基于BP神经网络的微波辅助提取无花果黄酮工艺优化
Process parameter optimization of microwave assisted extraction of flavonoids from Ficus carica Linn based on BP neural network
-
摘要: 为优化微波辅助提取无花果黄酮成分的工艺参数,在响应面实验数据的基础上,采用神经网络与遗传算法(BP-GA)相结合的方法优化微波辅助提取无花果黄酮的工艺参数,以提取时间、微波频率、料液比和提取温度作为网络输入,黄酮提取量作为网络输出对微波辅助提取无花果黄酮工艺进行分析模拟。结果表明,利用BP-GA得到微波辅助提取无花果黄酮的最优工艺条件为提取时间22 min,微波频率609 W,料液比1∶30 g/m L,提取温度50℃,此条件下无花果黄酮提取量为14.6610 mg/g;与BP-GA神经网络预测值相对误差为0.15%,表明神经网络优化无花果黄酮提取工艺具有很好的可靠性和实用价值,为其工艺优化提供了可行的途径。Abstract: In order to explore the optimum microwave-assisted extraction of flavonoids from the fruits of Ficus carica Linn, Back-Propagation neural network with genetic algorithm ( BP-GA) was used to optimize the process parameters of microwave assisted extraction of flavonoids from Ficus carica Linn based on the response surface experimental data.The simulation analysis of microwave assisted extraction of flavonoids from Ficus carica was carried out by using the extraction time, microwave frequency, the ratio of solid to ethanol and extraction temperature as the input of the network and flavonoids extraction amount as output.The optimized operation condition obtained by the neutral network method was extraction time of 22 min, microwave power of 609 W, the ratio of solid to ethanol of 1 ∶ 30 g/m L and extraction temperature of 50 ℃. And its maximum extraction yield climbed to 14.6610 mg/g under the optimal conditions, the deviation between verified and predicted values of yield was0.15%, which indicated the reliability and practicability in the optimized conditions.The research may provide a new method for optimizing the processing conditions for further development and utilization of the fig organization and build the foundations for better use of natural resources.