Abstract:
The induction motors are electro-mechanical devices used in industry applications. These
motors are robust and used for general purpose as well as in severe environment. Some of the
applications are conveyors, elevators, presses, packing apparatus, natural gas plant and coal
plant equipment. These motors require very low maintenance and are very highly reliable. In
terms of power it comes from hundreds to mega watts, which make them suitable for many
industrial processes. It is necessary to identify a fault in induction motor at the initial stage to
avoid severe damage of induction motor as well as unwanted shutdown of production. For the
detection of broken rotor bar fault and inter-turn short fault, an online detection technique
which is parameter estimation using three different non-linear estimators is implemented. All
the three estimators successfully estimate the parameters with low error in different load
conditions of induction motor. The estimated values are used to detect the occurrence of fault
by comparing with the reference values.
The non-model based technique, motor current signature analysis is also implemented. Fault
of broken rotor bar, bearing related faults and fault of inter-turn short in stator windings is
successfully detected using this technique. A GLRT detector is also designed for detecting
low amplitude signal in power spectrum of stator current. The detector is performing very
well and detector performance is shown using ROC curve.
