2017/03/29 by J. Coër, Hervé Laurent, Coer, Jeremy +8
Engineering · Materials Science · #Classical Physics (physics.class-ph) #FOS: Physical sciences #Metal Forming Simulation Techniques #Metallurgy and Material Forming #Microstructure and mechanical properties #Soft Condensed Matter (cond-mat.soft)
paper · pdf · doi:10.48550/arxiv.1703.10126
openalex publication_date 2017/03/29 · openalex created_date 2025/10/10 · openalex updated_date 2026/07/28
The Swift test was originally proposed as a formability test to reproduce the\nconditions observed in deep drawing operations. This test consists on forming a\ncylindrical cup from a circular blank, using a flat bottom cylindrical punch\nand has been extensively studied using both analytical and numerical methods.\nThis test can also be combined with the Demeri test, which consists in cutting\na ring from the wall of a cylindrical cup, in order to open it afterwards to\nmeasure the springback. This combination allows their use as benchmark test, in\norder to improve the knowledge concerning the numerical simulation models,\nthrough the comparison between experimental and numerical results. The focus of\nthis study is the experimental and numerical analyses of the Swift cup test,\nfollowed by the Demeri test, performed with an AA5754-O alloy at room\ntemperature. In this context, a detailed analysis of the punch force evolution,\nthe thickness evolution along the cup wall, the earing profile, the strain\npaths and their evolution and the ring opening is performed. The numerical\nsimulation is performed using the finite element code ABAQUS, with solid and\nsolid-shell elements, in order to compare the computational efficiency of these\ntype of elements. The results show that the solid-shell element is more\ncost-effective than the solid, presenting global accurate predictions, excepted\nfor the thinning zones. Both the von Mises and the Hill48 yield criteria\npredict the strain distributions in the final cup quite accurately. However,\nimproved knowledge concerning the stress states is still required, because the\nHill48 criterion showed difficulties in the correct prediction of the\nspringback, whatever the type of finite element adopted.\n