Research On Distribution Network Reconfiguration Based On Improved Particle Swarm Optimization Algorithm

Distribution network reconfiguration is an significant method to improve theoperation reliability and optimization of the power system, and plays a key role inDistribution Management System (DMS). In theory, network reconfiguration is acomplicated problem of multi-objective and nonlinear combinational optimization. Itis quite hard for the traditional mathematical optimization methods to deal with suchissues. Particle Swarm Optimization (PSO), a rising intelligence algorithm, has someamazing characteristics such as high calculation speed, good convergent performanceand tends to attain global optimization. These characteristics make it easier for PSO tosolve large number of nonlinear, multi-peak value and non-differentiable complicatedoptimization. PSO is successfully used in neural network training, functionoptimization and some other fields.In this paper, a model based on PSO with mutation operator is proposed fordistribution network reconfiguration. According to the actual network, binary PSO isstrongly recommended to deal with the operation of reconfiguration. On this basis, amathematical model, whose objective is loss minimization, is founded. Meanwhile, toconquer the defect that PSO is easy to fall in local optimization, mutation operatorbased on the theory of Genetic Arithmetic (GA) is introduced. Part of the particles inthe swarm will be initialized when they cannot get rid of local optimization.The result of distribution network testifies that the method in this paper canhandle the distribution network reconfiguration efficiently and get a satisfied result.