Identification for control: simple process models

Abstract

Identification for control in industrial practice most often means that a simple process model with two, three parameters are adjusted to a step response. These simple models are then used to tune PI- and possibly D-parameters in a basic controller. In this contribution we consider such simple process models from a system identification perspective. Some features of particular importance are the estimation focus and the input inter-sample behaviour, and we illustrate some of these aspects.