对二甲苯氧化反应器燃烧程度神经网络模型  

Development of an artifical neural network model for combustion reaction in p-xylene oxidation reactor

作　　者：颜学峰[1] 余娟[1] 钱锋[1] 

机构地区：[1]华东理工大学自动化研究所,上海200237

出　　处：《聚酯工业》2005年第1期14-17,共4页Polyester Industry

摘　　要：提出一种工业装置对二甲苯氧化反应器的燃烧程度模型,其燃烧程度通过反应器尾气中二氧化碳和一氧化碳总含量来表征。模型采用神经网络技术建立进料在反应器中停留时间、循环醋酸溶液的催化剂和溴促进剂浓度、反应器的反应温度、进入反应器料液的溶剂比等主要可调工艺操作参数与反应器尾气中二氧化碳和一氧化碳总含量的关联模型。建立模型的预测相对误差平均值为3.47%。The combustion reaction model of p-xylene oxidation reactor, in which the summation content of carbon monoxide and carbon dioxide in the reactor off-gas is regarded as the degree of the combustion reaction, was proposed. The artifical neural network technique was employed to model the relationship between the main operation conditions (i.e. residence time, catalyst concentration, promoter concentration, reaction temperature, and solvent ratio) and the summation content of carbon monoxide and carbon dioxide in the reactor off-gas. And, the mean of prediction relative error of the obtained model is 3.47%.

关 键 词：PTA 氧化反应 神经网络模型 

分 类 号：TQ245.12[化学工程—有机化工]