Peaking phenomenon inhibition in high gain observer design: Application to speed-sensorless induction motors

Abstract

Peaking phenomenon (pp) which usually takes place at transient regime, presents a serious hindrance that should be handled when output feedback controller is designed. To cope with the pp issue for uniformly observable class of multi-output nonlinear systems, we propose in this paper a slight modification on the correction term of the standard high gain observer (SHGO) by the incorporation of nested saturation functions. The design of peaking free observer is thus carried out in the case of sampled multi-outputs leading to a “continuous-discrete non-peaking HGO” (CD-NPHGO). Of fundamental interest, it is shown that the proposed scheme ensures a cancelation of large peaking at starting operation while preserving SHGO benefits in terms of arbitrarily fast convergence at the steady state. An application to the speed-sensorless control of induction motors (IM) highlight the features of the proposed approach in the ability to handle large ripples at transient as well as achieving accurate state estimation in very short time.