Problem of Scattering Field/Total Field tenique in FDTD

Problem of Scattering Field/Total Field technique in FDTD

Hi,

I want to calculate transmission problem by using of FDTD. The source is given by TF/SF technique. As we know, the scattering zone would seperate the UPML and structure(Scattering problem).

My question is, is it possible to set scattering region to be "zero" since the structure I simulate which have to "touch" the UPML boundary.

If it is incorrect if I set scattering zone to be zero, how should I do when defining the structure which would touch UPML under the TF/SF source technique? Or how to modify SF/TF technique for my purpose?

Thank you in advance

TFST with zero SF... That is simply a 3D problem isn't ?

It's a 2D problem. I code it and get some trouble.
Watch the red circle region, should I make the plate to touch the boundary? how?

I want to calculate the transmission of a single slit for a PEC plate.
But I have no idea if I should make the plate to touch the boundary of UPML.
In my opinion, I should make both the edges of the plate to touch the UPML boundary since it would make sense for a real structure.

But I have no idea about how to do it because the SF/TF source technique separate the structure from the UPML. (There is a scattering field zone between them.)

I can define the structure across scattering zone to touch UPML but I did not try it since I think it does not make sense.

How do you handle this problem? Or TF/SF just can be applied only for scattering problem, not for a transmission problem?

Thanks in advance.

Why not use the basic Yee cell with UPML ??

@kohi
It would also face the same problem even when I use basic PML since TF/SF technique always separate the PML layers from the total field zone by scattering field region.

Whatever, thanks for the kindly reply

I dont understand why you must use TFSF... Could you kindly explain? Cant a yee cell algorithim with pml work?

Kohi_boy said:

I dont understand why you must use TFSF... Could you kindly explain? Cant a yee cell algorithim with pml work?

Click to expand...

@kohi

From your post, I think TF/SF may not be the necessary condition for such transmission problem. May you provide other ways or hint about what kind of source technique I could use to give the plane wave?

BTW, I use TF/SF just because I learn it from Taflove's book and it seems lots of people apply the 2-D source by TF/SF;it's all, no other special purposes.

Why not begin with the simple point/line source?

You might also want to read:
1) The finite difference time domain for electromagnetic by Kunz
2) Electromagnetic simulation using FDTD method by Dennis Sullivan

@kohi:
I have tried it before;however,
the simple line(multi sin/cos points) source has to extend the start point and end point into the UPML region or it would radiate cylindrical wave at the two ends for the original given line source. And I have no idea about what to do when I extend the line source into the UPML region since the EM point/line source in the UPML would be different mathematical form with the sin/cos function in the free space.

Also, thank you very much for the providing books, it is useful to me since I have not read the first one by Kunz.
BTW, Merry X'mas!