Simple shear behavior of 2024-T351 aluminum alloy over a wide range of strain rates and temperatures: Experiments and constitutive modeling

Abstract

A recently developed single shear specimen (SSS) was used to investigate the thermo-viscoplastic behavior of 2024-T351 aluminum alloy (AA2024-T351) under simple shear stress state. Shear stress-shear strain relations over a wide range of strain rates, covering low (0.001 and 0.1 s−1), medium (10 s−1), high (3000 and 6000 s−1) and very high (14000-49000 s−1) regimes and at initial temperatures from 293 to 573 K were obtained experimentally. The results show that: (1) at 293 K, AA2024-T351 exhibited negligible strain rate sensitivity at strain rates below 6000 s−1, and the effect of strain rate became obvious at strain rates above 14000 s−1; (2) dynamic strain aging (DSA) occurred at temperatures between 373 and 573 K and strain rates below 6000 s−1; (3) strain rate sensitivity of the material was strongly affected by DSA, and bell-shaped strain rate sensitivity curves were observed; (4) with an increase in temperature, peak values of the bell-shaped strain rate sensitivity curves shifted to higher strain rates. A constitutive model taking the advantages of Johnson-Cook model, Khan-Huang-Liang model, Nemat-Nasser-Li model and Wang-Guo-Gao model was developed for deformation behavior description, with the effects of viscous drag on dislocation motion and DSA considered. The experimental flow stress curves were compared to the model predictions, and a good agreement was observed. To validate the constitutive model independently of the experiments used for the identification of the model parameters, numerical simulations of Taylor impact tests were performed. A good agreement between experimental and numerical post-test specimens, in terms of mushroom shaped specimen head and reduced specimen length, was observed. According to the numerical results, strain rates varied largely between 103 and 106 s−1 during Taylor impact tests. Besides, numerical simulations of Taylor impact tests were performed again using the constitutive model excluding the viscous drag component, and less accurate numerical results were obtained. It implies that for precise computations of engineering applications, the actual strain rate should be considered in the establishment of constitutive models.