Activity Coefficient Models

In this section, the activity coefficient models supported by the current version of SuperPro Designer are being presented.

Wilson

The Wilson equation is a two-parameter equation and requires two interaction parameters λ_ij and λ_ji per binary pair. It accurately predicts multicomponent mixture equilibria. For aqueous solutions, NRTL is recommended over Wilson, because NRTL better predicts phase equilibrium properties. Wilson model can not be used to predict liquid-liquid equilibria.

The partial molar volumes V_i and V_j are typically approximated, with satisfactory accuracy, by the molar volumes of the pure substances i and j. In the absence of user provided values for λ_ij and λ_ji, zero values are used by SuperPro.

NRTL

The NRTL model can be used to model vapor-liquid and liquid-liquid multi-component equilibrium systems. It performs better than Wilson for aqueous mixtures. The model requires two interaction parameters (g_ij, g_ji) and one binary parameter (a_ij).

In the absence of user provided values g_ij g_ji and a_ij, zero values are used by SuperPro.

SPT-NRTL

The SPT-NRTL (SMILEStoPropertiesTransformer NRTL) is a ten-parameter model that can be used to model the vapor-liquid equilibrium of mixtures with non-ionic components. It is implemented in SuperPro from the publication of Winter et al. in the Fluid Phase Equilibria journal. SPT-NRTL uses SMILES as molecular descriptors to predict activity coefficients. Specifically, the NRTL equation was integrated into a machine-learning model, trained with a very large set of experimental data, in order to estimate the required parameters. SPT-NRTL has eight “t” parameters (t_ij,1 to t_ji,4 and t_ji,1 to t_ji,4) and two “α” parameters (a₁ and a₂).

In the absence of user provided values for α₁, α₂ and τ_ij, zero values are used by SuperPro.