Detection of bearing faults based on band-pass filters and Fourier interpolation of the load torque

Abstract

Bearing faults are widely found in mechatronics systems especially those that are required to work with unbalance loads. A highly reliable mechatronics system requires high quality bearings and/or effective bearing fault diagnostic procedure. This paper deals with the detection of rolling bearing faults based on band-pass filters and Fourier interpolation of the load torque. The reference torque, which is the output of the speed controller, is considered to be an approximation of the load torque. The reference torque is band-pass filtered and then interpolated in terms of Fourier series by using a sliding window method. The Fourier coefficients associated with a healthy bearing are served as a baseline and are compared with faulty lines corresponding to faulty bearings to detect the presence of a bearing failure. The proposed diagnostic method can be conducted online and does not require any additional sensors. Furthermore, the proposed method is able to detect single-point defects whose faulty levels are located at level C of the ISO 10816 Part 3. A mechatronic system equipped with artificial bearing faults is built in the laboratory to verify the effectiveness of the proposed method