Investigation Of Compound Fire Detection System Based On Fuzzy Neural Network

Fire disaster has a catastrophic threat to people's lives and property.However,the traditional method used for fire detection is simply depended on a single sensor and a simple threshold method which is prone to false alarm in actual complex fire environment.Therefore,it is an urgent problem to reduce the false alarm rate and improve the performance of fire detection system.A novel composite fire detection system based on fuzzy neural network was studied in this paper.A composite fire detector was developed based on fire characteristic parameter such as carbon monoxide gas,smoke and temperature,also a multi-point fire detection system was constructed based on RS-485 bus.The carbon monoxide sensor MQ-7,the temperature sensor TC1047,the infrared tube PD243 B and the silicon photo-voltaic cell IR5123 C were employed as the key components of fire detection circuit.And a RS-485 bus circuit was built according to the chip of MAX1348.The shell of the composite fire detector was designed by the CATIA software and implemented by the 3D printing technology.The MCU program was designed to achieve the functions that the sub-computer can control the three points fire signal acquisition,A/D conversion and transmission.Combining fuzzy algorithm with neural network algorithm,a fire detection algorithm was constructed.The user monitor interface was designed by the Labview software to realize display,storage and reproduction of the fire characteristic parameter,the communication self-test and the fire real-time alarm.The novel composite fire detector was used to complete the functional verification experiments,fire response time experiments,anti-interference experiments,distributed simulation detection experiments.The results show that the novel composite fire detector was capable of detecting the wood smoldering fire,the cotton rope smoldering fire,the polyurethane open flame and the ethanol fire.The average response time of the novel composite fire detector is 19.8 second and it is 2.1 second faster than that of the commercially available gulf fire detector in response time.And the novel composite fire detector satisfiesthe national standard which does not exceed 30 second in the fire response time.The 800 times anti-interference experiments including Cigarette fire experiments,lighter fire experiments,heaters experiments and dust experiments show that the fire false alarm rate of the novel composite fire detector is 0.3%.The experimental system consist of three points composite fire detectors can be used for single point and multi-point fire detection.