The second industrial fluid properties simulation challenge

The industrial fluid properties simulation challenge was established in 2001 to provide a realistic assessment of the value of molecular simulation methods for predicting thermophysical properties of industrially important fluids. The organizing committee (the authors of this paper) wished to establish reliable comparisons between the available methods, to assess the state of the art, and to enhance alignment of academic efforts with industrial needs. The first contest was held in 2002. Commercial modeling companies, academic groups and government laboratories were challenged to predict vapor-liquid equilibria, densities, and viscosities for a specified set of organic fluids, mixtures and aqueous solutions. Based on the success of that endeavor a second contest was held, concluding in September 2004. Modeling groups from around the world attempted to predict vapor pressure and heats of vaporization, Henry's law constants, and heats of mixing using molecular simulation methods (the focus of this contest). The contestants applied a wide range of different methods, and different forcefields. Accurate benchmark values were obtained, based on experimental data, by a team from NIST and Dow Chemical and used to assess the accuracy of the predicted values. Predictions of Henry's constant were judged sufficiently accurate to be of value in an industrial environment. The results for vapor pressure and heats of vaporization were mixed. Reasonable qualitative predictions of heats of mixing were obtained for an organic/organic mixture. But results for aqueous solutions revealed an area where although, for the most part, the methods by which predictions were made are sound, the forcefield descriptions are inadequate.