Neural Network Modeling of Adsorption Equilibria of Mixtures in Supercritical Fluids

A neural network model was used to predict the ternary adsorption equilibria of 2,6- and 2,7-dimethylnaphthalene isomers dissolved in supercritical carbon dioxide on NaY-type zeolite. The neural network was trained using binary (pure solute dissolved in supercritical carbon dioxide onto NaY-type zeolite) and ternary adsorption equilibrium data. Despite a limited number of data points available for the training of the network, the model was capable of predicting the ternary adsorption equilibria using the binary adsorption equilibrium data very precisely.