LAUSR

Abstract In this study, the rate of CO2 mass transfer into various nanofluids in two different systems (packed bed column and hollow fiber membrane) has been predicted using an adaptive neuro-fuzzy inference system (ANFIS). Various effective parameters including nanoparticle diameter, nanoparticle concentration, liquid flow rate, gas flow rate and CO2 inlet concentration have been chosen as input variables. Moreover, in order to investigate the type of nanoparticle effect on mas transfer rate, density of nanoparticles is considered as an input for ANFIS model. The proportion of train data was 80% of all data and the remainder data (20%) has been chosen as test data. The obtained results for CO2 mass transfer rate in both of packed bed column and hollow fiber membrane show a good agreement with experimental data. The accuracy between real values and ANFIS model results were analyzed with mean square error (MSE) and regression coefficient (R2). The best R2 and MSE values for test data in packed bed column and hollow fiber membrane data sets are 0.99946, 4.598 × 10−7 and 0.99592, 4.0779 × 10−4, respectively. Therefore, the ANFIS model is capable of predicting the CO2 mass transfer rate into all of the nanofluids precisely two distinct absorption systems, including packed bed column and hollow fiber membrane.