A Method for Identifying the Friction-induced Vibration Based on The Maximum Information Coefficient

Abstract

A reciprocating running-in experiment is carried out on a friction-abrasion testing machine with disk-pin friction pair. The friction-induced vibration (FIV) signals measured in the experiments are identified by the maximum information coefficient (MIC) method. Experimental investigation shows that the association strength between the identified tangential and normal FIV signal is in a positive correlation with the coefficient of friction. The two-directional FIV signals distribute in the same frequency range, and their root mean square (RMS) variations are in similar accord to the changing of the coefficient of friction and can indicate the wear state evolution of the disk-pin friction pair from the running-in wear to stable wear. Therefore, the FIV signals can be identified by the MIC method.