基于计算晶粒法的颗粒增强金属基复合材料高性能仿真

Particulate-Reinforced Metal Matrix Composites (PRMMCs) are widely used in mechanical engineering. However, traditional analytical micromechanical models can only give a rough prediction of their mechanical parameters. Moreover, due to the complex preprocessing procedure and high computational burden of finite elements (FEM), it is difficult to efficiently model the random microstructure of PRMMCs using FEM. Computational Grains, recently developed by the applicant, have demonstrated significant advantages in micromechanical modeling of composite materials. However, the current version of Computational Grains is limited to linear elastic analyses. Moreover, due to the significantly different characteristic sizes of micro- and macro-structures, it is impractical to model both the material and the structural component in one scale. Thus, based on the applicant’s work in the past 5 years, it is proposed to extend the current linear-elastic version of Computational Grains, to further consider the mechanism of matrix yielding; take advantages of random packing algorithms, to quickly construct Representative Volume Elements (RVEs) with statistical features of realistic PRMMC microstructures; study methods of imposing periodic boundary conditions without symmetric discretization of RVEs; and study scale-coupling methods that are compatible with Computational Grains. The purpose of this proposal is to develop a high-performance tool for the elastic-plastic modeling of PRMMCs using Computational Grains, which can simulate the micro- and macro-scale mechanical behaviors of PRMMCs efficiently and accurately,in the interest of meeting the growing demands of numerical simulation in material developments.

颗粒增强金属基复合材料（PRMMC）在机械工程等领域已有广泛的应用。然而，传统的解析法微观力学模型只能较为粗略地估计这类材料的力学性能参数；有限元法则在前处理和计算代价两方面存在瓶颈，难以对其随机微观结构进行快速高效的仿真。申请者提出的计算晶粒法在复合材料仿真方面有较大的优势，但是目前只能进行线弹性分析。此外，由于宏、微观尺寸的差异，跨尺度地模拟材料和结构不符合工程实际。基于申请者近五年的研究工作，本项目拟扩展当前的线弹性力学计算晶粒方法，进一步考虑基体材料的塑性屈服；结合随机颗粒堆积技术，快速建立具有PRMMC材料微观结构统计特征的代表性体积元；发展不需要对称离散的周期性边界条件施加方式；研究适用于计算晶粒法的尺度耦合方法。目标是基于计算晶粒法建立高性能的PRMMC材料弹塑性力学仿真工具，对材料的宏观、微观力学性能进行快速、准确的模拟，以满足材料研发对数值计算方法日益增长的需求。

在国家自然科学基金委的支持下，本项目负责人及其团队对基于计算晶粒法的颗粒增强金属基复合材料(PRMMCs)高性能仿真进行了系统地研究，并建立了一套基于计算晶粒法建立高性能的PRMMC材料弹塑性力学仿真工具，可以对材料的宏观、微观力学性能进行快速、准确的模拟，能很好的满足材料研发对数值计算方法日益增长的需求。该项目取得了较好的研究成果，发表了6篇SCI论文并获得1项软件著作权。在研究项目同时，项目负责人培养2名博士研究生，并带领学生参加9次国内外会议。项目负责人及其团队扩展了当前的线弹性力学计算晶粒方法，进一步考虑了基体材料的塑性屈服；结合随机颗粒堆积技术，研发了一套快速建立具有PRMMC材料微观结构统计特征的代表性体积元算法；研发了一套不需要对称离散的周期性边界条件施加算法；研发了适用于计算晶粒法的尺度耦合方法。除了计划制定的主要研究内容，项目负责人团队还额外开发了含界面相计算晶粒，含界面应力计算晶粒以及纤维计算晶粒。因此，本项目完成了计划制定的主要研究内容，并且达到了预期科研目标。