局部均值分解方法及其在滚动轴承故障诊断中的应用研究

发布时间：2018-08-23 10:11

【摘要】：旋转机械是现代工业生产设备中极其重要的组成部分,是诸多机械设备中的一个重要燕类。在各类旋转机械中,滚动轴承扮演着十分关键的角色。为了保障机械设备的正常运行,针对旋转机械滚动轴承的状态监测和故障诊断已经成为重要的研究课题。局部均值分解(Local Mean Decomposition,LMD)是近年来出现的一种新的自适应的信号分析方法,由于其在处理非平稳非线性信号上的优势,而被广泛应用于旋转机械故障诊断领域。本文重点研究了局部均值分解方法及其在滚动轴承故障诊断中的应用,主要内容如下:首先,针对LMD分解过程中存在的模态混淆问题,研究了一种基于掩蔽信号和相关系数的掩蔽局部均值分解(Masking Local Mean Decomposition,MLMD),并通过仿真信号和实际的滚动轴承振动信号对该方法的可行性进行验证。其次,针对多尺度排列熵(Multi-Scale Permutation Entropy,MPE)粗粒化过程的局限,研究了复合多尺度排列熵(Composite Multi-Scale Permutation Entropy,CMPE),并与LMD、综合关联分析相结合,研究了一种基于LMD-CMPE和综合关联分析的滚动轴承故障诊断方法。该方法首先对故障信号进行LMD分解,然后选取包含故障信息较为丰富的前几个乘积函数(Product Function,PF)计算CMPE作为特征向量,在灰色相似关联度的基础上融入相近性因子,综合考虑相似性和相近性对CMPE特征向量进行模式识别,实现不同类型故障的判别。最后,从瞬时特征分析的角度出发,依据LMD算法的自身特点,定义了基于LMD的瞬时能量谱,并在此基础上结合信息熵的概念,提出了瞬时能量谱熵,并与数学形态学去噪相结合,应用到滚动轴承振动信号的特征提取当中。基于LMD的瞬时能量谱可以准确地反映故障特征频率,瞬时能量谱熵对不同运行状态的滚动轴承振动信号有稳定的区分度,将其作为故障特征参数可以有效地区分不同类型的故障。
[Abstract]:Rotating machinery is an extremely important part of modern industrial production equipment and an important swallow in many machinery and equipment. In all kinds of rotating machinery, rolling bearing plays a very important role. In order to ensure the normal operation of mechanical equipment, the condition monitoring and fault diagnosis of rolling bearings for rotating machinery has become an important research topic. Local mean decomposition (Local Mean) is a new adaptive signal analysis method in recent years. It is widely used in the field of rotating machinery fault diagnosis because of its advantages in dealing with non-stationary nonlinear signals. This paper focuses on the method of local mean decomposition and its application in rolling bearing fault diagnosis. The main contents are as follows: firstly, the modal confusion in the process of LMD decomposition is discussed. A masking local mean decomposition (Masking Local Mean DecompositionMLMD) based on masking signal and correlation coefficient is studied, and the feasibility of this method is verified by simulation signal and actual rolling bearing vibration signal. Secondly, considering the limitation of Multi-Scale Permutation Entropy Entropy (MPE) coarse-grained process, the composite Composite Multi-Scale Permutation Entropic permutation Entropy (MPE) is studied, which is combined with LMD and integrated correlation analysis. A method of rolling bearing fault diagnosis based on LMD-CMPE and comprehensive correlation analysis is studied. The method firstly decomposes the fault signal by LMD, and then selects the first several product functions (Product function / PF) which contain abundant fault information to calculate the CMPE as the eigenvector. The similarity factor is incorporated on the basis of the grey similarity correlation degree. Considering the similarity and similarity, the pattern recognition of CMPE feature vector is carried out to distinguish different types of faults. Finally, from the point of view of instantaneous feature analysis, according to the characteristics of LMD algorithm, the instantaneous energy spectrum based on LMD is defined, and based on the concept of information entropy, the instantaneous energy spectrum entropy is proposed. Combined with mathematical morphology denoising, it is applied to feature extraction of rolling bearing vibration signal. The instantaneous energy spectrum based on LMD can accurately reflect the characteristic frequency of the fault. The instantaneous energy spectrum entropy has a stable degree of discrimination for the vibration signals of rolling bearings in different running states. Using it as a fault characteristic parameter can effectively distinguish different types of faults.
【学位授予单位】：燕山大学
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：TH133.33

【相似文献】