Neuro-fuzzy control and particle swarm optimization on horizontal axis wind turbine

Abstract

The consumption of electric energy is increasing. This growth stimulates production that is entirely based on fossil fuels. However, for social, political or environmental reasons there is a need to change the energy source. Large, medium and low-scale generation by means of wind turbines is a viable solution. Similar to other generation methods, the wind turbine needs to be optimized and controlled to function as efficiently as possible. In this work a particle swarm optimization process optimized a profile of a wind turbine based on two stretching equations that had as objective the best coefficient of power for the average speed. We also used a fuzzy neuro controller based on the maximum for each ratio between speeds (𝑇𝑆𝑅). Three wind speed profiles were used to analyze the dynamics of the wind turbine. The controller was efficient and kept the rotation within the expected range. The results show that the controller prevented the generation above the maximum power, reducing the rotations by up to 12 [rad/s] above the maximum power, in cases of oscillation in the velocity of the view the control remained stable with a low standard deviation and reducing the power in at up to 8 [rad/s] for sine waves and up to 9 [rad/s] for random inputs.