Determining the residual formability of shear-cut sheet metal edges by utilizing an ML based prediction model

Abstract

When forming high-strength steel sheet material, premature failure can occur at shearcut component edges because formability of the base material is reduced due to work hardening caused by the previous punching process. Here, the digitization of production processes provides new possibilities for quality monitoring of such forming and stamping processes. In this context, the present paper deals with a novel machine learning (ML) based method for determining the residual formability of sheet metal materials from measured punching force curves. The specific objective of the study carried out was to develop an efficient and accurate method for predicting the residual formability of shear-cut edges. The methodology proposed for this purpose involves collecting a comprehensive dataset comprising experimental measurements of material properties, cutting conditions and punching-force curves measured during blanking. To determine the residual formability of the sheet metal materials investigated, hole tensile tests were performed and the maximum major and minor principal strain at initiation of cracking were measured. This dataset was then used to train and validate different AI prediction models, which employ machine learning algorithms to establish complex relationships between input parameters and residual formability.