Expert System for Bainite Design: the Approach to Enrich Physical Models with Information Derived from Knowledge Models

The development of a physical model begins with a knowledge model, initially existing as ideas in the mind of a researcher. A transition from knowledge models to strict mathematical formalisms is a challenging process, and may not always be feasible, particularly in the early stages of research. Another problem comes when many experts are participating in the development of new physical knowledge, which may result in inconsistency. To contribute to this domain, the paper presents the development of an expert system (ES), created to capture expert knowledge for the design of a new physical material, namely, the bainite steel. The ES combines physical properties and rules in a unique knowledge model and enriches them by derived from data probabilities. The proposed approach enables users to validate expert knowledge and find contradictions in the logical rules, giving the possibility of mapping them back to physical models.