超速离心机故障诊断专家系统研究

发布时间：2018-06-01 11:36

本文选题：超速离心机 + 故障诊断　；参考：《湖南大学》2012年硕士论文

【摘要】：离心机是依靠转子高速旋转产生的离心力进行分离、浓缩、提纯、分析等操作的仪器。一般将最高转速不低于30000转/分的离心机称为超速离心机。超速离心机目前已经广泛应用于医疗、化工、食品、制药和石油等领域，，是我国重点研究的科学仪器之一。超速离心机在运行过程中可能产生故障，这些故障因为发生在高速旋转的过程中，会造成极其严重的经济损失和人身危险。因此研究一种超速离心机的故障诊断专家系统，针对主要故障模式找出可能的原因是非常必要的。专家系统属于人工智能范畴，是具有包含某一领域大量的专家知识和经验，模拟人类专家的决策过程和思维方式的计算机程序。一个好的专家系统具有设备故障诊断速度快、准确率高等优点。本文首先对超速离心机和故障诊断的发展进行了阐述，并且详细介绍了超速离心机的工作原理、结构组成和参数要求等。为了建立完善的超速离心机故障诊断专家系统的知识库，本文对超速离心机在工作过程中可能出现的故障进行了总结，并采用多种传感器及其信号采集电路对信号参数进行提取，设计了合理的传输协议，使得系统能够实时的对测量信号进行检测，提供故障诊断的依据。本文采用产生式知识表示法和框架式知识表示法相结合的方法，对知识库里的专家知识进行表示，并结合Access2003对知识库进行分类构建，建立了完善的知识管理模块，并且对知识库的自学习功能进行了讨论。系统采用贝叶斯网络推理方法结合正向推理策略设计了系统的推理机制，能够推理出故障发生的原因及其概率，并采用MSBNx(Microsoft Bayesian Network Editor)对贝叶斯网络推理方法进行仿真验证，取得了预期的功效。MSBNx是微软公司出品的贝叶斯网络编辑器，专门用于贝叶斯网络的计算和仿真。系统采用预制文本法和追踪解释法相结合的方式建立了解释机制，能够对故障的处理过程进行较为完善的解释。论文最后具体分析了超速离心机故障诊断专家系统的软件结构，采用VC++2005作为软件开发工具，完成了包括状态检测、故障诊断和知识管理模块在内的软件设计。
[Abstract]:Centrifuge is an instrument for separation, concentration, purification, analysis and so on. Generally, the maximum speed of not less than 30000 rpm / min centrifuge is called an overspeed centrifuge. Ultra-speed centrifuge has been widely used in medical, chemical, food, pharmaceutical and petroleum fields. It is one of the most important scientific instruments in China. In the process of running, the ultra speed centrifuge may have faults, which will cause extremely serious economic loss and personal danger because of the high speed rotating process. Therefore, it is necessary to study an expert system for fault diagnosis of ultracentrifuge and to find out the possible causes for the main fault modes. Expert system belongs to the category of artificial intelligence. It is a computer program that contains a lot of expert knowledge and experience in a certain field and simulates the decision-making process and thinking mode of human experts. A good expert system has the advantages of fast equipment fault diagnosis and high accuracy. In this paper, the development of ultra speed centrifuge and fault diagnosis are described, and the working principle, structure composition and parameter requirement of ultra speed centrifuge are introduced in detail. In order to establish the knowledge base of fault diagnosis expert system for ultracentrifuge, this paper summarizes the possible faults in the working process of ultracentrifuge. The signal parameters are extracted by a variety of sensors and their signal acquisition circuits, and a reasonable transmission protocol is designed, which enables the system to detect the measured signals in real time and provide the basis for fault diagnosis. In this paper, the expert knowledge in the knowledge base is represented by the combination of the production knowledge representation method and the frame knowledge representation method, and the knowledge base is classified and constructed with Access2003, and a perfect knowledge management module is established. The self-learning function of knowledge base is also discussed. The system uses Bayesian network reasoning method and forward reasoning strategy to design the inference mechanism of the system. It can infer the reason and probability of the fault, and use MSBNx(Microsoft Bayesian Network Editor) to simulate the Bayesian network reasoning method. MSBNx is a Bayesian network editor produced by Microsoft, which is specially used in the calculation and simulation of Bayesian network. The system uses the prefabricated text method and the tracing interpretation method to establish the interpretation mechanism, which can explain the fault handling process perfectly. Finally, the software structure of fault diagnosis expert system for ultracentrifuge is analyzed in detail. VC 2005 is used as the software development tool, and the software design including status detection, fault diagnosis and knowledge management module is completed.
【学位授予单位】：湖南大学
【学位级别】：硕士
【学位授予年份】：2012
【分类号】：TH165.3;TP182

【参考文献】