Iterative Learning Control Of Time-delay Systems

In the practical process, phenomena of time-delay exist widely. That often results in poor performance even leads system to unstable behavior. It is meaningful and valuable to study the control problem for time-delay systems.In this thesis, a few iterative learning controllers are designed for several different types of time-delay systems, and the relevant theoretical analysis and simulation is studied. The main contents are concluded as follows:(1) The closed-loop PD type learning algorithm for iterative learning control systems with time-delay in state is proposed. The PD type learning algorithm for iterative learning control systems with time-delay in control is proposed. The PD type learning algorithm with initial rectifying action for iterative learning control systems with time-delay in both control and state is proposed. The convergence conditions of these proposed methods are investigated respectively. The effectiveness of these algorithms is verified by simulation.(2) The iterative learning control algorithm together with Smith predictor for batch processes with time-delay integrating object is proposed and analyzed in the frequency domain. Convergence conditions of the proposed method are stated and proven. The effectiveness in reference tracking and disturbance rejection of the proposed algorithm is verified by simulation.(3) For a class of first-order nonlinear systems with time-delay, unknown parameters and unknown disturbance, an adaptive iterative learning control (AILC) algorithm is proposed. Time-invariant and time-varying parameters are estimated in time-domain and iteration-domain respectively in the algorithm, which makes full use of time-domain and iteration-domain information, meanwhile, the disturbance is inhibited by robust control. The algorithm is then extended to the second-order nonlinear systems with time-delay by means of backstepping technology. The effectiveness of these algorithms is verified by simulation.