Optical Fiber Fire Monitoring And Software Design Based On Neural Network Algorithm

With the development of urban infrastructure and the demand for fire safety monitoring,it is of great practical significance to use optical fiber fire monitoring technology to monitor the safety of underground pipe network.In this thesis,based on the research of Raman scattering effect in optical fiber fire monitoring system,a neural network algorithm is proposed to realize the demodulation of optical signal to temperature,which can compensate the attenuation of optical signal and system error according to the actual situation,which has great practical significance for the application of optical fiber sensing monitoring technology.In this paper,the theory,system structure and neural network algorithm of optical fiber fire monitoring are deeply studied,and the monitoring software is designed and completed:(1)Based on the study of distributed optical fiber temperature sensor and neural network theory,a temperature demodulation algorithm based on genetic optimization BP neural network is proposed.(2)This paper studies the optical fiber fire monitoring system,including the design of system hardware structure,the calculation of parameters of pulse laser,data acquisition card and other key components,and completes the development of the monitoring system.(3)The temperature demodulation algorithm of BP neural network is realized by MATLAB.Raman ratio,distance and scale point Raman ratio are used as network input for network training.Genetic algorithm is used to optimize the neural network to realize the temperature demodulation and compensate the unstable factors of the system.(4)According to the requirements of distributed optical fiber fire monitoring,the software of temperature sensing optical fiber fire monitoring system is completed by using Lab VIEW,including data acquisition,temperature demodulation,data storage,human-computer interaction and other functions,so as to realize the real-time fire monitoring of underground pipe network.(5)The experimental test of optical fiber fire monitoring system is carried out,and the neural network algorithm is applied to the system.Without any calibration device,the system error caused by the instability of pulse laser,the change of detector gain and the nonuniformity of sensing optical fiber can be well reduced.