Bi-objective optimization in a wavelet identification procedure for fault detection in dynamic systems

This paper presents a wavelet-based fault detection approach where wavelet parameters are optimized in order to minimize the residue in nominal operating conditions and maximize it in faulty operating conditions. For this purpose a bi-objective optimization procedure is used. By taking into account the effects of the fault in the proposed optimization procedure, the sensitivity of the detector is improved and the detection delay is reduced. For illustration, a simulated example involving the detection of a sensor fault in a 747 aircraft is presented.