【摘要】 针对滚动轴承故障振动信号的非线性和非平稳特性的情况,提出了一种基于固有时间尺度分解和样本熵的新型故障特征提取方法,并与Tikhonov支持向量机相结合实现滚动轴承的故障诊断。该研究充分利用了固有时间尺度分解具有提取故障特征明显、计算简单等优点。首先采用固有时间尺度分解方法将振动信号分解为一序列固有旋转分量和一个基线分量之和,并计算每个固有旋转分量的瞬时幅值和瞬时频率。然后,提取上述瞬时数据的样本熵作为特征向量。最后将其作为Tikhonov支持向量机的输入,实现滚动轴承故障精确分类。经过实验验证,本文方法获取的不同类型故障样本特征差别较大,与小波能谱熵、时间小波能谱熵相比能够更精确和快速的识别轴承故障。更多还原

【Abstract】 According to the nonlinear and non-stationary characteristics of the vibration signals from rolling bearing,a new feature extraction method based on intrinsic time-scale decomposition( ITD) and sample entropy is proposed,then combined with Tikhonov support vector machine to achieve the fault diagnosis. This paper makes full use of the advantages of fault feature obvious different and low time complexity of ITD. Firstly,the bearing vibration signals are decomposed into a set of proper rotation components and a trend component by means of ITD. Then,the instantaneous amplitude and frequency information of each rotation component are calculated. Sample entropy of the faulted bearing is extracted from these instantaneous data. Finally,these sample entropy is used as the input feature vector of Tikhonov support vector machine to identify the different bearing faults precisely. The simulation results show that this method obtains great different fault samples. Compared with the wavelet energy spectrum and the time-wavelet energy spectrum diagnosis methods,the proposed method owns a higher accuracy and speed in identify roller bearing fault.更多还原