近红外光谱技术在线快速检测复烤片烟化学成分应用研究
Rapid Analysis of Chemical Components of Flue-cured Tobacco Strips with On-line Near Infrared Spectroscopy
胡芸 1刘娜 2姬厚伟 2黄锡娟 2彭黔荣 2邵学广3
作者信息
- 1. 贵州中烟工业有限责任公司技术中心,贵州贵阳 550009;南开大学化学学院,天津 300071
- 2. 贵州中烟工业有限责任公司技术中心,贵州贵阳 550009
- 3. 南开大学化学学院,天津 300071
- 折叠
摘要
[目的]实现打叶复烤生产过程中片烟化学成分的在线快速检测.[方法]采集2010年、2011年和2014年具有代表性的烟叶样品的在线近红外光谱,采用主成分马氏距离法和基于蒙特卡洛采样的奇异样本识别方法剔出异常光谱和化学异常样品,建立并优化复烤片烟6种化学成分(总植物碱、总糖、还原糖、总氮、钾和氯)的在线近红外分析模型.[结果]利用偏最小二乘方法建立的定量模型,其决定系数R2均在81%以上.通过模型外部检验发现,样本的近红外预测值与参考值的结果较为一致,氯的平均绝对误差小于0.1%,其他组分的平均相对误差小于5%.[结论]利用在线定量分析模型,可以实现复烤片烟化学成分的在线检测,为后期烟叶醇化、质量评价和配方设计提供数据支撑.
Abstract
[Objective]To on-line monitor the chemical components of tobacco leaf during the threshing and redrying process.[Method]The near infrared (NIR) spectra of representative samples from the years of 2010,2011,and 2014 were collected.The on-line NIR models of six chemical components (nicotine,total sugar,reducing sugar,total nitrogen,potassium and chlorine)were developed and optimized after outlier removal by principal component analysis-Mahalanobis Distance and Monte Carlo cross validation methods.[Result]Their coefficients of determination were above 81%.The relative average errors of five components(nicotine,total sugar,reducing sugar,total nitrogen,potassium) were less than 5%,the average absolute errors of chlorine were less than 0.1%.[Conclusion]The on-line NIR method is practicable for monitoring chemical components in threshing and redrying process,it provides data information for purifying,quality evaluation and producting formula design of tobacco leaves.
关键词
近红外光谱/复烤片烟/模型优化/在线检测/化学成分Key words
Near-infrared spectroscopy/Redried flue-cured strips/Model optimization/On-line monitoring/Chemical components引用本文复制引用
出版年
2017
国家科技期刊平台
NSTL
维普
万方数据