TE-1913 Web

IAEA-TECDOC-1913
IAEA TECDOC SERIES
IAEA-TECDOC-1913
Modelling of Fuel Behaviour

in Design Basis Accidents and
Design Extension Conditions
Proceedings of a Technical Meeting
@
IAEA SAFETY STANDARDS AND RELATED PUBLICATIONS
IAEA SAFETY STANDARDS

Under the terms of Article III of its Statute, the IAEA is authorized to establish or adopt
standards of safety for protection of health and minimization of danger to life and property, and
to provide for the application of these standards.
The publications by means of which the IAEA establishes standards are issued in the
IAEA Safety Standards Series. This series covers nuclear safety, radiation safety, transport
safety and waste safety. The publication categories in the series are Safety Fundamentals,
Safety Requirements and Safety Guides.
Information on the IAEA’s safety standards programme is available on the IAEA Internet
site
http://www-ns.iaea.org/standards/
The site provides the texts in English of published and draft safety standards. The texts
of safety standards issued in Arabic, Chinese, French, Russian and Spanish, the IAEA Safety
Glossary and a status report for safety standards under development are also available. For
further information, please contact the IAEA at: Vienna International Centre, PO Box 100,
1400 Vienna, Austria.
All users of IAEA safety standards are invited to inform the IAEA of experience in their
use (e.g. as a basis for national regulations, for safety reviews and for training courses) for the
purpose of ensuring that they continue to meet users’ needs. Information may be provided via
the IAEA Internet site or by post, as above, or by email to Oﬃ[email protected].
RELATED PUBLICATIONS
The IAEA provides for the application of the standards and, under the terms of Articles III
and VIII.C of its Statute, makes available and fosters the exchange of information relating
to peaceful nuclear activities and serves as an intermediary among its Member States for this
purpose.
Reports on safety in nuclear activities are issued as Safety Reports, which provide
practical examples and detailed methods that can be used in support of the safety standards.
Other safety related IAEA publications are issued as Emergency Preparedness and
Response publications, Radiological Assessment Reports, the International Nuclear Safety
Group’s INSAG Reports, Technical Reports and TECDOCs. The IAEA also issues reports
on radiological accidents, training manuals and practical manuals, and other special safety
related publications.
Security related publications are issued in the IAEA Nuclear Security Series.
The IAEA Nuclear Energy Series comprises informational publications to encourage
and assist research on, and the development and practical application of, nuclear energy for
peaceful purposes. It includes reports and guides on the status of and advances in technology,
and on experience, good practices and practical examples in the areas of nuclear power, the
nuclear fuel cycle, radioactive waste management and decommissioning.
MODELLING OF FUEL BEHAVIOUR
IN DESIGN BASIS ACCIDENTS
AND DESIGN EXTENSION CONDITIONS
The following States are Members of the International Atomic Energy Agency:
AFGHANISTAN GERMANY PAKISTAN

ALBANIA GHANA PALAU
ALGERIA GREECE PANAMA
ANGOLA GRENADA PAPUA NEW GUINEA
ANTIGUA AND BARBUDA GUATEMALA PARAGUAY
ARGENTINA GUYANA PERU
ARMENIA HAITI PHILIPPINES
AUSTRALIA HOLY SEE POLAND
AUSTRIA HONDURAS PORTUGAL
AZERBAIJAN HUNGARY
QATAR
BAHAMAS ICELAND
REPUBLIC OF MOLDOVA
BAHRAIN INDIA
BANGLADESH INDONESIA ROMANIA
BARBADOS IRAN, ISLAMIC REPUBLIC OF RUSSIAN FEDERATION
BELARUS IRAQ RWANDA
BELGIUM IRELAND SAINT LUCIA
BELIZE ISRAEL SAINT VINCENT AND
BENIN ITALY THE GRENADINES
BOLIVIA, PLURINATIONAL JAMAICA SAN MARINO
STATE OF JAPAN SAUDI ARABIA
BOSNIA AND HERZEGOVINA JORDAN SENEGAL
BOTSWANA KAZAKHSTAN SERBIA
BRAZIL KENYA SEYCHELLES
BRUNEI DARUSSALAM KOREA, REPUBLIC OF SIERRA LEONE
BULGARIA KUWAIT SINGAPORE
BURKINA FASO KYRGYZSTAN SLOVAKIA
BURUNDI LAO PEOPLE’S DEMOCRATIC SLOVENIA
CAMBODIA REPUBLIC SOUTH AFRICA
CAMEROON LATVIA SPAIN
CANADA LEBANON
SRI LANKA
CENTRAL AFRICAN LESOTHO
SUDAN
REPUBLIC LIBERIA
CHAD LIBYA SWEDEN
CHILE LIECHTENSTEIN SWITZERLAND
CHINA LITHUANIA SYRIAN ARAB REPUBLIC
COLOMBIA LUXEMBOURG TAJIKISTAN
CONGO MADAGASCAR THAILAND
COSTA RICA MALAWI TOGO
CÔTE D’IVOIRE MALAYSIA TRINIDAD AND TOBAGO
CROATIA MALI TUNISIA
CUBA MALTA TURKEY
CYPRUS MARSHALL ISLANDS TURKMENISTAN
CZECH REPUBLIC MAURITANIA UGANDA
DEMOCRATIC REPUBLIC MAURITIUS UKRAINE
OF THE CONGO MEXICO UNITED ARAB EMIRATES
DENMARK MONACO UNITED KINGDOM OF
DJIBOUTI MONGOLIA GREAT BRITAIN AND
DOMINICA MONTENEGRO NORTHERN IRELAND
DOMINICAN REPUBLIC MOROCCO UNITED REPUBLIC
ECUADOR MOZAMBIQUE
OF TANZANIA
EGYPT MYANMAR
UNITED STATES OF AMERICA
EL SALVADOR NAMIBIA
ERITREA NEPAL URUGUAY
ESTONIA NETHERLANDS UZBEKISTAN
ESWATINI NEW ZEALAND VANUATU
ETHIOPIA NICARAGUA VENEZUELA, BOLIVARIAN
FIJI NIGER REPUBLIC OF
FINLAND NIGERIA VIET NAM
FRANCE NORTH MACEDONIA YEMEN
GABON NORWAY ZAMBIA
GEORGIA OMAN ZIMBABWE
The Agency’s Statute was approved on 23 October 1956 by the Conference on the Statute of the
IAEA held at United Nations Headquarters, New York; it entered into force on 29 July 1957.
The Headquarters of the Agency are situated in Vienna. Its principal objective is “to accelerate and enlarge
the contribution of atomic energy to peace, health and prosperity throughout the world’’.
IAEA-TECDOC-1913
MODELLING OF FUEL BEHAVIOUR

IN DESIGN BASIS ACCIDENTS
AND DESIGN EXTENSION CONDITIONS
PROCEEDINGS OF A TECHNICAL MEETING
INTERNATIONAL ATOMIC ENERGY AGENCY

VIENNA, 2020
COPYRIGHT NOTICE
All IAEA scientific and technical publications are protected by the terms of
the Universal Copyright Convention as adopted in 1952 (Berne) and as revised
in 1972 (Paris). The copyright has since been extended by the World Intellectual
Property Organization (Geneva) to include electronic and virtual intellectual
property. Permission to use whole or parts of texts contained in IAEA publications
in printed or electronic form must be obtained and is usually subject to royalty
agreements. Proposals for non-commercial reproductions and translations are
welcomed and considered on a case-by-case basis. Enquiries should be addressed
to the IAEA Publishing Section at:
Marketing and Sales Unit, Publishing Section

International Atomic Energy Agency
Vienna International Centre
PO Box 100
1400 Vienna, Austria
fax: +43 1 26007 22529
tel.: +43 1 2600 22417
email: [email protected]
www.iaea.org/publications
For further information on this publication, please contact:
Nuclear Fuel Cycle and Materials Section

Vienna International Centre
PO Box 100
1400 Vienna, Austria
Email: [email protected]
© IAEA, 2020
Printed by the IAEA in Austria
June 2020
IAEA Library Cataloguing in Publication Data
Names: International Atomic Energy Agency.

Title: Modelling of fuel behaviour in design basis accidents and design extension conditions /
International Atomic Energy Agency.
Description: Vienna : International Atomic Energy Agency, 2020. | Series: IAEA TECDOC
series, ISSN 1011–4289 ; no. 1913 | Includes bibliographical references.
Identifiers: IAEAL 20-01317 | ISBN 978–92–0–108020–2 (paperback : alk. paper) |
ISBN 978–92–0–108120–9 (pdf)
Subjects: LCSH: Nuclear fuels. | Nuclear reactors – Computer programs. | Nuclear reactors
— Safety measures. | Computer simulation.
FOREWORD
Increased demands on fuel performance, including transient regimes, higher discharge

burnup and longer fuel cycles, have resulted in an increase of loads on fuel and core internals.
To satisfy these fuel performance demands while ensuring compliance with safety criteria, new
national and international programmes have been launched and advanced modelling codes are
being developed. The need for adequate analysis of all aspects of fuel performance to prevent
a failure, and to predict fuel behaviour in accident conditions, was evidenced by the accident at
the Fukushima Daiichi nuclear power plant.
In response to a recommendation by the Technical Working Group on Fuel Performance
and Technology, the IAEA organized a Technical Meeting on Modelling of Fuel Behaviour in
Design Basis Accidents and Design Extension Conditions. The meeting, hosted by the China
Nuclear Power Technology Research Institute of the China General Nuclear Power Group, in
Shenzhen, China, on 13–16 May 2019, was the tenth in the series of IAEA meetings on the
topic of fuel performance in accident conditions. The previous meetings were held in 1980
(jointly with OECD Nuclear Energy Agency, in Helsinki), 1983 (Risø, Denmark), 1986
(Vienna), 1988 (Preston, United Kingdom), 1992 (Pembroke, Canada), 1995 (Dimitrovgrad,
Russian Federation), 2001 (Halden, Norway), 2011 (Mito, Japan) and 2013 (Chengdu, China).
The two most recent meetings focused on modelling of fuel behaviour and modelling under
design basis accidents and design extension conditions, reflecting the importance of the subject
and the urgent need to accelerate the development and verification of accident related fuel
behaviour codes.
The current meeting provided a forum for international experts to review the state of the
art of code development for modelling fuel performance of nuclear fuel for water cooled
reactors with regard to steady state and transient conditions, for design basis and early phases
of severe accidents, including experimental support for code validation for traditional and
advanced technology fuels. The meeting was attended by 26 specialists in fuel performance
modelling and experimental support from 14 Member States and the Joint Research Centre.
Twenty papers were presented at five sessions that covered all topics of the meeting subject.
Meeting participants discussed in detail two recently completed coordinated research projects,
one on Fuel Modelling in Accident Conditions (FUMAC), held from 2014 to 2018, and one on
Analysis of Options and Experimental Examination of Fuels for Water Cooled Reactors with
Increased Accident Tolerance (ACTOF), held from 2015 to 2019. A round table discussion on
the information exchange in R&D on advanced technology materials helped to define a scope
and to outline a new CRP entitled Testing and Simulation for Advanced Technology and
Accident Tolerant Fuels (ATF-TS).
The IAEA wishes to thank all participants for their contributions, with special gratitude to
the CNPRI hosts for the excellent organization of the meeting including the technical visit to
the CNPRI testing facilities. The IAEA officer responsible for this publication was
M. Veshchunov of the Division of Nuclear Fuel Cycle and Waste Technology.
EDITORIAL NOTE
This publication has been prepared from the original material as submitted by the contributors and has not been edited by the editorial
staff of the IAEA. The views expressed remain the responsibility of the contributors and do not necessarily represent the views of the
IAEA or its Member States.
Neither the IAEA nor its Member States assume any responsibility for consequences which may arise from the use of this publication.
This publication does not address questions of responsibility, legal or otherwise, for acts or omissions on the part of any person.
The use of particular designations of countries or territories does not imply any judgement by the publisher, the IAEA, as to the legal
status of such countries or territories, of their authorities and institutions or of the delimitation of their boundaries.
The mention of names of specific companies or products (whether or not indicated as registered) does not imply any intention to
infringe proprietary rights, nor should it be construed as an endorsement or recommendation on the part of the IAEA.
The authors are responsible for having obtained the necessary permission for the IAEA to reproduce, translate or use material from
sources already protected by copyrights.
The IAEA has no responsibility for the persistence or accuracy of URLs for external or third party Internet web sites referred to in this
publication and does not guarantee that any content on such web sites is, or will remain, accurate or appropriate.
CONTENTS
SUMMARY ............................................................................................................................................................. 1
MAIN RESULTS OF THE CRP FUMAC (Session 1)
Simulation tools incorporated to DIONISIO in the frame of the FUMAC project ................................................ 15
M. Lemes, A. Soba, A. Denis
Qualification of FRAPCON/FRAPTRAN for LOCA fuel behaviour modelling and
safety evaluation ........................................................................................................................................... 29
J. Zhang, A. Dethioux, T. Drieu, C. Schneidesch
Fuel behaviour modelling in accident conditions in ALCYONE fuel performance code ...................................... 40
A. Bouloré, C. Struzik, P. Goldbronn, I. Guenot-Delahaie, J. Sercombe
Qualification of transuranus models for mixed core fuel based on the FUMAC outcome .................................... 50
M. Ieremenko, I. Ovdiienko
Main results of FUMAC from CNPRI and sensitivity
and uncertainty study of performance of ATF under transient condition ..................................................... 59
Y.R. Xie, Q.S. Ren, Y.D. Zhang
EXPERIMENTS ON FUEL BEHAVIOUR IN ACCIDENTAL CONDITIONS (Session 2)
Experimental and modelling results of the QUENCH-19 bundle tests with FeCrAl claddings ............................. 71
J. Stuckert, T. Hollands, K. Dolganov
Oxidation performance and failure behaviour of monolithic and coated
ATF claddings under severe accident conditions .......................................................................................... 86
C. Tang, M. Steinbrueck, M. Grosse
FUEL PERFORMANCE CODES APPLIED TO DESIGN BASIS ACCIDENTS (DBA) AND TO

DESIGN EXTENSION CONDITIONS (DEC) (Session 3)
Assessment of high conductivity ceramic fuel concept under normal and accident conditions ............................. 95
D.S. Gomes, A. Abe, A.T. Silva, R.O.R. Muniz, C. Giovedi, M.R. Martins
Modelling of accident scenarios with the FINIX fuel behaviour module ............................................................ 102
H. Loukusa, J. Peltonen, V. Tulkki
Safety analysis of fuel behaviour for implementation licensing .......................................................................... 111
H. Papp, K. Kulacsy
Fuel behaviour during DBA and DEC: analytical and experimental studies for
development of ATF ................................................................................................................................... 115
A.J. Gaikwad, A. K. Deo, P. Sharma, K. Obaidurrahman, S. Bera
An introduction to a multidimensional reactor simulation environment with
the capability of advanced modelling of nuclear fuel under accident conditions........................................ 124
R. Vadi, K. Sepanloo
Simulation of IFA650.5 experiment using the fully coupled SPACE-FRAPTRAN
code system ................................................................................................................................................. 136
H.C. Kim, S.W. Lee, K.D. Kim
Thermo mechanical performance of TRISO coated fuel particle used in
high temperature Gas Cooled Reactor ........................................................................................................ 145
S. Liu, Y. Zhou, P. Chen, W. Li, S. Gao, H. Pang, C. Tang, X. Qiu
3
MAIN RESULTS OF THE CRP ACTOF (Session 4)
Reactivity Initiated Accident assessment for ATF cladding materials................................................................. 155

C. Giovedi, M.R. Martins, A. Abe, R. Reis, A.T. Silva
Round Robin exercise of the candidate ATF cladding materials within the ACTOF project .............................. 162
M. Ševeček, J. Krejčí, A. Chalupová, J. Kabátová, F. Manoch, J. Kočí, L. Cvrček,
C. Tang, M. Steinbrück, M. Grosse, S. Penttilä, J.-M. Autio, J. Lydman,
A. Toivonen, B. Sartowska, W. Starosta, L. Waliś, A. Krześniak, E. Pańczyk,
P. Kalbarczyk, K. Kulisa, J. Smolik, J. Mizeracki, Z. Hózer, T. Novotny,
E. Perezné Feró, A. Pintér Csordás, M. Horváth, P. Szabó, C. Giovedi, P. Xu
Introduction of ATF development and main results of the CRP on ACTOF of KAERI ..................................... 174
C.H. Shin, H.C. Kim, Y.S. Yang
MODELLING OF TRADITIONAL UO2 AND ADVANCED ATF FUELS BEHAVIOUR IN INTEGRAL

SEVERE ACCIDENT (SA) CODES (Session 5)
Status of ATF modelling and application with ATHLET-CD ............................................................................. 183

T. Hollands, L. Lovasz, C. Bals
Advanced LOCA safety criteria: implementation in mechanistic fuel
performance codes and applications ........................................................................................................... 191
M. Veshchunov, A. Boldyrev, V. Shestak
List of Participants ............................................................................................................................................... 207
4
SUMMARY
1. INTRODUCTION
The Technical Meeting (TM) on Modelling of Fuel Behaviour in Design Basis Accidents and Design
Extension Conditions, was held on May 13-16, 2019, in Shenzhen, China. It was organised by the IAEA and
hosted by China Nuclear Power Technology Research Institute (CNPRI), a company of the China General Nuclear
Power Group (CGN).
The objectives were:
— To review Member States’ capabilities in modelling, predicting and improving their understanding of the
behaviour of nuclear fuel under accident conditions;
— To report on the main results and outcomes of the coordinated research projects (CRPs) on Fuel Modelling
in Accident Conditions (FUMAC) and Analysis of Options and Experimental Examination of Fuels with
Increased Accident Tolerance (ACTOF);
— To discuss a proposed new CRP on Testing and Simulation of Advanced Technology Fuels (ATF-TS)
(planned for 2020-2023).
Twenty-six (26) participants from 14 Member States and 2 International Organizations (IAEA and EC JRC)
attended the meeting. The TM was divided in 5 technical sessions:
— Main results of the CRP FUMAC (Chair: Mr. J. Zhang, Belgium);
— Experiments on fuel behaviour in accidental conditions (Chair: Mr. M. Ševeček, Czech Republic);
— Fuel performance codes applied to Design Basis Accidents (DBA) and to Design Extension Conditions
(DEC) (Chair: Mr. A. Boulore, France);
— Main results of the CRP ACTOF (Chair: Ms. T. Liu, China);
— Modelling of traditional UO2 and advanced ATF fuels behaviour in integral Severe Accident (SA) codes
(Chair: Mr. J. Stuckert, Germany).
A total of 20 papers were presented and discussed at the meeting. The main conclusions and
recommendations were summarized by the session chairs as follows.
2. SUMMARY OF THE TECHNICAL SESSIONS
2.1. Technical Session 1: Main results of the CRP FUMAC
The first session was chaired by Mr. Jinzhao Zhang, Tractebel (Engie), Belgium.
2.1.1.Overview
This session provided an overview of the recently completed the IAEA Coordinated Research Project on
Fuel Modelling in Accident Conditions (CRP FUMAC). Within the FUMAC project, the following achievements
were made:
— Verified experimental data set on fuel characteristics in accident conditions (in particular LOCA) and design
extension conditions (DEC), supporting fuel rod codes development and validation for potential extension
of the IFPE database;
— Better predictive capacities of fuel modelling codes: improved models, material properties and codes for the
simulation of nuclear fuel under DBA and severe accident conditions, statistical uncertainty and sensitivity
analysis method;
— Extended collaboration between some Member States organisations (examples: INL–POLIMI–JRC
collaboration agreement; IFE–IBRAE collaboration and joint presentation at the Enlarged Halden
Programme Group Meeting, 2017) beyond the FUMAC project timeframe.
Many participants have acknowledged the IAEA for organising this CRP, which provided an ideal platform
to compare their code results with others and especially with experimental data, to which they otherwise would
not have had access.
1
All experimental data analysed reveal a relatively large spread of the measured burst strains, which is not
only determined by the local conditions of temperature and pressure, but also by heterogeneities and composition
or micro-structural variations in the materials under investigations. An uncertainty analysis of the experimental
data was therefore recommended, including those data that have been used for the development and validation of
the codes applied.
It has been reiterated that this CRP triggered new collaborations, leading for example to the common
development of improved models (e.g. fission gas release model, large strain deformation model, axial fuel
relocation model), in particular for those that are used by a larger user group (e.g. FRAPTRAN,
TRANSURANUS). It has also enabled to point out differences in the interpretation of some experiments and
therefore in the use of the codes, the so-called “user effect”.
The continuous participation of the organisations providing the experimental data was also very instrumental
in clarifying various questions raised during the project.
2.1.2.Summary of presentations
A total of six papers were presented in this session.

The first presentation on overview of the CRP FUMAC described the organization and structure of the
Project was done by Mr. M. Veshchunov from the IAEA. The CRP was initiated in light of the Fukushima accident
under the umbrella of the IAEA Action Plan on Nuclear Safety, following a recommendation of the 2012 meeting
of the IAEA Technical Working Group on Fuel Performance and Technology (TWG FPT) to launch a new fuel
modelling CRP focusing on accident conditions. The CRP FUMAC was launched in 2014 and ended in 2018. The
objectives of the CRP were:
— Analysis and better understanding of fuel behaviour in accident conditions, with a focus on LOCA (DBA),
in line with the early stage of the scenario of the Fukushima accident (BDBA);
— Collection of well checked results of accident simulation experiments, and dissemination of experience in
the organizations of Member States;
— Identification of best practices in the application of physical models and computer codes used in different
Member States for modelling of fuel in accident conditions, and enhancement of predictive capacities of
these models and codes.
Twenty-six (26) organizations from 18 Member States participated in the CRP. Selected sets of accident
simulation experimental results were provided by the CRP participants and will be integrated into the International
Fuel Performance Experiments (IFPE) Database (developed in close co-operation and co-ordination between
OECD/NEA, the IAEA and the IFE/OECD/Halden Reactor Project). The codes used for benchmarking ranged
from fuel performance codes (ALCYONE, BISON, DIONISIO, FRAPTRAN, FTPAC, RAPTA,
TRANSURANUS) to system or severe accident codes (ATHLET-CD, MARS, MELCOR, SOCRAT). An
uncertainty and sensitivity analysis was also performed the on Halden LOCA test IFA-650.10. The results were
compiled and compared with the available experimental data.
The second paper on the simulation tools incorporated to DIONISIO was presented by Mr. M. Lemes from
ANEA, Argentina. Several models related to the cladding response in high temperature range: oxidation, hydrogen
uptake and release, ballooning and burst, were incorporated to the DIONISIO code within the frame of the IAEA
CRP FUMAC) to simulate accidental conditions, in particular the so-called LOCA. A specially designed thermal
hydraulic subroutine provides a simplified description of the rod environment in normal or accidental conditions.
The heat transfer coefficients corresponding to the different coolant regimes are activated as the corresponding
conditions are met. Simulation of different experiments has shown that, despite its simplicity this subroutine gives
adequate predictions of the conditions in the channel, not far from those given by specific thermal hydraulic codes.
With its incorporation, the description of the fuel rod atmosphere, which may consist of single (liquid or vapor) or
double phases, provides quite realistic boundary conditions for the simulation of the fuel rod behavior, without
requiring the intervention of external specific thermal hydraulic codes. The calculation scheme divides the rod
length into a user-defined number. The physical and chemical parameters in one representative pellet subjected to
the local conditions of the segment are solved, with the synchronous work of all the subroutines. Then, a
description of the whole rod is obtained by coupling all the segments. This strategy has yielded accurate
simulations of a wide variety of cases, either in normal or LOCA type conditions. Exhaustive comparisons were
carried out with several thermal hydraulic codes (COBRA-IV, RELAP5-Mod3.1, SOCRAT, ATHLET-Mod 1.1)
2
and with a number of experiments like those of the IFA–650 series (-1, -2,- 9, -10, -11), PUZRY, QUENCH-L0/L1
(for which a new working scheme was specially developed in DIONISIO), CORA-15, IAEA-SPE-4, among others.
The third paper on qualification of FRAPCON/FRAPTRAN for LOCA fuel behaviour modelling and safety
evaluation was presented by Mr. J. Zhang, from Tractebel (Engie), Belgium. He presented the model
improvements to FRAPTRAN-TE-1.5 code, such as the Quantum Technologies’ axial relocation model and errors
corrections in the adapted FRAPTRAN-TE-1.5 version. The impacts of the model improvements together with the
improvement in the thermal hydraulic modelling by using the imposed thermal hydraulic boundary conditions
from SOCRAT calculations and in the thermal boundary conditions (axial power profile, plenum temperature) on
the calculation results for the selected Halden LOCA tests IFA-650.9 and 10 were identified and discussed. In
addition, the statistical uncertainty and sensitivity analysis has been performed on the FRAPTRAN-TE-1.5
modelling of the selected Halden LOCA test IFA-650.10, which helped the identification of significant input
parameters for LOCA fuel behaviour modelling. The final objective was to apply the qualified fuel rod transient
analysis codes FRAPCON/FRAPTRAN to develop an efficient methodology for assessing the performance and
quantifying the margins for advanced technology fuel designs under design basis accident conditions (in particular
LOCAs). The perspectives for further model improvements and benchmarks, as well as the LOCA fuel safety
evaluation methodology development and applications were also discussed.
The fourth paper, fuel behaviour modelling in accident condition in ALCYONE fuel performance code, was
presented by Mr. A. Boulore (CEA, France). He presented the CEA development of the ALCYONE fuel
performance code for PWR fuel in the PLEIADES software environment. It is dedicated to normal, off-normal
and accident conditions such as RIA and LOCA. CEA’s participation to the IAEA FUMAC CRP led to an
improvement of the fuel modelling in LOCA conditions. Specific developments of the fuel performance code for
the LOCA conditions have been done regarding cladding behaviour modelling, fission gas release and stress
evaluation in the pellet before and during the tests. The improved code has been used to simulate some of the
experiments of interest of the FUMAC project (IFA650.10 and Studsvik 192 LOCA test). For IFA650.10, the
cladding outer temperature profile calculated with the SOCRAT code has been used. The results obtained with
ALCYONE were in a good agreement with the experimental data. In terms of uncertainty quantification, the
uncertainty on the determination of the boundary conditions like cladding outer temperature resulted in a large
uncertainty on the cladding deformation and the burst time.
In the fifth paper on qualification of TRANSURANUS models for mixed core fuel based on the FUMAC
outcome, Mr. M. Ieremenko (SSTC, Ukraine) presented results of testing of the TRANSURANUS code for
modelling the behaviour of WWER nuclear fuel in the LOCA accident conditions. A simulation of a part of the
experimental data (MTA-EK data, IFA-650.10, IFA 650.11 and Studsvik 192&198) was performed. Also
calculations of some of these data sets and KIT QUENCH-L1 set were carried out by other teams using the
TRANSURANUS code (INRNE, Bulgaria; JRC, Germany). TRANSURANUS code demonstrated good
possibility to predict behaviour of nuclear fuel rod cladding. Predicted geometry of cladding and time of burst for
both general types of cladding (PWR and WWER) showed good correlations with experimental data for such type
of regimes. In addition, the experimental data of the FUMAC project contained the results of post-irradiation
measurements after operational in commercial reactor. This data was used to test the capabilities of the
TRANSURANUS code for modelling of the irradiated fuel rod behaviour in the core.
The last paper on the main results of FUMAC from CNPRI and sensitivity and uncertainty study of
performance of ATF under transient condition was presented by Ms. Q. Ren (CGN, China). With the simulation
of Halden fuel rod LOCA behaviour tests IFA-650.9/10/11 using FRAPCON/FRAPTRAN, evaluation of
integrated performance of fuel rod under LOCA condition was performed. Generally, FRAPCON/FRAPTRAN
predicted well the fuel rod behaviour under LOCA condition. Sensitivity and uncertainty study on IFA-650.10 was
then performed using the DAKOTA code, for both conventional fuel (UO2 + Zry-4) and ATF fuel systems (UO2-
BeO + ODS FeCrAl). Based on assumed uncertainty distributions for selected input parameters (5 fuel rod design
data in uniform distribution, 12 physical properties and thermal hydraulic conditions), several global sensitivity
methods have been tested in order to determine the influence on the clad failure time under LOCA condition. It
was found out that the design parameters, especially the pellet diameter have important influence on fuel system’s
performance during LOCA. Proper adjustment of design parameters can optimize the fuel performance for ATF.
3
2.1.3.General conclusion
— All experimental data analyzed reveal a relatively large spread of the measured burst strains, which is not
only determined by the local conditions of temperature and pressure, but also by heterogeneities and
composition or micro-structural variations in the materials under investigations. An uncertainty analysis of
the experimental data was therefore recommended, including those data that have been used for the
development and validation of the codes applied;
— The fuel rod codes simulate reasonable agreement in the cladding and fuel thermal behaviours but large
dispersions in the cladding deformation and burst time;
— Further improvements in the plenum temperature, fuel relocation, cladding deformation and burst models
are needed. The thermal hydraulic boundary conditions are also important;
— Uncertainty and sensitivity analysis is helpful in improving the modelling of fuel behaviour during LOCA.
2.1.4.Further recommendations
— More analysis and cases for WWER fuel would be needed, especially in view of recent advanced fuel
developments;
— A general interest has been expressed to consider a similar analysis for advanced fuels and cladding materials,
including some of the so-called accident tolerant fuels (ATFs);
— For a successful CRP that involves so many cases and participants, the duration of the project meetings may
be extended, enabling for instance to analyse better model details and code changes as well as their
comparison with more detailed experimental data;
— Deeper analysis of failure criteria with advanced tools, and uncertainty analysis on experimental data is
needed;
— Need more quantitative information about fuel fragmentation.
It is also recommended to focus the future CRP in fuel modelling on more practical applications to support
the sustainability of nuclear technology, diversification of fuel supply, and innovation in fuel technology
development. The final objective is to apply the qualified fuel rod codes to develop a best estimate plus uncertainty
analysis (BEPU) type LOCA hot rod fuel safety evaluation methodology (FSEM) for assessment of the
performance of the evolutionary Advanced Technology Fuels (eATFs), and verification of the compliance with
the applicable LOCA safety criteria (e.g. the USNRC new LOCA criteria 10CFR50.46c).
2.2. Technical session 2: Experiments on fuel behaviour in accidental conditions
The second technical session focused mainly on the experimental investigation of new ATF cladding
concepts to support modelling and simulation activities. It was chaired by Mr. Martin Ševeček, Czech Technical
University, Czech Republic.
2.2.1.Overview
The session provides an overview of the experimental testing of different ATF cladding materials such as
silicon carbide, FeCrAl alloys, and coated cladding. The bundle test with FeCrAl alloy B136Y was supported also
by pre- and post test modelling activities using two severe accident codes. The sessions highlighted the expected
benefits as well as the remaining challenges of ATF cladding development.
Both contributions incorporated into this session were given by researches from the Karlsruhe Institute of
Technology (KIT), Germany. The first contribution was focused on the details of the first-ever integral bundle
experiment with ATF cladding material including modelling benchmark that followed. The second talk presented
experimental results from testing of four different ATF cladding candidates:
— Mr. Juri Stuckert (KIT, Germany) — Experimental modelling results of the QUENCH-19 bundle tests with
FeCrAl claddings;
4
— Mr. Chongchong Tang (KIT, Germany) — Oxidation performance and failure behaviour of monolithic and
coated ATF claddings under severe accident conditions.
Mr. Stuckert presented in his talk an overview of the bundle experiment with FeCrAl alloy in the well-known
QUENCH facility at KIT. The B136Y alloy was provided by Oak Ridge National Laboratory and the accident
scenario was chosen to allow direct comparison between older QUENCH experiments such as QUENCH-15 test
with ZIRLO alloy. B136Y alloy is advanced steel from the group of alloys usually titled as FeCrAl. The
composition of B136Y is as follows: Fe-80.85, Cr-12.99, Al-6.14, Y-0,003. It is considered as one of the main
near-term ATF candidates. There were several challenges found related to testing of this ATF material. These
challenges included also a new design of thermocouples that eventually initiated the melting of the bundle.
Additionally, there were steam leakages detected from the bundle that caused the wetting of the insulation and
related radial heat losses.
Substantial benefits of B136Y were confirmed by the test including reduced oxidation kinetics and hydrogen
production. The coping time that can be quantified for this specific design and scenario is on the order of several
thousand seconds. However, new issues related to FeCrAl deployment were identified. One issue is related to the
higher coefficient of thermal expansion of FeCrAl that is approximately 3 times higher than Zr. When reflooding
and quenching, the dimensional changes cause high stresses and failure of the cladding. Additionally, when FeCrAl
melts it is not covered with strong oxide layer such as traditional Zr alloys which allows higher mobility of melts.
It was also found that when locally melted, FeCrAl interacts with ZrO2 pellets that were used. This is not observed
for Zr alloys and poses an additional challenge and requires more investigation.
The pre- and post test calculations were performed with ATHLET-CD code by GRS. GRS models showed
good agreement with experiment, but serious challenges remain. The main inconsistency is related to the oxidation
model used. It strongly underpredicts oxidation and hydrogen release. The post test calculation was done also by
IBRAE using the SOCRAT code. The uncertainties revealed by calculation emphasize high needs for more
experimental data and separate effect experiments to support the development of new correlations.
In the second talk the team at KIT studied various types of ATF cladding candidates and mainly their
behaviour in DBA and DEC conditions. The studied materials are — Cr PVD coated Zry-2, Cr cold spray coated
Zry-4, FeCrAl alloys, MAX phase coated Zry-4 and SiC. Mr. C. Tang presented a range of tests with FeCrAl that
show that the oxidation resistance of this material depends on the heating scenario. This is due to the complex
oxide formation. It was shown that with high heating rates the oxidation of FeCrAl can be catastrophic.
Additionally, acceleration of oxidation was observed above 1350°C which is probably linked to Al(OH)3 melting.
MAX phase coated cladding showed the protective nature of the coating up to 1260°C in the transient test
but then local delamination and cracking were observed. Cr cold spray cladding showed good resistance against
oxidation at a lower temperature but at higher temperatures, local melting and cladding deformation were
observed. Tests with SiC provided by General Atomics and Westinghouse Electric Company were performed up
to 1750°C. It was found that silica bubbles formed above 1713°C and cladding locally failed at 1750°C. The
protective nature of SiC cladding is ensured only if the protective outer CVD SiC stays intact.
2.2.3.General conclusion
This session emphasized the need for more experimental data even for ATF concepts that have been already
inserted into commercial reactors. Both presentations revealed new failure modes and phenomena of ATF cladding
materials that were not known previously. The talks also highlighted the need for more high-quality experimental
data to support modelling and simulation activities.
The talks both confirmed the safety benefits of ATF cladding materials. Valuable separate effect test data
were presented but the claimed benefits of any ATF candidate should be confirmed by an integral bundle test. It
was also shown that ATF cladding material properties strongly vary based on the fabrication parameters. The
materials produced by fuel vendors and industry have been optimized in order to achieve the best possible
performance. Therefore, the transferability of experimental results between commercial and in-laboratory
produced materials is questionable. This was confirmed for both FeCrAl (B136Y alloy vs. laboratory alloy) as
well as for Cr coated cladding materials (disagreement with data published by fuel vendors). Additionally, new
phenomena such as SiC scale melting and excessive hydrogen release, FeCrAl failure during quenching or bending
5
of Cr cold-spray coated tubes were identified. This confirms that there is still a high need for experimental testing
of ATF cladding concepts including support for modelling and simulation activities.
2.3. Technical session 3: Fuel performance codes applied to design basis accidents (DBA) and to design
extension conditions (DEC)
The third technical session was chaired by Mr. Antoine Bouloré, CEA, France.
2.3.1.Overview
Whereas the first session was dedicated to the FUMAC CRP, mainly related to fuel behaviour modelling in
Design Basis Accidents, this session was focused on the fuel modelling and simulation in both Design Basis
Accidents (DBA), but also in Design Extension Conditions (DEC), which, according to IAEA standards, could
include conditions without important fuel degradation and conditions with core melting. During the session,
participants have presented their experience and their recent developments in relation with this topic. One objective
of this technical meeting was to prepare a proposal for the next CRP. So, in the general discussion at the end of
the session, some recommendations and general comments have been raised to help prepare this new CRP.
Seven presentations were made during this session:

— Assessment of high conductivity ceramic fuel concept under normal and accident conditions (Mr. A. Abe,
Brazil);
— Modelling of accident scenarios with the FINIX fuel behaviour module (Mr. H. Loukusa, Finland);
— Safety analysis of fuel behaviour for implementation licensing (Ms. H. Papp, Hungary);
— Fuel behaviour during DBA and DEC: analytical and experimental studies for development of ATF
(Mr. A. Gaikwad, India);
— An introduction to a multidimensional reactor simulation environment with the capability of advanced
modelling of nuclear fuel under accident conditions (Mr. R. Vadi, Islamic Republic of Iran);
— Simulation of IFA650.5 experiment using the fully coupled SPACE-FRAPTRAN code system (Mr. H. Kim,
Republic of Korea);
— Thermo mechanical performance of TRISO coated fuel particle used in high temperature Gas Cooled Reactor
(Mr. S. Liu, China).
Mr. A. Abe showed an example of assessment of an advanced fuel concept. This new fuel is made of UO2
particles surrounded by BeO matrix, in order to enhance the thermal conductivity of the pellet. A simple model
for the equivalent thermal conductivity is presented (mixing law) and considering the standard models in
FRAPCON/FRAPTRAN code, the effect of the lower temperature on the fission gas release is also presented.
Mr. H. Loukusa presented a multiphysics coupling scheme developed at VTT. The fuel behaviour module
FINIX has been coupled to the neutronics code SERPENT2 and to thermal-hydraulics codes TRAB1D, TRAB3D
and HEXTRAN. In the presentation, some elements of validation have been shown (IFA650.5) but also some
comparisons between stand-alone and coupled calculations. At the moment, the main limitations of the FINIX
module is in mechanical modelling, and this will be improved.
The presentation of Ms. H. Papp was an example of safety analysis, and an illustration of a conservative
approach in safety criteria assessment. This conservative approach consists in the selection of the most critical
rods in operation, and their deterministic simulation to make sure they verify the safety criteria.
A large number of examples of analytical and experimental studies made for ATF were given by Mr. A.
Gaikwad. Physical and chemical properties of many types of coated claddings and an evaluation of their effect in
simulations were presented.
The next two presentations were about the development of multiphysics coupling schemes. Mr. R. Vadi
presented a simulation environment consisting of three codes (a CFD based code for thermal-hydraulics, and two
neutronics codes for point kinetics equations and multigroup diffusion equations). Some results were presented
for LOCA conditions. In the presentation of H. Kim, a system analysis code (SPACE) was coupled to the fuel
performance code (FRAPTRAN). An illustration of this coupling scheme for LOCA test IFA650.5 was presented.
6
This test shows that the surface-to-surface radiation must be modelled precisely to predict the cladding
temperature. So, both presentations contained some elements of validation of the coupling schemes.
The last presentation of the session by Mr. S. Liu was about TRISO coated fuel particles with UN as the
kernel. Taking into account the physical and mechanical properties of the different materials involved in such
particles (kernel, coatings, …), the performance of this type of fuel in RIA conditions are assessed. The main
conclusion is that the thermal expansion is the dominant factor resulting in the failure of the particle in RIA
conditions.
2.3.3.General conclusions
In this session, several papers about fuel performance code developments and fuel behaviour modelling were
presented. The subjects covered the following: modelling of advanced fuel concepts, safety analysis example, link
between experimental and modelling activities for ATF and multiphysics coupling.
All presentations in this session contained some validation aspects of the codes and coupling schemes
developed (comparison between calculation and experimental data), which is a very good point if the computation
tools are used for safety analysis.
One objective of the present technical meeting was also to prepare the proposal for the new CRP following
FUMAC and ACTOF. In the discussion of this session, the link between the modelling capabilities of the Member
States and the requirements of this next CRP were listed and discussed.
The first important point was about the validity of the codes for transposition to real reactor cases. Fuel
performance codes presented in this session are all “single pin” codes, and the calibration of those codes is
performed on single pin experimental data. The transposition of the calculated results to real reactor cases raises
some questions. This point should be addressed to the next CRP.
In the second point of the discussion, it was mentioned that each fuel performance code has its own validation
domain, which is the domain where calculated results can be compared to experimental data. If the code is be used
out of its validation domain, extended margins have probably to be considered to take into account the lack of
knowledge of fuel behaviour out of this domain. The question of the validity of the extrapolation of the validation
domain to new concepts and new materials for ATF should be addressed in the next CRP.
In the session, several presentations introduced multiphysics simulation. Up to now, safety analysis is mainly
based on the single fuel pin analysis. But the question of the necessity of multiphysics coupling environment is
raised, which should be addressed in the next CRP. The two approaches presented (single fuel pin and multiphysics
coupling) have their pros and cons.
As mentioned in conclusion of the first session of this meeting, the next CRP will probably be dedicated to
the development of a method to quickly assess the performance of advanced technology fuel using simulation and
testing. In this framework, this raises the question of the material properties to be used for the new materials or the
new concepts tested. If these concepts are new, there will be probably very few data to consider the evolution of
those properties with irradiation. So experimental data are required on the new materials and new concepts to make
sure the proper material properties are considered in the simulation. This is why both testing and simulation have
to be considered in the definition of the next CRP. In the case of complex materials, such as composite or
heterogeneous materials, the modelling of the material properties should also be addressed.
The last point mentioned in the discussion for future work was about the uncertainties. In safety analysis,
margins must be defined, and a key point that should be addressed in the next CRP is how to take into account
epistemic uncertainties, which represent in fact a lack of knowledge of material properties, for example, in the
definition of these margins. The modelling of uncertainties is also a key point, as it was already mentioned in the
conclusions of the FUMAC project.
7
2.4. Technical session 4: Main results of the CRP ACTOF
The fourth session was chaired by Ms. Tong Liu, CNPRI, China.
2.4.1.Overview
This session provided an overview of the recently completed IAEA Coordinated Research Project on
Analysis of Options and Experimental Examination of Fuels for Water cooled Reactors with Increased Accident
Tolerance (ACTOF). The CRP dealed with the acquisition of data through experiments on new fuel types and
cladding materials, development of modelling capacity to predict the behaviour of the components and the integral
performance of accident tolerant fuel designs under normal and transient conditions and to demonstrate
improvements under severe accident conditions. The objective was to provide information to Member States to
support decision making on the choices available to improve the safety of Nuclear Power Plants under normal and
severe accident conditions.
Various ATF options were analysed resulting in improved understanding of ATF behaviours and
developments of ATF technologies in Member States. For example, several coated cladding materials were
produced, tested, characterized and analysed. The feedbacks are important for Member States to improve their
ATF concepts and development activities.
Fuel performance codes from participating organizations were significantly extended to the analysis of
several ATF concepts (including FeCrAl and SiC claddings, coated zircaloy claddings, and U3Si2 fuel).
Furthermore, a benchmark was organized and finalized where the extended codes were compared for the analysis
of fuel rod behaviour under both normal operation and design-basis accident conditions. Additionally, Monte Carlo
simulations were carried out to investigate the effect of new claddings and fuel on a potential cycle. Integral codes
(ATHLET-CD and SOCRAT) were applied and benchmarked against the severe-accident bundle test QUENCH-
19 with FeCrAl claddings of fuel rods.
In this session, four presentations were made:

— Main results of the IAEA Coordinated Research Project on Analysis of Options and Experimental
Examination of Fuels for Water cooled Reactors with Increased Accident Tolerance (ACTOF)
(Mr. M. Veshchunov, IAEA);
— Reactivity Initiated Accident assessment for ATF cladding materials (Ms. C. Giovedi, Brazil);
— Round Robin exercise of the candidate ATF cladding materials within the ACTOF Project (Mr. M. Ševeček,
Czech Republic);
— Introduction of ATF development and main results of the CRP on ACTOF of KAERI (Mr. C. Shin, Republic
of Korea).
Mr. M. Veshchunov gave an overview of IAEA CRP ACTOF by introducing the background, objectives,
participants and current status of ACTOF. The joint actions (including Round Robin tests, benchmark exercises,
severe accident experimental and modelling Exercise), the main outputs as well as the perspectives and
recommendations for the new CRP were presented.
Ms. Claudia Giovedi presented preliminary results for FeCrAl HbbBO-5 test case by using modified versions
of the FRAPCON-FRAPTRAN codes to simulate the fuel behaviour under steady state and RIA conditions.
According to the steady state study, FeCrAl showed better performance as compared to Zircaloy-4: higher fuel
temperatures, lower internal pressure, gap remained opened, and lower cladding hoop stress. And RIA simulation
results showed a good agreement with the experimental results.
In Mr. Martin Ševeček’s presentation, main results of Round Robin Test with ATF cladding candidates were
presented, including Cr-coated, MAX phase coated, ZrSi-Cr coated, AISI 348. According the experimental results,
only Cr coated concept seemed to be promising from perspective of both long term corrosion and high temperature
oxidation resistance.
Mr. Changhwan Shin introduced an overview of R&D activities for ATF in Republic of Korea. KAERI
focuses on surface modified Zr cladding with ODS treatments and ceramic or metallic microcell UO2 pellets as
the near-term choice. Mechanical Model of FRACAS-CT for multi-layered cladding has been developed and
implemented in FRAPCON. The fuel performance analysis with material properties of ATF developed by KAERI
8
shows ATF has significant advantage in the reduction of the fuel temperature, cladding oxidation thickness, fission
gas release, etc. Safety analysis was conducted to study the ATF benefits in LOCA and BDBA.
2.4.3.General conclusion and further recommendations
The following comments and conclusions were made via presentations and related discussions in the session:
— Development of standard methods and testing procedures are needed for all ATF products;
— Further experiments were proposed to carry out to better understand the failure mechanism and establish new
failure criteria for FeCrAl cladding under steady state and transient conditions;
— Most participants showed interests to carry out further round robin tests to enhance the database for better
understanding and evaluation of ATF behaviour and performance;
— A new CRP on testing and simulation of advanced technology fuels (ATF-TS) (2020-2023) was proposed.
2.5. Technical session 5: Modelling of traditional UO2 and advanced ATF fuels behaviour in integral
severe accident (SA) codes
The fifth session was chaired by Mr. Juri Stuckert, KIT, Germany.
2.5.1. Overview
This session provided an overview of the results of the benchmark on post test analyses of the CORA-15
bundle test with the system codes ATHLET-CD and SOCRAT, the status of ATF modelling and application in
ATHLET-CD, and advanced cladding embrittlement criteria, which could be implemented in mechanistic fuel
performance and safety codes and applications.
2.5.2. Summary of presentations
This session included three presentations.

The first paper presented by Mr. J. Stuckert (KIT) described the results of the benchmark on post test analyses
of the CORA-15 bundle test with the system codes ATHLET-CD and SOCRAT. The CORA-15 test was conducted
under transient conditions typical of many CORA-PWR tests, with bundle containing two (Ag, In, Cd) absorber
rods. In contrast to other CORA bundles, all fuel rod simulators with and without heater (16 + 7) were filled with
helium and were brought to a pressure of 60 bar before the transition. During the transition process, all the rods
were ballooned and burst. Balloons developed over about 100 s. Bursts occurred within 150 s in the temperature
range of 650 to 850°C. The most probable burst elevation is a bundle height of 750 mm (the hottest bundle level
on the onset of burst). At temperatures above 1800°C, molten Zr formed in the gap between UO2 and ZrO2 was
partially moved downwards with concurrent dissolution of oxides and partially released through the destroyed
zirconium oxide layer. The resulting melt, which mainly contains U, Zr, O, moves downwards in the form of a
slug and hardens between 400 and 550 mm in the form of a large porous structure. Above the Inconel lattice spacer
at 500 mm, a maximum bundle blocking was observed (almost 100%). The absorber melt, which have a much
lower solidus temperature, solidified at a lower bundle level of about 150 mm. Investigations performed after the
test showed only slight oxidation of claddings up to bundle elevation of 350 mm. The cladding tubes were
completely oxidized between 480 and 1000 mm. The maximum value of hydrogen rate (210 mg/s) was measured
at the transient end (4800 s). The total hydrogen mass released during the test was 145 ± 15 g.
The latest version of the ATHLET CD code was used to simulate the CORA-15 test, with very similar set of
input data and simulation options as for CORA-13 (pressure free rods; it was subject of the international problem
IS-31). In general, for both tests, a good agreement was obtained between the calculated and measured data
concerning the temperature evolutions, with a slight underestimation at the top of bundle and a slight
overestimation at the bundle bottom. The behavior of pressurized rods in CORA-15 was adequately predicted by
the code. The calculated time of the cladding burst, bundle elevation and temperature of the cladding tubes at the
burst position were within the experimental ranges. Hydrogen release was determined within the uncertainty of
the measured data, overestimating the generation rate during the oxidation acceleration between 4000 and 4500 s.
One discrepancy in the calculated results is due to the lack of model of melt retention at spacer grids, which leads
to a shift in the blockage profile and overestimation of temperatures at bundle levels below grid spacers.
9
To analyze the CORA-15 test, the current version V3 of the SOCRAT code was used. From a comparison
of the simulation results with the CORA-15 experimental data, it can be concluded that the SOCRAT code
reproduces quite well the temperature history of cladding tubes and tends to easily overestimate the temperature
of the claddings at the upper end of the heated region. The calculated parameters of bursts (time, temperature,
pressure, and bundle elevation), the total mass of hydrogen, and bundle blockage due to relocation and
solidification of melt are in good agreement with the data, taking into account the measurement accuracies.
Following the modeling of the experiment CORA-15, it is therefore established that the experimental data
are fairly good simulated (temperature history, parameters of bursts, total hydrogen release, final bundle
degradation). The results reached during the CORA-15 analysis by both system codes SOCRAT and ATHLET-
CD showed a good agreement with the measurements and with each other.
The materials of this presentation on the results of the CORA-15 test and the corresponding post-test analyses
with ATHLET CD and SOCRAT were recently published: J. Stuckert, H. Austregesilo, Ch. Bals, Th. Hollands,
A. Kiselev, D. Tomashchik, T. Yudina, “Post-test analyses of the CORA-15 bundle test with the system codes
ATHLET-CD and SOCRAT”, Nuclear Engineering and Design 342 (2019), pp. 320-335, and for this reason are
not included the current TECDOC.
The second paper, presented by Mr. L. Lovasz (GRS), described status of ATF modelling and application
with ATHLET-CD.
The QUENCH-19 bundle test with FeCrAl claddings was accompanied by post test analyses at GRS with
the AC² module ATHLET-CD. The QUENCH-19 test was conducted at KIT similarly to the QUENCH-15 test
for comparing the oxidation behaviour of FeCrAl and ZIRLO claddings at high temperatures. In contrast to the
test QUENCH-15, which showed a sharp escalation after start of quenching resulting in temperatures up to 1900°C,
no escalation occurred during QUENCH-19 and peak cladding temperatures of only about 1450°C were reached.
For the post test simulations an oxidation correlation for the KANTHAL APMT alloy was available. One
simulation used the oxidation kinetic defined by constants given for KANTHAL APMT, the other used the same
constants, but the reaction rate was multiplied by 300. Both approaches were implemented in ATHLET-CD with
the assumption that only Al2O3 layer was formed at the outer cladding surface, but no other oxides. Compared to
oxidation of Zr alloys both approaches gave orders of magnitude lower oxidation rates.
In general, the results of the post test simulation show good agreement of the thermal behaviour. While the
maximum temperature well reproduced an underestimation of only about 50°C, the radial temperature profile was
significantly underestimated by 150-200°C compared to the measured data. The comparison of the hydrogen
production of in total 9 g in the test and less than 1 g for both oxidation approaches in the simulation shows that
the oxidation model for FeCrAl in the code must be improved. A sharp increase in the hydrogen release was
observed at approximately 800 s before the quenching; one possible trigger for this event could be the failure and
melting of the steel cladding of thermocouples and following partial dissolution of claddings. This was not
considered in the post test calculations. For a detailed evaluation of the calculated hydrogen generation the post
test examination of the bundle is necessary to know which components contribute to the total value.
The third paper, presented by Mr. M. Veshchunov (IAEA), discusses advanced cladding embrittlement
criteria, which could be implemented in mechanistic fuel performance and safety codes and applications.
The analysis of experiments on direct quenching after isothermal oxidation at high temperatures showed that
the ECR 17% criterion notably underestimates cladding capability to withstand thermal shock during LOCA
reflood. On the other hand, the tests (impact, ring compression and 3-point bend) on the capability of the cladding
to withstand anticipated loads during fuel handling and transport of the assemblies showed that the 17% criterion
notably overestimates ductile-to-brittle transition oxidation level. For this reason, the alternative more advanced
and physically based Chung-Kassner criteria as well as the KAERI criterion for post-quench embrittlement are
considered for practical applications.
The advanced Chung-Kassner criteria are based on a more physical consideration (related to the thickness of
ductile β-Zr phase) and were validated by the authors of the criteria against a very wide set of quenching tests in a
wide temperature range (up to 1500–1900°C) that is far beyond LOCA conditions, and thus can be broadened to
Design Extension Conditions (DEC). For practical applications of these advanced criteria in fuel performance
codes, mechanistic diffusion models of cladding oxidation under high temperature transient conditions should be
applied. It allows modelling oxygen and hydrogen radial distributions in the cladding under normal and transient
conditions with consideration of cladding-fuel interactions. Utilization of the advanced criteria in the mechanistic
SVECHA/QUENCH code allows reasonable predictions not only for LOCA conditions, but also for oxidation and
10
quenching under severe accident conditions of KIT (with fresh Zry-4 claddings) and RIAR tests (with irradiated
WWER fuel fragments).
The second Chung-Kassner criterion on capability to withstand fuel handling, transport and storage should
be supplemented with an additional condition related to secondary hydriding, revealed in the KIT QUENCH-
LOCA tests (brittle fracture of claddings with local hydrogen content more than 1500 wppm during tensile tests).
The applicability of the KAERI advanced failure criteria (with some potential modifications) for oxidized Zr
alloy claddings with Cr coatings was confirmed in the recent ring compression tests performed within the IAEA
Coordination Research Project (CRP) ACTOF. On this base it is assumed that extension of the mechanistic
oxidation model to consideration of the coated claddings will allow verifying applicability of the advanced Chung-
Kassner and KAERI criteria (as well as the KIT complementary criterion) on the base of future tests on the coated
cladding quenching (which are foreseen, e.g. in the subsequent CRP on ATF in 2020–2023).
2.5.3. General conclusions and further recommendations
The development of mechanistic models for traditional and advanced fuel materials should be continued.
It would be very important to use results of conducted and planned bundle experiments with ATF materials
for the organization of benchmarks involving different codes.
A new CRP on ATF materials should include various single effect tests on:
— Determination of oxidation kinetics for reactor relevant FeCrAl alloys, especially near to the melting point
of these materials;
— Investigation of Cr coating behaviour at high temperatures for coated Zr alloy claddings.
3. CONCLUDING REMARKS AND RECOMMENDATIONS
In this Technical Meeting, participants exchanged their view on national perspectives, R&D progress and
results via paper presentations and discussion forums.
The general remarks were as follows:
— The IAEA CRPs on FUMAC and ACTOF provided an excellent platform to collect relevant experimental
data and compare the fuel performance and integral code calculation results with others and especially with
experimental data;
— The participation of the organisations providing the experimental data was very useful in clarifying various
questions raised during the projects;
— An uncertainty analysis of the experimental data was necessary, including those data that have been used for
the development and validation of the codes applied;
— Efforts are needed to clarify differences in the interpretation of some experiments and in the use of the same
codes, the so-called “user effect”;
— Efforts should be made to apply the validated computer codes and uncertainty analysis methods to evaluate
the fuel behaviours and performance of advanced technology fuels (ATFs) in nuclear power plants (NPPs).
It was recommended to focus the new CRP on the testing and simulation of evolutionary Advanced
Technology Fuel (ATF-TS), consisting of Round Robin mechanical tests of coated cladding and other concepts,
integral out-of-pile bundle tests with coated cladding, code adaptation and validation for ATFs, and development
and application of simplified methodology for ATF hot rod LOCA fuel safety evaluation.
Most participants showed interests to participate in this new CRP. It was recommended to hold a Consulting
Meeting to finalize the CRP programme and launch the proposal evaluation process by end 2019 and start this
project in 2020.
11
MAIN RESULTS OF THE CRP FUMAC
(Session 1)
Chairperson
J. ZHANG
Belgium
13
SIMULATION TOOLS INCORPORATED TO DIONISIO IN THE FRAME OF THE
FUMAC PROJECT
M. LEMES, A. SOBA, A. DENIS

Sección Códigos y Modelos,
Gerencia Ciclo del Combustible Nuclear,
Comisión Nacional de Energía Atómica,
Avenida General Paz 1499, 1650 San Martín,
Provincia de Buenos Aires, Argentina
Abstract
Several models were integrated to the DIONISIO code within the framework of the IAEA Research Project “Fuel
Modeling in Accident Conditions (FUMAC)”, to take account of accidental conditions, in particular the loss of coolant
accidents (LOCA). A specially designed thermal-hydraulic subroutine provides a simplified description of the rod environment
in normal or accidental conditions. The heat transfer coefficients that account for the different coolant regimes, in single or
double phases, are activated as the corresponding conditions occur. The simulation of a considerable number of experiments
has shown that, despite its simplicity this subroutine gives adequate predictions of the conditions in a vertical cooling channel,
quite similar to those given by the thermal-hydraulic codes. The description of the fuel rod atmosphere is improved with the
incorporation of this subroutine since it provides fairly realistic boundary conditions for the simulation of the fuel rod behavior,
without requiring the intervention of external specific codes. Models of high temperature oxide growth (ZrO2) and hydrogen
capture and release by the cladding in steam were also included. Moreover, the model of cladding creep predicts the conditions
for ballooning and eventually, those for catastrophic failure (burst) and its localization. The calculation scheme makes a
partition of the rod length into a number of segments defined by the user. In each segment the local conditions are considered
to calculate, with the synchronous work of all the subroutines, the physical and chemical parameters in one representative
pellet. Then, a description of the whole rod is obtained by coupling all the segments. This strategy has yielded accurate
simulations of a wide variety of cases, either in normal or LOCA type conditions. Exhaustive comparisons were carried out
with several thermal-hydraulic codes (COBRA-IV, RELAP5-Mod3.1, SOCRAT, ATHLET-Mod 1.1) and with a number of
experiments like those of the IFA–650 series (-1,-2,-9,-10,-11), PUZRY, QUENCH-L0/L1 (for which a new working scheme
was specially developed in DIONISIO), CORA-15, IAEA-SPE-4, among others.
1. INTRODUCTION
The DIONISIO code simulates the behavior of a typical PWR fuel rod under irradiation. It starts from the
linear power history and environmental temperature profiles to predict temperature, stress and strain (either in the
elastic or plastic regimes) distributions in the rod, creep, densification, swelling, release of fission products (noble
gases, cesium and iodine) to the internal free volume of the rod, gas mixing, pressure increase, cladding irradiation
growth, superficial oxide layer development, hydrogen capture and release, pellet restructuring and grain growth
and pellet-cladding mechanical interaction. The effects cladding corrosion due to the internal or external
atmosphere are also considered.
In its present version, the code is able to simulate a fuel rod in its whole length. To this end, the bar is axially
divided into a quantity of sectors defined by the user. In each one the linear power and external temperature are
assumed uniform. The domain formed by one pellet and the corresponding segments of cladding and gap is
considered. The differential equations describing all the features of the problem are solved in each rod segment
under the local external conditions, using the finite element method in a two-dimensional system, assuming
symmetry around the longitudinal axis and with respect to the pellet mid-plane. At every time step the code
provides the local values of temperature, stress, strain and every other significant parameter, conforming a net of
interconnected and mutually dependent subroutines. The results of the individual segments are then merged to
give the time evolution of the global rod features.
The wide temperature range displayed through the radius of the fuel pellet is responsible for the highly non-
linearity of most of the models employed to describe the diverse phenomena.
Several improvements have been carried out on DIONISIO during the recent years, with the intention of
widening its predicting capability. To give account of the high burnup range, several models were incorporated
represent the radial distribution within the pellet of power density, burnup, and concentration of U and Pu nuclides
[1–4].
15
Several types of accidents can occur in the core of a nuclear power reactor. One of considerable gravity is
that named Loss of Coolant Accident (LOCA). Fortunately, the safety systems normally act with short delay
interrupting the nuclear reaction, but the heat accumulated in the fuel rods and that still produced by radioactive
decay represent a large amount that has to be removed in possibly defective conditions.
During an excursion of this type, high temperature values can be reached in the rod. The consequent fast
cladding oxidation goes together with the release of a considerable amount of hydrogen, some of which dissolves
in the coolant and the rest in the metallic cladding, usually of Zry. Both situations involve severe risks, either due
to H2 accumulation in the reactor containment or to cladding material embrittlement due to hydrides precipitation
and justify the necessity of accurate predictions in this respect. The consequent coolant temperature increase can
give place to different conditions: turbulent flow of liquid water, birth of water vapor bubbles adjacent to the
cladding wall, breakup of the boundary layer and several more that can develop simultaneously at different places
along the rod and also at a given location during an accidental event until the equilibrium is restored.
Quite recently, a subroutine was developed and included in DIONISIO to analyze and quantify the coolant
behavior in terms of the system pressure and coolant velocity. A vertical channel containing one fuel rod and the
corresponding volume of surrounding coolant is modeled. The purpose is to describe in a simple manner the
conditions in the cooling channel, including single phase, either liquid or vapor, or two phase flow, thus avoiding
the use of specific thermal-hydraulic programs. The comparisons carried out with a considerable number of
experiments and with diverse reactor codes have given stimulating results, despite of the comparatively great
simplicity of the new subroutine.
2. MODELS AND PARAMETERS INVOLVED
To calculate the temperature distribution in the pellet-gap-cladding system, the heat equation in cylindrical
symmetry is solved. Fissions and radioactive decay in the pellet and oxidation of the external cladding surface are
considered as heat sources [1–5].
Heat is assumed to flow only in the radial direction. Simulation of the oxidation kinetics plays a significant
role since the oxide layer represents a thermal barrier between the rod and the coolant. Models allowing for either
normal (with unlimited oxygen supply) or accidental (with the rod in a steam starved atmosphere) conditions are
considered [5]. On the other hand, the convective cladding-coolant heat transfer coefficient, ℎ [W/(m K)] needs
also to be considered.
The thermal-hydraulic, complex codes describe the wide diversity of flow regimes that can be met during a
power excursion and/or along a vertical channel with water flowing upwards in typical PWR fuels [6–9]. The
objective of DIONISIO is focused on the simulation of the fuel behavior that requires realistic boundary conditions.
For this reason the thermal-hydraulic analysis is restricted to those modes regarded as the more representative [5]
and is condensed in a code subroutine.
— Forced single-phase convection to subcooled liquid;
— Saturated nucleate boiling;
— Post-critical transition boiling;
— Post-critical film boiling;
— Forced single-phase convection to superheated vapor.
The dissociation of water molecules caused by oxidation of the cladding, particularly rapid in accident
conditions, results in a considerable increase of dissolved hydrogen in the coolant and in the cladding material.
During a sudden cooling of the bar, a fraction of the hydrogen in solid solution can precipitate and form hydrides,
which can be the cause of serious damage to the material.
The model included in DIONISIO [10] involves two different mechanisms of hydrogen absorption by the
cladding: chemical reaction with H2 present in the coolant and rapid diffusion of H+ through the oxide layer and
the consequent discharge into the oxide-metal interface via the capture of an electron released in the oxidation
reaction [11–16].
To estimate the effective cladding creep rate, a temperature dependent empirical correlation [16] was
adopted. In a LOCA type scenario, the temperature, clad ductility and tangential stress can combine in such a way
that the increase of internal rod pressure and depressurization of the primary circuit, can lead to a local strain
sufficiently high as to make the cladding adopt a shape similar to a balloon. This phenomenon can eventually cause
catastrophic clad rupture, which is assumed to occur when the hoop stress exceeds a limit value for which the
16
empirical correlation [16, 17] was adopted. Specific subroutines have been included in the code to predict clad
ballooning and burst occurrence and localization.
The models expressing each phenomenon have been codified in subroutines that were separately tested prior
to their insertion in the code [5, 10].
3. RESULTS
After the introduction of the accident modules in DIONISIO, several complex experiments were simulated,
all of them taken from the IAEA database.
3.1. Simulation of IFA tests
The fuel designs and cladding materials developed in recent years, as well as the tendency to increasing
burnup, determined the need to re-examine and update the traditional safety criteria for LOCA accidents, which
had been developed on the basis of experiments performed in the 70s. This was the objective of the integral
experiments of IFA (Instrumented Fuel Assembly) that were carried out in the Halden Reactor (Norway), designed
to evaluate aspects such as the cladding temperature, the relocation of the fragmented sediment, cladding
ballooning, hydriding and oxidation, under simulation LOCA conditions. They included a sharp pressure drop and
low refrigerant flow [18].
The IFA 650-1 and -2 tests were performed in 2003 and 2004 respectively, with fresh fuel rods that were
electrically heated in a heavy water and high pressure circuit. Unlike these, subsequent experiments of the IFA
650 series used pre-irradiated fuel rods in PWR or BWR type reactor reactors that had reached intermediate or
high burnup levels [19–21].
To simulate the experimental conditions of both tests, the thermal-hydraulic model incorporated to
DIONISIO was applied, taking as input data the power history, the geometric characteristics of the rod (radii of
the pellet and cladding, rod length) and aspects of the coolant (flow, pressure, inlet temperature).
100
30 a.
90
25 80
linear power (10 W / m)
70
20
Pressure (10 Pa)
60
2
15 50
40
10 30
20
5
10
0 0
0 2000 4000 6000 8000 10000 12000
Time (s)
FIG. 1. IFA 650-1 experiment. a.) linear power (red) and pressure history (dark green); b.) temperature evolution:
experimental results in blue and simulations with DIONISIO in orange.
17
In Figure 1(a), the red curve shows the linear power in the IFA 650-1 experiment, which consisted of four
constant power steps. In the same interval, six pressure drops followed by stationary periods occurred, which are
shown in the dark green curve. Thermocouples placed at three axial positions recorded the temperature throughout
the experiment. They are shown in the blue curves in Figure 1(b). In the same figure, the temperature calculated
by DIONISIO at the 20 axial positions selected for the simulations [5] are shown in the orange curves. The very
good agreement between both sets of curves is visible. In particular, the small frame inserted in figure b shows a
magnification of step 2.
In IFA 650-2, the experimental device was a loop filled with heavy water containing a rod segment. It was
instrumented with a pressure sensor and thermocouples to record the external cladding wall temperature.
Accidental conditions were simulated by valves opening, thus leaving the channel almost empty in about 30 s. The
coolant pressure decreased to 3–4 bar and hence the temperature rose. The linear power was kept at an almost
constant value of 23 kW/m until 464 s after valves opening, when scram occurred to end the test. The steep external
pressure drop, that can be appreciated in Fig. 2, corresponds to the accident initiation (t = 0). For the simulations
the rod is divided into 20 axial segments [5]. The curves plotted in Fig. 2 represent the temperature calculated at
the axial segments that better coincide with the thermocouples locations. The figure shows that clad ballooning
and burst occurred during the heat up period. In this respect, a band is plotted to indicate the time interval for
which the code predicts an accumulated creep sufficient to produce tube ballooning. During the experiment, burst
occurrence triggered the injection of a water spray to stabilize the temperature, as evidenced by the change of slope
of the curves. The good coincidence between measured and calculated temperature results can be appreciated.
1600
scram
ballooning burst
1400 70
1200 60
Temperature (K)
Pressure (10 Pa)

1000 50
5
experim. calculated
temperature 40
800
ballooning
30
600 burst
20
400
10
200
0
0 100 200 300 400 500
Time (s)
FIG. 2. Comparison between the experimental data of IFA 650-2 and the calculations of DIONISIO. Cladding temperature at
three axial positions and pressure evolution; ballooning and burst times as well as the scram are indicated.
Figure 3(a) shows the measured and calculated internal rod pressure in IFA 650-2. The abrupt internal
pressure drop reveals the burst event. According to the experimental report [18], the He pressure should have then
dropped to about 3 bar, which is the rig pressure. The simulation is carried out under this assumption. The
experimental curve shows a different value which is probably due to the sensor design, which is unable to measure
pressures below 56 bar (indicated with the horizontal line in Fig. 3(a). Burst occurrence is predicted at t =110 s
while the experimental report indicates 99 s. The decrease in pressure evidenced in the experimental curve before
the outbreak is attributed to the effect of ballooning [18]. It needs to be mentioned that the mechanical models
included in the code assume that the deformations are small. DIONISIO is not able to reproduce this effect, since
it involves large strains. A model suitable for these cases will be incorporated in the future. Figure 3(b) shows the
experimental and calculated values of the cladding lengthening as functions of the time measured from the pressure
drop. The experimental curve evidences an axial contraction that is associated with agglomeration. The abrupt
interruption of the contraction is attributed to burst occurrence, at 99 s [18]. The images obtained after the
conclusion of the test show an important relocation that is surely responsible for the significant contraction of the
18
rod observed in the final part of the experiment. This phenomenon, which implies large deformations and involves
the entire bar, cannot be taken into account in DIONISIO for reasons similar to those mentioned in relation to
Fig. 3(a) This explains the significant deviation between simulation and measurement after 80 s. Before that, the
agreement is quite good.
90
a.
internal pressure (10 Pa) 80
70
5
60
50
40
30 experimental
DIONISIO
20
10
0
0 25 50 75 100 125 150 175 200
time (s)
3,0
2,5 b.
2,0
1,5
elongation (mm)
1,0
0,5
experimental
DIONISIO
0,0
-0,5
-1,0
-1,5
-2,0
-2,5
-3,0
0 25 50 75 100 125 150 175 200
time (s)
FIG. 3. Comparison between the experimental data of IFA 650-2 and the calculations of DIONISIO. a.) internal rod pressure
and b.) clad elongation.
TABLE 1. MAIN CHARACTERISTICS OF THE BASE IRRADIATIONS OF THE RODS IFA 650-9, 10
AND 11.
Experiment Fuel type Base irradiated at Up to a burnup of
(MWd/kgU)
IFA 650-9 PWR Gösgen NPP (Switzerland) 89.9

IFA 650-10 PWR Graveline 5 (France) 61
IFA 650-11 WWER-440 Loviisa NPP (Finland) 56
19
For the IFA 650-9, -10 and -11 experiments, short rods cut from commercial fuels that had been previously
irradiated in different reactors were instrumented and further irradiated. The main characteristics of the base
irradiations are listed in Table 1. The three experiments started with forced coolant circulation until the temperature
was stabilized. Then, the cooling was changed to natural circulation.
1400 ballooning burst scram a.
1200
temperature (°C) 1000
800
blowdown
600
400
experim. (3 thermocouples)
200 SOCRAT
DIONISIO
0
0 100 200 300 400 500
time (s)
90
b.
80
70
pressure (10 Pa)
60
5
50
experim.
40 DIONISIO
30
20
10
0 100 200 300 400 500
time (s)
1,8
experim. c.
1,6 DIONISIO
1,4
elongation (mm)
1,2
1,0
0,8
0,6
0,4
0,2
0,0
0 100 200 300 400 500
time (s)
FIG. 4. Comparison between the experimental data of IFA 650-9 and the calculations of DIONISIO a) Cladding temperature
at different axial positions vs. time after accident initiation; ballooning, burst and scram times are also indicated. The
predictions of SOCRAT are superimposed. b) Internal rod pressure and c) Cladding elongation.
20
To perform the simulations with DIONISIO, input data of these tests like power history and geometric rod
characteristics (pellet and cladding radii, rod length, etc.) are provided. The thermal hydraulic model described in
Section 2 was applied to simulate the experimental conditions in the coolant like flow, pressure and temperature.
The predictions obtained are compared with the experimental results reported in the data basis and with those
calculated with the integral severe accident code SOCRAT, in particular those corresponding to the rod
temperature.
1100 ballooning burst scram

1000 a.
900
800
temperature (°C)
700
600
500
blowdown
400
300
experim. (3 thermocouples)
200
SOCRAT
100 DIONISIO
0
0 100 200 300 400 500
time (s)
90
b.
80
70
pressure (10 Pa)
60
5
50
experim.
40 DIONISIO
DIONISIO with SOCRAT BC
30
20
10
0
0 100 200 300 400 500
time (s)
1,4
c.
1,2
1,0
elongation (mm)
0,8
0,6
0,4
0,2
0,0 experim.
-0,2 DIONISIO
-0,4
0 100 200 300 400 500
time (s)
FIG. 5. Idem Fig. 4. but for IFA 650-10.
21
The cladding temperature evolution predicted by DIONISIO for the IFA 650-9, -10 and -11 tests at different
axial positions is represented in Figs 4(a), 5(a) and 6(a) together with the experimental records of three
thermocouples. The evaluations of the SOCRAT code at similar locations are superimposed for comparison. These
figures also show the times for initiation of blowdown, ballooning, burst and scram, measured and predicted by
DIONISIO.
ballonning burst scram a.

1000
800
temperature (°C)
600
blowdown
400 experim. (3 thermocouples)
SOCRAT(central sectors)
200
DIONISIO(central sectors)
0
0 100 200 300 400 500
time (s)
65
60 b.
55
50
pressure (10 Pa)
45
40
5
35
30
25 experim.
20 DIONISIO
15
10
5
0
0 100 200 300 400 500
time (s)
1,4 c.
1,2
elongation (mm)
1,0
0,8
0,6
0,4 experim.
DIONISIO
0,2 DIONISIO with SOCRAT BC
0,0
0 100 200 300 400 500
time (s)
FIG. 6. Idem Fig. 4 but for IFA 650-11.
22
Figures 4(b), 5(b), 6(b) show the experimental and calculated internal rod pressures for the three tests. As in
Fig. 3(a), the steep internal pressure drop indicates burst occurrence. Figures 4(c), 5(c) and 6(c) show the
corresponding values of cladding elongation. A good coincidence is in general appreciated. In particular, in
Fig. 4(c) the agreement is good only before ballooning. After that, the calculations predict a slower elongation
rate. The steep contraction experimentally observed is probably due to relocation, as explained in connection with
Fig. 3(b), but this fact is not yet included in the simulation.
With the purpose of testing the quality of the thermal-hydraulic subroutine of DIONISIO, simulations of the
internal rod pressure and clad elongation in IFA 650-10 and -11 were performed in two different conditions: with
the subroutine incorporated and replacing its outputs with those provided by the SOCRAT code, which operated
as boundary conditions (BC) for the fuel behavior simulation with DIONISIO. Figs 5(a) and (b), and Figs 6(a) and
(b) show the comparison between the measurements and both types of calculations for the internal pressure and
axial elongation. The agreement is good in general, although the predictions seem to be better when DIONISIO is
run with the embedded thermal-hydraulic subroutine.
Table 2 shows the time elapsed since accident initiation until ballooning and burst occurrence, measured
[22–24] and predicted by SOCRAT and DIONISIO.
TABLE 2. COMPARISON OF THE PREDICTIONS OF SOCRAT AND DIONISIO WITH

EXPERIMENTAL DATA FOR THE TIME FOR BALLOONING AND BURST
IFA 650-9 IFA 650-10 IFA 650-11

Ballooning (s) Burst (s) Ballooning (s) Burst (s) Ballooning (s) Burst (s)
DIONISIO 79.89 121.2 163.7 256.2 178.4 251.4
SOCRAT — — 109.7 235.4 118.2 202.4
Experimental 105 130 199 255 180 210
3.2. Simulations of PMK tests
The PMK experiments were carried out in the mid-1980s on a scale model of the Paks Nuclear Power Plant
(Budapest, Hungary). They had been designed to study the safety margins of WWER reactors and validate the
ATHLET, CATHARE and RELAP 5 thermal-hydraulic simulation codes. The reactor core was simulated by a
bundle of electrically heated bars. In particular, the IAEA-PMK-2 test aimed to investigate processes after small
and medium size breaks (from 1% to 22%) [25]. For this end, a hole was opened in the primary circuit to initiate
the accidental sequence. Approximately 200 s afterward the cooling device acted by steam discharge. The coolant
temperature and pressure recorded during the experiment are shown in Fig. 7 together with the predictions of the
RELAP, ATHLET and DIONISIO codes.
23
900 6
a.
800 5
cladding temperature (K)
Coolant flow rate (Kg/s)

experimental
700 RELAP
4
ATHLET
DIONISIO sector 9
600
3
500
2
400
1
300
200 0
0 400 800 1200 1600
Time (s)
600 140
b.
120
550 experimental
DIONISIO
coolant temperature (K)
100
pressure (10 Pa)

500
5
80
60
450
40
400
20
350 0
0 200 400 600 800 1000 1200 1400 1600
Time (s)
FIG. 7. Comparison between the experimental data of the PMK tests and the calculations of RELAP, ATHLET and DIONISIO;
a.) Cladding temperature at different axial positions; b.) Coolant temperature and pressure history.
The different heat transfer modes that were described in Section 2 and are included in the DIONISIO thermal-
hydraulic subroutine are activated during the experiment and along the rods axis to provide a realistic simulation
of the conditions met in the transient.
3.3. Simulation of QUENCH experiments
In the QUENCH-LOCA facilities of the Karlsruhe Institute of Technology (KIT), a series of tests are carried
out to investigate the ballooning and burst behavior, as well as oxidation and hydrogen uptake in design basis
accident conditions. Instrumented fuel rods are used, internally heated with tungsten wires connected to DC
generators. In particular, results of the QUENCH-L0 and L1 experiments are shown in Figs 7, 9, 10 and compared
with the corresponding simulations performed with DIONISIO. The device consists of 21 rods, the 10 internal
ones alimented with high and the 11 external rods with low electrical power.
24
To investigate the influence of the pressure on the involved processes, the fuel rods in QUENCH-L0 were
loaded with internal pressures of 35, 40, 45, 50, and 55 bar, respectively [26]. The experimental determinations of
pressure evolution are shown in Fig. 8(a) along with the predictions of DIONISIO. In Fig. 8(b) the measured and
predicted time for burst are compared. It is seen that the rods subjected to high power, i.e., at high temperature,
fail before than those at low power. No significant effect of the initial internal pressure is visible. The experimental
determinations show that burst occurs in the temperature range 1049–1123K while for DIONISIO a similar range
of 1062–1123K is predicted.
experim DIONISIO
70 internal rods (lower power)
external rods (higher power)
60
pressure (10 Pa)
50
5
40
a.
30
20
10
0
0 100 200 300 400 500 600
time (s)
280
external rods 8%
+
internal rods
t
260 en
m %
r ee -8
ag
ct
Calculated (s)
erfe
p
240
b.
220
220 240 260 280

Experimental (s)
FIG. 8. Comparison between the experimental data of the QUENCH-L0 test and the calculations of DIONISIO. a.) Evolution
of the internal pressure. b.) Measured and predicted time for burst.
25
1600 sectors 4 5 6 7
experim.
1400 DIONISIO
1200
temperature (K) 1000
800
600
400
200
0 100 200 300 400
time (s)
FIG. 9. Comparison between the experimental data of the QUENCH-L0 bundle test and the calculations of DIONISIO.
Cladding temperature in selected sectors of rod 4.
Figure 9 shows the temperature evolution in different sections of rod 4 of the QUENCH-L0 experiment. A
partition in 10 segments is used for the simulation with DIONISIO. They are numbered starting from bottom in
such a way that sectors 5–7 fall in the middle of the rod. After burst, the temperature is drastically reduced by
ceasing the power supply and increasing the cooling. However, the residual heat is sufficient to provoke water
evaporation and a further temperature increase. This is counteracted with the injection of vapor with argon, which
is visible in the tail of the curves.
Measured [26] and calculated hydrogen release in rod 4 are compared in Fig. 10. The sudden liberation
accompanying the fast experimental temperature drop is well accounted for by the simulation. However, a
disagreement between both at the final portion of the curves is visible. It is surely originated by the mismatch in
temperature in that portion of the test. Moreover, while the experimental data correspond to the whole bundle, the
simulation with DIONISIO took into account only rod 4 and then the results were extrapolated to the entire bundle.
This fact can contribute to the observed disagreement, especially in the final stage of the experiment.
0,06 1,2
Experimental
0,05 DIONISIO 1,0
accumulated H2 release (g)
0,04 0,8
H2 release rate
0,03 0,6
0,02 Experimental 0,4

DIONISIO
0,01 0,2
0,00 0,0
0 100 200 300 400 500
time (s)
FIG. 10. Comparison between the experimental data of the QUENCH-L1, rod 4, and the calculations of DIONISIO of the
instantaneous hydrogen release rate and accumulated amount of hydrogen released vs. time.
26
4. CONCLUSIONS
Several subroutines have been introduced in DIONISIO in the recent times to give account of individual
phenomena related to the cladding response in the high temperature range: oxidation, hydrogen uptake and release,
ballooning and burst. They have been all subjected to separate testing plans and then incorporated to the general
code.
Moreover, with the purpose of providing realistic boundary conditions to simulate the fuel behavior, a
subroutine was developed to give account of the different scenarios that can be met in the coolant, particularly
during an accidental excursion. Although this is the specific task of the complex thermal-hydraulic codes, the
strategy was adopted of introducing a simplified calculation scheme that could be embedded within DIONISIO,
thus avoiding the use of external inputs. This challenging task has been adequately fulfilled as evidenced by the
considerable number of experiments simulated, many of which are shown in this report, where a good agreement
is verified with the existent experimental data or with the simulations provided by the specific codes.
ACKNOWLEDGMENTS
This work is performed within the frame of the IAEA Research Contract N°18536 which forms part of the
Research Project “Fuel Modeling in Accident Conditions (FUMAC)”, and is partially supported by
PICT 2018-01568, CONICET — Argentina.
REFERENCES
[1] SOBA, A., DENIS, A., DIONISIO 2.0: New Version of the Code for Simulating a Whole Nuclear Fuel Rod under
Extended Irradiation, Nuclear Engineering and Design 292 (2015) 213–221.
[2] SOBA, A., DENIS, A., ROMERO, L., VILLARINO, E., SARDELLA, F., A High Burn-up Model Developed for the
DIONISIO Code, Journal of Nuclear Materials 433 1–3 (2013) 160–166.
[3] SOBA, A., LEMES, M., DENIS, A., An Empirical Formulation to Describe the Evolution of the High Burnup Structure,
Journal of Nuclear Materials 456 (2015) 174–181.
[4] SOBA, A., LEMES, M., GONZÁLEZ, M.E., DENIS, A., ROMERO, L., Simulation of the Behavior of Nuclear Fuel
under High Burnup Conditions, Annals of Nuclear Energy 70 (2014) 147–156.
[5] LEMES, M., SOBA, A., DAVERIO, H., DENIS, A., Inclusion of Models to Describe Severe Accident Conditions in the
Fuel Simulation Code DIONISIO, Nuclear Engineering and Design 315 (2017) 1–10.
[6] U.S. NUCLEAR REGULATORY COMMISSION, International Agreement Report, Heat Transfer Processes During
Intermediate and Large Break Loss-Of-Coolant Accidents (LOCAs), 1986.
[7] MOCHIZUKI, H., Thermal Hydraulics in Nuclear Reactors, International graduate course, Tokyo Institute of
Technology, 2009.
[8] ODAR, F., Assessment of the TRAC-M Codes Using Flecht-Seaset Reflood and Steam Cooling Data, U.S.Nuclear
Regulatory Commission, 2001.
[9] TODREAS, N., KAZIMI, M., Nuclear Systems. Thermal Hydraulic Fundamentals, Massachusetts Institute of
Technology, USA, 1989.
[10] LEMES, M., DENIS, A., SOBA, A., Simulation of nuclear fuel behavior in accident conditions with the DIONISIO code,
Journal of Nuclear Engineering and Radiation Science 5 2 (2019).
[11] VESHCHUNOV, M., BERDYSHEV, A., Modelling of Hydrogen Absorption by Zirconium Alloys During High
Temperature Oxidation in Steam, Journal of Nuclear Materials 255 (1998) 250–262.
[12] COX, B., Oxidation of Zirconium and it's Alloys, Advances in Corrosion Science and Technology 5 (1976) 173–391.
[13] VESHCHUNOV, M., SHESTAK, V., Modeling of Zr Alloy Burst Cladding Internal Oxidation and Secondary Hydriding
under LOCA Conditions, Journal of Nuclear Materials 461 (2015) 129–142.
[14] COX, B., “Mechanisms of Hydrogen Absorption by Zirconium Alloys”, Atomic Energy of Canada Limited, AECL
(Report), (January 1985).
[15] ADAMSON, R., GARZAROLLI, F., COX, B., STRASSER, A., RUDLING, P., Corrosion Mechanisms in Zirconium
Alloys, Adv. Nuclear Technology International, Skultuna, Sweden (2007).
[16] MANNGARD, T., JERNKVIST, L.O., MASSIH, A., Evaluation of Loss-of-Coolant Accident Simulation Tests with the
Fuel Rod Analysis Code FRAPTRAN-1.4. Quantum Technologies AB, Report TR11-008V1 (2011).
[17] ROSINGER, H., A Model to Predict the Failure of Zircaloy-4 Fuel Sheathing during Postulated LOCA Conditions.
Journal of Nuclear Materials 120 1 (1984) 41–54.
[18] EK, M., LOCA testing at Halden; the second experiment IFA-650.2, HWR-813, (2005).
[19] LESTINEN, V., KOLSTAD, E., WIESENACK, W., LOCA testing at Halden, Trial runs in IFA-650, Nuclear Safety
Research Conference Washington, (2003).
27
[20] LESTINEN, V., LOCA testing at Halden, first experiment IFA-650.1 HWR-762, (2004)
[21] EK, M., LOCA testing at Halden; the second experiment IFA-650.2, HWR-813, (2005)
[22] CHOMONT, F., HWR-917 LOCA TESTING AT HALDEN, THE NINTH EXPERIMENT IFA-650.9, (2009)
[23] LAVOIL, A., HWR-974 LOCA testing at Halden, the tenth experiment IFA-650.10, (2010)
[24] LAVOIL, A., HWR-976 LOCA testing at Halden, the VVER fuel experiment IFA-650.11, (2010)
[25] SZABADOS, L., EZSÖL, G., PERNECZKY, L., KREPPER, E., PRASSER, H., SCHÄFER, F., Two Phase Flow
Behavior during a Medium Size Cold Leg LOCA Test on PMK-2 (IAEA SPE-4). KFKl Atomic Energy Research
Institute, Budapest, Hungary, (1995)
[26] HOLLANDS, T., BALS, C., AUSTREGESILO, H., Simulation of QUENCH-L0 with ATHLET-CD, International
Quench Workshop, November, (2011).
28
QUALIFICATION OF FRAPCON/FRAPTRAN FOR LOCA FUEL BEHAVIOUR
MODELLING AND SAFETY EVALUATION
J. ZHANG, A. DETHIOUX, T. DRIEU, C. SCHNEIDESCH

Tractebel (ENGIE),
Boulevard Simon Bolivar 34–36, 1000,
Brussels, Belgium
Abstract
As part of the Tractebel contribution to the IAEA FUMAC project, Tractebel has used the updated FRAPTRAN-TE-1.5
code for simulation of selected Halden LOCA tests IFA-650.9 and IFA-650.10, together with the improvement in the thermal
hydraulic modelling by using the imposed thermal hydraulic boundary conditions from SOCRAT calculations and in the
thermal boundary conditions (axial power profile, plenum temperature). In particular, the impacts of the model improvements
such as the Quantum Technologies’ axial relocation model and errors corrections in the adapted FRAPTRAN-TE-1.5 version
on the calculation results were identified and discussed. In addition, the statistical uncertainty and sensitivity analysis has been
performed on the FRAPTRAN-TE-1.5 modelling of the selected Halden LOCA test IFA-650.10, which helped the
identification of significant input parameters for LOCA fuel Behaviour modelling. The final objective is to apply the qualified
fuel rod transient analysis codes FRAPCON/FRAPTRAN to develop an efficient methodology for assessing the performance
and quantifying the margins for advanced technology fuel (ATF) designs under design basis accident conditions (in particular
LOCAs).
1. INTRODUCTION
As the owner’s engineer for the utility (Electrabel) that owns and operates 7 nuclear power plants (NPPs) in
Belgium, Tractebel needs to qualify and apply fuel rod codes for fuel behaviour modelling and safety evaluation
during normal and off normal conditions. For this purpose, Tractebel has been working on the qualification of the
FRAPCON and FRAPTRAN codes since 2010 [1]. This has been achieved partly by participation in international
benchmarks (such as the IAEA’s FUMEX-III and FUMAC coordinated research projects and the OECD/NEA’s
fuel rod codes RIA benchmarks), performing selected benchmark exercises using FRAPCON and FRAPTRAN
codes and sharing experience and information with other participants [1, 2].
The FUMAC (Fuel Modelling in Accident Conditions) project has been launched by IAEA as a new Co-
ordinated Research Project (CRP) in 2014–2017 [3]. It is a continuation of the previous FUMEX-III project with
the focus on the modelling of fuel behaviours in design basis and severe accidents, in particular, Loss-Of-Coolant
Accidents (LOCAs), which are of particular interests after the Fukushima accidents.
Tractebel participation in the FUMAC project consists partly in simulation of the integral single rod LOCA
tests IFA-650.9 [4] and IFA-650.10 [5] performed at the Halden reactor, using the USNRC fuel performance
codes, FRAPCON-3.5 [6, 7] and FRAPTRAN-1.5 [8, 9]. In particular, following the scoping studies [10], the
Quantum Technologies’ axial fuel relocation model [11] and errors corrections [12] were implemented in the
adapted FRAPTRAN-TE-1.5 version [13, 14], and the impacts of the model improvements on the calculation
results were identified and discussed. In addition, the statistical uncertainty and sensitivity analysis has been
performed on the FRAPTRAN-TE-1.5 modelling of the selected Halden LOCA test IFA-650.10.
The final objective is to qualify FRAPTRAN for simulating relevant behaviours of high burnup fuel under
LOCAs, in order to use the code for assessment of margins regarding to the new USNRC acceptance criteria for
LOCA (10CFR50.46c), for Belgian operating nuclear power plants (NPPs). For this purpose, an efficient
methodology will be developed in the future project for assessing the performance and quantifying the margins
for advanced technology fuel designs under design basis accident conditions (in particular LOCAs).
This paper will summarize the contribution of Tractebel to the IAEA’s FUMAC project and present the
proposal for future development.
29
2. MODEL IMPROVEMENTS IN FRAPTRAN-TE-1.5
The FRAPCON-3.5 [6] and FRAPTRAN-1.5 [8] fuel rod performance and transient analysis codes have
been developed in the framework of reactor safety research programme conducted for the U.S. Nuclear Regulatory
Commission (NRC) by Pacific Northwest National Laboratory (PNNL). FRAPCON-3.5 analyses the thermal
mechanical behaviour of LWR fuel rods under steady state and power ramp operating conditions, while
FRAPTRAN-1.5 code analyses the thermal mechanical behaviour of LWR fuel rods under transient and accident
conditions including LOCA and RIA. Experimental data are used for assessing the FRAPCON-3.5 and
FRAPTRAN-1.5 capabilities [7, 9]. The worldwide community also contributes to the development, testing and
assessment of models.
In the FUMAC project, Tractebel intends to improve and use the FRAPTRAN-1.5 code for better simulation
of the fuel behaviours during LOCA, including the axial relocation. TE has reviewed the capability of
FRAPTRAN-1.5 and made the preliminary assessment for LOCA simulation. The preliminary scoping parametric
analyses recommended improving certain models in FRAPTRAN-1.5, in particular the axial relocation model and
the thermal hydraulic boundary conditions [10].
Therefore, Tractebel requested Quantum Technologies to implement their axial fuel relocation model [11]
and error corrections [12] in the Tractebel version of FRAPTRAN-TE-1.5 [13, 14], in order to perform further
model validation.
3. SIMULATION OF THE HALDEN LOCA TESTS IFA-650.9 AND IFA650.10
3.1. Input models and assumptions
Detailed specifications on the Halden LOCA tests IFA-650.9 and IFA650.10 can be found in [4, 5]. In both
FRAPCON and FRAPTRAN input models, most of the data are set according to the specifications of IFA-650
tests (design data, initial data and measured test data) or the code manual recommended values.
In order to simulate the fuel rod conditions before the LOCA tests, FRAPCON input models are built to
generate initial conditions for FRAPTRAN analysis. The refabricated rodlet FRAPCON-3.5 input model was
validated by comparing with available measured data. The refabricated rodlet FRAPTRAN-1.5 initialization file
was the output of FRAPCON-3.5 input model which contained gas data from the initial father rod and subsequent
Fission Gas Release (FGR) history data. The number of moles of new gas mixture and the relative amount of each
gas species in the refabricated rodlet was adapted to match the calculated and measured initial rod internal pressure.
The measured rod power is imposed. The rod axial power profile in the rodlet is nearly symmetric, with an
axial peak to average power factor of 1.04–1.05. The heater is not simulated.
The most important and difficult is the modelling of the test rod thermal hydraulic boundary conditions [10].
As there exist too large uncertainties in the measured inlet coolant temperature and flowrate, and a lack of
validation of the FRAPTRAN built-in heat transfer models, it was recommended to use system thermal hydraulic
codes (such as RELAP5, TRACE) to calculate the realistic T/H boundary conditions for FRAPTRAN.
Within framework of the FUMAC project, IBRAE has used the integral system thermal hydraulic code
SOCRAT (System Of Codes for Realistic AssessmenT of severe accidents) to simulate the IFA-650 Halden
experiments, in order to provide the FUMAC project participants with the thermal hydraulic boundary conditions
for fuel rod code benchmarking [15]. In the current study, the calculated coolant temperatures and coolant-cladding
heat transfer coefficients form SOCRAT are imposed by using the “Heat” option, and FRAPTRAN will calculate
the cladding temperatures. The clad to coolant heat transfer coefficients are determined from the SOCRAT
calculated total heat flux (radiative + convective), namely 𝐻𝑇𝐶 = 𝑄 / (𝑇 – 𝑇 ), for each axial node. For
simplicity, only 3 pairs of the SOCRAT calculated thermal hydraulic conditions are imposed at bottom, mid and
top of the test rod, consistent with the cladding temperature measurements as in the tests.
The rodlet plenum gas temperature as imposed in the external volume (𝑇 ) are taken directly from the
measurement, or from the SOCRAT calculated coolant temperature at the pleneum level.
The father fuel rod and the re-fabricated rodlet have the same number of nodalization: 45 radial nodes and
40 axial nodes both in FRAPCON and FRAPTRAN.
A steady state period of at least 100 s is imposed for the simulations: the LOCA transient is assumed to occur
at 100 s and to end at the scram time (600 s). The maximum time step is 0.1 s during the steady state and 0.001 s
during the transient.
30
3.2. Parametric studies on the FRAPTRAN-TE-1.5 model options
The reference case (P0) considers the default properties and models in FRAPTRAN-TE-1.5, namely:
— The FRACAS-I rigid pellet and cladding deformation models (finite difference model) for fuel rod
mechanical response;
— The BALON2 high temperature clad failure model based on empirical strain and stress limits for the burst;
— The fuel thermal conductivity, thermal expansion, specific heat models;
— The cladding thermal conductivity, thermal expansion, specific heat models;
— The fuel initial radial deformation and relocation due to densification, solid swelling and gas swelling (from
FRAPCON);
— The cladding initial radial deformation (from FRAPCON);
— The Massih model for fission gas release (FGR);
— The gap conductance model due to gap gas conductivity, fuel/cladding emissivity and cladding hardness;
— The Cathcart-Pawel (C-P) model for high temperature oxidation;
— The non-protective initial oxide layer option.
In order to test the model improvements in FRAPTRAN-TE-1.5, the following parametric studies are
performed on the model options:
TABLE 1. PARAMETRIC STUDIES ON FRAPTRAN-TE-1.5 MODEL OPTIONS

Case Clad mechanical Clad burst failure Axial fuel relocation
P0 FRAGAS-I BALON2 (empirical strain and stress limits) Off
P1 FRAGAS-I noball (Instability strain) Off
P2 FRAGAS-I noball (Instability strain) On
P3 FEA (finite element analysis) noball (Instability strain) On
For the cases with the new “axial relocation” model, the following calibrated parameters are used in order to
obtain correct clad ballooning and burst behaviour and fuel packing fraction:
TABLE 2. SUMMARY OF THE AXIAL RELOCATION MODEL CALIBRATED PARAMETERS

Parameters IFA650.9 IFA650.10
Sigburststrain 1 0.45
Sigcladanneal 0.15 0.1
Sigcladyieldstress 1.05 1.30
Packing fraction 0.72 0.72
3.3. Parametric studies results and discussions
The calculated results of the parametric studies with FRAPTRAN-TE-1.5 (P0-P3) are compared with the
available experimental data (Measurement) in Fig. 1 for IFA-650.9, which indicated significant axial fuel
relocation. Fig.1 (a) and (b) show cladding temperatures, (c) and (d) coolant temperatures, (e) plenum temperature,
(f) rod internal pressure, (g) cladding elongation and (h) rod radius.
31
FIG.1. The parametric study results for IFA-650.9.
It can be observed that:
— The measured cladding temperatures at bottom (Fig. 1(a)) and top (b) are well predicted only by using the
FRAPTRAN-TE-1.5 axial relocation model (P2 and P3), the SOCRAT calculated coolant temperatures (Fig.
1(c) and (d)) and heat transfer coefficients. However, without the axial relocation model (P0 and P1),
FRAPTRAN-TE-1.5 fails to predict the decreased cladding temperature at top (b), even though the SOCRAT
calculated coolant temperatures are correct (d).
— The SOCRAT calculated coolant temperatures are quite close to the measured heater temperature at heater
top (d) and the rod plenum temperature (e), but significantly higher than the measured heater temperature at
bottom after 210 s (c). This can be attributed to the mechanistic relocation model used in the SOCRAT
calculation [15].
— The default BALON2 ballooning (with high temperature clad failure criterion based on empirical strain and
stress limits for the burst) (P0) predicts earlier burst than the default deformation model (with burst criterion
32
based on the cladding effective plastic strain) (P1) as shown in the rod internal pressure (f), lower
deformation as shown in the cladding elongation (g) and radius (h).
— The finite element analysis (FEA) model (P3) enables FRAPTRAN-TE-1.5 to better predict the cladding
deformation (g and h).
— The evolution of the rod internal pressure (f) during the cooldown and heatup period and following the burst
is poorly predicted in all cases. Since the plenum temperature is well imposed (e), on can attribute this to the
poorly predicted deformation (g, h) and the possible axial gas transportation from the plenum to the ballooned
and burst region (not simulated in FRAPTRAN-TE-1.5).
— FRAPTRAN-TE-1.5 fails to predict the “double balloons” as observed in the test (h). This may be attributed
to the coarse mesh for the thermal hydraulic boundary conditions. Indeed, with much refined axial meshes
in the test rod, SOCRAT is able to predict the “double balloons” [15].
Similar observations are made for IFA650.10, except that no differences are observed between the calculated
cases cladding temperature at top as there is almost no axial fuel relocation in this case.
It can be concluded that:
— The default mechanical model (FRACAS-I) with ballooning model under predicts the ballooning but predicts
better the rupture behaviour.
— The default mechanical model (FRACAS-I) without ballooning model over predicts the ballooning but
predicts later rupture behaviour.
— The default mechanical model (FRACAS-I) without ballooning but with relocation model predicts better the
cladding temperatures (in particular for IFA-650.9), the ballooning and rupture behaviour (in particular IFA-
650.10).
— Only the FEA model without ballooning but with relocation model predicts better the cladding elongation.
4. STATISTICAL UNCERTAINTY AND SENSITIVITY ANALYSIS ON IFA-650.10
4.1. Specifications
A detailed specification has been made for the uncertainty and sensitivity analysis on IFA-650.10, where
uncertain input parameters in fuel rod manufacturing data, operation and test boundary conditions, physical
properties and key physical models were identified, with their ranges and distributions defined [16].
The probabilistic input uncertainty propagation method is selected due to its simplicity, robustness and
transparency. A total of 200 FRAPTRAN code runs are made, and the input parameters are obtained following a
Simple Random Sampling (SRS) of the defined PDFs uncertainty ranges and distributions and assuming
independence between input parameters. A lower and upper uncertainty bound of the 5% and 95% percentiles at
confidence level higher than 95% are estimated by the 196th and 5th rank, respectively, by using the order statistics
method [16].
Besides statistical uncertainty analysis, a global sensitivity analysis is also performed to get qualitative
insights on the most influential input parameters, using the previously obtained 200 code runs. The identification
is based on selected sensitivity indices and significance thresholds [17].
4.2. Implementation on FRAPTRAN-TE-1.5 using DAKOTA
The DAKOTA (Design Analysis Kit for Optimization and Terascale Applications) tool [18], which is
developed by the Sandia National Laboratory and has been used by Tractebel for uncertainty analysis in multi
physics modelling [10], is used for performing the uncertainty and sensitivity analysis on the FRAPTRAN-TE-1.5
simulation of IFA-650.10. The above specifications were implemented except for some minor adaptations. In
particular, the relocation model related parameters based on the previously calibration for IFA-650.9 (see Table 2)
are kept for IFA-650.10, in order to have single calibrated model parameters for the future applications.
For each uncertain input parameter, the information includes a mean value, a standard deviation and a type
of distribution. In order to avoid unphysical numerical values, a range of variation (lower and upper bounds) is
also provided. The sampling is performed between the upper and lower bounds, i.e. the distributions are truncated.
In order to simplify the current benchmark application, a normal distribution is assigned to all the considered input
parameters. Their standard deviation is taken as the half of the maximum of the absolute value of the difference
between their nominal value and their upper or lower bound for all input parameters.
33
The DAKOTA tool has been used to sample these input parameters and generate corresponding 200 input
models for FRAPTRAN-TE-1.5. After running these 200 cases, the relevant output parameters were collected, the
lower and upper bounds were determined by using the order statistics, and the important input uncertainty
parameters were identified through the partial rank correlation coefficients (PRCC) obtained by the global
sensitivity analysis.
4.3. Uncertainty analysis results and discussion
The FRAPTRAN-TE-1.5 calculation results of the first 200 successful runs are collected, and the 5th and
th
196 ranks are chosen to estimate the upper/lower (95% / 5%) uncertainty bounds (LB, UB) of the output
parameter of interest. The reference case results (Best-Estimate or BE) are also presented, using all input
parameters at their mean (or nominal) values. The calculated reference, upper and lower bounds are compared
with the available measured data (Experimental) in Fig 2(a) for rod internal pressure, (b) cladding outer surface
temperature, (c) for cladding axial elongation, and (e) for rod diameter.
(a) (b)
(c) (d)
FIG.2. Uncertainty analysis results for IFA-650.10 with relocation model.
It can be observed that:

— The BE rod internal pressure agrees remarkably well with the experimental data, which is well bounded by
the uncertainty bands. The pressure is a little underestimated just after the blowdown, but both burst instant
and peak pressure are well predicted. This confirms that the fuel thermal aspects are well modelled by
FRAPTRAN. Since the fuel axial relocation and deformation are known to be much weaker in this test, the
weaknesses in the mechanical deformation models have no impact on the rod internal pressure.
— In the heatup phase, the calculated cladding outer surface temperatures (BE, LB and UB) are above the
experimental data, even when the uncertainties are accounted, due to probably the consideration of axial
relocation in the calculation. Indeed, it is suspected that some small axial relocations occurred during the test
34
but were not detected in the 3 cladding temperature measurements. More precise measurements are
recommended in the future LOCA tests.
— The predicted clad elongation has large initial uncertainty, which nearly disappears with blowdown. The
difference could be due to that the definition of the starting points are different in the calculations and in the
experiment. The measurements are overestimated by far and this confirms the need to work on the
mechanical deformation models. Use of FEA may also improve the prediction.
— The uncertainties on the cladding diameter are rather weak, except for in the bottom part. This means that
there are more phenomena of interest occurring in that bottom part, leading to larger uncertainties. The fact
that the uncertainty analysis does not cover the experimental data confirms that the models used are not
adequate for all phenomena. There is a balloon of the appropriate radial size, but of exaggerated axial size.
This confirms the results on the oversized elongation, but also the choice of the axial relocation model.
4.4. Sensitivity analysis results and discussion
The partial rank correlation coefficients (PRCC) measures the monotonic relation between the output
parameter and the input parameter, while adjusting for the effects of the other input parameters. As the
FRAPTRAN code have very complex models, and there are certain interactions between input parameters, these
coefficients can only be considered as qualitative and relative index for screening the important input parameters.
The PRCC values change during the transient. Based on the PRCCs at their maximum absolute values, the
important input parameters are identified for each output as shown in Table 3, where “H” (green) cell means that
the input parameter has a significant impact on at least one of the related output parameters (PRCC > 0.5), “M”
(orange) cell means a medium effect (0.25 ≥ PRCC ≤ 0.5), and “L” (white) cell means a low effect (PRCC < 0.25).
Note that the significance thresholds of 0.25 and 0.5 are chosen arbitrarily for simplicity, as specific guidance and
interpretation of the significance depends on the number of samples, number of variables, and analysis tolerance.
The following observations can be made:
— The fuel thermal behaviour (TFC, TFO) and rod internal pressure (RIP) are significantly impacted by the rod
gap geometry (fuel diameter and cladding inner diameter), gap gas conductivity, filling gas pressure, plenum
gas temperature, fuel thermal conductivity, heat capacity and axial relocation related parameters (clad yield
stress, fuel fragment packing ratio and clad annealing), test conditions (relative transient power, coolant
temperature, coolant to clad HTC).
— The cladding thermal behaviour (TCI, TCO) and cladding equivalent reacted (ECR) are significantly
impacted by clad thickness, steady state corrosion and high temperature oxidation, fuel heat capacity and
axial relocation related parameters, test conditions (relative transient power, axial power profile, coolant
temperature, coolant to clad HTC).
— The cladding and fuel mechanical behaviour (DCO, CES, ECT and EFT) are significantly impacted by rod
gap geometry, fill gas pressure, gap gas conductivity, fuel/clad related properties and models (fuel thermal
conductivity, densification, swelling, fuel/clad thermal expansion), steady state corrosion and high
temperature oxidation, axial relocation related parameters, test conditions (relative transient power, axial
power profile, coolant temperature, coolant to clad HTC).
It is clear that the clad, gap and fuel geometry, initial states, test conditions (relative transient power, axial
power profile, coolant temperatures and clad-to-coolant heat transfer coefficients) and axial fuel relocation related
parameters (clad yield stress, fuel fragment packing fraction and clad annealing) are important for nearly all output
parameters. This means that a good simulation of the fuel behaviours in Halden LOCA tests need a better
characterisation of the test rods, better measurements or calculations of the thermal hydraulic boundary conditions,
and a realistic axial relocation model.
35
TABLE 3. IMPORTANCE RANKING OF EACH PARAMETER FOR IFA-650.10
Input uncertainty parameter Fuel Thermal Clad Thermal Mechanical Overall
Behaviour + Behaviour + Behaviour
RIP ECR
Cladding outside diameter (mm) L H H H
Cladding inside diameter (mm) H H H H
Pellet outside diameter H M H H
Fuel theoretical density (kg/m3 at 20°C) M M L M
U235 enrichment (%) L L L L
Filling gas pressure (MPa) H M H H
Relative power during base irradiation L L L L
Relative power during test H H H H
Test rod power profile L H H H
Cladding temperature (°C)
Coolant temperature (°C) H H H H
Clad-to-Coolant heat transfer coefficient H H H H
Fuel thermal conductivity model H M M H
Clad thermal conductivity model L M L M
Fuel thermal expansion model L L H H
Clad thermal expansion model L L H H
Fuel densification model M L H H
Fuel solid swelling model M L H H
Fuel gaseous swelling model M L H H
Clad Yield stress H H H H
Fuel heat capacity H H M H
Cladding heat capacity L L L L
Cladding elastic modulus L L L L
Cladding corrosion model during steady state L H H H
operation
Cladding hydrogen pickup fraction during steady L L L L
state operation
L H H H
Cladding oxidation model at high temperature
Thermal conductivity of the oxide layer L L L L

Fission gas release L L L L
Gap gas conductivity H M H H
Fuel/cladding emissivity L L L L
Fuel radial relocation L L L L
Fuel fragment packing fraction H H H H
Cladding strain threshold for fuel mobility L L L L
Cladding Meyer hardness L L L L
Cladding annealing H H H H
Cladding burst criteria L L L L
Cladding burst strain criteria L L L L
Plenum gas temperature [°C] H M H H
36
5. PROPOSAL FOR DEVELOPMENT OF A METHODOLOGY FOR SAFETY EVALUATION OF
ADVANCED TECHNOLOGY FUEL DESIGNS UNDER LOCAS
The final objective is to apply the qualified fuel rod codes FRAPCON/FRAPTRAN to develop a best
estimate plus uncertainty analysis (BEPU) type hot rod LOCA fuel safety evaluation methodology (FSEM), in
order to verify the compliance with the applicable LOCA safety criteria (e.g. the USNRC new LOCA criteria
10CFR50.46c).
In this FSEM methodology, the important phenomena or input parameters that affect the LOCA fuel
behaviour modelling will be first identified by means of the PIRT (phenomena identification and ranking table)
technique, and the FRAPTRAN simulation model will be developed for a Pressurized Water Reactor (PWR) fuel
rod based on the thermal hydraulic boundary conditions (e.g. local coolant temperatures and cladding to coolant
heat transfer coefficients) from the qualified system thermal hydraulic codes (such as RELAP5 or TRACE,
CATHARE, MARS or SPACE), as obtained from the most limiting cases for the final safety analysis report (FSAR)
the design basis LOCA safety analysis.
The uncertainties related to the FRAPTRAN fuel rod simulation model input parameters and the thermal
hydraulic boundary conditions will be identified and defined. Their importance will be confirmed by using the
GSA (global sensitivity analysis) technique. The uncertainties on the important input parameters will be combined
statistically in the FRAPTRAN calculated upper bound values of peak cladding temperature (PCT) and maximum
local oxidation (MLO) or equivalent cladding reacted (ECR), by using the input uncertainty propagation and order
statistics methods (such as the method used in the FUMAC project).
This methodology will be used to quickly assess the LOCA performance and the margins of the advanced
technology fuel (ATF) rod, without the need to perform complex LOCA safety analyses.
6. CONCLUSIONS
As part of the Tractebel’s contribution to the IAEA FUMAC project, two selected Halden LOCA tests IFA-
650.9 and IFA-650.10 have been simulated using the updated FRAPTRAN-TE-1.5, in order to qualify the
FRAPTRAN-1.5 code for LOCA fuel behaviour modelling. In addition, uncertainty and sensitivity analysis has
been performed on the FRAPTRAN-TE-1.5 simulation of the IFA-650.10 test.
It was concluded that:
— The updated FRAPTRAN-TE-1.5 can well simulate the LOCA fuel thermal behaviours as observed in the
Halden LOCA tests IFA-650.9 and IFA-650.10, using adequate thermal hydraulic boundary conditions and
the axial relocation model.
— The updated FRAPTRAN-TE-1.5 does not well simulate the LOCA fuel mechanical deformation as
observed in the Halden LOCA tests IFA-650.9 and IFA-650.10, using the default FRACAS-I model.
— The uncertainty analysis on FRAPTRAN-TE-1.5 simulation of the IFA-650.10 test confirms that the
FRAPTRAN-TE-1.5 thermal models, axial relocation model, the heat transfer boundary conditions and gas
temperature model are adequate for simulating the thermal behaviour of the Halden LOCA tests, and their
impacts can be well considered by statistical uncertainty analysis. However, the mechanical models for the
cladding deformation still need to be improved for simulating the Halden LOCA tests, and the coverage of
the experimental data.
— The sensitivity analysis on the FRAPTRAN-TE-1.5 simulation of the IFA-650.10 test helps to identify the
most influential input fuel rod data, test conditions, physical properties or models. The parameters related to
the axial relocation model (when applied) are also important. However, the situation of strain threshold for
fuel mobility and burst criteria is more dubious, as they have low correlation coefficients, but should be of
importance according to the physics and the modelling. More detailed analysis is needed to examine this,
e.g. by considering the scatter plots or using other sensitivity indices (such as Sobol’s).
In conclusion, substantial progress has been made in achieving the FUMAC project objectives and the agreed
work plan for Tractebel. The following recommendations can be made:
— Further improvements or validation should be made to FRAPTRAN code for LOCA fuel mechanical
behaviour modelling: the ballooning and burst models, the finite element analysis model for cladding, and if
possible the axial gas transportation model.
37
— Further improvements could be made to FRAPTRAN code for LOCA fuel rod to coolant heat transfer
modelling: this may be overcome by coupling with a system thermal hydraulic code, using a refined nodding
scheme.
— The QT’s axial fuel relocation model could be improved by obtaining a unique set of model related
parameters based on all available Halden LOCA tests with significant fuel relocation.
— More detailed uncertainty and sensitivity analysis should be performed after the code improvements.
It is recommended to apply the validated FRAPTRAN code and the statistical uncertainty and sensitivity
analysis method to development of an efficient BEPU type LOCA fuel performance and safety evaluation
methodology for new or advanced technology fuel design assessment and licensing. This would enhance the
sustainability of nuclear technology and ensure the diversification of fuel supply. This development will be
proposed for the new IAEA CRP as a continuation of the FUMAC and ACTOF projects.
ACKNOWLEDGEMENTS
This work was funded by the Nuclear Development Division (NDD) of the ENGIE group as part of the IAEA
Coordinated Research Project Fuel Modelling in Accident Conditions (FUMAC). The authors are also grateful to
Dr. Lars Olof Jernkvist (Quantum Technologies) for his assistance in the implementation of the Quantum
Technologies axial relocation model and other error corrections in the Tractebel version of FRAPTRAN-TE-1.5
code, and to Dr. A.E. Kiselev (IBRAE) for providing the thermal hydraulic boundary condition calculations with
SOCRAT code.
REFERENCES
[1] UMIDOVA, Z., DETHIOUX, A., ZHANG, J., “Qualification Of FRAPCON/FRAPTRAN Codes for Fuel Rod Design
Verification and Fuel Safety Evaluation,” ENS Conference Proceedings TopFuel 2012, Manchester, UK, 2–6 September
2012.
[2] ZHANG, J., UMIDOVA, Z., DETHIOUX, A., “Simulation of fuel behaviours under LOCA and RIA using FRAPTRAN
and uncertainty analysis with DAKOTA”, in “Modelling of Water Cooled Fuel Including Design Basis and Severe
Accidents, Proceedings of a Technical Meeting Held in Chengdu, China, 28 October–1 November 2013”, IAEA-
TECDOC-CD-1775, 2015.
[3] VESHCHUNOV, M., STUCKERT, J., VAN UFFELEN, P., WIESENACK, W., ZHANG, J., “FUMAC: IAEA’s
Coordinated Research Project on Fuel Modelling in Accident Conditions,” ENS Conference Proceedings TopFuel 2018,
Prague, Czech Republic, 30 September–04 October 2018.
[4] BOLE DU CHOMONT, F., “LOCA Testing at Halden, the Ninth Experiment IFA-650.9”, HWR-917, OECD Halden
Reactor Project, October 2009.
[5] LAVOIL, A., “LOCA Testing at Halden, the Tenth Experiment IFA-650.10”, HWR-974, OECD Halden Reactor Project,
December 2010.
[6] GEELHOOD, K.J., LUSCHER, W.G., “FRAPCON-3.5: A Computer Code for the Calculation of Steady-State, Thermal
mechanical Behaviour of Oxide Fuel Rods for High Burnup”, NUREG/CR-7022, Vol. 1, Revision 1, USNRC, May 2014.
[7] GEELHOOD, K.J., LUSCHER, W.G, “FRAPCON-3.5: Integral assessment”, NUREG/CR-7022, Vol. 2, Revision 1,
USNRC, May 2014.
[8] GEELHOOD, K.J., LUSCHER, W.G, CUTA, J. M., “FRAPTRAN-1.5: A Computer Code for the Transient Analysis of
Oxide Fuel Rods”, NUREG/CR-7023, Vol. 1, Revision 1, USNRC, (May 2014).
[9] GEELHOOD, K.J., LUSCHER, W.G., “FRAPTRAN-1.5: Integral assessment”, NUREG/CR-7023, Vol. 2, Revision 1,
USNRC, (May 2014).
[10] ZHANG, J., DETHIOUX, A., OURY, L., SCHNEIDESCH, C., “FRAPTRAN 1.5 Modelling of the Halden LOCA Tests
IFA650.9 and 10: Scoping Parametric Studies and Perspectives”, in Proc. Enlarged Halden Programme Group Meeting
Sandefjord, Norway, 8th–13th May, 2016.
[11] JERNKVIST, L.O., MASSIH, A.R., ALVESTAV, A.: “Axial relocation of fragmented and pulverized fuel and its effects
on fuel rod heat load during LOCAs”, Top Fuel 2015, paper A0059, (2015).
[12] JERNKVIST, L.O., “Observed and corrected errors in source code and algorithms of FRAPTRAN-1.5,” Report TR15-
002V2, Quantum Technologies AB, (March 2016).
[13] ZHANG, J., “Code modification requirements specifications (CMR.S) for extension and adaptation of FRAPTRAN-1.5
for LOCA fuel behaviour modelling,” Internal report FUELROD/4DO/0437004/000/00, Tractebel, Brussels, Belgium,
February 2016.
[14] JERNKVIST, L.O., “The FRAPTRAN-TE-1.5 computer programme”, Report TR16-001, Quantum Technologies AB,
March 2016.
38
[15] KISELEV, A.E., “Short Information on the Results of IFA-650.9, IFA-650.10 and IFA-650.11 Calculations with
SOCRAT code”, Technical Note, version 2, IBRAE, December 2016.
[16] ZHANG, J., “Specifications for Uncertainty Analysis on Modelling of the Halden LOCA Test IFA-650.10”, Internal
report DDN/4NT/0520044/000/00, Tractebel, Brussels, Belgium, July 2017.
[17] MCKAY, M.D., “Sensitivity and uncertainty analysis using a statistical sample of input values,” Ch. 4, Uncertainty
Analysis, (Ronen, Y. Ed.), CRC Press, Florida, USA, (1988).
[18] ADAMS, B.M., Multilevel parallel object oriented framework for design optimization, parameter estimation, uncertainty
quantification, and sensitivity analysis: Version 5.4 user's manual. Technical Report SAND2010–2183, Sandia National
Laboratories, Albuquerque, NM, (Nov. 2013).
39
FUEL BEHAVIOUR MODELLING IN ACCIDENT CONDITIONS IN ALCYONE
FUEL PERFORMANCE CODE
A. BOULORÉ, C. STRUZIK, P. GOLDBRONN, I. GUENOT-DELAHAIE, J. SERCOMBE

CEA/DEN/DEC Cadarache,
13108 Saint-Paul-lez-Durance,
France
Abstract
CEA is developing the ALCYONE fuel performance code for PWR fuel in the PLEIADES software environment. It is
dedicated to normal, off-normal and accident conditions such as RIA and LOCA. CEA’s participation to the IAEA FUMAC
CRP led to an improvement of the fuel modelling in LOCA conditions. Specific developments of the fuel performance code
for the LOCA conditions have been done regarding cladding behaviour modelling, fission gas release and stress evaluation in
the pellet before and during the tests. The improved code has been used to simulate some of the experiments of interest of the
FUMAC project (IFA650.10 and Studsvik 192 LOCA test), the paper summarizes the results. For IFA650.10, the cladding
outer temperature profile calculated with the SOCRAT code and provided to the participants of the FUMAC CRP has been
used. The results obtained with ALCYONE are in a good agreement with the experimental data. In terms of uncertainty
quantification, it seems that the uncertainty on the determination of the boundary conditions like cladding outer temperature
results in a large uncertainty on the cladding deformation and the burst time. Recent developments have also been done in
ALCYONE to improve the modelling of fuel behaviour in RIA conditions, in particular about fuel mechanical behaviour and
the consequences on fission gas release.
1. INTRODUCTION
‶ALCYONE [1] is a multi dimensional fuel performance code co-developed in the PLEIADES [2] platform
by the CEA, EDF and FRAMATOME. It is dedicated to the modelling of the in-reactor Behaviour of PWR fuel
rods during normal (base irradiation) and off-normal (power ramps and accidental situations) operating conditions
and incorporates three calculation schemes. A one dimensional reference scheme, based on a one dimensional axi-
symmetric description of the fuel element associated to a discrete axial decomposition of the fuel rod in stacked
independent fuel slices, is used to study the Behaviour of the complete fuel rod [3]. A two-dimensional scheme,
which describes Pellet cladding Interaction (PCI) at the mid-pellet plane of a pellet fragment, is available to assess
precisely stress concentration in the cladding near a pellet crack tip [4]. A three dimensional model of the complete
pellet fragment and overlying cladding is also of interest when detailed studies of PCI at pellet–pellet interfaces
are required [5]. The different schemes use the Finite Element (FE) code Cast3M to solve the thermo mechanical
problem and share the same physical material models at each node or integration points of the FE mesh″ [22]
Figure 1 presents the flows chart of ALCYONE 1 D. On the figure the different tests of convergence loops are
identified in blue, the models in the thermo physical loop are in beige and the other models are in green if they are
calculated before the thermo physical loop or in grey pink if after.
Base irradiation code result has been extensively validated on PWR rods (UO2 and MOX fuel up to 80
GWj/tm, cladding Zry-4 or M5) irradiated in the frame of the French survey programme. Power ramp tests
performed in MTR are used to validate the Behaviour of the fuel rod in case of power transient regime. Calculated
results are confronted to experimental values at different scales, from overall measurements on the whole rod such
as overall fission gas release or geometrical changes to the local measurements on the fuel pellet by SIMS and
EPMA. Moreover, fuel temperature calculations are validated specifically on MTR experiments which present a
thermocouple in the fuel centre [3].
More recently, specific developments, summarized in this paper, have been done in the code in order to
simulate accident conditions such as LOCA and RIA transients. This improved version of the code has been used
in the FUMAC CRP benchmark to simulate two experiments. The paper summarizes the results of these
simulations.
40
FIG. 1. ALCYONE flow chart.
2. ALCYONE MODELLING FOR ACCIDENT CONDITIONS
2.1. Specific developments for LOCA conditions
Up to now, LOCA studies have been performed using 1.5 D scheme. For LOCA transient analysis,
ALCYONE is not able to calculate the cladding temperature evolution with time, so this information has to be
provided as input data. Nevertheless, the coupling of ALCYONE [6] with the system code CATHARE will be
available in a near future. For the LOCA studies point of view, the evaluation of the rod internal pressure is crucial.
To do so, it is necessary to have a good description of the fuel at the end of base irradiation. Then, it is essential to
have a correct evolution of the different components of the free volume (plenum, gap, …) and of the fission gas
Behaviour. On the other hand, the evaluation of the quantity of fuel which could be fractured in fine pieces during
the accident phase (i.e. at least the fuel zone which are restructured (HBS) or in restructuration) is also important.
This description is done by two fission gas models CARACAS [7] or MARGARET [8] (more phenomenological).
Both models describe the fission gas production, release and swelling and also the HBS formation. In the grain,
dissolved gas, nanometric bubbles, precipitated bubbles (up to 0.1µm) are considered; the gas atom diffusion and
bubbles diffusion are modelled. Trapping of gas atom by bubbles is taken into account. On the grain boundary gas
is considered to be precipitated into bubbles and the release could happen by inter-connection of bubbles or fracture
by overpressure on a weakened grain boundary. In each point of the pellet, the fraction of HBS is calculated by
the balance between defaults accumulation due to fission spike and thermal recovery. These entire phenomena are
very sensitive to fuel temperature, fission rate and local stresses.
2.1.1. Cladding behaviour modelling
The EDGAR model [9] describes the viscoplastic behaviour of a zirconium alloy tube under a pressure
imposed loading and at high temperature (more than 950 K). This model also gives the phase fractions evolution
(2 phases are considered: alpha and beta and a mixed transition domain, alpha + beta).
For its introduction into ALCYONE, the EDGAR model has been turned into a constitutive law suitable for
a mechanical analysis: basically, the EDGAR model appears as a modified Norton viscoplastic behaviour with the
total strain being split into an elastic and a viscoplastic part. The viscoplastic strain is then traceless which, for
41
small strains, guarantees that the cladding deformation is isochoric (no volume change), as it is classically assumed,
based on experimental facts, in metal plasticity and viscoplasticty.
The mechanical behaviour is expressed in the logarithmic strain framework proposed by Miehe et al [11].
Integration of this formalism in ALCYONE is discussed in a dedicated paper [10]. Once those changes were made,
the EDGAR experiments were simulated using ALCYONE to validate this implementation.
2.1.2. Stress distribution in the pellet
The 1D code generally considers that for a solid pellet, the radial displacement of the central point is not
authorised. This hypothesis has no impact in hot conditions during base irradiation or power ramp regime. Indeed,
the radial stress is in a compressive state in the whole pellet, due to the thermal gradient and that is all the truer
when the fuel to cladding gap is closed.
In case of LOCA, the thermal gradient is significantly lower, and the gap is calculated open (even if at very
high burnup, the gap, strictly speaking, is closed between cladding and fuel due to internal zirconia layer, but a
reopening occurs inside the pellet at about hundred microns from the edge). Then the calculated radial stress tends
to be in a tensile state in the central part of the pellet if the hypothesis of the central node blockage is done. This
induces wrong input data for fission gas behaviour modelling. We have then decided to consider that the central
point of the pellet is able to move radially.
Figure 2 presents stress distribution in the pellet when the gap is open in the cold conditions at the end of a
base irradiation with the two hypotheses. During a LOCA, stresses distribution is likely to be in the same shape,
as the thermal gradient remains low in the fuel pellet. Nevertheless, the actual value of the stress level depends on
the fission gas swelling contribution and then on the base irradiation power history.
FIG. 2 : Hydrostatic stresses distribution in the pellet at room temperature.
2.1.3. Fission gas release
The fission gas model used in the ALCYONE calculations is CARACAS [7], which considers the following
fission gas populations:
— Nanometric intragranular bubbles;
— Precipitated intragranular bubbles;
— Intergranular bubbles;
— Rim structure formation and rim bubbles evolution.
This model is validated in base irradiation conditions and power ramp conditions for UO2 up to 70 GWd/tM.
During the first part of the LOCA transient, fuel temperature remains low enough to affect only the
intergranular gas. A specific criterion for intergranular fracturation initiation has been developed. It is based on
the evaluation of stress applied to grain boundary resulting from macroscopic hydrostatic stress and from stress
induced by over pressurized intergranular bubbles (tensile stress). If the tensile stress on the grain boundary
42
induced by intergranular bubbles is higher than the grain boundary yield stress, partial or total fracturation of the
grain boundaries is supposed to be possible. This allows fission gas release.
We also assume that grain boundary yield stress is reduced by irradiation (fission gas atoms dissolved on the
grain boundary).
This criterion depends primarily on the amount of fission gas precipitated into bubbles after base irradiation
(thus before transient), and thus on the initial fuel state and the fission gas distribution between the different
populations.
It must be mentioned that this model has been developed to make most of the intergranular gas available for
release in transient conditions. Even if the model results in a fraction of opened grain boundaries, it must not be
used for now to evaluate a fuel fragment average size. The total fraction of opened grain boundaries being between
10 and 50%, it would result in very fine fragmentation (a few grains) which is not always observed, and particularly
has not been observed in the IFA650-10 experiment.
2.2. Specific developments for RIA conditions
ALCYONE pulse-irradiation simulation capability is based on [12]:

— the solving of the thermal heat balance equation and mechanical equilibrium for the pellet-gap cladding
system in non-steady state conditions;
— the solving of the thermal and mass balance equations for sodium or water coolant in non-steady state
conditions;
— material laws describing the nonlinear mechanical behaviour of irradiated claddings (Zry-4, M5) submitted
to RIA loading conditions;
— the addition, to fuel pellet creep and cracking modelling, of grain boundary cracking modelling;
— the use of a specific hypothesis as regards the release of fission gases of the high burnup structure (HBS)
zone.
Some of these features are summarized in the following sections.
2.2.1. Non-steady water heat and mass transport
‶Solving the heat and mass balance equations requires the estimation of the linear heat rate received by the
water coolant from the fuel rod, based on the heat exchange between the cladding outer surface and the water,
which involves the clad-to-water coolant heat transfer coefficient (HTC) in particular. The HTC is either given
together with the coolant bulk temperature or calculated using ALCYONE's built-in thermal-hydraulics models″.
The model is detailed in [12].
2.2.2. Suitable constitutive law for claddings
Loading conditions during a RIA are quite peculiar, especially in terms of clad temperature and strain rates.
Constitutive laws for Zry-4 [15] and M5 [14] cladding types and applicable to fast transient conditions and high
temperatures representative of RIA spectrum, have been incorporated in ALCYONE [12].
‶The models can account precisely for the impact of temperature, strain rate, and irradiation damage on the
ultimate stress, on the strain hardening exponent (up to uniform elongation) and on the plastic anisotropy of the
material. They are suitable for simulations of unfailed rods. The assumed isotropic elastic deformation is described
by Hooke’s law assuming a temperature-dependent Young’s modulus.″[12]
2.2.3. Suitable constitutive law for UO2 fuel
The model proposed by Salvo [16, 17] to describe the Behaviour of uranium dioxide within a range of
temperatures (1100−1700°C) and strain rates (10-4–10-1/s) representative of RIA loading conditions is used in
ALCYONE. This model consists of a hyperbolic sine model for the creep strain rate, completed by a Drucker –
Prager yield criterion with associated plastic flow to account for the porosity increase induced by grain boundary
cracking. The yield criterion is a temperature dependent function identified from the compression tests performed
on fresh UO2 at high strain rates and high temperatures that showed significant development of grain boundary
cracking. In this case, the samples’ porosity and diameter were found to increase significantly showing that grain
43
boundary cracking proceeds with pore volume increase. The latter is described in the model by the so-called
“plastic” porosity.
‶The creep and grain boundary cracking models are completed by a smeared crack model to describe pellet
cracking in tension. The resulting constitutive law is particularly relevant for RIA loading conditions where biaxial
compression and tensile stress states are commonly encountered.″[12]
2.2.4. Modelling of fission gas release
The fission gas model CARACAS [7] is so far not able to reproduce the FGR under RIA conditions. As this
phenomenon has been evidenced as able, especially for high burnup fuels [18], to contribute to clad straining
during RIAs, in particular through rod internal pressure increase potentially leading to clad ballooning, ALCYONE
related improvement has been done. It aims at relating the FGR to grain boundary cracking and thus to the local
mechanical damage of the oxide fuel generated by excessive compressive stresses. This coupling is based upon
the following assumptions:
— The cracking of the grain boundaries generated by excessive compressive stresses and described by the
plastic porosity (cf. Section 2.2.3) is used by the CARACAS model as the (mechanical) main criterion for
the instantaneous release of intergranular gas (i.e. at the grain boundaries) contained in the fuel pellet rings
where the criterion is met.
— In case of high burnup fuel where a HBS (characterized by an extremely fine grained structure and an
increased quantity of gases at the grain boundaries) has developed during base irradiation in the peripheral
regions of the pellet, a temperature criterion is adopted to be consistent with annealing tests which showed
FGR from the HBS at temperatures as low as 900 K [19]. In case the temperature threshold is reached, all
the fission gas content of the HBS (inter- and intragranular gas) is released instantaneously. The Intragranular
gas of the HBS is assumed to be released instantaneously because the grain size is too small for diffusion to
be rate-limiting.
— The plastic porosity which models the deformation induced by grain boundary cracking is added to the strains
in the thermo mechanical computation scheme of ALCYONE and contributes therefore to the clad loading.
3. SUMMARY OF THE RESULTS OBTAINED IN THE FUMAC CRP
3.1. IFA650.10
This experiment presented in the HWR report [20] has been performed in the IFA650 device. The whole
instrumentation of the experiment gives a large amount of valuable data which are used to build the input data for
calculation. Nevertheless, it remains difficult to evaluate the temperature axial profile of the cladding, even with
the use of fine thermohydraulic modelling.
About the cladding outer temperature profile during the experiment, two different sets of assumptions have
been made:
— The temperature difference between the coolant and the cladding is proportional to the fissile power. Then,
available measurements at two levels of the fissile rod (TCC1 (95 mm/bfc), TCC2 & TCC3 (360 mm/bfc))
allow the cladding temperature calculation for the whole rod and it is exact at the two TC levels. This set is
mentioned as “ALCYONE boundary conditions” in the following;
— The cladding temperature profile provided by SOCRAT calculations.
The considered creep behaviour and phase transformation laws are the ones for “fresh” material. For the
determination of the uniform strain limit at failure, we consider the value recommended for irradiated material.
The calculation is performed considering transient condition to evaluate the thermal exchanges. Cladding
burst is not calculated directly by ALCYONE but determined by post processing the calculated internal pressure,
cladding radius and cladding temperature.
The possibility of the grain boundary fracturation due to the gas bubble action is tested.
Figs 3 and 4 present the evolution of axial cladding elongation and pressure. Up to burst time, the agreement
between the experiment results and calculation is rather good, but both boundary conditions sets slightly delay the
burst time compared to the experiment. This result validates the hypothesis of perfect communication of the gases
between the plenum and the free volumes of the rod for this case (fast plenum depressurization).
44
FIG. 3. Cladding elongation evolution.
FIG. 4. Rod pressure evolution.
Figure 5 presents the cladding diameter calculated at burst time. Two times are considered: the experimental
burst time and the calculated one. It appears that the tendency is correct (calculated diameter is slightly
underestimated compared to the two measured diameters). With the SOCRAT thermal hydraulics boundary
conditions, the rod diameter is underestimated at experimental burst time. Also, a little depletion of the rod profile
is calculated at 2502 mm, and this result has to be related to the provided boundary conditions where this little
depletion exists.
45
FIG. 5. Cladding diameter at burst time.
Fuel fragmentation due to the over pressurized intergranular bubbles occurs in the simulation in the HBS
zones (outer parts of the pellet) and in some central parts. Fuel fragmentation also occurs after the cladding burst
which leads to a fast decrease of the inner pressure of the rod.
The amount of gas released during the test is about 8 cm3, and it has a negligible effect on the evolution of
the pressure during the test.
The parameters of the fragmentation model still have to be calibrated, to better assess the quantification of
fuel fragmentation after burst. Results of this fragmentation model will provide inputs for a proper fuel fragment
relocation model.
3.2. Studsvik 192 test
Same as for IFA650.10 experiment, the outer cladding axial temperature profile is imposed (given in the data
sheet) in the data file. So is the initial inner pressure in the rodlet.
The plenum temperature is imposed during the experiment. All the different components of the free volume
of the rodlet are assumed connected (plenum, gap, open porosity) so that the pressure in the rod is uniform during
all the experiment.
Figure 6 presents the calculated evolution of the pressure in the rodlet and the cladding temperature evolution
during the test. The calculated burst time is slightly delayed compared to the experiment (about 20 seconds later).
In the calculation, burst occurs for a cladding temperature higher of 100°C than the temperature at experimental
burst. This can be explained by the fact that the real cladding of the experiment is irradiated ZIRLO, whereas we
use the mechanical Behaviour law of the fresh Zry-4.
46
FIG. 6. Rod pressure evolution (Studsvik 192 test).
But the cladding profile calculated just before burst is in good agreement with the measured cladding profiles
after the test (Fig. 7).
FIG. 7. Cladding profile after test (Studsvik 192 test).
Fuel fragmentation before burst occurs only in high burnup structure zones (edge of the pellet). But after
burst, because of the fast decrease of the pressure in the rod, intergranular fragmentation occurs in HBS zones and
central zones of the pellet as shown on Fig. 8.
The fragmentation criterion is based on the calculation of the stress applied on the grain boundary due to the
overpressurized intergranular bubbles. So, it is closely related to the amount of gas on the grain boundary (in
47
bubbles in the CARACAS model) at the end of the steady state irradiation and to the fraction of bubbles
interconnected to the free volume of the rod.
But as the fission gas model is based on the description of an average grain of the microstructure with
spherical shape, the fractured grain boundary ratio cannot be used to quantify a fuel fragment size. In its current
development state, the “fragmentation model” can only be used as a fission gas release mechanism.
FIG. 8. Fuel fragmentation and fission gas distribution (Studsvik 192 test).
4. CONCLUSION
Some improvements have been done in the ALCYONE fuel performance code to better assess the fuel
Behaviour in LOCA conditions. Specific developments have also been done to better assess the behaviour of fuel
rod in RIA conditions.
In the framework of the FUMAC CRP, the simulation of IFA650.10 and Studsvik 192 test have been done
successfully until the cladding burst with ALCYONE. The calculated results are in relative good agreement with
the experimental data, especially if we consider the uncertainties on boundary conditions. Uncertainty
quantification and sensitivity analysis, performed using the URANIE software [21], not reported in this paper,
showed that for example burst time cannot be estimated more precisely than ± 10−15 seconds. Improvement in
the determination of the boundary conditions of such experiments is strongly required.
For future work, it would be interesting to integrate a fuel fragmentation/relocation model in ALCYONE in
order to be able to perform the simulation until the end of the experiment. Cladding oxidation during the
experiment has also to be represented.
REFERENCES
[1] MARELLE, V., et al., “Thermo mechanical modeling of PWR fuel with ALCYONE’. in proceedings of Top Fuel 2011,
Chengdu, China, (2011).
[2] MARELLE, V., et al., “New developments in ALCYONE 2.0 fuel performance code.” in proceedings of Top Fuel Boise
Idaho USA, (2016).
[3] STRUZIK, C., et al., “Validation of fuel performance CEA code ALCYONE, scheme 1D, on extensive data base.” in
proceedings of Top Fuel 2012, Manchester, United Kingdom of Great Britain and Northern Ireland, (2012).
[4] SERCOMBE, J., et al., Power ramped cladding stresses and strains in 3D simulations with burnup-dependent pellet-clad
friction, Nuclear Engineering and Design 242 (2012) 164.
48
[5] MICHEL, B., et al., 3D fuel cracking modelling in pellet cladding mechanical interaction, Engineering Fracture
Mechanics 75 (2008) 3581.
[6] MARELLE, V., Validation of PLEIADES/ALCYONE 2.0 fuel performance code. WRFPM2017, Jeju Island Corea,
(2017).
[7] JOMARD, G., et al., “CARACAS: an industrial model for the description of fission gas Behaviour in LWR-UO2 fuel.”
in Proceedings WRFPM 2014, Sendaï (Japan), (2014).
[8] NOIROT, L., et al., MARGARET: A comprehensive code for the description of ﬁssion gas behaviour, Nuclear
Enginering and design 241 (2011).
[9] FORGERON, T., et al., “Experiment and modeling of advanced fuel rod cladding Behaviour under LOCA conditions:
Alpha-beta phase transformation kinetics and EDGAR methodology”. Zirconium in the Nuclear Industry: Twelfth
International Symposium. ASTM International (2000, January).
[10] HELFER, T., Extension of monodimensional fuel performance codes to finite strain analysis using a Lagrangian
logarithmic strain framework, Nuclear Engineering and Design 288 (2015) 75–81
[11] MIEHE, C., APEL, N., LAMBRECHT, M., Anisotropic additive plasticity in the logarithmic strain space: modular
kinematic formulation and implementation based on incremental minimization principles for standard materials,
Computer Methods in Applied Mechanics and Engineering 191 (2002) 47–48.
[12] GUENOT-DELAHAIE, I., et al., Simulation of RIA transients on UO2-M5 fuel rods with ALCYONE V1.4 fuel
performance code”. Nuclear Engineering and Technology 50 (2018) 268–279.
[13] GOLDBRONN, P., et al., Avancées de la simulation du comportement du combustible nucléaire en 3D et en transitoire
rapide, Congrès français de mécanique, Bordeaux, France (2013).
[14] MOAL, A., et al., Nucl. Eng. Des. 280 (2014).
[15] LE SAUX, M., et al., J. Nucl. Mat. 378 (2008) 60.
[16] SALVO, M., et al., J. Nucl. Mat. 456 (2015) 54.
[17] SALVO, M., et al., J. Nucl. Mat. 460 (2015) 184.
[18] LEMOINE, F., J. Nucl. Mat. 248 (1997) 238.
[19] HIERNAUT, J-P., J. Nucl. Mat. 377 (2008) 313.
[20] LAVOIL, A., HWR 974, LOCA testing experiment at Halden (IFA650.10).
[21] GAUDIER, F., “URANIE : The CEA/DEN Uncertainty and Sensitivity Platform”. Procedia Social and Behavioural
Sciences 2 (2010) 7660–7661.
[22] SERCOMBE, J., MASSON, R., HELFER, T., Stress concentration during pellet cladding interaction: Comparison of
closed-form solutions with 2D (r, θ) finite element simulations, Nuclear Engineering and Design 260 (2013) 175-187.
49
QUALIFICATION OF TRANSURANUS MODELS FOR MIXED CORE FUEL
BASED ON THE FUMAC OUTCOME
M. IEREMENKO, I. OVDIIENKO
State Scientific and Technical Centre for Nuclear and Radiation Safety,
Kyiv, Ukraine
Abstract
This paper presents the results of TRANSURANUS [1–2] code testing for modelling the Behaviour of WWER nuclear
fuel in LOCA accident conditions. The code was tested using data of the IAEA Coordinated Research Project Fuel Modelling
in Accident Conditions (FUMAC, [4]). Fuel pins of Westinghouse and TVEL design are currently relevant for Ukrainian NPPs.
Fuel from both vendors is presented in the FUMAC project. A part of the experimental data (MTA-EK data, IFA 650.10,
IFA 650.11 and Studsvik 192&198) was simulated at SSTC NRS. Some of these data sets and KIT QUENCH-L1 set were
calculated by other teams using the TRANSURANUS code (INRNE, Bulgaria; JRC, Germany). TRANSURANUS code
demonstrated good capabilities for predicting the Behaviour of nuclear fuel rod cladding. The predicted cladding geometry and
time of burst for both general types of cladding (Westinghouse and TVEL) show good correlations with the experimental data
for such regimes. In addition, experimental data of the FUMAC project contained the results of post irradiation measurements
after operation in a commercial reactor. These data were used to test TRANSURANUS code capabilities for modelling the fuel
rod Behaviour in the core under burnup.
1. INTRODUCTION
The diversification of nuclear fuel is currently ongoing at Ukrainian operating NPPs in accordance with the
European Energy Security Strategy. Nuclear power supplies almost half of Ukraine’s electricity. Therefore, the
diversification of fuel supplies is a strategic area for ensuring the uninterrupted operation of nuclear facilities and
is one of the main aspects of energy security in Ukraine. As of today (May 2019), six of the 13 WWER-1000 units
have Westinghouse FAs (WFAs) in the core (the core of one unit is fully loaded with WFA and 2/4–3/4 of the
core is loaded by WFAs at other five units). Moreover, the operator of Ukrainian NPPs plans to extend
Westinghouse FA supplies also to Rivne NPP unit 3. The potential use of Westinghouse fuel for the WWER-440
reactor is also under consideration (outcome of the ESSANUF project [3]).
Significant efforts have been made to license alternative fuel designs for WWER NPPs in Ukraine. SSTC
NRS as the technical support organization of the Ukrainian regulatory authority takes part in the licensing process
for the implementation of new FA types. The TRANSURANUS code is used for independent confirmatory
calculations of fuel pin behaviour characteristics. Considering this fact, the adaptation and qualification of
TRANSURANUS models to new fuel types are a necessary part of up-to-date calculations, including fuel pin
modelling in LOCA. FUMAC project results were used to achieve this goal.
FUMAC — Fuel Modelling in Accident Conditions — IAEA’s Coordinated Research Activities (CRP),
Code T12028. The main objective of this project is to support the Member States in their efforts to analyze and
better understand the Behaviour of water cooled power reactor fuel in LOCA conditions by sharing experimental
data and best practices in the application of fuel modelling computer codes. FUMAC covers a wide range of fuel
characteristics and vendors. Fuel pins of Westinghouse and TVEL designs are relevant now for Ukrainian NPPs.
Fuel from both vendors is included into the FUMAC project.
According to the FUMAC project results, the TRANSURANUS code demonstrated a good capability to
predict the behaviour of nuclear fuel rod cladding. The predicted cladding geometry and time of burst for both
general cladding types (PWR and WWER) show good correlation with experimental data for such regimes.
SSTC NRS improved the capabilities for modelling nuclear fuel rods in LOCA conditions for both types of
cladding: typical WWER cladding and cladding for Westinghouse fuel design.
2. MODELLED CASES
An objective of the FUMAC project is to collect well checked results of accident simulation experiments
and distribute them among the participants. The following data were selected for this purpose:
— Ballooning and burst tests performed at MTA EK predecessor including 31 experiments [4], [5–8];
— Halden Reactor test series IFA-650 with fresh PWR rod, IFA-650.1 [9] and IFA-650.2 [10];
50
— Halden Reactor test series IFA-650 with irradiated PWR and WWER rod, IFA-650.9 [11], IFA-650.10 [12]
and IFA-650.11 [13];
— Two out-of-reactor LOCA simulation tests performed by Studsvik Nuclear AB, Sweden, under contract with
the U.S. Nuclear Regulatory Commission (U.S. NRC), NRC-192&198, [14–20];
— QUENCH-LOCA-1 test (QUENCH-L1, [21]) with electrically heated bundle (tantalum (Ta) heaters inside
each of the 21 rods) performed by KIT in accordance with a temperature/time scenario typical for LOCA.
This data list provided in the FUMAC project is incomplete. This list is formed in terms of the ability to test
the TRANSURANUS code. Additionally, the FUMAC project contains experimental data simulating fuel rod melt
(CORA-15). However, this is beyond the TRANSURANUS code capabilities and is not covered in this paper.
Table 1 presents common experimental data available in the FUMAC project and suitable for
TRANSURANUS code validation. The calculation results obtained by other codes are available (code-to-code
comparisons, see Table 2) in addition to the experimental data for TRANSURANUS validation.
TABLE 1. COMMON CHARACTERISTICS OF EXPERIMENTAL SETS
MTA IFA- IFA- IFA- IFA- IFA- NRC NRC QUENCH-

EK 650.1 650.2 650.9 650.10 650.11 192 198 L1
Zry-4
Cladding Zry-4 Zry-4 Zry-4 Zry-4 E110 ZIRLO ZIRLO Zry-4
duplex
Burnup, MWd/kgU 0 0 0 89.9 61 56 68.2 55 0
Westing
Westing house,
Vendor — — — AREVA EDF TVEL —
house IFBA
rod
700- 800–
T, C 1050 1100 850 935 1185 1185 1100
1200 1100
Clad elongation Y Y Y Y Y Y Y Y Y
Ballooning size Y Y Y Y Y Y Y Y Y
Pressure Y Y Y Y Y Y Y Y Y
Time of burst Y N Y Y Y Y Y Y Y
Position of burst — N Y Y Y Y Y Y Y
SSTC, SSTC, SSTC,

Modelling by
— — — JRC, JRC, SSTC SSTC JRC
TRANSURANUS JRC INRNE INRNE
TABLE 2. AVAILABLE DATA FOR CODE-TO-CODE COMPARISONS

IFA-650.9 IFA-650.10 IFA-650.11 NRC 192 QUENCH-L1
Integral fission gas release Y Y Y Y —
Cladding oxide thickness Y Y Y Y Y
Equivalent cladding reacted Y Y Y Y Y
Diametral gap Y Y Y Y —
Fuel fragmentation volume Y Y N Y —
Hydrogen content in cladding N N N N Y
51
The following experiments were selected for modelling by FUMAC participants:
— Six cases of MTA EK (PUZRY-8, 10, 12, 18, 26 and 30);
— IFA-650.9, IFA-650.10 and IFA-650.11;
— NRC-192;
— KIT QUENCH-L1.
Considering the complexity of IFA-650.9 (very high burnup and complicated T/H conditions, an axial
relocation model is also required), the participants agreed to use IFA-650.10 as the case for the uncertainty and
sensitivity analysis (UASA), see Annex I [4].
All 31 experiments of MTA EK sets and NRC-198 were modeled at SSTC NRS by the TRANSURANUS
code [5].
In addition to the FUMAC data obtained with LOCA experimental sets IFA-650.9, .10, and .11, NRC-192
and 198 contain data obtained by fuel rod re-fabrication: rod geometry, cladding oxide thickness, hydrogen
content.
3. MODELLING RESULTS
3.1. Separate Effect Tests (MTA-EK Burst tests)
The analyses of MTA-EK series tests using the TRANSURANUS (v1m1j11–SSTC NRS and v1m2j17–
JRC) code are presented in this section. The calculation results are compared with the measured data for the
following parameters:
— Time to burst failure;
— Pressure of cladding rupture.
SSTC NRS modelled all experimental data sets (31 experiments). JRC team modeled six experiments.
The results in terms of time to burst for the MTA EK cases are reported in Figure 1. The results in terms of
cladding inner pressure at the time of burst failure are presented in Figure 2. Experimental data are also included
in both figures. Predicted time and pressure of cladding burst show good correlation with the experimental data
for such regimes. Additional results of hoop strain at the axial peak position (burst location) on the cladding outer
surface at the time of burst failure are available in the FUMAC final report [4] and the participant final reports
(Annex II [4]).
5000 5000
JRC, FUMAC cases

JRC, FUMAC cases SSTC NRS, FUMAC cases
SSTC NRS, FUMAC cases SSTC NRS, not FUMAC cases
4000 4000
SSTC NRS, not FUMAC cases Experiment
Time burst, calculations, s
3000 3000
Time burst, s
2000 2000
1000 1000
0 0
0 1000 2000 3000 4000 5000 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31
Time burst, experimental, s Num of experiment
FIG.1. Calculation result. Time to burst failure.
52
11 12
JRC, FUMAC cases
10 11 SSTC NRS, FUMAC cases
SSTC NRS, not FUMAC cases
10 Experiment
9
JRC, FUMAC cases
SSTC NRS, FUMAC cases 9
Inner pressure, calculations, MPa
8
SSTC NRS, not FUMAC cases
8
7
Inner pressure, MPa

7
6
6
5
5
4
4
3
3
2 2
1 1
0 0
0 1 2 3 4 5 6 7 8 9 10 11 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31
Inner pressure, experimental, MPa Num of experiment
FIG.2. Calculation result. Pressure of cladding burst failure.
3.2. Halden LOCA tests IFA-650.10
The computer analyses of the considered Halden IFA 650.10 tests were carried out in two steps. Pre-
irradiation of a full (or father) fuel rod that was later re-fabricated into a test rodlet was modelled in the first step
using TRANSURANUS (v1m1j17). The results calculated by TRANSURANUS defining pre-test conditions for
the test fuel rodlet were used as input to the second analysis step [5]. Spent fuel storage in a water pool hardly
influences the characteristics of cladding and fuel pellets.
The input data for the test rodlet were developed using full (father) rod pre-irradiation history. The following
parameters for the father rod are used: linear power, in-cladding pressure, cooling temperature. Using the restart
option of the TRANSURANUS code, change of in-cladding gas composition before LOCA experiment was
simulated (not for all calculations, a part of tasks was performed with in-cladding gas pressure as boundary
conditions [5]).
It should be noted that no computer programme was used to calculate transient thermal hydraulic boundary
conditions required for fuel rod analyses with TRANSURANUS. These boundary conditions were derived from
measured temperatures or from pre-calculated data distributed by FUMAC participants (calculated by the
SOCRAT code).
Figure 3(a) shows the calculated and measured evolution of rod plenum gas pressure during Halden IFA-
650.10 test and diameter profiles at the end of the LOCA test. The calculated gas pressure is a little higher than
measurements for results of all teams. The calculated diameter profiles are close to measurements. Additional
results of the pre-irradiation history and LOCA transient (total fission gas release, clad axial elongation, cladding
oxide thickness, equivalent cladding reacted, etc.) are available in the FUMAC final report [4] and the participant
final reports (Annex II [4]).
10 15
9
14 INRNE
JRC
8 SSTC NRS
Measured
INRNE 13
7
Pressure (MPa)
JRC
SSTC NRS
6 Measured
diametr, mm
12
5
11
4
3 10
2
9
1
0 8
0 100 200 300 400 500
0 50 100 150 200 250 300 350 400 450 500 550 600
Length, mm
Time (s)
a) b)
FIG.3. Calculated evolution of rod plenum gas pressure (a) and diameter profiles (b) in comparison with measurements for
IFA-650.10 test rodlet.
53
3.3. Halden LOCA tests IFA-650.11
As for IFA 650.10, the computer analyses of the considered Halden IFA 650.11 tests were carried out in two
steps: pre-irradiation of a full (or father) fuel rod and test rodlet simulations at LOCA.
Figure 4(a) shows the calculated and measured evolution of rod plenum gas pressure during Halden IFA-
650.11 test and diameter profiles at the end of LOCA test. The calculated gas pressure is a little lower than
measurements for results of all teams. The calculated diameter profiles are close to measurements. Additional
results of the pre-irradiation history and LOCA transient (total fission gas release, clad axial elongation, cladding
oxide thickness, equivalent cladding reacted, etc.) are available in the FUMAC final report [4] and the participant
final reports (Annex II [4]).
10
INRNE 12
9 JRC
SSTC NRS
Measured
8
7 11
Pressure (MPa)
diametr, mm
5 10
3 9
2 INRNE
JRC
1 SSTC NRS
8
Measured
0
0 50 100 150 200 250 300 350 400 0 100 200 300 400 500
Time (s) Length, mm
a) b)
IFA-650.11 test rodlet.
3.4. Studsvik LOCA test NRC-192
As for IFA 650.10 and 11, the computer analyses of the considered NRC-192 tests were carried out in two
steps: pre-irradiation of a full (or father) fuel rod and test rodlet simulations during LOCA. Details and features of
pre-irradiation modelling by SSTC NRS are presented in [5].
Figure 5(a) shows the calculated and measured evolution of rod plenum gas pressure during NRC-192 test
and diameter profiles at the end of LOCA test. The calculated gas pressure is a little higher than measurements.
The calculated diameter profiles are smaller than measurements. Additional results of the pre-irradiation history
and LOCA transient (total fission gas release, clad axial elongation, cladding oxide thickness, equivalent cladding
reacted, etc.) are available in the FUMAC final report [4] and the participant final reports (Annex II [4]).
Time (s)
-20 0 20 40 60 80 100 120 140 160
13 17
12
16
11
SSTC NRS
10 Measured
15
9
Pressure (MPa)
8 14
diametr, mm
7
13
6
5 12
4 SSTC NRS
Measured 11
3
2
10
1
0 9
1060 1080 1100 1120 1140 1160 1180 1200 1220 1240 0 50 100 150 200 250 300
Time (s) Length, mm
a) b)
NRC-192 test rodlet.
54
3.5. Studsvik LOCA test NRC-198
As for the above cases, the computer analyses of the considered NRC-198 tests were carried out in two steps:
pre-irradiation of a full (or father) fuel rod and test rodlet simulations during LOCA. The details and features of
pre-irradiation modelling by SSTC NRS are presented in [5].
It should be noted that Westinghouse IFBA fuel rod was used for NRC-Studsvik 198 test. Fuel pellets of this
rod were coated by burnable absorber: zirconium diboride ZrB2. However, the characteristics of this burnable
absorber were not present in 198 test specifications [20]. This burnable absorber in the form of a thin layer works
through the (n, ) reaction; e.g. helium is produced and considered in the gas composition of free volume of fuel
rod. As a result, the in-cladding pressure depends on the characteristics of the zirconium diboride layer (thickness
or mass per fuel length), and so is cladding geometry for test rodlet. Figure 13(b) in [5] shows the influence of
ZrB2 thickness under in-cladding pressure during full rod pre-irradiation history. The thickness of ZrB2 multiplied
or divided by 2 results in the in-cladding pressure change of 20%.
Figure 6(a) shows the calculated and measured evolution of rod plenum gas pressure during NRC-198 test
and diameter profiles at the end of LOCA test. The calculated gas pressure is a little higher than measurements.
The calculated diameter profiles are close to measurements.
Time (s)
0 20 40 60 80 100 120 140
13 14
12
11
13
10
9
Pressure (MPa)
SSTC NRS
8 12 Measured
diametr, mm
5 11
4 SSTC NRS
Measured
3
10
2
0 9
1060 1080 1100 1120 1140 1160 1180 1200 0 50 100 150 200 250 300
Time (s) Length, mm
a) b)
NRC-198 test rodlet.
3.6. KIT QUENCH-L1 bundle test
Analyses of QUENCH-L1 test using the TRANSURANUS (v1m2j17–JRC) code are presented in this
section. The calculation results are compared with measured data for the following parameters:
— Time to burst failure;
— Pressure of cladding rupture.
JRC modelled QUENCH-L1 test.
Figure 7(a) shows the calculated and measured evolution of rod plenum gas pressure during QUENCH-L1
test and diameter profiles at the end of LOCA test. The calculated gas pressure is a little higher than measurements.
The calculated diameter profiles are close to measurements. Additional results of LOCA transient (cladding hoop
stress, cladding oxide thickness, equivalent cladding reacted, hydrogen content in cladding etc.) are available in
the FUMAC final report [4].
55
10 15
8 14
JRC
Measured
7 JRC
Pressure (MPa)
Measured
6 13
diametr, mm
5
4 12
2
11
0
0 50 100 150 200 250 300 10
200 400 600 800 1000 1200
Time (s)
Length, mm
a) b)
QUENCH-L1 test rodlet.
3.7. Common remarks on results
Table 3 presents the predicted failure time differences among the participants with the TRANSURANUS
code (JRC, INRNE and SSTC NRS). TRANSURANUS-predicted time of cladding burst for both general types of
cladding (PWR and WWER) shows good correlations with the experimental data for such regimes.
Table 4 presents differences in the cladding outer diameter after the test among the participants with the
TRANSURANUS code. TRANSURANUS results show good correlation with the experimental data for such
regimes.
The addition data of comparing TRANSURANUS results with the experimental data and code-to-code
comparisons are available in the FUMAC final report [4] and the participant final reports (Annex II [4]).
TABLE 3. BURST TIME. DIFFERENCES WITH MEASURED VALUE
MTA EK QUENCH-
IFA-650.10 IFA-650.11 NRC 192 NRC 198
6 cases L1
<13%
JRC (27% for 6% — — — 6%
PUZRY-18)
INRNE — 1% 12% — — —
<3%
SSTC (19% for 4% 11% 21% 23% —
PUZRY-18)
TABLE 4. CLADDING OUTER DIAMETER AFTER TESTS (FOR MTA EK - BURST PRESSURE).
DIFFERENCES WITH MEASURED VALUE
MTA EK
(burst QUENCH-
IFA-650.10 IFA-650.11 NRC 192 NRC 198
pressure) L1
6 cases
<11%
JRC (31% for 4% — — — 5%
PUZRY-18)
INRNE — 4% 3% — — —
<2%
SSTC (24% for 4% 4% 22% 1% —
PUZRY-18)
56
FUMAC results show some features in modelling fuel rod Behaviour in LOCA by the TRANSURANUS
code:
— Due to some differences in the fuel rod plenum design, test rodlet and fuel rod for commercial nuclear
reactors, the temperature of plenum gas in the TRANSURANUS model is described as boundary conditions
from measurement data or from calculations by other code. Considering this aspect, inn-cladding gas
pressure predicted by TRANSURANUS can be compared with the measured value with some limitations;
— TRANSURANUS model for prediction of fuel for plenum temperature in LOCA mode should be
additionally tested in view of fuel rod design for commercial nuclear reactors;
— Boundary conditions should be predicted by thermal hydraulic codes for modelling of nuclear fuel rods in
LOCA by TRANSURANUS code;
— TRANSURANUS model for Westinghouse IFBA fuel rod was not fully tested due to insufficient data in
FUMAC set.
4. CONCLUSION
TRANSURANUS code demonstrated good capability to predict the Behaviour of nuclear fuel rod cladding.
The predicted cladding geometry and time of burst for both general types of cladding (PWR and WWER) show
good correlation with the experimental data for such regimes.
SSTC NRS improved the capability for modelling the nuclear fuel rod in LOCA for both types of cladding:
typical WWER cladding and cladding for Westinghouse fuel design. The experience and the TRANSURANUS
models obtained during the FUMAC project will be used to perform calculations in the framework of licensing
mixed cores.
The FUMAC project contains experimental data and calculation results obtained by other participants of the
project to continue testing of the TRANSURANUS code.
REFERENCES
[1] TRANSURANUS HANDBOOK. Document Number Version 1 Modification 1 Year 2017 ('V1M1J17'). European
Commission. Joint Research Centre. Directorate G - Nuclear Safety & Security. PO Box 2340. D-76125 Karlsruhe,
(January 2017).
[2] LASSMANN, K., TRANSURANUS: A Fuel Rod Analysis Code Ready for Use, J. Nucl. Mater. 188 (1992) 295.
[3] ESSANUF Project.
http://www.essanuf.eu/
[4] INTERNATIONAL ATOMIC ENERGY AGENCY, Fuel Modelling in Accident Conditions (FUMAC). Final Report
of a Coordinated Research Project CRP T12028 (2014–2018), IAEA – TECDOC -1889, Vienna, (2019).
[5] IEREMENKO, M., Testing of the TRANSURANUS Computer Code by Joint Solution of Test Problems within FUMAC
Project. Annex II to [4] (draft, now preparing to publish)
[6] HÓZER, Z., GYUŐRI, C., HORVÁTH, M., NAGY, I., MARÓTI, L., MATUS, L., WINDBERG, P., FRECSKA, J.,
Ballooning Experiments with VVER Cladding, Nuclear Technology 152 3 (2005) 273–285.
[7] PEREZ-FERÓ, E., GYŐRI, C., MATUS, L., VASÁROS, L., HÓZER, Z., WINDBERG, P., MARÓTI, L., HORVÁTH,
M., NAGY, I., PINTÉR-CSORDÁS, A., NOVOTNY, T., Experimental Database of E110 Claddings Exposed to
Accident Conditions, Journal of Nuclear Materials 397 1–3 (2010) 48–54.
[8] PEREZ-FERÓ, E., HÓZER, Z., Experimental Database of E110 Claddings under Accident Conditions. EK-FRL-2011-
744-01/04 Budapest, (April 2012).
[9] LESTINEN, V., LOCA Testing at Halden, First Experiment IFA-650.1. Technical Report HWR-762, OECD Halden
Reactor Project, (2004).
[10] EK, M., LOCA Testing at Halden, the Second Experiment IFA-650.2. Technical Report HWR-813, OECD Halden
Reactor Project, (2005).
[11] DU CHOMONT, F.B., LOCA Testing at Halden, the Ninth Experiment IFA-650.9. Technical Report HWR-917, OECD
Halden Reactor Project, (2009).
[12] LAVOIL, A., LOCA Testing at Halden, the Tenth Experiment IFA-650.10. HWR-974. OECD HALDEN REACTOR
PROJECT. (December 2010).
[13] LAVOIL, A., LOCA Testing at Halden, the VVER Fuel Experiment IFA-650.11. HWR-976. OECD HALDEN
REACTOR PROJECT (December 2010).
[14] JERNKVIST, L.O., Computational Assessment of LOCA Simulation Tests on High Burnup Fuel Rods in Halden and
Studsvik. Technical Report. March (2017).
57
[15] HELIN, M., FLYGARE, J., NRC LOCA tests at Studsvik: Design and Construction of Test Train Device and Tests with
Unirradiated Cladding Material, Report STUDSVIK/N-11/130, Studsvik Nuclear AB, Studsvik, Sweden (2012).
[16] Studsvik LOCA test machine data sheet.xlsx. FUMAC website shared documents.
https://nucleus.iaea.org/sites/nefw-projects/fumaccrp/Shared%20Documents
[17] FLANAGAN, M., “NUREG-2160: Post test Examination Results from Integral, High-Burnup, Fueled LOCA Tests at
Studsvik Nuclear Laboratory,” U.S. Nuclear Regulatory Commission, (August 2013).
[18] FLANAGAN, M., ASKELJUNG, P., Observations of Fuel Fragmentation, Mobility and Release in Integral, High-
Burnup, Fueled LOCA Tests. Halden Workshop on LOCA. Lyon, France (29–30 May 2012).
[19] NRC-Studsvik LOCA test 192 data sheet. FUMAC’s website shared documents.
[20] NRC-Studsvik LOCA test 198 data sheet. FUMAC’s website shared documents
[21] STUCKERT, J., GROßE, M., RÖSSGER, C., STEINBRÜCK, M., WALTER, M., Results of the LOCA Reference
Bundle Test QUENCH-L1 with Zircaloy-4 Claddings. KIT-SR 7651, Karlsruher Institut für Technologie (2018).
58
MAIN RESULTS OF FUMAC FROM CNPRI AND SENSITIVITY AND
UNCERTAINTY STUDY OF PERFORMANCE OF ATF UNDER TRANSIENT
CONDITION
Y.R. XIE, Q.S. REN, Y.D. ZHANG

China Nuclear Power Technology Research Institute,
Shenzhen, China
Abstract
The safe, reliable and economic operation of nuclear power reactor fleet has always been a top priority for worldwide
nuclear industry. Typically, fuel rod Behaviour in accident conditions is one of the main concerns. After the severe Fukushima
accidents, enhancing the accident tolerance of light water reactor (LWRs) became a hot issue in the world. Particularly, the
fuel materials are expected to improve accident tolerance. Fuel performance is influenced by design parameters, physical
properties and thermal hydraulic condition. In this report, the main result on FUMAC which is in order to better understand
fuel rod Behaviour under LOCA condition and assess the predictive ability of the code, was present. Furthermore, take case
IFA-650.10 for example, sensitivity and uncertainty study is performed to evaluate the ATF fuel system’s (UO2-BeO + ODS
FeCrAl) performance under LOCA condition, and find out an optimal design by using the statistical tool DAKOTA coupling
with fuel performance analysis code.
1. INTRODUCTION
Fuel rod represents the primary and the essential safety barrier for nuclear safety. Safety criterion for accident
condition is of vital importance. Several LOCA experiments have been performed at Halden reactor to reexamine
the current criterion. According to the FUMAC work programme CNPRI was to perform simulations of IFA-650
Halden experiments with FRAPCON/FRAPTRAN to better understand fuel rod Behaviour under LOCA condition
and assess the predictive ability of the code. Main results of FUMAC are presented in this report.
After the sever Fukushima accidents in 2011, the concept of accident tolerant fuel (ATF) has been put
forward with the intention of improving the fuel performance under transient condition while maintaining or
improving the fuel performance during normal operation. Fuel performance is influenced by design parameters,
physical models and thermal hydraulic condition. Different factors have different impacts. Sensitivity and
uncertainty study can offer an efficient approach to evaluate the impacts of different factors and sort out the
important ones. Fuel Behaviour under transient condition is of vital important for the fuel assessment. In this
research, sensitivity and uncertainty study is applied to find out the important factors for typical ATF fuel system’s
(UO2-BeO pellet+ ODS FeCrAl cladding) failure time under LOCA condition and obtain an optimal design.
2. MAIN RESULTS OF FUMAC
Simulations of IFA-650.9/10/11 were performed in FUMAC work programme. The test rods used in IFA-
650.9/10/11 Halden experiments were segments cut from the standard PWR or WWER fuel rod of different burnup.
Main characteristics of the test rods are show in Table 1.
59
TABLE 1. MAIN CHARACTERISTICS OF THE TEST RODS [1–3]
IFA-650.9 IFA-650.10 IFA-650.11
Reactor type PWR PWR WWER

Burnup, MWd/kgU 89.9 61 56
Stack length, mm 480 440 480
Enrichment, wt%U235 3.5 4.49 3.6
Fuel density, % of T.D. 95 95.32 97
Fuel diameter, mm 9.13 8.21 7.55
Fuel length, mm 8 10 10
DX Zr2.5Nb (duplex)
Cladding type Zry-4 E110
Zry-4 (base)
Cladding outer diameter, mm 10.75 9.5 9.13

Cladding inner diameter, mm 9.3 8.36 7.77
Plenum volume, cm3 19 16–17 16–17
Gas pressure (RT), bar 40 40 30
Table 2 and Figs 1−3 illustrate the overall comparison of calculated results and corresponding measured
values in 3 cases. It implies that variation of cladding temperature, cladding outer diameter, rod internal pressure
as well as cladding failure are generally well predicted. The predicted maximum rod pressure and its corresponding
cladding hoop stress as well as cladding failure temperature are close to the measured value. Failure time predicted
by FRAPTRAN is close to the value that it was actually measured in case 650.10 and 650.11. Due to its very high
burnup, significant fuel relocation and the complicated thermal hydraulic boundary condition, the 650.9 case is
rather complex. Consequently, deviations between the predicted and measured results are relatively large in case
650.9.
TABLE 2. OVERALL COMPARISON OF FRAPTRAN CALCULATED RESULTS AND CORRESPONDING

MEASURED VALUES
650.9 650.10 650.11
MEAS. CAL. MEAS. CAL. MEAS. CAL.
Cladding deformation start time, s 106 78 228 211 183 212

Cladding failure time, s 133 87 249 233 207 217
Maximum rod pressure, MPa 7.73 7.08 7.16 7.00 5.57 5.36
Maximum rod pressure corresponding
47.2 47.1 49 48.6 29.5 29.0
hoop stress, MPa
Cladding failure temperature, K 1083 1066.7 1028 944.7 1112 1075.4
The sudden decrease of rod pressure indicates the failure of fuel rod. For the IFA tests, the total free volume
of gas was practically all located in the plenum, outside the heated region [4]. FRAPTRAN can model an external
plenum, and the external plenum gas temperature data is imposed as initial condition. However, the predicted
plenum gas temperature is higher than the measured values. It’s because that in FRAPTRAN, there is 2 options
for defining the plenum gas temperature: assume the gas temperature to be 10°F higher than the axial local coolant
temperature; a more detailed model to concern the interactions between the plenum gas and the top pellet surface,
hold down spring, and cladding wall [5]. In these temperature calculation processes, the input plenum gas
temperature hasn’t been used. The input plenum gas temperature history is only used for defining the sum of
60
volume over temperature terms in gas gap pressure equation. Difference in plenum gas temperature might cause
the deviation of other parameters.
The predicted cladding outside temperature shows similar trend with the measured values. The little
difference can be accounted by the difference in nodalization and the simplification of the models. Besides, using
the thermal properties of light water to replace those of heavy water might have some effect on the results.
The predicted axial variation of cladding outer diameter shows similar trend with the measured one.
Deviation of the ballooning node may be accounted by the difference in nodalization of boundary conditions. In
case 650.9, measured results indicate that cladding ballooning occurred on 2 portions of the rod while the predicted
result show that cladding ballooning occurred only at the upper part of the rod. This difference is due to the
calculation method of FRAPTRAN. Once the ballooning has been predicted in one node, no further strain
calculations would be performed for any nodes, the further calculation focuses only on the ballooning node [5].
Thus, it could only predict the primary ballooning.
FIG. 1. Variation of rod pressure, plenum gas temperature, cladding outside temperature, cladding outer diameter of case
650.9.
61
650.10.
650.11.
62
3. SENSITIVITY AND UNCERTAINTY STUDY OF ATF WITH IFA-650.10
3.1. Choice of input and output of uncertainty study
The uncertain parameters of uncertainty study are the variables with uncertainty, giving the upper bound,
lower bound as well as the distribution. The uncertain parameters can usually be quantified by the probability
distribution function (PDF), which often depends on the engineering experience or experimental results. Moreover,
we can use the correlation coefficient to represent how closely the two variables co-vary.
The ATF fuel system (ODS FeCrAl+UO2-BeO) studied in this report is planning to be applied in the present
commercial pressure water reactor, thus, the parameters of fuel rod configuration and thermal hydraulic condition
of reactor are not considered in this report. As shown in Table 3, 17 uncertain variables are chosen, including 5
design parameters and 12 physical property models.
As for the distribution of uncertain variables, we consider the uncertainty of physical property models is
normal distribution, and the bounds are ± 3𝜎 . Otherwise, the uncertainty of design parameters is uniform
distribution.
The uncertainty module in FRAPCON and FRAPTRAN are used for the physical property models analysis.
TABLE 3. INFORMATION OF UNCERTAIN VARIABLES
No. Variable Value Lower Bound Upper Bound Distribution
1 Pellet diameter 0.00821 0.0078 0.008333 Uniform
2 Cladding thickness 0.00057 0.000542 0.000579 Uniform
3 Plenum length 0.3006 0.28557 0.31563 Uniform
4 Pellet density 95.32 94.32 96.32 Uniform
5 Enrichment 4.49% 3.50% 5.50% Uniform
6 Fuel thermal conductivity 0 (Relative) 8.8% ±3σ Normal
7 Fuel thermal expansion 0 (Relative) 10.3% ±3σ Normal
8 Fuel swelling 0 (Absolute) 0.08% ±3σ Normal
9 Cladding creep 0 (Relative) 14.5% ±3σ Normal
10 Cladding corrosion 0 (Absolute) 5 um ±3σ Normal
11 Cladding hydrogen pickup 0 (Absolute) 23 ppm ±3σ Normal
12 Fuel thermal conductivity (Transient) 1 1.50% ±3σ Normal
13 Fuel thermal expansion (Transient) 1 1.50% ±3σ Normal
14 Fuel specific heat (Transient) 1 1.50% ±3σ Normal

Cladding thermal conductivity
15 1 1.50% ±3σ Normal
(Transient)
16 Cladding thermal expansion (Transient) 1 1.50% ±3σ Normal
Cladding surface heat transfer
17 1 1.50% ±3σ Normal
coefficient (Transient)
During LOCA accident, the evaluation of fuel rod Behaviour mainly focuses on: 1) variation of cladding and
fuel temperature; 2) cladding and fuel strain; 3) cladding oxide thickness; 4) fuel rod pressure and cladding
ballooning failure time. In this report, the cladding ballooning failure time is used to evaluate the fuel performance
under LOCA condition.
63
The statistical analysis is performed with the software Dakota-6.4.0 which is a free interface with open code
developed by the Sandia National Laboratories. Dakota includes methods for global sensitivity and variance
analysis, parameter estimation, uncertainty quantification, and verification, as well as meta level strategies for
surrogate based optimization, hybrid optimization, and optimization under uncertainty. It can be coupled with
engineering analysis software.
Based on the results shown in Section 2, case 650.10 uses the PWR fuel rod segment for test and can be well
predicted by FRAPCON/FRAPTRAN. Thus, the following study is performed with case 650.10.
3.2. Results
3.2.1. Reference calculation
Table 4 shows the overall comparison of different fuel systems under LOCA condition (case 650.10). All
the inputs are the same except the fuel and cladding material. It suggests that during LOCA, the ATF system
reduces the system temperature, pressure and delays the fuel rod failure. We take these results as reference to make
a comparison with the latter calculation.
TABLE 4. REFERENCE CALCULATION RESULTS

Fuel type UO2+Zry-4 ODS FeCrAl+UO2-BeO
Cladding deformation start time, s 211 280
Cladding failure time, s 233 302
Maximum rod pressure, MPa 7.00 3.78
Maximum fuel temperature, K 1013.21 1002.59
Maximum cladding temperature, K 861.00 856.73
3.2.2. Monte Carlo sampling and correlation coefficient
By applying the Wilks based sample size criterion, the number of samples in Monte Carlo random sampling
calculation is set to 93. The definition of mean, standard deviation, skewness and kurtosis of 𝑛 samples are as
follows:
𝑋 = 𝐸(𝑋) = ∑ 𝑋 , (1)
𝜎= 𝐸[(𝑋 − 𝑋) ] = ∑ 𝑋 − 𝑛𝑋 (2)
𝑠𝑘𝑒𝑤 = 𝐸 = ∑ (3)
( )
𝑘𝑢𝑟𝑡 = [( ) ]
= ∑ (4)
The results of 93 samples are shown in Table 5. Fig. 4 shows the Pearson correlation coefficient and
Spearman correlation coefficient of the 17 uncertain variables. According to the results, cladding failure time is
greatly influenced by pellet diameter, cladding thickness, plenum length as well as fuel specific heat. However,
material properties such as cladding corrosion, cladding hydrogen pickup, fuel thermal conductivity as well as
cladding thermal conductivity shows relative slight effect on cladding failure time. Under LOCA condition, fuel
rods fail due to large cladding hoop strain which is resulted from the high rod pressure caused by the high
temperature. Increase properly the thickness of cladding and pellet diameter can enhance the cladding mechanical
resistance and reduce the thermal gradient in fuel rod. Furthermore, high fuel specific heat can slow down the fuel
temperature increase so that reduce the fuel system temperature.
64
TABLE 5. RESULTS OF 93 SAMPLES
Mean Standard Deviation Skewness Kurtosis

Max fuel temperature, K 1007.00 -12.17 -18.63 -16.92
Max cladding temperature, K 857.11 -17.30 -17.39 -17.74
Max plenum pressure, MPa 4.22 0.51 0.01 0.01
Failure time, s 293.34 9.46 -1.11 0.15
1
0.8
0.6
Correlation Coefficient
0.4
0.2
0 Pearson
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 Spearman
-0.2
-0.4
-0.6
-0.8
-1
Variable
FIG. 4. Pearson & Spearman correlation coefficient.
3.2.3. Screening of low impact factor with Morris method
Morris one-at-a-time sensitivity study (elementary effects (EE) method) is a simple but effective way of
screening the important factors among the many that can be contained in a model. It proposes two sensitivity
measures: the overall influence assessed by the mean and the interaction effect with other factor estimated by
standard deviation [6].
By using Morris method with 108 samples, two sensitivity factors were obtained. The distribution of failure
time covers a large range as shown in Fig. 5. The average failure time is 286 s. It suggests that fuel performance
under LOCA condition is sensitive to the uncertainty of parameters.
0.25
0.2
0.15
Density
0.1
0.05
Failure Time, s
FIG. 5. Failure time distribution.
65
Figure 6 presents the overall influence (mean) and the interaction effect (standard deviation) calculated by
Morris method. In Fig. 6, Variable locates at the top right corner is the high impact factor while variable locates at
the low left corner is the low impact factor. For better illustrate, Table 6 shows the data calculated in order of the
overall influence, from high to low. The results show that pellet diameter has strong impact on the cladding failure
time. Cladding thickness, plenum length, fuel density, enrichment, fuel thermal conductivity (steady + transient),
fuel thermal expansion, cladding corrosion, fuel specific heat, cladding surface heat transfer coefficient (transient)
are the factors that have some influence. The other 6 factors are the factor of low interest and will not be considered
in the following study.
16 1
14
12
Interation Effect
10
6 6
2
4 7
2 5
14 3
417
10
12
0 813
16
915
11
0 5 10 15 20 25 30 35
Main Effect
FIG. 6. Overall influence (mean) and interaction effect (standard deviation) calculated by Morris method.
TABLE 6. OVERALL INFLUENCE (MEAN) AND THE INTERACTION EFFECT (STANDARD

DEVIATION) CALCULATED BY MORRIS METHOD (IN ORDER OF THE OVERALL INFLUENCE)
No. Variable Mean Standard Deviation
1 Pellet diameter 35.065 15.865
2 Cladding thickness 5.6975 4.9284
6 Fuel thermal conductivity 5.24 6.483
3 Plenum length 4.8425 0.99848
7 Fuel thermal expansion 3.575 4.4381
14 Fuel specific heat (Transient) 3.3825 1.1467
5 Enrichment 3.0325 1.7916
Cladding surface heat transfer coefficient
17 0.965 0.39965
(Transient)
4 Pellet density 0.6625 0.68288
12 Fuel thermal conductivity (Transient) 0.5875 0.26705
10 Cladding corrosion 0.52 0.46945
16 Cladding thermal expansion (Transient) 0.1125 0.036435
13 Fuel thermal expansion (Transient) 0.105 0.031464
8 Fuel swelling 0.035 0.029496
15 Cladding thermal conductivity (Transient) 0.015 0.018974
9 Cladding creep 0 0
11 Cladding hydrogen pickup 0 0
66
3.2.4. Sobol’ indices based on PCE method and quantitative analysis of uncertainty
Variance based decomposition is a global sensitivity method that summarizes how the uncertainty in model
output can be apportioned to uncertainty in individual input variables [7]. The Sobol’ indices are the variance-
based sensitivity indices, including the main effect sensitivity index Si and the total effect index Ti. The main effect
sensitivity index corresponds to the fraction of the uncertainty in the output that can be attributed to input xi alone
[7]. The total effects index corresponds to the fraction of the uncertainty in the output that can be attributed to
input xi and its interactions with other variables [7].
Traditional calculation method of Sobol’ indices is time consuming. For the case of N=100 samples and
k=10 input, it needs N(k+2) times of calculation. Usually, in order to reduce the computational expense, some
alternative methods would be applied, for example, the emulator, high dimensional model representations
(HDMR.), Fourier Amplitude Sensitivity Test (FAST). Here, the polynomial chaos expansion (PCE), which is an
emulator, was applied. It captures the functional relationship between a set of output response metrics and a set of
input random variables [7]. With the functional relationship established by PCE, we can perform a large-sample-
size sensitivity analysis in much less time and with high reliability.
Efficient Global Optimization (EGO), which is a global reliability method, was developed to search for
multiple points on or near the limit state throughout the random variable space [7]. It can evaluate the bounds of
the response functions’ distribution. This method is based on the iteration of the Gaussian process model.
With the PCE method based on 100 samples, 10000 calculations were performed and obtained the Sobol’
indices as shown in Table 7. Same as the above analysis, it demonstrates that pellet diameter is the most influential
factor for fuel rod failure during LOCA event. The others show similar effect.
TABLE 7. SOBOL’ INDICE CALCULATED BY THE PCE METHOD

No. Variable Main Effect Total Effect
1 Pellet diameter 0.822327328 0.853180919
2 Cladding thickness 0.016790807 0.036137951
3 Plenum length 0.030328453 0.049212338
4 Pellet density 0.000519525 0.017796019
5 Enrichment 0.004103068 0.018385835
6 Fuel thermal conductivity 0.005870458 0.045248346
7 Fuel thermal expansion 0.001270478 0.024899446
8 Cladding corrosion 0.002053417 0.01255415
9 Fuel thermal conductivity (Transient) 0.000495797 0.016435852
10 Fuel specific heat (Transient) 0.002025291 0.008276226

Cladding surface heat transfer 0.003789049 0.030197664
11
coefficient (Transient)
With further screening, fuel specific heat (transient), fuel thermal conductivity (transient) and cladding
corrosion were no longer considered. EGO analysis was performed with the rest 8 factors. It concludes that the
interval of failure time is (258.6 s, 306.93 s) by changing these 8 factors in the given range.
3.2.5. Optimal design
Based on the above analysis, it suggests that fuel rod design parameters play important roles in the fuel
performance under LOCA condition. Thus, in this section, EGO analysis is performed with the 5 design parameters
to find out an optimal design. It indicates that with the optimal design shown in Table 8, cladding failure time can
be delayed to 306.39 s.
According to the comparison of performance under steady state, shown in Table 9, the optimal design
improves also the fuel system performance under steady state that enhances the nuclear safety.
67
TABLE 8. OPTIMAL DESIGN GIVEN BY EGO METHOD
No. Variable Reference Value Optimal Value
1 Pellet Diameter, mm 8.21 8.33
2 Cladding Thickness, mm 0.57 0.5785
3 Plenum Length, mm 300.6 285.58
4 Pellet density, % 95.32 95.65
5 Enrichment, % 4.49 3.5
TABLE 9. COMPARISON OF PERFORMANCE UNDER STEADY STATE

Variable Optimal Design Reference Design
Max fuel rod pressure, MPa 3.90 4.14
Max fuel centerline temperature, °C 794.99 822.48
Fission gas release, % 2.02 2.22
4. CONCLUSIONS
With the simulation of Halden fuel rod LOCA behaviour tests IFA-650.9/10/11 using
FRAPCON/FRAPTRAN, evaluation of integrated performance of fuel rod under LOCA condition is performed.
Generally, FRAPCON/FRAPTRAN predicts well the fuel rod behaviour under LOCA condition. Sensitivity and
uncertainty study of ATF fuel system’s (UO2-BeO + ODS FeCrAl) failure time under LOCA condition, found out
that the design parameters, especially the pellet diameter have important influence on fuel system’s performance
during LOCA. Proper adjustment of design parameters can optimize the fuel performance.
ACKNOWLEDGEMENTS
The report performed under IAEA research Agreement No.18514, which forms part of the IAEA
Coordinated Research Project “Fuel Modelling in Accident Conditions (FUMAC)”. This research was supported
by the National Science and Technology Major Project (2015ZX06004001), and High safety ATF Engineering
Laboratory of Shenzhen.
REFERENCES
[1] DU CHOMONT, F.B., LOCA Testing at HALDEN, the Ninth Experiment IFA-650.9, Report on OECD HALDEN
Reactor Project No HWR-917, (October 2009).
[2] LAVOIL, A., LOCA Testing at HALDEN, the Tenth Experiment IFA-650.10, Report on OECD HALDEN Reactor
Project No HWR-974, (December 2010).
[3] LAVOIL, A., LOCA Testing at HALDEN, the VVER Fuel Experiment IFA-650.11, Report on OECD HALDEN Reactor
Project No HWR-976, (December 2010).
[4] GEELHOOD, K.J., LUSCHER, W.G., “FRAPTRAN-2.0: Integral Assessment,” U.S. Department of Energy under
Contract DE-AC05-76RL01830.
[5] GEELHOOD, K.J., LUSCHER, W.G., CUTA, J.M., PORTER, I.A., “FRAPTRAN-2.0: A Computer Code for the
Transient Analysis of Oxide Fuel Rods,” U.S. Department of Energy under Contract DE-AC05-76RL01830.
[6] SALTELLI, A., RATTO, M., ANDRES, T., CAMPOLONGO, F., CARIBONI, J, et, al., “Global Sensitivity Analysis.”
The Primer[M]. John Wiley & Sons, Ltd, (2008).
[7] DAKOTA, A., Multilevel Parallel Object oriented Framework for Design Optimization, Parameter Estimation,
Uncertainty Quantification, and Sensitivity Analysis: Version 6.8 Theory Manual.
68
EXPERIMENTS ON FUEL BEHAVIOUR IN ACCIDENTAL CONDITIONS
(Session 2)
Chairperson
M. ŠEVEČEK
Czech Republic
69
EXPERIMENTAL AND MODELLING RESULTS OF THE QUENCH-19 BUNDLE
TESTS WITH FeCrAl CLADDINGS
J. STUCKERT
Karlsruhe Institute of technology (KIT),
Germany
T. HOLLANDS
Gesellschaft für Reaktorsicherheit (GRS),
Germany,
K. DOLGANOV
Nuclear Safety Institute of the Russia Academy of Sciences (IBRAE),
Russian Federation.
Abstract
The QUENCH-19 test with the bundle including twenty four electrically heated fuel rod simulators with FeCrAl(Y)
claddings was conducted at KIT on 29th August 2018. Additional bundle parts were four FeCrAl(Y) spacer grids as well as
eight KANTHAL APM corner rods; the test bundle was surrounded by KANTHAL APM shroud. The objective of the
experiment was the comparison of FeCrAl(Y) and ZIRLO claddings under similar power and gas flow conditions, simulating
an early phase of hypothetical severe accident. Comparable with the previous QUENCH-15 test, the electrical power was the
same during both pre-oxidation and transient stages. After increase of the power to 18.12 kW, this power value was kept
constant during about 2000 s. At the end of this test stage the maximal peak cladding temperature of about 1500°C was reached.
Much lower heating rate in comparison to the QUENCH-15 test was measured. Sharp increase of hydrogen release rate was
observed above 1400°C. The experiment was terminated with reflood by injecting of about 48 g/s of water from the bundle
bottom. The total hydrogen mass measured during the whole test was 9.2 g compared to 47.6 g in the QUENCH-15 experiment
performed with much shorter high electrical power stage. The observation by endoscope showed the failure of cladding tubes
at the bundle elevations between 850 and 1000 mm. The claddings were failed either due to melting (central rods), or due to
interaction with melted thermocouples, or by spalling of small annular cladding segments. The post-test analyses were
performed at GRS and IBRAE by means of the integral system codes ATHLET-CD and SOCRAT. The results of these
simulations show that both codes are able to predict the temperature history. However, compared to the radial temperature
profile of up to 300 K observed in the experiment, both codes calculate less of 50 K. The hydrogen generation was strongly
underestimated due to lack of correct FeCrAl oxidation kinetics above 1300°C.
1. INTRODUCTION
The important accident management measure to terminate a severe accident progress in a Light Water
Reactor (LWR) is the water injection into the uncovered degraded core [1]. The QUENCH experiments are devoted
to the investigation of the hydrogen source term and materials interactions during LOCA and the early stages of
severe accidents including the reflood stage.
Zircaloy-4 was used as cladding tube material in 12 out of 19 QUENCH tests. QUENCH-12 was performed
with Nb-bearing E110 cladding material in WWER geometry and QUENCH-14 with M5 in the frame of the
Advanced Cladding Materials (ACM) experimental series in the standard PWR-type bundle arrangement.
QUENCH-15 as the second ACM experiment was conducted to check the effect of ZIRLO cladding material on
bundle oxidation and core reflood, in comparison with tests QUENCH-06, QUENCH-12, and QUENCH-14. The
arrangement of the QUENCH-15 bundle was different due to cladding outer diameter and pitch; the bundle was
consisted of 24 heated rods with ZIRLO claddings, eight corner rods, and a Zr-702 shroud. The QUENCH-15
experiment was performed with the same test protocol as QUENCH-06, -12 and -14, so that the effects of different
cladding materials could be observed [2].
The QUENCH-19 experiment was the worldwide first large-scale bundle test with Accident Tolerant Fuel
(ATF) claddings. It was conducted with FeCrAl(Y) claddings (alloy B136Y3 supplied by ORNL). The test
objective was the comparison of FeCrAl(Y) and ZIRLO claddings under similar power and gas flow conditions.
The QUENCH-15 experiment was used as the reference test.
71
2. EXPERIMENTAL RESULTS
2.1. Test facility
The scheme of the test section is given in Fig. 1. The arrangement of the test rods presented in Fig. 2. The
length of the test bundle is about 2.5 m; there are 24 heated fuel rod simulators. Heating is electric by central
tungsten heaters with 5 mm diameter and length of 1024 mm. Electrodes of molybdenum/copper are connected to
the tungsten heaters and to the cables leading to the DC electrical power supply. The tungsten heaters are
surrounded by annular ZrO2 pellets, which can simulate UO2 pellets concerning the similar heat capacity.
FIG. 1. QUENCH test section. FIG. 2. Cross section of bundle, top view.
The rods are held in position by five grid spacers (at elevations 200, 50, 550, 1050 and 1410 mm). The lower
grid spacer was the standard Inconel spacer; the other four spacers were manufactured by ORNL from the same
material as the cladding tubes. The cladding outside diameter was 9.52 mm, the inner diameter 8.76 mm. All rods
were filled with Kr at a pressure of 0.23 MPa after bundle heating to peak cladding temperature of 800 K.
Seven corner rods were installed at the bundle periphery. Four of them, i.e. rods “A”, “C”, “E”, and “G”, are
made of a solid rod (Kanthal APM, Ø6 mm) at the top and a tube at the bottom (Kanthal D, Ø 6 mm, wall thickness
0.4 mm). The bottom part was used for thermocouple instrumentation. The other three corner rods, i.e. rods “B”,
“D”, and “F” (solid rods with diameter 6 mm), could be withdrawn from the bundle during the test to check the
oxidation degree and hydrogen absorption.
‶The test bundle is surrounded by a shroud of Kanthal APM with a 34 mm thick ZrO2 fibre insulation
extending from the bottom (-300 mm) to the upper end of the heated zone (+1024 mm) and a double-walled cooling
jacket of Inconel 600 (inner)/stainless steel (outer) over the entire length. The annulus between shroud and cooling
jacket is purged and then filled with stagnant Ar. The annulus is connected to a flow- and pressure-controlled
argon feeding system in order to keep the pressure constant at the target of 0.23 MPa and to prevent a penetration
of steam into the annulus after shroud failure. The annulus of the cooling jacket is cooled by Ar flowing from the
upper end to the bottom of the bundle, and by water in the bundle head. Both the absence of porous zirconia
insulation above the heated zone and the water cooling allow to avoid too high temperatures in that region″ [2]
72
The thermocouple accuracies are:
- at bundle elevations between 0 and 500 mm (NiCr/Ni thermocouples): ± 2 K (up to 600 K), ± 0.005T
(above 600 K);
- at bundle elevations between 600 and + 1300 mm (W/Re thermocouples): ± 5 K (up to 700 K), ± 0.01T
(above 700 K).
The release of hydrogen is analysed by a quadrupole mass spectrometer Balzers “GAM300” connected to
the off-gas pipe of the facility. The ion currents representing the concentrations of the analysed gases are
determined. From these measurements, the production rates of different gases (especially hydrogen) are calculated
with the ratio of the partial pressure of the particular gas and that one of Ar (carrier gas) and multiplied by the Ar
flow rate through the bundle.
2.2. Test conduct and first results
In the QUENCH-19 experiment the test sequence can be distinguished in the following stages:
— Pre-oxidation 0000–6018 s (similar to QUENCH-15);
— Heat-up 6018–7127 s (similar to QUENCH-15);
— Extended period 7127–9100 s (constant electrical power);
— Quench ≈9115–9285 s with water flow rate 48 g/s (similar to QUENCH-15).
The power history for the QUENCH-19 experiment together with the measured peak cladding temperature
is shown in Fig. 3.
FIG. 3. QUENCH-19 and QUENCH-15test progresses.
Ar and steam flows, as well as system pressure values are shown in Fig. 4. QUENCH-15 and -19 tests have
similar gas inlet conditions except the inlet gas temperature: while the QUENCH-15 gas temperature was constant
during the whole test (720 K), the corresponding gas temperature during QUENCH-19 increased from 640 to
700 K.
73
FIG. 4. Comparison of QUENCH-15 and QUENCH-19 gas parameters at bundle inlet and outlet.
The boundary conditions were different for two bundle tests (Fig. 5). Whereas the porous heat insulation of
QUENCH-15 was filled with dry argon, the QUENCH-19 insulation contained water due to leakage of steam
through the small gap at the upper shroud flange during pre-tests.
FIG. 5. Readings of thermocouples at different bundle elevations during transient stage.
These different boundary conditions should be the main reason for the fact, that the peak cladding
temperature was about 210 K lower during the pre-oxidation stage in comparison to QUENCH-15 (Fig. 6).
Additionally, the radial temperature gradient across the QUENCH-19 bundle was larger than the radial gradient
for QUENCH-15. In contrast to QUENCH-15, no temperature escalation during extended transient stage of the
QUENCH-19 experiment was observed.
74
FIG. 6. Readings of thermocouples at 850 mm.
At the beginning of the transient stage, the lower temperatures were observed at all elevations of the
QUENCH-19 bundle in comparison to QUENCH-15 (Fig. 7). The hottest elevation of the QUENCH-19 was the
level 850 mm unlike 950 mm for QUENCH-15. At the end of transient, the temperatures were very similar
between -250 and 650 mm, whereas at higher elevations the QUENCH-15 temperatures were significantly higher
due to strong exothermal reaction ZIRLO/steam in this bundle at T>1500 K.
FIG. 7. Readings of thermocouples at 850 mm.
Due to higher temperatures at the onset of reflood, the initial evaporation rate was higher in QUENCH-15
(Fig. 8). Therefore, the duration of water level increases up to the bundle head was shorter in QUENCH-19 (270 s
instead 330 s instead 330 s in QUENCH-15).
75
FIG. 8. Reflood stage: evaporation of injected water and progress of collapsed water front propagation.
The wetting of cladding surface thermocouples TFS occurred early than the collapsed water front reached
the corresponding thermocouple elevation (Fig. 9). The reason is the relatively high region of two-phase fluid.
FIG. 9. QUENCH-19: wetting of thermocouples by two-phase fluid.
76
2.3. Post-test investigations
The inspection of some peripherical rods by videoscope showed absence of cladding circumferential
segments with the height up to 12 mm, probably spalled after thermal axial expansion followed by quench
shrinkage (Fig. 10). However, several other peripherical rods remained intact.
FIG. 10. QUENCH-19: videoscope observations of damaged claddings at upper part of heated zone.
The other damage character of claddings was found by the visual inspection of the post-test bundle. The
claddings of the thermocouples made of the AISI-304 stainless steel (melting range 1400–1450°C) and located at
the elevations 850, 950 and 1050 mm were melted. The relocated melt attacked the FeCrAl claddings and caused
partial melting of these claddings (Fig. 11).
FIG. 11. QUENCH-19 bundle at bundle elevations between 900 and 1100 mm: cladding damages by molten thermocouple
steel (AISI 304) sheaths.
The generation rate of hydrogen released due to oxidation of cladding tubes, corner rods, shroud, grid spacers
and thermocouples showed the maximal value of 280 mg/s at the end of the transient (QUENCH-15: 1830 mg/s).
The total hydrogen mas measured at the end of the test was 9.2 g, what corresponds about 20% of total hydrogen
in QUENCH-19 (Fig. 12).
77
FIG. 12. Hydrogen release during QUENCH-15 and QUENCH-19.
3. SIMULATION RESULTS
3.1. Simulation with ATHLET-CD
ATHLET-CD (Analysis of THermal-hydraulics of LEaks and Transients with Core Degradation) [5]
describes the thermal-hydraulic response of reactor coolant system during severe accidents, including core damage
progression as well as aerosol and fission product behavior, to calculate the source term for containment analyses,
and to evaluate accident management measures. This code is developed by GRS in cooperation with IKE,
University of Stuttgart.
3.1.1. Input data and model options
The post-test calculation used the same nodalization and as far as possible the same modelling options as for
the earlier QUENCH-15 simulation. The input model for the QUENCH facility comprises among others the bundle
fluid channel, subdivided into 20 axial nodes (10 nodes within the heated length). The rod bundle is simulated
within the code module ECORE by three concentric rings, an inner ring (ROD1) containing four heated rods, an
intermediate ring containing eight heated rods (ROD2) and an outer ring with twelve heated rods (ROD3). In
addition, the 5 spacer grids, the 8 corner rods, the shroud with its ZrO2 thermal insulation, and the outer cooling
jacket were simulated. The argon and steam flows, as well the quenching by water have been simulated by fill
junctions at the bottom of the bundle or at the top of the bundle (injection of argon after start of flooding).
For the simulation of QUENCH-19 input parameters and modelling options as recommended by the code
user’s manual have been applied. For the external resistance of each heated rod, which takes into account the
voltage drop across the sliding contacts at the rod extremities as well as at the cables connecting the sliding contacts
to DC power supply, a value of 5 mΩ was applied.
3.1.2. Assumptions and model changes for FeCrAl
In the input data deck for the simulation of the test the material properties of Zr were replaced by the values
for FeCrAl given by KIT for the Kanthal APM alloy with a composition of 69% Fe, 21.6% Cr, 4.9% Al (+ 4.5%
others) [6]. With the assumption that no other oxides are produced by oxidation beside Al2O3 and material data for
aluminum oxide taken from [7] the following material properties have been used (Table 1):
78
TABLE 1. MATERIAL PROPERTIES USED FOR ATHLET-CD
Material property FeCrAl Al2O3

3 3
Density (kg/m3) 7.1×10 3.9×10
Thermal conductivity
11 (323 K) → 35 (1673 K) 35
(W/mK)
Specific Heat
460 (293 K) → 800 (1673 K) 880
(J/kg K)
The modification of the oxidation model for FeCrAl components with this specific composition was
considered in ATHLET-CD within the ECORE module for the claddings and within the HECU module for corner
rods, spacer grids and shroud. Corresponding to the chemical reaction
Fe Cr Al + 3H O → Fe Cr Al O + 3H + 𝑧𝛥ℎ (1)
with an energy release of 𝛥ℎ = 9.3×105 J/mol = 9.32×106 J/kgFeCrAl (molar weight 𝑀 = 99.8×10-3
kg/mol) heat and hydrogen were produced. Due to the composition of FeCrAl with the molar masses of x=1.307,
y=0.404 and z=0.215 [6] the amount of hydrogen production was calculated. The mass of the resulting oxide Al2O3
(molar weight 𝑀 =102.0×10-3 kg/mol) determined the layer thickness of the oxide corresponding to its density
and the given geometry of the rods.
The oxidation rate was treated as a parabolic law as derived from the analytical solution of the diffusion
equation in the same way as used for the oxidation of Zr:
𝑑𝑊 = 𝐾 ∙ 𝑑𝑡 → = (2)
with 𝑊 = 𝑚 /𝑠𝑢𝑟𝑓. 𝑎𝑟𝑒𝑎 in kg/m2, 𝐾 = reaction rate in kg2/m4s, 𝑡 = time in s.

The reaction rate 𝐾 is defined by an Arrhenius formulation 𝐾 = 𝐴 exp (−𝐵/𝑅𝑇) ∙ 𝑔(𝑝 ) with
𝑅= 8.314 J/(mol∙K) (gas constant), 𝑇 is cladding temperature in K, 𝑔(𝑝 ) is reduction factor to consider steam
starvation (0 ≤ 𝑔(𝑝 ) ≤ 1) and the rate constants 𝐴, 𝐵 as given by KIT for KANTHAL APMT [6]: 𝐴 = 3.1 kg2/m4s,
𝐵 = 2.78519×105 J/mol. Additionally to this data, a second rate with a multiplication factor of 300 was used, which
was derived from [3]. Fig. 13 compares the resulting oxidation rates with the standard options of ATHLET-CD
used for the oxidation of Zr.
FIG. 13. Comparison of oxidation rates for Zr and FeCrAl in ATHLET-CD.
79
3.1.3. Main results of QUENCH-19 post-test simulation with ATHLET-CD
The main post-test calculation results for QUENCH-19 are depicted in the Fig. 14 (bundle temperatures) and
Fig. 15 (hydrogen release) [8]. The bundle temperatures for inner (ROD1: blue), intermediate (ROD2: red) and
outer ring of rods (ROD3: green) are satisfactorily reproduced by the code in comparison with data (black, olive,
grey) within the heated length, only for the position above the heated length at 1350 mm height the cladding
temperatures were overestimated in comparison to the experiment. Additionally, the comparison shows that there
exists a relatively high radial gradient between inner and outer temperatures with a difference up to ≈200 K in the
measured data (noticeable larger in comparison to QUENCH-15; the reason of this differences are not yet
completely clarified [4]) while the simulation yields a flat profile with a difference of at most 50 K.
Bundle temperatures at 250 mm Bundle temperatures at 550 mm
FIG. 14. Measured and simulated cladding temperatures for QUENCH-19.
In agreement with the measurement there is no temperature excursion in the simulation. The peak cladding
temperature in the experiment of 1728 K at an elevation of 850 mm is calculated in good concordance with 1685 K
at 950 mm height, a deviation of only 43 K. After start of quenching no further temperature increase occurs as
shown also for the test. The melting temperature of FeCrAl, which is 1773 K was not reached neither in test nor
80
in the simulation (horizontal black dashed line) but the videoscope observation of the bundle at the positions of
the withdrawn corner rods showed the failure of claddings at the bundle elevations between 850 and 1000 mm. It
is assumed that the claddings failed due to melting (central rods), or due to interaction with melted thermocouples
(peripherical rods), or by spalling of small annular cladding segments [4].
9g
Hydrogen generation rate Mass of generated hydrogen
FIG. 15. Hydrogen release for QUENCH-19.
The comparison of hydrogen generation rate and mass of total generated hydrogen (Fig. 15) shows a heavy
underestimation of oxidation in the simulation, especially for the reaction rate given by [6], but also with the
increased rate (factor 300) as given in [4], with 0.008 g resp. 0.35 g compared to 9 g in the test. The hydrogen
production in the test was comparable low up to the time when the maximum temperatures reached a level of
≈1400°C. Then a sharp increase of hydrogen release rate was observed at about 800 s before reflood. Probable
trigger of this event could be the melting of steel thermocouple claddings [4]. This effect was not considered in
the post-test calculation; nevertheless, the oxidation modelling for FeCrAl should be improved.
3.2. Simulation with SOCRAT
SOCRAT/V3 code [9] has been adapted for FeCrAl ATF claddings. Due to a lack of experimental data with
FeCrAl for validation of thermomechanical model in SOCRAT the calculated time of the FeCrAl claddings burst
corresponds to the moment when cladding started to melt, which overestimates the expected time of mechanical
failure.
3.2.1. Assumptions and model changes for FeCrAl
The system pressure is 2 bar. The team-argon mixture at the bundle inlet: has the following characteristics:
flow rate of Ar 0.003472 kg/s, flow rate of steam 0.00387 kg/s (initial value, decreases with time in accordance
with experimental data), temperature 504.7 K (calculated).
The side heat removal in the cooling jackets was chosen according to experimental data with the following
parameters: 1) Ar cooling with mass flow rate of 0.00559 kg/s, temperature 296.1 K, pressure 6.25 bar, 2) water
cooling with mass flow rate of 0.250 kg/s, temperature 341 K, pressure 6.5 bar. The heat removal from the cooling
jacket wall to containment atmosphere occurred by gas convection with heat transfer coefficient 5 W/(m2·K) at
gas temperature 335 K.
The reflood is simulated by water injection from the bundle bottom with flow rate in accordance with
experimental data and temperature 296.4 K.
3.2.2. Change of material properties
The basic temperature dependence of zirconia fiber insulation provided in a report for QUENCH-15
experiment was modified to account for supposed water presence in cold regions close to cooling jacket
(temperatures 300 and 404 K in Table 2). The heat capacity of zirconia insulation region was not modified and
was set in accordance with a two-point time dependence provided in QUENCH-15 report.
81
TABLE 2. MODIFIED DATA FOR FIBER INSULATION
Temperature, K 300 404 405 673 1373 1673 1923
Heat conductivity, W/(m·K) 0.68 0.68 0.12 0.11 0.14 0.19 0.24
No fiber insulation there is between 1024 and 1300 mm. The basic value of heat conductivity for this region
(calculated with account for radiation and convective heat transfer) was increased 3 times to assure the fast
decrease of temperature in the experimental data. The values implemented in calculations are shown in Table 3.
TABLE 3. HEAT CONDUCTIVITY ABOVE 1024 MM
Temperature, K 300 500 595 735 885 975 1040 1130 1200
Heat conductivity, W/(m·K) 0.42 0.99 1.56 2.70 3.90 4.95 6.00 7.39 9.00
The density of HfO2 powder in TC insulation was 1.5 times smaller than that in QUENCH-15 experiment.
Therefore, the basic values of density and heat conductivity provided in QUENCH-15 report were decreased 1.5
times.
3.2.3. Applied oxidation model
All FeCrAl alloys are modeled with a correlation based on data for the APMT alloy [10] for temperature
range T≤ 1723 K:
𝐾 (𝑇) = 0.3041 ∙ exp(−32850/𝑇), (3)

above 1773 K (melting temperature of FeCrAl), the correlation for stainless steel oxidation is used:
𝐾 (𝑇) = 3607 ∙ exp(−35210/𝑇),
where 𝐾 and 𝐾 are in g2/(cm4s). (4)

To assure a transition between these low-temperature and high-temperature oxidation kinetics 𝐾 and 𝐾 , a
special correlation is implemented for intermediate range of temperatures 1723–1773 K:
𝐾 (𝑇) = 𝑎 ∙ exp(−𝑏/𝑇), (5)

The values of coefficients a and b in (5) are defined by the parameters in correlations (3) and (4):
𝑎 = 𝑎 + (𝑎 − 𝑎 )(𝑇 − 1723)/50, (6)
𝑏 = 𝑏 + (𝑏 − 𝑏 )(𝑇 − 1723)/50, (7)
where 𝑎 = 0.3041, 𝑎 =3607, 𝑏 = 32850, 𝑏 = 35210.

The products of oxidation are specified taking into account that the initial metallic components Fe, Al, and
Cr are oxidized to FeO, Al2O3, and Cr2O3 respectively. The weight fractions of Fe, Al, and Cr in alloy are currently
fixed at 72%, 6%, and 22%, respectively.
The mass of oxidation products is used for calculating the thermal properties of oxidized FeCrAl. This
oxidation model was preliminary validated at a series of 7 separate effect test performed in ORNL [11].
4. MAIN RESULTS OF QUENCH-19 POST-TEST SIMULATION WITH SOCRAT
Figure 16 shows temperature results of in comparison with the experimental results for different bundle
elevations. Significant radial temperature gradient observed in experiment was not reproduced by code: for
peripherical rods (structure ROD 3) there is noticeable temperature overestimation. Figure 17 shows the modelling
results for the integral hydrogen release.
82
1400 1400 1400
1300 1300 1300
1200 1200 1200
Temperature, K
Temperature, K
Temperature, K
1100 1100 1100
1000 1000 1000
900 900 900
800 800 800
SOCRAT_ROD1 SOCRAT_ROD2
700 700 700 SOCRAT_ROD3
rod1 exp rod2 exp rod3 exp
600 600 600
0 2000 4000 6000 8000 10000 0 2000 4000 6000 8000 10000 0 2000 4000 6000 8000 10000
Time, s Time, s Time, s
ROD 1, 550 mm ROD 2, 550 mm ROD 3, 550 mm

1700 1700 1700
1600 1600 1600
1500 1500 1500
1400 1400 1400

Temperature, K
Temperature, K
Temperature, K
1300 1300 1300
1200 1200 1200
1100 1100 1100
1000 1000 1000
900 900 900

SOCRAT_ROD2 SOCRAT_ROD3
SOCRAT_ROD1
800 800 800
rod1 exp rod2 exp rod3 exp
700 700 700
0 2000 4000 6000 8000 10000 0 2000 4000 6000 8000 10000 0 2000 4000 6000 8000 10000
Time, s Time, s Time, s
ROD 1, 950 mm ROD 2, 950 mm ROD 3, 950 mm
FIG. 16. Cladding temperatures at elevations 550 and 950 mm.
10
9
SOCRAT_H2 Q19 H2 experiment
8
7
Hydrogen, g
6
5
4
3
2
1
0
0 2000 4000 Time, s 6000 8000 10000
FIG. 17. Comparison of experimental and calculation results for hydrogen release.
83
5. SUMMARY AND CONCLUSIONS
The QUENCH-19 experiment with bundle containing 24 heated rods with FeCrAl(Y) cladding and 4
FeCrAl(Y) spacer grids as well as 7 KANTHAL APM corner rods and KANTHAL APM shroud was performed
at KIT on August 29, 2018 with similar electrical power history as reference test QUENCH-15 (ZIRLO claddings).
Not similar conditions were 1) cooler steam-Ar flow, and 2) humid Ar inside the heat insulation for QUENCH-
19.
The QUENCH-19 test was performed in four test stages:
(a) pre-oxidation during about 6000 s (similar to QUENCH-15);
(b) transient during about 1130 s (similar to QUENCH-15);
(c) extended period with constant electrical power of 18.32 kW during 1970 s (to extend the temperature
increase stage);
(d) test termination by water flooding with rate of 48 g/s (similar to QUENCH-15).
The peak cladding temperatures during the pre-oxidation stage were about 200 K lower in comparison to
QUENCH-15. The radial temperature gradient was noticeable larger in comparison to QUENCH-15. The reasons
for these test differences could be the different properties of the bundle materials (lower thermal conductivity and
higher heat capacity and thermal expansion of FeCrAl) as well as the different boundary conditions (cooler gas
flow, humid heat insulation). Much lower heating rate was measured in comparison to QUENCH-15. A
temperature of about 1423 K (1150°C) was reached at the time point as a local melting of QUENCH-15 claddings
occurred. No temperature escalation was observed during the extended transient. Maximum cladding temperature
measured before reflood was about 1733 K (1460°C). The coping time was about 3200 s (≈ 1200 s for QUENCH-
15). Many claddings were damaged at elevations between 850 and 1000 mm: 1) by interaction with melted
thermocouples or 2) parts of claddings were spalled (probably due to thermal expansion followed by quench
shrinkage). Sharp increase of hydrogen release rate was observed about 800 s before reflood. Probable trigger of
this event could be the melting of steel thermocouple claddings. The maximum hydrogen release rate reached
before reflood was 280 mg/s (1830 mg/s for QUENCH-15). Total hydrogen production 9.2 g (47.6 g for
QUENCH-15).
After the implementation of two approaches for the oxidation kinetics of FeCrAl with the assumption that
only Al2O3 and no other oxides will be generated, the most recent ATHLET-CD code version has been used for
the simulation of the QUENCH-19 experiment. As far as possible the same input data deck and modelling options
were used compared to the reference case QUENCH-15, only boundary conditions of the test had been taken from
the data file and material data for FeCrAl instead of Zr (given by KIT) and the only considered oxidation product
Al2O3 instead of ZrO2 (from [7]) were changed.
The results of the post-test simulations show that ATHLET-CD is able to predict the thermal behaviour of
the experiment in good agreement to the measured values, especially within the heated length. Compared to the
observed radial temperature profile of up to ≈ 200 K in the experiment, which is different to the reference case
QUENCH-15 and the reason for this is not clear up to now, ATHLET-CD calculates a flat profile with a difference
of only ≈50 K between inner and outer rods. Different to QUENCH-15 there was no temperature excursion neither
in the test nor in the simulation; the temperatures decrease immediately after water injection. The maximum
temperature was obtained with high accuracy; there was a small underestimation of only ≈ 40 K.
The comparison of the hydrogen production with totally 9 g resulting from the test and less than 1 g for both
oxidation approaches in the simulation shows that the oxidation model of FeCrAl in the ATHLET-CD code does
not comprise all the effects which contribute to the hydrogen generation in the test. For a detailed evaluation of
the calculated hydrogen generation it is necessary to use the final results of the post-test examination of the bundle
to know which components account for the total value. Nevertheless, the implementation of the oxidation model
for FeCrAl was a first approximation and will be improved.
First validation analyses of the SOCRAT code for FeCrAl applications were performed on the basis of the
QUENCH-19. Due to absence of reliable oxidation kinetics of FeCrAl (Y) alloy at high temperatures, results of
the modelling have temporarily character.
Both ATHLET-CD and SOCRAT codes able to predict the thermal behavior of the internal bundle rods,
however temperatures for the periphery rods were significantly overestimated. The calculated radial temperature
gradient was about 50 K, whereas experimental gradient was more than 300 K. The cause of this was the not
simulated circumstance of the water penetration into the heat insulation through the leakage at the bundle head.
84
For simulation of the bundle oxidation, both codes used the correlation established early for the industrial
KANTHAL APMT alloy, whereas the test was performed with new B136Y3 alloy developed by ORNL for nuclear
applications (FeCrAl alloy with lower Cr content). New experimental data are needed for future correct modelling
of the oxidation of the B136Y3 alloy.
ACKNOWLEDGEMENTS
The QUENCH-19 experiment was supported by the KIT programme NUSAFE and partly sponsored by
ORNL. The cladding materials and thermocouples were provided by ORNL. The authors thank all colleagues
involved in the investigations, particularly M. Steinbrück and M. Große for many useful discussions, J. Moch and
U. Peters for bundle preparation, U. Stegmaier for mass spectrometer measurements.
The development and validation of ATHLET-CD are sponsored by the German Federal Ministry for
Economic Affairs and Energy (BMWi)
REFERENCES
[1] STUCKERT, J., STEINBRUECK, M., GROSSE, M., Experimental programme QUENCH at KIT on core degradation
during reflooding under LOCA conditions and in the early phase of a severe accident, IAEA-TECDOC-CD-1775 (Nov.
2015) 281–297.
[2] STUCKERT, J., BIRCHLEY, J., GROßE, M., JAECKEL, B., STEINBRÜCK, M., Experimental and calculation results
of the integral reflood test QUENCH-15 with ZIRLO cladding tubes in comparison with results of previous QUENCH
tests. Nuclear Engineering and Design 241 (2011) 3224–3233.
[3] NUCLEAR ENERGY AGENCY (NEA), State-of-the-Art Report on Light Water Reactor Accident-Tolerant Fuels, NEA
No. 7317, OECD (2018).
[4] STUCKERT, J., GROßE, M., LAIER, J., MOCH, J., PETERS, U., STEGMAIER, U., STEINBRÜCK, M., TERRANI,
K., “First results of the bundle test QUENCH-19 with FeCrAl claddings”, 24th International Quench Workshop,
Karlsruhe, (November 2018).
[5] AUSTREGESILO, H., et al., “ATHLET-CD 3.2 User’s Manual”, GRS Report, GRS-P-4/Vol. 1, Dec. 2018.
[6] STUCKERT, J., Mail to GRS from 26.7.2018.
[7] ACCURATUS, Ceramic Materials and Ceramic Components, 9.10.2018,
https://accuratus.com/alumox.html.
[8] HOLLANDS, T., BALS, C., “Pre- and Post-Test Simulations of the ATF Experiment QUENCH-19 with AC2”, 24th
International QUENCH Workshop, Karlsruhe, November 2018.
[9] BOLSHOV, L., STRIZHOV, V., SOCRAT — The System of Codes for Realistic Analysis of Severe Accidents. Proc.
of ICAPP’06. Reno, NV USA, June 4–8, 2006.
[10] UNOCIC, K.A., YAMAMOTO, Y., PINT, B.A., Effect of Al and Cr Content on Air and Steam Oxidation of FeCrAl
Alloys and Commercial APMT Alloy, Oxidation of Metals 87 3–4 (April 2017) 431–441.
ROBB, K.R., HOWELL, M., OTT, L.J., Parametric and Experimentally Informed BWR Severe Accident Analysis Using
FeCrAl, ORNL/SPR-2017/373, ORNL, (August 2017).
85
OXIDATION PERFORMANCE AND FAILURE BEHAVIOUR OF MONOLITHIC
AND COATED ATF CLADDINGS UNDER SEVERE ACCIDENT CONDITIONS
C. TANG, M. STEINBRUECK, M. GROSSE

Institute for Applied Materials (IAM-AWP), Karlsruhe Institute of Technology (KIT),
Hermann-von-Helmholtz-Platz 1, 76344 Eggenstein-Leopoldshafen,
Germany
Abstract
Various concepts of ATF claddings with enhanced tolerance to loss of active cooling in nuclear reactor core have been
proposed and intensively investigated worldwide. These potential solutions comprise mainly three categories: (1) state-of-the-
art zirconium based alloy cladding with high temperature oxidation resistance coatings, (2) new metallic cladding such as
FeCrAl alloy and (3) ceramic SiCf/SiC composite cladding. These innovative ATF cladding materials have shown excellent or
satisfactory performance during design-basis accident conditions. However, their degradation and failure behaviour under
simulated severe accident conditions are lacking assessment. In this paper, the oxidation performance and failure behaviour of
one type of monolithic ATF cladding (FeCrAl alloys) and two coated Zr alloys (Cr coated and Cr2AlC coated) in high
temperature steam extended to severe accident scenarios are explored. FeCrAl alloys are susceptible to breakaway oxidation
(formation of non-protective iron based oxide) at temperatures exceeding 1200°C, depending on heating schedules and
chemical compositions. The principal failure mechanisms of metallic (Cr) coated and ceramic (Cr2AlC) coated Zr alloys are
formation of liquid phase via eutectic reaction and cracking, respectively, at high temperatures.
1. INTRODUCTION
Zirconium based alloy fuel claddings used for current light water reactors (LWRs) possess desirable
performances during steady state normal operations and design basis accidents. The Fukushima Daiichi Accidents
in 2011 highlighted some detrimental performance characteristics of the fuel claddings, i.e. strong heat and
hydrogen releases via rapid exothermic zirconium-steam oxidation reaction, during severe accidents and the need
to develop advanced accident tolerant fuel (ATF) cladding for enhancing the safety margins of water cooled
reactors [1]. One of the crucial performance features of advanced ATF cladding materials is reduced oxidation
kinetics in high temperature steam with simultaneously minimized hydrogen generation and enthalpy input [2]. In
general, the high temperature oxidation resistance of materials relies on the growth of an external, protective and
slow-growing oxide scale. The most common protective oxide scales preventing underlying materials from fast
oxidation are Cr2O3, Al2O3 and SiO2. Various concepts of enhanced ATF claddings proposed and intensively
investigated currently are materials that can form one of those oxide scales during high temperature oxidation [2–
4]. These alternative solutions comprise mainly three categories: (1) coated Zr cladding, (2) metallic FeCrAl alloy
cladding and (3) ceramic SiCf/SiC composite cladding.
In case of coating materials, a variety of metallic and ceramic coatings has been examined. One of the most
intensively studied coatings is pure Cr [5]. Oxidation of Cr coatings in steam result in growth of a Cr2O3 layer,
which possesses good oxidation resistance up to 1200°C. Alumina-forming metallic or ceramic materials, for
instance FeCrAl alloys or Ti2AlC and Cr2AlC MAX phases, also have been proposed as protective coatings since
alumina scale is featured by a much lower growth rate than that of chromia [6]. FeCrAl-based alloys represent a
very promising candidate for replacing the Zr-based alloy claddings due to their excellent properties, i.e. excellent
high temperature oxidation resistance as well as good formability and mechanical properties [7]. However,
majority efforts on characterization and qualification of these enhanced ATF cladding materials are under
operational and design basis conditions. So far, only limited studies focused on their degradation and failure
behaviour under simulated severe accident conditions. A comprehensive understanding of the oxidation
performances and failure mechanisms of these ATF cladding candidates under severe accidents is lacking. In the
framework of the QUENCH programme at the Karlsruhe Institute of Technology, in cooperation with international
partners, the investigation of the performance of various ATF candidate materials during severe accidents has been
established in recent years. Numerous laboratory scale tests from design basis accidents extended to severe
scenarios (up to >1750°C in steam) have been performed. In this paper, the oxidation performance and failure
behaviour of one type of monolithic ATF cladding (FeCrAl alloys) and two coated Zr alloys (Cr coated and Cr2AlC
coated) in high temperature steam is explored and reported.
86
2. EXPERIMENTS
The FeCrAl alloys examined in this study consisted of one commercial alloys, Kanthal APM
(Fe20.5Cr5.8Al, wt.%), and one model alloy Fe12Cr5Al (wt.%). The commercial alloy and the mode alloy are in
tube and disk shape, respectively. Two types of Cr coated Zr alloys, including physic vapor deposition coated and
cold spraying coated, were tested. The thickness of the PVD Cr coating is around 25 µm and of the cold spraying
Cr coating is around 30 µm. The Cr2AlC coatings were synthesized via a two-step approach and more info can be
found in [8]. The thickness of the Cr2AlC coatings was around 6 µm with a 500 nm Cr interlayer as
bonding/diffusion barrier.
The majority oxidation experiments were conducted using the so-called BOX furnace with a horizontal
corundum tube under normal pressure. The specimen was located at the center of furnace on a corundum crucible
sample holder. A Bronkhorst flow controller and gas mixer system was used to define the oxidizing atmospheres.
The composition of the off-gas was in-situ analyzed by a mass spectrometer Balzers GAM300. The specimens
were oxidized in a flowing argon + steam atmosphere. Argon was used as the carrier gas and as the reference gas
for mass spectrometer analysis. Two different kinds of tests were conducted, isothermal tests and transient tests.
In the isothermal tests, samples were heated in high purity Ar atmosphere (6N) to the desired temperatures with
fixed heating rate 10 K/min. The atmosphere was then switched to 20 l/h Ar and 20 g/h H2O (≈55 vol.% H2O) for
steam oxidation with different dwell times. Finally, after the isothermal exposure, the specimens were cooled down
to room temperature by simultaneously changing the atmosphere back to high purity Ar. In the transient test, the
steam was injected at low temperatures, and the specimens were oxidized from low temperature to a pre-defined
high temperature with different heating rates; a succeeding isothermal exposure period was applied once reaching
the pre-defined temperatures. Few tests were performed using a single-rod quench facility (QUENCH-SR) with a
vertical quartz furnace heated by induction coils. After the test, the samples were characterized by X-ray
diffraction, scanning electron microscope (SEM), etc.
3. RESULTS
3.1. Monolithic FeCrAl alloys
Figure 1summarizes the oxidation behaviour of the Kanthal APM alloys during transient test from 500 to
1400°C with subsequent holding for 1 h at 1400°C with two different heating rates. At lower heating rate, 5 K/min,
the hydrogen production during the steam exposure was extremely low. Both XRD (Fig. 1(b)) and SEM analysis
(Fig. 1c) confirmed that a protective alumina scale established on the surface, inhibiting direct contact of steam
with the alloy matrix. However, at the slightly higher heating rate 10 K/min during the ramp period, surprising
results were observed. The hydrogen concentration increased abruptly to an extremely high value once the
temperature reaching around 1380°C (Fig. 1(a)), two magnitudes higher than at 5 K/min, accompanied by a strong
peak. The sample was disintegrated after the test and the final mass gains was around 170 mg/cm2 corresponding
to 47 wt.%. XRD and SEM results confirmed that the alloy was completely oxidized without alumina scale
surviving. These findings reveal that the alloy suffered catastrophic oxidation via rapid growth of non-protective
iron based oxide at temperature higher than ≈1380°C. Thus, an adequately thick alumina oxide layer is needed to
avoid the occurrence of catastrophic oxidation especially at elevated temperatures for this category of alloys [9].
87
FIG. 1. Oxidation performance of Kanthal APM alloys (Fe20.5Cr5.8Al) during transient test from 500 to 1400°C with
subsequent holding for 1 h at 1400°C, left column: 5 K/min heating rate and right column: 10 K/min heating rate. (a) and (a′)
hydrogen concentration in the off-gas with inserted appearance of the samples after test. (b) and (b′) XRD patterns. (c) and
(c′) SEM images of surface view. Note that alumina scale formed at lower heating rate and breakaway oxidation occurred at
higher heating rate.
Since commercial high Cr (≈20 wt.%) FeCrAl alloys sustain radiation-induced hardening and embrittlement
due to formation of the Cr-rich α' phase under irradiation, relatively low Cr alloys are currently optimized for
nuclear application [10]. A low Cr model alloy (Fe12Cr5Al) was further tested and Fig. 2 displays the hydrogen
concentration in the off-gas with inserted appearance of the samples during one transient test to 1450°C and one
isothermal test at 1200°C. In both tests, a high concentration peak of hydrogen appeared; together with appearance
of the samples after test clearly revealed that catastrophic oxidation occurred. The maximum tolerance temperature
of the alloy during the transient test was determined as around 1330°C. While, the alloy even cannot form an
alumina scale during isothermal oxidation at 1200°C, i.e. here without a pre-oxidation process. The alloys are
more susceptible to catastrophic oxidation in high temperature steam considering the absence of a pre-oxidation
process. In real accident scenarios, the heating rates in the nuclear core vary from a few K to dozens of K per
minute. The response of this alloys under simulated severe accident conditions needs detailed investigations.
88
FIG. 2. Hydrogen concentration in the off-gas with inserted appearance of the Fe12Cr5Al samples (a) transient test from 500
to 1400°C at 10 K/min heating rate with subsequent holding for 1 h at 1450°C and (b) isothermal test at 1200°C for 1 h. The
grey areas indicating steam oxidation period.
3.2. Coated Zr alloys
3.2.1. Cr coated Zr alloys
Figure 3 shows the cross sectional view of PVD Cr coated Zircaloy-2 after oxidation in steam at 1200°C for
30 min and 1300°C for 5 min, respectively. Under both conditions, a protective Cr2O3 grew on the surface and an
interdiffusion layer formed at coating/substrate interface. The coating possesses excellent adherence without any
sign of cracking and spallation. However, the oxidation rate and the interdiffusion rate both increased significantly
with increasing temperature. The oxide scale and interdiffusion layer thicknesses are thicker after oxidation at
1300°C for just 5 min (Fig. 3b) than those at 1200°C for 30 min (Fig. 3a). The diffusion of Cr into the Zr alloy
substrate has been proved leading to formation of brittle intermetallic compound (ZrCr2), which weaken the
strength of the alloy substrate [11]. Thus, during low term intermediate temperature (below the Cr-Zr eutectic
temperature) scenarios, the failure behaviour due to coating-substrate interdiffusion should be considered.
FIG. 3. Cross sectional view of PVD Cr coated Zircaloy-2 after oxidation (a) at 1200°C for 30 min (c) 1300°C for 5 min in
steam.
Figure 4 provides hydrogen release rate and surface SEM image of cold spraying Cr coated zircaloy during
a transient test up to 1500°C. Obviously, the hydrogen release rate accelerated once the temperature exceeded the
eutectic temperature (≈1330°C) and a steeply increased hydrogen release was observed at temperatures above
approx. 1450°C. Then the hydrogen rate was significantly higher than the hydrogen release rate of the uncoated
89
reference Zircaloy-4, indicating a much higher oxidation rate after failure of the coatings. Examination of the
sample surface clearly showed the formation of liquid protrusions and cracking of the coating, (Fig. 4b).
FIG. 4. Transient oxidation tests from 800 to 1500°C in steam with cold spraying Cr-coated and uncoated zircaloy (a) hydrogen
release rate (b) surface SEM image of coated zircaloy after test.
3.2.2. Cr2AlC coated Zr alloys
Cr2AlC belongs to a family of layered, ternary compounds known as MAX Phases. It possesses excellent
high temperature oxidation resistance and self-healing capability via fast growth of an alumina scale. As can be
seen in Fig. 5, a thin and dense alumina layer formed on the surface of the Cr2AlC coating after oxidation at 1000°C
in steam for 1 h, which protects the coating and the substrate from fast oxidation.
FIG. 5. Cross sectional SEM image of Cr2AlC coated zircaloy after oxidation at 1000°C in steam for 1 h.
Figure 6 shows the hydrogen release rate and cross sectional SEM images of the Cr2AlC coated and uncoated
Zircaloy-4 during transient oxidation tests from 500 to 1400°C in steam. For the coated sample, the hydrogen
release rate first increased slowly with the increasing of the oxidation temperature. The slope of the curve changed
and became much steeper at temperature ≈ 1260°C. The hydrogen release rate of the coated sample then surpassed
the uncoated one at around 1300°C. Combined with the cross sectional SEM image, it can be deduced that the
coatings cracked, and steam penetrated the unoxidized substrate, as shown in Fig. 6c. The uncoated Zircaloy-4
formed a uniform ZrO2 layer on top as seen in Fig. 6(b). In comparison, oxidation of the coated sample finally led
to the growth of a non-uniform oxide scale and thicker oxide layer was observed beneath the cracks. The thickness
of the oxide scale was still much thinner than that on uncoated sample. It is necessary to point out that no liquid
90
phase was found for the Cr2AlC coated sample up to 1400°C. The ceramic materials as coating can increase the
eutectic reaction temperature between the coating and substrate.
FIG. 6. Transient oxidation tests from 500 to 1400°C in steam with Cr2AlC coated and uncoated Zircaloy-4 (a) hydrogen
release rate, cross sectional SEM image of (b) uncoated and (c) coated zircaloy after test.
4. CONCLUSIONS
The oxidation performance and failure behaviour of one type of monolithic ATF cladding (FeCrAl alloys)
and two coated Zr alloys (Cr coated and Cr2AlC coated) in high temperature steam extended to severe accident
scenarios were investigated. Catastrophic oxidation, i.e. rapid and complete consumption of the FeCrAl alloy, can
be triggered above 1200°C in case of fast heating during transient tests in steam. The lack of the pre-oxidized
alumina scale declines the maximum tolerance temperature of the alloys to high temperature steam. Cr and Cr2AlC
coatings can establish protective Cr2O3 and Al2O3 scales during high temperature oxidation in steam. Significant
inward diffusion of the Cr into the Zr alloy substrate occurred during oxidation at temperatures lower than the Cr-
Zr eutectic reaction temperature, which may embrittle the fuel cladding. Formation of liquid phase via eutectic
reaction, leading to succeeding fast oxidation rate, dominates the failure feature at temperatures above the Cr-Zr
eutectic temperature. The Cr2AlC ceramic coating formed a protective alumina layer during oxidation, while it
cracked more easily at high temperatures. However, no eutectic reaction was found during oxidation up to 1400°C
in steam.
ACKNOWLEDGEMENTS
This work was sponsored by the HGF programme NUSAFE at Karlsruhe Institute of Technology. C. Tang
appreciates the fellowship supported by the China Scholarship Council (CSC) during his PhD. The authors thank
Koroush Shirvan from MIT, USA for providing the cold spraying Cr coated samples and U. Stegmaier and P.
Severloh for support during experiments.
REFERENCES
[1] ZINKLE, S.J., TERRANI, K.A., GEHIN, J.C., OTT, L.J., SNEAD, L.L., Accident tolerant fuels for LWRs: A
perspective, J. Nucl. Mater. 448 (2014) 374–379.
[2] TERRANI, K., Accident tolerant fuel cladding development: Promise, status, and challenges, J. Nucl. Mater. 501 (2018)
13–30.
[3] TANG, C., STUEBER, M., SEIFERT, H.J, STEINBRUECK, M., Protective coatings on zirconium based alloys as
accident-tolerant fuel (ATF) claddings, Corros. Rev. 35 (2017) 141–166.
91
[4] BRAGG-SITTON, S., Development of advanced accident-tolerant fuels for commercial LWRs, Nucl. News. 53 (2014)
83–91.
[5] BISCHOFF, J., DELAFOY, C., VAUGLIN, C., BARBERIS, P., ROUBEYRIE, C., PERCHE, D., et al., AREVA NP’s
enhanced accident-tolerant fuel developments: Focus on Cr-coated M5 cladding, Nucl. Eng. Technol. 50 (2018) 223–
228.
[6] TANG, C., GROSSE, M.K., TRTIK, P., STEINBRÜCK, M., STÜBER, M., SEIFERT, H.J., H2 permeation behavior of
Cr2AlC and Ti2AlC MAX phase coated Zircaloy-4 by neutron radiography, Acta Polytech. 58 (2018) 69–76.
[7] TANG, C., STEINBRUECK, M., STUEBER, M., GROSSE, M., YU, X., ULRICH, S., et al., Deposition,
characterization and high temperature steam oxidation behavior of single-phase Ti2AlC-coated Zircaloy-4, Corros. Sci.
135 (2018) 87–98.
[8] TANG, C., KLIMENKOV, M., JAENTSCH, U., LEISTE, H., RINKE, M., ULRICH, S., et al., Synthesis and
characterization of Ti2AlC coatings by magnetron sputtering from three elemental targets and ex-situ annealing, Surf.
Coatings Technol. 309 (2017) 445–455.
[9] TANG, C., JIANU, A., STEINBRUECK, M., GROSSE, M., WEISENBURGER, A., SEIFERT, H.J., Influence of
composition and heating schedules on compatibility of FeCrAl alloys with high temperature steam, J. Nucl. Mater. 511
(2018) 496–507.
[10] DRYEPONDT, S., UNOCIC, K.A., HOELZER, D.T., MASSEY, C.P, PINT, B.A., Development of low-Cr ODS FeCrAl
alloys for accident-tolerant fuel cladding, J. Nucl. Mater. 501 (2018) 59–71.
[11] KREJCI, J., SEVECEK, M., KABÁTOVÁ, J., MANOCH, F., KOČÍ, J., CVRČEK, L., et al., Experimental behavior of
chromium-based coatings, in: TopFuel (2018) 1–14.
92
FUEL PERFORMANCE CODES APPLIED TO DESIGN BASIS ACCIDENTS (DBA)
AND TO DESIGN EXTENSION CONDITIONS (DEC)
(Session 3)
Chairperson
A. BOULORÉ
France
93
ASSESSMENT OF HIGH CONDUCTIVITY CERAMIC FUEL CONCEPT UNDER
NORMAL AND ACCIDENT CONDITIONS
D.S. GOMES, A. ABE, A.T. SILVA, R.O.R. MUNIZ

Nuclear and Energy Research Institute–IPEN/CNEN, Nuclear Engineering Center–CEN,
Av. Prof. Lineu Prestes 2242, São Paulo, SP, Brazil,
C. GIOVEDI, M.R. MARTINS,

LabRisco, Universityof São Paulo,
Av. Prof. Mello Moraes 2231, São Paulo, SP, Brazil.
Abstract
After the Fukushima Daiichi accident, the high conductivity ceramic concept fuel has been revisited. The thermal
conductivity of uranium dioxide used as nuclear fuel is relatively low, as consequence fuel pellet centerline reaches high
temperatures, high fission gas release rate, increase of fuel rod internal pressure reducing the safety thermal margin. Several
investigations had been conducted in framework of ATF (Accident Tolerant Fuel) using different additives in ceramic fuel
(UO2) in order to enhance thermal conductivity in uranium dioxide pellets. The increase of the thermal conductivity of fuel
can reduce the pellet centerline temperature, consequently less fission gas releasing rate and the low risk of fuel melting, hence
improving significantly fuel performance under accident conditions. The beryllium oxide (BeO) has high conductivity among
other ceramics and is quite compatible with UO2 up to 2200°C, at which temperature it forms a eutectic. Moreover, it is
compatible with zircaloy cladding, does not react with water, has a good neutronic characteristics (low neutron absorption
cross-section, neutron moderation). This work presents a preliminary assessment of high conductivity ceramic concept fuel
considering UO2-BeO mixed oxide fuel containing 10 wt% of BeO. The FRAPCON and FRAPTRAN fuel performance codes
were conveniently adapted to support the evaluation of UO2-BeO mixed oxide fuel. The thermal and mechanical properties
were modified in the codes for a proper and representative simulation of the fuel performance. Theobtainedpreliminary results
show lower fuel centerline temperatureswhen compared to standard UO2 fuel, consequently promoting enhancement of safety
margins during the operational condition and under LOCA accident scenario.
1. INTRODUCTION
The uranium dioxide has been used as nuclear fuel for pressurized and boiling water reactors (PWR and
BWR) for a long time with an outstanding performance. Nonetheless, the research for accident tolerant fuels (ATF)
as a consequence of the Fukushima Daiichi accident brought the necessity to improve the performance of nuclear
fuels aiming to develop fuels which present enhanced accident tolerance in comparison with the standard and
existing UO2/zircaloy system widely used by the nuclear industry [1]. Nowadays, a significant effort has been
conducted in the cladding material research and investigation in order to accomplish the ATF criteria such as
corrosion resistance, good mechanical properties, lower hydrogen generation etc. Additionally, the performance
can be improved considering the fuel pellet itself, specially enhancing the thermal conductivity and reducing
fission gas release.
The low thermal conductivity of uranium dioxide used as nuclear fuel in PWR induces a high fuel pellet
centerline temperature, stores energy consequently reducing the safety thermal margins during steady state
irradiation, transients, and under accident scenarios, as for instance loss-of-coolant accident (LOCA). Due to the
poor thermal conductivity of UO2, the high fuel centerline temperature in the pellet promotes different phenomena
such as high fission gas release, steep temperature gradient resulting in high thermal stress, increase of the fuel
pellet swelling, plastic deformation, cracking, anticipation of pellet cladding mechanical interaction occurrence,
and consequently reduction of operational safety limit. In this sense, the increase of the uranium dioxide thermal
conductivity by means of the use of specific additives to the fuel matrix can reduce the fuel centerline temperature,
consequently the amount of fission gas release would be lower, less thermal stress, less swelling and deformation
allowing high burnup and enabling improved safety due to less stored energy.
In the last decades several investigations have been conducted in order to enhance the thermal conductivity
of the fuel pellet and one of the promising techniques is associated to doping the UO2 pellet with high conductivity
additive. The beryllium oxide (BeO) have shown as good candidate due to its high thermal conductivity among
oxides, high melting point, low neutron thermal absorption cross section, easily fabricated without significant
impacts in the conventional manufacturing process with an acceptable fuel cost. Additionally, to enhancing thermal
95
conductivity, beryllium oxide presents a large stability and chemical compatibility with uranium at high
temperatures as well as some neutron moderation capacity.
In order to perform a preliminary assessment of the effect of the BeO addition to enhance the nuclear fuel
behaviour, it was necessary to include the relevant properties, mainly the thermal conductivity into the FRAPCON
and FRAPTRAN fuel performance codes.
2. METHODOLOGY
2.1. Modelling Thermal and Mechanical Properties of UO2 with BeO addition
In the present paper, initially the necessary materials properties to be changed were identified considering
the MATPRO (Material Properties Data Library) existing materials properties for fuel pellet correlated directly to
temperature and a preliminary literature survey was conducted in order to assess the existing information about Be
(beryllium) and BeO (beryllium oxide), especially those related to thermal properties. Moreover, information
regarding fabrication process in order to verify the feasible amount of beryllium oxide to be mixed to the UO2
matrix [2–4]. As begin of assessment it was defined the enhanced fuel pellet with following composition: 10% of
BeO and 90% of UO2 in volume fractions, giving almost 3.10% of BeO in weight percent (wt%).
In order to define the new material properties correlations to be applied in the modified versions of the codes,
the first step consists in stablish the weight fraction of each fuel pellet component contribution, according to
following expressions [5]:
.
𝜔 = (1)
. .
.
𝜔 = (2)
. .
where 𝒗𝒊 are the volume fractions, 𝜌 are the densities and 𝜔 are the mass fractions of the two components.
2.1.1. Thermal Conductivity UO2-BeO
The thermal conductivity of UO2 pellet usually is a function of burnup, temperature, porosity, theoretical
density, and radiation damages due to change in the crystalline network or irradiation defects. The thermal
conductivity of UO2, 8.4 W‧m-1‧K-1, is significantly lower than that of BeO, 260 W‧m-1‧K-1, at room temperature.
Two type of correlations were considered before implementation in the FRAPCON code, the analytical model
proposed by D. Chandramouliand S. T. Revankar [7–9] and experimental correlations [10, 11] presented by
HALDEN (Norway) and Nippon Nuclear Fuel Development Co., Ltd (Japan). The thermal conductivity of the
UO2-BeO proposed in this work is a fitting curve of existing data considering a given BeO composition (10%) in
the fuel, this fitting curve became function only of temperature according to following expression:
.
𝐾 = 3348 𝑇 (3)
where 𝐾 is the UO2-BeO fuel pellet thermal conductivity in W‧m-1‧K-1, and T is the temperature in
Kelvin unit, the function was implemented in subroutine FTHCON of FRAPCON code.
2.1.2. Specific Heat Capacity
The heat capacity of UO2-BeO fuel pellet is given by the straightforward result of weight fraction applied to
both oxides specific heat capacities UO2 [13] and BeO [9, 12]:
𝐶𝑝 =𝜔 × 𝐶𝑝 +𝜔 × 𝐶𝑝 (4)
where:
𝐶𝑝 = 0.036 − 0.12 + 0.2 + 1.9 (5)
𝐶𝑝 was considered in equation 2.2-1 from [13].

The function above was implemented in subroutine FCP of FRAPCON code.
96
2.1.3. Enthalpy
The enthalpy of UO2-BeO fuel pellet is given by the straightforward result of weight fraction applied to both
oxides enthalpies as presented below [7] and the function was implemented in subroutine FENTHL:
𝐻 =𝜔 ×𝐻 +𝜔 ×𝐻 (6)
𝐻 = 11.1084 + 7.1245.10 𝑇 + − − 5453.21 (7)
𝐻 was considered in equation 2.2-2 from [13].
2.1.4. Melting Point and Heat of Fusion
The melting temperature of UO2 pellet is 3113 K and its heat of fusion is 2.74×105 J/kg [6, 15]; for UO2-
BeO fuel pellet, these values are 3104 K and 3.19×105 J/kg [7], respectively.
2.1.5. Thermal Expansion
Based on the weight fraction presented above, the thermal expansion of the UO2-BeO fuel pellet is given by
the following straightforward expression [7]:
=𝜔 × +𝜔 × (8)
The function was implemented in subroutine FTHEXP of FRAPCON code.
2.1.6. Mechanical Properties
The main mechanical properties for the UO2-BeO fuel pellet are obtained using volume fraction of each
component to calculate following properties: Young modulus, Shear modulus, Poisson coefficient, and strength,
as presented below [14]:
𝜌 =𝜔 ×𝜌 +𝜔 ×𝜌 (9)
𝜈 =𝜔 ×𝜌 (10)
Then, the mechanical properties for the UO2-BeO fuel pellet are given by the following straightforward
expressions:
𝑌 =𝜈 ×𝑌 +𝜈 ×𝑌 (11)
𝑆ℎ𝑒𝑎𝑟 =𝜈 × 𝑆ℎ𝑒𝑎𝑟 +𝜈 × 𝑆ℎ𝑒𝑎𝑟 (12)
𝑃𝑜𝑖𝑠𝑠𝑜𝑛 =𝜈 × 𝑃𝑜𝑖𝑠𝑠𝑜𝑛 +𝜈 × 𝑃𝑜𝑖𝑠𝑠𝑜𝑛 (13)
2.2. Fuel Performance Codes
The well-known FRAPCON and FRAPTRAN [15] codes, sponsored by the United States Nuclear
Regulatory Commission (U.S. NRC) for the licensing of PWR and BWR nuclear power plants, were conveniently
modified to support the evaluation of BeO/UO2 mixed oxide fuel containing 10 wt% of beryllium oxide. The
FRAPCON and FRAPTRAN codes have a material data packagecompilation, namely MATPRO [6], which
contains material properties, theoretical and experimental models and correlations considered in the simulation of
the fuel performance. The implemented modifications mainly addressed the role of the BeO to enhance the thermal
conductivity of the fuel pellet. The modifications of the previously identified subroutines related to UO2-BeO fuel
pellet properties in the fuel performance codes were carried out considering step by step manner in order to assess
the contribution of each property to the final fuel global performance. The first modified subroutine was related to
the thermal conductivity property, after following by the specific heat, the enthalpy, and finally, the mechanical
properties. For each property modification, new versions of the codes were generated and tests were carefully
performed. The verification of implemented modification was conducted by means of LOCA experiment (IFA
650-5) performed in HALDEN. It is worthwhile to mention that similar work [7] exist but consider only
97
FRAPTRAN code modification for UO2-BeO as fuel in the LOCA condition, this work considers modification on
both codes: FRAPCON and FRAPTRAN to consider the LOCA analysis.
2.3. Fuel Rod Analysis (Test Case: IFA 650-5)
The modified version of the codes (FRAPCON and FRAPTRAN) was evaluated using as test case, the data
available in the open literature related to the experiment IFA 650-5 performed in the framework of Halden Reactor
Project [16] to study the behaviour of UO2/zircaloy fuel rod under LOCA scenario.
The IFA 650-5 test fuel rod was re-fabricated from an irradiated PWR UO2/Zircaloy-4 fuel rod. The fuel had
a high average burnup of 83 MWd/kgU. The base irradiation of the full-length rod comprised 6 reactor cycles
corresponding to about 2000 effective full power days. The properties of the IFA-650.5 fuel rod are summarized
in Table 1 below. Initially, steady state condition was simulated using FRAPCON code for UO2 and UO2-BeO.
TABLE 1. FUEL ROD PROPERTIES OF IFA 650-5 TEST FUEL ROD

Fuel type PWR
Fuel material UO2
Fuel pellet diameter (mm) 9.132
Fuel pellet length (mm) 11
Fuel dish depth (mm) 0.28
Fuel dish width (mm) 1.2
Fuel density (% TD) 94.8
Fuel enrichment (w/o %) 3.5
Cladding material DX ELS0.8b
Cladding outer diameter (mm) 10.735
Cladding wall thickness (mm) 0.721
Fuel rod burnup (MWd/kgU) 83
Fuel rod total length (mm) 480
Fuel rod gap (mm) 0.0805
Fuel Rod plenum volume (cm3) 15
Fuel rod fill gas 90% Ar +10%He
Fill pressure (MPa) 4.0
2.4. Results and Discussion
The Figs 1, 2 and 3 present the evolution of the fuel centerline temperature, the internal pressure, and the
fission gas release as function of burnup for IFA 650-5 considering UO2 fuel pellet as reference and enhanced fuel
UO2-BeO pellet. For UO2-BeO fuel pellet, the results are shown according to the modified properties into
correspondent subroutines.
FIG. 1. Fuel centerline temperature for IFA 650-5 as function of burnup for reference UO2 fuel pellet and UO2-BeO fuel pellet
using FRAPCON codes (original and modified version) for steady state condition (burnup accumulation).
98
Figure 1 shows that the fuel centerline temperature is about 500°C lower for UO2-BeO fuel pellet compared
to the reference UO2 fuel pellet. The figure also shows that the governing property in the fuel pellet behaviour is
mainly due to the thermal conductivity; the modification of the other subroutines related to other properties
(specific enthalpy, specific heat and mechanical) does not promote significantly changes in the fuel pellet
centerline temperature. The result confirms that thermal conductivity plays a very important role in the fuel
temperature profile.
The Figure 1 additionally presents the Vitanza threshold [17] curve which is associated to fission gas release
rate due to the fuel centerline temperature. As it can be seen from the Vitanza curve, the fission gas release for
UO2-BeO fuel pellet do not exceed the threshold, consequently the dominant phenomena governing the fission gas
release process will be athermal up to approximatelly 40 MWd/kgU.
The evolution of the internal fuel rod pressure as function of burnup for IFA 650-5 presented in Figure 2
show that the internal pressure for the fuel rod with UO2-BeO fuel pellet is lower than that observed for the
reference UO2 fuel pellet during all the irradiation period, even considering the high burnup reached during the
steady state irradiation of the fuel rod.
FIG. 2. Internal fuel rod pressure for IFA 650-5 as function of burnup for reference UO2 fuel pellet and UO2-BeO fuel pellet
using FRAPCON codes (original and modified version) for steady state condition (burnup accumulation).
The evolution of the fission gas release as function of burnup presented in Figure 3 show that the amount of
fission gas released by UO2-BeO fuel pellet is significantly lower than that of the reference UO2 fuel pellet. Then,
the lower temperatures experienced by the UO2-BeO fuel pellet during the entire irradiation period enable the fuel
centerline temperature not to exceed the Vitanza threshold.
FIG. 3. Fission gas release for IFA 650-5 as function of burnup for reference UO2 fuel pellet and UO2-BeO fuel pellet using
FRAPCON codes (original and modified versions) for steady state condition (burnup accumulation).
99
Figure 4 shows the evolution of the internal fuel rod pressure as function of time during the LOCA transient.
Initially, this assessment was performed using coupled simulation (FRAPCON and FRAPTRAN), the original
version of FRAPCON and FRAPTRAN codes for the reference UO2 fuel pellet and, for UO2-BeO, the modified
versions of FRAPCON and FRAPTRAN codes taking to account new properties implementation. Moreover, the
steady state condition simulated using modified FRAPCON and original version of FRAPTRAN in order to verify
the consistency of FRAPTRAN modification. The curves in the figure show a slight increase of the cladding
rupture time for the UO2-BeO fuel pellet. The possible reason for a slight improvement shall be associated to
temperature boundary condition considered in the FRAPTRAN input, which is conservative assumption
considering the previous results (Figure 1) of fuel temperature. Again, it is worthwhile note the importance of
results presented by modified FRAPCON compared to the results obtained from modified FRAPTRAN in order
to evaluate the effects of the BeO addition in the global fuel performance during steady state irradiation prior to
LOCA scenario.
7
Internal Pressure (MPa)
3
UO2 – FRAPTRAN
2
UO2BeO – FRAPCON Modified / FRAPTRAN Standard
1
UO2BeO – FRAPCON Modified / FRAPTRAN Modified
0
0 50 100 150 200 250
Time (s)
FIG. 4. Internal fuel rod pressure for IFA 650-5 as function of time for reference UO2 fuel pellet and UO2-BeO fuel pellet using
FRAPCON-FRAPTRAN codes and the modified versions for LOCA condition.
3. CONCLUSION
All results obtained using modified version of FRAPCON code show an improvement of the parameters
directly associated to safety margins, specially the fuel centerline temperature reduction, consequently reduced
thermal gradient inside fuel pellet, small amount of fission gas release, lower internal pressure, all effects resulting
in the improvement of safety margins during the steady state operational condition. The results of modified
FRAPTRAN shown a slight improvement for burst time during the LOCA accident scenario. Some improvement
expected for UO2-BeO fuel during the LOCA was not clearly achieved mainly due to the assumption of same fuel
cladding temperature as boundary condition adopted for both simulations at beginning of transient. Moreover, the
future work can address the appropriate cladding temperature profile and the importance of burnup associated to
the thermal conductivity degradation and some evaluation of neutronic penalty due to reduction of uranium loading
in the UO2-BeO fuel pellet.
ACKNOWLEDGMENTS
The authors are grateful to the technical support of AMAZUL, USP and IPEN-CNEN/SP.
REFERENCES
[1] GOLDNER, F., Development Strategy for Advanced LWR Fuels with Enhanced Accident Tolerance,
http://energy.gov/sites/prod/files/061212%20Goldner%20%20NEAC%20Presentation%20(FINAL).pdf (2012).
[2] NISHIGAKI, S., MAEKAWA, K., Fabrication of BeO-UO2 Fuel Pellet, Journal of Nuclear Materials,14 (1964) 453-
458.
100
[3] BEAVER, W.W., THEODORE, J.G., BIELAWSKI, C.A., Effects of Powder Characteristics, Additives and Atmosphere
on the Sintering of Sulfate-Derived BeO, Journal of Nuclear Materials 14 (1964) 326.
[4] LI, D.S., GARMESTANI, H., SCHWARTZ, J., Modeling Thermal Conductivity in UO2 with BeO Additions as a
Function of Microstructure, Journal of Nuclear Materials 392 (2009) 22.
[5] SMITH, R. J., Enhanced Thermal Conductivity UO2-BeO Nuclear Fuel: Neutronic Performance Studies and Economic
Analyses, Master Thesis. Texas (2012).
[6] ALLISON, C.M., et alli., SCDAP/RELAP5/MOD3.1 Code Manual Volume IV: MATPRO - A Library of Materials
Properties for Light-Water-Reactor Accident Analysis, NUREG/CR-6150, EGG-2720 Washington (1993).
[7] CHANDRAMOULI, D., REVANKAR, S.T., Development of Thermal Models and Analysis of UO2-BeO Fuel during a
Loss of Coolant Accident, International Journal of Nuclear Energy (2014).
[8] ZHOU, W., LIU, R., REVANKAR, S.T., UO2-BeO Composite Fuel Thermal Property and Performance Modeling,
Journal of Energy and Power Engineering 8 (2014) 7.
[9] CHANDRAMOULI, D., REVANKAR, S.T., Study of Performances of UO2-BeO Composted Fuel Using FRAPTRAN,
Purdue University School of Nuclear Engineering Report PU/NE-12-16, Purdue University School of Nuclear
Engineering (2012).
[10] ISHIMOTO, S., HIRAI, M., ITO, K., KOREI, Y., Thermal Conductivity of UO2-BeO Pellet, Journal of Nuclear Science
and Technology 33 2 (1996) 134–140.
[11] MCGRATH, M.A., et. alli., In-reactor Investigation of the composite UO2-BeO fuel: background, results and
perspectives, The Nuclear Materials Conference, Montpellier, France, (2016).
[12] DOUGLAS, A.C., Thermodynamic Properties of Magnesium Oxide and Beryllium Oxide from 298 to 1200 K, Journal
of Research of the National Bureau of Standards A. Physics and Chemistry 67A 4 (1963) 325–329.
[13] LUSCHER, W.G., GEELHOOD, K.J., Material Properties Correlation: Comparison Between FRAPCON 3.4,
FRAPTRAN 1.4 and MATPRO, NUREG/CR-7024, (2010).
[14] ZHOU, W., and ZHOU, W., Thermophysical and Mechanical Analyses of UO2-36.4vol % BeO Fuel Pellets with
Zircaloy, SiC, and FeCrAl Claddings. Metals 8 (2018) 65.
[15] GEELHOOD, K.J., LUSCHER, W.G., BEYER, C.E., FLANAGAN, M.E., FRAPCON-3.4: A Computer Code for the
Calculation of Steady state Thermal mechanical Behavior of Oxide Fuel Rods for High Burnup, U.S.NRC, NUREG/CR-
7022, Washington (2011).
[16] MANNGARD, T., STENGARD, J., Evaluation of the Halden IFA-650 Loss-of- Coolant Accident Experiments 5, 6 and
7, Swedish Radiation Safety Authority, Sweden (2014).
[17] VITANZA, C., KOLSTAD, E., GRAZIANI, V., Analysis of Fission Gas Release Tests to High Burnup, Proc. ANS
Topical Meeting on Light Water Reactor Fuel Performance, Portland, Oregon (1979).
101
MODELLING OF ACCIDENT SCENARIOS WITH THE FINIX FUEL BEHAVIOUR
MODULE
H. LOUKUSA, J. PELTONEN, V. TULKKI

VTT Technical Research Centre of Finland Ltd.
Espoo, Finland
Abstract
The FINIX fuel behaviour module has been under development at VTT for six years for modelling traditional western
light water reactor and Russian WWER fuel. The module is designed to be implemented in various other reactor safety analysis
tools, such as neutronics and reactor dynamics codes. At VTT, it has been internally coupled with Serpent 2 to provide fuel
temperatures for neutronics, and to thermal hydraulics and reactor dynamics codes TRAB1D, TRAB3D and HEXTRAN.
FINIX will also be a part of the new VTT reactor analysis framework Kraken, the development of which is currently underway.
In such couplings, the fuel temperature distribution is one of the most important parameters passed on to the host code. From
the start, FINIX has been able to model the fuel temperature distribution in reactivity insertion accidents (RIA) comparably to
established codes such as FRAPTRAN. The irradiated state was typically taken into account with a restart file from FRAPCON.
Over the years, additional models have been implemented in the code, and long irradiation periods can now also be modelled
with FINIX. The average error in temperature predictions across several Halden irradiations is 6.6%, and FINIX predicts
typically 50 to 100 K higher temperatures at the highest temperatures compared to FRAPTRAN. Recently, the ability of FINIX
to model loss-of-coolant accidents (LOCA) has been investigated. The main limitation of the current FINIX version is in
mechanical modelling, as finite strain deformation cannot be accurately modelled, and the currently implemented failure
models are rudimentary. Temperature predictions of FINIX were compared to FRAPTRAN and found to be close to
FRAPTRAN until the blowdown phase, at which FINIX overpredicts the temperatures. FINIX LOCA predictions have also
been compared to results calculated by FRAPTRAN for the Halden IFA-650.5 LOCA test.
1. INTRODUCTION
The fuel behaviour models in many thermal-hydraulics or neutronics codes are typically based on simple
correlations, non-mechanical thermal elements, or even fixed values of temperature. Although they are quick to
understand and efficient to solve, such fuel models may be less-than-realistic in, for instance, transient conditions
or, in cases where fuel with extended burn-up should be considered.
The FINIX fuel behaviour module has been under development at VTT since 2012 [1], with the latest version
published in the beginning of 2019 [2]. The main use of FINIX is in coupling with other core physics codes, and
FINIX is to serve as the main fuel model in the upcoming Finnish reactor analysis framework Kraken [3]. The
advantages of developing an in-house code are numerous compared to only using codes developed elsewhere. The
code is more easily manageable and adapted to different usages. As a self-developed code does not suffer from
license limitations, the code is available to academic as well as commercial applications with ease. In addition,
code development serves as an interesting project and increases attractiveness of the field for new graduates.
The FINIX code has been designed so that it can be coupled on a source-code level, so that passing input
and output files between the codes is not necessary. FINIX includes a collection of built-in functions that can be
used for basic setup of the system, and for running the actual simulations, using a fairly high-level syntax. In
addition, FINIX has an error message system that can be used to detect beyond-normal operation of the code
without aborting programme execution. Because of the direct coupling on a source-code level, FINIX allows for
low-level control of its input and output. At VTT, FINIX has already been successfully coupled with the reactor
dynamics codes HEXTRAN, TRAB1D and TRAB3D [4] and the Monte Carlo neutronics code Serpent 2 [5].
2. DESCRIPTION OF FINIX
FINIX solves the one dimensional heat equation independently in several axial nodes, with fuel rod internal
pressure being equal across these nodes. This approximation is known as the 1.5-dimensional approach and is
typically used in fuel performance codes.
The current models are sufficient for simple simulation of reactivity insertion accidents as well as, to some
degree, longer irradiation periods. All models necessary for the accurate simulation of loss-of-coolant accidents
are not present in FINIX, but in later sections some results calculated by FINIX in these cases are shown. Such
102
simulations show the current status of FINIX development and serve as guidance to direct future development of
the code.
In Table 1, the different FINIX models for physical phenomena and material properties are listed with their
respective sources. All FINIX correlations are reported in public sources, and the code contains no proprietary
models. It must be noted that not all models are necessary in each coupling, and in coupled application some
models should be superseded by the models in the coupled code. For example, the coolant models are not necessary
in coupling to thermal hydraulics, and the radial distribution model is not necessary when coupling to neutronics
codes.
TABLE 1. SOURCES OF DIFFERENT FINIX MODELS.
FINIX model Reference
FRAPTRAN [6], FRAPCON-3.4 [7], FRAPCON-4.0 [8],

Gap conductance
MATPRO [9]
Cladding heat capacity, thermal conductivity,
FRAPTRAN [6]
thermal strain
Cladding Young's modulus, Meyer's hardness,
MATPRO [9]
Poisson's ratio
Cladding yield stress PNNL [10], Shestopalov et al. [11]
Cladding creep Limbäck and Andersson [12], FRAPCON [8]
Cladding failure model Simple overstrain criterion
Steady-state: Dyce [13], Kättö [14], FRAPCON [8]; transient:
Cladding oxidation
Baker-Just [15]
Cladding oxide heat capacity NIST-JANAF [16]
Cladding oxide thermal conductivity MATPRO [9]
Fission gas release Modified Forsberg-Massih [17, 18], modified SIFGRS [19]
Coolant heat transfer Dittus-Boelter [20], Jens-Lottes [21]
Coolant properties Liquid water: IAPWS-IF97 [22]
Radial power distribution TUBRNP [23]
Pellet heat capacity, thermal conductivity and
FRAPTRAN [6], MATPRO [9]
thermal strain
Pellet relocation FRAPCON [7]
FRAPCON [7], MATPRO [9],
Pellet swelling and densification
SIFGRS [19]
Pellet densification MATPRO [9]
Pellet grain growth Ainscough [24]
3. VALIDATION
FINIX temperature, FGR and oxidation predictions have been validated against experimental data [25]. At
the current development stage, the FINIX temperature predictions are very accurate. On average, FINIX
temperature predictions are within 6.6% of experimental data, which corresponds to an average absolute error of
70 K, see Fig. 1. However, high fission gas releases are generally underpredicted, as can be seen from Fig. 2.
Data from Halden tests IFA-515, -677 and -681 has been used in the temperature validation (see Table 2). In
IFA-515 rods A1 and B1 and IFA-681 rod 5, the temperature was measured with an expansion thermometer, and
this value was compared with the average fuel centreline temperature calculated by FINIX. In the rest of the rods,
the data was obtained from thermocouples and the fuel centreline temperature at the axial node containing the
thermocouple was used in the comparison. In the fission gas release validation, both Halden results and data from
the OECD/NEA IFPE database have been used.
103
TABLE 2. HALDEN IRRADIATIONS USED IN THE TEMPERATURE VALIDATION OF FINIX.
Rod Clad type Burnup at EOL
(MWd‧kgU-1)
IFA-515.10 rA1 Zry-2 86.4
IFA-515.10 rB1 Zry-2 95.4
IFA-677.1 r2 Zry-4 29.7
IFA-681 r1 Zry-4 41.1
IFA-681 r5 Zry-4 40
FIG. 1. FINIX steady state temperature predictions compared to experimental data from Halden, where the black diagonal
line corresponds to calculated temperature being equal with the measured temperature. The red line is a linear regression line
fitted to all of the data, and the red shaded area is the prediction interval. Black dashed lines correspond to 10% deviations
from the experimental values.
FIG. 2. FINIX FGR predictions compared to experimental data, where the black diagonal line corresponds to calculated FGR
being equal with the measured FGR. The red shaded area is the prediction interval, and the darker shade within it the
confidence interval of the mean.
104
4. MODELLING OF REACTIVITY INSERTION ACCIDENTS
From the beginning, the capability of FINIX to model fuel temperature in RIAs has been good [26]. FINIX
predictions in RIA cases have been verified against FRAPTRAN, and for the most part the FINIX results are
comparable to those calculated with FRAPTRAN. Currently, the main differences between the two codes relate to
the pellet thermal expansion in RIAs, cladding mechanical behaviour, rod internal pressure and plenum
temperature.
Results from two RIA tests, calculated with both the FINIX and FRAPTRAN codes are shown below. The
first results are from the simulation of the CABRI REP-Na3 test, and the other from the BIGR RT-8 test. The
details of these tests are presented in Table 3.
TABLE 3. RIA VERIFICATION CASES OF FINIX PRESENTED IN THIS ARTICLE.

Base irradiation Transient reactor Rod Rod type Burnup at EOL (MWd‧kgU-1) Pulse width (ms)
Gravelines-5 CABRI Na-3 PWR 17×17 53.8 9.5

Kolskaya BIGR RT-8 WWER-440 60 3
The pellet cladding gap conductance calculation requires the determination of the effective gap width. The
same correlation for the effective gap width is used in both FINIX and in FRAPTRAN. The use of this correlation
has been found to yield best results in the temperature validation. The results of the two codes for the gap
conductance are also very similar until the power pulse. As the gap reopens in FRAPTRAN but not in FINIX, the
FINIX gap conductances also remain higher than the respective values calculated by FRAPTRAN.
The main difference in the mechanical solution regarding the gap width in RIAs is the pellet thermal
expansion model. In FRAPTRAN, a special model taking into account the radial expansion due to circumferential
strain is present, which increases pellet radial expansion in a RIA. The effect of this model was investigated, and
for example in the CABRI REP-Na3 case the gap does not reopen in FRAPTRAN if this model is turned off. The
cladding stress and strain behaviours are very similar between FINIX and FRAPTRAN without the RIA pellet
thermal expansion model. (see Fig. 3.)
FIG 3. Gap heat transfer coefficient (left) and pellet cladding gap width (right) in the CABRI REP-Na3 case.
The behaviour of the pellet cladding gap is also evident from the cladding hoop stress predictions between
the two codes. (see Fig. 4.) As the gap reopens in FRAPTRAN but does not do so in FINIX, the FINIX cladding
hoop stress predictions are consistently higher for the remainder of the simulation after the power pulse. The peak
of the cladding hoop stress occurs in both codes at the time of the power pulse.
105
FIG. 4. Cladding hoop stress in the CABRI REP-Na3 case.
The fuel centerline temperatures in RIAs are typically slightly overpredicted in FINIX compared to
FRAPTRAN. The Figure 5 shows a very typical result, where the difference in centerline temperature is in the
range of 50 to 100 K.
The pellet cladding gap width in the BIGR RT-8 case is calculated very similarly in FINIX and FRAPTRAN.
As can be seen from the rod internal pressure in Figure 6, the cladding failure occurs at slightly after 0.4 seconds,
and this is seen as a bump in the pellet cladding gap width. A similar bump is observed also in FRAPTRAN;
however, no cladding failure was observed in FRAPTRAN.
FIG. 5. Fuel centerline temperature (left) and pellet cladding gap width (right) in the BIGR RT-8 RIA case.
The pellet surface temperatures between the codes show similar behaviour, although FINIX predicts slightly
lower temperatures. The rod internal pressure is first calculated by FINIX to be considerably higher than in
FRAPTRAN, and subsequently a cladding failure occurs. The cladding failure is observed from the Figure 6 as
the abrupt drop in internal pressure, as the internal pressure is then set to equal the coolant pressure
The moment of rod failure, evident from the rod internal pressure, is also visible in the pellet surface
temperature evolution. At the moment of rod failure, the fill gas composition in the rod is adjusted to consist of
water vapor, which has a lower thermal conductivity than the initial fill gas helium. This leads to higher pellet
surface temperatures after rod failure.
106
FIG. 6. Pellet surface temperature (left) and rod internal pressure (right) in the BIGR RT-8 RIA case.
5. MODELLING OF LOSS-OF-COOLANT ACCIDENTS
FINIX models are limited with regards to modelling LOCAs. For example, FINIX uses the typical
infinitesimal strain assumption, which breaks down at finite strains, such as during ballooning in a LOCA. In
addition, FINIX cladding failure models are very rudimentary, and currently aid in code convergence rather than
prediction of failure time or strain. In an ongoing research project the update of FINIX mechanical modelling is
planned.
However, FINIX capabilities with regards to modelling LOCAs were investigated by modelling several
Halden LOCA tests from the IFA-650 test series. In the following, some key results calculated for the IFA-650.5
test are shown. The test rod in IFA-650.5 was irradiated up to a burnup of 83 MWd‧kgU-1 in a commercial PWR
before the LOCA test at Halden.
The rod internal pressure and cladding hoop stress in FINIX follow similar behaviour than in FRAPTRAN.
However, the predictions of FRAPTRAN are peaked higher than those of FINIX, and the failure time in FINIX is
slightly later than in FRAPTRAN. The failure time in both figures below (Figs 7(a) and 7(b)) can be seen at the
moment where the rod internal pressure drops abruptly to coolant pressure, and when the cladding hoop stress
drops to zero.
FIG. 7. Rod internal pressure (left) and cladding hoop stress (right) calculated by FINIX compared to FRAPTRAN in the IFA-
650.5 LOCA test.
The fuel centerline temperature and pellet cladding gap width were also compared. The gap width in FINIX
is calculated to be much larger than in FRAPTRAN. However, even though the difference in Figure 8 is very large,
similar behaviour of the gap width is found with FRAPTRAN in other LOCA cases. The large gap width in FINIX
also leads to much higher temperatures in FINIX compared to FRAPTRAN. Before the increase in the pellet
cladding gap width, the fuel centerline temperatures in FINIX and FRAPTRAN match well.
107
FIG. 8. Fuel centreline temperature (left) and fuel-cladding gap width (right) calculated by FINIX compared to FRAPTRAN
in the IFA-650.5 LOCA test.
Cladding elongation was also compared between the two codes, as as can be seen from Fig. 9. The reference
points for strain differ between the two codes: In FINIX, the reference point of zero strain is the as-fabricated cold
state, whereas in FRAPTRAN the reference point is the irradiated cold state. As both codes use a FRAPCON
restart file for obtaining the irradiated state parameters, the irradiated state in both codes is the same. The difference
in the reference point can be obtained by comparing the as-fabricated cold state parameters to the irradiated state
parameters. Correcting for the difference, the cladding elongation results of the two codes are more similar. Before
cladding failure, the cladding elongation results perfectly match. After cladding failure, as the failure occurs at a
different time and a different strain, the elongation results show the similar behaviour but different absolute values.
FIG. 9. Cladding elongation in the IFA-650.5 LOCA test with FINIX and FRAPTRAN. The strain reference point used by
FINIX is different than in FRAPTRAN, and the solid blue line corresponds to FINIX cladding elongation with the FRAPTRAN
reference point.
6. SUMMARY
The FINIX fuel behaviour module has been under development at VTT for several years. Its aim is to provide
an easy-to-use fuel performance interface in multiphysics couplings without sacrificing accuracy. The most
important parameter to be passed on to other physics solvers is the temperature, which in steady state cases are
calculated accurately and in transient scenarios comparably to other, established codes. To be able to perform
steady state simulations over long irradiation periods accurately, many of the typical models present in fuel
performance codes have been implemented into FINIX, sometimes in some simplified form. The current
temperature predictions compared to Halden irradiations are very accurate, with an average relative error of 6.6%.
108
The FINIX source code in the C programming language is available for OECD/NEA members from the
OECD/NEA Data Bank, and the FINIX documentation is available from the VTT research information portal [27].
ACKNOWLEDGEMENTS
FINIX development has been funded by the Finnish Research Programmes on Nuclear Power Plant Safety
SAFIR2014 (2011–2014), SAFIR2018 (2015–2018) and SAFIR2022 (2019–2022).
REFERENCES
[1] IKONEN, T., LOUKUSA, H., SYRJÄLAHTI, E., VALTAVIRTA, V., LEPPÄNEN, J., TULKKI, V., Module for
thermomechanical modeling of LWR fuel in multiphysics simulations, Annals of Nuclear Energy 84 (2015) 111–121.
[2] LOUKUSA, H., PELTONEN, J., VALTAVIRTA, V., FINIX - Fuel behavior model and interface for multiphysics
applications - Code documentation for version 1.19.1. Technical Report VTT-R-00052-19. VTT Technical Research
Centre of Finland Ltd, 2019.
[3] VALTAVIRTA, V., HOVI, V., LOUKUSA, H., A. RINTALA, SAHLBERG, V., TUOMINEN, R., LEPPÄNEN, J.,
Kraken - An upcoming Finnish reactor analysis framework, ANS Topical Meeting on Mathematics and Computations,
Portland, Oregon, 2019. Abstract submitted.
[4] IKONEN, T., SYRJÄLAHTI, E., VALTAVIRTA, V., LOUKUSA, H., LEPPÄNEN, J., TULKKI, V., Multiphysics
simulation of fast transients with the FINIX fuel behaviour module, EPJ Nuclear Sciences & Technologies 2 37 (2016).
[5] VALTAVIRTA, V, LEPPÄNEN, J, VIITANEN, T, Coupled neutronics — fuel behavior calculations in steady state
using the Serpent 2 Monte Carlo code, Annals of Nuclear Energy, 100 (2017) 50–64.
[6] GEELHOOD, K.J., LUSCHER, W.G., BEYER, C.E., CUTA, J.M., Fraptran 1.4: A computer code for the transient
analysis of oxide fuel rods. Technical Report NUREG-CR-7023, Vol. 1, Pacific Northwest National Laboratory, 2011.
[7] GEELHOOD, K.J., LUSCHER, W.G., BEYER, C.E., Frapcon-3.4: A computer code for the calculation of steady state
thermal mechanical behavior of oxide fuel rods for high burnup. Technical Report NUREG-CR-7022, Vol. 1, Pacific
Northwest National Laboratory, 2011.
[8] GEELHOOD, K.J, LUSCHER, W.G, RAYNAUD, P.A, PORTER, I.E. FRAPCON-4.0: A computer code for the
calculation of steady-state, thermal mechanical behavior of oxide fuel rods for high burnup. Technical Report PNNL-
19418, Pacific Northwest National Laboratory, 2015.
[9] LUSCHER, W.G., GEELHOOD, K.J., Material property correlations: Comparisons between FRAPCON-3.4,
FRAPTRAN 1.4, and MATPRO. Technical Report NUREG-CR-7024, Pacific Northwest National Laboratory, 2011.
[10] GEELHOOD, K.J., BEYER, C.E., LUSCHER, W.G., PNNL stress/strain correlation for Zircaloy, Technical Report
PNNL-17700, Pacific Northwest National Laboratory, 2008.
[11] SHESTOPALOV, A., LIOUTOV, K., YEGOROVA, L., Adaptation of USNRC’s FRAPTRAN and IRSN’s SCANAIR
transient codes and updating of MATPRO package for modeling of LOCA and RIA validation cases with Zr-1%Nb
(VVER type) cladding. Technical Report NUREG/IA-0209, US NRC, 2003.
[12] LIMBÄCK, M., ANDERSSON, T., A Model for Analysis of the Effect of Final Annealing on the In- and Out-of-Reactor
Creep Behavior of Zircaloy Cladding, In Zirconium in the Nuclear Industry 11 (1996) 448–468.
[13] MASSIH, A.R., FORSBERG, K., LIMBFICK, M., A model for uniform Zircaloy clad corrosion in pressurized water
reactors, Nuclear Engineering and Design 154 (1995) 157–168.
[14] KÄTTÖ, J., Corrosion and its modeling in nuclear reactor fuel cladding. Master’s thesis, Aalto University Department
of Energy Technology (2013).
[15] JUST, L.C., BAKER, L. Jr., Studies of metal-water reactions at high temperatures III. Experimental and theoretical
studies of the zirconium-water reaction. Technical Report ANL-6548, Argonne National Laboratory, 1962.
[16] CHASE, M.W. Jr., NIST-JANAF Themochemical Tables, Fourth Edition, J. Phys. Chem. Ref. Data, Monograph 9.
Technical report, NIST, 1998.
[17] FORSBERG, K., MASSIH, A.R., Diffusion theory of fission gas migration in irradiated nuclear fuel UO2, Journal of
Nuclear Materials 135 (1985) 140–148.
[18] HERMANSSON, P., MASSIH, A.R., An effective method for calculation of diffusive flow in spherical grains, Journal
of Nuclear Materials 304 (2002) 204–211.
[19] PASTORE, G., LUZZI, L., DI MARCELLO, V., VAN UFFELEN, P., Physics-based modelling of fission gas swelling
and release in UO2 applied to integral fuel rod analysis, Nuclear Engineering and Design 256 (2013) 75–86.
[20] BOELTER, L.M.K., DITTUS, F.W., Heat Transfer in Automobile Radiators of the Tubular Type, University of
California Publications in Engineering 2 13 (1930) 443–461.
[21] JENS, P.A., LOTTES. W.H., Analysis of Heat Transfer, Burnout, Pressure Drop and DensityData for High-Pressure
Water. Technical Report ANL-4627, Argonne National Laboratory, 1951.
109
[22] IAPWS. Revised Release on the IAPWS Industrial Formulation 1997 for the Thermodynamic properties of Water and
Steam. Technical Report R7-97(2012), IAPWS, 2012.
[23] LASSMANN, K., O’CARROLL, C., VAN DE LAAR, J., WALKER, C.T., The radial distribution of plutonium in high
burnup UO2 fuels, Journal of Nuclear Materials 208 (1994) 223–231.
[24] AINSCOUGH, J.B., WARE, J.O., Isothermal grain growth kinetics in sintered UO2 pellets. Journal of Nuclear Materials
49 (1973) 117–128.
[25] LOUKUSA, H., Validation of the FINIX fuel behavior code version 0.13.9. Technical Report VTT-R-06565-13. VTT
Technical Research Centre of Finland Ltd, 2013.
[26] PELTONEN, J., FINIX - Fuel behavior model and interface for multiphysics applications - Validation of version 1.19.1.
Technical report VTT-R-00135-19. VTT Technical Research Centre of Finland Ltd, 2019.
[27] VTT, VTT Research Information Portal.
http://cris.vtt.fi.
110
SAFETY ANALYSIS OF FUEL BEHAVIOUR FOR IMPLEMENTATION
LICENSING
H. PAPP
Paks II. Ltd,
Paks, Hungary
K. KULACSY
Hungarian Academy of Sciences, Centre for Energy Research,
Budapest, Hungary
Abstract
Currently four WWER-440 nuclear power plant units operate in Hungary whose service life will expire in 2037.The Paks
II. NPP will be built with two new Russian-type WWER-1200 reactors.The Paks II. NPP is currently in the Implementation
Licensing stage. According to the Hungarian regulations, the Nuclear Safety Code (NSC) requires the following: ”Independent
calculations shall be carried out for parameters which are decisive from the safety point of view.” The following requirement
applies to the depth of the calculations: “The analyses used for the demonstration of safety shall be documented in a such way
and to such a depth that they may be repeated, independently reviewed and modified to an extent necessary for the evaluation
of modifications throughout the lifetime of the nuclear power plant; furthermore, the extent of conservatisms applied and the
extent of margins available based on the analysis may be reviewed and re-evaluated.” The NSC also prescribes that “For all
initiating events included in the design basis or in the extended design basis, the fulfilment of the relevant acceptance criteria
shall be demonstrated by deterministic safety analyses.” Based on these requirements the fuel behaviour analyses have been
performed for normal operation and accident situations.
1. INTRODUCTION
Most of the nuclear power plants in the world are pressurized type; such reactors are currently operating in
Paks. Unit 1 and Unit 2 of Paks NPP received 20 years operating license extension until 2032 and 2034. Lifetime
of Unit 3 and Unit 4 was extended by HAEA until 2036 and 2037. The Paks II. NPP will be built with two new
Russian-type WWER-1200 reactors. From a construction, commissioning and operation point of view, it is
important to build units which already have operation experiences in other countries.
The Paks II. NPP is currently in the implementation licensing stage. Nuclear power plants must be licensed
with very serious safety requirements. The new units must have the safety systems and technical solutions specific
to third generation power plants that will only have the consequences of sever accident events within the power
plant.
2. MAIN FEATURES OF WWER-440 AND WWER-1200 ASSEMBLY AND FUEL ROD
The Russian units of 1200 MW electric power are usually given with AES-2006 type indication, where the
2006 Figure refers to the year when the plans of the first block were completed. Such units were built in the
Novovoronyez and Leningrad nuclear power plants.
Like most power plant reactors in the world and the current Paks units, the new reactors will also operate
with uranium dioxide fuel.
The WWER-440 and WWER-1200 reactors have very similar fuel rods. Both have 7.6 mm outer diameter
UO2 pellets in 9.1 mm outer diameter zirconium tubes. The rods are provided with zirconium plugs at the top and
bottom [2] (Table 1).
111
TABLE 1. MAIN PARAMETERS OF WWER-440 AND WWER-1200 ASSEMBLY
WWER-440 WWER-1200
Fuel rod length 2600 mm 4033mm

Active length 2480 mm 3730 mm
Pellet outer/inner diameter 7.6/1.2 mm 7.6/1.2 mm
Clad inner/outer diameter 9.1/7.8 mm 9.1/7.8 mm
He gas pressure 6 bar 20 bar
Average linear heat rate 13.8 kW/m 16.7 kW/m
Maximum linear heat rate 32.5 kW/m 42 kW/m
Assembly length 3217 mm 4570 mm
Fuel rod numbers in the assembly 126 312
Control rods no 18ps
The WWER-1200 reactor core has the same main dimensions as the WWER-1000 core, so the external
dimensions of the assemblies are the same. The WWER-1200 assemblies have 534 kgUO2, which is 7 kg more than
the WWER-1000 type. In addition to the 312 pieces fuel rods, 18 control rods are provided. In the WWER-1200
core, assemblies with gadolinium and UO2 fuel elements are also designed. (Figure 1).
FIG.1. Fuel assembly.
3. REGULATORY BACKGROUND AND REQUIREMENTS
Based on the Hungarian Nuclear Safety Code the “The independent verification of the analyses containing
the characteristics of the designs, which are decisive from a safety point of view, shall also be carried out by
different calculation methods.”
The following requirement applies to the depth of the calculations: “The analyses used for the demonstration
of safety shall be documented in a such way and to such a depth that they may be repeated, independently reviewed
and modified to an extent necessary for the evaluation of modifications throughout the lifetime of the nuclear
power plant; furthermore, the extent of conservatisms applied and the extent of margins available based on the
analysis may be reviewed and re-evaluated.”
Based on the above-mentioned requirements as the part of the Implementation Licensing documentation
independent safety analyses must be submitted to the Hungarian Atomic Energy Authority. These analyses were
carried out by Technical Support Organization experts.
112
4. SAFETY ANALYSIS FOR LOSS OF COOLANT ACCIDENT
The LOCA calculations were made using the FRAPTRAN code [3] and according to international practice
with best estimate models, conservative initial and boundary conditions. The condition of the analyzed fuel rods
at the beginning of LOCA were defined based on the results of the FUROM (Fuel Rod Modelling) code [4]
calculations.
As a first step, conservative calculations with the FUROM programme were carried out in order to select the
rods that behave in the worst way from LOCA point of view. The states of the fuel rods were analyzed in 5
MWd/kgU burnup steps. The highest internal pressure and highest power rods have been defined for UO2, 5%
Gd2O3 and 8% Gd2O3 rods, as in the highest internal pressure rods the ballooning is the easiest, whereas the highest
clad temperatures are reached by the highest power rods at any given burnup.
Finally, the worst state rods have been selected and the LOCA calculations were performed for these rods.
(Table 2.)
TABLE 2. SELECTED RODS FOR LOCA CALCULATIONS (BURNUP, PRESSURE AND LHR
PARAMETERS ARE AT THE BEGINNING OF THE LOCA) [5]
Rod type Rod No. Burnup Average LHR Pressure
(MWd/kgU) (kW/m) (MPa)
297 0.00 30.1 7.8
28 19.99 29.1 9.4
UO2 302 34.53 27.4 9.9
302 54.99 23.8 11.1
302 65.00 22.8 13.4
194 4.99 14.2 7.2
194 24.89 19.9 9.5
UO2+5% Gd2O3
194 44.76 19.3 10.4
194 64.58 11.5 12.4
105 5.00 18.9 7.4
14 25.26 24.4 10.2
UO2+8% Gd2O3
14 44.96 22.7 10.7
202 60.01 8.5 10.8
In order to simulate the normal operating irradiation of the rods, the production parameters were adjusted to
obtain the highest possible gas emissions. To this end, the gap size and the burnup were maximized (maximum
internal radius of the cladding, minimum outer radius of the pellet) and the pressure of the filling gas was also
taken as high as possible.
One of the components of the conservatism of thermohydraulic calculations is the choice of axial power
profiles: profiles that lead to the highest clad temperature for a given power were selected.
5. CONCLUSIONS
Among the criteria, the maximum temperature of the clad and the fuel, and the maximum oxidation of the
clad at both the rod and core level, are far below the criteria under the given thermal hydraulics conditions.
The calculated maximum cladding temperature was 920°C, the maximum local ECR was approx. 3%, the
maximum fuel centerline temperature was 1726°C and the maximum cladding strain was below 5%, so no fuel
rod failed.
6. NOMENCLATURES
FUROM Fuel Rod Modelling

FRAPTRAN A Computer Code for the Transient Analysis of Oxide Fuel Rods
HAEA Hungarian Atomic Energy Authority
LOCA Loss of Coolant Accident
NPP Nuclear Power Plant
NSC Nuclear Safety Code
113
REFERENCES
[1] HÓZER, Z., PÁZMÁNDI T., Új blokkok a paksi atomerőműben, Budapest, Hungary, Nukleon VII (2014) 152.
[2] HÓZER, Z., Az új paksi reaktorok üzemanyag, Fizikai szemle 12 (2015).
[3] GEELHOOD, K.J., LUSCHER, W.G., CUTA, J.M., PORTER, I.A., “FRAPTRAN-2.0: A Computer Code for the
Transient Analysis of Oxide Fuel Rods”, PNNL-19400, Vol.1 Rev2, (2016).
[4] GADÓ, J., GRIGER, Á., KULACSY, K., The fuel behaviour code FUROM and its high burn-up simulation capabilities,
Nuclear Engineering and Design 327 (2018) 274–285.
[5] KULACSY, K., SOMFAI B., Nagycsőtöréses üzemzavar fűtőelem-viselkedési elemzései (internal report, 2018).
114
FUEL BEHAVIOUR DURING DBA AND DEC: ANALYTICAL AND
EXPERIMENTAL STUDIES FOR DEVELOPMENT OF ATF
A.J. GAIKWAD, A.K. DEO, P. SHARMA, K. OBAIDURRAHMAN, S. BERA

Nuclear Safety Analysis Division (NSAD), NSARG, Atomic Energy Regulatory Board (AERB),
Anushaktinagar, Mumbai 400094,
India
Abstract
Fuel heat-up, fuel degradation in an accident and the resultant fuel failure with release of the fission product (FP) into
the primary system during Design Basis Accident (DBA) and Design Extension Conditions (DEC) are the key aspects to
demonstrate the safety of the NPPs. Post Fukushima more emphasis is also laid on the development of the Accident Tolerant
Fuel Designs (ATFD) to avoid fuel degradation and hydrogen generation. Terms like practically eliminated and ATF, need to
be substantiated with physical and analytical evidence. Along with ATF development efforts should also be dedicated in
prevention of loss of heat removal and/or quick restoration and lining up of the emergency coolant inventories with in the
capabilities/survivability of the ATF. The aspects related to DBA and DEC fuel/core modelling are evolving specially the later
one. The expectations, development, is also discussed here along with the comparison of ATF with the existing fuels. The
development of ATF may be an iterative process shuffling from nuclear requirements to materials and safety performance
while testing out of pile and in-pile. ATF aspect w.r.t reactivity loads is also elaborated for PHWRs. The developments on RIA
and high burn-up fuel are also summarised. With improved ATF fuel performance improved MHT design and configuration
are also envisaged. It is possible to configure MHT, ECCS and passive safety system in such a manner that the possibility of
prolonged loss of cooling is reduced.
1. INTRODUCTION
Fuel heat-up, fuel degradation in an accident and the resultant fuel failure with release of the fission product
(FP) into the primary system during Design Basis Accident (DBA) and Design Extension Conditions (DEC) are
the key aspects to demonstrate the safety of the NPPs. Post Fukushima more emphasis is also laid on the
development of the Accident Tolerant Fuel Designs (ATFD) to avoid fuel degradation and hydrogen generation.
Adherent ceramic coatings are being considered to preventing diffusion of oxygen to the fresh metal surface of the
cladding to prevent degradation. It is also claimed that the ATFs are tolerant to loss of cooling during accidents,
this tolerance needs to be quantified in terms of time and highest temperature which the cladding can tolerate.
Ideally the melting point of the clad can be taken as the maximum temperature which an ATF can hold. For zircaloy
clad this can be around 1800C.
The safety in terms of post clad melting consequences also need to be assessed, in terms of the hydrogen
which may be released then. DBA domain for zircaloy fuels within 800 to 1200C. ATFs need to show an improved
performance in this temperature range and beyond up to melting of clad. Various operational requirement of fuel
in terms of materials aspects including corrosion, fuel fabrication, fuel handling, fuel economy including burn-up
need to be assessed.
Terms like practically eliminated and ATF, need to be substantiated with physical and analytical evidence.
Along with ATF development efforts should also be dedicated in prevention of loss of heat removal and/or quick
restoration and lining up of the emergency coolant inventories with in the capabilities/survivability of the ATF.
These aspects are related to the NPP core cooling designs and the generation of the NPPs, which dictates the
inherent and engineered safety systems. Confinement of fission products during accidents is also a very important
aspects for ATFs.
Analytical and experimental studies are necessary to validate and assess the ATF fuels performance for the
DBA and DEC behaviour of fuel pins and fuel assembly. These studies start with clad failure and further
degradation on heat-up, based on the type of fuel and core configuration. Various coatings are being considered
for ATF cladding on both the sides. Coating on the fuel pellet may also be envisaged to retain the fission products
even at high temperature, this may call for sufficient void spaces in the pellet to accommodate the fission products
based on the burn-up envisaged. These materials also have to with stand very high temperature so the choices will
be limited.
115
2. DBA AND DEC FUEL/CORE MODELLING ASPECTS
Lumping of several fuel assemblies in the reactor core modelling and the corresponding realistic simulation
of heat transfer phenomena including radiation modelling play an important role in DBA/DEC predictions. Also
the fuel models face challenges in modelling of the deformed, non-cool-able geometrises of the overheated fuel
and the hot/deformed primary system attained during the DEC core disassembly. Certain assumptions are made to
close the fuel as well as primary system models. These modelling limitations and assumptions mandate more
research, as experimental evidences and the detailed models are still evolving as more claims are made on the
performance of ATF. The safety analysis codes need to be updated for the particular ATF design being
envisaged/tested and included in the design of the NPP. The updating of the codes will require the performance
details based on physical and analytical studies. On the fuel failure resistance aspect the creep behaviour of the
cladding is very important. The coating also should not allow penetration/diffusion of oxygen and hydrogen into
the base metal of the cladding.
Typical fuel bundle for PHWR is as shown in Figure 4. Temperature predictions after channel voiding using
2-D transverse cross sectional model (COMSOL) and a lumped code (RELAP/SCDAP5) are shown in Figures 5
and 6 for voided channel. Two different modelling approach has been used for lumped parameter code calculation.
For the first case, view factor between bundle and PT is calculated assuming two concentric cylinder with inner
cylinder circumscribing fuel bundle. While in the second case, effective view factor between fuel various pins in
a bundle and PT is calculated using Monte-Carlo simulation and further used for the temperature prediction. When
the view factors are calculated using an approximate geometry, the lumped code under-predicts the temperature
values. However, it can be noted that steady state temperature predictions of lumped code closely matches with 2-
D model predictions when appropriate view factors are used in the analysis.
Fuel modelling for DBAs and DECs are different, in case of DBAs a conservative approach is followed i.e.
peak clad temperature is important parameter to compare with the acceptance criteria. However, for DEC analysis,
Severe Accident Management Guidelines (SAMG), source term estimation etc., a realistic modelling approach is
required as the conservative assumption may lead to misleading results. In the DEC analysis, core/fuel modelling
is divided in to various zones based on the relative power to the fuel assemblies with an appropriate lumping as
shown in Figure 1. Further accident management actions and their timings are important in SAMGs. The release
of FPs based on the fuel temperature predicted decide the dose (which may be one of acceptance criteria) for
DBAs, DEC.
Typical fuel temperature variation during first phase (till moderator boil-off) of LOCA initiated DEC-B event
is as shown in Figure 2. Nodalization of fuel in core for PWR is shown in Fig. 3(a). PWR core is divided into five
ring type regions by grouping similarly powered fuel assemblies together for vertical core disassembly. Whereas
PHWR horizontal core channels are clubbed for core disassembly dictated by moderator boil off in the Calandria
Vessel. Fuel temperature for a PWR DEC behaviour is shown in Figure 3(b). Temperature starts increasing when
the water inventory in RPV starts depleting and on exhaustion of ECCS passive systems. The fuel behaviour
depicted in Figs 2 and 3(b) will change based on the ATF characteristics and design as and when they are deployed.
Some ATFs may lead to delay in fuel failure which can altogether stop the accident progression if innovative active
and passive means for heat removal are restored with this period.
116
FIG. 1. Core lumping based on elevation. FIG. 2. Variation of Fuel Temperature of Three Axial Nodes
of top Channel 1st Phase.
511 Upper Plenum 2
100
200
300 510 Upper Plenum 1
400
To Hot Legs
108
208 1
308
408 508 410 411 412 413 414
From Cold Legs 505 10 10 10 10 10 10
9 9 9 9 9 9
2
8 8 8 8 8 8
7 7 7 7 7 7
6 6 6 6 6 6
3 5 5 5 5 5 5
4 4 4 4 4 4
3 3 3 3 3 3
Downcomer 4 2 2 2 2 2 2
Core
1 1 1 1 1 1
Lower Plenum
506
FIG. 3a. Fuel temperature for a typical PWR. FIG. 3b. Fuel temperature for a typical PWR.
FIG. 4. Typical 37 pin fuel bundle for PHWR. FIG. 5. Temperature variation after channel voiding using
COMSOL.
117
(a) View factor based on Monte-Carlo Simulation (b)View factor based on outer bundle
FIG. 6. Lumped code temperature predictions.
3. ATF FUEL DESIGN/STUDIES AND COMPARISON WITH EXISTING FUEL
Modelling of detailed fuel behaviour involves multi physics in terms of nuclear, physical and chemical
interactions. Fuel behaviour analysis gets inputs from various codes such as neutronics code, system thermal
hydraulics codes, at times coupled reactor physics and thermal hydraulics codes analysis is necessary along with
structural aspect for DEC scenarios. Both in-house developed and internationally available code are used for
analysis.
While designing/studying ATF the following aspects also should be considered:
— Nuclear Properties: Load on reactivity i.e. reduction in excess reactivity/burn-up, swelling;
— Physical and Metallurgical behaviour of clad and coating, adherence, uniformity;
— Chemical Integrations mainly corrosion with coolant and other core materials;
— Heat transfer characteristics: modes of HTC on cladding surface;
— Initial oxidation thickness of clad and variation during transient;
— Variation of gap conductance during transient & PCMI;
— Pressure difference across cladding, has bearing on clad burst fuel failure;
— Avoiding hydrogen and heat generation due to Zr–water reaction: This is an additional heat source to the
cladding, and hydrogen produced by the reaction can be absorbed by the cladding and the mechanical
properties of the cladding may be degrade further;
— Cladding deformation, pin holes, ductility and creep behaviour, ballooning or geometrical distortions of fuel
assemblies may degrade the cooling. In the case of a large pressure difference across the cladding, such as
in a large break LOCA, cladding burst can occur. Avoiding clad burst will improve safety and DEC
performance;
— Power distribution in fuel element can generate additional loads/stress;
— Fission gas release: In the course of the accident, due to the fuel temperature rise, a large fraction of the
fission gas accumulated in the fuel matrix will be released to the gap and the internal pressure will increase.
During DECs, even non volatiles/refractories may get released depending on the fuel temperatures reached.
The comprehensive comparison of the expectation from the ATF w.r.t existing zircaloy is summarized in
Table 1.
118
TABLE 1. COMPARISON OF ATF WITH EXISTING FUEL.
Requirements Existing zircaloy fuel (ZF) Expectation from Accident Tolerant Fuel (ATF)
Design Caters to AOO & DBA Caters to AOO, DBA & DEC
Nuclear Neutron economy, fuel cycle length May not Preserves neutron economy
AOO Heat Transfer Characteristics (HTC): LHR, Similar/Enhanced HTC: Does not hamper Normal
clad temperature (coolant boiling & plant operations & AOO performance.
efficiency), Higher corrosion resistance, more compatible with
Corrosion i.e. compatibility with coolant & coolant & fuel.
pellet (prolonged durations. Fabrication may be more involved
Ease of Fabrication May involve certain reactivity burden
Nuclear Characteristics New material may not have desired Nu Pro
Fuel Handling Fuel Handling, Adherence of coating, tribology
studies, anticipated loads
DBA Burst Criteria and pin-hole development. Similar/Enhanced burst characteristics avoids pin-hole
ECCS Acceptance Criteria development.
Fuel Failures (none/acceptable based on Similar/Enhanced behaviour beyond 1200 C
criteria) No Fuel Failures expected even beyond 1200 C
Poor performance beyond pin hole Similar/Enhanced resistant to pin hole development &
development & further heat-up further degradation beyond 1200 C
Rewetting Stable in Rewetting with no failures
Diffusion of Hydrogen & Oxygen Resistant to diffusion of Hydrogen & Oxygen
Hydrogen Generation Resistant to Hydrogen Generation
DEC Diffusion of Hydrogen, Oxygen and Resistant to diffusion of Hydrogen, Oxygen and
Hydrogen Generation resistant to Hydrogen Generation
Similar Enhanced clad ductility and creep Similar/Enhanced clad ductility and creep properties
properties Ability to retaining coolable fuel configuration
Retaining coolable fuel configuration Resistant to Fission Products Releases
Fission Products Releases
DEC MHT-core Core disassembly & fuel degradation ATF can survive higher temperatures, the core-MHT
Design Aspects Post Core-disassembly Cooling structures also need to sustain these high temperature
to retain cool-able geometries and avoid core
disassembly
Post fuel degradation and core-disassembly need to be
amenable to cooling.
It should not lead to sudden & high amount of
hydrogen generation
Pellet FP release Less FP releases
Swelling & Cracking Accommodate swelling for high burn-ups
PCMI Resistant to PCMI with various coatings
Internal Pressure in the clad Less pressurization expected even at HiBurn
Ease of Fabrication New materials development
MHT Design & Loss of cooling Resistant to loss of cooling for prolong periods. Based
Configuration on the PHT configuration and the ECCS design,
Natural Circulation, Power supply Reliability.
With improved ATF fuel performance improved MHT design and configuration are also envisaged. It is
possible to configure MHT, ECCS and passive safety system in such a manner that the possibility of prolonged
loss of cooling is reduced. Studies were carried out on a four loop channel type natural circulation boiling water
reactor where it was possible to configure the MHT, ECCS in such a manner that even a large LOCA without
SCRAM did not lead to any temperature rise for a prolonged period. The details are not published here to limit the
text.
119
4. ESTIMATION OF REACTIVITY LOAD IN PHWR FOR VARIOUS ATF COATINGS
ATF for PHWR700 can be developed though coating of the clad on both the sides of the cladding tube. The
coating materials add reactivity load on the core due to the absorption of neutron corresponding to the coating
thickness. Impact on the full core reactivity due to coating of various materials has been studied to arrive at possible
candidate coating material for the natural uranium fuelled PHWR. Based the analysis the coating materials are
classified in three categories (as shown in Table 2) w.r.t. reactivity loads: best suitable (with small reactivity load
0 to -5 mk), may be acceptable (with medium reactivity load -5 mk to -10 mk) and not suitable (with large reactivity
load more than -10 mk). The reactivity load for both side coating thickness of 10 μm, 20 μm, 30 μm and 40 μm
are used for estimating the reactivity load for potential 24 number of zircaloy based ATF clad coating materials
for PHWR700 (Figs 7–10).
FIG. 7. Coating thickness 10 μm. FIG. 8. Coating thickness 20 μm.
FIG. 9. Coating thickness 30 μm. FIG. 10. Coating thickness 40 μm.
120
TABLE 2. CATEGORISATION OF COATING MATERIAL BASED OF NEUTRON ECONOMY
Coating Thickness μm Category Material
10 Best suitable SiC, ZrC, Al2O3, Si, TiN, NbC, Y2O3, Cr2AlC
May be acceptable C35M, Kanthal-APMT, Kanthal-D, CrAlN, C36M, C06M,
Kanthal-A1, Kanthal-AF, Ti3AlC2, CrN, Cr, TiC, Tr2AlC,
TiAlN
Not suitable TaC
20 Best suitable SiC, ZrC, Al2O3, Si, TiN, NbC, Y2O3
May be acceptable Cr2AlC
Not suitable C35M, Kanthal-APMT, Kanthal-D, CrAlN, C36M, C06M,
Kanthal-A1, Kanthal-AF, Ti3AlC2, CrN, Cr, TiC, Tr2AlC,
TiAlN, TaC
30 Best suitable SiC, ZrC, Al2O3, Si, TiN, Y2O3
May be acceptable NbC
Not suitable Cr2AlC, C35M, Kanthal-APMT, Kanthal-D, CrAlN, C36M,
C06M, Kanthal-A1, Kanthal-AF, Ti3AlC2, CrN, Cr, TiC,
Tr2AlC, TiAlN, TaC
40 Best suitable SiC, ZrC, Al2O3, Si, TiN, Y2O3
May be acceptable NbC
Not suitable Cr2AlC, C35M, Kanthal-APMT, Kanthal-D, CrAlN, C36M,
C06M, Kanthal-A1, Kanthal-AF, Ti3AlC2, CrN, Cr, TiC,
Tr2AlC, TiAlN, TaC
Fuel behaviour analysis for accident conditions can be grouped into the three types of transients:
— Power excursion accident: When a large reactivity spike is inserted by a very fast ejection of a rod control, a
power excursion occurs. Since the time duration of the accident is quite short, fuel behaviour is determined
by how much energy is generated in the fuel during the short period in the local region. Therefore, one of
the acceptance criteria for this accident is usually defined as the maximum adiabatic enthalpy per unit mass
added to the fuel element;
— Power cooling mismatch accident: Mismatch of heat generation and heat removal may cause DNB, and the
cladding temperature may rise suddenly due to the small heat removal capability of transition boiling or film
boiling. Safety criteria for this type of accident are usually defined by the cladding temperature, the number
of rods entering DNB and the amount of cladding oxidation. Flow rate and flow distribution is an important
aspect;
— Decrease of reactor cooling inventory: A loss of coolant accident (LOCA) is caused by the loss of integrity
of the primary circuit or its associated pipes and devices. Safety criteria for this type of accident are usually
defined by cladding temperature and amount of cladding oxidation.
5. EXISTING EXPERIMENTAL STUDIES ON FUEL RIA BEHAVIOUR
Reactivity initiated accidents (RIA) leads to increase in reactor power, thermal as well as mechanical load in
fuel elements thereof. Nuclear design should ensure that the power pulse width and height are manageable with
combination of inherent feedbacks as well engineered safety features [1]. RIA simulation tests have been
performed on fresh un-irradiated fuel rods, using pulse reactors in various countries. To generate matrix of RIA
response of rods, simulation tests have been conducted on about 140 pre-irradiated light water reactor fuel rods in
six different power pulse reactors.
The results of these experiments though added significant value to the overall knowledge base on RIA fuel
behaviour, direct application of the results from the pulse irradiation tests to power reactors is difficult due to
difference in the test conditions and the conditions of the fuel pellet in test rodlets. The results have brought out
that PCMI and energy deposition as a possible failure mechanism. Evolving trend for increasing the fuel burn up
has also been assessed with regard to RIA failure. In general, fuel burn-up aggravates the mechanical loading of
the cladding, while corrosion, or better the hydrogen absorbed in the cladding as a consequence of corrosion, may
under some conditions make the cladding brittle and more susceptible to failure. Experiments have also shown
121
that corrosion impairs the fuel resistance for RIA transient occurring at cold conditions, though there is no evidence
of important embrittlement effects at hot conditions, unless the cladding was degraded by oxide spalling.
The cladding failure limit for RIA is based on the NRC standard review plan, which suggests a maximum
radially averaged fuel enthalpy of 170 cal/g for BWRs and the DNB criterion for PWRs. Though In Japan, the
threshold of PCMI failure in terms of enthalpy increase (cal/g UO2) has been determined in terms of burn-up
(GWd/t). Acceptable Incremental enthalpy increase is 110 Cal/g upto 25 GWD/t BU of fuel. This limit reduced to
40 Cal/g at burnup of the range 60GWD/t [2] as shown in Table 3.
TABLE 3. PCMI FAILURE THRESHOLD IN JAPAN

Burn-up (GWd/t) Enthalpy Increase (cal/g UO2)
< 25 110
25–40 85
40–65 50
65–75 40
ATF concepts include modifying the fuel, coatings on clad, new clad materials etc. Different types of coated
zircaloy claddings, Molybdenum (alloy) cladding coated with Zr or FeCrAl, different variants of FeCrAl (solid
tube), SiC-SiC composite, MAX phase (example Ti3SiC2) and various types of layered claddings are being tested
by various organizations (Table 4) [3–16]. In house, experimental studies on Cr coated Zircaloy-4 claddings are
being carried out.
TABLE 4. COMPARISON OF COATING MATERIAL BASED ON MECHANICAL PERFORMANCE

ATF Coating Advantage Limitations
Zircaloy cladding Sic Thermal Charaterstics Limiting Temperature 650°C, not as good as SiC
coatings cladding
Ti3AlC2, High MP, K, favorable material Oxidation limited to 1450°C, 1000°C
Ti3SiC2 properties
Coating required on both the side, Quality of coating, Adherence, Uniformity, etc
Hybrid Si Ceramic Sustains high Temperatures, Hydrogen diffusion, development of bonding

Metal High Strength, Hard CMC with techniques between SiC & Zry-4
Composite Zry-4
New cladding SiC Sustains high Temperatures, Requires large pellet clad gap, Joining techniques
High Strength, Compatible with to be developed
UO2
Engineered Steels & Advanced Zr-alloys are being developed & have a long way to go
To develop the burst criterion for zircaloy-clad tubes used in Indian PHWRs, experimental clad burst
facilities were developed. A generalized correlation for the burst stress using the present and previous experimental
data has been developed. The effect of pressure-rise during the clad tube heating prior to bursting is incorporated
in the burst criterion model developed. The effect of anisotropic parameters has been considered. Predictions from
the developed criterion model have been found in good agreement with experimental results. Facilities were further
developed to study the Influence of Hydrogen Content on Burst Characteristics of Zircaloy-4 Cladding.
The development of ATF may be an iterative process shuffling from nuclear requirements to materials and
safety performance while testing out of pile and in-pile. The nuclear criteria in terms of the neutron economy may
be the starting point for the non-enriched/enriched fuel NPPs. The LWRs also cannot afford much reactivity loads
to the clad coatings as it will affect the cycle length and cost. It is advisable to carry out major portion of testing
and elimination in out of pile tests. This may include testing at the elevated temperatures in steam environment
with pre and post adherence tribology studies to meet all the fuel handling and the safety
expectations/requirements. Certain material like grapheme may be more promising but the development and
deployment of these materials may be difficult and costly also.
122
6. CONCLUSIONS
The ATF development has gained importance especially after Fukushima. Aspects such as fuel design, fuel
coating, fuel coating stability, diffusion of O2 and H2 through coating, corrosion, for delaying/avoiding fuel failures
are being explored. The next step is related to studies on out of pile and in-pile for fuel heat-up, fuel degradation
and minimal release of the fission products (FPs) for future deployment. After this the fuel design and safety
analysis code will be updated to account for the ATFs. Along with the code capabilities, also the
modelling/lumping aspects related to DBA and DEC fuel/core modelling play wide role in estimating the fuel
temperature. The development of ATF may be an iterative process shuffling from nuclear requirements to materials
and safety performance while testing out of pile and in-pile. Reactivity loads and neutron economy which affects
the fuel cycle costs/duration, are also major concerns in the development of ATF especially for heavy water
reactors.
REFERENCES
[1] JERNKVIST, L., MASSIH, A., Nuclear fuel behaviour under reactivity-initiated accident (RIA) condition: State-of-
the-art report (2010).
[2] BECK, W., et. al, Nuclear Fuel Safety Criteria Technical Review, (2012).
[3] ALLEN, T., BUSBY, J., et. al, “Materials challenges for nuclear systems,” Elsevier.
[4] CHENG, B., KIM, Y., et. al, “Improving accident tolerance of nuclear fuel with coated Mo-alloy cladding,” Elsevier.
[5] M. K.-N. E. and Technology and undefined ,“Research and development methodology for practical use of accident
tolerant fuel in light water reactors,” Elsevier (2016).
[6] ZINKLE, S., TERRANI, K., et. al, “Accident tolerant fuels for LWRs: A perspective,” Elsevier.
[7] RAMANATHAN, L.V., Scope for High Temperature Coatings in the Nuclear Field, (2016).
[8] VAN NIEUWENHOVE, R., “Overview of ATF research...” EERA workshop on Materials resistant to extreme
conditions for future energy systems, Kyiv, Ukraine (2017).
[9] BISCHOFF, J., BLANPAIN, P., et. al, Development of fuels with enhanced accident tolerance (2016).
[10] BACZYNSKI, J., High temperature steam oxidation of titanium-coated Zircaloy-2 and Titanium-Zirconium Alloys
(2014).
[11] IDARRAGA-TRUJILLO, I., LE FLEM, M., BRACHET, J., LE SAUX, M., Assessment at CEA of coated nuclear fuel
cladding for LWRs with increased margins in LOCA and beyond LOCA conditions.
[12] INTERNATIONAL ATOMIC ENERGY AGENCY, Technical Meeting on Accident Tolerant Fuel Concepts for Light
Water Reactors 2014 : ORNL, "Accident tolerant fuel concepts for light water reactors" : proceedings of a technical
meeting held at ORNL, United States of America, 13-16 October 2014. .
[13] KIM, H., KIM, I., etal, High temperature oxidation behavior of Cr-coated zirconium alloy.
[14] BRAGG-SITTON, S., "Overview of international activities in accident tolerant fuel development for light water
reactors", Presentation given at the IAEA technical working group on fuel performance and technology, Vienna,
Austria (2014) 24–25.
[15] PINT, B.A., TERRANI, K.A., et. al, Material Selection for Accident Tolerant Fuel Cladding, Metallurgical and
Materials Transactions E,vol. 2 3 (Sep.2015) 190–196.
[16] BARRETT, K., BRAGG-SITTON, S., Advanced LWR nuclear fuel cladding system development trade-off study
(2012).
123
AN INTRODUCTION TO A MULTIDIMENSIONAL REACTOR SIMULATION
ENVIRONMENT WITH THE CAPABILITY OF ADVANCED MODELLING OF
NUCLEAR FUEL UNDER ACCIDENT CONDITIONS
R. VADI, K. SEPANLOO
Nuclear Science and Technology Research Institute (NSTRI),
Tehran, Islamic Republic of Iran
Abstract
In this paper, a Simulation Environment (SE) is introduced which has the capability of performing comprehensive and
accurate analysis of different types of nuclear reactors and behaviour of their fuel under various steady-state and accident
conditions. The SE currently consists of three main novel cods, a CFD-based, multidimensional code which works in transient
and steady states for both single and two-phase flows and two separate neutronic codes. First code solves the point kinetics
equations with multi-group delayed neutron precursors using four different methods. The second code solves multidimensional
multi-group diffusion equations in transient and steady states and isotropic or anisotropic forms. Using the finite volume method
for spatial discretization has provided considerable flexibility regarding geometry and boundary conditions, so that the most
sophisticated details of fuels geometry such as spacer grids could be modelled accurately. The most important innovations
incorporated into this SE are "simultaneous solution" approach which is a novel method for coupling the neutronic and
thermohydraulic analyses based on the use of the point-implicit solver, "multi-region porous media" model which is an
improvement over the original model by modifying the method of resistance coefficients evaluation and the approach by which
this model is applied to the core of nuclear reactor and an introduction to the "improved SIDK", a temporal discretization
method for accident analysis which has been obtained by combing the key feature of the original method that is decoupling of
the prompt and delayed neutron equations with the separated solution approach of the point-implicit solver. The advantages
provided by the mentioned innovations are partially indicated by presenting some practical examples and to demonstrate the
accident modelling capabilities of this novel SE some relevant results associated with the case studies on TRR are presented
and discussed which includes analyses of two DBAs i.e. RIA and LOCA.
1. INTRODUCTION
‶Insatiable growth in energy consumption has increased the demands on the rate and efficiency of nuclear
power production systems [1]. It means higher discharge burnup, longer operation cycles and transient regimes.
In order to fulfil these demanding criteria, the technology incorporated into the design of these systems has to
improve constantly, especially with regard to geometry optimization, so that the conventional one dimensional
numerical methodologies are losing their function as reliable and accurate tools for both design purposes and safety
assessment″ [2, 31].
‶Accordingly, Advanced Modelling and Simulation Office (AMSO) of the US Department of Energy
introduced the concept of ‘predictive simulation’ in 2008 for the first time [2]. The purpose of this concept is to
remove or minimize the use of empirical correlations (mathematical based methods) to develop reactor simulation
tools by incorporating the first-principles effects of the governing physics and transcending them to the traditional
"engineering" scale [1]. AMSO has initiated two parallel programmes [3, 4] to develop a new generation of
integrated modelling and safety software according to this concept. Similar efforts have been implemented in the
European Union and other countries as well [5–8]. This article is an introduction to a reactor simulation
environment, developed within the framework of the corresponding author’s Ph.D. dissertation, as an absolutely
independent academic effort, which pursues the same objective″ [9–11].
Structure of the SE which currently consists of three main codes is presented in Section 2 along with a brief
review of the primary and secondary numerical models and treatments incorporated in the development process,
in addition to the algorithms of these codes in the combined form for coupled thermohydraulic neutronic analyses
of nuclear reactors. In Section 3, some of the most noteworthy innovations associated with the SE are briefly
introduced along with a few selective outputs which indicate the practical advantages provided by these
innovations. Lastly, in order to demonstrate the accident modelling capabilities of the SE, selective outputs of
simulation associated with two DBAs, i.e. LOCA [9] and RIA [11] on TRR (Tehran Research Reactor) are
presented in Section 4. General conclusions on the matters discussed in this paper are presented in Section 5.
124
2. STRUCTURE OF THE SIMULATION ENVIRONMENT
The SE currently consists of three main codes which have been programmed from scratch. A CFD-based
code that solves the transient and steady states thermohydraulic sets of conservation equations (including the
Navier-Stokes equations for momentum balance) in two and three dimensions, for both single and two-phase flows.
“Table 1 gives a brief review of numerical methods and special treatments incorporated into the new CFD code
along with the references in which they are explained. Many subtle modifications are applied to the original
methods mentioned in the table. These modifications are mainly focused on two points; reducing computational
resources required and applying these methods to an unstructured and hybrid mesh grid. The most noteworthy
modifications are:
— Improving method of applying the least-square scheme for gradients evaluation;
— Modifying the PISO scheme for pressure-velocity coupling;
— Using geometric-based method for AMG accelerator instead of mathematic-based one” [10].
Further information has been presented in Ref. [9].
TABLE 1. BRIEF REVIEW OF NUMERICAL METHODS AND SPECIAL TREATMENTS

NCORPORATED INTO THE CFD CODE
Title Applied method

Spatial discretization method Finite volume [12]
Direct kinetic [1]
Temporal discretization method
HOBD [13]
Solver approach Point-implicit solver [14]
General Standard k-ε [15]
Simple
Wall modelling Standard wall-function [15]
Turbulence model
General Realizable k-ε [16]
Advanced
Wall modelling Enhanced wall-treatment [17]
Linearization shape function Second-order linear function [12]
Convection terms (except for pressure
Method of evaluation of dependant Second-order upwind [18]
term in the momentum equations)
variables on faces of a numerical cell
Diffusion terms Central-difference [19]
Pressure interpolation scheme Natural-convection dominated problems Body force weighted [20]
(including the momentum equations) Other problems Central-difference [19]
Gradients evaluation method Least squares cell-based [21]
Gradient (slope) limiter Differential type [22]
Pressure-velocity coupling scheme Steady state problems SIMPLE [23]
(pressure-correction equation) Time-dependant problems PISO [24]
Under-relaxation method Variables (explicit) relaxation scheme [12]
Solution accelerator Algebraic Multi-Grid (AMG) with V-cycle recursive procedure [25]
solution method of Linear Algebraic
LU decomposition with partial implicit pivoting [26]
Equations (LAE) set
Drift flux (Mixture) model [27]
General multiphase model
Two-flow (Eulerian) model [27]
wall boiling models in the context of the Sub-cooled boiling regime RPI nucleate boiling [28]
Eulerian multiphase model Departed nucleate boiling regime Extended RPI [29]
The second code solves the point kinetics set of neutronic differential equations with up to six groups of
delayed neutron precursors using three different numerical methods i.e. Runge-Kutta method of order 4, an open-
integral method of order 5 and the Gear method and one semi-analytical method i.e. the Spectral method [9].
Among these four prominent methods, the last one revealed to be the best for coupling with the CFD-code based
on the criterion that "It works with longest possible time-step while it keeps accuracy and stability of the solution".
Further information could be found in Ref. [9].
125
The third code solves transient and steady states sets of multi-group neutron diffusion equations in one to
three dimensions and isotropic or anisotropic forms. Table 2 presents a brief survey of numerical models
incorporated into the diffusion code along with the references in which they are explained. Note that the general
secondary numerical methods and treatments, which are similar to ones mentioned in Table 1, are omitted from
Table 2.
TABLE 2. BRIEF REVIEW OF NUMERICAL MODELS INCORPORATED INTO THE DIFFUSION CODE
Title Applied method
Spatial discretization method Finite volume [12]
Direct kinetic [1]
HOBD [13]
Temporal discretization method
Original SIDK [1]
Improved SIDK of first order and second order
Solver approach Point-implicit solver [14]
The solution algorithms of these codes are represented in Fig. 1 and 2 in the combined form for coupled
thermohydraulic neutronic analyses of nuclear reactors.
FIG.1. Coupled algorithm of point-kinetic and CFD codes. FIG.2. Coupled algorithm of diffusion and CFD codes.
126
3. A BRIEF REVIEW OF THE INNOVATIONS INCORPORATED INTO THE SIMULATION
ENVIRONMENT.
In this section some of the most noteworthy innovations associated with the SE are briefly introduced along
with a few selective outputs for the cases which are not going to be covered in the next section and indicate the
practical advantages provided by the innovations.
Most important innovation is named "simultaneous solution" which is a novel approach for coupling the
neutronic and thermohydraulic analyses. In the conventional method for this purpose, as represented in Fig. 3,
each analysis is conducted separately and then it is attempted to close the gap between them and exchange
feedbacks. This task is usually accomplished by employing a complicated and costly platform. However, in the
methodology introduced in this study, i.e. the simultaneous solution approach, the "point-implicit" solver is used
whereby all steps of solving the neutronic and thermohydraulic equations along with all steps of computing and
exchanging feedbacks are transferred into the numerical iteration cycle of this solver as depicted in Fig. 1 and 2.
Therefore, three iteration cycles of the conventional method (see Fig. 3) reduce to one and consequently,
computational cost reduces significantly. More importantly, the step of using platforms and the associated
problems is fundamentally removed which is very crucial. Furthermore, flexibility of the conventional method
with regard to mesh grid, time step and the basic method applied to the both analyses is completely preserved and
even, improved.
FIG.3. Typical algorithm for conducting the conventional neutronic-thermohydraulic coupled analysis.
Although, the new generation of advanced modelling codes (similar to the CFD and diffusion codes devolved
for the SE introduced in this study) are highly flexible to the details of geometry of solution domain and this
domain could be directly modelled up to its smallest details, but the computational cost of conducting a large-scale
simulation such as the whole core of a power nuclear reactor is way beyond the level of numerical resources which
is normally available to an average researcher. As a result, transcending the application of these coded to the
127
traditional "engineering" scale is an inevitable part of all programmes indicted in Setion 1 [2]. Employing the
porous media model is one the widely regarded approaches to achieve this goal [10]. However, in the conventional
method of applying this model to the CFD codes, whole reactor core simplifies to a single porous medium and
also, the resistance coefficients that are essential to use this model are constant values. These conditions impose
significant errors and restrict the applications of model to the cases like accident analysis and simulation of pure
natural convection in an accurate manner for which mass flow rate changes during the solution or it is basically
unknown. In the "multi-region porous media" model developed and incorporated into the SE, procedure for
calculating the coefficients is modified by introducing a practical algorithm. Using this algorithm will result in
obtaining each coefficient as a function of mass flow rate. Furthermore, method of applying these coefficients to
the reactor core is modified by dividing the core into several porous media each of which comprises a single fuel
assembly and then the data on the boundaries of each porous medium are extracted during the solution and applied
as boundary conditions to the detailed model of fuel assembly to retrieve the detailed information regarding the
fuel behaviour under any sort of conditions. Full information regarding this new model in available in Ref. [10].
Here, some selective outputs obtained by employing multi-region porous media model for a research reactor
(whole pool of TRR) and a power reactor (whole core of BNPP) are presented in Fig. 4 and 5, respectively.
FIG.4. An example of the multi-region porous media model application to TRR (temperature in Kelvin) [9, 10].
‶Semi-Implicit Direct Kinetic (SIDK) is an innovative method for temporal discretization of neutronic
equations proposed by J. Banfield [1]. The key approximation of SIDK method is to substitute a time-averaged
quantity for fission source term in the delayed neutron differential equations. Hence, these equations are decoupled
from prompt neutron equations and an explicit analytical representation of precursor concentrations is obtained
which lead to significant reduction in the computational cost. Since, fission source is not known in a time-step, the
original study suggested using the constant quantity pertaining to the previous time-step for this purpose and
reduction in the size of time-step was proposed to lessen the imposed errors. However, this remedy notably
diminishes the main advantage of SIDK method. In the Improved SIDK method incorporated into the SE, the key
feature of the point-implicit solver [14] which is the separated solution of the governing equations is employed to
mitigate the mentioned drawbacks. Accordingly, the SIDK’s analytical relations are inserted into the numerical
iteration cycle of the point-implicit solver as independent explicit steps, while the quantity of fission source is not
constant but it is being updated, according to a modified formulation, using the most recent values of neutron
128
group fluxes. This remedy has resulted in attaining a much more accurate estimation of the time-averaged fission
source term in the current time-step, while significant portion of the computational costs reduction has been
retained. Here, some selective outputs obtained by employing Improved SIDK method for the LRA transient
benchmark problem [30] are presented in Fig. 6 and 7″. Further information could be found in Ref. [31].
FIG.5. An example of the multi-region porous media model application to BNPP (temperature in Kelvin).
FIG.6. Variations of average power density of the BWR core versus time for LRA problem. (Reproduced courtesy of Springer
[31]).
129
FIG. 7. Temporal sequence of normalized contours of power density distribution on the central cross-section of the BWR core
associated with LRA problem during the accident using the improved SIDK method (Reproduced courtesy of Springer [31]).
4. CASE STUDIES ON TRR
In order to demonstrate the accident modelling capabilities of the SE, selective outputs associated with
simulation of two DBAs, i.e. RIA and LOCA in TRR, which is a pool-type material test research reactor, are
presented in this section. A brief review of the relevant specifications of TRR is presented in Table 3 [9].
TABLE 3. SPECIFICATIONS AND MAIN OPERATING CONDITIONS OF TRR [9].

Title Design specifications
Thermal power 5 MW
Fuel type and materials MTR, LEU 20%, U3O8-Al (Meat), Al (Clad)
Fuel assembly dimensions 7.6×8.01×87.8 cm
Each location cross sectional dimensions 7.71×8.1 cm
Grid plate locations for fuel assemblies 9×6
No. of fuel plates in standard fuel assembly 19
Coolant material/Operating pressure Light water/0.171 MPa
-Operational condition -Forced flow primary loop
Cooling method
-Full shutdown condition -Natural circulation
Primary coolant flow rate 500 m3/h
130
Both accidents have been modelled using the algorithm delineated in Fig. 1. The LOCA in TRR involves the
flow reversal phenomenon which characterizes as a sophisticated transition from fully-turbulent forced convection
flow regime to pure passive natural circulation combined with limited reactivity insertion due to thermohydraulic
feedbacks. On the other hand, the large positive reactivity induced in the case of RIA demanded to modify the
conventional reactivity feedback formulation in framework of the point-kinetics method, especially with regard to
the void generation effects in the two-phase flow realm. All details and information concerning the RIA and LOCA
modelling could be found in Ref. [11] and [9] respectively. Here, two major outputs of the RIA and LOCA
modelling are represented in Fig. 8 and 9 respectively, in order to focus on a very crucial point.
FIG. 8. Temporal sequence of temperature distribution contours (in [°C]) on the virtual cross-section that cuts through the
centre of FA perpendicular to the fuel plates for RIA (Reproduced courtesy of Elsevier [11]).
These three dimensional and detailed oriented outputs clearly indicate that the most vulnerable position of
the reactor core to effects of both accidents is on the outer side of the two fuel plates located in the outermost of
the fuel assembly. In all cases, the pressure loss in the side passage is higher than the central one, as well.
131
(a) X=0 [mm] (on the surface of fuel plate)
(b) X=0.2 [mm] (in the boundary layer region)
(c) X=0.675 [mm] (in the free stream region)

FIG. 9. Temporal sequence of temperature distribution contours (in [°C]) of coolant on three virtual sections parallel to the
fuel plate in the hot channel for the LOCA [9] (X refers to distance from the fuel plate) (Reproduced courtesy of Elsevier [9]).
132
The detailed results have also revealed that the reason for this vulnerability is 27% mass distribution
imbalance between the central and side passages due to very narrows diameter designated for the side passage (2.1
mm). Furthermore, 15% reduction in this diameter, halfway through the grid plate which apparently designed to
avoid the possibility of reversal flow formation at outlet of the passages [11] has clearly intensified the mass
imbalance issue. Accordingly, we repeated the simulation on the solution domain for which the mentioned
reduction in diameter was removed and then the total diameter is increased by only 9% (to the utmost safe extent)
[11]. The combined effects of these two adjustments resulted in more than 44% reduction in the amount of mass
imbalance at a very reasonable cost while the flow velocity calculated at outlet of the modified side passage has
been high enough to avoid any possibility of flow reversal occurrence [11].
This practical example profoundly demonstrates that how the detailed-oriented data and the consequent
significant physical insight provided by the new generation of codes developed based on the "predictive
simulation" concept could be used to specify the vulnerable positions of the reactor core under any accident
conditions and recognize the associated reasons whereby it provides the opportunity to achieve an accurate
estimation of the effects of various possible modifications, in order to select an optimized solution which could
resolve or mitigate the consequences of accidents in an informed manner. The crucial fact is that this essential
virtue could not be achieved using the conventional one-dimensional, mathematical-based codes such as the
PARET or Relap 5, simply because they could not include the required geometric details and also the empirical
correlations incorporated into them have many recognized restrictions [2].
5. CONCLUSIONS
In this paper, a novel SE developed based on the "predictive simulation" concept was introduced. Structure
of the SE which currently consists of three main cods was described and algorithms of these codes in the combined
form for coupled thermohydraulic-neutronic analyses of nuclear reactors were presented. Then, some of the most
noteworthy innovations associated with the SE were briefly explained using a few selective outputs which indicate
the practical advantages provided by these innovations. Lastly, in order to demonstrate the accident modelling
capabilities of the SE, selective outputs of simulation associated with two DBAs, i.e. LOCA and RIA on TRR
were presented and discussed. The varied and divers results represented in this paper indicated the capability of
the SE in performing comprehensive and accurate analysis of different types of nuclear reactors (including MTR,
VVER and BWR) and behaviour of their fuel under various steady-state and accident conditions. However, the
most noteworthy conclusion might be drawn from the practical examples mentioned in Sec. 4 which demonstrated
how the detail-orientated, three-dimensional outputs provided by the SE on the real geometry of the reactor instead
of the homogenized one used by the conventional one-dimensional codes led to specifying the vulnerable positions
of the reactor core to the accident, discovering its causes and thereby proposing informed and optimized
modifications to mitigate the consequences of the accident.
6. NOMENCLATURES
SE Simulation Environment
CFD Computational Fluid Dynamics
SIDK Semi-Implicit Direct Kinetic
TRR Tehran Research Reactor
DBA Design Basis Accident
LOCA Loss of Coolant Accident
RIA Reactivity Insertion Accident
AMSO Advanced Modelling and Simulation Office
MTR Material Test Reactor
BWR Boling Water Reactor
BNPP Bushehr Nuclear Power-Plant
AMG Algebraic Multi-Grid
PISO Pressure-Implicit with Splitting of Operators
FA Fuel Assembly
133
ACKNOWLEDGEMENTS
This article has been prepared based on the researches conducted in the Ph.D. dissertation of the
corresponding author and it has been entirely funded by the authors. This personal study has no connection with
any governmental or non-governmental entities.
REFERENCES
[1] BANFIELD, J.E., “Semi-implicit direct kinetics methodology for deterministic, time-dependent, and fine-energy neutron
transport solutions”, Ph.D. diss., University of Tennessee, Knoxville, 2013. http://trace.tennessee.edu/utk.graddiss/1694.
[2] AMSO, Advanced Modelling and Simulation Office. [Online] the USA DOE/NE (December 17, 2014)
http://energy.gov/ne/nuclear-reactor-technologies/advanced-modeling-simulation.
[3] NEAMS, Nuclear Energy Advanced Modelling and Simulation programme. [Online] the USA DOE/NE (December 17,
2014)
http://neams.ne.anl.gov.
[4] CASL, The Consortium for Advanced Simulation of Light Water Reactors. [Online] the USA (December 17, 2014)
http://www.casl.gov.
[5] F-BRIDGE, Framework of Basic Research for Innovative Fuel Design for GEn IV systems project, [Online] the
European Union. (December 17, 2014)
http://www.f-bridge.eu.
[6] SMART project, Integrated Modular Advanced Reactor project. [Online] Korea Atomic Energy Research Institute
(December 17, 2014)
https://www.kaeri.re.kr.
[7] IVNET, Innovative and Viable Nuclear Energy Technology Development project. [Online] Research Group for Thermal
and Fuel Energy. Japan Atomic Energy Agency (December 17, 2014)
nsec.jaea.go.jp.
[8] ASRC, Advanced Science Research Centre. Sector of Nuclear Science Research (SNSR). [Online] Japan Atomic Energy
Agency (December 17, 2014)
asrc.jaea.go.jp.
[9] VADI, R., SEPANLOO, K., Investigation of a LOCA in a typical MTR by a novel best-estimate code, Progress in
Nuclear Energy 86 (2016) 141–161.
[10] VADI, R., SEPANLOO, K., An improved porous media model for nuclear reactor analysis, Nuclear Science and
Techniques 27 (1) (2016) 1–24.
[11] VADI, R., SEPANLOO, K., Reassessment of the generic assumptions applied to the conventional analysis of the
reactivity insertion accident in the MTRs using a novel coupled code, Progress in Nuclear Energy 93 (2016) 96–115.
[12] LEVEQUE, R.J., Finite Volume Methods for Hyperbolic Problems, Cambridge University Press, UK, 2004.
[13] GINESTAR, D, High-order backward discretization of the neutron diffusion equation. Ann. Nucl. Energy 25 (3) (1998)
47–64.
[14] CHORIN, A. J., Numerical solution of Navier-Stokes equations, Mathematics of Computation 22 (1968) 745–762.
[15] LAUNDER, B. E., SPALDING, D. B., Lectures in Mathematical Models of Turbulence, Academic Press, London,
England, 1972.
[16] SHIH, T. H., LIOU, W. W., SHABBIR, A., YANG, Z., ZHU, J., A New k-ε Eddy-Viscosity Model for High Reynolds
Number Turbulent Flows - Model Development and Validation, Computers Fluids 24 (3) (1995) 227–238.
[17] CHEN, H. C., PATEL, V. C., Near-Wall Turbulence Models for Complex Flows Including Separation, AIAA Journal
26 (6) (1988) 641-648.
[18] BARTH, T.J., JESPERSEN, D., The design and application of upwind schemes on unstructured meshes, Technical
Report AIAA-89-0366, AIAA 27th Aerospace Sciences Meeting, Reno, Nevada, 1989.
[19] LEONARD, B.P., The ULTIMATE conservative difference scheme applied to unsteady one dimensional advection,
Comp. Methods Appl. Mech. Eng 88 (1991) 17–74.
[20] PATANKAR, S.V., Numerical Heat Transfer and Fluid Flow, Hemisphere, Washington, DC, 1980.
[21] ANDERSON, W., BONHUS, D.L., An Implicit Upwind Algorithm for Computing Turbulent Flows on Unstructured
Grids, Computers Fluids 23 (1994) 1–21.
[22] WANG, Z.J., A Fast Nested Multi-grid Viscous Flow Solver for Adaptive Cartesian/Quad Grids, International Journal
for Numerical Methods in Fluids 33 (2000) 657-680.
[23] RHIE, C.M., CHOW, W.L., Numerical Study of the Turbulent Flow past an Airfoil with Trailing Edge Separation, AIAA
Journal 21 (1983) 1525–1532.
[24] ISSA, R. I., Solution of Implicitly Discretized Fluid Flow Equations by Operator Splitting, Journal of Computational
Physics 62 (1986) 40–65.
134
[25] HUTCHINSON, B.R., RAITHBY, G. D., A Multi-grid Method Based on the Additive Correction Strategy, Numerical
Heat Transfer 9 (1986) 511–537.
[26] PRESS, W.H., TEUKOLSKY, S.A., VETTERLING, W.T., FLANNERY, B.P., “Numerical Recipes in C”, Cambridge
University Press, UK, 2th edition (1992).
[27] FERZIEGER, J.L., PERIC, M., “Computational Methods for Fluid Dynamics”, Springer-Verlag, Heidelberg. 1996.
[28] KURUL, N., PODOWSKI, M.Z., “On the modelling of multidimensional effects in boiling channels”, In Proceedings of
the 27th National Heat Transfer Conference, Minneapolis, Minnesota, USA. 1991.
[29] HOYER, N., Calculation of dryout and post-dryout heat transfer for tube geometry, International Journal of Multiphase
Flow 24 (1998) 319–334.
[30] COMPUTATIONAL BENCHMARK COMMITTEE OF THE MATHEMATICS AND COMPUTATION DIVISION
OF THE AMERICAN NUCLEAR SOCIETY, National energy software centre: benchmark problem book, Technical
Report, ANL-7416, supplement 3, Argonne, Illinois, USA, 1985.
[31] VADI, R., SEPANLOO, K., An improved semi-implicit direct kinetics method for transient analysis of nuclear reactors,
Nuclear Science and Techniques 30 11 (2019) 162.
135
SIMULATION OF IFA650.5 EXPERIMENT USING THE FULLY COUPLED
SPACE-FRAPTRAN CODE SYSTEM
H.C. KIM
Nuclear Fuel Safety Research Division,
Korea Atomic Energy Research Institute,
989–111, Daedeok-daero, Yuseong-gu, Daejeon, 34057,
Republic of Korea
S.W. LEE, K.D. KIM

Thermal Hydraulics and Severe Accident Research Division,
Republic of Korea
Abstract
A system analysis code, SPACE was coupled with a fuel transient analysis code, FRAPTRAN using a dynamic link
library (DLL) scheme. In this coupling scheme, FRAPTRAN handles the heat conduction model including fuel deformation of
single fuel rod and, SPACE deals with the calculation of all hydraulic cells and heat structures except for the fuel rod coupled
with FRAPTRAN. This coupled code has been developed for coping with the change of auditing requirements for the
emergency core cooling system (ECCS). To couple SPACE with FRAPTRAN, coupling scheme was proposed and coupling
variables were defined. For reflood phase, new boundary conditions for FRAPTRAN calculated by SPACE have been
developed in the view. In this work, IFA-650.5 experiments, which was a LOCA (loss of coolant accident) test without reflood
phase using a high burnup fuel in order to validate the SPACE-FRAPTRAN coupled code. From the simulation results of
SPACE-FRAPTRAN coupled system against Halden IFA-650.5 test, the cladding and heater temperatures predicted by code
agreed well with experimental data. In addition, it was revealed that the most important factor for the cladding and heater
temperature in the IFA-650 test is the surface-to-surface radiation heat flux and the fuel model had little effects on them.
Therefore, it is required that surface-to-surface radiation enclosure model should be simulated appropriately for the Halden
IFA-650 test to predict the proper behaviour of the cladding temperature.
1. INTRODUCTION
SPACE [1] is a system analysis code for design and safety analysis of nuclear power plants (NPPs),
developed by Korea Hydro and Nuclear Power (KHNP), Korea Atomic Energy Research Institute (KAERI), Korea
Electric Power Corporation Engineering and Construction (KEPCO-E&C), and Korea Electric Power Corporation
Nuclear Fuel (KEPCO-NF). SPACE is equipped with thermal hydraulic analysis models, heat structures analysis
models including fuel behaviour analysis model, trip and control systems etc. However, the fuel deformation model
of SPACE is not a delicate mechanistic model but is similar to that of RELAP5 [2] based on NUREG-0630 model
[3] developed in the early 1980s. Recently, revision of acceptance criteria for emergency core cooling system
(ECCS) has been proposed by Korea Institute of Nuclear Safety (KINS) in Korea [4]. In the revised criteria, it is
required to consider the improvement in our understanding of zirconium alloy cladding fuel behaviour during a
LOCA transient. For example, the fuel rod undergoes exothermic high temperature oxidation, cladding burst,
thermo mechanical deformation of cladding, and fuel fragmentation, relocation, and dispersion (FFRD) during
loss of coolant accident (LOCA). Therefore, to simulate proper fuel behaviour during LOCA, the system code
should incorporate transient fuel models in view of the best-estimated calculation. FRAPTRAN [5] developed by
U.S. Nuclear Regulatory Commission (NRC) is equipped with considerably delicate fuel model but limited thermal
hydraulics model compared with SPACE. Therefore, to compensate for the weakness and preserve the analyzing
capability of both codes, a linkage function between the two codes using the features of the Dynamic Link Library
(DLL) has been developed. In this coupling scheme, FRAPTRAN deals with heat conduction including fuel
deformation of single fuel rod and SPACE takes charge of the calculation of all hydraulic cells and heat structures
except the fuel rod coupled with FRAPTRAN. It is observed that the output of FRAPCON [6], which calculates
the steady state fuel characteristic related with fuel burnup, is incorporated to generate the initial condition for the
input of FRAPTRAN.
136
In this work, IFA-650.5 experiment [7], which was a LOCA (loss of coolant accident) test without reflood
phase using a high burnup fuel in order to validate the SPACE-FRAPTRAN coupled code. From the simulation
results of SPACE-FRAPTRAN coupled system against Halden IFA-650.5 test, the cladding and heater
temperatures predicted by code agreed well with experimental data. In addition, it was revealed that the most
important factor for the cladding and heater temperature in the IFA-650 test is the surface-to-surface radiation heat
flux and the fuel model had little effects on them. Therefore, it is required that surface-to-surface radiation
enclosure model should be simulated appropriately for the Halden IFA-650 test to predict the proper behaviour of
the cladding temperature.
2. DEVELOPMENT OF SPACE-FRAPTRAN COUPLED CODE SYSTEM
2.1. The coupling algorithm
In SPACE-FRAPTRAN coupling calculation, SPACE is a master programme and FRAPTRAN is coupled

with SPACE as a dynamic link library (DLL). Figure 1 shows the overview of the linkage between SPACE and
FRAPTRAN. FRAPTRAN DLL named as “M-TRAN” is called by SPACE and data transfer from SPACE to
FRAPTRAN is conducted through the functions name by “SET_”. Afterwards, FRAPTRAN calculates heat
conduction including fuel deformation and returns the heat flux, surface temperature, and outer diameter to SPACE
through “GET_” functions. Functions with a prefix “CALL_” are the functions that call the subroutines in
FRAPTRAN such as “CALL_SFRAPTRAN” for initialization and “CALL_CRANK6” for time advancement
during transient. If required, the FRAPCON output file can be used for burnup fuel simulation.
FRAPCON Burnup data

Output
FRAPTRAN M-TRAN DLL

Input (Fuel analysis)
, P, HTC, , Interface , ,
LHGR, variables Heat flux
SPACE SPACE
Input (System analysis)
Export & Import

M-TRAN.DLL DLL Functions SPACE.EXE
SET_DT
Global SET_HTCBULK Global
interface SET_PKW interface
SET_COOLPRS
variables variables
GET_NAXN
GET_HEATFLUX
GET_HEATRADIUS
GET_HEATTEMP_E
CALL_MTRAN
CALL_CRANK6
FIG. 1. Overview of SPACE-FRAPTRAN coupling scheme.
To couple SPACE/FRAPTRAN code system, coupling methodology should be defined because each code
system already was used and validated with their own methodology. As shown in Figure 2, we proposed coupling
methodology of two codes for steady state and transient maintaining each calculation flow and I/O (Input/Output)
system.
137
FIG. 2. Calculation sequence of SPACE-FRAPTRAN.
At the beginning of fully coupled calculation, SPACE performs steady state calculation with input file of
SPACE. For this calculation, SPACE employs its heat structure instead of fuel rod. We call the 1st SS (steady state)
calculation which performs null transient calculation without FRAPCON/FRAPTRAN. Once SPACE completes
the 1st SS, it calls M-TRAN which is modulized FRAPTRAN to be implemented as DLL. For the first calling, M-
TRAN initiates input variables and stores FRAPCON result file to apply burnup dependent variables. The M-
TRAN starts fuel stabilization which increase power gradually to stabilize fuel thermo mechanical behaviour. We
call the 2nd SS for fuel stabilization. Once fuel stabilization is completed, fully coupled SPACE/FRAPTRAN is
ready to start transient calculation for LOCA.
2.2. Coupling variables
For coupling calculation, several interface variables have been defined. First, the interface variables from
SPACE to FRAPTRAN are as follows: 1) total heat transfer coefficient (HTC), which is a simple summation of
all HTCs; 2) temporary heat flux, which is based on previous wall temperature and current HTCs; 3) linear heat
generation rate (LHGR); 4) coolant pressure of each cell. Among these variables, total HTC and temporary heat
flux are used for boundary condition of heat conduction in FRAPTRAN. The LHGR is also provided as a heat
source of the heat conduction model in FRAPTRAN. In addition, the coolant pressure is essential for fuel
deformation model. Second, the interface variables from FRAPTRAN to SPACE are as follows: 1) surface heat
flux; 2) fuel center and clad outer surface temperatures; 3) clad outer diameter. In SPACE, surface heat flux, clad
surface temperature, and clad outer diameter are used for calculating the energy conservation equations, wall HTCs,
and fluid deformation, respectively. The fluid deformation model deals with time dependent flow area and physical
models related with hydraulic diameter. Table 1 shows the summary of interface variables for the SPACE-
FRAPTRAN coupling method. Currently, the time increment for temporal advance is determined by SPACE and
provided for FRAPTRAN.
138
TABLE 1. COUPLING VARIABLES FOR SPACE-FRAPTRAN
Calling Variable name Content
module
Time increment Size of Time step
Power Linear Heat Generation Rate (LHGR)
M-TRAN CoolPress Coolant pressure
Htc Heat transfer coefficient of cladding surface
Tbulk Coolant T
Outdia Cladding outer diameter (incl. oxide thickness)
SPACE Heatflux Cladding heat flux
Tsurf Cladding surface T
2.3. Coupling of convection model for reflood phase
In coupled calculation, heat conduction inside the fuel rod and convection at the fuel rod surface are solved
using FRAPTRAN and SPACE, respectively. The boundary condition at the right surface in heat conduction
equation of FRAPTRAN.
It is observed that only one HTC and fluid temperature are used to solve the heat conduction equation in
FRAPTRAN. However, various HTCs and fluid temperatures exist in SPACE because SPACE deals with three-
field (droplet, vapour and liquid) conservation equations and thus, it is required that a representative HTC and bulk
fluid temperature should be provided by SPACE. Therefore, HTCs and fluid temperatures could be averaged for
over current time step at each axial position with a type of weighting factors but every averaging method exhibits
certain weak points in that singular point exists. To overcome the shortcomings, we modified the boundary
condition of the heat conduction equation for coupled calculation as follows [8]:
𝜃 =𝜃 +ℎ (𝑇 − 𝑇 ) = A𝑇 +𝐵 (1)
𝑇 = (2)
where 𝜃 is current heat flux, 𝜃 is a heat flux based on new HTCs and previous surface temperature, ℎ
is a total HTC, 𝑇 is surface temperature in current time step, 𝑇 is surface temperature in previous time step,
A and B are constant based on thermal properties.
Eq. (1) represents new boundary conditions for FRAPTRAN thermal calculation to obtain current surface
temperature. Therefore, temporary heat flux, 𝜃 and total HTC, ℎ is added to the interface variables for the
coupling of SPACE-FRAPTRAN.
3. SIMULATION OF IFA-650.5 EXPERIMENT
3.1. Thermal hydraulic modelling for space
An overall layout of SPACE modelling for IFA-650.5 [7] is presented in Figure 3. As for the fluid system,
the inlet flow from TFBC (Temporal Face Boundary Condition) of C100 enters lower plenum (C110) and is spilt
into two channels for fuel rod and heater through cross flow (C115). Both channels are mixed at the top of upper
plenum (C160) and finally exit to the outlet TFBC (C300) via the outlet pipeline (C200). Active fuel region and
heater region are divided into 9 axial nodes. Flow and pressure boundary condition are applied to inlet and outlet
TFBC, respectively. Blowdown valve to simulate the LOCA and spray injection are modelled as TFBC-999 and
TFBC-555, respectively. For the heat structures, there are three heat structure components to simulate a fuel rod
(H130), electric heater (H140) and pressure flask (H150), respectively. It was pointed out that the radiation heat
transfer played a very important role in behaviour of the cladding temperature in IFA-650 test [9], therefore, a
radiation enclosure model is applied into the facing surface of the fuel rod, heater and flask. A convective heat
transfer condition is applied to the outer surface of the flask and, the heat transfer coefficient and bulk fluid
139
temperature are assumed to be 3000 W/m2K and 235C which is the coolant temperature of heavy water in Halden
reactor. A schematic diagram of the test rig for SPACE is presented in Figure 2.
FIG. 3. SPACE Nodalization for test rig.
3.2. Fuel modelling for FRAPTRAN
The father rod is a UO2 rod with a SRA Zircaloy-4 cladding, irradiated during six annual cycles up to 83
MWd/kgU (rod average) under moderate power conditions (average liner heat rate used without taking into
account the gamma power generated in the coolant: 375, 280, 220, 200, 180, and 180 W/cm). At the end of the
irradiation on the tested part of the rod, the corrosion layer thickness measured are mean 65 µm and maximum
80 µm. The measured hydrogen content is around 650 ppm. Fig. 4 shows the calculation results for the base
irradiation of IFA-650.5. Table 2 shows fuel rod information of IFA-650.5. This experiment was performed using
an IFA-650 device, and the HWR report presents the experiment.
140
TABLE 2. FUEL ROD INFORMATION OF IFA-650.5
Item IFA-650.5
Active length (mm) 480

Rod diameter (mm) 10.73
Pellet diameter (mm) 9.144
Pellet height (mm) 11.0
Pellet dish depth (mm) 0.28
Pellet outer surface roughness (mm) 0.002
Cladding type Zry-4
Cladding inner surface roughness (mm) 0.0005
Cladding thickness (mm) 0.712
Number of coils in spring 20
Spring height (mm) 28.1
Total plenum volume (m3) 1.5×10-5
Fill pressure (bar, RT) 40
Gas composition He 10% Ar 90%
Number of axial node 9
For the simulation, the following assumptions were made:

— The plenum temperature was set as TOA plus 5.6 K;
— The initial free volume is 15 cm3.
To simulate the transient behaviour, a FRAPTRAN input is generated as follows: Burnup file required
(FRAPCON file); gap gas, 95 % Ar; 5%He/ 40 bar at Room Temperature; simulation time, 0 to 500 s (time step,
0.01 s); HT oxidation, C-P model; deformation, Balon2 model.
FIG. 4. Power and burnup history of IFA-650.5 rod.
141
4. SIMULATION RESULTS
4.1. Clad and Heater Temperature
Comparison of clad and heater temperature between coupled code and experiment are shown in Figure 5. As
shown in the figure, clad and heater temperature show a very good agreement with the experimental results. This
good agreement resulted from not the fuel deformation model but the radiation enclosure model. Figure 6 is a
simulation results of SPACE standalone calculation in which the fuel deformation model was not used but the
same radiation enclosure model as that of coupled calculation was applied. Comparing Figure 5 and Figure 6, it is
found that there is little difference between the simulation results. Therefore, it can be induced that the fuel
deformation model has little effect on the clad temperature in IFA-650 tests. Figure 7 shows a comparison of
predicted and experimental rod internal pressure. Fuel clad was ruptured at 155 s in the simulation and at 178 s in
the experiment. As shown in Figure 8, only the coupled code calculates hoop strain of each axial node because
heat structure of SPACE cannot take into account fuel deformation. Hoop strain at node 4 where rupture occurs is
the highest. After the rupture, the rod internal pressure decreased very slowly in the experiment, whereas predicted
one drastically fell off to the level of coolant pressure. This discrepancy is caused by high content of hydrogen in
the clad. Due to the high content of hydrogen, the clad became brittle and the rupture area was very small. Finally,
the rod internal pressure decreased slowly due to a small rupture area.
FIG. 5. Clad and heater temperature (coupled).
FIG. 6. Clad and heater temperature (SPACE standalone).
142
FIG. 7. Rod internal pressure.
FIG. 8 Hoop strain calculated by the coupled code.
5. CONCLUSIONS
SPACE/FRPATRAN fully coupled code system has been developed to take into account fuel behaviour for
safety analysis. The coupling algorithm and sequence were proposed to maintain own methodology of each code.
To overcome various HTCs and fluid temperature in SPACE, new coupling variables were proposed. In order to
validate the SPACE/FRAPTRAN code system, IFA-650.5 experiment was simulated. From the simulation results
of SPACE-FRAPTRAN coupled system against Halden IFA-650.5 test, the cladding and heater temperatures
predicted by code agreed well with experimental data. In addition, it was revealed that the most important factor
for the cladding and heater temperature in the IFA-650 test is the surface-to-surface radiation heat flux and the
fuel model had little effects on them. Therefore, it is required that surface-to-surface radiation enclosure model
should be simulated appropriately for the Halden IFA-650 test to predict the proper behaviour of the cladding
temperature.
143
REFERENCES
[1] HA, S.J., et al., Development of the SPACE Code for Nuclear Power Plants, Nuclear Engineering and Technology 43
(2011) 45–62.
[2] RELAP5/MOD3.3., Code Manual Volume I: Code Structure, System Models, and Solution Methods, NUREG/CR-
5535/Rev P4-Vol I, Information Systems Laboratories, Inc., Oct. 2010.
[3] POWERS, D.A., et al., Cladding Swelling and Rupture Models for LOCA Analysis, NUREG-0630, U. S. Nuclear
Regulatory Commission, April 1980.
[4] BANG, Y.S., et al., Technical Basis for the Revision of ECCS Acceptance Criteria of Domestic PWR Plants, KINS/RR-
1686, Korea Institute of Nuclear Safety, 2017.
[5] GEELHOOD, K.J., et al., FRAPCON-4.0: A Computer Code for the Calculation of Steady-State, Thermal mechanical
behavior of oxide fuel rods for high burnup, PNNL-19418, Vol. 1, 2015.
[6] GEELHOOD, K.J., et al., FRAPTRAN-2.0: A Computer Code for the transient analysis of oxide fuel rods, PNNL-19400.
Vol.1 Rev.2 (2016).
[7] KEKKONEN, L., LOCA Testing at HALDEN; The PWR Experiment IFA-650.5, HWR-839, OECD HALDEN Reactor
Project, Jan. 2007.
[8] LEE, S.W., et al., Coupled Calculation of SPACE and FRAPTRAN, Transactions of the Korean Nuclear Society Spring
Meeting, Jeju, Republic of Korea, May 17–18, 2018.
[9] CHOI, T.S., Code Assessment against IFA-650 LOCA Tests, S06NX08-F-1-TR-041 Rev.0, KEPCO-NF, July 2016.
144
THERMO MECHANICAL PERFORMANCE OF TRISO COATED FUEL
PARTICLE USED IN HIGH TEMPERATURE GAS COOLED REACTOR
S. LIU, Y. ZHOU, P. CHEN, W. LI, S. GAO, H. PANG, C. TANG, X. QIU

Nuclear Power Institute of China,
Science and Technology on Reactor System Design Technology Laboratory,
Chengdu 610200, China
Abstract
Three dimensional characterized unit was used for TRISO particle with UN as the kernel to analyze the influence of PyC
layers and internal pressure on the thermal mechanical performance of TRISO coated fuel particles by using the multi physics
coupling software COMSOL. The influence of operation conditions including steady state and reactivity introduced accident
(RIA) conditions on the thermal mechanical performance of TRISO particle was conducted. The results indicate that the
structure integrity of SiC layer was maintained but the IPyC layer was failed in steady state condition. Thermal expansion is a
dominant factor resulting in the loss of structure integrity of the TRISO coated fuel particle under RIA condition.
1. INTRODUCTION
TRISO coated fuel particles are the key parts of Fully Ceramic Microencapsulated (FCM) accident tolerant
fuel (ATF). As the pellets of FCM fuel, TRISO particles are designed by three buffer layers which can provide
more barriers for the released fission gas under reactivity introduced accident conditions. Since FCM fuel pellets
still have about 50% NITE-SiC matrix, in order to increase uranium loading with reasonable 235U enrichment to
achieve approximated cycle length compared with LWRs fuels, uranium mononitride (UN) with high uranium
density has became a research hotspot of kernel fuel in TRISO coated fuel particles [1]. For the FCM fuel, the
diameter of kernel fuel in TRISO coated fuel particles is usually 500–800 μm; the thickness of buffer layer which
can absorb fission fragment, release irradiation damage of other layers, accommodate kernel swelling and contain
fission gas is usually 50–100 μm; the typical thickness of IPyC, SiC and OPyC layers are 30, 40 and 30 μm,
respectively. The main function of SiC layer is the pressure bearing boundary and the barrier preventing the
diffusion of fission products. Based on the neutron economy analysis [2], when the TRISO particles loading
volume ration reaches 50% and the diameter of kernel fuel reaches about 800 μm, the FCM fuel with 19%
enrichment UN kernel could reach the comparative core power and lifetime in contrast with LWRs with UO2 fuels.
A 3D thermal mechanical coupling model was developed for TRISO coated fuel particle with UN as the
kernel by using the multi physics coupling software COMSOL in this study. Performance analysis of the TRISO
coated fuel particle with typical design parameters was conducted under steady state and reactivity introduced
accident (RIA) conditions, which is useful for further optimizing the FCM fuel design.
2. MATERIAL PHYSICAL AND BEHAVIOUR MODELS
2.1. Material physical properties
The thermal conductivity of UN (KUN) is the function of burn-up (Bu), temperature (T) and porosity (p) [3],
and thermal expansion coefficient (αUN) and specific heat (Cp, UN) are as follows [4]:
.
𝐾 = 1.37𝑇 (1 − 0.025𝐵𝑢) (1)
𝛼 =7.096×10-6 + 1.409 × 10-9 𝑇 (2)

.
.
𝐶 , = 205.3834 .
+ 0.0381𝑇 + 1.061 × 10 𝑇 𝑒𝑥𝑝 − (3)
The elastic modulus of the PyC layer is isotropic which can be written as follows [5]:
Ebuffer = 11.06(1 + 0.23Φ)∙(0.9560275 + 0.00015T) (4)
The elastic modulus of the PyC layer is anisotropic, and can be calculated by the following equation [5]:
145
EPyC = 25.5 0.384 + 0.000324ρPyC ∙(1 + 0.23Φ)(0.9560275 + 0.00015T) (5)
The elastic modulus, thermal conductivity and thermal expansion coefficient of SiC layer are achieved by
the following equations [5, 6]:
962
ESiC = 460 − 0.04Texp( − ) (6)
T
17885
KSiC = +2 (7)
T
0.0178
−1.8276 + − 1.5544 × 10-5 T2 + 4.5246 × 10-9 T3 , T < 1273K
αSiC = T (8)
5.0, T ≥ 1273K
where, Ф is the fast neutron fluence, and ρ is the density.

The basic thermal mechanical physical properties of each part in TRISO coated fuel particle are listed in
Table 1.
TABLE 1. BASIC THERMAL MECHANICAL PROPERTIES OF THE RELATIVE MATERIALS [7–9]

Material Density Poisson Thermal Thermal Specific Weibull
（kg·m-3） ratio conductivity expansion heat modulus
coefficient
UN kernel 13760 0.32 — — — —
Buffer layer 1000 0.23 0.5 5.7×10-6 720 —
IPyC layer 1900 0.23 4.0 5.7×10-6 720 9.5
SiC layer 3200 0.13 — — 620 6.0
OPyC layer 1900 0.23 4.0 5.7×10-6 720 9.5
2.2. Mechanical behaviour model
The irradiation swelling relation of UN kernel is expressed by the following equation [10]:
∆V
= 4.6473 × 10-11 T 3.12 Bu0.83 (9)
V
where, ∆V and V are the volume change by the swelling and the initial volume, respectively.
The irradiation deformation of PyC layer has significant effects on the mechanical properties of TRISO
particles and the choices of computation relations between them are various. In this study, the relations of radial
and hoop irradiation strain of PyC layers are expressed by the following equations, respectively [11].
εṙ = 0.181Φ2 − 1.066Φ (10)
εθ̇ = 0.140Φ2 − 1.324Φ (11)
where, εṙ and εθ̇ are the radial and hoop irradiation strain, respectively.
There is a saturation value when the irradiation deformation of PyC layer increase as the fast neutron fluence
increase. PyC layer would be considered as failed when the fast neutron fluence reaches 8.14×1025 m-2 , and there
would be no more irradiation deformation [11]. The irradiation deformation change of buffer layer is written as
follows [12]:
εṙ = εθ̇ = − 0.176 exp( − 1.75Φ) (12)
where, εṙ and εθ̇ are the radial and hoop irradiation strain, respectively. Meanwhile, since irradiation creep is
the significant mechanism of stress relaxation, irradiation creep models of PyC and buffer layer are considered in
this study as well [12].
146
2.3. Fission gas release model
The fission gas (Xe, Kr and He) release was modeled by two mechanisms, the recoil release and the diffusion
release. These two mechanisms would cause the release of Xe and Kr, but the release of He will be only considered
coming from diffusion release mechanism. The total release fission gas will be the sum of recoil and diffusion
release. The recoil release will be the predominant mechanism at low temperature and its portion of fission gas
can be calculated from the following empirical equation [13]:
𝑓= α (13)
where 𝑓 is the portion of Xe and Kr; 𝑆 is the superficial area of fuel particle in m2; V is the volume of fuel
particle in m3 and 𝛼 is the mean recoil range of fission gas atoms in m. The mean recoil range of Xe and Kr is 3.98
μm and 5.68 μm, respectively.
The main mechanism of fission gas release at high temperature is diffusion release and Booth classical
diffusion model was employed in this study to compute the final release fission gas of loose pyrolytic carbon and
gap through diffusion mechanism. The grain of UN was considered as the ideal sphere of 20 μm in diameter which
means the solving equation can be simplified as one dimensional form. The effective diffusion coefficient of fission
gas atoms within the fuel grains can be set from the following empirical relation [5]:
Dg = 6.66454 × 10-8 exp( − 19164/T) (14)
where Dg is the effective diffusion coefficient of fission gas atoms within the fuel grains in m2 ∙s-1 .
2.4. Fuel particle failure probability model
SiC layer is the main barrier preventing the release of radioactivity and is also the main pressure bearing
structure layer. The pressure vessel failure probability of SiC layer can be computed through equation (13) based
on the Weibull distribution theory [5].
Ψ = 1 − exp − ∫ 𝑑𝑉 (15)
where, Ψ is the failure probability; 𝜎 is the stress in MPa and the principal stress is employed to compute;
m is the irradiated Weibull modulus of SiC layer and 𝑚 = 6; σ0 is Weibull parameter and σ0 = 9.64 MPa∙m1/2 .
Besides, heat transfer across gaps between buffer layer and IPyC layer and internal pressure are also
considered in this study.
3. MODEL VERIFICATION AND APPLICATION
3.1. Performance analysis under steady conditions
The thermal power density of TRISO particles used in LWRs is higher compared with the typical TRISO
particles used in high temperature gas cooled reactor. The particle power was set as 0.5 W; the maximum fast
neutron fluency and burn-up were 18×1025 m-2 and 19% respectively and the simulation time was 1200 Effective
Full Power Days (EFPDs). The structure size of TRISO particle was set as follows: diameter of UN kernel was
800μm, thickness of buffer, IPyC, SiC and OPyC layers were 100 μm, 30 μm, 40 μm and 30 μm respectively. The
outside temperature of fuel particles was set at 1073 K as the boundary condition. The defect on SiC layer was set
in the model to analyze the effect of fabrication defect on the performance of TRISO particle. The fabrication
defect and model boundary condition were showed in Fig. 1.
147
FIG. 1. Schematic diagram of TRISO particle characteristic unit used in this work.
Figure 2 shows the variation of average radius at inner surface of buffer layer, IPyC layer and the gap width
with the increase of fast neutron fluence when computed the irradiation deformation of PyC layer. At the beginning
of operation, the radius of buffer layer would decrease rapidly due to the irradiation deformation while the radius
of IPyC would increase at the same time which caused the gaps between them. The buffer layer shrink because of
the densification of buffer layer, which was similar with the UO2 pellet. The shrinkage of IPyC layer was saturated
when the neutron flux reached at about 13×1025 n/m2, which was good agreement with the literature [14]. After
the increasing of gap size between buffer and IPyC layer, due to the irradiation creep and the expansion caused by
the radial irradiation deformation of IPyC layer, the radius of IPyC layer would start to decrease again until the
irradiation deformation reached saturation. The radius of buffer layer would start to increase as the irradiation
swelling of UN kernel. The maximum width of gaps would reach about 23 μm, the large gap was bad for the heat
transfer efficiency at the internal fuel particles.
FIG. 2. Variation of a) buffer and IPyC layer radius and b) gap size between buffer and IPyC layers.
The radial temperature distribution of fuel particles for different fast neutron fluence is shown in Figure 3.
Before the appearance of gaps, the internal fuel particles existed high temperature gradient which mainly due to
the low thermal conductivity of buffer layer. Since gaps would form between IPyC layer and buffer layer in later
period, fission gas would accumulate in the gaps and buffer layer which would dramatically decrease the thermal
conductivity efficiency of internal fuel particles. And there was up to 200 K temperature difference between buffer
layer and IPyC layer. The maximum temperature of UN kernel was about 1250 K when the neutron flux reached
25×1025 n/m2 which was lower than the melting point of UN ceramics. The gap between buffer and IPyC layer
offered the space which can contain the fission gas, and the amount of fission and internal pressure in the gap was
calculated.
148
FIG.3. Temperature distribution of TRISO particle along radial direction.
Figure 4 shows the variation of fission gas release and internal pressure with the fast neutron fluence. The
release of fission gas and internal pressure were approximately linear increasing as fast neutron fluence increased.
Since there is no CO release in UN kernel, gas production and internal pressure were much smaller compared with
TRISO particle with UO2 kernel. When burn-up reached 19% FIMA, the internal pressure was only 1.78 MPa
which conformed to the literature [14]. Compared with TRISO particles used in high temperature gas cooled
reactor, internal pressure makes less contribution to the pressure vessel failure probability of SiC layer. The stress
state and structure integrity of SiC layer are mainly influenced by the irradiation deformation of PyC layer.
FIG.4. Variation of a) fission gas release and b) internal pressure of TRISO particle with neutron flux.
Figure 5 plots the variation of average hoop stress at inner surface of PyC and SiC layers with fast neutron
fluence. At the beginning of operation, the radial and hoop shrinkage of IPyC and OPyC layers would squeeze SiC
layer which caused the increase of normal stress at hoop direction of SiC layer while tensile stress at hoop direction
would be applied on PyC layer which was a big threat to the structure integrity of PyC layer. With PyC layer
gradually change to rapid expansion at radial direction, the stress state of PyC layer at hoop direction changed to
pressure stress. Instead, pressure stress changed rapidly to tensile stress of SiC layer at hoop direction. Oversize
tensile stress might cause the pressure vessel failure probability of SiC layer and lost the barrier prevent fission
production leaking. With the irradiation deformation reached saturation, hoop stress of each layer would progress
to low stress which meant the irradiation deformation of PyC layer had great influence on the structure integrity
of fuel particle.
149
FIG. 5. Variation of hoop stress at inner surface of PyC and SiC layers with neutron flux, a) non defect surface, b) surface with
defect.
SiC layer is the main pressure bearing structure and safety barrier of inner parts of fuel particle. The failure
of SiC layer would have significant influence on the structure integrity of fuel particle. SiC is brittle material which
can use Weibull distribution model to compute its random failure probability and results is shown in Figure 6.
When Eq. 15 was employed to compute the irradiation deformation of IPyC layer, it would occur failure earlier
with higher probability. In fact, failed IPyC layer has great influence on the structure integrity of SiC layer and it
is essential to conduct more experiments and theoretical studies on the specified SiC layer alone which is also a
part of further investigations.
FIG. 6. Failure probability of the coated layers, a) IPyC layer, b) OPyC layer and c) SiC layer.
3.2. Performance analysis under RIA conditions
In order to analysis the performance of TRISO coated particle fuel under RIA condition, rod ejection accident
needs to be considered under hot shut down condition where reactivity is great caused by deep control rods.
Assume reactor operated at full power for 900 EFPD under thermal work condition and then turned into zero
power maintaining 10 EFPD. After that, rod ejection accident happened with high reactivity which would cause
the particle power of fuel assembly hottest area increasing to 30 times of rated power (15 W) and then decreasing
to zero power rapidly. Assume the time of this duration was 5 s. Finally, the reactor was shut down emergently.
The fuel particle central temperature variation of time was shown in Figure 7. Due to the appearance of gaps,
the central temperature increased rapidly and then stabilized gradually. The rod ejection accident happened at zero
power thermal work condition caused power increased and decrease rapidly at very short time so as the
corresponding central temperature. The highest temperature of UN kernel would reach 2200 K which was not
enough to cause fuel melting. In addition, the thermal shock caused by temperature rapid increasing would cause
great thermal expansion deformation which would cause the rapid contact between buffer layer and IPyC layer.
150
FIG. 7. Variation of kernel temperature with time under RIA condition.
Figure 8 plots the average hoop stress at inner of PyC and SiC layers variation of time which shows PyC and
SiC layers had huge stress shock caused by contract pressure and formed great tensile stress in a moment. PyC and
SiC layers might crack and lead structure failure. Therefore, thermal expansion is the dominant factor causing
structure failure of fuel particle under RIA condition.
FIG. 8. Hoop stress of PyC and SiC layers under RIA condition.
4. CONCLUSIONS
A computation model which is able to analysis the irradiation performance of TRISO coated fuel particle
was developed by using the multi physics coupling software COMSOL in this study and the main results are as
follows:
A three dimensional thermal mechanical coupling model was developed for TRISO coated fuel particle with
UN as the kernel which considered the physical properties and behaviour of materials. Under steady condition,
buffer and IPyC layers would separate and the gaps between them would affect the thermal conductivity efficiency.
The internal pressure caused by fission gas have little influence on the structure integrity of SiC layer while the
complicated irradiation deformation behaviour has a great influence on the stress of PyC and SiC layers. IPyC
layer have biggish structure failure risk while SiC layer can maintain the integrity of structure. Under RIA
condition, drastic thermal expansion would cause each layer contact and form great tensile stress which is the
dominant factor causing structure failure of TRISO coated fuel particle.
151
5. NOMENCLATURES
RIA Reactivity introduced accident

FCM Fully ceramic microencapsulated
ATF Accident tolerant fuel
TRISO Tri-isotropic
UN Uranium mononitride
Buffer porous pyrolytic carbon
IpyC Inner dense pyrolytic carbon
SiC Silicon carbide
OpyC Outer dense pyrolytic carbon
REFERENCES
[1] TERRANI, K.A., SNEAD, L.L., CEHIN, J.C., Fully ceramic microencapsulated fuels for LWRs. Transactions of the
American Nuclear Society 106 (2012) 107.
[2] MANCHANG, L., DAN, W., Neutronic feasibility analysis of fully ceramic microencapsulated fuel for commercial
PWRs, TOPFUEL2016. USA: Boise, 2016.
[3] ROSS, S.R., MATTHEWS, R.B., Thermal conductivity correlation for uranium nitride fuel between 10 and 1923 K.
Journal of Nuclear materials 151 (1988) 317.
[4] HAYES, S.L., THOMAS, J.K., Material property correlation for uranium mono nitride, I: Physique properties. Journal
of Nuclear materials 171 (1990) 270.
[5] MILLER, G.K., PETTI, D.A., PARFUM theory and model basis report, INL/EXT-08-14497, USA: INL, 2009.
[6] SNEAD, L.L., NOZAWA, T., Handbook of SiC properties for fuel performance modeling. Journal of Nuclear materials
371 (2007) 377.
[7] COLLIN, B.P. Modeling and analysis of FCM UN TRISO fuel using the PARFUM code, INL/EXT-13-30193, USA:
INL, 2013.
[8] HAYES, S.L., THOMAS, J.K., Material property correlations for uranium mono nitride, II: Mechanical properties.
Journal of Nuclear materials 171 (1990) 288.
[9] HALES, J.D., Williamson, R.L., Multidimensional multi physics simulation of TRISO particle fuel. Journal of Nuclear
materials 443 (2013) 543.
[10] GENK, M.S., ROSS, S.B., “Uranium nitride fuel swelling and thermal conductivity correlations”, Transactions of the
Fourth Symposium on Space Nuclear Power system. USA: Albuquerque (1987) 313-317.
[11] DEMANGE, P., MARIAN, J., TRISO fuel element thermo mechanical performance modeling for the hybrid LIFE
engine with Pu fuel blanket. Journal of Nuclear materials 405 (2010) 155.
[12] PETTI, D., MATEIN, P., Development of improved models and designs for coated-particle gas reactor fuels, INL//EXT-
05-02615, USA: INL, 2004.
[13] BESMANN, T.M., FERBER, M.K., Fission product release and survivability of UN-kernel LWR TRISO fuel. Journal
of Nuclear materials 448 (2014) 419.
[14] SCHAPPEL, D., TERRANI, K., POWERS, J., Thermo mechanical analysis of fully ceramic microencapsulated fuel
during in-pile operation, Top Fuel 2016, Boise, ID, September 11-15, 2016.
[15] COLLIN, B.P., Modeling and analysis of UN TRISO fuel for LWR application using the PARFUME code. Journal of
Nuclear materials 451 (2014) 77.
152
MAIN RESULTS OF THE CRP ACTOF
(Session 4)
Chairperson
T.LIU
China
153
REACTIVITY INITIATED ACCIDENT ASSESSMENT FOR ATF CLADDING
MATERIALS
C. GIOVEDI, M.R. MARTINS

LabRisco, University of São Paulo,
Av. Prof. Mello Moraes 2231, 05508-030, São Paulo, SP, Brazil,
A. ABE, R. REIS, A.T. SILVA

Nuclear Engineering Center — CEN, Nuclear and Energy Research Institute - IPEN/CNEN,
Av. Prof. Lineu Prestes 2242, 05508-000, São Paulo, SP, Brazil.
Abstract
Following the experience that came from the Fukushima Daiichi accident, one possible way of reducing risk in a nuclear
power plant operation would be the replacement of the existing fuel rod cladding material (based on zirconium alloys) by other
materials which could fulfill the requirements of the accident tolerant fuel (ATF) concept. In this sense, ATF should be able to
keep the current fuel system performance under normal operation conditions; moreover, it should present superior performance
than the existing conventional fuel system (zirconium based alloys and uranium dioxide) under accident conditions. The most
challenging and bounding accident scenarios for nuclear fuel systems in Pressurized Water Reactors (PWR) are Loss of Coolant
Accident (LOCA) and Reactivity Initiated Accident (RIA), which are postulated accidents. This work addresses the
performance of ATF using iron based alloys as cladding material under RIA conditions. The evaluation is carried out using
modified versions of the coupled system FRAPCON/FRAPTRAN. These codes were modified to include the material
properties (thermal, mechanical, and physics) of an iron based alloy, specifically FeCrAl alloy. The analysis is performed using
data available in the open literature related to experiments using conventional PWR fuel system (zirconium based alloys and
uranium dioxide). The results obtained using the modified code versions are compared to those of the actual existing fuel
system based on Zircaloy-4 cladding using the original versions of the fuel performance codes (FRAPCON/FRAPTRAN).
1. INTRODUCTION
Accident Tolerant Fuels (ATF) have been studied after the Fukushima Daiichi accident in order to improve
the nuclear safety under steady state irradiation and accident scenarios. The most challenge scenarios to be studied
are those related to postulated accidents, specifically LOCA and RIA. In this sense, in the framework of the ATF
programme, it is necessary to evaluate the performance under postulated accident scenarios of different cladding
materials which could be applied to replace the zirconium based alloys currently used. The cladding materials that
present higher potentiality to this are iron based alloys such as iron-chromium-aluminum (FeCrAl), and ceramic
materials such as silicon carbide (SiC).
There are two accident scenarios of particular interest considering inadvertent insertion of reactivity in
nuclear power reactors: the control rod ejection accident in Pressurized Water Reactors (PWR), and the control
rod drop accident in Boiling Water Reactors (BWR). These are design basis accidents, which may result in serious
consequences if they were not accounted properly in the design of the reactor and related safety systems. The cause
of accident is related to mechanical failure of a control rod mechanism housing and the primary coolant pressure
ejects a control rod assembly completely out of the core. The main consequence of the control rod ejection is a
rapid positive reactivity insertion, which shall occur within about 0.1 s in the worst possible scenario. The fuel rod
behaviour during a RIA is mostly influenced by characteristics of the power pulse (amplitude and pulse width),
coolant condition (pressure, temperature and flow rate), level of fuel burnup (degree of hydrogen pickup, cladding
corrosion, internal fuel pressure), and fuel rod design (cladding thickness, pellet geometry and initial fill pressure).
The rapid power excursion (0.1 s) leads to nearly adiabatic heating of the fuel pellets, which immediately deform
by solid thermal expansion and depending of burnup degree, the amount of gaseous fission products retained in
the fuel will contribute more to the solid pellet deformation. Due to pellet expansion the pellet cladding gap is
reduced or closed leading to pellet cladding mechanical interaction (PCMI) with significant mechanical loading
on the cladding tube [1].
Fuel performance codes are good tools to be applied to assess the behaviour of different materials under
irradiation; however, the conventional fuel codes need to be modified to introduce the properties of the materials,
which are being studied.
155
The FRAPTRAN code has two models to predict cladding failure, first failure model is related to RIA event,
where deformation is due to pellet cladding mechanical interaction and at relatively low temperature (< 700 K),
where PCMI is the driving force for cladding deformation, the model considers basically uniform plastic
elongation from irradiated cladding as function of temperature and hydrogen concentration. The second failure
model is applicable to LOCA events where deformation is relatively high due to gas overpressure and the
temperature of the cladding (> 700 K) [2].
This paper presents some preliminary results obtained using modified versions of the FRAPCON-
FRAPTRAN codes containing the properties related to the FeCrAl alloy to simulate the fuel behaviour under
steady state and RIA conditions.
2. METHODOLOGY
2.1. FeCrAl alloy
FeCrAl alloy presents better properties compared to zirconium based alloys specifically concerning to the
oxidation rates which are of 1 to 3 orders of magnitude lower [3]. Data from literature [3] based on computational
simulations also indicate that FeCrAl alloy maintains acceptable thermo mechanical properties, and fuel-clad
interactions under PWR conditions. The thermal, mechanical e physical properties of FeCrAl were obtained from
literature [3] and used to modify the fuel performance codes.
2.2. Fuel performance code modification
The basis for the code modification applied to study the steady state irradiation was the FRAPCON-3.4 code
[4], and for RIA evaluation the FRAPTRAN-1.5 [2]. The main subroutines related to the cladding in the codes
modified to introduce the properties in MATPRO [5] of FeCrAl alloy were: CELMOD, CORROS, CREEPR,
CSHEAR, and CTHCON. The material properties concerning to each of these subroutines are: CELMOD defines
the correlation for the cladding Young´s modulus; CORROS is related to the cladding waterside corrosion;
CREEPR; CSHEAR calculates shear modulus of cladding based on type and conditions; and CTHCON defines
the correlation for the cladding thermal conductivity.
2.3. Test case
The test case applied to compare the fuel performance of Zircaloy-4 and FeCrAl alloy was the HBO-5, which
data are available in the FRAPCON-3.5 assessment [6]. This case is part of a large scale RIA experiment carried
out at NSSR reactor by JAERI (Japan Atomic Energy Research Institute). The NSRR facility has a TRIGA type
reactor, which can generate significantly narrow power pulses, the RIA simulation tests on pre-irradiated fuel
segment (test rod) employ a capsule with stagnant water. The HBO-5 fuel rod was refabricated from a standard
1717 PWR fuel rod with uranium dioxide fuel pellet that was irradiated under steady state condition at OHI#1
reactor comprising a period of about 1360 days until an average burnup of 44 GWd/MTU. After this, the full-
length fuel rod was refabricated and submitted to the test under RIA conditions. The HBO-5 fuel rod experienced
failure at 0.2113 s.
The input for both codes were prepared following strictly according to the recommendations presented in the
assessment volume as presented in Table 1.
156
TABLE 1. HBO-5 FUEL ROD MAIN CHARACTERISTICS AND
BOUNDARY CONDITIONS
Characteristic/boundary condition PWR
Fuel material UO2
Fuel pellet outer diameter (mm) 8.05
Fuel pellet height (mm) 9
Fuel density (% TD) 95
Fuel enrichment (w/o %) 3.2
Cladding outer diameter (mm) 9.5
Cladding inner diameter (mm) 8.22
Cladding wall thickness (mm) 0.64
Fuel-cladding diametral gap (µm) 170
Fill rod pressure (MPa) 3.23
Coolant temperature (°C) 310
Coolant pressure (MPa) 15.5
The average power history and the rod average burnup as function of the irradiation time are presented in
Figs 1 and 2, respectively.
FIG.1. Average power history for HBO-5 test case.
FIG.2. Rod average burnup as function of irradiation time for HBO-5 test case.
157
3. RESULTS AND DISCUSSION
The main results obtained comparing the performance of Zircaloy-4 and FeCrAl alloy as cladding under
steady state irradiation for the HBO-5 fuel rod are presented in Figures 3 to 7.
Figure 3 shows that FeCrAl fuel rod presents fuel centreline temperatures higher than Zircaloy-4 fuel rod.
This occurs due to the higher thermal expansion of the FeCrAl. Then, the gap thickness is larger, as shown in
Figure 4, and, consequently, the fuel temperature reaches higher values.
FIG.3. Fuel centerline temperature evolution under steady state irradiation as function of time for HBO-5 test case considering
as cladding: Zircaloy-4, and FeCrAl alloy.
FIG.4. Gap thickness evolution under steady state irradiation as function of time for HBO-5 test case considering as cladding:
Zircaloy-4, and FeCrAl alloy.
Also, it can be observed in Figure 5 that FeCrAl fuel rod presents a lower plenum pressure compared to the
Zircaloy-4 fuel rod due to the higher internal free volume available that is a result of the higher thermal expansion
of the FeCrAl alloy.
158
FIG.5. Plenum pressure evolution under steady state irradiation as function of time for HBO-5 test case considering as
cladding: Zircaloy-4, and FeCrAl alloy.
Due to the gap closure around 600 irradiation days observed for the Zircaloy-4 fuel rod, the cladding hoop
stress, as shown in Fig. 6, changes from a compressive to a tensile state. This is not observed for the FeCrAl fuel
rod because the gap remains open during all the irradiation time, also, as a consequence of the higher thermal
expansion compared to Zircaloy-4.
FIG.6. Cladding hoop stress evolution under steady state irradiation as function of time for HBO-5 test case considering as
Regarding to the fission gas release, Figure 5 shows that the behaviour is exactly the same for both studied
materials. This shows that the differences observed in the fuel centerline temperatures are not enough to affect the
fission gas release behaviour for both studied materials.
159
FIG.7. Fission gas release evolution under steady state irradiation as function of time for HBO-5 test case considering as
After steady state irradiation, the RIA condition was simulated using FRAPTRAN code in order to reproduce
the experimental test for the HBO-5 fuel rod. Figure 8 presents the cladding hoop stress after RIA, considering
that the RIA starts at time equal to 0. It can be observed that the Zircaloy-4 fuel rod experiences failure 0.21 s after
starting the RIA, result that is very similar to that observed in the experiment. This confirms that the simulation
was appropriately carried out. The same simulation was performed under RIA conditions using the transient code
modified considering FeCrAl as cladding material. And, in these preliminary results the FeCrAl fuel rod
experiences failure at a time very close to that observed for Zircaloy-4 due to the fact that the failure criterion
adopted in the FRAPTRAN code for Zircaloy-4 cannot be appropriate to evaluate the FeCrAl fuel rod behaviour
under RIA conditions.
FIG.8. Cladding hoop stress after RIA (t=0) for HBO-5 test case.
4. CONCLUSIONS
The obtained results for the HBO-5 test case indicate that under steady state irradiation FeCrAl present a
good performance compared to Zircaloy-4 showing: higher fuel temperatures (about 100°C), lower internal
pressure, open gap during all the irradiation time, and lower cladding hoop stress.
The simulations carried out to under the RIA scenario reproduced exactly the experimental result registered
for the studied test case, occurring the failure of the fuel rod at the same time verified in the experiment.
160
Although the results of simulation obtained in the steady state condition have shown differences, mainly in
the fuel gap parameter, the fuel rod test case failure at similar time in the transient simulation, consequently it is
necessary to evaluate and establish a better and consistent failure criterion for FeCrAl cladding.
The strain-based failure criterion used in the original code version for Zircaloy-4 considers the effects of
hydride content and temperature [7]. In this sense, it is necessary to carry out experiments in order to obtain data
to better predict the cladding failure for FeCrAl behaviour under RIA conditions.
5. NOMENCLATURES
ATF Accident Tolerant Fuel

BWR Boiling Water Reactor
FeCrAl Iron-Chromium-Aluminum alloy
HBO-5 Experimental test case
LOCA Loss-of-Coolant Accident
PCMI Pellet Cladding Mechanical Interaction
PWR Pressurized Water Reactor
RIA Reactivity Initiated Accident
SiC Silicon Carbide
ACKNOWLEDGEMENTS
The authors are grateful to the technical support of USP and IPEN-CNEN/SP, and to the financial support
of IAEA to attend the Technical Meeting.
REFERENCES
[1] RASHID, Y.R., ZANGARI, A.J., LIN, C.L., Modeling of PCI Under Steady State and Transient Operating Conditions,
IAEA Technical Committee Meeting on Water Reactor Fuel Element Computer Modeling in Steady State, Transient and
Accident Conditions, IAEA-659/IWGFPT/32 Preston England, (1988) 91.
[2] GEELHOOD, K.J., LUSCHER, W.G., CUTA, J.M., FRAPTRAN- 1.5: A Computer Code for the Transient Analysis of
Oxide Fuel Rods, PNNL-19400, 1, Rev.1 (2014).
[3] GAMBLE, K.A., BARANI, T., PIZZOCRI, D., HALES, J.D., TERRANI, K.A., PASTORE, G., An Investigation of
FeCrAl Cladding Behavior Under Normal Operating and Loss of Coolant Conditions, Journal of Nuclear Materials 491
(2017) 55.
[4] GEELHOOD, K.J., LUSCHER, W.G., BEYER, C.E., FLANAGAN, M.E., FRAPCON-3.4: A Computer Code for the
Calculation of Steady state Thermal mechanical Behavior of Oxide Fuel Rods for High Burnup, U.S.NRC, NUREG/CR-
7022, Washington (2011).
[5] ALLISON, C.M., BERNA, G.A., CHAMBERS, R., CORYELL, E.W., DAVIS, K.L., HAGRMAN, D.L., et al.,
SCDAP/RELAP5/MOD3.1 Code Manual Volume IV: MATPRO―A Library of Materials Properties for Light-Water-
Reactor Accident Analysis, NUREG/CR-6150, EGG-2720, Washington (1993).
[6] GEELHOOD, K.J., LUSCHER, W.G., FRAPCON-3.5: Integral Assessment, PNNL-19418, 2, Rev.1 (2014).
[7] NEA/CSNI/R (2010) 1, Nuclear Fuel Behaviour under Reactive Initiated Accident (RIA) Conditions, OCDE (2010) 179.
161
ROUND ROBIN EXERCISE OF THE CANDIDATE ATF CLADDING MATERIALS
WITHIN THE ACTOF PROJECT
M. ŠEVEČEK, J. KREJČÍ, A. CHALUPOVÁ, J. KABÁTOVÁ, F. MANOCH, J. KOČÍ, L. CVRČEK

Czech Technical University in Prague/UJP Praha,
Prague, Czech Republic,
C. TANG, M. STEINBRÜCK, M. GROSSE

Institute for Applied Materials, Karlsruhe Institute of Technology,
Karlsruhe, Germany,
S. PENTTILÄ, J.-M. AUTIO, J. LYDMAN, A. TOIVONEN

VTT Materials for Power Engineering, Technical Research Centre of Finland,
Espoo, Finland,
B. SARTOWSKA, W. STAROSTA, L. WALIŚ, A. KRZEŚNIAK, E. PAŃCZYK, P.KALBARCZYK, K.

KULISA, J. SMOLIK, J. MIZERACKI
Institute of Nuclear Chemistry and Technology,
Warsaw, Poland,
Z. HÓZER, T. NOVOTNY, E. PEREZNÉ FERÓ, A. PINTÉR CSORDÁS, M. HORVÁTH, P. SZABÓ

Hungarian Academy of Sciences, KFKI Atomic Energy Research Institute,
Budapest, Hungary,
C. GIOVEDI
University of São Paulo,
São Paulo, Brazil,
P. XU
Westinghouse Electric Company,
Hopkins, SC, USA.
Abstract
The paper presents a summary of the round robin test activity organized within the IAEA ACTOF project. The test
conditions and sample matrix were finalized during fall 2017, production and transportation of samples started in Q1 2018 and
the tests have been performed during the past 14 months. Two fundamental experimental tests related to normal operating
conditions and accidental conditions of LWRs were defined: High temperature steam oxidation and Long term corrosion tests.
Originally, four laboratories/institutes joined the round robin exercise — CTU in Prague, KIT, VTT, and INCT. Later, also
MTA EK joined the activity. Westinghouse Electric Company provided Zircaloy-2 substrate. Two institutes applied their
protective coatings on the Zircaloy-2 substrate by PVD methods (CTU — pure Cr coating and INCT — ZrSi-Cr coating). KIT
applied MAX phase coating on Zircaloy-4 substrate. Additionally, AISI 348 steel provided by USP, Brazil was tested. The
geometry of the samples varied based on the needs of particular institutes. Long term corrosion tests were performed by three
laboratories in PWR and WWER chemistry for at least 63 cumulative days. High temperature steam oxidation tests were
performed by four institutes at three pre-defined conditions. Participants performed pre- and post-characterization of the
materials based on their standard procedures and available techniques. The paper summarizes motivation, plans, sample matrix,
materials, testing conditions and preliminary results including lessons learnt and recommendations for future joint projects.
1. INTRODUCTION
One of the key joint actions planned within the IAEA Coordinated Research Project on “Analysis of Options
and Experimental Examination of Fuels for Water cooled Reactors with Increased Accident Tolerance (ACTOF)”
was the round robin test (RRT) for cladding materials that are considered as Accident Tolerant Fuel (ATF) cladding
candidates. The tested cladding materials include three coated cladding concepts and AISI stainless steel. There
were two fundamental tests defined — high temperature steam oxidation test and long term corrosion test in
prototypical LWR conditions. The participants of the RRT are: Czech Technical University/UJP Praha (CTU),
162
Karlsruhe Institute of Technology (KIT), Institute of Nuclear Chemistry and Technology (INCT), VTT Technical
Research Centre of Finland (VTT), University of São Paulo (USP) and Hungarian Academy of Sciences Centre
for Energy Research (MTA EK). The RRT can be subdivided into three sub-activities:
— ATF cladding sample production;
— Long term corrosion test in WWER or PWR chemistry;
— High temperature oxidation.
The involvement of the parties in particular tasks of the RRT activity is summarized in Table 1.
TABLE 1. INVOLVEMENT OF PARTICIPANTS IN THE RRT

CTU KIT INCT VTT USP MTA EK
Sample production X X X — X —
Long term corrosion X — X X — —
High temperature oxidation X X — X — X
As can be seen in Table 1, there were four different types of ATF cladding samples produced by four
institutes. Three institutes performed long term corrosion tests in PWR or WWER chemistry and four institutes
performed high temperature oxidation testing.
2. SAMPLE PRODUCTION
Traditional Zr-based cladding materials were used as substrates (Zry-2, Zry-4 and E110) and three different
coatings were deposited on their surfaces using physical vapour deposition methods (PVD). Originally, standard
Zry-2 LK3 alloy was chosen as the substrate for all tests with two predefined geometries shown in Fig. 1. Due to
time and other limitations, it was decided to use also Zry-4 substrate provided by KIT and E110 substrate provided
by UJP Praha and change the sample geometry based on the requirements of particular participants (due to size
limitations in furnaces, to allow for mechanical testing or single-sided oxidation etc.).
FIG. 1. The geometry of samples for RRT as originally defined.
The coatings deposited by PVD methods are: pure Cr (CTU), ZrSi-Cr (INCT) and Cr/Cr2AlC/Cr (KIT).
Additionally, AISI 348 stainless steel provided by USP was tested. Their fabrication and specifications are
described further.
2.1. Cr PVD coated Zr-based alloys
Zr-based samples were Cr coated in two batches using unbalanced magnetron sputtering (UBM) in the
Hauzer Flexicoat 850 industrial system. It is a multi purpose device flexible for numerous applications. Depending
on the type, it can utilize a combination of different deposition technologies such as sputtering, HIPIMS, PACVD
and other. For the purposes of this work, UBM was used. It is a special type of sputtering with extended and
increased plasma density using coils in an unbalanced closed magnetic field. The images of the system, deposition
163
chamber and rotating sample holders with Cr target are shown in Figure 2. More details about the methods used
and coating parameters can be found in [1–5].
FIG. 2. Images of the Hauser Flexicoat 850 at CTU in Prague. Overview of the system (left); Deposition chamber (centre);
Rotating sample holders with Cr target in the back (right) (Reproduced courtesy of Elsevier [16]).
The coatings were deposited in a metal mode using the gas flow 90 sccm Ar (99.999 %) for about 18 hours.
The thickness of the coatings was measured with a Calotest (CSM, Switzerland) and was found to be 28 ± 0.1 µm
in the first batch and 16 ± 0.1 µm in the second batch.
2.2. ZrSi-Cr PVD coated Zry-2
ZrSi-Cr coated samples were produced at the Institute for Sustainable Technologies ITS (Radom, Poland).
Balzers system shown in Fig. 3 was used. The system contains three magnetron plasma sources, three power and
control panels. It allows to deposit materials from separate targets and to obtain multi elemental coatings [6].
Two magnetrons with separate, flat, circular targets were used. One was made of ZrSi2 and the second of
pure Cr. The resulting composition of the coatings as confirmed by XRD is — Zr40Si24Cr36 and its thickness is
about 2.5 microns.
a) b) c)
FIG. 3. Balzers system facility: a) general view, b) position of flat, circular magnetrons, c) chamber overview.
2.3. Cr/Cr2AlC/Cr (MAX phase) PVD coated Zry-4
A research group at KIT investigates the deposition of Ti2AlC, Cr2AlC and Zr2AlC MAX phases on Zircaloy-
4 by a PVD. Pre-tests have shown that the Cr2AlC MAX phases are the most promising with respect to their high
temperature oxidation behaviour. The coatings were synthesized via a two-step process, i.e. first deposited by
magnetron sputtering and subsequently thermally annealed in pure argon. The as-deposited coatings were
deposited using laboratory PVD equipment (Leybold Z 550 coater). Three high purity cylindrical elemental plates
of chromium (Cr: 99.95%, FHR Anlagenbau GmbH), graphite (C: 99.95%, Schunk Kohlenstofftechnik GmbH)
and aluminium (Al: 99.95%, FHR Anlagenbau GmbH) are used as targets for magnetron sputtering process. Figure
4 shows the schematic representation of the design of the as-deposited coatings on Zircaloy-4 substrates and the
arrangement of the targets and substrates during deposition.
After deposition, the coated Zircaloy-4 specimens were ex-situ annealed in pure argon (99.9999%) at
atmospheric pressure using a commercial thermal balance (NETZSCH STA-449 F3 Jupiter) to facilitate the growth
of MAX phases by solid reaction of the nanoscale elemental multilayers. The isothermal holding time was 10 min
164
at annealing temperature 550°C. Cr2AlC MAX phase was successfully obtained while no significant interdiffusion
between the Cr and Cr2AlC layer has been confirmed at such conditions [7, 8].
FIG. 4. Schematic representation of (a) the design of the as-deposited coatings on Zircaloy-4 substrates, and (b) arrangement
of the targets and substrates during deposition.
2.4. AISI 348 SS
USP produced austenitic stainless steel AISI 348 samples within the RRT joint activity. There were two
different geometries of austenitic stainless steel (AISI 348) samples supplied: tubes and plates. The composition
and properties of both materials were the same:
— Chemical composition: Fe-balance, C-0.055%, Mn-1.70%, P-0.017%, S-0.003%, Si-0.41%, Cr-17.5%, Ni-
11%, Nb-0.85%, N-0.0018%, Co-0.021%, Ta<0.005%, B-0.0008%;
— Micro-hardness: 150–200 HV;
— Surface roughness: ≤ 0.7 µm Ra;
— Mechanical properties at 20°C: tensile strength — 640 MPa, yield strength — 330 MPa, elongation of a 50
mm specimen — 47%;
— Mechanical properties at 370°C: tensile strength — 455 MPa, yield strength — 260 MPa, elongation of a 50
mm specimen — 26%;
— Non-metallic inclusions (ASTM E45, Table 1 [9]): ≤1;
— Niobium carbides evaluation (ASTM E407 [10]): continuous non-stabilized areas < than 0.05 mm of
equivalent diameter;
— Intergranular corrosion (ASTM A262, Practice A [11]): material does not present intergranular cracks
(magnification in the range 5x to 20x).
Plate samples were prepared from AISI 348 bars with a diameter of 22 mm. After machining, each plate
sample was polished, identified and cleaned. Tubular samples were prepared by cutting long tubes of AISI 348
tubes with a diameter of 9.8 mm and a thickness of 0.6 mm.
3. EXPERIMENTAL
3.1. Long term corrosion
The long term corrosion tests were done in two prototypical LWR chemistries — PWR (VTT, INCT) and
WWER (CTU) at 360°C. The period of 21 days was defined with a goal of three minimal periods (63 cumulative
days). Destructive testing was performed after 63 days but if there was enough material available, the tests
continued for a longer time. The long term corrosion tests performed by the participants are summarized in Table 1.
165
TABLE 1. LONG TERM CORROSION TESTS ORGANIZED WITHIN THE RRT PERFORMED BY
INDIVIDUAL PARTICIPANTS
CTU — Czech Republic VTT — Finland INCT — Poland
Long term corrosion (WWER Long term corrosion Long term corrosion (PWR
Cladding material
chemistry) (PWR chemistry) chemistry)
AISI348 164 days 63 days 63 days
Zry-2/Zry-2 164 days 63 days 63 days
Cr coated Zry-2 164 days 63 days 63 days
MAX phase coated Zry-4 147 days 63 days 63 days
ZrSi-Cr coated Zry-2 101 days 16.5 days 63 days
Cr coated E110 164 days — —
The experimental procedures vary depending on the available setups and standard procedures used at the
institutes. For example, CTU/UJP uses 360°C, 19.4 MPa, 4 dm3 static autoclave, VTT used standard operational
PWR parameters at 360°C with recirculation loop and online chemistry control whereas INCT used 360°C,
19.5 MPa, 1 dm3 static autoclave.
The samples were characterized before, during and after testing by non-destructive as well as destructive
methods. The characterization techniques include visual inspections, XRD, SEM, metallography, weight changes
and hydrogen and oxygen pickup measurements.
3.2. High temperature steam oxidation
Four participants performed high temperature (HT) oxidation tests. The three predefined conditions are:
— Flowing steam 1100°C, 60 minutes;
— Flowing steam 1200°C, 30 minutes;
— Flowing steam 1300°C, 5 minutes.
These parameters were defined based on estimates taking into account the standard DBA fuel safety criteria
for Zr-based cladding materials such as 17% ECR limit calculated using the Baker-Just correlation [12].
Additionally, previous experience with testing of ATF cladding materials was used to estimate potential additional
margins provided by the near-term ATF concepts tested and to determine reasonably long testing periods for all
samples [13]. Some of the participants decided to perform more tests if larger quantities of samples were available.
The summary of tested samples by particular institutes is shown in Table 2.
TABLE 2. STEAM OXIDATION TESTS PERFORMED BY INDIVIDUAL PARTICIPANTS

CTU — Czech VTT — Finland KIT — Germany MTA EK — Hungary
Republic
HT steam oxidation -
HT steam oxidation - HT steam oxidation - HT steam oxidation -
Cladding 1100,60'; 1100,180';
1100,60'; 1200,30'; 1100,60'; 1200,30'; 1100,60'; 1200,30';
material 1200,30';1200,45';
1300°C,5' 1300°C,5' 1300°C,5'
1200°C,60'
AISI348 X X X —
Zry-4 ref. — X X —
Cr coated Zry-2 X X X —
MAX phase —
X X X
coated Zry-4
ZrSi-Cr coated — — X —
Zry-2
E110 ref. X — — X
Cr coated E110 X — — X
166
The experimental procedures vary depending on the experimental setup used. For example, the heating rates,
steam flow rates, quench parameters, temperature measurements or sample holders are setup-dependent which
might slightly affect the results. In addition, different experimental setups are able to measure different parameters
such as hydrogen production or online weight changes. For example, only KIT was able to measure hydrogen
generation and on the other hand, only CTU and MTA EK performed post-quench mechanical testing.
The samples were characterized before and after testing. The characterization techniques include visual
inspections, XRD, SEM, microhardness, metallography, ring compression tests, hydrogen production, weight
changes or hydrogen absorption measurements.
4. RESULTS
4.1. Long term corrosion
The results of the long term corrosion test show the expected corrosion behaviour of Zr-based alloy. This
confirms that applied setups and procedures are valid and acceptable. Additionally, the data measured on Zr-based
alloys serve as a baseline for comparison with ATF candidate materials.
The results summarized in Figure 5 indicate drastic weight loss in the case of the MAX phase coated sample
suggesting dissolutions and spallation of the coating for both WWER and PWR chemistry. The weight
loss/spallation is slightly slower in WWER chemistry but still extreme. Since spallation was observed for this
sample, the weight change is not fully representative parameter in this case. It should be noted residuals of coating
system remained on the Zr substrate even after the drastic weight loss. This suggests that some parts of the coating
system are stable in LWR conditions.
ZrSi-Cr coating shows comparable corrosion kinetics to uncoated samples. It should be noted that with the
more complex systems, both weight loss and gain happen during the test. This was the case for ZrSi-Cr as well as
MAX phase coating. In the case of ZrSi-Cr coating, a high concentration of Si-based ions was found in the
autoclave after the tests. This confirms that weight gain evaluation standardly used for Zr-based alloys might not
be a suitable parameter for advanced materials. Additionally, the ZrSi-Cr PVD coated samples show very high
hydrogen pickup.
Both, AISI 348 and Cr coated samples show extremely low weight gains. However, hydrogen concentration
inside the Cr-coated is comparable to uncoated material suggesting higher H-pickup fraction. Figure 5 shows the
weight gains of the samples tested in WWER chemistry up to 164 days as a function of exposure time. The detail
of the weight gains of the Cr-coated samples is shown in Fig. 6 as well. Data measured on three samples as well
as their average are shown. The scattering is due to extremely low weight gains where slight uncertainty in
measurement results in higher scatter. But the absolute values are generally very low.
167
FIG. 5. Results of the weight gains from the corrosion test in WWER water chemistry. Top — summary of all WGs measured;
Bottom left — summary of weight gains measured in WWER chemistry; Bottom right — detail of the weight gain for Cr-coated
samples.
High resolution SEM micrographs of Cr-coated samples during the long term corrosion test are presented in
Figure 6. Micrographs show the top view of the columnar structure of the Cr coating during the corrosion test. The
grains are tightly coupled to each other with an average diameter of about 1 micron. The grains were clear and
smooth on the as-received material. Small oxide particles cover the pyramidal grains as the surface oxidizes.
INITIAL 21 days 42 days 63 days

FIG. 6. SEM micrographs — top view of the Cr-coated Zry-2 during the corrosion test in PWR chemistry (as-received, 21 days,
42 days, 63 days).
To summarize, AISI 348 and Cr-coated ATF cladding materials seem to be very promising from the
perspective of normal operation. They might provide also other benefits during normal operation such as improved
wear resistance or improved creep behaviour. Nevertheless, unresolved issues such as hydrogen pickup still remain.
168
4.2. High temperature oxidation
The weight gains of the oxidized samples as a function of calculated weight gains using Cathcart-Pawel
correlation [14] as measured at UJP Praha is shown in Figure 7. The results show a reduction of the weight gains
for all of the tested samples in comparison with uncoated Zr-based cladding materials.
FIG. 7. Experimentally measured weight gains in relation with calculated weight gain using Cathcart-Pawel correlation (valid
for most Zr-based alloys).
Additionally, the weight gains are directly linked with the oxidation kinetics for Zr-based alloys and the
weight gain reduction suggests lower hydrogen production and limitations of the exothermic HT reaction with
steam as well. These are the main required properties of ATF cladding materials.
However, the weight gain evaluation was established for Zr-based alloys and does not fully cover the
cladding performance during accidental conditions. The oxidation testing needs to be accompanied by mechanical
testing to quantify the accidental tolerance. Similarly, to long term corrosion tests, some of the samples showed
extensive oxide and material spallation (e.g. AISI 348). For that reason, the weight change is a summation of
oxidation weight gains and spallation (weight loss). The weight gain is from this perspective not a representative
parameter for oxidation kinetics evaluation.
When evaluating the oxidation kinetics based on the hydrogen production measured during the tests, the
results differ. The results of integral hydrogen production during the three predefined tests are shown in Figure 8.
It can be seen that for example, the Cr-coated samples produce very low hydrogen which was suggested also by
the low values of weight gain shown in Figure 7. On the other hand, the AISI 348 samples produce more hydrogen
that uncoated Zr-based alloys but its weight gain is low.
FIG. 8. Integral hydrogen production for 1100°C, 60min; 1200°C, 30min and 1300°C, 5min steam oxidation tests.
169
4.2.1. AISI 348 SPP 3, 1300°C, 5 min
The characterization of the AISI 348 sample after the steam oxidation test at 1300°C for 5 minutes shown in
Figs 9 and 10 was done by VTT. The SE and BSE images shown in Figure 9 present the outer oxide layer on the
top and the inner oxide layer. The bulk alloy can be seen on the bottom. The outer oxide layer is porous and partly
loosened as a result of sample preparation. The inner oxide layer has large cracks. The thickness of the whole
oxide layer varied from 520 μm to 1050 μm.
SE BSE
FIG. 9. SE and BSE images of the AISI 348 sample after exposure at 1300°C / 5 min.
SEM-EDS maps of O, Si, Cr, Mn, Fe and Ni are shown in Figure 10 where the detail of the outer and inner
oxide layer is on the left side. The outer layer contains mainly Fe with some Mn and the inner layer Fe, Mn, Cr
and Ni. The Fe content in the outer oxide layer has slightly increased when the test temperature has been increased.
The results of hydrogen production shown in Figure 8 together with the post-characterization prove that Fe-oxides
formed on this sample are not protective, spall off and do not enhance the oxidation resistance. This is in contrast
with other advanced steels such as FeCrAl that form protective chromia and alumina scales at high temperature
and considerably enhance the oxidation resistance [15]. Therefore, the HT oxidation resistance of the AISI 348 is
limited and does not enhance the accidental tolerance in severe accidental conditions.
O Si Cr
Mn Fe Ni
FIG. 10. SEM-EDS maps of AISI 348 SS after exposure at 1300°C / 5min showing the distribution of O, Si, Cr, Mn, Fe and Ni.
The outer oxide layer is on the left side, the inner oxide layer in the middle and the bulk is on the right side outside the images.
170
4.2.2. Cr-coated Zr1%Nb, 1200°C, 60 min
Some of the participants performed also more extensive testing of ATF candidate samples than originally
required. It includes for example also testing of samples in air atmosphere simulating severe accidental conditions.
The Cr PVD coated samples oxidized in air show higher weight gains (approx. 2.2 times) than Cr-coated samples
oxidised in steam at the same temperature and for the same time. The surface of the samples became dark (Figure
11), but not as dark as after oxidation in steam. Zirconium uncoated samples under similar conditions show about
10 x higher weight gains. This behaviour shows that the Cr coated samples are protective also in severe accidents
with air ingress that is another proven benefit of Cr coated materials.
FIG. 11. Cr-coated IAEA-01 specimen in the quartz boat before (left) and after oxidation (right) in air at 1200°C for 3600 s.
The close view of the Cr coated tube surface showed (Figs 12 and 13) that after oxidation in air a non-uniform
structure was developed. The kinetics of the oxidation in air needs to be further studied but the results obtained
within this activity suggest the protective nature of Cr coatings in air as well as in steam.
FIG. 12. Surface of the Cr-coated sample in as-received conditions.
FIG. 13. Surface of the Cr-coated sample IAEA-01 after oxidation in air (1200°C, 3600 s).
171
5. CONCLUSION
An experimental programme has been carried out with four types of ATF candidate cladding samples. Four
institutes produced their ATF candidate samples — Cr PVD coated Zr alloys, MAX phase PVD coated Zr alloys,
ZrSi-Cr PVD coated Zr alloys and AISI 348 SS. Two fundamental tests were defined — high temperature
oxidation and long term corrosion. Three participants performed long term corrosion testing for minimal 63 days
period in PWR or WWER chemistry. Four participants performed oxidation tests in steam and air atmosphere
between 1100°C and 1300°C including characterization and mechanical testing.
When evaluating the performance of tested samples, it can be seen that ZrSi-Cr coated samples show higher
weight gains and extremely high hydrogen pickup during long term corrosion test compared to uncoated samples.
Spallation and dissolution of MAX phase coated samples was observed in PWR as well as WWER chemistry
during long term corrosion tests. These two concepts are therefore not feasible for further development from the
perspective of normal operation.
AISI 348 SS, as well as Cr, coated samples show low weight gains and acceptable corrosion performance in
both WWER and PWR chemistries. The hydrogen pickup of Cr-coated samples should be, however, further
studied. In the accidental conditions, the AISI 348 performs much worse than Cr coated samples and also worse
than uncoated reference samples at extreme severe accidental conditions. AISI 348 does not form protective oxides
at HT and shows rapid oxidation with excessive hydrogen production. It can be concluded that only Cr-coated
cladding satisfies the requirements of ATF cladding materials when evaluating the four studied ATF concepts.
Additionally, it was found that standard methods (e.g. weight gain evaluation) might not be fully
representative for the evaluation and testing of new materials that differ from Zr-based alloys. The testing and
industry standards related to handling, cleaning as well as testing are not fully developed for these new materials
and new unexpected issues might appear in future.
It should be noted that not all activities originally planned were finished due to the limited time of the ACTOF
project. The activity was initiated during the second half of the project and there were several delays caused by
transport and regulatory issues, unavailability of required materials or funding. Despite the problems encountered
during the RRT activity, there was a high number of valuable data produced by all participants. The RRT activity
can be considered as very successful from the perspective of the involved parties as well as the community. The
full report presented in the ACTOF TECDOC will include all the details about the production, testing methods
and results that were produced within this RRT activity.
REFERENCES
[1] KREJCI, J., SEVECEK, M., CVRCEK, L., Development of Chromium and Chromium Nitride Coated Cladding for
VVER Reactors, 2017 WRFPM.
[2] KREJČÍ, J., et al., Experimental behavior of chromium-based coatings, Top Fuel 2018 - React. Fuel Perform., vol. 2018,
9.–4.10 2018, p. 13.
[3] KRAUSOVÁ, A., TŮMA, L., NOVÁK, M., CVRČEK, L., KREJČÍ, J., MACÁK, J., Chromium Coating as a Surface
Protection of Zirconium Alloys, Koroze Ochr. Mater. 61 5 (Dec. 2017) 169–172.
[4] HORNÍK, J., KRUM, S., TONDL, D., PUCHNIN, M., SACHR, P., CVRČEK, L., Multilayer coatings Ti/TiN, Cr/CrN
and W/WN deposited by magnetron sputtering for improvement of adhesion to base materials, Acta Polytech. 55 6 (Dec.
2015) 388–392.
[5] KREJČÍ, J., ŠEVEČEK, M., CVRCEK, L., KABÁTOVÁ, J., MANOCH, F., "Chromium and Chromium Nitride Coated
Cladding for Nuclear Reactor Fuel", in Proceedings of the 20th International Corrosion Congress, EUROCORR 2017,
Prague, Czech Republic (2017).
[6] STAROSTA, W., SEMINA, V.K., SMOLIK, J., WALIŚ, L., RYDZEWSKI, M., SARTOWSKA, B., Studies on
magnetron-sputtered zirconium-silicide coatings deposited on zirconium alloy for the enhancement of their high
temperature oxidation resistance, Nukleonika 63 3 (Sep. 2018) 73–79.
[7] TANG, C., et al., Synthesis and characterization of Ti2AlC coatings by magnetron sputtering from three elemental
targets and ex-situ annealing, Surf. Coat. Technol. 309 (Jan. 2017) 445–455.
[8] TANG, C., et al., Deposition, characterization and high temperature steam oxidation behavior of single-phase Ti2AlC-
coated Zircaloy-4, Corros. Sci., 135 (May 2018) 87–98.
[9] ASTM E45-18A, Standard Test Methods for Determining the Inclusion Content of Steel, 10.1520/E0045-18A, ASTM
International, West Conshohocken, PA (2018).
[10] ASTM E407-07(2015) E1, Standard Practice for Microetching Metals and Alloys, ASTM International, West
Conshohocken, PA (2015).
172
[11] ASTM A262-15, Standard Practices for Detecting Susceptibility to Intergranular Attack in Austenitic Stainless Steels,
ASTM International, West Conshohocken, PA (2015).
[12] BAKER, L., JUST, L., Technical Report ANL-6548, Argonne Natl. Lab., 1962.
[13] ŠEVEČEK, M., et al., Development of Cr cold spray–coated fuel cladding with enhanced accident tolerance, Nucl.
Eng. Technol. 50 2 (Mar. 2018) 229–236.
[14] PAWEL, R.E., CATHCART, J.V., McKEE, R.A., The Kinetics of Oxidation of Zircaloy‐4 in Steam at High
Temperatures, J. Electrochem. Soc. 126 7 (Jan. 1979) 1105–1111.
[15] REBAK, R.B., TERRANI, K.A., FAWCETT, R.M., FeCrAl Alloys for Accident Tolerant Fuel Cladding in Light Water
Reactors, p. V06BT06A009, (Jul. 2016).
[16] KREJČÍ, J., et al. Development and testing of multicomponent fuel cladding with enhanced accidental performance.
Nuclear Engineering and Technology (2019).
173
INTRODUCTION OF ATF DEVELOPMENT AND MAIN RESULTS OF THE CRP
ON ACTOF OF KAERI
C.H. SHIN, H.C. KIM, Y.S. YANG

Nuclear Fuel Safety Research Division,
111, Daedeok-daero 989 beon-gil, Yuseong-gu, Daejeon, 34057,
Republic of Korea
Abstract
As a candidate of the accident tolerant fuel (ATF) in the KAERI, the microcell UO2 pellet and the surface modified
cladding with coating are being developed as the near term technology. The microcell UO2 pellet is to enhance the fission
product retention and to increase thermal conductivity. The surface modified cladding is based on the coating technology on
the conventional Zirconium alloy cladding. It is to enhance the oxidation and deformation resistance of the fuel cladding. In
the IAEA-CRP ACTOF, the research objectives and anticipated outcomes of KAERI were the development and implementation
in the FRAPCON/FRAPTRAN code of ATF models for the coated cladding and metallic microcell UO2 pellet of KAERI.In
order to evaluate the mechanical behaviours of the ATF cladding under the normal operation conditions, a new analytical
module, FRACAS-CT, was developed based on the thick-wall theory to consider the multi-layered structure of the coated
cladding. The FRACAS-CT model was verified by comparison with an equivalent finite element model. And the module was
implemented into FRAPCON code with consideration of creep and stress relaxation behaviours of the multi-layered cladding.
The implemented FRACAS-CT can simulate the mechanical response and fuel performance of the multi-layered ATF cladding.
The preliminary analysis of the fuel performance for the KAERI’s ATF concept was summarized. The major material properties
of ATF such as the thermal conductivity and thermal expansion of the pellet and the corrosion behaviour of the CrAl-coated
cladding were modified based on out-of-pile test results. The differences compared to a conventional UO2-Zircaloy fuel was
assessed. From the FRAPCON results under the normal operation condition, ATF shows a significant advantage in the
reduction of the fuel centreline temperature, cladding oxidation thickness, fission gas release, and so on, because of the
increased thermal conductivity of the metallic microcell pellet and the oxidation resistance of the CrAl-coated cladding.
1. INTRODUCTION
After Fukushima Daiichi Nuclear Plant accident, the R&D has been conducted to improve the safety of the
nuclear fuel under the accident conditions. The accident tolerant fuel (ATF) is used as a representative expression
of the safety-enhanced nuclear fuel. The DOE-Programme suggested that the goal of accident tolerant fuel should
have a similar or better performance under normal operation conditions and during design basis accidents, as well
as in beyond design basis accidents compared to a conventional nuclear fuel [1, 2]
In Republic of Korea, the development of the accident tolerant fuel is classified into near term application
technology and long term development technology. As a near term technology, the microcell UO2 pellet with metal
or ceramic materials and the surface modified claddings with coating technology are being developed. The
microcell UO2 pellet is to enhance the fission product retention and to increase thermal conductivity. The surface
modified cladding is based on the coating technology on the conventional Zr alloy cladding. It is to enhance the
oxidation and deformation resistance of the fuel cladding [3].
In the IAEA-CRP ACTOF, the research objectives and anticipated outcomes of KAERI were the
development and implementation into the FRAPCON/FRAPTRAN code of accident tolerant fuel (ATF) models
for the Cr-coated cladding and the metallic microcell pellet, model validation using data from the out-of-pile
experiment. In this paper, the main results of the research conducted by KAERI in IAEA CRP ACTOF is
summarized.
174
2. DEVELOPMENT OF ATF IN KAERI
2.1. Fuel Pellet and Cladding for ATF
KAERI has been developing Cr-alloy coated cladding and metal microcell UO2 pellets as one of the
candidates of for ATF. As the coating material, CrAl or FeCrAl was coated on the zircaloy base cladding by Arc
ion plating or a 3D laser coating technique as shown in Figure 1. Various tests were conducted to evaluate the
coated cladding performance such as the corrosion, oxidation, creep, wear, and so on [4].
FIG. 1. Manufacturing techniques and samples for surface modified zircaloy cladding [4].
As the ATF pellet, ceramic and metallic microcell UO2 pellets are being developed as shown in Figure 2.
The enhanced pellet thermal conductivity is the distinct feature of a metallic microcell UO2 pellet to effectively
decrease the fuel temperature. The out-of-pile tests were performed to evaluate the thermo physical properties,
such as the thermal conductivity, the linear thermal expansion, steam oxidation, and so on [5].
FIG. 2. Fabrication process and micro structure of the metallic and ceramic microcell UO2 pellet (Reproduced courtesy of
Elsevier [3]).
2.2. Performance Analysis for ATF
A comparative performance analysis of the cores having different ATF fuel assemblies was performed to
show the effects of the new ATF fuels on the performances of the cores such cycle length, discharge burnup, and
reactivity coefficients. The candidate ATF fuel rods includes the UO2-Mo, UO2-Cr, UO2-SiO2-TiO2 microcell
pellets with zircaloy cladding coated with or without Cr layer. The reference core using UO2 fuel rods was designed
with the design data for the OPR1000 which is Korean standard reactor such that the core has 480 EFPDs cycle
175
length. The cores using Cr-coated zircaloy cladding have a slightly reduced cycle length, but the cores using
metallic microcell pellets have more reduced cycle length because of the higher neutron absorption by metal
composition and reduced loading amount of UO2.
For the fuel performance code, some material properties from the out-of-pile test were modified in
FRAPCON code. Increased thermal conductivity of a metallic microcell pellet was modified as function of
temperature for the fresh fuel condition. For the ATF cladding, the corrosion, oxidation, and hydrogen pick-up
model was modified based on the out-of-pile test. ATF performance results are compared to UO2 and zircaloy
reference fuel. A mechanical behaviour model for multi-layer structure analysis of coated cladding was developed
and verified, as an effort of developing ATF performance model.
Safety of ATF for the DBA and BDBA was evaluated using MARS, RELAP5 and SCDAP codes, comparing
ATF micro-cell pellet and CrAl-coated claddings to the reference fuel, UO2 pellet and zircaloy cladding. The target
reactor was OPR1000 and selected accident type was LBLOCA, SBLOCA plus addition failure, and so on.
3. ESTIMATION OF FUEL PERFORMANCE FOR ATF
3.1. Development of Mechanical Model for Multi-layered Cladding for FRAPCON
The fuel performance code FRAPCON have FRACAS-I model for the small displacement deformation of
the cladding based on the thin-wall theory. The analysis is performed to determine cladding stresses and strains
[6]. However this model based on the thin-wall cannot use for the multi-layered cladding such as a coated cladding.
In order to evaluate the mechanical behaviour of the ATF cladding under the normal operation conditions, a new
analytical module, FRACAS-CT, was developed based on the thick-wall theory to consider the multi-layered
structure of the coated cladding. The FRACAS-CT module was implemented into FRAPCON code for
consideration of creep and stress relaxation behaviours of the multi-layered cladding [7]. This module can model
the cladding consisting of 2- or 3-layers like the coated cladding, assign properties to each layer, and calculate the
stresses and strains for each layer of the coated cladding. To maintain the input structure of the original FRAPCON,
the FRACAS-CT module was added as one of mechanical model option, “mechan = 3”.
To verify the implemented module, the power condition was set to have a simple power history, and the
coating material was set to be identical to the substrate, Zr cladding. For code verification, the gap thickness
behaviour was compared in Figure 3. The calculated results of the gap thickness at top, center and bottom axial
node were highly consistent, which is reasonable because material properties and inelastic model used for the
FRACAS-CT are equal with ones of the original FRAPCON.
90 FRACAS Bottom
FRACAS Center
80 FRACAS Top
Gap thickness (m)
70 FRACAS-CT Bottom
FRACAS-CT Center
60 FRACAS-CT Top
50
40
30
20
10
0
0 2 4 6 8 10 12 14
Burnup (GWd/MTU)
FIG. 3. Comparison of the gap thickness behaviours for the verification of FRACAS-CT in FRAPCON (Reproduced courtesy
of Taylor and Francis [7]).
176
3.2. Fuel Performance Analysis under Normal Operation Condition
The preliminary analysis of the fuel performance for the KAERI’s ATF concept was conducted under the
normal operation condition in ACTOF. The thermal conductivity model for metallic microcell pellet and the
corrosion model for CrAl-alloy as coating material was modified based on out-of-pile test results as shown in
Figure 4. For the simple comparison, the simple power history and the cosine axial power profile were applied as
shown in Figure 5.
The thermal conductivity of the metallic microcell UO2 pellet was improved about twice as compared with
that of the conventional UO2 pellet. Therefore, it is expected to show a great benefit in the temperature of the pellet.
The fuel centreline temperature and average fuel temperature was shown in Figure 6. At the peak power node, the
centreline temperature was reduced from 220 to 340 K and the average fuel temperature was lowered by about
100 K.
The rod internal pressure was reduced by decreasing the temperature of the pellet as shown in Figure 7. In
the low burnup region, the internal void gap volume of rod was maintained relatively large due to the reduced
deformation amount of the pellet. In the high burnup region, the fission gas will be partially released into the rod
gap region. When the FRAPFGR model was applied, the amount of fission gas release was reduced for ATF,
which can be expected due to the decrease in the temperature of the fuel pellet. In this analysis condition, the rod
internal pressure was predicted to decrease by 20% at the end of life.
FIG. 4. The thermal conductivity model of metallic microcell UO2 pellet and the corrosion model for CrAl-alloy as
coating material.
FIG. 5. The power history and the axial power profile to analysis ATF performance.
177
FIG.6. The fuel centreline temperature at the peak power node and average fuel temperature.
FIG. 7. The behaviour of the fuel rod internal pressure, the total void volume and the fission gas release.
4. CONCLUSIONS
In the IAEA-CRP ACTOF, the research objectives and anticipated outcomes of KAERI were the
development and implementation in the FRAPCON/FRAPTRAN code of ATF models for Cr-coated cladding and
metallic microcell pellet, model validation using data from the out-of-pile experiment. In the ACTOF, the work
done of KAERI is as follows.
‶In order to evaluate the mechanical behaviour of the ATF cladding under the normal operation conditions,
a new analytical module, FRACAS-CT, was developed based on the thick-wall theory to consider the multi-
layered structure of the coated cladding. The FRACAS-CT model was verified by comparison with an equivalent
finite element model. And the module was implemented into FRAPCON code with consideration of creep and
stress relaxation behaviour of the multi-layered cladding. The implemented FRACAS-CT can simulate the
mechanical response and fuel performance of the multi-layered ATF cladding″ [7]
The preliminary analysis of the fuel performance for the KAERI’s ATF concept was summarized. The major
material properties of ATF such as the thermal conductivity and thermal expansion of the pellet and the corrosion
behaviour of the CrAl-alloy coated cladding were modified based on out-of-pile test results. The differences
compared to a conventional UO2-Zircaloy fuel was assessed. From the FRAPCON results under the normal
operation condition, ATF shows a significant advantage in the reduction of the fuel centreline temperature,
178
cladding oxidation thickness, fission gas release, and so on, because of the increased thermal conductivity of the
metallic microcell pellet and the oxidation resistance of the CrAl-alloy coated cladding.
ACKNOWLEDGEMENTS
This work was supported by the IAEA CRP ACTOF and the National Research Foundation of KOREA
(NRF) grant funded by the government (MSIT: Ministry of Science and ICT) (NRF-2017M2A8A5015064)
REFERENCES
[1] BRAGG-SITTON, S.M., et al., Advanced Fuel Campaign: Enhanced LWR Accident Tolerant Fuel Performance
Matrix, INL/EXT-13-000264, Feb., (2014)
[2] CARMACK, J., GOLDNER, F., BRAGG-SITTON, S.M., SNEAD, L.L., Overview of the U.S. DOE Accident
Tolerant Fuel Development Programme, TopFuel 2013, Charlotte, North Carolina, U.S., Sept., (2013)
[3] KIM, H.G., YANG. J.H., KIM W.J., KOO, Y.H., Development Status of Accident-Tolerant Fuel for Light Water
Reactors in Korea, Nuclear Engineering and Technology 48 (2016) 1–15.
[4] KIM, H.G., KIM, I.H., JUNG, Y.-I., PARK, D.-J., PARK, J.-H., YANG, J.-H., KOO, Y.H., Progress of Surface
Modified ZR Cladding Development for ATF at KAERI, 2017 Water Reactor Fuel Performance Meeting, Jeju,
Republic of Korea, Sep., (2017).
[5] KIM, D.J., KIM, K.S., KIM, D.S., OH, J.S., KIM, J.H., YANG, J.H., and KOO, Y.H., Development status of microcell
UO2 pellet for accident-tolerant fuel, Nuclear Engineering and Technology, 50, 253–258, (2018).
[6] GEELHOOD, K.J., LUSCHER, W.G., RAYNAUD, P.A., and PORTER, I.E. FRAPCON-4.0: A Computer Code for
the Calculation of Steady-State, Thermal mechanical Behaviour of Oxide Fuel Rods for High Burnup, PNNL-19418.
Vol.1 Rev.2, (2015)
[7] KIM, D.H., KIM, H.C., SHIN, C.H., KIM, H.S., Development of FRACAS-CT Module with FRAPCON4.0P01 for
Simulation of Mechanical behaviors for Accident Tolerant Fuel Cladding in a Reactor, Journal of Nuclear Science and
Technology 56 (2019) 671-683.
179
MODELLING OF TRADITIONAL UO2 AND ADVANCED ATF FUELS BEHAVIOUR IN
INTEGRAL SEVERE ACCIDENT (SA) CODES
(Session 5)
Chairperson
J. STUCKERT
Germany
181
STATUS OF ATF MODELLING AND APPLICATION WITH ATHLET-CD
T. HOLLANDS, L. LOVASZ, C. BALS

Gesellschaft für Anlagen und Reaktorsicherheit (GRS) gGmbH,
Garching, Germany
Abstract
To investigate the cladding material FeCrAl which is one option of Accident Tolerant Fuel (ATF), the bundle experiment
QUENCH-19 was performed at Karlsruhe Institute of Technology (KIT). ATF claddings could be used as one option to reduce
e.g. the thermal impact by oxidation and to increase time margins for accident management measures. On the other hand, the
neutron absorption is higher and the melting temperature is lower of FeCrAl compared to Zry based claddings. The test
QUENCH-19 was accompanied by post test analyses at GRS with the AC² module ATHLET-CD. The experiment QUENCH-
19 was conducted similarly to the experiment QUENCH-15 for comparing the oxidation behaviour of FeCrAl and Zry at high
temperatures. In contrast to test QUENCH-15, which showed a sharp escalation after start of quenching resulting in
temperatures up to ≈1900°C, no escalation occurred during QUENCH-19 and peak cladding temperatures of only ≈1450°C
were reached. For the post test simulations an oxidation correlation for KANTHAL APMT was available and also a second
approach derived from an OECD/NEA report. Both approaches were implemented in ATHLET-CD with the assumption that
only Al2O3 was generated but no other oxides. Compared to Zry oxidation both approaches give orders of magnitude lower
oxidation rates and are only valid for one special composition, while different materials were used in the experiment. In general,
the results of the post test simulation show good agreement of the thermal behaviour. While the maximum temperature well
reproduced an underestimation of only app. 50°C, the radial temperature profile was significantly underestimated by 150-
200°C compared to the measured data. The comparison of the hydrogen production of in total 9 g in the test and less than 1 g
for both oxidation approaches in the simulation shows that the oxidation model for FeCrAl in the code has to be improved. For
a detailed evaluation of the calculated hydrogen generation the post test examination of the bundle is necessary to know which
components contribute to the total value.
1. INTRODUCTION
Zirconium based claddings have been widely used all over the world for decades due to its good mechanical,
corrosion resistant and low neutron absorption properties at around operational temperatures. At higher
temperatures, however, the zirconium becomes more reactive, in case of a nuclear accident most likely with the
steam around it. The oxidation of zirconium by the surrounding steam unfortunately has several negative effects.
The oxidation itself is an exothermic process, which can accelerate the heat-up of the core significantly. This has
a high safety relevance, because it reduces the time available to mitigate or avoid the transition to a severe accident.
As a product of oxidation ZrO2 is formed, which degrades the mechanical properties of cladding. Also, large
amounts of hydrogen are generated, which can lead to. These negative characteristics at higher temperatures give
the nuclear community reasons and motivation to develop improved, more accident tolerant cladding material.
Research and development on accident tolerant fuel (ATF) and cladding (ATC) material began already
before the accident at the Fukushima Daiichi nuclear power plant, but this event highlighted the importance of the
development of such a new material. The new material should have significantly slower oxidation kinetics
compared to the Zr-based alloys that are used in a typical light water reactor. This should reduce the hydrogen
generation rate and slow down the core heat up after core uncovery [1]. One of the most promising accident tolerant
cladding materials under investigation is FeCrAl. This material has at 1200°C an oxidation rate which is 1000
times smaller than a typical Zirconium based material.
The objective of the QUENCH experimental programme [2] at the Karlsruhe Institute of Technology (KIT)
has been the investigation of oxidation characteristics of different fuel claddings during different accident scenarios
for more than 20 years. So far 19 experiments were conducted successfully, using fuel rod simulators that were
electrically heated. During the QUENCH-19 experiment the behaviour of FeCrAl cladding material was
investigated and compared to reference experiment QUENCH-15, where the behaviour of ZIRLO material was
tested.
ATHLET-CD has been used many times for pre- and post test calculations to support the QUENCH
experimental programme. The simulation of oxidation processes is very complex; oxidation models were
developed and validated over the years with a strong cooperation with KIT to be able to predict the fuel cladding
behaviour during a nuclear accident. Most of these models were, however, developed for the oxidation of
183
zirconium based claddings. New materials require adjustments and improvements also in the existing models. The
results of the QUENCH-19 experiment helped to develop the first version of FeCrAl oxidation model for
ATHLET-CD.
This paper presents the main findings of the QUENCH-19 experiment and shows the applicability of the
newly implemented correlations for FeCrAl material compared to the old oxidation models.
2. CONDUCT AND FIRST RESULTS OF THE QUENCH-19 EXPERIMENT [3]
“The QUENCH-19 bundle experiment was conducted at KIT on 29th August 2018. It was performed in
cooperation with the Oak Ridge National Laboratory (ORNL) and was supported by the KIT programme NUSAFE.
The test objective was the comparison of FeCrAl and ZIRLO claddings under similar configuration and similar
boundary conditions as the previous QUENCH-15 experiment [3].
Different to QUENCH-15, the new test QUENCH-19 had FeCrAl (Y) claddings and 4 FeCrAl (Y) spacer
grids as well as 8 KANTHAL APM corner rods and a KANTHAL APM shroud. For both tests the PWR-typical
bundle consisted of 24 heated rods and 8 corner rods inside a shroud, which was insulated by ZrO2 fiber and
surrounded by an Inconel cooling jacket (Fig. 1).
Fig.1. Cross section of test bundle QUENCH-19 (arrangement the same as for QUENCH-15).
To get the same temperature conditions as for the QUENCH-15 test similar gas flow conditions were adopted
(Fig. 2) and the same stepwise increase of electrical power was used up to a value of 18.12 kW (Fig. 3).
Inspite of these very similar boundary conditions there was a significant difference in the bundle heating
already during the pre-oxidation period. Fig. 3 shows that there was a ≈200°C lower temperature plateau for the
QUENCH-19 test. After reaching the maximum power value there was a strong escalation of the peak cladding
temperature up to ≈ 1900°C during QUENCH-15, but only a slow increase of maximum values for QUENCH-19.
Therefore, the power was kept constant for about 2000 s until a peak cladding temperature of ≈1455°C was reached.
The reflood was initiated at ≈9100 s; at the same time the electrical power was reduced to 4.1 kW which was kept
during the reflood. A temperature excursion was not observed.
184
FIG.2. Steam and argon flow rates of QUENCH-15 and QUENCH-19 [3].
FIG.3. Comparison of QUENCH-15 and QUENCH-19 [3].
The total hydrogen release during the whole test was 9.2 g compared to 47.6 g in the QUENCH-15 test with
a much shorter period of high electrical power. While during pre-oxidation a very small release rate resulted in
only 0.3 g hydrogen mass a sharp increase of hydrogen release rate was observed after the peak cladding
temperature exceeded 1400°C.” [4]
185
3. POST TEST CALCULATIONS AND COMPARISON
3.1. Description of the code ATHLET-CD
‶ATHLET-CD (Analysis of THermal-hydraulics of LEaks and Transients with Core Degradation) [5]
describes the reactor coolant system thermal hydraulic response during severe accidents, including core damage
progression as well as fission product and aerosol behaviour, to calculate the source term for containment analyses,
and to evaluate accident management measures. It is developed by GRS in cooperation with IKE, University of
Stuttgart. ATHLET-CD includes also the aerosol and fission product transport code SAFT and as a part of the
system code AC2 it is coupled with COCOSYS for modelling thermal-hydraulics and fission product behaviour
in the containment.
The code structure is highly modular to include a manifold spectrum of models and to offer an optimum
basis for further development. ATHLET-CD contains the original ATHLET models for comprehensive simulation
of the thermal-hydraulics in the reactor coolant system. The ATHLET code comprises a thermo-fluid-dynamic
module, a heat transfer and heat conduction module, a neutron kinetics module, a general control simulation
module, and a general-purpose solver of differential equation systems called FEBE. The thermo-fluid-dynamic
module is based on a six-equation model, with fully separated balance equations for liquid and vapor,
complemented by mass conservation equations for up to 5 different non-condensable gases and by a boron tracking
model. Alternatively, a five-equation model, with a mixture momentum equation and a full-range drift-flux
formulation for the calculation of the relative velocity between phases is also available. Specific models for pumps,
valves, separators, mixture level tracking, critical flow etc. are also included in ATHLET.
The rod module ECORE consists of models for fuel rods, absorber rods (AgInCd and B4C) and for the fuel
assemblies including BWR canisters and absorbers. It describes mechanical rod behaviour (ballooning), Zr-alloy
and B4C oxidation (Arrhenius-type rate equations), Zr-UO2 dissolution and melting of metallic and ceramic
components. Melt relocation (candling) is simulated by rivulets with constant velocity and cross-section, starting
from the node of rod failure. The models allow oxidation, freezing, re-melting, re-freezing and melt accumulation
due to blockage formation. Feedback to the thermal-hydraulics considers steam starvation and blockage formation.
Besides convective heat transfer, energy can be exchanged by radiation amongst fuel rods and to surrounding core
structures.
The release of fission products is modeled by rate equations or by a diffusion model within the module
FIPREM. The transport and retention of fission products and aerosols in the reactor coolant system are simulated
by the module SAFT. For the simulation of debris bed a specific model MEWA can be applied, with its own
thermal hydraulic equation system, coupled to the ATHLET fluid-dynamics on the outer boundaries of the debris
bed. The transition of the simulation of the core zones from ECORE to MEWA depends on the degree of
degradation in the zone. Finally, the code also comprises late phase models for core slumping, melt pool behaviour
in the lower plenum and vessel failure within the module AIDA.
The code validation is based on integral tests and separate effect tests, as proposed by the CSNI validation
matrices, and covers thermal-hydraulics, bundle degradation as well as release and transport of fission products
and aerosols. They include out-of-pile bundle experiments performed in the CORA and in the QUENCH facility
as well as in-pile experiments performed in the PHÉBUS or in the LOFT facility. The TMI-2 accident is used to
assess the code for reactor applications. ″ [4]
3.2. Model changes for simulating oxidation of FeCrAl with ATHLET-CD
Two types of changes were needed to enable a post test calculation of the QUENCH-19 experiment. First
the input deck for the simulation of the QUENCH-15 experiment had to be adjusted slightly. The material
properties of the zirconium cladding had to be replaced by the material properties of the FeCrAl cladding. The
used material properties were given by KIT for the Kanthal APM material with a composition of 69% Fe, 21.6%
Cr, 4.9% Al (+ 4.5% others). It is assumed that due to the oxidation only Al2O3 is produced. The used material
properties [6, 7] are summarized in Table 1.
186
TABLE 1. USED MATERIAL PROPERTIES
Material Property FeCrAl Al2O3
Density (kg/m3) 7.1×103 3.9×103
Thermal conductivity
11 (323 K) → 35 (1673 K) 35
(W/mK)
Specific Heat
460 (293 K) → 800 (1673 K) 880
(J/kg K)
Code changes were also necessary in ATHLET-CD to compute the different oxidation kinetics of the ATF
material. The oxidation of the chosen ATF material is governed by the following chemical reaction [8]:
𝐹𝑒 𝐶𝑟 𝐴𝑙 + 3 𝐻 𝑂 → 𝐹𝑒 𝐶𝑟 𝐴𝑙 𝑂 + 3 𝐻 + 𝑧𝛥ℎ (1)
where:
𝛥ℎ is energy release during oxidation = 9.3×105 J/mol = 9.32×106 J/kgFeCrAl;
𝑥, 𝑦, 𝑧 is composition of FeCrAl molar masses of 𝑥=1.307, 𝑦=0.404 and 𝑧=0.215
A parabolic law was used to determine the oxidation rate that was derived from the analytical solution of the
diffusion equation, similarly to the method for the oxidation of Zirconium:
𝑑𝑊 = 𝐾𝑑𝑡 → = (2)
where:
𝑊 𝑚 /(𝑠𝑢𝑟𝑓𝑎𝑐𝑒 𝑎𝑟𝑒𝑎) in (kg/m2);
𝑚 is mass of the resulting oxide (Al2O3);
𝐾 is reaction rate in (kg2/m4s);
𝑑𝑡 is time step (s).
The reaction rate is determined by the following Arrhenius equation:
𝐾 = 𝐴 ∙ exp (−𝐵/𝑅𝑇) ∙ 𝑔(𝑝 ) (3)
where:
𝑅 is the gas constant (8.314 J/molK);
𝑇 is cladding temperature (K);
𝑔(𝑝 ) reduction factor to consider steam starvation (0 ≤ 𝑔(𝑝 ) ≤ 1);
𝐴, 𝐵 are rate constants as given by KIT for KANTHAL APMT [6]: 𝐴 = 3.1 kg2/m4s, 𝐵 = 2.78519×105 J/mol.
3.3. Main findings of QUENCH-19 post test simulations
Altogether three post test simulations were performed, and their results were compared to the experimental
data. All of the simulations were based on the input deck for QUENCH-15 experiment, which represents the heated
part of the QUENCH facility with 10 axial nodes and three concentric rings. The innermost part (ROD1) contains
four heated rods, the second ring (ROD2) consists of eight heated rods, while the outermost ring (ROD3) has 12
heated rods. Five spacer grids, 8 corner rods and the shroud insulation with ZrO2 are also taken into account.
Besides the heated part of the facility, steam/ argon flows and the water quenching are simulated. A more detailed
description of the used input is given in [9]. All the inputs use the material constants of Table 1 for the cladding.
Differences are caused by the oxidation models. One simulation used the oxidation kinetic defined by constants
given by KIT for KANTHAL APMT [6] (Kanthal), the other used the same constants, but the reaction rate was
multiplied by 300 (Kanthal*300) derived from [1], and the last simulation used the oxidation kinetic model by
Cathcart/Prater/Courtright [5] for zirconium (Zr).
The results of the three simulations and experimental data are depicted in Figs 4–7. Figs 4 and 5 show the
evolution of temperatures in the innermost part of the bundle (ROD1) at elevation 550 mm and 950 mm,
respectively. It is clearly visible, as expected, that at lower temperatures the change of the oxidation model doesn’t
influence the simulation results, as visible at lower elevation in Fig.4. The bundle temperatures are satisfactorily
187
reproduced by ATHLET-CD with the newly implemented ATF oxidation models also at higher temperatures and
elevations, there is no significant temperature escalation visible. The simulation with the standard oxidation model
for Zr-based cladding resulted however in much higher maximal temperatures due to the extensive heat generation
during oxidation (Fig. 5.)
2000
1750 Experiment TFS

4/9
1500 Experiment TFS
8/9
Experiment TFS
Temperature (K)
1250 14/9
Experiment TFS
17/9
1000 Experiment TFS
22/9
750 A-CD ROD1
Kanthal*300
500
0 2000 4000 6000 8000 10000
Time (s)
FIG.4. Evolution of temperatures in experiment and in the different simulations at elevation 550 mm.
2000
Experiment
1750 TFS 1/13
Experiment
TFS 15/13
1500 Experiment
TFS 17/13
Experiment
TFS 21/13
Temperature (K)
1250 Experiment
TIT G/13
A-CD ROD1
1000 Kanthal*300
A-CD ROD1
Kanthal
A-CD ROD1
750 Zr
500
0 2000 4000 6000 8000 10000
Time (s)
FIG.5. Evolution of temperatures in experiment and in the different simulations at elevation 950 mm.
Figure 6 depicts the temperature evolution at different horizontal positions of a simulation at elevation 850
mm. The comparison shows that there is a relatively high radial gradient between internal and external
temperatures with a difference of up to  200°C in the measured data (significantly larger than QUENCH-15; the
reason for these differences is not yet fully clarified [3]), while all the simulations result in a flat profile with a
188
maximum difference of 50°C. The final report of the experiment could clarify the reasons for such a large
horizontal temperature gradient.
2000
Experiment TFS
2/12
1750 Experiment TFS
8/12
Temperature (K)
1500 Experiment TFS

16/12
Experiment TFS
1250 22/12
A-CD ROD1
1000
A-CD ROD2
750
A-CD ROD3
500
0 2000 4000 6000 8000 10000
Time (s)
FIG.6. Temperature evolution at different horizontal positions at elevation 850 mm (Kanthal).
10
8
Experiment
7
A-CD Kanthal*300
Hydrogen generation (g)
6 A-CD Kanthal
A-CD Zr
5
0
0 2000 4000 6000 8000 10000
Time (s)
FIG.7. Total hydrogen generated.
The total generated mass of hydrogen (Fig.7) shows a significant underestimation of oxidation in the
simulations, in particular with the reaction rate given by [6] (Kanthal), but also the Kanthal*300 underestimated
the generated hydrogen mass, with 0.008 g and 0.35 g respectively, compared to 9 g in the experiment. The
hydrogen generation in the experiment was comparably low until the maximum temperatures was reached
(≈1400°C). A sharp increase in the hydrogen release was observed at approximately 800 s before the quenching.
189
One possible trigger for this event could be the failure and melting of the cladding of the steel thermocouples [3].
This was not considered in the post test calculations.
4. SUMMARY AND CONCLUSIONS
In order to be able to more accurately simulate accident scenarios in nuclear power plants that are loaded
with accident tolerant materials, changes to the oxidation model were implemented in the code ATHLET-CD. The
extended models were tested on the QUENCH-19 experiment by comparing them with test data as well as the
simulation results with the previous modelling. The simulations showed that the oxidation models have a big effect
on the temperature evolution of the fuel cladding and therefore also on the whole accident scenario. The newly
implemented correlations for the oxidation of FeCrAl material deliver more reasonable results than the correlations
for Zr-based claddings, but their accuracy and applicability have to be further improved. GRS is plans to further
improve the oxidation models of ATF material in ATHLET-CD.
ACKNOWLEDGEMENTS
The development and validation of ATHLET-CD are sponsored by the German Federal Ministry for
Economic Affairs and Energy (BMWi) based on decisions of the German Bundestag.
REFERENCES
[1] NUCLEAR ENERGY AGENCY (NEA): State-of-the-Art Report on Light Water Reactor Accident-Tolerant Fuels, NEA
No. 7317, OECD 2018.
[2] STUCKERT, J., et al.: Experimental programme QUENCH at KIT on core degradation during reflooding under LOCA
conditions and in the early phase of a severe accident, IAEA-TECDOC-CD-1775, Nov. 2015.
[3] STUCKERT, J., et al.: First results of the bundle test QUENCH-19 with FeCrAl claddings, 24th International Quench
Workshop, Karlsruhe, November 2018.
[4] BALS, C., et al.: POST TEST SIMULATION OF QUENCH TEST 19 WITH ATHLET-CD, IAEA TECDOC, CRP
ACTOF, (to be published).
[5] AUSTREGESILO, H., et al., ATHLET-CD 3.2 User’s Manual, GRS Report, GRS-P-4/Vol. 1, Dec. 2018.
[6] STUCKERT, J., KIT, Mail from 26.7.2018.
[7] ACCURATUS, Ceramic Materials and Ceramic Components, 9.10.2018,
https://accuratus.com/alumox.html.
[8] US DOE: Handbook on the Material Properties of FeCrAl Alloys for Nuclear Power Production Applications,
ORNL/TM-2017/186.
[9] BALS, C., et al., Post test calculation of QUENCH-15 with ATHLET-CD, 15th International Quench Workshop,
Karlsruhe, November 2009.
190
ADVANCED LOCA SAFETY CRITERIA: IMPLEMENTATION IN MECHANISTIC
FUEL PERFORMANCE CODES AND APPLICATIONS
M.S. VESHCHUNOV
International Atomic Energy Agency,
Vienna International Centre,
PO Box 100, 1400 Vienna,
Austria
A.V. BOLDYREV, V.E. SHESTAK

Nuclear Safety Institute (IBRAE),
Russian Academy of Sciences,
52, B. Tulskaya, Moscow, 115191,
Russian Federation
Abstract
The advanced Chung-Kassner criteria on capability of fuel rods to withstand thermal shock during LOCA reflooding and
to withstand post-quench fuel handling, as well as complementary cladding embrittlement criteria of KAERI and KIT, are
compared with the traditional LOCA safety criterion related to 17% ECR oxidation. The advanced criteria were implemented
in the thermo mechanical code SVECHA/QUENCH (S/Q) and successfully applied to predicting the fuel rod failure behaviour
in a postulated LOCA scenario and in high temperature (1400C) pre-oxidation and quenching tests performed with fresh and
irradiated fuel rod simulators. Extension of the advanced criteria to coated Zr alloy claddings is discussed.
1. INTRODUCTION
In accordance with the IAEA Safety Standards [1] developed on the basis of international consensus by
revision of the Safety Guide No. NS-G-1.12 [2], to ensure that the structural integrity of the fuel rods is preserved
the following design limit should be defined and justified: the total oxidation of the cladding should remain below
limits such that the cladding can still withstand accident induced loadings (e.g. in the quenching phase of a loss of
coolant accident). Such limit should be determined by experiments that take into account pre-transient in-reactor
cladding oxidation and transient oxidation (outer side oxidation and possibly inner side oxidation), pre-transient
and transient hydrogen absorption, as well as chemical interactions between fuel pellets and cladding material
(§3.73b of [1]).
In addition to the thermal shock failure characteristics of zircaloy cladding during quenching, fuel rods
should be designed to withstand loadings resulting from post-transient fuel assembly handling, storage and
transport to a reprocessing or disposal facility (§3.73e of [1]). Following this requirement of the Safety Standards,
one must evaluate the impact and load-deflection properties of cladding that has been subjected to postulated
LOCA transients.
The existing regulatory LOCA safety criteria accepted in many countries establish the 1477 K (1204°C) peak
cladding temperature and 17% equivalent cladding reacted (ECR) maximum oxidation, calculated as function of
oxidation temperature and square root of time based on the Baker-Just correlation (see, e.g. [3]), with some
modifications in different countries. The purpose of these criteria, regarding maximum cladding temperature and
total oxidation, is to ensure that the cladding shall remain sufficiently ductile so that it does not shatter into pieces
during and after the quench phase of the LOCA transient.
In some countries (e.g. Japan) the current LOCA criterion on fuel safety is based on the failure threshold
value determined in the integral thermal shock tests. In other countries, although most of the attention is focused
on the events of the first few minutes after a postulated major cooling line break, the long term maintenance of
cooling would be equally important [4]. In particular, the U.S. Atomic Energy Commission (AEC) promulgated
Criterion 3 of the Interim Acceptance Criteria for ECCS for LWR [5] which states that: “The clad temperature
transient is terminated at a time when the core geometry is still amenable to cooling, and before the cladding is so
embrittled as to fail during or after quenching.”
191
2. CLADDING CAPABILITY TO WITHSTAND THERMAL SHOCK
The thermal shock failure properties of Zircaloy-4 cladding have been investigated for tubes oxidized at
controlled temperatures for various times and bottom flooded with water, or in a more advanced tests, ruptured in
steam under transient-heating conditions, oxidized at controlled temperatures and bottom flooded [3, 6]. The
failure map based on the ECR parameter and the oxidation temperature in steam, Fig. 1, typically shows that no
failures occurred for specimens oxidized to 17% of the total cladding thickness, based on the ECR parameter, and
temperatures to 1477 K, which are the present oxidation limits in the acceptance criteria for emergency-core-
cooling systems in LWRs. Open and closed symbols in Fig. 1 respectively denote the specimens which survived
and failed after a cold-water injection followed after a high temperature oxidation. However, these observations
highlighted a rather high conservatism of the currently accepted criterion. For instance, it was deduced in [6, 7]
that the critical ECR which is defined as the minimum point where a cladding fails, lies between 20% and 30%.
FIG. 1. Failure boundary of zircaloy cladding tubes after oxidation at high temperature and direct quenching from peak
oxidation temperature (left panel from [3] and [6]; right panel reproduced courtesy of Elsevier [7]).
To overcome this problem of the traditional ECR criterion, Chung and Kassner performed detailed analysis
of their own and former test results [6]. In these tests the mechanical response of zircaloy cladding under thermal
shock conditions typical for hypothetical LOCA situations in LWRs has been evaluated. Various failure maps for
fracture of the cladding by thermal shock were developed relative to the maximum oxidation temperature and
various time dependent oxidation parameters. The best correlation of the thermal shock failure data was obtained
relative to the thickness of the ductile -phase layer; i.e. cladding in which the thickness of the -phase, with  0.9
wt % oxygen, exceeded 0.1 mm remained intact irrespective of the specimen thickness, oxidation temperature and
total oxygen content of the cladding, Fig. 2.
FIG. 2. Failure map for Zircaloy-4 cladding by thermal shock relative to the wall thickness with  0.9 wt% oxygen after
isothermal oxidation and flooding with water at the oxidation temperature (from [6]).
192
Comparison of Figs. 1 and 2 shows that a high conservatism of the traditional 17% ECR criterion, that
notably underestimates a cladding capability to withstand thermal shock during LOCA reflooding, is essentially
reduced in the advanced Chung-Kassner criterion eventually formulated in [6] as:
— Capability to withstand thermal shock during LOCA reflooding: calculated thickness of the cladding
with  0.9 wt % oxygen should be greater than 0.1 mm.
In particular, it was noted in [6] that the absorbed hydrogen (analyzed by an inert-gas fusion technique) had
a minimal effect relative to that of oxygen on the thermal shock behaviour, assuming that the rapid cooling rate
below the Leidenfrost temperature suppresses hydride precipitation. This conclusion was confirmed in more recent
KAERI tests of J.H. Kim et al. [7]. In those tests, to simulate LOCA conditions specimens were oxidized in a
flowing steam at a desired temperature and time. To simulate the transient, the specimens were cooled at an
intermediate temperature of 700C for 100 s after being oxidized, and then quenched.
In the failure map obtained in these tests, Fig. 3, dotted lines correspond to different oxidation levels in terms
of ECR; closed (open) symbols represent failed (intact) claddings during the water quench. Post test absorbed
hydrogen content in the oxidized cladding was measured by a gas analysis method and represented by solid lines
in Fig. 3. This diagram shows that the effect of the absorbed hydrogen (up to 700 wppm) on the failure behaviour
during quenching of the Zircaloy-4 cladding was low, i.e. it practically did not influence the failure criterion in
terms of ECR.
FIG. 3. Changes of the material properties of the fuel cladding with the oxidation temperature and time, depending on absorbed
hydrogen content (< 700 wppm) (Reproduced courtesy of Elsevier [7]).
A similar conclusion was derived in the JAERI tests [8], also for pre-hydrided claddings, showing that the
influence of pre-hydriding, up to 600 ppm, is negligible on the thermal-shock fracture threshold under the
unrestrained condition, whereas the effect is more remarkable in the case of restrained test rods.
Furthermore, the oxygen content inside the prior -layer was additionally determined after oxidation and
quenching in the tests [7], Fig. 4. These data were not correlated with the thermal shock failure data; however,
they were further used for interpretation of cladding post-quench embrittlement, see Section 3 below.
193
FIG. 4. Changes of the material properties of the fuel cladding with the oxidation temperature and time, depending on oxygen
content inside the prior -layer (Reproduced courtesy of Elsevier [7]).
3. CLADDING POST-QUENCH EMBRITTLEMENT
In order to justify the fuel rod capacity to withstand loadings resulting from post-transient fuel assembly
handling, storage and transport (following the requirement, §3.73e, of the IAEA Safety Standards [1]), one must
evaluate the impact and load-deflection properties of cladding that has been subjected to postulated LOCA
transients.
In particular, critical fracture loads under conditions of impact, tension, and diametral compression were
determined in [6] as functions of the degree of oxidation of the material and microstructure produced by cooling
through the temperature range of  phase transformation at different rates. Quantitative correlations were
established between the failure-impact energy and the diametral-compression properties for cladding from the
integral tube-burst/thermal-shock tests. As with the thermal shock failure characteristics, the best correlation of
the room-temperature impact and diametral-compression properties was obtained relative to the thickness of the
transformed -phase layer with a maximum oxygen content. The results indicated that zircaloy cladding that could
survive an impact energy of 0.3 J at 300 K after rupture, oxidation in steam and cool-down would also withstand
either diametral deflections to the point of cladding-pellet contact or deflections of  3 mm before fragmentation
occurs, provided the thickness of the -phase layer containing  0.7 wt% oxygen is  0.3 mm irrespective of the
wall thickness produced by ballooning and rupture of the cladding and the maximum oxidation temperature, Fig. 5.
These results also showed that for a given thickness and a given oxygen content in transformed-beta layer,
hydrogen at concentrations up to  2000 wppm would not cause a significant change in the room-temperature
failure-impact behaviour.
On this base, the following embrittlement criterion was formulated by Chung and Kassner [6]:
— Capability to withstand fuel handling, transport and storage: calculated thickness of the cladding with 
0.7 wt % oxygen should be greater than 0.3 mm.
194
FIG. 5. Impact failure map as function of hydrogen content and thickness of the -layer containing ≤ 0.7 wt.% oxygen;
Zircaloy-4 tube burst, oxidized, slow-cooled and quenched (from [3, 6]).
A similar result was confirmed later in the KAERI [9] tests, which evidenced that after the simulated LOCA
the cladding lost its ductility abruptly when the prior -layer decreased below the threshold metal thickness of 0.3
mm, Fig. 6.
In the subsequent KAERI tests [7] (briefly described above in Section 2) at the same LOCA simulation
facility, the ductility of the oxidized cladding was evaluated by mechanical tests such as ring compression test and
3-point bend test at room temperature after the LOCA test. The absorbed energy during the mechanical test was
accurately determined as the area under the load–displacement curve and showed an abrupt decrease with the loss
of the cladding ductility. As a result of such studies the empirical failure diagram of the Zircaloy-4 cladding was
constructed in [7], Fig. 7.
FIG. 6. Comparison of the microstructural parameters presenting the ring-compression ductility and absorbed hydrogen
content vs. residual -layer thickness. An open symbol represents a specimen that survived and a closed symbol a specimen
that failed during the water quenching (Reproduced courtesy of Elsevier [9]).
195
“Brittle failure at a thermal shock” (red region in Fig. 7) indicates that the cladding is too brittle to withstand
a thermal stress during a water quenching. ‘Brittle fracture at the mechanical test’ (green region in Fig. 7) means
that although the cladding at first survived the water quenching, it has already lost its mechanical ductility so that
it could be failed during a handling, such as refuelling or transporting the fuel bundles to the spent fuel storage.
‘Ductile bending’ (blue region in Fig. 7) means that the cladding can assure its mechanical ductility after a thermal
shock.
FIG. 7. Failure map of Zircaloy-4 claddings at different oxidation temperatures and times (Reproduced courtesy of
Elsevier [7]).
By superposition of Figs 4 and 7, one can see that the brittle-ductile transition curve (blue line in Fig. 7)
practically coincides with the curve in Fig. 4 corresponding to the oxygen content in the prior -layer of 0.5 wt%.
This conclusion is also confirmed by the measurements of 3-point bending moment with respect to the
oxygen content inside beta layer presented in Fig. 8 (reproducing data from [7]). In this figure, it is clearly seen
that the absorbed energy of ductile and brittle cladding can be easily distinguished by the value of
1468±153 kgfmm and 47±35 kgfmm, where ductile and brittle claddings could be separated by the vertical
dashed line of  0.5 wt% of oxygen content inside the prior beta layer, regardless of test temperature and time.
FIG. 8. Plot of 3-point bending moment with respect to the oxygen content inside beta layer (by courtesy of J.H. Kim with
KAERI permission).
196
These important observations allowed the authors of [7] deriving a conclusion:
— When the absorbed oxygen content in the prior -layer was less than 0.5 wt%, the cladding maintained its
mechanical ductility after the LOCA test.
This conclusion can be considered as a new embrittlement criterion for Zry-4 claddings being complementary
to the Chung-Kassner criterion on capability to withstand fuel handling, considering that the latter criterion was
consistent with the results of the earlier KAERI tests [9] at the same test facility (see above). As shown in Fig. 3,
the cladding post test hydrogen content did not exceed 700 wppm, this value determines the applicability range of
the KAERI criterion. The observed weak dependence of the new criterion on the hydrogen content in this range
(as seen from comparison of Figs. 3 and 4) well correlates with the similar conclusion regarding the weak
dependence of the second Chung-Kassner criterion on the absorbed hydrogen content (below  2000 wppm).
On the other hand, by superposition of the constant hydrogen content curves from Fig. 3 with the brittle-
ductile curve in Fig. 7 one can deduce a strong dependence of the critical ECR-BJ (Baker-Just) for the ductile-
brittle transition on the hydrogen content, as shown in Fig. 9. This may substantiate new attempts of modification
of the post-quench embrittlement criterion in terms of the critical ECR dependence on the pre-quench hydrogen
content (see, e.g. Fig. 10 from [10, 11]); however, demonstrates practical inconvenience of this approach in
comparison with the KAERI and Chung-Kassner advanced criteria formulated purely in terms of oxygen
characteristics.
Unfortunately, predictions from Fig. 9 cannot be quantitatively compared with Fig. 10 (in terms of ECR-CP
(Cathcart-Pawel)), since no measurements of post-quench hydrogen content were presented in [10, 11]. However,
they seem qualitatively consistent with the results derived from Fig. 10 demonstrating a similar shift of the critical
hydrogen contents,  250-300 wppm, between pre-quench (from Fig. 9) and post-quench (from Fig. 10) data at
fixed values of ECR-CP (6 and 10%).
These predictions are also consistent with measurements in the recent QUENCH-LOCA tests at the KIT test
bundle facility [12], where post-quench brittle failure of ballooned and ruptured claddings was detected at narrow
axial bands above and below the burst opening with the local hydrogen concentrations above 1500 wppm.
Additional measurements demonstrated that ECR-BJ for those claddings near their burst opening was between 4
and 6% and the mean hydrogen content at these axial locations was in the range 720–750 wppm, in a fair agreement
with conclusions derived above from analysis of the KAERI tests in Fig. 9.
These KIT observations offer a complementary criterion for the post-quench cladding embrittlement in terms
of so called ‘secondary hydriding’ mechanism:
— Cladding maintains its mechanical ductility after the LOCA test when local hydrogen concentration (in
the secondary hydriding zones) is less than 1500 wppm.
This criterion is consistent with the Chung-Kassner and KAERI advanced criteria (as above explained);
however, its practical applications is complicated by necessity of implementation of advanced mechanistic models
describing the complex secondary hydriding phenomena and calculating local hydriding levels in the ruptured
claddings (as, e.g. realized in the mechanistic code SVECHA/QUENCH, see below Section 4.1).
197
FIG. 9. Superposition of the absorbed hydrogen content curves from Figure 3 with the ductile-to-brittle transition curve from
Figure 7, demonstrating a notable dependence of the critical ECR for post-quench embrittlement on absorbed hydrogen.
FIG. 10. Ductile-to-brittle transition oxidation level (in terms of ECR-CP) as a function of pre-test hydrogen content in
cladding metal for as-fabricated, prehydrided, and high-burnup cladding materials (from [11]).
In summary two qualitative conclusions concerning applicability of the traditional 17% ECR criterion may
be derived: (a) the criterion notably underestimates cladding capability to withstand thermal shock during LOCA
reflooding (i.e. too conservative), and (b) the criterion notably overestimates ductile-to-brittle transition oxidation
level, reflecting cladding post-quench ductility and capability to withstand fuel handling, transport and storage
(i.e. too excessive). For this reason, the alternative, more advanced and physically based Chung-Kassner and
KAERI criteria as well as the KIT complementary criterion may be used.
198
4. APPLICATION OF THE ADVANCED CRITERIA
For practical application of the Chung-Kassner criteria, a cladding oxidation model based on the diffusion
approach should be utilised, in accordance with the recommendation [6]. In particular, the mechanistic code
SVECHA/QUENCH (S/Q) designed for detailed modelling of fuel rod behaviour under re-flooding conditions can
be applied. This code, currently a sub-module of the fuel performance and safety code SFPR [13], was initially
developed at IBRAE [14] for severe accidents and then adapted for detailed analysis of the single-rod QUENCH
rig tests in close cooperation with FZK experimentalists [15, 16]. In the S/Q code, the main physical phenomena
occurring during degradation of fuel rods are mechanistically considered: cladding oxidation, cladding mechanical
deformation, hydrogen uptake and release by oxidized cladding, fuel rod liquefaction and downward relocation,
etc. [15, 16].
In the numerical scheme of S/Q, a fuel rod is divided into meshes along the axial direction. Each mesh
comprises a fuel pellet and Zr alloy cladding. For each mesh the cladding is considered as a cylindrical shell
consisting of three layers at each surface of the cladding (external and internal): oxide, -Zr(O) and -Zr layers.
The layers growth kinetics is calculated by the oxidation model which is based on the solution of the oxygen
diffusion problem across the multi-layered cladding structure. In particular, the developed diffusion approach
allows self-consistent consideration of internal cladding oxidation (and subsequent embrittlement) owing to
interactions with the fuel after the gap collapse.
As an example of application of the Chung-Kassner criteria to a LOCA-type scenario using the mechanistic
S/Q code, a WWER rod with typical geometry parameters is considered in calculations, Fig. 11.
1400 2.5
Burst
1200
2
1000
Overpressure, MPa
Temperature, °С
1.5
800
600
1
400
0.5
200
0 0
0 100 200 300 400 500 600
Time, s
Parameters of WWER fuel rod
Cladding material E110
Cladding OD 9.1 mm
Cladding thickness 0.69 mm
Pellet material UO2
Pellet OD 7.56 mm
Pellet ID 2.4 mm
FIG. 11. Temperature scenario of transient (LOCA) test with a WWER fuel rod for calculations.
The code predicts the cladding ballooning, burst and subsequent double-side oxidation at the axial location
of the burst opening, as presented in Fig. 12.
199
8 800
ZrO2
7 -Zr(O)
600
Layer radii positions, mm
-Zr
Layer thickness, mm
6 ZrO2
-Zr(O) 400
-Zr
5 gap
UO2 pellet
200
4
3 0
0 100 200 300 400 500 60 0 0 100 200 300 400 500 60 0
Time, s Time, s
FIG. 12. Calculated by the S/Q code evolution of the oxidized cladding layers radial positions (left) and thicknesses (right) at
the axial position of burst opening.
The corresponding evolution of the oxygen concentration profile in the cladding with time is presented in
Fig. 13 (left). From the final oxygen distribution in the -phase layer, Fig. 13 (right), one can see that the oxygen
profile in this layer is rather flat and corresponds to the oxygen content of  0.8 wt% and the thickness of
 0.19 mm. This implies that the calculated thickness of the cladding with  0.9 wt% oxygen is greater than 0.1
mm, whereas the calculated thickness of the cladding with  0.7 wt% oxygen is less than 0.3 mm. Therefore, in
accordance with the two Chung-Kassner criteria, the cladding withstands the thermal shock on cool-down but fails
during post-quench handling and/or transportation.
As expected, the same conclusion may be derived by application of the KAERI criterion [7]: since the
absorbed oxygen content in the prior -layer is more than 0.5 wt% (as seen from Fig. 13 (right)), the cladding
does not maintain its mechanical ductility after the LOCA test.
30 2.5
Time instants, s oxygen

300
400 2
20 500
Oxygen, wt %
Oxygen, wt %
1.5
10
1
0.9 wt %
0.7 wt %
0 0.5
6.7 6.8 6.9 7 7.1 7.2 7.3 6.8 6.9 7 7.1
Radial position, mm
Radial position, mm
FIG. 13. Evolution of oxygen distribution in cladding layers (left) and the final distribution in the -phase and -phase layers
(right) at the axial position of the burst opening.
4.1. Secondary hydriding criterion
A series of new out-of-pile QUENCH-LOCA bundle tests were performed at the QUENCH facility in
Karlsruhe Institute of Technology (KIT, Germany) [12], as mentioned above in Section 3. Neutron radiography of
cladding tubes showed that rather high hydrogen concentrations were attained in narrow bands on the boundaries
of the internal oxidized region around burst positions (secondary hydriding). Tensile tests at room temperature
evidenced brittle fracture at these bands for several inner rods. Corresponding maximal local hydrogen
concentrations measured at the fracture positions were about 1500 wppm and more. Peripheral rods had lower
hydrogen contents and were mostly fractured due to stress concentration at burst opening edges. Other tensile
tested claddings failed after necking far away from burst region.
In order to simulate the double-side oxidation and hydriding of burst cladding under LOCA conditions, the
single-rod SVECHA/QUENCH (S/Q) code has further been developed and verified [18] against various test series.
In application to the KIT QUENCH-L0 test [19] the code adequately predicted axial distributions of oxygen and
200
hydrogen in the cladding, Fig. 14. In particular, the code correctly calculated a rather high hydrogen content in the
secondary hydriding zones, > 1500 wppm, and thus, in accordance with the KIT criterion, predicted the brittle
fracture of the corresponding rods, in agreement with the observations (see [18]).
FIG. 14. Calculated axial distribution of the total oxygen content (left) and the total hydrogen content (right) at the moments
of 40, 60 and 80 s of the KIT QUENCH-L0 test (Reproduced courtesy of Elsevier [18]).
5. EXTENSION OF THE ADVANCED CRITERIA TO HIGH TEMPERATURES (DEC)
The Chung-Kassner criteria were developed on the base of the oxidation tests performed in a wide
temperature interval (far beyond LOCA conditions), up to 1500-1900°C (see Figs 1 (left) and 2). This allows
extending their applicability to Design Extension Conditions (DEC), in particular investigated in the FZK tests on
high temperature quenching of fuel rod simulators with zircaloy cladding by water or steam at a small-scale rig
[16]. The principal ‘input’ experimental parameters were the extent of pre-oxidation and the temperature of the
tube before cooldown. A total of 24 experiments has been performed with specimens pre-oxidized at 1400C with
oxide layer thicknesses varying from  100 to  350 µm which were rapidly cooled by water or steam from 1100,
1200, 1400 and 1600C. The main characteristics of the quenched cladding integrity was post test appearance of
cracks. Post test detection of through wall cracks with oxidized surfaces was interpreted as the cladding shattering
during quenching. In other samples non-oxidized cracks normally penetrated through the oxide and -phase layers
but were arrested in the -phase layer; these samples were interpreted as the cladding withstanding thermal shock.
S/Q calculated oxygen distributions along radial direction in the -phase layer of the cladding pre-oxidized
at 1400C to different extents are presented in Fig. 15. One can see that the first Chung-Kassner criterion (on
failure during quenching) is valid up to pre-oxidation of 130 µm. Hence, if pre-oxidation at any location does not
exceed 130 µm then the through-wall cracks are not predicted. This conclusion was fairly confirmed by
experimental observations [16].
201
1.7
100 mcm,1600C
1.5
1.3
200 mcm,1400C
Oxygen, wt.%
1.1
135 mcm,1400C
0.9
130 mcm,1400C
100 mcm,1400C
0.7
0.5
0.3
4.6 4.7 4.8 4.9 5.0 5.1 5. 2
Radial direction, mm
FIG. 15. S/Q simulations for various quenching temperatures and pre-oxidations; red (green) zone represents applicability of
the first (second) Chung-Kassner criterion to withstand thermal shock (post test handling).
Later the same approach was applied for interpretations of similar quenching tests with WWER-1000
irradiated fuel rod fragments with maximum burnup of 5065 MW·d/kgU, performed in RIAR (Dimitrovgrad)
[17]. In these samples, cladding oxide layers (external and internal) did not exceed a few µm, typically for the
E110 cladding material; this allowed neglecting them with a good accuracy in calculations.
In these tests depending on the extent of pre-oxidation (at 1400C), different samples were intact (green),
broken during quenching (red) or broken during post-quench handling (yellow), as shown in Fig. 16 and Table 1.
Predictions of the S/Q code for samples quenched from the pre-oxidation temperature 1400C, based on the
calculations from Fig. 15, are presented in Table 1 and reasonably interpret the observed failure modes using the
first and the second Chung-Kassner criteria.
Indeed, in accordance with the first criterion the code predicts that a rod withstands thermal shock (designated
in Table 1 as +1) or fails during thermal shock (designated as 1), depending on the oxide thickness (with a
threshold of  130 µm as explained above). In accordance with the second criterion, fuel rods survived during
quenching will withstand post test handling (designated as +2) or fail (designated as 2) with the oxide thickness
threshold of  100 µm. Some uncertainty of the fuel rod failure in the two tests with similar oxide layers of
 100 µm (tests #36 and #39) was probably connected with the proximity of their thicknesses to the threshold
value (as interpreted in Table 1).
202
Test: ##40, 46 # 36 # 39 # 54 # 55
FIG. 16. Post test appearance of irradiated fuel rods in the RIAR tests [17].
TABLE 1. CONDITIONS OF RIAR TESTS [17] WITH FUEL RODS PRESENTED IN FIGURE 16.
Test ##40, 46 # 36 # 39 # 54 # 55
Pre-oxidation 1400C/ 1400C/ 1400C/ 1400C/ 1400C/
temperature/ 0s  240 s  240 s 1050 s 1100 s
duration
Quenching 1400C 1400C 1400C 1400C 1400C
temperature
Oxide thickness 0 µm  100 µm  100 µm 280 µm 300 µm
Failure mode Intact Intact Brocken Brocken during Brocken during
during quenching quenching
handling
Failure criteria +1/+2 +1/+2 +1/2 1 1
6. EXTENSION OF THE ADVANCED CRITERIA TO COATED ZIRCONIUM CLADDINGS (ATF)
The traditional 17% ECR criterion cannot be directly applied to new designs of Zr alloy claddings with Cr
coatings, since the growth of the chromium oxide layer is strongly reduced, whereas oxygen continuously
penetrates into the Zr substrate, forms - and -phase layers and eventually embrittles the cladding (owing to
oxygen saturation of the ductile -phase layer). For this reason, one may expect, lacking the Zr oxide layer growth,
that the Zr cladding embrittlement occurs at ECR much lower than 17%. In particular, this was confirmed in the
recent Round Robin tests performed within the IAEA Coordination Research Project (CRP) on Analysis of Options
and Experimental Examination of Fuels with Increased Accident Tolerance (ACTOF) [20].
In this situation it is natural to expect that the failure criteria related to the ductility of the -phase layer are
applicable to the modified claddings. In particular, it was shown in the ring compression tests with oxidized
claddings (performed by Czech participants, CTU and UJP from Prague) [20] that the cladding embrittlement can
be well correlated with the oxygen content in the prior -phase, Fig. 17; this finding is in a fair accordance with
the KAERI criterion [7] presented in Section 4 (cf. Fig. 8). It was assumed in the new tests that the critical value
of the oxygen content in E110 alloy is 0.30.4 wt%, which slightly modifies the KAERI criterion [7] of ≈ 0.5 wt%
for Zry-4 presented in Section 4; however, the latter criterion seems also consistent with the new experimental
data [20] (Fig. 17).
203
FIG. 17. Residual ductility of the Cr-coated Zr1%Nb samples in relation with oxygen content in prior β-Zr layer (from [20]).
Lacking currently a diffusion model for oxidation of coated claddings, the second Chung-Kassner criterion
(on cladding embrittlement) was not verified against these test data; however, considering its mutual consistency
with the KAERI criterion (as discussed in Section 3) one may expect its applicability to the coated claddings as
well.
Extension of the mechanistic oxidation model to consideration of the coated Zr claddings will also allow
verifying applicability of the Chung-Kassner and KAERI criteria (as well as the KIT secondary hydriding
criterion) on the base of future tests on the coated Zr cladding quenching, e.g. foreseen in the new IAEA CRP on
Testing and Simulation of Advanced Technology Fuels (ATF-TS), planned for 2020–2023.
7. CONCLUSIONS
Basing on the analysis of the experimental data on Zr alloy cladding thermal shock and post-quench ductility
tests, it was emphasized that the traditional LOCA safety criterion related to 17% ECR: (a) notably underestimates
cladding capability to withstand thermal shock during LOCA reflooding (i.e. is too conservative), and (b) notably
overestimates ductile-to-brittle transition oxidation level, reflecting cladding post-quench ductility and capability
to withstand fuel handling, transport and storage (i.e. too excessive).
For this reason, the alternative, more advanced and physically based Chung-Kassner criteria as well as the
KAERI criterion for post-quench embrittlement are considered for practical applications.
Applicability of the advanced criteria in the majority of the currently existing codes is restricted owing to a
simplified approach for evaluation of the cladding oxidation using Baker-Just or Cathcart-Pawel correlations for
the weight gain realized in these codes, whereas these criteria require calculation of the oxygen concentration
profiles in the different phases (-Zr(O), -Zr and ZrO2-x) of the oxidized cladding. This problem can be overcome
in more mechanistic fuel performance codes using oxygen diffusion models, validated against isothermal and
transient oxidation tests.
As an example of the advanced criteria utilization, the mechanistic fuel performance and safety code SFPR
using the thermo mechanical module SVECHA/QUENCH (S/Q) with the oxygen diffusion model for Zr alloy
cladding oxidation, is considered. The code predicts cladding ballooning, burst and double-side oxidation during
LOCA scenario and allows natural application of the advanced failure criteria by calculation of the oxygen profiles
in the cladding.
Being extended to mechanistic consideration of the cladding secondary hydriding phenomenon, the S/Q code
allows applying the KIT complementary post-quench embrittlement criterion in practical calculations.
The Chung-Kassner criteria developed on the base of the oxidation tests performed in a wide temperature
range (up to 1500–1900°C) that is far beyond LOCA conditions, can be broadened to Design Extension Conditions
(DEC). This allowed reasonable predictions of the S/Q code for the fuel rod failure behaviour investigated in the
204
FZK tests (with fresh Zry-4 claddings) and in the RIAR tests (with irradiated WWER fuel fragments) on high
temperature (1400C) pre-oxidation and quenching by water or steam.
The applicability of the KAERI advanced failure criteria (with some potential modifications) for oxidized Zr
alloy claddings with Cr coatings was confirmed in the recent ring compression tests performed within the IAEA
Coordination Research Project (CRP) ACTOF. On this base it is assumed that extension of the mechanistic
oxidation model to consideration of the coated claddings will allow verifying applicability of the advanced Chung-
Kassner and KAERI criteria (as well as the KIT complementary criterion) on the base of future tests on the coated
cladding quenching (which are foreseen, e.g. in the subsequent IAEA CRP on ATF in 2020-2023).
ACKNOWLEDGEMENTS
The authors thanks Dr. J.H. Kim (KAERI, Republic of Korea) for presenting additional information on his
tests [7] and new figure (Fig. 8), as well as valuable comments to the text; Dr. J. Stuckert (KIT, Germany) for
presenting his unpublished data [12] and valuable discussions; Dr. M. Ševeček (CTU, Czech Republic) for sharing
his unpublished results (Fig. 17).
REFERENCES
[1] INTERNATIONAL ATOMIC ENERGY AGENCY, Design of the Reactor Core for Nuclear Power Plants, IAEA Safety
Standards, Specific Safety Guide Series SSG-52 (2019).
[2] INTERNATIONAL ATOMIC ENERGY AGENCY, Design of the Reactor Core for Nuclear Power Plants, IAEA Safety
Standards Series No. NS-G-1.12, IAEA, Vienna (2005).
[3] HACHE, G., CHUNG, H.M., The history of LOCA embrittlement criteria, No. ANL/ET/CP-102703, Argonne National
Lab., IL USA (2001).
[4] Interim Acceptance Criteria for Emergency-Core-Cooling Systems for Light-Water Power Reactor, U.S. Federal
Register 36 (125), (1971) 12247-12250.
[5] Atomic Energy Commission Rule-Making Hearing, Opinion of the Commission, Docket RM-50-1 (28 December 1973).
[6] CHUNG, H.M., KASSNER, T.F., Embrittlement criteria for Zircaloy fuel cladding applicable to accident situations in
light-water reactors. Summary report, No. NUREG/CR-1344; ANL-79-48, Argonne National Lab., IL USA (1980).
[7] KIM, J.H., LEE, M.H., CHOI, B.K. and JEONG, Y.H., Failure behavior of Zircaloy-4 cladding after oxidation and water
quench, Journal of Nuclear Materials 362 (2007) 36–45.
[8] NAGASE, F, FUKETA, T., Effect of pre-hydriding on thermal shock resistance of Zircaloy-4 cladding under simulated
loss-of-coolant accident conditions, Journal of nuclear science and technology 41 (2004) 723–730.
[9] KIM, J.H., LEE, M.H., CHOI, B.K., JEONG, Y.H., Embrittlement behavior of Zircaloy-4 cladding during oxidation and
water quench, Nuclear Engineering and Design 235 (2005) 67-75.
[10] YAN, Y., BURTSEVA, T.A., BILLONE, M.C., Letter Report from Argonne National Laboratory to NRC “Post-quench
Ductility Results for North Anna High-burnup 17×17 ZIRLO Cladding with Intermediate Hydrogen Content”, US NRC
ADAMS Accession Number ML091200702 (April 17, 2009)
[11] CLIFFORD, P.M., FLANAGAN, M.E., Proposed Modification of Cladding Embrittlement Criteria, Proceedings of Top
Fuel 2009, Paris, France, September 6-10, 2009, Paper 2120.
[12] STUCKERT, J., GROSSE, M., STEINBRÜCK, M., WALTER, M., WENSAUER, A., “Results of the QUENCH-LOCA
experimental programme at KIT”, ASTM Special Technical Publication in Proceedings of 19th International Symposium
“Zirconium in the Nuclear Industry”, STP-1622, Manchester, May 2019 (in press).
[13] VESHCHUNOV, M.S., BOLDYREV, A.V., OZRIN, V.D., SHESTAK, V.E., TARASOV, V.I., A new mechanistic
code SFPR for modeling of single fuel rod performance under various regimes of LWR operation, Nuclear Engineering
and Design 241 (2011) 2822-2830.
[14] VESHCHUNOV, M.S., PALAGIN, A.V., VOLCHEK, A.M., YAMSHCHIKOV, N.V., BOLDYREV, A.V.,
GALIMOV, R.R., KURCHATOV, S.YU., in Transactions of SMiRT-13 Conference, Vol.1 (1995) 159–163.
[15] HOFMANN, P., NOACK, V., VESHCHUNOV, M.S., BERDYSHEV, A.V., BOLDYREV, A.V., MATWEEV, L.V.,
PALAGIN, A.V., SHESTAK, V.E., Physico-Chemical Behavior of Zircaloy Fuel Rod Cladding Tubes During LWR
Severe Accident Reflood, Report FZKA 5846, Karlsruhe, Germany (1997).
[16] HOFMANN, P., MIASSOEDOV, A., STEINBOCK, L., STEINBRUECK, M., BERDYSHEV, A.V., BOLDYREV,
A.V., PALAGIN, A.V., SHESTAK, V.E., VESHCHUNOV, M.S., Quench Behavior of Zircaloy Fuel Rod Cladding
Tubes. Small-Scale Experiments and Modeling of the Quench Phenomena, Report FZKA 6208, INV-COBE (98)-D018,
Karlsruhe, Germany (1999).
[17] GORYACHEV, A.V., VESHCHUNOV, M.S., “Main results of the ISTC Project #1648.2”, 14th International QUENCH
Workshop, 2008, Karlsruhe, Germany.
205
[18] VESHCHUNOV, M.S., SHESTAK, V.E., Modeling of Zr alloy burst cladding internal oxidation and secondary
hydriding under LOCA conditions, Journal of Nuclear Materials, 461 (2015)129–142.
[19] STUCKERT, J., GROßE, M., RÖSSGER, C., KLIMENKOV, M., STEINBRÜCK, M., WALTER, M. QUENCH-LOCA
programme at KIT on secondary hydriding and results of the commissioning bundle test QUENCH-L0, Nuclear
Engineering and Design 255 (2013) 185–201.
[20] INTERNATIONAL ATOMIC ENERGY AGENCY, Analysis of Options and Experimental Examination of Fuels for
Water cooled Reactors with Increased Accident Tolerance (ACTOF), Report of a Coordinated Research Project 2015–
2019, IAEA-TECDOC (in press).
206
LIST OF PARTICIPANTS
Abe, A. Instituto de Pesquisas Energeticas e Nucleares – IPEN/CNEN – SP

Av. Prof. Mello Moraes 2231, 05508-030,
São Paulo, SP, Brazil,
Email:[email protected]
Boulore, A. CEA/DEN/DEC Cadarache,

13108 Saint-Paul-lez-Durance,
France
Chan, P. Royal Military College of Canada

Stn. Forces Kingston, ON K7K 7B4, PO Box 17000, Canada
Chen, M. China Nuclear Power Technology Research Institute

Yitian Road; 1312/A, Jiangsu Bldg., Shenzhen 518026 Futian, China
Gaikwad, A.J. Atomic Energy Regulatory Board

Mumbai, India
Giovedi, C. AMAZUL,
Av. Prof. Lineu Prestes 2468,
São Paulo, SP, Brazil.
Ieremenko, M. State Scientific and Technical Centre for Nuclear and Radiation Safety,
35–37, V. Stusa street, 03142, Kyiv, Ukraine
Khodadadi, H. Nuclear Reactors Fuel Company

Roshan Dasht, Isfahan, Islamic Republic of Iran
[email protected]
Kim, H.C. Nuclear Fuel Safety Research Division,

Republic of Korea
Lemes, M. Comisión Nacional de Energía Atómica,

207
Liu, S. Nuclear Power Institute of China,
Science and Technology on Reactor System Design Technology
Laboratory,
No. 328 Chang Shun Road, Hua Yang, Chengdu, Sichan, China
Liu, T. China Nuclear Power Technology Research Institute (CNPRI)

Loukusa, H. VTT Technical Research Centre of Finland Ltd.

Espoo, Finland
Lovasz, L. Gesellschaft für Anlagen und Reaktorsicherheit (GRS) gGmbH,

Boltzmann str. 14, 85748, Garching, Germany
Papp, H. Paks II. Ltd,

H-7030 Paks Gagarin u. l., Hungary,
Ren, Q. China Nuclear Power Technology Research Institute (CNPRI)

Ševeček, M. Czech Technical University in Prague,

Faculty of Nuclear Sciences and Physical Engineering
V.Holesovickach 2, Phaha 8, 180 00, Czech Republic
Shin, C. Nuclear Fuel Safety Research Division,

Republic of Korea
Soba, A. Comisión Nacional de Energía Atómica,

Stuckert, J. Karlsruhe Institute of Technology (KIT),

Germany
Tang, C. Institute for Applied Materials,

Karlsruhe Institute of Technology,
Germany
208
Vadi, R. Iran Nuclear Regulatory Authority
Atomic Energy Organization of Iran,
End of North Karegan Avenue, 14155–1339,
Tehran, Islamic Republic of Iran
Veshchunov, M. Nuclear Fuel Cycle and Materials Section

Vlahovic, L. European Commission,

Joint Research Centre,
Hermann-von-Helmholtz Platz 1, 76344, Eggenstein-Leopoldshafen,
Germany
Zhang, J. Tractebel (ENGIE),

Boulevard Simon Bolivar 34–36, 1000,
Brussels, Belgium
Zhou, Y. Nuclear Power Institute of China,

No. 328 Chang Shun Road, Hua Yang, Chengdu, Sichan, China
209
@ No. 26
ORDERING LOCALLY
IAEA priced publications may be purchased from the sources listed below or from major local booksellers.
Orders for unpriced publications should be made directly to the IAEA. The contact details are given at
the end of this list.
NORTH AMERICA
Bernan / Rowman & Littlefield
15250 NBN Way, Blue Ridge Summit, PA 17214, USA
Telephone: +1 800 462 6420 • Fax: +1 800 338 4550
Email: [email protected] • Web site: www.rowman.com/bernan
REST OF WORLD
Please contact your preferred local supplier, or our lead distributor:
Eurospan Group
Gray’s Inn House
127 Clerkenwell Road
London EC1R 5DB
United Kingdom
Trade orders and enquiries:

Telephone: +44 (0)176 760 4972 • Fax: +44 (0)176 760 1640
Individual orders:
www.eurospanbookstore.com/iaea
For further information:

Telephone: +44 (0)207 240 0856 • Fax: +44 (0)207 379 0609
Email: [email protected] • Web site: www.eurospangroup.com
Orders for both priced and unpriced publications may be addressed directly to:
Marketing and Sales Unit
Vienna International Centre, PO Box 100, 1400 Vienna, Austria
20-01339E
Telephone: +43 1 2600 22529 or 22530 • Fax: +43 1 26007 22529

Email: [email protected] • Web site: www.iaea.org/publications
Vienna
ISBN 978–92–0–108020–2
ISSN 1011–4289

TE-1913 Web

Uploaded by

Copyright:

Available Formats

TE-1913 Web

Uploaded by

Document Information

Original Description:

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

TE-1913 Web

Uploaded by

Copyright:

Available Formats

IAEA-TECDOC-1913

IAEA TECDOC SERIES

Modelling of Fuel Behaviour

IAEA SAFETY STANDARDS

AFGHANISTAN GERMANY PAKISTAN

MODELLING OF FUEL BEHAVIOUR

INTERNATIONAL ATOMIC ENERGY AGENCY

Marketing and Sales Unit, Publishing Section

For further information on this publication, please contact:

Nuclear Fuel Cycle and Materials Section

IAEA Library Cataloguing in Publication Data

Names: International Atomic Energy Agency.

Increased demands on fuel performance, including transient regimes, higher discharge

MAIN RESULTS OF THE CRP FUMAC (Session 1)

EXPERIMENTS ON FUEL BEHAVIOUR IN ACCIDENTAL CONDITIONS (Session 2)

FUEL PERFORMANCE CODES APPLIED TO DESIGN BASIS ACCIDENTS (DBA) AND TO

Reactivity Initiated Accident assessment for ATF cladding materials................................................................. 155

MODELLING OF TRADITIONAL UO2 AND ADVANCED ATF FUELS BEHAVIOUR IN INTEGRAL

Status of ATF modelling and application with ATHLET-CD ............................................................................. 183

List of Participants ............................................................................................................................................... 207

2. SUMMARY OF THE TECHNICAL SESSIONS

2.1. Technical Session 1: Main results of the CRP FUMAC

A total of six papers were presented in this session.

2.2. Technical session 2: Experiments on fuel behaviour in accidental conditions

Seven presentations were made during this session:

In this session, four presentations were made:

2.4.3.General conclusion and further recommendations

2.5.2. Summary of presentations

This session included three presentations.

2.5.3. General conclusions and further recommendations

3. CONCLUDING REMARKS AND RECOMMENDATIONS

M. LEMES, A. SOBA, A. DENIS

2. MODELS AND PARAMETERS INVOLVED

3.1. Simulation of IFA tests

Pressure (10 Pa)

IFA 650-9 PWR Gösgen NPP (Switzerland) 89.9

1400 ballooning burst scram a.

temperature (°C) 1000

0 100 200 300 400 500

1100 ballooning burst scram

0 100 200 300 400 500

FIG. 5. Idem Fig. 4. but for IFA 650-10.

ballonning burst scram a.

FIG. 6. Idem Fig. 4 but for IFA 650-11.

TABLE 2. COMPARISON OF THE PREDICTIONS OF SOCRAT AND DIONISIO WITH

IFA 650-9 IFA 650-10 IFA 650-11

3.2. Simulations of PMK tests

cladding temperature (K)

Coolant flow rate (Kg/s)

pressure (10 Pa)

3.3. Simulation of QUENCH experiments

220 240 260 280

temperature (K) 1000

0,02 Experimental 0,4

J. ZHANG, A. DETHIOUX, T. DRIEU, C. SCHNEIDESCH

3. SIMULATION OF THE HALDEN LOCA TESTS IFA-650.9 AND IFA650.10

3.1. Input models and assumptions

TABLE 1. PARAMETRIC STUDIES ON FRAPTRAN-TE-1.5 MODEL OPTIONS