No abstract

Purpose The purpose of this paper is to study the dynamic features of friction coefficient during running-in state based on recurrence analysis, so as to recognize the running-in state of crankshaft journal bearings. Design/methodology/approach The friction coefficient was measured in the friction experiments and the dynamic features are analyzed by recurrence plots (RPs), unthreshold recurrence plots (URPs) and recurrence quantification analysis. Findings During the running-in process, RPs have gone through disrupted patterns, drift patterns and homogeneous patterns successively. URP shows that the phase trajectory spirals in the disrupted pattern gradually converge in the drift pattern and remain stable in the homogeneous pattern. Three independent measures, recurrence rate, entropy and laminarity, are chosen to characterize friction coefficient from the perspective of point, diagonal line and vertical line structures of the RPs. Originality/value The results provide a feasible way to monitor the running-in process and recognize the running-in state.

Friction coefficient is difficult to measure in real application, whether it can be replaced by friction coefficient for dynamic analysis is an urgent issue to be solved. To study the correlation between the two signals, the friction tests were carried out under dry friction, starved and flooded lubrications, respectively. Cross recurrence plot (CRP) and its measures (recurrence rate, determinism, average diagonal line length and entropy) were used to explore the correlation. The results show that the more obvious the diagonal structures in CRPs, the stronger the correlation between the two signals, and average diagonal length is more sensitive to the degree of correlation. Irrespective of lubrication conditions, friction coefficient and vibration measured in a friction system have similar recursive characteristics. Thus, friction coefficient or vibration can be selected according to the convenience of measurement to monitor the running-in process.

The dynamic evolutions of friction force and worn surface profile were qualitatively analyzed by phase trajectory and recurrence plots and quantitatively characterized by fractal dimension and percent determinism. The results show that phase trajectories first shrink to a small volume, then stabilize at a minimum volume, finally expand to a large volume in the wear process. The white areas on the recurrence plots increase with the wear time. The fractal dimension first increases, then stabilizes at a high value, and finally decreases rapidly. The percent determinism first decreases, then fluctuates in a certain range, and finally increases. It demonstrates that friction force and worn surface topography derived from one tribology system evolve in a similar but not exactly the same way. They have the consistent evolution law in the wear process. Specially, friction force is much more sensitive to the variation of wear states than the worn surface.

No abstract