摘要
针对乙醛生产过程 ,建立关键过程变量粗乙醛浓度的软测量模型 ,并在此基础上建立粗乙醛的实时收率预测模型。对软测量实现中涉及到的回归变量选择、样本预处理、回归一致性分析、实时校正机制等关键技术进行讨论。该系统的预测值与离线分析值平均相对偏差为 1.2 %。
In this paper,a soft-sensing model of the coarse acetaldehyde is constructed,based on which the prediction model of the real-time yield is sequentially designed.Several key techniques in implementation,such as selection of the regression variables,preprocessing of samples,analysis of regression consistency and real-time rectification,have also been discussed.The whole system was implemented and was kept running for three months.The average deviation between the predicted values and the off-line laboratory analysis results is about 1.2 percent.
出处
《化工自动化及仪表》
CAS
北大核心
2004年第1期56-59,共4页
Control and Instruments in Chemical Industry
基金
国家"8 63"计划资助项目 (2 0 0 2AA412 110 )
关键词
乙醛生产过程
软测量
模型建立
在线校正
acetaldehyde production process
soft-sensing
model building
online rectification
