Simulation of metal punching and trimming using minimal experimental characterization

This paper presents a validated finite element modeling approach for simulating shear cutting, needing a minimal amount of experimental characterization. Only one uniaxial tensile test and one force–displacement relationship from a punching experiment are needed for calibration, with maintained prediction accuracy compared to more experimentally demanding approaches. A key ingredient is the observation that the Lode angle parameter is close to zero in the fracture region, postulating that the fracture strain only depends on stress triaxiality, with one free calibration parameter. The true stress–strain behavior is provided from inverse modeling of the tensile test, whereas the fracture model is calibrated using the punching test. The model is verified for different materials by comparing force–displacement curves for punching experiments not used in the calibration. The prediction error for the intrusion is below 4%. A validation is made for two setups. The local residual stresses are measured using Focused Ion-Beam Digital Image Correlation (FIB-DIC). The simulated values are within the experimental bounds. Cut edge morphology and plastic strains obtained by nano-indentation mappings are compared to simulation results, showing a decent agreement. For trimming, the cut edge morphology prediction performance decreases at 17% cutting clearance while it is maintained over the whole range for punching. The predicted hardness values hardness values have a mean absolute percentage error below 7.5%. Finally, the effect of element size and remeshing is discussed and quantified. The minimal experimental characterization and simulation effort needed, enables an efficient optimization of the cutting process in the industry.