Thermodynamics-Consistent Graph Neural Networks

TL;DR

This paper introduces a thermodynamics-consistent graph neural network model that accurately predicts activity coefficients in binary mixtures by directly modeling the excess Gibbs free energy, ensuring physical consistency without extra constraints.

Contribution

The novel GE-GNN architecture guarantees thermodynamic consistency by predicting excess Gibbs free energy and deriving activity coefficients through automatic differentiation, avoiding additional thermodynamic constraints.

Findings

High accuracy in activity coefficient predictions

Ensures thermodynamic consistency inherently

No extra loss terms needed for thermodynamic constraints

Abstract

We propose excess Gibbs free energy graph neural networks (GE-GNNs) for predicting composition-dependent activity coefficients of binary mixtures. The GE-GNN architecture ensures thermodynamic consistency by predicting the molar excess Gibbs free energy and using thermodynamic relations to obtain activity coefficients. As these are differential, automatic differentiation is applied to learn the activity coefficients in an end-to-end manner. Since the architecture is based on fundamental thermodynamics, we do not require additional loss terms to learn thermodynamic consistency. As the output is a fundamental property, we neither impose thermodynamic modeling limitations and assumptions. We demonstrate high accuracy and thermodynamic consistency of the activity coefficient predictions.