精馏塔产品组分的非线性主元回归估计器  

A Nonlinear Principal Regression Estimator for Product Composition of a Distillation Tower

作　　者：文忠俞 金寿 华向明[1] 

机构地区：[1]华东理工大学自动化研究所

出　　处：《石油化工自动化》1999年第1期23-26,共4页Automation in Petro-chemical Industry

基　　金：国家教委博士点基金

摘　　要：以一个理想二元精馏塔为对象，讨论基于塔板温度的产品组分估计问题。并指出，相对于线性部分最小二乘估计器，基于神经网络的非线性主元回归估计器具有更好的稳态精度和动态性能。To estimate the product composition for an ideal binary distillation tower based on tray temperatures is discussed.After comparing linear principal regress and partial least square estimator,it is found that the nonlinear principal regression estimator based on the neural networks is more accurate in steady state and has better dynamic properties.

关 键 词：非线性主元回归 精馏塔 估计器 产品质量 

分 类 号：TQ051.81[化学工程]