Investigation Of Wind Farm Layout Based On Particle Swarm Optimization Algorithm

Since the worldwide energy needs has been increasing for a long period and development of wind turbine technology has now become a hot spot issue, the advanced planning of wind farm has been more and more accounted of by many countries. One of the key links of advanced planning of wind farm is called'wind farm layout optimization', which influences the coefficient of utilization of the wind and land sources directly, as well as the economical efficiency of a wind farm. Currently it is generally realized that former engineering experience and business softwares are quite limited in wind farm layout optimization, so Particle Swarm Optimization is introduced into this crucial issue and further computer program is systematically developed in this thesis.Programmed the Wind Farm Layout Optimization problem by basic Particle Swarm Optimization algorithm, and validated it by comparing with foreign authority's result. Based on the investigation of the influence of computation grids on optimal result, took the "cost of energy" as the optimization target, and made a sensitivity analysis study on the influence of wind speed, wind speed probability distribution and roughness on the optimal configuration. The research shows that the cost of electricity can be significantly reduced after the Wind Farm Layout Optimization work even wind speed is less than turbine's rated speed. When the inflow is described by the Weibull distribution function, the results show that bigger scale factor c and shape factor k are helpful to reduce the cost of electricity at low wind speeds, at higher wind speeds, bigger shape factor k is helpful to improve wind farm economical efficiency, and the value of scale factor needs reasonable control.When modeling the Wind Farm Layout Optimization problem by basic PSO algorithm, we have to divide the terrain into several cells, which means the pre-set of wind turbine placement and the maximum number of turbines. To prevent the wind turbine installation location discontinuities caused by the pre-set of computational grids, the design variables in the Particle Swarm Optimization algorithm is increased, to act directly on the coordinates of the wind turbines. The result is compared with the optimization algorithm with grids, which shows that with the same number of wind turbine units, the continuous layout of wind turbines can effectively improve the wind farm output.Finally, in order to improve the ability of the basic Particle Swarm Optimization algorithm to solve the layout problem under irregular shape of wind farm, the binary-coded Particle Swarm Optimization algorithm is introduced into this optimization problem, which can operate on every wind turbine's location. This method is validated with several engineering cases, which expands the application area of the optimization method.