Table-Based Knowledge Integration for Graph-Assisted Enhancement of Process Understanding in Three Different Metal Forming Processes
Frontiers in artificial intelligence and applications
Abstract
Activities in the domain of metal forming rely on different types of expert knowledge. Starting from the simulations in early design stages to process engineering and ultimately the operational phase. Traditional document-based representations of expert knowledge cannot display the complex interrelations of parameters necessary to understand the processes. Knowledge graphs have emerged to systematically embed such relations. However, their implementation and subsequent accessibility require additional effort. This paper proposes an approach to systematically structure qualitative domain knowledge in tables to enable the automated generation of knowledge graphs, starting from a backward extrusion process. Furthermore, this paper examines the general applicability of the approach and the necessary adaptations for two additional metal forming processes: fine blanking, and deep drawing and cutting. The latter expands the types of integrated information artifacts from static PDF documents to the metadata of simulation datasets. An extended knowledge graph-based system aims to enhance process understanding through interaction with embedded domain knowledge. The evaluation highlights the system’s capabilities to provide information not readily available from the original information artifacts, while shedding light on remaining limitations. These include the remaining manual efforts in initial data curation and user interface customization, which should be addressed in future knowledge graph instantiations in the domain of metal forming.
Authors 7
-
Technical University of Munich
Affiliation as printed
Institute of Automation and Information Systems, Technical University of Munich, Germany, Email: josua.hoefgen@tum.de, vogel-heuser@tum.de
-
Friedrich-Alexander-Universität Erlangen-Nürnberg
Affiliation as printed
Institute of Manufacturing Technology, Friedrich-Alexander-Universitä Erlangen-Nürnberg, Germany, Email: arnold.harms@fau.de, marion.merklein@fau.de
-
Affiliation as printed
Manufacturing Technology Institute, RWTH Aachen University, Germany, Email: a.frigge@mti.rwth-aachen.de, p.niemietz@mti.rwth-aachen.de
-
Affiliation as printed
Manufacturing Technology Institute, RWTH Aachen University, Germany, Email: a.frigge@mti.rwth-aachen.de, p.niemietz@mti.rwth-aachen.de
-
Affiliation as printed
Institute of Industrial Automation and Software Engineering, University of Stuttgart, Germany, Email: sebastian.baum@ias.uni-stuttgart.de, michael.weyrich@ias.uni-stuttgart.de
-
Friedrich-Alexander-Universität Erlangen-Nürnberg
Affiliation as printed
Institute of Manufacturing Technology, Friedrich-Alexander-Universitä Erlangen-Nürnberg, Germany, Email: arnold.harms@fau.de, marion.merklein@fau.de
-
Technical University of Munich
Affiliation as printed
Institute of Automation and Information Systems, Technical University of Munich, Germany, Email: josua.hoefgen@tum.de, vogel-heuser@tum.de
Cited by 0 stored of 0
No patents citing this paper on Lens.org (checked 2026-10-06).