Dynamic analysis and control of sieve tray gas absorption column using MATALB and SIMULINK

The present work highlights the powerful combination of SIMULINK/MATLAB software as an effective flowsheeting tool which was used to simulate steady state, open and closed loop dynamics of a sieve tray gas absorption column. A complete mathematical model, which consists of a system of differential and algebraic equations was developed. The S-Functions were used to build user defined blocks for steady state and dynamic column models which were programmed using MATLAB and SIMULINK flowsheeting environment. As a case study, the dynamic behaviour and control of a sieve tray column to absorb ethanol from CO2 stream in a fermentation process were analysed. The linear difference equation relating the actual and equilibrium gas phase compositions was solved analytically to relate the actual gas phase composition to the liquid phase with Murphree tray efficiency as a parameter. The steady state mathematical model was found to be nonlinear (w.r.t. number of stages) due to the introduction of the Murphree tray efficiency. To avoid the solution of large linear algebraic system, a sequential steady state solution algorithm was developed and tested through the idea of tearing the recycle stream in the closed loop configuration. The number of iterations needed to achieve a given tolerance was found to be function of the Murphree tray efficiency. The open-loop dynamic analysis showed that the gas phase composition response was nonlinear with respect to the inlet gas flow rate, while it was linear with respect to inlet gas composition. The nonlinearity increased along the column height and was maximum at the top tray. On the other hand, the Murphree tray efficiency had little effect on the dynamic behaviour of the column. The controlled variable was found to exhibit fairly large overshoots due to step change in the inlet gas flow rate, while the PID controller performance was satisfactory for step change in the inlet gas composition. The closed-loop dynamic analysis showed that the controlled variable (outlet gas phase composition) had a fairly linear dynamics due to step changes in the set point.