Decision tree‐based modeling of CO2 equilibrium absorption in different aqueous solutions of absorbents

Abstract

In absorption‐based CO2 removal processes, one of the most important parameters that must be determined is the equilibrium solubility of CO2 in solvents. This study is the first time to employ the extremely randomized trees, Extra Trees, methodology to develop an Extra Trees model for the CO2 loading capacity of various absorbents. The ability of the proposed decision tree‐based model in estimation of CO2 solubility values was compared to that of the previously presented models on the basis of the adaptive neuro‐fuzzy inference system (ANFIS), least squares version of the support vector machine (LSSVM), and artificial neural network (ANN), using error analysis. According to the obtained results, the presented Extra Trees model is able to estimate the CO2 loading capacity of solvents with an absolute relative deviation in percent (AARD%) equal to 0.15. The calculated values of AARD% for the literature models, i.e., LSSVM, radial basis function‐artificial neural network (RBF‐ANN), multilayer perceptron‐ANN (MLP‐ANN), and ANFIS, are 2.00, 10.03, 6.18, and 14.15, respectively. Hence, the developed Extra Trees model provides much higher robustness and accuracy in estimating/representing the solubility of CO2 in different aqueous solutions of solvents. © 2019 American Institute of Chemical Engineers Environ Prog, 38: S441–S448, 2019