Optimization of forming parameters using AI and FEM

Due to the high complexity, the process parameters are usually not optimized with regard to product properties in real series production. The aim of the joint research project is to keep the production facilities in the cold forming of sheet metal stable by proactive control of the machine parameters. The aspired solution concept combines data-driven statistical (machine learning) and process-driven physical models (FEM) in order to map the behavior of the forming process and thus dynamically adjust the plant parameters. The material behavior is estimated by FEM calculations and combined with machine parameters and historical quality data. This allows correlations to be shown via the behavior in a model. In this way, the system learns which parameter combinations lead to a robust production process despite fluctuating environmental conditions (e.g. degree of oiling, geometry, roughness, etc.) and how it can avoid errors. The project combines the professorships for manufacturing engineering and for modeling and simulation of technical systems and processes. It is carried out together with the company Selmatec Systems GmbH and Gestamp Automoción S.L. and is supported by the European Fund for Regional Development.