Wallfunction problem RANS Spalart Allmaras
 

Hi everyone

I'm simulating a steady vortex generated by a delta wing towards wind.
I think I have problems with the initial conditions and it respective wallfunctions. What I need is that only the lower face (floor) and the wing surface to have a wallfunction for the lower face a wallfunction that simulate the snow roughness and a very soft wallfunction for the surfaceof the wing. But all the other surfaces (back,front and up (ceiling)) to be like open sky wind as the inlet outlet direction. I'm attaching the 0 folder and Boundary file.

Please help me if for my thesis.

Thanks!

Please guys please help me out with this one!

Can you please be a bit more clear.

As I understand you want to simulate the flow over a delta wing in ground effect . Am I right? it is just the wing and the ground and everywhere else it is free air.

Since you are using spalartAllmaras, you need wall functions only for the nut field.

I am really I cannot help you further unless you can tell me what exactly the problem is :-(

Kalyan

Hi kalyangoparaju

You're totally right, What I want to do is to find a realtion between Re and the cone formed by the generated vortex over the delta wing. So what I really need is to simulate the vortex generated by the delta wing so I can use the field data of p and U at some distance from the delta wing. So the only "walls" that are interacting is the ground and the delta wing everything else is free air.

I haven't been able to converge any simulation with spalart, it always crashes. I think my mesh it's quite accurate, i'm not intrested on the boundary layer so i'm using a high y+ value. The gamma and rho indicators of the mesh (i'm using gmsh) are over the 0,7 with minimun values over the 0,5.

I don't know what i'm missing here. Please helpme out

I had very very similar problem when I tried to simulate a roundtip airfoil inside a windtunnel and analyze the generated vortex.. there are couple of things to keep in mind.

1. It is a good thing to build the boundary layer on atleast the wing. Even though not having it might not lead to a blownup solution, I strongly suggest to build layers if you are analyzing vortices.

2. What order are you running? It is suggestible to drop down to first order and see what is happening .

3. Do checkMesh and see if FOAM thinks your mesh is compatible with it or not. If not, some changes might be required.

4. Initialize the solution by running potentialFoam. That can sometimes stabilize the solution.

5. Last but most probably something that might work - Decrease the relaxation factors ( to around 0.3 for all variables)

Hope this helps. If not, can you please share how exactly the solution blows up? Is it after 1-2 iterations or do the residuals gradually rise and diverge?

Kalyan

1. The problem is that i don't really know how to build the boundary, and if i make the cells so tinny in the wing the indicators go far away, because i don't know how to manipulate the cells so they go very little from the inside wheres anything to the outside growing in size.

2. I don't understand what do you mean by order. It's a 3D simulation if that's helpful.

3. I think the mesh is compatible because I passed to openfoam with the gmshToFoam function. And I don't know how to use checkmesh really.

4. Do you mean to first run potentialfoam and then with that solutions run a simplefoam?

5. I don't think it would be helpfull, I think it's a problem of parameters, I have just read in Verteeg-An indtroduction to CFD that for high reynolds numbers nut and nutilda are equal

Can we talk by skype? my user name is rafael.valenzuela.musura

thank you very much

4.

here is the error file, this is another simulation where i have reduce the U entrance to 4,78.