Learning Forum

[can.sumeyye]

Hello. I'm performing LES analysis using the vortex method.

 

• My vortex count is 1000 (I know 200 is normally recommended, but Ansys also states that the number of vortices should be 1/4 of the inlet mesh count, and mine is around 1100).

• I obtained the inlet velocity profile and inlet turbulence profiles (k and epsilon) from experimental data. More precisely, I calculated epsilon using lt (integral length scale) and k. So maybe the problem stems from epsilon, I don't know, but this epsilon didn't cause any problems in RANS analyses. Furthermore, the average of the profile I provided is not far from the epsilon average given in the experiment.

• Also, I didn't provide anything in the Reynolds-Stress Specification Method, so it probably assumed isotropic, but in reality, the setup/experiment is not isotropic. There is higher turbulence in the u' direction, i.e., the jet flow direction. Normally, the turbulence values(Figure 4.1) ​​should be lower in v' (the horizontal direction relative to the jet).

 

 

The problem stems from this: if you look at the graphs I've provided below, profile 1 is the velocity profile taken from the center in the direction of flow. It belongs to the dashed line experiment, and the blue line is my LES result. There's a sudden drop and increase.

 

 

ALSO, if you look at profile 2, the turbulent kinetic energy starts almost at the same level as the experiment and then decreases. 

 

ALSO, if you look at profiles 3 (left) and 4 (right) below, both start very high and then drop suddenly. But the value of 4 (i.e., v') shouldn't actually be so aligned with 3 (i.e., u'). Furthermore, I don't understand the reason for this sudden drop in the downstream.

 

 

1. Is the problem caused by the mesh? Chatgpt claims that mesh is not enough. However, I had already checked the mesh by asking Ansys. https://innovationspace.ansys.com/forum/forums/topic/les-resolution-quality-for-combustion/
2. Should I have used one of the other methods instead of the Vortex method, like the synthetic turbulence generator?
3. Was it the isotropic assumption that caused the problem? Should I have given something like (u')^2 for the Reynolds stress part?

Honestly, besides the three things I mentioned above, I can't think of any other reason. I hope you can help. Thanks in advance.