基于数学形态学与模糊C均值的滚动轴承故障诊断方法

发布时间：2018-07-05 00:13

本文选题：滚动轴承 + 故障诊断　；参考：《燕山大学》2012年硕士论文

【摘要】：随着现代化工业生产的不断发展，机械设备故障诊断技术近年来得到了广泛的重视，滚动轴承作为机械传动系统中的重要元件，，其运行的好坏直接影响机器的工作状况。针对滚动轴承振动信号噪声，建立了一种数学形态学组合滤波器，通过组合形态滤波器对其振动信号进行降噪处理；针对振动信号的非平稳性、非线性等特征，提出一种多尺度形态学分析方法对故障信号进行定性和定量分析，同时针对故障模式的模糊性问题，提出采用模糊C均值(Fuzzy Center Means，简称FCM)聚类算法的模糊故障识别方法，并将上述研究方法结合起来运用到滚动轴承的故障诊断中。首先，阐述了滚动轴承的故障主要形式和振动机理，给出振动信号常用降噪方法，如传统的滤波方法、小波变化消噪技术和经验模态分解(Empirical modedecomposition，简称EMD)降噪技术；同时阐述了传统的振动信号的分析方法，包括时域分析，频域分析等。其次，根据形态组合滤波器中结构元素目前尚无一确定的选取准则问题，分析了形态组合滤波器中结构元素的形状、宽度和幅度对形态滤波效果的影响。然后，分析了多尺度形态学在振动信号中的应用，通过分形维数和形态谱熵对故障信号进行特征描述，将其作为描述故障的特征参数引入到模糊C均值聚类算法中作为聚类分析的特征向量，为机械故障识别作准备。最后，针对来自美国凯斯西储大学的滚动轴承故障数据及宝钢1580SP轧机实测数据进行实验研究及分析，并给出结论。形态学滤波方法可以对滚动轴承振动信号达到很好的降噪效果；多尺度形态学方法可以对滚动轴承故障进行定性和定量描述，模糊C均值聚类可以取得良好的识别效果。
[Abstract]:With the development of modern industrial production, the fault diagnosis technology of mechanical equipment has been paid more and more attention in recent years. As an important component of mechanical transmission system, the running quality of rolling bearing directly affects the working condition of machinery. Aiming at the noise of rolling bearing vibration signal, a mathematical morphological combined filter is established, which is used to reduce the noise of the vibration signal, aiming at the non-stationary and nonlinear characteristics of the vibration signal. This paper presents a multi-scale morphological analysis method for qualitative and quantitative analysis of fault signals. Aiming at the fuzziness of fault mode, a fuzzy fault identification method using fuzzy C-means (FCM) clustering algorithm is proposed. The above research method is applied to the fault diagnosis of rolling bearing. Firstly, the main fault forms and vibration mechanism of rolling bearing are expounded, and the usual noise reduction methods of vibration signal are given, such as traditional filtering method, wavelet change noise elimination technique and empirical mode decomposition (EMD) noise reduction technology. At the same time, the traditional methods of vibration signal analysis, including time domain analysis, frequency domain analysis, etc. Secondly, the influence of shape, width and amplitude of structural elements on morphological filtering effect is analyzed according to the fact that there is no definite criterion for the selection of structural elements in morphological combinatorial filters. Then, the application of multi-scale morphology in vibration signal is analyzed. The fault signal is characterized by fractal dimension and morphological spectrum entropy. It is introduced into the fuzzy C-means clustering algorithm as the characteristic parameter to describe the fault, and it is used as the feature vector of the clustering analysis to prepare for the mechanical fault identification. Finally, the experimental research and analysis of rolling bearing fault data from case Western Reserve University and the measured data of Baosteel 1580SP mill are carried out, and the conclusions are given. Morphological filtering method can achieve good noise reduction effect on rolling bearing vibration signal, multi-scale morphological method can describe the rolling bearing fault qualitatively and quantitatively, and fuzzy C-means clustering can obtain good recognition effect.
【学位授予单位】：燕山大学
【学位级别】：硕士
【学位授予年份】：2012
【分类号】：TH165.3;TH133.33

【参考文献】