Re: [fluka-discuss]: Problem with biasing

From: Francesc Salvat-Pujol <>
Date: Wed, 31 Jul 2019 16:06:19 +0200

Dear Vladimir,

At first glance, I think you are experiencing the result of an
unfortunate choice of region importance values:

- Note that with the first biasing card you are setting importance 0.5
  from regions RmodL (number 5) to Rshell (number 6).

  Then, with the second card, you set importance 2 from region RmodR
  (number 4) to region Rbox (number 7), overriding what you intended to
  do with the previous card.

  Thus, regions in your geometry will have importance of either 1 or 2,
  but never 0.5 as you intended. This will allegedly just result in
  "wasting" a bit of CPU time.

- To make sure you control what importance is assigned to which region,
  search in the output file for "Particle importances" or, for a more
  visual approach, note that in Flair you can display the value of the
  importance assigned to each region in a new layer in the Geometry tab.
  Let us know if you need help with this.

- A few of your USRYIELD cards score neutrons from Rbox to subsequent
  regions Rcount1, Rcount2, Rcount3, respectively, at various positions
  X within Rbox. However, whereas Rbox has importance 2, your regions
  Rcount1, Rcount2, Rcount3 have importance 1. I fail to understand the
  virtue of this and would suspect it can actually be detrimental: some
  neutrons that were initially "split" then play Russian roulette when
  they cross into Rcount1, thwarting your efforts somewhat.

- Try a consistent scheme (assign increasing importance to subsequent
  regions) and maybe a bit more aggressive, i.e., importance 2 in RmodR,
  4 in Rshell, 8 in Rbox (and the counters). You may also want to
  effectively cut Rbox at various values of x (this is a very long
  region with considerable attenuation) and assign increasing
  importances to the farther regions.



On Mon, Jul 29 2019, at 08:46 +0300, babintsev wrote:
> Dear experts,
> I want to enlarge neutron statistics  on the right side
> of the target with the same number of beams:
>      lead box in moderator ... and 3 counters on the right side
> For this I use biasing : divide moderator into 2 parts : left and right
> I use W= 2 to the right  side
>        = 1/2 for the left side in moderator
> Then I compare results before biasing and after that.
> Results :
>   () Total number of low energy neutron interactions: 6025716
>                 with biasing
>                  ... it was                            3525982
>                                                     no biasing
>   () CPU  - approximately the same
>   () N neutrons in the counters is less with biasing !?
>      example with counter-3 (C3) :
>                    1.5200707E-02  +/-   2.631857     % (no biasing)
>                    1.3233015E-02  +/-   9.833910     %  with biasing
>              ------------------------
> so, I think that these are strange results
> May be I have errors somewhere ?
> thank you
> Vladimir

Francesc Salvat Pujol
CH-1211 Geneva 23
Tel: +41 22 76 64011
Fax: +41 22 76 69474
You can manage unsubscription from this mailing list at
Received on Wed Jul 31 2019 - 17:19:34 CEST

This archive was generated by hypermail 2.3.0 : Wed Jul 31 2019 - 17:19:40 CEST