Low order modeling for nonlinear process control

A key tenet of robust control theory is that robust performance is closely tied to model accuracy. However, detailed models lead to computationally intractable problems and simplified linear models cannot effectively capture the range of behavior exhibited by a nonlinear system such as a high-purity distillation column. Many distributed parameter systems including distillation columns and fixed bed reactors exhibit the phenomena of travelling waves in their transport behavior. Low order models based on the travelling waves of the composition profiles in the column are nonlinear, mathematically concise, and effective in capturing the behavior of a distillation column. These models are used in a nonlinear state feedback control scheme to obtain tight control of the column.