A Knowledge-Based Framework for Controller Structure Selection In Intelligent Process Control

Intelligent control principles are used to address the issue of combinatorial complexity in controller structure selection for plantwide control, i.e., the variable seection and pairing problem. A knowledge-based framework combining symbolic representation of heuristics and designer expertise, simple analytical "ranking" criteria, and more elaborate computational routines for selection and pairing is described in this paper. The use of an object-oriented structure leads to a system that is flexible, extensible, and multipurpose, which are features necessary if industrial engieering organizations are to adopt this tool on a routine basis.