登录 EN

添加临时用户

电子束选区熔化过程的多尺度多物理场建模

Multi-scale Multi-physics Modeling of Electron Beam Selective Melting Process

作者:闫文韬
  • 学号
    2012******
  • 学位
    博士
  • 电子邮箱
    yan******com
  • 答辩日期
    2017.03.20
  • 导师
    林峰
  • 学科名
    材料科学与工程
  • 页码
    127
  • 保密级别
    公开
  • 培养单位
    012 机械系
  • 中文关键词
    增材制造,电子束,多尺度模拟,传热,熔池流动
  • 英文关键词
    Additive Manufacturing, Electron Beam, Multi-scale Modeling, Heat Transfer, Molten Pool Flow

摘要

金属增材制造技术是目前制造业发展的前沿,被世界各国认为是一次新的技术革命。然而,增材制造过程包含多个时空尺度上的多种相互影响的物理现象,实验观测难度很大,对很多物理机制缺乏理解导致难以稳定控制成形过程,诸多的影响因素使得成形质量一致性差,这是目前金属增材制造技术发展的主要瓶颈。本论文针对电子束选区熔化技术,构建多尺度多物理场模型,模拟金属粉末熔化堆积的过程,以理解其中的物理机制,为工艺参数的选择、工艺过程的优化提供指导。本论文提出的电子束选区熔化过程的多尺度模型主要包含三部分: 微观尺度的电子束能量吸收模型,利用Monte Carlo方法模拟高能电子与材料原子的碰撞,追踪电子动能转化成材料原子振动能的分布,提出了全新的电子束热源模型——双高斯体热源,为细观和宏观尺度模拟提供了精确的输入热源,可以考虑材料特性和电子束特性等影响因素。 细观尺度的粉末熔化堆积模型,综合模拟了粉末颗粒在电子束扫描下的传热、熔化、流动、凝固等过程;研究了单道成形中的球化和不平直现象的形成机制,在表面张力的作用下,基板不熔化导致球化,熔池边缘与粉末的粘附造成单道不平直,降低粉末尺寸和层厚可有效改善单道成形质量;通过与铺粉过程模拟的耦合,实现了多层多道成形过程的三维模拟,揭示了扫描路径的影响和多层熔合的过程,明确了扫描间距应不大于单道成形时的基板熔化宽度,解释了零件边缘区域易出现孔洞等缺陷的机制,并量化讨论了成形过程的能量利用率和最终零件的表面粗糙度。另外,开发了简化的粉末传热模型,以快速判断给定工艺参数下粉末能否完全熔化,发现小粉末由于较高的比表面积而更容易熔化,且各粉末最先熔化区域并非烧结颈处。 宏观尺度的零件成形过程模型,基于细观尺度的模拟结果,将粉层简化为等效连续体,并忽略了熔池流动而只考虑传热,可以快速模拟整个零件的成形过程,与实验结果的对照验证了其预测熔池尺寸的精度,并可推测单道不平直程度等更多的质量指标。除揭示成形过程中的物理机制及影响因素的作用规律外,多尺度模拟还可以开展材料成形工艺参数筛选,快速找到工艺参数的大致范围以开展试验,以功能梯度材料为例展示了多尺度模型的应用,并得到了实验验证。

Metallic Additive Manufacturing technologies have proven to be very promising in recent years, which is believed to lead a new technological revolution. However, the manufacturing process, which consists of multiple complex physical phenomena over a broad range of time and length scales, poses a significant challenge for accurate experimental observations and measurement. Moreover, the process involves tens of parameters that could affect the manufacturing quality. Therefore, the quality inconsistency is the bottleneck for the wide industrialization of Additive Manufacturing technologies. In this dissertation, a multi-scale multi-physics modeling framework is proposed to simulate the Electron Beam Selective Melting (EBSM) process, which is a typical metallic powder-based Additive Manufacturing process. To the best of my knowledge, this is the first report of multi-scale model for the EBSM process. This framework mainly consists of three models ranging from micro-scale to meso-scale to macro-scale. micro-scale electron-atom interaction model using the Monte Carlo method, is aimed to deriving a new heat source model for the electron beam, by tracking the collision and energy transition between electrons and atoms. The heat source model, named as "double-Gaussian" heat source model, is material-dependent and experimental set-up specific, which is able to guide the process design and provide insight into uncertainty in experiments. A meso-scale powder evolution model,which incorporates multiple physical phenomena and takes into a variety of influencing factors, is capable of reproducing the complex melt-flow-solidify process of individual powder particles. The formation mechanisms of the single track defects, including balling effect and single track distortion, are systematically investigated, which reveals the dominance of surface tension. The melting processes along various scan paths in multiple powder layers are reproduced in 3D simulations to reveal the influence of successive tracks and layers, which is the first report. Thanks to the multi-layer multi-track simulation, the energy efficiency is quantitatively discussed, and the formation mechanism of surface roughness is studied. Additionally, a simplified powder-scale heat transfer model is developed to provide rapid prediction if powder particles can fully melt under given process parameters. A macro-scale heat transfer model, in which the loosely packed powder bed is simplified as an effective continuum material, is a powerful tool to rapidly reproduce the experimental fabrication process. To ensure a fair prediction accuracy, the simplifications are made based on the meso-scale simulations. Thus, the macro-scale model is more physically-enriched and predictive for more quality indices, which is validated by experiments.Besides shedding light on the fundamental physical mechanisms, the multi-scale model is feasible in design and optimization of the manufacturing process. A simulation-driven parameter selection scheme is proposed, and then demonstrated for an example case of the Functionally Graded Materials, which agrees well with experimental results.