Improved online condition monitoring using static eccentricity-induced negative sequence current information in induction machines

The detection of air-gap eccentricity-related motor faults is a very important function in online condition monitoring of induction machines. Low frequency and high frequency characteristic side-band harmonics are usually monitored in the stator current spectrum to detect the existence of air-gap eccentricities. However, in these techniques, only a single phase stator current spectrum is analyzed even though all three phase supply currents are usually available in the motor control center. In this paper, for the first time, the negative sequence components in supply currents due to the presence of a static eccentricity are proposed as a fault indicator. This paper also proposes a new technique to separate the negative sequence components from their positive sequence counterparts using qd-axis current spectra in a newly selected rotating reference frame. Analysis shows that in this way the oscillating load torque effect can be completely eliminated from the eccentricity effect. Based on an elementary, 2-pole, 3-phase induction motor model with sinusoidal winding distributions, all motor inductances are analytically derived and analyzed. Simulation and experimental results validate the existence of negative sequence fundamental component in supply currents resulting from a static air-gap eccentricity. The variation of magnitudes and phase angles associated with these negative sequence components are also investigated. By monitoring the magnitude of negative sequence components, one could determine the degree of static eccentricity in the presence of a mixed eccentricity. Therefore, the capability and reliability of online condition monitoring of induction machines are improved.