Keywords: artificial intelligence, digital twin, heat treatment, Industry 4.0, saw blades, ultra-fine-grain steel.

In the production of cold work tool steels, components are first subjected to heat treatment to set the desired parameters such as hardness and strength, and then ground to obtain the final geometric shape. Both processes are resource and energy intensive and highly complex due to many adjustable system and process parameters. Therefore, it is difficult to reliably control parameters such as stresses, microstructure, and dimensions due to the interaction between the individual processing steps. In industrial applications, this leads to a significant amount of rejected material that can only be detected late, which has an overall negative impact on overall efficiency. This work focuses on saw blades bodies that are made of 75Cr1 with microstructural grain sizes of approximately 20 ?m. Aided by artificial intelligence, the creation of a digital twin is presented to improve the production processes. It is intended to reduce the grain sizes to approximately 5 ?m in order to improve the quality and the performance of the final products. The data is fed to the digital twin by data collection from the manufacturing machines, simulations, laboratory investigations, and metallographic results. In addition, a basis is created to be able to map other processes and industrial sectors. By creating multi-scale and cross-process simulations with industrial machining data using artificial intelligence, it is possible to increase the understanding of the entire process chain. All process data is to be recorded and simulations are carried out to improve the process parameters so that defective parts can be reduced and detected earlier. Furthermore, the component quality and the energy efficiency of the process shall be increased by process optimization. This will lead to improved competitiveness in economic, ecological, and technological terms, especially for small and medium enterprises.

download the full-text of this paper (PDF, 5 pages, 445 Kb)

go to the previous paper
go to the next paper
return to the table of contents
return to the volume description