An accessible statistical regression approach for the estimation of spent nuclear fuel compositions and decay heats to support the development of nuclear fuel management strategies

Abstract

Computational methods are essential to support and advance nuclear technologies due to the hazards of handling and analysing highly radioactive materials such as spent nuclear fuel (SNF). However, many such methods, including those that can predict SNF compositions and decay heat parameters, often require expensive, proprietary software, alongside significant computational experience and power for utilisation, severely limiting availability of data, and hampering research throughput. Although some datasets are available, including some where post-irradiation analyses of SNF have been carried out and published in literature or technical reports, many are incomplete, or only cover certain fuel systems for older reactor types. Research investigating new methods for SNF recycling, for example, requires compositional and decay heat data for fuel systems not covered by published computational or experimental SNF composition data, though analogous source data may be available. With this in mind, we have developed a simple, accessible, and flexible method for estimating isotopic, elemental, and decay heat compositions for SNF at discharge and following decay storage before recycling. This semi-empirical method uses relatively simple physical and mathematical principles and can be performed using software accessible to all researchers. This provides outputs accurate to within 1% of reference values interpolated within the range of available data for isotopic compositions, with sensible extrapolations at higher burnups beyond those reported, with overall elemental outputs accurate to within 0.1% of expected totals. In this publication, we present the developmental methodology, some sample data, the present limitations, and options for future development and expansion of functionality.