Machine Learning In Incremental Sheet Forming

Konferenz: ISR 2016 - 47st International Symposium on Robotics
21.06.2016 - 22.06.2016 in München, Germany

Tagungsband: ISR 2016

Seiten: 7Sprache: EnglischTyp: PDF

Persönliche VDE-Mitglieder erhalten auf diesen Artikel 10% Rabatt

Stoerkle, Denis Daniel; Seim, Patrick; Thyssen, Lars; Kuhlenkoetter, Bernd (Chair of Production Systems at Ruhr Universität Bochum, Bochum, Germany)

Within this article, the authors propose a new methodology to increase the geometric accuracy in the robot-based, incremental sheet forming process ROBOFORMING. This process addresses the production of sheet metal components in small lot sizes and prototypes. In ROBOFORMING, two cooperating industrial robots are applied for the kinematic-based forming of sheet metal workpieces. Hereby, workpiece-dependent tooling and dies are omitted. This offers very high flexibility for the geometrical design of the sheet metal workpieces. One of the major drawbacks of incremental sheet forming processes is the low geometrical accuracy, which limits the widespread industrial application of these. Responding to these constraints, the authors propose the application of machine learning techniques to increase the geometric accuracy in incremental sheet forming processes. In this context, they present a learning approach which applies reinforcement learning as a flexible and promising solution.