Using local damage models to predict fracture in additively manufactured specimens

Abstract

This paper explores the efficacy of employing local damage models, normally applied to ductile material systems manufactured by subtractive techniques, to additively manufactured laboratory specimens. While these specimens were ductile and metallic, their additive character (i.e. porosity and surface roughness) could have had potential to activate multiple life-limiting failure paths, thus obfuscating failure prediction. Herein, two damage models are considered and compared: the micromechanical Gurson–Tvergaard–Needleman model and a Crack Band model of the strain-based, phenomenological genre. Simulations used to calibrate elastic and plastic material properties and predict damage in a novel, non-standard specimen were quasi-static, explicit. Both damage models proved capable in resolving the experimentally-observed failure path and associated loading conditions. The analyses described herein were made as part of the Third Sandia Fracture Challenge.