The fourth industrial fluid properties simulation challenge

Abstract

a b s t r a c tThe primary goal of the fourth industrial fluid properties simulation challenge was to test the transferability of molecular simulation methods and intermolecular potentials (force fields) when applied to a wide variety of physical properties for a given industrially relevant small molecule. Force field parameters are often developed for, tested with, and applied to a relatively limited range property types. Methods that are able to predict a broad range of properties may be more useful for practical application in an industrial setting. To address this primary goal, entrants were challenged to develop new intermolecular potential models for ethylene oxide (EO) and to us they predict a variety of physical properties. The predictions were judged by comparison to a set of benchmark values established based on peer-reviewed scientific literature and highly regarded property databases, and the results were generally excellent. A secondary objective was to assess the variability of those property predictions when performed by different researchers using the same force field (a "round-robin" experiment). To address this goal, entrants were challenged to calculate the same set of properties for a specific EO potential model from the literature, and good consistency was found between the predictions of most of the entrants.

Fundamental equation of state for ethylene oxide based on a hybrid dataset

“…The residual Helmholtz energy was determined by Widom's test particle insertion (Widom, 1963). The simulations were based on a molecular interaction model by Eckl et al (2008), which won the Fourth Industrial Fluid Properties Simulation Challenge in 2008 (Case et al, 2008). It consists of three Lennard-Jones (LJ) sites and a point dipole located at the geometric center of the molecule.…”

Gibbs ensemble Monte Carlo simulation using an optimized potential model: pure acetic acid and a mixture of it with ethylene

“…With an accurate intermolecular potential, it is possible to predict vapor-liquid equilibria of pure fluids and mixtures with great accuracy in many cases [2][3][4][5][6][7][8][9]. Substantial efforts have been made in recent years to study the vapor-liquid equilibrium properties of relevant fluid systems [10][11][12][13][14]. Many molecular simulation methods, such as Gibbs ensemble Monte Carlo [15], GibbsDuhem integration [16], histogram reweighting grand canonical Monte Carlo [17], and the NPT + test particle method [18], have been proposed.…”

The fifth industrial fluid properties simulation challenge

“…The first four challenges are described in earlier issues of this journal [1][2][3][4]. The background and history of the challenges has been described earlier [4] and will not be repeated in detail here.…”

“…The background and history of the challenges has been described earlier [4] and will not be repeated in detail here. In summary, the simulation challenge is aligned with the vision of the Industrial Fluid Properties Simulation Collective:…”