I am new to Simscale and have only really done the compressible golf ball simulation. I used the guide to try simulate a 9x19 mm Parabellum (which I built in freecad and exported as an .stl file) with a muzzle velocity of 460 m/s. I am trying to test whether or not turbulators like the dimples on a golf ball, actually reduce the drag coefficient on things like a bullet or a car (my next model) for my high school thesis. This is then the base model of the 9 mm, which I will compare to the half dimpled one.
Simscale link: SimScale (Tried to attach Freecad file, but Simscale does not accept the format).
Error log:
Maximum number of iterations exceeded when calculating temperature from a thermodynamic potential. This may be caused by low-quality mesh producing unrealistic pressure in a few cells or inappropriate boundary conditions, fluid properties or time step. Inspect fields for large values in the last time step.
At this point I have tried multiple times with different projects, yet I still get the same error. I have read the article that is recommended (and several forum-posts), and found that my meshes non-orthogonality is really high, its max being 73. But I have no idea how to decrease it (I found a really long and complicated forum post, but I am too much of a beginner to even understand what to do, Reducing Non-Orthogonality through CAD & Mesh settings with help from the Mesh Quality Feature), or if it really is the problem? I would be very thankful if someone could help tell me what is wrong with this project.
I am however not at all experienced with compressible simulations but i can make some other universal recommendations for your meshing.
I would look into and change your boundary layer settings. You need to find the correct Y+ value so that you get accurate results. The current settings you have are more of a standard and you will be measuring in the log-law region. You need to be measuring the viscous sub-layer. this means a Y+ of less then 1.
I would also recommend increasing the boundary box layer in front of and to the side of the bullet so you dont have and problems with being too close to a wall
You could also put a finer mesh surface on the bullet itself. What you have will work but if you want to compare to a dimpled bullet. It would be better to have a similar mesh size for both runs and the dimples will definitely require a finer mesh to accurately describe the geometry. I would use a local surface refinement to do this.
Compressible flow simulations (Mach > 0.5) are not trivial. I am not fully experienced in those, in addition to @dschroeder suggestions, I would add the following:
The domain is very small compared to the bullet. If you are restricting values in those boundaries, it may generate conflict (strong oscillations → divergence) with the solution inside the domain.
Start with a lower velocity and see how the case works.
Check out this post Post. Again, such solutions easily diverge.
