[shivasluzz]

Hi all,

I am trying to simulate a 3D gear pump, which has 25 microns axial clearance and radial clearance. I use the Overset method for this problem. I used prism layers with 5 microns as the first node distance. I consistently get the divergence error with velocity shooting up to an infinite value.

I tried the following:
1. Running steady state simulation with zero rotational speed.
2. Running with Invicid flow option
3. Checked and ensured mesh quality (no skewness Or orthogonality warning in fluent)
4. Tried Hex mesh and Poly mesh
5. Tried mass flow as inlet and also as outlet
6. Checked for any domain connectivity by checking 'auto mesh interfaces'
7. Reducing the boundary conditions with very minimum values
8. Reducing the timestep Or time scale factor values
9. Reducing the under relaxation factors to 0.3
10. Tried Couple, PISO and SIMPLE pressure velocity coupling
11. Checked and ensured the fluid properties.
12. Tried standard, hybrid and patch initialization
13. Ensured a minimum of 4 elements in the overlap region
14. Ensured the minimum no. of orphan cells (200 orphan cells out of 1.7 crore mesh elements)
15. Ensured no orphan cells at critical region
16. Ensured there are no gaps or cut or overlaps after initialization
17. Tried idea gas in place of actual fluid
18. Tried gap modeling for the clearance with the artificial viscosity by defining the threshold value of gap value


All these I tried both for the steady state and transient cases. However always the simulation fails within a few iterations. I observed the mass flow out and average velocity shooting to infinite values and this starts from the iteration when there was reversal flow warning at the outlet.

I believe the problem is due to the very fine mesh and also I guess the fluid is unable to flow over the small gap, but I am unable to find the reason even after several attempts. My Reynolds no. in the gap is coming around 160000.

Any help would be highly appreciated!!!

Thank you!

[shivasluzz]

Also when I use the turbulence model, I observed that k and omega equations diverge from the very first iteration and the turbulence viscosity ratio is reaching its limits! Then it is failing with the floating point exception error!!

[shivasluzz]

And my one more doubt is, can the transition ratio cause this kind of problem? I used uniform layers of 0.005 mm and the body element max size is 0.1 mm!! I understand the transition ratio should be closer to 1 but I couldn't maintain it because of the geometry!!