topic Re: Validation of NACA 4412 Hydrofoil in CFD Forum

Validation of NACA 4412 Hydrofoil

Anonymous — Wed, 17 Apr 2019 10:43:23 GMT

Hi there, I am trying to validate the NACA 4412 hydrofoil in a sea water domain, but my lift and drag coefficients are not even close.

I have followed the Autodesk airfoil modelling webinar and other forum posts and used all these tips.

I would appreciate if someone could quickly review my setup and let me know if they have tips to improve the validate results.

CAD: Sketched the NACA4412 hydrofoil using Autodesk Fusion 360 with airfoil sketch from file plugin. Closed tip with a small tangent arc. Sketched 5 refinement regions scaled outwards from the foil. Patched the sketches into 2D surfaces and simplified using Simulation "simplify" then opened in CFD.

Material:

Sea water at 293K
Fixed conditions (not variable)

Boundary conditions:

Inlet: Velocity V=0.68 m/s (Re = 563,000)
Angle of attack: 4 degrees (velocity components modified at Vx=V*cos(AoA), Vy=V*sin(AoA) )
Top/bottom edge slip symmetry
Outlet pressure 0Pa gage

Initial conditions:

Same as inlet boundary condition for whole domain

Turbulence:

SST k-omega RC Hellsten

Meshing:

Hydrofoil edge element length 0.25cm.
First region surface element length 0.5cm.
Following surface regions 1cm, 2.5cm, 5cm, 15cm, 50cm in outer region
Refinement region following wake to outlet, surface length 2.5cm
Layer factor: 0.84
Layer gradation: 1.75
Wall layers: 15
After 1000 iterations max y+ is 0.009
I've used all the flags in the Autodesk webinar e.g. resid_bdry_force_calc = 1
Total nodes: 80,000

Solver:

Convergence - tight, + 1 order of magnitude (1e-6/0.0001/0.0001/1e-7) for the four convergence settings
Incompressible flow

The best results I have achieved is within 10% error of one coefficient, but then 50%+ error in the other. And vice versa.

Are there some specials settings that should be changed when modelling in a water domain versus an air domain?

Could someone please let me know what I'm doing wrong here, or how to improve the results?

Thanks, Alastair

Mesh:

Re: Validation of NACA 4412 Hydrofoil

Anonymous — Wed, 17 Apr 2019 18:42:54 GMT

I've refined the mesh in each region to a total of 230,000 nodes, and only a very small increase in accuracy.

Nodes	CD Error	CL Error
80000	21%	28%
226000	18%	22%

The finer mesh takes already 5 hours to converge so I'm not sure whether increasing the mesh further is a feasible solution...

Re: Validation of NACA 4412 Hydrofoil

frederic.gaillard.7 — Thu, 18 Apr 2019 18:16:36 GMT

I see only 3 layer in this picture

Re: Validation of NACA 4412 Hydrofoil

Anonymous — Thu, 18 Apr 2019 21:00:52 GMT

No, I have 15 layers. Most of them are very close to the wall since I used the layer gradation of 1.75. I decided on this so it produces very thin layers near the wall and a good transition from the size of the rectangular cells to tetrahedral cells in the first region.

I would like to use 30 or 60 layers, but when I set this mesh_enhance_layers=60 flag in the Flag Editor, there is suddenly no boundary layer at all. Just tetrahedral mesh right to the surface. I wonder if this is a software glitch or if I've made a mistake?

Re: Validation of NACA 4412 Hydrofoil

Anonymous — Mon, 22 Apr 2019 07:50:52 GMT

Update on this, I have now got 60 wall layers working by lowering the gradation factor.

After 6000 iterations the results are 30% off for lift coefficient and 25x high for drag. This is baffling! The velocity contours look fine for what is expected at 0 deg AoA. I've tried the sst-kw DES turbulence model this time as I'm modelling water. Does anyone have any advice?

Re: Validation of NACA 4412 Hydrofoil

frederic.gaillard.7 — Fri, 26 Apr 2019 17:17:45 GMT

Can you share your CFZ with us ?

thx

Fred

Re: Validation of NACA 4412 Hydrofoil

Anonymous — Fri, 26 Apr 2019 17:36:09 GMT

Hi Fred,
Here is my CFZ file for my latest model. I'm now using SST k-omega turbulence model because it seems to produce better results for this 500,000 reynolds number flow with water.

Re: Validation of NACA 4412 Hydrofoil

Anonymous — Fri, 26 Apr 2019 17:38:05 GMT

Hi Fred,

Here is my CFZ file for my latest model. I’m now using SST k-omega turbulence model because it seems to produce better results for this 500,000 reynolds number flow with water.

Re: Validation of NACA 4412 Hydrofoil

frederic.gaillard.7 — Fri, 26 Apr 2019 17:46:25 GMT

Thx

Re: Validation of NACA 4412 Hydrofoil

Anonymous — Fri, 26 Apr 2019 17:46:18 GMT

Sorry about that, here it is again

Re: Validation of NACA 4412 Hydrofoil

frederic.gaillard.7 — Thu, 09 May 2019 14:32:15 GMT

Hello @Anonymous ,

Have you try to uncheck the box ''intelligent wall formulation'' in the advanced turbulence parameter.

If your log law region is resolved by the mesh this button must be unchecked.

https://www.simscale.com/forum/t/what-is-y-yplus/82394

Fred