矿井极性混合气体红外光谱定量分析研究.pdfVIP

摘 要 对煤矿生产期间和灾变后的井下极性混合气体进行快速准确的分析，能够保障煤矿企 业安全生产，为应急救援提供决策性建议，确保井下工作人员生命健康。对于矿井气体的 分析监测，目前主要采用气相色谱法、传感器法和一般光学法，而上述方法在分析速度、 适用范围、使用寿命和可靠性等方面都有不足。应用红外光谱法对气体进行分析，使用方 便，信噪比高，稳定性好，分析速度快。本文采用傅里叶变换红外光谱法对甲烷、乙烷、 丙烷、异丁烷、正丁烷、乙烯、丙烯、乙炔、六氟化硫、一氧化碳和二氧化碳 11 种煤矿 井下极性气体进行检测，建立基于支持向量机的非线性定量分析模型。通过合理选择光谱 区间，偏最小二乘法降维，粒子群优化算法调整模型参数，使校正模型预测精度基本达到 煤矿气体检测要求，初步实现了基于红外光谱技术的煤矿井下极性混合气体定量分析。 关键词：红外光谱；矿井极性混合气体；支持向量机；定量分析 - I - Abstract Quantitative analysis of the air component of coal mine during production and after the disaster is very useful for guaranting the safety of coal mining enterprises, providing decision-making proposals for emergency management and ensuring life and health of mine workers. At present coal mine gas is monitoried mainly by gas chromatography, transducer method and normal optical method, which all have disadvantages in analysis speed, scope, service life and reliability, etc.. There are many advantages when analyzing the gas by infrared spectroscopy, such as convenience, high signal to noise ratio, good stability and fast response. Eleven different kinds of Polarity gases in the coal mine, including CH , C H , C H , C H , 4 2 6 3 8 4 10 C H , C H , C H , SF , CO and CO , were chosen for quantitative analysis by fourier transform 2 4 3 6 2 2 6 2 infrared spectroscopy. Support vector machine (SVM) nonlinear regression model was established and corrected after appropriate spectrum region choosing, data pre-treating with partial least squares (PLS) and parameters settin