how to get E, H distribution for arbitrary frequency

my method now is: after FDTD calculation, we can get E. H data in time domain. To get E.H distribution for arbitrary frequency in post-processing, we have to store all the E.H data in time domain followed by FFT.

However, if i got a big structure, say 1000,000 cells, it is not so practicable to store all the data in time domain.

anybody has any expeeriences regarding this issue? thanks a lot in advance.

regards,
tenderne[/b]

There is an easy solution to this. You will need to continually perform an integration that computes your frequency-domain field values. This way you only need to store one more set of field values for each frequency of interest. You can down a copy of my PhD dissertation. I summarize FDTD in Chapter 6 and describe this Fourier transform procedure in section 4.2.9 starting on page 98. The link to the dissertation is:

https://www.edaboard.com/download.php?id=124081

-Tip

hi, rrumpf,

thank you very much for the reply. but if we don't know the frequency before FDTD caculation, then how? thanks.

regards,
tenderne