[gmpatzer]

Hello all,

I am attempting to run variations of the same STL sphere in the same environment to compare lift, drag, etc.

I am running into a lot of trouble converting my STL files into SU2 files -- I have even tried using third party software (gmsh) to do the conversions with no luck. In my attempts so far, the converted SU2 file has either failed to converge, or is missing markers.

Ultimately I would like to write my own code to convert between the files. But right now, I would really appreciate it if anyone could point me in the right direction.

Frankly, I'm not very experienced with this kind of software, but after months of research, reading, and attempts I realize I have made very little progress. I know this is a broad request, but I have checked this forum and have read the documentation many times. If anyone has any pointers (a guide, post, etc) or could even convert the STL into an SU2 file, put it in a wind tunnel, and share it as an example file, it would be greatly appreciated.

Here is a link to the STL file: https://drive.google.com/file/d/1TyB...ew?usp=sharing

[pcg]

Hello,

Short answer:
For a 3D geometry it is not possible to convert STL to SU2 because STL is a format where surfaces are described as a collection of triangles (almost like a surface mesh) and SU2 needs a description of a volume (using tetrahedrons, hexahedrons, prisms, pyramids)

Long answer:
It is not easy to write an elementary (not needing prior knowledge) description of what you need to do, so I highly recommend that you look for some presentation/book/video that covers the basics of CFD at the user level, namely what are the major steps involved. I say this because it seems that you are trying to go from geometry to simulation without creating a mesh of the domain where you want to solve the governing equations.
Although there are physics models and solution methods that do not require a mesh, the Finite Volume Method in SU2 is not one of them.
A mesh is a discretization of a volume (the domain) into simpler geometric shapes. This division into small elementary shapes is what allows the partial differential equations of your physics model to be simplified into algebraic equations that can be solved.
So for you specific case, where you want to simulate flow around a sphere, the key word is "around", your domain is not the sphere, it is the volume of fluid around it, you need to subtract the sphere from a larger shape (cube, sphere) whose boundaries are far enough from your inner geometry to not have an influence in the solution (say 50 times the radius). In other word you need the negative of your geometry.
You then take this geometry and use a meshing software (such as gmsh) to divide it into small volumes, and that is what you feed to SU2 or other CFD solver.

Cheers,
Pedro