Fault-Tolerant Robust Model-Predictive Control of Uncertain Time-Delay Systems Subject to Disturbances

• Published in: IEEE transactions on industrial electronics (1982) Vol. 68; no. 11; pp. 11400 - 11408
• Main Authors: Khan, Owais, Mustafa, Ghulam, Khan, Abdul Qayyum, Abid, Muhammad, Ali, Muhammad
• Format: Journal Article
• Language: English
• Published: New York IEEE 01.11.2021; The Institute of Electrical and Electronics Engineers, Inc. (IEEE)
• Subjects: Actuator faults; Actuators; Continuously stirred tank reactors; Control stability; Disturbances; Fault tolerance; Fault tolerant systems; fault-tolerant control; Feedback control; Optimization; Parameters; Performance degradation; Performance indices; Predictive control; Predictive models; Process controls; Robust control; robust model-predictive control; robust stability; Time delay systems; Tracking control; Tracking errors; uncertain systems; Uncertainty; Upper bounds
• ISSN: 0278-0046; 1557-9948
• DOI: 10.1109/TIE.2020.3029469

Time delays, model uncertainties, faults, and disturbances are frequently the main causes of performance degradation and instability in industrial process control. This article presents a fault-tolerant robust model-predictive control design for processes that involve the above effects and process constraints. The uncertainties are modeled into a polytopic affine form based on variations in the system's parameters. A new model based on augmentation of the state variables and tracking error is used that provides increased degrees of freedom for the control design and guarantees tracking performance. A parameter-dependent Lyapunov-Krasovskii functional is used to design a state-feedback control that reduces the conservatism of the conventional approach and ensures robust stability and tracking performance of the closed-loop system. At each sampling instant, a control action is computed by solving an online constrained optimization problem that minimizes the upper bound of the "worst-case" performance index. Finally, the proposed framework is employed for the fault-tolerant control of a continuous stirred tank reactor to demonstrate its effectiveness in mitigating actuator faults and tracking the desired set point.