A bearing fault detection method based on compressive measurements of vibration signal

The general method for bearing fault detection is achieved by using bearing vibration signals which sampled in the frame of Shannon sampling theory. So it is necessary to sample and save abundant original vibration data in the process of uninterrupted monitoring, and this will generate masses of original data which would burden the storage and transmission. For this issue, a fault detection method based on compressed sensing theory is proposed in this paper. It only needs to sample and save fewer compressive measurements of bearing vibration signal directly compared to original signal. There is no need to recover the original signal accurately for detecting bearing faults, while it just requires referring to the prior training result and reconstructing the overall energy distribution of the original signal in some transform domain. The availability and effectiveness of the method proposed is validated with bearing vibration signals sampled in practice.