【作者】 徐河杭；

【导师】 顾新建；

【作者基本信息】 浙江大学 ， 机械制造及其自动化， 2010， 博士

【摘要】 面对全球经济一体化,客户需求个性化以及产品交货高速化的经济时代,现代制造企业正面临着快速响应客户需求、缩短产品交货期、提高产品质量和改进产品服务的压力。在这种压力下,企业需要从产品全生命周期管理(PLM)的角度,对企业各个环节的数据进行分析管理和挖掘,形成满足各部门需要的、满足企业高层管理者需要的信息和知识,以支持他们的决策,快速响应制造企业内外部环境的变化,实现由“中国制造”向“中国创造”的转变,并促进制造企业向现代制造服务业的转型。因此,论文以国家863项目和浙江省重大科技项目为背景,提出了面向PLM的数据挖掘技术和应用研究。主要内容有：第一章主要阐述了论文的研究背景,提出制造企业对数据挖掘的需求,特别是面向PLM的数据挖掘的需求,总结了目前制造企业应用数据挖掘存在的问题。介绍了数据挖掘的基本概念、相关理论和国内外的研究现状,总结了数据挖掘在产品生命周期各阶段的应用。给出了论文研究的内容和体系结构,并讨论了论文研究的意义。第二章提出了面向PLM的四层数据挖掘体系结构,包括数据层、方法层、结果层和应用层。讨论了面向PLM的数据挖掘的关键技术,包括分类、编码、有序化和集成技术等等。分析阐述了面向PLM的数据挖掘过程中需要用到的数据、数据挖掘方法、数据挖掘结果和应用及其它们的各种分类模型、编码模型和集成模型。第三章从PLM的需求出发,将各种挖掘方法模型分为通用模型和专用模型。通过对数据挖掘方法模型的字典描述和程序文件库,采用“模型字典+模型程序库”形式的模型存储方法和模型库构架,利用面向对象的方法构建了一个数据挖掘方法模型库。并通过模型库管理系统对模型进行统一管理和维护。第四章设计了一个面向PLM的数据挖掘流程,通过挖掘需求与数据属性和方法模型属性描述的匹配程度进行数据和模型选择,快速找到合适的数据挖掘方法模型和数据源。然后从面向PLM的数据挖掘流程出发,研究开发了一个面向PLN的数据挖掘系统。对系统的功能流程进行了阐述；对系统实现所使用的工具技术和相关功能模块进行分析；实现了一个数据挖掘流程和对过程中数据挖掘需求、模型和结果的管理。第五章对数据挖掘在产品生命周期中的应用进行了研究。首先从数据挖掘在产品生命周期各阶段的应用研究出发,详细描述了不同阶段的各种应用需求、数据处理以及相关挖掘方法,并得出了相关挖掘结果。如设计阶段的专利技术分析等等,生产制造阶段的质量因素分析等等,销售服务阶段的客户分类和销售预测等等挖掘应用。然后以质量数据为例,对面向PLM的数据集成挖掘应用进行了研究。最后对面向PLM的数据挖掘结果集成应用进行了研究。主要包括产品生命周期内各部门之间以及面向企业高层管理者需求的数据挖掘结果集成。第六章对全文工作进行了总结,并对后续研究工作进行了展望。更多还原

【Abstract】 The modern manufacturing enterprises are faced with pressures on responding to customer’s demand rapidly, shortening product delivery time, improving product quality and services, because of economic era with global economic integration, personal demand and high-speed of product delivery. In this pressure, data analysis and mining for product lifecycle management (PLM) are needed for enterprises, to meet the needs of various departments and leaders for decision-makings, which makes the enterprise to achieve changes on "Made in China" to "Created in China", and promotes transition to manufacturing services enterprises by responding to the changes of internal and external environment rapidly. A research on technologies and application of data mining for product lifecycle is provided in this dissertation, based on the national 863 project and the important technical project. The main content is as follows:In chapter 1, the background of research is introduced. The demand of data mining in manufacturing enterprise especially for PLM (Product Lifecycle Management) is advanced. The current problems in data mining applications in manufacturing enterprise are summed up. The basic concepts, related theory and research status of data mining are introduced, and the application of data mining in product lifecycle stages is summarized. The content and structure of the thesis are given, and the significance of the research paper is discussed.In chapter 2, the four-layer system frame of the data mining for PLM is proposed, including the data layer, method layer, the results layer and application layer. The key technology of data mining for PLM is discussed, including classification, coding, ordering and integration technologies. The data, methods, results and applications of data mining and their various classification models, coding model and integrated model are expounded, which is needed in the data mining process for PLM.In chapter 3, a variety of mining models are divided into general models and specific models for the demand of PLM. A data mining model libraries is constructed using object-oriented method and "model dictionary+model program libraries" form. The model is managed and maintained uniformly by Model Base Management System.In chapter 4, a data mining process for PLM is designed. The appropriate data mining models and data sources can be found quickly through the matching degree of the attributes describe of data and method models with data mining needs. A data mining system for PLM is researched and developed starting from the data mining process for PLM. The function processes of the system are described. The technologies and tools which used by the system and the relevant function modules of system are analyzed, which manages the data mining process, demands, models and results.In chapter 5, the data mining application of product lifecycle is researched. First, a various needs in application, data processing and related mining method models in the different stages are described and the related mining results are obtained. Such as the analysis of patents in the design phase, the production quality factors in the manufacturing stage, customer segmentation and sales forecast in the sales service stage, etc. And then the data mining applications based on quality data integration for PLM are studied. Finally, the application of data mining results integration for PLM is studied. It includes the integration of data mining results for the demand of various departments of product lifecycle and business leaders.In chapter 6, all achievements and innovations of this dissertation are summarized and the future research direction is put forward.更多还原