节点文献
制造装备远程监控故障诊断系统研究
Study on Manufacturing Equipment Remote Monitoring&Fault Diagnosis System
【作者】 王海;
【导师】 王宛山;
【作者基本信息】 东北大学 , 机械制造及其自动化, 2012, 博士
【摘要】 随着Internet/Intranet技术的迅速发展,现代制造模式正朝着集成化、网络化、全球化方向发展,涌现出智能制造、网络化分散制造等先进制造理念。以数控机床、PLC控制的制造装备、机器人等为代表的现场设备作为各种现代化制造系统的最基础装备,其正常运行是现代制造企业能够健康运转的根本保证。现场设备技术迅猛发展,结构日趋复杂,使得针对这些设备及其工艺过程的故障诊断工作越来越重要。本文分析了现代化装备制造系统在新形式下面临的挑战,在现有支撑技术条件下及其可见的发展趋势下,针对以CNC装置和PLC为控制器的制造装备的远程监控故障诊断进行深入研究。研究工作从构建由现场控制器到远程分布式系统的故障诊断功能的体系结构及关键使能技术出发,以系统的通用性为重要原则,提出远程监控故障诊断系统(EMID)的结构功能模型;提出实时数据平台RDSP功能模型,通过其向上层提供集成数据服务,并对实时数据的采集、管理、服务功能进行深入研究;提出“结构-故障树”方式的装备诊断知识组织模型,在其基础上,采用基于智能引导的融合诊断方法实现高效的诊断维护工作。本论文的主要研究工作具体如下:(1)分析现代制造环境下的装备故障诊断工作的功能需求,提出面向生产现场的远程监控故障诊断系统结构框架和功能模型,对其作出了系统化、全局性阐述,分析了实时数据服务和融合诊断方法等关键技术,为系统的具体实现奠定基础。(2)分析了生产现场实时数据接口特点和信号采集内容,分析了在诊断系统中实时数据服务功能的需求,提出基于扁平结构的实时数据平台RDSP,针对基于广域网分布式实时数据访问问题,采用散列映射访问机制和Socket线程池等技术,解决数据服务的实时性、并发性问题,通过实际检测达到预计效果。RDSP的研究包括平台的结构、内存实时数据维护、现场实时数据采集、实时数据服务、历史数据维护等具体功能的设计思想和实现方法。(3)充分分析制造装备固有特点和其故障特点,同时分析了应用于实际的各种故障诊断技术方法的特点。针对现有的各诊断理论功能单一的局限,提出基于“结构-故障树”知识组织方法和基于智能引导的层次融合诊断推理方法,解决了诊断知识和诊断方法在融合诊断工作中组织问题。在收集和分析大量数控机床等故障实例基础上,提出基于诊断过程性知识引导的诊断方法,基于设备-故障特征知识引导的诊断方法。同时,在经验知识不够充分时,提出基于RDSP实时/历史数据的ANFIS知识发现方法,实现面向数据的故障诊断功能。EMID层次诊断功能从纵向配置角度,提出基于现场控制器故障诊断的实施原则;在讨论和分析网络化协同工作环境下,诊断系统基本结构、功能实现的基础上,给出分布式故障诊断工作的流程及模糊层次评价方法。(4)在上述原理研究的基础上,在教育部博士基金项目“基于网络化的数控加工智能化研究”(145015)和教育部重点科研项目“基于CSCW远程控制的机械装备协调设计技术研究”(105057)的资助下,结合企业的CIMS及管控一体化开发等大量横向课题,综合运用CNC、PL、OPC、.NET、XML Web Service、AJAX、数据库等技术,以建立开放式远程监控故障诊断平台为基础,开发了面向网络化的EMID远程监控故障诊断原型系统,实现对上述复杂设备的状态监测与故障诊断的远程化、智能化,验证了论文提出的远程监控故障诊断系统的思想和方法的可行性、实用性。
【Abstract】 With the rapid development of Internet/Intranet technology, modern manufacturing mode is moving towards the integration networking and globalization. The intelligent manufacturing, distributed network manufacturing and other advanced concepts come forth successively.The manufacturing equipments include the CNC machines, the PLC controlled equipments and the robots etc. These were the most basic equipment of the modern manufacturing systems. The health running of the equipments is the fundamental guarantee for the manufacturing enterprises. The technologies of equipments are developing rapidly, so the increasingly complex structure&process for these equipments make the fault diagnosis more and more important.This paper analyzed the challenge of modern equipment manufacturing system in present. With the support of the existing technologies and visible technologic trend, the remote monitoring and fault diagnosis system was studied deeply.The research starts from the system architecture and key enabling technology. The versatility of system as an important principle, paper proposed the structure and function model of the diagnosis system, raised real-time data service platform (RDSP) features model, which provided the data services to the upper integration. The real-time data acquisition, data management and service function were in-depth studied. The paper proposed diagnostic knowledge model of structure-fault tree that organize the knowledge effectively. A hierarchy integral intelligent diagnosis process was researched, which can be directed by the experience knowledge and data from RDSP real-time system. The collaborative and evaluation methods of coordination in the network diagnosis environment were also discussed. The main research works in this paper as follows:(1) Analyzed the function requirements of the fault diagnosis system in the modern manufacturing environment, proposed integration framework and functional model, and made a systematic, elaborated description, which was the foundation for forward implementation. (2) Analyzed the feature of real-time data content and acquisition interface from the production site. Analyzed the demand for real-time data services, proposed the flat structure based of RDSP. For real-time data application in WAN, the use of the hash map and Socket thread pool technology to solve real-time data services, concurrency issues, practice testified that it obtained the desired effect. RDSP research includes the platform structure, the memory maintenance of the real-time data, real-time data acquisition, historical data maintenance, data services and other specific features.(3) Analyzed the inherent characteristics of manufacturing equipment and its fault characteristics adequately, analyzed the characteristics of various practical diagnosis methods, in order to solve the singularity limitations of the existing theory, proposed structure-fault tree methods of knowledge organization. An intelligent hierarchy integral diagnosis inference model was researched. After the collection and analysis of a large number of CNC machine failure samples, the inference methods that were based on the diagnosis process and the characteristics of fault was extracted respectively. Meanwhile, in the case of knowledge is insufficient, proposed a data-oriented ANFIS inference method.In the vertical view of structure, the diagnosis implementation in field controller was explained. In the distributed collaborative environment, the basic structure of a fault diagnosis, function and evaluation was discussed. The integrated diagnostic functions make diagnosis maintenance process accurate and efficient.(4) On the basis of the above-mentioned principle, under the auspices of the Ministry of Education Doctor Foundation,"The CNC-based Intelligent Network Research"(145015) and "Remote Control Coordination Design Technology Machinery and Equipment Based on CSCW "(105057), and other practice CIMS projects, using the CNC, PLC, OPC,. NET, XML Web Service, AJAX, database technologies, developed the remote monitoring and diagnosis prototype system, that was suitable for both Web Client and WinForm Client appliances, analyzed the key technologies to system development, which has checked the feasibility and practicality of the proposed idea.
【Key words】 remote monitoring fault diagnosis; CNC equipment; PLC system; real-time dataservice platform; structure-fault tree; guide diagnosis; fuzzy inference; ANFISinference;