基于概率神经网络的光纤智能结构承载定位

广西师范大学学报（自然科学版） ›› 2011, Vol. 29 ›› Issue (2): 223-226.

基于概率神经网络的光纤智能结构承载定位

李鹏

华东交通大学机电工程学院,江西南昌330013

收稿日期:2011-05-10 发布日期:2018-11-19
通讯作者: 李鹏(1976—),男,江西南昌人,华东交通大学讲师,博士。E-mail:ecjtulipeng@126.com
基金资助:
国家自然科学基金资助项目(61063037);华东交通大学校立课题(09102005)

Load Localization of Optical Fiber Smart Structures Based on Probabilistic Neural Networks

LI Peng

School of Mechanical and Electronical Engineering,East China Jiaotong University,Nanchang Jiangxi 330013,China

Received:2011-05-10 Published:2018-11-19

摘要/Abstract

摘要： 本文提出一种用于识别光纤智能结构承载位置的方法,研究构建光纤传感网络,设计相应的监测系统,获得用于表征结构承载信息的特性向量。同时,运用神经网络方法构造了四层结构的概率神经网络,并以特征向量作为样本,对结构的承载定位进行研究,重点分析训练样本数和平滑系数对定位效果的影响。实验结果表明:采用训练样本数为30,平滑系数为0.25的概率神经网络对承载位置的定位正确率可达93.2%,研究最终给出用于承载定位的概率神经网络模型。

关键词: 光纤智能结构, 承载定位, 概率神经网络, 平滑系数

Abstract: In this paper,the load location of optical fiber smart structures is identified based on probabilistic neural networks.The optical fiber network and the monitoring circuit were designed to acquire structural load information as feature vectors.The probabilistic neural networkstructure is brought forward by employing the method of neural network.The feature vectors as the sample is presented to the probabilistic neural network.Theeffect of smoothing parameter and number of training samples on the localization performance of the probabilistic neural network is studied.The result shows that good selection of the number of training samples (30)and smoothing parameter (0.25) boost the load localization accuracy (93.2%).The localization model of probabilistic neural was presented finally.

Key words: optical fiber smart structures, load localization, probabilistic neural networks, smoothing parameter

中图分类号:

TP391.4

李鹏. 基于概率神经网络的光纤智能结构承载定位[J]. 广西师范大学学报（自然科学版）, 2011, 29(2): 223-226.

LI Peng. Load Localization of Optical Fiber Smart Structures Based on Probabilistic Neural Networks[J]. Journal of Guangxi Normal University(Natural Science Edition), 2011, 29(2): 223-226.

参考文献

[1] 王彦,梁大开,周兵.强度解调的光纤光栅智能结构动态监测技术[J].光电子·激光,2008,19(3):361-364.
[2] CAPEZZUTO F,CIAMPA F,CAROTENUTO G,et al.A smart multifunctional polymer nanocomposites layer for the estimation of low velocity impact damage in composite structures[J].Composite Structures,2010,92(8):1913-1919.
[3] SPECHT D F.Probabilistic neural networks[J].Neural Networks,1990,3(1):109-118.
[4] ANCONA F,COLLA A M,ROVETTA S,et al.Implementing probabilistic neural networks[J].Neural Comput and Applic,1997,5(3):152-159.
[5] 李苏梅,韩国强,周咏梅.一种基于BP神经网络的颜色空间量化方案[J].广西师范大学学报:自然科学版,2008,26(1):232-235.
[6] PARZEN E.On estimation of a probability density function and mode[J].Annals of Mathematical Statistics,1962,33(3):1065-1076.

Metrics

Viewed

Full text

Abstract

Cited

Shared

Discussed