主管：中华人民共和国应急管理部
主办：应急管理部天津消防研究所
ISSN 1009-0029  CN 12-1311/TU
消防科学与技术

消防科学与技术 ›› 2020, Vol. 39 ›› Issue (11): 1490-1494.

• • 上一篇    下一篇

基于循环神经网络的飞机货舱火灾快速识别算法

何永勃，李明伟   

  1. 中国民航大学电子信息与自动化学院，天津300300
  • 出版日期:2020-11-15 发布日期:2020-11-15
  • 通讯作者: 李明伟（1994-），男，辽宁鞍山人，中国民航大学电子信息与自动化学院硕士研究生。
  • 作者简介:何永勃（1971-），男，陕西蒲城人，中国民航大学电子信息与自动化学院副教授，博士,主要从事航空检测技术及智能化仪表方面的研究，天津市东丽区中国民航大学北院，300300。
  • 基金资助:
    民航科技项目“大型客机座舱空气环境适航审定技术研究”(MHRD20150220)

A fast classification algorithm of aircraft cargo compartment fire based on recurrent neural network

HE Yong-bo, LI ming-wei   

  1. School of Electronic Information and Automation, Civil Aviation University of China, Tianjin 300300, China
  • Online:2020-11-15 Published:2020-11-15

摘要：

针对飞机火灾探测系统对不同火源和干扰源类型的快速准确识别问题，提出一种基于长短期记忆的循环神经网络火灾探测算法，可考虑火灾信号的时间动态信息。通过对不同的真火源和干扰源进行试验，将复合式火灾探测器测取的双波长、CO和温度等信号，按照时间连接成特征序列来训练网络分类模型。通过数据集时间序列长度的不同选取方法，模拟燃烧过程的剧烈程度，验证算法的鲁棒性、快速性和准确性。结果表明，该网络在真火源和干扰源快速识别问题上取得了良好的效果。

关键词: 火灾探测, 循环神经网络, 长短期记忆, 快速识别

Abstract:

Against the problem of rapidly and accurately classifying different fire sources and interference sources by aircraft fire detection system, a recurrent neural network fire detection algorithm based on long short- term memory was proposed,which can consider the vital time dynamic information in the fire signal. Through the experiment of different fire and interference sources, and based on the signals of dual wavelength, CO and temperature measured by the composite fire detector, the network classification model was trained by connecting them into a characteristic sequence according to time. Through the different selection methods of length of time series data sets, the combustion intensity was simulated, and the robustness, rapidity and accuracy of the algorithm were validated. The results showed that the network is effective in the problem of fast classification of true ignition source and interference source.

Key words:

fire detection, recurrent neural network, long short - term memory, fast classification