Data-Driven Process Optimization in Forming

Efforts towards interconnected and intelligent production are being promoted at national level in various initiatives, be it 'Industry 4.0' in Germany, 'Society 5.0' in Japan, 'Smart Manufacturing' in the US, 'Industry of the Future' in France, 'Made in China 2025' in China or 'Intelligent Factory' in Italy [1]. An essential component of all these endeavors is the data collected, for example from product design and process development, but especially from ongoing production. This database initially paints an incomplete and disruptive picture of reality in digital form. By curating and statistically analyzing the data, linking it with suitable models makes it possible to derive causal relationships that ultimately lead to decisions and actions. In this context, the term digital twins is used, which are utilized for a variety of tasks. Data-driven approaches are increasingly being used to build models that underlie digital twins, although there are still cross-domain and domain-specific challenges, particularly in the production context.