APRC-based decentralised model predictive control for parallel splitting systems with a matrix annihilation

Abstract

A decentralised model predictive control strategy for interconnected process systems having parallel-splitting structure based on the asymptotically positive realness constraint (APRC) is presented in this paper. Parallel masking and transform descriptor approaches have been employed in previous work for this type of interconnection processes. A robust control perspective has been brought to light in this work to resolve the issue of multiple subprocess parallelised-ly decoupled in a mixed connection configuration of dynamically coupled units. An annihilation is employed to cancel out the interactive vectors between interconnected processing units. Simulation for a parallel redundant process system in mining industry is provided to demonstrate the effectiveness of the presented robust control approach to parallelised interconnected systems.