摘要
浓缩液化学质量的稳定直接影响着再造烟叶产品的质量,传统方法分析其常规化学成分费时费力,反馈相对滞后。利用近红外透射技术结合偏最小二乘法建立了造纸法再造烟叶生产过程中浓缩液烟碱(Nic)、总糖(Ts)及还原糖(Rs)定量分析模型。通过考察样本划分、光谱区域选择、光谱预处理及最佳主因子数选择等方面对模型进行不断优化,最终确定了最佳建模参数。此外,通过选取未参与建模的30组样品进行外部验证并通过T检验-成对双样本均值分析得出所建立的模型预测性及准确性较好,可以应用于浓缩液化学质量的快速检测。
The quality of reconstituted tobacco products was directly affected by the chemical quality of concentrated liquid. The traditional analysis method of routine chemical components was time-consuming and got data with delay. A rapid quantitative analysis method was set up for routine chemical components （nicotine, total sugar and reducing sugar） in reconstituted tobacco process concentrated liquid based on the combination of near-infrared transmission technology and partial least squares （PLS） method. The best model parameters were finalized by sample division, spectral region selection, spectrum pretreatment and PLS factors selection, etc. Moreover, these three prediction models were established with 30 different samples by external verification and T-test, and had been verified as accurate and good prediction. These prediction models could be applied to chemical quality detection for concentrated liquid.
出处
《红外技术》
CSCD
北大核心
2016年第1期64-70,共7页
Infrared Technology
关键词
近红外透射
再造烟叶浓缩液
定量分析
near-infrared transmittance, reconstituted tobacco concentrated liquid, quantitative analysis
