枝晶生长的介观尺度三维数值模拟

为预测合金介观组织的形成与演变,提出了改进的三维元胞自动机(CA)模型。该模型采用枝晶形状函数,简化描述介观区域的枝晶生长轮廓;利用基于Euler角的空间坐标转换,描述随机晶向的多晶粒生长;采用Hash表加速算法,耦合了溶质场计算,采用数值算法处理固液界面的溶质再分配。该模型可模拟大范围随机取向的多晶粒生长过程,并具有较高的计算效率。模拟结果表明:不同模拟区域和网格剖分尺寸对晶粒体积的分布规律没有明显影响;较快冷却速度下,存在溶质富集和枝晶偏析。对Al4Cu合金的砂型和金属型试样的显微组织分别进行模拟,结果与实验符合较好。

Three-dimensional numerical simulation of meso-scale dendritic growth in alloys

A modified three-dimensional cellular automaton (CA) model was developed to predict the formation and evolution of mesoscopic structures in alloys. Dendrite shape functions were used to approximately describe the actual meso-scale dendrite profile. Space coordinate transformations based on the Euler angle were used to calculate the multiple grain growth with stochastic preferential orientations. A hash table was used to accelerate the computations. The numerical model predicts the solute concentration distributions and the solute redistribution at the solid/liquid interface. The improved model can efficiently simulate the 3D growth of multiple grains with various preferential orientations in large domains. The simulations show that the domain and mesh sizes have little effect on the grain size distribution and the solute enrichment and dendrite segregation that occur with rapid cooling rates. The predicted mesoscopic structure of a cast Al-4.0%Cu alloy cast in a sand and metal mold agrees well with experimental results.