Application of Neural Network Combined Taguchi Method on Wind Turbine Parameters

This paper is to solve the frequency problem of wind power grid-connected. For the control strategy of rotor speed and pitch angle of wind turbine, the selection of reference speed and pitch angle is determined by Taguchi method. Since the primary Taguchi method can only obtain a parameter under a fixed wind speed, and the actual wind speed is constantly changing, the Taguchi method is used to calculate the rotational speed and pitch angle of 10 groups of different wind speeds. The data is trained by neural network to realize the application of Taguchi method in the case of variable wind speed. From the simulation results, the parameters obtained by Taguchi method are within the range of $\boldsymbol{50 \pm 1 ~\text{Hz}}$ after simulation, which meets the requirements and has better stability. The trained neural network can also meet the frequency requirement under variable wind speed. Therefore, Taguchi method combined with neural network can be used in the selection of wind turbine control parameters.