[gimson]

I'm currently simulating a 3D natural convection problem in a closed room with fin-coil heat exchangers providing the cooling for the air. The fin spacing of the heat exchanger is around 10 fins/inch thus it is not possible to mesh it as a solid.

I used Fluent Meshing to mesh both the room and to create the heat exchanger cells. This is because the documentation stated that the heat exchanger cells have to be structured with Hex/Wedge cells and Fluent Meshing provides the option to create cells that are suitable for the Heat Exchanger Model.
Next, I used mesh interface to match the inlets and outlets of the heat exchanger to the air in the room.
I applied the Ungrouped Macro Heat Exchanger Model in Fluent and gave a row/pass value to match the number of fins per pass.
The solution seems to converge after checking for the energy balance of the room, as all the heat load in the room has been taken up by the heat exchangers. However, when I viewed the velocity vector plot of the cross section within the heat exchanger, I see that some parts of the flow have vectors pointing in the direction of the fins. Realistically, this shouldn't occur since the fins are meant to be walls and should at least channel the flow in 1 direction. This probably affected the accuracy of my results for the final room temperature and the cooling capacity of the heat exchangers.

I read through the documentation for the heat exchanger model in Fluent and they did mention that the primary fluid streamwise direction must be aligned with the core. So in order to achieve such flow, may I know if it is possible to mesh baffles within the heat exchanger cells using Fluent Meshing such that the flow through the heat exchanger cells are directed through in the same direction?

Currently, my other attempt at this problem to validate the Macro Heat Exchanger Model is to mesh the heat exchanger core with baffles in air to simulate fin walls at a fixed temperature 1K higher than the water temperature running through the heat exchanger. The results of this simulation shows a high difference in the cooling capacity and a poorer cooling performance as compared to using the Macro Heat Exchanger Model, however the flow pattern looks a lot more realistic for a heat exchanger.
May I know if this simplification is acceptable?

Please let me know if you have any suggestions for simplifying the model, or if I should provide any additional information. Thanks in advance.