Experimental analysis and modeling of anisotropic ductile damage in non-proportional extreme low-cycle biaxial loading with shear-tension histories

Abstract

This paper discusses the ductile damage and fracture behavior based on newly designed and performed non-proportional, non-reverse, extremely low-cycle experiments. In contrast to most extremely low-cycle experiments, which involve reverse loading histories or are restricted to a limited small plastic strain range, this study proposes novel non-proportional tension-to-shear (TS) and shear-to-tension (ST) loading patterns. Different combinations of displacement increments are applied within individual cyclic loading patterns, ensuring that specimen failure is governed by ductile damage and fracture under large plastic deformations. Numerical calculations are based on an advanced cyclic plastic-damage constitutive model with combined hardening laws. A novel non-proportionality parameter incorporating the effective back stress tensor is introduced into the combined hardening formulation to account for non-proportional hardening, allowing for a more accurate characterization of plastic behavior under non-proportional cyclic loading conditions. Digital image correlation (DIC) is used to analyze the global load–displacement curves and local strain fields, enabling comparison with the numerical results at both macroscopic and microscopic levels. Scanning electron microscope (SEM) and light optical microscope (LOM) images were taken from the fracture surfaces as well as both fractured and unfractured notch areas of the specimen, respectively. A novel quantitative analysis was introduced to evaluate the obtained SEM images using a convolutional neural network (CNN) approach, whereas LOM images were analyzed with the open-source software ImageJ. The present work highlights that non-proportional loading histories and shear-tension cyclic loading sequences with various plastic amplitudes significantly influence on the material’s plastic and ductile damage behavior.