FA-PMA-VMD方法及其在齿根裂纹故障诊断中的应用  

FA-PMA-VMD method and its application in gear tooth root crack fault diagnosis

作　　者：程军圣[1] 李梦君 欧龙辉 杨宇[1] CHENG Junsheng, LI Mengjun, OU Longhui, YANG Yu(State Key Laboratory- of Advanced Design and Manufacture for Vehicle Body, Hunan University, Changsha 410082, China)

机构地区：[1]湖南大学汽车车身先进设计制造国家重点实验室,长沙410082

出　　处：《振动与冲击》2018年第15期27-32,67,共7页Journal of Vibration and Shock

基　　金：国家科技支撑计划课题(2015BAF32B03);国家自然科学基金(51575168;51375152);智能型新能源汽车国家2011协同创新中心资助;湖南省绿色汽车2011协同创新中心资助

摘　　要：针对变分模态分解(VMD)中难以确定分解分量个数k和惩罚参数α的问题。提出一种改进的变分模态分解方法—基于萤火虫算法及主模态分析法的变分模态分解(FA-PMA-VMD)方法。该方法用主模态分析(PMA)对VMD分解的带限内禀模态函数(BIMF)分量进行排序;用萤火虫算法对变分模态分解的最佳影响参数[k,α]组合进行搜索,以新提出的正交低峰值作为萤火虫算法的优化目标,得到的最佳的惩罚参数α和分量个数k组合;根据预先设定的故障特征参数自适应地将信号分解为k个BIMF分量。通过对仿真信号和齿轮齿根裂纹实际故障信号进行分析,分析结果表明FA-PMA-VMD具有良好的分解效果。Aiming at the decomposed components ＇ number k and the penalty parameter α in the method of variational mode decomposition（ VMD） being difficult to determine,an improved VMD method,based on the firefly algorithm（ FA） and the principle mode analysis（ PMD）,named the method of FA-PMA-VMD was proposed. Firstly,the new method used PMA to sort band-limited intrinsic mode function（ BIMF） components obtained with VMD. Then,FA was used to search the optimal influence parameter combination k and α in VMD with a new proposed orthogonal low peak value taken as the objective of FA,the optimal combination of k and α was gained. Finally,a signal was decomposed into k BIMF components adaptively with FA-PMA-VMD according to the preset fault characteristic parameters. The analysis results of simulated signals and actual fault signals of gear tooth root cracks showed that the proposed method of FA-PMAVMD has a good decomposition effect.

关 键 词：变分模态分解(VMD) 主模态分析(PMA) 萤火虫算法(FA) 齿根裂纹 故障诊断 

分 类 号：TP242[自动化与计算机技术—检测技术与自动化装置]