基于ROSST的列车轴承轨边声学信号校正与诊断研究

摘要

由列车道旁麦克风采集的列车轴承故障声学信号存在多普勒畸变现象,需要在信号处理过程中进行校正和相关诊断.利用重分配算子同步压缩变换(Reassignment Operator and SynchroSqueezing Transform,ROSST)获取高分辨率时频分布,通过脊线提取、莫尔斯声学理论和非线性拟合得到信号的多普勒畸变参数,再对原信号重采样,消除多普勒畸变.将其应用于仿真信号和列车轴承多普勒畸变故障实验信号的校正与诊断,验证了该方法的有效性.%The phenomenon of Doppler distortion in the acoustic fault signal of moving train needs to be regulated during signal preprocessing.A Doppler distortion removal method based on ROSST (Reassignment Operator Synchro Squeezing Transform) was proposed to solve the problem,and was applied to the wayside fault diagnosis of moving train bearings.The time-frequency distribution with high resolution was obtained by using the ROSST.A resampling method was then carried out to remove the Doppler distortion based on the ridge extraction,Morse acoustic theory and nonlinear data fitting.A simulation signal and experimental acoustic signals of train bearing with defects on the outer race and inner race were utilized to verify the availability.The results indicate the effectiveness of the proposed method.