Study On Distribution Network Reconfiguration Considering Multi-load Mode

Distribution network reconfiguration is an important method for operationoptimization and network reliability, and plays an important part in DistributionManagement System. In theory, network reconfiguration is a complicated problem ofnonlinear combinational optimization. The traditional mathematical optimizationmethods have great limitations in dealing with such issues.A model considering multi-load mode for distribution network reconfiguration ispresented in this paper, which can simulate the variation of load with multi-load mode.It can achieve network reconfiguration for a period of time, and is closer to the actualoptimal operation. Also, this model can solve the limitation of the staticreconfiguration model which can't reflect the variation of load and the defect of thedynamic reconfiguration model which requires frequent operation on switchesBinary Particle Swarm Optimization (BPSO) is applied to solve the problem ofdistribution network reconfiguration in this paper. Dynamic inertia weight adaptationstrategy is introduced for BPSO, which uses the evolution information of thepopulation and can improve velocity and precision of convergence. Also, an improvedencoding mode based on switches of loops of network is introduced according to thecharacteristics of distribution network, which can reduce the probability of infeasiblesolutions. And based on these, a method based on improved BPSO for distributionnetwork reconfiguration considering multi-load mode is presented in this paper.A simple method for identifying radial structures of distribution network isapplied in reconfiguration, and an improved algorithm based on branch current andnode voltage for power flow calculation of distribution network is presented in thispaper, which can be applied to network of variational structure.The results of two testing systems of distribution network justify that the methodproposed in this paper can solve the problem of distribution network reconfigurationefficiently, which has better convergent performance and can get better result.