 -
0.40 -
7 n.3- 0.35 -
peak pressure increase due to HPME = 0.208 MPa E
I 0.30 -
i...........................................................................
- 0.25 -
0.20 -
0.15 i
i i
i i
i 0
1 2
3 4
5 6
7 8
9 10 Time (s)
Figure 31. Vessel pressure in the CE-2 experiment from 0 to 10 s.
73 NUREG/CR-6469 1
=
Experimental Results 0.60 O.55 -
0.50 -
0.45 -
a g
peak pressure increase due to HPME = 0.253 MPa 30.40 -
3 a.
'! 0.35 -
}<
0.30 -
~-
~-
peak pressure increase pri r to HPME = 0.137 MPa 0.25 -
0.20 -
t,,, = -29 s 0.15 i
-60 0
60 120 180 240 300 360 420 480 540 600 Time (s)
Figure 32. Vessel pressure in the CE-3 experiment from -60 to 600 s-NUREG/CR-6469 74
Experimental Results 0.60 O.55 -
' " " ' " " " ' " " " " ' " " * " ~
~
0.50 -
l 0.45 -
j iii i
n.I i
l
~
2 0.40 -
- s a.
$ 0.35 -
peak pressure increase 3
due to HPME = 0.253 MPa
.8 i
l 0.30 -
peak pressure increase prior to HPME = 0.137 MPa 0.25 -
0.20 -
t,
= -29 s 0.15 i
i
-30
-20
-10 0
10 20 30 Time (s)
Figure 33. Vessel pressure in the CE-3 experiment from -30 to 30 s.
Experimental Resulta i
i 0.60 l
i
.eeeeeeee 5eeeeeeeeeeee e eee eeeeeeeee eeeeeeeeeeeeeeeeeeeeeeeeee e4eeeei l
0.50 -
0.45 -
7c.3 l 0.40 -
peak pressure increase due to HPME = 0.253 MPa I
ne I 093 5 -
}4 0.30 -
i 1
0.'15 -
0.20 -
091 5 i
,.. i i
i i
i i
0 1
2 3
4 5
6 7
8 9
10 Time (s)
Figure 340 Vessel pressure in the CE-3 experiment from 0 to 10 s.
Expcrimental Results 0.55 0.50 -
0.45 -
g 0.40 -
peak pressure increase due to HPME = 0.217 MPa
'.E e=
0.35 -
n.
2a
< 0.30 -
peak pressure increase prior to HPME = 0.125 MPa 0.25 -
w 0.20 -
t o. = -29 s 0.15
-60 0
60 120 180 240 300 360 420 480 540 600 Time (s)
Figure 35. Vessel pressure in the CE-4 experiment from -60 to 600 s.
Experimental Results l
l 0.55 1
1 0.50 -
0.45 -
g 0.40 -
n.
E.
peak pressure increase 3
due to HPME = 0.217 MPa 3 0.35 -
n.
3
- s 14 0.30 -
peak pressure increase prior to HPME = 0.125 MPa 0.20 -
t,
= -29 s 0.15
-30
-20
-10 0
10 20 30
. Time (s)
Figure 36. Vessel pressure in the CE-4 experiment from -30 to 30 s.
Experimental Results 4
0.55 0.50 -
N 0.45 -
g 0.40 -
n.
3 peak pressure increase due to HPME = 0.217 MPa 23 f 0.35 -
i n.
b.30-0 J..................
l 0.25 -
0.20 -
0.15 i
i i
i i
i i
i 0
1 2
3 4
5 6
7 8
9 10 Time (s)
Figure 37. Vessel pressure in the CE-4 experiment from 0 to 10 s.
i Experimental Results 1
i 0.60
--o-- CES-1
--o-- CES-2
+ CES-3
--+- CE-1 i
0.55 -
1
--*-- C E-2 4
--o-- CE-3
--o-- CE 4
}
0.50 -
)
i j
0.45 -
h 7n.
y
- . 0.40 -
s 4
e h
i j
n.
1 e
j 5 0.35 -
o e
4
.o 4
'N 0 30 -
0.25 -
k
^ "~ ~
A 0.20 -
L --
0.15
-50 0
50 100 l
Time (s) i Figure 38. Vessel pressure In the CE DCH experiments.
t i
Experimental Results I
0.60
-o-- CES-1
- o-CES-2
+ CES-3
+ CE-1 O.55 -
CE-2
-o-CE-3
-*- CE-4 4
l 0.50 -
~
i j
0.45 -
T 2
l 2 0.40 -
l
- s i
n.
$ 0.35 -
s
.I i
4 i
0.30
= - -
1 0.25 -
H D
0.20 -
0.15 i
i i
i i
i i
-1 0
1 2
3 4
5 6
7 8
9 10 Time (s)
Figure 39. Vessel pressure during the HPME in the CE DCH experiments.
Experimental Results i
i i
i 4
-*- cavity
--*-- vessel i
i #'
3-
<>i' 4
<i
<i<
g 0-
<i E,
d f <<,
e y
l2-a-
3 a
'5
.8
<i
< I 1-g
['
E 0
-0,1 0.0 0.1 0.2 0.3 0.4 Time (s)
Figure 40. Cavity pressure and vessel pressure in the CES-1 experiment.
Experimental Results 4
-*- cavity
-*- vessel i [,
i i,
o<>
I>
3-t>
i' i,
7 i i n.
1-i,
i, il ii rrr- -
0
-0.1 0.0 0.1 0.2 0.3 0.4 Time (s)
Figure 41. Cavity pressure and vessel pressure in the CES-2 experiment.
Experimental Results f
1 i
i i
i i
4
--+-- cavity
--+-- vessel k
t i
4, 3-is 4
,I I
I q
4 1
+
n.
l i
E i
e j
m f 2-a.
3 2
oj i
o 4
<i Ii 1-
}
qi i
i l
I l
il W
O
-0.1 0.0 0.1 0.2 0.3 0.4 Time (s) i Figure 42. Cavity pressure and vessel pressure in the CES-3 experiment.
i NUREG/CR-6469 84 1
Experimental Results i
1 2.0
---*- cavity
-e-vessel i
i 1.5 -
I I
di I
e a.
E e=
l1.0-4 n.
e 33
.3 4
ii 4i 4
it 0.5 -
Y, 0.0 i
-0.1 0.0 0.1 0.2 0.3 0.4 Time (s)
Figure 43. Cavity pressure and vessel pressure in the CE-1 experiment.
Experimental Results 2.0
--*- cavity
--+-- vessel 1
1.5 -
l 1
7 k
e 3el1.0-n.
o
.8<
0.5 -
a.
-~i i
00 i
-0.1 0.0 0.1 0.2 0.3 0.4 Time (s)
Figure 44. Cavity pressure and vessel pressure in the CE-2 experiment.
.~
t l
I l
Experimental Results l
2.0
--+-- cavity
--*- vessel 1
1.5 -
1 I
7 e
3
\\
el1.0-o.
33 o
E l,'
iI I
(l i
- ^ * -
0.5 -
i r-; -
e-
<i il
in 0.0 -
-0.1 0.0 0.1 0.2 0.3 0.4 Time (s)
Figure 45. Cavity pressure and vessel pressure in the CE-3 experiment.
Experimental Results 2.0
--+-- cavity
-*- vessel 1.5 -
R E
~
1.0 -
a.
e 1ll o
E<
0.5 -
~~
0.0
-0.1 0.0 0.1 0.2 0.3 0.4 Time (s)
Figure 46. Cavity pressure and vessel pressure in the CE-4 experiment.
2 NUREG/CR-6469 88
. ~..
Experimenail Results 4
4
-o-- CES-1 4
-o-CES-2
+ CES-3 i
CE-1 CE-2
-<>- CE 3
--*- C E-4 i
3-i 4
b o,
5.
e 3*
l2-1 n.
i
(
'5
.E 1
o O
1-t.
M i
-r
)
0 i
i i
i i
-0.05 0.00 0.05 0.10 0.15 0.20 0.25 0.30
]
Time (s) 1 j
Figure 47. Cavity pressure in the CE DCH experiments.
i 89 NUREG/CR-6469 4
Experimental Results i
l i
i 800
-+- Dome average
--+- Subcompartment average
+ Vessel average l
700 -
l o
j 600 -
o 11 2a 3 500 -
E.
E l'
400 -
N
{
l 300 1 200
-10 0
10 20 30 40 50 Time (s)
Figure 48. Average gas temperatures in the CES-1 experiment.
l 1
Experimental Results l
1 1000
--*-. Dome average
--+-- Subcompartment average
+ Vessel average 900 -
!l d
a 800 -
o o
700 -
o 2
2o o
3 600 -
E.
E 500 -
4 Jl
_ 2 400 -
~
F lM 300 200 i
-10 0
10 20 30 40 50 Time (s)
Figure 49. Aserage gas temperatures in the CES-2 experiment.
Experimental Results 800
-+- Dome average
-e-Subcompartment average
+ Vessel average 700 -
600 -
2 k
n 3 500 -
8.!
H 400 -
2 ;I : : : :
300 :s 200 i
i i
-10 0
10 20 30 40 50 Time (s)
Figure 50. Average gas temperatures in the CES-3 experiment.
Experimental Results 1
4 1
700
--*- Dome average
-e-Subcompartment average
+ Vesselavetage j
q' 1
e 1
l, i
600 -
o i,
i I
i o
< i
< 1 e
i; i
E o
Ei
<i f F30 -
e E
E o
t
\\
j m _
i 400 -
n l E I
i n i
l i
4 1
300 i
i i
i i
-30
-20
-10 0
10 20 30 l
Time (s)
I i
i
{
Figure 51. Average gas temperatures in the CE-1 experiment.
4 1
)
1 93 NUREG/CR-6469 1
1
Experimental Results 700
--+- Dome average
-+- Subcompartment average Vessel average 1
600 -
2 8
- s E 500 -
8.!
h h
'ji l
4 4
i 400 -
4 lE imusamma 300 i
-50
-40
-30
-20
-10 0
10 20 30 Time (s)
Figure 52. Average gas temperatures in the CE-2 experiment.
l I
Experimental Results i
1 k
900 4
4
--*- Dome average
-e-Subcompartment average
+ Vesselaverage j
)
800 -
ii 1
d, 700 -
d M
400 -
,w,
,wwv Y
Y Y
i i
iIlilllIIIIIIIIIilililillIlifillilIllilIlillifil!!Ill!lilillIll!IllIll!!IlllflIllillif tl!1!I!!!ll!ll!1iilllll!I
(
300 -
1 l
{
l l
l
'I
-50
-40
-30
-20
-10 0
10 Time (s) l l
Figure 55. Comparison of the dome average gas temperatures in the CE crperiments l
during the thermite reaction interval.
l l
r I
l 97 NUREG/CR-6469
Fxperimental Results 1
1 i
f f
900 -
800 -
700 -
E f)
^
E f, 600 -
-e c.
E I
p J
I 500 - -
I O
O 400 -
O CES-1 O CES-2 A CES-3 h
y CE-1 h
+ CE-2
-d 0 CE-3 300 -
e CE4 1
I I
I I
O 1
2 3
4 5
Time (s)
Figure 56. Comparison of the dome average gas temperatures in the CE experiments during the IIPME.
Experimental Results l
700 l
O CES-1 l
0 CES-2 a CES-3 y CE-1 4>
CE-2 O CE-3 e CE-4 l
600 -
L
'l b"
)3 3
5 500 -
I j
4 I
\\
Temperat.ure (K) 1 4
)
4 400 -
v v v if!!!IIil'Il!)I:!!!!!:Iliili!!i'!!I11Il!!!! :fI!!111i:!;P fII!I!IIIIIll:lfl!!!!!IIl!!l!il!Ii!!IlI? H!!!!!!!!!IIH a
300 -
l I
I I
I
-50
-40
-30
-20
-10 0
10
.'is) l l
Figure 57. Comparison of the vessel ma!c aserage gas temperatures in the CE experiments
(
during the thermite reaction intersal.
[
l 99 NUREO/CR-6469
Experimental Results 700 600 -
1 52 2
- s
'd 500 -
8.
E 4
/
o h
r W
400 -
O CES-1 s
Y O CES-2 a CES-3 y CE-1
+ CE-2
.N O CE-3
. CE-4 g
300 0
1 2
3 4
5 Time (s)
Figure 58. Comparison of the vessel mole-average gas temperatures in the CE experiments during the HPME.
NUREG/CR-6469 100
Experimental Results l
7000
- 0.50
- 700
--G-moles
- pressure
--A-temperature
- 650 I
- 0.45
- 600 6000 -
- 0.40
- 550
- 0.35
- 500 2 5
^
e s
E
~
- 0.30 2 2 5000 -
U" y
- 450 f O
- 0.25
- 40C 4000 -
- 0.20
- 350
^ ^ ^ ^ "
- 0.15
,,,,,,e
- 300 3000
- 0.10
- 250
-10 0
10 20 Time (s) i Figure 59. Vessel gas pressure, average remperature, and moles in the CES-1 experiment.
j I
l 101 NUREG/CR-6469
Experimental Results s
7000
- 0.60
- 700
-*- moles
-m-pressure
-A-- temperature
- 0.55
- 650 ll
- 0.50
- 600 6000 -
- 0.45
- 550
- 0.40 a
- 500 -
E 2 5000 -
- 0.35 2 3
E E
f
- 450 5
- 0.30
- 400
- 0.25
- 350
- 300
- 0.15 3000
- 0.10
- 250 10 0
10 20 Time (s)
Figure 60. Vessel gas pressure, average temperature, and moles in the CES-2 experiment.
NUREG/CR-6469 102
Experimental Results j
i l
8000 h
- 0.55
- 700
] moles
-m-pressure j,
--e-temperature
- 0.50
- 650 i
1 7000 -
- 0.45
-600 R
i Rg S
- 0.40
- 550 a n s
6000 -
- 0.35 2
- 500 2
^
S 2
8 O
v 3
2 e
3 E
E E
U
- 0.30
- 450 $
c.
r i
5000 -
- AAAn - 0.25
- 400 n
gasma
- 0.20
- 350 4000 1
^A4*A#
- 0.15
- 300 3000
' 0.10
- 250 i
-10 0
10 20 Time (s)
Figure 61. Vessel gas pressure, average temperature, and moles in the CES-3 experiment.
103 NUREGr;:R-6469
Experimental Results i
1 4
5 7000
--*- moles
- 0.55
- 700 4
-> pressure
-A-- temperature J
4
- 650 l
- 0.50 6000 -
- 600 l
- 0.45 i
<l
- 550 i
\\
~ 0.40 g 2
\\
E o.
~
T>
2 2
i 2 5000 -
a 8
2
- 500 4
O 2
5
- 0.35 0-F-
- 450
~ 0.30 4000 -
- 400 I
aaaaaA I
d
- 0.25 i
i 3000
- 0.20
- 300 j
-30
-20
-10 0
10 20 Time (s)
}
I Figure 62. Vessel gas pressure, average temperature, and moles in the CE-1 experiment.
a k
NUREG/CR-6469 104
Experimental Results i
7000
- 0.55
- 700
-G-moles pressure
' - temperature 4
- 650
- 0.50 l
l 6000 -
I
- 600 1
~
j i
- 0.45 f
I l
- 550 i
i - 0.40 g 2
a.
2 2
s i
jt 2
5000 -
2
- 500 e 3
g g
i O
g
- 0.35 1 H
- 450 4 >
- 0.30 4000 -
- 400 EM h
1
- 0'25 4W p
- 350 1
3000
- 0.20
- 300 i
-50 40 30
-20
-10 0
10 20 Time (s)
Figure 63. Vessel gas pressure, average temperature, and moles in the CE-2 experiment.
105 NUREG/CR-6469
Experimental Results 7000
- 0.60
- 700 moles pressure
--A-temperature I
- 0.55
- 650 I I a6 6000 -
- 0.50
- 600
- 0.45
- 550 7
2
~
m a.
E 2 5000 -
- 0.40 2
- 500 E E
E O
E E
l a
4 1
- 0.35
- 450
' '1 I
l l 4000 -
I
' - 0.30
- 400
- 0.25
- 350 3000
- 0.20
- 300 j
i i
i i
-30
-20
-10 0
10 20
)
Time (s) j Figure 64. Vessel gas pressure, average temperature, and moles in the CE-3 experiment.
NUREG/CR-6469 106
- _. =
Experimental Results i
i 1
7000
- 0.55
- 700
+ moles j
-m-pressure - temperature I
- 650
- 0.50 II 6000 -
- 600
- 0.45 2
- 550 t
- 0.40 g 2
e a
E h
2 5000 -
2
- 500 e E
h 1
o e
E
- 0.35 5 4
- 450 l
l k
l q
- 0.30 4000 -
i
- 400 I
l p
- 0.25
- 350 i
- 0.20
- 300 3000 i
i r.
30
-20
-10 0
10 20 Time (s)
Figure 65. Vessel gas pressure, average temperature, and moles in the CE-4 experiment.
t 107 NUREG/CR-6469
Experimental Results 10 "I"
--o-- CE$ff"
...--o-... CES-2..
. 3...
...--*-.. CES-3..
T..CE1.,
i
--+- CE-2 L
. 4..
--o---C63 -
-+-- CE-4 i
l 4
1-i-
+
E
.. j...
E M
.4
..t..
.. l..
3 i
g
. j...
. j...
..j..
.9 i
M
.. j...
. 4..
<SMMD> = 0.3 mm !
. 4..
i 0.1 -
. 3...
r
+
.. p..
.. j...
..t.
..j..
.. j...
4..
. j_..
.. j..
..i,...
. 4,..
..i..
..i..
n s
a s
1 10 30 50 70 90 99 Gummulative Percentage Figure 66. Sleve analysis of debris recovered from the operating deck.
NUREG/CR-6469 108
e i
t l
4.0 ANALYSES 4.1
- Debris Dispersal Prior to Water Delivery Some scoping calculations of the dispersal phenomena that could occur prior to water delivery to the cavity were perfonned. Specifically, water delivery will be preceded by a " burp" j-of N gas. The dispersal consequences of this burp are:
2 i
1.
that the burp may only be partially effective at displacing melt away from the d
impingement region, I
i 2.
that the burp has a tendency to levitate melt up the annulus, and i.
i 3.
that only trivial quantities of melt will be dispersed from the cavity prior to water.
injection into the cavity.
l The detaUs of these calculations follow.
t i
Figurc 67 schematically illustrates the test geometry.
A large accumulator and i
connecting piping are partially filled with cold water. The water column is separated from the
" empty" melt generator by a set of rupture disks. The empty melt generator (actually a pipe passes through the melt generator) communicates with the cavity and Surtsey atmospheric conditions prior to rupture disk failure because the orifice remains unplugged. Thermite (33 kg) is bumed in the cavity.
After ignition and complete burning of thermite, the rupture disks are blown. Water accelerates into the empty crucible, compressing the small amount of gas. This gas quickly discharges (i.e., burps) into the cavity before the water slug. The purpose of these scoping calculations are: (1) to determine if the ejected water will jet into a stagnant melt pool, and (2) to determine if the gas can entrain significant quantities of melt from the cavity prior to water injection into the cavity.
)
4.1.1 Initial Displacement of Melt Figure 68 illustrates the layout of the reactor cavity. The thermite powder is poured into a thin (1 mil. thick) Teflon or polyethylene bag or tube, forming a donut-like configuration in the bottom of the cavity. A pyrofuse sealed in the bag is used to ignite the theunite. When fully molten, the thermite (8.3 x 10 m') forms a layer 0.026 m deep on the cavity flooc.
Table 16 summarizes some key geometry numbers used in the following calculations.
3 Table 16 also lists the initial conditions prior to failure of the rupture disks. Neglecting flow out the orifice, we assume that the small gas volume in the melt generator compresses isentropically after failure of the rupture disks. Table 17 also summarizes these key conditions.
j 109 NUREG/CR-6469 i
Analyses The burp of gas impinges on the melt surface, Depression of the liquid surface is a measure of the gas's ability to remove the melt from the impingement zone. The depression depth of a gas jet on liquid surface was correlated by Davenport et al. (1966) and later confirmed (Chatterjee and Bradshaw,1972). Their correlation for the depression depth is given by j
-1+h'*
h' 115 h
(4.1)
=
h, <
h,>
x pi gh' i
where j
h distance between orifice and liquid surface
=
1 14 depth of depression
=
h jet momentum flow
=
pt liquid density
=
g acceleration dtte to gravity.
=
The correlation has received validation in the metals industries over the range A
10" <
3 < 10-'.
(4.2) pi gh Thejet momentum f1cw is given by A = p* V,**A,,=
(4.3) p, A,,
where r
7.o 2 v
rh, = C p, A, MW' y ' 2
- i'i s
(4.4)
RJ, s y + 1) is the characteristic flow rate from the compressed gas volume in the melt generator to the cavity, and where t
' 2 ' iI P, = P, y + li (4.5) is the gas density in the orifice. For nitrogen (MW, = 0.028 kg/g-mole, y = 1.33), the orifice density is p*g =_23.3 kg/m', the mass flow rate is th, ~9 kg/s, and the jet momentum flow is h= 2.76 x 10' N.
NUREG/CR-6469 110
Analyses i
The jet momentum flow is well outside the valid range of the correlation, but applying the correlation anyway predicts a depression of h, ~ 1.4 m. This depth may not be realized in a deep pool, but the melt layer in the experiment is only 0.026 m deep. Clearly, the gas burp will punch through to the cavity floor and displace liquid radially so that the water jet will not
' impinge directly on a melt surface. This does not preclude, however, that melt and water can mix violently in the corners of the cavity after the waterjet is redirected by the floor.
The previous argument ignores the dynamics of the process, i.e., ignoring the time to move the molten thermite in comparison to the short lived burp. The burp time constant is m'-- = 6.61x10-'s.
(4.6) rg ~ #s The time required to move a liquid plug (orifice diameter x pool depth) at least one orifice diameter is,
'2D pt A 6'
3 3
r,,,,, ~
(4.7)
F where the force is given by 1
II*'
1 F=1 p, V,' As = 2
= - $f.
(4.8) 2 p, A 2
3 The motion time constant is about 2.76 x 10s, which is only about half the burp time constant.
Consequently, the burp is only partially effective at moving liquid away from the impingement region before water delivery to the cavity.
4.1.2. Entrainment Prior to Water Delivery The threshold for particle levitation up the annulus is based on the Kutateladze number, A$
Ku =
> I4-(4 9)
(Pt 80),,
(Pt 80),2 The gas density in the cavity is a function of the gas temperature (which we assume equilibrates with the debris, T,,,,, ~ 2500 K) and the cavity pressure. The orifice flow area, cavity flow area, and driving pressure do not differ too greatly from the Zion IET tests where representative cavity 111 NUREG/CR-6469
l Analyses 2
pressures were P. ~ 0.8 MPa. Under such conditions, the gas density is p,,,~ 1.08 kg/m and i
. the Kutateladze number is ~ 400, which exceeds the threshold for dispersal.
The potential exists to levitate melt up the annulus, so we must examine the dynamics of
.the dispersal process by computing the coherence ratio, 3 1/2 r
3 1/4 f T
Ms A V,g/3 1
t R' = o = 1.7 Cs (4.10)
{
m
< Ts*,
M, V
v y
where the lead constant (1.7, Section 4.5) is indicative of CE geometry. Substituting values, we compute R, ~.14. Consequently, the time required to disperse all the melt from the cavity is an order of magnitude longer than the burp time. As a result, we expect that only negligible j-quantities of melt will be dispersed from the cavity prior to water injection.
4.2 Pre-HPME Vesscl Behavior i
Combustion of DCH-produced hydrogen played an important role m the ' lion IET tests.
This is illustrated in Figure 69, where the pressure rise in tests with r. reactive atmosphere are a j
factor of~2.5 times greater than pressure rises in tests with inert atmospheres. Reactive atmosphere j
tests were performed ' tith and without hydrogen preexisting in the containment atmosphem. We
).
concluded, therefore, that the observed differences were due to combustion of DCH-produced hydrogen (produced during cavity dispersal) and that hydrogen (~3%) preexisting in the atmosphere had no observable impact on DCH loads.
1 We anticipated that hydrogen combustion could also be a significant contributor to DCH l
loada in some of the CE tests, but the experiment data clearly shows that this was not the case.
Figure 70 compares two steam-driven tests. The CES-2 experiment was conducted with a fully inert atmosphere (N ) while the CE-3 test had a reactive air / steam atmosphere with -4% hydrogen 2
preexisting in the atmosphere prior to the thermite ignition event. Because of the different atmosphere compositions in the CES-2 and CE-3 tests, the pressure rise is normalized by y-1 for comparison. This normalization is suggested by the single cell equilibrium model AP = y - 1 E AE'.
(4.11)
V 1+ y Figure 70 shows that hydrogen combustion, whether DCH-produced hydrogen or preexisting hydrogen had a negligible impact on the HPME pressure rise in the vessel. Some small amount of hydrogen combustion may have occurred in CE-3 to offset a somewhat lower dispersal, but the effect is not large. Figure 71 shows that similar conclusions are derived from the saturated water tests. Dispersal was nearly identical in all the tests driven with saturated water.
Table 17 shows key information regarding hydrogen production and combustion in the CES/CE tests. Row 13 shows that substantial hydrogen burned some time during the test. Itis NUREG/CR-6469 112
Analyses 1
significant that the number of moles burned (130-323 mole) far exceeds the amount of hydrogen L
-(~80 moles) preexisting in the dome (subcompartment temperatures are never high enough to burn hydrogen at test conditions). Consequently, the production and combustion must be associated with either cavity phenomena or some long term production and combustion. There is no evidence of the latter in the gas samples, so production and combustion must be associated with cavity phenomena. In fact, the combustion numbers are more closely correlated with the production i
i number than they are with available hydrogen in the dome.
1 Assume for the moment that all the hydrogen was produced and bumed on the HPME time p
scale (~0.1 - 1 s). The potential pressurization, AP=I Nuzde :,
(4.12) n V
e 5
resulting from adiabatic combustion on the HPME time scale is comparable to or substantially greater than the total HPME pressure rise measured in the tests. This is inconsistent with Figures 70 and 71, which clearly show that potential hydrogen combustion has no impact on observed DCH loads. We, therefore, conclude that the bulk of the hydrogen production and combustion occurred during the thermite bum prior to the HPME event. Such phenomena are not prototypic of a NPP i
l accident where melt is forcibly ejected into the cavity.
i Figure 72 shows the pressure in the Surtsey vessel prior to the HPME event. The i
atmosphere is i a and the cavity is dry in the CES tests. Negligible pressurization of the vessel occurs when V.uermite is ignited at ~-10s.
l Figure 7J shows that substantial vessel pressurization occurs prior to the HPME event in the l
four CE tests. These tests all had reactive atmospheres; consequently, significant quantities of j
hydrogen were produced and bumed prior to the HPME event. Videos all show large flames jetting i
from the RPV annulus into the refueling canal; however, the videos show that this process is 2
noticeably less vigorous in the CE-1 experiment, which had no hydrogen preexisting in the atmosphere. This video observation is consistent with the reported number of moles of hydrogen j
bumed in CE-1 relative to the latter tests.
[
A lower bound to the number of hydrogen moles produced in the cavity and bumed prior to I
the HPME event can be estimated from the observed pre-HPME pressurization, i
V(AP- 0.014x10')
i.
Nu2 =
(4.13) j (y -1)Aen1 by assuming a complete and adiabatic bum of any hydrogen produced during the thermite bum.
Here we subtract out a small contribution to pressurization due to debris / gas heat transfer as inferred from the CES tests. Table 17 (row 17) shows that a minimum of ~30 - 80 moles of i
hydrogen could have been produced and burned prior to the HPME event in the CE tests. This amount of hydrogen is substantially less than the quantities of hydrogen reported as produced and i
113 NUREG/CR-6469
. Analyses -
p bumed in the tests. The combustion of preexisting hydrogen, either before or after the HPME event, cannot explain the shortfall in the amount of hydrogen produced. It is likely then, that there
' are additional processes occurring during the thermite bum that produce additional hydrogen and combustion processes that do not contribute to loads.
i Complete oxidation to FeO of all the Fe in the thermite could produce ~330 moles of i
, hydrogen; consequently, the potential source of hydrogen is limited by other processes. The CE tests had air / steam atmospheres such that condensed steam accumulated in the cavity. However, a j
drain was intended to limit the depth to ~3 - 6 mm. This translates into ~53 - 106 moles of water i
j (and potential hydrogen) in the iron oxidation reaction noted above.
l Thermite powders can be hydroscopic to the extent of ~0.3 moles-H 0/kg-thermite 2
(Gronager et al. (1986). Experiment procedures require baking the thermite to reduce this value by i
about half; consequently, the water vapor driven from the thermite could produce ~5 moles of i
i hydrogen. Thermal decomposition of the cavity concrete could also release both bound and unbound water as an additional source of steam for iron oxidation. Unbound water, however, is the
{
more likely source since the concrete has been dehydrated from multiple uses while unbound water can be reabsorbed from standing water following each CE test. The decomposition velocity is -
~0.35 mm/s for thermite on concrete; consequently, ~10 - 15 moles of unbound water can be driven j
from the concrete. These numbers are roughly consistent with the inferred amount of hydrogen j
combustion based on pre-HPME pressurization as noted above. We conclude, therefore, that there i
is adequate metal and an adequate steam supply in the CE tests to produce sufficient hydrogen to i-explain the pre-HPME pressurization.
I i
Figure 72 shows that the three tests (CE-2, 3, 4) with preexisting hydrogen in the atosphere all have higher pre-HPME pressurizations than CE-1, which did not have preexisting i
hydrogen in the atmosphere. This strongly suggests that preexisting hydrogen burned prior to the HPME event. There are two mechanisms by which preexisting hydrogen can burn prior to the j-HPME event: entrainment into the burning hydrogen jet venting from the cavity, and by inducing a deflagration in the dome. These mechanisms are quantified next.
. Hot hydrogen jets venting from the cavity are observed to bum when they meet oxygen in
{
the refueling canal or dome. All the oxygen for combustion must be supplied by the dome atmosphere. Entrainment, which supplies oxygen to thejet, also carries preexisting hydrogen into l
the burning jet. This additional hydrogen can also burn. Pilch et al. (1994c), has shown that the
]
moles ofpreexisting hydrogen that can be entrained and bumed is given by N
1 Nu7, ~ Nl12.s.
?
H lh i
N 2A,o, - A,n2 o
L From CE-1, we estimate No ~29.3 moles because all hydrogen must have been supplied by the jet since there was no preexisting H Using test specific input, Table 17 (Row 22) shows that only 2
j
~6 moles of preexisting hydrogen need be entrained into the jet. This is an insignificant depletion 2
N i
NUREG/CR-6469 114 l
Analyses of the ~80 moles of hydrogen preexisting in the dome atmosphere. It could be that more hydrogen was produced in the cavity in the CE-2,3, and 4 tests relative to CE-1, but a factor of two increase
-would imply that ~12 moles of preexisting hydrogen would be entrained into the jet. This is an upper bound to possible depletion of dome hydrogen because some of the oxygen required to bum the jet hydrogen could have been sucked in through the nozzle cutouts from the subcompartment.
Consequently, jet combustion alone is not likely to significantly deplete the hydrogen concentration in the dome.
Hydrogen concentrations (--4%) in the vessel prior to ignition of the thermite are below the flammability limits for the atmosphere composition. However, the production of hydrogen during the bum interval and its subsequent combustion as it vents to the dome could heat the dome possibly to the point where the mixture is no longer inert. It is also possible that not all the H2 produced in the cavity burned in the jet; consequently, it is possible that the preexisting hydrogen concentrations were increased from their initial value.
Figure 73 shows the pre-HPME dome temperatures in the CE-1 experiment. Huge variations, with temperatures ranging from -440K to -430K, are observed throughout the dome region. Subcompartment temperatures show almost no response suggesting that all potential pre-HPME combustion is confined to the dome. We note, however, that there is ~30% reduction in subcompartment hydrogen based on posttest gas analyses, so there must have been some combustion in the subcompartment. However, it cannot be determined if this occurred pre-or post-HPME. The large temperature variations in the dome may be indicative of large composition variations also. For instance, the hottest regions may be composed predominately of combustion products of the buming jet while the cooler regions are more representative of the preexisting atmosphere. The potential for deflagrations is impossible to quantify in such situations because deflagrations are both composition and temperature dependent. We can scope the problem, however, by assessing the-potential for deflagrations using the test-specific average dome temperature in conjunction with the test-specific atmosphere composition prior to the thennite bum.
This analysis uses the constitutive relations recommended by Pilch et al. (1994b, Appendix E). The upward and downward flammability limits are given by 4
Xu2 (up) = 0.037 + 0.0238 X,,,,, - 5x10 (T-373)
(4.15)
Xu2 (dwn) = 0.075 + 0.02381 X,,, - 1.0135x10" (T-373) in terms of composition and temperature. Deflagrations will not propagate if the hydrogen concentration is below the upward flammability limit. Complete combustion of all hydrogen is expected when the hydrogen concentration exceeds the downward flammability limit. Between these extremes, the combustion completeness can be approximated by Xu, - X, (up) g (4.16) r7 = Xg2 (dwn)- Xg2(up) 115 NUREG/CR-6469
Analyses Table 17 suggests that there is some potential for deflagrations (rows 24 and 25) in CE-2,3, and 4, but maybe not in CE-1. The video records show a brief pulsating orange glow at the end of the thermite bum in CE-2,3, and 4, suggesting that some deflagration may have occurred. CE-1 did not exhibit this behavior. Table 17 (row 27) shows that ~0 - 15 moles of hydrogen may have burned as a deflagration in the dome. Considering ~6 moles from entrainment plus ~15 moles from j
deflagration would represent ~25% of the hydrogen preexisting in the dome. As an' upper bound, the deflagration potential was recomputed using the maximum dome temperature. In this case, nearly complete combustion of all dome hydrogen is predicted.
]
DCH produced hydrogen burned and contributed significantly to DCH loads in the Zion IET tests (Figure 69). Figures 70 and 71 show that this is not the case in the CES/CE tests regardless of whether the melt was steam-driven or water-driven. It is plausible that coejected water could quench the melt or inert the combustion. For instance, there is ~2.8 kg of water j
coherent with debris dispersal in the water-driven CE tests. Vaporization of all this water would quench the melt by ~160K. The Al 0 phase may become partially solidified but the Fe phase 2 3 would still have a superheat of ~600K; consequently, Fe steam reactions should still be efficient.
Fine melt fragmentation (~0.3 mm mass mean) suggests that both phases remained molten, so quenching is not likely the major mitigator.
Inerting of DCH-produced hydrogen is also a distinct possibility in the water-driven tests, but the steam-driven tests also exhibited no signs of significant combustion of DCH-produced hydrogen. In the steam-driven tests, ~50 moles of steam is coherent with dispersing melt; and if all 50 moles of steam is converted to hydrogen which burns, then an additional -0.077 MPa of pressure should be realized. In reality, the Fe reaction will not go to completion because of thermodynamic limitations, consequently, the potential for pressurization is even less.
The dispersal interval is quite short (~0.1 s) in all the tests, so we ask if there is sufficient
' time to consume the ~50 mole of steam by metal oxidation during dispersal.
1 The characteristic time constant, N
Na mo g2
'k s
n2o m
6k' PsDsR, T (4.17) k, ~ Sh B Ds 4 2 can be computed for cavity conditiens (Pmo ~0.8 MPa,3 ~10 m /s) and the posttest measured particle size of 0.3mm. For these conditions and Sh ~2, the reaction time, % -0.07, is comparable to the ~0.ls time required to disperse the melt, so it is possible that processes are too rapid to convert coherent steam to hydrogen. The interaction times will be ten times longer at plant NUREG/CR-6469 116
Analyses scale so it is reasonable that dispersal will be accompanied by more complete hydrogen production and combustion.
_ The potential for deflagrations in the dome is small following the HPME event in water-driven CE tests. Although peak dome temperatures range from 650-900K in the tests, these temperatures rapidly drop below ~500K during the ~3 - 4 seconds of blowdown. In addition, blowdown adds a minimum of ~2000 moles of steam from the flashing water, which raises the steam concentration in the dome to ~69 percent while reducing the hydrogen concentration
. (assuming no pre-HPME depletion) to ~2 percent. It would be impossible to induce a deflagration in the dome under these conditions.
The steam-driven test, CE-3, is not subject to these quenching and inerting processes to nearly the same extent as the water-driven tests. For the limiting conditions T
~800K, Xm ~0.04, and X,,, ~0.42, we find that combustion could be nearly complete, y ~0.88, in the dome. However, the energy release rate is less than half of what is required to overcome heat losses to structures.
Consequently, any possible deflagration in the dome would not contribute to containment loads.
This conclusion is also true for all pre-HPME deflagrations.
This simple analysis is consistent with the observation (Figure 70) that any possible hydrogen combustion in the dome does not contribute significantly to loads. It is interesting to note that posttest gas analyses show only ~1 percent hydrogen (on a dry basis, no mixing fans) in the dome for CE-3 while CE-4 showed ~3.4 percent hydrogen under similar conditions. Both tests had similar pre-HPME behavior, so the difference likely is due to differences in post-HPME combustion. Thus, post-HPME deflagrations are possible, but they may not contribute to containment loads for the test conditions.
In summary, experiment results show that significant quantities of hydrogen were produced and bumed in the CE tests. Significant pressurization of the vessel during the thermite burn, but prior to the HPME event, can be attributed to some of this hydrogen combustion. The video records, in conjunction with scoping analyses, suggest that significant quantities, and possibly all, of the preexisting hydrogen in the dome bumed prior to the HPME event. These processes, which are not prototypic, altered the temperature and composition of the atmosphere prior to the HPME event. Experiment results also suggest that DCH-produced hydrogen and any post-HPME deflagrations did not contribute significantly to loads in the HPME; however, this is probably due to the fact that most of the hydrogen burned prior to the HPME, which is an artifact of the experimental method used. Bounding analyses for post-HPME deflagrations suggest that, if they did occur, they would not be expected to contribute to loads. Uncertainties in the timing of hydrogen production and combustion preclude definitive conclusions concerning mitigative processes in NPP analyses.
4.3 Thermite Reaction and HPME Interval The thermite burn was executed in the cavity rather than the crucible in all CES/CE tests.
Rupture disks were then intentionally failed in order to initiate blowdown of water or steam into the 117 NUREG/CR-6469
t l
Analyses I
cavity. It is imperative that the thermite reaction be complete' prior to blowdown and melt dispersal from the cavity. To ensure this, two scoping tests were performed outside Surtsey using a 33 kg tube of thermite laying on the concrete floor of a test cavity (dimensions similar to the scaled Calvert Cliffs cavity). The purpose of the tests was to determine the minimum delay time after
. ignition of the thermite prior to commencing the HPME (to ensure reaction compkteness). The concrete floor was dry in the first experiment. The thermite appeared to be fully reacted after 12 s.
A dark slag appeared to form on the surface of the melt pool by 30 s. A 0.6 cm deep pool of water was placed on the floor in the second test. Again, the thermite appeared to be fully reacted by 12 seconds. However, the bum seemed more intense and slag was not seen until about 50 s after ignition.
'l Thermite was ignited ~15 - 45 s prior to failure of the rupture disks in the CES/CE tests.
This time interval, by design, was sufficiently long to ensure complete reaction of the thermite prior to blowdown. In some cases, however, the delay was sufficiently long to raise questions about energy loss from the thermite and partial freezing of the melt prior to blowdown. Figure 74 shows no decisive correlations of melt retention with the correlation delay time prior to blowdown, Scoping calculations were performed to estimate the magnitude of such energy losses.
Figure 75 depicts the relevant geometry. The bulk averaged temperature can be calculated i
from a simple energy balance, dT l
me, - = - h, A, (T - T,,,) - hu ( A, + A,) (T-T,_,),
(4.18)
.where energy is lost upwards.by radiation and downwards by conduction into the concrete.
Thermal resistances in the melt pool are also represented so that the upward and downward effective heat transfer coefficients are given by l
1 - 1 1
1 0.678
+
=
h, h,
h
,o, acT' + K,,o, a
(4.19)
=-+1 2(a,t)u2 0338 1
1
=
+
h h,
h K,
Kn u
n respectively. Note that the thermal boundary layer in the concrete is time dependent.
These equations apply provided the alumina or iron phases are not freezing. Alumina is the only real concem here because it freezes at 2300K. Once alumina freezing is initiated, the temperature is assumed to remain constant at 2300K until the alumina's heat of fusion has been completely extracted,
,]h, A, (T-T,. ) + hu(A, + A,)(T-T,.,)}dt = 0.45M h,,n2o,.
(4.20)
NUREG/CR-6469 118
4 Analyses L
1 Note that thermite is composed of 45% alumina by weight.
i
[
Table 18 lists the material properties used in the analysis. The equations have been solved i
numerically and the results plotted in Figure 76 for two bounding cases. The first case considers both radiation from the top surface and conduction into concrete. As a bound, the top surface is assumed to radiate to a cold environment at 300K. This situation more closely represents the two j
scoping tests that were performed outside Surtsey.
i Radiation from the top surface is expected to be greatly reduced in the actual experiments carried out inside Surtsey. This is because the top surface no longer radiates to a cold open environment. Instead, the top surface radiates to concrete walls and the melt generator, both of which are splattered by thermite during the early phases of the bum. As a second bound, the heat I
losses from the top surface are set to zero and conduction into concrete dominates heat losses from
(
the thermite.
Figure 76 shows that the thermite cools to approximately 2300K during the interval i
(~15-45 s) between thermite ignition and the start of blowdown. Depending on how effective i-radiation was from the top surface, some freezing of the alumina might have been initiated. At most, ~25% of the alumina or 11% of the total thermite mass may have solidified for those tests j
with the longest cooling period prior to blowdown (29-45 s). The iron phase will still have ~600 K j
of superheat at 2300 K, so we expect that it is fully molten and easily oxidized.
l 4.4 Accumulator Depressurization j
i 4.4.1 Introduction i
i l
Accumulator depressurization histories for representative CE tests have been examined.
i Analyses were performed to quantify the extent to which experiment observations are i
understood. Blowdown in the SNL/CES/ ICE tests can be logically grouped into one of three i
categories as noted in Table 19. The following sections analyze blowdown records for each
- category, i
}
4.4.2 Accumulator Depressurization During and After Ejection of a Nonnashing Liquid
?
Figure 77 depicts the situation to be analyzed. Such a situation exists only in the CES-1 experiment. A pressurized accumulator is partially filled with water, and water ejection is initiated at t = 0.
Depressurization occurs in two phases: liquid ejection followed by gas discharge. During the liquid ejection phase, the gas mass in the accumulator remains constant but its volume increases. Single phase gas discharge begins when all the water is discharged from the accumulator. Gas expansion during both phases is assumed to be isentropic and that gas flow through the hole is also assumed to be isentropic during the gas discharge phase.
Simple analytic expressions for the depressurization history are developed below for each phase.
I19 NUREG/CR-6469
l l
Analyses Liquid Discharge Phase A three-step procedure is used:
l 1.
The governing equation for the depressurization transient is developed from (a) water continuity, (b) a state equation for isentropic gas expansion, and (c) constancy of total i
accumulator volume, i
2.
The governing equation is normalized to develop an expression for the characteristic depressurization time constant during water ejection, and 3.
Simple analytic solutions are derived for the water ejection time and the depressurization transient during water ejection.
The water continuity equation is given by dV*
- C A, 2 (P-P,)' "'
@.M)
=
s dt p,
There are two time dependent unknowns here: the water volume (V,) and the accumulator pressure (P). An equation for P alone is sought, so the instantaneous water volume must be related to the instantaneous accumulator pressure. Constancy of accumulator volume requires that V, = V," + V," - V, (4.22) where the instantaneous gas volume (V,) can be related to instantaneous pressure through a state equation for isentropic expansion, V'
'p** h (4.23)
--=
V,"
Combining equations 4.21-4.23 yields the governing equation for depressurization during the water discharge phase, y Cs Au ' P
- h
'2(P-P,)* "' dP P (4.24) dt V," r P* > p, subject to the initial condition P (t = 0) = P. NUREG/CR-6469 120 Analyses This equation is now normalized such that P and its derivative are of order unity. This is 3 achieved by defining normalized variables: l I t*= (4.25) pe, P - P,w P-P,w P' - P,g AP 1 The pressure normalization can be written in a number of useful altemative forms. First, P = P,3 + AP P * (4.27) i i from the basic definition (Eq. 4.26). Divide and multiply the right hand side (RHS) by P' = P,a + AP, so that P = P' 1+ x'P * = P'P+ (4.28) 1 + x, where AP (4.29) xi = P,s r is a " coupling factor" and P+ = 1 + xiP* (4.30) 1 + xi is a transient term that varies from one to (1+ni)' as P* varies from one to zero. In this way, the physical pressure (Eq. 4.28) is written as a product of a constant term, P*, that carries the j magnitude of the transient and a second term of order unity, P', that defines the transient. Physically, when ni 1, then the accumulator pressure is essentially constant (P ~ P,a ~ P ), and the accumulator pressure is decoupled from the water ejection transient. This is the origin of the name " coupling factor." The utility of these manipulations becomes obvious when the governing equation is normalized, _ y Cs A, P, '2(P' - P,)' "'(P ) Y ' P* - x,' "' AP dP* (4.31) 4 p, s < 1-x 1 dt
- V,'
2 > 121 NUREG/CR-6469 .~. Analyses P where x = p. 2 The intent of normalization is that dP*/dt* and P* be transient terms of order unity; consequently, AP y C, A P, 2(P' - P,)' 3 (4.32) ~ t,,, V,' p, if both the left hand side (LHS) and the RHS are the same order of magnitude, as they must be. With this recognition, the normalized equation for the pressure transient reduces to 8 gp, cl ' p+ _ g 2 = - (P+ ) ' (4.33) dt * < 1-x 2, which is subject to the initial condition P *(t* = 0) = 1 or P+(t* = 0) = 1 (4.34) The characteristic depressurization time during water ejection is then given by 1 AP V'* t,,,=--- (4.35) yP ,2(po - p, y,,,, CA s3 P, which can also be written as 1 AP,' m* V t"' = (4.36) y P' V,' at*, where. ' 2(P" - P, )' rir" = p.C, A (4.37) 3 P, is the characteristic water discharge rate. An analytic solution to Eq. 4.33 is easily obtained when n2 = P/P 1, which is the common case of practical interest. Note also that NUREG/CR-6469 122 Analyses dP* 1+x, dP* (4.38) = dt
- x, dt
- from Eq. 4.28. With the former approximation and the latter substitution, the governing equation becomes U
dP* x' 1+ x, (P') 2' (4.39) =- dt
- which is easily integrated yielding P* = P 1
(4.40) = P* y+2 K s 9 s 27 1+ x j i or alternatively, 2r / Sg P*= P 1 (4.41) = P* 1+y+2 x, P* V m*,t 2 1+x AP V," m*, t i j when written in real time. Note that P' goes to (1+ni) as P approaches P,a. This observation in conjunction with Eq. 4.36 can be used to find the precise time that marks the end of depressurization during water
- ejection, d
7 1+ x' ~I t*= (1+ x ) 2r -1 (4.42) i y+2 x, or alternatively AP V** m' '(1 + x ) 2r d 2 1+ x i -1 (4.43) t= i y+2 x, P* V: m* g when expressed as real time. 123 NUREG/CR-6469 ~.. _. _ _ _. Analyses i All parameter evaluations should be written in terms ofinitial and boundary conditions. This is explicitly obvious, except in the case of 1 '$p. p * - p""" p* i - 1, (4.44) x= = = P,i, P,a P,a i 1 where we require an estimate of P*/P,i,. Evaluation of P*/P,i, falls naturally into one of two phenomenological regimes: inadequate gas volume' resulting in incomplete water ejection or 1 adequate gas volume with complete water ejection. 4 Consider the first regime. The gas volume is small so that the accumulator pressure i e drops to ambient terminating water ejection before its completion. Under these conditions, i AP P* 1. (4.45) x= =-- i P,1, P, The water remaining in the crucible is given by V,, = V,' + Vl - V,, (4.46) J where the final gas volume can be expressed in terms of the pressures with the isentropic state equation (Eq. 4.23),. ) t pas; r V,j = Vl - Vl < -1 (4.47) \\ aA The second phenomenological regime has adequate gas to completely eject all the water from the accumulator before the pressure is reduced to ambient. The final pressure is given by the isentropic state equation, r ar p" y'o (4.48) --= P' < V,' + V *, so that J Ap ' V,* + Vl * ' -1 and $p =1-(4.49) V,' x= = i P,1, Y,' P' < V,' + Vl, NUREG/CR-6469 124 Analyses The accumulator temperature at the end of water ejection is also obtained from an isentropic state j equation -= - $ V'* \\ i T 'P (4.50) 1 T* <P*> < V,' + Vl, The second regime, which is more relevant to our applications, is assured provided >r r V'o P 2:,. (4.51) < V,* + Vl, P* Gas Discharge Phase Isentropic depressurization and flow after the water ejection phase are assumed. The depressurization transient is given in gas dynamics text as 3h r P 1 (4.52) -= P* 1 + y - 1 h,' -t 2 m,' g where g nn mw' ,2 r-n (4.53) m,' = C A P* R,T,, y < y + 1, is the characteristic gas flow rate. Care must be taken when evaluating these expressions because P* = P, and T' = Tr are the accumulator conditions at the end of the water ejection phase. Likewise, time here is also referenced to me end of the water ejection phase. Application to the CES-1 Test Table 20 summarizes input parameters for the depressurization models. The volume of the flow nozzle region below the rupture disks is 0.009 m', which is 3.54% of the total volume above the rupture disk. Following failure of the rupture disk, the gas volume above the rupture disk expands as water pushes gas from the flow nozzle. Assuming isentropic expansion of the cover gas, modified initial conditions for the calculation are given by 125 NUREG/CR-6469 Analyses P V,* ' T' h (4.54) -= = P* < V * + V,*,,, < T* > The modified initial conditions are in parenthesis in Table 20. Figure 78 compares experiment data with model predictions. Using a standard orifice discharge coefficient 0.6, the model predicts transition from water to gas discharge reasonably well but tends to overpredict the pressure during any given instant during water discharge. The actual nozzle is better characterized as a reducer from a 10 cm pipe to 30 cm of 5 cm pipe. For such situations, one might expect the discharge coefficient to exceed the orifice value but be less than ideal. Figure 78 shows model predictions for C, = 0.8. Predictions better track the pressure during water discharge, but transition to gas discharge occurs too soon. Simple hand calculations can be performed to gain insight into the better value. If the pressure were constant, the water discharge time would be given by m* m*" t = "- = (4.55) CA(2p,(P-P,))"'.
- 8 Using the smaller of the candidate discharge coefficients (C, ~ 0.6), the discharge time would be i
~0.62s for a constant pressure of 8 MPa and ~0.88s for a constant pressure of 4 MPa. The actual ^ discharge time should be somewhere in between, ~0.75s. Figure 78 snows that the actual transition time is ~0.9s, which is outside the plausible range and' noticeably larger than a 2 reasonable estimate. This apparent inconsistency would be even worse for a discharge { j. coefficient of 0.8. j. The explanation probably lies in the ambiguous transient phenomena that occurs in the [ first 0.ls. These phenomena are largely associated with pressurization and voiding of the nozzle i region, and Figure 78 shows that wild pressure oscillations occur during this period before a l quasi-steady behavior is observed. Quantitative resolution of this early behavior is outside the scope of this effort. Analysis of the other tests indicates that a discharge coefficient of-- 0.6 does a good job predicting the data, so it is recommended that C, ~ 0.6 be used in the an ~ lysis of CES-1 also. For our purposes, the model developed here provides confidence that our basic interpretation of water discharge followed by gas discharge is adequate. 4.4.3 Steam Only Blowdown The CES-2 experiment had no water in the accumulator and was pressurized with steam. The blowdown history for such a situation is given by Eq. (4.52). Table 21 summarizes the relevant initial conditions. Figure 79 shows that model predictions are in good agreement with experiment data for a discharge coefficient of 0.6. These predictions provide confidence that steam only blowdowns are adequately understood. NUREG/CR-6469 126 Analyses 4.4.4 Ejection of a Flashing Liquid Followed by Steam Blowdown Four tests (CES-3, CE-1, CE-2, and CE-4) were performed with 100 kg of saturated water in the accumulator. The test procedure was to pour 100 kg of cold water into the accumulator. The sealed accumulator was then heated until the partial pressure of steam reached the target conditions of the test. The procedure was reliable and repeatable; consequently, only CES-3 and CE-3 need be analyzed as they are representative of the others. The CE-4 experiment was conducted with about half the dciving pressure as the other tests. Table 22 lists the key initial conditions for CES-3. Water / Steam Inventory The accumulator pressurization realized in these tests comes as a consequence of vaporizing some of the initial water inventory. Here, we quantify the relative portions. The volume fraction of gas at saturation is given by f, = Pr., -- P' (4.56) Pt.s ~ Ps., where p* ~ 100/.254 = 393.7 kg/m' is the initial water inventory averaged over the entire accumulator volume (i.e., the small air mass at one atmosphere is ignored), and where the liquid and gas densities are taken as saturated at the target pressure of 8.33 MPa. The steam volume fraction is 0.473 from which the gas and liquid volumes are computed to be 0.120 m' and O.134 m', respectively. Using the appropriate densities, the steam and liquid masses are computed to be 4.37 kg and 95.63 kg, respectively. Consequently, very little of the water inventory is vaporized to achieve the target pressure in the accumulator. Table 22 summarizes these results for later use. Water Discharge Rate and Flux Figure 80 shows the depressurization history for CES-3, CE-1, and CE-2. It is obvious that flashing water is being discharged for - 2.4s. The mass discharge rate (39.8 kg/s) and the l 2 mass flux (3.17 x 10' kg/(m s)) are then quantified and listed in Table 22 as experiment values. We now compare these values with predictions for flashing two phase flow from a nozzle. The mass flux for critical flashing flows is bounded by G < G < G,,.n. (4.57) um The Bernouli mass flux is given by G,,,.i, =.61 (2p,,,(P' - P,,,, ))", (4.58) m i 127 NUREG/CR-6469 i Analyses i l where the pressure difference driving the flow is controlled by critical pressure (Todreas and Kazimi,1990; pp. 513). Figure 11-25 (page 511-in Todreas and Kazimi) gives the critical pressure ratio (Pg/P ) as a function of the nozzle L/D ratio. For the experiment nozzle, L/D ~ 5.7 and Pw/P* ~ 0.45. The Bernouli mass flux is then 3.49 x 10 kg/(m s). 4 2 For sufficiently long nozzles, Lahey and Moody (1993) state that the critical mass flux can be reasonably predicted with the homogenous equilibrium model (HEM). This is illustrated .t by Lahey and Moody in Fig. 9-20, on page 457 in their text. Figure 9-10a in Lahey and Moody E shows that G can be read from a plot with the stagnation enthalpy (h ') of the outflow and u I system pressure (P*) as parameters. The stagnation enthalpy of the saturated liquid is 1.35 MJ/kg 4 2 so that Ga ~ 2.93 x 10 kg/(m s). 1 in summary, the experimentally determined mass flux of flashing water is tightly a bounded by predictions using the Bernouli discharge model and the homogeneous equilibrium model. i Depressurization Rates i i Figure 80 shows that the depressurization rate is nearly constant at 1.16 MPa/s during the [ period of flashing discharge. Lahey and Moody (1993, pp. 475) express the depressurization rate as i, dP 4 G(P*,h*,) -hf - f(P* ) (4.59) -=- dt M F(P',Y / M) j where F(P*,V/M) and f(P ) can be read from Figures 9-27 and 9-28 in Lahey and Moody (1993). i Using the measured mass flux, the depressurization rate is estimated to be ~ 0.84 MPa/s.. This is lower than the experimentally determined rate of ~ 1.16 MPa/s, and the reason is not readily apparent. 1 Eq. (4.52) can be used to predict the gas blowdown phase provided the time of transition i is properly taken into account. The initial pressure and temperature are 6.31 MPa and 552 K as i estimated from a predicted depressurizatian rate of 0.34 MPa/s up to 2.4s. A discharge coefficient of 0.6 is used. 4 Figure 80 compares model predictions with the experiment data. The models generally i-overpredict the pressure during discharge of flashing liquid; however, these predictions are l-adequate to demonstrate a basic understanding of the key processes. The gas discharge phase is l well predicted if the somewhat higher initial conditions are acknowledged. Figure 80 also shows a prediction assuming the water is discharged as a nonflashing liquid. These predictions are in ll gross disagreement with the data. e i i NUREG/CR-6469 128 1 1 5 Analyses 4.5 Coherence of Debris Dispersal and Blowdown The TCE model assumes that debris / gas interactions in the cavity are limited to that portion of the blowdown gas that is coherent with the dispersal process. The ratio of the characteristic 3 dispersal time to the characteristic time constant of blowdown is termed the coherence ratio. Smaller values of the coherence ratio means that the primary heat sink for debris / gas thermal interactions is smaller and that metal / steam reactions are more likely to be steam limited. The notion that noncoherence (between debris dispersal and RCS blowdown) can limit DCH interactions is not unique to the TCE model. Ginsberg and Tutu (1987) were the first to suggest this limitation. Early CONTAIN calculations (Williams and Louie,1988) also exhibited some sensitivity to coherence, though the effect found was not large. The CLCH model (Yan and Theofanous,1993) also considers noncoherence as a basic modeling process. These analytic reflections all have a solid basis in experiment observations. Unpublished real-time flash x-rays taken at SNL show that dispersal is complete well before blowdown. In addition, many experiments have been conducted (e.g., Allen et al.,1991; Allen et al.,1992a,b) with pyrometers focused on the cavity exit. Pyrometer signals also confirm the notion of noncoherence, and they suggest that cavity pressurization records can also be used to define the coherent interval. Despite this physical evidence, no systematic experiments have ever been performed for the purpose of directly validating the impact of noncoherence on DCH loads. Pilch (Appendix E in Pilch et al.,1994b) developed a correlation for the coherence ratio based on momentum considerations. The Pilch correlation can be expressed as R,=v'- = C,,f, T"* * * 'C M
- A, V'
(4.60) o M,, V r, < Ts i acs > where Ca, is determined from experiment data. For an isentropic blowdown of the RCS, the fraction of blowdown gas that is coherent with debris dispersal is given by M'* ' P' "' h = 1-1+ y-1 R, h h (4.61) fg = 1 - =1- \\ 1 M,,, < Pacs > for R, s 0.5, fa ~ R,, so that R, is directly proportional to the amount of blowdown gas that can react with the debris. The coherence ratio determines how much blowdown gas has been vented from the RCS (or accumulator) on the same time scale as debris dispersal. Figure 81 shows a conceptual cavity pressurization record and a conceptual accumulator blowdown curve. The entrainment interval is determined primarily from the cavity pressurization record and is defined as the interval of time when cavity pressure exceeds vessel pressure. Pyrometers and video cameras placed at the cavity exit show that the cavity pressure joins the containment pressure at the end of debris dispersal. 129 NUREG/CR-6469 _._ _ _ ~ Analyses e 5 Despite these independent measures, cavity pressure and pyrometers, the interpretation of the entrainment interval is still inherently subjective. The end of the entrainment interval is marked by ( and corresponds to a final pressure P in the RCS. The decline in RCS pressure over the entrainment interval is a direct measure of the amount of gas vented into the cavity, and for an isentropic expansion within the RCS, j l AM' M"*= 1 - P' ' "' f, M,' M,,, < P' > (4.62) For computational convenience, it is useful to idealize the blowdown process as a single-phase gas discharge from a fixed size orifice (after any possible ablation). For isentropic expansion in the RCS and isentropic nozzle flow, the ideal blowdown history can be approximated by .-2 7 P 1+ y -1 t'R (4.63) -= P' 2 rv. r where the characteristic blowdown time is given by M',* (4.64) r, = s Physically, the pressure will be reduced to 29 percent ofits initial value (for y = 1.33) after one time constant, and 61 percent of the initial gas mass will have been vented from the RCS. The " measured" coherence ratio, assuming the idealized blowdown and consistent with the measured depressurization over the entrainment interval, is obtained from Equation 4.65, 2 ' P' ' Tr~ R, = r' = -1 (4.65) r, y-1 < P,, The real utility of this formulation is that the idealized ficw will predict the same quantity of vented gas (using this value of the coherence ratio) as the actual flow will vent by the end of the measured entrainment interval. Measured values of the coherence ratio are presented in the last column of Table 23. The supporting experiment information is listed in the prior two columns. Figure 82 shows a correlation of the measured coherence ratio with Equation 4.60 for those i cases where steam alone is the dispersing medium. The more extensive database for the Zion and Surry cavities is also shown for comparison. The solid lines represent least squares regressions through the data. The cavity-specific constant, Co, is also determined from the least squares NUREG/CR-6469 130 i Analyses i 1 analysis, with the Ca,. values being listed in Table 24. Table 24 also shows two statistical measures for the correlation: the relative bias and the relative RMS error (standard deviation) referenced to i the bias line. The statistical measures are defined by j 'enn - d.l 'Unn - M, 7 7 d=' (4.66) e=- N N-1 The statistics for the Calvert Cliffs cavity are not very good because of the limited database i and because the " measured" values of R, are sensitive to the selected values of P /Pe when they are - close to unity. The key observation, however, is that the Calvert Cliffs cavity is approximately six times 'more dispersive than Zion-like or Surry-like cavities even with the acknowledged l uncertainties. t i Columns 2-8 of Table 23 provide all the information necessary to evaluate the RHS of Equation 4.60. Consistent with the blowdown transients, the discharge coefficient is taken at 0.6. { Table 25 examines the applicability of the existing database to Calvert Cliffs applications. It is l clear that the database largely overlaps typical reactor applications. We note that ignition of thermite in the cavity (rather than pressure driven ejection into the cavity) probably maximizes i coherence. In the more prototypic case some melt may have already exited the cavity under its own momentum prior to gas blowdown. Furthermore, we note that complete oxidation of the zirconium in-core debris is predicted in NPP applications, even with this much reduced coherence. I It is useful to examine the Zion and Surry databases for validation insights on parameters j other than cavity design. This is because the database for Calvert Cliffs is too,"mited. Zion and j Surry experiments have been conducted at 1/40*,1/30*,1/20*,1/10*, and 1/5.75* scale. The data l confirm model predictions that there is no effect of physical scale. Experiments have been i j conducted at driving pressures ranging from 4 to 13 MPa, with hole sizes ranging from 0.4 m to j 1.0 m (full scale equivalent), and melt densities ranging from 4000 - 8000 kg/m'. These dependencies are adequately accounted for by the model. Cavity design is the sole systematic deviation of the data from the correlation. Variations with cavity design, which are not. fully i accounted for by the model, are not surprising given that such variations are well documented in j experiments that derme the low pressure dispersal curve. This implies that different values of Ca, are required in Equation 4.60 for each cavity design. I j Table 26 shows how many moles of steam are coh rent with the dispersal process, N,- N' R,, (4.67) 'for the two steam driven tests. It is instructive to compute the amount of coherent steam (flashing) and water in the water-driven tests. We can ignore the accumulator steam (or nitrogen) in the i water-driven tests because the dispersal interval is only ~0.1 s while the water ejection interval is ~2.5 s. j 131 NUREG/CR-6469 i 4!. Analyses The moles of coherent steam and water in the water-driven tests can be computed from X'if*I' N* ~ (4.68) 1 Af % 1 N~ ~ (1-X,) XI, t, (4.69) Af % The water ejection rate (if,) is estimated by dividing the initial 100 kg of water by the ejection interval as observed in accumulator blowdown plots. If the water is superheated relative to the containment pressure, then some of the water will flash to steam when it discharges into the cavity. The thermodynamic quality (X,) can be estimated by assuming isenthalpic blowdown to containment pressure. The dispersal interval is estimated from the cavity pressurization interval. With this input, the moles of coherent steam and water can be computed with Equations 4.68 and 4.69. The results are presented in Table 26. The amount of coherent steam in tests with saturated water is approximately double that in the steam-driven tests with the same driving pressure. Coherent steam in the steam-driven tests is only a small fraction (~15%) of 330 moles required to oxidize all the Fe in the thermite. In the water-driven tests, there is sufficient coherent steam from flashing and coherent saturated water to oxidize 50-90% of the iron. 4.6 Flow of Dispersed MaterialInto the Subcompartments The Calvert Cliffs geometry favors debris dispersal into the dome. In the CES/CE experiments, approximately 75% of the dispersed melt passed into the dome. This is obtained directly from the posttest mass balance. This can be contrasted to the 10 - 20% dome transport observed in the Zion and Surry test. Figure 83 depicts the Calvert Cliffs geometry. There are two paths by which ~25% of the dispersed melt can enter the subcompartments: through the small manway on the cavity floor and through the six nozzle cutouts in the biological shield wall. Our intent here is to estimate how much material enters the subcompartment through the nozzle cutouts. Below the RPV, melt may enter the annular gap around the RPV or may disperse out the manway opening. Pilch (1994b, Appendix I) has shown for Zion and Surry cavities that the fraction of dispersed melt entering the gap, A' f, ~ A, + A,,, (4.70) isjust the area fraction for that particular flow direction. Table 27 shows that ~95% of the dispersed material is expected to enter the gap. This is because the manway area is small compared to the gap
- NUREG/CR-6469 132
R Analyses p ' area, and because a plug (in the experiment) in the manway fully opened as intended in only the CES tests. The manway area was partially plugged in the CE tests. To first order, almost all the material collected in the subcompartments must pass through the nozzle cutouts. This is quantified more carefully as follows. The fraction of material entering the dome )' fu = f,(1-f.) (4.71) is determined largely by how much material passes up the gap, subject only to some attenuation as material is diverted through the nozzle cutouts. The dome transport fraction, f, is determined i. _ experimentally, and f, has been estimated previously; consequently, the inferred nozzle split fraction can be computed as j f, = 1 - (4.72) /s. Table 27 shows that approximately 22% of the melt passing through the gap is diverted by the nozzles into the subcompartment. Comparing the steam-driven tests (CES-2, CE-3) with the others suggests that f, is not sensitive to the dispersing medium. Comparing CE-4 (4 MPa) with the other tests (8 MPa) suggests that f, is not a strong function of driving pressure. This lack of pressure sensitivity contradicts observations by Bertodano (1993) using a similar geometry. Bertodano performed 1/20* scale experiments with Woods metal as the melt simulant and N or He as the dispersing medium. Bertodano found that ~21% of the dispersed material passed 2 through the nozzle cutouts. This fraction is similar to our results, except that Bertodano reports some sensitivity to flow parameters. Bertodano found that increasing the pressure from 4.1 MPa to-6.8 MPa decreased the amount of material dispersed through the nozzles from 31% to 13%. Bertodano cautions that the experiments were not scaled for this phenomena. In particular, the amount of finely fragmented entrained material was far less than expected at plant scale. The measured sieve mass mean size is ~0.3 mm in the CES/CE tests suggests that more prototypic conditions were achieved, perhaps because of the increase in physical scale, the higher driving pressure, and the feedback of gas heating in the cavity. There are two physical mechanisms by which dispersing melt can be diverted into the nozzle cutouts. The first involves splashing off the underside of the nozzles. As an upper bound, we assume that everything striking the bottom of the nozzles is diverted into the cutouts ^ *j""*". (4.73) f% ~ e i where A,,is the projected area of the nozzles on the gap looking up from the bottom. Table 27 shows that this approach overpredicts f. by more than a factor of two; however, it is reasonable that 133 NUREG/CR-6469 Analyses not all the projected area of the nozzles in the gap is effective at diverting the dispersing melt. Since our experiments are geometrically scaled, we expect that this contribution to flow into the cutouts will be preserved. The second mechanism for diverting melt into the nozzle cutouts is for small particles to follow the gas flow into the cutouts. As an upper bound, we can assume that all the particles are efficient at following the gas flow. In this bound, A" f,,p ~A,+A, (4.74) Table 27 shows that this simple expression also overpredicts f, by more than a factor of two. This is not surprising since it is well known that dispersing melt does not follow gas streamlines very well. For a geometrically scaled experiment such as ours, the relevant scaling i parameter is P' L x= (4.75) Ps d. 1 Paiticles follow gas streamlines for larger values of this scaling group. Thermite has half the density of corium, but this distortion is overcompensated for by a factor of ten increase in length scale when extrapolating to plant scale. Since we expect similar particle sizes at plant scale, there is the potential for enhanced diversion through the cutouts at plant scale. In summary, the experiments suggest that ~20% of the melt passing up the gap will be diverted into the nozzle cutouts and that number is not sensitive to driving pressure or driving medium. Scaling arguments imply that comparable or greater diversion into the cutouts can be expected at plant scale. Simplistic arguments based on splashing areas or perfect coupling of debris and gas are inadequate for predicting debris diversion into the cutouts. 4.7 Validation of the TCE Model in Open Geometry Experiments The Calvert Cliffs cavity geometry is such that most melt will be dispersed directly to the 1 upper dome of the containment. Thus, the Calvert Cliffs geometry does not favor the mitigating processes resulting from debris trapping in subcompartment structures as observed in Zion and Surry geometries. The TCE model was successful in predicting experiment results in Zion and Surry geometries, and it is our desire here to validate TCE for application to the Calvert Cliffs geometry. 1 Experiment results clearly show that DCH loads are lower in the water-driven CES/CE tests compared to the steam-driven tests. Since TCE has no model for water interactions, we validate the TCE model against the more bounding steam-driven tests: CES-2 and CE-3. We r.ote that thermite was burned in the cavity rather than forcibly ejected from the melt generator. This r NUREG/CR-6469 134 l i Analyses i i nonprototypicality is not expected to influence significantly the experimental pressure rises because debris / gas interactions in the dome dominate debris / gas interactions in the cavity. He experiment is conservative in this regard (i.e., more time for cavity interactions) because in the more prototypic pressure driven case, some melt may exit the cavity under its own momentum prior to gas blowdown. Figure 84 compares TCE loads predictions with experiment observations. Le model conservatively overpredicts loads by 21% and 24% for the CES-2 and CE-3 tests, respectively. The Calvert Cliffs experiment geometry faithfully modeled subcompartment structures (~40% of the total volume), although most debris is dispersed to the dome. Experimentally observed dispersal fractions and the experimentally determined distribution of mass between the dome and subcompartment were used in the input. Experimentally observed coherence ratios were also used . in the input. The thermite bum altered containment conditions prior to the HPME event. The bulk average vessel temperature just prior to the HPME event was used in the TCE input. Atmosphere composition was also altered during the thermite burn in some unquantifiable manner. The calculations were performed with the initial atmosphere composition prior to the thermite bum; however, TCE predicted that deflagrations would not contribute to the observed experiment loads. The specification of these quantities for plant calculations is addressed in other sections. Figure 84 also shows TCE validation against other experiments in the database where all-debris was dispersed directly to the dome. These experiments differ from the CES/CE experiments in that no subcompartment structures whatsoever existed in the test chamber. Figure 84 shows that TCE (taken in a limiting case of only one cell) often overpredicts these older experiments by ~100%. This is somewhat surprising since it might be expected that TCE's equilibrium assumptions might be more closely approached in these open geometry tests. This apparent discrepancy is addressed below to gain better confidence in model predictions. Two potential mitigating processes have been identified: freezing of dispersed melt on the dome, and time-of-flight (TOF) limitations to debris / gas interactions in the dome. Table 28 summarizes the assessment of these processes for the various test series. The " measured" efficiency, AP"" (4.76) n,m = Y pred is taken as the ratio of measured to predicted loads. It is seen that the CES/CE tests are more efficient than the earlier tests. Freezing of dispersed melt on the dome was important in the SNUDCH/rDS tests. The settling time for debris to fall from the roof to the floor is about six times longer than the time for debris to rise from the cavity and freeze on the dome. In addition, the rising debris travels in a tight plume and does not significantly interact with the atmosphere as a whole. Consequently, we expect that DCH loads in the SNIJDCH/FDS tests will be greatly mitigated due to excessive freezing on ) 135 NUREG/CR-6469 . Analyses the dome. The TCE predictions for these tests used the entire dispersed mass and not just what settled to the floor. A measure of this effect is given by M"'. n_ = 1 - (4.77) M, Table 28 summarizes the appropriate numbers. Almost half the dispersed mass froze on the bare steel dome in the SNUDCH/FDS tests. This was not the case in the latter tests. A painted concrete slab intercepted debris just prior to impact with the dome in the SNUWC/LFP8 tests. Melt did not adhere to this surface. In the SNUCE-3 test, melt did not adhere to structures because they were wet from condensing steam. Debris did not adhere to structures in the SNUCES-2 test, but for reasons that are not readily identifiable. Time-of-flight limitations to ' debris / gas interactions is the second possible mitigating j mechanism. Debris may settle to the floor before debris / gas interactions can achieve equilibrium. Debris that has settled on the floor is not very efficient at heating the bulk atmosphere because heat is lost to the floor, because debris velocities are reduced to zero, and because the settled debris can interact with only a thin layer of the atmosphere near the floor. An analytical expression, nu = 1 - exp (-r,, / rg7), (4.78) is easily derived for the kinetic efficiency, which describes how much debris / gas heat transfer actually occurs prior to debris settling relative to how much must occur in order to achieve equilibrium. The time for debris to fall at terminal velocity from the roof to the floor is given by H r,,= (4.79) v, where the terminal velocity is i , v2 < 3 c, p, g D (4.80) v, = The time constant for all airbome debris to reach thermal equilibrium with the atmosphere is given by rur = (4 81) 1+y kg7 NUREG/CR-6469 136 Analyses where the heat capacity ratio, [ y/= (4.82)
- ^#
m, c,,, i appears because the debris / gas equilibrium temperature may be high and not all the latent and sensible heat in the debris will be transferred to the gas in order to achieve equilibrium. The characteristic energy exchange rate, kr=6h, (T,' - T,* ), (4.83) is the sum of the energy exchange rates for all airborne particles. Lastly, radiation and convection contribute as' parallel processes to the effective debris / gas heat transfer coefficient h4 = h, + h, = o,c,Tj + (2 + 0.6 Re 0.5 Pr 0.33). (4.84) It should be noted that the kinetic efficiency (tlu) is a function of particle size, and a broad spectrum of particle sizes are observed in the tests. Bigger particles fall faster (small TQ and have less efficient debris / gas heat transfer (longer Tur). The kinetic efficiency is computed for each test H 4 series using the atmosphere conditions appropriate to the tests and using the sieve mass mean particle size appropriate to the individual test. In general, the sieve mass mean particle size was ~1 mm for all the tests except the SNL/CES/CE tests where the sieve mass mean particle size was -0.3 mm. l The computed kinetic efficiency for each test series is listed in Table 28 where the mass i mean particle size (50%) is used as the basis. For the lower half of the observed particle size distribution, complete or nearly complete energy exchange can be expected. The process was completed again for particle sizes representative of the midpoint of the upper half (75%) of the particle size distributions. At the upper end of the particle size distribution, Table 28 shows that i debris / gas interactions can be inefficient. The overall efficiency of debris / gas interactions in the tests can be estimated by taking a weighted average of the kinetic efficiency and considering only the debris mass that does not freeze quickly on the roof of the test chamber, 4,,a=4,,,,,(0.5 Un (50%) + 0.5 Uy (75%)]. (4.85) l Table 28 shows that a combination of freezing on the dome or TOF limitations for some dispersed particles can explain the relatively low measured efficiencies observed in some of the open geometry tests. 137 NUREG/CR-6469 - Analyses - All structure surfaces in a NPP are painted and are expected to wet from condensed steam; consequently, freezing of dispersed melt on the dome is not expected to be a significant mitigator of DCH loads. Dome heights are a factor of 5 - 10 bigger in a NPP relative to the experiments, while particle sizes are expected to be comparable. Thus, TOF limitations are not expects! to be a . significant mitigator of DCH loads at plant scale. i In summary, it is expected that near equilibrium conditions were achieved in the SNUCES-2 and SNUCE-3 tests. Significant mitigation in earlier open geometry tests can be j explained in terms of a combination of freezing on the dome and TOF limitations. The TCE model is not expected to be overly conservative in NPP applications. 4.8 Impact of Coejected Water on DCH Loads in Calvert Cliffs Geometry 4 Water is expected to be in the lower plenum of the reactor pressure vessel in DCH-relevant accident sequences. The quantity of water present is a function of the accident sequence and potential operator interventions. Should the lower head fail while the RCS is still pressurized, melt and water would be coejected from the RPV into the reactor cavity. The TMI-II accident came close to this situation. Water coejected from the RPV, unlike water that might be present in the cavity, has the potential to pressurize the containment even in the absence of DCH contributions. This is because coejected water (at high pressure) will be superheated relative to containment pressure, so some of the water will flash to steam during the blowdown. With an RCS full of saturated water at system setpoint pressure (16 MPa), blowdown into l the containment (the large LOCA design basis accident) could pressurize the containment by ~0.3 l MPa.' The question then arises as to whether DCH loads are additive to this pressurization. The flashing water could be an additional source of hydrogen, but hydrogen combustion may be less likely because of additional steam inerting. ' The possible interactions are complex and beyond current modeling capabilities; consequently, we rely on experiments for a more direct answer. Table 29 summarizes the observed effects of coejected water on DCH loads in Calvert Cliffs geometry. Drivir.g with saturated water-reduces loads by 10 - 20% relative to the steam-driven tests, and driving with room temperature water reduces loads by ~25%. Consequently, we conclude that large amounts of coejected water reduces DCH loads (when hydrogen combustion is not significant) in Calvert Cliffs geometry. 7 These observations are from experiments where any potential hydrogen combustion had an insignificant impact on DCH loads. Situations in the NPP can exist where preexisting hydrogen concentrations can be well above the flammability limits. In these cases, coejected water can partially quench the atmosphere (i.e., keep it cooler) and the flashing component can significantly ' increase the steam concentration in the dome. Taken together, it is likely that coejected water could mitigate hydrogen combustion; however, this mitigation has not been demonstrated decisively in NUREG/CR-6469 138 l Analyses the tests; consequently, a more conservative approach to hydrogen combustion is recommended that takes no credit for coejected water's potential to mitigate hydrogen combustion. Figure 71 compares the post-HPME pressure rise for tests driven with saturated water. It can be seen that the bulk of the pressurization occurs on the debris dispersal time scale (~0.1 - 0.2s). Following this rapid pressure rise (which is associated with rapid heating of the dome atmosphere) the containment pressure holds constant for the ~3s of flashing discharge from the RPV. Gas temperatures in the dome are observed to drop significantly during the blowdown period when the containment pressure remains constant. Thus, containment pressure is being maintained by additional moles from the flashing fraction of the blowdown and possibly from the vaporization of the liquid fraction of the blowdown by the hot atmosphere. The processes that contribute to the maintenance of containment pressure can be expressed as V dP 7_ j 7 = X,m,h, + f,(1 - X,)m,(h, - h ) = 0 (4.86) 3 where the first term is the enthalpy addition to the atmosphere resulting from the flashing fraction (X, ~0.37 in the test driven with 8 MPa) of the blowdown. The second term represents additional enthalpy addition to the atmosphere resulting from vaporization of some fracdon (f.) of the liquid portion of the blowdown. Note that X, is a weak function of RCS and containment pressures and that h, (0.744 MJ/kg-w) and hr, (2.26 MJ/kg-w) are only weak functions of the containment pressure. In this formulation, we note that the reference temperature for h, is zero degrees Kelvin so that h,-c,T. If all the blowdown water (f = 1) was tightly coupled with the atmosphere temperature, then dP/dt would be strongly negative and the containment pressure should decline rapidly following the initial pressure rise during the dispersal interval. For the pressure to remain constant, only ~29% of the liquid water (f. ~0.29) must be vaporizing as the hot atmosphere is quenched. Thus, there is not much potential for coejected water to enhance loads unless substantially larger quantities of hydrogen are produced by metal / steam reactions during dispersal, and this HPME-produced hydrogen burns. Figures 70 and 71 show that this is not the case. At reactor scale, the time available for water interactions with the hot atmosphere are increased significantly, so that a greater fraction of the water may couple with the atmosphere and participate in its quenching. Thus, the potential for blowdown of flashing water to mitigate loads is increased at plant scales. In conclusion, the SNUCES/CE tests demonstrate that DCH loads with coejected water are comparable to or less than DCH loads for steam-driven tests. The potential for mitigation is expected to increase at plant scale. ] 139 NUREG/CR-6469 2 Analyses 4.9 Consistency in Pressure, Tem perature, and Moles i The ideal gas law, PV l Z = NRT = 1, (4.87) f requires consistency in pressure, temperature, and moles in the containment atmosphere. We test. l - this consistency for each test at three different times. The first time is at the end of the pre-HPME thermite bum interval. This is just prior to failure of the rupture disks and after the thermite bum. ] The time of peak pressure is examined next. It is difficult to get a representative average temperature during this period, so we make the last assessment at 20 s after failure of the mpture disk. j I The pressure measurements are reliable and consistent everywhere in the vessel. i Temperature is more problematic because strong temperature gradients exist in the vessel, ~ ) - especially during the blowdown period. - This requires that we use a mole average temperature in i i the vessel._ We have only point measurements (thermocouple arrays) of temperature in the dome and subcompartment, so we approximate the mole average by taking the arithmetic average within the dome and within the subcompartment and then performing a mole average of the dome and j subcompartment. This is only an approximation because strong temperature gradients exist within 3-the dome and subcompattment separately. 1 i The number of moles in the vessel prior to the thermite bum is well characterized; however, l processes occurring during the thermite bum can alter the molar content of the atmosphere prior to J the HPME event.' These processes include: vaporization of the bag containing the thennite, offgasing of the thermite, concrete decomposition, vaporization of condensate water in the cavity, hydrogen combustion, and oxygen uptake by the metals. All these processes occur in the CE tests. l The CES tests are simpler in that only the first three processes can occur because the cavity is dry and because the atmosphere is inert. In Table 30, consistency is tested based on the initial number ] of moles in the vessel so any significant deviation of Eq. 4.87 from unity is a measure of the importance of these other processes. i j Table 30 shows that Z ~1 for the pre-HPME period. It is interesting to note that Z is slightly greater than unity for the CES tests. The number of additional " produced" moles necessary to force Z=1 is also listed. For the CES tests, these numbers follow closely the reported number of l H moles produced. This suggests that a large portion of the H may have been produced prior to i 2 2 the HPME event; however, uncertainties in the vessel average temperature preclude a definitive interpretation. The CE tests have many more processes that can alter the number of moles in the vessel. In . spite of this, Z more closely approaches unity in the CE tests. This may be because we have a . limited ability to estimate an average temperature or it may be because there are processes in the i CE tests that both add or consume moles in the vessel, so there may be some tradeoff. NUREG/CR-6469 140 Analyses 1 ] Peak pressure in the CES/CE tests usually occurs at the end of the blowdown period. We q tested consistency based on the initial moles in the vessel plus any moles added during blowdown. This includes that fraction of any coejected water that flashes to steam during blowdown. It does not include water that does not flash but subsequently vaporizes through heat transfer from the melt or the hot atmosphere. We did not consider the impact of hydrogen combustion or metal oxidation on the mole basis. Table 30 shows that there is not good consistency during the blowdown period, but we suspect that this may be due to particularly poor estimates of the average atmosphere j temperature. We examined consistency at a later time (~20 s) when things are perhaps more settled in the vessel. The basis for vessel moles is the same as the analysis at peak pressure. Consider first the two steam-driven tests (CES-2, CE-3) at 20 s. The Z value is near unity as expected. The Z value is also near unity for the four tests driven with flashing water (CES-3, CE-1, CE-2, and CE-4) at 20 s. The large amount of flashing blowdown is comparable to about half the number of moles initially in the vessel. Second order effects like additional water vaporization or molar changes due to combustion would be difficult to resolve with so many moles in the vessel. Consider now CES-1 at 20 s. The Z value exceeds unity by a potentially significant amount, suggesting that excess moles (-509) were created by water vaporization. This 509 moles of additional vaporization (in addition to flashing during blowdown) represents about 15 percent of the water that did not originally flash upon blowdown. The 33 kg of thermite has sufficient thermal energy to vaporize ~2000 moles of saturated water. 141 NUREG/CR-6469 Analyses Table 16. Debris dispersal prior to water delivery Parameters Accumulator Melt Generator Orifice Cavity Geometry l V'(m') 0.17 2.5x10 2 0.113 D,(m) 0.04 l A (m ) 1.26x10-' 2 Ammn. (m ) 8.96x10 2 Ann (m') 0.317 Initial State P (MPa) 8.17 0.2 0.2 P(K) 283 283 283 p*, (kg/m') M, (kg) 5.95x10' Ma (kg) 33 Ta(K) 2500 .After Rupture Disk Failure and Gas Compression P (MPa) 8.17 T(K) 710 V (m') 1.54x10 3 p,(kg/m') 38.8 M,(kg) 5.95x10-2
- p, (kg/m')
23.3 Gas Flow Through Cavity P 0.8 T 2500 p, 1.08 NUREG/CR-6469 142 l Table 17. Hydrogen combustion and analyses in CES/CE tests ROW TEST NUM. PARAMETER CES-1 CES-2 CES-3 CE-1 CE-2 CE-3 CE-4 INITIAL CONDITIONS 1 Driving Pres. (MPa) 8 8 8 8 8 8 4 2 Driving Media Cold Water Steam Sat. Water Sat. Water Sat. Water Steam Sat. Water 3 Atmosphere Inert Inert Inert Reactive Reactive Reactive Reactive 4 Cavity Water? N N N Y Y Y Y 5 Mixing Fans? N 30 min 30 min N 0 min 2 min 30 min 6 Total Moles in Vessel 4091 3851 3986 3346 3506 3700 3730 7 % Steam in Atmosphere 0 0 0 49.2 473 48.9 51.2 8 % O in the Vessel 0 0 0 12 11.6 Ii 10.9 2 9 % H in Atmosphere 0 0 0 0 3.7 3.8 4 2 10 Moles H Preexisting in Vessel 0 0 0 0 129 139 149 2 II Moles H Preexisting in Dome 0 0 0 0 77.4 83.4 89.4 2 [ EXPERIMENT RESULTS 12 Moles H Produced 75 141 145 191 191 269 215 2 13 Moles H Burned 0 0 0 130 203 323 256 2 14 AP Pre HPME (MPa) 0.015 0.014 0.014 0.059 0.121 0.137 0.125 15 AP HPME (MPa) 0.234 0316 0.293 0.242 0.208 0.253 0.217 INSIGHTS 16 Pred AP(MPa) based on Row l3 0 0 0 0.200 0312 0.496 0.393 I 17 Pred moles based on preHPME AP 0 0 0 293 69.7 80.1 723 I8 Max Moles H form Complete Fe 330 330 330 330 330 330 330 2 Z Oxidation 19 Moles of H from Cavity Water 0 0 0 53-106 53-106 53-106 53-106 2 m 20 Moles H from Hermite Burn 5 5 5 5 5 5 5 2 k 21 Moles H form Concrete 8.9 8.9 8.9 8.9 15.4 11.4 11.4 2 Decomposition g k S 5 h Table 17. Hydrogen combustion and analyses in CES/CE tests g ROW TEST Q NUM. PARAMETER CES-1 CES-2 CES-3 CE-1 CE-2 CE-3 CE-4 y h PREHPME H COMBUSTION 3 g 22 Moles H Entrained into Jet 0 0 0 0 5.5 6.0 6.6 2 g 23 Average (Meas) Dome Temperature 0 0 0 500 600 640 640 24 Upward Flammability Limit 0 0 0 0.0424 0.0369 0.0353 0.0358 25 Downward Flammability Limit 0 0 0 0.0738 0.0633 0.05 % 0.0601 26 Reaction Completeness 0 0 0 0 0.003 0.111 0.171 27 Moles of Dome H Burned in 0 0 0 0 03 93 153 2 Deflagration 28 Maximum (Meas) Dome Temperature 0 0 0 630 850 780 800 29 Upward Flammability Limit 0 0 0 0.0359 0.0244 0.0283 0.0278 30 Downward Flammability Limit 0 0 0 0.0607 0.0379 0.0454 0.0439 31 Reaction Completeness 0 0 0 0 0.932 0.568 0.756 p 32 Moles of Dome 11 Bumed in 0 0 0 0 72.1 473 67.6 2 A Deflagration 6 Analyses i I Table 18. Material properties for thermite coohng Property Al O Fe Thermite 2 3 c 0.8 hr(MJ/kg) 1.16 k (w/m/k) 5 20 1178 C,(J/kg/K) Table 19. Categories for accumulator blowdown histories Tests Distinguishing Features CES-1 100 kg of room temperature water expelled with pressurized nitrogen, i.e., nonflashing water discharge CES-2 No water, steam blowdown only CE-3 CES-3 100 kg of saturated water expelled by pressurized steam, i.e., CE-1 flashing water discharge CE-2 CE-4 Table 20. Input parameters for the CES-1 experiment Parameter Value
- p. (kg/m')
960 V,' (m') 0.150(0.159) V.* (m') 0.104 MW,(kg/g mole) 0.028 y 1.40 D (m) 0.0525 P* (MPa) 8.27 (7.62) T"(K) 311 (304) P,(MPa) 0.2024 C. 0.70 Values in ( ) represent modified initial conditions as noted in text. 145 NUREG/CR-6469 i Analyses Table 21. Input parameters for the CES-2 experiment Parameter Value V,* (m') 0.254 j V.* (m') 0.0 MW,(kg/g mole) 0.018 i y 1.10 D,(m) 0.0523 P (MPa) 8.52 l T*(K) 607 P (MPa) 0.2027 C, 0.6 Table 22. Initial conditions and computations for CES-3 Parameter Value Initial Conditions P (MPa) 8.33 i T"(K) 571 M.,,aa* ('kg) 100 M,* (kg) 4.37 M.* (kg) 95.63 V,' (m') 0.120 V.* (m') 0.134 D,' (m) 0.04 Properties 3 p (kg/m ) 393.7 3 p,., (kg/m ) 36.4 pt,(kg/m') 714.5 h,'(MJ/kg) 1.35 t Water Discharge G,(kg/m'/s) 3.17 x 10' 0%n (kg/m'/s) 3.49 x 10' Gw (kg/m'/s) 2.93 x 10' Depressurization Rate P, (MPa/s) 1.16 o F(P', VIM) (J/kg/Pa) 0.05 f(P )(MJ/kg) 1.25 NUREG/CR-6469 146 . ~ _ _ _ Table 23. Experiment data on coherence ratio for Calvert Cliffs cavity geometry Test f C P, T,,, M*, D, V, V, t', L-R, u (K) (K) (kg) (m) (m') (m') (s) SNUCES-2 .852 .6 607 33.2 9.76 .05 .22 .254 .1 1.11 .081 SNUCES-3 .644 .6 622 33.2 8.59 .04 .22 .254 .1 1.14 .104 Table 24. Key parameters characterizing the coherence correlation [ Zion Surry Calvert Cliffs l Na 21 13 2 C,, 9.618 12.2 1.717 I e 0.014 -0.014 -0.184 u o,,,,, 0.240 0.154 0.585 t Table 25. Applicability of the Calvert Cliffs database to reactor applications -a" f P/I", M*/M A, V,"5 w v. Database 3x10 -4.7x10-8 4 CES-2, CE-3 .64.85 0.26 3.7 NPP - Scenario VI ~.9 0.25 7.7 - 16.3 2.2x10-3 P* = 8 MPa D, = 0.4m P = 1000K i g M*, = 40 - 85 mt g NPP - Scenario V ~.9 0.36 1.3 - 3.5 2.2x10-' o P* = 16 MPa } D, = 0.4m g P = 700K g y M*, = 25 -65 mt y 2 C m Table 26. Coherent steam and water during dispersal in the SNUCES and SNUCE tests 4-Q Parameter CES-1 CES-2 CES-3 CE-1 CE-2 CE-3 CE-4 5 m
- n Steam Driven Tests T
N*,(moles DF 477 R, 0.09 0.104 Coh. stm. (moles) 44 50 Coh. water (moles) 0 0 Water Driven Tests % (kg/s) 166 41.7 37.0 41.7 31.3 0 0.37 0.37 0.37 0.26 X, 0.15 0.10 0.15 0.10 0.15 t, Coh. sim. (moles) 0 86 114 86 45 Coh. water (moles) 1383 146 194 146 129 A Analyses Table 27. Assessment of debris flow through nozzle cutouts Parameter Test Avg. CES-1 CES-2 CES-3 CE-1 CE-2 CE-3 CE-4 2 A,(m ) 0.0825 0.0825 0.0825 0.0825 0.0825 0.0825 0.0825 A,,,, (m') 0.0058 0.0058 0.0058 0.0029 0.0029 0.0029 0.0029 A (m') 0.0878 50878 0.0878 0.0878 0.0878 0.0878 0.0878 I A, bottom (m ) 0.091 0.091 0.091 0.091 0.091 0.091 0.091 2 f, pred 0.934 0.934 0.934 0.966 0.966 0.966 0.966 fe meas 0.792 0.720 0.760 0.763 0.740 0.713 0.742 f, inferred 0.152 0.229 0.187 0.210 0.234 0.262 0.232 0.22 i f splash 0.524 0.524 0.524 0.524 0.524 0.524 0.524 0.52 f,,, flow 0.516 0.516 0.516 0.516 0.516 0.516 0.516 0.52 4 N Table 28. Loads mitigation in open geometry experiments Test Series y ym. agt(50%) ngr(75%) qu SNUDCH 0.40 0.52 0.95 0.62 0.41 SNL/fDS 0.50 0.60 0.97 0.52 0.45 SNUWC/LFP8 0.51 0.93 0.73 0.28 0.47 SNUCES/CE 0.78 50.97 1.00 0.79 0.87 i Table 29. Experiment insights on coejected water i Driving Test Medium Atmosphere AP m CES-2 steam inert 0.316 CES-3 sat. water inert 0.293 CES-1 cold water inert 0.234 CE-3 steam reactive 0.253 CE-2 sat. water reactive 0.208 s 4 4 a 149 NUREG/CR-6469 Analyses Table 30. Test of P, T, N consistency Pre-HPME Post-HPME t=-Os t @ Pmax t=20s CES-1 P (MPa) 0.215 0.450 0.360 T(K) 309 495 425 N (g-moles) 4091 4585 4585 PV/(NRT) 1.02 1.19 1.11 Nprod(g-moles) 93 882 509 CES-2 P (MPa) 0.217 0.530 0.380 T(K) 325 540 500 N (g-moles) 3851 4389 4389 PV/(NRT) 1.04 1.34 1.04 Nprod(g-moles) 164 1514 182 CES-3 P (MPa) 0.215 0.510 0.395 T(K) 310 430 400 N (g-moles) 3986 6231 6231 PV/(NRT) 1.05 1.14 0.95 Nprod(g-moles) 185 902 -292 CE-1 P (MPa) 0.257 0.495 0.395 T(K) 440 445 430 N (g-moles) 3346 5597 5597 PV/(NRT) 1.05 1.20 0.99 Nprod(g-moles) 167 1093 -73 CE-2 P(MPa) 0.265 0.47 0.395 T(K) 455 425 415 N (g-moles) 3506 5751 5751 PV/(NRT) 1.00 1.16 1.00 Nprod(g-moles) -3 900 -27 CE-3 P(MPa) 0.3 0.545 0.37 T(K) 485 650 555 N (g-moles) 3700 4177 4177 PV/(NRT) 1.01 1.21 0.96 Nprod(g-moles) 20 865 -168 CE-4 P(MPa) 0.295 0.512 0.42 T(K) 475 500 450 N (g-moles) 3730 5299 5299 PV/(NRT) 1.00 1.16 1.06 Nprod(g-moles) 5 859 314 NUREG/CR-6469 150 1 Analyses i j 4 Accumulator f
- j Connecting l
Pipe s i 4 1 J i Rupture i Disks ] Empty Melt i Generator 1 i Flow Nozzle Exit l Hole or Orifice -~ 1 i I j Figure 67. Conceptuallayout of the system. i k 151 NUREG/CR-6469 i 8 Analyses il l missile shield refueling canal to accumulator burst diaphragms nev 5 i 1 annular gap i access hatch \\ thermite flow nozzle j Figure 68. Details of the melt generator / cavity layout. l NUREG/CR-6469 152 Analyses 4 j i i i O SNUIET-1 --G-SNUIET-1R O SNUIET-5 7 SNUIET-3 -m-SNUIET-6 -y-SNUIET-4 SNUIET-7 ^ i 0.50 Hydrogen Combustion i 0.45 (Reactive Atmosphere) 0.40 m 0-0 E .35 E 0.30 8 1U W; " 2 0.25 U 1 No Hydrogen Combustion 0.20 (Nonreactive Atmosphere) i l 0.15 i i i i i i 0 10 20 30 40 50 60 70 80 Experiment Time (s) l Figure 69. Hydrogen combustion in Zion geometry SNIAET tests. 153 NUREG/CR-6469 Analyses 10. g; E ES-2, inert 2 8 :: g 7i E-3, Reactive Atm. with 4% H 2 6-2 5 :: a. 44 m =i 32 E 8 2 :~ 3 1' 0 '.. ,,i..,,, 0 1 2 3 4 5 Time (s) 0.9 i CES-2, Inert 0.8 : \\ _ 0.7 d m 0- ~ \\
- 2. 0.6 -
CE-3, Reactive Atm. with 4% H 0.5 g 2 I 0.42 c 2 0.3i < 0.2 :: 0.1 - 0.0~,,,,,....i,,,,, ,,,i.... 0 1 2 3 4 5 Time (s) Figure 70. Hydrogen combustion in steam-driven tests. NUREG/CR-6469 154 .... _ _ = i Analyses 10 - - g 9i a
- ii 8-4 3
2 7i CES-3; CE-1; CE-2 o 6-E 5 :i o. 45 a m 3i CE-4 E 3 2-8 1i O ~....i....i........i....i........i........i.... 0 1 2 3 4 5 6 7 8 9 10 Time (s) 0.9 CES-3* Inert 0.8 : x gE-1, Reactive Atm. ~ with 0% H _ 0.7 i f 2 a \\ $'0.62. A ~ 'CE-2, Reactive Atm with 4% H, 0.5. n i 1 0.42 D
- c. O.3 _
- 0.2 i 0.1 i 0.0 ~....i....i............i...,.. i...i... i....
0 1 2 3 4 5 6 7 8 9 10 Time (s) Figure 71. Hydrogen combustion in saturated water-driven tests. 155 NUREG/CR-6469 Analyses 0.20 ~ 1 0.15 - CE-3 i CE-2 cc n. CE-4 2 0.10 - i a.< 0.05 i CE-1 j CES-1,2,37 0.00 ........i.<...........i........,..i........ 45 35 25 15 -10 -5 0 Time (s) Figure 72. Vessel pressurization during thermite burn in cavity and prior to HPME event. NUREG/CR-6469 156 Analyses i l 800 - --+- d o m e I ~ --+-- subcompartment 700 -' + vessel average l j { g F, 600 - l e B i i 6 /( I i g, >1 g 500 - {_ j g [ l l h s = =. l ~ 400 - i j 300 ' -30 -20 -10 0 10 20 30 ] Time (s) i I i i i Figure 73. Pre-HPME dome temperatures in the CE-1 experiment. 157 NUREG/CR-6469 Analyses 1.0 ) 5 b 0.9 - i m 0.8-o { .c. 0.7 - ] 0.6 - .E 0.5 - .5
- CE-4 gE 0.4 -
8 CE-3 5 0.3 - .a ~ o 0.2-G CE-2 2 S CE-1 0 CES-1,2,3
- u. 0.1 -
0.0 ........i........i................i....i.... 45 35 25 15 -10 -5 0 Time (s) Figure 74. Dependence of melt retention on the delay between thermite ignition and blowdown. NUREG/CR-6469 158 Analyses 0.63m A j hr 4 N A A hAi203 g 0.67 S Al 0 2 3 m T m V 9 = C A hp, pe l v 0.33 S y y j A Concrete 2Ga t h c c V V Figure 75. Heat losses from thermite on the cavity floor prior to blowdown. 159 NUREG/CR-6469 Analyses 1 -J l 2500 i 4 s 2400 - ConducUon E \\ \\ 2 2300 - 4 2 \\ G s 1 2200 l E s i Radiation +/ ', 2100 ; Conduction i r 2000 - i 0 50 100 150 200 l Time (s) Delay Prior to Blowdown 4 d l Figure 76. Predicted temperature history of thermite in the cavity prior to blowdown. i NUREG/CR-6469 160 j Analyses i 4 i Vo 8 Accumulator ^ l 1 i 9 1 Connecting Pipe Vo w l 4 i Rupture Disks ) i Flow Vo sa Nozzle i i i j Figure 77. Conceptuallayout of the system. .1 i ,.i i 161 NUREG/CR-6469 i Analyses Prediction O Data 10 {g 9 39 8y 7 7] Q.2 6-l v g 5] 8 O =0.6 g c) - J[O E 4-d[ 3-2f O = 0.8 g 1: 0 ....i....i....i.... 0.0 0.5 1.0 1.5 2.0 l Time (s) Figure 78. Comparison of the predicted depressurization history for CES-1 with experiment data. NUREG/CR-6469 162 Analyses 1 Prediction O Data 10 9-8- ~ W 7-n. 2 6-2 5-2 4-e d[ 3-C = 0.6 d 2-1- C = 0.8 g 0 i i i 0.0 0.5 1.0 1.5 2.0 Time (s) Figure 79. Comparison of the predicted depressurization history for CES-2 with experiment data. 163 NUREG/CR-6469 Analyses d I l 1 O CES-3 O CE-1 A CE-2 Prediction i 10 i 9 Flashing j/ Discharge 8 ? 7 Gity[A a. m E 4 0 m E 3 Nonflashing 2 Discharge 1 0 i....i...., ...i.... 0 1 2 3 4 5 Time (s) Figure 80. Accumulator depressurization in the CES-3, CE-1, and CE-2 experiments. NUREG/CR-6469 164 Analyses 1 i P Accumulator Pressure 4 I' Cavity Pressure 1 Pe 8 i J rt Time e J 4 Figure 81. Procedure for estimating the coherence interval. 165 NUREG/CR-6469 Analyses 2.0 Surry 1.5 - O og Zion 0: 08 1.0 - O e 8 0 U 0.5 - h Calvert Cliffs 0.0, i i i 0.00 0.05 0.10 0.15 0.20 I* T* *"*' M ' A, V,"' ( Ts* i f* M,* V acs > Figure 82. Coherence of dispersed debris and blowdown during cavity dispersal. NUREG/CR-6469 166 Analyses 1 b l i i l missile shield i l refueling canal 1 i i to accumulator l burst diaphragms f "f. 4 RPV i i RPV nozzle cutout j fu, Anz i i I f annular gap l ~ access hatch f "1-IE A., 4 thermite flow nozzle j Figure 83. Validation of the TCE model for Calvert Cliffs geometry i and other open geometry tests. } } i 167 NUREG/CR-6469 Analyses a SNUDCH O SNLTTDS + SNULFP O SNUWC 0 SNUCES/CE 0.8 0.7 - [0.6-E ~ 0.5-E y 0.4 - 0 1 ~ + 0.3 _ u1 L< 0.2 - 5 0.1 - 0.0 i 0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 AP Measured (MPa) Figure 84. Validation of the TCE modelin "open geometry" experiments. NUREG/CR-6469 168 j l
5.0 CONCLUSION
S The Surtsey Test Facility at Sandia National Laboratories (SNL) was used to perfonn DCH tests with scaled models of the Calvert Cliffs NPP structures. These structures were intended to be representative of Combustion Engineering plants with a Bechtel annular cavity design (Calvert Cliffs 1 and 2, Millstone 2, Arkansas Nuclear One Unit 2, and Palisades). The model included the reactor pressure vessel (RPV), cavity, refueling canal, operating deck, control rod drive missile shield, and crane wall. A corium simulant (thermite) was reacted on the floor of a 1/10*
scale Calvert Cliffs cavity. The melt was entrained out of the cavity into the Surtsey vessel by a i
high-velocity steam or water / steam blowdown of an accumulator modeling the RCS. Seven tests l
were conducted: three tests with inert atmospheres (CES-1,2,3) and four tests with reactive l
atmospheres (CE-1 with air / steam and CE-2, 3, 4 with air / steam / hydrogen).
The major i
conclusions are summarized below.
1.
Driving melt from the cavity with saturated water reduced DCH peak loads in the tests by J
10-20% relative to steam-driven tests.
Driving melt from the cavity with room j
temperature water reduced loads by 25%. The containment atmosphere was rapidly quenched during the period of water ejection into the atmosphere. Therefore, we l
conclude that large amounts of coejected water slightly mitigates loads in the Calvert Cliffs geometry.
i 2.
Significant amounts of hydrogen, preexisting in the Surtsey atmosphere and also produced by the thermite reaction with condensate water, burned in the reactive i
atmosphere tests prior to the HPME event and pressurized the vessel. Experiment data suggest that any potential hydrogen combustion during the HPME event did not contribute to loads. However, the bulk of the hydrogen production and combustion probably occurred during the thermite burn prior to the HPME event. This production and combustion mechanism is an artifact of the experimental method and, therefore, is not prototypic of a NPP accident. However, the combustion of the preexisting hydrogen is considered in a NPP analysis for DCH issue resolution (NUREG/CR-6475).
3.
The coherence of the melt dispersal with the steam and water / steam blowdown was small.
The debris entrainment interval was about 0.1 s. This was much less than the ~0.4 to
)
0.8 s seen in the DCH tests that involved Westinghouse cavities with instrument tunnels.
Comparable coherence ratios were seen in all CE tests, whether the melt was ejected from the cavity using steam or using water. Most of the DCH load on the vessel occurred very
~
quickly, in the melt dispersal time frame (0.1-0.2 s) and not in the blowdown time frame l
(2-4s).
4.
Approximately 78%
12% of the melt in the 1/10* scale Calvert Cliffs reactor cavity was dispersed into the Surtsey vessel. This was comparable to the previous Surry and Zion DCH tests. In the 1/6* scale Surry tests there was 81% dispersal into the Containment Technology Test Facility vessel and in the 1/10* scale Zion tests there was 77%* dispersal into the Suitsey vessel.
1 169 NUREG/CR-6469 i
Conclusions 5.
Results of all seven Calvert Cliffs experiments indicated that 58% of the total debris recovered posttest was transported to the upper dome. In the Zion and Surry tests without the annular gap modeled, only 7%' to 10% of the total debris recovered w~as found in the upper dome.
6.
Substantial cavity pressures were measured in some of the tests.
However, the experimental pressure may have resulted from the nonprototypic contact of blowdown and melt in the cavity.
i Data from IET-4, IET-8A, and IET-8B were excluded when calculating these averages due to nonprototypic damal;e to structures.
NUREG/CR-6469 170
. ~. -.
6.0 REFERENCES
Allen, M.D., M. Pilch, R.T. Nichols and R.O. Griffith, Oct.1991, Experiments to Investigate the Efect ofFlight Path on Direct Containment Heating (DCH) in the Surtsey Test Facility - The Limited Flight Path (LFP) Experiments, NUREG/CR-5728, SAND 91-1105, Sandia National Laboratories, Albuquerque,NM.
Allen, M.D. et al. (1992a). Experiments to Investigate the Efect of Water in the Cavity on Direct Containment Heating (DCH) in the Surtsey Test Facility - The WC-1 and WC-2 Tests, 5AND91-1173, Sandia National Laboratories, Albuquerque, NM.
Allen, M.D. et al. (1992b). Experimental Results of Tests to Investigate the Efects of Hole Diameter Resultingfrom Bottom Head Failure on Direct Containment Heating (DCH) in the i
Surtsey Test Facility - The WC-1 and WC-3 Tests,5AND91-2153, Sandia National Laboratories, Albuquerque, NM.
Allen, M.D., M. Pilch, T.K. Blanchat, R.O. Griffith, and R.T. Nichols, May 1994, Experiments to Investigate Direct Containment Heating Phenomena with Scaled Models of the Zion Nuclear Power Plant in the Surtsey Test Facility, NUREG/CR-6044, SAND 93-1049, Sandia National Laboratories, Albuquerque,NM.
Bertodano, M. Lopez de (l993). Direct Containment Heating DCHSource Term Experimentfor Annular Reactor Cavity Geometry, Ninth Proceedings of Nuclear Thermal Hydraulics,1993 j
ANS Winter Mtg., Nov. 14-18,1993, San Francisco, CA, p. I11-120.
j Blanchat, T.K., M.D. Allen, M. Pilch, and R.T. Nichols, June 1994, Experiments to Investigate Direct Containment Heating Phenomena with Scaled Models of the Surry Nuclear Power Plant, NUREG/CR-6152, SAND 93-2519, Sandia National Laboratories, Albuquerque, NM.
Blanchat, T.K., January 1995, "Results of the Oxide Thermite Scoping Tests," letter report to R. Lee (USNRC), Sandia National Laboratories, Albuquerque, NM.
Chatterjee, A. and A.V. Bradshaw, March 1972," Break-up of a Liquid Surface by an lmpinging Gas Jet," J. ofthe Iron and SteelInstitute, pp. 179-187.
Davenport, W.G., D.H. Wakelin, and A.V. Bradshaw,1966, " Interaction of Both Bubbles and Gas Jets With Liquids," J. ofHeat and Mass Transfer in Process Metallurgy, pp. 207-244.
Ginsberg, T. and N.K. Tutu (l987). Safety Research Programs Sponsored by Ofice ofNuclear Regulatory Commission, Quarterly Progress Report, NUREG/CR-2331, Vol. 7, No. 2.
Gronager, J.E. et al. (l986).
TURC1:
Large Scale Metallic Melt-Concrete Interaction Erperiments and Analysis, SAND 85-0707, NUREG/CR-4420, Sandia National Laboratories, Albuquerque, NM.
171 NUREG/CR-6469
I References i
Lahey, Jr., R.T. and FJ. Moody (1993). The Thermal Hydraulics of Boiling Water Reactors,2nd Edition,' American Nuclear Society, La Grange Park, IL.
1
- Pilch, M.M., Oct.1991, " Adiabatic' Equilibrium Models for Direct Containment Heating,"
j
' SAND 91-2407C, presented at the 19th Water Reactor Safety Information Meeting, Washington, DC.
1 Pilch, M.M., (1994a), " Plant Selection for Confirmatory Testing for CE Plants," Letter report to R. Lee (USNRC), Sandia National Laboratories, Albuquerque, NM.
Pilch, M.M. et al. (1994b). The Probability of Containment Failure by Direct Containment Heating in Zion, NUREG/CR-6075, Supplement 1, Sandia National Laboratories, Albuquerque, NM, Pilch, M.M. et al. (1994c). The Probability of Containment Failure by Direct Containment l
Heating in Zion, SAND 93-1535, NUREG/CR-6075, Sandia National Laboratories, Albuquerque, NM.
1 Pilch, M.M., M.D. Allen, and E.W. Klamerus,- October 1995, Resolution of the Direct i
Containment Heating Issue for All Westinghouse Plants with Large Dry Containments or Subatmospheric Containments, NUREG/CR-6338, SAND 95-2381, Sandia National
)
i Laboratories, Albuquerque,NM.
l
'Pilch, M.M. et al. (1996).
Resolution of the Direct Containment Heating issue for all
)
Combustion Engineering Plants and Babcock & Wilcox Plants, NUREG/CR-6475, Sandia National Laboratories, Albuquerque, NM.
j Todreas, N.E. and M.S. Kazimi (1990).
NUCLEAR SYSTEMS I Thermal Hydraulic i
Fundamentals, Hemisphere Publishing Corporation.
I l
Williams, D.C. and D.L.Y. Louie (1988). "CONTAIN Analyses of Direct Containment Heating Events in the Surry Plant," ANS/ ENS Intemational Meeting, ANS Thermal Hydraulics Division,-
Washington, DC.
Yan, H. and T.G. Theofanous (1993). "The Prediction of Direct Containment Heating," ANS Proceedings,1993 National Heat Transfer Conference, Atlanta, GA, p. 294-309.
Zuber, N. et al.,1991, An Integrated Structure and Scaling Methodologyfor Severe Accident l
TechnicalIssue Resolution, NUREG/CR-5809, EGG-2659.
I i
j i
172
N?:Ceoxv 335 U S. NUCLE AR f;EGUL AToR Y CoMMIS51ohl
- 1. REPORT NUMBE R e5, t N,Y '
CM 1102.
ene m i 32oi. nu BIBLIOGRAPHIC DATA SHEET tsee,nsteuct.ons on une reveses NUREG/CR-6469 SAND 96-2289
- 2. TITLE AND SU8 TITLE Experiments to Investigate Direct Containment Heating Phenomena With Scaled Models of the Calvert Cliffs Nuclear Power Plant 3
oATE REPORT PUBU$wED j
wom uu i
February 1997
- 4. FIN oR GR ANT NUMBE R W6162
- 6. AUTHOR (Si
- 6. TYPE oF REPORT T.K. Blanchat (SNL)
Technical M.M. Pilch (SNL)
- 7. PE R i Cov t R E D uactu - Os, M.D. Allen (SNL) gFo,RujN RG ANIZ ATioN - N AM E AN D AoDR ESS (n 4#C. mv ar 0,v s,on. Onece or Meteoa u.1 hac4er Aepuderory coaimemen, ered meelnap eadreu at sentrecer, m..e aP Sandia National Laboratories Dept. 6402/MS1137 P.O. Box 5800 Albuquerque, NM 87185-1137
- e. sPogoRgoRG ANIZATioN - N AME AND ADO R E$$ fit Nec. type "seme se ecove". # coneracror. pwee 4 Ac 0,veen. Onsce or Aarson. U & hucmer Mepuietory commewon.
Division of Systems Technology Office of Nuclear Regulatory Research U.S. Nuclear Regulatory Commission Washington, DC 20555-0001
- 10. SUPPLEMENTARY NOTES R. Lee, NRC Proiect Manacer
- 11. ABSTRACT (200 wores or eus Tha Surtsey Test Facility at Sandia National Laboratories (SNL) is used to perform ecaled experiments for the Nuclear Regulatory Commission (NRC) that simulate High Pressure Melt Ejection (HPME) accidents in a nuclear power plant (NPP). These experi-mints are designed to investigate the effects of direct containment heating (DCH) phsnomena on the containment load.
In previous experiments, high-temperature, chemi-cally reactive (thermitic) melt was ejected by high-pressure steam into a scale model of either the Zion or Surry NPP. The results from the Zion and Surry experiments were extrapolated to other Westinghouse plants. This report describes tests performed with Combustion Engineering plant geometries (in particular, Calvert Cliffs-like) and the im-pact of codispersed water as part of the overall DCH issue resolution.
Integral effects tests were performed with a 1/10th scale model of the Calvert Cliffs NPP inside the
)
Surtsey test vessel. The experiments investigated the effects of codispersal of water, steam, and molten core simulant materials on DCH loads under prototypic accident condi-tions and plant configurations. The results indicated that large amounts of coejected wstar reduced the DCH load by a small amount. Large amounts of debris were dispersed from the cavity to the upper dome (via the annular gap).
- 12. E Y wor DS/DESOR:PToR S fits: worss or pareses taet wan ems # rewerraen sa tocersag the repen. #
- 13. A V A' 6 Ain ti f v.T A f t ML N1 Unlimited conjected water
...ecv iii. ctassa. CAT os Combustion Engineering (CE)
, ra,,,,,,,
Direc; Containment Heating (DCH)
Unclassified high pressure melt ejection
, ra,,,,,u asvare accident Unclassified pressurized water reactor is. NuMeER oF PAGES Calvert Cliffs 16 PRICE j
wxceoav m ra n
i i
l l
}
t l
i i
i l
l i
l i
i l
l l
Printed on recycled paper Federal Recycling Program
. _..... -.. -.. ~..... - -
.~
UNITED STATES '
N STANDARD E -
NUCLEAR REGULATORY COMMISSION POSTAGE AND FEES PAfD USNRC WASHINGTON, DC 20E50001 PERMIT NO. G67.
4 OFFICIAL BUSINESS PENALTY FOR PRfVATE USE $300 12055513o531 1 laN1P3 US NRC-0ADM DIV FOIA ', DUBLICATIONS SVCS TPS-PDR-NUREG 2WFN-6E' WASHINGTON DC
?"555 L
L ti i
L I
m.
.'