基于LMD自适应多尺度形态学和Teager能量算子方法在轴承故障诊断中的应用

为了从故障轴承信号中提取包含故障信号的特征频率，提出了基于LMD自适应多尺度形态学和Teager能量算子解调的方法。首先，采用LMD将目标信号分解成有限个PF（Product function，PF）分量，分别对其进行多尺度形态学滤波，利用峭度准则优化形态学结构元素尺度，自适应寻求最优解，最后用Teager能量算子计算各PF分量的瞬时幅值，通过瞬时Teager能量的Fourier频谱识别轴承的故障特征频率。为了验证理论的正确性，进行了数字仿真实验和轴承故障模拟实验，并与EMD形态学和包络解调方法进行了比较，结果表明该算法明显优于其他两种方法，对滚动轴承外圈、内圈和滚子故障的检测精度更高，能够清晰地提取出故障信号的频率特征。

Bearing fault feature extraction method based on LMD adaptive multiscale morphological and energy operator demodulating

In order to extract the characteristic frequencies from bearings fault signals containing fault information, the adaptive morphology method was proposed based on LMD and energy operator demodulating. LMD was used to decompose multi-component AM-FM signal into number of production functions(PF) firstly, the PFs containing fault information were multiscale morphologically filtered respectively. The kurtosis criterion was used to adaptive optimize the structural elements of morphology, Finally, the energy energy operator demodulating is applied to each PF and the amplitudes and frequencies of a multi-componentAM-FM signal are extracted for bearing fault diagnosis. In order to verify the correctness of the theory, through numerical simulations and bearings fault simulation tests,the proposed method was compared with EMD and envelope demodulation. The results showed that the proposed method is superior to the other two, the higher detection accuracy of the outer and inner ring of rolling bearing and roller faults, it can be used to clearly extract various characteristic frequencies of bearings faults.