Application of ANFIS soft computing technique in modeling the CO2 capture with MEA, DEA, and TEA aqueous solutions

Abstract

As continuous process, CO2 capture from natural gas is crucial for enhancing the heating value of the gas and reduction of CO2 emissions as greenhouse gas, as well. A commonly applied CO2 capture technology is based on absorption by amine solutions. A novel study of adaptive neuro-fuzzy inference system (ANFIS) technique is performed in this communication for modeling the CO2 loading capacity of MEA, DEA, and TEA aqueous solutions as function of system's temperature, partial pressure of CO2, and concentration of amines at aqueous phase. Development of ANFIS models are on the basis of the Gaussian membership function. The results obtained by the presented ANFIS models are analyzed by employing average relative deviation (%ARD), absolute average relative deviation (%AARD), root mean square error (RMSE), and coefficient of determination (R2) as performance indicators. The prediction capabilities of the proposed ANFIS models have also been compared to the previously developed LSSVM models. With accordance to the error analysis results, applying ANFIS soft computing approach leads to significant improvement for modeling the CO2 capture with amine solutions compared to the available LSSVM models.