Evolution behavior on friction-induced vibration and lubrication state identification

Abstract

In this study, the tribological performances and vibration behaviors under varying lubrication states and loads of GCr15 steel were investigated, revealing the mechanism of friction-induced vibration. As the results, lubrication state yielded the lowest friction coefficient, minimal roughness with mild adhesive wear, corresponding to vibration exhibited the lowest RMS values and low-frequency spectrum. While under insufficient lubrication and dry friction, tribological parameters increased significantly with evident abrasive wear even fatigue wear, the RMS values raised, and the spectrum shifted toward higher frequencies. With the increase in loads, tribological parameters exhibited an upward trend, leading to more severe friction-induced vibrations. Furthermore, CNN-SVM hybrid algorithm was proposed, the accuracies of single-state identification and multi-load multi-state datasets were 100 % and 88.53 %, respectively.