Combined Noise Reduction Method for Collected Voiceprint Data of Traction Fans of HXD1 Locomotive
Traction fans in the HXD1 locomotive workshop are key electrical and mechanical equipment.Voiceprint analysis technology can be used to monitor the status and diagnose the faults of the fans.As there are many mechanical and electrical components in the locomotive workshop,the collected voiceprint monitoring data of traction fans would be affected by various noise data.A combined noise reduction method based on ICEEMDAN and SANC is proposed for collected voiceprint data of traction fans of HXD1 locomotive to reduce noise data interference and effectively extract the required voiceprint eigen values for analysis.This method involves:decomposing the noisy signal and the reference signal by ICEEMDAN respectively to obtain two groups of corresponding intrinsic modal components(IMFs);using SANC to perform adaptive noise reduction on IMFs containing the noisy signal to obtain the collected voiceprint data after combined noise reduction.In the time domain indexes of the simulated electric whistle signal and impact signal,the SNR value,MSE value,and correlation coefficient of the proposed algorithm are better than those of the SANC and ICEEMDAN algorithms.In the time domain indexes of the measured electric whistle signal and impact signal,the SNR value,MSE value,and correlation coefficient of the proposed algorithm are also better than those of the SANC and ICEEMDAN algorithms.Therefore,this combined noise reduction method is adaptive,feasible,and effective.It can facilitate noise reduction of collected voiceprint monitoring data of traction fans in the machinery workshop,laying a foundation for status monitoring and fault diagnosis analysis of traction fans.
traction fans of locomotivevoiceprint monitoringdata processingICEEMDANSANCadaptive noise reduction
NETL
NSTL
万方数据
