FINITE-ELEMENT MODELING OF MULTISOLUTE ACTIVATED CARBON ADSORPTION

Multisolute modelling of activated carbon adsorption in a fixed bed reactor can be complex and expensive computationally. The computational techniques that have been utilized over the years are finite differences and a global method of orthogonal collocation. The global method of orthogonal collocation, which is currently gaining popularity over the more traditional finite difference techniques, has some inherent disadvantages. In an attempt to eliminate these disadvantages, finite element models are developed herein by using the characteristic-Galerkin method and the Pertov-Galerkin method. Both the models were found to be sensitive to changes in film transfer coefficients but were relatively insensitive to changes in diffusion coefficients. The models were also found to be stable and computationally efficient. The model predictions were compared to bisolute adsorption data, and it was found that the Pertov-Galerkin model gives a better description of experimental data and, in addition, gives convergent results using a significantly larger time step than the characteristic-Galerkin method.