ML010310397

NUREG/CR-6150, Vol. 5, Rev.
2
INEL-96/0422
SCDAP/RELAP5/MOD 3.3
Code Manual
Assessment of Modeling of
Reactor Core Behavior During Severe Accidents
Idaho National Engineering and Environmental Laboratory
U.S. Nuclear Regulatory Commission

Office of Nuclear Regulatory Research
Washington, DC 20555-0001
AVAILABILITY OF REFERENCE MATERIALS
IN NRC PUBLICATIONS
NRC Reference Material Non-NRC Reference Material
As of November 1999, you may electronically access Documents available from public and special technical
NUREG-series publications and other NRC records at libraries include all open literature items, such as
NRC's Public Electronic Reading Room at books, journal articles, and transactions, Federal
www.nrc.gov/NRC/ADAMS/index.html. Register notices, Federal and State legislation, and
Publicly released records include, to name a few, congressional reports. Such documents as theses,
NUREG-series publications; FederalRegister notices; dissertations, foreign reports and translations, and
applicant, licensee, and vendor documents and non-NRC conference proceedings may be purchased
correspondence; NRC correspondence and internal from their sponsoring organization.
memoranda; bulletins and information notices;
inspection and investigative reports; licensee event Copies of industry codes and standards used in a
reports; and Commission papers and their substantive manner in the NRC regulatory process are
attachments. maintained at
The NRC Technical Library
NRC publications in the NUREG series, NRC Two White Flint North
regulations, and Title 10, Energy, in the Code of 11545 Rockville Pike
FederalRegulations may also be purchased from one Rockville, MD 20852-2738
of these two sources.
1. The Superintendent of Documents These standards are available in the library for
U.S. Government Printing Office reference use by the public. Codes and standards are
P. 0. Box 37082 usually copyrighted and may be purchased from the
Washington, DC 20402-9328 originating organization or, if they are American
www.access.gpo.gov/sudocs National Standards, from
202-512-1800 American National Standards Institute
2. The National Technical Information Service 11 West 421 Street
Springfield, VA 22161-0002 New York, NY 10036-8002
www.ntis.gov www.ansi.crg
1-800-533-6847 or, locally, 703-805-6000 212-642-4900
A single copy of each NRC draft report for comment is

available free, to the extent of supply, upon written The NUREG series comprises (1) technical and
request as follows: administrative reports and books prepared by the
Address: Office of the Chief Information Officer, staff (NUREG-XXXX) or agency contractors
Reproduction and Distribution (NUREG/CR-XXXX), (2) proceedings of
Services Section conferences (NUREG/CP-XXXX), (3) reports
U.S. Nuclear Regulatory Commission resulting from international agreements
Washington, DC 20555-0001 (NUREG/IA-XXXX), (4) brochures
E-mail: DISTRIBUTION@nrc.gov (NUREG/BR-XXXX), and (5) compilations of legal
Facsimile: 301-415-2289 decisions and orders of the Commission and
Atomic and Safety Licensing Boards and of
Some publications in the NUREG series that are Directors' decisions under Section 2.206 of NRC's
posted at NRC's Web site address regulations (NUREG-0750).
www.nrc.gov/NRC/NUREGS/indexnum.html
are updated periodically and may differ from the last
printed version. Although references to material found
on a Web site bear the date the material was
accessed, the material available on the date cited may
subsequently be removed from the site.
i
DISCLAIMER: This report was prepared as an account of work sponsored by an agency of the U.S. Government.
Neither the U.S. Government nor any agency thereof, nor any employee, 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 publication, or represents that its use by such third
party would not infringe privately owned rights.
NUREG/CR-6150, Vol. 5, Rev. 2
INEL-96/0422
SCDAP/RELAP5/MOD 3.3
Code Manual
Assessment of Modeling of
Manuscript Completed: August 2000
Date Published: January 2001
Prepared by
L. J. Siefken, E. W. Coryell, E. A. Harvego,
J. K. Hohorst

P.O. Box 1625
Idaho Falls, ID 83415-3129
S. A. Arndt, NRC Project Manager
Prepared for
Division of Systems Technology
NRC Job Code W6095
ABSTRACT
The SCDAP/RELAP5 code has been developed for best-estimate transient simulation of light water
reactor coolant systems during a severe accident. The code models the coupled behavior of the reactor
coolant system and reactor core during severe accidents as well as large and small break loss-of-coolant
accidents, operational transients such as anticipated transient without SCRAM, loss of offsite power, loss
of feedwater, and loss of flow. The coolant system behavior is calculated using a two-phase model
allowing for unequal temperatures and velocities of the two phases of the fluid, and the flow of fluid
through porous debris and around bockages caused by reactor core damage. The reactor core behavior is
calculated using models for the ballooning and oxidation of fuel rods, the meltdown of fuel rods and
control rods, fission product release, and debris formation. The code also calculates the heatup and
structural damage of the lower head of the reactor vessel resulting from the slumping of reactor core
material. A generic modeling approach is used that permits as much of a particular system to be modeled
as necessary. Control system and secondary system components are included to permit modeling of plant
controls, turbines, condensers, and secondary feedwater conditioning systems.
The assessment of the MOD3.3 version of SCDAP/RELAP5 showed that its new models for severe
fuel damage and thermal hydraulic behavior result in calculated behavior of fuel assemblies under severe
accident conditions in agreement with measurements. The new models that were assessed using a wide
range of fuel damage experiments and the TMI-2 accident included; (1) integral diffusion model for
oxidation of fuel rod cladding, (2) stress-based model for failure of oxide layer containing melted metallic
cladding, (3) model for re-slumping of previously slumped and frozen cladding, (4) model for cracking of
cladding oxide layer during reflood conditions and affect of cracking on oxygen transport, (5) models for
flow losses and heat transfer in porous debris, (6) model for heat transfer in molten pool that has stratified
into oxidic and metallic parts, and (7) model for break-up ofjets of molten material slumping into a pool of
water. The assessment showed that the SCDAP/RELAP5 calculations of the axial distribution in oxidation
and meltdown of fuel assemblies and the behavior of fuel assemblies under reflood conditions were
improved by the new models in MOD3.3. The assessment showed that MOD3.3 calculates a more rapid
onset of severe fuel damage under severe accident conditions than that calculated by the previous version
of the code. The assessment also showed that MOD3.3 can analyze both severe fuel damage experiments
and nuclear power plants with a single set of models for the phenomena causing damage to fuel
assemblies.
iii NUREG/CR-6150-Rev 2, Vol 5

CONTENTS
A B STRA C T ................................................................................................................................ iii
FIGURES .................................................................................................................................. vii
TA B LES ....................................................................................................................................... ix
EXECUTIVE SUMMARY ......................................................................................................... xi

ACKNOWLEDGMENTS ........................................................................................................... xiii
1. INTRODUCTION .................................................................................................... 1-1

1.1 References ..................................................................................................... ... 1-1
2. Summary of New Models in MOD3.3 and Test Matrix for Assessment ................................... 2-1
2.1 Reference ...................................................................................................................... 2-6
3. Summary of Assessment Using Results of Severe Fuel Damage Tests .................................... 3-1
3.1 Assessment of Models for Temperature Behavior of Fuel Assemblies ........................ 3-1
3.1.1 Temperature Behavior of PWR Fuel Assemblies with Gradual Uncovery ........ 3-1
3.1.2 Temperature Behavior of PWR Fuel Assemblies with Complete Uncovery ...... 3-2
3.1.3 Temperature Behavior of BWR Fuel Assemblies with Complete Uncovery ..... 3-7
3.1.4 Temperature Behavior of Fuel Assemblies During Reflood ............................... 3-8
3.2 Assessment of Calculation of Hydrogen Production .................................................... 3-9
3.3 Assessment of Model for Ballooning of Fuel Rod Cladding ..................................... 3-10
3.4 Assessment of Models for Oxidation and Meltdown Using PIE Results ................... 3-11
3.5 Conclusions ................................................................................................................ 3-13
4. Summary of Assessment of Late-Phase Models Implemented into MOD3.3 ........................... 4-1
4.1 Assessment of Models for Flow Loss and Heat transfer in Porous Debris .................. 4-1
4.2 Assessment of Model for Heat Transfer In Stratified Molten Pool .............................. 4-4
4.3 Assessment of Model for Flow of Liquefied Material Through Porous Debris ........... 4-8
4.4 Assessment of Model for Molten Fuel-Coolant Interaction ....................................... 4-10
4.5 Assessment Using TMI-2 Accident ............................................................................ 4-12
4.6 Conclusions .................................................. 4-18
4.7 References .................................................................................................................. 4-20
APPENDIX A SCDAP/RELAP5/MOD3.3 Assessment ................................................................ A-1
V NUREG/CR-6150- Rev 2, Vol 5

FIGURES
Figure 3-1. Comparison of calculated and measured temperature at 2.37 m
elevation for FLHT-5 ........................................................................................ 3-2
Figure 3-2. Comparison of calculated and measured temperatures at 0.7 m
elevation for PBF SFD ST test .......................................................................... 3-3
Figure 3-3. Comparison of calculated and measured temperature at elevation
of 0.7 m for PBF SFD 1-1 test .......................................................................... 3-3
and 0.54 m elevation for PBF SFD 1-4 test ...................................................... 3-4
Figure 3-5. Comparison of calculated and measured temperature at 0.75 m
for CORA -5 test ............................................................................................ 3-4
Figure 3-6. Comparison of calculated and measured temperature at 0.75 m f
or C O RA -7 test................................................................................................. 3-5
Figure 3-7. Comparison of calculated and measured temperatures of unheated
fuel rod at elevation of 0.75 m for CORA- 13 test ............................................ 3-5
Figure 3-8. Comparison of calculated and measured fuel temperatures at 0.7 mn
elevation for PHEBUS B9+ test ........................................................................ 3-6
Figure 3-9. Comparison of calculated and measured cladding temperatures at 0.7 m
elevation for PHEBUS FPTO ............................................................................ 3-6
Figure 3-10. Comparison of calculated and measured temperatures of unheated rod at
elevation 0.75 m for CORA-17 test .................................................................. 3-7
Figure 3-11. Comparison of calculated and measured temperatures of fuel rod
at elevation of 0.254 m for DF-4 test ................................................................ 3-7
elevation for PBF SFD ST test .......................................................................... 3-8
Figure 3-13. Comparison of calculated and measured temperatures of unheated
fuel rod at elevation of 0.55 m for CORA-13 test ............................................ 3-9
Figure 4-1. Schematic of debris bed analyzed for assessment of flow
loss calculations ................................................................................................ 4-2
Figure 4-2. Schematic of BNL quenching experiment ........................................................ 4-3
Figure 4-3. Comparison of calculated and measured transient temperature
distribution in debris bed .................................................................................. 4-4
Figure 4-4. Molten pool system analyzed for assessment of MOD3.3 models for
heat transfer in stratified molten pool ............................................................... 4-5
Figure 4-5. Temperature histories of oxidic and metallic parts of molten pool
before and after stratification ............................................................................ 4-7
Figure 4-6. Ratio of heat flux to critical heat flux on external surface of lower
head before and after stratification of molten pool ........................................... 4-7
Figure 4-7. Schematic of system analyzed and nodalization of system for
calculation of melted material flowing through porous debris ......................... 4-8
Figure 4-8. Distribution of debris bed saturation and temperature after melting
of 30% of core plate (time of 1500 s) ............................................................... 4-9
vii NUREG/CR-6150-Rev. 2, Vol. 5

Figure 4-9. Ratio of heat flux to CHF on external surface at bottom
center of lower head ........................................................................................ 4-10
Figure 4-10. Schematic of FARO Test Arrangement ........................................................... 4-11
Figure 4-11. Comparison of calculated and measured pressure histories for
FARO Test L-08 .............................................................................................. 4-11
Figure 4-12. History of calculated maximum temperature in reactor core .......................... 4-15
Figure 4-13. History of calculated effective radius of molten pool ..................................... 4-15
Figure 4-14. Calculated hydrogen production during TMI-2 accident ................................ 4-16
Figure 4-15. MOD3.3 calculated temperature histories of fuel rods in upper part
of outer most fuel assemblies in reactor core .................................................. 4-17
Figure 4-16. MOD 3.3 calculated temperature history of location with porous debris (2.7 m
elevation of fourth ring of fuel assemblies) ................................................... 4-18
viii
NUREG/CR-6150-Rev. 2, Vol. 5
TABLES
Table 2-1. New models in MOD3.3 of SCDAP/RELAP5 ................................................. 2-2
Table 2-2. Test matrix for developmental assessment of SCDAP/RELAP5/MOD3.3 ...... 2-3
Table 2-3. Definition of test problems listed in Table 2-3 .................................................. 2-4
Table 3-1. Comparison of calculated and measured hydrogen production ........................ 3-9
Table 3-2. Comparison of calculated and measured ballooning behavior of fuel rods.... 3-10
Table 3-3. Comparison of calculated and measured axial distribution in
fuel rod oxidation ............................................................................................ 3-11
Table 4-1. Comparison of SCDAP/RELAP5 calculated flow losses for porous
debris with those calculated by other benchmarked models ............................. 4-2
Table 4-2. Comparison of SCDAP/RELAP5/MOD3.3 and benchmarked
model calculations for heat transfer behavior of stratified molten
pool w ith deep metallic layer ............................................................................ 4-5
Table 4-3. Comparison of SCDAP/RELAP5/MOD3.3 and benchmarked
model calculations for heat transfer behavior for case of shallow
metallic pool ...................................................................................................... 4-6
Table 4-4. Calculated timing and sequence of core damage progression ........................ 4-13
ix NUREG/CR-6150-Rev. 2, Vol 5
EXECUTIVE SUMMARY
The specific features of SCDAP/RELAP5/MOD3.3 are described in this five volume set of manuals
covering the theory, use, and assessment of the code for severe accident applications. This set replaces the
SCDAP/RELAP5/MOD3.2 Code Manuals, NUREG/CR-6150, Rev. 1.
The SCDAP/RELAP5 computer code is designed to calculate for severe accident situations the
overall reactor coolant system (RCS) thermal-hydraulic response, core damage progression, and reactor
vessel heatup and damage. The code was developed at the Idaho National Engineering and Environmental
Laboratory (INBEL) under the primary sponsorship of the Office of Nuclear Regulatory Research of the
U.S. Nuclear Regulatory Commission (NRC). The code is the result of merging the RELAP5 and SCDAP
codes. The models in RELAP5 calculate the overall RCS thermal-hydraulics, control system interactions,
reactor kinetics, and the transport of noncondensable gases. The RELAP5 code is based on a two-fluid
model allowing for unequal temperatures and velocities of the fluids and the flow of fluid through porous
debris and around blockages caused by reactor core damage. The models in SCDAP calculate the
progression of damage to the reactor core. These models calculate the heatup, oxidation and meltdown of
fuel rods and control rods, the ballooning and rupture of fuel rod cladding, the release of fission products
from fuel rods, and the disintegration of fuel rods into porous debris and molten material. The SCDAP
models also calculate the heatup and structural damage of the reactor vessel lower head resulting from the
slumping to the lower head of reactor core material with internal heat generation. Although previous
versions of the code have included the analysis of fission product transport and deposition behavior, this
capability has been removed from SCDAP/RELAP5, and the analysis of fission product behavior is now
performed using the detailed fission product code, VICTORIAa, in an effort to reduce duplicative model
development and assessment.
The SCDAP/RELAP5 code includes many generic component models from which general systems
can be simulated. The component models include fuel rods, control rods, pumps, valves, pipes, reactor
vessel, electrical fuel rod simulators, jet pumps, turbines, separators, accumulators, and control system
components. In addition, special process models are included for effects such as form loss, flow at an
abrupt area change, branching, choked flow, boron tracking, and noncondensable gas transport. The code
also includes a model for reactor kinetics.
Several new capabilities and improvements in existing capabilities were implemented into the
MOD3.3 version of SCDAP/RELAP5. The new capabilities include; (1) an integral diffusion method to
calculate oxygen and hydrogen uptake accounting in mechanistic manner for steam starvation and rapid
changes in temperature, (2) calculation of the relocation in the circumferential direction of melted metallic
cladding retained by the oxidic portion of cladding, (3) calculation of the re-slumping of cladding that
previously slumped and froze, (4) calculation of heat transfer in porous debris using correlations specific
to porous debris, (5) calculation of flow losses in porous debris locations based on Darcy's Law and
applying relative permeabilities and passabilities based on local debris conditions and volume fractions of
the liquid and vapor phases of the coolant, (6) calculation of oxidation of both intact and slumped cladding
under reflood conditions, (7) calculation of the heatup of the lower core structures and its interaction with
slumping core material, (8) calculation of the behavior of jets of core material penetrating into a pool of
a. N. E. Bixler, "VICTORIA2.0: A Mechanistic model for Radionuclide Behavior in a Nuclear Reactor

Coolant System Under Severe Accident Conditions," NUREG/CR-6131, SAND93-2301, December 1998.
xi NUREG/CR-6150-Rev. 2, Vol. 5
water, (9) calculation of the permeation of melted core plate material into porous debris in the lower head
of reactor vessel and affect of this permeation on lower head heatup, and (10) calculation of heatup of
lower head containing melted core material and accounting for whether the melted material is well-mixed
or stratified into oxidic and metallic pools. The improvements in existing modeling capabilities include;
(1) a semi-mechanistic stress-based model instead of a wholly empirical model for failure of the oxidic
portion of cladding retaining melted metallic cladding, and (2) more simplistic but accurate models for
calculating position, configuration, and oxidation of melted fuel rod cladding that slumped to a lower
location and froze. The MOD3.3 version of the code retains all of the capabilities of the previous version,
namely MOD3.2.
This volume, Volume 5, contains detailed comparisons of measurements and calculations for a wide
range of severe fuel damage experiments and for the TMI-2 accident. The comparisons showed that the
new models in SCDAP/RELAP5 for severe fuel damage and thermal hydraulic behavior result in
calculated behavior of fuel assemblies under severe accident conditions in agreement with measurements.
The assessment showed that the SCDAP/RELAP5 calculations of the axial distribution in oxidation and
meltdown of fuel assemblies and the behavior of fuel assemblies under reflood conditions were improved
by the new models in MOD3.3. The assessment showed that MOD3.3 calculates a more rapid onset of
severe fuel damage under severe accident conditions than that calculated by the previous version of the
code. The assessment also showed that MOD3.3 can analyze both severe fuel damage experiments and
nuclear power plants with a single set of models for the phenomena causing damage to fuel assemblies.
NUREG/CR-6150-Rev. 2, Vol. 5 xii

ACKNOWLEDGMENTS
Acknowledgments are made to those who made significant contributions to this and earlier versions
of SCDAP/RELAP5, B. D. Reagan, D. L. Knudson, F. Griffin, J. L. Rempe and M. Sohal. The authors also
acknowledge the former RELAP5 development team, specifically R. A. Riemike and J. Tolli for their
contributions to SCDAP/RELAP5.
The SCDAP/RELAP5 Program is indebted to J. Schaperow and S. Arndt of the U. S. Nuclear

Regulatory Commission, who were responsible for directing the overall program. Finally,
acknowledgment is made of those many code users who have been very helpful in stimulating correction
of code deficiencies and suggesting improvements.
.°.i NUREG/CR-6150-Rev. 2, Vol. 5

Introduction
1. INTRODUCTION
This report describes the developmental assessment of MOD3.3 of the SCDAP/RELAP5 code. 1
The developmental assessment achieved four objectives: First, it applied the code to a broad range of
problems in order to reveal and resolve errors not uncovered in its previous limited range of application.
Second, it verified that the code is capable of modeling the most important phenomena that occur during
severe accidents. Third, it verified that the new models implemented into MOD3.3 perform correctly in the
context of severe accidents in Light Water Reactors (LWRs). Fourth, it evaluated the capability of
MOD3.3 relative to the previous version of the code (MOD3.2). The developmental assessment of the
previous version of the code is described in Reference 1-2.
This report has three additional sections that correspond with the three major elements in the
developmental assessment effort. Section 2 summarizes the models that were not in the previous version of
the code and includes the test matrix devised to assess these new models. Section 3 summarizes the
assessment of the code using the results of a broad range of severe fuel damage experiments. Section 4
summarizes the assessment of the models implemented into MOD 3.3 for late-phase damage progression.
This section also evaluates the capability to calculate accident sequences such as the TMI-2 accident,
TMLB' accident sequences in a PWR, and a severe accident sequence in a BWR. The detailed results of
each developmental assessment test problem are presented in Appendix A.
1.1 References
1-1. L. J. Siefken, E. W. Coryell, E. A. Harvego and J. K. Hohorst, SCDAP/RELAP5/MOD3.3 Code

Manual, Volume 2: Modeling of Reactor Core and Vessel Behavior During Severe Accidents,
NUREG/CR-6150, Revision 2, INEL-96/0422, September 2000.
1-2. The SCDAP/RELAP5 Development Team, SCDAP/RELAP5/MOD3.2 Code Manual, Volume V.

DevelopmentalAssessment, NUREG/CR-6150 Rev. 1, INEL-96/0422, July 1998.
1-1 NUREG/CR-6150-Rev. 2, Vol. 5

Summary of New Models/Error Corrections in MOD3.3
2. Summary of New Models in MOD3.3 and Test Matrix for Assessment
Several new models were implemented into SCDAP/RELAP5/MOD3.3. The new models and the
capabilities added to the code by the models are summarized in Table 2-1. The new models include; (1) an
integral diffusion method for calculating oxidation and hydrogen uptake in fuel rod cladding, (2) a model
for the configuration of melted metallic cladding retained by an oxide layer and accounting for size of fuel
cladding gap, (3) a stress-based model for calculating failure of cladding oxide retaining melted metallic
cladding, (4) heat transfer correlations specific to porous debris, (5) a model for calculating flow losses in
porous debris that is based on Darcy's Law and relative permeabilities and passabilities (6) a core plate
model for calculating heatup of the core plate and its interaction with material that slumps from the core
region, (7) a model for calculating heat transfer for the case of a molten pool in the lower head stratified
into oxidic and metallic pools, (8) a model for calculating the movement of melted core plate material
through porous debris in the lower head and (9) a model for calculating the behavior of jets of core
material slumping into a pool of water in the lower head of the reactor vessel.
Several models in the previous version of the code were also improved. The calculations of the
oxidation of in situ and slumped cladding were extended to account for the effective surface areas for
oxygen uptake. Corrections were made to the models for calculating the internal gas pressure in fuel rods
and the conductance of the fuel-cladding gap.
MOD3.3 also has changes made to improve the maintenance and use of the code. Several features of
Fortran 90 programming were implemented into the code so as to reduce the size of files used for restart
and plotting. The Fortran 90 features also improve the efficiency of code maintenance operations. The
output of the code was extended to include tables summarizing the timing and spatial distribution of
various damage progression events.
A broad assessment was required of MOD3.3 due to the broad scope of the new models added to the
code and other changes made to the code. The calculations of almost every aspect of reactor core behavior
during the early-phase of a severe accident (phase before beginning of slumping of material to lower head)
have been impacted by the new models and other changes. The results of severe fuel damage tests provide
data for assessing the models calculating reactor core behavior during the early-phase of a severe accident.
Some new models have also been added for calculating late-phase behavior, such as lower head heat
transfer and the behavior of slumping core material. Although only a limited number of experiments have
been performed for investigation of the late-phase of a severe accident, data exist for assessing some of
these models, including the model for the behavior of jets slumping into a pool of water and the model for
heat transfer in a stratified molten pool. Other new late-phase models, such as the model for the movement
of melted core plate material through porous debris, can be assessed only by evaluating for internal
consistency.
A test matrix for MOD3.3 with twenty two test problems was compiled to achieve the objective of a
broad assessment of MOD3.3. The test matrix is shown in Table 2-2. The test matrix includes problems
ranging from the analyses of several severe fuel damage experiments to the analysis of the behavior of
nuclear power plants during a severe accident. Each row in Table 2-2 corresponds with a model or code
update unique to MOD3.3. Each column in Table 2-2 corresponds with a test problem for which a solution
by MOD3.3 provides information for assessing one or more models in MOD3.3. A short description of
each test problem is given in Table 2-3 and a more detailed description of most of the test problems is
given in Reference 2-2. For each column, an "X" is placed in the rows corresponding with models or code
updates that can be assessed with that test problem. The problems are listed from left to right in the order
of ascending fuel assembly damage. This order of solution is based on the logic that the calculation of late-
2-1 NUREG/CR-6150-Rev. 2, Vol 5

phase damage is dependent on the calculation of the early-phase damage progression, which in turn is
dependent upon the calculation of the rate of boiloff and convective cooling in fuel assemblies with
deficient cooling. So the assessment began with the bundle boiloff test problem, then advanced to test
problems solving experiments involving early-phase fuel damage progression and for which a significant
amount of measurements are available, and then advanced to test problems calculating late-phase damage
progression in LWRs. The broad range of the test problems resulted in an assessment of code updates such
as the conversion of the code from Fortran 77 to Fortran 90, and improvements in code architecture, which
involved changes in models whose mathematical basis was not changed but whose Fortran programming
was changed.
Table 2-1. New models in MOD3.3 of SCDAP/RELAP5.
Name Extensions in capability provided by model
Integral diffusion 2 -3 (1) Thinning of oxide layer at steam-starved locations, which was not
calculated by the previous parabolic kinetics model, (2) applicability to
rapidly changing temperature and fluid conditions, which was not the
case for the previous parabolic kinetics model, (3) representation of
acceleration in oxygen uptake caused by cracks occurring in the oxide
layer during reflood, which was represented in a more simplified fash
ion by the previous parabolic kinetics model, and (4) calculation of
hydrogen uptake and associated heatup and embrittlement, which was
not calculated by the previous parabolic kinetics model
Circumferential (1) Melted metallic cladding retained by oxide layer configured as seg
relocation 2-4 ment of annulus, which is consistent with PIE results of severe fuel
damage tests, instead of as axisymmetric cylinder.
Oxide layer failure2 -4 (1) Timing of failure of oxide layer retaining melted metallic cladding
based on stress in oxide layer instead of on empirical model based on
cladding temperature and fraction of oxidation. As a result, differences
between timing of oxide failure in LWR and in severe fuel damage
tests are consistent with differences in the basic mechanism causing
failure of the oxide layer. This basic mechanism is the stress induced in
the oxide layer by the temperature gradient in it, which is much larger
for severe fuel damage tests than for fuel bundles in a LWR, because in
severe fuel damage tests the nuclear heat generation rate is atypically
large to compensate for heat loss from the relatively small test bundles
(2) timing of failure of oxide layer based on whether it retains melting
metallic cladding with no space for expansion during phase change
from solid to liquid, as is the case for cladding oxidized on both outer
and inner surfaces. In the case of double-sided oxidation, expansion of
the metallic portion during the phase change causes stresses that fail
the oxide layer.
Heat transfer in porous (1) Convective heat transfer at locations with porous debris calculated
debris2-5 by correlations specific to debris, and which calculate heat transfer as
function of debris characteristics, debris temperature, and local coolant
conditions.
NUREG/CR-6150-Rev. 2, Vol. 5 2-2

Table 2-1. New models in MOD3.3 of SCDAP/RELAP5. (Continued)
Name Extensions in capability provided by model
Flow loss in porous (1) Flow losses at locations with porous debris calculated by correla
debris 2-5 tions specific to debris, and which calculate flow losses for liquid and
vapor phases as function of debris characteristics and local volume
fractions of liquid and vapor phases.
Core plate2- 7 (1) Calculation of heatup of core plate and extent of its interaction with
core material that slumps below bottom elevation of core.
Heat transfer in stratified (1) Heat transfer in molten pool in lower head calculated based on
pool2 -6 whether constituents of pool are well-mixed or stratified into oxidic
and metallic pools.
Material movement in (1) Calculation of movement of melted core plate through porous
porous debris 2 "8 debris in lower head, and affect of this movement on heatup of lower
head.
Fuel-Coolant Interaction (1) Calculation of behavior of jets of material slumping into pool of
(FCI)2 -9 water in lower head. Aspects of behavior calculated include; (1) frac
tion of jet that breaks up into drops and fraction that slumps to bottom
of lower head without break up, (2) convective and radiative heat trans
fer from drops resulting from break up, and (3) increase in system pres
sure caused by vapor generation, which in turn is caused by heat
transfer from drops, and (4) motion and settling to bottom of lower
head of drops resulting from break up of jets.
Table 2-2. Test matrix for developmental assessment of SCDAP/RELAP5/MOD3.3.
Test problem
Model or pbf11 cora5 pdropl sur

code update boff pbfst pbfl4 cora7 b9+ pdrop2 qstrat mltdb fci tmi
flht5 fptO coral3 pdrop3 bf
df4 coral7 debqn
Integral diffu- X X X X X
sion
Circumferen- X X X X
tial relocation
Oxide layer X X X X X
failure
Heat transfer X X
in porous
debris
2-3 NUTREG/CR-6150-Rev. 2, Vol 5

Table 2-2. Test matrix for developmental assessment of SCDAP/RELAP5/MOD3.3. (Continued)
Test problem
Model or pbfl 1 cora5 pdropl

code update boff pbfst pbfl4 cora7 b9+ pdrop2 qstrat mltdb fci tmi
flht5 fptO coral3 pdrop3 bf
df4 coral7 debqn
Flow loss in X X
porous debris
Core plate x
Heat transfer X X
in stratified
pool
Material move- X
ment in porous
debris
FCI X
Fortran 90 X X X X X X X X X X
Other code X X X X
improvements
Table 2-3. Definition of test problems listed in Table 2-3.
Abbreviation
of test problem Description of test problem
boff Boiloff of water in bundle of fuel rods with deficient cooling. Boiloff and
heatup of fuel rods as calculated with fuel rods represented by SCDAP heat
structures compared with that calculated with fuel rods represented by
RELAP5 heat structures.
pbfst PBF SFD ST test; nuclear heatup of PWR fuel assembly. Top of fuel assembly
uncovered due to coolant boiloff. Measurements of rate of boiloff, temperature
and damage of fuel rods, and H 2 production.
flht5 FLHT-5 test; nuclear heatup of full-length PWR fuel assembly. Gradual boiloff
of coolant. Measurements of rate of boiloff, temperature and damage of fuel
rods, and H2 production.
pbfl 1 PBF SFD 1-1 test; nuclear heatup of PWR fuel assembly. Steam flow through
assembly. Measurements of temperature and damage of fuel rods, and H2 pro
duction.

Table 2-3. Definition of test problems listed in Table 2-3. (Continued)
Abbreviation
pbf14 PBF SFD 1-4 test; nuclear heatup of PWR fuel assembly. Flow of steam and Ar
through assembly. Measurements of temperature and damage of fuel rods, and
H2 production.
fptO PHEBUS FPTO test; nuclear heatup of PWR fuel assembly. Flow of steam
through assembly. Measurements of temperature and damage of fuel rods, and
H2 production.
df4 ACRR DF-4 test; nuclear heatup of BWR fuel assembly. Flow of steam
through assembly.
cora5 FzK CORA-5 test; PWR fuel assembly with electrical heater rods.
cora7 FzK CORA-7 test; PWR fuel assembly with electrical heater rods. Flow of
steam and Ar through assembly. Measurements of temperature and damage of
fuel rods, and H2 production.
coral3 FzK CORA-13 test; PWR fuel assembly with electrical heater rods. Flow of
steam and Ar through assembly, followed by rapid reflood of hot assembly.
Measurements of temperature and damage of fuel rods, and H2 production.
coral7 FzK CORA-17 test; BWR fuel assembly with electrical heater rods. Flow of
steam and Ar through assembly, followed by rapid reflood of hot assembly.
Measurements of temperature and damage of fuel rods, and H2 production.
b9+ PHEBUS B9+ test; nuclear heating of PWR fuel assembly. Flow of steam
through assembly until assembly hot, then flow of He through assembly so as
to represent situation of extreme steam starvation. Measurements of tempera
ture and damage of fuel rods, and H2 production.
pdropl Steam flow through porous debris bed. SCDAP/RELAP5 calculations com
pared with pressure drop in steam obtained from benchmarked model described
in literature.
pdrop2 Flow of liquid water through porous debris bed. SCDAP/RELAP5 calculations
compared with pressure drop in water obtained from benchmarked model
described in literature.
pdrop3 Flow of two-phase water though porous debris bed with internal heat genera
tion. SCDAP/RELAP5 calculations compared with pressure drop in two-phase
water obtained from benchmarked model described in literature.
debqn BNL debris quenching experiment. Quenching from bottom of porous debris at
initial temperature of 775 IC Measurements of transient temperature of several
locations in debris bed.

Table 2-3. Definition of test problems listed in Table 2-3. (Continued)
Abbreviation
qstrat Molten pool stratified into oxidic and metallic parts, and contained in herni
spherical lower head; SCDAP/RELAP5 calculations compared with solution
obtained from literature of temperature of oxidic and metallic parts, and heat
flux at boundaries of molten pool
mltdb Theoretical situation of core plate melting on top of ceramic porous debris bed
in lower head of reactor vessel. Solution of the posed problem presents results
that can be evaluated for reasonableness and consistency.
fci ISPRA FARO-8 test. Slumping of mixture of molten U0 2 and ZrO2 into pool
of water in a vessel. Measurement of transient pressure response in vessel, tem
perature of plate supporting the settled slumped material, fraction of breakup of
the slumped material, and particle size distribution of the slumped material.
sur Theoretical situation of TMLB' accident in PWR. Solution of the posed prob
lem can be compared with solution of MOD3.2, and evaluated for reasonable
ness and consistency.
trni TMI-2 accident. Measurements of transient system pressure, heatup of parts of
reactor system piping, and final state of reactor core. Indirect measurement of
H2 production.
bf Unmitigated large break Loss of Coolant Accident in Browns Ferry BWR
2.1 Reference
2-1. The SCDAP/RELAP5 Development Team, "SCDP/RELAP5/MOD3.2 Code Manual - Damage

Progression Model Theory," NUREG/CR-6150, Vol. 2, Rev. 1, INEL-96/0422, July 1998.
2-2. The SCDAP/RELAP5 Development Team, "SCDP/RELAP5/MOD3.2 Code Manual

Developmental Assessment," NUREG/CR-6150, Vol. 5, Rev. 1, INEL-96/0422, July 1998.
2-3. L. J. Siefken, "Calculation of Hydrogen and Oxygen Uptake in Fuel Rod Cladding During Severe
Accidents Using the Integral Diffusion Method," INEEJEXT-99-00571, May 1999.
2-4. L. J. Siefken, "Models for the Configuration and Integrity of Partially Oxidized Fuel Rod Cladding
at High Temperatures," INEEIJEXT-98-0066 Rev. 1, January 1999.
2-5. L. J. Siefken, E. W. Coryell, S. Paik, and H. Kuo, "SCDAP/RELAP5 Modeling of Heat Transfer
and Flow Losses in Lower Head Porous Debris," INEIJEXT-98-00820, Rev. 2, July 1999.
2-6. L. J. Siefken and M. S. Sohal, "A Heat Transfer Model for a Stratified Corium-Metal Pool in the
Lower Plenum of a Nuclear Reactor'" INEL/EXT-99-00763, Rev. 1, January,2000..

2-7. E. W. Coryell and F. P. Griffin, "SCDAP/RELAP5 Lower Core Plate Model," INEEIJEXT-99
01029, September 1999.
2-8. L. J. Siefken, "SCDAP/RELAP5 Modeling of Movement of Melted Material Through Porous

Debris in Lower Head," INEELfEXT-98-01178 Rev. 2, October 1999.
2-9. M. S. Sohal and L. J. Siefken, "Design Report on Model Development for Molten Fuel Coolant
Interaction During Melt Slumping," INELEXT-97-00084, Rev. 1, March 1998.

Assessment of Modeling of Early Phase Phenomena
3. Summary of Assessment Using Results of Severe Fuel Damage

Tests
This section summarizes the assessment of SCDAP/RELAP5/MOD3.3 using the results of the ten
severe fuel damage experiments identified in Section 2. The experiment results provided information over
a broad range of conditions for assessing the MOD3.3 calculations of the following important early-phase
damage progression events; (1) heatup due to uncovery and oxidation of fuel assembly, (2) oxidation of
fuel assembly during uncovery and during fuel assembly reflood, (3) fuel rod ballooning and rupture, and
(4) fuel assembly meltdown. The assessment was performed by comparing measured results of the
experiments with the corresponding calculated results of MOD3.3 and MOD3.2. Appendices Al through
A7 describe in detail each of these ten experiments and the MOD3.3 and MOD3.2 calculations for these
experiments.
The comparison of both MOD3.3 and MOD3.2 calculations with the measured results provided
information for assessing the affects of the modeling improvements in MOD3.3. The modeling
improvements to be assessed included; (1) integral diffusion method for calculating the oxidation of fuel
rod cladding, (2) model for configuration of melted metallic cladding retained by oxide layer and
accounting for size of fuel-cladding gap, (3) stress-based model for calculating the time at which failure
occurs to a fuel rod oxide layer retaining melted metallic cladding, and (4) model for re-slumping of
previously frozen slumped cladding. These new modeling improvements impact the MOD3.3 calculations
of the heatup, oxidation, and meltdown of fuel assemblies. Experiment results applicable to the assessment
these modeling improvements included transient measurements of test fuel assembly temperatures and test
fuel assembly hydrogen production and Post-Irradiation Evaluations (PIE) of the axial distribution in
oxidation of the test fuel assembly and the location of blockages due to meltdown in the test fuel assembly.
3.1 Assessment of Models for Temperature Behavior of Fuel

Assemblies
The assessment of the MOD3.3 model for fuel rod heatup showed that the calculated rates of heatup
were in good agreement with the measured rates of heatup for most of the severe fuel damage tests. The
comparisons of calculated and measured rates of heatup were divided into four categories of test fuel
assemblies; (1) heatup of PWR fuel assembly with gradual uncovery of fuel assembly, (2) heatup of PWR
fuel assembly with complete uncovery of fuel assembly throughout the heatup period of the test, (3)
temperature behavior following reflood of hot PWR fuel assembly, and (4) heatup of BWR fuel assembly
with complete uncovery of fuel assembly throughout the heatup period of the test. The experiments in the
first category were; (1) FLHT-5, (2) PBF SFD ST, and (3) PBF SFD 1-1. The experiments in the second
category were; (1) PBF SFD 1-4, (2) CORA-5, (3) CORA-7, (4) CORA-13, (5) PHEBUS B9+, and (6)
PHEBUS FPTO. The experiments in the third category were; (1) PBF SFD ST test, and (2) PBF CORA-13
test. The experiments in the fourth category were; (1) DF-4, and (2) CORA-17. For each of these tests, the
calculated and measured rates of heatup were compared at the location 3/4 of the fuel pellet stack height
above the bottom of the fuel pellet stack. This location was generally near the first location in the test fuel
assembly to undergo rapid oxidation and a temperature measurement was generally available. Also, the
temperature at this location was usually not influenced by blockages caused by fuel rod meltdown or
influenced by cold structures above the test fuel assembly.
3.1.1 Temperature Behavior of PWR Fuel Assemblies with Gradual Uncovery
The MOD3.3 calculated rate of heatup of PWR fuel assemblies with a gradual boiloff of coolant was

in general agreement with the measured rate of heatup. The comparisons of calculated and measured
temperature are shown in Figure 3-1, Figure 3-2, and Figure 3-3 for the FLHT-5, PBF SFD ST, and PBF
SFD 1-1 tests, respectively. The MOD3.2 calculations are also shown in these figures. In general, the
temperature measurements fail before the fuel rod temperature exceeds 2000 K and before melting of the
fuel rod cladding has occurred. For the PBF SFD ST test and the PBF SFD 1-1 test, the calculated and
measured rate of heatup are in good agreement. For the FLHT-5 test, the calculated rate of heatup is
somewhat less than the measured rate of heatup. The reason for this underprediction of the rate of heatup
was not identified. As shown in Figure 3-1, Figure 3-2, and Figure 3-3, MOD3.2 calculated a temperature
behavior similar to MOD3.3. Since the measurements did not fail until after rapid oxidation of the fuel rod
cladding had begun, the generally good agreement in temperatures of the MOD3.3 calculations and the
measurements indicates that the integral diffusion model for oxidation is performing correctly.
3000.0 1i I -I- I I
Failure of measurement
Slumping of cladding
.................................
2000.0
E
Measured at 2.37 m S1000.0
---- .........
- Calculated, MOD3.3
. ..............Calculated, MOD3.2
0.0 1 1 1 1
0.0 1000.0 2000.0 3000.0 4000.0 5000.0
Time (s)
Figure 3-1. Comparison of calculated and measured temperature at 2.37 m elevation for FLHT-5.
3.1.2 Temperature Behavior of PWR Fuel Assemblies with Complete Uncovery
The MOD3.3 calculated rates of heatup of PWR fuel assemblies with a constant and complete
uncovery were found to be in good agreement with the measured rates of heatup. The calculated and
measured rates of heatup are compared in Figure 3-4, Figure 3-5, Figure 3-6, Figure 3-7, Figure 3-8, and
Figure 3-9 for the PBF SFD 1-4, CORA-5, CORA-7, CORA-13, PHEBUS B9+, and PHEBUS FPTO tests,
respectively. The MOD3.2 calculations are also shown in these figures. The agreement between calculated
and measured rates of heatup ranged from very close agreement for the CORA-7, CORA-13 and PHEBUS
B9+ tests to a significant overprediction of the rate of heatup for the PHEBUS FPTO test. For the PBF SFD
1-4 test, MOD3.3 calculated a relatively slow rate of heatup at the 0.74 m elevation in the period of 1800 s
to 1900 s due to steam starvation, while the measurement indicated a rapid increase in temperature at this
elevation during this period. At the 0.54 m elevation and where steam starvation was calculated to not
occur, the MOD3.3 calculated temperature and the measured temperature were in close agreement. Since
for these six tests the measurements of the temperatures did not fail until after rapid oxidation of the fuel
rod cladding had begun, the generally good agreement of the MOD3.3 calculations and the measurements
indicates that the integral diffusion model for oxidation is performing correctly.

3000.0
- 2000.0
:3
(D
CL
E
'- 1000.0
0.0
0.0 5000.0 10000.0 15000.0
Time (s)
Figure 3-2. Comparison of calculated and measured temperatures at 0.7 m elevation for PBF SFD ST
test
3000.0
Measured at 0.35 m
..............
Calculated, MOD3.3 ', .
MOD3.2
-Calculated, i( ,
- 2000.0
CL
E
'- 1000.0
0.0
0. 0 1000.0 2000.0 3000.0 4000.0
Time (s)
Figure 3-3. Comparison of calculated and measured temperature at elevation of 0.7 m for PBF SFD 1-1
test.

3000.0
Failure of measurements
-. 2000.0
a)
E
- 1000.0 v
Measured, 0.74 m
<>- O alculated, MOD3.3, 0.74 m
............. Measured, 0.54 m
-
V ---- VCalculated, MOD3.3, 0.54 m
0.0 10.0
1004 2000.0 3000.0
Time (s)
Figure 3-4. Comparison of calculated and measured temperatures at 0.74 m and 0.54 m elevation for
PBF SFD 1-4 test.
3000.0
2 2000.0
CD
E
g 1000.0
0.0 1
3000.0 3500.0 4000.0 4500.0 5 000.0
Time (s)
Figure 3-5. Comparison of calculated and measured temperature at 0.75 m for CORA-5 test.

3000.0
2000.0
1000.0
0.0
3000.0 4000.0 5000.0
Time (s)
Figure 3-6. Comparison of calculated and measured temperature at 0.75 m for CORA-7 test.
3000.0
£ 2000.0
a.
E
I-
1000.0
0.0
3000.0 4000.0 5000.0
Time (s)
Figure 3-7. Comparison of calculated and measured temperatures of unheated fuel rod at elevation of
0.75 m for CORA-13 test.

3000.0
Measured at 0.7 m
Calculated, MOD3.3
Calculated, ------------..
MOD3.2
2000.0
0
Q
Ca.
E
I-
1000.0
.7 Start of He flow
0.0
0. () 2500.0 5000.0 7500.0 10000.0
Time (s)
Figure 3-8. Comparison of calculated and measured fuel temperatures at 0.7 m elevation for PHEBUS
B9+ test.
3000.0
2000.0
2
0.•
E
0-
1000.0
0.0 1
0.0 5000.0 10000.0 15000.0 20000.0
Time (s)
Figure 3-9. Comparison of calculated and measured cladding temperatures at 0.7 m elevation for
PHEBUS FPTO.
NUREGICR-6150-Rev. 2, Vol. 5 3-6

3.1.3 Temperature Behavior of BWR Fuel Assemblies with Complete Uncovery
The MOD3.3 calculated rate of heatup of BWR fuel assemblies with a constant and complete
uncovery were found to be in generally good agreement with the measured rate of heatup. The calculated
and measured rates of heatup are compared in Figure 3-10 and Figure 3-11 for the CORA-17 and ACCR
DF-4 tests, respectively.
3000.0
2000.0
ca
0.
E
a)
I-1000.0
0.0 1
3000.0 4000.0 5000.0
Time (s)
Figure 3-10. Comparison of calculated and measured temperatures of unheated rod at elevation 0.75 m
for CORA-17 test.
3000.0
2000.0
W
E
1000.0
0.0
5000.0 6000.0 7000.0 8000.0
Time (s)
Figure 3-11. Comparison of calculated and measured temperatures of fuel rod at elevation of 0.254 m for
DF-4 test.

3.1.4 Temperature Behavior of Fuel Assemblies During Reflood
An acceleration in heatup after the start of reflood was calculated by MOD3.3 to occur at locations
that were hot and partially oxidized at the start of reflood. The PBF SFD ST test and the CORA-13 test
were the only tests in the set of assessment problems that involved the reflood of a hot fuel assembly. For
both of these tests, the location 1/2 of the fuel pellet stack height above the bottom of the fuel pellet stack
was hot and partially oxidized at the beginning of reflood. Locations higher in the test fuel assembly
generally did not undergo an acceleration in heatup during reflood due to the cladding at those locations
either being completely oxidized or the metallic part of the cladding having slumped before reflood. The
calculated heatup at the location 1/2 of the fuel pellet stack height above the bottom of the fuel pellet stack
for the PBF SFD ST test and the CORA-13 test are shown in Figure 3-12 and Figure 3-13, respectively.
For the PBF SFD ST test, the temperature was calculated to decrease to 1500 K in response to the decrease
in nuclear heat generation and the corresponding increase in water level, and then increase to 1800 K in
response to the increase in oxidation caused by cracking of the oxide layer on the fuel rod cladding. For the
CORA-13 test, the temperature was calculated to decrease to 1750 K due to the increased steam flow
caused by reflood and then increase to 2300 K due to the acceleration in oxidation resulting from cracking
of the cladding oxide layer. For both of these tests, the temperature measurements in the test fuel assembly
failed before the beginning of reflood and thus comparisons of calculated versus measured temperature
during the reflood period could not be made. Nevertheless, measurements of the rate of hydrogen
production indicated that an acceleration in oxidation occurred during the reflood period of these tests, and
thus the calculated increase in the rate of heatup was consistent with experiment results.
2000.0
Measured at 0.35 m
Calculated, MOD3.3
--------..---. Calculated, MOD3.2
£
a) 1000.0 F-
C)
E
(D
Beginning of decrease in
Cracking of oxide layer nuclear heat generation
due to nse in water level and rise in water level
0.0 O. I m . .
.0 5000.0 10000.0 15000.0

Time (s)
Figure 3-12. Comparison of calculated and measured temperatures at 0.5 m elevation for PBF SFD ST
test.

3000.0
£ 2000.0
E
(D
1000.0
0.0
3000.0 3500.0 4000.0 4500.0 5000.0
Time (s)
Figure 3-13. Comparison of calculated and measured temperatures of unheated fuel rod at elevation of
3.2 Assessment of Calculation of Hydrogen Production

The MOD3.3 calculation of hydrogen production was generally in good agreement with the
measured hydrogen production. Comparisons of calculated and measured hydrogen production are shown
in Table 3-1 for the nine experiments for which hydrogen production was measured. MOD3.2 calculations
are also shown in order to evaluate the affect of the improvements implemented into MOD3.3. The
MOD3.3 calculated hydrogen production was less than the measured hydrogen production for four of the
tests, greater than the measured hydrogen production for two of the tests, and within the measured
uncertainty in hydrogen production for the other tests. The deviations in calculated hydrogen production
ranged from being 50% too large for the PBF SFD 1-1 test to 15% too small for the PHEBUS FPTO test.
Table 3-1. Comparison of calculated and measured hydrogen production.
Hydrogen production (g)
Measured MOD3.3 MOD3.2

Experiment
Pre- Pre
Pre-reflood Total reflood Total re-
reflood Total
FLHT-5 300 ± 30 - 231 - 265
PBF SFD ST 150 ± 35 98 125 121 121

PBF SFD 1-1 64 ± 7 - 105 - 94
PBF SFD 1-4 86 ± 12 82 - 94

Table 3-1. Comparison of calculated and measured hydrogen production. (Continued)
Hydrogen production (g)
Experiment Measured MOD3.3 MOD3.2

Pre- Pre
Pre-reflood Total re-
reflood Total reflood Total
CORA-7 - 114 - 91 - 79
CORA-13a 97 161 78 158 145 175
PHEBUS B9+ - 39 - 46 - 49
PHEBUS FPT0 90 77 - 63
ACRR DF-4 40 - 29 -
a. Estimated hydrogen production from oxidation of structures above the test fuel rods was subtracted
from the measured hydrogen production to obtain estimate of measured hydrogen production from oxi
dation only of structures within the heated elevation interval of the test.
3.3 Assessment of Model for Ballooning of Fuel Rod Cladding
The assessment of the MOD3.3 model for the ballooning and rupture of fuel rod cladding showed a
possibility for improvement of this model. The results of the assessment of this event in severe accident
early-phase damage progression are summarized in Table 3-2. Only a few of the experiments provided
information on the timing of cladding rupture. For the FLHT-5, PBF SFD ST and PHEBUS FPTO tests, the
calculated and measured times of cladding rupture and temperature at rupture were in fairly good
agreement. For the PBF SFD 1-1 test, the calculated time of cladding rupture was significantly later than
the measured time of cladding rupture. For the PBF SFD 1-4 test, the calculated time of cladding rupture
was significantly earlier than the measured time of cladding rupture.
Table 3-2. Comparison of calculated and measured ballooning behavior of fuel rods.
Rupture temperature Time of cladding rupture
Test Measured MOD3.3 Measured MOD3.3
(K) (K) (s) (s)

FLHT5 1950a 1900 1010 1230
PBF SFD ST 1150-1200 1050 5850-6260 6015-6385
PBF SFD 1-1 1538-1632 1785-1798 1150-1280 1785-1795
PBF SFD 1-4 1720-1900 1360-1530 1300-1600 1100-1150
FPT0 973 1060- 1070 -- 7350-7410
a. Since the coolant pressure was greater than the internal gas pressure, the cladding failed by
compression instead of by ballooning.
NUREG/CR-6150-Rev. 2, Vol. 5 3-10

3.4 Assessment of Models for Oxidation and Meltdown Using PIE

Results
The integral diffusion model for oxidation and the stress-based model for breach of the cladding
oxide layer in MOD3.3 contributed to a significant improvement in the calculated oxidation and meltdown
of fuel rod cladding. The comparison of calculated and measured axial distribution in oxidation of fuel rod
cladding provided a basis for assessing these models. The timing of cladding slumping is as important of a
factor in determining the total extent of oxidation of the cladding as the rate of oxidation. The calculated
and measured axial distribution cladding oxidation are compared in Table 3-3 for the six severe fuel
damage tests for which this information was available. The calculations of both MOD3.3 and MOD3.2 are
shown in the table. In this table, elevation is expressed as fractional height, wherein a fractional height of
0.85 indicates the elevation of the point located the distance of 85% of the height of the fuel pellet stack
above the bottom of the fuel pellet stack. As shown in Table 3-3, the almost complete oxidation of the
upper half of the cladding of the test fuel rods occurred for the three tests involving a slow boiloff of
coolant or a slow heatup in steam flow, namely FLHT-5, PBF SFD ST, and PHEBUS B9+. For these tests,
the calculated and measured fraction of oxidation are in good agreement. Only a small amount of slumping
of the metallic part of the cladding was measured to occur for these tests, and only a small amount of
slumping of the metallic part of the cladding was calculated to occur. For the PBF SFD 1-1 test, PBF SFD
1-4 test, and CORA-13 test, the extent of oxidation of the cladding in the upper half of the test fuel rods
was limited by meltdown of the cladding and the extent of oxidation of the lower part of the test fuel
assembly was accelerated by the meltdown of the cladding. MOD3.2 tended to overpredict the extent of
oxidation in the upper half of the test fuel rods and underpredict the extent of oxidation of the fuel rod
cladding in the lower half of the test fuel rods. While the MOD3.3 calculations of the axial distribution in
oxidation for these three tests was not in close agreement with the measured results, nevertheless the
MOD3.3 calculations were an improvement over the MOD3.2 calculations. For example, at the fractional
elevation of 0.85 for the PBF SFD 1-4 test, the measured, MOD3.3 calculated, and MOD3.2 calculated
values of fraction of cladding oxidation were 0.22, 0.24, and 0.86, respectively. For a second example, at
the fractional elevation of 0.45 for the CORA-13 test, the measured, MOD3.3 calculated, and MOD3.2
calculated values of fraction of cladding oxidation were 1.0, 1.0, and 0.10, respectively. For a third
example, at the fractional elevation of 0.25 for the PBF SFD 1-4 test, the measured, MOD3.3 calculated,
and MOD3.2 calculated values of fraction of cladding oxidation were 1.0, 0.63, and 0.11, respectively. The
better agreement of calculations with measurements was achieved primarily due to implementation of the
following models into MOD3.3; (1) integral diffusion model for oxidation, (2) stress-based model for
breach of the cladding oxide layer, and (3) model for re-slumping of previously frozen slumped material.
Table 3-3. Comparison of calculated and measured axial distribution in fuel rod oxidation
Elevation Fraction of cladding oxidized
(Fractional height) FLHT-5 PBF PBF PBF CORA- PHEBUS

SFD ST SFD 1-1 SFD 1-4 13 B9+
0.85 Measured 1.0 1.0 0.20a 0.22a -<1.0
MOD3.3 1.0 1.0 0.16a 0.24a 0.30a
MOD3.2 - - 0.34 0.86 0.43a -
3-11 NUREG/CR-6150-Rev. 2, Vol. 5

Table 3-3. Comparison of calculated and measured axial distribution in furl rod oxidation
Elevation Fraction of cladding oxidized
(Fractional height) FLHT-5 PBF PBF PBF CORA- PHEBUS

SFD ST SFD 1-1 SFD 1-4 13 B9+
0.75 Meas. 1.0 1.0 O.lla 0.37a 1.00 0.75
MOD3.3 1.0 1.0 0.18a 0.22a 0.30a 0.90
MOD3.2 - - 0.46 0.26a 0.20a 0.52
0.65 Meas. 1.0 1.0 0.45 - 1.00 1.00

MOD3.3 0.43a 1.0 0.25a 0.33 1.00 0.99
MOD3.2 - - 0.38 - 0.36a 1.00
0.55 Meas. 1.0 1.0 - <0.30a 1.00 0.70

MOD3.3 1.0 1.0 0.36a 0.67 1.00 1.00
MOD3.2 - - - 0.37 0.16a 1.00
0.45 Meas. 1.0 1.0 0.65 - 1.00 0.40

MOD3.3 1.0 1.0 0.38a 0.68 0.69 0.70
MOD3.2 - - 0.36 - 0.l0a 0.86
0.35 Meas. 1.0 1.0 0.25 0.50a 0.60 0.25
MOD3.3 0.97 1.0 0.78 0.67 1.00 0.49

MOD3.2 - - 0.16 0.52 0.70 0.71
0.25 Meas. 0.0 1.0 1.00 1.00 0.07 0.10
MOD3.3 0.04 0.03 0.80 0.61 1.00 0.30
MOD3.2 - 0.51 0.11 0.43 0.21
0.15 Meas. 0.0 0.03 1.00 1.00 0.0 0.03
MOD3.3 0.0 0.0 0.01 0.03 1.00 0.07
MOD3.2 - - 0.0 0.02 0.17 0.06
0.05 Meas. 0.0 0.0 0.60 0.10 0.0 0.0
MOD3.3 0.0 0.0 0.01 0.0 0.03 0.02
MOD3.2 - - 0.0 0.0 0.21 0.02
a. Metallic part of cladding slumped.
NUREG/CR-6150-Rev. 2, Vol. 5 3-12

3.5 Conclusions
The assessment of MOD3.3 showed that its calculations of fuel rod behavior during severe accident
conditions were in general. agreement with measured behavior. The assessment was performed using
various results from a wide range of severe fuel damage tests. The MOD3.3 calculations of fuel rod
meltdown and the axial distribution in fuel rod oxidation were significantly improved over MOD3.2
calculations. MOD3.2 tended to overpredict the extent of oxidation of fuel rod cladding in the upper part of
a fuel assembly due to the lack of a model for the dissolution of the oxide layer in a steam-starved region.
MOD3.2 also tended to underpredict the extent of oxidation of fuel rod cladding in the bottom part of a
fuel assembly due to the lack of a model to calculate the re-slumping of previously slumped cladding and a
tendency to underpredict the rate of oxidation in a steam-rich region. The MOD3.3 calculations of the
behavior of fuel assemblies under reflood conditions was also improved over the MOD3.2 calculations.
The improvements in the calculation of fuel assembly meltdown, axial distribution in oxidation, and
reflood behavior were the result of implementing the following models into MOD3.3; (1) integral
diffusion model for oxidation and dissolution, (2) model for calculating the timing of failure of an oxide
layer that is based on stress and accounts for the affect of oxide dissolution on stress, (3) model for re
slumping of previously slumped cladding, and (4) models for cracking of oxide layer during reflood
conditions and affect of cracking on oxygen transport.
3-13 NUREG/CR-6150-Rev. 2, Vol. 5

Assessment of Modeling of Late Phase Phenomena
4. Summary of Assessment of Late-Phase Models Implemented into

MOD3.3
This section summarizes the assessment of models implemented into MOD3.3 for improving
SCDAP/RELAP5's calculation of the behavior of a reactor core during the late-phase of a severe accident.
Five models were implemented into MOD3.3 for this purpose; (1) model for flow losses in porous debris,
(2) model for heat transfer in porous debris, (3) model for heat transfer from molten pool that has stratified
into oxidic and metallic parts, (4) model for flow of liquefied core plate material through porous debris in
the lower head of the reactor vessel, and (5) model for break-up of jets of molten core material penetrating
into a pool of water in the lower head. These five models were assessed using experimental results and
numerical solutions available from the literature. Appendices AS through All describe in detail the
assessment of these five models. These models along with the early-phase models were also assessed by
calculating the TMI-2 accident, which is presented in Appendix A12. Models in MOD3.3 for heat transfer
from a homogeneous molten pool and for structural failure of the crust supporting a molten pool are the
same as in MOD3.2, so these models were not assessed for MOD3.3. The assessments of these models are
described in the developmental assessment report for MOD3.2. 4
A summary of the assessment of the five late-phase models implemented into MOD3.3 and the TMI
2 calculation are summarized in the following sections of the report. The assessment of the models for
flow loss and heat transfer in porous debris is described in Section 4.1. The assessment of the model for
heat transfer from a stratified molten pool is described in Section 4.2. The assessment of the model for
flow of liquefied material through porous debris is described in Section 4.3. The assessment of the model
for break-up of jets of slumping molten material is described in Section 4.4. The assessment of the code
using benchmarking data obtained during and after the TMI-2 accident is described in Section 4.5.
Conclusions are presented in Section 4.7 and the references are presented in Section 4.8.
4.1 Assessment of Models for Flow Loss and Heat transfer in Porous
Debris
The models for flow loss and heat transfer in porous debris were assessed by comparisons of
SCDAP/RELAP5/MOD3.3 calculated thermal hydraulic behavior of debris with measurements and the
calculations of benchmarked models presented in the literature. The details of the assessment are described
in Appendix A8.
The assessment of the flow loss models showed that MOD3.3 is calculating in an acceptable manner
the flow losses in porous debris. The MOD3.3 flow loss models were assessed by comparing MOD3.3
calculated flow losses with the calculated flow losses of other benchmarked models. A schematic of a
system analyzed for the assessment is shown in Figure 4-1. The system consisted of a 1-m deep bed of
porous debris with a porosity of 0.4 and particles with a diameter of 3 mm. The assessment was performed
for coolant conditions ranging from subcooled liquid to superheated steam. The assessment included
coolant conditions involving an axial distribution in two-phase coolant conditions typical of a covered
debris bed resulting from a severe accident in a LWR.4"2 A summary of the assessment of the flow loss

models is shown in Table 4-1. For all the cases analyzed, the coolant pressure was 6.9 MPa. The MOD3.3
calculated flow losses are in approximate agreement with the values calculated by benchmarked models.
Table 4-1. Comparison of SCDAP/RELAP5 calculated flow losses for porous debris with those
calculated by other benchmarked models
Superficial Pressure drop due to flow losses
velocity at (Pa)
Coolant conditions bottom of bed
(mis) MOD3.3 Benchmarked

4 4.3
moe 4-2, 3
model
superheated steam 0.132 1.03x10 3 1.14x10 3
subcooled liquid 1.39x10-2 1.00x10 3 1.28x10 3

two-phase water in debris bed with 1.47x 10-2 10.6x10 3 12.6x10 3
internal heat generation
4 4 4 •Fluid leaving}debris bed
00 •e) 00 , Insulated, ri. ;id surface
eo o•@
0 0 ( (D \\ Debris withi
a 0 0 and particle Porosity of 0.4
diameter of 3 mm
®0
e0 0o a @0 -"
m
0 f0u v
4 4 4 4 Defined fluid veloity
Figure 4-1 .Schematic of debris bed analyzed for assessment of flow loss calculations.

The MOD3.3 modeling of heat transfer in porous debris was assessed using the results of a BNL
debris experiment involving the quenching from the bottom of a hot and porous debris bed.4-4 During this
experiment, measurements were obtained of the transient temperature at various locations within the
debris bed. Since the transient temperatures were a function of flow losses, this experiment in an indirect
manner also provided an assessment of flow loss models. A schematic of the experiment is shown in
Figure 4-2.
"0.108m
Fixed screen 0 ýj Heated stainless steel tube
(D0 (2 Stainless steel spherical particles;

e 00 diameter = 3.175 mm
(•) (porosity = 0.39
0.422 m initial temperature = 775 K
QOQ
Fixed screeno- Water;
velocity = 4.42 mm/s
pressure= 0.1 MPa
temperature = 373 K
Figure 4-2. Schematic of BNL quenching experiment.
The calculated transient temperatures in the debris bed were in general agreement with the measured
transient temperatures. The calculated and measured transient temperatures at the elevations of 0.025 m
and 0.24 m are compared in Figure 4-3. The elevation of 0.025 m is near the bottom of the debris bed,
where reflood began, and the elevation of 0.24 m is slightly above the midplane of the debris bed. The
overprediction of the temperature at the 0.24 m elevation in the period of 40 s to 50 s is considered to be
due to two-dimensional hydrodynamic behavior, wherein the liquid phase moved up along the wall,
formed a pool at the top, and then some of the water flowed down the center region of the debris bed. Since
such behavior is not expected in a debris bed resulting from a severe accident in a LWR due to the much
larger size of such a debris bed and due to such a debris bed not having cold walls, the calculated and
measured temperature comparisons indicate that MOD3.3 is calculating in an acceptable manner the heat
transfer in porous debris.

800.0
700.0
" 600.0 ........ -
E
a 500.0
- Calculated,0.025 m
Measured,0.025 m
- --- Calculated,0.24 m-
I - -- Measured,0.24 m
400.0
300.0,0.0 50.0 , 100.0

Time (s)
Figure 4-3. Comparison of calculated and measured transient temperature distribution in debris bed
4.2 Assessment of Model for Heat Transfer In Stratified Molten Pool

The model for heat transfer in a stratified molten pool was assessed by comparing its steady state
calculations with those of a model4- 5 benchmarked using the results of experiments with simulant
materials. The model was also assessed for internal consistency by examining its calculations of the
transient analysis of a stratified molten pool. The lack of experimental data for LWR materials and
geometry excluded the assessment of the model by direct comparison with experimental results. A detailed
description of the assessment is described in Appendix A9.
A schematic of the system analyzed for assessment purposes is shown in Figure 4-4. The system
consisted of a lower head of a reactor vessel containing a pool of molten reactor core material. The lower
head was submerged in a pool of water. The lower oxidic part of the molten pool was composed of a
mixture of U0 2 and ZrO 2 with a liquidus temperature of 3000 K. The upper metallic part of the molten
pool was composed of stainless steel with a liquidus temperature of 1630 K. The oxidic part of the molten
pool had a volumetric heat generation rate of 1.4 MW/m 3 and the metallic part did not have any internal
heat generation. No heat transfer was assumed to occur from the top surface of the metallic pool. The
boundary conditions and the liquidus temperatures were fixed to these values in order to be consistent with
the calculations of the benchmarked model with which the MOD3.3 model was being compared.
The MOD3.3 calculated steady state heat transfer behavior was similar to that calculated by the
benchmarked model. The MOD3.3 and the benchmarked model calculations for the heat transfer from the
stratified molten pool are compared in Table 4-2. The elevation on the external surface of the lower head
with the minimum margin to C-F may be the location with the maximum heat flux from the molten pool
to the lower head. For stratified molten pools, the maximum heat flux into the lower head generally occurs
near the interface of the metallic pool with the lower head; so the most important result in the calculations

Centerline of reactor vessel
Adiabatic surface
Molten metallic
layer
ier head of
U
ctor vessel
Molten oxidic pool
Figure 4-4.Molten pool system analyzed for assessment of MOD3.3 models for heat transfer in stratified
molten pool.
is the sideward heat flux into the lower head at this elevation. The MOD3.3 and the benchmarked model
calculations for the sideward heat flux at this interface were 0.66 MW/m 2 and 0.55 MW/m 2 , respectively.
Several other aspects of heat transfer behavior are also compared in Table 4-2. The MOD3.3 and
benchmarked model calculations of the heat flux into the lower head at the elevation of the interface of the
oxidic and metallic molten pools were 0.83 MW/m 2 and 0.77 MW/m 2 , respectively. Both models
calculated that this elevation was the elevation of the maximum heat flux into the lower head from the
molten pool. The MOD3.3 and the benchmarked values of the bulk temperature of the oxidic part of the
molten pool were 3102 K and 3110 K, respectively. The MOD3.3 and benchmarked model calculations for
the thickness of the crust on top of the molten pool were 3.7 mm and 5.5 mm, respectively.
Table 4-2. Comparison of SCDAP/RELAP5/MOD3.3 and benchmarked model calculations for heat
transfer behavior of stratified molten pool with deep metallic layer.
Aspect of heat transfer MOD3.3 Benchmarked

model
Sideward heat flux of metallic pool (MW/rn 2 ) 0.66 0.55

Sideward heat flux at elevation of interface of oxidic and 0.83 0.77
metallic pools (MW/m 2 )
Downward heat flux at bottom center of oxidic pool (MW/m 2 ) 0.06 0.13
Bulk temperature of oxidic pool (K) 3102 3110
Bulk temperature of metallic pool (K) 1691 1680
Thickness of crust on top of metallic pool (umm) 3.7 5.5

MOD3.3 and the benchmarked model also calculated similar heat transfer for a molten pool with a
shallow metallic layer. For the case of a shallow metallic pool, the depth of the metallic pool was 0.22 m
instead of 0.80 m and the depth of the oxidic pool was 1.18 m instead of 1.60 m. The other parameters
were the same as for the deep metallic pool case. The MOD3.3 and benchmark model calculations for this
case are compared in Table 4-3. Both models calculated a significantly greater heat flux into the lower
head from the metallic pool for this case than for the case of a deep metallic pool. The MOD3.3 and
benchmarked model calculations for this heat flux were 1.91 MW/m 2 and 1.22 MW/m 2 , respectively.
Table 4-3. Comparison of SCDAP/RELAP5/MOD3.3 and benchmarked model calculations for heat
transfer behavior for case of shallow metallic pool
Benchmarked
MOD3.3 model
Aspect of heat transfer model
Sideward heat flux of metallic pool (MW/m 2 ) 1.91 1.22

Sideward heat flux at elevation of interface of oxidic and 0.55 0.50
metallic pools (MW/m2 )
Downward heat flux at bottom center of oxidic pool (MW/M2 ) 0.04 0.14
Thickness of crust on top of metallic pool (mm) 4.6 6.2
A temporary decrease in the margin to CHF may occur during the transition of a molten pool from
the condition of being well-mixed to being stratified. The MOD3.3 transient analysis of a well-mixed
molten pool that switches to a stratified molten pool was performed for the case of a molten pool with a
deep metallic pool after stratification. Except for being initially well-mixed, the molten pool parameters
were the same as for the case presented in Table 4-2. The molten pool was assumed to be well-mixed for
the first 2000 s of the analysis. At 2000 s, the metallic part of the molten pool was assumed to segregate
from the oxidic part and form a metallic pool on top of the oxidic pool. The MOD3.3 calculated transient
temperatures of the oxidic and metallic parts of the molten pool are shown in Figure 4-5. After segregation,
the metallic pool was calculated by MOD3.3 to cool from 2970 K to 1690 K in about 500 s. The margin to
CHF on the external surface at the elevation of the metallic pool was calculated to be a minimum at the
time about midway through the cool down period of the metallic pool. Figure 4-6 is a plot as a function of
time of the MOD3.3 calculated ratio of heat flux to critical heat flux for the external surface of the lower
head at the elevation of the interface of the oxidic and metallic parts of the molten pool. As shown in the
figure, at the instant of stratification, the heat flux ratio was calculated to be 0.26. After stratification, the
heat flux ratio was calculated to increase to 0.33 in 320 s, and then gradually decrease to a steady state
value of 0.26. MOD3.3 applied the Cheung model4 "6 for the critical heat flux model for the external
surface of the lower head.
NUREG/CR-6150-Rev. 2, Vol. 5 4--6

4000.0
3000.0
2000.0
E
0)
F
1000.0
0.0
0.0 1000.0 2000.0 3000.0 4000.0 5000.0
Time (s)
Figure 4-5. Temperature histories of oxidic and metallic parts of molten pool before and after
stratification.
0.5
0.4
0.3
.2
"U
Ib
0.2
Time at which molten pool stratified

0.1
S I,
nln
1000.0 2000.0 3000.0 4000.0 5000.0
Time (s)
Figure 4-6. Ratio of heat flux to critical heat flux on external surface of lower head before and after
stratification of molten pool.

4.3 Assessment of Model for Flow of Liquefied Material Through

Porous Debris
The assessment of the model for flow of liquefied core plate material through porous debris indicated
that this model was functioning in MOD3.3 in an acceptable manner. No experimental results or
calculations of a benchmarked model were available for assessment. So the assessment was limited to a
test of the operation of the model and to an evaluation of the internal consistency of the model and the
reasonableness of its calculations. A detailed description of the assessment is presented in Appendix A10.
The assessment was performed by analysis of a system composed of porous debris with liquefied
core plate material permeating through the porous debris. The porous debris was supported by the lower
head of a reactor vessel. A schematic of the system analyzed is shown in Figure 4-7. The porous debris bed
had a porosity of 0.4 and was composed of U0 2 particles with a diameter of 3 mm. The debris bed had a
volumetric heat generation rate of 1 MW/m 3 . The depth of the debris bed along the center line of the
reactor vessel was 0.9 m and its initial temperature was 1727 K. The core plate above the debris bed was
melting due to heat transfer by radiation from the top surface of the debris bed. The core plate was
composed of stainless steel and had a mass of 25,000 kg. The outer surface of the lower head supporting
the debris bed was in contact with a pool of water. The dashed lines in Figure 4-7 represent the finite
elements and control volumes used in the numerical solution to represent the debris, lower head, and the
core plate material permeating through the debris.
Centerline of
reactor vessel Radiation
heat transfer
Core plate
,Melted core
plate material
1 Top surface
S//// of debris
0.9mL,
Outer surface
7 ,of lower head
Inner surface
"oflower head
Porous debris bed composed
of (U, ZrO 2 ) particles
Figure 4-7 Schematic of system analyzed and nodalization of system for calculation of melted material,
flowing through porous debris.

The melted core plate material was calculated to permeate to within 0.05 m of the bottom surface of
the lower head and then freeze. The distribution of core plate material within the debris bed along its center
line at the time of 1500 s, when 30% of the core plate had melted, is shown in Figure 4-8. The temperature
distribution in the debris bed is also shown in Figure 4-8. The debris near the inner surface of the lower
head was calculated to be significantly cooler than the debris located several cm above the inner surface.
The permeating core plate material was calculated to be blocked by a frozen crust of previously frozen core
plate material. The crust caused a region saturated with core plate material to accrete near the inner surface
of the lower head.
0.6 2500.0
- -= Saturated
region
.0"\-2000.
= 0.4 2000.0
""D Temperature N
Saturation
02ocation S0.2 of -1500.0-- C

C.
frozen crust 150.
Bottom of
debris bed
0.0, I , I , 1 1000.0
0.0 02 0.4 0.6 0.8 1.0
Distance below top surface of debris bed (m)
Figure 4-8. Distribution of debris bed saturation and temperature after melting of 30% of core plate (time
of 1500 s).
The permeating core plate material was calculated to not significantly decrease the margin to CHF at
the external surface of the lower head. The MOD3.3 calculated ratio of heat flux to critical heat flux for the
external surface of the lower head at its bottom center is shown in Figure 4-9. The critical heat flux was
calculated using the Cheung model.4"6 The figure also shows the heat flux ratio for the case of no core plate
material permeating through the debris bed. At 2500 s, the heat flux ratios for the case of permeating core
plate material and no permeating material were 0.18 and 0.16, respectively. If the core plate material had
permeated to the inner surface of the lower head, the heat flux ratio may have increased significantly due to
filling in of voids at the interface of the debris and lower head. The effective thermal conductivity of the
region saturated with core plate material was a factor of four greater than that for a region with no core
plate material. This increase in effective thermal conductivity was the main contributor to the increase in
heat flux ratio.

Assessment of Modeling of Late Phase Phenomena'
0.40 ' I *I I
Permeation of core plate material

0.30 ......... No permeation of core plate material
LL
"r 0.20
- ---- ----------------------
-- -------------------------
Q 0.10
0.00
-0.10 JI
500.0 1000.0 1500.0 2000.0 2500.0
Time (s)
Figure 4-9Ratio of heat flux to CHF on external surface at bottom center of lower head.
4.4 Assessment of Model for Molten Fuel-Coolant Interaction
The assessment of the Fuel-Coolant Interaction (FCI) model in MOD3.3 indicated that this model
was representing the basic features of FCI and possibly needs improvements for representation of some of
the finer features of FCI. The model was assessed using the results of two FCI experiments and by testing
the operation of the model in the context of severe accident analysis. A detailed description of the
assessment is presented in Appendix All. The model has application to the calculation of the extent of
break-up of jets of molten material slumping from the core region into a pool of water in the lower head of
the reactor vessel. The extent of break-up has a significant influence on the rate of heat transfer from the
slumped material to the fluid in contact with it. The extent of break-up also has an influence on the rate of
heat transfer from the slumped material to the lower head supporting the slumped material.
The MOD3.3 calculated transient pressure in a vessel with FCI was in general agreement with the
measured transient pressure. The FARO L-08 experiment4"7obtained a measurement of the transient
pressure in a vessel in which FCI occurred. This experiment was used to assess the FCI model in MOD
3.3. A schematic of this experiment is shown in Figure 4-10. In this experiment, a molten mixture of 80
wt% U0 2 and 20 wt% ZrO 2 was injected into a 1.0 m deep pool of water at an initial temperature of 536 K
and a subcooling of 10 K. The initial pressure in the test vessel was 5.8 MPa. The MOD3.3 calculated and
measured pressure history in the reactor vessel are compared in Figure 4-11. The calculated and measured
pressure histories in the test vessel are similar in shape but the calculated rate of increase of pressure is
about 50% greater than the measured rate of increase in pressure. The maximum calculated increase in
vessel pressure was about 12% greater than the measured maximum increase in pressure. The MOD 3.3
calculated and measured extents of break-up of the jet of slumping material were 70% and 68%,
respectively. The MOD3.3 calculated and measured maximum increases in temperature of the bottom
plate of the test vessel on which the slumped material settled were 275 K and 352 K, respectively.
NUREG/CR-6150-Rev. 2, Vol. 5 4.-10

.. Molten material
Nozzle with
hinged-flap
for release
Water
SDebris catcher
a-Bottom plate
Figure 4-10. Schematic of FARO Test Arrangement.
8.0
7.5
,-, 7.0
a(L
2 6.5
"CO
Cu
6.0
5.5
5.0
0.0 1.0 2.0 3.0 4.0 5.0
Time (s)
Figure 4-11. Comparison of calculated and measured pressure histories for FARO Test L-08.
4-11 NUREG/CR-6150-Rev. 2, Vol. 5

A significant discrepancy in the MOD3.3 calculation was the calculated size of the particles resulting
from FCI; the MOD3.3 and measured calculated sizes of particles were 0.56 mm. and 3.8 mm, respectively.
One reason for this discrepancy may be the use of an incorrect value for the interfacial tension of a mixture
of U0 2 and ZrO 2 in contact with water and steam. The particle size is calculated to be proportional to this
material property for which there is large amount of uncertainty in value.
4.5 Assessment Using TMI-2 Accident
The TMI-2 accident involved a wide range of damage progression and thus is useful for assessing
severe accident computer codes. The damage progression included; (1) ballooning of fuel rods, (2) intense
oxidation and rapid heatup to temperatures that melted a part of the reactor core, (3) reflood of a hot,
partially oxidized core, (4) formation of a molten region across the entire diameter of the reactor core, and
(5) slumping of a significant amount of molten material to the lower head.4- 8 During and following the
TMI-2 accident, measurements and analyses based on measurements have provided a basis for a
quantitative assessment of some of the most important variables calculated by a severe accident computer
code. These variables include the increase in primary coolant system pressure following reflood of a hot
core, cumulative hydrogen production before and after reflood of a hot and partially oxidized core,
maximum mass of molten material and its location within the reactor core, and timing of slumping of
molten material from the core region to the lower head. Since most of the processes causing damage
progression are interdependent, these limited number of measurements and values inferred from
measurements provide a basis for assessing most of the damage progression models in the SCDAP/
RELAP5/MOD3.3 code. The detailed description of the assessment of MOD3.3 using data from the TMI
2 accident is presented in Appendix A12.
The MOD3.3 calculation of the progression of damage in the TMI-2 reactor during its severe
accident was in general agreement with the measurements of damage progression and the post-accident
observation of the state of the reactor core. Table 4-4 summarizes the comparison of calculated and
measured damage progression. The MOD3.2 calculations of the TMI-2 accident are also included in this
table. After core uncovery was calculated to occur at 6270 s, core damage was calculated to occur quite
rapidly. Cladding failure due to ballooning was calculated by MOD3.3 to begin at 8445 s. MOD3.2
calculated cladding failure to be due to chemical attack by the Inconel spacer grids and to not occur until
9418 s. After cladding failure, double-sided oxidation of the cladding was calculated by MOD3.3 to occur
in the vicinity of the cladding failure. The melting of fuel and the formation of a molten pool were
calculated by MOD3.3 to begin at 9530 s. MOD3.2 calculated molten pool formation to begin at 10,330 s.
The start-up of the 2B-pump at 10,446 s was calculated by both MOD3.3 and MOD3.2 to cause a rapid
increase in the pressure of the primary coolant system and to accelerate the rate of hydrogen production.
These calculations are in agreement with the measurements. MOD3.3 calculated 15,000 kg of molten
material in the core region at the start of the 2-B pump and 27,600 kg of molten material in the core just
before the slumping of core material to the lower head at 12,890 s. MOD3.2 calculated 1734 kg of molten
material in the core region at the start-up of the 2B-pump and 37,400 kg of molten material in the core just
before the slumping of core material to the lower head at 13,379 s. The post-accident examination of the
TMI-2 reactor indicated that 40,800 kg of the reactor core was molten at some time. Inferences from the
measured system pressure and other measurements indicated that 15,800 kg of molten material slumped to
the lower head at 13,500 s. Both the MOD3.3 and MOD3.2 calculated masses of molten material and the
time of slumping are in general agreement with the measured values.
The MOD3.3 calculated location of molten core material was in good agreement with the post
accident observation of the TMI-2 reactor. The elevations of the bottom surface of the in-core molten pool
NUREG/CR-6150-Rev. 2, Vol. 5 4-12

at the center line of the core were calculated and observed to be 1.10 m and 0.71 m, respectively. The
highest location in the core to become molten was calculated to be 2.9 m and to be located along the center
line of the core. The molten pool was calculated to extend in the radial direction from the center line of the
reactor vessel to the periphery of the reactor core. The bottom surface of the molten pool at the periphery
of the core was calculated by MOD3.3 to be at the elevation of 2.2 m. These calculated results were in
agreement with the post-accident observation of the TMI-2 core.
Table 4-4. Calculated timing and sequence of core damage progression.
Damage progression Measured

parameter or inferred
Beginning of long term core - 6,390 6,270

uncovery (s).
Beginning of cladding failure due to - 9,417 8,445
ballooning (s).
Beginning of spacer grid slumping - 9,418 9,112
(s).
Beginning of molten pool (s). - 10,330 9,530
Cumulative hydrogen production at 300 275 365
start-up of 2B-pump at 10,446 s
(kg).
Primary coolant system pressure at 8.20 5.03 6.96
start-up of 2B-pump (MPa).
Mass of molten material at start-up - 1734 15,000
of 2B-pump (kg).
Increase in primary coolant system 6.30 5.21 7.00
pressure after start-up of 2B-pump
(MPa).
Final cumulative hydrogen 460 453 417
production (kg).
Mass of core material that was 40,800 37,400 27,600
molten during some period of
accident (kg).
Elevation of bottom of molten 0.71 1.46 1.10
region relative to bottom of core
(m).
Smallest distance from side of 0.0 0.0 0.0
molten pool to periphery of core
(m).
4-13 NUREG/CR-6150-Rev. 2, Vol. 5

Table 4-4. Calculated timing and sequence of core damage progression. (Continued)

Time at which bulk of material in 13,500 13,379 12,890

molten material slumped to lower
head (s).
Mass of molten material that 15,800 37,400 27,600
slumped to lower head (kg).
Percent of molten material that 39 100 100
slumped to lower head (%).
MOD3.3 calculated severe core damage to begin about 800 s earlier than MOD3.2. The onset of
melting of the reactor fuel and the beginning of molten pool formation is a mark of the beginning of severe
core damage. The timing of damage progression is indicated by plots of the history of the maximum
temperature in the reactor core and of the effective radius of the molten pool, as shown in Figures 4-12 and
4-13, respectively. The effective radius is the radius of a hemisphere with a volume equal to the calculated
volume of molten material. The MOD3.3 and MOD3.2 calculated timings for the beginning of fuel
melting were 9530 s and 10,330 s, respectively. MOD3.3 calculated a significantly more rapid heatup of
the reactor core after the beginning of oxidation than did MOD3.2. In the temperature range of 1800 K to
2200 K, when rapid oxidation occurs, MOD3.3 calculated a heatup rate of nearly 50 K/s. The MOD3.2
calculated rate of heatup in this range of temperature was significantly less than that calculated by
MOD3.3. Both MOD3.3 and MOD3.2 calculated that activation of the 2B-pump at 10,446 s did not result
in any reduction in the maximum core temperature. A reduction in temperature was calculated to not occur
because a significant part of the core was in the form of a large molten pool and because of an increase in
oxidation of fuel rod cladding due to cracking of oxide layers and more flow of steam. The activation of
IPIS at 12,012 s was calculated to not cause any cooling of the molten part of the reactor core. MOD3.3
calculated no further melting of fuel after 10,480 s (34 s after activation of 2B-pump), while MOD3.2
calculated fuel melting to occur until 13,000 s. MOD3.3 calculated the maximum effective radius of the
molten pool to be 1.2 m. This value is in general agreement with the corresponding measured value of 1.4
m. When using a special model for failure of an oxide layer retaining melted metallic cladding, MOD3.2
also calculated the amount of molten core material to be similar to the measured value. But when using the
same oxide failure model as used for the analyses of severe accident experiments, MOD3.2 underpredicted
the amount of molten material by a factor of two.
NUREG/CR-6150-Rev. 2, Vol. 5 4-14

3000.0
- MOD 3.3
MOD 3.2
Beginning of
fuel melting Slumpin
.2 2000.0 molten i
Cuto lowef
0)
CL
E
E
E
x
cc
E 1000.0
S,
8 ("
Start-up
(10,446
of 2B-pump
s)\
0.0 5000.0 10000.0 15000.0

Time (s)
Figure 4-12. History of calculated maximum temperature in reactor core.
1.5 ,
- MOD 3.2 (0.6 durability)
-.... MOD3.2 (0.2 durability) --.
- MOD 3.3
0C.
C-
1.0 Molten pool
2 slumped to
0
E lower head
"6n
0.5
Start-up of 2B-pump
Lid
Start-up of
S~HPIS
0(12,000s)
000.0 11000.0 13000.0 15000.0

Time (s)
Figure 4-13. History of calculated effective radius of molten pool.
4-15 NUR.EG/CR-6150-Rev. 2, Vol. 5

The calculated hydrogen production was in general agreement with the hydrogen production
estimated from post-accident observations and inferences. The MOD3.3 calculated cumulative hydrogen
production is compared with the measured hydrogen production in Figure 4-14. The MOD3.2 calculated
hydrogen production is also shown in the figure. MOD3.3 calculated the rapid production of hydrogen to
begin sooner than MOD3.2. Nevertheless, both the MOD3.3 and MOD3.2 calculations of hydrogen
production are in approximate agreement with the measured hydrogen production. At the start-up of the
2B-pump, the MOD3.2 and MOD3.3 calculations of cumulative hydrogen production were 275 kg and
365 kg, respectively. The measured hydrogen production at the start-up of the 2B-pump was 300 kg. The
MOD3.3 calculated and measured total hydrogen productions were 417 kg and 460 kg, respectively. The
MOD3.2 total hydrogen production was 453 kg. Both MOD3.3 or MOD3.2 calculated that no significant
amount of hydrogen production occurred 50 s after the start-up of the 2B-pump (10,500 s). Hydrogen
production was calculated to not occur after 10,500 s because the portions of the core with intact fuel rods
and some metallic cladding were too cool to rapidly oxidize. This behavior is shown in Figure 4-15, where
the MOD3.3 calculated temperature histories are shown for the fuel rods in the outer most fuel assemblies
of the reactor core at the elevations of 3.11 m and 3.47 m, respectively. As shown in this figure, hydrogen
production at the 3.11 m elevation and at the 3.47 m elevation was calculated to stop due to the cooling
caused by the start-up of the 2B-pump. The oxide layer at these locations was calculated to be too thin to
crack during the reflood caused by start-up of the 2B-pump.
500.0
.• 400.0
"o
0. 300.0
CL
>' 200.0
E
: 100.0
0.
- -. 9000.0 11000.0 13000.0 15000.0

Time (s)
Figure 4-14. Calculated hydrogen production during TMI-2 accident.
The MOD3.3 calculation of the disintegration of fuel rods into porous debris was in agreement with
the post-accident observation of the locations in the reactor core with porous debris. Porous debris regions
NUREG/CR-6150-Rev. 2, Vol. 5 4-16

3000.0 I I
- 3.11 m elevation
........ 3.47 m elevation
- 2000.0
OD
E
'- 1000.0
.................................. V44444......
Start-up of 2B3-pump
(10,446s)
0.0 I I
0.0 5000.0 10000.0 15000.0
Time (s)
Figure 4-15. MOD3.3 calculated temperature histories of fuel rods in upper part of outer most fuel
assemblies in reactor core.
were calculated to form in the outer most fuel assemblies in the elevation interval of 0.5 m to 1.2 m and
across the entire diameter of the reactor core in the elevation interval of 2.6 m to 3.6 in. The calculation of
porous debris in the elevation interval of 2.6 m to 3.6 m was consistent with the post-accident observation
of the state of the reactor core.
The porous debris thermal hydraulic models in MOD3.3 performed properly after activation of the
HPIS, which resulted in two-phase coolant conditions in porous debris in the upper part of the reactor core.
The calculated temperature history at a location with porous debris is shown in Figure 4-16. The plot
applies for the location 2.7 m in elevation and in the fourth ring of fuel assemblies. This location
disintegrated from intact fuel rods to porous debris soon after start-up of the 2B-pump, and thus was
porous debris at the time of start-up of HPIS. As a result, the flow losses and heat transfer at this location
during the reflood period beginning with the start-up of HPIS were calculated with the porous debris
thermal hydraulic models implemented into MOD3.3. The debris had a porosity of 0.46 and a particle
diameter of 3.5 mm. Since the molten pool was located below this location and blocked the upward flow of
water from the IPIS, the debris was flooded from the top down. The calculated temperature history of this
location following reflood of the reactor core beginning at 12,012 s was consistent with that seen in
experiments on the quenching of porous debris, as described in Appendix 8.
The implementation into MOD3.3 of the integral diffusion model for fuel rod oxidation caused it to
calculate a more rapid progression of damage to the reactor core than calculated by MOD3.2. The onset of
severe damage was calculated to begin at 9530 s by MOD3.3 and at 10,330 s by MOD3.2. This difference
is due to the integral diffusion model for oxidation in MOD3.3 calculating a more rapid heatup due to
oxidation at locations with a rich supply of steam than that calculated by the parabolic kinetics model for
oxidation in MOD3.2. These differences in calculated behavior also occurred in the analyses of severe fuel
damage experiments described in Appendices A2 through A4.
The MOD3.3 calculation of the TMI-2 accident was improved by the implementation into MOD3.3
4-17 NUREG/CR-6150-Rev. 2, Vol. 5

1000.0 I
900.0
Quenching due to temporary
reflood following start-up of
2B-pump
.j 800.0
E 700.0
I.
Quenching
600.0 600.0 sfollowing
u tart-up
Start-up
s/start-up of -XnIS
of HPIS
500.0
11500.0 12000.0 12500.0
Time (s)
Figure 4-16. MOD 3.3 calculated temperature history of location with porous debris (2.7 m elevation of
fourth ring of fuel assemblies).
of the stress-based model for calculating the time of failure of an oxide layer retaining the melted metallic
part of the fuel rod cladding. While MOD3.2 applied one model for failure of the oxide layer for analysis
of severe fuel damage experiments and another model for the analysis of the TMI-2 accident, MOD3.3
applied the same oxide failure model for analysis of severe fuel damage experiments and the TMI-2
accident. The calculations of the oxide failure model have a strong influence on the calculations of
hydrogen production and the extent of melting of fuel assemblies. With the stress-based model for
calculating oxide failure, MOD3.3 calculated hydrogen production and extent of melting in general
agreement with measurements for both severe fuel damage experiments and the TMI-2 accident. On the
other hand, when using the same oxide failure model for the TMI-2 analysis as used for analysis of severe
fuel damage experiments, MOD3.2 underpredicted by a factor of two the extent of melting of the reactor
core, and did not predict any slumping of molten core material to the lower head.
Another difference in calculated reactor core behavior between MOD3.3. and MOD3.2 was in the
calculated ballooning and rupture of fuel rods; MOD3.3 calculated ballooning and rupture of fuel rod
cladding to occur significantly earlier than MOD3.2. This difference in calculated behavior is due to
corrections made in MOD3.3 to the model for ballooning of the fuel rod cladding.
4.6 Conclusions
The assessment of MOD3.3 showed that the five new models implemented into the code for
calculating the behavior of LWRs during the late-phase of a severe accident are fully operational and
extend the range of applicability of the code. The first two new models, namely the models for flow loss
NUREG/CR-6150-Rev. 2, Vol. 5 4-18

and heat transfer in a porous debris, furnish MOD3.3 with the capability to correctly calculate the blockage
to flow caused by the formation of porous debris and to calculate the temperature behavior of porous
debris for a wide range of coolant conditions. The third new model, namely the model for heat transfer
from a stratified molten pool, furnishes MOD3.3 with the capability to calculate the heat transfer from
molten pools that stratify into oxidic and metallic parts. This model calculated that stratification of a
molten pool decreases the margin to CHF on the external surface of a flooded lower head. The fourth new
model, namely the model for the flow of melted core plate material through porous debris in the lower
head, extends the range of application of the code to the situation in which the core plate is melting above
a porous debris bed. This model calculated that melted core plate material does not significantly decrease
the margin to CHF for a flooded lower head. The fifth new model, namely the model for the interaction of
jets of molten fuel with a pool of water in the lower head, calculated a transient pressure increase due to
fuel-coolant interaction in general agreement with the measured pressure increase.
The MOD3.3 calculation of the TMI-2 accident showed that its new models result in calculated
behavior of the reactor core and primary coolant system in general agreement with measurements and
post-accident observations. The new models that were an integral part of the MOD3.3 calculation of the
TMI-2 accident included; (1) integral diffusion model for oxidation of fuel rod cladding, (2) stress-based
model for failure of oxide layer containing melted metallic cladding, (3) cracking model for cladding oxide
layer during reflood conditions, and (4) porous debris thermal hydraulic models. The calculated and
measured behavior of the TMI-2 reactor during its accident are summarized as follows. The calculated and
measured hydrogen productions were 417 kg and 460 kg, respectively. The calculated and measured
masses of molten material in the core region were 27,600 kg and 40,800 kg, respectively. The calculation
of the locations of porous debris regions in reactor core was for the most part in agreement with the
observed locations of porous debris. The calculated temperature behavior of the porous debris was
consistent with temperature behavior expected during debris quenching. The calculated and measured
rapid increase in primary coolant system pressure following the start-up of the 2B-pump were in general
agreement. The calculated location of molten material in the core region and the timing of the slumping of
this molten material to the lower head were in general agreement with inferences from measurements and
the post-accident observation of the reactor core.
While MOD3.3 and MOD3.2 calculated similar final damage states for the TMJ-2 reactor,
nevertheless the two versions of the code calculated transient aspects of behavior to be significantly
different. MOD3.3 calculated damage progression in the reactor core to occur significantly more rapid
than MOD3.2. This difference is due to oxidation of fuel rod cladding in MOD3.3 being calculated by the
integral diffusion model instead of by the parabolic kinetics model in MOD3.2. MOD3.3 calculated the
ballooning and rupture of fuel rod cladding to occur significantly earlier than MOD3.2. This difference is
due to corrections to the ballooning model implemented into MOD3.3. The stress-based model in MOD3.3
for the failure of an oxide layer retaining melted cladding resulted in good agreement of calculations with
measurements for both the analysis of the TMI-2 accident and the analyses of severe fuel damage
experiments. On the other hand, MOD3.2 underpredicted the extent of core melting by a factor of two
when using for the TMI-2 analysis the same oxide failure model as used for the analyses of severe fuel
damage experiments. In the modeling of phenomena causing damage to fuel assemblies during severe
accident conditions, MOD3.3 does not require a distinguishing of models for the analyses of severe fuel
damage experiments from the models for the analyses of nuclear power plants; one set of models applies
for both types of analyses, and all of the models used for nuclear power plant analyses have been assessed
using severe fuel damage experiments.
4-19 NUREG/CR-6150-Rev. 2, Vol. 5

4.7 References
4-1. The SCDAP/RELAP5 Development Team, "SCDAP/RELAP5/MOD3.2 Code Manual,

Developmental Assessment," NUREG/CR-6150, Vol. 5, Rev. 1, INEL-96/0422, July 1998
4-2. V. X. Tung, V. K. Dhir, and D. Squarer, "Forced Flow Cooling Studies of Volumetrically Heated
Porous Layers," Second International Topical Meeting on Nuclear Reactor Thermal-Hydraulics,
Santa Barbara, California, USA, January 11-14, 1983.
4-3. S. Ergun, "Fluid Flow Through Packed Columns," Chem. Eng. Prog. 48 (1952) pp. 89-94.
4-4. N. K. Tutu et al, "Debris Bed Quenching Under Bottom Flood Conditions (In-Vessel Degraded
Core Cooling Phenomenology)", NUREG/CR-3850, 1984.
4-5. T. G. Theofanous, C. Liu, S. Additon, S. Angelini, 0. Kymalainen, and T. Salmassi, "In-Vessel

Coolability and Retention of Core Melt," Report DOE/ID-10640, Vol. 1, July 1995.
4-6. F. B. Cheung, K. H. Haddad, and Y. C. Liu, "A Scaling Law for the Local CHF on the External
Bottom Side of a Fully Submerged Reactor Vessel," NUREG/CR-0157, Vol. 2, February 1997, pp.
253-277.
4-7. D. Magallon and H. Hohmann, "Experimental Investigations of 150-kg Scale Corium Melt Jet
Quenching in Water," Proceedings of the 7-th International Meeting on Nuclear Reactor Thermal
Hydraulics (NURETH-7), Saratoga Springs, NY, September 1995, NUREG/CP-0412, pp 1688
1711.
4-8. J. M. Broughton, P. Kuan, D. A. Petti, and E. L. Tolman, "A Scenario of the Three Mile Island Unit
2 Accident," Nuclear Technology, 87, August 1989, pp. 34-53.
NUREGICR-6150-Rev. 2, Vol. 5 4-20

APPENDIX A-SCDAP/RELAP5/MOD3.3
ASSESSMENT
APPENDIX A CONTENTS
CONTENTS ............................................................................................................................... Ai
FIGURES .................................................................................................................................. Av
TABLES ................................................................................................................................... Axi

Al. ASSESSMENT OF BOILOFF IN BUNDLE BY COMPARISON
W ITH BENCHM ARKED M ODEL ....................................................................................... A l-i
A l.1 References ........................................................................................ ............. Al-6
A2. ASSESSM ENT USING FLHT-5 TEST ................................................................................. A2-1
A2.1 Nodalization .............................................................................................................. A2-1
A2.2 Boundary Conditions ................................................................................................. A2-2
A2.3 Comparison of Calculations and M easurements ....................................................... A2-2
A2.4 References ................................................................................................................. A2-7
A3. PBF SFD SERIES OF EXPERIM ENTS ................................................................................ A3-1

A3.1 Assessment Using PBF SFD-ST Test ........................................................................ A3-2
A3.1.1 Nodalization ................................................................................................. A3-2
A3.1.2 Boundary Conditions .................................................................................... A3-3
A3.1.3 Results .......................................................................................................... A3-5
A3.2 Assessment Using the PBF SFD 1-1 Test ............................................................... A3-10
A3.2.1 Nodalization ............................................................................................... A3-10
A3.2.2 Comparison of Calculations and Measurements ........................................ A3-12
A3.3 Assessment Using the PBF SFD 1-4 Test ............................................................... A3-17
A3.3.1 Nodalization ............................................................................................... A3-17
A3.3.2 Initial Conditions ........................................................................................ A3-19
A3.3.3 Boundary Conditions .................................................................................. A3-19
A3.3.4 Comparison of Calculations and Measurements ........................................ A3-19
A3.3.5 References .................................................................................................. A3-25
A4. ASSESSM ENT USING THE CORA TESTS ........................................................................ A4-1

A4.1 General Description of the CORA Test Facility ........................................................ A4-1
A4.2 Nodalization of the CORA Test Facility ................................................................... A4-1
A4.3 Test Conditions for the CORA Tests ......................................................................... A4-2
A4.4 Assessm ent Using CORA-5 TEST ............................................................................ A4-3
A4.4.1 Description of the Experimental Bundle ...................................................... A4-3
A4.4.2 Comparison of Calculations and M easurments ............................................ A4-4
A4.5 Assessment Using the CORA-7 Test ......................................................................... A4-8
A4.5.1 Description of the CORA-7 Experimental Bundle and Test ........................ A4-8
A4.5.2 Assessment Results ...................................................................................... A4-9
Ai
A4.6 Assessm ent Using the CORA -13 Test ..................................................................... A4-12
A4.6.1 References .................................................................................................. A4-16
A4.7 Assessment Using the CORA-17 Test ..................................................................... A4-17
A4.7.1 Nodalization ............................................................................................... A4-17
A4.7.2 Boundary Conditions .................................................................................. A4-18
A4.7.3 Comparison of Calculations and M easurements ........................................ A4-18
A5. ASSESSMENT USING THE PHEBUS B9+ TEST .............................................................. A5-1
A5.1 Bundle Nodalization .................................................................................................. A5-1
A5.2 Boundary and Initial Conditions ............................................................................... A5-3
A5.3 Com parison of Calculations and M easurem ents ....................................................... A5-3
A5.4 Reference ................................................................................................................... A5-8
A6. ASSESSMENT USING THE PHEBUS FPTO TEST ............................................................ A6-1
A .6.1 Introduction ............................................................................................................... A6-1
A.6.2 Test D escription ......................................................................................................... A6-1
A.6.3 Nodalization of Test Fuel Assembly ......................................................................... A6-3
A.6.4 Comparison of Calculations and M easurem ents ....................................................... A6-4
A.6.5 References ................................................................................................................. A6-8
A7. ASSESSMENT USING THE ACRR DF-4 TEST ................................................................. A7-1
A.7.1 Nodalization of Test Fuel Assembly and Boundary Conditions ........................... A7-2
A7.2 Comparison of Calculations and Measurem ents ...................................................... A7-4
A7.3 References ................................................................................................................. A7-5
A8. ASSESSMENT OF MODELS FOR FLOW LOSS AND HEAT TRANFER
IN POROUS DEBRIS ............................................................................................................ A8-1
A.8.1 Flow Loss ......................................................................................................... A8-1
A8.2 Heat Transfer ............................................................................................................. A8-4
A8.3 Thermal Hydraulic Behavior in Lower Head ............................. ;.............................. A8-6
A 8.4 Conclusions ............................................................................................................... A8-9
A 8.5 References ................................................................................................................. A8-9
A9. ASSESSMENT OF MODEL FOR HEAT TRANSFER IN STRATIFIED
M OLTEN POOL .................................................................................................................... A9-1
A9.1 Com parison of Calculations W ith Benchm arks ........................................................ A9-1
A9.2. CONCLUSION S .................................................................................................................... A9-6
A9.3. REFERENCES ....................................................................................................................... A9-7
Aii
A10. ASSESSMENT OF MODEL FOR FLOW OF MELTED MATERIAL
THROUGH POROUS DEBRIS ....................................................................................... A10-1
A10.1 Testing By Application To PWR Severe Accident Situation .................................. A10-1
A10.2 Testing By Evaluation of Internal Consistency and Reasonability ..................... A10-5
A 10.3 References ............................................................................................................... A 10-6
All. ASSESSMENT OF MODEL FOR MOLTEN FUEL-COOLANT INTERACTION .......... A 1-1

A11.1 Assessment with a MAGICO-2000 experiment ...................................................... All-i
Al 1.2 Assessment with a FARO Experiment ................................................................ Al1-4
Al1.3 Application of FCI Model to Severe Accident Analysis ..................................... A11-7
A11.4 Summary of Assessment Results ........................................................................... Al1-12
A ll.5 References ............................................................................................................. A 1l1-14
A.12. ASSESSMENT USING TMI-2 ACCIDENT ....................................................................... A12-1
A 12.1 Introduction ............................................................................................................ A 12-1
A12.2 Description of TMI-2 Accident Problem ................................................................ A12-1
A12.2.lBoundary Conditions ................................................................................ A12-10
A12.3 Assessing Using TMI-2 Accident ......................................................................... A12-12
A 12.4 C onclusions ........................................................................................................... A 12-21
A 12.5 R eferences ............................................................................................................. A 12-22
A13. BROWNS FERRY BWR LOCA SEVERE ACCIDENT ..................................................... A13-1
A 13.1 Nodalization ............................................................................................................ A 13-1
A13.2 Description of the Modeled BWR Severe Accident ................................................ A13-5
A 13.3 R esults ...................................................................................................................... A 13-5
A14. SURRY PWR SEVERE ACCIDENT .................................................................................. A14-1
A14.1 Introduction .............................................................................................................. A 14-1
A14.2 Model Description and Nodalization ....................................................................... A14-1
A14.3 Boundary Conditions for TMLB' Severe Accident ................................................. A14-6
A 14.4 References ................................................................................................................. A 14-11
Aiii
FIGURES
Figure Al-i Schematic of test problem and nodalization of fuel bundle ........................... A1-2
Figure A1-2 History of flow rate at inlet of bundle ............................................................ A1-3
Figure A1-3 Comparison of SCDAP and RELAP5 calculated cladding surface
temperatures at elevations of 0.40 m and 0.86 m ........................................... A1-4
Figure A 1-4 Bundle uncovery histories as calculated by SCDAP and RELAP5
heat structures at elevations of 0.40 and 0.86 m ........................................ A1-5
Figure A1-5 Heat transfer rate to coolant at axial node 5 .................................................. A1-6
Figure A2-1 RELAP5 nodalization and bundle configuration for FLHT-5 ........................ A2-2
Figure A2-2 Comparison of calculated and measured temperature at 2.37 m
elevation for FLHT-5 ..................................................................................... A 2-3
Figure A2-5 Comparison of calculated and measured cumulative hydrogen
production for FLHT-5 test ............................................................................ A2-6
Figure A2-6 Comparison of calculated and measured hydrogen production rate .............. A2-6
Figure A3-1 RELAP5 nodalization of the test train for the SFD-ST experiment .............. A3-2
Figure A3-2 SCDAP core representation of the SFD-ST experimental bundle ................. A3-3
Figure A3-3 Total bundle power for the SFD-ST experiment ............................................ A3-4
Figure A3-4 Comparison of calculated and measured cladding temperatures
at the 0.7 m elevation for PBF SFD ST test ................................................... A3-6
Figure A3-5 Comparison of calculated and measured temperatures at 0.5 m
elevation for PBF SFD ST test ....................................................................... A3-7
elevation for PBF SFD ST test ....................................................................... A3-7
Figure A3-7 Calculated cumulative hydrogen production for PBF SFD ST test ............... A3-9
Figure A3-8 Comparison of calculated and measured collapsed liquid level
in test fuel assembly for PBF SFD ST test .................................................... A3-9
Figure A3-9 RELAP5 nodalization of the test train for the SFD1I -1 experiment ............. A3-11
Figure A3-10 SCDAP nodalization of the PBF SFD1-1 experimental bundle .................. A3-12
elevation for PBF SFD 1-1 test .................................................................... A3-13
Figure A3-12 Comparison of calculated and measured temperature at elevation
of 0.5 m for PBF SFD 1-1 test ..................................................................... A3-14
Figure A3-13 Comparison of calculated and measured temperature at elevation
of 0.7 m for PBF SFD 1-1 test ..................................................................... A3-14
Figure A3-14 Comparison of calculated and measured rate of hydrogen
production for PBF SFD 1-1 test ................................................................. A3-16
Figure A3-15 SCDAP nodalization of the PBF SFD 1-4 test bundle ................................. A3-18
Figure A3-16 RELAP5 nodalization of the PBF SFD 1-4 test bundle ............................... A3-18
Av
Figure A3-20 Comparison of calculated and measured temperatures at shroud
mid-wall at 0.91 m for PBF SFD 1-4 test .................................................... A3-22
Figure A3-21 Comparison of calculated and measured rates of hydrogen
production for PBF SFD 1-4 ........................................................................ A3-24
Figure A3-22 Calculated cumulative hydrogen production for PBF SFD 1-4 test ............ A3-24
Figure A4-1 General CORA RELAP5 nodalization used for all PWR tests ...................... A4-2
Figure A4-2 CORA-13 bundle cross-section basic design used for all PWR tests ............ A4-3
Figure A4-3 Comparison of calculated and measured temperatures of
unheated fuel rod at 0.35 m elevation for CORA-5 test ............................ A4-4
Figure A4-4 Comparison of calculated and measured temperatures of
unheated fuel rod at 0.55 m elevation for CORA-5 test ................................ A4-5
Figure A4-5 Comparison of calculated and measured temperatures of unheated fuel
at 0.75 m elevation for CORA-5 test ............................................................. A4-5
Figure A4-6 Comparison of calculated and measured temperature of unheated fuel
rod at 0.85 m elevation for CORA-5 test ....................................................... A4-6
Figure A4-7 Comparison of calculated and measured temperature of unheated fuel
rod at 0.95 m elevation for CORA-5 test ....................................................... A4-6
Figure A4-8 Comparison of calculated and measured temperature of control rod at
elevation of 0.55 m elevation for CORA-5 test ............................................. A4-7
Figure A4-9 CORA-7 large bundle cross-section .............................................................. A4-8
Figure A4-10 Comparison of calculated and measured temperature at the 0.35 m
elevation for CORA-7 test ............................................................................. A4-9
elevation for CORA-7 test ........................................................................... A4-10
Figure A4-14 Calculated cumulative hydrogen production for the CORA-7 test.............. A4-12
Figure A4-15 Comparison of calculated and measured temperatures of unheated
fuel rod at elevation of 0.35 m for CORA-13 test ....................................... A4-13
fuel rod at elevation of 0.55 m for CORA-13 test .................. A4-14
fuel rods at elevation of 0.75 m for CORA-13 test ...................................... A4-14
Figure A4-18 Comparison of calculated and cumulative hydrogen production
for C O RA -13 tests ....................................................................................... A4-16
Avi
Figure A4-19 RELAP5 nodalization of the CORA-17 experiment ................................... A4-17
Figure A4-20 CORA - 17 bundle cross-section ................................................................... A4-18
fuel rod at elevation of 0.55 m for CORA-17 test ....................................... A4-19
rod at elevation of 0.75 m for CORA-17 test ............................................. A4-19
Figure A4-23 Comparison of calculated and measured temperatures of channel
box at 0.55 m for CORA-17 test .................................................................. A4-20
Figure A4-24 Comparison of calculated and measured temperatures of control blade
at 0.55 m for CORA -17 test ......................................................................... A4-20
Figure A5-1 RELAP5 nodalization used for the PHEBUS B9+ experiment ..................... A5-2
Figure A5-2 SCDAP representation of the fuel bundle ...................................................... A5-3
Figure A5-3 Comparison of calculated and measured cladding temperatures at
0.2 m elevation for PHEBUS B9+ test .......................................................... A5-4
Figure A5-4 Comparison of calculated fuel temperatures at 0.4 m elevation for
PHEBU S B9+ test ...................................................................................... A 5-5
Figure A5-5 Comparison of calculated and measured fuel temperatures at 0.6 m
elevation for PHEBUS B9+ test .................................................................. A5-5
Figure A5-6 Comparison of calculated and measured cladding temperatures at 0.7 m
elevation for PHEBUS B9+ test ..................................................................... A5-6
Figure A5-7 MOD3.3 calculated nuclear and oxidation heat generation histories
in test fuel assembly for PHEBUS B9+ test .................................................. A5-6
Figure A5-8 Calculated cumulative hydrogen production for PHEBUS B9+ test ...... A5-8
Figure A6-1 Cross-section of bundle of test rods for PHEBUS FPT0 experiment ............ A6-1
Figure A6-2 Inlet flow rate history of test bundle .............................................................. A6-2
Figure A6-3 Power history of test bundle ........................................................................... A6-2
Figure A6-4 Schematic of nodalization of test rods, insulated shroud, and flow
channels .......................................................................................................... A 6-3
elevation for PHEBUS FPTO ......................................................................... A6-5
elevation for PHEBUS FPTO test ................................................................... A6-5
elevation for PHEBUS FPTO test ................................................................... A6-6
Figure A6-8 Calculated cumulative hydrogen production for PHEBUS FPTO test ........... A6-7
Figure A7-1 Cross section of the DF-4 experiment bundle ................................................ A7-1
Figure A7-2 SCDAPIRELAP5 nodalization for the DF-4 experiment analysis ................ A7-3
Figure A7-3 Comparison of calculated and measured temperatures of fuel rod at
elevation of 0.254 m for DF-4 test ............................................................. A7-4
Figure A7-4 Comparison of calculated and measured temperatures of fuel rod at
elevation of 0.096 m for DF-4 test ............................................................. A7-4
Figure A8-1 Schematic of debris bed analyzed for assessment of flow loss
calculations ................................................................................................. A 8-1
Avii
Figure A8-2 Nodalization of debris bed analyzed for assessment of flow loss
calculations .................................................................................................... A 8-3
Figure AS-3 Schematic of BNL quenching experiment ..................................................... A8-4
Figure A8-4
Comparison of calculated and measured transient temperature
A 8-5
Figure A8-5 distribution in debris bed ...............................................................................
head of reactor vessel .......................... A8-6
Figure A8-6 Nodalization of debris bed in lower
RELAP5 nodalization for analysis of porous debris in lower head ............... A8-7
Figure A8-7
Volume fraction of liquid in RELAP5 control volume containing
the flooded debris bed ................................................................................ A 8-7
Figure A8-8
Transient temperature at center of debris bed ............................................ A8-8
Figure A8-9
Comparison of debris bed power and heat transfer to coolant ....................... A8-8
Figure A9-1
Molten pool system analyzed in literature and by SCDAP/RELAP5 ............ A9-1
Figure A9-2
Temperature histories of oxidic and metallic parts of molten pool
before and after stratification ......................................................................... A9-5
Figure A9-3
Ratio of heat flux to critical heat flux on external surface of lower
head before and after stratification of molten pool ........................................ A9-5
Figure A10-3
Schematic of system analyzed and nodalization of system ......................... A10-1
Figure A10-2
Distribution of debris bed saturation and temperature after melting
of 30% of core plate (time of 1500 s) ..................................................... A10-3
Figure A10-3
Temperature history of debris bed along its centerline ................................ A10-3
Figure A1O-4
Ratio of heat flux to CHF on external surface at bottom center
of low er head ................................................................................................ A 10-4
Figure A10-5 Axial distribution in velocity of melted material at 1500 s as
calculated by melt movement model .......................................................... A1O--6
Figure Al 1-1 Schematic of MAGICO - 2000 Test Arrangement ...................................... A11-1
Figure A11-2 Comparison of calculated and measured advancement of the front
of cloud of particles for MAGICO-2000 experiment .................................. A 11-3
Figure A 11-3 Comparison of calculated and measured volume fraction of liquid at
elevation of 0.65 m for MAGICO-2000 experiment .................................... A11-3
Figure A 1-4 Comparison of calculated and measured water level swell for
MAGICO-2000 experiment ..................................................................... A11-4
Figure Al1-5 Schematic of FARO Test Arrangement ........................................................ A11-5
Figure A 11-6 Comparison of calculated and measured pressure histories for FARO
Test L-08 .................................................................................................. A 11-6
Figure A 11-7 Schematic of application of FCI model to severe accident analysis ............ A11-7
Figure A 11-8 Calculated transients pressures in vessel for Case 1 (1 large jet) and
Case 2 (25 sm all jets) ................................................................................. A ll-10
Figure A11-9 Calculated transient pressure in vessel for Case 3 (1 small jet),
Case 4 (2 small jets) and Case 5 (4 small jets) .......................................... All-10
Figure A11-10 Transient volume fractions of liquid water at various elevations for
Case 1 (single large jet).............................................................................. A ll-11
Figure A11- 11 Transient volume fractions of liquid at various elevations for
Case 2 (25 sm all jets) ................................................................................. A ll-11
Aviii
Figure A 12-1
RELAP5 nodalization of the reactor vessel and core .................................. A12-2
Figure A12-2RELAP5 nodalization of primary coolant loop A ....................................... A12-3
Figure A12-3RELAP5 nodalization of steam generator A ............................................... A12-4
Figure A 12-4
RELAP5 nodalization of the pressurizer ..................................................... A12-5
Figure A12-5Cross-sections of core showing fuel assembly grouping and radial
peaking factors ............................................................................................. A 12-7
Figure A12-6 Makeup flow history for TMI-2 calculation .............................................. A12-11
Figure A12-7 Reactor power versus time curve used for TMI-2 calculation ................... A 12-11
Figure A12-8 Timing of closure of PORV block valve and history of calculated
rate of flow through PORV valve............................................................... A12-12
Figure A 12-9 History of calculated collapsed liquid level in reactor vessel .................... A12-15
Figure A12-10 Comparison of calculated and measured liquid level in pressurizer ........ A12-15
Figure A12-11 History of calculated maximum temperature in reactor core .................... A12-16
Figure A12-12 History of calculated effective radius of molten pool ............................... A12-17
Figure A 12-13 Calculated and measured pressures of primary coolant system ................ A12-17
Figure A12-14 Comparison of calculated and measured pressure histories of primary
coolant system for case of boundary conditions from Annunziato ............ A12-18
Figure A12-15 Calculated hydrogen production during TMI-2 accident .......................... A12-19
Figure A12-16 MOD3.3 calculated temperature histories of fuel rods in upper
part of outer most fuel assemblies in reactor core ..................................... A12-20
Figure A12-17 Temperature history of location with porous debris (2.7 m
elevation of second ring of fuel assemblies) .............................................. A12-21
Figure A13-1 Nodalization diagram of Browns Ferry ........................................................ A13-1
Figure A13-2 Nodalization diagram of lower reactor vessel .............................................. A13-2
Figure A 13-3 Component guide for detailed Browns Ferry SCDAP model ...................... A13-3
Figure A 13-4 Browns Ferry radial power profile for 4 radial rings ............................... A13-4
Figure A13-5 Browns Ferry power profiles for 13 axial nodes ...................................... A13-4
Figure A13-6 Predicted radial temperature distribution across reactor core of
control blades ............................................................................................... A 13-6
channel boxes ............................................................................................... A 13-7
fuel rods at the 0.5334 m elevation .............................................................. A13-7
fuel rods at the 0.9906 m elevation .............................................................. A13-8
fuel rods at the 2.22 10 m elevation .............................................................. A 13-8
Figure A 13-11 Predicted Browns Ferry system pressure .................................................... A13-9
Figure A13-12 Predicted integral hydrogen production .................................................... A13-10
Aix
TABLES
Table Al-i Axial power profile in fuel bundle ................................................................. A1-3
Table A2-1 Comparison of calculated and measured axial distributions in oxidation
and m eltdown for FLHT-5 test ....................................................................... A2-5
Table A3-1 History of inlet flow rate to the test fuel assembly for PBF SFD ST Test ..... A3-4
and meltdown for PBF SFD ST Test ............................................................. A3-8
and meltdown for PBF SFD 1-1 test ............................................................ A3-15
and meltdown for PBF SFD1 1-4 test ............................................................ A3-22
and meltdown for CORA-13 test ................................................................. A4-15
and meltdown for PHEBUS B9+ test ............................................................ A5-7
Table A6-1 MOD3.3 calculated axial distribution in oxidation and meltdown for
PHEBU S FPTO test ........................................................................................ A 6-6
Table A8-1 Characteristics of debris bed and coolant conditions for assessment of
pressure drop calculations .............................................................................. A 8-2
Table A8-2 Comparisons of SCDAP/RELAP5 calculated flow losses with those
presented in literature for corresponding cases .............................................. A8-3
Table A8-3 Summary description of BNL quenching experiment ............................... A8-4
Table A9-1 Characteristics of system analyzed that were same for each test problem.... A9-2
Table A9-2 Characteristics of molten pool that varied with test problems ....................... A9-2
Table A9-3 Comparison of SCDAP/RELAP5 and Reference A9-1 solutions for
Test Problem 1 (deep metallic pool) .............................................................. A9-3
Table A9-4 Comparison of SCDAP/RELAP5 and Reference A9-2 solutions for
Test Problem 2 (shallow metallic pool) ......................................................... A9-4
Table A9-5 Sensitivity of calculated molten pool behavior to correlations for flux
shape and sideward heat transfer in metallic pool ......................................... A9-6
Table A10-1 Initial conditions and boundary conditions ............................................ A10-2
Table A10-2 Effect of level of debris saturation on thermal conductivity of debris ......... A10-5
Table A 1l-1 Characteristics of Run Z1500/0.2 of MAGICO-2000 Experiment .............. Al 1-2
Table A 11-2 Characteristics of Test L-08 of FARO Experiments .................................... Al 1-4
Table A11-3 Summary of comparison of measured and calculated values for
Test L-08 of FARO .................................................................................... A 1l1-6
Table A 11-4 Characteristics of typical severe accident conditions with possibility
for FC I ...................................................................................................... A l1 -8
Table A 11-5 Summary of results for FCI under conditions typical for
severe accident in PW R ............................................................................... Al 1-9
Table A11-6 Experiment parameters for FARO series of experiments .......................... Al1-13
Table A11-7 Summary of FARO experimental results . ......................... .................... Al1-13
Table A11-8 Summary of CCM-5 and CCM-6 experimental parameters ..................... Al 1-14
Axi
Table A12-1 TMI-2 axial power peaking factors .............................................................. A12-6
Table A 12-2 Total fuel assemblies, fuel rods, and control rods in each core region ........ A12-7
Table A12-3 SCDAP input parameters ............................................................................. A12-8
Table A 12-4 TMI-2 initial conditions at turbine trip ........................................................ A12-9
Table A12-5 Steam generator initial conditions ............................................................. A12-10
Table A12-6 Calculated timing and sequence of core damage progression ................... A12-13
Axii
Appendix Al
Al. ASSESSMENT OF BOILOFF IN BUNDLE BY COMPARISON WITH

BENCHMARKED MODEL
The Boiloff in Bundle test problem is used to assess the models for convective heat transfer and
coolant boil off. The assessment was performed by comparing the calculations of the temperature behavior
of a fuel assembly during coolant boiloff using the SCDAP fuel rod model with that calculated using a
previously benchmarked model.
The Boiloff in Bundle Test Problem requires for solution the steady-state heat transfer in a bundle of
fuel rods and then the transient uncovery of the bundle and transient heat up of the bundle in response to a
reduced rate of flow through the bundle. A schematic and the nodalization of the test bundle are shown in
Figure Al-i. The bundle consists of 32 identical fuel rods. The bundle is 0.9144 m in height. The fuel rods
in the bundle have an outer diameter of 9.63 mm and a pitch of 12.80 mm. Except for height, the design of
the fuel rods is typical of PWR fuel rods. The flow area of the bundle of fuel rods is equal to 3.685 x 10-3
mi2 . The bundle of fuel rods is surrounded by an adiabatic boundary through which no flow of heat occurs.
The SCDAP/RELAP5 code represented the fuel bundle as eight equally sized axial nodes and eight equally
sized hydrodynamic control volumes.
The bundle power was constant with time and equal to 58.5 KW. The bundle has a cosine shaped
axial power distribution. The peak power in the bundle was 1.33 times the axially averaged power and
occurs at an elevation of 0.46 m above the bottom of the fuel stack. The axial power profile of the fuel rod
bundle is shown in Table AI-1.
AI-I NUREG/CR-6150-Rev. 2, Vol. 5

Appendix A I
- Time dependent
-.4 volume
Single junction
o3 Bundle of
32 fuel rods
0.9144 m
G0 Pipe component
Node number
4
L do Time dependent
Junction
Time dependent
volume
Figure A1-1. Schematic of test problem and nodalization of fuel bundle.
NUREG/CR-6150-Rev. 2, Vol. 5 AI-2

Appendix Al
Table Al-1. Axial power profile in fuel bundle.
Elevation above bottom Axial power factor

of fuel stack (m)
0.0571 0.5941
0.1714 0.9406
0.2858 1.1881
0.4000 1.3267
0.5143 1.3267
0.6287 1.1881
0.7430 0.9109
0.8573 0.5248
The coolant condition history in the fuel bundle was designed to show calculated convective heat
transfer under steady-state pre-CHF conditions and then show transient convective heat transfer under the
conditions of a boil off of the coolant. The pressure in the bundle was constant and equal to 6.89 MPa. The
flow rate history at the inlet of the bundle is shown in Figure A 1-2. The inlet flow rate was 1 kg/s in the
period of 0.0 s to 50 s, 0.1 kg/s in the period of 50 to 100 s, and zero from 100 s through the end of the
problem. The inlet temperature of the water was 558 K, which was equal to the saturation temperature at
the bundle pressure of 6.89 MPa.
1.5
1.0 -
_o
0
0.5
0.01o. 0 -7--I
50.0 100.0 150.0 200.0
Time (s)
Figure A1-2. History of flow rate at inlet of bundle.
AI-3 NUREG/CR-6150-Rev. 2, Vol. 5

Appendix Al
The SCDAP modeling for convective beat transfer was assessed by using as a benchmark the
solution for the test problem obtained by representing the fuel bundle by RELAP5 heat structures.All The
test problem was solved with the fuel bundle represented by RELAP5 heat structures and twice with the
fuel bundle represented as SCDAP heat structures, once using the explicit calculation of coolant conditions
(Table 4 variable nhysol equal to 3) and once using the implicit calculation of coolant conditions (Table 4
variable equal to 7). The SCDAP calculation of convective heat transfer coefficients was assessed by
comparing the SCDAP calculated cladding surface temperature histories with those calculated using
RELAP5 heat structures. The SCDAP calculation of the rate of coolant boil off was assessed by comparing
the SCDAP calculated histories of the liquid void fraction of the coolant in the bundle with those obtained
using RELAP5 heat structures. The SCDAP convective heat transfer modeling was also checked using an
energy balance.
The calculated convective heat transfer coefficients through a range of two-phase coolant conditions
corresponding with coolant boil off were virtually identical for the two SCDAP models and the RELAP5
heat structure model. This evaluation is based on comparisons of the cladding surface temperature histories
as calculated by SCDAP and the RELAP5 heat structure models. Figure A 1-3 is a plot of the calculated
surface temperatures at the elevations of 0.40 m and 0.86 m above the bottom of the fuel stack. For the 0.86
m elevation, the explicit and implicit SCDAP models and the RELAP5 heat structure model calculate that
film boiling begins at 107 s, which is 7 s after the time at which the fuel bundle inlet flow rate drops to
zero. After initiation of film boiling, all three models calculate an almost identical rate of heat up of the
cladding surface. This similarity in rate of heatup indicates that all three models are calculating almost
identical convective heat transfer. Similar results were obtained for the 0.4 m elevation. For this elevation,
all three models calculate that film boiling begins at 135 s. After film boiling occurs, all three models
calculate an almost identical rate of heat up. At 180 s, the calculated rate of heat up is 8.4 K/s.
1500.0 ....
"---e SCDAP exp, 0.86
o-oSCDAP imp, 0.86 m
1300.0 -*-*RELAP5 0.86
exp, 0.40 m SSCDAP
u-. SCDAP imp, 0.40 m
-. S1100.0
RELAP5,0.40m
E 900.0
700.0
700.0
500.0, , , , , , ,
100.0 120.0 140.0 160.0 180.0 200.0
Time (s)
Figure A1-3. Comparison of SCDAP and RELAP5 calculated cladding surface temperatures at
elevations of 0.40 m and 0.86 m.
NUREG/CR-6150-Rev. 2, Vol. 5 AI-4

Appendix Al
The SCDAP explicit and implicit convective heat transfer models and the RELAP5 heat structure
model calculate identical void fraction histories for the fuel bundle. Figure AI-4 shows plots of the
histories of the volume fraction of liquid at the elevations of 0.86 m and 0.40 m. In the period of 60 s to
100 s, the three models calculate a liquid volume fraction of 0.6 at the 0.86 m elevation and a liquid
volume fraction of 0.72 at the 0.40 m elevation. For the 0.86 m elevation, bundle uncovery is calculated to
begin at 101 s and is completed by 107 s. For the 0.40 m elevation, bundle uncovery is calculated to begin
at 124 s and is completed by 136 s.
1.0 I * I * I * I
08 k
v.v
0
.06
E
*-* SCDAP exp, 0.40m
_E3DSCDAP exp, 0:86m \
.4
0.4
eeSCDAP imp, 0.86m
0.2 -* RELAP5, 0.40m
*-*RELAP5, 0.86m
0..
0 80.0 Te - )- M
Time (s)
Figure A1-4. Bundle uncovery histories as calculated by SCDAP and RELAP5 heat structures at
elevations of 0.40 and 0.86 m.
The SCDAP convective heat transfer modeling results in an energy balance between fuel rod power
and heat transferred to the coolant. This energy balance was checked by comparing fuel rod power with the
calculated rate of heat transfer to the coolant. The nuclear heat generation rate at axial node 5 is 9,730 W.
The elevation span of this node is from 0.457 m to 0.572 m above the bottom of the fuel bundle. The heat
transfer rate to the coolant at this axial node is plotted in Figure Al-5. The heat transfer rate to the coolant
is also equal to 9,730 W until bundle uncovery occurs at 113 s. The two SCDAP convective heat transfer
models and the RELAP5 heat structure model calculate similar histories for the heat transfer rate to the
coolant.
The CPU time required for solution of the test problem was almost identical for the implicit and
explicit convective heat transfer models in SCDAP. The CPU time for the solution of the test problem was
50 s on a DEC 3000 work station. The CPU time was evenly distributed through the time span of the test
problem.
A1-5 NUREG/CR-6150-Rev. 2, Vol. 5

Appendix A I
15000.0
•:10000.0 •
1Bundle uncovery-
at axial node 5
0
, 5000.0
E
"0 0.0- Implicit SCDAP
;> Explicit SCDAP
S..... RELAP5 heat structure
-5000.0 ' I
0.0 50.0 100.0 150.0 200.0
Time (s)
Figure Al-5. Heat transfer rate to coolant at axial node 5.
In summary, this test problem has produced two results. First, it has shown that SCDAP heat
structures accurately calculate the boil off and heatup that occurs in a bundle of fuel rods subjected to a
deficiency in coolant flow. Second, the implicit convective heat transfer model does not result in any
increase in computational efficiency or accuracy relative to the explicit convective heat transfer model.
Since the explicit model has been applied and assessed much more than the implicit model and is as
efficient and accurate as the implicit model, this result is a basis for recommending the explicit convective
heat transfer model as the default model for convective heat transfer in SCDAP/RELAP5/MOD3.3.
A1.1 References
A1-1. The RELAP5 Development Team, RELAP5/MOD3 Code Manual, Volume IV, NUREG/CR-5535,
INEL-95/0174, August 1995.
NLUREG/CR-6150-Rev. 2, Vol. 5 A1-6

Appendix A2
A2. ASSESSMENT USING FLHT-5 TEST

FLHT-5, the fourth in a series of full length, high temperature severe core damage tests, was
conducted by Pacific Northwest National Laboratory (PNNL) in the National Research Universal (NRU)
reactor at Chalk River Nuclear Laboratory (CRNL), Ontario, Canada.A2 -'The objectives of this test, which
simulated conditions in a light water reactor (LWR) core during a loss-of-coolant accident, were; (1) to
heat up fuel rods to the point of severe fuel damage at 30 kW nuclear power during a boildown and the
subsequent heatup, (2) to determine the time at which oxidation and hydrogen production ceased within
the test assembly and characterize the cause of the cessation of oxidation, and (3) to characterize the
release, transport, and deposition of gaseous and volatile fission products. To accomplish the goals of the
test, the bundle was held at a maximum temperature of 2,500 K for 60 minutes, instrumented with
numerous thermocouples, flow meters and pressure transducers, and examined visually after completion of
the test.
The FLHT-5 test assembly consisted of 12 fuel rods surrounded by a multilayer low density zirconia
shroud, with a zircaloy liner. The fuel bundle, a modified 4 x 4 array, is shown in Figure A2-1. The comer
rod positions in the array were replaced with angled zircaloy carrier pieces that routed and protected
instrument wires in the test bundle. Ten of the remaining 12 positions in the fuel bundle were fresh fuel
rods, fabricated at PNNL. Eight of the nonirradiated rods were instrumented. One of the remaining
positions in the array contained a three-cycle irradiated PWR fuel rod from the H. B. Robinson reactor, the
other a stainless steel rod centered in a zircaloy guide tube. The irradiated rod had an initial U0 2 pellet
enrichment of 2.55%, a nominal rod-average burnup of 28 MWd/kgU, and a cooling time of 13 years. The
3.6 m in height fuel rods in the test bundle were held in place by 8 instrumented grid spacers, four
composed of inconel and four zircaloy (the bottom three and top grid spacers were inconel, the other four
were zircaloy).
The test was conducted in five phases; (1) equipment installation and checkout, (2) commissioning,
which included the injection of metered hydrogen and nitrogen gas flows into the effluent piping to check
the response and calibration of the various hydrogen sampling and analysis systems and to perform leak
checks on the system, (3) preconditioning of the assembly to ensure adequate fuel pellet cracking of the
fresh rods, (4) the boilaway transient, where 30-kW of nuclear power and a coolant makeup flow of 1.26 g/
s were maintained for 60 minutes beyond the first cladding temperature escalation to 2,500 K, and (5) post
test activities, which included deposition rod removal and gamma scanning, visual inspection of the
damaged test assembly, sectioning of the test assembly, stack release analyses, and flux wire gamma
scanning.
A2.1 Nodalization
As shown in Figure A2-1, the nodalization of the FLHT-5 test assembly was represented by SCDAP/
RELAP5/MOD3.3 using ten, 0.3648 m in height, RELAP5 hydrodynamic volumes in the test bundle
region and a single RELAP5 volume to represent the bypass. Five SCDAP components were used to
represent rods and structures in the core region. Component 1 represented the unirradiated fuel rods,
Component 2 represented the single irradiated H. B. Robinson fuel rod, Component 3 represented the
zircaloy guide tube containing a centered stainless steel rod; input in the assessment deck as a control rod
with an extremely small quantity of control material. Component 4 represented the comer carrier tubes.
Component 5 represented the insulating shroud.

Appendix A2
otet
Zircaloy tube
--------
shroud with
ited Zr liner on
inner surface
Bypass channel
0 Fresh fuel rods
Irradiated fuel rods
0
G Stainless steel rod
Source-
S---------- Carrier rods
Bypass
source
Shroud
Figure A2-1. RELAP5 nodalization and bundle configuration for FLHT-5.
A2.2 Boundary Conditions

Prior to initiating the test, steady-state conditions were established at a system pressure of 1.38 MPa
with a coolant flow rate of 11.3 g/s and an initial liquid level at the elevation of 2.9 m above the bottom of
the test bundle. After the establishment of steady-state conditions in the test bundle, the test was initiated
by decreasing the inlet flow of the coolant to 1.26 g/s and applying 30 kW nuclear power to the test
assembly. The initiation of power started the boilaway phase of the experiment.
A2.3 Comparison of Calculations and Measurements

The results for the FLHT-5 experiment provide benchmarks for the early-phase damage progression
models in a severe accident analysis computer code. The experiment subjected the test fuel rods to
conditions that caused rapid oxidation and some meltdown of the cladding of the test fuel rods. The
temperature history of the test fuel assembly at three different elevations was obtained during the period in
which oxidation was causing a rapid heatup of the test fuel assembly. A measurement was obtained of the
NUREG/CR-6150-Rev. 2, Vol. 5 A2-2

Appendix A2
amount of hydrogen produced by oxidation of the cladding of the test fuel rods and other structures. The
rate of hydrogen production was also measured. A limited post-irradiation examination (PIE) obtained
some information on the extent of oxidation of the test fuel assembly and the location of slumped fuel rod
material.
The MOD3.3 calculated temperature behavior of the test fuel assembly was generally in good
agreement with its measured temperature behavior. The MOD3.3 calculated and measured temperature
history of a test fuel rod at the 2.37 m elevation (referenced from bottom of fuel rod) are shown in Figure
A2-2. The 2.37 m elevation is the highest elevation at which a temperature measurement was obtained for
3000.0
2000.0
0.
E
i--
1000.0
0.0 L
0.0 1000.0 2000.0 3000.0 4000.0 5000.0
Time (s)
Figure A2-2. Comparison of calculated and measured temperature at 2.37 m elevation for FLHT-5.
the test fuel rods. A temperature increase due to uncovery of the test bundle occurred first at this location.
The calculated rate of temperature increase for a temperature greater than 1500 K was about 16 K/s. The
calculated rate of temperature increase is slightly slower than the measured rate of temperature increase.
The temperature measurement failed after the temperature exceeded 1600 K. The temperature history
calculated by MOD3.2 at this elevation is also shown in Figure A2-2. The temperature histories calculated
by MOD3.2 and MOD3.3 are similar. The calculated and measured temperature histories of the test fuel
rods at the 1.64 m and 1.22 m elevations are compared in Figure A2-3 and Figure A2-4, respectively. The
calculated rate of temperature increase was slightly less than the measured rate of increase. A maximum
temperature of 2670 K was calculated at the elevations of 1.22 m and 1.64 m. Based on measurements of
the temperature of the outside of the insulated flow shroud surrounding the test fuel assembly, the
maximum fuel assembly temperature was estimated to exceed 2600 K.A2- The temperature histories
calculated by MOD3.2 are also shown in Figure A2-2 through Figure A2-4. The MOD3.3 and MOD3.2
calculations are similar for the 1.64 m elevation, but at the 1.22 m elevation the MOD3.3 calculations are
in better agreement with the measured temperature history than the MOD3.2 calculations.

Appendix A2
3000.0 'I * I * I I '

2000.0
1-- ------------
E
I-
1000.0 -t Measured
Calculated, MOD3.3
-------------.
........... Calculated, MOD3.2
SI tI ,I ,*
0.0 L-
0.0 1000.0 2000.0 3000.0 4000.0 5000.0
Time (s)
3000.0 I I * I
2 2000.0
.. ..-
--..---------.. ..... ...-\
E S.
I" M
'- 1000.0 easured
............. C,alculated, MOD3.3
alculated, MOD3.2
I I I I
0.0
0.0 1000.0 2000.0 3000.0 4000.0 5000.0
Time (s)

Appendix A2
The calculated axial distribution in oxidation of the cladding of the test fuel rods was in general
agreement with the axial distribution obtained from the PIE performed on the test fuel assembly. Although
the PIE was limited in scope, nevertheless information was obtained indicating almost 100% oxidation of
the cladding of the test rods above an elevation of 1.3 m. The measured and the MOD3.3 and MOD3.2
calculated axial distributions in the fraction of the fuel rod cladding oxidized are tabulated in Table A2-1.
As shown in this table, MOD3.3 calculated almost 100% oxidation of the cladding above the 1.3 m
elevation. The only exception was at the 2.37 m elevation, where cladding oxidation was limited to 43%
due to slumping of the melted metallic cladding at that location. Double-sided oxidation was calculated to
not occur at any location in the test fuel assembly. Since meltdown has an influence on the axial
distribution in oxidation, the calculated locations of cladding slumping and blockages are also shown in
Table A2-1. The blockage is defined to be the fraction of coolant space filled with slumped material. The
Table A2-1. Comparison of calculated and measured axial distributions in oxidation and
meltdown for FLHT-5 test.
Fraction
blockage due
Elevation Fraction cladding to meltdown/
oxidized slumping
from location
(yes or no)
(m) MOD3.3 Measured MOD3.3

0.18 0.0 -0.0 0.0/no
0.55 0.0 -0.0 0.0/no
0.91 0.04 -0.0 0.15/no
1.28 0.97 -1.0 O.O/no
1.64 1.00 -1.0 0.0/nO
2.00 1.00 -1.0 O.O/no
2.37 0.43 -1.0 O.O/yes
2.75 1.00 -1.0 0.0/no
3.10 1.00 -1.0 0.0/no
3.47 1.00 -1.0 O.O/no
PIE indicated that some slumping of melted cladding occurred but that the slumping was limited and did
not cause any extensive blockage. The limited scope of the PIE precluded a quantitative measurement of
fuel rod meltdown. MOD3.3 also calculated only a small amount of cladding slumping, and thus the
calculated amount of fuel rod meltdown was consistent with the measured amount.
The calculated hydrogen production was somewhat less than the measured hydrogen production. The
calculated and measured hydrogen production are compared in Figure A2-5. The MOD3.3 calculated and
measured hydrogen productions were 231 g and 300 g + or - 30 g. The hydrogen production calculated by
MOD3.2 was 249 g. The general agreement of calculated and measured axial distribution in oxidation
indicates that the underprediction in hydrogen production may be due to an underprediction of the

Appendix A2
0.40
0.30
0
-c
CM
0.20
0
"{0
CU
0.10
E
0.00 1
0.0 1000.0 2000.0 3000.0 4000.0 5000.0
Time (s)
Figure A2-5. Comparison of calculated and measured cumulative hydrogen production for FLHT-5 test.
0.00050
0.00040
C
0
0.00030
0._
:3
0
"C
VO
Cu
o 0.00020
0
0.00010
0.00000 L
0.0 2000.0 4000.0 6000.0
Time (s)
Figure A2-6. Comparison of calculated and measured hydrogen production rate.

Appendix A2
oxidation of other components in the test train for the test fuel assembly, such as the shroud with a Zr inner
liner and Zircaloy carriers for instrument wires. The calculated and measured rate of hydrogen production
are compared in Figure A2-6. The measured rate was obtained from the Noncondensable Turbine
Flowmeter. The measured rate of hydrogen production lags the calculated rate of hydrogen production
because the measurement was performed a considerable distance downstream of the test fuel assembly.
Both the calculations and the measurements showed that most of the hydrogen production occurred in the
period of 1000s to 3000 s. During this period, most of the uncovered portions of the test fuel rods were at a
temperature greater than the melting temperature of the metallic part of the fuel rod cladding, as shown in
Figure A2-2 through Figure A2-4. The calculated and measured maximum rates of hydrogen production
were about 1.3x10"4 kg/s. This rate of hydrogen production corresponds with oxidation in the test
assembly consuming 100% of the makeup coolant flowing into the test bundle.
In summary, the calculated and measured behavior of the test fuel assembly for the FLHT-5 test are
in good agreeement. Both the calculations and the measurements showed an oxidation driven heatup rate
in excess of 15 K/s in the portion of the test assembly uncovered by the boiloff of water. The calculations
and measurements indicated that cladding melting did not reduce the the rate of oxidation. The
calculations and measurements indicated 100% consumption of the makeup coolant flow during the high
temperature period of the test and almost 100% oxidation of the cladding of the fuel rods uncovered by
coolant boiloff. Both the calculations and measurements indicated that the meltdown of the test fuel
assembly was limited even though the maximum temperature in the test fuel assembly exceeded 2600 K,
and that no significant flow blockage occurred.
A2.4 References
A2-1. D. D. Lanning et al., "Data Report: Full-Length High Temperature Experiment 5," PNL6540,
April 1988

Appendix A3
A3. PBF SFD SERIES OF EXPERIMENTS

A severe fuel damage (SFD) research program was initiated by the U.S. Nuclear Regulatory
Commission (NRC) to develop a data base and models to predict (a) the overall response of a reactor core
and associated structures, (b) the rate of release of fission products; their chemical forms, and
characteristics of their transport and deposition in the primary system, (c) the rate of hydrogen generation
from the interaction of coolant with the fuel, cladding, and reactor structure, and (d) the coolability of the
damaged fuel following reflood.
Several severe fuel damage experiments were performed at the PBF test reactor at the INEEL. This
test reactor had a uranium dioxide driver core and a central flux trap contained in an open tank reactor
vessel. An independent, pressurized water coolant loop gave the capability to provide a wide range of
thermal hydraulic conditions within the reactor core. A complete description of the Power Burst Facility
can be found in the test results reports for each of the individual experiments. A3M, A3 "3The PBF core was a
right-circular annulus, 1.3 m in diameter and 0.91 m in height, enclosing a vertical flux trap, 0.21 m in
diameter. The reactor core was designed for both steady-state and power burst operation and contained
eight control rods for reactivity control during steady-state operation and four transient rods for dynamic
control during rapid reactivity transients. Each control and transient rod consisted of a stainless-steel
canister which contained a cylindrical annulus of boron carbide. The control and transient rods were
operated in an air-filled shroud.
An in-pile tube, placed in the central flux trap region, contained the test assembly. The in-pile tube
used for the PBF reactor was a thick walled, Inconel, high strength pressure tube designed to contain the
steady-state operating pressure and subsequent pressure surges resulting from the failure of any test fuel
rod. Therefore, any conceivable failure of the test fuel during the experiment such as clad failure, gross
fuel melting, fuel-coolant interactions, fuel failure propagation, fission product release or metal-water
interactions, would be contained by the in-pile tube without damage to the driver core.
A flow tube was positioned in the in-pile tube. This tube directed coolant flow during the
experiment. The coolant entered the top of the in-pile tube above the reactor core and flowed down the
annulus between the in-pile tube wall and the flow tube. The flow through the reactor reversed at the
bottom, flowed upward around the outside wall of the test train assembly maintaining the appropriate heat
transfer and keeping the shroud properly cooled during an experiment.
The objectives of the PBF SFD test series were to obtain data to better understand fuel behavior,
fission product release, deposition and transport, and hydrogen generation during severe fuel damage
accidents. The data from each of these well controlled and instrumented tests was verified and qualified by
an independent data review committee. Two additional purposes of these tests were to reconstruct the
TMIl-2 accident scenario and to investigate other risk-dominant accident scenarios that could lead to severe
core damage during an incident in a nuclear power plant.
Three PBF SFD tests were used to assess SCDAPIRELAP5/MOD3.3. These tests were; (1) PBF
SFD-ST test, which involved a gradual boiuloff of coolant, (2) PBF SFD 1-1 test, which involved heatup
in a steam environment, and (3) PBF SFD 1-4 test, which involved heatup of a bundle of fuel rods with
burnup and with a control rod. The assessment results obtained for each of these three experiments
follows.
AM- NUREGICR-6150-Rev. 2, Vol. 5

Appendix A3
A3.1 Assessment Using PBF SFD-ST Test
The PBF SFD Scoping TestA3-I was used to assess code calculated reflood behavior and evaluate the
models for cladding meltdown and oxidation.
The SFD-ST test train incorporated a fuel bundle containing 32, 0.92 m in length pressurized, trace
irradiated fuel rods with an average burnup of 91 MWd/MtU in a 6 x 6 array with the comers removed.
The spacing between the fuel rods in the test bundle was maintained by three inconel spacer grids, 40 mm
in height, located at the 0.05, 0.45 and 0.86 m elevations. The fuel bundle was surrounded by an insulating
shroud to minimize heat losses during the test. The shroud consisted of an inner zircaloy liner, a layer of
porous zirconia, and an outer wall constructed of zircaloy. The fuel rods and shroud were instrumented
with numerous thermocouples and pressure sensors. All test data was verified or categorized by a data
integrity review committee.
A3.1.1 Nodalization
The entire test assembly for Scoping Test was modeled with SCDAP/RELAP5. Ten (10) 0.1 m in
length RELAP5 hydrodynamic volumes were used to represent the core region, a single RELAP5 volume
to represent the bypass region, and time dependent volumes to represent a sink and source for the test
bundle flow channel and core bypass. The RELAP5 nodalization scheme used for the analysis is shown in
Figure A3-1. The bundle components, fuel rods and insulating shroud are represented by four SCDAP
1.0
Bypass Bundle
0.
Figure A3-1. RELAP5 nodalization of the test train for the SFD-ST experiment.

Appendix A3
components as shown in Figure A3-2. The fuel rods in the 36 rod test bundle were represented by three
SCDAP components as follows: (1) component 1, the four fuel rods of the inner ring, (2) component 2, the
twelve fuel rods of the middle ring, and (3) component 3, the sixteen fuel rods of the outer ring. A fourth
SCDAP component was used to represent the flow shroud. The SCDAP components were modeled with
ten axial nodes 0.1 m in height which interfaced with the 10 corresponding RELAP5 hydrodynamic
control volumes. The external surface of the bundle shroud was connected to the single RELAP5
hydrodynamic control volume that was used to represent the bypass. The fuel rods for this test were
modeled using 4 radial nodes and the insulating shroud modeled using 19 radial nodes.
O Outer rods
Middle rods
o Inner rods
Figure A3-2. SCDAP core representation of the SFD-ST experimental bundle.
A3.1.2 Boundary Conditions
The nuclear power used for the SFD-ST assessment calculations was obtained from reactor power
measurements and reactor physics calculations. The total bundle power is shown in Figure A3-3. At 11960
s, the bundle power was ramped down attaining a value of zero at 12520 s. The uncertainty envelope for
the total measured power was + 15%. During the test the axial power profile varied in response to changes
in the water level in the reactor core, therefore three discrete power profiles were used, one profile
represented the partially water filled bundle, a second a 10% water filled bundle, and the third a steam
filled bundle. The measured initial conditions for the water inlet flow rate, 16 g/s at 525 K, and a system
pressure of 6.65 MPa were used for the calculation. The pre-test measured thermal conductivity for the
shroud material was used for the analysis even though there was evidence that variations in the shroud
thermal conductivity occurred during the test.
The flow rate of coolant at the inlet of the test fuel assembly varied with time. From 1889 s to 11760
s, the inlet flow rate was 16 g/s. At 11913 s, the inlet flow was reduced to 15 g/s. At 12230 s, the inlet flow
was reduced to 12 g/s. At 12739 s, the inlet flow rate was increased to 30 g/s to reflood the hot test fuel
assembly. The history of the inlet flow rate to the test fuel assembly is tabulated in Table A3-1.

Appendix A3
100000.0
80000.0
= 600030.0
40000.0
u 20000.0
0.00
0.0 5000.0
Time (s)
Figure A3-3. Total bundle power for the SFD-ST experiment.
Table A3-1. History of inlet flow rate to the test fuel assembly for PBF SFD ST Test
Inlet flow
rate
(s) (g/s)
0.0 12.75
1889. 16.20
2752. 16.92
4517. 16.74
5000. 16.79
6892. 16.60
7870. 16.57
9000. 16.52
10000. 16.20
11739. 16.20

Appendix A3
Table A3-1. History of inlet flow rate to the test fuel assembly for PBF SFD ST Test (Continued)
Inlet flow
Time e
rate
(s) (g/s)
11913. 15.02
12091. 15.02
12230. 12.24
12326. 12.51
12400. 15.05
12509. 16.20
12557. 15.18
12617. 15.32
12739. 29.51
15000. 30.00
5000. 16.79
A3.1.3 Results
The results for the PBF SFD Scoping Test provide benchmarks for the modeling of early-phase
damage progression in fuel assemblies and for the modeling of the affect of reflood on the behavior of hot
partially oxidized fuel assemblies. The experiment also involved heat transfer and flow losses in porous
debris. The experiment subjected the fuel rods with only a trace amount of burnup to conditions that
caused oxidation of the cladding of the test fuel rods. The boiloff of coolant in the test fuel assembly
occurred over an extended period of time. After a significant oxidation and heatup of the upper part of the
test fuel assembly, the test fuel assembly was reflooded. The temperature history of the test fuel assembly
at three different elevations was obtained during the period in which oxidation was causing a heatup of the
test fuel rods. A measurement was obtained of the amount of hydrogen produced by oxidation of the
cladding of the test fuel rods and other structures. An instrument obtained the time of rupture of the fuel
rod cladding due to excessive stress and ballooning. A post-irradiation examination (PIE) obtained
information on the extent of oxidation of the test fuel assembly and the location of slumped fuel rod
material.
The MOD3.3 calculated temperature behavior of the test fuel assembly was in good agreement with
its measured temperature behavior. The MOD3.3 calculated and measured temperature histories of a test
fuel rod in the center of the test fuel assembly at the 0.7 m elevation (referenced from bottom of fuel rod)
are shown in Figure A3-4. The MOD3.2 calculated temperature history is also shown. The temperature
measurement failed after the temperature exceeded 2000 K. The MOD3.3 calculated temperature history
and the measured temperature history are in good agreement. The temperature was calculated to begin to
decrease beginning at 11960 s due to the rapid decrease in the nuclear heat generation in the test fuel
assembly. The calculated and measured temperature histories of the test fuel rods at the 0.5 m and 0.35 m

Appendix A3
elevations are compared in Figure A3-5 and Figure A3-6, respectively. The 0.5 m elevation is near the
midplane of the test fuel rods. For both the 0.5 m and 0.35 m elevations, MOD3.3 calculated a significant
increase in temperature to occur after the beginning of reflood. MOD3.2 did not calculate any increase in
temperature after the beginning of reflood. The PIE results indicated that the maximum temperature in the
test fuel assembly exceeded 2670 K. The maximum temperature of the test fuel rods was calculated by
MOD3.3 to be 2585 K and to occur at 12375 s, which was during the beginning of the reflood period of the
test.
3000.0
2000.0
CL
E
I-
1000.0
0.0 L_
0.0 5000.0 10000.0 15000.0
Time (s)
Figure A3-4. Comparison of calculated and measured cladding temperatures at the 0.7 m elevation for
PBF SFD ST test.
agreement with the axial distribution obtained from the PIE performed on the test fuel assembly. The
measured and the MOD3.3 and MOD3.2 calculated axial distributions in the fraction of the fuel rod
cladding oxidized are tabulated in Table A3-2. Double-sided oxidation of the fuel rod cladding was
calculated to occur in the elevation interval of 0.8 m to 0.9 m. The PIE results showed 100% oxidation of
the cladding of the test rods in the elevation interval of 0.35 m to 0.95 m (top of fuel rods). MOD3.3
calculated 100% oxidation of the cladding in the interval of 0.25 m to 0.95 m. Since cladding meltdown
has an influence on the axial distribution in oxidation, the calculated locations of cladding slumping and
blockages are also shown in Table A3-2. The blockage is defined to be the fraction of coolant space filled
with slumped material. The PIE indicated some slumping of fuel and cladding in the elevation interval of
0.4 m to 0.6 m, and some slumped material at the elevation of 0.15 m. The PIE indicated that a large
amount of material slumping did not occur. The slumping of cladding was calculated by MOD3.3 to not
occur. The only material calculated by MOD3.3 to slump was the Inconel spacer grid material and the Zr
cladding melted due to eutectec reaction with the spacer grids.

Appendix A3
3000.0
Measured at 0.5 m Failure of measurement

Calculated, MOD3.3
.... .......... Calculated, MOD3.2
2000.0
E.
E
I-
1000.0
in water
nuclear heat generation
0.00.C
1
0 5000.0 10000.0 15000.0
Time (s)
Figure A3-5. Comparison of calculated and measured temperatures at 0.5 m elevation for PBF SFD ST
test.
3000.0
Measured at 0.35 m ailure of measurement

Calculated, MOD3.3
-----------.... Calculated, MOD3.2
2000.0
SD
(D
0.6
E
(D
1000.0
Cracking of oxide/
........... layer due to rise
. . . .. in water level
nuclear heat generation
0.0
0.0 5000.0 10000.0 15000.0
Time (s)
Figure A3-6. Comparison of calculated and measured temperatures at 0.35 m elevation for PBF SFD ST
test.

Appendix A3
Table A3-2. Comparison of calculated and measured axial distributions in oxidation and meltdown for
PBF SFD ST Test.
Fraction blockage due

Fraction cladding to meltdown/slumping
oxidized from location (yes or
no)
(m) MOD3.3 Measured MOD3.3 Measured
0.05 0.0 0.0 0.0/no 0.0/no

0.15 0.0 0.03 0.0/no >0.0/no
0.25 0.03 1.0 0.0/no 0.0/no
0.35 1.0 1.0 0.0/no 0.0/no
0.45 1.0 1.0 0.0/no 0.0/yes
0.55 1.0 1.0 0.0/no 0.0/yes
0.65 1.0 1.0 0.0/no 0.0/no
0.75 1.0 1.0 0.0/no 0.0/no
0.85 1.0 1.0 0.0/no 0.0/no
0.95 1.0 1.0 0.0/no 0.0/no
The MOD3.3 calculated hydrogen production was in good agreement with the measured hydrogen
production. A hydrogen production of 150 g + or - 35 g was measured. MOD3.3 calculated a hydrogen
production of 125 g. The hydrogen production calculated by MOD3.2 was 130 g. The MOD3.3 calculated
cumulative hydrogen production is shown in Figure A3-7. The bulk of the hydrogen production was
calculated to occur in the period of 8500 s to 12400 s. About 25% of the hydrogen production was
calculated to occur after the beginning of reflooding of the test fuel assembly, when the nuclear heat
generation rate in the test fuel assembly was decreasing rapidly and the flow of coolant into the test fuel
assembly was increasing. These MOD3.3 results are consistent with the measured rate of hydrogen
production, which had a large amount of uncertainty but nevertheless provided a measurement of the trend
of the hydrogen production. The MOD3.2 calculated cumulative hydrogen production is also shown in
Figure A3-7. MOD3.2 did not calculate any hydrogen production after the beginning of reflooding.
The ballooning and rupture of the fuel rod cladding affects the temperature history, oxidation and
meltdown of the fuel rods. The fuel rod cladding was measured to rupture in the period of 5850 s to 6260 s.
The fuel rod cladding was calculated by MOD3.3 to rupture in the period of 6015 s to 6325 s and at a
cladding temperature of about 1050 K. The location of cladding rupture was calculated by MOD3.3 to be
at the elevation of 0.85 m. The severely damaged post-test state of the test fuel rods precluded a
measurement of the elevation interval of cladding ballooning.
The calculated boiloff of water in the test fuel assembly was generally in good agreement with the
measured boiloff. The boiloff at water was mapped by the collapsed liquid level in the test fuel assembly.

Appendix A3
0.20 'I I
I
Calculated, MOD3.3
..............
Calculated, MOD3.2
._ 0.15 Measured: 0.15 kgt0.035 kg
0
"o0 Accelerated oxidation due to
0
cracking of oxide layers
0.
75 0.10
CO
Beginning of decrease
in nuclear heat generation
E
I I I.
0.00
0.i0 5000.0 10000.0 15000.0
Time (s)
Figure A3-7. Calculated cumulative hydrogen production for PBF SFD ST test.
The MOD3.3 calculated and measured collapse liquid levels are compared in Figure A3-8.
1.0
0.8
E
_• 0.6
S0.4
0
0.
0.2
0.0 1
0.0 5000.0 10000.0 15000.0
Time (s)
Figure A3-8. Comparison of calculated and measured collapsed liquid level in test fuel assembly for
PBF SFD ST test.

Appendix A3
The comparison of calculated and measured test fuel assembly pressure differentials indicates that
the model in MOD3.3 for flow losses in porous debris is consistent with experimental results. The
experimental results were useful only for identifying trends in the thermal hydraulic behavior of the test
fuel assembly, in part due to an uncertainty in measured pressure caused by failure of the shroud. Both the
measured pressure differential and the MOD3.3 calculations indicated a significant increase in flow losses
after the test fuel assembly disintegrated into porous debris in the period of 12300 s to 12600 s. The
MOD3.3 calculated and measured increases in test fuel assembly pressure differentials during the period
of disintegration of the test fuel assembly were 6 kPa and 20 kPa, respectively.
In summary, this test problem showed that MOD3.3 calculates early-phase damage progression in
agreement with experimental results. The MOD3.3 calculated time of rupture of the cladding of the test
fuel rods was in good agreement with the measured time of cladding rupture. The calculated boiloff of
coolant and oxidation of the test fuel assemly resulted in a calculated heatup of the test fuel assembly in
good agreement with the measured heatup. The calculated and measured axial distribution in the fraction
of oxidation of the cladding of the test fuel assembly were in good agreement. MUD3.3 calculated that the
meltdown of the test fuel assembly was limited to meltdown resulting from the eutectic reaction of Inconel
spacer grids with Zircaloy cladding. The PIE indicated a limited amount of slumping of liquefied cladding
with dissolved fuel. While the measured extent of meltdown was greater than the calculated extent of
meltdown, nevertheless the differences in calculated and measured meltdown were not large. Both the
MOD3.3 calculations and the measurements indicated a significant oxidation excursion and heatup after
reflood of the test fuel assembly began and its nuclear heat generation rate was rapidly decreasing. The test
problem also provided a limited test of the model in MOD3.3 for calculating the flow losses in porous
debris. Both the MOD3.3 calculations and the measurements indicated a significant increase in flow losses
after the beginning of reflood and the disintegration of part of the test fuel assembly into porous debris.
A3.2 Assessment Using the PBF SFD 1-1 Test
PBF SFD 1-1 TestA3 -2 was the second in the series of severe fuel damage experiments performed in
the Power Burst Facility at the INEEL. The SFD 1-1 test train incorporated a fuel bundle containing 32, 1
m in length pressurized trace-irradiated fuel rods with an average burnup of 70 MWd/MtU surrounded by
a multi-layer insulating shroud. The fuel rods in the test assembly were held in position by three 40 mm in
height inconel spacer grids located at the 0.05, 0.45, and 0.86 m elevations above the bottom of the test
bundle. The insulating shroud consisted of an inner zircaloy liner, a porous zirconia insulating layer, and a
zircaloy wall. The experimental bundle was positioned in a zircaloy flow tube, which was inserted into the
test facility in-pile-tube. The insulated region was pressured with argon and the gap between the inner and
outer shroud walls was pressurized with helium. Numerous thermocouples recorded fuel rod and shroud
temperatures during the test and pressure sensors recorded changes in system pressure. All test data was
verified, validated and qualified by an independent data integrity review committee.
A3.2.1 Nodalization and Boundary Conditions
The entire test assembly for PBF SFD 1-1 was modeled with SCDAP/RELAP5. Ten (10) 0.1 m in
length RELAP5 hydrodynamic volumes were used to represent the core region, a single RELAP5 volume
to represent the bypass region, and time dependent volumes to represent a sink and source for the test
bundle flow channel and core bypass. The test facility nodalization used RELAP5 hydrodynamic volumes
and heat structures to represent the entire test facility including the deposition rod, steam line, condenser,
and knockout drum. The RELAP5 nodalization scheme used for the analysis is shown in Figure A3-9. The
bundle components, fuel rods and insulating shroud are represented by four SCDAP components as shown
in Figure A3-10. The fuel rods in the 36 rod test bundle were represented by three SCDAP components as

Appendix A3
Steam line
Check valve
Condenser
Heat structure
-Deposition rod
Upper
plenum
Stack of ten RELAP5
RELAP5 control
volume representing
coolant in bypass
0.
M2,2-BDA-1 113-00
Figure A3-9. RELAP5 nodalization of the test train for the SFDI-1 experiment.
follows: (1) component 1 represented the four fuel rods of the inner ring, (2) component 2 represented the
twelve fuel rods of the middle ring, and (3) component 3 represented the sixteen fuel rods of the outer ring.
A fourth SCDAP component was used to represent the insulating shroud. Each SCDAP component was
divided into ten axial nodes with each axial node 0.1 mm height. The axial nodes interfaced the 10
corresponding RELAP5 hydrodynamic control volumes. The external surface of the insulating shroud
interfaced with the RELAP5 control volume that was used to represent the coolant in the bypass. The fuel
rods for this test were modeled using 4 radial nodes and the insulating shroud was modeled using 19 radial
nodes.

Appendix A3
w 'r# • •ei •' w • shroud

is •012 0(component4)
* Inner rods (component 1)

) Middle rods (component 2)
O Outer rods (component 3)
Figure A3-10. SCDAP nodalization of the PBF SFDI-1 experimental bundle.
The bundle nuclear power generation and bundle coolant boundary conditions were defined
according to experimental results. An estimation of nuclear power generated in the test bundle, as well as
the uncertainty envelop of+ 7% and -15%, was obtained from the measured reactor power and uncertainty
calculations. This estimation was used for the power input for the assessment calculations. The single axial
power profile used for the calculations was determined from measurements recorded from the fission
chambers located on the outer wall of the shroud and reactor physics calculations. The measured water
inlet flow rate of 0.64 g/s at 531 K with a measured system pressure of 6.8 MPa was defined in the input
deck. Argon injected at a flow rate 1.3 g/s, beginning at 3,200 s and continuing to the end of the
experiment, was also modeled.
A3.2.2 Comparison of Calculations and Measurements
The results for the PBF SFD 1-1 experiment provide benchmarks for the early-phase damage
progression models in a severe accident analysis computer code.A3"2 The experiment subjected the test
fuel rods to conditions that caused rapid oxidation and meltdown of the cladding of the test fuel rods. The
temperature history of the test fuel assembly at three different elevations was obtained during the period in
which oxidation was causing a rapid heatup of the test fuel rods. A measurement was obtained of the
amount of hydrogen produced by oxidation of the cladding of the test fuel rods and other structures. An
instrument obtained the time of rupture of the fuel rod cladding due to excessive stress and ballooning. A
post-irradiation examination (PIE) obtained information on the extent of oxidation of the test fuel
assembly and the location of slumped fuel rod material.
The MOD3.3 calculated temperature behavior of the test fuel assembly was generally in good
agreement with its measured temperature behavior. The MOD3.3 calculated and measured temperature
history of the test fuel rods in the middle row of test fuel assembly at the 0.35 m elevation (referenced from
bottom of fuel rod) are shown in Figure A3-1 1. The calculated rate of temperature increase at the
temperature of 1700 K was about 13 K/s. The calculated rate of temperature increase was in agreement

Appendix A3
with the measured rate of temperature increase. The temperature measurement failed after the temperature
exceeded 1700 K. The calculated temperature history has an inflection point at the time of 1976 s
(temperature of 2200 K) due to the relocation of the cladding in the circumferential direction, which in turn
was due to melting of the cladding. The temperature history calculated by MOD3.2 at this elevation is also
shown in Figure A3- 1. The temperature histories calculated by MOD3.2 and MOD3.3 are similar. The
calculated and measured temperature histories of the test fuel rods at the 0.5 m and 0.7 m elevations are
compared in Figure A3-12 and Figure A3-13, respectively. The 0.5 m elevation is near the midplane of the
test fuel rods. The calculated temperature history at the 0.5 m elevation has an inflection point at 1960 s
due to the slumping of melted cladding. The calculated temperature history at the 0.7 m elevation has a
similar inflection point due to the slumping of melted cladding. The maximum calculated temperature of
the test fuel rods was 2818 K in the elevation interval of 0.3 m to 0.5 m. The maximum measured
temperature of the test fuel rods, obtained from PIE results, was 2890 K at the elevation of 0: 15 m.A3"1 The
temperature histories calculated by MOD3.2 are also shown in Figure A3-12 through Figure A3-13. The
MOD3.3 and MOD3.2 calculations are similar.
3000.0
2000.0
0
CL
E
0.
E
1000.0
0.0 I
0.0 1000.0 2000.0 3000.0 4000.0
Time (s)
Figure A3-1 1. Comparison of calculated and measured temperatures at 0.35 m elevation for PBF SFD
1-1 test.

Appendix A3
3000.0 I 'I 'I'

Failure of measurement , .. ,,
2000.0
2
CU
IP
CL "/#•
E
a)
1000.0
Measured at 0.5 m
-------- Calculated, MOD3.3

.............. Calculated, MOD3.2
0.0
0.(0 1000.0 2000.0 3000.0 4000.0
Time (s)
Figure A3-12. Comparison of calculated and measured temperature at elevation of 0.5 m for PBF SFD
1-1 test.
3000.0
2000.0
2
CD
ca
E
I-CD
1000.0
0.0 L
0.0 1000.0 2000.0 3000.0 4000.0
Time (s)
Figure A3-13. Comparison of calculated and measured temperature at elevation of 0.7 m for PBF SFD
1-1 test.

Appendix A3
cladding oxidized are tabulated in Table A3-3. The entries in the table without any data are due to
PBF SFD 1-1 test.
Fraction blockage due to meltdown!
Elevation Fraction cladding oxidized slumpin floca to (estorno
slumping from location (yes or no)
(m) MOD3.3 MOD3.2 Measured MOD3.3 MOD3.2 Meaured.
0.05 0.01 0.0 0.60 0.28/no 0.0/no 0.0/no

0.15 0.80 0.51 1.00 0.61/no 1.0/no -0.5/no
0.25 0.78 0.13 1.00 0.0/no 4- 0.0/no
0.35 0.78 0.16 0.25 0.0/no 4- 0.0/yes
0.45 0.38 0.36 0.65 0.06/yes 4- 0.0/no
0.55 0.36 - - 0.0/yes 4
0.65 0.25 0.38 0.45 0.0/yes 4- 0.0/no
0.75 0.18 0.46 0.11 0.0/yes 0.0/no 0.0/yes
0.85 0.16 0.34 0.20 0.0/yes 0.0/no 0.0/yes
0.95 0.63 0.0/no 0.0/no
measurements not being obtained at that elevation. Double-sided oxidation of the fuel rod cladding was
calculated to occur in the elevation interval of 0.4 m to 0.6 .n. The PIE results showed almost 100%
oxidation of the cladding of the test rods in the elevation interval of 0.1 m to 0.3 m. MOD3.3 calculated
about 80% oxidation of the cladding in the interval of 0.1 m to 0.4 m. MOD3.2 underpredicted
significantly the extent of oxidation in the interval of 0.1 m to 0.4 m. Both MOD3.3 and MOD3.2
underpredicted somewhat the extent of oxidation in the elevation interval of 0.4 m to 0.7 m. In the 0.7 m to
0.9 m interval of elevation, the MOD3.3 calculations are in generally good agreement with the measured
results, while the MOD3.2 calculations significantly overpredicted the extent of oxidation. Since
meltdown has an influence on the axial distribution in oxidation, the calculated locations of cladding
slumping and blockages are also shown in Table A3-3. The blockage is defined to be the fraction of
coolant space filled with slumped material. The PIE indicated a significant blockage due to meltdown in
the interval of elevation from 0.1 m to 0.2 m. Both MOD3.3 and MOD3.2 calculated a significant blockage
in the same interval of elevation as indicated by the PIE results. The PIE results indicated the slumping of
melted cladding in the elevation interval of 0.7 m to 0.9 m. MOD3.3 also calculated slumping in this
interval of elevation but MOD3.2 did not. This improvement by MOD3.3 in modeling the slumping of
cladding is the reason MOD3.3 did not overpredict the extent of oxidation of cladding in this interval of
elevation, as did MOD3.2.
The MOD3.3 calculated hydrogen production was greater than the measured hydrogen production. A
hydrogen production of 64 + or - 7 g was measured by analysis of the contents of the collection tank. A
hydrogen production of 99 g was measured from PIE results. MOD3.3 calculated a hydrogen production of
105 g. The hydrogen production calculated by MOD3.2 was 94 g. The calculated and measured rate of
A3-15 NLTREG/CR-6150-Rev. 2, Vol. 5

Appendix A3
hydrogen production are compared in Figure A3-14. The bulk of the hydrogen production was calculated
to occur in the period of 1800 s to 2300 s. The bulk of the hydrogen production was measured to occur in
the period of 2100 s to 2600 s. The calculated rate of hydrogen production was somewhat greater than the
measured rate of hydrogen production. Since the rate of hydrogen production was measured down stream
of the test fuel assembly, the measured rate of hydrogen production lagged the rate of hydrogen production
in the test fuel assembly and the peak measured rate of hydrogen production was a round off of the peak
rate of hydrogen production in the test fuel assembly. The lag time is estimated to be 300 s.A31 Taking
into account this lag in the measured rate and round off of the measured rate, the calculated and measured
rates are in fair agreement. Both the calculations and measurements show most of the hydrogen production
occuring within a 500 s period. Both the calculations and the measurements show a sharp decrease in
hydrogen production occurring almost right after the maximum rate of production. According to the
calculations, this sharp decrease in the rate of hydrogen production was due to slumping of melted
cladding.
0.00025
Measured (downstream o on)
---------- Calculated, MOD3.3
-- 0.00020
(D
0.00015
._C
0
CD
D 0.00010
-1-
I- 0.00005
0.00000 ' ' "

0.0 1000.0 2000.0 3000.0 4000.0
Time (s)
Figure A3-14. Comparison of calculated and measured rate of hydrogen production for PBF SFD 1-1
test.
The fuel rod cladding was calculated to rupture in the period of 1785 s to 1798 s. Although the severely
damaged post-test state of the test fuel rods precluded an accurate measurement of the elevation interval of
cladding ballooning, nevertheless sufficient evidence was present to estimate the elevation interval of
cladding ballooning to be from 0.3 m to 0.7 m above the bottom of the fuel rods. The calculated elevation
interval of cladding ballooning was from 0.4 m to 0.6 m above the bottom of the fuel rods. In reference to
the elevation interval of cladding ballooning, ballooning is defined to have occurred where the cladding
hoop strain exceeded 0.12. In the calculations, double-sided oxidation occurred where the cladding hoop
strain exceeded 0.12 and the cladding had ruptured.

Appendix A3
agreement with experiment results. The calculated oxidation driven heatup of the test fuel assembly was in
good agreement with the measured oxidation driven heatup of the test fuel assembly. The calculated
meltdown of the test fuel assembly was in good agreement with the measured meltdown of the test fuel
assembly. The rapid production of hydrogen was both calculated and measured to occur within a 500 s
period. The rate of production of hydrogen was calculated to decrease rapidly after the time of the peak
rate of hydrogen production. This characteristic of hydrogen production was also in the measured rate of
hydrogen production. The MOD3.3 calculations of the axial distribution in oxidation of the test fuel
assembly was an improvement over the MOD3.2 calculations of axial distribution in oxidation. MOD3.3
correctly calculated a significant oxidation of the lower part of the test fuel assembly and a meltdown of
the upper part of the test fuel assembly. Both MOD3.2 and MOD3.3 underpredicted somewhat the extent
of oxidation of the slumped cladding in the blockage area near the bottom of the test fuel assembly. The
MOD3.3 calculated time of rupture of the cladding of the test fuel rods was later than the measured time of
cladding rupture. Nevertheless, the overall calculation of early-phase damage progression was in good
agreement with the measurements of early-phase damage progression.
A3.3 Assessment Using the PBF SFD 1-4 Test
The SFD 1-4 test A3-3 was the fourth in a series of severe core damage experiments performed in the
Power Burst Facility at the Idaho National Engineering and Environmental Laboratory (INEEL). The SFD
1-4 test train incorporated a fuel bundle containing 26 pre-irradiated fuel rods with an average burnup of
36 GWd/MtU, two fresh, instrumented fuel rods and four silver-indium-cadmium control rods in a 6 x 6
array with the comer removed. All of the fuel rods were pressurized. One of the four control rods was
instrumented. The absorber material in the control rods was an alloy consisting of 80% Ag, 15% In, and
5% Cd by weight. The spacing between fuel rods and control rod tubes in the test assembly was
maintained by three inconel spacer grids -40 mm in height located at the 0.09, 0.49, and 0.94 m elevations
above the bottom of the test bundle. The active length of the fresh and irradiated fuel rods was 1 m.
The fuel bundle was housed in an insulated shroud to minimize radial heat losses. The layered shroud
was composed of a zircaloy liner, insulating material (porous zirconia), a zircaloy wall region surrounding
the shroud. The shrouded test bundle was inserted in a zircaloy flow tube, which in turn was centered in the
in-pile tube. The insulated region was pressurized with argon, and the gap between the inner and other
shroud walls was pressurized with helium.
A3.3.1 Nodalization
As shown in Figure A3-15, the test bundle was represented by seven SCDAP components;
(1) component 1 represented the three irradiated fuel rods in the inner ring of bundle fuel rods, (2)
component 2 represented the fresh fuel rod in the inner ring, (3) component 3 represented the fresh fuel
rods in the middle ring, (4) component 4 represented the seven irradiated fuel rods in the middle ring, (5)
component 5 represented the four control rods in the middle ring, (6) component 6 represented the 16
irradiated fuel rods in the outer ring, and (7) component 7 represented the insulating shroud. The rods fuel
components for this experiment were modeled radially with four radial nodes, the control rods were
modeled using five radial nodes, and the shroud was modeled using 19 radial nodes.
As shown in Figure A3-16, all components were modeled using 10 axial nodes, each node being 0.1
m in height and interfacing with one of the ten corresponding RELAP5 hydrodynamic control volumes.
The external surface of the shroud was connected to a single RELAP5 hydrodynamic volume representing
the bypass.

Appendix A3
Shroud
(component 7)
Fresh rod 3B (component 2)

SFresh
rod 4D (component 3)
o Irradiated rods, inner ring (component 1)
* Irradiated rods, outer ring (component 6)
(Vi) Irradiated rods, middle ring (component 4)
C) Control rods, (component 5)
Figure A3-15. SCDAP nodalization of the PBF SFD 1-4 test bundle.
Stack of ten RELAP5

control volumes
RELAP5 control representing coolant
volume representl In test bundle
coolant In bypass
Shroud
Figure A3-16. RELAP5 nodalization of the PBF SFD 1-4 test bundle.
NLTREG/CR-6150-Rev. 2, Vol. 5 A3-18

Appendix A3
A3.3.2 Initial Conditions
All fuel rods were pressurized with helium to 2 MPa at room temperature, with the exception of fresh
fuel rod (4D), which was pressurized to 3.8 MPa. The calculations started at the experiment time of 1,000
s, at which was the time for completion of the boildown phase and when the two-phase level in the bundle
was -0.20 m. At this time, the mean temperature of the bundle fuel rods was 810 K.
The best-estimate fuel rod power history was used in the calculations. The transient axial power
profiles were represented in the calculations using three discrete curves. The calculations were started
using a profile corresponding to an intact bundle partially filled with water, then shifted to a profile
representing an intact bundle filled with steam (1,020 seconds) and finally to a profile representing a
damaged bundle with relocated control rod materials and filled with steam (2,050 seconds).
A constant water inlet flow rate of 0.6 g/s with a temperature of 532 K and a pressure of 6.95 MPa
was used over the experimental transient. Throughout the experiment, argon injected from the bottom
flowed through the test bundle. For the first 1965 seconds of the experiment argon flow was maintained at
a rate of 0.26 g/s; from 1,965 to 3,280 seconds the argon flow rate was maintained at 0.58 g/s; from 3,280
seconds to the termination of the experiment the argon flow rate was maintained at 1.88 g/s.
A3.3A Comparison of Calculations and Measurements
The results for the PBF SFD 1-4 experiment provide benchmarks for the modeling of early-phase
damage progression in fuel assemblies with a significant amount of burnup.A 3 The experiment subjected
the test fuel rods with a burnup of 36 MWd/kgU to conditions that caused rapid oxidation and meltdown of
the cladding of the test fuel rods. The temperature history of the test fuel assembly at three different
elevations was obtained during the period in which oxidation was causing a rapid heatup of the test fuel
rods. A measurement was obtained of the amount of hydrogen produced by oxidation of the cladding of
the test fuel rods and other structures. An instrument obtained the time of rupture of the fuel rod cladding
due to excessive stress and ballooning. A post-irradiation examination (PIE) obtained information on the
extent of oxidation of the test fuel assembly and the location of slumped fuel rod material.
its measured temperature behavior. The MOD3.3 calculated and measured temperature histories of the
fresh test fuel rod in the center of the test fuel assembly at the 0.4 m elevation (referenced from bottom of
fuel rod) are shown in Figure A3-17. The MOD3.2 calculated temperature history is also shown. The
temperature measurement failed after the temperature exceeded 2000 K. The MOD3.3 calculated
temperature history and the measured temperature history are in good agreement. The rate of heatup
calculated by MOD3.3 is slightly faster than that calculated by MOD3.2. The MOD3.3 calculated
temperature history at the time of 2040 s (temperature of 2350 K) has an inflection point due to a reduction
in the rate of oxidation caused by the presence of dissolved fuel in the metallic part of the cladding. The
rate of heatup was calculated by MOD3.3 to begin to decrease at 2350 s due to complete oxidation of the
cladding and melting of the fuel. The calculated and measured temperature histories of the test fuel rods at
the 0.54 m and 0.74 m elevations are compared in Figure A3-18 and Figure A3-19, respectively. The 0.54
in elevation is near the midplane of the test fuel rods. The calculated rate of heatup at the 0.74 m elevation
was limited due to steam starvation. The rate of heatup at the this elevation was calculated to be
significantly less than that at the 0.4 m elevation due to steam starvation. The rate of heatup at 0.74 m
elevation began to decrease at 2600 s due to the decrease at this time in the nuclear heat generation rate in

Appendix A3
3000.0 1 I '
Failure of measurement ........- ,
S2000.0 , / "
a) I I •".
,,. 7.
E - .
a)
- 1000.0 Measured at 0.4 m
--------------- Calculated, MOD3.3
............... Calculated, MOD3.2
0.0 I
1000.0 2000.0 3000.0 4000.0
Time (s)
Figure A3-17. Comparison of calculated and measured temperatures at 0.4 m elevation for PBF SFD 1-4
test.
the test fuel bundle. The maximum temperature of the test fuel rods was calculated by MOD3.3 to be 2870
K and to occur in the outer row of fuel rods in the elevation interval of 0.4 m to 0.6. The PIE results
indicated the maximum temperature of the test fuel assembly to be greater than 2800 K and to occur in the
elevation interval of 0.5 m to 0.6 mn.A3 -3 A measurement was obtained of the temperature of the shroud
mid-wall through out the high temperature period of the experiment. The MOD3.3 calculated and
measured temperature of the shroud mid-wall at 0.91 m are compared in Figure A3-20. The good
agreement of the calculated and measured temperatures indicates that MOD3.3 correctly calculated the
temperature behavior of the test fuel assembly during the high temperature period of the test.
cladding oxidized are tabulated in Table A3-4. The entries in the table without any data are due to
calculated to occur in the elevation interval of 0.6 m to 0.8 m. The PIE results showed almost 100%
about 70% oxidation of the cladding in the interval of 0.2 m to 0.6 m. MOD3.2 underpredicted
significantly the extent of oxidation in the interval of 0.2 m to 0.3 m. MOD3.3 overpredicted by about a
factor of two the extent of oxidation in the elevation interval of 0.5 m to 0.6 in. The overprediction of the
extent of oxidation in this interval of elevation may be due to the code not calculating the slumping of
cladding in this interval of elevation. The PIE results indicate that cladding in this interval of elevation
slumped. In the 0.7 m to 0.9 m interval of elevation, the MOD3.3 calculations of the fraction of cladding
oxidation are in generally good agreement with the measured results. Both the calculations and the
measurents indicate that the extent of oxidation of the cladding in this interval of elevation was limited by
slumping of cladding. MOD3.2 did not predict the slumping of cladding in the elevation interval of 0.8 m
to 0.9 in, and as a result the extent of oxidation of the cladding in this interval of elevation was
NUREG/CR-6150-Rev. 2. Vol. 5 A3-20

Appendix A3
3000.0
2000.0
0
Cu
E
_
1000.0 --- Measured at 0.54 m
Calculated,
-------. MOD3.3 .-...-
.............. Calculated, MOD3.2
I I
0.0
100( 0.0 2000.0 3000.0 4000.0
Time (s)
Figure A3-18. Comparison of calculated and measured temperatures at 0.54 m elevation for PBF SFD
1-4 test.
3000.0
2000.0
CL
E.
E
I-
1000.0 Measured at 0.74 m
.............. Calculated, MOD3.3
.............. Calculated, MOD3.2
0.0 I I
100(0.0 2000.0 3000.0 4000.0
Time (s)
Figure A3-19. Comparison of calculated and measured temperatures at 0.74 m elevation for PBF SFD
1-4 test.

Appendix A3
1000.0
-s
800.0
-. . - -
£
600.0
0.
E 400.0
(D
I-
Measured, 0.91 m
Calculated, MOD3.3 .------------..
200.0
0.0
1000.0 2000.0 3000.0 4000.0
Time (s)
Figure A3-20. Comparison of calculated and measured temperatures at shroud mid-wall at 0.91 m for
PBF SFD 1-4 test.
overpredicted. Since cladding meltdown has an influence on the axial distribution in oxidation, the
calculated locations of cladding slumping and blockages are also shown in Table A3-4. The blockage is
defined to be the fraction of coolant space filled with slumped material. The PIE indicated a significant
blockage due to meltdown in the interval of elevation from 0.1 m to 0.2 m. Both MOD3.3 and MOD3.2
calculated a significant blockage in the same interval of elevation as indicated by the PIE results. The PIE
results indicated the slumping of melted cladding in the elevation interval of 0.3 m to 0.9 m. MOD3.3
calculated slumping of the cladding in the interval of elevation from 0.6 m to 0.9 m. MOD3.2 did not
predict slumping of the cladding in the elevation interval of 0.8 m to 0.9 m.
meltdown for PBF SFD 1-4 test.
Fraction blockage due to meltdown/

(m) MOD3.3 MOD3.2 Measured MOD3.3 MOD3.2 Measured
0.05 0.0 0.0 0.10 0.0/no 0.0/no 0.0/no

0.15 0.03 0.02 1.00 0.5/no 0.5/no -0.5/no
0.25 0.61 0.11 1.00 0.0/no 0.0/- -0.1/no
0.35 0.67 0.52 0.50 0.0/no 0.0/- 0.0/yes

Appendix A3
meltdown for PBF SFD 1-4 test.
Elev. Fraction cladding oxidized Fraction blockage due to meltdown/

0.45 0.68 - - 0.0/no 0.0/- 0.0/yes
0.55 0.67 0.37 <0.30 0.0/no 0.0/- 0.0/yes
0.65 0.33 - - 0.0/yes 0.0/- 0.0/yes
0.75 0.22 0.26 0.37 0.0/yes 0.0/- 0.0/yes
0.85 0.24 0.86 0.22 0.0/yes 0.0/no 0.0/yes
0.95 0.90 0.0/no 0.0/no
The MOD3.3 calculated hydrogen production was in good agreement with -the measured hydrogen
production. A hydrogen production of 86 + or - 12 g was measured by analysis of the contents of the
collection tank. MOD3.3 calculated a hydrogen production of 81 g. The hydrogen production calculated
by MOD3.2 was 94 g. The PIE results did not provide a quantitative measurement of the extent of
oxidation of the slumped material in the large blockage region in the elevation interval of 0.1 m to 0.2 m,
but did indicate that a significant fraction of the slumped material in this region was oxidized. MOD3.3
calculated that 30% of the slumped cladding in the blockage region was oxidized. The MOD3.3 calculated
and measured rates of hydrogen production and cumulative hydrogen production are compared in Figure
A3-21 and Figure A3-22. The bulk of the hydrogen production was calculated to occur in the period of
1900 s to 3000 s. The bulk of the hydrogen production was measured to occur in the period of 2100 s to
3300 s. The calculated rate of hydrogen production during the first 120 s of intensive oxidation was greater
than the measured rate of hydrogen production by about a factor of two. Since the rate of hydrogen
production was measured down stream of the test fuel assembly, the measured rate of hydrogen production
lagged the rate of hydrogen production in the test fuel assembly and the peak measured rate of hydrogen
production was a round off of the peak rate of hydrogen production in the test fuel assembly. The lag time
is estimated to be about 300 s.A3 3Taking into account this lag in the measured rate and round off of the
measured rate, the calculated and measured rates are in fair agreement. Both the calculations and
measurements showed that most of the hydrogen production occurred within a 1000 s period. The
measured rate of hydrogen production showed a leveling off at 2100 s, when intensive oxidation was just
beginning and which may have been due to slumping of cladding at the onset of intensive oxidation. The
rate of hydrogen production was calculated to decrease due to the slumping of cladding after about 120 s of
intensive oxidation.
The fuel rod cladding was calculated to rupture in the period of 1360 s to 1530 s and at cladding
temperatures in the range of 1100 K to 1150 K. The severely damaged post-test state of the test fuel rods
precluded a measurement of the elevation interval of cladding ballooning. The calculated elevation
interval of cladding ballooning for the fuel rods with burnup was from 0.5 m to 0.8 m above the bottom of
the fuel rods. In reference to the elevation interval of cladding ballooning, ballooning is defined to have
occurred where the cladding hoop strain exceeded 0.12. In the calculations, double-sided oxidation

Appendix A3
0.00020
-- Measured (downstream locations)
Calculated, MOD3.3 I --------.
0 0.00015 F
cc
C
0
0.00010 F
0
I-N
0
0
0.00005 I-
/
I
0.00000 1 I T mI mI ...
___ I -d. • t.. I -
0.0 1000.0 2000.0 3000.0 4000.0

Time (s)
Figure A3-21. Comparison of calculated and measured rates of hydrogen production for PBF SFD 1-4.
1.0
0.08
0
V
0r
0.06
"0D
CM
0.04
E
0
0.02
0.00
1000.0 2000.0 3000.0 4000.0
Time (s)
Figure A3-22. Calculated cumulative hydrogen production for PBF SFD 1-4 test.

Appendix A3
occurred where the cladding hoop strain exceeded 0.12 and the cladding had ruptured.
agreement with experiment results. The calculated oxidation driven heatup of the test fuel assembly was
in good agreement with the measured oxidation driven heatup of the test fuel assembly. The calculated
meltdown of the test fuel assembly was in fair agreement with the measured meltdown of the test fuel
assembly. The calculated and measured cladding meltdown were in good agreement in the lower and
upper parts of the test fuel assembly. The cladding meltdown at the midplane of the test fuel assembly was
underpredicted. The rapid production of hydrogen was both calculated and measured to occur within
about a 1000 s period. The maximum calculated rate of hydrogen production was somewhat greater than
themaximum measured rate of hydrogen production. This discrepancy may be due to an underprediction
of the measured maximum rate of hydrogen proiduction caused by the downstream position of the
measuring devises. The MOD3.3 calculations of the axial distribution in oxidation of the test fuel
assembly was an improvement over the MOD3.2 calculations of axial distribution in oxidation. MOD3.3
correctly calculated a significant oxidation of the lower opart of the test fuel assembly and a meltdown of
the upper part of the test fuel assembly. The MOD3.3 calculated time of rupture of the cladding of the test
fuel rods was sooner than the measured time of cladding rupture. Nevertheless, the overall calculation of
earlyphase damage progression was is good agreement with the measurements of early-phase damage
progression.
A3.3.5 References
A3-1. A. D. Knipe, S. A. Ploger, and D. J. Osetek, "PBF Severe Fuel Damage Scoping Test - Test
Results Report," NUREG/CR-4683, EGG-2413. August 1986.
A3-2. Z. R. Martinson, D. A. Petti, and B. A. Cook, "Volume 1: "PBF Severe Fuel Damage Test 1-1
Test Results Report," NUREG/CR-4684, EGG-2463, October 1986.
A3-3. D. A. Petti, Z. R. Martinson, R. R. Hobins, C. M. Allison et al., "Power Burst Facility (PBF)
Severe Fuel Damage Test 1-4 Test Results Report," NUJREG/CR-5163, EGG-2542, April 1989.

Appendix A4
A4. ASSESSMENT USING THE CORA TESTS
A series of bundle heating and melting experiments were conducted in the CORA test facility at
Forschungszentrum, Karlsruhe, (FzK) to examine core damage progression under severe accident
conditions. These steam-fed out-of-pile tests used a combination of electrically heated and unheated
bundle fuel rods to simulate decay heating in a reactor core. Due to the ease in configuring the test bundle
and the ability to systematically control test conditions, key severe accident phenomena, such as fuel rod
ballooning, bundle meltdown, and quench-induced fragmentation for both PWR and BWR-configured
bundles were studied. The results of these experiments are useful for assessing models in severe accident
computer codes. A general description of the test facility, a description of the experiments used to assess
SCDAP/RELAP5/MOD3.3, and assessment results are presented in the following sections.
A4.1 General Description of the CORA Test Facility
The CORA test facility consisted of an experimental fuel bundle (25 or 56 rods, heated, unheated,
and optional control rods for the PWR tests, and 16 or 48 unheated and heated fuel rods with a centered
stainless steel clad boron carbide control blade for the BWR tests), an insulating zirconia shroud, a quench
tank, and a high temperature shield. The experimental bundle consisted of a representative 2-m high fuel
rod bundle with a 1-rn electrically heated region to simulate decay heat. The inlet for superheated steam
and argon was on a side near the bottom of the heated section of the experimental bundle, and the hot gases
flowed through the experimental bundle and then into a condenser. The non-condensable gases, hydrogen
formed during zircaloy oxidation and the argon cover gas, are expanded to atmospheric pressure in a
mixing chamber. For safety reasons the hydrogen fraction was diluted by adding compressed air. After
dilution, the composition of the released gas and the mass of hydrogen produced during the experiment
was determined. The fuel rod bundle was surrounded by a porous zirconia insulating shroud. The insulated
test bundle was surrounded by a ceramic high temperature radiation shield.
A steam generator positioned within the containment provided the steam supply to the bundle. A
superheater was used to heat the steam and argon cover gas to temperatures in excess of 800 0 C. The
maximum allowed pressure in the facility was 10 bars, though all experiments used for the MOD3.2
assessment were performed at 2 bar (1 bar overpressure). The temperature of the gases leaving the
superheater dropped about 200 0 C as it flowed through the pipe connecting the superheater to the test
bundle.
The experimental test apparatus had viewing ports through which the melt progression in the test
bundle could be continuously observed. Video recording and high-resolution still photography were used
to provide a permanent visual record of the melt progression and core damage events occurring in the test
bundle during the experiment. Thermocouples mounted on the zircaloy cladding of the heated and
unheated fuel rods, at the fuel centerline of some unheated fuel rods, on the control rod or blade, and at
various locations in the insulating shroud and high temperature shield recorded each component's
temperature change during the test. Mass flow rates of gases, system pressure, and power were
continuously monitored and recorded during the test.
A4.2 Nodalization of the CORA Test Facility
The RELAP5 nodalization diagram of the CORA test facility used to model the CORA PWR
experiments for the assessment is shown in Figure A4-1. The bundle nodalization used for calculations
employed a second shroud component to model the high temperature shield. The two SCDAP shrouds

Appendix A4
calculated the radiation heat transfer between the insulating flow shroud and the high temperature shield.
A cross-section of a typical small bundle, CORA-13, is shown in Figure A4-2. The ceramic high
temperature shield was represented by a second SCDAP shroud component surrounding the test bundle.
The rod configuration of each small bundle PWR test is similar to that used for the CORA-13 test with
minor modifications such as the presence or absence of control rods in the test bundle. The initial
conditions used in all assessment input decks were obtained from test data supplied to the INEEL by FzK.
Insulated flow shroud 2
55
92 9
32
9 5Top of stack of
o RELAP5 control
050 volumes that
High temperature represent coolant
shield 89 49 29 in test bundle
Figure A4-1. General CORA RELAP5 nodalization used for all PWR tests.
A4..3 Test Conditions for the CORA Tests
Each CORA tests was divided into three distinct phases. The test phases were (a) the gas preheat
phase, (b) the transient phase, and (c) the cooling phase. During the gas preheat phase a low constant
power of 0.65 KW was input to the heater rods while preheated argon flowed through the bundle. During
this phase, the bundle insulation was equilibrated at temperatures sufficiently high to prevent steam
condensation during the test. System pressure during this phase was maintained at the desired test pressure
of 0.22 MPa. During the transient phase, power to the heater rods was increased linearly with time to

Appendix A4
High temperature shield,

radial node 1
High temperature shield,
radial node 17 -
Insulated flow shroud, Component #2

radial node 17 Component #3
Insulated flow shroud,

radial node 1 - Component #1
(oute surace)Insulated flow shroud,
(component #4)
Component #5
® Heated fuel rods 0 Unheated fuel rods
() Control rods
Figure A4-2. CORA-13 bundle cross-section basic design used for all PWR tests.
develop in the bundle a heatup rate equal to 1 K/s up to the onset of oxidation. Power to the heater rods for
most tests was maintained at the maximum value for several hundred seconds. Throughout each test, argon
flow through the bundle was maintained at a constant flow rate. Steam for most of the tests was introduced
into the bundle 300 seconds after the initiation of power to the heater rods. The superheated steam flowed
through the bundle for a predetermined period of time, usually until termination of power to the heater
rods. The argon or argon/steam mixture, heated to - 1,000 K in a superheater, cooled as it flowed through
the 40 foot connecting pipe to the bundle inlet. The bundle inlet temperature remained near 800 K during
the early phase of the simulated accident, increasing by approximately 200 K 500 seconds after the
initiation of power to the heater rods. In quench tests, such as CORA-13, where the addition of cooling
water was used to quickly cool the test bundle, the cooling phase was initiated by the raising of a water
filled cylinder at a rate of 1 cm/s to quench the bundle. In non-quench tests the test bundles were allowed
to cool slowly in flowing argon after termination of power to the heater rods.
A4.4 Assessment Using CORA-5 TEST
A4.4.1 Description of the Experimental Bundle
The CORA-5 experiment was a small bundle PWR test containing a single control rod centered in
the bundle. The CORA-5 bundle consisted of twenty five (25) rods in a 7 x 7 array with the comers
removed. The bundle, similar to the one shown in Figure A4-2, contained 8 unheated fuel rods, 16 heated
rods and a single stainless steel clad silver-indium-cadmium control rod centered in the bundle. The
unheated fuel rods used in this CORA experiment were prototypical KWU fuel rods. The argon flow rate
through the test bundle for this experiment was maintained 8.2 g/s. Steam flowed through the bundle at a
rate of 6.8 g/s starting 300 s after the initiation of power to the heater rods and continued until the end of
the experiment. While most CORA experiments had open viewing windows cut into the insulating shroud,
A4-3 NUREGICR-6150-Rev. 2, Vol. 5

Appendix A4
the viewing windows for the CORA-5 experiment were made of quartz, therefore eliminating the
possibility of gas or heat leakage through the bundle windows.
The results of the CORA-5 test were applied to the assessment of MOD3.3 by comparing calculated
temperature histories for several different elevations in the test fuel assembly with the measured
temperature histories at those elevations.
The MOD3.3 calculated temperature behavior of the test fuel assembly was in generally good
agreement with the measured temperature behavior. The MOD3.3 calculated and measured temperature
histories are shown in Figure A4-3 through Figure A4-7 for the elevations of 0.35 m, 0.55 in, 0.75 m, 0.85
in, and 0.95 in, respectively. The elevations are referenced with respect to the bottom of the fuel rods. The
MOD3.2 calculated temperature histories are also shown. At the 0.35 in elevation, the MOD3.3 calculated
heatup of an unheated rod was more rapid than the measured heatup. The heatup was calculated to be
accelerated by the slumping of hot material from above to a location just above the 0.35 m elevation and
then the conduction of heat in the axial direction to the location of the 0.35 in elevation. At the 0.55 in
elevation, the MOD3.3 calculated heatup was slightly faster the measured heatup. At the 0.75 in and 0.85
in elevations, the MOD3.3 calculated heatup was slightly slower than the measured heatup. At the 0.95 in
elevation, the measured temperature decreased due to contact with water from the condensation of steam
in the upper part of the test fuel assembly. The calculation did not represent the upper part of the test fuel
assembly and the condensation of steam, and as a result the calculated rate of heatup was more rapid than
the measured rate of heatup. The MOD3.3 calculated and measured temperature histories of a control rod
at the elevation of 0.55 in are compared in Figure A4-8. The calculated and measured temperature histories
are in good agreement.
3000.0
2000.0
2
E
09
1000.0
o0n I
3000.0 3500.0 4000.0 4500.0 5000.0
Time (s)
Figure A4-3. Comparison of calculated and measured temperatures of unheated fuel rod at 0.35 in
elevation for CORA-5 test.
The MOD3.3 calculated hydrogen production was 25% greater than that calculated by MOD3.2. The

Appendix A4
3000.0
2000.0
0
(D
a.
E
I-D
1000.0
0.0 1
3000.0 3500.0 4000.0 4500.0 5000.0
Time (s)
Figure A4-4. Comparison of calculated and measured temperatures of unheated fuel rod at 0.55 m
3000.0
2000.0
0
a.
0.
E
F- 1000.0
0.0
3000.0 4000.0 5000.0
Time (s)
Figure A4-5. Comparison of calculated and measured temperatures of unheated fuel at 0.75 m elevation
for CORA-5 test.

Appendix A4
3000.0
2000.0
2'
a)
0.
E
I- 1000.0
0.0 L
3000.0 3500.0 4000.0 4500.0 5000.0
Time (s)
Figure A4-6. Comparison of calculated and measured temperature of unheated fuel rod at 0.85 m
3000.0
2000.0
2
a)
a)
0.
E
aD
1000.0
0.0 I
3000.0 3500.0 4000.0 4500.0 5000.0
Time (s)
Figure A4-7. Comparison of calculated and measured temperature of unheated fuel rod at 0.95 m

Appendix A4
3000.0
2000.0
E
I- 1000.0
Measured
Calculated, MOD3.3
Calculated, MOD3.2
0.0 1 1
3000.0 3500.0 4000.0 4500.0 5000.0
Time (s)
Figure A4-8. Comparison of calculated and measured temperature of control rod at elevation of 0.55 m
MOD3.3 and MOD3.2 calculations of hydrogen production were 125 g and 100 g, respectively. A
measurement of hydrogen production was not obtained.
The elevation of the primary blockage caused by fuel rod meltdown was calculated to be lower than
the measured elevation. The calculated and measured elevations of primary blockage were 0.05 m and
0.45 m, respectively. MOD3.2 calculated the primary blockage to be located at the elevation of 0.20 m.
The discrepancy between calculated and measured elevations of blockage may be due to the calculated
behavior of the test fuel assembly not accounting for condensation occurring above the test fuel assembly
and the reflux caused by the condensation. MOD3.2 calculated a higher location of blockage than MOD3.3
because it does not have a model for calculating the re-slumping of cladding that has slumped and frozen.
In summary, this problem showed that the integral diffusion model in MOD3.3 correctly calculates
the heatup of fuel rods due to oxidation. The MOD3.3 calculated rate of fuel rod heatup were generally in
good agreement with the measured rate of heatup. MOD3.3 calculated about 25% more hydrogen
production than MOD3.2. MOD3.3 may overpredict the distance of slumping of melted fuel rod cladding.

Appendix A4
A4.5 Assessment Using the CORA-7 Test
A4.5.1 Description of the CORA-7 Experimental Bundle and Test
CORA-7 was the first of two large bundle experiments performed in the CORA test facility. The
utilization of the larger bundle resulted in smaller radial heat losses. The 56 rod CORA-7 bundle, Figure
A4-9, contained 32 electrically heated simulator rods, 19 fresh PWR-type fuel rods, and 5 typical PWR
silver-indium-cadmium control rods. The test was conducted in three stages. During the pre-conditioning
phase, the hot argon gas flowed through the bundle for 12 hours prior to establishing a thermal equilibrium
in the test assembly. The transient phase of the experiment was initiated by applying power to the heater
rods while the argon gas flow continued. Five minutes (300 s) after the initiation of power, superheated
steam was mixed with the flowing argon and power was linearly increased. Steam flow was terminated at
4,180 s. Power was applied to the bundle until the desired maximum temperature of 2,100 K was attained.
At 4195 s, power to the heater rods was then terminated and the bundle cooled gradually. Due to smaller
radial heat losses, the time needed to reach the maximum desired temperature was considerably less than
for the small bundle tests. Bundle temperatures were monitored and recorded by numerous thermocouples
attached to components in the test bundle, heated and unheated fuel rods, control rods, the insulating
shroud, and the high temperature shield. The damage progression in the bundle during the test was viewed
on video displays. Steam and argon mass flow rates, system pressure, fuel rod pressure and hydrogen
production were also measured during the test.
PWR Bundle
Legend
* Heater rods
Q Unheated rods
(a Control rods
M350-WHT-797-r7
Figure A4-9. CORA-7 large bundle cross-section.

Appendix A4
A4.5.2 Assessment Results
The MOD3.3 calculated temperature behavior of the test fuel assembly was in good agreement with its
measured temperature behavior. The MOD3.3 calculated and measured temperature histories of an unheated
fuel rod near the center of the test fuel assembly at several different elevations are shown in Figure A4-10
through Figure A4-13. The elevations are referenced with respect to the bottom of the fuel rods. The
MOD3.2 calculated temperature histories are also shown. In general, the temperature measurements failed
after the temperature exceeded about 2000 K. Double-sided oxidation was calculated by MOD3.3 to occur
in the elevation interval of 0.7 m to 1.0 m. The rate of heatup calculated by MOD3.3 at the 0.95 m elevation
was somewhat more rapid than the measured temperature. This difference may be due the calculations not
accounting for cooling affects at the top of the test fuel assembly. At the 0.75 m elevation, the MOD3.3
calculated rate of heatup is slightly slower than the measured rate of heatup. The rapid increase in
temperature at the 0.55 m elevation in the period of 4220 s to 4230 s was calculated by MOD3.3 to be partly
due to hot material from above slumping into that location and partly due to oxidation. During this 10 s
period, the cladding at this location was calculated to increase in temperature from 1520 K to 1970 K. The
maximum temperature of the test assembly was calculated to be 2650 K. Overall, the calculated and
measured rates of heatup of the test fuel assembly were in general agreement.
2000.0
1500.0
1000.0
0.
E
9
500.0
0.0 1
3000.0 4000.0 5000.0
Time (s)
Figure A4-10. Comparison of calculated and measured temperature at the 0.35 m elevation for CORA-7
test.

Appendix A4
3000.0
.- 2000.0
0.
E 100o
1000.0
0.0
3000.0 4000.0 5000.0
Time (s)
Figure A4-1 1. Comparison of calculated and measured temperature at 0.55 m elevation for CORA-7
test.
3000.0
2- 2000.0
CD
E
I
1000.0
0.0 1
3000.0 4000.0 5000.0
Time (s)
Figure A4-12. Comparison of calculated and measured temperature at 0.75 m elevation for CORA-7
test.

Appendix A4
3000.0 1 I I
Measured
- Calculated, MOD3.3
-------- Calculated, MOD3.2
22000.0
E
o Failure of measurement
1000.0
0.0
3000.0 4000.0 5000.0
Time (s)
Figure A4-13. Comparison of calculated and measured temperature at 0.95 m elevation for CORA-7
test.
The MOD3.3 calculated location of the blockage caused by meltdown of the fuel rods was in fairly
good agreement with the measured location of blockage. Both the PIE and the calculations indicated that
the meltdown and the blockage resulting from meltdown were limited in extenL The location of blockage
was determined from the PIE to be in the elevation interval of 0.35 m to 0.50 m. MOD3.3 calculated the
blockage to be in the elevation interval of 0.50 m to 0.60 m. In the elevation interval of 0.5 m to 0.6 m,
MOD3.3 calculated that 49% of the pre-blockage coolant space was filled with slumped material. MOD3.2
calculated the blockage to be in the elevation interval of 0.70 m to 0.91 m, which was significantly higher
than the PIE indicated location of blockage. The improvement in the MOD3.3 calculation with respect to
the MOD3.2 calculation is due to MOD3.3 having a model for the re-slumping of slumped fuel rod
material that froze and then re-melted, while MOD3.2 does not have such a model.
The hydrogen production calculated by MOD3.3 was in fairly good agreement with the measured
hydrogen production. The calculated cumulative production in hydrogen is shown in Figure A4-14. Most
of the hydrogen was calculated to be produced in the period of 3800 s to 4250 s. Intense oxidation was
calculated to occur for about 50 s after the termination of electrical heat generation in the test fuel
assembly. The total hydrogen production was calculated by MOD3.3 to be 80 g. MOD3.2 calculated a total
hydrogen production of 79 g. The total hydrogen production was measured to be 114 g. The measured
hydrogen production includes hydrogen produced by oxidation of the 0.4 m long extension of fuel rod and
simulator Zircaloy cladding above the heated length of the test fuel assembly. Thus, the hydrogen
produced in the heated portion of the test fuel assembly was somewhat less than 114 g. Since the
calculations represented only the heated portion of the test fuel assembly, the MOD3.3 and measured
hydrogen production are in fairly good agreement.

Appendix A4
0.12 I 1
Measured (0.114 kg)
S0.10
"-
C::
0.1Calculated, MOD3.3
.0
.)
S0.08
"0.
0
o•
0
0.06
® 0.04
:3
E Termination of electrial
0.02 heat generation
0.00
3000.0 4000.0 5000.0
Time (s)
Figure A4-14. Calculated cumulative hydrogen production for CORA-7 test.
In summary, the MOD3.3 calculated behavior of the test fuel assembly is in good agreement with the
measured behavior. The calculated heatup and meltdown of the test fuel assembly are in general agreement
with the measurements of these aspects of the behavior of the test fuel assembly. The calculation of
meltdown of the fuel assembly was improved by the addition to SCDP/RELAP5 of a model to calculate
the re-slumping of slumped fuel rod material that froze and then re-melted. The heatup of the lower part of
a fuel assembly was calculated to be accelerated by the slumping of hot material from above. After
accounting for hydrogen production from oxidation of the parts of the test fuel assembly above its heated
portion, the calculated and measured hydrogen production are estimated to be in fairly good agreement.
A4.6 Assessment Using the CORA-13 Test
In the CORA-13 experiment, a PWR fuel bundle was heated to the point of severe damage and then
was quenched. The experiment was used to define International Standard Problem 31. The configuration
of the CORA-13 bundle is shown in Figure A4-2. Viewing windows were cut in the insulating shroud at
three elevations, 390, 590, and 790 mm. The dimensions of these windows were 30 x 40 mm. The high
temperature shield was represented by a second SCDAP shroud component. The use of a second shroud to
represent the high temperature shield results in the calculation of radiation heat transfer from the outer
surface of the flow shroud to the inner surface of the high temperature shield. Beginning at 4,869 s, the test
bundle was reflooded by raising the quench tank at a rate of 0.01 m/s.
its measured temperature behavior. The MOD3.3 calculated and measured temperature histories of an
unheated fuel rod near the center of the test fuel assembly are shown in Figure A4-15 through Figure A4
17 for the elevations of 0.35 m. 0.55 m, and 0.75 mn,respectively. The elevations are referenced with
respect to the bottom of the fuel rods. The MOD3.2 calculated temperature histories are also shown. In

Appendix A4
general, the temperature measurements failed after the temperature exceeded about 2000 K. At the 0.35 m
elevation, MOD3.3 calculated a small increase in temperature at the beginning of reflood. At the 0.55 m
elevation, the MOD3.3 calculated temperature in the period of from 4100 s to 4300 s, when the
temperature increased from 1250 K to 2400 K, was influenced by contact with hot slumped material. The
unevenness in the measured curve during this period may also be due to the affect of slumping material. At
this elevation, MOD3.3 calculated the temperature to increase from 1750 K to 2500 K during the reflood
period. At the 0.75 m elevation, the MOD3.3 calculated temperature and the measured temperature are in
excellent agreement. Double-sided oxidation was calculated by MOD3.3 to occur at this location and at all
locations in the elevation interval of 0.7 m to 1.0 m. The maximum temperature of the test assembly was
calculated to be 2650 K. Overall, the calculated and measured rates of heatup of the test fuel assembly are
in general agreement.
3000.0
Measured Melting of claddin
Calculated, MOD3.3 M n l
------------.. Calculated, MOD3.2
2000.0
a)
E
Start of reflood
0.0 I
3000.0 4000.0 5000.0
Time (s)
Figure A4-15. Comparison of calculated and measured temperatures of unheated fuel rod at elevation of
cladding oxidized are tabulated in Table A4-1. The elevations in the table without any data are due to
calculated to occur in the elevation interval of 0.7 into 1.0 m. The PIE results showed almost 100%
almost 100% oxidation of the cladding in the interval of 0.1 m to 0.7 m. MOD3.2 underpredicted
significantly the extent of oxidation in the interval of 0.4 in to 0.8 m, while MOD3.3 overpredicted the
extent of oxidation in the elevation interval of 0.1 m to 0.4 in. The overprediction of the extent of oxidation
in this interval of elevation may be due to an incomplete modeling of the boundary conditions at the
bottom of the test fuel bundle and the cooling affect of these boundary conditions. Since cladding

Appendix A4
3000.0
2000.0
£
CL
E
a)
1000.0
0.0 --
3000.0 4000.0 5000.0
Time (s)
3000.0
2000.0
C.
E
a)
F- 1000.0
0.0 L
3000.0 4000.0 5000.0
Time (s)
Figure A4-17. Comparison of calculated and measured temperatures of unheated fuel rods at elevation
of 0.75 m for CORA-13 test.

Appendix A4
slumping and blockages are also shown in Table A4-1. The blockage is defined to be the fraction of pre
blockage coolant space filled with slumped material. The PIE indicated a significant blockage due to
meltdown in the interval of elevation from 0.4 m to 0.5 m and minor blockages in the elevation interval of
0.1 m to 0.3 m. The location and extent of blockage calculated by MOD3.3 is in good agreement with the
test results. The location of blockage calculated by MOD3.2 is significantly higher than the measured
location of blockage. The improvement by MOD3.3 in the calculation of the location of blockage is due to
the model added to MOD3.3 to calculate re-slumping of previously frozen slumped material.
CORA-13 test.
Fraction blockage due to meltdownl
Elevation Fraction cladding oxidized slumpin floca to mestorno
0.05 0.03 0.21 0.0 0.02/no 0.0/no 0.0/no

0.15 1.00 0.17 0.0 0.0/no 0.0/no 0.15/no
0.25 1.00 0.43 0.07 0.02/no 0.0/no 0.15/no
0.35 1.00 0.70 0.60 0.0/no 0.0/no 0.0/no
0.45 0.64 0.10 1.00 0.40/no 0.0/no 0.43/no
0.55 1.00 0.16 1.00 0.0/yes 0.0/no 0.0/yes
0.65 1.00 0.36 1.00 0.0/no 0.0/no 0.0/yes
0.75 0.30 0.20 1.00 0.0/yes 0.5/no 0.0/yes
0.85 0.30 0.43 < 1.0 0.0/yes 0.0/yes 0.0/yes
0.95 0.34 0.40 - 0.0/yes 0.0/yes 0.0/yes
The MOD3.3 calculated hydrogen production was in good agreement with the measured hydrogen
production. The measured cumulative hydrogen production from the start of the experiment to the start of
reflood was 126 g. The measured hydrogen production during the reflood period of the experiment was 84
g, with the total cumulative hydrogen production of 210 g. Some of the produced hydrogen came from the
oxidation of the cladding of the test rods that extended above the heated elevation interval of the
experiment. The PIE analysis showed no metallic material left in the first 0.3 m of the test fuel rod
cladding extending above heated elevation interval of the test fuel assembly.A.4-1If the cladding in this 0.3
m elevation did not slump away, then oxidation of this portion of the test fuel assembly produced 49 g of
hydrogen and the oxidation of the heated portion of the test fuel assembly produced 161 g. Assuming the
hydrogen production above the heated portion was of the test fuel assembly was divided between pre
reflood and reflood like the overall hydrogen production, then the amount of hydrogen production before
reflood in the heated portion of the test fuel assembly is estimated to be 97 g. MOD3.3 and MOD3.2
calculated the hydrogen production only in the heated portion of the test fuel assembly; the test rod
cladding above the heated elevation interval was not modeled. MOD3.3 calculated a hydrogen production
of 78 g for the pre-reflood period of the experiment and a total hydrogen production of 158 g. The

Appendix A4
cumulative hydrogen production calculated by MOD3.3 is shown in Figure A4-18. The MOD3.2 hydrogen
production is also shown in this figure. The MOD3.3 calculation of hydrogen production is in good
agreement with the test results, while the MOD3.2 calculation overpredicted somewhat the hydrogen
production in the pre-reflood period and underpredicted the hydrogen production during the reflood
period.
0.20
C
:3
0 0.15
0
.L
C:
7)
0 0.10
.E 0.05
E
"C-
0.00 L_
3000.0 4000.0 5000.0
Time (s)
Figure A4-18. Comparison of calculated and cumulative hydrogen production for CORA-13 tests.
In summary, the MOD3.3 calculated behavior of the test fuel assembly was in good agreement with
the measured behavior. The calculated transient temperature distribution in the test fuel assembly was in
good agreement with the measured transient temperature distribution. The calculated meltdown of the test
fuel assembly was in good agreement with the measured meltdown. Except for an overprediction of the
extent of oxidation of the cladding in the bottom portion of the test fuel assembly, the calculated extent of
oxidation was in good agreement with the measured extent of oxidation. The discrepancy in extent of
oxidation for the bottom portion of the test fuel assembly may have been due to an incomplete modeling of
the boundary conditions at the bottom of the test fuel assembly. The MOD3.3 calculated hydrogen
production before reflood and during reflood were in good agreement with the measured hydrogen
production. The MOD3.3 calculation of the behavior of the test fuel assembly was improved over the
MOD3.2 calculation in four areas; (1) location of blockages from meltdown of the test fuel assembly, (2)
axial distribution in the oxidation of the test fuel assembly, (3) hydrogen production during the reflood
period of the experiment and (4) temperature history in upper part of test fuel assembly where heatup may
be limited by steam strarvation. These improvements are due to the implementation into MOD3.3 of the
following models; (1) model for re-slumping of previously frozen slumped cladding, (2) model for
cracking of cladding during reflood, and (3) integral diffusion model for oxidation of cladding.
A4.6.1 References
A4-1. S. Hagen et al., "Results of SFD Experiment CORA-13 (OECD International Standard Problem
31)" KfK 5054, February 1993.
Appendix A4
A4.7 Assessment Using the CORA-1 7 Test
CORA-17 was one of five BWR configured heating and melting tests performed in the CORA test
facility at FzK and the only BWR configuration test bundle cooled by quenching the hot degraded test
bundle after termination of power to the heater rods. The test procedure for the BWR configured tests was
similar to that used for the PWR tests with the following exceptions. The rate of steam and argon flow
through the bundle during the test was considerably less than those for the PWR tests. Steam flow was
maintained at a rate of approximately 2.5 g/s and the argon flow was 8 g/s.
A4.7.1 Nodalization
The input model for the CORA-17 test used two SCDAP shroud components to model the test train,
one shroud component modeled the insulated flow shroud surrounding the test bundle, the other shroud
component modeled the high temperature shield. The RELAP5 nodalization of the BWR bundle, as shown
in Figure A4-19, used three RELAP5 flow channels. One flow channel contained the fuel rods, heated and
unheated, a second flow channel contained the BWR control blade and channel box, and the third flow
BWR
Nodalization
CORA-17
Bypass
channel Fuel rods
Blades/
box channel
Figure A4-19. RELAP5 nodalization of the CORA-17 experiment.

Appendix A4
channel represented the bypass. Each flow channel was modeled using a single RELAP5 pipe component
divided into 12 equal volumes. The lower ten volumes were connected with the heated zone of the test
bundle, the upper two volumes were connected with the volumes immediately above the top of the heated
simulator rods that contained molybdenum and copper electrodes. The CORA-17 BWR bundle was
modeled using five SCDAP components. Figure A4-20 shows a cross-section of the CORA-17 test bundle.
The bundle components were; Component 1 to represent the unheated fuel rods, Component 2 to represent
the simulator rods, Component 3 to represent the control blade and channel box, Component 4 to represent
the insulating shroud, and Component 5 to represent the high temperature shield.
• Shroud
Control blade/
Schannel box
2-Unheated rod
S• Heated rod
Figure A4-20. CORA-17 bundle cross-section.
The test was performed using a system pressure of 2.2 MPa (1.2 Bar overpressure). Argon flow was
maintained at 8 g/s, until reflood. As the quench tank moved upward in the bundle, argon flow was
terminated. Once the quench tank surrounded the test bundle, argon flow at a rate 8 g/s was resumed.
Steam at a rate of 2 g/s was mixed with the argon gas beginning at 100 s after the test. One hundred
seconds after the termination of electrical power to the heater rods, at 4,800 s, steam flow was stopped.
Power to the heater rods was ramped up over a period of 1,500 s to 22 KW and held at 22 KW for 300 s.
The maximum power was somewhat less than the 27 KW power needed in CORA-13 test to reach
maximum desired temperature. At 5,080 s, reflood of the test bundle began by raising the quench tank at a
rate of 0.01 m/s.
The results of the CORA-17 test were applied to the assessment of MOD3.3 by comparing calculated
temperature histories for several different elevations in the test fuel assembly with the measured
temperature histories at those elevations. The calculated and measured hydrogen production were also
compared.The results of the CORA-17 test were applied to the assessment of MOD3.3 by comparing
calculated temperature histories for several different elevations in the test fuel assembly with the measured
temperature histories at those elevations. The calculated and measured hydrogen production were also
compared.
NUREG/CR-6150-Rev. 2, Vol. 5 A4--18

Appendix A4
the measured temperature behavior. The MOD3.3 calculated and measured temperature histories of an
unheated fuel rod are shown in Figures A4-21 and A4-22 for the elevations of 0.55 m and 0.75 rn,
3000.0
2000.0
0
0
0U
E
I-
1000.0
0.0
3000.0 3500.0 4000.0 4500.0 5000.0
Time (s)
3000.0
2000.0
lie
E
(D
1000.0
0.0
3000.0 3500.0 4000.0 4500.0 5000.0
Time (s)
Figure A4-22. Comparison of calculated and measured temperatures of unheated rod at elevation of 0.75
m for CORA- 17 test.

Appendix A4
respectively. The elevations are referenced with respect to the bottom of the fuel rods. The MOD3.2
calculated temperature histories are also shown. The temperature measurement for the 0.55 m elevation
beyond the time of 4500 s has a high degree of uncertainty. The MOD3.3 calculated and measured
temperatures are in good agreement. The comparison of calculated and measured temperatures indicates
that MOD3.3 calculates more accurately than MOD3.2 the reduction in rate of heatup caused by cladding
melting and steam starvation of oxidation. The calculated and measured temperatures of the channel box
for the elevation of 0.55 m are compared in Figure A4-23. The calculated and measured temperatures of
the control blade for the elevation of 0.55 m are compared in Figure A4-24. The calculated and measured
temperatures of these components of the test fuel assembly are also in good agreement.
3000.0 1 1
Measured
Calculated, MOD3.3
------------ Calculated, MOD3.2
£ 2000.0
E• • Sl umpinga \
E Of material
I- 1000.0
0.0 I
3000.0 4000.0 5000.0
Time (s)
Figure A4-23. Comparison of calculated and measured temperatures of channel box at 0.55 m for
CORA-17 test.
3000.0
g2000.0 - Measured
(D Calculated, MOD3.3
". . .Calculated, MOD3.2 .'
•_1000.0.
0.0 ,clwn40 0. ..0...00.
Slumping of control blade material

4000.0 45000, 5000.
0
Time (s)
Figure A4-24. Comparison of calculated and measured temperatures of control blade at 0.55 rn for
CORA- 17 test.

Appendix A4
The calculated hydrogen production was about 9% less than the measured hydrogen production. The
calculated and measured hydrogen production were 132 g and 150 g, respectively.
In summary, this test problem confirms that the integral diffusion model in MOD3.3 correctly
calculates the heatup of fuel rods due to oxidation. The MOD3.3 calculated rates of heatup of the fuel rods
in the test fuel assembly were in good agreement with the measured rates of heatup. The calculated heatup
of the channel box and control blade were also in good agreement with the measured heatup. The
calculated hydrogen production was 9% less than the measured hydrogen production.

Appendix A5
A5. ASSESSMENT USING THE PHEBUS B9+ TEST

The PHEBUS B9+ experiment, the fifth in a series of six severe fuel damage experiments using fresh
PWR fuel rods, was performed by the Institut de Protection et de Surete Nucleaire at the PHEBUS test
facility, Cadarache, France, in January of 1 9 8 9 .A5-1 The purpose of this test was to study phenomena
occurring during a PWR severe core damage accident. The test specifically focused on the following areas;
(1) cladding oxidation, (2) the mechanical behavior of cladding which had an oxide layer of variable
thicknesses and contained molten zircaloy, (3) the dissolution of U0 2 and ZrO 2 by molten zircaloy, and
(4) core degradation and melt progression phenomena. This experiment was designated as International
Standard Problem 28 by OECD/CSNI. The experiment had well defined thermal-hydraulic and boundary
conditions and has been used to the assess the ICARE Severe Accident computer code for nearly all
aspects of core degradation. The phenomena that can be assessed and verified by the PHEBUS
experiments are (1) zircaloy oxidation and the resulting hydrogen production, (2) cladding mechanical
damage, (3) material interactions occurring between zircaloy and fuel pellets, zircaloy and inconel spacer
grids, fuel rod cladding and Ag-In-Cd absorber material, and (4) the relocation of molten material and
blockages formed in the core region.
The test bundle for the PIEBUS B9+ experiment consisted of 21 fresh U0 2 fuel rods in a 12.6 mm
square pitch matrix. The bundle configuration was a 5 x 5 array with the corners removed. The fuel rods
were held in position by two 38 mm in height inconel grid spacers located at the 0.138 and 0.661 m
elevations. Each fuel rod had a stack of fuel pellets that was 0.8 m in height. The fuel bundle was
surrounded by a multilayer insulating shroud consisting of a 0.6 mm thick zircaloy liner, a 21.25 mm thick
porous zirconia insulating layer, and a 1 mm thick layer of high density zirconia. The shrouded bundle was
inserted in an 8 mm thick stainless steel tube. The stainless steel shrouded bundle was then inserted in the
driver core of the test reactor. The driver core supplied nuclear power to heat the test bundle. Thermal
hydraulic conditions close to what would be expected in a commercial reactor core in a beyond design
based accident were replicated during the test. Numerous temperature, flow, pressure, and hydrogen
measurements were recorded during the test.
The PHEBUS B9+ test was conducted in three separate phases; (1) oxidizing, (2) heatup in pure
helium, and (3) a slow cool down. During the oxidizing phase, superheated steam flowed through the
bundle at a rate of 2.0 g/s. The steam flow continued until the zircaloy cladding at the mid-plane of the
bundle was completely oxidized. The second phase of the experiment was initiated when bundle
temperatures exceeded 1,800 K. The flow of steam was replaced with helium at a flow rate of 0.5 g/s. The
purpose of this phase of the experiment was to produce fuel dissolution by pure zircaloy under steam
starved conditions. The nuclear power was increased in steps over a period of 5,500 s until the desired
maximum core temperature of 2,750 K was reached. The third and final phase of the experiment was a
slow cool down. The cool down was accomplished through step reductions in the nuclear power as helium
continued to flow through the test bundle. The purpose of the slow cool down was to assure that the bundle
geometry at the maximum test temperature was not changed during the cooling phase.
A5.1 Bundle Nodalization
The PHEBUS B9+ test section was modeled using a single thermal-hydraulic channel to represent
the test facility. Two time-dependent volumes connected by time-dependent junctions to a single mixing
volume were used to control the steam and helium flow through the bundle. The subcooled water source
on the outside of the shroud was also modeled using RELAP5 volumes. The RELAP5 system nodalization
is shown in Figure A5-1. The test bundle was represented by four SCDAP components; (1) a central fuel

Appendix A5
rod, (2) a middle ring of fuel rods consisting of 8 fuel rods, (3) an outer ring of fuel rods consisting of 12
fresh fuel rods, and (4) the shroud. Figure A5-2, a bundle cross-section, shows the SCDAP representation
of the test bundle. The 0.8 m high active fuel zone containing the fuel rods and insulating shroud was
divided into 10 axial nodes each 0.08 m high. Six radial nodes were used at each axial node to represent
the fuel rods radially and 16 radial nodes were used at each axial node to represent the insulating shroud.
All geometric data used to model the test facility were taken from the test specification report.
Subcooled
water ---p
301
Figure A5-1. RELAP5 nodalization used for the PHEBUS B9+ experiment.

Appendix A5
Subcooled water
ýo0
0000 0
000 0
Shroud
*Central rod
O Middle rods
0 Outer rods
Figure A5-2. SCDAP representation of the fuel bundle.
A5.2 Boundary and Initial Conditions
The inlet fluid temperature was defined to be 528 K, and the system pressure was defined to be 1.2
MPa. The steam flow rate during the oxidizing phase of the experiment was 2.0 g/s and the helium flow
rate during the fuel dissolution stage of the experiment was 0.5 g/s. The helium flow replaced the steam
flow at 8370 s. A mass flow rate of 10 kg/s was defined for the subcooled water used to cool the outside of
the shroud. The thermal conductivity of the insulating shroud as supplied by the PHEBUS organizers was
input as a user defined material property. The calculations were started at 640 s, one second before the
initiation of nuclear power to the test bundle, and continued until fuel dissolution was predicted to occur.
its measured temperature for the upper part of the test fuel assembly and overpredicted the temperature for
the lower part of the test fuel assembly. The MOD3.3 calculated and measured temperature histories are
shown in Figure A5-3 through Figure A5-6 for the elevations of 0.2 m. 0.4 m, and 0.6 mn,and 0.7 m,
respectively. The elevations are referenced with respect to the bottom of the fuel rods. The MOD3.2
calculated temperature histories are also shown. Before the stop of steam flow and the start of helium flow

Appendix A5
at 8370 s, the calculated and measured temperatures are in good agreement at all elevations. After the stop
of steam flow, the calculated temperature overpredicts the measured temperature at the elevations of 0.2 m
and 0.4 m. Since MOD3.3 calculated a complete cessation of oxidation after 8380 s due to the absence of
steam in the test fuel assembly, this overprediction of the temperature increase is not due to the calculation
of oxidation after 8380 s. The nuclear heat generation and the oxidation heat generation in the test fuel
assembly as a function of time are shown in Figure A5-7. The overprediction in calculated temperature
may be due to some combination of the following factors; (1) retaining too much insulation capability in
the flow shroud after the input of helium into the test fuel assembly, (2) underprediction of heat loss by
conduction from the blockage area to the part of the flow shroud in contact with the blockage, (3)
underprediction of the cooling of the test fuel assembly by the helium flowing through the test fuel
assembly, and (4) incorrect power history after 8380 s. After 9370 s, the calculated temperature at the
elevation of 0.2 m and 0.4 is calculated to increase significantly due to the increase in nuclear heat
generation in the test fuel assembly.
3000.0
Measured Heatup due to further

Calculated, MOD3.3 increase in nuclear
-------------. Calculated, MOD3.2 heat generation
2000.0
h..
E -
'- 1000.0
Start of He flow
0.0
0.0 2500.0 5000.0 7500.0 10000.0
Time (s)
Figure A5-3. Comparison of calculated and measured cladding temperatures at 0.2 m elevation for
PHEBUS B9+ test.
cladding oxidized are tabulated in Table A5-1. The locations in the table without any entry are due to data
not being available for that location. Double-sided oxidation of the fuel rod cladding was calculated to
occur in the elevation interval of 0.56 m to 0.64 m. The PIE results showed almost 100% oxidation of the
cladding of the test rods in the elevation interval of 0.48 m to 0.76 m. MOD3.3 calculated almost 100%
oxidation of the cladding in this same interval of elevation. MOD3.3 overpredicted somewhat the extent of
oxidation of the bottom part of the test fuel assembly and MOD3.2 underpredicted the extent of oxidation
at the top of the test fuel assembly. The MOD3.3 overprediction of the extent of oxidation in the bottom
part of the test fuel assembly may be due to an incomplete modeling of the boundary conditions at the
bottom of the test fuel bundle and the cooling affect of these boundary conditions. Since cladding

Appendix A5
3000.0
Measured at 0.4 m Heatup due to further

Calculated, MOD3.3 increase in nuclear
- - ------------ Calculated, MOD3.2 heat generation -.
2000.0
0
2'
E
(D
1000.0
( -flow- H-
Startf
I Start of He flow
0.0
0.0 2500.0 5000.0 7500.0 10000.0
Time (s)
Figure A5-4. Comparison of calculated fuel temperatures at 0.4 m elevation for PBEBUS B9+ test.
3000.0
Measured at 0.6 m
Calculated, MOD3.3
Calculated, MOD3.2
-----.---.---..
2000.0
0
------------
02.
E
I-
1000.0
Start of He flow
0.0
0.0 2500.0 5000.0 7500.0 10000.0
Time (s)
Figure A5-5. Comparison of calculated and measured fuel temperatures at 0.6 m elevation for PHEBUS
B9+ test.

Appendix A5
3000.0
Measured at 0.7 m
Calculated, MOD3.3
Calculated, MOD3.2
-----..------.
2000.0
2'
E
(D
1000.0
Start of He flow
0.0
0.0 2500.0 5000.0 7500.0 10000.0
Time (s)
PHEBUS B9+ test.
30000.0
Nuclear heat generation

Oxidation heat generation
20000.0
._o
0
cc
( 10000.0
I
Stop of steam flow and
start of He flow
0.0 , I,
0.0 2000.0 4000.0 6000.0 8000.0 10000.0
Time (s)
Figure A5-7. MOD3.3 calculated nuclear and oxidation heat generation histories in test fuel assembly
for PHEBUS B9+ test.

Appendix A5
slumping and blockages are also shown in Table A5-1. The blockage is defined to be the fraction of pre
blockage coolant space filled with slumped material. The PIE indicated a significant blockage due to
meltdown in the interval of elevation from 0.2 m to 0.3 m and minor blockages in the elevation interval of
0.0 m to 0.2 m. The location and extent of blockage calculated by MOD3.3 is in good agreement with the
test results. After the stop of steam flow into the test fuel assembly, MOD3.3 calculated a significant
dissolution of the oxide layers on the fuel rod cladding in the elevation interval of 0.0 in to 0.40 m of the
test assembly, and as a result, slumping of the metallic part of the cladding in this interval of elevation after
it was heated to its melting temperature.
PIEBUS B9+ test.
Fraction blockage due to

Elev. Fraction cladding oxidized meltdown/slumping from location
(yes or no)
0.04 0.02 0.02 0.0 0.5/yes -/yes
0.12 0.07 0.06 0.03 0.0/yes -/yes
0.20 0.16 0.11 0.05 1.0/yes 1.0/yes
0.28 0.30 0.21 0.10 1.0/yes 1.0/yes 1.0/
0.36 0.49 0.31 0.25 0.0/yes 0.0/yes
0.44 0.70 0.86 0.40 0.0/no 0.0/yes
0.52 1.00 1.00 0.70 0.0/no 0.0/yes
0.60 1.00 1.00 1.00 0.0/no 0.0/yes
0.68 0.99 1.00 1.00 0.0/no 0.0/no
0.76 0.90 0.52 0.75 0.0/no 0.0/no
The MOD3.3 calculated hydrogen production was about 15% greater than the measured value. The
cumulative hydrogen productions calculated by MOD3.3 and MOD3.2 are shown in Figure A5-8. The
MOD3.3 and MOD3.2 cumulative hydrogen productions were 46 g and 39 g, respectively. The measured
hydrogen production was 39 g. MOD3.3 calculated that no production of hydrogen occurred after the flow
of steam into the test fuel assembly stopped. The overprediction of hydrogen production by MOD3.3 was
due to an overprediction of the extent of oxidation of the bottom part of the test fuel assembly.
In summary, the B9+ test problem tested the capability of MOD3.3 to calculate the behavior of fuel
assemblies for the condition of extreme steam starvation. For the upper half of the test fuel assembly, the
extent of oxidation of the fuel rod cladding was in good agreement with the measured extent. For the
bottom half of the test fuel assembly, MOD3.3 overpredicted by about a factor of two the extent of
oxidation of the fuel rod cladding. MOD3.3 correctly calculated that all oxidation in the test fuel assembly
stopped after the flow of steam was replaced by the flow of helium. The total hydrogen production
calculated by MOD3.3 was in fairly good agreement with the measured value. MOD3.3 correctly

Appendix A5
0.050 1
Calculated, MOD3.3
"-------------- Calculated, MOD3.2
0.040 - ------
.o0
0 -"" Measured
0.030
0
0.020
0 "
E 0.010
0.000
0.0 2500.0 5000.0 7500.0 10000.0
Time (s)
Figure A5-8. Calculated cumulative hydrogen production for PHEBUS B9+ test.
calculated the interval of elevation at which a blockage of almost 100% occurred due to liquefaction of
fuel. The code overpredicted the heatup of the lower part of the test assembly after the flow of steam was
stopped. This overprediction may have been due to a degradation to the insulation capability of the shroud
after the flow of helium began, but which was not taken into account in the calculations.
A5.4 Reference
A5-1. B. Adroguer et al., ComparisonReportfor the OECD/CSNI InternationalStandard Problem 28,

OECD NEA CSNI Report No. 28, December 1992.

Appendix A6
A6. ASSESSMENT USING THE PHEBUS FPTO TEST
A.6.1 Introduction
The PHEBUS FPT0 test problem assessed the models for early phase damage progression and the
modeling of the transition into late phase damage progression. The test problem is based on the PHEBUS
FPTO experiment,A6"1 which was conducted in the PHEBUS reactor in France. This experiment was one
of a few experiments that heated fuel to ceramic melting by nuclear heat generation under severe accident
conditions. The heat up and damage phase of the experiment occurred through a period of 18,000 s, which
is a long period of time compared with most other severe accident experiments.
A.6.2 Test Description
The test bundle for the PHEBUS FPTO experiment consisted of 20 typical PWR fuel rods and one
Ag-In-Cd control rod. The cross-section of the test bundle is shown in Figure A6- 1. The flow area of the
ZIRCALOYSPACER GRID
* 12.2/11.4 rnm
20 FUEL
RODSDO rO
INSTUET LEADS 0I 81.5/73 rnen
0 97/33 rrh
,I CONTR•OL
ROD
0 112/100 mm
Figure A6-1. Cross-section of bundle of test rods for PHEBUS FPTO experiment.
test bundle was 0.002556 m 2 . The rods were arranged as a 5 x 5 array with the comer rods removed. The
control rod was located in the center of bundle of fuel rods. The test fuel rods had a 1.0 m high stack of fuel
pellets. The fuel rods had a trace of burnup. The fuel rods had a fill gas of He at a pressure of 2.8 MPa at a
temperature of 293 K. The test bundle had two spacer grids composed of Zr. One spacer grid was 0.29 m
above the bottom of the fuel stack and the other was 0.81 m above the bottom of the fuel stack. The array
of rods were surrounded by an insulated shroud. The insulated shroud was composed of two layers of low
porosity ZrO2 . Unlike the shroud in the PBF SFD tests, the FPTO shroud surface that.faces the fuel rods did
not have a Zr liner.The test fuel rods were cooled by steam during the heat up and damage phase of the
experiment. The steam was at a pressure of 0.22 MPa. The temperature of the steam at the inlet varied from

Appendix A6
452 to 484 K. The inlet flow rate varied from 0.5 x 10-3 to 3 x 10-3 kg/s during the heat up and damage
phase of the experiment. A plot of the inlet flow rate history is shown in Figure A6-2.
0.0040 I ,
0.0030
S0.0020
o 0.0010
LT•
0.0000
-0.00100,
0.0 5000.0 10000.0 15000.0 20000.0
Time (s)
Figure A6-2. Inlet flow rate history of test bundle.
The insulated flow shroud was cooled on its outside surface by an annular coolant channel with a
constant flow rate of 10 kg/s. The pressure of the coolant was 8 MPa and the inlet temperature of the
2
coolant was 438 K. The flow area of the channel was 0.00166 m .
The bundle power was increased in steps during the heat up and damage phase of the experiment.
The bundle power was produced by nuclear heat generation. The bundle power increased from 0 to 50.7
kW over a period of 18,000 s. The bundle power history is shown in Figure A6-3. The bundle power was
rapidly decreased after reaching its maximum value of 50.7 kW.
60000.0
50000.0
i. 40000.0
0
. 30000.0
20
20000.0
Ca)
10000.0
0.0o.
0.0
Time (s)
Figure A6-3. Power history of test bundle.

Appendix A6
A.6.3 Nodalization of Test Fuel Assembly

The 21 test rods, insulated shroud and coolant channels were represented by the SCDAP/RELAP5
MOD3.2 model of the experiment. The input file for the model was constructed at ENEA in Italy by
G. Bandini.A- 2 A schematic of the nodalization of the test bundle, insulated shroud and flow channels is
shown in Figure A6-4. The test rods were divided into three components; (1) component number 1
1 control rod
(SCDAP component #2)
8 fuel rods
-/ -(SCDAP component #1) 12 fuel rods
F
1 (SCDAP component #3)
--- Insulated shroud

120 120 230
230 (SCDAP component #4)
.4 + -
' ,x A
I
- 4~ - Flow channel next to

outside surface of
-4 - shroud (RELAP5
200 control volume 200
1m Main bundle flow
A. -'-c
/.
"100-11) "volumes 100-01 to
Bypass flow channel in

e WV;1
., _(RELAP5 event of bundle blockage
control volumes
102-01 to 102-11)
. " Height of an
_0_ _ axial node
- 1. TTime-dependent junction 185
904• RELAP5 control volume number
----..... j.------ Time dependent junction 85
80
Figure A6-4. Schematic of nodalization of test rods, insulated shroud, and flow channels.
represented the eight inner fuel rods in the test bundle, (2) component number 2 represented the single
control rod at the center of the bundle, and (3) component number 3 represented the twelve fuel rods in the
outer part of the fuel bundle. The insulated shroud was represented by component number 4. The test
bundle and insulated shroud were divided into eleven axial nodes with the middle nine axial nodes having
a height of 0.1 m and the bottom and top axial nodes having a height of 0.05 m. The two fuel rod

Appendix A6
components had six radial nodes at each axial node; the fuel was represented by four radial nodes and the
cladding by two radial nodes. The shroud was represented by nineteen radial nodes at each axial node. The
coolant in the test bundle and the annular flow channel next to the outside surface of the test bundle were
represented by RELAP5 control volumes. Three parallel flow channels were modeled. The first flow
channel represented the flow of coolant past the test rods and next to the inside surface of the insulated
shroud. This flow channel was represented by the RELAP5 control volume 100 and its eleven subvolumes
which corresponded in location with the eleven axial nodes of the test rods and insulated shroud. The
second flow channel represented flow past a blockage in the bundle of test rods. This flow channel was
represented by the RELAP5 control volume 102 and its eleven subvolumes. The third flow channel
represented the annular flow channel next to the outside surface of the insulated shroud. This flow channel
was represented by RELAP5 control volume 200. Crossflow was modeled between the two flow channels
inside the insulated shroud. The conditions of coolant into the bundle was controlled by the RELAP5 time
dependent volume 80. The rate of flow into the bundle was controlled by RELAP5 time-dependent
junction 85. The conditions of the coolant next to the outside surface of the insulated shroud was controlled
by RELAP5 time-dependent volume 180 and the rate of flow of this coolant was controlled by RELAP5
time-dependent junction 185.
A.6.4 Comparison of Calculations and Measurements
The results of the PHEBUS FPTO test were applied to the assessment of MOD3.3 by comparing
calculated results with measured results. First, the calculated temperature histories for three different
elevations in the test fuel assembly are compared with the measured temperature histories at those
elevations. Second, the calculated meltdown and axial distribution in oxidation of the test fuel assembly
are described. Third, the calculated hydrogen production is compared with the measured hydrogen
production.
The MOD3.3 calculated temperatures of the test fuel assembly were in good agreement with the
measured temperatures for the first and latter parts of the test and overpredicted the rate of heatup during
the intermediate part of the test. The calculated and measured temperatures are compared in Figure A6-5,
Figure A6-6, and Figure A6-7 for the elevations of 0.7 m, 0.4 m, and 0.3 m, respectively. The MOD3.2
calculated temperatures are also shown for the elevations of 0.7 m and 0.4 m. As shown in Figure A6-5,
the calculated and measured temperatures are in good agreement for the period of 0.0 s to 11000 s, wherein
the fuel rod cladding temperatures increased from 645 K to 1500 K and rapid oxidation of the fuel rod
cladding had not yet begun, and for the period of 11750 s to the end time of the measurement (14350 s),
wherein oxidation was reduced due to meltdown. The rate of heatup in the period of 11000 s to 11750 s,
when a rapid heatup from 1500 K to 2200 K was calculated to occur, was overpredicted by MOD3.3. The
maximum calculated rate of heatup was 19 K/s. The maximum measured rate of heatup during this period
was about 9 K/s. At 11120 s, the cladding was calculated to melt and relocate in the circumferential
direction. This relocation of the cladding decreased its surface area in contact with steam and resulted in a
decrease in the rate of heatup. The same trend between calculated and measured temperatures was seen at
the 0.4 m and 0.3 m elevations. Double-sided oxidation was calculated to occur in the elevation interval of
0.45 m to 0.65 m. Since double-sided oxidation was not calculated to occur at the elevations of 0.4 m and
0.3 in, the overprediction in the rate of heatup at these locations is due to factors other than double-sided
oxidation. Three possibilities for the overprediction of temperature in the 11000 s to 11750 s period are;
(1) overprediction of the insulation capability of the shroud around the test fuel assembly, (2) prediction of
double-sided oxidation in the elevation interval of 0.45 m to 0.65 in, although the rate of heatup was also
overpredicted in elevation interval of 0.25 m to 0.45 m, where double-sided oxidation was calculated to
not occur, and (3) phase diagram for the FPTO fuel rod cladding different from that applied by MOD3.3.

Appendix A6
3000.0
- 2000.0
0.
E
S1000.0
0.0
0.0 5000.0 10000.0 15000.0 20000.0
Time (s)
PHEBUS FPTO.
3000.0
£ 2000.0
E.
E
I-
1000.0
0.0
0.0 5000.0 10000.0 15000.0 20000.0
Time (s)
PHEBUS FPTO test.

Appendix A6
3000.0 J. .....--
Measured at 0.3 m iý %
Calculated, MOD3.3
--------- Calculated, MOD3.2
2000.0
,a
0. Failure mof
E measurement
I-
1000.0
-------------
0.0
0 .0 5000.0 10000.0 15000.0 20000.0
Time (s)
PHEBUS FPTO test.
Metallic meltdown of the cladding in the test fuel assembly was calculated to occur at the locations
in the test fuel assembly with double-sided oxidation. These locations were in the elevation interval of 0.45
m to 0.65 m. The axial distribution in oxidation and meltdown in the test fuel assembly show some of the
basis for the calculated temperature behavior of the test fuel assembly.The calculated axial distribution in
oxidation and meltdown are shown in Table A6-1. Measurements of these aspects of behavior are for the
most part unavailable. One available result is the measured elevation interval with a 100% blockage. The
PIE indicated a 100% blockage in the elevation interval of 0.16 m to 0.23 m. The MOD3.3 calculated
elevation interval with a 100% blockage was 0.0 m to 0.45 m. The overprediction of the elevation interval
with 100% blockage may be due to MOD3.3 overpredicting the insulation capability of the flow shroud
and thus overpredicting the amount of ceramic melting that occurred in the test fuel assembly.
Table A6-1. MOD3.3 calculated axial distribution in oxidation and meltdown for PHEBUS FPTO test.
Elevation Fraction Slumping of Melting and

cladding Fraction blockage metallic cladding slumping of fuel at
cladding due to meltdown from location (yes location (yes or
(in) oxidized o o o
or no) no)
0.025 0.01 1.00 no no
0.10 0.89 1.00 no yes
0.20 0.86 1.00 no yes
0.30 0.84 1.00 no yes

Appendix. A6
Table A6-1. MOD3.3 calculated axial distribution in oxidation and meltdown for PHEBUS FPTO test.
Elevation Fraction Slumping of Melting and

cladding Fraction blockage metallic cladding slumping of fuel at
cladding due to meltdown from location (yes location (yes or
(in) " oxidized
or no) no)
0.40 0.86 1.00 no yes
0.50 0.54 0.0 yes no
0.60 0.56 0.0 yes no
0.70 0.76 0.0 no no
0.80 0.90 0.0 no no
0.90 0.89 0.0 no no
0.975 0.96 0.0 no no
MOD3.3 underpredicted by about 15% the amount of hydrogen produced by oxidation of the
metallic parts of the test fuel assembly. The calculated cumulative hydrogen production is shown in Figure
A6-8. The total calculated and measured hydrogen production were 77 g and 90 g, respectively. The
underprediction in hydrogen production was probably due to metallic meltdown being calculated to occur
where in reality it did not occur. The overprediction of metallic meltdown, in turn, was probably due to
double-sided oxidation being calculated to occur where in reality it did not occur. The overprediction of
the elevation interval with double-sided oxidation, in turn, was probably due to an overprediction of
cladding ballooning.
0.10
0.08
CL
.o_ 0.06
0
"-0
"o0
0.04
C.)
0
0.02
0.00 L.
0.0 5000.0 10000.0 15000.0 20000.0
Time (s)
Figure A6-8. Calculated cumulative hydrogen production for PHEBUS FPTO test.

Appendix A6
In summary, the PHEBUS FPTO test problem showed that MOD3.3 has the capability to calculate
the behavior of fuel assemblies through the beginning of late-phase damage progression, when blockages
occur due to the melting of fuel. The comparisons of MOD3.3 calculated versus measured temperatures
showed good agreement during the early and latter stages of fuel assembly heatup under severe accident
conditions. For heatup in the intermediate stage, where cladding temperatures range from 1500 K to 2200
K, the rate of heatup calculated by MOD3.3 was about a factor of two greater than the measured value.
Three possibilities for the overprediction of the rate of heatup during the intermediate stage are; (1)
overprediction of the insulation capability of the shroud around the test fuel assembly, (2) overprediction
of cladding ballooning, and (3) phase diagram for fuel rod cladding different from that applied by the
oxidation model in MOD3.3. The total hydrogen production for FPTO was underpredicted by about 25%.
This underprediction in hydrogen production was probably due to an overprediction of the metallic
meltdown of the test fuel assembly, which in turn was probably due to an overprediction of cladding
ballooning. MOD3.3 overpredicted the amount of fuel melting which occurred during the test. This
overprediction may have been due to an overprediction of the insulation capability of the flow shroud
surrounding the test fuel assembly.
A.6.5 References
A6-1. N. Hanniet et al., PHEBUS PF FPTO Test - Preliminary Report, DRS/SEA/LERES IP/94/211,
Institut de Protection et Surete Nucleaire, France, May 6, 1994.
A6-2. G. Bandini, "Sensitivity Calculations for FPT-0 Bundle with SCDAP/RELAP5/MOD3.1 Code,"
Presentationto the 4-th Meeting of the Bundle DegradationInterpretationCircle, ISPRA, June 28,
1995, SAWG 95-031/1, Joint Research Centre of European Commission, Safety Technology
Institute.

Appendix A7
A7. ASSESSMENT USING THE ACRR DF-4 TEST
The DF-4 experiment was designed to examine phenomena associated with the heatup, oxidation,
and meltdown of a BWR fuel assembly.A7-1, A7-2 The experiment was conducted in the Annular Core
Research Reactor (ACRR) at Sandia National Laboratories. The experimental bundle, designed to
represent a small section of a General Electric D-lattice core, consisted of fourteen 0.5-m-long fuel rods, a
Zircaloy channel box enclosing a representation of the tip region of a BWR control blade, and an insulated
shroud consisting of porous ZrO 2, which contained a fully dense, ceramic, ZrO 2 tube.
Seven Zircaloy-4-clad fuel rods were situated on either side of the channel box structure. The cross
section of the experiment bundle is shown in Figure A7- 1. The spacing between the fuel rods in the bundle,
the fuel rod cladding, and the channel box wall were prototypical of a BWR, but the fuel rod diameter was
not. The fuel rods used were pressurized water reactor (PWR)-type rods, somewhat smaller in diameter
than BWR rods, and were situated in a triangular 0.5-in. (12.7-mm) pitch to improve the effectiveness of
heating the channel box.
Containment boundary Zircaloy channel box
ZrO 2 fiber insulator

SU02 fuel
Fully dense ZrO2 tube
""Control blade
stainless steel
Presueb~zday
(stainless steel) "4~" • ~ ""P' Absorber-834 C powder
bo rBpw
Figure A7-1. Cross section of the DF-4 experiment bundle.
The channel box consisted of a Zircaloy-4 rectangular box that represented the fuel canister walls on
either side of the control element in an actual BWR. The offset-from-center control blade situated in the
channel box consisted of five stainless steel tubes containing boron carbide (B4 C) powder surrounded by a
stainless steel sheath. The blade was positioned in the channel box to represent a tip region of a reactor
control blade.
A high-temperature oxidation transient was initiated by injecting superheated steam into the bundle.
Fission product decay heat was simulated by fission heating of the 10% 23 5 U-enriched fuel rods using the
ACRR driver core to drive the heatup. Steam was fed to the two separate flow regions in the DF-4

Appendix A7
experiment bundle through the use of orifices. To duplicate the steam flow in a BWR core, 80% of the
steam was directed into the flow channel containing the fuel rods, and the remaining 20% was directed into
flow channel containing the B 4 C control blade. The steam flow rate and distribution were consistent with
those anticipated to occur during a station blackout in a large plant.
Test thermal behavior was monitored through the use of platinum/rhenium and tungsten/rhodium
thermocouples mounted at various location in the bundle. The hydrogen production rate and the quantity
produced, including hydrogen produced during the Zircaloy oxidation phase of the experiment, were
derived from temperature measurements of a CuO-H 2 bed through which the test section effluent flowed.
A.7.1 Nodalization of Test Fuel Assembly and Boundary Conditions
Figure A7-2 shows the SCDAP/RELAP5 nodalization diagram of the ACRR test train used in the
analysis of the DF-4 experiment. The RELAP5 thermal-hydraulic modeling used for the analysis
employed two parallel channels. Each channel contained five equal axial volumes in the 0.5-m test section.
In Figure A7-2, the outermost channel, represented by volume 20, modeled the region bounded by the
outer wall of the channel box and the stainless steel pressure boundary. This region contained the 14 fuel
rods and the ZrO 2 insulation, both porous and ceramic. The innermost channel, represented by volume 30,
modeled the Zircaloy channel box that contained a control blade. Pressure boundaries at the inlet and
outlet of the test section were modeled by the use of two time-dependent volumes (10 and 60). The steam
flow was modeled by the use of two time-dependent junctions (111 and 112).
The components in the test fuel assembly were represented using the following SCDAP component
models: (1) fuel rod model for the Zircaloy-clad 10% 2 3 5U-enriched U0 2 fuel rods, (2) BWR control blade
model for the B4 C stainless steel-clad control blade, (3) general structural component model for both the
porous and ceramic ZrO2 -insulated shroud surrounding the fuel rods and channel box and the Zircaloy
channel box surrounding the B4 C control blade. Each component represented in the bundle consisted of
five axial nodes in the 0.5-m active zone. Radial detail varied between the components, from two nodes for
both the channel box and control blade to nine for the insulating shroud.
The power history input into SCDAP/RELAP5 duplicated the experimental power history and the
experimentally measured axial power profile during the time frame considered important for a complete
analysis of the DF-4 experiment: 5000 to 8000 s. The power history input into the code contained the
following: (1) representation of the initial nuclear heating phase, the rapid increase in reactor power to 1
MW to initiate heating of the fuel rods, (2) equilibration phase, the decrease in reactor power to 0 -.4 MW
as soon as fuel rod temperatures attained 1000 K to allow the temperatures of the channel box walls and
stainless steel control blade to equilibrate with the fuel, (3) preoxidation phase, a stepwise boost in reactor
power to full, (4) nuclear phase when full reactor power was maintained for 400 s to simulate a boiloff
resulting in core meltdown.
The steam flow history used as input for the analysis duplicated the steam flow used during the
experiment to simulate the boiloff. The steam flow ratio in the flow channels was kept constant throughout
the experiment. During the initial stages of bundle heatup, the steam flow was half of that delivered during
the nuclear phase of the experiment.

Appendix A7
50
*1 Upper plenum
IIIIIE150
50 Spring
!
1140
!
40 !
e
o
j Spacer
Thermocouple !
It
elevations
(M) 130 120
0.495
0368
0.254 b 0
30 20 0.505-m
'j. test bundle
0.096 -
20% steam flow-:: 112 % steam flow

1V-8
10 Lower plenum
Figure A7-2. SCDAP/RELAP5 nodalization for the DF-4 experiment analysis.

Appendix A7
The results of the DF-4 test were applied to the assessment of MOD3.3 by comparing calculated
temperature histories and hydrogen production with the measured temperature histories and hydrogen
production. The calculated and measured hydrogen production were also compared. The MOD3.3
calculated temperature behavior of the test fuel assembly was in good agreement with the measured
temperature behavior. The MOD3.3 calculated and measured temperature histories of a fuel rod are shown
in Figure A7-3 and Figure A7-4 for the elevations of 0.254 m and 0.096 m, respectively. The elevations
3000.0 I 'I
Measured
Calculated, MOD3.3
£ 2000.0
(.
E
I- 1000.0 I
An I I * I
" 0oo.o 6000.0 7000.0 8000.0

Time (s)
Figure A7-3. Comparison of calculated and measured temperatures of fuel rod at elevation of 0.254 m
for DF-4 test.
3000.0
, 2000.0
(D
E
( 1000.0
6000.0 7000.0 8000.0

Time (s)
Figure A7-4. Comparison of calculated and measured temperatures of fuel rod at elevation of 0.096 m
for DF-4 test.
are referenced with respect to the bottom of the fuel rods. The elevation of 0.254 m is about midway

Appendix A7
between the bottom and top elevations of the fuel pellet stack. At both the 0.254 m and 0.096 m elevations,
the calculated and measured temperature are in good agreement.
The calculated hydrogen production was about 30% less than the measured hydrogen production.
The calculated and measured hydrogen production were 28.5 g and 40 g, respectively. The underprediction
of hydrogen production was due in part to the calculations not being performed out to the time of the
cooldown of the test fuel assembly.
In summary, this test problem confirms that the integral diffusion model in MOD3.3 correctly
calculates the heatup of fuel rods due to oxidation. The MOD3.3 calculated rates of heatup of the fuel rods
in the test fuel assembly were generally in good agreement with the measured rates of heatup. The
calculated hydrogen production was 30% less than the measured hydrogen production.
A7.3 References
A7-1. R.O. Gauntt, R. D. Gasser, and L. J. Ott, "The DF-4 Fuel Damage Experiment in the ACRR with a
BWR Control Blade and Channel Box," NUREG-/CR-4671, SAND86-1443, Sandia National
Laboratories, November 1989.
A7-2. J. K. Hohorst and C. M. Allison, "DF-4 Analysis Using SCDAP/RELAP5," Nuclear Technology,
Vol. 98, May 1992.

Appendix A8
A8. ASSESSMENT OF MODELS FOR FLOW LOSS AND HEAT

TRANSFER IN POROUS DEBRIS
The models for thermal hydraulic behavior in porous debris were assessed by comparisons of
SCDAP/RELAP5 calculated behavior of debris with that evaluated by measurements and benchmarked
models presented in the literature. In particular, the models for flow loss were assessed by comparing the
pressure drop calculated by SCDAP/RELAP5 with that calculated by benchmarked pressure drop models.
The models for heat transfer were assessed by comparing calculated and measured temperatures for the
case of reflood from the bottom of an initially hot debris bed. The implementation of the porous debris
thermal hydraulic models into the COUPLE model was assessed by evaluating the behavior calculated for
a flooded porous debris bed located in the lower head of a reactor vessel.
A.8.1 Flow Loss

The assessment problems for pressure drop involved the steady state analyses of the coolant
conditions in a porous debris bed with forced flow at the bottom boundary of the debris bed. A schematic
of the system analyzed is shown in Figure A8-1. The assessment was performed for the following coolant
44 "Fluid leavinc debris bed
0 Q O . Insulated, ri
gid surface
0 000 Debris withi porosity of 0.4

e 0O and particle diameter of 3 mm
\Z1
1\1 0 0 E)"
im 04 4
s'
• -•Defined fluid velocity S•
Figure A8-1. Schematic of debris bed analyzed for assessment of flow loss calculations.

Appendix A8
conditions; (1) superheated steam, (2) subcooled liquid, and (3) two-phase water. The third case analyzes a
debris bed representative of a debris bed resulting from a severe accident in a LWR.A8-1 The debris and
coolant conditions for the three cases are described in Table A8-1.
Table A8-1 . Characteristics of debris bed and coolant conditions for assessment of pressure drop
calculations.
Case
Parameter steam subcooled two-phase
porosity of debris 0.4 0.4 0.4
size ofparticles in debris bed (mm) 3.0 3.0 3.0
height of debris bed (m) 1.0 1.0 1.0
heat generation in debris bed (MW/m3 ) 0 0 7.5

coolantpressure (Mpa) 6.9 6.9 6.9
2 1.47x10-2
superficialvelocity offluid at bottom of 0.132 1.39x1]0
debris bed (m/s)
temperatureoffluid at bottom of debris 1050 400 558.
bed (K)
quality offluid at bottom of debris bed 1.0 0.0 0.0
density of heaviestphase offluid (kg! 15.2 940.7 740

m3)
hydrostatichead (Pa) 149 9.22x10 3 3.4x10C
For all three cases, two boundary conditions were defined. The first boundary condition was the
velocity of the fluid at the bottom of the debris bed. The second boundary condition was the temperature
and quality of the fluid at the bottom of the debris bed. The heat generation in the debris bed was defined to
be zero for the two cases involving single-phase coolant and was defined to be 7.5 MW/m 3 for the two
phase fluid case. The nodalization diagram for the analyses is shown in Figure A8-2. For each case, the
debris bed was divided into ten nodes and the fluid in the debris was represented by a stack of ten RELAP5
control volumes. The relative permeabilities and passabilities of the liquid and vapor phases were
calculated using the Catton and Chung model.A8-4 The calculation of the flow losses in porous debris was
assessed by comparing SCDAP/RELAP5 calculated pressure drops with independent calculations
published in the literatureASl, A8-2 and shown to agree with experimental results.
The assessment of the modeling of flow losses showed that the SCDAP/RELAP5 is correctly
calculating the flow losses in porous debris. The SCDAP/RELAP5 flow losses for the three cases
identified in Table A8-1 and those presented in the literature for the corresponding cases are compared in
Table A8-2. The SCDAP/RELAP5 calculated flow losses are in approximate agreement with the values
presented in the literature.

Appendix A8
200
(Time-dependent volume)
1-i120
(Z 100 - 10 ,, Debris node #10
0.1 m
C I0 "- D
0 C>o 0-Q
0C0 100 000
- 09°
0100- 0
Z0 O9 0
b 0C
eo o"
o 100-06 0
1.0 m 0C 0)0
0 100 - 05-0 dentifier number fc lr
._.____g 0 RELAP5 control vo lume
00~00
0 100 -04o0
0100- 03 0
O00000
0 100 - 02 0
0.1 m C
o 0u0-0u, O Debris node #1
0 0 0 (
"F 56 (time-dependent junction)
10
(Time-dependent volume)
Figure A8-2. Nodalization of debris bed analyzed for assessment of flow loss calculations.
Table A8-2. Comparisons of SCDAP/RELAP5 calculated flow losses with those presented in literature
for corresponding cases.
Case Pressure drop due to flow losses (Pa)
SCDAP/RELAP5 Literature
superheatedsteam 1.03x]03 1.14x103
subcooled liquid 1.OOxl 1.28x1O3

two-phase water 10.6x)& 12.6x103

Appendix A8
A8.2 Heat Transfer
The modeling of heat transfer in porous debris was assessed using the results of a BNL debris
83
experiment involving the quenching from the bottom of a hot porous debris bed.A " During this
experiment, the transient temperature distribution in the debris bed was measured. Since the transient
temperature distribution is a function of the flow losses in the debris bed, this experiment in an indirect
manner also assessed the modeling of flow losses. A schematic of the experiment is shown in Figure A8-3.
Except for representing the fluid in the debris bed by a stack of twelve instead of ten RELAP5 control
volumes, the nodalization for the experiment is as shown in Figure A8-2. Each RELAP5 control volume
represented a segment of the debris bed that was 35 mm in height. A summary description of the
experiment is shown in Table A8-3
0.108 m
Fixed scree e) Heated stainless steel tube
Stainless steel spherical particles;

eoj• • diameter = 3.175 mm
00 (porosity = 0.39
0.422 m • initial temperature = 775 K
Fixed screen 0e T
0V**-* Water;,
velocity = 4.42 mm/s
pressure= 0.1 MPa
temperature = 373 K
Figure A8-3. Schematic of BNL quenching experiment.
Table A8-3. Summary description of BNL quenching experiment.
Parameter Value
Porosity of debris 0.39
Composition ofparticles in debris bed stainless steel
Diameterof particlesin debris bed (nun) 3.175

Height of debris bed (m) 0.422

Appendix A8
Table A8-3. Summary description of BNL quenching experiment. (Continued)
Parameter Value
Diameterof debris bed (m) 0.108
initial temperatureof debris bed (K) 775
Coolant water
Temperature of coolant at bottom of debris bed (K) 373
Quality of coolant at bottom of debris bed 0.0
System pressure (MPa) 0.1
Superficial velocity of coolant at bottom of debris bed (mmn/s) 4.42
The calculated transient temperatures at two different elevations along the centerline of the debris
bed were in general agreement with the measured transient temperatures. The calculated and measured
transient temperatures for the two elevations are compared in Figure A8-4. The elevation of 0.025 m is
near the bottom of the debris bed, where reflood began, and the elevation of 0.24 m is slightly above the
midplane of the debris bed. The overprediction of the temperature at the 0.24 m elevation in the period
from 40 s to 50 s is considered to be due to MOD3.3 not modeling two-dimensional hydrodynamic
behavior, wherein the liquid phase moved up along the wall, formed a pool at the top, and then some of the
water flowed down the center region of the debris bed. Nevertheless, the calculated and measured trend in
quenching are in good agreement.
800.0
700.0
2 600.0
a. - Calcuated,0.02,1Im
E 500.0 - Measured,0.025 im
0! --- - Caculattd,0.24 m•.
-- - Measured,0.24 nl
400.0
300.0 '
0.0 50.0 100.0
"lime (s)
Figure A8-4. Comparison of calculated and measured transient temperature distribution in debris bed

Appendix A8
A8.3 Thermal Hydraulic Behavior in Lower Head

The implementation of the porous debris thermal hydraulic models into the COUPLE model was
assessed by the analysis of a porous debris bed in the lower head of a reactor vessel A schematic of the
system analyzed is shown in Figure A8-5. This figure also shows the COUPLE model nodalization of the
0.19 m-
[4-- 2.01 n - p
Lower head
of reactor
vessel
in debris bed
Figure A8-5. Nodalization of debris bed in lower head of reactor vessel.
porous debris. The porous debris had a total internal heat generation rate of 16.3 MW. The porosity of the
debris is 0.4. The debris particles were composed of U0 2. The particles were spherical in shape and had a
diameter of 3 mm. The initial temperature of the particles was 600 K. The debris particles slumped into the
lower head in the 10s interval of tume beginning at 11378 s, which was the start time of the analysis. The
RELAP5 nodalization for the analysis is shown in Figure A8-6. At the start of the analysis, the lower head
was empty of water. Beginning at the start of the analysis, the lower head was flooded with water at the
rate of 80.3 kg/s. The system pressure was 0.2 MPa.
An assessment was performed of the energy balance in the debris bed and the coolant that flooded
the debris bed. A plot is shown in Figure A8-7 of the transient volume of liquid in the RELAP5 control
volume containing the flooded debris bed. The volume fraction of liquid increased from 0.0 to 0.5 during
the first 200 s of flooding. The transient temperature at the center of the debris bed is shown in Figure A8
8. At this location, the debris cooled in about 200 s from 600 K to a temperature near the saturation
temperature of the coolant. The debris bed power and the heat transfer to the coolant. are compared in
Figure A8-9. After the initial internal energy was removed from the debris, the debris bed power and heat

Appendix A8
Junction 9 Volume 901-03

I_______
Volume 901-02
Volume 901 -01
Junction 50
(Time-dependent)
IVol. 10 j
source volume -N d
fnis of.oroude1
Figure A8-6. RELAP5 nodalization for analysis of porous debris in lower head
0.8
0.6
*0
0~
._o
:3
0.4
0E
0.2
0.0L 1
11400.0 11600.0 11800.0
Time (s)
Figure A8-7. Volume fraction of liquid in RELAP5 control volume containing the flooded debris bed.

Appendix A8
600.0
500.0
E
400.0
40.0 11500.0 11600.0

Time (s)
Figure A8-8. Transient temperature at center of debris bed.
300.0
- Debris bed power (w)

Heat transfer to coolant-----..
200.0
0
a
100.0
.
r'n
#L
11400.0 11500.0 11600.0

Time (s)
Figure A8-9. Comparison of debris bed power and heat transfer to coolant

Appendix A8
transfer to the coolant converge to the same value. This equality indicates that the COUPLE model for
porous debris has been properly interfaced with the RELAP5 model.
A8.4 Conclusions
The SCDAP/RELAP5/MOD3.3 calculations of flow losses and heat transfer in porous debris are in
general agreement with the calculations of benchmarked models and with experiment results. The
SCDAP/RELAP5 calculated flow losses for three different conditions of coolant were similar to those
presented in the literature for corresponding conditions. The SCDAP/RELAP5 calculated transient
temperature distribution in a hot debris reflooded from the bottom was similar to the measured transient
temperature distribution. The COUPLE model representation of heat transfer in porous debris flooded with
water results in a correct energy balance in the overall system analyzed. As a result of the models that have
been implemented, SCDAP/RELAP5/MOD3.3 calculates in a more thorough manner the heatup of porus
debris in the core region and in the lower head.
A8.5 References
A8-1. V. X. Tung, V. K. Dhir and D. Squarer, "Forced Flow Cooling Studies of Volumetrically Heated
Porous Layers," Second International Topical Meeting on Nuclear Reactor Thermal-Hydraulics;
Santa Barbara, California, USA, January 11-14, 1983.
A8-2. V. X. Tung, "Hydrodynamic and Thermal Aspects of Two-Phase Flow Through Porous Media,"
Ph. D. Thesis, University of California, Los Angeles, 1988.
A8-3. N. K. Tutu et al, "Debris Bed Quenching Under Bottom Flood Conditions (In-Vessel Degraded
Core Cooling Phenomenology)," NUREG/CR-3850, 1984.
A8-4. I. Catton and M. Chung, "Two-Phase Flow in Porous Media with Phase Change: Post-Dryout Heat
Transfer and Steam Injection," Nuclear Engineering and Design, Vol. 151, pp. 185 - 202.

Appendix A9
A9. ASSESSMENT OF MODEL FOR HEAT TRANSFER IN STRATIFIED

MOLTEN POOL
The model for heat transfer in a stratified molten pool was assessed by comparing its steady state
9
calculations with those of another modelAglthat was benchmarked against experiment results using
simulant materials. The model was also assessed by examining for internal consistency its calculation of
the transient analysis of a stratified molten pool and by examining the sensitivity of its calculations to
various combinations of heat transfer correlations. The lack of experimental data for LWR materials and
geometry excluded the assessment of the model by direct comparison of calculations with measurements.
A9.1 Comparison of Calculations With Benchmarks

A total of three test problems were solved to produce information for comparing the model
calculations against benchmarks and for evaluating the calculations for internal consistency. All three test
problems involved the analysis of a stratified molten pool in the lower head of a PWR reactor vessel
submerged in a deep pool of water. A schematic of the system analyzed is shown in Figure A9-1. Each of
Centerline of reactor vessel
Adiabatic surface
, • IJ"" it'-layer Molten metallic
8 , ,,reactor
"Lowerhead of
vessel
Molten oxidic pool
2. m_ .Water
1.6 ,,.
Frozen oxidic material
Figure A9-1. Molten pool system analyzed in literature and by SCDAP/RELAP5.

the test problems analyzed this basic configuration. The characteristics of the system that were the same in
each of the test problems are shown in Table A9-1. These characteristic included; (1) lower head of reactor
vessel composed of carbon steel, inner surface with radius of 2.0 in, and thickness of 0.158 m, (2) oxidic
pool composed of 86 wt% U0 2 and 14 wt% Z7O 2 , (3) metallic pool composed of stainless steel, and (4)
volumetric heat generation in oxidic pool of 1.4 MW/m 3 . The liquidus temperature of the oxidic pool was
3000 K and the liquidus temperature of the metallic pool was 1630 K. The liquidus temperatures were
fixed to these values to be consistent with the solutions for Test Problems 1 and 2 presented in Reference
A9-1. No heat transfer occurred from the top surface of the metallic pool. The depths of the metallic and
oxidic parts of the molten pool and the time for molten pool stratification varied with the test problems.
The values of these characteristics for each of the test problems is shown in Table A9-2. The depth of the
metallic pool was relatively large for Test Problems I and 3, and was relatively shallow for Test Problem

Appendix A9
2. The molten pool was stratified into oxidic and metallic parts from the start for Test Problems 1 and 2,
and became stratified at 2000 s for Test Problem 3. Thus, Test Problems 1 and 2 involve a steady state
solution and Test Problem 3 involves a transient solution.
Table A9-1. Characteristics of system analyzed that were same for each test problem.
Characteristic Units Value
Radius of inner surface of lower head m 2.00

Thickness of wall of lower head m 0.158
Volumetric heat generation rate in oxidic pool MW/m 3 1.40
Volumetric heat generation rate in metallic pool MW/m 3 0.0
Composition of oxidic pool wt% 86%UO 2 ,

14% ZrO 2
Composition of metallic pool wt% 100% Stain

less steel
Liquidus temperature of oxidic pool K 3000
Liquidus temperature of metallic pool and lower head K 1630
Table A9-2. Characteristics of molten pool that varied with test problems.
Test Test Test

Characteristic Units Problem Problem Problem
1 2 3
Depth of oxidic pool m 1.60 1.18 1.60

Depth of metallic pool m 0.80 0.22 0.80
Mass of U0 2 in oxidic pool kg 93.9x103 58.3x10 93.9x103
Mass of stainless steel in metallic pool kg 70.OxI03 19.4x103 70.OxI03
Time of stratification into oxidic and metallic s 0.0 0.0 2000.
parts
The SCDAP/RELAP5 steady state solutions for Test Problems 1 and 2 were compared to the
solutions in Reference A9-1. The most important results in the solutions were the sideward heat fluxes into
the lower head at the location of the metallic pool. For stratified molten pools, the heat flux into the lower
head is expected to be maximum at this location, and thus the margin to CHF on the external surface of the
lower head may be a minimum at this location. Other important results from a calculation of the system
behavior include; (1) heat flux to lower head at various locations in the oxidic pool, (2) temperature of the
metallic pool, (3) temperature of the oxidic pool, and (4) thickness of the crust of top of the oxidic pool.
The MOD3.3 steady state solutions for Test Problems 1 and 2 were similar to the solutions in
NURBG/CR-6150-Rev. 2, Vol. 5 A9-2

Appendix A9
Reference A9- 1. The SCDAP/RELAP5 calculated steady state behavior for Test Problem 1 is compared in
Table A9-3 with the behavior for this case presented in Reference A9-1. For the sideward heat flux, the
SCDAP/RELAP5 and Reference A9-1 values were 0.66 MW/m 2 and 0.55 MW/m 2, respectively. For the
heat flux in the oxidic pool at a polar angle of zero degrees (bottom center), the SCDAP/RELAP5 and
Reference A9-1 values were 0.06 MW/m 2 and 0.13 MW/m 2, respectively. For the heat flux in the oxidic
pool at the location of the interface of the oxidic and metallic pools (polar angle of 78 degrees), the
SCDAP/RELAP5 and Reference A9-1 values were 0.83 MW/m 2 and 0.77 MW/m 2, respectively. The
SCDAP/RELAP5 and Reference A9-1 calculations of the temperature of the metallic pool were 1691 K
and 1680 K, respectively. The SCDAP/RELAP5 and Reference A9-1 calculations of the temperature of
the oxidic pool were 3102 K and 3110 K, respectively. For the thickness of the crust at the interface of the
oxidic and metallic pools, the SCDAP/RELAP5 and Reference A9-1 values were 3.7 mm and 5.5 Him,
respectively.
The SCDAPIRELAP5 and Reference A9-1 calculations for Test Problem 2 both show a significantly
greater sideward heat flux in the metallic pool compared with Test Problem 1, where the depth of the
metallic pool was 0.80 m instead Of 0.22 m. The two calculations are compared in Table A9-4. The
SCDAPIRELAP5 and Reference A9-1 values for the sideward heat flux were 1.91 MW/rn 2 and 1.22 MW/
m 2 , respectively. The SCDAP/RELAP5 and Reference A9-1 calculations of the steady state oxidic and
metallic pool temperatures are similar-, the values for the oxidic pool are 3080 K and 3086 K, respectively,
and the values for the metallic pool are 1799 K and 1736 K, respectively.
The margin to CIF on the external surface of the lower head was calculated to not be significantly
greater for the shallow metallic pool (Test Problem 2) than for a deep metallic pool (Test Problem 1). This
result may in part be due to relatively more conduction of heat in the axial direction in the lower head after
transfer from the metallic pool for the case of the shallow metallic pool than the deep metallic pool. For
both cases, the minimum margin to CHF was calculated to be at the elevation of the interface of the oxidic
and metallic pools.
The margin to C-F is a function of the critical heat flux model applied for the external surface of the
lower head; SCDAP/RELAP5 used the Cheung model, A92 which is dependent on the depth of the pool of
water in which the lower head is submerged. Other factor influencing the calculation of CHF margin
include the modeling of the ablation of melted reactor vessel material and the modeling of heat transfer
across the interface of frozen oxidic material and the inner surface of the lower head.
Table A9-3. Comparison of SCDAP/RELAP5 and Reference A9-1 solutions for Test Problem 1 (deep
metallic pool).
Value
Variable SCDAP/
RELAPSReference 5
RELAP5
Sideward heat flux of metallic pool (MW/m2 ) 0.66 0.55
Heat flux in oxidic pool at bottom center (MW/m 2) 0.06 0.13
Heat flux in oxidic pool at interface with metallic pool (MW/m 2 ) 0.83 0.77

Appendix A9
Table A9-3. Comparison of SCDAP/RELAP5 and Reference A9-1 solutions for Test Problem 1 (deep
metallic pool). (Continued)
Value
Variable SCDAP/
RELAPSReference 5
RELAP5

Thickness of crust on top of oxidic pool (mm) 3.7 5.5
Minimum CHF margin on external surface of vessel (heat flux/critical 1.30
heat flux)
Table A9-4. Comparison of SCDAP/RELAP5 and Reference A9-2 solutions for Test Problem 2 (shallow
metallic pool).
Value
Variable SCDAP/
RELAP5
RELAP5 Reference 5
2
Sideward heat flux of metallic pool (MW/m ) 1.91 1.22
2
Heat flux in oxidic pool at bottom center (MW/m ) 0.04 0.14
2 0.55 0.50
Heat flux in oxidic pool at interface with metallic pool (MW/m )
Thickness of crust on top of oxidic pool (mm) 4.6 6.2
CHF margin on external surface of vessel at elevation of metallic pool 1.35
(heat flux/critical heat flux)
The stratification of a molten pool may result in a temporary decrease in the margin to CHF on the
external surface of the reactor vessel. This aspect of molten pool behavior is indicated by the SCDAP/
RELAP5 solution for Test Problem 3. In this test problem, the molten pool is well-mixed for the first 2000
s of the analysis. Then, the metallic material in the molten pool was defined to instantly separate from the
oxidic material and form a pool on top of the oxidic pool. The metallic pool was initially at the temperature
of the well-mixed pool, and then gradually cooled down to the temperature at which all the heat received
from the oxidic pool was transferred to the reactor vessel. The transient temperatures of the metallic and
oxidic parts of the molten pool are shown in Figure A9-2. At the instant of stratification of the pool (2000
s), the temperature of the metallic and oxidic materials was 2970 K. After stratification, the temperature of
the oxidic material increased to 3110 K in about 500 s and the temperature of the metallic material
decreased to 1690 K in about 500 s. After 2000 s, the molten pool was again in steady state condition. The
change in margin to CHF caused by stratification of the molten pool materials is shown in Figure A9-3. In
this figure, the ratio of the heat flux on the external surface of the reactor vessel to the critical heat flux is
plotted. The heat flux ratio is plotted for the location at which this ratio had its maximum value, which was

Appendix A9
the external surface at the same elevation as the interface of the oxidic and metallic pools. At the instant of
stratification, the heat flux ratio was calculated to be 0.26. After stratification, the heat flux ratio was
calculated to increase in 320 s to a maximum value of 0.33, and then gradually decrease to a steady state
value of 0.26.
4000.0
3000.0
c 2000.0
E
1000.0
0.0 1
0.0 1000.0 2000.0 3000.0 4000.0 5000.0
Time (s)
Figure A9-2. Temperature histories of oxidic and metallic parts of molten pool before and after
stratification.
0.5
0.4 F
.2 0.3
"1
LL
0
0.2
Time at which molten pool stratified

0.1
I
AAil
11000.0
0
2000.0 3000.0 4000.0 5000.0
Time (s)
Figure A9-3. Ratio of heat flux to critical heat flux on external surface of lower head before and after
stratification of molten pool.

Appendix A9
A study was also made of the sensitivity of calculated results to models for which two correlations
are available. One of these models is the correlation for the shape of the heat flux on a hemisphere
containing oxidic molten material with internal heat generation. The two available correlations are; (1)
9 3
mini-ACOPO correlationA9 1 and (2) UCLA correlation.A - Another model with two available
correlations is the model for sideward heat flux from a metallic molten pool. The two available
correlations are; (1) Churchill and Chu correlation,A9 "4 which contains the Prandfl number of the metallic
pool, and (2) modified Globe and Dropkin.A9 5 . Calculations were performed using various combinations
of these sets of correlations and for the molten pool configurations corresponding with Test Problem 1 in
Table A9-2, namely the case of the deep metallic pool. The results of the sensitivity study are presented in
Table A9-5. The results show that the calculated behavior of the molten pool is not sensitive to the
selection of correlations for flux shape and sideward heat transfer in the metallic pool.
Table A9-5. Sensitivity of calculated molten pool behavior to correlations for flux shape and sideward
heat transfer in metallic pool
Maximum
Sideward heat flux at Maximum
Flux shape Sideward heat heat flux of interface of ratio of heat
correlation for transfer correlation metallic pool oxidic pool flux to CHF
oxidic pool for metallic pool and wail on external
surface
(MW/m2 ) (MWnim2 )
mini-ACOPO Modified Globe and 0.66 0.83 1.30

Dropkin
UCLA Modified Globe and 0.70 0.88 1.30
Dropkin
mini-ACOPO Churchill-Chu 0.64 0.83 1.25
A9.2. Conclusions
The SCDAP/RELAP5 model for heat transfer in a stratified molten pool was assessed by; (1)
comparison of its calculations with the calculations of a steady state model presented in the literature, (2)
applying the model to the transient analysis of a stratified molten pool, and (3) examining the sensitivity of
the model calculations to various combinations of heat transfer correlations. The SCDAP/RELAP5 and
literature solutions of the steady state heat behavior of a stratified molten pool were generally in agreement
for important aspects of heat transfer behavior such as the sideward heat flux from the metallic pool and
the location of minimum margin to CHF on the external surface a hemispherical-shaped lower head
containing the molten material. Both SCDAP/RELAP5 and the literature solutions calculated the
minimum margin to CHF to occur at the elevation of the interface of the oxidic and metallic pools. Both
SCDAPIRELAP5 and the literature solutions calculated that the sideward heat flux in a shallow metallic
pool is about twice as large as that in a pool four times deeper. The SCDAP/RELAP5 calculations showed
that the margin to CHF was not significantly affected by the depth of the metallic pool. The transient
analysis of a molten pool showed that the occurrence of stratification after heatup of a previously well-
NUREGICR-6150-Rev. 2, Vol. 5 A9-6

Appendix A9
mixed pool results in a significant temporary decrease in the margin to CHF. In the case analyzed, the
margin to CHF decreased by 27% in the first 320 s after stratification, and then the margin to CHF
gradually returned to a value nearly the same as that before stratification. The sensitivity study on heat
transfer correlations showed that the calculated behavior of a stratified molten pool is not sensitive to the
flux shape applied to the heat transfer from the oxidic pool or to the correlation for sideward heat transfer
from the metallic pool.
A9.3. References
A9-1. T. G. Theofanous, C. Liu, S. Additon, S. Angelini, 0. Kymalainen, and T. Salmassi, "In-Vessel
Coolability and Retention of a Core Melt," Report DOE/ID-10460, Vol. 1, Chapter 5, July 1995.
A9-2. F. B. Cheung, K. H. Hadded, and Y. C. Liu, "A Scaling Law for the Local CHF on the External
Bottom Side of a Fully submerged Reactor Vessel, NUREG/CR-0157, Vol. 2, February 1997, pp.
253-277.
A9-3. F. J. Afsia and V. K. Dhir, "An Experimental Study of Natural Convection in a Volumetrically
Heated Spherical Pool Bounded on Top with a Rigid Wall," in In-Vessel Coolability and Retention
of a Core Melt, edited by Theofanous et al., Report DOE/ID-10460, Vol. 1, Appendix C, July 1995.
A9-4. Churchill, S. W. and H. S. Chu, Int. J., Heat Mass Transfer,18,1323 (1975).
A9-5. S. Globe and D. Dropkin, "Natural Convection Heat Transfer in Liquid Confined by Two
Horizontal Plates and Heated from Below," J. Heat Transfer, 81, 1959, pp. 24-28.

Appendix A 10
A10. ASSESSMENT OF MODEL FOR FLOW OF MELTED MATERIAL

THROUGH POROUS DEBRIS
A10.1 Testing By Application To PWR Severe Accident Situation
The model for movement of melted core plate material through porous debris was tested by applying
it to the analysis of a situation that may occur in a PWR during a severe accident. A schematic of the
situation is shown in Figure A10-1. In this situation, melted core plate material is slumping onto the top of
a porous debris bed in the lower head of the reactor vessel. This situation may occur after break up of jets
of molten core material slumping into a pool of water in the lower head, and then boil off of all the water in
the lower head by the broken up core material configured as a porous debris bed. The evaluation of the
integrity of the lower head in this situation requires the analysis of the movement of the melted core plate
material through the porous debris and the affect of this movement on the heatup of the lower head. If the
ratio of the heat flux to the critical heat flux on the external surface of the lower head is not significantly
increased by the movement of melted core plate material through the porous debris, then the structural
integrity of the lower head is not adversely affected by a melting core plate.Al-I
Centerline of
reactor vessel - Radiation
heat transfer
- Core plate
Melted core
plate material
- Top surface
of debris
Outer surface
of lower head
Porous debris bed composed

of (U, ZrO2 ) particles
Figure A-1 0-1. Schematic of system analyzed and nodalization of system.
In the assumed severe accident situation in a PWR, a 0.9 m deep porous debris bed is resting on the
inner surface of the lower head of the reactor vessel. Decay heat in the debris has boiled all water out of the
AIO-1 NUREG/CR-6150-Rev. 2, Vol. 5

Appendix A1O
vessel. As a result, the vessel is filled with steam at a pressure of 0.2 MPa. Heat radiating from the top
surface of the debris bed is melting the bottom of the stainless steel core plate located above the debris bed.
The melted core plate material is slumping to the top of the debris bed and then permeating into the debris
bed. The dashed lines in Figure A10-1 describe the finite elements and control volumes used in the
numerical solution to represent the debris, lower head, and the core plate material permeating the debris.
The intersections of the dashed lines are the nodes at which temperature was calculated. The bottom of the
debris bed interfaces with the inner surface of the lower head. The outer surface of the lower head is in
contact with a pool of water. The initial conditions and boundary conditions are summarized in Table A10
1. The initial temperature of the debris bed is 1727 K, which corresponds with the liquidus temperature of
the core plate. The entire lower head is assumed to have an initial temperature of 400 K. The debris has an
internal heat generation rate due to decay heat in the U0 2 of 1 MW/m 3 . The debris bed has a porosity of
0.4 and is composed of U0 2 particles with a diameter of 3 mm. The analysis starts at a time of 0.0 s.
Table Al 0-1. Initial conditions and boundary conditions.
Characteristic Value
Core plate material stainless steel

Mass of core plate (kg) 25,000
Debris bed material U0 2
Debris bed porosity 0.4

Maximum depth of debris bed (in) 0.9
Initial temperature of debris bed (K) 1727
3 1
Nuclear heat generation rate in debris bed (MW/m )
Diameter of particles in debris bed (mm) 3
Initial temperature of lower head of reactor vessel (K) 400
Thickness of lower head (in) 0.158
Composition of lower head carbon steel
Heat transfer coefficient at interface of debris and lower head (W/m2 • K) 500
The melted core plate material was calculated to permeate to within 0.05 m of the bottom of the
debris bed. Figure A10-2 is a plot of the axial distribution in debris bed saturation and temperature along
its centerline for the time of 1500 s, when 30% of the core plate was calculated to have melted and
slumped onto the debris bed. In general, the liquefied stainless steel was calculated to move downward
through the debris at a rate of about 2 mm/s. Due to the cold temperature of the debris near the lower head,
a crust of frozen core plate material formed about 0.05 m above the surface of the lower head. Above this
crust, an ever increasing region of the debris bed was calculated to become saturated with core plate
material. At 2000 s, when 50% of the core plate was calculated to have melted, a 0.4 in deep region along
the centerline of the debris bed was calculated to be saturated.
The temperature of the debris bed was calculated to continually increase withtime due to inadequate
cooling. The calculated transient temperature distribution in the debris bed is plotted in Figure AlO-3 for
two locations along the centerline of the debris bed. The first location is at the top of the debris bed and the
NUTREG/CR-6150-Rev. 2, Vol. 5 A10-2

Appendix AIO
0.6 1 2500.0
- .~ - Sturated
"region
0.4 2000.0 M
CD Temperature
Saturation
C
S0.2 Location --
of 1500fro
0 .
Z rzen crust - 10. E
CO)
Bottom of
debris bed
0.0 I -I ITr I 1000.0

0.0 02 0.4 0.6 0.8 1.0
Distance below top surface of debris bed (m)
Figure A10-2. Distribution of debris bed saturation and temperature after melting of 30% of core plate
(time of 1500 s).
3 000.01
3.0............ . . ........
2000.0
a.o
E
S1000.0
Surface of debris
bed
----......... 0.7 m below surface of debris bed
0.0 ! * , I I
0.0 500.0 1000.0 1500.0 2000.0 2500.0
Time (s)
Figure A10-3. Temperature history of debris bed along its centerline.

Appendix A10
second location is 0.7 m below the top of the debris bed. At the location 0.7 m below the top surface, the
temperature of the debris was calculated to increase at a rate of about 0.35 K/s. The temperature of the
location at the top surface was calculated to increase at a slower rate due to contact with the relatively cool
liquefied core plate material and due to convective and radiative cooling. At 2500 s, the temperatures of
these two locations were calculated to be 2115 K and 2540 K, respectively.
The permeation of core plate material into the porous debris bed did not significantly increase the
heat flux on the external surface of the lower head. Plots of the ratio of heat flux to critical heat flux (CHF)
at the bottom center of the lower head for the cases of core plate material permeating through the debris
bed and core plate material staying above the debris are compared in Figure A10-4. The CHF was
calculated using the Cheung model.Al° 2
0.40 Pr o c pI
Permeation of core plate materal

0.30 permeation.Noof core plate material
LL
0.20
0
--
(D 0.10
0.00
•I .I I -
-5.1u
500 ).0 1000.0 1500.0 2000.0 2500.0
Time (s)
Figure Al 0-4. Ratio of heat flux to CHF on external surface at bottom center of lower head.
Generally, the value of CHF is smallest at the bottom center. As shown in this figure, the heat flux
ratios for the two cases are nearly equal until 1000 s, after which the heat flux ratio becomes about 2%
greater for the case of core plate permeation. The somewhat higher heat flux for the case with permeation
is due to the increase in effective thermal conductivity caused by the presence of core plate material in the
debris. The calculated effective thermal conductivity for three levels of debris saturation are shown in
Table A 10-2. The effective thermal conductivity ranges from 2.5 W/m . K for a saturation value of zero to
9.4 W/m . K for fully saturated debris. This increase in thermal conductivity of the debris increases the rate
of heat transfer from the interior of the debris to the lower head supporting the debris. Nevertheless, the
heat flux is considerably less than the critical heat flux even for the case of permeation of core plate
NUREG/CR-6150-Rev. 2, Vol. 5 AI0-4

Appendix AlO
material into the debris bed.

Table Al 0-2. Effect of level of debris saturation on thermal conductivity of debris.
Level of debris saturation Effective thermal conductivity (W/m 2

- K)
0.0 2.5
0.20 4.7
1.0 9.4
A10.2 Testing By Evaluation of Internal Consistency and

Reasonability
While experimental results do not exist as a benchmark for a direct assessment of the model for
movement of melted core plate through porous debris, nevertheless an indirect assessment of the model
can be performed by evaluating for internal consistency and reasonableness the calculations of the model.
First, an evaluation of the capability of the model to conserve mass can be made by comparing the mass of
stainless steel in the porous debris at any time with the cumulative mass of melted core plate material that
slumped onto the top of the debris bed. Second, the implementation of the momentum equation for melted
material can be checked for reasonableness by comparing the calculated rate of permeation of melted core
plate material with an independent calculation of the rate of permeation. Third, an evaluation of the
capability of the model to conserve energy can be made by checking the affect of the permeating core plate
material on the temperature distribution in the porous debris.
Figure A10-2 has information for evaluating the capability of the melt movement model to conserve
mass. This figure describes the distribution of melted core plate material in the porous debris bed at 1500 s.
At the centerline of the lower head, a layer of debris with a thickness of 0.21 m is saturated with porous
debris. By projecting a similar thickness of saturated layer under the entire core plate, the mass of core
plate material is the saturated zone of the core plate is estimated to be 7300 kg. The mass of core plate
material in the unsaturated zone of the debris bed is estimated to be 500 kg. Thus, the total mass of core
plate material in the debris bed is estimated to be 7800 kg. The cumulative mass of melted core plate
material that slumped onto the top of the debris bed at 1500 s was calculated to be 7600 s. The approximate
agreement of the estimated mass in the debris bed with the cumulative mass slumping onto the top of the
debris bed indicates the melt movement model is conserving mass.
The implementation of the momentum equation into the melt movement model was checked for
reasonableness by comparing its calculation of the rate of permeation of melted material with an
independent analysis of the rate of permeation. A plot as a function of elevation of the velocity of melted
material as calculated by the melt movement model is shown in Figure A10-5. The velocity is plotted for
the time of 1500 s. The calculations were part of the same calculations as shown in Figure A10-2 through
Figure A10-4. The plotted velocity is equal to the calculated superficial velocity divided by the product of
the debris bed porosity times the debris bed saturation. The independent analysis was performed for a
debris bed with the same porosity as that analyzed by the melt movement model but with a smaller particle
diameter and somewhat different debris bed saturation.Al°- 3 For the calculations shown in Figure A10-5,
the debris bed porosity was 0.4, the diameter of the debris particles was 6 mm, and the average debris bed
saturation was 0.07. For the independent analysis, the porosity of the debris bed was 0.4, the diameter of
the debris particles was 2 num, and the debris bed saturation was 0.1. As shown in Figure A10-5, the
average velocity in the upper 0.6 m of the debris bed was about 7.5 mm/s. For a similar debris bed except

Appendix A10
1.0 I '
0.8 -Top
of debris bed
C
S0.6
0.4
0.2
0.0 2.0 4.0 6.0 8.0 10.0
Velocity (minis)
Figure A10-5. Axial distribution in velocity of melted material at 1500 s as calculated by melt
movement model.
for a smaller particle size, the independent analysis calculated a velocity of 2.4 mm/s for the rate of
permeation of the melted material. This difference in calculated velocities is consistent with the difference
expected for the difference in particle sizes in the debris beds analyzed by these two calculations. Thus, the
comparison of the velocity calculated by the melt movement model in SCDAP/RELAP5/MOD3.3 with the
independent analysis indicates the momentum equation for the melt movement model was correctly
implemented into the code.
The reasonableness of temperature histories in Figure A10-3 indicates that the energy equation for
the melt movement model was correctly implemented. The melting and slumping core plate material has a
temperature of 1700 K and thus is cooler than the debris. The core plate material permeating into the debris
is heated by the debris and thus has more influence on cooling the debris at the top surface than at the
elevation of 0.7 m below the top surface. At the 0.7 m elevation, the temperature increase corresponds with
almost an adiabatic heatup and the influence of the core plate material on the temperature of the debris is
slight. These trends in the temperature histories at the two elevations indicate a correct implementation of
the energy equation for the melt movement model into SCDAP/RELAP/MOD3.3 model.
A10.3 REFERENCES
A10-1. T. G. Theofanous, C. Liu, S. Addition and A. Angelini, 0. Kymalaien, and T. Salmani, "In-Vessel
Coolability and Retention of Core Melt," DOE/ID-10460, Vol. 1.
A10-2. F. B. Cheung, K. H. Hadded, and Y. C. Liu, "A Scaling Law for the Local CHF on the External
Bottom Side of a Fully Submerged Reactor Vessel," NUREG/CP-0157, Vol. 2, February 1997, pp.
253-277.
A10-3. L. J. Siefken, "SCDAP/RELAP5 Modeling of Movement of Melted Material Through Porous

Debris in the Lower Head," LNEEI/EXT-98-01178 Rev. 2, October 1999.

Appendix AI I
All. ASSESSMENT OF MODEL FOR MOLTEN FUEL-COOLANT

INTERACTION
The Fuel-Coolant Interaction (FCI) model was assessed using the results of two FCI experiments.
The first FCI experiment used for the assessment was one of the MAGICO-2000 series of experiments
performed at the University of California, Santa Barbara, California.AIl This experiment involved
slumping of a cloud of small hot particles into a column of water and measuring the transient void fraction
in the water and the transient position of the particles. The second FCI experiment used for assessment was
one of the FARO experiments performed by Magallon and Hohmann at the Safety Technology Institute,
Ispra, Italy.Al1-2 This experiment involved slumping a jet of hot material into a column of water and
measuring the transient pressure in the water and the fraction of break up of the jet of slumping material.
The applicability of the FCI Model to severe accident analysis was tested by calculating the affect of
FCI on reactor coolant system behavior for several different ambiences.
Al 1.1 Assessment with a MAGICO-2000 experiment
An experiment identified as Run Z1500/0.2 in the MAGICO-2000 series of experiments was used to
assess the capabilities of the FCI model to calculate the motion and heat transfer of a cloud of hot particles
made of ZrO2 .AI1-1 A schematic of the experiment is shown in Figure A11-1. The characteristics of Run
~;7IICloud of
hot particles
Heated container
600 00
000000
I
*00000
UOOS
Water at
atmospheric
pressure
2.0 m
T
0.8.M
0.6 ,m
--
Figure A11-1. Schematic of MAGICO - 2000 Test Arrangement.

Z1500/0.2 are shown in Table Al1-1. An important feature of this experiment is its accurate
characterization of the cloud of hot particles plunging into a column of water. This feature is the result of
the absence of break-up of the material interacting with water, which in turn is a result of using a cloud of
hot particles instead of a jet of molten material. The measurements obtained during the experiment
All-1 NUREG/CR-6150-Rev. 2, Vol. 5

Appendix A ll
included: (1) void fraction for a small period of time at one location in the test facility, (2) location of the
cloud of particles as a function of time, and (3) transient swell of water level.
Table Al 1-1. Characteristics of Run Z1500/0.2 of MAGICO-2000 Experiment.
Initial height of column of water (m) 0.8
Diameter of test vessel (m) 0.6
Distance of free fall of particles (distance from 1.2
initial position to initial surface of water) (m)
Initial pressure in test vessel (MPa) 0.1
Water subcooling (K) 0.0
Velocity of particles at surface of water (m/s) 4.8
Initial temperature of particles (K) 1648
Mass of particles (kg) 6.2
Pour equivalent diameter (m) 0.225
Pour duration (s) 0.33
Chemical composition of particles ZrO 2
SCDAP/RELAP5/MOD3.3 with the FCI model was used to calculate the particle-coolant interaction
observed in Run Z1500/0.2 of the MAGICO-2000 series of experiments and the calculated results were
found to be in fairly good agreement with the measured results. The calculated and measured location of
the front of the cloud of plunging particles are compared in Figure A 11-2. The good agreement of
calculations with the measurements indicates that the FCI model is correctly calculating the motion of the
particles in response to gravity and drag from contact with water. The calculated and measured liquid
volume fraction in the test vessel at an elevation of 0.65 m above the bottom of the test vessel are
compared in Figure A 11-3. The only available measurement of liquid volume fraction occurred at 0.45 s
(0.35 s after initial contact of the particles with water). At this time, the calculated liquid volume fraction is
significantly greater than the measured value. The calculated and measured water level swell as a function
of time are compared in Figure A 11-4. As shown in Figure A 11-4, the approximate agreement of the
calculations with the measurements indicates that the calculations for the macroscopic hydrodynamic
behavior of the coolant account for the basic phenomena.
The comparisons shown in Figure Al 1-3 and Figure A 11-4 indicate that the drag force on vapor may
be overpredicted for the situation of vapor production during fuel-coolant interaction. In Figure Al 1-3, the
calculated volume fraction of liquid at the 0.65 m elevation at the time of 0.45 s is greater than the
measured value. This discrepancy may be due to vapor produced below this elevation being calculated to
flow too slowly upward due to an overprediction of the drag force on the vapor. In Figure A 11-4, the
calculated swell of water is greater than the measured swell after 0.4 s. This discrepancy may be due to an
underprediction of the rate at which vapor leaves the top of the water column, and in turn due to an
overprediction of the drag force on the vapor.
NUREG/CR-6150-Rev. 2, Vol. 5 All-2

Appendix AI I
100.0
80.0
E
E
02
.0 60.0
M
-c
E
0
40.0
(D
0
20.0
0.0 ' I1 1
0.00 0.20 0.40

Time (s)
Figure Al 1-2. Comparison of calculated and measured advancement of the front of cloud of particles for
MAGICO-2000 experiment.
1.0
0.8
0
"6 0.6
0
0J
0.2
0.0 L.
0.0 0.5 1.0
Time (s)
Figure Al 1-3. Comparison of calculated and measured volume fraction of liquid at elevation of 0.65 m
for MAGICO-2000 experiment.
Al11-3 NUREG/CR-6150-Rev. 2, Vol. 5

Appendix All
110.0
100.0
Q
90.0
80.0 F Measured (upper bound) N

--- Measured (lower bound)
0 Calculated
1I I
70.0
0.0 0.2 0.4 0.6 0.8
Time (s)
Figure A11-4. Comparison of calculated and measured water level swell for MAGICO-2000
experiment.
All .2 Assessment with a FARO Experiment
An experiment identified as Test L-08 in the FARO series of experiments was used to assess the
capabilities of the FCI model to calculate the break-up of a jet of molten material that plunges into a
column of water and to calculate the heat transfer between the material and the water with which it
interacts.Al 1-2 The characteristics of Test L-08 are shown in Table A 11-2. A schematic of the experiment
is shown in Figure A11-5. An important feature of this experiment is the measurement of the fraction of
the jet of molten material that broke up and the size of the particles resulting from break-up. The
measurements obtained during the experiment that are useful for comparison with calculations include; (1)
transient pressure in the test vessel, (2) fraction of break-up of the jet of molten material, (3) size of
particles resulting from break-up, and (4) transient temperature of the plate onto which the slumping
material settled.
Table Al 1-2. Characteristics of Test L-08 of FARO Experiments
Initial height of column of water (m) 1.0
Diameter of test vessel (m) 0.71

Diameter of jet nozzle (m) 0.1
Distance of free fall of jet (distance from jet noz- 1.53

zle to initial surface of water) (m)
Initial pressure in test vessel(MPa) 5.8
NUREG/CR-6150-Rev. 2, Vol. 5 AlI1-4

Appendix AI I
Table Al 1-2. Characteristics of Test L-08 of FARO Experiments (Continued)
Volume of gas (H 20, Ar) above initial surface of 0.875
water (m 3 )
Initial temperature of column of water (K) 536
Water subcooling (K) 10
Velocity of jet at surface of water (m/s) -10
Initial temperature of material in jet (K) 3023
Mass of material that slumped (kg) 44
Composition of material that slumped (wt %) 80% U0 2 , 20% ZrO 2
Pour duration (s) 0.37
Molten material
Nozzle with
hinged-flap
for release
0.1 m
0.71 m
2.53 m
Water
1.0 - -
Debris catcher
Bottom plate
Figure A11-5. Schematic of FARO Test Arrangement.
SCDAP/RELAP5/MOD3.3 with the FCI model was used to calculate the fuel-coolant interaction
occurring in Test L-08 of the FARO series of experiments and the calculated results were found to be in
fairly good agreement with the measured results. The calculated and measured pressure in the test vessel
are compared in Figure Al 1-6. The calculated and measured pressure histories were similar in shape but
the calculated rate of increase in pressure was about 50% more rapid than the measured rate of increase in

Appendix A 1l
8.0
7.5 000 Oo
0 00 0 0 0 0 0 0 0 0
. 7.0 0
o_
2 00
IT6.50
Cn0
U,o
CL 6.00
0 Measured
5.5 - Calculated
5.0 1
0.0 1.0 2.0 3.0 4.0 5.0
Time (s)
Figure Al1-6. Comparison of calculated and measured pressure histories for FARO Test L-08.
pressure. The maximum calculated increase in pressure was 112% of the measured maximum increase in
pressure. No other transient measurements were available for comparing with calculations. A summary of
the experimental data and the corresponding calculated values is shown in Table A 11-3. A break-up of
70% of the jet of slumping material was calculated, which is slightly greater than the 68% break-up that
was measured. The mean size of the broken-up material based on weight was calculated to be 0.56 mim,
which is significantly less than the measured mean size of 3.8 umm. The plate at the bottom of the test
vessel was heated by material that settled on the plate. The calculated and measured maximum increases in
temperature of this plate were 352 K and 275 K, respectively. The level swell of water is a measure of the
rate of steam production and of the spatial distribution of steam production. The calculated and measured
level swells of water were 0.25 m and 0.15 m, respectively.
Table Al 1-3. Summary of comparison of measured and calculated values for Test L-08 of FARO.
Variable Measured Calculated

Fraction of break-up of slumping material 0.68 0.70
Mean size of broken up material based on weight (mm) 3.8 0.56
Maximum pressure increase due to FCI (MPa) 1.70 1.90
Maximum temperature increase of bottom plate (K) 275 352
Level swell of water (m) 0.15 0.25
The greatest discrepancy between calculated and measured values for Test L-08 of the FARO
experiments was in the mean size of particles resulting from break-up of the slumping material. The
calculated size was significantly smaller than the measured size. This discrepancy was a factor in the
NUREG/CR-6150-Rev. 2, Vol. 5 Al11-6

Appendix All
calculated rate of increase in pressure being about 50% more rapid than the measured rate of increase in
pressure. The discrepancy between calculated and measured particle size may be due to one or more
reasons. One reason may be the use of an incorrect value for the interfacial surface tension of a mixture of
U0 2 and ZrO2 that is in contact with liquid water and steam. The calculated mean particles size is
proportional to the value of this material property. A value of 0.45 N/m was used in the calculations, which
is the most appropriate value available from MATPRO. This value is applicable for a mixture of Zr-U-O
that is in contact with Zr cladding. This value has a large expected standard error; its upper bound value is
estimated to be 1.45 N/re. The use of the upper bound value in the calculations would result in a significant
increase in the calculated mean size of the broken-up material.
Al 1.3 Application of FCI Model to Severe Accident Analysis
This section evaluates the applicability of the FCI model to severe accident analysis. Five cases of a
test problem are calculated that involve calculations of FCI under conditions that are possible in the event
of a severe accident in a PWR. For each case, a core meltdown in a PWR is assumed to have occurred that
results in molten material slumping into a pool of water in the lower head of the reactor vessel. A
schematic of the test problem is shown in Figure Al 1-7.
Flow to containment
I--""" " "--IUO2and Mixture of molten

ZrO2
Plate with holes

4.0m .Y'
•• 1 Jets of slumping
material
-I Particles resulting from

S,,break up of slumping
"110% 1 material
0
2.0m 1. o0 :
Settled particles
66
II *
0*
FS i F m
0 ~~2.0 m--.
Figure Al11-7. Schematic of application of FCI model to severe accident analysis.

Appendix All
Table A 11-4 identifies the slumping parameters for each of the five cases that were analyzed.
Table All -4. Characteristics of typical severe accident conditions with possibility for FCI.
Case Case Case Case Case

Characteristic1 1 2 334 4 5
Initial height of column of water (m) 2.0 2.0 2.0 2.0 2.0
Diameter of vessel (m) 2.0 2.0 2.0 2.0 2.0
Diameter of jets (mm) 137. 27.5 27.5 27.5 27.5
Number of jets 1 25 1 2 4
Velocity of jet (m/s) 5.0 5.0 6.25 6.25 6.25
Initial pressure in vessel (MPa) 0.17 0.17 0.17 0.17 0.17
Volume of steam above initial surface of water 50 50 50 50 50

(M3)
Initial temperature of pool of water (K) 388.3 388.3 388.3 388.3 388.3
Water subcooling (K) 0.0 0.0 0.0 0.0 0.0
Initial temperature of material in jet (K) 3023. 3023. 3023. 3023. 3023.
Total mass flow rate of slumping material (kg/s) 600 600 30 60 120
Wt% U0 2 in slumping material 66.7 66.7 66.7 66.7 66.7
Wt% ZrO 2 in slumping material 33.3 33.3 33.3 33.3 33.3
Duration of slumping (s) 10 10 10 10 10
In each of the cases, the system pressure at the start of slumping was 0.17 MPa and the water had no
subcooling. Also, the initial height of the column of water, the diameter of the reactor vessel, and the initial
volume of steam above the water level were the same for all of the cases. The initial volume of steam
above the water level was assumed to be 50 m 3, which is a volume that is of the same order of magnitude
as the volume of a typical PWR reactor vessel. The number of jets varied from 1 to 25 for the five cases
and the diameter of the jets was 27.5 mm for four of the cases and 137.3 mm for one case. The dividing of
a single large jet into several small jets is possible when a single jet slumps onto the core plate and then the
material in the jet spreads out over the plate and flows through holes in the core plate. The diameter and
number of jets were adjusted so as to maintain a jet velocity at the time of initial contact with water that
was in the range of 5.0 m/s to 6.25 mis. The total mass flow rate of the jets varied from 30 kg/s to 600 kg/
s. The mass flow rate of 600 kg/s is representative of the mass flow rate of the slumping core material in
the TMI-2 accident.A 11-3 In all cases, the slumping material was at a temperature of 3023 K and was
composed of 66.7 wt% U0 2 and 33.3 wt% ZrO2 . Each case may have about an equal probability of
occurring. For each case, the material was assumed to slump at a constant rate for 10 s. For each case,
NUREG/CR-6150-Rev. 2, Vol. 5 Al11-8

Appendix A 11
calculations were performed of the behavior of the slumping material and of the behavior of the water and
steam in the reactor vessel. Calculation results are presented that provide a basis for evaluating the internal
consistency and realism of the model and that show the sensitivity of FCI to parameters such as the rate of
slumping and the number and size of the jets of slumping material.
SCDAP/RELAP5/MOD3.3 with the FCI model was used to calculate the behavior of the slumping
material for the five cases, and the results showed that the extent of break-up of the slumping material is a
strong function of the rate of slumping of material. The results of the calculations for the five cases are
summarized in Table Al 1-5. Time in this table refers to time since start of slumping. The fraction of break
Table Al 1-5. Summary of results for FCI under conditions typical for severe accident in PWR.
Variable describing FCI Case Case Case Case Case

1 2 3 4 5
Fraction of break-up of slumping material 0.15 0.28 1.0 0.95 0.74
Pressure increase after 10 s of slumping (MPa) 0.10 0.17 0.01 0.08 0.11
Swell of water level (in) 0.65 2.00 0.15 0.35 0.85
up varied from 1.0 for the case of a single small diameter jet with a mass flow rate of 30 kg/s to 0.15 for the
case of a single large diameter jet with a mass flow rate of 600 kg/s. The increase in pressure caused by
FCI was a function of the number of jets of slumping material. For the cases with small jets (diameter of
27.5 mm). the maximum increase in pressure for the case of four jets was approximately nine times greater
than the maximum increase in pressure for the corresponding case of a single jet.
Plots of the calculated pressure in the vessel provide additional insights into the performance of the
FCI model. The calculated transient pressures in the vessel for the first two cases (600 kg/s rate of
slumping) are compared in Figure A 11-8. This figure shows that slumping material which divides into
several jets will produce a significantly more rapid rate of steam production and thus a more rapid rate of
increase in pressure than slumping material that remains a single jet. The dividing of a single large jet into
several small jets is possible when a single large jet slumps onto the core plate and then the material in the
jet spreads out over the plate and flows through holes in the core plate. For both the single jet and 25 jet
cases in Figure Al 1-8, the pressure history indicates that the calculated FCI was most vigorous during the
first 4 s of slumping. The calculated transient pressures for the last three cases are compared in Figure
A11-9. For these three cases, the mass flow rate per jet was 30 kg/s and the number of jets varied from a
single jet for Case 3 to four jets for Case 5. The pressure increase for the four jet case was about double the
pressure for the single jet case.
The calculated behavior of the water is indicated by plots of the calculated volume fraction of liquid
water as a function of elevation. The transient volume fractions of liquid for a range of elevations above
and below the initial elevation of the top of the column of water are plotted in Figure A11-10 for Case 1
(single large jet). This figure shows that a swell of 0.5 in occurred in the column of water. Figure A11-11
shows the calculated transient volume fractions of liquid for Case 2 (25 small jets) for elevations 1.0 m
above and below the initial elevation of the top of the column of water. This figure shows that a significant
amount of water was pushed upwards a distance of 1.0 m for about 8 s.

Appendix A 1l
.40
.30
co
0(
n3
PD
(0
.20
.10 L
0.0 2.0 4.0 6.0 8.0 10.0
Time (s)
Figure All-8. C alculated transients pressures in vessel for Case 1 (1 large jet) and Case 2 (25 small
jets).
.30
.25
20
.15
0.0 2.0 4.0 6.0 8.0 10.0
Time (s)
Figure Al 1-9. Calculated transient pressure in vessel for Case 3 (1 small jet), Case 4 (2 small jets) and
Case 5 (4 small jets).

Appendix All
1.0 •
\ i
- 0.5 m above initial liquid level

......... 1.0 m below initial liquid level
0.8
/ '/ .'
0-'
S-i
• •
S
/
i
/
A
.o_
Cu
0
0.6
\/
E
0
0.4
0.2
0.0 1
0.0 2.0
I I
4.06.0
k . r\\"- , Aý
8.0 10.0
Time (s)
Figure Al 1-10. Transient volume fractions of liquid water at various elevations for Case 1 (single large
jet).
1.0
0.8
.o_
C) 0.6
E
0* 0.4
-J
0.2
0.0
0.0 2.0 4.0 6.0 8.0 10.0
Time (s)
Figure Al1-11. Transient volume fractions of liquid at various elevations for Case 2 (25 small jets).

Appendix AI 1
Al 1.4 Summary of Assessment Results
The FCI model implemented into the SCDAP/RELAP5/MOD3.3 code was subjected to an
assessment that was limited in scope but which indicates that the model and its implementations are
complete enough for preliminary calculations for severe accident analysis. The model and its
implementation were assessed using the results of MAGICO-2000 and FARO experiments. The
assessment showed that the model is calculating several important features of FCI fairly well, including
the motion of particles resulting from FCI, the fraction of break-up of slumping material, the rate of steam
production from FCI, and the maximum pressure increase in a region in which FCI is occurring. The FCI
model calculates that a jet that divides into several jets due to contact with a porous structure such as a core
plate will break up significantly more than when a division into several jets does not occur. This result is
consistent with the results of experiments on FCI performed at ANL.Al 1-4
The assessment also showed that the FCI model and its implementation and material property data
base are open to improvement. The assessment showed that the FCI model underpredicts the size of
particles resulting from FCI and underpredicts the volume fraction of steam in regions where FCI is
occurring. The resolution of these discrepancies between calculated and measured results may require the
following actions. First, a more accurate correlation for the interfacial surface tension of a mixture of
molten U0 2 and ZrO 2 in contact with water needs to be obtained. Second, the RELAP5 calculation of flow
regime mapping and steam velocity in a region with FCI may need to be improved. The first two steps to
this improvement are suggested to be; (1) impose a flow regime that is consistent with the configuration of
steam in a region where FCI is occurring, and (2) eliminate in the implementation of the FCI model the
assumption that the momentum change of the slumping material is small compared to the momentum of
the water and steam.
A significant number of experiments other than those used in this report are available for assessing
the FCI modeling in SCDAP/RELAP5/MOD3.3 and for providing insights on how to best improve the
2
modeling. These experiments include other FARO experimentsA10-5'AI0"6,A10-7'Al1-8,A10"I and the
fragmentation and quenching experiments performed at ANL.AI0-3, AlO 9Table A1O-6 lists some of the
experimental conditions from the FARO test series and Table A 10-7 gives results obtained from the tests.
The tests performed at ANL differed somewhat from those performed in the FARO test series. The ANL
experiments had a lower system pressure, water subcooling, metal contents in the melt, and melt jet
leading edge information. The data from the ANL experiments CCM-5 and CCM-6 are given in Table
A10-8. Several other sources of data for assessing the FCI model are also available. There are test results
available from KROTOS series of experimentsA1°-1 (number 33 and 37, without steam explosion)
performed by the Joint Research Center, Ispra, Italy, FITS series of experiments performed by Sandia
National Laboratory,Al°'l° and other MAGICO experiments performed by Theofanous and
coworkers.A10-11

Appendix All
Table Al 1-6. Experiment parameters for FARO series of experiments
Melt:
Composition (wt%)
U0 2 + ZrO 2 + Zr 80+20+0 80+20+0 76. +19.2+ 4.1 80+20+0
Mass (kg) 18 44 151 125
Temperature (K) 2923 3023 2823 3073 + 50
Delivery nozzle diam. (m) 0.1 0.1 0.1
Metallostatic head in 0.1
release vessel (m) 0.34 0.47 0.59 0.57
Free fall in gas (m) 1.66 1.53 1.09 1.04
Water:
Mass (kg) 120 255 608 623
Depth (m) 0.87 1.00 2.00 2.05
Initial mean temp. (K) 539 536 535 537
Gas phase:
Composition (wt(%)
steam + argon 83+17 70+30 77+23 77+23
Volume (m3 ) 0.464 0.875 1.280 1.260
Initial pressure (MPa) 5.0 5.8 4.9 5.0
Initial mean temp. (K) 543 536 536 536
Water container:
Diameter (m) 0.470 0.710 0.710 0.710
Volume (M3 ) 0.640 1.3 1.3 1.3
Table A11-7. Summary of FARO experimental results.A 14
L-06 L0
L-08
(Scoping t) (Quenching L-11 L-14
test)
Melt:
Delivery time (s) 0.28 0.37 - 1 - 1
Broken-up particles (kg) 12 30 151 105
Collected at the bottom plate
(kg) 6 14 0 20
Particles mean diam. (mm) 4.5 3.8 3.5 4.8
Pressure increase (MPa):
Before melt-water contact 0.4 0.3 0.3 0.1
Shortly after contact 1.1 (at 1.2 s) 1.8 (at 1.2 s) 5.1 (at 2.15 s) 2.8 (at 2.4 s)
Long-term maximum 1.6 at 12 s) 1.8 (at 20 s) 2.8 (at 22 s) 3.4 (at 30 s)
Maximum rate (MPa/s) 1.6 3.3 4.8 2.4
All1-13 NUREG/CR-6150-Rev. 2, Vol. 5

Appendix All
Table Al 1-7. Summary of FARO experimental results.A1 4 (Continued)
L-06 L-08
(Quenching L-11 L-14
(Scoping test)
test)
Temperature increase (K):

Gas phase (top of vessel) 40 43 67 77
Water 15 23 27 28
Level swell:
Maximum indicated by level
meters (m) 0.130 0.410 1.000 1.100
Bottom plate:
Maximum temperature
increase (K) Not available 275 20 330
Table Al 1-8. Summary of CCM-5 and CCM-6 experimental parameters
CCM-5 CCM-6
Melt:
Composition (wt%)
-U0 2 +ZrO 2 +stainless steel 60+ 16+24(67+21+ 12) 60+ 16+24(67+21+ 12)
(Fe + Cr + Ni)
Mass (kg) 10.87 11.34
Initial melt superheat (K) 157 157
Delivery nozzle diam. (m) 0.0508 0.0508
Free fall in gas (m) 0.34 0.34
Water:
Depth (m) 1.07 1.07
Initial water subcooling (K) 49.0 2.50
Gas phase:
Composition (mole%) - steam
+ argon 17+83 96+4
Volume (m3) 0.46 1.14
Initial pressure (MPa) 0.117 0.114
Initial mean temperature (K) 363 404
Water container:
Pressure at onset of melt
release (MPa) 0.117 0.202
A11.5 REFERENCES
Al1-1 S. Angelini, T. G. Theofanous and W. W. Yuen, "The Mixing of Particle Chloride Plunging into
Water," Proceedings of the 7-th InternationalMeeting on Nuclear Reactor Thermal-Hydraulics,
(NURETH-7), SaratogaSprings, NY, September 1995, NUREG/CP-0142, pp. 2359-2389.

Appendix All
All-2 D. Magallon and H. Hohmann, "Experimental Investigation of 150-kg scale Corium Melt Jet
Quenching in Water," Proceedingsof the 7th InternationalMeeting on Nuclear Reactor Thermal
Hydraulics (NURETH-7), Saratoga Springs, NY, September 1995, NUREG/CP-0142, pp. 1688
1711.
All-3 J. M. Broughton, P. Kuan, D. A. Petti, and E. L Tolman, "A Scenario of Three Mile Island Unit 2
Accident," Nuclear Technology, 87, August 1989, pp. 34-53.
Al1-4 C. C. Chu, J. J. Sienicki, and B. W. Spencer, "Validation of the THERMAL-i Melt-Water

Interaction Code," Proceedings of the 7th International Meeting on Nuclear Reactor Thermal
Hydraulics (NURETH-7), Saratoga Springs, NY, September 10-15, 1995, NUREG/CP-0142, pp.
2359-2389.
A11-5 D. Magallon and H. Hohmann, "High Pressure Corium Melt Quenching Tests in FARO." Nuclear
Engineeringand Design, 155, 1995, pp. 253-270.
Al1-6 H. Hohmann, D. Magallon, H. Schins, and A. Yerkess, "FCI Experiments in the Aluminum Oxide/
Water System," Nuclear Engineeringand Design, 155, 1995, pp. 391-403.
Al1-7 D. Magallon and H. Hohmann, "Experimental Investigation of 150-kg-Scale Corium Melt Jet
Quenching in Water," Proceedingsof the 7th InternationalMeeting on Nuclear Reactor Thermal
Hydraulics (NURETH-7), Saratoga Springs, NY, September 1995, NUREG/CP-0142, pp. 1588
1711.
Al 1-8 I. Huhtiniemi, H. Hohmann, and D. Magallon, "FCI Experiments in the Corium/Water System,"
Proceedingsof the 7th InternationalMeeting on NuclearReactor Thermal-Hydraulics(NURETH
7), SaratogaSprings,NY, September 1995, NUREG/CP-0142, pp. 1712-1727.
Al1-9 B. W. Spencer, S. K. Wang, C. A. Blomquist, L. M. McUmber, and J. P. Schneider,

"Fragmentation and Quench Behavior of Corium Melt Streams in Water," Argonne National
Laboratory Report ANL-93-32, NUREG/CR-6133, February 1994.
Al 1-10 B. W. Marshall, Jr., "Recent Fuel-Coolant Interaction Experiments Conducted in the FITS Vessel,"
ANS Thermal Hydraulic Division Proceedings, 25th ASME/AIChE National Heat Transfer
Conference, Houston, Tx, July 24-27, 1988, pp. 265-275.
All-l1 S. Angelini, E. Takara, W. Yuen, and T. G. Theofanous, "Multi-Phase Transients in the Premixing
of Steam Explosions," Proceedings of the Fifth International Meeting on Nuclear Reactor
Thermal-Hydraulics(NURETH-5), Salt Lake City, UT, September 21-24, 1992, pp. 471-478.
All-12A. Annunziato, C. Addabbo, A. Yeikess, R. Silverii, W. Brenka, and G. Leva, "OECD/CSNI

International Standard Problem 39 on FARO Test L-14 on Fuel Coolant Interaction and Quenching
- Analysis of the Results," EUR17736EN, 1998.

Appendix A 12
A.12. ASSESSMENT USING TMI-2 ACCIDENT
A12.1 Introduction
The TMI-2 Accident Test Problem provides under full scale conditions an assessment of virtually
every in-core damage progression model in the SCDAP/RELAP5/MOD3.3 code. Since the accident was
not an experiment, the measured behavior of the reactor during the accident was limited. Nevertheless, the
limited amount of information obtained during the accident and the post-accident examinations of the
reactor provide a significant amount of information for assessing at least in an indirect manner most of the
in-core damage progression models in the code. The TMI-2 accident involved spacer grid meltdown,
cladding ballooning, control rod meltdown, fuel rod oxidation, hydrogen production, cladding meltdown,
fuel melting, molten pool formation, quenching of hot and embrittled fuel rods, and molten pool slumping.
The measurements obtained during the accident and inferences made from observations after the accident
only provide quantitative assessment of a few of the damage progression models, such as models that
calculate total hydrogen production, location of previously molten frozen material, and total amount of
molten material. Nevertheless, since one damage progression event in the overall chain of damage
progression events is dependent upon all the other previous damage progression events in the chain, the
correct calculation of a few of the damage progression events cannot be made without a correct calculation
of the other damage progression events. So an indirect assessment can be made of virtually every in-core
damage progression model in the code.
A12.2 Description of TMI-2 Accident Problem
All major components of the TMI-2 primary system were represented in the TMI-2 Accident Test
Problem. The RELAP5 module was used to simulate the thermal-hydraulics of the reactor vessel, primary
coolant loops, steam generators, and pressurizer. Steam generator secondary side coolant levels, pressures,
and feedwater temperatures, and primary side makeup and letdown flow rates were supplied as boundary
conditions. The SCDAP module was used to simulate the reactor core, which was divided into five radial
regions by grouping similarly powered fuel assemblies together.
The TMI-2 accident is generally divided into four distinct phases for analysis purposes.A 12-1 Phase 1
(0 - 100 min) is a small-break loss-of-coolant accident (LOCA) through the stuck-open pilot-operated
relief valve (PORV). One or more reactor coolant system (RCS) pumps operated continuously during
Phase 1 of the accident, thereby providing adequate core cooling. Phase 2 (100 - 174 min) is a continuation
of a small break LOCA without the RCS pumps. Core uncovery, heatup, and initial melting occurred
during Phase 2. Phase 3 (174 - 200 rin) begins with a restart of reactor coolant pump 2B. Approximately
30 m 3 of coolant was injected into the reactor vessel in less than one minute, cooling the peripheral fuel
assemblies and forming an upper core debris bed with significant zircaloy oxidation. Heatup of the
degraded core region, with the formation and growth of a pool of molten material, continued during Phase
3. Phase 4 (200 - 300 min) begins with the initiation of high pressure injection (HPI). The central region of
the partially molten core material was not coolable by HPI even through the water level reached the level
of the hot legs by 207 min. Between 224 and 226 min, the crust encasing and supporting the molten core
region is believed to have failed, allowing molten material to relocate to the lower plenum. Summaries of
the measured and observed reactor core damage are given in Reference A 12-2 and Reference A 12-3.
The RELAP5 portion of the TMI-2 model was derived from an Oconee plant model described in
Reference A12-4. Both TMI-2 and Oconee are PWR's having a two-by-four coolant loop configuration,
i.e., two primary coolant loops, each containing one hot leg and two cold legs. Both plants were built by

Appendix A 12
Babcock & Wilcox in the 1970's and have nearly identical design and operating characteristics.
Consequently, the Oconee RELAP5 model was easily adapted to represent TMI-2. Figure A 12-1 through
Figure A12-4 are nodalization diagrams of the reactor vessel, primary piping, steam generators, and
pressurizer respectively.
Shot
WWO
FfrA CM *9S
w~o4e4-1
Figure Al 2-1. RELAP5 nodalization of the reactor vessel and core.
The RELAP5 vessel model (Figure A12-1) represent all major components of the reactor vessel,
including the inlet annulus, downcomer, lower plenum, core, core bypass, upper plenum, upper head,
reactor vessel vent valves, and the control rod guide tube brazements. The core is divided into five parallel
channels, each consisting of ten subvolumes (branch components 10 through 59). Lateral flow between
adjacent core channels is simulated using the RELAP5 crossflow model. Annulus component 570
represents the downcomer and pipe component 510 the core bypass. Branch components 505 and 575
represent the lower plenum. The upper plenum is also divided into five parallel regions that are connected
laterally by crossflow junctions. This arrangement allows for the development of in-vessel natural
circulation under appropriate conditions. Valve component 542 represents the reactor vessel vent valves
and pipe components 580 through 584 the guide tube brazements. Fifty-one heat structures were used to
model the thermal behavior of reactor vessel metal structures.
Figure A 12-2 is a nodalization diagram of the primary coolant loop A, which consists of one hot leg
(components 100 through 114), one steam generator (discussed subsequently), two pump suction legs
(pipe components 135 and 165), and two cold legs (components 140 through 151 and 170 through 181).
Primary loop B is identical to loop A, except it does not contain a letdown flow path (time-dependent
junction 193) or connections to the pressurizer spray and surge lines. The component numbers for loop B
are also increased by 100 (e.g., the primary pumps are numbered 235 and 265 rather than 135 and 165).
NUREGICR-6150-Rev. 2, Vol. 5 A12-2

Appendix A12
111o
To plenumin
spra
OT0
Vessel
downomer
HPI
8074WANG-W
Figure A12-2. RELAP5 nodalization of primary coolant loop A.

The high pressure injection (HPI) system is represented by time-dependent volumes 710 and 715, which
are connected to the cold legs by time-dependent junctions 711 and 716. HPI flow is assumed to be split
equally between the A and B loops while makeup flow is injected only into the B loop. Eighteen heat
structures (per loop) were used to model the thermal behavior of the primary piping.
The nodalization of steam generator A is shown in Figure A12-3. Steam generator B is identical
except that all component numbers are increased by 100. The boiler region is divided into two parallel
flow channels: an inner channel (volumes 310 through 323), connected to 90% of the steam generator
tubes, and an outer channel (volumes 360 through 373), connected to 10% of the steam generator tubes.
Crossflow junctions connect the two boiler regions. Auxiliary feedwater is normally injected into the top
of the 10% region. Pipe component 120 represents the primary side of the steam generator tube bundle,
while branch components 115 and 125 represent the inlet and outlet plena. The steam generator
downcomer is modeled by component 305, and components 345 and 350 represent the steam line. To
preheat the feedwater, a portion of the steam flow is bled into the downcomer through an aspirator near
mid-boiler (modeled with a junction between component 365 and 305). Valve component 821 represents
the main steam valve. Forty-three heat structures were used to model the thermal behavior of steam
generator metal structures (including the tube bundle).
It should be noted that for all the calculations reported here, auxiliary feedwater was injected into the
steam generator downcomer rather than into the tube bundle as indicated in Figure A12-3. A previous

Appendix A12
T T Mm•MWas.
Figure A12-3. RELAP5 nodalization of steam generator A.

TMi-2 analyses using SCDAPIRELAP5a indicated that feedwater injection directly onto the steam
generator tubes resulted in too much primary-side condensation, which in turn caused the primary system
pressure to be underpredicted. It should also be noted that the nodalization of the steam generators differs
somewhat from that recommended in the SCDAP/RELAP5 user's guide.A 12 -5 Although the boiler is
divided into two parallel regions, the tube bundle (pipe 120) is not. Dividing the tube bundle into a 10%
region and a 90% region may alleviate some of the condensation problems encountered previously.
Figure A12-4 is a nodalization diagram of the pressurizer. The pressurizer upper head is modeled
with branch component 615 and the pressurizer cylindrical body and lower head with pipe component 610.
Valve 801 represents the pilot operated relief valve (PORV). Pipes 600 and 620 represent the pressurizer
surge and spray lines, respectively, and valve 616 models the spray valve. Single volume component 949
represents the containment building, which is initially filled with air at 101 kPa. Twelve heat structures are
used to model the thermal behavior of the pressurizer shell, upper and lower heads, and the surge line; one
heat structure is used to simulate operation of the pressurizer heaters; and five heat structures are used to
a. C. A. Dobbe, private communication, EG&G Idaho, Inc., March 15, 1994.

Appendix A 12
sRV PORV
S7Spray
616/ valve
610 - - -
------- 620 Sp ray
line
ROMn 110
Surge line
:_F=m 170
Figure A12-4. RELAP5 nodalization of the pressurizer.

model the thermal behavior of the containment building.
It should be noted that critical flow through the PORV is modeled using the homogeneous (single
velocity) two-phase flow option in RELAP5. Previous TMI-2 calculations using SCDAP/RELAP5a have
shown that this option better predicts the PORV flow history reported in the TMI-2 initial and boundary
conditions (ICBC) data baseA1 2 - 6 than other options. [The PORV flow rates reported in Reference A12-7
were calculated using the Henry-Fauske critical flow model for subcooled conditions and the
homogeneous equilibrium critical flow model (HEM) for two-phase conditions.] It should also be noted
that for all calculations reported here, a servo valve was installed between the pressurizer and the surge
line at 117 min to prevent the pressurizer from draining. In preliminary calculations, the pressurizer
drained completely after the PORV block valve was closed at 139 min, which effectively terminated core
heatup. More accurate representations of the surge line and pressurizer might eliminate some of the
problems encountered in this and previous TNfl-2 analyses. For example, the junction connecting the surge
line to hot leg A should be oriented horizontally rather than vertically (to reflect its true alignment) and the
countercurrent flow limitation (CCFL) model should be activated at the junction connecting the surge line
to the pressurizer, rather than at the hot leg junction. Also, the CCFL input parameters (currently set to
default values) should be reviewed for applicability.
a. C. A. Dobbe, private communication, EG&G Idaho, Inc., March 15, 1994.

Appendix A 12
The TMI-2 core was divided into five regions for this analysis by grouping similarly powered fuel
assemblies together. Figure A12-5 is a cross-section of the core illustrating each region and its average
radial power peaking factor. Table A12-1 lists the average axial power peaking factors for each region.
Both the axial and radial peaking factors were derived from detailed peaking factor data presented in
Appendix A of Reference A 12-6.
Table A12-1. TMI-2 axial power peaking factors.
Power factor
Distance
from Region I Region 2 Region 3 Region 4 Region 5
bottom of
fuel (m)
01.183 0.665 0.674 0.729 0.690 0.670

0.549 0.933 0.919 0.962 0.951 0.944
0.914 1.134 1.099 1.112 1.132 1.145
1.280 1.216 1.164 1.112 1.168 1.213
1.646 1.248 1.202 1.138 1.192 1.238
2.012 1.262 1.221 1.153 1.206 1.248
2.377 1.225 1.232 1.251 1.241 1.222
2.743 1.078 1.124 1.174 1.131 1.083
3.109 0.792 0.853 0.880 0.834 0.794
3.475 0.448 0.512 0.488 0.455 0.442
One SCDAP fuel rod component is used to represent all the fuel rods in each core region. One
SCDAP control rod component is used to represent all the full and part-length control rods, all the guide
tubes (including those containing burnable poison rods), and all the instrument tubes in each core region
(except region five which contains no control rods). The control rod radii in regions one through four have
been adjusted so that the total mass of zircaloy, Ag-In-Cd absorber, and stainless steel is conserved (the
burnable poison mass is neglected). In core region five, a dummy fuel rod component is used to represent
all the guide and instrument tubes. By specifying a small fuel diameter and zero power, this component
essentially behaves as a hollow zircaloy tube. The SCDAP grid spacer model is used to represent the eight
inconel spacer grids that are uniformly distributed along the length of each fuel assembly.
Much of the SCDAP input data was obtained from Reference A12-8 and is summarized in Table
A 12-2. Table A 12-3 lists the total number of fuel assemblies, fuel rods, control rods, burnable poison rods,
and orifice rods in each core region.

Appendix A 12
I -1rI L-F I I I 1
:X
1 12 5-
I I I r] [71 1 1 1j7E
4- -* - - * - q- U- - -
- *- S - * - * -4-P - I- b-
I
Radial
Region Peaking Factor
1 1.245
2 1.136
3 1.074
4 1.061
5 0.733
M028-BDR-0494-001
Figure A12-5. Cross-sections of core showing fuel assembly grouping and radial peaking factors.
Table A12-2. Total fuel assemblies, fuel rods, and control rods in each core region.
Full- Part- Burnable

Core Fuel Fuel length length poison Instrument Orifice
regions assemblies rods control control rods tubes rods
rods rods
1 13 2704 144 0 64 13 0
2 28 5824 256 0 192 28 0
3 40 8320 192 128 320 40 0
4 48 9984 384 0 384 48 0
5 48 9984 0 128 48 640
Total 177 36816 976 128 1088 177 640

Appendix A12
Table A12-3. SCDAP input parameters.

Parameter Value
Fuel rods
Active height (in) 3.568
Rod pitch (in) 1.443 x 10-2
Cladding inner radius (in) 4.788 x 10-3
Cladding outer radius (in) 5.461 x 10-3
Fuel pellet radius (in) 4.699 x 10-3
Fuel density (% T.D.) 92.5
Mass of He fill gas (estimated) (kg) 1.265 x 104

3
Upper and lower plenum void volume (mi ) 1.490 x 10-5
Control rods
Guide tube inner radius (in) 6.325 x 10-3
Guide tube outer radius (in) 6.731 x 10-3
Cladding inner radius (in) 5.055 x 10-3
Cladding outer radius (in) 5.588 x 10-3
Absorber radius (m) 5.004 x 10-3
Instrument tubes
Tube inner radius (in)
Tube outer radius (in)
Grid spacers I
Grid spacer mass (kg)
Grid spacer height (in)
Grid spacer thickness (in)
Table A12-4 compares the initial conditions in the SCDAP/RELAP5 model to those recommended
in the ICBC data base.a with the exception of steam generator pressures and temperatures, the calculated
(or specified) initial conditions are in good agreement with the data base. For steady-state calculations, a
control system is used in the SCDAP/RELAP5 model to automatically adjust steam generator pressures
a. All initial conditions correspond to the time of turbine trip: 04:00:37 hours on March 28, 1979.

Appendix A12
(by varying the flow areas of the main steam valves) until user-specified cold leg temperatures are
obtained. For simplicity, the target coolant temperature for all four cold legs was specified to be 565 K.
Table A12-5 compares the calculated initial conditions on the secondary side of each steam generator to
the initial conditions recommended in Reference A12-9. It is seen that the calculated steam generator
pressures are in much better agreement with the Reference A12-9 data than with the ICBC data base.a
Calculated steam generator coolant levels, however, differ considerably from those reported in Reference
A12-9. For future calculations, it is recommended that the steam generator models should be adjusted to
better represent the Reference A 12-9 data. One way to accomplish this may be to increase the pressure
drop across the tube support plates as was done for a TMI-2 analysis performed with the CATHARE
code.AI2-10
Table A12-4. TMI-2 initial conditions at turbine trip.
Parameter ICBC data SCDAP/
base RELAP5
Reactor power (MW) 2700 2700

Primary system pressure (MPa) 15.2 15.2
Pressurizer level (m) 5.77 5.76
Pressurizer heater power (MW) 1.39 1.39
Cold leg temperature 1A (K) 561 565
Cold leg temperature 2A (K) 548 565
Hot leg temperature loop A (K) 592 593
Hot leg temperature loop B (K) 592 593
Makeup flow (kg/s) 5.44 0.0
Letdown flow (kg/s) 4.18 0.0
PORV flow (kg/s) 2.59 0.0
Feedwater temperature (K) 513
Steam generator A pressure (MPa) 7.31 6.34
Steam generator B pressure (MPa) 7.24 6.28
Steam generator A steam temperature (K) 586 576
Steam generator B steam temperature (K) 585 582
a. The pressures reported in Reference A12-9 are average steam line pressures measured 10 to 0.1 min before
turbine trip.

Appendix A12
Table A12-5. Steam generator initial conditions.

SCDAP/
Reference A 12-9 RELAP/
Parameter RELAP5
Steam generator A feedwater flow (kg/s) 722 723
Steam generator B feedwater flow (kg/s) 718 717
Steam generator A pressure (MPa)a 6.38 6.34
Steam generator B pressure (MPa)a 6.24 6.28
Steam generator A steam temperature (K) 586 576
Steam generator B steam temperature (K) 586 582
Steam generator A riser level (cm) 526 197
Steam generator B riser level (cm) 538 183
Steam generator A downcomer level (cm) 660 559
Steam generator B downcomer level (cm) 669 543
Steam generator A power (MW) 1346 1332
Steam generator B power (MW) 1339 1378
All boundary conditions, except HPI/makeup flow rates, were obtained from the ICBC data base
(Reference A12-6). The HPI/makeup flow rate history reported in Reference A12-11 was adjusted until
the time of core uncovery (as inferred from hot leg temperature measurements), the time of initial fuel rod
cladding failure (as inferred from containment radiation measurements), and the primary system pressure
history were predicted reasonable well.a Figure A12-6 compares the HPI/makeup flow rate history used
for the best-estimate SCDAP/RELAP5 calculation discussed subsequently to that recommended in
2 2
Reference A 12-6. In a previous SCDAP/RELAP5 analysis of the TMI-2 accident,A1 "1 using a previous
version of the code, better results were obtained by reducing the makeup flow rate from 4 to 2 kg/s
13
between 100 and 174 min. In an analysis performed with the MELPROG/TRAC code,A12- it was
concluded that the makeup flow rate recommended in Reference A 12-6 was too high between 12 and 100
min. For that analysis, the flow was reduced from 6.5 to 1 kg/s between 12 and 100 min (which was the
nominal value given in the original issue of the ICBC data base). Core power as a function of time for the
first 400 s following reactor scram was estimated using the reactor (point) kinetics and decay heat models
in the RELAP5 code. The decay power from 617 minutes onward was obtained from Reference A12-14.
Figure A 12-7 shows the reactor power versus time curve in the SCDAP/RELAP5 model.
a. The uncertainty in HPI/makeup flow is large, particularly between 100 and 174 nin. A12-11 Consequently, as
noted in Reference A12-12, one set of assumptions is probably as good or bad as another.
NUTREG/CR-6150-Rev. 2, Vol. 5 A12-10

Appendix A 12
80.0
60.0
40.0
0
M
20.0
0.0 L
0.0 100.0 200.0 300.0
Time (s)
Figure A12-6. Makeup flow history for TMI-2 calculation.
100
10
0
A.
0.1 4
[0a C
I
0
.
g
40
I
go
C
I
120
C
I
160
I
200
I
I
240
I
280 220
Time (min)
Figure A12-7. Reactor power versus time curve used for TMI-2 calculation.

Appendix A 12
A12.3 Assessing Using TMI-2 Accident
The timing of the opening and closing of the PORV block valve is an important boundary condition
for the TMI-2 accident. Figure A12-8 is a plot that shows the times at which the block valve was opened
and closed and the calculated rate of mass flow through the valve for the periods of time that it was open.
The plot also shows the calculated rate of flow through the PORV valve. The calculated rate of flow
through the open PORV valve ranged from 5 to 35 kg/s.
40.0
30.0
cc
20.0
LL
0
10.0
0.0 L1
0.0 5000.0 10000.0 15000.0
Time (s)
Figure A12-8. Timing of closure of PORV block valve and history of calculated rate of flow through
PORV valve.
The SCDAP/RELAP5 MOD3.3 solution to the TMI-2 Accident test Problem is presented next. The
calculated timing of the damage progression is described and calculated results are presented for which
measurements are available for comparison and assessment. These calculated results are; (1) location of
core material, (2) maximum amount of molten material, (3) reactor system pressure, and (4) hydrogen
production. Default values were used for all the modeling parameters subject to being defined by the code
user.
The damage to the TMI-2 reactor core began with the ballooning and rupture of the cladding of the
fuel rods and advanced to the slumping of a significant amount of molten core material to the lower head
of the reactor vessel. The MOD3.3 calculated and measured progression of damage are compared in Table
A12-6. The results of the MOD3.2 calculation of the TMI-2 accident are included in the table. After core
uncovery was calculated to occur at 6270 s, core damage was calculated to occur quite rapidly. Cladding
failure due to ballooning was calculated by MOD3.3 to begin at 8445 s. MOD3.2 calculated cladding
failure to be due to chemical attack by the Inconel spacer grids and to not occur until 9417 s. After
cladding failure, double-sided oxidation of the cladding was calculated by MOD3.3 to occur in the vicinity
of the cladding failure. The melting of fuel and the formation of a molten pool were calculated by MOD3.3
to begin at 9500 s. MOD3.2 calculated molten pool formation to begin at 10,330 s. The start-up of the 2B
pump at 10,446 s was calculated by both MOD3.3 and MOD3.2 to cause a rapid increase in the pressure of
the primary coolant system and to accelerate the rate of hydrogen production. These calculations are in

Appendix A 12
agreement with the measurements. MOD3.3 calculated 9,530 kg of molten material in the core region at
the start of the 2-B pump and 27,600 kg of molten material in the core just before the slumping of core
material to the lower head at 12,890 s. MOD3.2 calculated 1734 kg of molten material in the core region at
the start-up of the 2B-pump and 37,400 kg of molten material in the core just before the slumping of core
material to the lower head at 13,379 s. The post-accident examination of the TMI-2 reactor indicated that
40,800 kg of the reactor core was molten at some time. Inferences from the measured system pressure and
other measurements indicate that 15,800 kg of molten material slumped to the lower head at 13,500 s. Both
the MOD3.3 and MOD3.2 calculated masses of molten material and the time of slumping are in
approximate agreement with the measured values. The MOD3.3 calculated location of molten core
material was in fair agreement with the post-accident observation of the TMI-2 reactor. The elevations of
the bottom surface of the in-core molten pool at the centerline of the core were calculated and observed to
be 1.10 m and 0.11 m, respectively. The highest location in the core to become molten was calculated to be
2.9 m and to be located along the centerline of the core. The molten pool was calculated to extend in the
radial direction from the centerline of the reactor vessel to the periphery of the reactor core. The bottom
surface of the molten pool at the periphery of the core was calculated by MOD3.3 to be at the elevation of
2.2 m. These calculated results are in agreement with the post-accident observation of the TMI-2 core.
Table Al 2-6. Calculated timing and sequence of core damage progression.

Beginning of long term core 6,390 6,270

uncovery (s).
Beginning of cladding failure due to - 9,417 8,445
ballooning (s).
Beginning of spacer grid slumping - 9,418 9,112
(s).
Beginning of molten pool (s). - 10,330 9,530
Cumulative hydrogen production at 300 275 365
start-up of 2B-pump at 10,446 s
(kg).
Primary coolant system pressure at 8.20 5.03 6.96
start-up of 2B-pump (MPa).
Mass of molten material at start-up - 1734 15,000
of 2B-pump (kg).
Increase in primary coolant system 6.30 5.21 7.00
pressure after start-up of 2B-pump
(MPa).
Final cumulative hydrogen 460 453 417
production (kg).

Appendix A12
Table Al 2-6. Calculated timing and sequence of core damage progression. (Continued)

Mass of core material that was 40,800 37,400 27,600

molten during some period of
accident (kg).
Elevation of bottom of molten 0.71 1.46 1.10
region relative to bottom of core
(in).
Smallest distance from side of 0.0 0.0 0.0
molten pool to periphery of core
(in).
Time at which bulk of material in 13,500 13,379 12,890
molten material slumped to lower
head (s).
Mass of molten material that 15,800 37,400 27,600
slumped to lower head (kg).
Percent of molten material that 39 100 100
slumped to lower head (%).
MOD3.3 calculated a greater rise of the water level in the reactor core after activation of the 2B
pump than did MOD3.2. Both MOD3.3. and MOD3.2 calculated damage progression to be a strong
function of the calculated collapsed liquid level in the reactor vessel. Figure A12-9 are plots of the
MOD3.3 and MOD3.2 calculated collapsed liquid levels as a function of time. A temporary core uncovery
was calculated by MOD3.3 to begin at 4800 s and the long term uncovery was calculated to begin at 6330
s. After 6330 s, the core continued to uncover until the 2-B pump activation at 10,446 s. The collapsed
liquid level was calculated by MOD3.3 to be 0.3 in above the bottom of the reactor core just before
activation of the 2-B pump. MOD3.2 calculated the collapsed liquid level just before start-up of the 2B
pump to be 0.2 m above the bottom of the reactor core. After activation of the 2-B pump, MOD3.3
calculated the water level to rise 3.0 in and MOD3.2 calculated the water level to rise 0.8 in. This
difference in calculated rise in water level is due to MOD3.3 calculating a significantly greater fraction of
the reactor core to be molten at the start-up of the 2B-pump than MOD3.2. Both MOD3.3 and MOD3.2
calculated activation of the HPIS at 12,012 s to cover the entire reactor core with water within 600 s.
The level of water in the pressurizer has an influence on the level of water in the reactor vessel.
Figure A 12-10 is a plot of the measured water level in the pressurizer and the levels calculated by MOD3.3
and MOD3.2. In the period of 10,000 s to 13,000 s, both MOD3.3 and MOD3.2 calculated a lower level of
water in the pressurizer than the measured level.
MOD3.3 calculated severe core damage to begin about 800 s earlier than MOD3.2. The onset of
melting of the reactor fuel and the beginning of molten pool formation is a mark of the beginning of severe
core damage. The timing of damage progression is indicated by plots of the history of the maximum
temperature in the reactor core and of the effective radius of the molten pool, as shown in Figures A 12-11

Appendix A 12
12.0
10.0
8.0
0
o
0_ 6.0
Q.
0
4.0 -7o
2.0o 0
5000.0 10000.0 15000.0
Time (s)
Figure Al 2-9. History of calculated collapsed liquid level in reactor vessel.
12.0
10.0
E
N
(D
8.0
CO
0.
.C
6.0
"0
._V
,_1
4.0
2.0 L0
0.0 5000.0 10000.0 15000.0
Time (s)
Figure A12-10. Comparison of calculated and measured liquid level in pressurizer.

Appendix A12
and A12-12, respectively. The effective radius is the radius of a hemisphere with a volume equal to the
calculated volume of molten material. The MOD3.3 and MOD3.2 calculated timings for the beginning of
fuel melting were 9530 s and 10,330 s, respectively. MOD3.3 calculated a significantly more rapid heatup
of the reactor core after the beginning of oxidation than did MOD3.2. In the temperature range of 1800 K
to 2200 K, when rapid oxidation occurs, MOD3.3 calculated a heatup rate of nearly 50 K/s. The MOD3.2
calculated rate of heatup in this range of temperature was significantly less than that calculated by
MOD3.3. Both MOD3.3 and MOD3.2 calculated that activation of the 2B-pump at 10,446 s did not result
in any reduction in the maximum core temperature. A reduction in temperature was calculated to not occur
because a significant part of the core was in the form of a large molten pool and because of an increase in
oxidation of fuel rod cladding due to cracking of oxide layers and more flow of steam. The activation of
HPIS at 12,012 s was calculated to not cause any cooling of the molten part of the reactor core. MOD3.3
calculated no further melting of fuel after 10,480 s (34 s after activation of 2B-pump), while MOD3.2
calculated fuel melting to occur until 13,000 s. MOD3.3 calculated the maximum effective radius of the
molten pool to be 1.20 m. This value is in general agreement with the corresponding measured value of
1.40 m.
3000.0
" 2000.0
CL
0,.
E
SE
E
E
Ca
E 1000.0
0
0
0.0 " 5000.0 10000.0 15000.0

0.0
Time (s)
Figure Al 2-11. History of calculated maximum temperature in reactor core.
The MOD3.3 calculated and measured histories of the primary coolant system pressure were in
partial agreement. The measured and calculated pressure histories are compared in Figure A12-13. The
MOD3.2 calculated pressure history is also shown. The calculated and measured pressures were in good
agreement from the start of the accident until 7,500 s. In the period of 7,500 s to start-up of the 2B-pump at
10,446 s, the calculated pressure was significantly less than the measured pressure. The coolant system
pressure was calculated to increase about 7 MPa after start-up of the2-B pump. This calculated increase in
pressure was in general agreement with the measured pressure increase of 6.3 MPa. The increase in
pressure was calculated to be due to steam generated by pumping of water into the hot reactor core and by

Appendix A 12
1.5 1
- MOD 3.2 (0.6 durability) 0 Measured
.- - MOD3.2 (0.2 durability)
- MOD 3.3
0Ca- 11..0 •olMolten pool
E
E
Sslum ped to
lower head
16
.2
0 0.5
F I pu pps
Start-up of 2B-pump
Start-up of
• / (12,000 s)
0 00.0 11000.0 13000.0 15000.0
Time (s)
Figure A12-12. History of calculated effective radius of molten pool.
20.0
- MOD 3.3
--- MOD3.2
Measured
15.0
0
CL
10.0
U)
5.0 HPIS (112,000s

Start-up of 2B-pump
(10,446s) Z
0.%.0 5000.0 10000.0 15000.0
Time (s)
Figure A12-13. Calculated and measured pressures of primary coolant system.

Appendix A 12
heatup of the reactor core by an acceleration in the oxidation of fuel rod cladding. In the period of 10,500 s
to activation of HPIS at 12,012 s, the pressure was calculated and measured to gradually decrease a few
MPa. After activation of the HPIS, the pressure was calculated and measured to increase a few MPa.
The MOD3.3 calculated pressure history of the primary coolant system was improved using the
boundary conditions for the TMI-2 accident applied by Annunziato et al. A12.17 The TMI-2 accident has a
large degree of uncertainty in boundary conditions that influence the pressure of the primary coolant
system. The accident has uncertainties in boundary conditions such as make-up flow rate, performance of
the 2B-pump, flow rate through the PORV, and time of its closure. Since there are uncertainties in these
boundary conditions, the TMI-2 accident was also calculated by MOD3.3 using the boundary conditions
used by Annunziato at al. The pressure history calculated by a slightly earlier version of MOD3.3 with
these boundary conditions is compared with the measured pressure history in Figure A12-14. The
calculated pressure history in the period of 7500 s to 10,000 s was in general agreement with the measured
pressure history. Other aspects of reactor behavior, such as the total hydrogen production and total mass of
molten material, were also in general agreement with measurements and the post-accident observation.
20.0
- MOD 3.3
•- Measured
15.0
2? 10.0 \,
5.0
0.0 15000.0
0.0 5000.0 10000.0
Time (s)
Figure A12-14. Comparison of calculated and measured pressure histories of primary coolant system for
case of boundary conditions from Annunziato.
The calculated hydrogen production was in general agreement with the hydrogen production
estimated from post-accident observations and inferences. The MOD3.3 calculated cumulative hydrogen
production is compared with the measured hydrogen production in Figure A12-15. The MOD3.2
calculated hydrogen production is also shown in the figure. MOD3.3 calculated the rapid production of
hydrogen to begin sooner than MOD3.2. Nevertheless, both the MOD3.3 and MOD3.2 calculations of
hydrogen production are in approximate agreement with the measured hydrogen production. At the start
up of the 2B-pump, the MOD3.3 and MOD3.2 calculations of cumulative hydrogen production were 275
kg and 365 kg, respectively. The measured hydrogen production at the start-up of the 2B-pump was 300
kg. The MOD3.3 calculated and measured total hydrogen productions were 417 kg and 460 kg,

Appendix A 12
500.0 1
MOD 3.3 Measured--*-*
--- MOD 3.2
400.0
.0
M asured
2 300.0
C
." 200.0
e-0
Start-up of
:3 HPIS (12,000 s)
E
E 100.0 Start-up of
/ 2B-pump
- (1 0;446s)
0) 0 : 9000.0 11000.0 13000.0 15000.0

Time (s)
Figure A12-15. Calculated hydrogen production during TMI-2 accident.

respectively. Both MOD3.3 or MOD3.2 calculated that no significant amount of hydrogen production
occurred 50 s after the start-up of the 2B-pump (10,500 s). Hydrogen production was calculated to not
occur after 10,500 s because the portions of the core with intact fuel rods and some metallic cladding were
too cool to rapidly oxidize. This behavior is shown in Figure A12-16, where the MOD3.3 calculated
temperature histories are shown of the fuel rods in the outer most fuel assemblies of the reactor core at the
elevations of 3.11 m and 3.47 m, respectively. As shown in this figure, hydrogen production at the 3.11 m
and 3.47 m elevations was calculated to stop due to the cooling caused by the start-up of the 2B-pump. The
oxide layer at this location was calculated to be too thin to crack during the reflood caused by start-up of
the 2B-pump.
The MOD3.3 calculation of the disintegration of fuel rods into porous debris was in agreement with
the post-accident observation of the locations in the reactor core with porous debris. Porous debris regions
were calculated to form in the outer most fuel assemblies in the elevation interval of 0.5 m to 1.2 m and
across the entire diameter of the reactor core in the elevation interval of 2.6 m to 3.6 m. The calculation of
porous debris in the elevation interval of 2.6 m to 3.6 m was consistent with the post-accident observation
of the state of the reactor core.
The porous debris thermal hydraulic models in MOD3.3 performed properly after activation of the
HPIS, which resulted in two-phase coolant conditions in porous debris in the upper part of the reactor core.
The calculated temperature history at a location with porous debris is shown in Figure A12-17. The plot
applies for the location 2.7 m in elevation and in the fourth ring of fuel assemblies. This location
disintegrated from intact fuel rods to porous debris soon after start-up of the 2B-pump, and thus was
porous debris at the time of start-up of HPIS. As a result, the flow losses and heat transfer at this location
during the reflood period beginning with the start-up of HPIS were calculated with the porous debris
thermal hydraulic models implemented into MOD3.3. The debris had a porosity of 0.46 and a particle

Appendix A12
3000.0 1 1
3.11 m elevation
........ 3.47 m elevation
2000.0
(D
c
E
'- 1000.0
-. o...oo..oo..o...o .... , .p°...............
Start-up of 2B-pump
(10,446 s)
0.0 1
0.0 5000.0 10000.0 15000.0
Time (s)
Figure A12-16. MOD3.3 calculated temperature histories of fuel rods in upper part of outer most fuel
assemblies in reactor core.
diameter of 3.5 mm. Since the molten pool was located below this location and blocked the upward flow of
water from the HPIS, the debris was flooded from the top down. The calculated temperature history of this
location following reflood of the reactor core beginning at 12,012 s was consistent with that seen in
experiments on the quenching of porous debris, as described in Appendix 8.
The implementation into MOD3.3 of the integral diffusion model for fuel rod oxidation caused it to
calculate a more rapid progression of damage to the reactor core than calculated by MOD3.2. The onset of
severe damage was calculated to begin at 9530 s by MOD3.3 and at 10,330 s by MOD3.2. This difference
is due to the integral diffusion model for oxidation in MOD3.3 calculating a more rapid heatup due to
oxidation at locations with a rich supply of steam than that calculated by the parabolic kinetics model for
oxidation in MOD3.2. These differences in calculated behavior also occurred in the analyses of severe fuel
damage experiments described in Appendices A2 through A4.
The MOD3.3 calculation of the TMI-2 accident was improved by the implementation into MOD3.3
of the stress-based model for calculating the time of failure of an oxide layer retaining the melted metallic
part of the fuel rod cladding. While MOD3.2 applied one model for failure of the oxide layer for analysis
of severe fuel damage experiments and another model for the analysis of the TMI-2 accident, MOD3.3
applied the same oxide failure model for analysis of severe fuel damage experiments and the TMI-2
accident. The calculations of the oxide failure model have a strong influence on the calculations of
hydrogen production and the extent melting of fuel assemblies. With the stress-based model for calculating
oxide failure, MOD3.3 calculated hydrogen production and extent of melting in general agreement with
measurements for both severe fuel damage experiments and the TMI-2 accident. On the other hand, when
using the same oxide failure model for the TMI-2 analysis as used for analysis of severe fuel damage
experiments, MOD3.2 underpredicted by a factor of two the extent of melting of the reactor core, and did
not predict any slumping of molten core material to the lower head.
NUIREG/CR-6150-Rev. 2, Vol. 5 A12-20

Appendix A12
1000.0
900.0 Quenching due to

temporary
reflood following start-up of
2B-pump
S 800.0
E 700.0
i-CD
Quenching
600.0 following S
start-up Of HPIS
of HPIS
500.0 i•
11500.0 12000.0 12500.0
Time (s)
Figure A12-17. MOD 3.3 calculated temperature history of location with porous debris (2.7 m elevation
of fourth ring of fuel assemblies).
Another difference in calculated reactor core behavior between MOD3.3. and MOD3.2 was in the
calculated ballooning and rupture of fuel rods; MOD3.3 calculated ballooning and rupture of fuel rod
cladding to occur significantly earlier than MOD3.2. This difference in calculated behavior is due to
corrections made in MOD3.3 to the model for ballooning of the fuel rod cladding.
A12.4 Conclusions
The MOD3.3 calculation of the TMI-2 accident showed that its new models result in calculated
behavior of the reactor core and primary coolant system in general agreement with measurements and
post-accident observations. The calculated and measured hydrogen productions were 417 kg and 460 kg,
respectively. The calculated and measured masses of molten material in the core region were 27,600 kg
and 40,800 kg, respectively. The MOD3.3 calculation of the locations of porous debris regions in reactor
core was for the most part in agreement with the observed locations of porous debris. The calculated
temperature behavior of the porous debris was consistent with temperature behavior observed in debris
quenching experiments. The calculated and measured rapid increase in primary coolant system pressure
following the start-up of the 2B-pump were in good agreement. Except for an intermediate period of the
accident, the calculated primary coolant system pressure was in good agreement with the measured
pressure for all periods of the accident An adjustment of boundary conditions within their range of
uncertainty resulted in good agreement of calculated and measured pressure also during this intermediate
period. The calculated location of molten material in the core region and the timing of the slumping of this
molten material to the lower head were in general agreement with inferences from measurements and the
post-accident observation of the reactor core.

Appendix A12
The MOD3.3 calculations of the TMI-2 accident differed from the MOD3.2 calculations in some
aspects of behavior. MOD3.3 calculated damage progression in the reactor core to occur significantly
more rapid than MOD3.2. This difference is due to oxidation of fuel rod cladding in MOD3.3 being
calculated by the integral diffusion model instead of by the parabolic kinetics model in MOD3.2. MOD3.3
calculated the ballooning and rupture of fuel rod cladding to occur significantly earlier than MOD3.2. This
difference is due to corrections to the ballooning model implemented into MOD3.3. The stress-based
model in MOD3.3 for the failure of an oxide layer retaining melted cladding resulted in good agreement of
calculations with measurements for both the analysis of the TMI-2 accident and the analyses of severe fuel
damage experiments. On the other hand, MOD 3.2 underpredicted the extent of core melting by a factor of
two when using for the TMI-2 analysis the same oxide failure model as used for the analyses of severe fuel
damage experiments. In the modeling of phenomena causing damage to fuel assemblies during severe
accident conditions, MOD3.3 does not require a distinguishing of models for the analyses of severe fuel
damage experiments from the models for the analyses of nuclear power plants; one set of models applies
for both types of analyses, and all of the models used for nuclear power plant analyses have been assessed
using severe fuel damage experiments.
Al 2.5 References
A12-1. D. F. Guessing, "The Three Mile Island Analysis Exercise," Nuclear Technology, 87, August
1989, pp. 298-301.
A12-2. J. M. Broughton, P. Kuan, D. A. Petti, and E. L. Tolman, "A Scenario of the Three Mile Island
Unit 2 Accident," Nuclear Technology, 87, August 1989, pp. 34-53.
A12-3. J. R. Wolf et al., TMI-2 Vessel Investigation ProjectIntegrationReport, NUREG/CR-6197, TMI

V(93)EG10, EGG-2734, March 1994.
A12-4. D. A. Brownson, L. N. Haney, and N. D. Chien, Intentional Depressurization Accident

Management Strategy for Pressurized Water Reactors, NUREG/CR-5837, EGG-2688, April
1993.
A12-5. E. W. Coryell et al., SCDAP/RELAP5/MOD3.2 Code Manual Volume III: User's Guide and
Input Manual, NUREG/CR-6150, INEL-95/6422, Revision 1, November 1996.
A12-6. R. W. Brower L. J. Fackrell, D. W. Golden, M. L. Harris, and C. L. Olaveson, JCBC Version 3.1,
TMI-2 Initial and Boundary Conditions Data Base, GEND-INF-078, January 1988.
A12-7. Y. Nomura, PORV DischargeFlow Duringthe TMI-2 Accident, EGG-TMI-7825, July 1987.
A12-8. D. Coleman, "As-Built Design and Material Characteristics of the TMI-2 Core," PartIII in TMI
2 Accident Core HeatupAnalysis, A Supplement, NSAC-25, June 1981.
A12-9. J. L. Anderson, TMI-2 Once Through Steam GeneratorSecondary Level Analysis, EGG--TMI
7359, January 1987.

Appendix A12
A12-10. P. Dumaz, "Three Mile Island Unit 2 Analysis Exercise: CATHARE Computations of Phases 1
and 2 of the Accident," Nuclear Technology, 87, December 1989, pp. 946-955.
A12-11. J. L. Anderson, Recommended HPI Rates for the TMI-2 Analysis Exercise (0 - 300 Minutes),
EGG-TMI-7833, September 1987.
A12-12. D. W. Golden and N. Ohnishi, SCDAP/RELAP5 Demonstration Calculation of the TMI--2

Accident, EGG-TMI-8473, March 1989.
A12-13. F. E. Motley and R. P. Jenks, "Modeling of the Three Mile Island Unit 2 Accident with
MELPROG/TRAC and Calculation Results for Phases 1 and 2," Nuclear Technology, 87, August
1989, pp. 302-309.
A12-14. T. R. England and W. B. Wilson, TMI-2 Decay Power: LASL Fission-Productand Actinide
Decay Power Calculationsfor the President'sCommission on the Accident at Three Mile Island,
LA-8041-MS, Revised, March 1980.
A12-15. E. L. Tolman et al., TMI-2 Accident Scenario Update, EGG-TMI-7489, December 1986.
A12-16. J. Y. Hohorst et al., TMI-2 Analysis using SCDAP/RELAP5/MOD3.1, INEL-94/0157,

November 1994.
A12-17. A. Annunziato, A. Franceschini, and C. Addabbo, "TMI-2 Simulation by RELAP5/SCDAP and

COMETA Codes," 8th International Conference on Nuclear Engineering, April 2-6, Baltimore,
MD USA.

Appendix A13
A13. BROWNS FERRY BWR LOCA SEVERE ACCIDENT
The large break loss-of-coolant accident (LOCA) model for the Browns Ferry Nuclear Plant was
developed from a short term station blackout input deck developed by Oak Ridge National Laboratory
(ORNL). To model a LOCA, the short-term station blackout input deck was modified to include a break in
one of the two recirculation loops in the power plant. The ECCS was assumed to not be activated.
The conversion of the ORNL short-term station blackout input deck to a LOCA deck for use in the
current assessment was done to take advantage of model improvements and error corrections implemented
by ORNL.
A13.1 Nodalization
The nodalization diagram used for the Browns Ferry reactor vessel and associated piping is shown in
Figure A 13-1. The Browns Ferry reactor coolant system is modeled from the feedwater inlet to the turbine
inlet and includes the reactor pressure vessel, two recirculation loops, the feedwater piping, the control rod
drive cooling water, and the steam piping. The Browns Ferry containment is not explicitly modeled. Safety
relief valves are used to discharge coolant and release gases into a single large volume representing the
containment. This volume is maintained at a constant pressure of 0.31 MPa (45 psia) during the postulated
severe accident.
Figure A13-1. Nodalization diagram of Browns Ferry.

Appendix A13
Figure A 13-2 shows a detailed nodalization of the lower half of the reactor pressure vessel. The
hydrodynamic volumes are represented by open boxes and the solid structures represented by shaded
boxes in the diagram. The active core is divided into four radial rings and thirteen axial nodes. The center
most ring of the core represents approximately 55 percent of the fuel assemblies in the core, while each of
the other three rings represent approximately 15 percent of the fuel assemblies in the core. The power
density in the peripheral ring is greatly reduced from that in the central ring. The lower 1.07 m (3.5 feet of
the active core is divided into seven axial nodes, while the upper 2.74 m (9.0 ft.) of the active core is
divided into six axial nodes. The lower head of the Brown Ferry Nuclear Power Plant reactor vessel is
modeled with a COUPLE mesh.
Figure Al 3-2. Nodalization diagram of lower reactor vessel.
In each of the four radial rings, there is one pipe volume that represents the coolant flow inside the
fuel assemblies, and one pipe volume that represents the coolant flow outside the fuel assemblies in the
interstitial region surrounding the control blade. The interstitial volumes outside the fuel assemblies are
connected by crossflow junctions to the volumes surrounding the fuel so that coolant is able to flow
horizontally between the center and the periphery of the core. Only the crossflow junctions at the top of the
core are represented in the diagram, even though all internal volumes are connected by crossflow
junctions.
The primary coolant flow through the core, approximately 90 percent, is from the lower plenum
through the fuel support pieces and lower tie plates into the fuel assemblies. The remaining 10 percent of
the core coolant flow is through the interstitial region. Coolant enters the interstitial region by flowing
through holes machined in the lower tie plates, leaking past the core plate, or flowing through the control
rod guide tubes from the control rod drive cooling water pumps.

Appendix A 13
The jet pumps in the Browns Ferry reactor are represented by three volumes in each of the
recirculation loops. In a BWR, the upper mixing sections of the jet pumps are connected to the lower
diffuser sections by mechanical slip fits for easy removal during maintenance. The leakage through the slip
joints is represented by junctions that connect the jet pumps to the annulus region.
The SCDAP components representing the core structures are shown in Figure A13-3. Eight SCDAP
components were used to represent the core. Component 1 represents 432 fuel assemblies at the center of
the core; Component 2 represents the 108 control blades in the central core region; Component 3
represents the 104 fuel assemblies in ring 2; Component 4 represents the 26 control blades in ring 2;
Component 5 represents the 128 fuel assemblies in ring 3; Component 6 represents the associated 32
control blades in ring 3; Component 7 represent the 100 fuel assemblies in the peripheral ring; and
Component 8 represents the 19 control blades in the periphery of the reactor core.
Channel Box
Segment 2
Control Blade
Sheath
Segment 2
Rods
Segment I
Rods
Hot Rods
Channel Box
Segment 1
Figure A13-3. Component guide for detailed Browns Ferry SCDAP model.
The radial and axial power profiles used for the Browns Ferry accident simulation are shown in
Figure A13-4 and Figure A13-5. As shown in these figures, the peripheral fuel assemblies produce only
one third as much power as the center fuel assemblies. As shown in Figure A 13-5, the peak power in the
center fuel region is near the bottom of the core.

Appendix A13
1.3
1.2
1.1
1.0
0.9
w
UC 0.8
0.7
3c
0
a 0.6
0.5
,a
0.4
0.3
0.2
0.1
0.0
Ring 4 Ring 3 Ring 2 Ring 1
(periphery) (center)
Figure Al 3-4. Browns Ferry radial power profile for 4 radial rings.
1.6 . . . . . .. I . I . I . I . i . I . I . I . I I I
1.4
1.2
0
U
a 1.0
U
a
3 0.8 -- Ring 1 (center)
0
0
a 0.6 -A- Ring 2
--- Ring 3
0.4
-v- Ring 4 (periphery)
0.2
a
Sm I L
m I . I . I
0.0
0 1 2 3 4 5 6 7 8 9 10 11 12
Core Elevation (ft)
Figure A13-5. Browns Ferry power profiles for 13 axial nodes.

Appendix A 13
A13.2 Description of the Modeled BWR Severe Accident
The simulated LBLOCA was initiated with a break in one of the recirculating loops 300 seconds
after the start of the transient. The reactor was scrammed 0.5 seconds after initiation of the break.
Immediately after initiation of the accident the following events occurred. The water level in the annulus
surrounding the reactor core dropped from approximately 15 to 4 m within a short period of time, fuel
temperatures in the core dropped rapidly, and the pressure in the upper plenum of the reactor dropped from
7 MPa at the start of the accident to 0.40 MPa by the end of the accident. Immediately after the reactor
scram, the main steam isolation valve (MSIV) opened and a recirculation pump tripped on. The voiding of
the cooling water and the lowering of pressure in the reactor vessel triggered core uncovery. In the
modeled accident the core uncovered over a period of -25 seconds. Once a region in the core uncovered
temperatures began rising, initially quite fast and later at a gradual rate. The upper elevations of the core
heated faster and to higher temperatures than the lower elevations.
The containment for the Browns Ferry Nuclear Power plant was modeled as filled with saturated air
at a constant pressure of 0.1 MPa. Before the start of the accident, the initial system pressure was 7.03
MPa. The initial system temperature was 558 K and the power was 3,000 MW. The initial inlet flow to the
hot region in the reactor core was 2,900 kg/s.
A13.3 Results
This accident scenario, though not probable for a BWR, demonstrates that SCDAP/RELAP5/
MOD3.3 is able to predict heating, cladding oxidation and melt relocation of the fuel rods, heating,
oxidation and the subsequent relocation of the control blade and channel box into the lower head of the
reactor, and eutectic interactions that occur between various core materials.
SCDAP/RELAP5/MOD3.3 predicts the first damage progression event, the formation and rupture of
localized balloons in the fuel rod cladding, to occur 791 seconds into the simulated accident. The
temperature of the fuel rod cladding at time of rupture is 1120 K. The first ballooning incident is predicted
to occur in the center-most channel near the top (2.20 m) of the fuel rods in the reactor core. The
ballooning and rupturing of fuel rods near the 2.20-m elevation of the modeled core is predicted to proceed
radially outward through approximately 85% of the reactor core over a period of about 174 seconds as fuel
rod cladding temperatures approach 1,200 K in each channel. Ballooning and rupture of fuel rods at the
periphery of the core occur approximately 640 seconds later. The power produced in this peripheral region
is approximately one third less than the power produced in the central region of the core, resulting in
slower heating rates and the predicted delayed ballooning and rupture of the fuel rods.
As the reactor core continues to heat during the course of the simulated accident, the control blades
begin melting as the zircaloy channel boxes oxidize. Eutectic interactions occur between the zircaloy of
the channel box and the control blade stainless steel and boron carbide resulting in the melting and
relocation of the control blades and finally the channel box. Figure A13-6 shows the predicted midplane
radial temperature distribution across the reactor core of the stainless steel clad boron carbide control
blades, and Figure A 13-7 the predicted midplane radial temperature distribution across the reactor core of
the channel boxes in the reactor core. As the temperature of the control blade in each channel reached
1,500 K, the blade melted and relocated into the lower head of the reactor. The relocation time for each
blade channel is represented in Figure A13-6 as the transition from rapidly rising temperatures to a
constant temperature of 1,500 K blade failure temperature. The control blade in the center-most channel
begins relocating into the lower head of the reactor approximately 1,139 seconds into the simulated
accident, when the blade temperature in the channel reaches 1,500 K. The blade in Channel 2 relocated

Appendix A 13
300 seconds later. Approximately 500 seconds after the blade in Channel 2 melted and relocated to the
lower head, the blade in Channel 3 began to melt and relocate into the lower head of the reactor vessel. As
shown in the figure the blade in the low power peripheral channel heated at a considerably slower rate and
never completely relocated during the simulated LBLOCA. Once the blade attains 1,500 K at a specified
elevation, the control material at that elevation relocates quickly into the lower head of the reactor. During
the calculation the predicted relocation into the lower head is shown by increasing height of the debris bed.
As shown in Figure A 13-7, the temperature histories of the channel boxes in each radial ring of the
reactor show a behavior similar to the control blades. The disappearance of the channel box at midplane in
the reactor core is shown by the transition to a constant temperature of 2,250 K, representative of the
melting temperature of the zircaloy channel box material. Temperatures in the peripheral channel tend to
be cooler due to the lower developed power, therefore, very little melting and relocation of the control
blade and channel box material occurs.
2000.0
1500.0 F
I
Cu
1000.0 F
I
a.
500.0[ - Center channel

- Channel 2
- Channel 3
Peripheral channel -----.
0.%oL- 1000.0 2000.0 3000.0

Time (s)
Figure Al 3-6. Predicted radial temperature distribution across reactor core of control blades.
Figure A13-8 to Figure A13-10 show predicted radial temperature distributions across the reactor
core of fuel rods at three elevations, 0.5334, 0.9906, and 2.2210 m. As shown in the figures, the predicted
radial temperature distribution follows the heating pattern established for the control blade and channel
box. The temperatures in the inner two channels exceed 2,830 K, the temperature where the transition from
intact geometry to molten material occurs, at the three elevations shown in the figures. The figures also
show that rapid temperature changes in the two outermost channels are in response to the arrival of hot
relocating material from higher elevations in the channel. The arrival of relocating material is indicated at
a location by the sharp changes in the temperature of the fuel rod cladding.

Appendix A13
3000.0
- Center channel
- Channel 2
2500.0 - Channel 3
-------Peripheral channel
,-, 2000.0
S1500.0 t
0..
500.0
0.0. 1000.0 2000.0 3000.0

Time (s)
Figure A13-7. Predicted radial temperature distribution across reactor core of channel boxes.
3000.0 1 1
- Center channel
- Channel 2
o-O Channel 3
0-0 Peripheral channel
,- 2000.0
0..0 1000.0 2000.0 3000.0

Time (s)
Figure Al3-8. Predicted radial temperature distribution across reactor core of fuel rods at the 0.5334 m
elevation.

Appendix A 13
3000.0
,-,2000.0
I-I
S1000.0
0.0
0.0 1000.0 2000.0 3000.0

Time (s)
Figure Al 3-9. Predicted radial temperature distribution across reactor core of fuel rods at the 0.9906 m
elevation.
3000.0
,- 2000.0
2
F-, 1000.0
0.00.0 1000.0 2000.0 3000.0

Time (s)
Figure Al 3-10. Predicted radial temperature distribution across reactor core of fuel rods at the 2.2210 m
elevation.

Appendix A 13
The calculated system pressure is shown in Figure A13-11. Immediately after reactor scram
(approximately 300 seconds into the transient), the system quickly depressurized to approximately 0.4
MPa. The depressurization coincided with the uncovery phase of the accident. Once the core uncovered,
depress urization terminated and all elevations of the core entered a period of rapid heating. The period of
zircaloy oxidation for all elevations indicated that the core had completely dried out. Fuel rod heating was
mostly due to nuclear heat generation, as shown by the considerably slower heating rates in the peripheral
regions of the core at all elevations.
8.0
6.0
24.0
2.0
0.0 1000.0 2000.0 3000.0

Time (s)
Figure A13-11. Predicted Browns Ferry system pressure.
The initial formation of molten material occurs near the upper middle of the reactor core and
progresses upward and downward with time. The initial hot spot for each modeled component occurred
near the axial center of the fuel assembly. The first area to begin the transition from intact geometry to
molten material was in the highest power central channel.
Figure A 13-12 shows the predicted hydrogen production during the simulation of the LBLOCA. As
shown in this figure, approximately 350 kg of hydrogen were produced during this simulated severe
accident.
In summary, the Brown's Ferry BWR LOCA Severe Accident test problem demonstrated the
capability to calculate the behavior of a BWR during a severe accident. Damage progression in the BWR
was analyzed from its beginning out to the time of slumping of most of the reactor core to the lower head
of the reactor vessel. All of the major damage progression that can occur in a BWR reactor core during a
severe accident was calculated to occur during the analyzed accident sequence. The calculated damage
progression included oxidation and meltdown of fuel rods, control blades, and channel boxes. The late
phase damage progression was calculated to occur more rapidly than is the case for a TMLB' type of a
severe accident in a PWR. The more rapid progression of damage was due to the rapid depressurization of
the reactor core resulting from the imposed large break in the primary coolant system piping. The
calculated damage progression also differed from that in a PWR in that damage to the bottom of the reactor
core was calculated to occur almost as rapidly as damage to the top of the core.

Appendix A 13
400.0
(U
o 300.0
C
S200.0
0
7
S100.0
U
0.0,
0.0 1000.0 2000.0 3000.0
Time (s)
Figure Al 3-12. Predicted integral hydrogen production.

Appendix A14
A14. SURRY PWR SEVERE ACCIDENT
A14.1 Introduction
The analysis of a severe accident in the Surry nuclear power plant was performed to exercise and
evaluate damage progression models in SCDAP/RELAP5/MOD3.3 under conditions typical of those
expected in a typical plant severe accident analysis. The information in this section is arranged as follows:
The design of the Surry nuclear power plant and the SCDAP/RELAP5/MOD3.3 nodalization of the plant
are described in Section A14.2. The severe accident sequence that was selected for analysis is described in
Section A14.3 and the calculation results are presented in Section A14.4.
Al 4.2 Model Description and Nodalization
The Surry nuclear power plant is a typical Westinghouse three-loop pressurized water reactor (PWR)
with a rated thermal power of 2,441 MWt. The core consists of 157 15 x 15 fuel assemblies with an active
fuel height of 3.66 m. Each of the three primary coolant loops contains a U-tube steam generator, a reactor
coolant pump (RCP), and associated piping. A single pressurizer is attached to the hot leg piping of one of
the three primary coolant loops. The pressurizer surge line and the hot/cold leg piping are constructed of
stainless steel. The steam generator tubes are made of Inconel. The power operated relief valves (PORV) is
located at the top of the pressurizer and can be used to relieve excess pressure in Reactor Coolant System
(RCS). Pressurizer Safety Relief Valves (SRV) are also available to handle pressure excursions in excess
of the pressurizer PORV capacity. An accumulator, containing 29,100 kg of borated water at 322 K, is
attached to each cold leg. The accumulators, which are initially pressurized to 4.24 MPa with a nitrogen
cover gas, are the only operational part of the emergency core cooling system (ECCS) during a station
blackout. A large, dry, subatmospheric pressure containment building surrounds the reactor system.
The RELAP5 nodalization of the reactor core and vessel is shown in Figure A 14-1. In this figure the
reactor core is represented using 5 radial segments (flow channels) and ten axial nodes. One assessment
case was performed using 5 flow channels and 20 axial nodes in the reactor core as part of a sensitivity
study. Each flow channel is connected to its neighboring flow channels by crossflow junctions, thus flow
in the lateral direction through the core is modeled. Separate RELAP5 control volumes are used to
represent the core bypass and the downcomer. The core bypass is represented by a RELAP5 control
volume (pipe) containing five sub-volumes and the downcomer is represented by a RELAP5 control
volume (pipe) with seven sub-volumes. Two RELAP5 control volumes are used to represent the lower
plenum and seventeen RELAP5 control volumes to represent the upper plenum region of the reactor
vessel.
The RELAP5 nodalization of the coolant loop with the pressurizer connected to the hot leg is shown
in Figure A14-2. The non-pressurizer loops were modeled with nodalizations similar to that shown in
Figure A14-2. The hot leg is represented by five RELAP5 control volumes, the cold leg by four RELAP5
control volumes, the surge line by three RELAP5 control volumes, and the pressurizer by eight RELAP5
control volumes. The primary side of the steam generator is represented by eight RELAP5 control volumes
and the secondary side by twenty RELAP5 control volumes. The loop seal is represented by five RELAP5
control volumes. The three coolant loops in the reactor are represented by a total of 140 RELAP5 control
volumes. At the time that core uncovery begins, the hot leg and steam generator are renodalized to allow
single-phase countercurrent flow in the hot leg.
A radial cross-section of the nodalization of the reactor core is shown in Figure A14-3. The first

Appendix A14
Figure A14-1. RELAP5 nodalization of the Surry reactor core and vessel for 5 x 10 cases.
radial segment represents the 5 fuel assemblies at the center of the core. The second radial segment begins
at the outer boundary of the first radial segment and extends to the inner boundary of the third segment.
This segment represents 20 fuel assemblies. The third and fourth radial segments represent fuel assemblies
progressing radially outward from the second radial segment and the fifth radial segment represents the
fuel assemblies at the periphery of the core. The third segment represents 36 fuel assemblies, the fourth
represents 60 fuel assemblies, and the fifth represents 36 fuel assemblies. Each of the radial segments

Appendix A14
Figure Al 4-2. RELAP5 nodalization of the Surry primary system loop

corresponds to a flow channel represented by one of the RELAP5 pipe volumes for the reactor core. The
numbering for the pipe volumes is as follows: radial segment 1 is 111, segment 2 is 112, segment 3 is 113,
segment 4 is 114, and segment 5 is 115.
The fuel rods in each fuel assembly were represented by a SCDAP fuel rod component and the
control rods were represented by a SCDAP PWR control rod component. Each individual fuel assembly
was represented as having 204 fuel rods and 21 control rods. Each fuel rod and control rod was divided
into ten axial nodes. The water rods in the outermost radial segment were represented by a SCDAP control
rod component model containing an extremely small quantity of control material and the empty zircaloy
guide tubes in the center ring were represented by a fuel rod component model with an extremely small
fuel pellet. The fuel rods have a fuel pellet stack that is 3.6576 m in length. The fuel pellets in each fuel rod
have a radius of 4.634 umn. The outer radius of the fuel rod cladding was 5.359 mm with a cladding
thickness of 0.617 mm. The reactor core was represented by a total of ten SCDAP components; two fuel

Appendix A14
First (center) flow channel:

CL AI
5 fuel assemblies
oI
relative power = 1.19
Second flow channel:

20 fuel assemblies
Hs1H1-)7
IR1E
HH HiHL9
Third flow channel:
36 fuel assemblies
1.711D IF0 l6HIF9
Fourth flow channel:
60 fuel assemblies
Fifth (outer) flow channel:
36 fuel assemblies
Figure Al 4-3. Radial cross-section of Surry reactor vessel showing nodalization of the core.
rod component models in the center radial segment and a fuel rod and control rod component in each of the
four other radial segments. The core of the reactor contained six Inconel spacer grids located at the
following elevations, 0.0 m, 0.73 m, 1.46 m, 1.83 m, 2.19 m, and 2.93 m, respectively. These spacer grids
were represented using the SCDAP spacer grid model.
The severe accident analyzed was a station blackout without operator actions. This type of accident
is designated at a TMLB' sequence. The accident is initiated by loss of off site power. Onsite ac power is
also unavailable because the diesel generators fail to start or fail to supply power. Decay heat removal
through the steam generators cannot be maintained in the long term because there is no ac power for the
electrical pumps, and the steam driven auxiliary feedwater pumps also fail to supply water. As the RCS
depressurizes during the accident, the RCP loop seals may clear.
When natural circulation began in the reactor vessel and the hot leg began draining, a renodalization
of the hot leg and steam generators was performed. Figure A14-4 and Figure A14-5 show the ex-vessel
nodalization used for Surry analyses after the hot leg drained. The new nodalization provided the
additional flow paths needed to simulate hot leg countercurrent natural circulation, which develops only
after the hot legs are voided. A horizontal section of the modeled surge line in the Surry PWR connects to
the centerline of the hot leg piping. This orientation required the modeling of a split surge line to represent
potential countercurrent flow in a manner similar to that used for the rest of the loop piping. The associated
valves (component number 463 and 465 in Figure A14-5) open whenever the pressurizer PORV opens so
that fluid from both halves of the hot leg will be drawn into both halves of the surge line during valve
cycles.

Appendix A14
Figure A14-4. Nodalization for modeling natural circulation in hot leg and steam generator.
The nodalization changes shown in Figure A14-4 were needed to accommodate both full-loop and
hot leg countercurrent natural circulation. These changes were accomplished by using servo valves to
connect (a) the split hot leg to the split steam generator inlet plenum, (b) the split steam generator inlet
plenum to the split steam generator tube bundle, and (c) the split steam generator tube bundle to the steam
generator outlet plenum. The valves were configured with two sets of loss coefficients, one set consistent
with those used to model normal plant operation and one set appropriate to model hot leg countercurrent
natural circulation. The other loss coefficients in the loop network were unaltered relative to those used for
normal plant operation. These loss coefficients were equivalent to those used prior to the draining of the
hot leg. If the horizontal portion of the RCP loop seal in a given loop is full of liquid, control logic directs
the code to use the countercurrent flow loss coefficients for all servo valves. in the loop. Alternately, if
voids begin to form in the horizontal portion of the RCP loop seal, control logic directs the code to use the
loss coefficient values for normal plant operation for the servo valves in the loop. As a result of the use of
A14-5 NLJREG/CR-6150-Rev. 2, Vol. 5

Appendix A14
of split hot leg
Figure A14-5. Arrangement of junction and valves for modeling natural circulation.
this modeling logic, both full-loop and countercurrent natural circulation flow should be calculated
correctly in each loop as the code responds to changing loop conditions during the accident.
The Surry PWR model used a simple two volume representation for the reactor containment
(Volumes 449-1 and 449-2 in Figure A14-2 and Figure A14-4). This nodalization was not expected to
adequately calculate condensation of the primary effluent. Without a reasonable prediction of primary
effluent condensation, a reasonable prediction of containment pressure cannot be expected. To resolve this
problem, the pressure in the reactor containment was not allowed to rise higher than 0.2 MPa. Input was
included in the model to calculated creep rupture of the stainless steel surge line (Volume 455-2 in Figure
A14-4), the three stainless steel hot legs (Volume 400-1 in Figure A14-4 and Volumes 200-1 and 300-1 in
areas of the non-pressurizer loops which are not shown in the nodalization diagrams), the hottest of the
Inconel steam generator tubes (Volume 408-1 in Figure A 14-4), and Volumes 208-1 and 308-1 in areas of
the non-pressurizer loops which are not shown in the nodalization diagrams).
A14.3 Boundary Conditions for TMLB' Severe Accident

A Surry TMLB' with a failed steam generator (SG) relief valve and intentional RCS depressurization
was analyzed. The accident sequence was similar to that performed for the resolution of the issue of SG
tube integrity.Al4"l Reference A.14-1 provides details on the severe accident sequence analyzed, denoted
Case 20. The sequence is basically a postulated station blackout transient in Surry with concurrent failure
of the diesel generators and steam driven auxiliary feedwater pumps. Additionally, the pressurizer loop SG
secondary side atmospheric dump valve (ADV) is assumed to fail open on the first challenge. Also, the
operator is assumed to latch open the pressurizer PORV when the core exit steam temperature reaches 922
K. The sequence produces RCS depressurization by way of the latched open PORV, resulting in
accumulator injection, core melting and relocation to the lower head with subsequent lower head failure.
RCS piping was monitored for creep rupture failure but a break in the ruptured pipe was not imposed.
The MOD3.3 and MOD3.2 calculations of the timing of creep rupture of the surge line were similar.
The calculated primary coolant system thermal hydraulic behavior and the overall damage progression for
the two code versions were also similar. Comparisons of the calculated timing of events in the primary
coolant system and in damage progression are shown in Table A14-1. The calculated timing of creep

Appendix A14
rupture of the surge line for MOD3.3 and MOD3.2 were 24,205 s and 22,515 s, respectively.
Table A14-1. Comparison of SCDAP/RELAP5/MOD3.3 and SCDAP/RELAP5/MOD3.2 calculations of
timing of events in primary coolant system and in damage progression for Surry TMLB' with failed SG
ADV and intentional RCS depressurization.
Time (s) for

Event
MOD3.3 MOD3.2
TMLB' initiation 0 0
Collapsed liquid level falls below top of fuel rods 8,916 8,915
Collapsed liquid level falls below bottom of fuel rods 10,600 10,618
Onset of fuel rod oxidation 11,590 11,045
Pressurizer surge line fails by creep rupture 24,205 22,515
Accumulators/core flood tank empty 25,000 25,865
The MOD3.3 coolant condition histories for the reactor core and primary coolant system are shown
in Figures 14-6, 14-7, and 14-8, respectively. The primary system pressure response, reactor vessel
collapsed liquid level, and the liquid inventory for the pressurizer loop accumulator (nonpressurizer loops
accumulators are shown. The calculated thermal hydraulic response is similar to that calculated by
MOD3.2.
20.0
15.0
10.0
(D
cc
$D
CL
5.0
0.0 L.
0.0 10000.0 20000.0 30000.0
Time (s)
Figure A14-6. MOD3.3 calculation of pressurizer steam dome pressure for the Surry TMLB' with failed
SG ADV and intentional depressurization.

Appendix A14
15.0
-i 10.0
0
(D
5.0
0
0.0
0.0 10000.0 20000.0 30000.0
Time (s)
Figure A14-7. MOD3.3 calculation of reactor vessel collapsed liquid level for the Surry TMLB' with
failed SG ADV and intentional depressurization.
30.0
E
E 20.0
0~
75 10.0
E
0
0
0.0
0.0 10000.0 20000.0 30000.0
Time (s)
Figure A14-8. MOD3.3 calculation of pressurizer loop accumulator liquid inventory for the Surry
TMLB' with failed SG ADV and intentional depressurization.

Appendix A14
MOD3.3 calculated that accumulator injection delayed the onset of rapid oxidation of the reactor
core. A plot of core maximum temperature is shown in Figure 14-9. A plot of the cumulative hydrogen
production is shown in Figure 14-10. The cumulative hydrogen production was calculated to be 152 kg at
the time of surge line failure. A plot of the temperature history of the surge line is shown in Figure A 14-11.
3000.0
2000.0
E
16:
0
0
O
E 1000.0
Co
E
0
0
0
0.0 L
0.0 10000.0 20000.0 30000.0
Time (s)
Figure A14-9. MOD3.3 calculation of core maximum peak cladding temperature for the Surry TMLB'
with failed SG ADV and intentional depressurization.

Appendix A14
200.0
ia
150.0
'0
"0
C0
100.0
0
76
Z
50.0
0.0
0.0 10000.0 20000.0 30000.0
Time (s)
Figure A14-10. MOD3.3 calculation of total hydrogen production for the Surry TMLB' with failed SG
ADV and intentional depressurization.
2000.0
1500.0
0.
E
0) 1000.0
"CO 500.0
0.0
0.0 10000.0 20000.0 30000.0
Time (s)
Figure A14-11. MOD3.3 calculation of surge line nozzle volume average temperature for the Surry
TMLB' with failed SG ADV and intentional depressurization.

Appendix A14
Al 4.4 References
A14-1. D. L. Knudson and C. A. Dobbe, SCDAP/RELAP5 Evaluation of the Potential for Steam
Generator Tube Rupture when the Reactor Coolant System is Depressurized, DLK-4-97, INEEL,
May 23, 1997.

NRC FORM 335 U.S. NUCLEAR REGULATORY COMMISSION 1. REPORT NUMBER
(2-89) (Assigned by NRC, Add Vol.,
NRCM 1102. BIBLIOGRAPHIC DATA SHEET Supp., Rev., and Addendum
320)1.3202 (See Instructions on the reverse) Numbers, if any.)
NUREG/CR-6150, Rev. 2
Volume 5
INEL-96/0422
2. TITLE AND SUBTITLE 3. DATE REPORT PUBLISHED
SCDAP/RELAP5/MOD3.3 Code Manual, Volume 5: Assessment of Modeling of MONTH YEAR
January 2001
4. FIN OR GRANT NUMBER
W6095
5. AUTHOR(S) 6. TYPE OF REPORT
Technical
L. J. Siefken, E. W. Coryell, E. A. Harvego, J. K. Hohorst 7. PERIOD COVERED (nclusi,, Datr•)
8. PERFORMING ORGANIZAITON - NAME AND ADDRESS (If NRC, provide Division, Office or Region, U.S. Nuclear Regulatory Commission,
and mailing address; if contractor, provide name and mailing address.)

P.O. Box 1625
Idaho Falls, ID 83415-3129
9. SPONSORING ORGANIZATION - NAME AND ADDRESS (If NRC, type "Same as above"; If contractor, provide NRC Division, Office or Region,
U.S. Nuclear Regulatory Commission, and mailing address.)
Division of Systems Technology
10. SUPPLEMENTARY NOTES

S. A. Arndt, NRC Project Manager
11. ABSTRACT (200 words or less)
The SCDAP/RELAP5 code has been developed for best-estimate transient simulation of light water reactor coolant systems
during a severe accident. The code models the coupled behavior of the reactor coolant system and reactor core during severe
accidents as well as large and small break loss-of-coolant accidents, operational transients such as anticipated transient
without SCRAM, loss of offsite power, loss of feedwater,.and loss of flow.
Volume 5 describes the assessment of the MOD3.3 version of SCDAP/RELAP5, which showed that its new models for
severe fuel damage and thermal hydraulic behavior result in calculated behavior of fuel assemblies under severe accident
conditions in agreement with measurements. The new models were assessed using a wide range of fuel damage experiments
and the TMI-2 accident, and the assessment showed that the SCDAP/RELAP5 calculations of the axial distribution in
oxidation and meltdown of fuel assemblies and the behavior of the fuel assemblies under reflood conditions were improved
by the new models in MOD3.3. The assessment showed that MOD3.3 calculates a more rapid onset of severe fuel damage
under severe accident conditions than that calculated by the previous version of the code. The assessment also showed that
MOD3.3 can analyze both severe fuel damage experiments and nuclear power plants with a single set of models for the
phenomena causing damage to fuel assemblies.
12. KEY WORDS/DESCRIPTORS (List words or phrases that will assist researchers in locating the report.) 13. AVAILABILITY STATEMENT
SCDAP/RELAP5; severe fuel damage; thermal hydraulics; LWR; coolant system; reactor core Unlimited
14. SECURITY CLASSIFICATION
(This page)
Unclassified
(This report)
Unclassified
15. NUMBER OF PAGES
16. PRICE
NRC FORM 335 (2-89)

Federal Recycling Program
UNITED STATES
NUCLEAR REGULATORY COMMISSION
WASHINGTON, DC 20555-0001
OFFICIAL BUSINESS
PENALTY FOR PRIVATE USE, $300

ML010310397

Uploaded by

Copyright:

Available Formats

ML010310397

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

ML010310397

Uploaded by

Copyright:

Available Formats

NUREG/CR-6150, Vol. 5, Rev.

Idaho National Engineering and Environmental Laboratory

U.S. Nuclear Regulatory Commission

A single copy of each NRC draft report for comment is

Idaho National Engineering and Environmental Laboratory

S. A. Arndt, NRC Project Manager

iii NUREG/CR-6150-Rev 2, Vol 5

FIGURES .................................................................................................................................. vii

EXECUTIVE SUMMARY ......................................................................................................... xi

1. INTRODUCTION .................................................................................................... 1-1

V NUREG/CR-6150- Rev 2, Vol 5

vii NUREG/CR-6150-Rev. 2, Vol. 5

a. N. E. Bixler, "VICTORIA2.0: A Mechanistic model for Radionuclide Behavior in a Nuclear Reactor

NUREG/CR-6150-Rev. 2, Vol. 5 xii

The SCDAP/RELAP5 Program is indebted to J. Schaperow and S. Arndt of the U. S. Nuclear

.°.i NUREG/CR-6150-Rev. 2, Vol. 5

1-1. L. J. Siefken, E. W. Coryell, E. A. Harvego and J. K. Hohorst, SCDAP/RELAP5/MOD3.3 Code

1-2. The SCDAP/RELAP5 Development Team, SCDAP/RELAP5/MOD3.2 Code Manual, Volume V.

1-1 NUREG/CR-6150-Rev. 2, Vol. 5

2. Summary of New Models in MOD3.3 and Test Matrix for Assessment

2-1 NUREG/CR-6150-Rev. 2, Vol 5

Table 2-1. New models in MOD3.3 of SCDAP/RELAP5.

Name Extensions in capability provided by model

NUREG/CR-6150-Rev. 2, Vol. 5 2-2

Table 2-1. New models in MOD3.3 of SCDAP/RELAP5. (Continued)

Name Extensions in capability provided by model

Table 2-2. Test matrix for developmental assessment of SCDAP/RELAP5/MOD3.3.

Model or pbf11 cora5 pdropl sur

2-3 NUTREG/CR-6150-Rev. 2, Vol 5

Table 2-2. Test matrix for developmental assessment of SCDAP/RELAP5/MOD3.3. (Continued)

Model or pbfl 1 cora5 pdropl

Table 2-3. Definition of test problems listed in Table 2-3.

NUREG/CR-6150-Rev. 2, Vol. 5 2-4

Table 2-3. Definition of test problems listed in Table 2-3. (Continued)

2-5 NUREG/CR-6150-Rev. 2, Vol 5

Table 2-3. Definition of test problems listed in Table 2-3. (Continued)

2-1. The SCDAP/RELAP5 Development Team, "SCDP/RELAP5/MOD3.2 Code Manual - Damage

2-2. The SCDAP/RELAP5 Development Team, "SCDP/RELAP5/MOD3.2 Code Manual

NUREG/CR-6150-Rev. 2, Vol. 5 2-6

2-8. L. J. Siefken, "SCDAP/RELAP5 Modeling of Movement of Melted Material Through Porous

2-7 NUREG/CR-6150-Rev. 2, Vol 5

3. Summary of Assessment Using Results of Severe Fuel Damage

3.1 Assessment of Models for Temperature Behavior of Fuel

3.1.1 Temperature Behavior of PWR Fuel Assemblies with Gradual Uncovery

3-1 NUREG/CR-6150-Rev. 2, Vol. 5

3.1.2 Temperature Behavior of PWR Fuel Assemblies with Complete Uncovery

NUREG/CR-6150-Rev. 2, Vol. 5 3-2

3-3 NUREG/CR-6150-Rev. 2, Vol. 5

NUREG/CR-6150-Rev. 2, Vol. 5 3-4

3-5 NUREG/CR-6150-Rev. 2, Vol. 5

NUREGICR-6150-Rev. 2, Vol. 5 3-6

3.1.3 Temperature Behavior of BWR Fuel Assemblies with Complete Uncovery

3-7 NUREG/CR-6150-Rev. 2, Vol. 5

3.1.4 Temperature Behavior of Fuel Assemblies During Reflood

.0 5000.0 10000.0 15000.0

NUREG/CR-6150-Rev. 2, Vol. 5 3-8

3.2 Assessment of Calculation of Hydrogen Production

Hydrogen production (g)

Measured MOD3.3 MOD3.2