Integration of a digital twin for data-driven modeling of punch-bending processes using the asset administration shell

Within the punch-bending process semi-finished products of strip or wire material are formed and punched in several subsequent steps into a finished product like brackets, mounts, contacts or spring elements. In the context of those multi-stage straightening and bending processes, cross-stage and quantity-dependent effects significantly leads to undesired component deviations. To optimize the punch-bending process with regard to these component deviations and thus the waste rate, the concept of a hybrid data-driven model is presented. To automatically acquire and process this hybrid data while also enable the usage by multiple clients, a digital twin has to be developed. In this paper the communication infrastructure between the punch-bending system and the digital twin is presented, using the Asset Administration Shell as specification. This automated communication is validated using exemplary data from the punch-bending system.