摘要
Abstract
In order to discuss the feasibility of predicting the nicotine, total sugar, reducing sugar and chlorine components in reconstituted tobacco raw materials(stem) with Near Infrared Spectroscopy , these four components models were established by Near-Infrared diffuse reflection spectroscopic combined with partial least squares (PLS) method. The spectrum of the 130 samples, supplied by Shanghai reconstituted tobacco S、Guangdong reconstituted tobacco raw materials and mixed stem materials , were collected by Fourier transform near-infrared analyzer. And their predicted effects, stability and accuracy were evaluated. The results indicate that:①these prediction models show approximate and lower Root Mean Square Error of Calibration(RMSEC) and Root Mean Square Error of Prediction (RMSEP), and their Corr. Coeff. are both above 0.97. Otherwise, the mean relative deviations of the predicted values of nicotine, total sugar, reducing sugar are3.47%, 1.23%, 1.31%and 2.34%respectively. ②By the verification of these prediction models, their values RMSEP/RMSEC are close to 1 while their RPD are both above 3, showing that quantitative analysis by Near-Infrared Spectroscopy has presented itself as an accurate, reliable and no significant difference with Flow Analysis method.关键词
近红外/常规化学组分/烟梗Key words
near-infrared spectroscopy/routine chemical components/tobacco stem分类
轻工纺织
