Reactor core-structure re-location modelling for severe nuclear accidents

Modern computational codes can be very useful in assessing the behaviour of nuclear power facilities and ensuring that they present minimal hazard to; the public, facility workers and the environment when they enter unintended operating scenarios. This proposal is to develop such methods, as well as to establish a simulation tool that is; accurate, robust, efficient and validated, and able to determine the levels of confidence that we can place on the models. It seeks to help establish advanced computational methods to address problems in fault conditions, as well as to investigate aspects of system behaviour in severe accident situations. It will help provide accuracy that is beyond what is currently possible, and will allow the physics to be explored that cannot be reproduced through experiment. The work proposed here seeks to achieve this by developing a basis for the verification and validation of computational tools against benchmark cases that will then be used to simulate more complex/realistic scenarios. The project will combine the expertise from the UK and Japan, both within academia and industry.

The specific situation that this project intends to investigate is in a civil reactor's response to severe accident scenarios following a loss of its coolant. Under such situations internal structures can be compromised and the melting of control rods and/or fuel pins may occur. In extreme situations whole sections of the core may melt resulting in large quantities of molten materials accumulating in the vessel's lower head. In all these situations the relocation of the core's materials will affect the functioning of the reactor. There is a possibility of achieving a sustained critical reaction, resulting in extensive heating, and the coolant flows will be diverted thus preventing the heat removal from parts of its core. In addition, chemical reactions can occur, some of which pose significant hazards. Examples include oxidation processes between the air and/or steam with the fuel cladding (zirconium in particular) or control rod materials resulting in hydrogen production - as what occurred in Fukushima.

The aims of this project will be to develop a generic framework for accident modelling and validate it through the study of control rod and fuel pin melt. A computer model will be built capable of resolving, in unprecedented detail, the melting, relocating and re-solidification process of the pins when overheated. The objectives are to build the specialist numerical tools that enable the complex physics and chemistry to be resolved. The objective also includes a validation process by a comparison with experiment. This will be through collaboration with academics and industry in Japan.

The outcomes of this work will help scientists and engineers understand the processes during accident and melting scenarios. They will help improve future designs and aid operators' responses to such events. In addition, they will help to enhance safety, limit damage and inform policy makers on design integrity. Importantly, the outcomes of this work will demonstrate to the public our commitment to safety in order to regain or strengthen their confidence in nuclear technology.

This work will benefit those scientists, government bodies and industries concerned with nuclear power safety. It will also be of interest to those interested in multi-phase flows, structural models, damage models, geological safety of waste repositories and state-of-the-art computational modelling. Specific organisations that will benefit from the modelling and increased understanding include AWE, AREVA, NNL, NDA, HSE, Rolls-Royce, AMEC and EDF (now running UK nuclear reactors). Example areas that will benefit include: nuclear safety, core design analysis, training and decommissioning. It is the research outputs of the kind planned here that is important in view of the central role that nuclear power is expected to play over the coming decades. New reactors and reactor types, and those undergoing life extensions, must meet ever more stringent economic and safety criteria, and assessing their ability to meet these must increasingly rely on advanced computational modelling. The outcome of this research will help determine the strategy for nuclear accident modelling which aims to enhance safety in nuclear reactors. Designers, assessors, regulators and operators of nuclear plants will benefit through better analysis tools, properly validated, which are essential to realize the continuing benefits of nuclear new build. In the unlikely event of an accident the framework proposed here will have helped the scientist and engineers in predicting the events that follow, thus improving their preparations for recovery. It would provide a key component in enabling optimized mitigation strategies to be developed during an evolving accident.

Society will also benefit from the economical and carbon-free energy that nuclear plants provide as well as reassurance that the science and tools behind their safety are of the highest quality. This in turn will help improve public confidence and perception of nuclear safety. The economic impacts will be seen from the potential use of the software and associated consultancy, both within the UK and abroad, and in the prevention of overseas codes dominating the market. The impacts also include addressing the urgent need to train scientists and engineers, at all levels including doctoral and postdoctoral levels, to undertake the necessary activities for the next generation of nuclear power. Besides the research outputs, programmes like this are valuable in training such people.

This work will also be of interest to wider EPSRC and NERC communities. For the EPSRC community, areas where our work will be of benefit include: plasma physics, nuclear waste modelling, industrial processing, combustion, chemical and catalyst applications, and oil and gas. NERC communities will benefit from this next generation multi-phase technology for research in predictive flows in atmospheres and oceans. Many of the techniques and tools we will be working on are also of interest to the wider computational physics community. There is strong potential for their re-application to resolve other physical phenomena using the general adaptive discretisations and solver technologies. There is also a strong interest in using this software for multi-phase simulations of nuclear systems in National Laboratories including those in the USA, France and Japan, where our researchers have acted as consultants often using our models.