Vibration modeling of rolling bearings considering compound multi-defect and appraisal with Lempel-Ziv complexity

The vibration modeling is a key link to explore the cognitive rules in fault diagnosis. In order to foresee the vibration characteristics accurately and efficiently, the transfer path and dynamic contact among joint interfaces were well studied. A novel methodology for dynamic modeling of 6DOF systems consisting of bearing inner race, outer race and housing was presented based on the theories of Hertzian contact. The progressive processes of single and compound multi-defect were provided according to the practical kinematics of rolling bearings. The model describes the vibrations in time and frequency domains. Computed results from the model were according to the validated with experimental results, which were generated on defective deep groove ball bearings. The complexities of vibration signal change with the emergence and growth of faults. Based on nonlinear dynamics theory, the Lempel-Ziv complexity measure was used to characterize the complexity of single and compound multi-defect signals. The results show that the complexity measure used as a quantitative criterion can effectively appraise the running condition of rolling bearings.