Wind power system control based on least square support vector machines algorithms

In this paper, least square support vector machines (LS-SVM) intelligent technique is proposed for the control of a wind energy system based on permanent magnet synchronous generator (PMSG). The proposed LS-SVM is “trained” off-line to forecast at each wind speed the optimal duty cycle which is considered as control variable. In order to maximize wind turbine generated power, an inserted DC/DC power converter into DC bus is implemented by the LS-SVM forecasted control. The wind energy system is designed with SIMULINK environment, and Simulation results are finally presented.