Quad Mesh issue

[Ankur Gupta]

I am just about to give up on this software. I have run out of any idea why my file does not simulate. The polygons are perfect. I would be enlightened and motivated again if an expert can look into this.

[Robin Mitchell]

I will take a look at it tomorrow

Robin

On Sun, Apr 26, 2020, 10:47 AM Ankur Gupta <aagup...@gmail.com> wrote:

I am just about to give up on this software. I have run out of any idea why my file does not simulate. The polygons are perfect. I would be enlightened and motivated again if an expert can look into this.

--
You can view the THERM forum online at: http://windows.lbl.gov/software/THERM/forum
---
You received this message because you are subscribed to the Google Groups "THERM" group.
To unsubscribe from this group and stop receiving emails from it, send an email to LBNL-THERM+...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/LBNL-THERM/e45595ef-770f-4500-88c8-dc0d6193a36d%40googlegroups.com.

[Robin Mitchell]

What version of THERM were you running for this file? 

Robin 

--

Robin Mitchell

Building Technology and Urban Systems Division
Energy Technologies Area
Lawrence Berkeley National Laboratory

[Robin Mitchell]

It looks like all your polygons are defined as Frame Cavities, even the polygons that I assume should be some other material like aluminum or steel?

Also, this is a very large file of a symmetrical design. I am not sure the THERM mesher will be able to handle it. I would suggest just modeling one of the repeating modules. Possible cutting it something like this (others on the thread may have different ideas of where to cut the repeating pattern, which would be interesting to hear)

Also, looking at the polygons, there is too much detail in some of the curves. You will probably need to simplify them quite a bit to get the model to run. 

On Sun, Apr 26, 2020 at 4:15 PM Robin Mitchell <rdmit...@lbl.gov> wrote:

[Robin Mitchell]

I simplified the model to only have one repeating symmetrical section, and the file runs without issue even without simplifying the curves, etc. However, I have no idea what the materials really are, and that could potentially change the mesh, depending. But I did assume that all the thin elements were steel, which is what the mesher would usually have a hard time with, ie, very small, thin, highly conductive elements in a large model. But that didn't seem to be the case here. 

I have attached my model -- given that it meshed without simplifying the curves at mesh parameter = 8, you could potentially expand the model a bit, but I don't think it's needed, since the isotherms (at least with the materials I assigned which are probably completely wrong) become parallel in the smaller segment modeled.

My file is attached