Page 1 of 1

Normalization of far field

Posted: Fri 14 Oct 2016, 22:58
by Fabrice
Hi,

I'm using openEMS to design a multi-band antenna. Starting form a scientific paper that shows simulation results with Ansoft HFSS, I have tried to verify the antenna behavior using openEMS. The results match pretty well and it gives me again more confidence in openEMS.

The only mismatch appears when displaying the radiation pattern. After investigation it seems that there is a small issue in the normalization (that appears in several antenna example/tutorial).

For example, in the Simple Patch Antenna (http://openems.de/index.php/Tutorial:_S ... ch_Antenna), the normalization code is

E_far_normalized = nf2ff.E_norm{1} / max(nf2ff.E_norm{1}(:)) * nf2ff.Dmax;

However, Dmax is a power ratio and its square root has to be used to normalize the fields if my understanding is correct.

I replaced the code by:
E_far_normalized = nf2ff.E_norm{1} / max(nf2ff.E_norm{1}(:)) * sqrt(nf2ff.Dmax);

and get coherent results when displaying radiation patterns in Paraview in dBi.

Could you confirm my interpretation?

I hope it helps. Best regards,

Fabrice

PS: thanks for providing to the community such nice simulation tool. I'm amazed how easy it is to use openEMS, to show results in very powerful ways (graphs, 3d diagrams, field animation).

Re: Normalization of far field

Posted: Fri 14 Oct 2016, 23:21
by thorsten
Hi Fabrice,

yes you are correct, this formula was wrong.
If you look some lines above:

Code: Select all

% conventional plot approach
% plot( nf2ff.theta*180/pi, 20*log10(nf2ff.E_norm{1}/max(nf2ff.E_norm{1}(:)))+10*log10(nf2ff.Dmax));
you see a correct approach using 20*log10 for the fields and 10*log10 for Dmax ...

But the export to vtk indeed needs a square root...
I guess I will have to revise all examples and Tutorials and look for that mistake. After all my lazy self uses copy and paste far to often ;)

regards
Thorsten