Wheel bearing fault isolation and prognosis using acoustic based approach

Abstract

Wheel bearing fault detection, isolation and failure prognosis are critical to improve perceived quality and customer experience for retail vehicles, and to reduce the repair cost and down time for fleet vehicles. Currently, most of the research in bearing failure and degradation diagnosis focus on vibration signal analytics. However, these techniques are rarely applied in automotive industry due to the high sensor cost, installation space limitation, and limited communication bandwidth. In this work, an acoustic based approach for wheel bearing fault detection and isolation is developed to overcome these limitations. Since the bearing noise is a precursor of bearing failure, the proposed method is a prognosis solution. The whole solution is verified using the data collected from a production vehicle. The results show that the proposed method can predict the wheel bearing failure with promising accuracy and robustness.