A novel optimization method for maximizing wind farm performance through turbine positioning and yaw angle estimation

Abstract

This study presents novel approaches to optimize wind farm performance by utilizing smart initialization techniques combined with optimization algorithm Levenberg–Marquardt to enhance position of the wind turbines, although fast yaw angle estimation technology with the Levenberg–Marquardt algorithm to enhance the yaw angle estimation of the turbines in order to increase the power generation of wind farm has been implemented. The k-means clustering was used for the evaluation of wind parameters, allowing for a precise estimation of wind speed and direction while significantly reducing computational time. The proposed combination of smart initialization with Levenberg–Marquardt algorithm technologies obtained the best results in the position of the turbines compared to other optimization algorithms. Although, fast yaw angle estimation can be independently or combined with the optimization algorithm improve the yaw angle estimation. The yaw angle of the first wind turbine (in a given wind direction) that maximizes the power generated by the remaining wind turbines is determined, then estimating the yaw angle of the second wind turbine that maximizes the power generated will continue and so on; then the results refined by the Levenberg–Marquardt algorithm. The results demonstrate significant improvements in overall wind farm performance by 44.9 MW in a 80 wind turbine farm compared to the non-optimized wind farm and by 5.6 MW when smart initialization added to Levenberg–Marquardt, highlighting the effectiveness of combining these advanced techniques for better resource management and energy efficiency in wind energy systems.