论文部分内容阅读
基于滚动轴承振动信号的各种三维或二维谱图中包含的不同故障信息的客观现实,通过二维小波变换在不同尺度空间下构造的小波共生矩阵提取了谱图的纹理特征向量.利用灰关联分析表征这些纹理特征的不同发展态势从而实现滚动轴承的故障诊断.对实测滚动轴承不同状态故障数据的分析表明:该方法具有较高的故障模式分类精度;随着故障尺寸的增加,由于轴承各部件的相互影响诊断正确率会有所降低.同时研究表明对于特定的诊断方法是否进行特征向量归一化需区别对待.
Based on the objective reality of different fault information contained in various three-dimensional or two-dimensional spectra of rolling bearing vibration signals, the texture feature vectors of the spectra are extracted by the wavelet co-occurrence matrix constructed by two-dimensional wavelet transform in different scales. The analysis of the different development trend of these texture features in order to realize the fault diagnosis of rolling bearing.The analysis of the fault data of different states of the measured rolling bearing shows that this method has a higher classification accuracy of the fault mode.With the increase of the fault size, The accuracy of mutual diagnosis will be reduced.At the same time, research shows that for a particular diagnostic method whether the eigenvector normalization needs to be treated differently.