Bearing fault diagnosis of a wind turbine using maximum likelihood detection

Bearings are the most frequently used components in a wind turbine. As such, bearing Fault Detection is an imperative part of preventive maintenance procedures of a wind turbine. This paper presents a Maximum likelihood method to implement bearing fault diagnosis. This set extracts the amplitude and frequency modulations of the vibration signals measure from a wind turbine system. As the amplitude demodulation is inherent in this set, the fault frequency can be detected from the spectrum of the transformed signal. The effectiveness of this method has been validated by using simulated signal and experimental data.