Butterfly outdoor airflow in 2d and 3d


I am pretty new to Butterfly and I am looking at the outdoor airflow at ground level around a height of 1.1 meters. I was originally planning on using the Butterfly example file of outdoor airflow in 2d as the calculation time is significantly shorter than the 3d example. However, when comparing both of them, with the same legend parameters and same windspeed input for the wind tunnel, the results in 2d have blue arrows and lower windspeed around the edges of the left building while the 3d one has different results as you can see in the illustration.

My interpretation of this is that the 2d file is not considering the vertical variable of the wind leading to smaller wind vectors when the wind is directed more vertically. So if I am looking at the ariflow in a park or between these two boxes I need to look at the 3d one which is more reliable, am I right on this ?

Hi @skh9892

I would definitely use 3D CFD models for this one. I typically used 2D simulations in smaller contexts, something like openings in a facade or a singular panel, and even then it was more to understand flow than validate / properly simulate.

I’m sure there might be ways for you to get better results than what is in the image, but I don’t think you’ll ever capture the effects properly. The model you are showing, or those types of inputs, should not take too long anyways (well in CFD terms :slight_smile: ).


Thanks a lot for your reply
Would you know why on the left image the wind is faster on the sides of the grid ? Because I can’t think of any reason why this is happening

My first question would be: are you sure the scales are the same?

Between the both of them ? yes they are
However, I decided only to go with the 3d one on the left but I don’t understand why the wind is getting faster on the edge of the grid, it’s as if the wind is coming from the center and is realeased from a less large “corridor” increasing the speed on the sides

Around the building edges at the windward side, the speed should be higher as per illustrated as there are flow separation.

For your domain boundary conditions, I would advise you use a larger domain as It is quite weird for the boundary to be accelerated in terms of air flow.

Perhaps you can check the turbulence model that you are using, perhaps checking to use k-epsilon model (standard or realizable). External wind have a logarithmic wind profile, so wind speed at low heights are low, and it will be higher at higher altitudes/elevation.