ECCOMAS 2024

Simulation of the Viscoplasticity of Porous Polycrystals with FFT-Based Methods

  • Védrine, Louis (CNRM, Centre d'Études de la Neige)
  • Hagenmuller, Pascal (CNRM, Centre d'Études de la Neige)
  • Löwe, Henning (WSL, SLF)
  • Montagnat, Maurine (Univ. Grenoble Alpes, IGE)
  • Gélébart, Lionel (Université Paris-Saclay, CEA)

Please login to view abstract download link

Porous polycrystals comprise manufactured materials, such as metallic foams or polycrystalline nuclear fuel, and natural resources, such as snow. Porous polycrystals are considered here as solids consisting of a set of crystals with pores of similar size. Their microstructures can take many different shapes, which have a strong influence on the mechanical response. Knowledge of their mechanical properties, including their viscoplasticity, is essential to optimize them for industrial purposes or to better understand their role in the natural environment. However, systematic relationships between the microstructure of porous polycrystals and their viscoplastic behavior are rarely spell out in the literature. Indeed, one difficulty that arises when dealing with varying porosity and microstructural patterns is the fact that the intrinsic deformation mechanisms active at the crystal scale may depend on the degree of confinement related to surrounding grains, which constrains the dislocations activity. For instance, in the frame of viscoplasticity modeling, Védrine et al., 2022 suggested that the role of non-basal slip systems (harder systems) in hexagonal ice crystals is different between snow and dense ice. This study aims to investigate the conceptual transition between crystals acting as isolated mono-crystals in highly porous materials and crystals with a more neighborhood-constrained behavior in dense materials. The case of a material with a strong viscoplastic anisotropy, ice, will be considered as strong strain incompatibilities are expected between crystals. We examined this transition through numerical full-field simulations. We conducted three-dimensional simulations on synthetic porous structures based on a crystal plasticity model, using a Fast Fourier Transform-based numerical solver, AMITEX_FFTP. This solver incorporates a modified discrete Green operator and Anderson's convergence acceleration technique, enabling efficient simulations with high contrast. Based on these numerical experiments, we investigated the influence of the microstructural pattern and the anisotropy of the viscoplastic behavior on the macroscopic homogenized behavior, with a specific focus on the creep exponent. This study comprises a first step towards a unified model of porous polycrystal viscoplasticity.