Lengthscale effects in optical strain measurement for fracture characterization in simple shear

Abstract

Digital image correlation (DIC) strain measurement has transformed the study of shear fracture of automotive sheet metals over the past decade. Novel specimen designs have been proposed within the literature that can minimize the occurrence of premature edge fracture in uniaxial tension with DIC providing the local shear failure strains. However, the appropriate methodology for measurement of the local fracture strain and its strong dependence upon the lengthscale remain open questions. Shear localization can rapidly occur prior to fracture when the shear deformation zone collapses into a narrow shear band with extremely large local strains. Consequently, the shear failure strains reported using DIC can significantly vary based upon the choice of lengthscale and settings used in the DIC analysis. This study aims to perform a systematic investigation into strain measurement under simple shear for an automotive structural steel, DP1180. To by-pass the spatial averaging inherent in DIC strain measurement, the so-called Rotation method is proposed that employs the DIC displacement field data to directly determine the shear fracture strains. The choice of DIC settings that define the virtual strain gauge (VSG) is studied and contrasted with the measurements of the Rotation method and the fracture strains obtained from grain measurements. The VSG was found to be superior to the virtual strain gauge length (VSGL) to define the lengthscale by accounting for the subset size. The non-local averaging within the DIC strain measurements was observed to underestimate the local failure strains compared to the Rotation method. The variations in fracture strains from different strain measurement techniques revealed a notable contrast in measured values that was more remarkable than the differences in strain due to noise or sample-to-sample variation. The choice of DIC parameters used in the analysis was found to substantially alter the fracture strains and can result in significantly different stress-state dependent fracture loci with implications for fracture modeling. Depending upon the VSG and strain measurement method, the von Mises equivalent fracture strain in simple shear varied from 0.60 to 1.54 for respective gauge lengths of 0.625 mm to 0.020 mm. To reduce variability between analysts and enable consistent comparisons for the shear fracture strains the recommended parameter settings for the DIC analysis are provided and discussed.