Estimation of Texture-Dependent Stress-Strain Curve and <i>r</i>-Value of Aluminum Alloy Sheet Using Deep Learning

Deformation of an aluminum alloy sheet which is affected by its underlying crystallographic texture has been widely studied using crystal plasticity finite element method (CPFEM). Numerical material test based on the CPFEM allows us to quantitatively estimate stress-strain curve and the Lankford value (r-value) of aluminum alloy sheets that depend on the texture. However, to utilize the numerical material test as a means of optimizing the texture to design aluminum alloys, the CPFEM is computationally expensive. In this study, we propose a methodology for rapidly estimating the stress-strain curve and r-value of aluminum alloy sheets using the deep learning with a neural network. We train the neural network with synthetic texture and stress-strain curves calculated by the numerical material test. To capture the feature of synthetic texture from an {111} pole figure image, the neural network incorporates the convolution neural network. Using the trained neural network, we can estimate the uniaxial stress-strain curve and the in-plane anisotropy of r-value for various textures which contain Cube and S components. The results demonstrate that the neural network
trained with the results of the numerical material test is a promising methodology for rapidly estimating the deformation of aluminum alloy sheets.