Towards estimating absorption of major air pollutant gasses in ionic liquids using soft computing methods

Abstract

Capture of air pollutant gases using novel and green solvents is obtaining widespread attention. Accurate estimation of this process is complex. We have estimated the absorption of CO2, CH4, H2S, N2O, SO2 and CO gases in ionic liquids (ILs). We have applied Multilayer Perceptron-Artificial Neural Networks (MLP-ANN), Hybrid-Adaptive Neuro Fuzzy Inference System (Hybrid-ANFIS), Particle Swarm Optimization-Adaptive Neuro Fuzzy Inference System (PSO-ANFIS) and Coupled Simulated Annealing-Least Squares Support Vector Machine (CSA-LSSVM). We have gathered 3060 data of 72 IL-Gas mixtures for 40 types of ILs. The inputs of these models are: Temperature (T), pressure (P), IL molecular weight (MwIL), IL critical temperature (Tc, IL), IL critical pressure (Pc, IL), IL acentric factor (ωIL), gas molecular weight (Mwgas), gas critical temperature (Tc, gas), gas critical pressure (Pc, gas), gas kinetic diameter (d) and the acentric factor (ωgas). The CSA-LSSVM model produces best estimation with an Average Absolute Relative Deviation (AARD) of 8.7%. The results suggest the solubilities of the gases in ILs are correlated with structural factors of ILs. Estimation of the equilibrium behaviors in ionic liquids is of importance in simulation and design of solvent-based pollutant gas capture processes.