Practical Aspects of Hammerstein Models for Nonstationary and Nonlinear Processes

Practical experience shows that in real modeling situations an engineer has to cope with several issues associated with non-stationary behavior, non-linearities, data deficiencies and identification process limitations caused by technology. The paper presents observations and practical solutions developed in such a situation. The goal of the modeling is to derive nonlinear model that captures most of the process characteristics and can be a base for process simulations and further design of alternative control structures. The model under consideration is of Hammerstein type. Proposed modified structure covers process nonlinearities and has embedded self-adaptation mechanism enabling satisfactory operation in non-stationary environment.