Operation Security Prediction for Wind Turbines Using Convolutional Neural Networks: A Proposed Method

Abstract

A wind turbine rotor system is a typical networked industrial control system. The security of its operation is very important to energy systems and users. In this article, the artificial intelligence algorithm is used to predict the security operation of a wind turbine rotor system, and a prediction method of system security monitoring based on a convolutional neural network (CNN) is proposed. First, the dynamic analysis of the operation principle of the wind turbine rotor system is carried out, and the industrial control model of the rotor system is established by using the relevant data of the wind turbine. The relevant data required for the security prediction of the wind turbine rotor system are obtained, and its dataset is established. Then, the CNN is trained with limited datasets, and the trained CNN is used to accurately predict the pitch angle. The residual information is obtained by comparing the predicted pitch angle with the real output pitch angle of the wind turbine rotor changing system. Finally, the security prediction results are obtained according to the residual and the decision index. The proposed security trend prediction method for wind turbine rotor systems can accurately and effectively predict the change of the fault amplitude, provide detection and estimate decision results, and improve the system security.