A New Observer-based Model- Reference Approach for Nonlinear Discrete-Time Systems with Application to Power Systems

Abstract

In this paper, a novel observer-based modelreference technique is developed to control the performance of nonlinear discrete-time systems. Firstly, a linear or nonlinear model is chosen for which the dynamics are designed to satisfy the pre-specified performance behaviours of the nonlinear system as stated by the designer. The proposed approach is then used with the system to fulfil a point-by-point tracking of the pre-designed model trajectories. To achieve this goal, an observer is firstly used to predict the states and the output(s) at the next sampling point of the system to be controlled from the received set of measurements. The differences between the achieved predicted estimates and those of the model are then used to generate the necessary online closed-loop control actions to force the system trajectories to track those of the model at the next sampling instant of time, as desired. The stability of the closed-loop control system is investigated. To check the performance of the developed approach, it is then used to control the performance of a synchronous machine and a permanent magnet synchronous motor. The achieved results are presented to illustrate the effectiveness, simplicity, and applicability of the proposed design procedure