Treffer: Intelligent extraction of manufacturing system components from natural language using transformer models.

Writing control programs for manufacturing systems is time-consuming and requires expert knowledge. Traditional methods, whether heuristic or formal, often fail to scale with the complexity and flexibility of modern production systems. A key challenge limiting progress is the lack of labelled data tailored to the domain of control logic. This gap hinders the development of AI systems capable of automating control program generation from natural language specifications. In this work, we introduce AutoFactory, a new dataset specifically designed for this task, along with AutoLabel-NER, a custom labelling tool that supports efficient annotation. We frame component extraction as a Named Entity Recognition problem and fine-tune several transformer-based models to evaluate their ability to detect relevant entities in requirement specifications. We benchmark thirteen transformer-based models on this dataset and compare their performance. Our results show that language models can detect control components accurately. To continue toward automation, we propose a pipeline that maps extracted components to predefined local control logic blocks. This pipeline generates IEC 61131-3 Structured Text code without manual programming. Our work creates the first benchmark and method for AI-assisted control program generation. It provides a solid foundation for future research in intelligent industrial automation. [ABSTRACT FROM AUTHOR]

Copyright of International Journal of Production Research is the property of Taylor & Francis Ltd and its content may not be copied or emailed to multiple sites without the copyright holder's express written permission. Additionally, content may not be used with any artificial intelligence tools or machine learning technologies. However, users may print, download, or email articles for individual use. This abstract may be abridged. No warranty is given about the accuracy of the copy. Users should refer to the original published version of the material for the full abstract. (Copyright applies to all Abstracts.)

Volltext ist im Gastzugang nicht verfügbar. Login für vollen Zugriff.