Space nuclear reactor fuel design based on dynamic analysis and GA & PSO optimization

Abstract

A nuclear reactor is an attractive choice in space applications such as planetary and deep space exploration. This paper deals with an optimal design for a 100 kWe integrated space nuclear reactor core based on the dynamic analysis. This type of reactor has advantages such as compact structure, high power density, lower cost, higher safety and reliability, and proven technology. In this study, neutronic analysis is performed considering different fuel compositions to obtain excess reactivity, radial power peaking factor, fuel reactivity temperature coefficient, and coolant reactivity temperature coefficient. An artificial neural network is trained considering the extracted data of neutronic codes. Genetic Algorithm (GA) and Particle Swarm Optimization (PSO) are implemented to find an optimal design of the fuel composition based on an objective function. The trained neural network is considered in the optimization process. The simulation result shows that the optimized fuel composition has better performance in comparison with the existing design. Also, the optimal design is validated using neutronic calculations. © 2021 Elsevier Ltd