滚动轴承振动信号降噪方法研究

发布时间：2018-01-17 03:07

本文关键词：滚动轴承振动信号降噪方法研究　出处：《华北电力大学(北京)》2016年硕士论文　论文类型：学位论文

【摘要】：大型旋转机械设备中的滚动轴承是旋转机械中关键且容易损坏的部件,旦出现故障,将会导致机械设备运行失稳等一系列严重后果,而故障部位提取的振动信号往往表现出波动特征。滚动轴承振动信号包含了大量的轴承运行状态信息,是对轴承开展故障分析的主要依据之一。因此,论文针对大型旋转机械的滚动轴承振动的仿真信号和实测信号分别开展了信号降噪方法研究,将对进一步处理振动监测信号、预防安全事故发生都具有重要的理论指导意义和工程实际应用价值。论文研究的主要目的是利用大型旋转机械设备滚动轴承振动信号进行降噪处理,提取出所需要的振动信号。针对振动信号非平稳的特点,信号降噪研究了基于改进的小波半软阂值降噪方法。特征提取是实现机械故障诊断的关键步骤,对与滚动轴承振动故障诊断至为关键。论文首先介绍了小波分析的基本原理,给出了小波变换的基本定义和基本性质,介绍了常见小波变换的基本算法。然后介绍了传统的基于小波变换的硬阈值降噪方法和基于小波变换的软阈值降噪方法,在分析它们的优点和缺点基、础上,针对它们存在的缺点和不足,提出了改进的基于小波变换的半软阈值降噪方法。最后给出了评价各降噪方法的性能评价准则。针对滚动轴承振动降噪问题采用改进的基于小波变换的半软阈值降噪方法分别进行了仿真信号测试和实际信号测试分析,最后通过性能评价指标、频谱分析对比了这三种方法的性能优劣,结果表明所提出的基于小波变换的半软阈值降噪方法比传统的基于小波变换的硬阈值降噪方法和基于小波变换的软阈值降噪方法能达到更好的降噪效果。最后研究了滚动轴承振动信号降噪问题的故障特征提取问题,介绍了基于Winger分布和奇异值分解相结合的特征提取方法,得到Winger时频谱后,接着对Winger进行奇异值分解,在四种不同工况下分别得到特征向量,可以便于进一步开展针对滚动轴承振动信号降噪处理。
[Abstract]:Rolling bearings of large rotating machinery is the key equipment in the easy damage of rotating machinery parts, once the failure will lead to mechanical equipment running instability and a series of serious consequences, and the fault vibration signal extraction often shows the fluctuation characteristics of rolling bearing vibration signal contains the state information bearing running large, is one of the main basis to carry out fault analysis of bearing. Therefore, the rolling bearing vibration of large rotating machinery simulation signal and measured signal are respectively carried out the research of signal denoising method, the further processing of vibration signals to prevent safety accidents has important theoretical significance and practical value in engineering application. The main purpose of this thesis is the noise reduction of vibration signal of rolling bearing using large rotating machinery, vibration signal to extract needed. Needle The vibration signal characteristic of non-stationary signal denoising, the improved wavelet threshold denoising method based on semi soft. Feature extraction is the key step to realize mechanical fault diagnosis, and the rolling bearing fault diagnosis is the key. This paper firstly introduces the basic principle of wavelet analysis, this paper gives the definition of wavelet transform and basic introduces the basic properties of the common algorithm of wavelet transform is introduced. Then based on the traditional hard threshold denoising method based on wavelet transform and soft threshold denoising method based on wavelet transform, the analysis of advantages and disadvantages, on the basis of them, according to their disadvantages and deficiencies, proposed an improved semi soft threshold denoising method based on wavelet transform based on the transformation of the performance evaluation criterion. Finally, the noise reduction method is presented. Aiming at the problem of rolling bearing vibration and noise reduction based on wavelet transform semi soft threshold value by the improved Noise reduction methods are analyzed respectively the simulation test and practical test signal signal, finally the performance evaluation index, spectrum analysis compared the performance of the three methods, the results show that the proposed semi soft threshold denoising method based on wavelet transform than the traditional hard threshold denoising method based on wavelet transform and soft threshold denoising method based on wavelet transform based on can achieve better noise reduction effect. Finally the fault feature of vibration signal of rolling bearing extraction, based on the Winger distribution and the singular value decomposition combined feature extraction method, Winger spectrum, and singular value decomposition on Winger, under four different operating conditions are characteristic vector, can facilitate the further development of the rolling bearing vibration signal denoising.

【学位授予单位】：华北电力大学(北京)
【学位级别】：硕士
【学位授予年份】：2016
【分类号】：TH133.33

【参考文献】