Application of Modified Genetic Algorithm in Source Term Inversion of Nuclear Accident on Offshore Platform

In this paper a source term inversion model is built to study the source release during severe accident in nuclear power plant. The model is based on the combination of genetic algorithm and Nelder-Mead simplex algorithm. It uses the off-site monitor data and the meteorological data to estimate the amount of radiological materials released into environment during severe accident. This method can support the rapid obtain of source term during nuclear emergency and become a supplement to the core damage evaluation method. Based on numerical analysis experiments, indoor wind tunnel experiments and field tracing experiments, the established inversion model is verified. The verification results show that the source term inversion model established in this paper can quickly and accurately give the release source term and the max error belows 48%. This model can be used in the consequences evaluation of nuclear accidents, provide suggestions to accident emergency decision-making.