节点文献
面向复杂产品网络化制造的知识集成与应用关键技术研究
Research on Key Technologies of Knowledge Integration and Application Supporting Networked Manufacturing for Complex Products
【作者】 刘骄剑;
【作者基本信息】 南京航空航天大学 , 航空宇航制造工程, 2012, 博士
【摘要】 复杂产品网络化制造通过企业间的协作和资源共享实现产品生产过程的快速响应,以及企业资源的优化配置。制造过程技术复杂、知识密集,包括一系列的协同过程和复杂决策活动,需要有效利用各企业已有的知识资源,来提高复杂产品制造过程的快速决策能力。本文在分析复杂产品制造过程特点及制造企业知识管理和应用现状的基础上,提出将企业内外部知识资源进行有机组织并与制造过程进行集成,来达到制造过程的敏捷化、知识化,对知识集成机制及其关键技术进行了研究。本文的主要内容和研究成果如下:(1)对复杂产品网络化制造系统知识集成机制进行了研究。运用多视图建模方法对复杂产品网络化制造系统的功能结构、组织方式、运行过程、资源与数据管理进行了全面细致的描述;在此基础上,从复杂产品网络化制造过程对知识集成与应用的需求出发,建立了知识集成的总体模型,指出实现知识集成需要解决的关键技术,包括企业间分布异构知识的集成、知识间的关联集成以及知识与制造过程的集成;最后,提出了基于知识集成的复杂产品网络化制造系统体系结构,并介绍了系统的工作原理及特点。(2)对网络化制造企业间分布、异构知识的集成与共享机制进行了研究。提出了以本体、知识网络、知识元三层结构间的映射为特征的知识多粒度层次模型,将制造知识元作为制造过程知识管理和使用的基本单位;构建了制造任务、企业、资源、工艺、过程知识本体,并实现了基于OWL的本体形式化表达,为知识的共享提供语义基础,在此基础上,提出了基于扩展Web服务的知识集成方法;针对协作企业知识共享存在的风险和冲突,为了优化知识共享决策,运用博弈理论建立了知识共享完全信息博弈模型。(3)对网络化制造知识的关联分析和评价技术进行了研究。分析了现有的知识关联揭示方法在复杂产品网络化制造过程中应用的局限性,提出了将复杂网络理论运用于制造知识的关联分析与评价。通过对网络化制造知识的应用过程进行抽象,建立了制造知识复杂网络,并分析了复杂网络在网络化制造过程中的应用思路;提出了知识节点重要性评价方法,利用网络结构的“社团性”,提出了基于网络局域信息的关联知识凝聚算法,通过工艺协同设计实例对网络进行了统计和计算,并验证了算法的准确性和有效性。(4)对网络化制造过程中的知识搜索和推送技术进行了研究。提出了制造活动知识情境的三维度定义,在此基础上,使用本体图结构对知识需求进行了表达,并采用面向本体的SPARQL查询对需求进行语义扩展;针对知识推送引擎的实现,首先建立了知识推送过程的控制结构,并使用本体图对制造知识元的语义进行了描述,在此基础上,提出了基于本体图相似度计算的知识匹配算法,并以实例验证了该方法的有效性;最后,对推送结果的访问控制和可视化输出进行了研究。(5)依托相关项目研究背景,开发了基于知识集成的复杂产品网络化制造系统,介绍了系统开发背景,提出了系统的开发思路和技术体系结构,给出了系统运行环境、功能结构,并以某航天复杂产品的网络化制造为例,进行了系统验证。
【Abstract】 Networked manufacturing of complex products achieves rapid response of production processand optimal allocation of resources through collaboration and resource sharing among enterprises.The manufacturing process, which is technically complex and knowledge-intensive, includes a seriesof collaborative processes and complex decision-making activities. In order to improve the rapiddecision-making capacity in the manufacturing process, existing knowledge resources of enterprisesneed to be used effectively. In this thesis, the characteristics of manufacturing process of complexproducts, present situation of knowledge management and application in manufacturing enterprisesare analyzed. On this basis, in order to achieve agile and knowledge-based manufacturing process, theidea of organizing internal and external knowledge resource effectively and integration withmanufacturing process is put forward. The mechanism and key technologies of knowledge integrationare studied. The main contents and research findings are as follows:1. The knowledge integration mechanism of networked manufacturing system for complexproducts is studied. First, a comprehensive and detailed description of functional structure,organization, running process, resource and data management of the system is provided. Then, basedon the requirement for knowledge integration and application in networked manufacturing, the overallmodel of knowledge integration is built. Key technologies need to be addressed in knowledgeintegration are analyzed, including the integration of distributed and heterogeneous knowledge fromdifferent enterprises, associated integration among knowledge, and knowledge integration withmanufacturing process. Finally, the networked manufacturing system architecture based on knowledgeintegration is put forward, and then the working principle and characteristics of the system aredescribed.2. The integration and sharing mechanism of distributed and heterogeneous knowledge fromdifferent enterprises is studied. A hierarchical model based on multi-granularity knowledge isproposed, which is characterized by mapping among ontology, knowledge network and knowledgeelements. Manufacturing knowledge element is regarded as the basic unit of knowledge managementand use in manufacturing process. In order to provide semantic foundation for knowledge sharing,domain ontology of manufacturing task, enterprise, resource, technology and process is established,and then formalized expression of the ontology based on OWL is realized. On this basis, a knowledgeintegration method based on extended Web Service is proposed. Considering the risk and conflictduring collaborative enterprise knowledge sharing, aiming at optimizing the process, a complete information game model of knowledge sharing based on game theory is established.3. Knowledge evaluation and association analysis technology in networked manufacturing isstudied. First, the application limitations of existing method of knowledge association analysis innetworked manufacturing process are analyzed, and then applying complex network theory to theassociation analysis and evaluation of manufacturing knowledge is put forward. By abstracting theapplication process of networked manufacturing knowledge, the complex network of manufacturingknowledge is built, and its application mechanism is analyzed. On this basis, the importanceevaluation method of knowledge nodes is proposed. Utilizing the community characteristics ofnetwork, an associated knowledge cohesion algorithm based on local information is proposed. Finally,a collaborative process design instance is used to illustrate the statistics and calculations, and to verifythe accuracy and validity of the algorithm.4. Knowledge search and push technology in networked manufacturing are studied. First, theknowledge contexts of manufacturing activities are described from three dimensions. On this basis,knowledge requirement is expressed using ontology graph structure, and extended semantically withSPARQL query. For the realization of knowledge push engine, the process control structure ofknowledge push is established first, and then the semantics of manufacturing knowledge elements isdescribed using ontology graph. On this basis, a knowledge matching algorithm based on similaritycomputation of ontology graph is put forward, and proved to be effective by application to an instance.Finally, access control and visual output of push results are studied.5. Relying on the above studies and project background, the networked manufacturing system forcomplex products based on knowledge integration is developed. Development background,development ideas, technical architecture, running environment and function structure of the systemare introduced. Finally, the system is verified through a networked manufacturing example of anaerospace complex product.
【Key words】 complex products; networked manufacturing; knowledge integration; ontology; complex network; knowledge push;