基于激光诱导击穿光谱的燃煤挥发分定量分析

摘要

21 coal samples with known volatile components were used for laser-induced breakdown spectroscopy (LIBS)ex-periment.According to structure characteristics of coal quality,wave band strength of characteristic spectral lines of coher-ent elements was selected as input for regression analysis and multiple regression calibration method and partial least squares (PLS)was respectively introduced to realize quantitative analysis on volatile component in fired coal.By comparing calibra-tion results of the two methods,conclusion was drawn as that indexes including model coefficients,average relative error of predictive samples,root-mean-square error of predictive samples and so on from PLS algorithm were superior to those from multiple regression calibration method which might be more suitable for quantitative analysis on volatile component based on LIBS and obtain higher analysis precision.%利用21个已知挥发分的煤样样品，进行激光诱导击穿光谱实验，根据煤质的结构特点，选择相关元素的特征谱线波段强度，作为回归分析的输入量，分别引入多元回归定标法和偏最小二乘法，实现燃煤中挥发分的定量分析。通过比较两种方法的定标结果得出结论：偏最小二乘法算法下模型系数、预测样品平均相对误差、预测样品均方根误差等指标皆优于多元回归定标法，更适用于激光诱导击穿光谱定量分析挥发分中，并且能够得到较高的分析精度。