Search

In this article, a novel active fault‐tolerant control (AFTC) design method is developed by using the switching linear parameter varying (LPV) controller with inexact fault‐effect parameters. The LPV controller is not only considered as a gain‐scheduled controller varying with the operating points, but also an AFTC controller varying with the fault‐effect parameters, which can be obtained via the fault estimator in real‐time. The inexact measurement of fault‐effect parameters is considered for the first time in this article. To cover large parameters variation, a class of switched LPV fault‐tolerant controller is designed to work in the multiple partitioned parameters subregions. The dynamic output feedback controller model is constructed to perform the switched LPV fault‐tolerant control task under the mode‐dependent average dwell time constraint. By using the multiple parameter‐dependent Lyapunov function approach, sufficient conditions with less conservatism are obtained, such that the corresponding closed‐loop systems are globally uniformly exponentially stable and satisfy an upper bound of weighted l2‐gain performance indexes. Finally, effectiveness and applicability of the developed approaches are validated via an active magnetic bearing system. [ABSTRACT FROM AUTHOR]

Abstract: Based on a two-dimensional (2D) Fornasini–Marchsini system description of a batch process in industry, a robust state feedback integrated with an iterative learning reliable control (FILRC) scheme is proposed. The scheme is intended for batch processes with uncertain perturbations and state delay subject to actuator failures, which is dependent on the upper and lower delay bounds of the interval time-varying delay. The relevant concepts of fault-tolerance are introduced. The proposed control law can guarantee closed-loop convergence along both time and cycle directions to satisfy H ∞ performance even with unknown disturbances and actuator failures. By introducing a new 2D Lyapunov–Krasovskii functional candidate and adding a differential inequality without introducing redundant free-weighting matrices to the difference Lyapunov functional for 2D systems possessing two directions, conditions for the existence of the proposed FILRC scheme are established in terms of linear matrix inequalities (LMIs). By solving these LMIs, the FILRC law is explicitly formulated together with an adjustable robust H ∞ performance level. Applications to injection velocity control show that the proposed FILRC achieves the design objectives well. [Copyright &y& Elsevier]