A combined fault-tolerant and predictive control for network-based industrial processes

This paper investigates the tracking and optimization problems for a class of industrial processes by utilizing output feedback fault-tolerant control (FTC) and predictive compensation strategy. At device layer, the tracking problem for device layer subsystems which subject to random failures and random network-induced delays is investigated. These two different random processes are modeled as Markovian chains. Device layer controllers are designed to guarantee the tracking performance at $H-{\infty}$ disturbance attenuation level. At operation layer, a nonlinear model predictive control (NMPC) strategy is proposed to stabilize the upper operation layer system. Then by considering the effect of radial basis function (RBF) performance index and random packet dropout phenomena, a predictive compensator is designed to guarantee the input-to-state practical stability (ISpS) of the resulting system. In addition, networked flotation processes are considered in the simulation part, and the simulation results further demonstrate the effectiveness of the proposed method.