## Fluids

#### Creating baffles within cells of Macro Heat Exchanger Model to simulate presence of fins

• gimson
Subscriber

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.

• RK
Ansys Employee
Hello,
It is possible to model baffles in Fluent meshing. However, you will have to be careful while bringing in the scdoc file into FM. Please refer to this post on suggestions to bring baffles into FM : https://forum.ansys.com/discussion/26065/ansys-insight-spaceclaim-settings-for-creating-3d-surfaces-baffles-for-cht-analysis/p1?new=1
Please try the Fluent meshing approach before making simplifications to the model.
• gimson
Subscriber
Hello, thank you for your prompt response.

I attempted to replicate the settings shown in the link you supplied, however there were a few problems I have encountered. My steps for geometry creation and meshing are as follows:
1. Create air domain and hxc core
2. Create rectangular surfaces and use Linear Pattern tool to create 199 surfaces which will act like fins, and thus dividing the core into 200 division per length.
4. Group air and hxc core as separate components and labelling them as Merge in the properties. Group all fin surfaces as a single component and label it as None in properties.
5. Next, I used the Share Topology tool in SpaceClaim and it identified the edges and baffles correctly and I applied sharing.
6. After that, I named all the appropriate surfaces with Named Selection, for the inlet, outlet, fin, hxc-walls, hxc-inlet, hxc-outlet.
7. Transferred the Geometry to Fluent Meshing via Workbench by linking the modules.
8. In Fluent Meshing, I followed the Watertight Geometry workflow. Didn't add local sizing. Created surface mesh. Described geometry with fin and hxc-wall as walls, hxc-top and hxc-bot as internal.
9. I'm not sure how to proceed after this step. I tried setting the hxc-core region as dead such that Volume mesh does not occur within this area. Instead I used Fluent Meshing's Mesh > Create > Heat Exchanger, to create the mesh for heat exchanger within this area and then Create Volume Mesh after this.

After this step, I am not able to get the baffle surfaces in between the Heat Exchanger cells, and I lost all the surface meshes for the baffles, I think this is because I set the hxc-core to dead. I think I should be using Fluent Meshing's Mesh > Create > Heat Exchanger to create the cells for the Macro Heat Exchanger Model since this option is provided. However, I am not familiar with manipulating the mesh as I ended up having 2 separate cell zones as shown in the Outline View here.

Is it possible to extract the surfaces within the Heat Exchanger cells to create the baffles instead?
My goal was to create the vertical baffle surfaces, an example of 1 baffle surface is shown in the attached picture, between the structured mesh that was created with the Heat Exchanger option. And this structured mesh is embedded within an unstructured mesh for the air domain.

Can you advise me if I have gotten any steps incorrect and how I could proceed with the settings in Fluent Meshing to obtain the baffles? Thank you.

• RK
Ansys Employee
I think the issue was in step 5. You do not need to add share topology, instead in the properties tab, under shared topology, select share (just like how you selected the merge option). This should do the trick.