Thermodynamic models for CO2absorption

Abstract

The most mature technology for capturing CO2 in industrial power generation processes is absorption by aqueous alkanolamines. These processes are governed with both phase and chemical equilibrium; the representation of these phenomena is thus crucial for the modelling of absorption plants. In this work the thermodynamics of H2O-alkanolamine-CO2 systems is addressed.

The scope of the work has been somewhat multipartite. Since CO2 loaded aqueous alkanolamine solutions contain electrolytes, some fundamentals of electrolyte theory are presented. In particular, emphasis is put on the Debye- Huckel model. Further the theory regarding the frameworks of statistical mechanics for electrolyte solutions is explored and discussed in terms of thermodynamic consistency.

The major part of the thesis deals with the development of equilibrium models for H2O-alkanolamine-CO2 systems. An equilibrium model for the representation of vapour-liquid and chemical equilibria in aqueous alkanolamine solutions was developed. The model is based upon the refined electrolyte-NRTL model for calculating the liquid phase non-idealities. Vapour phase non-ideality is taken into account by utilizing the Soave-Redlich-Kwong equation of state. For comparison with the refined e-NRTL model, also the original e-NRTL and the extended UNIQUAC models were implemented. Chemical equilibrium is solved using a Gibbs energy minimization routine. Interaction parameters for the binary water-alkanolamine systems were re- gressed to experimental data on total pressures, excess enthalpies and freezing point depression. Interaction parameters involving the ionic species were fitted to VLE data for the CO2 loaded aqueous alkanolamine solutions. The implemented models were used to predict CO2 and amine partial pressures, liquid phase composition and heats of absorption for the MEA and MDEA based systems. The results of the refined e-NRTL model correspond well with experimental data.

Finally, the implemented equilibrium models were implemented into a rate based process simulator, CO2SIM. Process simulations were performed and compared to pilot plant data.