Improved Networked Iterative Learning Fault-tolerant Control Algorithm for Systems with Time-delays, Random Packet Losses, Limited Communication and Actuator Failure

This paper is devoted to develop an improved networked iterative learning fault-tolerant control (NILFTC) algorithm for a class of systems with time-delays, random sensor-controller-channels packet losses, limited communication, external disturbances and random actuator failure. Firstly, the state variable, the Markov chain process of random packet losses, network communication delay, limited communication and actuator failure are introduced to establish an extended state-space model of the system. Secondly, based on the dynamics model of the system and combination of stability theory for linear repetitive processes and linear matrix inequality (LMI) technique, robust monotonic trial to trial convergence for NILFTC is specified in terms of LMIs and satisfy H ∞ performance. Thirdly MATLAB optimization based on Yet Another LMI Parser (YALMIP) is applied to improve the performance of the NILFTC scheme, in which the convergence of system output tracking error is converted to a convex optimization problem and the constraints are convex, and the optimal iterative learning gain of the NILFTC is determined by solving the optimization problem. Finally, the feasibility and effectiveness of the proposed design method is illustrated on a dynamic hydro-turbine governing system model based on Matlab/Simulink and TrueTime toolbox.