The Detection of Unbalanced Faults in Inverter-Fed Induction Machines

Fault diagnosis of electrical machines has primarily been focused on steady state operation of the machine due to the limitations in signal processing techniques. However, the advancement in this area has led to the development of a new class of algorithms which are time-frequency based signal processing techniques. In this paper we explore the concept of identifying faults by analyzing the transient machine currents. We apply a new transient technique which extracts the nonstationary fundamental current component and analyzes the residual currents using wavelets. It will be shown that the technique shows a considerable improvement in detecting unbalanced faults in drive-connected induction machines.