(80d) A Critical Comparison of Manual and Automated Equation of State Production

Chemical process simulation and design hinges on the availability of accurate thermophysical and transport properties for all chemical species present. By far the most useful representation of thermophysical properties is an accurate equation of state. However, the production of accurate equations of state is a complex process that depends on the availability of a large set of high quality, evaluated data. The traditional path to an equation of state is for a data evaluator to produce parameters that fit a complex equation of state to a chosen data set. This method requires a highly experienced thermodynamicist and can be very time consuming. Recently, the Thermodynamics Research Center at the National Institute of Standards and Technology has introduced a software system for the evaluation of thermophysical property data called ThermoData Engine (TDE). While the initial release of TDE included only single property data correlation capabilities along with subsequent consistency checks, TDE 2.0 has the additional capability of automatically fitting all property data for a given compound to a chosen equation of state. A comparison of the performance of TDE's automated equation of state production against a set of equations of state produced at NIST using more traditional methods will be presented. The various performance metrics chosen to evaluate the quality of the fit will be discussed and the results for both equation of state production methods when fitting a specific Span-Wagner equation of state will be given. Additionally, automated equation of state production for compounds for which data are relatively sparse and no current critically evaluated equation of state is available will be presented.