Data-Driven PID Controller of Wind Turbine Systems Using Memory-Based Smoothed Functional Algorithm

Stochastic nature of wind speed and turbulence generated between turbines are commone factors that cause stress in wind turbine. Hence, it is important to regulate the rotor speed of wind turbine based on the desired reference speed. Consequently, employing a PID-based controller is crucial for maintaining wind turbine system performance. Recently, there has been a growing interest in utilizing better optimization tools to tune the PID control parameters, providing an advantage in enhancing wind turbine system output response while maintaining the robustness and simplicity of PID controller. However, existing optimization tools for tuning PID controllers, particularly those based on multi-agent optimization, often involve a large number of function evaluations (NFE), resulting in high computational burdens. This study introduces a novel approach using a memory-based smoothed functional algorithm (MSFA) to tune the PID controller in wind turbine systems. The MSFA, which is in the class of single-agent based optimization techniques, requires fewer number of function evaluations per iteration, addressing the computational burden issue. Simulation analyses, encompassing the fitness function convergence curve, step response time specification analysis, Bode plot stability analysis, and computational effort analysis based on NFE, are performed to evaluate the effectiveness of the suggested PID controller for wind turbine systems based on MSFA. The results indicate that the proposed MSFA-based PID controller is highly effective in producing better integral absolute error (IAE) with less values of NFE in comparison with other existing based methods.