以乙醇体积分数、提取时间、提取次数、液固比以及粒子大小为试验因素,提取率为试验指标,采用均匀设计法安排试验,建立了人工神经网络模型,对系统进行辨识。利用已训练成功的神经网络模型,采用自主提出的多因素多水平邻近分析方法,对产品提取工艺的提取时间进行了较为系统的研究,对枇杷叶中熊果酸的提取工艺进行了优化,试验结果的一致性好。得到乙醇体积分数为65%,提取时间为15 min,液固比值为10,提取次数为2,颗粒目数为40目,产率为91.40%(模型计算为91.38%)。采用该工艺从枇杷叶中提取熊果酸时,时间对产率影响是非线性的,且在其他条件不同时表现出不同的影响规律。

The extraction process was studied by taking the concentration of ethanol,extraction time,times of extraction,solid-liquid ratio and particle size as test factors and extraction rate(productivity) as test target.The test was performed by uniform design method and the model was established by Artificial Neural Networks(ANN).The paper studied systematically the time characteristic in extraction process based on the trained neural networks model by our neighboring analysis of multifactor and multilevel.Through the model instruction,a number of verifying tests were done and the results were well consistent with the model calculation.A set of optimized process parameters was obtained: the concentration of ethanol is 65%,extraction time 15min,times of extraction 2,solid-liquid ratio 10,and particle size 40 meshes.The yield is 91.40%(value of model calculation is 91.38%).The conclusion is that the influence of extraction time on the yield of ursolic acid extraction is nonlinear and this influence is varied with other extracting conditions.