Distribution Network Reconfiguration Based On A Cooperative Co-Evolutionary Algorithm

Distribution network reconfiguration is a non-linear and high dimensional optimization model. The introduction presents the basic content, characteristics and current worldwide research level of network reconfiguration. It also compares advantages and disadvantages of different optimization algorithms, and lists a couple of objective functions of distribution network reconfiguration.One of the substantial restrictions in reconfiguration model is the connection and radiation of topology structure. It is essential to make each solution keep in line with the restriction during the whole solving process. Based on analysis of several usual connection modes of distribution network and topology theory, this dissertation presents the criterion of connection and radiation and its realization in programming. The forward/backward sweep method based on branch current is applied to compute the power flow of distribution network.A weighted average objective function with voltage balance index and power loss is established. The updating strategy of PSO and the annealing schedule of SA are modified to satisfy the traits of distribution network. CPSOSA is applied in cases with different scale and compared with PSO and SA.The basic principles and characteristics of Particle Swarm Optimization (PSO) and Simulated Annealing (SA) are analyzed. And then Cooperative Co-evolutionary algorithm (CPSOSA) is proposed, which involves a lock of local optimum area based on Gaussian distribution and a self-adaptive cooling schedule. The particles in local optimum area are performed with parallel simulated annealing to strengthen the climbing ability and global search ability of PSO. The results testify the superiority and practicability of CPSOSA.Finally, a mathematic model for fault restoration is established. Family tree structure is applied to represent the connection mode of distribution network so as to accelerate topology analysis. CPSOSA is applied in the phased fault restoration method. The phased method can provide optimum restoration strategy according to different network structures and various fault locations. The results of case study testify the practicability and efficiency of the fault restoration algorithm for distribution network.