Difference between 2D and 3D simulation results

Tuck · June 17, 2016, 01:26

Hi all,

I am running simulations on racecar wings. I have done both 2D and 3D simulation. 2D simulation is done with "Fluid Flow(Fluent)" while 3D simulation is being done in "Fluent (with TGrid meshing)".

I now face a problem whereby the 2D results do not correspond with the 3D results. As I increase the AOA of the wings, the 2D results show flow seperation while the 3D results remain attached and gives me more downforce. Both simulations I run include the same wing geometry, only difference is that the 3D simulation includes endplates.

My chord length of the wing is about 1m. Both models were ran using Pressure based, K-Omega SST models, 12.5m/s velocity inlet with a pressure outlet, solved with coupled solvers, all solution controls parameters were default.

For the 2D flow, the mesh was created with a C-shaped inlet, meshed with ansys meshing with sphere of influence, edge sizings and inflation layers. Under solution methods: Gradient = Least Squares Cell Based, Pressure = Second Order, Momentum = Second Order Upwing.

As for the 3D flow, the mesh was created with meshing provided by Fluent(with Tgrid meshing), mesh was created with a size function with small curvature and proximity functions and a refinement box for the rear wing. It was then wrapped and volume meshed, allowing prisms to grow. Solution Methods are default except for TDE and TKR changed to "Second Order Upwind"

I am curious as to what may cause the deviation for both results? And which result should be trusted more? I'll appreciate any inputs and will provide more information if needed. Thank you.

Tuck · June 19, 2016, 23:52

Bump. Can someone enlighten me about why there is a difference between the two methods? Because there seem to be no point in doing 2D analysis if the difference would be different in 3D anyway.

Bodo1993 · January 9, 2021, 18:43

Quote:

Originally Posted by Tuck

Bump. Can someone enlighten me about why there is a difference between the two methods? Because there seem to be no point in doing 2D analysis if the difference would be different in 3D anyway.

Hi, I am wondering if you got an answer to your question above. Thanks

June 17, 2016, 01:26	Difference between 2D and 3D simulation results	#1
Tuck New Member Join Date: Jun 2016 Posts: 4 Rep Power: 10	Hi all, I am running simulations on racecar wings. I have done both 2D and 3D simulation. 2D simulation is done with "Fluid Flow(Fluent)" while 3D simulation is being done in "Fluent (with TGrid meshing)". I now face a problem whereby the 2D results do not correspond with the 3D results. As I increase the AOA of the wings, the 2D results show flow seperation while the 3D results remain attached and gives me more downforce. Both simulations I run include the same wing geometry, only difference is that the 3D simulation includes endplates. My chord length of the wing is about 1m. Both models were ran using Pressure based, K-Omega SST models, 12.5m/s velocity inlet with a pressure outlet, solved with coupled solvers, all solution controls parameters were default. For the 2D flow, the mesh was created with a C-shaped inlet, meshed with ansys meshing with sphere of influence, edge sizings and inflation layers. Under solution methods: Gradient = Least Squares Cell Based, Pressure = Second Order, Momentum = Second Order Upwing. As for the 3D flow, the mesh was created with meshing provided by Fluent(with Tgrid meshing), mesh was created with a size function with small curvature and proximity functions and a refinement box for the rear wing. It was then wrapped and volume meshed, allowing prisms to grow. Solution Methods are default except for TDE and TKR changed to "Second Order Upwind" I am curious as to what may cause the deviation for both results? And which result should be trusted more? I'll appreciate any inputs and will provide more information if needed. Thank you. yoser and sprengeler like this.

June 19, 2016, 23:52		#2
Tuck New Member Join Date: Jun 2016 Posts: 4 Rep Power: 10	Bump. Can someone enlighten me about why there is a difference between the two methods? Because there seem to be no point in doing 2D analysis if the difference would be different in 3D anyway.