CEEMDAN-PE-TFPF降噪法在齿轮故障诊断中的应用  

Application of CEEMDAN-PE-TFPF De-Noising Method in Gear Fault Diagnosis

作　　者：白丽丽 韩振南[1] 任家骏[1] BAI Li-li;HAN Zhen-nan;REN Jia-jun(College of Mechanical Engineering,Taiyuan University of Technology,Shanxi Taiyuan030024,China)

机构地区：[1]太原理工大学机械工程学院

出　　处：《机械设计与制造》2020年第1期80-83,88,共5页Machinery Design & Manufacture

基　　金：国家自然科学基金资助项目（50775157）;国家自然科学基金资助项目（51805355）

摘　　要：针对齿轮故障诊断过程中,大量噪声使得故障特征难以完全提取的情况,提出了一种完整的自适应噪声集成经验模态分解(Complete Ensemble Empirical Mode Decomposition with Adaptive Noise,CEEMDAN)、排列熵(Permutation Entropy,PE)和时频峰值滤波(Time-Frequency Peak Filtering,TFPF)相结合的去噪方法。由于TFPF方法在窗长问题上的局限性,引用CEEMDAN和PE对此进行改进,使信号在噪声抑制和信号保真方面得到了很好的权衡。首先利用CEEMDAN算法得到原始信号的本征模态函数(Intrinsic Mode Functions,IMFs),计算每个IMF的PE值来判断IMF是否需要滤波,然后针对不同的IMF选择不同的窗口长度进行TFPF滤波,最后将滤波后的IMFs和剩余IMFs重构得到最终的降噪信号。通过模拟仿真信号和实测齿轮信号验证了该降噪方法的可行性,且降噪后的信号可以有效地揭示故障的特征信息。最后与多种降噪方法对比,体现了所提方法的有效性和优越性。In the process of gear fault diagnosis,it is difficult to extract the fault features completely due to a large amount of noise. A new method based on Complete Ensemble Empirical Mode Decomposition with Adaptive Noise(CEEMDAN)、Permutation Entropy(PE)and Time-Frequency Peak Filter(TFTP)is proposed in this paper. Due to the limitation of the TFPF method in the window length problem,CEEMDAN and PE are applied to improve this method,which makes the signal get a good balance both noise suppression and signal fidelity. Firstly,the Intrinsic Mode Function(IMFs)of the original signal are obtained using CEEMDAN algorithm,and calculate the PE value of each IMF to determine whether the IMF requires filtering. Then,the different window length is selected to filter using TFPF for different IMFs. Finally,the signal is reconstructed as the sum of the filtered and residual IMFs. The feasibility of the proposed method is verified by the simulation signal and the measured gear signal,and the de-noised signal can effectively reveal the fault characteristic information. Finally,the proposed met hod is compared with other de-noising methods,which shows the effectiveness and superiority of the proposed method.

关 键 词：CEEMDAN 排列熵 时频峰值滤波 齿轮 降噪 

分 类 号：TH16[机械工程—机械制造及自动化] TH132.41