Dear Alex,
a second thought on the evaluation of USRBIN errors: you can
actually make them to coincide between a cartesian and region based
scoring, by taking a very coarse mesh in which bins coincide with
regions. This is very simplistic and not realistic, of course, but
it is useful to understand.
I send you a modified version of the standard Fluka example in
which the target is divided in two regions: I have defined both
a cartesian (1 bin in x/y and 2 bins in z) and a region scoring
and you can see that the outputs coincide both as average values
and as statistical errors (please note the different normalization
of the Fluka output: with region scoring you have to divide
manually the value by the region volume).
Best,
Andrea
Il 01/11/2018 16:57, Ševčik Aleksandras ha scritto:
> Dear Andrea,
>
> Thank you very much. As bins are much smaller than a typical region, the error gets much higher, which makes sense,
>
> Alex
>
> -----Original Message-----
> From: Andrea Fontana <andrea.fontana_at_pv.infn.it>
> Sent: Thursday, November 1, 2018 16:17
> To: Ševčik Aleksandras <aleksandras.sevcik_at_ktu.edu>; fluka-discuss_at_fluka.org
> Subject: Re: [fluka-discuss]: regarding relative error in USRBIN X-Y-Z and REGION
>
> Dear Alexander,
> when you request a region based or a cartesian based scoring in Fluka, you are evaluating two different quantities:
>
> - in the first case (region), you are globally evaluating the
> average value of your score (let us say the deposited energy) over
> the whole region;
> - in the second case (cartesian binning), you are locally evaluating
> your score in each single bin of your mesh.
>
> The statistical errors calculated by Fluka reflect this difference.
>
> Perhaps this is not clearly stated in the manual, but you can find some (hopefully) helpful details in these old discussions:
>
> http://www.fluka.org/web_archive/earchive/new-fluka-discuss/10023.html
> http://www.fluka.org/web_archive/earchive/new-fluka-discuss/6342.html
>
> Andrea
>
> Il 01/11/2018 10:44, Ševčik Aleksandras ha scritto:
>> Hello all,
>>
>> I am doing the simulations in which scoring USRBIN Dose/DepEnergy in
>> X-Y-Z and Region. When I look in the scored Region output files,
>> typically I see very small error % value (<1%) for almost all the
>> regions, however, when I check X-Y-Z bin file (usually through
>> SimpleGeo) I see that error % is typically very high for the most of
>> regions except maybe only in areas directly affected by the beam. I
>> understand that I missed some important information about how Fluka
>> calculates relative error for different scoring tools. Could someone
>> please comment or give the link where to read further on this matter
>> as I, unfortunately, cannot find explanation ,
>>
>> Attached two files for illustration, usrbin region scoring txt and
>> illustration of usrbin xyz scoring for error values
>>
>> Regards
>>
>> Alexander
>>
> --
> ========================================================================
> Dr. Andrea Fontana tel: +39 0382 987991
> Istituto Nazionale fax: +39 0382 423241
> di Fisica Nucleare
> Sezione di Pavia e-mail: andrea.fontana_at_pv.infn.it
> Via Bassi 6 web : www.pv.infn.it/~fontana
> 27100 PAVIA, Italy
> ========================================================================
>
>
--
========================================================================
Dr. Andrea Fontana tel: +39 0382 987991
Istituto Nazionale fax: +39 0382 423241
di Fisica Nucleare
Sezione di Pavia e-mail: andrea.fontana_at_pv.infn.it
Via Bassi 6 web : www.pv.infn.it/~fontana
27100 PAVIA, Italy
========================================================================
__________________________________________________________________________
You can manage unsubscription from this mailing list at
https://www.fluka.org/fluka.php?id=acc_info
Received on Thu Nov 01 2018 - 18:44:27 CET