Flaw Identification Based on Layered Multi-subnet Neural Networks
摘 要
针对单一神经网络在涡流无损检测中存在识别精度低、训练时间长和识别范围小的缺点,提出了一种适用于实时在线检测的神经网络结构——层次式多子网神经网络。该网络包括一个总网和各层子网,可以将一个复杂的任务分成多个小任务去完成,能快速识别出缺陷有无、走向以及大小。由于每个网络采用改进的径向基神经网络优化隐含节点数,利用小波多尺度边缘检测方法提取输入信号的特征值以简化输入节点数,网络结构得到极大简化。结果表明,层次式多子网神经网络适用于实时在线检测。
Abstract
Pointed to the disadvantages such as low recognizing precision, long training time and limited recognizing range of single neural network in eddy current testing, layered multi-subnet neural network was presented. It was composed by a sumnet and several layered subnets, and could divide a complex task into a series of subtasks, so it could quickly identify whether the defect was existed, and also the defect location and dimension. Because of the improved RBF and wavelet multi-scaling edge detecting were used in each network, the network structure was simplified much. The result showed that layered multi-subnet neural network was suitable to online eddy current testing.
中图分类号 TG115.28
所属栏目 科研成果与学术交流
基金项目 河北省自然科学基金资助项目(602378);河北省教育厅博士基金资助项目(B2001206)
收稿日期 2006/3/6
修改稿日期
网络出版日期
作者单位点击查看
备注刘东辉(1971-),男,博士,主要从事信号处理和无损检测方面的研究。
引用该论文: LIU Dong-hui,SUN Xiao-yun. Flaw Identification Based on Layered Multi-subnet Neural Networks[J]. Nondestructive Testing, 2007, 29(5): 251~254
刘东辉,孙晓云. 基于层次式多子网神经网络的缺陷识别[J]. 无损检测, 2007, 29(5): 251~254
共有人对该论文发表了看法,其中:
人认为该论文很差
人认为该论文较差
人认为该论文一般
人认为该论文较好
人认为该论文很好
参考文献
【1】Kreis T, Jueptner W, Biedermann R. Neural network approach to holographic nondestructive testing[J]. Applied Optics,1995,34(8):1407-1415.
【2】Arkadan A, Chen Y, Subramaniam S, et al. NDT identification of a crack using ANNs with stochastic gradient descent[J]. IEEE Transactions on Magnetics,1995,31(3):1984-1987.
【3】孙晓云,袁 斌,盛剑霓.神经网络方法在涡流无损检测定量分析中的应用[J].西安交通大学学报,2000,34(6):6-10.
【4】孙功星,朱科军,戴长江,等.层次式多子网级联神经网络[J].电子学报,1999,27(8):49-51.
【5】丁 晖,刘君华,申忠如.基于分布式多子网神经网络的多组分混合气体识别[J].仪器仪表学报,2001,22(6):592-594.
【6】Uesaka M, Hakuta K, Miya K, et al. Eddy current testing by flexible microloop magnetic sensor array[J]. IEEE Trans Magne,1998,34(4):2287-2297.
【7】陈德智.涡流无损检测中数值模拟与信号处理的研究[D].西安:西安交通大学,1998.
【2】Arkadan A, Chen Y, Subramaniam S, et al. NDT identification of a crack using ANNs with stochastic gradient descent[J]. IEEE Transactions on Magnetics,1995,31(3):1984-1987.
【3】孙晓云,袁 斌,盛剑霓.神经网络方法在涡流无损检测定量分析中的应用[J].西安交通大学学报,2000,34(6):6-10.
【4】孙功星,朱科军,戴长江,等.层次式多子网级联神经网络[J].电子学报,1999,27(8):49-51.
【5】丁 晖,刘君华,申忠如.基于分布式多子网神经网络的多组分混合气体识别[J].仪器仪表学报,2001,22(6):592-594.
【6】Uesaka M, Hakuta K, Miya K, et al. Eddy current testing by flexible microloop magnetic sensor array[J]. IEEE Trans Magne,1998,34(4):2287-2297.
【7】陈德智.涡流无损检测中数值模拟与信号处理的研究[D].西安:西安交通大学,1998.
相关信息