MCSA-based fault diagnosis for rotary parts in servo motion system: Approach and experimental study

Abstract

Motor current signature analysis (MCSA), as a non-invasive diagnostic method, is robust to environmental noise and of less sensor cost than vibration-based monitoring. Although large amount of progress has been achieved, little attention is paid to the MCSA-based diagnosis task in the Servo motion systems (SMS). An approach using Park vector demodulation, comb filtering and the ensemble empirical mode decomposition (EEMD) is proposed for cyclic fault event detection in this paper. Experiment studies reveal that the MCSA approach could extract the fault signature under extremely low-speed or noisy working condition, which has potential in the scenarios where the accelerometers placement is limited or affected by interferences.