Damage assessment of mechanical systems based on recurrence quantification analysis

This paper presents a new application of recurrence quantification analysis (RQA) for damage severity assessment of mechanical systems. As a nonlinear time series technique, the recurrence quantification analysis derives quantitative measures from the recurrence plot (RP) to quantitatively characterize changing dynamics of a given system in the reconstructed phase space. In the present study, four quantitative measures, recurrence rate (RR), determinism (DET), laminarity (LAM), and RP entropy (ENTR), are investigated as quantitative indicators for damage severity assessment of rolling bearings. Based on analysis results of experimental data acquired from a series of SKF 6205 bearings with different damage locations and under different operation conditions, ENTR and LAM are found to be suited for evaluating their damage severities.