基于STOA-VMD结合相关峭度及2.5维谱的行星齿轮箱磨损故障诊断  

Fault diagnosis of planetary gearbox based on STOA-VMD combined with correlation kurtosis and 2.5-dimensional spectrum

作　　者：张家帅 蒋章雷[1] 刘秀丽[1] 吴国新[1] ZHANG Jiashuai;JIANG Zhanglei;LIU Xiuli;WU Guoxin(MOE Key Laboratory of Modern Measurement and Control Technology,Beijing Information Science&Technology University,Beijing 100192,China)

机构地区：[1]北京信息科技大学现代测控技术教育部重点实验室,北京100192

出　　处：《北京信息科技大学学报（自然科学版）》2023年第1期56-61,69,共7页Journal of Beijing Information Science and Technology University

基　　金：北京信息科技大学“勤信人才”培育计划项目(QXTCP C202120)。

摘　　要：针对行星齿轮箱振动信号的非线性特征明显、故障特征难以识别的问题,提出了一种基于乌燕鸥算法(sooty tern optimization algorithm, STOA)优化变分模态分解(variational modal decomposition, VMD),并结合相关峭度以及2.5维谱的故障诊断方法。由于变分模态分解的分解能力取决于二次惩罚因子α和本征模态分量个数K两个参数,所以首先运用乌燕鸥算法对VMD参数进行优化,确定最优的参数,然后以相关峭度为依据重构本征模态分量,最后利用2.5维谱对重构信号进行分析。搭建了行星齿轮箱磨损故障加速疲劳寿命试验台,获得了故障数据,运用STOA-VMD结合相关峭度及2.5维谱的方法对故障数据进行分析。结果表明该方法能够准确提取磨损故障特征频率。To solve the problems of obvious nonlinear characteristics and difficult identification of fault characteristics of planetary gearbox vibration signals, a fault diagnosis method was proposed based on sooty tern optimization algorithm(STOA) optimized variational modal decomposition(VMD) and combined with correlation kurtosis as well as 2.5-dimensional spectrum.Because the decomposition ability of VMD hinges on the secondary penalty factor α and the number of eigenmodal components K,so firstly, the STOA was used to optimize the parameters of the VMD to determine the optimal parameters.Then the eigen modal components were reconstructed based on the correlation kurtosis, and finally the reconstructed signals were analyzed using 2.5-dimensional spectrum.The accelerated fatigue life test rig for planetary gearbox wear fault was built, and the fault data were obtained.The method of STOA-VMD combined with correlation kurtosis and 2.5-dimensional spectrum was used to analyze the fault data.The outcomes indicate that the raised method can accurately recover the characteristic frequency of wear fault.

关 键 词：行星齿轮箱 磨损故障 变分模态分解 乌燕鸥算法 相关峭度 2.5维谱 

分 类 号：TH132.4[机械工程—机械制造及自动化] TH17