Analysis And Application Of Machine Learning Algorithm In Wind Turbine Fault Prediction

With the rapid development of information and communication technology,the power system has realized advanced automatic management through the introduction of network technology and control technology,and many distributed management systems have been developed.The power system has accumulated huge amounts of data and information.However,a large amount of data hiding fault information is not fully utilized.In reality,fault detection is usually based on monitoring and diagnosis rather than prediction.In this paper,we mine hidden information from these massive data through machine learning algorithms to make predictions.In this paper,Machine learning algorithms are applied to the fault prediction of the induced draft fan in the power plant,the fan operating data are applied to mine information through machine learning algorithms,and to verify the validity and feasibility of the algorithms in fault prediction.The main study of three aspects of the problem.First of all,fault prediction is usually faced with high-dimensional data problems,after some feature extraction methods for analysis and research,two feature extraction methods are proposed to reduce the feature dimension,which are Principal Component Analysis(PCA)and Partial Least Square(PLS).The experimental results show that both principal component analysis and partial least squares support most of the information of the original data,and the cumulative variance contribution rate is high,but partial least squares are better explained in the choice of principal components ability.Secondly,the key problem of fault prediction technology is how to improve the detection effect.A Support Vector Machine(SVM)regression algorithm is proposed to improve the detection accuracy.Experiments show that the performance of SVM regression model based on partial least square feature extraction is better than the SVM regression model based on principal component analysis,and the mean square error of fan fault prediction is small.Finally,the algorithm of partial least squares and SVM regression are applied to the fan fault prediction system.