基于概念驱动的变速箱故障诊断机理研究

发布时间：2018-03-11 09:45

本文选题：概念驱动　切入点：规则　出处：《吉林大学》2011年硕士论文　论文类型：学位论文

【摘要】：本文在总结前人研究的基础上,研究了基于概念驱动的变速箱故障诊断机理。概念驱动诊断是一种根据过去知识和经验进行的诊断。研究了如何模拟人利用感知过程中的信息处理方式来建立诊断行为模型,发掘各种可能的经验信息介入方式、概念驱动诊断算法的逻辑框架和算法的实现方法。给出一个新的智能诊断理论算法框架,满足在生产节拍要求下对变速箱生产线终端快速异常噪声问题的检测。在复杂的工作环境下,装配线上工人所定义的敲响、异响、叉响、噪音等感知性故障状态是固定的诊断结论。所以,将人的认知能力辨别的故障类型作为诊断概念来指导变速箱故障诊断,就可解决故障结论界定不统一、特征参数提取缺乏概念分类依据等不一致问题。本文基于认知心理学关于人类认知过程的信息加工原理,探索了基于概念驱动满足生产节拍需要的变速箱异常噪声问题。研究中首先利用成熟的信号采集分析技术收集变速箱敲响、异响、叉响、噪音等典型故障的振动信号,并计算其振动特征参数值。研究各种故障与相应故障状态下齿轮、轴承、轴以及箱体等质量问题相关联的振动特征值,区分哪些特征对哪些问题敏感,以及敏感条件是什么。变速箱故障的振动与噪声表现有时域特征和频域特征,需要研究这些特征之间的关系,以及这些关系变化与变速箱故障状态问题的关联性,并研究这些关系变化与变速箱制造质量问题的关联性。其次,建立一个可供后续变速箱异常噪声问题诊断系统开发研究的知识库,这个知识库将记录变速箱典型故障振动特征与变速箱零部件制造质量问题相关联的基本知识。其次,模拟人的认知加工过程,依据振动特征之间的关系对变速箱异常噪声进行了诊断识别。为此,本文提出了采用变速箱故障特征维分数维分析方法、支持向量机与基于规则的推理相结合的故障诊断方法来建立变速箱故障诊断模型。本文首先研究了具有概念驱动意识的推理规则；研究了各种模式的触发规则,以及使用这些规则进行诊断的诊断规则,建立一个可供后续变速箱典型质量问题研究的知识库。这个知识库将记录变速箱典型故障振动特征与变速箱零部件制造质量问题相关联的基本知识。知识库的建设将为后续实施基于概念驱动的诊断算法研究提供基础数据支撑,并在此基础上,对基于概念驱动的变速箱故障诊断的机理进行了研究。其次,运用人类知觉过程的信息加工原理、数据挖掘技术、人工智能原理以及信息论等理论方法研究了变速箱敲响、异响、噪音、叉响等质量问题的“概念”描述方法。提取了诊断过程中的概念驱动方法规则,阐述了故障诊断过程中的信息加工与概念驱动机制,构建了具有概念驱动意识的变速箱质量问题智能诊断算法框架。
[Abstract]:On the basis of summing up the previous studies, In this paper, the mechanism of gearbox fault diagnosis based on concept driving is studied. Conceptual driven diagnosis is a kind of diagnosis based on past knowledge and experience. This paper explores various possible methods of empirical information intervention, the logical framework of concept-driven diagnosis algorithm and the implementation method of the algorithm. A new framework of intelligent diagnosis theory algorithm is presented. To meet the requirements of the production beat to the transmission production line terminal fast abnormal noise detection. In a complex work environment, the perceived fault states defined by workers on the assembly line, such as ringing, abnormal, fork, and noise, are fixed diagnostic conclusions. Taking the fault type identified by human cognitive ability as the diagnosis concept to guide the gearbox fault diagnosis can solve the problem that the fault conclusion is not uniform. Based on the information processing principle of cognitive psychology on human cognitive process, this paper explores the abnormal noise problem of gearbox based on concept driven to meet the needs of production rhythm. In the research, the vibration signals of typical faults such as sound, abnormal noise, fork noise, noise and so on are collected by using the mature signal collection and analysis technology, and the characteristic parameters of vibration are calculated, and the gears under various faults and corresponding faults are studied. Vibration eigenvalues associated with mass problems such as bearings, shafts and boxes, distinguishing which features are sensitive to which problems and what are the sensitive conditions. The vibration and noise of gearbox faults exhibit time-domain and frequency-domain characteristics. It is necessary to study the relationship between these characteristics and the relationship between the change of these relations and the problem of the fault state of the gearbox, and the relationship between the change of these relations and the problem of the manufacturing quality of the gearbox. Secondly, A knowledge base can be established for the development and research of the subsequent gearbox abnormal noise diagnosis system. This knowledge base will record the basic knowledge of the vibration characteristics of the transmission typical fault and the manufacturing quality problem of the gearbox parts and components. Secondly, according to the relationship between vibration characteristics, the abnormal noise of gearbox is diagnosed and identified by simulating human cognitive processing process. In this paper, a fractal dimension analysis method of gearbox fault feature is proposed. The fault diagnosis model of gearbox is established by combining support vector machine and rule-based reasoning. In this paper, we first study the reasoning rules with concept driven consciousness, the trigger rules of various patterns, and the diagnostic rules using these rules for diagnosis. To establish a knowledge base which can be used to study the typical quality problem of gearbox. This knowledge base will record the basic knowledge of the vibration characteristics of the gearbox typical fault and the manufacturing quality problem of the gearbox parts. It will provide the basic data support for the further research on the concept driven diagnosis algorithm. On this basis, the mechanism of gearbox fault diagnosis based on concept drive is studied. Secondly, using the principles of information processing, data mining, artificial intelligence and information theory of human perceptual process, the paper studies the sound, noise and noise of gearbox. In this paper, the concept driving method rules in the diagnosis process are extracted, and the information processing and concept driving mechanism in the fault diagnosis process are expounded. An intelligent diagnosis algorithm framework of gearbox quality problem with concept driven consciousness is constructed.
【学位授予单位】：吉林大学
【学位级别】：硕士
【学位授予年份】：2011
【分类号】：TH165.3

【参考文献】