Flutter Suppression of Wind Turbine Blade Based on RBF Neural Network Compensation Backstepping Control

Abstract

Aiming at the critical flutter problem of wind turbine blades, which is between classical flutter and stall flutter, a flutter suppression scheme based on radial basis function (RBF) neural network friction compensation backstepping is presented. The structure model is based on the typical 2D section of bending and twist model of spring-mass-damper, and the rotor variable exciter second-order model with friction disturbance is incorporated to control the rotor variable blade. A modified quasi - steady - state aerodynamic model was used for aerodynamics actuation. RBF compensation backstepping control scheme is a block-controlled backstepping controller designed based on the stability theorem of Lyapunov function, which approximates the frictional interference with nonlinear characteristics through RBF network, and cancels the friction existing in the actuator of variable rotor. Four wind speed environments were selected to analyze the response of blades under different wind speeds, and the flutter suppression effects under two wind speeds were selected to verify the ability of RBF network to approach the nonlinear function. The results show that the RBF backstepping control scheme can improve the robustness to suppress the critical flutter problem of wind turbine blades.