采用EEMD算法与互信息法的机械故障诊断方法

Mechanical Fault Diagnosis Method Using EEMD Algorithm and Mutual Information Method

提出一种总体经验模态分解(EEMD)算法与互信息法相结合的Hilbert-Huang变换机械故障诊断改进的方法.仿真与实例结果表明:EEMD算法能克服模态混叠弊端,获得具有实际物理含义的固有模态函数(IMF);互信息法能有效剔除虚假分量,使最终IMF分量更加精准且集中突显故障信号特征;所提出方法能有效表征机械故障特征,并进行精确诊断.

An improved Hilbert-Huang transform method for fault diagnosis was proposed which combined ensemble empirical mode decomposition(EEMD)algorithm and mutual information method.The simulation and example results showed that EEMD algorithm can overcome the drawbacks of modal aliasing and obtain the intrinsic modal function(IMF)with practical physical meanings and the mutual information method can effectively eliminate the false components,which makes the final IMF components more accurate and more concentrated on the fault signal characteristics.The proposed method can effectively characterize the mechanical failure and realize accurate diagnosis.