Multiple Linear Regression Approach for Sensorless MPPT of PMSG Wind Power Generation Systems

Abstract In this paper, a novel approach for extracting the maximum power from a wind energy conversion system (WECS) eqquiped with a permanent magnet synchronous generator (PMSG) is presented. Wind power generation systems have gained prominence as an alternative to conventional power plants worldwide, being the renewable energy source with the best growth prospects. However, the intermittent nature of wind power reduces energy reliability and hinders the development of technologies in this area. Therefore, a sensorless controller for a WECS with a PMSG is developed to enhance the conversion efficiency of the wind enegy system. The proposed control algorithm predicts the wind speed from the wind turbine power-speed characteristics. It uses the predicted speed to determine the optimum reference rotating speed to extract the maximum power point (MPPT) based on the optimum tip-speed ratio. The proposed control approach is based on a multiple linear regression (MLR) algorithms which is used to predict the wind speed value by calculating the regression coefficients from the predetermined training samples of the turbine power, rotating speed, and wind speed. The MPPT operating mode is maintained as long as the available power is less than the rated power of the converter. Otherwise, the MPPT algorithm is disabled, and a mechanical power control loop guarantees nominal power. A complete theoretical analysis and the experimental results considering the prototype with the nominal power of the 3 kW wind turbine used are presented. The experimental results has proven the effectiveness and robustness of the proposed control approach.