相关维数在大机组故障诊断中的应用

Application of Correlation Dimension in Large RotatingMachinery Fault Diagnosis

文中提出了将描述混沌运动的特征参数———相关维数（ｃｏｒｅｌａｔｉｏｎｄｉｍｅｎｓｉｏｎ）用于大机组振动信号的分析．一些带有低频噪声的复杂故障信号的分析结果表明，信号在伪相空间中具有不同的相关维数，反映了不同故障的动力学形成机制的不同．因此，相关维数可用于诊断传统方法难以区分的故障．

In this paper, we intend to introduce the correlation dimension as an experimental tool for analyzing irregular vibration signals of large machinary whits. Some faulty vibration signals with subharmonic noise are analyzed by means of correlation dimension. Results show that these noiselike signals possess different correlation dimensions in the reconstructed pseudo phase space, thus the correlation dimension can be used to intelligently recognize different faults, which can be hardly discriminated by means of Fourier spectrum analysis.