Research on a novel online condition monitoring technique for induction machinery

Condition monitoring of induction machinery is traditionally implemented by conducting spectral analysis of either vibration or stator current signals, because it is believed that once a fault occurs, it will produce a unique characteristic frequency in the signals. Hence, the fault and its further growth can be detected and traced through observing the variation tendency of the amplitude at this fault-related frequency. Such a condition monitoring approach requires significant pre-knowledge about the machine and its components, however does not always work very well in practice, in particular when the fault is in its infancy. In addition, to date there has not been convincing proof demonstrating that electrical faults can be readily detected by the means of vibration analysis, although much effort has been done to prove the validity of stator current analysis in detecting the mechanical faults occurring in induction machinery. In view of this, a new online condition monitoring technique is developed in this paper dedicated for induction machinery, which is based on detecting the phase angle of the stator current with respect to the corresponding voltage. In the paper, the proposed technique is verified through both simulated and practical experiments. It is shown that the proposed technique is not only valid in detecting electrical and mechanical faults occurring in the induction machine, but is also able to distinguish stator winding faults from rotor winding faults without requiring any pre-knowledge about the machine. Moreover, the proposed technique uses relatively simple calculations and is therefore ideally suited for performing the condition monitoring task online. (5 pages)