Friction fault detection methods based on impact signal for plain bearing

Aiming at the problem of plain bearing acoustic emission (AE) signal interference seriously, fault signal separation difficulty, friction fault monitoring difficulty, based on the pulse characteristic of impact signal, a method based on shock signals to monitor plain bearing faults was proposed, which established the relation between impulse signal and bearing friction fault. Firstly, the Kmeans singlular value decomposition (KSVD) dictionary was used to capture impact characteristics, and the time of impact pulse generation was determined. In order to sharpen and enhance impact signal, the impact envelope signal was demodulated by the Hilbert envelope. Secondly, the bell pulse parameter was introduced to fit the single impact envelope signal, and the simplified formula of the strength of the impact envelope signal was further deduced, which was applied to the contact friction fault identification of the plain bearing. Finally, the friction fault of plain bearing was simulated by cutting off oil supply channels. Comparing with the traditional characteristics of AE, the proposed algorithm were easier to identify early friction fault and higher computational efficiency and can completly describe bearing friction process. So, the proposed algorithm provides a new way for the plain bearing contact friction fault diagnosis.