4-12. Materials intelligent design: software, databases and case studies

Yong Du¹，Shuhong Liu¹，Yuling Liu¹，Changfa Du²，Shiyi Wen¹, Taibai Fu²，Qi Huang¹, Hong Mao¹，Yi Kong¹，Jiong Wang¹, Sang Tan¹, Kai Li¹，Lianchang Qiu¹，Zhoushun Zheng²，

Bo Li¹, Yingbiao Peng³，Cong Zhang⁴，and Peng Zhou⁵

1. State Key Laboratory of Powder Metallurgy, Central South University, Changsha 410083

2. School of Mathematics and Statistics, Central South University, Changsha 410083

3. College of Metallurgy and Materials Engineering, Hunan University of Technology, Zhuzhou 412007

4. Hunan Provincial Key Defense Laboratory of High Temperature Wear-Resisting Materials and Preparation Technology, Hunan University of Science and Technology, Xiangtan 411201

5. Collaborative Innovation Center of Steel Technology, University of Science and Technology Beijing, Beijing 100083

Abstract: Material Science is a process of continuously understanding of the evolution of microstructure and establishment of the quantitative relationship between microstructure and performance. Phase diagrams and phase transitions are important foundations for studying the structure, properties and applications of materials. This report mainly introduces our research work on phase diagram, phase transformation, database construction and machine learning and shows several examples of intelligent material design. For material genetic engineering and intelligent design, we developed an efficient and stable program CALTPP (CALculation of Thermophysical Properties) with independent intellectual property rights to calculate thermophysical properties of multicomponent alloys. It can provide thermophysical parameters (including diffusion coefficient, interfacial energy, thermal conductivity, viscosity and molar volume) that vary with composition and temperature. some progress has been made in the calculation program of phase diagram, and a novel calculation method of binary phase diagram has been proposed. The practical applications show that the method can effectively improve the efficiency and stability of phase diagram calculation. Besides, multi-scale phase field models were developed to quantitatively simulate the microstructural evolution of materials with different scales. After nearly 20 years' efforts of the team, we have established our own database of thermodynamic and thermophysical (diffusion coefficient, viscosity, thermal conductivity, volume) databases of light and cemented carbides based on key experiments, phase diagram calculation method and first-principles calculations. For Al-alloys, alloy composition and the alloy microstructure can be optimized and controlled based on phase diagram calculations. As a result, a high-strength high-toughness 6xxx aluminum alloy with better comprehensive performance than other similar alloys was obtained. For cemented carbides, we designed the alloy composition and sintering process based on the thermodynamic and kinetic calculations on the gradient layer microstructure of gradient cemented carbides, and phase-field simulations of the liquid phase migration phenomenon during sintering process, the morphology of WC grains and the microstructure evolution of nano-precipitates in a composite binder phase. On CVD coating, TiSiCN coatings with a certain Si content and grain size were prepared from gaseous mixtures of TiCl4, SiCl4, CH3CN, NH3/N2 and H2 by a low pressure chemical vapor deposition (CVD) process under the guidance of calculations of phase diagrams. they displayed superior cutting performance compared to MT-Ti(C,N) and state-of-the-art thicker MT-Ti(C,N) +Al2O3 multilayer coating.

Keywords: Intelligent design of materials,; Software; Database; Industrial product design

材料智能设计: 软件、数据库及工业产品设计实例

杜 勇¹，刘树红¹，刘钰玲¹，都昌发²，文诗艺¹，傅太白²，黄奇¹，毛鸿¹，孔毅¹，汪炯¹,唐塞¹，李凯¹，邱联昌¹，郑洲顺²，李波¹, 彭英彪³，张聪⁴，周鹏⁵

1. 中南大学粉末冶金国家重点实验室，长沙410083

2. 中南大学数学与统计学院，长沙410083

3. 湖南工业大学冶金工程学院, 株洲412007

4. 湖南科技大学高温耐磨材料及制备技术湖南省国防科技重点实验室, 湘潭411201

5. 北京科技大学钢铁共性技术协同创新中心, 北京100083

摘要：材料学研究是对组织结构演变规律的不断理解及对组织结构与性能之间量化关系建立的探求过程。相图和相变是研究材料结构、性能及应用的重要基础。本报告以相图、相变、数据库及机器学习为核心系统介绍我们在软件开发、数据库建立方面的研究成果，并展示材料智能设计的几个实例。（1）面向材料基因工程和智能设计，我们自主开发了高效稳定的多元合金热物性计算通用的程序CALTPP（CALculation of ThermoPhysical Properties），可提供随成分和温度变化的热物性参数(包括扩散系数，界面能，热导率，粘度和摩尔体积)。（2）在相图计算程序上提出了一种新的二元相图计算方法，可有效提高相图计算的效率和稳定性。开发了多尺度的相场模型即可针对不同尺度的材料微结构演变现象进行定量化模拟。（3）经过近二十年的努力，研究团队基于关键实验、相图计算方法和第一性原理计算建立了具有自主知识产权的轻合金和硬质合金热力学、热物性质（扩散系数、粘度、热导率、体积）数据库。针对铝合金，团队基于相图热力学计算优化铝合金成分、调控合金微观结构，由此获得了一种综合性能优于其他同类合金的高强韧6 xxx系铝合金。（4）在硬质合金研究领域，团队基于硬质合金热力学、热物性质计算和相场方法模拟梯度硬质合金梯度层微观组织结构、WC晶粒形貌及复合粘结相中纳米析出相的微结构演变等，从而设计硬质合金成分和烧结工艺，并实现了产业化生产。（5）在CVD耐磨涂层方向，团队基于热力学计算和计算流体力学采用低压化学气相沉积技术以TiCl4、SiCl4、CH3CN、NH3/N2和H2为前驱体制备了一定Si含量和晶粒尺寸的TiSiCN涂层，其硬度和切削性能优于MT-Ti(C,N)和商用MT-Ti(C,N) +Al2O3多层厚涂层，并实现了产业化生产。

关键词：材料智能设计、软件、数据库、工业产品设计