A customisable computational pipeline for understanding formability using virtual materials testingWednesday (06.11.2019) 09:55 - 10:15 Part of:
During forming processes of light alloys, the mechanical anisotropy evolves in a potentially complex way. Precisely understanding this evolution would enable manufacturers to gain confidence in the properties of their components. An important part of this understanding is the identification of the yield surface and how it evolves during deformation. To do this experimentally, the material would need to be subjected to an extensive testing regime in which arbitrary stress states are probed. A yield function could then be fit to the yield data at different yield points to provide a phenomenological model of the yield surface evolution. Such a testing programme is expensive and time-consuming and so, in recent years, so-called virtual materials testing has been developed, whereby the yield data is provided by simulation rather than experiment.
Full-field crystal plasticity models, using either a finite element method or a fast Fourier transform method can be used for this purpose and such methods explicitly account for the effects of microstructure, which have been shown to be crucial for determination of the evolving anisotropy. However, these methods are computationally expensive, and their use intractable for the simulation of lab-scale components. On the other hand, cheaper phenomenological models can be calibrated using full-field models.
In this work, we developed a customisable pipeline for virtual materials testing that enables us to gain insight into which computationally cheaper models may be employed without loss of accuracy. We focus on identifying which, from a range of models (from Taylor models to full field crystal plasticity simulations), offer the best trade-off between accuracy and computational expense. We demonstrate this pipeline with Surfalex (AA6016A), which has been developed specifically to exhibit high formability, and is designed for use as skin in automotive applications. For calibration and comparison, we undertook a comprehensive experimental characterisation of the material, including uniaxial tensile tests with digital image correlation and electron backscatter diffraction analyses. We developed this work openly and aim for it to be beneficial to other researchers interested in efficient virtual materials testing.