期刊文献+

基于BP神经网络的乳酸脱水制丙烯酸仿真模拟 被引量:12

Simulation of preparation of acrylic acid from lactic acid dehydration based on BP neural network
下载PDF
导出
摘要 通过生物质发酵产生的乳酸(2-羟基丙酸)在食品和医药领域具有广泛的商业应用价值,作为可再生资源的生物质乳酸已成为备受关注的化学原料。本文采用BP神经网络对乳酸脱水制丙烯酸的催化过程进行了仿真研究。采用正交实验设计确定实验点,主要考察原料液pH、原料液流速、载气流量和温度4个因素对丙烯酸产率的影响。针对正交实验的缺陷,将正交实验设计和神经网络结合起来,采用基于DOS界面并能方便调节BP神经网络计算的程序包对正交实验结果进行训练,用训练好的网络模拟催化反应体系的动态过程。结果表明,由神经网络仿真模拟出的三维图可以直观地体现各个反应条件对丙烯酸产率的影响,并用穷举法求出最佳反应条件,在该条件下的神经网络模拟产率为27.45%,与实验结果较吻合,相对误差约为-0.4391%。 Lactic acid (LA) is a commercial fine chemical, used in food and medicine, which is readily available by biomass fermentation. As a chemical feedstock, LA is widely used due to the renewable biomass resources. In this paper, BP neural network was applied to simulating the dehydration reaction of lactic acid. The optimal reaction conditions were determined through orthogonal design that could be used to investigate the interrelated effects of pH, feed liquid flow rate, carrier gas flow rate and temperature on the yield of acrylic acid (AA) . With the training samples from experimental data of orthogonal design, a new model that could simulate the process of the dehydrated reaction was established based on BP neural network theory. Then the network system trained was used to anticipate the effects of different ingredients and their interactions on AA yield. The three-dimensional graphs produced by the network could effectively express the relationships between reaction conditions and catalytic activity. Finally, with the help of the network, the highest AA yield of 27.45% was obtained, which was very close to the experimental result with the relative error of -0. 4391%.
出处 《化工学报》 EI CAS CSCD 北大核心 2009年第1期83-88,共6页 CIESC Journal
基金 中石化科研开发项目(206003)
关键词 神经网络 丙烯酸 乳酸 脱水 neural network acrylic acid lactic acid dehydration
  • 相关文献

参考文献6

二级参考文献44

共引文献37

同被引文献113

引证文献12

二级引证文献21

相关作者

内容加载中请稍等...

相关机构

内容加载中请稍等...

相关主题

内容加载中请稍等...

浏览历史

内容加载中请稍等...
;
使用帮助 返回顶部