Slip transfer mechanism is studied based on in-situ tensile tests performed on an Al-Mg alloy and high-throughput computing. A statistical analysis of slip transfer is performed for over 1180 grain boundaries and 155,250 slip pairs while considering the local lattice rotation, grain boundaries, and slip system geometries. Two new slip transfer parameters, $N$ and $B$, representing the alignment of the slip planes and the activation of slip systems in adjacent grains, are proposed. A decision tree model is built to evaluate the prediction accuracy of the slip transfer parameters. The optimized results of the decision tree classifiers show that the two new parameters are more effective than the Luster–Morris parameter, $m^{\prime }$. The improvement in the classification accuracy is attributable to the information regarding the slip plane geometry and orientation evolution contained in $N$ and $B$. The new slip transfer parameters improve our understanding of the slip transfer mechanism and can be integrated into plasticity models to predict the deformation and fracture behaviors of polycrystalline materials.

基于对铝镁合金进行的原位拉伸试验和高通量计算，研究了滑移传递机制。在考虑局部晶格旋转、晶界和滑移系统几何形状的同时，对超过 1180 个晶界和 155,250 个滑移对进行了滑移转移统计分析。两个新的滑动传递参数，$否$和$乙$，代表滑移平面的排列和相邻晶粒中滑移系统的激活，被提议。建立决策树模型评价滑移传递参数的预测精度。决策树分类器的优化结果表明，这两个新参数比 Lustre–Morris 参数更有效，${米}^{‘}$. 分类精度的提高归因于有关滑移面几何形状和方向演化的信息$否$和$乙$. 新的滑移传递参数提高了我们对滑移传递机制的理解，并且可以集成到塑性模型中以预测多晶材料的变形和断裂行为。