汽车工程 ›› 2018, Vol. 40 ›› Issue (10): 1172-1178.doi: 10.19562/j.chinasae.qcgc.2018.010.008

• • 上一篇    下一篇

柴油机振动信号自适应分解方法的比较*

贾继德1, 任刚2, 贾翔宇2, 韩佳佳2   

  1. 1.陆军军事交通学院军用车辆系,天津 300161;
    2.陆军军事交通学院研究生管理大队,天津 300161
  • 收稿日期:2017-12-12 出版日期:2018-10-25 发布日期:2018-10-25
  • 通讯作者: 贾继德,教授,博士,E-mail:jide@ustc.edu。
  • 基金资助:
    *陆军装备部重点项目(WG2015JJ010008)资助。

Comparison of Adaptive Decomposition Methods for Diesel Engine Vibration Signals

Jia Jide1, Ren Gang2, Jia Xiangyu2, Han Jiajia2   

  1. 1.Military Vehicle Department, Army Military Transportation University, Tianjin 300161;
    2.Postgraduate Training Brigade, Army Military Transportation University, Tianjin 300161
  • Received:2017-12-12 Online:2018-10-25 Published:2018-10-25

摘要: 变分模态分解是一种新的自适应分解方法,为检验其对柴油机信号的适用性,建立多分量、调幅调频和含噪仿真信号,采用变分模态分解法对其进行分解,并与其它自适应分解方法从分解效果、抑制模态混叠和端点效应能力等方面进行比较;接着分解柴油机瞬变工况的振动信号,发掘曲轴轴承磨损信号变化规律,提取故障特征;最后利用支持向量机进行故障类型识别,进一步验证该方法的有效性。结果表明:变分模态分解在分解效果、抑制模态混叠和端点效应能力等方面均优于其他自适应分解方法,适用于柴油机状态监测和故障诊断。

关键词: 柴油机, 振动信号, 自适应分解, 变分模态分解, 故障诊断

Abstract: Variational mode decomposition (VMD) is a new adaptive decomposition method. In order to check its suitability for diesel engine signals, a multi-component, AM-FM simulation signal is built with white noise added, which is then decomposed by using VMD and compared with other adaptive decomposition methods in terms of decomposition effects and the capability to suppress modal aliasing and endpoint effects. Then, the vibration signals of diesel engine under transient conditions are decomposed, the changing pattern of wear signals of crankshaft bearing is explored and fault features are extracted. Finally fault types are identified by using support vector machine, to further verify the effectiveness of the method adopted. The results show that VMD method is better than other adaptive decomposition methods in decomposition results and the capability to suppress modal aliasing and endpoint effects, suitable for the state monitoring and fault diagnosis of diesel engine

Key words: diesel engine, vibration signal, adaptive decomposition, VMD, fault diagnosis