基于ANN和RS的超硬材料及制品知识库的研究  

Research on superhard materials and its products knowledge base with ANN and RS

作　　者：宋冬冬[1] 杨波[2] 栗正新[1] 

机构地区：[1]河南工业大学材料学院,河南郑州450007 [2]湖南大学计算机与通信学院,湖南长沙410082

出　　处：《超硬材料工程》2010年第3期25-28,共4页Superhard Material Engineering

基　　金：国家中小企业创新基金资助;<磨具CAD软件>06C26214101442

摘　　要：介绍了基于知识工程原理开发超硬材料及制品知识系统的思路。系统包括知识库(基于教材、文献、经验)、数据库(国家标准、产品、原材料、工艺、工装、被加工材料、应用参数)和图像库(石墨-金刚石相图,BN相图、晶体结构、岩相图、金相图、工装结构图)。通过人工神经网络对大量的研究数据进行知识挖掘来补充完善超硬材料及制品设计的配方、工艺和相图数据。应用粗糙集理论,对大量复杂的知识数据进行简化以便去除冗余属性,为配方设计中原料选用提供了一种新方法。研究结果形成了一个初步的基于ANN和RS的超硬材料知识库系统和专家系统,提高了超硬材料及制品(人造金刚石和cBN及制品)的研发效率及准确性。This paper systematically introduces the idea of establishment of a superhard materials and its products knowledge system.In order to improve the developing efficiency and accuracy of superhard materials and related products,a knowledge system was developed according to the principle of Knowledge-Based Engineering,which includes the knowledge base（relevant textbooks,literatures and experiences）,database（standards,industrial products,raw materials,manufacturing technology,equipments,processed objects and parameters） and image base（graphite-diamond phase diagram,boron nitride phase diagram,crystal structure,ceramic phase diagram,metal phase diagram,mould image）.The ANN（Artificial Neural Network） method was carried out to analyze the experiential data.Parameters for the preparation of superhard materials and related products were optimized.RS（Rough Set） method was applied to develop a new method for raw materials selection.A preliminary knowledge-based system and expert system of superhard materials and related products was proposed on the bases of the ANN and RS investigation.

关 键 词：超硬材料制品 人造金刚石 立方氮化硼 知识库 数据库 人工神经网络 粗糙集 

分 类 号：TQ164[化学工程—高温制品工业]