Why turbulence-model at very high Re-numbers?

[thschrader]

Hi,
did some cfd test-runs with my little house.
goal: compute the resulting drag-force at 115 km/h wind-speed
according to german DIN-manuals. When using cfmesh/mesh-refinement and
k-omega turbulence model, I get a resulting drag force of approx 37,4 kN.
When doing the same with snappyhexmesh/mesh-refinement and NO turbulence model,
I get nearly the same force. Checking the results "by hand" gives a good accordance.
In reality there is turbulence (turbulence-intensity behind house approx 20%), but
in the models there is practically no difference. So why should I use time-consuming
turbulence calculations?
Comments welcome.
Thomas

turbulence.JPG (23.63 KiB) Viewed 796 times

resulting_force_x.JPG (118.42 KiB) Viewed 796 times

cfmesh.JPG (119.31 KiB) Viewed 796 times

[oliveroxtoby]

Hi,
did some cfd test-runs with my little house.
goal: compute the resulting drag-force at 115 km/h wind-speed
according to german DIN-manuals. When using cfmesh/mesh-refinement and
k-omega turbulence model, I get a resulting drag force of approx 37,4 kN.
When doing the same with snappyhexmesh/mesh-refinement and NO turbulence model,
I get nearly the same force. Checking the results "by hand" gives a good accordance.
In reality there is turbulence (turbulence-intensity behind house approx 20%), but
in the models there is practically no difference. So why should I use time-consuming
turbulence calculations?
Comments welcome.

Off the top of my head: Turbulence modelling will mostly influence how accurately the friction drag is calculated. However, for a big blunt body like a house, the force will mostly be composed of a pressure drag component, with the friction drag being less significant. That would typically not be the case for a car or an aircraft.