Sinuous Antenna simulation

How to use openEMS. Discussion on examples, tutorials etc

Moderator: thorsten

Felix_Yusopov
Posts: 13
Joined: Sat 08 Sep 2018, 01:05

Sinuous Antenna simulation

Post by Felix_Yusopov » Sat 29 Sep 2018, 04:02

Hi all,

I was trying earlier today and yesterday to simulate a 2 arm, balanced feed, sinuous antenna with the following pattern dimensions (in meters):
2arm_large_sinuous.PNG
2 arm Sinuous Antenna, as designed in MATLAB
2arm_large_sinuous.PNG (26.11 KiB) Viewed 7696 times
I am having a lot of trouble with the geometry. (I'm still new to openEMS)

The program that was used to design the antenna generates 1857 different (x,y) coordinate pairs to define each arm. I used spline() to reduce to this about 300 coordinate pairs, as this was around the minimum I needed to correctly define the antenna.

In attempting to define the geometry, I used AddPolygon() and sent it the 600 pairs of coordinates for both arms. It got stuck running (quoting from the MATLAB command window, in order):

SmoothMeshLines2>calc_gaps (line 132)
idx = interp1( temp_lines, 1:numel(temp_lines), lines(n), 'nearest' );

In SmoothMeshLines2 (line 77)
gap = calc_gaps(lines,old_gap,max_res);

In AutoSmoothMeshLines (line 92)
out_lines{m} = methods{m}(lines, max_res, ratio, 'CheckMesh', false);


In SmoothMesh (line 54)
mesh.x = AutoSmoothMeshLines(mesh.x, max_res(1), ratio(1), varargin{:});

After several hours I force quit MATLAB and attempted to render just a single arm. Another 3-4 hours later, it still had made no progress.

The shape simply seems too complex to render using AddPolygon().

My question is, in my ignorance, am I going about rendering this geometry the wrong way? Or is it simply impossible for openEMS to render such a complicated shape?

I've enclosed my MATLAB simulation code below, along with a .txt file with the coordinates of the 2 arms.
Attachments
LargeSin.txt
File containing coordinates for the arms
(116.07 KiB) Downloaded 288 times
Sinuous_antenna_sim.m
simulation file for the sinuous antenna
(4.86 KiB) Downloaded 297 times

thorsten
Posts: 1425
Joined: Mon 27 Jun 2011, 12:26

Re: Sinuous Antenna simulation

Post by thorsten » Sat 29 Sep 2018, 08:41

Hi,

I hopefully will find some time to have a closer look. In theory a large polygon like this should not be an issue. But I have never tried one so far.
My question would be. At which part do you get stuck? At simulation? Showing the structure with AppCSXCAD (which you not really need to be able to)?
Can Matlab show the poly as a figure? Does the automesh stuck (create just a homogeneous manually!)?

regards
Thorsten

thorsten
Posts: 1425
Joined: Mon 27 Jun 2011, 12:26

Re: Sinuous Antenna simulation

Post by thorsten » Sun 30 Sep 2018, 18:01

I tried to run your script but the line with

Code: Select all

A1 = load('LargeSin');
Does not load the txt file (at least not with Octave). Adding the .txt suffix did not help...

In any case I guess a "plot" of the loaded data (that you want to add to AddPoly) does look ok?
At what point does it stop working?
Can you maybe attach the xml file created by "WriteOpenEMS" ?

regards
Thorsten

Felix_Yusopov
Posts: 13
Joined: Sat 08 Sep 2018, 01:05

Re: Sinuous Antenna simulation

Post by Felix_Yusopov » Mon 01 Oct 2018, 05:29

Hi Thorsten,

Thank you so much for your response. The Plot of the data (the xy coordinates of the antenna arms) is definitely fine. Where I'm running into problems is the SmoothMesh() function. The MATLAB compiler is unable to get past that stage. The window showing the geometry never opens up. Another unfortunate result of this is that I cannot include an .xml file of the sim, as the compiler never reaches the stage at which one is generated.

And I'm sorry about the text file not opening correctly. Enclosed below is an updated version of the sim code that includes the code I used to generate the sinuous antenna coordinates in the first place.

Considering the complexity of the structure, is it still possible to simulate it?

Best Regards,
Felix
Attachments
Sinuous_antenna_sim.m
Updated simulation file
(8.73 KiB) Downloaded 310 times

HexAndFlex
Posts: 11
Joined: Sat 06 Oct 2018, 09:11

Re: Sinuous Antenna simulation

Post by HexAndFlex » Sat 06 Oct 2018, 11:19

This has some similarities to a spiral antenna I have recently designed. See link for info https://hexandflex.com/2018/10/05/spira ... enems/amp/

First, my guess is cylidrical coordinates are better than cartesain for antennas like this. Though I confess i didnt try cartesian.

Second, I dont think detect edges doesnt work well for this type of model.

Im still learning this myself, so dont take this as gospel.

Felix_Yusopov
Posts: 13
Joined: Sat 08 Sep 2018, 01:05

Re: Sinuous Antenna simulation

Post by Felix_Yusopov » Mon 08 Oct 2018, 20:38

Hi Hex,

Thanks for you response. Any and all help is welcome. I'll give it a shot with cylindrical, and leave out detectEdges()

thorsten
Posts: 1425
Joined: Mon 27 Jun 2011, 12:26

Re: Sinuous Antenna simulation

Post by thorsten » Mon 08 Oct 2018, 20:42

Hi,

I disagree, the cylindrical mesh does not help here. It is only useful if your structure has some really cylindrical shapes. From your screenshots you should just use a Cartesian homogeneous mesh over your structure....

br
Thorsten

thorsten
Posts: 1425
Joined: Mon 27 Jun 2011, 12:26

Re: Sinuous Antenna simulation

Post by thorsten » Mon 08 Oct 2018, 20:49

Use only one or two cells for your substrate, leave out the detect edges on your arms (just crazy) and everythings starts to look fine... Just use a fine homogeneous mesh over the entire antenna area...

Felix_Yusopov
Posts: 13
Joined: Sat 08 Sep 2018, 01:05

Re: Sinuous Antenna simulation

Post by Felix_Yusopov » Mon 15 Oct 2018, 20:17

Thank you so much for your help Thorsten and Hex! The simulation runs great now, and the results align very closely with what I got from the VNA

PaulUK
Posts: 79
Joined: Wed 28 Oct 2015, 14:23

Re: Sinuous Antenna simulation

Post by PaulUK » Tue 20 Nov 2018, 16:54

Hello Felix,

Are the links to your files still active? I cant seem to download the files. Thanks.

Post Reply