【作者】 刘刚；

【导师】 李兆前；

【作者基本信息】 山东大学 ， 机械工程， 2012， 博士

【摘要】 制造业是国家的基础性、前沿性、支柱性与战略性的产业,竞争的焦点是产品创新和服务。但是,受数据获取和处理机制的影响,多数企业不能建立产品全生命周期信息库,无法真正实现产品数据的无缝集成,以及新数据的快速获取和旧数据的高效复用,阻碍了产品创新和服务的发展。本文基于数据交换、元模型、数据建模、数据追踪追溯等技术和理论,从产品数据获取、生命周期统一数据模型和产品数据闭环算法等方面进行了研究,建立了产品数据闭环管理模式,实现了数据的有效获取、集成、供给和复用,扩大了数据使用范围,提升企业的产品数据管理能力和效率,支持产品创新和服务的发展。在数据获取的研究中,分析了产品数据内容、形式和状态以及BOL、MOL和EOL三个阶段的数据特点,总结了四种数据获取方式,建立了数据交换模式,形成了数据获取框架。在数据分析的基础上,为了消除数据交换中的异构性,利用XML技术,建立了产品数据交换模式：树、队列、哈希表和元数据等数据结构用来管理数据状态格式和数据逻辑格式,共同支持结构化数据和非结构化数据的交换。搭建了数据获取框架,涵盖数据存储、管理、传输和获取,支撑主动提交、被动提交、集成获取和自动获取四种获取方式,为数据获取提供了可行的方案。在产品数据模型研究中,提出了产品形态,建立了产品形态分治的统一数据模型,给出了数据模型的定制机制。产品在其生命周期中经历了模型产品、设计产品和物理产品三种形态,产品形态的数据特点及其演化规律是产品生命周期数据统一管理的决定因素。利用面向对象和分治的思想、UML技术和元数据技术,建立了产品生命周期统一数据模型,由模型产品数据模型、设计产品数据模型和物理产品数据模型组成,这三个模型分别由单一数据模型、“核心数据模型+阶段规则过滤器”、“核心数据模型+阶段专有数据模型”的形式实现；底层产品形态对象间映射管理产品进化关系,集成三个产品形态模型为一个整体。消除了产品形态数据管理模式的差异,实现整个生命周期数据的统一管理,保证数据的一致性和完整性,有利于控制版本向上蔓延。为了适应业务不断发展的需要,基于元模型理论,提出了产品数据模型的定制机制。分析了可定制数据模型的内涵,研究了数据模型定制的实现技术。该机制扩大了数据模型的适用范围,减少了二次开发工作量,帮助产品生命周期数据管理系统快速实施。在产品生命周期数据闭环管理模式研究中,建立产品形态进化树,提出了跨产品形态的闭环管理模式,给出了数据追踪追溯的算法,实现产品数据同源和借用管理。产品形态进化树管理模型产品、设计产品到物理产品的进化过程,利用产品形态进化矩阵,描述产品内部的结构和对象信息,产品之间的零部件映射信息和同源、借用信息。建立了涵盖模型产品、设计产品和物理产品的数据闭环管理模式,提出产品形态间属性和结构独立映射机制,形成相应的映射规则。研究数据追踪和追溯的算法,包括产品对象和结构分别追踪和追溯的数学模型以及映射矩阵关联模型。给出了同源矩阵和借用矩阵,实现了设计零部件的同源和借用管理,以及物理零部件的模型同源和设计同源管理,使产品生命周期数据管理系统能自动为相关的业务提供数据支持。基于上述研究,探讨了设计产品和物理产品典型数据相互支持的方法。在数据闭环管理模式的支持下,同源物理产品故障数据反馈到设计产品,形成了海量故障样本,通过统计分析,获得每种故障模式的故障频率等级、危害程度等级、可检测性等级和风险优先数,实现设计故障模式及影响分析,支持设计改进。该方案将故障实例数据和潜在故障模式数据有效融合,提高分析的客观性和准确性,帮助提升产品的可靠性。同时讨论了维修对象到设计数据源头的追溯方法。最后,设计研发了支持产品生命周期数据闭环管理的PLM系统—InforCenter,并面向船舶行业进行了示范应用。InforCenter统一管理模型船舶、设计船舶和物理船舶的数据,支持数据的前馈和反馈,提高了企业数据管理能力,实现了设计数据和服务数据的相互支持。更多还原

【Abstract】 Manufacturing industry is the fundamental, advanced and strategic cornerstone industry, and its competitiveness focuses on innovation and service of product. Influenced by acquisition and handling mechanism of product data, full lifecycle information store, seamless integration of data, rapid acquisition of new data and effective reuse of old data are hardly achieved. And the development of product innovation and service is blocked. Therefore, this paper studied the data acquisition, lifecycle unified data model and closed-loop data management based on data exchange, meta-data, meta-model, data modeling, data tracking and reverse tracing. The proposed closed-loop management of product data facilitates the effective management, supply and utilisation of product data, improves the product quality and sharpens the competitiveness edge of enterprises.Study of data acquisition-XML was used to build data exchange model, in order to eliminate the heterogeneity of data to be exchanged. The formats of structured data and unstructured data were managed by meta-data, trees, queues and hash tables. Structured data were divided into static data and dynamic data. Dynamic data comprised time-based, event-based and message-based data. Thus product data could be exchanged among several PLM-centered agents. To support data acquisition and exchange modes, a data acquisition framework comprising4layers of data storage, management, transmission, and acquisition was built. The framework was a feasible scheme for data acquisition in that it enabled positive/negative upload, and integrated/automatic access of product data. Study of unified data model-The product morphology including model product, design product and physical product were proposed. Data characteristic and evolution laws of product morphology are the key factors for unified data management. Object Oriented (00), UML and meta-data were used to build unified data model of product lifecycle, including model-product data model, design-product data model and physical-product data model. Model-product data model was a single data model. Design-product data model was made up of core data model and phase ruler filters. Physical-product data model comprised core data model and phase private data models. All the data were unitively managed, and effectively stored using mapping of fundamental objects to manage relations of evolution. In order that data model fits development of business, a customized approach of data model was exhibited. Its framework and customized method were established. The approach extended the application range of data model, reduced the workload of customized and further development, and achieved rapid implementation of PLM system.Study of data closed-loop management-Product-morphology evolvement tree, which described the evolvement among three product forms, managed structre and object information of one product form, and mapping, homology and borrow information of components among product forms. Data closed-loop management model supporting data integration and share among three product forms was put forward. Respective mapping of attribute and structure was discovered, and the mapping rules were built too. Mathematical models of data tracking and data reverse tracing were formulated, including object tracking, structure tracking, object reverse tracing, structure reverse tracing and mapping matrix tracking. Homology matrix and borrow matrix were used to achieve homology and borrow management of design components, and homology management of physical components. Then the capacity of PLM automatically supplying data for business was improved.Based on the studies above, an interactive scheme of product design and service was proposed. With data closed-loop management model, failure cases from homologous physical product could feed back to design phase and then Severity degree, Occurrence degree, Detectability degree and Risk Priority Number (RPN) of every failure mode were got. Design Failure Mode and Effects Analysis (DFMEA) were conducted, and design improvement and data fusion of failure cases and failure mode were achieved. The objectivity and precision of DFMEA were improved. Then the approach of data reverse tracing from service object to its design provenance was discussed.Finally, a PLM system-InforCenter achieving data closed-loop management was developed and implemented on shipbuilding plant. InforCenter managed the data of model ship, design ship and physical ship, and their data could feed forward and back. Then the data of design and service could support each other, and the data-management capacity of shipyard was improved.更多还原