摘要
针对木塑复合材料五种典型的缺陷及损伤机制,选择合适的木塑试样,应用三点弯曲的加载方法采集声发射信号。对主损伤区附近的声发射事件,应用小波变换提取特征参数,确定五类主要损伤机制所对应的声发射信号特征。采用B-P型反向传播神经网络构成的智能化模式分类器,对此五类声发射信号进行识别,获得了满意的效果.
Proper wood-plastic composite samples are chosen for five kinds of typical flaws or damages. Acoustic emission （AE） signals axe collected in the three-point flexural tests. AE events near the main damage section are studied. Wavelet transform is applied to extract characteristic parameters from major AE events. An intelligent pattern classifier with B-P neural network is used in recognition of those five kinds of AE signals successfully.
出处
《应用声学》
CSCD
北大核心
2007年第6期352-356,共5页
Journal of Applied Acoustics
关键词
声发射
小波变换
神经网络
模式识别
Acoustic emission, Wavelet transform, Neural network, Pattern recognition
