In spite of the numerous recent methodological advances in simulation methodologies, an engineering-oriented framework has not been established. If successful, the proposed work will provide general guidelines and specific algorithms to construct such an engineering-oriented framework. In the short term, this work will provide efficient methods that scientists and engineers can use to generate any type of phase diagram for fluid mixtures that could be needed to understand and unveil a particular phenomena or to facilitate engineering calculations in the design of separation processes. In the long term, the results of this work could pave the way to the development of process simulation software having a molecular simulation "module" for thermodynamic data generation. Applications to the study relatively new systems and processes will unquestionable be explored in the future once better force-fields and faster computers become available.
The first step is the development of methods for the simulations of different classes of phase diagrams through single-stag simulations. Novel Gibbs ensemble variants and novel pseudo-ensemble methods will be developed to accomplished this. Pseudo-ensembles provide a general framework to connect arbitrary phase equilibrium specifications with the parameters needed to simulate a selected ensemble; these methods thus provide a convenient tool to generate arbitrary projections (phase diagrams) of the thermodynamic phase space of the system. The second step involves the development of new approaches to map out specific sub-domains of phase space that are relevant to the simulation of phase equilibria in multi-stage processes for systems with many components. These approaches combine simulation data and engineering thermodynamic models in order to harness the predictive power of molecular simulation and the efficiency of engineering models. Novel quasi-Gaussian analytical models will be investigated which, by design, synthesize more information about the underlying ensemble density distribution function (available from simulation output) than conventional engineering models. Several systems will be investigated to test the usefulness of the proposed methods, in particular, the separation of oligomeric mixtures by supercritical antisolvent fractionation.
