...

ASSESSMENT OF SIZE ASPECTS IN MODELLING INTERACTION MOLTEN FUEL

by user

on
Category: Documents
1

views

Report

Comments

Transcript

ASSESSMENT OF SIZE ASPECTS IN MODELLING INTERACTION MOLTEN FUEL
ASSESSMENT OF SIZE ASPECTS IN MODELLING MOLTEN
FUEL COOLANT INTERACTION
melt
jet
drops
debris + cake
Student:
PhD Director:
August 2003
Patricia Pla Freixa
Francesc Reventós Puigjaner
Assessment of Size Aspects in Modelling Molten Fuel Coolant Interaction
ACKNOWLEDGEMENTS
ACKNOWLEDGEMENTS
After all the years of research at the Joint Research Centre (JRC) of Ispra and at the Technical University of
Catalonia (UPC), I would like to thank many people who contributed to the research and helped finalising
this work. In particular, I would like to thank to:
Dr. Francesc Reventós, for being my supervisor and for his assistance, support, kindness and patient despite
my absence. To all colleagues at the University for their company during my stays in Barcelona.
The JRC FARO team for welcoming me to their project and let me participate in the real laboratory
experiences. Especially to Alessandro Annunziato, who has been my trainer and guide in the knowledge of
MFCI phenomena; for his help during all these years, support, kindness, sincere interest and significant
discussions and revisions to the manuscript. “Grazie mille Alessandro!!”
To all JRC FARO team: Carmelo Addabbo, Daniel Magallon, Giuseppe Leva, Werner Brewka, Rocco
Silverii, Arnold Yerkess, Ilpo Huhtiniemi, Horst Weisshaeupl, Arturo Romor, Massimo Anselmi, Giovanni
Sciamanna, Jean Ezan, George Nicol, Olivier Hubert, Colin Pantry, Antonio Leppeni, Alfredo Nalesso, Gian
Piero Marzi, Zdzislaw Dzbikowicz, Gianpaolo Fraguglia, Raoul Kiefer, L. Bonomi, A. Renoldi, Valerio
Ragazzoni, Arturo Signorelli for their enthusiasm and for engaging me in the FARO experiments activities,
Karin Faber, Sara Brusamolin, Annie Mignot for their kindness and help, especially to Carola Piatti and
Monica Leva for their friendship, company and free coffees.
To Chris Allison and Dick Wagner for their valuable help and constructive discussions about
RELAP/SCDAP code use and understanding.
My parents, who always fully supported my work despite my stay abroad. All my Spanish friends living in
Spain and around Europe, who supported my decision, and came to visit me in some occasions. And all
friends and colleagues in and around Ispra and Milano, who created a pleasant atmosphere and social
environment during my stay.
Assessment of Size Aspects in Modelling Molten Fuel Coolant Interaction
ABSTRACT
ABSTRACT
Severe accidents in light water nuclear reactors occur when reactor vessel water inventory decreases and
there is no available additional water coolant to be delivered into the core. In general, during an extended
severe accident sequence a period exists in which the reactor core, after a partial or total melt down, is
poured into the lower plenum that can have some water present. The study of the interaction of the melt fuel
with the water is the objective of MFCI (Melt Fuel Coolant Interaction) activities.
MFCI is one of the most important issues awaiting resolution in water cooled reactor safety analysis. The
progression of a severe accident in a water cooled reactor can lead to energetic (steam explosion) or nonenergetic (melt quenching) interactions as the molten fuel relocates and eventually interacts with the coolant
either in the vessel lower head (in vessel) or in the cavity (ex-vessel).
The MFCI experiments at JRC Ispra site were conducted in the FARO (Furnace And Release Oven) test
facility under realistic melt composition and prototypical accident conditions to provide basic information on
underlying phenomena. The experimental programme was complemented by comprehensive pre-test and
post-test analytical activities based on the development and application of the thermalhydraulic COMETA
(COre MElt Thermalhydraulic Analysis) code. The code is developed and assessed on the basis of
experimental information acquired in the FARO facility tests, and there are some limitations and
uncertainties in their application to the full plant, which need to be identified and possibly quantified.
In general the main objective of the PhD research was achieved expanding the general knowledge in Melt
Fuel Coolant Interaction. The knowledge was complemented collaborating and complementing the
application of COMETA code under conditions not experimented before, developing and improving
COMETA code sources and verifying the code consistency, analysing and unifying the COMETA
simulations carried so far.
Also a further analytical study was carried out in order to illustrate the MFCI inside the general overview of
a NPP (Nuclear Power Plant) severe accident sequence.
Assessment of Size Aspects in Modelling Molten Fuel Coolant Interaction
CONTENTS
CONTENTS
1.
2.
3.
INTRODUCTION .....................................................................................................................................................1
MELT FUEL COOLANT INTERACTION AND QUENCHING ............................................................................3
STATE OF THE ART ...............................................................................................................................................5
3.1.
SEVERE ACCIDENT UNDERSTANDING ..................................................................................................5
3.2.
MELT FUEL COOLANT INTERACTION IN LOWER WATER REACTOR SEVERE ACCIDENT
PHENOMENOLOGY CONTEXT.................................................................................................................................5
3.3.
WORLD WIDE SEVERE ACCIDENT RESEARCH ....................................................................................9
3.3.1.
EXPERIMENTAL FACILITIES................................................................................................................9
3.3.2.
ANALYTICAL TOOLS ...........................................................................................................................13
3.3.3.
PRESERVATION OF SEVERE ACCIDENT DATA .............................................................................16
4. OBJECTIVES ..........................................................................................................................................................19
5. THE FARO TEST FACILITY.................................................................................................................................21
5.1.
INTRODUCTION .........................................................................................................................................21
5.2.
FARO FACILITY COMPONENTS..............................................................................................................23
5.3.
SIMULATION OF ACCIDENT PHENOMENA IN THE FARO FACILITY.............................................24
6. THE COMETA CODE ............................................................................................................................................27
6.1.
INTRODUCTION .........................................................................................................................................27
6.2.
THERMALHYDRAULIC MODEL..............................................................................................................27
6.3.
MELT FUEL FRAGMENTATION MODEL ...............................................................................................32
7. RESEARCH ACTIVITIES SUMMARY ................................................................................................................39
7.1.
COMETA PRE-TEST CALCULATION OF FARO TEST L-29 .................................................................40
7.1.1.
INTRODUCTION ....................................................................................................................................40
7.1.2.
RESULTS OF SUBCOOLED TEST OPTION FOR THE FARO TEST L-29 ........................................41
7.1.3.
RESULTS OF SATURATED TEST OPTION FOR THE FARO TEST L-29 ........................................46
7.1.4.
ADDITIONAL CALCULATIONS PERFORMED FOR THE SUBCOOLED CASE ............................47
7.1.5.
CONCLUSIONS.......................................................................................................................................50
7.2.
COMETA PRE-TEST CALCULATION OF FARO TEST L-33 .................................................................51
7.2.1.
INTRODUCTION ....................................................................................................................................51
7.2.2.
RESULTS OF TEST OPTION FOR THE FARO TEST L-33 BEFORE TRIGGERING .......................52
7.2.3.
TRIGGER CALCULATION TESTS WITH WATER.............................................................................61
7.2.4.
CALCULATION OF THE TRIGGERING EVENT FOR THE FARO TEST L-33 ................................64
7.2.5.
CONCLUSIONS.......................................................................................................................................67
7.3.
ANALYSIS OF FUEL-COOLANT QUENCHING PHENOMENA BY COMETA CODE IN REACTOR
GEOMETRY ................................................................................................................................................................68
7.3.1.
INTRODUCTION ....................................................................................................................................68
7.3.2.
INITIAL AND BOUNDARY CONDITIONS. ANALYSED CASES.....................................................68
7.3.3.
CALCULATION RESULTS ....................................................................................................................73
7.3.4.
CONCLUSIONS.......................................................................................................................................91
7.4.
COMETA CODE CALCULATIONS OF THE FARO QUENCHING TESTS............................................92
7.4.1.
INTRODUCTION ....................................................................................................................................92
7.4.2.
EXPERIMENTS IN THE FARO FACILITY...........................................................................................92
7.4.3.
COMETA CODE SIMULATIONS OF THE FARO TESTS...................................................................94
7.4.4.
EFFECT OF THE NODALIZATION SCHEMES ON THE RESULTS .................................................95
7.4.5.
CONCLUSIONS.......................................................................................................................................96
7.5.
CHARACTERIZING MELT FUEL-COOLANT INTERACTION IN A NPP CONTEXT WITH
RELAP5/SCDAP 3.2 CODE ........................................................................................................................................98
7.5.1.
INTRODUCTION ....................................................................................................................................98
7.5.2.
THE LOSS OF COOLANT ACCIDENT (LOCA) ..................................................................................98
7.5.3.
THE RELAP5/SCDAP 3.2 CODE ...........................................................................................................99
7.5.4.
NODALIZATION. ANALYSED CASES..............................................................................................100
7.5.5.
CALCULATION RESULTS ..................................................................................................................102
7.5.6.
INITIAL CONDITIONS FOR A MFCI DETAILED STUDY ..............................................................132
7.5.7.
CONCLUSIONS.....................................................................................................................................134
8. COMETA CODE MODIFICATIONS...................................................................................................................135
8.1.
PREVIOUS COMETA LOGIC AND ACTUAL NEEDS ..........................................................................135
8.2.
MODIFIED COMETA. CHANGES INTRODUCED ................................................................................137
8.3.
CALCULATIONS WITH MODIFIED COMETA LOGIC AND PREVIOUS COMETA LOGIC ...........141
8.4.
CONCLUSIONS .........................................................................................................................................154
Assessment of Size Aspects in Modelling Molten Fuel Coolant Interaction
CONTENTS
OVERALL CONCLUSIONS .........................................................................................................................................155
BIBLIOGRAPHY AND REFERENCES .......................................................................................................................157
PUBLICATIONS............................................................................................................................................................161
LIST OF FIGURES ........................................................................................................................................................163
LIST OF TABLES..........................................................................................................................................................166
Assessment of Size Aspects in Modelling Molten Fuel Coolant Interaction
Chapter 1 - INTRODUCTION
1. INTRODUCTION
The present PhD Thesis in Nuclear Engineering deals with the methodology for calculate the characteristics
and phenomenology of the interaction between core molten or melt material and residual coolant liquid
(light water) located into the PWR (pressurized water reactor) vessel lower plenum during a severe accident
progression. The study of the interaction of the melt fuel with the coolant water is the objective of MFCI
(Melt Fuel Coolant Interaction) activities.
The core molten material, also called melt is formed by the core elements materials (U and Zr mainly) at
about 3000 K of temperature. The melt relocation transient starts with the melt fall down from the core
molten pool and ends when all molten material has reached the vessel bottom.
The exhaustive comprehension of the characteristic Melt Fuel Coolant Interaction phenomena is very
important for the immediate effects assessment, for example sudden pressure increase, and also for the
physical state of the material that reaches the bottom plenum, for example its temperature or fragmentation
degree. This fragmentation defines the initial conditions of the lower plenum vessel thermal attack and the
likely subsequent ex-vessel severe accident transient.
The Three Mile Island (TMI) reactor accident has been an important experience for the study and evolution
of the light nuclear reactors (LWR) safety analyses. Theoretical studies supported by experimental facility
tests, built up after the TMI accident, have allowed the expansion and verification of complex calculation
codes, which are able to simulate a severe accident progression in a LWR. In this context the TMI
experimental data are fundamental for the codes assessment.
The PhD research work was performed in its first part in the European Commission (EC) Joint Research
Centre (JRC) of Ispra (Italy) as host site by means of a EC Marie Curie Research Training Grant (from
march 1998 to march 2000) and a EC Auxiliary contract (from march 2000 to march 2001). The JRC
researchers Alessandro Annunziato and Carmelo Addabbo supervised the work. Work was finished at the
Technical University of Catalonia (UPC), Barcelona (Spain). The director of the PhD Thesis is the UPC
professor Dr. Francesc Reventós.
In the field of reactor severe accident research, the European Commission carried out the FARO Research
Programme specifically devoted to the characterization of MFCI, melt quenching as well as melt spreading
phenomena. The MFCI experiments at JRC-Ispra site were conducted in the FARO test facility under
realistic melt composition and prototypical accident conditions. The COMETA thermalhydraulic code was
developed in the JRC-Ispra site in order to support FARO test preparation and execution with pre-test
calculations and assist test results interpretation with post-test calculations.
PhD Thesis includes all the following items carried out during the stage in the Joint Research Centre - Ispra
site: a) Complementing the application of COMETA code to the prediction of tests performed in the FARO
facility under new conditions not experimented before (subcooled and triggered conditions). b) Simulation
with COMETA of an extended MFCI accident in a hypothetical real size reactor with arbitrary geometry. c)
Verification, analyses and unification of the COMETA simulations carried so far simulating again all the
FARO experiments with the last available version of the COMETA code. All those subjects included the
identification of simulation problems in the application of the COMETA code, which lead to a further
development, and improvement of the code.
The general overview of the MFCI phenomena included in the PhD research was completed in the Technical
University of Catalonia (UPC) with a study of a large LOCA accident in the overall view of a NPP with the
RELAP5/SCDAP code. Focusing its interest in clarify how MFCI is placed in the context of a severe
accident.
1
Assessment of Size Aspects in Modelling Molten Fuel Coolant Interaction
Chapter 1 - INTRODUCTION
The document is structured in the following way:
Chapter 2 presents an explanation of the Melt Fuel Coolant Interaction (MFCI) phenomena.
Chapter 3 explains severe accidents phenomena and places the MFCI into the general overview of the severe
accidents, this chapter also gives a brief of the severe accident research, mainly MFCI research, around the
world, in terms of experimental facilities and analytical tools.
Chapter 4 presents the objectives of the PhD Thesis.
Chapter 5 presents the FARO facility, the main components and the simulation of accident phenomena into
the facility.
Chapter 6 presents the models of the thermalhydraulic and melt fragmentation COMETA code.
Chapter 7 is a review of the PhD research activities. In particular:
COMETA pre-tests calculations are explained in chapters 7.1 and 7.2. In addition to give information for test
experiment planning and execution in the FARO facility, the calculations gave results to improve COMETA
code logic.
Chapter 7.3 summarizes the application of the COMETA code to reactor configurations, larger than scaled
test facilities. This application was not performed before. The results lead to the identification of scale-up
problems in the logic of the COMETA code. They also solved and explained one of MFCI related
phenomena: the relation between the void fraction and the quenching rate.
To analyse and unify the experiment simulations with the last available version of the COMETA code
another work is presented in chapter 7.4. Along the years the COMETA code was improved and changed to
fit experimental results. This purpose was fulfilled simulating again tests from L-14 to L-31. This work gave
some indications that help to choose the nodalization scheme in order to reduce the COMETA computer
calculation time. The post-test analyses of the FARO experimental data increased the information for the
assessment and qualification of the COMETA code.
In chapter 7.5 the calculations with RELAP5/SCDAP code complete the general overview of the MFCI
phenomena included in the PhD research.
The application of COMETA code to the prediction of tests performed in the FARO facility and to reactor
configurations, larger than scaled test facilities, lead to improve and develop COMETA code. Chapter 8
explains an important COMETA code improvement, particularly in the modification of the code to adjust
drops radial movement, which is necessary in reproducing some local phenomena.
2
Assessment of Size Aspects in Modelling Molten Fuel Coolant Interaction
Chapter 2 - MELT FUEL COOLANT INTERACTION AND QUENCHING
2. MELT FUEL COOLANT INTERACTION AND QUENCHING
The MFCI (Melt Fuel Coolant Interaction) is the sudden evaporation of a coolant liquid when interacts
with another liquid, molten fuel, less volatile and hotter. Phenomena present in nature could verify also this
happening, for example falling volcanic magma emission in water. In a MFCI there is a rapid heat exchange
from the molten fuel to the coolant through the large increase of the heat exchange surface due to the molten
fuel fragmentation in a short period of time.
In a postulated severe accident in which the core melt-down partial or totally has occurred jets of molten
material are poured into the lower plenum through the core support plate. The presence of residual water in
the lower plenum can determine a fragmentation of the melt leading to interaction of melt particles with the
coolant. This thermal interaction, depending on the fragmentation degree and the fragments size, could lead
to a strong local vapour production that could eventually threaten the vessel integrity (steam explosion).
However, if no steam explosion occurs, a strong vapour production and high system pressurization exists. On
the other hand, if the melt reaches the bottom of the lower plenum unquenched potential exists for lower
head penetration due to the residual power still produced in the debris bed.
The study of the interaction of the melt fuel with the water MFCI (Melt Fuel Coolant Interaction)
activities is one of the most important issues awaiting resolution in water cooled reactor safety analysis. The
progression of a severe accident in a water cooled reactor can lead to energetic (steam explosion) or nonenergetic (melt quenching) interaction as the molten fuel relocates and eventually interacts with the coolant
either in the vessel lower head (in-vessel) or in the cavity (ex-vessel). Referring to potential steam explosion,
it is important to determine the melt-coolant mixture conditions in order to evaluate whether an explosion is
conceivable or not. For melt coolability, the physical state of the molten fuel reaching the lower plenum
bottom can seriously affect the possibility of coolability and vessel failure.
In MFCI experimental problems one of the main matters is the determination of the conditions of the melt
when it arrives to the bottom plate of the vessel. If great fragmentation and quenching is produced in the
water no big problems for vessel attack would occur. If instead a significant amount of melt would remain as
a solid molten cake, potential for lower head penetration would exist.
In-vessel progressions take place into the pressure reactor vessel during a severe accident where initially core
is degraded loosing geometry and melted down, the transient ends with molten pool formed slumping and
interacting into the residual water in the lower plenum. Depending on the conditions of the molten material
when it reaches the bottom vessel an energetic (steam explosion) or non-energetic (melt quenching)
interaction could happen. Experimental studies have so far demonstrated that in-vessel steam explosion is
retained very unlikely to occur. Only in cases when primary system pressure runs down to very low values
steam explosion could be considered. In the working pressure ranges non-energetic (melt quenching)
interaction is likely to occur.
Ex-vessel progressions take place outside pressure reactor vessel. They start when molten material slumps
outside broken reactor vessel towards containment building. Molten material could interact with coolant
water present in the containment pavement; water delivered from a primary circuit LOCA (Loss of coolant
Accident) or present due to some emergency system procedure. Energetic interaction could occur because of
low pressure in the containment building but in general is not considered due to the limited water quantity
and only spreading of molten material over the water occurs.
Since the signing of the EURATOM treaty in 1957 the European Commission (EC) has been engaged in the
support of reactor safety research activities, principally in the synchronization safety practices and
methodologies among the EC member counties. The Commission promotes and conducts direct action
research programmes performed in the JRC laboratories in close collaboration with research organizations.
In the severe accident research field, the MFCI experiments at JRC-Ispra site were conducted in the FARO
test facility to provide basic information on fundamental phenomena. The Commission of the European
Communities in order to characterize MFCI, melt quenching as well as melt spreading phenomena
3
Assessment of Size Aspects in Modelling Molten Fuel Coolant Interaction
Chapter 2 - MELT FUEL COOLANT INTERACTION AND QUENCHING
established the FARO Programme in September 1990. A first series of corium melt water quenching
experiments was proposed in collaboration with US Nuclear Regulatory Commission (NRC), the Electric
Power Research Institute (EPRI), and the Ente Nazionale per l’Energia Elettrica (ENEL, Italy). This
programme was approved by the EC Member States and was followed by a group of EC experts.
KROTOS was another facility located in the JRC-Ispra site, it was a relatively small scale experimental
installation dedicated to the study of molten fuel coolant pre-mixing with little masses of prototypic reactor
melts or simulants as alumina (Al2O3). The progression of spontaneous and triggered energetic fuel coolant
interactions (steam explosions) was also studied.
The experimental facilities programme was complemented by comprehensive pre-test and post-test analytical
activities based on the development and application of the thermalhydraulic COMETA code.
From the MFCI experiences it is retained that a quite high pressurization rate has to occur in the interaction
zone to have an energetic interaction (steam explosion); this would disturb the stability of the vapour film
around the melt particles and possibly cause liquid-to-liquid contact with explosive consequences. The
pressurization rate that can be achieved spontaneously or with a trigger is function of the pulse strength and
the local void fraction.
FARO and KROTOS experiments shown that hydrogen generation due to several oxidation processes
during the premixing phase of an interaction process of UO2/ZrO2 melts with water, would lead to increase
void fraction formation within the mixture, which could suppress the propagation of triggered pulses and
thus an eventual energetic escalation. Although there are considerable quantitative uncertainties on the
behaviour of oxidic melts at high temperatures which need to be further assessed, COMETA calculations and
KOTOS tests suggested that hydrogen generation and thus void formation in the interaction of Al2O3 base
melts with water is less pronounced leading to thermalhydraulic conditions favourable to a steam explosion
with energetic escalation.
In general, COMETA results indicated that non-condensable gases as H2 have to be accounted for in the
calculation for a proper representation of melt fuel coolant interaction and quenching processes.
4
Assessment of Size Aspects in Modelling Molten Fuel Coolant Interaction
Chapter 3 - STATE OF THE ART
3. STATE OF THE ART
3.1. SEVERE ACCIDENT UNDERSTANDING
The accident of Three Mile Island (TMI)-2 reactor in the United States on 28th march 1978 was the first
nuclear accident with severe damage to the core. A severe accident in light water nuclear reactors occurs
when the reactor vessel water inventory decreases and there is no available additional water coolant to be
delivered into the core. In general during an extended severe accident sequence a period exists in which the
reactor core, after a partial or total melt down, is poured into the lower plenum that can have some water
present.
This accident shows that diagnostic and operation failing together with malfunctions in the operational
systems could lead to serious consequences that were not expected in the design base accident scenarios. It is
because severe accidents have now been studied worldwide for more than twenty years. Since TMI accident
great effort has been devoted by manufacturers, exploiters and regulatory organizations to prevent and
mitigate severe accidents: safeguards design, fission products confinement barriers, post-accident
instrumentation, inspection and maintenance, emergency operation procedures, degraded accident
management, operators training and formation among other subjects were generated or revised.
Multidisciplinary international projects have been raised to improve best understanding in the computational
models that simulate phenomena correlated to severe accidents: core degradation, containment response, as
well as to estimate the behaviour of gases, vapours and aerosols released in the accident progression.
3.2. MELT FUEL COOLANT INTERACTION IN LOWER WATER REACTOR SEVERE
ACCIDENT PHENOMENOLOGY CONTEXT
A wide range of physical and chemical phenomena characterizes severe accidents. Knowledge of those
phenomena and its interconnection can be acquired through experimental research at small and large scale
and through analyses of the real experience.
Damage mechanisms exist during a severe accident in an LWR. They can change a reactor core from
cylindrical fuel rods located in the centre of the reactor vessel to a deep debris bed located in the bottom of
the vessel. The reactor vessel itself becomes vulnerable to heatup and damage after the core has relocated to
the lower head.
The TMI-2 accident and severe fuel damage experiments have shown that reactor core damage proceeds
through several stages before the core slumps to the lower head. These stages of damage progression include:
(a) embrittlement of cladding due to oxidation, (b) melting of metallic cladding and dissolution of fuel in
contact with liquefied cladding, (c) slumping of liquefied cladding and dissolved fuel due to failure of the
oxide shell containing the liquefied mixture, (d) solidification of the slumped mixture at a lower and cooler
location in the core and concurrent formation of a nonporous debris region that blocks the flow of coolant,
(e) meltdown of the reactor core into a molten pool supported by the frozen previously molten ceramic
material, and (f) melt-through or structural failure of the crust of frozen material that supported the molten
pool and slumping of the molten pool to the bottom of the reactor vessel. An hour or more of time may
elapse before damage has progressed through these six stages. This order of damage progression is
established by the differences in melting temperatures of the metallic and ceramic parts of the reactor core.
Main severe accident associated well-known phenomena are represented in the schematic drawings of Figure
3.1.
Among the in-vessel phenomena:
5
Assessment of Size Aspects in Modelling Molten Fuel Coolant Interaction
Chapter 3 - STATE OF THE ART
Figure 3.1 a) represents cladding degradation when cooling system is not able to extract thermal energy
produced in the fuel. Clad rod heats up, looses its mechanical properties, suffers ballooning, breaks due to
internal pressure or does not break. Previous breaking sometimes occurs without ballooning. In any case,
cladding zircalloy oxidizes in vapour presence, at temperatures above about 1500 K. Oxidation produces
hydrogen and energy (Zr + 2H2O ? ZrO2 + 2H2 + 586 KJ/mol), which increases core heating rate and could
produce temperature peaks that causes cladding melting. The oxidation increases also melting temperature of
a-Zr(O) up to 2250 K and 2970 K for ZrO2.
Figure 3.1 b) shows advance of cladding and fuel degradation. Oxidation energy and channels blocking due
to cladding ballooning increase the fuel and cladding heating. A Zr-Fe and Zr-Ni eutectic is the first liquid
formed. The In-Ag alloy also attacks Zr at 1500 K. Zirconium forms a eutectic with UO2 that melts at 2500
K. The molten material drips over the intact clad rods and freezes in the cooler core parts as the plate that
supports fuel rods.
The TMI-2 reactor core autopsy as well as the series of integral experiments SFD, CORA and PHEBUSCSD have demonstrated that a metal solidified crust is formed, generally located on the lower plate (Figure
3.1 c)). This crust encloses a molten material pool, commonly oxides UO2, ZrO2 and steel components that
form eutectic mixtures with low melting points (1670 K). Another crust, that contains oxides, is formed
above the molten material pool. A porous debris bed, which is difficult to cooling, is formed above the crust.
Before damage has occurred, the configuration of the core is characterized by parameters such as rod spacing
and rod diameter. After in-vessel damage has occurred, the configuration is characterized by parameters such
as depth and porosity of debris. If the porosity is large and the debris is covered with water, then most of the
decay heat in the debris can be removed by convective cooling. But, if the porosity is small and the debris
bed is deep, then a large molten pool may develop. At this stage, there is the potential for rapid slumping of a
large amount of hot material into the lower head of the reactor vessel and the possibility of a vigorous
thermal attack of the lower head.
The damage progression can be either abated or intensified by the injection of water into the reactor vessel.
On one hand, the injected water may cool the damaged core and stop damage progression. On the other hand,
the injected water may break embrittled fuel rods so that the fuel rod fragments slump into the configuration
of a porous debris bed. The injected water may also increase an excursion in the oxidation of fuel rod
cladding in the parts of the reactor core where the oxidation was steam starved.
The last phenomenon recognized into the reactor vessel is the corium molten pool slumping into the bottom
vessel (Figure 3.1 d)). In the TMI-2 accident the disintegration heat made the crust unstable due to thermal
and mechanical stresses. The fluid pours in coherent shape into the bottom vessel. If there is residual water in
the vessel at the slumping time, molten fuel interacts with water (Melt Fuel Coolant Interaction),
increasing pressure and continuing oxidation of zircalloy. A vapour explosion cannot be excluded. In other
sequences hotter parts of the core can fall into the lower plenum due to lower plate destruction. These hot
rests can form a liquid pool directly in the bottom vessel.
Among the ex-vessel phenomena:
Figure 3.1 e) shows molten corium expulsion outside pressure vessel. The interaction of fused and
superheated corium with the bottom vessel steel can break it provoking fluid expulsion. Experts predict 3
methods, depending on pressure: (1) One vessel zone partial fusion. (2) Instrumentation tubes (PWR or
BWR) fusion or control rods devices (BWR) fusion. (3) Bottom vessel break due to fluence at high
temperature. TMI-2 vessel did not suffer degradation and only some tubes were melting. Probably melt fuel
coolant interaction formed a fragmented debris bed, which was easy to cool. TMI-2 VEP project from
NEA/OECD studied this important issue.
Figure 3.1 f) represents when molten corium is expulsed outside pressure vessel like a jet at high pressure
and collides with the concrete pavement on the containment floor. Experts predict corium fragmentation in
small hot particles and the particles diffusion in the containment environment. The non-oxidized zirconium
6
Assessment of Size Aspects in Modelling Molten Fuel Coolant Interaction
Chapter 3 - STATE OF THE ART
can still react with the water steam present, releasing chemical energy and hydrogen. The atmosphere heating
increases pressure and temperature on the containment walls threatening its integrity. The generated
hydrogen increases the detonation risks. The detailed analyses of the associated phenomena and the
containment break are being studied.
The hydrogen generated in the oxidation reactions; mainly from zirconium oxidation is accumulated in the
containment room. It could reach the detonation threshold of the mixtures hydrogen-water steam-air. The
shock waves generated can threaten vessel integrity (Figure 3.1 g)). Devices are installed for H2 combustion
in a controlled way to avoid the explosion risks. All the phenomena associated to hydrogen generation and
transport, the detonation threshold, the transition from deflagration to detonation, the flames propagation, the
shock waves physics, etc. is now object of deep study.
Figure 3.1 h) represents corium-concrete interaction. Corium poured on the containment floor produces
concrete thermal disintegration, with water steam, CO2 and silicon oxides release. These gases react with the
corium metallic components producing hydrogen and CO. Oxides released in the concrete disintegration are
incorporated to the corium reducing the melting temperature of the mixture. Corium molten mass cooling
with water has not been demonstrated experimentally. Phenomena related to corium-concrete interaction has
been studied in the EPRI-MACE project.
Among the gases, vapours and aerosols behaviour:
Gases, vapours and aerosols release (Figure 3.1 i)): 1) Cladding break releases gas and vapour that were
accumulated in the gap, mainly noble gases and elements or chemical volatile substances. 2) Fuel melting
releases vapours of the elements and components with lower boiling point, as caesium, iodine, tellurium,
which lately become aerosols by condensation. 3) During the corium-concrete interaction, the higher
temperatures and the gas bubbling lead to refractory oxides release, oxides that become aerosols of
mechanical origin.
Gases, vapours and aerosols transport through primary system (Figure 3.1 j)): Gases, vapours and aerosols
generated in the core, travel by pipes and components and arrive to the containment or outside the system.
During transport gases could remain trapped by condensation, adsorption or precipitation in metallic
surfaces. The gases, vapours and aerosols in the same way could break off from the walls when the hydraulic
or thermal conditions change in the pipes and components or when physical properties change forming new
chemical components.
Retention phenomena of gases, vapours an aerosols in water pools are represented in Figure 3.1 k). The
bubbling of gases, vapours and aerosols in water pools is an efficient method to retain them. The document
NUREG-1150 states decontamination factors from 1.2 to 4000. In PWR reactors these phenomena can
happen in relief tanks, secondary side of steam generators and pressurizer. In BWR reactors it occurs in the
relief pool. Gases vapours and aerosols in the bubbles can pass to the water through the separation
membrane. The pass rate depends on the thermalhydraulic conditions of bubbles and mass of water. The
LACE project investigated this subject.
Gases, vapours and aerosols retention phenomena in containment building are represented in Figure 3.1 l).
Containment buildings have systems to retain gases vapours and aerosols released in accidents. Spray and
filtration systems are the most commons. Water drops are created by the aspersion system and dissolve easily
some components of iodine and caesium, while the high efficiency filters are much effective in retaining
aerosols. Precipitation and adsorption are also used. Iodine can form inert organic molecules, which are
difficult to retain. Radiation and disintegration heat introduce additional phenomena.
7
Assessment of Size Aspects in Modelling Molten Fuel Coolant Interaction
Chapter 3 - STATE OF THE ART
In-vessel phenomena:
a) Cladding oxidation and
ballooning
b) Fuel dissolution and
melting
c) Crust, melt and debris
bed formation
d) Corium
relocation in the
bottom vessel
Ex-vessel phenomena:
e) Expulsion of melt corium f) Containment direct heating g) Hydrogen explosions
h) Corium-concrete
interaction
Gas, vapour and aerosol behaviour:
i) Gases, vapour and aerosol j) Gases, vapour and aerosol k) Gases, vapour and aerosol l) Gases, vapour
release
transport through primary
retention phenomena in
and aerosol
system
water pools
retention
phenomena
in containment
buildings
Figure 3.1 - Severe accident associated phenomena [From [34]]
8
Assessment of Size Aspects in Modelling Molten Fuel Coolant Interaction
Chapter 3 - STATE OF THE ART
3.3. WORLD WIDE SEVERE ACCIDENT RESEARCH
3.3.1.
EXPERIMENTAL FACILITIES
In 1996, the Senior Group of Experts on Nuclear Safety Research Capabilities and Facilities (SESAR/CAF)
[28], which was investigating the ability of OECD Member Countries to sustain an adequate level of
research, identified a number of facilities and programmes that were important for continuing research
needed by the safety community during the coming decade. They also pointed out that many of these
facilities and programmes were facing increasing budgetary constraints, and that some would cease to be
supported at the national level in the near or medium-term future. Some of the facilities were of interest to
more than one country. It therefore seemed logical to investigate the possibility of operating the facilities in
an international context, in order to share the costs and the expertise, and to promote quicker and deeper
international consensus on safety issues. It was in this context that in 1997 the NEA Committee on the Safety
of Nuclear Installations (CSNI) decided to set up a Senior Group of Experts on Nuclear Safety Research
Facilities and Programmes (SESAR/FAP). The new Senior Group of Experts was asked to identify facilities
of potential interest for present or future international collaboration, to make specific recommendations
regarding facilities, research programmes, and joint projects, and to discuss other possible forms of
international collaboration. For efficiency, the group was restricted to the countries running the widest and
most advanced research programmes. Some of the conclusions of the studies are presented in this chapter.
Severe accidents join three main topic areas: in-vessel phenomena, ex-vessel phenomena and fission
products.
Due to the enormous extension of research topics in severe accidents only a little brief of the current state of
the art following ex and in-vessel phenomena research facilities is given. Some of the in main subjects
related to the in-vessel phenomena are:
1. Core Degradation and Melt Progression: The facilities currently in operation conducting core melt
progression experiments for PWRs are the QUENCH facility in Germany, PHEBUS in France and CODEX
in Hungary.
QUENCH facility investigates the early phase of core melt progression, with emphasis on reflooding and the
resulting fragmentation and H2 generation. It was planned to operate in the near term.
PHEBUS had three tests remaining, primarily directed toward fission product transport; however, some
information on early and late phase melt progression will be obtained, although at a small scale. Work on
irradiated fuel is undertaken in PHEBUS. The continued conduct of core melt progression experiments will
complement analytical code development and assessment and may be useful in addressing any concerns
related to the changing operational characteristics of operating plants (e.g. higher burn-up fuels, higher
power density).
2. Melt Fuel Coolant Interaction (MFCI): Currently there are several facilities in operation investigating
MFCI.
Small-scale facilities (using simulant materials) exist in the US (University of California-Santa Barbara MAGICO and SIGMA), Japan (ALPHA), Korea (CONVEX), Sweden (MIRA), COTELS (Kazakhstan,
sponsored by Japan) and France (MICRONIS and TREPAM).
The ECO (Experiments on energy COnversion during a steam explosion) facility (FZK, Germany), housed
inside the large FAUNA steel vessel, was designed for investigating energy conversion ratios up to 20%,
related to 10 kg of melt under well-defined conditions after an energetic steam explosion. Alumina from a
thermite re-action is used as a simulating material instead of corium.
The QUEOS (QUEnching Of Spheres) facility (FZK, Germany) serves to study premixing phenomena with
solid spheres, i.e. without the danger of a steam explosion and the complication of melt fragmentation. In
9
Assessment of Size Aspects in Modelling Molten Fuel Coolant Interaction
Chapter 3 - STATE OF THE ART
order to simulate melt jets as closely as possible, the spheres are released as a cylindrical jet into a threedimensional test vessel.
The PREMIX (FZK, Germany) experiments have been performed to study the premixing of sizable amounts
of very hot oxidic melts with water when being released as a jet in a reasonably characterized way and with
full optical access. Alumina at 2600 K from a thermite reaction was used to simulate the corium melt.
PREMIX involves the full physics of the mixing process including jet break-up and melt drop fragmentation.
Each of these facilities, with the possible exception of ALPHA, is projected to continue operation in the near
term or be available, if needed.
The MISTEE (Micro Interactions STeam Explosion Experiment) (RIT, Sweden) facility is laboratory scale
research on Molten Fuel-Coolant Interactions (MFCI) using a single molten drop. A new test facility was
constructed. The objectives of this activity on single drop steam explosions are to investigate the
triggerability and explosivity in a well-controlled facility with high temperature melts with an external
trigger, to identify the influence of melt thermo-physical properties on triggerability, fragmentation and
explosivity of the melts, to acquire quantitative data on the volume fractions of melt, coolant and vapour in
the interaction zone and on the fine fragmentation process during the explosions, and eventually to develop
scaling methodology for the steam explosions.
BILLEAU (CEA Grenoble, France) facility studies premixing phase of the vapour explosion phenomenon;
study of the dispersion of hot plane jets into water; use of spheres (diameter 5 to 20 mm) heated up to
2200°C (analogies with QUEOS experiments which treat hot axisymetric jets). Measurement of the fraction
of fragments, liquid water and steam.
The FARO (Furnace And Release Oven) and KROTOS in the European Commission JRC-Ispra (Italy) were
shut down in 1999. The loss of FARO will eliminate large scale experimental data using prototypic material
as well as useful data on melt quenching and spreading. KROTOS facility devoted to the study of molten
fuel-coolant pre-mixing either with prototypic reactor melts or simulants such as alumina up to 5 kg and
progression and energetics of spontaneous and triggered fuel-coolant interactions (vapour explosions) is
relocated in France, where it will again be available for use.
In-vessel fuel coolant interactions can challenge reactor vessel integrity, if they are large enough. Knowledge
sufficient to predict their occurrence and energy release is not mature and analytical tools based on MFCI
fundamental physics need further development. However, there is general consensus that as long as the
primary system is pressurized, large-scale in-vessel MFCI is very unlikely and additional research on this
topic is not warranted. For low-pressure conditions there remains the potential for energetic MFCI.
Therefore, the fundamentals of mixing and triggering (Trigger refers to potential steam explosion, it is
important to determine the melt-coolant mixture conditions in order to evaluate whether an explosion is
conceivable or not) MFCI still need investigation to develop adequate models to predict MFCI during
unpressurized conditions (in-vessel and ex-vessel) and this is an area where additional experimental work is
warranted to characterize under what conditions and to what extent MFCI will occur, including under what
reactor vessel re-flood conditions. Also, confirmatory testing using prototypic materials are highly desirable
to completely assess the analytical tools and resolve the remaining issues.
Work is continuing on the development of analytical tools and small-scale MFCI experiments. This work is
directed toward understanding the fundamentals of MFCI. As discussed above, additional work on MFCI
under low pressure conditions is warranted since MFCI is an important consideration in reactor vessel
integrity and accident management (e.g. should the reactor cavity be flooded before or after vessel failure?)
and in developing new designs. With the loss of the FARO facility at the end of 1999, large-scale
experimental facilities, data and expertise in handling large amounts of prototypic materials needed to
address MFCI will be lost. Therefore, there is a need to consider whether or not additional confirmatory
experiments using prototypic materials under low-pressure conditions are needed. This would also help
maintaining expertise in handling and understanding the effects of prototypic materials.
10
Assessment of Size Aspects in Modelling Molten Fuel Coolant Interaction
Chapter 3 - STATE OF THE ART
MFCI issues are relevant to safety aspects and related accident management strategies in current reactors as
well as to the development of advanced safety features to be incorporated in new reactor concepts. The
European Commission carried out for many years the FARO experimental facility program, which was
specifically devoted to the characterization of MFCI, melt quenching as well as melt spreading
phenomenologies. The program consisted of experimental and analytical activities at JRC-Ispra (Italy) and
KFK-Kalsruhe (Germany). These activities support test preparation and execution with pre-test simulations,
assist test results interpretation with post-test calculations and perform sensitivity analyses to point out the
various parameters influence.
In Canada, a new programme is being established to investigate in-vessel fuel/water interactions in CANDU
reactors for accident scenarios (e.g. flow blockage in a fuel channel) that can lead to molten fuel being
forcibly ejected from fuel channels into the surrounding D2O-filled Calandria vessel. A new facility, the
Molten Fuel Moderator Interaction (MFMI) facility is currently being designed for this purpose at AECL’s
Chalk River Laboratories. The experimental programme will investigate the energetics associated with the
ejection of up to 25 kg of prototypic molten CANDU corium into the moderator. This programme is
expected to be completed by the end of 2005.
Among other Fuel Coolant Interaction facilities in present and past years:
At the Argonne National Laboratory the COEXIT facility allowed for FCI studies of 1-10 kg of corium.
Series of experiments were focused on FCIs occurring in the ex-vessel reactor cavity and the associated
containment pressurization from high and low pressure melt ejections into the cavity region, also tests were
performed to investigate the coarse mixing and melt jet breakup for an in-vessel geometry.
At Winfrith large scale experiments were performed using corium thermite in the MFTF and MIXA
facilities. MFTF facility was a large pressure vessel equipped with a thermite charge container. Melt was
released in two ways: free and restricted release modes. Studies were focused on the effects of melt mass,
ambient pressure and subcooling water, and explosion energetics. MIXA facility was designed for studying
the mixing phase of FCI, in particular the corium melt poured through a droplet format, which produced
various diameters of melt droplets before melt-water contact.
At the Sandia National Laboratories many large scale experiments were conducted in the EXO-FITS and
FITS facilities, focusing on estimating the FCI energy conversion ratio, determining the triggering behaviour
and explosion threshold.
At the University of Wisconsin it was constructed the WFCI facility to support FCI research as originally
sponsored by the US NRC, in particular to perform fundamental experiments in energetic FCI. The
experimental test chamber was designed as an approximate one dimensional geometry (L/D>10). It allows
many measurements (fuel-coolant mixing and explosion propagation/escalation, mixture level swell, spatial
and temporal shock pressure histories, explosion propagation velocities, expansion work and fragmented fuel
debris distributions).
3. Debris Interaction with the reactor vessel Lower Plenum: Currently, work in this area is being
conducted in the following places:
RASPLAV (Kurchatov Institute/Russia) - OECD co-operative project to investigate load to the reactor vessel
lower head and chemical interaction between the reactor vessel steel and molten core debris, using prototypic
materials. Separate effect tests and large scale real materials experiment up to 200 kg of UO2 / ZrO2 / Zr /
Stainless steel heated up to 2300 - 2700 °C. Melt pool behaviour in the lower head with or without cooling.
Corium lower head interaction.
COPO2 (Finland) - to investigate heat load to the reactor vessel using water as a simulant.
SNL-LHF (USA) - OECD co-operative project to investigate the mechanical behaviour failure modes of the
reactor vessel lower head under high temperature pressurized conditions.
11
Assessment of Size Aspects in Modelling Molten Fuel Coolant Interaction
Chapter 3 - STATE OF THE ART
BALI (France) - to investigate the heat load to the reactor vessel using water as a simulant.
ALPHA (Japan) - to investigate in-vessel debris coolability.
FOREVER (Failure Of REactor VEssel Retention) (RIT, Sweden) experiment is an integral test investigating
the creep failure of a 1:10 scale reactor pressure vessel under the combined thermal and pressure loadings. It
simulates the late stages of in-vessel-melt progression under nuclear reactor severe accident scenario. The
objectives of the EC-FOREVER tests are to obtain multi-axial creep deformation and vessel failure mode
data for the prototypical vessel geometry (scaled 1:10), under prototypical thermal and pressure loading
conditions.
The objective of the POMECO (POrous MEdia COolability) experiment is to investigate the effect of
downcomer configuration on the dryout heat flux for an internally-heated particulate debris bed.
SONATA (Korea) - investigating interactions with the lower head. Currently using simulant materials.
RUPTHER (France) - investigating mechanical behaviour (creep rupture) under accident conditions.
The DISCO-C (DIspersion of Simulated COrium with Cold fluids) (FZK, Germany) facility serves to
investigate melt dispersal from the reactor pit when the reactor pressure vessel lower head fails at low system
pressure of less than 2 MPa. The fluid dynamics of the dispersion process is studied using model fluids,
water or bismuth alloy instead of corium, and nitrogen or helium instead of steam. The effects of different
breach sizes and locations, and different failure pressures on the dispersion can be studied.
The DISCO-H (DIspersion of Simulated COrium with Hot fluids) (FZK, Germany) test facility was set up to
perform scaled experiments that simulate melt ejection scenarios under low system pressure in Severe
Accidents in Pressurized Water Reactors (PWR). These experiments are designed to investigate the fluiddynamic, thermal and chemical processes during melt ejection out of a breach in the lower head of a PWR
pressure vessel at pressures below 2 MPa with an iron-alumina melt and steam.
The experimental programme named KAJET (Karlsruhe Jet Experiments, FZK, Germany) is being
performed to investigate features of a pressurized melt jet and the interaction with substratum material
(MCCI). Compact melt jets, rather than a spray-type melt release, are simulated using iron and aluminum
oxide instead of corium. The experiments provide general information about erosion processes and data for
the validation of computer codes (or, if possible, simplified correlations), which then are able to transfer the
results to reactor conditions.
At present, in the general context of a nuclear severe accident (ex and in-vessel phenomena) and in the way
to prevent radioactive release, the French Atomic Energy Commission (CEA) is pursuing an important
programme on nuclear severe accidents for many years. It encompasses the development of models and
codes, performance of experiments in simulant and prototypic materials and the analysis of international
experiments.
The CEA severe accident studies on corium behaviour address the following topics: Molten Core Concrete
Interaction (MCCI), Fuel Coolant Interaction (FCI), and In Vessel Retention with reactor pit flooding and
Corium-Ceramic Interaction. The experiments with prototypic corium are performed in the PLINIUS
experimental platform at CEA Cadarache (France).
The European Commission who financially supports Transnational Access to this Research Infrastructure
has selected the PLINIUS platform. The 4 facilities of the PLINIUS experimental platform where are
performed the experiments dedicated to the understanding of the corium behaviour are: VULCANO,
COLIMA, VITI and KROTOS. Material analyses and code validation activities complement the
experimental work, which is directed towards the analysis of new phenomena.
PLINIUS is an acronym that means: PLatform for Improvements in Nuclear Industry and Utility Safety.
Four facilities are devoted to the corium behaviour and to the physical properties studies:
12
Assessment of Size Aspects in Modelling Molten Fuel Coolant Interaction
Chapter 3 - STATE OF THE ART
The VULCANO facility (Versatile UO2 Lab for Corium ANalysis and Observation): it is a rotating plasma
arc furnace able to melt about 80 kg of corium at temperatures of up to 3000°C (in or ex-vessel corium) and
to pour the melt according to different configurations: spreading, interaction, solidification studies.
The COLIMA facility (COrium LIquid and MAterials): it is a 1.5 m 3 controlled atmosphere vessel, with an
internal pressure, which can rise to 0.3 MPa. Induction heating can maintain some kilograms of corium at
very high temperature (up to 3000°C) to study thermal exchanges, aerosol release and thermo-physicals
interactions studies.
The VITI facility (VIscosity Temperature Installation) has been developed to perform viscosity and surface
tension measurements on corium by aerodynamic levitation up to 2500°C. Samples of a few mg corium can
be implemented.
The KROTOS facility: it is dedicated to steam explosion phenomenon studies. About 5 kg of corium at more
than 2850°C are dropped in water. Thermal, optical and pressure instrumentation, along with fast imaging,
constitute the instrumentation. The KROTOS facility for corium-water interaction tests has been transferred
from JRC Ispra (Italy) and is being reinstalled at CEA-Cadarache on the PLINIUS platform.
3.3.2.
ANALYTICAL TOOLS
Computer codes have been developed around the world to analyse nuclear severe accident phenomena and to
support and simulate pre- and post- test carried out in the experimental facilities. Some examples:
The ICARE/CATHARE code [41] was developed by French Institut de Protection et de Sûreté Nucléaire
(IPSN) in the framework of security research studies. The aim of the code is to predict the PWR core
behaviour during an hypothetical severe accident. The severe accident can be result of a combined failure of
reactor standard cooling system and security cooling system, and could lead to reactor core melting. ICARE
code is used for the preparation and interpretation of the PHEBUS PF international program experiments.
Sandia National Laboratories (SNL) developed MELCOR for the USA Nuclear Regulatory Commission
(NRC). It is a fully integrated, relatively fast-running code that models the progression severe accident in
LWR. An entire spectrum of severe accident phenomena is modelled. Characteristics of severe accident
progression that can be treated with MELCOR include the thermalhydraulic response in the reactor coolant
system, reactor containment and confinement buildings; core heatup and degradation; radionuclide release
and transport, hydrogen production, transport, and combustion; core concrete attack; and the impact of
engineering safety features on thermalhydraulic and radionuclide behaviour.
The Idaho National Engineering and Environmental Laboratory (INEEL) developed the RELAP5/SCDAP
code for the USA NRC. It was designed for best-estimate transient simulation of light water reactor coolant
systems during a severe accident. The code models the coupled behaviour of the reactor coolant system, the
core, fission products released during a severe accident transient as well as large and small break loss-ofcoolant accidents, operational transients such as anticipated transient without SCRAM, loss of offsite power,
loss of feedwater, and loss of flow. A generic modelling approach is used that permits as much of a particular
system to be modelled as necessary. Control system and secondary system components are included to
permit modelling of plant controls, turbines, condensers, and secondary feedwater conditioning systems.
The integral code ASTEC (Accident Source Term Evaluation Code) [12] [42] is being developed by IPSN
(Institut de Protection et de Sûreté Nucléaire), France, and GRS (Gesellschaft für Anlagen und
Reaktorsicherheit), Germany, since 1994. The aim of this close co-operation of both companies is the
creation of a fast running integral code which allows the calculation of the entire sequence of a severe
accident in a light water reactor from the initiating event up to the release of fission products into the
environment, covering all important in-vessel and ex-vessel phenomena. The main fields of application of
this code are probabilistic safety analysis level 2 studies, accident sequence studies, uncertainty and
sensitivity studies and support to experiments.
13
Assessment of Size Aspects in Modelling Molten Fuel Coolant Interaction
Chapter 3 - STATE OF THE ART
Since the 1980s, a two tier approach has been applied by IPSN and GRS based on the simultaneous but
independent development of both integral and detailed mechanistic codes. During this time IPSN has
developed the integral code ESCADRE and GRS has modelled the containment behaviour using two codes,
RALOC for the thermalhydraulics and the hydrogen distribution and FIPLOC for the aerosol behaviour. For
the first ASTEC version (called V0), it has been decided to gather in the same system the best candidates,
which can be provided by the two companies. Thus, ASTEC V0 consists in a combination of some modules
of ESCADRE (for the reactor cooling system, core degradation, fission product release and transport, corium
ejection from the vessel, direct containment heating and iodine chemistry in the containment), and of the
module CPA (Containment Part of ASTEC), which combines the RALOC and FIPLOC codes.
The main code requirements are: best modelling available in both developing organizations, fast-running
code, sufficient validation to cover the main physical phenomena, to account for safety systems and
procedures, easy use to perform sensitivity analyses.
The European Validation of the Integral Code ASTEC (EVITA) project (5th EC FWP) has as main objective
to distribute the severe accident integral code Accident Source Term Evaluation Code (ASTEC) to European
partners in order to apply the validation strategy issued from the VASA project (4th EC FWP). Partners
evaluate the code capability through validation on reference experiments and plant applications accounting
for severe accident management (SAM) measures, and compare results with reference codes.
The basis version V0 of ASTEC commonly developed and basically validated by GRS and IRSN was made
available in late 2000 for the EVITA partners on their individual platforms. The actual version V1 has been
released to the EVITA partners by the end of June 2002. It allows simulating the front-end phase by two new
modules.
Following modules are forming ASTEC V1: CESAR for thermal hydraulics in the reactor coolant system;
DIVA for core degradation up to vessel lower head failure; ELSA for fission product release from fuel rods;
SOPHAEROS for fission product vapour and aerosol transport in the reactor coolant system; RUPUICUV
for direct containment heating (DCH); CORIUM for heat transfer between containment atmosphere and
corium entrained out of the cavity by DCH; WEX (based on WECHSL) for molten-corium-concrete
interaction (MCCI) in the cavity; Containment part of ASTEC (CPA) for thermal hydraulics, aerosol and
fission product behaviour inside the containment; IODE for iodine behaviour in the containment; ISODOP
for calculation of activity and decay heat in the reactor zones; SYSINT for management of engineered safety
systems.
Among the specific codes for the Melt Fuel Coolant Interaction study:
At JRC-Ispra the development and application of the thermalhydraulic COMETA (Core Melt
Thermalhydraulic Analysis) code [4] complemented the FARO experimental program by comprehensive
pre-test and post-test analytical activities.
The COMETA code is composed of a two-phase flow field and a three phase corium field. The two-phase
flow field is described in Eulerian coordinates by 6 equations (mass, momentum and energy conservation
equations for each phase) and ‘n’ mass conservation equations for the non-condensable gases. The corium
field, which is composed by the jet, the droplets and the debris bed are described in Lagrangian coordinates.
Melt fragmentation and relocation model treats the jet, the droplets and the fused debris separately. The melt
is assumed to be released in the form of a coherent jet, which has a conical shape, based on the initial
discharge diameter. During its descent the jet accelerates and reduces in size as particles leave the jet surface;
the erosion rate develops as function of the jet breakup length (Saito, Epstein-Fauske or mixed correlations
are available in the current version of the code). The drops are created with an initial diameter to satisfy the
Weber number criterion. Up to 500 groups of drops can be followed; each group is characterized by
diameter, mass, temperature, velocity and position providing good basis for the analysis of the statistical
distribution of the debris particles. The residual part of the jet, which reaches the bottom, relocates as a cake
that then forms a conglomerate with the loosed debris. Although the heat transfer from the bottom
conglomerate may be negligible during the initial phase, it becomes important and has to be properly
14
Assessment of Size Aspects in Modelling Molten Fuel Coolant Interaction
Chapter 3 - STATE OF THE ART
accounted for the long-term energy balance. A simplified model to describe hydrogen generation during the
interaction phase is included in the current version of the code.
The IKE version of the COMETA code builds up on the original version of the code and includes an
additional jet fragmentation model developed at University of Stuttgart, Germany, which assumes as decisive
mechanisms wave growth and wave crest stripping along the jet surface. Thereafter an erosion rate is
determined leading to a coherent core of molten material surrounded by a cloud of fragments.
MC3D is a multicomponent 3D code developed by CEA, Grenoble, France, which describes within the
Eulerian approach the behaviour of water, steam, hydrogen, fuel droplets and corium jet. The code has five
continuity equations, four momentum equations and three internal energy equations. Steam and hydrogen are
mixed in both momentum and energy equations. Each field can exchange momentum and energy with the
other field, the exchange of mass being limited in this version between water and steam and between corium
jet and droplets. The corium droplet population will undergo fragmentation described through an interfacial
area transport equation. Jet fragmentation is calculated from an external analytical model dealing with the
small scale instabilities initiated by the hot vapour film.
The IVA code, developed at Forschungszentrum Karlsruhe and at Siemens, Germany, models transient
multiphase flows consisting of water, steam, non-condensable gases, microscopic and macroscopic solid
particles and/or molten materials. Three material fields are modelled by the code; the first one contains a
mixture of steam and air, the second one liquid water, and the third one a normally hot, heavy material
(corium) in the liquid or solidus state. A Cartesian mesh is used. The model can be up to 3D, rectangular or
cylindrical, and is to be subdivided in a number of cells in a rectangular grid. Time dependent volumes
fractions of the three material fields (and their state variables, velocities, etc.) are calculated for each cell. A
flow regime is assigned to each cell, depending on the volume fraction of the fields and in part on their
physical state, and the modelling of the transfer of mass, momentum and energy depends on this flow
regime. With reference to melt fragmentation/coalescence, the code does not contain an explicit jet model,
but as long as the melt has not solidified a flow regime with continuous melt can occur. When changing from
this regime to one with discontinuous melt, an initial particle size of 0.1 m is assumed. Further fragmentation
and coalescence are calculated by the code until the melt cools down to 40% liquid, where solidification is
assumed. Fragmentation tends to dominate as long as the melt moves, whereas coalescence tends to
dominate in melt settling down. Particle sizes do not change any more after solidification except through
thermal shrinking and mixing with particles from other calculation zones.
The IFCI (Integrated Fuel Coolant Interaction) code was developed under the auspices of the USNRC at
Sandia National Laboratories, USA, to investigate fuel coolant interactions in an as mechanistic as possible
manner. The code is intended to address all aspects of FCI phenomena, including coarse fragmentation and
mixing of molten material with water, triggering, shock wave propagation and fine fragmentation, and
expansion of the melt water system. The ultimate objective of the code is to predict rates of steam and
hydrogen formation, melt fragmentation and dispersion, fission product release, shock wave generation and
propagation, and system loading for explosive and non-explosive FCIs in reactor systems. To add generality,
the fuel is divided in two fields: solid particles (corium debris) and molten corium (melt), while the gas
component is a more properly a mixture of steam and hydrogen. The melt is described by a Eulerian field,
interacting with the water and steam fields, which are also described in Eulerian coordinates.
The present version of JASMINE-pre (2.01) is a premixing simulation code for steam explosion analysis
developed at JAERI, Japan. JASMINE-pre solves three sets of mass, momentum and energy conservation
equations for steam, water and melt. These fundamental equations are formulated on Eulerian field,
discretized by FDM. A simple flow regime map based on volume fractions was used in the analysis of ISP39. The melt was always assumed to be a dispersed phase. Water was considered to be continuous if the void
fraction was less than 0.3, while steam was considered to be continuous if the void fraction was more than
0.7. Constitutive models for water-continuous and steam-continuous regimes were averaged if the void
fraction was between 0.3 and 0.7. The present version of JASMINE does not have any physical-based debris
settling/packing model. The option used assumes that the melt is instantaneously spread on the floor if the
temperature is above the melting point, so that the melt surface area is the same as the floor area. However, if
15
Assessment of Size Aspects in Modelling Molten Fuel Coolant Interaction
Chapter 3 - STATE OF THE ART
the fully spread melt layer on the floor is thinner than the value defined (5 mm), the melt surface area was
reduced keeping the defined layer thickness.
The TEXAS computer model was developed at the University of Wisconsin, USA, for the simulation of fuel
coolant interaction during its mixing, triggering and explosion phases. TEXAS is based on a onedimensional hydrodynamics code developed at Sandia National Laboratories and modified for fuel coolant
interactions. The code solves the 1-D, three-field equations describing the fuel, coolant vapour and liquid.
Two fields represent the coolant as a separate liquid and vapour in a Eulerian control volume, while one field
models the fuel as discrete material volumes or master particles in a Lagrangian formulation within the
Eulerian region. The ‘Lagrangian’ treatment for the fuel makes it quite straightforward to track the fuel
particle movement and thus eliminates the numerical diffusion difficulties encountered in pure Eulerian
codes. The two key constitutive relations involve hydrodynamic fragmentation during the mixing phase, and
thermal fragmentation of the fuel and rapid quench during the explosion phase. The TEXAS thermal
fragmentation model is a semi-empirical formulation based on the concept of vapour film boiling and coolant
jet impingement on the fuel surface. The current model reflects the key features for the rapid escalation and
propagation of the vapour explosion. As the pressure shock wave directly contributes to rapid fuel
fragmentation, the fragmented fuel is quenched by the coolant generating more vapour which in turn
increases the pressure sustaining the propagation of the shock wave to adjacent fuel coolant mixture regions.
The THIRMAL code has been developed at the Argonne National Laboratories, USA, to specifically address
the physical processes related to melt breakup, quenching, steam generation, hydrogen generation and debris
particle formation when the melt enters water as a circular stream. The code predicts conditions of melt
stream impingement, debris collection at the bottom of a water pool and premixing conditions for the
evaluation of steam explosion loadings. The code models are based upon observations of
breakup/intermixing phenomena in visualization experiments performed by EDL and others using hightemperature reactor material melts and water. The melt stream is treated in Lagrangian coordinates with
time-dependent release diameter, temperature and velocity.
The VAPEX code has been developed at the Electrogorsk Research and Engineering Center (Russia) to
analyse steam explosion processes under severe accident conditions. The code permits to study both invessel and ex-vessel fuel coolant interaction processes. Premixing and propagation models of the VAPEX
code are based on the three-dimensional multifluid approach. Three phases (melt droplets, water and steam)
are considered. Melt droplets are described with the Lagrangian approach. Unsteady 3-D governing
equations of mass, momentum and energy conservation for all species with a common pressure are used in
the VAPEX model. The VAPEX melt fragmentation model is based on the critical Weber number concept;
other correlations are, however, available as an option. Moving continuum is modelled by a system of liquid
particles, coincident with Eulerian computational cells at the old time step. A transition to a new time level is
realized by three stages: 1) determination of preliminary values of all the parameters without effect of flow
movement, 2) mass flows through cell boundaries are calculated, 3) final new values for all parameters are
determined. An explicit finite difference scheme is used for numerical solution of the governing equations.
The scheme conserves mass, momentum and energy in each computational cell.
3.3.3.
PRESERVATION OF SEVERE ACCIDENT DATA
All the considerable amount of resources devoted at the international level during the last few decades to the
generation of experimental databases in order to provide reference information for the understanding of
reactor safety relevant phenomenologies and for the development and/or assessment of related computational
methodologies can be coupled to new technologies and modern methods to storage all these enormous
information.
The extent to which these databases are preserved and can be accessed and retrieved is an issue of major
concern. In particular, it has been recognized that new working methods and rapid advancement of computer
hardware and software technologies require continuous upgrading of storage methods which can otherwise
render access to and retrieve of the data unpractical and in some cases impaired. Similarly, code
16
Assessment of Size Aspects in Modelling Molten Fuel Coolant Interaction
Chapter 3 - STATE OF THE ART
documentation including code sources and input decks as well as relevant assessment cases need to be
properly preserved.
Within this overall context, the JRC-Ispra originally developed from 2001 the STRESA (Storage of Reactor
Safety Analysis Data) [45] web-based informatics platform (http://asa2.jrc.it/stresa) in order to provide a
secure repository of JRC Reactor Safety Facilities as well as COMETA code databases exploiting modern
computer information technologies for access and retrieve of the data.
Some time later and in the field of Severe Accidents the European Commission EURSAFE (EUROPEAN
NETWORK FOR THE REDUCTION OF UNCERTAINTIES IN SEVERE ACCIDENT SAFETY ISSUES)
[11] project included also a JRC developed web-based informatics platform (http://asa2.jrc.it/eursafe), which
preserves database and information about the main European severe accident facilities currently working.
17
18
Assessment of Size Aspects in Modelling Molten Fuel Coolant Interaction
Chapter 4 - OBJECTIVES
4. OBJECTIVES
The work has been focused in the prediction through computer codes of the progression and consequences of
severe accidents in water cooled reactors.
Taking into account the State of the Art of severe accident phenomena and particularly Melt Fuel Coolant
Interaction, FARO facility experiments and COMETA code calculations have demonstrated to be worth
mentioning in MFCI experience.
The main objective of the PhD research was focused in expanding general knowledge in Melt Fuel
Coolant Interaction by means of the experience obtained in the three last FARO facility tests carried out at
the Joint Research Centre of the Ispra site. The knowledge was complemented with the pre and post-test
analytical activities based on the development and application of the COMETA code to the prediction of
tests performed in the FARO facility. Also the application of the code to a full plant hypothetical severe
accident increased MFCI knowledge. A further analytical study was carried out in order to illustrate the
MFCI inside the general overview of a NPP severe accident sequence. Results were expected to be helpful in
order to clarify the importance of the melt fuel-coolant interaction in the context of a severe accident.
To fulfil the above mentioned general objective some detailed goals were established as follows:
- Collaborating and complementing the application of COMETA code to the prediction of tests performed
in the FARO facility under new conditions not experimented before (subcooled and triggered conditions).
- Verifying the code consistency, analysing and unifying the COMETA simulations carried so far.
- Developing and improving COMETA code. Both the application of COMETA code to the prediction of
tests performed in the FARO facility and the identification of problems in scale-up to reactor configurations,
larger than scaled test facilities, leaded to find COMETA code lacks and improvements. Solving this issue is
part a matter of engineering analyses and part a matter of computer programming analyses.
- Application of the RELAP5/SCDAP 3.2 thermalhydraulic code in order to obtain a deeper understanding
of the code MFCI modelling theory and its possible relationship with other detailed MFCI thermalhydraulic
codes.
The following research activities were the tools to develop the above objectives:
Test L-29 was the first test performed in subcooled conditions. The general objective of the analysis was to
investigate the behaviour of the facility for the selected test conditions and for a variety of parametric
variations in order to provide reference information for test planning and execution. More specifically, the
analysis aimed at providing a benchmark prediction for the verification and validation of the COMETA
code.
Test L-33 was the third test performed in subcooled conditions and the first test including triggering. In this
test an external trigger was applied to enhance the possibility of an energetic melt coolant interaction. In the
case of failed triggering the test would have been considered as another quenching test.
Moreover in order to verify the consistency, analyse and unify the COMETA simulations carried so far, it
was decided to simulate again all the FARO experiments with the last available version of the COMETA
code and keeping as far as possible the same basic nodalization scheme for all the tests. Only changes due to
different boundary conditions of facilities arrangements were performed. Tests calculations L-14, L-19, L24, L-27, L-28, and L-31 were simulated again.
In order to study the phenomenological behaviour of the thermalhydraulic system in a large-scale accident,
the work was focused in the simulation (with COMETA thermalhydraulic code) of an extended MFCI
accident in a hypothetical reactor with arbitrary geometry. The objective was to correlate the initial
19
Assessment of Size Aspects in Modelling Molten Fuel Coolant Interaction
Chapter 4 - OBJECTIVES
conditions of the reactor with the phenomena evolution. The application of COMETA code to a full plant
severe accident was not developed before. The reactor geometry chosen was typical of the Spanish reactor
ASCO-1, a 3-loops 966 MWe (2686 MWth) Westinghouse PWR.
Finally an analytical activity reproducing a severe accident following a LOCA break in a generic PWR NPP
(3 loops, 1000 MWe) with the RELAP5/SCDAP 3.2 thermalhydraulic code was performed in order to obtain
a better understanding of the MFCI modelling theory of RELAP5/SCDAP 3.2, to study the influence of the
“low pressure injection system” (LPIS) in the LOCA sequence and to obtain the initial conditions just before
the molten pool slumping into the lower plenum in order to introduce them in a specific MFCI code input so
a detailed MFCI study could be later performed.
20
Assessment of Size Aspects in Modelling Molten Fuel Coolant Interaction
Chapter 5 - THE FARO TEST FACILITY
5. THE FARO TEST FACILITY
5.1. INTRODUCTION
FARO (Furnace And Release Oven) was an experimental facility operated by the Institute for Systems,
Informatics and Safety (ISIS) at the JRC (Joint Research Centre) Ispra site of the European Commission.
The primary objectives of the FARO research program carried out were aimed at:
a) The acquisition of a reference experimental data base from tests performed in the FARO installation with
realistic melt composition and under reactor typical accident conditions. The experimental reference is
essential for the development and improvement of analytical models and the assessment of system codes
used in LWR safety analysis.
b) The investigation of basic phenomena relevant to the progression of severe accidents in water cooled
reactors with particular emphasis on the interaction of molten fuel with coolant and/or structures under both
in-vessel and ex-vessel postulated severe accident conditions.
The test facility became operational in 1987 and was initially dedicated to the investigation of severe
accident phenomena in liquid metal fast breeder reactors such as melt relocation and molten fuel-sodium
interaction. Thereafter, the test facility was reconfigured to address emerging safety issues pertinent to fuel
coolant interaction and melt quenching in light water reactors (LWR). FARO facility began the experiments
or tests of the LWR-MFCI phenomena in 1990 in collaboration with several reactor safety research
organizations from EC member countries and with the participation of the US Nuclear Regulatory
Commission (USNRC) in the context of a Technical Exchange Agreement established with the Commission.
The interaction of large masses of prototypical corium melt mixtures (e.g.; UO2/ZrO2, UO2/ZrO2 /Zr) in the
water under a variety of realistic accident conditions was studied. The reference scenario is relevant to a
postulated in-vessel core melt down accident when jets of molten corium penetrate into the lower plenum
water pool, fragment and settle on the lower head. Test conditions are such as system pressure, water
subcooling, water pool depth and melt composition.
In the context of the investigation relevant to the fragmentation and quenching of molten material into the
water coolant at different initial pressure and water subcooling, 12 quenching tests were performed in the
FARO facility (Figure 5.1): 5 at 50 bar initial pressure, 1 at 20 bar and 6 tests at pressure lower than 5 bar. In
the last test L-33 performed in July 1999 an external trigger was applied.
Another configuration was used, SARCOFAGO, (Figure 5.2) to investigate the impact on the core catcher of
corium ejected after reactor pressure vessel failure during a core meltdown accident. The way melt spreads
on the core catcher surface is important because of its effect on the long-term coolability of the melt. Two
tests have been performed in the FARO facility, one with a dry surface and one with 1 cm of water layer.
The test facility is properly instrumented in order to characterize the evolution of the interaction process.
21
Assessment of Size Aspects in Modelling Molten Fuel Coolant Interaction
Chapter 5 - THE FARO TEST FACILITY
Figure 5.1 - Outside view of FARO Test facility with FAT release vessel for melt quenching experiments [From
[25]]
Figure 5.2 - The FARO SARCOFAGO test vessel for melt spreading experiments [From [25]]
22
Assessment of Size Aspects in Modelling Molten Fuel Coolant Interaction
Chapter 5 - THE FARO TEST FACILITY
5.2. FARO FACILITY COMPONENTS
In its initial configuration (Figure 5.3), the FARO test facility comprised five main major components that
include the furnace, the intersection valve unit, the release vessel or melt catcher, the interaction test section
and the venting system. From Test L-14 to L-24 the TERMOS interaction test section was used (0.71 m of
diameter and about 1.5 m3 of volume) (Figure 5.3). From Test L-27 it was changed to FAT test vessel
(Figure 5.4), which consists of a pressure vessel of 1.5 m internal diameter and 2 m high, designed for a
pressure of 8 MPa and a temperature of 300°C. To compare the FAT results with the TERMOS results, an
internal cylinder was inserted in the FAT vessel. The cylinder internal diameter was the same of the previous
TERMOS vessel, 0.71 m. This cylinder is filled with water and the outer annular space is part of the free
board volume.
The FARO furnace consists of a pressure container, a melt mixture container, two electrodes and the release
tube. A quantity in the order of up to 200 kg of prototypic oxide fuel melts can be produced in the FARO
furnace possibly mixed with metallic components. Melt is generated by direct heating of a UO2/ZrO2
granulated variable mixture (in general consisted of 80% weight of UO2 and 20% weight of ZrO2) compacted
between the electrodes. The lower electrode is provided with an orifice in the centre to transfer the melt
through the release tube into the release vessel; during the heat up and melting phase, the orifice is closed by
a tungsten disk mounted on the graphite cap of the lower electrode. Relatively large masses of oxide fuel
type melts (up to 200 kg and 3000 °C) can be generated in the FARO furnace; as appropriate, the melt can be
eventually mixed with metallic components in the release vessel. The heating process is depending on the
mass quantity but generally is of about some hours. The corium is melted at low pressure (0.1-0.2 MPa)
while the pressure and temperature in the test section TERMOS or FAT is as required by the experiment.
The intersection valve unit provides the means to isolate the furnace from the test section during the
interaction phase. It consists of two slide valves, which are closed sequentially after the melt is transferred
into the release vessel. An optical device is mounted in the upper part of the inter-section valve unit to survey
the tungsten disk and to qualitatively monitor the melting process.
The release vessel is located inside the dome shaped upper head of the TERMOS or FAT test vessel. It is
designed to hold the melt for the time necessary to isolate the furnace and balance the release vessel and the
TERMOS or FAT test vessel pressures thus ensuring a gravity release of the melt. The release vessel is
equipped with two flaps; the lower flap allows the discharge of the melt whereas the side flap ensures against
built up of pressure differences between the release vessel and the test vessel during the release phase. The
release orifice depending on the test was of diameter 5 cm or 10 cm.
At the end of the corium melting phase the melt is released from the furnace to the release vessel via the
release tube and the valve S02 is closed. The release vessel is used as lock-chamber for pressure
equalization. This is obtained bursting a double disk mounted on the communication line, which acts as a
quick opening valve. Upon pressure balancing, the release vessel hinged-flap automatically opens, and the
melt is delivered by gravity.
The interaction test section TERMOS test section consists of a pressure vessel designed for a pressure of 10
MPa at a temperature of 300°C and a debris catcher mounted on the lower part of the vessel. From Test L-27
it was changed to FAT test vessel (Figure 5.4), which consists of a pressure vessel of 1.5 m internal diameter
and 2 m high, designed for a pressure of 8 MPa and a temperature of 300°C and a debris catcher mounted on
the lower part of the vessel, the rest of the original facility was maintained. The test vessel is heated from the
outside by trace heaters in order to establish the initial test conditions and is thermally insulated to minimize
heat looses to the environment. It is connected to a steam-water separator during all the phases of the test.
A venting system connects the steam-water separator to a condenser through a set of four pressure relief
valves to accommodate over-pressures in excess of the test vessel design pressure.
23
Assessment of Size Aspects in Modelling Molten Fuel Coolant Interaction
Chapter 5 - THE FARO TEST FACILITY
Three cameras are placed outside the test vessel to film the initial instants of the melt release and water
interaction. These films are very important to analyse interaction phenomena and melt jet shape, melt
fragmentation, etc.
A total of about 250 signals are generally recorded during a typical FARO test. They include pressures and
temperatures both in the steam and water regions, vessel wall and debris catcher bottom plate temperatures
as well as dedicated sensors for level swell measurement and hydrogen detectors.
5.3. SIMULATION OF ACCIDENT PHENOMENA IN THE FARO FACILITY
The severe accident phenomena can be divided in some stages, which are reproduced in the FARO facility:
The core molten material of the reactor (melt) pours by gravity into the lower plenum, in this first stage in the
FARO facility melt is poured in jet shape through an orifice and falls by gravity in the gas space.
The jet enters into lower plenum water pool or in the FARO release vessel. The jet into the water fragments
in little drops and melt fragments (debris) fall to the vessel bottom (debris catcher), at the same time in the
vessel bottom a solid and dense mass (cake) is accumulated. This stage is characterized by high pressure and
temperature increase and high heat exchange between drops and water, which is the quenching rate. The
high quenching rate provokes void fraction and vessel mixture level increase.
Last stage is a medium and long term cooling, the high void fraction leads to a fragmentation (or drops
production) decreasing, quenching rate, pressure and temperature increase become stable. Drops fall to the
vessel bottom and cake increases in the debris catcher.
The FARO test objectives were to evaluate the steam generation rate associated to the melt heat transfer
(quenching), to evaluate the hydrogen production associated to Zr oxidation, the heating of the bottom vessel
structures and to evaluate the debris bed structure in the bottom plate.
The FARO tests with pure oxidic UO2/ZrO2 melts are characterized by a considerable amount of hydrogen
generation which could be the result of either water dissociation at high temperature, reduction of UO2/ZrO2
during melt generation and subsequent oxidation in contact with steam/water, oxidation of UO2 to U3O8,
oxidation of the vessel material, a combination of the above or any other still unknown reason. Although
pure oxidic melts such as UO2/ZrO2 or Al2O3 should form stable chemical compounds, at high temperatures
chemical reduction reactions could result in changes to their molecular composition. In the performed FARO
tests the quantity of hydrogen generated ranges between 1.2 and 3.4 g per kg of mixture.
Evidence of hydrogen generation in FARO has been given by mass spectrometer analysis; for some tests a
quantitative evaluation has been performed comparing the saturation pressure to the actual system pressure
or by venting the mixture of steam-argon-hydrogen from TERMOS into the condenser and evaluating the
amount of hydrogen through the pressure increase in the condenser under the assumption of homogeneous
conditions at the mean pressure of discharge. The evaluation of hydrogen generation in all FARO tests using
the two methods has been performed.
24
Assessment of Size Aspects in Modelling Molten Fuel Coolant Interaction
Chapter 5 - THE FARO TEST FACILITY
FARO furnace
Lower electrode
Release tube
closing disc (W)
γ1, 2 detectors
Release tube
(Ø = 50 mm, h = 2.5 m)
A
Mirror system drive
Videocam
Depressurizer
B
Pressure equalisation
for melt release
V.420
(Ar)
Protection valve S01
Main isolation valve S02
(Ø =120 mm)
Steam venting
Pressure equalisation
during quenching
3645
(for 150 kg)
Inj level
Dome
Release vessel
(volume up to S02 = 0.056 m3)
Melt
Instrumentation ring
Hinged-flap for melt release
(Ø nozzle= 100 mm)
W level
TERMOS vessel
(Ø
= 800 mm)
ext
(Ø
= 710 mm)
int
Water
Heating sections
Debris catcher
(Ø = 660 mm)
Elevation (mm)
0.00
-240
-390
Bottom plate
(thickness = 40 mm)
Figure 5.3 - FARO Test facility with TERMOS release vessel [From [17]]
25
Assessment of Size Aspects in Modelling Molten Fuel Coolant Interaction
Chapter 5 - THE FARO TEST FACILITY
Water
level
Figure 5.4 - FARO Test facility with FAT release vessel [From [20]]
26
Assessment of Size Aspects in Modelling Molten Fuel Coolant Interaction
Chapter 6 - THE COMETA CODE
6. THE COMETA CODE
6.1. INTRODUCTION
COMETA (COre MElt Thermalhydraulic Analysis) is a coupled thermalhydraulic and melt fragmentation
code specifically conceived for the computational analysis of melt coolant interaction and quenching
processes in LWR severe accident conditions.
It was written at the JRC Ispra in order to analyse with sufficient detail both the thermal-hydraulics and the
fuel fragmentation phenomena of melt quenching tests as executed and/or planned in the FARO facility,
including all the facility components considered necessary. Thus the immediate objective of the code is the
prediction of the thermalhydraulic behaviour of FARO facility for design, definition of operational
procedures and tests interpretation, and the long term objective of the code is to predict the behaviour of
MFCI phenomena, including steam explosion events in any geometry 1d or 2d.
The overall structure of the COMETA code is complemented by the WinGraf data processing package.
The current version of the code has been extensively assessed against the FARO experimental data for the
analysis of the integral aspects of melt-coolant interaction and against the LOBI test data for the verification
of the related thermal-hydraulic. Also lately the code was also applied to real vessel reactor geometries.
The code runs both on a personal computer and on a workstation and it produces on-line plots (up to 9) and
automatic nodalization drawing schemes showing volumes void fractions.
COMETA is structured in a thermalhydraulic two-phase flow model resolved in Eulerian coordinates and a
melt fuel fragmentation model resolved in Lagrangian Coordinates.
The two-phase flow field is organized in a number of lumped volumes connected with junctions. A 2D
nodalization can be built up connecting a number of macro-volumes (containing radial and axial volumes)
and macro-junctions. Thermalhydraulic components, valves, pumps, separators and accumulators can be
defined in order to represent the overall system configuration.
The melt field is described by the jet, the drops and the fused debris bed.
COMETA includes a model for hydrogen generation from metallic components based on validated
correlations and also an empirical model for hydrogen generation from oxidric melts based on the
experimental evidence provided by the FARO experiments.
6.2. THERMALHYDRAULIC MODEL
The two-phase flow field is described by “6+n” equations: mass, momentum and energy conservation
equations for the liquid and vapour phases and mass conservation equations for “n” non-condensable gas
present in the mixture. The model is derived for a 1d system (x direction) and is extended to 2d with use of
radial junctions.
a) Conservation Mass equations for the liquid, vapour and non-condensable phases:
Liquid phase:
∂ (α f ρ f
∂t
)
+
1 ∂ (v f α f ρ f A)
= −Γg (1)
A
∂x
27
Assessment of Size Aspects in Modelling Molten Fuel Coolant Interaction
Chapter 6 - THE COMETA CODE
Steam phase:
∂ (α g ρ g )
∂t
+
1 ∂ (v g α g ρ g A)
= Γg (2)
A
∂x
where:
g (subscript) = vapour
f (subscript) = liquid
a = void fraction
? = density (kg/m3)
v = velocity (m/s)
A = flow area (m2)
t = time (s)
x = coordinate (m)
? = volumetric vapour generation rate (kg/m3s)
The equations considered in the code are the sum of (1) and (2) leading to the mixture conservation equation:
(1)+(2)
∂ (α g ρ g + α f ρ f
)
∂t
+
1 ∂ (v gα g ρ g A + v f α f ρ f A)
= 0 (3)
A
∂x
and the difference between equations (1) and (2). The choice of the sum and difference equations is in order
to prevent a zero row in the resolution matrix when only one phase is present in a certain volume:
(1)-(2)
∂ (α g ρ g − α f ρ f
∂t
)
+
1 ∂ (v g α g ρ g A − v f α f ρ f A)
= 2Γg (4)
A
∂x
The mass conservation equations are integrated over a volume which is connected to other volumes via
junctions that can be located at the inlet or outlet side (Figure 6.1 a) or at internal or external side in case of a
2D arrangement (Figure 6.1 b). The volume is assumed to have constant flow area.
28
Assessment of Size Aspects in Modelling Molten Fuel Coolant Interaction
Chapter 6 - THE COMETA CODE
Figure 6.1 - Integration volumes [From [6]]
The non-condensable mass conservation equation can be written as:
(
∂ α nci ρ nci
∂t
)
+
(
)
1 ∂ v g α nci ρ nci A
= S nci (5)
A
∂x
for nci non-condensable gas
where:
S = volumetric source for non-condensable gas (kg/m3s)
b) Conservation Momentum equations for the liquid and vapour phases:
The steam and liquid momentum equations are defined at the junctions and also in this case are the sum and
difference forms. The momentum conservation equations are integrated between the centres of the volumes
connected by the junction, using the donor cell technique to define the properties at the junctions.
Sum equation:
α g ρg
∂v g
∂t
+α f ρ f
∂v 2f
∂v g2 1
1
∂P
=−
+ ρ g − α g ρ g FWGvg + α f ρ f FWFv f − Γg (v g − v f
+ α g ρg
+ αf ρf
2
∂x 2
∂x
∂x
∂t
∂v f
(6)
Difference equation:
∂vg
∂t
−
∂v f
∂t
+
2
2
 1
Γ
Γ
1 ∂vg 1 ∂v f
1  ∂P
−
= − −  − FWGvg + FWFv f + g (vI − vg ) − g (vI − v f ) − (ρFI )(vg − v f
ρ

2 ∂x 2 ∂x
α g ρg
αf ρf
 g ρ f  ∂x
(7)
where:
g (subscript) = vapour
f (subscript) = liquid
P = pressure (Pa)
29
)
Assessment of Size Aspects in Modelling Molten Fuel Coolant Interaction
Chapter 6 - THE COMETA CODE
FWF, FWG = wall drag coefficients (1/s) for liquid and vapour
ρ = mean density (kg/m3) = [ a ?g + (1- a ?f)]
? = interphase density (kg/m3) = [ aI ?g + (1- aI ?f)]
FI = interphase liquid/vapour friction factor (m3/kg s)
vI = interphase velocity (m/s)
c) Conservation Energy equations for the liquid and vapour phases:
The energy conservation equations are kept separately for void fraction a = 0 or a = 1. In case of single
phase the energy equation is substituted by the saturation equation. The energy conservation equations are
also integrated over the control volumes.
In general, for the vapour phase:
∂ (α g ρ gU g )
∂t
+
∂α g P ∂ (α g v g A)
1 ∂ (α g ρ g v gU g A)
= −P
−
+ Qwg − Qwcond + Qig + Γg wall hg + Γg bulk h g + DISS g
A
∂x
∂t
A
∂x
(8)
For the liquid phase:
∂ (α f ρ f U f
∂t
)
+
∂α f
1 ∂ (α f ρ f v f U f A )
P ∂ (α f v f A )
= −P
−
+ Q wf − Q wboil + Q if + Γg wall h f + Γg bulk h f + DISS f
A
∂x
∂t
A
∂x
(9)
where:
g (subscript) = vapour
f (subscript) = liquid
U = internal energy (J/kg)
Qif = interphase heat transfer rate to liquid (W)
Qig = interphase heat transfer rate to gas (W)
Qwboil = wall boiling heat transfer rate (W)
Qwcond = wall condensation heat transfer rate (W)
Qwf, Qwg = wall heat transfer rate to liquid/gas (W)
h = enthalpy (J/kg)
DISS = dissipative terms in energy equation (W)
30
Assessment of Size Aspects in Modelling Molten Fuel Coolant Interaction
Chapter 6 - THE COMETA CODE
The unknown variables are:
- pressure
- void fraction
- gas and liquid velocities
- gas and liquid internal energies
The resolution is obtained by approximating the differential equations with finite differences and inverting
the resultant matrix by a Gaussian elimination method. A system is used to reduce the initial 6n x 6n matrix
to a n x n matrix by appropriate row linear combinations. The solution is therefore limited to the pressure
solution that is then back substituted to determine the other unknowns.
The mass error is controlled via the time step reduction according to the actual error. The mass error is
defined as the ratio between the mass obtained by integration of the partial differential equations and the
mass obtained by the new temperature, pressure and void fraction values.
Additional relations are needed to characterize two-phase flow.
For the vapour generation rate the following formulation is used:
Γg = Γgbulk + Γg wall (10)
where:
Gg = volumetric vapour generation rate (kg/m3 s)
The term Ggwall is the vapour generated directly at the wall by wall heat transfer and it is determined
according to the heat transfer mode.
The term Ggbulk is the vapour created in the bulk fluid due to the heat exchange between the liquid and the
steam phase (depressurization, steam superheating or liquid subcooling):
hig S ig
Γg bulk = −
V
(T
sat
)
− Tg −
hif S if
hg − h f
V
(T
sat
−Tf
)
(11)
where:
if = liquid interphase
ig = gas interphase
sat (superscript) = saturation
T = temperature (K)
The terms hS/V are the heat transfer coefficient per unit volume and unit surface of interphase, Ci. A simple
correlation is adopted for Ci:
where:
C = heat transfer coefficient per unit volume (W/m3K)
if the void fraction is lower than 0.25
31
Assessment of Size Aspects in Modelling Molten Fuel Coolant Interaction
Chapter 6 - THE COMETA CODE
(
)
C if = max 10 8 α g ,1.5 ⋅ 10 5 (T f − Tsat )
C ig = 10 α g
8
(12)
if the void fraction is higher than 0.75
C if = 10 8 α f
(
)
C ig = max 10 8 α f ,1.5 ⋅ 10 5 (Tsat − Tg )
(13)
a constant value Cig=Cif=Ci=2.3 105 is used in case the void fraction is between 0.25 and 0.75.
For the interphase friction factor a model similar to the one present in RELAP5/MOD2 code is adopted.
Radial junctions can be specified in order to set up a 2D nodalization. The radial junctions are characterized
by a slightly different momentum equation in which the gravity term is not considered and the interphase
friction factor is minimized in order to allow mixing in the horizontal connected zones.
Heat slabs with different material composition can be specified in which the conduction and therefore the
internal temperature distribution is accounted for. The Fourier equation is resolved in plane, cylindrical or
spherical geometry and the geometry can be selected by input.
The heat exchange between structures (slabs, jet, drops, debris) and the coolant is described by the following
correlations:
Single phase liquid convection
Condensation model
Nucleate boiling
Transition/film boiling
Single phase vapour convection
Dittus Boelter
Collier
Chen
Bromley-Pomeranz
Dittus Boelter
Additional models regard separator model and valve discharge/control model. These models allow the
description of a complete test facility like FARO and the representation of a full complex transient.
6.3. MELT FUEL FRAGMENTATION MODEL
The melt field is composed of three phases for the description of melt: jet, drops and fused debris bed also
called cake (Figure 6.2).
The melt is released from a tank (melt catcher) with an orifice and is fragmented during the fall keeping
constant the ratio L/D, L is the distance from the injection orifice to the point where jet disappears and D is
the initial diameter. The melt height is calculated at each time step according to the real actual dimensions of
the melt catcher; it was found that the specification of the real dimensions of the melt catcher is important for
the derivation of the discharge velocity:
Vrelease (t ) =
2 ghmelt (t )
(14)
K or
where:
hmelt = Melt height in the melt catcher (m)
Kor = Melt orifice discharge coefficient (-)
32
Assessment of Size Aspects in Modelling Molten Fuel Coolant Interaction
Chapter 6 - THE COMETA CODE
g = gravity acceleration (9.81 m/s2)
The melt is assumed to be released in the form of a coherent jet, which is conical shaped with a wavy and
rough surface.
melt
jet
drops
debris + cake
Figure 6.2 - Schematic melt field release description
a) Jet description:
Three models for jet fragmentation and fragments creation are included: the original COMETA model, based
on the Jet Break-up Length concept, the Corradini-Tang model, similar to the model present in the TEXAS
code (Chu and Corradini, 1989) and the IKEJET model (Bürger et al., 1995) developed by the University of
Stuttgart.
The original COMETA model for jet melt fragmentation and erosion is based on an interpolated Jet Breakup Length criterion (Figure 6.3) with L/D evaluated at each position and time step. The local erosion rate is
function of L/D.
33
Assessment of Size Aspects in Modelling Molten Fuel Coolant Interaction
Chapter 6 - THE COMETA CODE
Figure 6.3 - Jet Break-up Length definition [From [6]]
The jet accelerates along its descent and reduces its size. Meanwhile some particles leave the jet surface at
.
different axial positions. The erosion rate m is established to satisfy some of the available correlations for the
Jet Break-up Length.
.
vjρ j
md
(15)
=
Sj
L
2 
 D
 L 

 = f (We, Fr , x,...) (16)
D 
 j
where:
j (subscript) = melt jet
? = density (kg/m3)
v = velocity (m/s)
S = surface (m2)
L = length (m)
D = diameter (m2)
We = Weber number
Fr = Froude number =
v 2j
gD j
The correlation actually used in COMETA is a combination of two correlations: the Saito [14] correlation
and the Epstein-Fauske [44] correlation.
The Saito L/D correlation is:
34
Assessment of Size Aspects in Modelling Molten Fuel Coolant Interaction
Chapter 6 - THE COMETA CODE
ρ j 0.5
L
= 2 .1
Fr (17)
D
ρc
where:
j (subscript) = melt jet
c (subscript) = coolant
? = density (kg/m3)
The Epstein-Fauske correlation for L/D is independent from the jet velocity:
L
3  ρ v  ρ j
=
1+
D
2  ρ j  ρ v
(18)
The combined correlation is defined (Figure 6.4):
1 - Saito correlation is used up to Weber number 50, Epstein-Fauske above 100 and an interpolation in the
transition between the two.
2 - In the two phase flow the Saito correlation is used for each Weber number.
Figure 6.4 - Jet Break-up Length interpolated model [From [6]]
Both the jet and the drop particles released are moved in a Lagrangian framework of equations. The position
is tracked along the transient and the corresponding Eulerian cell is identified and receives the heat source
from the jet and drop particles.
35
Assessment of Size Aspects in Modelling Molten Fuel Coolant Interaction
Chapter 6 - THE COMETA CODE
The jet equations are represented by:
Jet mass conservation equation:
dM j
dt
.
= − m d (19)
Jet momentum equation:
 dv j 
 dt 


ρ jV j 
(
)
= V j ρ j − ρ g − F j (20)
n
Fj =
(
)
1
(cd W j )ρS j v j − v | v j − v | (21)
8
Wj is a multiplier that takes into account the surface roughness; at present however Wj=1.
Jet energy conservation:
ρ jV j c p j
dT j
dt
= (q" jf S jf + q" jg S jg )W j (22)
where:
j (subscript) = melt jet
g (subscript) = vapour
f (subscript) = liquid
q" = heat flux (W/m2)
M = mass (kg)
V = volume (m3)
T = temperature (K)
Wj = jet roughness parameter (-)
cp = Specific heat (J/kg K)
cd = Drag coefficient (-)
b) Drops description:
The melt drops are created with an initial diameter, which can be progressively halved to satisfy the Weber
number criterion: i.e. the Weber number has to be lower than 12.
COMETA standard model is the fragmentation model with coefficients 6 for the diameters of the drops (it
multiplies the equilibrium Weber number 12, and influences the final size of the fragments) and no multiplier
for the Jet break-up length.
36
Assessment of Size Aspects in Modelling Molten Fuel Coolant Interaction
Chapter 6 - THE COMETA CODE
Up to 500 groups of drops characterised by diameter, mass, temperature, velocity and position can be
accounted for. A unique drop temperature is assumed in each group. The residual part of the jet reaching the
vessel bottom plate forms a fused debris bed cake, which can agglomerate eventual drops arriving with
higher temperature than the solidification one.
If the drops are still liquid the diameter is controlled at each time step and if the Weber number increases
above the limiting value (due to drops acceleration or different medium density) and the temperature is
greater than solidification, the diameter is halved to satisfy Weber criterion.
The governing equations are:
Drop momentum equation:
 dvd 

 dt 
ρ d Vd 
(
)
= Vd ρ d − ρ g − F (23)
n
(
)
1
F = c d ρ S d v d − v | v d − v | (24)
8
The value of cd is obtained by a relation containing the Reynolds number.
Drop energy equation:
ρ d Vd c pd
dTd
= (q"df S df + q"dg S dg ) (25)
dt
where:
d (subscript) = melt drops
A unique drop temperature is adopted. This hypothesis is sufficient for small drops (<2mm) but it is
probably not correct for larger drops. The only reason not to calculate the real temperature distribution is to
avoid excessive iterations of the Fourier equation between time steps.
c) Fused-debris description:
The residual part of the jet that reaches the vessel bottom relocates in the so-called fused-debris bed. In this
context the fused-debris bed is considered as the fused part of debris bed that represents only a portion of
what is normally intended as debris bed, therefore not including the fragments.
The equations considered are:
Mass conservation equation:
dM deb
= M j( Z =0 ) + M d ( Z =0 ,T >T ) (26)
d melt
dT
Energy conservation equation:
ρ debVdeb C pdeb
dTdeb
= q"debf S debf + q"debg S debg (27)
dt
37
Assessment of Size Aspects in Modelling Molten Fuel Coolant Interaction
Chapter 6 - THE COMETA CODE
where:
deb = melt fused-debris
The surface area of the fused-debris is assumed to be equal to 2.5 times the area of a hemisphere having the
same volume; it is also possible to define a flat cake. The energy conservation equation takes into account
the heat released to the coolant. Due to the different surfaces taking part in the heat exchange process during
the initial phase of the transient the heat transferred from the fused-debris is negligible, while in the long
term, after the quenching of the small drops, it becomes important for the total energy balance.
A model for H2 generation based on the available end-of-test data of FARO tests was introduced into the
code after L-20 post-test calculations. A first simple model is based on the consideration that oxidation
potential is proportional to fragmented mass:
dM H 2 / dt = 0.003dM fragm / dt [kg / s ] (28)
The second model takes into account the surface generation and the chemical kinetics:
dM H 2 / dt = d ( S 2k H 2 t ) / dt [kg / s ] (29)
and relates the rate of H2 mass production to the fuel surface S generated by the interaction and the reaction
kinetic K H 2 that is obtained from the O2 kinetic using the relationship:
K H 2 = K O2 / 64 [kg2m-4s-1] (30)
where the oxygen kinetic law is an exponential one:
 −b
K O2 = a exp
 (31)
 RT 
The constants a and b are related to the type of the material considered; code uses the values present in the
RELAP5/SCDAP MOD2 for metal oxidation. The maximum H2 mass produced is limited to the
experimental value, that is 2 10-3 kg of H2 per kg of fragmented mass.
COMETA code includes the possibility to simulate vapour explosions. The threshold for the initiation of an
explosive interaction in COMETA is based on a cell pressurization rate higher than 2 MPa/s; when this value
is attained either in a spontaneous or triggered fashion, the melt present in the interested cell is fragmented in
finer particles with a consequent increase of heat transfer and steam generation rate, enhancement of
pressurization which eventually propagates to adjacent cells escalating the interaction.
The coupling between the fragmentation models and the thermal-hydraulic models is explicit, but the time
step chosen is the minimum between the requirements of both fields. In particular the thermal-hydraulic
models use as boundary condition the power transferred by the fuel fragments into the coolant, while the fuel
fragmentation model use the local density, temperature and void fraction to calculate the fragmentation rate
and the heat transmitted to the coolant. The volume occupied by the fuel is generally small compared to the
coolant volume and therefore no volume reduction is taken into account. However some particular
circumstances (as for example some KROTOS tests) could require taking into account the volume reduction.
38
Assessment of Size Aspects in Modelling Molten Fuel Coolant Interaction
Chapter 7 - RESEARCH ACTIVITIES SUMMARY
7. RESEARCH ACTIVITIES SUMMARY
The PhD research activities were mainly concentrated in severe accident simulations with JRC-COMETA
code, in one hand performing pre and post-test analysis of the FARO tests, in another hand performing code
calculations applied to a large NPP reactor configuration. The application of RELAP5/SCDAP to a general
NPP configuration completed the research activities.
Two important tests were performed in the FARO facility in 1998 and 1999: L-29 and L-33 under new
conditions not experimented before (subcooled water for L-29 and triggered conditions for L-33). COMETA
pre-tests calculations are explained in chapters 7.1 and 7.2. In addition to give information for test
experiment planning and execution in the FARO facility, the calculations gave results to improve COMETA
code logic.
Chapter 7.3 summarizes the application of the COMETA code to reactor configurations, larger than scaled
test facilities. This application was not performed before. The results lead to the identification of scale-up
problems in the logic of the COMETA code. They also solved and explained one of MFCI related
phenomena: the relation between the void fraction and the quenching rate.
To analyse and unify the experiment simulations with the last available version of the COMETA code
another work is presented in chapter 7.4. Along the years the COMETA code was improved and changed to
fit experimental results. This purpose was fulfil simulating again tests from L-14 to L-31. This work also
gave some indications that help to choose the nodalization scheme in order to reduce the COMETA
computer calculation time.
In chapter 7.5 the calculations with RELAP5/SCDAP complete the general overview of the MFCI
phenomena included in the PhD research.
39
Assessment of Size Aspects in Modelling Molten Fuel Coolant Interaction
Chapter 7 - RESEARCH ACTIVITIES SUMMARY
7.1. COMETA PRE-TEST CALCULATION OF FARO TEST L-29
7.1.1.
INTRODUCTION
Test L-29 was scheduled to be performed in the FARO facility in July 1998. It was the second test performed
with the new FAT vessel. Test L-29 was the first test performed in the FARO facility in subcooled
conditions. The general objective of the analysis was to investigate the behaviour of the facility for the
selected test conditions and for a variety of parametric variations in order to provide reference information
for test planning and execution. More specifically, the analysis aimed at providing a benchmark prediction
for the verification and validation of the COMETA code.
There was a great interest in performing a subcooled test in the FARO test facility due to the potential for an
energetic melt coolant interaction. This is supported by the recent results of the KROTOS Test 58, in which,
for an UO2 / ZrO2 melt mixture, with conditions of 5 bar, 130°C subcooling a mild steam explosion occurred.
Since there were uncertainties in the real initial test conditions, determined by the peculiarity of the test
performed, a number of different hypotheses have been explored as initial conditions of the test, including
saturation conditions cases.
The following hypotheses have been considered:
Subcooled Test
Pressure:
Mass:
1, 2, 5 bar
30, 50, 70, 150 kg (only for comparison purposes)
Saturated Test
Pressure:
Mass:
2, 5 (only as a recalculation of L-28), 20 bar
150 kg
Additional cases have been analysed to highlight the influence of enhanced fragmentation model, H2
generation and the nodalization type (1d/2d).
Figure 5.4 shows the configuration of the facility for this test. An inner vessel with the same diameter as the
TERMOS will be installed inside the FAT vessel. The venting lines to the separator will be closed.
40
Assessment of Size Aspects in Modelling Molten Fuel Coolant Interaction
Chapter 7 - RESEARCH ACTIVITIES SUMMARY
7.1.2.
RESULTS OF SUBCOOLED TEST OPTION FOR THE FARO TEST L-29
Initial and boundary conditions
In order to cover the possible range in the experimental conditions, calculations have been performed with
different values of melt injected mass (Tmelt = 3073K) and 5 cm of discharge orifice. The first case,
L29PR30N should be seen as a lower bound and case L29PR150N should be seen as an upper bound.
The so-called standard fragmentation model has been used for this test, i.e.: Cd=6, Cj=1, fH2=1.810-3kgH2/kg
fragmented.
The following table summarises subcooled cases at 5 bar, 50°C and different melt mass:
Descriptor
Melt
Mass (kg)
L29PR30N
L29PR50N
L29PR70N
L29P150N
30
50
70
150
Melt
temperature
(K)
3073
3073
3073
3073
Initial
Pressure
(bar)
5
5
5
5
Initial water
level (m)
1.5
1.5
1.5
1.5
Initial water
temperature (K) and
subcooling
323 (100°C subc.)
323 (100°C subc.)
323 (100°C subc.)
323 (100°C subc.)
Fragmentation
Model
Stand.
Stand.
Stand.
Stand.
Table 7.1 - Performed L-29 pre-test subcooled calculations at 5 bar, 50°C
Other subcooled calculations have been performed to analyse effect of subcooling. Here, we consider four 50
kg melt mass cases with the subcooled water at different initial pressures and temperatures:
Descriptor
Melt
mass (kg)
L291BSUB
L292BSUB
L29P2B20
L29P5020
50
50
50
50
Melt
temperature
(K)
3073
3073
3073
3073
Initial
Pressure
(bar)
1
2
2
5
Initial water
level (m)
1.5
1.5
1.5
1.5
Initial water
temperature (K) and
subcooling
323 (50°C subc.)
323 (70°C subc.)
293 (100°C subc.)
293 (130°C subc.)
Fragmentation
Model
Stand.
Stand.
Stand.
Stand.
Table 7.2 - Performed L-29 pre-test subcooled calculations
Due to the amount of figures it is impossible to present here all the case results, only the main results
obtained in the calculations are graphically represented.
Results of the base case calculation (50 kg melt subcooled water)
Base case is L29PR50N: standard fragmentation model case of 50 kg melt mass at 3073K falling into 1.5 m
of subcooled water at 50°C at 5 bar.
The void fraction is not very high and is mainly due to the non-condensable gas (H2) and not to steam
(Figure 7.1). The low void fraction causes great fragmentation and therefore low cake mass accumulation
(Figure 7.2).
The jet leading edge position (Figure 7.3) shows that the melt falls down according to melt gravity release
with 1m/s of initial velocity until 0.7 m. What happens is that the jet parcels are released at progressive
increasing velocity from 0 to about 3 m/s. Some parcel can overcome the leading edge parcels, becoming
leading edge. Therefore, the leading edge can become faster than the theoretical gravity release with initial
velocity zero. This is a point that needs improvement in the code. It was not occurring with 10 cm jet
because the velocity increase was smaller due to higher inertia. After the height of 0.7 m the friction
becomes more and more important and deviation of the jet leading edge position from the theoretical gravity
discharge curve occurs.
41
Assessment of Size Aspects in Modelling Molten Fuel Coolant Interaction
Chapter 7 - RESEARCH ACTIVITIES SUMMARY
WinGraf - 07-06-1998
.40
.35
XXX
YYY
ZZZ
.30
L29PR50N
L29PR50N
L29PR50N
AVGVOIDML
AVGVOID10
VOID10
X
Void fraction (-)
.25
X
.20
XXXX
Avg. void fraction below mixture level
X
X
X
X
Y
Y
Y
Y
.15
Y
Y
X
Y
.10
Y
Y
.05
X
Y
Avg. void fraction in water zoneY
X
Y
X
Y
X
X
Y
X
Y
Y
Z
Z
Y
0 XY
ZZ ZZ
ZX
Z
Z
Z
Z
Z
Z
Z
Z
Z
Avg. void fraction of steam in water zone
Z
Z
X
Y
Z
Y
Z
-.05
0
2.0
4.0
6.0
Time (s)
8.0
10.0
12.0
Figure 7.1 - Global average void fraction below mixture level (AVGVOIDML), global average void fraction in
the water region (AVGVOID10) and average void fraction of steam in the water region (VOID10) in Test L-29
WinGraf - 07-06-1998
70.0
XXX
YYY
ZZZ
60.0
L29PR50N
L29PR50N
L29PR50N
50.0
Y
Y
Y
Y
Y
Y
Y
Y
Y
Drops
Y
Y
Y
Y
Y
40.0
Mass (kg)
TOTMJET
TOTMDROPS
AMDEBRIS
30.0
Y
20.0
Y
10.0
X
XX
YXX
Jet
Y
X
0 XY
ZY
ZZZZ Z
Cake
Z XZ X Z XZ X Z X Z X Z X Z XZ X Z X Z XZ X Z XZ
-10.0
0
2.0
4.0
6.0
Time (s)
8.0
10.0
12.0
Figure 7.2 - Jet, Drops, and Molten Cake mass in the base case calculation in Test L-29
42
Assessment of Size Aspects in Modelling Molten Fuel Coolant Interaction
Chapter 7 - RESEARCH ACTIVITIES SUMMARY
Jet behavior
2,5
Standard Model
Gravity release (V=0)
2
Height (m)
Gravity release (V=1)
1,5
Initial Water level
1
0,5
0
0
200
400
600
800
1000
1200
1400
1600
Time (ms)
Figure 7.3 - Jet leading edge in the base case calculation in Test L-29
Results of the bounding calculations
In order to highlight the possible minimum and maximum energy release in this test three additional
calculations have been performed with subcooled water (50°C) at 5 bar changing the quantity of melt mass
injected. The minimum case was performed reducing the injected mass to 30 kg. In the other cases the mass
was increased to 70 kg and 150 kg.
The mass reduction and increase of 20 kg do not show great differences. The increase in mass of 100 kg
shows a higher pressure increase and also the energy released is greater in the upper boundary case with 150
kg. As injected mass increases more fragmentation is occurring. The cake mass produced is very limited, less
than 5% of the total mass in the cases with 30, 50 and 70 kg. In the upper boundary case the cake mass
increases up to 18 kg (12% of the total mass).
Global average void fraction is mainly due to H2, it is higher as mass increases. Void fraction due to steam is
practically zero in all cases.
Effect of subcooling and initial pressure
In order to check the possible influence of initial subcooling the following 50 kg mass melt cases are
analysed. The base case L29PR50N (50°C at 5 bar, 100°C of subcooling), L291BSUB case (50°C at 1bar,
50°C of subcooling), L292BSUB case (50°C at 2bar, 70°C of subcooling), L29P2B20 case (20°C at 2bar,
100°C of subcooling), and L29P5020 case (20°C at 5bar, 130°C of subcooling).
Pressure is not affected by subcooling in the range 70°C to 100°C. The 100°C of subcooling case at 2 bar
pressure is similar than case with 70°C of subcooling. The same occurs with the two subcooled cases at 5
bar.
Global void fraction is affected by pressure but not by subcooling. In Figure 7.4 it is possible to observe
average void fraction of non-condensable gas (AVGVOISNC10) and of steam (AVGVOISST10) in the 10
lower volumes of FAT. Average void fraction of steam is very low in all cases; it increases as pressure and
43
Assessment of Size Aspects in Modelling Molten Fuel Coolant Interaction
Chapter 7 - RESEARCH ACTIVITIES SUMMARY
subcooling decrease (Figure 7.5). Average void fraction of H2 is higher and decreases as pressure increases;
However it is quite similar in the cases at the same pressure with different subcooling.
The fragmentation is very high in all cases and practically equal for the subcooled cases at the same pressure.
It increases as pressure increases, due to lower void fraction.
Quenching rate shows moderate decreasing with pressure; and moderate increase as initial water temperature
is lower at the same pressure, as subcooling increases. The quenching rate in the 1 bar subcooled case
becomes unstable due to CCFL phenomena. High oscillations in the void fraction of the lowest volume are
produced in the FAT vessel, which is continuously voided and refilled with water (Figure 7.6)
WinGraf - 07-07-1998
.80
.70
XXX
YYY
ZZZ
VVV
JJJ
HHH
###
OOO
AAA
BBB
.60
1 bar
2 bar
.50
Void fraction (-)
5 bar
.40
X
.30
X
ZJ
X
Z
J
.20
X Z
J
Z
#
XJZ
#A
#
A
.10
Z
#
A
XZJ#A##
Y
Y
#YY Y
J#A
Z
HV O B
HO
VHO
YO
A
VO
HOH
BB
#O
VVVV
OO
X
Y
BH
O
BB
0 X
VVVV
Y
#JO
H
ZZZ
X
X
Z J Z JZ
#A # A#
X
Z
J
Z
XJ Z
JXZ
A
# # #AY
# AY
Y
Y
Y
H V H
H
H
H
VOB O BOV OBV
O VO
B VO B V
X
ZJ
A
Y
H
V
B
L291BSUB
L291BSUB
L29P2B20
L29P2B20
L292BSUB
L292BSUB
L29P5020
L29P5020
L29PR50N
L29PR50N
AVGVOISNC10
AVGVOISST10
AVGVOISNC10
AVGVOISST10
AVGVOISNC10
AVGVOISST10
AVGVOISNC10
AVGVOISST10
AVGVOISNC10
AVGVOISST10
X
X
ZJ
AY
H
VB
JX
A
Y
H
B
X
X
J
J
J
A Y
H
B
A
J
Y
H
B
A Y A
H
H
B
B
-.10
avgvoid of steam
-.20
0
1.0
2.0
3.0
4.0
5.0
6.0
7.0
8.0
Time (s)
Figure 7.4 - Effect of subcooling and pressure on void fraction of non-condensable and steam in Test L-29
44
Assessment of Size Aspects in Modelling Molten Fuel Coolant Interaction
Chapter 7 - RESEARCH ACTIVITIES SUMMARY
WinGraf - 07-07-1998
.15
X
X
XXX
YYY
ZZZ
VVV
JJJ
.13
Void fraction (-)
.10
L291BSUB
L29P2B20
L292BSUB
L29P5020
L29PR50N
AVGVOISST10
AVGVOISST10
AVGVOISST10
AVGVOISST10
AVGVOISST10
.08
X
X
1 bar
2 bar
X
X
.05
.03
0
X
X
XX
XX
Z
X
Z
X
ZZ
X Z Z JYJY VJY
Z YJY
JY JYVJV VVV
YZVV
JX
Y
X VV
Z
YJXYZV
70°C sub
-.03
Z
Z
V
JY
Y
J
V
J
V
Z
Y
Z
Z
V
J
Y
V
J
VY
100°C sub
Z
Z
Y
J
V Y VJ
130°C sub
5 bar
Z
Z
Y
J Y
J
J
100°C sub
-.05
0
.50
1.00
1.50
2.00
2.50
3.00
3.50
4.00
4.50
5.00
Time (s)
Figure 7.5 - Effect of subcooling and pressure on void fraction of steam in Test L-29
WinGraf 2.4 - 06-30-1998
1.40
XXX
YYY
1.20
L291BSUB
L291BSUB
VOID001
ALFANC001
void fraction of steam
1.00
X
X
Void fraction (-)
X
.80
X
.60
.40
X
.20
Y
Y
Y
0 XY XXXX
YYY X X
Y
X
X
XY
Y
Y
X
Y
Y
Y
Y
X
Y
X
Y
Y
X
Y
X
Y
void fraction of non condensable
-.20
0
1.0
2.0
3.0
4.0
5.0
6.0
7.0
8.0
9.0
10.0
Time (s)
Figure 7.6 - Void fraction of steam and non-condensable in the FAT lowest volume for the 1 bar subcooled case
in Test L-29
45
Assessment of Size Aspects in Modelling Molten Fuel Coolant Interaction
Chapter 7 - RESEARCH ACTIVITIES SUMMARY
7.1.3.
RESULTS OF SATURATED TEST OPTION FOR THE FARO TEST L-29
The following table summarises three 150 kg melt mass cases (5 cm discharge orifice) with saturated water
at different initial pressures. It has been shown with Test L-28 prediction that is necessary to use the
enhanced fragmentation model (Cd=3 instead of 6, Cj=0.5 instead of 1) to overcome the increased voiding
caused by the lower pressure; therefore, the enhanced model has been used for these cases.
Descriptor
Melt
mass (kg)
L292BAR
L29P150S
L2920BAR
150
150
150
Melt
temperature
(K)
3073
3073
3073
Initial
Pressure
(bar)
2
5
20
Initial water
level (m)
1.5
1.5
1.5
Initial water
temperature (K) and
subcooling
Tsat.
Tsat.
Tsat.
Fragmentation
Model
Enhanced
Enhanced
Enhanced
Table 7.3 - Performed L-29 pre-test saturated calculations
Results of 150 kg mass saturation cases at different pressures
The saturated cases were performed at different pressures and 150 kg melt mass (3073K). In order to have an
idea of the influence of the subcooling the 150 kg saturated case at 5 bar (L29P150S) is compared with 150
kg subcooled (50°C at 5 bar, 100°C of subcooling) case: L29PR150N.
Cases are the following:
At 2 bar: L292BAR case; at 5 bar: L29P150S case and at 20 bar: L2920BAR case.
Pressure increases more as initial pressure is higher; it increases about 20 bar in the 20 bar initial pressure
case. It is shown that the pressure is practically constant in the subcooled case with 150 kg mass at 5 bar.
Also, quenching rate increases as initial pressure is higher and causes more fragmentation and less cake mass
accumulation.
The quenching rate in the subcooled case is lower than the saturated one and becomes unstable due to high
oscillations in the void fraction of the lowest volume in the FAT vessel, which is continuously voided and
refilled with water (CCFL phenomena).
The void fraction below the mixture level increases as initial pressure decreases and so it does the
accumulated cake mass. There is not a great influence of pressure on the global void fraction, which reaches
about 70%. Void fraction in the subcooled case is much lower: the heat is transferred to subcooled water,
which increases its temperature.
As conclusion it can be said that pressure increase in saturated test affects quenching rate and causes
accumulated cake mass reduction as void fraction decreases.
46
Assessment of Size Aspects in Modelling Molten Fuel Coolant Interaction
Chapter 7 - RESEARCH ACTIVITIES SUMMARY
7.1.4.
ADDITIONAL CALCULATIONS PERFORMED FOR THE SUBCOOLED CASE
Additional cases have been analysed to highlight the influence of enhanced fragmentation model, H2
generated and the nodalization type (1d/2d) in the subcooled case.
Effect of fragmentation model
The base case is compared with a case in which the enhanced fragmentation model has been used:
Descriptor
Melt
Mass (kg)
L29PR50N
L29PR50E
50
50
Melt
temperature
(K)
3073
3073
Initial
Pressure
(bar)
5
5
Initial water
level (m)
1.5
1.5
Initial water
temperature (K) and
subcooling
323 (100°C subc.)
323 (100°C subc.)
Fragmentation
Model
Stand.
Enhanced
Table 7.4 - Performed L-29 pre-test base case calculations with standard and fragmentation model
Case L29PR50N: standard model case of 50 kg melt mass at 3073K falling into 1.5 m of subcooled water at
50°C at 5 bar, and L29PR50E case (same characteristics but using enhanced model).
The pressure starts to increase as the melt reaches the water at 0.27 s. In both cases, standard and enhanced,
pressurisation are similar. In the enhanced fragmentation model case the pressurisation rate is initially
greater, between 1.8 and 3 seconds pressure in the standard case is higher, but after 3 seconds and in the
longer term the pressure in the enhanced case is maintained slightly greater.
It can be noted that in the standard model case the released energy is less than 65 MJ at 12 s. In the enhanced
fragmentation case the amount is a little greater, about 70 MJ at 12 s. This little difference results in little
differences in pressurisation rates.
In both cases, the average void fraction is between 12% and 18%, but is higher in standard fragmentation
case. The reason is that in the enhanced fragmentation case the void fraction of single volumes of FAT
increases much more in upper volumes, meanwhile is practically zero in the bottom volumes, in the standard
case, it increases also in bottom volumes, and this makes the average value to increase.
Void fraction is mainly due to generation of non-condensable gas (H2) and not to void fraction of the steam.
The melt components (jet, drops and cake) are presented in for the standard and enhanced fragmentation
models. In the standard model most of the melt is fragmented and a little part is collected on the bottom as
cake, with the enhanced model this trend is emphasized and the cake mass is practically zero. With the
enhanced model the heat transfer surface is almost double than the other case.
To summarise, it can be said that calculations have shown that little difference exists in the application of the
standard model or the enhanced fragmentation model when using initial subcooled water in the FAT vessel.
A great quantity of drops is produced in both cases and little pressurisation and lower void fraction.
47
Assessment of Size Aspects in Modelling Molten Fuel Coolant Interaction
Chapter 7 - RESEARCH ACTIVITIES SUMMARY
Effect of H2 model, H2 production
Also other case has been performed in order to analyse influence of H2 production.
The base standard case without hydrogen production (L29P50NH) is compared with base case (L29PR50N):
Descriptor
Melt
Mass (kg)
L29PR50N
L29P50NH
50
50
Melt
temperature
(K)
3073
3073
Initial
Pressure
(bar)
5
5
Initial water
level (m)
1.5
1.5
Initial water
temperature (K) and
subcooling
323 (100°C subc.)
323 (100°C subc.)
Fragmentation
Model
Stand.
Stand.No H2
Table 7.5 - Performed L-29 pre-test base calculation using model with and without H2 production
Two cases are compared to verify the influence of H2 production during a subcooled transient. The base
standard case L29PR50N (50 kg of melt mass, 1.5 m subcooled water at 50°C and 5 bar) and standard case
L29P50NH (50 kg of melt mass, 1.5 m subcooled water at 50°C and 5 bar) using a model without H2
production.
In both cases it is possible to observe similar quenching rate and fragmentation. Global void fraction (Figure
7.8) is much lower in the case without H2 production. Pressure (Figure 7.7) is also lower in this case, and
practically constant during the entire transient.
Summarising: it can be observed similar quenching rate and fragmentation with or without H2 production
model, but much lower void fraction (due to non-condensable gas and not to steam) and pressure in the case
without H2 production.
Effect of nodalization
In order to analyse the effect of the nodalization a case in 2d was run. This case was characterized by 5 radial
volumes.
The pressure and quenching rate are very similar in the two cases. The only remarkable difference is in the
void fraction distribution (Figure 7.9), which, in the 2d case, shows a central void fraction in the order of
30% and lateral void fraction about 8%, whereas in the 1d case is about 16%.
48
Assessment of Size Aspects in Modelling Molten Fuel Coolant Interaction
Chapter 7 - RESEARCH ACTIVITIES SUMMARY
WinGraf - 06-26-1998
6.00
5.80
XXX
YYY
Pressure (bar)
5.60
5.40
with H2 model
Y
Y
Y
Y
L29P50NH
L29PR50N
Y
Y
P020
P020
Y
Y
Y
5.20
Y
YY
YY
Y
XY X YXYXX XXXXX X
5.00
X
X
X
X
X
X
without H2 model
X
Y
X
4.80
4.60
4.40
0
.50
1.00
1.50
2.00
2.50
3.00
3.50
4.00
4.50
5.00
Time (s)
Figure 7.7 - Effect of the H2 production on pressure in Test the L-29
WinGraf - 06-29-1998
.40
XXX
YYY
ZZZ
VVV
JJJ
HHH
.35
.30
Void fraction (-)
V
V
.25
L29P50NH
L29P50NH
L29PR50N
L29PR50N
L29P50NH
L29PR50N
V
.20
VVV
VV
V
.05
0
Z
Z
Z
.15
.10
Z
V
Z
V Z X
XXXXXX
X
Z
Z
Z
Y
Y
Z YY YY
Y
Z Y
JHJHJHJ JHJHJH J HJ
Y
JHXY
XYZX
VY
JH
Z
XV
AVGVOIDML
AVGVOID10
AVGVOID10
AVGVOIDML
VOID10
VOID10
V
V
with H2 model
Y
H
J
X
X
X
X
without H2 model
Y
Y
Y
Y
H
J
V
V
V
Z
Z
X
V
X
Z
Z
Z
Z
X
Y
Y
H
HJ
H
J
JH
J H J
JH
avgvoid of steam w and w/o H2 model
H
-.05
0
.50
1.00
1.50
2.00
2.50
3.00
3.50
4.00
4.50
5.00
Time (s)
Figure 7.8 - Effect of the H2 production on the average void fraction in Test L-29
49
Assessment of Size Aspects in Modelling Molten Fuel Coolant Interaction
Chapter 7 - RESEARCH ACTIVITIES SUMMARY
WinGraf - 07-13-1998
.40
.35
.30
X
XXX
YYY
ZZZ
Central void fraction
L292D1
L292D1
L29PR50N
AV_VOID10(R1)
AV_VOID10(R5)
AVGVOID10
X
Void fraction (-)
X
.25
Lateral void fraction
X X
.20
Z
X
Z
X X X
Z
Z
Z Y Y
Y Y
Z
Y Y Y
Z
.15
Z
.10
X
Z
Y Y Y
Z
.05
Y
Z
X
Z
0
0
Y
X
X X
X
2.0
Z
Y
Y
Y
ZY Y Y Y
1.0
Global void fraction 1d
Z
X X
X
3.0
4.0
Z
Z
Z
Z
X
5.0
6.0
7.0
Time (s)
Figure 7.9 - Void fraction in the 2d calculation in Test L-29
7.1.5.
CONCLUSIONS
The COMETA code was successfully applied to all the configurations prospected for the test L-29, that were
basically:
Subcooled test with a reduced mass
Saturated test with the maximum available mass
In the subcooled test option it was shown the effect on the global void fraction of the selection of different
pressure and/or temperature. The results indicate that the greater influence on the void fraction is caused by
the pressure level (1 to 5 bar), which compressing the non-condensable void fraction, induces remarkable
change in the global void fraction. No major influence is present in the quenching rate. Also, in subcooled
conditions no significant difference resulted from the adoption of the standard or enhanced fragmentation
model.
In the saturated test it was shown that the initial pressure influences the quenching rate and the amount of
accumulated cake.
The code was able to calculate the subcooled conditions without innerving in major stability problems.
However a point that needs improvement in the code is the jet parcels description in order to properly
account for velocity differences among them.
Contrast between the 1-d and 2-d calculations showed no large overall influence of the nodalization scheme;
difference in void fraction distribution between the central and pheripherical zones was, however, noticed.
50
Assessment of Size Aspects in Modelling Molten Fuel Coolant Interaction
Chapter 7 - RESEARCH ACTIVITIES SUMMARY
7.2. COMETA PRE-TEST CALCULATION OF FARO TEST L-33
7.2.1.
INTRODUCTION
Test L-33 was scheduled to be performed in the FARO facility in July 1999. There was a great interest in
performing a subcooled test in the FARO test facility due to the potential for an energetic melt coolant
interaction. In this test an external trigger was applied to enhance the possibility of this occurrence. This was
supported by the results of the KROTOS Test 58, in which, for an UO2 / ZrO2 melt mixture, with conditions
of 5 bar, 130°C subcooling a weak steam explosion occurred. In the case of failed triggering, the test will
have been considered as another quenching test, the second one in subcooling conditions. Since there were
uncertainties in the real initial test conditions, determined by the peculiarity of the test performed, a number
of different hypotheses were explored as initial conditions of the test.
It is noted, however, that no spontaneous nor triggered steam explosions have been recorded so far with
UO2/ZrO2 mixtures contrary to Al2O3 mixtures in the KROTOS facility which have exhibited spontaneous
explosive interactions in subcooled conditions and triggered explosive interactions in nearly saturated
conditions.
The following hypotheses were considered:
Subcooled Test
Initial Pressure:
Mass:
Initial Temperature
H2 Production rate:
3, 4, 5 bar
176 kg
30°C
100% of what assumed for FARO tests (1.8·10-3 kg H2 / kg frag)
20% of what assumed for FARO tests (0.36·10-3 kg H2 / kg frag) for
the 4 bar case.
Trigger Conditions
First trigger on the FAT vessel bottom, in the central volume, at 0.50 s.
Second trigger on the FAT vessel, at height 0.663 m, in the lateral volume, at 3 s.
Initially it was not yet decided if to perform the test with or without the internal cylinder in the FAT test
section. Calculations were therefore performed in order to check the different behaviour in the two
conditions. The relatively low influence of the internal cylinder and, even more important, safety reasons
suggested to include the internal cylinder in the final test specifications. Only results with the internal
cylinder in the FAT test section are here presented.
Figure 5.4 shows the configuration of the facility for this test. An inner vessel with the same diameter as the
TERMOS will be installed inside the FAT vessel. The venting lines to the separator will be closed.
51
Assessment of Size Aspects in Modelling Molten Fuel Coolant Interaction
Chapter 7 - RESEARCH ACTIVITIES SUMMARY
7.2.2.
RESULTS OF TEST OPTION FOR THE FARO TEST L-33 BEFORE TRIGGERING
Initial and boundary conditions
The 2d nodalization adopted for the pre-test calculations for FARO Test L-33 is shown in Figure 7.10
FAT UPPER VOLUME (VOLUME 1)
STEAM DOME ABOVE NOZZLE (VOLUMES 91 TO 100)
STEAM DOME BELOW NOZZLE (VOLUMES 81 TO 90)
INTERNAL VESSEL VAPOR ZONE (VOLUMES 66 TO 80)
FAT EXTERNAL ZONE (VOLUMES 2 TO 15)
MELT CATCHER
INTERNAL VESSEL LIQUID ZONE (VOLUMES 16 TO 65)
Figure 7.10 - COMETA 2d nodalization for FARO Test L-33
52
Assessment of Size Aspects in Modelling Molten Fuel Coolant Interaction
Chapter 7 - RESEARCH ACTIVITIES SUMMARY
In order to cover the possible range in the experimental conditions, calculations have been performed with
different values of initial pressure, with 176 kg of melt injected mass (Tmelt = 3073K) and 5 cm of discharge
orifice. The case, L33INT3B should be seen as a lower bound and case L33INT5B should be seen as an
upper bound.
The so-called standard fragmentation model has been used for this test, i.e.: Cd=6, Cj=1,fH2=1.8·10-3kgH2/kg
fragmented. For 4 bar case another case was performed in which the H2 production rate was imposed to 20%
of what assumed for standard tests (fH2=0.36·10-3kgH2/kg fragmented). The reason for this change was due to
the lower H2 amount found in the first subcooled tests.
Calculations have been performed up to 3 s; in the following chapters several initial conditions for central
and lateral triggering at 0.5 s and 3 s are show.
The following table summarises cases with 176 kg, at 30°C and different initial pressure:
Descriptor
Melt
Mass (kg)
L33INT3B
L33INT4B
L33INT5B
L33I4BNH
176
176
176
176
Melt
temperature
(K)
3073
3073
3073
3073
Initial
Pressure
(bar)
3
4
5
4
Initial water
Level (m)
1.5
1.5
1.5
1.5
Initial water
temperature (K)
and subcooling
303 (103ºC subc.)
303 (113ºC subc.)
303 (121ºC subc.)
303 (113ºC subc.)
Fragmentation
Model
Stand.
Stand.
Stand.
Stand. 20% H2
production
Table 7.6 - Performed pre-test calculations for Test L-33
Results of the cases calculations up to 3 s
Cases L33INT3B, L33INT4B, L33INT5B and L33I4BNH: standard fragmentation model cases of 176 kg
melt mass at 3073K falling into 1.5 m of subcooled water at 30°C at 3, 4 and 5 bar. The 4 bar case was also
performed with the H2 production rate imposed to 20% of what assumed for standard tests.
The pressure increases about 1 bar in L33INT3B and L33INT4B cases and 2 bar in L33INT5B case at 3
seconds. In the case with 20% H2 production this increase is only about 0.5 bar. The energy released is
between 50 and 65 MJ at 3 s for all cases. Most of the melt is fragmented and a little part is collected on the
bottom as cake.
The central void fraction is calculated in the central volumes of FAT vessel, where jet descends and
fragments (vol.16 to 25). The global average void fraction (Figure 7.11) is calculated as the average void
fraction of the 50 lower volumes (initially full of water) in the FAT vessel.
Void fraction is an important quantity in the study of possible steam explosions, since as the global void
fraction is lower than a certain value, the steam explosion is likely to occur. In the three cases with higher H2
production, the global void fraction reaches about 0.3 (30%) at 3 seconds, increasing as the initial pressure
decreases. The void fraction in the central volumes is much less influenced by the initial pressure, being
about 0.4 in all cases. A great influence is due to the amount of H2 produced. Reducing it to 20% the void
fraction in the central volumes reduces from 0.4 to 0.2.
The temperature increase is strongly dependent on the amount of energy released and therefore follows the
same trend of this quantity. It can be noted that temperature increases faster in the bottom zone than in the
upper zone.
The jet leading edge position (Figure 7.12) shows that initially the melt falls down according to melt gravity
release as for Test L-31. What happens is that the jet parcels are released at progressive increasing velocity.
Some parcel can overcome the leading edge parcels, becoming leading edge. Therefore, the leading edge can
become faster than the theoretical gravity release with initial velocity zero (Figure 7.13).
53
Assessment of Size Aspects in Modelling Molten Fuel Coolant Interaction
Chapter 7 - RESEARCH ACTIVITIES SUMMARY
In order to remove this discrepancy a modification in the logic of the COMETA logic was introduced and if
one parcel reaches the one before it, the interference part between the two is broken up and the faster jet is
slowed down. Therefore the leading edge proceeds with the gravity discharge trend (Figure 7.14). With this
new logic the comparison with L-31 is much better. Figure 7.15 shows that the leading edge position with
the modification of the COMETA logic for 4 bar case (performed with the H2 production rate imposed to
20% of what assumed for standard tests) is similar to jet behaviour for Test L-31.
Pressure behaviour increases in the first seconds for case with this modification in the COMETA logic,
because the jet travels slowly and produces more fragmentation: in this case the heat transfer surface is
higher.
Conditions at the expected time of activation of the Central Trigger (0.50 s)
Figure 7.16 to Figure 7.18 show the curves of some quantities obtained in the four calculations at 0.50
seconds versus the height in the FAT vessel up to the point of injection of the melt. These would be the
initial conditions for the trigger at 0.5 seconds.
Figure 7.16 shows central void fraction; it is zero in the bottom volumes because the jet has not yet arrived at
that location. It would be better to trigger before the jet arrives to the bottom to avoid the vapour to absorb
the impulse of the trigger.
Figure 7.17 shows the lateral void fraction, only different from zero in the upper volumes. Figure 7.18
presents the mass of drops in the central volumes; drops are mainly present in the volumes at medium height.
The jet mass at 0.50 s has not arrived at the bottom.
Conditions at the expected time of activation of the Lateral Trigger (3 s)
Figure 7.19 to Figure 7.22 present the initial conditions for lateral trigger at 3 seconds and at height 0.663 m.
Central and lateral void fractions are higher in the upper volumes. In this case also the mass of drops in the
lateral volumes is represented (Figure 7.22). The jet has arrived at the bottom except for case at 5 bar. The
reason is due to a greater fragmentation with this higher pressure (more drops present, Figure 7.21).
54
Assessment of Size Aspects in Modelling Molten Fuel Coolant Interaction
Chapter 7 - RESEARCH ACTIVITIES SUMMARY
WinGraf - 07-15-1999
.40
GLOBAL VOID FRACTION
.35
.30
X
X
XXX
YYY
ZZZ
VVV
.25
.20
L33INT3B
L33INT4B
L33INT5B
L33I4BNH
.10
XY
Z
.05
XYZV
0
0
XY
Z
V
XYZ
V
Z
XYV
.50
V
X
3 bar
AVGVOID
AVGVOID
AVGVOID
AVGVOID
X
X
X
.15
X
YZ
X
YZ
V
X
Y
X
Z
Y
Z
V
V
X
Y
YZ Z
V
X X
YZ YZ
Y
Z
Y
Z
Y
Z
Y
Y
Y
Z
Z
5 bar
4 bar
V
V
V
V
V
V
V
V
V
V
20% H2
1.00
1.50
Time (s)
2.00
2.50
3.00
Figure 7.11 - Global void fraction for Test L-33
2,5
50
2,0
40
1,5
30
1,0
20
0,5
10
0,0
0
200
400
600
800
1000
Mass (kg)
Jet behavior 176 kg, 5 cm.
Height (m)
Void fraction (-)
X
3 bar
4bar
5bar
Test L-31
4bar 20%
Mass
0
1200
Time (ms)
Figure 7.12 - Jet leading edge for Test L-33
55
Assessment of Size Aspects in Modelling Molten Fuel Coolant Interaction
Chapter 7 - RESEARCH ACTIVITIES SUMMARY
CURRENT COMETA LOGIC
2
2
1
1
1
2
The faster jet piece (2) can
overpass the slower one (1)
Figure 7.13 - Current COMETA logic for leading edge behaviour
MODIFIED COMETA LOGIC
2
This piece is broken
2
1
2
1
Piece 2 takes the
velocity of 1
1
The faster jet piece (2)
cannot overpass the slower
one (1)
Figure 7.14 - Modified COMETA logic for leading edge behaviour
56
Assessment of Size Aspects in Modelling Molten Fuel Coolant Interaction
Chapter 7 - RESEARCH ACTIVITIES SUMMARY
2,5
50
2,0
40
1,5
30
1,0
20
0,5
10
0,0
0
200
400
600
800
1000
Mass (kg)
Height (m)
Jet behavior 176 kg, 5 cm.
Test L-31
4bar 20% modif
4 bar 20%
Mass
0
1200
Time (ms)
Figure 7.15 - Jet leading edge for Test L-33 with modified COMETA logic
Central Void Fraction 0.50 s
1.2
Void Fraction (-)
1
0.8
3bar
5bar
0.6
4bar
4 bar 20% H2
0.4
0.2
0
0
0.5
1
1.5
2
2.5
H (m)
Figure 7.16 - Central void fraction vs. height at 0.5 s for Test L-33
57
Assessment of Size Aspects in Modelling Molten Fuel Coolant Interaction
Chapter 7 - RESEARCH ACTIVITIES SUMMARY
Lateral Void Fraction 0.50 s
1.2
Void Fraction (-)
1
0.8
3bar
5bar
4bar
4 bar 20% H2
0.6
0.4
0.2
0
0
0.5
1
1.5
2
2.5
H (m)
Figure 7.17 - Lateral void fraction vs. height at 0.5 s for Test L-33
Drops in the Central Volumes 0.50 s
1.4
Mass of drops (kg)
1.2
1
3bar
5bar
4bar
4bar 20% H2
0.8
0.6
0.4
0.2
0
0
0.5
1
1.5
2
2.5
H (m)
Figure 7.18 - Drops in the central volumes vs. height at 0.5 s for Test L-33
58
Assessment of Size Aspects in Modelling Molten Fuel Coolant Interaction
Chapter 7 - RESEARCH ACTIVITIES SUMMARY
Central Void Fraction 3 s
1.2
Void Fraction (-)
1
0.8
3bar
5bar
4bar
4bar 20% H2
0.6
0.4
0.2
0
0
0.5
1
1.5
2
2.5
H (m)
Figure 7.19 - Central void fraction vs. height at 3 s for Test L-33
Lateral Void Fraction 3 s
1.2
Void Fraction (-)
1
0.8
3bar
5bar
4bar
4bar 20% H2
0.6
0.4
0.2
0
0
0.5
1
1.5
2
2.5
H (m)
Figure 7.20 - Lateral void fraction vs. height at 3 s for Test L-33
59
Assessment of Size Aspects in Modelling Molten Fuel Coolant Interaction
Chapter 7 - RESEARCH ACTIVITIES SUMMARY
Drops in the Central Volumes 3 s
Mass of drops (kg)
24
20
16
3bar
5bar
4bar
4bar 20% H2
12
8
4
0
0
0.5
1
1.5
2
2.5
H (m)
Figure 7.21 - Drops in the central volumes vs. height at 3 s for Test L-33
Drops in the Lateral Volumes 3 s
Mass of drops (kg)
6
4
3bar
5bar
4bar
4bar 20% H2
2
0
0
0.5
1
1.5
2
2.5
H (m)
Figure 7.22 - Drops in the lateral volumes vs. height at 3 s for Test L-33
60
Assessment of Size Aspects in Modelling Molten Fuel Coolant Interaction
Chapter 7 - RESEARCH ACTIVITIES SUMMARY
7.2.3.
TRIGGER CALCULATION TESTS WITH WATER
Several shapes of trigger were tested in the FAT vessel with the internal cylinder and without releasing melt,
to choose the appropriated one. In the experiment it was planned to use a plastic charge whose form in terms
of pressurization was quite different from previously used gas triggers like in KROTOS. Figure 7.23 shows
the triangular mass flow shape of the vapour trigger initially tested, injecting vapour at 1 bar and 293 K, with
a maximum peak of 167 kg/s. This trigger causes pressure behaviour in the central volumes shown in Figure
7.24. This impulse was strong enough but the shape of the pressure peak was not as short as in the case of the
planned plastic trigger.
WinGraf - 09-08-1999
180
VAPOUR TRIGGER SHAPE
160
140
Mass flow (kg/s)
120
XXX
L331SNM fillfl01
X
100
80
60
X
X
40
X
X
20
X
0
.999
1.000
1.001
1.002
1.003
Time (s)
1.004
1.005
1.006
Figure 7.23 - Vapour trigger mass flow shape for Test L-33
WinGraf 3.2 - 09-08-1999
70.0
PRESSURE IN THE CENTRAL VOLUMES
X
XXX
YYY
ZZZ
VVV
JJJ
HHH
###
OOO
AAA
BBB
60.0
Y
Z
50.0
Pressure (bar)
X
V Y
40.0
X
Z YX
Y
J
30.0
10.0
0
1.000
O #O
#A
OAB
HB A
H#O
AB
#
H
O J J BJH
X
Z
A#
YX
AB
# HJ
V V VJ
BOH J V V
YX
ZV Z
JV Z
ZBOH
#
J
V
Z
A
X
Y
#
Z
Z
J
X ZX Z Z Z
O
JYHV
A# V
V
Z Z B
V
Y
#
Z Z YXYXY Y Y Y Y Y
HJ V
H
VJA
HX V
X X X B
O
OA
O
V VJHB
X
# HBJ J BJ B A#
B
J
OA# HB A
O
A#
H#OH#
OA#OA
B HB
OA
#
1.002
1.003
1.004
1.005
V Z
H
20.0
1.001
X
Y
L331snm
L331snm
L331snm
L331snm
L331snm
L331snm
L331snm
L331snm
L331snm
L331snm
p016
p017
p018
p019
p020
p021
p022
p023
p024
p025
X
Y
1.006
Time (s)
Figure 7.24 - Pressure in the central volumes triggering with vapour mass flow
61
Assessment of Size Aspects in Modelling Molten Fuel Coolant Interaction
Chapter 7 - RESEARCH ACTIVITIES SUMMARY
Another trigger shape (Figure 7.25) was tested in the same conditions but injecting water instead of vapour at
200 bar and 293 K, during 0.251 ms in a rectangular shape, and a mass flow maximum of 1350 kg/s. This
impulse causes the pressure behaviour shown in Figure 7.26. A high pressure of 250 bar was reach in the
bottom central volume. This was the selected trigger shape for the calculations.
WinGraf - 09-09-1999
1600
WATER TRIGGER SHAPE
1400
XXX X
X
XXX
1SNMLIQ fillfl01
Mass flow (kg/s)
1200
1000
800
600
400
200
0
.999
1.000
1.000
1.001
1.001
1.002
1.002
Time (s)
Figure 7.25 - Water trigger mass flow shape for Test L-33
WinGraf 3.2 - 09-08-1999
300
PRESSURE IN THE CENTRAL VOLUMES
250
X
XXX
YYY
ZZZ
VVV
JJJ
HHH
###
OOO
AAA
BBB
Pressure (bar)
200
150
1snmliq
1snmliq
1snmliq
1snmliq
1snmliq
1snmliq
1snmliq
1snmliq
1snmliq
1snmliq
p016
p017
p018
p019
p020
p021
p022
p023
p024
p025
X
100
Y
50
Y
Y
Z
X
0
1.000
X
ZV
YZ VJ
X
X
1.001
Y Z V JH
B OA B O A B OAB# O A B # O
Y
ZZ
VV
JJ
YY
HH
ZO
V JX H
Y#
ZO
VA
JX H
Y#
ZO
VA
JX B
Y#
H
ZO
VA
JX H
YZ
Y#
Z V JX H
Y#
Z VJXH
YZ V JX H
Y ZV
XXX
##
# V JX H
1.002
1.003
1.004
1.005
Time (s)
Figure 7.26 - Pressure in the central volumes triggering with water mass flow
62
Assessment of Size Aspects in Modelling Molten Fuel Coolant Interaction
Chapter 7 - RESEARCH ACTIVITIES SUMMARY
Figure 7.27 shows pressure behaviour in the central bottom volume 17 triggering with the triangular vapour
mass flow and the rectangular water mass flow. A thinner shape is obtained triggering with water.
Figure 7.28 shows the final trigger shape used for all calculations, injecting water at 200 bar and 293 K
during 0.251 ms in the bottom central volume of the FAT vessel at 0.5 s and in the lateral volume at 3 s.
WinGraf - 09-09-1999
60.0
Y
X
50.0
XXX
YYY
X
Pressure (bar)
40.0
L331SNM p017
1SNMLIQ p017
Y
X
Triangular vapour trigger
X
30.0
X
Y
X
X
Rectangular water trigger
20.0
X
Y
X
X
10.0
X X
X X X X X X
X
X
Y
0
1.000
1.001
YYY
Y
1.002
Y
Y
Y
1.003
Y
Y
1.004
Y
Y
1.005
Time (s)
Figure 7.27 - Pressure in two central volumes triggering with vapour and water mass flows
WinGraf - 09-09-1999
1600
1400
FINAL WATER TRIGGER SHAPE
XXXXX XX
XXX
3B_05C fillfl01
Mass flow (kg/s)
1200
1000
800
600
400
200
0
.499
.499
.500
.500
.500
Time (s)
.500
.501
.501
.501
Figure 7.28 - Selected water trigger mass flow shape for Test L-33
63
Assessment of Size Aspects in Modelling Molten Fuel Coolant Interaction
Chapter 7 - RESEARCH ACTIVITIES SUMMARY
7.2.4.
CALCULATION OF THE TRIGGERING EVENT FOR THE FARO TEST L-33
Results of 176 kg mass cases at different pressures
Two different triggers have been simulated with COMETA code. The first one at 0.5 seconds of calculation,
simulated injecting water at 200 bar and 293 K during 0.251 ms in the bottom central volume of the FAT
vessel (volume 16) and the second one at 3 seconds injecting water with the same characteristics in the
lateral volume at height 0.663 m (vol. 60).
1
Melt catcher
92
94
96
98 100
91
93
95
97
99
Insulation
82
84
86
88
90
Vessel steel
81
83
85
87
89
68
71
74
77
80
15
67
70
73
76
79
14
66
69
72
75
78
13
25
35
45
55
65
12
24
34
44
54
64
11
23
33
43
53
63
10
22
32
42
52
62
9
21
31
41
51
61
8
20
30
40
50
60
7
19
29
39
49
59
6
18
28
38
48
58
5
17
27
37
47
57
4
16
26
36
46
56
3
2
Central trigger
(vol. 16)
Lateral trigger
(vol. 60)
Figure 7.29 - COMETA 2d nodalization and triggering position
The trigger causes pressure pick that should be transmitted to the upper and lateral volumes.
Pressure behaviour is studied after central triggering in the central (vol. 16 to 25) and lateral volumes (vol.
56 to 65) initially full of water. For the case at 3 bar of initial pressure (L33INT3B), after the central trigger,
the pressure reaches 250 bar in the central zone and about 60 bar in the lateral zone. But this impulse is not
propagated, is absorbed almost completely by the vapour present in the central volumes.
In case at 4 bar initial pressure (L33INT4B), and case at 5 bar initial pressure (L33INT5B) occur the same
problem, pressure reaches a high value (about 250 and 60 bar in central and lateral bottom volumes,
respectively) but this impulse is not propagated.
64
Assessment of Size Aspects in Modelling Molten Fuel Coolant Interaction
Chapter 7 - RESEARCH ACTIVITIES SUMMARY
Case at 4 bar initial pressure with the H2 production rate reduced to 20% of what assumed for FARO
standard tests (L33I4BNH) is presented in Figure 7.30 and Figure 7.31 (central and lateral pressure
evolution). It is observed that a propagation is present and it is possible to see the progression of the pressure
along the mixture.
Central and lateral pressures in the four cases after lateral trigger at medium height (0.663 m) are studied. All
cases show low increase of pressure and little transmission of the impulse (Figure 7.32 and Figure 7.33).
In the standard cases the pressure increase only in the bottom volume and the void fraction is sufficiently
high to suppress any propagation. In the case with reduced H2 production, instead, a propagation is present
and it is possible to see the progression of the pressure along the mixture. The calculation could not be
continued after 20 ms because the pressure of 250 bar has been reached and the code stops at that value.
WinGraf - 06-30-1999
300
PRESSURE IN THE CENTRAL VOLUMES
250
X
XXX
YYY
ZZZ
VVV
JJJ
HHH
###
OOO
AAA
BBB
Pressure (bar)
200
150
100
50
0
.495
V
Y
H
4B_05CNH
4B_05CNH
4B_05CNH
4B_05CNH
4B_05CNH
4B_05CNH
4B_05CNH
4B_05CNH
4B_05CNH
4B_05CNH
p016
p017
p018
p019
p020
p021
p022
p023
p024
p025
B
Z
Y
Y
A
Z
Z J
VX
O
XY
J
Z OABVV
J JVV
J VY#
JVJ Z
ZO
H#
#VJXH
J
#O
H
A
H
A
HHHHH
J
B
B
#
O
V
#
##
#B#
ZAA
A
B
OOOOO
Y
Z
Z
AA
JX
XB
VV
#
BBB
H
AB
YY
#V
ZOJA
VJA
H
OJX
H#
H
Y
Y#
ZO
VJA
YZ
VJAB
ZVJAH
H
H# O
B#O
AB
XB
XB
ZO
ZO
VJA
VJA
Y#
H
H
XYZX X
YYYY
XXXXX
XB
Y#
XB
Y#
ZO
ZZ
.500
.505
.510
.515
.520
Time (s)
.525
Figure 7.30 - Pressure behaviour in the central volumes after triggering at 0.5 s in the 4 bar case calculation with
20% of H2 production
65
Assessment of Size Aspects in Modelling Molten Fuel Coolant Interaction
Chapter 7 - RESEARCH ACTIVITIES SUMMARY
WinGraf - 06-30-1999
90.0
PRESSURE IN THE LATERAL VOLUMES
80.0
70.0
Pressure (bar)
60.0
50.0
X
40.0
Y
30.0
20.0
10.0
0
.495
V
Z
XXX
YYY
ZZZ
VVV
JJJ
HHH
###
OOO
AAA
BBB
4B_05CNH
4B_05CNH
4B_05CNH
4B_05CNH
4B_05CNH
4B_05CNH
4B_05CNH
4B_05CNH
4B_05CNH
4B_05CNH
p056
p057
p058
p059
p060
p061
p062
p063
p064
p065
YZ
V
H
#
V
X
A
O
J
VA
Y
JO
H#
VB
J
Z
V
BZJH#O
Y H
H#
#OAYZ X
JH
A
#
V
V
J
X
H
XBO
JHJHH
X X B
A
#VV
JB
Z
O
B
#
B
H
#
O
Y
A
#
B
A
H
J
#
B
B
B
J
X
H
X
B
A
V
A
B#O
Z OJ
JB
H#OO
A #O H
ZVJH
ABB
ZVA
#
VZ ZO
OOO
Y#
AZVAB
Y#
AB
J
VA
YZVJ
VA
ZO
VX
JH
ZO
XY #
JH
YA
Y#
YZX
YYYYY
ZXXXXX
XYZ
XH
.500
.505
.510
.515
.520
Z J
.525
Time (s)
Figure 7.31 - Pressure behaviour in the lateral volumes after triggering at 0.5 s in the 4 bar case calculation with
20% of H2 production
WinGraf - 06-30-1999
8.00
PRESSURE IN THE CENTRAL VOLUMES
XXX 4B3S20%L
YYY 4B3S20%L
ZZZ 4B3S20%L
VVV 4B3S20%L
JJJ 4B3S20%L
Y
HHH 4B3S20%L
### 4B3S20%L
OOO 4B3S20%L
Z
X
AAA 4B3S20%L
BBB 4B3S20%L
X
X
Y V
Y Z
ZV
XY
J YJ V
Y YZ
V
X
J
Z
JV
HZ
Y
H#OA
#
H
#
Z
Y
Z
Y
H
H
X
YAB
V
J
#
Y
Z
J
H
Z
Z
V
#
#
O
Z
#
X
H
Z
V
J
Y
J
J
J
X
OAB
O
O
JB
#
HA
JV
VAB
#
#
#
#
HA
OAB
O
HA
HA
VA
V
Z
O
O
O
HA
JV
O
HA
H
XO
O
V
XB
JB
O
V
YA
O
OAB
B
YAB
#B
#
JO
#B
HA
#
Z
VJ J HJXB
XB
XB
YZB
H#
X
H
Y
X
X Y
X V
Z V V
Z
Z
Y
Y
X
X
X
7.00
Pressure (bar)
6.00
5.00
4.00
3.00
2.00
2.990
2.995
3.000
3.005
3.010
Time (s)
3.015
3.020
p016
p017
p018
p019
p020
p021
p022
p023
p024
p025
3.025
3.030
Figure 7.32 - Pressure behaviour in the central volumes after triggering at 3 s in the 4 bar case calculation with
20% of H2 production
66
Assessment of Size Aspects in Modelling Molten Fuel Coolant Interaction
Chapter 7 - RESEARCH ACTIVITIES SUMMARY
WinGraf - 06-30-1999
18.0
PRESSURE IN THE LATERAL VOLUMES
16.0
XXX
YYY
ZZZ
VVV
JJJ
HHH
###
OOO
AAA
BBB
14.0
Pressure (bar)
12.0
10.0
8.0
4B3S20%L
4B3S20%L
4B3S20%L
4B3S20%L
4B3S20%L
4B3S20%L
4B3S20%L
4B3S20%L
4B3S20%L
4B3S20%L
p056
p057
p058
p059
p060
p061
p062
p063
p064
p065
6.0
4.0
Y
X
XYX
XYZ
XYZ
Z
XV
V
Y
Z
Y
YX
V
X
V
XY H# O
Y
V
Y
V
V
Y
V
V
V
Y
Z
Z
X
Y
X
Y
V
V
V
V
Y
Y
Z
YJZ
BJZ
A
H
BJZ
A
B
O
O
O
A
O
A
H
H
H
H
Z
#O
#
#
#
#
#
XH
XH
BJA
O
A
O
BJ#
A
O
BJ#
AH
BJ#
A
BJ#
AH
BJ#
B
BJA
B
BJA
H
BJZ
A
BJA
H
B
#O
#H
#H
V
H
Z
Z
#O
#O
VJZ
XO
VJA
ZO
XH
YJA
XO
H
XO
XO
Y
XH
YJZ
ZVJZ
XV
2.0
0
2.990
2.995
3.000
3.005
3.010
Time (s)
3.015
3.020
3.025
Figure 7.33 - Pressure behaviour in the lateral volumes after triggering at 3 s in the 4 bar case calculation with
20% of H2 production
7.2.5.
CONCLUSIONS
The COMETA code has been used to calculate the conditions foreseen for the FARO Test L-33. According
to the initial specifications this facility configurations were prospected: one with the internal vessel and one
without it. The calculations indicated no major differences as far as regards the central part of the test
section. Therefore the solution with the internal cylinder was chosen.
The code has been also applied to simulate the triggering and the response due to a possible explosion. No
propagation was calculated if the H2 generation model was the same as applied for previous saturated tests. If
the amount is reduced to values typical of the last subcooled Test L-31 or even less a propagation of the
explosion seems possible.
67
Assessment of Size Aspects in Modelling Molten Fuel Coolant Interaction
Chapter 7 - RESEARCH ACTIVITIES SUMMARY
7.3. ANALYSIS OF FUEL-COOLANT QUENCHING PHENOMENA BY COMETA
CODE IN REACTOR GEOMETRY
7.3.1.
INTRODUCTION
During the progression of a severe accident in water cooled reactors, molten fuel resulting from overheating
the core, in absence of adequate cooling, can relocate and interact with coolant in the reactor pressure vessel
lower plenum or in the reactor cavity.
The document reports some simulation results of an extended melt coolant interaction accident in a reactor
geometry typical of the Spanish reactor ASCO-1, a 3-loops 966 MWe (2686 MWth) Westinghouse PWR
(Asociación Nuclear ASCÓ, A.I.E - “Private communication. Data trasmission” - 1998) [33] [49]. 1-d and 2d calculations have been performed with the last version of JRC-Ispra COMETA code, simulating 30.000 kg
of melt that fall into the central part of the core inlet and lower plenum. The melt mixture consisted of 80%
weight of UO2 and 20% weight of ZrO2 at 3073K released by a 10 cm diameter discharge orifice into
saturated water at a pressure of 50 bar and at different water levels. Boundary cases in which completely full
or completely empty conditions were also considered.
Emphasis is placed on overall thermal-hydraulic system response and melt fragmentation behaviour in order
to correlate pertinent phenomena to a range of accident initial and boundary conditions such as residual
water level depth, melt discharge rate and exchanged power.
7.3.2.
INITIAL AND BOUNDARY CONDITIONS. ANALYSED CASES
The following case was analysed: a molten pool is formed in the central core region and starts to be released
through a hole of diameter 10 cm into the core inlet and the lower plenum that are considered full of water or
partially full, depending on the case. The melt is assumed to be discharged in the central part of the vessel
and eventual fragmented pieces are kept in the central zone for the 2-d cases, which is 0.37 m wide. In 1-d
cases fragmented pieces can be found in all core inlet and lower plenum volumes. No heat exchange is
assumed above the injection point. The melt does not occupy volume, i.e. no reduction of the volume due to
the presence of the melt is assumed in the lower plenum.
The following conditions were considered:
Pressure:
Temperature:
Melt mass:
Melt temperature:
Orifice size:
1-D Descriptor
ACQPL2
ACQ2M37
ACQ2M
ACQ129M
ACQ1M
ACQ05M
VAPOR
BOUND
68
50 bar.
Tsaturation.
30.000 kg.
3073K.
10 cm.
Initial
Water level (m)
2.969 (vol.: 19 to 33, 34 to 45 and 46 to 57)
2.37 (vol.: 19 to 33, 34 to 38 and 46 to 50)
2 (vol.: 19 to 33, 34 and 46)
1.29 (vol.: 19 to 28)
1 (vol.: 19 to 26)
0.5 (vol.: 19 to 22)
0.26 (vol. 19 and 20)
All lower plenum (vol.: 19 to 33), core inlet (vol.:
46 to 57), downcomer
(vol.: 1 to 8 and 34 to 45), riser and upper plenum
(vol.: 11 to 18) full of water
Nodalization
1D (57 nodes)
1D (57 nodes)
1D (57 nodes)
1D (57 nodes)
1D (57 nodes)
1D (57 nodes)
1D (57 nodes)
1D (57 nodes)
Assessment of Size Aspects in Modelling Molten Fuel Coolant Interaction
Chapter 7 - RESEARCH ACTIVITIES SUMMARY
2-D Descriptor
2DASCOFULL
2DASCOHALF
2DASCOBASE
2DASCO1M
Initial
Water level (m)
8.1 (vol.: 1 to 8, 11 to 18, 19 to 52, 53 to 57 and 58
to 77)
5.51 (vol.: 1 to 4, 11 to 14, 19 to 52, 53 to 57 and
58 to 77)
2.969 (vol.: 19 to 52, 53 to 57, and 58 to 77)
1 (vol.: 19 to 48)
Nodalization
2D (77 nodes)
2D (77 nodes)
2D (77 nodes)
2D (77 nodes)
The initial conditions assumed in the calculations are depicted in Figure 7.34 for 1-d calculations: only
extreme cases VAPOR (0.26 m water level) and BOUND (Vessel full, 8.1 m water level). The rest of the
cases are between those ones (0.5 m, 1 m, 1.29 m, 2 m, 2.37 m and 2.969 m). Initial conditions for 2-d
calculations are shown in Figure 7.35: Vessel full (2DASCOFULL, 8.1 m), half vessel full (2DASCOHALF,
5.51 m), base case (2DASCOBASE, 2.969 m) and 1 m case (2DASCO1M).
In specific sequences like TMI, however, after core melt down, relocation into the lower plenum occurred
with a vessel full condition since the primary system water inventory was being replenished during the
evolution of the accident. So it was decided to perform cases with all vessel full of water up to upper head
level: BOUND and 2DASCOFULL cases in 1-d and 2-d nodalization, respectively. One case with half vessel
full of water in 2-d nodalization (2DASCOHALF) was also performed.
The nodalization chosen for all calculations is presented in Figure 7.36 and Figure 7.37 for the 1-d and for
the 2-d cases; great detail has been devoted to the simulation of the reactor vessel lower downcomer, core
inlet and lower plenum.
69
Assessment of Size Aspects in Modelling Molten Fuel Coolant Interaction
Chapter 7 - RESEARCH ACTIVITIES SUMMARY
a) 0.26 m water (VAPOR)
b) Vessel Full (BOUND)
Figure 7.34 - Initial conditions for extreme 1-d calculations in COMETA reactor calculation
a) 1 m water
(2DASCO1M)
b) Lower plenum
and core inlet full
(2DASCOBASE)
c) Half vessel full
(2DASCOHALF)
d) Vessel Full
(2DASCOFULL)
Figure 7.35 - Initial conditions for 2-d calculations in COMETA reactor calculation
70
Assessment of Size Aspects in Modelling Molten Fuel Coolant Interaction
Chapter 7 - RESEARCH ACTIVITIES SUMMARY
REST PRIMARY
(VOLUME 10)
V=143.85 m 3 ; H=5.68 m
UPPER HEAD
(VOLUME 9)
V=30.8 m 3 ; H=3.08 m
UPPER DOWNCOMER
(AXIAL VOLUMES 1 TO 8)
V=10.16 m 3 ; H=5.105 m
RISER + UPPER PLENUM
(AXIAL VOLUMES 11 TO 18)
V=36.572 m 3 ; H=5.105 m
CORE INLET
(AXIAL VOLUMES 46 TO 57)
V=8.215 m 3 ; H=1.024 m
LOWER DOWNCOMER
(AXIAL VOLUMES 34 TO 45)
V=2.037 m 3 ; H=1.024 m
LOWER PLENUM
(AXIAL VOLUMES 19 TO 33)
V=13.98 m 3 ; H=1.945 m
Figure 7.36 - COMETA 1-d nodalization in reactor calculation
71
Assessment of Size Aspects in Modelling Molten Fuel Coolant Interaction
Chapter 7 - RESEARCH ACTIVITIES SUMMARY
REST PRIMARY
(VOLUME 10)
V=143.85 m3 ; H=5.68 m
UPPER HEAD
(VOLUME 9)
V=30.8 m3 ; H=3.08 m
UPPER DOWNCOMER
(AXIAL VOLUMES 1 TO 8)
V=10.16 m3 ; H=5.105 m
RISER + UPPER PLENUM
(AXIAL VOLUMES 11 TO 18)
V=36.572 m3 ; H=5.105 m
CORE INLET
(VOLUMES 58 TO 77)
V=8.215 m3 ; H=1.024 m
LOWER DOWNCOMER
(AXIAL VOLUMES 53 TO 57)
V=2.037 m3 ; H=1.024 m
LOWER PLENUM
(VOLUMES 19 TO 52)
V=13.98 m3 ; H=1.945 m
Figure 7.37 - COMETA 2-d nodalization in reactor calculation
72
Assessment of Size Aspects in Modelling Molten Fuel Coolant Interaction
Chapter 7 - RESEARCH ACTIVITIES SUMMARY
7.3.3.
CALCULATION RESULTS
All calculations, 1-d and 2-d, were performed until about 130 seconds. The melt mass released (30000 kg),
increases up quickly to a constant injection period of about 50 seconds with a mass flow rate of 250 kg/s and
after, it reduces to 0 at about 200 seconds.
The case with water filling lower and upper plenum, riser and downcomer (BOUND) was prolonged up to
300 seconds to observe the behaviour in the longer term.
Results of 1-d cases at different water levels (from 2.969 m down to 0.26 m)
The next table indicates distribution mass at 40 seconds of transient:
Case
Initial Water
level (m)
Fragmented mass
(kg)
Molten Cake mass
(kg)
ACQPL2
2.969
6203
3627
Molten Cake
mass/Total mass
(%)
37
ACQ2M37
2.37
6161
3712
38
ACQ2M
2
6008
3945
40
ACQ129M
1.29
5729
4182
42
ACQ1M
1.0
5245
4712
47
ACQ05M
0.5
67
9805
99.3
VAPOR
0.26
50
10034
99.5
The pressure (Figure 7.38) increases initially in a similar way for 2.969, 2.37, 2, 1.29 and 1 m level cases. In
the long term in the first three cases it continues to increase while in the 1.29 and 1 m level cases trends to a
stabilisation. The stabilisation is due to the completely voiding of the lower plenum, for case with 1 m level
and for case with 1.29 m level, from a certain time no fragmentation and heat exchange occurs.
The pressurisation in the 0.5 m level case (ACQ05M) and in the case with 0.26 m of initial level (VAPOR) is
much lower than in the other cases, due to the lower quenching rate, as explained later. In these two cases
about 70 and 60 bar are reached respectively at 130 seconds.
The mass of fragmented melt (drops) grows initially in a similar way for cases with 2.969, 2.37, 2, 1.29 and
1 m initial water level. In the 0.5 m and 0.26 m water level cases it is very low, less than 100 kg. In 1.29 and
1 m level the fragmentation mass is stabilised after about 100 and 80 seconds, respectively.
The production rate of drops (kg/s) can be observed in Figure 7.39 for the lower plenum. In the core inlet, in
the first seconds, it becomes null in all cases. In the 1.29 m and, 1 m level cases, the production rate in the
lower plenum increases rapidly due to the earlier arrive of the jet to the water. In those cases the production
rate becomes null and consequently fragmentation mass is stabilized.
In cases with initial water level above the lower plenum height (2.969 m, 2.37 m and 2m) manometric flow
oscillations occur between core inlet and lower downcomer.
From Figure 7.40 can be verified these phenomena observing velocity oscillations (gas and liquid) in
junctions 19 (from lower plenum towards core inlet) and 20 (from lower plenum towards lower downcomer)
for 2 m initial water level cases.
73
Assessment of Size Aspects in Modelling Molten Fuel Coolant Interaction
Chapter 7 - RESEARCH ACTIVITIES SUMMARY
The heat exchange (Figure 7.41), which is related to hot drops production, is almost equal for all cases with
higher initial water level (2.969 m, 2.37 m and 2 m). As the initial water level is reduced (0.5 m and 0.26 m)
the quenching rate decreases due to the lower drops production. After 40 seconds it decreases, in all cases, as
melt mass injection flow is reduced.
The mixture level, in the lower plenum and core inlet, is calculated up to a maximum of 2.969 m, the
injection point. It reaches the core bottom level or injection point (2.969 m) for cases with 2.969, 2.37 and 2
m initial water level in the first seconds of transient and remains at that value or above for a longer period as
the initial level is higher. After about 20 seconds (2 m level case), 60 seconds (2.37 m level case), and 150
seconds (2.696 m level case), the mixture level decreases below the injection point. The mixture level in the
rest cases reaches levels below the injection point and also below the lower plenum height (1.945 m).
To resume, Figure 7.42 shows the mixture level at 10 seconds transient versus initial water level. If the initial
water level is above 1.9 m, the heat exchange is able to increase the mixture level above the injection point.
The global mean void fraction in lower downcomer, core inlet and lower plenum (Figure 7.43) increases
quickly up to an equilibrium value near 0.4 for the cases with higher initial water level (2.969 m, 2.37 m and
2 m). As the mixture level decreases in all these cases, the void fraction starts to increase (at about 150
seconds, 60 seconds and 20 seconds, depending on the case).
The 1.29 and 1 m level cases start from a higher mean value and reach stability at the maximum value of
void fraction, 1.0. For the 0.5 m and 0.26 m water level cases, initial value is almost 1.0 and after small
increase, they reach equilibrium.
The un-fragmented mass production ratio (cake), calculated as un-fragmented mass/injection melt mass flow,
is high for the cases with lower initial water level (0.5 m and 0.26 m), where lower drops are produced, it fast
increases, reaching almost 100% of the discharged melt mass in these cases. The cases with higher initial
level increase up to about 40% (Figure 7.44).
The void fraction of the upper volumes indicate that when the mixture level exceeds the lower downcomer,
the core inlet or the lower plenum volumes the water is delivered to the upper downcomer or riser or upper
volumes of lower downcomer, core inlet or lower plenum.
As conclusion: All these cases show decreasing fragmentation and energy production as the initial water
level in the core inlet and lower plenum is reduced. With an initial water level below a value between 1 m
and 0.5 m, the fragmentation and the quenching rate progressively decrease and the cake mass accumulation
increases (Figure 7.45).
74
Assessment of Size Aspects in Modelling Molten Fuel Coolant Interaction
Chapter 7 - RESEARCH ACTIVITIES SUMMARY
WinGraf - 02-25-1999
160
PRESSURE UPPER HEAD
Z
XXX
YYY
ZZZ
VVV
JJJ
HHH
###
120
Pressure (bar)
2.37 m
1.29 m
140
ACQPL2
ACQ2M37
ACQ2M
ACQ129M
ACQ1M
ACQ05M
VAPOR
p009
p009
p009
p009
p009
p009
p009
100
V
Z
YXJ
V
Z
ZYX J
VYV
V
YZ
Z
XJ XJ
V
Y
X
Z
J
V
Z YXJ
V
XJYXJ
V
Z
JZ
Y
XY
H
V
Z
J
X
Y
V
J
Z
HH H
Z
XJYX
YV
HHHH# # #
#
X
Y
X
###
80
60
Y
Z
2m
Z
Z
Z
V
V
Z YX
J
YX
J
Z
YV
X
Z
V
Y
X
Y
X
V
V
X
V
X
J
J
J
J
Y
YV
X
2.969 m
1m
0.5 m
H
#
Y
#
#
#
#
H
H
H
H
H
H
H
#
#
#
All void
40
0
20.0
40.0
60.0
80.0
100.0
120.0
140.0
Time (s)
Figure 7.38 - Pressure in the upper head at different initial water levels
WinGraf - 02-26-1999
300
Drops production (kg/s)
250
XXX
YYY
ZZZ
VVV
JJJ
HHH
###
2m
ACQPL2
ACQ2M37
ACQ2M
ACQ129M
ACQ1M
ACQ05M
VAPOR
V
V
V
J VJ
J
JV
JV
Z
Z
X
Z
Z
Z
ZX
XYX
XY YXYXYXYXYXYXJYX
X
Y
Z
V
Y
Z
Y
X Y
ZZ
Y
Z
Z
X Y
Z Z ZJ V
V YV
X
X
J
V V V VJV
J
J
J
J J
200
150
100
PRODROPLWPL
PRODROPLWPL
PRODROPLWPL
PRODROPLWPL
PRODROPLWPL
PRODROPLWPL
PRODROPLWPL
2.37 m
Z
Z
Z
Y
V
X
Z
Y
X
V
Y
X
J
All void
0.5 m
0
20.0
Y Z
Y
X
X
V
50
0
Y
1m
#
40.0
60.0
2.969 m
J 1.29 m
80.0
J
100.0
V
120.0
140.0
Time (s)
Figure 7.39 - Drops production rate in lower plenum at different initial water levels
75
Assessment of Size Aspects in Modelling Molten Fuel Coolant Interaction
Chapter 7 - RESEARCH ACTIVITIES SUMMARY
WinGraf - 02-01-1999
2.00
Gas velocity
1.50
Z
Z
Junction velocity (m/s)
1.00
Z
V
Z
Z
XX
XXXXXXZ
XVVX
Z
X
Z Z XX
Z
V Y
Y
Y YYYY
Z Y
V
YV
YYY ZV
VV
.50
0
-.50
X
Z
Y
V
V
Y
Y
X
V
Z
X
Z
X
Z
Y
V
X
Z
Y
V
V
V
V
Z
X
Y
V
Y
Y
V
-1.00
ACQ2M velgj019
ACQ2M velfj019
ACQ2M velgj020
ACQ2M velfj020
Z
X
X
Z
Y
X
XXX
YYY
ZZZ
VVV
Liquid velocity
-1.50
-2.00
0
20.0
40.0
60.0
80.0
100.0
120.0
140.0
Time (s)
Figure 7.40 - Velocity in junctions 19 and 20 in 2 m initial water level case
WinGraf - 02-25-1999
225
200
XXX
YYY
ZZZ
VVV
JJJ
HHH
###
Quenching Rate (MW)
175
2m
1.29 m
X
150
Z
XYX Y Z
XYZ
XYZ
XYZ
XYZ
YXYZ
XYZ
XYZ
XYZ
XYZ
X YZ
XYZ
V
V
V
X
V
V
V
VV V V
V
V
J
1m V
J J J
J
J J
J J
J J
125
100
75
0.5 m
Z
YX
Z
Z
Y
2.37 m
Z
Y
X
Y
X
All void
V
V
2.969 m
V
#####
0
0
20.0
V
J
H
#
40.0
H
#
H
#
H
#
60.0
H
#
80.0
H
#
J
H
#
100.0
H
#J
V
H
#
120.0
Time (s)
Figure 7.41 - Quenching rate at different initial water levels
76
Y
X
J
J
HHH
Y Z
X
J
HHH
Y Z
X
J
50
25
QUENRATE
QUENRATE
QUENRATE
QUENRATE
QUENRATE
QUENRATE
QUENRATE
Z
Y
X
V
ACQPL2
ACQ2M37
ACQ2M
ACQ129M
ACQ1M
ACQ05M
VAPOR
V H
#
140.0
Assessment of Size Aspects in Modelling Molten Fuel Coolant Interaction
Chapter 7 - RESEARCH ACTIVITIES SUMMARY
core bottom level
3.5
3
lower plenum level
Mixture level (m)
2.5
2
1.5
Mixture Level
1
0.5
0
0
0.5
1
1.5
2
2.5
3
3.5
Initial W ater level (m)
Figure 7.42 - Mixture level vs. initial water level
WinGraf - 02-25-1999
1.60
XXX
YYY
ZZZ
1m
VVV
0.5 m
JJJ
HHH
###
H
HV
#
#
#
#
#
J
J
H
#
#
#
H
#
#
V
J
H
H
V
J
J
HHHHH
V
J J V V
J
J
V
JJ J J
V
J J J JJ
VV
VVV
2m
V
V
Z
V
V
VV
Z
1.29 m
Z
Z
Z
Y
Z
Y
ZZ
Z
Y
Z
Z
ZXY
ZXY
ZXY
YXY
XYXYXYXYX Y X Y X Y
X
X
X
X
X
All void
1.40
Void fraction (-)
1.20
1.00
.80
.60
.40
ACQPL2 AVGVOID
ACQ2M37 AVGVOID
ACQ2M
AVGVOID
ACQ129M AVGVOID
ACQ1M
AVGVOID
ACQ05M AVGVOID
VAPOR
AVGVOID
#H
#H
#H
#H
V
Z
Z
Z
Z
2.37 m
Y
Y
Y
#H
Z
2.969 m
X
X
X
X
X
.20
0
0
20.0
40.0
60.0
80.0
100.0
120.0
140.0
160.0
180.0
200.0
Time (s)
Figure 7.43 - Global mean void fraction in lower downcomer, core inlet and lower plenum at different initial
water levels.
77
Assessment of Size Aspects in Modelling Molten Fuel Coolant Interaction
Chapter 7 - RESEARCH ACTIVITIES SUMMARY
WinGraf - 02-25-1999
1.60
XXX
YYY
ZZZ
VVV
JJJ
HHH
###
#
#
H
H
1.40
All void
0.5 m
Cake mass ratio (%)
1.20
1.00
#
H
H
H#
H#
H #
H#
H#
H#
#
H
#
H
ACQPL2
ACQ2M37
ACQ2M
ACQ129M
ACQ1M
ACQ05M
VAPOR
#
H
CAKE
CAKE
CAKE
CAKE
CAKE
CAKE
CAKE
#
#
H
H
#
H
2.696 m
.80
2.37 m
1m
J J J
JJ
.60
J J
J
J
VVVV J J
V
J
V
VV J
J
VV
V
Z Z ZYZYZYX
Z YX
J
ZYX
Z
ZYX
J
J YX YX
ZJYX
ZYX
X
YXYX X X
V YX
ZYX
Z YX
Y X
Y V
Z
ZYX
X
V
Z
Y
Z
V
V
X
Z
Y
Z
V
V
V
X
1.29 m
2m
.40
.20
Y
V
X
Y
Z
0
0
20.0
40.0
60.0
80.0
100.0
120.0
140.0
Time (s)
140
100
120
90
100
80
80
70
60
60
40
50
Cake formation ratio (%)
Quenching rate (MW)
Figure 7.44 - Cake mass ratio at different initial water levels
Quenching rate
20
40
0
30
0
0.5
1
1.5
2
2.5
3
3.5
Cake mass flow
formation per
Melt mass flow
injected
Initial Water level (m)
Figure 7.45 - Quenching rate and cake mass flow per melt mass flow versus initial water level
78
Assessment of Size Aspects in Modelling Molten Fuel Coolant Interaction
Chapter 7 - RESEARCH ACTIVITIES SUMMARY
Boundary case: downcomer, riser, upper plenum, core inlet and lower plenum full of water
For the boundary case (riser, upper plenum and upper downcomer full of water) the calculation has been
continued up to 300 seconds to analyse the behaviour in the longer term, with particular interest on
quenching and mean void fraction behaviour.
Pressure (Figure 7.46) increases slower than in the other cases. The reason for the lower pressurisation is not
due to the smaller quenching rate achieved, but is due to the greater void condensation in the upper volumes.
Pressure tends to a stabilisation. As the injected mass flow rate decreases fragmentation stabilises, producing
constant quantity of drops.
Fragmentation in boundary case is similar to cases with higher initial water level.
The quenching rate achieved (Figure 7.47) is initially similar to cases with higher initial water level, after
about 5 seconds it can be observed an increasing of quenching due to drops accumulation in the core inlet
(Figure 7.48). Heat exchange does not disappear; it reaches a minimum steady value close to 12 MW at
about 200 seconds.
Figure 7.49 shows that natural circulation is established in this case; water is moving from lower downcomer
to core inlet through junctions number 20 (from lower downcomer to lower plenum) and junction number 19
(from lower plenum to core inlet). Liquid mass flow in those junctions cause drops accumulation in volumes
of core inlet, as it was observed in Figure 7.48. Natural circulation continues up to about 210 seconds. After
that time void fraction in lower downcomer and core inlet are balanced and no more natural circulation is
produced.
The global mean void fraction (Figure 7.50) shows for the boundary the same initial steady value of 0.4 than
cases with higher initial water level. In this case, the initial level (8.074 m) is much higher than the injection
point (2.969 m) and the mixture level, due to boiling, never goes below that point. Therefore, global mean
void fraction continues to decrease. A minimum steady value close to 0.07 is obtained at about 200 seconds.
The cake mass production is similar to cases with higher initial water level, increasing up to about 40% of
discharged mass.
Water is transferred mainly to upper head (volume 9) and some to the rest primary system (volume 10). The
void fraction in the upper head (volume 9) and in the rest of primary system (volume 10) is shown in Figure
7.51 for the boundary case in the longer term. The water, which had been transferred in the upper head
(volume 9), turns back into the riser, core inlet and lower plenum as the quenching rate and steam
evaporation decrease.
To summarise: In the boundary case pressure increase is lower due to condensation, quenching rate is higher
and drops mass, void fraction and cake mass accumulation are similar to cases with higher initial water level.
The difference in this case is the onset of natural circulation in the vessel that causes a redistribution of the
drops mass in the system.
79
Assessment of Size Aspects in Modelling Molten Fuel Coolant Interaction
Chapter 7 - RESEARCH ACTIVITIES SUMMARY
WinGraf - 02-25-1999
160
PRESSURE UPPER HEAD
140
XXX
YYY
ZZZ
VVV
JJJ
HHH
###
OOO
Pressure (bar)
120
100
80
ACQPL2
ACQ2M37
ACQ2M
ACQ129M
ACQ1M
ACQ05M
VAPOR
BOUND
Z
p009
p009
p009
p009
p009
p009
p009
p009
Y
Z
Z
Z
ZJ
VYV
V
ZYZ
XJ X
V
Y
Z
X
J
VYXJ
V
XJ
Z YZ
O
V
Z
Y
XJYXJ O O O
V
Z
J
Y
X
Z
V
J
O
X
H
Y
H
O
Z
H
XJ
XY
JV
V
YZ
HHHH# # #
XYX
###
60
Y
Z
V
Z
YXJ
Z
V
Z YX
V
J
YX
J
O
O
H
#
H
#
X
V
V
X
V
X
boundary
J
O
O
O
O
O
Y
J
J
J
O
O
YV
X
YV
X
Z
V
Y
X
Y
H
H
H
H
H
H
H
#
#
#
#
#
#
#
All void
40
0
20.0
40.0
60.0
80.0
100.0
120.0
140.0
Time (s)
Figure 7.46 - Pressure in the upper head in boundary case
WinGraf - 02-25-1999
250
225
Quenching Rate (MW)
200
175
X
XXX
YYY
ZZZ
VVV
JJJ
HHH
###
OOO
ACQPL2
ACQ2M37
ACQ2M
ACQ129M
ACQ1M
ACQ05M
VAPOR
BOUND
Z
Z
QUENRATE
QUENRATE
QUENRATE
QUENRATE
QUENRATE
QUENRATE
QUENRATE
QUENRATE
150
O
O
ZYZ
ZYX
Z
YX
ZOYX
ZO
Z YX
ZYX
YXYX
YX
X
O XO Y X
Y
Y
Z
X
X
Z
Y
ZZ
O
V
V
V
V
V
V
VV V V
V
V
J
J J J
J
J J
J J
J J
125
100
75
boundary
Z
YOX
Y
OX
Y
O X
V
V
Z
Y
O X
V
J
All void
50
Z
YO
X
HHHHHH
#####
0
20.0
Y Z
X
V
V
J
H
#
40.0
H
#
H
#
H
#
60.0
H
#
80.0
Y
X
O
V
H
#
J
H
#
H
#J
100.0
Time (s)
Figure 7.47 - Quenching rate in boundary case.
80
Y Z
J
J
0
O
O X
J
25
Z
Y
V
H
#
120.0
V H
#
140.0
Assessment of Size Aspects in Modelling Molten Fuel Coolant Interaction
Chapter 7 - RESEARCH ACTIVITIES SUMMARY
WinGraf - 02-25-1999
140
120
X
XXX
BOUND DROPSENUC
X
100
Mass of drops (kg)
X
80
X
X
X
X
60
X
X
X
X
40
X
X
X
X
X
X
X
X
X
X
20
0
0
5.0
10.0
15.0
20.0
25.0
30.0
35.0
40.0
45.0
50.0
Time (s)
Figure 7.48 - Drops mass in core inlet in boundary case
WinGraf - 02-01-1999
6000
4000
X
X
XXX
YYY
Junction Flow (kg/s)
X
BLONG mflowfj019
BLONG mflowfj020
Junction from lower plenum to core inlet
2000
X
X
X
X
Y
Y
Y
X
X
0
Y
Y
X
Y
Y
X
Y
Y
X
X
YX YX YX YX YX YX
Y
X
Y
Y
-2000
Y
Junction from lower plenum to lower downcomer
-4000
Y
-6000
0
50.0
100.0
150.0
200.0
250.0
300.0
Time (s)
Figure 7.49 - Liquid mass flow in junctions 19 and 20 in boundary case in the longer term
81
Assessment of Size Aspects in Modelling Molten Fuel Coolant Interaction
Chapter 7 - RESEARCH ACTIVITIES SUMMARY
WinGraf - 02-25-1999
1.80
XXX
YYY
ZZZ
VVV
JJJ
HHH
###
OOO
1.60
1.40
Void fraction (-)
1.20
1.00
ACQPL2
ACQ2M37
ACQ2M
ACQ129M
ACQ1M
ACQ05M
VAPOR
BOUND
AVGVOID
AVGVOID
AVGVOID
AVGVOID
AVGVOID
AVGVOID
AVGVOID
AVGVOID
V J HV # VH # H # H # H #
J H
VJ #
J V#
Z H
H#H#H#H # H #JH J#VH
Z
Z
J J V V
Z
J
J
J
Z
JJ
JJ JJJ VVVV V
Z
VV
V
Z
V
V
VV
Z
Y
All void
Z
Y
Z
Y
Z
Y
Z
Y
ZZ
Z
O
Y
Z
O
Z
O
ZZ Z
XXXXXXX
O
YYYYYYYY
XX
Y O
Y X OY XOY XO
XO X
O
X
X
X O X
X
X
OX
O
boundary
O
O
.80
.60
.40
.20
# H
#
O
0
0
25.0
50.0
75.0
100.0
125.0
150.0
175.0
200.0
225.0
250.0
Time (s)
Figure 7.50 - Global mean void fraction in lower downcomer, core inlet and lower plenum in boundary case
WinGraf - 02-01-1999
1.20
Rest primary system
1.00
Y
Y
Y
Y
Y
Y
Y
Void fraction (-)
.80
X
X
X
Y
X
X
X Y
YX Y
X
Y
X
X
X
Y
Y
Y
X
X
X
Y
Y
Y
X
Y
Y
X
.60
X
XXX
YYY
X
.40
X
X
BLONG voialf009
BLONG voialf010
Upper head
.20
0
0
50.0
100.0
150.0
200.0
250.0
300.0
Time (s)
Figure 7.51 - Void fraction in the upper head (9) and rest primary system (10) in boundary case in the longer
term
82
Assessment of Size Aspects in Modelling Molten Fuel Coolant Interaction
Chapter 7 - RESEARCH ACTIVITIES SUMMARY
Results of 2-d cases at different water levels (from 8.1 m down to 1m)
Pressurisation due to power exchange in 2-d cases (Figure 7.52) is higher in 2.969 m case (Base case),
reaching about 150 bar at 140 seconds. In “full case” and “half full case” the greater void condensation in the
upper volumes leads to a slower pressurisation as it occurred in 1-d boundary case. The pressure in the lower
level case (1 m case) after an initial increase trends to a stabilisation, at about 60 seconds, due to voiding of
lower plenum as it occurred in 1-d case.
The heat exchanged (Figure 7.53) increases quickly up to a similar value for “full”, “half full” and “base”
case, reaching about 160 MW. In the 1 m case the lower plenum is rapidly voided and fragmentation
produced is lower, corresponding to a lower quenching rate achieved.
In the cases “full”, “half full” and “base”, cases with the mixture level during the melt discharge up to or
above the injection point, the reduction of quenching rate occurs after about 50 seconds, when the melt mass
flow injected is also reduced. In the 1 m case this reduction of power exchange starts before, due to the lower
mixture level reached, below the injection point, and subsequent more void produced.
The fragmented mass (total drops mass, Figure 7.54) increases in a similar way for cases “full”, “half full”
and “base”. In 1 m case, fragmented mass increases up to 6000 kg and at about 60 seconds it stabilises due to
voiding of the lower plenum.
For 2-d cases the mean void fraction in lower plenum, core inlet and lower downcomer is presented as global
void fraction, which is the result of the heat transferred to all lower plenum, core inlet and lower downcomer.
It has been also calculated a local mean void fraction for volumes 19 to 26 (central lower plenum) and
volumes 58 to 62 (central core inlet) as central void fraction, which is responsible for the fragmentation
produced (fragmentation is assumed only in central part of lower plenum and central part of core inlet).
In Figure 7.55 and Figure 7.56 mean void fraction (global and central) for 2-d cases is shown. Central void
fraction increases rapidly until a steady value of 0.5 due to the heat transferred by drops fragmented in the
central volumes. The void fraction does not increase above this value and equilibrium is reached in which the
void fraction, the power exchanged and the pressurisation rate are almost constant. In the 1m case the initial
void fraction starts from a higher value and continues to increase, the mixture level is reduced and the lower
plenum is voided. In the base case after about 100 seconds void fraction starts to increase because the
mixture level becomes lower than the injection point. In “full” and “half full” cases the mixture level never
becomes lower than the injection point and void fraction does not increase.
Production rate of drops (kg/s) is observed in Figure 7.57 for the lower plenum and for the core inlet. In the
core inlet, in the first seconds, it becomes null in all cases. In “full”, “half full” and “base” cases production
rate is similar in the lower plenum and in 1m case it is lower, becoming zero at about 60 seconds, and
stabilising fragmented mass (Figure 7.54).
The un-fragmented mass produced (cake ratio, Figure 7.58) is similar for “full”, “half full” and “base” cases,
reaching about 50% of discharged mass in the first seconds. 1 m case reaches a higher cake mass production,
about 70%.
To summarise it can be said that “full”, “half full” and “base” cases showed similar energy exchange, drops
production and cake mass accumulation. In these three cases, the mixture level is up or above the injection
point (for base case up to about 100 seconds), allowing equilibrium in void fraction, heat exchange and
pressurisation rate. In 1 m case the mixture level is continuously reduced and lower plenum is voided, it
shows lower heat exchange and fragmentation than other cases and higher cake mass accumulation.
83
Assessment of Size Aspects in Modelling Molten Fuel Coolant Interaction
Chapter 7 - RESEARCH ACTIVITIES SUMMARY
WinGraf - 02-25-1999
160
Z
140
XXX
YYY
ZZZ
VVV
Pressure (bar)
120
2DASCOFULL
2DASCOHALF
2DASCOBASE
2DASCO1M
Z
p009
p009
p009
p009
Z
Z
Z
Z
Half full case
Z
Base case
Z
Z
100
Z
80
60
Y
Z
XV
X
Y
Z
V
X
Y
V
Z
X
Y
V
ZV Y
X
Z XVY
Z Y
V
X
Z
V
X
Z
VY
V
Y X
Y
Y
Y
Z
VX
VX
Y
V
V XV X
YV
V
Y
X
V
X
VY
X
X
Full case
Y
Y
Y
X
X
1 m case
40
0
20.0
40.0
60.0
80.0
100.0
120.0
140.0
Time (s)
Figure 7.52 - Pressure in the upper head in “full”, “half full”, “base” and 1 m cases with 2-d nodalization
WinGraf - 02-25-1999
225
Half base case
200
XXX
YYY
ZZZ
VVV
Base case
Quenching Rate (MW)
175
X
Y
150
ZX
X
125 Z
Y
Y
100
X
V
V
Z
75 V
50
Z
YZ
X
YZ
Z
X
Y
XY
Z
XYZ
XYZ
XYZ
XYZ
2DASCOFULL
2DASCOHALF
2DASCOBASE
2DASCO1M
XYZ
XY
Full case
V
V
V
ZX
Y
QUENRATE
QUENRATE
QUENRATE
QUENRATE
ZXY
ZX Y
V
ZX YZ
V
V
V
Y
XY
Z
V
V
X
1 m case
V
25
V
V
V
V
V
V
0
0
20.0
40.0
60.0
80.0
100.0
120.0
140.0
Time (s)
Figure 7.53 - Quenching rate in “full”, “half full”, “base” and 1 m cases with 2-d nodalization
84
Assessment of Size Aspects in Modelling Molten Fuel Coolant Interaction
Chapter 7 - RESEARCH ACTIVITIES SUMMARY
x 10 4
2
WinGraf - 02-25-1999
Base case
1.8
XXX
YYY
ZZZ
VVV
1.6
Mass of drops (kg)
1.4
2DASCOFULL
2DASCOHALF
2DASCOBASE
2DASCO1M
totmdrops
totmdrops
totmdrops
totmdrops
Z
1.2
1
.8
.6
.4
.2
Y
Y
Z
Z
X
XV
V
Y
V
V
X
Z
Y
0 X
0
Y
Z XV
ZX
V
Y
ZX
V
Y
ZX
Y
Z
XY
ZX
Y
ZX
V
V
Y
Y
ZXY
ZXY
ZX Y
Full case
Half full case
V
V
V
V
V
V
V
V
V
1 m case
V
20.0
Z
XY
ZX
ZX Y
ZX Y
ZXY
40.0
60.0
80.0
100.0
120.0
140.0
Time (s)
Figure 7.54 - Total drops mass in “full”, “half full”, “base” and 1 m cases with 2-d nodalization
WinGraf - 02-26-1999
1.20
GLOBAL VOID FRACTION
1 m case
1.00
Void fraction (-)
VV
.80 VV
V
V
V
V
V
V
V
V
V
XXX
YYY
ZZZ
VVV
V
V
V
V
2DASCOFULL
2DASCOHALF
2DASCOBASE
2DASCO1M
V
V
AVGVOID
AVGVOID
AVGVOID
AVGVOID
.60
Base case
Full case
.40
Z
.20
Y
Z
X
Y
Z
X
Y
X
Z
Y
0 X
0
Z
X
X
YZ
Y
X
Z
YZ
X
YZ
X
Y ZX Y Z X Y Z X Y Z
XY
XY
Z
XY
Z
Z
XY
XY
Z
Z
Z
XY
XY
Half full case
20.0
40.0
60.0
80.0
100.0
120.0
140.0
Time (s)
Figure 7.55 - Global mean void fraction in “full”, “half full”, “base” and 1 m cases with 2-d nodalization
85
Assessment of Size Aspects in Modelling Molten Fuel Coolant Interaction
Chapter 7 - RESEARCH ACTIVITIES SUMMARY
WinGraf - 02-26-1999
1.20
CENTRAL VOID FRACTION
1.00
1 m case
Void fraction (-)
.80 VV
VV
.60
.40
V
V
V
V
V
V
V
V
V
XXX
YYY
ZZZ
VVV
V
V
V
2DASCOFULL
2DASCOHALF
2DASCOBASE
2DASCO1M
V
V
V
AVGVOIDCENTR
AVGVOIDCENTR
AVGVOIDCENTR
AVGVOIDCENTR
Base case
X
Y
Z
Z
YX
X
Y Z Y ZX Y Z Y Z X Y X Y
Z
X
Z
Z X Y ZX Y Z X Y X Y Z
X
X Y ZX Y ZX Z
Z
Y XY X
Z
Full case
Z
Y
Z
XY
Half full case
.20
Y
0 X
0
20.0
40.0
60.0
80.0
100.0
120.0
140.0
Time (s)
Figure 7.56 - Mean void fraction in the central side of lower plenum and core inlet in “full”, “half full”, “base”
and 1 m cases with 2-d nodalization
WinGraf - 02-25-1999
250
225
Full case
XXX
YYY
ZZZ
VVV
Drops production (kg/s)
200
2DASCOFULL
2DASCOHALF
2DASCOBASE
2DASCO1M
PRODROPLWPL
PRODROPLWPL
PRODROPLWPL
PRODROPLWPL
175 X
V
Half full case
150
Z
125
100
Y
X
Z
Y
75
Z
Z
XY V XY
Y ZXY
V
V
V
XZY
V
V
Z
XY
YZ X Z
Y X
Y ZX Z
Z
Y
XY
XY
1 m case
V
V
V
Y
V
50 Z
25
Base case
XZ
Y X YZ X
Z
XY
Z
XY
XZY
XX
Z
Z
V
V
0
0
20.0
40.0
60.0
V
80.0
100.0
120.0
140.0
160.0
180.0
Time (s)
Figure 7.57 - Drops production rate in lower plenum in “full”, “half full”, “base” and 1 m cases with 2-d
nodalization
86
Assessment of Size Aspects in Modelling Molten Fuel Coolant Interaction
Chapter 7 - RESEARCH ACTIVITIES SUMMARY
WinGraf - 02-26-1999
1.20
XXX
YYY
ZZZ
VVV
Cake mass ratio (%)
1.00
2DASCOFULL
2DASCOHALF
2DASCOBASE
2DASCO1M
CAKE
CAKE
CAKE
CAKE
.80
1 m case
V
.60
V
V
V
V
V
V
V
V
V
V
V
V
V
V
V
Half full case
V
ZX Y ZX Y Z X Y Z X Y Z X Y Z X Y ZX Y Z X Y X Y
Z
Y
.40
.20
V
X
Z
Y
X
Z X Y Z X Y Z X Y ZX Y Z X Y X Y
Z
Z
Base case
Full case
Z
Y
ZX
V
0
0
20.0
40.0
60.0
80.0
100.0
120.0
140.0
Time (s)
Figure 7.58 - Cake mass ratio in “full”, “half full”, “base” and 1 m cases with 2-d nodalization
The thermal-hydraulic equilibrium plot
In this chapter the relation between the void fraction and the quenching rate is analysed to show why the
various cases trend to an equilibrium value, arriving to it in two different ways depending on the initial
conditions.
Let’s consider a system in which the power is imposed and the resulting mixture level is arbitrary. The power
is imposed only below the mixture level, Figure 7.59. At a certain power corresponds a certain void fraction
in the mixture. As the power applied increases, the void fraction increases. It is therefore possible to draw a
line relating the various power states and the related void fraction. However, if the same power is applied
and the mixture level is reduced, the global void fraction is higher for two reasons; firstly, because locally
the heat transfer coefficient is higher and secondly, because if the mixture level is lower the specific power
will be higher. Therefore; another line can be drawn corresponding to a different mixture level. Therefore a
family of curves exist corresponding to the various mixture levels. The upper bound curve is the one
corresponding to the maximum mixture level, i.e. up to the injection point.
Let us suppose now a different situation, in which melt is injected into a constant density medium. Imposing
an arbitrary void fraction to a steam-water system, the melt injected with a certain mass flow rate, will
fragment and create drops that will exchange heat with the medium. Therefore, it is possible to draw a line
corresponding to the power exchanged by the melt at a fixed injection mass flow rate and at a fixed void
fraction (independent of the power exchanged), Figure 7.60. If the mass flow rate is reduced, the
fragmentation is reduced and also the power exchanged; therefore another family of curves corresponding to
various injected mass flow rates can be generated.
The two simultaneous processes described in Figure 7.59 and Figure 7.60 can be merged in the chart of
Figure 7.61. At higher mass flow rate and higher mixture level corresponds the equilibrium point A. If the
system is so that a reduction of the mass flow rate occurs with a constant mixture level, the system will move
from point A to point B', on the left-down of point A. If instead the mixture level is decreasing with a
constant injection flow, the point will move from point A to point B", i.e. right-down. Of course, if both the
injection flow and the mixture level are reduced, the representative point will move from A to B"'.
87
Assessment of Size Aspects in Modelling Molten Fuel Coolant Interaction
Chapter 7 - RESEARCH ACTIVITIES SUMMARY
The cases analysed so far indicate that equilibrium in terms of injected mass, quenching rate and void
fraction is reached after some initial seconds of stabilisation. This process can be well seen in a plot showing
the observed quenching rate versus mean void fraction.
Figure 7.62 and Figure 7.63 show the calculated behaviour for the reactor in the plane void fraction-power
for cases performed in 1-d and 2-d, respectively.
In Figure 7.63 the curve A represents the power that is determined by the fragmentation process as a function
of the void fraction at an injected mass flow rate of 250 kg/s, while curve B represents the equilibrium void
fraction at various imposed power levels (distributed uniformly in the central part of a lower plenum) with
the maximum mixture level in saturation conditions.
Figure 7.62 represents in the plane void fraction-power all the 1-d cases. The three cases with higher initial
water level, boundary case (BOUND), all lower plenum and core inlet full (ACQPL2) and 2.37 m initial
water level (ACQ2M37) behave in the following way: they start with void fraction and power null (BOUND
and ACQPL2) and void fraction 0.25 and power null (ACQ2M37), as the transient proceeds, the power
exchanged and the void fraction increase tending to an equilibrium point in all cases (void fraction ≈ 0.4 and
power ≈ 120 MW).
Since the mixture level is constant in the period of constant injection, the three cases will move down-left
(like A to B'), because the reduction of the injected mass flow rate occurs. In the boundary case, the mixture
level never becomes lower than the injection point and the behaviour continues along the curve B, while in
the 2.969 m initial water level (ACQPL2) case after about 150 s the representative point moves down (like A
to B"'), because the mixture level became lower than the injection point and void fraction below this point
starts to increase. The same occurs for 2.37 m initial level case, but earlier, at about 60 seconds.
In the other cases, those with lower initial water level (2 m, 1.29 m, 1 m 0.5 m and 0.26 m), at the beginning
the representative point on the thermal-hydraulic plot is a not null void fraction and power 0. As the transient
proceeds, the power exchanged and the void fraction increase tending to an equilibrium point but on different
B-type curves. All the representative points move down-right (like A to B"), because, due to boiling the
mixture level continues to decrease and void fraction increases.
Figure 7.63 represents in the plane void fraction-power all the 2-d cases. In the “full” case, “half full” case
and “base” case, at the beginning the representative point on the thermal-hydraulic plot is void fraction and
power null; as the transient proceeds, the power exchanged and the void fraction increase tending to an
equilibrium point in all three cases (void fraction ≈ 0.5 and power ≈ 160 MW), then they will move downleft (like A to B'), because the reduction of the injected mass flow rate occurs. The base case, after 100 s,
moves down (like A to B"'), because the mixture level became lower than the injection point. The half and
full cases will continue along the so called curve B, because the mixture level never becomes lower than the
injection point.
In 1 m initial water level case, at the beginning the representative point on the thermal-hydraulic plot is the
void fraction 0.7 and power 0. The representative points in this case of low level move down-right (like A to
B"), because due to boiling, the mixture level continues to decrease.
To summarise: All the cases with mixture level higher than the orifice injection will behave similarly with an
equilibrium point which moves down-left (like A to B'), while the cases with lower mixture level will move
down-right (like A to B") because of reduced inventory in the interaction zone. Depending on the case, in
some cases with initially mixture level higher than injection point, this mixture level starts to reduce and void
fraction starts to increase after a particularly time depending on the initial water level and then they move
down right (like A to B"').
88
Assessment of Size Aspects in Modelling Molten Fuel Coolant Interaction
Chapter 7 - RESEARCH ACTIVITIES SUMMARY
Increasing
mixture level
A, higher mixture level
Power (MW)
A
A
B
Average Void fraction (-)
B, lower mixture level,
same power
Figure 7.59 - Resulting void fraction by applying a constant power
Power (MW)
A
A
Increasing melt injection
mass flow rate
B
Average Void fraction (-)
B, lower injection mass flow,
same void fraction
Figure 7.60 - Resulting power exchanged fragmenting melt in a constant void fraction medium
89
Assessment of Size Aspects in Modelling Molten Fuel Coolant Interaction
Chapter 7 - RESEARCH ACTIVITIES SUMMARY
Power (MW)
Higher injected
mass flow rate
A
B'
Higher mixture
level
B''
B'"
Average Void fraction (-)
Figure 7.61 - Power exchanged versus void fraction plan
250
Power (MW)
200
BOUND
ACQPL2
ACQ2M37
ACQ2M
ACQ129M
ACQ1M
ACQ05M
VAPOR
150
100
60 s
50
150
0
0
0.2
0.4
0.6
0.8
1
Void Fraction (-)
Figure 7.62 - Quenching rate versus global mean void fraction in cases 1-d
90
Assessment of Size Aspects in Modelling Molten Fuel Coolant Interaction
Chapter 7 - RESEARCH ACTIVITIES SUMMARY
Power (MW)
250
CURVE B
Void fraction resulting
from power imposed
200
100 s
CURVE A
Power resulting at
fixed void fraction
with 250 kg/s of
melt injected
150
FULL CASE
HALF FULL
CASE
BASE CASE
100
1 M CASE
50
0
0
0.2
0.4
0.6
0.8
1
Void fraction (-)
Figure 7.63 - Quenching rate versus global mean void fraction in cases 2-d
7.3.4.
CONCLUSIONS
As conclusions of the study, the following items can be pointed:
All the cases show decreasing fragmentation and energy production as the initial water level in the core inlet
and lower plenum is reduced. With an initial water level below a value between 1 m and 0.5 m, the
fragmentation and the quenching rate progressively decrease and the cake mass accumulation increases.
In the boundary case pressure increase is lower due to condensation, quenching rate is higher and drops
mass, void fraction and cake mass accumulation are similar to cases with higher initial water level. The
difference in this case is the onset of natural circulation in the vessel that causes a redistribution of the drops
mass in the system.
The relation between the void fraction and the quenching rate has been clarified. An equilibrium value is
generally attained and is possible only if increase and decrease of void fraction is allowed; that is if the
mixture level is at or above the injection point (flooded regime). On the contrary, if the mixture level is too
low, the power exchanged will continuously decrease and tend to zero as the water level decreases (depleted
regime).
91
Assessment of Size Aspects in Modelling Molten Fuel Coolant Interaction
Chapter 7 - RESEARCH ACTIVITIES SUMMARY
7.4. COMETA CODE CALCULATIONS OF THE FARO QUENCHING TESTS
7.4.1.
INTRODUCTION
The work has been focused in the simulation (with COMETA thermalhydraulic code) of the MFCI test series
carried out in the FARO facility. The objective is to analyse and to unify the experiment simulations with the
last available version of the COMETA code.
The interaction of big melt masses in the water under realistic accident conditions was the objective of the
FARO experiments. FARO facility began the experiments of the LWR-MFCI phenomena in 1991. The last
test was performed in 1999.
In this context the COMETA thermalhydraulic code was developed in order to support test preparation and
execution with pre-test calculations and simulation, assist test results interpretation with post-test
calculations and perform sensitive analyses to point out the influence of various parameters on the test
facilities response.
During the time the COMETA code models and input decks changed as a result of additional analyses and in
some cases of errors correction.
In order to analyse and unify the COMETA simulations carried so far, it was decided to simulate again all
the FARO experiments with the last available version of the COMETA code and keeping as far as possible
the same basic nodalization scheme for all the tests. Only changes due to different boundary conditions of
facilities arrangements were performed. This chapter summarises the Tests calculations (L-14, L-19, L-24,
L-27, L-28, and L-31). Test L-33 was not included in this document because, due to the specific nature of the
test (steam explosion) the analysis was performed in a separate document.
7.4.2.
EXPERIMENTS IN THE FARO FACILITY
The series of experiments in the FARO facility were part of the research activities in severe accidents in
LWR carried out by the Institute for Systems, Informatics and Safety (ISIS) of the JRC. The reference
situation is an accident in which one or more mass melt jets pour into the water of the lower plenum,
fragment and settle down in the bottom. The objectives of the tests were to evaluate the steam generation rate
associated to the melt heat transfer (quenching), to evaluate the hydrogen production associated to Zr
oxidation, the heating of the bottom vessel structures and to evaluate the debris bed in the bottom plate.
The experiments simulated with the last version of the COMETA code are presented in Table 7.7, which
summarises the experimental conditions for each experiment. From Test L-14 to L-24 the TERMOS test
section was used (0.71 m of diameter) (Figure 5.3). From Test L-27 it was changed to FAT test vessel
(Figure 5.4), which consists of a pressure vessel of 1.5 m internal diameter and 2 m high, designed for a
pressure of 8 MPa and a temperature of 300°C. To compare the results with the ones of the previous
configuration TERMOS, an internal cylinder was inserted in the FAT vessel. The cylinder internal diameter
was the same of the previous TERMOS vessel, 0.71 m. This cylinder is filled with water and the outer
annular space is part of the free board volume.
Tests L-14, L-19 and L-20 were performed under a relative medium and high pressure, lower pressure was
imposed from Test L-24 until L-33. Test L-31 and L-33 were the only ones performed under subcooled
conditions in the pool of water. In Test L-33 an external trigger was applied at 1.1 s. Due to the peculiar type
of phenomena, this test has not been included in this study.
Test L-28, L-31 and L-33 were performed with a 5 cm melt orifice discharge, the rest with a larger one (10
cm). This reduction determined a longer discharge time of the melt.
92
Assessment of Size Aspects in Modelling Molten Fuel Coolant Interaction
Chapter 7 - RESEARCH ACTIVITIES SUMMARY
EXPERIMENT
L-14
L-19
L-20
L-24
L-27
L-28
L-31
L-33
(**)
Initial pressure (bar)
50.4
50.4
20.4
5.1
4.9
5.1
2.2
4.1
Mass (kg)
125
157
96
177.1
128.7
174.9
92
100
Melt temperature (K)
~3073
3073
3173
3023
3023
3052
2988
3073
Initial water level (m)
2.05
1.1
1.97
2.02
1.47
1.44
1.45
1.62
Initial water temperature
Tsat
Tsat
Tsat
Tsat
Tsat
Tsat
290.8 K
294 K
Free fall in gas (m)
1.04
1.99
1.115
1.065
0.73
0.89
0.77
0.77
Injection elevation (m)
3.09
3.09
3.085
3.085
2.2
2.33
2.22
2.39
Orifice diameter (m) (*)
0.092
0.092
0.092
0.094
0.094
0.044
0.046
0.044
TERMOS
TERMOS
TERMOS
TERMOS
FAT
FAT
FAT
Vessel geometry
Test date
June
23rd1994
nd
June 22
1995
January
30th 1996
December
5th 1996
rd
Dec. 3
1997
nd
Apr 2
1998
FAT
th
Nov 11
1998
July 1st
1999
Table 7.7 - Experimental conditions in the series of experiments in the FARO facility
(*) The initial orifice diameters are 0.1 m and 0.05 m but due to the formation of crusts before the melt
injection, at the time of release, the effective orifice diameter is smaller.
(**) This test has not been included in this study.
93
Assessment of Size Aspects in Modelling Molten Fuel Coolant Interaction
Chapter 7 - RESEARCH ACTIVITIES SUMMARY
7.4.3.
COMETA CODE SIMULATIONS OF THE FARO TESTS
Calculations were performed with the last version of the COMETA code. The original jet fragmentation
model was applied. The IKE jet fragmentation model, developed at the University of Stuttgart, Germany,
was also tested.
Other calculations were performed for low pressure tests to account for the initial higher fragmentation rate.
The Enhanced fragmentation model, in which a lower critical Weber number multiplier and a lower jet
break-up length coefficient were used during the initial phase of calculation (i.e. less than 1s). After that
time, the standard values adopted for high pressure calculation were used (*).
The following Table describes the COMETA calculations performed:
EXPERIMENT
Descriptor 1d
original
COMETA model
Descriptor 2d
original
COMETA model
Descriptor 1d
IKEJET
fragmentation
model
Descriptor 2d
IKEJET
fragmentation
model
Descriptor 1d
ENHANCED
fragmentation
model
Descriptor 2d
ENHANCED
fragmentation
model
L-14
L-19
L-20
L-24
L-27
L-28
L-31
L14N_1D
L191D05A
L20N_1D
L24N_1D
L27N_1D
L28N_1D
L31N_1D
L14N_2D
L192D05A
L20N_2D
L24N_2D
L27N_2D
L28N_2D
L31N_2D
L141DIKE
L191DI0A
L201DIKE
L241DIKE
L271DIKE
L281DIKE
-
L142DIKE
L192DI0A
L202DIKE
L242DIKE
L272DIKE
-
-
-
-
-
L241DENH
1D_PROV3
L281DENH
-
-
-
-
L242DENH
PROVA3
L282DENH
-
Table 7.8 - Experiment calculations with COMETA code
The detailed calculation results [Publication [3]] are not presented in this document due to the big amount of
information, only summary results, general conclusions and overall recommendations extracted from the
results are commented.
(*) It was never clarified if the initial higher fragmentation, which was measured from Test L-24, was due to
the lower pressure conditions or due to the melt release mode. The priorities in other research topics and the
anticipated closure of the FARO project prevented a clear understanding of this issue, which could have been
resolved by performing a new test at high pressure with the new melt release device.
94
Assessment of Size Aspects in Modelling Molten Fuel Coolant Interaction
Chapter 7 - RESEARCH ACTIVITIES SUMMARY
7.4.4.
EFFECT OF THE NODALIZATION SCHEMES ON THE RESULTS
The next table shows a first assessment summary of the overall results after performing all the Test
simulations. The objective is to give some general conclusions and to help to choose the nodalization scheme
to reduce the computer calculation time. It may be helpful to simulate other similar quenching cases.
IKEJET
FRAGMENTATION
MODEL
1d ? 2d: effect of
vapour-liquid
junction velocity
PRESSURE
ORIFICE
DIAMETER
WATER
CONDITIONS
COMETA
ORIGINAL MODEL
High
10 cm
Saturated
1d nodalization ~ 2 d
nodalization
High
5 cm
Saturated
Not assessed
Not assessed
-
High
5 cm
Subcooled
Not assessed
Not assessed
-
L-24 and
L-27
Low
10 cm
Saturated
1d ? 2d: effect of
vapour-liquid
junction velocity
TERMOS and
FAT
L-28
Low
5 cm
Saturated
Only performed 1d
nodalization
FAT
L-31
Low
5 cm
Subcooled
Not performed
FAT
TEST
L-14,
L-19 and
L-20
FARO
Test
missing
FARO
Test
missing
Very good results
using 2d Enhanced
frag.
1d not defined
Better results with 2d
nodalization
Good calculation
results.
1d nodalization ~ 2 d
nodalization
VESSEL
GEOMETRY
TERMOS
Table 7.9 - Assessment summary of the simulation results
•
Tests performed at high pressure, 10 cm orifice diameter and saturated initial water conditions: no great
differences on the main quantities (pressure, energy, quenching fragmented mass) are observed between
calculations using 1d or 2d nodalization scheme with the original COMETA model. With the same initial
conditions, the COMETA-IKEJET model shows differences between 1d and 2d nodalization scheme:
calculations with the 2d COMETA-IKEJET model show an initial pressure high jump due to the effect
on this 2d model of the liquid-vapour junction velocity.
The influence on the results of the melt discharge orifice diameter was not performed at high pressure, in
saturated or subcooled conditions because no FARO Test with 5 cm orifice diameter was available in these
conditions.
•
Tests performed at low pressure, 10 cm orifice diameter and initial water saturated conditions: Very
good results are obtained using the 2d Enhanced fragmentation model. 1d nodalization does not show
differences respect to 2d nodalization in Test L-27. Calculations with the 2d COMETA-IKEJET model
show an initial pressure high jump due to the effect on this 2d model of the liquid-vapour junction
velocity, 1d and 2d nodalization show differences.
•
Tests performed at low pressure, 5 cm orifice diameter and initial water saturated conditions: There are
great differences on the main quantities between 1d and 2d COMETA original model, 2d calculations are
much better. The same is true using the Enhanced fragmentation model.
•
Tests performed at low pressure, 5 cm orifice diameter and subcooled initial water conditions: In this
Test only COMETA original model was used for the simulations. Using 1d or 2d nodalization scheme it
shows no major differences in the main results although they are, in some cases, seen amplified by the
small plotting scale i.e. pressure.
So, in general it seems better to use 2d nodalization scheme for Test performed at low pressure, saturated
conditions, and small diameter. At high pressure and saturated conditions it is better to use a simpler 1d
95
Assessment of Size Aspects in Modelling Molten Fuel Coolant Interaction
Chapter 7 - RESEARCH ACTIVITIES SUMMARY
nodalization, because there are not great differences versus 2d and it allows reducing the computer
calculation time.
7.4.5.
CONCLUSIONS
The work was focused in the simulation of the quenching test series carried so far in the FARO facility (L14, L-19, L-24, L-27, L-28, and L-31) with the last version of the COMETA thermalhydraulic code. The
objective of the simulation was to keep as far as possible the same basic nodalization scheme code models
and options for all the tests. Only changes due to different boundary conditions were included. Test L-33 was
not included in this document because, due to the specific phenomena included in the test (steam explosion)
the analysis was performed in a separate document.
The COMETA original jet fragmentation model was applied to all Tests. The COMETA-IKEJET
fragmentation model, developed at the University of Stuttgart, Germany, was also tested.
The Tests L-14, L-19 and L-20 were performed with water at the beginning in saturated conditions above 20
bar in the TERMOS vessel.
a) In 1d calculations the rising ramp and the overall trend for pressure are well predicted; long term values
are also close to experimental value, except for L-19 that are slightly higher. Energy release and
quenching rate overall trend are well predicted by the calculations; long term values for L-20 are slightly
lower. The calculated fragmented mass is slightly underpredicted in some cases.
b) In 2d calculations pressure overall trend is well predicted; calculations with the COMETA-IKEJET
model show an initial high jump due to the effect on this model of the liquid-vapour junction velocity.
Energy release and quenching rate overall trend are well predicted by the original COMETA and
IKEJET-COMETA models; long term values for L-20 are slightly lower. The fragmented mass in the
COMETA-IKEJET calculations shows values closer to the experimental ones.
In Tests at high pressure, with 10 cm orifice diameter and saturated water initial conditions, 1d and 2d
nodalization differences are shown in Table 7.9.
The Tests L-24, L-27, L-28 and L-31 were performed below 20 bar: L-24 was performed in the TERMOS
vessel, the others in the FAT vessel. All these Tests were performed with water at the beginning in saturated
conditions, with the exception of Test L-31, in which the water was subcooled. Some calculations were
performed to account for the initial higher fragmentation rate (Enhanced fragmentation model).
a) In 1d calculations the initial pressure rising ramp is very well predicted with the Enhanced fragmentation
model. In Test L-27, the long term behaviour with the Enhanced fragmentation model is also very well
predicted. COMETA original calculation for subcooled Test L-31 shows pressure value very close to the
experimental one. In general, the pressure increase for subcooled tests is much lower than for saturated
cases. In Test L-31 the differences between experimental and calculated values are less than 0.2 bar.
Energy release and quenching rate show values very well predicted by the Enhanced fragmentation
model in Test L-27 and by the original COMETA model in Test L-31.
b) In 2d calculations the pressure overall trend is very well predicted with the Enhanced fragmentation
model; calculations using the COMETA-IKEJET model show in Test L-24 pressure an initial high jump
due to the effect of the liquid-vapour junction velocity. As it was shown for 1d case, 2d COMETA
original model calculation for subcooled Test L-31 shows pressure value very close to the experimental
one. Energy release, quenching rate and fragmented mass show calculated values closer to the
experimental ones with the Enhanced fragmentation model and COMETA-IKEJET model in Tests L-24,
L-27 and L-28.
In Tests at low pressure, 10 or 5 cm orifice diameter and saturated or subcooled water initial conditions, 1d
and 2d nodalization differences are shown in Table 7.9.
96
Assessment of Size Aspects in Modelling Molten Fuel Coolant Interaction
Chapter 7 - RESEARCH ACTIVITIES SUMMARY
In general, as a main conclusion: the COMETA code was successfully applied to all the FARO
configurations reaching the main objective: to analyse and unify the COMETA calculations carried so far.
Also some indicative advises, related to the model or the 1d - 2d nodalization, are given. The post-tests
simulations also gave significant information for the assessment and qualification of the COMETA code.
97
Assessment of Size Aspects in Modelling Molten Fuel Coolant Interaction
Chapter 7 - RESEARCH ACTIVITIES SUMMARY
7.5. CHARACTERIZING MELT FUEL-COOLANT
CONTEXT WITH RELAP5/SCDAP 3.2 CODE
7.5.1.
INTERACTION
IN
A
NPP
INTRODUCTION
After concluding FCI studies from the reactor vessel standpoint, a wider overview of these phenomena and
others related to a severe accident may be of interest. Because of this it was decided to complete the research
studying the severe accident progression in a nuclear power plant. This work may help to clarify the
importance of the melt fuel-coolant interaction (MFCI) in the context of a severe accident.
The study presented aims to illustrate the MFCI inside the general overview of a severe accident sequence
and the overall view of the NPP.
RELAP5/SCDAP 3.2 is the code chosen for the calculations [31] [32] [48] [50]. It is being used by the
Technical University of Catalonia (Universitat Politècnica de Catalunya, UPC) thanks to the agreement
between the Spanish Nuclear Security Council (Consejo de Seguridad Nuclear, CSN) and the American
Idaho Innovative Systems Software (ISS) Company.
Three objectives are aimed to reach:
a) Deeper understanding of the MFCI modelling theory of RELAP5/SCDAP 3.2 and how the code models
all the associated phenomena to the melt fuel coolant interaction.
b) Study the influence of the “low pressure injection system” (LPIS) in the sequence of a large “loss of
coolant accident” (LOCA) in the cold leg, followed by core uncover and degradation. It is also interesting to
study up to what point it is possible to minimize core degradation and to avoid its meltdown (with the LPIS)
and in the same way to quantify the LPIS effect on the cladding oxidation, hydrogen production and core
additional heating due to the exothermic reaction.
c) Obtain the initial conditions (pressure, coolant temperature, molten mass temperature, molten mass
composition, water level in the lower plenum, etc.) just before the molten pool slumping into the lower
plenum in order to introduce them in a specific MFCI code input (as JRC-COMETA [6]) so a detailed MFCI
study could be later performed.
A generic PWR NPP (3 loops, 1000 MWe) is chosen to perform the calculations.
7.5.2.
THE LOSS OF COOLANT ACCIDENT (LOCA)
In the event called “loss of coolant” [26], the reactor cooling fluid is escaping from a leak or a break within
the primary cooling system and can cause an imbalance between heat production and heat removed leading
to a potential risk of inadmissible high fuel temperatures.
A loss of coolant caused by a major rupture within the primary cooling system is one of several “limiting
fault conditions occurrences”. These are faults, which are not expected to occur but are postulated because
their consequences could include the risk of a major release of radioactive material.
The loss of coolant accident (LOCA) caused by a double ended break of a main coolant inlet pipe (cold leg
between pump and reactor pressure vessel, RPV) of the primary cooling system of a PWR is the most
dangerous fault against which the safety systems have to be designed; it is a design basis accident (DBA)
determining the specifications for the limiting, safety related design of the reactor plant, particularly of the
emergency core cooling system (ECCS).
98
Assessment of Size Aspects in Modelling Molten Fuel Coolant Interaction
Chapter 7 - RESEARCH ACTIVITIES SUMMARY
The severity of a LWR LOCA in the primary cooling system and its possible consequences is essentially
determined by the rupture characteristics in terms of break location and break size.
Possible break sizes range from a double ended or guillotine rupture of a main coolant pipe which is referred
to as a 2x100% (of the main coolant pipe flow cross section) break, down to a break of a single steam
generator tube (about 0.1%).
Within this wide rupture size spectrum, a distinction can be made between small (<2%, <80 cm2 break area),
intermediate (2%-10%, 80-400 cm2 break area) and large (>10%, >400 cm2 break area) LOCAs according to
break size. The quantitative values for the break size area and percentage should not be considered fixed;
they certainly vary between plants from different manufacturers.
In addition, for the case of a rupture of the main coolant piping, distinction is made between cold leg
ruptures, pump and reactor pressure vessel (RPV) or between steam generator (SG) and pump, and hot leg
ruptures between RPV and SG.
In general the ECCS are designed to cope any of these type of accident.
For the calculations here considered the loss of coolant accident (LOCA) is caused by a break with the 60%
flow area (0.23 m2) of a main coolant inlet pipe (cold leg between pump and reactor pressure vessel, RPV) of
the primary cooling system of a PWR in which not all the emergency systems are in operation. As a
consequence the core experiences a severe accidental sequence.
7.5.3.
THE RELAP5/SCDAP 3.2 CODE
The Idaho National Engineering and Environmental Laboratory (INEEL) initially developed the
RELAP5/SCDAP code [38][39][40] for the USA NRC. It was designed for best-estimate transient
simulation of light water reactor coolant systems during a severe accident.
The RELAP5/SCDAP code models the coupled behaviour of the reactor coolant system, the core kinetics,
fission products released during a severe accident transient 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; and SCDAP models the core melting phases up to melt relocation in the lower
head.
The core degradation includes models of cladding oxidation, materials relocation in the core, cladding
rupture, molten pool formation and its relocation in the lower head.
Calculation results (at the moment of the relocation) performed with codes as RELAP5/SCDAP 3.2 could be
seen as initial conditions for JRC-COMETA code calculations to simulate detailed MFCI phenomena.
RELAP5/SCDAP 3.2 modelling theory in the late phase damage progression
The uncertainties involved in modelling the late phase damage progression make it useful to perform
bounding studies on the calculated times of molten pool slumping and failure of the lower head, that include
the next areas: strength and configuration of solidified material that supports a pool of molten core material,
fragmentation temperature of embrittled fuel rods that are quenched, configuration of slumping molten
material, heat transfer coefficient between debris and the lower head of the reactor vessel.
The code user can define a parameter for each of these areas of modelling, so that a series of analyses can be
performed to bind the possible behaviour of the reactor.
From the code version 3.1, in the models that determine when material from a molten pool in the core region
slumps from this region to the lower head of the reactor vessel three events are considered to trigger this
slumping:
99
Assessment of Size Aspects in Modelling Molten Fuel Coolant Interaction
Chapter 7 - RESEARCH ACTIVITIES SUMMARY
1- The first event is the melting of structural failure of the vertically oriented crust at the periphery of the
core that was containing the molten pool and keeping it from slumping into the core bypass region.
2- The second event is the structure failure of the crust at the top of the molten pool that was supporting core
material above the molten pool. This event is considered to cause slumping of the molten pool to the lower
head only when the molten pool had already spread to the periphery of the code.
3- The third event is the propagation of the molten pool to the elevation of the bottom of the core.
The calculated time of slumping of a molten pool to the lower head should be regarded as having a large
degree of uncertainty, it is very difficult to predict even for a well defined manufactured structure. This large
uncertainty in the calculation of crust failure requires that the severe accident analyst examine the effect of
the failure criteria upon the calculated course of a severe accident. It is also possible to set this slumping time
and the duration of the slumping (10s by default) by the code user as an input parameter.
The code user also defines the degree of interaction of the slumping material with the water through which it
falls.
If the user defines that no interaction takes place, then the slumping core material is considered to fall with
no transfer of internal energy from the slumping material to water. The slumped material that accumulates in
the lower head is considered to have no porosity.
If the user defines a complete interaction between slumping material and the water through which it falls,
then the slumping material is considered to break up into small particles and transfer all of its internal energy
instantly to the water, then the molten material is cool when it impacts the lower head. By default the
slumped material that accumulates in the lower head is considered to have a porosity of 0.5 and a particle
size of 10 mm. These values are an input parameter than could be settled by the code user. The transfer of
internal energy from the slumping material to water may cause a large amount of vapour generation and may
significantly increase the pressure in the reactor vessel.
Since that time interval of the slumping is arbitrary, the pressure increase caused by the energy addition may
or may not be conservative. However, the total energy added to the water is still conserved. The option to
predict the minimum heat transfer to the water assumes that all of the stored energy remains in the melt and
is only removed by heat transfer from the surfaces of the resulting debris in the lower plenum. This option is
the most conservative in a possible study about lower head failure and want the maximum amount of energy
in the melt when it interacts with the vessel.
7.5.4.
NODALIZATION. ANALYSED CASES
For the calculations the loss of coolant accident (LOCA) chosen is caused by a break with the 60% flow area
(0.23 m2) of a main coolant inlet pipe of the primary cooling system. Break transient time is 460s.
The nodalization adopted in all calculations for some NPP parts (the reactor vessel, break and three loops
LPIS) is shown in Figure 7.64. It includes a detailed core nodalization in 5 thermalhydraulic radial channels,
two for the centre core (number 111 and 114), two for the middle core (112 and 115) and one for the outer
core (113) connected by cross-flow junctions and divided in 10 axial nodes, numbered from 1 (core bottom)
to 10 (core top). Each core channel includes two SCDAP components, one that simulates fuel rod
(components number 1, 3, 5, 7, 9) and other that simulates control rod (2, 4, 6, 8, 10). The total reactor
thermal power is 2.686e9 W.
In all cases performed, accumulators are available; they open below about 5 MPa. In the so-called “Base
Case” (DEBRIS_NOBK, DEBRIS_BK) low pressure injection system (LPIS) is not available, so core
degradation and meltdown occurs. The other cases include the manual action of LPIS system at different
times after the break (from 15 min up to 34 min).
100
Assessment of Size Aspects in Modelling Molten Fuel Coolant Interaction
Chapter 7 - RESEARCH ACTIVITIES SUMMARY
For the “Base Case” two calculations were performed (DEBRIS_NOBK, DEBRIS_BK) in order to
highlight the difference in the degree of interaction (of the slumping material with the water through which it
falls) that could be obtained setting two different values in an input parameter.
267
264
261
267
264
108
220
218
901
106
222
Break
913
Figure 7.64 - Vessel, LPIS and break RELAP5/SCDAP nodalization
The following Table 7.10 summarises the performed cases. These cases provide the initial conditions that
could be set in an input code for a detailed MFCI study, like JRC-COMETA code [6]. Initially cases with
LPIS were performed in order to study which is the maximum LPIS action time that prevents core meltdown.
Descriptor
Break size (%)
LPIS
LPIS injection time after
transient beginning (t=0s)
LPIS injection time
after break (t=460s)
DEBRIS_NOBK
(Base Case)
60
NO
-
-
DEBRIS_BK
(Base Case)
60
NO
-
-
LPIS_2000
60
YES
2000s
1540s
LPIS_1500
LPIS_1300
LPIS_1200
LPIS_1150
LPIS_1125
LPIS_1121
LPIS_1120
LPIS_1119
LPIS_1115
LPIS_1110
LPIS_1100
LPIS_900
60
60
60
60
60
60
60
60
60
60
60
60
YES
YES
YES
YES
YES
YES
YES
YES
YES
YES
YES
YES
1500s
1300s
1200s
1150s
1125s
1121s
1120s
1119s
1115s
1110s
1100s
900s
1040s
840s
740s
690s
665s
661s
660s
659s
655s
650s
640s
440s
Calculation purpose
Understand
RELAP5/SCDAP 3.2
model for MFCI.
Initial conditions for
COMETA code
Understand
RELAP5/SCDAP 3.2
model for MFCI
LPIS influence. Initial
conditions for
COMETA code
"
"
"
"
"
"
"
"
"
"
"
"
Table 7.10 - Performed RELAP5/SCDAP calculations
101
200
300
400
174
163
173
183
118
104
LPIS
182
322
114
112
115
113
320
318
172
267
364
171
151
361
111
LPIS
152
102
181
100
153
422
162
420
418
164
190
161
261
154
LPIS
Assessment of Size Aspects in Modelling Molten Fuel Coolant Interaction
Chapter 7 - RESEARCH ACTIVITIES SUMMARY
7.5.5.
CALCULATION RESULTS
Calculation results are presented in four different chapters:
The first one (a) discusses the two calculations of the “Base Case” (DEBRIS_NOBK, DEBRIS_BK) in
order to study the SCDAP models for a possible MFCI of the slumping material towards the lower plenum
with the water through which it falls and to show the difference in the degree of interaction that could be
obtained setting two different values in an input parameter.
The second and third chapter (b) and (c) show calculation results for all cases in which LPIS does not avoid
(after core degradation) molten pool formation and subsequent slumping to the lower plenum. That is up to a
LPIS injection time of 660s (1120s from t = 0). In chapter (b) all cases that are clearly far from core recovery
are presented: from case LPIS_1150, injection time 690s after the break (1150s from t = 0) up to LPIS_2000,
injection time 1540s after the break (2000s from t = 0). In chapter (c) the three cases that are closer to the
core recovery are presented together (LPIS_1120, LPIS_1121 and LPIS_1125) because they are a little
different from the rest after the LPIS injection time.
Chapter (d) presents cases in which LPIS could avoid core degradation and melting.
Figure 7.65 shows that if LPIS injection time is lower than 659s after the break (case LPIS_1119) core
temperature increase can be mitigated.
WinGraf - 05-20-2003
3500
XXX
YYY
ZZZ
VVV
JJJ
HHH
###
OOO
AAA
BBB
CCC
DDD
EEE
FFF
3000
Temperature (K)
2500
2000
1500
LPIS_2000
LPIS_1500
LPIS_1300
LPIS_1200
LPIS_1150
LPIS_1125
LPIS_1121
LPIS_1120
LPIS_1119
LPIS_1115
LPIS_1110
LPIS_1100
LPIS_900
DEBRIS_NOBK
1000
X Y Z
X V
Y JZ
X H
V
Y#
JZ
X O
H
V
Y A
JZ
#
X
INITIATION OF CORE DEGRADATION
bgmct0
bgmct0
bgmct0
bgmct0
bgmct0
bgmct0
bgmct0
bgmct0
bgmct0
bgmct0
bgmct0
bgmct0
bgmct0
bgmct0
Y
Y JZ V
Y JZ V
V
Z
O
JZ V
Y J#
O#
J V JZ V JZ
X F
O
X F
# V
F #
O
Y #
X O
F
X
F
X
Z
J
F
Y
X
F
Y
H
X
F
Y
H
X O
Z
V
O
F
V
Y
H
JX
C
Z
#
A
E F
B
V
H
O
D
Y
E
JZ
X
#
A
C
B
H
O
V
YA
J
X
C
#
Z
F
B
V
H
O
D
Y
E
E
JZ
X
C
#
H
O
B
V
D
Y E
F
V
H
O
B
Y A
JZ
XD
#
A
JZ
#
A
C
X B
F
H
O
V
D
Y C
500
F
B
V
H
O
D
Y
LPIS action
Break time 460s
JZ
X
#
A
C
H
H
H
D
B C
#
A
H
#
D
B C
E
E
O
H
O
# H
CORE RECOVERY
A D
E
B C
E D
A
BE
A D
C
B E
A D
C
B E
A D
C
B E
A D
C
BE
A D
C
B
ACCUM action
0
0
200.
400.
600.
800.
1000.
Time (s)
1200.
1400.
1600.
1800.
2000.
Figure 7.65 - Core maximum temperature in all SCDAP calculations
102
Assessment of Size Aspects in Modelling Molten Fuel Coolant Interaction
Chapter 7 - RESEARCH ACTIVITIES SUMMARY
a)
“Base Case” (DEBRIS_NOBK, DEBRIS_BK cases). MFCI with RELAP5/SCDAP 3.2
Two “Base Case” calculations were performed with the code. The chosen scenario is a LOCA without LPIS
leading the vessel to a progressive voiding and the core to a complete uncover. Core melting and molten pool
formation that pours into the lower plenum will be also observed.
The first case DEBRIS_NOBK is a calculation performed with the input option in the code (w4 = 1 on card
51010000) that predicts the minimum heat transfer from the slumped material to the water in the lower
plenum. This option is the most conservative for a study about lower head failure because the melt contains
the maximum amount of energy when it interacts with the vessel.
The second case DEBRIS_BK is a calculation performed with the input option in the code (w4 = 0 on card
51010000) that predicts that maximum heat transfer from the debris (as it slumps) to the water is caused
when the debris breaks up in little particles. All the stored energy from the melt is added to the water over a
time interval of about 10 seconds. Linked to this option two parameters can be settled in the core input,
debris porosity and particle size. By default the debris material is considered to have a porosity of 0.5 and a
particle size of 10 mm and these are the values used to perform the calculation.
After LOCA break (at 460s) some parameters are plotted in both cases. Obviously both calculations are
identical up to the moment of slumping. LOCA break mass flow is presented in Figure 7.69. It reaches a
peak of about 3200 kg/s 15 seconds after the break. In Figure 7.66 pressure in the primary system is shown.
It decreases in few seconds from 16 MPa up to stabilization at about 0.2 MPa. Figure 7.67 shows water
temperature in the primary system after break. In Figure 7.68 downcomer water level is presented, it can be
also observed the initial increase of water level due to accumulators injection.
At the time of the slumping, the higher interaction between molten material and water in the lower plenum
and higher energy transferred to it causes a large amount of vapour generation. This may increase pressure
and temperature significantly in the reactor vessel. Figure 7.70 and Figure 7.71 show lower plenum pressure
in both cases. The higher pressure peak during slumping is obtained in case DEBRIS_BK (from 0.25 to 0.9
MPa) where molten material transfers instantly all the energy to the water. The calculation is performed in an
open system (break) so the pressure increase is small (Figure 7.71).
Void fraction (Figure 7.72) in case DEBRIS_BK reaches the value 1 in few seconds, almost instantly lower
plenum is voided due to the heat transfer from the fragmented little particles to the water, in case
DEBRIS_NOBK lower plenum is voided progressively due to the debris heat transfer. The first slumping of
control material (some kg of Ag) before the real so-called molten pool slumping is shown.
Figure 7.73 shows lower plenum temperature, higher during slumping for the DEBRIS_BK case.
From Figure 7.74 and Figure 7.75 case DEBRIS_NOBK shows that power and energy still stored in the
slumped debris are transferred by convection from top surface and to fluid at boundaries of debris and
structural material. In DEBRIS_BK both values are small because most molten pool energy was transferred
form the fragmented particles to the water and less debris mass is expected to be in the lower plenum bottom.
The two calculations have shown the MFCI study possibilities with RELAP5/SCDAP 3.2. However in
general it is more interesting to study the grade of interaction (a point between total interaction, total
fragmentation of the slumping material instead of no interaction or no fragmentation) as a result from the
initial conditions in the time just before the slumping of the molten pool and the evolution of the interaction
transient instead of to impose this interaction grade as an input parameter. More accuracy can be achieved
using detailed MFCI codes but on the other hand these codes do not have the capacities to run full detailed
NPP transients as RELAP5/SCDAP 3.2 has. So the combined use of two codes could be a good solution to
study MFCI phenomena in a full NPP transient.
103
Assessment of Size Aspects in Modelling Molten Fuel Coolant Interaction
Chapter 7 - RESEARCH ACTIVITIES SUMMARY
WinGraf - 05-27-2003
22.5
20.0
17.5
X
Y
X
Y
X
Y
X
Y
X
Y
X
X
Y
X
Y
X
Y
X
Y
X
Y
Pressure (MPa)
15.0
12.5
XXX
YYY
DEBRIS_BK
DEBRIS_NOBK
p220010000
p220010000
10.0
7.5
X
Y
5.0
2.5
X
Y
0
0
100.0
200.0
300.0
400.0
X
Y
500.0
Y
X
X
Y
600.0
X
Y
X
Y
700.0
Time (s)
Figure 7.66 - Pressure of primary system after break
WinGraf - 05-27-2003
400
350
300
Temperature (°C)
X
X
Y
X
Y
X
Y
X
Y
X
Y
X
Y
X
Y
X
Y
X
Y
X
Y
X
Y
250
XXX
YYY
DEBRIS_BK
DEBRIS_NOBK
tempf220010000
tempf220010000
200
150
X
Y
X
Y
X
Y
X
Y
X
Y
100
X
Y
50
0
0
100.0
200.0
300.0
400.0
Time (s)
500.0
600.0
Figure 7.67 - Temperature in primary system after break
104
700.0
Assessment of Size Aspects in Modelling Molten Fuel Coolant Interaction
Chapter 7 - RESEARCH ACTIVITIES SUMMARY
WinGraf - 05-27-2003
4.50
4.00
X
Y
X
XY X Y X YX
Y
X
XY X Y X YX
Y
X
3.50
Level (m)
3.00
XXX
YYY
2.50
DEBRIS_BK
DEBRIS_NOBK
cntrlvar130
cntrlvar130
X
Y
2.00
X
Y
X
1.50
Y
X
XY
1.00
X
X
Y
Y
X YX
Y
X
.50
0
0
100.0
200.0
300.0
400.0
500.0
600.0
700.0
800.0
900.0
Time (s)
Figure 7.68 - Downcomer level after the break
WinGraf - 05-27-2003
4000
3500
XXX
YYY
DEBRIS_BK
DEBRIS_NOBK
mflowj901000000
mflowj901000000
Mass flow (kg/s)
3000
2500
2000
X
Y
1500
1000
X
Y
500
X
Y
0
300.0
X
Y
350.0
400.0
450.0
500.0
Time (s)
X
Y
X
Y
550.0
X
Y
X
Y
600.0
X
Y
X
Y
X
Y
650.0
X
Y
700.0
Figure 7.69 - Break mass flow
105
Assessment of Size Aspects in Modelling Molten Fuel Coolant Interaction
Chapter 7 - RESEARCH ACTIVITIES SUMMARY
WinGraf - 04-12-2003
1.20
1.00
XXX
YYY
Pressure (MPa)
.80
DEBRIS_BK
DEBRIS_NOBK
p106010000
p106010000
slumping
.60
X
.40
Y
Y
X
.20
0
3500.0
Y
X YX YX YX YX YX YX YX YX
3600.0
3700.0
3800.0
Y
X
X
X YX YX YX YX YX
3900.0
4000.0
4100.0
Time (s)
Figure 7.70 - Pressure in the lower plenum at the slumping time in MFCI SCDAP calculations
WinGraf - 04-12-2003
20.0
18.0
16.0
XY
XXX
YYY
XY
DEBRIS_BK
DEBRIS_NOBK
p106010000
p106010000
Pressure (MPa)
14.0
12.0
10.0
8.0
6.0
slumping
4.0
2.0
XY
0
0
500.
XY XY XY XY XY XY XY XY XY XY XY XY XY XY XY XY XY
1000.
1500.
2000.
2500.
3000.
3500.
4000.
4500.
5000.
Time (s)
Figure 7.71 - Pressure in the lower plenum in MFCI SCDAP calculations
106
Assessment of Size Aspects in Modelling Molten Fuel Coolant Interaction
Chapter 7 - RESEARCH ACTIVITIES SUMMARY
WinGraf - 06-02-2003
1.00
X
Y
X
.90
.80
XXX DEBRIS_BK
YYY DEBRIS_NOBK
voidg106010000
voidg106010000
slumping
Void Fraction (-)
.70
.60
Y
.50
Control rod slumping
Y
.40
.30
.20
Y
.10
0 XY XY XY
1000.
1500.
X
2500.
2000.
Y
X
3000.
Y Y
3500.
4000.
4500.
5000.
5500.
Time (s)
Figure 7.72 - Void fraction in the lower plenum at the slumping time in MFCI SCDAP calculations
WinGraf - 06-02-2003
260
240
XXX
YYY
220
DEBRIS_BK
DEBRIS_NOBK
tempf106010000
tempf106010000
Temperature (°C)
slumping
200
180
160
140
Y
X
YX
YX
Y
X
120
100
3200.
3400.
Y
X
Y
X
Y
X
3600.
Y
X
Y
X
Y
X
Y
X
3800.
X
Y
X
Y
X
4000.
Y
X
Y
X
Y
X
4200.
Y
X
Y
X
Y
X
4400.
Time (s)
Figure 7.73 - Water temperature in the lower plenum at the slumping time in MFCI SCDAP calculations
107
Assessment of Size Aspects in Modelling Molten Fuel Coolant Interaction
Chapter 7 - RESEARCH ACTIVITIES SUMMARY
x 10 7
4.5
WinGraf - 04-13-2003
Tot rate of htransf from top surf debris (W)
4
XXX
YYY
3.5
3
DEBRIS_BK
DEBRIS_NOBK
debqup1
debqup1
Y
Y
2.5
Y
2
Y
Y
1.5
Y
1
Y
.5
Y
0
3400.
Y
Y
3600.
Y
Y
3800.
X
4000.
X
X
X
X
X
X
Y
Y
4200.
4400.
4600.
YX YX YX YX YX
X
4800.
5000.
5200.
Time (s)
Figure 7.74 - Total rate of heat transfer by convection from top surface of debris (W) in MFCI SCDAP
calculations
x 10 10
3.5
WinGraf - 04-13-2003
Integral with of total power in debris (J)
3
XXX
YYY
DEBRIS_BK
DEBRIS_NOBK
intq1
intq1
Y
2.5
2
Y
Y
Y
Y
Y
Y
Y
Y
Y
1.5
X
X
X
1
X
X
X
X
Y
.5
X
0
0
1000.
2000.
Y
3000.
X
X
Y
4000.
5000.
Time (s)
6000.
7000.
8000.
9000.
Figure 7.75 - Integral with respect to time of total transfer heat from debris and structural material to fluid at
boundaries of debris and structural material (J) in MFCI SCDAP calculations
108
Assessment of Size Aspects in Modelling Molten Fuel Coolant Interaction
Chapter 7 - RESEARCH ACTIVITIES SUMMARY
b)
Core degradation, molten pool and slumping occurring. Cases with LPIS injection time =
1150s, 1200s, 1300s, 1500s and 2000s
Some sensibility cases are shown in this chapter in order to study the impact of LPIS. As it will be seen LPIS
does not avoid (after core degradation) molten pool formation and subsequent slumping to the lower plenum.
Cases are clearly far from core recovery: from case LPIS_1150, injection time 690s after the break (1150s
from t = 0) up to LPIS_2000, injection time 1540s after the break (2000s from t = 0).
Base Case DEBRIS_NOBK (without LPIS) is shown in all plots together with the rest in order to verify if a
late LPIS injection has similar consequences to a case without LPIS.
Bgmct0 (Figure 7.76) represents maximum temperature in the 5 fuel core components at all axial nodes, so it
is an indicator of the hottest point in the core. From the figure it is shown that maximum temperature grows
higher in LPIS cases compared to case without LPIS (DEBRIS_NOBK) because LPIS water powers the
cladding oxidation producing H2 at temperatures greater than about 1500K. This oxidation reaction is
exothermic producing a lot of energy (586 KJ/mol) that makes maximum core temperature to reach 3000K in
few seconds. Temperature decreases when molten pool slumps to lower head.
H2 production rate (Figure 7.77) is more important in cases in which LIPS water arrives to the core when it
has arrived to the threshold temperature for oxidation, about 1500K. Case without LPIS (DEBRIS_NOBK)
produces much less H2 at the beginning because the reaction it is not powered by additional LPIS water. It is
also interesting to observe the integral of this quantity, which gives the H2 accumulated (Figure 7.78).
In LPIS cases molten pool in the core (Figure 7.79, molten pool radius) is formed at about 1600s-1700s
except for LPIS at 2000s that is similar to case without LPIS (DEBRIS_NOBK). It seems that LPIS delays
molten pool slumping, except for LPIS at 2000s and LPIS at 1300s.
From Figure 7.80 to Figure 7.83 component masses of molten pool that slumped to lower plenum are
represented, UO2, ZrO2, Zr and Ag. Although it is difficult to calculate exact masses and exact times of
slumping, the figures purpose is to give an idea of the principal amounts that form the molten pool. These are
principally UO2 (about 60000-70000 kg), Zr (less than 5000 kg) and ZrO2 (about 9000 kg). Ag mass (about
2000 kg) represents control material slumping the first. It is deduced that the rest of components (UO2, ZrO2,
Zr) fall down later in the so-called slumping of molten pool and that once it occurs RELAP5/SCDAP 3.2
considers that almost all molten pool mass slumps. Almost all mass slumped to the lower plenum are present
in the liquid phase (Figure 7.84).
From Figure 7.85 to Figure 7.90 the mass amounts are represented case by case. LPIS at 1300s is the case
with less mass involved, molten pool is smaller and smaller slumping mass falls, ZrO2 and Ag masses are
similar to the rest. Zr mass is higher than ZrO2 in all cases, different from what was experimented in the
FARO facility, where molten pool was generally simulated with more ZrO2 % (in mass) than Zr %. The first
slump occurs when control material (Ag) falls into the lower plenum.
Figure 7.91 and Figure 7.92 show downcomer water level, which is something like water level in the core
region. In the first part of the transient, after the break (460 s) downcomer level decreases quickly in all
cases. At about 500s accumulators start to inject water recovering the water leak for some seconds, but it is
the effect of LPIS injecting water in every case that recovers core cooling and downcomer level reaches its
top. Obviously in case without LPIS (DEBRIS_NOBK) level continues to decrease. Figure 7.92 shows the
consequence of the slumping. In case without LPIS (DEBRIS_NOBK) level increases because residual water
in the lower plenum is pressed towards the downcomer. In the rest of cases level decreases because voiding
produced in the lower plenum is transferred to downcomer.
Figure 7.93 shows void fraction in the lower plenum. Lower plenum is voided in case without LPIS and void
fraction reaches 1.0 in few seconds. In the rest of cases the lower plenum is continuously filled up with LPIS
water so void fraction oscillates.
109
Assessment of Size Aspects in Modelling Molten Fuel Coolant Interaction
Chapter 7 - RESEARCH ACTIVITIES SUMMARY
Figure 7.94 and Figure 7.95 show pressure and water temperature in the lower plenum. Case without LPIS is
almost in saturation conditions. The rest of cases are in subcooled conditions due to the LPIS that injects
subcooled water into the vessel. When slumping occurs all cases reach saturation temperature in the lower
plenum. Pressure at the moment of the slumping reaches (from 2.5 bar) about 4-5 bar in all cases.
WinGraf - 05-20-2003
3500
jump due to cladding oxidation
H H H H X
X
Y JY
YV Y
Y V HY JX Z XYZ YZ XY JHY
J YZVJ YZVJX
H Z J ZV Y VJ VJX VJ ZV ZVJ Z J ZVJ ZV ZVJ
H
X
ZVJ YZV H
X
3000
H
X
J
2500
Temperature (K)
Y
H
X
H
H
2000
H
XXX
YYY
ZZZ
VVV
JJJ
HHH
V
1500
Z
Y
H
X
J
1000
V
JH
X
Z
Y JH
XY
ZV
XYZVJH
XYZV
500
LPIS_2000
LPIS_1500
LPIS_1300
LPIS_1200
LPIS_1150
DEBRIS_NOBK
bgmct0
bgmct0
bgmct0
bgmct0
bgmct0
bgmct0
H
X
X
X
X
slumping
0
0
500.
1000.
1500.
2000.
2500.
Time (s)
3000.
3500.
4000.
4500.
5000.
Figure 7.76 - Maximum core temperature in LPIS SCDAP calculations
WinGraf - 05-20-2003
4.00
3.50
XXX
YYY
ZZZ
VVV
JJJ
HHH
H2 production (kg/s)
3.00
2.50
LPIS_2000
LPIS_1500
LPIS_1300
LPIS_1200
LPIS_1150
DEBRIS_NOBK
bgth0
bgth0
bgth0
bgth0
bgth0
bgth0
2.00
Y
1.50
1.00
.50
0
0
500.
J Z
Y
H X
VX
Z
HY X
H X
H ZV H H
HYZVJHYZVJH ZVJHYZVJHYZVJH ZVJ
HYZVJX
HYZ
YZVJX
HY VJX
HYZVJX
JX
V
YZVJ Y JXYZVJX
1000.
1500.
2000.
2500.
3000.
3500.
4000.
4500.
5000.
Time (s)
Figure 7.77 - H2 production rate in LPIS SCDAP calculations
110
Assessment of Size Aspects in Modelling Molten Fuel Coolant Interaction
Chapter 7 - RESEARCH ACTIVITIES SUMMARY
WinGraf - 04-13-2003
300
J
J
V
V
250
H2 produced (kg)
200
V
V
V
V
J
J
J
J
J
J
J
J
J
V
V
V
V
V
V
V
J
J
VH VH
H
H
H
H
H
H
H
H
H
150
Z
Z
Z
Z
Z
H
H
Z
Z
Z
JH
100
Z HZ
J Y
Y
Y
Z
Z
Y
Y
Z
Z YZ Y
X
XY X
X
X
XXX
YYY
ZZZ
VVV
JJJ
HHH
X
50
HX
YZ J
X V
H
YZVJ
1000.
0
0
X
2000.
3000.
XY
XY
XY
XY
LPIS_1150
LPIS_1200
LPIS_1300
LPIS_1500
LPIS_2000
DEBRIS_NOBK
4000.
XY
XY
XY
H2_ACCUM
H2_ACCUM
H2_ACCUM
H2_ACCUM
H2_ACCUM
H2_ACCUM
5000.
6000.
Time (s)
Figure 7.78 - H2 accumulated in LPIS SCDAP calculations
WinGraf - 05-20-2003
2.50
XXX
YYY
ZZZ
VVV
JJJ
HHH
2.25
2.00
LPIS_2000
LPIS_1500
LPIS_1300
LPIS_1200
LPIS_1150
DEBRIS_NOBK
repool0
repool0
repool0
repool0
repool0
repool0
1.75
Radius (m)
X
1.50
XH
SLUMPING
X
H
XH
Y
Y
Y
1.25
Y
J
Y
JZ J V
Y
V
ZV
J
V
Y ZJ Z
ZV
J
Z
X
1.00
.75
Y
J
V
Y
JV
Y
V
Y
Y
Z
Z
Z
Y
Y
JV J V
Z
VJ
J
V
Z
Y
JV
Z
Z
Z
H
.50
.25
0
0
1000.
2000.
3000.
4000.
5000.
Time (s)
6000.
7000.
8000.
9000.
Figure 7.79 - Radius of molten pool in LPIS SCDAP calculations
111
Assessment of Size Aspects in Modelling Molten Fuel Coolant Interaction
Chapter 7 - RESEARCH ACTIVITIES SUMMARY
x 10 4
8
WinGraf - 05-20-2003
H
7
H X HX XH X HX XH X H
X X
X XH
H
J
H
H
J
H
Y
JV
V
H
Y
V
H
Y
H
Y
V
UO2 mass (kg)
6
H XXX
YYY
ZZZ
VVV
JJJ
HHH
5
4
LPIS_2000
LPIS_1500
LPIS_1300
LPIS_1200
LPIS_1150
DEBRIS_NOBK
masuo21
masuo21
masuo21
masuo21
masuo21
masuo21
3
2
Z XZ Z Z Z Z Z Z Z Z Z Z Z Z
1
0
2000.
3000.
4000.
5000.
6000.
7000.
8000.
9000.
10000.
Time (s)
Figure 7.80 - Total UO2 mass slumped in lower plenum in LPIS SCDAP calculations
WinGraf - 05-20-2003
7000
6000
XXX
YYY
ZZZ
VVV
JJJ
HHH
LPIS_2000
LPIS_1500
LPIS_1300
LPIS_1200
LPIS_1150
DEBRIS_NOBK
maszo21
maszo21
maszo21
maszo21
maszo21
maszo21
5000
ZrO2 mass (kg)
X X X X X X X X X X X
4000
Y
H
3000
H
H
H
H
H
H
Y
H
H
H
H
V
J
V
V
J
Y
H
Y
H
H
H
2000
Z Z Z Z Z Z Z Z Z Z Z Z Z Z
J
X
1000
0
2000.
H
3000.
4000.
5000.
6000.
Time (s)
7000.
8000.
V
9000.
10000.
Figure 7.81 - Total ZrO2 mass slumped in lower plenum in LPIS SCDAP calculations
112
Assessment of Size Aspects in Modelling Molten Fuel Coolant Interaction
Chapter 7 - RESEARCH ACTIVITIES SUMMARY
WinGraf - 05-20-2003
10000
H
9000
H
H
H
H
H
HJ
H
J
H
H
J
H
V
V
H
Y
H
Y
H
Y
H
Y
X X X X X X X X X X X
8000
V
V
H
Zr mass (kg)
7000
XXX
YYY
ZZZ
VVV
JJJ
HHH
6000
5000
LPIS_2000
LPIS_1500
LPIS_1300
LPIS_1200
LPIS_1150
DEBRIS_NOBK
masszr1
masszr1
masszr1
masszr1
masszr1
masszr1
4000
3000
Z XZ Z Z Z Z Z Z Z Z Z Z Z Z
2000
1000
0
2000.
3000.
4000.
5000.
6000.
7000.
8000.
9000.
10000.
Time (s)
Figure 7.82 - Total Zr mass slumped in lower plenum in LPIS SCDAP calculations
WinGraf - 05-20-2003
3000
2500
H
H
Ag mass (kg)
2000
V
H
H
H
H
H
H
H
H
H
H
H H H HY H
Y V Y V YV Y
X X X X X ZX ZYX ZV Y VJY JV Y JV
V XV V V V V J J
ZY
X
Y
X
J
Z
J
1500
VX XY
Y J Y J YJ
J
1000
J
XXX
YYY
ZZZ
VVV
JJJ
HHH
LPIS_2000
LPIS_1500
LPIS_1300
LPIS_1200
LPIS_1150
DEBRIS_NOBK
massag1
massag1
massag1
massag1
massag1
massag1
500
0
2000.
3000.
4000.
5000.
6000.
Time (s)
7000.
8000.
9000.
10000.
Figure 7.83 - Total Ag mass slumped in lower plenum in LPIS SCDAP calculations
113
Assessment of Size Aspects in Modelling Molten Fuel Coolant Interaction
Chapter 7 - RESEARCH ACTIVITIES SUMMARY
x 10 4
9
WinGraf - 05-20-2003
H
H
H
8
H
H
H
X
7
H
J
H
J
V
V
H H H
X X X X X X X X X X X X X X
5
H
H
H
H
Y
J
6
Liquid mass (kg)
H
HY H
V
V YV Y
Y
J
XXX
YYY
ZZZ
VVV
JJJ
HHH
4
3
LPIS_2000
LPIS_1500
LPIS_1300
LPIS_1200
LPIS_1150
DEBRIS_NOBK
masliq1
masliq1
masliq1
masliq1
masliq1
masliq1
2
Z Z Z Z Z Z Z Z Z Z Z Z Z Z Z Z Z
1
0
3000.
Z
H
V
4000.
Y
J
5000.
J
6000.
7000.
8000.
9000.
10000.
Time (s)
Figure 7.84 - Total Liquid mass in lower plenum in LPIS SCDAP calculations
x 10 4
8
WinGraf - 04-14-2003
7
XXX
YYY
ZZZ
VVV
JJJ
6
Mass (kg)
5
LPIS_1150
LPIS_1150
LPIS_1150
LPIS_1150
LPIS_1150
J
V
J
V
massag1
masszr1
maszo21
masuo21
masliq1
VJ
V
4
3
2
1
Y
0
3000.
X
3500.
4000.
X
X JX
4500.
5000.
X
X
5500.
X
X
6000.
X
X
6500.
X
Y
Y
X
XZ XZ XZ
J Z
7000.
7500.
8000.
Time (s)
Figure 7.85 - Mass distribution in lower plenum in case LPIS_1150
114
Assessment of Size Aspects in Modelling Molten Fuel Coolant Interaction
Chapter 7 - RESEARCH ACTIVITIES SUMMARY
x 10 4
7
WinGraf - 04-14-2003
J
V
6
XXX
YYY
ZZZ
VVV
JJJ
Mass (kg)
5
LPIS_1200
LPIS_1200
LPIS_1200
LPIS_1200
LPIS_1200
V
V
V
V
J
J
J
J
massag1
masszr1
maszo21
masuo21
masliq1
4
3
2
1
Y
0
3000.
X
X
J
4000.
X
X
X
X
X
X
X
J
6000.
Time (s)
5000.
X
X
Y
Y
Y
Y
XZ XZ XZ XZ XZ
X
7000.
8000.
9000.
Figure 7.86 - Mass distribution in lower plenum in case LPIS_1200
x 10 4
1.8
WinGraf - 04-14-2003
V
V
V
V
V
V
V
V
V
V
V
V
V
V
V
V
V
1.6
V
1.4
J
J
J
J
J
J
J
J
J
J
J
J
J
J
J
J
J
J
Mass (kg)
1.2
1
XXX
YYY
ZZZ
VVV
JJJ
.8
LPIS_1300
LPIS_1300
LPIS_1300
LPIS_1300
LPIS_1300
massag1
masszr1
maszo21
masuo21
masliq1
.6
.4
YZ YZ YZ YZ YZ YZ YZ YZ YZ YZ YZ YZ XYZ XYZ XYZ XYZ XYZ
X
.2
0
3000.
Z
3500.
4000.
4500.
5000.
5500.
6000.
6500.
7000.
Time (s)
Figure 7.87 - Mass distribution in lower plenum in case LPIS_1300
115
Assessment of Size Aspects in Modelling Molten Fuel Coolant Interaction
Chapter 7 - RESEARCH ACTIVITIES SUMMARY
x 10 4
8
WinGraf - 04-14-2003
J
7
XXX
YYY
ZZZ
VVV
JJJ
6
LPIS_1500
LPIS_1500
LPIS_1500
LPIS_1500
LPIS_1500
VJ V
massag1
masszr1
maszo21
masuo21
masliq1
V
J
V
V
V
V
V
V
V
V
J
J
J
J
J
J
J
J
J
Mass (kg)
5
4
V
3
2
1
Y
X
0
7000.
X
X
X
X
X
7500.
X
X
X
8000.
Y
Z
X
Y
Z
X
Y
Z
X
Y
Z
X
8500.
Time (s)
Y
Z
X
Y
Z
X
Y
Z
X
9000.
Y
Z
X
Y
Z
X
Y
Z
X
Y
Z
9500.
X
10000.
Figure 7.88 - Mass distribution in lower plenum in case LPIS_1500
x 10 4
8
WinGraf - 04-14-2003
J
7
V
V
V
V
V
V
V
V
V
V
V
V
V
V
V
6
J
J
J
J
J
J
J
J
J
J
J
J
J
J
J
Mass (kg)
5
4
XXX
YYY
ZZZ
VVV
JJJ
3
LPIS_2000
LPIS_2000
LPIS_2000
LPIS_2000
LPIS_2000
massag1
masszr1
maszo21
masuo21
masliq1
2
V
1
0
3000.
3500.
Y
Y
Y
Y
Y
Y
Y
Y
Y
Y
Y
Y
Y
Y
Y
Z
Z
Z
Z
Z
Z
Z
Z
Z
Z
Z
Z
Z
Z
Z
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
4000.
4500.
5000.
5500.
6000.
6500.
Time (s)
7000.
Figure 7.89 - Mass distribution in lower plenum in case LPIS_2000
116
Assessment of Size Aspects in Modelling Molten Fuel Coolant Interaction
Chapter 7 - RESEARCH ACTIVITIES SUMMARY
x 10 4
10
WinGraf - 04-14-2003
9
J
8
J
J
J
7
Mass (kg)
J
J
6
V
V
J
J
VJ
V
J
V
J
V
J
V
V
V
V
V
XXX
YYY
ZZZ
VVV
JJJ
5
V
J
V
J
J
J
J
J
V
V
V
V
V
DEBRIS_NOBK
DEBRIS_NOBK
DEBRIS_NOBK
DEBRIS_NOBK
DEBRIS_NOBK
massag1
masszr1
maszo21
masuo21
masliq1
4
3
2
1
Y
0
3000.
Y
Y
Y
Y
Y
Y
Y
Y
Y
Y
Y
Y
Y
Y
Y
Y
X J XZ XZ XZ XZ XZ XZ XZ XZ XZ XZ XZ XZ XZ XZ XZ XZ XZ XZ
4000.
5000.
6000.
7000.
8000.
9000.
10000.
Time (s)
Figure 7.90 - Mass distribution in lower plenum in case DEBRIS_NOBK
WinGraf - 05-20-2003
8.00
XXX
YYY
ZZZ
VVV
JJJ
HHH
7.00
Downcomer level (m)
6.00
LPIS_2000
LPIS_1500
LPIS_1300
LPIS_1200
LPIS_1150
DEBRIS_NOBK
cntrlvar130
cntrlvar130
cntrlvar130
cntrlvar130
cntrlvar130
cntrlvar130
LPIS
5.00
4.00
XYZVXJYHZVXJY
VJ Z JYZVJYZVJY VXJYZVXJYZVXJYZVXJYZVXJYZVXJYZVXJYZVJ
Z
V
J
V
3.00
2.00
H
XJ
Y
HZ
Z
V
VX
J
YH
Z
V
Y
ACCUM XJYHZV
XJYHZ X
Z
YH X H X
H XH XH
1.00
0
0
500.
1000.
1500.
2000.
Time (s)
H
H
H
H
2500.
3000.
H
3500.
Figure 7.91 - Downcomer level before slumping in LPIS SCDAP calculations
117
Assessment of Size Aspects in Modelling Molten Fuel Coolant Interaction
Chapter 7 - RESEARCH ACTIVITIES SUMMARY
WinGraf - 11-09-2003
7.00
XXX
YYY
ZZZ
VVV
JJJ
HHH
Downcomer Level (m)
6.00
5.00
LPIS_2000
LPIS_1500
LPIS_1300
LPIS_1200
LPIS_1150
DEBRIS_NOBK
cntrlvar130
cntrlvar130
cntrlvar130
cntrlvar130
cntrlvar130
cntrlvar130
SLUMPING
4.00
YV
VJXZ ZV
Y
YZ
YJ ZVJYZX JZ
VY ZJ VZYJ ZV Y
JVZ XZ
Z JXZY
J X
ZJV
X JZ
XVY Z
JX VX
J VY VY
YZVJXZYX
JYVXZJ X
Z
X
YV
YV
V
3.00
X
X
2.00
X
H
H
1.00
X
X
J
Y
J
V
Y
Y
SLUMPING
H
0
2000.
J
H
H
3000.
4000.
5000.
6000.
7000.
8000.
9000.
10000.
Time (s)
Figure 7.92 - Downcomer level after slumping in LPIS SCDAP calculations
WinGraf - 05-20-2003
1.00
XXX
YYY
ZZZ
VVV
JJJ
Z
HHH
.90
.80
LPIS_2000
LPIS_1500
LPIS_1300
LPIS_1200
LPIS_1150
DEBRIS_NOBK
voidg106010000
voidg106010000
voidg106010000
voidg106010000
voidg106010000
voidg106010000
W/O LPIS
SLUMPING
.70
Void Fraction (-)
J
J
.60
X
.50
Y
V
Y
X
H
X
.40
.30
X
Y
V
.20
X
V
.10
0
0
X X
X
Z
Y
Z JV Y
X
ZY H VYZXJ VZ J ZVYJZ VZ
YV
XH
Z JX
ZHX
JZYV ZJ Y
V
ZJ X
ZV
YJ Z YXH JX
JY ZV
JY ZJVY
1000.
2000.
3000.
4000.
5000.
6000.
7000.
Time (s)
Y
Y
8000.
V
9000.
10000.
Figure 7.93 - Void fraction in the lower plenum in LPIS SCDAP calculations
118
Assessment of Size Aspects in Modelling Molten Fuel Coolant Interaction
Chapter 7 - RESEARCH ACTIVITIES SUMMARY
WinGraf - 05-20-2003
.70
XXX
YYY
ZZZ
VVV
JJJ
HHH
.60
LPIS_2000
LPIS_1500
LPIS_1300
LPIS_1200
LPIS_1150
DEBRIS_NOBK
p106010000
p106010000
p106010000
p106010000
p106010000
p106010000
Pressure (MPa)
.50
Y
Z
.40
Z
Y
J VJ
V Z YZ X J
ZY
ZV Z
Y
Z X X
J
Y
X Y XV
JY ZV
JXY ZJVY
JX JV
XJ ZXVYJZX VZ
Z VY V Y V Y V Y
Y JV ZXJ VZY
X
V
H
Z
J J
X
X X
JY
XV
ZHX H
V Y
X
H H H H H
H H H H H H H H H H H
JV
.30
J
.20
.10
0
1000.
2000.
3000.
4000.
5000.
Time (s)
6000.
7000.
8000.
9000.
10000.
Figure 7.94 - Pressure in the lower plenum in LPIS SCDAP calculations
WinGraf - 05-20-2003
160.0
XXX
YYY
ZZZ
VVV
JJJ
HHH
tempf106010000
tempf106010000
tempf106010000
tempf106010000
tempf106010000
tempf106010000
X
Y
V
J
Z
X X
Z
J J
V
X X
Y
Z YZ V
ZH H H
Y
ZHX H
XV
X
H
JY
J JHZ J
X H H XH X H H
Z
H
Y
XH
Y
JZV
H HV H H H
X
Y
Z
V
V
V
J
Z
Z Z
J
Y
Z
Z
X
Z
Y
X
Z
Y
X
Y
V
Z
VY
X
VY
X
J
Y
Y
V
J
V
VY
V
V J
slumping
J Y VYJ J J
150.0
140.0
Temp (°C)
130.0
120.0
110.0
100.0
90.0
80.0
LPIS_2000
LPIS_1500
LPIS_1300
LPIS_1200
LPIS_1150
DEBRIS_NOBK
VJ
70.0
60.0
0
1000.
2000.
3000.
4000.
5000.
6000.
7000.
8000.
9000.
10000.
Time (s)
Figure 7.95 - Water temperature in the lower plenum in LPIS SCDAP calculations
119
Assessment of Size Aspects in Modelling Molten Fuel Coolant Interaction
Chapter 7 - RESEARCH ACTIVITIES SUMMARY
c)
Core degradation, molten pool and slumping occurring. Cases with LPIS injection time =
1120s, 1121s and 1125s
These three cases are closer to the core recovery. They are presented together (LPIS_1120, LPIS_1121 and
LPIS_1125) in order to discuss the transient evolution after the LPIS injection time. Base Case
DEBRIS_NOBK (without LPIS) is shown in all plots together with the rest.
Maximum core temperature (Figure 7.96) increases due to H2 production at temperatures greater than about
1500K, powered by LPIS water. Although LPIS reaches to stop temperature increase at the beginning, the
core recovery does not take place. The second increase (in cases LPIS_1120, LPIS_1121) at about 1600s is
due to node covering by molten pool. Later on core degradation continues up.
Since bgmct0 quantity represents maximum temperature in the 5 fuel core components at all axial nodes, in
case LPIS_1120 the exact maximum temperature (Figure 7.97) corresponds to axial position 9 fuel
components 5 and 7, that are in the hydrodynamic channels 112 and 115 respectively (Figure 7.64), both in
the radial external point. Temperature increase at about 1200s in component 7, axial position 9, is due to H2
local production, as it can be observed from Figure 7.98. Figure 7.99 shows damage levels in these
components, the damage starts when ballooning of cladding geometry, second step is fragmentation and they
finally become molten pool.
In the same way for case LPIS_1121 maximum temperature (Figure 7.100) corresponds to radial external
point fuel component 7, axial position 9 and fuel component 9, axial position 8. Those are included in the
hydrodynamic channels 115 and 113 respectively. Figure 7.101 shows damage levels in these components,
high temperature increase up to about 3000K in these components corresponds when they become molten
pool.
H2 production rate (Figure 7.102) as previous cases shows high peak production due to LPIS water injection.
Radius of molten pool (Figure 7.103) shows that slumping is delayed in cases LPIS_1120 and LPIS_1125 up
to about 10000s compared to case without LPIS, which occurs at about 3900s as in LPIS_1121 case.
From Figure 7.104 to Figure 7.108 composition masses of molten pool slumped to the lower plenum are
presented in all cases. LPIS_1121 case shows much less mass in the slumped molten pool than the rest.
Figure 7.109 shows downcomer water level, after the break, downcomer level decreases quickly in all cases.
It is shown the effect of LPIS injecting water in every case that recovers core cooling and downcomer level
reaches its top. After slumping level decreases in LPIS cases because voiding produced in the lower plenum
is transferred to downcomer.
Figure 7.110 shows void fraction in the lower plenum. Lower plenum is voided in case without LPIS in few
seconds. In LPIS_1120 and LPIS_1125 cases the lower plenum is continuously filled up with LPIS water so
void fraction oscillates. In LPIS_1121 case only a small amount of mass slumps to the lower plenum so void
fraction peak is also small.
Figure 7.111 and Figure 7.112 show pressure and water temperature in the lower plenum. When slumping
occurs all cases reach saturation temperature in the lower plenum. Pressure jump at the time of the slumping
reaches (from 3 bar) about 3.5-4 bar in all cases.
120
Assessment of Size Aspects in Modelling Molten Fuel Coolant Interaction
Chapter 7 - RESEARCH ACTIVITIES SUMMARY
WinGraf - 05-20-2003
3500
LPIS
V
V
V
Y V YZ XY XYZV X Z X
XZ
V Y
Y
Z
YZ Y
YZV YZV ZV YZ X Z X Z X
3000
V
Temperature (K)
2500
X
X
H2 production
V
2000
X
Z
V
Molten Pool cover
X
V
1500
1000
XYZV XYZV
X
V
Y
Z XYZ
V
V
Z
Y
Y
X
Y
X
YZ
X
XXX
YYY
ZZZ
VVV
500
LPIS_1125
LPIS_1121
LPIS_1120
DEBRIS_NOBK
V
bgmct0
bgmct0
bgmct0
bgmct0
slumping
0
0
500.
1000.
1500.
2000.
2500.
3000.
3500.
4000.
4500.
Time (s)
Figure 7.96 - Maximum core temperature in LPIS SCDAP calculations
WinGraf - 05-20-2003
3500
XXX LPIS_1120
YYY LPIS_1120
ZZZ LPIS_1120
3000
cadct60905
cadct60907
bgmct0
ZX
ZX
ZX
ZX
ZX
ZX
Z
Z
component 5 axial 9
Y
Temperature (K)
2500
Y
Y
2000
Y
Y
Y
Y
1500
ZX
1000
ZY
X
X Y ZX Y ZX Y ZX Y
ZX
Y ZX Y
ZX
Y
ZX
Y
ZX
Y
X
Y
Z
Z Y X
Y
component 7 axial 9
X
500
0
0
250.
500.
750.
1000.
1250.
1500.
1750.
2000.
2250.
Time (s)
Figure 7.97 - Maximum core temperature in LPIS_1120 case
121
Assessment of Size Aspects in Modelling Molten Fuel Coolant Interaction
Chapter 7 - RESEARCH ACTIVITIES SUMMARY
x 10 -5
5
WinGraf - 04-21-2003
XXX
YYY
H2 production (kg/s)
4
3
LPIS_1120
LPIS_1120
h2oxd2905
h2oxd2907
Y
component 7 axial 9
Y
Y
2
component 5 axial 9
Y
1
Y
X
0
1000.0
X
1050.0
X
X
X
X Y X Y X Y X
1100.0
1150.0
1200.0
Time (s)
X
X
1250.0
X
1300.0
1350.0
Figure 7.98 - Local H2 production rate in case LPIS_1120
WinGraf - 04-21-2003
1.20
molten pool
XXX
YYY
1.00
LPIS_1120
LPIS_1120
damlev905
damlev907
X
X
XY XY XY XY XY
component 5 axial 9
.80
Damage Level
X
.60
Rubble (fragmented)
Rupture (ballooning)
.40
component 7 axial 9
.20
XY XY XY
Y
Y
Y
XY XY
0
0
500.
1000.
1500.
2000.
Time (s)
2500.
3000.
3500.
Figure 7.99 - Damage state in components 5 and 7 in LPIS_1120 case
DAMLEV Damage state:
0.0 Intact geometry
0.1 Rupture due to ballooning
0.2 Rubble (fragmented)
0.4 Cohesive debris
1.0 Molten pool
122
Assessment of Size Aspects in Modelling Molten Fuel Coolant Interaction
Chapter 7 - RESEARCH ACTIVITIES SUMMARY
WinGraf - 04-21-2003
3500
XXX LPIS_1121
YYY LPIS_1121
ZZZ LPIS_1121
3000
bgmct0
cadct60907
cadct60809
XY
XY
XY
XY
XY
XY
XY
XY
XY
Y
YZ XYZ
X
Temperature (K)
2500
2000
component 7 axial 9
Z
1500
X
1000
X
Y
XYZ XYZ
500
Z
Y
X
XY
X
component 9 axial 8
Y
Y
X
Z
Z
Z
Z
Z
Z
Z
Z
Z
Z
Z
Z
Z
0
0
500.
1000.
1500.
2000.
2500.
3000.
3500.
4000.
4500.
Time (s)
Figure 7.100 - Maximum core temperature in LPIS_1121 case
WinGraf - 04-21-2003
1.20
XXX
YYY
LPIS_1121
LPIS_1121
molten pool
damlev907
damlev809
1.00
X
X
X
X
X
X
X
X
XY
XY
XY
XY
Damage Level
.80
component 7 axial 9
.60
component 9 axial 8
.40
Rubble (ballooning)
Rupture (fragmented)
.20
X
X
X
Y
Y
Y
Y
Y
Y
X
0
0
500.
1000.
1500.
2000.
2500.
3000.
3500.
4000.
4500.
5000.
Time (s)
Figure 7.101 - Damage state in components 7 and 9 in LPIS_1121 case
DAMLEV Damage state:
0.0 Intact geometry
0.1 Rupture due to ballooning
0.2 Rubble (fragmented)
0.4 Cohesive debris
1.0 Molten pool
123
Assessment of Size Aspects in Modelling Molten Fuel Coolant Interaction
Chapter 7 - RESEARCH ACTIVITIES SUMMARY
WinGraf - 05-20-2003
1.80
1.60
XXX
YYY
ZZZ
VVV
H2 production (kg/s)
1.40
LPIS_1125
LPIS_1121
LPIS_1120
DEBRIS_NOBK
bgth0
bgth0
bgth0
bgth0
1.20
1.00
.80
Y
.60
.40
Z
.20
0
500.
ZV
Y
X
YZVX
1000.
V
V
X Z
X
1500.
V
X
V
V
V
V
ZV
Z
V
V
YZV YZV Y
YZ YZV Y
Y
YZ XYZVXYZV
2000.
2500.
3000.
3500.
4000.
Time (s)
Figure 7.102 - H2 production rate in LPIS SCDAP calculations
WinGraf - 05-20-2003
1.80
V
1.60
V
XXX
YYY
ZZZ
VVV
LPIS_1125
LPIS_1121
LPIS_1120
DEBRIS_NOBK
V
1.40
Radius (m)
Z
Z
1.00
Y
.80
Y YZ
.60
X
Z
Y Y
X
Z
X
X
Z
X
Z
X
Z
X
X
X
Z
Z
1.20
repool0
repool0
repool0
repool0
X
Z
X
Z
X
X
Y
Y Y
X
Y Y
Y Y Y
slumping
Y
Z
.40
.20
0
0
.2
.4
.6
Time (s)
.8
1
1.2
X 10 4
Figure 7.103 - Radius of molten pool in LPIS SCDAP calculations
124
Assessment of Size Aspects in Modelling Molten Fuel Coolant Interaction
Chapter 7 - RESEARCH ACTIVITIES SUMMARY
x 10 4
8
WinGraf - 05-20-2003
7
V V V V V V
V V V V V V V
X
X
X
X
X
Z
Z
Z
Z
Z
X
X
Z
Z
6
Z
Mass (kg)
5
Z
4
XXX
YYY
ZZZ
VVV
3
LPIS_1125
LPIS_1121
LPIS_1120
DEBRIS_NOBK
masuo21
masuo21
masuo21
masuo21
2
YYYYYYYYYYY
1
0
0
.2
.4
.6
.8
1
1.2
1.4
1.6
1.8
X 10 4
Time (s)
Figure 7.104 - Total UO2 mass slumped in lower plenum in LPIS SCDAP calculations
WinGraf - 05-20-2003
4000
V
3500
3000
V
XXX
YYY
ZZZ
VVV
2500
Mass (kg)
V
V
V
V
V
LPIS_1125
LPIS_1121
LPIS_1120
DEBRIS_NOBK
V
V
V
V
V
V
V
V
V
maszo21
maszo21
maszo21
maszo21
2000
1500
X
X
X
X
1000
Z
Z
Z
Z
Z
Z
Z
500
Y Y Y Y Y Y Y Y Y Y Y Y Y Y Y
0
.2
.4
.6
.8
Time (s)
1
1.2
1.4
X 10 4
Figure 7.105 - Total ZrO2 mass slumped in lower plenum in LPIS SCDAP calculations
125
Assessment of Size Aspects in Modelling Molten Fuel Coolant Interaction
Chapter 7 - RESEARCH ACTIVITIES SUMMARY
x 10 4
1.2
WinGraf - 05-20-2003
1
V V V V V V
Z
Z
V V V V V V V
Z
X
Z
X
Z
X
XZ
XZ
Z
Z
X
X
Mass (kg)
.8
.6
XXX
YYY
ZZZ
VVV
.4
.2
LPIS_1125
LPIS_1121
LPIS_1120
DEBRIS_NOBK
masszr1
masszr1
masszr1
masszr1
YYYYYYYYYYY
0
0
.2
.4
.6
.8
1
1.2
1.4
1.6
1.8
X 10 4
Time (s)
Figure 7.106 - Total Zr mass slumped in lower plenum in LPIS SCDAP calculations
WinGraf - 05-20-2003
3000
XXX
YYY
ZZZ
VVV
LPIS_1125
LPIS_1121
LPIS_1120
DEBRIS_NOBK
2500
V
V
V
V
V
V
V
V
V
X
V
2000
Mass (kg)
massag1
massag1
massag1
massag1
V
1500
V
X
V
X
X
X
Z
X
X
V
Z
X
1000
500
Z
Z
Z
0
0
.2
.4
.6
Time (s)
.8
1
1.2
X 10 4
Figure 7.107 - Total Ag mass slumped in lower plenum in LPIS SCDAP calculations
126
Assessment of Size Aspects in Modelling Molten Fuel Coolant Interaction
Chapter 7 - RESEARCH ACTIVITIES SUMMARY
x 10 4
9
WinGraf - 05-20-2003
V V
8
V V V V V V V
V
7
X
V
Mass (kg)
6
V V
5
XXX
YYY
ZZZ
VVV
4
LPIS_1125
LPIS_1121
LPIS_1120
DEBRIS_NOBK
masliq1
masliq1
masliq1
masliq1
Z
Z
X
X
Z
Z
Z
X
X
Z
Z
3
2
Y
1
Y Y Y Y Y Y Y Y Y Y
V
0
0
.2
.4
.6
.8
Time (s)
1
1.2
1.4
1.6
X 10 4
Figure 7.108 - Total Liquid mass in lower plenum in LPIS SCDAP calculations
WinGraf - 05-20-2003
5.00
LPIS
4.50
SLUMPING
4.00
3.50
Y YX Y XYZ
XZ Y XZY X
Y
Z YXZ Y XZ
Y YXZ Y XZY YXZ Y XZY Y
Z
X
Z
XYZV
XZ
XZ
XZ
X
XZ
XZ
Level (m)
3.00
2.50
X
2.00
Z
V
1.50
1.00
Y
Y
.50
0
0
XXX
YYY
ZZZ
VVV
V
V w/o LPIS
V
V
.2
V
LPIS_1125
LPIS_1121
LPIS_1120
DEBRIS_NOBK
cntrlvar130
cntrlvar130
cntrlvar130
cntrlvar130
Z
V
.4
.6
.8
1
Time (s)
1.2
X 10 4
Figure 7.109 - Downcomer level in LPIS SCDAP calculations
127
Assessment of Size Aspects in Modelling Molten Fuel Coolant Interaction
Chapter 7 - RESEARCH ACTIVITIES SUMMARY
WinGraf - 05-20-2003
1.00
w/o LPIS
.90
XXX
YYY
ZZZ
VVV
.80
LPIS_1125
LPIS_1121
LPIS_1120
DEBRIS_NOBK
voidg106010000
voidg106010000
voidg106010000
voidg106010000
Void Fraction (-)
.70
Y
.60
X
X
.50
X
V
.40
SLUMPING
X
Y
.30
Z
Z
Z
.20
Z
X
.10
0
0
Y Y
Y
Y XZ
Y Y XZ
Y Y XZ
YV XZ XZY Y ZY Y ZY XZY YXZ X
VXY
Z V
.2
.4
.6
.8
Time (s)
Z
X
XZ
Z
Z
X
1
1.2
1.4
1.6
X 10 4
Figure 7.110 - Void fraction in the lower plenum in LPIS SCDAP calculations
WinGraf - 05-20-2003
.60
.55
.50
XXX
YYY
ZZZ
VVV
Pressure (MPa)
.45
LPIS_1125
LPIS_1121
LPIS_1120
DEBRIS_NOBK
p106010000
p106010000
p106010000
p106010000
.40
X
Y Z X
Y Y Y
Z
.35
.30
V
X
Y
Z
.25
Z
X
YZ
Y
X
Z
Y Y
V V V V V V
Y
Y
Y Y
Y Z X YXZ XY Z
YZ Y
X
Z
XZ
XZ
X
Z
X
Z
X
X
Z
V V V V V V V V V V V V
X
Z
Z
X
X
Z
X
.20
.15
.10
0
.2
.4
.6
.8
Time (s)
1
1.2
1.4
1.6
X 10 4
Figure 7.111 - Pressure in the lower plenum in LPIS SCDAP calculations
128
Assessment of Size Aspects in Modelling Molten Fuel Coolant Interaction
Chapter 7 - RESEARCH ACTIVITIES SUMMARY
WinGraf - 05-20-2003
150.0
slumping
140.0
Z
120.0
Temp (°C)
Z
Y
Y X
YZ X
V V Y
VZ V VY Z
X
V
Y
V V V V
ZY
V
Z YZ
V Y
X
VY
Y Y
V
Y
X
Z
X
X
Y
Y
Y
Y
Y
Z
Z
Z
Z
130.0
110.0
100.0
90.0
XXX
YYY
ZZZ
VVV
70.0
LPIS_1125
LPIS_1121
LPIS_1120
DEBRIS_NOBK
Z
V
V
V
X
V
Z
X
Z
X
80.0
V
X
Z
X
XZ
Z
X
X
tempf106010000
tempf106010000
tempf106010000
tempf106010000
X
X
X
X
60.0
0
.2
.4
.6
.8
Time (s)
1
1.2
1.4
X 10 4
Figure 7.112 - Water temperature in the lower plenum in LPIS SCDAP calculations
129
Assessment of Size Aspects in Modelling Molten Fuel Coolant Interaction
Chapter 7 - RESEARCH ACTIVITIES SUMMARY
d)
Core degradation, molten pool and slumping avoided. Cases with LPIS injection time = 900s,
1100s, 1110s, 1115s and 1119s
In all these cases core is recovered due to the core emergency LPIS system, as it is observed in Figure 7.113.
Only axial nodes 9 of components 1 to 7 in LPIS_1119 suffer break due to clad ballooning (Figure 7.114).
H2 is produced (Figure 7.115 and Figure 7.116) due to water injected but it is much less than in the core
unrecovered cases.
WinGraf - 05-20-2003
3500
H
3000
H
H
H
H
H
w/o LPIS
H
H
H
H
H
2500
Temperature (K)
H
2000
XXX
YYY
ZZZ
VVV
JJJ
HHH
H
1500
XYHZ
J
HZ
X ZV
VX
H
Y
JY VJ
HZ
XYZVXJYHZVXJY
1000
500
JY
V Z
LPIS_900
LPIS_1100
LPIS_1110
LPIS_1115
LPIS_1119
DEBRIS_NOBK
bgmct0
bgmct0
bgmct0
bgmct0
bgmct0
bgmct0
V
X
XJYZVXJYZVXJYZ VXJYZVXJYZVXJYZVXJYZVXJYZ VXJYZVXJYZVXJYZVXJYZVJ
0
0
500.
1000.
1500.
2000.
2500.
3000.
3500.
Time (s)
Figure 7.113 - Maximum core temperature in LPIS SCDAP calculations
WinGraf - 04-21-2003
1.00
.90
XXX
YYY
ZZZ
VVV
.80
Damage Level
.70
LPIS_1119
LPIS_1119
LPIS_1119
LPIS_1119
damlev901
damlev903
damlev905
damlev907
.60
.50
.40
.30
Rupture (ballooning)
.20
.10
XYZVXYZVXYZVXYZVXYZVXYZVXYZVXYZVXYZVXYZVXYZVXYZVXYZVXYZVXYZV
0
0
500.
1000.
1500.
2000.
Time (s)
2500.
3000.
3500.
4000.
Figure 7.114 - Damage state in components 1, 3, 5 and 7 in LPIS_1119 case
DAMLEV Damage state:
0.0 Intact geometry
0.1 Rupture due to ballooning
0.2 Rubble (fragmented)
0.4 Cohesive debris
1.0 Molten pool
130
Assessment of Size Aspects in Modelling Molten Fuel Coolant Interaction
Chapter 7 - RESEARCH ACTIVITIES SUMMARY
WinGraf - 05-20-2003
.25
.23
XXX
YYY
ZZZ
VVV
JJJ
HHH
.20
H2 production (kg/s)
.18
LPIS_900
LPIS_1100
LPIS_1110
LPIS_1115
LPIS_1119
DEBRIS_NOBK
bgth0
bgth0
bgth0
bgth0
bgth0
bgth0
H
.15
H
H
.13
H
.10
H
H
H
H
H
H
H
H
.08
.05
J
Y Z
V
H
.03
0
400.
X
800.
600.
Z
J
X V
H
Y
1000.
Y
V Z
1200.
H
1400.
Time (s)
1600.
1800.
2000.
2200.
2400.
Figure 7.115 - H2 production rate in LPIS SCDAP calculations
WinGraf - 05-20-2003
6.00
H2 produced (kg)
5.00
4.00
3.00
Z
V
Z
V
Z
V
Z
V
Z
V
Z
V
Z
V
Z
V
Z
V
Z
V
Z
V
Z
V
Z
V
Z
V
Z
V
Z
V
Z
V
Y
Y
Y
Y
Y
Y
Y
Y
Y
Y
Y
Y
Y
Y
Y
Y
Y
XXX
YYY
ZZZ
VVV
JJJ
HHH
2.00
1.00
LPIS_1119
LPIS_1115
LPIS_1110
LPIS_1100
LPIS_900
DEBRIS_NOBK
H2_ACCUM
H2_ACCUM
H2_ACCUM
H2_ACCUM
H2_ACCUM
H2_ACCUM
V
0
0
J
Z
J Y
1000.
J
J
2000.
J
J
3000.
J
J
J
J
J
4000.
5000.
Time (s)
J
J
6000.
J
J
7000.
J
J
J
8000.
J
9000.
Figure 7.116 - H2 accumulated in LPIS SCDAP calculations
131
Assessment of Size Aspects in Modelling Molten Fuel Coolant Interaction
Chapter 7 - RESEARCH ACTIVITIES SUMMARY
7.5.6.
INITIAL CONDITIONS FOR A MFCI DETAILED STUDY
From the performed RELAP5/SCDAP 3.2 calculations it is possible to obtain the conditions (system
pressure, coolant temperature, molten mass temperature, molten mass composition, water level in the lower
plenum, etc.) just before the molten pool slumping towards the lower plenum. These conditions can be used
as initial conditions in a specific MFCI code input (as JRC-COMETA [6]) so a detailed MFCI study could be
later performed.
Figure 7.117, Figure 7.118 and Figure 7.119 show volume, liquid and UO2 mass in core region at the molten
pool slumping time in all performed calculations, in which core degradation leads to a molten pool. Data
have been retrieved from RELAP5/SCDAP 3.2 output files, that give every print out time the status,
temperatures, pressures, mass components, etc. of the degraded core region.
Table 7.11 presents some quantities of cases DEBRIS_NOBK (Base case, without LPIS), LPIS_1120 and
LPIS_1300 because they are the most representative. LPIS_1300 is different from the common trend, with
very small molten pool mass. The last column shows system conditions used in the FARO melt fuel coolant
interaction experiment Test L-11, where 151 kg of molten melt at 2800 K were poured in a pool of water at
4.9 MPa and saturation temperature.
Total mass of melt in the Base case is about 80000 kg, 69710 kg of which being UO2, 3586 kg ZrO2 and
9191 kg Zr. In general total mass or total volume of molten pool varies from about 60000 kg in the case with
lower mass up to 80000 kg in the Base case. Melt mass is in calculated cases constituted most of UO2 and Zr,
FARO Tests were in general performed with a mixture of about 80% UO2 and 20% ZrO2, that it is a little
different from results obtained in the proportion of Zr and ZrO2. Melt mass temperature is similar in all cases
about 2900 K. System pressure, about 0.2 MPa is result of LOCA depressurization.
It seems from figures that the effect of LPIS delays slumping of the molten pool from about 1 hour in the
Base case up to about 11000s (3 hours) in case LPIS_1125, another effects are less molten mass produced
and some degrees of subcooling of the residual water in the lower plenum. Cases LPIS_1121 and LPIS_1300
are far from the common trend, with early slumping time and small molten pool, reasons of the early
slumping are not yet well identified and may be related to the large degree of uncertainty of the slumping
time with RELAP5/SCDAP 3.2 code.
Specific MFCI calculation will also be useful to compare some quantities after slumping time, as lower
plenum pressure, temperature or void fraction, but these detailed calculation are beyond the objectives of the
presented study.
Mass UO2 (kg)
Mass ZrO2 (kg)
Mass Zr (kg)
% UO2
%ZrO2
%Zr
Tot mass melt (kg)
T melt (K)
T system (dcmer) (K)
P lower plenum (Pa)
T lower plenum (K)
Subcooling (°C)
Slump at (minutes)
Without LPIS
BASE
69710
3586
9191
84.51
4.35
11.14
82487
2973.00
sat
2.4E+05
sat
0
64.2
LPIS at 1300s
LPIS at 1120s
FARO Test L-11
16700
1891
1945
81.32
9.21
9.47
20536
2943.50
387.65
2.8E+05
381.15
22.24
58.8
51980
707
9535
83.54
1.14
15.32
62222
2977.60
sat
2.8E+05
391.34
12.05
164.7
115
29
7
76.16
19.21
4.64
151
2800
49E+05
sat
0
-
Table 7.11 - Initial conditions for a MFCI detailed calculation at the moment of slumping
132
Assessment of Size Aspects in Modelling Molten Fuel Coolant Interaction
Chapter 7 - RESEARCH ACTIVITIES SUMMARY
Volume in Molten Pool
12.00
Volume_LPIS_2000
3
Volume (m )
10.00
8.00
Volume_LPIS_1500
Volume_LPIS_1200
6.00
Volume_LPIS_1150
Volume_LPIS_1125
4.00
Volume_LPIS_1121
Volume_LPIS_1120
2.00
Volume_w/o_LPIS
Volume_LPIS_1300
0.00
0.0
2000.0
4000.0
6000.0
8000.0
10000.0 12000.0
Time (s)
Figure 7.117 - Molten pool volume up to the slumping time in all SCDAP calculations
Liquid in Molten Pool
30000.00
Liquid_LPIS_2000
Mass (kg)
25000.00
20000.00
Liquid_LPIS_1500
Liquid_LPIS_1200
15000.00
Liquid_LPIS_1150
Liquid_LPIS_1125
10000.00
Liquid_LPIS_1121
Liquid_LPIS_1120
Liquid_Without_LPIS
Liquid_LPIS_1300
5000.00
0.00
0.0
2000.0
4000.0
6000.0
8000.0 10000.0 12000.0
Time (s)
Figure 7.118 - Molten pool liquid phase up to the slumping time in all SCDAP calculations
133
Assessment of Size Aspects in Modelling Molten Fuel Coolant Interaction
Chapter 7 - RESEARCH ACTIVITIES SUMMARY
UO2 in Molten Pool
80000.00
UO2_LPIS_2000
70000.00
UO2_LPIS_1500
Mass (kg)
60000.00
UO2_LPIS_1200
50000.00
UO2_LPIS_1150
40000.00
UO2_LPIS_1125
UO2_LPIS_1121
30000.00
UO2_LPIS_1120
20000.00
UO2_w/o_LPIS
10000.00
UO2_LPIS_1300
0.00
0
2000
4000
6000
8000
10000
12000
Time (s)
Figure 7.119 - Molten pool UO2 mass up to the slumping time in all SCDAP calculations
7.5.7.
CONCLUSIONS
Calculations show that the RELAP5/SCDAP 3.2 modelling theory in the late damage phase (as it is MFCI)
reproduces phenomena as a function of the assumed option. The code assumes either interaction between
molten mass and the residual water in the lower plenum or no interaction at all. In the latter the stored
energy remains in the molten mass and is only removed by heat transfer from the surfaces of the resulting
debris in the lower plenum.
Calculations have demonstrated that in order to perform a detailed MFCI study it is better to use a MFCI
thermalhydraulic code in which the interaction grade is result from the initial conditions of the system, the
molten pool conditions and the own interaction transient evolution and it is not an input parameter set by the
code user. RELAP5/SCDAP 3.2 can be used to obtain the final conditions as initial conditions in a MFCI
thermalhydraulic code. In that way RELAP5/SCDAP 3.2 and a detailed MFCI code can be seen as
complementary tools.
Figure 7.65 shows that if LPIS injection time is lower than 659s after the break (case LPIS_1119) core
temperature increase can be mitigated. The maximum time available for the operator to restore cooling
conditions via the LPIS is about 10 minutes after the break (659s). Above this time meltdown cannot be
avoided. Core temperature is powered due to the effect of the exothermic H2 production when LPIS water
reacts with cladding at temperatures above about 1500K (oxidation temperature).
Finally conditions up to the moment of the slumping are presented in order to select them as initial
conditions for a detailed MFCI study.
134
Assessment of Size Aspects in Modelling Molten Fuel Coolant Interaction
Chapter 8 - COMETA CODE MODIFICATIONS
8. COMETA CODE MODIFICATIONS
8.1. PREVIOUS COMETA LOGIC AND ACTUAL NEEDS
COMETA code was initially planned and developed for prediction of the thermalhydraulic behaviour of
FARO facility, for design, definition of operational procedures and tests interpretation. The analytical
activities were based on broad pre-test and post-test studies. The main objective of the code is to predict the
behaviour of MFCI phenomena, including steam explosion events in any geometry 1d or 2d. The code is
written in FORTRAN language and is structured in modules and subroutines depending on the functions and
phenomena to deal with or simulate.
COMETA code is organized in a number of lumped volumes connected with junctions in the so-called 1d
nodalization. A 2d nodalization can be built up connecting a number of macro-volumes (containing an
arbitrary number of radial and axial volumes) and macro-junctions (Figure 6.1). The real conditions in every
thermalhydraulic volume are used to set the drops conditions while they are in the volume.
An example of COMETA 1d nodalization used in the simulations for FARO geometry is presented in Figure
8.1 as a vertical cut. In some studies, however, for accuracy reasons a 2d nodalization is used. For example,
FARO test L-33 uses such a nodalization to set the exact position of the trigger event. Figure 7.10 shows the
real shape of 2d nodalization of FARO Test L-33 and Figure 7.29 the corresponding vertical cut.
A particular important issue concerning molten melt behaviour and its simulation with COMETA is the
movement of the little drops formed due to the jet fragmentation when it enters into an hypothetical lower
plenum water pool or in the water FARO release vessel. These particles float without any definite path in the
water pool and fall down towards the vessel bottom, following the motion dynamic laws (gravity and
friction).
In the previous COMETA logic the drops are considered to move or fall down with only a one-direction
velocity, i.e. vertical coordinate and towards the vessel bottom. This occurred for any geometry 1d or 2d. In
1d geometry they could only fall down passing through the thermalhydraulic volumes (Figure 8.2 a). In the
2d geometry the original model assumed that drops were falling down in the central volumes without any
radial component (Figure 8.2 b).
In FARO tests in which the initial jet size was relatively large (10 cm, versus a diameter of 70 cm) the 1d
assumption was reasonable. As the jet size was reduced to 5 cm or in a real plant case, this assumption was
no more applicable and it was necessary to model also the 2d flow and therefore it was necessary to include
also the radial drops displacement.
It was therefore decided to modify the COMETA code subroutine that determines drops movement to
include a radial velocity component (Figure 8.3). Changes are explained in the following subchapter.
135
Assessment of Size Aspects in Modelling Molten Fuel Coolant Interaction
Chapter 8 - COMETA CODE MODIFICATIONS
18
Melt catcher
17
16
Insulation
15
14
Vessel steel
13
12
11
10
9
8
7
6
5
4
3
2
1
32
31
30
29
28
27
26
25
24
23
22
21
20
19
Figure 8.1 - COMETA nodalization 1d for FARO Test L-29
136
Assessment of Size Aspects in Modelling Molten Fuel Coolant Interaction
Chapter 8 - COMETA CODE MODIFICATIONS
z
19
29
4
18
3
28
2
17
27
a) 1d
b) 2d
Figure 8.2 - Drops movement in thermalhydraulic volumes in 1d and 2d nodalization in previous COMETA logic
r
z
19
29
18
28
17
27
Figure 8.3 - Drops movement in thermalhydraulic volumes in 2d nodalization in present COMETA logic
8.2.
MODIFIED COMETA. CHANGES INTRODUCED
The application of COMETA code to the prediction of tests performed in the FARO facility and the
identification of problems in scale-up to reactor configurations, larger than scaled test facilities, lead to
improve COMETA code lacks. The modifications performed in the code included part of engineering and
part of computer programming analysis.
The drops movement subroutine (“mapfrag”) was modified in order to include the radial component (r). The
aim of the subroutine is to determine from the drop position (coordinates x, r) the thermalhydraulic volume
number in which the drop is positioned in the correspondent time step.
In that way for the 2d nodalization drops movement is not only restricted to central volumes but also to all
lateral volumes.
137
Assessment of Size Aspects in Modelling Molten Fuel Coolant Interaction
Chapter 8 - COMETA CODE MODIFICATIONS
The COMETA code is organized in explicit modules or subroutines that send and receive information from
each other to calculate thermalhydraulic conditions in each time step (as a pure thermalhydraulic code) but
also melt conditions are calculated in each time step for the three melt shapes of the melt field (jet, drops and
debris).
Mainly in each global time step (t1 = t0+? t) both the thermalhydraulic equations and the melt fragmentation
equations are solved separately in two modules (THSOLVER and FRASOLVER):
The so-called THSOLVER (built up of thermalhydraulic subroutines) solves equations obtaining 6
thermalhydraulic quantities for the time step t1: pressure, internal energy for each phase (liquid and vapour),
void fraction for each phase (liquid and vapour) and velocities for each phase (liquid and vapour). To solve
equations it uses melt heat transfer calculated in the previous time step t0.
The thermalhydraulic quantities found are used for the so-called FRASOLVER to solve melt conditions in
the time step t1 (3 melt quantities are obtained: heat transfer for drops, heat transfer for jet and heat transfer
for debris). In general the time step interval ? t is equal for THSOLVER and FRASOLVER modules, but in
some cases it is necessary to reduce global ? t for the melt conditions solving the melt quantities in shorter
time steps.
The drops movement subroutine (“mapfrag”) is implicit to the FRASOLVER module, basically for each
drop and its exact position the subroutine calculates the thermalhydraulic volume that contains the drop in
the particular time step. Initially the subroutine was only taken into account the vertical volumes (z
direction), central volumes in case of 2d nodalization. During the modifications radial component was added
(r direction) so radial and not only central volumes were included in a 2d nodalization.
“mapfrag” receives as parameters for the calculation the following: position of the drop (components r and z
). The returned variables are: the middle points of each thermalhydraulic volume (zmed, rmed), the number
of the volume (jvol) where the drop is calculated to be in the time step and rmax that is the maximum radial
distance of the volume (jvol).
In the subroutine first part, or preparation part, running a loop a matrix stores all limit coordinates (rmin,
rmax, zmin and zmax) (Figure 8.4) of each thermalhydraulic volume (j) in the geometry (cylindrical or
plenum shape).
In the second part or running part the code searches for the position of the drop using the drops coordinates
received (r, z), i.e. running the loop for each thermalhydraulic volume (j) it is checked if the coordinate r is
between rmin and rmax and coordinate z between zmin and zmax of the volume (j), if so the drop is in the
volume (j) and the subroutine returns jvol=j and the middle points of the jvol volume (zmed, rmed).
(0,0)
rmin
rmax
zmin
jvol
Drop (r,z)
(rmed, zmed)
zmax
Figure 8.4 - Limit coordinates of the jvol volume and drop coordinates
138
Assessment of Size Aspects in Modelling Molten Fuel Coolant Interaction
Chapter 8 - COMETA CODE MODIFICATIONS
The subroutine “movedrrad” is the one that really calculates drops relative movement and drops relative
velocity.
This procedure calls “mapfrag” for every drop present and takes also into account if the drop position (x1)
has overpassed the dimensions of the geometry (r>rmax), laterally or vertically and sets it symmetrically to
the radial volume or to the upper volume inverting drop absolute radial velocities in both cases. This is the
correspondent part of the code:
c check if reached the side, if yes vel=-vel
c
if(x1.ge. rmax) then
x1 = rmax ! max extension
veldropRad(k) = -veldropRad(k)
endif
If the drop position (x1) has reached the bottom it takes velocity=0. This is the correspondent part of the
code:
c check if reached the center , if yes vel=0
c
if(x1.le.0.0) then
x1=0.e-4
veldropRad(k)=0.0
endif
Drops arrived to the bottom could be included in the fused debris conglomeration depending on its
temperature.
The radial velocity of the (k) drop in the liquid phase (vfr) and in the gas phase (vgr) is calculated as:
vgr=vgradn(iv)+(rdrops(k)-deltar/2)*dvgdrn(iv) (Eq. 1)
vfr=vfradn(iv)+(rdrops(k)-deltar/2)*dvfdrn(iv) (Eq. 2)
where:
rdrops(k) is the radial position of the (k) drop
vfradn(iv) is the radial velocity of the volume (iv) liquid phase
vgradn(iv) is the radial velocity of the volume (iv) gas phase
deltar=rade(iv)-radi(iv) where rade is the external radius of the volume (iv) and radi the internal radius.
dvgdrn(iv) and dvfdrn(iv) are calculated in the subroutine “velrad”
dvfdrn(iv)=(vfo-vfi)/(rade(iv)-radi(iv)) (Eq. 3)
dvgdrn(iv)=(vgo-vgi)/(rade(iv)-radi(iv)) (Eq. 4)
where:
subscript o means out junctions
subscript i means inlet junctions
vfi is the liquid velocity of all inlet junctions of volume (iv)
139
Assessment of Size Aspects in Modelling Molten Fuel Coolant Interaction
Chapter 8 - COMETA CODE MODIFICATIONS
vfo is the liquid velocity of all outlet junctions of volume (iv)
vgi is the gas velocity of all inlet junctions of volume (iv)
vgo is the gas velocity of all outlet junctions of volume (iv)
These four velocities are calculated following the relationship between the velocities of inlet and outlet
junctions in the gas or liquid phase and the mass flow that passes through the volume (iv):
v=
M
Aαρ
where:
v is the junction velocity (m/s)
M is the mass flow (kg/s)
A is the inlet or outlet junction area (m2)
a is the void fraction of liquid or gas
? is the density of liquid or gas (kg/m3)
(vgr) and (vfr) are calculated solving (Eq.1 and Eq.2) and the average velocity is calculated as a weighted
mean value of (vgr) and (vfr) in the gas, liquid and non-condensable phases in the volume (iv):
velmed=alfagn(iv)*vgr+alfafn(iv)*vfr +alfanc(iv)*vgr (Eq. 5)
where:
alfagn(iv) is the vapour void fraction in the volume (iv)
alfafn(iv) is the liquid void fraction in the volume (iv)
alfanc(iv) is the non-condensable void fraction in the volume (iv), non-condensable gases are assumed to
have the vapour velocity
Finally the drop relative velocity (velr) is the sum of the velocity of the drop (veldropRad(k)) and the
average velocity of the fluid in the volume (velmed) calculated before:
velr=veldropRad(k)+velmed (Eq. 6)
In the following subchapter the new version of the code including the modifications explained is tested
versus FARO experimental data and compared to previous COMETA code non modified.
140
Assessment of Size Aspects in Modelling Molten Fuel Coolant Interaction
Chapter 8 - COMETA CODE MODIFICATIONS
8.3. CALCULATIONS WITH MODIFIED COMETA LOGIC AND PREVIOUS COMETA
LOGIC
L-28 was the FARO test chosen to perform two comparative calculations in order to demonstrate the
different results achieved using the previous COMETA logic and the modified one. This was the first FARO
test in which the jet size was reduced to 5 cm nozzle diameter from 10 cm (versus a vessel diameter of 70
cm) so the 2d nodalization assumption was the most reasonable, including the 2d flow and therefore the
radial drops displacement. Figure 8.5 shows 2d nodalization for COMETA calculations. The conditions of
the performed test are the following:
Mass:
Melt Temperature
Initial water Temperature
Initial water level
Delivery nozzle diameter
Initial Pressure:
Vessel geometry
174.9 kg
3052 K
Tsat (saturation conditions)
1.44 m
0.044 m (5 cm jet versus FAT vessel 70 cm diameter)
5.1 bar
FAT
Melt
catcher
1
168
170
172 174 176
167
169
171 173 175
150
149
148
147
138
137
120
119
118
117
36
35
34
33
32
31
30
29
28
27
26
25
24
23
22
21
20
19
18
17
154
153
152
151
140
139
124
123
122
121
56
55
54
53
52
51
50
49
48
47
46
45
44
43
42
41
40
39
38
37
158
157
156
155
142
141
128
127
126
125
76
75
74
73
72
71
70
69
68
67
66
65
64
63
62
61
60
59
58
57
Insulation
Vessel steel
Water
level
162
161
160
159
144
143
132
131
130
129
96
95
94
93
92
91
90
89
88
87
86
85
84
83
82
81
80
79
78
77
166
165
164
163
146
145
136
135
134
133
116
115
114
113
112
111
110
109
108
107
106
105
104
103
102
101
100
99
98
97
16
15
3
2
Figure 8.5 - COMETA 2d nodalization for FARO test L-28
141
Assessment of Size Aspects in Modelling Molten Fuel Coolant Interaction
Chapter 8 - COMETA CODE MODIFICATIONS
Case L28denm is the calculation performed with the previous COMETA logic. In this case the drops
generated in the jet interaction with the water present in the vessel are retained to move vertically in the
central volumes (volume 17, bottom vessel, to volume 36 from Figure 8.5).
Case L28denh is the calculation performed with the modified COMETA logic, in this case the drops
generated in the interaction have also the radial movement therefore they can travel from one internal
volume to the laterals and vice versa. Their movement is free in all water volumes (volume 17, bottom
vessel, to volume 116 from Figure 8.5).
Results of the calculations are presented in the following figures. Global pressure is presented in Figure 8.6.
Pressure behaviour evolutions are compared to the experimental data, they follow the overall trend of the
experimental data and both are similar in absolute values. The changes introduced in the code have not
disturbed the general prediction of the experiment itself.
WinGraf - 10-28-2003
35.0
30.0
XXX
YYY
ZZZ
L28C
PT.350.1665
L282DENM p001
L282DENH p001
COMETA MODIFIED
25.0
Pressure (bar)
COMETA NON MODIFIED
20.0
15.0
10.0
Y
ZX
Y
Y
Z
Z
Y
Z
X
Z
X
Y Y
Z
X
Y Z
X
Y Z
X
YZ
X
YZ
YZ
X
Y
Z X
Y
Z X
Y
Y
Z Z
X
X
X
Z
ZY
Y
Z
X
X
EXPERIMENT
X
X
X
Z
Y
ZXY
5.0
0
0
2.0
4.0
6.0
8.0
10.0
12.0
Time (s)
Figure 8.6 - Pressure behaviour in COMETA calculations for FARO test L-28
The comparison between both calculations to evidence the improvements in the code should be regarded in
the local phenomena, i.e. local temperatures, local void fraction and local fragmented melt or drops
distribution.
The local void fraction and fuel distribution inside a mixture strongly affects the pressure wave propagation
and thus it has a strong importance in the evaluation of the explosivity and the effects of the explosion. Since
the fuel distribution also affects the void distribution (steam + non-condensable gases) it is of decisive
importance the correct local distribution of the fuel drops.
In a mixture where the drops fragment and distribute uniformly the explosion potential is higher if a trigger
is applied. If the fuel fragments less, it is more likely that the application of a trigger does not determine an
explosion. However a so called "stratitied explosion" can occur. That is when a trigger is applied to a layer of
fuel with the water above.
From Figure 8.7 to Figure 8.14 water temperatures calculation results are plotted for different vertical and
two radial positions in the vessel. They are compared to experimental data retrieved from the thermocouples
placed into the FARO vessel during the experiment.
142
Assessment of Size Aspects in Modelling Molten Fuel Coolant Interaction
Chapter 8 - COMETA CODE MODIFICATIONS
Results are compared in the so-called “stable melt discharge and quenching phase” of the experiment [Figure
9 of [37]] that is from 2s to about 6s. That period is the real situation in which a possible spontaneous
explosion could take place.
It can be observed for most figures that the temperatures calculated with the COMETA modified are higher
(due to the presence of drops in the lateral volumes) and closer to the experimental results in that period (2s6s). So the code with the modifications included to take account of the drops radial movement has improved
also the general prediction of the experiment in terms of temperature. This gives a further result of
COMETA code assessment.
As it was discussed before, modifications in the code are very important for the prediction of local quantities
(void fraction and drops distribution) critical in the evaluation of the explosivity and the effects of a possible
steam explosion.
So void fraction and drops mass calculated quantities (in the correspondent volumes in which where
compared the temperatures) are also presented in the report during the “stable melt discharge and quenching
phase”. These quantities cannot be compared to experimental values but are very important in some studies.
Local temperatures were compared and are in agreement with experimental values. As temperatures are
partly a result of the local drops distribution and the void fraction, they contribute to assess code capabilities
in the prediction of void fraction and drops distribution.
From Figure 8.15 to Figure 8.22 void fraction calculation results are plotted for different lateral volumes at
four heights and two radial positions in the vessel. In general it can be observed that modified COMETA
calculations reach a higher void fraction in all these volumes. Void fraction is influenced by a lot of
parameters but also by the vapour produced by hot drops created in the central volumes, which with the new
version of the code can travel to lateral volumes.
Finally the distribution of drops mass is plotted to verify the lateral presence of drops in the calculations
performed with the modified code version. From Figure 8.23 to Figure 8.26 the mass of drops is shown in
volumes at four heights and two radial positions in the vessel during the “stable melt discharge and
quenching phase” (2s-6s).
143
Assessment of Size Aspects in Modelling Molten Fuel Coolant Interaction
Chapter 8 - COMETA CODE MODIFICATIONS
WinGraf - 10-29-2003
210.0
XXX
YYY
ZZZ
VVV
200.0
L28C
L28C
L282DENM
L282DENH
TW.295.0420.150
TW.205.0420.150
tf062
tf062
Y
Modif
Temp (°C)
190.0
X
Y
Y
Y
Y
Experiment
Y
YX
Y
160.0
X
Y
X
X
X
Z
YX
V
V
Z
V
Z
Z
Z
V
V
Z
X
Z
V
V
X
X
Y
V
X
ZVX ZVYXZV ZV ZV ZV Z V Z VZ
150.0
V
V
X
Y
X
Y
Y
YX
V
X
Y
Y
X
180.0
170.0
X
Y
X
YX
Z
Z
Z
No Modif
140.0
0
1.0
2.0
3.0
4.0
5.0
6.0
7.0
Time (s)
Figure 8.7 - Water temperature at height 0.420 m and 0.150 m radial vessel position
WinGraf - 10-29-2003
220.0
210.0
XXX
YYY
ZZZ
VVV
200.0
L28C
L28C
L282DENM
L282DENH
TW.205.0420.330
TW.025.0420.330
tf102
tf102
Modif
Y
Temp (°C)
190.0
180.0
X
V
X V V
X X ZY Z
Y ZY
Experiment
Y
Y
YX
V
X
170.0
Y
X
160.0
Y
X
Y
ZVX ZVYXZV ZV YZXV Y
ZV YZXV Z VZ
YX
X
150.0
YZ
X
X
V Z
V
V
Z
Z
V YX
YX Z
VYX
YX
V
No Modif
X Z
V
Z
140.0
0
1.0
2.0
3.0
4.0
5.0
6.0
7.0
Time (s)
Figure 8.8 - Water temperature at height 0.420 m and 0.330 m radial vessel position
144
Assessment of Size Aspects in Modelling Molten Fuel Coolant Interaction
Chapter 8 - COMETA CODE MODIFICATIONS
WinGraf - 10-29-2003
220.0
210.0
XXX
YYY
ZZZ
VVV
200.0
L28C
L28C
L282DENM
L282DENH
TW.295.0690.150
TW.205.0690.150
tf066
tf066
Modif
X
Y
Temp (°C)
190.0
170.0
Y
Experiment X
Y
Y
X
Y
X
Y
160.0
X
YX
X
V
X
ZVX ZVY ZV ZV ZV ZV Z V Z Z
150.0
X
Y
V
Z
YX
V
V
Z
V
V
V
V
Z
V
YX
X
VY
X
YX
X
V
X
Y
Y
X
X
YX
180.0
Y
YX
Y
Y
Z
V
Z
Z
Z
Z
Z
No Modif
Z
140.0
0
1.0
2.0
3.0
4.0
5.0
6.0
7.0
Time (s)
Figure 8.9 - Water temperature at height 0.690 m and 0.150 m radial vessel position
WinGraf - 10-29-2003
220.0
210.0
XXX
YYY
ZZZ
VVV
200.0
L28C
L28C
L282DENM
L282DENH
Temp (°C)
190.0
TW.115.0690.330
TW.295.0690.330
tf106
tf106
X
Modif
180.0
170.0
Experiment
V
YX
X
Y Z
X
X Y
VZ
Y
X
V
ZVX ZVYXZV YXZV YZXV Y
ZV Z Z
160.0
150.0
YV
X
Z
X Y
X
Y Y VZV
V
Z
X X
X Y YZ
V
V
Z X YV ZV
Y
Z
Z
X YX
X ZY V
YV
X
Y
No Modif
140.0
0
1.0
2.0
3.0
4.0
5.0
6.0
7.0
Time (s)
Figure 8.10 - Water temperature at height 0.690 m and 0.330 m radial vessel position
145
Assessment of Size Aspects in Modelling Molten Fuel Coolant Interaction
Chapter 8 - COMETA CODE MODIFICATIONS
WinGraf - 10-29-2003
210.0
XXX
YYY
ZZZ
VVV
JJJ
200.0
L28C
L28C
L28C
L282DENM
L282DENH
TW.295.0960.150
TW.205.0960.150
TW.025.0960.150
tf070
tf110
Modif
Y Z X ZJ
J Z J Y YX
X J
ZY
X
Temp (°C)
190.0
Y
ZX
X
YX X Y Z J
Experiment
Y
Z
J
V
J
J
Y
V
JX
Z JZ
Z
ZYX YX Z J X X Z
V
V
Y Y
Y
Y
X
V
No Modif
Z
V
X
Z
YX
J
V V
V
J VJ V
VX
J VJ VJ V
180.0
170.0
160.0
150.0
J
ZY ZY
X X
J
Z Z
YX YX
V
V
V
V
140.0
0
1.0
2.0
3.0
4.0
5.0
6.0
7.0
Time (s)
Figure 8.11 - Water temperature at height 0.960 m and 0.150 m radial vessel position
WinGraf - 10-29-2003
210.0
XXX
YYY
ZZZ
VVV
200.0
L28C
L28C
L282DENM
L282DENH
V
TW.205.0960.330
TW.025.0960.330
tf110
tf110
Experiment
190.0
Temp (°C)
V
170.0
VX
Y
X
Y Z
V YX Z
Z
YX
160.0
YV
X
YX
YX
Z
Modif
180.0
Z
VYX
V
V
VX
Y YX ZY
Z
X
YX
Z
V
V
Y ZYX
X
Y YX V
XV
Z
Z
Z
No Modif
V Z
X
Z
ZVX ZVYXZV YXZV YZXV Y
150.0
140.0
0
1.0
2.0
3.0
4.0
5.0
6.0
7.0
Time (s)
Figure 8.12 - Water temperature at height 0.960 m and 0.330 m radial vessel position
146
Assessment of Size Aspects in Modelling Molten Fuel Coolant Interaction
Chapter 8 - COMETA CODE MODIFICATIONS
WinGraf - 10-29-2003
220.0
210.0
V
J
J
X
Z
ZYX ZYX YX ZY
ZYX
200.0
ZY
VJX
Z
VYX
J
ZX
X Y YJ
X
X ZY Z J V
J
Y
X Z J
X
V V
Z ZY ZYJ
ZY
J
X JX
V
ZY
Y
V
XXX L28C
J
V V
X
YYY L28C
V
ZZZ L28C
V
VVV L282DENM
No Modif
JJJ L282DENH
Modif
190.0
Temp (°C)
J
V
Experiment
180.0
170.0
X
Y
J
Z
VJ V
ZYX
Y
Z
YX X
J
V
VX
J VJ VJ
160.0
150.0
TW.115.1230.150
TW.205.1230.150
TW.025.1230.150
tf074
tf074
140.0
0
1.0
2.0
3.0
4.0
5.0
6.0
7.0
Time (s)
Figure 8.13 - Water temperature at height 1.230 m and 0.150 m radial vessel position
WinGraf - 10-29-2003
220.0
XXX
YYY
ZZZ
VVV
210.0
200.0
L28C
L28C
L282DENM
L282DENH
TW.115.1230.330
TW.295.1230.330
tf114
tf114
Z
Z
Temp (°C)
190.0
Modif
180.0
V
170.0
V ZX
Z Y
160.0
YXV
YX
V Z
V
Y
Z
X
Z
YX V
ZV
Z V V Z V YX
X
Y
ZX
YX Y
V
YX
V
Y
X
YX
V
YX
YX
Experiment
Y
VYX VX
Z
Y
Z
ZX Z
No Modif
ZV
ZVX ZVYXZV YX YX
150.0
140.0
0
1.0
2.0
3.0
4.0
5.0
6.0
7.0
Time (s)
Figure 8.14 - Water temperature at height 1.230 m and 0.330 m radial vessel position
147
Assessment of Size Aspects in Modelling Molten Fuel Coolant Interaction
Chapter 8 - COMETA CODE MODIFICATIONS
WinGraf - 10-29-2003
1.00
.90
XXX
YYY
.80
L282DENM voialf062
L282DENH voialf062
Void Fraction (-)
.70
.60
.50
.40
Modified
.30
Y
Non Modified
Y
.20
Y
0
XY XY
X
0
1.0
XY X YX
2.0
YX
XY
XY
3.0
X
X
X
X
.10
Y
Y
XY
XY X Y
4.0
5.0
6.0
7.0
Time (s)
Figure 8.15 - Void fraction in volume 62 at 0.42 m height and 0.150 m radial position
WinGraf - 10-29-2003
1.00
.90
XXX
YYY
.80
L282DENM voialf102
L282DENH voialf102
Void Fraction (-)
.70
Modified
Y
.60
.50
Y
.40
Non Modified
X
.30
Y
Y
.20
X
Y
X
X
.10
0
X
0
XY
1.0
XY X
2.0
YX
YX
XY
XY
3.0
X
Y
Y X YX
4.0
5.0
6.0
7.0
Time (s)
Figure 8.16 - Void fraction in volume 102 at 0.42 m height and 0.330 m radial position
148
Assessment of Size Aspects in Modelling Molten Fuel Coolant Interaction
Chapter 8 - COMETA CODE MODIFICATIONS
WinGraf - 10-29-2003
1.00
.90
XXX
YYY
.80
L282DENM voialf066
L282DENH voialf066
Void Fraction (-)
.70
.60
.50
Modified
.40
Y
Y
.30
Y X
.20
Y
XY
X
0
0
Y XY XY XY
1.0
X
XY X Y
YX
2.0
X
Y
X
Y
X
X
X
.10
Y
Y
Non Modified
X
Y
3.0
4.0
5.0
6.0
7.0
Time (s)
Figure 8.17 - Void fraction in volume 66 at 0.69 m height and 0.150 m radial position
WinGraf - 10-29-2003
1.00
.90
XXX
YYY
.80
L282DENM voialf106
L282DENH voialf106
Void Fraction (-)
.70
Modified
Y
.60
.50
X
.40
Y
.30
Y
.20
X
Y
X
.10
XY XY
0
0
1.0
X
X Y
XY
2.0
X
Y
XY
Y
Y
Y
X
Y
X
X
X
X
Non Modified
Y
3.0
4.0
5.0
6.0
7.0
Time (s)
Figure 8.18 - Void fraction in volume 106 at 0.69 m height and 0.330 m radial position
149
Assessment of Size Aspects in Modelling Molten Fuel Coolant Interaction
Chapter 8 - COMETA CODE MODIFICATIONS
WinGraf - 10-29-2003
1.00
.90
XXX
YYY
.80
L282DENH voialf070
L282DENM voialf070
Modified
Void Fraction (-)
.70
.60
.50
X
.40
Y
XY
Y
YX
.30
YX
0
0
Y
Y XY X
Y X
X
1.0
Y
X
Y
Y
YX Y
.10
X
X
Y
X
X
.20
Y
X
X
X
Y
Non Modified
X
Y
2.0
3.0
4.0
5.0
6.0
7.0
Time (s)
Figure 8.19 - Void fraction in volume 70 at 0.96 m height and 0.150 m radial position
WinGraf - 10-29-2003
1.00
X
.90
XXX
YYY
.80
L282DENH voialf110
L282DENM voialf110
Y
Void Fraction (-)
.70
X
Modified
Y
X
.60
X
Non Modified
.50
.40
X
Y
X
.30
Y
.20
Y
YX
Y
X
X
Y
Y
Y X
YX
XY
.10
0
0
X
Y
Y XY
X
1.0
X
Y
Y
X
2.0
3.0
4.0
5.0
6.0
7.0
Time (s)
Figure 8.20 - Void fraction in volume 110 at 0.96 m height and 0.330 m radial position
150
Assessment of Size Aspects in Modelling Molten Fuel Coolant Interaction
Chapter 8 - COMETA CODE MODIFICATIONS
WinGraf - 10-29-2003
1.00
.90
XXX
YYY
.80
Modified
L282DENM voialf074
L282DENH voialf074
Y
Y
Void Fraction (-)
.70
.60
Y
Y
.50
XY
.40
Y
X
.30
Y
.20
X
Y
Y
Y
X
Y X
Y
X
X
X
Y
X
X
X
Y
Y
X
X
X
Y
X
X
Y
X
Non Modified
.10
XY
0
0
1.0
2.0
3.0
4.0
5.0
6.0
7.0
Time (s)
Figure 8.21 - Void fraction in volume 74 at 1.23 m height and 0.150 m radial position
WinGraf - 10-29-2003
1.00
Y
Y
.90
XXX
YYY
.80
Y
.70
Void Fraction (-)
Y
X
L282DENM voialf114
L282DENH voialf114
Modified
X
Y
X
X
.60
.50
Y
X
Y
.40
Y
X
.30
X
.20
Y
X
Y
Y
X
X
X
X Y
X
Y
Y
Y
Y
X
X
X
Non Modified
.10
0
XY XY
0
1.0
2.0
3.0
4.0
5.0
6.0
7.0
Time (s)
Figure 8.22 - Void fraction in volume 114 at 1.23 m height and 0.150 m radial position
151
Assessment of Size Aspects in Modelling Molten Fuel Coolant Interaction
Chapter 8 - COMETA CODE MODIFICATIONS
WinGraf - 10-29-2003
.70
.60
XXX
YYY
L282DENH vmdrop062
L282DENH vmdrop102
Vol 62
Mass of drops (kg)
.50
Vol 102
.40
Y
.30
X
Y
Y
.20
Y
Y
.10
Y
Y
X
X
X
0
0
1.0
X
X
Y
2.0
3.0
4.0
X
5.0
X
6.0
7.0
Time (s)
Figure 8.23 - Mass of drops distribution in lateral water volumes 62 and 102 at 0.42 m height in modified
COMETA calculation
WinGraf - 10-29-2003
.70
.60
XXX
YYY
Y
L282DENH vmdrop066
L282DENH vmdrop106
Mass of drops (kg)
.50
Vol 66
.40
Vol 106
.30
Y
.20
Y
Y
Y
X
.10
X
X
X
X
Y
X
0
X
XY
Y
0
1.0
2.0
3.0
4.0
5.0
X
Y
6.0
7.0
Time (s)
Figure 8.24 - Mass of drops distribution in lateral water volumes 66 and 106 at 0.69 m height in modified
COMETA calculation
152
Assessment of Size Aspects in Modelling Molten Fuel Coolant Interaction
Chapter 8 - COMETA CODE MODIFICATIONS
WinGraf - 10-29-2003
.70
.60
XXX
YYY
L282DENH vmdrop070
L282DENH vmdrop110
Vol 70
Mass of drops (kg)
.50
Vol 110
.40
X
.30
Y
.20
Y
X
X
Y
X
Y
X
.10
Y
X
Y
X
Y
0
0
1.0
2.0
X
Y
X
X
X
3.0
4.0
Y
5.0
6.0
7.0
Time (s)
Figure 8.25 - Mass of drops distribution in lateral water volumes 70 and 110 at 0.96 m height in modified
COMETA calculation
WinGraf - 10-29-2003
2.00
1.80
Y
XXX
YYY
1.60
L282DENH vmdrop074
L282DENH vmdrop114
Mass of drops (kg)
1.40
Vol 114
1.20
1.00
Vol 74
X
.80
.60
.40
XY
XY
.20
0
0
1.0
X
Y
2.0
Y
X
Y
Y
X
XY
3.0
4.0
Time (s)
Y
5.0
6.0
7.0
Figure 8.26 - Mass of drops distribution in lateral water volumes 74 and 114 at 1.23 m height in modified
COMETA calculation
153
Assessment of Size Aspects in Modelling Molten Fuel Coolant Interaction
Chapter 8 - COMETA CODE MODIFICATIONS
8.4. CONCLUSIONS
COMETA code source has been modified in order to include the radial movement of the melt drops
produced in the interaction when the melt jet penetrates into the water and along its travel to the bottom
vessel through the vessel central section. In the old version of the code drops movement was simulated
restricted only vertically in the central volumes.
This modification has been found to be a good improvement in the prediction and assessment of the code and
essential in some studies where local quantities are critical in the evolution of the transient (as local void
fraction and drops distribution are for vapour explosions studies).
The modified COMETA code has been tested against FARO test L-28 experimental data and compared to
old version. In the results 3 local parameters were presented: temperatures were compared to experimental
values and gave improved results in the prediction. Void fraction and drops distribution were plotted in order
to evidence the differences using the old and the new version of the code.
154
Assessment of Size Aspects in Modelling Molten Fuel Coolant Interaction
OVERALL CONCLUSIONS
OVERALL CONCLUSIONS
In the context of the prediction through computer codes of the progression and consequences of severe
accidents in water cooled reactors the following achievements were reached:
•
The COMETA code was successfully applied for the prediction of FARO tests under conditions not
experimented before: subcooling in test L-29, vapour explosion in test L-33.
In the subcooled test option it was shown the effect on the global void fraction of the selection of
different pressure and /or temperature. The results indicate that the greater influence on the void
fraction is caused by the pressure level (1 to 5 bar), which compressing the non-condensable void
fraction, induces remarkable change in the global void fraction. No major influence is present in the
quenching rate. Also, in subcooled conditions no significant difference resulted from the adoption of
the standard or enhanced fragmentation model.
The code was able to calculate the subcooled conditions without innerving in major stability
problems. It was found a point that needed improvement in the code: the jet parcels description in
order to properly account for velocity differences among them.
The COMETA code was able to calculate the conditions foreseen for the FARO Test L-33. The code
was also applied to simulate the triggering and the response due to a possible explosion. No
propagation was calculated if the H2 generation model was the same as applied for previous saturated
tests. If the amount is reduced to values typical of the last subcooled Test L-31 or even less a
propagation of the explosion seems possible.
•
The code was applied to large scale configurations (ASCÓ NPP reactor) clarifying phenomena
(thermalhydraulic equilibrium) and size effects. All the cases shown decreasing fragmentation and
energy production as the initial water level in the core inlet and lower plenum was reduced. The
relation between the void fraction and the quenching rate was clarified. An equilibrium value is
generally attained and is possible only if increase and decrease of void fraction is allowed; that is if
the mixture level is at or above the injection point (flooded regime). On the contrary, if the mixture
level is too low, the power exchanged will continuously decrease and tend to zero as the water level
decreases (depleted regime).
Problems in scale-up to reactor configurations were identified in the application of COMETA code
and leaded to improve COMETA code lacks. For example COMETA logic related to drops
movement in 2d nodalization was changed and adapted to large reactor configurations. The
implementation of two-dimensional equational model has revealed as absolutely necessary for
reproducing the detailed movement of the drops in particular applications where detailed knowledge
of local phenomena is absolutely needed. The hypotheses used for the improvement implemented in
the COMETA code, formulated in the equations that this Thesis proposes, have resulted useful for
the fitting description of the fragmented melt during the MFCI.
•
Assessment and qualification of the COMETA code: Simulations performed with old versions were
unified with the present code version. Results were improved and consolidated. Guides to perform
simulations were given to advise code users when selecting 1d or 2d geometries.
•
Calculations to understand the RELAP5/SCDAP 3.2 modelling theory in the late damage phase were
performed. They revealed that in order to perform a detailed MFCI study it is better to use a MFCI
thermalhydraulic code in which the interaction grade is result from the initial conditions of the
system, the molten pool conditions and the own interaction transient evolution and it is not an input
parameter set by the code user. RELAP5/SCDAP 3.2 can be used to obtain the final conditions as
initial conditions in a MFCI thermalhydraulic code. In that way RELAP5/SCDAP 3.2 and a detailed
MFCI code can be seen as complementary tools. The maximum time available for the operator to
155
Assessment of Size Aspects in Modelling Molten Fuel Coolant Interaction
OVERALL CONCLUSIONS
restore cooling conditions via the LPIS after a LOCA was acquired from the calculations. Above this
time meltdown could not be avoided. Finally conditions up to the moment of the slumping were
presented in order to select them as initial conditions for a detailed MFCI study.
In general the main objective of the PhD research was achieved expanding the general knowledge in Melt
Fuel Coolant Interaction. The knowledge was complemented collaborating and complementing the
application of COMETA code under conditions not experimented before, developing and improving
COMETA code sources and verifying the code consistency, analysing and unifying the COMETA
simulations carried so far.
A further analytical study was carried out in order to underline the MFCI inside the general overview of a
NPP severe accident progression.
156
Assessment of Size Aspects in Modelling Molten Fuel Coolant Interaction
BIBLIOGRAPHY AND REFERENCES
BIBLIOGRAPHY AND REFERENCES
[1]
A. Annunziato, S. Barsotti - “Assessment and Improvements of the COMETA Condensation Model
in Presence of Non-Condensable Gases” - JRC - Technical Note No.I.94.90. June 1994.
[2]
F. Reventós, J. Sánchez-Baptista, A. Pérez-Navas, P. Moreno - “Assessment and application of
blackout transients at Ascó nuclear power plant with relap5/mod2” - International agreement report
nureg/ia-119-June 1993.
[3]
C. Addabbo, A. Annunziato, R. Silverii, D. Magallon - “COMETA analysis of system pressure
effect in FARO Melt Quenching tests” - Proceedings of the ICONE 6 - 6335 - May 10-14, 1998.
[4]
A. Annunziato, C. Addabbo - “COMETA (Core Melt Thermalhydraulic Analysis) a computer code
for Melt Quenching Analysis” - International Conference on New Trends in Nuclear System
Thermalhydraulics. Pisa, May-June 1994.
[5]
A. Annunziato, C. Addabbo, H. Hohmann, D. Magallon - “COMETA code calculation of FARO
Melt Quenching Tests” - International Conference on New Trends in Nuclear System
Thermalhydraulics. Pisa, May-June 1994.
[6]
A. Annunziato, C. Addabbo - “COMETA (Core Melt Thermalhydraulic Analysis): Code User’s
Manual” - JRC - Special Publication No.I.98.130. October 1998.
[7]
A. Annunziato - “COMETA v0.4 Pre Test Calculation of FARO Base Case Test” - JRC Technical Note No.I.93.114. August 1993.
[8]
S. Bocci, A. Cirigliano, G. Forasassi, G. Fruttuoso, A. Manfredini and F. Oriolo - “Core
coolability and corium retention in a LWR” - Università di Pisa. Dipartimento di Ingegneria
Meccanica, Nucleare e della Produzione.
[9]
A. Annunziato - “The effect of H2 production on the propagation of Energetic Fuel Coolant
Interaction in UO2/ZrO2 Mixtures” - JRC - Technical Note No.I.96.108. June 1996.
[10] A. Annunziato, C. Addabbo - “Effect of uncondensables in Fuel Coolant Interaction and
Quenching” - Proceedings of the ICONE 5 - 2593 - May 26-30, 1997 Nice, France.
[11] G. Albrecht (FZK-D), A. Annunziato (JRC-EC), P. Giordano (IRSN-F), D. Magallon (CEA-F), L.
Meyer (FZK-D), S. Park (RIT-SE), P. Piluso (CEA-F), R. Seghal (RIT-SE), W. Schütz (FZK-D).
“EUROPEAN NETWORK FOR THE REDUCTION OF UNCERTAINTIES IN SEVERE ACCIDENT
SAFETY ISSUES (EURSAFE). ASSESSMENT OF CURRENT PRACTICES FOR PRESERVATION
AND MAINTENANCE OF Severe Accident DATA”. January 2003.
[12] H.J. Allelein, K.Neu, J.P. Van Dorsselaere, K. Müller, P. Kostka, M. Barnak, P. Matejovic, A.
Bujan, J. Slaby. - “European validation of the integral code ASTEC (EVITA)”. - Nuclear
Engineering and Design 221 (2003) 95 - 118.
[13] A. Annunziato - “Evaluation of melt final superheat with Metal Zr mixtures” - Internal JRC Note.
October 1998.
[14] M. SAITO et al. - “Experimental study on penetration behaviors of Water Jet into Freon-11 and
Liquid Nitrogen” - ANS Proc. 1988 Nat.l Heat Transfer Conf., Houston.
[15] A. Annunziato, A. Yerkess, C. Addabbo - “FARO and KROTOS code simulation and analysis at
JRC Ispra” - Nuclear Engineering and Design 189 (1999), 359 - 378.
157
Assessment of Size Aspects in Modelling Molten Fuel Coolant Interaction
BIBLIOGRAPHY AND REFERENCES
[16] A. Annunziato, C. Addabbo - “FARO Base Case Post-Test Analysis by COMETA Code” - Seventh
International Topical Meeting on Nuclear Reactor Thermal Hydraulics (NURETH-7). Saratoga,
1995.
[17] D. Magallon - “FARO LWR Programme. Scoping Test Data Report” - JRC - Technical Note
No.I.92.135. 1992.
[18] M. Dehn, D. Magallon - “FARO LWR Programme Base Case Test L-11 Data Report” - JRC Technical Note No.I.94.147. 1994.
[19] A. Benuzzi, D. Magallon - “FARO LWR Programme Base Case Test L-11 Quick Look Report” JRC - Technical Note No. I.94.55. 1994.
[20] R. Silverii, D. Magallon - “FARO LWR Programme Test L-27 Data Report” - JRC - Technical
Note No.I.99.194. 1999.
[21] R. Arroyo, J. A. Martínez, C. Vicente, P. Villalibre - “Fenomenología de degradación del núcleo.
Comparación con una simulación predictiva para un PWR comercial” - Revista Sociedad Nuclear
Española, January 1991.
[22] M. Alconero, A. I. Alvárez, J. A. Bagües, J. C. García, J. González, A. Lantaron, J. Saiz, O.
Villegas - “Herramientas de análisis e interpretación de fenomenología de degradación del núcleo”
- Revista Sociedad Nuclear Española, January 1991.
[23] A. Franceschini - “Interazione Combustibile Refrigerante in un reattore nucleare ad acqua legera
in condizioni inzidentali severe” - Università degli Studi di Pisa. Facoltà di Ingegneria. Nuclear
Engineering Final Project. 1998 - 1999.
[24] A. Alonso - “La investigación sobre accidentes severos” - Revista Sociedad Nuclear Española,
October 1989.
[25] A. Annunziato, C. Addabbo - “JRC STRESA Web site: http://asa2.jrc.it/stresa”.
[26] “Loss of Coolant Accident by W.L. Riebold”. - Nuclear Reactor Safety Transfer. Ed. Mc. GrawHill. ISBN 0-89116-224-0.
[27] A. Maréchal, A. Annunziato - “Modification of the Corium Physical Properties in the COMETA
Code” - JRC - Technical Note No.I.94.98. July 1994.
[28] NEA Committee on the Safety of Nuclear Installations (CSNI) - Senior Group of Experts on
Nuclear Safety Research Facilities and Programmes (SESAR/FAP) “Nuclear Safety Research in
OECD Countries: Summary Report of Major Facilities and Programmes at Risk” - 2001 Report.
[29] A. Annunziato, C. Addabbo, A. Yerkess, R. Silverii, W. Brewka, G. Leva - “'OECD/CSNI
International Standard Problem 39 on FARO Test L-14 on Fuel Coolant Interaction and Quenching
- Comparison Report, Volume I: Analysis of the Results” - NEA/CSNI/R (97)31. 1997.
[30] “Oecd-loft lp-fp-2 code comparison report vol i: thermal-hydraulic comparisons compiled by:
d.c.mecham, m.merilo t.sirkia, j.j.peña, s.enciso, f. reventós, f.oriolo, e.coryell, s.güntay” - OECDLOFT project report - April 1989.
[31] F. Reventós, J. J. Peña, S. Enciso - “Oecd-loft experiment lp-fp-2 calculation using relap5/mod2”
- International code assessment and application program specialist meeting. Winfrith- UK - January
1987.
158
Assessment of Size Aspects in Modelling Molten Fuel Coolant Interaction
BIBLIOGRAPHY AND REFERENCES
[32] C. Pretel, F. Reventós - “Prediction of unusual occurrence of core dryout in sbloca scenarios of
the Ascó npp” - 6th international conference on nuclear engineering-icone6 - San Diego-CaliforniaUsa - May 1998.
[33] F. Reventós, J. Sánchez-Baptista, A. Pérez-Navas, P. Moreno - “Primary feed&bleed thermalhydraulic analysis for asco nuclear power plant” - Relap5 international users seminar - BostonMassachusetts-Usa - July 1993.
[34] A. Alonso, L. Rebollo, R. Salve - “El proyecto Phebus - CSD en el marco de la investigación
sobre accidentes severos” - Revista Sociedad Nuclear Española, January 1991.
[35] Comité de Dirección del Consorcio - “El proyecto Phebus - CSD y el consorcio Phebus- España”
- Revista Sociedad Nuclear Española, January 1991.
[36] B. Adroguer, C. Gonnier, J. A. Martínez - “El proyecto Phebus. La instalación, la matriz
experimental y el análisis de los resultados” - Revista Sociedad Nuclear Española, January 1991.
[37] A. Annunziato, C. Addabbo, D. Magallon - “FARO Test L-28 Quick Look Report” - JRC Technical Note No.I.99.74. 1999.
[38] The SCDAP/RELAP5 Development Team - “SCDAP/RELAP5/MOD3.2 CODE MANUAL
VOLUME II: DAMAGE PROGRESSION MODEL THEORY” - NUREG/CR-6150 INEL-96/0422
Revision 1 Volume II. October 1997.
[39] The SCDAP/RELAP5 Development Team - “SCDAP/RELAP5/MOD3.2 CODE MANUAL
VOLUME III: USER’S GUIDE AND INPUT MANUAL” - NUREG/CR-6150 INEL-96/0422
Revision 1 Volume III. November 1997.
[40] The SCDAP/RELAP5 Development Team - “SCDAP/RELAP5/MOD3.2 CODE MANUAL
VOLUME V: DEVELOPMENTAL ASSESSMENT” - NUREG/CR-6150 INEL-96/0422 Revision 1
Volume V. October 1997.
[41] Emmanuell Gailliez, Jean-Calude Micaelli - “Séminaire ICARE/CATHARE” - Institut de
Protection et de Sûreté Nucléaire (IPSN).
[42] H.J. Allelein, J. Bestele, K. Neu, F. Jacq, M. Kissane, W. Plumecocq, J.P. Van Dorsselaere. “Severe accident code ASTEC development and validation”.
[43] J. C. Latché - IPSN/DRS, A. Breest, H.J. Allelein - GRS - “Status of the Simulation of Melt
Spreading and Melt Coolability within the Containment”.
[44] M. Epstein, H.K. Fauske - “Steam film instability and the Mixing of the Core-Melt Jets and
Water” - ANS Proc., 1985 Nat.l Heat Transfer Conf., Denver
[45] A. Annunziato, C. Addabbo, W. Brewka - “The STRESA (Storage of Reactor Safety) Database
(Web page: http://asa2.jrc.it/stresa)” - ICONE 9 - 818 - 2001.
[46] J. J. Peña, S. Enciso, F. Reventós - “Thermal-hydraulic post-test analysis of lp-fp-2 experiment” OECD loft project-Spanish consortium final report-November 1987.
[47] J. J. Peña, S. Enciso, F. Reventós - “Thermal-hydraulic post-test analysis of oecd-lp-fp-2
experiment” - International agreement report nureg/ia-0049-April 1992.
159
Assessment of Size Aspects in Modelling Molten Fuel Coolant Interaction
BIBLIOGRAPHY AND REFERENCES
[48] L. Batet; C. Pretel; C. Llopis, A. Cuadra, JL. Gago, F. Reventós - “Thermal-hydraulic code
assessment activities of the technical university of Catalonia ” - 2002 relap5 international users
seminar - Park City-Utah-Usa - September 2002.
[49] F. Reventós, J. Sánchez-Baptista, A. Pérez-Navas, P. Moreno - “Transient analysis for Ascó npp
using relap5/mod2” - Relap5 user's seminar college station-Texas-Usa, February 1989.
[50] F. Reventós, J. Sánchez-Baptista, A. Pérez-Navas, P. Moreno - “Transient analysis in the Ascó npp
using relap5/mod2” - Nuclear Technology volume 90, number 3, June 1990.
[51] H. S. Park, R. Chapman, M. L. Corradini (University of Wisconsin, Madison) - “Vapor Explosions
in a One-Dimensional Large-Scale Geometry with Simulant Melts”. NUREG/CR-6623. October
1999.
[52] M. L. Corradini (University of Wisconsin, Madison) - “Vapor Explosions: A Review of
Experiments for Accident Analysis”. Nuclear Safety, vol. 32.No 3, July-September 1991.
160
Assessment of Size Aspects in Modelling Molten Fuel Coolant Interaction
PUBLICATIONS
PUBLICATIONS
[1] “Characterizing Melt Fuel - Coolant Interaction in a NPP context with RELAP5/SCDAP 3.2 code”.
CD-Rom for the 29th Annual Meeting of the Spanish Nuclear Society, Zaragoza, October 2003. P.
Pla, F. Reventós, A. Annunziato.
[2] “Use of STRESA (Storage of Reactor Safety) Web platform for Remote Codes Input deck
Preparation and Execution”. CD-Rom for the 27th Annual Meeting of the Spanish Nuclear Society,
Valencia, October 2001. P. Pla, A. Annunziato, C. Addabbo.
[3] “COMETA Code Calculations of the FARO Quenching Tests”. JRC EUR 19761 EN, March 2001.
P. Pla, A. Annunziato, C. Addabbo.
[4] “Simulación mediante el código de cálculo COMETA de los experimentos de interacción
combustible fundido-refrigerante realizados en la instalación FARO”. CD-Rom ISSN. 1575-3204 for
the 26th Annual Meeting of the Spanish Nuclear Society, León, October 2000. P. Pla, A. Annunziato,
C. Addabbo.
[5] “COMETA code Analysis of Fuel Coolant Interaction Phenomena in a Reactor Geometry” - Ninth
International Topical Meeting on Nuclear Reactor Thermal Hydraulics (NURETH-9), San Francisco,
California, October 3 - 8, 1999. A. Annunziato, C. Addabbo, P. Pla.
[6] “COMETA v.1 Pre-Test Calculation of FARO Test L-33”. JRC Technical Note No I.99.163,
September 1999. A. Annunziato, P. Pla, C. Addabbo.
[7] “COMETA v.1 Pre-Test Calculation of FARO Test L-29”. JRC Technical Note No I.99.44, March
1999. A. Annunziato, P. Pla, C. Addabbo.
[8] “Analysis of Fuel-Coolant Quenching Phenomena by COMETA code in Reactor Geometry”. JRC
Technical Note No I.99.48, March 1999. P. Pla, A. Annunziato, C. Addabbo.
[9] “Análisis de la fenomenología de interacción combustible fundido-refrigerante mediante el código
COMETA”. CD-Rom ISSN. 1137-2885 for the 24th Annual Meeting of the Spanish Nuclear Society,
Valladolid, October 1998. P. Pla, A. Annunziato.
161
162
Assessment of Size Aspects in Modelling Molten Fuel Coolant Interaction
LIST OF FIGURES
LIST OF FIGURES
Figure 3.1 - Severe accident associated phenomena [From [34]] .......................................................................................8
Figure 5.1 - Outside view of FARO Test facility with FAT release vessel for melt quenching experiments [From [25]]22
Figure 5.2 - The FARO SARCOFAGO test vessel for melt spreading experiments [From [25]] ....................................22
Figure 5.3 - FARO Test facility with TERMOS release vessel [From [17]] ....................................................................25
Figure 5.4 - FARO Test facility with FAT release vessel [From [20]].............................................................................26
Figure 6.1 - Integration volumes [From [6]].....................................................................................................................29
Figure 6.2 - Schematic melt field release description .......................................................................................................33
Figure 6.3 - Jet Break-up Length definition [From [6]]....................................................................................................34
Figure 6.4 - Jet Break-up Length interpolated model [From [6]] .....................................................................................35
Figure 7.1 - Global average void fraction below mixture level (AVGVOIDML), global average void fraction in the
water region (AVGVOID10) and average void fraction of steam in the water region (VOID10) in Test L-29 ......42
Figure 7.2 - Jet, Drops, and Molten Cake mass in the base case calculation in Test L-29 ...............................................42
Figure 7.3 - Jet leading edge in the base case calculation in Test L-29 ............................................................................43
Figure 7.4 - Effect of subcooling and pressure on void fraction of non-condensable and steam in Test L-29 .................44
Figure 7.5 - Effect of subcooling and pressure on void fraction of steam in Test L-29....................................................45
Figure 7.6 - Void fraction of steam and non-condensable in the FAT lowest volume for the 1 bar subcooled case in Test
L-29..........................................................................................................................................................................45
Figure 7.7 - Effect of the H2 production on pressure in Test the L-29 .............................................................................49
Figure 7.8 - Effect of the H2 production on the average void fraction in Test L-29.........................................................49
Figure 7.9 - Void fraction in the 2d calculation in Test L-29 ...........................................................................................50
Figure 7.10 - COMETA 2d nodalization for FARO Test L-33.........................................................................................52
Figure 7.11 - Global void fraction for Test L-33 ..............................................................................................................55
Figure 7.12 - Jet leading edge for Test L-33 .....................................................................................................................55
Figure 7.13 - Current COMETA logic for leading edge behaviour ..................................................................................56
Figure 7.14 - Modified COMETA logic for leading edge behaviour ...............................................................................56
Figure 7.15 - Jet leading edge for Test L-33 with modified COMETA logic...................................................................57
Figure 7.16 - Central void fraction vs. height at 0.5 s for Test L-33.................................................................................57
Figure 7.17 - Lateral void fraction vs. height at 0.5 s for Test L-33 .................................................................................58
Figure 7.18 - Drops in the central volumes vs. height at 0.5 s for Test L-33....................................................................58
Figure 7.19 - Central void fraction vs. height at 3 s for Test L-33....................................................................................59
Figure 7.20 - Lateral void fraction vs. height at 3 s for Test L-33 ....................................................................................59
Figure 7.21 - Drops in the central volumes vs. height at 3 s for Test L-33.......................................................................60
Figure 7.22 - Drops in the lateral volumes vs. height at 3 s for Test L-33........................................................................60
Figure 7.23 - Vapour trigger mass flow shape for Test L-33............................................................................................61
Figure 7.24 - Pressure in the central volumes triggering with vapour mass flow .............................................................61
Figure 7.25 - Water trigger mass flow shape for Test L-33 ..............................................................................................62
Figure 7.26 - Pressure in the central volumes triggering with water mass flow ...............................................................62
Figure 7.27 - Pressure in two central volumes triggering with vapour and water mass flows ..........................................63
Figure 7.28 - Selected water trigger mass flow shape for Test L-33 ................................................................................63
Figure 7.29 - COMETA 2d nodalization and triggering position .....................................................................................64
Figure 7.30 - Pressure behaviour in the central volumes after triggering at 0.5 s in the 4 bar case calculation with 20% of
H2 production ...........................................................................................................................................................65
Figure 7.31 - Pressure behaviour in the lateral volumes after triggering at 0.5 s in the 4 bar case calculation with 20% of
H2 production ...........................................................................................................................................................66
Figure 7.32 - Pressure behaviour in the central volumes after triggering at 3 s in the 4 bar case calculation with 20% of
H2 production ...........................................................................................................................................................66
Figure 7.33 - Pressure behaviour in the lateral volumes after triggering at 3 s in the 4 bar case calculation with 20% of
H2 production ...........................................................................................................................................................67
Figure 7.34 - Initial conditions for extreme 1-d calculations in COMETA reactor calculation........................................70
Figure 7.35 - Initial conditions for 2-d calculations in COMETA reactor calculation .....................................................70
Figure 7.36 - COMETA 1-d nodalization in reactor calculation ......................................................................................71
Figure 7.37 - COMETA 2-d nodalization in reactor calculation ......................................................................................72
Figure 7.38 - Pressure in the upper head at different initial water levels ..........................................................................75
Figure 7.39 - Drops production rate in lower plenum at different initial water levels ......................................................75
Figure 7.40 - Velocity in junctions 19 and 20 in 2 m initial water level case ...................................................................76
Figure 7.41 - Quenching rate at different initial water levels ...........................................................................................76
Figure 7.42 - Mixture level vs. initial water level.............................................................................................................77
163
Assessment of Size Aspects in Modelling Molten Fuel Coolant Interaction
LIST OF FIGURES
Figure 7.43 - Global mean void fraction in lower downcomer, core inlet and lower plenum at different initial water
levels. .......................................................................................................................................................................77
Figure 7.44 - Cake mass ratio at different initial water levels ..........................................................................................78
Figure 7.45 - Quenching rate and cake mass flow per melt mass flow versus initial water level.....................................78
Figure 7.46 - Pressure in the upper head in boundary case...............................................................................................80
Figure 7.47 - Quenching rate in boundary case. ...............................................................................................................80
Figure 7.48 - Drops mass in core inlet in boundary case ..................................................................................................81
Figure 7.49 - Liquid mass flow in junctions 19 and 20 in boundary case in the longer term ...........................................81
Figure 7.50 - Global mean void fraction in lower downcomer, core inlet and lower plenum in boundary case...............82
Figure 7.51 - Void fraction in the upper head (9) and rest primary system (10) in boundary case in the longer term .....82
Figure 7.52 - Pressure in the upper head in “full”, “half full”, “base” and 1 m cases with 2-d nodalization....................84
Figure 7.53 - Quenching rate in “full”, “half full”, “base” and 1 m cases with 2-d nodalization .....................................84
Figure 7.54 - Total drops mass in “full”, “half full”, “base” and 1 m cases with 2-d nodalization...................................85
Figure 7.55 - Global mean void fraction in “full”, “half full”, “base” and 1 m cases with 2-d nodalization....................85
Figure 7.56 - Mean void fraction in the central side of lower plenum and core inlet in “full”, “half full”, “base” and 1 m
cases with 2-d nodalization ......................................................................................................................................86
Figure 7.57 - Drops production rate in lower plenum in “full”, “half full”, “base” and 1 m cases with 2-d nodalization 86
Figure 7.58 - Cake mass ratio in “full”, “half full”, “base” and 1 m cases with 2-d nodalization ....................................87
Figure 7.59 - Resulting void fraction by applying a constant power ................................................................................89
Figure 7.60 - Resulting power exchanged fragmenting melt in a constant void fraction medium....................................89
Figure 7.61 - Power exchanged versus void fraction plan ................................................................................................90
Figure 7.62 - Quenching rate versus global mean void fraction in cases 1-d ...................................................................90
Figure 7.63 - Quenching rate versus global mean void fraction in cases 2-d ...................................................................91
Figure 7.64 - Vessel, LPIS and break RELAP5/SCDAP nodalization ...........................................................................101
Figure 7.65 - Core maximum temperature in all SCDAP calculations ...........................................................................102
Figure 7.66 - Pressure of primary system after break .....................................................................................................104
Figure 7.67 - Temperature in primary system after break...............................................................................................104
Figure 7.68 - Downcomer level after the break ..............................................................................................................105
Figure 7.69 - Break mass flow........................................................................................................................................105
Figure 7.70 - Pressure in the lower plenum at the slumping time in MFCI SCDAP calculations ..................................106
Figure 7.71 - Pressure in the lower plenum in MFCI SCDAP calculations....................................................................106
Figure 7.72 - Void fraction in the lower plenum at the slumping time in MFCI SCDAP calculations ..........................107
Figure 7.73 - Water temperature in the lower plenum at the slumping time in MFCI SCDAP calculations ..................107
Figure 7.74 - Total rate of heat transfer by convection from top surface of debris (W) in MFCI SCDAP calculations.108
Figure 7.75 - Integral with respect to time of total transfer heat from debris and structural material to fluid at boundaries
of debris and structural material (J) in MFCI SCDAP calculations .......................................................................108
Figure 7.76 - Maximum core temperature in LPIS SCDAP calculations .......................................................................110
Figure 7.77 - H2 production rate in LPIS SCDAP calculations ......................................................................................110
Figure 7.78 - H2 accumulated in LPIS SCDAP calculations ..........................................................................................111
Figure 7.79 - Radius of molten pool in LPIS SCDAP calculations ................................................................................111
Figure 7.80 - Total UO2 mass slumped in lower plenum in LPIS SCDAP calculations.................................................112
Figure 7.81 - Total ZrO2 mass slumped in lower plenum in LPIS SCDAP calculations ................................................112
Figure 7.82 - Total Zr mass slumped in lower plenum in LPIS SCDAP calculations ....................................................113
Figure 7.83 - Total Ag mass slumped in lower plenum in LPIS SCDAP calculations ...................................................113
Figure 7.84 - Total Liquid mass in lower plenum in LPIS SCDAP calculations............................................................114
Figure 7.85 - Mass distribution in lower plenum in case LPIS_1150.............................................................................114
Figure 7.86 - Mass distribution in lower plenum in case LPIS_1200.............................................................................115
Figure 7.87 - Mass distribution in lower plenum in case LPIS_1300.............................................................................115
Figure 7.88 - Mass distribution in lower plenum in case LPIS_1500.............................................................................116
Figure 7.89 - Mass distribution in lower plenum in case LPIS_2000.............................................................................116
Figure 7.90 - Mass distribution in lower plenum in case DEBRIS_NOBK....................................................................117
Figure 7.91 - Downcomer level before slumping in LPIS SCDAP calculations ............................................................117
Figure 7.92 - Downcomer level after slumping in LPIS SCDAP calculations ...............................................................118
Figure 7.93 - Void fraction in the lower plenum in LPIS SCDAP calculations .............................................................118
Figure 7.94 - Pressure in the lower plenum in LPIS SCDAP calculations .....................................................................119
Figure 7.95 - Water temperature in the lower plenum in LPIS SCDAP calculations .....................................................119
Figure 7.96 - Maximum core temperature in LPIS SCDAP calculations .......................................................................121
Figure 7.97 - Maximum core temperature in LPIS_1120 case .......................................................................................121
Figure 7.98 - Local H2 production rate in case LPIS_1120 ............................................................................................122
Figure 7.99 - Damage state in components 5 and 7 in LPIS_1120 case .........................................................................122
Figure 7.100 - Maximum core temperature in LPIS_1121 case .....................................................................................123
Figure 7.101 - Damage state in components 7 and 9 in LPIS_1121 case .......................................................................123
164
Assessment of Size Aspects in Modelling Molten Fuel Coolant Interaction
LIST OF FIGURES
Figure 7.102 - H2 production rate in LPIS SCDAP calculations ....................................................................................124
Figure 7.103 - Radius of molten pool in LPIS SCDAP calculations ..............................................................................124
Figure 7.104 - Total UO2 mass slumped in lower plenum in LPIS SCDAP calculations...............................................125
Figure 7.105 - Total ZrO2 mass slumped in lower plenum in LPIS SCDAP calculations ..............................................125
Figure 7.106 - Total Zr mass slumped in lower plenum in LPIS SCDAP calculations ..................................................126
Figure 7.107 - Total Ag mass slumped in lower plenum in LPIS SCDAP calculations .................................................126
Figure 7.108 - Total Liquid mass in lower plenum in LPIS SCDAP calculations..........................................................127
Figure 7.109 - Downcomer level in LPIS SCDAP calculations .....................................................................................127
Figure 7.110 - Void fraction in the lower plenum in LPIS SCDAP calculations............................................................128
Figure 7.111 - Pressure in the lower plenum in LPIS SCDAP calculations ...................................................................128
Figure 7.112 - Water temperature in the lower plenum in LPIS SCDAP calculations ...................................................129
Figure 7.113 - Maximum core temperature in LPIS SCDAP calculations .....................................................................130
Figure 7.114 - Damage state in components 1, 3, 5 and 7 in LPIS_1119 case ...............................................................130
Figure 7.115 - H2 production rate in LPIS SCDAP calculations ....................................................................................131
Figure 7.116 - H2 accumulated in LPIS SCDAP calculations ........................................................................................131
Figure 7.117 - Molten pool volume up to the slumping time in all SCDAP calculations...............................................133
Figure 7.118 - Molten pool liquid phase up to the slumping time in all SCDAP calculations .......................................133
Figure 7.119 - Molten pool UO2 mass up to the slumping time in all SCDAP calculations...........................................134
Figure 8.1 - COMETA nodalization 1d for FARO Test L-29.........................................................................................136
Figure 8.2 - Drops movement in thermalhydraulic volumes in 1d and 2d nodalization in previous COMETA logic....137
Figure 8.3 - Drops movement in thermalhydraulic volumes in 2d nodalization in present COMETA logic..................137
Figure 8.4 - Limit coordinates of the jvol volume and drop coordinates........................................................................138
Figure 8.5 - COMETA 2d nodalization for FARO test L-28..........................................................................................141
Figure 8.6 - Pressure behaviour in COMETA calculations for FARO test L-28 ............................................................142
Figure 8.7 - Water temperature at height 0.420 m and 0.150 m radial vessel position ...................................................144
Figure 8.8 - Water temperature at height 0.420 m and 0.330 m radial vessel position ...................................................144
Figure 8.9 - Water temperature at height 0.690 m and 0.150 m radial vessel position ...................................................145
Figure 8.10 - Water temperature at height 0.690 m and 0.330 m radial vessel position .................................................145
Figure 8.11 - Water temperature at height 0.960 m and 0.150 m radial vessel position .................................................146
Figure 8.12 - Water temperature at height 0.960 m and 0.330 m radial vessel position .................................................146
Figure 8.13 - Water temperature at height 1.230 m and 0.150 m radial vessel position .................................................147
Figure 8.14 - Water temperature at height 1.230 m and 0.330 m radial vessel position .................................................147
Figure 8.15 - Void fraction in volume 62 at 0.42 m height and 0.150 m radial position ................................................148
Figure 8.16 - Void fraction in volume 102 at 0.42 m height and 0.330 m radial position ..............................................148
Figure 8.17 - Void fraction in volume 66 at 0.69 m height and 0.150 m radial position ................................................149
Figure 8.18 - Void fraction in volume 106 at 0.69 m height and 0.330 m radial position ..............................................149
Figure 8.19 - Void fraction in volume 70 at 0.96 m height and 0.150 m radial position ................................................150
Figure 8.20 - Void fraction in volume 110 at 0.96 m height and 0.330 m radial position ..............................................150
Figure 8.21 - Void fraction in volume 74 at 1.23 m height and 0.150 m radial position ................................................151
Figure 8.22 - Void fraction in volume 114 at 1.23 m height and 0.150 m radial position ..............................................151
Figure 8.23 - Mass of drops distribution in lateral water volumes 62 and 102 at 0.42 m height in modified COMETA
calculation ..............................................................................................................................................................152
Figure 8.24 - Mass of drops distribution in lateral water volumes 66 and 106 at 0.69 m height in modified COMETA
calculation ..............................................................................................................................................................152
Figure 8.25 - Mass of drops distribution in lateral water volumes 70 and 110 at 0.96 m height in modified COMETA
calculation ..............................................................................................................................................................153
Figure 8.26 - Mass of drops distribution in lateral water volumes 74 and 114 at 1.23 m height in modified COMETA
calculation ..............................................................................................................................................................153
165
Assessment of Size Aspects in Modelling Molten Fuel Coolant Interaction
LIST OF FIGURES
LIST OF TABLES
Table 7.1 - Performed L-29 pre-test subcooled calculations at 5 bar, 50°C .....................................................................41
Table 7.2 - Performed L-29 pre-test subcooled calculations.............................................................................................41
Table 7.3 - Performed L-29 pre-test saturated calculations ..............................................................................................46
Table 7.4 - Performed L-29 pre-test base case calculations with standard and fragmentation model ..............................47
Table 7.5 - Performed L-29 pre-test base calculation using model with and without H2 production ...............................48
Table 7.6 - Performed pre-test calculations for Test L-33 ................................................................................................53
Table 7.7 - Experimental conditions in the series of experiments in the FARO facility...................................................93
Table 7.8 - Experiment calculations with COMETA code ...............................................................................................94
Table 7.9 - Assessment summary of the simulation results ..............................................................................................95
Table 7.10 - Performed RELAP5/SCDAP calculations..................................................................................................101
Table 7.11 - Initial conditions for a MFCI detailed calculation at the moment of slumping ..........................................132
166
Fly UP