[matteo.monti]

Hello everyone,
I hope this is not such a dumb question but I spent several days trying to look through the wiki and making trial&error (mostly errors!) to understand how to do this, but I seem not to be able to understand it on my own.

Here is my problem (i think this is quite a basic problem) that I would like to solve with SU2.

I have a 3D mesh of an object. As far as I understand (I have a basic background in 3D graphics) a 3D mesh should be just a set of coordinates of points in a 3D space, and a set of triplets (A, B, C) of points representing triangles. These triangles usually join together to form a discretization of the surface (mesh) of the object represented.

So, I have a mesh of this object, and I would like to use SU2 to just determine, given a velocity v, what the drag force on that object would be if air was to flow against that object at that velocity, with v parallel to the positive x axis.

I read how the configuration file is formed, and I was able to use SU2 to solve the flow for the example meshes. I also think I understood how to interpret the surface_flow file to get the force (correct me if I am wrong!): it is just a set of points representing the pressure as a scalar field over all the vertices specified in the mesh. Now, since pressure has a meaning only when applied to a surface, I would loop over all the triangles in the mesh, compute the average pressure over them (just by computing pressure on (A + pressure on B + pressure on C)/3), and then get the force as a sum weighted by the area of each triangle of the force vector given by the pressure multiplied by the normal versor to the triangle. Does that sound right?

Now, I wanted to understand how I could generate the mesh. I read the wiki (input and output files section) but I couldn't understand. As far as I understood from the wiki, an .su2 file contains somehow the description of some internal structure of the 3D objects, so that if I wanted to describe an object I wouldn't be able to just describe its surface (which sounds to me like the only thing that should be relevant to a CFD simulation!) but I have to come out with some (irrelevant?) internal structure made of tetrahedrons or similar 3D elements. Why that? How can I do it? I have also noticed that one has to write the description of marker elements that actually sound like the mesh I was looking for! But that feels quite redundant: if I have to describe the whole internal structure (again, it sounds irrelevant to a CFD simulation!) the border of that object should automatically be evident! Why repeat that?

So I made this very naive test and tried to just ignore the 3D elements part and just write down one marker for a tetrahedron, on which I tried to compute the flow.

Here is the .su2 file:

NDIME= 3
NELEM= 0
NPOIN= 4
0.0 0.0 0.0 0
1.0 0.0 0.0 1
0.5 0.0 0.8660254 2
0.5 0.75 0.4330127 3
NMARK = 1
MARKER_TAG= tetrahedron
MARKER_ELEMS= 4
5 0 1 2 0
5 0 1 3 1
5 1 2 3 2
5 0 3 2 3

I setup some basic .cfg file that looks to long to paste here, and tried to run the simulation. I got a SEGMENTATION FAULT! That scared me a lot and I gave up with trying to understand this on my own.

Can someone explain to me, just from scratch, how to do what I need to do?

- I have a set of triangles (mesh) on which I would like to run a simulation
- The simulation consists in air flowing along the positive x direction at some speed v
- I need to know the drag force D that acts on that object

Thank you very much,
and sorry for the long post!

Matteo Monti