基于BP神经网络的乙醇制备烯烃反应条件最优化设计  

Optimization Design of Olefin Reaction Conditions For Ethanol Preparation Based on BP Neural Network

作　　者：莫裕俊 Mo Yu-jun(Yancheng Institute of Technology,Yancheng Jiangsu 224000,China)

机构地区：[1]盐城工学院,江苏盐城224000

出　　处：《江西化工》2023年第2期88-91,共4页Jiangxi Chemical Industry

基　　金：盐城工学院《乙苯脱氢制苯乙烯实验装置等设备》(ZB-20170926-033)。

摘　　要：设计乙醇制备烯烃正交实验,提高C4烯烃收率和选择性。将A_(1)~A_(4)四组不同的Co/SiO_(2)催化剂组合控制在200mg,并使四组催化剂浓度不相等,可得250~400℃时乙醇转化率、C_(4)烯烃选择性与温度成正相关。建立方程分析组分数据,并比较5个因素,可得出C_(4)烯烃收率最大为29.06%。建立BP神经网络预测其余未知组合,350℃时,装料方式为B则C_(4)烯烃最大收率为52.66%;低于350℃时,装料方式为B则C_(4)烯烃最大收率29.49%,最优反应温度为350℃。Orthogonal experiment of ethanol preparation was designed to improve C_(4)Oene yield and selectivity.Four different groups of Co/SiO_(2) catalysts from A_(1) to A_(4) were controlled at 200mg,and the catalyst concentration was unequal,allowing for the ethanol conversion rate and C at 250~400℃_(4).The olefin selectivity is positively correlated with the temperature.Establishing equation analysis,component data,and comparing 5 factors,yields C_(4)The maximum yield of olefin was 29.06%.The BP neural network was established to predict the remaining unknown combinations,with the loading mode being C at B at 350℃_(4)The maximum yield of olefins was 52.66%.Below 350℃,0 B_(4)The maximum yield of olefins was 29.49%,and the optimal reaction temperature was 350℃.

关 键 词：相关性分析 PCA BP神经网络 最优化设计 正交实验 

分 类 号：TP278[自动化与计算机技术—检测技术与自动化装置]