While most of the problems related to the feed- back control of AC motors have been successfully solved, the condition monitoring and the detection of their most common faults is still an open problem.
In this work a novel approach to the fault diagnosis of AC motors is introduced. The detection and the identification of a fault is performed through the analysis of the tracking errors (errors between reference and observed quantities) evaluated by a non- linear observer.
The description of the diagnostic method will be given and extensive simulation results concerning different control schemes will be reported. Finally some experiments on a real induction motor drive will be described, that confirm the results obtained in the simulations.
