Chaotic artificial neural network in reactive power optimization of distribution network

This thesis mainly focused on the study of reactive power optimization of distribution network by using Chaotic Artificial Neural Network. It established the optimization model of power distribution from the comparison with high voltage power system, and implemented it by Chaotic Artificial Neural Network. It depend on Chaotic Artificial Neural Network imitating the pain point of the grid, that once there was unsuitable reactive power on the line, the nerve cell could fell pain and had a pain output. Then we could find out where there was unsuitable reactive power, and redistribute reactive power once more, till the optimization of reactive power result had been found out. To overcome the poorly computational efficiency of simple Artificial Neural Network, we combined with Perceptron Network, Self-organizing ANN and Hopfield ANN to rebuild a new Chaotic Artificial Neural Network. So we also made an improvement of it to enhance its searching efficiency. At last, the calculation results showed that the presented method was effective through a calculation of a real distribution network with 30 nodes in IEEE. This method was also suitable for Smart Grid.