摘要
旋转机械振动信号中的冲击特征通常代表着轴承损伤、齿轮损伤等常发故障的出现,为了准确提取信号中的冲击分量,提出一种基于改进群延迟估计的同步压缩变换时频分析方法。分析时间重分配同步压缩变换原型算法在处理实际强频变信号时的特性,发现其易导致明显的时频模糊问题。构建基于局部最大搜索算法的改进型经典群延迟估计方法,以克服TSST在分析强频变信号时带来的时频模糊问题,并在此基础上提出了群延迟自适应估计策略。形成一种基于改进群延迟估计的自适应同步压缩变换方法,在其基础上提出一种振动信号中脉冲特征提取方法。仿真信号和试验数据分析结果表明,该方法可较准确地提取出振动信号中的冲击特征,相较其他常用时频分析方法能够生成更为聚集的时频表示。
The impact features in the vibration signal of rotating machinery usually represent the occurrence of common faults such as bearing damage and gear damage. In order to accurately extract the impact component in the signal, a time-frequency analysis method based on improved time-reassignment synchrosqueezing transform(TSST) is proposed. Firstly, the characteristics of TSST prototype algorithm in dealing with actual strongly frequency varying signals are analysed, and it is found that it is easy to cause evident time-frequency ambiguity. Then, an improved group delay estimation method based on local maximum search algorithm is constructed to overcome the time-frequency ambiguity problem caused by TSST. On this basis, an adaptive group delay estimation strategy is proposed. Finally, an adaptive synchrosqueezing transform method based on improved group delay estimation is formed,and a pulse feature extraction method in vibration signal is developed. The results of simulation and experimental data show that the proposed method can extract impulse features of vibration signals more accurately, and generate a more concentrated time-frequency representation than other time-frequency analysis methods.
作者
贺雅
胡明辉
卢子元
明煊
贾彦飞
HE Ya;HU Minghui;LU Ziyuan;MING Xuan;JIA Yanfei(Key Lab of Engine Health Monitoring-Control and Networking of Ministry of Education,Beijing University of Chemical Technology,Beijing 100029;Hangli(Group)Industrial Co.,Ltd,Chengdu,Chengdu 611936)
出处
《机械工程学报》
EI
CAS
CSCD
北大核心
2022年第4期22-33,共12页
Journal of Mechanical Engineering
基金
博士后创新人才支持计划资助项目(BX20180031)。
作者简介
贺雅,女,1995年出生,博士研究生。主要研究方向为航空发动机转子系统故障诊断。E-mail:heya@mail.buct.edu.cn;通信作者:胡明辉,男,1990年出生,博士,副教授。主要研究方向为航空发动机故障诊断与振动抑制。E-mail:humh2008@163.com。
