Design of Steam Thermal Deicing Control System for Wind Turbine Blades Based on BP Neural Network Algorithm

With the growth of wind power industry and the increase of single unit capacity, wind power generation plays a more important role in the national economy, which also puts forward higher requirements for the reliable and stable operation of wind turbines in complex natural environment. Ice coating on the blade surface will change its cross-sectional shape and structure, so that its inherent aerodynamic characteristics will be changed and its aerodynamic performance will be reduced. The increased ice weight will change the natural frequency of the fan parts, which may lead to the stall of the fan blades ahead of time or later, resulting in the loss of wind power generation. Compared with other anti-icing technologies, the blade gas heating method has the engineering and technical advantages of easy transformation, easy maintenance, high deicing efficiency, flexible design and operation. In this article, a back propagation neural network (BPNN) based control algorithm for wind turbine blade thermal de-icing is proposed. This method is used to establish the model of wind turbine, which only needs the data of input and output training samples, and does not need the inherent parameters of the fan and the calculation and derivation process, thus making the modeling and simulation process simpler and more accurate.