Based on the Lyapunov stability theorem, an adaptive variable structure observer and a controller are proposed in this thesis for a class of mismatched perturbed multi-input multi-output (MIMO) dynamic systems with unmeasurable states to solve regulation and tracking problems. In order to estimate the unmeasurable states, a design methodology of variable structure observers is presented first. Then the controller is designed so that the trajectories of the controlled systems are able to reach sliding surface in a finite time. Some adaptive mechanisms are embedded in the sliding surface function and sliding mode controllers, so that not only the mismatched perturbations are suppressed effectively during the sliding mode, but also the information of upper bounds of some perturbations are not required. When the controlled system is the sliding mode, the stability or asymptotical stability is guaranteed. A numerical example and a practical example are given to demonstrate the feasibility of the proposed design technique.

No description available.

No description available.

Thesis (Ph. D.)--University of Texas at Austin, 2006. / Vita. Includes bibliographical references.

No description available.

This thesis presents a torque control mechanism for switched reluctance machine (SRM) drives. The proposed mechanism is capable of maintaining ripple free torque control while minimizing the copper loss or mode-0 radial force or both at a fixed switching frequency. In the proposed approach, the torque control problem is addressed by splitting it into two parts. The first part consists of identification of optimum phase current references while the second part incorporates the design of an efficient current controller. For the identification of optimum phase current references, three algorithms are presented in the form of a developmental process. The nature of the online optimization problem is demonstrated using a simple 2-dimensional gradient descent method. Subsequently, a parametric form gradient descent algorithm is presented which transforms the original optimization problem into two 1-dimensional problems, viz. torque error minimization and identification of optimum search direction. This method yields improved computational efficiency and accuracy. The third algorithm incorporates projection using equality constraint on the phase torque contributions to achieve a 1-dimensional solution process. Although this algorithm takes more iteration as compared to the parametric form gradient descent algorithm, it demonstrates greater accuracy and computational efficiency. A comparative analysis of these algorithms is performed in at different operating conditions in terms of the torque ripple magnitude and computational effort. The thesis also presents a comprehensive analysis of well known control techniques for application in SRM current control in discrete-time domain. This analysis also presents a comparative evaluation of these control techniques under different operating conditions. On account of this analysis, several recommendations pertaining to the performance improvement are presented. Finally, a digital sliding-mode based model-free current controller suitable for fixed switching frequency operation is presented. The proposed controller is capable of providing a consistent dynamic response over wide operating range without utilizing any model information. The reference current tracking performance of this controller is verified through simulation studies in MATLAB/Simulink® environment and over a 1.2kW, 100V, 2500RPM, 12/8 experimental SRM drive. / Thesis / Doctor of Philosophy (PhD)

No description available.

No description available.

No description available.

No description available.