基于改进的HHT边际谱齿轮箱故障诊断  

Gearbox Fault Diagnosis Based on Improved Hilbert-Huang Transform Marginal Spectrum

作　　者：马百雪[1] 潘宏侠[1] 杨素梅[1] 

机构地区：[1]中北大学机械工程与自动化学院,太原030051

出　　处：《煤矿机械》2013年第10期251-254,共4页Coal Mine Machinery

基　　金：国家自然科学基金项目(50875247)

摘　　要：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边际谱相结合的方法对齿轮箱故障进行故障诊断。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,提高EEMD分解的精确度;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2种工况信号进行Hilbert变换并计算得到边际谱,确定故障信号的故障频率。研究表明该方法在避免EMD分解带来的模态混叠现象方面具有可行性,能提高齿轮箱故障诊断的准确率。A signal analysis technique for gearbox fault diagnosis based on ensemble empirical mode decomposition （EEMD） and Hilbert--Huang transform （HHT） is presented in this paper to restrain the mode mixing of Empirical Mode Decomposion（EMD）. Firstly, wavelet threshold analysis is carried out to preprocess the gearbox vibration signal which contains large of backgroun accuracy of EEMD. Secondly, the EEMD method is used to decompose the d noise. It can improve the gearbox vibration signal into many of intrinsic mode function（IMF） components and contrast the difference between the normal signal and fault signal.Then the Hilbert transform is applied to each intrinsic mode function.Therefore the total marginal spectrum of gearbox vibration signal is obtained. It will be easier to recognize the failure frequency of gearbox fauh signal. The results show that this method can effectively avoid thephenomenon of mode mixing and improve the accuracy of gearbox fauh diagnosis.

关 键 词：聚合经验模态分解 Hilbert边际谱 小波降噪 故障诊断 

分 类 号：TB112[理学—数学]