- Contemporary messages sorted: [ by date ] [ by thread ] [ by subject ] [ by author ] [ by messages with attachments ]

From: Ševčik Aleksandras <aleksandras.sevcik_at_ktu.edu>

Date: Tue, 25 Feb 2020 19:23:34 +0000

Hello again

Unfortunately my lack of skill in programming and deeper understanding prevented me from implementing the divergence into source.f for rectangular beam cases. In cases anyone has used/developed similar or is able to provide step-by-step (literally) solution I would be insanely grateful.

Rgds

Alex

-----Original Message-----

From: Francesc Salvat-Pujol <francesc.salvat.pujol_at_cern.ch>

Sent: Thursday, July 25, 2019 20:29

To: Ševčik Aleksandras <aleksandras.sevcik_at_ktu.edu>

Cc: fluka-discuss_at_fluka.org; owner-fluka-discuss_at_mi.infn.it

Subject: Re: [fluka-discuss]: how to include the divergence in the source.f / photons

Hey Alex,

If I get you right, you want to sample beam particle initial directions with a rectangular divergence (i.e. cosine uniformly sampled in a given

interval) around an arbitrary direction theta_f and phi_f in your modified source.f.

On paper, what one would ordinarily do is first sample the initial direction with the desired rectangular divergence around the z axis and, then, apply a rotation onto the final direction. So you're almost there.

- I believe for a rectangular beam the divergence is to be understood as

a full width. Thus, to sample the (cosine of the) polar angle one

would rather do

CDH = COS ( HLFHLF * DIV )

COSTH = ONEONE - ( ONEONE - CDH ) * FLRNDM ( TTT )

alternatively

COSTH = CDH + ( ONEONE - CDH ) * FLRNDM ( TTT )

where DIV is the divergence in radians, and the azimuthal angle you

sample uniformly from 0 to 2*pi like you already do. By the way, the

DIVBM variable should still have the value you passed in WHAT(3) of

your BEAM card, already converted to radians, but double check in your

particular case.

- You then construct T{X,Y,Z}FLK and ensure it is unit normalized. If

you chose to do smth a la TZFLK(NPFLKA) =SQRT(...) keep an eye on its

sign (for big DIV).

- The resulting T{X,Y,Z}FLK should be fine, but they are still sampled

around the z axis. Let d_i=(tx,ty,tz). To get the "final" direction,

d_f=(tx',ty',tz'), one would apply on d_i a rotation of angle theta_f

around the y axis, R(theta_f,y), and, next, a rotation of angle phi_f

around the z axis, R(phi_f,z).

To be explicit, see the (hopefully not too far off) attached recipe, without any particular effort to be numerically robust (which is a bad idea... and exactly the reason why director cosines are preferrable instead of angles...), but it will get you started. At the very least, make sure that the final T{X,Y,Z}FLK components are normalized (such that the sum of their squares is ONEONE).

Cesc

On Tue, Jul 23 2019, at 18:36 +0000, Ševčik Aleksandras wrote:

*>
*

*>Thank You, Sougata,
*

*>
*

*>I meant BEAMPOS WHAT(4) & (5) , i.e. the ones which gives the direction cosines of the beam itself.
*

*>The divergence works itself great when beam is straight, with BEAMPOS WHAT(4)&(5) = 0...
*

*>
*

*>In case anyone can help with this question, also with the coding for
*

*>the rectangular beam divergence, kindly please,
*

*>
*

*>Regards
*

*>Alex
*

*>
*

*>-----Original Message-----
*

*>From: Sougata <sougata_at_barc.gov.in>
*

*>Sent: Tuesday, July 23, 2019 13:02
*

*>To: Ševčik Aleksandras <aleksandras.sevcik_at_ktu.edu>
*

*>Cc: fluka-discuss_at_fluka.org; owner-fluka-discuss_at_mi.infn.it
*

*>Subject: RE: [fluka-discuss]: how to include the divergence in the
*

*>source.f / photons
*

*>
*

*>Dear Alex,
*

*>
*

*>For your query (1).
*

*>Any parameter from source card may be passed to source.f file through whasou(#).
*

*>
*

*>e.g
*

*>Coding for 200 mrad divergence beam where WHAT(3)=0.98007 im source card.
*

*>
*

*> * Cosines (tx,ty,tz)
*

*> COSTH = ONEONE-(1-whasou(3))*FLRNDM(TTT)
*

*> IF(COSTH.GT.ONEONE) COSTH=ONEONE
*

*> IF(COSTH.LT.-ONEONE) COSTH=-ONEONE
*

*> PHI = TWOPIP*FLRNDM(SSS)
*

*> UBEAM=SQRT(ONEONE-COSTH*COSTH)*COS(PHI)
*

*> VBEAM=SQRT(ONEONE-COSTH*COSTH)*SIN(PHI)
*

*> WBEAM=COSTH
*

*> TXFLK (NPFLKA) = UBEAM
*

*> TYFLK (NPFLKA) = VBEAM
*

*> * TZFLK (NPFLKA) = WBEAM
*

*> TZFLK (NPFLKA) = SQRT ( ONEONE - TXFLK (NPFLKA)**2
*

*> & - TYFLK (NPFLKA)**2 )
*

*>
*

*>Please check.
*

*>
*

*>Regarding your query (2). I am not so experienced. Better consult some expert person.
*

*>
*

*>Regards,
*

*>--
*

*>SOUGATA RAKSHIT / सौगत रक्षित
*

*>Scientific Officer(D) / वैज्ञानिक आधिकारीक(D) RSSD / आर एस एस डि Bhabha
*

*>Atomic Research Centre / भाभा परमानु अनुसंधान केंद्र
*

*>Mumbai-400 085 / मुंबई-400 085
*

*>Tel: +91-022-2559 2214/ दूरभाष: +91-022-2559 2214
*

*>
*

*>On 2019-07-23 12:57, Ševčik Aleksandras wrote:
*

*>> Dear Sougata Rakshit,
*

*>>
*

*>> Wonderful, it works indeed!
*

*>> But may I ask two more moments here:
*

*>>
*

*>> 1) I need to set up the certain values ( 0.2 ) for BEAMPOS WHAT(4) &
*

*>> (5) - direction cosine of the beam with respect to the x & y-axis. It
*

*>> does not take values from the card now when I use this modified
*

*>> routine. What additional line should be included for this?
*

*>> 2) also this code works for angular flat beam only, but not for
*

*>> rectangular flat beam - does the code differs much in such case?
*

*>>
*

*>> If you or anybody else could find an opportunity to help with this, I
*

*>> will greatly appreciate such help,
*

*>>
*

*>> Regards
*

*>> Alex
*

*>>
*

*>> -----Original Message-----
*

*>> From: Sougata <sougata_at_barc.gov.in>
*

*>> Sent: Tuesday, July 23, 2019 07:47
*

*>> To: Ševčik Aleksandras <aleksandras.sevcik_at_ktu.edu>
*

*>> Cc: fluka-discuss_at_fluka.org; owner-fluka-discuss_at_mi.infn.it
*

*>> Subject: Re: [fluka-discuss]: how to include the divergence in the
*

*>> source.f / photons
*

*>>
*

*>>
*

*>> Dear Alex,
*

*>>
*

*>> Please note following if it can help you.
*

*>>
*

*>> cos 200 mrad=0.98007
*

*>> So we have to randomize direction cosine along beam axis from
*

*>> 0.98007 to 1 (i.e. cos (0 degree) to cos (200 mrad))
*

*>>
*

*>> you can try following code as
*

*>>
*

*>> * Cosines (tx,ty,tz)
*

*>> COSTH = ONEONE-(1-0.98007)*FLRNDM(TTT)
*

*>> IF(COSTH.GT.ONEONE) COSTH=ONEONE
*

*>> IF(COSTH.LT.-ONEONE) COSTH=-ONEONE
*

*>> PHI = TWOPIP*FLRNDM(SSS)
*

*>> UBEAM=SQRT(ONEONE-COSTH*COSTH)*COS(PHI)
*

*>> VBEAM=SQRT(ONEONE-COSTH*COSTH)*SIN(PHI)
*

*>> WBEAM=COSTH
*

*>> TXFLK (NPFLKA) = UBEAM
*

*>> TYFLK (NPFLKA) = VBEAM
*

*>> * TZFLK (NPFLKA) = WBEAM
*

*>> TZFLK (NPFLKA) = SQRT ( ONEONE - TXFLK (NPFLKA)**2
*

*>> & - TYFLK (NPFLKA)**2 )
*

*>>
*

*>> I think it will work.
*

*>>
*

*>> Regards,
*

*>> --
*

*>> SOUGATA RAKSHIT / सौगत रक्षित
*

*>> Scientific Officer(D) / वैज्ञानिक आधिकारीक(D) RSSD / आर एस एस डि
*

*>> Bhabha Atomic Research Centre / भाभा परमानु अनुसंधान केंद्र
*

*>> Mumbai-400 085 / मुंबई-400 085
*

*>> Tel: +91-022-2559 2214/ दूरभाष: +91-022-2559 2214
*

*>>
*

*>>
*

*>> On 2019-07-23 00:21, Ševčik Aleksandras wrote:
*

*>>> Dear experts,
*

*>>>
*

*>>> I am using the attached source.f to sample the designed photon
*

*>>> spectrum, however, I noticed that BEAM WHAT(3) does not respond to
*

*>>> the values if source routine is activated - just as stated in the manual.
*

*>>>
*

*>>> Is there any "simple" way to make my source beam with divergence 200
*

*>>> mrad, or only additional piece of code must be included in the
*

*>>> source.f?
*

*>>>
*

*>>> Looking through old forum archives I found similar code for that
*

*>>> purpose:
*

*>>>
*

*>>> Divbm_Sigma = Divbm*0.2
*

*>>>
*

*>>> CALL FLNRR2 (RGAUSSDIVX,RGAUSSDIVY)
*

*>>>
*

*>>> DIV_VECTORX = RGAUSSDIVX*Divbm_Sigma
*

*>>>
*

*>>> DIV_VECTORY = RGAUSSDIVY*Divbm_Sigma
*

*>>>
*

*>>> TXHLP = TAN( DIV_VECTORX )
*

*>>>
*

*>>> TYHLP = TAN( DIV_VECTORY )
*

*>>>
*

*>>> THELP = SQRT( TXHLP*TXHLP + TYHLP*TYHLP + ONEONE )
*

*>>>
*

*>>> TXFLK (NPFLKA) = TXHLP / THELP
*

*>>>
*

*>>> TYFLK (NPFLKA) = TYHLP / THELP
*

*>>>
*

*>>> TZFLK (NPFLKA) = SQRT ( ONEONE - LTXFLK (NPFLKA)**2 - TYFLK
*

*>>> (NPFLKA)**2 )
*

*>>>
*

*>>> TXFLK (NPFLKA) = UBEAM
*

*>>>
*

*>>> TYFLK (NPFLKA) = VBEAM
*

*>>>
*

*>>> TZFLK (NPFLKA) = WBEAM
*

*>>>
*

*>>> TZFLK (NPFLKA) = SQRT ( ONEONE - TXFLK (NPFLKA)**2
*

*>>>
*

*>>> & - TYFLK (NPFLKA)**2 )
*

*>>>
*

*>>> But simply applying it the code is not compiled, and my lack of
*

*>>> experience with user routines is a big obstacle to move forward from
*

*>>> this point. In addition, my BEAMPOS is defined with WHAT(1)-(WHAT(3)
*

*>>> and also rotation value for WHAT(5) is used. Also the beam is
*

*>>> directed to positive Z direction.
*

*>>>
*

*>>> Could anyone share a similar working source.f file which would
*

*>>> directly use BEAM WhHAT(3) value, or maybe guide me through this
*

*>>> coding problem?
*

*>>>
*

*>>> Both source.f routines - the original and the one I tried to modify
*

*>>> - are attached.
*

*>>>
*

*>>> Thank you very much,
*

*>>>
*

*>>> Rgds
*

*>>>
*

*>>> Alex
*

*>>
*

*>>
*

*>>
*

*>> _____________________________________________________________________
*

*>> _ ____ You can manage unsubscription from this mailing list at
*

*>> https://www.fluka.org/fluka.php?id=acc_info
*

*>
*

*>
*

*>
*

*>_______________________________________________________________________
*

*>___ You can manage unsubscription from this mailing list at
*

*>https://www.fluka.org/fluka.php?id=acc_info
*

*>
*

--

Francesc Salvat Pujol

CERN-EN/STI

CH-1211 Geneva 23

Switzerland

Tel: +41 22 76 64011

Fax: +41 22 76 69474

__________________________________________________________________________

You can manage unsubscription from this mailing list at https://www.fluka.org/fluka.php?id=acc_info

Received on Tue Feb 25 2020 - 23:21:51 CET

Date: Tue, 25 Feb 2020 19:23:34 +0000

Hello again

Unfortunately my lack of skill in programming and deeper understanding prevented me from implementing the divergence into source.f for rectangular beam cases. In cases anyone has used/developed similar or is able to provide step-by-step (literally) solution I would be insanely grateful.

Rgds

Alex

-----Original Message-----

From: Francesc Salvat-Pujol <francesc.salvat.pujol_at_cern.ch>

Sent: Thursday, July 25, 2019 20:29

To: Ševčik Aleksandras <aleksandras.sevcik_at_ktu.edu>

Cc: fluka-discuss_at_fluka.org; owner-fluka-discuss_at_mi.infn.it

Subject: Re: [fluka-discuss]: how to include the divergence in the source.f / photons

Hey Alex,

If I get you right, you want to sample beam particle initial directions with a rectangular divergence (i.e. cosine uniformly sampled in a given

interval) around an arbitrary direction theta_f and phi_f in your modified source.f.

On paper, what one would ordinarily do is first sample the initial direction with the desired rectangular divergence around the z axis and, then, apply a rotation onto the final direction. So you're almost there.

- I believe for a rectangular beam the divergence is to be understood as

a full width. Thus, to sample the (cosine of the) polar angle one

would rather do

CDH = COS ( HLFHLF * DIV )

COSTH = ONEONE - ( ONEONE - CDH ) * FLRNDM ( TTT )

alternatively

COSTH = CDH + ( ONEONE - CDH ) * FLRNDM ( TTT )

where DIV is the divergence in radians, and the azimuthal angle you

sample uniformly from 0 to 2*pi like you already do. By the way, the

DIVBM variable should still have the value you passed in WHAT(3) of

your BEAM card, already converted to radians, but double check in your

particular case.

- You then construct T{X,Y,Z}FLK and ensure it is unit normalized. If

you chose to do smth a la TZFLK(NPFLKA) =SQRT(...) keep an eye on its

sign (for big DIV).

- The resulting T{X,Y,Z}FLK should be fine, but they are still sampled

around the z axis. Let d_i=(tx,ty,tz). To get the "final" direction,

d_f=(tx',ty',tz'), one would apply on d_i a rotation of angle theta_f

around the y axis, R(theta_f,y), and, next, a rotation of angle phi_f

around the z axis, R(phi_f,z).

To be explicit, see the (hopefully not too far off) attached recipe, without any particular effort to be numerically robust (which is a bad idea... and exactly the reason why director cosines are preferrable instead of angles...), but it will get you started. At the very least, make sure that the final T{X,Y,Z}FLK components are normalized (such that the sum of their squares is ONEONE).

Cesc

On Tue, Jul 23 2019, at 18:36 +0000, Ševčik Aleksandras wrote:

--

Francesc Salvat Pujol

CERN-EN/STI

CH-1211 Geneva 23

Switzerland

Tel: +41 22 76 64011

Fax: +41 22 76 69474

__________________________________________________________________________

You can manage unsubscription from this mailing list at https://www.fluka.org/fluka.php?id=acc_info

- application/octet-stream attachment: source_mod_rect.f

*
This archive was generated by hypermail 2.3.0
: Tue Feb 25 2020 - 23:22:05 CET
*