In this paper, we present the results of 426 anisotropic grain growth simulations in two-dimensions using a diverse set of initial microstructures. We consider anisotropy by using a realistic 5D grain boundary (GB) energy function (Bulatov et al., 2014). We sample from the space of all possible crystallographic textures (the texture hull) to generate initial microstructures. The resulting diversity of initial states enables us to evaluate the impact of the initial state on the evolutionary trajectory and identify general trends. We run anisotropic grain growth simulations using these microstructures as the initial points. Additionally, we analyze the evolution of the texture and GB network of these microstructures. We use the orientation distribution function (ODF) and the triple junction distribution (TJD) as the descriptors of the microstructures. We find that at the end of the simulations, there is a universal increase in the sharpness of the ODFs. We identify both a trivial and non-trivial sharpening effect and quantify their contributions to the overall texture sharpening. We also observe a positive correlation between texture strength and its sharpening rate during grain growth. Finally, the evolution of the GB network showed an increase in the low-angle GB fraction in most of the simulations, while spatial correlations in the GB network remained dominated by crystallographic constraints around TJs without the development of longer-range effects.

中文翻译：

---

各向异性晶粒生长过程中晶体织构和晶界网络结构的演变

在本文中，我们展示了使用一组不同的初始微观结构进行 426 个二维各向异性晶粒生长模拟的结果。我们通过使用真实的 5D 晶界 (GB) 能量函数来考虑各向异性（Bulatov 等人，2014）。我们从所有可能的晶体纹理（纹理外壳）的空间中采样以生成初始微观结构。由此产生的初始状态的多样性使我们能够评估初始状态对进化轨迹的影响并确定总体趋势。我们使用这些微观结构作为初始点进行各向异性晶粒生长模拟。此外，我们还分析了这些微观结构的纹理和 GB 网络的演变。我们使用取向分布函数（ODF）和三重结分布（TJD）作为微观结构的描述符。我们发现，在模拟结束时，ODF 的锐度普遍增加。我们识别了平凡和重要的锐化效果，并量化了它们对整体纹理锐化的贡献。我们还观察到晶粒生长过程中纹理强度与其锐化率之间呈正相关。最后，GB 网络的演化表明，在大多数模拟中，低角度 GB 部分有所增加，而 GB 网络中的空间相关性仍然由 TJ 周围的晶体学约束主导，而没有产生长期效应。