<div class="section abstract"><div class="htmlview paragraph">Thermoplastics find application in many automotive components. Off late, hardware testing is supplemented by analysis using finite element (FE) codes. One of the factors determining the analysis accuracy is the representation of the components with suitable material models. While a uniaxial tensile test on the specimens typically provides engineering stress-strain data, material plasticity models in commercial FE solvers, such as LS-DYNA and ABAQUS, require equivalent plastic strain versus true stress as input. Engineering stress and strain can be converted to the corresponding true stress and true strain using equations based on the constant volume assumption; however, these equations are valid only up to the point of necking. A conventional approach of estimating the true stress and true strain beyond necking involves a trial and error method in which several true stress-strain curves are generated, and a finite element simulation is used to find the best possible fit with the test data. This approach is time consuming and provides an approximate true stress-strain curve beyond necking. The objective of this paper is to provide an experimental alternative to the conventional approach. Digital image correlation (DIC) was used to measure volumetric strain in the necking region of a tensile test specimen by employing two cameras simultaneously, one focusing on the front (width, length) and the other focusing on the side (thickness, length) surface of the test specimen. True strain measured by DIC in the longitudinal direction and the true stress, estimated from the applied force and measured true transverse strains, were used to generate material models for use in FE analysis without modification, i.e. without a need for correction factors. Material models were then validated by comparing the FE simulation with the tensile test results.</div></div>
Read full abstract