基于EEMD和二维边际谱熵的齿轮箱故障诊断

针对信号经验模态分解（EMD）过程中存在波形混叠现象，提出一种基于聚合经验模态分解（EEMD）和Hilbert二维边际谱熵相结合的方法对齿轮箱故障进行分类故障诊断．首先使用小波阈值分析对背景噪声较大的齿轮箱振动信号进行预处理；其次对预处理信号进行分解，得到IMF分量，对比正常信号与故障信号的区别；最后对3种工况信号进行Hilbert变换并计算得到边际谱，并且提取二维边际谱熵作为支持向量机（SVM）的输入量，建立故障诊断模型．经测试该方法在齿轮箱故障诊断方面有着较强的分类能力和诊断精度，具有一定的可行性．

Gearbox Fault Diagnosis based on EEMD and Two-Dimensional Marginal Spectrum Entropy

In order to restrain the mode mixing in the process of Empirical Mode Decomposition （EMD）, a signal analysis method for gearbox fault diagnosis is presented based on both Ensemble Empirical Mode Decomposition （EEMD） and Hilbert two-dimensional marginal spectrum entropy. Firstly, wavelet threshold analysis is used to preprocess the vibration signals with bigger background noises from the gearbox. Secondly, the preproeessed signals are decomposed into many Intrinsic Mode Function （IMF） components by means of the EEMD method, comparing the fault signals with the normal signals. Finally, the Hilbert transform is applied to each intrinsic mode function, obtaining total HHT marginal spectrum of gearbox vibration signal. The two-dimensional marginal spectrum entropy is extracted and input into support vector machine （SVM） classification for establishing a diagnosis model. The results show that this method has a good performance in fault classification and diagnosis precision.