摘要
长大隧道火灾一旦发生,往往造成严重后果。在采集火灾参数信息时,传统的火灾预警是基于火灾探测器,通过阈值比较法对火灾的发生进行判断和识别。由于火灾发生具有极强的随机性,这样的预警技术往往存在适应力、决策力和识别力弱的缺陷。针对这些问题,通过比较分析现有的隧道火灾预警技术,引入了基于温度的多传感器信息融合的火灾智能预警方法,这种方法根据信息融合的信息、特征、决策三大层次,通过模块化的算法识别融合了多层次信息,并基于遗传BP神经网络算法优化模糊控制模块快速响应,弥补了目前预警算法的不足。
Once the fire occurs in long tunnels,it often causes serious consequences.When collecting the fire parameter information,traditional fire warning methods are depended on the fire detectors,where the threshold comparison method is widely adopted for judgement and recognition.Due to the strong randomness of fire occurrence,such warning method often has the defects of weak adaptability,decision-making and identification.After comparing and analyzing the existing tunnel fire early warning methods,a new one is introduced based on temperature multi-sensor information fusion.According to the three levels(information,characteristic and decision)of information fusion,this method integrates multi-level information through modular algorithm identification.It optimizes the fast response of the fuzzy control module based on the genetic BP neural network algorithm,which offsets the current deficiency in traditional techniques.
作者
姚宇
方忠强
田野
YAO Yu;FANG Zhongqiang;TIAN Ye(China Design Group Co.Ltd.,Nanjing 210014;Department of Geotechnical Engineering,Tongji University,Shanghai 200092)
出处
《现代隧道技术》
CSCD
北大核心
2022年第S01期676-683,共8页
Modern Tunnelling Technology
基金
江苏省交通科学研究计划项目(长大水下隧道智能防灾救援关键技术研究)