Estudo de otimização de padrões de recarga do combustível para um reator PWR utilizando redes neurais feedforward multicamadas

Abstract

Fuel recharge management in a power reactor has as its main focus the recharge pattern in the core in order to achieve better efficiency in the cycle, observing all adopted safety parameters. The possibility of fuel distribution, considering a core in equilibrium, with 1/3 recharge of this and N fuel elements, arrives at NN combinations. In this article we present the strategy of an algorithm based on artificial neural networks to optimize the fuel recharge in the core of a nuclear reactor. The idea is to develop a methodology capable of choosing the best combinations that satisfy the peak radial power factor and maximize the effective multiplication factor at the beginning of the cycle, and also satisfy the ratio of minimum critical power and maximum rate of heat generation at the end of the cycle. In this work we present the development of the multilayer feedforward neural network based on multivalued neurons that will be used for the development of the methodology.