基于制造大数据的产品研发知识集成与服务机制研究

项目介绍

AI项目解读

基本信息

批准号：
71671057
项目类别：
面上项目
资助金额：
49.3万
负责人：
余本功
依托单位：
合肥工业大学
学科分类：
G0112.信息系统与管理
结题年份：
2020
批准年份：
2016
项目状态：
已结题
起止时间：
2017-01-01 至2020-12-31

项目参与者：
杨颖；褚伟；赵树平；张强；邵臻；陈意；朱星；张连彬；
关键词：
知识服务机制产品研发异质知识集成制造大数据数据可用性评估

项目摘要

Big data of manufacturing can extend the source of the product research and development data and also provide developers with more accurate and comprehensive integration of research and development of knowledge, but its repetition, fragmented, scattered and heterogeneous characteristics bring challenges to carrying out product research and development of knowledge integration and knowledge service for enterprises. According to the openness, dynamic evolution of big data of manufacturing and complex correlation characteristics, this proposal starts with aggregate management of big data of manufacturing and comes out research in dynamic hierarchical knowledge integration methods and Real-time dynamic knowledge service mechanism, in addition, carrying out a case study on development of intelligent network of automobile products. Specific research contents include: (1) On the basis of analysis of the availability of heterogeneous big data of manufacturing, research is aimed at the management methods of panoramic data aggregation based on a heterogeneous big data of manufacturing. (2)The research explores the heterogeneous knowledge integration methods adapting to the need of the product development drive, against the features of shortening the product innovation cycle and constantly changing in research and development of knowledge system. (3) On the basis of analysis of research and development staff personalized knowledge service requirements, research is to study personalized real-time dynamic mechanism of knowledge service based on big data of manufacturing. The research provides product research and development of knowledge management under the environment of internet and big data with new means and approaches, enriches and deepens the research system of knowledge integration theory and method and also contributes to improving product innovation ability of enterprise.

结项摘要

制造大数据能够扩展产品研发数据的来源，为研发人员提供更为准确和全面融合的研发知识，但其重复、碎片化、分散、异构等特征，为企业开展产品研发知识集成与知识服务带来了挑战。本课题针对制造大数据的开放性、动态演化性和复杂关联性特点，从制造大数据的聚合管理入手，开展动态层次化知识集成方法和实时动态知识服务机制研究，以复杂产品研发开展案例研究。具体研究内容包括：（1）分析异构制造大数据可用性的基础上，研究基于异构制造大数据的全景式数据聚合管理方法。（2）针对产品知识领域不断细分和研发知识体系不断变化更新等特征，研究面向产品研发需求驱动的异质性知识集成方法。（3）分析研发人员个性化知识服务需求基础上，研究基于制造大数据的个性化实时动态知识服务机制。.项目研究了基于条件随机场的知识文本数据转换和语义标记，能够有效地标记和预测大规模文档集和语料库中的文本序列。针对知识文本的高层次数据特征的提取方法，能有效提高知识文本关键词等特征提取的效率。考虑用户需求特征建模的问题分类方法，从语义层面挖掘用户需求间的逻辑关系和内涵，为研发人员提供更准确的知识。考虑知识服务匹配机制的选择和评估方法，将需求层次结构的各节点要素与协同交互的知识单元进行匹配，包括了概念相似度、属性相似度和结构相似度计算，从而形成候选知识单元。考虑用户偏好的知识服务推送方法和决策支持系统，实现对知识需求的内容过滤，提供相似用户组的用户偏好个性化知识推送和决策支持。并将研究成果进行了应用。项目在国内外重要期刊上发表论文22篇，其中SCI/SSCI检索期刊论文6篇，CSSCI检索期刊论文10篇，获省部级科技进步一等奖2项。申请国家发明专利9 项。.以上研究为互联网和大数据环境下产品研发知识管理提供新的方式与途径，丰富和深化知识集成理论与方法的研究体系，也有助于提升企业产品创新能力。