Help needed in flow over a bullet-shaped object

[ankita-nag17]

Hello, I wanted to simulate flow over bullet shaped object, nose is elliptical (2:1) and total length of body is 2D, D is diameter. I was using the STL file to make my epsilon.bin and other files. I have already done it for different objects, and it worked. But in this case, with the bullet, I am not getting proper results. First of all, only the grid of 151 for the Y and Z directions is working; any other grid size, like 161,181,301, for Y and Z, is getting diverged in the initial 200 steps. Even the case that completely ran has issues, as I am observing a sudden velocity increase after the object. I am using iibm=2. I have tried changing dt, but that also did not work. I am attaching my epsilon file slice. Can anyone suggest any help?

[slaizet]

Maybe try iibm=1 (no flow reconstruction inside the bullet) to see if it is the problem.

[ankita-nag17]

Thank you for the suggestion, but I tried using iibm=1, but it is giving zero value of velocity and force. It is not solving the problem. Do you know what the possible issue can be?

[slaizet]

No clear what the issue can be. I would try to change the spatial resolution to see if it is an issue with this. The issue might be with the sharp corners.

[ankita-nag17]

I feel like the issue is coming from the stl approach along with the Lagrange polynomial approach.Is it possible to combine the facility of stl importer with the cubic spline? As lagrange polynomial can overfit and give rise to grid-level oscillation.

[slaizet]

there is no compatibility issue with the two approaches so not sure what the problem can be here. It could be maybe that the stl file was not properly generated with small holes in the geometry (this is not unusual for stl files).