[sah]

I have a problem in a simulation for a university project,

I have a rectangular canal ( 4*4m), the medium is air. The inlet is set to velocity-inlet and has a velocity of 0,3 m/s , The flow is therefore laminar in the air duct. The outlet is set to pressure-outlet.

In the middle of the canal I have a half-sphere (60cm diameter), which is facing the incoming air from the inlet. This will obviously generate wake behind the flat surface downstream. This would most likely generate turbulence there.

I don’t know two things

1-Should I select laminar flow model or turbulent flow one?
2- Should I select steady or unsteady? or by just selecting the unsteady model will cover all the possibilities?

I am using Star CCM+ in case this is necessary to be mentioned.

Could someone help me by this please?

[me3840]

If you think the flow will be laminar, then laminar should be what you select.

Whether or not you run steady or unsteady depends on what you want out of the simulation. Do you want to find a shedding frequency? Can't do that with steady. Do you want to find just the overall drag? Steady will do okay. Either way you sound pretty new, so I would start with steady, it's far easier to understand when you're just getting started.

[sah]

Thank you for your reply.

Actually, I must measure the velocities at several points in the wake region under a spherical body (a part of a sphere).

I have a room with a laminar flow inlet covering the entire ceiling. And the outlet is a perforated plate outlet covering the entire room’s floor. The inlet and the outlets are properly balanced and adjusted to create a laminar flow in the room. The inlet velocity is 0.3 m/s

In the room there is a spherical body. I have measured the velocity at several points under the body using very accurate anemometers.

I simulated the room using steady stead and turbulence models. But the results were way difference from the measurement results. Actually the results were more than halve the value lower than the results that I measured in the room especially in the lower points, which are the furthest from the body.

The points of interest are shown in the attached picture.

Could you suggest any possible simulation-obstruction that might be causing the results to be way far from the measured values?

[me3840]

Your mesh is not remotely fine enough, especially around the shear layer of the wake. Why would you coarsen the surface of the body as you get closer to the tip where the wake is produced?

Is the bottom the outlet pressure boundary in that picture? That's a very ill posed condition, it's sitting in the wake of the object itself.

[ashokac7]

me3840 is right. Your mesh looks quite course. Do some wake refinement or use volumetric control in wake region. And one more thing is I think you can relate your problem with flow over a cylinder. They use SST model there. (Note- Please search and confirm first).