Study On Active Distribution Network Reconfiguration Based On Modified Invasive Weed Optimization Algorithm

In recent years,distributed generation(DG)such as wind turbine,photovoltaic and micro gas turbine are gradual integrated in distribution network.Distributed generation has the advantages of low carbon,environmental protection,renewable and flexible control.Meanwhile,the output power of wind turbine and photovoltaic are effected by wind speed and solar irradiance with the characteristics of random variability.It makes the operation of active distribution network more complex.Distribution network reconfiguration is one of the effective means to improve the performance of distribution network.It is easy to operate without investment of additional equipment.Distribution network reconfiguration is to change network topology by recombination of switching states under system constraints to optimize objective functions.Currently,active distribution network reconfiguration is usually solved by intelligent algorithm due to its good global search ability and simple operation.Based on the analysis of distribution network research background and significance,this paper introduces and combs the research status of distribution network reconfiguration.Based on the research of network reconfiguration model,DG uncertainty and original invasive weed optimization algorithm,this paper proposes an active distribution network reconfiguration method using modified invasive weed optimization algorithm.From aspects of operation benefit and environmental benefit,a multi-objective reconfiguration model is established with the objective of minimizing active power loss,node voltage deviation,load balancing and environmental cost of pollutant emission.Firstly,multi-scenario model of DG was established by scene partition and scene reduction.Then,based on original invasive weed algorithm,this paper proposed initial population selection mechanism,which is to optimize initial solution by electric distance judgment of load power supply path and branch exchange;introduced process of Lévy flight and set up quasi hamming distance determination operation to improve population diversity;designed seed quantities adjust strategy based on the proportion of optimal individual to improve convergence.Modified invasive weed optimization algorithm can balance population diversity and convergence performance to enhance global search ability,convergenceand stability of algorithm.Finally,IEEE 33-bus system was tested by Matlab simulation.The simulation results confirmed the validity of the proposed model and algorithm.Multi-objective active distribution network reconfiguration not only considers the impact of DG access on system operation benefits,but also pays attention to the environmental benefits of DG.Compared with original invasive weed optimization algorithm,particle swarm optimization algorithm and genetic algorithm,modified invasive weed optimization algorithm has better global search ability,faster convergence speed,and stronger stability while solving the problem of distribution network reconfiguration.