基于Design-Expert构建烟碱浸提率预测模型的研究

Construction of Prediction Model on Nictoine Extraction Rate Based on Design-Expert

通过析因实验，用Design-Expert构建涉厦浸提时间（A）、浸提温爱（B）、液料比（C）、烟末粒径（D）的烟碱浸提率预测模型。当A=10—120min，B=20—60℃,C=20～60mL／g，D=O．3～1．25mm时得到烟碱浸提率预测模型R（％）=0．5020＋0．0037C，特征参数液料比（C）对烟碱浸提率的影响最为显著，贡献值述91.75％．此模型的拟合度达99．99％．校正决定系数AdjR^2=0．92。烟碱实际浸提率与模型预测浸提率据对误差在5％以内，表踞该模型预测烟碱浸提率可信度高，具有很好的工程应用价值。

The prediction model on nictoine extraction rate involving extraction time （A）, extraction temperature （B）, liquid -solid ratio （C） and particle size （D） was constructed by Design -Expert using factorial design. This model, R （ % ） = 0. 5020 ＋ 0.0037C , was achieved as A = 10 ～ 120 rain, B = 20 ～ 60 ℃, C = 20 - 60 mL/g and D = 0.3 ～ 1.25 ram. The fitting degree af this model on the special factor C reached 99.99% and a high correlation eoeffieient of 0.92 was also obtained. These indicated the liquid - solid ratio was a critical factor for improving nictoine extraction rate and its contribution reached 91.75% compared with other factors and their interactions. Meanwhile, the relative error between actual and prediction extraction rate is tess than 5% which indicates this prediction model has a good prospect in engineering application.