Matti Schneider

Flavia Gehrig, Matti Schneider

Although FFT-based methods are renowned for their numerical efficiency and stability, traditional discretizations fail to capture material interfaces that are not aligned with the grid, resulting in suboptimal accuracy. To address this issue, the work at hand introduces a novel FFT-based solver that achieves interface-conforming accuracy for three-dimensional mechanical problems. More precisely, we integrate the extended finite element (X-FEM) discretization into the FFT-based framework, leveraging its ability to resolve discontinuities via additional shape functions. We employ the modified abs(olute) enrichment and develop a preconditioner based on the concept of strongly stable GFEM, which mitigates the conditioning issues observed in traditional X-FEM implementations. Our computational studies demonstrate that the developed X-FFT solver achieves interface-conforming accuracy, numerical efficiency, and stability when solving three-dimensional linear elastic homogenization problems with smooth material interfaces.

Binh Huy Nguyen, Matti Schneider

Solving cell problems in homogenization is hard, and available deep-learning frameworks fail to match the speed and generality of traditional computational frameworks. More to the point, it is generally unclear what to expect of machine-learning approaches, let alone single out which approaches are promising. In the work at hand, we advocate Fourier Neural Operators (FNOs) for micromechanics, empowering them by insights from computational micromechanics methods based on the fast Fourier transform (FFT). We construct an FNO surrogate mimicking the basic scheme foundational for FFT-based methods and show that the resulting operator predicts solutions to cell problems with arbitrary stiffness distribution only subject to a material-contrast constraint up to a desired accuracy. In particular, there are no restrictions on the material symmetry like isotropy, on the number of phases and on the geometry of the interfaces between materials. Also, the provided fidelity is sharp and uniform, providing explicit guarantees leveraging our physical empowerment of FNOs. To show the desired universal approximation property, we construct an FNO explicitly that requires no training to begin with. Still, the obtained neural operator complies with the same memory requirements as the basic scheme and comes with runtimes proportional to classical FFT solvers. In particular, large-scale problems with more than 100 million voxels are readily handled. The goal of this work is to underline the potential of FNOs for solving micromechanical problems, linking FFT-based methods to FNOs. This connection is expected to provide a fruitful exchange between both worlds.