A robust intelligent control for a variable speed wind turbine based on general regression neural network

Abstract

In this paper, a robust general regression neural network sliding mode (GRNNSM) controller is designed for a variable speed wind turbine. The objective of the proposed control is defined in relation with the trade-off between the wind energy conversion maximization and the minimization of the stress on the drive train shafts. Sliding mode control approach (SMC) emerges as an especially suitable option to deal with variable speed wind turbine. However, for large uncertain systems, the SMC produces chattering problems due to the higher needed switching gain. In order to reduce this gain, general regression neural network (GRNN) is used for the prediction of model unknown component and hence enable a lower switching gain to be used. In the present work, back-propagation (BP) algorithm will be used to train online the GRNN weights. A robust control term with low switching gain is added to compensate the neural network errors. The stability is shown by the Lyapunov theory and the control action used did not exhibit any chattering behavior. The effectiveness of the designed method is illustrated in simulations by the comparison with traditional SMC.