An investigation into the characterization of the hardening response of sheet metals using tensile and shear tests with surface strain measurement

Abstract

The adoption of full-field digital image correlation (DIC) strain measurement in recent years has fundamentally transformed conventional analysis procedures for the mechanical characterization of sheet metals. The wealth of local strain data that is now available until fracture has enabled experimentalists to propose new analysis techniques for tensile and shear tests to obtain the hardening response to large strain levels. Although the tests show great potential, significant gaps remain surrounding their applicability. The choice of test geometry and suitability of the analysis across a broad range of sheet metals with diverse hardening and anisotropic behaviour remain open questions. Critically, the assumptions within the tensile and shear methodologies in applying surface strain measurements to large strains while omitting through-thickness gradients have not been verified. The objective of the present study is to perform a critical evaluation of constitutive characterization techniques for sheet metals under quasi-static, ambient conditions. A series of virtual experiments were developed to numerically replicate the testing and analysis procedures followed in lab-scale experiments to extract the hardening response. A simple shear and two standard tensile test geometries were evaluated across a broad range of plasticity characteristics including different hardening rates and anisotropic behaviour. For each case, the input hardening curve in the virtual experiment is treated as the exact solution and compared to the hardening response obtained from each analysis method. The so-called area reduction method (ARM) used for tensile specimens was found to be relatively insensitive to the specimen geometry, yield exponent, and plastic anisotropy but overestimated the flow stress. The performance of the ARM method hinges upon the aspect ratio of the cross-section to avoid shear band formation and requires empirical correlations to improve its predictions at large strains. Despite their simplicity, tensile analysis methods that use small extensometers or local DIC point data are shown to systematically overestimate the hardening behaviour. Simple shear tests can provide accurate estimations of the hardening response, but the choice of geometry is material dependant. The study concludes by considering a strongly anisotropic mild steel to evaluate the tensile and shear characterization strategies with the results of a hydraulic bulge analysis.