Fault Diagnosis of Wind Turbine Based on PCA and GSA-SVM

A fault diagnosis method based on principal component analysis (PCA) and support vector machine (SVM) model is proposed to solve the problem of high dimension and large sample size of wind turbine fault data. Firstly, The PCA is used to extract low-dimensional fault features from high-dimensional fault data to eliminate the correlation between features. Then, the grid search algorithm (GSA) is used to optimize the loss parameters and kernel function parameters of the SVM model. Secondly, low-dimensional fault features are used as input training classifiers for SVM. Finally, fault diagnosis is carried out through feature classification. Simulation results have shown that the diagnostic accuracy could reach 100% when Polynomial kernel function and two-dimensional principal component analysis were used, indicating that this method can quickly and effectively diagnose various faults.