论文部分内容阅读
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 fact that the non-stationary acoustic damage (AE) signals of plywood damage and the characteristics of damage categories overlap each other, a signal feature extraction and recognition method based on empirical mode decomposition (EMD) and singular value decomposition (SVD) is proposed. Firstly, the AE signal is decomposed by EMD, and the IMFs containing the main information are screened by using the cross-correlation coefficient and the variance contribution rate. Secondly, SVD is performed on the initial eigenmatrix constructed by each IMF component. The singularities As the eigenvector to characterize each damage signal.Finally, Mahalanobis distance discriminant function is established to identify each damage signal.The measured data of five plywood damage shows that this method can extract the characteristics of AE signal and make effective its damage type Recognition.