This paper investigates the sensitivity of machine electrical quantities when employed as a means of bearing fault detection in wound rotor induction generators. Bearing failure is the most common failure mode in rotating AC machinery. With the widespread use of wound rotor induction machines in modern wind power generation, achieving effective detection of bearing faults in these machines is becoming increasingly important in order to minimize wind turbine maintenance related downtime. Current signature analysis has been demonstrated to be an effective technique for achieving detection of different fault types in ac machines. However, this technique lacks sensitivity when used for detection of bearing failures and therefore sophisticated post processing techniques have recently been suggested to improve its performance. As an alternative, this paper investigates the sensitivity of a range of machine electrical quantities to bearing faults, with the aim of examining the possibility of achieving improved bearing fault detection based on identifying a clear fault spectral signature. The reported signatures can be subjected potentially to refined processing techniques to further improve fault detection.
A Comparison Study on EEG Signal Processing Techniques Using Motor Imagery EEG Data
