Note: This discussion is about an older version of the COMSOL Multiphysics® software. The information provided may be out of date.

Discussion Closed This discussion was created more than 6 months ago and has been closed. To start a new discussion with a link back to this one, click here.

Far field plotting needs a very long time

Please login with a confirmed email address before reporting spam

Hello,

I'm trying to simulate a antenna over a ground plane with far field condition around, like it is in the examples.

The attached model works fine. But when I try to show the emw.normEfar on a surface, my computer utilize all cores with 100% (HP xw6600, 2xXeonE5450, 16GB). This is fine for some time. But after half an hour I canceled the process, because there is no cancel button at the GUI.

So I have two question:
First: Is there a mistake in my model or preferences?

And second: I lost some models on this way, because the emw.normEfar plot was the last active view when I saved. And COMSOL tries to recalculate, when I open it. Is there a way to prevent the recalculating, when COMSOL gets opened?

Thanks in advance

--
stefan


6 Replies Last Post May 7, 2012, 11:49 a.m. EDT

Please login with a confirmed email address before reporting spam

Posted: 1 decade ago May 4, 2012, 3:14 p.m. EDT
There is an option somewhere in the preferences to disable recalculation of the plots. It you choose it, plots will update only when you press "plot" button.

Also you can try increasing Java heap size (www.comsol.com/support/knowledgebase/830/) to speed up post-processing
There is an option somewhere in the preferences to disable recalculation of the plots. It you choose it, plots will update only when you press "plot" button. Also you can try increasing Java heap size (http://www.comsol.com/support/knowledgebase/830/) to speed up post-processing

Ivar KJELBERG COMSOL Multiphysics(r) fan, retired, former "Senior Expert" at CSEM SA (CH)

Please login with a confirmed email address before reporting spam

Posted: 1 decade ago May 6, 2012, 1:55 a.m. EDT
i

Indeed the plot on demand helps for heavy models, but I can also confirm tha if you have a heavy model, then some of the calculations can take for ever, it happens quite often for my multiple parametric multiple integration Derived varales I try to generate.

And indeed there is no "stop" for these calculations, I have reported this once to support, an I hope they will add it in a future version, but if we are several to ask, sometimes it comes quicker ;)
On the other hand I have heared next release is for the end Mai or June it might already be there ;)

--
Good luck
Ivar
i Indeed the plot on demand helps for heavy models, but I can also confirm tha if you have a heavy model, then some of the calculations can take for ever, it happens quite often for my multiple parametric multiple integration Derived varales I try to generate. And indeed there is no "stop" for these calculations, I have reported this once to support, an I hope they will add it in a future version, but if we are several to ask, sometimes it comes quicker ;) On the other hand I have heared next release is for the end Mai or June it might already be there ;) -- Good luck Ivar

Please login with a confirmed email address before reporting spam

Posted: 1 decade ago May 7, 2012, 3:49 a.m. EDT
Hi Stefan,

I have the same experience about far field calculation. However I found that calculation times get reasonable when you avoid 3D plots. If you just calculate the usual Azimuth and Elevation plots in polar coordinates you have normal postprocessing time demand.
The 3D plots are looking nice but what is mostly needed in antenna design are the polar plots.

Cheers
Edgar
Hi Stefan, I have the same experience about far field calculation. However I found that calculation times get reasonable when you avoid 3D plots. If you just calculate the usual Azimuth and Elevation plots in polar coordinates you have normal postprocessing time demand. The 3D plots are looking nice but what is mostly needed in antenna design are the polar plots. Cheers Edgar

Please login with a confirmed email address before reporting spam

Posted: 1 decade ago May 7, 2012, 5:07 a.m. EDT
I have had a look into the model: isn't your port just a short circuit at the antenna rod? Shouldn't it connect the vertical rod and the ground plane?

Cheers
Edgar
I have had a look into the model: isn't your port just a short circuit at the antenna rod? Shouldn't it connect the vertical rod and the ground plane? Cheers Edgar

Please login with a confirmed email address before reporting spam

Posted: 1 decade ago May 7, 2012, 11:33 a.m. EDT
Hi,

Thank you for answers.

I found the option to disable the recalculation "Update plot when selected". And so I can open such models again. I increased the memory options for the VM.

And about the 3D plot. I tried to recalculate results from a other simulation program, see attached picture. I want to compare the simulation of the both programs. And actually it should be same.

@Edgar: You are right my first idea was to set the port vertical between the antenna rod and the ground plane, like in the attached model. But then I get the errors from the solver. Is there anything missing or wrong?

Thanks so far

--
stefan
Hi, Thank you for answers. I found the option to disable the recalculation "Update plot when selected". And so I can open such models again. I increased the memory options for the VM. And about the 3D plot. I tried to recalculate results from a other simulation program, see attached picture. I want to compare the simulation of the both programs. And actually it should be same. @Edgar: You are right my first idea was to set the port vertical between the antenna rod and the ground plane, like in the attached model. But then I get the errors from the solver. Is there anything missing or wrong? Thanks so far -- stefan


Please login with a confirmed email address before reporting spam

Posted: 1 decade ago May 7, 2012, 11:49 a.m. EDT
Stefan,

I think the port should rather be a horizontal piece that connects the vertical rod at its ground end and the ground plane.

Cheers
Edgar
Stefan, I think the port should rather be a horizontal piece that connects the vertical rod at its ground end and the ground plane. Cheers Edgar

Note that while COMSOL employees may participate in the discussion forum, COMSOL® software users who are on-subscription should submit their questions via the Support Center for a more comprehensive response from the Technical Support team.