关键词: 响应曲面法, 遗传算法-人工神经网络, 超声波, 水提取, 多酚

Abstract: To investigate the extraction of polyphenols using water assisted by ultrasonic cell grinder,the effects of temperature,duty cycle and solid-liquid ratio on the yield of polyphenols(significance,interaction,primary and secondary) were visually analyzed using response surface methodology(RSM) with orange peel as raw material.Subsequently,the RSM data were taken as samples and deeply learned by genetic algorithm-artificial neural network(GA-ANN).The results showed that high temperature(70 ℃) effectively promotes the release of polyphenols but the degradation of polyphenols was obviously induced by high duty cycle in aqueous system.The optimum conditions of GA-ANN were as follows:temperature 67 ℃,duty cycle 1/2,solid-liquid ratio 1∶25 g/mL.Total phenolic content was 8.21 mg GAE/g,which reached 84% of the extraction yield in 80% methanol aqueous.This study proved the feasibility of ultrasound-assisted aqueous extraction of polyphenols and the superiority of RSM and GA-ANN in optimum extraction.

Key words: RSM, GA-ANN, ultrasound, aqueous extraction, polyphenols