基于统计分布模型的滚动轴承故障特征提取方法研究

发布时间：2018-04-29 22:25

本文选题：特征提取 + 威布尔分布　；参考：《湖南工业大学》2011年硕士论文

【摘要】：滚动轴承是旋转机械中使用最为广泛和最易受损的零部件之一,其工作状态直接影响到整个机械系统的性能,对其进行故障诊断具有重要的实际意义。基于振动信号分析的滚动轴承特征提取方法是国内外使用最多、也是最有效的方法之一。统计分布模型参数在可靠性工程中已被广泛应用于反映机械产品的疲劳寿命和疲劳强度,但在机械特别是轴承状态监测和故障诊断中用于特征提取的研究尚未多见。为了充分挖掘滚动轴承运行中蕴含的有效状态变化信息,提出了一种基于威布尔分布模型参数及其数字特征的故障特征提取新方法。在对滚动轴承原始振动信号建立Weibull分布模型的基础上,分别提取模型的尺度参数以及中位数作为表征轴承运行状态的一种新特征向量。仿真试验结果证明了该特征提取方法的有效性。针对滚动轴承振动信号的非高斯特性,提出了一种基于对数正态分布模型的故障特征提取新思路,提取其模型参数的对数均值作为表征滚动轴承运行状态的新特征量。有效地解决了振动信号的非高斯问题。针对上述方法无法全面准确描述滚动轴承振动信号的非平稳特性问题,提出了一种基于小波域对数正态模型的滚动轴承故障特征提取新方法。首先,对滚动轴承振动信号进行小波、小波包分析,将非平稳信号转化为平稳信号,在平稳信号的基础上建立典型的非高斯分布模型—对数正态分布模型,最后提取每个尺度上的对数正态分布模型参数作为表征轴承运行状态的新特征量。试验证明了所提特征提取方法有效地解决了滚动轴承振动信号的非平稳、非高斯问题。
[Abstract]:Rolling bearing is one of the most widely used and easily damaged parts in rotating machinery. Its working state directly affects the performance of the whole mechanical system, so it is of great practical significance to diagnose its faults. Feature extraction of rolling bearings based on vibration signal analysis is one of the most widely used and effective methods at home and abroad. Statistical distribution model parameters have been widely used in reliability engineering to reflect fatigue life and fatigue strength of mechanical products. However, the research on feature extraction in mechanical, especially bearing condition monitoring and fault diagnosis has not been widely seen. A new fault feature extraction method based on Weibull distribution model parameters and its digital features is proposed in order to fully excavate the effective state change information contained in rolling bearing operation. Based on the Weibull distribution model of the original vibration signal of the rolling bearing, the scale parameters and the median of the model are extracted as a new characteristic vector to characterize the running state of the bearing. Simulation results show the effectiveness of the feature extraction method. Aiming at the non- characteristic of rolling bearing vibration signal, a new idea of fault feature extraction based on logarithmic normal distribution model is proposed, and the logarithmic mean of model parameters is extracted as a new characteristic quantity to characterize rolling bearing running state. The problem of non-Gao Si vibration signal is solved effectively. In view of the problem that the above methods can not accurately describe the non-stationary characteristics of rolling bearing vibration signals, a new method for fault feature extraction of rolling bearings based on wavelet domain logarithmic normal model is proposed. Firstly, the vibration signal of rolling bearing is analyzed by wavelet and wavelet packet, and the non-stationary signal is transformed into stationary signal. On the basis of stationary signal, a typical non- distribution model, logarithmic normal distribution model, is established. Finally, the parameters of the logarithmic normal distribution model on each scale are extracted as the new characteristic variables to characterize the running state of the bearing. The experimental results show that the proposed feature extraction method can effectively solve the non-stationary and non- problem of rolling bearing vibration signal.
【学位授予单位】：湖南工业大学
【学位级别】：硕士
【学位授予年份】：2011
【分类号】：TH165.3

【引证文献】