粒子群优化神经网络船舶火灾数据融合研究

The study of ship fre data fusion technology of particle swarm optimization neural network

船舶火灾发生后,受到其环境和空间的限制,容易造成严重的财产损失甚至危及人员生命安全。传统的单一型火灾探测器(包括温度、二氧化碳、烟雾等)存在较高的误报率,通过多传感器数据融合技术,可以有效减少误报率,帮助判断火灾发生。笔者采用粒子群优化神经网络的方法,对火灾数据进行模拟训练,判断火灾的发生情况,符合船舶火灾探测系统的要求。并对船用火灾探测器架构设计作简要叙述。

After a ship fire occurs,due to the limitation of its environment and space,it is easy to cause serious property damage and even endanger the safety of personnel.Traditional single-type fre detectors(including temperature,CO,smoke,etc.)have a high false alarm rate.Through multi-sensor data fusion technology,the false alarm rate can be efectively reduced and the fre can be judged.In this paper,the method of particle swarm optimization neural network is used to simulate the fre data to judge the occurrence of fre,which meets the requirements of ship fre detection system.And the architecture design of marine fre detector is briefy described.