Faire avancer la sûreté nucléaire

La Recherchev2


Annals of Nuclear Energy - Numéro spécial Ermsar 2013 sur Sarnet



Email :

Mot de passe :

​Annals of Nuclear Energy / Volume 74, décembre 2014

Type de document > *Article de revue

Mots clés >

Unité de recherche > IRSN/PSN-RES

Auteurs > VAN DORSSELAERE Jean-Pierre, [et al.]

Date de publication > 01/12/2014


Despite the highly efficient accident prevention measures adopted for the current Generation II nuclear power plants (NPP) and the still more demanding ones for the Generation III ones, some accident scenarios with a low probability may result in severe accidents, as recently emphasized by the Fukushima-Daiichi accidents in Japan. This can result in core melting, plant damage and dispersal of radioactive materials outside of the plant containment, thus threatening public health and the environment.

The ERMSAR 2013 conference (European Review Meeting on Severe Accident Research) was hosted by the French Institut de Radioprotection et de Sûreté Nucléaire (IRSN) in the Palais des Papes in Avignon (France) from 2 to 4 October 2013. It was organized under the auspices of the SARNET (Severe Accident Research NETwork of excellence) network that is coordinated by IRSN. This conference concluded a cycle of life of twice 4 years of the network in the frame of the 6th and 7th Framework Programmes (FP6 and FP7 respectively) of Research and Development of the European Commission (EC) that co-funded the network activities. ERMSAR 2013 gathered 137 participants (including about 30 from outside Europe) from 61 organizations and 25 countries, confirming its status as a leading international event on major nuclear reactor accidents. Fifty-two papers and eleven posters were presented and significant time was allocated after each presentation for questions and open discussions.

Forty-four organisations (research centres, universities, industries, utilities, safety authorities and Technical Safety Organizations) – from 22 countries worldwide – have networked since 2004 their capacities of research in SARNET to resolve the most important open severe accident issues in existing and future Gen. II–III NPPs. The main objective was to optimize the use of the available research budgets and to constitute a sustainable consortium in which common research programmes would be defined and performed. The network partners were, from 2008: Tractebel (Belgium), INRNE and Technical University of Sofia (Bulgaria), UJV (Czech Rep.), VTT (Finland), AREVA NP SAS, CEA, EDF, IRSN and Thermodata (France), AREVA GmbH, University of Bochum, GRS, FZ Jülich, KIT and University of Stuttgart (Germany), Demokritos (Greece), MTA-EK, NUBIKI and University of Budapest (Hungary), ENEA, RSE and University of Pisa (Italy), LEI (Lithuania), NRG (the Netherlands), INR (Romania), CIEMAT (Spain), IVS, UJD SR and VUJE (Slovakia), JSI (Slovenia), Chalmers University and University of Stockholm (or KTH) (Sweden), PSI (Switzerland), NNL, University of Newcastle and Imperial College of London (UK), JRC/IET and ITU (EC), AECL (Canada), USNRC (USA), KAERI and KINS (Korea) and BARC (India). Contacts were taken with JAEA and JNES from Japan in the last two years to start a further collaboration.

The overall work carried out by the network members, involving more than 230 researchers and 30 PhD students, represents an equivalence of about 40 full-time persons per year. Most key European R&D actors in the severe accident area are members and the network takes also benefit from the knowledge and contribution of important non-European organizations from USA, Canada, Korea and India.

The outcomes of the FP7 project are very positive:
  • Good efficiency of networking and sharing of knowledge among the partners,
  • Significant progress of knowledge on the high-priority severe accident issues through collaborative work (analysis of experiments, benchmarks among numerical simulation codes, state-of-the-art reports) on corium, containment and source term phenomena,
  • Capitalisation of knowledge in the ASTEC IRSN-GRS integral code that simulates whole severe accident scenarios and in the DATANET database for storage of experimental data (using the JRC STRESA tool),
  • Dissemination of knowledge1 through education courses for young researchers and students or managers and senior scientists, edition of a textbook on severe accident, mobility of researchers and ERMSAR periodic conferences.

The ERMSAR-2013 conference was the opportunity to present the main outcomes of R&D (i.e. mainly new knowledge and new physical models for computer codes) from 2008 to 2013 on the five following high-priority topics: corium and debris coolability, molten corium concrete interaction (MCCI), hydrogen mixing and combustion in containment, source term, and ASTEC integral code. All topics included both experimental and modelling activities, and assessment of the applicability of simulation codes to plant accident scenarios. This special issue of Annals of Nuclear Energy includes a selection of 20 ERMSAR 2013 joint synthesis papers.

An essential and continuous task of the network was the ranking of R&D priorities in order to define common future severe accident research programs. The Severe Accident Research Priority (SARP) group has updated the R&D priorities on the basis of the SARNET outcomes and on the recent international R&D (see the paper “Conclusions on severe accident research priorities”, by W. Klein-Heßling et al.). In particular it took into account OECD/NEA/CSNI projects, the Level 2 Probabilistic Safety Analysis (PSA) studies and the ASAMPSA2 FP7 project on PSA2 best-practice guidelines that ended in 2013. The group also accounted for the preliminary understanding of the Fukushima-Daiichi accidents, which resulted only in slight reorientations of priorities because most involved phenomena were already addressed as being of highest priority.

Topic "Corium and Debris Coolability"

The R&D aims at reducing the remaining uncertainties on the possibility of cooling the reactor core structures and materials during a severe accident, either in the core region or in the vessel lower head or in the reactor cavity after vessel failure, so as to limit the progression of the accident. This could be achieved either by ensuring the retention of corium within the reactor vessel by water injection or at least a slow corium progression and small flow rates of corium release into the cavity. These issues are linked to severe accident management (SAM) for current reactors and also for design and safety evaluation of future reactors.

The specific objectives were to enhance the database on the remelting of debris and corium pool formation in the lower head (relying on the LIVE experiments with simulant materials at KIT), to develop and validate the models and computer codes for simulation of in-vessel melt pool behaviour, to perform reactor-scale analysis for in-vessel corium coolability and to assess the influence of SAM measures on in-vessel coolability. Work comprised experimental and modelling activities, with strong cross-coupling between the tasks.

Substantial knowledge exists now concerning cooling of fuel in large intact rod-like geometries, but efforts had to be made to address modelling of air ingress into an overheated core following an earlier partial oxidation in steam (see the paper “SARNET2 Benchmark on Air Ingress Experiments QUENCH-10, -16” by L. Fernandez-Moguel et al.).

The main experimental efforts concentrated on debris bed formation and cooling by water to demonstrate effective modes and establish cooling rates and limits (see the paper “Evaluation of an effective diameter to study quenching and dry-out of complex debris bed” by N. Chikhi et al.). Several experiments were performed in the following facilities: DEBRIS at the University of Stuttgart for top and bottom flooding of hot debris, POMECO at the University of Stockholm for boil-off conditions with emphasis on basic laws and specific 2D effect, and PRELUDE at IRSN for the effect of power distribution inside a larger debris bed with multi-dimensional effects (see the paper “Experimental investigation on reflooding of debris beds” by S. Leininger et al.). Other facilities were more adapted to ex-vessel debris beds such as COOLOCE in VTT and DEFOR in KTH (see the paper “Analyses on Ex-Vessel Debris Formation and Coolability in SARNET frame” by G. Pohlner et al.).

In parallel, efforts have aimed at assessing and validating the models in system-level and in detailed codes for core degradation, oxidation and debris behaviour, before improving these models. Modelling of porous media has indicated a strongly favoured coolability by inflow of water from lateral water-filled regions of the core with higher porosities.

In view of the vulnerability of nuclear fuels that are stored before their evacuation and final disposal or possible reprocessing, as highlighted by the Fukushima-Daiichi accidents, work also addressed the accidents in spent fuel pools of different geometries (see the paper “Syntheses of spent fuel pool accident assessments using Severe Accident codes” by J. Fleurot et al.).

Topic "Molten Corium Concrete Interaction"

The addressed situation is, after a postulated vessel lower head failure, the corium presence in a reactor pit which is initially dry but with the possibility of water injection later during MCCI. A major question was why an anisotropic ablation of basemat was observed with siliceous concretes and an isotropic one with carbonaceous concretes. The work was based on the analysis of recent 2D real material MCCI-OECD (in Argonne National Labs. in USA) and VULCANO (in CEA) experiments in an oxidic pool and of analytical experiments with simulant materials on 2D heat convection in a bubbling pool (see the paper “Towards an European consensus on possible causes of MCCI ablation anisotropy in an oxidic pool” by M. Cranga et al.). The scaling issue was also investigated to support reliable plant calculations (see the paper “Transposition of 2D Molten Corium Concrete Interactions from experiment to reactor” by C. Spengler et al.).

Another potential factor that might strongly affect MCCI is the possible metal/oxide stratification and the spatial distribution of metallic and oxide phases. Experiments provided new insights at the large-scale MOCKA facility (in KIT) (∼1 ton of corium) with simulant materials and at the VULCANO (in CEA) and the SICOPS (in AREVA GmbH) facilities with prototypic materials (see the paper “Experiments on MCCI with oxide and steel” by J.J. Foit et al.).

As for corium coolability, recent data suggest that early water flooding (i.e. at a time when the concrete molten fraction is low) has a good coolability potential but this must be confirmed by further experiments. The efficiency of late top flooding seems more doubtful, especially for siliceous concretes. Bottom flooding is being studied as an efficient alternative to top flooding.

Several benchmarks were performed among simulation codes, either on VULCANO and COMET experiments (the latter in KIT) or on VVER-1000 reactor MCCI scenarios, in order to evaluate the code capabilities.

The corium composition and properties determine the interactions both with the reactor vessel and in the later phases with the concrete basemat. Analytical and larger scale tests supplied new data on melt-liquidus composition for the NUCLEA thermodynamic database (see the paper “Quality Improvements of Thermodynamic Data applied to Corium Interactions for Severe Accident Modelling in SARNET2” by P.D.W. Bottomley et al.).

Topic "Containment Issues"

The focus is on the threat to containment integrity caused by highly energetic phenomena, particularly steam explosion and hydrogen combustion.
Steam explosion may be caused by fuel–coolant-interaction after corium pouring into the flooded reactor cavity. The work, aiming at completing the SERENA2 OECD/NEA/CSNI project, was based on experiments in the facilities KROTOS (CEA), TROI (KAERI), MISTEE and DEFOR/PULiMS (KTH), and DISCO-FCI (KIT) and on the simulation codes MC3D (IRSN) and JEMI-IDEMO (IKE) (see the paper “Status of steam explosion understanding and modelling” by R. Meignen et al.).

Hydrogen combustion (deflagration and detonation) may be caused by ignition of a gas mixture with high local hydrogen concentrations, which may be due to the imperfect mixing of the containment atmosphere. Benchmarks have been performed to assess the effect of specific phenomena and/or safety systems with the potential to affect the containment atmosphere behaviour and to evaluate the code capabilities and improvements:
  • On containment sprays (see the paper “Achievements of spray activities in nuclear reactor containments during the last decade” by J. Malet et al.),
  • On hydrogen combustion (see the paper “SARNET hydrogen deflagration benchmarks – main outcomes and conclusions” by A. Bentaib et al.),
  • On steam condensation on walls (see the paper “Lesson learned from the SARNET Wall Condensation Benchmarks” by W. Ambrosini et al.),
  • On passive autocatalytic recombiners.

Additionally, a theoretical benchmark based on a generic NPP containment allowed an in-depth comparison of thermal–hydraulic models in Lumped-Parameter codes (see the paper “Generic containment: Detailed comparison of containment simulations performed on plant scale” by S. Kelm et al.). This led to recommend that results of a single simulation of a transient in a NPP containment with a lumped-parameter code cannot be trusted before at least an additional independent simulation (preferably with a different code) of the same transient.

Topic "Source Term"

The activities focused on reducing the uncertainties associated to estimates of radioactive releases to the NPP external environment. Two major issues were investigated: the impact of oxidising conditions on source term and the iodine chemistry within the reactor coolant system (RCS) and containment.

The main emphasis was placed on iodine and ruthenium, given their high radio-toxicity and volatility. The transport and chemical transformation of these two elements through the RCS and inside the containment were investigated (see the paper “Transport of ruthenium in primary circuit conditions during a severe NPP accident” by T. Kärkelä et al.). Of particular importance is the prediction of volatile iodine and ruthenium species in the containment atmosphere since they would be only partly removed by containment sprays or by filtration when the venting of the containment system is actuated. Full advantage was taken of cooperation with international programmes such as Phébus FP, ISTP (International Source Term Programme) and OECD/NEA/CSNI projects (i.e. STEM, BIP, BIP2 and THAI). Benchmarks were organised on the basis of the Phébus FPT3 (organised by IRSN) and THAI experiments (organized by GRS).

Interactions between iodine and containment paint (e.g. adsorption) have been known to likely happen in case of a severe accident. However, recent insights concerning chemical mechanisms seem to be the key for a comprehensive understanding of organic iodides formation and, as a consequence, for an enhancement of predictability of gas phase iodine concentration in the longer term (see the paper “Iodine-Paint Interactions during nuclear reactor severe accidents” by L. Bosland et al.).
Nonetheless, there are other processes involved in the formation of gaseous iodine. The experiments on iodine oxidation demonstrated also that gaseous iodine species such as molecular iodine and volatile organic iodide can reversibly react with the products of air radiolysis to form iodine oxide particles (see the paper “Experimental and modelling studies of iodine oxide formation and aerosol behaviour relevant to nuclear reactor accidents” by S. Dickinson et al.).

Chemical revaporisation or physical resuspension of fission product deposits from the primary circuit might be a major source term in the late phase of severe fuel degradation. The work conducted on Phébus FP post-test analysis at JRC/ITU, EXSI-PC experiments at VTT, and CHIP experiments at IRSN support this statement (see the paper “Revaporisation of fission product deposits in the primary circuit and its impact on accident source term” by P.D.W. Bottomley et al.).

Topic "ASTEC code"

The ASTEC integral simulation code, jointly developed by IRSN and GRS, plays a key role in the network by capitalizing severe accident knowledge through the implementation of the new physical models that are being developed in each above Topic. The code assessment was done by partners through comparison with more than 50 experiments and applications to severe accident scenarios in diverse NPP types (PWR, VVER, CANDU…), as well as benchmarks with other codes (see the paper “Synthesis of the ASTEC integral code activities in SARNET. Focus on ASTEC V2 plant applications” by P. Chatelard et al.). The current development efforts for the next major version ASTEC V2.1, planned end of 2014, address the reflooding of degraded cores and the adaptation of core degradation models to BWR and CANDU.

Conclusions and SARNET perspectives

The ERMSAR periodic conference organized by the network is becoming the major worldwide conference on severe accident research. ERMSAR 2013 was a cornerstone since the FP7 SARNET2 project ended in March 2013, and the SARNET network is now fully integrated in the new international NUGENIA association ( on R&D on Gen. II and III NPPs. The NUGENIA Technical Area N°2, entitled “Severe accidents”, covers the SARNET scope and extends to the issues of emergency preparedness and response, as well to the impact of severe accidents on the environment. In particular, the SARP research priorities have been totally used to build the NUGENIA roadmap for severe accident R&D in the next years. The next ERMSAR conference will be hosted by CEA in Marseille (France) on 24–26 March 2015.

All the participants to the network must be acknowledged, and in particular the members of the Management Team (besides the guest editors below, A. Auvinen, VTT, D. Beraha, GRS, P. Chatelard, IRSN, C. Journeau, CEA, I. Kljenak, JSI, A. Miassoedov and W. Tromm, KIT, and R. Zeyen, JRC-IET). The EC must also be acknowledged for the co-funding of the FP6 and FP7 projects, and particularly Michel Hugon, EC coordinating officer for SARNET, for his very strong interest and support. Special thanks also to the IRSN staff in Cadarache and the Avignon Tourism Office for the practical organization of the conference. Last but not least, this special issue would not have been possible without the help of the Elsevier staff.