Fault-tolerant Model Predictive Control Strategy Applied to Industrial Processes

Abstract

This paper presents a strategy to address the income of actuators faults using a model-based predictive controller scheme, which allows to track a reference even in the presence of actuator faults. The proposed fault-tolerant control system adopts a model predictive control technique to design a reconfigurable fault-tolerant controller and a reduce-order observer to achieve the fault detection and diagnosis function. Simulation results, based on two typical chemical industries processes, are given to illustrate the use and performance of such approach.