Quantitative damage diagnosis for rolling bearing based on contact vibration model and optimum kurtosis-Laplace wavelet

The positive and inverse problems of quantitative damage diagnosis for rolling bearing are studied. The physical model was characterized according to the vibration transmission path by studying the interactions of bearing interface based on dynamical contact theory and tribology. The influences of the load distribution and play, and the contact slippage, separation, local fault enlargement depicted by surface profile changes were considered in the model. Model parameters were modified based on the measured signals, thus integrating active service with numerical simulation in the case of real operation. By solving the positive problem the optimum kurtosis-Laplace wavelet was constructed to obtain the dimensionless index, thus forming the damage matrix database. The inverse problem could be transferred into multi-dimension optimization one. Using optimization for grey degree of association the samples which are the most close to the inputs were obtained and furthermore the damage location and size could be determined. Practical experimental data verifies the validity and robustness of the quantitative diagnosis method, which is easy to popularize for engineering application to diagnose the defects in bearing system.