Dear Raksha,
for an unshielded isotropic source, the particle fluence
(and thus also the Dose equivalent) is inversely proportional to the surface
of a sphere around the point of origin of the source (so it goes with
1/r^2). Therefore, moving your scorer farther away will reduce the
statistics in the bins, and correspondingly increase the uncertainty if
the binsize remains unchanged. Also with a shielding geometry this
argument in principle still holds (the farther you go away, the less
statistics you get), but the exact proportionality now depends on the
geometry and is difficult to estimate (that's why we use the Monte Carlo
method to calculate the particle distributions across the geometry).
As to how to chose which binning is optimal - as a rule of thumb, it is
the one for which you can achieve an uncertainty of 5% or better in a
reasonable amount of computation time. This may depend on how detailed
your geometry is modeled, in which part of the geometry you want to have
statistical convergence (in other words, an acceptable uncertainty) and
what quantity you are scoring.
To get an idea where your input is converging, it is often helpful to have
a 2- or 3-dimensional USRBIN and plot the uncertainty (checking
the "errors" box in FLAIR's plotting section will display the uncertainty
in % for each bin).
"Variance reduction" techniques can help to improve statistical
convergence, see the lecture material on "Biasing" on one of the FLUKA
courses. In your case, it may be worthwhile to divide your shielding sphere
into several spherical shells and increase the importance gradually
towards larger radii using the "BIASING" card for each shell region (using
factors of 2. in the WHAT(3) is usually a good start, e.g. innermost shell
get 1., next shell gets 2., next shell gets 4., etc.).
Cheers,
Stefan
--
Stefan E. Mueller
Department of Information Services and Computing - Computational Science
and Institute of Radiation Physics
Helmholtz-Zentrum Dresden-Rossendorf
Tel: +49 (0351) 260 3847
Stefan.Mueller_at_hzdr.de
http://www.hzdr.de
Vorstand: Prof. Dr. Sebastian M. Schmidt, Dr. Diana Stiller
Vereinsregister: VR 1693 beim Amtsgericht Dresden
On Fri, 28 May 2021, Raksha Rajput wrote:
>
> Dear Experts,
>
> I have scored Dose Equivalent at 100 cm from the surface of the shield with
> an isotropic source at centre - please find attached the input.
>
> I am getting different outputs for different volume of the Cartesian bins
> used (Cartesian binning in USRBIN) - which seems logical as it is averaged
> over the volume of the bin - but the question is - how to decide the correct
> values ? The statistical errors in all the cases (scores with different
> volume bins) are less than 5 % - so all are acceptable. Is there any other
> check to decide the correctness of the output, that the result is more
> reliable out of many ? Please find attached the .lis files for the scores
> obtained at 100 cm Left and 100 am Right from the shield.
>
> I also found that near surface the outputs are obtained for 8 cc and 16 cc
> bin volumes with acceptable statistical error, but at 100 cm the volume of
> scoring bin has to be bigger, say upto 96cc /128 cc / 400 cc to obtain the
> expected output. Does the isotropic nature of the source has a role to play
> here ? Scattering is air outside the shield ?
>
> Please explain.
>
> Thanks and regards,
> Raksha.
>
>
>
>
>
__________________________________________________________________________
You can manage unsubscription from this mailing list at
https://www.fluka.org/fluka.php?id=acc_info
Received on Sun May 30 2021 - 14:13:57 CEST