Status and challenges of intelligent plant control

Abstract

Intelligent plant control is a rapidly expanding research area that is nourished by the industrial interest in improving the efficiency and safety of plant operations, by the general academic interest in developing theories for control of complex systems and by results from AI research. Currently a variety of techniques for building advanced controls has been developed but there is no general consensus on the meaning of basic terms and there seems to be a limited recognition of the importance of common sense engineering knowledge in realising controls. In the paper it is argues that common sense knowledge of the purpose of plant components and subsystems play a significant role in building dynamic models, in diagnosis and planning and in the design of supervisory displays for the human supervisor. The lack of understanding and formalisation of common sense engineering knowledge within the control domain limits the levels of plant automation that can be achieved and impede the development of methods for design of supervision and control systems. Recent research results of the author on the modelling of plant goals and functions (Multilevel Flow Modelling) show that modelling research that supplement and expands the traditional modelling paradigms of control theory may lead to significant advances in the design of intelligent plant controls.