基于经验模态分解的单通道轴承故障诊断方法应用研究毕业论文
2022-03-02 21:56:25
论文总字数:19271字
摘 要
在工业生产中,轴承是易损器件,据统计,机械的故障有30%是由轴承引发的,轴承的运行状况关系到整台设备的性能,其缺陷会导致装置产生异常振动及噪声,所以对轴承进行故障诊断研究是有一定意义的。
单通道轴承故障诊断的整个流程分别是轴承振动信号的采集、信号的降噪处理、特征值提取及模式识别几个部分构成。其中,信号的降噪处理是轴承故障诊断的关键部分。因为不能对信号进行有效的降噪处理就会很难提取出有用信息,进而更难进行后续的处理,由于经验模态分解(Empirical Mode Decomposition, EMD)具有自适应性、局部性等优点,特别适合处理非线性非平稳信号,所以把EMD方法引入到本课题来应用,给轴承故障信号进行处理分析。通过对西储大学的轴承振动数据进行实验之后,可以得出基于EMD方法对轴承故障诊断是有效的。对于特征值提取这部分,由于轴承故障特征频率在振动频谱上很难检测到与轴承损伤的特征频率,所以采用了特征值能量熵的提取方法来实现,实验证实了能量熵这个特征值能够比较好的体现轴承运行状况。由于在小样本问题上支持向量机(Support Vector Machine,SVM)能够得到很好的应用,所以在模式识别方面,本文运用SVM来给轴承进行故障分类,西储大学轴承振动信号通过以上几个过程之后,再经SVM分类发现此方法是可行的。在进行SVM分类实验时,着重分析归一化和核函数问题,实验结果表明,未经归一化处理的准确率远远低于经归一化后的,核函数的选择是根据不同的数据来决定的。
最后,基于MATLAB开发设计轴承故障诊断系统,并融入EMD算法和能量熵提取算法及SVM分类算法,仿真实现了轴承振动信号的采集、EMD分解、能量熵提取及SVM分类。
关键词: 故障诊断 EMD EEMD 能量熵 SVM
Application of Fault Diagnosis Method for Single Channel Bearing Based on Empirical Mode Decomposition
Abstract
In industrial production, the bearing is a wearable device, according to statistics, mechanical failure is caused by 30% of the bearing, the bearing operating conditions related to the performance of the entire device, the defect will lead to abnormal vibration and noise, It is of some significance to study the fault diagnosis of bearing.
Single-channel bearing fault diagnosis of the entire process are bearing vibration signal acquisition, signal noise reduction, eigenvalue extraction and pattern recognition of several parts. Among them, the signal noise reduction is a key part of bearing fault diagnosis. It is difficult to extract useful information and make it more difficult to carry out subsequent processing because of the fact that the empirical mode decomposition (EMD) has the advantages of adaptability, locality and so on. Suitable for dealing with non-linear non-stationary signal, so the EMD method into the subject to the application, to the bearing fault signal processing analysis. By experimenting with the bearing vibration data of Xidu University, it can be concluded that EMD method is effective for bearing fault diagnosis. For the eigenvalue extraction of this part, because the characteristic frequency of the bearing fault is difficult to detect the characteristic frequency of the bearing damage in the vibration spectrum, the eigenvalue energy entropy extraction method is used to realize the energy entropy. Good performance of the bearing. (SVM) can be used to achieve good application, so in the pattern recognition, this paper uses SVM to fault classification of bearings, Xichu University bearing vibration signal through the above process After this, by SVM classification found that this method is feasible. In the SVM classification experiment, the problem of normalization and kernel function is analyzed. The experimental results show that the accuracy of the normalized processing is much lower than that of the normalized kernel. The choice of the kernel function is based on different data To decide.
Finally,based on MATLAB development and design of bearing fault diagnosis system, and into the EMD algorithm and energy entropy extraction algorithm and SVM classification algorithm, simulation of bearing vibration signal acquisition, EMD decomposition, energy entropy extraction and SVM classification.
Key words: fault diagnosis; EMD; EEMD; energy entropy; SVM
目录
摘要 I
Abstract II
第一章 绪论 1
1.1 故障诊断的背景及意义 1
1.2 故障诊断技术概况 1
1.3 EMD的研究现状 2
1.4 本课题的主要安排 3
第二章 单通道轴承故障诊断技术概述 5
2.1 轴承的主要损伤形式及其成因 5
2.2 轴承振动类型及其振动机理 5
2.3 轴承故障诊断流程 5
2.4 轴承故障诊断主要目的 6
2.5 本章小结 6
第三章 基于EMD的单通道轴承故障信号处理 7
3.1 EMD的原理和作用 7
3.2 IMF分析 7
3.3 EMD方法的优越性 8
3.3.1 局部性 8
3.3.2 完备性 8
3.3.3 正交性 8
3.3.4 自适应性 9
3.4 EMD执行过程 9
3.5 单通道轴承四种状态振动信号经EMD实验仿真实验 11
3.6 本章小结 15
第四章 基于能量熵的单通道轴承振动信号特征提取 16
4.1 轴承的固有频率 16
4.2 轴承故障频率的理论值计算 16
4.3 能量熵 17
4.3.1 能量熵定义 17
4.3.2 能量熵性质 18
4.4 单通道轴承四种状态提取出的能量熵 18
4.5 本章小结 21
第五章 基于SVM的单通道轴承故障诊断 22
5.1 SVM算法实现 22
5.2 SVM一般特性总结如下 22
5.3 多分类SVM简述 23
5.4 SVM中归一化问题处理分析 23
5.5 SVM中不同核函数分析 25
5.6 单通道轴承故障诊断 27
5.7 本章小结 28
第六章 单通道轴承故障诊断界面设计 29
6.1 基于MATLAB的轴承故障诊断界面 29
6.2 本章小结 35
第七章 结论与展望 36
7.1 论文结论 36
7.2 研究展望 36
参考文献 37
致谢 40
第一章 绪论
1.1 故障诊断的背景及意义
如今,当机械装置出现故障时,会导致很大经济损失的,还可能对人们生命安全构成威胁。随着科学技术的飞速发展和社会生产水平的不断提高,多个机械装置联合工作变为常态,机械设备中的每一个元件在内部不仅相互联系,同时也影响其它装置。机械装置在结构上变得越来越复杂,其内部的各个元器件彼此之间的耦合作用也变得越来越强,对装置的运行影响因素也随之变多。机械装置出现故障频率在不断增加,所以我们要对轴承故障进行预示和诊断。
请支付后下载全文,论文总字数:19271字