Research On Fire Alarm Detector In Large-Space Building Based On Multi-sensor Information Fusion

Fire in the buildings is one of the most dangerous disasters to humans.It can endanger people’s lives and property.With the rapid development of the social economy,a variety of large space buildings have been constructed.These buildings are more complex and easy to catch fire with high density of people and equipment.A greater risk of fire with these special features brings more challenges in the design of fire detection and raises widespread concern.The security protection architecture of large space fire protection is an essential factor.This paper studies the mechanism of fire,the fire process and the by-products of the fire.The overview of the development of the fire detection is presented.What’s more,the classification of fire detectors based on different fire features and some typical fire detectors are introduced in detail.Besides,the fire features of the large space buildings are expounded.Based on the features of the Large-Space Building,this paper proposes a fire signature combination method and the fire optimum combination of features which are optimal for the detection of the Large-Space Building in the multi-sensor fire detection systems.Also,the results of the selected features for fire detection and the optimum combination of features are analyzed.This selection can increase the efficiency of the fire detection and decrease the cost of fire detection.Besides,this paper proposes the module of fire feature distribution determine system,including the distribution of the main fire features.Some fire simulations of the large-space building models based on FDS are conducted.The change rules and distribution of the fire features(smoke,CO and temperature)at different fire stages are analyzed and the optimal positions of the sensors are determined.Finally,BP neural network theory and DS evidence theory are used for multi-sensor information fusion and this secondary method of fusion is a good combination of the merits of the above theory which can improve the accuracy of large space fire detection.Finally,the effectiveness of the proposed method is verified.