Beschreibung
This dissertation addresses a key limitation of current industrial automation systems: their inability to flexibly interpret, plan, and execute variable user tasks under changing conditions. Traditional rule-based approaches are too rigid for modern production environments, which increasingly require adaptability, fast reconfiguration, and intuitive human-machine interaction. To overcome this limitation, the dissertation proposes a three-layer framework that integrates large language models (LLMs), digital twins, and automation systems into an autonomous system. It further introduces the Task-Process-Service-Resource (TPSR) model as a unified mechanism for transforming user tasks into executable processes, and identifies four functional roles of LLMs in task automation. The proposed concepts are developed and refined through five design science research studies and demonstrated in case studies and prototype implementations. The results show that the approach enables adaptive task planning, event-driven control, simulation-based parameterization, and digital model generation, while automating a substantial share of manual engineering work. Overall, this dissertation contributes a generalizable framework for embedding LLM-based reasoning into industrial automation systems, improving their adaptability and usability in areas such as process planning, system control, and information model generation. The framework improves adaptability and usability in applications such as process planning, system control, and information model generation. Limitations include reliance on accurate digital models, high computational demands, and the need for human oversight in safety-critical scenarios.
Herstellerkennzeichnung:
Shaker Verlag GmbH
Am Langen Graben 15a
52353 Düren
DE
E-Mail: info@shaker.de
