箱式烘烤烟叶高温蒸汽回潮工艺参数优化  

Optimization of high temperature steam moisture recovery process parameters for box curing tobacco leaves

作　　者：董文帅 胡彬彬 闫伸 姜永雷 宋正熊 遆晋松 宋朝鹏[1] 赵浩宾 DONG Wen-shuai;HU Bin-bin;YAN Shen;JIANG Yong-lei;SONG Zheng-xiong;TI Jin-song;SONG Zhao-peng;ZHAO Hao-bin(College of Tobacco Science,Henan Agricultural University,Zhengzhou 450046,China;Yunnan Academy of Tobacco Agricultural Sciences,Kunming 650021,China;China Tobacco Corporation Staff Education College,Zhengzhou 450008,China;Henan Company,China National Tobacco Corporation,Zhengzhou 450018,China)

机构地区：[1]河南农业大学烟草学院,郑州450046 [2]云南省烟草农业科学研究院,昆明650021 [3]中国烟草总公司职工进修学院,郑州450008 [4]中国烟草总公司河南省公司,郑州450018

出　　处：《西南农业学报》2025年第1期200-208,共9页Southwest China Journal of Agricultural Sciences

基　　金：河南省烟草公司资助项目(2023410000240027);云南省烟草公司科技计划项目(2023530000241022)。

摘　　要：【目的】确定箱式烘烤烟叶关键工艺参数,为箱式烘烤烟叶回潮工艺优化提供参考。【方法】利用高温蒸汽回潮房回潮设备,对箱式烘烤烟叶高温蒸汽回潮工艺进行研究。以前期炉内温度、转换点含水率和后期炉内温度为试验因素,以单位能耗、烟叶硬度为试验指标,采用加权综合评分法进行赋权计算综合评分值,并结合响应面法(RSM)和遗传算法-反向传播(GA-BP)神经网络进行参数优化及预测。【结果】前期炉内温度、转换点含水率和后期炉内温度对烟叶综合评分值的影响显著,其中后期炉内温度对烟叶综合评分值的影响最大,其次为转换点含水率和前期炉内温度。RSM和GA-BP神经网络模型对箱式烘烤烟叶最佳回潮工艺的预测性能均较佳,其中RSM模型的R2=0.9792,GA-BP神经网络模型的R2=0.9692。经RSM优化的箱式烘烤烟叶回潮工艺参数为:前期炉内温度111.00℃、转换点含水率10.00%、后期炉内温度114.00℃,综合评分值的预测值为23.17,试验实际值为22.91;GA-BP神经网络优化得到的箱式烘烤烟叶回潮工艺参数为:前期炉内温度112.00℃、转换点含水率11.00%、后期炉内温度114.00℃,综合评分值的预测值为22.77,试验实际值为22.95;GA-BP神经网络预测值与实际值的相对误差为0.80%,优于RSM的相对误差1.12%。【结论】GA-BP神经网络模型可有效确定箱式烘烤烟叶最佳回潮工艺,该研究可为箱式烘烤烟叶回潮工艺优化提供参考。【Objective】The study aimed to identify the key process parameters for box curing tobacco leaves and offer a reliable reference for optimizing the moisture recovery process in this context.【Method】The experiment was conducted to study on the high temperature steam moisture recovery process of box curing tobacco leaves using specialized equipment.The study focused on early furnace temperature, water content at the conversion point and late furnace temperature as test factors.Specific power consumption and tobacco hardness were used as indicators.A weighted comprehensive scoring method was employed to assign weights and calculate the comprehensive score.The response surface method(RSM) and the genetic algorithm-back propagation(GA-BP) neural network were combined to optimize and predict the parameters.【Result】The early furnace temperature, water content at conversion point and late furnace temperature have a significant impact on the comprehensive score.Among them, the late furnace temperature had the greatest impact on the comprehensive score, followed by the water content at conversion point and early furnace temperature.Both RSM and GA-BP neural network models had good predictive performance for the optimal moisture recovery process of box curing tobacco leaves, with R~2=0.9792 for RSM model and R~2=0.9692 for GA-BP neural network model.The optimization results for the moisture recovery process parameters obtained through RSM were: early furnace temperature of 111.00 ℃,water content at conversion point of 10.00%,and late furnace temperature of 114.00 ℃;The predicted comprehensive score value was 23.17,and the actual experimental value was 22.91.For the GA-BP neural network, the optimized parameters were: early furnace temperature of 112.00 ℃,water content at conversion point of 11.00% and late furnace temperature of 114.00 ℃;The predicted comprehensive score value was 22.77,and the actual experimental value was 22.95.The experimental results demonstrated that the GA-BP neural network's predict

关 键 词：箱式烘烤 高温蒸汽 回潮工艺 响应面法 遗传算法 BP神经网络 

分 类 号：S572[农业科学—烟草工业]