Non-intrusive methods to detect air-gap eccentricity faults in three-phase induction motor

The focus of this paper is model-based behavioural analysis for the diagnosis of the air-gap eccentricity fault condition in the three-phase induction machine. To facilitate reliable and sensitive diagnosis, an online condition monitoring system has to depend on multiple signals and signatures. Air-gap mixed eccentricity fault signatures derived from three non-invasive methods such as stator line current, instantaneous power, and estimated air-gap torque are presented. Modified winding function theory and multiple coupled circuit approaches are used to model the machine. The sidebands present around the fundamental frequency in the spectrum of stator current, and around double the supply frequency in the spectrum of instantaneous power indicate the presence of air-gap mixed eccentricity. The low frequencies which exist near the DC component in the spectra of the air-gap torque and instantaneous power also indicate the same. These specific components were also observed in the high-frequency spectra around the principal slot harmonics. The modelling approach, the torque estimation approach, the simulation results at different severity and load conditions, and the experimentation result in a motor with prefabricated eccentricity are presented.