基于神经网络的滚动轴承故障诊断技术.docx-资料库

luck410-7405283-4744302543385892859.docx.pdf-第1页.png

第1页 / 共73页

luck410-7405283-4744302543385892859.docx.pdf-第2页.png

第2页 / 共73页

luck410-7405283-4744302543385892859.docx.pdf-第3页.png

第3页 / 共73页

luck410-7405283-4744302543385892859.docx.pdf-第4页.png

第4页 / 共73页

luck410-7405283-4744302543385892859.docx.pdf-第5页.png

第5页 / 共73页

luck410-7405283-4744302543385892859.docx.pdf-第6页.png

第6页 / 共73页

luck410-7405283-4744302543385892859.docx.pdf-第7页.png

第7页 / 共73页

luck410-7405283-4744302543385892859.docx.pdf-第8页.png

第8页 / 共73页

摘要摘要滚动轴承是机械中应用最广泛的一种通用机械部件，轴承特定的使用环境造成其寿命的随机性较大，目前还无法准确预测其寿命。因此，轴承故障诊断就显得非常重要。本文对滚动轴承的振动数据分别在时域和频域进行了分析，由神经网络处理，对结果采用“判决区间+举手表决”方式，获得最终判断结果。最后，本文对实测的轴承振动数据进行结论验证，结果表明，本文提出的以神经网络为基础的轴承故障诊断方法具有准确、实用的特点，有广泛的应用前景。关键词：滚动轴承，故障诊断，神经网络，判决区间，举手表决 I

Abstract Abstract Rolling bearing is one of the most widely used in mechanical general machinery parts, bearing the use of specific environment caused by the randomness of life, it is unable to accurately predict its life.Therefore, bearing fault diagnosis is very important.In this paper, the vibration of rolling bearings data in time domain and frequency domain are analyzed respectively, by the neural network processing, the results by adopting the working style of "ruling range + show of hands" in the end to judge the results. Finally, this paper verifies the conclusion of the bearing vibration data, the results show that the proposed neural network based bearing fault diagnosis method has the characteristics of accurate, practical, have broad application prospects. Key words: Rolling bearing fault diagnosis, neural network, decision interval, a vote II

目录目录目录.................................................................................................................................5 第 1 章引言................................................................................................................. 1 1.1 滚动轴承故障诊断技术的研究现状与发展态势.............................................1 1.2 滚动轴承故障诊断技术的基本内容.................................................................1 1.3 滚动轴承故障诊断技术的意义.........................................................................2 1.4 神经网络理论在故障诊断技术中的应用.........................................................3 第 2 章滚动轴承振动机理与故障诊断..................................................................... 4 2.1 滚动轴承的结构.................................................................................................4 2.2 滚动轴承失效的基本形式.................................................................................4 2.3 滚动轴承故障的检测方法.................................................................................6 2.4 滚动轴承的振动机理和故障特征频率.............................................................7 2.4.1 滚动轴承的振动机理..................................................................................7 2.4.2 滚动轴承各元件单一缺陷的频率..............................................................7 2.5 本章小节.............................................................................................................8 第 3 章神经网络理论................................................................................................. 9 3.1 人工神经网络概述.............................................................................................9 3.1.1 人工神经网络的基本组成..........................................................................9 3.1.2 神经网络的诊断方法................................................................................10 3.1.3 神经网络在故障诊断中的应用................................................................11 3.2 BP 神经网络.....................................................................................................12 3.2.1 BP 神经网络学习规则..............................................................................12 3.2.2 BP 算法......................................................................................................13 3.3 BP 网络的不足及改进.....................................................................................14 3.3.1 BP 网络的局限性......................................................................................14 3.3.2 BP 网络的改进..........................................................................................15 3.4 本章小结...........................................................................................................16 第 4 章数据处理....................................................................................................... 17 4.1 轴承振动数据的获取.......................................................................................17

电子科技大学学士学位论文 4.1.1 实验系统....................................................................................................17 4.1.2 轴承故障诊断实验....................................................................................18 4.2 原始信号数据预处理.......................................................................................19 4.2.1 各态历经性................................................................................................19 4.2.2 时域分析....................................................................................................19 4.2.3 频域分析....................................................................................................22 4.3 神经网络预结果再处理...................................................................................23 4.4 本章小结...........................................................................................................24 第 5 章轴承故障诊断模型的构建与仿真............................................................... 25 5.1 轴承故障诊断神经网络建模...........................................................................25 5.1.1 输入层节点数............................................................................................25 5.1.2 输出层节点数............................................................................................25 5.1.3 隐层数和隐层节点数................................................................................26 5.1.4 学习方式的确定........................................................................................27 5.1.5 神经元传递函数的选取............................................................................27 5.1.6 学习率的确定............................................................................................27 5.1.7 初始权重的设置........................................................................................28 5.2 时域仿真結果及分析.......................................................................................28 5.2.1 普通 BP 网络时域仿真.............................................................................29 5.2.2 自适应修改学历率算法时域仿真............................................................30 5.2.3 有弹回的 BP 算法时域仿真.....................................................................32 5.2.4 新样本时域仿真........................................................................................34 5.3 频域仿真结果及分析.......................................................................................36 5.3.1 自适应修改学习率算法频域仿真............................................................36 5.3.2 有弹回的 BP 算法频域仿真.....................................................................38 5.3.3 新样本频域仿真........................................................................................39 5.4 本章小结...........................................................................................................41 第 6 章结束语...............................................................................................................42 参考文献.........................................................................................................................43 致谢.................................................................................................................................44 翻译文章.........................................................................................................................45 原文.............................................................................................................................45 6

译文.............................................................................................................................57 目录

资料库

基于神经网络的滚动轴承故障诊断技术.docx

相关推荐

操作系统

热门标签

最新资料