BP神经网络在白酒高级醇生成条件研究中的应用  

Application of BP neural network for formation conditions of higher alcohols in liquor production

作　　者：罗惠波[1,2] 宗绪岩[2] 霍丹群[1] 王毅[2] 江文涛[2] 

机构地区：[1]重庆大学生物工程学院,重庆400030 [2]四川理工学院生物工程学院,自贡643000

出　　处：《食品科技》2012年第4期90-93,共4页Food Science and Technology

基　　金：四川省应用基础研究项目(2006J13-006)

摘　　要：为了更好的解决白酒生产中高级醇含量问题,采用BP神经网络研究了曲样、加曲量、加水量、加糠量和投料量对高级醇生成量的影响。结果表明:采用BP神经网络建立模型,预测偏差范围在-3.69%～6.39%之间。BP神经网络预测高级醇生成的模型效果较好,能符合实际情况,应用于实际生产中。In order to reduce the content of higher alcohols in liquor production,BP neural network is adopted to study the influence of higher alcohols with kind of distiller＇s,amount of distiller＇s,amount of water,amount of chaff and amount of material.The results showed：The range of prediction error was from-3.69% to 6.39% with BP neural network model.Model of higher alcohols production predicted with BP neural network is better,and can accord with the practical situation.It can be applied in practical production.

关 键 词：白酒 高级醇 BP神经网络 数学模型 

分 类 号：TS262.3[轻工技术与工程—发酵工程]