This project aims to enhance open-source implementations of COSMO-RS by refining the representation of physical interactions and optimizing computational algorithms for solving its core equations.
Accurate thermodynamic modeling is crucial for phase equilibrium predictions in chemical and biochemical engineering. This project improves open-source implementations of the COSMO-RS model, focusing on both improving its physical accuracy and computational efficiency.
To enhance efficiency, our recent work introduces a novel algorithm for solving COSMOSPACE equations, significantly accelerating calculations. On the physical side, we focus on dispersion interactions, a critical but challenging aspect of molecular interactions. We explored multiple approaches with varying degrees of reliance on experimental data fitting.
Our latest development employs a first-principles quantum chemical descriptor—atomic polarizabilities—which, when combined with a minimal set of general model parameters, substantially improves phase equilibrium predictions. This is particularly impactful for halocarbon systems, which are highly sensitive to dispersion effects and notoriously difficult to model with COSMO-RS.
Starting with halocarbons, we extend our work to a diverse range of chemical systems, broadening the applicability and predictive power of COSMO-RS.
Main supervisor: Wei Yan
Co- supervisor: Erling H. Stenby