Jump to content

Dortmund Data Bank: Difference between revisions

From Wikipedia, the free encyclopedia
Content deleted Content added
m Fix links to Salt (via WP:JWB)
No edit summary
Line 1: Line 1:
{{Infobox data structure
| name = Dortmund Data Bank
| type = [[Data Bank]]
| invented_by =
| invented_year =
| space_avg =
| space_worst =
| search_avg =
| search_worst =
| insert_avg =
| insert_worst =
| delete_avg =
| delete_worst =
| peek_avg =
| peek_worst =
| find_min_avg =
| find_min_worst =
| delete_min_avg =
| delete_min_worst =
| decrease_key_avg =
| decrease_key_worst =
| merge_avg =
| merge_worst =
}}

The '''Dortmund Data Bank'''<ref>Onken U., Rarey-Nies J., Gmehling J., "The Dortmund Data Bank: A Computerized System for Retrieval, Correlation, and Prediction of Thermodynamic Properties of Mixtures", Int.J.Thermophys., 10(3), 739-747, 1989
The '''Dortmund Data Bank'''<ref>Onken U., Rarey-Nies J., Gmehling J., "The Dortmund Data Bank: A Computerized System for Retrieval, Correlation, and Prediction of Thermodynamic Properties of Mixtures", Int.J.Thermophys., 10(3), 739-747, 1989
</ref> (short '''DDB''') is a factual data bank for [[thermodynamic]] and thermophysical data. Its main usage is the data supply for [[process simulation]] where [[experimental data]] are the basis for the design, analysis, synthesis, and optimization of [[chemical process]]es. The DDB is used for fitting parameters for thermodynamic models like [[Non-Random Two Liquid model|NRTL]] or [[UNIQUAC]] and for many different equations describing pure component properties like e.&nbsp;g. the [[Antoine equation]] for vapor pressures. The DDB is also used for the development and revision of predictive methods like [[UNIFAC]] and [[PSRK]].
</ref> (short '''DDB''') is a factual data bank for [[thermodynamic]] and thermophysical data. Its main usage is the data supply for [[process simulation]] where [[experimental data]] are the basis for the design, analysis, synthesis, and optimization of [[chemical process]]es. The DDB is used for fitting parameters for thermodynamic models like [[Non-Random Two Liquid model|NRTL]] or [[UNIQUAC]] and for many different equations describing pure component properties like e.&nbsp;g. the [[Antoine equation]] for vapor pressures. The DDB is also used for the development and revision of predictive methods like [[UNIFAC]] and [[PSRK]].

Revision as of 15:11, 10 February 2020

Dortmund Data Bank
TypeData Bank
Time complexity in big O notation
Operation Average Worst case
Space complexity

The Dortmund Data Bank[1] (short DDB) is a factual data bank for thermodynamic and thermophysical data. Its main usage is the data supply for process simulation where experimental data are the basis for the design, analysis, synthesis, and optimization of chemical processes. The DDB is used for fitting parameters for thermodynamic models like NRTL or UNIQUAC and for many different equations describing pure component properties like e. g. the Antoine equation for vapor pressures. The DDB is also used for the development and revision of predictive methods like UNIFAC and PSRK.

Contents

Mixture properties

Double Azeotrope of Benzene and Hexafluorobenzene, taken from Dortmund Data Bank

The mixture data banks contain currently (April 2007) approx. 308,000 data sets with 2,157,000 data points for 10750 components building 84870 different binary, ternary, and higher systems/combinations.

Pure component properties

Vapor pressure diagrams of Water

The pure component properties data bank contains currently (April 2007) approx. 157,000 data sets with 1,080,000 data points for 16700 different components.

Data sources

The DDB is a collection of experimental data published by the original authors. All data are referenced and a quite large literature data bank is part of the DDB, currently containing more than 92,000 articles, books, private communications, deposited documents from Russia (VINITI), the Ukraine (Ukrniiti) and other former USSR states, company reports (mainly from former GDR companies), theses, patents, and conference contributions.

Secondary sources like data collections are normally neglected and only used as a literature source. Derived data are also not collected with the main exception of the azeotropic data bank which is built partly from evaluated vapor–liquid equilibrium data.

History

The Dortmund Data Bank was founded in the 1970s at the University of Dortmund in Germany. The original reason for starting a vapor–liquid phase equilibria data collection was the development[2] of the group contribution method UNIFAC which allows to estimate vapor pressures of mixtures.

The DDB has since been extended to many other properties and has increased dramatically in size also because of intensive (German) government aid. The funding has ended and the further development and maintenance is performed by DDBST GmbH, a company founded by members of the industrial chemistry chair of the Carl von Ossietzky University of Oldenburg, Germany.

Additional contributors are the DECHEMA, the FIZ CHEMIE (Berlin), the Technical University in Tallinn, and others.

Availability

The Dortmund Data Bank is distributed by DDBST GmbH as in-house software. Many parts of the Dortmund Data Bank are also distributed as part of the DETHERM data bank which is also available online.

See also

References

  1. ^ Onken U., Rarey-Nies J., Gmehling J., "The Dortmund Data Bank: A Computerized System for Retrieval, Correlation, and Prediction of Thermodynamic Properties of Mixtures", Int.J.Thermophys., 10(3), 739-747, 1989
  2. ^ Gmehling J., Weidlich U., "Die Dortmunder Datenbank. Basis für die Weiterentwicklung der UNIFAC-Methode", Chem.Ing.Tech., 57(5), 447-449, 1985