Semantic Networks Based Design of Electric Drives

Abstract

Based on the study of projects of electric drives with complex mechanisms and frequency converters (FC), artificial intelligence tools are proposed for modeling and analyzing the consistency and validity of technical solutions in different engineering domains. The developed design methodology takes into account design features of the mechanisms and describes the process of creating and transforming descriptive semantic models of electric drives. Different forms of such models on the one hand provide a visual representation of design factors, on the other allowing the use of a known mathematical apparatus to automate the search and analysis of design decisions. For electric drive requirements modeling SysML language is used. Further obtained model is converted into a semantic network in the form of a conceptual graph (CG). This graph has a mathematical description in the form of logical expressions and can be used by an inference machine for automatic analysis and synthesis of technical solutions. Other knowledge models of electric drives design are considered. To take into account the inconsistency of the requirements, the model in form of conceptual graph is transformed into description necessary for applying the GORE methodology tools. To take into account the incomplete certainty of design factors, a Bayesian network is built on the basis of the conceptual graph. The proposed set of semantic modeling tools is aimed at ensuring the completeness of the technical description of electric drives and automating the stage of their conceptual design.