Simulation-based model optimization for a steam-filled chamber. Part I: Open-loop identification

Abstract

This paper presents a study of simulation-based model optimization for an industrial application of a steam-filled chamber under different operating conditions. Open-loop input/output process data are used to optimize selected linear models of different complexity in the form of continuous transfer functions. The optimization is performed using the MATLAB computing system and its toolboxes for simulation and optimization. More specifically, the nonlinear programming solver fmincon has been fruitfully utilized in this study for the task. Based on the suggested criteria, a suitable model in the form of a second-order astatic system with time-delay has been chosen as a trade-off between its simplicity and fidelity, which is confirmed by experimental comparison. As the modelled process is non-linear in nature, the resultant model parameters vary for different process conditions. The results of this work are further usable for both building a simulation testing model of the system and for the subsequent step – control system design and tuning purposes. The suggested solution enables to obtain a relatively simple but control-relevant model of the investigated process, linking directly controlled and control input variables, which is advantageous from the control system design point of view. © 2021 Praise Worthy Prize S.r.l.-All rights reserved.