Windfarm optimization using Nelder-Mead and Particle Swarm optimization

Wind energy is a source of green and clean energy and has become the focus of environmentalists across the globe. As companies and governments invest and expand on wind farms it becomes crucial to implement ways to optimize its layout. Wind farms are generally plagued by the wake effect which causes reduced output and can affect the generation capacity and profits of a company. As wind farms can have up to hundreds of turbines, the problem of optimizing its performance to reap maximum energy with minimum area arises. With the goal of solving this problem, this paper suggests the combined use of genetic algorithm, Particle Swarm optimization (PSO) and Nelder-Mead Simplex Method for optimization. This approach uses genetic algorithm for greedy search, and PSO and NelderMead for local and fine search. The approach as will be seen in subsequent sections can help improve the annual production by 4-7%. The wake effect considered in the optimization problem is calculated using the Jensen's model. Final result obtained for a windfarm of dimension 4000x4000 m2 and 50 turbines to have an AEP of ry 526GWh improved from 505GWh.