首页|基于CEEMDAN-WSVD组合串扰消除法车内噪声源识别

基于CEEMDAN-WSVD组合串扰消除法车内噪声源识别

扫码查看
原文链接
  • 万方数据
  • 维普
为解决车内噪声源识别中结构路径易受外部因素干扰,以及多源振动串扰影响,导致采集的工况数据存在噪声等问题,提出基于自适应噪声的完备集合经验模态分解(Complete Ensemble Empirical Mode Decomposition with Adaptive Noise,CEEMDAN)的CEEMDAN-WSVD组合去噪法,该方法利用自适应加噪特征避免模态混叠现象发生,引入样本熵对高频含噪分量进行小波变换(Wavelet Transform,WT),实现一层降噪后进行重构;并采用奇异值分解(Singular Value Decomposition,SVD)对重构信号获取主分向量,同时使用主分量衰减方法剔除较小主分量,实现二层降噪.运用模拟仿真信号验证上述方法对复杂含噪信号有降噪效果.通过对采集的工况数据降噪处理,计算路径传递率并得到贡献量.将各降噪方法应用于工况传递路径模型中对比分析,发现经过本文方法降噪后模型的合成响应与实测响应准确性较高,降噪效果较优.
Vehicle's Internal Noise Source Identification Based on CEEMDAN-WSVD Combined Crosstalk Cancellation Method
The structural path in the vehicle's internal noise source identification is prone to be disturbed by external factors and impacted by multi-source vibration crosstalk,which leads to the existence of the noise in the collected working condition data.In order to solve this problem,the Complete Ensemble Empirical Mode Decomposition with Adaptive Noise(CEEMDAN)method combined with WSVD is proposed.In this method,the adaptive noise addition method is used to avoid the modal overlap phenomenon.And the sample entropy is introduced into the noisy high-frequency components to perform wavelet transform,so as to realize one-layer noise reduction and reconstruction.The singular value decomposition(SVD)is carried out on the reconstructed signal to obtain the principal components.Meanwhile the smaller principal components are removed by using the principal component attenuation method to realize two-layer noise reduction.The denoising effect of the above method on the complex noisy signal is verified by the simulation signal.Through the denoising processing of the collected working condition data,the path transmission rate is calculated and the noise contribution is obtained.The noise reduction methods are applied to the operational transfer path model for comparison and analysis.It is found that the accuracy of the synthetic response of the proposed model after noise reduction is higher than that of the measured response,and a better noise reduction effect is obtained.

acousticscomplete ensemble empirical mode decomposition with adaptive noisewavelet transformsingular value decompositionoperational transfer path analysisnoise source identification

李艺江、陈克

展开 >

沈阳理工大学 汽车与交通学院,沈阳 110159

声学 完备集合经验模态分解 小波变换 奇异值分解 工况传递路径 噪声源识别

2022年教育厅基本科研资助项目

LJKMZ20220603

2024

10.3969/j.issn.1006-1355.2024.04.034

噪声与振动控制
中国声学学会

噪声与振动控制

CSTPCD北大核心
影响因子：0.622
ISSN：1006-1355
年,卷(期)：2024.44(4)
  • 11