Dynamical Characterization Of An Industrial Machine Through Impedance Spectroscopy Linear Estimation Of Damages

Impedance spectroscopy is a technique widely used in Condensed Matter Physics to characterize the dielectric response of materials. This technique is very sensitive to many types of defects inside the material. We show in this study that it can be used to probe ceramic based ball bearings, part of a tooling machine. A model taking into account both the dielectric properties of the balls and the mechanical conditions at the interface between the balls and other parts of the machine is proposed. The model allows an electrical representation of the mechanical state of the machine. We finally show how the transition to the probabilistic language through Fokker-Planck's dynamics allows to predict the ball bearings impedance and the way it is influenced by the fatigue and/or defect processes. Routes to the building of a linear estimation tool (Wiener-Hopf filtering) are described with the final goal of a predictive maintenance tool.