The Time Series method and Statistical Process Control strategy is applied to predict failures of wind turbine gearboxes. First, based on the real-time temperature data of gearboxes measured by temperature sensors, the temperature prediction model under normal operating conditions is established by ARIMA model. The analysis of the predicted values and the actual values of gearbox temperature is done, and proves that its residuals are normally distributed; then combined with statistical process control (SPC) methods, the big number of temperature data is used to calculate the standard deviation(σ) of residuals, and the gearbox failure threshold will be identified; Finally, the temperature data are analyzed both in normal operating condition and the failure condition to determine the operation status of the gearbox, statistical analysis and residual charts are carried out for gearbox failure prediction, verifying the feasibility and effectiveness of the proposed method.
Wind Turbine Fault Diagnosis and Predictive Maintenance Through Statistical Process Control and Machine Learning
