Learning Forum

[Gregory Harrington]

I am having some issues creating some surfaces in Fluent that are intended to be internal surfaces that permit flow through them so I can monitor temperature, pressure, mass flow etc.

I created these surfaces in Discovery and they are a circular pattern that has 6 total surfaces. The circular pattern therefore has 6 component parts inside it and each component part holds a surface. The 6 surfaces are called vane-1-in, vane-1-out, vane-2-in, vane-2-in, vane-2-out, vane-3-in, and vane-3-out. I used performed shared topology in Discovery and completed the geometry pre-check successfully with no errors. I then created a named selection for each surface so there are 6 named selections. 

I imported the geometry into Ansys fluent watertight mesh workflow and checked all my labels while creating local sizings. As you can see in the picture below, vane-1-in, vane-1-out, and vane-2-in all have an associated label and rolling over them in the workflow highlights the correct surface. Vane-2-out was skipped and the 2 surfaces for vane-3 are bundled into vane-3-in:vane-3-out. Also, my entire model is highlighted when I roll my mouse over this surface.

When I look at the zone list, there is a zone for origin-vane-1-in, origin-vane-1-out, origin-vane-2-in, and the other 3 surfaces are bundled into origin-vane-2-out:vane-3-in:vane-3-out.

Another weird thing is if I don't include the vane 3 surfaces (just vane 1 and vane 2) everything works as intended meaning I have a separate label for the 4 surfaces.

I really need help with this because my model is actually going to have 20 inlets and outlets (40 surfaces total).

[Gregory Harrington]

I want to add, while these named selections have -in and -out in their names, they are not boundary conditions of the computational domain. Flow comes from one space, splits 20 times and goes through each vane inlet. Then flow goes through each vane and comes back together. I have 40 named selections (20 for vane inlets and 20 for vane outlets) so that I can measure parameters on each one. For example I want to be able to determine how much flow goes to each vane and the temperature of the flow going to each vane.

I am coming from a star-ccm background where I can have one region with internal interfaces where I measure parameters.

Do I need to use multiple regions here?

[Federico]

Hello Gregory, 

this can be done if your internal BCs (the vanes) separate different regions. You cannot have internal face zones with the same cell zones on both sides in Fluent.

[Gregory Harrington]

So I am going to have to have 20 different regions? Aghhh that's a pain.

I am coming from a star-ccm background where you can have 1 fluid region and separate it by surfaces to measure quantities of interest on those surfaces like pressure/temperature etc.

Thank you.

 

[Federico]

If you need those surfaces only for monitoring purposes, then you can take advantage of Imprint Surcaces 38.1. Using Surfaces (ansys.com) (see section 38.1.3 Imprint Surfaces).

You can have a single fluid cell zone. Create a new file that has only the vane surfaces and save it in .stl format. You will then be able to read this file in Fluent Solution and imprint them onto your fluid cell zone. You can then use those surfaces for monitoring or creating plots.

Let me know if this addresses your question or not.

[Gregory Harrington]

Federico that sounds like the best option for me. I figured I would have to do all that set up prior to meshing.

If I create the imprinted surface after there is a solution then isn't the mesh not conformal on the imprint? or does the solver do some type of interopolation. I am thinking the imprinted surface will "cut through" some cells instead of the cells on either side of it be conformal.

[Federico]

Yes, you bring up a good point. I believe Fluent will cut through and interpolate data from the nearest cells on the imprinted surfaces but it would be worth testing!

[Gregory Harrington]

That is why I wanted to get these surfaces in before the volume mesh is created so that wont happen, but I also dont want to have a ton of regions. 20 is a lot! If the design changes I can have maybe 30-50 and then that makes the model exponentially more difficult to deal with.

I just want to have 1 region and monitor quantities at intermediate points within the region.

I guess the best way for me to explain it is to have a pipe and I want to monitor quantities at various stations down the length of the pipe. That would be the easiest scenario where I could just make planes and monitor on the planes. But in my case my interrogation surfaces aren't nice easy shapes to work with.

[Federico]

I understand that. but I think your best option remains surface Imprints so that you can provide the desired shape and location. It is also a scalable option if you grow your model.

If your mesh is fine enough, I wouldn't worry too much about interpolation error.

[Gregory Harrington]

I think I am going to try that. But I guess I still don't understand why the named selections are doing that. I thought if you make a named selection then a label will automatically be created for each one.