A practical method for fault diagnosis of wind turbine gearbox using multi-source information fusion

Current methods for fault diagnosis of wind turbine gearbox are mainly based on three types of condition monitoring data, namely temperature, vibration and shock. In practice. These three types of data are often used individually, with each one having its own limitations. In this study, based on the multi-source information fusion and an improved D-S (Dempster-Shafer) evidence theory, we proposed a practical method for fault diagnosis of wind turbine gearbox using synchronous or quasi-synchronous sampling data of temperature, vibration, and shock. Through real experiments, it is shown that the proposed method can improve the accuracy of diagnosis, and reduce the false alarm rate comparing to method using single type of data.