Re: Questions about EM interactions and scoring

From: Alberto Fasso' <fasso_at_SLAC.Stanford.EDU>
Date: Sat, 17 Nov 2012 03:24:40 -0800 (PST)

Dear Michele,

electron impact ionization with subsequent atomic relaxations and
characteristic x-ray production is not available in the present FLUKA version,
but will be implemented soon.

A point detector (or next event estimator) is not what you describe, which is
what is commonly called "range rejection". Neither of them is available in
FLUKA, as a deliberate choice of the developers. Both techniques in fact
present a number of undesirable aspects, which are discussed at length in
Monte Carlo textbooks.
In short, range rejection neglects important physical effects such as multiple
scattering and bremsstrahlung, and in any case could not be implemented for
energies larger than one or two MeV.
The next event estimator produces a contribution at every collision point,
by calculating the probability that the particle would scatter in the direction
of the scoring point and would reach it without further interaction. That
direction is different from the actual direction followed by the particle.
This technique can be applied only to low-energy photons and neutrons, has an
infinite variance and a very skewed distribution, due to the fact that it
produces many small (often negligible) contributions to the tally, and a few
very large ones. Its results converge to the true average, but mainly from
the lower side, so it is likely to lead to underestimates.

Alberto

On Tue, 13 Nov 2012, Michele Marziani wrote:

> Dear Fluka experts,
>
> I'm evaluating the suitability of Fluka for EMF-only simulations below
> 100 keV. I would like confirmation on a couple of items:
>
> 1) There is no modelling of electron impact ionization with subsequent
> atomic relaxations and characteristic x-ray production. Is it correct?
>
> 2) Is there any way to mimic a "point detector" in MCNP/MCNPX jargon?
> i.e. a partially-deterministic variance reduction technique called "next
> event estimator" because it is a fluence tally at a point if the next
> event is a trajectory without further collision directly to the point
> detector. It would be useful as an alternative to a track-length fluence
> estimator for very small volumes far from the source, at specific
> directions.
>
> Thank you,
>
> Michele
Received on Sun Nov 18 2012 - 20:40:21 CET

This archive was generated by hypermail 2.2.0 : Sun Nov 18 2012 - 20:40:22 CET