Set-up assistance system for transfer presses based on AI

The Institute of Forming Technology and Machines and the Institute of Measurement and Control Technology are involved in the DFG priority program SPP 2422 with a joint research project " Set-up assistance system for transfer presses based on AI".

Due to complex internal interactions with not fully identified interdependencies, creating and maintaining stable process conditions during production on transfer presses can only be achieved with a great deal of set-up effort and by incorporating implicit knowledge. If changes to the process parameters occur in one stage, which lead to changes in process forces, for example, this influences the process sequence in other stages, which makes the restoration of good part production a lengthy process, depending on the complexity of the tool set.

As part of the new research project, AI-based methods will therefore be used to gain a deeper understanding of the relationships between process influencing variables on multi-stage presses and to identify significant influencing variables with regard to component quality. To this end, a demonstrator component and suitable geometric quality parameters are first defined and a multi-stage forming process is developed. Based on a data acquisition system, the resulting tool set is equipped with various sensors for measuring system variables (such as process forces, temperatures, structure-borne noise). Furthermore, the quality of the process is quantified by comparing the measured and optimum quality parameters.

To model the system interrelationships, an approach consisting of two interlocking AI models is developed, which first identify the system-inherent interactions and also predict the geometric quality characteristics depending on the system configuration. On the basis of non-linear optimization methods, which run through the two-stage AI model, a recommendation for action is derived for an optimal machine setup to regain good part production.