Incipient Bearing Fault Detection via Stator Current Noise Cancellation using Wiener Filter

Current-based monitoring can offer significant economic savings and implementation advantages over traditional vibration monitoring for bearing fault detection. A key issue in current-based bearing fault detection is to extract bearing fault signature from motor stator current. In this paper bearing fault signature in motor stator current is detected by estimating and removing non-bearing fault components via a noise cancellation method. In this method, the components of the stator current that are not related to bearing faults are regarded as noise and are estimated by a Wiener filter. These noise components are then cancelled by their estimates in a real time fashion and a fault indicator is established based on the remaining components that are related to bearing faults. Machine parameters, bearing dimensions, nameplate values, or stator current spectrum distributions are not required in the method. The results of on-line experiments with a 20-horsepower induction motor have verified the effectiveness of this method.