Distrituion Network Service Restoration Based On Muli-Objective Co-Evolutionary Algorithm

In modern life, both social production and people life demand uninterrupted supply and high quality of power supply. However, the complex structure and large scale of distribution network make its service failure unavoidable. There is of great significance to research distribution network service restoration and apply them.Topology analysis and power flow calculation are the basis of distribution network service restoration and other advanced functions. According to the structure and operating characteristics of distribution network in our country, this paper proposes a real-time higher topology analysis and power flow calculation method, in which the conception of node-depth is combined with breadth-first search method and forward and backward substitution method. If only part of the distribution network structure changed, carried out fast topology analysis method on the part of the topology, and get the entire network topology structure according to the results of the fast topology analysis results, power flow calculation can be made without global topology analysis. Based on this method, parallel power flow calculation can be made on multi-source distribution network without power island analysis. Simulations on IEEE 16 and IEEE33 node network system show that the proposed topology analysis and power flow calculation method based on the node-depth can effectively save the calculation time.According to the characteristics of distribution network service restoration, this paper designed a multi-objective co-evolutionary algorithm, which includes the co-evolution of two populations. In order to further maintain the diversity of individuals in the population, selection method of multi-objective individual was improved, reducing the risk of repeat individuals in a population is selected. Simulations of distribution network service restoration based on the multi-objective co-evolutionary algorithm on IEEE33 nodes distribution system show that the algorithm can find the global optimal solutions quickly.