首页 | 本学科首页   官方微博 | 高级检索  
     检索      

基于MSA的智能火灾报警系统
引用本文:贾年,罗晓晖,成华友.基于MSA的智能火灾报警系统[J].西南交通大学学报,2007,42(4):452-455.
作者姓名:贾年  罗晓晖  成华友
作者单位:1. 西华大学数学与计算机学院,四川,成都,610039
2. 成都威斯特消防机械有限公司,四川,成都,611743
基金项目:四川省教育厅重点资助项目(2006A138);四川省教育厅自然基金预研项目(03226181)
摘    要:针对传统的采用单一火灾传感器的自动消防系统易产生误报警、漏报警的问题,提出基于MSA(mu lti-step-alarm)的智能火灾报警系统.该系统采用基于BP神经网络的多传感器数据融合技术融合各种火灾传感器的信号,以减少报警时间、误报警和漏报警.提出了相关判定规则,以确认火灾报警信号,克服BP算法容易陷入误差局部最小、单纯使用训练值会产生振荡的缺陷.实验结果表明,与3种单一传感器的平均值比较,基于MSA的智能火灾报警系统的误报率和漏报率分别减少了80%和92%,报警时间缩短了50%.

关 键 词:火灾探测  报警  系统  MSA  多传感器数据融合  BP神经网络
文章编号:0258-2724(2007)04-0452-04
修稿时间:2007-03-23

Intelligent Fire Alarm System Based on MSA System
JIA Nian,LUO Xiaohui,CHENG Huayou.Intelligent Fire Alarm System Based on MSA System[J].Journal of Southwest Jiaotong University,2007,42(4):452-455.
Authors:JIA Nian  LUO Xiaohui  CHENG Huayou
Institution:1. School of Mathematics and Computer Eng. , Xihua University, Chengdu 610039, China; 2. Chengdu Weisite Fire Protection Co., Ltd., Chengdu 611743, China
Abstract:An automatic fire alarm system based on MSA(multi-step-alarm) was proposed to overcome the shortcomings of a conventional fire alarm system composed of single type sensors.The proposed system integrates the signals from all fire alarm sensor by a multi-sensor data fusion system that is based on an BP neural network to reduce the possibility of failure for correct alarming.Criteria were presented to further confirm the alarm signals and to overcome the local optimum and oscillation in output of the BP neural network based solely on training.Experimental results show that,compared with the averaged results of 3 single type fire alarm sensors,the proposed system reduces false alarm by 80%,failure for alarming by 92%,and alarm time by 50%.
Keywords:fire  detection  alarm  MSA  multi-sensor data fusion  BP neural network
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