Transient analysis of the external magnetic field via MUSIC methods for the diagnosis of electromechanical faults in induction motors

In the induction motor predictive maintenance area there is a continuous search for new techniques and methods that can provide additional information for a more reliable determination of the motor condition. In this context, the analysis of the external magnetic field has drawn the interest of many researchers. The simplicity, low cost and potential of this technique makes it attractive for complementing the diagnosis provided by other well-established methods. More specifically, the study of this quantity during transient operation of the motor (e.g. under the starting) has been recently proposed as a valuable tool for the diagnosis of certain electromechanical faults. Despite this fact, the research in this approach is still incipient and the employed signal processing tools must be still optimized for a better visualization of the fault components and, therefore, for a better determination of the machine condition. This paper presents an advanced algorithm based on MUSIC for enhancing the visualization of the harmonics caused by different motor failures in the electromotive force signals induced by the external magnetic field. Two faults are considered in the work: rotor problems and misalignments. Also, different positions of the external coil sensor are studied. The results prove the potential of the MUSIC algorithm for the reliable diagnosis of electromechanical faults.