Assessment of the rotor condition in soft-started induction motors through the Hilbert transform of transient stray flux signals

In soft-started induction motors, the identification of the signatures caused by fault harmonics in the electrical signals for fault diagnosis purposes is far more difficult than in line-started induction machines. This is partially due to the higher harmonic content introduced by these drives, as well as by the amplification of certain components that are already present in motors supplied from the grid. This paper proposes a non-invasive method that greatly enhances the identification of fault components caused by rotor damages in soft-started induction motors. The method is based on the computation of the envelope of stray-flux signals that are captured under the motor starting using the Hilbert transform and on the subsequent application of time-frequency transforms to the obtained envelope. Unlike other methods, the results of the proposed approach reveal clear signatures linked with the fault that appear for different models of soft-starters and operating conditions. Moreover, the intensities of these signatures increase as the level of fault gets worse, so that reliable fault indications can be introduced for determining the fault severity level. The results shown in this work, obtained with four different models of soft-starters from well-known manufacturers confirm the potential of the proposed methodology.