3-D and Anisotropic Effects on the Prediction of Burst in Aluminum Tube Hydroforming

Abstract

Thin‐walled Al‐6260‐T4 aluminum tubes were hydroformed in a custom testing facility [1,8]. The major mode of failure observed in the experiments was bursting, despite the simultaneous application of axial compression while inflating the tubes. At the same time, a series of FE models were developed in the nonlinear code ABAQUS to simulate the experiments; however, initial computations failed to yield accurate predictions of burst. This was attributed to the adoption of the classical J2 plasticity, which is unsuitable for an anisotropic aluminum alloy, and led to an extensive study of the constitutive behavior of Al‐6260‐T4 and of its forming limits (see [2–4]). With the benefit of this improved understanding of the material behavior, the hydroforming simulations were revisited and models of different degrees of sophistication were developed. Starting with shell element models, the anisotropic yield functions calibrated earlier in [2–4] were shown to improve predictions over the J2 plasticity, but were found to still be deficient in predicting the failures observed in the experiments. This was in turn attributed to the fact that shell elements cannot capture the stress triaxiality associated with the gradual evolution of necking encountered in hydroforming. In addition, despite the relatively thin‐walled geometries involved, significant through‐thickness stresses develop in the regions of the tube in contact with the die. These stresses are again missed by a shell element discretization. Both of these observations point to the use of solid element models, to capture the stress triaxiality. We will show that when these models are run in conjunction with non‐quadratic anisotropic constitutive models, accurate predictions of failure in tube hydroforming are obtained. The conclusion that solid elements are required for failure calculations in tube hydroforming and that shell element models are deficient in that respect, contrasts sharply the current industrial practice in hydroforming simulations.