【作者】 李真；

【导师】 程卫东；

【作者基本信息】 北京交通大学 ， 机械制造及其自动化， 2014， 硕士

【摘要】 滚动轴承是旋转机械的关键组成部分,也是极容易损坏的部件之一,长期以来,对旋转机械滚动轴承的故障诊断一直是业内研究的重点和难点。通过分析旋转机械运行过程中产生的振动信号可以实现对滚动轴承的故障诊断。绝大多数旋转机械都是在变转速工况下运行的,因此所产生的振动信号大部分都是非平稳信号。当滚动轴承发生故障时,这些变转速信号将会表现的更加明显,并且其中包含丰富的故障特征信息。然而现有的信号处理方法并不能有效地完成变转速工况下滚动轴承的故障诊断。本文在对现有的信号处理方法进行研究的基础上,提出了熵选择IMF分量的滚动轴承故障诊断方法。该方法将EMD分解、香农熵、傅里叶变换、阶比跟踪以及包络解调等方法进行了有机结合,实现了对变转速工况下振动信号的处理与分析,最终利用分析结果完成了对变转速工况下滚动轴承的故障诊断,该方法的具体步骤如下。首先,利用EMD分解方法对采集到的振动信号进行预处理,将其自适应的分解为一系列的本征模函数,即IMF分量。该方法可以将原振动信号分解为不同频率成分的振动信号,起到了滤波的作用。其次,选择合适的IMF分量分别提取滚动轴承的转速信息与进行滚动轴承的故障诊断。为了更好的实现对故障信号的提取分析,首先需要求出EMD分解之后得到的各IMF分量的香农熵,然后根据香农熵值的大小判断信号的无序性程度。香农熵值越小,信号越有序；香农熵值越大,信号越无序。当滚动轴承发生故障时,其故障表面与其它元件相接触时会发生周期性的撞击,从而产生间隔均匀的脉冲,因此本文选择香农熵值最小的IMF分量提取滚动轴承的故障特征。转速信息一般位于信号的低频段,仅仅利用香农熵值的大小很难判断信号频率的分布情况,所以需要对IMF分量做FFT变换,了解各IMF分量的频率分布,进而选择合适的分量来提取振动信号中的转速信息。最后,提取故障特征,对滚动轴承进行故障诊断。利用提取的转速信息对熵值最小的IMF分量做等角度重采样,将其从时域非平稳信号转换为角域平稳信号,对角域平稳信号做包络解调,得到包络阶比谱,判断滚动轴承的故障类型。本文以多功能转子试验台为实验平台,对上述方法进行了实验验证,实验结果表明该方法能够准确可靠地实现变转速工况下滚动轴承的故障诊断。更多还原

【Abstract】 Rolling bearing is the key component as well as one of the quick-wear parts of the rotating machinery. For a long time, the fault diagnosis of the rolling bearing used in rotating machinery is the priority and difficulty in the industry research. The fault diagnosis of the rolling bearing can be realized by analyzing the vibration signals which are produced during the running of the rotating machinery. As most of the rotating machinery is running under variable speed condition, so the vibration signals generated in this process are mostly non-stationary signal. These variable speed signals will become more obviously when the fault is happened on rolling bearing. In addition, there are rich fault characteristic information contained in these variable speed signals. However, the existing signal processing methods cannot effectively complete the fault diagnosis of the rolling bearing under the variable speed condition.In this paper, a fault diagnosis method of the rolling bearing based on entropy selecting IMF component is proposed based on the study of existing signal processing method. Methods of EMD decomposition, Shannon entropy, Fourier transform, Order tracking and Envelope demodulation are organic combined in this method and the processing and analysis of the vibration signals are achieved under the variable speed condition. Finally, the fault diagnosis of the rolling bearing is accomplished by utilizing the analysis results. The specific steps are as follows.Firstly, the collected vibration signals are preprocessed by using EMD decomposition and they are adaptively resolved in a series of intrinsic mode function (IMF). This method plays the role of the filtering by resolving the original vibration signal into different frequency components.Secondly, choose the appropriate IMF component to extract the rotating speed information from the vibration signals and diagnose the fault of rolling bearings respectively. In order to better extract and analyze the fault signal, the Shannon entropy of each IMF component obtained after the EMD decomposition needs to be calculated first, then according to the Shannon entropy, the disorder degree of the signal can be determined. The smaller the Shannon entropy, the more orderly the signal; the greater the Shannon entropy, the more chaotic the signal.When the rolling bearing malfunctions, the periodic impact would occur when the fault surface contacted with other components, which would produce evenly spaced pulse. Thus, the IMF component with minimum Shannon entropy is chosen to extract the fault features of the rolling bearing in this paper. Speed information is generally located in the low frequency signal, it is hard to judge the signal frequency distribution by using the Shannon entropy, so the FFT transform for IMF component is needed in order to know the frequency distribution of each IMF component in this paper. Then, the appropriate component is chosen to extract the speed information from the vibration signals.Finally, the fault features are extracted to conduct the fault diagnosis of the rolling bearing. The uniform angle reampling processing is conducted to the IMF component with minimum entropy by utilizing the extracting speed information and it is converted from the time domain non-stationary signal into angle domain stationary signal. Then the fault type of rolling bearing could be judged by the envelope order spectrum which are got from the envelope demodulation which is done to the angle domain stationary signal.At the end of this paper, the experiments are carried out to verify the above method by using the multifunctional rotor test bench. The experimental results show that this method can realize the fault diagnosis of the rolling bearing accurately and reliably under the variable speed condition.更多还原