Fault Warning Method of Wind Turbine Based on Multivariate Data Anomaly Detection

Lei Song, Li li Guo, Jun Rao, and Cheng bing He


fault feature extraction, multiple linear regression, multivariate outlier detection, fault warning


Due to the complexity of operating condition and coupling of component, the timeliness and accuracy of fault warning for wind turbine is difficult to guarantee, causing lots of mistakes and delays of fault warning. The paper proposes a novel method based on multivariate data anomaly detection. The paper firstly carries out the fault feature extraction methods of multivariate data. Secondly, the paper utilizes the MLR (Multiple Linear Regression) method to establish the accuracy fault warning model of wind turbine. Finally based on MOD (Multivariate outlier detection) method, the paper realizes the efficient and accuracy fault warning of wind turbine. Cases verify the validity of the method in this paper.

Important Links:

Go Back