•Description of the DRACCAR multi-physics code dedicated to analysis of LOCA in LWR.•Computational analysis of multi-rod ballooning, contact and fuel relocation.•Details on the associated ...thermo-mechanics and thermalhydraulics coupling.•Dedicated modeling to simulate reflooding with channel blockage during LOCA.
Computational predictions concerning ballooning of multiple fuel pin bundles during a loss of coolant accident with a final reflooding phase are now more than ever of interest in the framework of light water reactor nuclear safety. To carry out these studies, two difficulties have to be overcome. First, the modeling has to take into account many coupled phenomena such as heat transfer (heat generation, radiation, convection and conduction), hydraulics (multidimensional 2-phase flow, blockage), mechanics (thermal expansion, creep, embrittlement) and chemistry (oxidation, hydriding). Secondly, there are only a few experimental investigations that can help to validate such complex coupled modeling. Over several years, IRSN has developed the 3D computational tool DRACCAR to investigate rod bundle strain during LOCA transients including prediction of the reflooding phase. DRACCAR code is dedicated to study complex configurations such as the deformation and possible contact between neighboring rods and the associated blockage of thermalhydraulic channels in the ballooned zone of the fuel assembly. Modeling efforts have been devoted to the assessment of the coolability of deformed geometries by coupling the thermo-mechanical behavior of the fuel assembly to the thermalhydraulics. The physical modeling available in the current version of DRACCAR V2.3.1 as well as its flexibility are depicted. As a conclusion, some prospects regarding the development of the future version DRACCAR V3 are provided, in particular accounting for the knowledge acquired through IRSN R&D project PERFROI.
•Validation status of DRACCAR, a multi-physics code dedicated to LOCA in LWR.•Computational analysis of multi-rod ballooning, contact and fuel relocation.•Details on the associated thermo-mechanics ...and thermalhydraulics validation.•Assessment of the modeling capabilities to simulate reflooding and channel blockage.
Computational predictions concerning ballooning of multiple fuel pin bundles during a loss-of-coolant accident with a final reflooding phase are now more than ever of interest in the framework of light water reactor nuclear safety. To carry out these studies, two difficulties have to be overcome. First, the modeling has to take into account many coupled phenomena such as heat transfer (heat generation, radiation, convection and conduction), hydraulics (multidimensional 2-phase flow, blockage), mechanics (thermal expansion, creep, embrittlement) and chemistry (oxidation, hydriding). Secondly, there are only a few experimental investigations that can help to validate such complex coupled modeling. Over several years, IRSN has developed the 3D computational tool DRACCAR to investigate rod bundle strain during LOCA transients including prediction of the reflooding phase. The DRACCAR code is dedicated to study complex configurations such as the deformation and possible contact between neighboring rods and the associated blockage of thermalhydraulic channels in the ballooned zone of the fuel assembly. To accompany the development of DRACCAR, efforts have been devoted to the validation of the coupling between the thermo-mechanics and thermalhydraulic models – including reflooding – through a comparison to integral experiments dedicated to LOCA. The DRACCAR capabilities and validation status are depicted for the version DRACCAR V2.3.1. DRACCAR provides an interesting insight on LOCA by simulating multi-rod and fluid interaction which cannot be investigated with a classical single rod approach. As a conclusion, some prospects regarding the development and validation of the future version DRACCAR V3 are mentioned. In particular significant evolutions are expected regarding the cladding rupture prediction, the contact simulation and the assessment of the coolability of deformed geometries. These evolutions will be based on the knowledge acquired through the R&D project PERFROI, a project dedicated to LOCA, launched by IRSN in association to other partners and supported by the French National Research Agency (ANR).
•An OECD/NEA PWR spent fuel pool experimental program simulated with DRACCAR v2.3.•New models developed for radiative HT in the bundle and for Zr nitrides production.•Air flow rates well calculated ...in cells with the hypothesis of oxides swelling.•Clad temperature response for a 5 kW assembly in a uniform pattern is well assessed.•1 × 4 pattern bundles still depending on unchecked hypotheses.
This paper describes simulations of two ignition tests performed at full power in the frame of the Sandia Fuel Project (SFP) with the thermo-mechanical code DRACCAR v2.3.
The OECD/NEA Sandia Fuel Project was built on an agreement between 12 countries from OECD, the Nuclear Energy Agency (NEA) and the US-NRC for the characterization of thermal-hydraulic and zirconium fire phenomena in pressurized-water reactor (PWR). The experimental program was split in two phases to focus at first on axial heating and burn propagation in one prototypic fuel assembly (Phase I), and then on axial and radial heating and burn propagation in 1 × 4 fuel assemblies (Phase II).
DRACCAR is a simulation tool, developed at IRSN, for fuel assembly mechanical behavior and coolability assessment during a LOCA transient. The flexibility of DRACCAR allows the modeling of many kinds of geometries. Because the code is based on a 3D non-structured meshing, it can be used to model any non-axisymmetric geometry, like the 1 × 4 fuel assemblies geometry of the Phase II of the SFP program.
In order to check the consistency of the modeling, we have optimized the code options to get best results for the Phase I, and applied the same options to the Phase II. Most of the DRACCAR results have been successfully checked against experimental ones, using additional code improvements. Air oxidation and breakaway modeling of the zircaloy claddings were successfully tested against the experimental results. Nevertheless parts of the experimental results of Phase II have been difficult to reproduce. As many causes could be involved in these difficulties, such as detailed evolution of the air convective loop, radiative heat transfers in the bundles, and the modeling of additional reactions of zirconium with nitrogen in places where oxygen is lacking, there is still room for improvement in the work of interpretation and modeling of the SFP tests.
•Recent advances in the development and validation of the DRACCAR code are presented.•Experimental programs dealing with reflooding of an intact or ballooned bundle are simulated.•Ways of improvement ...have been identified and are in progress such as a new reflooding model and a 6 equation version of the thermal–hydraulics code.•Spent-fuel-pool draining accidents are addressed and the modeling flexibility of the DRACCAR code to model non axis-symmetric systems is emphasized.
To meet the simulation needs of its LOCA R&D program, the IRSN is developing a multi-rod computational tool named DRACCAR. In order to realistically describe the behavior of the reactor core during a Loss Of Coolant Accident (LOCA), modeling has to take into account many coupled phenomena such as thermics (heat generation, radiation, convection and conduction), hydraulics (multi dimensional 1–3 phase flow, shrinkage), mechanics (thermal dilatation, creep, embrittlement) and chemistry (oxidation, oxygen diffusion, hydriding,...). This paper presents several aspects of the DRACCAR code abilities: first to handle thermal–hydraulics during reflooding of an intact and of a partially ballooned bundle and secondly the simulation of the OECD SFP phase II experiment dealing with the instantaneous draining of a spent fuel pool.
Under limited core damage accidents (LCDAs) of Pressurized Heavy Water Reactor (PHWR), coolable geometry of the channel might be retained thanks to the presence of moderator heat sink. Indeed, the ...pressure tube is amenable to creep deformation at high temperature due to internal pressure and fuel bundles weight. Partial or complete circumferential contact between pressure tube and calandria tube aids heat dissipation to the moderator. A new module has been developed by Bhabha Atomic Research Centre (BARC) for simulating this phenomenon which is specific to horizontal-type of reactors. It requires additional calculation of pressure tube sagging/ballooning and temperature field in the circumferential direction. The module is well validated with available experimental results concerning pressure tube deformation and the associated heat transfer in the area of contact. It is then used in analysing typical LCDAs scenarios in Indian PHWR under low and medium internal pressure conditions. This module is implemented in the ASTEC IRSN-GRS severe accident code version under development and will thus be available in the next major version V2.1.
This paper summarizes the work done in the SARNET European Network of Excellence on Severe Accidents (6th Framework Programme of the European Commission) on the capability of the ASTEC code to ...simulate in-vessel corium retention (IVR). This code, jointly developed by the French Institut de Radioprotection et de Sûreté Nucléaire (IRSN) and the German Gesellschaft für Anlagen und Reaktorsicherheit mbH (GRS) for simulation of severe accidents, is now considered as the European reference simulation tool.
First, the DIVA module of ASTEC code is briefly introduced. This module treats the core degradation and corium thermal behaviour, when relocated in the reactor lower head. Former ASTEC V1.2 version assumed a predefined stratified molten pool configuration with a metallic layer on the top of the volumetrically heated oxide pool. In order to reflect the results of the MASCA project, improved models that enable modelling of more general corium pool configurations were implemented by the CEA (France) into the DIVA module of the ASTEC V1.3 code.
In parallel, the CEA was working on ASTEC modelling of the external reactor vessel cooling (ERVC). The capability of the ASTEC CESAR circuit thermal-hydraulics to simulate the ERVC was tested. The conclusions were that the CESAR module is capable of simulating this system although some numerical and physical instabilities can occur. Developments were then made on the coupling between both DIVA and CESAR modules in close collaboration with IRSN. In specific conditions, code oscillations remain and an analysis was made to reduce the numerical part of these oscillations. A comparison of CESAR results of the SULTAN experiments (CEA) showed an agreement on the pressure differences.
The ASTEC V1.2 code version was applied to IVR simulation for VVER-440/V213 reactors assuming defined corium mass, composition and decay heat. The external cooling of reactor wall was simulated by applying imposed coolant temperature and heat transfer coefficient (HTC). The obtained results (pool temperatures, heat flux distribution, reactor wall ablation) were compared with available predictions of other codes. The agreement was correct, in particular on the shape and depth of ablation, as well as the maximum heat flux in case of a thick metallic layer, while ASTEC calculated a lower maximum heat flux for a thin metallic layer.
The French Institut de Radioprotection et de Sûreté Nucléaire (IRSN) and the German Gesellschaft für Anlagen und Reaktorsicherheit mbH (GRS) have been jointly developing for several years a system of ...calculation codes (or "integral" code), ASTEC (Accident Source Term Evaluation Code), to simulate the complete scenario of a hypothetical severe accident in a nuclear light water reactor from the initiating event through the possible radiological release of fission products out of the containment, the so-called "source term." Very intensive validation work has been performed in recent years by IRSN and GRS on the V1 versions by comparison of code calculations with results of more than 160 international experiments. Complementary validation was performed by 30 partners of the SARNET European Network of Excellence in the 6th Framework Programme of the European Commission, where ASTEC is considered the European reference code. The global status of validation is good for most phenomena, as shown by several examples that are described in this paper, and even very good on fission product behavior. The main need for modeling improvement concerns reflooding of a degraded core, due to the lack in ASTEC V1 of any dedicated model, and intensive efforts will focus on this topic in the next years. Molten core concrete interaction models are at the state of the art, but new experiments under way in the international frame and a better understanding of physical mechanisms are necessary to make further progress. Version V2.0 of the new ASTEC series, released mid-2009, takes benefit of the previous very intensive validation of the ICARE2 IRSN mechanistic code since its core degradation models have now been implemented. Validation will continue in the SARNET network from 2009 to 2013.
PDF
