The application of modern signal processing techniques for use in rotor fault detection and location within three-phase induction motors

A commonly used technique for the detection of faults which may occur in three-phase induction motors is to carry out a spectral analysis of the supply current to the motor under investigation. The presence of certain frequency components within the spectral analysis has been shown to be indicative of a fault condition (Hargis et al., 1982). Such techniques are becoming well established and are used regularly to monitor the health of large induction motors which are operated in critical applications within, for example, the nuclear and oil industries. This technique is generally applied to machines when they are operated under steady-state conditions. Recent work however has suggested that greater information may be obtained from a faulty motor if the same parameters which are monitored in the steady state are measured and analysed during transient conditions, such as the initial starting and acceleration periods of the motor. The frequency components which are indicative of motor faults are functions of speed, and hence, under these transient conditions are non-stationary in nature. This paper presents a review of modern signal processing techniques applicable to the analysis of signals whose frequency content is non-stationary. The individual techniques are compared using both test and actual data. Results are presented which identify the technique most appropriate for the task of fault detection in an induction motor under transient conditions. This technique is further investigated with regard to the possibility of locating as well as detecting machine faults.