Improved Rotating Speed Estimation and Bearing Fault Diagnosis Using Multi-Channel Vibration Signals

A rotation speed estimation algorithm for motor bearing fault diagnosis based on multi-channel vibration signals analysis under variable speed conditions is proposed in this study. First, the vibration signals are acquired by a tri-axial accelerometer installed on the motor housing. Multiple instantaneous frequencies (IFs) are extracted from the vibration signals using time-frequency analysis, and then multiple rotating angle curves are calculated based on the IFs. The rotating angle curves are selected and fused into a new one with high accuracy. The original vibration signal is resampled in accordance with the accumulated rotating angle curve. Bearing fault type can be determined from the spectrum of the demodulated resampled vibration signal. The method proposed can realize rotating speed estimation and fault diagnosis using only the vibration signals. Hence, it provides a new solution to realize bearing fault diagnosis under speed variation conditions without using a tachometer.