首页 | 本学科首页   官方微博 | 高级检索  
     

基于LMBP算法的罐底腐蚀声发射信号模式识别
引用本文:邢菲菲,李一博,靳世久.基于LMBP算法的罐底腐蚀声发射信号模式识别[J].计算机测量与控制,2008,16(12):1945-1947.
作者姓名:邢菲菲  李一博  靳世久
作者单位:天津大学精密仪器与光电子工程学院,天津,300072
基金项目:国家自然科学基金  
摘    要:在建立储罐罐底腐蚀实验平台的基础上,研究了一种基于LM(Levenberg-Marquardt)BP算法的罐底腐蚀信号模式识别方法;选取上升时间、计数、能量、持续时间、幅度这5个声发射信号特征参数作为BP神经网络的输入构建区分腐蚀信号和其他两类声发射信号的模式识别系统;由传统的BP算法与LMBP算法的对比分析比较得到:LMBP算法解决了传统BP算法收敛速度慢,容易陷入局部极小点的问题;实验结果表明,LMBP算法应用于储罐罐底声发射腐蚀信号的模式识别,效果良好。

关 键 词:罐底腐蚀  神经网络  LMBP算法  模式识别

Pattern Recognition Analysis of Tank Bottom Corrosion Acoustic Emission Signal Based on LMBP Arithmetic
Xing Feifei,Li Yibo,Jin Shijiu.Pattern Recognition Analysis of Tank Bottom Corrosion Acoustic Emission Signal Based on LMBP Arithmetic[J].Computer Measurement & Control,2008,16(12):1945-1947.
Authors:Xing Feifei  Li Yibo  Jin Shijiu
Affiliation:Xing Feifei Li Yibo Jin Shijiu College of Precision Instruments and Optic-electronics Engineering,Tianjin University,Tianjin 300072
Abstract:Investigated a pattern recognition analysis of tank bottom corrosion acoustic emission signal based on LMBP arithmetic by simulating tank bottom corrosion progress under the laboratory conditions.Choose Rising-time,Counts,Energy,Lasting-time and Amplitude as the inputs of the BP neural network to establish a pattern recognition system which can distinguish the corrosion acoustic emission signal with another two differ- ent signals.Contrastive experiments proved that LMBP arithmetic solved the inherent disadvantage of traditional BP arithmetic and it is an effective method applying to the pattern recognition analysis of tank bottom corrosion acoustic emission signal.
Keywords:tank bottom eorrosion  neural network  LMBP arithmetic  pattern recognition
本文献已被 CNKI 维普 万方数据 等数据库收录!
设为首页 | 免责声明 | 关于勤云 | 加入收藏

Copyright©北京勤云科技发展有限公司  京ICP备09084417号