Simulating wind driven natural ventilation

[Anonymous]

Hi there!

I am a Building Science student at Victoria University of Wellington, and I am looking to use Simulation CFD to simulate wind driven natural ventilation. I am new to CFD software and so have been following the help files and online tutorials on how to do this. Generally they've been really useful, however I've come across a bit of an issue.

Using Revit to make the geometry, my initial test was to set up an external study of wind speeds and pressures at the external faces of the building. To do this I followed the "AEC best practices" instructions for natural ventilation here: http://help.autodesk.com/view/SCDSE/2014/ENU/?guid=GUID-F6011744-B69B-47D0-BB2B-50E1BA8DC32D

These said to set up the model in a large void space, with the dimensions shown on the screen capture in attachment 1. The air inlet into the void was defined by assigning a velocity and a temperatre as boundary conditions, and the outlet was defined with a pressure of 0. Slip symmetry was also assigned to all "open" faces in the void. I took a screen capture of the boundary conditions as well- they're shown in attachment 2. I assigned a wind speed velocity of 20m/s and a temperature of 23C.

However, after the simulation runs, the change in velocity of teh air is measured in m/s x10^-9 which seems unrealistic.

Is there something I have missed in the process? Or have I assigned some of these things incorrectly? I've attached a .cfz support file as well.

Cheers,

Riley

[Jon.Wilde · Accepted solution]

Hi Riley,

I can make a few points:

1. Suppress the conrete volumes around your air domain - internal boundary conditions are a no-no
2. Don't have a slip condition on your outlet (I think it might actually do nothing rather than cause an issue but just incase)
3. Extend the outlet - it needs to be far longer, enough so that there is no recirculation over the P=0
4. I would suppress any buildings where you don't need to consider heat transfer through the walls (there are no heat loads so maybe you can suppress all walls?

Hope that helps,

Kind regards,

Jon

[Anonymous]

Ah thank you Jon!

I think assigning the bounday condition on the internal side of the wall was the big issue, but thanks for the other points as well- I'll keep them in miond in future. Suppressing components from the mesh was a handy tool to learn!

Cheers again