Chemical engineers need reliable property data for process design and optimization. When available, experimental measurements are the most reliable source for many properties and the basis to validate theoretical calculations or develop predictions. For any of these applications, it is critical to know quantitatively how reliable those values are, i.e., to assess uncertainty. The Thermodynamics Research Center (TRC) at NIST performs such analysis in the course of providing reliable thermophysical and thermochemical property data for pure compounds, binary and ternary mixtures, and chemical reactions. The growth of our data resource and breadth of modern modeling technologies has allowed us to develop and apply new methods and approaches to uncertainty evaluation. Experimental manuscripts often only consider statistical variation, providing only repeatability or what is termed a Type A evaluation in the Guide to the Expression of Uncertainty in Measurement (International Organization for Standardization: Geneva, Switzerland, 1993). Historically, TRC estimates uncertainty from factors such as experimental method, conditions, sample purity, and chemical behaviors of the substance, resulting in uncertainties that often significantly exceed the values reported by the authors or equipment manufacturers. However, this approach only provides an estimate of the lowest possible uncertainty. Systematic review of modern publications shows that actual deviations frequently exceed that level, sometimes by two or three orders of magnitude. Full evaluation of the uncertainty (Type B) requires validation against independent additional measurements (preferably performed with a different method), cross-property consistency, and agreement with theoretical calculations or correlation with similar compounds. Such analysis requires a comprehensive collection of property data and a variety of computational and correlation methods individual for each property. TRC maintains such a data archive (SOURCE) and an expert system (ThermoData Engine, SRD 103b) to interpret it. TRC constantly reassesses data quality as new data and methods become available, and uncertainties are updated as inconsistencies are revealed, in some cases even requiring additional experimentation to resolve. This iterative process allows TRC to support the research community to maximize impact and boost cost-efficiency. The presentation will give examples and discuss such opportunities.

Reliability of data: A meaningful and comprehensive assessment

254th National Meeting (2017)

The official archive and repository for all ACS Meetings & Expositions abstracts and select posters or presentations.