【作者】 周学秋； 劳彩莲； 陈都洪； 李力； 严衍禄； 刘惠颍； 陈吾峰； 戴百龄； 蒋晓春；

【Author】 Zhou Xuequi Lao Cailian Chen Duhong Li Li Yan Yanlu (Department of Applied Physics,China Agriculturial University,Beijing 100084) Liu Huiying Chen Wufeng Dai Bailing (Beijing Oil Research Institute,Beijing 102300) Jiang Xiaochun (Beijing Office,Bio-Rad China,Beijing 100088)

【机构】 中国农业大学应用物理系； 北京油料研究所； 北京伯乐分析生化仪器有限公司；

【摘要】 <正>引言近红外光谱(NIR)分析是一种理想的多指标光谱分析技术,近红外光谱的信息量丰富,图谱稳定性高,获得图谱容易;偏最小二乘法(PLS)是一种很有效的多元统计方法,近年来,NIR与PLS结合受到化学计量学工作者的重视,裴鑫德等应用NIR和PLS方法分析了谷物样品多种氨基酸的含量,取得了满意的效果。汽油的理化指标能对汽油发动机的功率和使用寿命产生很大的影响。所以测定汽油的更多还原

【Abstract】 This paper introduced the basic principle and arithmetical method of partial least squares(PLS).Detemined the octane number,density and near infrared spectrometry(NIRs) of the gasoline sample from difference areas in China. The results of calibration with PLS showed:the correlation coefficients(R) of octane number and density between the values determined by NIRs and classical method were 0.999 and 0.998 respectively,with the coefficients variant(CV ) being 0.003 and 0.001 correspondingly.Prediction results showed the correlation coefficients (R) of the above two indexes were 0.989 and 0.991 with the coefficients variant(CV) were 0.017 and 0.004 respectively.All of the results passed through T test and F test.更多还原