基于改进HHT与SVM的滚动轴承故障诊断方法研究

发布时间：2018-12-10 19:59

【摘要】：随着科学技术的进步和现代工业的发展,滚动轴承作为机械设备中的重要旋转零件,在工业生产领域应用广泛,也是主要故障源之一。许多机械设备的故障都与滚动轴承有关,轴承的工作状态正常与否直接影响到机械设备的性能。因此开展滚动轴承故障诊断技术的研究具有重要的理论价值和现实意义。本文从轴承振动信号的处理着手,对滚动轴承故障诊断技术所涉及的理论方法和关键技术进行了深入的研究,主要内容如下：阐述了滚动轴承故障诊断技术的内容和意义,系统介绍了滚动轴承的振动机理及其故障特征频率,总结了滚动轴承故障诊断中特征提取和模式识别的研究现状。介绍了希尔伯特-黄变换的基本原理和Hilbert谱和Hilbert边际谱的概念,针对现有经验模态分解方法模态混叠的缺点,引入完备的集合经验模态分解方法,通过对周期性冲击信号进行经验模态分解和完备的集合经验模态分解对比分析,验证了完备的集合经验模态分解得到的内禀模态函数更接近理想值。针对滚动轴承故障特征易被噪声干扰淹没的特点,提出了一种基于小波包和完备的集合经验模态分解相结合,并且附带内禀模态函数筛选过程的特征提取方法。将该方法用于仿真信号和实验数据的分析,结果表明,改进的HHT方法能有效提取故障特征频率成分并抑制噪声。采用最小二乘支持向量机模型对滚动轴承的故障进行模式识别,并用粒子群优化算法对参数进行寻优。将提取到的内禀模态函数的奇异值作为支持向量机的特征向量进行训练,对测试样本进行识别验证了支持向量机的分类效果。
[Abstract]:With the progress of science and technology and the development of modern industry, rolling bearing, as an important rotating part in machinery and equipment, is widely used in the field of industrial production, and is also one of the main fault sources. Many mechanical equipment faults are related to rolling bearings, the normal working condition of bearings directly affects the performance of mechanical equipment. Therefore, the research of rolling bearing fault diagnosis technology has important theoretical value and practical significance. In this paper, the theoretical methods and key technologies of rolling bearing fault diagnosis technology are deeply studied from the view of bearing vibration signal processing. The main contents are as follows: the content and significance of rolling bearing fault diagnosis technology are expounded. The vibration mechanism and fault characteristic frequency of rolling bearing are introduced systematically. The research status of feature extraction and pattern recognition in rolling bearing fault diagnosis is summarized. In this paper, the basic principle of Hilbert-Huang transform and the concepts of Hilbert spectrum and Hilbert marginal spectrum are introduced. Aiming at the shortcomings of the existing empirical mode decomposition methods, a complete set empirical mode decomposition method is introduced. By comparing the empirical mode decomposition and complete set empirical mode decomposition of periodic shock signal, it is verified that the intrinsic mode function obtained by complete set empirical mode decomposition is closer to the ideal value. In view of the fact that the fault features of rolling bearings are easily submerged by noise, a feature extraction method based on wavelet packet and complete set empirical mode decomposition (EMD) with intrinsic modal function selection process is proposed. The method is applied to the analysis of simulation signals and experimental data. The results show that the improved HHT method can effectively extract the fault characteristic frequency components and suppress the noise. The model of least square support vector machine (LS-SVM) is used to identify the fault of rolling bearing and the particle swarm optimization (PSO) algorithm is used to optimize the parameters. The singular value of the intrinsic modal function is trained as the eigenvector of the support vector machine, and the classification effect of the support vector machine is verified by the recognition of the test samples.
【学位授予单位】：北京交通大学
【学位级别】：硕士
【学位授予年份】：2016
【分类号】：TH133.33

【相似文献】