Amec Foster Wheeler is a global leader in the field of nuclear safety advice and analysis, and the company has built a powerful suite of software to model and simulate such facilities, and assess the safety and operation of nuclear reactors.
This includes tools for criticality calculations (determining whether or not a given geometric configuration is capable of sustaining a chain reaction or not); and reactor physics calculations (analysing the physical properties of a reactor that's currently sustaining a chain reaction). These computations are essential for reactor analysis and assessing the safety limits for fuel transportation and storage.
Increasing safety standards and demands for improved plant performance have led to requirements for improved fidelity of modelling and simulation tools, and to extend the domain of applicability of modelling tools to cover a wider range of plant scenarios. This was the case with one of the tools for perturbation analysis, which did not converge reliably when used over an extended range of scenarios. Reliability is paramount in the nuclear industry, which sets very high safety standards, so Amec Foster Wheeler approached mathematicians at Bath to investigate and suggest remedies.
Our researchers drew on their knowledge of relevant areas of numerical analysis, including numerical linear algebra, partial differential equations and eigenvalue problems - the fundamental mathematical building blocks used in modelling nuclear reactors. The team identified conditions under which the company's Monte Carlo perturbation module was not guaranteed to converge and, as a result, suggested a method to extend the range of scenarios for which it would. This prompted the company to recode parts of the module to extend its range of applicability.
Amec Foster Wheeler's nuclear modelling software is used throughout the UK nuclear industry and in many countries around the world. The research carried out by mathematicians at Bath had a clear impact on its robustness, and thus the ability of the company to compete as a world leader in the development of these software tools. They have now started a new programme of research with us.
Paul Smith from Amec Foster Wheeler said: "The work performed for the CASE project provided considerable insight into the convergence of the perturbation scheme and suggested improvements to the method. Our reactor physics code has been modified in the light of the improved understanding, resulting in improved robustness of the scheme over a wider range of applications. Without such improvements to the existing method, it would have been necessary to develop alternative methods to satisfy industry needs, at considerable cost to the nuclear industry."