基于改进BP网络的管道安全监测技术研究

详细信息 本馆镜像全文    |  推荐本文 | | 获取馆网全文

英文篇名：Research on pipeline secure monitor based on improved BP neural networks 中文刊名：仪器仪表学报 英文刊名：Chinese Journal of Scientific Instrument 作者：孙洁娣 ; 温江涛 ; 靳世久 英文作者：Sun Jiedi Wen Jiangtao Jin Shijiu (State Key Laboratory on Precise Measuring and Testing Technique and Instrument ; Tianjin University ; Tianjin 300072 ; China) (Department of Electronics and Communication ; Yanshan University ; Qinhuangdao 066004 ; China) (Electrical Engineering Institute ; Yanshan University ; Qinhuangdao 066004 ; China) 中文关键词：管道安全监测 ; 地震动信号 ; 数据融合 ; 神经网络 ; 目标识别 英文关键词：pipeline secure monitor seismic detection data fusion neural network target recognition 出版日期：2006-12-30 机构：天津大学精密测试技术及仪器国家重点实验室,燕山大学电气工程学院,天津大学精密测试技术及仪器国家重点实验室 天津300072燕山大学信息学院电子与通信工程系秦皇岛066004,秦皇岛066004,天津300072 年：2006 期：S2 出版单位：仪器仪表学报

摘要

管道安全监测技术能够对管道安全构成威胁的行为进行早期监测、定位和预警,对管道周围目标进行正确的分类识别是其中的关键。采用地震信号传感器对管道周围的人、车辆等目标发出的震动信号进行采集,通过对信号特征的分析,实现震动信号的特征提取和关联,在特征级实现目标特性融合,并提出一种改进的神经网络方法进行融合及分类识别,在对实际目标信号的处理中得到较好的效果。
Pipeline secure monitor can finish early monitor, location and advance alarm for the threatening actions to pipeline security. The key technology is to accurately classify and recognize the targets around the pipeline. Because moving targets over ground generate a succession of impacts, the soil disturbances propagate away from the source as seismic waves. The data characters were investigated and processed by means of multi-sensors data fusion, feature extraction, data association and etc. An improved BP algorithm is presented to fuse, classify and recognize the targets generating the seismic signals. The experiments indicate this method is effective to recognize these different targets.

引文

[1]王占山,张化光,冯健,等．长距离流体输送管道泄漏与定位技术的现状与展望[J]．化工自动化及仪表, 2003．30(5):5-10．
    [2]孙涛,张宏健．目标识别中的信息融合技术[J]．自动化仪表,2001,22(2):1-4．
    [3]聂伟荣,朱继南．运动目标地震动信号的时频特征分析[J]．南京理工大学学报,2002,26(5):478- 481,485．
    [4] Nie Weirong,Zhu Jinan,Guo Yajun,et al．Seismic Signals analysis And Identification[J]．电子科技大学学报,2003,32(1):26-30．