铸造充型过程宏观数值模拟研究进展  

Research Progress in Macroscopic Numerical Simulation of Casting Filling Process

作　　者：牛晓峰[1] 王天成 葛涛涛 张鸣宇 马巨 赵静雨 Niu Xiaofeng;Wang Tiancheng;Ge Taotao;Zhang Mingyu;Ma JU;Zhao Jingyu(School of Materials Science and Engineering,Taiyuan University of Technology)

机构地区：[1]太原理工大学材料科学与工程学院

出　　处：《特种铸造及有色合金》2023年第10期1312-1318,共7页Special Casting & Nonferrous Alloys

基　　金：中央引导地方科技发展资金资助项目(YDZJSX2022A012)。

摘　　要：综述了铸造充型过程宏观数值模拟研究的现状。给出了铸造充型过程宏观模拟所涉及的物理模型,并以三维带障碍物溃坝模型、平板件充型流动场为例,基于光滑粒子流体动力学(SPH)方法进行了模拟计算。阐述了常见的网格划分类型和数值求解方法,对比了不同数学方法的特点。说明了铸造充型多相流模拟的研究情况,并指出进行流动场气、液、固三相流宏观数值模拟,能够准确地预测夹渣、冷隔、浇不足、气孔等铸造缺陷,可考虑采用多数学方法耦合的思路处理。SPH的公式构造并不受粒子分布的随意性的影响,可以自然地处理一些具有极大变形的问题。通过图形处理器技术(GPU)能够大幅提高SPH的计算效率。将人工智能机器学习技术和确定性物理建模相耦合,有助于优化传统的数值求解算法,提高精度,使模拟结果更接近于实际生产。The research status of macroscopic numerical simulation of casting filling process was reviewed.The physical models involved in the macroscopic process simulation of casting flling were presented.Taking the three-dimensional dam-break model with obstacles and the filling flow field of flat plate as instance,the simulation cal-culation was carried out based on SPH method.The common meshing types and numerical solution methods were expounded,and characteristics of different mathematical methods was analyzed comparatively.The situation of multiphase flow simulation of casting filling was illustrated,and it is pointed out that the macroscopic numerical simulation of gas-liquid-solid three-phase flow in flow field can accurately predict casting defects such as slag in-clusion,cold shut,misrun and porosity,which can be processed by the coupling of multiple mathematical meth-ods.The formula construction of SPH is not affected by randomness of particle distribution,which can naturally deal with some problems with great deformation.The computational efficiency of SPH can be greatly improved by Graphics Processing Unit(GPU)technology.Coupling artificial intelligence machine learning technology with de-terministic physical modeling helps to optimize traditional numerical solution algorithms and improve accuracy,making simulation results closer to actual production.

关 键 词：宏观过程 数值模拟 流动场 多相流 人工智能 

分 类 号：G353.11[文化科学—情报学] TP311[自动化与计算机技术—计算机软件与理论] O242[自动化与计算机技术—计算机科学与技术]