Estimation of Metal Hardening Models at Large Strains

Simulation accuracy of large strain deformation of sheet metals, such as that which occurs during hemming and vehicle crash situations, is limited because existing hardening laws (true stress vs. true strain relationships) are extrapolated from uniform elongation data and applied for post-uniform deformation. In this paper, a reverse-engineering method was developed to predict metal hardening laws at large strains beyond uniform elongation for sheet metals. The method required a standard uniaxial tensile test and finite element analyses (FEA), and was implemented as a custom computer code called GMSS (General Motors Stress-Strain). The true stress vs. true strain data pairs are determined when the load and displacement history of a tensile test specimen matches the FEA results using GMSS. Test cases showed that the true stress vs. true strain relationships at very large strains (75% for AA6111 aluminum, and 85% for DP600 steel) could be automatically generated using GMSS. This reverse-engineering method will provide General Motors with an easy-to-use tool for generating very accurate metal hardening laws for post-uniform deformation that can greatly improve the accuracy of FEA for formability (including hemming), and crashworthiness simulations.