OpenFoam Box Domain Wind Simulation: Null vectors or no vectors in some grid points?

[AVBBS]

Hello,

I am working with Eddy3D for outdoor wind simulations and would like to clarify a few issues regarding wind-vector output and terrain setup.

I am using the following software versions:

• Eddy3D: v1.14.0.827
• Rhino: 8 SR34
• Operating System: Windows 10 — 10.0.26200 x64

I repeatedly encounter the following issues:

• Wind vectors are missing at some grid points.
  I can have grid points distributed across the terrain, but not every point receives/displays a wind vector. This happens on both flat and sloped terrain, while neighbouring points can have valid vectors. The simulation otherwise appears to run correctly and produces vectors in many other locations. What determines whether a grid point receives a velocity vector? Could this be related to the CFD cell size, probe location, interpolation, terrain mesh, buildings, or something else?

• What is the recommended relationship between grid spacing and CFD mesh resolution?
  Could missing vectors occur when grid points are smaller/more numerous than the CFD cells? Is there a recommended Domain Cell Size / refinement / probe spacing relationship to obtain vectors consistently across the terrain?

• Sloped terrain sometimes produces this warning:
  The terrain does not cover the whole domain floor (635 of 644 floor samples lie beyond it). Past its edge the floor is a flat plate at the terrain's lowest level, and the flow can pass under the terrain sheet there. It reaches about 110 m from its centre; the domain needs 970 m — set Land Cover Roughness's Terrain Radius to at least 970, or extend the terrain.

  Why does Eddy3D require such a large terrain radius, and what is the correct way to solve this permanently? Should I extend the terrain, increase Terrain Radius, or adjust the Box Domain?

• Could the geometry type be responsible?
  Does Eddy3D require the terrain and buildings to be Meshes, Breps, or specific geometry types? Can the resolution/topology of the terrain mesh affect whether wind vectors are returned?

• Could simulation settings cause missing vectors?
  Are iterations, number of CPUs, Domain Cell Size, Box Domain dimensions or refinement capable of causing missing vectors, or should missing vectors normally be considered a probing/mesh issue rather than a solver issue?

I have attached a Grasshopper file with the relevant components, together with screenshots of the results, so the setup can hopefully be reproduced.

Thank you for your help.

Best regards,

AV

WindSimulation_Eddy3D.zip

[kastnerp]

It looks like you are probing below the terrain, which yields NaN (or a very large number) in OpenFOAM. You'd have to offset your probing points from the terrain surface. Please let me know if this works for you.

Patrick

[AVBBS]

Hi Patrick

This probing was done with the probe points positioned 2 m above the terrain. I also ran the simulation and probing several more times at different elevations. As you can see, more arrows become visible at higher elevations. However, at 5 m and 7 m, the simulation seems to barely recognize the buildings as boundaries anymore, which is not what I want. I would like to analyse how the wind behaves behind the buildings at 2 m above the terrain. I would expect to see more detailed wind patterns behind the buildings, for example recirculation or flow recovery. There is not a single probe point located below or in the terrain at the 2 m offset. What could therefore be causing this issue, and what would be a possible solution?

Thanks a lot for your help.

AV

Gridpoints above the surface:

2m (also weird that some buildings are not seen as boundaries):

5m:

7m: