摘要
一种时频数据处理方法被应用于复合材料损伤检测。根据小波变换的框架重构理论和小波变换时频相空间理论 ,提取了信号的时频域特征 ,通过比较原信号的时频空间和小波变换相空间的相同部分 ,得到了能反映原信号同样时频特征的小波级数展开项和 ,应用 Gram- schmidt正交化方法 ,按照一定准则 ,对所得到的小波级数展开项的线性组合进行正交化处理 ,用代数数值方法 ,从已知采样数据分布集合得到了对应于曲线本身时频特征。经过处理得到的曲线 ,能够很好的逼近原来采样曲线 ,故而能保证和原采样曲线有相同的时间频率性质 ,这样的复合材料损伤缺陷判别方法结果较好。
A new data processing method is proposed for the damage detection of anisotropic composite materi- als.The concept of the wavelet transform and its phase space are discussed.The expansion of the wavelet se- ries,determined by the time- frequency wavelet transform of wavelet transform ,has the same local time- fre- quency properties as the original signal.We study the signal reconstruction by Gram - schmidt orthogonalization procedure to reduce redundancy.According to frame of wavelet and the theory of the phase space,itis interest- ed in presenting the features of the signals in the field of the time- frequency.We obtain the terms of the expan- sion of the wavelet series which have the same property as the original signal by comparing the time- frequency space of the signal with the phase space of the wavelet transform. Because the number of possible combina- tions,which are used to construct the wavelet network,is often too large for a reasonable search,the conver- gence is too slowly in the training process,even so.We apply Gram- schmidtorthogonalization for removing the redundant wavelets in series by the characters of the sample data,and through linear equation obtain the values of features which are thought as the weights of wavelet network.The reconstruction of curves obtained by the theories can well approximate the original curves.The experiments on the damage detection of composites with health,various size cracks,and delamination are conducted to demonstrate the feasibility of the proposed method.
出处
《仪器仪表学报》
EI
CAS
CSCD
北大核心
2000年第5期449-454,463,共7页
Chinese Journal of Scientific Instrument
基金
国家自然科学基金资助项目
关键词
小波变换
损伤检测
复合材料
时频特性
采样信号
Wavelet transform Phase space of time- frequency Gram- schmidt orthogonalization Reconstruc- tion of curves Damage detection Composites
