Decentralized observer-based control of uncertain dynamic systems

Karanjit Kalsi, Purdue University

Abstract

A mathematical model of a dynamic system that accounts for the system uncertainties or disturbances is called an uncertain system model. One class of uncertain systems are large-scale interconnected systems that comprise of linear subsystems and unknown nonlinear interconnections. The proposed research deals with the design of decentralized observer-based controller architectures for the aforementioned class of systems. In the first stage, a decentralized sliding-mode observer based control strategy is proposed to achieve asymptotic stabilization for a class of nonlinear interconnected systems. Sliding-mode observers are also used to reconstruct the unknown nonlinear interconnections between the subsystems. However, the design of the controller and observer are interdependent. In the second stage, a decentralized reduced-order observer-controller compensator architecture is developed where the design of the controller and observer is decoupled. The controller design presented in the first two stages is dependent on a fairly restrictive condition, which, for practical systems may not always be satisfied. To overcome this problem, in the next two stages, Linear Matrix Inequality (LMI) based dynamic output feedback controllers are proposed. The new controller architecture is shown to guarantee the transient stability of a multi-machine power system. In the final stage, a novel approach is proposed for relaxing one of the necessary and sufficient conditions for the design of the sliding-mode and reduced-order observers proposed thus far. This broadens the applicability of the decentralized observer based controllers proposed herein.

Degree

Ph.D.

Advisors

Zak, Purdue University.

Subject Area

Electrical engineering

Off-Campus Purdue Users:
To access this dissertation, please log in to our
proxy server
.

Share

COinS