Dear Anna
The "fake" double spots are due to incorrect tracking settings.
Unfortunately the transport in magnetic field inside materials involves
the combination of angular deflections due to the magnetic field and the
multiple scattering. This combination is unavoidably inexact (see the
discussion below taken from a different case, but same problematic,
which unfortunately was not forwarded to the list), and, in order to get
it "correct" it requires care in setting step lengths in regions where
magnetic fields and materials co-exist, eg your target.
In your case in absence of any other constrain (the 25 deg constrain
implies huge step lengths). the target is span in one single long step
which on top gets cut at the border (-> multiple scattering
inaccuracies). Just forcing the step length in the target to 1 cm solves
the problem (see attached input, stepsize cards, and plot).
Further reducing the steplength does no longer change the results.
As a comment, when I refereed to the accuracy in finding boundaries wrt
the region thickness I meant that what you had was too large not too
small. With a couple of 100 um thick regions the bounding finding
accuracy should be << 100 um at least for those regions. I corrected it
in your input, even though it is not the cause of your problem.
In the future we'll try to devise a (hopefully) automatic strategy for
setting those step sizes in materials in absence of user input and when
magnetic fields are present, however it is not a straightforward task.
Below the old discussion of a couple of years ago with another user (in
that case there was a long gas region with magnetic field).
Alfredo
...
there is nothing really wrong with the treatment of multiple scattering
in gases in presence of a magnetic field. Only, the approximation used,
which is the same for regions without magnetic field, which consists in
applying the cumulative effect of multiple Coulomb scattering over the
step
length at the end of the step, is significantly more sensitive,
accuracy-wise,
to the step length, given that the magnetic field "turns" the particles
on its own. In order to achieve better accuracy, you can proceed in two
ways
a) reducing the default step length, 10 cm in magnetic field
corresponding roughly to 500 individual scatterings in your case, eg to
1 cm.
This can be accomplished by adding in your example the following cards:
MULSOPT 1.0 1.0
GLOBAL
STEPSIZE 1.0 GASVOL GASVOL
(see the manual for their meaning)
b) invoking single scattering treatment everywhere in your gas regions.
This possibility assures the maximum accuracy at the prize of a
significant CPU penalty, but if you are sensitive to sub-millimeter
accuracy over tens of metres it is the best choice. For your example,
this can be achieved adding:
MULSOPT 999999990 BEAMAT BEAMAT
I checked that both a) and b) give reasonable results in your example,
with a) being similar to b) (but b) is guaranteed to give the "correct"
result withing the limit of our Coulomb scattering implementation).
On 2024-03-14 10:19, Senger, Anna Dr. wrote:
> Dear Alfredo.
>
> The pictures present beam profile. I checked energies of these hot
> spots: the are exact proton beam energy.
>
> I checked other settings for magnetic field and used simple
> implementation: Bx=0, By=-0.45, Bz=0. The behavior is the same (see
> picture). Argon and PROPANEL (and H-LIQUID) are standard materials
> from FLAIR:
>
> * Propane liquid C3_H8
> *
> MATERIAL 0.43
> PROPANEL
> COMPOUND -0.182855 HYDROGEN -0.817145 CARBON
> PROPANEL
> MAT-PROP 52.0 PROPANEL
> STERNHEI 3.5529 0.2861 2.6568 0.10329 3.562
> PROPANEL
>
> I add new input with new settings. It is simple to use. May be I need
> to add some physics or/and transport parameters, but I don't
> understand what exactly.
>
> Regards
> Anna
>
> -------------------------
>
> Von: aferrari <alfredo.ferrari_at_mi.infn.it>
> Gesendet: Mittwoch, 13. März 2024 21:43
> An: Senger, Anna Dr.
> Cc: 'fluka-discuss_at_fluka.org'
> Betreff: Re: [fluka-discuss]: Proton beam profile after liquid target
> in magnetic field
>
> Dear Anna
>
> I'll try to help understanding what you see, but first of all I have a
>
> few questions:
>
> a) the plots you sent, do they refer to the BEAMPART or the ALL-PART
> binnings?
> b) could you provide the magnetic field routine, or at least some
> guess
> about its intensity and direction?
> c) is the ARgon target Argon gas and if so at which pressure/density?
> d) what is PROPANEL (composition? density?)?
>
> As an observation, possibly unrelated to your question, a maximum
> deflection per step of 40 deg is quite large and it could result in
> relatively sloppy tracking. Also the requested boundary accuracy of
> 0.05
> cm -> 500 um which is comparable or larger than the the thickness of
> some of your volumes, again it is likely unrelated to what you observe
>
> but it can result in inaccurate tracking.
>
> Ciao
> Alfredo
>
> On 2024-03-12 10:48, Senger, Anna Dr. wrote:
>> Dear FLUKA experts.
>> I'm not so familiar with simulation of the proton beam and liquid
>> target, and I would like to ask: if this beam profile in magnetic
>> field is artefact of the FLUKA implementation for gas and liquid
>> targets or there are some problems.
>> I used simple input: only vacuum as a cave, different target
> materials
>> and magnetic field in target region. I got different beam profiles
> at
>> 15 m from the target for different target materials. I expected to
> get
>> some smooth distribution. Could you please help me to understand my
>> results.
>>
>> Best regards
>> Anna Senger
>>
>> ______________________________________
>>
>> Dr. Anna Senger
>>
>> Office KBW 2.01
>> Tel. +49(0)615971 1844
>> E-Mail A.Senger_at_gsi.de
>>
>> FAIR - Facility for Antiproton and Ion Research in Europe GmbH
>>
>> Planckstr. 1
>> 64291 Darmstadt
>> DEUTSCHLAND
__________________________________________________________________________
You can manage unsubscription from this mailing list at
https://www.fluka.org/fluka.php?id=acc_info
Received on Fri Mar 15 2024 - 13:41:28 CET