铝合金管弯曲成形表面粗化宏细观耦合建模仿真

铝合金管弯曲成形中的表面粗化缺陷严重影响了其表面质量和使役性能，迫切需要探明该成形过程中的表面粗化形成规律以避免此类缺陷的产生。管材弯曲成形宏观有限元模型无法准确描述细观尺度不均匀变形导致的表面粗化行为，而细观力学模型又难以应用到具有复杂加载边界条件的实际成形过程分析。将基于Voronoi图和晶体塑性理论的代表性体积单元（Representative volume element，RVE）细观模型引入宏观有限元模型，建立了铝管弯曲三维宏细观耦合模型，模拟研究5052-O铝管在弯曲过程中的表面粗化特征，系统分析了细观材料参数对弯曲过程表面粗化行为的影响规律。结果表明，晶粒尺寸越小，晶粒内部与晶界附近的应变差值越小，其变形越均匀，能获得越小的表面粗糙度；对于Cube织构和随机织构混合的模型，随着Cube织构含量的增多，弯曲变形后表面粗糙度呈下降趋势；织构分布与加载方式耦合影响表面粗化行为。

Macro-meso Scale Coupling Modeling and Simulation of Surface Roughening in Aluminum Alloy Tube Bending

The conventional Arrhenius constitutive model is limited by its low accuracy in predicting flow stress. The phenomenological Arrhenius constitutive model coupling true strain is constructed for superalloy Inconel 617B. Furthermore, the BP-ANN (Back propagation-artificial neural network) constitutive model is constructed. The results show that the flow stress of the alloy is sensitive to the deformation temperature, deformation rate and true strain. The relative error (δ) of the flow stress predicted by the phenomenological model is 9.020% and the correlation coefficient (R) is 0.981 4, while the relative error of the BP-ANN model (the optimum structure is determined as 3×20×1) is only 1.527% and R reaches 0.998 9. The flow stress of Inconel 617B is more accurately predicted by the BP-ANN model. Using BP-ANN to construct constitutive model can improve the prediction accuracy, which can describe the constitutive relationship between flow stress and deformation parameters of superalloy Inconel 617B more rationally.