ABSTRACT
Analysis of process characteristics and inter-variable relationships is of paramount importance in prediction, control, monitoring, design and innovation of process systems. A key step in these analyses is the development of a (mathematical) description of the process under study, known as the model. Two contrasting approaches are generally followed for model development: (i) a theoretical (first-principles) approach that is based on fundamental laws of matter and energy, and (ii) an empirical approach that is based on analysis of observations (experimental or operating data). The latter approach is a highly practical alternative to the former and widely followed since most processes are too complex to be understood at a fundamental level. Observations potentially carry a wealth of information that remains otherwise obscure in a first-principles approach. The subject of System Identification is concerned with the means and techniques for studying a process system through observed / experimental data, primarily for developing a suitable (mathematical) description of that system.
