基于贝叶斯网络的双参数火灾探测系统  

Two-parameters fire detection system based on Bayesian network

作　　者：刘全义 朱博 胡林 邓力 梁光华 LIU Quanyi;ZHU Bo;HU Lin;DENG Li;LIANG Guanghua(College of Civil Aviation Safety Engineering,Civil Aviation Flight University of China,Guanghan 618307,China;Hefei Institute for Public Safety Research,Tsinghua University,Hefei 230601,China)

机构地区：[1]中国民用航空飞行学院民航安全工程学院,四川广汉618307 [2]清华大学合肥公共安全研究院,安徽合肥230601

出　　处：《南京工业大学学报（自然科学版）》2022年第6期684-690,共7页Journal of Nanjing Tech University(Natural Science Edition)

基　　金：国家自然科学基金(U2033206,U1933105);四川省科技计划项目(2018GZYZF0069,2020YFG0447);中国民用航空飞行学院基金(J2020110,J2020120,X2019022);中国民用航空飞行学院大学生双创建设项目(S202010624137)。

摘　　要：为解决单一火灾特征参数引发的误报、漏报现象,设计并建立实验平台,以航空煤油燃烧产生的烟气颗粒物质量浓度以及烟气成分浓度为火灾报警参数,建立基于贝叶斯网络模型的火灾探测系统并评估其可行性及准确率。结果表明:该区域环境中可吸入颗粒物(PM_(10))浓度与CO浓度最高值远超过空气中该值的一般含量,表明该区域内污染严重且存在燃烧现象,可作为火灾检测参数;基于贝叶斯网络的双参数火灾探测系统与真实结构相吻合,且真实输出值准确率达到80%,表明该系统可用于火灾探测。可见,利用贝叶斯网络评估火灾探测系统是有效的,为火灾探测提供算法支撑。To solve false alarm and missed alarm caused by single fire characteristic parameter,an experimental platform was designed and established.A fire detection system taking the mass concentration of smoke particle and smoke composition concentration by aviation kerosene as fire alarm parameters was established and its feasibility and accuracy were evaluated based on Bayesian network.Results showed that the peak of particulates(PM_(10))and carbon monoxide(CO)concentrations in the environment of this area exceeded the general amount in the air,indicating that the area was seriously polluted and there was a burning phenomenon,and it was used as fire test parameters.The two-parameter fire detection system based on Bayesian network was consistent with the real structure and the accuracy of the real output value reached 80%,which could be used for fire detection.It can be seen that it is feasible to use Bayesian networks for fire detection,which provides algorithm support for fire detection.

关 键 词：火灾探测 烟气成分浓度 烟气颗粒物质量浓度 贝叶斯网络 

分 类 号：X932[环境科学与工程—安全科学]