摘要
针对电动汽车用电机输出轴直接与太阳轮连接的电机-行星齿轮集成系统。开展电机-行星齿轮系统故障信号重构及识别分析。设计了一种特征提取与故障识别流程,通过筛选互信息比定阈值更大的IMF并完成信号重构,输入PSO-SVM中并识别行星齿轮故障模式。研究结果表明:对信号重构扩散熵进行量化分析,评估振动复杂性。嵌入维数设定在2或3,散布模式的数量设置成6。每种状态下扩散熵发生大幅变化,剥落时发生重叠结果。利用VMD分解重建可以提高识别效果,能显著减少噪声成分,准确率为99.99%。该研究有助于提高电机输出传动系统的运行稳定性,具有很高的实际指导意义。
Aiming at the motor-planetary gear integration system,i.e.the output shaft of electric motor for electric vehicles directly connect with the sun wheel,this paper carries out the analyses on the fault signal reconstruction and identification of the system.A process of feature extraction and fault identification is designed.The IMF whose mutual information is larger than the fixed threshold is screened,the signal is reconstructed and input into PSO-SVM,and the fault mode of planetary gear is identified.The research results show that the diffusion entropy changes greatly in each state,such as quantitatively analyzing the signal reconstruction diffusion entropy,evaluating the vibration complexity,setting the embedding dimension to 2 or 3 as well as setting the number of scatter modes to 6,and the overlapping result occurs during spalling.The decomposition and reconstruction by VMD can improve the identification effect and significantly reduce the noise components,which accuracy is 99.99%.This research is helpful to improve the operation stability of the motor output drive system and has a very practical guiding significant.
作者
罗志华
Luo Zhihua(School of Automotive Engineering,Henan Vocational College of Industry and Trade,Zhengzhou 450053,China)
出处
《防爆电机》
2025年第4期113-116,共4页
Explosion-proof Electric Machine
基金
河南省高等学校重点科研项目(编号:22A470005)。
关键词
电机-行星齿轮系统
故障识别
变分模态分解
支持向量机
准确率
Motor-planetary gear system
fault identification
variational mode decomposition
support vector machine
accuracy rate
