Enhancing Model-Based Networked Control of Uncertain Process Systems Through Error-Triggered Parameter Re-Identification

摘要

This work focuses on the problem of stabilization of process systems controlled over resource-constrained communication networks, subject to process parametric uncertainty, limited state measurements and sensor-controller communication constraints. A methodology for augmenting model-based output feedback control with error-triggered parameter reidentification is developed. The goal is to maintain closed-loop stability in the presence of varying levels of plant-model mismatch during periods of process drift, while simultaneously keeping the rate of sensor-controller communication over the network to a minimum and maintain acceptable levels of closed-loop performance. The model-based controller utilizes observer-generated state estimates to periodically update the model states at the sensor-controller communication times. An error monitoring scheme with a time-varying instability alarm threshold is devised to determine on-line when model parameter updates are needed. Upon breach of the instability threshold at some time, process operation is temporarily safe-parked by increasing the sensor-controller communication rate to counteract the destabilizing influence of increased plant-model mismatch. The input and output data collected during the safe-parking period are used to obtain new estimates of the plant parameters, and a post-identification characterization of the networked closed-loop stability region associated with the new model is performed to determine the appropriate post-drift sensor-controller communication rates and controller/observer design parameters. The implementation of the developed methodology is illustrated using a chemical process example.