Self-learning Expert Systems in Textile Technology: Development of a Socio-Technical Approach

Abstract

Digital learning factories prepare a showcase to demonstrate the utilization of new digital technologies in production. The production of technical and domestic textiles in European countries requires a lean and error free approach in order to maintain competency against production in low wage countries. Digital expert systems can support setting and troubleshooting processes in a textile production. A procedure model for the introduction of a self-learning expert system (SES) based on AI developed and validated in a socio-technical approach will be introduced in this article. The introduction of the SES can be practiced in the model textile production line in DCC of ITA Group in Aachen, Germany. The procedure model considers all required social and technical components in a textile production. The development and design processes take place in cooperation with all relevant groups of actors like employees and managers. Its technical core is a process model which in a first step enables companies to identify the most important quality-relevant defects and the process input variables that cause them. A SES suitable for the company will then be introduced in order to support the employees in setting the process input variables, thus avoiding defects and simplifying their elimination by suggesting most relevant causes. The procedure model still contains approaches for the iterative-participative development of the SES in order to ensure acceptance by the production employees.