梗签含量和尺寸对卷烟物理指标的影响及预测研究OA
Effects of stem piece content and size on cigarette physical indexes and their prediction
[目的]探究梗签含量和尺寸对卷烟物理指标的影响及其内在关联.[方法]以常规卷烟为研究对象,设置梗签含量梯度组及尺寸梯度组,利用三维重构量化卷烟内部组分含量,采用相关性分析、基于指标间相关性的权重确定方法(CRITIC)等评价梗签参数对卷烟物理指标的影响,并建立极端梯度提升(XGBoost)模型预测卷烟物理指标.[结果]梗签尺寸越大,卷烟中梗签含量稳定性越低.梗签含量与卷烟单重、硬度和孔隙率呈极显著正相关,与圆周呈显著正相关,与开放吸阻呈显著负相关.当梗签含量为 4%,卷烟机风门全关时,卷烟物理指标稳定性和综合得分最优.卷烟对单重、硬度和孔隙率的 XGBoost 优化模型 R2 分别为0.953 1、0.950 3 和0.875 3,精度较高且优化效果显著.[结论]梗签含量与卷烟单重、硬度和孔隙率存在正相关关系,并可通过 XGBoost 模型对三者进行预测.
[Objective]To investigate the effects of stem piece content and size on cigarette physical indicators and their underlying relationships.[Methods]Conventional cigarettes were used as the research object,and gradient groups varying in stem piece content and size were prepared.Three-dimensional reconstruction was employed to quantify the internal component content.Correlation analysis and the CRITIC(Criteria Importance Through Intercriteria Correlation)weighting method were applied to evaluate the effects of stem piece parameters on cigarette physical indicators.Additionally,an extreme gradient boosting(XGBoost)model was constructed to predict these indicators.[Results]Larger stem piece size resulted in lower stability of stem piece content in cigarettes.Stem piece content exhibited highly significant positive correlations with cigarette weight,hardness,and porosity,a significant positive correlation with circumference,and a significant negative correlation with open draw resistance.When the stem piece content was 4%and the air damper of the cigarette maker was fully closed,the cigarette physical indicators showed optimal stability and comprehensive scores.The coefficients of determination(R2)of the optimized XGBoost models for cigarette weight,hardness,and porosity were 0.953 1,0.950 3,and 0.875 3,respectively,indicating high predictive accuracy and significant model improvement.[Conclusion]Stem piece content was positively correlated with cigarette weight,hardness,and porosity,and these three indicators could be accurately predicted using the XGBoost model.
叶政宏;姬会福;段鹍;杨琦;翁嘉凯;顾甲甲;邵惠芳;丁美宙
河南农业大学 烟草学院,河南 郑州 450046河南农业大学 烟草学院,河南 郑州 450046河南中烟工业有限责任公司 技术中心,河南 郑州 450003河南中烟工业有限责任公司 技术中心,河南 郑州 450003河南农业大学 烟草学院,河南 郑州 450046河南中烟工业有限责任公司 技术中心,河南 郑州 450003河南农业大学 烟草学院,河南 郑州 450046河南中烟工业有限责任公司 技术中心,河南 郑州 450003
轻工纺织
梗签含量梗签尺寸物理指标预测模型XGBoost
stem piece contentstem piece sizephysical indexespredictive modelXGBoost
《轻工学报》 2026 (4)
94-103,10
中国烟草总公司重点研发项目(110202202010)河南中烟工业有限责任公司科技项目(AN2023018)河南省研究生教育改革与质量提升工程项目(YJS2026YBGZZ21)
评论