Distributed output feedback control for multi-unit system with delayed multirate measurements

Abstract

The aim of this work is to analyze output feedback distributed control framework based on model predictive control (MPC) for multi-unit systems with multirate and delayed measurements. Primary controlled variable(s) is sampled infrequently, and the measurement is available after a delay. The distributed, unit-specific estimators and controllers communicate through selective information transfer based on process knowledge. The distributed state estimator is designed based on sampled state augmentation method to efficiently compute improved state estimates on arrival of the delayed infrequent measurements. The multirate estimators are appropriately modified to incorporate the disturbance model to handle model-plant mismatch (MPM). The proposed implementation is analyzed using two case studies. A linearized, reactor-separator system is used to demonstrate that multirate estimation is required for offset-free tracking. A more complex variant of the reactor-separator system with recycle demonstrates the ability of nonlinear distributed estimator-MPC framework to incorporate infrequent and delayed primary measurements.