Fault warning of wind turbine gearbox
                                         HUANG Xiaoguang, PAN Donghao, SHI Xiaoming, WANG Xing, WANG Xin
                                                         (Zhejiang Windey Co., Ltd., Hangzhou 310012, China)

Abstract: In order to provide early anomaly warnings for wind turbine gearbox, the model of bearing temperature and lube oil pressure of gearbox was established. The model was based on the fact that the behavior mode of the gearbox would change when faults existed. A moving window statistical method was used to detect the change of the distribution of residual errors, thus the fault of wind turbine gearbox could be predicted in time by monitoring this change. The residual error overrun ratio was defined and calculated within a day. The alarm would be sent out if the ratio of the previous day exceeded the threshold set based on the historical gearbox operational data. The operation and maintenance personnel could be triggered by the alarm to check the status of the gearbox. The model was tested on real wind turbines on site, and an early anomaly warning of the gearbox temperature control valve was detected and validated. The experimental results indicate that the model can be applied satisfactory to provide the early warning of wind turbine gearbox anomalies.

Key words: condition monitoring; gearbox; nonlinear state estimation; wind turbine; fault diagnose