[MercuryCFD]

Hello guys,

I'm facing a natural convection problem that I can't seem to find a proper similar solution/paper to figure out where my rationale is wrong.

For a experiment that is making use of a heated (155°C (428.15K)) block on top of an isolator that sits on a simplified geometry that just needs to model the approximated geometry as shown in this picture:



I just want to examine the flow field in the near proximity of the sphere (it is a reflective lens, that is very sensitive to changes in the refractive index of air:

I tried to understand the flow field in 2D at first, even though its not exactly a 3D problem, but I am fine with that for the start.



On this picture you can see the side view of my geometry and a fairly small enclosure. Now my question are basically:

Informations:

- I tried ideal gas and bousinessq material model and both seem to give me aproximately the same converged velocity for that one specific enclosure size.
- I am using k-epsilon turbulence model (so far I don't really care about my boundary layer contributing to the results if that makes sense).
- I tried using p-inlet at the bottom/side and p-outlet at the top with delta_p=0 to my operating pressure. The left edge is symmetry.
- When I added delta_p at inlet of 100pa, and delta_p at outlet of 100pa the velocities went up too much imo.

1. Why are my boundary conditions not inducing a flow field that make sense? If I define a bigger enclosure (no matter in x- or y-direction) the velocity gets bigger. What is even worse that no matter how far away from my geometry the velocity is almost the same as it is close to my geometry.
2. How does the solver is increasing the velocitys so much if I just increase my delta_p at in- and outlet a tiny bit?
3. If you have questions that might help you understand my problem and further helping me in understanding, feel free to ask and I will hand it in later.

[MercuryCFD]

Originally Posted by MercuryCFD

Hello guys,

Moderator can close this. I got it done and have evidence for correct results.

Sincere regards