DEEP LEARNING IN SHEET-BULK METAL FORMING PART DESIGN
Editor: Marjanović D., Štorga M., Škec S., Bojčetić N., Pavković N.
Author: Sauer, Christopher; Schleich, Benjamin; Wartzack, Sandro
Section: SYSTEMS ENGINEERING AND DESIGN
DOI number: https://doi.org/10.21278/idc.2018.0147
Within the Transregional Collaborative Research Centre 73, a self-learning engineering workbench is being developed. It assists product developers in designing sheet-bulk metal formed (SBMF) parts by computing the effects of given product and process characteristics on the product properties. This contribution presents a novel approach to using deep learning methods for the properties prediction. By making use of a parameter study of 20 SBMF part designs, a metamodel is trained and used to predict the total equivalent plastic strain on local level as an indicator for part manufacturability.