Rolled aluminum alloys are known to beanisotropic due to their processing histories. This paper focuses onmeasuring and modeling monotonic and cyclic strengthanisotropies as well as the associated anisotropy of theelastic/elastic-plastic transition of a commercially-available rolledplate product. Monotonic tension tests were conducted onspecimens in the rolling plane of 25.4 mm thick AA 7075-T6 platetaken at various angles to the rolling direction(RD). Fully-reversedtension/compression cyclic experiments were also conducted. Asexpected, we found significant anisotropy in the back-extrapolatedyield strength. We also found that the character of theelastic/elastic-plastic transition (knee of the curve) to be dependenton the orientation of the loading axis. The tests performed in RDand TD(transverse direction) had relatively sharp transitionscompared to the test data from other orientations. We found thecyclic response of the material to reflect the monotonic anisotropy.The material response reached cyclic stability in IO cycles or lesswith very little cyclic hardening or softening observed. For thisreason, we focussed our modeling effort on predicting themonotonic response. Reckoning that the primary source ofanisotropy in the rolled plate is the processing-inducedcrystallographic texture, we employed the experimentally-measured texture of the undeformed plate material in continuumslip poly-crystal plasticity model simulations of the monotonicexperiments. Three types of simulations were conducted, upperand lower bound analyses and a finite element calculation thatassociates an element with each crystal in the aggregate. We foundthat all three analyses predicted anisotropy of the back-extrapolatedyield strength and post-yield behavior with varying degrees ofsuccess in correlating the experimental data. In general. the upperand lower bound models predicted larger and smaller differencesin the back-extrapolated yield strength. respectively, than wasobserved in the data. The finite element results resembled those ofthe upper bound when initially cubic elements were employed. Wefound that by employing an element shape that was moreconsistent with typical rolling microstructure. We were able toimprove the finite element prediction signficantly. The anisotropyof the elastic/elastic-plastic transition predicted by each model wasalso different in character. The lower bound predicted sharpertransitions than the upper bound model. capturing the shape of theknee for the RD and TD data but failing to capture the otherorientations. In contrast, the upper bound model predictedrelatively long transitions for all orientations. As with the upperbound, the FEM calculation predicted gentle transitions with lesstransition anisotrogy predicted than that of the upper bound
展开▼
