Artificial neural network approach to network reconfiguration for loss minimization in distribution networks

Abstract

Network reconfiguration of distribution systems is an operation in configuration management that determines the switching operations for a minimum loss condition. An artificial neural network (ANN)-based network reconfiguration method is developed to solve the network reconfiguration problem to reduce the real power loss in distribution networks. Training-sets for the ANN are generated by varying the constant P-Q load models and carrying out the off-line network reconfiguration simulations. The developed ANN model is based on the multilayer perceptron network and training is done by the back propagation algorithm. The trained ANN models determine the optimum switching status of the dynamic switches along the feeders of the network, which thereby reduce real power loss by network reconfiguration. The proposed ANN method is applied to the 16-bus test system. Test results indicate that the developed ANN models can provide accurate and fast prediction of optimum switching decisions for minimum loss configuration. The proposed ANN method is compared with Kim's method [IEEE Transactions on Power Delivery 8, 1356–1366 (1993)] and a comparative study is presented. The proposed method can achieve minimum loss configuration with drastic reductions in the number of ANNs and less computational time.