An investigation of anisotropic RANS turbulence closures for the heat transfer prediction in ribbed cooling passages
Seiten
2021
Dr. Hut (Verlag)
9783843948005 (ISBN)
Dr. Hut (Verlag)
9783843948005 (ISBN)
- Keine Verlagsinformationen verfügbar
- Artikel merken
In this thesis, closures for the Reynolds-averaged momentum and energy conservation equations are investigated, with the goal of improving the heat transfer prediction accuracy in ribbed cooling passages for gas turbine blades in comparison with simple models commonly used in industrial applications. Most of the Reynolds stress and scalar flux models are explicit and algebraic, and are able to accurately represent the Reynolds stresses in the vicinity of solid walls. The models have been implemented in the open source CFD code OpenFOAM. They are evaluated against the current industry standard for the simulation of complex flows with heat transfer, the k-omega SST model using a constant turbulent Prandtl number.
The first part of this work is focused on conventional closure methods. The turbulence models have been investigated using several periodic channel cases of increasing complexity. This study identified a model combination which provides improved heat transfer predictions with respect to the k-omega SST model and the turbulent Prandtl number. The investigation of a realistic geometry for a ribbed cooling channel using the new model combination showed a general good agreement with the experimental data.
The second part of this thesis is focused on a novel modeling concept based on machine learning algorithms. It is applied here to derive a tensor representation for the turbulent thermal diffusivity and a nonlinear extension for eddy-viscosity models. The former has been applied in planar channel flows only, since at the time of its creation reference data for more complex flows were not available. The second has been tested in planar channels, a two-dimensional ribbed channel and a square duct. The scalar flux model and the nonlinear extension for eddy-viscosity models provided predictions in good agreement with the reference data.
The first part of this work is focused on conventional closure methods. The turbulence models have been investigated using several periodic channel cases of increasing complexity. This study identified a model combination which provides improved heat transfer predictions with respect to the k-omega SST model and the turbulent Prandtl number. The investigation of a realistic geometry for a ribbed cooling channel using the new model combination showed a general good agreement with the experimental data.
The second part of this thesis is focused on a novel modeling concept based on machine learning algorithms. It is applied here to derive a tensor representation for the turbulent thermal diffusivity and a nonlinear extension for eddy-viscosity models. The former has been applied in planar channel flows only, since at the time of its creation reference data for more complex flows were not available. The second has been tested in planar channels, a two-dimensional ribbed channel and a square duct. The scalar flux model and the nonlinear extension for eddy-viscosity models provided predictions in good agreement with the reference data.
| Erscheinungsdatum | 01.07.2021 |
|---|---|
| Reihe/Serie | Thermodynamik |
| Verlagsort | München |
| Sprache | englisch |
| Maße | 148 x 210 mm |
| Gewicht | 281 g |
| Themenwelt | Naturwissenschaften ► Physik / Astronomie ► Thermodynamik |
| Technik | |
| Schlagworte | Neural • OpenFOAM • Turbulence |
| ISBN-13 | 9783843948005 / 9783843948005 |
| Zustand | Neuware |
| Informationen gemäß Produktsicherheitsverordnung (GPSR) | |
| Haben Sie eine Frage zum Produkt? |
Mehr entdecken
aus dem Bereich
aus dem Bereich
zur Materialproblematik der modernen Elektronik
Buch | Hardcover (2025)
Alfred Kröner Verlag
CHF 27,85
Von Energie und Entropie zu Wärmeübertragung und Phasenübergängen
Buch | Softcover (2023)
De Gruyter Oldenbourg (Verlag)
CHF 83,90
