Research Of Distribution Network Planning Optimization Based On Modified Particle Swarm Algorithm

Distribution network is one of the most important parts of the power system, it distributes the power that comes from power source and transmission network to users at different Voltage Levels, The planning and optimization of network structure are the focus of distribution network planning. If this job can be done well, the investment efficiency of distribution can be greatly improved and it will also lay a wonderful foundation to the flexible and reliable operation of distribution.At first, this thesis had done several researches about the model of distribution network structure and proposed an uncertainty model. What's more, Particle Swarm Optimization was studied. Though the analysis of improvement strategy, I proposed the Advanced Particle Swarm Optimization. Then, we had a depth research about optimization of network structure by the Advanced Particle Swarm Optimization. We also had a preliminary study about Distribution network Reconfiguration and recovery of distribution network failure isolation. At last, the feasibility of the method was proved by projects. The main research and details of this thesis is as follow:(1)Considering about the uncertainty of the distribution network planning, an uncertain distribution network structure optimization model which based on load changes was proposed in this thesis. This model could make the distribution network structure optimization model closer to projects, and it laid a great foundation to the choice of distribution network structure optimization program.(2)In this thesis, flatness information of fitness function was used to change dynamic inertia factor in Particle Swarm Optimization. The information of objective function was fully used by the improved algorithm, which made by the inspiration of search direction. The concept in Immune algorithm was lead into this thesis to improve the diversity of particle. The POS was improved by concentration of the selection mechanism. Though the simulation, this improved algorithm could enhance the diversity of particle effectively, and the convergence rate was speeder. It also avoid that Particle Swarm Optimization fall into local optimal solution.(3) Modified Particle Swarm Optimization was used into actual planning and optimization of network structure. Though the analysis of planning structure, we could find that the convergence rate of this algorithm was fast and the accuracy of it was high. This algorithm could overcome"precocious"problem in iterative process optimization too.(4)Distribution Network Reconfiguration based on modified Particle Swarm Optimization and the recovery of failure isolation was researched in this thesis. The feasibility was tested by practical example.