Line losses calculation in distribution network based on RBF neural network optimized by hierarchical GA

In this paper, we propose a practical method of calculating line losses in distribution network based on RBF neural network (RBFNN) optimized by hierarchical genetic algorithm (HGA). The first step is to determine three parameters of RBFNN, namely the number of hidden layer nodes, the width and the center of the basis function. In the second step, RBFNN is adopted to map the complex nonlinear relation between energy losses and characteristic parameters of distribution net so that the net can learn the trend of energy losses under varying distribution net structure and operation parameters. The simulation results have verified that the method presented is reliable and effective.