Issues calculating Cd of a simple cube (beginner)

[zifcani]

Hi everyone, I'm new to CFD in general and I'm trying to calculate the coefficient of drag of a cube but I cannot get the right value (1.05 at standard conditions according to Wikipedia). I found this Autodesk tutorial about this but no matter what I did I was getting wrong values. I also used the drag formula from that article but its just a rearrangement of the regular drag formula.

Since my mesh was extremely simple, the program autosized it to 10k elements. I tried increasing element count and mesh refinement up to like 200k which did change values but it still didn't give me the correct ones. I tried plugging in the Z axis (upwards) force from the wall calculator on top, bottom, top+bottom and all sides of the cube with the same cross sectional area (10x10cm cube so 0.01m2) and I also tried plugging in the values on the whole "tunnel" but none of that worked. I tried changing the size of the "tunnel" and other things I forgot at this point but all of that to no avail.

This is one of those runs right here with all of the forces on all sides of the cube. Note that the arrows should be the other way around, Z force is greater on the bottom face.

I have attached my projects so it would be great if someone could take a look and tell me what I'm wrong.

Thank you.

[marwan_azzam]

Hello Nuff Lee,

What kind of Cd values are you getting?

The rar file you sent does not open here.  Please make sure it is not corrupt and share again.

Marwan

[zifcani]

Hi, I'm sorry for that, I will reupload the file. 

For example, on the screenshot I included, for the Z force on the bottom face (Boundary ID 6), Cd=~1.45, for the force on the top face (Boundary ID 1), Cd=~0.42 and for the total Z force, Cd=~1.89.

The following is the code I'm using to calculate Cd:

rho = 1.20473
v = 10
A = 0.01

def calculate_Cd(Fd, rho, v, A):
  return (2 * Fd) / (rho * v**2 * A)

[marwan_azzam · Accepted solution]

Thank you for sending the model.

I applied symmetry boundary conditions around the air domain.  This resulted in a reaction force of 0.654N.

Based on that, Cd comes out to 1.09 which is pretty close to the 1.05 number you are looking for.

I'm attaching the cfz file.

Marwan

[zifcani]

Thank you for the quick response.
I will download your project and play around with it but do you mind explaining why adding a symmetry boundary condition would make such a big difference, especially since my mesh is already perfectly symmetric? And what would happen if my mesh wasn't symmetric? Would I still be able to get correct results?

[zifcani]

I finally managed to open your project and I realized you were not referring to "mirror symmetry" but to a "Slip/Symmetry" boundary condition which makes a lot more sense. Do you mind explaining what this does exactly? Does it just make the walls of the tunnel frictionless? I also saw that some other things may be different so do you mind quickly explaining why you did those changes and what exactly do they mean?

Thank you.

---

[marwan_azzam]

Hello,

You are right, Slip/Symmetry boundary conditions mean that the walls are frictionless and flow parallel to the walls is allowed.  This does not affect the mesh other than not creating a wall layer.  It makes the model behave as if there is a continuum to the air domain (it is larger than it actually is).

That is the only change I recall making.  I'm not sure which other changes you refer to.

Marwan

[zifcani]

Ok, I see, thank you. 

Do you just mind explaining how mesh sizing and element counts and stuff works (or link some resources)? Because mesh resolution certainly makes a big difference but I'm not sure how to adjust it properly and know how many elements I'd want for a good enough sim.

Also, I'm only getting the correct Z force of ~0.65 N on the bottom face (the one facing the flow), but there is also Z force (something like 0.2 N) on the top face and most tutorials tell me to use the sum of Z forces which would throw off Cd by a lot? Is that wrong and should I just be taking the force on the face(s) facing the flow into account?

[zifcani]

Just wanted to clarify the question from my last post since I cannot edit it - I was trying to ask how to know if I have enough elements/mesh resolution to be confident that my results are actually correct because I usually won't have a reference to compare against.