Based On Particle Swarm Optimization For Reactive Power Optimization

Reactive power optimization in power system is an effective means to improve voltage quality, reduce network losses and ensure the secure and economical operation of power system. Based on this, much effort has been made on it by domestic and foreign experts for many years. Whereas, so far, the problem can not be settled properly.Essentially, reactive power optimization problem is a non-linear complicated optimization problem with non-continuum, multi-variables, multi-restraint, plentiful local minima. Particle swarm optimization(PSO) is defined as a kind of heuristic algorithm based on swarm intelligence, which is characterized as convenient operation, fast convergence, high optimized efficiency, less sensitive property to population size, good robustness etc. And which can be applied to solve complicated optimization problems with discrete variables, non-continuum, multi-variables, multi-restraint, non-linear property. Therefore, PSO is quite suitable for reactive power optimization problem.First, this paper introduces the category of reactive power optimization problem, and summarizes many methods, key problems on it. Second, the Newton-Raphson method flow calculation is emphatically introduced and the reactive power optimization problem's mathematic model for PSO is established as well. Third, the paper studies the mathematical model, the procedure flow, the parameter analysis and the thinking of improvement on PSO and summarizes the characteristic of PSO. In addition, which introduces in detail the present applications of PSO in power system. Consequently, a modified PSO(MPSO) which can change inertia weight freely is proposed. Meanwhile, stochastic particle swarm optimization(SPSO) and simulated annealing SPSO(SASPSO) are introduced as well. Those three algorithms have been respectively utilized in reactive power optimization problem, and simulation results from standard IEEE14 and IEEE30 bus systems are presented explicitly. Moreover, the contrastive analysis between their results and the outcomes of PSO proves that in reactive power optimization problem the properties of global convergence and convergence precision of MPSO and SASPSO are more advanced than those of PSO.