摘要
针对胶合板损伤声发射(AE)信号的非平稳性和损伤类别特征相互重叠的实际情况,提出了基于经验模态分解(EMD)和奇异值分解(SVD)相结合的信号特征提取与识别方法.首先对AE信号进行EMD分解,运用互相关系数和方差贡献率筛选出包含主要信息的本征模态函数(IMF)分量;其次对各IMF分量构建的初始特征矩阵进行SVD分解,将得到的奇异值作为表征各损伤信号的特征向量;最后建立Mahalanobis距离判别函数对各损伤信号进行识别分类.五层胶合板损伤的实测数据表明,该方法能够方便地提取出AE信号特征并对其损伤类型进行有效的识别.
Aiming at the non-stationary features of acoustic emission（ AE） signals generated from plywood damage and overlapping of damage features in practice,a method of feature extraction and classification was proposed based on empirical mode decomposition（ EMD） and singular value decomposition（ SVD）. Firstly,the original AE signals were decomposed by EMD,and intrinsic mode function（ IMF） including the main feature information were selected by cross correlation coefficient and variance contribution ratios. Secondly,the initial feature vector matrix constructed of IMFS was decomposed by SVD,and the decomposed singular values served as the characteristic vectors of damage AE signals types. Finally,the Mahalanobis distance criterion function was used to identify all AE signals types generated from plywood damage. The measured data of five-plywood damage show that the method can extract AE signals characteristics easily and identify damage types efficiently.
出处
《应用基础与工程科学学报》
EI
CSCD
北大核心
2014年第6期1238-1247,共10页
Journal of Basic Science and Engineering
基金
江苏省第七批"江苏省六大人才高峰"项目(DZXX-149-148)
江苏高校优势学科建设工程一期项目(苏政办发2011-137-5)
南京林业大学科技创新基金(163070080)
南京林业大学"十五"人才基金(163070505)
关键词
声发射
经验模态分解
奇异值分解
特征提取
MAHALANOBIS距离
acoustic emission
empirical mode decomposition
singular value decomposition
feature extraction
Mahalanobis distance
作者简介
徐锋（1977-），男，在职博士生，讲师．E-mail：xufeng@njfu．com．cn
通信作者：刘云飞（1962-），男，博士，教授．E-mail：lyf@njfu．com．cn
