CAJ | 학술논문

为实现烟叶原料焦油和烟碱的快速检测，分别用烟丝（111个）和烟末（204个）样品建立了原烟卷烟主流烟气中焦油和烟碱的近红外模型，研究表明两种样品状态均能建立其近红外速测模型，且烟气烟碱的校正模型较好。用烟末建立的焦油和烟碱的校正模型略好于用烟丝建立的模型，其内部交叉验证均方差（RMSECV）分别为0.211和1.90，烟丝内部交叉验证均方差（RMSECV）分别为0.257和2.04。并对样品量较大的烟末模型进行了外部验证，2个模型预测值与标准值的平均相对偏差分别为5.13和5.93，t-检验表明预测值和标准值之间没有显著性差异，且系统精密度良好，可以用于大量样品的快速检测。
위실현연협원료초유화연감적쾌속검측，분별용연사（111개）화연말（204개）양품건립료원연권연주류연기중초유화연감적근홍외모형，연구표명량충양품상태균능건립기근홍외속측모형，차연기연감적교정모형교호。용연말건립적초유화연감적교정모형략호우용연사건립적모형，기내부교차험증균방차（RMSECV）분별위0.211화1.90，연사내부교차험증균방차（RMSECV）분별위0.257화2.04。병대양품량교대적연말모형진행료외부험증，2개모형예측치여표준치적평균상대편차분별위5.13화5.93，t-검험표명예측치화표준치지간몰유현저성차이，차계통정밀도량호，가이용우대량양품적쾌속검측。
In order to realize rapid detection of tar and nicotine in the cigarette,the near infrared(NIR) detection model for tar and nicotine in raw tobacco was established by 111 samples of cut tobacco and 204 samples of tobacco powder. The research showed that such NIR rapid detection model for two kinds of tobacco samples could be established, and the nicotine calibration model was better. The calibration model established through tobacco powder was better than which established through cut tobacco. The root mean square error of cross-validation (RMSECV) of the former was 0.211 and 1.90, and the RMSECV of the latter was 0.257 and 2.04. To verify the model established through tobacco powder, the mean relative deviation between the calculated value and the actual value for nicotine and tar was 5.13 and 5.93 respectively. The t-test showed good systematic precision, and there was no significant difference between the calculated value and the actual value, so the NIR model could be applied to rapid detection of mass sample.