The Design Of Optical Methane Detection System And Research On Signal Processing Method

Methane is one of the main pollutants in the atmosphere, which cause serious damage to people’s life safety. It is understood that when the concentration of methane in air reaches 10 percent, the human body can feel severe discomfort, and even affect the life health. Therefore, find a convenient, efficient and practical testing method to determine the concentration of methane is very importance. By the classic Beer-Lambert law, the spectral absorption method for the detection of methane has its unique advantages, which can realize the low concentration and high precision measurement of methane. Based on the spectral absorption method as the main line, launched a series of research. The optical path module design, hardware circuit module design and signal processing software three connects in turn, in order to study design, eventually formed a complete set of methane concentration detection system. The following work is studied respectively:(1) Analyze the theory of spectral absorption of methane, depending on the nature of the methane molecule spectrum theory, choose suitable for linear and line width of methane detection, and then according to the determination of spectral absorption band, choose the appropriate light source, and design of the optical path differential detection.(2) Build the overall pattern of methane concentration detection system, it is mainly divided into three modules: optical circuit module, circuit module, a signal processing module. To design of these three modules respectively, the use of photonic crystal fiber and the effective combination of the full prism is designed which can realize gas concentration measurement of double light path air chamber. And the hardware circuit and software programming of received signal processing, improve the accuracy of experimental results.(3) The wavelet denoising and EMD denoising applied to the measurement of results, using different evaluation comparison of their advantages and disadvantages, so as to determine the empirical mode decomposition for denoising effect even better. And de-noised experiments analyzed the data, verify the validity of the data, and then verify the reliability of the system.(4) Support vector machine neural network is applied to implement the methane concentration level forecast, through the experiment verify the reliability of its predictions, implements in don’t need to know accurate concentration, just need to know accurate prediction under a certain concentration range. And design a GUI-based methane concentration levels predict the human-computer interaction interface, to facilitate a more intuitive observation methane concentration level changes, take timely measures to reduce the loss.