| 초록 |
In order to predict the binary and ternary azeotrope formation between components in mixtures and the types of formed azeotropes, a hierarchical knowledge-based expert system, called AZEOPERT was constructed. The specific knowledge included in AZEOPERT and the implemention of AZEOPERT are also described along some of the several components of expert system applications : knowledge representation strategy and levels of knowledge abstraction, inference machine, user interface, explanation facilities. The AZEOPERT knowledge base can be largely divided into the knowledge base for the prediction of binary azeotrope formation and the knowledge base for the prediction of ternary azeotrope formation. Each of the knowledge bases consists of the different knowledge levels, arranged in a hierarchical structure. The predictive capabilities and generality of AZEOPERT can be highly enhanced through the integration of different kinds of domain knowledge into hierarchical structure. When faced with binary or ternary systems that are not predicted with compiled reasoning, AZEOPERT can successfully predict the azeotrope formation with model-based reasoning. In this manner, AZEOPERT may predict the azeotrope formation in a wide variety of mixtures, especially, in unknown azeotropic systems.
|
