Rigorous modeling of CO2 equilibrium absorption in MEA, DEA, and TEA aqueous solutions

• A machine learning system is presented for modeling CO2 absorption in alkanolamine aqueous solutions.
• Monoethanolamine (MEA), diethanolamine (DEA), and triethanolamine (TEA) are studied herein.
• Different levels of temperatures over wide range of CO2 partial pressure are investigated.
• The CSA-LSSVM approach has been employed to develop models for the application of interest.
• The models predictions are in satisfactory with corresponding target values for all studied systems.

The main objective of the presented communication is to utilize a Machine Learning System for modeling equilibrium CO2 absorption in monoethanolamine (MEA), diethanolamine (DEA), and triethanolamine (TEA) aqueous solutions at various alkanolamine concentrations at different levels of temperatures over wide range of CO2 partial pressure. Least Squares Support Vector Machine (LSSVM) approach has been employed to develop intelligent models for the application of interest. The required data for modeling purposes include the experimentally measured equilibrium data of (H2O + MEA + CO2), (H2O + DEA + CO2), and (H2O + TEA + CO2) systems reported in the literature. The optimum parameters of the proposed models have been obtained using Coupled Simulating Annealing (CSA) technique. According to the error analysis results, the models predictions are in satisfactory agreement with corresponding target values with R-squared of greater than 0.98 and the absolute average relative deviation percent (%AARD) to be less than 6.5% for all studied systems.