Support Vector Regression Inverse System Control for Small Wind Turbine MPPT with Parameters’ Robustness Improvement

Abstract

With the increasing penetration of the permanent-magnet direct-drive wind power system, the maximum wind-energy capture and the generation speed control are more and more important. In the literature, the dynamic performance of the generator speed is well documented by the inverse system method. However, conventional inverse system methods have parameter dependency that is not sufficient to meet the dynamic requirements for permanent magnet synchronous generator (PMSG) speed tracking. Therefore, this paper introduces a support vector regression machine (SVR) method, especially for the inverse system model, which could solve the inaccurate parameters problems. As the SVR has the nonlinear approximation ability to identify and adjust the parameters online, thus, the system robustness could be improved. Finally, the dynamic performance of generator speed is evaluated by using the SVR method. Proposed theoretical developments are verified by the Simulink Test and experimental test.