Abstract
A generalized form of the semi-implicit Runge-Kutta method proposed by Michelsen was developed removing the necessity of defining new variables which place system differential equations in state variable form (ordinary differential equation with only one derivative). The application of the generalized semi-implicit Runge-Kutta method, the semi-implicit Runge-Kutta method proposed by Michelsen, and Gear's method in solving systems of coupled differential and algebraic equations which describe dynamic separation processes were investigated. One of the dynamic separation models developed in the investigation of the numerical methods was expanded such that two control systems are included. In addition, changing from one control system to the other is possible while the column is responding to a given disturbance. Sparse matrix operation techniques, such as LU factorization and packed form of storage, were used to solve the equations in the dynamic separation models. These techniques had been coded into two FORTRAN subroutines and were tested on a steady state distillation column operating at the condition of minimum reflux ratio before being applied to the solution of dynamic models. The objective in doing so was twofold: (1) to develop two models for columns operating at minimum reflux ratio, and (2) to check the validity of the subroutines.
Feng, A. (1984). Dynamics of separation processes. Texas A&M University. Texas A&M University. Libraries. Available electronically from
https : / /hdl .handle .net /1969 .1 /DISSERTATIONS -555734.