On Mon, 10 Apr 2006, X. Y. AO wrote:
How to get the integral of the normal component of the time-averaged
Poynting vector over a plane? Meep runs with complex fields, drived by
As described in the manual, Meep uses real fields unless you set
force-complex-fields? to true (see the Meep Reference).
multiple continuous-wave electric dipoles with random phase factors. The
number given by (add-flux) is increased with TIME (run-until TIME). It
The add-flux thing is for getting flux spectra, which is not what you want
it sounds like.
seems that (meep-fields-flux-in-box fields dir box) can work, but what
does "fields" here means?
Yes, that will work, although you will want to set force-complex-fields?
in order that it give you the time average flux rather than a snapshot.
fields is an abstract data type (corresponding to the C++ type
meep::fields*) describing all of the fields in the current simulation.
There is a global variable named "fields" (described in the Meep
Reference) that is initialized when you start the simulation run), which
you will want to use here.
(In principle, you can have multiple meep::fields objects to run multiple
simulations at once, but that is not the normal mode of operaton.)
The next version of Meep will have (flux-in-box dir box) convenience
functions that automatically use the global fields object.
Another question, how to get the steady-state field distribution
(complex amplitude)? (output-hfield-z) etc. can only output fields in a
certain time.
(To others on this list: "steady-state" response is the usual term for the
exp(-iwt) response field from an exp(-iwt) source, after all transients
have died away.)
To get the steady-state response from a CW source, currently the best way
is to simply run for a long time after smoothly turning on a source (you
will want to set the fwidth or width parameter of the continuous-src to
make it turn on smoothly instead of suddenly).
If you want the complex amplitudes, you need to set force-complex-fields?
to true as mentioned above. Of course, there is an arbitrary phase
depending on what time instant you look at, but you can simply look an
integer number of periods after your source starts if you want to look at
a time when J has zero phase.
There is an hidden experimental feature to solve for the steady-state CW
response directly by solving the associated linear equations via a
biconjugate gradient algorithm. This is not really documented yet because
its convergence is still problematic, but you can play with it by running:
(meep-fields-solve-cw fields tol)
where "tol" is a relative tolerance (e.g. 1e-4) in the fields and "fields"
is the global "fields" variable. You'll want to do this after running for
at least a few periods to initialize the fields to something reasonable.
Steven
_______________________________________________
meep-discuss mailing list
[email protected]
http://ab-initio.mit.edu/cgi-bin/mailman/listinfo/meep-discuss
