光学式甲烷检测系统设计及其信号处理方法研究

发布时间：2018-02-04 20:55

本文关键词： 甲烷光谱吸收法经验模态分解支持向量机光子晶体光纤　出处：《燕山大学》2015年硕士论文　论文类型：学位论文

【摘要】：甲烷是大气的主要污染物之一,严重危害人们的生命安全,据悉,当空气中甲烷的浓度达到百分之十时,人体就会产生严重不适,甚至影响生命健康。因此,找到一种便捷,高效且实用的检测甲烷浓度的方法极为重要,由经典的郎伯-比尔定律可知,光谱吸收法对于甲烷的检测具有其独特的优势,能够实现甲烷的低浓度、高精度测量。本文以光谱吸收法为主线,展开系列研究,将光路模块设计、硬件电路模块设计、软件信号处理三者相衔接,依次进行研究设计,最终构成了一套完整的甲烷浓度检测系统。分别对以下工作进行研究:(1)对甲烷光谱吸收理论进行分析,根据甲烷分子的性质和光谱理论,选择适合甲烷检测的线性和线宽,然后根据光谱吸收波段的确定,选择合适的光源,并进行差分式检测的光路设计。(2)搭建整体模式的甲烷浓度检测系统,将它主要分为三个模块:光路模块、电路模块、信号处理模块。分别对三个模块进行设计,利用光子晶体光纤和全反棱镜的有效结合,设计出可实现气体双光路浓度检测的气室,并对接收信号进行硬件电路和软件编程的处理,提高实验结果的精确性。(3)将小波去噪和经验模态分解去噪分别应用到测量的实验结果中,选取不同的评价指标对比它们的优劣性,从而确定经验模态分解的去噪效果更优。并且将去噪后的数据进行多次实验分析,验证了数据的有效性,进而验证了系统的可靠性。(4)应用支持向量机神经网络实现甲烷浓度的等级预测,通过实验验证其预测的可靠性,实现了在不需要知道精确浓度,只需知道一定浓度范围下的准确预测,并设计出基于GUI的甲烷浓度等级预测的人机交互界面,便于更加直观的观测甲烷浓度等级变化,及时采取措施,降低损失。
[Abstract]:Methane is one of the major pollutants in the atmosphere, seriously endangering the safety of people's lives. It is reported that when the concentration of methane in the air reaches 10:00, the human body will have serious discomfort and even affect the health of life. It is very important to find a convenient, efficient and practical method to detect methane concentration. According to the classical Langbo-Beer law, the spectral absorption method has its unique advantages for methane detection. In this paper, spectral absorption method is the main line, a series of research will be carried out, optical circuit module design, hardware circuit module design, software signal processing. A complete methane concentration detection system was designed in turn. The following work was studied: 1) the theory of methane spectral absorption was analyzed. According to the properties of methane molecule and spectrum theory, the linear and linewidth suitable for methane detection are selected, and then the suitable light source is selected according to the spectrum absorption band. And the differential detection of the optical path design. 2) build the overall mode of methane concentration detection system, it is mainly divided into three modules: optical circuit module, circuit module. Signal processing module. The three modules are designed respectively. Using the effective combination of photonic crystal fiber and full inverse prism, the gas chamber can be designed to detect the concentration of gas double optical path. In order to improve the accuracy of the experimental results, wavelet de-noising and empirical mode decomposition de-noising are applied to the experimental results respectively. Different evaluation indexes are selected to compare their advantages and disadvantages, so as to determine that empirical mode decomposition is more effective in de-noising, and the de-noised data are analyzed many times to verify the validity of the data. Furthermore, the reliability of the system is verified. 4) the support vector machine neural network is used to realize the prediction of methane concentration, and the reliability of the prediction is verified by experiments, and the accurate concentration is realized without knowing the exact concentration. We only need to know the accurate prediction under a certain concentration range, and design the man-machine interface of methane concentration grade prediction based on GUI, which is convenient to observe methane concentration grade change more intuitively and take measures in time. Reduce losses.
【学位授予单位】：燕山大学
【学位级别】：硕士
【学位授予年份】：2015
【分类号】：X84;TP274

【参考文献】