摘要
针对单一信号通道反映故障信息不全面、不准确的问题,提出利用多元变分模态分解(MVMD)处理多通道信号提取故障特征,实现故障诊断。首先通过构建多分量调制仿真信号,分析比较MEMD、NAMEMD和MVMD的分解效果,然后利用MVMD对柴油机4个通道振动信号进行自适应分解,提取每层分量的能量分布作为故障特征,最后利用支持向量机对不同失火故障进行了识别。结果表明,MVMD在抑制模态混叠和分解效率上均优于其他两种算法,且能够有效识别柴油机不同类型失火故障。
For the problems of incomplete and inaccurate fault information of univariate signal,multivariate variational modal decomposition(MVMD)was proposed to process multi-channel signals to extract fault features and hence realize fault diagnosis.A multi-component modulation simulation signal was constructed and the decomposition effects of MEMD,NAMEMD and MVMD were analyzed and compared.The adaptive decomposition of the vibration signals from the four channels for diesel engine was then conducted by using MVMD and the energy distribution of each IMF component was extracted as the fault feature.Finally,different misfire faults were identified by support vector machine.The results show that MVMD is better than the other two algorithms in suppressing modal aliasing and decomposition efficiency and can effectively identify different misfire faults of diesel engine.
作者
顾程
乔新勇
靳莹
韩立军
GU Cheng;QIAO Xinyong;JIN Ying;HAN Lijun(Department of Vehicle Engineering,Army Academy of Aromed Forces,Beijing 100072,China;Urumqi Campus of Engineering University of PAP,Urumqi 830049,China)
出处
《车用发动机》
北大核心
2020年第6期83-89,共7页
Vehicle Engine
基金
国防973项目(6312520302)
军内科研计划项目(2018ZB58)。
关键词
多元变分模态分解
振动信号
故障诊断
multivariate variational modal decomposition(MVMD)
vibration signal
fault diagnosis