材料数据库和数据挖掘技术的应用现状-中国材料进展

文章信息/Info

作者:: 杨丽13; 苏航12; 柴锋1; 罗小兵1; 段琳娜12; （1. 钢铁研究总院工程用钢研究所，北京 100081）
（2. 北京钢研新材科技有限公司，北京 100081）
（3. Institute of Iron and Steel, RWTH Aachen University, Aachen 52072, Germany）

Author(s):: YANG Li1; 3; SU Hang1; 2; CHAI Feng1; LUO Xiaobing1; DUAN Linna1; 2; (1. Department of Structural Steel, CISRI, Beijing 100081, China)
(2. Beijing CISRI New Material Technology Co., Ltd., Beijing 100081, China)
(3. Institute of Iron and Steel, RWTH Aachen University, Aachen 52072, Germany)

Keywords:: material database; materials genome initiative; data mining; integrated computing; properties prediction

摘要:: 数据库作为材料基因工程的三大重点研究领域之一，对材料的加速设计有重要的作用和意义。从20世纪70年代起材料数据库开始在国内外逐步发展，到目前为止各国都已形成了一定数量的离线和在线数据库。早期材料数据库的基本功能是数据存储、数据管理以及数据检索服务。随着材料基因工程理念的提出，材料数据库开始关注发展数据共享、数据自动收集和输出等系列新功能。通过在线集成计算软件或程序、数据分析工具已逐步建立了一些基于数据库的材料智能设计平台。数据挖掘不考虑复杂的物理和化学意义，而是直接从材料数据库中挖掘出有价值的知识或模式的过程，它能够充分发挥材料数据库甚至小数据量在材料设计中的作用。目前已经被广泛应用到材料性能预测和优化、缺陷质量预测和生产监控、微观组织识别等相关领域。未来，材料数据库和数据挖掘技术将更加紧密地结合，协同向材料集成计算和智能设计的需求方向发展。

Abstract:: Database as one important research direction of materials genome initiative (MGI) has a meaningful role in the accelerated design of material. The material database was gradually developed worldwide since 1970s, and then a huge number of offline and online databases have been built up until now on. The basic function of early database is to store, manage and search data. With the concept of MGI put forward, material database begins to focus on a lot of advanced functions, such as data sharing, automatic collection and data output. Intelligent design platform for material based on the material database has been developed by online integrated computing software or program and data analysis software. Data mining is a process to find the valuable knowledge from the database without considering the complex mechanism of physics and chemistry. It plays an important role in material design in database or small data set. Data mining has been widely applied to the prediction and optimization of material properties, defects or quality prediction and production monitoring, recognition of microstructures. In future, the material database and data mining technology will be integrated tightly to satisfy the requirements of the integrated computing and intelligent material design.