The Hierarchy of Models in Chemical Process Control

Chemical processes are large-scale systems that cannot be described adequately by a single model. A three-tiered hierarchy of process models and control problems is presented to illustrate the nature-of the industrial control system design problem and the interrelations among various solutions. The process phenomena under consideration generally exhibit nonlinear, time-varying distributed-parameter dynamics occurring over a continuum of time scales. The key question, then, is "what constitutes an adequate process model?". This paper makes the case for more complex models than the "first-order plus deadtime" types commonly invoked in practice. This case is based on considerations of process nonlinearities and dynamics that occur on multiple time scales.