基于Box-Behnken响应面法与BP神经网络的油菜花粉总黄酮提取工艺研究

Study on the Extraction Process of Total Flavonoids from Rapeseed Pollen Based on Box Behnken Response Surface Methodology and BP Neural Network

目的 优化油菜花粉总黄酮的提取工艺。方法 以干膏得率、DPPH·清除率、总峰面积和指标峰面积为评价指标,通过熵权法赋以权重,得到综合评价指标。在单因素试验的基础上,建立Box-Behnken响应面法油菜花粉提取物HPLC指纹图谱-DPPH·清除率的谱效关系并计算关联度,利用关联度对指纹图谱峰面积进行校正,通过比较Box-Behnken响应面法与BP神经网络获得最佳提取工艺。结果 通过Box-Behnken响应面法获得的最佳提取工艺为乙醇浓度70%、料液比1∶25(g∶mL)、回流时间30 min。BP神经网络预测的最佳提取工艺为乙醇浓度70%、料液比1∶30(g∶mL)、回流时间为40 min。经验证,BP神经网络的综合评价指标为1.120 4,优于Box-Behnken响应面法的综合评价指标0.894 7。结论 优化后的油菜花粉总黄酮提取工艺稳定可靠,可为油菜花粉深度开发提供理论依据。

ObjectiveTo optimize the extraction process of total flavonoids from rapeseed pollen.MethodUsing the dry paste yield,DPPH · clearance rate,total peak area,and indicator peak area as evaluation indicators,a comprehensive evaluation index is obtained by assigning weights using the entropy weight method.On the basis of single factor experiments,a spectral efficiency relationship was established for the HPLC fingerprint of rapeseed pollen extract using Box Behnken response surface method-DPPH · clearance rate,and the correlation degree was calculated.The peak area of the fingerprint was corrected using the correlation degree.The optimal extraction process was obtained by comparing Box Behnken response surface method with BP neural network.ResultThe optimal extraction process obtained through Box Behnken response surface method was ethanol concentration of 70%,solid-liquid ratio of 1∶25(g∶mL),and reflux time of 30 minutes.The optimal extraction process predicted by BP neural network is ethanol concentration of 70%,solid-liquid ratio of1∶30(g∶mL),and reflux time of 40 minutes.After verification,the comprehensive evaluation index of BP neural network is 1.120 4,which is better than the comprehensive evaluation index of Box Behnken response surface method,which is 0.894 7.Conclusion The optimized extraction process of total flavonoids from rapeseed pollen is stable and reliable,providing a theoretical basis for the deep development of rapeseed pollen.