Hi @nicholas Thanks for sharing all this. A uniform inlet will not do, certainly not for an urban case. If there is a way to script the logarithmic profile in Autodesk I would look for that (much like one can script it in StarCCM+).
Concerning the results, firstly we should note that the ground is not really flat. We are inputting a ground roughness to the model, most likely 1.0 which reflects a city area. This kind of creates a ground that is not considered flat even though there is no obstacle in the way.
As for the specific accelertion I think Abraham covered it nicely. You generally want to allow for a large extension on the leeward side (after the wind hits the area of interest), I would input a number between 10-15 in the leeward extension on BF windTunnelParameters component. This is because when we construct any CFD model, at least with our default settings in BF, we make an assumption. We assume that the pressure in the outlet is 0 (you can validate that by checking the boundary condition of the outlet patch in the p file, inside the 0 folder). To achieve that therefore, we allow a large extension that lets the freestream converge. A nicely converged study will always show that, a pressure that is near 0 close (and not only at) the outlet boundary.
Also you generally want to allow for a side extension, although the one you have now seems ok. You can try extending it a bit, maybe try 6 in the side extension input. I have a habit from the days of hand calculating grid expansion values to use the same number (6 in this case) for windward and side extensions but any number (usually equal or smaller than that) will do for your case. I think a best practice is somewhere around 3-4.
If your model gets too big from the choices above, you can try playing around with the blockMesh grading component. It can allow you to put a lot of refinement near your region and much less at the edges of your mesh. I typically have values like 200m even for individual cells near the end of my tunnel for example. My typical grading mesh settings is to set a blockMesh cell size of 1-2m at my region of interest and an expansion of 1.2 (max) from there to the edges (sides and outlet). For the inlet, I make sure the initial cell size is not too large. A value of 10m for the first cell would be good.
I finally get some time, I am planning to upload a number of BF definitions on Hydra, with the updated components, that show the logic of all this. Bare with me for a few days.