Options
Analyzing the link between GE-model parameter regression and optimal process design
Publikationstyp
Conference Paper
Date Issued
2018-06
Sprache
English
TORE-URI
First published in
Number in series
43
Start Page
103
End Page
108
Citation
Computer Aided Chemical Engineering 43: 103-108 (2018)
Contribution to Conference
Publisher DOI
Scopus ID
Publisher
Elsevier
ISBN
978-0-444-64235-6
The importance of accurate thermodynamic models, capable of describing vapor-liquid and liquid-liquid equilibria, is generally acknowledged for the design of chemical processes. However, the parameterization of thermodynamic models and the development of chemical processes are usually treated as two different disciplines. More importantly, the objectives of each discipline do not necessary align. While the quality of a thermodynamic model is often judged purely on the basis of its average mean deviation from experimental data, process design aims at minimizing the overall process costs, assuming that the underlying thermodynamic models (especially model parameters) are reliable/accurate. Assessing the uncertainty of the thermodynamic model parameters is the link between both disciplines. However, this link is rarely established, and even the effect of the consideration of uncertainties in parameter regression by means of a deterministic process design is rarely investigated. Within this work this effect was taken into account for vapor-liquid equilibrium modeling and optimal design of a distillation column. The results indicate the necessity of a consistent consideration of uncertainties both during parameter regression and process optimization. Furthermore, the results highlight that a special treatment of experimental data and uncertainty information is required for specific applications (e.g. distillation of tangent pinch systems) in order to obtain reliable results in process design.
Subjects
Distillation
Optimization
Parameter regression
Thermodynamic equilibria
DDC Class
600: Technology