Research And Application Of Fuzzy Neural Network In Fire Detection Of Shopping Malls

In order to meet people's daily needs,more and more large-scale buildings have been built,and some safety hazards have emerged,especially fire safety.At present,due to the special nature of the environment and the stability of the steps caused by human factors,the current fire detection system is also in an unstable state.Constant false alarms will not only bring huge losses to production and life,but also numb the nerves and make people Loss of trust in the fire detection system,resulting in greater losses.Among various fire prevention methods,fire detection technology is particularly prominent,and it is an easy and simple method for preventing and reducing fire occurrence.Because the early traditional fire detection technology is limited to the monitoring of single parameters,it is prone to false positives and false negatives.Therefore,based on the traditional fire detection technology,I focus on the analysis of intelligent fire detection system based on fuzzy neural network system.And slightly improved the fire detection algorithm,and applied to large-scale shopping mall fire detection technology.The main work is as follows:(1)Study the principle of fire detectors with single parameters in the early stage,and the characteristics of fire occurrence,and summarize the shortcomings of traditional fire detection technology based on its detection principle,and propose information fusion technology to detect the temperature and smoke concentration at the scene of the fire.The CO content is used together as an input to the system.(2)Fuzzy logic has strong processing ability on nonlinear problems.Using this advantage,fuzzy logic reasoning is used to fuzzify,fuzzy reason and defuzzify the three feature quantities input by the system.64 fuzzy rules,the simulation process of fuzzy reasoning can be realized by MATLAB,and its output is an important auxiliary criterion.(3)The neural network has super self-learning ability,can imitate the human brain to learn and summarize the law.According to this great advantage,the neural network model is used to simulate and train the sample data.The most common and typical BP neural network and RBF neural network are compared for comparison.The optimal neural network is selected by training the same sample data.(4)Fuzzy logic reasoning and neural network have their own advantages and disadvantages.In order to make full use of their respective advantages and complement each other's defects,fuzzy reasoning and neural network are organically integrated,and a decision layer is proposed.A new criterion,comprehensive fuzzy logic reasoning and the results of the neural network system,to make a reasonable scientific judgment,in order to improve the accuracy of the fuzzy neural network system for fire detection.(5)Apply the fuzzy neural network system with reasonable design to the fire detection system of large shopping malls,and compare the judgment results with the known results to verify the authenticity and reliability of the fuzzy neural network system.Figure [24] table [9] reference[71]...