Adsorption-diffusion model with neural network-based equilibrium relationship Article uri icon

abstract

  • In this work, we propose a novel approach to predict adsorption equilibrium using artificial neural networks. The equilibrium model (particularly the derivative of such approach) is employed into a surface diffusion model to interpret the concentration decay curves during the adsorption of pyridine onto activated carbon in aqueous solution. Moreover, we estimated the external mass transfer and surface diffusion coefficients through an experimental-based inverse problem formulation. Also, we predict and compare with lab measurements the adsorption equilibrium curve and concentration decay dynamics. We found that results obtained by an artificial neural networks-based equilibrium model coupled with diffusional model agree well with experimental data. Thus, the artificial neural networks capabilities suggest that using a subrogate approach to predict the equilibrium relationship is an appropriate alternative when standard isotherm models fail. © 2018 Desalination Publications. All rights reserved.

publication date

  • 2018-01-01