Development of an AI-based expert system for the part- and process-specific marking of materials

Abstract

Due to regulations and industry-specific standards, parts of machines, for example in the health care sector, have to be labelled. However, only a few engineering experts possess the knowledge in marking technology within manufacturing companies. Hence, the article introduces a hybrid expert system which reproduces given problem-solving abilities from experts with regard to their technological and process knowledge during the marking of materials. For the implementation, a specific method for expert system development is explained. Through a detailed analysis of the process for part- specific marking of materials at a special machine manufacturer, requirements for the expert system and the knowledge base are created. Based on the results, the architecture and functionality of the classic expert system and extension of the system by machine learning algorithms are explained and evaluated.