A machine-learned expression for the excess Gibbs energy
This work addresses a long-standing problem in chemical engineering and chemistry for modeling thermodynamic properties, with incremental improvements through novel integration techniques.
The authors tackled the challenge of predicting the excess Gibbs energy for multi-component liquid mixtures from molecular structures by integrating physical laws as hard constraints in a neural network, resulting in HANNA, which outperforms state-of-the-art methods in accuracy and scope.
The excess Gibbs energy plays a central role in chemical engineering and chemistry, providing a basis for modeling the thermodynamic properties of liquid mixtures. Predicting the excess Gibbs energy of multi-component mixtures solely from the molecular structures of their components is a long-standing challenge. In this work, we address this challenge by integrating physical laws as hard constraints within a flexible neural network. The resulting model, HANNA, was trained end-to-end on an extensive experimental dataset for binary mixtures from the Dortmund Data Bank, guaranteeing thermodynamically consistent predictions. A novel surrogate solver developed in this work enabled the inclusion of liquid-liquid equilibrium data in the training process. Furthermore, a geometric projection method was applied to enable robust extrapolations to multi-component mixtures, without requiring additional parameters. We demonstrate that HANNA delivers excellent predictions, clearly outperforming state-of-the-art benchmark methods in accuracy and scope. The trained model and corresponding code are openly available, and an interactive interface is provided on our website, MLPROP.