基于盲源分离的滚动轴承复合故障诊断方法研究
发布时间：2018-08-14 16:07
【摘要】：滚动轴承作为一种承受和传递载荷的重要部件,在机械设备中得到了广泛应用,其运行状态会直接影响设备整体性能,滚动轴承部件一旦发生故障,可能导致设备损毁,甚至造成灾难性事故。因此,对滚动轴承进行故障诊断具有非常重要的意义。在实际生产环境中,滚动轴承某一部位出现故障时常伴随着其它部位的故障,即出现滚动轴承的复合故障。这种情况下多个故障的振动源信号及噪声相互耦合,振动信号呈现复杂化,故障类型的诊断变得尤为困难。而盲源分离是解决复合故障源信号分离问题的有效方法之一。因此,论文以滚动轴承复合故障振动信号为研究对象,结合盲源分离理论以及时频分析和模式识别方法,对滚动轴承复合故障诊断过程中遇到的常规和极端条件下的故障源分离、特征提取、故障类型诊断等问题展开研究。论文的主要研究工作可概述如下:对盲源分离方法的基本理论进行研究,并通过仿真实验对几种经典盲源分离算法的分离效果进行分析比较,在此基础上,将JADE盲源分离算法用于分离滚动轴承复合故障信号。滚动轴承故障信号是非平稳和非线性的,而传统的时域和频域信号分析方法难以兼顾非平稳信号的时变特性,不能准确体现滚动轴承各个故障类型的故障特性。针对这一问题,提出一种基于EMD时频分析的样本熵和能量比特征提取方法,该方法在时频分析基础上进行信号特征提取,能够更全面、更准确地揭示滚动轴承振动信号中的故障特征信息。此外,结合盲源信号分离、特征提取以及支持向量机构建了一种有效的滚动轴承复合故障诊断机制。传统的盲源分离方法大多基于观测信号数不少于源信号数的假设,不能适应于单通道条件下的复合故障信号分离。因此,将变分模态分解方法引入到盲源分离领域,通过变分模态分解将单通道盲源分离的极端欠定问题转化为适定或超定问题,为单通道条件下的复合故障诊断实现提供一种有效的解决方案。实验结果表明,提出的基于VMD的单通道盲源分离方法相对于传统方法更具优越性。
[Abstract]:As an important component to bear and transfer load, rolling bearing has been widely used in mechanical equipment. Its running state will directly affect the overall performance of the equipment. Once the rolling bearing component fails, it may lead to equipment damage. And even catastrophic accidents. Therefore, the fault diagnosis of rolling bearings is of great significance. In the actual production environment, the fault of one part of the rolling bearing is often accompanied by the fault of other parts, that is, the compound fault of the rolling bearing. In this case, the signal of vibration source and noise of multiple faults are coupled with each other, the vibration signal is complicated, and the diagnosis of fault type becomes more and more difficult. Blind source separation is one of the effective methods to solve the problem of complex fault source signal separation. Therefore, the paper takes the complex fault vibration signal of rolling bearing as the research object, and combines the blind source separation theory with the method of timely frequency analysis and pattern recognition. The problems of fault source separation, feature extraction, fault type diagnosis and so on in the process of rolling bearing composite fault diagnosis are studied. The main research work of this paper can be summarized as follows: the basic theory of blind source separation method is studied, and the separation effect of several classical blind source separation algorithms is analyzed and compared by simulation experiments. JADE blind source separation algorithm is used to separate composite fault signals of rolling bearings. The fault signals of rolling bearings are non-stationary and nonlinear, but the traditional time-domain and frequency-domain signal analysis methods are difficult to take into account the time-varying characteristics of non-stationary signals and can not accurately reflect the fault characteristics of each fault type of rolling bearings. In order to solve this problem, a method of extracting feature of sample entropy and energy ratio based on EMD time-frequency analysis is proposed. The fault characteristic information in the vibration signal of rolling bearing is revealed more accurately. In addition, combining blind source signal separation, feature extraction and support vector mechanism, an effective fault diagnosis mechanism for rolling bearing is proposed. Most of the traditional blind source separation methods are based on the assumption that the number of observed signals is not less than the number of source signals. Therefore, the variational mode decomposition method is introduced into the field of blind source separation, and the extreme underdetermination problem of single channel blind source separation is transformed into a suitable or overdetermined problem by variational mode decomposition. It provides an effective solution for the realization of complex fault diagnosis under the condition of single channel. Experimental results show that the proposed single channel blind source separation method based on VMD is superior to the traditional method.
【学位授予单位】：重庆大学
【学位级别】：硕士
【学位授予年份】：2016
【分类号】：TH133.33
[Abstract]:As an important component to bear and transfer load, rolling bearing has been widely used in mechanical equipment. Its running state will directly affect the overall performance of the equipment. Once the rolling bearing component fails, it may lead to equipment damage. And even catastrophic accidents. Therefore, the fault diagnosis of rolling bearings is of great significance. In the actual production environment, the fault of one part of the rolling bearing is often accompanied by the fault of other parts, that is, the compound fault of the rolling bearing. In this case, the signal of vibration source and noise of multiple faults are coupled with each other, the vibration signal is complicated, and the diagnosis of fault type becomes more and more difficult. Blind source separation is one of the effective methods to solve the problem of complex fault source signal separation. Therefore, the paper takes the complex fault vibration signal of rolling bearing as the research object, and combines the blind source separation theory with the method of timely frequency analysis and pattern recognition. The problems of fault source separation, feature extraction, fault type diagnosis and so on in the process of rolling bearing composite fault diagnosis are studied. The main research work of this paper can be summarized as follows: the basic theory of blind source separation method is studied, and the separation effect of several classical blind source separation algorithms is analyzed and compared by simulation experiments. JADE blind source separation algorithm is used to separate composite fault signals of rolling bearings. The fault signals of rolling bearings are non-stationary and nonlinear, but the traditional time-domain and frequency-domain signal analysis methods are difficult to take into account the time-varying characteristics of non-stationary signals and can not accurately reflect the fault characteristics of each fault type of rolling bearings. In order to solve this problem, a method of extracting feature of sample entropy and energy ratio based on EMD time-frequency analysis is proposed. The fault characteristic information in the vibration signal of rolling bearing is revealed more accurately. In addition, combining blind source signal separation, feature extraction and support vector mechanism, an effective fault diagnosis mechanism for rolling bearing is proposed. Most of the traditional blind source separation methods are based on the assumption that the number of observed signals is not less than the number of source signals. Therefore, the variational mode decomposition method is introduced into the field of blind source separation, and the extreme underdetermination problem of single channel blind source separation is transformed into a suitable or overdetermined problem by variational mode decomposition. It provides an effective solution for the realization of complex fault diagnosis under the condition of single channel. Experimental results show that the proposed single channel blind source separation method based on VMD is superior to the traditional method.
【学位授予单位】：重庆大学
【学位级别】：硕士
【学位授予年份】：2016
【分类号】：TH133.33
【相似文献】
相关期刊论文 前10条
1 林秋华,殷福亮;盲源分离自适应算法的统一形式[J];大连理工大学学报;2002年04期
2 李广彪;张剑云;;基于变步长等变化自适应盲源分离算法[J];电子信息对抗技术;2006年01期
3 苏中元;贾民平;;周期平稳信号盲源分离算法及其应用[J];机械工程学报;2007年10期
4 陈锡明;黄硕翼;;盲源分离综述——问题、原理和方法[J];电子信息对抗技术;2008年02期
5 刘秀芳;艾延廷;张[，
本文编号：2183395
本文链接：https://www.wllwen.com/jixiegongchenglunwen/2183395.html
