Research And Application Of Tunnel Fire Monitoring Method Based On Data Fusion

With the rapid development of economic society and urban construction, the transportation plays a more and more important role in the traffic, and the effect of tunnel can not be ignored. The urban tunnel is a part of city road. It is built according to the urban topographic features and current situations of traffic, in order to alleviate the traffic congestion. In recent years, a lot of city tunnels have been projected and built in Wuhan, such as Wuhan Yangtze River Tunnel、Shuiguohu tunnel, Bayi Road tunnel, Zhongshan Road tunnel, Huanglongshan tunnel,Donghu tunnel. The usage of these tunnels alleviates the traffic pressure in Wuhan to some degree, improves the traffic efficiency significantly and promotes the economic development. This thesis, based on the Bayi Road tunnel project,introduces the research status of intelligent monitoring system of city tunnel, summarizes the advantages of tunnel monitoring system abroad, analyzes the implementation scheme of intelligent monitoring system of city tunnel which fits the specific situation in China. This thesis classifies the common fire detection sensors in tunnel and studies the output information form of multiple kinds of sensors in order to get a more reliable fire condition estimation.This thesis introduces the meaning of data fusion firstly, and analyze some primary data fusion algorithm(such as Bayes and D-S), then introduces the sensors’ output format and classifies them by uncertainty information in the fusion of decision-making level and the processing methods, including uncertainty information processing method and integration method. This thesis introduces the problems of multi-sensor fusion method in tunnel monitoring system, puts forward specific implementation integration schemes, and applies the D-S theory and BP neural network in data fusion scheme, and runs simulation analysis by actual data fusion algorithm. It is found that the theory can effectively eliminate the abnormal data from the whole data, and make a significant improvement in the veracity of linkage for fire monitoring system.In the end, this thesis realizes the urban tunnel intelligent monitoring system by four layer structures, and uses the D-S theory in actual project and designs the implementation of fire alarming system for urban tunnel monitoring particularly.