摘要
通过自行设计的超声波检测系统对模拟分层试样进行实验,提取出缺陷信号,并对缺陷信号进行频域分析,在此基础上进行了基于小波包变换的缺陷信号特征提取和BP神经网络识别。通过"能量—缺陷"的信号特征提取方法,比较不同分层位置之间的信号特征,得到一些有用的数据和信息,为复合材料的质量评估提供了重要的参考依据。
Adopt self designed ultrasonic testing system to test the simulation delamination sample, extract defects signal, and make frequency spectral analysis on defects signal. On that basis, it extracted defects signal characteristic based on wavelet packet transform and identified the defects by means of the BP neural network. Through ＂energy-defect＂ method of signal characteristic extraction, it compared signal characteristic among different delamination parts, gets some useful data and information, provided a consulting evidence for the composite material quality estimating.
出处
《兵工自动化》
2009年第11期56-58,68,共4页
Ordnance Industry Automation
关键词
复合材料
超声波检测
分层
小波包变换
BP神经网络
Composite
Ultrasonic testing
Delamination
Wavelet packet transform
BP neural network
