基于BP神经网络对汽油辛烷值损失预测模型的构建

汽油燃烧尾气中含有的硫、烯烃等混合物对环境造成了极大的污染,但企业脱硫降烯的过程也会降低代表企业利润的辛烷值含量.通过数据关联或机理建模,可以刻画化工过程与辛烷值含量的关系,为解决传统的数据关联模型中变量相对较少、机理建模对原料的分析要求高、对过程优化的响应不及时等问题,本文利用Matlab软件,基于粒子群优化算法,通...

Prediction of gasoline octane loss based on BP neural network model

The mixture of sulfur and olefin in the exhaust gas of gasoline combustion causes great pollution to the environment,but the process of desulfurization and detene will also reduce the octane number which represents the profit of the enterprise.Through data correlation or mechanism modeling,the relationship could be depicted between the content of chemical process and octane content.In order to solve the problems of relatively few variables in traditional data association models,high requirements for raw material analysis in mechanism modeling,and untimely response to process optimization,the paper uses Matlab,based on the Particle Swarm Optimization algorithm,the data collected in the production process of the factory is mined through the BP neural network model.Consequently,the prediction model of octane loss is established,225 data samples are selected to train the octane loss prediction model,and 100 samples are used to verify the octane loss model.The model is highly fitting to the prediction of the target value and solves the related problems well.