Robust adaptive neural network event-triggered compensation control for continuous stirred tank reactors with prescribed performance and actuator failures

Abstract

This paper studies the event-triggered compensation tracking control problem of the Continuous Stirred Tank Reactor (CSTR) with actuator failures. By actuator redundancies, a novel adaptive neural network (NN) event-triggered controller (ETC) design scheme is proposed based on the switching threshold event-triggering mechanism (SWT-ETM). To constrain the maximum overshoot and the convergence rate within given specifications, a prescribed performance function (PPF) error transformation is employed. It is shown that the tracking error will exponentially converge to an adjustable neighborhood of zero with prescribed transient performance, despite the presence of failures, system nonlinearities, parametric uncertainties and external disturbances. Besides, the system’s burden is significantly alleviated requiring less system resources for signal transmission and actuator execution to decrease the occurrence rate of the actuator failures; and the minimum inter-event interval is guaranteed to be positive to avoid the Zeno phenomenon. Simulation results illustrate the effectiveness of the proposed adaptive NN ETC scheme.