Input of electric fields

[GoogleCodeExporter]

Implement reading of initial field for iterative solver and incident
electric field from file. This will enable any incident field and allow
more flexibility in complex simulations. For instance, to adaptively
decrease dipole size reusing previous results.

Original issue reported on code.google.com by yurkin on 26 Nov 2008 at 6:27

[GoogleCodeExporter]

is this feature difficult to implement? we would like to test vectorial 
Gaussian beams in Adda. Reading the incident field from a file would be 
sufficient for this task.

Original comment by vita...@gmail.com on 1 Sep 2011 at 1:48

[GoogleCodeExporter]

There are several complexity levels for this issue. 

The simplest approach is to read the file in the same format as is currently 
used for reading shape (using six real numbers for complex vector instead of 
one integer, which is now used for domain number). Then the grid in the file 
should perfectly match the one inside ADDA, which is a complete responsibility 
of the user. And the user should perform a transformation of integer 
coordinates (in numbers of dipoles) in the shape file into real coordinates 
(e.g. in nm) to compute the incident field. 

Of course, this is not a critical limitation. One can for example run a test 
run with 
"-prognosis -save_geom -store_int_field" which will produce two files with 
one-to-one correspondence (e.g. by line numbers) between these two types of 
coordinates.

A harder issue is to implement some kind of consistency check or even to accept 
incident field on arbitrary grid with interpolation to a dipole grid if 
necessary.

Finally, since you are interested in this issue, I will try to implement the 
simplest approach by next release. The time frame is, hopefully, about one 
month.

Original comment by yurkin on 1 Sep 2011 at 5:26

• Added labels: Milestone-1.1

[GoogleCodeExporter]

Hi Maxim, I appreciate your prompt reaction. If you think that I could be (with 
very limited C++ programming skills) of any help, please let me know. With 
thanks, vita.

Original comment by vita...@gmail.com on 2 Sep 2011 at 9:59

[GoogleCodeExporter]

Hi Vita,
It seems to be really easy to implement. So I just need to sit down and do it. 
But I will definitely use your help for testing.

Original comment by yurkin on 2 Sep 2011 at 5:49

[GoogleCodeExporter]

r1083 adds new command line option '-beam read <filenameY> [<filenameX>]', 
which should be sufficient for playing with incident field. The beam is 
specified in the particle reference frame exactly in the same format as from 
'-store_beam'. Therefore '-prop ...' and '-orient ...' has only minor effect:
a) by defining the reference frame for calculation of amplitude and Mueller 
matrices. However, this should not effect values obtained by integration over 
the whole scattering angles, e.g. Csca.
b) by slightly affecting of polarizability value (for LDR, CLDR, and possibly 
SO formulations)

Since ADDA expects exact correspondence (including order of lines) of the input 
beam file to the internal dipole distribution, the recommended workflow is like 
this
1) adda <options defining scattering problem> -store_beam [-sym no]
'-sym no' can be used to enforce production of two incident beams even for 
symmetric shapes.
2) copy the files IncBeam-Y (and IncBeam-X) from the output directory to the 
working folder (usually one level up)
3) replace the values of plane wave in the obtained files by the values 
computed by yourself
4) adda <options defining scattering problem> -beam read IncBeam-Y [IncBeam-X] 
[-sym enf]
'-sym enf' can be used to indicate that both shape and the beam are 
rotationally symmetric. Then second beam file is not needed.

Current limitations are:
1) issue 30 - also workarounds are possible (see also '-sym enf' above). 
Moreover, one can just run ADDA with single beam file. ADDA will correctly 
calculate all values stored in *-Y files, but then exit with an error.
2) issue 134 (although if tilted plane wave is supplied, Cext would be 
correct), and issue 135.

r1083 - 2d45e2be79ea912d377c075862c99e50c4de0222

30, #134, #135

Original comment by yurkin on 23 Nov 2011 at 4:35

[GoogleCodeExporter]

Half of the issue (input of incident beam) is implemented (see above), at least 
in beta stage. So I shift the milestone for the remaining parts to the next 
release.

Original comment by yurkin on 21 Mar 2012 at 4:56

• Added labels: Milestone-1.2
• Removed labels: Milestone-1.1

[GoogleCodeExporter]

This issue was closed by revision r1223.

r1223 - c8154735ef2a3957da176acdcfe28d7ec55d1d6a

Original comment by yurkin on 4 May 2013 at 5:16

• Changed state: Fixed

[GoogleCodeExporter]

In addition to previous comment - r1223 also changed handling of errors. Now if 
only one field is given (when two are required), the error will appear 
immediately (before processing Y polarization). So use of "-sym enf" in 
workaround described above is now mandatory.

r1223 - c815473

Original comment by yurkin on 4 May 2013 at 5:19

[myurkin]

/cc @vitakar