scattering cross section comsol
does anyone know how to get the total scattering cross section of various kind of objects in the Electromagnetic module of COMSOL?
Thanks in advance.
Hi!
I have had a similar problem some time ago. I guess your first problem is to separate the scattered field from the incident field. That's easy. Then you have the scattered field and you only need to find a contour over which you should integrate the scattered power. I did it 'manually' - I exported scattered field data from COMSOL and then processed that data (knowing the geometry) separately from COMSOL.
I guess there should be an more straightforward way, but I didn't have time to play around.
Regards
COMSOL itself has the integration function, so maybe you can draw
some contour manually but integrate automatically.
Thanks all guys! :D
Yes, you are rigth wandering heart - had I known this, it would save me quite some time :)
Is it possible to integrate flux e.g. for 2D ?(i.e. is it possible that the integration recognizes the line segments as vectors)? This is necessary if a irregularly contour is used...
Hi all,
as I need to calculate the scattering cross sections, I have to integrate the far field.
Can anyone tell me how to get the far field? Is there some postprocessing to do that?
Cheers
What do you need the far-field for?
The only reason why the scattering cross section is calculated using the asymptotic far-field expressions is to obtain analytic expressions. If you already have the numerical solution for the fields, I see not point converting them to far-field. I even don't know if it is possible... You will get the same result for the scattered power flux (and hence the scatt. cross section) if you integrate it along any contour outside the scatterer.
Regards
Hi goran 1901,
could you please give me more details on the procedure to follow to calculate the scattering cross section of an object in COMSOL?
Thanks
Hi iaia,
I just figured out that you're the same person from this other topic. So, you're doing cloaks?
Ok, this is how I did the calculation of the power flux for the first time:
1. draw a contour around the scatterer (in my case, this was a square)
2. do the simulation
3. go to Export and choose Postprocessing data
4. there you can choose which quantity you want to export (Ez, Hx, Hy) and, I think, whether you want the coordinate data or the mesh data
5. export all the quantities you need (I needed all three field components and I exported separately the real and imaginary part)
6. write the program which calculates the Poynting vector (ie the component perpendicular to the correspongin surface ie line in this 2D case)
So, this was the way I did it and, as you can see, it was quite an effort. I am telling you this because this is exactly what I did and I got results which were in very good agreement with the theory (the difference was ~1% or less) - for the scattering cross section. The way below should be just the same, however I haven't tried it because now I am not working on cloaks and scattering any more.
Now, I would do it otherwise: you can tell Comsol to perform contour integration (postprocessing -> boundary integration) and in doing so you can choose which quantities to integrate. So, if you have a simple contour (like a square) you can easily calculate the perpendicular components of the Poynting vector. For example: for the horizontal lines of the square contour, the perpendicular component of P is Py=(1/2)* Re(Ez* (Real(Hx)-i*Imag(Hx))) (its Hx conjugated but I am not sure how can you do conjugation in Comsol - the above expression is one possibility. This way, you can get the segments of the integral for each of the four sides of the square. You can further improve this to do everything in one step if you play a bit with angles: phi=atan2(y,x), and integrate (1/2) Py*sin(phi)+(1/2)Px*cos(phi), etc.
I hope this is clear enough.
Regards
To iaia,
I saw there is an option in Physcis in COMSOL, but I haven't touched that now.
To goran 1901:
thank you for your very detailed answer! I'm sure this will help me :D
To wandering heart:
thank you, I'll also try this option.
Cheers