[Antoine_faye]

Hello,


I am working on a project where I need to simulate (in OF v21.06) a body moving in a two phases turbulent flow. This body will be defined in an overset mesh and wall functions are used on the body.

I ran some simple cases in 2D with only 1 fluid (to do so, I set alpha.water=0 in all the domain). My test case is a cylinder (in an overset mesh) in a domain (background mesh)(overset_mesh.png). I firstly ran a simulation where the meshes (background+overset) are fixed and submitted to a uniform flow (Re=1000). In this case, the flow is continuous and correctly simulated (fixed_cylinder.png), we can even see the formation of a Von Karman vortex street.


In my second test case, the meshes where moving in a fluid without velocity. This simulation should be equivalent to the first one but it is not. We can see some discontinuities (for the pressure) between the overset mesh and the background mesh, so the flow is not correctly simulated (moving_cylinder_t_0.5s.png and moving_cylinder_t_3s.png).


To impose the motion of the meshes, I used the solidBody solver of the dynamicOversetFvMesh. The solid body motion is imposed with a timeDataFile (cylinderMotion.dat). I also modified the U boundary condition for the cylinder, in the fix case, it was fixedValue (=0) and in the dynamic case, it is now movingWallVelocity. The only other change I made was in the fvSolution, I added the following lines in the PIMPLE section:


ddtCorr yes;
momentumPredictor no;
correctPhi yes;
oversetAdjustPhi yes;


I investigated this problem but I can't find any solution, is there any mistake in my set up (which is linked to this post) ?

In the constant/polyMesh, there is only the boundary file, otherwise, the fileSize of my project would have been to important.



Thanks,
Antoine

[Mid_Night]

Did you manage to deal with overset mesh?

[shishirgautam]

With the new release, we don't need some of those keywords like ddtCorr, correctPhi, and oversetAdjustPhi in fvSolution file(see this link for more info: https://www.openfoam.com/news/main-n...v2212/numerics).