ANN-GA法优化石吊兰素提取工艺

Optimization of the nevadensin extraction conditions from Fewflower Lysionotus using ANN-GA

摘要: 为了研究人工神经网络技术（ANN）、遗传算法（GA）相结合的化学计量方法在石吊兰素回流提取过程中的应用,在单因素实验基础上,采用Box-Behnken实验设计和ANN-GA法研究乙醇浓度、提取时间和提取次数、固液比对提取液中石吊兰素含量的影响。得到石吊兰中石吊兰素的最佳提取工艺为:乙醇浓度84%,提取2.8h,固液比1∶17,提取2次。按照该条件进行验证,得到提取液中石吊兰素含量为2.35mg/g,与预测值误差为1.88%。结果表明,神经网络遗传算法模型拟合度较好,这一方法在工艺优化过程中具有广泛的应用前景。

Abstract: In order to explore the extraction process of nevadensin from Fewflower Lysionotus, the single factor experiment and artificial neural network-genetic algorithm (ANN-GA) methods were applied to optimize the concentration of ethanol, extracting time, extracting times and solid-liquid ratio.The optimal extraction process of nevadensin was as follows :ethanol concentration 84%, extraction time 2.8h, solid-liquid ratio 1∶17, and extracting 2 times, respectively.Based on this combined conditions, the predicted maximum was 2.35mg/g, and the error of the predicted value of 1.88%.Results indicated that ANN-GA method had wide application prospect in the optimization process.