Learning Forum

[colintho]

Hello,

I am currently modelling a geothermal earth air heat exchanger entering a room (a box). There is then a vent on the other side of the room that is attached to a solar chimney which will accelerate the air out the house. 

I have modelled so far in design modeller the house attached to geothermal pipes. I was looking to test a simulation at this stage but have come across the enclosure error. I tried using the enclosure tool but this failed to work.

As I am modelling the heat transfer to the air through the pipes and then the mixing in the room I need to create an enclosure as a fluid domain. I was hoping to then do a boolean subtraction to leave me with the domain of air to be meshed. 

 

I tried manually creating a frozen fluid body around the model and tried a boolean subtraction but got the "Operation would result in non manifold bodies" error.

Any advise on how to proceed would be appreciated!

 

[Keyur Kanade]

go to tools - analysis tools and do fault detection for bodies. if there are any errors for bodies, please correct them. 

if any of the bodies are having tangential contact, then please remove these tangential contact and the try boolean. 

[colintho]

Hi kkande,

Tested for any errors and non appeared. 

 

Can you explain a little what you mean by the tangential contact? Tangential contact to what?

 

Thanks

[peteroznewman]

When two spheres touch at a point, that is tangential contact and it is very difficult for a mesh to fill that space, even if the geometry boolean to make a solid body enclosure to represent the fluid is successful.  The corrective action is to either very slightly increase or decrease the sphere radius to either make a finite area of overlap or a small gap between the spheres. This will make a successful mesh much more likely.  The same goes for cylindrical faces.

[colintho]

Just so I can check I am understanding properly, increasing the diameter of my cylindrical pipes should hopefully allow for the successful use of a boolean subtraction? It looks like it is the pipes causing the issue as I can perform a boolean subtraction on the house itself but not on the hollow inside the pipe.

[Rob]

Possibly - it depends on what is causing the non-manifold warning. What geometry is there inside the room (red-ish box)? 

[colintho]

 

The inside of the box is hollow, I am trying to fill the inside of the pipes and box with a fluid body "air"

[Rob]

Not sure what's causing that unless it's confused which side of some of the surfaces is "inside". I'm not entirely sure I know either.  Are the pipes & red-ish box in the first image the fluid region? 

Given the shapes, you may be quicker building the model of the fluid region from scratch: not ideal but we may struggle to help as we can't open attachments (other, non-ANSYS, community members can).  

[colintho]

The pipes and red-ish box are the walls shall we say and then my idea was to do a boolean subtraction from the fluid region which is the box that has it all inside it! I can upload the file if need be but I understand you are unable to open it

[Rob]

The idea is good, I'm just not sure why it's not working. Thanks for understanding our limitations on here: it's all down to US export law!  

Try making another volume that's a bit bigger than the solid domain and then try the subtraction again. It's possible you've got some misalignment so it's not a clean cut: ie you'd leave a volume that's too thin to exist due to tolerances with the current set up. 

[peteroznewman]

I downloaded your archive and tried recreating the enclosure solid so that it was larger on all sides than the body that was the hollow cube with three cylinders on the side. I got the same error.  I also tried the DM Enclosure feature with a margin on all sides and it similarly failed.  There was something about this topology that the geometry engine could not handle. I tried suppressing all but the box and 3 pipes then using a Boolean to Unite the four bodies. It showed an error.

The corrective action was to extrude the pipes in two directions, out where you wanted them and in just a little bit to create an overlap between the pipes and the box so that the boolean would work.  Due to a tiny roundoff error in the computation, the pipes were not touching the box. That is what caused the non-manifold error.

I reconstructed the rest of the features from this point on and now the enclosure works.  Attached is the ANSYS 19.2 archive.

If this post answers your question, please check Is Solution to mark the discussion as Solved or ask a follow-up question.

[Rob]

Thanks Peter. Was the CAD an import or native from DM?

[colintho]

Thanks for your help, that seems to have solved it! Such a small thing but looks like that's what caused it!

The CAD was Native in DM and not imported!

 

Might come back to you with further questions if I get stuck but thanks for the help!