作者：李达¹, 李娥贤², 李晓科¹, 许仁杰¹, 蔡昊城¹, 袁鹏¹, 李超³

作者单位：1. 红云红河烟草（集团）有限责任公司，云南 昆明 650231;
2. 云南省农业科学研究院生物技术与种质资源研究所，云南 昆明 650201;
3. 云南中烟工业有限责任公司技术中心，云南 昆明 650024

关键词：感官品质;回潮区温度;清香风格特征;逐步线性回归

摘要：

为研究不同回潮区温度对卷烟清香风格特征的影响，采用GC-MS/MS 对50种国内市售成品卷烟烟丝中30种致香成分进行分析建模，以得到能够表征清香风格特征的模型，并根据该模型对3种回潮温度下的烟丝样品的清香风格特征的感官品质进行定量分析。研究发现：在30种致香成分中，23种进入回归方程，方程的决定系数为0.942，调整决定系数为0.932，该方程对于多种卷烟的清香风格特征具有良好的拟合效果。X₇~X₁₆这10种物质与清香评分呈现正相关，其中γ-十一内酯与清香评分呈现最大正相关；其余13种物质与清香评分呈现负相关，其中δ-十一内酯与清香评分呈现最大负相关。模型对测试集清香评吸指标具有较好的拟合效果，预测值与实际值误差较小，MAE值（平均绝对误差）为0.1007分。随着回潮区温度的升高，烟丝中的致香成分含量均有相应的降低，烟丝的致香成分数目也有降低。根据卷烟烟丝中的30种致香成分含量进行分析建模，建立清香风格特征感官品质与致香成分含量的联系，得到的回归模型预测的卷烟清香风格特征感官评吸得分与实际评吸结果吻合度较高，对于清香这一感官品质指标具有较好的预测能力。根据建立的模型研究回潮区温度对卷烟感官品质的影响，回潮区温度控制在50 ℃时更有利于得到感官评分更高的烟丝。

Research on the construction of a prediction model for cigarette fragrance style characteristics
LI Da¹, LI Exian², LI Xiaoke¹, XU Renjie¹, CAI Haocheng¹, YUAN Peng¹, LI Chao³
1. Hongyun Honghe Tobacco (Group) Co., Ltd., Kunming 650231, China;
2. Institute of Biotechnology and Germplasm Resources, Yunnan Academy of Agricultural Sciences, Kunming 650201, China;
3. Technology Center of China Tobacco Yunnan Industrial Co., Ltd., Kunming 650024, China
Abstract: To study the influence of different temperature of the regain zone on the characteristics of fen-flavour style tobacoo, 30 aroma components in 50 kinds of domestic commercial cut tobacco were analyzed and modeled by GC-MS/MS, so as to obtain a model that can characterize the fragrance characteristics. According to the model, the sensory quality of cut tobacco samples at 3 different moisture regain temperatures was quantitatively analyzed. The results showed that: among the 30 aroma components, 23 was selected into the regression equation, the determination coefficient of the equation is 0.942, and the adjustment coefficient is 0.932, indicating the good fitting effect of the regression equation. The 10 substances of X₇ to X₁₆ were positively correlated with the scent score, and the γ-undecalactone was the most positively correlated with the scent score. The other 13 substances were negatively correlated with the scent score, and δ-undecalactone was the most negatively correlated with the scent score. The model has a good fitting effect on the scent evaluation index of the test set. The error between the predicted value and the actual value is small, and the MAE value (average absolute error) is

points. With the increase of temperature in the moisture regain zone, the content of aroma components in cut tobacco has decreased accordingly, and the number of aroma components in cut tobacco has also decreased. Based on the analysis and modeling of 30 kinds of aroma components in cigarette tobacco, the relationship between sensory quality of fen-flavor style characteristics and content of aroma components was established. The results of regression model predicted that the sensory evaluation score of cigarette fen-flavor style characteristics was consistent with the actual results, which had a good prediction ability for the sensory quality index of fragrance clearing. According to the model, the influence of the resurgence temperature on the sensory quality of cigarette was studied. The 50 ℃ of temperature was controlled more favorable to get tobacco with higher sensory score.
Keywords: sensory quality;temperature in the resurgence zone;fragrance style characteristics;stepwise linear regression
2021, 47(7):64-70  收稿日期: 2020-09-29;收到修改稿日期: 2020-11-15
基金项目: 红云红河烟草（集团）有限责任公司科技项目（HYHH2016GY04）
作者简介: 李达(1980-),男,云南昆明市人,工程师,硕士,主要从事烟草化学、信息化研究
参考文献
[1] 刘秀明, 王艳芳, 李超, 等. GC-MS指纹图谱结合多元统计方法用于香精质量差异分析[J]. 中国测试, 2018, 44(6): 42-47
[2] 景延秋, 宫长荣, 张月华, 等. 烟草香味物质分析研究进展[J]. 中国烟草科学, 2005(2): 44-48
[3] 段胜智, 李军营, 杨利云, 等. 烟叶致香成分及其环境影响因子的研究进展[J]. 贵州农业科学, 2015, 43(1): 45-52
[4] 刘字, 颜合洪. 烟叶致香物质的研究进展[J]. 作物研究所, 2006, 20(5): 470-474
[5] 国家质量监督检验检疫总局. 卷烟第4部分: 感官技术要求: GB 5606.4—2005[S]. 北京: 中国标准出版社, 2005.
[6] 国家烟草专卖局. 卷烟中式卷烟风格感官评价方法: YC/T 497—2014[S]. 北京: 中国标准出版社, 2014.
[7] 沈宁, 赵娟, 于静洋, 等. 基于PLSR分析常规化学成分对市售卷烟感官品质的影响[J]. 食品与机械, 2015, 31(6): 47-52
[8] 吴宁宁, 杨俊, 张天栋, 等. 卷烟的感官特征与其烟丝提取物GC/MS 数据之间的关系[J]. 烟草科技, 2013, 11: 31-35
[9] 郭春生, 李力群, 纪旭东, 等. 基于BP神经网络的常规化学成分预测烟气成分和感官得分预测模型研究[J]. 农产品加工, 2017(2): 51-55
[10] 薛超群, 王建伟, 奚家勤, 等. 烤烟烟叶理化指标与浓香型风格程度的关系[J]. 烟草科技, 2012(1): 52-56
[11] 林顺顺, 张晓鸣. 基于PLSR分析烟叶化学成分与感官质量的相关性[J]. 中国烟草科学, 2016, 37(1): 78-82
[12] 周芳芳, 张晓龙, 詹军, 等. 典型清香型烤烟产地间致香物质含量及组成差异分析[J]. 中国烟草学报, 2016, 22(2): 34-42
[13] 席元肖, 魏春阳, 宋纪真, 等. 不同香型烤烟化学成分含量的差异[J]. 烟草科技, 2011(5): 29-33
[14] 周泽弘, 曹淋海, 王昌全, 等. 基于RBF神经网络建立库存烟叶香型的预测模型[J]. 中国烟草科学, 2016, 37(2): 65-70
[15] 邵惠芳, 赵蓉蓉, 范磊, 等. 松散回潮回风温度对烟叶化学成分与中性致香物质的影响[J]. 中国农业科技导报, 2016, 18(6): 138-145
[16] 李达, 李晓科, 许仁杰, 等. 回潮温度对中试试样烟丝中致香物质的影响[J]. 食品与机械, 2019, 35(5): 202-207
[17] 钟洪祥. 烟草化学中挥发性香味成分的分析研究[D]. 合肥: 中国科学技术大学, 2004.
[18] 林翔云. 调香术[M]. 北京: 化学工业出版社, 2008.
[19] 何晓群. 多元统计分析[M]. 北京: 中国人民大学出版社, 2012: 143-154.