基于CCSO-VMD的声发射脱粘信号特征提取  

Feature extraction of acoustic emission debonding signals based on CCSO-VMD

作　　者：李玉珠 金永[1] Li Yuzhu;Jin Yong(College of Information and Communication Engineering,North University of China,Taiyuan 030051,China)

机构地区：[1]中北大学信息与通信工程学院,太原030051

出　　处：《电子测量技术》2024年第19期96-103,共8页Electronic Measurement Technology

基　　金：自然科学研究面上项目(202103021224202)资助。

摘　　要：声发射检测方法目前被广泛应用在设备的缺陷检测方面,针对大部分设备产生的声发射信号幅度小,噪声大且特征难以提取的问题,本文提出了一种信号处理方法:基于皮尔逊相关系数-包络熵最小原则的CCSO算法优化VMD中参数的处理方法。该方法在经典的鸡群优化算法之上融入了交叉算法,并通过改进的CCSO算法来精确优化VMD中的关键参数,即模态数K和惩罚因子α。通过采用基于新适应度函数的CCSO-VMD方法,对模拟信号进行分析,信噪比达到了25.8141 dB。这一结果证明,基于新目标函数的CCSO-VMD算法能够显著降低噪声水平,同时最大程度地保留信号中的有效信息。此外,本文提出了一种综合频谱差异指标,CSDI值能有效区分不同状态下的声发射信号。Acoustic emission detection methods are widely used in the defect detection of equipment,for most of the equipment generated acoustic emission signal amplitude is small,large noise and features are difficult to extract the problem,this paper proposes a signal processing method:the CCSO algorithm based on the Pearson correlation coefficient-envelope entropy minimum principle optimizes the processing method of parameters in VMD.In this method,the cross algorithm is integrated on top of the classical flock optimization algorithm,and the key parameters in the VMD,namely the modal number K and the penalty factorα,are accurately optimized by the improved CCSO algorithm.By using the CCSO-VMD method based on the new fitness function,the analog signal was analyzed,and the signal-to-noise ratio reached 25.8141 dB.This result proves that the CCSO-VMD algorithm based on the new objective function can significantly reduce the noise level while retaining the valid information in the signal to the greatest extent.In addition,this paper proposes a comprehensive spectral difference index,and the CSDI value can effectively distinguish the acoustic emission signals in different states.

关 键 词：声发射信号 脱粘检测 交叉鸡群优化算法 变分模态分解 适应度函数 综合频谱差异指标 

分 类 号：TB535[理学—物理] TN06[理学—声学]