Publikationsrepositorium - Helmholtz-Zentrum Dresden-Rossendorf

1 Publikation

Quality assurance for chemical databases using the example of the thermodynamic reference database THEREDA

Bok, F.; Moog, H. C.; Gaona, X.; Freyer, D.; Wissmeier, L.

Abstract

Quality assurance for thermodynamic databases is essential to increase confidence in predictive models based on such data. Demands are numerous and range from the review and selection of data, their correct input into a digital database, internal calculations and conversions, to the output of the data to the users - as formatted- parameter files specific for geochemical modeling programs as well as a display on a website with open access to the public. Both technical procedures and reciprocal audits by the editors are used in this process.
An essential part of data quality checks are comparative calculations against experimental measurements. This is accompanied by internal, automated test calculations before each new data release. Moreover, the uncertainty information is given for all data, which are also evaluated and accordingly labelled in terms of quality using a simple scheme. All this information helps users to evaluate the sensitivity of individual parameters and their influence on the result of geochemical calculations in their work.
Maximum transparency is also intended to demonstrate the trustworthiness of the data: For each datum included, the primary source is given in detail, and the results of all comparative calculations are openly presented on the project website (https://www.thereda.de/).
The reliability of the database also includes its long-term availability. THEREDA's system is completely based on free and open source software and extended by self-developed, fully documented components. This guarantees independence from software manufacturers and at the same time ensures the future availability of the data.

Keywords: thermodynamic database; THEREDA; quality assurance

  • Eingeladener Vortrag (Konferenzbeitrag)
    BASE-Workshop „Trust in Models“ (TiM 2022), 17.-18.11.2022, Berlin, Deutschland

Permalink: https://www.hzdr.de/publications/Publ-35047