Study Of Service Restoration Reconfiguration In Distribution System Based On Artificial Intelligence

Network reconfiguration is not only an important method in circulating and controlling of distribution system, but also an important part of distribution management system. With the development of intelligent technology, it is possible to reduce the loss of energy in the electric network through intelligent algorithm.The background, the situation and the characteristics of the reconfiguration of the distribution network are discussed in this paper, and the popular algorithms for reconfiguration of distribution network nowadays are researched. Furthermore, the objective functions and constraint conditions of the reconfiguration are analyzed. On this basis, according to the branch-exchange algorithm, its principle is investigated and is improved efficiently itself. And then, through the software of PSASP power system, actual examples and simulation results of the distribution networks of different structures are given to verify the practicability of this method.Artificial intelligent algorithm can avoid repeatedly power flow calculation of ordinary algorithms, and it is advantageous in function approximation. In allusion to the problems of its low learning speed and local convergence, work has been done to improve the BP algorithm.Following by the knowledge of graph, three mathematical models are established. Through the constraints of tree-topology of the network, the error function is improved, and determination of ring-web and isolated vertex are made to efficiently solve the problem of local convergence to minimum of the BP algorithm. Proceeding from actual conditions, the improved algorithm is tested to confirm that it is speedy in output and efficient in function approximation.With the most important and impressive feature of self-learning, BP neural network is very suitable for long-running distribution network. Therefore, this algorithm should have a cheerful prospect in application.