A knowledge based system for the selection of thermodynamic models

Abstract An expert system for choosing optimal physical property and phase equilibria models for chemical process design and simulation is presented. Based on user specified problem type, chemical substances and expected thermodynamic state conditions, a set of selection indices are extablished using rules developed from theory and practical experience. The indices provide paths through taxonomies to select appropriate models and to access required parameters and data. A description of the methodology and illustrative examples are given using current techniques and data banks. Straight forward extension to new chemicals and models, as well as ways to recognize model inadequacies, are discussed.