A Deep learning Model-based Wind Power Generation System for the control of Permanent Magnet Synchronous Generators

Wind power is a rapidly growing and cost-effective renewable energy source. This research introduced a novel method for controlling wind turbines equipped with permanent magnet synchronous generators (PMSGs) to maximize electricity generation from the wind and minimize energy waste. This innovative approach optimizes how fast the turbine blades spin and the electrical currents involved maximizing wind power generation and minimizing overall losses in the system. The relevant tip speed ratio ‘ $\delta$ ‘ and quadrature current ‘ $i_{q}$ ‘ are specified as targets for each scenario when the rotor speed is varied from $50,100, \ldots, 1000$. The rotor speed and the accompanying core loss are fed into the deep neural network (DNN). After setting the target, the corresponding data, such as ‘ $\lambda$ ‘ and ‘ $i_{q}$ ‘ were predicted with minimum error.In the proposed study, the quadrature axis current, rotor speed, and tip speed ratio are tuned to improve wind power output. The error minimization among the actual and predicted data is considered as the objective function. The implementation of the presented control strategy outcomes has demonstrated superior dynamic performance and enhanced power efficiency for the wind generation system