Information Entropy Analysis of Frictional Vibration under Different Wear States

To reveal the chaotic characteristics of friction vibration, pin-plate wear experiments were carried on a wear tester. Different wear states were designed by changing the amount of lubricating oil and distinguished with the friction coefficient. In order to improve the reliability of the collected data, the fictional vibration signals were denoised using an ensemble empirical mode decomposition (EEMD) method. The method of information entropy was proposed to measure the disorder of frequency energy distribution of friction vibration signals. The information entropy of friction vibration decreases in the running-in wear state, fluctuates steadily in the stable wear state, and increases rapidly in the severe wear state. The variation of information entropy of friction vibration signals is closely related to the surface topography of friction pairs. The results indicate that the information entropy of friction vibration can be applied to monitor and recognize the wear state of friction pairs.