Research On Combustible Gas Detection System Based On RBF Neual Network

Combustible gas has been used more and more widely with the development of nationaleconomy. But because of the characteristics of inflammable, explosive, the concentration ofcombustible gas in the air reaches the explosion limit, encountered fire will be burning evenhappen the serious accident such as explosion,it is a huge threat to the safety of people’s lifeand property. However, due to some reasons, our country has not formed a complete systemfor efficient detection of combustible gas. Especially in the petrochemical industry, falsepositives, false security problems that often occur. Therefore, under this background,thispaper put forward scheme that aimed at the demand of combustible gas detection technology.This paper aims at the shortcomings of the existing sensor in the field of gas detection,this paper puts forward a method by using multi gas sensor array to collect gas physicalinformation,and design the gas detection system. According to the system structure this paperdesign the hardware circuit of the corresponding mainly consists of gas information collectionmodule, microcontroller module, wireless communication module and power module.For the characteristics of huge amount of data in the field of gas detection,put forward toachieve the gas sensor array to collect the information fusion by using RBF neural network.Atthe same time, aiming at the shortage of traditional wired network, I compare and contrast theadvantages and disadvantages of three kinds of wireless communication technology,combining the advantages of ZigBee Technology, put forward to build the combustible gasdetection network by using ZigBee wireless sensor network.Through the simulation experiments,this paper demonstrated the feasibility of theanalysis of selected RBF neural network for single gas qualitative identification, quantitativeanalysis and mixed gas quantitative, at the same time prove that the method used in this paperhas a good application value.