Sensitivity and Robustness with Constant Controllers

Abstract

The preceding controller design methods assumed that the process model is exactly known. However, this is never the case in practice. In theoretical modelling as well as in experimental identification one must always take into account both the small and often the large differences between the derived process model and the real process behaviour. If, for simplicity, it is assumed that the structure and the order of the process model are chosen exactly then these differences are manifested as parameter errors. Moreover, during most cases of normal operation, changes of process behaviour arise for example through changes of the operating point (the load) or changes of the energy- mass- or momentum storages or flows. When designing controllers one must therefore assume

• the assumed process model is inexact;

• the process behaviour changes with time during operation.