Lubrication State Monitoring of Journal Bearings Based on Vibration Features

Journal bearings are important components of rotating equipment, and wear is one of the important factors that affect the performance of bearings and trigger early failures. Although vibration monitoring is widely used in bearing fault diagnosis, it has been less developed in wear analysis. To detect the friction state of journal bearings, the research aims to use vibration features to characterize the real-time friction dynamics of journal bearings. The acceleration data of different radial loads, rotational speeds, and oil viscosity were collected on a rotor bearing test rig. Relying on the vibration analysis of short-time Fourier transform (STFT) and wavelet packet decomposition (WPD), the time-domain feature parameters kurtosis and root mean square (RMS) are analyzed. According to the strong correlation of vibration generations with the Stribeck curve, the experimental results show that the feature parameters can well reflect the lubrication regimes under different working conditions, which provides important information for online monitoring of the wear analysis of journal bearings.