Investigating Galilean invariance in CFD
When characterizing a body moving in a quiescent flow, like the case of a cruising aircraft, CFD simulations and wind tunnel tests rely on the Galilean invariance principle, which assumes that the moving body can be equivalently modelled by a stationary body with a freestream velocity equal to the m...
Saved in:
Main Author: | |
---|---|
Other Authors: | |
Format: | Thesis-Master by Research |
Language: | English |
Published: |
Nanyang Technological University
2023
|
Subjects: | |
Online Access: | https://hdl.handle.net/10356/164694 |
Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
Institution: | Nanyang Technological University |
Language: | English |
Summary: | When characterizing a body moving in a quiescent flow, like the case of a cruising aircraft, CFD simulations and wind tunnel tests rely on the Galilean invariance principle, which assumes that the moving body can be equivalently modelled by a stationary body with a freestream velocity equal to the motion velocity of the body. However, the validity of this principle for fluid dynamics has not been comprehensively examined before. To this end, this thesis considers numerical simulations of both scenarios and compares the wake and drag coefficient at two Reynolds numbers of 40, where the flow produces a steady wake, and 400, where an unsteady laminar von Karman vortex street is present. The steady laminar regime represented by the lower
Reynolds number shows negligible differences in drag coefficient of the two frames. However, visible differences in the wake of the cylinder were observed. For the unsteady laminar regime represented by the higher Reynolds number, the mean drag coefficient calculated in both reference frames differs by approximately 6%. Additionally, the wake in both frames differs in characteristics and length. Further numerical investigations confirm that these differences are not due to added mass or compressibility effects, thus challenging the assumed invariance of the two reference frames. Future experimental validation is required to prove that the discrepancies are physical and not just numerical artefacts. The Galilean invariance assumption should also be investigated for turbulent flow regimes, for which the change of reference frames could possibly amplify the differences found here due to the large number of timescales involved. |
---|