To investigate the nonlinear low-order models and functional models behavior of high-purity distillation columns

• Main Authors: Wu Nu Lee, 李文能
• Other Authors: Shu wei Yu
• Format: Others
• Language: zh-TW
• Published: 1994
• Online Access: http://ndltd.ncl.edu.tw/handle/77334766116192268000

碩士 === 國立中央大學 === 化學工程研究所 === 82 === Due to the nonlinear dynamic behavior of high-purity distillationmns , it is normally difficult to achieve satisfactory control ofe units using nominal model based linear control algorithms . Theary objective of this study is to investigate the feasibility ofizing simplied low-order nonlinear models in multiloop feedbackgn . Linear transfer function models obtained at different oper-ition of a methanol / water dynamic simulation model prov-the basis for the low- order models proposed in thsis . The first approach is to correlate parameter such as stead-statecess gain and time constants of the linear transfer functions asctions of the overlead and bottoms product compositions . The se-d approach is to include manipulated variables in parameter esti-ion in addition to product compositions . To improve accuracy oflow- order models , variables transformations of product composi-ns is also employed . Low-order models based on absolute composition normalized compo-ion and variable transformation are companed extensively withamic responses of the rigorous model under open-loop and closed-p conditions . Test results reveal that low-order model bulit oniable transformation matches well with rigorous model . Limitati-and deficiencies of the variable low-order models are also prese-d .