Robust actuator fault detection for an induction motor via genetic-algorithm optimisation

In this study, an application of genetic algorithm optimization-based fault detection for an induction motor with actuator faults is addressed. The frequencies of the modelling errors are identified from the Fourier transform of the measurement outputs, and a multi-objective cost function is constructed for minimizing the effects from the dominant modelling errors components while maximizing the influences from the actuator faults to the residual. Abrupt faults and incipient faults are both investigated, also the effectiveness of the proposed fault detection methods are demonstrated by using experimental/simulation studies of the induction motor.