Analysis showing non-symmetric results to symmetric problem

[jones79]

Hello! I am new to Algor and encountered a problem while I was trying to calculate a rather simple steady-state heat transfer. The assembly was created using Inventor 2011 and consists of two parts (cuboids). The result shows a clearly non-symmetric solution to a symmetric problem. The mesh size does influence the non-symmetry and is much worse with a wide mesh size. I attached a picture that shows the problem (was created from a fine mesh, 10% autosize) Loads -------- Heat Flux: front side of thin cuboid (simulates sunshine) Convection: front and back side of thin cuboid I hope you can help me with that. Kind regards, Walter

[John_Holtz]

Hi Walter, welcome to Autodesk Simulation (a.k.a. Algor)

I am having trouble interpreting the image that you attached. A 3-D view may be helpful. I do not see the symmetry in the model, but I am guessing that the symmetry that you expect to see is that the bottom green and yellow contours should be straight like the top band. Correct?

What result are you showing? Temperature? Heat Flux? Heat Rate of Face?

It also depends on whether you are looking at an external face, or a face on a slice plane, and so on.

If permissible, you can create an archive of the model ("File > Archive > Create") and attach the resulting .ach file to this forum.

[jones79]

Hello!

Thanks for your quick response. I edited the picture to make the symmetry clear.

> I am guessing that the symmetry that you expect to see is that the bottom green and yellow contours should be straight like the top band. Correct?

Yes, exactly. The graph shows the temperature.

> It also depends on whether you are looking at an external face, or a face on a slice plane, and so on.

It is an external face.

 > If permissible, you can create an archive of the model ("File > Archive > Create") and attach the resulting .ach file to this forum.

I'am not at work right now, but I if neccessary, I will attach the file tomorrow.

Any ideas what could cause the problem?

Kind regards,

Walter

[xli]

Hi Walter,

If mesh is not symmetry there, that could be one of possibilities to get unsymmetric results. Turn on mesh showing you would see. Even when mesh is symmetry, you also might want to check out the connection between that front panel and behind part is symmetric too. This is just a guess.

-xli

[jones79]

Xli, thanks, your hint to check the connection led to the solution: The contact was set to "surface". After I changed this to "Bonded" the result is symmetric, yet I don't know why. Can you tell me?

I have another small question, maybe you can help me there, too, so I don't have to start another thread.

See the attached picture. This are the two parts combined in an Inventor assembly.

I want the surface with the pink dots (vertical party) to emit radiation. But I need to exclude the area where the two parts are combined. How I exclude this area? Can it be done using Algor or Inventor?

Kind regards,

Walter

[xli]

Now, I can assume you have a heat transfer model. It looked like auto-made thermal contact elements had a problem. My guess is the mesh between two parts having one or two triangles there. Currently our heat transfer analysis dose not support the wedge element that is with two triangle faces. That could be the reason. You can hide one part and see the mesh betweenfor triangle to varify.

Without thermal contact elements that allow you to set almost "no emission" in the touching portion of two parts, it is a hard problem, even just to image. Well, you probably can set contact pair as free so they not connected as "bonded", and also set that portion in different surface# and then you can define the radation not including that surface# pair. This is just the idea, good luck. (And probably in near future Algor heat transfer will support wedge contact elements, I will let our thermal guys knowing).

I am not sure if Inventor thermal analysis can do your problem or not. Sorry I am not too familar to that package yet.

regards,

-xli

[jones79]

Ok, I will try to figure that out.

Thanks for your help.

I am a bit confused about the naming of the Algor software. Currently I am using Autodesk Algor Professional 2011. Is the 2012' version renamed to Autodesk Simulation Multiphysics?

I am looking forward to see the Algor interface as consistent and user-friendly as Inventor's is, and I am quite sure that you guys are working on it 🙂 Keep up the good work!

Kind regards,

Walter

[xli]

yeah, I am not very undertanding whole naming thing too. Probably "Algor" inside there does not help too much in sale. Without it making the sense to let software in autodesk more consistent and for an easy UI. It is our goal. Thanks.

-xli

[John_Holtz]

Walter,

Knowing that there is thermal contact, my guess is that the non-symmetric results are explained by this paragraph from the help ("Autodesk Algor Simulation > Meshing Overview > Mesh Overview > Creating Contact Pairs > Types of Contact" under "Surface Contact" paragraph)

Tip: A thermal resistance of 0 --- indicating a perfect contact between the parts --- is not a good choice. A resistance of 0 implies a conductance of infinity, which leads to inaccuracies during the numerical solution. Use a small, nonzero value to simulate such conditions when they exist.

Please check what thermal resistance you entered in your analysis. (Bet you didn't know there was a thermal resistance input )

For the second problem of excluding the contact area from the load, you can either split the face in Inventor, or use surface knitting when importing the model into Algor. This page in the Help describes your type of problem: "Autodesk Algor Simulation > Opening Models > Introduction > Open CAD Files > Surface Knitting".

[jones79]

Hello John,

thanks for your detailed answers. So much to learn 🙂

But I don't know what to mean saying "split the face in Inventor". If I fit two parts together to form an assembly, where can I split the face?

Kind regards,

Walter