摘要
采用逐步多重线性回归（ＳＭＬＰ）技术解析近红外（ＮＩＲ）光谱数据来测定汽油中乙醇含量。在乙醇浓度为６％～２０％范围内，间隔０．４％，其选取３６个浓度值进行逐步回归计算。通过考察校正标准误差、预测标准误差、校正相关系数和预测相关系数，以及考察最佳实验条件建立了最优回归方程，其校正标准误差和预测标准误差分别为：０．０８４９４和０．０８４９２。实验和预测结果表明：该方法简单、快速、准确，为测定汽油中乙醇加入量提供了一种切实可行的方法。
Stepwise multiple linear regressionnear infrared spectroscopy is used to determinate ethanol in gasolines. 36 density values are calculated regressively under the conditions of ethanol desity 20% and interval 0.4%. The optimal regression equation has been built by investigating the standard error of calibration, the standard error of prediction, the multiple correlation coefficient, the multiple prediction coefficient and the best experimental conditions. The standard error of calibration and the standard error of prediction are 0.084 94 and 0.084 92, respectively. The prediction results show that the method is simple, accurate and rapid.
出处
《抚顺石油学院学报》
1998年第2期1-3,6,共4页
Journal of Fushun Petroleum Institute
关键词
汽油
乙醇
MIR光谱
测定
SMLP
Near infrared
Stepwise multiple linear regression
Gasoline
Ethanol
