Distribution Network Reconfiguration Algorithm For Grid Computation

With the development of grid computation technology, its application extends to the complex electrical power system. The distribution network takes an electrical power system's important part, reducing the line loss and Optimizing distribution network operation has always been an important inspection content of Power Supply Company. Among measures for distribution network'economy optimization, distribution network reconfiguration can full display distribution network's potential and enhance system's security and efficiency without extra investment.On the basis of the predecessor'discussion about the power distribution network reconfiguration algorithm, this paper has analyzed the power distribution network reconfiguration algorithm for grid computation. The research involved the introduction of CIM to the distribution model, code optimization and genetic strategy of genetic algorithm for network reconfiguration, and the application of parallel model of genetic algorithm for network reconfiguration computation under grid computation environment.This paper introduces the distribution network model based on CIM for topology analysis of the distribution network. The introduction of CIM makes the algorithm and the program module good readbility or system's data exchange and the information sharing, enhancing the algorithm's openness.As for the characteristics of the distribution network reconfiguration computation, this paper makes discussion about the code optimization, cross and mutation strategy of the genetic algorithm for distribution network reconfiguration. Based on the discussion, this paper determined the encoding, cross and mutation strategy for GA.This paper has analyzed the genetic algorithm's parallelism and has established the master to slave parallel model and the community level peer to peer parallel model for genetic algorithm, and then carried on the network reconfiguration's heavy computation from the function level master to slave parallel model and the community level peer to peer parallel model on which organized machines are in a local area network for the parallel computation. The simulation results indicated that two kinds of parallel computations models have obtained the better optimized effects, remarkably enhanced the computation speed.