Modelling and validation for control design uncertain systems, monotone bode plots and process control

Publisher Website
Google Scholar
Abstract
The interplay between modelling and control design is studied in the context of typical plant dynamics in industrial process control. An abstract way of summarizing the plant dynamics in most examples in textbooks on process control and in many applications is provided by the concept of stable models with monotone Bode plots. It is shown that for this class of systems approximate knowledge of a small part of the plant frequency response provides sufficient information for successful control design from experimental data. The role of adaptive information is studied in experimental modelling of systems with monotone Bode plots for control design. Frequency domain data is considered in the present tutorial style paper only. Both open-loop and closed-loop experimental conditions are included. In fact, it turns out that a strong form of adaptive information can be useful, reminescent of dual control where the input signal is to guarantee a learning property, too. Furthermore, the role of uncertain systems, model validation and control design validation are discussed.

This publication has 12 references indexed in Scilit: