Thermal comfort on room - cooling design - Nozzle boundary condition

[matthieu.grivelet-INEX]

Hello all,

I try to simulate the air distribution system of a theater

In these circumstances, I realize the 3D model of the room and started some studies. 

To make it simple, i have some nozzle and air suction hole along airs ducts.

The model works, but  I have a problem of boundary condition.

I don't know if my models is correct and where put boundary condition.

My question is can we put inlet (air velocity in) and outlet (air velocity out) or 0 pressure boundary condition on air materials but in contact on a solid matérial ?

I m not realy sure that possible

Thank you for your answer

[Jon.Wilde]

Hi Sylvie,

Correct, we cannot have internal boundary conditions. If you need to do this, it should be attached to a suppressed solid part (to create an internal void).

Thanks,

Jon

[matthieu.grivelet-INEX]

Thank you for your answer,

I made ​​the changes according to your advice.

The model works, the temperature distribution and air velocity appears logical

However, the temperature is too hot (80-120°C and more) in the room.

I think, the issue comes from the fact I have 30 kw by the cooling systems against 32 kW from people + 30 kW from equipment in an adiabatic room in stationaryoperating conditions.

Can we include the notion of time in the calculation ?

Or can you tel me where is my error ?

Thank you

[matthieu.grivelet-INEX]

This is my model if you need :

Dowload.zip

Thank you

[Jon.Wilde]

No problem. Are the internal BC's now attached to suppressed volumes? It looks to me like the solid parts are still meshed, so you still have internal conditions?

Try refining the mesh, it looks a little coarse.

Yes, this is essentially the result after an infinite time. You could run transiently, check out the help 🙂

[Jon.Wilde]

Hi,

You still have internal boundary conditions. I suggest you suppress (from the mesh) the parts that these are attached to. Even better would be to have them recessed a little, so that they represent reality more.

Kind regards,

Jon

[matthieu.grivelet-INEX]

The BC that I incorporated are detailed below

It not a supressed matérials (I did not quite understand how to do), the matérials is just air in a pipe whose end is neutral pressure.

I reduced a little mesh size

I run  transiently with 60 second (the target is to simulate during 5 hour with 300 time step) 

Inner iteration is define to 10.

My BC are the sames but with an transient constant fonction (i think is the same thing that to leave them in permanent condition)

Initial condition is 21°C for all my materials.

The results is wrong

[Jon.Wilde]

Check out meshing in the help, that will detail it 🙂

You 100% need the boundary conditions attached to a suppressed volume. It is also much better to have the boundary recessed into a tube - as reality.

How about running steady state until you have the flow converged and then re-initialising initial conditions and running thermal only (flow off so it remains fixed at the end of the steady state solution) for the transient solution? Then you can run with timesteps like 60s.

If you want to run flow and thermal together this you will need a much smaller timestep (0.25 or 0.5s maybe)

Hope that helps!