 |
| Titel |
Towards CFD modeling of turbulent pipeline material transportation |
| VerfasserIn |
Amir Shahirpour, Nicoleta Herzog, Cristoph Egbers |
| Konferenz |
EGU General Assembly 2013
|
| Medientyp |
Artikel
|
| Sprache |
Englisch
|
| Digitales Dokument |
PDF |
| Erschienen |
In: GRA - Volume 15 (2013) |
| Datensatznummer |
250077145
|
|
|
|
|
|
| Zusammenfassung |
| Safe and financially efficient pipeline transportation of carbon dioxide is a critical issue in the
developing field of the CCS Technology. In this part of the process, carbon dioxide is
transported via pipes with diameter of 1.5 m and entry pressure of 150 bar, with Reynolds
number of 107 and viscosity of 8Ã10(-5) Pa.s as dense fluid [1]. Presence of large and small
scale structures in the pipeline, high Reynolds numbers at which CO2 should be transferred,
and 3 dimensional turbulence caused by local geometrical modifications, increase the
importance of simulation of turbulent material transport through the individual components
of the CO2 chain process.
In this study, incompressible turbulent channel flow and pipe flow have been modeled
using OpenFoam, an open source CFD software. In the first step, simulation of a turbulent
channel flow has been considered using LES for shear Reynolds number of 395. A simple
geometry has been chosen with cyclic fluid inlet and outlet boundary conditions to simulate a
fully developed flow. The mesh is gradually refined towards the wall to provide values close
enough to the wall for the wall coordinate (y+). Grid resolution study has been conducted
for One-Equation model. The accuracy of the results is analyzed with respect to
the grid smoothness in order to reach an optimized resolution for carrying out the
next simulations. Furthermore, three LES models, One-Equation, Smagorinsky and
Dynamic Smagorinsky are applied for the grid resolution of (60 Ã 100 Ã 80) in (x,
y, z) directions. The results are then validated with reference to the DNS carried
out by Moser et al.[2] for the similar geometry using logarithmic velocity profile
(U+) and Reynolds stress tensor components. In the second step the similar flow is
modeled using Reynolds averaged method. Several RANS models, like K-epsilon and
Launder-Reece-Rodi are applied and validated against DNS and LES results in a similar
fashion.
In the most recent step, it has been intended to generate an optimized LES solver to model
turbulent pipe flow for larger Reynolds numbers. The validations are carried out using
experiments conducted in Cottbus Large Pipe Test Facility at BTU as a reference [3]. In the
mentioned experimental research, evolution of statistical pipe flow quantities, such as
turbulence intensity, skewness and flatness are investigated to clarify the development
length needed to achieve fully developed turbulence. These observations take place
in a relatively large pipe test facility with an inner pipe diameter of Di = 0.19 m
and a total length of L = 27 m where a bulk Reynolds number of 8.5Ã105 can be
reached.
1. CO2 pipeline Infrastructure: An analysis of global challenges and opportunities, Final
Report For International Energy Agency of Greenhouse Gas Program (2010)
2. J. Kim, P. Moin, R. Moser, Turbulence statistics in fully developed channel flow at low
Reynolds number, J.Fluid Mech. 177, 133-166, (1987)
3. F. Zimmer, E.-S. Zanoun and C. Egbers, A study on the influence of triggering pipe flow
regarding mean and higher order statistics, Journal of Physics: Conference Series 318 (2011)
032039 |
|
|
|
|
|
|