###
食品研究与开发:2024,45(23):100-107
本文二维码信息
码上扫一扫!
基于响应面和人工神经网络-遗传算法优化液态发酵制备藕渣可溶性膳食纤维
(1.南京农业大学食品科技学院,江苏 南京 210095;2.江苏省农业科学院农产品加工研究所,江苏 南京 210014)
Optimization of Soluble Dietary Fiber from Liquid Fermented Lotus Root Residues Based on Response Surface and Artificial Neural Networks-Genetic Algorithms
(1.College of Food Science and Technology,Nanjing Agricultural University,Nanjing 210095,Jiangsu,China;2.Institute of Agricultural Processing,Jiangsu Academy of Agricultural Sciences,Nanjing 210014,Jiangsu,China)
摘要
图/表
参考文献
相似文献
本文已被:浏览 53次   下载 14
投稿时间:2024-03-05    
中文摘要: 为实现莲藕副产物的高值化应用以及藕渣可溶性膳食纤维(soluble dietary fiber,SDF)含量的提高,该文以藕渣为原料,接种副干酪乳杆菌PC18 发酵改性制备藕渣可溶性膳食纤维,通过单因素试验及Box-Behnken 试验设计,探讨料液比、发酵时间、接种量对藕渣可溶性膳食纤维含量的影响,在此基础上,选用响应面(response surface methodology,RSM)和人工神经网络-遗传算法(artificial neural networks and genetic algorithms,ANN-GA)构建发酵模型,并对两者优化结果及模型参数进行比较。单因素及Box-Behnken 试验结果表明,影响藕渣SDF 含量主要因素主次顺序依次为发酵时间、料液比、接种量。通过验证试验及模型参数与优化结果对比分析,发现RSM 与ANN-GA 构建的模型整体拟合效果均较好,但ANN-GA 模型预测值、试验值更高,相对误差值更低。基于ANN-GA 确定最佳工艺参数为料液比1∶38(g/mL)、发酵时间48 h、接种量4.0%,经验证得到此条件下SDF 含量为(5.97±0.73)%。
Abstract:To realize the high-value application of lotus root by-products and improve the soluble dietary fiber(SDF)content of lotus root residues,this paper used lotus root residues as raw materials and inoculated Lactobacillus paracasei PC18 to prepare SDF from lotus root residues through fermentation and modification.The effects of material-liquid ratio,fermentation time,and inoculation amount on the SDF content of lotus root residues were investigated through single-factor experiments and Box-Behnken experiments.On this basis,the fermentation model was constructed by using response surface methodology(RSM),as well as artificial neural networks and genetic algorithms(ANN-GA),and the optimization results and model parameters of the two methods were compared. The results of the single-factor and Box-Behnken experiments showed that the main factors affecting the SDF content of lotus root residues,in descending order of significance,were fermentation time,material-liquid ratio,and inoculum amount. Through the validation experiments and the comparative analysis of model parameters and optimization results,it was found that the overall fitting effect of models constructed by RSM and ANN-GA were both good,but the ANN-GA model had higher predictive and experimental values and lower relative error values.Based on ANN-GA,the optimal process parameters were determined as a material-liquid ratio of 1∶38(g/mL),a fermentation time of 48 h,and an inoculum amount of 4.0%,and the SDF content was verified to be(5.97±0.73)% under these conditions.
文章编号:202423014     中图分类号:    文献标志码:
基金项目:江苏省科技项目(BE2022367)
引用文本:


用微信扫一扫

用微信扫一扫