Fault-tolerant control by means of moving horizon virtual actuators: Concepts and experimental investigation

摘要

This work presents a moving horizon virtual actuator (MHVA) as an extension of the classical virtual actuator technique. Featuring the same properties of the well-known model predictive control (MPC), the MHVA represents a viable candidate to provide practical use of fault-tolerant control in the process industry, where MPC technology is widely applied. Experimental application of the proposed technique was demonstrated in a pH neutralization process subject to actuator fault scenarios such as lock in position and loss of effectiveness.The proposed strategy allowed the system to recover performance under fault conditions and restore the overall nominal behavior, while coping with operational constraints and allowing the nominal controller to remain functional.