Wind Farm Layout Optimization Considering Obstacles Using a Binary Most Valuable Player Algorithm

Wind farms are developed and implemented in many places around the globe. Designing a wind farm is becoming more and more complex especially with the recent trend towards large farms. Finding the optimal locations of wind turbines inside a wind farm to reduce energy cost is a highly challenging task, as it requires the handling of conflicting criteria and depending on the number of turbines considered it can turn to a large scale-optimization problem. Therefore, the aim of this paper is to place efficiently wind turbines inside a given area considering all constraints. This problem formulated as an optimization problem is referred to as the wind farm layout optimization (WFLO) problem. This real-world problem is nonlinear and difficult to solve using classical optimization algorithms and it has to take into consideration wind scenarios, power curve and wake effects. For this purpose, a binary version of the most valuable player algorithm (MVPA) called BMVPA is developed and implemented. Furthermore, ten scenarios were investigated using different wind speeds, terrain sizes with and without obstacles. For the same terrain but including obstacles, it was found that the energy cost increased due to the presence of obstacles that could limit the search space and consequently reduces the number of available options. The empirical results obtained using BMVPA were compared with those obtained using other well-known algorithms like the binary particle swarm optimization and genetic algorithm. BMVPA showed better results in solving the WFLO problem than the comparative algorithms. The optimum design of the wind farm obtained will allow an efficient and economic exploitation of wind resource.