Exploring the effect of zeolite's structural parameters on the CO2 capture efficiency using RSM and ANN methodologies

Abstract

Kaolin-based zeolites have high adsorption capacity due to their combination of mesoporous and microporous structures. In this research, artificial neural networks (ANN) and response surface methodology (RSM) were utilized to model and optimize the carbon dioxide (CO2) adsorption capacity of different zeolites. Three types of zeolites synthesized from kaolin were employed. Structural characterization of zeolites and operational conditions were used to investigate their impacts on CO2 adsorption capacity. The RSM method employs the central composite design (CCD) to depict performance conditions through a model created via the least-squares technique. The RSM analysis shows that the R2 and adjusted R2 values are 0.9087 and 0.9103, respectively. As the RSM does not provide any information about the quality of the solution obtained, the ANN method was used as a global substitute model in optimization problems. The ANNs are versatile tools that can model and predict various nonlinear and complex processes. This article discusses the validation and enhancement of an ANN model, including the most commonly used experimental designs, their limitations, and general applications. The ANN weight matrix successfully predicted CO2 adsorption under different process conditions. Multilayer perceptron (MLP) and radial basis function (RBF) were employed in this study. The MLP model with two layers is recommended for adsorption simulation models, with a mean square error (MSE) of 0.003021. It is a valuable tool for optimizing CO2 capture processes with zeolites.

Keywords

• Zeolite
• CO2 capture
• Adsorption
• RSM
• ANN