基于小波包熵的轴承故障信号解调方法研究
发布时间：2018-09-04 20:43
【摘要】：在当代中国各类经济行业中,尤其是在制造、运输、能源、冶炼、石油、国防科技等行业中的核心零部件,其工作环境大多具有腐蚀、高温、高压等复杂、恶劣的环境特点,设备中的核心零部件和重要机械结构不可避免的会发生不同程度的故障。轴承由于其零件所在位置的特殊性,往往容易发生疲劳剥落和点蚀等故障,一旦发生故障,则会导致严重的经济财产损失,减少设备的运行寿命。因此本文主要针对滚动轴承故障信号作为研究对象,对故障信号的降噪与解调方法进行研究和讨论,论文的主要工作内容包括:(1)分析了滚动轴承故障信号处理中常用到的几种降噪方法,对比了小波与小波包相关的降噪理论,针对故障信号中往往含有大量背景噪声难以去除的情况,提出了基于小波包熵值与EMD分解相结合的降噪方法,该方法在经过小波包熵值的有效降噪后再进行EMD分解,能够自适应的从故障信号中提取出微弱的故障成分。(2)提出了基于小波包熵值与自相关分析相结合的降噪方法,利用自相关分析能够突出故障信号周期性的特性,将小波包熵值降噪法与之相结合,在小波包熵值降噪去除大量噪声的同时利用自相关分析进一步抑制噪声,在保留原有故障调制信息的基础上突出信号的周期性。(3)对比了各类信号解调方法的优缺点,分析了在相同降噪方法下,能量算子解调法与Hilbert解调法的解调效果,结合以上两种信号降噪分析方法,提出了基于小波包熵值与EMD的能量算子解调法;基于小波包熵值与自相关分析的能量算子解调法,从而准确的判断故障位置。(4)引入EMD与EEMD多分量分析,将小波包熵值降噪,自相关分析,能量算子解调法与之结合起来,提出了一种基于多分量分析与自相关分析相结合的故障信号能量算子解调法,该方法在小波包熵值有效降噪基础上,能够对故障信号进行有效的解调,实现对故障位置的判别。本文的研究工作为旋转机械故障信号的解调分析和诊断提供了一条新的方法思路,对于滚动轴承故障信号的解调处理上具有一定的参考。
[Abstract]:In contemporary China's various economic industries, especially in manufacturing, transportation, energy, smelting, petroleum, national defense science and technology and other industries, the core components, its working environment is mostly corrosion, high temperature, high pressure and other complex, harsh environmental characteristics. The core parts and important mechanical structure of the equipment will inevitably break down in varying degrees. Because of the particularity of the location of the bearing parts, it is easy to occur fatigue spalling and pitting. Once the failure occurs, it will lead to serious economic property losses and reduce the service life of the equipment. Therefore, in this paper, the noise reduction and demodulation methods of rolling bearing fault signals are studied and discussed. The main contents of this paper are as follows: (1) several noise reduction methods used in rolling bearing fault signal processing are analyzed, and the noise reduction theory related to wavelet and wavelet packet is compared. In view of the fact that it is difficult to remove a large amount of background noise in the fault signal, a method based on the combination of wavelet packet entropy and EMD decomposition is proposed, which decomposes the EMD after the effective denoising of the wavelet packet entropy. It can self-adaptively extract weak fault components from fault signals. (2) A noise reduction method based on wavelet packet entropy and autocorrelation analysis is proposed, and the periodicity of fault signals can be highlighted by using autocorrelation analysis. The wavelet packet entropy denoising method is combined with the wavelet packet entropy value to remove a large amount of noise, and the autocorrelation analysis is used to further suppress the noise. On the basis of preserving the original fault modulation information, the periodicity of the signal is highlighted. (3) the advantages and disadvantages of various signal demodulation methods are compared, and the demodulation effects of the energy operator demodulation method and the Hilbert demodulation method under the same noise reduction method are analyzed. The energy operator demodulation method based on wavelet packet entropy value and EMD, the energy operator demodulation method based on wavelet packet entropy value and autocorrelation analysis, and the energy operator demodulation method based on wavelet packet entropy and autocorrelation analysis are proposed. In order to accurately judge the fault location. (4) EMD and EEMD multicomponent analysis are introduced to reduce the noise of wavelet packet entropy, autocorrelation analysis, energy operator demodulation method and the combination of them. A fault signal energy operator demodulation method based on multi-component analysis and autocorrelation analysis is proposed. Based on the effective denoising of wavelet packet entropy, the fault signal can be demodulated effectively and the fault location can be distinguished. The research work in this paper provides a new method for the demodulation analysis and diagnosis of the fault signals of rotating machinery, and has a certain reference for the demodulation and processing of the fault signals of rolling bearings.
【学位授予单位】：内蒙古科技大学
【学位级别】：硕士
【学位授予年份】：2016
【分类号】：TH133.3
本文编号：2223277
[Abstract]:In contemporary China's various economic industries, especially in manufacturing, transportation, energy, smelting, petroleum, national defense science and technology and other industries, the core components, its working environment is mostly corrosion, high temperature, high pressure and other complex, harsh environmental characteristics. The core parts and important mechanical structure of the equipment will inevitably break down in varying degrees. Because of the particularity of the location of the bearing parts, it is easy to occur fatigue spalling and pitting. Once the failure occurs, it will lead to serious economic property losses and reduce the service life of the equipment. Therefore, in this paper, the noise reduction and demodulation methods of rolling bearing fault signals are studied and discussed. The main contents of this paper are as follows: (1) several noise reduction methods used in rolling bearing fault signal processing are analyzed, and the noise reduction theory related to wavelet and wavelet packet is compared. In view of the fact that it is difficult to remove a large amount of background noise in the fault signal, a method based on the combination of wavelet packet entropy and EMD decomposition is proposed, which decomposes the EMD after the effective denoising of the wavelet packet entropy. It can self-adaptively extract weak fault components from fault signals. (2) A noise reduction method based on wavelet packet entropy and autocorrelation analysis is proposed, and the periodicity of fault signals can be highlighted by using autocorrelation analysis. The wavelet packet entropy denoising method is combined with the wavelet packet entropy value to remove a large amount of noise, and the autocorrelation analysis is used to further suppress the noise. On the basis of preserving the original fault modulation information, the periodicity of the signal is highlighted. (3) the advantages and disadvantages of various signal demodulation methods are compared, and the demodulation effects of the energy operator demodulation method and the Hilbert demodulation method under the same noise reduction method are analyzed. The energy operator demodulation method based on wavelet packet entropy value and EMD, the energy operator demodulation method based on wavelet packet entropy value and autocorrelation analysis, and the energy operator demodulation method based on wavelet packet entropy and autocorrelation analysis are proposed. In order to accurately judge the fault location. (4) EMD and EEMD multicomponent analysis are introduced to reduce the noise of wavelet packet entropy, autocorrelation analysis, energy operator demodulation method and the combination of them. A fault signal energy operator demodulation method based on multi-component analysis and autocorrelation analysis is proposed. Based on the effective denoising of wavelet packet entropy, the fault signal can be demodulated effectively and the fault location can be distinguished. The research work in this paper provides a new method for the demodulation analysis and diagnosis of the fault signals of rotating machinery, and has a certain reference for the demodulation and processing of the fault signals of rolling bearings.
【学位授予单位】：内蒙古科技大学
【学位级别】：硕士
【学位授予年份】：2016
【分类号】：TH133.3
【相似文献】
相关期刊论文 前4条
1 狄鹏,黎放,胡剑;熵值双基点多目标决策法在AIP选型中的应用[J];船海工程;2005年03期
2 黄健宇;“熵”论及其它[J];机械工业标准化与质量;1998年10期
3 王首绪;吴岳;;基于IAHP-Entrop-Topsis模型的公路采购决策方法[J];长沙理工大学学报(自然科学版);2013年01期
4 ;[J];;年期
相关会议论文 前2条
1 李国良;李忠富;;基于聚类的企业绩效熵值评价方法研究[A];第十一届中国管理科学学术年会论文集[C];2009年
2 Bakhdavlatov saidsho;毛羽忻;龚萍;周沛然;毛征;;基于局部熵值图的目标检测分割及质心计算[A];系统仿真技术及其应用学术论文集（第15卷）[C];2014年
相关重要报纸文章 前2条
1 董旭光 韩晓伟;信息系统中的“熵”[N];解放军报;2000年
2 李亚民;为企业思想注入特质 实现企业管理的“熵减”[N];中国航空报;2008年
相关硕士学位论文 前1条
1 王戈;基于小波包熵的轴承故障信号解调方法研究[D];内蒙古科技大学;2016年
，本文编号：2223277
本文链接：https://www.wllwen.com/jixiegongchenglunwen/2223277.html
