[ajonuz]

Dear openfoamers,

I am having troubles simulating a tube bank heat exchanger unit cell with cyclic boundary conditions. I wish to model a 2D turbulent transient case using buoyantBoussinesqPimpleFoam as the solver and kOmegaSST as the turbulence model. I have attached the picture of the mesh and the boundaries are set up as follows:

- Left Inlet: Cyclic, neighboured to right outlet
- Right outlet: Cyclic, neighboured to left inlet
- Top left: Cyclic, neighboured to bottom right
- Top right: Cyclic, neighboured to bottom left
- Bottom left: Cyclic, neighboured to top right
- Bottom right: Cyclic, neighboured to top left
- Cylinder surfaces: zero velocity and fixed temperature

As it is set up now with an initial velocity, the model runs for less than a second real time. As I can see the velocity drops to 0, it seems like there is no driving force on the fluid. It makes sense because of the zero velocity cylinder surfaces.

My problem is therefore to understand how I can apply a driving force when using cyclic BC on all boundaries except the cylinders. I assume i have to apply a pressure gradient somehow?

The case is attached as a zip file.