基于模糊神经网络的舰船火灾早期探测报警系统设计  

Design of fire early detection and alarm system based on Fuzzy Neural Network

作　　者：郑珊珊 李田烽 蔡欣欣 ZHENG Shan-shan;LI Tian-feng;CAI Xin-xin(The 726 Research Institute of CSSC,Shanghai 201108,China)

机构地区：[1]中国船舶集团有限公司第七二六研究所,上海201108

出　　处：《舰船科学技术》2021年第12期161-165,共5页Ship Science and Technology

摘　　要：结合当前舰船火灾探测报警系统的现状和对于火灾早期探测的需求,提出一种将主动吸气式空气采样感烟探测器和阴燃类电气火灾燃烧产物特征量CO及传统温度信号通过模糊神经网络和遗传算法进行融合判定的系统设计方法,详细描述系统的构成,通过数据采集、模糊推理、火灾判定的软件实现过程。同时通过仿真计算表明该系统对于提高火灾探测报警的关键性能指标"响应时间"和"误报率"的有效性。该系统设计方法为电气类阴燃火的早期探测提供有效的技术解决途径,同时提高了系统抗误报警能力。Combined with the current status of ship fire detection and alarm system and the demand for early fire detection, this paper presents a method to fuse the active aspirating air sampling smoke detector, the characteristic quantity CO of smoldering electrical fire products and traditional temperature signal through fuzzy neural network and genetic algorithm,The article describes the system structure in detail, through data acquisition,fuzzy reasoning, fire descision software implementation process. At the same time, the simulation results show that the system is effective to improve the detection and alarm performance. So as to achieve the goal of early fire detection and alarm, and improve the anti-false alarm ability of the system.

关 键 词：舰船火灾早期探测 空气采样感烟探测器 CO探测器 模糊神经网络 遗传算法 抗误报警 

分 类 号：U664.88[交通运输工程—船舶及航道工程]