"Array overflow in Mesher" Programm gives this error. Can't fix it.

[Kiakonov]

I have a double glazed window with a frame. I tried everything, but when calculating it gives an error. Help me please. I don't understand how to solve it.

Attached the file to the request

[Robin Mitchell]

You need to simplify your file. 

I will try some simplifications and see if I can get it to run. 

Others can do the same if they are inclined (!)

Robin

On Wed, Mar 6, 2024 at 11:46 AM Kiakonov <kiak...@gmail.com> wrote:

I have a double glazed window with a frame. I tried everything, but when calculating it gives an error. Help me please. I don't understand how to solve it.

Attached the file to the request

--
You can view the THERM forum online at:
https://groups.google.com/g/LBNL-THERM
---
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/f5a41316-8a29-4d1d-bd16-62ba779a2232n%40googlegroups.com.

--

Robin Mitchell

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

[Jonathan Spencer]

The solution here is to to turn down the mesh and chase the red circle until it works. I turned it down to 8 and then followed the circle, making 2 or 3 small changes until it calculated. Basically, cut off chunks of polygons and fill them back in. Its all completely arbitrary, keep going until you get something that works.

I'm curious about the airspaces being broken up into sections like you have them. Is there a reason behind that. It's not causing the problem here. Just wondering.

[Jonathan Spencer]

The solution here is to turn down the mesh parameter and then chase the red circle until it works. I turned it down to 8 and then followed the red circle, making 2 or 3 small changes to the model until it calculated. Basically, cut off chunks of polygons and fill them back in. Its all completely arbitrary, you just have to keep working at it until it runs for you.

On Wednesday, March 6, 2024 at 2:49:37 PM UTC-5 rdmit...@lbl.gov wrote:

[Robin Mitchell]

Jonathan 

That is a good approach. 

In terms of the frame cavities, they should not be broken up like that. They should be modeled as one cavity, breaking them up only as shown in the THERM manual

https://cdn.ymaws.com/nfrccommunity.org/resource/resmgr/simmanual/nfrcsim7-mar2023.pdf

See description starting on page 6-8.

Robin

To view this discussion on the web visit https://groups.google.com/d/msgid/LBNL-THERM/36002db9-f381-4b6b-8a08-11dc05281d7dn%40googlegroups.com.

[Marek Bożydaj]

It looks like with bigger cross cections and fine detailes in them the error: Array overflow in Mesher is inevitable. What is the problem to increase the capability of the software to be able to work with larger mesh data bases ?

[Robin Mitchell]

THERM 8.0.33, which is on our public website (under "Research versions of THERM 8"), has a new mesher that should be able to handle the larger and more complicated THERM models. 

https://windows.lbl.gov/therm-software-downloads

You can set the program to use either the "old" mesher or the new mesher, by clicking on the File menu, Properties option, and then under the Calculation Options tab, select the "Simmetrix Mesher" option. 

This new mesher hasn't been widely tested, as far as I know, so if you want to try it out on some of your complex models, that would be great. 

And let us know if you find any issues using this new mesher. 

Robin 

To view this discussion on the web visit https://groups.google.com/d/msgid/LBNL-THERM/b6ba92a6-1f55-4907-b039-d67163e53abcn%40googlegroups.com.

[Marek Bożydaj]

 Thank you rdmit...@lgl.gov.

The new version of Therm did the calculations without any problem. You can simplify the cross sections of the profiles changing slightly the corners, arks or deleting small voids. But with spacer it will be senseless. Creating models of spacers we are sometimes modeling layers or fields of the materials with thicknesses less than 0,01 mm. So in this case simplification is not possible. Below is the cross section of the vertical meeting rail with TGI-Spacer M and the spacer.

 

Changing the Mesh Parameter in Therm 7.8 created vicious circle. Decreasing the parameter you get bad points and errors increasing the parameter you are getting array mesh overflow. Simplification of the spacer cross section is rather sensless. I did manage to calculate the cross section with Therm 7.8 but I have to replace the detailed cross section of the spacer with the two-box model method (Rosenheim ift-Guideline WA-08engl/3).

 

But I like alredy the Therm 8 version. I hope it soon will be official for use.