[Steve Fraser]

Hi
I'm a reasonably experienced, if occasional, user of CFX. I normally use CFD to predict dispersion of pollution from stacks attached to unconventional buildings that more basic dispersion models can't adequately consider, or where the source is located in complex terrain e.g. in a quarry with steep sided walls, or in a river valley combined with complicated building layouts. So pretty much all, simple, one domain type problem. So imagine my consternation when it turns out I need to use domain interfaces.

My problem is to examine how air pollution from an underground car park will affect occupants of the building above.

So, I've created two domains, one for the garage and one above ground, with two interfaces: one (pathway in) is the domain interface between the above ground and the garage; and the second pathway out. These are shown in the pdf attached. The flow is meant to be drawn through the open door down through the car park by four fans and exhausting from a chimney structure with brick vents on the outer wall.

The solver reports that there are 6 isolated regions. The error message suggests that the problem may lie with the domain interface setup, but I've checked and rechecked the setup. The model runs when the expert parameter for checking for isolated regions is set to f, but the solution is incorrect.

I'd be grateful for any advice about what to try next.

Any suggestions

[Opaque]

Have you visualized the 6 isolated volumes the software reported? I did not see them in the pdf file.

Hopefully, that will help to understand where the problem is.

[Steve Fraser]

Hi
Thanks for the response. I did look at the isolated volumes in the post processing, which helped me identify where the problem isolated regions were located. I've figured out the problem: it was due to do with the way I had constructed the below ground elements in design modeller. I've extruded a mold below ground and cut the elements from that. Now running sweetly.