基于EEMD分解的阶次跟踪方法研究  

Order tracking method based on EEMD decomposition

作　　者：魏仕华[1] 蔺梦雄 WEI Shihua;LIN Mengxiong(School of Mechanical and Electrical Technology,Taizhou Polytechnic College,Taizhou 225300,China;College of Urban Transit and Logistics,Beijing Union University,Beijing 100101,China)

机构地区：[1]泰州职业技术学院机电技术学院,江苏泰州225300 [2]北京联合大学城市轨道交通与物流学院,北京100101

出　　处：《机电工程》2024年第9期1604-1612,共9页Journal of Mechanical & Electrical Engineering

基　　金：泰州职业技术学院技术开发项目(HXKT-2023-0146)。

摘　　要：摆线针轮减速器组成零部件繁多、构成复杂,工作时噪声干扰大且多在变转速、往复的复杂工况下工作,因此,难以准确提取其内部的故障特征。针对这一问题,提出了一种基于集合经验模态分解(EEMD)与阶次跟踪分析的方法,对摆线针轮减速器进行了故障诊断。首先,对采集到的时域振动信号和转速信号进行了等角度域差值采样,得到了振动信号的等角域平稳信号;然后,对等角域信号进行了集合经验模态分解,得到了若干个固有模态分量(IMFs),计算了各个固有模态分量的峭度值,选取目标模态分量进行了信号重构;接着,采用快速傅里叶变换得到了故障信号的阶次图;最后,根据减速器的传动方式、各零部件的模数,计算出了各主要部件的故障阶次,对比减速器在故障前后阶次图的能量峰值进行了故障诊断。研究结果表明:该方法能够准确提取包含故障信息的固有模态分量,实现从等时域信号到等角域信号的转换,并提取摆线针轮减速器的滚针故障阶次(8.37阶),故障准确率达到99.6%,可实现摆线针轮减速器在非平稳工况下的故障特征识别,并验证该方法的可行性和有效性。A method for fault diagnosis of cycloidal gear reducers based on ensemble empirical mode decomposition(EEMD) and order tracking analysis was proposed,in response to the problems of complex components,high noise interference during operation,and difficulty in accurately extracting internal fault characteristics of cycloidal gear reducers,which often operate under complex working conditions such as variable speed and reciprocating.Firstly,the collected time-domain vibration signal and speed signal were subjected to equal angle domain difference sampling to obtain the equal angle domain stationary signal of the vibration signal.Secondly,the ensemble empirical mode decomposition on angular domain signals was performed to obtain several intrinsic mode functions(IMFs),by calculating the kurtosis values of each intrinsic mode component,the target mode component was selected for signal reconstruction.Then,the order diagram of the fault signal was obtained through fast Fourier transform.Finally,based on the transmission mode of the reducer and the modulus of each component,the fault order of its main components was calculated,and the energy peak of the order graph before and after the fault of the reducer for fault diagnosis was compared.The research results show that the proposed method can accurately extract the intrinsic mode components containing fault information,achieve the conversion from equidistant time domain signals to equiangular domain signals,extract the roller fault order(8.37 orders) of the cycloidal pin wheel reducer,and achieve a fault accuracy of 99.6%.It realizes the fault feature recognition of the cycloidal pin wheel reducer under non-stationary working conditions,and verifies the feasibility and effectiveness of the proposed method.

关 键 词：摆线针轮减速器 集合经验模态分解 阶次跟踪分析 故障诊断 变转速工况 固有模态分量 

分 类 号：TH132.46[机械工程—机械制造及自动化]