Process Control-Relevant and Closed-Loop Identification

摘要

The identification of dynamical systems on the basis of data, measured under closed-loop experimental conditions, is a problem which is highly relevant in many (industrial) applications. When using models as a basis for model-based robust controldesign both nominal models and model uncertainty bounds are required. In this paper it is shown how -in particular- model uncertainty bounds can be obtained from closed-loop experimental data in the classical prediction error identification framework. The considered uncertainty structure is adjusted so as to allow direct evaluation of the performance robustness of both the actual and a to-be-designed controller.