Distribution Network Planning Based On Improved Genetic Algorithm

As the connection between transmission network and consumers facilities, distributionnetwork has been paid more and more attentions about it’s planning. A scientific and efficientoptimization project for distribution network planning can save investment, reduce powercutting and the line loss, as well as improve the quality of electric power, which is currentlyone of the important task of the departments of power system．Distribution network structureplanning is a combine optimization problem. With the enlargement scale of distributionnetwork and the increment of the possible roads dimensions, it becomes increasingly complex.The thesis carries on to the model and algorithm of distribution network structure planningand reactive power planning deeply with improved Immune Genetic Algorithm. Main contentsare as follows:To begin with,a11kinds of methods on power flow calculation in distribution networkswas discussed in this dissertation, especially highlighting the technique of backward／forward iterative algorithm. In order to find the objective function of optimal networkconfiguration in network planning，topology analysis is needed to change the run mode ofnetwork. This paper selected a best method of power flow calculation and improved it in useof distribution network planning.Repair the unfeasible solutions for distribution networkstructure planning. The results indicate that the algorithm has a good astringencyBased on biological immune system’s antigen recognition，maintaining the diversity ofantibodies and immune memory and other features. Put forward an improved immune geneticalgorithm, by vaccination, immune selection and each population evolves in coordination.thealgorithm will introduce the thinking of biological systems immune to the genetic algorithm.Uses the mathematical model whose goal is to make the lines integrated cost of theplanning year the least in the study of the distribution network structure planning，and appliesimproved immune genetic algorithm to solve the model with single and multiple distributionsubstation．Through the simulation for some planning examples，it is proved that computingspeed and convergence performance of improved immune genetic algorithm are better thanthe other algorithms. Then，it uses the mathematical model whose goal is to make the annualintegrated economic benefits the most in the study of the distribution network reactive powerplanning．Two approaches are studied for the problem．The one is to choose the reactivepower compensation points with the impedance reactive power moment method firstly，andthen apply the immune genetic algorithm to solve the compensation capacity；the other is that the position and the compensation capacity of reactive power compensation points are allsolved by the immune genetic algorithm．Digital simulation for IEEE33-bus test systemdemonstrates that the two approaches are effective．And the compensation benefits andeconomic benefits of the two algorithms were compared．...