In this paper, the grain microstructure and strain partitioning in a polycrystalline NiTi wire subjected to tensile loading was reconstructed from an experimental 3D-XRD dataset. The reconstruction of a volume containing more than 8,000 stressed grains involved optimization with respect to both the geometrical features and material elastic properties. The geometrical features of the microstructure were reconstructed using Laguerre tessellations based on the experimental 3D-XRD dataset. Two different algorithms fitting Laguerre tessellations were applied in order to assess the sensitivity of the reconstruction to the choice of the algorithm. The material properties in terms of elastic anisotropy were refined from an initial published value to minimize the mismatch between experiment and simulation using an optimization algorithm based on linear elasticity simulations. As a result of this, we constructed a numerical microstructure model that statistically matches the experimentally probed material in terms of positions and sizes of grains as well as partitioning of elastic strain and stress in the microstructure (average elastic properties and standard deviations of piecewise constant components of elastic strain and stress tensors in grains).

中文翻译：

---

从 3D-XRD 数据重建的 NiTi 线的数值微观结构模型

在本文中，从实验 3D-XRD 数据集重建了承受拉伸载荷的多晶 NiTi 线中的晶粒微观结构和应变分配。包含超过 8,000 个受力晶粒的体积的重建涉及几何特征和材料弹性特性的优化。微观结构的几何特征使用基于实验 3D-XRD 数据集的 Laguerre 细分重建。应用了两种适合 Laguerre 曲面细分的不同算法，以评估重建对算法选择的敏感性。材料在弹性各向异性方面的特性从最初公布的值中细化，以使用基于线性弹性模拟的优化算法最大限度地减少实验和模拟之间的不匹配。因此，我们构建了一个数值微观结构模型，该模型在晶粒的位置和尺寸以及微观结构中弹性应变和应力的分配（分段常数分量的平均弹性性能和标准偏差）方面与实验探测的材料在统计上匹配晶粒中的弹性应变和应力张量）。