[dasith0001]

Thank you for the quick reply, let me back-off a little bit here to explain this better.

Domain : wedge vertical cylinder

Components : please see the attached figure, hollow solid in the middle (S1), fluid domain all around, vertical heater, hollow cylinder as the outer boundary (S2)

Physics : Heater, heats the fluid, creates buoyancy driven flow and distribute heat all around. convection BC set to outer BC of S2. ( note that heater also radiates heat to S1 and S2

Issue : Experimentally measured temperature difference is 5-10 C between S1 and adjacent fluid, but simulation results says 25-30 C.

Note1 : I understand that here radiation plays a part, which is calibrated to an extent. This is a engineering application and only required to close up the temperature gap between the fluid the S1.

Note2: I got a good mesh, i have only 5 elements with high orthogonality out of 100,000. Residual remains under 1e-5. running first order Euler.

Note3 : Mesh sensitivity done, played with inflation layers in both solid/s and fluid, emissivity tweaked, turbulence added( does not make difference as velocity of the buoyancy fluid is only 0.1- 1 m/s), y+ is checked ( below 0.4).

Note 4: after reading this post, i've reduce the time stepping, increase 'n' correctors, moved to CrankNickolson , waiting for results on those.

I am running out of options to reduce the temperature gap between the S1 and the fluid. At this point if there is anything even artificially increase the heat transfer rate, that would be helpful. Because one could argue it will balance out the errors from the radiation heat transfer.

OF12 - OpenFOAM Version 12.

Please let me know if anything is not clear or need more info.
Thank you
D

[ghorrocks]

This is the CFX forum. For answers specific to Open FOAM you should try that forum.

One thing I can see straight away which could reduce heat transfer is that you are running a wedge of a vertical cylinder in a buoyancy driven flow. For most practical applications buoyant flows are 3D and transient, and if you reduce it to 2D (or a wedge with symmetry) you will underestimate the heat transfer as you cannot model the flapping of the buoyant plume.

Don't artificially increase the heat transfer. That is not a useful way forwards.

[dasith0001]

Thank you for the input, unfortunately 3D is not an option as the domain becomes too big to handle in a practical time frame. Thank you anyway.

[ghorrocks]

In that case then the simulation you propose is not practical in the computing resources you have, so it cannot be done.