Feature Extraction for Bearing Diagnostics Based on the Characterization of Orbit Plots With Orthogonal Functions

This paper introduces a new feature extraction method based on the topology of the phase space of the dynamical systems. The method is based on transforming the phase space into the density space and approximating the density with orthogonal functions. The proposed method has been applied to vibration data of a rotating machine measured by proximity probes in order to identify and diagnose bearing defects. As will be shown, the proposed feature extraction method has outstanding capability in characterizing the system response and diagnosing the system, is robust to noise, does not depend on expert knowledge about the system, requires no feature ranking or selection and can easily be applied in an automated process.