【作者】 张晓阳；

【导师】 吴慧中； 孙宇；

【作者基本信息】 南京理工大学 ， 计算机应用技术， 2005， 博士

【摘要】 复杂系统在国民经济中扮演着重要角色,但其日益复杂的结构和控制系统造成了其故障的多发性特点。对于用户来说,如何提高和保障系统的工作可靠性与有效性是一个亟待解决的问题;而对于装备制造企业来说,产品质量的竞争焦点集中到了设计开发阶段和售后服务阶段,为客户提供高可靠性的产品、迅速有效的诊断维护服务是众多装备制造企业所面临的共同问题。这些问题的解决具有重要的实际意义。 基于上述动机,本文提出了面向复杂系统生命周期的诊断维护理念,详细阐述了其内涵和三维视图,目的是以网络和远程通信技术为支撑,实现装备系统生命周期时间内故障诊断维护数据信息的集成和管理,建立、完善和保持复杂系统生命周期时间内的信息流,并以此为基础,开展相关的应用研究。本文的主要成果与创新为: 首先,提出了面向复杂系统生命周期的诊断维护理念,以此为指导研究了面向复杂系统生命周期的诊断维护系统,提出了其描述模型、总体结构和逻辑结构,分析了该系统与企业管理信息系统相融合的必要性、可行性及意义,对构建相关系统具有指导意义。 其次,提出了诊断维护知识的应用模型,定义了知识建模和知识单元的概念,建立了知识单元模型,并提出了基于知识单元的知识共享和创新应用方法。通过知识单元和知识编码实现了诊断维护知识的组织、管理和重用,提升了企业诊断维护知识资源的管理和应用水平,有利于产品的设计优化、可靠性提高和诊断系统的开发,架起了系统设计和诊断系统开发之间的桥梁。 再次,提出了从知识单元生成诊断规则和贝叶斯网络的方法思路,以知识单元为基础,设计了集成智能诊断系统总体结构。并提出了用粗糙集确定其提取的诊断规则的可信度的方法,给出了规则推理、实例推理和贝叶斯网络推理的集成诊断策略。其中贝叶斯网络很好地解决了故障的不确定性推理问题。 第四,指出了复杂系统健康管理的内涵及其监控模式和信息视图。提出了基于生物免疫机制的分布式多Agent健康监测体系,建立了免疫Agent模型和基于免疫机制的监测系统工作模型。提出了复杂系统健康评估的方法,并以故障的危害性大小来衡量其对系统健康状态的影响。 最后,提出了面向复杂系统生命周期的诊断维护系统的实施策略,以轨道交通自动门的诊断维护系统为例,介绍了其总体方案,以及计算机网络、无线通信等支撑技术的集成和系统功能的实现方法,并给出了系统的测试运行结果。更多还原

【Abstract】 The Complex System plays an essential role in the national economy. But the increasing complicated structure and control system make it more easily fault. How to improve and support its reliability and validity becomes an urgent problem for customers. On the other hand, product competition focuses on the development stage and after sale stage, many manufacturers have to face the problem to provide more reliable product and rapid effective diagnosis and maintenance service for customer. How to resolve these problems is very valuable.Based on the above motivation, this dissertation proposes the Lifecycle-Oriented Diagnosis and Maintenance Philosophy for complex system, expounding its connotation and three-dimensional view in detail, which supported by network and remote communication technology aims to integrate and manage all fault data and information in the complex system’s lifecycle, build and keep the lifecycle information flow, and develop relative applications based on the flow. The main contents and innovations are as following:Firstly, it proposes the Lifecycle-Oriented Diagnosis and Maintenance Philosophy, which guiding the research on Lifecycle-Oriented Diagnosis and Maintenance System for Complex System. And then the system’s description model, architecture and logical structure are presented. And the necessity, feasibility and meaning of the system’s integration with enterprise management system are analyzed, which can be used for reference when developing similar system.Secondly, the Application Model of diagnosis and maintenance knowledge is proposed, and the concepts of Knowledge Unit and Knowledge Modeling are defined. The methods of knowledge sharing and innovative using are presented based on the built Knowledge Unit Model. By using the Knowledge Unit Model and Knowledge Code, the diagnosis and maintenance knowledge’s organization, management and reusing are implemented, which improves the management and application level of enterprise’s knowledge resources. This work not only benefits product’s design and reliability but also bridges the system design and development of fault diagnosis system.Thirdly, the methods of producing rules and Bayesian network from Knowledge Unit are presented, the architecture of an integrated intelligent diagnosis system is designed based on Knowledge Unit. The method of using rough sets to acquire rules from condition data and ascertaining rules’ creditability are also put forward. Then the integration strategy to combine rule-based reasoning, case-based reasoning and Bayesian Network reasoning is given, Bayesian Network can resolve the faults’ uncertainty problem effectively.Fourthly, the meaning of Complex System Health Management is discussed, and its monitoring mode and information view are presented. Then the Distributed Multi-Agent Health Monitoring Architecture and its functioning mechanism based on immune更多还原