Abstract
This paper presents the mathematical model for multi-stage distribution substation planning. In order to decide the substation-planning scheme in each stage simultaneously, this work put forward a new method for getting initial feasible solution. The initial feasible solution makes load points in every stage into consideration by classifying new substations into several classes. Based on the initial feasible solution, we present Hop field neural network to calculate each substation's planning scheme in every stage simultaneously. The energy function of the neural network and calculation algorithm are given in this paper. A real multi-stage planning example is also given, and the comparison between the calculated scheme and the scheme that is done by power institution shows that the presented algorithm is effective.