The numerical prediction of ductile fracture of martensitic steel in roll forming

In this study, a ductile fracture model is used to generate a fracture locus for martensitic steel. Two sets of experiments are used to calibrate the model. The first method included a shear specimen, a specimen with a central hole and a notched specimen while in the second approach the notched specimen is replaced by a V-bend test to calibrate for plane strain conditions. The calibrated fracture model was integrated in a commercial FEA software and used to model fracture for a simple roll forming processes. Experimental roll forming trials were performed and forming strains analysed with a DIC system to verify the numerical results. The results show that the level and final shape of the fracture locus depends on the calibration method that is used. Fracture in the roll forming process is accurately predicted when plane strain was calibrated with the V-bend test. In contrast calibration with the notched specimen led to an overestimation of the fracture limit. The inaccurate calibration of the fracture locus with the notched specimen, which follows the conventional approach, is due to a non-linear strain path during material deformation.