Assessment of mass transfer correlations used in post-combustion CO2 capture by piperazine activated 2-amino-2-methyl-1-propanol (a-AMP)

The prevalent solvents used in CO2 capture processes have such drawbacks as degradability, high energy consumption for solvent regeneration, low absorption capacity, and corrosive nature which have convinced the researchers to suggest a new combination of alkanolamines in place of the commonly-used solvents like MEA. Since Rate-Based model, compared to Equilibrium model, gives more accurate simulations of CO2 capture processes by amine solvent, it is used in this study. Mass transfer correlations are significant parameters in Rate-Based model, and using them improperly will lead to considerable disagreement between real and simulated results. In the present work, mass transfer coefficients in CO2 capture processes are evaluated using an aqueous blend of piperazine activated 2-amino-2-methyl-1-propanol (a-AMP) solution for the first time. It is done according to Khan et al. (2016) experimental results in 36 points and for different operational conditions. In the first evaluation, 12 sets of data are achieved following the changes of the operational parameters, gas flow rate and CO2 partial pressure, while other parameters such as operational absorption temperature, solvent flow rate and solvent blends of AMP thorn PZ remain constant. The studies indicate that considering the amount of CO2 absorbed, the Mean Absolute Errors (MAE) of Onda et al., Bravo-Fair, and Billet-Schultes mass transfer coefficients are 12%, 7.63%, and 3.2% respectively in comparison with Khan et al. results. The second evaluation is also done in 12 operational points regarding the variations of PZ concentration and operational absorption temperature. Other parameters remain unchanged in this evaluation, too. This study demonstrates that Billet-Schultes correlations own higher accuracy compared to the Onda et al. and Bravo-Fair. Mean Absolute Errors (MAE) of Onda et al., Bravo-Fair, and Billet-Schultes are 12.74%, 8.26%, and 2.04% respectively. The third evaluation of this work is of rich solvent loading (absorber loading). Similarly, this evaluation is performed in 12 operational points in regard to the changes of CO2 partial pressure and the ratio of AMP to PZ, while other operational parameters are kept constant. This evaluation likewise shows that, compared to two other ones, Billet-Schultes correlations have more precision when utilized in CO2 capture processes by the a-AMP solvent. Mean Absolute Error (MAE) of Onda et al., Bravo-Fair, and Billet-Schultes in predicting absorber loading and in accordance with experimental data are 24.21%, 16.44%, and 6.23% respectively.