Physics-informed machine learning for predicting temperature-dependent chemical properties

Physics-informed machine learning model decouples chemistry from thermodynamic state, including temperature, to accurately predict fluid properties across thermodynamic conditions. It enables reliable design of new cooling fluids.