【作者】 王兆卫；

【导师】 余忠华；

【作者基本信息】 浙江大学 ， 机械制造及自动化， 2010， 博士

【摘要】 近年来,我国在轴承、齿轮和紧固件等基础件生产取得了突飞猛进的发展,其总体质量也得到了不断的提升。然而,由于受到人员、设备、技术与管理等多重因素的制约,许多企业在其制造过程中质量问题频发,特别重复性的质量事件屡见不鲜,致使企业穷于应付,不但带来巨大的经济损失,而且产品质量无法得到有效的保证与改善,这已成为我国制造企业目前迫切需要解决的难点。在国家自然基金和武器装备预研基金的资助下,本文围绕问题驱动的制造企业质量改进这个主题,开展以问题驱动为核心的制造企业质量改进技术研究。第一章阐明了课题的背景,综述了国内外质量技术的研究现状,总结了企业质量改进发展的趋势和热点,从而引出了本文的研究内容是以“问题”为核心的制造企业质量改进技术及实现技术的研究;第二章分析我国制造企业在质量问题管理领域存在的困惑,提出质量问题特征分类管理的需求。在对目前特征分类管理方法开展对比研究的基础上,构建了基于正交分类方法的制造过程的QP-DOC质量问题特征分类管理模型。通过质量问题特征分类管理模型的构建,为问题驱动的质量改进活动的开展奠定了基础;第三章在问题特征分类管理模型的基础上,提出了问题驱动的质量改进模型,讨论基于ExGQM的业务过程描述模型、结构化质量改进模式的功能模块规划及知识资源管理等内容。从而为问题驱动的质量改进的实现奠定了基础;第四章讨论了基于知识地图的质量改进案例知识整合方法。重点讨论了知识地图系统形式化描述、知识节点识别和描述、知识导航及知识节点检索向量构建等内容;第五章在对制造企业生产过程监控研究成果进行分析的基础上,提出基于质量指标之间误差传播关系的生产过程监控点配置决策方法,使监控点的配置更加符合企业的实际情况;第六章分析贝叶斯网络在质量问题诊断方面的优势及其存在的瓶颈,面向工序质量问题诊断的贝叶斯网络解构方法研究。重点讨论了案例的贝叶斯诊断网络模型的分解、模型库构建和基于子网模型的贝叶斯网络重构等内容,力图为面向工序质量问题诊断的贝叶斯网络模型的构建提供一个借鉴的思路和方法;第七章在前面研究的基础上,以轴承企业为对象,研究问题驱动的质量改进在制造企应用。重点讨论了生产过程监控网络平台的构建等内容;第八章总结了全文的研究内容及创新点,并展望了未来的研究工作。更多还原

【Abstract】 In China, repetitive quality incidents continue to occur in the manufacturing process of our basic parts, which often forced companies to be engaged in event handling passively. Not only are caused huge economic losses, but also product quality can not be effectively improved. China’s manufacturing enterprises are urgent to resolve these difficulties. From the grants of the National Natural Science Foundation and the weapons and equipment pre-research Fund, this paper researched on the technologies for quality improvement of issue-driven manufacturing companies.In chapter one, the background of research, surveys the current research situation of quality improvement home and abroad is introduced and summarizes the development trends and hot spots of quality improvement, which leads to this dissertation research contents. Based on "issues", the research content includes quality improvement theory and key techniques for manufacturing companies.In chapter two, the perplexities in quality control field of our manufacturing companies is analyzed and puts forward the requirement for the management of classification for the quality problems features. Based on the comparative study of current management of features classification and orthogonal classification, the dissertation proposes the QP-DOC model, which lay the foundation for the issue-driven quality improvement..In chapter three, based on the model for features classification management, the dissertation proposes a issue-driven quality improvement model and discusses description model of business process based on ExGQM, structured function module plan of quality improvement mode and knowledge resource management. This lays the foundation for the implement of issue-driven quality improvement.In chapter four, the method of knowledge integration for quality improvement case based on knowledge map is disscussed. It includes formal description of knowledge map system, identification and description of knowledge nodes, knowledge navigator and knowledge node retrieval vector construction.In chapter five, the decision-making method of production process monitoring points based on error propagation between quality indexes is put forward. And the studies of researches on production process monitoring.In chapter six, the strengths and bottleneck of Bayesian network in quality issues diagnosis and the deconstruction method of Bayesian networks in process quality problem diagnosis oriented is analyzed. It emphasizes on the decomposition of Bayesian network model for diagnosis, model base construction, and sub-net model reconfiguration based on Bayesian network. It tries to provide a draw on ideas and methods for the construction of Bayesian network model for process quality problem diagnosis.Based on the researches of the front, chapter seven studies the application of issues-driven quality improvement in manufacturing enterprises with the bearing company as target.In chapter eight, the contents and innovations of the dissertation is summarized and lookedto the future research.更多还原