文献类型：期刊文献

中文题名：声发射神经网络模式识别

英文题名：The Neural Network of Acoustic Emission for Pattern Recognition

作者：刘国光[1];程青蟾[1];李燮里[1];张月兰[1]

机构：[1]华东理工大学,上海200237

年份：2003

卷号：24

期号：Z2

起止页码：406

中文期刊名：仪器仪表学报

外文期刊名：Chinese Journal of Scientific Instrument

收录：CSTPCD;;Scopus;北大核心:【北大核心2000】；CSCD:【CSCD2011_2012】；

语种：中文

中文关键词：声发射;神经网络;质量检测

外文关键词：Acoustic emission;;Neural network;;Quality evaluation

摘要：利用工件在加载与卸载时所产生的声发射信号,作为神经网络模式识别的输入特征,并对输出量进行模糊化隶属度计算,用BP学习法对连接权重系数进行优化,成功地实现了对工件样本组的训练和分类。
The acoustic emission signals produced when iron workpiece is loaded and unloaded are used as inputcharacteristics of pattern recognition for neural network. Meanwhile the fuzzy subordinate degree is calculatedfor network output. The connection weight factors are optimized through BP learning. The training and classifi-cation for workpiece sample group are successfully fulfilled.