摘要
针对钢丝绳漏磁检测信号中存在的未知频率噪声干扰问题,提出一种基于多尺度样本熵(Multi-scale Sample Entropy,MSE)的自适应噪声完备集合经验模态分解(Complete Ensemble Empirical Mode Decomposition with Adaptive Noise,CEEMDAN)联合小波包阈值降噪(Wavelet Packet Threshold Denoising,WPTD)方法。将钢丝绳损伤信号分解为一系列本征模态函数(Intrinsic Mode Functions,IMF)分量,分别计算各IMF分量的多尺度样本熵,通过阈值对比筛选出含有噪声的IMF分量,进而利用小波包阈值法对这些分量进行降噪,并与保留的IMF分量进行重组,获取降噪后的损伤信号。实验结果表明,该联合降噪方法比其他降噪方法的信噪比提高了9.2%,均方根误差明显提高,滤波后的损伤信号曲线更加平滑,余弦相似度也大于0.85,能够有效降低检测信号中的噪声分量。
For the problem of unknown frequency noise interference in steel wire rope leakage detection signals.a joint denoising method which combines the complete ensemble empirical mode decomposition with adaptive noise(CEEMDAN)and the wavelet packet threshold denoising(WPTD)based on the multi-scale sample entropy(MSE)is proposed.The steel wire leakage damage signal is decomposed into a series of intrinsic mode functions(IMF),and the MSE of each IMF component is calculated respectively.By comparing the threshold,the IMF components that contain noises are selected,and the wavelet packet threshold is employed to denoise these components.Reconstruct these components by reuniting the retained components to obtain the denoised damage signals.Experiment results show that the SNR of the proposed joint denoising method is 8.9%higher than the other algorithms,and has significantly increased the root mean square error.After the filtering,the denoised signal curve is smoother,and the cosine similarity is also greater than 0.85.It can effectively reduce the noise component in the detection signal.
作者
杨春杰
李聪聪
刘满仓
YANG Chunjie;LI Congcong;LIU Mancang(School of Automation,Xi’an University of Posts and Telecommunications,Xi’an 710121,China;Quality and After-sales Center,Shaanxi Yu Chen Aviation Equipment Co.,Ltd,Xi’an 710089,China)
出处
《西安邮电大学学报》
2023年第4期102-110,共9页
Journal of Xi’an University of Posts and Telecommunications
基金
陕西省技术创新引导计划项目(2020CGXNG-011)
陕西省教育厅服务地方专项项目(20JC0)。
关键词
信号处理
钢丝绳损伤检测
自适应噪声完备集合经验模态分解
多尺度熵
小波包阈值降噪
signal processing
steel wire rope damage detection
complete ensemble empirical mode decomposition with adaptive noise
multiscale entropy
wavelet packet threshold denoising
