基于EEMD与模糊信息熵的旋转机械故障诊断方法

针对旋转机械故障识别率偏低的问题,提出一种基于EEMD与模糊信息熵的旋转机械故障诊断方法.该方法结合EEMD分解和模糊信息熵在特征提取方面的优势,构造出一种能够精细度量不同类别振动信号故障概率复杂度的特征集合.首先将原振动信号进行EEMD分解,获得若干个本征模态函数(IMFs);计算出前5个高频IMF分量的模糊信息熵组成高维特征集;利用LPP对高维特征集进行维数约简剔除冗余不相关特征;最后将约简后的样本集输入到KNN分类器中进行故障识别.用双跨转子实验台采集的数据对所述方法进行验证,并与EMD模糊熵、EMD模糊信息熵、EEMD模糊熵方法进行故障识别率对比,结果表明该方法能够有效提取转子振动信号的故障特征,并且具有更高的故障识别率.

Fault diagnosis method of rotating machinery based on both EEMD and fuzzy information entropy

Aimed at the lower accuracy of fault recognition of rotary machinery, the fault diagnosis method is proposed based on both EEMD and fuzzy information entropy. By using this method and incorporating with the superiority of EEMD and fuzzy information entropy in connection with feature extraction, a feature set is constructed, which can finely measure the complexity of the fault probability of different category of vibration signals. Firstly, the original vibration signal is decomposed with EEMD to obtain several intrinsic modality functions (IMFs), the fuzzy information entropy of the first 5 high-frequency IMF components is calculated to compose high-dimensional feature set then the LPP is used to reduce the dimensionality of the high-dimensional feature set and eliminate redundant irrelevant features, and finally, the reduced sample set is input into the KNN classifier to identify the faults. Foregoing method is validated by data collected from a double-span rotor test rig and compared with the methods of EMD fuzzy entropy, EMD fuzzy information entropy and EEMD fuzzy entropy in connection with the fault recognition accuracy. The result show that this method is able to extract effectively the fault feature of rotor vibration signals.It have higher fault recognition accuracy.