Complete simulation of severe accidents
The purpose of the ASTEC software package (Accident Source Term Evaluation Code) is to simulate all the phenomena that occur during a severe accident in a water-cooled nuclear reactor, from the initiating event to the possible release of radioactive products (the 'source term') outside the containment. ASTEC has been developed jointly over a number of years by the IRSN and its German counterpart, the Gesellschaft für Anlagen und Reaktorsicherheit mbH (GRS).
The main applications of the software package are safety analyses for nuclear reactors (e.g. the European Pressurised Reactor - EPR), source term evaluations (e.g. the re-evaluation of the S3 source term for the French pressurised water reactors - PWR), and the development of severe accident management guidelines. ASTEC is widely used in IRSN level 2 probabilistic safety assessments (PSA2) for 900 and 1300 MWe PWR and for EPR. It is also used in the preparation and interpretation of experimental programmes, in particular the Phébus FP integral test programme (now finished) and in the tests carried out as part of the International Source Term Program (ISTP).
ASTEC is the European reference software within the European Commission SARNET network of excellence. It is also used by organisations in Canada, Russia, India, Korea and China.
ASTEC covers the entire phenomenology of severe accidents except steam explosion (for which the IRSN uses the MC3D software) and the mechanical integrity of the containment (for which the IRSN uses the CEA CAST3M software package). Its modular structure (see Figure 1) simplifies qualification by comparing the simulated results with those obtained experimentally.
Each module simulates the phenomena occurring in one part of the reactor or at one stage of the accident. These include:
The two-phase thermal-hydraulics of coolant flows in the reactor coolant primary and secondary systems using a numerical approach based on five equations (see Figure 2);
The degradation of the core geometry as the residual power causes a temperature rise resulting in chemical reactions between the constituent materials and even their melting, up to the formation of a mixture of molten materials named "corium". A dynamic management approach (appearance, disappearance, transformation, relocation) of the various components is used within a control volume of the core;
The release of fission products (FP), particularly iodine, from fuel in the core, together with their transport and chemical behaviour in the reactor coolant system and subsequently within the containment.
The thermal-hydraulics and transport of aerosols within the containment using a lumped-parameter or 0D volumes approach (see Figure 3).
The molten corium-concrete interaction (MCCI) in the reactor cavity following rupture of the reactor vessel, using a 0D volumes or layers approach (see Figure 4)
ASTEC also simulates other phenomena, associated with the accident, including direct containment heating (DCH) by the transfer of hot gases and corium droplets from the reactor cavity, following the rupture of the vessel; the combustion of hydrogen accumulated within the containment and the associated risk of explosion; and the radioactivity of the isotopes and the associated residual power in all parts of the reactor.
The most recent version of ASTEC (V2.0rev2), delivered to all the partners at the end of 2011, contains models that incorporate the latest state-of-the-art knowledge. The models of fission product behaviour in particular have been based largely on knowledge derived from the results of the Phébus FP experiments. The scientific excellence of these models is one of the major advantages of ASTEC compared with other international software packages. New models have been developed for representing the EPR core-catcher with its spreading features.
The core degradation models, an essential component of the ASTEC package, have been considerably improved in the new V2 family of versions. These models are the ones of the ICARE2 mechanistic software resulting from a major IRSN investment dating from the early 1990s (see Figure 5). Similarly, the containment models (thermal-hydraulics and aerosols) are based on work by the GRS in the development of the earlier RALOC and FIPLOC softwares and the more recent COCOSYS software.
In common with the majority of softwares available worldwide, the model relating to the reflooding of degraded cores is still considered to be inadequate and R&D work on this model is continuing within the the SARNET network.
The ASTEC package has been validated by over 160 tests, including:
Analytical tests, either separate-effect or coupled-effect tests. Examples include the VERCORS tests (CEA) relating to fission product release and transport, BETA (KIT –former FZK- Germany) concerning the molten corium-concrete interaction, and the ACE-RTF tests (Canada) on iodine behaviour in the containment;
Integral tests, including the Phébus FB
in-pile tests (IRSN) simulating an entire accident with real materials up to the source term in the containment, and the out-of-pile CORA and QUENCH tests (KIT) representing a bundle of electrically heated fuel rods in the core, using simulant materials.
Among these 160 tests, the OECD/NEA ISP exercises (International Standard Problem) have been selected as often as possible as these constitute an international reference by virtue of the high quality of the measurements and their use as benchmarks when comparing software packages (PACTEL, VANAM, BETHSY, LOFT tests, etc.). The test matrix is continually being expanded by the results of international programmes, including: CCI-OECD (Argonne National Laboratories, USA), ISTP (EPICUR and CHIP at the IRSN), ThAI-OECD (Becker Technologies, Germany), and ARTIST (PSI, Switzerland).
The software is also regularly tested against the Three Mile Island reactor accident in the USA (TMI-2) in 1979 with the aim of consolidating the results prior to application to actual reactor configurations.
The Figures 6 to 10 illustrate five of the numerous validation results; note that these calculation examples have been selected in a way to cover most aspects of severe accident phenomenology, i.e. to cover both in-vessel and ex-vessel processes, in order to provide a good picture of the current ASTEC V2 capabilities:
The LOFT-LP-FP2 fuel rod bundle degradation test carried out by the Idaho National Laboratories in the USA. The red diamonds indicate the measured temperatures and the almost superimposed blue, red and green curves show the ASTEC results for three different fuel rods. The simulated and measured results are in good agreement (see Figure 6)
The Phébus FPT4 debris bed test carried out by IRSN. The symbols indicate the measured temperatures at four different elevations in the bed centreline and the straight lines refer to the corresponding ASTEC results. The simulated and measured results are in very good agreement (see Figure 7)
The STORM SR aerosol transport and deposition tests carried out by the JRC/ISPRA in Italy. Four of the tests SR09 to SR12 (horizontal axis) are shown. The blue columns represent the fraction of the aerosols deposited in the experimental system. The other colours represent the results of the ASTEC simulations using three different models. These models encompass the measured results but improvements to the models are needed (see Figure 8)
The Marviken blowdown experiment carried out in the Marviken Boiling Heavy Water Reactor nuclear power plant (located in Sweden). The symbols indicate the measured pressure in the containment discharge room while the straight lines refer to the corresponding ASTEC results for two successive code versions. The simulated and measured peak pressure are in very good agreement (see Figure 9)
The OECD-CCI2 corium-concrete interaction tests. The red curve represents the mass of concrete ablated as a function of time. The blue curve represents the corresponding ASTEC simulation. The simulated and measured results are in good agreement (see Figure 10)
The high level of the software is evident from the validation work carried out by the IRSN, the GRS and their international partners.
The ability of the software to provide accurate simulations of any accident scenario involving currently operating reactors has been demonstrated by the level 2 PSA studies carried out by the IRSN for 900 and 1300 MWe PWRs, and by a large number of benchmark tests in comparison with other software (see the Section on International Collaboration below). ASTEC can provide good simulations of the vast majority of safety systems and the actions or procedures followed by the operators of current reactors, including the depressurisation of the reactor coolant system and use of the containment spray system. The trends and orders of magnitude of the results of these sequence simulations are generally in agreement with those from the MELCOR and MAAP4 software packages widely used throughout the world (see Figure 11).
Code-to-code comparisons are also regularly made with mechanistic software packages such as CATHARE (French reference code for circuit thermal-hydraulics) for thermal-hydraulics phenomena before core degradation (see Figure 12) and ICARE/CATHARE (IRSN) or ATHLET-CD (GRS) for emphasis on core degradation processes (see Figures 13 and 14).
About thirty European member organisations of the SARNET network are evaluating ASTEC, either through validation against the results of experimental programmes (see above), or by means of benchmark comparisons with other software packages for accident scenarios in various types of reactor (900 MWe PWR, Konvoi 1300, Westinghouse 1000, VVER 440 and VVER 1000). For many years, the IRSN has also been working in close collaboration on similar work with a number of organisations outside Europe, including the Kurchatov Institute (Russia), Atomic Energy of Canada Limited (AECL, Canada), Bhabba Atomic Research Centre (BARC, India) and Atomic Energy Regulatory Board (AERB, Indian safety authority).
The work done in SARNET has shown that ASTEC V2 is today able to simulate a large part of severe accident scenarios in boiling water reactors (BWR) and heavy water CANDU reactors, except for the stage of core degradation. Evaluations of the behaviour of fission products and aerosols in the primary circuit and in the containment of these reactors are already possible.
Maintenance and user support
With almost a hundred users of ASTEC throughout the world, several engineers are employed to provide an efficient user support ensuring a rapid response to user requests. Users can easily download any new code version or revision as well as any update of the code documentation through the ASTEC web portal. Besides, a dedicated internet system, MARCUS, has been established to manage the exchange of information between the developers of the software and the users, particularly bug reports and fixes (see Figure 15).
The ASTEC users' club meets about every 18 months bringing together the various organisations involved. These meetings provide an ideal forum to discuss users’ experience of the software, their requirements for future software evolutions, and the development plans proposed by the IRSN and GRS.
Future developments of ASTEC
Improvements of the ASTEC models continue based on the results of international R&D work, especially within the SARNET network, the ISTP programme and the OECD projects. In particular, to answer one of the main safety analysts requirements, significant efforts are currently paid at IRSN to the development of a dedicated model to deal with the reflooding of degraded cores, taking full benefit of the on-going PRELUDE and PEARL IRSN experimental programmes. Other priorities include corium-concrete interactions in the reactor cavity, and gas chemistry kinetics in the reactor coolant system with a particular emphasis on iodine and ruthenium behaviour.
The validation of ASTEC will continue intensively on the OECD exercises, together with other current international programmes (MCCI-OECD, ThAI, ARTIST, etc.) and IRSN programmes (PEARL, ISTP and STEM-OECD using the facilities EPICUR and CHIP relating to fission products).
In 2012-13, the core degradation models will be developed for BWR and CANDU reactors. First ASTEC applications to the accidents in the Fukushima-Daiichi Japanese reactors are also foreseen in parallel.
Work has also been performed for use of parts of the software package in emergency response tools. IT must be underlined that, beyond the continuous improvement of the relevance of physical models that remains an essential objective of IRSN, the other objective is to extend the efforts of numerical optimization and computing time reduction that have started. An extension to sevre accident simulators could be planned in the future.
Beyond the numerous code applications to Gen.II and Gen.III reactors, some model have been adapted to the conditions of accidental scenarios in nuclear fusion installations like ITER. Efforts will continue on this topic in the next years.
Longer term objectives include modifications to suit Generation IV reactors, especially the SFR (Sodium-cooled Fast Reactor).