Research On Room Fire Detection System Based On Information Fusion

Fire is a sudden and uncontrollable violent combustion reaction.It may affect people’s normal life and destroy their homes.The uncertainty caused by fire will lead to the use of a single detection criterion for fire identification,which can not achieve the expected accuracy.In view of these defects,this paper adopts the test method based on information fusion and establish test system to make more accurate discrimination on fire.By consulting data and collecting information,the conditions of fire occurrence and characteristics of combustion process are analyzed,and the development process of fire is divided.Temperature value,CO mass fraction and smoke mass fraction are selected as the criteria for fire identification,and three kinds of fire are predicted: no fire,smoldering fire and open fire.The fire simulation software pyrosim is used to establish the room model,and divides it into grids.Based on the three conservation equations,the smoldering and open fire tests of polyurethane combustion are carried out.The test data of fire criterion are obtained as the training input of detection algorithm.The mathematical principle of support vector machine(SVM)is studied,and three kinds of meta heuristic optimization algorithms,genetic algorithm(GA),particle swarm optimization(PSO)and equilibrium optimizer(EO),are described.GA-SVM,PSO-SVM and EO-SVM are designed,firstly,the mata heuristic algorithm is used to optimize the parameters of SVM,and then the simulation training is carried out on the experimental data,so that the algorithm can complete the classification of no fire,smoldering fire and open fire.In the design of the whole system,stm32f4 is used as the core control chip to build the whole system.Three kinds of sensors are used to collect the temperature value,CO mass fraction and smoke mass fraction respectively.The TFT color screen is used to show the fire,and NRF24L01 is used to realize the communication between master and slave controllers.The control software is programmed,the combination of software and hardware is realized,and the comprehensive test of the system is completed.The experimental results show that EO-SVM has the best prediction accuracy and the fastest convergence speed,and the fire parameter data fusion method based on EO-SVM is established.Through the communication of master-slave controller,the real time transmission of fire information is realized.The research results can be used for reference to improve the accuracy and timeliness of fire identification,and also play a positive role in improving the speed of disaster rescue.